Document xdeMw0pM9N0d5Vq0b3qrEJwBQ

By*E<5ward E. Shea The last two decades have seen major changes in the law relating to product liability--the development of strict lia bility; long-arm jurisdiction; elimina tion or limitation of privity in most states; expansion of pretrial discovery; extension of design and warning duties to include foreseeable misuse; and in terpretation of statutes of limitations to run from injury or from discoverability of injury or its cause rather than the time of sale of the product. Recent court decisions adopting new theories of industry liability and ex panding insurance coverage for latent injuries and diseases are making still more fundamental changes. The result is billions of dollars of potential liabil ity exposure without clear answers where the liability may fall or what should be done to minimize the risk. As a result, there is growing interest in federal legislation to clarify the rights of claimants injured by products and the duties of product manufacturers and their insurers. This article dis cusses the court decisions and recent and pending legislation and gives rea sons why adoption of national uniform product liability legislation is timely. A traditional requisite for proof of product liability based on negligence, breach of warranty, or strict liability is sufficient evidence to identify the manufacturer or seller of the product. In recent years, however, the courts have been presented with a major vol ume of cases involving latent diseases or injuries. Best known are the cases seeking to recover for asbestosis, mesothelioma, and bronchogenic car cinoma alleged to have resulted from exposure to asbestos and the cases seek ing to recover for cancer suffered by daughters alleged to have resulted from their mothers' use of diethylstilbestrol during pregnancy. Under recent in terpretations of statutes of limitations, these claims are not necessarily barred because the products were sold many years ago. Plaintiffs' counsel, however, found they were faced with a further barrier--the major difficulty or impos sibility of proving the identity of the manufacturers and sellers of the asbes tos and DES used. To overcome this barrier, plaintiffs' counsel sought relief from the burden of manufacturer identification and sought to impose joint liability on manu facturers of asbestos and DES as indus tries Two theories were argued initially. concerted action liability and alternate liability Precedent for concerted action liability was found in Hail v. E./. du Pont de Nemours 6- Company, 345 F.Supp. 353 (E.D. N.Y. 1972), in which the court ruled that manufacturers of products alleged to be essentially iden tical (blasting caps) could be held jointly liable for injuries if they er gaged in concerted action by conspir acy or agreement to control the risks of the product, for example, through membership in an industry association. Alternate liability was argued on precedent found in Summers v. Tice. 199 P. 2d 1 (Calif. 1948), in which the court ruled that two hunters who fired guns simultaneously could be held jointly li able for injury, although the plaintiff could not identify which hunter fired the shot causing injury. An article. "DES and a Proposed Theory of Enterprise Liability," pub lished in 46 Fordham Low Review 963 (1978), has been cited frequently in subsequent court decisions. The article pointed out that concerted action liabil ity. as outlined in Hall, might not apply to a product manufactured by a large Changing liability theories have produced a situation that cannot be measured by product manufacturers, sellers, and insurers. Process of 576 American Bar Association Journal ou'mber of widespread manufacturers, only some of whom were named as de fendants. and might require proof of a conspiracy, agreement, or understand ing beyond mere parallel practices. The article noted that the events in Sum mers occurred simultaneously at one location and both hunters potentially responsible were defendants. The arti cle proposed adoption in the pending. DES cases of a hybrid theory called "enterprise liability." Under this proposal the burden of manufacturer identifica tion would be shifted to manufacturers named as defendants on proof of sev eral elements, the most important being proof that a generically similar defec tive product was manufactured by all the defendants, the joined defendants accounted for a high percentage of the defective products on the market at the time of the plaintiffs injury, and there existed an insufficient industry-wide standard of safety as to manufacture of the product. The courts so far have been cautious about adopting industry liability and none has embraced "enterprise liabil ity" as proposed in the law review arti cle. In Lyons v. Premo Pharmaceuticals Labs. Inc- 406 A. 2d 185 (1979). a New Jersey intermediate appellate court ruled that theories of industry liability do not apply when the manufacturer can be identified. In Abel v. Eli Lilly & Company, 289 N.W. 2d 20 (1979), a Michigan intermediate appellate court rejected enterprise liability. In a twoto-one decision, however, it found precedent in Michigan law to relieve the plaintiffs of the burden of manufac turer identification and allow them to proceed on theories of concert of action liability and alternate liability. In Davis v. Yearwood. 612 S.W. 2d 917 119801. the Tennessee Court of Appeals refused to adopt theories of industry liability in a lawsuit against 102 defendants, in cluding 59 manufacturers, resulting* from a fire in a jail that began in a pad ded cell alleged to have contained a va riety of materials including nylon, latex, polyvinyl chloride, and flexible polyurethane foam. Some lower courts have rejected all theories of industry liability in DES cases, and some have stated that adoption of enterprise liabil ity would be proper only by a state su preme court or legislature. In 198Q the Supreme Court of California announced a new theory of several industry liability. This theory was adopted in S/ndell v. Abbott Labo ratories, 607 P. 2d 924, in a DES case after finding lengthy allegations of par allel manufacturer activities insuffi cient to support concerted action liabil ity. finding joint alternate liability un justified because only five of 200 manu facturers were before the court, and re jecting enterprise liability because of the widespread decentralized nature of the industry and the fact that many standards followed in the industry are imposed by government regulation. The court formulated a new theory under which the action could proceed provided that manufacturers having a "substantial percentage" of the market are named as defendants with several liability allocated among them based on their market shares. This theory is now being called '`market share liabil ity." and it was followed in 1961 by a New Jersey court in Ferrigno v. Eli Lilly & Company, 420 A. 2d 1305. New theories abound in product liability The Appellate Division, First De partment, New York in Bichier v. Eli Lilly Br Company, 436 N.Y.S. 2d 625 (1981), upheld jury instructions that concerted action liability could be im posed on a DES manufacturer without proof of an actual conspiracy if the jury found either conscious parallelism or independent actions substantially aid ing or encouraging failure to test the product. If upheldjhis decision would considerably expand joint industry lia bility. A federal district court in Massa chusetts, however, granted partial summary judgment in favor of manu facturers of DES after careful analysis showed that the evidence failed to demonstrate any actual agreement amounting to concert of action, aiding or abetting, or joint venture. Payton v. Abbott Laboratories. 512 F.Supp. 1031, (1981). As in Sindell. the court specif ically rejected efforts to establish con certed action on the basis of alleged conscious parallelism. Plaintiffs' counsel have moved quickly to employ the new theories of Industry liability, and the result has been shock waves as manufacturers find themselves being named in large defendant groups alleged to be "indus tries." The courts that adopted industry lia bility sought to remove a barrier to trial on the merits based on the allegations in the asbestos and DES cases, As a precedent, however, industry liability raises serious difficulties. Among the problems are definition of industries and identification of their members; de termination of the essential identity of products sold for differenl markets and uses or containing multiple compo nents and raw materials; determination of market shares or other means of al locating liability; unavailability of de fense evidence on the merits in older cases; and potential adverse effects on the activities of industry associations in developing product standards and test methods and acting as clearinghouses for product safety data. In addition, there is a problem of duplication of defense cost in cases de fended by multiple manufacturers. Each manufacturer retains its own counsel and wishes to defend its own product Sharing of defense work and cost is possible within limits, but that co-operation tends to reinforce plain tiffs' theory that the defendants com pose an industry group, If defendants responsible for vital portions of the investigation and defense should settle, theJoss of continuity to other defend ants can be a severe blow. Thus, there is considerable pressure on manufactur ers to make cost-of-defense settlements regardless of the merits to avoid years of effort and expense. Product liability insurance is part of the coverage commonly provided in comprehensive general liability policies insuring against liability for bodily injury and property damage. The coverage is accomplished by "products hazard" and "completed op erations" provisions. A general liability policy obligates the insurer with respect to bodily injury or property damage caused by an "oc currence" during the policy period, which is ordinarily one year. Limits of liability are stated per occurrence and in the aggregate. For example, a policy may limit liability of the insurer to $500,000 for each occurrence and $1,000,000 for all occurrences within the policy period, less a deductible amount of $100,000 per occurrence. A fairly typical policy defines an "occur rence" as an accident, including con tinuous or repeated exposure to condi tions, that results in bodily injury or property damage neither expected nor intended from the standpoint of the in sured during the policy period. Some manufacturers carry "excess" policies with "drop down" provisions in the event that liability exceeds the limits of the primary policy, but many do not, Most product accidents (such as ve hicle collisions or pressure vessel ex- % 'B.i.Ul.r- Km May. 1982 Volume 68 577 URL 03890 plosions) occur suddenly. The event al leged as the basis of the manufacturer's liability is clearly an "occurrence" under the policy for the year in which the accident happens. Thus, the cover* age of that policy, including limitations and deductibles, applies. Although the product is alleged to have been defective since manufacture, it has been accepted that an "occur* rence" does not happen until the defect results in bodily injury or property damage -- perhaps a number of years later. This method of coverage'has im portant practical consequences. ^When it issues or renews a policy for a par ticular year, an insurer assumes the risk of sudden occurrences involving prod ucts sold in prior years when the manu facturer may have been self-insured or covered by another insurer. Insurers found product liability cover age increasingly difficult to underwrite as the courts expanded product liabil ity. A "crisis" developed in the middle 1970s. Premiums increased sharply and coverage was restricted or denied en tirely to some manufacturers. This led the Department of Commerce to spon sor a study and propose legislation, to which 1 refer later. As the 1970s progressed, the difficul ties of product liability coverage again were magnified as litigation to recover for nonsudden injuries began to grow. Lawsuits were commenced against manufacturers of pharmaceuticals, chemicals, construction materials, and other products alleging that the prod ucts caused disease or injury either gradually over a period of years or caused harm that remained apparently dormant for years and then manifested itself as disease or injury. These cases claimed liability that far exceeded the customary limits afforded for "occur rences" within the annual policy cov erage, including "excess" coverage, maintained even by major manufactur ers. Because of the nature of the harm claimed, it was not clear when an "oc currence" had taken place. Faced with enormous potential lia bility, manufacturers contended that coverage should be provided under policies issued in prior years. Insurers asserted claims against earlier insurers and against manufacturers with respect to earlier periods of self-insurance seek ing to shift to them at least a propor tionate share of liability for latent dis eases and injuries. Two principal theories have been argued: "manifesta tion" and "exposure." The manifestation theory was adopted by the Supreme Court. New York County, in American Motorists Insurance Company v. E.R. Squibb & Sons, Inc.. 406 N.Y.S. 2d 658 (1978), and the insurer whose policies were in effect during the years when injuries from diethylstilbestrol manifested themselves was held liable. The expo sure theory was adopted by a federal district court in Michigan in Insurance Company of North America v. FortyEight Insulations, Inc., 451 F.Supp. 1230 (1980). The decision was affirmed on appeal, and the Supreme Court de nied certiorari. The result was that in surers providing coverage during the years of worker exposure to asbestos in sulation sold by the defendant manu facturers were held obligated for a proportionate share of the liability to claimants suffering from asbestosis, al though the harm did not manifest itself untir many years Later. The exposure rule also was later applied by the court to claims resulting from mesothelioma and bronchogenic carcinoma. What theory governs the asbestos cases? In subsequent decisions the exposure theory was adopted by the Fifth Circuit in Porter v. American Optical Corpora tion, 641 F. 2d 1128 (1981), and by the U S District Court in the District of Co lumbia in Keene Corporation v. Insur ance Company of North America, 513 F.Supp. 47 (1981). At that point, a po tential trend seemed possible toward the exposure theory, which, the courts observed, tends to allocate liability widely among insurers and reduces the difficulty experienced by manufactur ers, distributors, and retailers of asbes tos and other products in obtaining cur rent coverage. Recent decisions, however, have not followed the potential trend and have opened possible new directions. In Eogle-Picher Industries v, Liberty Mutual Insurance Company, 523 F. Supp. 110 (D. Mass. 1981). the court adopted a manifestation theory. Placing emphasis on analysis of the policy lan guage, Judge Rya Zobel held that injury occurs when exposure produces clini cally evident diseases that can be diag nosed. On appeal of the district court's decision in Keene Corporation, the Court of Appeals for the District of Co lumbia Circuit held that insurers who had policies outstanding during periods of inhalation exposure, expo sure in residence, and maniiestfenen. must each provide full coverage to the insured manufacturer (667 F. 2d 1034 (1981)). With one dissent, the court also held that the coverage must be provided by the insurers without prora tion for periods when the manufacturer was uninsured. The court stated that liability among the insurers should be allocated according to the "other insur ance" provisions of their policies, which contain formulas for contribu tion by equal shares and by limits. The only consistency now found in the courts' decisions is a tendency to decide in favor of theories that provide the greatest insurance coverage on the facts of the case before them. (Keene and American Motorists involved the relatively unusual situation in which the manifestation theory provided greater coverage than the exposure theory.) Resolution may have to await review by the Supreme Court, but its decision on March 8,1982. to decline to review the Keene case may mean a lengthy wait. Faced with growing exposure to lia bility and expense during the product liability crisis, manufacturers and in surers have sought legislative relief. State legislatures have adopted a va riety of reform laws. The most common restriction has been a statute of limita tions requiring that any product liabil ity action be commenced within a rela tively short period (two to six years) after a product defect is discovered or became discoverable, but in no event later than a somewhat longer period (ten to 12 years) after the product was first delivered to its initial purchaser. Other common provisions clarify the availability of defenses based on (1) the state of the art at the time the product was made, (2) modification of the product subsequent to sale, and (3} use of a product beyond its reasonable an ticipated life. The Department of Commerce spon sored an interagency task force on product liability that included repre sentatives of several government agen cies supplemented by an advisory committee of representatives of indus try, labor, law firms, insurers, and con sumer organizations. The task force published its final report in 1977, and the department published a draff model code for adoption by state legislatures. After public comments, the Model Uni form Product Liability Act was published by the department in 44 Federal Register 62714 (1979), together with an analysis of its provisions. A statute con- 578 American Bar Association Journal 1 lJR\_ 038^2 THE TOXICITY OF FOODS OF NATURAL ORIGIN* THE ROYAL SOCIETY OF SOUTH AFRICA SECRETARY: A. V. Hall PRESIDENT: N. Sapcika VICE-PRESIDENTS: W. J. Ltttjeharms S. M. Naudi EDITOR: A. C. Brown TREASURER: J. Staz . by N. Sapeika Medical School, Observatory, Cape Town (With 17 text figures) A food is defined as a substance that is taken to replace physiological waste of tissue, to supply energy and heat, and to build up tissues. Pharmaco-active substances of natural origin, and additives or contaminants of many different kinds introduced by man, occur in a large variety of foods. Many of them are generally not significant, but there arc a few that may on occasion seriously affect a single individual, groups of individuals, and sometimes hundreds of animals or human beings, as for example in the case of poisoning with deadly mushrooms, mycotoxins, and pesticides. Even long-established foods may reveal unexpected dangers, as in the case of cheese, which is dangerous when the consumer is receiving treatment with antidepressant drugs of the type classified as mono amine oxidase inhibitors. It has been known to mankind since the dawn of history that there are natural products which are acutely toxic, but it is only relatively recently that it has been recognized that chronic effects may follow the ingestion of certain materials which do not necessarily produce acute effects. Also, there is scanty information about the possible adverse effects produced by the ingestion of mixtures or combination of foods. Bacterial Food Poisoning There are specific bacteria well known as causes of food poisoning; such infec tions are not considered here. ! Food Allergy 1 H t 5 i ' . ,1 '' " Prices ofearlierpints may be obtainedfrom the Hon. Secretary, Royal Society of South Africa, University of Cape Town, Rondebosch, Cape Town '-,4 '> /1 Vt ! Some persons react abnormally to specific foodstuffs that do not affect the vast majority of people; for example there is the well-recognized sensitivity of some persons to milk or eggs. Almost all allergic reactions, including the most unusual manifesta tions, can be evoked by foods. It is possible also that attacks of migraine may be produced in certain susceptible individuals by tyramine and other amines present in foods such as cheese, fish, beans, milk, dairy products and chocolate, and in alcoholic beverages. Individuals may occasionally show hypersensitivity to drugs present in food, as for example in the case of cow'? milk containing`penicillin used in the treatment of mastitis in the animal. While there is evidence that food components can act as antigens and allergens there is a need for more investigation regarding the antigenicity and allergenicity of bacterial products and of additives and contaminants in food. Presidential Address (1973) delivered to the Royal Society of South Africa, in Cape Town and Pietermaritzburg. I Trim*, rrrt. ,tVvr Afr , * r Pr t Vf-rrh intr URL2 TRANSACTIONS OF THE ROYAL SOCIETY OF SOUTH AFRICA Bioactive Substances Foods may cause untoward effects because of their content of `naturallyoccurring' pharinaco-aclive substances. A review of many foods and the substances present as natural constituents or which they derive from contaminating sources will be considered later in this address. Food additives and contaminants Food may be harmful because of additives or contaminants introduced by man. Food additives arc foreign non-nutritive substances intentionally incorporated to enhance the taste, the structure, or the storage life of food. Food contaminants are harmful inan-made foreign substances such as pesticides and radionuclides that have accidentally become included in food. The requirements of our developed society have necessitated that the foods we cat arc in many instances prepared and preserved by the addition of chemical substances which are regarded as harmless. Some additives have been used for many years and arc generally regarded as safe, but these as well as newer chemicals used in foods need thorough evaluation, regarding their safety, through modern scientific methods of investigation. It is not possible to be completely certain that these addi tives will never produce harmful effects. The use of chemicals has nevertheless become a necessary part of modern food production and processing. The world is short of food, and the shortage grows worse as the population increases. Substances which prolong the shelflife of food or which diminish wastage are very important. It is not the substances that produce severely acute toxic effects but rather those which have a slow or cumulative effect that are probably of greatest practical importance. It seems possible that with increasing recognition and knowledge of chemicals consumed by man and animal that the causes of some diseases may be revealed. In addition to food additives, food contamination also is a subject of great importance. Antibiotics, for example, may come to be present in foods in a number of different ways. There has been some concern about the storage of pesticides in the Fat of mammals, especially the chlorinated hydrocarbon insecticides such as DDT; residues of this persistent insecticide have been found, sometimes hundreds of kilo metres from the nearest spraying operation, in the bodies of animals throughout the world. Such residues have accumulated to levels that are catastrophic for certain animal populations, especially carnivorous birds. Contamination of cow's milk and human milk and of hen's eggs with pesticides, especially with organochlorine residues is also welt recognized. Food contamination has recently come to the fore with the recognition of the production of afiatoxms and other mycotoxins by moulds such as Aspergillus growing in groundnut meal and otlu r food material. Tlic possibility of this type of carcinogen, and other fungal toxins, being toxic to man has caused concern about the me of certain foods, for example the use of groundnut meal as a milk substitute and an econo mical food in certain countries. The relationship of foods, feeds, food additives and THE TOXICITY OF FOODS OF NATURAL ORIGIN OO 3 contaminants and food degradation products to carcinogenesis has been much dis cussed in recent years. In addition to the hepatocarcinogenic aflatoxins present in fungus-contaminated grain and peanuts and in the milk of cows given such contaminated food, other examples of potential carcinogens for man and animals include the following: the carcinogen responsible for hepatoma in rainbow trout; the hepatotoxic carcinogens lutcoskyrin and islanditoxin which are present as a common contaminant of rice and rice products used by Asiatic peoples (these substances are produced by a mould Lutcoskyrin Islanditoxin Penicillium islandicum which infects rice, producing yellow rice or yellowsis, and may be a cause of primary liver carcinoma in Japan); the very potent carcinogen cycasin present as a natural ingredient, not as a contaminant, in unrefined cycad meal will be considered later. The possible significance of N-nitrosamines as causal factors in human carcino genesis is based on the widespread occurrence of nitrosamine precursors in biological CHj CH, N NO Dime thy [nitrosamine systems and the remarkable way in which various nitrosamines can affect different organs in the same species of experimental animal. This is important in the current 4 TRANSACTIONS OF THE ROYAL SOCIETY OF SOUTH AFRICA URL 03894 search for nitrosamines or other agents which may induce oesophageal and other cancers in man. These carcinogens can arise during the storage of fish meal preserved with nitrite. Distilled alcoholic drinks such as Kachasu in Zambia have been found to contain nitrosamines. It has been suggested that the high incidence of oesophageal carcinoma in certain African areas (Zambia; Natal Zulu; Xhosa) may be due to the presence of nitrosamines in certain foods and drink. The presence of radionuclides in foods and the problem of safety levels are regularly studied in many countries. Everyone is now exposed not only to natural radiation but to some additional radiation such as that from strontium-90 present as a contaminant in milk and other foods. Ordinary Foods Harmful in Certain Individuals 1 Foods that are well known and widely used without causing harmful effects may cause disturbances in certain individuals who have basic disorders that may be aggravated or precipitated by an oHending food substance. The following are examples. () Allergic disorders, as already mentioned, such as urticaria, asthma and possibly migraine may result from the ingestion of certain foods by predisposed subjects. () Coeliac disease is a gluten-induced enteropathy and the basis of treatment is a gluten-free diet plus certain supplements. The condition occurs in young children in whom there is interference with intestinal absorption of neutral fat (steatorrhoea), carbohydrate, vitamin D and other ft>od factors. Gluten is a mixture of the proteins gliadin and glutelin present, foe example, in wheat flour and rye flour and many breakfast cereals, also in some proprietary preparations such as Ovahine, The normal intestinal mucosa is capable of hydrolysing polypeptides of gliadin, those with molecular weight under 2 000, whereas the coeliac mucosa is unable to do so, thus permitting the polypeptides (in some unexplained manner) to damage the mucous membrane of the proximal part of the intestine in these patients. Many food manufacturers, for example in Britain, label packages of food pre pared for use by patients with coeliac disease with a symbol (an ear of wheat with a line through it) to indicate the food is gluten-free. (c) Patients who have hepatic cirrhosis with liver failure should not be given protein while the special treatment for this complication is being given. Other THE TOXICITY OF FOODS OF NATURAL ORIGIN 5 nitrogen-containing substances also should be excluded otherwise hepatic encephalo pathy is likely to be produced. (d) Maple syrup urine disease, so called from the characteristic odour of the urine, is an inborn error of the metabolism of branched-chain amino acids; the withdrawal of valine, leucine, isoleucine and methionine from the diet has produced some benefit. () Favism is a haemolytic disorder often associated with haemoglobinuria, which is induced by the ingestion of the fava bean or by contact with the pollen of the plant. The red blood cells in these palirnts are deficient in the enzyme glucose-6phosphate dehydrogenase (G-6PD) necessary for the continued integrity of these cells. The cells are also sensitive to a number of drugs, and to vitamin-K analogues. (/) Phenylketonuria is an inborn metabolic error of amino acid metabolism. The basic defect is the absence of the hepatic enzyme, phenylalanine hydroxylase, which normally converts into tyrosine the dietary phenylalanine that is not used up in the synthesis of protein. Mental deficiency will occur unless a diet sufficiently low in phenylalanine content such as casein hydrolysate is started early. The brains of these infants are not damaged at birth and they will develop normally if a diet low in phenylalanine content is started early. There is a simple test for phenylpyruvic acid (one of the degradation products of phenylalanine) in the urine. Most states of America require a PKU test at birth. (g) Galactosaemia is a rare inborn error of metabolism in which the body is unable to convert galactose to glucose. The accumulation of galactose-1-phosphate causes damage to the brain, liver, lens of the eye and the kidneys. In this case the manifestations of the disease may be prevented or reversed in many of the affected infants if the diagnosis is made early and a lactose-free diet given, for example, the elimination of milk products from the diet. (A) Porphyria genetica (porphyria variegata), an important latent hereditary disease in South Africa, may be produced in certain individuals by alcohol and by drugs such as barbiturates and sulphonamides. Hypermiaminosis When pure potent vitamins, especially those of the fat-soluble type, are given in excess, they may act as poisons. Vitamin A intoxication is recognized as two forms. There is an acute type due to the ingestion of a single large dose. The livers of the polar bear and certain other animals are toxic in this way because they are extremely rich in vitamin A, and potent vitamin A preparations have caused serious poisoning when given in the treatment of young infants and children. Chronic poisoning may result from the daily ingestion of moderately toxic amounts over a considerable period, for example in the treatment of certain diseases. The disorder may occur in food faddists. Vitamin D administered to children in potent preparations has also caused serious ill-health. Vitamin K analogues in large and repeated doses should not be given parentcrally to newborn infants as serious disturbances have been produced. Also, vitamin K 6 TRANSACTIONS OF THE ROVAL SOCIETY OF SOUTH AFRICA URL 03895 taken by the mother late in pregnancy can adversely affect the foetus. Nicotinic acid ingested in large amounts may produce pharmacological and harmful effects. Poisoning has occurred from the consumption of meat in which nicoiinic acid was an additive to prevent darkening of the meat. Taken in large doses for months this vitamin can aggravate diabetes mellitus. Classification of Pharmaco-active Foods Pharmacologically active substances that may be present in foods may be con sidered under the following headings; Pharmaco-active substances in foods of plant origin Pharmaco-active substances in foods of land animal origin Pharmaco-active substances in foods of marine animal origin Food additives and food contaminants introduced by man. The limited time available for this address does not permit detailed consideration of the many active substances naturally present in foods of plant or animal origin. Attention will therefore be given to some of the more important examples in the above-mentioned groups and particularly with regard to those important for man. Foods of Plant Origin Many vegetable materials used as food contain potentially harmful substances such as the following: Amines Dopamine, noradrenaline, adrenaline, 5-hydroxytryptamine and tyramine are present in certain fruits and vegetables used as human food, for example in bananas and plantains, tomatoes, avocado and pineapples. Certain amines are also found in OH OH OH 1 1 THE TOXICITY OF FOODS OF NATURAL ORIGIN AMOUNT EXPRESSED AS MC/lOO OM FRESH FRUIT Bananas (peel) Bananas (pulp) Plantain (pulp) Pineapples (ripe) .. Avocado .. Tomato Plum, red .. Plpm, blue Eggplant .. Potato Orange Cyanogens 5-Hydroxytryptamine 5~5 2,8 .. a-10 .. 2 t 1,2 .. 1 .. 0 . 0,2 .. 0 0 Tyramine 6,5 0,7 _ a,3 0,4 0,6 0,3 0,1 1,0 Dopamine 70 o.a 5 0 0 0 0 0 Noradrenaline 12,2 0,2 0,2 O O + O 0,2 + Many plant foodstuffs contain cyanogenic glycosides. Such compounds have been found, for example, in maize, sorghum millet (kaffircorn), lima bean, ktdaey bean (haricot), sweet potato, manioc (cassava), almond, and in many other plant materials such as the seeds of apples, apricots, cherries, pears, plums, prunes and other fruit. Bruising of such foodstuffs causes the interaction ofan enzyme and glycoside and the Telease of hydrocyanic acid. Poisoning is rarely produced by most of these food stuffs because the cyanogens are present in small amount or in parts of the food (seeds, pips) which are not eaten. However, poisoning has not infrequently occurred from the ingestion of manioc and lima beans. The seed of sorghum, the common sorghum or great millet known in South Africa as kaffircorn used in the preparation of Bantu beer, is edible, but the plant is toxic to stock due to the production of hydrocyanic acid from the cyanogenic glyco sides. Hepalotoxins A number of plant and fungal poisons can produce liver damage in animals and man. As already mentioned, Aspergillus aflatoxins and other inycotoxins may be present as contaminants in certain foods. The Amanita cyclopeptides in mushrooms Norodrcnaline Tyramine Adrenaline cheese especially cheddar, Camembert, and Stilton cheese, in a number of other foods, and in certain wines. It has been demonstrated that arylalkylamines are more widespread in the plant kingdom than had previously been realized. The banana contains the largest amounts of these substances; banana pceis contain enormous quan tities. PholJoldin Phollom Phallacidin RI . R, R, OH CH, H CH OH CH(CH^, R. CH, CH, COOH CH,---------CH------------CH, CH, N CH, CH, CH, Pyrrolitidinc H TRANSACTIONS OP THE ROYAL SOCIETY OF SOUTH AFRICA URL 03896 such as the Death Caps, Amanita phalloides and Amanita capensis, and the Sentcio alka- ] loids (pyrrolizidines) are severe liver poisons. Many members of the Senecio genus of j plants arc eaten as spinach (imfino) or arc used for medicinal purposes by the Bantu t in some areas of South Africa. Alcoholic beverages are also important causes of j pathological changes in the liver and other organs in man. j Cycad seeds or kernels, also called cycad nuts, are used as food especially in the Pacific area. Cycad flour prepared by soaking the seeds in water removes carcino genic substances, but ground unwashed cycad seeds are extremely toxic; malignant tumours have been induced in the liver and kidneys of experimental animals and resemble those produced by nitrosamincs. The toxic substances are azoxy glycosides such as cycasin and macrozamin in which the aglycone, melhylazoxymethanol, is the o t CH3N=NCH2--O -- glucose Cycasin O t CH3--N-- NCHa--O -- primverose ! j i j Macromazin O t ch3n = nch2oh ; i Methylazomethano! carcinogenic moiety; cycasin is decomposed to the aglycone by bacterial enzymatic action.' Cycad meal may also be contaminated with fungi that produce mycotoxins like the afiatoxins so that carcinogens both from the meal and the fungi may be the cause of liver cancer and other malignancies in people who eat this food, ! < j j [ THE TOXICITY OF FOODS OF NATURAL ORIGIN 9 Nitrates A number of plants growing under certain soil conditions have been found to contain dangerous amounts of nitrate (saltpetre). Some have caused serious losses in pigs and other animals from nitrate poisoning, The nitrate content of spinach has on occasion caused nitrate poisoning in infants who had eaten spinach. Nitrates are reduced to nitrite in the gastrointestinal tract and these convert haemoglobin to methaemoglobin. Oxalates Certain plants, for example rhubarb, spinach, celery, and green plantain, con tain oxalic acid usually in the form of acid potassium oxalate; they have occasionally given rise to fatal poisoning when used in cooking or for medicinal purposes. Normally 80 to go per cent of oxalate in the diet is not absorbed but is eliminated in the faeces as insoluble calcium salts. The oxalic acid content (mg per 100 gm fresh edible uncooked portion) of certain foods is shown in the following list. Rhubarb Spinach Celery (stalk) Almond Plaintain (green) Plaintain (ripe) Sorrel. . Cashew nut .. Parsley Beetroot Currants Prunes Beans .. Soyabean ioo -900 320 -900 620 407 5,7>5-5a4 2,2 100 -500 3*8*4 too -200 100 --200 25 -100 25 -IOO 23 -IOO 77 Goitrogens A number of plants contain an antithyroid (goitrogenic) principle which has been identified as /-5-vinyl-thiooxazolidone, known as goitrin which is a cyclized thiocyanate. It has been found in high concentration in extracts from the seeds of Metals Vegetables and animal foods normally contain minute traces of many elements and at times metals such as copper, arsenic, and iron, usually in small amounts. Selenium poisoning occurs in certain areas in the United States and in other countries where fodder and grain are rich in selenium due to the unusual amount in the soil. In South Africa numerous species of plants in the Karoo area have been found to contain dangerous amounts of selenium, and correlation has been claimed between selenium content of the vegetation and the incidence of gceldikkop and enzootic icterus, which are important ovine diseases in this country. Thiourea Goitrin most Brassitae, for example cabbage seed and the edible parts of turnip. However, the goitrogens are present in such low concentrations, except in seeds, that it is doubtful whether human beings would eat enough for goitres to be produced; cooking destroys the goitrogenic activity. Cabbage also contains a thioglucoside, glucobrassicin, which n--oso; Ogr CH.CN + S URL 03897 -> HtSSOO," * glucose Ogr CH.OH+ SC N" flavouring agent in root beer. Its use as a flavouring agent in food has long been pro hibited.) Onion and other species of AUium, for example garlic, contain irritant volatile oils and other substances. Onion has been incriminated as a goitrogen. CH,0 Glucobrassicin 3 --hydroxymethyl Indole on enzymatic cleavage yields thiocyanate and other substances (indole acetonitrile and 3-hydroxymethyl indole); thiocyanate ion is goitrogenic. Ijilhyrus poisons The seeds of the chick pea (Lathyrus salivas L.) are allied to beans and peas. They grow wild in certain countries and are used as food and fodder. Lathyrus peas have caused disease (lathyrism) in thousands of people, for instance in 1'ranee, Spain, Italy and North Africa, also in concentration camps and prison-of-war camps during World War II. Two syndromes have been attributed to the consumption of these small peas: (a) ostcolathyrism is one type, in which bone deformities and some other anomalies are produced. Certain nitriles are regarded as causing these lesions in stock HOOCCH C H,C H,CON HC H,CH,CN NH, y-glut amyl---- amino nitrile CH,CH=CH, H,C--O Myristicin Safrole Some other substances Certain plants contain small amounts of oestrogens, enzymes, and lectins (phyto agglutinins) Some contain unusual amino acids, e.g. the hypoglycins present in CH,: C-CH. CH, . CH.COOH \/ I CH, NH, Hypoglycln A CH' C -- CH. CH,. CH. COOH \/ I CH. HN-CO. CH..CH.. CH COOH in U.S.A. by inhibiting collagen maturation; (b) neurolathyrism is another disorder, in which pain, and weakness, and spasticity of the leg muscles, occur. Neurotoxic factors such as diaminobutyric acid and bcta-cyanoalanine that may be implicated have been isolated from various species of lathyrus peas. NH.CH.CH.CHCOOH I NH, Diaminobutyric acid N = C-CH,-CHCOOH I NHt /3~cyanoalanine NH, Hypoglycln B Ackce fruit common in the West Indies and South America; they produce hypoglycaemia and fatty changes in the liver of certain laboratory animals that resemble the changes caused by phalloidin and a-amanitin. Djenkoiic acid present in the seeds of a plant eaten with relish as a food in certain parts of Indonesia is an amino acid that may produce oliguria, albuminuria and renal colic with blood, casts and crystals in the urine; the condition may resemble an abdominal emergency. Spices S------- CH,--------- CHNH,------- COOH Flavouring agents may occur naturally in food, but many foods require some added substance to make them more appetizing. Among these are the essential oils S-------CH,--------- CHNH,-------- COOH and carminatives used as spices. These arc complex mixtures obtained from various i plant sources. They are mild irritants, but they can produce toxic effects. Nutmeg contains myristicin, which may have psychotropic properties, and a Djenkoiic acid number of other active principles. Myristicin is closely related in chemical structure to Sanguinarine (in Argemone mexicana) which is one of a number of substances safrole, which is hepatotoxic and even hcpatocarcinogenic in rats. (Safrole is a com suggested as a causal factor in oesophageal carcinoma, and ergot alkaloids which can ponent of several essential oils. It was once extensively used in the U.S.A. as a cause ergotism (St. Anthony's Fire), have occasionally been contaminants in bread. 12 TRANSACTIONS OF THfc RUVAL SOClfcl* Of SrtJUlM AfKlt-A o=c HaC-------- CH, P 0C CH. NH o=i- nn /c= yc\ , HR URL 03898 H Ergotamine, R=-CHS. R=-CHa-0 Honey This has on rare occasions been the cause of poisoning when it contained toxic material collected by the bee from plants such as rhododendrons, Nerium oleander, datura, and some other poisonous plants. Honey and honeycombs may contain fumigants such as calcium cyanide, methyl bromide and paradichlorobcnzene. Foods of Animal Orioin Pharmacological activity of food of land animal origin may be considered from two points of view. Firstly, the intrinsic composition of the food may temporarily or permanently make it unsuitable and even unsafe for certain individuals, Secondly, the presence of additives and especially of contaminants may render it potentially or actually harmful for all persons who consume such food. Meat The majority of mankind has a natural appetite for meat, and the flesh of more than one hundred different species of animal is regularly eaten by man. Reference has already been made to the fact that protein in the diet, and other nitrogen-containing substances including a number of drugs, produce an abnormally increased concentration of ammonium ions in the blood, which must be avoided in patients who have cirrhosis of the liver. Otherwise, neurological symptoms may develop that progress to stupor, coma (and death); the condition is referred to as hepatic coma or portacaval encephalopathy and was previously known as meat poisoning or ptomaine poisoning. tn certain sensitive persons the ingestion of cooked crab meat has produced mild to severe anaphylactic shock. Meat extracts, for example Bovril, contain amines that can enter the blood stream when certain drugs, the monoamine oxidase inhibitors are taken by patients suffering from certain types of depression. A severe hypertensive reaction may arise, such as may occur when tyraminc-containing cheese and certain other foods are ingested concurrently with these drugs. THE TOXICITY OF FOODS OF NATURAL ORIOIN *3 Additives and contaminants Nitrate or nitrites added to meat and meat products as preservatives have been the cause of poisoning (methaemoglobinaemia). Sodium nicotinate or nicotinic acid used for the preservation of the red colour in meat, to prevent its darkening, has caused flushing and itching of the face and neck, nausea, abdominal cramps and sweating. Cyclamates (sweetening agents) have been used in meats such as bacon and ham. In large doses they are carcinogenic in animals; the metabolite cyclohexylamine is an inducer of bladder cancer. Meat grilled on skewers made from the wood of the ornamental shrub Nerium oleander L. (Ceylon rose) has caused death. All parts of this plant are poisonous due to the presence of digitalis principles. Certain plants of the genus Cotyledon contain neurotoxic and cardiotoxic principles. Poisoning has occurred in dogs that have eaten the meat, blood, or entrails ofpoisoned animals. Human beings, for example Bushmen, have developed symptoms after eating raw or undercooked meat or the blood of such animal. The meat of the quail, a bird related to the partridge and grouse, has on some occasions been poisonous. This has been attributed to the ingestion by the birds of toxic plants such as hemlock which does not affect them but which renders their meat toxic. During their travels in the desert the Hebrews ate manna and quail. The quail `fell in great profusion around the camp of the Israelites' and the numerous deaths attributed to the eating of these birds is mentioned in the Old Testament (Numbers 11). In recent years some peopic on the island of Lesbos have developed an acute myogtobinuria shortly after eating quail, from sensitivity to some substance present ih the quail. Muscular fatigue is regarded as important since the disorder rarely occurs in persons at rest. This observation has been suggested as an explanation for the deaths of the Hebrews. However, myoglobinuria also occurs occasionally in normal subjects, for example in military trainees after severe exertion, also in Haff disease (in Eastern Europe) possibly here due to an unidentified toxic substance present in contaminated fish, and in a variety of disorders of muscle. The widespread use ofpesticides and dissemination ofradio-active materials over large areas has led to contamination of soil, water and plant and animal material. Meat and milk from cattle that have eaten this type of contaminated herbage present a danger to man that may become serious. Broiled meat (steaks) and barbecued ribs, and smoked fish, contain complex polynuclear (polycyclic) hydrocarbons, some of which are carcinogens; these, how ever, are present in small quantities and are not regarded as dangerous to the con sumers of such food. It has been demonstrated that polynuclear hydrocarbons are present in higher concentrations in steaks and barbecued ribs prepared in a restaurant, but the amounts present in these broiled foods are small and do not necessarily con stitute a danger to the consumers of such foods. In Iceland it is customary in rural areas to preserve mutton and fish by a heavy smoking process; these foods then contain among other compounds the carcinogenic 4 TRANSACTIONS OF THE ROYAL SOCIETY OF SOUTH AFRICA URL t substance 3,4-benzpyrene. This is an example of contamination of a food by simple traditional processes. Gastric cancer is common among the Iceland population. \ 3,4- benzpyrene Liver Liver is a useful food containing naturally a variety of substances including certain vitamins and haematinic substances. However, there are circumstances where liver can be harmful and dangerous to life. The livers of polar bears and certain other animals such as the Arctic fox, seal, whale and shark are extremely rich in vitamin A and acute intoxication (hypervitaminosis A) has occurred in man, for example in Arctic explorers and in their dogs, from the consumption of this material. VITAMIN A (lU PE* I OO GM WEIGHT, FRESH) Cod>liver oil .. Halibut-liver oil Shark-liver oil Polar bear liver Seal liver Tunny (tuna) Tunny (tinned) Sardines Sardines (tinned) Herring Liver (sheep and ox) Butter.. 200 000 4-6 million ( + ) 3 million ( + ) 1,8 million 1,3 million 0,8-8 million 70-200 4 800-54 000 136-290 9 000 4 0OO-45 000 2400- 4000 Chicken liver eaten by patients taking the so-called monoamine oxidase inhibitor drugs for the treatment of depression may induce hypertension. As already mentioned, this may occur when certain other foods that contain amines are taken with these drugs. i1 Milk Milk has abundantly been used in the management of gastric and duodenal ulcer. When ingested in large doses for months with much soluble absorbable alkali such as sodium bicarbonate, hypercalcemia, inelastic calcification in the arteries and kidneys (ncphroculcinosis) and other changes have been produced, a condition known as Burnett's syndrome. THE TOXICITY OF FOODS OF NATURAL ORIGIN 5 Cow's milk may contain pesticides, and pesticide residues may be found in human milk. Also antibiotics, radionuclides, fungal toxins and active principles from certain plants may be present in cow's milk. Penicillin in milk has caused skin and allergic reactions in persons sensitive to this antibiotic. Streptomycin and possibly chloramphenicol may also be incriminated in this respect. Milk sickness (trembles; alkali disease) is a disorder of cattle in the United States, caused by eating white snake-root or raylcss goldenrod. The poison in the milk (or meat) of animals has also caused illness in man. Several members of the Lincoln family, including the mother of Abraham Lincoln, and many early settlers in the United States died of this `milk sickness'. The leaves and stem of the plants mentioned contain `tremetol', which is a mixture of several compounds. The poison is not destroyed by pasteurization of contaminated milk or milk products. Goitrogenic substances may be present in cow's milk, following the ingestion of cabbage, turnips, and certain other plants. Goitrogens (such as goitrin) in cow's milk have been suspected as the cause of goitres in Tasmania and Australia. Death has occurred in an infant who drank milk from a cow that had eaten the foliage of Nerium oleander L. Milk from cows fed very toxic aflatoxin-containing groundnut meal can induce jiver lesions in ducklings. Cheese Various cheeses, for example chcddar, Camembert, and Stilton, have been found to contain tyramine. There may be enough amine present to produce a marked rise in blood pressure and other cardiovascular changes in patients receiving drugs that are mono-amine oxidase inhibitors. Sudden severe headache lasting for twenty minutes to several hours is associated with a marked rise in blood pressure. Death may occur from intracranial haemorrhage. An attack has followed the ingestion of as little as one ounce (27,5 gm) of cheddar cheese. The tyramine present in the cheese enters the general circulation and releases noradrenaline from local stores in nerve endings. It has been suggested that in certain susceptible individuals attacks of migraine are induced by tyramine and other amines present in cheese, milk, dairy products and certain other foods. Moulds may be present in cheese, and many produce toxic substances from which the danger cannot on present knowledge be completely excluded. Radionuclides may be present in cheese. Atherosclerosis The possibility that certain dietary constituents such as animal fats, sugar and alcohol may be causal factors in the production of atherosclerotic changes has been much investigated for many years. More proof is still required that dietary factors and dietary measures influence atherogencsis in man. Other hypotheses have also been put forward regarding atherogencsis, for example that there is a deficiency of |6 TRANSACTIONS OF THE ROYAL SOCIETY OF SOUTH AFRICA essential fatty acids, that prostaglandins may piay an important role in the effect of dietary fats. Drinking water has also been incriminated; some association has been found between softness of drinking water and increased death-rates due to cardio vascular diseases although the foods consumed may be more important in determining the ultimate effects than the type of drinking water. Foods of Marine Animal Orioin There are marine animals which regularly produce or which temporarily contain active principles that may cause poisoning when they are ingested by man. Shellfish Poisonino I Mussels may occasionally cause illness, and even death. The paralytic type of poisoning is most dangerous, caused by a potent toxin in the shellfish, which is known ' as clam toxin, mussel toxin, and saxitoxin (from its presence in the Alaskan butterclam Saxidomus giganteus). Paralytic shellfish poisoning occurs when certain molluscs (mussels, clams, scallops) that have ingested toxic dinofiagellates are eaten by man. Under certain conditions `blooms' of such toxic plankton may cause the phenomenon known as red water or red tide, visible in the daytime and luminescent at night; this heralds an outbreak of shellfish poisoning. Toxic shellfish have been found sporadically in many parts of the world. Certain coastal regions of South Africa, for example the False Bay and St. Helena Bay areas, have been affected in recent years. Poisonous mussels are indistinguishable from normal ones. They are themselves unaffected by the toxin which can only be detected by animal toxicity tests; an easy, quantitative chemical test for the poison is desirable. 'Hie usual methods of cooking do not reliably remove the poison. The clinical features of poisoning have been described by many authors. The pathognomonic symptoms arise within thirty minutes and include a tingling or burning sensation on the lips, gums, tongue and face, spreading to other parts of the body; this paraesthesia changes to numbness. In severe cases ataxia and interference with speech and swallowing occur. In the terminal stages there is motor weakness, and paralysis becomes progressively more severe. Death occurs from respiratory paralysis usually within twelve hours. Treatment is symptomatic as there is no specific antidote. Paralytic shellfish toxin is one of the most lethal biological substances known. Man is very sensitive to it; the minimal lethal oral dose of purified toxin may be between i to 4 mg. Poisonous Fishes Approximately five hundred species of marine fish are known to be toxic, and may be poisonous when eaten by man. Some are poisonous at all times, others only at certain periods or only in certain areas. THE TOXICITY OF FOODS OF NATURAL ORIGIN 7 In conclusion it may be said that every meal very likely contains deliberate or accidental addition of many possibly injurious non-nutritive compounds such as additives and contaminants as well as pharrnaco-active compounds in foods obtained from plants and from land and marine animals, and harmful synthetic substances in water, soft drinks and alcoholic beverages. This review has been concerned mainly with naturally-occurring substances in foods that may cause mild but occasionally severe and sometimes fatal reactions in man or animals eating such foods. Examples of toxic principles in certain natural foods Ofplant origin Ackee fruit Almond Mananas, and tome other fruits Beans, broad Beans, soya B^ans, Lima Brassica seeds, and some other vegetables Cycad seeds Fungi Jcnghol seed Lathyrus seeds (chick pea) Legumes Manioc (cassava) Millet (Kaffircorn) Nutmeg Potato (green) Rhubarb Ofanimal origin Hypoglycins (unusual amino acids) Cyanogenic glycoside 5-Hydroxytryptamine, noradrenaline, adrenaline Dopa, dopamine Trypsin inhibitor, haemagglutinins, oeatrogens Cyanogenic glycoside Thiooxazolidone, thiocyanates Azoxy glycosides which yield methylazoxymethanol Various active principles Djenkolie acid (amino acid) Beta ammopropionitriie Haemagglutinins Cyanogenic glycoside Cyanogenic glycoside Mymticin, and other active agents Solanine Oxalate Cheese Fish Liver, polar bear and sonic others Meat, broiled Meal extracts Mussels Quail Tyramine Various toxic principles Vitamin A in large amount Hydrocarbons (carcinogenic) Amines (adrenergic) Saxitoxin Hemlock (?) References National Academy of Sciences; National Research Council, Washington, D.C. 1966. Toxicants Occurring Naturally in Foods. Sapeika, N. 1969. Food Pharmacology. Springfield, Illinois, U S.A.; Charles C. Thomas. Liener, I. E. 1969. Toxic Constituents of Plant Foodstuffs. New York and London; Academic Press. Liener, I. E. 1973. Toxic Constituents nf Animal Foodstuffs. New York and London; Academic Press. LETTERS Cancer and Diet Bruce N. Ames' remarkable summary (23 Sept., p. 1256) of the evidence that cancer and cardiovascular and other de generative diseases are of metabolic ori gin comes as welcome support for the hiiherto-little-noticed contentions by Totter (/). Handler |2), and Fridovich (j) that oxygen radicals may be an impor tant proximate cause of cancer. The thrust of Ames' article would seem to be that cancer is essentially a natural aging process. No matter what we eat, the huge flood of oxygen radicals produced in many metabolic processes overwhelms all but the most heavy ex ternal carcinogens, such as tobacco in heavy smokers. To be sure, anticarcinogenic substances are of benefit, but to choose a noncarcinogenic diet would probably be equivalent to starving to death. The implications of Ames' findings are broad and fall into three categories. First, our preoccupation with small efflu ents of carcinogens resulting from vari ous industrial processes represents a se rious misdirection of resources. This was revealed by Totter in 1980 (/). when he showed that overall cancer mortality in 19 countries, when corrected for com pleting risks, was not correlated with degree of industrialization, as measured by per capita energy use. Second, the Delaney amendment, which seeks to eliminate the last trace of artificial carcinogen in food, seems to be targeting a tiny part of the carcinogenic burden and ignoring the major carcino gen. the ubiquitous oxygen radical. In short, the Delaney amendment may be based on wrong science and be wrong policy. Finally, if we concede that cancer, like death itself, is "natural,'' then our pri mary focus in cancer research ought to shift far more toward earty detection and extirpation of tumors. This, I believe, would require rethinking of the National Cancer Institute's underlying strategy, which at present seems to be dominated by the belief that cancer, unlike death itself, is a preventable disease. Alvin M. Weinberg Institute for Energy Analysis. Oak Ridge Associated Universities, P.O. Bo.x II7. Oak Ridge. Tennessee 37830 References and Nirtes l. J R. Toller. Prnc. Suit. Acuil. Sci. U.S.A. T7, J7M i IW1>. 2- P. Handler, dedication address presented at the Northwestern University Cancer Center. Ev anston. III.. IK May 1979. 3. I. Fridnvich. Stie/ur 201, K75 (I97X): subse quent personal communication. r On page 1260 of his article "Dietary carcinogens and anticarcinogens'' Ames refers to "dietary selenium (usually sele nite)." The term "selenite" could cause confusion because geologists know the mineral selenite, a variety of gypsum (CaSOa 2H:0) containing no selenium. Ames must be referring to a compound of selenium when he uses the term "sele nite." I doubt that the mineral selenite would inhibit tumor or counter the oxi dative toxicity of mercuric salts. Dale E. Ingmanson Department of Natural Science, College of Sciences, San Diego State University. San Diego, California 92182 Perhaps the most provocative thing in Ames' provocative article is his placing of the figs in the family Umbelliferae. It is often said chat disciplinary crossovers can reinvigorate static fields by bringing new insights unfettered by conventional wisdom. It remains to be seen if Ames' dietary pre- and proscriptions have as much impact on American life-styles as his creative taxonomy is likely to have on the family Moraceae. Arthur M. Shapiro Department of Zoology, University of California, Davis 95616 In his article "Dietary carcinogens and anticarcinogens," Ames states (p. 1258) that "high dietary fat is a promoter and a presumptive carcinogen," citing among his references the recent National Acad emy of Sciences (NAS) report Diet, Nu trition, and Cancer (/). The NAS report does state that "most of the data suggest that dietary fat has promoting activity"; however, this report does not refer to fat as "a presumptive carcinogen." Rather, the report notes that "there is not enough evidence to warrant the com plete exclusion of an effect on initia tion." Ames writes that "the amount of in gested oxidized fat may be appreciable." Neither of the references he cites (2, J), however, provides direct evidence in support of this statement. Shorland et at. (2) demonstrate that vitamin E supple mentation to calves retarded lipid oxida tion of some muscle tissue but not others during frozen storage. No estimates are provided regarding how much oxidized fat humans typically ingest. A perusal of Autoxidation in Food and Biological Systems, edited by Simic and Karel (J), reveals the same facts. There are no direct estimates of the amounts of oxi dized fat ingested by humans, and it is well recognized that the unpalatable na ture of rancid fats precludes (heir inges tion in significant quantities. We strongly disagree with Philip Abelson's assertion that "the colon and digestive tract are exposed to many fatderived carcinogens" (Editorial, 23 Sept., p. 1249). Unsaturated fatty acids in dietary fats are subject to chemical reactions (oxidation, polymerization, hy drolysis) that can occur to a limited extent during deep-fat frying. The extent of these reactions, however, depends largely on frying conditions, principally the temperature, aeration, and duration. Many of the studies used to support the implication that oxidation that can occur during cooking "formfsj mutagens, pro moters. and carcinogens" were per formed under exaggerated conditions that are unrealistic and not indicative of actual conditions. It is the usual practice of restaurants to discard frying fat when prolonged frying causes excessive foam ing of the hot fat or when undesirable flavor or dark color develop. This being the case, Abelson's statements that "rancid fats are possible causative agents of colon and breast cancer in humans" and that "rancid fats should not be part of the diet" are unnecessarily alarming to prudent users of heated fat or other fat-containing products. in support of the safety of fats heated under more realistic conditions, a 2-year animal feeding study by Nolen et al. (4) showed that animals consuming used frying fats as the sole source of fat in the diet throughout their life-span thrived as well as control animats consuming the same fat that had not been subject to frying conditions. Furthermore, if "the colon and digestive tract are [truly being] exposed to many fat-derived carcino gens." we should be seeing increasing colon and breast cancer mortality in the United States as a result of the marked increases in vegetable oil (much of it highly unsaturated) consumption since the early 1900`s (5). In fact, however, data from the American Cancer Society indicate that age-adjusted mortality rates for both colon and breast cancer have remained essentially unchanged since 1940(6). Ames notes that "fs]ome fatty acids, such as C;; i and certain truns fatty acids, appear to cause peroxisomal pro liferation because they are poorly oxi dized in mitochondria and are preferen tially oxidized in the peroxisomes.'1 Data are not accumulating, however, to substantiate such a theory. Citing the paper of Enig et al. i7). which has been criticized i8). Ames states that "Americans consume about 12 g of irons fatty acids a day and a similar amount of unnatural cis isomers. . . . mainly from hydrogenated vegeta ble fats." We believe these estimates of d r--JJ ac(oOo < URL 03902 consumption are excessive and are not supported by reliable data. A more rea sonable estimate of consumption of "un usual" cis and tram positional isomers has been suggested by Hmken to be around 9 grams per day (not 24 grams per day, as suggested by Enig et al.) or about 6 to 8 percent of total fatty acid intake (9). When one considers that the fatty acid composition of adipose tissue re* fleets that of the diet and that a range of from 2.0 to 5.8 percent trans fatty acids has been reported in human adipose tissue (10), an adult male consumihg his recommended dietary allowance of 2700 calories per day (//) of a diet providing 40 percent of the calories as fat would ingest around 2.4 to 7 grams of trans fatty acids per day. Contrary to the disputed hypotheses of Enig et al. (7), there are no reliable data relating trans fatty acids to tumor development. A study by Brown (12) not cited by Ames indicated no unusual inci dence of tumors in mice treated with dimethylhydrazine (or with saline) and then fed a diet high in tram fatty acids for 17 months. Finally. Ames mentions the disagree ment between the NAS report (/) and a critique of this report (13) by the Council for Agricultural Science and Technology (CAST) on the appropriateness of rec ommending reduced fat consumption to the American public. In doing this, he indirectly quotes the CAST report as saying that, "until we know more . . . a- bout which types offats are dangerous, it is premature to recommend dietary changes" (13). However, the CAST re port does not state that certain "types of fats are dangerous." We believe there are insufficient reliable data to justify the suggestion that certain fats in the current American diet represent a substantial cancer risk. J. Edward Hunter Biological Subcommittee, Institute of Shortening and Edible Oils, Inc., 1750 New York Avenue. NW. Washington, D.C. 20006 References t. National Research Council, Diet. Nutrition, and Cancer (National Academy Press. Washington. D.C.. 19821. 2. F. B. Shorland et al..J. Auric. Food Ctiem. 29. 863 (1981). 3- M. G. Simic and M. Karel, Eds.. Autoxidution in Food and Btoitwhui Systems (Plenum. New York. 1980). 4 G. A. Nolen et al.. J. Nutr. M, 337 11967). J. R. M. Marston and S. O. Welsh. Natl. Food Rev 2|, 17 (1983). 6. American Cancer Society. Canter Fans A Fit,- ures--I9K3 (American Cancer Society. New York. 1982). 7. M. G. Enig et at.. Fed. Froc. Fed. Am. Soc. Exp. Biol. 37. 2215 (1978). 8. T. H. Applewhite, ibid. 38. 2435 11979): J. C. Bailar III. ibid., p. 243S; W. H. Meyer, ibid., p. 2436. 9. E. A. Emken. J. Am. Oil Chem. Soc. 58 , 278 (1981). 10. J. B. Ohlrogge el al., J. Lipid Res. 22. 955 (1981). 11. Food and Nutrition Board. Recommended Di etary Allowances (National Academy of Sci ences. ed. 9. Washington. D C.. 1980). 12. R. R. Brown, Cancer Res. 41. 3741 (1981). 13. Council for Agricultural Science and Technolo gy. Diet. Nutrition, and Cancer: A Critique (Special Publication 13. Council for Agricultural Science and Technology. Ames, Iowa. 1982). We commend Ames for his review of natural dietary toxins, but not for con cluding that, rather than reducing expo sure to environmental and occupational carcinogens, "dietary practices are the most promising area to explore" for re ducing cancer risks. Ames' article, moreover, is flawed by substantial er rors, omission of relevant data, and reli ance on tenuous hypotheses. These limi tations are more significant in view of the major public health implications of Ames' article and the accompanying edi torial by Abelson, press release, and publicity in the mass media. Ames' position that there is no evi dence for generalized recent increases in U.S. or U.K. cancer rates, other than for cancers attributed to tobacco, is based on epidemiological analyses that, with tenuous justification, exclude people over the age of 65 and also blacks of all ages and attribute a near exclusive to bacco etiology to cancers of various or gans in addition to the lung (/). In fact, overall cancer rates have increased sharply since 1970 (2). Incidence and mortality rates in the United States, age standardized to 1970, have risen sharply since the late 1960's particularly for per sons over 60, blacks of all ages, and a wide range of occupational subgroups (2-4). From 1969 to 1976, mortality rates increased for white and black males by 8 percent and 17 percent, respectively, and for white and black females by 4 percent and 6 percent, respectively. While this increase was pronounced for lung cancer--21 percent and 32 percent for white and black males, respectively, and 74 percent and 56 percent for white and black females, respectively--in creases also occurred in other organs, including, for whites, the prostate (11 percent), male and female kidney (5 per cent), and female breast (4 percent); sharper increases were noted for less common cancers, including those of brain, liver, esophagus, and multiple my eloma. Incidence rates rose more rapidly than mortality on an overall basis and for cancers of various organs, such as the colon, bladder, kidney, skin (melano ma). uterus, female breast, and prostate, besides lung (2); for whites, cancers of sites other than the lung accounted for approximately 70 percent of the in crease. The most recent data show per sistence of these trends through 1980 (5). These trends are consistent with the the ory that past exposure to industrial car cinogens, whose production have in creased exponentially since the 1940's, are responsible for recently increasing cancer burdens (3, 4). The assertion that smoking is respon sible for essentially all lung cancer, and thus accounts for almost all recent in creases in cancer rates, is negated by substantial evidence (J), including (i) the more than doubling of lung cancer rates among nonsmokers over the last two decades, with the proportion of these cancers in nonsmokers approaching 20 percent (3. 6); (ii) the sharply increasing incidence of adenocarcinoma of the lung, which is less closely related to smoking than are squamous and oat celt carcino mas (7); (iii) over the last three decades' (5), the decline in the proportion of smoking males and the tar content of cigarettes, while lung cancer mortality increased at a rate that cannot be ac counted for by cohort effects; (iv) the strong positive associations, largely in dependent of smoking habits, between lung cancer and exposure to a wide range of occupational carcinogens, including vinyl chloride, mustard gas and chloro- y methylmethylether. and carcinogenic processes, such as copper smelting and uranium, zinc, and lead mining (3, 4)\ (v) lung cancer rates in black men that are now about 40 percent higher and have been increasing more rapidly than in whites over the Last 30 years, although blacks smoke less and start smoking later in life (4, 9); (vi) lung cancer rates that are almost equal in white and black women, although the proportion of whites smoking more than one pack a day is twice that of blacks (9); (vii) a threefoldincrease in lung cancer rates among women between 1950 and 1975, a steeper increase than could be accounted for by the modest rise in their smoking preva lence (5); (viii) the major geographic variations in mortality rates due to can cers of the lung (besides other organs) that have been associated with work place and community air pollution U0) and are not explainable by differences in smoking patterns; (ix) the shift of the highest lung cancer rates from northeast ern to southeastern and southcentral states after World War II industrializa tion of the South; and (x) the divergent trends and directions observed between cancers of the lung, on the one hand, and, on the other hand, of other organs, including the esophagus, buccal cavity, and pharynx (4), which have also been strongly associated with cigarette smok ing (/). These considerations in no way 660 SCIENCE. VOL. 224 How To A Bath Without Ta ing One Learn how constant temperature happens. Precision'1* baths show you how. V&ve made more constant C appliances for the lab than anyone else in the world. FREE GUIDE on how to buy a bath plus literature on general purpose, circu lating. and shaker baths. Write GCA/ Precision Scientific Group. 3737 W. Cortland St.. Chicago. III. 60647. LEDs show oscillation rate and *C direct. No waiting, no average. Model 25 shaker bath, 37*0 0.05aC uniformity. Precision" Baths Stainless steel chamber. Block insulation. Baked silicone enamel finish. Quiet, frictionless rollers keep shaker tray and samples uni formly in motion. Digital temperature setpoint. 0.1 *C precise. Platinum probe immersed in the liquid is sensitive to .0.03C. GCA Solid state controller cycles heat near setpoint for better balance, less drift GCA CORPORATION Precision Scientific Group For literature circle reader service number 194 For a sales representative to call circle reader service number 195 detract from the critical importance of tobacco as a major cause of preventable disease and death. In his statement that high-dose expo* sure to occupational carcinogens "might also turn out to be important for particu lar groups of people" (emphasis added], Ames does not acknowledge the sub stantial literature on occupational can cer. According to a 1978 federal esti mate. occupational exposure just to as bestos and live other carcinogens could, on a worst case basis, account for 18 to 38 percent of all male cancers in coming decades {11). Even outspoken critics of these estimates, whose analyses Ames cites, concede that "the minimum pro portion of all current cancer deaths at tributable to occupation can hardly be less than 2% or (/), 4000 to 6000 male deaths per annum. Asbestos and coke plant workers both have lung can cer rates five to ten times those of appro priate controls Uf). Some 10 million workers are now potentially exposed to II "high volume human carcinogens." and there are major excesses of cancers throughout a wide range of occupational groups, including oil refinery and petro chemical workers, rubber and tire work ers, welders and metal-trades workers (4), and atomic plant workers {12). These studies are all the more important as two- to fivefold excesses in cancer rates have generally been necessary before they could be detected by standard epi demiological techniques {13). Contrary to Ames, substantive studies have documented the carcinogenic ef fects of urban air pollution or some relat ed urban factor. Accordingly, the World Health Organization concluded that "it is probable that some urban atmospheric factor is involved [in the etiology of lung cancer], resulting from the air pollution from car exhausts, fumes from heating systems and industrial fumes" (14)\ automobile exhaust contains a wide range of carcinogens, many common to tobacco smoke. In addition, many epide miological studies have documented large geographical variations in stan dardized cancer mortality rates, on an overall and organ-specific basis, with higher rates in communities located near smelters, petrochemical plants and facili ties producing nuclear weapons, and in communities with high levels of atmo spheric pollution {10, I5)\ definitive epi demiological evidence of carcinogenic and reproductive hazards from proximi ty of residence to hazardous waste land fills or industrial impoundments is not yet available, although preliminary data from sites such as Woburn, Massachu setts, are highly suggestive {16). Ames dismisses the possibility that carcinogenic synthetic pesticides, mar keted since the 1940's, may contribute substantially to cancer rates, as their dietary intake is claimed to be 10,000 times lower than that of age-old "na ture's pesticides." There is, however, much evidence to the contrary. For ex ample. a number of widely used chlori nated hydrocarbon pesticides have accu mulated by many orders of magnitude in certain foods to levels comparable to those inducing cancer in small groups of experimental animals {17). Chub and trout in Lake Michigan have been found with aidrin and dieldrin residues above 0.3 part per million, and similar residues of chlordane and heptachlor have been found in the Great Lakes and in Long Island and New York City lakes: in 1983 Montana health officials warned against eating game contaminated with concen trations of heptachlor epoxide more than 100 times the Environmental Protection Agency's (EPA's) "acceptable intake level." Aidrin and dieldrin were found to be carcinogenic at dietary concentra- C 33 oOJ to CCOD How To Buy A C02 Incubator Without Choking Up Learn how a constant environment happens. PrecisionTM COz incubators show you how. We've made more constant aC appliances for the lab eo>nnajnMMrfOrQW*t OnflOtaoMur than anyone el& in the world. FREE GUIDE on how to buy a COz incubator plus literature. Write GCA/Precision Scientific Group. 3737 West Cort/and Street. Chicago. Illinois 60647. Precision' COs Incubators Digital pushbutton control to set *C and COz%. Bright LED display for each. Shelves and supports pop out to easily clean and decontami nate chamber. Hookups for remote alarms and monitoring of *C and COz%. Dual purge cycles to swiftly adjust for changes in CO2 tension. -All solid state, automatic control eliminates COz waste of continuous flow Stainless steel chamber Waterjacket insulation. Polyurethane enamel finish. COz monitored at point of infusion for immediate response. Fan optimally positioned to prevent COz stratification. PrecisionTM automatic COz incubator. 0.1 C. 10.1% COz uniformity GCA CORPORATION Precision Scientific Group For a Sales Representauve to call circla reader service number 196 For Literature arete reader service number 197 URL 03904 tions of between 0.1 and 20 parts per million in five separate rodent bioassays, and residues of chlordane and heptachior have been found in concentrations in human fat similar to those found in rats in whom carcinogenic effects had been induced by these pesticides (/). By ail principles of extrapolation, such expo sures would be expected to result in a significant excess of human cancers. The widespread use of chlordane and hepta chior for termite treatment represents additional major carcinogenic expo sures. Indoor chlordane concentrations greater than an arbitrary interim guide line of 5 micrograms per cubic meter have led to the evacuation of more than 1500 contaminated homes at Air Force bases across the country (19) and to the petition by a New York State citizens" group, after the finding in April 1983 that 63 percent of 443 treated homes were contaminated, to ban the use of chlor dane for termite treatment. Exposure to 5 micrograms per cubic meter of chlor dane. approximately 50 micrograms per day for an average adult, according to EPA extrapolations that considerably underestimate risk for several reasons, including neglect of high-dose flattening, would be expected to increase lifetime cancer risks by as much as 0.1 to 0.5 percent (20). Ames' position on the significance of dietary burdens of carcinogenic synthet ic pesticides is not supported by recent data on ethylene dibromide (EDB) resi dues, with concentrations up to 5000 parts per billion in fiour and citrus pulp. EPA estimated, again using procedures that minimize risk, that lifetime expo sures to "realistic worst case" dietary concentrations of 31 parts per billion of EDB would result in cancer risks of from 10"4 to I0_> (2D. about 300 to 3000 deaths per year, occupational risks were estimated to be as high as 40 percent. Ames has also objected to (he regulation of EDB. saying that the "trace of the carcinogen EDB now allowed in food is insignificant" (22): this in spite of the fact that available noncarcinogenic alter natives include aluminum phosphide for grains and cold storage for fruits and vegetables. The minimal references by Ames to problems of poorly regulated exposures to a wide range of environmental and occupational carcinogens are in contrast to his exaggerated emphasis of the roles of high-fat and low-fiber diets and of charred foods as ``major risk factors," although evidence for such risks, where not negative, is generally inconclusive. A recent report concludes that "in the only human studies in which the total fiber consumption was quantified, no association was found between total fi ber consumption and colon cancer" (73). The position that high fat consumption is a major cause of breast and colon cancer is based on experimental and epidemio logical studies (/. 24). However, this evidence is weak and inconsistent (2. 25). There appear to be no data on the correlation between the proportion of fat in the diet, the critical variable examined in the animal experiments, and rates of colon and breast cancers on a nation-by nation basis: while those rates are strongly correlated with absolute fat con sumption. this correlation is equally good with other measures of industrial ization. such as per capita energy pro duction U). Moreover, up to 20-fold in creases in dietary fat were generally nec essary to increase tumor yields in ro dents after the administration of carcinogens, whereas between-country differences in total fat consumption are generally less than a factor of 2 (3;. Finally, no evidence was found in two major case control studies of an associa- tion between fat consumption and breast cancer rates (26). These considerations do not denigrate the importance of a prudent diet in the promotion of health nor the need for research in this area which could lead to future cancer pre vention strategies; a low-fat and highfiber diet not only decreases intake of fat-soluble synthetic carcinogenic con taminants but also reduces risks of car diovascular disease and diverticulitis. Evidence on the qualitative and quan titative significance in generalized diets of Ames' examples of "nature's pesti cides" and on their carcinogenicity is unimpressive. For instance, conclusions about the carcinogenicity of pepper are based on the results of a single question able study (27), and the inference that mushrooms are carcinogenic is based on the identification in certain mushroom extracts of unstable diazonium com pounds that are carcinogenic in mice only after artificial in vitro stabilization. The implicit identification of mutagens with carcinogens, the implication of an identity in their underlying mechanisms, the blurring of the distinction between different types of mutagens', the identifi cation of quantitative mutagenicity with the results of Ames' bacterial assay, and the derivation of carcinogenic potency from quantitative mutagenicity data are all of questionable validity (28). Many mutagens are inactive in carcinogenesis tests, and many carcinogens are inactive in short-term tests for mutagenicity (29); glutathione is positive in the Ames test (JO), although Ames recognizes it as an anticarcinogen and an antimutagen. Fur thermore. recent evidence has suggested that gross mutagenic events, such as chromosome translocations, are more likely to be crucial in carcinogenesis than are the point mutations or deletions de tected in the Ames assay (28). Moreover, while somatic mutations are likely to be involved in carcinogenesis, epigenetic events also appear critical. Ames' discussion of free radicals and the potential anticarcinogenic effects of antioxidants is speculative and of dubi ous relevance. Even one of the authors cited in support of the thesis that carot enoid antioxidants are protective in smokers has admitted that various stud ies revealed only "a slightly lower than average incidence of cancer among peo ple with above average intake of 3-caro tene" and that even this slim association may be artifactua! <3i). A recent largescale case control study <J2) produced no evidence "relating intake or serum levels of antioxidant vitamins to a re duced cancer risk." Evidence for major carcinogenic ef- 666 fects of trace natural components of U-S. diets is speculative. Strategies based on this hypothesis offer little hope for cancer prevention, and the hypothe sis affords no basis for Ames' trivializ ing the importance of reducing exposure to occupational and other environment al carcinogens. Understandably, such strategies are applauded by corporations resisting regulation of their carcinogenic products and processes and seeking, with others, to explain away cancer cau sation largely in terms of diet and faulty life-style (/). Strangely, Ames' current proposals appear at variance with his strongly argued recent positions (JJ). These include warnings that EDB is "a potent carcinogen" whose presence as an impurity in tris-BP [tris (2,3-dibromo- propyl) phosphate] is one of the reasons why this flame retardant "should not be used"; that there are "enormous possi ble [carcinogenic] risks" from inade quately tested industrial chemicals, such as flame retardants; that a "steep in crease in the human cancer rate from [industrial] chemicals may soon oc cur ... as the 20- to 30-year lag time for chemical carcinogenesis in humans is almost over"; that "tens of thousands of man-made chemicals have been intro duced into the environment in the last few decades--with widespread human exposure--to low but disturbing doses of these carcinogens" and that such chemi cals should be tested for mutagenicity and carcinogenicity; and that priorities must be established to "minimize human exposure to these chemicals" (S3). Clearly there is substantial evidence that, besides smoking, involuntary expo sures to occupational and industrial envi ronmental carcinogens are major and generally avoidable contributors to the burgeoning national cancer burden and to a wide range of other chronic dis eases. Vigorous public health measures are essential to reduce such exposures. Samuel S. Epstein Joel B. Swartz* Department of Preventive Medicine and Community Health, University of Illinois Medical Center, Chicago 60680 `Cosignatories: Eula Bingham. University of Cin cinnati Medical School; Donald DahKtcn, Universi ty of California. Berkeley; Susan Daum. Albert Einstein College of Medicine. New York: John Gofman. University of California. Berkeley: Robert Hams, Princeton University; Joseph Highland. Princeton University; Ruth Hubbard. Harvard Uni versity: Marvin Legator. University of Texas Medi cal Branch. Galveston; Kenneth Miller, Oil, Chemi cal and Atomic Workers. Washington. D.C.: Rafael Moure and Michael Silverstcm. International Union, United Auto Workers. Detroit, Michigan; Glenn Paulson, National Audubon Society, New York; Marvin Schneidennan. Bethesda. Maryland: Joseph Wagoner, Springfield. Virginia; George Wald. Har vard University: and Bailus Walker. Commonwealth of Massachusetts. Reference* 1. R. Doll and R. Peto, J. Sait. Cancer !nst. 66. 1191 (1981). 2. Trends in Mortality 1971-1975 (Series DHI, U K. Office of Population Censuses and Sur veys. London. 1978). No. 3; E. Pollack and J. Horm. /. Sail. Cancer Inst. M. 1091 (1980). 1. S. S. Epstein. The Politics of Cancer lAnchor. New York. !979j; S. S. Epstein and J. B. Swam. Suture (London) 289. 12? (1981). 4. D. L. Davis. K. Bridbord, M. Schnciderman. Teratogen. Carcinogen. Mumgen. 2. 105 11982). 5. E. Sondik (National Institutes of Health), personal communication to Congressman J. Con yers. 6. J. Enstrom. J. Salt. Cancer Inst. 62. 755 (t979). 7. C. Harris. CancerChemother. Rep. 4. 59(1973). pan 3; E. Wyndcr and 5. Stellman. Cancer Res. 37. -M08 (]977>; J. Wagoner. P. Infante. D. Barbs. rnvx>n. Res. 21. 15 (JSBO). 8. Smoking omd Health: A Report of the Surgeon General (Government Printing Office. Washing ton. D.C.. 1979J; J. Harris. J. Sail. Cancer Inst. 71. 473 (1983). 9. Advance Data from Vital and Health Statistics ofSattona! Centerfor Health Statistics I No. 52. Department of Health. Education, and Welfare. Washington. D.C.. 19 September 1979): Health United Stotts (No. {PHSi 83-1232. Department of Health and Human Services. Washington, D.C.. December 1982). 10. J. Mason and F. McKay. Atlas of Cancer Mor tality for (J.S. Counties: /950-/969 (Department of Health. Education, and Welfare. Washington, D.C.. 1975); Atlas of Cancer Mortalin for U-S. Counties among US. Son-Whites (Department of Health. Education, and Welfare. Washington. D.C., 1976); W. J. Biott, L. A. Brinton. J. F. Fraumeni, Jr.. B. J. Stone. Science 198. 51 (1977); B. Camow. Environ. Health Perspect. 22. 17 (1978); L. S. Robensoo. ibid. 36. 197 11980). 11. K. Bridbord et a!., "Estimates of the fraction of cancer in the United States related to occupa tional factors'* (National Cancer Institute. Na tional Institute of Environmental Health Sci ences. and National Institute for Occupational Safety and Health. Washington, D C.. I9?8l. 12. G. W. KneaJe. T. . Stancuso. A. M. Stewart. Br J. Ind. Med. 38. 156U981). 13. J. Stellman and S. Stellman. paper presented at the annual meeting of the American Society for Preventive Oncology, Chicago, III.. 6 and 7 March 1980. 14. U- Mohr. D. SchmahJ. L. Tomatis. lARCUnt. Agency Res. Concert Sci. Pual. 16 119773. 15. C. Johnson, paper presented at the AAA5 An nual Meeting. Washington. D.C., 1982. 16. S. W. Lakagos, B. J. Wessen. M. Zelen. unpub lished manuscript. 17. Office of the General Counsel. Environmental Protection Agency, "Respondent's brief, pro posed findings, and conclusion on suspension" in re Shell Chemical Company et al.. Consoli dated AldritvDieldnn Hearing (FIFRA Dockets 145 et at., Washington. D.C., 16 September 1974); S. Epstein. Set. Total Environ. 4. I (19751. 18. Office of the General Counsel. Environmental Protection Agency. "Respondent's final brief' m re Velsicol Chemical Corporation et al.. Chlordare and Heptachlor (FIFRA Docket 384. Washington. D.C.. 8 December 1975): S. Ep stein, Sci. Total Environ. 6. 103 (1976). 19. An Assessment of the Health Risks of Seven Pesticides Used for Termite Control (National Academy Press, Washington. D C.. 1982). 20. Ambient Water Quality Criteria for Chlordune (Environmental Protection Agency. Washing ton. D.C., 1980). 21- "Ethylene dibromide: Position document 2/3" (Environmental Protection Agency. Washing ton. DC., 1980). 22. B. Ames, Wail Street J.. 14 February 1984. p. 30. 23. Diet. Sutrition. and Cancer (Saiionui Academy Press, Washington. D C.. I9K2). pp. 8-5. 24. B. Armstrong and R. Doll. Ini. J. Canter 15.617 11975). 25. J. Berg, in Persons at High Risk of Canter: An Approttth to Cancer Etiology and Control. J. Fraumeni. Jr., Ed. I Academic Press. New York. 1975). pp. 201-224; K. Carroll. Cancer Res. 35, 3374(1975). 26. R. Phillips, Cancer Res. 35, 3513 (19751; A. B. Miller et al.. Am. J. Epidemiol. 107. 499 (1978). 27. J. M. Concon. D. S. Newburg, T. W. Swerczck. Sutr. Cancer l, 22 (1979). 28. I. Weinstein. M. Wigler. C. Pictropaolo, in Origin of Human Cancer, H. Hiatt. J. Watson. J. Winston. Eds. (Cold Spring Harbor Labora tory. Cold Spring Harbor, N.Y.. 1977), pp. 751- SCIENCE. VOL. 224 c 3D OOJ CoO i 772:!. Weinstein el al.. in Carcinogens: Identifi cation and Mechanisms of Acr/on. A. Orvffin and C. Shaw. Eds., (Raven. New York, 1979), pp. 399-418; H. Rubin. J. Natl. Cancer Inst. 64, 995 (1980); J. Cairn, Nature (London) 2*9, 353 (1981); R. Sager, Chromosome Mutation and Neoplasia (Liss, New York. 1983), p. 333. 29. S. J. RmVus and M. S. Legator. Cancer Res. 39, 3289(1979). 30. H. Clatt, M Protid-Sabljid. F. Oesch, Science 220. 961 (1983). 31. R. Peto, R. Doll, J. D. Buckley. M, B. Spom, Nature (London) 290, 201 (1981). 32. W. C. Willett et at., N. Engl. J. Med. 310, 430 (1984). 33. A. Blum and B. N- Ames, Science 195, 17 (1977); B. N. Ames, ibid. 204, 587 (1979). Weinberg and I are in agreement that cancer and heart disease appear to be in large part degenerative diseases of old age, that oxygen radicals generated dur ing normal metabolism are likely to be major contributors to this aging process, and that it is unlikely we are going to eliminate them. I also agree that Fridovich. Totter, and Handler have made major contributions to the field. We also agree that every meal con tains natural carcinogens, and it is un likely we are going to eliminate all of them. However, I do not think that this knowledge makes it any less important to work toward cancer prevention. By identifying smoking as a major cause of lung cancer and heart disease, we have furnished people with the knowledge that they can live 8 years longer on average by not smoking heavily. The incidence of stomach cancer is high in Japan and low in the United States, while colon and breast cancer incidence are high in the United States and low in Japan. This may be due to a limited number of dietary components, and if we could identify them, we might be able to prolong the life span of the people affect ed in both countries. Understanding some of the main causes of cancer may be the first step in preventing cancer, and although causes and mechanisms are complex, with more knowledge we should be able to sort out some of the major risks in our diet and intervene in many ways, both to minimize significant carcinogens and to maximize amicarci* nogenic defenses. I also agree that the preoccupation with tiny amounts of man made pollution has been blown up out of proportion. lngmanson rightly points out that sele nite is another name for the crystalline form of the mineral gypsum (CaSOj 2H;0). The etymology is from the Greek for moonstone, "probably an allu sion to the soft moon-like reflection of light from some of its faces" </). The selenite I meant is the ScOj2~ anion (analogous to sulfite and tellurite). The name of the element selenium is also derived from the Greek word for moon, 668 selene. Selenite also means "a supposed inhabitant of the moon" (2); presumably there will not be any confusion with this last meaning. Hunter says that "there are insuffi cient reliable data to justify the sugges tion that certain fats in the current Amer ican diet represent a substantial cancer risk." The situation is confusing because there are so many types of fat and the evidence so far does not prove cause and effect. I referenced the considerable epi demiological literature associating high fat consumption with colon and breast cancer and the considerable body of ani mal experiments implicating high fat with cancer. The National Academy of Sciences committee (which also re viewed the field), and more recently the American Cancer Society, have advised the American public that it would be prudent to reduce their fat intake to lower cancer risk. It was important to point out the controversy as to where prudence begins and to mention the dis agreements with this view. Recent re views on nutrition and cancer also dis cuss fat (J). Most scientists would em phasize that the evidence linking fat to cancer is much less secure than that implicating cigarettes, alcohol, or asbes tos with cancer. I had hoped that a discussion of plausible molecular mecha nisms for a fat-cancer connection might provide some testable hypotheses. I dis cussed cyclopropenoid fatty acids, ran cid fat. and peroxisome oxidation of cer tain fats. Newmark et al. (4) and Welsch and Aylsworth (J) have other explana tions. Ail of these mechanisms are plau sible, but we do not know which, if any, are correct. Fat rancidity products in the diet still appear to be a possible source of muta gens and carcinogens that could contrib ute to colon and breast cancer. 1 listed in my article some of the carcinogenicity and mutagenicity data on the variety of hydroperoxides, enals, epoxides, and other reactive chemicals produced by the rancidity reaction. Appreciable amounts of lipid oxidation products may exist in palatable food. For example, Tsai et al. (6) have found significant amounts of cholesterol epoxide (a mix ture of a and &) in commercial dried eggs, scrambled egg mix, and dried whole egg products containing additives, each averaging about 20 parts per million (ppm), although some samples reached eight times this. Cholesterol epoxide is a weak alkylating agent, induces sarcomas at the injection site in rats and mice, is positive in a sister chromatid exchange test, transforms hamster embryo cells. induces chromosome damage in human fibroblast cultures, and is mutagenic in hamster cells (7, 8). Concentrations in human breast fluid, prostate secretions, or serum samples from particular people can be enormous (5. 9), although it is not dear whether the source is endogenous oxidation or the diet. We need more research on the extent of epoxide de struction by the acid in the stomach. In addition, lipid hydroperoxides are pres ent in heated fat that is reused. A nu mber of hydroperoxides have been shown to be mutagens and carcinogens, and others are likely to be, due to their generation of oxygen radicals. I discussed ionizing ra diation as a mutagen-carcinogen that is active because it generates oxygen radi cals, and I also referred to the carcinoge nicity of hydrogen peroxide and fatty acid hydroperoxides. Even a small amount of oxidation (for example, a per oxide number as low as 2) which could be found commonly in cooking oil in restaurants and in fat (/0), would repre sent a level of 1200 ppm (if it were a triglyceride hydroperoxide). Meat can also have a fair amount of rancidity. I mentioned Shorland's article ()/) be cause it reviews some of the literature on rancidity in meat; "In contrast to fresh intact meat, cooked and uncooked ground meat becomes rancid within 48 hours at 4C. . . . This phenomenon has been described ... as `warmed over' flavor. . . ." Rancidity products (as measured by malondialdehyde reaction) were found to be increased in ground meat stored in the refrigerator and in the urine of people who consumed the meat {12). I gave references to both sides of the trans fatty acid controversy, and I find Hunter's additional comments use ful. The letter from Epstein and his co signatories implies that my inquiry into natural dietary carcinogens and anticar cinogens trivializes the importance of reducing exposure to the carcinogens of occupation and pollution and that, there fore, I am aiding the corporations, which Epstein et al. imply are the true causes of cancer. They also criticize me for changing my mind and seem to misun derstand the chief purposes of my arti cle. One way in which biology advances is by the formulation of new hypotheses which can then be tested and either rejected, accepted, or (more commonly) modified and converted into the next generation of more specific and more testable hypotheses. It is through this process that scientists change their minds, which is, in fact, desirable. I was prompted to write the article in order SCIENCE. VOL. 224 URL 03906 1 * to draw together five areas of research: 1) The standard epidemiological view that dietary factors may be important in the etiology of certain types of cancer (3, U. 14). 2) The awareness that increasing numbers of natural products are being identified as carcinogens in rodent stud ies and the realization that many of these compounds are present in very large amounts in the diet relative to the amounts of man-made carcinogens. 3) The finding in recent years of a large number of mutagens formed on cooking food and also among the group of natural pesticides present in plants and molds, many of which appear in the human diet in large amounts. A number of these mutagens have now been tested and shown to cause cancer in laborato ry animals. These mutagenicity findings may be much more representative of the dietary hazards to which humans are exposed than the findings of animal can cer tests, in which very few chemicals are examined each year, almost all of which are man-made. 4) The relation between cancer and aging in animals of widely different life spans, such as rodents and humans, sug gesting that cancer is a degenerative dis ease of old age. Of relevance to aging is the recent interest in the generation of oxygen radicals, a destructive process in normal metabolism, which leads to DNA damage and other damage in cells and could be a major force in both aging and cancer. Also relevant are the contribu tions of the radiobiologists who have demonstrated that the oxygen radicals produced by radiation appear to account for a good part of its mutagenicity and carcinogenicity. 5) The finding by many cancer re searchers that a variety of nutritional factors can have a marked anticarcinogenic effect in rodents (for example, an azo dye gives 90 percent of the rats cancer, but only 14 percent of the ani mals get cancer if 4 ppm of a seleniumcontaining salt is added to the diet) (/5). Many of the substances found to be anticarcinogenic in rodent experiments, such as selenium. {3-carotene, vitamin E, and ascorbate, are components of our normal antioxidant defenses. Many epi demiological studies now implicate di etary factors as being possibly protective against cancer, with some evidence pointing toward antioxidants such as se lenium and (3-carotene. I restricted my article to dietary car cinogens and anticarcinogens because, as I stated, "whether or not any recent changes in lifestyle or pollution in indus trialized countries will substantially af 670 fect future cancer risks, some important determinants of current risks remain to be discovered among long-established aspects of our way of life.'* In addition, I felt that there were certain unifying con cepts in this area. I was not addressing future risks or (hat fraction of cancer today which might be caused by viruses, hormones, occupation, or pollution. I was not attempting to belittle these ar eas, as these and other causes are of concern (13, 14, 16). Doll and Peto (13, 16), for example, ascribe to occupational factors about 12 percent of all lung can cer deaths, plus a smaller percentage of other cancer deaths (totaling about 4 percent of ait cancer deaths), and the emergence of such factors may have contributed to certain cancer trends--for example, over the last few decades the annual number of asbestos-induced U.S. lung cancer deaths has risen from per haps a few hundred to perhaps a few thousand and is likely to continue to increase for some time yet. This increase is caused by the delayed effects of past heavy exposures and not the present asbestos levels, which are much re duced. Whether occupation causes about 4 percent of current cancer, as indicated by Doll and Peto and other leading epidemiologists (13, IS), or even double that, is irrelevant to my article. As indicated by Epstein et al., I have previously advocated, and still do, vigi lance in the area of man-made carcino gens and mutagens stemming from occu pation and pollution (17). Epstein and his cosigners state that a generalized increase in cancer not relat ed to tobacco is in progress, even though the most distinguished epidemiologists who have studied the available data on national trends have come to exactly the opposite conclusion. The question of trends appears to have arisen because, in arguing that some important determi nants of cancer are likely to await dis covery among long-established aspects of ihe American way of life, I noted that "there is no convincing evidence of any generalized increase in U.S. (or U.K.) cancer rates other than what could plau sibly be ascribed to the delayed effects of previous increases in tobacco usage." a conclusion that I drew after discussing the work of Doll and Peto (13) with many leading epidemiologists. If lung cancer is not included, the overall cancer death rates have declined, not only according to Doll and Peto (13, 16), who, inciden tally, did evaluate blacks and people over 65 (13, p. 1272), contrary to Ep stein's statement, but also according to both the American Cancer Society (18) and the recent thorough study by the Environmental Protection Agency and ; the National Cancer Institute (NCI) of three decades of cancer (19). The recent NCI SEER data (20) also show no con- *. vincing evidence of any increases in the major cancers except for lung cancer, while liver, stomach, and uterine cervix cancer are declining. The argument by Epstein et al. for an increase is based on. earlier and superseded SEER data. The statements by Epstein et al. that 20 percent of lung cancer cases occur in nonsmokers and that lung cancer is in creasing in this group both appear to be incorrect. Two recent representative studies of the very high lung cancer areas of South Louisiana and the petrochemi cal area near Houston show that only about 3 to 6 percent of lung cancer patients are people who have never smoked (21). The arguments by Epstein et al. about blacks and lung cancer are not supported by these studies; for ex ample, for black males in South Louisi ana. 97.6 percent of the lung cancer patients were smokers or ex-smokers, while the figure for white males was 97.7 percent. The study in Louisiana also showed a sizable modifying factor of diet: smokers who rarely ate fresh Fruits and vegetables had a 30 percent higher risk from lung cancer for a given amount of smoking, in agreement with the results of Hirayama (22) in Japan on the influ ence of diet on cancer induced by ciga rette smoking, which I quoted. There is no good evidence for any appreciable increase in lung cancer in nonsmokers (23): The conclusion by Epstein et al. is based on an earlier, flawed (/J) study. Thus, only one major cancer rate, that of lung cancer, is increasing, and this appears to be largely attributable to smoking; the rates of liver, stomach, and uterine cervix cancer are decreasing. Breast and colorectal cancer rates (both associated with high fat) have been fairly constant for decades. Some less com mon types of cancer are becoming more prevalent, but the causes still have to be determined. We cannot assume that these are due to occupation or pollution, although some may turn out to be. In the case of esophageal cancer in blacks, for example, an epidemiological investiga tion by NCI implicated high alcohol con sumption as a major risk factor and a good diet as a protective factor (24). The most authoritative and thorough study of causes of cancer in America (16) suggests that diet and life-style are major contributors to cancer and that the con tribution of occupation and pollution are only a very small percentage of the cause of the major human cancers. This study (Continued on page 757) SCIENCE. VOL. 224 URL 03907 LETTERS ers, and so forth, until precise adjust ments are made for both duration and (Continued from page 6?0f amount of smoking. I also pointed out that the major human intake of browned discredits many of the arguments of Ep- and burnt material, even more than stein et al. in detail. The estimates by smoking, comes from cooked food, Doll and Peto agree with those of other which is also full of mutagens and carcin leading epidemiologists (13) and also are ogens. although it is ingested rather than consistent with the analyses of cancer in inhaled, and we do not know the human different countries and immigrant groups risk from this. (14). although everyone agrees there are Epstein et al. dismiss the whole area uncertainties in the figures. The range of of dietary anticarcinogens, despite the human intake of fat is more than Epstein fact that many leading scientists think it et al. state, and the many animal tests is one of the most promising areas in showing an effect of fat were often over cancer research. There is a vast litera a comparable range (3). Breast cancer ture on oxygen radicals in pathology, rates are not so closely correlated with and there is a large and striking literature industrial society. Modem industrial so* on anticarcinogens in animal tests and on cieties such as Japan and the United dietary protective factors in people. Al States have an ever-increasing life span though this is an inherently difficult area, and a lower overall cancer rate than the impact on prevention ofcancer could many less industrialized societies, such be great. We hope our own work on as Finland and New Zealand (25). noninvasive measurement of the high Czechoslovakia (after Luxembourg) has endogenous flux of oxidative DNA dam the highest cancer rate in the world (25) age in individual humans (and the much and does not have any corporations. The higher rate in rats) may help to open up calculations that Epstein et al. refer to new approaches to measuring the effects stating that IS to 38 percent of the cancer of antioxidant modifying factors in hu in coming decades will be caused by mans and to the understanding of the asbestos and five other carcinogens ap contribution of the aging process to can pear to inappropriately assume that any cer (29). one with any exposure to asbestos would Epstein et al. seem to have a double have the same risk as a person with standard about carcinogens. They em massive exposure. These estimates have phasize only industrial chemicals or pol been shown to be erroneous in other lution and belittle or ignore the evidence ways as well (13, 15) and have been concerning the many natural carcino effectively criticized by a good number gens, mutagens, and teratogens dis of the world's leading epidemiologists cussed in my article, as evidenced by (26). their comments on mushrooms, pepper, Epidemiologists studying urban air and fat. Toth, an expert on the carcino pollution have had a difficult time dem genicity of man-made hydrazines, has onstrating any measurable effect on can also published numerous papers on the cer because of the larger effects of small carcinogenicity of many mushroom hy differences in smoking (27). In Contra drazines (30). Gyromitrin, a carcinogen Costa County, California, a possible in mice (31), is present in large amounts connection between air pollution due to (500 ppm, dry weight) in the widely eaten several refineries and excess cancer was false morel and has recently been shown widely publicized for several years. The to massively alkylate the DNA of rats connection evaporated when a more de (31). The mushroom contains several tailed epidemiological study showed that other carcinogenic hydrazines, including the excess cancer was explainable by the N-methyl N-formylhydra2ine, which is higher smoking rate in blue-collar work both potent and stable (30). The common ers, who made up a higher percentage of commercial mushroom Agaricus bi- the population in the area (28). In this sporus contains agaritine, a hydroxy study, as in those discussed above, only methyl phenylhydrazine derivative (3000 3 percent of the males with lung cancer ppm, 45 milligrams per mushroom (50). had never smoked, and green vegetable A metabolic product of agaritine, a dia- consumption showed a marked protec zonium compound, is highly reactive and tive effect. I pointed out in my article mutagenic and was found to cause stom that one would have to breathe in Los ach tumors in mice at low doses of both Angeles smog for about a year to inhale the sulfate and tetrafluoroborate salt an amount of burnt material equal to that (30). Agaritine is not appreciably de inhaled by a smoker in one day. There stroyed by cooking and, when eaten, is fore, one has to be suspicious of superfi distributed in tissues, where it is con cial associations with refineries, smelt verted to the reactive mutagenic diazoni- IS MAY 1984 um metabolite by cytochrome P-450 (50). The diazonium metabolite is also present in the mushroom (about 2 ppm) as it is formed from agaritine by enzymes in the mushroom (50). Epstein et al. dismiss the mushroom work because the diazoni um compound was found to be a carcino gen when tested as a salt, an irrelevant argument in view of ail of the findings about agaritine and the fact that the diazonium ion would be the chloride salt in the stomach in any case. Toth has recently identified another carcinogen in Agaricus, p-hydrazinobenzoic acid (10 ppm, 150 micrograms per mushroom) (3(7). Gyromitrin and some other natural pesticides cause lung tumors when fed to mice (50). Lung tumors may conceivably be caused in humans by natural carcino gens as well as by smoking and by the occupational hazards listed by Epstein et al. The positive cancer test on pepper was done by skin painting and is not a very elegant lest, but it is statistically significant for each of three sites: skin, lung, and liver. Pipeline, the major natu ral pesticide in black pepper (present at 10 percent of its weight), is closely relat ed in structure to the known natural carcinogens safrole, estragole, and methyteugenol, which are also widely distrib uted in spices and plant oils. The test should be repeated, although, unfortu nately, no government agency seems very interested in doing cancer tests on natural products. Toth has also shown that capsaicin, the pungent material in hot pepper, is a mutagen, and he has some preliminary evidence for its carci nogenicity in mice (52). Epstein et al. call all of these "trace natural components." My article was full of numbers showing that the amounts are not traces. Nature's pesti cides are present at pans per hundred and parts per thousand, while man-made pesticides are present at parts per million and parts per billion. We are eating more than 10,000 times more of nature's pesti cides than of man-made pesticides. The arguments of Epstein et al. about man made pesticides do not invalidate this calculation. His calculations are based on rare, highly contaminated foods or people, while mine are based on Food and Drug Administration values for man made pesticide residues. Epstein et al. criticize my discussion of mutagens as potential carcinogens. There are 3000 laboratories using our test alone, and there are many other kinds of mutagenicity tests in which cells from mammals are used. Over the last 10 years, more than 5000 compounds, natu ral as well as man-made, have been 757 URL 03908 tested, and an amazing variety of muta gens in the natural world have surfaced. These include superoxide, hydrogen per oxide, and aldehydes generated by our normal metabolism; products of lipid rancidity; products of cooking food such as the brown color on our toast; many of nature's pesticides; and the aroma of butter (diacetyl). We cannot ignore this new information, as it is telling us some thing important about nature. There are good reasons for thinking that mutagens are potential carcinogens and that DNA damage is of concern in itself. Our test is successful at detecting carcinogens as mutagens (80+ percent) and has been improved since validation (33). In many cases its mutagens have later been found to be carcinogens in rodents (for exam ple, natural metabolic products such as hydrogen peroxide and formaldehyde, natural pesticides such as ally! isothio cyanate from mustard, many pyrolysis products from cooking, nitropyrenes from diesel exhaust, and synthetic sub stances such as ethylene dibromide, hair dye components, and the flame retardant tris-BP). Recent evidence on oncogenes also supports a mutation-DNA damage hypothesis as one aspect of cancer cau sation (34). The identification of muta gens aids epidemiology in its search for hypotheses to test, serves as a bioassay for active principles in complex mixtures (for example, the mutagenic pyrolysis products from cooking, later shown to be carcinogens), and is a way of investi gating active forms of carcinogens. Of course, it is far from a perfect guide. We need to understand promotion as well, although I pointed out recent evidence for an oxygen radical connection. We know that some chemical carcinogens are missed by mutagenicity tests, but the battery of short-term test systems agree with each other quite well and detect a remarkable percentage of the known carcinogens. Agents that are known to cause deletions, translocations, and chromosomal rearrangements, such as xrays, are also detected by our test sys tem (33). Whether there are many signifi cant carcinogens that cause transloca tions and are not detected as mutagens in short-term tests remains to be seen. Glu tathione. as pointed out by Epstein et al., is convened to a mutagen by kidney homogenate (but not liver homogenate). This may be due to an enzyme specific for kidney that generates oxygen radicals from O2 and glutathione (35), but this is not relevant to the role of glutathione as an antioxidant or the value of mutagenic ity tests. Epstein and his colleagues may be 758 drawing the wrong conclusion from the results coming out of animal cancer tests. The National Cancer Institute-Na tional Toxicology Program animal can cer bioassay test program, which is the most thorough and extensive source of tests, is looking at only a small portion of the chemicals in the world. The current cost is more than 5500,000 per chemical, so it is difficult to test many chemicals. The tests are almost exclusively done on man-made chemicals and therefore, of course, find man-made carcinogens. Out of about 200 chemicals tested by NCI in 8 years, 60 percent were judged carci nogenic, 33 percent noncarcinogenic, and 7 percent inadequately tested (i6). The high percentage of carcinogens found is somewhat disturbing, as the conventional wisdom is that carcinogens are very rare. This discrepancy could be accounted for by the fact that more sus picious chemicals are being tested. It also could be that carcinogens are more common than we think. We have no idea of what the true percentage of carcino gens is among chemicals in general (in cluding natural ones) when tested at the maximum tolerated dose in rodents. Even if it is 10 percent, our current regulatory policies, which assume car cinogens are rare, are in trouble. I point ed out that we are ingesting enormously more in both number and amount of natural pesticides and other natural toxic molecules (and traditional mixtures such as cooked food) than we are of man made substances. Plants have been de vising nasty chemicals to kill off insects and animals throughout ail of evolution. There is no reason to think nature is any more benign than man. We already know about natural carcinogens, such as psoralens, afiatoxins, sterigmatocystins, pyrrolizidine alkaloids, safrole, asbes tos, radioactive potassium in our body, radon coming up from the ground into our houses, mushroom hydrazines, hy drogen peroxide made during normal metabolism, and sunshine. The same de toxification and activation mechanisms appear to operate on both man-made and natural chemicals. Animal cancer testing will not be able to catch up with the large number of mutagens being uncovered. Therefore, it might be reasonable to plan on doing animal cancer testing on those natural mutagens that we eat in largest amounts to see how many are rodent carcinogens and to provide a benchmark of the natural hazard for setting priorities relative to man-made carcinogens. Much fear of traces of man-made car cinogens is based on ignoring the natural background of carcinogens and using "worst case'* assumptions in extrapolat ing risk from the most sensitive rodent (when the chemical is given at the maxi mum tolerated dose) to low-dose human exposure. This quantitative extrapola tion is viewed with great unease by much of the toxicological and epidemiological community because there is little scien tific support underlying it (13). It is an extrapolation based on ideas of pru dence, not on firm science US, 37, 38). This is true, of course, for both natural and man-made carcinogens. We cannot validate these extrapolations. In a few cases, individual extrapolations from rats to man can be examined, although not very satisfactorily (j). In cases such as ethylene dibromide (EDB) (39), aflatoxin (J, 40), and vinyl chloride (38), the extrapolations appear to markedly over estimate the risk to man. We assume a linear response with dose, but we do not know if this is true. We have no basis for assuming metabolism in rodents and hu mans is the same. We do not know if we can extrapolate cancer risk from short lived species such as rodents to longlived species such as man because of the cancer-longevity connection (my article discusses this). Antioxidant defenses can differ markedly from rodents to humans (41). We know that some dietary changes make great differences in cancer inci dence in rodent tests, but we do not understand which are the dietary pro tecting factors that appear to influence cancer risks so markedly in human epidemiology studies. We do not under stand promotion or the interaction of carcinogens. Thus, it is time to do the same types of worst case risk calcula tions on natural chemicals as well as man-made chemicals before deciding what our priorities are. I emphasized that To identify a substance, whether natural or man-made, as a mutagen or a carcinogen, is just a first step. Beyond this, it is necessary to . . . quantitate the approximate magnitude of the risk. . . . [T]he rapid progress of sci ence and technology . . . should help to dis pel confusion about how important health risks can be identified among the vast number of minor risks. Epstein ei at. distort my statement on EDB (42), which was in favor of the EPA standards and much more stringent oc cupational standards. Aluminum phos phide (phosphine gas), the suggested al ternative of Epstein et al. to EDB for grain fumigation, is both extremely toxic to humans and flammable, and. despite their statement, it has not been tested for carcinogenicity (43). Epstein and his cosigners offer sup- SCtENCE, VOL. 224 URL 03909 r port for the idea that cancer is basically a political problem, that it is "corporate cancer," and that the solutions are not scientific, since we already know what we need to know to solve the problem politically. Yet their fundamental hypotheses and beliefs are wrong or like ly to be wrong, If we follow their advice, we will ignore the major area of diet as a source of both protective factors and risk factors for cancer. We will continue to be preoccupied instead only with indus trial sources of cancer, trying to elimi nate smaller and smaller risks even though the alternatives to these have unknown risks. Pollution is not being neglected when the budget of the EPA is equal to that of NCI, nor are occupation al hazards when we have large govern ment agencies such as the Occupational Safety and Health Administration and the National Institute of Occupational Safety and Health. Smoking, which causes 30 percent of cancer and 25 per cent of heart disease, is largely neglect ed: we subsidize tobacco farmers. Un derstanding cancer mechanisms may turn out to be among the most costeffective ways to reduce the burden of cancer that is not related to tobacco, and for the present the key questions are scientific, not political. I would like to correct some errors and omissions in my article. I omitted a pa per reviewing the carcinogenicity of the herbs comfrey and coltsfoot {44), which contain the very potent carcinogenic pyrTolizidine alkaloids that are wide spread in plants. The use of comfrey is increasing markedly with the new inter est in returning to natural herbs and the increase in health food stores. I over looked the extensive work ofJanzen and others {45) on the biological role of plant toxins and of Morton (46) on the connec tion between tannins in plants and esophageal cancer in a number of coun tries. I also overlooked a paper on the plasma levels in humans of the potato toxins in which the toxicology of these compounds was reviewed {47). Also overlooked were the natural mutagens in ginger {48) and in com. rye, and wheat (49) and the work on the formation of mutagens by the reaction of sugars with amino acids (nonenzymatic browning or Maillard reaction) during the cooking of food (50). It was a single dose of 400 micrograms per gram, not 400 nano grams per gram, of the mushroom diazonium derivative that caused the appre ciable cancer in mice. I would also like to call attention to the paper of Bimboim (5/) on the role of oxygen radicals in strand breaks and promotion, which l overlooked, and to the book in which it 760 appears, which has many papers of rele vance. I also apologize to Shapiro and my numerous botanist correspondents for misdassifying figs. Bruce N. Ames Department of Biochemistry, University of California, Berkeley 94720 References uid Notes !. Encyclopedia Brittanica. Nth Edition (1910). 2. Oxford Universal Dictionary (Oxford Univ. Press. New York. 1953). 3. W. C. Willett and B. MacMahon. N. Engl. J. Med. 319. 697 <1984); J. H. Weisburger, B. S. Reddy. W. S. Barnes. E. L. Wynder. Environ. Heahk Perspect. SO, 101 41983): E. Wynder. Cancer Res. 43. 3024(1993): S. Graham. Epide miol. Rev. S. 38 (1983). See also K. K. Carroll. J. Environ. Pathol. Toxicol. 3, 253 (1980). 4. H. Newmark, M. C. Wargovich, W. R. Bruce. J. Natl. Cancer Inst., in press. 5. C. W. Welsch andC. F. Aylswonh. ibid. 70.21J (1983). 4. L. S. Tsai. K. Ijichi. C. A. Hudson. J. J. Meehan. Lipids 15. 124 (1980); Proc. 39 Inst. Food Tech. Meet, (abstr.) 164 (1979); ibid.. unpublished manuscript. This work is discussed to some extent by L. L. Smith [in Autoxidation in Food and Biological Systems. M. G. Simic and M. Karel. Eds. (Plenum. New York. 1980). pp. 119-132). Tsai et al. also show a solution to the problem. They found that direct gas-fired heating of eggs gave much higher epoxide values than indirect heating. NOr produced by gas flames, is known to be an effective inducer of oxidation reactions in lipids, and it would be simple for the industry to switch over to indirect heating; The beer industry switched from direct to indirect heating of malt promptly after nitros- arnines were discovered in beer and were found to be caused by the NO. reacting with the malt. The changeover was perhaps speeded up by the full-page advertisements of Coors. in which they compared the nitrosamine content of their com petitors' beers with their own. which had negli gible levels because it was produced by an indi rect heating method. 7. F. Bischoff, in Advances in Lipid Research, R. Paoletti and D. Kritchevsky. Eds. (Academic Press, New York. 1949). vol. 7. pp. 165-244; A. Sevanian and A. R. Peterson. Proc. Nail. Acad. Sei. U S A., in press. 8. N. L. Petralcis. L. D. Oruenke. J. C. Craig. Cancer Res. 41. 2563 (1981): N. L. Petrakis et al. in Banbury Report 13. Indicators of Ceno- toxic Exposure (Cold Spring Harbor Labora tory, Cold Spring Harbor. N.Y., 1982), pp, 67- 82. These papers review the toxicology. 9. M. F. Gray. T. D- V. Lawrie. C. J, W. Brooks, Lipids 6. 836(1971). 10. I. A. Thompson et al.. Food Technol. fChicagol 21. 405 (1947): C- W. Fritsch. D. C- Egbert, J. S. Magnuson. J. Am. Oil Chem. Soc. 56. 746 (1979). 11. F. B. Shortand et Agric. Food Chem. 29, 863 (1981). 12. E. A. Jacobson. H. L. Newmark. R. P. Bird. W. R. Brace. Nutr. Rep. Int. 28. 509 (1983). 13. R. Doll and R. Peto. The Causes of Cancer (Oxford Univ. Press. New York, 1981). Also available in J. Natl. Cancer Inst. 66. 1191 (1981). 14. J. Higginson, Environ. Mutagen. 5, 929 (1983). 15. A. H. Oaoud and A. C. Griffin, Cancer Lett. 9, 299 (1980). 16. R. Peto and M. Schnciderman. Eds.. Banbury Report 9. Quantification of Occupational Can cer (Cold Spnng Harbor Laboratory. Cold Spring Harbor. N.Y.. 1981). 17. B. N. Ames. Science 204, 587 (1979). 18. Cancer Facts and Figures. 1964 (American Can- cerSociety, New York. 1984). 19. W. B. Riggan er at.. Cancer Mortality Rates and Trends. 195(3-1979 ) Environmental Protection Agency-National Cancer institute. Washington. DC.. 1984). 20. S. S. Devesa. E- S. Pollack, J. L. Young, Jr.. Am. J. Epidemiol. 119. 274 (1984). 21- P. Cornea. L. W. Pickle. E. Fontham. Y. Lin. W. Haenszd. Lancet 1983-11. 593 (1983): P Correa, L. W. Pickle. E. Fontham. N. Dalager. Y. Lin. W. Haenszei. W. D. Johnson, in Pro ceedings, International Lung Cancer Update Conference. M. Mizell and P. Correa. Eds. (Vcrlag Chcmie International. Deerfield Beach, Fla., in press): P. Butfler. L. W. Pickle. T. J. Mason. C. Contant, in ibid. 22. T. Hirayama, Nutr. Cancer I. 67 (1979). 23. L. Garnnkel, J. Natl. Cancer Inst. 64. 106! (1981). 24. R. G. Ziegler er al.. ibid. 67. 1199 (1981). 25. M. Segi, Age-Adjusted Death Rates for Cancer for Selected Sites (A-ClassificationI in 52 Coun tries in 1973 (Segi Institute of Cancer Epidemiol ogy. Nagoya. Japan. 1978). 26. E. Efron. Am. Sped. 17. 10 (1984). 27. F. E. Speizer. Environ. Health Perspect. 47, 33 (1983). Speizer estimates that air pollution causes less than 2 percent of all lung cancers. 28. D. F. Austin. Epidemiological Study of the Incidence of Cancer as Related to Industrial Emissions in Contra Costa County. California (Final Report to the Environmental Protection Agency. Washington. D.C.. 1982). 29. R. Cathcart, E. Schwiers, R. L. Saul, B. N. Ames. Proc. Natl- Acad. Sci. U.SA., in press. 30. B. Toth, in Carcinogens and Mutagens in the Environment. H. F. Stich, Ed. (CRC Press. Boca Ratoct. Fla.. 1984). pp. 99-108: T. Lawson and B. Toth. Am. Assoc. Cancer Res. Abstr. 24. 77 (1983): Eur. J. Cancer Clin. Oncol. 19, 1296 (1983); B. Toth, personal communication. In one cancer test by Toth, agaritine itself was not found to be carcinogenic in mice. 31. A. Meier-Bratschi et al., J. Agric. Food Chem. 31. 1117 (1983). 32. B. Toth and E. Rogan, Fed. Proc- Fed. Am. Soc. Exp. Biol. 43. 594 (1984). 33. D. E. Levin. L. } Mamett, B. N. Ames. Proc. Natl. Acad. Sci. U.SA.. in press; D. E. Levin et al.. ibid. 79. 7445 (1982). 34. E. Samos et al.. Science 223, 661 (1984). 35. C. H. Schmelzer tt al.. Arch. Biochtm. Biopkys. 228.681 (1984). 36. K. C. Chu. C. Cueto. Jr.. J. M. Ward. J. Toxicol. Environ. Health 8. 25) (1981): D. Salsburg. Fund. Appl- Toxicol. 3, 63 (1983). 37. F. C. Roe. Nature (London) 303. 657 (1983). 38. Assessment of Technologies for Determining Cancer Risks from the Environment (Office Of Technology Assessment. Washington, D.C., 1981). 39. J. C. Ramsey. C. N. Park. M. G. Ott. P. J. Gehring. Toxicol. Appl. Pharm. 47. 411 (1978). 40. L. Stoloff. Nutr. Cancer S. 165 (1983). 41. B. N. Ames. R- Cathcart. E. Schwiers. P. Hochstem. Proc. Natl. Acad. Sci. USA. 71. 685B (1981). 42. What I said in the reference Epstein cites is: "As pesticides go. it is a potent carcinogen in rodents. The new standards set for EDB are useful and overdue and should keep industry from getting sloppy. The Ruckelshaus standards seem pretty reasonable while EDB is being phased out. On the other hand, the risk to workers using EDB could be significant. The government's air standards until recently were such that workers were allowed to breathe in an amount of EDB roughly equivalent to eating a million ounces of bread a day with EDS residues of 30 pans per billion (the new Ruckelshaus standard for bread). Two studies of workers who were breathing in close to this amount of EDB for decades showed no increase in cancer. But because of the limitations of these studies. I and others fought to get California to lower the allowable level for workers. In 1982. California lowered the allowable level for workers by more than 100-fold. Our experience with asbestos and radium has taught us we can't ignore occupa tional carcinogenic hazards." 43. U. Kackenberg (Toxicol. Appl. Pharmicol. 23. 147 (19721) has done a cancer test only on food Runigated with phosphine gas. not on phosphine f> at high dose. Hirono, H. Mori. M. Haga. J. Natl. Cancer Inst. 61. 865 II97BI. 45. G. A. Rosenthal and D. H. Janzen. Eds.. Herbi vores: Their Interaction with Secondary Plant Metabolites (Academic Press. New York'. 1979). This book has many interesting articles on de fense toxins in plants: D. H. Janzen, Am. Nat. 111. 691 (1977). 46. J. F. Morton, in Recent Advances in Phyto chemistry. vol. |4, The Resource Potential in Phytochemistry. T- Swain and R. Kleiman. Eds. (Plenum. New York. 1980). chap. 3, pp. 53-73. 47. J. A. Matthew et al.. Food Chem. Toxicol. 21. 637 i 1983). 48. H. Nakamura and T. Yamamoto, Mutat. Res. 122. 87 (1983). 49. T. (shizaki. Y. Hashimoto, K. Shudo, T. Okamoto. Tetrahedron Lett. 23. 4055 UY82): Y. Hashimoto. K. Shudo. T. Okamoto. Mutat. Res. 66. 191 (1979). 50. K. Shinohara*t a/..iWur<ji, Res. 122.279(1983). 51. H. C. Bimboim. in Radioprotectors and Anti- carcinogens, O. F- Nygaard and M. G. Simic, Eds. (Academic Press, New York, 1983). pp. 539-556. SCIENCE. VOL. 224 UHL U3910 i NOTICE, This Material nay be protected by copyright 17; r^Jow. (THIs US. Codri Carcinogens Occurring Naturally in Food by Doris M. Hilker URL 03911 Abstract Carcinogens naturally occurring in food in clude metabolites of microorganisms which in fect plants during harvesting and storage; products of the normal metabolism of plants and products of chemical changes occurring during processing, preservation, preparation or which are picked up from the environment, including contaminated water or soil. Some examples of these types of carcinogens are: fungal metabolites such as aflatoxins in pea nuts stored under improper conditions; safrol from the oil of various plants; tannin in tea, grain and grapes; and polycyclic hydrocar bons including benzo (a) pyrene formed by smoking meat and fish. Epidemiological studies indicate that there is a high incidence of gastric cancer in the areas of the world where smoked fish are common in diets. Vi tamin A may play a rote in preventing the carcinogenic action of polycyclic hydrocar bons. It is well established that certain chemical com pounds have carcinogenic action in humans and in experimental animals. Recently, attention has focused on carcinogens in food as potential causes of human cancer. This attention has resulted largely from con sumer concern about chemicals added to food during processing and preservation. However, much less at tention has been paid to carcinogens occurring natur ally in food material. D.M. Hilker is affiliated with the Department ot Food Science and Human Nutrition. University of Hawaii. Henke Hall 224. t800 East-West Road. Honolulu. HI 96822. This report will deal with the role of carcinogens as metabolites of microorganisms which infect plants dur ing harvesting and storage; as products of the normal metabolism of plants; and as products of chemical changes occurring during processing, preservation or preparation, which are picked up from the environment through contaminated water or soil. Carcinogens as Fungal Metabolites Miller and Miller31 have published several reviews on the subject of naturally occurring carcinogens in food (Table 1) , some of the most potent carcinogens being fungal metabolites. Aflatoxin B,, a product of certain strains ot Aspergilfus and Penicilfium, can induce tumors in livers of rats fed Aflatoxin Bi at 1 ppb. Other aflatoxins, B2, G, M,, Q,. and P,, are less active. The production of aflatoxins is greatest when crops such as peanuts are grown or stored under warm, humid condi tions. Contamination of food with aflatoxin can be a serious health hazard in some areas of the world and has been related to a high incidence of cancer in cer tain parts of Asia and Africa. Alcoholic beverages made from such products as grapes, maize, rice and barley may also become con taminated with aflatoxin by field or storage micro fungi.4* In Kenya, epidemiological studies have im plicated beer made from maize in the high incidence of esophageal cancer.1* The other fungal products, 4-ipomeanol, luteoskyrin, methylhydrazine and ethyl carbamate have been studied much less than the aflatoxins: however, it is known that they may occur in plants in amounts up to 5% of the dry weight. The pyrrolizidine alkaloids, which occur in many plant species, have been found in honey samples from the Pacific Northwest13 in amounts up to 3.9 ppm. The alkaloids tend to concentrate in flowers. Vol. 2, No. 4 Dupiicattoo of Nutrition and Ctncor, m who** or in port by any manna for any mapoaa ** magat. 217 URL 03912 Table 1. Natural Carcinogens in Food" Carcinogen Funaal origin Aflatoxin B, 4-lpomeanol Luteoskyrin Methylhydrazine Ethyl carbamate Plant Droducts Pyrrolizidine Alkaloids Safrol Oil of Calamus Cycasin Thiourea Tannin Polvcyclic aromatic hvdrocarbons Food source peanuts, grains sweet potatoes rice mushroom fermented foods, beer, wine, yogurt, soy sauce herbs, honey sassafras, sesame oil, nutmeg, pepper bitters, liqueurs cycad nuts turnip, cabbage tea, wine, betel nut smoked foods, plants, seafoods Site of action liver liver liver liver, lung lung, lymphoid, skin, liver liver, skin, lungs liver mesenchymal liver, kidney thyroid, liver liver, buccal gastrointestinal constituting 0.15-0.3 percent of the dried flower ma terial. Safrol and related compounds occur in the oils of a wide variety of plants that were used in the US as flavorings until 1960, when they were reported to cause hepatic tumors in rats and mice. Safrol is also present in the oil of sassafras, yet the bark of the sassa fras tree is used for tea in certain parts of the US. Carcinogenic Activity of Polyphenol Compounds Tannins are polyphenol compounds present in many plants including tea. grains such as sorghum, and grapes. The betel nut also contains tannin, and the high incidence of buccal cancer in betel nut chewers in Asia has been attributed to this compound. Bhide and colleagues2 induced gastrointestinal tumors in mice by giving them an aqueous extract40 of betel nut quid. In addition, oral and gastric cancers were produced in the cheek pouch of hamsters by combinations of betel nut, lime and tobacco.40 Bracken fern, which contains a high tannin content, is used as a food in some countries and is also con sumed by animals. Neoplasms have resulted from feeding this material to rats and cows.* Experiments by Pamukcu and co-workers* involving fractionation of the bracken fern, however, indicated that tannin was not the carcinogenic factor. Other foods and bever ages containing tannin (e.g., tea) may be potentially hazardous, but have been largely ignored. Several studies, however, have shown-that tannin-containing fractions of plants used for tea can induce cancer in rats given injections of the extract.1** Plant Carcinogens Oil of calamus contains compounds related to safrol which also have carcinogenic activity when fed to rats. Cycasin from the palmlike cycad tree is one of the more potent plant carcinogens. Nuts from this tree are con sumed by people in the South Pacific Islands, but there appears to be no evidence of human cancer caused by consuming the cycad nut. Thiourea and related com pounds fed to rats as 0.1 % of their diet have produced various types of tumors. Here again, we have no evidence that these compounds cause cancer in humans. More recently, a review by Weisburger* listed additional compounds from plant material with carcin ogenic properties, including: 0-asarone (found in Aspris calamus, or sweet flag), which was used as a flavoring in American vermouth before 1967 when it was banned by the FDA; estragole, a component of oil of tarragon which has been used a a flavoring agent in gourmet types of vinegar; and N-methyl-N-formylhydrazine, which is found in the wild mushroom Gyromitra esculenta. Another recent review by 8rown# gives extensive information on the genetic effects of naturally occurring flavanoids, anthroquinones and related compounds. In recent years there has been a growing interest in the carcinogenic properties of N-nitrosamines. This interest has generally been concerned with nitrite-- 218 of Nutrition and Cancmr. in nftoNor m pan by mny imm for anypurpoeo m iBrngrt. Nutrition and Cancer URL 03913 cured meats. However, nitrosamines have been report ed in many other products, including cheese and fish.11 Scantan and co-workers42 analyzed commercially available bottled and canned beer for volatile nitro samines and found w-nitrosodimethylamine (NQMA) in 23 of 25 samples with levels ranging from 0-14 ppb. Other workers4* found NDMA in six or seven brands of Scotch whiskey, in additionto beer. Mutagens in Meat The development of the Ames method1 for detecting mutagens has led to numerous studies on the mu tagenic properties of various foods. Potent mutagenic activity that is distinguishable from benzo {a) pyrene (BP) activity has been found in cooked beef and beef extract.*-1* The charred surface of cooked fish has also been found by the Ames test to possess high mutagen ic activity.** Attempts to define the mutagenic products in cooked meat have led to studies showing that pyrolysis products of certain amino acids, includ ing phenylalanine and tryptophan, have mutagenic activity.4*-51.*** Rappaport and colleagues41 reported that volatile mutagens from cooked meat exceeded those present in the cooked product prepared in an open container, but that this volatile material could remain on the meat cooked in a closed container. The subject of mutagens from cooked meat will be dealt with in more detail in another article in this volume. Cof fee and tea,34 as well as preserved fish,17 have also been shown by the Ames test to have mutagenic activity. Polycyclic Hydrocarbons The remainder of this report will be concerned with polycyclic aromatic hydrocarbons, particularly ben zo (a) pyrene (BP). There are at least 15 compounds of this class that are commonly found in food. For the sake of brevity this discussion will be limited to BP, even though several other compounds, including ben zo (c) phenanthrene, 7,12,dimethyl benzo (a) athra- cene and 3 methylcholanthrene may be more carcin ogenic and may be present in higher amounts in some foods. It should be mentioned that methods for analyz ing the polycyclic aromatic hydrocarbons in foods differ from one laboratory to another with respect to extraction, clean up and detection procedures, so values may not be comparable. Table 2 shows some of the reported values for BP analyzed in plant foods. In most cases, the values were below 1 ppb. The highest values were found in seaweed, but spinach, soy sauce and tea leaves also have levels above 1 ppb. Presumably, the major source of BP in plants is the soil. Fritz and Engst1* in East Germany analyzed soil at various distances from industry, and found that BP ranged from 0.B-60 ng/kg at the greatest distance from the emitter, 199-1100 Mg/kg at medium distance and 3,000-1,500.000 Mg /kg close to the emitter. Vegetables grown close to the emitter had ten times more BP than those grown at greater distances from the emitter. In a study in Hungary,* the BP content depended upon the depth at which the soil sample was taken. The highest con tent of BP was obtained from samples taken at 0.5 cm and the lowest content was from samples taken at 90100 cm. The deeper soil sample contained 1.6-3.4 Mg/kg. Not all studies are consistent in showing that BP can be taken into plant tissue from the soil. Blum and Swarbrick3 investigated the translocation of 1dC labeled BP into plants grown in a liquid nutrient medium saturated with this compound. The plants selected were green beans, cantaloupes and cotton, which have different growing times. This study indicat ed that the BP was not taken up by the plants. How ever, a Russian study7 in which oats were grown in sandy loam soil enriched with 3H labeled BP, showed that the label was taken into the plant tissue, with the Table 2. Benzo(a)Pyrene in Plant Foods Food Benzo(a)pyrene (Pbb) Spinach Seaweed Carrots Head lettuce Tea leaves Soy Sauce Coffee Miso Bread Rice Potatoes 3.3 31.3 0.7-1.9 0.2-1.3 6.8-7.6 0.2-0.6 0.95 0.18-.26 0.03-.08 2.4-3.0 Source Shiraishi, et al., 1974 Shiraishi, et al., 1973 Side! and Happel, 1975 Sidel and Happel, 1975 Yamaguchi.etal., 1978 Strobe), 1974 Miyabe, et al., 1976 Miyabe, et al., 1976 Miyabe, et al., 1976 Miyabe, et al., 1976 Vol.2, No. 4 Duplication of Nutrition and Cancdr. >n wholo or m port by anti moon* for ony purpoto i$ illegal. 219 URL 03914 Table 3. Benzo(a)Pyrene in Seafoods Food Lobster tail (fresh) Lobster tail (after impoundment) Fish and shellfish Fresh fish, Estonia Whitefish, smoked Whiting, smoked Cod, smoked Herring Smoked salmon Smoked haddock Katsuobushi (bonito) Honobushi (bonito) Sababushi (mackerel of sardine) Urunebushi (mackerel of sardine) Benzo(a)pyrene (pbb) 0.76 281 0.8-9.85 0.11-5.96 4.3 6.9 4.5 1.0 1.0 0.3 12-27 9 7 2 Source Dunn and Fee, 1979 Dunn and Fee. 1979 Shiraishi, etal., 1975 Bogovskii, et al., 1978 Howard and Fazio, 1969 Howard and Fazio, 1969 Howard and Fazio, 1969 Howard and Fazio, 1969 Lijinsky and Shubik, 1965 Lijinsky and Shubik, 1965 Masua and Kuratsune, 1971 Masuaand Kuratsune, 1971 Masua and Kuratsune, 1971 Masuaand Kuratsune, 1971 Table 4. Benzo(a)Pyrene in Meats Food Grilled meat (electric) Grilled meat (charcoal) Grilled meat (fircones) Smoked ham Mutton sausage Bologna sausage Charcoal broiled steak Barbecued ribs Barbecued pork Liquid smoke flavor Benzo(a)pyrene (pbb) 0.18 0.5-50.5 50-190 26 0.07-.15 0.05-.08 8 10.5 5.0 0.1 Source Toth and Blaas, 1973 Toth and Blaas, 1973 Toth and Blaas, 1973 Potthast, 1977 Fretheim, 1976 Fretheim, 1976 Lijinsky and Shubik, 1964 Lijinsky and Shubik, 1964 Malanoski, et a).. 1968 Hirokado.etal., 1975 highest amounts taken into the stalk. The BP was not specifically identified in this study, and the measure ments may have included metabolites or breakdown products of the BP, Blumer and Youngblood4 analyzed various types of New England soil, including continen tal, coastal, marsh and subtidal, in order to obtain in formation on the origin of polycyclic hydrocarbons. The investigators found a high degree of similarity in the molecular weight distribution among the various soil types and suggested that the hydrocarbons may have been formed in forest fires and transported by air. Thus, it would seem that industry and automobile ex haust may not be the only sources of carcinogenic hydrocarbons in plants. The high incidence of gastric cancer in areas where smoked fish is common in diets and also in areas where fish are caught in contaminated waters has led to a number of studies of polycyclic hydrocarbons in seafood (Table 3). The amounts of BP in seafood are quite varied, but in general are higher than those in plant foods. Of particular interest is a study in which lobster was found to contain less than 1 ppb BP in freshly caught samples but contained much higher levels after commercial impoundment in pens con structed with creosoted timber. In this study,1* shrimp, crab and salmon had no detectable BP, and the range of values for smoked fish was no higher than that for fresh fish. The BP in meat products is shown in Table 4. Grilled, broiled and barbecued as well as smoked meats con tain this compound well in excess of 1 ppb, which is considered to be the maximum level in commercially smoked meat in West Germany, Lijinsky and Shubik*1 state that the 8 n$/kg of BP in charcoal-broiled steak is approximately equivalent to the amount in the smoke of 600 cigarettes. In view of the rapid growth of "fast food" industries in the US as well as in other countries, with an increasing proportion of diets consisting of grilled, broiled and barbecued meat, the possibility that these foods may contribute to human cancer should be considered. While there have been many studies on the content of carcinogenic polycyclic hydrocarbons in food and also on the possible mechanisms of carcinogenesis of 220 OvpUcmtion of Nutrfttoo and Connor. in toholo or in port by ony moon* for ony porpooo it Mogot. Nutrition and Cancer these compounds, there have been few studies linking the consumption of foods containing these com pounds to the development of cancer in experimental animals or humans. In 1946, Sulman and Sulman53 studied the carcinogenic effects on rats and mice of wood soot from the chimney of a smoked sausage fac tory. They fed the smoked sausage to rats for two years ad libitum, but none of the rats developed tumors. Rats implanted with soot subcutaneously or in jected with soot extracts did develop sarcomas. The authors gave no information on the quantity of sausage ingested or on other components of the diet. Dungal14 carried out a series of animal experiments in which rats were fed smoked mutton, smoked trout or salted cod fish. Malignant tumors developed in some of the rats fed the smoked mutton and the smoked trout, but none of those fed the codfish developed tumors. Analyses of BP showed that the mutton, trout and codfish con tained 1.3. 2.1 and 0.5 ppb, respectively. Dungal states that smoked mutton is consumed in consider able quantities, 1-2 lbs at a meal in Iceland. Codfish is smoked commercially for short periods of time, while the mutton and trout may hang in the smoke for weeks or months. On the basis of geographical distribution of cases of stomach cancer. Dungal states that "it is not only possible, but quite likely that the consumption of smoked meat, and eventually salted meat, may 6b of importance in this connection."14 In a review of carcinogens in the environment, Hueper*4 makes the point that animal experiments do not provide convincing evidence of the carcinogenic action of smoked foods since gastric cancer has not consistently been produced by feeding these foods to rats or mice, nor has it definitely been established that BP or other similar compounds are carcinogenic in man. Fark14 fed bacon to mice and obtained tumors in 8 to 20 of the animals but failed to confirm these find ings in another study. In a recent study, Hect and colleagues30 fed charcoal-broiled beef containing 52.7 ng/kg of BP to rats and humans and analyzed for BP in feces, using 14C labeled BP as the internal standard. Their study found that unchanged BP amounted to 11 % of the total dose in the feces of rats but was not detected in the human feces. The fate of BP in the human was uncertain, since the urine was not analyzed. Michels and Einbrodt31 did analyze human urine for BP and found an average of 0.546 pq/\ in urine from inhabitants of an industrial area and 0.416 *tg/l in urine from an uncontaminated area. It was not stated whether these differences were statistically sig nificant. It has been reported that vitamin A and certain of its analogues can prevent the carcinogenic activity of BP and some other polycyclic aromatic hydrocarbons. Chu and Malmgren* showed that 0.5% vitamin A plus BP or 7,12 dimethylbenzanthrene added to diets of hamsters prevented the development of gastrointestin al carcinoma which developed without the vitamin A. The work of Hill and Shin31 showed that BP is me tabolized to a product (BP-X) in the presence of mi crosomal mixed function oxidases from liver and lung tissue from rats and hamsters. The formation of this product is strongly inhibited by retinol and some of its derivatives. Vitamin A becomes more important in rela tion to gastric cancer when one considers that the diets of people in areas that have a high incidence of that cancer are very low in vitamin A.5*The interrelated ef fects of vitamin A-poor diets, consumption of carcin ogenic hydrocarbons and the incidence of cancer need further study. In summary, carcinogenic chemicals are present in many, if not most, of the foods we eat. Epidem iological, animal and chemical studies provide some evidence that the consumption of foods containing these compounds may be related to human cancer. Much research, however, is needed to improve our ability to predict the relative toxicity of these chem icals. Information is also needed on the amount of car cinogenic chemicals an individual must consume in order to show carcinogenesis and on the interrelated effects of other dietary constituents. References 'Ames, BN. McCann, J. and Yamasaki, E; "Methods for Detect ing Carcinogens and Mutagens With the Salmonella/ MammaiianMicrosome Mutagenicity Test." MutatRes 31, 347-364, 1975. lBhide. SV. Shivapurkar. NM, Gothoskar. SV. et al.: "Carcin- <'q ogenicity ot Betel Quid Ingredients. Feeding Mice With Aqueous Extract and the Polyphenol Fraction of Betel Nut." BrJ Career 40. 922-926, 1979. *Blum. SC. and Swarbrick, RE: "Hydroponic Growth of Crops in Solutions Saturated with ("C) Benzo (a) Pyrene." J Agnc Food Chem 28. 1093-1096, 1977. Blumer. M. and Youngblood, WW: "Polycyclic Aromatic Hydrocarbons in Soils and Recent Sediments." Science 180, 52- 55, 1975. Bogovskii, PA, Veldre, I, Itra. A, et al.: "Benzo (a) Pyrene in Fish of Estonian Waters." GigSamt 4,111-113, 1978. Brown, JP: "A Review of the Genetic Effects of Naturally Oc curring Flavanoids. Anthroquinones and Related Compounds." MutatRes 75,243-277, 1900. TChernichenko. IA, and Pavlova, NA: "Some Data on the Tran sport of Benzo (a) Pyrene H* From the Soil to the Plant. " In Slepyan. El (ed): Rast Khim Kantserogeny (Simp) 1st 1976. Len ingrad. USSR: Izd. Mauka, 1979, pp. 89-90. Chu, EW, and Malmgren, RA. "An Inhibitory Effect of Vitamin A on the Induction of Tumors ot Forestomach and Cervix in the Syrian Hamster by Carcinogenic Polycyclic Hydrocarbons." Cancer Res 25.304-895. 1965. Commoner, B, Vithayathil. AJ. Dolara. P, et al.: "Formation of Mutagens in Beef and Beef Extracts During Cooking." Science 20, 913-916. 1978. 'Cook. P: "Cancer of the Oesophagus in Kenya. A Summary and Evaluation of the Frequency o' Occurrence and a Preliminary Indication ot the Possible Association With the Consumption of Al coholic Drinks Made From Maize." BrJ Cancer 25. 853-879, 1971. "Crosby, NT, Foreman. JK. Palframan, JF, et al.: "Estimation of Steam-Volatile N-Nitrosamines in Food at the M/kg Level." Nature 283.342, 1972. URL 0391 Vol.2, No. 4 Ouphcstion of Nutrition and Cancer. in wttoio or in port by on? moon* for purpo-- i* 221 "Deinzer, ML, Thomson, PA. Burgett. DM, et al.: "Pyrrolizidine Alkaloids: Their Occurrence in Honey From Tansy Ragwort (Senecia jacobaea L.)Science 195, 497-499. 1977. "Dolara, P, Commoner. B. Vithayathil, A, et al.: "The Effects of Temperature on the Formation of Mutagens in Heated Beef Stock and Cooked Ground Beef." Mutation Res 80. 231-237, 1979. "Oungal, N: "The Special Problem of Stomach Cancer in Iceland." JAMA 25. 789-793, 1961. "Dunn, BP. and Fee. J: "Polycyclic Aromatic Hydrocarbon Carcinogens in Commercial Seafoods." J Fish Res Board Canada 38. 1469-1476, 1979. Fark, G: "Carcerogene Wirkung von Gerauchertem Spek." Ztschr Krebsforsch 56. 583-586. 1950. 17Fong, LYY. Ho. JHC. and Huang, DP: "Preserved Foods and Possible Cancer Hazards: WA Rats Fed Salted Fish Have Mutagen ic Urine." intJCancer23. 542-546. 1979. "Fritz, W, and Engst, R: "New Results on the Environmentally Caused Contamination of Foods With Cancer Inducing Hydrocar bons." Tech Umwettschutz 10. 140-172, 1975. "Fretheim, K: "Carcinogenic Polycyclic Aromatic Hydrocarbons in Norwegian Smoked Sausages." J Agric Food Chem 24. 976979, 1976. "Hecht, S. Grabowski. W, and Groth, K: "Analysis of Feces for Benzo (a) Pyrene After Consumption of Charcoal-Broiled Beef by Rats and Humans." Fed Cosmet Toxicol 17, 223-227, 1979. "Hill, DU. and Shih. T-W: "Vitamin A Compounds and Analogues as Inhibitors of Mixed-Function Oxidases That Me tabolize Carcinogenic Polycyclic Hydrocarbons and Other Com pounds." Cancer Res 34, 564-570. 1974. "Hirokado, M. Endo, F Usami, H, et al.: "Polynuclear Aromatic Hydrocarbons Benzo (a) Pyrene in Food Additives. II. Liquid Paraf fin and Liquid Smoke Flavor." Tokyo Toreisu Eisei Kenkyshua KenkyuManpo 26. 220-224. 1975. "Howard, JW. and Fazio, T: "A Review of Polycyclic Aromatic Hydrocarbons in Foods." J Agric Food Chem 17. 527-531, 1969. "Hueper, WC; "Carcinogens in the Human Environment." Arch Pathol 72, 355-380, 1961. "Kapadia. GJ. Paul. BD, Chung. EB, et al.: "Carcinogenicity of 'Camellia senensis' (Tea) and Some Tannin-Containing Folk Medicinal Herbs Administered Subcutaneously in Rats." JNCI 57. 207-209, 1976. Lijinsky, W, and Shubik, P: "Polynuclear Hydrocarbon Carcin ogens in Cooked Meat and Smoked Food. " Ind Med Surg 34. 152154, 1965. "Lijinsky, W. and Shubik. P: "Benzo (a) Pyrene and Other Poly nuclear Hydrocarbons in Charcoal-Broiled Meat." Science 145. 53-55, 1964. Malanoski, AJ, Greenfield, EL. Barnes, CJ. et al.: "Survey of Polycyclic Hydrocarbons in Smoked Food." JOAC 51. 114-121, 1968. "Masua. Y. and Kuratsune. M: "Polycyclic Aromatic Hydrocar bons in Smoked Fish. Katsuobushi." GANN 62, 27-31, 1971. Medve, F: "Air Pollution of the Ecosystem Area on an Oakwood and Natural 3.4-Benzopyrene Content of the Soil and Plant Samples in the Ecosystem." Acta Bot AcadSciHung 22. 393-402, 1976. "Michels, S. and Einbrodt, HJ: "Polycyclic Aromatic Hydrocar bons in Human Urines Collected in a Large Industrial City-An Epidemiological Study." Wiss Umwelt. 107-112, 1979. "Miller. EC, and Miller, JA: "Naturally Occurring Chemical Carcinogens That May be Present in Foods." Internal Rev Biochem 27, 123-165, 1979. "Miyabe, M, Asanoma. M, and Sakabe, Y: "Survey of 3,4 Ben zopyrene in Foods on the Market." Nagoya-shiEiseiKenkyusho Ho 22, 54-58, 1976. Nagao, M. Takahashi. Y, Yamanaka, H. et al.: "Mutagens in Coffee and Tea." MutationRes 58. 101-106,1979. Nagao, M. Honda, M. Seino. Y. et al.: "Mutagenicities of Smoke Condensates and the Charred Surface of Fish and Meat." Cancer Lett 2. 221-226. 1977, Pamukcu. AM, Wang. CY, Hatcher, J, et al.: "Carcinogenicity of Tannin and Tannin-Free Extracts of Bracken Fern (Pteridium aquilinum) in Rats." JNCI&S, 131-136,1980. "Potthast, K: "Polycyclic Aromatic Hydrocarbons in Smoked Meat Products: An Application of a New Method." Acta Aliment Pot %, 195-201, 1977. Pradhan, SN. Chung, FB. Ghosh, B, et al.: "Potential Carcin ogens. I. Catcinogenicity of Some Plant Extracts and Their TanninContaining Fractions in Rats." JNCI 52.1579-1582, 1974. Price. JM, and Pamukcu, AM: "The induction of Neoplasms of the Urinary Bladder of the Cow and Small Intestines of the Rat by Feeding 8racken Fern (Pteris aquilina)Cancer Res 25. 2247- 2251, 1968. Ranadine, KJ. Randive, SN, Shivapurkar. NM. et al.: "Betel Quid Chewing and Oral Cancer: Experimental Studies on Ham sters." IntJCancer 24.635-843. 1979. "Rappaport. SM. McCartney, MC. and Wei. ET.: "Volatilization of Mutagens From Beef During Cooking." Cancer Lett 8, 139-145, 1979. Scanlan, RA. Barbour, JF, Hotchkiss. JH, et al.: -N-Nitrosodimethylamine in Beer." Fed Cosmet Toxicol 18, 27-29, 1980. Schoental. R: "Relationships of Fusarium Mycotoxins to Dis orders and Tumors Associated With Alcoholic Drinks." Nutr Cancer 2.88-91, 1980. Side!. K. and Happel. H: "Effect of Refuse Compost on 3.4 Benzopyrene Content in Carrots and Head Lettuce." Naturwissenschaften 62. 300. 1975. Shiraishi, Y, Shirotori. T. and Takabatake. E: "Determination of Polycyciic Aromatic Hydrocarbons in Foods. II. 3,4 Benzopyrene in Japanese Daily Foods." ShokEisZass 14,173-178.1973. "Shiraishi. Y. Shirotori. T. and Takabatake. E: "Determination of Polycyclic Aromatic Hydrocarbons in Foods. III. 3.4 Benzopyrene in Vegetables." ShokEisZass 15. 18-21,1974. ^Shiraishi, Y. Shirotori, T. and Takabatake. E: "Determination of Polycyclic Aromatic Hydrocarbons in Foods. IV. 3,4 Benzopyrene in Fish and Shellfish." ShokEisZass 18. 178-181. 1975. Spiegelholder, B. Eisenbrand. G. and Preussman, R: "Con tamination of Beer With Trace Quantities of N-Nitroso dimethylamine." Fed Cosmet Toxicol 17, 29. 1979. Spingarn, NE, Kasai, H, Vuolo, LL, et al.: "Formation of Mu tagens in Cooked Foods. III. Isolation of a Potent Mutagen From Beef." Cancer Lett 8, 177-183, 1980. Strobel, RGK: "Determination of 3,4 Benzopyrene in Coffee Products." CofloqIntChimCafes6, 128-134, 1974. "Sugimura, T, Nagao. M, Kawachi. T, et al.: "MutagenCarcinogens in Food With Special Reference to Highly Mutagenic Pyrolytic Products in Broiled Foods. " In Origins of Human Cancer, Cold Spring Harbor Laboratory Symposium. New York: Cold Spring Harbor Laboratory, 1977, pp. 1561-1577. Sugimura. T, Kawachi, T, Nagao, M, etal.: "Mutagenic Princi ple (s) in Tryptophan and Phenylalanine Pyrolysis Products." Proc Japan Acad 53. 58-61, 1977. "Sulman. E, and Sulman, F: "The Carcinogenicity of Wood Soot From the Chimney of a Smoked Sausage Factory." Cancer Res 5. 366-367, 1946. Toth. L. and Blaas, W: "Carcinogenic Hydrocarbon Contents in Grilled Meat Products." Fleischwirtschaft 53, 1456-1459. 1973. "Vetdre, J. Raher, M, lira. AR. et al.: "Content ot Benzo (a) Py rene in the Soil and Plants of Estonia." In Slepyan. El (ed): Past URL 03916 222 OvpffMrfeft o'Nutrition and Caneor, m or m parr by tny moans for any purpoM 11 iffopof. Nutrition and Cancer L Ktim Kantserogeny (Simp) 1st 1976. Leningrad, USSR: Izd Nauka, 1979. **Weisburger. EK: ``Natural Carcinogenic Products." Environ$a Tech 13, 278-281, 1979. S7Wynder, EL. Kmet. J. Oungal, N. et al.: "An Epidemiological In vestigation of Gastric Cancer." Cancer 18. 1461-1493. 1963. ^amaguchi. K. Zenda. H. and Shudo. K: "Ultramicro Analysis of Benzo (a) Pyrene in Food." Chem Pharm Bull 26, 3600-3602. 1978. "Yamaizumi. Z. Shiomi, T, Kasai. H. et al.: "Detection of Potent Mutagens Tryp-p-l and Tryp-p-2 in Broiled Fish." Cancer Lett 9, 75-83. 1980. URL 0391 7 Vol.2, NO. 4 OufMc4tiA of Nutrition Cancsr. in or in pmn by any mmmnm for fry purpo-- m illegal. 223 Nutrition and the cancer patient C \ S'U'A- 2. Food and the aetiology of cancer J. W. T. Dickerson, PhD, reader in human nutrition, Jepertment ofbiochemistry. University ofSurrey, GuMdford NOTiCCs This mamnnl may *Ml 5rrV eprtiBt low (Ttto It UA. Cod#> URL 03918 has been suggested that about 90 per cent of human ameers are caused by chemical substances, the remainder bong caused by viruses and radiation. There are three feasible routes by which exogenous chemical carcinogens say enter the body. Thus, they'may be absorbed through he skin, from the lungs, or from the gastrointestinal tract, ft is this latter route that we are concerned with here. Of potentially carcinogenic compounds that may enter the body in food, some are intended additives such as colourings ac flavourings, while others are accidental contaminants derived, for instance, from pw^fgjng materials or from o2s used to lubricate machinery. Still other chemical carcinogens that may be found in food are mycotoxws, such as-afi-atoxin, which result from the contamination offood by micro-organisms. A further possibility is that carcinogenic Compounds may be formed in the intestine from otherwise innocuous constituents of food by the action of intestinal flora. |[ Chemical carcinogens that enter the body in food, or tie formed from compounds that do so, may be involved in & aetiology of cancer in only certain organs, "the target organs". Thus, they may be involved in cancer ofthe gastro , intestinal tract, the bladder, or the liver. A consideration - a the incidence of cancer in different parts of the g, *xo- unestinal tract (Table 1) shows chat the incidence is higher isjthoae parts of the t ct in which the luminal cements are 3fiy to remain for a longer time in contact with the epithelial lining. In the intestine the is highest in shortest part of the tract, the rectum, where the contents ormoved only at the time of evacuation. A longer contact 3ox also accounts for the higher incidence of cancer of the aamach in the prepyloric region, and of cancer of the tal tract at the base of the bladder. Primary cancer liver may also result from its function as a storage of the stomach Ipidemiological studies have shown that cancer of the aonach is about twice as common in men as it is in women, ft:-incidence in the general population varies considerably nm one pan of the world to another. For instance in 958-59 it was about 70 per 100,000 of the population in ipanese men whereas the comparable value for this country about 30; among the white population in the United tites it was about 15. Such variations in geographical ffrxibution suggest that an exogenous environmental iw is involved, and a dietary one has been sought. In ne countries the dietary habits may be so complex as to ike it impossible retrospectively to identify a dietary Vtor. Japan, however, with its generally high inridence of inter of the stomach together with an uneven distribution a the country appears more promising. One study in Japan, * fret, has suggested that the high incidence ofcarcinoma ihe stomach in certain areas is associated with the ^sumption of larger amounts of salty food and smaller fcoonnts of milk. The occurrence of a carcinogenic polyr i cyclic hydrocarbon, 3,4 benzpyrene, os and in smoked and roasted foodstuffs such as flsh or coffee, has also attracted attention. Similar compounds may also be present in fried food. 3,4-benzpyrene was isolated from pitch in 1931 at the Royal Cancer Hospital. About years before this, in 1875, von Volkmann at Halle in Germany, had described occupa tional skin cancers among workers in the tar and paraffin industry and studies oo experimental animals had shown that repeated application of coal tar to the skin caused cancer. It is interesting to note that this same compound is also present in tobacco and in the polluted atmosphere of many of our cities. Its presence in die air and particularly in cigarette smoke is thought to be responsible for cancer of the lung. Thus, the same carcinogen, 3,4-benzpyrene,* nay enter the body by two different routes and cause cancer in two different sites, the stomach and the lungs. Food additives (a) Inumional: A wide range of compounds are intentionally added to foods for various' purposes such as colouring, flavouring, sweetening, and as preservatives and emulsifiers. These compounds are submitted to very stringent tests to make reasonably certain thai' nose of 'hem is potentially carcinogenic. It is interesting to note that one compound that might have been used as a colouring agent for butter, "butter yellow" (4-diniethylamino-azobenzene) was shown in 1936 to be a very effective liver carcinogen in rats! * Sodium nitrate may be used as a preservative for fish and meat. Some years ages an outbreak of severe liver disease in sheep in Norway was traced to the consumption of fish ttwI preserved with sodium nitrite. This compound had reacted with one called dimethylamine /which had been formed as the result of slight decomposition of the fish, to form dimethyl-nitrosamine, which is a potent liver toxin that eventually produces liver cancer. The use of cyclaxnstes as artificial sweeteners in foods was banned in 1969. Tins ban followed a demonstration that it is changed to cydohexylamine in some individuals and accreted as such in the urine. Other work had shown TABLE 1. Distribution of canams of the gnatro-intasthial tract (Adopt*ifnm Hu*p*r and Conway, 1964) Stomach -- prepylork region Small intestine Duodenum jejunum-ileum Large immune Caecum colon Sigmoid Rectum Parcentage f oancera 60% of all gastric cancers 2.5-10% of intestinal cancers 2.0- 3% of intestinal cancers 0.5-7% of miestinal cancers 90.0- 973% of intestinal cancers 30.0- 35.0% of intestinal cancers 20.0- 25-0% of intestinal cancers 40.0- 45.0% of intestinal cancers URL 03919 Um o# tfcn teM ealtpM to mooturo nhrittonw tat that cyclohexylaraihe produced cancer f the bladder in the rat. It should be emphasised, however thar there has been no evidence whatever that cyelamatesfeve caused cancer in humans. (b) Unintentional; A variety of potestU carcinogens such as herbicides, insecticides, fertilisers, onbiorics, detergents, metals, besides products derived from fungi and moulds, zsay contaminate human food. One ofshe most widely used insecticides, DDT, was shown to predate hepatomas after prolonged feeding to rats in amounts greatly in excess of those likely to be consumed by man. However, DDT and other chlorinated pesticides pass ia milk and therefore even infants are likely to have some degree of exposure to these compounds. It is also known itat these compounds accumulate in various tissues, partkadariy fat tissue. Both natural and synthetic oestwguu are known to be potent carcinogens in a variety of ^ecies, and to cause cancer in a variety of organs such as the breast, uterus, etc In women, cancer of the breast ^ten regresses after ovariectomy, adrenalectomy or bgpophysectomy. It is therefore evident that hormones can stimulate the growth of a breast cancer, even if they cans* actually induce it. It is possible that women may receive oogenous hormones of the same type in food, for they are some times implanted in domestic animals as growth proaaten. It is not incon ceivable, therefore, that in such ascumstances food may play a role in the growth of human cancer. In 1960 a large number of turkeys died as the result of a liver disease called Turkey X. The xmue of this disease was traced to the Brazilian groundnuts that had been used as mash for feeding the birds. The wma were contaminated with a mould, Aspergillus flaws, and subsequent work showed that this mould produces a^oup of extremely toxic substances, which were called aflatomes. Many investigations have now been done on these sttetances, and they have been shown to be carcinogenic in several animal specie*. There is no direct evidence that they are carcinogenic a man, although indirect evidence would suggest that they are. Thus, in comparison to the peoples of the Western * world, the incidence of primary hepatoma is high among the negriform races, and particularly those in Africa. It a not inconceivable that at least some of this higher incidence is the result of the ingestion of aflatoxins. These substances have been reported, for instance, as a contaminant ofmaras meal in Ethiopia. The incidence of hepatoma in is among the highest in the world, and in that country fermented peanut press cake, a popular fir-d and a good source of dietary protein, has been found to contain aflatoxins. !^ It is important to realise that both generic and nutritional factors may interact with mycotoxins, or other toxins of plant origin. In developing countries where people have been living on a low protein diet, an increase in the protein intake may act as a kind of catalyst for the induction of cancer in a precancerous dormant liver. The problem may be still further complicated by the fact that the additional protein ingested may have come from cereals or other products that are readily contaminated with mycotoxins. There is evidence, too, from animal studies, that hepato- toxins such as aflatoxin may show their maximal carcino genicity when given with a diet that is deficient of pyridoxin (vitamin B6). Thus, it could be reasoned that the high incidence of primary liver cancer in African groups implies that the diets are deficient in pyridoxin, or alternatively, contain antagonists to pyridoxin. lntrahmanal formation of catcinogms V Recently, attention has been focused on the possibilfr that carcinogens may be formed in the intestine as tb< .wu, M> JUUC Ifii > result of ingesting dietary constituents that are themselves h innocuous. An epidemiological study of environmental factors that may be responsible for cancer of the colon and rectum in Japanese living in Japan has shown that cancer of the colon is significantly less common in Japan than in the United . States, whereas the incidence of cancer of the rectum is similar in both countries. Moreover, Japanese patients with cancer of the colon were found to come from a higher socio-economic group of the population than those with rectal cancer, and this higher socio-economic background was associated with the eating of a more Western-type diet. These workers suggested that the proportion of fat in the diet might affect the pathogenesis of cancer of the colon through the action of the bacterial flora. A study ofthe effect of vegan diets on bacterial flora and faecal steroid concentrai don has led Dr Frey Ellis and his collaborators to postulate a similar correlation between diet in Western countries and the relatively high incidence of cancer of the colon. The incidence of cancer of the breast is higher in North America and North-West Europe than it is in Africa, Asia and South America, and within a cc-antry is highest, like that of cancer of the colon, in people of the higher socio economic classes. This led Wynder to postulate that there is a relationship between dietary fat and breast cancer. He .suggested that dietary fat may affect hormone production ;or the mention of hormones by adipose tissue. The effect may, however, be mediated by the action of the gut flora on bile steroids resulting in the formulation of oestrogens. This work can only be considered at the moment as suggesdve for we would like to know, among other thing*, the potency of the particular oestrogen produced by the bacteria as compared with the normally occurring oestrogens. 35 It may be that da. type of fat in the diet is also important. Dietary saturated fkts have been incriminated in the aetiology of arhcwrlcrosis, and the results have been reported of an eigtc-ycar controlled el*cai trial in which men had been assigned randomly to a conventional diet or to one similar in A respects except for a substitution of vegetable oils, rich in polyunsaturated fats, for saturated fat. Fatal heart diaase was found to be more common in the men having coeremional diet, whereas cancer was more common in the gtmp having vegetable oils. Further work is obviously neededlefore too much is made ofthese results, but taken in conjimetion with what has been said about the possible relahonsh^pof dietary fat to cancer of the colon and breast, it may not he too early to advise certain patients to reduce their fax snake and particularly the intake of polyunsaturated fim. Attention has amend? been drawn to the similarity in the epidemiology afT diverticulitis and carcinoma otf the colon. The incidence of diverticulitis has apparently increased over the gut 100 years or so, and it has been suggested that this increase is associated with the ingestion of a diet rich in rrfnd' carbohydrate and consequently with a low roughage conat It is always sad that "prevention is better than cure'*. Preventive measim have greatly reduced the incidence of infectious diwn, including tuberculosis. It .may be that at least a oartaf the answer to the cancer problem lies in more strenuou measures to reduce our dietary intake of potentially cardnapnic chemicals. The comparatively recently appreciate possibility that even the "0011081" constituents of oar diet in certain circumstances may play a role in caiiJaagt m, iii, is one which opens up a whole new area for invesapDon. Next meek: Mttabotic effects ofcancer Horrrone treatment instead of prostatectomy? Penign prostatic hypertrophy . or prostate medrogestone adMily has any effect on the enlargement is a common condition in males antibacterial fac*te found in prostatic and over 50. seminal fluids. Ha is important, to avoid In one in five, the prostate enlargement is compromising the patient's natural defences severe enough to obstruct the urine against bacteriaiinfKtions. passageway, resulting in inability to empty the Esriier research found that seminal and \> bladder. The most common treatment for this prostatic fluids ontain a potent antibiotic ? , k ' condition is prostatectomy, a simple but costly surgical procedure for removal of the prostate. Thousands of these operations are performed every year. A new treatment, still in the experimental stage -- the result of which may play a significant role in the development of a new, substance which ftfltb bacteria responsible for urinary tract infective*. This discovery not only accounted for the absence of urinary tract infection in healthy males, but explained how the body fought off Bacteria which caused such infections. non-surgicai treatment of the prostate -- may If further study entirms that medrogestone is replace the now common surgery. A female effective in humans and has no effect on the hormone, medrogestone, given orally or by antibacterial factor is prostatic fluid, prostate injection, shrinks the diseased prostate, making surgery for this bavin disease may become a an operation unnecessary. thing of the past. Research urologists in the USA believe that if Additionally, the IK research urologists plan clinical trials demonstrate its effectiveness, to study the bioche--aal factor in prostatic and medrogestone may become the treatment of seminal fluids, and at- clinical importance as a choice for this disease in the future. This defence mechanise against urinary tract particular hormone appears to be free of the infections; they alee intend to investigate the undesirable side-effects associated with possibility that the yarv-urethrai glands of the female hormone treatment female may secrete a. similar substance, and Before recommending hormone treatment for that this agent is re^nstbie for the absence of 2 the prostate on a routine basis, however, the harmful bacteria on the urethra and external research urologists want to know if vagina of normal femribs. L X) <UrLoD-J Q The mutagens that cooking produces John S. Wishnok -JU-si4r- c? ^ ^~(_ t - f URL 03921 Can carcinogenic substances be formed during the cooking of food? The answer is yes. Experimentally, activity is found in simple extracts of surface char or smoke condensates (1). However, the greatest specific activity is usually found in the extract of a basic fraction of tfib char or smoke condensate. Substantially lower mutagenic activities are observed if NaOH is used in the work-up rather than NH4OH (2). The origin of this effect has not yet been elucidated, although it may be due to the formation of mutagenic nitrogen heterocycles via condensation reactions of ammonia with carbonyl-containing pyrolysis products (3,4). Much of the early work was done with NH4OH, and in some cases the base was not specified. Thus the quantitative aspects of all these reports remain questionable. Soon after Ames developed mutagenicity assays via bacterial tester strains, reports began to appear noting mutagenic activity in crude extracts of smoke condensates or charred surfaces obtained from broiled fish or meat (5), and these early reports have been followed by a steady flow of related publications. Much of this research can be loosely divided into three general areas: screening of various types of cooked food for mutagens or for compounds that alter mutagenic activity chemical characterization of the mutagens biochemistry of the food mutagens (cancer initiation, metabolism, adduct formation, etc.). Following the initial experiments with cooked whole fish and meat, several food components or model systems were studied. Smoke condensates from the pyrolysis of lysozyme, histone, DNA, RNA, starch, and vegetable oil were tested. The greatest activity was found in the protein pyrolysate following microsomal activation (6). These observations, along with solubility and chromatographic behavior, indicated that the mutagens were organic bases. This in turn suggested that proteins might be major precursors of the mutagens, and a series of amino acids were then pyrolyzed. The smoke condensate from tryptophan showed high Ames test activity. Two mutagens (1 and 11) arising from tryptophan and one from phenylalanine (III) have been reported (7). m To date, a wide variety of foods have been tested for mutagenic activity, but beef has received the most attention, followed by fish, with noticeably less attention having been paid to other foods. Benzo(<z)pyrene is formed during the grilling of beef, pork, lamb, and turkey (5). Most of the recent research, however, is focused on compounds other than benzo(u)pyrene, which are formed under generally milder cooking procedures. Such mutagens have been observed, for example, in beef extract after refluxing for several days or from frying ground beef at surface temperatures of less than 250 C (9-13). When beef is cooked isothermally, there appears to be an induction period prior to which there is relatively low mutagenic activity (13). A sharp rise in activity occurs between 140 C and 180 C (12). When beef patties are fried between 140 C and 250 C as much as 99% of the mutagenic activity appears in the cooking vapors rather than on the meat surface (13). Interestingly, the total mutagenic activity obtained from the broiling of ground beef is increased by adding proline to the meat before cooking. Of 20 common amino acids tested, only proline had this effect (14). To date, only one of the beef mutagens, Trp-P-1, has been unequivocally characterized (15). The homologous mutagens IQ and methyl-IQ have been isolated in broiled (16) and roasted (17) fish. 348 CHEMTECH JUNE 1984 Y NH, of* CH, Me-IQ Pyrolyses of glutamic acid (18, 19) and lysine (19) gave additional mutagenic nitrogen-containing heterocycles, and provided a starting point for more systematic screening of the mutagens produced from cooking foods. Even though food mutagens were produced in extremely low concentrations, it was possible to follow the mutagenic activity of various fractions and to compare the fractionation characteristics with those of the known amino acid pyrolysis products. These comparisons, however, were often frustratingly negative. The high-pressure liquid chromatography (HPLC) of the major mutagenic fraction obtained from broiled sardines, for example, differed from all the amino acid pyrolysis mutagens known at the time, as well as from some structurally related nonmutagens (20). At least one of the mutagens obtained from broiled squid, on the other hand, has the spectral and chromatographic properties of a mutagen (Glu-P-2) obtained from the pyrolysis of glutamic acid (21). Glu-P-2 Mutagenic activity has been detected in bakery and cereal products (22, 23), coffee and tea (24), and Japanese pickles (25), but the mutagenic compounds have only been characterized in Japanese pickles where they were shown to be flavonoids (25). At least one author has stated that at least some mutagenic activity could be observed in all types of food that were tested (23). Mutagens in model systems Because product mixtures from foods are chemically and biologically complex, most of our information comes from CHEMTECH JUNE 1984 349 IJR I03922 mytn V studies of various model systems. Tryptophan pyrolysate was the most mutagenically active, but 11 of 20 pyrolysates of common amino acids gave positive Ames tests (18). The mutagens of interest in all cases were distinct from benzo(a)pyrene, although many amino acids do yield benzo(a)pyrene upon pyrolysis, usually at somewhat higher temperatures (650-850 C) than would be used for cooking (26). The next highest mutagenic activity among common amino acids, following pyrolysis, occurred with glutamic acid. Again, two mutagenic components were isolated from the pyrolysate and characterized as Glu-P-1 and Glu-P-2. GIu-P-1 In most cases in which mutagens have been identified, the compounds are isolated in pairs differing only by a methyl group. Pyrolysis of proteins also yields mutagens. There is reasonably persuasive evidence for the presence of Glu-P-2 in casein pyrolysates, but only about 10% of the mutagenic activity of the pyrolysate could be attributed to Glu-P-1 and Glu-P-2 (27). The amount of Trp-P2 in various protein pyrolysates roughly corresponds to the percent of tryptophan in the protein (28). Soy globulin also produced two additional mutagens, which have also been shown to be present in the tryptophan pyrolysate (28-30): These compounds are much weaker mutagens than Trp-P-1 and Trp-P-2. Note that they too constitute a methylated-nonmethylated pair. Mutagenically significant non-mutagens All of the mutagenic compounds that have been characterized are nitrogen-containing aromatic heterocycles with an exocyclic a-amino group. Some structurally related pyrolysis products (e.g., harman and norharman) without the exocyclic amino group have also been observed (31). Neither harman nor norharman is mutagenic, but they enhance the activity of several mutagens (32-34). Norharman increases the mutagenicity of 4-dimethylaminoazobenzene 40-fold, while harman has no effect. Harman does, however, enhance the mutagenicity of 2-aminoanthracene, 2-aminofluorene, and 2-acetylaminofluorene (34) while it lowers the mutagenicity of several polycyclic aromatic hydrocarbons such as 7,2-dimethylbenzanthrene, benzo(a)pyTene, chrysene, and dibenzo{<j)anthracene (33). The complexity of this picture is increased by the observation that mixtures of norharman and aniline or norharman and o-toluidine are mutagenic even though none of these compounds are mutagenic when tested alone. m-Toluidine and p-toluidine remain nonmutagenic in the presence of norharman (35). It's not known how norharman and harman elicit these effects, although it has been shown that these compounds interact with DNA. Harman apparently interacts more strongly with DNA, yet norharman is the more effective co-mutagen (36). There is evidence that mutagens can influence each other's behavior (27-29,37). For example, various mixtures of four of mutagens obtained from tryptophan pyrolyses exhibit synergistic activity (35). We found that the mutagenic activity of the basic fraction from boiled beef extract with added IQ is essentially the sum of the two activities (38a), while the addition of Trp-P-2 to the basic fraction depresses mutagenicity (38b). Metabolism and structure--activity variations Despite the complexity, patterns of metabolism and of some aspects of structural effects of mutagenic activities are beginning to emerge. It was shown fairly early, for example, that the exocyclic a-amino group is necessary for mutagenic activity, and that three-ring heterocyclic systems are more mutagenic than two-ring systems. The fusion position also has a pronounced effect on the relative potency (39). The presence or absence, as well as the position, of methyl groups also affects the mutagenicity of the pyrolysis mutagens. Glu-P-1 is a stronger mutagen than its six methylated isomers or the nonmethylated Glu-P-2 (40). Trp-P-2 is a stronger mutagen than the dimethyl Trp-P-1, or the monomethyl isomers (41-43). Mutagenic activity roughly parallels the degree of noncovalent interactions with DNA as measured by fluorescence quenching and equilibrium dialysis (41, 42, 44). Similar results were obtained for Glu-P-1 and Glu-P-2 (45). In these experiments, as well as a variety of others, activation by induced liver microsomal fractions is required (46). For Trp-P-2, binding to DNA occurs following treatment with either microsomes or cytosol; NADPH and ATP are required. The degree of binding depends on the nature of the inducer (47-49). Both Trp-P-2 and Glu-P-1 URL 03923 u , > ' KSEO Ittfi systems, these compounds appear as homologous pairs differing only by the presence of one methyl group. The relative mutagenic potency within a series of these mutagens is strongly affected by substitution, but no general rational pattern of structure-activity relationships has been developed. The mutagenicity of mixtures of these compounds often differs from that expected on the basis of individual potencies, but no rational pattern for this behavior is evident. Nonmutagenic compounds, which enhance the activity of mutagens, are also formed during cooking procedures. All of the cooking-related mutagens that have been studied require microsomal activation in order to exert their mutagenicity. This microsomal activation also requires induction of the liver microsomal systems. The most effective compounds are known inducers of mixed-function oxidases. The active metabolites are apparently the N-hydroxylated derivatives at the exocyclic amine. Adducts at the Cg position of guanine have been ' form Cg-guanine derivatives when treated with microsomes characterized. ` in the presence of DNA, apparently via the N-hydroxyl Some of these compounds have been individually shown derivatives (45, 50-52). In the case of Trp-P-2, four to be carcinogenic in laboratory animals but it is not clear metabolites, two of which are direct-acting mutagens, can how they may behave when administered in be observed via HPLC. The major mutagenic metabolite combination. ' from Trp-P-2 was isolated and characterized as the 3-amino hydroxy derivative. Acknowledgments Carcinogenicity of food mutagens Some of these mutagenic food pyrolysis products also have carcinogenic activity in laboratory animals, although their potency is much less than that of 3-methylcholanthrene (53-55). Both Trp-P-1 and Trp-P-2 produce chromosome aberrations, with Trp-P-1 being 5-10 times more effective than Trp-P-2 (57, 58). Because of the complex and unpredictable synergism among these compounds and the fact that human exposure to them will invariably be part of a complex food matrix, it is probably premature to attempt to assess the potential importance of these findings to human cancer epidemiology. Summary Research on food mutagens is still developing and many current results are contradictory or confusing. Nonetheless, several conclusions are emerging from this work: Strongly mutagenic compounds can be formed from most types of food during normal cooking procedures. The mutagens that have been characterized all appear to be of the same structural type, i.e., 3-ring nitrogen-containing heterocycles with an exocyclic amino group on a carbon adjacent to a ring nitrogen. In model This investigation was supported by PHS Grant No. 1-P01-CA26731-03, awarded by the National Cancer Institute, DHHS, and by Grant No. 2-P01-ES00597-11, awarded by the National Institute of Environmental Health Sciences. References (1) McCann, J. CHEMTECH 1978, November. 682. (2) Iwaoka, W. T. et al. Cancer Lett. 1981,11, 225. (3) Yoshida, D.; Okamoto, H. Agric. Biol. Chem. 1980,44, 2521. (4) Yoshida, D.; Okamoto, H. Biochem. Biophys. Ret. Commun. 1980, 96 844. (5) Nagao, M. et al. Cancer Lett. 1977,2, 221. (6) Nagao. M. et al. Cancer Lett. 1977,2, 335. (7) Sugimura, T. et aL Proc. Jpn. Acad. 1977,53, 58. (8) Doremire, M. E.; Hannon, G- E-; Pratt, D E. J. Food Set 1979,44, 627. (9) Commoner, B. et al. Science 1978,201, 913. (10) Spingam, N. E.; Weisburger, J. H. Cancer Lett. 1979,2, 259. (11) Dolara. P et al. Mutat. Ret. 1979, 80, 231. (12) Pariza, M. W. et al. Cancer Lett. 1979,7, 63. (13) Rappaport, S. M.; McCarthy, M. C-; Wei, E T. Cancer Lett. 1979, 8,139. (14) Ashoor, S. H. et al. Life Sci. 1980,26, 1801. (15) Yamaguchi, K. et al. Gann 1980,71,745. (16) Kasai, H. et ai. Proc. Jpn. Acad. (B) 1980,56, 278. (17) Kasai, H. et aL Seikagaku 1980,52, 804. (18) Yamamoto, T. et al. Proc. Jpn. Acad. 1978,54,248. (19) Kosuge. T. et al. Chem. Pnarm. Bull. 1978,26,611. (20) Kasai, H. et al. Cancer Lett. 1979,7, 343. (21) Yamaguchi, K. et al. Gann 1980, 71, 743. (22) Pariza, M. W; Ashour. S. H.; Chu, F. S. Food Cosmet. Toxicol. 1979, 17. 429. CHEMTECH JUNE 1984 351 (23) Spingam, N. E , Slocum, L. E-; Weisburger, J. H. Cancer Lett. 1980, 9. 7. (24) Nagao, M. et af. Mutat. Res. 1979,68, 101. (25) Takahashi, Y. et ai. Mutai. Ret. 19T9,88, 1 IT. (26) Patterson, J. M. et al. J. Agric. Food Chem. 1978,26, 266. (27) Yamaguchi, K. et ai. Gann 1979,70,849. (28) Yoshiaa. D ; Matsumoto, T.; Nishigata, H. Agric. Biol. Chem. 1980, 44,253. (29) Yoshida, D.; Matsumoto, T. Agric. Biol. Chem. 1979,48,1155. (30) Yoshida, D.; Nishigata, H.; Matsumoto, T. Agric. Biol. Chem. 1979, 43, 1769. (31) Nishigata, H.; Yoshida. D.; Matsumoto, T. Agric. Biol. Chem. 1980, 44, 209. (32) Nagao, M. et al. Cancer Lett. 1977,3, 339. (33) Matsumoto, T.; Yoshida, D.; Mizusald. S. Mutat. Res. 1977,56,85. (34) Umezawa, K. et al. Proc. Natl. Acad. Sci. USA 1978,75, 928. (35) Nagao. M. et al. Proc. Nat. Acad. 1977,53. 34. (36) Hayashi, K.; Nagao, M.; Sugimura, T. Nucleic Acids Res. 1977, 4, 3679. (37) Yoshida, D.; Matsumoto, T.i Okamoto, H. Mutat. Res. 1979, 68, 175. (38) (a) Turnkey, R. J. etai. Carcinogenesis 1983,47,863. (b)Tureskey. R., unpublished observations- (39) Matsumoto, T. et al. Agric. Biol. Chem. 1978,42, 861. (40) Takeda, K. et aL Carcinogenesis 1980,1, 889. (41) Pezzuto, J. M.; Moore, P. D ; Hecht, S. M. Biochemistry 1981, 20, 298. (42) Nagao, M. etal. Carcinogenesis 1980,1, 451. (43) Hasnimoto, Y., Shudo, K.; Okamoto, T. Biochem. Biophys. Res. Commun. 1980,92, 971. (44) Pezzuto, J. et al. Proc. Nat. Acad. Sci. USA 1980, 77, 1427. (45) tmamura, M. et al. Biochem. Biophys. Res. Commun. 1980, 96, 611. (46) Nebert, D. W. et al. Proc. Nat. Acad. Sci. USA 1979,76. 5929. (47) Nemoto, N. etal. Chem -Biol. Interact. 1979,27, 191. (48) Mita, S. et al. Proc. 39th Annual Meet. Jpn. Chem. Assoc. 1980. (49) Ishii, K. et al. Proc, 39th Annual Meet. Jpn. Chem. Assoc. 1980. (50) Hashimoto. Y.; Shudo, K/, Okamoto, T. Biochem. Biophys. Res. Common. 1980,96,355. (51) Hashimoto, Y., Shudo, K.; Okamoto, T. Proc. 39th Meet. Jpn. Chem. Assoc. 1980. (52) Yamazoe, Y. et al. Chem.-Biol. Interact. 1980,30. 125. (53) Sugimura, T. et ai. Nucleic Acids Res. Special Pub. 1977,3, 246. (54) Takayama, S. et ai. Proc. Jpn. Acad. 1977,53, 126. (55) Takayama, S-; Hirakawa. T.; Sugimura, T. Proc. Jpn. Acad. 1977,54, 126. (56) Ishikawa, T. et al. J. Cancer Res. Clin. Oncol. 1979,95, 221. (57) Sasaki, M. et al. Proc. Jpn. Acad- (B) 1980,56, 332. (58) Matsukura, N. et al. Science 1981,213,346. John S. Wishnok is a research scientist in the Department of Nutrition and Food Science at MIT (Cambridge, Mass. 02139; 617-253-6795). A graduate of the College of Wooster and Brown University (M.A.T. and Ph.D., organic chemistry), he was a post-doctoral fellow with Paul Schleyer at Princeton, and taught organic chemistry at Boston University before moving to MIT. His research has included vacuum ultraviolet photochemistry, pyrolytic rearrangements, biological structure-activity relationships, | carbonium-ion chemistry, and carcinogen j metabolism. THE ACADEMIC HIERARCHY URL 03925 THE PRESIDENT: Leaps tall buildings at a single bound. Is more powerful than a locomotive, Is faster than a speeding bullet, Walks on water. Gives policy to God. The ACADEMIC VICE-PRESIDENT: Leaps short buildings at a single bound, Is more powerful than a switch engine. Is just as fast as a speeding bullet. Walks on water if sea is calm, Talks with God. PROFESSOR: Leaps short buildings with a running start and favorable winds, Is almost as powerful as a switch engine. Can fire a speeding bullet. Walks on water in an indoor swimming pool. Talks with Cod if special request is approved. ASSOCIATE PROFESSOR: Barely clears a quonset hut. Loses tug of war with locomotive. Misfires frequently. Swims well. Is occasionally addressed by God. ASSISTANT PROFESSOR: Makes high marks on walls when trying to leap tall buildings. Is run over by locomotive. Can sometimes handle a gun without inflicting self-injury, Dog paddles, Talks to animals. GRADUATE STUDENT: Runs into buildings. Recognizes locomotives two out of three times. Is not issued ammunition. Can stay afloat with life jacket. Talks to walls. UNDERGRADUATE STUDENT: Falls over doorstep when trying to enter building. Says, "Look at the choochoo," Wets himself with a water pistol. Plays in mud puddles. Mumbles to himself. DEPARTMENT SECRETARY: Lifts tall buildings and walks under them. Kicks locomotives off tracks, Catches speeding bullets in her teeth and eats them, Freezes water with a single glance. Is God. CHEMTECH Book Review Editor Gil Mains of Oklahoma State University sent us this from "Pople's Gaussian 80 Q.M." CHEMTECH JUNE 1.984 Diet and the Chemical Environment as Modifiers of Carcinogenesis A. E. M. McLEAN1 We are faced with the problem that, on the one hand, chemical carcinogens such as benzo [a] pyrene, and also possibly oncogenic viruses, such as the Epstein-Barr virus, are ubiquitous, and that, on the other hand, each culture, social group and region of the world has its different characteristic pattern of tumour incidence (Doll, 1967; Lijinsky & Epstein, 1970; Ringertz, 1971; Shabad et ai, 1971). The studies of migrant peoples suggest that the differing patterns of cancer incidence are mostly determined by environmental rather than genetic factors (Haenszel 6C Kurihara, 1968). The environment has social, biological, chemical and physical components. It seems to me most likely that the differing patterns of cancer incidence have their origin largely in. differing chemical environments. It is difficult to imagine mechanisms whereby social or interpersonal factors, per se, could modify carcinogenesis, though no doubt they might do so by altering endocrine states, as through anxiety or pregnancy. Again, there seems little evidence of differences in bact erial, parasitic or viral status for most of the groups that differ in cancer incidence. Genital herpes and endemic malaria could be excep tions. In a country like England, practically all social classes share the same respiratory virus infections, intestinal parasites are rare and seem unlikely to be involved in tumorigenesis, but social gradients of cancer still exist. The physical environment, temperature, radiation and the like can account for varia 1 University College Hospital Medical School, London, UK. tions in a few instances, as with skin cancers, but rhe bulk of the variation in cancer inci dence must be attributed to variation in the chemical input to the members of the popula tions concerned. The chemical environment of man, chat is all the molecules that impinge upon him, can be arbitrarily divided up into the following groups: (1) air, and inhaled substances, such as smoke and dust; (2) water; (3) food and drink; (4) substances ingested for social or thera peutic purposes, or for pleasure (medications, herbs, drugs, betel chewing); (5) substances concerned in surface ap plications cosmetics, soaps, oils, etc., ap plied either by individual choice or through industrial exposure. The chemical environment is indivisible and, from the point of view of a liver cell, it does not matter whether a molecule arrives via an anal suppository or an industrial smoke. We breathe about 11 kg of air per day, which is the carrier of important contami nants in cigarette smoke and industrial ef fluents Drinking-water is consumed in variable amounts, and the 1--3 litres that we consume every day may well modify car cinogenesis, as it seems to modify coronary heart disease, by virtue of its variable mineral content. Little is known about this factor (Crawford, Gardner & Morris, 1971). -- 223 The bulk of the molecules that come into contact with cells come as food. We eat 1--3 kg of food per day, and food is inextricably a part of man's culture. It has great symbolic significance, and food habits merge gradually into social customs, such as dancing, drinking, the use of herbs and spices, washing or cosmetics, making it difficult to isolate single factors in food. The components of the diet can be'looked at from three points of view: (1) general nucritional adequacy; (2) contamination with carcinogenic subs- stances; and ,, (3) factors in the food chat alter metabolic networks, and so alter response to carcinogens. NUTRITIONAL ADEQUACY AND CARCINOGENESIS The nutritional adequacy of a diet is usually measured by its ability to support growth and reproduction, without producing signs of deficiency disease. It can be cor related with, and summarized by, its total nutrient content (i. e., carbohydrate, aminoacids, essential fatty acids, vitamins, minerals, etc.). The exact means by which an adequate nutrient intake is achieved seems to matter very little. For instance, in rats the best growth rates are achieved with about 20% of the calories as fat, but whether this is olive oil, com oil, or herring oil matters very little. In a number of experimental studies, it has been found that a reduction in calorie intake and of protein intake tends to reduce the incidence of cancers in laboratory animals (White, 1961; Tannenbaum & Silverstone, 1953; Ross 8c Bras, 1971). But human popula tions vary more in the kinds of foods that are taken in than in the values found when these foods are reduced to "calories*" or "protein". In some laboratory instances, specific diet ary deficiencies lead co sensitivity to chemical carcinogens. For instance, riboflavine defici ency leads to a marked sensitivity to azo dye carcinogenesis, associated with low levels of one pathway of metabolism of the carcinogen (Miller et al., 1948). However, human defi ciency diseases are not in general associated with carcinogenesis. For instance, Jamaica has a high incidence of childhood malnutrition, but no marked excess of cancer of the liver or other organs. In Africa, the areas of high oesophageal or liver cancer incidence are areas in which malnutrition is prevalent, but equally malnou rished areas without these cancers are found nearby (Oettle, 1964). CARCINOGENS IN FOOD The problem about carcinogenic molecules in food is that they seem to be a .niversal background; just as cosmic radiation falls on us all, so do carcinogenic molecules enter us. This means that, while they may play a part in carcinogenesis, they cannot be held respon sible for the differences between individuals, and often provide no explanation for the differences between populations. Benzo [a] pyrene can be found wherever rhere is fire and industry. Low ievels of nitrosamines can be detected in all kinds of foods, and ic seems possible chat they are synthesized in vivo (Lijinsky 8c Greenblatt, 1972; du Plessis, Nunn 8c Roach, 1969). There are some instances where carcinogenic substances are found in foods in amounts that approach the carcinogenic levels for animals. Thus, aflatoxin contamination of market produce is widespread in Africa and other tropical countries (Lopez 8c Crawford, 1967). The brackens eaten in Japan seem to be carcinogenic, (Hirono er al., 1972), and smok ed foods in Iceland may be related to a high stomach cancer incidence. ENZYME INDUCTION AND FOOD Diets that are nutritionally adequate can be made up in an immense variety of ways. The protein content can vary widely. The foodstuffs that make up the diet can be purely vegetatian or almost exclusively animal products. Fat content can range from 10% of calories in Japan to 42% in Britain. Such varied diets can support life and growth because the enzyme composition of the body is matched to the chemical input by the process of induction of enzyme synthesis. URL 03927 DIET AND THE CHEMICAL ENVIRONMENT AS MODIFIERS OF CARCINOGENESIS 225 Thus t}je alanine amino transferase (glut amine pyruvic transaminase -- GPT) of the liver is raised whenever there is a rapid flow of amino-acids through the liver. This occurs either when high protein diets are fed, or in starvation, when muscle protein is broken down for gluconeogenesis. When a low-pro tein high-calorie diet is fed, the levels of GPT in the liver fall, both in rats and in man (Schimke, 1962; McLean, 1966) (see Table 1). Table 1. Effect of low protein diet on enzyme activity in human and rat liver1 Diet Alanine transaminase* IliM mg protein per hour) Malate dehydrogenase* (pM.'mg protein per hour) Adults: hospital diet Children: 3 g protein/kg per day Children: 1 g protein/kg per day {isocalorie) flats: 30/o casein Rats: no proteir (2--3 weeks) 12.5 25 11.8 2.2 6.5 1.2 12.5 * 3.5 1.2 0.8 14.2 2.9 16.2 2.2 20.4 1.1 30 . 6 40 4 * Source: McLean (19661 and unpu&liehed data. *> Mean = standard deviation. THE TRIPLE SYSTEM OF DETOXICATION One of the biochemical requirements of any animal is a method for the disposal of the absorbed molecules from food that are either toxic or inert and do not fit easily into metabolic pathways. Steroids, for example, are readily absorbed but poorly excreted. To prevent the accumulation of such compounds, both exogenous and endo genous in origin, a triple system of disposal has evolved. In the liver, it consists first of proteins specialized for the uptake of molecules such as bilirubin and steroids, the Y and Z proteins of Arias (1971). This leads on to the microsomal hydroxylation system, centred on cytochrome P 450, which is capable of hydroxylating an immense variety of substrates. Drugs, such as phenothiazines, bar biturates and many others, are attacked, while steroids, polycyclic hydrocarbons, and fatty acids are also drawn into this system. Some compounds, such as nitrosodimethylamine and carbon tetrachloride, are converted into highly reactive metabolites by the system (McLean, 1971), but in the majority of instances hydroxylation leads either to excre tion of water-soluble metabolites, or else to immediate conjugation with chemical groups, such as glucuronides. The conjugation step is almost invariably an inactivation that des troys pharmacological activity, and as a result, the triple system is, as a rule, an effective detoxication mechanism. Poisons are substances that escape this system, often because they form reactive metabolites that may even attack distant sites. For Instance, tetraethyl tin and lead are dealkylated to triechyl compounds by liver microsomes, and the triethyl compound then attacks brain mitochondria, of which it is a potent inhibitor (Cremer, 1959). The hydroxylation system is found not only in the liver, but also, to a lesser extent, in the lung, kidney, gastro-intestinal tract, and skin. Hepatic synthesis of the hydroxylation sys tem and its central cytochrome P 450 is stimulated about six-fold by phenobarbitone and other inducers. As a result, the ability of animals to withstand compounds that are either detoxicated or activated by the hydrox ylation system of one or other tissue depends very greatly on the previous exposure to inducing drugs. HYDROXYLATION AND CARCINOGENESIS Compounds that are toxic or carcinogenic have to react with cell components in order to exert their effects. This means chat a highly reactive molecule, such as a free radi cal or epoxide, is usually the effective agent. Such reactive molecules cannot live long in a natural environment, so that the reactive molecules must be generated in the body from nonreactive precursors. The most usual en zyme system for generating such reactive molecules is the hydroxylation system. U R L 03928 > URL 03929 226 HOST ENVIRONMENT INTERACTIONS Microsomal hydroxylation seems to be responsable for the activation of polycyclic hydrocarbons, carcinogenic amines, pyrrolizidine alkaloids, nitrosoamines, and aflatoxin (Miller, 1970; E. K. McLean, 1970; Garner er al., 1971; Butler, Mattocks Sc Barnes, 1970). In many instances, the interaction of carci nogen and microsomal enzymes is complex, since multiple hydroxylations may inactivate, as will further conjugation. So while a certain level of hydroxylation generates a reactive molecule from, say aflatoxin, higher levels protect animals from the toxic and carcino genic action of aflatoxin in vivo (McLean Sc A. K. Marshall, 1971). This means that the interaction between carcinogen and target animal will depend on the exact level of hydroxylation enzymes in the animal. The hydroxylation activity of the different tissues, is, in turn, determined by the diet. The effect of increasing microsomal activity is often an increased resistance by the animal to carcinogenesis. The induction of the entire triple system of detoxication is a more effec tive mechanism of detoxication than one would predict if only hydroxylation were increased. The protection given by induction is not absolute, but rather a shift of the doseresponse curve so that a smaller proportion of the exposed population develops tumours (Miller et al., 1958). (Vesell Sc Page, 1969). For mice, we know that the genetic factor influences the activity of the enzymes in liver, kidney and intestine (Nebert & Gelboin, 1969). In rats, we have shown that dietary protein and the dietary fat content exert a predo minant influence on cytochrome P 450 levels and microsomal hydroxylation activity. These effects are found not only in the liver (Kato, Oshima & Tomiyawa, 1968; McLean Sc McLean, 1966, 1972), but also the kidney and lung (Paine1). Table 2. Effect of diet and phenobarbitone on cytochrome P450 levels in rat liver microsomes* No protein, no fat No protein + 15% herring oil 20% casein. no fat 20% casein + 15% herring oil Stock pellets Cytochrome P 450 level* (n M/g) in liver, with: Tap water Phenobarbitone in drinkinq-water 81 24 6 13 1 10 3 53 17 60 7 24 + 7 30 3 123 + 19 . 124 + 7 Source: Marshall & McLean (19711. Mean standard deviation. DETERMINANTS OF HYDROXYLATION The activity of hydroxylation enzymes in animals and man is determined by three main factors: (1) genetic background; (2) dietary protein and fat intake; and (3) intake of non-nutrient enzyme inducers. Studies in various strains of animals and human studies of twins have shown that hydroxylation activity and cytochrome P 450 levels in the liver are under strong genetic influence. The genetic factors determine not only the basal level of hydroxylation, but also -the response to inducers like phenobarbitone Table 2 shows that both protein and a source of unsaturated fat (in this case, herring oil) are required for synthesis of basal and induced levels of cytochrome P 450 in liver. In previous work, (McLean 6c McLean, 1966) we showed that 6/o protein in the diet reduc ed drug metabolism to about half maximum and that, as protein levels became adequate for growth, so they supported synthesis of drug metabolizing enzymes. This suggests that in all those parts of the world where protein supplies are rate limi ting for growth, they are also rate limiting for hydroxylation. 1 Unpublished data. ie DIET AND THE CHEMICAL ENVIRONMENT AS MODIFIERS OF CARCINOGENESIS 227 Similarly, the experiments showed that maximal cytochrome P 450 levels are achieved only if a source of poly-unsaturated fatty acid is incorporated in the diet. Many substances are present in natural diets that are either inducers or synergists of inducers. Wattenberg (1971) and Wattenberg, Page & Leong (1968) have shown such inducing accivity for liver, lung, intestine and kidney by flavones, and by some factor present in cabbages and many other vege table foods. We have found that benzofujpyrene hydroxyiation in kidney responds markedly to the fat content of the diet, but not to protein. In contrast, the lung hydroxylase responds to protein and not to fat (see Table 3). In this respect, the benzo[a]pyrene hydroxylase differs from the mass or cyto chrome P 450 in the liver, which depends on both fat and protein (Marshall Sc McLean 1969, 1971). The hydroxyiation system consists of a number of elements. Coon showed that it could be split into a reductase (NADPH-P450 reductase), a lipid component, and a terminal haemoprotein (Coon, Autor & Strobel, 1971). The terminal haemoprotein portion gives the substrate specificity. The hydroxylase system from different organs, or after different dietary or induction treatments, has dif ferent substrate specificity. After induction with methylcholanthrene, the haemoprotein that is synthesized has an absorption peak at 488 nm (i. e., P 448) and is highly active in benzo[a]pyrene hydroxyiation, while phenobarbitone induces the P 450 haemoprotein, which has a wider substrate range and a lower activity towards benzo^r] pyrene (Kuntzman ef al., 1971). NUTRITION AND NITROSODI METHYLAMINE Nitrosodimethylamine (DMNA) is meta bolized by liver microsomes (Magee 8C Vandekar, 1958), and there is good evidence that metabolism is required for it to exert its toxic and carcinogenic effects (Magee Sc Swann, 1969; Lijinsky, Loo 8c Ross, 1968; McLean, 1971). The toxic and carcinogenic effects of DMNA are profoundly altered by changes of diet. When a low-protein, highly purified diet is fed, the acute hepatotoxic effects of DMNA are greatly reduced (McLean & Verschuuren, 1969), while the renal carcinogenic effect is enhanced (McLean & Magee, 1970). Metabolism of DMNA by the liver of rats fed low-protein diets is reduced to about one half that found when stock pellets are fed. As a result, DMNA persists in the blood for a long time after a single dose. Renal meta bolism of DMNA is not affected by the diet, and renal exposure to DMNA is effectively increased (Swann Sc McLean, 1971). We have recently been measuring the meta bolism of DMNA to formaldehyde by liver microsomes and postmitochondrial factions. (One mole of formaldelyde is produced per mole of DMNA metabolized). We find that rats fed purified low-protein diets have low levels of DMNA metabolism. Table 3. Effect of diet on benzo(aJpyrene hydroxyiation in liver, kidney and lung* Diet Benzo{a]pyrene hydroxyiation* (pM/mg protein per 30 min) Liver Kidney Lung Low protein, no fat Low protein, 5% fat* 20% casein, no fat 20% casein, 5% fat* 1163 214 1461 131 2755 573 3353 1218 75 26 152 54 101 11 221 100 36 19 40 7 65 * 32 71 19 * Source: Paine & McLean (1972) (unpublished data). * Mean standard deviation. c Olive oil. 228 HOST ENVIRONMENT INTERACTIONS but this can be restored to the levels found in rats fed stock pellets by giving phenobarbitone. Benzofc] pyrene has very much less effect on the rate of DMNA metabolism. However, phenobarbitone has no effect on DMNA toxicity, while benzofajpyrene greatly increases the hepatotoxic effect (Table 4). It seems likely that there are several pos sible pathways of DMNA metabolism, only some of which lead to toxic effects; we do not know whether carcinogenic effects are similarly affected. Arcos and his co-workers suggest that inducers, like 3-methylcholanthrene or phenobarbitone, inhibit DMNA metabolism in their rats; in fact, what they show is a reduced specific activity (DMNA metabolism /mg microsomal protein), when inducers cause, a large increase in microsomal protein, while DMNA metabolism per 100 g rat remains relatively constant (Venkatesan, Argus 8c Arcos, 1970). Stock rat diet seems to contain factors that stimulate DMNA metabolism in the maximally "toxic" direction. It would seem of importance to know more about these alternative pathways of DMNA metabolism and the dietary factors controlling them. THE SIGNIFICANCE OF DIETARY INDUCERS OF HYDROXYLATION ACTIVITY We now have a series of loosely linked observations: (1) carcinogenic chemicals are widespread in the environment; (2) populations differ greatly in their pat terns of cancer incidence; (3) microsomal hydroxylation seems neces sary for the activation of carcinogens, but if taken far enough can then inactivate these compounds; (4) dietary factors alter microsomal hydro xylation activity in many tissues in experi mental animals; and (5) populations differ widely in their dietary habits in ways that we would predict should alter their microsomal hydroxylation activity. These observations seem to me to lead to two conclusions. First, we badly need to develop methods to enable us to measure microsomal hydro xylation in large numbers of people in order to see whether patterns of hydroxylation cor relate with cancer patterns, and whether diet influences microsomal hydroxylation in man. Second, we should perhaps look at diets made up to resemble those used by human populations, to see whether the diets alter hydroxylation and whether they confer resistance or susceptibility to carcinogens, when given to experimental animals. URL 03931 Table 4. Effect of diet on metabolism of, and liver injury from nitrosodimethylamine Oiet DMNA metabolism" (y,M/g liver per hour) Indices of liver injury* Plasma isocitrate dehydrogenase (nM/ml min) Plasma bilirubin (mg/100 ml) Stock pellets 3% Casein, 5% olive oil 3% Casein, 5% olive oil + phenobarbitone 3% Casein, 5% olive oil + benzo[a]pyrene [ ! 1.41 0.2 0.62 0.2 1.31 0.1 0.75 j 1000 135 180 851 1.54 0.32 0.36 1.0 * Measured 24 hours after i p. administration of OMNA (90 mg/kg). b Mean standard deviation. DIET AND THE CHEMICAL ENVIRONMENT AS MODIFIERS OF CARCINOGENESIS REFERENCES Arias. I. M. (1971) On the structure and function of Y protein. Chem.-Biol. Interactions., 3, 237-238 Butler, W. H., Mattocks, A. R. & Barnes, J. M. (1970) Lesions in the liver and lungs of rats given pyrrole derivatives of pyrrolizidine alkaloids. J. Path., 100, 169-176 Coon, M, J,, Aucor, A. P. & Strobe!, H. W. (197!) Role of phospholipid in electron transfer in a reconstituted liver microsomal enzyme system containing cytochrome P450. Chem.-Biol. Interactions, 3, 248-250 Crawford, M. D., Gardner. M, J. Sc Morris, J. N. (1971) Changes in water hardness and local death rates. Lancet, iii, 327-329 Cremer, J. E. (1959) Biochemical studies on the toxicity of tetraethyl lead and other organolead compounds. Brit. J. industr. Med., 16, 191-199 Doll, R. (1967) The prevention of cancer. Pointers from epidemiology, London, Nuffield Provincial Hospitals Trust, p. 143 Du Plessis, L. S., Nunn, J. R. 8c Roach, W. A. (1969) Carcinogen in a Transkei Bantu food additive. Nature (Lond.), 222, 1198 Garner, R. C., Miller, E. C., Miller, J. A., Garner, J. V. Sc Hanson, R. S. (1971) Formation of a factor lethal for $. typhimurium TA 1530 on incubation of Aflatoxin B, with rat liver microsomes. Biachem. biophys. Res. Commun., 45, 774-780 Haenszel, W. 8c Kurihara, M. (1968) Studies of Japanese migrants. 1. Mortality from cancer and other diseases among Japanese in the United States. J. nat. Cancer Inst., 40, 43-68 Hirono, L, Shibuya, C., Shimiyu, M. Sc Sushimi, K. (1972) Carcinogenic activity of processed bracken used as human food. J. nat. Cancer Inst., 48, 1245-1250 Kato, R., Oshima, T. & Tomiyawa, S. (1968) Toxiciry and metabolism of drugs in relation of dietary protein. Jap. J. Pharmacol., 18, 356-366 Kuntzman, R-, Lu, A. Y. H., West, S., Jacobson, M. 8c Conney, A. H. (1971) The importance of cytochrome P 450 and P 448 in determining the specificity of the reconstituted liver micro somal hydroxylation system. Chem.-Biol. In teractions, 3, 287-288 Lijinsky, W, Sc Epstein, S. S. (1970) Nitrosamines as environmental carcinogens. Nature (Lond.), 225, 21-23 Lijinsky, W. 8C Greenblatt, M. (1972) Carcinogen dimethyl nitrosamine produced in vivo from nitrite and aminopvrine. Nature (Lond.) Nev; Biology, 256, 177-178 Lijinsky, W., Loo, J. 3c Ross, A. E. (1968) Mechanism of alkylation of nucleic acids by nitrosodimethylamine. Nature (Lond.), 218, 1174-1175 Lopez, A. 8c Crawford, M. A. (1967) Aflatoxin content of groundnuts sold for human con sumption in Uganda. Lancet, i, 1351-1354 Magee, P. N. 8c Swann, P. F. (1969) Nitroso compounds. Brit. med. Bull., 25, 240-244 Magee, P. N. & Vandekar, M. (1958) Metabolism of dimethylnitrosamine in vitro. Biochem, 70, 60Q-6Q5 Marshall, W. J. 8c McLean, A. E. M. (1969) The effect of oral phenobarbitone on hepatic microsomal cytochrome P 450 and demetnylation activity in rats fed normal and low protein diets. Biochem. Pharmacol., 18, 153- 157 Marshall, W. J. 8c McLean, A. E. M. (1971) Re quirement for dietary lipid for induction of cytochrome P 450 by phenobarbitone in rat liver microsomal fraction. Biochem. ]., 122, 569-573 McLean, A. E. M. (1966) Enzyme activity in the liver and serum of malnourished children in Jamaica. Clin. Sci., 30, 131-139 McLean, A. E. M. (1971) Conversion of inactive into toxic molecules by the liver. In: Aldridge, W. N., ed., Mechanisms of toxicity, London, MacMillan, pp. 219-228 McLean, A. E. M. 8c McLean, E. K. (1966) The effect of diet and DDT on microsomal hydro xylation and on sensitivity of rats to carbon tetrachloride poisoning. Biochem. /., 100, 564-571 McLean, A. E. M. Sc McLean, E. K. (1972) Pharmacogenetic and other regulatory factors in drug metabolism. In: Buriand, W. L. 8e Laurance, B. M., eds. The therapeutic choice in paediatrics, London, Churchill, pp. 84-95 McLean, A. E. M. 8c Magee, P. N. (1970) In creased renal carcinogenesis by dimethylnitros amine in protein deficient rats. Brit. /. exp. Path., 51, 587-590 McLean, A. E. M. 8c Marshall, A. K. (1971) Reduced carcinogenic effects of aflatoxin in rats given phenobarbitone. Brit. ]. exp. Path., 52, 322-329 McLean, A. E. M. 8c Marshall, W. J. (1971) Effect of linoleic acid, peroxidation and anti oxidants on induction of cytochrome P 450 in rat liver. Chem.-Biol. Interactions, 3, 294-295 URL 03932 230 HOST ENVIRONMENT INTERACTIONS McLean, A. E. M. Sc Verschuuren, H. G. (1969) Effect of diet and microsomal enzyme induc tion on the toxicity of dimethyJnitrosamine. Brit. J. exp. Path., 50, 22-26 McLean, E. K. (1970) The toxic actions of pyrrolizidine (senecio) alkaloids. Pharmacol. Rev., 22, 429-483 Miller, J. A. (1970) Carcinogenesis by chemicals: an overview. Cancer Res., 30, 559-576 Miller, E. C., Miller, J. A., Brown, R. R. Sc MacDonald, J. C. (1958) On the protective action of certain polycyclic aromatic hydro carbons against carcinogenesis by amino-azo dyes and 2 acetyl amino fluorene. Cancer Res., 18, 469-477 Miller, E. C., Miller, J. A., Kline, B. E. Sc Rush, H. P. (1948) Correlation1 of the level of hepatic riboflavine with the appearance of liver tumours in rats fed azo dyes. ]. exp. Med., 88, 89-98 Nebert, D. W. & Gelboin, H. V. (1969) In vivo and in vitro induction of aryl hydrocarbon hydroxylase in mammalian cells of different species, tissues, strains and developmental and hormonal states. Arch. Biochem., 134, 76-89 Oettle, A. G. (1964) Cancer in Africa, especially in regions south of the Sahara. J. nat. Cancer Inst., 33, 383-439 Ringertz, I. V. (1971) Cancer incidence in Fin land, Iceland, Norway and Sweden. Acta path, microhioi. scand., suppl. 224, pp. 11-35 Ross, M. H. 8c Bras, G. (1971) Lasting influence of early calorie restriction on prevalence of neoplasms in the rat. /. nat. Cancer Inst., 47, 1095-1113 Schimke, R.T. (1962) Differential effects of fasting and protein free diets on levels of urea cycle enzymes in the rat liver. ]. biol Cheyn., 237, 1921- Shabad, L. M., Cohan, T. L., Ilnitsky, A. P., Klesiva, A. Y., Shcherbak, N. P. fic Smirnov, G. A. (1971) The carcinogenic hydrocarbon benzo-a-pyrene in the soil. ]. nat. Cancer Inst., 47, 1179-1191 Swann, P. F. 8t McLean, A. E.M. (1971) The effect of a protein free diet on metabolism of dimethylnitrosamine. Biochem. }., 124, 2S3-28S Tannenbaum, A. 8c Silverstone, H. (1953) Nutri tion in relation to cancer. Advanc. Cancer Res., 1, 451-501 Venkatesan, N., Argus, M. F. Sc Arcos, J. C. (1970) Mechanism of 3 methylcholanthrene induced inhibition of dimethyl nitrosamine demethylase in rat liver. Cancer Res., 30, 2556-2562 VeselL E. S. Sc Page, J. G. (1969) Genetic control of the phenobarbital induced shortening of plasma anripyrene half lives in man. /. clin. Invest., 48, 2202-2209 Wattenberg, L. W. (1971) Studies of polycyclic hydrocarbon hydroxylases of the intestine possibly related ro cancer. Effect of diet. Cancer, 28, 99-102 Wattenberg, L. W., Page, M. A. Sc Leong, J. L. (1968) Induction of increased benzpyrene hydroxylase activity by flavons and related compounds. Cancer Res., 28, 934-937 White, F. R. (1961) The relation between under feeding and tumour formation, transplantation and growth in rats and mice. Cancer Res., 2t, 281-290 URL 03933 Transactions of the Illinois State Academy of Science (1983). Volume 76. 1 and 2. pp. 181-194 MNbnCE: by This rvat^rial mav be orotftcted 'Co ' o- w u /> >i> Toxicities of Natural Foods in Man: A Survey Harold Kaplan Department of Physiology School of Medicine, Southern Illinois University Carbondale, Illinois Foods are considered to be substances that nourish cells and enable them to grow, multiply and repair themselves. Many foods may produce bodily harm, how ever, by some intrinsic toxicity or by excess consumption, contamination, prepara tion, inherited inability to process them, hypersensitivity of the individual, or Other. The literature is scattered and voluminous and the early reports are untrust worthy. In the present discussion, man rather than animals is emphasized as well as food toxicity encountered chiefly in the United States. We have not included any commentary on federal policies of increasing toleration of food practices dan gerously leading to toxicities (see Marshall, 1982). FOODS CONTAINING CARCINOGENS Saccharin Saccharin is the most widely used sugar substitute. In large doses it incites bladder tumors in laboratory animals and the FDA has placed it on a regulated list. Tea The black teas from Ceylon and India contain considerable tannin and gallic \ These are controversially suspect as causative factors in esophageal cancer vis and Elvin-Lewis, 1977, p. 119). CONTAMINANTS FAVORING CARCINOGENICITY Fungus Contaminants (Mycotoxins) Some strains of Aspergillus, a mold, produce ajlatoxins. The one found most in food is aflatoxin Br It inhibits protein synthesis, causing the death of liver particularly. is one of the most potent hepatocarcinogens known. It is id chiefly in figs, peanuts and other nuts, and cereals (Osborne, 1982). There is a toxic substance called patulin which can occur in apple juice due to lamination by molds. The substance is suspected of being carcinogenic (Crosby, 979). 181 (-- & %<P 182 Toxins can be present in food that is not visibly moldy and they can persist even if the food is cooked or otherwise processed. Polycyclic Aromatic Hydrocarbons (PCHs) PCHs are an important class of carcinogens because they are being widely disT persed in the atmosphere, soil and water from the smoke and smog of industrial cities and some have tumorigenic potency. Dangerous levels of PCHs are nowjg detected in many foods, particularly leafy vegetables (Harvey, 1982). CONTAMINATION OF DRINKING WATER Water is a food. Contamination by industrial wastes possibly leading to car-J cinogenicity is a huge problem and will not be considered here. There is atmo-! pheric contamination of an incidental nature, however, nitrates posing a problem. It. is contended that a high nitrate concentration in water supplies provides a risk gastric cancer (Walters and Smith, 1981). A concern about nitrates in water peripheral to the cancer issue centers on' their reduction to nitrates by bacteria which can live in an infant's stomach due to low acidity therein. Hemoglobin, especially the persisting fetal type in the infant, is susceptible to oxidation by nitrites and methemeglobin is formed. Oxygen delivery J&f is impaired. Cyanosis in infants who are given water from wells is a manifestation of nitrate toxicity (Gleason et al., 1969, p. 100). Water per se can produce ill effects. Massive overdrinking (which has been used in torture) produces the syndrome of "water intoxication". Cerebral edema, con* vulsions and coma can follow. Water causes gastric distress in many persons because it can excessively stimu late gastric juice, which irritates the stomach lining. That is why "club soda", which contains sodium bicarbonate, may produce no such symptoms. FOODS DEVELOPING CARCINOGENIC PROPERTIES BECAUSE OF PREPARATION Charcoal broiling can result in the formation of carcinogenic hydrocarbons. Smoke imparts a flavor derived from phenolic materials produced when wood is burned in conditions of oxygen deficiency. While the phenolic flavors are formed, polynuclear aromatic hydrocarbons are also produced. Some of these, especially 3, 4-benzpyrene and benzanthracene, are carcinogenic. These occur in smoked fish, caramelized sugar and roasted coffee (Crosby, 1979). Benzpyrene is the ubi quitous carcinogen in city smog. Its metabolites, particularly epoxides, are active in producing malignancy in animals and possibly in man. Fats and oils subjected to prolonged overheating by frying are controversially thought to produce carcinogens. The toxicity is claimed to increase if an oil is used repeatedly and if a shallow pan allowing oxidation is used. Yannai (1980) presents a succinct review of toxicity induced by food processing. Thus, a cured meat such as bacon contains trace amounts of N-nitrosodimethylamine (NDMA) and on frying bacon N-nitrosopyrrolidine is formed. Both sub stances are claimed to be carcinogenic (Gough et al., 1976). Fortunately, both nitrosamines get lost to a considerable degree in the vapor during cooking, although some remain in the rasher and in the cooked-out fat (Holmes and Woodburn, 1981). Wasserman et al. (1978) state that frying bacon at low or medium 4s ,, at for less tb ' nsabene et .mperature a: The nitrite t Clostridium :uce nitrosam ::umals (Piers .^nts (Widdu In 1978 it nfH and tljat . > modify th< markedly. Mutagens nbution in tl mutagenic act There are cor Markedly to tl C Although i IX--I a claim that CODJ :!ro and that uC>O Charred n iid glutamic ; incisions arise i stroys patho a\e a high m There is ; SU77) cite a r.ide poisonin The amin <'ntjudalin. T "xygen carria ire. The less< children, espi Apricot p should be avr as a cancer c persist lely disdustrial ire now ; to cars atmosblem. It a risk of nters on ?h due to e infant. . deliver, festation een used ma, con- iy stimub soda". ,S xiarbons. n wood i>. e formed, especial]} n smoked is the ubiare active oversially oil is used irocessine idimethu' Both sul>tely, botli ; cookine. nd Woodir mediun. i.i heat for less than 10 minutes can result in the least nitrosamine levels, although Pensabene et al. (1974) report that N-nitrosamine formation is dependent on temperature and not time. The nitrite in bacon processing imparts flavor and color and prevents growth of Clostridium botulinum. Since the nitrites react with amines in the body to pro duce nitrosamines which are inconclusively implicated in cancer of laboratory animals (Pierson and Smoot, 1982), there are proposed substitutes or sparing agents (Widdus and Busta, 1982). In 1978 it became known that NDMA was present in domestic and imported beers and that barley malt was the source (Havery et al., 1981). FDA regulations to modify the malting process have reduced the present nitrosamine levels markedly. FOODS CONTAINING MUTAGENS Mutagens are agents that produce genetic mutations and they have a wide dis tribution in the diet. There are mutagenic flavonoids in many vegetables and mutagenic activity is said to exist in extracts from coffee and tea (Maugh, 1982). There are controversial opinions as to whether mutagens in foods contribute markedly to the risk of cancer (Committee, 1982). i A. Caffeine Although there is no evidence that caffeine is involved in carcinogenesis, there is a claim that it induces chromatid breakage (genetic damage) in human cells in vitro and that it could prove to be a dangerous mutagen (Kahlmann et al., 1968). B. Charring of Meat and Fish Charred meat and fish can develop mutagenic potential. Tryptophan, serine and glutamic acid may produce mutagens at high temperature pyrolysis. Difficult decisions arise. Although cooking of meat and fish may produce mutagens, it also destroys pathogenic organisms. Further, some vegetables that contain mutagens have a high nutritive value. FOODS CONTAINING TOXICANTS (MAINLY CYANOGENIC GLYCOSIDES) A. Fruits Apple (Malus spp.) There is a toxic part of apples, mainly in the seeds. Lewis and Elvin-Lewis (1977) cite a case where one cup of seeds was eaten and caused death, from cya nide poisoning. The amino acid phenylalamine in the seed converts to the cyanogenic glycoside amygdalin. This hydrolyzes on digestion to hydrocyanic acid. The cyanide poisons oxygen carriage by cytochrome oxidase. Victims may die within minutes of expos ure. The lesson here is (1) that all parts of a food are not harmless, and (2) that children, especially, must learn this early. Apricot {Pruiius armeniaca) Apricot plants contain amygdalin in the seeds, bark and leaves and ingestion should be avoided (Lewis and Elvin-Lewis, 1977, p. 41). Laetrile, legal in Mexico as a cancer cure, is an extract of apricot pits. The theory is that the extract yields URL 03936 cyanide*, which preferentially attacks tumor cells. Cyanide is released by betaglucuronidase common to tumor cells. Elderberry (Samhucus spp.) The fresh leaves, stems and roots contain a poisonous alkaloid, and a cyano- genic glycoside called sambunigrin causing cyanide poisoning. Red elderberries are very toxic. The fully ripe fruit is harmless and is used in pies, wine, jelly and pancakes. Other parts of the plant are used as drugs. > Children using peashooters made of elderberry stems have been poisoned (Lewis and Elvin-Lewis, 1977, p. 50). ' -*? Cherries (Prunus spp.) Seeds and leaves contain the cyanogenic glycoside amygdalin. The seeds are the most toxic. The fleshly part of sweet and of sour cherries is edible. That of the laurel cherry (P. laurocerasus) contains amygdalin. Pear (Pyrus communis) Pear seeds contain amygdalin and are potential poisons. Peach (Prunus persica) All parts of the peach contain amygdalin, concentrated mainly in the bark, leaves and seeds. Poisoning has followed ingestion of peach kernels (Liener, 1980, p. 152). EggP1 -terns and , ii-nce, to Being pot nlasodinc Fava - mo is cal nit severe Sensit ilehydrog* 1980, p. : The p !rom eatii Lima 'in amarin <:\ anic aci !<*stroy th ixic lecti Plum (common is Prunus domestica; wild is P. americana) The plum is the least toxic member of the genus Prunus, but the leaves and seeds contain cyanogenic glycosides that convert to hydrocyanic acid. Prune (Prunus domestica) The prune is the partially dried fruit of a plum. The laxative effect of prunes is well known. Prunes contain derivatives of hydroxyphenylisatin which acts as a gastrointestinal irritant. Almond (Prunus amygdalus) % < % s The 1 ^'astrointe Bender a .mylase f a hen taki Blockers f vj whetht I 9 Almonds are the fruit of the almond tree. The seeds of the sweet varity are edible. It is difficult to separate the edible sweet almond from the bitter almond (Prunus When t amygdalus var. amara). The seeds of the bitter plant contain high yields of bread ba> amygdalin (Abrol, 1967). Lewis and Elvin-Lewis (1977, p.41) state that 50 to 70 destroy th f. bitter almond nuts are fatal for adults and 7 to 10 for children. 1 cereals (L j Breac FOODS CONTAINING TOXICANTS tion in t 1 Goodhar B. Vegetables clue in pa Onion (Allium cepa) A toxic principle thought to be an alkaloid may cause an anemia in man if onions are eaten in "large" amounts for one or more weeks (Schmutz et al., 1979, p. 155). The bulb contains the toxic material. Other related suspected foods are chives, garlic, leeks and shallots. Onions and other plants in the genus Allium contain sulfur compounds called glucosinolates. These give rise to thiocyanates, isothyocyanates, nitriles, goitrin and progoitrin. Clucosinolates substitute for or displace iodine, preventing its up take or accumulation in the thyroid gland. Heat reduces the direct action of these substances. Normal patterns of consumption cause no harm. [ } The 1< of calciur precipitat Lewis, not ordin Betw* gestion of tains muc than thre t 'etu- an<>. jrrit^ akc1'Onu: Js art cf th* bark. lftv. s ari': mev. i s .. ediliit Prunu elds it0 to 71 man :: . 197 ' ods ars calif-'.Eoitr::. : its ut of thes is.? Eggplant (Solanum melongena) } Eggplant is a fruit eaten as a vegetable. Botanically, it is a berry. Its leaves, stems and fruit contain solanaceous alkaloids that are claimed, without solid evi dence. to be poisonous. The plant belongs to a nightshade family long suspected as being potentially toxic. Glycosides of solasodtne are common in eggplant and solasodine is teratogenic in hamster fetuses (Kinghorn. 1979, p. 75). Fava Bean (Vicia jaba) Fava (broad) beans may produce acute febrile hemolytic anemia. The condi tion is called favism. Crises due to dry or cooked beans are of moderate intensity but severe attacks follow ingestion of fresh, raw fava seeds. Sensitivity is hereditary and involves a deficiency of glucose-6-phosphate dehydrogenase. Causative toxic substances may be divicine and i&ouramil (Liener, , 1980, p. 283). The plant is widely cultivated in parts of the United States. Most cases result from eating fresh beans and thus peak incidence occurs at harvest. Lima Beans (Phaseolus lunatus) Lima beans contain a lectin (protein hemagglutinin) called phaseolunatin or linamarin (Lewis and Elvin-Lewis, 1977, p. 44). This hydrolyzes in part to hvdrocvanic acid. Thorough cooking reduces the danger of poisoning, but may not fully destroy the toxic principle. Preliminary soaking before cooking helps eliminate the f toxic lectin. Kidney (Navy) Beans (Phaseolus vulgaris) The kidney bean contains lectins (hemagglutinins) associated with human gastrointestinal disturbances. The toxin is completely destroyed by boiling (Bender and Reaidi, 1982). Of interest, the bean also contains inhibitors of alphaamylase from the saliva and pancreas. The active inhibitor, called phaseolamin, when taken orally blocks starch digestion and this is the principle behind starchblockers for weight control (Bo-Linn et ai., 1982). The FDA is now in litigation as to whether these agents are drugs or foods, Wheat (Triticum aestivum) Wheat also contains the alpha-amylase inhibitor. This may persist through bread baking and infants or adults having impaired digestive enzymes may not destroy the inhibitor. A similar problem also arises in some wheat-based breakfast cereals (Liener, 1980, p. 450), Bread cannot be made without the giuten of wheat and some uncertain frac tion in the gluten exacerbates the symptoms of patients with celiac disease (Goodhart and Shils, 1973, p. 423). Wheat is aso a very common food allergen, due in part to albumin and gliadin. Spinach (Spinacia oleracea) The leaves of spinach contain salts of oxalic acid that could become precursors of calcium oxalate crystals. Excesses have been controversially claimed to cause precipitates in the kidney. Spinach also lowers blood sugar (Lewis and ElvinLewis, 1977, p. 218). The concentrations of oxalates or hypoglycemic factors are not ordinarily toxic. Between 1959 and 1965 in Germany, 15 cases of nitrite poisoning followed in gestion of spinach by babies (Simon, 1966). Poisoning occurs only if spinach con tains much nitrate, which bacteria of the stomach convert to nitrite. Infants less than three months of age should not be given spinach because of bacterial action. W?L 0 3 9 38 136 Arena (1979, p. 718) cites the death of a five-year old girl who ate a large number of raw spinach leaves. Epstein (1969) states that vegetables in highly fertilized soil contain 1,000 to 5,000 ppm of sodium nitrate, well exceeding the maximum per missible amount in foods. , Nitrates can convert to nitrite in cans and the nitrite reacts with the metal to produce nitrogen which swells the can. Metal cans have caused toxicity for other reasons, essentially because of dissolved tin or the lead solder used and, for that matter, even plastic containers contain toxic polymers (Liener, 1980, p. 417). Potato (Solanum cuberosum) ; The "Irish" potato is edible so long as it is not very green. Since it is a modified stem, a potato exposed to light for a period of time becomes green and it may ac cumulate toxic amounts of a glycoside alkaloid called solanine. The toxic parts of the potato include green areas under the potato skin and the sprouts. The skid is harmless when eaten in normal quantities. A concentration of 20 mg solanine/lOOg fresh weight is thought to be the maximum safe level for market potatoes, to pre vent gastrointestinal and neurologic disorders. Fatalities are rare because sprouted potatoes are peeled. Also, solanine in boiling is leached out into the water (Koel- ling, 1982), although cooking does not destroy it. Solanine is a cholinesterase inhibitor. Rhubarb (Rheum rhaponticum) The leaves and edible stalks of rhubarb contains rather high quantities of oxalic acid. This was once thought to be the cause of vomiting and enlargement of the liver that follows ingestion of raw leaves and stems. It is much more likely that the toxicity is due to anthraquinone glycosides. These are present in leaves, especially in early summer. About 10 to 20 grams of fresh leaves are highly irritating (Fairbairn, 1976, p.15). Death has sometimes followed ingestion of rhubarb (Leiner, 1980, p. 453). Mushroom (Amanita spp) Mushrooms are unpredictable. Some species are only occasionally poisonous. Some are poisonous only if eaten raw; others if eaten in certain seasons or stages of maturity. Cooking does not destroy all the toxins. Some people become ill if they drink alcoholic beverages when eating mush rooms. Use of certain drugs may make a harmless mushroom toxic for some people. The most common poisoning in the United States is from Amanita verna and A. brunnescens. Poisoning by these and the more rare A. phalloides is due to ther mostable toxins that destroy cells. Toxins are cyclopeptides which include phalloin, phalloidin, and amanitins. Poisoning by A. muscaria is due to ibotenic acid and muscimol (Kinghorn, 1979, p. 25). There are other mushrooms that contain psilo cybin and psilocin which produce hallucinations. Additional toxicants have been extracted from still other species. Yams (Dioscorea spp.) Yams are among compounds suspected of causing teratogenesis in man and other animals (Lewis and Elvin-Lewis, 1977, p. 95). Certain yams (D. hirsute) contain an alkaloid, dioscorine, which produces paralysis of the central nervous system (Glasby, 1975, p. 490). Sweet Potato (Ipomoea batatas) Sweet potatoes produce metabolites called phytoalexins in response to stress (Salunkhe and Wu, 1977). An abundant metabolite is ipomeamarone. This is a ' j ' i ! 1 e te to 'M O .u. m l! U. r* via Tl <1U at c\ K; : ir- bo< an< lesi Ini wr lae to h> ] chi ze number 'tilizedsoil imum per- ie metal to >' for other d, for that p. 417). a modified it may m.. xic parts o; The skin o .anine; lOooes, to pn se sprouter ater (Ko* )ninestera- ies of oxaii Tient of tii* ely that th. especial; iting (Fai: rb (Letn< ; po.-'onoo or stage': iting mu'.: ime peopit. t vertia an:: due to ther ie phalliiu. ic acid am! ntain p.n/<" s have bet :. n man am1 D. hiruit.. ral nervou lse to stre r. This i' r 187 i hepatotoxin in mice and a suspected toxin in man. It is not eliminated by baking or boiling and it is best to avoid potatoes with blemishes or discolorations. C. Food Colors; Spices Nearly 65 synthetic colors are"known to be in use around the world. Khanna and his co-workers (1980) review these substances and they deplore the fact that some appear to be highly toxic and nevertheless are insufficiently regulated. In the United States there are about eight color additives and the FDA certifies them. Monosodium Glutamate This is the responsible agent for the digestive upset known as the "Chinese restaurant syndrome." It is especially dangerous in baby foods because of reported damage to the brain in newborn animals (Grollman and Grollman, 1970). Hot Peppers (Capsicum spp.) Red chili peppers (C- annuum) can irritate the entire gastrointestinal tract. The symptoms are more severe with the tobacco pepper (C. frutescens) which is considered to be a hepatocarcinogen in rats. Nutmeg (Myristica fragrans) Nutmeg is the ripe seed of M. fragans. It is a powerful hallucinogen. This is due to a combination of volatile oils, myristicin, elemicin and the weak carcinogen safrole. Ingestion of large quantities of nutmeg causes stupor and death (Merck Index, 1960, p. 700). Sage (Salvia officinalis) Sage represents the dried leaves of the plant. Large doses may cause nausea, vomiting and constipation (Merck Index, 1960, p. 918). D. Milk Bovine milk is the main source of orotic acid in the human diet. Levels in cow's milk increase from 21 to 84 ppm from days 2 to 45 of lactation. Because only 10 ppm may affect liver metabolism and cause liver necrosis in rats, there is need to evaluate its possible toxicity in milk powders used in humans (Okonkwo and Kinselia, 1969). Robinson (1980) does not think that orotic acid poses a serious f problem in man, but believes that research is needed on this point. FOODS CONTAMINATED BY TOXICANTS Milk There is considerable concern about the level of pesticide residues in both 'bovine and human milk. Dillon et al. (1981) present a survey of this problem. Lee [tad Lorenz (1979) note other hazards of milk, i.e., in the cardiac-prone hyperchoItsterolemic person, in the lactose-intolerant person, and in the allergic person. ^Infants with hereditary disturbance of galactose metabolism develop serious ^symptoms with ingestion of milk. Such galactosemia demands rigid exclusion of lactose and galactose from the diet. Speckmann et al. (1981) claim that intolerance to lactose in an 8-ounce serving of milk is rare. Rye (Secale cereale) Ergot is the toxic chemical constituent of the sclerotium (branching filaments; ig byphae) of Claviceps purpurea, a fungus that parasitizes kernels of rye (and wheat) chiefly. Ergot contains ergonovine, an alkaloid causing human ergotism. This is vs** .r4fc\- URL 03940 188 highly controlled today, but it killed millions of people in the Middle Ages. There are burning sensations and gangrene in the extremities, and also abortion and con* vulsions. Epidemics in the Middle Ages were referred to as "St. Anthony's Fire" because gangrenous limbs looked as though they were burned. Fish and Shellfish Fish muscle contains histidine which can form histamine by bacterial action if left standing. Canned tuna fish has been a source of histamine poisoning (Rice et al., 1976), and the disorder called scombroid poisoning is caused chiefly by histamine. Shellfish such as clams eat the dinoflagellates Gonyaulax catenella or C. tarn* arerwis and thereby acquire a nitrogenous compound called tetrodotoxin which does not harm the shellfish, but which is often a fatal neurotoxin to man. At cer tain times of year the dinoflagellates so multiply as to produce a dangerous "red tide" poisonous to clams, crabs, oysters, shrimp and other marine organisms. A neurotoxin is involved called saxitoxin which is highly poisonous when man ingests the shellfish and this toxin is not destroyed by cooking. Fish eggs can also contain toxicants (see Liener, 1974). Meat Grilled on Wooden Skewers Wooden skewers used to hold "hot dogs" and other meat for grilling can in troduce hazards. There are skewers made from the stems of an evergreen, Nerium oleander, which contain cardioactive glycosides that have caused deaths (Liener, 1974). The poisons are oleandroside and nerioside. Beer Treated with Cobalt At one time cobalt was added to beer to stabilize the foaming qualities. It took many years before cobalt was incriminated as the cause of severe heart attacks in heavy beer drinkers and this additive was discontinued (Alexander, 1969). FOODS ACTING AS TOXICANTS BECAUSE OF INHERITED DEFECTS IN THE INDIVIDUAL For a person having severe pancreatic diabetes, where from birth insulin function may be deranged, a single ordinary meal containing cane sugar (sucrose) may set off a train of metabolic events leading to disastrous results. Fructose occurs in many fruits and is the predominant sugar in honey. Some people are born without an enzyme, aldolase B, and for them fruits (and also cane sugar) have to be completely excluded to prevent gastrointestinal pain, vomiting and hypoglycemia (Cox et al,, 1982). The disorder is acquired from a double dose of a recessive gene. Persons who by gene defect lack the enzyme lactase fail to break down milk sugar (lactose) and they may not tolerate milk. Fatty foods, especially french fried potatoes and onions, are anathema to most people who suffer by inheritance or otherwise from disorders of the biliary tract. This is expected since many nutrients affect the quality and amount of the bile, the movements of the gallbladder, and the tone of the sphincter of Oddi. These so-called inborn errors of metabolism also include protein disturbances. In phenylketonuria, there is a deficiency of phenylalanine hydroxylase and pheny lalanine intake must be sharply reduced by feeding artificial protein substitutes. In "maple syrup urine disease," the infant must ingest an artificial formula free of \ ulim fnxici \ < i i).05 nf bit t ists (< f>ern are a ribot dair -cen C XI hav coo - ita for 1 rom afte rna* is a I :Cf. i udu t five flic \V it to tei Tht-r. Jon and cM. bony's Klr, rial action :: : (Rico et j\ histamin. i or C. tan, toxin whui. nan. At ti i yerous "rt.o rganism.v A man inycvtilso contair. line can it, en. Xeriun' :hs (Liencr ties. It took t attacks in 969). )UAL rth insulin r (sucrose; nev. Some 1 also cane . vomiting ouble dose Iown milk ia to most ary tract. ' the bile, di. jrbances. id phenybstitutes. la free of 189 valine, leucine and isoleucine. These are only examples of a class of amino acid tonicities. FOODS AS VITAMIN ANTAGONISTS A. Biotin Antagonist -- Raw Eggs Raw eggs contain a glycoprotein called acidin, present in the egg white up to q 05% f total protein content. One molecule of avidin binds four molecules of biotin (vitamin "H"). This deprives the person ingesting raw eggs of an essential growth factor (Comments, 19S2-83). B- Thiamin Antagonists -- Seafoods, Vegetables. Fruits Fish such as carp, herring, swordfish and lake trout contain thiamin antagon ists (Goodhart and Shils. 1973, p. 415). There is also anti-thiamin activity in black berries, black currants, red beets, Brussels sprouts, red cabbage and spinach. There are antagonists for niacin in corn and millet, for pantothenic acid in yeast and for riboflavin in the akee plum. C. Multivitamin/Mineral Preparations Multivitamin/mineral preparations contain analogues for B|2 which are i claimed to be harmful to human vitamin Bi: metabolism (Herbert et al., 19S2). It seems advisable to omit B,. from vitamin/mineral preparations. Vitamin Poisoning -- Hvpervitaminosis 1 Only those vitamins are considered herein whose known or probable toxicides have strong documentation. Gilftian (1980, p. 1553) lists recommended daily vitamin allowances for the reference man, woman and child. Quantitative values for toxic thresholds are controversial. Poisoning does not occur with ordinary dailv consumption. The recommended adult allowance for vitamin A is 5,000 I.U. Poisoning occurs after ingesting up to 100.000 units (30 mg) daily for several months. Food faddists may ingest toxic levels. Infants need about 1,500 I.U, and doses of 18,500 daily for 1 to 3 months are reported to be toxic (Goodhart and Shils. 1973, p. 153). For vitamin D, the daily requirement for infants and adults is 400 I.U. There > is a danger in giving supplements such as 4,000 I.U, in fortified foods to infants (Goodhart and Shils, 1973, p. 160). Ingestion of 50,000 units daily to children and adults should be avoided for any long periods of time. For other vitamins, toxicities are either unknown or they occur at such repeti tively high values that they may be considered relatively safe. The reports are con flicting. GOITROGENIC FOODS Soybean (Glycine max) Goiters have occurred in infants fed soy milk or soy flour (Hydovitz, I960; van Wyk et al., 1969). There are also many goitrogenic foods in the genus Allium, but it is only excess consumption that may be associated with toxicity. Soybeans are toxic in other regards. One possible toxic factor, which is common to other legumes,.is a trypsin inhibitor. Trypsin is essential in the small bowel for pro tein digestion. The inhibitor, present in the raw bean, is inactivated by cooking. Soybeans are also hemagglutinins (phytoagglutinins or lectins) and this may be partly responsible for the gastrointestinal disturbances if raw soybeans are eaten. URL 03942 190 A valid explanation is still needed to explain the growth-inhibiting action in raw soybeans (Coodhart and Shils, 1973, p. 418). f TOXICITY OF FOOD ADDITIVES Preservatives and Sweeteners are Emphasized Below, Preservatives These substances extend the shelf life of foods. Shtenberg and Ignat'ev (1970) urge that sulfites and benzoates should be restricted, substituting sorbic acid. Sorbic acid and potassium sorbate are considered safe by the FDA. Sodium nicotinate, a fresh meat color extender, has caused food poisoning, expressed particularly in edema of the extremities. This additive is not legalized in some states. Sodium nitrite is used to prevent growth of Clostridium botulinum and to preserve.the color of meat and fish in salting and pickling. Nitrites are potentially poisonous in several ways, one involving methemeglobin formation. Another is in formation of N-nitrosamines, formed either in processing foods or in the digestive tract by reaction with amino compounds. By similar mechanisms, nitrites react with phenolic compounds, forming aromatic nitroso substances, some of these nitrophenols being suspected of carcinogenicity (Challis, 1973). Sweeteners There are some toxic synthetic sweeteners. Cyclamates were banned in 1969 because they produced bladder cancer in rats. Saccharin is now a regulated sub stance with suspected carcinogenicity. Crosby (1976) reviews several sweetening agents. Cane sugar as a natural sweetener is in a sense an additive and its local dental effects are of interest. Meals high in cane sugar favor colonization of cariogenic organisms on teeth. Restriction of cane sugar reduces bacteria such as Streptococcus mutans in dental plaques (Newbrun, 1982). Bibby and Mundorff (1975) blame acid candies, beverages and fruits for the highest dissolution of enamel. ^ . *1'p` '* it POISONING BY EXCESS MINERAL INGESTION A. Sodium Chloride (Table Salt) In the United States, the daily sodium intake per person is three or more grams. An estimated safe and adequate daily intake should range from 1.1 to 3.3 grams (48 to 143 mEq.) (National Research Council, 1980). A moderate restriction is one gram. The hypertensive influence of table salt is firmly documented and the FDA proposes labelling the sodium content of processed foods. Salt poisoning has occurred in hospitalized infants where salt was mistaken for cane sugar. Renal damage and convulsions followed. In the neonate, the repeated intake of skim milk has produced salt intoxication. B. Iron The daily iron requirement is 10 to 20 mg in adult males and 8 to 15 in females (Gilman et al., 1980). Finch (1982) states that a normal person can control the accumulation of body iron despite ingestion of 5 to 10 times the normal amount. Other writers note that excesses can be dangerous because excretion hardly increases, so that iron deposits in vital organs. Some wines of southern France contain 16 mg of iron per liter and even \meric m iron male of '.t'eondi Th mends demen \ itamir from ot Th. The pr Levels daily q called ' uuantit Ma dude f ithpi ildre r-- ater j S 3-*-}. ith di . ppm Th. p. 155* develop is clairr. from in 393: Sa Th< plants ; formini eiencv : Th influer not be Th (1977. tion in {1970 Sorbic ing. e\lized in and to entialk ner is ir, igesthi es reac >f thev in 1 Obed sul. etenm. 1 dent.i. logom f*roct i blanit ir mon 1 to 3, i .friction and tin aken In: epeatcr. feniair trol ti;' .mount hard!-. id eve;. 191 American wines supply several mg daily. The Bantus of South Africa brew a beer * in iron kettles and the beer contains more than 40 mg of iron per liter. A Bantu male often consumes 100 mg of iron per day in beer (Emery, 1982). A disease called secondary hemochromatosis, often fatal, occurs. The Food and Nutrition Board of the National Academy of Sciences recom mends a daily allowance of 10 mg daily for children in the first six months to sup plement milk. This may be too high. Also, it is inadvisable for an adult to take ' vitamin/mineral preparations containing iron if iron is being ingested sufficiently from other sources. . C. Calcium The minimal adult requirement is 800 mg per day (Gilman, 1980, p. 1553). The predominant.source is dairy products. Healthy serum contains 10 mg/dl. , Levels exceeding 14 are dangerous and 18 mg/di may cause death. Ingestion of ; daily quantities of calcium salts is ordinarily unlikely to cause toxicity. The so\ called "milk-akali syndrome" can be exacerbated, however, by ingestion of large quantities of milk. D. Fluoride Man obtains fluoride from ingesting plants and water. Incidental sources int elude food additives, e.g., baking powder. Combinations of fluoridated water, toothpastes and certain foods could conceivably produce a problem especially in -> children. This is a developmental problem in teeth where 1.7 ppm of fluoride in water produces mottled enamel before the teeth have erupted (Gilman, 1980, p. 1547). At 4 to 6 ppm, mottling approaches 100%. Severe mottling is associated with disorders in bone mineralization. Fluoridated drinking water is restricted to 1 ppm optimally. E. Selenium The estimated safe daily intake of selenium is 0.05 to 0.02 mg (Gilman, 1980, p. 1555). Where the soil and vegetation have a high selenium content, livestock develop "alkali disease" or "blind staggers." The human population in such areas is claimed, controversially, to develop gastrointestinal symptoms and liver disease from ingestion of meat, eggs, milk and vegetables (Goodhart and Shils, 1973, p. 393; Salunkhe and Wu, 1977). FOODS INDUCING MINERAL DEFICIENCIES There is a large literature on phytates as toxins. These commonly occur in food ts and they interfere with the intestinal absorption of essential minerals by Ing insoluble complexes with them. This produces the hazard of mineral defi ciency in parts of the world that rely on cereal proteins (Reddy et ai., 1982). FOODS AFFECTING CARDIOVASCULAR FUNCTION S'* The literature is voluminous concerning the development of atherosclerosis as Influenced by dietary excess of cholesterol and low density lipids, but these will liot be considered further here. A. The Effects of Coffee The cardiac effects of coffee are highly controversial. Lewis and Elvin-Lewis *3J1977' p- 181) cite data to the effect that the risk of developing myocardial infarc- ill % C20 r" OLtoO -U 192 tion is about twice as great for heavy coffee drinkers than for non-drinkers. This does not imply that caffeine is the cause since tea does not produce the same effect. B. Orange Oil Orange oil is expressed from the fresh peel of the ripe fruit of the orange. Citral is a constituent of this oil and is thus present in marmalade, fruit juices flavored with the oil, and orange drinks made by compressing the whole fruit. There is a question whether high consumption of such foods might damage blood vessels, as seen in experimental animals (Liener, 1980, p. 442). Citral is found to greater ex tent in oil of lemon, which is a flavoring agent. Goodhart and Shils (1973, p. 414) discuss the controversial aspects of citral toxicity. C. Licorice (Glycirrhiza glabra) Hypertension associated with salt and water retention leading to edema follows licorice ingestion to excesses over time (Koster and David, 1968). Koster's patient ingested 100 grams per day. The noxious agent is a steroid-like compound called glycirrhetinic acid. D. Cheese, Drugs, Wine and Beer Monoamine oxidase inhibitors are in widespread use to elevate the mood. Tyramine, present in certain brands of cheese, wine and beer, can provoke a hypertensive crisis in patients treated with monoamine oxidase inhibitors. It takes only 6 mg of tyramine to provoke hypertension in such situations (Sen, 1969), and 25 mg (which occur in many foods or alcoholic beverages) is a very dangerous level (Hutchison, 1968). The ripening of cheese breaks down casein, one product being tryosine. Bacteria involved convert tyrosine to tyramine. A cheese and wine party could be disastrous if high tyramine substances such as cheddar cheese and Chianti wine were freely consumed. There are other high tyramine cheeses as well as wines and beers (Sen, 1969; Bassuk and Schoonover, 1977, p. 41; Maga, 1978). Tyramine releases pressor catecholamines (Editorial, 1965). It is interesting that some varieties of ripe cheddar and Cheshire cheese, plain chocolate, and to a small extent some wines, precipitate migraine attacks and this is due to 2-phenylethylamine (Chaytor et al., 1975). FOODS DISTURBING NEUROLOGIC MECHANISMS Certain foods are cholinesterase inhibitors and thus disturb nerve impulse transmission. Such foods include potatoes (solantne effect), roots of tomatoes, beets, apples and eggplants (Salunkhe and Wu, 1977). Bananas in excess affect transmission of impulses because they contain serotonin and catecholamines. Some foods produce hallucinations. Examples are nutmeg seeds and mace, which contain myristicin. The methylated xanthines in tea, coffee, cocoa and cola drinks overstimulate sensitive individuals. The drugs include caffeine, theophylline and theobromine. The neurologic effects of coffee are very controversial. A 5-ounce cup of coffee contains up to 90 mg of caffeine which can produce nervous excitement and cardiac stimulation, but only in a sensitive individual. Ingestion of one gram (11 cups) could seriously affect such an individual. The potential toxicity of caffeine is offset by the fact that when ingested in ordinary amounts it does not accumulate excessively in the body. O' Our discussior .ind drug interacts anticoagulants, tr< .is allergens is exte Food contaminati poisoning of anim man is under perj Miral. Y.P.. 1967. Stu< iFrunus amygdalus Menander, C.S.. 1969 Kfena. J.M.. 1979. Po Itxvsuk. E L. and S?h Medical Book Co.. Bonder. A.E. and Be; reference to lectins. Bhide. S.V.. Pratap. r uf nitrosamines in a Ihlihv. B.G. and Munc :;."-Linn.C.. Santa At C mi colorie absorptic 3D 1 Tallis. B.C.. 1973. F r" nitrites. Nature 244 1 li4>tor. J.P.. Crathr (O U'thvlamine in food ^ 'mments. Winter i( : "mmittee on Diet. N National Academy 1 "*. T.M.. Camilleri. if fructose intoierai f.rosbv. G.A.. 1976. 1 Crosby. W.H.. 1979 < rimt. J.R. and Sjoer amines in bananas i>ilion. J.C.. Martin Toxicol. 19-.437-44 Oiitoral. 1965. Pressi Emery, T., 1980. In Epstein. S.. 1969. H' rheumatism- Ann. jI i-'jirbairn. JAW (Ed finch, C.A.. 1982. 1 I Gilman. A.C.. Cood Basis of Therapeut \ Glushy. J.S.. 1975. fc * Gleason. M.N.. Goss* Products. William Goodhart. R.S. and t Philadelphia. Gough. T.A., Good; cooked bacon. J. h Groilman. A. and C Philadelphia. . Harvey. R.C.. 1982 Kavery. D C.. Hotc Sci. 46:501-505. Herbert. V.. Dmes. vitamin Bt, may irinkers. This - same effect. -range. Citral lices flavored it. There is a od vessels, as to greater ex1973, p, 414j ;dema follows >ster's patient ipound called te the mood an provoke a itors. It taken. 1969). and ry dangerous isine. Bactern. >e disastrous if e were free id beers (Srr nine releast cheese, pla tacks and tl >MS erve impulse of tomatoes, excess affect olamines. Is and mace. >verstimu]ut< heobrojnim cup of cofte< t and cardiu; 1 cups) coulu e is offset h\ te excessive 193 OTHER ASPECTS OF FOOD POISONING Our discussions hardly scratch the surface. There is a large literature on food and drug interactions, making physicians take caution when prescribing antibiotics, anticoagulants, tranquilizers, and certain laboratory tests. The literature on foods as allergens is extensive. Problems in irradiation of food are still greatly unsolved. Food ^contamination by microorganisms is a continual and severe threat. The poisoning of animals by food toxicants and the transmission in the food chain to man is under perpetual surveillance, REFERENCES Abrol. Y.P . 1967. Studies on the biosynthesis of amvgdalin, the cyanogenic glycoside of bitter almonds {Prunus amygdalus Stokes). Indian], Biochem. 4:54-55. Alexander. C.S.. 1969. Cobalt and the heart. Ann. Int. Med. 70:411-413. Arena. J.M.. 1979. Poisoning. 4th ed., C.C. Thomas. Springfield. IL. Bassuk, E.L. and Schoonover. S.C.. 1977. The Practitioner's Cuide to Psychoactive Drugs. Plenum Medical Book Co.. NY. Bender. A.E. and Reaidi. G B.. 1982. Toxicity of kidney beans (Phaseolus vulgaris) with particular reference to lectins. J. Plant Food 4:15-22. Bhide, S.V., Pratap, A.!., Shivapurka, N.M.. Sipahimaiani, A.T. and Chadha. M.S., 1981. Detection of nitrosamines in a commonly used chewing tobacco. Fd. Cosmet. Toxicol. 19:481-483. Bibby, B.G. and Mundorff. S.A., 1975. Enamel demineralization by snack foods. J. Dent. Res. 54:461470. Bo-Linn. G.. Santa Ana. C.A.. Morawski. S.C. and Fordtran, J.S.. 1982. Starch blockers -- their effect on colorie absorption from a high-starch meal- N.E.J. Med. 307(23):1413-1416. Challis. B.C., 1973. Rapid nitrosation of phenols and its implications for health hazards from dietary nitrites. Nature 244:466. Chavtor. J.P.. Crathorne, 8. and Saxbv, M.J., 1975. The identification and significance of 2-pheny- lethvlamine in foods. J. Sci. Food Agric. 26:593-598. Comments. Winter 1982-83. U S. Biochemical Corp., pp. i-2. Committee on Diet. Nutrition and Cancer. 1982. Diet. Nutrition and Cancer. National Research Council. National Academy Press. Washington. D C. Cox. T.M.. Camilleri. M.. O'Donnell, M.W.. and Chadwick. V.S.. 1982. Pseudodominant transmission of fructose intolerance in an adult and three offspring. N.E.J. Med. 307(9):537-540. Crosby, G.A., 1976. New sweeteners. CRC Crit. Rev. Fd. Sci. Nutr. 7:297-324. Crosby. W.H.. 1979. Carcinogens in food and the Delaney clause. J.A.M.A. 24l(6).-6I7-619. Crout. J.R. and Sjoerdsma. A., 1959. The clinical and laboratory significance of serotonin and catechol amines in bananas. N.E.J. Med. 261(0:23-26. Dillon, J.C.. Martin, G.B. and O'Brien, H.T., 1981. Pesticide residues in human milk. Fd. Cosmet. Toxicol. 19:437-442. Editoral, 1965. Pressor attacks during treatment with monoamine-oxidase inhibitors. Lancet 1:945-946. Emery, T., 1980. Iron metabolism in humans and plants. Amer. Scientist 70:626-632. Epstein, S., 1969. Hypersensitivity to sodium nitrate: A major causative factor in case of palindromic rheumatism. Ann. Allergy 27:343-349. Fairbairn, J.W, (Ed.). 1976. The Anthraquinone Laxatives. S. Karger. NY. :h, C.A., 1982. The detection of iron overload. N.E.J. Med. 307(27):1702-1703. Hmin, A.C., Goodman. L.S- and Gilman, A., 1980. Goodman and Gilman's The Pharmacological Basis of Therapeutics. Macmillan Publishing Co., NY. Gtasby, J.S., 1975. Encyclopedia of the Alkaloids. Plenum Press, NY, Vol. 1. Gleason. M.N., Gosselin, R.E.. Hodge, H.C. and Smith, R.P., 1969. Clinical Toxicology of Commercial Products. Williams and Wilkins Co., Baltimore. art, R.S. and Shiis, M.E., 1973. Modern Nutrition in Health and Disease. 5th ed. Lea & Febiger, Philadelphia. igh. T.A., Goodhead, K. and Waltere, C.L., 1976. Distribution of some volatile nitrosamines in cooked bacon. J. Sci. Fd. Agric. 27:181-185. CroHman, A. and Crollman, E.F., 1970. Pharmacology and Therapeutics. 7th ed. Lea and Febiger. Philadelphia, Harvey, R.G., 1982. Polycyclic hydrocarbons and cancer. Amer. Scientist 70:386-393. Havery, D.C., Hotchkiss, J.H. and Fazio, T.. 1981. Nitrosamines in malt and malt beverages. J. Food - Sd. 46:501-505. Herbert, V. , Drives, G., Foscaldt, R. et al.. 1982. Multivitamin<'mineral food supplements containing vitamin B,2 may also contain analogues of vitamin B12. N.E.J. Med. 307(4):255:256. cn 194 Huang, DP. and Ho. 1981. Volatile nitrosamines in salt-preserved fish before and after cooking. Food Cosmet. Toxicol. 19:167-171. Hydovitz, J.D.. I960. Occurrence of goiter in an infant on a soy diet. N.E.J. Med. 262/7) 351-353. Johnson. F.C.. 1971. A critical review of the safety of phenolic antioxidants in foods. CRC Crit. Rev. Fd. Technoi. 2:267-304. Kerdel-Vegas. F.. 1964. Generalized hair loss due to the ingestion of "Cocode Mono" (Lecythis ollaria). J. Invest. Dermatol. 42:91-94. Khanna, S. K.. Singh. G.B. and Krishna Murti, C.R., 1980. Toxicity profile of some commonly encoun tered food colors. J. Food Sci. and Technoi. 17(1 and 2):95-103. Kinghorn. A.D. (Ed.). 1979. Toxic Plants. Columbia University Press. NY. Koster. M. and David, C.K., 1968. Reversible severe hypertension due to licorice ingestion. N.E.J. Med. 278(25): 1381-1383. Koelling. A.C., 1982. Please don't eat the daffodils or the Datura or the Dieffenbachia. The Living Museum 44(3):36-39. Kuhlmann. \V.. From me. H.G., Heege, E.M. and Ostertag, W., 1968. The mutagenic action of caffeine in higher organisms. Cancer Res. 28:2375-2389. Lee. V. A. and Lorenz. K., 1979. The nutritional and physiological impact of milk in human nutrition. CRC Crit. Rev, Fd. Sci. Nutr. 11:41-116. Lewis, VV.H. and Elvin-Lewis. M.P.F., 1977. Medical Botany. John Wiley and Sons, NY. Liener. I.E.. 1962. Toxic Factors in edible legumes and their elimination. Am. J. Clin. Nutrition 11:281-298. Liener. I.E.. 1974. Food Science and Technology. Academic Press. NY. Liener, I.E.. 1980. Toxic Constituents of Plant Foodstuffs. 2nd ed., Academic Press. NY. Maga. J.A., 1978. Amines in foods. CRC Crit. Rev. Fd. Sci- Nutr. 10:373-403. Marshall. E.. 1982. EPA's High-risk carcinogen policy. Science 2I8(4576):975-97S. Maugh. T.H.. 1982. Cancer is not inevitable. Science 217(4554):36-37. Merck Index. 1960. Merck and Company, Rahway. NJ. National Academy of Sciences. 1973. Toxicants Occurring Naturally in Foods. 2nd ed.. Washington, D.C.. pp. 255-257. National Research Council. Committee on Dietary Allowances of the Food and Nutrition Board. 1980. Recommended Dietary Allowances. National Academy of Sciences. Washington. D.C. Newbrun. E.. 1982. Sugar and dental caries: A review of human studies. Science 2l7{4558):4l8-423. Okonkwo. P.D. afid Kinsella. J.E., 1969. Orotic acid in food milk powders. Amer. J. Clin. Nutr. 22:532-534. Osborne. B.C., 1982. Mycotoxins and the cereals industry -- a review. J. Food Technoi. 17:1-9. Pensabene, JAV., Fiddler. W.. Cates. R.A.. Fagan. J C. and Waswerman, A.E., 1974. Effects of frying and other cooking conditions on nitrosopyrrolidine formation in bacon. J. Food Sci. 39:314-316. Pierson. LA. and Smoot. L.A.. 1982. Nitrite, nitrite alternatives, and the control of Clostridium botu- linum in cured meats. Crit. Rev. Fd. Sci. Nutr. 17:141-187. Reddy. N.R.. Salunkhe, D.K. and Sathe. S K., 1982. Biochemistry of black gram (Phoseolus mongo L.). CRC Crit. Rev. Fd. Sci. Nutr. 16:49-114. Rice. S.L.. Eitenmiller, R.R. and Koehler, P.E.. 1976. Biologically active amines in food: A review. J. Milk Food Technoi. 39(5):353-358. Robinson. JL.. 1980. Bovine milk orotic acid: variability and significance for human nutrition. J. Dairy Sci. 63:865-871. Salunkhe. D.K. and Wu, M.T., 1977. Toxicants in plants and plant products. CRC Crit. Rev. Fd. Sci. Nutr. 9:265-324. Schmutz. E.M. and Hamilton, L.B.. 1979. Plants that Poison. Northland Press. Flagstaff. Arizona. Seenappa. M. and Nyagahungu. I.K.. 1982. Retention of aflatoxin in Ugaii and bread made from con taminated maize flour. J. Food Sci. and Technoi. 19(2):64-65. Sen. N.P.. 1969. Analysis and significance of tyramine in foods. J. Food Sci. 34:22-26. Shtenberg, A.J. and Ignat'ev. A.D.. 1970. Toxicological evaluation of some combinations of food pre servatives. Fd. Cosmet. Toxicol. 8:369-380. Speckmann. E.V., Brink, M.F. and McBean, L.D.. 1981. Dairv foods in nutrition and health. J. Dairy Sci. 64:1008-1016. Van Wyk. J.J.. Arnold, M.B., Wynn, J. and Pepper, F., 1959. The effects of a soybean product on thyroid function in humans. Pediatrics 24:725-760. Walters. C.L. and Smith. P.L.R.. 1981. The effect of water-borne nitrate on salivary nitrite. Fd. Cosmet. Toxicol. 19:297-302. Warin. R.P, and Smith. R.J.. 1976. Challenge test batterv in chronic urticaria. British J. Derm. 94:401-405. Wasserman. A.E.. Pensabene. J.W. and Piotrowski, E.G.. 1978. Nitrosamine formation in home-cooked bacon. J. Food Sci. 43:276-277. Widdus, R. and Busta. F.F., 1982. Antibotulinal alternatives to the current use of nitrites in foods. Food Technoi. 36(12):105-106. Yannai. S.. Toxic factors induced by processing, pp. 371-427. In Liener. I.E., Toxic Constituents of Plant Foodstuffs. 1980. 2d ed.. Academy Press. NY. . -- } * ? ` * ; Transactions of the Illinc 1983). Volume 76, 1 am R. Cult Proper soybean Auction costs may b field trial was condi -ovbean cultivars oi -ultivars `Amsoy 7 respectively) were s in 73-cm rows. The messic Tvpic Ochra lowest pod height, densities at maturit each year as a resul greatest for the sm the cultivar. The c1 height and lowest [ <et its lowest pods There was a consist number per plant .fields were influen Essex', the highest largest pod number the lowest vields, bt Soybeans [Gfi/c different plant spat ever, it is important The quantity of se cultivar, seed viab (Morse et al. 1949 URL 03948 liMi / MKKVCINI In till olhei instances, very hide is known .< I n u i die sequcine ol events taking between exposuie to (lie pesticide .mil appr o a m e ol minors in animals, ( he chronic toxicity ol 1)1)1', tIn- major metabolite o| 1)1)1 in liiniiaii Ih iiif-s, li.is not Fein tin..........:dil\ umlied li seems 11 .i I jn < ompai is< m lo liiniiaii beings, in miie ami tuts 1)1) I is converted more readily io 1)1)1) llian lo DDK. No iiiinoi pioilin lion was obtained in nine with either pp'-DI)D or op'-DDI) nndei (omlitions compatahle lo ihose which led lo die appeaiame of hepatomas InMowing pp' DD I Knowledge on iniiat ellnlai binding ol 1)1)1 is extremely sc ai c e. The estrogriticity ol' DDT, particularly that of the op'-isomrr, is cs lahlished ami is probably parity responsihle lot the (llects of DDK on wild life. F.nviionmcni.il estrogens are considered to have a iole in the total can inogenic load, ft is not known whether in the experiments with DDT leading to the appearance' of tumors also a hormonal imbalance was induced. In addition, op'-DDI) has long been Known lo piodme ad renal changes in dogs and man; an effect on cortisol iiieiaholi.sm hy op"- 1)I)T and technical DDT lias also iccenlly been reported. Other |K)ssible effects of DDK on lahoramiv animals which under completely didereni conditions have been postulated to be somehow re lated to can biogenesis aie iimmmoileptrssioji ami inlei leiem c vvilh vitamin A inclabolism. Effects Possibly Related to Carcinogenesis Exerted by Pesticides Which Have Not Produced Tumors in Long-term Experiments Mechanisms of carcinogenesis for man ami animals aie laigely un known so that, in general, it is not possible to assume dial < hcnmals hav ing pari Millar proper! ics represent a eaic inogenic hazard. For instance, biological alleviating agents or mutagens desene piimilv in a piogiani of evaluation of carcinogenicity, but it cannot lie siaied that these prop erties per sc indicate carcinogenic ity. A cliHer'cut < omlilion is goit togenic ily, since this clfect can be (it into one theory on the mec hanisms of car cinogenesis. Several dithiocarbamates used as fungicides (lor instance, maneh and zirani) have produced thyroid enlargement vvilh nodular hy perplasia in rats and occasionally in dogs, similar in ihose produced by ihiouiacil. Tinselled occ lured at liigh doses hill (lie quantitative aspects ol the phenomenon are not sufficiently known. Finally, enzyme induction and enzyme inhibition can be elicited by pesticides. They are not likely to be directly coiielated with carcinogene sis. There is some experimental evidence that these' compounds can in terfere with the metabolism and larrinogcnit ity ol some known carcino gens. Evaluation a fnioii of risk or henclit is exliemely dillu tilt i<ll> ( il\l VMIXVXlS VMH.VKC liSOl.KNt-sh *r,i Food (contaminants and Carcinogenesis R (I SSL LI, <). SINNIllJliKR l>, /'in/mint a) hun I Siltntr ailii I if y. ( ' mrri \tly. (-on'iilln, (higun, IS I Wfr I in nn i or Mom kn in.ui ami his animals is composed ol a wide v ai icl) >l loods gathered Jiom all pails of (he world. These foods have hern giovvn, haivested, and piocessed, sometimes hy rather primitive means. 1 he hazards of transportation and storage, coupled with the factor ol lime have an impact on the quality and wholcsoincness of ihc final prod uct. One needs only to walk through a supermarket to lie avvaie of the complexities which make up man's food today. On the shelves and in die hce/et cases, we have a vast array of foods com.lining a mixture' of seasonings and additives, and piocessed in a multitude of ways. Many ol these; are combinations of loods wdiicli have been piocessed, packaged, stored, ami later sold, perhaps lo lie stored again and finally he ,iled and consumed--'with other processed loods, equally as complex. The evolution that has taken place in our Inotl preparation assumes that the producer and processor have taken the necessary steps lo assure llie safety of all the ingredients which comprise the food and in the li nal product itself. This responsibility requires the combined efforts, tal ents, and knowledge of many specialists trained to iccngni/c the poten tial hazards which may exist or develop in food and its components. Unexpected incidents involving lood contaminants have occ lined even with our advanced technology and close attention to lood salei). A case in point, foiinnately with fish and not man, was (he IH4iO-l4J(il epizootic of liver cancer in rainliow trout. This incidence of liver can ccr was correlated with a change in dict'lrom one composed largely ol raw meat and (ish to a computerized, highly nutritious diet of dry in gredients processed into conveniently sized pellets. One of the ingicdients, cottonseed meal, contained the lood contaminant allatoxin, and rainbow trout proved to have a liver that was highly sensitive to this mold metabolite.80 Food contaminants to be discussed ate those lor which their is some evidence of carcinogenic activity. These may he hioadly classified in the following groups: K (,onl,iiiiin,mls initnchucd dining pioduciioii <>r processing, e.g., ben/o (a)pyrene. 2. Contaminants lormed hy the interac tion of substances in lood, e.g., niliosummcs. 3. Contaminants lesulting fiom the action of biological processes in or on foods, e.g., mycotoxins. 2 111rlH / MXXIII II URL 03949 Food Contaminants Introduced during Production or Processing A immmIhi nl polii yi lit .11 <tiii.il 11 \ i In x ,i 11" ms li.ivr I m in <h iimoM i ,il ed in i;nhc ciikci in expel inicnl.d animals. ll mm Ik- dial lirn/o (,i) pyrene is flic most universally disiribnicd lammigrn in unr cmiionrnent. hxhousl guscs limn the limning <>l lo.ssil I mis .iml ihe minims l inn engine < on tribute i lie inaji*i share >1 lien/n (.i)|hiciic aim li Milne i|ticillly finds jis wav in man. Ids animals, and their Innd. A intent it: port by Shabad7K o| studies (nntlueted in the (I.S.S.K. show (he distiibulinn, riit ulatinn, ami late nl hvilint at linns, pal tit ul il ly ben/o (a)pymie. in the model n cm imimttin. For com, sime man has used lire In took his Innd, hen/n (a)p\ i cm lias Ikcii pieseiil as an unrecognized loud tmilainiiiani. Smoking nl meal and jish, min nl (lie oldest incihnds nl Innd piesei \ at inn. m;i) al so contribute ben/o (a)pyrene anti nlhei poiytytlit hydrocarbons. The popularity of ihe backyard b;n hectic anti charcoal-broiled meals lea tnred in American restaurants disregard ihe ret out findings that these procedures may toimibute measurable amounts nf tartinngens in nut fond. Table 6-5 piesenls a list nl sumketl anti hatlietticd IoihIs willi ihe levels of ben/o (a)pytene found. An extensive review nf pnfytvefic aro matic bytlrocarhnns in foods was rctendy prepared by Flowaid and Fa zio.:m ShabadTM repotted that people who tmisiune huge tpianlities nl smoked food or work in meat smoking hit lories show a pat lit ularh high incidence of digestive tamers. In Midi, l)imgal,H suggested that a correlation existed between the mnsutnpfinri <1 smoked million and trout and the incidence nl gastric tauter in belaud. Dungal's nhserva lion, probably mote than any ntini, served in .stimulate a series ol inves tigations on the possible iclalionship of polyivrlii aromatic compounds in food and gasitoiniesiinal lancer.77 KM In Japan, where the incidence of gastric tamer is nnitpiely high, workers have suggested that the common piatiite of (hartoal broiling fish may |irocliice carcinogenic compounds.11-1 l.ijinskv and Sliul)ikr>-' r,:' found hen/o (a)pyrene in hat bet net! spaicribs ami t hart oal-broiled steaks. In a later report, f.ijinksy and Boss-'*1 * i*nvestigated ihe ellecl ol looking methods on (he production of benzo (a)p\icn<: and polynucleoi hydrocarbons in meal. Thc-s related the piodmiion ol hydi ot ai bons with: (1) the pyrolysis of fat, (2) the fat tnnieui ami exposure of the food to tlve Haines, and (H) the closeness to die heal sounc. Ben/o(a)pyrene was not produced when die heat souite was above the fond or il (he melted fat was kept Iroin dripping into the heat source. Howard e( al.:i7 an.dy/cd a number of smoked food products hn 1k`U/o (a)p\icue and lomid levels varying limn O.H to `5.2 pg/kg. Kc cently, investigators in the II. S. Department of Agriculture and l1 * S. Food and Drug Administration'1- conducted an extensive survey in astei lain the types and amounts of jiolycydtc aromatic liytlrotailxms preseni in 60 smoked products and miscellaneous foods. They selected samples 250I tlOII CONTAMINANTS ANIl CAKCINtMtt'.NKMS / 5 0 1| Alt! I- li .--Hl.N/ (,\) 'lKI M IN SmOKIK I'IKHi 1 t I'l <11 1 <11111 M mo i <ii Trmii toil Hnllish VlitUOli 1 l.uldm k Satisa^i loti M 1111 tilt (lll'llU' |lll|l.) Million (i<mnm-[< i.ih 1-isfi, illicit lulling listi, \ini|>t-i)n 1 (.mi IlnlM- lll.irkciil .Nltet?|> heads (sinned) Sea hints (silled) 1 lam ( .od Whitiiij; li.tl Ih i m d heel Sausage Hamburger fork chops (liicki.ii Sirloin T-bone slcak (gas) T-honc slcak (chauoal) HknzoOJi'j mst. r<./*'.) R f H HI Nl t.H 7 2.1 .7 0.3 a 0.R ;7 1 .7:i ' 0.3 r.:i t.'i-ior. :i 17- 7.5 :to L'ti 77 i 77 i <i :i7 (IK 37 .'1.2 37 (i r> 0.7 Hint: 28 HH 20-99 88 0.7 02 4 0 02 n.o 02 ,s.:i 02 o.i 02 2fi .71 7.9 .71 3.7 ,31 11.1 31 4 1 31 70.4 51 to lepresent the wide variety of treatments used in producing smoked foods ami found values of ben/o (a) pyrene from 0.1 to 7.0 flg/kg of footl. On tlic outsitle (skin) of smoked whiling and whitelisli, they re ported levels ol 70 and 53.6 pg/kg, respectively. Ii stems apparent from these studies that ben/o (a) pyrene is detived liom the iueoni|)lete tomhiistinn of wood products in the case of smoked foods, ami liom the pyrolysis of animal fat in the barbecued samples. In addition to smoked products, other foods have been shown to tonlain polynuclear hydrocarbons. Benzo (a)pymic,!i was found in heav ily roasted toilet: anti ii was reported to accumulate in oysters taken from polhued waters.11 Jung ant! Moraml;iU reported the occurrence of beu/n (a)pyrcne as well as other polycyclic hydrocarbons in vegetable oils. This was later confirmed by llownni et ui.,B who described an assay meth od and reported ben/o (a)pyrene levels from 0.4 to 1.4 ^g/kg. The ori gin of these compounds in vegetable oils remains to be determined, bin it appeals that the environment may be ibe source. Ben/o (a)pyreiu: has been found in soil at levels up to 21,150 /ig/kg, presumably derived liom iIk lalloui ol (oinlmslinii products, hugely liom engine exhaust.7'1 We have, at this time, no direct evidence tliai the ingestion of benzo 13 3 ^ It -- s3 2Z-= = - 3 = r i- s S Ia.srl-SSSsi -"S' 3*7-3 ? 3 a i ? S II * s* - a I g ^ 1_ n S] wl 3 i|sa>o w2 E-s. a. re -;=>i < 3 " s i 5' --ft i-s-ii 5* 35O - 3' ft , 1s 2 , ^i3 I !: H!t3H3 3S' T ? , ? * x, -a = sH-3?sa.3;s-:? s'ss is n g 1 = ? ^ r = n ~2 2 3 2-'Ja w x-f?nr2"rsse*,t---f--3S-1"=x<-':t, j3-->"afSa5_r2Zn<eCt-.. "-3nf^OB~=o'^=.ak--f"S"t3.-'.53--snz--fM--f--S0--_a^t?t;)i.' 3--3H"-Q--ffraa3<3ett?*-,3-f"-^xa3--=t3fiyr2----rtt2es-I;`-3J4=53302r--c"M*' 5a3--?a!--5a>2i=' J="aS3_ar.S'5a'3Zi23o3c--3-2e.'3233--307Or^a_ss>3J-1?".s|=5ijfr>x<;t^<-,,,-?>2<=I--^J<--*-sTXS.a_?3Cf--n----^>t--,i.y?a7=a2r;u"<f"e't5r.2' *s--^o-3sf"Ot' 3-=2a3S3i.a'---i--Ss3fof5tg2t^'?.a2.<_5w2-?^f3vu>t 3`^s=--5f-aft".}*^a--93fS--3t'~1i_3a353^i'i.-3Ss53-f'fa3tet*2i2-rSq.ff5^OS3i~tt,1--''o--f1"a3uJf--_frttu_.t?^3"fSS_~f3=iCtt` =2f<f_=tt5i'--xSx5-Xfs<t. sr-~s25z^-f3fxStt"i_".'azc*'a'ss--^^S:Sia=^~=a---:s=32-P,--xs^f--y-t--=--:a=z-~2ir-^-2;a--r*aG.--aZ?-= 2,_ 3 - - " a3--^S :3o,^3=- =2S -. a g-S'O 3iV33s?9*a--*"c*-c5553 3#g'*vSB`-g3J21g--reS----<s.--i"^--EPC-3ftZ5{*25a <'Si^'JO _M3. -- 3 3 w,, X3^jHj TM32^1 iX-ii^A*33-v "2 -5 =---- I -S 3- a (i 2S * 1 g 5 * a ? >i^r-=asa.s s'! s.r-za.5. s 2. 3. X = O oft3a a3 re:re 3 >< A -- = 1 ffi J6 n -. "^ = r: 2 3 jq -- *** 2 ^ - S. - - -= 3 = = i 2, s =' 7 =| ; ^ 3 5 =.3S' ft W jS` _. = a .. _r r ft 3 ^ 5*I = S ='5 i -3 ^ = a-" 1 5= 2. 2 = ' =: " ?=3"'^r i"l>p s*5 J'* is 22i-i r I E 3 I 1 - i. - = ' f fj K 2 a21'?3 = ~:?rXft* , ^ ^ ft = 3 * a- 5' 3. 3 3 ~ ^ 1 .5I' " w 2. - i *x> 5" 5 I i1 * 3I 2 Z f 15 3 5- =-r^ .??! =' f ! | ST . I X = J= 3 3/ ~ Er s c 5 = WS~ '-S. ~ L zs == s J-- - = * 3 =11=1 --3 M s a = = 1 3 ?3 r = ft -e 7ft - 3 i. O ft =.2 2isirft =~ 5vi z_ba - nwz%3?z ^ixr *S*=?r aS-sflSi 22 055'5*3`-- ' -- ft ; 1 : ?o : 77* S 2. fi?3l.35?""s jrZiB-a.=3-_^=` ^` ^2 5- 2.o_: r.s_Sa S-f-tr-i"3a"a.soos|l;?C-c' =Hx?5^^sn5s'i^*s?1" 5j; 52S,3 ii . ? :- 22 _. == H=. ^i's.HJI.JI!: 13-3 3^- =Z----. -5 a 3S".^1* i5-Z^ = = = r-s--"2-ft2s**^h8"5.C --ftr==ft* "55=:2?.==*.sf_X'itI--55=2 5- 2 < a. P -. -5_2--ft~=Z3^3a "3r5s^XftXEsE'ft ~h~S *iZ2 ~: K z: zl r, I --^ i 5 i ^ 262 / SINNHI/HKR U&L 0395? lems ol analysis were discussed. Levels of N-nitrosamincs ranging fmm 10 to 100 ^g/kg were ited hut remain to he (onfii med. l ire |R>ssil)ilily of die loinuiiion cd N nitiosamiru's in roasted foods was suggested hv Devik " lie heated 0.01 mol o| amino arid and 0.005 mol ol D-ghuosc absorbed onio 20 g potato statili lot several hours and reported a yield ol N-MMiosamines from 0 to 1.000 /ig. The N-nitrosamines were detected hy a poiarographic procedure. Workers at Oregon .State Univeisiiy7- were unable to confirm these liudiugs an<l a report ol their work is in preparation. It is apparent front these reports that methods for the detection of ainiiies and N-niitos.miim-s are lar king in sensitivity and specificity. The increasing occurrence ol nitrates in mn environment, particularly in our water and food supply as a direct result of agricultural practices has been the cause of growing concern, In a review of the subject, Fassett-7 recounted a number of cases of prisoning which were attributed to toxic levels ol nitrates and nitrites in food and water. Later Phillips"" discussed the nitrate problem and its public health implications. He re viewed data citing nitrate levels of 1,071 to 1,668 ppm for spinach and 034 to 2,165 ppm for beets in prepared infant foods. The preservation of fist) and meat with salt has its origin in Anlupiily Later rr it rate was added to meat to retard the growth of anaerobic bacteria and to produce a desirable red or pink color. Nitr ile has been permitted in meal as a food additive since 1925. In 1939-1910 Tarr and Sunder land"8* fl6 recommended nitrites as a preservative for wet fish and suggest ed levels of 200 ppm. Its use was |>ermitied in Canada until 1956 when it was rcplaicd hy tetracycline antibiotics. The United Slates restricted the entry of Canadian fish which had been treated with nitrite, but per mitted 10 ppm sodium nitrite in smoke-cured tuna fish products proc essed in the United States. Ai the present time, up to 200 ppm is per mitted in smoke-cured sablefish, salmon, and shad. These three species may also contain up to 500 ppm of sodium nitr ate in the finished prod uct. 'l ire use of sodium nitrite of not less than 100 ppm and not more than 200 ppm has been proposed for smoked cluib. Lite purpose of ni trite addition to smoked lish is to give an extra measure of protection against the formation of botulinum toxin. It is difficult to reconcile the use of nitrite in lish products in view of the happenings in Norway. F.uder and Cch32 stared in 1967, "The prac tice of dipping fish fillets in diluted NaN(>2-solutions in order to avoid spoilage--should Ik: deemed highly risky." Since it is known that fresh water species contain a lesser amount ol trimethylamine oxide and amines than marine species,this may provide some jusiification for the use of nitrites on fresh water varieties. Re cently, Japanese workers demonstrated that a number of species of fish contain enzymes capable of converting the tertiary amine, trimethyla- inine to dimcthylamine and formaldehyde.81 Other secondary amities are present in fish such as the amino acids pvo- IOOIMONI V \l | \ \\|N WIICVKCIXCX.I-Nlsls / 2n:J line, h)dio\ypioliiH. hisi idinc, 115 pioplr an, and aiginine, Imi liule is known about the nticiiingeim properties ol iheii ,\ ntlniMi derivatives. I lypoxanlhinc, a secondary amine fotined from nucleotide degradation in lish dur ing sloragc may accumulate ai high levels'4- and a nilrosamiiic may result from I In- leaiiiou with nilrile. Iloylaiul el al.H irpoiird iiilin sopiper idine to he an active carcinogen ami niirosopinlinc to be prac tically inactive, though no experimenial data were presented for nitrnsopioljne. High temperatures, such as those encountered in the cooking ol lish, as well as the addition of arid condiments would favor the reaction. However, Fndei ci ai.-- reported the lonuaiion ol niliosamine from di methylamine and nitrite will occur even at temperatures of OnC and at a pH of 6.5. *1 lie recognition that N-niliosamines may lie lorined horn a varied of comjioumls, many of which could he present in meal ami othei foods, adds a new dimension to the use of nitrite as a food additive. As far as I am aware, no N-iiitrosainincs have been reported to occur in meat products as the result of nitrite addition. Much of the niniic binds with hemoglobin and myoglobin; however, some remains as nitrite and would be free to react with secondary amines. In 1963, Dnukrey Cl al.,T recognized the possibility that the acidic con dition of the stomach c:outd provide the necessary criteria for reaction between nitriles and secondary amines. In a scries of experiments with rats, they were unable to induce earner or cause any teratogenic: effects after feeding sodium nitrite and dieiliylamiiic lor an entire life span and over two generations. || was suggested by them that tbe [ill of I he i at stoma h wide h is c lose to I ami llie basic ii y ol din hy lamine could be an explanation for the lack of biological activity. Sander et ai.71 in 1968 detected N-niimsainine when they incubated nitrite and various second ary amines with human gastric juice which had a pi I of 1.3. Sen et al.Tr* in an improved method for the determination of N-iiitrosainincs in foods confirmed the findings or Sander. I.ater they74 demonstrated dial dielliylnitrosamine was formed when sodium nitrile and diethylainine were incubated with gastric juice fmm rats, rabbits, cats, dogs, and man. Gastric: juices from man and rabbit with a more acid p11 were- teporn-d to give a higher yield of N-nitrosamine. They were also aide u> delect N-nitrosainines formed in vivo alter feeding lire reactants lo cats ami rabbits. It was suggested that other species which have gastric juice with a pH similar to that of man be used instead of rats as experimental ani mals lor these studies. Sensitive* procedures lor die detection, character ization and quantitation of N-nitrosaniincs ami their precursors ate needed, particularly in food systems. A combination of gas chromatog raphy and mass spectrometry would appear to offer a promising ap proach to ihe analytical problem. In conclusion, it seems from ifie accumulated evidence of the past decade that the existence of N-nitrosamines in our [ockI r emain: a stmng possibility, 'i he mode of formation in food and in vivo is less dear and -Ill / sINNllllHI'k URL 03952 will requite the icsc.ii (11 I'lioiis of s< it*i11isis vcisi il in many disciplines Only then (an the benefits to lie derived from niliiic addition lo lood be weighed against the possible ha/atd. Contaminants Resulting from Action of Biological Processes in Foods Olds timing I lie past ilrr adc* tv lien it was estal dished that pi < id lit Is dcli v cd It oin molds tan indue c < am er and may lie pi esernt as t oni.iminan ts in out loads have lhc*\ in fixed die comeilcil altenlion of investigators concerned with puhlie health. 'I he olten quoted turkey im idem which occurred in Knglamf in !%0 anti the simultaneous epizootic of liver ran ter in rainbow trout dramatically em|ihasi/cd out lack of knowledge about molils. their metabolites and particulatlv their public health sig nificance. One needs only to examine the literatme prior to and after the above episodes to apptcriatc the imiirnclons scientific effort dial was being din e ted toward solving this problem. l-orgac s and ('ai)l-" icvirwnl the problem <>| m\< olos i, nses anti em phasized the potential hazard of molds anil their piochccts. Outing this period, a mniimiing series of symposia wen* held, and reviews and books were published on the .subject ol molds, mold toxins, and their implica tions.1- fM`"n- ln an entile book devoted to the subjec t of allaloxin, a name un known It) years ago, was edited by (oldhlait.-n It would be pointless foi this writer to attempt to review that which has been done so altlv and elfeet ively in this hook. There ate a number of mold metabolites whic h have been demonstrat ed to be tarcinogenic to animals, 'ibis discussion will hiiclly examine sierigmatocystin and allatoxin M, as examples ol mycotoxins which mav be present as food contaminants and may have the potential to induce cancer. Steriginatoryslin, a metabolite of Aspergillus .etsiiolru1* and Asptigilhis nidttlnns*4 which hears a close structural relationship to aflatoxin 15,. lias been found in peanuts and cereals. It has been reported lo he less tarcinogenic lo rats than allatoxin 15,,r* and to have an l,l):>(, about six times higher than allatoxin/^ Tire oral l.!):,,, lor mite lor sicrigmatoc vstin5'* was determined and found to he ol a low order of toxicity and in excess of H00 mg/kg. With tests on ducklings, using bile ductal hyperplasia as the criterion, aflatoxin 15, was 125 nines more toxic: than sierigmatocystin.,w Purchase* and van dor \Vuit,i:* fed stcrigmatoc vstin in a standard diet at levels bom 10 to 150 ppm. Two male rats led 10 ppm lor six months ami 15 ppm for the next six months developed hepato cellular carcinomas within a year. They suggest that siet igm.uocystiii is about one-tenth as carcinogenic as aflatoxin 15, when given oiallv. However, in their opinion, sterigmatocystin may be just as important a carcinogen to human health ns aflatoxin since it is eiahoralcd bv several molds and in gteater abundance than allatoxin 15,. 1IIIUIMUM \MIM AN S AMIC \KC ilNDOKNISIs / 2<>f Sterigmatocystin I he sli uc tine ol st c i igmatex yslin is shown a Inn e and it can lie noted that it contains the 7,H ctihydrofuro (2,3-1)) luran structure which is also present in aflatoxin 15,. Our studies with rainbow trout, which compile aflatoxin 15, with aflatoxin K., indicate that unsaliualion at the 2,5 posi tion is important for carcinogenicity.4 Preliminary studies with rainbow trout, which I will report elsewhere; show sterigmatocystin to be rarcino genic even at levels of 20 ppb if fed with cylopropenoid fatty acids. These cyclic fatly acids which occur naturally in certain vegetable oils, im hiding cottonseed, have been demonstrated to increase lire incidence and growth of aflatoxin 15, induced liver cancel in rainbow trout. Allatoxin M, or 3-hyclroxy-G-itiethoxydil'urmmnn.none was repotted by Allcroft and Cat naglmn2 to be present in the milk of cows which consumed groundnut meal containing aflatoxin 15,. Subsequent studies showed the milk to be toxic when tested on ducklings. The chemical '.nurture was reputed by Hol/apl'cl el al.3!i This toxic factor, M,. was al so purified and isolated from sheep urine.*1- u4 Ilol/aplcl ct ;d.:`r* reported that the oral Ll)r,,, dose lor day-old Pekin ducklings for allatoxin 15, and M, was 12 and Kid ,ig, respectively. Smd ics by de longh et al.13 established that aflatoxin M, was also present in toxic groundnut meal and that the lariating rat converted allatoxin 15, to M,. butler and Clifford"' demonstrated thru aflatoxin f5, was partially convened lo the milk toxin M, in the rat liver. In studies with rainbow trout, we found that trout liver microsomes will convert allatoxin 15, to M,.4 " In a teview ol allatoxicosis in farm animals, Allcroft1 staled their was a litteat relationship belween the amount of allatoxin 15, ingested .Hid the concentration of allatoxin M, found in the milk. She concluded URL 03953 ifa 15 = ?5 ?SgS 5,8-^. '!lf$r r < r = <?.*-* s --t ^ ^ t -- Will l I a^a>i-j| -- i. fO " K _t I * isT | s ** = li.rM *r - f = K 5 ---= = 5- g 4 = * = " Eii -- = s -s =I II r. ^ zS T ^ c. ^ - 3? 2 -- ~ V> = S -I ?4 ^ 5 = s Z= s5Si- Ifif g* q: I ill' -- X J- 5 - =- 5* s y- = 2 G -' _. *:>-! e A ft. <0 -zo ai -- - --s y ?2 55 iT ^ E. ;r -. - r? = -=^:- = Z S 3 ~S : * 2 5 z 5 r = i ft-- -- s x -- ' iH=:^?- LXsft ; I" f = M.n -- *! == " " --= X s^ss-g-ms _ =. X 3?S ^?S's? T5 ir:s?1 5 !' ---- --- m!!m c - J -I EsL l--ss a--aj j-= S^ sc"-_Xi_n`_x W. 7 "= ft 1 o s* p- = -rt l e" ? ft -r. l x=^ 5 3 "= "S / 1 - :='2 = R1' = *5 -- -S's-ft-EE* i* <~pQS = ^.r a_ ._ c. rs ^ * = = = r = *=ft 2=r. I WU: 2E *C1 "x r> aea. s 30 55s = = S s 3 | N s ri * ___ ' it r r= E-- "z " "5 X*' X~ S 5S. x "Z X d =' ~ z2 - ft i~ ; E = L =. n, -- -- s=955` -70 / SI N N11 < i til-1( URL 03955 37. IImva t d. | VV . R I . I c-ague. |t , K II M luii- ,iiiiI It ( li', 11 I- \ii,n in .iml rslimjlii'ii ill |ml\i(.lu .iimii.ii ii > 11 < I......... I u >i i s m niiokeil lim.ls | lllcllind. J. Ass. < Hlli . \nal. ( hcmlsis I11 565 Ii I I, I'Mili ........I 38. IIuwjkI, J. W. F. VV. Tnriiihi, R II. While, and 1 Fa/in: Fsti.niinii and I'sliinalinii ill pnlpyihi aimii.iiir hyilior.iilinns in u*i. I|* nils. | Vs\ otlir. \.11 ( In miM.v I'l; |Ilf M 1966 39 Jung. I... ami lv. Mm.mil. I'i CSCIMC ill p)MIIC. ill Ii. ll/'ll | | \ I III ll ||, lull/.. 3.1 |> idle dans ilillili nles lillilcs \cgi l.di >. < U. Mail Si i. (I'aiis). 257 Kilts lldll, 1963. in. K < IJ >( .l K Ji- V: \n imlhir.ik <9 Inxir ||\<| IH||||. Ill i ilillili.nils. \(I||| \.l Mill. 16:365 322, 1961. M. Ko|>|uiij", N.: A sewn- pmgn-ssii-e livci disease m [in animals I. Simpi-nn-. anil mgyn ( halites. Ni.nl Vci Med., 18:265 209, I'Mili. 12. Rnppang, V. anil A. I Ii Igebnsiad: 'Inxir hepaiusis In lm nnm.ilv II I In- .in,I og' ihuidaled In li-rding, cxpeiiiuents. \.nd V el Mid. IK-2M2l5 I9trf> II Roppaug. V. and V. Melgclmslad: I nvn hepaiusis in lm animals III (m.dilimis .dli-iliiig die lm maiii >n <>l ill. Iu\i. l.n lm m In mih^ meal \md Vei. Mill 18:216-225, 1966. II. Knppang. N., I*. Mags* old, M. A. i l.niscn, I'.. Siigncn anil R, Svi nkeitid: feeding experiments with meal pimluecil fiom hciiing prescind mill sodium ninilc and fminalin. Nmd. Vcl. \lal., 16:313 Hli2. 1961. In. Kuralsune, M., and U'. (.. Ilucpci: Polsoihc aiom.iiii liydnicailmiis in roasted coffee, f. Nat. rancor lnsl . 24:463 469, I960. It*. laipn-ur, (. I... (>. Miikelseii, M. C. Whiling, and | . I . Kniland; Can inogi iiie ptopettirs nf mils limn llynn imma/n I.. indigenous to (main. J \;U Canter In si., 31:919-951, 1963. I/. Ivtc, D. J., J. N, Rodim. I. ( . Vn. and R <>. Siimlinlxi; Idln I ol w.`i laity j< ids on tile giosilli rale id lainhow nom, Sulino gmulmiti. J Null.. 92:9.1-96, 1967. 18. 14-c, D. J,. and R. (). Similmhei: t npulilishnl obsei r a lions. I 969. III. Iwe, II. J., J. II. Males. |. I,. A)lcs, and K. II Sinidmlni: Suuigixm lielween i ycloprupcuoid lalli at ids and ilieiniial laiiiuogens in i.iinhmv limn iilm grttn/nrri). Caiiui Ris., 28:2312 2318, MHirt. ;>tl. I.ijinsky, W., and S. S. Kpslcin: Niltosninines as ein.iimiiiicni.il lamiiogins Na ture, 225:21 23, 1970. 51. Lijinsky, VV., and A. I-.. Ross: Prndmiinii nl tan biogenic: |HiIynm le.n hydmiar bons in ihe cooking of food. Food Cosmci loxiiol.. 5.313 317, 1967. >2. Lijinsky, VV., and I'. Slmliik: lleii/oia)|i\iciie and niher bydrocailiims in rliaicoal broiled meat. Siience. 14.5:',H-55, ItKil. 5H. Lijinsky, VV., ami 1'. Miubtk: 1`nlynmleai lixdimailxm tarcinogons in moked meat and smoketl food, Ind. Med. -Niiir.. 81:152 151. 1905. 54. Lijinsky, VV.. and I*. Sliuhik: I lie detedion ol poluyc lit arnmaiir liydroiailions in liipiid smoke and some footls. I'oxiiol. Appl. ]`li.iunaro|,, 7:5,17 HIH, I9(i5. 55. Lillelmj, F.. B., and A. Cieglcr: Biological aelivin, ol stei i-finatocy slin. Myionaih Mycol. Appl., 35:878 370, 1LK3H. 56. Magee, 1'. N., Nitrites, nitrosamines. and (jiner. ...... . 1071 1072. I%H. 57. Magee. P. V: hi vivo u-.uiions of nitioso (mnpmunls. Ann. N V' Arad Sri 103 717-730, 1909, 58. Magee, P, N., and J. M. Banns: Ihe piodnciion o| niali^nani pinnuiy hepatic tumours in Ihe rat by feeding dimelhyhiitiosarnine Biii I ramci IO IM-I'^2 1956. >9, Magee, I'. N., and |. M. B.niiis: Inihnlion of kidm-v liunmiis in die ral with dimelh)Initrosamine. J Paili Barleriol.. HI; 19 31. IW.'. ' 00. Magee, IV N,, and J, M. Barnes: (iai<inogenic nilmso imiipmimh. In: Advanu s in ( ancer Researth (\. Iladtlow, ami S, VVeinlmnse. eds.), A<adeniic I'li ss N V Vol, 10, 1907, pp. 163 246. 61. Magee, I*. N . and R. Sdioental: Carcinogenesis hy nilmso ioninmiiids Biii Med Bull., 20.102-196. 196-1. l-IUMl CON I AMIN 5N IS AND t: \RCI Nt Mil.NI-.StS / 27 l lal.uioski. A J I- l. t.ieenli. Id. C. J. Haim-s, j. M. VVoilhingioii, and I 1 |i,,. ; suite) ol |*ol\coli< aimnalic hsdimarhmis in smoked hnnls. | Ass. (>lln . Anal, t heniisls. 51:1 II 121. 1968. 61. M.mjiiaidi, P.. and I.. Iledh-r: ('In-i il.is Vmkmnnun um Niliovnninen m Wei /eiiinehl. Aoneimiilellmsi linng lt>.77H 77fk 1906. 61 M.isii. M. S . R I-. I.imdin, |. R Page, and V <;. t.aii i.i: (tyslalline allaiosin M, 11 mu in ine and milk N.il n 11 , 21 : 751 < > >. I'M>,. 6. Masnda. Y. k Mmi, and M. Kmalsiun: Polo win aimnaiii h) dim .o bons in (oinnimi |;,pamse londs. !. Ilimled lish. roasi.-d l.ailev. shorn ail'd laiamel .......... ',7:133 I 12. 1901. (. Mali It's. U l.. and t. N. Mogan. l-.ds . IlnuhemisU) ol Some loodl.oiue Mum hial I o\ins. the M.l I. 1'iess, ( amln idge, Massat Imselis. IIM>7. 07. Peine, VV. F. II.: I iniioin pimimimn In lime oil in die mouse lmeslom.K Ii Na line, IH9;497 196. 1901. 06. Phillips. VV. K. J.; Niliale u>iikiiI ol londs- -public. Iicallli impli( aiions. (..mad. Inst. Fond lechnol. )., 1:98-103. I9i>6. 09. Pnuhase. I. F. II.. and (. J. sail del Wan: < jm iimgenu ily ol sici igmaloc) sim Food Cosmel. 'I'oxiiol., 6:555 550, 1968. 70. Sakshaug, \., t. Sognen. M. A. llanscii, and N. Koppang: Dimelhy liiiimsainim-; iis hepainioxic elfeil in sheep and its occunence in toxic batches of liening meal. Nalmc, 206:1261-1262, 1905. 71. Sander, J., F. Scltweinshetg. and II. P. Men/: Ihiurisuchungen iiIk-i die F.iilslelnuig cancerogcnei Nitrosamine im Magen. Hoppe-Seylei's /.ciisihr. Physiol, ('hem., 119: 1691 -1697, IIMiK, ,2. Sianl.in, R. A.: Peisonal loiiiiiiiiiiicalion, I9(i9. 73. Schmiilil, 11., and II. Oxswald: (am inogt nesis in dilleiein ;iiiiui:il spei n-s In <1. etliylniliosiiinim-. I spelKilli.i, 23: 197 500, l!M>7. 71. Sen, N P , II. C. Smith, and L. Schwinghamer: Voimdimn ol N nitiosamims bom sixondai y amines ami nitrile in human and animal gasliic jnite l oud Cosmel. I'oxiiol.. 7:301-307. 1909. 75 Sen, N. P,, II. C .Smith, I.. Sclnvingh.imei, and |. |. Maileau: Dieihylnili osainine and olliet N nilussamines in limits. |. Ass. (lllu. Anal, (.hem,, 52.17 52, l`Hi9. 70 Sliabad. I.. M.: Snulics in the P.S.S.R. on ihe distiilmtion, iinutatioii, and fanof carcinogenic hydiocailions in die human ciiviionment and die nde of ihcii deposition in tissues in can inngenesis: A lcvicw. Cancer Res., 27:1 132-1187, 1907. 77. Sigmjnnssoii, J : Oitnpational vatialions in tnnitalily fmm gastric i.mrei in rrla lion to dietary dillcicnies. litit. J. Caiuer. 21:651-656, 1967. 76. Sinnhuber, R. ().. I), j. I.ec, j. II. Wales, anil J. |.. Ayies: Dietary faitois and hepatoma in lainlmw limn (Sithno <iirthteri). II. CiKaicinogenesis h) tyilopiopeuoid Tally acids, and the HTett of gossypol and nlieied lipids on allamxin in dined liter cancer. J. Nat. Cancer Inst., 41:1293-1301, 1968. 79 Smnliubei. R. O. D. J. I.ce. | It. Wales. M. K. I.andeis. and A. ( . Keyl: Alla toxin M,. a polenl liui laiiinogen lm rainbow iroul. Fed. Pick., 29:568, 1970. Hi Sioiilmbet. R. <., f. II. Wales, J. L. Ayres, R. II. Eogebiei lit. and II. I.. Amend: Dietaiy failors and lupaimna in rainliow trout (Sulmo gniiduei'l. I. Vflaioxins in ii'gelalile pioleio leedsinlTs. J, Nat. Caiuei Inst.. 41:711-716. MHi6. Soudan. F.: 'Ihe nainral formol conleiil in hslieiies pimlinis. In: lisli in Nuni non (F. Heen and R, kieii/er, eds.) , Fishing News (Books) l.ld.. London. Fnglaod, 1962, pp. 76 79. 82. -Spinelli, j., M. Fkluiid, and 1). Miyaurhi: Measurement of hypoxantfime in hsh as a method of assessing fieshness. |. Fimd Set., 29:710-714. 1961. 83. Si out. F. M |. Adair, and | K Oldtielil: Hcpatotnxicosis in mink associaled nidi feeding toxic hening meal. Amer. fur Breeder, 41:12 11, 196R. 81. Sugai, M., L. A. Willing, H. 'Lsnchiyama. and F. A. Kuinmerow: The effect of heated hit on flu- raicinogenir ariiviiy of 2 arclylaminiifhioieiie. Cancer Res. 22: 510 519. 1962. S5. larr. II I. A: Bai u iios|.ui< aition i>{ nitiitcs. Naime. 1(7:117, I'HO rsoO ic z 7 S S r Z s = 7 z 5 z TZ As ~ e c 3D G> OJ (O </1 CD Z c tr O u* < > 2: o e o sc*a O o 73 > TJ ffi < G URL 03957 I (H'\ It 1<. 11 I l*i,l ll1* W. Ut IH m ) k Ml lilt \l I I 111 IMII'RS, IM.. .Ml riglu-. iumiu.iI. ,\<i j.;m <i| Miis | I>111.1 i > > mo lie rcpioiliitcd. Mmid 111 .1 ietiici.il >)M<iii. (>i 11.111 si 11 i I > ic I. iii .in) limit or by any lur.ms. rliittunir. 1011l1.1ni1.il, | .lmlo<n|o ing. |<< .imIjh,;. <>i mltri wise, without |)iim wiiiien jut mission hum Hie jnihlishei. I'linlcil in Mir United Sin(i*s of .Amcrici, Library of Congress C!;*r;log" Card Number: 7(Mril)2f>- [liliTli.ilinnil Manil.ml Itniit \ 11 iiiIjci : |7H7 | Acknowledgmen ts l lic following numBcrs ol (lie- Dcpailmnit of Publications of Tlu: University of Texas M. 1). Anderson Hospital and Tumor Institute at Houston partici pated in the editing and preparation of this volume for publication: Associate Editor DoKOHIY M. Bi-ANE, B.A. Assistant Editors Susan Hikki 1.. B.A. Lynda G. Burgner Janina M. Lev. A.B. Simri.i y J. Hartman, B.B A. Pamfla IIester Barnara E. Johnson. B.A. Deborah L. Kensh.. B.A. Judith Wibi.e Lktieney, B.A. M. Lucinda Marinis, B.A. Deborah L. Ryi.andi-r, M.A. Diane Siioquist, B.A. Katiii.f.kn S. Yacu/yo, B.S. Acknowledgment is made of the kind services of Dr. Clifton D. Howe, Asso ciate Director for Clinics and Chief of Clinics, Dr. Darrell N. Ward, Head. Department of Biochemistry, and Dr. Felix L. Haas, Head. Department of Biology, The University of Texas i\T. D. Anduson Hospital and Tumor Insti tute' at Houston for their assistance in arranging die contents of the volumes. 1 lie help of Miss Donna MtComiick and Mrs. Marilyn Cavanagh of die Congress' Secretariat is gratefully acknowledged. LINDA HALL LIBRARY VO..- \a ITPY MO Briefing Carcinogens in Scotch The Scottish whiskey exporters de nounced it as an American conspiracy to ruin their trade but, in fact, there was nothing unexpected in the report issued by the Nationaf Science Foun dation on 7 August saying that a car cinogen known as /V-nitrosodimethylamine (NDMA) has been discovered in six out of seven brands of Scotch tested this year. NDMA was found in 16 brands of beer as well, but not in any of the wines, sherries, liqueurs, brandies, gins, vodkas, or rums that came under review. The research, done by David Fine and E. Ulku Goff of the New England Institute for Life Sciences in Waltham, occurs when nitrosopyrrolidine con centrations reach to pans per billion (ppb) att^r. frying. A decade ago, be fore meal producers were asked to take remedial action, it was not un common lo find concentrations as high as 100 ppb. Hie compound in bacon, researchers say, is considered less carcinogenic than NDMA, the one found in beer and Scotch. NDMA has been shown to cause cancer in nearly every laboratory animal tested, and in one experiment, adding 10 ppb of NDMA to the water of tumor-prone mice tripled the incidence of lung tu mors. The important work on beer was done by B. Spiegeihalder at the Ger man national institute for chemical and cancer research in Heidelberg. In the summer of 1976 Spiegeihalder Cartoon by Ohpnant. Copyngnt t 1979 Tft* Washington Sui Massachusetts, did little more than fill a gap in earlier work, most of it done by German researchers, showing that many varieties of beer contain NDMA. Beer and Scotch producers use a sim ilar process of barley malting, and the research on Scotch merely strength ens what was suspected already-- that the problem is in the malt. NDMA is one of a family of carcino gens called nitrosamines, all of which occur widely in the environment. The best known until recently was A/-nitrosopyrrolidme. a suspected carcinogen m fried bacon. The Department of Ag riculture has put into effect a food monitoring program to keep track of nitrosamines. and the "action level" at wh'Ch bacon is taken off the market and his colleagues published a report that 70 percent of the 158 European beers they had analyzed contained NDMA in amounts ranging from around 1 to 68 ppb. In general, dark beer contained more than light beer, and the highest concentrations were found in "rauchbier." made from smoked malt. This was deemed a sig nificant health risk for Germans, whose diet is one-quarter beer. Spiegelhalder calculated that, based on average figures, a young German beer drinker might consume as much as 1.2 micrograms of NDMA a day. Other nitrosamines were found, but none so potent or abundant as this one. Several possible sources of con 'V itfN-u'r.uMJO 5o ll Copyright s |V9 AAAS cJ URL 03S: tamination are under study; the Ger mans have decided that the point at which NDMA is introduced is in the process of drving or "kilning" the barfey malt. The use of hot ar drawn di rectly from a fire is thought to create NDMA in the malt, and indirect heat ing is thought to inhibit it. Malt for Scotch rs prepared in the same way, suggesting that the one Scotch with out NDMA in the seven tested--White Label--may use an indirectly heated mall. (The others were Chivas Regal, Black and White, J & B, Ballantine's Sandy Scot, and Cutty Sark.) It is not known which compound in the barley reacts to produce NDMA, although people have suggested it may be a fungicide, pesticide, or anticorrosive agent. The American work has confirmed the German research and given more weight to the theory that the malting process is the key problem. It also suggests the health risk is lower for Scotch drinkers than for beer drinkers. The concentrations of NDMA are lower m Scotch to begin with, and the volume consumed is lower. The con centrations of NDMA found in the 18 foreign and domestic beers tested at the New England research tab ranged from 0.4 to 7 ppb. The Distilled Spirits Council, a trade t n association >n the United States, co stresses that nitrosamines have not beer proved a tnreat to human health, and a spokesman. Sam Chilcote. claimed recently that it is "practically impossible" to eliminate these com pounds from the human diet. The dis tillers have not decided whether to take any action. However, shortly af ter the news about nitrosamines in European beer came out last year, the U.S. Brewers Association set up a re search committee chaired by William Hardwick of Anheuser-Busch Inc. and asked him to come up with some sug gested solutions A spokesman, re fusing to give details, said the brewers have spent1 a lot of money" and com missioned research at three labs. Lacking definitive proof that this is a public health hazard, the companies are reluctant to order a complete over haul of the barley malting process for. as one company official said, that would cost hundreds of millions ot dol lars and bankrupt half the maltsters m the country. Worried annkers always have the science. 1.1)1. :o.c aiocsi h?v option of avoiding beer and Scotch. Sut they should be advised to stay clear of carrot juice and beet juice as well, for they also contain relatively high concentrations of nitrosammes. _______________ i Congress Relents, Spares OES Bureau After a harrowing summer, the State Departments Bureau of Oceans and International Environmental and Scientific Affairs (OES) is ex pected to glide safely home to port this fall when Congress returns from its August recess and takes up the ap propriations bill for the State Depart ment. The OES, which handles nego tiations for fishing agreements and scientific and technological ex changes, earlier this year ran afoul of Representative William Alexander (DArk.). He slashed both the budget and the staff of the OES in half when they came before a subcommittee on which he sits (Science, 8 June 1979). j The funds were restored after a , 1 '/2-hour debate on the floor of the i House on 12 July m which Alexander | reported thirdhand that Henry Kis singer had once told another official that OES "is where the Department of State places its incompetents." He continued: "I would rather have one good horse than a whole team of lame nags that sit grazing at the trough of public expense. . . When the harangue was over, sev eral congressmen chided Alexander for failing to hold any hearings on his charges. Representative Clement Zablocki (D-Wis.) spoke of the "unfor tunate record" and said the budget slash was adopted without any con sideration and chiefly at the urging of one individual, our good friend, the gentleman from Arkansas." The House then voted to restore the bud get: a conference report cleared the House on 2 August: and ail that is lacking is the Senate s approval. An OES official, Leslie Brown, said, A lot of us were surprised not only by tne depth but by the breadth of sup port" that came through for OES at the last moment. Rep. Alexander had boasted that he would win the battle if it ended up on the House floor. He plainly underestimated the opposition. ----------------------------------------------Eliot Marshall -- vt oiST ivn i( I'Tlirjui-J Irnni page 7fi7i would be doubly concerned about an asthmatic with a common cold." Greatest concern centers on pilots-- and understandably so. The picture of a flight crew incapacitated by ozone is not a very pretty one. Yet most pilots refuse to talk about any problems they may have encountered. This too is under standable. Mere mention of "chest pains" might be equated with "heart trouble." which could mean the loss of a pilot's medical certification to fly. The most vocal complainers are flight attendants. If they seem to have more problems with ozone than passengers and pilots. Tierney notes, it is probably because they work harder. A person at rest breathes about 5 liters of air per min ute. A person pushing a heavy can up and down an aisle breathes 15 to 20 liters per minute--and thus takes in many times more ozone. Though flight attendants raise a fuss, the companies they work for arc often noncommittal. Trans World Airlines, which operates 11 regular 747'a. says it is waiting to see if the FAA regulations go into effect. It will then add ozone-remov ing equipment. Pan American says that its 29 regular 747's will get catalytic conveners to break down ozone sometime in 1980. and that the ozone problem has already been solved on its very long-range, high-alti tude planes. Charcoal filters (each weighing 800 pounds) were installed on all ten of Pan American's 747SP's in March 1978. Since then, there have been only two passenger complaints. The 645 complaints by flight crews in 1978. says a Pan American spokesperson in New York, have to do with politics involving union wage settlements. "As far as we re concerned, there are many more passengers than crew." says James Are>. "Based on the complaint figures, we have done the job. There are no more problems." Flight attendants, however, say that a wage contract has long been reached-- and that complaints are still being sent to New York. "In fact." says Carmen Azzopurdi of the Independent Union of Flight Attendants. "Pan Am just settled out of conn with two flight attendants for ozone-related injuries. One had devel oped bronchitis, the other asthma." She also notes that passengers have never sent in many complaints, for few under stand the problem. A spokesperson for Boeing says that Pan American is the only airline putting in filters and caialyuc converters. The other airlines can get (hem as optional equipment, he says, but so far none have. He also says that the FAA regula tions will probably never come out. since the problem is so rare. "I've traveled around the world in a 747 and never had a single problem." says Leonard Weiss. "This ozone thing has been vastly over played by the news media." To give credence to their side of the story in the face of industry skepticism, flight attendants presented a 1330-person survey at the House hearing in July. Per formed by Dway ne Reed, an epidemiolo gist with the California State Department of Health Services, the survey looked at flight attendants from Pan American, which flies high-altitude international flights: from Trans World Airlines, which flies both international and loweraltitude flights: and from Pacific South west Airlines (PSA), which flies only lowaltitude flights within California. The survey tabulated symptoms for five con secutive flight days. With fatigue, back ache. nausea, and vomiting, there was little difference between airlines. Chest pain, however, was experienced by 19 percent of the Pan American flight attendants. 13 percent of those on TWA, and 5 percent of those on PSA. What worries many is lack of data about long-term effects. Ozone, for in stance. is a mutagen and is known to cause biochemical changes in the blood stream of exposed persons. But it is not known if high rates of miscarriages and children with birth defects among flight attendants are a result of ozone exposure. Even if the FAA does make a ruling on ozone. ALPA fears that the airlines will not take the time and money to put on equipment that really works. As part of the FAA regulation, therefore. ALPA wants a requirement for on-board sys tems to monitor ozone levels --thus checking the effectiveness of ozone re moval equipment. A few months ago the FAA said the ozone ruling, if it is made, would come in September. Now it has pushed that back to October. Asked how long the air lines would have to comply with a ruling. Ray Ramakis of the FAA's safety regula tions division said: "The notice of pro posed rule-making said 6 months. It could go to a year. I don't know. There is no set formula." If the airlines have their way. it w ill take a w hile. In January the Air Transport Association, which repre sents the nation's airline companies, told the FAA how long it thought the con version to ozone-removal equipment would take. The estimate came to some where between 4 and 6 years. J-- William Bkovd 7AV URL 03959 mc of the oe e ASSESSMENT OF TECHNOLOGIES FOR DETERMINING CANCER RISKS FROM THE ENVIRONMENT ta JUNE 1981 C 13 O oJ iO cn C3 OTA Reports are the principal documentation of formal assessment projects. These projects are approved in advance by the Technology Assessment Board. At the con clusion of a project, the Board has the opportunity to review the report but its re lease does not necessarily imply endorsement of the results by the Board or its indi vidual members. CONGRESS OF THe UNITED STATES Wee ot Technology Assessment WtiflingtO'V 0 C. 20910 the number of retail outlets. (Recently, sev eral States have raised drinking ages.) 10 percent before 1966 to 19 percent be tween 1966 and 1975. Around 1960, total per capita sales of ab The importance of alcohol as a health hazard I solute ethanol began to rise significantly is not limited to its association with a relatively 1 registering a 30-percent gain between 1961 small percentage of cancer. Estimates of the an and 1971. Since 1971, there has been vir nual number of deaths related to alcohol range tually no change in per capita sales. The ef from 37,000 to 205,000. Cirrhosis of the liver, fect of the overall increase may not yet the seventh leading cause of death in the United have fully manifested itself in cancer rates. States, was responsible for 30,066 deaths in There is particular concern over increased 1978. As an contributory cause of homicides, alcohol consumption in youths. This is suicides, and accidents, alcohol's toll is even heightened by the observation that early greater. The Institute of Medicine (178) recently drinking behavior predicts drinking habits completed a study of alcoholism and related in later life. problems and indicated several opportunities o> Pc ln< pr Al The proportion of high school students who reported ever having been drunk in for research. These included further research on alcohol metabolism, development of appro creased dramatically from 19 percent priate animal models, and further efforts before 1966 to 45 percent between 1966 and through epidemiologic studies to explore the 1975. The proportion reporting being in toxicated at least once a month rose from link between alcohol and its adverse health con sequences. :p A <jj> ir tc DIET Introduction Studying the relationship of diet and health is a continued source of frustration and excite ment. Food affects all body functions and comes into direct contact with the digestive system and indirect contact with all other organs. Cancer rates for digestive sites vary considerably around the world and have prompted studies of diet's role in cancer causation. As discussed here, diet encompasses those items ingested as food, including substances added to food, those produced during normal cooking, storage, and digestion, but excluding drinking water and alcohol (discussed in this chapter under Air and Water Pollution and Alcohol, respectively). The amounts and balance of the major com ponents of diet are generally believed to be responsible for the lion's share of diet-related cancers. Deficiencies or excesses of micro elements, and the presence of additives or con taminants, are probably less important. The various means by which diet may influence the development of cancer are listed in table 16. Diet also plays a role in the treatment of cancer illustrating the pervasive role of diet with respect to all aspects of the disease. Unfortunately, dietary studies are plagued with methodological problems and conflicting evidence exists for almost every specific ques tion that has been investigated. The overall association of cancer with diet exists but there is no reliable indication of exactly what dietary changes would be of major importance in reduc ing cancer incidence and mortality. The strongest positive associations identified through correlations of dietary patterns and cancer rates are those between total fat intake, particularly animal fat consumption and can cers of the breast and endometrium; and be tween total protein intake and cancer of the co lon. The most dramatic change observed in a diet-related cancer site has been the reduction in incidence and mortality from stomach cancer. In the United States, the 19S0 age-adjusted mortality rate from stomach cancer was 24.4/ 100,000 for males and 13.1/100,000 for females b V a c t I a percent be- jalth hazard a relatively ?s of the ancohol range 5f the liver, t the United > deaths in homicides, oil is even 78) recently ind related >portunities research on of approter efforts xplore the health con- table 16. t of cancer diet with e plagued :onflicting rific ques\e overall ut there is at dietary in reduc- identified :ems and at intake, and can- and beof the co~ved in a luction in i cancer, -adjusted /as 24.4/ r females Ch. 3--Factors Associated With Cancer 77 Table 16.--Some Currently Attractive Hypothetical or Actual Ways in Which Diet May Alleet the incidence of Cancer Possible ways or means Ingestion of powerful, direct-acting carcinogens or their precursors Affecting the formation of carcinogens in the body Example* Carcinogens in natural foodstuffs (plant products) Carcinogens produced in cooking Carcinogens produced in stored food by microorganisms (bacterial and fungal) Providing substrates for the formation of carcinogens in the body (e.g., nitrites, nitrates, secondary amines) Altering intake or excretion of cholesterol and bile acids (and hence the production of carcinogenic metabolites in the bowel) Altering the bacterial flora of the bowel (and hence the capaci ty to form carcinogenic metabolites) Affecting transport, activation, or deactivation of carcinogens Altering concentration in, or duration of contract with, feces Altering transport of carcinogens to stem cells Induction or inhibition of enzymes (which affect carcinogen metabolism or catabolism) Deactivation, or prevention of formation, of short-lived in tracellular species (e.g., by use of selenium, vitamin E, or otherwise trapping free radicals; by use of b-carotene or otherwise quenching singlet oxygen; by use of other antiox idants) Affecting "promotion" of cells (that are already initiated^ Overnutrition Vitamin A deficiency (clinical orsubciinical) Retinol (Binding Protein] (hormonal and other factors deter mine blood RBP, though vitamin A intake may not affect it much) Otherwise affecting stemcell differentiation (carotenoids' determinants of lipid "profile"?) Age at menarche Adipose-tissue-derived estrogens Other effects *Thr my M considerable ovariao between many of the enlriea In lhi table ^Of, more generally. aMeetlno the proMtitity tht a partially transformed it*m cell will become fully trinaformed and will proliferate suecetafuily into cancer. SOURCE. Doll and Palo (93). while in 1977 the rate was 8.8/100,000 for males and 4.0/100,000 for females. This decrease has occurred in many other countries, including those with high initial rates, such as Japan and Iceland, and those with lower initial rates, such as Canada and New Zealand. The factors be lieved to have contributed to these decreases in clude: reduction in use of salt and pickling, lower consumption of smoked foods, and in creased use of refrigeration, increased con sumption of milk, green vegetables, fruit, and antioxidants (237). In general, as a result of increased intake of calories, proteins, and certain other nutrients, Americans have been growing taller and reach ing sexual maturation earlier. A great improve ment in the Nation's health has resulted from this change, but increased risk for certain cancer sites may accompany the improvement. For ex ample, earlier sexual maturation in women is associated with higher risk of breast cancer later in life, though as yet, no increases in breast cancer has been attributed directly to improved nutrition (see Sexual Development, Reproducfive Patterns, and Sexual Practices for a more complete discussion of this topic). Many estimates of the importance of diet to cancer have been put forth. Doll and Peto (93) estimated that altering dietary practices may reduce cancer by as much as 35 percent (stom ach and large bowel cancer by 90 percent; en dometrium, gallbladder, pancreas, and breast by 50 percent; lung, larynx, bladder, cervix, mouth, pharynx, and esophagus by 20 percent; others by 10 percent). The great uncertainty of this estimate is indicated by the range of 10 to 70 percent which they attach to their estimate. Wynder and Gori (368) estimate, by calculating URL 03962 -- . t^l 78 Technologies for Determining CancerRiskt From the Environment the percent difference between U.S. mortality rates and the lowest rates reported worldwide and by considering specific case-control studies, that an even larger proportion of cancer, ap proximately 40 percent in males and 60 percent in females, could be attributable to diet. In testimony before the Subcommittee on Nutrition of the Senate Committee on Agri culture, Nutrition, and Forestry, Dr. Arthur Upton (353), then-Director of the National Can cer Institute (NCI), succinctly discussed the ex tent of involvement of diet and cancer when he stated: Despite the impression that cancers are linked with dietary patterns and the inability to pinpoint specific dietary carcinogens, scientists generally agree that factors in diet and nutrition--including drinking water contaminants--appear to be related to a large number of human cancers, perhaps approaching 50 percent. Dietary Intake Fat/Meat Intake Examination of cancer rates in different coun tries show positive correlations between colon cancer and consumption of meat and animal protein and between cancer of the breast and en dometrium with total fat consumption (17). These cancers are common in the United States, Canada, and Western Europe and rarer throughout the developing world. The most widely held theory is that fat in the diet has a promotional effect on the devel opment of cancer. One suggested mechanism is that fats affect hormone levels. Several cancers, including breast, ovarian, and endometrial are, in turn, influenced by hormones. This associa tion of dietary fat with higher cancer rates is, however, not found uniformly. Breast and col orectal cancers are not uncommon among vege tarians, and the observed incidence in Seventh Day Adventists who are largely vegetarian, is matched by the same incidence in Mormons who eat meat (293). This finding and the fact that meat intake among Mormons is not mark edly lower than among the general population is often cited as a rebuttal to the meat/fat cancer causation hypothesis. These interpretations should be viewed cautiously for the studies may not have been adequate to reflect any promo tional effect in these low risk groups who are less exposed to many other types of carcino genic stimuli, such as, cigarettes and alcohol, than is the general population. Additional studies in this area are needed. More and more, studies have focused on specific types of fat. Evidence suggests that diets with a high ratio of polyunsaturated fats (main ly of vegetable origin) to saturated fats (mainly of animal origin) may increase the risk of can cer. Paradoxically, this type of diet is recom mended to lower the risk of heart disease. Re sults from epidemiological studies (98,290) and animal trials (53,305) have been inconsistent and confusing, and have shown different effects on different cancer sites. In rats, polyun saturated fats need be only a small proportion of a total high fat intake to promote breast tumor incidence. If the animal model is ap plicable to humans, virtually all high-fat diets will exert a promotional effect (52). Serum cholesterol has been investigated in many studies as a risk factor - for various cancers. Serum levels are directly affected by in take of cholesterol and fat. A diet with a low saturated fat to polyunsaturated fat ratio decreases the amount of ingested cholesterol that will appear in the serum. The discovery that the stools of colorectal cancer patients ob tain an abnormally high proportion of add steroids, derived from bile salts, and A* lesterol, supports the hypothesis that cert^ types of fat play a role in the production of c* orectal cancer. A recent epidemiologic study found an iaa riation between high-density lipoproteins cancer risk (199). On the other hand, in a spective epidemiologic study of heart disease* Framingham, Mass., serum cholesterol were inversely associated with the incidence* colon cancer and other sites in men. Men ^ the lowest serum cholesterol a colon cancer rate which was than men with the higher c-----(361). A similar negative association ported in data from the Paris Prospective of Coronary Heart Disease (45). ac Fit 'aiAiitt 'mr'rr* Ch. 3--Factors Associated With Career * 79 mam* lainly f can- 15 diets The hypothesis that dietary cholesterol plays a direct role in the production of colon cancer is supported by some animal studies. For instance, the addition of cholesterol to the diet of rats'On a cholesterol-free liquid diet, had a promotional effect and cholesterol enhanced the carcinogenic effect of a known carcinogen (1,2 dimethylhydrazine). Fiber Intake Burkitt (40) observed that several intestinal diseases that are common in developed coun tries are rare in rural Africa where unprocessed food is consumed, and the stools tend to be soft, bulky, and frequent, it is suggested that dietary fiber may reduce colorectal cancer by decreas ing the time stools remain in the bowel, by in creasing bulk (thereby decreasing the concentra tion of carcinogens in stool), or by perhaps altering the distribution of bacteria, some of which may produce or destroy carcinogenic me tabolites. The effects of fiber on cancer incidence have attracted much interest but remain problemati cal. As in the case of fats, dietary fiber is a term Much covers a multitude of different sub stances, each of which may have different influ ences on carcinogenesis. The methods by which types of fiber are chemically characterized are till in a primitive state, and analyses of the composition of dietary fiber in many foods is beking. Despite the drawbacks, some associa tions have emerged. A close inverse correlation hos been found between the pentose fiber con tent of the diet and mortality from colon cancer part of the United Kingdom (93). No correla tion was shown with any other fiber types, or teth dietary fiber as a whole. and Vitamins Several metals and vitamins are linked with development. They can either act as car themselves or they may biologically -"pete with other dietary constituents to ***** or suppress a cancerous response. ^*^<J-drficient syndrome was shown to be with a high risk of developing can to* pharyngeal and esophageal mucosa in northern Scandinavia. The incidence of gastric cancer is found to be 4 to 5 times higher in countries where iron deficiency is prevalent than in the United States (356). Selenium, an effective antioxidant, is often found at higher levels in plants, milk, or human blood in sections of the United States with low cancer rates. Selenium deficiency in rats con sistently increases the carcinogenic effect of known chemical carcinogens, particularly in animals fed on high polyunsaturated fat diets (190). Supplementation with selenium above dietary requirements decreases tumor yield in animals on both low- and high-polyunsat-rated fat diets (188). Stocks and Davies (24) found that a high zinccopper ratio in soils was associated with elevated rates of human stomach cancer while Strain et al. (24) reported an association of elevated cancer rates with low zinc-copper ratios. Marginal zinc deficiency is associated with increased esophageal cancer in animals (122) and esophageal cancer patients had lower levels of zinc in their blood, hair, and tumor tissue than controls in a study of Chinese men (211). Zinc deficiency may interact synergistically with alcohol to enhance esophageal car cinogenesis (135). The overall relationship between cancer and vitamins is not well understood. Vitamin C has been shown to reduce carcinogen formation in experimental animals (359), and this activity may be important in reducing cancer occur rence. Vitamin A (retinol) has been more ex haustively studied than any other vitamin (for review see 292) and it has been suggested that vitamin A, or, more particularly, its vegetable precursor, beta-carotene, may decrease the sus ceptibility of a variety of epithelial tissues to the development of cancer. Retinol (or its analogs retinoic add and various retinoids) has been repeatedly demonstrated to diminish the risk of experimentally induced cancer in laboratory animals (93). These results are particularly in triguing because the protective effect is observed even when these substances are fed long after the animal is treated with an initiating cardnogen, and the vitamin and its analogs appear to be effective at a wide variety of sites. _* y <0 Technologies for Determining Cancer Risks From the Environment People with a history of consuming above average amounts of provitamin A (beta-caro tene) have a slightly lower incidence of several different types of cancer than people who give a history of consuming less. Beta-carotene is found in carrots, green leafy vegetables and in red palm oil which is used for cooking in many tropical countries. Further epidemiologic studies are now in progress in areas where red palm oil is habitually used. The role of beta-carotene in cancer prevention is still uncertain and this par ticular hypothesis has been mentioned to illus trate the potential importance of diet. Local deficiency of folate (a B-compJex vita min) has been demonstrated in the abnormal cervical tissue of some women taking oral con traceptives. The observed abnormalities are often precursors of cervical cancer. Supplemen tation with folic acid in these women can re verse the abnormality and appears to prevent progression to carcinoma (41). Deficiencies of lipotropes (choline, methionine, and folic acid) increased the susceptibility of animals to a variety of environment carcinogens in several studies (311). Dietary Balance Studies in laboratory animals have shown that altering gross aspects of diet can have sub stantial impact on the risk of an animal de veloping cancer. Jose (195) reviewed several re ports which demonstrate that restricting calorie intake without modifying the proportion of the individual constituents reduces the incidence of spontaneous tumors and of a variety of cancers produced by exposure to known carcinogens. Not only did calorie restriction result in de creased incidence of tumors, but it also delayed the time of appearance of tumors, and when tumors appeared, they grew and metastasized more slowly. The life spans of animals on restricted diets were often increased up to 50 percent compared to normally fed controls. However, decreased calorie intake, when ac companied by inadequate protein intake makes laboratory animals more susceptible to many environmental carcinogens (47). A suspected association between obesity and the risk of cancer was strengthened in an epi demiologic study conducted by ACS which fol lowed 750,000 people from 1959 to 1972 (207). Overall cancer mortality was found elevated for those individuals who were more than 40 per cent heavier than average. Mortality from can cers of the colon and rectum were increased among men while mortality from cancers of the gallbladder and biliary passages, breast, cervix, endometrium, and ovary were increased among women. The meaning to be attached to the re sults is not entirely dear since weight is asso ciated with a variety of social and behavioral characteristics that affect the risk of cancer in other ways, including smoking habits and socio economic status. Another concern is that diag nosing cancers may be more difficult in obese individuals, and cancers may generally be more advanced when they are detected. In addition to the quantity and composition of food, the timing of intake has also been shown to be important. For example, Roe and Tucker (cited in 93) randomized mice with a high spontaneous incidence of mammary tu mors, between continuous feeding, in which the mice were fed 6 g of food each day, which they consumed in frequent small amounts, and inter mittent feeding, in which food was limited to 5 g per day which was eaten at once. No clear dif ference in longevity was observed, but nonfatal spontaneous mammary tumors arose in 64 per cent of the continuously fed mice and in only 8 percent of those fed intermittently. Immune function in both humans and ani mals can be severely compromised by deficien cies of certain dietary nutrients--in particular, protein, methionine, choline, folate, vitamin Bu, vitamin A, zinc, and pyridoxine (147). Defi ciencies in certain of these nutrients, with a con comitant depression of immune system function have been demonstrated in animals and are sus pected of increasing the susceptibility of de veloping certain cancers (356). Naturally Occurring Carcinogens and Precursors Along with the major components, thousands of chemical substances occur in small quantities in foods. Most naturally occurring carcinogens and mutagem tant class, nr plants and an Nitrates and ' Nitrate an. vegetables, fi their presen82), and ar residues in f chemicals ir samines ani are among cinogens in mors at a va lungs, esop nasal cavit: (221). Hxp* shown to ir (359). Epidemic nitrates to ; itionshij C e conter r- s notab S main s g itect ag<- ssiblybi i preser n of die i produt Other Sut Cycasir pyrrolizic tracts of naturally carcinoge bracken cancer in in Japan, three tim esophagi (93). Ith cancer ra tion of c. of the c. food (24 A Aft. which fol1972 (207). levated for an 40 perfroin can- increased icers of the ist, cervix, sed among 1 to the reht is assobehavioral cancer in and sociothat diagt in obese [y be more imposition also been , Roe and ce with a imary tuwhich the vhich they and interdted to 5 g , clear dif* t nonfatal in 64 per- in only 8 and aw* r deficien* (articular, vitamin 47). Deh* ithaconi function dares*v of de- Or. 3--Factors Associated With Cancer 81 and mutagens are of plant origin, but an impor tant class, nitrates and nitrites, occurs in both plants and animals. Nitrates and Nitrites Nitrate and nitrite salts naturally occur in vegetables, fish, and meat, are added to food for their preservative properties (see Additives p. 82), and are present in pesticide and drug residues in food (127). They react with other chemicals in the body to produce N-nitrosamines and N-nitrosamides. N-nitrosamines are among the most powerful chemical car cinogens in laboratory animals, producing tu mors at a variety of sites including liver, kidney, lungs, esophagus, bladder, pancreas, trachea, nasal cavities, and peripheral nervous system (221). Experimentally, vitamin C has been shown to inhibit the formation of nitrosamines (359). Epidemiological evidence linking nitrites and nitrates to cancer is fragmentary. Evidence Cpr a relationship between gastric cancer and the ni trite content of the diet is not wholly consistent. It is notable that vegetables, which are usually the main source of dietary nitrates, appear to protect against the development of the disease, possibly because of their vitamin C content (93). On present knowledge, the possible contribu tion of dietary nitrates and secondary amines to the production of cancer is uncertain. Other Substances Cycasin in the cycad nut, safrole in sassafras, pyrrolizidine alkaloids in some plants, and ex tracts of coltsfoot and bracken fern are a few naturally occurring compounds which exhibit carcinogenic properties in animals. Only bracken fern has been demonstrated to cause car'cer in man. Bracken fern is commonly eaten Japan, and Japanese who eat it daily have ree times the risk of developing cancer of the **phagus as Japanese who do not eat it at all has been postulated that the high bowel c*ftcer rates in Scotland may be due to the inges- of cattle fed on bracken fern or the leaching the carcinogenic components into water or (24). Range-fed cattle and sheep may pass pyrrolizidine alkaloids along to humans in their meat and milk (215). Certain plant-derived preparations are asso ciated with cancer. Extracts of some plants used to make herbal infusions, for use as home rem edies, tonics or beverages, are carcinogenic in animals, and their tannin-containing fractions are particularly active. Some population groups who use these products have high rates of esophageal cancer, suggesting an association (297). Perhaps of greater concern, because of its widespread use, coffee has been associated with human bladder (24) and pancreatic cancer (220), but whether the associations are causal is not yet known. Studies in animal cell cultures have shown caffeine, a constituent of both coffee and tea, to potentiate the effect of carcinogenic substances (96). Mutagenic substances have also been iden tified in cruciferous plants (cabbage, broccoli, etc.), from cereal grains, and some grazing range plants in the Southwestern United States (MacGregor, 1980). Other not-yet-identified carcinogens and mutagens may occur naturally in food, but on present evidence, naturally oc curring carcinogens are not regarded as an im portant cause of cancer in the United States (93). Carcinogens and Precursors Produced by Cooking Another possible source of carcinogens is their production in cooking. Humans are the on ly animals that cook their food, and it has been known for many years that carcinogenic chem icals such as benzo(a)pyrene (B(a)P) and other polycyclic hydrocarbons are produced when meat or fish is broiled or smoked or when food is fried in fat which has been used repeatedly. Sugimura et al. (336) demonstrated that broiling also produces powerful mutagens that cannot be accounted for by the production of B(a)P alone. Few people eat more broiled foods than Ameri cans and while the declining stomach cancer rate provides some assurance that cancer at that site is not related to broiled food, the possibility remains that colorectal cancer, which has not decreased materially, might be related. Recent URL 03966 '* iS 'in.s , Contaminants N*i. 'i: C ..^rurcr*? A less obvious source or can mogenir activity and one that vas overlooked altogether unHl tbf early 1960V is the ocoduntii-vp.of r^f. Si- - iii.i-/- uiliis t:. '.. iCd tOod. 'V'vl:OXif , J product cf the fungus Aspergillus rlu-us : rhe most powt fvi liver carcinogen known `or 'orrv animal sp ^es In addition, human liver cells cont-ii" \ .\>zymes necessity t > prcduu" the metj '< : i \iucts 'hat ippi-.\r tc ?; r'sponst ble l> i:: r . witv. fhe: - is ev;di .ice lor b;)iev u:g that afVuixin is a major factor :n .the pro- dm tio:i ot !i rei cancer in certain rropu al coun- -i ...i, 'V'j. "->vt -W* / ;* ' - ' ' vii ,,-. re , et*' - > States p-' uary ... - 'cr cf : he >t`.c is a rate disease accu-.u'iting f-.u less than 1 percpri: of rancer <ieaths [2,796 deaths in 1978). The amoi-m or afia`oxin in the American diet is nr-cll and only one among seveial other possi- 1,K causes of liver cancer. In the American con vex* , the chief importance of the discovery of the r !i... r --Kit : S'.'- - *- { J--. ,'l--C '>>* '. v-. c i-v.'---:,t-^tr.V %cer*i , th<- . . Jo that 0w._. n? Vbute to the high incidence of esoph- icer in parts of China by increasing the . .Mitenc or ccnrammared food \iUSM 5^v>iAxnn>:-ula* Contaminants hi : V* 1979, OTA < i:V'Orv .ippc.. ch ev - v .'`mpov--. 1. lorenvi;-. rn u.. .......___ rT;amf>irh re*' in ...latei food. A wide products ranging from ticide residues, to subof packaging, such as .an pollute food. Organic > ose the greatest potential od contamination based on > and toxicity of organics nited States. p,1-` -:A'- *<>. hydrocarbons (PAH) are \iund wi.J*-!y in .cs of foods, sometimes at the ame lev A an be present in charcoal- Vriile-i. u--'ah and iked ham, i.e., up to 15 ug W.a'iP/'ty. weight Jome shellfish and finfish from pc'1 e ' wa' - have been reported to con- =. up 'P/kg. The importance of ii >;< tea '* i t' in r induction is uncertain. Fri'A.-mio' ">i -aP- -elationship has been es- ^',1:^0.' - * -! ncer is produced experi- . .Pi . animals, it either affects :' 'c- ?.<i in man or occurs by a V. appear relevant (93). >> . ' 1 iraip. cf ..d^ ' u- . -r.. u-. -i kr-. v r , wa, b? v-.m ' residues lr. e fo . constitute haza- imines * /drocarbons that were DDT, aldrin, dieldrin) jpatomas) in mice, and to be carcinogenic in ompounds accumulate in rease in liver tumors aplanied their introduction .he latent period is not ibility exists that effects ars from now. Pesticide ,-f secondary amines may the formation of nitros- proves to be a cause of 1 ci ' V. d $ fall into this category minants. The extent to buted to the production rficult to evaluate, but V es i - ' reused preserve food and give i< -ol. ur, and nsistency. Food dyes - '<v he first < nicals investigated be- <.x - ;n:ciur-v milarities to accepted Consumption of chethe; foo^ intr sere usei S whi wer in a yell in I the' Petper diti t the atic sac and C ar 'w X) :or n- tha O co der cCnO stra --j twe the OT fro- son cha hur last twe rele beg Ion froi the C ing' ere. 1.6 mei as r a ri the Ch. 3--Factors Associated With Cancer S3 chemicals by consumers who are unaware of their presence is partially responsible for the food safety laws. The laws require that newly introduced direct food additives be carefully screened in the laboratory before they may be used. Some definitely carcinogenic chemicals, which have now been withdrawn from foods, were used for a time before their carcinogenicity in animals was discovered. These include butter yellow, thiourea, and a food preservative used in Japan. The number of cancers, if any, which these additives produced is unknown. Doll and Peto (93) estimated an attribution of less than 1 percent of total cancer mortality to food ad ditives. Of the many food additives presently used in the United States, three require special consider ation: the artificial sweeteners cyclamates and saccharin, butylated hydroxytoluene (BHT), and nitrites and nitrates. Cyclamates were shown to produce bladder cancer in animals (24) and were removed from commerce in 1969. OTA (282) reviewed data that show saccharin caused cancer of the blad der in rats in two circumstances: 1) in straightforward feeding studies when given over two generations, and 2) when given following the administration of a powerful carcinogen. OTA also reported positive and negative results from a number of short-term tests. Therefore, some evidence supports the conclusion that sac charin causes cancer in defined conditions. The human evidence that has been collected over the last few years fails to distinguish clearly be tween saccharin and cyclamates, but it is more relevant to the former as the use of saccharin began earlier (in 1902) and has continued longer. An increase in incidence or mortality from bladder cancer could not be attributed to the introduction of saccharin (16). One epidemiologic study (175) showed that ingestion of artificial sweeteners by males in creased their relative risk of bladder cancer to 1.6. (The incidence of bladder cancer among men who did not consume saccharin was taken as relative risk equal to 1.0; saccharin users had a risk 60-percent higher). The results reported in the study were consistent with those from the laboratory. In both cases cancer occurred in the bladder in males. Subsequent epidemiologic studies have failed to confirm the relative risk of 1.6. The large case-control study of bladder cancer in the United States conducted by NCI revealed rel ative risks of 0.99 for males and 1.01 for females among users of artificial sweeteners in any form (174). However, Wilson (362) points out that one projection from animal studies suggests that only about 500 bladder"Cancer cases annually are to be expected from saccharin consumption in the United States (see also 282). The relative risk represented by 500 cases could not have' been detected in the case-control study. Doll and Peto (93) reviewed five case-control studies which examined saccharin consumption and bladder cancer. With the exception of Howe . et al. (175), the relative risks in all the experi ments lie very close to 1.0; some slightly higher; some lower. They conclude that the "human evidence could hardly be more null," at least for cancer of the bladder, which was the anatomic site affected by saccharin in rats (93). BHT has been used extensively as an antioxi dant for many years. It is not carcinogenic by itself, but has been reported to enhance the pro duction of lung tumors by urethan in mice (93). Conversely, its antioxidant effect is found to in hibit the formation of active carcinogens in the laboratory (93), and similar effects might be ex pected to occur in vivo. It has been postulated that its use-- and perhaps that of the more wide ly used butylated hydroxyanisole--have con tributed to the decline in mortality from stom ach cancer (24). Nitrites have been used to preserve meat since the last century (also see above). According to Shubik (328), nitrites added to food constitutes only 10 percent of the total nitrite reaching the stomach in vegetables and saliva. However, if the formation of nitrosamines and nitrosamides in the intestinal tract proves to be of practical importance, dietary nitrite may play a role in cancer formation. The National Research Coun cil's Panel on Nitrates (cited in 93) was unable to reach any conclusions about their quantitative effect, but advised that reasonable measures be URL 03968 -T 84 Technologies for Determining Cancer Risks From the Environment taken to minimise human exposure to N-nitroso . compounds, including the restriction of the amounts of nitrate and nitrite added to meat t products. 1 There is great uncertainty regarding the con tribution of the compounds discussed above to " the formation of cancer. The possibility also ex ists that other additives might have detrimental . effects. Diet Summary Dietary components discussed above, and many others, are currently the subjects of inten sive research, from which some results should be'known within the next few years. The out comes may show diet to be a factor in determin ing cancer occurrence at many sites, particularly the stomach, large bowel, endometrium, gall bladder and in tropical countries, the liver. Diet may also be shown to affect the incidence of cancers of the breast and pancreas, and, through the antipromoting effects of retinoids, to reduce incidence of epithelial cancers in many other tissues. If these or other hypotheses are proven, practicable means of dietary modifica tions may eventually be identified to reduce can cer rates. Cairns (43) draws attention to the probable difficulty of changing cancer incidence even if a direct link is shown between a diet constituent and cancer incidence: Cancer of the lung is due to a pleasant and highly addictive habit, cancer of the large in testine and breast are most common in affluent countries and so are presumably associated with some desirable habit, such as a diet high in ani mal fats, that the rich nations can afford and the others cannot .... These are signs, therefore, that the campaign to prevent cancer may come into some conflict with people's immediate desires. OCCUPATIONAL EXPOSURES The last several years have seen a heated dis_ cussion concerning the contribution of occu pational exposures to cancer in the United States. Since the formation of the Occupational Safety and Health Administration (OSHA) in 1970, 20 regulations have been promulgated and two more proposed relating to suspect carci nogenic factors in the workplace. Labor unions 4 and public interest groups have criticized OSHA for moving slowly, while industry and their _ tirade associations claim unnecessary irrational regulation and undue expense. In an effort to implement a comprehensive and rational policy for the regulation of carcinogens in the work place, OSHA held 2 months of hearings in 1978 which attracted the participation of labor unions, environmental groups and spawned a . major new trade association, the American In dustrial Health Council (A1HC). Subsequently, OSHA promulgated "a general policy for the identification and regulation of physical and chemical substances that pose a potential oc cupational carcinogenic risk to humans" (278,279). The first recognized industrial cancer was identified by Percival Pott, a British surgeon, who observed that scrotal cancer occurred more frequently in men who had been employed as chimney sweeps as boys. This led to the iden tification of soot as the first chemical and oc cupational carcinogen. In the ensuing years, many other groups of workers have been found to suffer from occupationally induced cancer. For the purpose of this assessment, occupational exposure is defined as exposure to a substance or physical agent through any route during the course of employment. The workplace setting has proved to be the single most productive source of information in the discovery of carcinogenic substances. This is because of the defined populations involved and higher exposures which can be more easily monitored and identified. Table 17 lists carcino gens and processes found in the workplace which are associated with increased cancer risk. Occupations known to produce an elevated risk of cancer, though the specific agents responsible - Agent Acrylon 4-afTitftor Arsenic contpo Asbestc Auramir of aura Benzen. Benzio Beryliu comp.- Cad mu compc Carbor Chloro' (CMM Chrom chrof* Coat to Coke c Dimett- Epicht "`Vie mat d me 3r~3 oro. ini o no staipf- (0 CD <0 tei 1r ciB jia Radia Soots Tfiorii Vinyl Agerv SOUR TRENDS URL 03910 The Reaganomics of Toxic Chemicals The Toxic Substances Control Act (TSCA) broke new ground when it became law six years ago, directing the Environ mental Protection Agency to balance risks with benefits in managing hazardous chem icals. Though this risk-benefit analysis makes TSCA consistent with the Reagan administration's philosophy of regulation, the government's campaign against waste and overregulation is nevertheless bringing changes to the EPA's stewardship of TSCA. As a result, the mixed reviews that TSCA earned from the chemical industry during its first five years are giving way to growing enthusiasm for the EPA's initia tives. One TSCA effort that everyone ap plauds is the basic inventory--the list of 55,000 chemicals now being manufactured in the U.S. that EPA has assembled since 1977. "No similar document exists in any other country," E. Hamilton Hurst, vicepresident for environmental health and safety of Nalco Chemical Co., told the American Chemical Society's 1982 spring meeting. The list is "a valuable tool" for both EPA and the industry, and many for eign nations are busy assembling similar data. There is also industrial enthusiasm for EPA's changing philosophy concerning the TSCA requirement that chemical compa nies give "premanufacturing notice" (pmn) 90 days before they put new chemicals into production. EPA is to use that 90-day peri od to determine if the chemicals have risks greater than their potential benefits. Douglas G. Bannerman. acting director of EPA's Office of Industry Assistance, told the ACS that over 1,000 pmns have come to EPA since 1979, when the require ment took effect.The result, he said, "is the only complete and accurate record of the development and commercialization of new chemicals ever compiled ... a wealth of information." But C.W. Umland of Exxon Chemical Americas said the pmn requirement has caused "substantial disruption of new chemical development and introduction." and J.R. Yost of the Muskegon Chemical Co. called the pmn requirement "a unique burden for small companies," where most U.S. innovation is supposed to take place. Dr. Bannerman's analysis shows that just over 70 percent of the pmns have come from large companies (with over S500 mil lion in annual sales). And he admits that by adding costs and delays. TSCA may have caused "a significant reduction in new- product innovation among smaller special ty chemical firms." Of more than 1,000 pmns received. EPA has kept only 9 chemicals off the market. But Dr. Bannerman pointed out that TSCA's greater achievement has been to keep other new chemicals off the market by encouraging companies to take "a more responsible look" on their own. This low rejection rate, coupled with the Reagan administration's initiatives for reg ulatory relief, have encouraged EPA to issue some exemptions: pmns will no longer be required for three classes of new chemi cals: high-molecular-weight polymers. chemicals that are intermediates to other products and used only within one manu facturing plant, and some very low-produc tion specialty items. And Dr. Bannerman brought the ACS a special message from Washington: "Don't let the pmn require ments limit your creative spirit in the development and commercialization of new chemicals." But J. Clarence Davies of the Conserva tion Foundation is not entirely comfortable with these initiatives. The long latency period of chronic effects from chemicals means that the pre-Reagan pmn system simply hasn't been proven; even if a dan gerous chemical has slipped past EPA's sieve, we wouldn't know it yet. EPA's other job under TSCA is to study the 55,000 existing chemicals already in production for possible, hazards--clearly a prodigious effort that is beyond any reason able resources EPA might have had even during the Carter administration. That en terprise has now been largely abandoned as just too ambitious. In doing that. Dr. Davies told the ACS, the EPA "surrendered what should be one of the main benefits of TSCA--the oppor tunity to comprehensively review the uni verse of commercial chemicals." But he lamented that this retrenchment might be necessary. "Under Reaganomics, asking EPA to look for new chemical problems is akin to complaining about the lack of good French restaurants in San Salvador. If sur vival is in question, some desirable goals must be sacrificed."--J.M. Regulating Food: Cancer on Your Plate? U.S. food safety laws mandate "zero risk," in contrast to risk-benefit comparisons re quired by the Toxic Substances Control Act (see above). The Delaney Clause says that no food product can be sold in the United States if it contains a measurable quantity of any known carcinogen. What is a "measurable quantity"? That depends on when you ask. and of whom, for methods of chemical analysis are constant ly becoming better. Analysts are able to identify ever smaller quantities of pollu tants in a food, down to parts per billion or even trillion. 76 Technology Review It was a somewhat different world when Representative James J. Delaney first heard in 1950 the judgment of Dr. W.C. Heuper, chief of the Environmental Can cer Section of the National Institutes of Health: "I do not believe that one can establish a safe dose of carcinogens. ... ft would be a wise precautionary measure not to add any chemicals to our food supply that produce cancer either in man or in experimental animals." Thus, the Delaney Clause was bom. All this explains why major revisions in the Food Additive Amendments of 1958 to August/Seotember 1982 ' " "........................................... --` '- W the Food, Drug, and Cosmetic Act are now cy. Dr. Horwitz cited analyses of identical manufacturer would be free to market the before Congress--and why members of the materials by different laboratories that var product unless FDA had objected. American Chemical Society devoted so ied by 55 percent at the 40-parts-per-bil- A different approach, sponsored by Rep much of their 1982 spring meeting to the lion level. resentative Albert Gore, would permit food issue. The problem of aflatoxins--the carcino impurities in "de minimus amounts." Rep The issue arises in part because of the genic molds sometimes found on corn and resentative Gore doesn't define de mini increasing use of plastics in food packag peanuts--provides another example of the mus, but Mr. Taylor says the implication is ing--and the fact that minute quantities of difficulties, expressed by one participant: clear: "de minimus refers to an amount of a the polymers in the containers migrate to "If you can see a carcinogen, it will hurt substance that, taking into account what is their contents. It's now possible to detect you; but if you can't see it, it won't." known about [its] toxicity . . . poses risks such migration, and so the offending con When analytical chemists could detect that are so small or so improbable that they tainers are outlawed under present law. But aflatoxins at the level of 20 parts per billion can reliably be said to be de minimus. or Howard R. Roberts, former acting director in consumer products such as peanut but simply not worthy of government con of the Bureau of Foods in the Food and ter, the FDA lowered the limit accordingly. cern." Drug Administration (FDA) who is now But by now it's possible to detect I part per But the question of risk, absent in with the National Soft Drink Association, billion of aflatoxin in peanut butter, and today's "technical-feasibility" criteria, told the ACS that many of these constitu ents "occur at levels so minute as to be toxicologically insignificant." enforcing that limit would in some years opens the issue to personal rather than legal foreclose marketing as food nearly half of or scientific decisions. Judith Rae Brunton the nation's peanut butter, said Joseph V. of the Washington law firm of Hyman and He joined Michael R. Taylor, formerly a Rodricks. vice-president of Clement Asso Phelps, specialists in food, drug, and cos lawyer for the FDA. in appealing for a ciates, Inc., who was associated with FDA metic law, told the ACS that "there is no change. "The public health benefit derived for 13 years. single acceptable risk level for all foods from closely regulating such minute Accordingly, FDA has decided with among all consumers." If the "technical- amounts of substances is frequently out some uneasiness to base its regulatory deci feasibility" criteria are abandoned, "one of weighed by the tremendous resources be sion not only on analytical detectability but the choices among the acceptable levels of ing expended--by government and indus "on the technical capability of manufactur risk--no risk---has been eliminated." In try--in chasing the elusive zero," said Mr. ers to control the aflatoxin content of their that case, she warned, the public will turn ( Taylor. products," said Dr. Roberts. The question to the scientific community "to identify That search for the "elusive zero" can be of risk is not dealt with explicitly. and quantify the risks that must be ac expensive. William Horwitz, acting direc Two bills now in Congress would change cepted." And so far, "Science hasn't been tor of FDA's Science Policy Staff, de all this. Senator Orrin Hatch would require able to answer those questions." scribed a new method for determining the food manufacturers to tell the FDA in Ms. Brunton concluded that the present dioxins in fish down to a level of about 50 advance what chemicals, and how much of controversy is merely "the opening argu parts per trillion--at a cost of about $ 1,000 them, would be in new food products, and ment in what promises to be a long debate per determination. they would be asked to provide appropriate about fundamental social change."-- And then there's the problem of accura toxicological data. Ninety days later the J.M. URL 03971 World's Fastest t Uranium There was jubilation when final confirma tion came at 6:00 p.m. Tuesday, May 11. A giant atom smasher at the University of California's Lawrence Berkeley Laborato ry--the Bevalac--had become the world's first machine to accelerate uranium parti cles to nearly the speed of light. Long a goal of nuclear scientists, this feat provides a vast new capability for exploring the fundamental properties of matter under the extreme conditions of heat and pressure created when nuclei of heavy elements collide. "These collisions," says Herman Grunder, director of LBL's accelerator division, "will momentarily cre ate concentrations of neutrons and protons far bigger, hotter, and denser than any seen previously." The achievement results from S6 million Traeks lelt by soma of the first uranium nuclei ever to be accelerated to nearly the speed of light. The dark streaks are the laet one-half millimeter of three tracks as they came to rest in a special photographic emulsion. The bottom track shows a nucleus splitting into two lighter nuclei. The work was done at the Bevalac located at the University of California's Lawrence Berkeley Laboratory. August/September 1982 Technology Review 77 URL 03912 **' 'oth species face a common enemy at high frequency, (ii) the mechanism by which they compete displays density- dependent effectiveness, and (iii) the two species harm each other less than either is harmed by the common enemy. That this suggestion is sufficiently general to accommodate cooperation between spe cies is supported by observations of plant defense guilds (9), aggregations of tube-building polychaetes (10), associa tions of hydroids and bryozoans (11), and my own observation that B. simplex, a treelike congener of B. turrita, settles preferentially into dense stands of young B. turrita colonies (12). I have shown here that a density de pendence in interference competition can lead to group living and that the formation of groups in a sessile organism necessitates intraspecific cooperation. Yet cooperation and competition are generally viewed as being virtually oppo site extremes of organism interactions. My results suggest the reverse. In cer tain cases, interspecific competition may provide the very selective pressures that lead to the evolution of cooperation. Leo W. Buss Department ofBiology. Yale University, New Haven, Connecticut 06511 Rcfemca and NMm 1. W. C. Allee, Animal Aggregations (Univ. of Chicago Press. Chicago, 1931). p. 431; Cooper ation Among Animals (Schumaa. New York, 1931). p. 233; R. D. Alexander, Anna. Rev. Ecol. Svst. 5. 325 (1974). . 2. W. C. Allee, J. Exp. ZotA, ii. 269(1912): H. S. Horn. Ecology 49 . 682 (1968): A. Zahari. Ibis 113. 203 (1971); S. T. Emlen and N. I. Demong, Science 188. 1029 (1975): E. B. Gilt and L. L. Wolf. Ecology 36. 333 (1975); D. I. Rubenstein. R. J. Barnett'. R. S. Ridley. P. H. Klopfer. Ibis 119, 10 (1977): B. C. R. Bertram, in Behavioral Ecology An Evolutionary Approach, ). Krebs and N. Davies, Eds. (Blackwell. Oxford, 1978). p. 245; D. I. Rubenstein. Persped. Ethol. 3. 205 (1978). 3. Fertile colonies were collected from Eel Pond. Woods Hole. Mass., and induced to release their larvae by light shock. A minimum of 100 larvae were released into clean Nalgene contain ers. Each container was covered with a piece of plastic mesh (Vexar) divided into six equalized (100 cm*) regions. In each region colonies of B. turrita were attached in different densities in a modified Latin Square by looping the stalks through the holes in the mesh. After a period of 6 to e hours, the mesh was removed and the number of larvae settled in each region was counted. 4. Colonies that had previously settled on Vexar mesh were attached to a new (10 by 10 cm) Vexar substratum with cable ties and monofila ment line. Attachment by monofilament failed in some instances, resulting in the loss of some colonies. Only one loss, however, could not be unambiguously attributed to a failure in the attachment method. All weight gains are ex pressed as averages of those colonies that re mained attached. 5 G. Lutand. Ann. Soc. Zool. Belg. 91. 157(1961). 6. Clearly the evolution of group living will be dependent upon the frequency at which interfer ence competition occurs. 7. Cooperation may be defined as "a dynamic ecological state of organisms living in aggrega tion characterized by sufficient mutual benefit to outweigh disadvantages associated with crowd ing" (Webster's Collegiate, ed. 7). Allee (/) was the first to use the word in this general sense. Recent workers have restricted the use of the term cooperation to apply only when there is a risk of fitness loss if others do not participate iSi. Allee was aware of such risks although he did not incorporate them in his definition of cooper ation: for example, see "the costs of under crowding" [W. C. Allee, The Social Life of Animals (Norton. New York. 1939). p. 293. figure 2). The gregariousness of B. turrita is a cooperative act under either definition. 8. R. Trivers. Q. Rev. Biol. 46. 35 (f 971); I. D. Chase. Am. Nat. 115. 827 (I9B0): R. Axelrod and W. D. Hamilton. Science 211. 1390 (1981). 9. P. R. Atsatt and D. J. O'Dowd. Science 193. 24 H976). 10. J. B. C. Jackson. Am. Nat. 111. 743 (1977); S A. Woodin. J. Mar. Res. 34 . 25 (1976). 11. R. W. Osman and J. A. Kaugsness,Science 211. 846(1981). 12. An experiment designed exactly like that de scribed in the text was attempted with B. sim plex larvae and fl. turrita residents. All B. turrita colonies were 0.05 g or less in weight. Results are presented as B. turrita density (colonies per square centimeter), and the percentage of B. simplex larvae settling: 0.0i colonies. 15 per cent: 0.05. 10:0.08. 35; 0.13, 28:0.17. 8:0.22. 2. 13. 1 thank R. Grosberg, G. E Hutchinson. I. B. C. Jackson. B. Keller. N. Knowlton. J. Moore. C. Wahle. and J. Wulff for technical assistance or comments (or both). Supported by a Steps To ward Independence fellowship from the Marine Biological Laboratory, Woods Hole. Mass. 3 February 1981; revised 5 May 1981 / Milk of Dairy Cows Frequently Contains a Leukemogenic Virus Abstract. Milk or viable milk cells collectedfrom 24 dairy cattle naturally infected with bovine leukemia virus were inoculated into lambs, which were subsequently examined for the development of infection. With this bioassay, infectious virus was demonstrated in the milk of 17 ofthe cows. Bovine leukemia virus is leukemogenic in at least two mammalian species, is widespread in commercial dairy herds, and can infect a wide range of hosts in vo,oi and cells, including human cells, in vitro. Bovine leukemia virus (BLV) is a hori zontally transmitted, probably insectborne oncornavirus that differs from the leukemia viruses of other species in sev eral important ways (/, 2). In vivo, BLV is usually present in the lymphocytes in a covert, nonproductive state; the infected lymphocytes do not show viral RNA, viral antigens, or viral particles unless they are grown in vitro for a few hours (2-4), BLV is universally regarded as the Table 1. Detection of BLV in lambs inoculat ed with whole milk from naturally infected cows. Each milk sample was injected intraperitoneally into one or. in most cases, two lambs less than 7 days old. The lambs were subsequently examined for the presence of BLV antibodies by means of the immunodif fusion test with BLV glycoprotein antigen (23) and. in most cases, for the presence of infectious BLV by means of the syncytia induction assay (14). Symbols: 7, one or both sheep positive for BLV 12 or more months after injection; --, sheep negative for BLV, Milk donor Day of lactation on which milk was collected 10 30 50 G-245 +- BF-157 + - BF-138 + - G-142 - 4. G-266 - BF-269 - BF-291 - G-255 - G-189 - G-43 -+ G-257 -- G-256 + G-263 G-24 - + G-265 - G-198 - causative agent of the adult enzootic form of bovine leukemia (lymphosarco ma), the most common fatal malignancy of dairy cattle (/). Under experimental conditions BLV infects sheep (5, 6), goats (7), and apparently chimpanzees (). BLV-infected sheep frequently de velop leukemia (5, 6). There is evidence that BLV can cross species barriers un der natural conditions and infect both sheep and capybaras (9). In vitro, BLV infects cells of various origins, including human and simian cells (10, 11). More than 20 percent of the dairy cows and approximately 60 percent of the herds surveyed in the United States were infected with BLV (/). Bovine leu kemia viru* infection is also prevalent among cattle of most other countries studied. Thus the question of whether dairy cows naturally infected with BLV release infectious virus into milk is an important public health consideration. Several in vitro infectivity assays for BLV have been developed, but they are not suitable for detecting the virus in milk, mainly because of the composition and frequent bacterial content of this secretion. Sheep provide an alternative means for determining the presence of BLV in milk because they are highly susceptible to BLV infection and seem to resist the bacteria present in various bovine secretions and excretions (/2). In a preliminary experiment, BLV was demonstrated in one sheep injected with the milk of one naturally infected cow (12). We used the sheep bioassay to deter mine the frequency with which infec tious BLV is released into the milk of cows naturally infected with BLV. In our first experiment. BLV-free lambs iou 1X06-80^ 8| 0828-I0I4S01.000 Copyright t 1981 AA\S SCIENCE. VOL. 213. 28 AUGUST 19X1 were injected with fresh milk (100 ml) collected from 16 BLV-infected cows belonging to a dairy herd (13) maintained at the University of Pennsylvania (New Bolton Center) and to a commercial dairy herd. The milk from six cows was assayed on two different lactation days, but, owing to the short lambing season, recipient sheep of appropriate age were not available for use in assaying milk from any one cow more frequently. Sev en lambs injected with milk samples from a BLV-free cow and seven lambs that were not inoculated served as con trols. All lambs were obtained from the same BLV-free flock and were raised in the same facilities. Transmission of BLV from sheep to sheep has not been ob served (5, 6). As shown in Table 1, the injection of milk from 8 of the 16 infected cows induced BLV antibodies in one or both recipient lambs. These antibodies were usually detected 8 months after inocula tion and persisted at least until the sheep were 12 to 16 months old. In no instance was the milk from a cow positive for BLV on more than one of the lactation days. In several instances only one of a pair of sheep inoculated with a milk sample developed BLV antibodies. It appears, therefore, that the quantity of infectious BLV in 100 ml of milk is close to the sheep bioassay's limit of sensitiv ity. If this is the case, it cannot be determined whether the failure to detect BLV in milk of an iofected cow at successive lactation dates is due to inter mittent release of the virus or to a con stant release at low levels. The persistence of the antibody re sponse of the recipient sheep indicated that the response was due to infection rather than to immunization by virion antigens in the milk inoculum. Further more, infectious virus was detected in the blood lymphocytes of six of nine antibody-positive sheep examined with the syncytia induction assay for BLV {14). None of the control sheep showed BLV antibodies or infectious BLV dur ing the 16 months of observation. In cattle, BLV seems to infect only the lymphocytes, and BLV-infected lym phocytes do not usually synthesize BLV panicles in vivo (2-4). We therefore as sumed that the infections in milk-inject ed sheep are due mainly to the presence in milk of BLV-infected lymphocytes rather than to free extracellular virus. Thus, in an attempt to increase the sensi tivity of the sheep bioassay, we conduct ed a second experiment in which lambs were injected with viable cells isolated from milk rather than with whole milk. Of the nine infected dairy cows used as 28 AUGUST 1981 donors, only one (BF-138) had been used in the first experiment. As shown in Table 2, milk cells collected from the cows at one or more of three sampling times induced BLV infection or BLV antibodies in the recipient lambs. Most of the sheep developed antibodies before their lymphocytes became positive for BLV. However, by 8 to 10 months after inoculation there was a close correlation between the presence of antibodies and the presence of BLV. There were no significant differences among lactation days in the frequency with which BLVinfected cells were present in the cows' milk. Nine sheep injected with lympho cytes from BLV-free cows and eight uninjected sheep housed with the infect ed sheep remained negative for BLV for at least 16 months. In most instances, BLV was detected in both sheep inoculated with milk cells from a BLV-positive cow (Table 2). In contrast, only half of the infected wh'iu milk samples in the first experiment in duced BLV infection in both recipient lambs. Moreover, whereas most of the sheep inoculated with milk cells became infected within 3 months, BLV was usu ally detected only after 8 months in the sheep inoculated with whole milk. Thus the sensitivity of the sheep bioassay was indeed greatly increased when milk cells were used as the inoculum. It is likely that in the first experiment the percent age of cows identified as having infec tious BLV in their milk would have been greater had milk cells rather than whole milk been assayed. The fact that infectious BLV or BLVinfected cells are present in the milk of most naturally infected cows does not necessarily mean that calves become in fected with BLV by the oral route. Some Table 2. Detection of BLV in lambs inoculated with milk cells from naturally infected cows. The cells were isolated by centrifugation, washed twice, and injected subcutaneously (10* viable cells per lamb); N.D.. not determined. Milk donor Day of laciation Recipient sheep 3 to 4 months after inoculation BLV antibodies* BLVt 8 to 10 months after inoculation BLV antibodies BLV G-4IB2 G-149 G-312 0-314 G-281 BF-138 BF-283 BF-306 BF-316 4 1 to 2 4 to 6 10 to 13 1 to 2 4 to 6 10 to 13 1 to 2 1 to 2 4 to 6 i to 2 4 to 6 10 to 13 1 to 2 4 to 6 10 (o 13 1 to 2 4 to 6 10 to 13 4 10 to 13 14 23 24 3 60 $4 62 64 18 55 59 65 52 67 57 66 53 63 8 15 19 7 9 11 16 12 yy 51 58 4 2 l 10 6 w + ++ + - N.D. N.D. 4- + + + +- + + - +- + - ++ 4- -- + + - ++ +- + + - ++ + - ++ + 4- + + - - ++ - - ---- - - -- - - -- + - 4- + + - ++ + ++ ++ 4- + - * 4. + + * + -+* + + -+ + + T-- ++ + + - 4- + + - +- -- ++ + N.D. N.D, N.D. + + N.D. N.D. 4- N.D. N.D. + ++ ++ + As determined by the immunodiffusion test with BLV glycoprotein antigen. +As determined by the competitive radioimmunoassay for BLV p25 (24). Peripheral blood lymphocytes were cultured for 48 hours with phytohemagglutinin iji. washed by centrifugation, resuspended in buffer (0.02M iris. 0.1 A/ NaCl. 0.001W tiDTA (pH 7,5). 0 5 percent Nomdet P40. 0.2 percent sodium deoxycholaie, and 2-phenylmcthylsul> fanilyl fluoride] to a final concentration of 3 * 10* cells per milliliter, and frozen and thawed three times. After incubation at 37C for 15 minutes, the extract was clarified by centrifugation and tested with the competitive radioimmunoassay. An extract was considered to be positive for BLV when 4ft ul [equivalent to 10* lymphocytes! displaced more than 20 percent of the labeled antigen. 1015 URL 03973 'searchers in Europe have concluded that bovine leukemia, and therefore BLV, is transmitted via milk (15, 16). However, owing to the experimental de sign and methodology used, the data on which this conclusion was based do not indicate whether the calves became in fected prenataily, by milk ingestion, or by contact. Studies conducted under ap propriate experimental conditions have shown that milk-borne transmission of BLV, as compared with contact trans mission. is rare if it occurs at all (17, 18). The resistance of calves to milk-borne infection with BLV is probably due mainly to the maternal virus-neutralizing antibodies that all calves nursed on BLV-positive dams acquire through co lostrum (17-19). The present study and the data on the prevalence of BLV infection in dairy herds indicate that humans are often orally exposed to BLV. Although the infectivity of BLV is apparently de stroyed by pasteurization (20), people in many countries consume unpasteurized milk. Moreover, it is not known whether pasteurization destroys the biological ac tivity of the proviral BLV DNA in the infected cells of milk. There is no evi dence that BLV can infect humans, but neither do the data exclude this possibili ty. Although attempts to demonstrate BLV antibodies, particles, and antigens in humans have been negative (/), the findings are not conclusive because of the limited sensitivity of the assays used and because some BLV-infected cells do not synthesize virus particles or viral antigens (2-4, 10, 11). Molecular hybrid ization studies have failed to demon strate BLV-related sequences in human tumors (21), but only a few tumors were examined and the probe used was only partially representative of the viral genome. While earlier epidemiological surveys showed no association between human and bovine leukemia (7), the most recent survey, involving a large number of cases, showed a statistically signifi cant increase in human acute lymphoid leukemia in areas with a high incidence of bovine leukemia and BLV infection (22). Clearly, the question of whether BLV poses a public health hazard de serves thorough investigation with the most sensitive virological and immuno logical techniques available, particularly highly representative molecular probes. Jorge F. Ferrer Simon J. Kenyon Phalguni Gupta University of Pennsylvania School of Veterinary Medicine. New Bolton Center. Kennett Square 19348 Reference* and Notes 1. J. F. Ferrer. Adv. Vet. Sci. Comp. Med- 24. I (1980). 2. C. Cabradilla. P. Gupta. Cold Spring Harbor Symp. Cell Prolif. 7. 987 (1980). 3. N. D. Stock and J. F. Ferrer. J. Sail. Cancer Inst. 48. 985 (1972). 4. V. Baliga and J. F. Ferrer. Proc. Joe. Exp. Biol. Med. 156. 388 (1977). 5. M. J. Van der Maaten and J. M. Miller. Bibl. Haematol. (Baseh 43. 377 (1976). 6. S. J. Kenyon. J. F. Ferrer. R. A. McFeely. D. C. Graves. J. Natl. Cancer hut., in press. 7. H- E. Hoss and C. Olson. Am. J. Vet. Res. 35. 633(1974). 8. M. J. Van der Maaten and J. M. Miller. Vet. Microbiol. 1. 331 (1976). 9. C Maria. N. Lopez, L. Alvarez. H. Castanos. W. Espaoa. A. Leon. A. Bello, paper presented at the Fourth international Symposium on Bo vine Leukosis, Bologna. Italy. 3 to 7 November I960. 10. C- A. Diglio and J. F. Ferrer. Cower Res. 36. 1936 (1976). 11. D. C. Graves and J. F. Ferrer, ibid., p. 4132. 12. J. M. Miller and M. J. Van der Maaten. J. Nad. Cancer Inst 62. 423 (1979). 13. 1. F Ferrer. D. A. Abt. D. M. Bhatt. R. R. Marshak. Cancer Res. 34. 893 0 974). 14. J. F. Ferrer, C. Cabradilla, P. Gupta. Am J Vet. Res. 42. 9 (1981). 15. A. Burny et al,, Adv. Cancer Res. H. 23! (1978). 16. O. C. Straub. F. Weiland. B. Frcnzel. Dtsch. Tieraerztt. Wochenschr. 81. 581 (1974). 17. J. F. Ferrer and C. E. Piper. Ann. Recti. Vet. 9, 803 0 978). 18. C. E. Piper, J. F. Ferrer. D. A. Abt, R. R. Marshak. J. Natt. Cancer Inst. 62. 165 (1979). 19. J. F. Ferrer. C. E. Piper. D. A. Abt, R. R. Marshak, Am. J. Vet. Res. 38. 1977 0 977). 20. L. Baumganener. C. Olson. M. Onuma, J. Am. Vet. Med. Assoc. 169, 1189 (1976). 21. R. Kettman er al.. Ann. Reck. Vet. 9. 837 (1978). 22. K. J. Donham. J. W. Berg, R. S. Sawin. Am. J. Epidemiol. 112, 80 (1980). 23. P. Gupta and J, F. Ferrer. Ann. Reck. Vet. 9. 683 0978). 24. H. C. McDonald and J. F. Ferrer, J. Natl. Cancer Inst. 57. 875 0976). 25. We thank D. Worth and B. Thompson for tech nical and secretarial assistance, respectively. Supported by National Cancer Institute grant 3P0I-CA-14J93. the Wetterberg Foundation, and Pennsylvania Department of Agriculture grant ME-24. 10 March 1981; revised-26 June 1981 Emergence of Posttetanic Potentiation as a Distinct Phase in the Differentation of an Identified Synapse in Aplysia Abstract. The developmental time coarse ofpostteianic potentiation was studied at an identified chemical synapse. In stage 11 juveniles (3 weeks after metamorpho sis), the synaptic connections made by cholinergic neuron L/q onto postsynaptic neurons Ls to L were present but showed no posttetanic potentiation. In stage 13 adults (12 weeks after metamorphosis), the same tetanus resulted in an increase of 300 percent in the synaptic potential. A similar pattern was observed at two other identified synapses in the abdominal ganglion. Thus, the initial steps in synapse formation do not include the expression of this plastic capability. Rather, at least 10 weeks is required between the onset ofsynaptic function and the final expression of mature synaptic properties. Many chemical synapses in the adult animal can be altered for long periods as a result of previous stimulation. It is important in the study of neural differen tiation to determine whether the capabil ity for plastic change is part of the ma chinery that is present in the initial estab lishment of a functioning synapse, or whether these capabilities represent sep arate regulatory processes that emerge independently later in development. Although the initial development and maintenance of synapses have been studied extensively (1-4), little is known about how neurons acquire plastic capa bilities. Knowledge about the develop ment of synaptic plasticity may help to elucidate the mechanisms of normal be havioral maturation, as well as the mech anisms that contribute to disorders in perception and motor coordination that occur when animals are reared in abnor mal perceptual or motor environments (5). We examined posttetanic potentia tion (PTP), a common form of plasticity, at an identified chemical synapse in the abdominal ganglion of the marine mollusk Aplysia californica. We found that PTP develops as a late step in the matu ration of the synapse and is independent of transmission. The gradual emergence of PTP at an* identifiable synapse pro vides an opportunity for studying the detailed mechanisms underlying this form of synaptic plasticity. We used laboratory-reared animals at various stages of development, from stage 11 juveniles weighing 3 mg at 3 weeks after metamorphosis to reproductively mature, stage 13 adults weighing 100 g or more at 12 weeks after metamor phosis (6). We examined a specific, iden tified inhibitory synapse between presynaptic neuron L|0 and one of its five identified follower cells (L2 to L$) locat ed in the rostral quadrant of the left abdominal ganglion (7). Even in the youngest animals studied, both the preand postsynaptic neurons were clearly identifiable and large enough <15 p.m in diameter in animals weighing 3 mg) to permit successful intracellular record ings. To obtain reliable postsynaptic po tentials. we hyperpolarized the follower cells to at least 50 mV below the reversal potential of the inhibitory postsynaptic 1016 0036-807?.'8]/082 8-1016501.00/0 Copyright 1981 AAAS SCIENCE. VOL. 213. 28 AUGUST 1981 W6E0 nan 1 \ DIET, NUTRITION, AND CANCER: INTERIM DIETARY GUIDELINES Sushma Palmer, Project Director Kulbir Bakshi, staff scientist National Academy of Sciences 2101 Constitution Ave., N.W. Washington, D C., 20418 Received November 9, 1982; accepted February 4, 1983 Address reprint requests to Dr. Palmer IJRL 03975 1152 Palmer and Bakshi Acknowledgments The authors and the members of the Committee on Diet, Nutrition, and Cancer wish to acknowledge the assistance of the many scientists who drafted reports or who served as consultants for this study and of those scientists who participated in workshops or conferences designed to aid the committee in the preparation of the report. The authors are grateful to Dr. Andrew Chiarodo and Dr. Diane Fink, the former project officers of the National Cancer Institute, for their constant support of this project. C u r" O GJ --J3J a? JNCl, VOL. 70, NO. 6, JUNE 1983 TABLE OF CONTENTS ABSTRACT .............................................................................................................. . 1153 INTRODUCTION .......................................................................................................................1153 THE SCOPE OF WORK AND THE APPROACH TO EVALUATING THE LITERATURE ... 11 54 MAJOR CONCLUSIONS AND INTERIM DIETARY GUIDELINES ....................................... 1154 Interim Guidelines ................................................................................................................ 1154 Lipids (Fats and Cholesterol): Basis for Conclusions....................................................... 1155 Fruits and Vegetables ......................................................................................................... 1157 Salt-Cured. Salt-Pickled, and Smoked Foods .................................................................. 1159 Food Additives and Contaminants .................................................................................... 1160 Mutagens in Food ............................................................................................................. 1161 Alcohol ..................................................................................................................................1162 Other Dietary Components ................................................................................................1163 Contribution of Diet to Overall Risk of Cancer.................................................................. 1163 COMMENTS ............................................................................................................................. 1163 REFERENCES ......................................................................................................................... 1163 ABSTRACT--Th Commit!* on Diet, Nutrition, and Canear of tht National Academy of Sciences recently evaluated the role of diet in carcinogenesis. Both epidemiological and laboratory avidanca sug gests that a high intaka of total fat incraasas susceptibility to cancer of different sites, particularly tha braast and coion. In apidamiologIcal studies frequent consumption of certain fruits and vegetables and in laboratory experiments some components of fruits and vagatabias, especially cruciferous vegetables, appear to dacraasa tha incidence of cancers at various sites. In contrast, fraqgant con sumption of salt-curad, salt-pickled, or smoked foods, possibly because they may contain nitrosamines or polycyclic aromatic hy drocarbons, appears to incraas# the risk of ssophageal or stomach cancer. Excessive alcohol consumption among smokers appears to be associated with an elavated risk of cancars of the oral cavity, esophagus, larynx, and respiratory tract Interim dietary guidalines to reduca tha risk of cancer were proposed in accordance with these conclusions. No definitive conclusions were reached for other dietary factors, including total calorias, cholesterol, fiber, and se lenium, nor could the quantitative contribution of diet to overall cancer risk be estimated___JNCl 1983; 70:1151-1170. INTRODUCTION On June 16, 1982, a committee of scientists*1 appointed by the National Research Council-National Academy of Sciences (the Academy) and operating within the Assembly of Life Sciences completed an analysts of the literature concerning the role of diet in the etiology and prevention of cancer.* This article summarizes the basis for the major conclusions and recommendations of the committee's report "Diet, Nutrition, and Cancer" (/). The search for the causes of cancer has been an important branch of cancer research. Considerable effort has been devoted to studying the influence of both environmental and genetic factors on the incidence of cancer. In the course of this research, it became clear that most cancers have external causes and in principle, therefore, should be pre ventable. For example, blacks and Japanese residing in the United States develop the spectrum ofcancers that is typical for the United States but different from that in Africa or Japan. Epidemiologists have demonstrated a correlation between diets consumed in modern affluent societies and the inci dence of cancers in such organs as the breast and colon. However, to establish causal relationships and to determine which, if any, of the dietary- components is responsible have proved to be difficult. Similar difficulties are encountered in laboratory experiments. Like humans, most animals have a significant incidence of cancer in old age. and the rates of these cancers often tend to be affected by changes in diet. However, the influence of diet on spontaneous and experi mentally induced cancers is not easily investigated, partly because the underlying mechanisms and molecular biology of the cancers still are not fully understood. Indeed, the effects of diet were often regarded as a nuisance, i.e., yet another variable standing between the investigators and their measurement of carcinogenicity. As a consequence, researchers only recently have returned to the study of diet as a factor in carcinogenesis. Despite these difficulties, considerable research effort has been devoted in the last 3-4 decades to understanding whether and how dietary factors affect carcinogenesis. In June 1980, the National Cancer Institute (NCI) commis sioned the Academy to evaluate the implications of this research. Lack of consensus among scientists about the interpretation of the research data, and as a consequence heightened confusion and concern among the public and the Congress about the implications of this knowledge for public health, served as a major impetus for the Academy's study. 1 The Committee on Diet. Nutrition, and Cancer is chaired by Clifford GrohMcin. Vice-chairman is John Cairns. Members of the committee are Kolscrt Berliner. Sclwyn Bmitnian, T. Colin Campbell. Joan D. Guvwiw. I-uurcncc N. Kolond. David Kritchcv-tky, Walter Men*. Anthony B Miller, Michael J. Prival. Thomas Slai^a, and lx* Waticnbcr^. Takahi Sutpinura serves as an adviser to the committee. 1 This work was conducted pursuant to Public Health Service contract N0lCPJ)3ri03 from the Division of Cancer Cause and Prrven lion. National Cancer Institute. 1153 JNCl, VOL. 70, NO. 6, JUNE 1983 1154 Palmer and Bakshi The Academy established a multidisciplinary 14-member panel, the Committee on Diet, Nutrition, and Cancer, to address NCI's request to 1) "review . . . the state of knowl edge and information pertinent to diet/nutrition and the incidence of cancer," 2) "develop a series of recommenda tions related to dietary components (nutrients and toxic contaminants) and nutritional factors which can be com municated to the public," and 3) "based on the above stateof-the-art appraisals and the identification of gap areas, develop a series of research recommendations related to dietary components and nutritional factors and the inci dence of cancer." The committee's 500-page report summarizes scientific evidence concerning nutritive and nonnutritive dietary fac tors in carcinogenesis and offers to the public several interim dietary guidelines to reduce the risk of cancer. THE SCOPE OF WORK AND THE APPROACH TO EVALUATING THE LITERATURE Scientifically valid data on diet and nutrition in relation to cancer are provided by three major sources: epidemiolog ical studies on human populations, experimental studies on animals, and in vitro tests for genetic toxicity. However, there are a number of limitations to the interpretation of data from each type of study. The strengths and weaknesses of epidemiological studies, of dietary methodology used in epidemiological studies, and of laboratory experiments to study the relationship between dietary factors and carcino genesis are discussed in the report (1). Epidemiological studies of the association between diet and cancer have examined many common forms of cancer, although most research has been directed to cancers of the gastrointestinal tract and the breast and other tissues sus ceptible to hormonal influence and less research has been directed to cancers of the respiratory tract and urinary bladder. The committee examined the evidence for each of these sites. In the laboratory, investigators have attempted to shed light on the mechanisms by which diet may influence carcinogenesis. In this pursuit, they have examined the ability of individual nutrients, food extracts, or nonnutritive components of food to enhance or inhibit carcinogenesis and mutagenesis, thereby providing epidemiologists with testa ble hypotheses regarding specific components of the diet. Because the data from both epidemiological and laboratory studies are generally grouped according to dietary constit uents or cancer sites, (he committee decided to organize its report in a similar fashion, focusing on all major nutritive and nonnutritive constituents of the diet and on each major site. Although the committee considered the evidence from all tvpes of epidemiological studies, it had the most confidence in data derived from case-control studies and from the few cohort studies thai have been reported. Instead of relying on aggregate correlation data, these studies are based on the collection and analysis of data on individuals, and the investigators attempt to control for confounding variables. Thus these studies provide more definitive evidence indica tive of meaningful associations than do data derived from correlation and descriptive studies. Greater emphasis was given to the results of case-control or cohort Studies that were designed to examine a specific hypothesis. Because most epidemiological studies suffered from methodological weaknesses, the consistency of results among studies of dif ferent types and in different settings assumed particular importance in the evaluation of the epidemiological evi dence. In evaluating laboratory evidence, the committee placed more confidence in data derived from studies on more than one animal species or test system, in results that have been reproduced in different laboratories, and in the few data that indicated a dose response. Overall, the strength and preponderance of data and the degree of concordance between the epidemiological and laboratory evidence determined (he certainty of the conclu sions in the report. Concurrence between epidemiological and laboratory evidence was considered a prerequisite to proposing dietary guidelines to the public. MAJOR CONCLUSIONS AND INTERIM DIETARY GUIDELINES Overall, the committee concluded that "the differences in the rates at which various cancers occur in different human populations are often correlated with differences in diet. The likelihood that some of these correlations reflect cau sality is strengthened by laboratory evidence that similar dietary patterns and components of food also affect the incidence of certain cancers in animals." This general statement was prefaced with the warning that it is not yet possible to be absolutely certain about some of the associations between diet and cancer because scientists have yet to attain a complete understanding of how diet influences carcinogenesis. The committee likened the cur rent state of knowledge on this subject to the knowledge on cigarettes and lung cancer approximately 20 years ago. However, the evidence was judged to be sufficiently con vincing to propose interim dietary guidelines regarding total fat, fruits, vegetables, and whole grain products; salt-cured, salt-pickled, and smoked foods; and alcohol. In addition, some research and regulatory recommendations were pro posed concerning food additives, contaminants, and muta gens. A discussion of the ma.jor conclusions follows the six interim guidelines presented below. Interim Guidelines 1) .. the committee recommends that the consumption of both saturated and unsaturated fats be reduced in the average U.S. diet. An appropriate and practical target is to reduce the intake of fat from its present level (approximately 40%) to 30% of total calories in the diet. The scientific data do not provide a strong basis for establishing fat intake at precisely 30% of total calories. Indeed, the data could be used to justify an even greater reduction. However, in the judgment of the committee, the suggested reduction (i.e., one-quarter of the fat intake) is a moderate and practical target, and is likely to be beneficial.'* 2) "The committee emphasizes the importance of includ ing fruits, vegetables, and whole grain cereal products in the W L 03973 jNCt, VOL. 71), NO. 6, JUNE 1983 Diet, Nutrition, and Cancer 1155 daily diet.'* However, the committee cautioned against the use of high and potentially toxic doses of nutrient supple ments. 3) "The committee recommends that the consumption^/ food preserved by salt-curing (including salt-pickling) or smoking be minimized." 4) "The committee recommends that efforts continue to be made to minimize contamination of foods with carcino gens from any source. Where such contaminants are una voidable. permissible levels should continue to be established and the food supply monitored to assure that such levels are not exceeded. Furthermore, intentional additives (direct and indirect) should continue to be evaluated for carcinogenic activity before they are approved for use in the food supply." 5) "The committee recommends that further efforts be made to identify mutagens in food and to expedite testing for their carcinogenicity. W'here feasible and prudent, mu tagens should be removed or their concentration minimized when this can be accomplished without jeopardizing the nutritive value of foods or introducing other potentially hazardous substances into the diet." 6) "The committee recommends that if alcoholic bever ages are consumed, it be done in moderation." These interim guidelines were presented with the proviso that "it is not now possible and may never be possible to specify a diet that would protect everyone against alt forms of cancer." However, the guidelines are likely to reduce the risk of cancer, and they concur with good nutritional prac tices and with dietary recommendations issued by other organizations concerned with public health, e.g., the Amer ican Heart Association (?), the NCI (J), and the U.S. Department of Agriculture and the U.S. Department of Health, Education, and Welfare (4). Lipids (Fats and Cholesterol): Basts for Conclusions Evidence supporting an association between fat intake and cancer is derived both from epidemiological studies and laboratory experiments. Epidemiological evidence:fats.--Epidemiological studies have repeatedly shown an association between dietary fat and the occurrence of cancer at several sites, but most of the evidence pertains to cancers of the breast, colon, and prostate gland. Breast cancer is known to be affected by hormonal status, but diet has also been suggested as a contributory factor (J). Several international correlation studies have shown strong direct associations between per-capita fat intake and breast cancer incidence or mortality rates {6-JO). In general, the correlations in these studies were higher for total fat than for other dietary factors, such as total protein, animal pro tein, and specific fat components. Intracountry data sets in the continental United States, England, and Hawaii have generally supported these findings {II-13). For example, on the basis of dietary interview data from a population survey in Hawaii, Kolonel et al. (13) correlated ethnic sex-specific consumption of fat with corresponding patterns of breast cancer incidence and found significant associations for total fat. animal fat, and both saturated and unsaturated fats. The findings of three case-control studies and one cohort study, in which husbands of wives with and without breast cancer were interviewed, further support a role for dietarv fat in the risk for breast cancer {14-17). An increased risk of large bowel cancer has also been asso ciated with higher fat intake in both correlation and casccontrol studies. Correlations of international incidence and mortality rates for large bowel cancer in relation to dietary variables suggest that colon cancer, and to a lesser extent rectal cancer, are strongly associated with total dietary fat (6, 10). In contrast to the Strong correlations observed in international studies, there is a striking lack of a good correlation within countries. For example, Bingham et ai. (IS) calculated average intakes of nutrients by populations in different regionsofGreat Britain and found no significant association between high fat intake and mortality from colon and rectal cancers. Lyon and Sorenson (Iff) also reported little difference in fat intake between the popula tion of Utah (with a low risk for colon cancer) and that of the United Staves as a whole. One possible explanation for this discrepancy is that there may be insufficient variability in food intake patterns within a country to demonstrate a strong association with the risk of colon or rectal cancer. In contrast, the variation in fat intake among countries is much greater, and so the overall association is much stronger. Several case-control studies have shown a direct associa tion between frequent consumption of high-fat foods, total fat intake, especially saturated fat, and colon cancer (14, 20-22). Jain et ai. (22), for example, reported a gradient in response between the risk of both colon and rectal cancers and the level of dietary fat, especially saturated fat. Several investigators also found a direct association be tween meat consumption (meat is an important source of fat, especially saturated fat) and colon cancer (23-25). However, these results were not reproduced in other studies (26-28). Some of the conflicting data on fat and colon cancer may be explained by the possible confounding effects of dietary fiber, which has been suggested as exerting a protective effect. The incidence of colon cancer in Denmark and NewYork is higher than in Finland, although the per-capita fat intake of these three populations is similar. In contrast, the intake of fiber-containing foods among Finns appears to be higher than in the other 2 groups (29, 30). In an ongoing cohort study in Minnesota, Bjelke (26) has found a reduced risk of colorectal cancer in subjects with a high index of vegetable consumption. (Vegetables are good sources of fiber and micronutrients.) However, no such effect has yet been observed in a parallel cohort in Norway (26). Some investigators have suggested that the protective efTect of dietary fiber may be due to its effect on adsorption, dilution, or metabolism of carcinogens or promoters or to decreased fecal transit times and thus decreased contact between carcinogens and the colon mucosae (30, 31). However, contrary to many reports, differences in fecal transit times due to differences in fiber intake have not been consistently reported (32, 33). The data on prostate cancer are limited. However, several correlation studies indicate an association between high fat intake and increased prostate cancer incidence and mortal ity rates (6, 13, 34-36). The results of two case-control studies also support these findings (37, 38). International URL 03919 JNCI, VOL. 70, NO. 6, JUNE 1983 URL 03980 1156 Palmer and Bakshi incidence and mortality data also indicate that there are significant correlations between prostate cancer and cancers of other sites associated with a high-fat diet, including cancers of the breast, colon, and corpus uteri (34, 39, 40). However, there are important exceptions; c-g., the Mormons in Utah have a high incidence of prostate cancer but a low incidence of breast cancer {41). Other cancer sites that have been associated with dietary fat include testis, corpus uteri, ovary, and pancreas (5, 13, 36, 42, 43). The epidemiological data relating fat intake to different cancers are not entirely consistent, and the data for breast and prostate cancer are more consistent than for colon cancer. In the case of breast cancer, the magnitude of the association appears greater in the correlation data than in the case-control studies; e.g., the results of the most thor ough case-control study yet reported by Miller et al. {15) were only weakly positive, perhaps partly reflecting the fact that recent dietary intake was assessed, whereas dietary patterns much earlier in life may have had a greater influ ence on breast cancer risk {44). Studies of changing breast cancer incidence among Japanese migrants to Hawaii and their descendents support this hypothesis {45). The data for colon cancer also are not altogether consistent. Some studies of large bowel cancer may have failed to demonstrate an association with dietary fat, either because they focused on relatively homogeneous populations or because they were not specifically designed to examine the association of fat intake with cancer, Indeed, the studies that were designed specifically to test this hypothesis tended to show the most striking direct associations, especially when the possible confounding efTects of dietary fiber were taken into consid eration {20, 22, 29, 30). Another difficulty is distinguishing clearly between the effect of protein and fat in epidemiological studies. Several of the same international correlation studies and case-con trol studies that have shown an association between high dietary fat intake and cancer at several sites have also suggested possible associations between high intake of total protein or animal protein and increased risk of cancers at a number of different sites, including the breast (5, 3-10, 12, 13, 15), large bowel (6, 22, 46), pancreas (6, 42), prostate gland (13, 36). and endometrium {6, 13). However, the data for protein, especially from analytical studies, are much more limited than for fats. In addition, the more consistent and often stronger association of these cancers with fat intake makes it likely that dietary fat is the more active component. Nevertheless, because of the very high correla tion between fat and protein intake In Western diets, the evidence does not completely preclude an independent effect of protein. Epidemiological evidence: cholesterol.--A relationship between dietary or serum cholesterol and cancer, especially colon cancer, has been reported by many investigators. Most of these studies were designed to examine the association be tween serum cholesterol and cardiovascular disease. Several prospective cohort studies and some intervention trials indicate an inverse association between serum choles terol and cancer, usually mortality from colon cancer in males but not in females {47-55). However, in general, in the intervention trials, low cholesterol levels per se do not appear to be associated with increased mortality from cancer (54, 55). Elimination of early cases of cancer in some of the cohort studies suggests that the lower serum cholesterol levels may be the result of metabolic change accompanying tumor growth rather than the cause of the disease (52, 56, 57). The only case-control study (58) designed specifically to examine the relationship of serum cholesterol to cancer reported a significant inverse association between serum cholesterol levels among controls and male cases with ad vanced tumors only; the study, therefore, supports the hy pothesis that lower serum levels may be due to metabolic changes associated with tumor growth. However, in other studies, the inverse association, especially with colon cancer, persisted despite adjustment for this possibility (48-50). Five cohort studies and three intervention trials demonstrated no association between serum cholesterol and cancer (59-62). Dietary and serum cholesterol levels are frequently not well correlated in individuals, and dietary cholesterol has usually been associated directly with colon cancer (22, 63). Experimental evidence: fats.---The first demonstration that fat could influence tumorigenesis was reported in 1930 by Watson and Mellanby (64). Since then, numerous experi ments have been conducted in which different levels and components of fat were used to study their effects in various tissues and animal species. In studies of mammary carcinogenesis, the influence of lipid nutriture on tumorigenesis appears to follow a consistent pattern, regardless of whether the tumors are chemically induced, occur spontaneously, or result from tumor cell implantation. Tannenbaum (55) demonstrated that high dietary fat enhanced the development of either spontaneous or chemi cally induced mammary tumors in mice. Tannenbaum and Silverstone (55) noted that tumor incidence was greater in obese mice than in normal mice and that caloric restriction inhibited mammary tumorigenesis in normal mice. Feeding mice isocaloric high- and low-fat diets provided evidence that fat rather than calories per se was responsible for enhancing tumorigenesis (55). However, Lavik and Bau mann (67) found chat calories appear to have a greater effect than fat on 3-methylcholamhrene-induced skin tu mors. High-fat diets appear to enhance breast carcinogenesis induced by several carcinogens, e.g., 7,12-dimethvlbenz(ajanthracene (68-71), iV-nitro$o-,V-mcthylurea (NMU) (72), and X-rays (73). Increasing the level of dietary fat up to 20% by weight of the diet appears to increase the inci dence or multiplicity of chemically induced breast tumors, depending on the carcinogen used (74-76). Breast tumori genesis appears to be enhanced when high-fat diets are fed after, but not before, tumor initiation (75). This Is consistent with the concept that dietary fat exerts a promoting effect on tumorigenesis rather than an efTect at the initiation stage. The levels of fat and of the carcinogen appear to be important determinants of the outcome. In some experi ments that used low doses of the carcinogen, high-fat diets decreased the latent period and increased tumor incidence. At high doses of the carcinogen, high-fat diets increased tumor incidence and tumor multiplicity but appeared to JNCI, VOL. 7(1, NO. 6, JUNE 1983 r URL 03981 Diet, Nutrition, and Cancer 1157 have no effect on latency (75, 76). The type of fat also has been shown to be an important factor in breast tumorigenesis. Mammary tumor incidence appears to be uniformly high with all types of fat at 20% of the diet, but diets containing 20% polyunsaturated fat en hance tumorigenesis more effectively than do diets contain ing 20% saturated fat, provided that the fat serves as an adequate source of essential fatty acids (63). Supplementa tion of a diet high in saturated fat (17%) with 3% polyun saturated fat provides the same tumor-enhancing effects as a 20% polyunsaturated fat diet (77, 78). Thus the possible promoting effects of high-lipid diets on breast carcinogenesis appear to depend on the total quantity of fat in the diet (with the maximum effect achieved at 20%) and sufficient polyunsaturated fat to serve as an adequate source of essen tial fatty acids. Lipid requirements for maximum rate of growth of trans plantable breast adenocarcinomas are essentially the same as those described above (79, 80). Limited data also indicate that polyunsaturated fats enhance the growth of both nor mal and neoplastic mammary epithelial cells from rats (8!, 82). In studies of intestinal cancer, increasing the quantity of dietary fat (generally from 5 to 20% by weight) increases the incidence of bowel tumors induced by a variety of carcinogens: 1,2-dimethylhydrazine (DMH) (83-86), metfi- ylazoxymethanol (MAM) acetate, 2\3-dimethyl-4-ammobiphenyl (DMAB), or NMU (86, 87). A promoting effect of dietary fat was suggested by studies in which high-fat diets increased the frequency of small and large bowel tumors when fed to rats after administration of azoxymethane (AOM) but not before or during administration of the carcinogen (88). However, in studies in which the effects of high and low levels of dietary fat on bowel tumorigenesis were compared, the animals rarely were "pair fed" or fed isocaloric diets. Consequently, it is possible that the tumor enhancing effect of high-fat diets on the bowel is related to increased caloric intake. An intake of 5% polyunsaturated fat appears to have a greater tumor-enhancing effect than does the equivalent level of saturated fat. At a 20% fat level, no clear difference was apparent between the effects of polyunsaturated fat and saturated fat (83, 84). Some studies suggest that bowel tumorigenesis is associ ated with increased concentration of fecal bile acids that accompanies high levels of fat intake. Through the efforts of many investigators, the concept has evolved that fecal bile acids and metabolites of cholesterol may function as cocar cinogens. carcinogens, or promoters in tumorigenesis of the large bowel (84, 89-93). To date, however, no active carcin ogen derived from bile acids has been isolated from human or animal feces, and, therefore* the effect of bile acids on bowel tumorigenesis is not fully resolved. Increasing dietary lipids appears to increase the incidence of carcinogen-induced hepatomas in rats (74, 94). Differing effects of the quality and quantity of dietary fat depend on the stage of hepatoma development at the time of exposure (74). Hepatic tumorigenesis, unlike breast carcinogenesis, is enhanced by a deficiency of lipotropic factors in rats fed high-fat diets (95). Limited data indicate that the yield of pancreatic adenocarcinomas is increased more effectively by polyunsaturated fat than by saturated fat (95, 97). x Experimental evidence: cholesterol.--Some investigators hat e suggested that dietary cholesterol is cocarcinogenic in the bowel (98), but other studies have not confirmed these findings (85, 99). Lipids (fats and cholesterol): conclusion.--The committee con cluded that of all the dietary components it considered, the combined epidemiological and experimental evidence is most suggestive for a causal relationship between high-fat intake and the occurrence of cancer. Both epidemiological studies and experiments in animals provide strong evidence that increasing the intake of total fat increases the incidence of cancer at certain sites, particularly the breast and colon and to a lesser extent the prostate gland, and, conversely, that the risk is lower with lower intakes of fat. Laboratory data suggest that when fat intake is low, polyunsaturated fats are more effective than saturated fats in enhancing tumorigenesis, whereas the data on humans do not permit a dear distinction to be made between the effects of different components of fat. In general, however, the evidence from epidemiological and laboratory studies is consistent and. therefore, strengthens confidence in conclusions about the importance of fat in carcinogenesis. Although most of the data suggest that dietary fat has promoting activity, there is not enough evidence to warrant the complete exclusion of an effect on initiation. Data concerning the association between serum choles terol levels and colon cancer are inconsistent. An inverse association between serum cholesterol levels and colon can cer in males, but not in females, has been noted in several studies, but not in others, and alternative explanations of these findings have been suggested. Furthermore, the exper imental data on cholesterol and cancer risk also are too limited for any inferences to be drawn. Fruits and Vegetables Fruits, vegetables, and whole grain products are impor tant sources of many nutrients, especially vitamins and their precursors (such as 0-carotene) and minerals. They are also a good source of dietary fiber and nonnutritive inhibitors of carcinogenesis. The recommendation emphasizing their daily consumption is based on the following epidemiological and laboratory evidence. Epidemiological evidence: cruciferous vegetables.--Several casecontrol and a few cohort studies point specifically to an inverse association between the consumption of vegetables in general, raw vegetables (e.g., lettuce and celery), or cruciferous vegetables in particular (e.g., cabbage, cauli flower, brusseis sprouts, and broccoli) and cancers of the alimentary tract [e.g., cancer of the stomach (26, 36, 100102) and iarge bowel (27) J. Epidemiological evidence: vitamin A and carotene.--A growing accumulation ofepidemiological evidence from several casccontrol and a few cohort studies indicates an inverse rela tionship between vegetable consumption or estimated "vitamin A" intake and cancers of the lung (103-106), larynx (107), bladder (108), esophagus (109-112), stomach (36, 100, !0l), colorectum (26), and prostate gland (38). With few exceptions, the estimates of vitamin A intake were JNCJ, VOL. 70, NO. 6, JUNE 198J 1158 Palmer and Bakshi derived from the frequency of ingestion of groups of foods (e.g., green and yellow vegetables) known to be rich in precursors of vitamin A (notably, )3-carotene) and a few foods such as whole milk and liver containing preformed retinol (vitamin A). Thus to a large extent, these Studies have measured indirect indices of vitamin A intake. Two case-control studies found an inverse relationship between serum vitamin A levels and the risk of bronchial carcinoma (1/3, 114), and, recently, three cohort studies in the United States and England have also reported a similar relationship for scrum vitamin A and subsequent overall cancer risk (49, 52, 1/5). Epidemiological evidence: vitamin C--As in the Studies for vitamin A, several case-control studies and a few correlation studies suggest an inverse association between fresh-fruit consumption or estimated vitamin C intake and cancers of different sites, especially cancers of the stomach (13, 26, 116-/18), esophagus (1/0-112, 1/9, /20),and larynx (107), and uterine cervical dysplasia (121). In contrast, for colon cancer Jain et al. (22) found no association with vitamin C consumption in a case-control study. The observations for gastric cancer are consistent with the hypothesis that vitamin C offers protection by blocking the reaction of secondary and tertiary amines with nitrite to form nitrosamines (122). As described later in this paper, several of the same studies have indicated a direct associa tion between consumption of cured or smoked foods and stomach cancer (26, 116, ! 17). These studies mostly provide indirect evidence of a protective role for vitamin C against gastric cancer, inasmuch as they are based on the consump tion of fresh fruits and vegetables, known to contain high concentrations of vitamin C, rather than on actual measure ments of vitamin C intake. Epidemiological evidence: fiber.--Some evidence concerning the protective role of fiber-containing foods in colon cancer was also discussed earlier in this paper. Some correlation data support the hypothesis that dietary fiber affords pro tection against cancer of the large bowel (29, !23). However, other correlation studies have not supported this hypothesis (7, 19, 63). Bingham and colleagues (!8) found no signif icant correlation between total fiber intake and mortality from colon cancer, although they observed that the mean intake of the pentosan fraction of total dietary fiber was inversely correlated with mortality from colon cancer. Casecontrol studies also have yielded mixed results; some studies indicated a protective effect of fiber on colon carcinogenesis (20, 25, 124), whereas others found no inverse association (22) or reported a direct association (21). Laboratory evidence.--In laboratory experiments, a number of nutrients (especially vitamin A and vitamin C) and other constituents of fruits and vegetables (e.g., fiber components, indoles, and isothiocyanates} have been tested for their effect on chemically induced carcinogenesis in vivo. Laboratory evidence: nonnutntive components offruits and vegeta bles.--Besides the presence of vitamins in fruits and vegeta bles, which may offer protection, there is some evidence that nonnutritive components of some plant products, especially cruciferous vegetables such as cabbage, broccoli, brusseh sprouts, and cauliflower, afford protection against chemi cally induced neoplasia in vivo (125, 126). These compounds include aromatic isothiocyanates (127), flavones (128), in doles (129), protease inhibitors (130),^-sitosterol (13/, 132), and naturally occurring phenols (/33). The possible modes of action of these nonnutritive substances have been re viewed by W'attcnbcrg (125, 126, 134). Laboratory evidence: vitamin A, retinoids, and (3-carolene.-- Stud ies on animals indicate that a deficiency of vitamin A can result in an increased susceptibility to carcinogen-induced neoplasia (/35-139), whereas higher levels of vitamin A or its synthetic analogues, the retinoids, appear to have a protective effect against the induction of cancer by chemical carcinogens (135, 140-145) or to lead to regression of skin papillomas (146). In contrast, some investigators observed an increase in hormone-induced or chemically induced neo plasia by an enhanced intake of the vitamin (138, 147, !43). In vitro experiments in organ cultures support the concept that a deficiency of vitamin A enhances neoplastic response to chemical carcinogens (149) and that its addition inhibits carcinogenesis (150). Recently, carotenoids (e.g., /S-carotene) have been shown to protect against the development of L'Y light-induced skin tumors in hairless mice (151), whereas an opposite effect of/?-caroienc was observed earlier by Shamberger (141) for chemically induced carcinogenesis. Laboratory evidence: vitamin C.--Ascorbic acid in vitro and in vivo can prevent the reaction of nitrites with amino compounds, thus inhibiting the formation of carcinogenic N-nitroso compounds and indirectly inhibiting tumor for mation (152-/56). In contrast, the effect of vitamin C on in vivo carcinogenesis from already formed carcinogens has been inhibitory in some studies (157-159) but not in others (160-/62). In cell cultures, ascorbic acid appears to prevent chemically induced transformation of cells or to induce reversion of transformed cells (163). Laboratory evidence: fiber.--Studies on the role of dietary fiber in chemically induced carcinogenesis have produced mixed results; e.g., bran administered by different routes appears to protect against DMH-induced colon cancer in animals (164-166). Cellulose also was found to protect rats against DMH-induced tumors (167, 168) but not against tumors induced by AOM or NMU (169, 170). Pectin did not inhibit DMH-induced colon tumors (168). Recently, Giauert et al. (/"/) reported that dietary agar enhanced DMH-induced colon cancer in mice. Although some data suggest that some types of fiber (e.g., bran and cellulose) can inhibit the action of certain chemical carcinogens in vivo, the data from different studies are difficult to compare, primarily because of lack of uniform experimental protocols. Fruit and vegetables: summary and conclusions. -- Numerous epi demiological studies indicate an inverse association between the consumption of foods containing carotene, vitamin A, or vitamin C or of vegetables in the cruciferous family and the risk of cancer of the urinary bladder, large bowel, lung, stomach, or esophagus. The epidemiological evidence was derived from several case-control studies as well as cohort studies in different parts of the world, including the United States. Hawaii, Japan, Singapore, Norway, and Iran. In laboratory experiments, vitamin A, its synthetic analogues (the retinoids), indoles, and other nonnutritive chemicals in cruciferous vegetables tend to suppress the induction of URL 0398 ro JNCI, VOL. 7(J, No. 6, JUNE 198:1 Diet, Nutrition, and Cancer 1159 URL 03983 chemically induced tumors. The evidence for an inhibitory effect of /3-carotene is limited, and that for fiber is inconclu sive. There is evidence that vitamin C can inhibit the formation of some carcinogens in vitro and in vivo. Because investigators have not yet clearly identified which compo nents of vegetables or fruits are responsible for the protective effect observed in epidemiological studies and because some nutrients like vitamin A can be toxic in high doses, the committee recommended against the use of high-dose sup* plements of individual nutrients. Salt-Cured, Salt-Pickled, and Smoked Foods The recommendation to minimize the consumption of cured, pickled, and smoked foods is based on evidence from some parts of the world (especially China, Japan, Hawaii, Hungary, Norway, and Iceland) that populations frequently consuming salt-cured (including salt-pickled) or smoked foods have a greater incidence of cancer, especially esopha geal or stomach cancer. Laboratory investigations support the epidemiological observations by demonstrating carci nogenicity and/or mutagenicity of polycyclic aromatic hy drocarbons (PAH) or N-nitroso compounds that are pro duced bv some methods of smoking or curing foods. These compounds cause mutations in bacteria and cancer in ani mals and are suspected of being carcinogenic in humans. Epidemiological evidence: pickled, smoked, and cured foods.-- Correlation studies in China on increased susceptibility to esophageal cancer have implicated trace mineral deficiencies (especially molybdenum), diets low in animal products, high intakes of pickles, pickled vegetables, and moldy foods con taining N-nitroso compounds (possibly produced by fungal contaminants), and hot foods (172, 173). However, studies in Iran have shown no differences in nitrosamine levels in foods in regions of high and low risk for esophageal cancer U/2). In contrast to the limited evidence for esophageal cancer, a number of studies in different parts of the world have associated frequent consumption of cured, pickled, or smoked foods with stomach cancer, e.g., salted fish in Nor way (26), home-smoked meat products containing high levels of benzojajpyrene (BP) in the Vend population in Hungary (174, 173), smoked trout and smoked mutton containing high levels of PAH in northwestern Iceland (176), and occupational exposure to curing smoke and frequent consumption of smoked fish in coastal regions of Latvia (177). Studies of migrants to Canada and the United States have supported this evidence. Choi et al. (178) reported that mortality from gastric cancer among Icelandic immi grants who consumed large amounts of smoked and pickled foods in Manitoba was twice as high as that among people born in Manitoba. Kriebel andjowett (179) attributed the disproportionately high rates of stomach cancer in Minne sota, Wisconsin, and the Upper Peninsula of Michigan to high rates among migrants from northern Europe and sug gested that native-born Americans have shared the increased risk of the migrants, possibly by adopting the "high-risk" diet of the foreign-born residents. In a case-control study, Haenszel and colleagues (101) studied Japanese living in Hawaii. They reported an ele vated stomach cancer risk for both Issei (migrants) and Nisei (offspring) who frequently used pickled vegetables and dried and or salted fish. These findings were not confirmed in a similar study in Japan. The authors interpreted this differ ence in findings to difficulty in demonstrating associations with dietary factors in nonmigrant populations and the frequent failure of case-control studies to establish the sig nificance of small differences in food habits among cases and controls (102). However, other case-control studies have reported more frequent consumption of bacon in the Netherlands (H6), fried foods, especially bacon drippings, in the United States (117), and frequent use of salted pickled foods in Japan (ISO) among patients with gastric cancer, as compared to consumption by controls. Epidemiological evidence: salt.--Except for the few instances in which salt, nitrate, nitrite, N-nitroso compounds, or PAH have been labeled as the potential cancer-causing agents, in ~ general, the specific causative agents in cured, pickled, or smoked foods have not been identified. The evidence relat ing salt intake per se to gastric cancer is tenuous. Joossens and Geboers (181) suggested that both gastric cancer and cerebrovascular disease are related to salt intake. However, salt is generally used in combination with nitrate or nitrite to preserve food (182). Furthermore, Okamura and Matsuhisa (183) found no correlation between the salt content of salted fish and the death rate for gastric cancer in Japan. Epidemiological evidence: nitrate, nitrite, and Pl-nitroso com pounds.--Because many N-nitroso compounds are strong carcinogens under a variety of conditions and in many species (184, 185), much concern has been expressed during the past 10-15 years about their role in the etiology of human cancer. The most important sources of nitrosamines in the American diet currently are cured meats, especially bacon (186). Humans can be exposed to N-nitroso com pounds indirectly from the use of nitrate and nitrite as preservatives in meats and other cured products. Nitrate can be reduced to nitrite, which can interact with dietary substrates such as amines or amides to produce N-nitroso compounds. Vegetables contribute most of the nitrate in gested, but they also contain inhibitors of nitrosation such as ascorbic acid. More than one-third of the average daily intake of nitrite is contributed by the ingestion of cured meats, approximately one-third by baked goods and cereals, and less than one-fifth by vegetables (186). Findings from epidemiological studies conducted in Col ombia, Chile. Japan, Iran, China, England, and the United States (Hawaii) have been consistent with the hypothesis that exposure of humans to high levels of nitrate or nitrite in the diet or drinking water is associated with increased incidence of cancers of the stomach and the esophagus. [5*r review by the Academy (186); for examples, see (101, 116, 117, 187-189).j The associations with cancer in these studies were based either on correlations of high-risk population groups with corresponding exposures to food and water supplies or on comparisons of the frequency of consumption of foods containing these substances (plus secondary amines) by gastric cancer patients and controls rather than on direct measures of exposure. The Academy's Committee on Nitrite and Alternative Curing Agents in Food (186) recently con- JNCI, VOL. 70, NO. 6, JUNE 1983 T160 Palmer and Bakshi eluded that these studies do not provide conclusive evidence for a causal relationship between nitrate and nitrite and human cancer. However, because nitrate can be converted to nitrite and subsequently to N-nitroso compounds that are suspected human carcinogens, the committee recommended that exposure to these agents be reduced {186). A number of investigators have suggested a possible mech anism for the role of N-nitroso compounds in human gastric cancer {182, 190-192): bacterial colonization of the stom ach, especially in cases of achlorhydria, consequent reduc tion of dietary nitrate to nitrite, and endogenous synthesis of N-nitroso compounds. Weisburger et al, {182} postulated and Ohshima and Bartsch {193) recently demonstrated that reducing agents such as ascorbic acid may interfere with the endogenous production of N-nitroso compounds that occurs when dietary nitrite reacts with amines or amides. Such a mechanism could explain the protective effect of green or yellow vegetables and of other vitamin C-containing foods mentioned earlier in this paper. Experimental evidence: nitrate, nitrite, and N-nitroso compounds. -- The few experiments conducted in animals have provided no evidence that nitrate is carcinogenic {194-/96) or mu tagenic or that nitrite is directly carcinogenic {186, 19720/). However, nitrite can interact with dietary amines or amides or with endogenous metabolites to produce N-nitroso compounds that induce cancer. Nitrite is mutagenic, at (east in microbial systems. A majority of the approximately 300 different N-nitroso compounds tested, with the use of any of several routes of administration, have been shown to induce cancer in various tissues of one or more species of laboratory animals {/86). Experimental evidence: PAH.--Small amounts of PAH are present as contaminants in a variety of grilled, roasted, or conventionally smoked foods. Vegetables can become con taminated by PAH from air, soil, or water: fish and shellfish can assimilate such chemicals from their marine environ ments {202). However, unless vegetables or seafood are obtained from highly contaminated environments, the ma jor sources of PAH appear to be the smoking, cooking, or other methods of processing foods {202, 203). Of the more than 100 PAH found in the environment, approximately 20 arc carcinogenic in laboratory animals. Of the five PAH found to be carcinogenic when adminis tered orally, three--BP. dibcnz[a, h janthracene (DBA), and benz[<j]amhraccne (BA)--may occur in the diet. The car cinogenicity of BP in various animal species, including mice, rats, and hamsters, has been well established {204-206). BP is mutagenic in microbial and mammalian cell systems {207, 208). DBA is carcinogenic in mice {209) and mutagenic in bacteria {2/0). BA also is carcinogenic in mice {211) and mutagenic in a variety of cell systems {208, 2/0). Pickled, smoked, and cured foods: summary and conclusions. -- Evidence is limited that the consumption of pickled foods, possibly containing N-nitroso compounds, is associated with esophageal cancer in China. The evidence for gastric cancer is more extensive. A number of correlation studies in differ ent parts of the world, a few studies of migrants to Canada and the United States, and some case-control studies suggest that a high risk for gastric cancer is associated with frequent consumption of smoked foods (which may lead to increased exposure to PAH), with frequent ingestion of salt-pickled foods or cured foods containing N-nitroso compounds or nitrate and nitrite (which may result in subsequent in vivo production of nitrosamines), and possibly with ingestion of other carcinogens produced by food processing, including cooking. PAH and N-nitroso compounds cause mutations in various systems, arc carcinogenic in animals, and are suspected of causing cancer in humans. Food Additives and Contaminants Lack of adequate data on a large number of substances precluded a comprehensive assessment of the risk to humans from exposure to additives and contaminants in the diet. The report (/) contains an evaluation of the carcinogenicity of selected chemicals in both categories. Naturally occurring carcinogens.--In addition to nutrients, a variety of nonnutritive substances (e.g.. hydrazines) are natural constituents of foods. Furthermore, metabolites of molds (e.g.. mycotoxins such as the potent carcinogen ana toxin) and of bacteria (e.g.. carcinogenic nitrosamines) may contaminate foods. Many of these are occasional contami nants, whereas others are normal components of relatively common foods. Aflatoxin, a mvcoloxin that occurs in grains and other food commodities, is carcinogenic in several species of ani mals, including mice, hamsters, rats, trout, and ducks, and there is evidence ofa dose response (/, 2/2, 2/3). In addition, anatoxin has been shown to be mutagenic in bacterial and mammalian species {2/4, 2/5). Several other mycotoxins that are found in food are carcinogenic and/or mutagenic in laboratory tests (/, 2/6). However, with the exception of aflatoxin, which has been implicated in liver cancer in humans in some pans of the world (/, 217, 2/8), there is no reliable epidemiological evidence relating other mycotoxins to neoplasia in humans. The experimental and epidemiological evidence concern ing nitrate, nitrite, and N-nitroso compounds was discussed earlier in this paper. Several hydrazine derivatives of two mushrooms--Agancus bisporus and Cyromitra esculenta, both of which are con sumed in various pans of the world--appear to be carcino genic in mice and, under certain conditions, in hamsters, and they are mutagenic in bacteria {2/9-221). However, the significance of these findings for risk to humans cannot be determined inasmuch as there are no epidemiological data. Several pyrrolizidine alkaloids (e.g., symphytine and senkirkine) are carcinogenic in animals and/or mutagenic in several test systems {2/6, 222). Tumors develop in rats fed plants such as coltsfoot and comfrey, which contain these alkaloids (/, 2/6). Cycad nuts, which are eaten in some parts of the world, contain cycasin (methylazoxymcthanol/J-o-glucosidc), a compound known to be carcinogenic in animals (2/6'). Cycasin is also mutagenic in the Ames test after addition of /3-t>-glucosidase (22J). However, no direct evidence has been presented for the carcinogenicity of pyr rolizidine alkaloids and cycasin in humans (/, 216). Consumption of bracken fern, which occurs widely in nature, has been associated with a higher risk of esophageal URL 03984 jnci, voi.. 70, no. 6, junk IU8J URL 03985 Diet, Nutrition, and Cancer 1161 cancer in Japan {224). However, in a case-control study no association was observed between consumption of fiddlehead greens (related to bracken fern) and bladder cancer {225). Coffee drinking has been associated with an elevated risk for bladder cancer in several case-control studies {225-227), but, with some possible exceptions (228), there is no evi dence of a dose response and, therefore, the evidence is inconclusive. CofTee drinking also has been associated with pancreatic cancer {229, 230), but its association with other cancer sites is inconsistent (231-234). Coffee, whether brewed, instant, or decaffeinated, has been found to be mutagenic (2J5-2J7),but not carcinogenic in animal bioas says (238, 239). Other plant constituents, such as methylxanthmes, thi ourea, tannins, coumarin, parasorbic acid, safrote, and estragole, are carcinogenic in laboratory animals and muta genic in bacterial or mammalian cell systems (/, 216). However, the significance of these findings for human health is not known because there are no epidemiological data. In conclusion, many of the naturally occurring substances discussed above have been found to be carcinogenic in laboratory' animals and mutagenic in bacterial and other systems, thereby posing a potential risk to humans. However, no pertinent epidemiological studies have been made except for those on cofTee and afiatoxins and those on nitrate, nitrite, and N-nitroso compounds. The compounds thus far shown to be carcinogenic in animals have been reported to occur in small amounts in the average U.S. diet; however, there is no evidence that any ofthese substances individually makes a major contribution to the total risk of cancer in the United States. This lack of sufficient data should not be interpreted as an indication that these or other compounds subsequently found to be carcinogenic do not present a hazard. Further investigations are necessary. Food additives.--Nearly 3,000 substances are intentionally used to process foods in the United States. Another estimated 12,000 chemicals, such as vinyl chloride and acrylonitrile, which are used in food production and food processing, are classified as indirect additives. Some additives, such as sugar, are consumed in large amounts by the general population. However, the annual per-capita exposure to most of these substances constitutes a minute portion of the diet (240). Although the Food Safety Provisions and, in many cases, the Delaney Clause of the Federal Food, Drug, and Cosmetic Act (241) prohibit the addition of known carcinogens to foods, only a small proportion of substances added to foods has been tested for carcinogenicity according to protocols that are considered acceptable by current standards. More over, except for the studies on nonnutritive sweeteners (242), very few epidemiological studies have been conducted to examine the efTcct of food additives on cancer incidence (/). Of the few direct food additives that have been tested and found to be carcinogenic in animals, all except saccharin have been banned from use in the food supply. Minute residues of a few indirect additives that are known either to produce cancer in animals [e.g., vinyl chloride and acrylo nitrile (243, 244)] <>r to be carcinogenic in humans [e.g., vinyl chloride (2-^5)] occasionally are detected in foods [(246, 247): Breder CV: Personal communication). There is no evidence to suggest that the increasing use of food additives has contributed significantly to the overall risk of cancer for humans. However, the lack of a detectable efTcct may be due to the relatively recent use of these substances, to their lack of carcinogenicity, or to the inability of epide miological techniques to detect the effects of additives against the background of common cancers from other causes. Therefore, no definitive conclusion can be reached until more data become available (/). Environmental contaminants.--Very low levels of a large and chemically diverse group of substances--environmental con taminants--may be present in a variety of foods. The dietarylevels ofsome of these substances are monitored through the Total Diet Studies conducted by the Food and Drug Ad ministration (248). Many of these contaminants have been extensively tested for carcinogenicity and mutagenicity. The results of standard chronic toxicity tests indicate that a number of environmental contaminants (e.g., some organochlorine pesticides and polychlorinated biphenyls) cause cancer in laboratory animals. There is no epidemio logical evidence to suggest that these compounds individ ually make a major contribution to the risk of cancer in humans. However, the possibility that they act synergistically and may thereby create a greater carcinogenic risk cannot be excluded (/). Mutagens in Food It is generally agreed, but not without some controversy, that there U a high degree of correlation between the mutagenicity of compounds in the 5cfmonr//c-microsome assay and their carcinogenicity in laboratory animals (210, 249-252). Recent studies show that this correlation is de pendent on the class of chemical being investigated, and one must exercise judgment in evaluating the significance of negative or positive tests. A chemical that is mutagenic to bacteria or other organisms is generally suspccred of being carcinogenic, although carcinogenicity must be confirmed in long-term tests in whole animals. Considerable attention recently has been directed toward the presence of mutagenic activity in foods. Many categories of mutagenic compounds have been identified in the diet, including food additives, plant constituents such as flavonoids, and contaminants such as mycotoxins. nitrosamines, residues of pesticides, and PAH. In addition, cooking and other processing of food lead to the production of numerous mutagens, many of which have not yet been identified. Most of the studies to identify mutagens in food have used the Ames Salmonella assay. Naturally occurring mutagens infoods.--Among the most wide spread of the known naturally occurring mutagens that are norma) constituents of many foods are the mutagenic fiavortoids (253). Among the flavonol aglyconcs that have been shown to be mutagenic are quercetin, kaempferol, and galangin (253-256). The mutagenic activity of many foods or drinks (e.g., green tea, Japanese pickles, and dillweed) has been attributed to the flavonoids contained in them (257-259), and the mutagenicity of onions and raisins has been shown to be due to quercetin (260). Quercetin was found to be carcinogenic in one study (26t) but noncarcinogenic in three other studies (262-264). JNCl, VOL. 70, NO. 6, JUNE 1983 URL 03986 1162 Palmer and Bakshi The mutagenicity of coffee has been mentioned earlier. Although caffeine has been reported to be mutagenic in bacteria (265), it probably is not responsible for the muta genicity of coffee because decaffeinated cofTee is reported to be as mutagenic as regular coffee under test conditions (237). Crape juice and several varieties of tea also have been shown to be mutagenic (237, 266). Most of the alcoholic beverages from China (267) and whiskies from Japan and North America (268) and some French brandies have been shown to be mutagenic (269). The mutagenicity of some mycotoxins and some plant constituents was discussed earlier in this paper. Alutagens resulting from cooking offoods.--Different methods of cooking and processing foods appear to affect their mu tagenic activity. Charcoal-broiled and grilled beef contains PAH, many of which are mutagenic and/or carcinogenic (270) . The smoking of foods and roasting ofcofTee are major sources of PAH activity (202). Much of the mutagenicity of welt-cooked foods has been shown to be due to pyrolysis of amino acids in proteins (271) . Pyrolysates of most amino acids tested yielded some mutagenic activity, the pyrolysate of tryptophan being the most mutagenic (272, 273). Several of the mutagenic pyrolysates of amino acids or proteins have been tested for carcinogenicity in vivo. For example, pyrolysates of tryptophan--3-amino- 1,4-dimethyl5//-pyrido(4,3-4]indole (Trp-P-1) and 3-amino-l-methyl5//-pyrido[4,3-ijindole (Trp-P-2)--pyrolysates of glutamic acid--2-amino-6*meihyldipyrido[l,2-a:3',2'-rf]imidazole (Glu-P-1) and 2-aminodipyrido( 1 ^-arS'^'-^jimidazole (Glu- P-2)--and pyrolysates of soybeans--2-amino-9//-pyrido[2,3-A]indole (AaC) and 2-amino-3-methyl-9//-pyrido[2,3-]indoIe (MeAaC)--which are all produced during pyrolysis of proteins and amino acids, were found to be carcinogenic (274, 275). Thus it appears that the identifi cation of several of the mutagenic compounds found in pyrolysates of proteins and amino acids was an accurate prediction of carcinogenicity. However, the presence of a carcinogenic chemical in a pyrolyzed amino acid or protein mixture does not necessarily imply that the carcinogen also will be present in normally cooked, uncharted food. Modifiers of mutagenic activity.--A number of substances in food have been reported to either enhance or diminish the mutagenic activity of other substances. Most of these effects have been observed in vitro, but their relevance to effects on intact mammals is unknown. Norharman and harman present in some foods (e.g., broiled beef and sardines) appear to act as comutagens (276-280) in in vitro test systems. However, juices prepared from some common vegetables, fruits, and spices (including cabbage, broccoli, green pepper, mint leaf, ginger, apple, and pineapple) reduced the mutagenic activity of trypto phan pyrolysates (281, 282). Antimutagenic activity also has been found in extracts of wheat sprouts, leaf lettuce, parsley, brussels sprouts, mustard greens, spinach, cabbage, broccoli, and other vegetables (282, 283). A number of antioxidants, such as butylated hydroxytolucnc (BHT) and butylated hydroxyanisole (BHA), were antimutagenic in the presence of a series of mutagens that require metabolic activation in vitro but not in the presence of direct-acting mutagens (284, 285). However, the muta genicity of direct-acting mutagens, such as /?-propiotactone and .V-methyl-Ar-nitrosoguanidine, also was reported to be inhibited by antioxidants such as ascorbate (286, 287). These findings probably reflect an underlying common mechanism concerning mutagenicity inhibition by antioxi dants. Afutagens in food: summary and conclusions.--Many foods con tain mutagenic fiavonoids, such as quercetin, kaempferol, and their glycosides, and some substances found in foods can enhance or inhibit the mutagenic activity of other compounds. Mutagens in charred meat and fish are pro duced during the pyrolysis of proteins that occurs when foods are cooked at very high temperatures. Mutagens can also be produced during normal cooking of meat at lower temperatures. Smoking of foods as well as charcoal broiling results in the deposition of mutagenic and carcinogenic polynuclear organic compounds such as BP on the surface of the food. Most mutagens detected in foods have not been ade quately tested for their carcinogenic activity. Thus the committee concluded that it is not yet possible to assess whether such mutagens are likely to contribute significantly to the incidence of cancer in the United States (/). Alcohol In certain populations throughout the world there appears to be an association between esophageal cancer and the consumption ofcertain locally prepared alcoholic beverages, e.g., in Japan (288), in some African populations (289, 290), in France (29/, 292), and in China (293, 294). Gastric cancer (295) has been associated with red-wine consumption in France. An association between beer drinking and colorectal can cers appears to be statistically significant in certain coun tries. Excessive beer consumption has been correlated with colorectal or rectal cancer in the mainland United States (296, 297), with colorectal cancer in 20 countries (298), and with bowel cancer in males in Great Britain (299). Two cohort studies ((J00); Bjelke E: Unpublished Ph.D. thesis. 1973) reported a direct relationship between colorectal can cer and the frequency of beer or liquor consumption in Dublin and Norway. However, case-control studies of bowel cancer in Finland (301), in the United States (117), and in Norway [(302)', Bjelke E: Unpublished Ph.D. thesis] showed no significant relationship with beer drinking. Similarly, no associations with bowel cancer were observed for beer drink ing in a correlation analysis by Hinds et al. (303) or in a cohort study by Jensen (304). An association between alcohol abuse and cancer at var ious sites (e.g., esophagus, tongue, pharynx, hypopharynx, larynx, and lung) has been recognized for some time (305309). In the United States, case-control studies have estab lished that excessive consumption of alcoholic beverages increases the risk of incurring cancer of the oral cavity (excluding the lip, glottis, and supraglottic region), larynx, and esophagus (3/0-321). Excessive alcohol consumption has also been linked with the development of hepatomas (322, 323), although hepatomas generally are found in JNCt, VOL. 70, NO. 6, JUNE I9#:l Diet, Nutrition, and Cancer 1163 URL 03987 individuals with cirrhosis. In addition to alcohol, hepatitis B virus and aflatoxin have been associated with liver cancer. Because all three of these agents produce hepatic injury that may lead to cirrhosis, they may contribute to the develop ment of hepatomas. Synergism between alcohol and smoking.--Alcohol consumers (especially abusers) are, more often than not, smokers. Flamant et al. {324), assessing both variables, stressed the interaction between alcohol and smoking in cancers of the oral cavity and the esophagus. Studies completed since then have confirmed these findings of an interactive role between tobacco and alcohol in tumorigenesis of the oral cavity, / larynx, lung, and esophagus {3/1, 314, 317, 3/9, 325-329). Avoidance of tobacco and alcohol by males could effect a marked reduction of these cancers (3/7, 328). Excessive alcohol consumption may also augment other disease proc esses such as impaired nutritional status, which may be associated with the development of cancer at these sites (326). Alcohol; summary and conclusions.--In some countries, includ ing the United States, beer drinking has been correlated with an increased risk of colorectal cancer, especially rectal cancer. Some evidence exists that excessive alcohol con sumption may contribute to the formation of hepatomas by a mechanism involving hepatic injury and cirrhosis. Alcohol appears to act svncrgistically with smoking in increasing the ! risk for cancers of the mouth, larynx, esophagus, and respi-j ratory tract. Other Dietary Components The extensive analysis of the literature, only part of which is cited in this paper, encompassed all of the major nutrients and nonnutritive components of diet. However, because of insufficient or inconsistent data or because of lack of con currence between the epidemiological and laboratory evi dence, no definitive conclusions could be drawn about the role in carcinogenesis of many dietary components: total caloric intake, protein, dietary fiber, carbohydrates, choles terol, most minerals (including selenium), vitamin E. and the B-vitamins. Discussion of the literature on these subjects can be found in (/). Contribution of Diet to Overall Risk of Cancer The NCI had requested that the Academy attempt to determine the percentage of cancers that are diet related. Higginson and Muir (330) had suggested that in some regions of England and Wales possibly 30% of cancers occurring in men and 60% in women could be attributed to life-style (including diet). Wynder andGori (331) were more specific, concluding that in the United States more than 40% of cancers in men and almost 60% in women could be related to dietary factors. Using data on mortality but employing a similar approach to that of Wynder and Cori, Doll and Peto (332) provided a wide range (i.e., 10-70%; average, 35%; for the proportion of deaths from cancer that could be reduced by practical dietary means. They con cluded that dietary modification would have the greatest effect on the incidence of cancers of the stomach and large bowel and would affect, to a lesser extent, the incidence of cancers of the breast, endometrium, and lung. The Committee on Diet, Nutrition, and Cancer concluded that cancers of most major sites are influenced by dietary patterns but that the data are not sufficient to quantitate the contribution of diet to the overall cancer risk or to determine the percent reduction in risk that might be achieved by dietary modifications (/). COMMENTS One of the major objectives in examining the literature on diet, nutrition, and cancer was to determine whether any guidance could be provided to the public. The committee held the strong conviction that dietary recommendations are justified only if the evidence is sufficiently convincing that guidelines to reduce the risk of cancer have a high likelihood of benefit and are without discernible risk. The committee was also aware of the potential impact of its report on scientists, the public, and the food industry. However, having been persuaded that the evidence was sufficiently convincing to warrant certain conclusions, the committee believed that not to offer certain interim guide lines would be a dereliction of its responsibility. Further more, the guidelines are consistent with good nutrition and with other recommendations from public health authorities. The committee called for the assistance of the NCI, other federal agencies concerned with public health, and the food industry in disseminating and implementing the interim guidelines. Although in the judgment of the committee these guidelines are likely to reduce the risk of certain diet-related cancers, the committee concluded: "Since (he current data base is incomplete, future epidemio logical and experimental research is likely to provide new in sights into the relationship between diet and cancer. Therefore, the committee suggests (hat the National Cancer Institute estab lish mechanisms to review these dietary guidelines at least every 5 years." The final report, to be published by the Academy in June 1983, will be concerned with future research on diet, nutri tion, and cancer. REFERENCES (/) National Academy of Sciences. Diet, nutrition, and cancer Com mittee an Diet, Nutrition, and Cancer, .Assembly of Life Sciences. National Roearch Council. Washington, D.C.: National Acad emy Press. 1982. (2) American Heart Association. Committee on Nutrition. Rationale of the diet-heart statement of the American Heart Association. Circulation 1982; 65:839A-854A. fj) Upton AC, Fink D, witnesses. National Cancer Institute. Diet. Nutrition and Cancer Program, National Cancer Institute. In: Hearing before the Subcomittee on Nutrition of the Committee on Agriculture, Nutrition, and Forestry. United States Senate, 96th Congress, 1st Session, October 2, 1979. Washington, D.C.: U S. Govt Print Off, 1980:4-10. (4) U.S. Department of Agriculture and U S. Department of Health, Education, and Welfare. Nutrition and your health--dietary guidelines for Americans. Washington, DC : USDA-DHEW, 1980. (5) MacMaHon B, Coif P, Bkown J. Etiology of human brcxti cancer: A review. J Natl Cancer Inst 1973; 50:21-42. JNCl, VOL. 70, NO. 6, JUNE 1983 URL 03988 1164 Parmer and Bakshi (5) Armstrong B, Dull R. Environmental factors and cancer incidence and mortality in different countries, with special reference to dictarv practices. Int J Cancer 1975; 15:617-631. (7) D*vsar BS,1\ inc D. Environmental factors and cancer of the colon and breast. Br J Cancer 1973; 27:167-172. (8) Gray GF., Ptkf MC, Henderson BE. Breast-cancer incidence and mortality rates in different countries in relation to known risk factors and dietary practices. Br J Cancer 1979; 39:1-7. (P) I Ir.sis G. The contributions of diet and childbearing to breast-cancer rates. Br J Cancer 19*8: 37:974-902. (10) Knox F.G. Foods and diseases. Br J Prev Soc Med 1977; 31:71-90. (//) Caikill SP, McGuire WL. Osborne CK, Stern MP Breast cancer mnrtalitv and diet in the United States. Cancer Res 1979; 39 3628-3637. (12) Hems G. Associations between breast-cancer mortality rates, child bearing and diet in the United Kindgom. Br J Cancer 1980; 41:429-437. (13) Kotosei. I.N. Hankin JH. Lee J, Chu SY, Nomura AM. Hinds MU' Nutrient intakes in relation to cancer incidence in Hawaii. Br J Cancer 1981; 44 332-339. (/VI Phillips RL. Role of life-style and dietary habits in risk of cancer among Seventh-Day Adventists. Cancer Res 1975; 35:3513-3522. (/5) Mit ler AB. Kelly A. Choi NW. et at. A study or diet and breast cancer. Am J Epidemiol 1978, 107:499-509. (16) Lukin JH. Bt rns PE, Blot WJ, Ziecler RG, Lees AW, Fraumeni JF Jr. Dictarv factors and breast cancer risk. Int J Cancer 1981; 28:685-689. (17) Nomura A. Henderson BE. Lee J. Breast cancer and diet among the Japanese in Hawaii. Am J Clin Nutr 1978, 31:2Q20-2025. (!8) Bincham S. Williams DR, Cole TJ, James WP. Dietary fibre and regional large-bowel cancer mortality in Britain. Br J Cancer 1979; 40:456-463. (19) Lyon JL. Sorenson AW. Colon cancer in a low-risk population. Am J Clin Nutr 1978; 3LS227-S230. (20) Dales LG, Friedman GD, Ury HK, Grossman S. Wiluams SR. A case-control study of relationships of diet and other traits to colorectal cancer in American blacks. Am J Epidemiol 1978; 109:132-144. (21) Martinez I, Torres R, Frias Z, Colon JR, Fernandez N. Factors associated with adenocarcinomas of the large bowel in Puerto Rico. In: Birch JM. ed- Advances in medical oncology, research and education. Vol 3. Epidemiology. New York: Pergamon Press. 1979:45-52. (22) Jain St. Cook GM. Davis FG, Grace MG, Howe GR, Miller AB. A case-control studv of diet and colo-rectal cancer. Int J Cancer I960; 26:757-768. (23) Berc JW. Howell MA. The geographic pathology of bowel cancer. Cancer 1974; 34 807-814. (24) Howell MA. Diet as an etiological factor in the development of cancers of the colon and rectum. J Chronic Dis 1975; 28:67-80. (25) Haesszel W, Berc JW, Scci M. Kurihara M, Locke FB. Largebuwel cancer in Hawaiian Japanese. J Natl Cancer Inst 1973: 51:1765-1779. (26) Bjei.ke E. Dietary factors and the epidemiology of cancer of the stomach and large bowel. Aktuel Ernaehrungsmed Klin Prax 1978. 2 (suppl): 10- 17. (27) Graham S. Dayal H, Swanson M, Mittleman A, Wilkinson G. Diet in the epidemiology of cancer of the colon and rectum. JNCI 1978: 61:709-714. (28) Haf.nszel W, Locke FB, Seci M. A case-control study of large bowel cancer in Japan. JNCI 1980; 64:17-22, (29) MacLennan R. Jensen OM, Mosaics* J, Vuoat H. Diet, transit time, stool weight, and colon cancer in two Scandinavian popu lations. Am J Clin Nutr 1978. 31 :S239-S242. (30) Reudy BS, Hedges AR, Laakso K, Wvnder EL. Metabolic epide miology of large bowel cancer: Fecal bulk and constituents of high-risk North American and low-risk Finnish population. Can cer 1978; 42:2832-2838. (31) Bi kkitt DP, Trowei.l HC. Refined carbohydrate foods and disease. Some implications of dietary fibre. New York: Academic Press. 1975. (321 Ci user GA, Klein KL, Moore JO. Abba BC. Bowel transit-times in two populations experiencing similar colon-cancer risks. Lancet 1974, 2:80-81. (JJ) Closer GA. Nomura A, Kamiyama $, Shimaoa A, A asa BC Bowel tratuit-time and stool weight in populations with different colon- cancer risks. Lancet 1977; 2:110- 111. (JV) Howell MA. Factor analysis of international cancer mortality data and per eapita food consumption. Br J Cancer 1974; 29 328-336. (33) Blair A, Fraumeni JF Jr. Geographic patterns of prostate cancer in the United States. JNCI 1978, 61:1379-1384. (36) HtRAYAMA T. Changing patterns of cancer in Japan with special reference to the decrease in stomach cancer mortality In: Hiatt HU, Watson JD. Winsten JA. eds. Origins of human cancer. Book A: Incidence of oncer in humans. Cold Spring Harbor, N.Y : Cold Spring Harbor Laboratory, 1977:55-75. (37) Rutkin ID. Studies in the epidemiology of prostatic cancer: Ex panded sampling. Cancer Treat Rep 1977; 61; 173-180. (58} Schlman LM, Mangel JS, Raokc A, Seal U. Halserc F. Some selected features of the epidemiology of prostatic cancer: Minneapolis-St. Paul, Minnesota, case control study, 1976-1979 In: Magnus K. ed. Trends in cancer incidence: Causes and practical implications. Washington, DC.: Hemisphere Publishing. 1982:345-354. (39) Berc JW. Can nutrition explain the pattern of international epide miology of hormone-dependent cancers? Cancer Res 1975, 35:3345-3350. (40) Wvnder EL. Mabuchi K, Whitmore WF Jr. Epidemiology of cancer of the prostate. Cancer 19*1; 28:344-360. (V/) Lyon Jl_ Klauber MR. Gardner JW, Smart CR Cancer incidence in Mormons and non-Mormons in Utah, 1966-1970. N Engl J Med 1976: 294:129-133. (V2) In AJ. Neoplasms and environmental factors. Ann R Coll Surg Engl 1967: 41:432-438. (43) Linceman CM, Etiology of cancer of the human ovary: A review. J Natl Cancer Inst 1974:53.1603-1618. (^V) Miller AB. Bulbrook RD.The epidemiology and etiology of breast cancer. N Engl J Med 1960: 303; 1246-1248 (Vi) Kolonel LN. Hinds MW. Hankin JH. Cancer patterns among migrant and native-born Japanese in Hawaii in relation to smok ing. drinking, and dietary habits. In: Gelboin HV, MacMahon M, Matsushima T, Sugimura T. Takayama S. Takebe H. eds. Genetic and environmental factors in experimental and human cancer. Tokyo: Japan Scientific Societies Press. 1980:327-340. (46) Creoor O, Toman R, Prusova F. Gastrointestinal cancer and nutrition. Gut 1969; 10:1031-1034. (47) Rose G, Blackburn H, Keys A. et al. Colon cancer and bloodcholesterol. Lancet 1974; 1:181-183. (4d) William RR. Sorue PD, Funlxib M. McNamara PM. Kannel WB, Dawber TR. Cancer incidence by levels of cholesterol. JAMA 1981; 245:247-252. (V9) Karx JD. Smith AH, Hamks CC. The relationship of serum choles terol to the incidence of cancer in Evans Countv, Georgia. J Chronic Dis 1980: 33:311-322. (50) Garcia-Palmieri MR, Sorue PD, Costas R Jr. Havlik RJ. An apparent inverse relationship between serum cholesterol and can cer mortality in Puerto Rica. Am J Epidemiol 1981; 114:29-40. (51) Kacan A, McGee DL, Ya.no K, Rhoads GG, Nomura A. Scrum cholesterol and mortality in a Japanese-American population. Am J Epidemiol 1981; 114:11-20. (52) Cambien F. Ducimetif.re P. Richaid J. Total serum cholesterol and cancer mortality in a middle-aged population. Am J Epidemiol I960: 112:388-394. (55) Kozarevic D. McGee D, Vojvooic N, et al. Serum cholesterol and mortality. The Yugoslavia cardiovascular disease study. Am J Epidemiol 1981; 114:21-28. (54) Committee of Principle Investigator?. A co-operative trial in the primarv prevention of ischaemic heart disease using clofibrate. Br Heart J 1978; 40:1969-1118. (5.5) Pearce MI- Davtun S. Incidence of cancer in men on a diet high in polyunsaturated fat. Lancet 1971: 1:464-467. (56) Rose G, Shirley MJ Plasma lipids and mortality: A source of error. Lancet 1980; 1:523-526. (57) Stemmf.rmann GN. Nomura AM. Heilbrdn LK. Pollack US. Kacan A. Serum cholesterol and colon cancer incidence in Ha waiian Japanoe men. JNCI 1981; 67:1179-1182. (55) Miller SR, Tarttfji PI. Pamtfjtas AE, Slater G, Aufscs Al I Jr. Semm cholesterol and human colon cancer. JNCI 1981: 67:297- JNCI, VOL. 70, NO. 6, JUNE 1983 300. (59) Westli'.nd K. Nicolayjen R. Ten-year mortality and morbidity related to scrum cholesterol. Scand J Clin Lab Invest 1972; 27 (suppl 127): 1-24. (60) Ederer F. Lerem P, Tlrpeinen O, Frantz ID J*. Cancer among men on cholesterol'lowering diets: Experience from five clinical trials. I-ancet 1971; 2:203-206. (61) Dier AR, Stamler J. Paul O, et al. Serum cholesterol and risk of death from cancer and other causes in three Chicago epidemiolog ical studies. J Chronic Dts 1981; 34:249-260. (62) Stamler J. Berkson DM, Linrer HA. et al- Does hypercholesterol emia increase risk of lung cancer in cigarette smokers? Circulation 1968; 38 (suppl 6): 188. (tfj) Liu K, Stamler J, Moss D, Garside D, Peesky V, Soltero !. Dietary cholesterol, fat, and fibre, and colon-cancer mortality. Lancet 1979; 2:782-785. (64) Watson AF, Mcllanby . Tar cancer in mice; condition of skin when modified by external treatment or diet, as factors in influ encing cancerous reaction. Br J Exp Pathol 1930: 11:311-322. (65) Tannesbai m A. The genesis and growth of tumors. III. Effects of a high-fat diet. Cancer Res 1942: 2:468-475. (66) Tasnenbaum A, SiLvtRrro.NC H. Nutrition and the genesis of tu mours. In: Raven RW, ed. Cancer. Voi 1. London: Butterworth, 1957:306-334. (67) Lavik PS. Baumann CA. Further studies on tumor-promoting action of fat. Cancer Res 1943; 3:749-756. (63) Carroll KK. Kkor HT. EfTecis of level and type of dietary fat on incidence of mammary tumors induced in female Sprague-Dawley rats by 7,l2-dimethylbenz(a)amhracene. Lipids 1971; 6:413-420. (69) Hopkins GJ. West CE. Possible roles ofdietary fats in carcinogenesis. Life Sci 1976; 19:1103-1116- (70) Ip C. Ability of dietary fat to overcome the resistance of mature female rats to 7,12-dimethylbcnz()anthracene-induced mamman- tumorigenesis. Cancer Res 1980; 40:2785-2789. (71) King MM, Bailey DM. Gibson DD, Pitha JV. McCay PB. Inci dence and growth of mammary turnon induced by 7,12-dimcthylbenz(a [anthracene as related to the dietary content of fat and antioxidant. JNCI 1979; 63:657-663. (72) Ckan PC, Heao JF, Cohen LA. Wvnoer EL. Influence of dietary fat on the induction of mammary tumors by .V-nitrosomethylurca; Associated hormone changes and differences between SpragueDawlcy and F344 rats. J Natl Cancer Inst 1977; 59:1279-1283. (73) Silverman J. Sheilabarcsr CJ, Holtzman $, Stone JP, W'euburcer JI1. Effect of dietary fat on X-ray-induced mammary cancer in Sprague-Dawley rats. JNCI 1980; 64:631-634. (74) McCay PB. King M, Rika.ns LE. Pitha JV. Interactions between dietary fats and antioxidants on DMBA-induced mammary car cinomas and on AAF-induced hyperplastic nodules and hepato mas. J Environ Pathol Toxicol I960: 3:451-465. (75) Carroll KK. Lipids and carcinogenesis. J Environ Pathol Toxicol 1980; 3:253-271. (76) Carroll KK. Khor HT. Effects of dietary fat and dose level of 7,i2-dimeihylbenz(<namhracene on mammary tumor incidence in rats. Cancer Res 1970; 30:2260-2264. (77) Carroll KK. Hopkins GJ. Dietary polyunsaturated fat versus saturated fat in relation to mammary carcinogens!*. Lipids 1979; 14:155-138. (78) Hopkins GJ. Carroll KK. Relationship between amount and type of dietary fat in promotion of mammary carcinogenesis induced bv ?,l2-dimeihy!benz(e|amhraccne. J Nail Cancer Inst 1979; 62:1009-1012. (79) Hopkins GJ. West CE. Effect of dietary polyunsaturated fat on the growth of a transplantable adenocarcinoma in C3HA**fB mice. JNCI 1977; 58:753-756. (80) Abraham S. Rao CA. Lipids and lipogenesis in a murine mammary neoplastic system. In: Crixs WE. Ono T, Sabine JR. eds. Progress in cancer research and therapy. Vol l. Control mechanisms in cancer. New York: Raven Press. 1976:363-378. (81) Kiuwell WR, Monaco ME, Wicha MS, Smith GS. Unsaturatcd fatty acid requirements for growth and survival of a rat mammary tumor ceil line. Cancer Ra 1978; 38:4091-4100. (82) Wicha MS. Liotta LA. Kidwell WR. Effects of free fatty acids on the growth of normal and neoplastic rat mammary epithelial cells. Cancer Res 1979; 39:426-435. Diet, Nutrition, and Cancer 1165 (63) Redoy BS, Weisburcer JH, Wynder EL Effects of dietary fat level and dimethylhydrazine on fecal acid and neutral sterol excretion and colon carcinogenesis in rats. J Natl Cancer Inst 1974; 52:507511. (84) Reddy BS, Mancat S, Sheinpil A, Weisburcer JH, Wynder EL Effect of type and amount of dietary fat and 1,2-dimethylhydrazine on biliary bile acids, fecal bile acids, and neutral sterols in rats. Cancer Res 1977; 37:2132-2137. (65) Broitman SA. Vitale JJ. Vavrousek-Jakuba E, Gottlieb LS. Polyunsaturated fat, cholesterol and large bowel tumorigenesis. Cancer 1977; 40:2455-2463. (66) Reddy BS, WaTanasi K, Weisbubgkk JH. Effect of high-fat diet on eolon carcinogenesis in F344 rats created with 1,2'dimethylhydra* zinc, methylazoxvmethanol acetate, or meihyinitrosourea. Cancer Res 1977; 37:4156-4159. (67) Reody BS, Cohen LA. McCoy GD, Hill P, Weisburcer JH, Wynder EL Nutrition and its relationship to cancer. Adv Cancer Res 1980; 32:237-345. (66) Bull AW, Soullier BK, Wilson PS, Hayden MT. Nicro ND. Promotion of azoxymethane-induced intatinal cancer by high-fat diet in rats. Cancer Ra 1979; 39:4956-4959. (39) litll SiJ, Drasar BS, Aries V, Crowther JS. Hawksworth G, Williams RE. Bacteria and aetiology of cancer of large bowel. Lancet 1971; 1:95-100. (99) Mastromarino A, Redoy BS, Wynder EL Metabolic epidemiology of colon cancer: Enzymic activity of fecal flora. Am J Clin Nutr 1976. 29:1455-1460. (91) Narisawa T, Macaoia NE. Weisburcer JH, Wynder EL Promot ing effect of bile acids on colon carcinogenais after imrarectal instillation of ,V-methvl-.V*-nitro-V-nit rasoguanidine in rats. J Natl Cancer Inst 1974; 53:1093-1097. (92) Nicro ND, Bkadrachari N, Chomchai C. A rat model for studying coionic cancer Effect of cholestyramine on induced tumors. Dts Colon Rectum 1973; 16:438-443. (93) Redoy BS. Wynder EL Large-bowel carcinogenesis. Fecal constit uents of populations with diverse incidence rata of colon cancer. J Natl Cancer Inst 1973: 50:1437-1442. (94) Sucai M. Wrm.NO Lv Tsuchiyama H. Kummerow FA. The effect of heated fat on the carcinogenic activity of 2-acetvlaminofluorene. Cancer Re 1962; 22:510-519. (95) Newbekne PM. Zeicer E. Nutrition, carcinogensis. and mutagene sis. In; Flamm WC, Mchiman MA, eds. Advances in modern toxicology. Vol 5. Mutagenesis. New York: Wiley, 1978:53-84. (96) Loncnecker DS, Roebucr DB. Yacer JD Jr. Lilja HS. Siecmuno B. Pancreatic carcinoma in azaserine-treated rau: Induction, clas sification and dietary modulation of incidence. Cancer 1981; 47:1562-1572. (97) Roebuck BD. Yacer JD Jr, Loncnecker DS. Dietary modulation of azascrine-induced pancreatic carcinogenais in the rat. Cancer Ra 1961; 41:388-893. (96) Cruse JP, Lewin MR, Ferulano GP. Clark CG. Cocarcinagenic effects of dietarv cholesterol in experimental colon cancer. Nature 1978; 276:822-825. (99) Reody BS, Watanaie K. Effect of cholaicrol metabolites and promoting effect of lithocholie acid in colon carcinogenesis in germ-free and conventional F344 rats. Cancer Ra 1979; 39:15211524. (100) Graham S, Scmotz W, Martino P. Alimentary factors in the epi demiology of gastric cancer. Cancer 1972; 30:927-938. (lOt) Haenszzl W, Kurihara M. Segi M, Lee RK. Stomach cancer among Japanese in Hawaii. J Nail Cancer Inst 1972, 49:969-988. (102) Haenszel W, Kurihara M, Locke FB, Shimuzu K, Seci M. Stom ach cancer in Japan. J Natl Cancer Inst 1976, 56:265-274. (103) Bjelke E. Dietary vitamin A and human lung cancer. Int J Cancer 1975; 15:561-565. (104) MauLennan R. Da Costa J, Day NE, Law CH, Nu YK, Shanmucaratnam K. Risk factors for lung cancer in Singapore Chinese, a population with high female incidence rata- Int J Cancer 1977; 20:854-860. (105) Greuor A, Lee PN, Rot FJ, Wilson MJ, Melton A. Comparison of dietary histona in lung cancer cases and controls with special reference to vitamin A. Nutr Cancer 1980; 2:93-97. (106) Smith PG, Jicx H. Cancers among users of preparations containing vitamin A. Cancer 1978; 42:808-811. URL 03989 JNCI, VOL. 70, NO. 6, JUNE 1983 1166 Palmer and Bakshi URL 03990 (/O') Graham S, Mettlin C, Marshall J, Priorc R, Rzerka T, Skeoo D Dietary factors in the epidemiology of cancer of the larynx. AmJ Epidemiol 1081; 113:675-680. (JOS) Mettlin C. Graham S. Dietary risk factors in human bladder cancer. AmJ Epidemiol 1979; 110:255-263- (109) Wv nder EL, Bro&s IJ A study of etiological factors in cancer of the esophagus. Cancer 1961; 14:389-413. (110) Mettlin C. Graham S, Priore R. Marshall J, Swanson M. Diet and cancer of the esophagus. Nutr Cancer 1981, 2:143-147. (111) l loMMozoiARi 11, Day NE. Aramesm B, Mahboubi E. Dietary factors and esophageal cancer in the Caspian littoral of Iran. Cancer Res 1073, 35:3493-3498. (112) Joint Iran-International Agency for Research on Cancer Study Group. Esophageal cancer studies in the Caspian littoral of Iran: Results of population studies--a prodrome. J Natl Cancer Inst 1977;59:1127-1138. (1/3) Basi: TK, Donaldson D, Jenner M, Williams DC, Sakula A. Plasma vitamin A in patients with bronchial carcinoma. 8r J Cancer 1976, 16:297-298. (IN) Saklla A. Letter to the Editor: Vitamin A and lung cancer. Br Med J 1976; 2:298. (IIS) Wald M, Idle M, Boriham J, Bailey A. Low serum-vitamin A and subsequent risk of cancer--preliminary results of a prospective study. Lancet I960, 2:813-815. (//O') Mfinsma L. Nutrition and cancer. Voeding (The Hague) 1964: 25:357-365 (in Dutch: English summary). (1/7) (Iiluinwn J. Etiological factors in gastrointestinal cancer in man. J Nail Cancer Inst 1966; 37:527-545. (118) Haf.nseel W, Correa P. Developments in the epidemiology of stomach cancer over the past decade. Cancer Ra 1975; 35:34523459. (//9) Couk-Mozaffari PJ, Azordecan F, Dav NE, Ressicaud A, Sasai C. Aramesm B. Oesophageal cancer studies in the Caspian littoral of Iran. Results of a case-control study. Br J Cancer 1979: 39:293- 309 (/20) Cook-Mozaffar: PJ. The epidemiology ofcancer ofthe oesophagus. Nutr Cancer 1979; 1:51-60. (121) Wa**ertheil-S muller S, Romney SL. Wvlie-Rosett J. et al. Di etary vitamin C and uterine cervical dvsplasia. Am j F.pidemioi 1981; 114:714-724. (122) Correa P. 1 Lfnszel W, Cuello C. Tannenrai'm S, Archer M A model for gastric cancer epidemiology. Lancet 1975: 2:58-60. (123) Mai hutra SI. Dietary* factors in a study of cancer colon from Cancer Registry, with special reference to the role of saliva, milk and fermented milk products and vegetable fibre. Med Hypotheses 1977; 3:122-126. (124) Modas B. Barell V. Lubin F, Mooan M, Greenberg RA, Graham S. Low-fiber intake as an etiologic factor in cancer of the colon. J Nall Cancer Inst 1975:55:15-18. (125) W.vttenberu LW. Naturally occurring inhibitors of chemical carci nogenesis. In: Miller EC, Miller JA. Hirono I, Sugimura T. Takavama S, eds. Naturally occurring carcinngcns-mutagms and modulators of carcinogenesis. Tokyo and Baltimore: Japan Sci entific Societies Press and University Park Press, respectively. 1979.313-J29. (125) W.vttenberu LW, Lam LK., Speier JL, Loub WD, Burschert P. Inhibitors of chemical carcinogenesis. In: Hiatt it!!, Watson JD, Wiimen JA. eds. Origins of human cancer. Cold Spring Harbor, N Y : Col<l Spring Harbor Laboratory, 1977:785-800. (127) Wattf.nrf.ri; LW. Inhibition of carcinogenic effects of polycyclic hydrocarbons bv benzyl isothiocyanate and related compounds. J Natl Cancer Inst 197?`; 58:395-398. (128) Wattf.nbfru I.W, Lf.onc JL. Inhibition of the carcinogenic action of 7,l2-dimcthylbenz(e)anihracene by beta-naphthafiavone. Pmc Sue Exp Biol Med 1968; 128:940-943. (I2'J) Wattenberu LW. Lous WD. Inhibition of polycyclic aromatic hvdrucarbon-induced neoplasia by naturally occurring indoles. Cancer R.-i 1978; 38:1410-1413. (130) Thom. W. Blocking of tumor promtMion by protease inhibitors. In: Hurchenal JH, Oettgen UK. eds. Cancer: Achievements, chal lenges and prospect* for the 1980's. Vol 1. New York: Crune, 1981:349-553. (131) Cohf.n Bl, Raicht RK. Plant sterols: Protective role in chemical carcinogenesis. In: Zedeck MS, Lipkin M.eds. Inhibition of tumor induction and development. New York: Plenum Prc*i, 1981.189- 201. (132) Rvicht RK, Cohen Bl, Fazeini EP. Sarwai AN, Takam.vj.hi M. Protective effect of plant sterols against chemically induced colon tumors in rats. Cancer Res 1980, 40:403-405. (133) W vtteniierc. LW, Cocoa JB, Lam LK. Inhibitory effects of phenolic compounds on bcnzo(a)p\ renc-induced neoplasia. Cancer Res 1980: 40:2820-2823. (/34) WATTENnERC I.W. Inhibitors of chemical carcinogenesis. In: Bur- chenal J11, Oct t gen |IK, eds. Cancer: Achievements, challenges, and prospects for the 1980 s. Vol l. New York: Crane, 1981:31 7- 539. (135) N'ettesheim P. Williams ML. The influence of vitamin A on the susceptibility of the rat lung to 3-methytcholamhrene. Int J Can cer 1976; 17:351-357. (136) Cohen SM. Wittenberg JK, Bryan CT. Effect of avitaminosis A and hvpervitaminosis A on urinary bladder carcinogenicity of ,V- (4-(5-nitro-2-furyl)-2-ihiazol>l|rnrmamidc Cancer Res I9"t, 36:2334-2339. (137) Nf.wberne PM. Ruuers AE. Rat colon carcinomas associated with aflatoxin and marginal vitamin A. J Natl Cancer lax 1973: 50439-448. (138) Rogers AE. Herndon BJ. Newrerne PM. Induction by dimeihvl- hydrazine ofintestinal carcinoma in normal rats and rats fed high or low levels of vitamin A. Cancer Res 1973. 33:1003-1001*. (139) Narisawa T. Reddy BS. Wunis CQ. Weisim-riser JH. Efleet of vitamin A deficiency on rat colon carcinogenesis bv .V-meihyi-.V'- nitro-.Y-nitrosoguamdinc. Cancer Res 1976; 36:1379-1383. (140) Chu EW, Malmcren RA. An inhibitory effect of vitamin A on the induction of tumors of forestomach and cervix in the Syrian hamster bv carcinogenic poivcvciic hydrocarbons. Cancer Res 1965: 23:884-895. (141) Shambercer RJ. Inhibitory efTect of vitamin A on carcinogenesis. J Natl Cancer Inst 1971:47:667-673. (142) Sapfiutti U. Muntemno R. Sellakimar AR. Burg 5A Expert- mental cancer of the lung: Inhibition bv vitamin A of the induc tion of tracheobronchial squamous metaplasia and squamous cell tumors. Cancer 1967; 20:837-864. (143) Shorn MB. Newton DL. Chcmoprcvenuon of cancer with retinoids. (144) Fed Proc 1979; 38:2328-2534. . Recent advances in ihc use of retinoids for cancer prevention. In: Burchcnal JH. Oettgen 11K. eds. Cancer: Achievements, chal lenges and prospects for the 1980 s. Vol 1. New York: Grunc. 1981:541-548. (/4J) Shorn MB. Dunlop NM. Newton DL Smith JM. Prevention of chemical carcinogenesis by vitamin A and its synthetic analogs (retinoidsl. Fed Proc 1976: 35:1332-1338. (146) Bollau W. Therapy of chemically induced skin tumors uf mice with vitamin A pahnitate and vitamin A acid. F-xperieniia 1971. 27.`H)- 92. (147) Smith DM, Rogers AE. Hf.rnixjn BJ. N'f.wbek.vf PM Vitamin A (retinvl acctatcl and benzotaIpyrene-induced respiratory tract carcinogenesis in hamsters fed a commercial diet. Cancer Res 1973; 33:11-16. (148) Wf.uch CW. Goodrich-Smith M. Brown CK. Crow e N. Enhance ment by retinvl acetate uf homiune-induced numnun luiuori- gcncsis in female GR/A mice. JNCI 1981: 67.935-938. (149) Lasnittki I. Growth pattern uf the mouse prostate gland in organ culiurc and its response to sex hormones, vitamin A. and Jmethylcholanthrene. Natl Cancer Inst Munogr 1963; 12:381 -40.1. (150) Crixjkf.r TT. Sanuer* LL Influence of v itamin A jnd 3,7-dimethvl- 2.6xKtadienal (citral) on ihc elTcvt of benzoin ipyrene on hamster trachea in organ culture. Cancer Ra 1970: JO: 1312 -1318. U5I) Mathews-Roth MM. Pathak MA. Fitzpatrick TB. Harder I.II. K.uvs EH. Beta carotene thcrapv for ervthrojxiirtic protnporphv ria and other photosensitivity diseases. Arch Dermatol 1977. 113:1229-1232. (152) Ivankovic S. pRF.fs.sviANN R. Sgiimahi. D. Zei.lf.r JW. Prevention hv .Lscorhic acid of in vivo formation of .Y-niinmo compound*. In: Bogovxki P. Walker F.A. eds. .V-Niiroso compounds in the environ ment. lARCSci Publ 1975; 9:101-102. (153) Mirvish SS. Inhibition of the formation of carcinogenic N-nitroso compounds by ascorbic acid and other compounds. In: Burchcnal JH. Oettgen UF, eds. Cancer: Achievements, challenges, and JNCI, VOL. 7f), NO. 6, JUNE 1983 Diet, Nutrition, and Cancer 1167 URL 03991 pruvpect* for the 1980V Vot l. New York; Grune, 1981:557-587. (154) Mnvi'H SS. Wallcave L, Eagen M, Shubik P. Ascorbate-nitrite reaction: Possible means of blocking ihe format ion or carcinogenic N-nitroso compounds. Science 1972; 177:65-68. (755) Mirvish SS, Carocsa A. Wallcave L, Shubik P. Induction'of mouve lung adenomas by amines or ureas plus nitrite and by Nnitroso compounds: Effect of ascorbate, gallic acid, thiocyanate, and caffeine. J Nall Cancer Iny 1975, 55:633-636. (156) Rustia SI. Inhibitory effect of sodium ascorbate on ethylurea and sodium nitrite carcinogcnsis and negative findings in progeny after intestinal inoculation of precursors into pregnant hamsters, J Natl Cancer Tost 1975, 55:1389-1394. (757) Locvs T, Frommbi D. The influence of oral vitamin C supplements on experimental colorectal rumour induction. Aust N2 J Med I960: 10:588. (158) Reddy BS, Hirota N. EfTeet of dietary ascorbic acid on 1,2-dimethvlhydrazine-induced colon cancer in rats. Fed Proc 1979; 38.714. (159) Kallistratos G. Fasske E. Inhibition of benzo{a)pvrene carcino genesis in rats with vitamin C. J Cancer Res Clin Oncol I960; 97:91-96. (160) Soloway MS, Cohen SM, Oekernion JB. Persxv L. Failure of ascorbic acid to inhibit FAN'FT-induced bladder cancer. J Urol 1975. 113:483-486. (/$/} Basic S. Vitamin C acts as a cocarcinogen to methylcholanthrene in guinea-pigs. Cancer Lett 1981; 11:239-242. (162) Russell WO, Ortcca LR, Wynne ES. Studies on methylcholan threne induction of tumors in scorbutic guinea pip. Cancer Rea 1952: I2:2l&-216. (763) Benedict WF, Wheatley WL, Jones PA. Inhibition of chemically induced morphological transformation and reversion of the trans formed phenotvpe bv ascorbic acid in C3H/10T-1/2 cells. Cancer Res i960 : 40:2796-2801. (764) Barbolt TA, Abraham R. The effect of bran on dimethylhydrazine- induced colon carcinogenesis in the rat. Proc Soc Exp Biol Med 1978: 157:636-659. (165) Che.y WF. Patchefsky AS, Goldsmith HS. Coionic protection from dimethvlhvdruine by a high fiber diet. Surg Gynecol Ohstet 1978; 147:503-506. (166) Wilson RB. Hutcheson DP, Wiocman L. Dimethylhydrazine-induced colon tumors in rats fed diets containing beef fat or corn oil with and without wheat bran. Am J Clin Nutr 1977; 30:176-181. (767) Freeman HJ. Sfiller GA, Kim YS. A double-blind study on the effect of purified cellulose dietary fiber on 1,2-dimethylhydrazineinduced rat colonic neoplasia. Cancer Res 1978; 38. 2912-2917. (768) --. A double-blind study on the effects of differing purified cellulose and pectin fiber diets on 1,2-dimethylhydrazine-induced rat colonic neoplasia. Cancer Res 1980; 40:2661-2665. (169) Ward JM. Yamamoto RS, Weisburger JH. Briefcommunication: Cel lulose dietary bulk and azoxymcthane-induced intestinal cancer. J N'atl Cancer Inst 1973:51:713-715. (770) Watanabe K. Reddy BS, Wong CQ, Weisbubcee JH. Effect of dietarv undegraded carrageenan on colon carcinogenesis in F344 rats treated with azoxymethane or methylnitrosourea. Cancer Res 1978; 38:4427-4430. (777) Glauert HP. Bennink MR. Sander CH. Enhancement of 1,2dimethylhydrazinc-induced colon carcinogenesis in mice by di etarv agar. Food Cosmet Toxical 1981. 19:281-286. (!72) Coordinating Group for Research on Etiology of Esophageal Cancer in North China. The epidemiology and etiology of esophageal cancer in North China. A preliminary report. Chin Med J {Engl) 1975: 1:167-183. (77J) Yang CS Research on esophageal cancer in China: A review Cancer Res 1980; 40:2633-2644. (77-7) Gabor II. A "Vcndvidckcn" eluforduio gyomorrik a a Kiiionleges fustolcsi mod <Mszcfuggeae. Medicus Universalis 1974; 7:287-289 (in Magyar). (175) Sous K. The occurrence of carcinogenic polycyclic hydrocarbons in foodstuffs in Hungary. Arch Toxicol 1980; suppl 4:446-448. (176) Duncai. N. The special problem of stomach cancer in Iceland. With particular reference to dietary factors. JAMA 1961; 178:789-798. (177) Voitaluvich F.A. Df.ekoqn PP, Deemarsky LU, Sharad LM. Com parative studv of malignant tumor frequency in Tookoom District of thr Ijtvian SSR. Vopr Onkol 1947; 3:351-357 (in Russian; English summary). (178) Choi NW. Entwistle DW. Michaluk W, Nelson N. Gastric cancer in Icelanders in Manitoba. Isr J Med Sci 1971; 7:1500-1508. (779) Kriebel D. Juwett D. Stomach cancer mortality in the North Central States: High risk is not limited to the foreign-bom. Nuir Cancer 1979; 1:8-12. (780) lltRAVAMA T. The epidemiology of cancer of the stomach in Japan with special reference to the role of diet. In: Hams RJ, ed. Proceedings of the ninth international cancer congress. Union Internationale Contre le Cancer Monograph 10. Berlin: SpringerVerlag, 1967:37-48. (18/) Joossens JV, Geboers J. Nutrition and gastric cancer. Proc Nutr Soc 1981: 40:37-46. (782) VS'ejsiurcer JH. Reddy BS, Fiala ES, et al. Dietary factors in the causation and prevention of neoplasia. In: Burchenal J, Oetigcn H, eds. Cancer: Achievements, challenges, and prospects for the 1980's. Vol l. New York: Crone, 1981:595-612. (183) Okamura T, MAnvursA F. Fluorine and other related components ofJapanese foods. V. Fluorine and salt contents of miso and other foods rich in salt, and their geographical correlation with mortality from gastric cancer. Eiyo To Shokuryo 1965; 18:253-257 (in Japanese; English summary). (184) Magee PN, Barnes JM. Carcinogenic nitroso compounds. Adv Cancer Res 1967; 10:163-246 (/85) Magee PN, Montesano R, Preussman R. N-nitroso compounds and related carcinogens. In: Scarle CE. ed. Chemical carcinogens. American Chemical Society Monograph 173. Washington, D.C.: Am Chcm Soc, l976:49t-625. (186) National Academy of Science. The health efreeu of nitrate, nitrite, and Mniirao compounds. Part I of a 2-part study by the Com mittee on Nitrite and Alternative Curing Agents in Food. Assem bly of Life Sciences. Washington, D C: National Acadcmv Press, 1961. (187) Armijo R. Coulson AH. Epidemiology of stomach cancer in Chile-- the role of nitrogen fertilizers. Ini J Epidemiol 1975; 4:301 -309. (188) Armijo R, Gonzalez A. Orellana M. Coulson AH. Sayre JW, Dctels R. Epidemiology of gastric cancer in Chile. U. Nitrate exposures and stomach cancer frequency. Int J Epidemiol 1981: 10:57-62. (789) Cuellq C, Correa P. Hacnszel W, et al. Gastric cancer in Col ombia. I. Cancer risk and suspect environmental agents. J Natl Cancer Inst 1976; 57:1015-1020. (790) Weisburger JH. Rainerj R. Dietary factors and the etiology of gastric cancer. Cancer Res 1975; 35:3469-3474. (79/) Ruodell W5. Bone S, Hill MJ, Walters CL. Pathogenesis of gastric cancer in pernicious anaemia. Lancet 1978; 1:521-523. (792) Broitman SA, Velez H, Vitale JJ. A possible role of iron deficiency in gastric cancer in Colombia. Adv Exp Med Biol 1961; 135:155181. (793) Ohshima H. Bartsck H. Quantitative estimation of endogenous nitresation in humans by monitoring .V-nitrosoproline excreted in the urine. Cancer Res 1981; 41:3658-3662. (194) Gheenblatt M, Mirvuh SS. Dose-response studies with concurrent administration of piperazine and sodium nitrite to strain A mice. J Natl Cancer Inst 1972; 50:119-124. (795) Lijinsky W, GreenblaTT M, Kommineni C. BritfevnmimuaUM: Feed ing studies of nitrilotriacetic acid and derivatives in rats. J Natl Cancer Inst 1973; 50:1061-1063. (796) Suciyama K, Tanaka T, Mori H. Carcinogenicity examination of sodium nitrate in mice. Gifu Daigaku tgakubu Kiyo 1979: 27:1-6 (in Japanoe: English summary). (797) U.S. Fond and Drug Administration. Re-evaluation of the pathology findings of studies on nitrite and cancer Histologic lesions in Spraguc-Dawlcy rats. Final report, submitted by Universities Associated for Research and Education in Pathology, Inc. Wash ington. D C.: FDA, DHHS. i960. (/98) Greenriatt M. Kommineni VR. Ljjinsky W, Briff communication: Null efTcct of concurrent feeding of sodium nitrite and amino acids to MRC rats. J Natl Cancer Inst 1973; 50:799-802. (799) Lijinsky W, Reuber MD, Manning WB. Potent carcinogenicity of nitroaodiethanolamine in rats. Nature 1981}; 288:589-590. (200) NewnF.RNE PM. Nitrite promotes lymphoma incidence in rats. Sci ence 1979; 204:1079-1081. (201) Shank RC, Newbf.rnr PM. Do-response study of the carcinogenic- JNCI, VOL. 70, NO. 6, JUNE 1983 1168 Palmer and Bakshi ity of dietary sodium nitrite and morpholine in rats and hamsters. Fond Cosmet Toxicol 1976; 14:1-8. (202) I low vrd JW, Fazio T. Review of polycyclic aromatic hydrocarbons in funds. Analytical methodology and reported findings of poly cyclic aromatic hydrocarbons in foods. J Assoc OfT Anal Chert) 1980; t>3:1077-11 14. (203) Lo MT. Sanih E. Polycyclic aromatic hydrocarbons (polynuclears) in foods. Residue Rev 1978; 69:35-86. (204) Riuuov RU. Neal J. Relationship oflcukemia to lung and stomach tumors in mice fed bcnzo(a)pvrenc. Proc Soc Exp Biol Med (969; 130:146-148. (205) Huggins C. Yang NC. Induction and extinction of mammary cancer. Science 1962: 137:257-262. (206) Chu EW, Malmcren RA. An inhibitory effect of vitamin A on the induction of turnon of forcstomach and cervix in the Syrian hamster by carcinogenic polycyclic hydrocarbons. Cancer Ret 1965: 25:884-895. (207) Nauao M. Sugimura T. Mutagenesis: Microbial systems. In: Celboin HV. TVo PQ, eds. Polycyclic hydrocarbons and cancer. Vol 2. Molecular and cell biology. New York: Academic Press, 1978:99- 121. (203) Tonc C. Laspia MF, Telanc S, Williams CM. The use of adult rat liver cultures in the detection of the genotoxicitv of various polycyclic aromatic hydrocarbons. Environ Mutagen 1981; 3:477487. (209) Biascihori C, Caschexa F. The relation between pseudopregnancy and the chemical induction by four carcinogens of mammary and ovarian tumours in BALB/c mice. Br JCancer 1962; 16:722-730. (2/0) McCann J, Choi E, Yamasaki E. Ames BN. Detection of carcinogens as mutagens in the Salmane/la/cn'icrotame test: Assav of 300 chem icals. Proc Natl Acad Sci USA 1975; 72:5135-5139. (2!!) Klein M. Susceptibility of strain B6AF1/J hybrid infant mice to tumorigenesis with 1,2-benzanihracene, dcoxycholic acid, and 3methylcholanthrene. Cancer Res 1963; 23:1701-1707. (212) Wocan CN. Aflatoxin carcinogenesis. In: Busch H, ed Methods in cancer research. Vol 7. New York: Academic Proa, 1973:309-344. (2/3) VESseLiNovrrcH SD, Mihailovich N, Wocan CN, Lombard LS. Rao KV. Aflatoxin Bi, a hepatocareinogen in the infant mouse. Cancer Res 1972; 32:2289-2291. (214) Uemo Y. Kubota K, Ito T, Nakamura Y. Mutagenicity of carci nogenic mycotoxins in Salmonella tjphimnnum. Cancer Res 1978; 38:536-342. (2/5) Umeoa M, Tsimui T, Saito M. Mutagenicity and inducibility of DNA single-strand breaks and chromosome aberrations by various mvcotoxins. Gan 1977; 66:619-625. (2/6) International Agency for Research on Cancer. Some naturally oc curring substances. IARC Monogr Eval Carcinog Risk Chem Man. 1976: 10:1-353. (2/7) Linsell CA, Peers FG. Aflatoxin and liver cell cancer. Trans R Soc Trop Med Hvg 1977; 71:471-473. (2/8) Van Rensburc SJ, van oer Watt JJ, Purchase IF, Coutinho LP, Markham R. Primary liver cancer rate and aflatoxin intake in a high cancer area. S Afr Med J 1974; 46:2508A-2S08D. (2/9) Toth B. Smith JW. Patil KO. Cancer induction in mice with acetaldehvde meth vI formvlh vdrazone of the false morel mush room. JNCi 1981,67 881-887. (220) Rgcan EG. Walker BA. Cingell R. Nacel DL. Toth B. Microbial mutagenicity of selected hydraiines. Mutat Res 1982; 102:41342*. (22/) Toth B, Patil K. Carcinogenic effects in the Syrian golden hamster of .Y-methyl-.Y-formylhydrazine of the false morel mushroom Gy romitta ticultnia.J Cancer Res Clin Oncol 1979; 93:109-121. (222) Hirono I. Mori H, Haca M, et al. Edible plants containing carci nogenic pyrrolizidine alkaloids in Japan. In: Miller EC, Miller JA, Hirono I. Sugimura T, Takayama S, eds. Naturally occurring carcinogens--mutagens and modulators of carcinogenesis. Tokyo and Baltimore: Japan Scientific Societies Press and University Park Press, respectively. 1979:79-87. (22J) Matsushima T. Matsumoto H. Shirai A, Sawamura M. Sugimura T. Mutagenicity of the naturally occurring carcinogen cycasin and synthetic meihylazoxvmethanol conjugates in Salmonella tt- p/umunum. Cancer Res 1979; 39:3780-3782. (224) IIiravama T. F.pidemiolngical evaluation of the role of naturally occurring carcinogens and modulators of carcinogenesis. In: Miller EC, Miller JA, Hirono I, Sugimura T, Takayama S, eds. Naturally occurring carcinogens--mutagens and modulators of carcinogen esis. Tokyo and Baltimore: Japan Scientific Societies Press and University Park Press, respectively. 1979:359-380. (225) Howe GR. Bi rch JD. Miller AB, ei al. Tobacco use, occupation, coffee, various nutrients, and bladder cancer. JNCI 1980, 64:701- 713. (226) Fraumeni JF Jr. ScottoJ. Dunham LJ Letter to the Editor: Coffee drinking and bladder cancer. Lancet 1971; 2:1204. (227) Miller CT.Neutbl Cl. Nai* RC, MarrettLD, Last JM, Collins WE. Relative importance ofrisk factors ia bladder carcinogenesis. J Chronic Ois 1978; 31:51-56. (228) Wi nder EL. Colosmfth R. The epidemiology of bladder cancer. A second look. Cancer 1977; 40:1246-1268. (229) MacMahon B, Yen S, Trichopoulos D. Warren K, Naroi G. Coffee and cancer of the pancreas. N Engl J Med 1981; 304:630- 633. (230) Lin RS, Kessler II. A muEiifactoria! model for pancreatic cancer in man: Epidemiological evidence. JAMA 1981; 245:147-152(23/) Shcnnan Dll. Letter 10 the Editor: Renal carcinoma and coffee consumption in 16 countries. Br J Cancer 1973; 28:473-474. (222) Martinez I. Factors associated with cancer of the esophagus, mouth, and pharynx in Puerto Rico. J Natl Cancer Inst 1969; 42:10691094. (233) Stocks P. Cancer mortality in relation to national consumption of cigarettes, solid fuel, tea and coffee. Br J Cancer 1970; 24:215225. (234) Armstrong B, Garrod A, Doll R. A retrospective study of renal cancer with special reference to coffee and animal protein con sumption. Br J Cancer 1976; 33:127-136. (235) Aeschbacher HU, Wubzmbb HP. An evaluation of instant and tegular coffee in the Ames mutagenicity test. Toxicol Leu 19811: 5:139-143. (236) Aeschbacher HU. Cmappuis C, Wurzner HP. Mutagenicity toting of coffee: A study of problems encountered with the Ames Salmo nella test system. Food Costner Toxicol I960; 18:605-613. (237) Nacao M. Takahasmi Y. Wakabavashi K. Sucimlra T. Mutagens in coffee and tea. Mutat Res 1979; 68:101-106. (238) ZtrrttN BR. Letter to the Editor: Coffee and bladder cancer. Lancet 1972: 1:1066. (239) Wuizner HA, Lindstrom E, Vuatai L, Lucinbuhl H. A 2-year feeding study of instani coffee in rats. II. Incidence and types of neoplasms. Food Cosmet Toxicol 1977; 15:289-296. (240) Roberts HR. Food safety in perspective. In: Roberts HR, ed. Food safety. New York: Wiley, 1981:1-13. (24/) Food and Drug Administration. U.S. Department of Health and Human Servics. Food and Drug Administration Acts. Federal Food, Drug, and Cosmetic Act, as amended January 1980: Public Health Service Act. Biological Products; Radiation Control for Health and Safetv Act; Fair Packaging and Labeling Act. Rock ville. Md: FDA. DHHS, 1980 ([DHHS publication No. <FDA)801051}. (242) International Agency for Research on Cancer. Some non-nutritive sweetening agents. IARC Monogr Eval Carcinog Risk Chcm Humans. I960: 22:171-185. (243) Feron VJ, Henmiksen CF, Speck AJ, Til HP. Spit BJ. Lifespan oral toxicitv study of vinvl chloride in rau. Food Cosmet Toxicol 1981; 19:317-333. (244) Norris JM. Status report on the 2 year study incorporating acrylo nitrile in the drinking water of rats. Health and environmental research. Midland, Mich: Dow Chemical Co., 1977. (243) Tabershaw IR, Gapfev WR. Mortality study of workers in the manufacture of vinyl chluride and its polymers. J Occup Med 1974; 16:509-518. (246) Rusu M. Zjmmerli B, Marek B. Ruckstandc von VinylchloridMonomer in Speiseolen. Mitt Gris Lebensmiirelumers Hyg 1975; 66:507-511 (in German; English summary). (247) U.S. Environmental Protection Agency. Preliminary assessment of suspected carcinogens in drinking water. Report to Congress. EPA-560/4-75-005. PB-250961. Washington, DC.: Office ofToxic Substances, El'A, 1975. (248) U.S. Food and Drug Administration, U.S. Department of Health. URL 03992 JNCI, VOL. 70, SO. 6, JUNE 1983 Diet, Nutrition, and Cancer 1169 Education, and Welfare FDA Compliance Program Report of Findings. FY 77, Total Diet Studies--adult (7320.73). Washing ton. DC.: FDA, DHF.W. 1980. (249) McCann J. Avies BN. Detection of carcinogens as mutagens in the Sd/snarir/fa/niicrusomc test: Assay of 300 chemicals: Discussion. Proc Natl Acad Sci USA 1976; 73:930-93-1. C-7Jf7) Purchase IF, Longstaff E, Ashby J, et al. An evaluation of 6 short- term tests for detecting organic chemical carcinogens. Br J Cancer 197B; 37:873-939. (251) Simmon V'F. In vitro mutagenicity assays of chemical carcinogens and related compounds with Salmonella tj/pkimunam. JNCi 1979, 62:893-899. (232) Sucimura T. S*to S. Nagao M. et al. Overlapping of carcinogens and mutagens. In: Magee PN, Takayama S. Sugimura T, Mat- suhima T, eds. Fundamentals in cancer prevention. Batiimore: University Park Press, 1976:191-213. (2JJ) Brown JP. A review of the genetic effects of naturally occurring flavonoids, amhraquinoncs and related compounds. Mutat Res 1980: 73:243-277. (254) Bjei.oanes I.F, Chang GW. Mutagenic activity of quercetin and related compounds. Science 1977; 197:377-378. (255) Hardigree AA, Epler Jl.. Comparative mutagenesis of plant fla- vonoids in microbial systems. Mutat Res 1978; 38:231-239. 2( jo) MacGrecor JT, Jurd L. Mutagenicity of plant flavonoids: Struc tural requirements for mutagenic activity in Salmonella yphimurium. Mutat Rr> 1978: 34:297-309. (257) Uyeta M, Tuue S, Mazaxi M. Mutagenicity of hydrolysates of tea infusions. Mutat Res 1981: 88:233-240 (253) Takahashi Y, Nagao M, Fujrso T, Yamauumi Z, Sucimura T. Mutagens in Japanese pickle identified as flavonoids. Mutat Res 1979: 68:117-123. ^ (259) Fukuoka M, Yoshihira K, Natori S, et al. Characterization of mutagenic principles and carcinogenicity tat of dill weed and seeds. J Pharmacobiodyn 1980; 3:236-244. (260) Stoltz DR. Stavric B, Klassen R. Muise T. The health significance of mutagens in foods. In: Stitch HF. ed. Carcinogens and mutagens in the environment. Yo) 1. Food products. Boca Raion, Fla: CRC Press 1982:185-197. (261) Pamuckcu AM. Yal(inee S, Hatcher JF, Brva.n GT. Quercetin, a rat intstinai and bladder carcinogen present in bracken fern (Plertdium aqutltnum) Cancer Res I960; 40:3468-3472. (262) Satto D. Shirai A. Matsushima T, Hirono I. Tot of carcinogenicity of quercetin, a widely distributed mutagen in food. Teratogenesis Carcmog Mutagen 1960; l:2l3-22l. (263) Hirono I, Ue.no (, FIosaka S, et al. Carcinogenicity examination of quercetin and rutin in ACI rats. Cancer Lett 1981; 13:15-21. (264) Morino K. Matsukura N, Kawachi T, Ohcaki H, Sugimura T. Hiro.no 1. Carcinogenicity test of quercetin and rutin in golden hamsters bv oral administration. Carcinogenesis 1982: 3:93-97. (265) Clarke CH. Wade MJ. Evidence that caffeine, 3-mechoxypsoralen and steroidal diamines are frameshift mutagens for E. toii K-12. Mutat Res 1973: 28123-123. (266) Tamcra G, Gold C, Ferro-Luzzi A. Ames BN. Fecalasc: A model for activation of dietary glycosides to mutagens bv intestinal flora. Proc Natl Acad Sci USA 1980; 77:4961-4965. (26?) Lee JS. Fonc LY. Mutagenicity of Chinoe alcoholic spirits. Food Cosmet Toxicol 1979; 17:575-578. (263) Nagao M. Takahashi Y, Wakaravashi K, Sucimura T. Mutage nicity of alcoholic beverages. Mutat Res 1981; 88:147-154. (269) Lo^urr C, Toussaint G, LeTalaer JY. Studies on mutagenic constituents of apple brandy and various alcoholic beverages collected in Western France, a high incidence area for oesophageal cancer Mutat Res 1981; 88:155-164. (270) Lijivsky W, Shlbik P Benzolsipyrene and other polynuclear hy drocarbon* in charcoal-broiled meat. Science 1964. 145:53-35. (271) Nagao M. Yahagi T, Kawachi T, et al. Mutagens in foods, and especially pvrolvxi* products of protein. In: Scott D. Bridges BA, Nobels FI I, eds. Progress in genetic toxicology. New York; Elsevier/ Nnrih-I toiland. 1977.259-264. (272) Matsl-muto T. Yoshuia D, Mieusaki S, Okamoto tl. Mutagenic activiiv of amino acid pvrolvzaies in Salmonella yphimuriitm TA 98. Mutat Res 1977; 48:279-286. (273) Sugimura T. Nagao M. The use of mutagenicity to evaluate carci nogenic hazards in our daily lives. In: Heddle JA, ed. Mujage- nicitv: New horizons in genetic toxicology. New York: Academic Press. 1982:73-83. (274) K,Matsukura N, Kawachi T, Morino Ohcaki II. Sucimura T. Takayama S. Carcinogenicity in mice of mutagenic compounds from a tryptophan pyrolvzaie. Science 1981; 213:346-347. (275) Sucimi'ea T. Mutagens in cooked food*. In: Fleck R. Hollaender A, eds. Genetic toxicology: An agricultural perspective. New York: Plenum Pros, 1982:243-270. (276) Tamuchi T, Ocawa K, Iimuma H. et al. Monoamine oxidase inhibitors isolated from fermented broths. J Antibiot (Tokyo) 1973;26:162-167. (277) Yasuda T, Yamaizumi 2. Nishimura S, et al. Detection of comuta- genic compounds, harman and norharman in pyrolysis product of proteins and food by gas chromatograph-mass spectrometry Nip pon Gan Cakkai Sokai Kiji 1978; 37:6. (278) Nacao M, Yahagi T. Sucimura T, et al. Comucagenic action of norharman and harman. Proc Jpn Acad 1977; 53:95-98. (279) Nacao M, Yahagi T. Honda M, Seino Y, Matsushima T. Sucimura T. Demonstration of mutagenicity of aniline and e-toluidine by norharman. Proc Jpn Acad 1977; 53:34-37. (280) Sugimura T, Nacao M. WakaBayashi K. The metabolic aspects of the eomutagenic action of norharman. In: Snyder R. Parke D, Kocsis JJ. Jollow DJ. Gibson CG, Witmer CN. eds. Biological reactive intermediata 2. Pan B. New York: Plenum Press, 1982:1011-1025. (281) Morita K. Hara M. Kaoa T. Studies on natural dsmutagens: Screening for vegetable and fruit factors active in inactivation of mutagenic pyrolysis products from amino acids. Agric Biol Chern 1978; 42:1235-1238. (282) Lai CN, Butler MA. Matnrv TS. Antimutagenic activities of common vegetables and their chlorophyll content. Mutat Res 1980; 77:245-250. (283) Lai CN. Chlorophyll: The active factor in wheat sprout extract inhibiting the metabolic activation of carcinogens in vitro. Nutr Cancer 1979; 1:19-21. (284) McKee Rll. Tometsko AM. Inhibition of promutagen activation by the antioxidants butvlated hydroxvanisoie and butylated hv- droxytoluene. JNCI 1979.63:473-477. (285) Katoh Y, Tanaea M. Umezawa K. Takayama S. Inhibition of mutagenesis in Chinese hamster Y-79 cells bv antioxidants. Tox icol Lett 1980: 7:125-130. (266) Skamrercer RJ, Coruett CL, Beaman KD, Kastzn BL. Antioxi dants reduce the mutagenic effect of malonaldehyde and -pro- piolactone. Part IX. Antioxidants and cancer. Mutat Res 1979; 66:349-353. (287) Guttenflan JB. Mechanisms of inhibition by ascorbate of microbial mutagenesis induced bv .Y-nitroso compounds. Cancer Res 1978: 38:2018-2022. (288) Kono S. Ikeoa M. Correlation between cancer mortality and alco holic beverage in Japan. Br J Cancer 1979; 40:449-455. (289) Cook P. Cancer of the oesophagus in Africa: A summary and evaluation of the evidence for the frequency of occurrence, and a prelimmarv indication of the possible association with the con sumption of alcoholic drinks made from maize. Br J Cancer 1971: 25:853-880. (290) Collis CM. Cook PJ. Foreman JK. Palframan JF. Letter to the Editor: Cancer of the oesophagus and alcoholic drinks in Ea-vt Africa. Lancet 1972; 1:441. (291) Tutyns AJ. Pequicnot G, AaaATUCCt JS. Oesophageal cancer and alcohol consumption; importance of type of beverage. Int J Cancer 1979; 23:443-447. (292) Lamy L Etude de statixtique clinique de 134 cas de cancer de I'oesophage du cardia. Arch Mai Appar Dig Mai Nutr 1910; 4:451-475 fin French). (291) Kwan KW. Carcinoma of the esophagus; statistical studv. Chin Med J 1937;52:237-254. (294) Wu YK. Lugcks 1111. Carcinoma of the esophagus or cardia of the stomach: An analysis of 172 cases with 81 resections. Ann .Surg 1951; 134:946-956. (295) Murv J, Muntvrrnay C, Lambert R. Wine and tobacco: Rivk factors for gastric cancer in France. Am J Epidemiol 1981: 113:668-674. URL 03993 JNCI, VOL. 70, NO. 6, JL'NE 1983 1170 Palmer and Bakshi (296) 8rrslow NF., E.nstrox* JE. Geographic correlations between cancer mortality rata and alcohol-tobacco consumption in the United States. J Natl Cancer Inst 1974, 53:6.31-639. (297) EnstromJE. Colorectal cancer and beer drinking. Br J Cancer 1977; 33:G74-6ti3. (298) Vital* JJ. Broitman SA, Gottlie* LS. Alcohol and carcinogenesis. In: Newell OR, Ellison NM, eds. Nutrition and cancer; Etiology and treatment. New York: Raven Press, 1961:391-301. (299) Stocks P. Cancer incidence in North Wales and Liverpool region in relation to habits and environment. British Empire cancer cam paign thirty-fifth annual report. Supplement to pan 11. London: Cancer Research Campaign. 1957:1-156. (300) Dean G, MacLc.nnan R, McLoucmlin H, Shelley E. Causa of death of blue-collar workers at a Dublin brewerv, 1954-1973. Br J Cancer 1979; 40:581-589. (301) Pernu J. An epidemiological study on cancer of the digestive organs and respiratory system. A study based on 7076 cases. Ann Sled lntem Fenn I960; 49 (suppl 33)1-117. (302) Bjelke E. Case-control study of cancer of the stomach, colon, and rectum. In: Clark RL, Cumley RC. McCay JE. Copeland MM, cds. Oncology 1970, Voi 5. Chicago: Yearbook Medical Publ, 1971.320-334. (303) Hinds MW, Kolonel LN, LeeJ, Hirohata T. Associatiorubetween cancer incidence and alcohol/cigarette consumption among live ethnic groups in Hawaii. Br J Cancer i960; 41:939-940. (304) Jensen OM. Cancer morbidity and causa of death among Danish brewery workers, lnt J Cancer 1979; 23:454-463. (303) Piquet J, Tison. Alcool et cancer de Poesophage. Bull Acad Med (Paris) 1937; 117:236-239 (in French). (306) Schwartz D. Alcool ec cancer. Etude de geographie pathologique. Cancro 1966; 19:200-209 (in French; English summary). (307) Schwartx D. Lellouch J, Flamant R, De.noix PF. Alcool et cancer. Resultacs d'une equete retrospective. Rev Fr Etud Clin Biol 1962; 7:590-604 (in French; English summary). (3Q8) Hakllinen T, Lihtiuaki L, Lxhtonen M, Terro L. Cancer mor bidity among two male cohorts with increased alcohol consump tion in Finland. J Natl Cancer I tut 1974; 52:1711-1714. (309) World Health Organization. Cancer agents that surround us. World Health 1964 Sept: 16-17. (3f0) Bross ID. Coo*<as J. Early onset of oral cancer among women who drink and smoke. Oncology 1976; 33:13&-139. (J//) Birch JD. Howe GR. Millxb AB. Semincjw R. Tobacco, alcohol. asbestos, and nickel in the etiology of cancer of the larynx: A casecontrol study. JNCI 1981.67:1219-1224. (3/2) Graham S. Daval H, RoKVEa T, ei aJ. Dentition, diet, tobacco, and alcohol in the epidemiology of oral cancer. J Natl Cancer Inst 1977:39:1611-1618. (313) Kamionkowski MD, Fleshler B. The rale of alcoholic intake in esophageal carcinoma. Am J Med Sci 1965; 249:696-700. (314) Keller AZ. Terris M. The association of alcohol and tobacco with cancer of the mouth and pharynx. Am J Public Health 1965; 35:1570-1585. (315) Keller M. Proxusel DM, Spiccler D. Lic.ht L, Davies MN, eds. Alcohol and cancer. In; Second special report to the U S. Congress on alcohol and health. Washington, D C.: U.5. Govt Print Off. 1977:53-67 [OIIEW publication No. (ADM)75-212). (316) Moore C. Smoking and cancer of the mouth, pharynx, and larvnx. JAMA 1963: 191:283-286. (3/7) Rothman K. Keller A. The effect of joint exposure to alcohol and tobacco on risk of cancer of the mouth and pharvnx. J Chronic Dis 1972; 25:711-716. (3/8) Schottenulo D. Alcohol as a co-factor in the etiology of cancer. Cancer 1979; 43:1962-1966. (2/9) ScHorrtNPELD D. Gantt RC, Wynoe EL The role of alcohol and tobacco in multiple primary cancers of the upper digestive system, larynx and lung: A prospective study. Prev Med 197-4; 3 277-29). (320) Vincent RG, Marchetta F. The relationship of the use of tobacco and alcohol to cancer of the oral cavity, pharynx or larynx. Ant J Surg 1963; 106:501-505. (32/) Wyndci EL, Baoas IJ, Feldman RM. A study of the etiological factors in cancer of the mouth. Cancer 1957; i0:1300-1323. (322) MacDonald RA. Cirrhosis and primary carcinoma of the liver: Changes in their occurrence at the Boston City Hospital, 18971954. N Eng! J Med 1936; 253:1179-1183. (323) Purttlo DT, CorruEi LS. Cirrhosis and hepatoma occurring at Boston City Hospital (1917-1968). Cancer 1973; 32:458-462. (324) Flamant R, Lasserri O, Lazar P, Lecuerinais J, Denoix P, Schwartz D. Differences in sex ratio according to cancer site and possible relationship with use of tobacco and alcohol. Review of 65,000 cases. J Natl Cancer Inu 1964; 32:1309-1316. (325) Martinez 1. Retrospective and prospective study of carcinoma of the esophagus, mouth, and pharynx in Puerto Rico. Bol Assoc Med PR 1970; 62:170-178. (326) McCoy GD, Chen CB, Heckt SS, McCoy EC. Enhanced metabo lism and mutagenesis of nitrosopyrrolidine in liver fractions iso lated from chronic et Hanoi-consuming hamsters. Cancer Res 1979: 39:793-796. (327) Pottery LM. Morris LE, Blot WJ. Ziegler RG, Fracmeni JF Jr. Esophageal cancer among black men in Washington. DC I. Alcohol, tobacco, and other risk factors. JNCI 1981; 67:777-783. (328) RothXian KJ. The proportion of cancer attributable to alcohol consumption. Prev Med 1980: 9:174-1*9. (J29) Wafxicx 5. Castle W, Nichollc D. Zanamwe LN, Gelsand M. Cigarette smoking, alcohol and caneer of the oesophagus. S Afr Med J 1972; 46:2023-2026. (330) HicctxaoN J. Mur CS. Guat editorial: Environmental carcinogen esis: Misconceptions and limitations to cancer control. JNCI 1979; 63:1291-1298. (JJ/) Wynoer EL, Coat CB. Cuesi editorial. Contribution of the environ ment to cancer incidence: An epidemiologic exercise. J Nail Cancer Inst 1977: 56:825-032. (332) Doll R. Prro R. The causes of cancer: Quantitative estimates of avoidable risks of cancer in the United States todav. JNCI 1981: 86:1191-1308. URL 03994 JNCI, VOL. 70, NO. 6, JUNE 1983 23 Sepiember 1983, Volume 221. Number 4617 URL 03995 AMERICAN ASSOCIATION FOR THE ADVANCEMENT OF SCIENCE Si ivm c serve, ns readers .is a forum for the prcsciYtunon and discussion of important issues related to the ad'anccmcnl of science, including (he prescntalion of minority or conflictitiF points of slew, rather than by publishing only material on v, hich a consensus has been reached. Accordingly, all articles published tn Sci ence---including editorials, news and comment, and book reviews--are signed and reflect the individual views of the authors and not official points of view adopted by the AAAS or the institutions with which the authors are affiliated. Editorial Board iW* RFrederick BlaTTncr, Bernard F. BlRKE. Charles L FDrake. Arthur Findeis. E. Peter Geiduschek. Glynn Isaac. Milton Russell. Wil liam p Slicitter. John Wood tv*4- Arnold Demain. Neal E. Miller. Freder ick Mosteller. Allen Newell. Rlth Patrick. Bryant W. Kossiter. Veka C Robin. Solomon H. E.Snyder. Pall Waggoner Publisher William D. Carey Associate Publisher: ROBERT V. OkMES Editor: Philip H. ABELSON Editorial Staff Assistant Managing Editor: John E Ringle Production Editor: Ellen E, MlRPHY Business ttanager HanS Nl'SSBaLM di/nr Barbara J. ClllitOn Vfnj and Comment: Colin Norman ideputy editor!. Jeffrey L Fox. Constance Holden. Eliot Mar shall. R. Jeffrey Smith. Marjorie Son. John Walsh European Correspondent: David Dickson Comribuuni: Writer. Luther J. Carter Research \<*><\: Roger Lew in ideputy editori. Rich ard A. Kerr. Gina Koi.ay a. Ji.an L. Marx. Thomas H. Malgh II. Arthlr L. Robinson. M. Mitchell Waldrop Administrative Assistant. Sens: SchERRainc Mack: Editorial Aisifiant, Xen-s: Fannie Groom Senior Editors: ElEanORE Bltz. Mary DoRFMaN. Ruth Kllstad A.rjrKTtiic Editors: Sylvia Eberhart. Caitilin Gor don. Lois Schmitt Assistant Editors: Martha Collins. Stephen Kepple. Edith Meyers Book Renews: Katherine Limsgston. Editor: LtN- OA HEISERSIan. Janet Kegg Letters: Christine Gilbert Cop\ Editor. Isabella Bouldin Production: John Baker. Susannah Borg; Holly Bishop. Eleanor Warner; Jean Rockwqqd. Shar. on Ryan. Beverly Shields Coi i'rr. Reprints, and Permissions: GravcE Finger. Editor: Geraldine Crump. Corrine Harris Guide to Scientific Instruments. Richard G. Sommer Assistant to the Editor: Susan Elliott /UiMia/ir to the Associate Publisher: Rose Lowery A.uKHinr in the Managing Editor. Nancy HartnaGEL Membership Recruitment: Gwendolyn Hldole Member and Subscription Records: Ann RaGLaND EDITORIAL CORRESPONDENCE; 1515 Massachu setts Ave.. NW, Washington. D.C. '0005. Area code 202. General Editorial Office. 467-4350; Book Reviews. 467-4367; Guide to Scientific Insiruments. 467-44B0; News and Comment. 467-4430; Reprints and Permis sions, 467-4483; Research News. 467-4321. Cable: Advancesci. Washington. For '`Information for Contribu tors," write to (he editorial office or see page xi, Sfirnre. 24 June 1983. BUSINESS CORRESPONDENCE: Area Code 20'. Membership and Subscriptions: 467-4417, Advertising Representatives Director. Earl J. Scheraoo PriXfurfui" Manacc'. Gina REILLY Advertising Sales Manager: Richard L. Charles Marketing Manager: HcRBFRT L. Blrklu.nO Sales- New-York. N.Y. KX136: Steve Hamburger. 1515 Broadway (212-730-1050); Scotch Plains. n.J. 0707&: C. Richard Callis. 12 Linami Lane (20I-8H9-4K73I; Chi cago, III. 60611; Jack Rvun. Room 2107. 919 N. Michigan Ave. (312-337-4973); Beverly Hills. Calif. 90211: Winn Nance. 111 N. La Cicnega Blvd. (213-6572772); Dorset. Vt. 05251: Fred W. DicfTenhach, Kent Hill Rd. (802-867-5'8l>. ADVERTISING CORRESPONDENCE: Tenth floor. 1515 Broadway. New York. N.Y. 10036. Phone: 212730-1050. Dietary Carcinogens In this issue of Science. Bruce Ames reviews the increasing body of evidence that large numbers of potent carcinogens arise from natural processes. Mutagens are present in substantial quantities in fruits and vegetables. Carcinogens are formed in cooking as a result of reactions involving proteins or fats. Dietary practices may be an important determi nant of current cancer risks. Ames describes the role of plant materials as follows; Plants in nature synthesize toxic chemicals in large amounts, apparently as a primary defense against the hordes of bacterial, fungal, and insect and cither animal predators. Plants in the human diet are no exception. The variety of toxic chemicals is so great that organic chemists have been characterizing them for over 100 years, and new plant chemicals are still being discovered. Recent widespread use of short-term tests for detecting mutagens and theincreased testing of plant substances for carcinogenicity in animals have contributed to the identification of many natural mutagens, teratogens, and carcinogens in the human diet. Safrole and related compounds are present in many edible plants. Safrole is a carcinogen in rodents and some of its metabolites are mutagens. Oil of sassafras, once used id flavor some root beer, is about 75 percent safrole. Black pepper contains about 10 percent by weight of a closely related compound, piperine. Extracts of black pepper at a dose equivalent to 4 milligrams of dried pepper per day cause tumors in mice at many sites. Many hydrazines are carcinogens and mutagens, and large amounts of them are found in edible mushrooms. One carcinogenic hydrazine is present in the false morel at a concentration of 50 milligrams per 100 grams. It causes lung tumors in mice at a level of 20 micrograms per mouse per day. Carcinogens and mutagens are present in mold-contaminated foods such as corn. nuts, peanut butter, bread, cheese, and fruit. Some of these contaminants, such as aflaioxin, are among the most potent know n carcino gens and mutagens. Nitrosamines and nitroso compounds are suspect as causative agents of stomach and esophageal cancer in humans. In the digestive system these nitrogen compounds arc formed from nitrate and nitrite. Beets, celery, lettuce, spinach, radishes, and rhubarb all contain about 200 milligrams of nitrate per 100-gram portion. Rancid fats are possible causative agents of colon and breast cancer in humans. These forms account for a substantial fraction of all the cancer deaths in the United Stales. Unsaturated fats are easily oxidized on standing and in cooking to form mutagens, promoters, and carcinogens. Among the numerous products of such oxidations are fatty acid hydroperoxides and cholesterol epoxide. Thus the colon and digestive tract are exposed to many fat-derived carcinogens. Human breast fluid can contain high levels of cholesterol epoxide. Burnt and browned materials formed by heating proteins during cooking are highly mutagenic. Chemicals isolated from such products have been found to be carcinogenic when fed to rodents. In addition, the browning reaction products from caramelizution of sugars or the reaction of amino acids and sugars during cooking contain a large variety of DNA-damaging agents. The view that dietary practices might be a causative factor in cancer is not new. Epidemiologists have noted marked differences in cancer rates be tween population groups. Effects from changes in diet following migration have also been observed. Results of current studies are beginning to delineate more, sharply specific causative agents. When more definitive information is available, it should be possible for prudent people to choose fruits and vegetables that present minimal hazards. In the meantime, there is persuasive evidence that charred meats and rancid fats should not be part of the diet.--Philip H. Abelson . /L/ URL 03996 amounts of safrole and large amount7"\ (close to 10 percent by weight! of the l closely related compound piperinc (26). } Extracts of black pepper cause tumors in mice at a variety of sites at a dose of ^ Dietary Carcinogens and extract equivalent to 4 mg of dried pep: per peraay (about 160 mg/kg per davU'or Anticarcinogens T montt\g; an estimate ofthe average human intake of black pepper is over 140 mg per day (about 2 mg'kg per day! for Oxygen radicals and degenerative diseases life (26). 2) Most hydrazines that have been Bruce N. Ames tested are carcinogens and mutagens, and large amounts of carcinogenic hy drazines are present in edible mush rooms. The widely eaten false morel (Gyromitra escalenta) contains I I hydra Comparison of data from different Natural Mutagens and zines. three of which are known_oarcino- countries reveals wide differences in the Carcinogens in Food rates of many types of cancer. This leads gens (23). One of these. iV-mcthvl-iVformylhydrazine. is present at a concen to hope that each major type of cancer may be largely avoidable, as is the case for cancers due to tobacco, which consti tute 30 percent of the cancer deaths in y Plant material. Plants in nature syn- j tesize toxic chemicals in large amounts.i fpparently as a primary defense against' the hordes of bacterial, fungal, and in tration of 50 mg per 100 g and causes tung tumors in mice at the extremely low dietary level of 20 ng per mouse per dav (28). The most common commercial the United States and the United King sect and other animal predators f5-40). mushroom. Aanrictts bisporus, contains dom (/). Despite numerous suggestions IPlants in the human diet are no excep-]j about 300 me of aearinne. the 8-glutamyl to the contrary, there is no convincing [tion. The variety of these toxic chemi derivative of the mutagen 4-hydroxy- evidence of any generalized increase in U.S. (or U.K.) cancer rates other than what could plausibly be ascribed to the Summary. The human diet contains a great variety of natural mutagens and delayed effects of previous increases in carcinogens, as well as many natural antimutagens and anticarcmogens. Many of tobacco usage Thus, whether or these mutagens and carcinogens may act through the generation of oxygen radicals. not any recent changes in life-style or Oxygen radicals may also play a major role as endogenous initiators of degenerative pollution in industrialized countries will processes, such as DNA damage and mutation (and promotion), that may be related substantially affect future cancer risks, lo cancer, heart disease, and aging. Dietary intake of natural antioxidants could be an some important determinants of current important aspect of the body's defense mechanism against these agents. Many risks remain to be discovered among antioxidants are being identified as anticarcinogens. Characterizing and optimizing long-established aspects of our way of such defense systems may be an important part of a strategy of minimizing cancer life. Epidemiologic studies have indicat and other age-related diseases. ed that dietary practices are the most promising area to explore (/. 4). These studies suggest that a general increase in cals is so great that organic chemists methylphenylhydrazine. per 100 g of consumption of fiber-rich cereals, vege have been characterizing them for over mushrooms, as well as smaller amounts tables. and fruits and decrease in con 100 years, and new plant chemicals are of the closely related carcinogen A'-ace- sumption of fat-rich products and exces still being discovered (12. 24. 25). How tyl - 4 - hydroxymethylphenylhydrazine sive alcohol would be prudent </. 4). ever. toxicological studies have been (28). Sorpe agaritine is metabolized by There is still a lack of definitive evidence completed for only a very small percent the mushroom to a diazonium derivative about the dietary components that are age of them. Recent widespread use of which is a very potent carcinogen (a critical for humans and about their mech short-term tests for detecting mutagens single dose of 400 ng/g gave 30 percent of anisms of action. Laboratory studies of (41. 42) and the increased number of mice stomach tumors) and which is also natural foodstuffs and cooked food are animal cancer tests on plant substances present in the mushroom in smaller beginning to uncover an extraordinary (6) have contributed to the identification amounts (23). Many hydrazine carcino variety of mutagens and possible carcin of many natural mutagens, teratogens, gens may act by producing oxygen radi ogens and anticarcinogens. In this article and carcinogens in the human diet <5- cals |4J). I discuss dietary mutagens and carcino 40). Sixteen examples are discussed be 3) Linear ftirmoinnarins such as psor gens and anticarcinogens that seem of low. alen derivatives are potent light-activat importance and speculate on relevant I) Safrole. e'itrai'olc. methylenyenol. ed carcinogens and mutagens and are biochemical mechanisms, particularly and related compounds are present in widespread in plants of the Umbelliferae the role of oxygen radicals and their many edible plants (5). Safrole. estra- family, such as celery, parsnips, figs, inhibitors in the fat-cancer relationship, gole. and methyleugenol are carcinogens and oarslev (for instance. 4 mg per 100 g promotion, anticarcinogenesis, and ag in rodents, and several of their metabo of parsnip) (17. 19. 44). The level in ing. lites are mutaecns (J). Oil of sassafras, celery (about 100 jxg per 100 g) can The author is chairman of ihe Dorurtmcni of Biochemistry, University of California, Berkeley 94720. which had been used in "natural'' sarsa parilla root beer, is about 75 percent safrole. Black pepper contains small increase about 100-fold if the celery is stressed or diseased (19). Celery pickers and handlers commonly develop skin i:j* science, vol. m i.6 b 0 W rashes on their arms when exposed to diseased celery U9). Oil of bergamot, a citrus oil. is very rich in a psoralen and was used in the leading suntan lotion in France i/7). Psoralens, when activated by sunlight, damage DNA and induce tanning more rapidly than the ultraviolet "component of sunlight, which is also a carcinogen (.17). Psoralcnsv(plus light) are also effective in producing oxygen radicals (18). 4) The potato glycoalkaloids solcinine and chaconine are strong cholinesterase inhibitors and possible teratogens and are present at about 15 mg per 200 g of potato (12. 13). When potatoes are dis eased. bruised, or exposed to light, these and other l-4i glycoalkaloids reach lev els that can be lethal to humans i/2>. Plants typically respond to damage by making more land often different) toxic chemicals as a defense against insects and fungi [19. 24. 25). The different cultivars of potatoes vary in the concentra tion of these toxic glycoalkaloids (the concentration is a major determinant of insect and disease resistance): one cultivar bred for insect resistance had to be withdraw n from use because of its toxici ty to humans <> 40 mg of glycoalkaloids in a 200-g potato is considered to be a toxic level) (12). 5) Quercetin and several similar flavonoids are mutagens in a number of short-term test systems. Plavonoids are extremely widespread (daily levels close to I g) in the human diet (8. 16. 20. 21). There is evidence for the carcinogenicity of quercetin in two strains of rats (). although it was negative in other experi ments [21). 6) (jQuinones and their phenol precursors (9. 14. 16. 23. 45) are widespread in the human diet. Quinones are quite toxic as they can act as electrophiles or accept a single electron to yield the semiquinone radical, which can either react di rectly with DNA (/4, 46) or participate in a redox cycle of superoxide radical gen eration by transferring the electron to O; (47). The superoxide radical and its met abolic product H;0; can, in turn, lead to the oxidation of fat in cellular mem branes by a lipid peroxidation chain re action, thus generating mutagens and carcinogens, as discussed hglnw A mim, ber of quinones and dietary phenols have_ been shown to be mutagens (7, 9. 16. 23. 44). Mutagenic anthraquinone derivalives are found in plants such as rhubarb and in mold toxins (7. 16, 48). Many dietary phenols can spontaneously autoxidize to quinones, generating hydro gen peroxide at the same time [examples are catechol derivatives such as the caf* feic acid component of chlorogenic acid (9). which is present at about 250 mg per cup of coffee]. The amounts of these phenols in human urine fand in the diet) are appreciable (45). Catechol, for exam ple. is excreted in urine at about 10 mg per day and appears to be mainly derived from metabolism of plant substances (45). Catechol is a notent promoter of carcinogenesis (45), an ind"r*r pf nisia~~ damage, a likely active metabolite of the carcinogen benzene M6). and a toxic agent in cigarette smoke (45\. t_aiecnosamtne ifi55cTt5?f^r^ar(!IoInyopathy is thought to occur through generation of oxygen radicals (49). 7) Theobromine, a relative of caffeine, has been shown to be genotoxic in a variety of tests, to potentiate (as does caffeine) DNA damage bv various car cinogens in human cells, and to cause testicular atrophy and spermatogenic cell abnormalities in rats (27). Cocoa powder is about 2 percent theobromine, and therefore humans may consume hundreds of milligrams of theobromine a day from chocolate. Theobromine is also present in tea.8 9 10 8) Pxrrolizidine alkaloids are carcino genic. mutagenic, and teratogenic and are present in thousands of nlapt species I often at > 1 percent by weight), some of which are ingested by humans, particu larly in herbs and herbal teas and occa sionally in honey (7, 29). gyrrolizidine alkaloid poisoning's in humans (as well as in other mammals) cause lung and liver lesions and are commonly misdiagnosed (29). 9) The broad (fava) bean (Wc/a faba). a common food of the Mediterranean region, contains the toxins vicine and com icine at a level of about 2 percent of the dry weight (30). Pythagoras forbade his followers to eat the beans, presum ably because he was one of the millions of Mediterranean people with a deficien cy of glucose-6-phosphate dehydrog enase. This deficiency results in a low glutathione concentration in blood cells, which causes increased resistance to the malarial parasite, probably accounting for the widespread occurrence of the mutant gene in malarial regions. Howev er, the low glutathione concentration also results in a marked sensitivity to agents that cause oxidative damage, such as the fava bean toxins and a varie ty of drugs and viruses. Sensitive indi viduals who ingest fava beans develop a severe hemolytic anemia caused by the enzymatic hydrolysis of vicine to its aglycone, divicine, which forms a quinone that generates oxygen radicals (30). 10) Ally! isatltiocyanute. a major flavor ingredient in oil of mustard and horse radish, is one of the main toxins of the mustard seed and has been shown to cause chromosome aberrations in ham ster cells at low concgrumlinn (50\ and to be a carcinogen in rats (31). 11) possypoi is a major toxin in cot tonseed and accounts for about I percent of its dry weight (32). Gossypol causes pathological chances in rat and human testes, abnormal sperm, and male steril ity (32. 33). Genetic damage has been observed in embryos sired by gossvpolrreated mule rats: dominant lethal muta tions in embryos were measured after males were taken off gossypol treatment and allowed to mate (22). Gossypol ap pears to be a carcinogen as well: it has been reported to be a potent initiator and also a promoter of carcinogenesis in skin painting studies with mice (34). Crude, unrefined cottonseed oil contains consid erable amounts of gossypol (100 to 750 mg per 100 ml). Thus human consump tion may be appreciable in countries, such as Egypt, where fairly crude cot tonseed oil is commonly used in cooking, Gossypol is being tested as a male con traceptive in over 10.000 people in China (at an oral dose of about 10 mg per person per day), as it is inexpensive and causes sterility during use (22). Gossypol's mode of action as a spermicide may be through the production of oxygen radicals (25). Plant breeders have developed "'gland less cotton." a new strain with low levels of gossypol. but seeds from this strain are much more susceptible to attack by the fungus AspergillusJiavus. which pro duces the potent carcinogen aflatoxin (261. 12) Siernilic acid and malvalic acid are widespread in the human diet. They are toxic cvclopropenoid fatty acids present in cottonseed oil and other oils from seeds of plants in the family Malvaceal (for instance, cotton, kapok, okra, and durian) (51). Another possible source of human exposure is consump tion of fish, poultry, eces. and milk from animals fed on cottonseed 157). Cvclo propenoid Tatty acids are carcinogens in trout, markedly potentiate the carcinoge nicity of aflatoxin tn trout, cause athero sclerosis in rabbits, are mitogenic in rats. and have a variety of toxic effects in farm animals (51). The toxicity of these fatty acids could be due to (heir ease of oxida tion to form peroxides and radicals (51). 13) Leguminous plants such as lupine contain very potent teratogens (22). When cows and goats forage on these plants, their offspring may have severe teratogenic abnormalities; an example is 23 SEPTEMBER 1983 1257 URL 03998 the characteristic "crooked calf abnor (25). There are health costs for the use of number of animal studies have shown mality due to the ingestion of arutgyrine these natural pesticides, just as there are that high dietary fat is a promoter and a from lupine [22). In addition, significant for man-made pesticides (41, 54). and presumptive carcinogen (4. 67, 68). Co amounts of these teratogens are transr these must be balanced against the costs lon and breast cancer and lung cancer ferred to the animals* milk, so that drink of producing food. However, little infor (which is almost entirety due to cigarette ing the milk during pregnancy is a seri mation is available about the toxicology smoking) account for about half of all ous teratogenic hazard [22). In one rural of most of the natural plant toxins in our U.S. cancer deaths. In addition to the California family, a baby boy, a litter of diet, despite; the laree doses we are ex cyclopropenoid fatty acids already dis puppies, and goat kids all had posed to. Many, if not most, of these cussed. two other plausible mechanisms "crooked*' bone birth-defect abnormali plant toxins may be "new" to humans in involving oxidative processes could ac ties. The pregnant mother and the dog the sense that the human diet has count for the relation (69) between high had both been drinking milk obtained chanced drastically with histone times fat and both cancer and heart disease. from the family goats, which had been By comparison, our knowledge of the_ 1) Rancidfat. Fat accounts for over 40 foraging on lupine (the main forage in toxicological effects of new man-made percent or the calories in (he U.S. diet winter) (22). It was at first mistakenly pesticides is extensive. andeenecaLejt^. (67), and the amount of ingested oxidized thought that the birth defects were pnoire is exceedingly lnw (SO fat may be appreciable (70. 71). Unsatu caused by spraying of 2.4-D. Plants also contain a variety of anticar rated fatty acids and cholesterol in fat 14) Sesquiterpene lactones are wide cinogens (55), which are discussed be are easily oxidized, particularly during spread in many plants (27), although low. cooking (70, 71). The lipid peroxidation because they are bitter they are not eaten Alcohol. Alcohol has long been associ chain reaction (rancidity) yields a variety in large amounts. Some have been ated with cancer of the mouth, esopha (71-73) of mutagens, promoters, and car shown to be mutagenic (27). They are a gus. pharynx, larynx, and, to a lesser cinogens such as fatty acid hydroperox major toxin in the white sap of Lactuca extent, liver (/ 56), and it appears to be ides (62). cholesterol hydroperoxide virosa (poison lettuce), which has been an important human teratogen, causing a (74). endoperoxides, cholesterol and fat used as a folk remedy. Plant breeders are variety of physical and mental defects in ty acid epoxides (74-77), enals and other now transferring genes from this species babies of mothers who drink (57). Alco-_ aldehydes (44, 59, 78). and alkoxy and to commercial lettuce to increase insect hoi drinking causes ahnnrnnnlitiff* in hydroperoxy radicals (44. 72). Thus the resistance (25). mice (57a) and is a synergist forchromo-_ colon and digestive tract are exposed to 15) The phorbol esters present in the some damage in humans (55). Alcohol, a variety offal-derived carcinogens. Hu Euphorbiacea. some of which are used metabolism generates acetaldehyde. man breast ffuid can contain enormous as folk remedies or herb teas, are potent which is a mutagen and teratogen (59). a levels (up to 780 ia.W) (75) of cholesterol promoters of carcinogenesis and may cocarcinogen, and possibly a carcinogen epoxide (an oxidation product of choles have been a cause of nasopharyngeal (60). and also radicals that produce lipid terol). which could originate from either cancer in China and esophageal cancer in hydroperoxides (6/) and other mutagens ingested oxidized fat or oxidative pro Curasao (29). and carcinogens (62; see below). In some cesses in body lipids. Rodent feeding 16) Alfalfa snrnnK contain canavan- epidemiologic studies on alcohol (56). it studies with oxidized fat (79) have not ine, a highly toxic arginine analog that is has been suggested that dietary green yielded definitive results. incorporated into protein in place of argi vegetables are a modifying factor in the 2) Peroxisomes oxidize an appreciable nine. Canavanine. which occurs in alfal _ reduction of cancer risk. percentage of dietary fatty acids, and fa sprouts at about 1.5 percent of their / Mold carcinogens. A variety of mold j removal of each two-carbon unit gener dry weight (40). appears to be the active rarcinogens and mutagens are present in ates one molecule of hydrogen peroxide agent in causing the severe lupus ervthe- mold-contaminated food such as com. (a mutagen, promoter, and carcinogen) matosus-like syndrome seen when mon grain, nuts, peanut butter, bread, (50. 57). Some hydrogen peroxide es keys are fed alfalfa sprouts (40). Lupus cheese, fruit, and apple juice (15. 62). capes the catalase in the peroxisome (80, in man is characterized bv a defect in the Some of these, such as sterigmatocystin 82.83). thus contributing to the supply of immune system which is associated with and afiatoxin, are among the most potent oxygen radicals, which also come from autoimmunity, antinuclear antibodies, carcinogens and mutagens known (15. other metabolic sources (72. 52-55). Hy chromosome breaks, and various types 63). Dietary glutathione has been report droperoxides generate oxygen radicals in ot pathology (40). The chromosome ed to counteract afiatoxin carcinogenic the presence of iron-containing com breaks appear to be due to oxygen radi cals as they are prevented by superoxide dismutase (52). The canavanine-alfalfa ity. Nitrite, nitrate, and nitrosamines. A number of human cancers, such as stom pounds in the cell (72). Oxygen radicals, in turn, can damage DNA and can start the rancidity chain reaction which leads sprout pathology could be due in part to ach and esophageal cancer, may be relat- to the production of the mutagens and the production of oxygen radicals during ed to nitrosamines and other nitroso carcinogens listed above (72). Drugs phagocytization of antibody complexes compounds formed from nitrate and ni such as clofibrate. which cause lowering with canavanine-containing protein. trite in the diet 164, 65). Beets, celery. of serum lipids and proliferation of per The 16 examples above, plus coffee"") lettuce, spinach, radishes, ^pfl rhnharh oxisomes in rodents, result in age pig (discussed below), illustrate that the hu all contain about 200 me of nitrate_ner. ment (lipofuscin) accumulation (a sign of man dietary intake of "nature's pesti 100-g portion (65). Anticarcinogens in lipid peroxidation in tissues) and liver cides" is likely to be several grams per the diet may be important in this context tumors in animals (80). Some fatty acids, day--probably at least 10.000 times high as well (66). such as Cj2;i and certain irons fatty er than the dietary intake of man-made Fat and cancer: possible oxidative acids, appear to cause peroxisomal pro pesticides (52). mechanisms. Epidemiologic studies of liferation because they are poorly oxi Levels of plant toxins that confer in cancer in humans suggest, but do not dized in mitochondria and are preferen sect and fungal resistance are being in prove, that high fat intake is associated tially oxidized in the peroxisomes, al creased or decreased by plant breeders with colon and breast cancer(7, 4. 67). A though they may be selective for heart or 1ISK SCIENCE, VOL. 221 URL 03999 liver (56). There has been controversy about the role of trans fatty acids in cancer and heart disease, and recent evidence suggests that trans fatty acids might not be a risk factor for atheroscle rosis in experimental animals (87). Americans consume about 12 g of trans fatty acids a day (87) and a similar amount of unnatural cis isomers [which need further study (#)], mainly from hydrogenated vegetable Tats. Dietary fatty acids are also obtained from rapeseed oil and fish oils (86). Thus oxidation of certain fatty acids might generate grams of hydrogen peroxide per day within the peroxisome (86). Another source of fat toxicity could be perturba tions in the mitochondrial or peroxisom al membranes caused by abnormal fatty acids, yielding an increased flux of su peroxide and hydrogen peroxide. Mito chondrial structure is altered when rats are fed some abnormal fatty acids from partially hydrogenated fish oil ($9). Di etary C;;:i fatty acids and clofibrate also induce ornithine decarboxylase (56). a common attribute of promoters. A recent National Academy of Sci ences committee report suggests that a reduction of fat consumption in the American diet would be prudent (4). although other scientists argue that, until we know more about the mechanism of the fat-cancer relation and about which types of fat are dangerous, it is prema ture to recommend dietary changes (90). Cooked Food as a Source of Ingested Burnt and Browned Material Work of Sugimura and others has indi cated that the burnt and browned materi al from heating protein durinp cooking is highly mutagenic (21, 91). Several chem icals isolated on the basis of their mu tagenicity from heated protein or pyro- lyzed amino acids were found to be carcinogenic when fed to rodents (21). In addition, the browning reaction products from the caramelization of sugars or the reaction of amino acids and sugars dur ing cooking (for instance, the brown ma terial on bread crusts and toasted bread) contain a large variety of DN A-damaging agents and presumptive cardnngens (23 38. 92). The amount of burnt and browned material in the human diet mav be several grams per day. By compari son about 500 me of burnt material is inhaled each day bv a smoker using two packs of cigarettes fal J>n lQr p"r cigarette) a dav. Smokers have more easily detectable levels of mutagens in their urine than nonsmokers (93). but so do people who have consumed a meal of fried pork or bacon (94). In the evalua tion of risk from burnt material it may be useful fin addition to carrying out epide miologic studies) to compare the activity of cigarette tar to that of the burnt mate rial from cooked food (or polluted air) in short-term tests and animal carcinoge nicity tests involving relevant routes of exposure. Route of exposure and com position of the burnt material are critical variables. The risk from inhaled cigarette smoke can be one reference standard: an average life shortening of about 8 years for a two-pack-a-dav smoker. The amount of burnt material inhaled from severely polluted cilv air, on the other -hand, is relatively small: it would be necessary to breathe smoggy Los Ange- [es air (III fig/m1 total particulates: 3l p.g/m3 soluble organic matter) for 1 to ^ weeks to equal the soluble organic mat ter of the particulates or the mutagenic ity from one cigarette (20 mg of tar) (95). Epidemiologic studies have not shown significant risks from city air pollution alone (/. 96). Air in the houses of smo}^_ ers is considerably more polluted than city air outside (97). Coffee, which contains a considerable amount of burnt material, including the mutagenic pyrolysis product methylgjy- oxal. is mutagenic (21. 98). However, one cup ofgoffee also contains about 250 mg of the natural mutagen chlorogenjc acid (9) [which is also an antinitrosating agent (66)]. highly toxic atractylostdes (10). the glutathione transferase inducers kahweal palmitaie and cafestol palmitate (//Land about 100 mg of caffeine [which inhjbits a DNA-repair and can increase tumor yield (99) and cause birth defects at high levels in several e*r^rL_ mental species- (/00)j,_There is prelimi nary. but not conclusive, epidemiologic evidence that heavy coffee drinking is associated with cancer of the ovary, bladder, pancreas, and large bowel U0I). Cooking also accelerates the rancidity reaction ot cooking oils and fat in meat (70. 71). thus increasing consumption of mutagens and carcinogens. Anticarcinogens We have many defense mechanisms to protect ourselves against mutagens and carcinogens, including continuous shed ding of the surface layer of our skin, stomach, cornea, intestines, and colon (102). Understanding these mechanisms should be a major goal of cancer, heart, and aging research. Among the most important defenses may be those against oxygen radicals and lipid peroxidation if, as discussed here, these agents are major contributors to DNA damage (103). Ma jor sources of endogenous oxygen radi cals are hydrogen peroxide (83) and su peroxide (72. 104) generated as side products of metabolism, and thq oxygen radical burst from phagocytosis after vi ral or bacterial infection or the inflamma tory reaction (105). A variety of environ mental agents could also contribute to the oxygen radical load, as discussed here and in recent reviews (72. 106). Many enzymes protect cells from oxida tive damage: examples are superoxide dismutase (104). giutathione peroxidase (107). DT-diaphorase (108). and the glu tathione transferases (109). In addition, a variety of small molecules in our diet are required for anlioxidative mechanisms and appear to be anticarcinogens; some of these are discussed below. !) V/raoim E (tocopherol) is the major radical trap in lipid membranes (72) and has been used clinically in a variety of oxidation-related diseases (110). Vitamin E ameliorates both the cardiac damage and carcinogenicity of the quinones adriamycin and daunomycin. which are mutagenic, carcinogenic, cause cardiac damage, and appear to be toxic because of free radical generation (///). Protec tive effects of tocopherols against radia tion-induced DNA damage and mutation and dimethylhydrazine-induced carcino genesis have also been observed (112). Vitamin E markedly increases the endur ance of rats during heavy exercise, which causes extensive oxygen radical damage to tissues (113). 2) (3-Carotene is another antioxidant in the diet that could be important in pro tecting body fat and lipid membranes against oxidation. Carotenoids are freeradical traps and remarkably efficient quenchers of singlet oxygen (114). Sin glet oxygen is a very reactive form of oxygen which is mutagenic and particu larly effective at causing lipid peroxida tion (114). it can be generated by pig ment-mediated transfer of the energy of light to oxygen, or by lipid peroxidation, although the latter is somewhat contro versial. 3-Carotene and similar polyprenes are present in carrots and in all food that contains chlorophyll, and they appear to be the plants' main defense against singlet oxygen generated as a by product from the interaction of light and chlorophyll (115). Carotenoids have been shown to be amicarcinogens in rats and mice (116). Carotenoids (in green and yellow vegetables) may be anticar cinogens in humans (/, 56, 117). Their protective effects in smokers might be related to the high level of oxidants in both cigarette smoke and tar (45, 118). Carotenoids have been used medically in 23 SEPTEMBER 1983 1259 the treatment for some genetic diseases, such as porphyrias, where a marked pho tosensitivity is presumably due to singlet oxygen formation {119). 3) Selenium is another important di etary anticarcinogen. Dietary selenium (usually selenite) significantly inhibits the induction of skin, liver, colon, and mammary tumors in experimental ani mals by a number of different carcino gens. as well as the induction of mamma ry tumors by viruses (120). It also inhib its transformation of mouse mammary cells (121). Low selenium concentrations may be a risk factor in human cancer {122). A particular type of heart disease in young people in the Keshan area of China has been traced to a selenium deficiency, and low selenium has been associated with cardiovascular death in Finland 1123). Selenium is in the active site of glutathione peroxidase, an en zyme essential for destroying lipid hy droperoxides and endogenous hydrogen peroxide and thus helping to prevent oxygen radical-induced lipid peroxida tion (107). although not all of the effects of selenium may be accounted for by this enzyme (120). Several heavy-metal tox ins. such as Cd2' (a known carcinogen) and Hg2'. lower glutathione peroxidase activity by interacting with selenium (107). Selenite (and vitamin E) has been shown to counter the oxidative toxicity of mercuric salts (124). 4) Glutathione is present in food and is one of the major antioxidants and anti mutagens in the soluble fraction of cells. The glutathione transferases (some of which have peroxidase activity) are ma jor defenses against oxidative and alkyl ating carcinogens (109). The concentra tion of glutathione may be influenced by dietary sulfur amino acids (125. 126). /VAcetylcysteine. a source of cysteine, raises glutathione concentrations and re duces the oxidative cardiotoxicity of adriamycin and the skin reaction to radi ation (127). Glutathione concentrations are raised even more efficiently by l-2oxothiazolidine-4-carboxylate. which is an effective antagonist of acetamino phen-caused liver damage U26). Acet aminophen is thought to be toxic through radical and quinone oxidizing metabo lites (128). Dietary glutathione may be an effective anticarcinogen against aflatoxin a 29). 5) Dietary ascorbic acid is also impor tant as an antioxidant. It was shown to be anticarcinogenic in rodents treated with ultraviolet radiation. benzoMpyrene. and nitrite (forming niiroso carcin ogens) (64. 65. 130). and it may be in versely associated with human uterine cervical dysplasia (although this is not proof of a cause-effect relationship) (131). It was recently hypothesized that ascorbic acid may have been supple mented and perhaps partially replaced in humans by uric acid during primate evo lution (132). 6) Uric acid is a strong antioxidant present in high concentrations in the blood of humans </22). The concentra tion of uric acid in the blood can be increased by dietary purines; however, too much causes gout. Uric acid is also present in high concentrations in human safiva (132) and may play a role in de fense there as well, in conjunction with lactoperoxidase. A low uric acid level in blood may possibly be a risk factor in cigarette-caused lung cancer in humans (133). 7) Edible plants and a variety of sub stances in them, such as phenols, have been reported to inhibit (cabbage) or to enhance (beets) carcinogenesis (II. 55. 134) or mutagenesis (23 . 66. 92. 135) in experimental animals. Some of these substances appear to inhibit by inducing cytochrome P-450 and other metabolic enzymes [(134): see also (//)], although on balance it is not completely clear whether it is generally helpful or harmful for humans to ingest these inducing sub stances. The hypothesis that as much as 80 percent of cancer could be due to envi ronmental factors was based on geo graphic differences in cancer rates and studies of migrants (136). These differ ences in cancer rates were thought to be mainly due to life-style factors, such as smoking and dietary carcinogens and promoters (136). but they also may be due in good part [see also (/)J to less than optimum amounts of aniicarcinogens and protective factors in the diet. The optimum levels of dietary antioxi dants. which may vary among individ uals. remain to be determined; however, at least for selenium 1)20). it is important to emphasize the possibility of deleteri ous side effects at high doses. Oxygen Radicals and Degenerative Diseases Associated with Aging Apinp. A plausible theory of aging holds that the major cause is damage to DNA 1/02. 137) and other macromol ecules and that a major source of this damage is oxygen radicals and lipid per oxidation 143. 84. 103. 138-141). Cancer and other degenerative diseases, such as heart disease < 102). are likely to be due in good part to this same fundamental de structive process. Age pigment (lipofuscin) accumulates aging in all mammalian species and has been associated with lipid peroxidation (72, 84. 138, 139). The fluorescent products in age pigment are thought to be formed by malondialdehyde (a mutagen and carcinogen and a major end product of rancidity) crosslinking protein and lipids (138). Metabol ic rate is directly correlated with the rate of lipofuscin formation (and inversely correlated with longevity) (139). Cancer increases with about the fourth power of age. both in short-lived species such as rats and mice (about 30 percent of rodents have cancer by the end of their 2- to 3-year life-span) and in longlived species such as humans (about 30 percent of people have cancer by the end of their 85-year life-span) (142). Thus, the marked increase in life-span that has occurred in 60 million years of primate evolution has been accompanied by a marked decrease in age-specific cancer rates; that is. in contrast to rodents. 30 percent of humans do not have cancer by the age of 3 (142). One important factor in longevity appears to be basal metabol ic rate (139. 141). which is much lower in man than in rodents and could markedly affect the level of endogenous oxygen radicals. Animals have many antioxidant de fenses against oxygen radicals. In creased levels of these antioxidants, as well as new antioxidants, may also be a factor in the evolution of man from short-lived prosimians (143). It has been suggested that an increase in superoxide dismutase is correlated (after the basal metabolic rate is taken into account) with increased longevity during primate evolution, although this has been disput ed (74/). Ames et al. proposed (132) that as uric acid was an antioxidant and was present in much higher concentrations in the blood of humans than in other mam mals. it may have been one of the inno vations enabling the marked increase in life span Snd consequent marked de crease in age-specific cancer rates which occurred during primate evolution. The ability to synthesize ascorbic acid may have been lost at about the same time in primate evolution as uric acid levels be gan to increase (144). Cancer and promotion. Both DNAdamaging agents (initiating mutagens) (21. 41, 42) and promoters (145) appear to play an important role in carcinogene sis (2/. 146). It has been postulated that certain promoters of carcinogenesis act by generation of oxygen radicals and resultant lipid peroxidation (72. 146149). Lipid peroxidation cross-links pro teins (43. 150) and affects all aspects of cell organization (72). including mem brane and surface structure, and the mi- 1260 SCIENCE. VOL. 221 URL 04001 totic apparatus. A common properly of promoters may be their ability to pro duce oxygen radicals. Some examples are fat and hydrogen peroxide (which may be among the most important pro moters) (67. 68. 81). TCDD (151). lead and cadmium {152). phorbol esters (147, 149, 153). wounding of tissues (154). asbestos </5J). peroxides (156). catechol (45) (see quinones above), mczcrcin and teleocidin B (147). phenobarbital 057). and radiation (72. 158). Inflammatory reactions involve the production of oxy gen radicals by phagocytes (105). and this could be the basis of promotion for asbestos 055) or wounding (154). Some of the antioxidant anticarcinogeps (dis cussed above) are also antipromoters (73. 121. 146. 159, 160). and phorbol ester-induced chromosome damage (149) or promotion of transformation (159) is suppressed by superoxide dismutase. as would be expected if promoters were working through oxidative mecha nisms. Many ``complete*' carcinogens cause the production of oxygen radicals (73. 161): examples are nitroso com pounds, hydrazines, quinones, polycy clic hydrocarbons (through quinones). cadmium and lead salts, nitro com pounds. and radiation. A good part of the toxic effects of ionizing radiation damage to DNA and cells is thought to be due to generation of oxygen radicals (103. 162). although only a tiny part of the oxygen radical load in humans is likely to be from this source. Recent studies give some clues as to how promoters might act. Promoters dis rupt the mitotic apparatus, causing hemizygosity and expression of recessive genes (163). Phorbol esters generate oxy gen radicals, which cause chromosome breaks (164) and increase gene copy number (/65). Promoters also cause for mation of the peroxide hormones of the prostaglandin and leukotriene family by oxidation of araehidonic acid and other C'o polyenoic fatty acids, and inhibitors of this process appear to be antipro moters 060). These hormones are inti mately involved in cell division, differen tiation. and tumor growth (166) and could have arisen in evolution as signal molecules warning the cell of oxidative damage. Effects on the cell membrane have also been suggested as the impor tant factor in promotion, causing inhibi tion of intercellular communication (167) or protein kinase activation (167a). Heart disease. It has been postulated that atherosclerotic lesions, which are derived from single cells, are similar to benign tumors and are of somatic muta tional origin (102. 168). Fat appears to be one major risk factor for heart disease as well as for colon and breast cancer (69). In agreement with this, a strong correla tion has been observed between the fre quency of atherosclerotic lesions and adenomatous polyps of the colon (69). Thus, the same oxidative processes in volving fat may contribute to both dis eases. Oxidized forms of cholesterol have been implicated in heart disease 1/69), and atherosclerotic-like lesions have been produced by injecting rabbits with lipid hydroperoxide or oxidized cholesterol (169). The anticarcinogens discussed above could be anti-heart dis ease agents as well. As pointed out in the preceding section, vitamin E amelio rates both the cardiac damage and carci nogenicity of the free-radical-generating quinones adriamycin and daunomycin: N-acetylcysteine reduces the cardiotoxicity of adriamycin: and selenium is an antirisk factor for one type of heart dis ease. Other diseases. The brain uses 20 per cent of the oxygen consumed by man and contains an appreciable amount of unsaturated fat. Lipid peroxidation (with consequent age pigment) is known to occur readily in the brain (72). and possi ble consequences could be senile demen tia or other brain abnormalities (84). Sev eral inherited progressive diseases of the central nervous system, such as Batten's disease, are associated with lipofuscin accumulation and may be due to a lipid peroxidation caused by a high concentra tion of unbound iron (170). Mental retar dation is one consequence of an inherit ed defective DNA repair system (XP complementation group D) for depurinated sites in DNA (17]). Senile cataracts have been associated with light-induced oxidative damage (/72). The retina and an associated layer of cells, the pigment epithelium, are ex tremely sensitive to degeneration in vita min E and selenium deficiency (173). The pigment epithelium accumulates massive amounts of lipofuscin in aging and dietary antioxidant deficiency (173). The eye is well known to be particularly rich in antioxidants. The testes are quite prone to lipid peroxidation and to the accumulation of age pigment. A number of agents, such as gossypol, which cause genetic birth defects (dominant lethals) may be active by this mechanism. The various agents known to cause cancer by oxidative mechanisms are prospective mutagenic agents for the germ line. Thus, vitamin E, which was discovered 60 years ago as a fertility factor (72), and other antioxi dants such as selenium (174), may help both to engender and to protect the next generation. Risks There are large numbers of mutagens and carcinogens in every meal, ail per fectly natural and traditional [see also 421. Nature is not benign. It should be emphasized that no human diet can be entirely free of mutagens and carcino gens and that the foods mentioned are only representative examples. To identi fy a substance, whether natural or man made. as a mutagen or a carcinogen, is just a first step. Beyond this, it is neces sary to consider the risks for alternative courses of action and to quantitate the approximate magnitude of the risk, al though the quantification of risk poses a major challenge. Carcinogens differ in their potency in rodents by more than a millionfold (175). and the levels of partic ular carcinogens to which humans are exposed can vary more than a billionfold. Extrapolation of risk from rodents to humans is difficult for many reasons, including the longevity difference, anti oxidant factors, and the probable multicausal nature r>f mnst human cnnc&r Tobacco smoking is, without doubt, a major and wxll-understood risk, causing about 30 percent of cancer deaths and 21. perrenr Af fatal heart ap;>rtfs fas well as other degenerative diseases) in the United States 0). These percentages may increase even more in the near future as the health effects of the large increase in women smokers become apparent (/). Diet, which provides both carcinogens" and anticarcinogens, is extremely likely to be another major risk factor. Exces sive alcohol consumption is another risk. although it does not seem to be of the same general importance as smoking and diet. Certain other high-dose exposures might also turn out to be important for particular groups of people--for in stance, certain drugs, where consump tion can reach hundreds of milligrams per day: particular cosmetics: and cer tain occupational exposures (23_ where workers inhale dusts or solvents at hiefc concentration. We must also be prudent about environmental pollution (41. 54). Despite all of these risks, it should be emphasized that (he overall trend in life expectancy in the United States_U_con? tinuing steadily upward (176). The understanding of cancer and de generative disease mechanisms is being aided by the rapid progress of science and technology, and this should help to dispel confusion about how important health risks can be identified among the vast numher of minor risks. We have many methods of attacking the problem of environmental carcinogens (and anticarcinogens), including human epidemi- 23 SEPTEMBER 1983 1261 ology (/), short-term tests (41, 42, 177), and animal cancer tests (175). Powerful new methods are being developed [for instance, see (58, 177)] for measuring DNA damage or other pertinent factors with great sensitivity in individuals. These methods, which are often noninvasive as they can be done on blood or urine (even after storage), can be com bined with epidemiology to determine whether particular factors are predictive of disease. Thus, more powerful tools will be available for optimizing antioxi dants and other dietary anti-risk factors, for identifying human genetic variants at high risk, and for identifying significant health risks. References and Notes 1. R. Doll and R. Peto. /. Natl. Cancer Inst. 66. 1192(1981). 2. R. Peto and M. Schneiderman. Eds.. Banbury Report 9. Quantification of Occupational Can cer (Cold Spring Harbor Laboratory. Cold Spring Harbor. N.Y.. 1981). 3. Cancer Facts and Figures. 1963 (American Cancer Society. New York. 1982). 4. National Research Council. Diet. Nutrition and Cancer (National Academy Press, Wash ington. D.C.. 1982). 5. E. C. Miller. J. A. Miller, I. Hirono. T. Sugi- mura. S. Takayama, Eds.. Naturally Occurring Carcinogens-Mutagens and Modulators of Carcinogenesis (japan Scientific Societies Press and University Park Press. Tokyo and Baltimore. 1979): E. C. Miller et at.. Cancer Res. 43. 1124 (1983): C. Ioannides. M. D<la- forge. D. V. Parke. Food Cosmet. Toxicol. 19, 657(19811. 6. G- J- Kapadia. Ed., Oncology Overview on Naturally Oecuring Dietary Carcinogens of Plant Origin < international Cancer Research Data Bank Program. National Cancer Institute, Bethesda. Maryland, 1982). 7. A. M. Clark, in Environmental Mutagenesis, Carcinogenesis, and Plant Biology, E. i. Kle- kowski. Jr.. Ed. (Praeger, New York. 1982). vol. 1. pp. 97-132. 8. A. M. Pamukcu. S. Yalciner. J. F. Hatcher. G. T. 8ryan. Cancer Res. 40, 3468 (1980); J. F. Hatcher. A. M. Pamukcu. E. Erturk, G. T, Brvan, Fed. Proc. Fed. Am. Soc. Exp. Biol. 42, 786 (1983). 9. H. F. Such. M. P. Rosin. C. H. Wu. W. D. Powrie. Muiai. Res. 90. 201 (1981): A. A. Aver yanov. BioUtirmya 46. 256 (1981); A. F. Hanham. 8. P. Dunn. H. F. Stich. Mutat. Res. 116. 333 (19831. 10. K. H. Pegel, Chem. Eng. News 59. 4 (20 July 1981). 11. L. K. T. Lam. V. L. Spamins, L. W. Watten- berg. Cancer Res. 42. 1193 (1982). 12. S. J. Jadhav, R. P. Sharma. D. K. Salunkhe. CRC Crit. Rev. Toxicol. 9. 2] (1981). 13. R. L. Halt, Nun. Cancer 1 (No. 2). 27 (1979). 14. H. W. Moore and R. Czcmiak. Med. Res. Rev, 1. 249(1981). 15. I. Hirono. CRC Crit. Rev. Toxicol. 8, 235 (1981). 16. J. P. Brown. Mutat. Res. 75. 243 0980). 17. M. J. Ashwood-Smith and G. A. Poulion. ibid. 85. 389 0 981). 18. A. Ya. Potapenko. M. V. Moshnin, A. A. Krasnovsky. Jr.. V. L. Sukhomkov. Z. Natur- forsch. 37. 70 0982). 19. G. W. (vie. D. L. Holt, M. C. Ivey. Science 213. 909 0981): R. C. Beierand E. H. Oertli. Phytochemistry. m press: R. C. Beier. G. W. I vie. E. H. Oertli. in "Xenobiotics in Foods and Feeds," ACS Svmp. Ser., in press; D. L. Holt. Food Chem. Toxicol. 21. 163 0983). 20. G. Tamura. C. Gold. A. Feno-Luzzi. B. N. Ames, Proc. Natl. At ad. Set. US.A. 77, 4961 0 980). 21. T. Sugimura and S. Sato. Cancer Res. fSuppl.l 43. 2415s 0983); T. Sugimura and M. Nagao. in Mutagenicity: New Horizons in Genetic Toxi cology. J. A. Heddle. Ed. (Academic Press. New York, 1982). pp. 73-88. 22. W. W. Kilgore, D. G. Crosby, A. L. CraigmiJI, N. K. Poppcn. Calif. Agric. 35 (No. II) (No vember 1981); D. G. Crosby. Chem. Eng. 50. A. Kasamaki et al., Mutat. Res. I0S, 387 News 61. 37 (U Apnl 1983); C. D. Warren, (1982). ibid,, p. 3 03 June 1983). 31. J. D. Hendricks. R. 0. Sinnhuber. P. M. Love 23. H. F. Stich. M. P. Rosin. C. H. Wu. W. D. land. N. E. Pawiowsfci. J. E. Nixon. Science Powrie. in Mutagenicity; New Horizons in Ge 208. 309 (1980): R. A. Phelps. F. S. Shenstone. netic Toxicology. J. A. Heddle. Ed. (Academic A. R. Kemmerer. R. J. Evans. Poult. Sci. 44. Press, New York. 1982). pp. 117-142;_______ 358(I964);N. E. Pawlowski. personal commu W. D. Powrie. Cancer Lett. 14. 251 0981). nication. 24. N. Katsui. F. Yagihashi. A. Murai. T. Masa- 52. 1. Emerit. A. M. Michelson. A. Levy, J. P. mune. Bitil. Chirm. Soc. Jpn. 55. 2424 (1982); Camus. J. Emerit. Hum. Genet. 55. 341 09801. _______ ibid., p. 2428; R. M. Bostock. R. 5A3.. FDA Compliance Program Report of Findings. Laine. J. A. Kuc. Plant Physiol. 70. 1417 FY79 Total Diet Studies--Adult iNo. (1982). 7305.002): available from National Technical 25. H. Griesebach and J. Ebel. Angew. Chem. Int. Information Service. Springfield. V*.), It is Ed. Engl. 17.635 (1978). estimated that the daily dietary intake of syn 26. J. M. Concon. D. S. Newburg. T. W. Swer- thetic organic pesticides and herbicides is czefc.. Nun. Cancer 1 (No. 3). 22 (1979). about 60 iLg. with chlorpropham. malathion. 27. H. W. Renner arid R. Munzner. Mutat. Res. and DDE accounting for about three-fourths of 13. 275 09823; H. W. Renner. Experientia 38, this. An estimate of ISO ug of daily exposure in 600(1982);D, Mourdatos, J. Dozi-Vassiliades. Finland to pesticide residues has been made by A. Graniisas, Mutat. Res. 104, 243 (1982); J. H. K. Hemm/mki. H. Vainio. M. Sorsa. S. Sal- Gans. Toxicol. Appf. Pharmacol. 63, 312 minen [J. Environ. Sci. Health Cl (No. I). 55 (1982). 0983)]. 28. B. Toth, in Naturally Occurring Carcinogens- 54. N. K. Hooper. 8. N. Ames, M. A. Saleh. J. E. Mutagens and Modulators of Carcinogenesis, Casida. Science 205. 591 (1979). E. C. Miller. J. A. Miller. I. Hirono, T. Sugi- 55. L. W. Wattenberg, Cancer Res. (Suppl.) 43. mura. S. Takayama. Eds. (Japan Scientific 2448s (1983). Societies Press and University Park Press. 56. J. Hoey, C. Montvemay. R. Lambert. Am. J. Tokyo and Baltimore. 1979). pp. 57-65: A. E. Epidemiol. 113. 668 (1981): A. J. Tuyns, G. Ross, D. L. Nagel, B. Toth. /. Agric. Food Chem. 30. 521 (1982); B. Toih and K. Pacil. Pequignot, M. Gignoux. A. Valla. Int. J. Can cer 30. 9 (1982); A. Tuyns, irt Cancer Epidemi Mycopathologia 78, 11 (1982); B. Toth. D. ology and Prevention. D. Schoitenfeld and J. Nagel. A. Rdss. Br. J. Cancer 46. 417 0982). F. Fraumem. Jr.. Eds. (Saunders. Philadel 29. R. Schoental. Toxicol. Lett. 10, 323 (1982): R. phia. 1982), pp. 293-303: R. G. Ziegler et at.. J. J. Hustable. Perspect. Biol. Med. 24, I (1980); Natl. Cancer Inst. 67. 1199 (1981); W. D. H. Niwa. H. Ishiwata. K. Yamada. J. Chroma- Flanders and K. I. Rothman. Am. J. Epidemi togr. 257. 146 (1983). ol. 11S. 371 (1982). 30. M. Cheviot! and T. Navok, Anal. Biochem. 57. E. L. Abel. Hum. Bud. 54. 421 (1982); H. L. 128. 152 0983); V. Lattanzio. V. V. Bianco, D. Rosset. L. Weiner. A. Lee. B. Zuckerman. E. Lafiandra. Experientia 38. 789 11982): V. L. Dooling. E. Oppenheimer. Obstet. Gynecol. Flohe, G. Niebch, H. Reiber, Z. Klin. Chem. 61. 539(19831. Klin. Biochem. 9. 431 0971|; J. Mager. M. 57a.R. A. Anderson. Jr.. B. R. Willis. C. Oswald. Cheviot!. G- Glaser, in Toxic Constituents of Plant Foodstuffs. I. E. Liener. Ed. (Academic Press. New York. 1980). pp. 265-294. L. J. D. Zaneveld. J. Pharmacol. Exp. Ther. 225. 479 (1983). 58. H. F. Stich and M. P. Rosin. Int. J. Cancer 31, 31. J. K. Dunnick eta/.. Fundam. Appl. Toxicol. 2. 305 (1983). 114 0 982). 32. L. C. Berardi and L. A. Goldblatt. in Toxic Constituents of Plant Foodstuffs. I. E. Liener. 59. R. P. Bird. H. H. Draper. P. K. Basrur. Mutat. Res. SOI. 237 (1982); St. A. Campbell and A. G. Fantel. Life Sci. 32. 2641 (I9B3). Ed. (Academic Press, ed. 2. New York. 1980). 60. V. J. Feron. A. Kruysse. R. A. Woutcrsen. pp. 183-237. Eur. J. Cancer Clin. Oncol. 18. 13 (1982). 33. S- P. Xue. in Proceedings, Symposium on Recent Advances tn Fertility Regulation (Bei 61. T. Suematsu et at.. Alcoholism: Clin. Exp. Res. 5. 427 (1981); G. W. Winston and A. 1. Ceder- jing. 2 to 5 September. 1980). p. 122. baum, Biochem. Pharmacol. 31. 2301 (1982); 34. R. K. Haroz and J. Thomasson. Toxicol. Lett. L. A. Videla. V. Fernandez. A. dc Marinis. N. Suppl. 6. 72 0980). 35. M. Cobum. P. Sinsheimer. S. Segal. M. Bur gos, W. Troll. Biol. Bull. (Woods Hole. Mass.) Fernandez. A. Valenzuela. Biochem. Biopbys. Res. Commun. KM, 965 (1982); T. E. Siege. Res. Commun. Chem. Pathol. Pharmacol. 36. 159. 468 (1980). 287 (1982). 36. C. Campbell, persona) communication. 62. M. G. Cutler and R. Schneider. Food Cosmet. 37. G. D. Manners. G. W. Ivie. J. T. MacGregor. Toxicol. Appl. Pharmacol. 45 . 629 0 978); G. W. Ivie and 0. A. Witzef. in Plant Toxins, vo). I. Encyclopedic Handbook of Natural Toxins. Toxicol. 12. 451 (19741. 63. Y. Taxima. in Environmental Mutagenesis. Carcinogenesis and Plant Biology. E. J. Klekowski, Jr.. Ed. (Praeger. New York. 1982), A. T. Tu and R. F. Keeler. Eds. (Dekker. New vol. I, pp. 68-95- York, in press). 64. P. N. Magee. Ed.. Banbury Report 12. Nitros- 38. J. C. M. Van der Hoeven et al., in Mutagens in atmnes and Human Cancer (Cold Spring Har Our Environment. M. Sorsa and H. Vainio. bor Laboratory. Cold Spring Harbor, N.Y., Eds. (Liss. New York. 1982), pp. 327-338: J. 1982); P. E. Hartman, in Chemical Mutagens. C. M. van der Hoeven. W. J. Lagerwei), J. M. F. J. de Serres and A. Hollaender. Eds. (Ple Bruggeman, F. G. Voragen. J. H. Koeman, J. num. New York. 1982). vol. 7, pp. 211-294-. P. Agric. Food Chem.. in press. E. Hartman. Environ. Mutagen. 5. HI (1983). 39. T. Hirayama and Y. Ito. Prev. Med. 10. 614 65. Committee on Nitrite and Alternative Curing (1981); E. Hecker.y. Cancer Res. Clin. Oncol. Agents in Food. Assembly of Life Sciences. 99. 103 0981). National Academy of Sciences. The Health 40. M. R. Matinow, E. i. Bardana, Jr.. B. Pir- Effects ofNitrate. Nitrite, and N-Nttroso Com ofsky. S. Craig, P. McLaughlin. Science 216. pounds (National Academy Press. Washing 415 (1982). ton. DC.. 1981). 41. B. N. Ames. ibid. 204. 587 0979). "Mutagen" 66. H. F. Stich, P. K. L. Chan. M. P. Rosin. Int. J. will be used in its broad sense to include Cancer 30. 719 0 982): H. F. Stich and M. P. clastogens and ocher DNA-damaging agents. Rosin, in Nutritional and Metabolic Aspects of 42. H. F- Stich and R. H. C. San. Eds.. Short- Food Safety. M. Fnedman. Ed. (Plenum, New Term Tests for Chemical Carcinogens (Spring- York, in press). er-Vertag. New York. 1981). 67. L. J. Kinlen. Br. Med. J. 286. 1081 (198J); D. J. 43. P. Hochstein and S. K. Jain. Fed. Proc. Fed. Fink and D. Kritchevsky. Cancer Res. 41.3677 Am. Soc. Exp. Biol. 40. 133 0981). (1981). 44. D. E. Levin, M- Hollstein. M. F. Christman, 68. C. W. Welsch and C. F. Aylswonh. J. Natl. E. Schwiers. B. N. Ames. Proc. Natl. Acad. Cancer Inst. 70. 215 (19831. Sci. V.S.A. 79. 7445 (1982). Many additional 69. P. Correa. J. P. Strong, W. D. Johnson. P. quinones and aldehydes have now been shown Pizzoiato. W Haenszel, J. Chronic Dis. 35. to be mutagenic. 3D (1982). 45. 5. G. Carmelia. E. J. LaVoie, S. S. Hecht, 70. F. B. Shoriand ft al..J. Agric. Food Chem. 19, Food Chem. Toxicol. 20. 587 (1982). 863 (19811. 46. K. Morimoto. S. Wolff. A. Koizumi. Mutat. 7|. M. G. Simic and M. Karel. Eds.. Autoxidation Res. Lett. 119. 355 0983); T. Sawahata and R. in Food and Biological Systems (Plenum. New A. Neal. Mol. Pharmacol. 23. 453 II9H3). York. 1980). 47. H. Kappus and H. Sies. Experientia SI, 1233 72. W. A. Pryor. Ed.. Free Radicals in Biology (19811. (Academic Press. New York. 1976 to 1982). 48. L- Tikkanen. T. Matsushima. S. Natori. Mu vols. I to 5. tat. Res. 116. 297 0983). 73. H. B. Demopoulos. D. D. Pietronigro. E. S. 49. P. K. Singal. N. Kapur. K. S. Dhillon. R. E. Flamtn. M. L. Seligman. J. Environ. Pathol. Beamish. N. S. Dhalta. Can. J. Physiol. Phar Toxicol. 3. 273 (1980). macol. M. 1390 (1982). 74. F. Bischoff. AuV. Lipid Res. 7, 165 (1969). 1262 SCIENCE. VOL. 221 URL 04003 75. N. L. Petrak/s. L. D. Grac/ike, J. C. Craig, Cancer Re*. 41. 1563 (mi 1. 76. H. S. Black and D. R. Douglas, ibid. 32. 2630 (1972). 77. H. Imai. N. T. Wcrthessen. V. Subramanyam, P. W. LcQuesne. A. H. Soloway. M. Kanisawa. Science 207. 65! (19X0). 78. M. Fcrrali. R. Fulccri. A. Benedetti, M. Comporti. Res. Comman. Chem. Pathol. Pharma col. 30. 90 09X0). 79. N. R. Anman. Adv. Lipid Re*. 7. 245 (1969). 80. J. K. Reddy. J. R. Warren. M. K. Reddy. N. D. Lalwam. Ann. N.Y. Acad. Sci. 386. 81 (1082): J. K. Reddy and N. D. Lalwuni. CRC Crit. Rev. Toiicol.. in press. 81. H. L. Plainc. Genetic* 40. 268 (1955); A. Ito. M. Nailo. Y. Naito. H. Watartabc. Gann 73. 315 0982); G. Speii. W. Vogel. M. Walt. Environ. Mutagen. 4. 135 (m2); H. Tsuda. Jpn. J. Genet. 56. 1 f 19811; N. Hiroia and T. Yokoyama. Gann 72. 811 0981). 82. S. Horie, H. Ishii. T. Suga. J. Biochem. iTokvo) 90. 1691 09811; D. P. Jones. L. Eklow. H. Thor. S. Orremus. Arch. Biochem. Biophvs. 210.505 0981). 83. B. Chance. H. Sies. A. Boveris. Phvtioi. Rev. 59. 527 0979). 84. D. Hannan. Proc. Natl. Acad. Sci. U.S.A. 78. 7124 (1981); in Free Radicals in Biology, W. A. Pryor. Ed. (Academic Press. New York. 1982). vol. 5. pp. 255-275. 85. I. Emerit. M. Keck. A. Levy. J. Femgold. A. M. Michelson. Murat. Res. 103.*165 0 982). 86. C. E. Neat .MS. Thomassen. H. Osmundsen. Biochem. J. 196. 149 0 981); J. Bremer and K. R. Norum. J. Lipid Res. 23. 243 11982); M. S. Thomassen. E. N. Chrtsltanscn. K. R. Norum. Biochem. J. 206, 195 0982): H. Osmundsen. Int. J. Biochem. 14. 905 0 982): J. Norsclh and M. S. Thomassen. Biochim. Biophys. Acta, in press. 87. M. G. Enig. R. J. Munn. M. Keeney. Fed. Proc. Fed. Am. Joe. Exp. Biol. 37.2215119781; J. E. Hunter. /. Natl. Cancer Inst. 69. 3J9 0982): A B. Awad. ibid., p. 320; H. Ruttenberg. L. M. Davidson. N. A. Lillie, D. M. Klurfeld. D. Kritchevsky. J. Nut/. 113. 835 (1983). 88. R. Wood. Lipids 14. 975 (1979). 89. E. N. Christiansen. T. Flatmark. H. Kryvi, Ear. J. Ceil Biot. 26. II (1981). 90. Council for Agncultural Science and Technolo gy, Diet. Nutrition, and Cancer: A Critique (Special Publicaiion 13. Council for Agricultur al Science and Technology. Ames. Iowa. 1982). 91. L. F. Bjeldanes et at.. Food Chem. Toxicol. 20. 357 0982): M. W. Pariza. L. J. Loretz, J. M, Storkson. N. C. Holland. Cancer Res. ISuppl.) 43. 2444s 0983). 92. H. F. Stich. W. Such. M. P. Rosin. W. D. Powrie. Muiat. Res. 91. 129 0981): M. P. Rosin. H. F. Stich. W. D. Powne, C- H. Wu. ibid. 101. 189 0 9821; C.-l. Wei. K. Kilamura. T. Shibamoto. Food Cosmet. Toxicol. 19. 749 0 981). 93. E. Yamasaki and B. N. Ames. Proc. Natl. Acad. Sci. U.S.A. 74. 3555 0 977). 94. R. Baker. A. Arlauskas, A. Bonin, D. Angus. CancerLeu. 16. 81 (1982). 95. D. Schuelzle. D. Cronn. A. L. Cnllenden. R. J. Charlson. Environ. Sci. Technol. 9. 838 097$); G. Gartrel! and S. K. Friedlander. At mos. Environ. 9. 279 (1975); L. D. Kier. E. Yamasaki. B. N. Ames. Proc. Natl. Acad. Sci. U.S.A. 71. 4159 (1974); J. N. Pitts. Jr.. Envi ron. Health Perspeet. 47. 115 (1983). 96. J. E. Vena. Am. J. Epidemiol. 116. 42 0982); R. Cederlof. R. Doll. 8. Fowler, Environ. Health Perspeet. 22, 1 (1978): F. E. Spcizer, ibid. 47. 33 (1983). 97. 8. Brunekreef and J. S. M. Boleij, Int. Arch. Occur. Environ. Health 50. 299 (1982). 98. H. Kasai et al.. Gann 73. 681 0982). 99. V, Armuih and I. Berenblum. Carcinogenesis 2. 977 (1981). 100. S. Fabro. Reprod. Toxicol. I. 2 (1982). 101. D. Trichopoulos, M. Papaposiolou. A. Poly- chronopoulou. Int. J. Cancer 28. 691 (1981): P. Hartge, L. P. Lesher. L. McGowan. R. Hoo ver. ibid. 30. 531 09821; B. MacMahon. Can cer {Brussels] 50.2676 0982): H. S. Cuckle and L. J. Kinlen. Br. J. Cancer 44.760 0 981); R. L. Phillips and D. A. Snowdon, Cancer Res. tSuppl.) 43. 2403s (1983): L. D. Marrett. S. D. Waller. J. W. Meigs. Am. J. Epidemiol. 117. 113 0983): D. M. Weinberg. R. K. Ross. T. M. Mack. A. Paganini-Hill. B. E. Henderson. Cancer (Brussels) 51, 675 0 983). 102. P. E. Hartman. Environ. Mutagen., in press. 103. J. R. Toller. Proc. Natl. Acad. Sci. U.S.A. 77, 1763 0980). 104. I. Fridovich, in Pathology of Oxygen. A. Au- lor. Ed. (Academic Press. New York. 1982). pp. 1-19: L. W. Obcrley. T. D. Oberlev. G. R. Buettner. Med. Hypotheses 6. 249 (1980). 105. B. Halliwell. Cell Biol. hu. Rep. 6. 529 (1982); s.A. 1. Tauber. Trends Biochem. Sci. 7. 411 (1982): A. B. Weiibcrg, A. Wcilzman. M. Dcsrrempcs. S. A. Latt. T. P. Slosvel. N. Engl. J. Med. 30*. 26 (1983). Neutrophils also pro duce HOC), which is both a chlorinating and oxidizing agent. 106. M. A. Trash. E. G. Mimnuugh. T. E. Gram. Biochem. Pharmacol. 31. 3335 11982). 107. L. Flohe? m Free Radicals in Biology. W. A. Pryor. Ed. (Academic Press. New York. 19821. vol. S. pp. 223-254. 108. C. Lind. P. Hochstein. L. Ernstcr. Arch. Bin- chem. Biophys. 216. 178 0982). 1Q9. M. Warholm, C. Gwthcnbcrg. B. Manncrvik. C. on Bahr, Biochem. Awtphvj. Res. Com mon. 98. 512 0 98 0. 110. J. G. Bicri. L- Corash. V. $. Hubbard. N. Engl.J. Med. 308. 1063 0 983). 111. Y. M. Wang et al.. in Molecular Interrelations of Nutrition and Cancer. M. S. Amott. J. van Eys. Y.-M. Wang. Eds. (Raven. New York. 19821. pp. 369-379. 112. C. Beckman. R. M. Roy. A. Sproule. Mmat. Res. 105. 73 11982); M. G. Cook and P. McNa mara. Cancer Res- 40. 1329 0980). 113. K. J. A. Davies. A. T. Quintanilha. G. A. Brooks, L. Packer. Biochem. Biophvs. Res. Commun. 107. 098 (1982). 114. C. S. Foote, in Pathology ofOxygen. A. Autor. Ed. (Academic Press. New York. 1982). pp. 21-44; J. E- Packer. J. S. Mahood. V. O. MoraArellano, T. F. Slater. R. L. Willson. B. S. Wolfenden. Biochem. Biophvs. Res. Commun. 98. 901 0981/: W. Bors. C- Michel. M. Saran. Bull. Eur. P/tvi/opor/ioi. Resp. 17 <5upp(.). 13 (1981). 115. N. I. Krinskv and S. M. Deneke. J. Natl. Cancer Inst. 69. 205 0 982): J. A. Turner and J. N. Prebble. J. Gen. Microbiol. 119. 133 0980): K. L. Simpson and C. 0. Chichester. Anna. Rev. Nutr. I. 351 (1981). 116. G. Rettura. C. Dattagupta. P Listowsky. S. M. Levenson. E. SeiFler. Fed. Proc. Fed. Am. Joe. Exp. Biol. 42, 786 0 983), M. M. Mathews- Roth. 0roJo*v 39. 33 (1982). )17. R. Peto. R. Doll. J. D. Buckley, M. B. Spent. Nature (London) 290. 201 09811: R- 8. Shekelle et al.. Lancet 1981-11. 1185 (1981); T. Hirayama. Stttr. Cancer 1. 67 0979); G. Kvalc. E. Bjelke. J. J. Carl. Im. J. Cancer 31. 397 0 983). 118. W. A. Pryor. M. Tamura. M. M. Dooley. P. 1. Premovic. D. F. Church, in Oxy-Rudicah and Their Scavenger Systems: Cellular and Medi cal Aspects. G. Cohen and R. Greenwald. Eds. (Elsevier. Amsterdam. 1983). vol. 2. pp. 185192: W. A. Pryor. B. J. Hales. P. I. Premovic. D. F. Church. Science 220. 425 0 983). 119. M. M. Mathews-Roth.y. Natl. Cancer Inst. 69. 279 0 982). 120. A. C. Griffin, in Molecular Interrelations of Nutrition and Cancer. M. S. Amolt. J. Vafleys. Y. M. Wang. Eds. (Raven. New York. 19821. pp. 401-408; D. Medina. H. W. Lane.C. M. Tracey. Cancer Res. tSuppl.) 43. 2460s (19831; M. M. Jacobs, Cancer Res. 43. 1646 0983): H. J. Thompson. L. D. Meeker. P. J. Bccci. $. Kokoska. ibid. 42. 4954 11982): D. F. Birt. T. A. Lawson. A. D. Julius. C. E. Runice. S. Salmasi. ibid., a. 4455: C. Witting. U. Wit ting. V. Krieg./. Cancer Res. Clin. Oncol. 1W. 109 (1982). 121. M. Challenge and M. R. Banetje*. Cancer Leu. 17. 187 0 982). 122. W. C. Willett et al.. Lancet, in press. 123- J. T. Salonen, G. Alfthan. J. Pikkarainen, J. K. Huttunen. P. Puska. ibid. 1982-1!. 175 (1982). 124. M. Yonaha, E. Itoh. Y. Qhbayashi. M. Uchiyama, Res. Commun. Chem. Pathol. Pharmacol. 28. 105 (1980): L. i. Kling and J. H. Soares. Jr.. Nutr, Rep. Int. 24. 39 (1981). 125. N. Tateishi. T. Higashi. A. Naruse. K. Hikita, Y. Sakamoto, J. Biochem. (Totvo) 90, 1603 0981). 126. J. M. Williamson. B. Boettcher. A. Meister, Proc. Natl. Acad. Sci. U.S.A. 79. 6246 (1982). 127. CME Symposium on "/V-Acetylcystcine (NAC): A Significant Chemoproiective Ad junct," Sem. Oncol. 10 (SuppJ. I). ) (1983). 128. J. A. Hinson. L. R. Pohl. T. J. Monks. J. R. Gillette. Life Sci. 29, 107 0 981). 129. A. M. Novi. Science 212, 541 (1981). 130. W. B. Dunham et al.. Proc. Natl. Acad. Set. U.S.A. 79. 7532 (1982); G. Kallistratos and E. Fasske, J. Cancer Res.'Clin. Oncol. 97. 91 (1980). 131. S. Wassertheil-Smoller et aL. Am. J. Epidemi ol. 114.714(1981). 132. B. N. Ames. R. Cathcart, E. Schwiers. P. Hochstein. Proc. Natl. Acad. Sci. V.S.A. 78, 6858 0 981). 133. A. Nomura. L. K. Heilbrun, G. N. Sicmmermann, in preparation. 134. J. N. Boyd. J. G. Babish. G. S. Stocwsand, Food Chem. Toxicol. 20. 47 (1982). 135. A, W. Wood, el al.. Proc. Natl. Acad. Sci. U.S.A. 79. S313 0982). 136. T. H. Maugh II. Science 205. 1363 (1979) (interview with John Higginsonl. 137. H. L. Genslcr and H. BcrnMcin. Q. Rev. Biol. 6, 279 (1981). 138. A. L. Tappcl. in Free Radicals in Biolngv. W. A. Pryor. Ed. (Academic Press. New York. )9Kfl). vol. 4. pp. 1-47. 139. R. S. Sohal. in Age Piaments. R. S. Sohal. Ed. (Elscvier/Nonh-Holland. Amsierdam. 1981), pp. 303-316. 141^-0. E. Fleming, J. Miquel. S. F. Cottrell. L- S. Ycngoyan. A. C. Economos. Gerontology 28. 44 0982). 141. 1. M. TolmasofF. T. Ofto. R. G. Cutler. Proc. Natl. Acad. Sci. U S.A. 77. 2777 0 980): R. G. Cutler. Gerontology, in press: J. L. Sullivan. ibid. 28 . 242 0 982). 142. R. Peto. Proc. R. Soc. London Ser. B 205. 1) 1 11979); D- Dix. P. Cohen. J. Flannery. J. Theor. Biol. 83. 163 0980). 143. R. G. Cutler, in Testing the Theories of Aging. R. Adelman and G. Roth. Eds. (CRC Press. Boca Raton. Fla., in press). 144. D. Hersh. R. G. Cutler. B. N. Ames, in prepa ration. 145. E. Boyland, in Health Risk Analysis (Franklin Institute Press. Philadelphia. 1980). pp. 181-- 193; E. Boyland. in Cancer Campaign, vol. 6. Cancer Epidemiology, E. Grundmann, Ed. (Fischer. Stuttgart. 19821. pp. 125-128. 146. 2. L. Marx. Science 219. 158 (1983). 147. B. D. Goldstein. G. Witz. M. Amoruvo. D. S. Stone. W. Troll. Cancer Leu. 11. 257 0 981); W. Troll, in Environmental Mutagens and Car cinogens. T. Sugimura. S. Kondo. H. Takebe. Eds. (Univ, of Tokvo Press. Tokyo, and Liss, New York. 1982). pp. 217-222. 148. B. N. Ames. M. C. Hollstein. R. Cathcart. in Upid Peroxide in Biology and Medicine. K. Yagi. Ed. (Academic Press. New York. (9821, pp. 339-351. 149. 1. Emerit and P. A. Cerutti. Proc. Natl. Acad. Sci. U.S.A. 79. 7509 (1982); Nature (LondonI 293. 144 0981); P. A. Cerutti. 1. Emerit. P. Amstad. in Genes and Proteins in Oncogene sis. I. B. Weinstein and H. Vogel. Eds. (Aca demic Press. New York, in press); I. Emerit. A. Levy. P. Cerutti. Muiat. Res. 110. 327 09831. 150. J. Funes and M. Karel. Lipids 16. 347 11981). 151. S. J. Stohs. M. Q. Hassan. W. J. Murray. Biochem. Biophys. Res. Commun. 111. 854 0983). 152. C. C. Reddy. R. W. Scholz. E. J. Massaro, Toxicol. Appl. Pharmacol. 61. 460 (f98l). 153. H. Nagasawa and J. B. Lmle. Carcinogenesis 2. 601 (1981); V. Solanki. R. S. Rana. T. J. Slaga. ibid., p. 114|; T. W. Kensler and M. A. Trash. Cancer Res. 41. 216 0981). 154. R. H. Simon. C. H. Scoggin. D. Patterson./. Biol. Chem. 256. 7181 0981); T. S. Argyris and T. J. Slaga. Cancer Res. 41. 5193 (1981). 155. G. E. Hatch. D. E. Gardner, D. B. Menzel. Environ. Res. 23. 121 (1980). 156. A. J. P. Klcin-Szanto and T. J. Slaga, /. Invest. Dermatol. 79. 30 0 982). 157. C. C. Weddle, K. R. Hombrook. P. B. McCay, /. Biol. Chem. 251. 4973 (1976). 158. A. G. Lurie and L. S. Culler. J. Natl. Cancer Inst. 63. 147 0 979). 159. C. Borek and W. Troll. Proc. Natl. Acad. Sci. U.S.A. 80. 1304 (1983): C. Borek. in Molecular Interrelations of Nutrition and Cancer, M. S. Amott. J. van Eys. Y.-M. Wang, Eds. (Raven, New York. 19821. pp. 337-350. 160. T. J. Slaga r al.. in Carcinogenesis: A Com prehensive Treatise (Raven. New York, 1982), vol. 7. pp. 19-34; K. Ohuchi and L. Levine, Biochim. Biophys. Acta 619, 11 (1980); S. M. Fischer. G. D. Mills. T. J. Slaga. Carcinogene sis 3. 1243 (1982). 161. R. P. Mason, in Free Radicals in Biology. W. A. Pryor. Ed. (Academic INess. New York, 1982). vol. 5. pp. 161-222. 162. G. McLennan. L. W. Oberley. A. P. Autor, Radial. Res. 84. 122 (1980). 163. J. M. Pany, E. M. Parry. J. C. Barrett. Nature (London) 294. 263 (1989); A. R. Kinsella, Carcinogenesis 3. 499 (1982). 164. H. C. Birnboim, Can. J. Phvsinl. Pharmacol. 60. 1359 (1982). 165. A. Varshavsky. CW/25, 561 0981). 166. T. J. Powles er al.. Eds, Prostaglandins and Cancer: First International Conference (Liss. New York. 1982). 23 SEPTEMBER 1983 1263 167. J. E. Trosko. C.-C. Chang. A. Medcalf. Cancer Invest., in press. 167a.I. B. Weinstein, Mature {London) 302. 750 (1983). 168. i. A. Bond, A. M. Gown. H. L. Yang, E. P. Benditt. M. R. Juchau. /. Toxicol. Environ. Health 7, 327 (19811. 169. Editorial. Lancet 19804. 964. (19801; K. Yagi. H. Ohkawa, N. Ohishi. M. Yamashita. T. Nakashima. J. Appl. Biochem. 3, 58 (19811. 170. J. M. C. Guttcridge. B. Halliwell. D. A. Row. ley. T. Wcstermarck, Lancet 1982-11. 459 (19821. 171. J. E. Cleaver, in Metabolic Basis of Inherited Disease. J. B. Stanbury. J. B. Wyngaarden, D. S. Fredrickson. J. L. Goldstein. Eds. (McGraw-Hill, ed. 5, New York, 19831, pp. 1227-12J0. 172. K. C. Bhuyan. D. K. Bhuyan. S. M. Podos. IRCS Med. Sci. 9. 126 (I98J); A. Spector. R. Scotto. H. Weissbach. N. Brot. Btochent. Biophys. Res. Comtnun. 108. 429 119821; S. D. Varma. N. A. Beachy. R. D, Richards. Photochem. Phatobial. 36. 623 (19821. 173. M. L. Katz. K. R. Parker, G. J. Handelman. T. L. Bramel. E. A. Dratz. Exp. Eve Res. 34. 339 (19821. 174. D. Behne, T. Hofer, R. von Berswordt-Wall- rabe. W. Eiger. J. Mute. 112. 1682 (19821. 175. B. N. Ames. -L. S. Gold, C. B. Sawyer. W. Havender. in Environmental Mutagens and Carcinogens. T. Sugimura. S. Rondo. H. Ta kebe. Eds. (Univ. of Tokyo Press. Tokyo, and Liss. New York. 1982). pp. 663-670. 176. National Center for Health Statistics. Advance Report. Final Mortality Statistics. 1979. Monthly Vital Statistics Report 31. No. 6. suppl. (DHHS publication (PHS) 82-1120. (Public Health Service. Hyansville, Md.. I982[; Metropolitan Life Insurance Company Actuarial Tables. April 1983. 177. B. A. Bridges. B- E. Butterworth, I. B. Wein stein. Eds.. Banbury Report 13. Indicators of Cenotoxic Exposure (Cold Spring Harbor Lab oratory. Cold Spring Harbor. N.Y., 198!); R. Montesano. M. F. Rajewsky. A. E. Pegg. E. Miller, Cancer Res. 42. 5236 (19821; H. F. Stich. R. H. C. San. M. P. Rosin, Ann. M.Y. Acad. Set-, in press: L B. Weinstein. Annu. Rev. Public Health 4. 409 (1983). 178. I am indebted to G. Ferro-Luzzi Ames. A. Blum. L. Gold, P. Hartman. W. Havender. N. K. Hooper, G. W. Ivic, J. McCann. J. Mead. R. Olson. R Peto. A. Tappel. and numerous other colleagues for their criticisms. This work was supported by DOE contract DE-AT0376EV70I56 to B.N.A. and by National Insti tute of Environmental Health Sciences Center Grant ES0I896. This article has been expanded from a talk presented at the 12th European Environmental Mutagen Society Conference. Espoo. Finland. June 1982 Jin A/i/t/iveiwwi Our Environment, M. Sorxa and H. Vainio. Eds. (Lrss. New York. I982)|. I wish to dedicate this article to the memory of Philip HanJter. pio neer in the field of oxygen radicals. URL 04004 RESEARCH ARTICLE Imaging Dopamine Receptors in the Human Brain by Positron Tomography Henry N. Wagner, Jr., H. Donald Burns. Robert F. Dannals Dean F. Wong, Bengt Langstrom, Timothy Duelfer, J. James Frost Hayden T. Ravert. Jonathan M. Links. Shelley B. Rosenbloom Scott E. Lukas, Alfred V. Kramer, Michael J. Kuhar cose, ["Clcarboxyhemoglobin, ionic rubidium-82, MGa-tabeled EDTA, and oth er radiopharmaceuticals, and subsequent imaging of the distribution of the radio active label in the brain by means of the tomographic method, based on detection of the annihilation radiation produced during positron emission (9). The butyrophenone neuroleptic drug spiperone has been useful in binding studies for measuring dopamine recep tors both in vitro (10) and in vivo (11). We now report initial results obtained with 3-;V-[llC]methylspiperone (nCNMSP). a spiperone derivative, in PET One of the most intriguing problems in as a result of neuroleptic therapy (4). biomedical research today is that of re The development of positron emission lating manifestations of neuropsychiatric tomography (PET) and appropriate ra scanning studies to visualize the distribu diseases to chemical processes in differ dioactive tracers labeled with positron- tion of dopamine receptors in the brains ent parts of the brain. The neurotrans emitting radionuclides has now made it of baboons and a human being. All stud ies were performed with a NeuroECAT scanner (Ortec. Inc., Oak Ridge. Ten Abstract. Neurotransmitter receptors may be involved in a number of nettropsy nessee). which has a spatial resolution of chiatric disease states. The ligand 3-N-[nC]methylspiperone. which preferentially approximately 8 mm (full width at half binds to dopamine receptors in vivo, was used to image the receptors by positron maximum) in the plane of the slice. The emission tomography scanning in baboons and in humans. This technique holds distance between slices is 3 cm. promise for /rom/rvtuive clinical studies of dopamine receptors in humans. The newly developed tracer "C- NMSP was synthesized by <Valkylation of spiperone with ["'Clmethyl iodide: the mitter dopamine appears to be associat possible to relate regional biochemistry iodide was produced from uCO;. which ed with abnormalities related to disor within the human brain to measurements in tum had been produced with an in- ders such as Parkinson's disease and of behavior in normal subjects and to hospital cyclotron (model RNP-16, Scan- schizophrenia. The highest density of elucidate abnormalities in patients with ditronix Cyclotron. Sweden). Carbon-11 dopamine neurons occurs in the nigro- Alzheimer's disease (5), Huntington's is a positron-emitting isotope with a striatal dopamine pathway which degen disease (6), depression (7), and multiple physical half-life of 20 minutes. The en erates in Parkinson's disease (1). Neuro infarct dementia (5). The technique con tire synthesis was accomplished with leptic drugs elicit extrapyramidal parkin sists of intravenous injection of a sub material ready for injection within 55 sonian side effects by blocking dopamine stance such as ,8F-labeled deoxyglu- minutes after the end of the cyclotron receptors in the corpus striatum and also exert antischizophrenic action by block ing dopamine receptors, perhaps in lim bic areas (2). Numbers of dopamine re ceptors are increased by chronic neuro leptic treatment (J) and are also in creased in some schizophrenics, perhaps Henry N. Wagner. Jr., H. Donald Burns. Robert F. Dannals, Dean F. Wong. Timothy Duclfcr. J. James Frost. Hayden T. RavcR, Jonathan M. Links, and Alfred V. Kramer are in the Division of Nuclear Medicine. Johns Hopkins Medical Institutions. Baltimore. Maryland 21205. Bengt Langstrom is in the Institute of Chemistry. University of Uppsala. S75121 Uppsala 1. Sweden. Shelley B. Rosenbloom is in the Division of Neuroradiology. Johns Hopkins Medical Institutions. Scott E. Lukas is at the NIDA Addiction Research Center. Baltimore City Hospitals. Baltimore. Maryland 21224. Michael J- Kuhar is m the Departments of Neuroscience. Pharmacology and Experimental Therapeutics, and Psychiatry and Behavioral Sciences. Johns Hopkins University School of Medicine. Baltimore. Maryland 21205. Correspondence should Pc sent to Henry N. Wagner. Jr. Natural carcinogenic products Not all cancer-causing substances are man-made. Nature does a fair share ofproducing compounds that are potential toxins when ingested; fortunately, many of these toxins are weak carcinogens URL 0^005 the cobalt miners of the Schneeberg in uranium and its daughters is combined j i ii Elizabeth K. VVeisburger National Cancer Institute Bethesda, Md. 20014 Saxony, was described by Paracelsus in 1531, and by Agricola and others who called it "Bergkrankheiten" (mountain disease). At that time the with smoking. Thus, an occupational disease related to a "natural" product was first noted almost 450 years ago. i disease was mistaken for tuberculosis Culprits: natural organics There are many statements in the- - of the lung. This ailment, which also Other naturally occurring inorganic popular press that cancer results from occurred in the miners working in the compounds are associated with an in exposure to pollutants emanating from Joachimsthal of eastern Germany, was crease in cancer rate. Examples are the increase in technological develop caused by exposure to uranium, radi arsenic and several other metals ments. However, certain historical um, and their daughter products present in drinking water in certain documents indicate that lung cancer, present in the ores mined in these lo areas of the world. However, naturally for example, often associated with air cations. occurring organic compounds are pollution or excess smoking, was also It is currently evident that modern much more prevalent among the lists known in Germany during the Middle miners on the Colorado Plateau en of natural carcinogenic products. Ages. counter the same risks as the German A research group showed in 1942 Lung cancer, which occurred among miners, especially when exposure to that crude ergot was tumorigenic in 278 Environmental Science & Technology This article not subject to U.S. Copyright. Published 1979 American Chemical Society URL 04006 animals fed this compound for six months or more. Since the tumors generally regressed when the ergot was discontinued, this was indicative of a relatively benign effect. The constitu ents of ergot responsible for this action have not been investigated In the early 1950's, it was reported that subcutaneous injections of tannic acid or tannins induced an appreciable number of liver tumors in both rats and mice. The data were somewhat con tradictory and have not been fully confirmed, especially in view of the somewhat artificial method of ad ministration used. For this reason, the carcinogenicity of tannins and tannic acid for a route pertinent to human exposure is questionable. During the 1960s, it was first re ported that safrole caused liver tumors in animals fed the substance at high dose levels. Safrole, which occurs in many natural oils such as anise, cam phor or sassafras oil, was previously allowed as a synthetic flavoring agent in beverages and foods. As a result of the tumor-inducing properties of safrole, reexamination of the GRAS (Generally Recognized as Safe) list, which includes safrole, was deemed necessary by the FDA. Safrole is partially metabolized by hydroxylation to l'-hydroxysafrole, which is further metabolized to l-acetoxysaf- role, which is considered the likely active carcinogenic intermediate. In an attempt to identify the agent responsible for the high incidence of amyotrophic lateral sclerosis in Guam, the National Institute of Neurological Diseases and Blindness tested various plant products used on that island, in cluding the nuts of the cycad plant. These nuts were used as an emergency starch when typhoons had destroyed other food sources. Many different varieties of the cycad plant (Cycas circinalis), which Is a member of an ancient plant family, grow in tropical or subtropical areas. The nuts, about the size of a lemon, and the outer husk are the parts used in Guam. The husk is chewed by peo ple working in the fields; the nut is used as a starch after it is soaked in water to remove the toxin. The toxic material is cycasin, the glucoside of methylazoxymethanol, which is the active moi ety. Cycasin causes liver, kidney and intestinal tumors in rats and other laboratory animals. It also has neuro logical effects since cattle grazing on cycads, often the only green forage in times of drought, may show paraly sis. In Kenya, the kernels of the plant Encephalartos, also a member of the cycad family, are consumed both by humans and by cattle. In this very large kernel, the toxic compound is macrozamin, which is also carcino genic in rats, causing kidney, liver and lung tumors similar to the effect of cycasin. Cycasin. The nuts and husks (above) of the cycad (below) plant, which contain the toxin cycasin, are used as food by the people of Guam; husks are chewed, and the nuts are soaked and used as a starch Peanuts. Moldy peanuts contain the deadly toxin aflatoxin Edible but toxic In many areas of the world poison ing of domestic animals by plants of the Senecio, Heliotrope and Crotalaria species has been reported. Senecio species are quite widespread, and some forms, such as "golden rag wort" or Senecio aureus, are found in North America. These aster-like plants are quite toxic to the liver be cause detoxification of ammonia from dietary nitrogen is blocked. Thus grazing animals which eat these plants often die from ammonia intoxica tion. In many African countries the Senecio plants, especially S. jacobaea, are used as herbal medicines despite their toxicity. Most of these plants contain one or the other of the pyrrol- izidine alkaloids such as retrorsine, isatidine and lasiocarpine. Honey made from such plants may be con taminated by these alkaloids, some of which are also liver carcinogens in rats. In Japan, plants known as coltsfoot are used as herbal remedies or as veg etables. Several of these are also car cinogenic to the livers of rats. Appar ently, one compound responsible is an alkaloid, petastenine, which has a structure similar to that of the pyrrolizidine alkaloids isolated from the Senecio plants. Also, there is a recent report that a plant called comfrey, which is used as a poultice or as a tea, has a carcinogenic effect in animals. One may question the use of herbal remedies or sassafras tea in modern society. Nevertheless, until several years ago one of the materials used to flavor vermouth, a constituent of popular drinks such as martinis or Manhattans, was a root of an Indian (Jammu) variety of calamus. The American variety of this plant is Acorus calamus or sweet flag. This plant was used by the American Indi ans and others as a flavoring agent or folk medicine. In 1967 it was found that j3-asarone, about 85% of the fla voring material in Jammu calamus, caused intestinal tumors in rats. Thus the use of calamus as a flavoring ma terial for American vermouth was no longer allowed by the FDA. Because of certain structural similarities between safrole and 3-asarone, one might as sume analogous pathways for activa tion of /3-asarone and safrole. Volume 13. Number 3. March 1979 279 Another compound with similar features is estragole, a component of oil or estragen or oil of tarragon, ob tained from the tarragon plant, Art emisia dracunculus. Recent studies have shown that estragole has a defi nite but weak liver carcinogenic effect in young male mice. Since tarragon is often used as a flavoring agent in gourmet types of vinegar, naturally occurring carcinogens may be con sumed by those living sophisticated lifestyles. Another gourmet-type food is a wild mushroom, the false morel (Cyromitra esculenta). This mushroom contains N-methyl-N-formylhydrazine, which causes several types of tumors when given to mice in their drinking water for their lifespan. Active ingredient unknown In the previously mentioned plants, the carcinogenic components have generally been identified. However, there are other carcinogenic plants in which the active component has not been characterized. One of these is the bracken fern, a common inhabitant of open woods and waste places. The young shoots, called the fiddleheads or croziers, are eaten as a springtime delicacy in many countries. Cattle that often graze on bracken fern are found to have bladder or stomach cancer; this is especially so in Turkey, Scotland and Wales. To some extent the effect of bracken fern in experimental animals can be inhibited by simultaneous administration of the enzyme-inducer phenothiazine. Various compounds have been pro posed as the active component in bracken fern. However, numerous ef forts to fractionate and characterize the active material(s) have not been productive. Of all carcinogenic plants, the one with the greatest exposure is probably the tobacco plant. There continues to be much controversy over the excess risk of lung and bladder cancer from smoking cigarettes. Since tobacco smoke contains dimethylnitrosamine, benzo(a)pyrene and numerous other known carcinogens, the probability of a risk from continued long-term ex posure would seem high. The gas phase contains nitrosamines, hydrazine, vinyl chloride, and several other toxic materials. The particulate phase contains numerous polycylic aromatic hydrocarbons of varying carcinogenic potency. In ad dition, some of these substances, at least in animal experiments, have an organ-specific effect, especially for the esophagus, lung, pancreas and blad der. Further studies indicate that at least four nitrosamines are present in un burned tobacco itself. One of these, N-nitrosodiethano!amine, was found previously to induce liver cancer in rats. Endogenously formed carcinogens may also represent a hazard. There has been much publicity and controversy lately over N-nitroso compounds formed in the digestive system from nitrite and secondary or tertiary amines. Another case of interest is the for mation of ethyl carbamate or urethane in foods or- beverages processed by fermentation. Ethyl carbamate is a weak carcinogen but does cause lung tumors in mice, or liver tumors in mice or newborn rats, although adult rats are resistant to its action. Foods and beverages such as beer, wine, bread, olives and yogurt all-contain measur able levels of the compound, which is probably formed by reaction of en dogenous carbamyl phosphate with ethanol to yield ethyl carbamate. Microbes make carcinogens Even microorganisms can synthe size carcinogens. Some of these toxins, when ingested, affect more people than any other unintentionally ingested natural carcinogen. Fungus. Aspergillus fiavus on moldy grains produces afiatoxin. which is a potent liver carcinogen In 1960, an outbreak in England of a fatal liver disease in turkey poults was traced to peanut meal in their feed which had been infested with the fun gus Aspergillusfiavus. This fungus is capable of producing some very toxic compounds such as afiatoxin Bi and B2, lesser quantities of afiatoxin Gi and G2, and several other compounds which have not been thoroughly in vestigated. Afiatoxin B is one of the most po tent liver carcinogens in rats or rain bow trout. In certain strains of rats fed 0.1 ppm of afiatoxin Bj for 50-80 weeks there was a 50% incidence of liver cancer; even 0.015 ppm was ef fective after 68-80 weeks. Rainbow trout were affected by levels as low as 0.1 ppb. On the other hand, adult mice are quite resistant to afiatoxin. The fungus can contaminate many types of grains or oil seeds such as rice, corn, cottonseed meal, and copra. Cows fed meal containing afiatoxin Bi or B; excrete afiatoxin Ms or M2 in the milk. These compounds represent de toxification products; they are much less toxic to rats, and do not induce tumors in rats under conditions where afiatoxin does. In warm, humid areas of the world the afiatoxin contamination of food crops can often be correlated with liver cancer incidence in the population. Data supporting this supposition have been obtained from Swaziland, Kenya and Thailand. Another compound, similar to af iatoxin, is sterigmatocystin, produced by Aspergillus versicolor or Asper gillus nidulans. Although this com pound is carcinogenic in rats and mice, it is considered for use as a chemo therapeutic agent in the treatment of neoplastic (tumor-related) diseases. Certain other microorganisms, however, are capable of synthesizing quite unique structures. Elaiomycin, produced by Streptomyces hepaiicus has an azoxy structure similar to that of cycasin, and has weak carcinogenic activity in rats. Griseofulvin from Penicillium griseofulvum is useful in the treatment of fungal skin infections. This compound has induced liver tu mors and other liver-specific effects such as cirrhosis and protoporphyrin in mice. Streptozotocin, from Streptomyces achromogenes possesses an N-nitrosourea structure, a most un usual type of compound. While this compound induces kidney and lung tumors in rats, it is a useful drug in treatment of neoplastic disease. Still other mycotoxins which may exhibit some carcinogenicity are luteoskyrin (from P. islandicum) and Actinomycin D (from S. chrysomallus or S. antibioticus). Precautions to take What measures can be taken to protect against some of these naturally occurring carcinogens? As an exam ple, afiatoxin can be considered. URL 04007 280 Environmental Science & Technology URL 04008 Naturally occurring carcinogens Toxin Source - Speclee affected J.-` -33 Organ A cj developing ' v tumors l Actinomycin D Aflatoxin Bt Aflatoxin B2 Aflatoxin G, Aflatoxin G2 Asarone Cycasin Elaiomycin Ergot (crude) Estragofe Ethyl carbamate Griseofuivin Luteoskyrin Macrozamin < Streptomyces chrysomatius Mouse or S. antibioticus Rat !Local reaction Aspergillus flavus from moldy grains Man? 1Liver Monkey 1Liver Mouse ILiver (infant) Rainbow Liver trout - Rat Calamus root Rat v: Cycads L Guinea pig " Mouse ..Rat .; 1 Streptomyces hepaticus ' Rat / Claviceps purpurea , fungus on rye <- , Rat ' . I Oil of tarragon : Mouse l Fermented products . Mouse Penicillium griseofulvum Mouse Penlcillium islandicum ' > Mouse >Rat.` t Encephalartos'and zamia plants : Rat N-Methyl-N-formyl. hydrazine False morel - .. Mouse Nitrosamines r Tobacco : Rat > Petastenine ' Coltsfoot plant Rat ':.. Pyrrolizidine alkaloids Senecio plants r Rat Safrole Natural plant oils -f, -y. .<! Mouse -ji Rat Liver Sterigmatocystin Streptozotocin Aspergillus versicolor - 'Mouse -V ,V -iffi'Rat ` , -'' Streptomyces achrorpo- ;^RafV- - - genes rl'y:-. ,-f^v. Unknown ; ^Bracken fern \.vv'If-.'.' Cattle :>: Mouse -Intestine '.'JcRat Lung G rowth of the fungus on seed crops can be inhibited by proper harvesting, by preventing insect damage, by avoiding damage to the kernels and by ensuring proper drying and storing. Many simple compounds such as p-aminobenzoic acid; potassium sul fite; potassium fluoride; and acetic, citric and lactic acids inhibit the growth of the fungus or production of the toxin. Furthermore, an extensive search by the Southern Regional Re search Laboratory has shown that aflatoxin B| can be inactivated by vari ous treatments. A promising one involves treatment with ammonia (at two or three atmo spheres of pressure) to cleave the lac tone ring of aflatoxin with the loss of one carbon. Further degradation can occur with complete loss of two rings in the aflatoxin molecule. Compounds of this type are not carcinogenic when tested under conditions where afla toxin B| is active. Furthermore, the ammonia treatment does not seem to affect the protein level of treated pea nut Tneal, for example. Thus the pro cess offers a means to salvage large lots of contaminated oil seeds for animal feed. Although there are many naturally occurring carcinogens in the environ ment--except for tobacco and afla toxin--data relating these substances to cancer in people are lacking. How ever, the normal mammalian detoxi fication systems can usually handle small quantities of many toxins fairly well, especially if the nutritional quality of the diet is adequate. Fur thermore, repair enzymes are available to handle isolated damage. Except for aflatoxin and cycasin, many of the naturally occurring car cinogens are relatively weak. Also, many of our foodstuffs contain sub stances that can inhibit the action of thesd carcinogens. Vegetables such as brussels sprouts, alfalfa, and cabbage contain certain indole derivatives that protect against carcinogens. Thus, a moderate, nutritionally adequate diet affords protection against some toxins. In addition, the avoidance of obviously deleterious materials such as tobacco or moldy foods is advantageous in protecting one against naturally occurring carcino gens. Additional reading Sunderman, F. W., Jr., Carcinogenic ef fects of metals, Fed. Proc.. 37, 40 (1978). ' Wogan, G. N., Naturally occurring car cinogens, In The Physiopathology of Cancer. Vol. 1. Biology and Biochemistry. Karger, Basel, 1974, p 649. Miller, J. A., Miller, E. C., Carcinogens occurring naturally in foods. Fed. Proc.. 35, 1316(1976). Toth, B., Nagel, D., Tumors induced in mice by N-methyl-N-farmylhydrazine of the false morel GyromUra esculenta. J. Nat. Cancer Inst., 60, 201 (1978). Schmeltz, I., Hoffmann, D., Nitrogencontaining compounds in tobacco and to bacco smoke, Chem. Rev., 77, 295 (1977). Ough, C. S., Ethylcarbamate in fermented beverages and foods: Naturally occurring ethylcarbamate, J. Agric. Food Chem., 24, 233 (1976). Brekke, 0. L., Sinnhuber, R. O., Peplinski, A. J., Wales, J. H., Putnam, G. B,, Lee, D. J., Ciegler, A., Aflatoxin in corn: Ammonia inactivation and bioassay with rainbow trout, Appl. Environ. Microbiol., 34, 34 (1977). Elizabeth K. Weisburger is chief. Labora tory of Carcinogen Metabolism. Division ofCancer Cause and Prevention. National Cancer Institute. Dr. Weisburger is a consultant to the National Academy of Sciences/National Research Council, and is assistant editor-in-chiefofthe Journal of the National Cancer Institute. Coordinated by LRE Volume 13, Number 3, March 1979 281 -----***^*---^ --- -------------- ------ -- til URL 04009 /up amounts of safrole and targe amounts (close to 10 percent by weight) of the closely related compound piperine (26). Extracts of black pepper cause tumors in mice at a variety of sites at a dose of Dietary Carcinogens and extract equivalent to 4 mg of dried pep per per day (about 160 mg/kg per day) for , Anticarcinogens 3 months; an estimate of the average human intake of black pepper is over 140 mg per day (about 2 mg/kg per day) for Oxygen radicals and degenerative diseases life (26). 2) Most hydrazines that have been Bruce N. Ames tested are carcinogens and mutagens, and large amounts of carcinogenic hy drazines are present in edible mush rooms. The widely eaten false morel (Cyromitra esculenta) contains 11 hydra Comparison of data from different Naturaf Mutagens and zines. three of which are known carcino countries reveals wide differences in the Carcinogens in Food rates of many types of cancer. This leads gens (28). One of these. N-methyl-;Vformylhydrazine, is present at a concen to hope that each major type of cancer Plant material. Plants in nature syn tration of 50 mg per 100 g and causes may be largely avoidable, as is the case thesize toxic chemicals in large amounts, lung tumors in mice at the extremely low for cancers due to tobacco, which consti apparently as a primary defense against dietary level of 20 jig per mouse per day tute 30 percent of the cancer deaths in the hordes of bacterial, fungal, and in (25). The most common commercial the United States and the United King sect and other animal predators (5--*0). mushroom. Agaricus bisporus. contains dom (/). Despite numerous suggestions Plants in the human diet are no excep about 300 mg of agariiine. the &-g!utamyl to the contrary, there is no convincing tion. The variety of these toxic chemi derivative of the mutagen 4-hydroxy- evidence of any generalized increase in U.S. (or U.K.) cancer rates other than what could plausibly be ascribed to the Summary. The human diet contains a great variety ot natural mutagens and delayed effects of previous increases in carcinogens, as well as many natural antimutagens and anticarcinogens. Many of tobacco usage U-3). Thus, whether or these mutagens and carcinogens may act through the generation of oxygen radicals. not any recent changes in life-style or Oxygen radicals may also play a major role as endogenous initiators of degenerative pollution in industrialized countries will processes, such as DNA damage and mutation (and promotion), that may be related substantially affect future cancer risks, to cancer, heart disease, and aging. Dietary intake of natural antioxidants could be an some important determinants of cutTent important aspect of the body's defense mechanism against these agents. Many risks remain to be discovered among antioxidants are being identified as anticarcinogens. Characterizing and optimizing long-established aspects of our way of such defense systems may be an important part of a strategy of minimizing cancer life. Epidemiologic studies have indicat and other age-related diseases. ed that dietary practices are the most promising area to explore (/, 4). These studies suggest that a general increase in cals is so great that organic chemists methylphenythydrazine. per 100 g of consumption of fiber-rich cereals, vege have been characterizing them for over mushrooms, as well as smaller amounts tables, and fruits and decrease in con 100 years, and new plant chemicals are of the closely related carcinogen .V-acc- sumption of fat-rich products and exces still being discovered (12. 24. 25). How tyl - 4 - hydroxymethylphenylhydrazine sive alcohol would be prudent (/. 4). ever. toxicological studies have been (28). Some agaritine is metabolized by There is still a lack of definitive evidence completed for only a very small percent the mushroom to a diazonium derivative about the dietary components that arc age of them. Recent widespread use of which is a very potent carcinogen (a critical for humans and about their mech short-term tests for detecting mutagens single dose of 400 ng/g gave 30 percent of anisms of action. Laboratory studies of (41. 42) and the increased number of mice stomach tumors) and which is also natural foodstuffs and cooked food are animal cancer tests on plant substances present in the mushroom in smaller beginning to uncover an extraordinary- (6) have contributed to the identification amounts (25), Many hydrazine carcino variety of mutagens and possible carcin of many natural mutagens, teratogens, gens may act by producing oxygen radi ogens and anticarcinogens. In this article and carcinogens in the human diet (5- cals (43). I discuss dietary mutagens and carcino 40). Sixteen examples are discussed be 3) Linear furocoumarms such as psor gens and anticarcinogens that seem of low. alen derivatives are potent light-activat importance and speculate on relevant 1) Safrole. estragole. rnethyletigenol. ed carcinogens and mutagens and are biochemical mechanisms, particularly and related compounds are present in widespread in plants of the Umbclliferue the role of oxygen radicals and their many edible plants (5). Safrolc. cstra* family, such as celery, parsnips, figs, inhibitors in the fat-cancer relationship, gole, and methyleugenol are carcinogens and parsley (for instance. 4 mg per 100 g promotion, anticarcinogenesis, and ag in rodents, and several of their metabo of parsnip) (17, 19. 44). The level in ing. lites are mutagens (5). Oil of sassafras, celery (about 100 ^.g per 100 g) can The author is chairtnnn of the Department of Biochemistry. University of California. Berkckv 94720. which had been used in ''natural'* sarsa parilla root beer, is about 75 percent safrole. Black pepper contains small increase about 100-fold if the celery is stressed or diseased 1/9). Celery pickers and handlers commonly develop skin 1256 SCIENCE. VOL. 221 rashes on their arms when exposed to diseased celery (/9). Oil of bergamot, a citrus oil. is very rich in a psoralen and was used in the leading suntan lotion in France {17). Psoralens, when activated by sunlight, damage DNA and induce tanning more rapidly than the ultraviolet component of sunlight, which is also a carcinogen {17). Psoralens (plus light) are also effective in producing oxygen radicals (18). 4) The potato giycoalkaloids sotanine and diuconinc are strong cholinesterase inhibitors and possible teratogens and are present at about 15 mg per 200 g of potato (12. IS). When potatoes arc dis eased. bruised, or exposed to light, these and other (24) giycoalkaloids reach lev els that can be lethal to humans (12). Plants typically respond to damage by making more (and often different) toxic chemicals as a defense against insects and fungi (19. 24. 25). The different cultivars of potatoes vary in the concentra tion of these toxic giycoalkaloids (the concentration is a major determinant of insect and disease resistance); one cultivar bred for insect resistance had to be withdrawn from use because of its toxici ty to humans (> 40 mg of giycoalkaloids in a 200-g potato is considered to be a toxic level) (12). 5) Quercetin and several similar flavonoids are mutagens in a number of short-term test systems. Fluvonoids are extremely widespread (daily levels close to I g) in the human diet (8. 16. 20. 21). There is evidence for the carcinogenicity of quercetin in two strains of rats (5), although it was negative in other experi ments (2/). 6) Quinones and their phenol precur sors (9. 14. 16. 23. 45) are widespread in the human diet. Quinones are quite toxic as they can act as electrophiles or accept a single electron to yield the semiquinone radical, which can either react di rectly with DNA U4.46) or participate in a redox cycle of superoxide radical gen eration by transferring the electron to O' (47). The superoxide radical and its met abolic product H:0' can, in turn, lead to the oxidation of fat in cellular mem branes by a lipid peroxidation chain re action. thus generating mutagens and carcinogens, as discussed below. A num ber of quinones and dietary phenols have been shown to be mutagens (7. 9. 16. 23. 44). Mutagenic anthraquinone deriva tives are found in plants such as rhubarb and in mold toxins (7, 16, 48). Many dietary phenols can spontaneously autoxidize to quinones. generating hydro gen peroxide at the same time (examples are catechol derivatives such as the caf- 23 SEPTEMBER 1983 feic acid component of chlorogenic acid (9). which is present at about 250 mg per cup of coffee]. The amounts of these phenols in human urine (and in the diet) are appreciable (45). Catechol, for exam ple, is excreted in urine at about 10 mg per day and appears to be mainly derived from metabolism of plant substances (45). Catechol is a potent promoter of carcinogenesis (45), an inducer of DNA damage, a likely active metabolite of the carcinogen benzene (46), and a toxic agent in cigarette smoke (45). Catechol amine induction of cardiomyopathy is thought to occur through generation of oxygen radicals (49). 7) Theobromine, a relative of caffeine, has been shown to be genotoxic in a variety of tests, to potentiate (as does caffeine) DNA damage by various car cinogens in human cells, and to cause testicular atrophy and spermatogenic cell abnormalities in rats (27). Cocoa powder is about 2 percent theobromine, and therefore humans may consume hundreds of milligrams of theobromine a day from chocolate. Theobromine is also present in tea. 8) Pyrrolizidine alkaloids are carcino genic. mutagenic, and teratogenic and are present in thousands of plant species (often at > 1 percent by weight >. some of which are ingested by humans, particu larly in herbs and herbal teas and occa sionally in honey (7, 29). Pyrrolizidine alkaloid poisonings in humans (as well as in other mammals) cause lung and liver lesions and are commonly misdiagnosed (29). 9) The broad (fava) bean (Vicia faba), a common food of the Mediterranean region, contains the toxins vicine and convicine at a level of about 2 percent of the dry weight (JO). Pythagoras forbade his followers to eat the beans, presum ably because he was one of the millions of Mediterranean people with a deficien cy of glucose-6-phosphate dehydrog enase. This deficiency results in a low glutathione concentration in blood cells, which causes increased resistance to the malarial parasite, probably accounting for the widespread occurrence of the mutant gene in malarial regions. Howev er. the low glutathione concentration also results in a marked sensitivity to agents that cause oxidative damage, such as the fava bean toxins and a varie ty of drugs and viruses. Sensitive indi viduals who ingest fava beans develop a severe hemolytic anemia caused by the enzymatic hydrolysis of vicine to its aglycone. divicine, which forms a quinone that generates oxygen radicals (30). 10) Allyl isothiocyanate, a major flavor ingredient in oil of mustard and horse radish. is one of the main toxins of the mustard seed and has been shown to cause chromosome aberrations in ham ster cells at low concentration (50) and to be a carcinogen in rats (31). H) Gossypol is a major toxin in cot tonseed and accounts for about I percent of its dry weight (J2). Gossypol causes pathological changes in rat and human testes, abnormal sperm, and male steril ity (32. 33). Genetic damage has been observed in embryos sired by gossypoltreated male rats: dominant lethal muta tions in embryos were measured after males were taken off gossypol treatment and allowed to mate (JJ). Gossypol ap pears to be a carcinogen as well: it has been reported to be a potent initiator and also a promoter of carcinogenesis in skin painting studies with mice (J4). Crude, unrefined cottonseed oil contains consid erable amounts of gossypol (100 to 750 mg per 100 ml). Thus human consump tion may be appreciable in countries, such as Egypt, where fairly crude cot tonseed oil is commonly used in cooking. Gossypol is being tested as a male con traceptive in over 10.000 people in China (at an oral dose of about 10 mg per person per day), as it is inexpensive and causes sterility during use (JJ). Gossypol's mode of action as a spermicide may be through the production of oxygen radicals (35). Plant breeders have developed "gland less cotton." a new strain with low levels of gossypol. but seeds from this strain are much more susceptible to attack by the fungus Aspergillus ftavus. which pro duces the potent carcinogen aflatoxin (J6>. 12) Sterculic acid and mttlvalic acid are widespread in the human diet. They are toxic cyclopropenoid fatty acids present in cottonseed oil and other oils from seeds of plants in the family Malvacea! (for instance, cotton, kapok, okra, and durian) (51). Another possible source of human exposure is consump tion of fish, poultry, eggs, and milk from animals fed on cottonseed (51). Cyclopropenoid fatty acids are carcinogens in trout, markedly potentiate the carcinoge nicity of aflatoxin in trout, cause athero sclerosis in rabbits, are mitogenic in rats, and have a variety of toxic effects in farm animals (51). The toxicity of these fatty acids could be due to their ease of oxida tion to form peroxides and radicals (5/). 13) Leguminous plants such as lupine contain very' potent teratogens (22). When cows and goats forage on these plants, their offspring may have severe teratogenic abnormalities: an example is 1257 URL 04010 iTiimtfiifrlfcfi hrirtnwn rnlfi.il. * ^[hiyittTTTiir nflffi URL 04011 the characteristic "crooked calf" abnor mality due to the ingestion of anagyrme from lupine (22). In addition, significant amounts of these teratogens are trans ferred to the animals' milk, so that drink ing the milk during pregnancy is a seri ous teratogenic hazard (22). In one rural California family, a baby boy, a litter of puppies, and goat kids all had "crooked" bone birth-defect abnormali ties. The pregnant mother and the dog had both been drinking milk obtained from the family goats, which had been foraging on lupine (the main forage in winter) (22). It was at first mistakenly thought that the birth defects were caused by spraying of 2,4-D. 14) Sesquiterpene lactones are wide spread in many plants (37), although because they are bitter they are not eaten in large amounts. Some have been shown to be mutagenic (37). They are a major toxin in the white sap of Lactuca virosa (poison lettuce), which has been used as a folk remedy. Plant breeders are now transferring genes from this species to commercial lettuce to increase insect resistance (J8). 15) The phorbol esters present in the Euphorbiacea. some of which are used as folk remedies or herb teas, are potent promoters of carcinogenesis and may have been a cause of nasopharyngeal cancer in China and esophageal cancer in Curasao (39). 16) Alfalfa sprouts contain canavanine. a highly toxic arginine analog that is incorporated into protein in place of argi nine. Canavanine, which occurs in alfal fa sprouts at about 1.5 percent of their dry weight (40), appears to be the active agent in causing the severe lupus erythematosus-Iike syndrome seen when mon keys are fed alfalfa sprouts (40). Lupus in man is characterized by a defect in the immune system which is associated with autoimmunity, antinuclear antibodies, chromosome breaks, and various types of pathology (40). The chromosome breaks appear to be due to oxygen radi cals as they are prevented by superoxide dismutase (52). The canavanine-alfalfa sprout pathology could be due in part to the production of oxygen radicals during phagocytization of antibody complexes with canavanine-containing protein. The 16 examples above, plus coffee (discussed below), illustrate that the hu man dietary intake of "nature's pesti cides" is likely to be several grams per day--probably at least 10,000 limes high er than the dietary intake of man-made pesticides (55). Levels of plant toxins that confer in sect and fungal resistance are being in creased or decreased by plant breeders (58). There are health costs for the use of these natural pesticides, just as there are for man-made pesticides (41, 54), and these must be balanced against the costs of producing food. However, little infor mation is available about the toxicology of most of the natural plant toxins in our diet, despite the large doses we are ex posed to. Many, if not most, of these plant toxins may be "new" to humans in the sense that the human diet has changed drastically with historic times. By comparison, our knowledge of the toxicological effects of new man-made pesticides is extensive, and general ex posure is exceedingly low (5i). Plants also contain a variety of anticar cinogens (55), which are discussed be low. Alcohol. Alcohol has long been associ ated with cancer of the mouth, esopha gus, pharynx, larynx, and, to a lesser extent, liver (I, 56). and it appears to be an important human teratogen, causing a variety of physical and mental defects in babies of mothers who drink (57). Alco hol drinking causes abnormalities in mice (57a) and is a synergist for chromo some damage in-,humans (58). Alcohol metabolism generates acetaldehyde, which is a mutagen and teratogen (59). a cocarcinogen, and possibly a carcinogen (60). and also radicals that produce lipid hydroperoxides (61) and other mutagens and carcinogens (62: see below). In some epidemiologic studies on alcohol (56), it has been suggested that dietary green vegetables are a modifying factor in the reduction of cancer risk. Mold carcinogens. A variety of mold carcinogens and mutagens are present in mold-contaminated food such as com. grain, nuts, peanut butter, bread, cheese, fruit, and apple juice (/5. 63). Some of these, such as sterigmatocystin and afiatoxin. are among the most potent carcinogens and mutagens known (15. 63). Dietary glutathione has been report ed to counteract afiatoxin carcinogenic ity. Nitrite, nitrate, and nitrosamines. A number of human cancers, such as stom ach and esophageal cancer, may be relat ed to nitrosamines and other nitroso compounds formed from nitrate and ni trite in the diet (64, 65). Beets, celery, lettuce, spinach, radishes, and rhubarb ail contain about 200 mg of nitrate per !00-g portion (65). Anticarcinogens in the diet may be important in this context as well (66). Fat and cancer; possible oxidative mechanisms. Epidemiologic studies of cancer in humans suggest, but do not prove, that high fat intake is associated with colon and breast cancer (/. 4. 67). A number of animal studies have shown that high dietary fat is a promoter and a presumptive carcinogen (4. 67, 68). Co lon and breast cancer and lung cancer (which is almost entirely due to cigarette smoking) account for about half of all U.S. cancer deaths. In addition to the cyclopropenoid fatty acids already dis cussed, two other plausible mechanisms involving oxidative processes could ac count for the relation (69) between high Cat and both cancer and heart disease. 1) Rancidfat. Fat accounts for over 40 percent of the calories in the U.S. diet (67). and the amount ofingested oxidized fat may be appreciable (70, 71). Unsatu rated fatty acids and cholesterol in fat are easily oxidized, particularly during cooking (70. 71). The lipid peroxidation chain reaction (rancidity) yields a variety (71-73) of mutagens, promoters, and car cinogens such as fatty acid hydroperox ides (62), cholesterol hydroperoxide (74). endoperoxides. cholesterol and fat ty acid epoxides (74-77). enals and other aldehydes (44, 59. 78), and aikoxy and hydroperoxy radicals (44, 72). Thus the colon and digestive tract are exposed to a variety of fat-derived carcinogens. Hu man breast fluid can contain enormous levels (up to 780 iAf) (75) of cholesterol epoxide (an oxidation product of choles terol). which could originate from either ingested oxidized fat or oxidative pro cesses in body lipids. Rodent feeding studies with oxidized fat (79) have not yielded definitive results. 2) Peroxisomes oxidize an appreciable percentage of dietary fatty acids, and removal of each two-carbon unit gener ates one molecule of hydrogen peroxide (a mutagen, promoter, and carcinogen) (80. 81). Some hydrogen peroxide es capes the catalase in the peroxisome (80. 82, 83). thus contributing to the supply of oxygen radicals, which also come from other metabolic sources (72, 83-85). Hy droperoxides generate oxygen radicals in the presence of iron-containing com pounds in the cell (72). Oxygen radicals, in turn, can damage DNA and can start the rancidity chain reaction which leads to the production of the mutagens and carcinogens listed above (72). Drugs such as clofibrate. which cause lowering of serum lipids and proliferation of per oxisomes in rodents, result in age pig ment (lipofuscin) accumulation (a sign of lipid peroxidation in tissues) and liver tumors in animals (80). Some fatty acids, such as C:2:1 and certain trims fatty acids, appear to cause peroxisomal pro liferation because they are poorly oxi dized in mitochondria and are preferen tially oxidized in the peroxisomes, al though they may be selective for heart or 1258 SCIENCE. VOL. 221 liver (56). There has been controversy fried pork or bacon (9*/). In the evalua contributors to DNA damage (/0j). Ma about the role of trans fatty acids in tion of risk from burnt material it may be jor sources of endogenous oxygen radi cancer and heart disease, and recent useful (in addition to carrying out epide cals are hydrogen peroxide (5J) and su evidence suggests that trans fatty acids miologic studies) to compare the activity peroxide (72, 104) generated as side might not be a risk factor for atheroscle of cigarette tar to that of the burnt mate products of metabolism, and the oxygen rosis in experimental animals (87). rial from cooked food (or polluted air) in radical burst from phagocytosis after vi Americans consume about 12 g of trans short-term tests and animal carcinoge ral or bacterial infection or the inflamma fatty acids a day (57) and a similar nicity tests involving relevant routes of tory reaction (105). A variety of environ amount of unnatural cis isomers (which exposure-. Route of exposure and com mental agents could also contribute to need further study (55)1, mainly from position of the burnt material are critical the oxygen radical load, as discussed hydrogenated vegetable fats. Dietary variables. The risk from inhaled cigarette here and in recent reviews (72. 106). Cj2:| fatty acids are also obtained from smoke can be one reference standard: an Many enzymes protect cells from oxida rapeseed oil and fish oils (86). Thus average life shortening of about 8 years tive damage; examples are superoxide oxidation of certain fatty acids might for a two-pack-a-day smoker. The dismutase (104), glutathione peroxidase I generate grams of hydrogen peroxide per amount of burnt material inhaled from (107). DT-diaphorase (108). and the glu 1 day within the peroxisome (56). Another severely polluted city air. on the other tathione transferases (109). In addition, a source of fat toxicity could be perturba hand, is relatively small: it would be variety of small molecules in our diet are tions in the mitochondrial or peroxisom necessary to breathe smoggy Los Ange required for aniioxidative mechanisms al membranes caused by abnormal fatty les air (111 p.g/m3 total particulates; 31 and appear to be anticarcinogens: some acids, yielding an increased flux of su- M-g/m1 soluble organic matter) for 1 to 2 of these are discussed below. peroxidc and hydrogen peroxide. Mito weeks to equal the soluble organic mat 1) Vitamin (tocopherol) is the major * chondrial structure is altered when rats ter of the particulates or the mutagenic radical trap in lipid membranes (72) and i \ are fed some abnormal fatty acids from ity from one cigarette (20 mg of tar) (9J). has been used clinically in a variety of partially hydrogenated fish oil (59). Di Epidemiologic studies have not shown oxidation-related diseases (f 10). Vitamin etary C;2:| fatty acids and clofibrate also significant risks from city air pollution E ameliorates both the cardiac damage induce ornithine decarboxylase (56). a alone (/, 96). Air in the houses of smok and carcinogenicity of the quinones common attribute of promoters. ers is considerably more polluted than adriamycin and daunomycin. which are A recent National Academy of Sci city air outside (97). mutagenic, carcinogenic, cause cardiac ences committee report suggests that a Coffee, which contains a considerable damage, and appear to be toxic because reduction of fat consumption in the amount of burnt material, including the of free radical generation (///). Protec American diet would be prudent (d), mutagenic pyrolysis product methylgly- tive effects of tocopherols against radia although other scientists argue that, until oxal. is mutagenic (21. 95). However, tion-induced DNA damage and mutation we know more about the mechanism of one cup of coffee also contains about 250 and dimethylhydrazine-induced carcino the fat-cancer relation and about which mg of the natural mutagen chiorogenic genesis have also been observed (112). types of fat are dangerous, it is prema acid (9) [which is also an antinitrosating Vitamin E markedly increases the endur ture to recommend dietary changes (90). agent (66)]. highly toxic atractylosides ance of rats during heavy exercise, (10). the glutathione transferase inducers which causes extensive oxygen radical kahweal palmitate and cafestol palmitate damage to tissues U13). Cooked Food as a Source of (//). and about 100 mg of caffeine (which 2) 3-Carotene is another antioxidant in Ingested Burnt and Browned Material inhibits a DNA-repair system and can the diet that could be important in pro increase tumor yield (99) and cause birth tecting body fat and lipid membranes Work of Sugimura and others has indi defects at high levels in several experi against oxidation. Carotenoids are free- cated that the burnt and browned materi mental species (100)]. There is prelimi radical traps and remarkably efficient al from heating protein during cooking is nary. but not conclusive, epidemiologic quenchers of singlet oxygen (114). Sin highly mutagenic (2/, 91). Several chem evidence that heavy coffee drinking is glet oxygen is a very reactive form of icals isolated on the basis of their mu associated with cancer of the ovary, oxygen which is mutagenic and particu tagenicity from heated protein or pyro- bladder, pancreas, and large bowel (101). larly effective at causing lipid peroxida lyzed amino acids were found to be Cooking also accelerates the rancidity tion (114). It can be generated by pig carcinogenic when fed to rodents (21). In reaction of cooking oils and fat in meat ment-mediated transfer of the energy of addition, the browning reaction products (70. 71). thus increasing consumption of light to oxygen, or by lipid peroxidation, from the caramelization of sugars or the mutagens and carcinogens. although the latter is somewhat contro reaction of amino acids and sugars dur versial. 3-Carotene and similar poly- ing cooking (for instance, the brown ma prenes are present in carrots and in all terial on bread crusts and toasted bread) Anticarcinogens food that contains chlorophyll, and they contain a large variety of DNA-damaging appear to be the plants' main defense agents and presumptive carcinogens (22. We have many defense mechanisms to against singlet oxygen generated as a by 28. 92). The amount of burnt and protect ourselves against mutagens and product from the interaction of light and browned material in the human diet may carcinogens, including continuous shed chlorophyll (115). Carotenoids have be several grams per day. By compari ding of the surface layer of our skin, been shown to be anticarcinogens in rats son about 500 mg of burnt material is stomach, cornea, intestines, and colon and mice (116). Carotenoids (in green inhaled each day by a smoker using two (102). Understanding these mechanisms and yellow vegetables) may be anticar- packs of cigarettes (at 20 mg of tar per should be a major goal of cancer, heart, cinogens in humans (/. 56. 117). Their cigarette) a day. Smokers have more and aging research. Among the most protective effects in smokers might be easily delectable levels of mutagens in important defenses may be those against related to the high level of oxidants in their urine than nonsmokers (9J). but so oxygen radicals and lipid peroxidation if. both cigarette smoke and tar (45. 118). do people who have consumed a meal of as discussed here, these agents are major Carotenoids have been used medically in URL 04012 23 SEPTEMBER I9ftt 1239 URL 04013 the treatment for some genetic diseases, such as porphyrias, where a marked pho tosensitivity is presumably due to singlet oxygen formation (119). 3) Selenium is another important di etary anticarcinogen. Dietary selenium (usually selenite) significantly inhibits the induction of skin, liver, colon, and mammary tumors in experimental ani mals by a number of different carcino gens. as well as the induction of mamma ry tumors by viruses (120). It also inhib its transformation of mouse mammary cells (121). Low selenium concentrations may be a risk factor in human cancer (122). A particular type of heart disease in young people in the Keshan area of China has been traced to a selenium deficiency, and low selenium has been associated with cardiovascular death in Finland (123). Selenium is in the active site of glutathione peroxidase, an en zyme essential for destroying lipid hy droperoxides and endogenous hydrogen peroxide and thus helping to prevent oxygen radical-induced lipid peroxida tion (107). although not all of the effects of selenium may be accounted for by this enzyme (120). Several heavy-metal tox ins. such as Cd:* (a known carcinogen) and Hg:". lower glutathione peroxidase activity by interacting with selenium (107). Selenite (and vitamin E) has been shown to counter the oxidative toxicity of mercuric salts (124). 4) Glutathione is present in food and is one of the major antioxidants and antimutagens in the soluble fraction of cells. The glutathione transferases (some of which have peroxidase activity) are ma jor defenses against oxidative and alkyl ating carcinogens (109). The concentra tion of glutathione may be influenced by dietary sulfur amino acids (125. 126). N- Acetylcysteine, a source of cysteine, raises glutathione concentrations and re duces the oxidative cardiotoxicity of adriamycin and the skin reaction to radi ation (127). Glutathione concentrations are raised even more efficiently by l-2oxothiazolidine-4-carboxylate. which is an effective antagonist of acetamino phen-caused liver damage (126). Acet aminophen is thought to be toxic through radical and quinone oxidizing metabo lites < 128). Dietary glutathione may be an effective anticarcinogen against afiatoxin (129). 5) Dietary ascorbic acid is also impor tant as an antioxidant. It was shown to be anticarcinogenic in rodents treated with ultraviolet radiation, benzofolpyrenc. and nitrite (forming nitroso carcin ogens) (64. 65. 130). and it may be in versely associated with human uterine cervical dysplasia (although this is not proof of a cause-effect relationship) (131). It was recently hypothesized that ascorbic acid may have been supple mented and perhaps partially replaced in humans by uric acid during primate evo lution (132). 6) Uric acid is a strong antioxidant present in high concentrations in the blood of humans (132). The concentra tion of uric acid in the blood can be increased by dietary purines; however, too much causes gout. Uric acid is also present in high concentrations in human saliva (132) and may play a role in de fense there as well, in conjunction with lactoperoxidase. A low uric acid level in blood may possibly be a risk factor in cigarette-caused lung cancer in humans (133). 7) Edible plants and a variety of sub stances in them, such as phenols, have been reported to inhibit (cabbage) or to enhance (beets) carcinogenesis (//. 55. 134) or mutagenesis (23. 66. 92. 135) in experimental animals. Some of these substances appear to inhibit by inducing cytochrome P-450 and other metabolic enzymes (UJ4); see also (//)], although on balance it is not completely clear whether it is generally helpful or harmful for humans to ingest these inducing sub stances. The hypothesis that as much as 80 percent of cancer could be due to envi ronmental factors was based on geo graphic differences in cancer rates and studies of migrants (136). These differ ences in cancer rates were thought to be mainly due to life-style factors, such as smoking and dietary carcinogens and promoters (136). but they also may be due in good part (see also (/)] to less than optimum amounts of anticarcinogens and protective factors in the diet. The optimum levels of dietary antioxi dants. which may vary among individ uals. remain to be determined; however, at least for selenium (120). it is important to emphasize the possibility of deleteri ous side effects at high doses. Oxygen Radicals and Degenerative Diseases Associated with Aging Aging. A plausible theory of aging holds that the major cause is damage to DNA (102. 137) and other macromoleculcs and that a major source of this damage is oxygen radicals and lipid per oxidation (43 . 84. 103. 138-141). Cancer and other degenerative diseases, such as heart disease (102). are likely to be due in good part to this same fundamental de structive process. Age pigment (lipofuscin) accumulates aging in all mammalian species and has been associated with lipid peroxidation (73. 84, 138. 139). The fluorescent products in age pigment are thought to be formed by malondialdehyde (a mutagen and carcinogen and a major end product of rancidity) crosslinking protein and lipids (138). Metabol ic rate is directly correlated with the rate of lipofuscin formation (and inversely correlated with longevity) (139). Cancer increases with about the fourth power of age, both in short-lived species such as rats and mice (about 30 percent of rodents have cancer by the end of their 2- to 3-year life-span) and in longlived species such as humans (about 30 percent of people have cancer by the end of their 85-year life-span) (/*'), Thus, the marked increase in life-span that has occurred in 60 million years of primate evolution has been accompanied by a marked decrease in age-specific cancer rates; that is. in contrast to rodents. 30 percent of humans do not have cancer by the age of 3 (142). One important factor in longevity appears to be basal metabol ic rate (139. 141). which is much tower in man than in rodents and could markedly affect the level of endogenous oxygen radicals. Animals have many antioxidant de fenses against oxygen radicals. In creased levels of these antioxidants, as well as new antioxidants, may also be a factor in the evolution of man from short-lived prosimians (143). It has been suggested that an increase in superoxide dismutase is correlated (after the basal metabolic rate is taken into account) with increased longevity during primate evolution, although this has been disput ed (141). Ames et al. proposed (132) that as uric acid was an antioxidant and was present in much higher concentrations in the blood of humans than in other mam mals. it may have been one of the inno vations enabling the marked increase in life span and consequent marked de crease in age-specific cancer rates which occurred during primate evolution. The ability to synthesize ascorbic acid may have been lost at about the same time in primate evolution as uric acid levels be gan to increase (144). Cancer and promotion. Both DNAdamaging agents (initiating mutagens) (21. 41, 42) and promoters (145) appear to play an important role in carcinogene sis (21. 146). It has been postulated that certain promoters of carcinogenesis act by generation of oxygen radicals and resultant lipid peroxidation (73. 146149). Lipid peroxidation cross-links pro teins (43. 150) and affects all aspects of cell organization (72). including mem brane and surface structure, and the mi- 1260 SCIENCE. VOL. 221 totic apparatus. A common property of promoters may be their ability to pro duce oxygen radicals. Some examples are fat and hydrogen peroxide (which may be among the most important pro moters) (67, 68, 81), TCDD (151). lead and cadmium (152). phorbol esters (147, 149, 153), wounding of tissues (154), asbestos (155), peroxides (156). catechol (-/5) (see quinones above), mezerein and teleocidin B (147). phenobarbital (157). and radiation (72, 158). Inflammatory reactions involve the production of oxy gen radicals by phagocytes (105), and this could be the basis of promotion for asbestos (155) or wounding (154). Some of the antioxidant anticarcinogens (dis cussed above) are also antipromoters (73. 121. 146, 159, 160). and phorbol ester-induced chromosome damage (149) or promotion of transformation (159) is suppressed by superoxide dismutase, as would be expected if promoters were working through oxidative mecha nisms. Many ''complete" carcinogens cause the production of oxygen radicals (73, I6!)\ examples are nitroso com pounds. hydrazines, quinones, polycy clic hydrocarbons (through quinones). cadmium and lead salts, nitro .com pounds, and radiation. A good part of the toxic effects of ionizing radiation damage to DNA and cells is thought to be due to generation of oxygen radicals (103, 162), although only a tiny part of the oxygen radical load in humans is likely to be from this source. Recent studies give some clues as to how promoters might act. Promoters dis rupt the mitotic apparatus, causing hemizygosity and expression of recessive genes (163). Phorbol esters generate oxy gen radicals, which cause chromosome breaks (164) and increase gene copy number (165). Promoters also cause for mation of the peroxide hormones of the prostaglandin and leukotriene family by oxidation of arachidonic acid and other C'o polyenoic fatty acids, and inhibitors of this process appear to be antipro moters (160). These hormones are inti mately involved in cell division, differen tiation, and tumor growth (166) and could have arisen in evolution as signal molecules warning the cell of oxidative damage. Effects on the cell membrane have also been suggested as the impor tant factor in promotion, causing inhibi tion of intercellular communication (167) or protein kinase activation (I67a). Heart disease. It has been postulated that atherosclerotic lesions, which are derived from single ceils, are similar to benign tumors and are of somatic muta tional origin (102. 168). Fat appears to be one major risk factor for heart disease as 23 SEPTEMBER 1983 well as for colon and breast cancer (69). In agreement with this, a strong correla tion has been observed between the fre quency of atherosclerotic lesions and adenomatous polyps of the colon (69). Thus, the same oxidative processes in volving fat may contribute to both dis eases. Oxidized forms of cholesterol have been implicated in heart disease (169), and atherosclerotic-like lesions have been produced by injecting rabbits with lipid hydroperoxide or oxidized cholesterol (169). The anticarcinogens discussed above could be anti-heart dis ease agents as well. As pointed out in the preceding section, vitamin E amelio rates both the cardiac damage and carci nogenicity of the free-radical-generating quinones adriamycin and daunomycin: /V-acetylcysteine reduces the cardiotoxicity of adriamycin: and selenium is an antirisk factor for one type of heart dis ease. Other diseases. The brain uses 20 per cent of the oxygen consumed by man and contains an appreciable amount of unsaturated fat. Lipid peroxidation (with consequent age pigment) is known to occur readily in the brain (72), and possi ble consequences could be senile demen tia or other brain abnormalities (3-/). Sev eral inherited progressive diseases of the central nervous system, such as Batten's disease, are associated with lipofuscin accumulation and may be due to a lipid peroxidation caused by a high concentra tion of unbound iron (170). Mental retar dation is one consequence of an inherit ed defective DNA repair system (XP complementation group D) for depurinated sites in DNA (171). Senile cataracts have been associated with light-induced oxidative damage (172). The retina and an associated layer of cells, the pigment epithelium, are ex tremely sensitive to degeneration in vita min E and selenium deficiency (173). The pigment epithelium accumulates massive amounts of lipofuscin in aging and dietary antioxidant deficiency (173). The eye is well known to be particularly rich in antioxidants. The testes are quite prone to lipid peroxidation and to the accumulation of age pigment. A number of agents, such as gossypol, which cause genetic birth defects (dominant lethals) may be active by this mechanism. The various agents known to cause cancer by oxidative mechanisms are prospective mutagenic agents for the germ line. Thus, vitamin E. which was discovered 60 years ago as a fertility factor (72), and other antioxi dants such as selenium (174). may help both to engender and to protect the next generation. Risks There are large numbers of mutagens and carcinogens in every meal, all per fectly natural and traditional^fsee also (21. 23)\. Nature is not benign. It should be emphasized that no human diet can be entirely free of mutagens and carcino gens and that the foods mentioned are only representative examples. To identi fy a substance, whether natural or man made, as a mutagen or a carcinogen, is just a first step. Beyond this, it is neces sary to consider the risks for alternative courses of action and to quantitate the approximate magnitude of the risk, al though the quantification of risk poses a major challenge. Carcinogens differ in their potency in rodents by more than a millionfold (175). and the levels of partic ular carcinogens to which humans are exposed can vary more than a billion fold. Extrapolation of risk from rodents to humans is difficult for many reasons, including the longevity difference, anti oxidant factors, and the probable multicausal nature of most human cancer. Tobacco smoking is, without doubt, a major and well-understood risk, causing about 30 percent of cancer deaths and 25 percent of fatal heart attacks (as well as other degenerative diseases) in the Unit ed States (/). These percentages may increase even more in the near future as the health effects of the large increase in women smokers become apparent (/). Diet, which provides both carcinogens and anticarcinogens, is extremely likely to be another major risk factor. Exces sive alcohol consumption is another risk, although it does not seem to be of the same general importance as smoking and diet. Certain other high-dose exposures might also turn out to be important for particular groups of people--for in stance, certain drugs, where consump tion can reach hundreds of milligrams per day: particular cosmetics: and cer tain occupational exposures (2), where workers inhale dusts or solvents at high concentration. We must also be prudent about environmental pollution (41, 54). Despite all of these risks, it should be emphasized that the overall trend in life expectancy in the United States is con tinuing steadily upward (176). The understanding of cancer and de generative disease mechanisms is being aided by the rapid progress of science and technology, and this should help to dispel confusion about how important health risks can be identified among the vast number of minor risks. We have many methods of attacking the problem of environmental carcinogens (and anticarcinogens), including human epidemi- i:m URL 04014 ology (I), short-term tests (41, 42, 177), and animal cancer tests (175). Powerful new methods are being developed [for instance, see (.5<3, 177)) for measuring DNA damage or other pertinent factors with great sensitivity in individuals. These methods, which are often nonin- vasive as they can be done on blood or urine (even after storage), can be com bined with epidemiology to determine whether particular factors are predictive of disease. Thus, more powerful tools wilt be available for optimizing antioxi dants and other dietary anti-risk factors, for identifying human genetic variants at high risk, and for identifying significant health risks. Reference* and .Note* 1. R. Doll and R. Peto, J. Natl. Cancer Inst. 66, ii9: i iwi i. 2. R. Peio and M. Schneidcrman. Ed*.. Banbury Report 9. Quantification of Occupational Can cer (Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.. 1981). 3. Cancer Facts and Figures, I9&J (American Cancer Society. New York, 1982). 4. National Research Council. Diet. Nutrition and Cancer (National Academy Press. Wash ington. D.C.. 1982). 5. E. C. Miller. J. A. Miller, I. Hirono. T. Sugimura. S. Takayama, Eds.. Naturally Occurring Careinogens-Mutagens and Modulators of Carcinogenesis (Japan Scientific Societies Press and University Park Press. Tokyo and Baltimore, 1979V. E. C. Miller et al.. Cancer Res. 43. 1124 (1983): C. loannides. M. Delaforge. D. V. Parke. Food Cosmet. Toxicol. 19. 657 119811. 6. G. J. Kapadia. Ed.. Oncology Overview on Naturally Occurine Dietary Carcinogens of Plant Origin 4 Internationa) Cancer Research Data Bank Program. National Cancer Institute. Bethesda. Maryland. 1982). 7. A. M. Clark, in Environmental Mutagenesis, Carcinogenesis, and Plant Biology. E. J. Klekowski. Jr.. Ed. tPraeger. New York. 1982). vol. I. pp. 97-132. 8. A. M. Pamukcu. S. Yalciner. J. F. Hatcher. G. T. Bryan. Cancer Res. 40. 3468 (1980): J. F. Hatcher, A. M. Pamukcu. E. Enurk. G. T. Brvan. Fed. Proc. Fed. Am. Soc. Exp. Biol. 42. 786(1983). 9. H. F. Stich. M. P. Rosin. C. H. Wu. W. D. Powrie. Mutat. Res. 90. 201 (198!): A. A. Aver'yanov. Biokhimtya 44. 256 (1981); A. F. Hanham, B. P. Dunn. H. F. Such. Mutat. Res. 116. 333 (1983). 10. K. H. Pegel. Chem, Eng. News 59. 4 (20 July 1981). 11. L. K. T. Lam. V. L. Spamins. L. W. Wattenberg. Cancer Res. 42. 1193 (1982). 12. S. J. Jadhav. R. P. Sharma, D. K. Salunkhe. CRC Crit. Rev. Toxicol. 9, 21 (1981). 13. R. L. Hall. Nurr. Cancer 1 iNo. 2). 27 (1979). !4. H. W, Moore and R. Czcmiak. Med. Res. Rev. I, 249 (1981). 15. I. Hirono. CRC Crit. Rev. Toxicol. 8, 235 (19811. 16. J. P. Brown. Mutat. Res. 75. 243 (1980). 17. M. J. Ashwood-Smnh and G. A. Poulton. ibid. S5. 389 (1981). 18. A. Ya. Potapenko. M. V. Moshnin. A. A. Krasnovsky, Jr.. V. L. Sukhorukov. 2. Naturforsch. 37. 70(1982). 19. G. W. I vie. D. L. Holt, M. C. Ivey. Science 213. 909 (1981); R. C. Beierand E. H. Oenli. Phytochemistry, in press; R. C. Beier. G. W. Ivte, E. H. Oertli. in "Xenobiotics in Foods and Feeds." ACJ Symp. 5e'.. in press: D. L. Holt. Food Chem. Toxicol. 21. 163 (1983). 20. G. Tamura, C. Gold. A. Ferro-Luzzi. B. N. Ames. Proc. Natl. Acad. Sci. U.S.A. 77, 4961 (1980). 21. T. Sugimura and S. Sato, Cancer Res. (Suppl.) 43. 2415s 11963); T. Sugimura and M. Nagao, in Mutagenicity New Horizons in Genetic Toxi cology. J. A. Hcdille, Ed. (Academic Press. New York. 1982), pp. 73-88. 22. W. W. Kilgore. 0. G. Crosby. A. L. Craigmill. N. K. Puppen, Calif. Agric. 35 (No. ID (No vember 198D; D. G. Crosby, Chem. Eng. 50. A. Kasamaki et al., Mutat. Res. 105. 387 News 61. 37 (II April 1983); C. D. Wanen. (1982). ibid., p. 3(13 June 1983). 51. J. D. Hendricks. R. O. Sinnhuber, P. M. Love 23. H, F. Stich. M. P. Rosin. C. H. Wu, W. D. land. N. E. Pawlowski, J. E. Nison. Science Powrie, in Mutagenicity: New Horizons in Ge 208. 309(1980): R. A. Phelps. F. S. Shenstone. netic Toxicology, J. A. Heddle. Ed. (Academic A. R. Kemmerer. R, J. Evans. Poult. Sri. 44. Press. New York. 1982). pp. 117-142; 358 (1964); N. E. Pawlowski. personal commu W. D. Powrie. Cancer Lett. 14. 251 (1981). nication. 24. N. Katsui. F. Yagihashi. A. Murai, T. Masa- 52. I. Emerit, A. M. Michelson. A. Levy. J, p, mune. Bull. Chem. Soc. Jpn. 55. 2424 (1982); Camus. J. Emerit. Hum. Genet. 55, 3-4J (1980). _______ ibid., p. 2428; R. M. Bostock. R. 5A3.. FDA Compliance Program Report of Findings. Laine. J. A. Kuc. Plant Phvsiol. 70, 1417 FY79 Total Diet Studies--Adult (No. (1982). 7305.002); available from National Technical 25. H. Griesebach and J. Ebei. Angew. Chem. Int. Information Service. Springfield. Va.). It is Ed. Engl. 17,6)5 (1978). estimated that the daily dietary intake of syn 26. J. M. Concon. D. S. Newburg. T. W. Swer- thetic organic pesticides and herbicides is czek. Nulr. Cancer t (No. 3). 22 Il9m. about 60 ug. with chlorpropham. malathion. 27. H. W. Renner and ft. Munzner. Murat. Res. and DDE accountlag for about three-fourths of 103. 275 (1982): H. W. Renner. Experientia 38. this. An estimate of 150 *g of daily exposure in 600(1982); D. Mourelatos. J. Dozi-Vassiliades. Finland to pesticide residues has been made by A. Granitsas. Mutat. Res. 104. 243 (1982); J. H. K. Hemmimki. H. Vainio. M. Sorsa. S. Sal- Gans. Toxicol. Appl. Pharmacol. 63. 312 (1982). minen (J. Environ. Sci. Health Cl (No. I), 55 <I983)|. 28. B. Toth, in Naturally Occurring Carcinogens- 54. N. K. Hooper. B. N. Ames. M. A. Saleh. J. E. Muiagens and Modulators of Carcinogenesis. Casida. Science 205, 591 (1979). E. C. Miller. J. A. Miller, I. Hirono. T. Sugi 55. L. W, Wallenberg. Cancer Res. (Suppl.) 43. mura. S. Takayama. Eds. (Japan Scientific 2448s (1983). Societies Press and University Park Press. 56. J. Hoey. C. Montvemay, R. Lambert. Am. J. Tokyo and Baltimore. 1979). pp. 57--65; A. E. Epidemiol. 113. 668 (1981); A. J. Tuyns. G. Ross. D. L. Nagel. B. Toth. J. Agric. Food Pequignot. M. Gignoux. A. Valla. Int. J. Can- Chem. 30, 521 (1982); B. Toth and K. Paiil. cer 30.9 (1982); A. Tuyns. in Cancer Epidemi Mycopathologia 78. II (1982); B. Toth. D. ology and Prevention. D. Schottenfeld and J. Nagel. A. Ross. Br. J. Cancer 46. 417 (1982). F. Fnumeni. Jr.. Eds. (Saunders, Philadel 29. R. SchoemaJ, Toxicol. Lett. 10. 323 (1982): R. J. Huxtable, Ptrspect. Biol. Med. 24. t (1980): phia. 1982). pp. 293-303. R. G. Ziegler et al.. J. Natl. Cancer Inst. 67. 1199 (1981); W. D. H. Ntwa. H. lshiwaa. K. Yamada. J. Chromatogr. 257, 146(1983). Flanders and K- J. Rothman. Am. J. Epidemi ol. 115, 371 (1982). 30. M. Chevion and T. Navok. Anal. Biochem. 57. E. L. Abel. Hum. Biol. 54. 421 (1982): H. L. 128. 152 (1983); V. Latlanzio. V. V. Bianco. D. Rosset. L. Weiner. A. Lee. B. Zuckerman, E. Lafiandra. Experientia 38 . 789 (1982): V. L. Flohe, G. Niebch. H. Reiber. Z. Klin. Chem. Dooling, E- Oppenheimer. Obstet. Gynecol. 61. 539 (1983). Klin. Biochem. 9. 431 (1971); J. Mager. M. 57a.R. A. Anderson. Jr.. B. R. Willis. C. Oswald. Chevion. G. Glaser, in Toxic Constituents of L. J. D. Zaneveld. /. Pharmacol. Exp. Ther. Plant Foodstuffs. I. E. Liener. Ed. (Academic 225. 479 (1983). Press. New York. 1980), pp. 265-294. 31. J. K. Dunnick et al.. Fundam. Appl. Toxicol. 2, 58. H. F. Stich and M. P. Rosin. Int. J. Cancer 31, 305 (1983). 1)4 1)982). 59. R. P. Bird. H. H. Draper. P. K. Basnir. Mutat. 32. L. C. Berardi and L. A. GoldWatt. in Toxic Constituents of Plant Foodstuffs. 1. E. Liener. Rex. lilt. 237 < 1982). M. A. Campbell and A. G. Fantel. Life Sci. 32. 2641 (1983). Ed. (Academic Press, ed. 2. New York. 1980). 60. V. J. Feron. A. Krvysse. R. A. Woutersen. pp. 183-237. Eur. J. Cancer Clin. Oncol. 19. 13 (1982). 33. S. P. Xue. in Proceedings. Symposium on 61. T. Suematsu el al.. Alcoholism: Clin. Exp. Res. Recent Advances in Fertility Regulation (Bei 5. 427 (1981); G. W. Winston and A. I. Ceder- jing. 2 to 5 September. I960), p. 122. baum, Biochem. Pharmacol. 31, 2301 (1982): 34. R. K. Haroz and J. Thomasson. Toxicol. Lett. L. A. Videla. V. Fernandez. A. de Marinis. N. Suppl. 6. 72 (1980). Fernandez. A. Valenzuela. 8iochem. Biophys. 35. M. Cobum. P. Sinsheimer. S. Segal. M. Bur gos. W. Troll. Biol. Bull. (Woods Hole. Mass.I Res. Commun. 104. 965 (1982): T. E. Siege, Res. Commun. Chem. Pathol. Pharmacol. 36. 159, 468 (1980). 287 (19821. 36. C. Campbell, personal communication. 62. M. G. Cutler and R. Schneider. Food Cosmet. 37. G. D. Manners. G. W. Ivie. J. T. MacGregor. Toxicol. Appl. Pharmacol. 45, 629 (1978); G. Toxicol. 12. 451 (1974). 63. Y. Tazima. in Environmental Mutagenesis. W. Ivie and D. A. Witzel. in Plant Toxins, vol. Carcinogenesis and Plant Biology, E. J. Kle- I, Encyclopedic Handbook of Natural Toxins, kowski. Jr., Ed. (Praeger. New York. 1982), A. T. Tu and R. F. Keeler. Eds. (Dekker. New vol. I, pp. 68-95. York, in press). 64. P. N. Magee. Ed.. Banbury Report 12. Nitrot- 38. J. C. M. Van dcr Hoeven et ai.. in Mutagens in amines and Human Cancer (Cold Spring Har Our Environment, M. Sorsa and H. Vainio, bor Laboratory, Cold Spring Harbor, N.Y.. Eds. (Liss. New York, 19821. pp. 327-338: J, 1982): P. E. Hartman, in Chemical Mutagens. C. M. van der Hoeven. W. J. Lagerweij. 1. M. F. J. de Senes and A. Hollaender. Eds. (Ple Bruggeman. F. G. Voragen. J. H. Koeman. J. num. New York. 1982). vol. 7. pp. 211-294; P. Agric. Food Chem.. in press. E. Hartman. Environ. Mutagen. $. 111 (1983). 39. T. Hirayama and Y. (to. prev. Med. 18. 614 65. Committee on Nitrite and Alternative Curing f 19811: E. Hecker. J. Cancer Res. Clin. Oncol. Agents in Food. Assembly of Life Sciences. 99. 103 (I98D. National Academy of Sciences. The Health 40. M. R. Malinow, E. J. Bardana. Jr.. 8. Pir- Effects of Nitrate. Nitrite, and S-Nitroso Com ofsky. S. Craig, P. McLaughlin. Science 216. pounds (Naiionai Academy Press. Washing 4)5 11982). ton. D.C.. 1981). 41. B. N. Ames. ibid. 204. 587 (1979). "Mutagen" 66. H. F. Stich. P. K. L. Chan. M. P. Rosin. Int. J. will be used in its broad sense to include Cancer 30. 719 (1982); H. F. Stich and M. P. clastcgens and other DNA-damaging agents. Rosin, in Nutritional and Metabolic Aspects of 42. H. F. Such and R. H. C. San. Eds.. Short- Food Safety. M. Friedman. Ed. (Plenum, New Term Tests for Chemical Carcinogens (Spring- York, in press). er-Veriag. New York. I98D. 67. L. J. Kinlen. Br. Med.J. 286. 1081 (1983); D. J. 43. P. Hochstein and S. K. Jain. Fed. Proc. Fed. Fink and D. Kritchevsky, Cancer Res. 41.3677 Am. Soc. Exp. Biol. 40. 183 (1981). (1981). 44. D. E. Levin. M. Hollstein. M. F. Christman. 68. C. W. Welsch and C. F. Aylsworth. J. Natl. E. Schwiers. B. N. Ames. Proc. Natl. Acad. Cancer Inst. 70. 215 (1983). Sci. U.S.A. 79. 7445 (1982). Many additional 69. P. Correa. J. P. Strong. W. D. Johnson. P. quinones and aldehydes have now been shown Pizzolato. W. Huenszel, J. Chronic Dis. 35. to be mutagenic. 313 (1982). 45. S. G. Carmella. E. J. LaVoie. S. S. Hechi. 70. F. B. Shoriand et al..J. Agric. Food Chem. 29. Food Chem. Toxicol. 20. 587 (19821. 863 (1981). 46. K. Morimoto, S. Wolff. A. Koizumi, Mutat. 71. M. G. Slink and M. Karel. Eds., Autoxidation Res. Lett. 119. 355 (1983): T. Sawahata and R. in Food and Biological Systems (Plenum. New A. Neal. Mol. Pharmacol. 23. 453 (1983). York. 1980). 47. H. Kappus and H. Sics, Experientia 37. 1233 72. W. A. Pryor. Ed., Free Radicals in Biology (1981). (Academic Press. New York. 1976 to 1982). 48. L. Tikkanen. T. Matsushima. S. Natori, Mu vois. 1 to 5. rat. Res. 116. 297 (1983). 73. H. B. Demopoulos. D D. Pietronigro. E. 5. 49. P. K. Singal. N. Kapur, K. S. Dhillon. R. E. Flamm, M. L. Sdigman. J. Environ. Pathol. Beamish, N. S. Dhalla, Can. J. Physiol. Phar Toxicol. 3. 273 (1980). macol. 60. 1390 (1982). 74. F. Bischoff. Adv. Lipid Res. 7. 165 (1969). 1262 SCIENCE, VOL. 221 URL 04015 * i s URL 04016 75. N. L. Petrakis, L. D. Gruenke, J. C. Craig. Cancer Res. 41, 2563 (19811. 76. H. S. Black and D. R. Douglas, ibid. 32. 2630 (19721. 77. H. Imai. N. T. Werthessen, V. Subramanyam, P. W. LcQuesne. A. H. Soloway, M. Kan. isawa. Science 207 . 651 (19801. 78. M. Ferrali. R. Fulceri. A. Benedetti, M. Comporti. Res. Commit*. Chem. Pathol. pharmacot. 30. 99 (1980). 79. N. R. Altman. Adv. Lipid Res. 7, 245 (1969). 80. J. K- Reddy, J. R. WaiTen, St. K. Reddy. N. D. Lalwant, Ann. N. Y. Acad. Sci. 386 . 81 (1982); S. K. Reddy and N. D. Lalwani. CRC Cm. Rev. Toxicol., in press. 81. H. L. Plains, Generics 40. 268 (1955): A. ho, M. Naito. Y. Nailo. H. Waianabe. Gunn 73. 315 (19821: G. Speic. W. Vogel. M Wolf. Environ. Mutagen. 4. 135 ((9821: H. Tsuda. Jpn. J. Genet. 56, f (1981); N. Hiroia and T. Yokoyama, Gann 72. 811 (19811. 82. S. Hone. K, tshii. T. Suga, /. Biochem. iTokyo/Xt. 1691 (1981): D. P. Jones. L. Ektow. H. Thor. S. Orrenius. Arch. Biochem. Biophys. 210. 505 (1981). 83. B. Chance. H. Sies. A. Boveris, Physiol. Rev. 59. 527(1979). 84. D. Harman. Proc. Nad. Acad. Sci. U.SA. 78. 7124 (1981): in Free Radicals in Biology, W. A. Pryor. Ed. (Academic Press. New York, 1982). vol. 5. pp. 255-275. 85. I. Emeric. M. Keck. A. Levy. I. Feingold. A. M. Michelson. Mutai. Res. 1QJ. 165 (1982J. 86. C. E. Neat. M. S. Thomassen. H. Osmundsen. Biochem. J. 196. 149 < 19811: J. Bremer and K. R. Norum. J. Lipid Ret. 23. 24} <19821: M. S. Thomassen. E. N. Christiansen. K. R. Norum. Biochem. J. 206. 195 (1982); H. Osmundsen. /nt. J. Biochem. 14. 905 (1982): J. Norseth and M. S Thomassen. Biochim. Biophys. Acta, in piess. 87. M. C. Enig. R. J. Munn. M. Keeney. Fed. Proc. Fed. Am. Soc. Exp. Biol. 37. 2215 (19781: J. E. Hunter. /. Natl. Cancer Inst. 69. 319 (1982T. A. B. Awad. ibid., p. 320: H. Rutten- berg. L. M. Davidson. N. A. Little, P. M. Klurfeld. D. Kritchevslcy. J. Nutr. 113. 835 (1983). 88. R. Wood. Lipids 14. 975 (1979). 89. E. N- Christiansen, T. Flatmark. H. Kryvi, Ear. J. Ceil Biol. 26. II (1981). 90. Council for Agricultural Science and Technolo gy. Diet. Nutrition, and Cancer: A Critique (Special Publication 13, Council for Agricultur al Science and Technology. Ames. Iowa, 1982). 91. L. F. Bjeldanes eta!.. Food Chem. Toxicol. 20. 357 (1982): M. W. Pania. L. J. Loretz. J. M. Storkson. N. C. Holland. Cancer Res. iStippl.i 43. 2444s (1983). 92. H. F. Stich. W. Scich. M. P. Rosin. W. D. Powrie. Mutat. Res. 91. 129 11981): M. P. Rosin. H. F. Stich. W. D. Powrie. C. H. Wu. ibid. 101. 189 (1982): C.-I. We. K. Kitamura. T. Shibamolo, Food Cosmet. Toxicol. 19. 749 (1981). 93. E. Yamasaki and B. N. Ames. Proc. Natl. Acad. Sci. U.S.A. 74. 3555 (1977). 94. R. Baker. A. Ariauslus, A. Bonin. D. Angus, Cancer Lett. 16. 81 (19821. 95. D. Schuetzie, O- Cronn. A. L. Crittenden. R. J. Charlson. Environ. Sci. Technot. 9, 838 (1975): G. Gartrcll and S. K. Fnedtander. At mos. Environ. 9. 279 (19751: L. D. Kier. E. Yamasaki. B. N. Ames. Proc. Natl. Acad. Sci. U S A. 71. 4159 (1974): 1. N. Pitts. Jr.. Envi ron. Health Perspecl. 47. 115 11983). 96. J. E. Vena. Am. J. Epidemiol. 116. 42 (1982); R. Cederlof. R. Dolt, B. Fowler. Environ. Health Perspect. 22. I (1978); F. E. Speizer. ibid. 47. 33 (1983). 97. B. Brunekreef and J. S. M. Boleij. Int. Arch. Occup. Environ. Health 50. 299 (19821. 98. H. Kasai ei a!., Gann 73. 681 (1982). 99. V. Armuth and l. Berenblum. Carcinogenesis 1. 977 (1981). 100. S. Fabro. Reprod. Toxicol. 1. 2 (1982). 101. D. Tnchopoulos. M. Papapostolou. A. Poly. Chronopoulou. Int. J. Cant er 28. 691 (1981); P. Hartge. L. P. Lesher. L. McGowan. R. Hoo ver. ibid. 30. 531 (1982); B. MaeMahon. CancerlBrusseisi 50.2676(19821: H. S. Cucklc and L. J. Kinlen. Br. J. Cancer 44, 760 (1981); R. L. Phillips and D. A. Snowdon. Cuncer Res. (Suppl-) 43. 2403s (1983): L. D. Marrett. S. D. Walter. J. W. Meigs, Am. j. Epidemiol. 117. 113 119831: D. M. Weinberg. R. K. Ross.T. M. Mack. A. Paganini-Hill. B. E. Henderson. Cancer IBrusselsI SI. 675 < 19831. 102. P. E, Hartman, Environ. Mutagen., in press. 103. J. R. Totter. Proc. Natl. Acad. Sci. U.S.A. Tt, 1763 11980). 104. I. Fridovich, in Pathology of Oxygen. A. Au- lor. Ed. (Academic Press. New York, 1982). pp. 1-19, L. W. Oberley. T. D. Oberiey. G. R, Buettner. Med. Hypotheses 6. 249 (1980). 105. B. Halliwdl. Cell'Biol. Int. Rep. 6. 529(1982): A. [. Tauber. Trends Biochem. Sci. 7. 411 (1982); A. B. Weitberg, S. A. Weit2man. M. Destrempes. S. A. Lalt. T- P. Stossel. N. Engl. J. Med. 308. 26 (1983). Neutrophils also pro duce HOC), which is both a chlorinating and oxidizing agent. 106. M. A. Tnish, E. G. Mimnaugh. T. E. Gram. Biochem. Pharmacol. 31, 3335 11982). 107. L. Flohe. in Free Radicals in Biology. W. A. Pryor. Ed. (Academic Press. New York. 1982), vol. 5. pp. 223-254. 108. C. Lind. P. Hochsiein. L. Emstcr. Arch. Bio chem. Biophys. 216. 178(1982). 109. M Warholm. C- Guthenberg. B. Mannervik. C. von Bahr, Biochem. Biophss. Res. Coxnmun. 98. 512 (1981). 110. 1. G. Bieri. L. Carash, V. S. Hubbard, N. Engl. J. Med. 308. 1063 (19831. 111. Y. M. Wang et al.. in Molecular Interrelations of Nutrition and Cancer. M. S. Amotl. J. van Eys. Y.-M. Wang. Eds. (Raven. New York. 1982). pp. 369-379. 112. C. Beckman. R. M. Roy. A. Sproule. Mutat. Res. IM. 73 (1982): M. G. Cook and P. McNa mara. Cancer Res. 40. 1329 (1980). 113. K. J. A. Davies. A. T. Quintanilha. G. A. Brooks, L. Packer. Biochem. Biophvt. Res. Common. 107, 1198 0982). 1)4. C. S. Foote, in Pathology of Oxygen. A. Autor. Ed. (Academic Press. New York. 1982). pp. 21-44; J. E. Packer. J. S. Mahood. V. O. MoraArellano. T. F. Slater, R, L. Wiltson. B. S. Wolfenden. Biochem. Biophvt. Res. Commun. 98.901 (1981); W. Bors. C. Michel. M. Saran. Bull. Ear. Phvsiopathol. Resp. 17 (Suppl.), 13 (1981). 115. N. I. Krinsky and S. M. Deneke. J. Natl. Cancer Inst. 69, 205 (1982): J. A. Turner and J. N. Prebble. J. Gen. Microbiol. 119. 133 (1980): K. L. Simpson and C. 0- Chichester. Anna. Rev. Nutr. 1. 351 (1981). 116. G. Rettura, C. Dattagupta. P. Listowsky. S. M. Levenson. E. Seifter. Fed. Proc. Fed. Am. Soc. Exp. Biol. 41. 786(1983): M. M. MathewsRoth. Oncology 39 . 33 (1982). 117. R. Peto. R. Doll. J. D. Buckley. M. B. Spent. Nature ILondonI 290. 201 11981): R. B. She- kelle et al.. Lancet 1981-11. 1185 <19811; T. Hirayama. Nutr. Cancer l. 67 (1979); G. Kvale. E. Bjelke. J. J. Gan. Int. J. Cancer SI. 397 11983). 118. W. A. Pryor, M. Tamura. M. M. Dooley. P. I. Premovic, D. F. Church, in Oxy-Radicals and Their Scavenger Systems: Cellular and Medi cal Aspects. G. Cohen and R Greenwald. Eds. (Elsevier, Amsterdam. 1983). vol. 2. pp. 185- 192; W. A. Pryor. B. J. Hales. P. I. Premovic. D. F. Church. Science 220. 425 (1983). 119. M. M. Mathews-Roth. J. Natl. Cancer Inst. 69. 279 (1982). 120. A. C. Griffin, in Molecular Interrelations of Nutrition and Cancer. M- S. Arnott. J. Vanevs. Y. M. Wang. Eds. (Raven. New York. 1982). pp. 401--108; D. Medina. H. W. Lane. C. M. Tracey. Cancer Res. iSuppl.) 43. 2460s (1983): M. M. Jacobs. Cancer Res. 43, 1646 (1983); H. J. Thompson. L. D. Meeker. P. J. Becci, S. Kokoska. ibid. 42. 4954 (1982); D. F. Bin. T. A. Lawson. A. D. Julius. C. E. Runice. S. Salmas*, ibid., p. 4455; C. Witting, (j. Wit ting. V. Krieg./. Cancer Res. Clin. Oncol. 104. 109 (1982). 121. M. Chatterjee and M. R. Banerjee. Cancer Leu. 17. 187(1982). 122. W. C. Willett et al.. Lancet, in press. 123. J. T, Salonen. G. Alfthan. J Ptkkaramen. J. K. Huttunen. P. Puska. ibid. 1982-11. 175 (1982). 124. M. Yonaha. E. Itoh. Y. Ohbayashi. M. L'chiyama, Res. Common. Chem. Pathol. Pharmacol. 28. 105 (1980); L. J. Kling and J. H. Soares, Jr., Nutr. Rep. Int. 24. 39 (1981). 125. N. Tateishi. T. Higashi. A. Narusc. K. Hikita. Y. Sakamoto, J. 8iochem. (Tokrvol 90. 1603 (1981). 126. S. M. Williamson. B. Boettcher.-A. Meister. Proc. Natl. Acad. Sci. U.S.A. 79. 6246 (1982). 127. CME Symposium on "/V-Acetylcysteine (NACl: A Significant Chcmoprotective Ad junct." Sem. Oncol. 10 (Suppl. 1), I (1983). 128. J. A. Hinson. L. R. Pohl. T. J. Monks. J. R. Gillette. Life Sci. 29. 107 11981). 129. A. M. Novi. Science 212. 541 (19811. 130. W. B. Dunham et al.. Proc. Natl. Acad. Sci. U.S.A. 79. 7532 (1982): G. Kallistratos and E. Fasske. J. Cancer Res. Clin. Oncol. 97. 91 (1980). 131. S. Wassertheil-Smoller et til.. Am. J. Epidemi ol. 114. 714(1981). 132. B. N. Ames. R. Cathcart. E. Schwiers. P. Hochstein. Proc. Natl. Acad. Sci. U-S-A. 78. 6858 11981). 133. A. Nomura. L. K. Heilbrun. G. N. Stemmerrnann. in preparation. 134. J. N. Boyd. J. G. Babish. G. S. Sioewsand. Food Chem. Toxicol. 20 . 47 (1982). 135. A. W. Wood, et al., proc. Natl. Acad. Sci. U SA. 79, 5513 (1982). 136. T. H. Maugh II. Science 205. 1363 (1979) (interview with John Higginson). 137. H. L. Genslerand H. Bernstein. O. Rev. Biol. 6, 279(1981). 138. A. L. Tappet, in Free Radicals in Biology. W. A. Pryor. Ed. (Academic Press, New York. 1980). vol. 4. pp. 1-47. 139. R. S. Sohal, in Ace Pigments. R. S. Sohal. Ed. (Elsevier/North-Holland. Amsterdam. I98li. pp. 303-316. 14ft/I. E Fleming. J. Miqucl. S. F. Cottrell. L. S. Yertgovan. A. C. Economos. Gerontology 28. 44(1982), 141. J. M. Tolmasoff. T. Ono. R. G. Cutler. Proc. Natl. Acad. Sci. U.S.A. 77 . 2777 (1980). R. G. Cutler. Gerontology, in press: J. L. Sullivan. ibid. 28. 242(1982). 142. R. Peto. Proc. R. Soc. London Ser. 8 205. 111 (1979); D. Dix. P. Cohen. J. Flannery. J. Theor. Biol. 83. 163 (1980). 143. R. G. Cutler, in Testing the Theories ofAging. R. Adelman and G. Roth. Eds. (CRC Press, Boca Raton. Fla., in press). 144. D. Hersh. R. G. Cutler. B. N. Ames, in prepa ration. 145. E. Boyland, in Health Risk Analysis (Franklin Institute Press, Philadelphia. 19801. pp. 181193. E. Boyland. in Cancer Campaign, vol. 6. Cancer Epidemiology. E. Grundmann. Ed. (Fischer. Stuttgart, 1982). pp. 125-128. 146. J. L. Man. Science 219. 158 (1983). 147. B. D. Goldstein. G- Witz. M. Amoruso. D. S. Stone. W. Troll. Cancer Lett. II, 257 (1981); W. Troll, in Environmental Mutagens and Car cinogens. T. Sugimuru. S. Kondo. H. Takebe. Eds. (Univ. of Tokyo Press. Tokyo, and Lisa. New York. 1982). pp. 217-222. 148. B. N. Ames. M. C. Hollstem. R. Cathcart. in Lipid Peroxide in Biology and Medicine. K. Yagi. Ed. (Academic Press. New York, 1982), pp. 339-35). 149. I. Emerit and P. A. Cerutti. Proc. Natl. Acad. Sci. U.S.A. 79. 7509 (1982); Nature ILondoni 293, 144 (1981): p. A. Cerxmi. I. Emerit. P. Atnstad. in Genes und Proteins in Oncogene sis. I. B. Weinstein and H. Vogel. Eds. lAcademic Press. New York, in press); I. Ement. A. Levy. P. Cerutti. Mutat. Res. 110. 327 (1983). 150. J. Funes and M. Karel. Lipids 16, 347 (1981). 151. S. J. Siohs. M. Q- Hassan. W. J. Murray. Biochem. Biophys. Res. Commun. 111. 854 (1983). 152. C. C. Reddy. R. W. Schotz. E. J. Massaro. Toxicol. Appl. Pharmacol. 61. 460 (19811. 153. H. Nagasawa and J. B. Lillie. Carcinogenesis 2. 601 (1981): V. Solanki. R. S. Rana. T. J. Slaga. ibid., p. 1141; T. W. Kenslcr and M. A. Trush, Cancer Res 41. 216 (1981), 154. R. H. Simon. C. H. Scoggin. D. Patterson. J. Biol. Chem. 256.7181 (1981 j:T. S. Argyris and T. J. Slaga. Cancer Res. 41. 5193 (1981>. . 155. G. E. Hatch. D. E. Gardner. D. B. Menzel. Environ. Res. 23. 121 (1980)- 156. A. J.P. Klein-Szanto and T. J. Slaga. J. Insest. Dermatol. 79 . 30 (l982l. 157. C. C. Weddle. K. R. Hombrook. P. B. McCav. /. Biol. Chem. 251. 4973 (1976). 158. A. G. Lurie and L. S. Cutler, J. Natl. Cancer last. 63. 147 (1979). 159. C. Borek and W. Troll. Proc Natl. Acad. Sci. U.S.A. 80. 1304 11983); C. Borek. in Molecular Interrelations of Nutrition and Cancer. M. S. Amott. J. van Eys. Y.-M. Wang. Eds. (Raven. New York. 1982). pp. 337-350. 160. T. J, Slaga et al.. in Carcinogenesis: A Com prehensive Treatise tRaven. New York. 1982). vol. 7. pp. 19-34; K. Ohuchi and L. Levine. Biochim. Biophys. Acta 619, II (1980): S. M. Fischer. G. D. Mills. T. J. Slaga. Carcinogene sis 3. 1243 (1982). 161. R. P. Mason, in Free Radicals in Biology. W. A- Pryor. Ed. (Academic Press. New York. 1982). vol. 5, pp. 161-222. 162. G. McLennan. L. W. Oberley. A. P. Autor. Radial. Res. 84. 122 (1980). 163. I- M. Parry, E. M. Parry. J. C. Barrett. Nature (London/ 294. 263 (1981): A. R. Kinselta. Car cinogenesis 3. 499 (19821. 164. H. C. Bimboim, Can. J. Physiol. Pharmacol. 60. 1359 (1982). 165. A. Varshavsky. Cell 25. 561 (1981). 166. T. J. Powles et al., Eds. Prostaglandins and Cancer.- First International Conference iLlSS. New York. 1982). 23 SEPTEMBER 1983 1263 f 167. J. E. Trosko, C.*C. Chang. A. Medcalf, Cancer Invest., in press. 167a.I. B. Weinstein, Nature (London/ 302, 750 (1983). 168. J. A. Bond. A. M. Gown. H. L. Yang. E. P. Bendilt. M. R. Juchau, J. Toxicol. Environ. Health 7. 327 (1981). 169. Editorial, Lancet 1980-1. 9M. (1980). K. Yags. H. Ohkawa. N. Ohishi. M. Yamashita. T. Nakashima. J. Appl. Biochem. 3. 58 11981). 170. J. M. C. Gutteridge, B. Halliwell, D. A. Row- ley. T. W<stermarck. Lancet >982-11. 459 (1982). 171. I. E. Cleaver, in Metabolic Basis of Inherited Disease. I. B. Sianbury. J. B. Wynguarden, D. S. Fredrickson. J. L. Goldstein. Eds. (McGraw-Hill. ed. 5, New York. 1983). pp. 1227-1250. 172. K. C. Bhuyan, D. K. Bhuyan. S. M. Podos, IRCS Med. Sci. 9. 126 (1981); A. Spector. R. Scotto. H. Weissbach. N. Brot, Biochem. Biophxs. Res. Commun. 108. 429 (1982); S. D. Varma, N. A. Beachy, R. D. Richards, Photo- Chem. Photohiol. 36, 623 (1982). 173. M. L. Katz. K. R. Parker. G. 1. Handelman. T. L. Bramel. E. A. Dratz. Exp. Eve Res. 34. 339 (1982). ' 174. D- Behne. T. Hofer. R. von Berswordt-Wall- rabe. W. Eiger, /. Nuu. l>2. 1682 (1982). 175. B. N. Ames. L. S. Gold, C. B. Sawyer. W. Havender. in Environmental Mutagens and Carcinogens. T. Sucimura, S. Kondo. H. Ta- kebe. Eds. (Univ. of Tokyo Press. Tokyo, and L-iss. New York. 1982), pp. 663-670. 176. National Center for Health Statistics. Advance Report. Final Mortality Statistics. 1979. Monthly Vital Statistics Report 31. No. 6. suppt. [DHHS publication (PHS) 82-1120, (Public Health Service. Hyattsville. Md., 1982]; Metropolitan Life Insurance Company- Actuarial Tables. April 1983. 177. B. A. Bridges. B. E. Butterworth. I. B. Wein stein. Eds., Banbury Report IJ. Indicators of Ctnotoxic Exposure (Cold Spring Harbor Lab oratory. Cold Spring Harbor. N.Y.. 1982): R. Montesarto. M. F. Rnjewsky. A. E. Pega. E. Miller. Cancer Res. 42. 5236 (1982); H- F. Stich, R. H. C. San. M. P Rosin. Ann. N.Y. Acad. Sci-. in press; t. 8. Weinstein. Annu. Rev. Public Health 4. 409 (1983). 178. 1 am indebted to G. Ferro-Luzzi Ames. A. Blum, L. Gold, P. Hartman, W. Havender. N. K. Hooper. G. W. Ivie. J. McCann. J. Mead. R. Olson. R. Peto. A. Tappel. and numerous other colleagues for their criticisms. This work was supported by DOE contract DE-AT0376EV70I56 to B.N.A. and by National Insti tute of Environmental Health Sciences Center Grant ES01896. This article has been expanded from a talk presented at the 12th European Environment! Mutagen Society Conference. Espoo. Finland. June 1982 [in .Wiuncfnj in Our Environment. M. Sorsa and H Vainio. Eds. (Liss. New York. 1982)]. [ wish to dedicate this article to the memory' of Philip Handier, pio neer in the held of oxygen radicals. URL 04017 RESEARCH ARTICLE Imaging Dopamine Receptors in the Human Brain by Positron Tomography Henry N. Wagner, Jr., H. Donald Bums, Robert F. Dannals Dean F. Wong, Bengt Langstrom, Timothy Duelfer, J. James Frost Hayden T. Ravert, Jonathan M. Links, Shelley B. Rosenbloom Scott E. Lukas. Alfred V. Kramer. Michael J. Kuhar cose. ["CJcarboxyhemoglobin. ionic nibidium-82. ^Ga-labeied EDTA. and oth er radiopharmaceuticals, and subsequent imaging of the distribution of the radio active label in the brain by means of the tomographic method, based on detection of the annihilation radiation produced during positron emission (9), The butyrophenone neuroleptic drug spiperone has been useful in binding studies for measuring dopamine recep tors both in vitro [10) and in vivo (//). We now report initial results obtained with 3-<V-[nC|methyIspiperone (MCNMSP), a spiperone derivative, in PET One of the most intriguing problems in as a result of neuroleptic therapy (4). biomedical research today is that of re The development of positron emission lating manifestations of neuropsychiatric tomography (PET) and appropriate ra scanning studies to visualize the distribu diseases to chemical processes in differ dioactive tracers labeled with positron- tion of dopamine receptors in the brains ent parts of the brain. The neurotrans emitting radionuclides has now made it of baboons and a human being. All stud ies were performed with a NeuroECAT scanner (Ortec, Inc.. Oak Ridge. Ten Abstract. Neurotransmitter receptors may be involved in a number of nettropsy- nessee), which has a spatial resolution of chiairic disease states. The ligand 3-N-["C]methyispiperone. which preferentially approximately 8 mm (full width at half binds to dopamine receptors in vivo, was used to image the receptors by positron maximum) in the plane of the slice. The emission tomography scanning in baboons and in humans. This technique holds distance between slices is 3 cm. promise for noninvasive clinical studies of dopamine receptors in humans. The newly developed tracer llC- NMSP was synthesized by iV-alkylation of spiperone with [nC]methyl iodide; the mitter dopamine appears to be associat possible to relate regional biochemistry iodide was produced from "CO;, which ed with abnormalities related to disor within the human brain to measurements in turn had been produced with an in- ders such as Parkinson's disease and of behavior in normal subjects and to hospital cyclotron (model RNP-16, Scan- schizophrenia. The highest density of elucidate abnormalities in patients with ditronix Cyclotron, Sweden). Carbon-11 dopamine neurons occurs in the nigro- Alzheimer's disease (J), Huntington's is a positron-emitting isotope with a striatal dopamine pathway which degen disease (6). depression (7). and multiple physical half-life of 20 minutes. The en erates in Parkinson's disease (/). Neuro infarct dementia (<S). The technique con tire synthesis was accomplished with leptic drugs elicit extrapyramidal parkin sists of intravenous injection of a sub material ready for injection within 55 sonian side effects by blocking dopamine stance such as l8F-labeled deoxyglu- minutes after the end of the cyclotron receptors in the corpus striatum and also exert antischizophrenic action by block ing dopamine receptors, perhaps in lim bic areas (2). Numbers of dopamine re ceptors are increased by chronic neuro leptic treatment (J) and are also in creased in some schizophrenics, perhaps Henry N. Warner, Jr.. H. Donald Burn*, Robert F. Dannals. Dean F. Wong, Timothy Duelfer. J. James Front. Hayden T. Ravert, Jonathan M. Links, and Alfred V. Kramer are in [he Div ision of Nuclear Medicine. Johns Hopkins Medical Institutions. Baltimore. Maryland 21205. Bengt Langstrom is in the Institute of Chemistry. University of Uppsala. S75I2I Uppsala I. Sweden. Shelley B. Rosenbloom is m the Divimuo of Neuroradiology, Johns Hopkins Medical Institutions. Scott E. Lukas is at the NIDA Addiction Research Center. Baltimore City Hospitals. Baltimore. Maryland 21224. Michael J. Kuhar is in the Departments of Neuroscience. Pharmacology and Experimental Therapeutics, and Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine. Baltimore. Maryland 212D5. Correspondence should he sent to Henry N. Wagner. Jr. ............................ .......... ' t Reprint Series 23 September 1983, Volume 221, pp. i:j9 Dietary Carcinogens Philip H. Abelson Copyright 1983 by the American Association for the Advancement of Science URL 04018 UOW W i Dietary Carcinogens In this issue of Science. Bruce Ames reviews the increasing body of evidence that large numbers of potent carcinogens arise from natural processes. Mutagens are present in substantial quantities in fruits and vegetables. Carcinogens are formed in cooking as a result of reactions involving proteins or fats. Dietary practices may be an important determi nant of current cancer risks. Ames describes the role of plant materials as follows: Plants in nature synthesize toxic chemicals in large amounts, apparently as a primary defense against the hordes of bacterial, fungal, and insect and other animal predators. Plants in the human diet are no exception. The variety of toxic chemicals is so great that organic chemists have been characterizing them for over 100 years, and new plant chemicals are still being discovered. Recent widespread use of short-term tests for detecting mutagens and the increased testing of plant substances for carcinogenicity in animals have contributed to the identification of many natural mutagens, teratogens, and carcinogens in the human diet. Safrole and related compounds are present in many edible plants. Safrole is a carcinogen in rodents and some of its metabolites are mutagens. Oil of sassafras, once used to flavor some root beer, is about 75 percent safrole. Black pepper contains about 10 percent by weight of a closely related compound, pipeline. Extracts of black pepper at a dose equivalent to 4 milligrams of dried pepper per day cause tumors in mice at many sites. Many hydrazines are carcinogens and mutagens, and large amounts of them are found in edible mushrooms. One carcinogenic hydrazine is present in the false morel at a concentration of 50 milligrams per 100 grams. It causes lung tumors in mice at a level of 20 micrograms per mouse per day. Carcinogens and mutagens are present in mold-contaminated foods such as corn. nuts, peanut butter, bread, cheese, and fruit. Some of these contaminants, such as aflatoxin. are among the most potent known carcino gens and mutagens. Nitrosamines and nitroso compounds are suspect as causative agents of stomach and esophageal cancer in humans. In the digestive system these nitrogen compounds are formed from nitrate and nitrite. Beets, celery, lettuce, spinach, radishes, and rhubarb all contain about 200 milligrams of nitrate per 100-gram portion. Rancid fats are possible causative agents of colon and breast cancer in humans. These forms account for a substantial fraction of all the cancer deaths in the United States. Unsaturated fats are easily oxidized on standing and in cooking to form mutagens, promoters, and carcinogens. Among the numerous products of such oxidations are fatty acid hydroperoxides and cholesterol epoxide. Thus the colon and digestive tract are exposed to many fat-derived carcinogens. Human breast fluid can contain high levels of cholesterol epoxide. Burnt and browned materials formed by heating proteins during cooking are highly mutagenic. Chemicals isolated from such products have been found to be carcinogenic when fed to rodents. In addition, the browning reaction products from caramelization of sugars or the reaction of amino acids and sugars during cooking contain a large variety of DNA-damagtng agents. The view that dietary practices might be a causative factor in cancer is not new. Epidemiologists have noted marked differences in cancer rates be tween population groups. Effects from changes in diet following migration have also been observed. Results of current studies are beginning to delineate more sharply specific causative agents. When more definitive information is available, it should be possible for prudent people to choose fruits and vegetables that present minimal hazards. In the meantime, there is persuasive evidence that charred meats and rancid fats should not be pan of the diet.--Philip H. Abelson Reprint Series 23 September 1983. Volume 221. pp. 1256-1264 Dietary Carcinogens and Anticarcinogens Bruce N, Ames Copyright 1983 by the American Association for the Advancement of Science IJRL 04020 amounts of safrole and large amounts (close to 10 percent by weight) of the closely related compound piperine (36). Extracts of black pepper cause tumors in mice at a variety of sites at a dose of Dietary Carcinogens and extract equivalent to 4 mg of dried pep per per day (about 160 mg'kg per day) for Anticarcinogens 3 months: an estimate of the average human intake of black pepper is over 140 mg per day (about 2 mg-kg per day) for Oxygen radicals and degenerative diseases life (26). 2) Most hydrazines that have been Bruce N. Ames tested are carcinogens and mutagens, and large amounts of carcinogenic hy drazines are present in edible mush rooms. The widely eaten false morel (Gyromitru escnlenta) contains 11 hydra Comparison of data from different Natural Mutagens and zines. three of which are known carcino countries reveals wide differences in the Carcinogens in Food rates of many ty pes of cancer. This leads gens (28). One of these. .V-methyJ-.Vformylhydrazme. is present at a concen to hope that each major type of cancer Plant material. Plants in nature syn tration of 50 mg per 100 g and causes may be largely avoidable, as is the case thesize toxic chemicals in large amounts, lung tumors in mice at the extremely low for cancers due to tobacco, which consti apparently as a primary defense against dietary level of 20 p.g per mouse per day tute 30 percent of the cancer deaths in the hordes of bacterial, fungal, and in (25). The most common commercial the United States and the United King sect and other animal predators (5-40). mushroom. Agariciis bi.sporas. contains dom (/). Despite numerous suggestions Plants in the human diet are no excep about 300 mg of aguricine. the 6-glutamy I to the contrary, there is no convincing tion. The variety of these toxic chemi derivative of the mutagen 4-hydroxy- evidence of any generalized increase in U.S. lor U.K.l cancer rates other than what could plausibly be ascribed to the Summary. The human diet contains a great variety of natural mutagens and delayed effects of previous increases in carcinogens, as well as many natural antimutagens and anticarcinogens. Many of tobacco usage (/-J). Thus, whether or tese mutagens and carcinogens may act through the generation of oxygen radicals. not any recent changes in life-style or Oxygen radicals may also play a major role as endogenous initiators of degenerative pollution in industrialized countries will processes, such as DNA damage and mutation (and promotion), that may be related substantially affect future cancer risks, some important determinants of current risks remain to be discovered among to cancer, heart disease, and aging. Dietary intake of natural antioxidants could be an important aspect of the body's defense mechanism against these agents. Many antioxidants are being identified as anticarcinogens. Characterizing and optimizing long-established aspects of our way of such defense systems may be an important part of a strategy of minimizing cancer life. Epidemiologic studies have indicat and other age-related diseases. ed that dietary practices are the most promising area to explore (/. 4). These studies suggest that a general increase in cals is so great that organic chemists methylphenylhydrazine. per 100 g of consumption of fiber-rich cereals, vege have been characterizing them for over mushrooms, as well as smaller amounts tables. and fruits and decrease in con 100 years, and new plant chemicals are of the closely related carcinogen A'-ace- sumption of fat-rich products and exces still being discovered 112. 24. 25). How tyl - 4 - hydroxymethylphenylhydrazine sive alcohol would be prudent (/, 4). ever. toxicological studies have been (28). Some agaritine is metabolized by There is still a lack of definitive e\ idence completed for only a very small percent the mushroom to a diazomum derivative about the dietary components that are age of them. Recent widespread use of which is a very potent carcinogen (a critical for humans and about their mech short-term tests for detecting mutagens single dose of 400 ng ggave 30 percent of anisms of action. Laboratory studies of (41. 42) and the increased number of mice stomach tumors) and which is also natural foodstuff's and cooked food are animal cancer tests on plant substances present in the mushroom in smaller beginning to uncover an extraordinary- (6) have contributed to the identification amounts i25). Many hydrazine carcino variety of mutagens and possible carcin of many natural mutagens, teratogens, gens may act by producing oxygen radi ogens and anticarcinogens. In this article and carcinogens in the human diet (5- cals (43). 1 discuss dietary mutagens and carcino 40). Sixteen examples are discussed be 3) Linear furocoumarins such as psor gens and anticarcinogens that seem of low. alen derivatives are potent light-activat importance and speculate on relevant 1) Safrole. estragole. methylettgeno!. ed carcinogens and mutagens and are biochemical mechanisms, particularly and related compounds are present in widespread in plants of the Umbelliferae the role of oxygen radicals and their many edible plants (5). Safrole. estra- family, such as celery, parsnips, figs, inhibitors in the fat-cancer relationship, gole. and methyleugenol are carcinogens and parsley (for instance. 4 mg per 100 g promotion, anticarcinogenesis, and ag in rodents, and several of their metabo of parsnip) (17. 19. 44). The level in ing. lites are mutagens (5i. Oil of sassafras, celery (about 100 p-g per 100 g) can The author is ch'uirmnn of the Dupurtmenl of Biochemisirv. L'niversuv of California. Berkeley which had been used in "natural" sarsa parilla root beer, is about 75 percent safrole. Black pepper contains small increase about 100-fold if the celery is stressed or diseased (19). Celery pickers and handlers commonly develop skin 1256 SCIENCE. VOL. ::t rashes on their arms when exposed to' diseased celery (19). Oil of bergamot, a citrus oil. is very rich in a psoralen and was used in the leading suntan lotion in France 1/7). Psoralens, when activated by sunlight, damage DNA and induce tanning more rapidly than the ultraviolet component of sunlight, which is also a carcinogen 1/7). Psoralens (plus light) are also effective in producing oxygen radicals 118). 4) The potato glycoalkaloids solanine and chaconine are strong cholinesterase inhibitors and possible teratogens and are present at about 15 mg per 200 g of potato (12. 13). When potatoes are dis eased. bruised, or exposed to light', these and other (24) glycoalkaloids reach lev els that can be lethal to humans (12). Plants typically respond to damage by making more (and often different) toxic chemicals as a defense against insects and fungi {19. 24. 25). The different cultivars of potatoes vary in the concentra tion of these toxic glycoalkaloids (the concentration is a major determinant of insect and disease resistance); one cultivar bred for insect resistance had to be withdrawn from use because of its toxici ty to humans (> 40 mg of glycoalkaloids in a 200-g potato is considered to be a toxic level) (12). 5) Quercetin and several similar flavonoids are mutagens in a number of short-term test systems. Flavonoids are extremely widespread (daily levels close to l g) in the human diet IS. 16. 20. 21). There is evidence for the carcinogenicity of quercetin in two strains of rats (), although it was negative in other experi ments (21). 6) Quinones and their phenol precur sors (9. 14, 16, 23. 45) are widespread in the human diet. Quinones are quite toxic as they can act as electrophiles or accept a single electron to yield the semiquinone radical, which can either react di rectly with DNA (14. 46) or participate in a redox cycle of superoxide radical gen eration by transferring the electron to O; (47). The superoxide radical and its met abolic product H:0: can. in turn, lead to the oxidation of fat in cellular mem branes by a lipid peroxidation chain re action. thus generating mutagens and carcinogens, as discussed below. A num ber of quinones and dietary phenols have been shown to be mutagens (7. 9, 16, 23, 44). Mutagenic anthraquinone deriva tives are found in plants such as rhubarb and in mold toxins (7. 16, 48). Many dietary phenols can spontaneously autoxidtze to quinones. generating hydro gen peroxide at the same time [examples are catechol derivatives such as the caf- 23 SEPTEMBER 1983 feic acid component of chlorogenic acid (9). which is present at about 250 mg per cup of coffee). The amounts of these phenols in human urine (and in the diet) are appreciable (45). Catechol, for exam ple. is excreted in urine at about 10 mg per day and appears to be mainly derived from metabolism of plant substances (45). Catechol is a potent promoter of carcinogenesis (45). an inducer of DNA damage, a likely active metabolite of the carcinogen benzene (46). and a toxic agent in cigarette smoke (45). Catechol amine induction of cardiomyopathy is thought to occur through generation of oxygen radicals (49). 7) Theobromine, a relative of caffeine, has been shown to be genotoxic in a variety of tests, to potentiate (as does caffeine) DNA damage by various car cinogens in human cells, and to cause testicular atrophy and spermaiogenic cell abnormalities in rats (27). Cocoa powder is about 2 percent theobromine, and therefore humans may consume hundreds of milligrams of theobromine a day from chocolate. Theobromine is also present in tea. 8) Pyrrolizidine alkaloids are carcino genic. mutagenic, and teratogenic and are present in thousands of plant species (often at > 1 percent by weight), some of which are ingested by humans, particu larly in herbs and herbal teas and occa sionally in honey (7. 29). Pyrrolizidine alkaloid poisonings in humans (as well as in other mammals) cause lung and liver lesions and are commonly misdiagnosed (29). 9) The broad (fava) bean (Vicia faba), a common food of the Mediterranean region, contains the toxins vieme and convicine at a level of about 2 percent of the dry weight (30). Pythagoras forbade his followers to eat the beans, presum ably because he was one of the millions of Mediterranean people with a deficien cy of glucose-6-phosphate dehydrog enase. This deficiency results in a low glutathione concentration in blood cells, which causes increased resistance to the malarial parasite, probably accounting for the widespread occurrence of the mutant gene in malarial regions. Howev er. the low glutathione concentration also results in a marked sensitivity to agents that cause oxidative damage, such as the fava bean toxins and a varie ty of drugs and viruses. Sensitive indi viduals who ingest fava beans develop a severe hemolytic anemia caused by the enzymatic hydrolysis of vicine to its aglycone. divicine. which forms a quinone that generates oxygen radicals (30). 10) Allyl isothiocyoncite. a major flavor ingredient in oil of mustard and horse radish. is one of the main toxins of the mustard seed and has been shown to cause chromosome aberrations in ham ster celts at low concentration (50) and to be a carcinogen in rats (31). 11) Gossypol is a major toxin in cot tonseed and accounts for about 1 percent of its dry weight i52). Gossypol causes pathological changes in rat and human testes, abnormal sperm, and male steril ity (32. 33). Genetic damage has been observed m embryos sired by gossypol* treated male rats: dominant lethal muta tions in embryos were measured after males were taken off gossypol treatment and allowed to mate (55). Gossypol ap pears to be a carcinogen as well: it has been reported to be a potent initiator and also a promoter of carcinogenesis in skin painting studies with mice (34). Crude, unrefined cottonseed oil contains consid erable amounts of gossypol (100 to 750 mg per 100 mi). Thus human consump tion may be appreciable in countries, such as Egypt, where fairly crude cot tonseed oil is commonly used in cooking. Gossypol is being tested as a male con traceptive in over 10,000 people in China (at an oral dose of about 10 mg per person per day), as it is inexpensive and causes sterility during use (55). Gossypol's mode of action as a spermicide may be through the production of oxygen radicals (35). Plant breeders have developed '*glandless cotton." a new strain with low levels of gossypol. but seeds from this strain are much more susceptible to attack by the fungus Aspergillus flavus. which pro duces the potent carcinogen aflatoxin (56). 12) Sterculic odd and malvalic odd are widespread in the human diet. They are toxic cyclopropenoid fatty acid? present in cottonseed oil and other oils from seeds of plants in the family Malvaceal (for instance, cotton, kapok, okra, and durian) i5/L Another possible source of human exposure is consump tion offish, poultry, eggs, and milk from animals fed on cottonseed (51). Cyclo propenoid fatty acids are carcinogens in trout, markedly potentiate the carcinoge nicity of aflatoxin in trout, cause athero sclerosis in rabbits, are mitogenic in rats, and have a variety of toxic effects in farm animals (51). The toxicity of these fatty acids could be due to their ease of oxida tion to form peroxides and radicals (51). 13) Leguminous plants such as lupine contain very potent teratogens (22). When cows and goats forage on these plants, their offspring may have severe teratogenic abnormalities: an example is 1 URL 040 ro po the characteristic "crooked calf abnor mality due to the ingestion of unavyrirle from lupine (22). In addition, significant amounts of these teratogens are trans ferred to the animals' milk, so that drink ing the milk during pregnancy is a seri ous teratogenic hazard (22). In one rural California family. a baby boy. a litter of puppies, and goat kids all had "crooked" bone birth-defect abnormali ties. The pregnant mother and the dog had both been drinking milk obtained from the family goats, which had been foraging on lupine (the main forage in winter) (22). It was at first mistakenly thought that the birth defects were caused by spraying of 2.4-D. |4i Sesquiterpene lactones are wide spread m many plants (37). although because they are bitter they are not eaten in large amounts. Some have been shown to be mutagenic (57). They are a major toxin in the w hite sap of Lactuca \irosa (poison lettuce), which has been used as a folk remedy. Plant breeders are now transferring genes from this species to commercial lettuce to increase insect resistance (38). 15) The phorhol esters present in the Euphorbiacea. some of which are used as folk remedies or herb teas, are potent promoters of carcinogenesis and may have been a cause of nasopharyngeal cancer in China and esophageal cancer in Curasao (39). 16) Alfalfa sprouts contain cunavanine. a highly toxic arginine analog that is incorporated into protein in place of argi nine. Canavanine. which occurs in alfal fa sprouts at about 1.5 percent of their dry weight (40). appears to be the active agent in causing the severe lupus erythematosus-Uke syndrome seen when mon keys are fed alfalfa sprouts (40). Lupus in man is characterized by a defect in the immune system which is associated with autoimmunity, antinuclear antibodies, chromosome breaks, and various types of pathology (40). The chromosome breaks appear to be due to oxygen radi cals as they are prevented by superoxide dismutase (52). The canavanine-alfalfa sprout pathology could be due in part to the production of oxy gen radicals during phagocytization of antibody complexes with canavanine-conlainmg protein. The 16 examples above, plus coffee idiscussed below), illustrate that the hu man dietary intake of "nature's pesti cides" is likely to be several grams per day--probably at least 10.000 times high er than the dietary intake of man-made pesticides (55). Levels of plant toxins that confer in sect and funeal resistance are being in creased or decreased by plant breeders 1258 (38). There are health co>ts for the use of these natural pesticides, just as there are for man-made pesticides (41. 54). and these must be balanced against the costs of producing food. However, little infor mation is available about the toxicology of most of the natural plant toxins in our diet, despite the large doses we are ex posed to. Many, if not most, of these plant toxins may be "new'" to humans in the sense that the human diet has changed drastically with historic times. By comparison, our knowledge of the toxicological effects of new man-made pesticides is extensive, and general ex posure is exceedingly low (55). Plants also contain a variety of anticarcinogens (55). which are discussed be low. Alcohol. Alcohol has long been associ ated with cancer of the mouth, esopha gus. pharynx, larynx, and. to a lesser extent, liver (/. 56). and it appears to be an important human teratogen, causing a variety of physical and mental defects in babies of mothers who drink (57). Alco hol drinking causes abnormalities in mice (57a) and is a synergist for chromo some damage in humans i58). Alcohol metabolism generates acetaldehyde, which is a mutagen and teratogen (59). a cocarcinogen, and possibly a carcinogen (601. and also radicals that produce lipid hydroperoxides (61) and other mutagens and carcinogens (62; see below). In some epidemiologic studies on alcohol (56). it has been suggested that dietary green vegetables are a modifying factor in the reduction of cancer risk. Mold carcinogens. A variety of mold carcinogens and mutagens are present in mold-contaminated food such as corn, grain, nuts, peanut butter, bread, cheese, fruit, and apple juice (15. 63). Some of these, such as sterigmatocystin and aflatoxin. are among the most potent carcinogens and mutagens known U5. 63). Dietary glutathione has been report ed to counteract aflatoxin carcinogenic ity. Nitrite, nitrate, and nitrosamines. A number of human cancers, such as stom ach and esophageal cancer, may be relat ed to nitrosamines and other nitroso compounds formed from nitrate and ni trite in the diet (64. 65). Beets, celery, lettuce, spinach, radishes, and rhubarb all contain about 200 mg of nitrate per 100-g portion (65). Anticarcinogens in the diet may be important in this context as well (66). Fat and cancer: possible oxidative mechanisms. Epidemiologic studies of cancer in humans suggest, but do not prove, that high fat intake is associated with colon and breast cancer (/, 4. 67). A number of animal studies have shown that high dietary fat is a promoter and a presumptive carcinogen 14. 67. 68). Co lon and breast cancer and lung cancer (which is almost entirely due to cigarette smoking) account for about half of all ITS. cancer deaths. In addition to the cyciopropenoid fatty acids already dis cussed. two other plausible mechanisms involving oxidative processes could ac count for the relation (69) between high fat and both cancer and heart disease. J) Rancid fat. Fat accounts for over 40 percent of the calories in the U.S. diet (67). and the amount of ingested oxidized fat may be appreciable 170. 711. Unsaturated fatty acids and cholesterol in fat are easily oxidized, particularly during cooking (70. 71). The lipid peroxidation chain reaction (rancidity i yields a variety (71-73) of mutagens, promoters, and car cinogens such as fatty acid hydroperox ides (62). cholesterol hydroperoxide (74). endoperoxides. cholesterol and fat ty acid epoxides 174-77). enals and other aldehydes (44. 59. 78). and alkoxy and hydroperoxy radicals (44. 72). Thus the colon and digestive tract are exposed to a variety of fat-derived carcinogens. Hu man breast fluid can contain enormous levels (up to 780 p.,V/) (75) of cholesterol epoxide (an oxidation product of choles terol). which could originate from either ingested oxidized fat or oxidative pro cesses in body lipids. Rodent feeding studies with oxidized fat (79i have not yielded definitive results. 2) Peroxisomes oxidize an appreciable percentage of dietary fatty acids, and removal of each two-carbon unit gener ates one molecule of hydrogen peroxide (a mutagen, promoter, and carcinogen) (80, HI). Some hydrogen peroxide es capes the catalase in the peroxisome (80. 82. 83). thus contributing to the supply of oxygen radicals, which aiso come from other metabolic sources (72. 83-85). Hy droperoxides generate oxygen radicals in the presence of iron-containing com pounds in the cell (72). Oxygen radicals, in turn, can damage DNA and can start the rancidity chain reaction which leads to the production of the mutagens and carcinogens listed above (72). Drugs such as clofibrate. which cause lowering of serum lipids and proliferation of per oxisomes in rodents, result in age pig ment (lipofuscin) accumulation (a sign of lipid peroxidation in tissues) and liver tumors in animals (80). Some fatty acids, such as C;;. i and certain trans fatty acids, appear to cause peroxisomal pro liferation because they are poorly oxi dized in mitochondria and are preferen tially oxidized in the peroxisomes, al though they may be selective for heart or science, vol. ::i URL 04023 * URL O4024 liver (56). There has been controversy about the role of tran.s fatty acids in cancer and heart disease, and recent evidence suggests that trans fatty acids might not be a risk factor for atheroscle rosis tn experimental animals (#7). Americans consume about 12 g of iron', fatty acids a day (57) and a similar amount of unnatural ct'.v isomers [which need further study (55>j. mainly from hydrogenated vegetable fyts. Dietary C:; i fatty acids are also obtained from rapeseed oil and fish oils (56). Thus oxidation of certain fatty acids might generate grams of hy drogen peroxide per day within the peroxisome (56). Another source of fat toxicity could be perturba tions in the mitochondrial or peroxisom al membranes caused by abnormal fatty acids, yielding an increased flux of su peroxide and hydrogen peroxide. Mito chondrial structure is altered when rats are fed some abnormal fatty acids from partially hydrogenated fish oil (59). Di etary C:: i fatty acids and clofibrate also induce ornithine decarboxylase (56). a common attribute of promoters. A recent National Academy of Sci ences committee report suggests that a reduction of fat consumption in the American diet would be prudent (-/). although other scientists argue that, until we know more about the mechanism of the fat-cancer relation and about which types of fat are dangerous, it is prema ture to recommend dietary changes (90). Cooked Food as a Source of Ingested Burnt and Browned Material Work of Sugimura and others has indi cated that the burnt and browned materi al from heating protein during cooking is highly mutagenic (21. 91). Several chem icals isolated on the basis of their mu tagenicity from heated protein or pyrolyzed amino acids were found to be carcinogenic w hen fed to rodents (21). In addition, the browning reaction products from the caramelization of sugars or the reaction of amino acids and sugars dur ing cooking (for instance, the brown ma terial on bread crusts and toasted bread) contain a large variety of DNA-damaging agents and presumptive carcinogens (23. 38. 92). The amount of burnt and browned material in the human diet may be several grams per day. By compari son about 500 mg of burnt material is inhaled each day by a smoker using two packs of cigarettes (at 20 mg of tar per cigarette) a day. Smokers have more easily detectable levels of mutagens in their urine than nonsmokers (95). but so do people who have consumed a meal of :? SEPTEMBER 1983 fried pork or bacon i94). In the evalua tion of risk from burnt material it may be useful (in addition to carrying out epide miologic studies) to compare the activity of cigarette tar to that of the burnt mate rial from cooked food (or polluted air) in short-term tests and animal carcinoge nicity tests involving relevant routes of exposure. Route of exposure and com position of the burnt material are critical variables. The risk from inhaled cigarette smoke can be one reference standard: an average life shortening of about 8 years for a tw'o-pack-a-day smoker. The amount of burnt material inhaled from severely polluted city air. on the other hand, is relatively small: it would be necessary to breathe smoggy Los Ange les air (111 fAg/nv total particulates: 31 p.g/rrv' soluble organic matter) for 1 to 2 weeks to equal the soluble organic mat ter of the particulates or the mutagenic ity from one cigarette (20 mg of tar) (95). Epidemiologic studies have not shown significant risks from city air pollution alone (/, 96). Air in the houses of smok ers is considerably more polluted than city air outside (97). Coffee, which contains a considerable amount of burnt material, including the mutagenic pyrolysis product methylglyoxal. is mutagenic 12/. 95). However, one cup of coffee also contains about 250 mg of the natural mutagen chlorogenic acid (9) [which is also an antinitrosating agent (66)\. highly toxic atractylosides (10). the glutathione transferase inducers kahweal palmitate and cafesioi palmitate l / /). and about 100 mg of caffeine [which inhibits a DNA-repair system and can increase tumor yield (99) and cause birch defects at high levels in several experi mental species (100)J. There is prelimi nary. but not conclusive, epidemiologic evidence that heavy coffee drinking is associated with cancer of the ovary, biadder, pancreas, and large bowel (iOI). Cooking also accelerates the rancidity reaction of cooking oils and fat in meat (70. 7/1. thus increasing consumption of mutagens and carcinogens. Anticarcinogens We have many defense mechanisms to protect ourselves against mutagens and carcinogens, including continuous shed ding of the surface layer of our skin, stomach, cornea, intestines, and colon (702). Understanding these mechanisms should be a major goal of cancer, heart, and aging research. Among the most important defenses may be those against oxygen radicals and lipid peroxidation if. as discussed here, these agents are major contributors to DNA damage (103). Ma jor sources of endogenous oxygen radi cals are hydrogen peroxide )5Ji and su peroxide (72, 104) generated as side products of metabolism, and the oxygen radical burst from phagocytosis after vi ral or bacterial infection or the inflamma tory reaction 1105). A variety of environ mental agents could also contribute to the oxygen radical load, as discussed here and in recent reviews (72. 106). Many enzymes protect cells from oxida tive damage; examples are superoxide dismutase (1041. glutathione peroxidase (107). DT-diaphorase (108). and the glu tathione transferases (109). In addition, a variety of small molecules in our diet are required for antioxidative mechanisms and appear to be anticarcinogens: some of these are discussed below. 1) Vitamin E (tocopherol) is the major radical trap in lipid membranes (72) and has been used clinically in a variety of oxidation-related diseases I I10). Vitamin E ameliorates both the cardiac damage and carcinogenicity of the quinones adriamvcin and daunomyein. which are mutagenic, carcinogenic, cause cardiac damage, and appear to be toxic because of free radical generation (/!/). Protec tive effects of tocopherols against radia tion-induced DNA damage and mutation and dimethylhydrazine-induced carcino genesis have also been observed \!I2). Vitamin E markedly increases the endur ance of rats during heavy exercise, which causes extensive oxygen radical damage to tissues 1113). 2) 3-Carotene is another antioxidant in the diet that could be important in pro tecting body fat and lipid membranes against oxidation. Carotenoids are freeradical traps and remarkably efficient quenchers of singlet oxygen (1/4). Sin glet oxygen is a very reactive form of oxygen which is mutagenic and particu larly effective at causing lipid peroxida tion (114). It can be generated by pig ment-mediated transfer of the energy of light to oxygen, or by lipid peroxidation, although the latter is somewhat contro versial. 3-Carotene and similar polyprenes are present in carrots and in all food that contains chlorophyll, and they appear to be the plants' main defense against singlet oxygen generated as a by product from the interaction of light and chlorophyll (115). Carotenoids have been shown to be anticarcinogens in rats and mice I//6). Carotenoids (in green and yellow vegetables) may be anticar cinogens in humans (/, 56. i/7). Their protective effects in smokers might be related to the high level of oxidants in both cigarette smoke and tar (45. 1/8). Carotenoids have been used medically in 1259 the treatment for some genetic diseases, such as porphyrias. where a marked pho tosensitivity is presumably due to singlet oxygen formation \ll9t. 3) Selenium is another important di etary anticarcinogen. Dietary selenium (usually selenite) significantly inhibits the induction of ->kin. liver, colon, and mammary tumors in experimental ani mals by a number of different carcino gens. as well as the induction of mamma ry tumors by viruses i i20\. It also inhib its transformation of mouse mammary cells (/'/). Low selenium concentrations may be a risk factor in human cancer 1122). A particular type of heart disease in young people in the Keshan area of China has been traced to a selenium deficiency, and low selenium has been associated with cardiovascular death in Finland {123). Selenium is in the active site of glutathione peroxidase, an en zyme essential for destroying lipid hy droperoxides and endogenous hydrogen peroxide and thus helping to prevent oxygen radical-induced lipid peroxida tion 1107). although not all of the effects of selenium may be accounted for by this enzyme [120). Several heavy-metal tox ins. such as Cd:~ la known carcinogen) and Hg:~. lower glutathione peroxidase activity by interacting with selenium 1107). Selenite (and vitamin El has been shown to counter the oxidative toxicity of mercuric salts 1124). 4 Glutathione is present in food and is one of the major antioxidants and antimutaeens in the soluble fraction of cells. The glutathione transferases (some of which have peroxidase activity) are ma jor defenses against oxidative and alkyl ating carcinogens (109). The concentra tion of glutathione may be influenced by dietary sulfur amino acids (125. 126). .\Acetylcysteine. a source of cysteine, raises glutathione concentrations and re duces the oxidative cardiotoxicity of adriamvcin and the skin reaction to radi ation 1127). Glutathione concentrations are raised even more efficiently by l-2oxothiazolidine-4-carboxylate. which is an effective antagonist of acetamino phen-caused liver damage [126). Acet aminophen is thought to be toxic through radical and quinone oxidizing metabo lites (128). Dietary glutathione may be an effective anlicarcinogen against aflatoxin (129). 5) Dietary ascorbic acid is also impor tant as an antioxidant. It was shown to be anticarcinogenic in rodents treated with ultraviolet radiation. benzoMpyrene. and nitrite (forming mtroso carcin ogens) 164, 65. 130). and it may be in versely associated with human uterine cervical dysplasia (although this is not i:) proof of a cause-effect relationship) t!31). It was recently hypothesized that ascorbic acid may have been supple mented and perhaps partially replaced in humans by uric acid during primate evo lution [132). 6) Uric acid is a strong antioxidant present in high concentrations in the blood of humans il32). The concentra tion of uric acid in the blood can be increased by dietary purines: however, too much causes gout. Uric acid is also present in high concentrations in human saliva 1132) and may play a role in de fense there as w-ell. in conjunction with lactoperoxidase. A low uric acid level in blood may possibly be a risk factor in cigarette-caused lung cancer in humans )I33). 7) Edible plants and a variety of sub stances in them, such as phenols, have been reported to inhibit icabbage) or to enhance (beets) carcinogenesis (//. 55. 134) or mutagenesis [23 . 66 . 92. 135) in experimental animals. Some of these substances appear to inhibit by inducing cytochrome P-450 and other metabolic enzymes [U34)\ see also (//)], although on balance it is not completely clear whether it is generally helpful or harmful for humans to ingest these inducing sub stances. The hypothesis that as much as 80 percent of cancer could be due to envi ronmental factors was based on geo graphic differences in cancer rates and studies of migrants U36). These differ ences in cancer rates were thought to be mainly due to life-style factors, such as smoking and dietary carcinogens and promoters (136). but they also may be due in good part (see also (/)] to less than optimum amounts of anticarcinogens and protective factors in the diet. The optimum levels of dietary antioxi dants. which may vary among individ uals. remain to be determined; however, at least for selenium 1120). it is important to emphasize the possibility of deleteri ous side effects at high doses. Oxygen Radicals and Degenerative Diseases Associated with Aging Aging. A plausible theory of aging holds that the major cause is damage to DNA (102. 137) and other macromol ecules and that a major source of this damage is oxygen radicals and lipid per oxidation (43. 84. 103. 138-141). Cancer and other degenerative diseases, such as heart disease [102). are likely to be due in good part to this same fundamental de structive process. Age pigment ilipofuscin) accumulates aging in all mammalian species and has been associated with lipid peroxidation I 73. 84. 138. 1391. The fluorescent products in age pigment are thought to be formed by malondialdehyde (a mutagen and carcinogen and a major end product of rancidity) crosslinking protein and lipids < 138). Metabol ic rate is directly correlated with the rate of iipofuscin formation (and inversely correlated with longevity) 1139). Cancer increases with about the fourth power of age. both in short-lived species such as rats and mice (about 30 percent of rodents have cancer by the end of their 2- to 3-year life-span) and in longlived species such as humans (about 30 percent of people have cancer by the end of their 85-year life-span) 1142). Thus, the marked increase m life-span chat has occurred in 60 million years of primate evolution has been accompanied by a marked decrease in age-specific cancer rates: that is. in contrast to rodents. 30 percent of humans do not have cancer by the age of 3 (142). One important factor in longevity appears to be basal metabol ic rate (139. 141). which is much lower in man than in rodents and could markedly affect the level of endogenous oxygen radicals. Animals have many antioxidant de fenses against oxygen radicals. In creased levels of these antioxidants, as well as new antioxidants, may also be a factor in the evolution of man from short-lived prosimians (i43). It has been suggested that an increase in superoxide dismutase is correlated (after the basal metabolic rate is taken into account) with increased longevity during primate evolution, although this has been disput ed (141). Ames er at. proposed U32) that as uric acid was an antioxidant and was present in much higher concentrations in the blood of humans than in other mam mals. it may have been one of the inno vations enabling the marked increase in life span and consequent marked de crease in age-specific cancer rates which occurred during primate evolution. The ability to synthesize ascorbic acid may have been lost at about the same time in primate evolution as uric acid levels be gan to increase (144). Cancer and promotion. Both DNAdamaging agents (initiating mutagens) (21. 41. 42) and promoters (145) appear to play an important role in carcinogene sis (21. 146). It has been postulated that certain promoters of carcinogenesis act by generation of oxygen radicals and resultant lipid peroxidation (73. 146149). Lipid peroxidation cross-links pro teins (43. 150) and affects all aspects of cell organization (72). including mem brane and surface structure, and the mi- SC1ENCE. VOL. 221 URL 04025 totic apparatus. A common property of promoters may be their ability to pro duce oxygen radicals. Some examples are fat and hydrogen peroxide (which may be among the most important pro moters! (67. 68. 81). TCDD 1/5/). lead and cadmium (152). phorbol esters 1147. 149. 153). wounding of tissues [154). asbestos 1/5.5). peroxides it56). catechol {45) (see quinones above), mezerein and teleocidin B </-<7l. phenobarbital 1/57). and radiation (72. 158). Inflammatory reactions involve the production of oxy gen radicals by phagocytes t/05>. and this could be the basis of promotion for asbestos 1155) or wounding (154). Some of the antioxidant anticarcinogens (dis cussed above) are also antipromoters (73. 121. 146. 159. 160). and phorbol ester-induced chromosome damage (149) or promotion of transformation 1/59) is suppressed by superoxide dismutase. as would be expected if promoters were working through oxidative mecha nisms. Many "complete" carcinogens cause the production of oxygen radicals (73. 161): examples are nitroso com pounds, hydrazines, quinones. polycy clic hydrocarbons (through quinones). cadmium and lead salts, nitro com pounds. and radiation. A good part of the toxic effects of ionizing radiation damage to DNA and cells is thought to be due to generation of oxygen radicals (103. 162). although only a tiny part of the oxygen radical load in humans is likely to be from, this source. Recent studies give some clues as to how promoters might act. Promoters dis rupt the mitotic apparatus, causing hemizygosity and expression of recessive genes (163). Phorbol esters generate oxy gen radicals, which cause chromosome breaks (164) and increase gene copy number (/65). Promoters also cause for mation of the peroxide hormones of the prostaglandin and leukotriene family by oxidation of arachidonic acid and other C:o poiyenotc fatty acids, and inhibitors of this process appear to be antipro moters 1/60). These hormones are inti mately involved in cell division, differen tiation. and tumor growth (166) and could have arisen in evolution as signal molecules warning the cell of oxidative damage. Effects on the cell -membrane have also been suggested as the impor tant factor in promotion, causing inhibi tion of intercellular communication < 167) or protein kinase activation (167a). Hean disease. It has been postulated that atherosclerotic lesions, which are derived from single cells, are similar to benign tumors and are of somatic muta tional origin (102. 168). Fat appears to be one major risk factor for hean disease as 23 SEPTEMBER 1083 well as for colon and breast cancer (691. In agreement with this, a strong correla tion has been observed between the fre quency of atherosclerotic lesions and adenomatous polyps of the colon (69). Thus, the same oxidative processes in volving fat may contribute to both dis eases. Oxidized forms of cholesterol have been implicated in heart disease 1/69). and atherosclerotic-like lesions have been produced by injecting rabbits with lipid hydroperoxide or oxidized cholesterol (/69). The amicarcmogens discussed above could be anti-heart dis ease agents as well. As pointed out in the preceding section, vitamin E amelio rates both the cardiac damage and carci nogenicity of the free-radical-generating quinones adriamycin and daunomycin: .V-acetylcysteine reduces the cardiotoxicity of adriamycin: and selenium is an antinsk factor for one type of heart dis ease. Other diseases. The brain uses 20 per cent of the oxygen consumed by man and contains an appreciable amount of unsaturated fat. Lipid peroxidation (with consequent age pigment) is known to occur readily in the brain (72). and possi ble consequences could be senile demen tia or other brain abnormalities (84). Sev eral inherited progressive diseases of the central nervous system, such as Batten's disease, are associated with lipofuscin accumulation and may be due to a lipid peroxidation caused by a high concentra tion of unbound iron t/70). Mental retar dation is one consequence of an inherit ed defective DNA repair system (XP complementation group D) for depurinated sites in DNA (171). Senile cataracts have been associated with light-induced oxidative damage (172). The retina and an associated layer of cells, the pigment epithelium, are ex tremely sensitive to degeneration in vita min E and selenium deficiency (173). The pigment epithelium accumulates massive amounts of lipofuscin in aging and dietary antioxidant deficiency U73). The eye is well known to be particularly rich in antioxidants. The testes are quite prone to lipid peroxidation and to the accumulation of age pigment. A number of agents, such as gossypol. which cause genetic birth defects (dominant lethals) may be active by this mechanism. The various agents known to cause cancer by oxidative mechanisms are prospective mutagenic agents for the germ line. Thus, vitamin E. which was discovered 60 years ago as a fertility factor (72). and other antioxi dants such as selenium (174). may help both to engender and to protect the next generation. Risks There are large numbers of mutagens and carcinogens in every meal, all per fectly natural and traditional [see also (21. 221J. Nature is not benign. It should be emphasized that no human diet can be entirely free of mutagens and carcino gens and that the foods mentioned are only representative examples. To identi fy a substance, whether natural or man made. as a mutagen or a carcinogen, is just a first step. Beyond this, it is neces sary to consider the mks for alternative courses of action and to quantitate the approximate magnitude of the risk, al though the quantification of risk poses a major challenge. Carcinogens differ in their potency in rodents by more than a millionfold (175). and the levels of partic ular carcinogens to which humans are exposed can vary more than a billiontold. Extrapolation of risk from rodents to humans is difficult for many reasons, including the longevity difference, anti oxidant factors, and the probable multicausal nature of most human cancer. Tobacco smoking is. without doubt, a major and well-understood risk, causing about 30 percent of cancer deaths and 25 percent of fatal heart attacks (as well as other degenerative diseases) in the Unit ed States (/>. These percentages may increase even more in the near future as the health effects of the large increase in women smokers become apparent (/). Diet, which provides both carcinogens and anticarcinogens, is extremely likely to be another major risk factor. Exces sive alcohol consumption is another risk, although it does not seem to be of the same general importance as smoking and diet. Certain other high-dose exposures might also turn out to be important for particular groups of people--for in stance. certain drugs, where consump tion can reach hundreds of milligrams per day: particular cosmetics; and cer tain occupational exposures (2). where workers inhale dusts or solvents at high concentration. We must also be prudent about environmental pollution (41. 54). Despite all of these risks, it should be emphasized that the overall trend in life expectancy in the United States is con tinuing steadily upward 1176). The understanding of cancer and de generative disease mechanisms is being aided by the rapid progress of science and technology, and this should help to dispel confusion about how important health risks can be identified among the vast number of minor risks. We have many methods of attacking the problem of environmental carcinogens (and anticarcinogens). including human epiderm ic URL 04026 URL 04027 ology (/). short-term tests [41, 42. 177). and animal cancer tests [175). Powerful new methods are being developed [for instance, see idtf. 177)J for measuring DNA damage or other pertinent factors with great sensitivity in individuals. These methods, which are often nonin- vasive as they can be done on blood or urine (even after storage), can be com bined with epidemiology to determine whether particular factors are predictive of disease. Thus, more powerful tools will be available for optimizing antioxi dants and other dietary anti-risk factors, for identifying human genetic variants at high risk, and for identifying significant health risks. References and Notes 1 R. Doll and R. Pete. J. Sail. Cancer ln.il. 66. 119: (1981). R. Peto and M. Schneiderman. Eds.. Banbury Report 9. Quantification of Occupational Can cer (Cold Spring Harbor Laboratory. Cold Spring Harbor, N Y . 19811. 3. Cancer Facts and Futures. i'ISj (American Cancer Society. New York. 1982). 4. National Research Council. Diet. Nutrition and Cancer (Naiional Academy Press. Wash ington. D C., 1982). 5. . C. Miller. J. A. Miller. I Hirono. T Sugi mura. S. Takavama. Eds.. Naturally Occurring Carcinogens-Murayens and Modulators of Carcinogenesis (Japan Scientific Societies Press and University Park Press, Tokyo and Baltimore, |979i: E. C. Miller et at. Cancer Res. 43. 1124 0983); C. [oannides. M. Dela1'orge, D V. Parke, Food Cosmet. Toxicol. 19. 657 1)981). 6. G- J. Kapadia, Ed.. Oncology Overvieh- on Naturally Occunng Dietary Carcinogens of Plant Origin (International Cancer Research Data Bank Program. National Cancer Institute, Bethesda. Maryland. 1982], 7 A. M. Clark, in Environmental Mutagenesis. Carcinogenesis. and Plant Biology. E. J Klekowski. Jr , Ed. (Praeger. New- York. !982). vol. 1, pp. 97-132. 8, A. M. Pamukcu. S. Yalciner. J. f. Hatcher, G. T. Bryan, Cancer Res. 40. 3468 11980); J. F. Hatcher. A. M. Pamukcu. E. Erturk. G. T. Brvan. Fed. Proc Fed. Am Sac. Exp. Biol. 42. ?86(I983(. 9 H. F. Stich. M. P. Romo. C. H. Wu. W. D. Powrie. Mutat. Res. 90. 201 119811; A. A. Aver'yanov. Biokmmiya 46. 256 (|98li, A. F. Hanham. B. P. Dunn, H. F. Stich. Mutat. Res. 116. 333 (1983) 10. K H. Pegel. Chem. Eng. Sews 59. 4 (20 Julv 1981). 11. L K. T. Lam. V. L. Sparmns. L W. Wattenberg. Cancer Res 42. I 19) 11982). 12. S. J. Jadhav. R. P. Sharma. D K. Salunkhe, CRC Cm. Rev Toxicol. 9. 21 (198 0. 13. R. L. Hail. Sutr. Cancer J (So. 2l. 27 i 1979). U H. W Moore and R Czermak. Med. Res Rev 1. 249(1980. 15. I. Hirono. CRC Cnt. Rev. Toxicol. 8, 235 (1980. 16. J P. Brown. Mutat Res. 75. 24? 119801. 17. M. J. Ashuood-Smith and G. A. Poulton. ibid. 85. 389 0 98 0. IS. A. Ya. Potapenko. M V Moshmn. A. A. Krasnovskv. Jr.. V L. Sukhorukov. Z. Nuturforich. 37. 70(1982) 19. G. W. Ivie. D. L. Holt, M. C. tvev. Science 213. 909 H98H: R. C. Beier and E. H. Oertli. Phytochemistry, m press-. R. C. Beier. G. W. [vie. E- H. Oertli. m "Xenobiotics m Foods and Feeds." ACS Symp. Ser.. in press:_______ D. L. Holt. Food Chem. Toxicol. 21. 163 09831 20. G Tamura. C. Gold, A. Ferro-Luzzi. B. N. Ames. Proc. Sail. Acad. Sci. U S A. 77, 496! (19801. 21. T. Sugimura and S. Sato. Cancer Res. tSuppl.i 43. 2415s (1983): T. Sugimura and M. Nagao, m Mutagenicity: New Horizons m Genetic Toxi cology. J. A. Heddle. Ed. iAcademic Press. New York. 1982). pp 73-88. 22. W. W Kilgore. D. G. Crosby. A. L- Craigmill. N. K. Poppen. Calif. Avne. 35 <No. 11) (No vember 1980: D. G. Crosby. Chem. Eng. Sens 61. 37 (II April 1983): C. D Warren. ibid-. p 3 (13 June !983>. 23. H F. Stich. M. P. Rosin. C. H. Wu. W. D Powrie. in .Mutagenicity. New Horizons in Ge netic Toxicology. J. A. Heddle. Ed. (Academic Press, New York. 1982). pp, 117-142;_______ W. D. Powne. Cancer Lett. 14. 25! ' 1980. 24. N Katsui. F. Yagihashi. A. Murat. T. Masamune. Bull. Chem. Soc Jpn. 55. 2424 (1982): ----------- ibid., p. 2428: R. M. Bostock. R. A. Lame. J. A. Kuc. Plant Physiol. 70. 14)7 (1982). 25. H. Griesebach and J. Ebel. AngeChem. /nr. Ed. Engl. 17, 635 11978). 26. J. M. Concon. D. S. Newburg. T. W. Swerczek. Sutr. Cancer I (No. 3). 22 (1979). 27. H. W. Renner and R, Munzner. .Mutat Res. 103. 275 (1982): H. W. Renner. Experientia 38. 600 O982):D. Mourelatos.J. Dozi-Vassiliades. A. Granitsas. .Mutat. Res. 104. 243 (1982); J. H. Cans. Toxicol. Appl. Pharmacol. 63. 312 t1982) 28. B. Toth, in Naturally Occurring Carcinogens- Mutagens and .Modulators of Carcinogenesis. E. C. Miller. J. A. Miller, I. Hirono, T. Sugi mura. S. Takayama, Eds. (Japan Scientific Societies Press and University Park Press. Tokyo and Baltimore. 1979), pp. 57-65: A. E. Ross. D. L. Nagel, B. Toth. J. Ague. Food Chem. 30. 521 (1982); B. Toth and K. Patil. Mycopathologia 78. II (1982); B. Toth. D. Nagel. A. Ross. Br. J. Cancer 46. 4|7 (1982). 29. R. Schoental, Toxicol. Lett. 10. 323 11982); R. J. Huxtable. Perspect. Biol. Med 24, I 11980): H. Niwa. H. Isluwata. K. Yamada./. Chromatogr. 257. 146 09831. 30. M. Cheviot) and T. Navok. Anal. Biochem. 12*. 152119831; V. Lattanzio. V. V Bianco. D. Laltandra. Experientia 3*. 789 0982); V. L. Flohe. G. Niebch. H. Retber. Z. Klin. Chem. Klin. Biorhem. 9. 431 (I97|); j. Mager. M. Chevion. G. Glaser, in Toxic Constituents of Plant Foodstuffs. I. E. Liener. Ed. (Academic Press. New York, 19801. pp. 265-294. 31. J. K- Dunnick etai.. Fundam. Appl. Toxicol. 2. 114 0 982). 32. L. C. Berardt and L. A. Goldblatt, in Toxic Constituents of Plant Foodstuffs. I. E Liener. Ed. (Academic Press, ed. 2, New York, 19801. pp. 183-237. 33. S. P Xue. in Proceedings. Symposium on Recent Advances in Fertility Regulation (Bei jing. 2 to 5 September. 1980). p. 122. 34. R. K- Haroz and J. Thomasson. Toxicol. Lett. Suppl. 6. 72 0 980). 35. M. Coburn. P. Sinsheimer, S. Segal, M. Bur gos. W. Troll. BioI Bull. tWoodi Hole. Mass.) 159. 468 0980). 36. C. Campbell, personal communication. 37 G. D. Manners, G. W. Ivte. J. T. MacGregor, Toxicol. Appl. Pharmacol. 45 . 629 0 978); G W. Ivie and D. A. Wjtzel, in Plant Toxins, vol. I. Encyclopedic Handbook of Natural Toxins. A. T. Tu and R. F. Keeler, Eds. (Dekker. New York, in press). 38. J. C- M. Van der Hoeven et at., m Mutagens in Our Environment. M. Sorsa and H. Vainio. Eds. (Uss. New York. 1982). pp. 327-338; J. C. M. van der Hoeven, W. J. Lagerweij. I. M. Bruggeman. F. G. Voragen. J. H. Koeman. J. Agric. Food Chem.. in press. 39. T. Hirayama and Y Ito. Prev. Med. 10. 614 09811: E. Hecker. J. Cancer Res. Clin. Oncol. 99, 103 0980. 40. M. R- Malinow, E. J. Bardana. Jr., B. Pirofsky. S. Craig. P. McLaughlin, Science 216. 415 0 982). 41. B N. Ames. ibid. 204, 587 (1979). "Mutagen" will be used in its broad sense to include clastogens and other DNA-damaging agents. 42. H. F. Stich and R. H. C. San. Eds.. ShortTerm Tens for Chemical Carcinogens iSpringer-Verlag. New York, 1981). 43 P. Hochstem and S. K. Jain. Fed. Proc. Fed. Am. Soc. Exp. Biot. 40. 183 0 981). 44. D. E. Levin, M. Hollstein. M. F. Christman, E. Schwiers, B. N. Ames, Proc. Natl. Acad. Sci. C/.S.A. 79. 7445 0 982). Many additional quinones and aldehydes have now been shown to be mutagenic. 45. S. G. Carmeila. E. J. LaVoie. S. S. Hecht. Food Chem. Toxicol. 20. 587 0 982). 46. K. Morimoto, S. Wolff. A. Koizumi. Mutat. Res Leu. 119. 355 11983); T. Sawahata and R. A. Neal. Mol. Pharmacol. 23. 453 0983). 47. H. Kappus and H. Sies. Experientia 37. 1233 098!). 48. L. Tikkanen. T. Matsushima. S. Natori. Mutat. Res. 116. 297 0 983). 49. P. K. Singal. N. Kapur. K. S. Dhillon. R. E. Beamish. N. S Dhalla. Can. J. Physiol. Phar macol 60. 1390 0 982). 50. A. Kasamaki et at.. Mutat. Res. tOS. 387 0982). 51. J D. Hendricks. R.O. Sinnhuber. P. M. Love land. N. E. Pawtowski. J. E. Nixon. Science 20*. 309 11980); R. A. Phelps. F. S. Shenstone. A. R. Kemmerer. R. J. Evans. Poult. .Sci. 44. 358 i )9Mr. N. E. Pawlowski, personal commu nication. 52. I Ement. A. M MicheKon. A. Levy. J. P. Camus. 1. Ement. Hum Genet. 55. 341 09801. 53. FDA Compliance Program Report of Findings. FY?9 Total Diet Studies--Adult iNo, 7305 002); available from Narional Technical Information Service. Springfield. Va.). It is estimated that the daily dietary intake of syn thetic organic pesticides and herbicides is about 60 ng. with chlorprepham. malatbion. and DDE accounting for about three-fourths of this. An estimate of 150 u-gof datlv exposure in Finland to pesticide residues has been made by K. Hemmimki, H. Vainio. M. Sorsa. S. Saltmnen [/. Environ. Sci. Health Cl (No. !i. 0983)]. 54. N K. Hooper. B. N. Ames. M. A. Saleh. J. E. Casida. Science 295. 591 1)979). 55. 1_. W. Wattenberg. Cancer Res. iSuppl.i 43, 2448s 0983). 56. J Hocy. C Montvernay. R. Lambert. Am. J. Epidemiol. 113. 668 0981): A. J. Tuyns. G. Pequignot. M. Gignoux. A. Valla. In/. J Can cer 30. 9 (19821; A. Tuyns. in Cancer Epidemi ology and Prevention. D. Schottenfeld and J. F- Fraumeni. Jr.. Eds. (Saunders. Philadel phia. 19821, pp. 293-303: R. G. Ziegler et at.. J Sail. Cancer Inst. 67. 1199 0981); W. D. Flanders and K, J. Rothman. Am. J. Epidemi ol. 115. 371 0982/. 57. E. L. Abel. Hum. Biol. 54. 421 0982), H. L Rosser. L. Weiner. A. Lee. B. Zuckerman. F. Pooling, E. Oppenheimer, Obstet Gvnecoi. 61. 539 11983). 57a.R. A. Anderson. Jr.. B R. Willis. C. Oswald. L. J. D. Zaneveid. J Pharmacol. Exp. Ther. 225. 479 09831. 58. H F. Stich and M. P. Rosin. Ini J. Cancer 31, 305 (19831. 59. R. P. Bird. H. H. Draper. P. K. Basrur. Mutat. Res. 19). 237 0982). M. A. Campbell and A. G. Fame). Life Sci. 32. 2641 0983). 60. V. J. Feron. A. Kruysse. R. A. Woutersen. Eur. J. Cancer Clin Oncol. 18. 13 11982). 61. T. Suematsu et al.. Alcoholism. Clin Exp.P.j. 5. 427 11981): G. W. Winston and A. I. Cederbaum, Biochem. Pharmacol. 31. 2301 0 982), L- A. Videla. V. Fernandez. A. de Marinis, N. Fernandez. A. Valenzuela. Biochem. Biophys. Res. Commun. 104. 965 0982); T. E. Siege. Res. Commun. Chem. Pathol. Pharmacol. 36. 287 0 982). 62. M G. Cutler and R. Schneider. Food Cosmet Toxicol. 12. 451 0 974). 63. Y. Tazima. in Environmental Mutagenesis. Carcinogenesis and Plant Biology. E. J. Klekowski. Jr.. Ed. (Praeger. New York. 1982). vol. I. pp. 68-95. 64. P. N. Magee. F-d.. Banbury Report 12. Nitrosamines and Human Cancer (Cold Spring Har bor Laboratory, Cold Spring Harbor. N.Y.. 1982): P. E. Hartman, in Chemical Mutagens. F. J. de Setres and A. Holiaender. Eds. (Ple num, New York. 1982). vol. ~. pp 211-294; P. Hartman. Environ. Mutagen. 5. Ill 0983). 65 Committee on Nitrite and Alternative Curing Agents in Food. Assembly of Life Sciences. National Academy of Sciences. The Health Effects of Nitrate. Nitrite, and S-Nuroso Com pounds (National Academy Press, Washing ton, D C.. I98D. 66. H. F. Stich. P. K. L. Chan, M. P. Rosin. Int. J Cancer 30. ?I9 (1982); H. F Stich and M. P. Rosin, in Nutritional and Metabolic Aspens of Food Safety. M. Friedman. Ed. (Plenum. New York, in press!. 67. L. J. Kinlen. Br. Med. J. 2*6. 1081 0983): D. J. Fink and D. Kmchevsky. Cancer Res. 41.3677 09811. 68. C. W. Welscb and C. F. Aylswonh. J. Nad. Cancer Inst. 7ft. 215 (1983). 69. P. Correa. J. P. Strong. W. D Johnson. P. Pizzolato. W. Haenszei. J. Chronic Pis. 35. 313 0982). 70. F- B-Shorland ei ai..J Agric. Food Chem 29. 863 0981). 71. M- G. Simic and M. Karel. Eds.. Autoxidation in Food and Biological Svstems (Plenum. New York. 1980). 72. W. A. Pryor. Ed., Free Radicals in Biology (Academic Press. New York. 1976 to 1982'. vols. l to 5. 73. H. B. Demopoulos. D. D. Pietronigro. E, S. Flamm. M. L. Setigman. J. Environ. Pathol. Toxicol 3. 273 0980). 74. F. Bischoff. Adv. Lipid Res. 7. 165 0 969). 1262 SCIENCE. VOL. 221 URL 04028 7< N .L. Petrakis. L. D. Gruenke. J. C Craig. Cancer Res. 41. 2563 II98H. 7ft H. S. Black and D R Douglas, ibid. 32. 2630 i IV72). 77. H. Imai. N T. Werthessen. V. Subramarnam. P. W. LeQuesne. A. H. Soloway. M. Kanisawa. Science 207 65 1 (I98t)i. "8. M. Ferrali. R. Fulcen. A. Benedetti. M Com port). Res Comma*. C>um. Pitthol. Pharma col. 30. 99 119801. T9 N R. Artman. Ad\. Lipid Re*. 7. 245 (1969). 80. J. K. Reddy. J. R. Warren. M. K. Reddv. N. D. Lalwani. Ann. S.Y. Acad. Su. 386. 81 (1982); J. K. Reddv and N. D. lalwani. CRC Crit. Rec Toxicol., m press. 81. H L. Plume. Genetics 40. 268 11955): A Ito. M. Nano. Y, Naito. H. W'aianabe. Gann 73, 315 (1982): G. Spelt. W. Vogel. M Wolf. Environ. Mutagen. 4. 135 (1982). H. Tsuda. Jpn J. Genet. 56. 1 H9R]i: N. Hirota and T. Yokoyama. Gann 72. 8(1 r(98h, ` 82. S Hone. H. Ishu. T. Suga. J. Biochem. -To- iv,,90. 1691 119811; D. P. Jones. L. Eklow. H Thor. S. Orremus. Anh. Biot hem Biophw. 210. 505 119811. 83 B. Chance. H. Stes, A. Bovens. Phvstol. Ret 59. 527 i [979). 84. D Harman, Proc Sad. Acad. Sri. C.J.A. 78. 7124 119811: m Free Radicals m Biologv. W. A. Pryor. Ed. (Academic Press. New York. 1982). vol. 5. pp 255-2"5. 85. I Emerit, M. Keck. A. Levy. J. Feingold. A. M Mtchelson. Murat. Res. 10J. 165 119821 86. C. E. Neat. M. S. Thomassen. H. Osmundsen. Bioihem. J. 1%. 149 11981): J. Bremer and K. R. Norum. J. Lipid Res 2J. 243 i I9K2): VI. S Thomassen. E. N. Christiansen. K R. Norum. Biochem. J. 206. 195 (1982). H. Osmundsen. Ini. J. Run Hem. 14, 905 11982): i. Norseth and M. S. Thomassen. Btoihim Biophys. Acta, m press. 7 M. G. Enig, R. J Mann, M Keeney. Fed. Proc Fed. Am. Sac. Exp Biol. 37. 22 IS i 1978): J. E. Hunter. J. Sad. Cancer Inst. *9. X19 H982): A. B. Awad. ibid., p. 320: H. Rutten- berg. L. M. Davidson. N. A. Little, D M. Klurt'eld. D. Kritchevskv. J. Star, i 13. 8JS (1983) 88. R. Wood. Lipids 14. 97S i 1979i. 89. E. N. Chnsiiansen. T. Flatmark. H. Krwi. Ear. J. Cell Biol. 26. II 11981 > 90 Council lor Agricultural Science and Technolo gy. Diet Sutrition. and Cancer. A Critnjiie (Special Publication 13. Council tor Agricultur al Science and Technology. Ames. Iowa. 1982i. 91. L. F. Bjeldanes et ai.. Food Chem Taxi< ol. 20. 357 (1982): M. W Pariza. L. J. Loretz. J. M. Slorkson. N . C. Holland. Cancer Res. ti'iippi. i 43 . 2444s (1983). 92. H F. Such. W. Such. M. P Rosin. W. D. Powrie. Mutat. Res. 91. 129 il981): M. P. Rosin. H. F. Such. W. D. Powne. C. H. Wu. ibid. 101. 189 (19821; C -l. Wei. K Kitamura. T Shibamoto. Fond Cosmet. Toxicol. 19. 749 11981). 93 E. Yamasaki and B. N. Ames. Proc. Sctd. Acad. .Vtv. U.S A. 74. 3555 (1977). 94. R. Baker. A. Arlauskas. A. Bonin. D. Angus. Cancer Lett. 16. 81 I 1982). 95 D. SchuerzJe. D Cronn, A. L. Crittenden. R. J. Charlson. Environ. Set. Technol. 9. 838 (1975): G. Gartrell and S K. Fnedlaniler. At mos Environ. 9. 279 (1975). L D Kier. E. Yamasaki. B. N Ames. Proc. Sail. Acad. Su. L'.S A. 71. 4159 M974); J. N. Pius. Jr.. Ennron Health Perfect. 47. 115 (J9H.H 96. J. E. Vena. Am. J. Epidemiol. 116. 42 11982). R. Cederlof. R. Doll. B Fowler. Environ. Health Perspeit 22. I M978i; F. E. Speizer. ibid. 47. 33 i 1983'. 97 B. Brunekreef and J. S M. Bolei.i. Int. Anh Occttp Environ. Health 50. 299 (1982). 98 H. Kasai et at.. Gann 73. 681 I |9K2i. 99. v. Armuth and I. Berenblum. Carcinogenesis 2. 977(!98h. 100. S. Fabro. Reprud. Toxicol. I. 2 11982). 101. D. Tnchopoulos. M. Papaposiolou. A. Poly- chronopoulou. Int. J Cancer 28. 691 11981); P. Hartge. L. P. Lesher. L McGowan. R. Hoo ver. ihid 30. 531 (1982); B. MacMahun, Cancer <Briisselsi SO. 2676 11982). H. 5. Cuckle and L. J. Kinlen, fir. J. Cancer 44. 76011981 >. R. L. Phillips and D. A. Snowdon. Cancer Res. < Suppl. >43. 2403s (19831: L. D. Marrett. S. D. Walter. J. W Meigs. 4m. J. Epidemiol. 117. II3H983): D. M Weinberg. R. K. Ross. T. M. Mack. A. Pagamni-Hill. B. E. Henderson. Cancer 'Brusselsi 51. 675 H983). 102. P. E Hartman. Em-i'mn. Mutagen.. in press. 103 J. R Totter. Proc Sail Acad. Su. U.S.A. 77, |763 11980). 104. [. Fridovich. m Pathology ot Oxygen. A. Au- lor. Ed. (Academic Press. New York. 1982). pp. 1-19; L. W. Oberley. T. D. Oberley. C. R. Buettner. Med. Hipotheies 6. 249 (1980i ins B. Halliwell. Cell Biol. Int. Rep. 6. 529(1982). A. I. Tuuber. Trends Biochem. Sci. 7. 411 119821: A. B. W'eitberg. S. A. Weitzman. Vi. Destrempes. S. A. Latt, T P. Sto'sel. V. Engl. J Med 308. 26 i]983). Neutrophils also pro duce HOCi. which is both a chlorinating and oxidizing agent 106. M. A. Trush. E. G. Mimnaugh. T. E. Gram. Biochem. Pharmacol. 31. 3335 11982). 107. L. Flohe. in Free Radicals in Biology. W. A. Pryor. Ed. (Academic Press. New York. 1982i. voJ. 5. pp. 223-254 108. C. Lind. P Hochstem. L. Ernster. Arch. Bio chem. Biophys 216. I ?8 1)982). 109 M. Warholm. C. Guthenberg. B. Mannervik. C. von Bahr, Bioihem. Biophw. Res. Com- mun. 98, 512 (198D. MO. J. G. Biert. L. Corash. V. S. Hubbard. V Eng! J. Med. 308. 1063 M983). 111. Y. M. Wang et iif. in .Molec idar Interrelations of Sutrition and Canter. M. S. Arnott. J. van Evs. Y.-M. Wang. Eds. (Raven. New York. 19821. pp 369-3',9. | 12. C. Beckman. R. M Rov. A. Sproule. Murat Res. 105. 73 (1982). M G Cook and P. McNa mara. Cancer Res 40.1329 (1980). 113. K. /. A. Davies. A. T. Quintandha. G. A. Brooks. L. Packer. Biochem. Biophw. Res. Commit*. 107. 1198 11982). 114. C. S. Foote, in Pathology ofOxvgen. A. Autor. Ed. i.Academic Press. New York. 1982). pp. 21-44. J. E. Packer. J 5. Mahood. V O. Mora- Areilano. T. F. Slater. R. L. Willson. B S. Wollenden. Biochem. Bioplos. Res. Commun. 98. 901 11981). W. Bors. C. Michel. M. Saran. Bull Ear. Phv.uoparhtd. Resp. 17 (Suppl ). 13 (1981). 115. N. I. Krmsky and S. M. Deneke. 1. Sail- Cancer Inst. <9. 205 i 19K2). J. A. Turner and J. N. Prebble. J. Gen. .Microbiol. 119. 133 11980): K. L. Simpson and C. O. Chichester. Annu. Rei. \utr. 1. 351 (198M. 116. G. Rettura. C. Dattagupta. P. Listowsky.S. M. Levenson, E. Seifter. Fed. Proc. Fed. Am. Sot. Exp. Biol. 42. 78611983); M. M. MathewsRoth. Oncology 39. 33 (1982). IP. R. Pelo. R. Doll. J. D. Buckley. M. B. Spom. Suture <London/ 29(1. 20! (1981): R. B. She kels et at.. Lancet 1981-11, 1185 119811; T. Hirayama. Xutr. Cancer 1. 67 d979i. G. Kvale. E. Bjelke. J. J. Gart. Int. ] Cancer 31. 397 11983). 118. W. A. Pryor. VI. Tamura. M. M. Dooley. P. I. Premovie. D. F Church, in Osy-Radiuds and Their Scavenger Systems Cellular and Medi co/ Aspects. G. Cohen and R. Greenwald. Eds. (Elsevier. Amsterdam. 1983). vol. 2. pp. 185192. W. A. Prvor. B. I Hales. P. 1. Premovic. D. F. Church. Science 220. 425 (I983i. 119. M. M. Mathews-Roth.y. Sutl. Cancer Inst. 69. 279 H982). 120. A. C. Griffin, in Molecular Interrelations ol Sutrition and Cancer. M. S. Arnott. J Vaneys. Y. M. Wang. Eds. (Raven. New York. 1982). pp. 401--408. D. Medina. H. W Lane. C. M. Tracey. Cancer Res iSuppl ) 43. 2460s (1983): M. M. Jacobs, Cancer Res. 43. 1646 (1983): H. J. Thompson. L. D. Meeker, P. J. Becci.S. Kokoska. ibid. 42. 4954 11982); D. F. Bin. T A. Lawson. A. D Julius. C. E. Runice. S. Salmasi. ibid., p. 4455: C. Wilting. L'. Wit ting, V. Kneg, J. CanierRes. Clin. One id. 104. 109(1982). 121. M. Chattetjee and M. R. Banerjee. Cancer Lett 17. 187 11982). 122. W C. Willett et ai.. Lancet, in press. 123 J. T. Salonen. G. Alfthan. J. Ptkkaramen. J. K. Huttunen. P. Puska. ibid. 1982-11. 175 11982). 124. M. Yonaha, E Itoh, Y. Ohhayashi. M. Lchiyama. Res. Commun. Chem. Pathol. Pharmacol. 28. 105 (19801; L. J Kling and J. H. Soares. Jr.. Star. Rep Int. 24, 39 < 19811. 125 N. Tateishi. T. Higashi. A. Naruse. K. Hikita. Y. Sakamoto. J. Bioihem. iToksoi 90. 1603 (1981). 126. J. M. Williamson, B. Boettcher. A. Meister. Proc Sail A cud. Sa. L'.S.A. 79. 6246 (1982). 127. CME Symposium on " V-Acetylcysteme iN'ACl; A Significant Chemoprotective Ad junct," Sem. Oncol. 10 (Suppl. 1). 1 (1983). 128. J. A. Hinson. L. R. Pohl. T. J. Monks. J. R. Gillette. Life Su. 29. J07 11981). 129. A. M. Novi. Science 212. 541 (1981). 130. W. B. Dunham et at.. Proc. Salt. Acad. Sci. U.S.A. 79. 7532 M982): G. Kallistratos and E. Fasske, J. Cancer Res. Clin. Or.ioL 97, 91 (1980). 131 S. Wassertheil-Smoller et a!.. Am. J. Epidemi ol. 114. 7l4(198h. 132. B N. Ames, R. Cathcart. E. Schwiers. P. Hochstem. Proc. Sail. A uni Sci. {. ' 5 A 78. 6858 11981). 133. A. Nomura. L. K Heilbrun, G. N. Stemmer- mann. in preparation. 114. J. N. Bo\d. J. G. Babish. G. S. Stoewsand. Food Chem. Toxicol. 20. 47 11982). 135 A. W Wood, et ai.. Prm. Sad. Acad. Sci. C S A 79. 5513 119821. 136. T. H. Maugh II. Science 20S. 1363 H9"9) (interview with John Higginson). 137. H. L. Gensler and H. Bernstein. Q. Rev Biot 6. 279 11981). 138. A. L. Tappel. in Free Radicals in Biology. W A Prvor. Ed. 1 Academic Press. New- York 1980). vol. 4. pp. 1-17. 139. R. S. Sohai. mAce Pigments. R. S. Sohat. Ed (ElsevierNonh-Holland. Amsterdam. 1481) pp. 303-316. 140^-J. E. Fleming. J. Miquel. S. F Cottreil. L. S Yengovan. A. C. Economos. Gerontology 28 44 11982). 141. J. M. Tolmasoff. T. Ono. R. G. Cutler. P'uc Sail. Acad. Su. U.S.A. 77, 2777 11980): R. G Cutler. Gerontologs. in press; J. L. Sullivan ibtd. 28. 242 (19821. 142 R. Peto. Proc R. Sin . London Ser. B 205. 11 (19791 D. Dix. P. Cohen. J. Flannery. J Theor. Bud. 83. 163 119801. 143. R. G. Cutler, in Testing the Theories ol Aging R. Adeltnan and G. Roih. Eds. (CRC Press Boca Raton. Fla., in press) 144 D. Hersh. R. G. Cutler. B. N Ames, in prepa ration. 145. E. Boyland. in Health Risk Analysis iFranklin Institute Press. Philadelphia. 1980). pp. 181- 193, E. Boyland. in Canter Campaign, vol 6. Cancer Epidemiology. E. Grundmann. Ed. 1 Fischer. Sluttaart. 1982). pp. 125-128. 146. J. L. Mar\. Science 219. 158 1 1983i. 147 B. D Goldstein, G. Witz. M. Amoruso. D. S. Slone. W. Troll. Cancer Lett. U. 257 i|981i; W. Troll, in Environmental Mutagens and Carlinogens. T. Sugimura. S. Kundo. H. Takebe, Eds. (L'niv. of Tokvo Press. Tcikvo. and Liss. New York. 1982). pp. 217-222. 148. B. N. Ames. M. C. Hollstein. R. Cathcart. in Lipid Peroxide in Biology and Medicine. K. Yagi. Ed. I.Academic Press. New York. 1982). pp 339-351. 149 1. Emerit and P. A. Cerutti. Pmc. Sail. Acad Sci. U.S.A. 79. 7509 (1982): Saltire Hondom 293. 144 (1981): P. A. Ceruiti. I. Emerit. P. Amstad. in Genes and Proteins in Om agene sis. !. B. Weinstein and H. Vogel. Eds. 1 Aca demic Press. New York, in press): 1. Emerit. A. Lew. P. Cerutti. Aftortr. Res. U0. 327 11983). 150 J. Funes and M. Karel. Lipids 16. 347 119811. 151 S. J. Stohs. M. Q. Hassan. W. J. Murray. Biochem. Biophvs. Res. Commun. 111. 854 (1983i. 152. C. C. Reddy. R. W. Scholz. E. J. Massaro. Toxicol. Appl. Pharmacol. 61. 460(19811. 153. H. Nagasawa and J. B. Little. Carcinogenesis 2. 60l'il98l); V. Solanki. R. S. Rana. T. J. Slaga. ibid . p. 1141; T. W. Kensler and M. A Trush. Cancer Res. 4t. 216 1 198lt. 154. R. H. Simon. C. H. Scoggin. D- Patterson. J Sio/. Chem. 256. 7|fi| (19811: T. S. Argvns and T. J. Slaga. Cancer Res 41. 5193 H98D. 155. G. E. Hatch. D. E. Gardner. D. B. Menzel. Environ. Res. 23. 121 (19801. 156. A.J P Klem-Szanto and T. J. Slaga. J Invest. Dermatol. 79. 30 11982). 157. C. C. Weddle. K. R. Hombrook. P B. VfcCav. J. Biol. Chem. 251. 4973 11976). 158. A. G. Lurie and L. S. Cutler. J. Still. Cancer Inst. 63. 147 ( 19791. 159. C. Borek and W. Troll. Proc. Sail. Acad. Sci. 0 S .A 80. 1304 11983): C. Borek. in Molecular Interrelations of Sutrition and Cancer. M. S. Arnott. J. van Evs. Y.-M. Wang. Eds. iRaven. New York. 1982). pp. 337-350. 160. T. S. Slaga c>r a!., in Carcinogenesis- A Com prehensive Treatise iRaven. New York, 1982). vol. 7. pp. 19-34; K. Ohuchi and L. Levine. Biochtm. Biophys. Acta 619. II (1980): S. M. Fischer, G. D. Mills. T J. Slaga. Carcinogene sis 3. 1243 (1982). 161. R. P. Mason, in Free Radicals m Biology. W A. Prvor. Ed. (Academic Press. New- York. 1982). vol. 5. pp. 161-222. 162. G. McLennan, L. W. Oberiev, A. P. Autor. Radiat. Res. 84. 122 (1980). 163. J. M. Parry. E. M. Parry. J. C. Barrett. Sature lLondon) 294. 263 (1981): A. R. Ktnsella. Car1 inogenesis 3. 499 (19821. 164. H C. Birnboim. Can. J Phvsiol. Pharmacol. 60. 1359 (1982). 165. A. Varshavsky. Cell 25, 561 11981). 166. T. J. Powles et at.. Eds. Prosraglandtns and Cancer- First International Conference 'Liss. New York. 1982). 23 SEPTEMBER 1981 1263 167. I E. Trosko. C.-C. Chang. A. Medcalf, Cancer Invest., in press. 167a.I. B. Weinstein. Nature iLondoni 302, 750 119831. 168. ! A. Bond. A. M. Gown. H. L. Yang. E. P. Benditt. M. R. Juchau, J. Toxicol. Environ. Health 7. 327 119811. 169. Editorial. Lancet 1980-1, 964. (1980); K. Yagi. H. Ohkawa. N Ohishi. M. Yamashita. T. Sakashima. J. Appl. Biochern. 3. 58 (1981). (70. J. M. C. Gutteridge. B, Halhwell. D. A. Rowley. T. Weslermarck. Lancet 1982-11. 459 (1982). [Tf. I. E. Cleaver, in Metabolic Basis of Inherited Disease. J. B Stanbury. J. B. Wyngaarden. D. S. Fredrickson. J. L. Goldstein. Eds. (McGraw-Hill. ed. 5. New York. (983). pp. 1227-1250. ]72. K. C. Bhuyan. D. K. Bhuvan, S. M. Podos, [RCS Med. Sci. 9. 126 (1980; A. Spector. R. Scotto. H. Weissbach. N. Brot. Biochern. Biophys. Res. Commun. 108, 429(19821; S. D. Varma. N A. Beachy. R. D. Richards. Photo- chem. PHoiobiol. 36. 623 H9B2). 173. M. L. Katz. K. R Parker. G. J Handeiman. T. L. Bramel. E. A. Dratz. Exp Eve Res. 34. 339 (1982). 174. D. Behne. T. Hofer. R. von Berswordt-Wallrabe, W. Eiger. J. Nutr 112. 1682 (1982). 175. B. N. Ames. L. S. Gold, C. B. Sawyer. W. Havender. in Environmental Mutagens and Carcinogens. T. Sugimura. S. Kondo. H. Takebe, Eds. (L'niv. of Tokyo Press. Tokyo, and Liss. New York. 1982). pp. 663-670. 176. Nanonal Center for Health Statistics. Advance Report. Final Mortality Statistics. 1979. Monthly Vital Statistics Report 31. No. 6. suppl. [DHHS publication iPHS) 82-1120. (Public Health Service. Hyattsvillc. Md.. 1982]; Metropolitan Life Insurance Company Actuarial Tables. April 1983 177. B. A. Bridges. B. E. Butterworth, I. B. Wein stein. Eds.. Banbury Report 13. Indicators of Genotoxic Exposure (Cold Spring Harbor Lab oratory, Cold Spring Harbor. NY.. 1982); R. Montesano. M. F. Rajewskv. A. E. Pege. E. Miller. Cancer Res. 42 . 5236 (19821: H. F Stich, R. H. C. San. M. P. Rosin. Ann. .V. Y Acad. Sci.. in press: 1. B. Weinstein. Annu. Rev. Public Health 4. 409 ((983). 178. 1 am indebted to G- Ferro-Luzzi Ames. A. Blum. L- Gold. P. Hartman. W. Havender. N. K. Hooper. G. W. (vie, J. McCann. J. Mead. R. Olson. R. Peto. A. Tappel. and numerous other colleagues for their criticisms. This work was supported by DOE contract DE-AT0376EV70156 to B.N.A. and by National Insti tute of Environmental Health Sciences Center Grant ES01896. This article has been expanded from a talk presented at the 12th European Environmental Mutagen Society Conference. Espoo. Finland. June 1982 (in Mutagens in Our Environment. M. Sorsa and H Vainio. Eds. <L;ss. New York. 1982)]. I wih to dedicate this article to the memory of Philip Handier, pio neer in the field of oxygen radicals. URL 04029 I .CANCER EPIDEMIOLOGY and PREVENTION DAVID SCHOTTENFELD, M.D. Chief of Epidemiology and Preventive Medicine Memorial Sloan-Kettering Cancer Center New York. New York JOSEPH F. FRAUMENI, JR., M.D. Chief, Environmental Epidemiology Branch National Cancer Institute National Institutes of Health Bethesda, Maryland U/VJV. of CONN. M 2 4 1982 health CENTER LMRARY 1982 W. B. SAUNDERS COMPANY Philadelphia London Toronto Mexico City Rio de Janeiro Sydney Tokyo URL 04030 CHAPTER EIGHTEEN Occupation Pierre Decoufle URL 04031 INTRODUCTION Epidemiologic studies of occupational groups provide a unique opportunity to observe the experience of people in continual and intimate contact with a host of potentially hazardous substances. Occupational exposures to physical and chemical agents are usually heavy and extend over a relatively long period of time. These circumstances suggest that if a particular substance has the potential for inducing a car cinogenic effect in humans, the best opportuni ty of observing it is in an occupational setting. Occupational exposures are often associated with well-defined and circumscribed popula tions that make natural study groups. For exam ple, one might define a study group from the work force of a particular industry, plant, or labor union if there is information on duration of employment in specific jobs and, possibly, a qualitative or quantitative assessment of expo sures to particular agents. In a cohort approach, these individuals are followed forward to deter mine the status of their health at subsequent intervals and its relationship to prior expo sures. There are at least four reasons for conducting epidemiologic studies of industrial populations. One is the monitoring or surveillance of various groups to detect previously unsuspected haz ards by noting unusual disease patterns. This has been done in the steel and rubber industries in the United States (Uoyd and Ciocco, 1969; McMichael et al, 1974; Monson and Nakano, 1976; Andjelkovich et al, 1976). Epidemiologic studies also assist in the re-evaluation of current thinking on what are "safe" levels for known . hazards. As was the case with vinyl chloride \/ {Creech and Johnson, 1974), our thinking may change because new and more serious health effects are found from observations in industry. Epidemiologic studies of workers can provide leads for experimental research. Thus, identifi cation of high-risk occupational groups suggests a set of potentially carcinogenic substances that can be tested individually and jointly in labora tory animals. Finally, there will always be work environments in the real world of industrial processesthatwill bedifficult to re-create under experimental conditions. Human exposure to complex mixtures of different materials pro duced during industrial processes may defy replication in the laboratory. HISTORY The first realization that contact with materi als in the work environment could cause malig nant disease was Percivail Pott's observation of an unusually high frequency of scrotal cancer among London chimney sweeps in 1775. Al though heavy exposure to soot due to the lack of protective clothing and the poor personal hygiene of the young workers were considered causal factors by contemporary observers, ex- 318 OCCUPATION -- 319 perimental verification of the carcinogenicity of epidemiologic study of the chromate-producing soot derived from bituminous coal did not come industry was not instigated until the 1940s, until this century (Passey, 1922). A hundred when the management of one of the large years after Pott's report, it was recognized that United States firms became concerned about an skin cancer (including scrotal) could be induced apparently high incidence of lung cancer by a variety of coal tar products in different in among its employees (Machle and Gregorius, dustries (Volkmann, 1875; Manouvriez, 1876; 1948). Ball, 1885; Butlin, 1892; Lueke, 1907). An analysis of occupations reported on death Early in the 19th century, Paris described certificates between 1921 and 1932 in England cancers of the scrotum among Cornish smelter and Wales identified several groups of coal workers, presumably due to arsenic liberated in carbonization and by-products workers to be at the smelting process (Paris, 1822). However, high risk of lung cancer (Kennaway and Ken- there are no subsequent reports confirming his naway, 1936). These findings were later con observations, and later investigators have firmed by conventional epidemiologic studies. doubted the accuracy of this report (Butlin, Tumors of the urinary bladder were recog 1892; Henry, 1946). It was not until this century nized as having an industrial origin when Rehn that inorganic arsenic, in the form of a pesticide, studied workers in the German dyestuffs in was definitely established as a cause of occupa dustry in 1895. The process involved produc tional skin cancer (Hill and Faning, 1948). tion of synthetic dyes utilizing various chemi Certain forms of mineral oil constitute the cals produced from coal tar. For many years, third major occupational exposure capable of these malignancies were mistakenly referred to causing skin cancer. The earliest report was Dr. as "aniline cancers" after the principal dye Joseph Bell's description of scrotal cancel; cases intermediate that was used. Only after repeated among "paraffin pressmen" in the Scottish epidemiologic and experimental studies were no, shale oil industry (Bell, 1876). Early in the 20th two specific carcinogenic aromatic amines, gic ent century, the risk of scrotal cancer from mineral benzidine and /3-naphthylamine, discovered. oil was found among mule spinners in the C 33 wn cotton industry (Southam and Wilson, 1922). ide >ay Ith ry. de ifists Subsequently, cutting oils used in the machine tool industry were associated with an unusual frequency of skin cancer (Cruickshank and Squire, 1950). In 1879, Harting and Hesse identified pulmo nary cancer as an occupational disease among metal miners in central Europe -- apparently CURRENT STATUS OF EPIDEMIOLOGIC FINDINGS To date, most of the known occupational causes of cancer have been identified through epidemiologic observations with subsequent confirmation by laboratory studies, although g o i(sjjj tat rajrk the first instance of an occupationally induced tumor of this organ system. In the early part of this century, radioactivity was discovered in arsenic and benzene still lack experimental evidence. In the cases of 4-aminobiphenyl, mustard gas, and vinyl chloride, the risks were v ial ler to o*fy these mines and was eventually conceded to be the causal agent (Hueper, 1966b). Between 1930 and 1950, there was increased awareness of occupational environments associated with high risks of respiratory cancer. During this detected in man after the substances had been shown to induce tumors in laboratory animals, although little attention was given to the experi mental studies when first reported. Table 1 outlines well-established relation period, nickel refining (Annual Report of the ships between specific sites of cancer and vari Chief Inspector of Factories and Workshops, ous chemical and physical agents or industrial 1933), coal carbonization processes (Kuroda processes based on direct observations of vari and Kawahata, 1936), arsenic (Hill and Faning, ous occupational groups. For some industrial n- 1948), chromates (Machle and Gregorius, processes, discrete carcinogenic agents have 1948), and asbestos (Merewether, 1949) were not been identified yet, such as iron ore mining gof r .1- k al implicated as respiratory cancer hazards. In the case of asbestos, suspicions were first aroused when an unusually high frequency of lung cancer appeared at autopsy among asbes tos workers dying of asbestosis, a form of pul monary fibrosis induced by the fibrous dust and the manufacture of the dyes auramine and magenta. Coal carbonization processes (e.g., coke ovens) liberate a multitude of chemicals (including polycyclic aromatic hydrocarbons, some of which are carcinogenic in experimen tal animals), but the specific compound or d (Merewether, 1949). Although there were early combination of compounds responsible for the case reports in the German literature, the first excess of lung cancer among exposed workers 33JP 3Q J9 IU 3 Q fiv a H URL 04033 TABLE 1. Cancer Sites for Which Relationships with Occupational Exposures are Well-Established in Human Studies 64 adder Site Agant or Industrial Process Benzidine. 0-naphthyiamine. 4-Amihobiphenyl (xenylamine) Blood (leukemia) Bone Larynx Liver (angloMrcoma) Manufacture of certain dyes (e g., suramin# and magenta) Gsa retorts Rubber and cable making Industries Benzene X-radiation Radium. Meeothorium Elhanol (alhyl alcohol) manufacture by strong add process (dielhyl sulfate?) Isopropyl alcohol manufacture by strong acid process (dflsopropyl sulfate?) Mustard gaa Arsenic (inorganic compounds) Vlnyt chloride Lung. Bronchus Arsenic (Inorganic compounds) Asbestos 8ls(chloromethyf) ether Chromium compounds References Rehn. 1895; Ferguson et al. 1934; Scott. 1952; Case et at. 1954. Meilck et al. 1955; Vigliani and Baraotti. 1961; Lieben. 1963; Goldwater et al. 1965; Mancuso and El-Attar, 1967; Meilck et al. 1971; Zavon et al. 1973; Tsuchiya at al. 1975 Case and Pearson. 1954 Doll et al, 1965; Doll et al. 1972 Casa and Hosker. 1954; Davies. 1965 Girard et al. 1971; Aksoy et al. 1974; Vigilant. 1976; infante el al. 1977 Warren. 1956; Seltser and Sarhven, 1965; Warren and Lombard. 1966; Maienoski ei a>. 1975 Martland. 1931: Polednak et al. 1978 Lynch al al. 1979 Wail et al. 1952; Eckardt. 1974 Wada at al, 1966 Roth. 1958 Creech and Johnson, 1974, Waxweiter et al. 1976; Spirtas and Kaminski. 1978 Hill and Facing. 1948; Roth. 1958; Gaiy et al, 1963; Lee and Fraumeni, 1969: Oil el al. 1974; Milham and Strong. 1974; Kuratsune et al. 1974; Tokudome and Kuratsune. 1976; Rancher at al. 1977; Pinto et al. 1978: Axelson et el, 1978; Mabuchi et al. 1980 Merewether. 1949; Doll. 1955; Mancuso and Coulter, 1963; Selikoff el al. 1964; Jacob and Anspach, 1965; Lieben. 1966; Enterline and Kendrick. 1967; Selikoff et al. 1968; Knox el al. 1968: Newhouse. 1969: Tabershaw at al, 1970, Elmes and Simpson. 1971; Fletcher. 1972; Selikoff et al. 1972; Newhouse et al. 1972; Enterline et al, 1973, Selikoff et al. 1973: Edge. 1976; Martischnig et al. 1977; Peto et al, 1977; Robinson at al. 1979; McDonald et al. 1980; Selikolf et at. 1980 Figueroa et al. 1973; Lemen at al. 1976; DeFonso and Kelton, 1976; Pasternack at al. 1977 Machle and Gregorius, 1948; Baetier, 1950a; Baeljer. 1950b; Mancuso and Hueper, 1951: Brinton et al, 1952; Bldstrup and Case. 1956; Enterline. 1974; Langard and Norseth, 1975; Royle. 1975. MichelBriand and Simonln. 1977; Davies. 1976; Hayes et at, 1979; Datager at al, I960 THE CAUSES OF CANCER W O Coal carbonization processes (coke ovens on* Kurort" "nrt Kawo*'"*" in'** k rv-r _ ......... .. uuyaiu aiiu rrursetn. itrrs; Hoyle, 1975, Michel- Brland and Slmonln. 1977; Davies. 1976; Hayes at al. 1979; Datager el al, 1980 04034 Coal carbonization processes (coke ovens, gas retorts, producer gas manufacture) Coal far pitch volatiles (roofing materials. aluminum reduction plants) Iron ore (hematite) mining Mustard gas Nickel refining Radiation (radioactive ores) Nasal Cavity, Sinuses Peritoneum (mesothelioma) Pharynx Pleura (mesothelioma) Isopropanol (isopropyl alcohol) manufacture by strong acid process (diisopropyl sulfate?) Mustard gas Nickel refining Radium. Mesolhorium Shoe manufacturing (leather dust?) Woodworking (wood dual?) Asbestos Mustard gas Asbestos Skin (including seroturn) (Epilhei)oma*) Arsenic (inorganic compounds) Coal ter products (mainly coal tar, creosote, pitch, soot) Coal hydrogenation Mineral oils (from coal, petroleum, shale) X-radiation Kuroda and Kawahata, 1936; Kennaway and Kennaway. 1936; Kennaway and Kennaway, 1947; OoM. 1952; Christian. 1962; Doll at al, 1965. Kawai at al, 1967; Lloyd. 1971, Redmond et al. 1972. Doll et al, 1972; Mazumdar et al. 1975 Hammond et al. 1976, Gibbs and Horowitz. 1979 Faulds and Stewart, 1956; Boyd el al, 1970 Wada el al. 1968 Doll. 1958; Morgan. 1958; Masiromatteo. 1967; Dollet al. 1970: Pedersen et al. 1973; Doll et al. 1977 Hading and Hesse. 1879; Peller, 1939; Lorenz, 1944; DeViltiars and Windish. 1964; Wagoner et al. 1965; Lundin el al. 1969; Lundin et al, 1971 Weil et al. 1952; Eckardt 1974 Wada et si. 1968 Annual Report. 1933; Doll. 1956: Morgan, 1958; Mastramatteo. 1967; Doll et al. 1970; Pedersen el al, 1973; Doll et al, 1977 Hasterlik al al. 1964 Acheson at al, 1972; Acheson, 1976 Acheson at al. 1972; Acheson. 1976; Brlnton et al, 1977 Selikoff at al. 1965; Newhouse and Thompson. 1965; Selikoff at al. 1972; Selikoff et al. 1973; Greenberg and Lloyd Davies. 1974 Wada at al. 1968 Wagner et al, 1960; Sellkolf et al. 1965; Newhouse and Thompson, 1965; Seiikolf et al. 1972; Selikoff et al, 1973; Greenberg and Lloyd Davies. 1974 Paris. 1822; Kennaway. 1925, Henry. 1946; Hill and Fanlng, 1948; Roth, 1958: Mabuchi et al. I960 Pott. 1775. Votkmann. 1875; Manouvriez. 1876: Ball. 1885. Butlin. 1892; Lueke. 1907; Kennaway. 1925; Henry, 1946; Henry. 1947; Kipling and Waldron. 1976; Doll et al. 1972 Sexton. 1960 Volkmann. 1875; Bell. 1876. Scott. 1922; Southern and Wilson, 1922; Kennaway. 1925; Henry. 1946: Henry. 1947; Cruickshank and Squire, 1950; Cruickshank and Gourevitch. 1952: Hendricks et at. 1959; Kipling, 1968: Kipling and Waldron. 1976. Jarvhotm et al. 1981 Court Brown and Doll, 1956; Hunter. 1969; Matanosk! et al. 1975 OCCUPATION -- 321 THtm-ww * 322 --THE CAUSES OF CANCER U R l 04035 is not known. Similarly, the precise constituents recovering paraffin from petroleum (Hendricks of wood and leather dusts that produce tumors et al, 1959) were associated with the occur of the nasal cavity and sinuses are unknown, rence of skin cancer among workers exposed to although in the case of woodworking the use of certain fractions of unrefined oils. hardwoods in the furniture industry seems to be Skin cancer has also been seen among ma associated with the greatest risk (Acheson, chinists who came into intimate and prolonged 1976). contact with oils used in cutting and grinding The original report of upper respiratory sys operations (Cruickshank and Squire, 1950; tem cancers among workers engaged in the Cruickshank and Gourevitch, 1952; Kipling, manufacture of isopropyl alcohol implicated 1968). Experimental work has shown that the isopropyl oil, a volatile by-product of the proc carcinogenic component of mineral oils can be ess, as the likely carcinogen (Weil et al, 1952). greatly reduced by the method and degree of However, a recent report of laryngeal tumors refining they undergo (Bingham et al, 1965). In among workers engaged in a similar process general, the more refined a mineral oil product suggests that alkyl sulfates may be the common is, the less is its carcinogenic potential. Thus, denominator (Lynch et al, 1979). Great Britain adopted a highly refined "white" There is disagreement on exactly which com lubricating oil as the lubricant of choice in the pounds of chromium are carcinogenic for man. cotton industry to control the "epidemic" of Studies in the chromate-producing industry skin cancer among mule spinners (Kipling, (i.e., alkaline roasting of chromite ore) and in 1968). i the manufacture of chromate pigments (lead Epidemiologic studies of industrial popula and zinc chromate) provide convincing evi tions have implicated additional substances as dence that at least certain hexavalent chromium "suspect" carcinogens. A sample of these is compounds are human carcinogens (Enterline, listed in Table 2. The quality and quantity of 1974; Langard and Norseth, 1975; Davies, epidemiologic evidence vary from fairly 1978). Experimental work indicates that some strong in the cases of cadmium with prostatic chromium compounds show very little carcin cancer and coke oven emissions with kidney i J ogenic activity while others (e.g., calcium chro cancer to suggestive for lead, ethylene oxide, i mate) are very active (Sunderman, 1976). Re and acrylonitrile. In fact, some of these agents cently, several reports suggest that exposure to might appear on other lists of "recognized" car chromic acid mist in chromium plating opera cinogens. tions may be associated with an unusual fre In addition to expanding the previous list of quency of respiratory cancer (Royle, 1975; agents, Table 2 indicates other cancer sites that i Michel-Briand and Simonin, 1977). may be affected by the materials shown in Table The situation with nickel and its compounds 1. Thus, asbestos is linked with cancers of the is much the same as with chromium. Epidemio esophagus, larynx, stomach, and^arge bowel in logic studies have established beyond a doubt addition to the three sites previously listed, a that the nickel refining process is associated remarkable finding. However, it should be with very high risks of respiratory cancer. How noted that all seven sites have not necessarily ever, workers in this industry who developed been seen in every asbestos-exposed popula cancers of the nasal passages or lung were tion studied. Coke oven emissions are associat exposed to mixtures of several different nickel ed with increased mortality from kidney cancer compounds as well as other carcinogenic (hypernephroma) and possibly prostatic cancer. agents (e.g., arsenic). Animal testing has shown An unusual pattern of cancers of the brain and at least one nickel compound (nickel subsul respiratory system has been seen in workers fide) to be a potent carcinogen (Sunderman, exposed to vinyl chloride, now a well- 1976). As yet there are no definitive reports of established human liver carcinogen. Mineral increased respiratory cancer among nickel- oil, in the form of cutting oils, is related to exposed workers outside the refining industry. cancers of the lung and digestive organs among Mineral oils represent a complex class of workers exposed to mists arising from machin materials ranging from crude petroleum through ing operations. For most of these materials, various fractions produced at refineries to prod there are supporting data from studies in experi ucts used as coolants and lubricants in various mental animals, although the animal tumors industries. Not all mineral oils exhibit carcin may not be at the same site as observed in ogenicity for the skin either in man or in labora humans. tory animals. Extraction and refining of oil from With the proliferation of epidemiologic stud Scottish shale (Scott, 1922) and old methods of ies of different occupational groups during the URL 04036 TABLE 2. Industrial Materials lor Which Epidemiologic Studies Suggest Carcinogenicity Acrylonitrile Asbestos Material Beryllium Cadmium Coke Oven Emissions Cutting Oils Site(s) Lung Colon. Rectum Esophagus Larynx Stomach Lung Lung Prostate Kidney Prostate Lung. Digestive organs Stomach. Large intestine Stomach References O'Berg. 1980 Selikoff et at. 1973 Selikoff et al. 1973 Steil and McGill, 1973; Shettigara and Morgan. 1975 Selikoff et al. 1973 Mancuso. 1980; Wagoner et al. 1960 Lemen et al. 1976 Potts. 1965; Kipling and Waterhouse, 1967; Lemen et al. 1976 Redmond et al. 1972 Redmond et al. 1972 Waterhouse. 1972 Decouf*.1978 Jarvholm et al.. 1981 Ethylene Oxide/Ethylene Dichloride Lead Polychlorinated Biphenyls Vinyl Chloride Blood (leukemia) Stomach Lung Skin (melanoma) Brain Lung Hogstedt et al, 1979 Hogstedt et al, 1979 Cooper and Gaffey, 1975 Bahn et al, 1976 Waxweiler et al. 1976 Waxweiler et al. 1976 OCCUPATION -- 323 aMULi 'iuva ml--H ypiiapuuon jp URL 04037 324 -- THE CAUSES OF CANCER TABLE 3. Occupational Groups Associated with High Risks for Cancer, With No Specific Agents Identified Occupational Group Benzoyl Chloride Manufacture Chemists Cosl Miners Coke By-Product Plant Workers Foundry Workers Leather Workers Metal Miners Oil Refinery/Petrochemical Workers Painters Printing Workers Rubber Industry Work Areas Textile Workers Woodworkers Sifefs) Lung Brain Lymphatic and Hematopoietic Tissues Pancreas Stomach Colon Pancreas Lung Bladder Larynx Mouth. Pharynx Lung Brain Brain. Leukemia, Multiple Myeloma. Stomach Esophagus. Lung, Stomach Blood (leukemia) Lung Mouth. Pharynx Sadder Blood (leukemia) Brain Lung Prostate Stomach Nasal Cavity and Sinuses Lymphatic Tissue (Hodgkin's disease) Reterencas Sakabe et al, 1976 Olio and Ahlbom, 1960 Li et al. 1969; Olin and Ahlbom. 1980 Li et al. 1969 Rockette. 1977 Redmond et al. 1976 Redmond at al, 1976 Turner and Grace. 1938; McLaughlin and Harding. 1956; Koskeia el al. 1976; Gibson et al, 1977. DecouftA and Wood, 1979; Egan el al. 1979; Tola et al. 1979 Henry et al. 1931: Versluys. 1949; Wynder el al. 1963; Cole et ai. 1972; DecoulM. 1979 DecoufM. 1979 Decouf*. 1979 Wagoner et al. 1963 Theriault and Goulet. 1979; Alexander et ai. 1980 Thomas et al. I960 Hants et al, 1979 Viadana and Brass, 1972 Moss el al. 1972: Greenberg. 1972 Lloyd et al. 1977 Monson and Nakano. 1976; McMichaei et al. 1976. Monson and Fine. 1976 Monson and Nakano. 1976. McMichaei et al. 1976: Andjelkovich et al. 1977; Monson and Fine. 1976 Monson and Nakano. 1976; Monson and Fine. 1978 Monson and Nakano. 1976; McMichaei et al, 1976; Fox and Collier, 1976; Andjelkovich et al, 1977 Monson and Fine. 1978 McMichaei et al. 1976. Andjelkovich et al. 1977 Monson and Nakano. 1976. McMichaei et al. 1976 Andjelkovich el al. 1977; Monson and Fine. 1978 Acheson et al, 1972 Milham and Hesser, 1967; Petersen and Milham. 1974; Gnitterman at al. 1976; Greene et al. 1976 OCCUPATION --325 URL 04038 last 10 to 15 years, many industrial populations have been shown to be at high risk for certain forms of cancer. Table 3 lists some of these groups. Although the reports suggest that occu pational factors are responsible for the observed excesses, causative agents in the work environ ment have not been identified yet. Additional epidemiologic studies may be needed to con firm some of these findings, as in the case of painters with leukemia for which there is only one known report (Viadana and Bross, 1972). An important reason for the inability to pin point discrete carcinogenic agents in these stud ies is the multiplicity of exposures encountered in the subject industries. In addition to materials known to be used in various work settings, many other substances generated as by products of industrial processes have not been characterized completely. Industrial hygiene in vestigations of the past and present work envi ronments in high-risk industries can be impor tant, since they may uncover the unsuspected presence of known carcinogens. In addition, experimental studies in which test animals are exposed to the different materials found in high-risk industries may assist in sorting out likely carcinogenic substances. The foundry industry is a particularly interest ing example. Since early reports of an unusual frequency of respiratory cancer among foundry workers and metal grinders (Turner and Grace, 1938; Kennaway and Kennaway, 1947; McLaughlin and Harding, 1956), iron oxide has been considered a possible etiologic agent. However, there are now data to'suggest a link with polycyclic aromatic hydrocarbons present in ctrtain casting processes (Tola et al, 1979). Furthermore, exposure to iron oxide in the absence of known carcinogens produced a neg ative association with lung cancer in a recent study (Axelson and Sjoberg, 1979). Perhaps both these materials play a role in the cancer experience of foundry workers. CHARACTERISTICS OF OCCUPATIONAL CANCER Latent Period Tumors of occupational origin become clini cally manifest only after a definite time period (measured in years) has elapsed from first expo sure to the carcinogenic agent. Latent period is the term commonly used to characterize this phenomenon, and it can be viewed in several ways. For tumors of very low normal frequency (e.g., angiosarcoma of the liver, mesothelioma, osteosarcoma) or those of moderate frequency (bladder cancer, leukemia) when the relative risk is very high, virtually every case in an exposed group can be considered occupationally induced. Thus, the individual appearance times of the cases can be analyzed and a minimum or "average" latent period deter mined, keeping in mind that cases with the longest latency may not have been observed yet. On the other hand, for the more common tumors (e.g., lung cancer) when the relative risk is not very large, the latent period may be viewed as the time interval during which a significant increase in risk first occurs, or the time period of maximum risk, if one exists. The minimum latent period is usually five years for noncutaneous cancers, with a large proportion of cases appearing between 10 and 30 years after first exposure. Occupational skin cancers exhibit a wide span of latent periods ranging from intervals of less than one year (Sexton, 1960) to well over 50 years (Henry, 1947). An interesting illustration of exposurespecific latent periods is Henry's data on cuta neous epithelioma of occupational origin. He noted that the "average" latent period for min eral oil-induced lesions was significantly longer than for cases produced by pitch, tar, and tar products (Henry, 1947), suggesting that the latter substances are more potent carcinogens. The latency phenomenon has an important impact on the design of epidemiologic studies of occupational groups exposed to potential or recognized carcinogens. Thus, studies covering a limited number of years since the work force was first exposed to the material (e.g., 10 years) may fail to show an increase in risk simply because enough time has not elapsed for occu pationally related tumors to become manifest. This problem may be compounded if the work ers are very young, since they will not have reached the peak years of normal cancer in cidence during the course of a relatively short follow-up study. Dose-Response Relationships Where data are available indicating the amount of exposure to a carcinogenic sub stance received by workers, whether it be a qualitative or quantitative estimate, an increas ing risk with increasing exposure is generally observed. For example, this has been seen for lung cancer in relation to arsenic (Lee and 326--THE CAUSES OF CANCER URL 04039 Fraumeni, 1969), asbestos (Knox et al, 1968; Newhouse, 1969; Enterline et al, 1973; McDonald et al, 1980), uranium mining (Wa goner eta), 1965; lundin et al, 1971), and coke oven emissions (Lloyd, 1971; Mazumdar et al, 1975). Thus, reduction in environmental levels of carcinogenic agents should reduce the rela tive risk of disease. Epidemiologic studies have failed to provide adequate evidence for the existence of "thresholds" for occupationally related cancers, i.e., environmental levels of exposure below which there is no detectable increase in the incidence of cancer related to specific agents. Both biologic and statistical considerations enter the discussion of this point (Schneiderman et al, 1979). Low levels of exposure to particular agents may not appear to be associated with an unu sual incidence of cancer within a given period of time, but with additional follow-up a definite pattern of increased incidence may be noticed. For example, men who worked at the sides of coke ovens, who presumably had much less exposure to the emissions than topside workers, did not at first exhibit a significant increase in lung cancer (Lloyd, 1971), but after a longer follow-up period, they did experience a real increase (Redmond et al, 1972). Thus, it may be that dose and latent period are inversely related, at least for some carcinogenic agents. Multiple Sites Industrial materials capable of causing cancer in exposed workers may produce tumors at more than one anatomic site. This appears to be true for asbestos, coal carbonization processes, mustard gas, nickel refining, and radiation. Sometimes this is not immediately evident be cause epidemiologic studies are not always designed to examine a variety of cancer sites simultaneously. An excess risk for two or more cancers may be due to interactions of the putative agent with nonoccupational factors or with other agents present in the work environ ment. For certain agents, evidence of an association with more than one cancer site is either weak or so recent that confirmatory data from other studies are desirable. Thus, arsenic has been linked to tumors of the lymphatic tissues based on a study of pesticide makers (Ott et a), 1974), although such a link has not been reported among smelter workers (Lee and Fraumeni, 1969). Cancers of the upper respiratory tract have been mentioned among chromate work ers, but the data are very sparse (Hueper, 1966a). The fact that nonmalignant nasal condi tions (e.g., perforation of the nasal septum) do occur frequently among heavily exposed chro mate workers (Mancuso, 1951) may have lent credence to this possibility. It should be noted that exposure to aromatic amines is associated with cancers of the upper urinary tract (ureter, kidney), although the number of reported cases is small in relation to bladder tumors (Zavon et al, 1973; Tsuchiya et al, 1975). Age at Diagnosis or Death Generally, the age at which occupational tumors appear is a function of the age at first exposure and the attained age distribution of the particular group under study. Latency and amount of exposure are determining factors also. Table 4 shows data pertaining to age at diagnosis (or death) and latent period for select ed occupational cancers. It can be seen that tumors have arisen as early as the fourth decade of life, some even in the twenties. Part of the apparent shift to younger ages in these data (compared to the general population) is due to the method of case ascertainment. Most of these cases are persons who were actively employed in the industry when the cancer was first diagnosed, i.e., the easiest cases to identify. Cases appearing after retirement or among per sons who left the industry ate generally more difficult to trace. Occupational cancers are still manifest after age 65 among individuals who have retired from hazardous industries. Thus, increased lung cancer has been seen among retirees from asbestos plants (Enterline et al, 1973), smelters (Pinto et al, 1978), and coke plants (Collins and Redmond, 1976). On the other hand, studies confined to pensioners may not always demon strate the true magnitude of risk associated with specific work environments because of selec tive employment factors related to health (Reid and Buck, 1956). Histologic Types Several occupational cancers are associated with particular histologic types. The most ex treme example is that of vinyl chloride with liver cancer where angiosarcoma (hemangioendotheliosarcoma) is the exclusive cell type (Creech and Johnson, 1974; Spirtas and Ka- *8? 13-?5i 5Zjc q<das 2 uas 2 9 gQ--**- o = o ? a'S.3 i<ga.f s?3sl`9tsgy8-ft^ft Ss.ira3 ^RL 04040 Sn,M3.,ft3 a 8 3 a.gLS 9 8-^3 ! OCCUPATION -- 327 TABLE 4. Ages at Diagnosis or Death, and Latency Periods lor Selected Occupational Cancers Agent and Cancer Site Chromate* Respiratory System Number in Series 35 Bia(chloromethyl) ether Lung. Bronchus 36 Vinyl chloride Angiosarcoma of Liver 64 Aromatic amines Urinary Bladder 81 66 100 13 Age at Diagnosis or Death RAMOS AVERAGE 33-72 Mean ~ 517 (at death) 31-66 Mean = 48 3 (at diagnosis or death) 37-71 Median = 49 (at diagnosis) 25-72 Mean -* 52.1 25-66 <30 Mean - 44.9 to 60+ 34-56 Mean - 45 6 (at diagnosis) Latency Period (Years) RANGE AVERAGE References 4-47 Mean = 17.2 (as ol death) Machle and Gregorius. 1946 6-26 Mean -14 5 (as of diagnosis or death) Lemen et al. 1976 DeFonso and Kellon. 1976 9-38 Median -- 21 (as of diagnosis) Spirtas and Kaminski, 1976 2-42 Mean - 229 5-32 Mean = 16 5 <5 Mean = 16.0 to 36+ 9-29 Mean -- 16.6 (as ot diagnosis) Goidwater at at, 1965 Scott. 1952 Tsuchiya el al. 1975 Zavon el al. 1973 328--THE CAUSE5 OF CANCER minski, 1978). Myeloid leukemia {mainly acute) is the predominant type found among workers exposed to benzene (Vigliani, 1976; Infante et al, 1977) and among the early radiologists (Warren and lombard, 1966). Most nasal can cers seen among furniture workers are adeno carcinomas, whereas other cell types are preva lent in the shoe industry (Acheson, 1976). The following lung carcinogens appear to be associated with an unusual frequency of a par ticular cell type; Agent Arsenic Bis (chloromeihyl) ether Mustard gas Radiation Ce// Type Poorly differentiated epidermoid carcinomas Small cell undiffer entiated carcinomas Undifferentiated and squamous cell carcinomas Smalt cell undiffer entiated carcinomas Reference Newman et al, 1976 Figueroa et al, 1973 Wadaetal, 1968 Saccomanno et al, 1971 Particularly striking in these data is the predomi nance of small cell, undifferentiated carcinomas attributable to bis(chloromethyl) ether and radi ation. Virtually all of the occupational skin cancers are epitheliomas (carcinomas), with the possi ble exception of melanoma following exposure to polychlorinated biphenyls (Bahn et al, 1976). Mesotheliomas of the pleura and peritoneum have been consistently observed in populations exposed to asbestos (Wagner et al, 1960; Seli- koff et al, 1965; Newhouse and Thompson, 1965). Moreover, all three major commercial varieties of asbestos (amosite, chrysotile, and crocidolite) have been linked to these tumors (Selikoff et al, 1972; Greenberg and Lloyd Davies, 1974). EXPERIMENTAL CARCINOGENESIS AND EPIDEMIOLOGY Many industrial materials found to be carcin ogenic in laboratory animals have not been confirmed in humans (Tomatis et al, 1978). Table 5 shows a group of such chemicals. Part of the reason for a deficiency in epidemiologic data is related to problems of finding appro priate study groups with adequate information to characterize their exposure. Also, there may be a lack of personal data needed to trace workers over extended time periods to observe long-term health effects. Furthermore, human exposure to a chemical of interest may be confounded by simultaneous exposure to other TABLE S. Substances Showing Carcinogenic Activity in Laboratory Animals' SuPsfance Carbon Tetrachloride Chloroform Hydrazine and Derivatives Perchtoroethylene Polyehiorineted Biphenyls Trichloroethylene Urathana Industrial Uses Solvent, degreasing agent, fire extinguishing agent Solvent, fumigant, insecticide ingredient Oxygen scavenger in boiler water systems, drug manufacture (ieoniazid). solder fluxes Dry cleaning, vapor degreasing Dielectrics in capacitors and transformers, heat exchange fluids, hydraulic fluids Metal degreasing: extraction solvent for oils, fats, waxes: dry cleaning Chemical intermediate tor manufacture of pharma ceuticals. pesticides, fungi cides. textile finishes `No definitive human evidence is availably howavtr. toxic agents. In other instances, an inordinate amount of effort is required to obtain sufficient numbers of exposed workers, or it may be that no information exists to define who has been exposed. Table 6 outlines important differences be tween experimental and epidemiologic studies. Three general features of each approach are highlighted: selection of exposed and control groups, exposure to the hazardous material, and determination of health effects. In human stud ies, the first and third phases are implemented in a less than ideal fashion scientifically, whereas the exposure period is exactly the reality we wish to evaluate. On the other hand, the first and last steps are carried out to the ultimate degree of scientific precision in the laboratory, whereas the exposure phase is a contrived but, hopefully, useful approximation to the human counterpart. Thus, rt seems that for some industrial materi als epidemiologic studies will be invaluable for risk assessment, while for others laboratory test ing will be the most expeditious means of evaluating carcinogenicity because epidemio logic studies would be impossible or very diffi cult to carry out. lie firs de sing id IS >n late ient :hat een beies. are trol and udited lly, the nd, the the sa :ion :enfor estof liolifft- OCCUPATION -- 329 TABLE 8. V F*9turm Experimental and Epidemiologic Approaches to the Study of Occupational Cancer Animal Tasting Human Studiaa Selection of exposure and control groups Random allocation of homogeneous populations of animals to "exposure" or "no exposure" conditions Usually, a nonrandom selection of a heterogeneous group of "exposed" individuals and. if you are lucky, an "unexposed" group approximating a true control group might be available Exposure period Administration of well-defined levels of single agents through unique routes of entry. Careful control of other factors (e.g., diet) Individual exposure levels variable and usually unknown. Multiple exposures may compli cate the situation. Other factors may be operating (e.g.. smoking). Determination of health effects Detailed pathologic examination of all body systems In studies of fatal disease (e.g., cancer) the only health effect information may be death certification diagnosis. DATA NEEDS Exposure Data In planning an epidemiologic study of occu pational exposure to a suspected carcinogenic material, many questions should be raised con cerning the nature and degree of exposure in the workplace. Figure 1 outlines some of these considerations. Not all of the questions may have answers, and the prospects for initiating the study may depend on whether or not the key ones can be addressed. These include number of exposed workers, level of exposure, and Who were, and are, the producers? How long has it been produced? How is it made? Is there any significant axposure to X during its production? Have there been changes over time in the tech nology of production? How extensive is (or was) the production (e.g., one small plant or several large ones)? Where Nature How Extensive Any Exposure Duration in what industrial or commercial activities is it used? Is it always found together with other toxic agents or are thera industrial uses where it Is the only exposure? How many uses? How many people likely to be exposed for each use? Do people who use it actually get exposed to it? How long has each use been going on? Figure 1. Considerations in planning an epidemiologic study of axpoaure to agent X 330 --THE CAUSES OF CANCER duration of use. After going through this proc ess, one must make a decision as to whether an exposed working population really does (or did) exist. This involves a judgment as to what constitutes an exposure that is likely to produce a detectable effect, if indeed the material is carcinogenic for man. Input from several dis ciplines, including industrial hygiene and toxi cology, is valuable at this point. Worker Data Weir, 1968). Other unions maintain valuable information on their members in the form of dues records. These files depict weekly or monthly payments of dues so that approximate dates and duration of employment may be estimated. In addition, they may provide suffi cient identifying information to trace persons for current vital status. The existence and availabili ty of these records could be a major determin ing factor in undertaking a study of an industry whose work force is distributed among thou sands of small establishments. The conduct of epidemiologic studies of can cer in industry requires data that specify who worked where and when, and to what were they exposed. Thus, records must exist that define the population at risk in terms of demo graphic and occupational factors. Industry is in the best position to provide this information from data contained in administrative files such as personnel and employment records. These records have been used to define groups of workers employed in the past in suspect indus tries (historical cohort studies). Personal data are needed to trace terminated employees so their subsequent disease experience can be analyzed through ascertainment of morbidity or mortality. Employment information, such as duration of employment in specific jobs and departments, is utilized to characterize an employee's potential exposure to various hazardous agents. In addition, industry often has access to death information (sometimes even death certificates) for individuals who die while still on the employment rolls or as pen sioners. These data can be analyzed in their own right using the proportionate mortality approach for preliminary indication of an unu sual disease pattern. Labor unions are another potential source of data that should not be overlooked. In particu lar, trade unions may represent the only central source of information on their membership, because skilled tradesmen and construction workers tend to be scattered throughout many industries. Several trade unions operate death benefit plans for their members, which provide a mechanism for compiling cause of death information on large numbers of workers in a particular trade. Studies of several such groups have been reported based on these data systems (Decoufle et aJ, 1977; Lloyd et at, 1977; Egan et al, 1979). Furthermore, membership mailing lists have proved useful in establishing contem porary cohorts for forward studies (Dunn and Identification of High-Risk Croups A number of data sources exist that provide valuable information on relationships between cancer at specific sites and occupation on a broad-scale basis. Vital Statistics Data Henry and the Kennaways examined the oc cupational distribution of persons dying from selected cancers based on death certificate en tries for England and Wales during the 1920s and 1930s (Henry et al, 1931; Kennaway and Kennaway, 1936, 1947). Among lung cancer deaths in those two decades, they identified several high-risk occupations exposed to coal tar products, which were subsequently con firmed by formal epidemiologic studies. Analy sis of bladder cancer deaths also showed an association with coal tar occupations. Leather tanning and dyeing were two other trades linked to high rates of bladder cancer in these data, and both processes have come under recent scrutiny. Since those early studies, similar data have been available for analysis every ten years in England and Wales (Registrar General, 1958, 1971, 1978) and represent a valuable resource for screening purposes, keeping in mind that interpretation must be done with caution. In other countries, investigators have used death certificate occupational data in the same way (Versluys, 1949; Morrison, 1957), but only once has it been done on a national scale in the United States (Guralnick, 1963). However, a comprehensive analysis of occupation and cause of death for the State of Washington has been completed (Milham, 1976). Furthermore, surveys limited to a specific cancer site in an area where the mortality rate is unusually high are useful (Gottlieb et al, 1979). able n of v or mate ' be is for ibiliminustry hou- vide 'een >n a ocrom en)20s and ncer ified to :onlalyi an ither ades hese nder have rs in 958, urce that i. In eath way >nce the r, a and has ore, i an ligh OCCUPATION-- 331 Cancer Surveys but may spread into the general environment. In this way, large populations may be at unusual As part of the Third National Cancer Survey risk of cancer from such agents as arsenic, in the United States, occupational histories asbestos, polycyclic hydrocarbons, radioactive were obtained on a random sample of incident materials, and vinyl chloride. Although the ex cases. These data were analyzed in such a way tent of the hazard resulting from environmental as to draw attention to strong associations be pollution or contaminated consumer products is tween specific cancer sites and broad occupa not clear, there is no question that occupational tional groups (Williams et al, 1977). Thus, high- groups are invaluable as sentinels for identifying risk occupations could be identified for further, and evaluating risks to the general population. in-depth study. Unfortunately, the amount of The importance of epidemiologic studies available data limited the specificity of occupa within occupational groups can be appreciated tional information that could be examined. in the case of asbestos, in which regulatory agen cies have been forced to evaluate the potential Cancer Registries public health impact of the ever increasing instances of general population exposure to this Unfortunately, cancer registries in the United material. A long series of epidemiologic studies, States have not systematically gathered occupa dating back to 1955, have documented the tional information on their cases. Hopefully, in carcinogenic effects of exposure to asbestos in the future, steps will be taken to collect and the workplace. These data provide the basis for analyze job histories in at least certain types of assessing the potential impact of general en cancer registry systems. ^ vironmental asbestos exposures in the form of (1) crushed stone used on roads and public Hospital-based Data recreation areas, (2) insulation sprayed on struc tural supports for new buildings, (3) decaying Special surveys of bladder cancer in various insulation materials in already existing build localities have produced clues to occupational ings, and (4) asbestos-contaminated water, relationships (Wynder et al, 1963; Dunham et drugs and beverages. al, 1968; Anthony and Thomas, 1970; Cole et It would be difficult to document, directly, al, 1972}. In addition, studies of lung cancer the carcinogenic effects of asbestos attributable cases and prior work history have provided to these intermittent or low-level exposures. leads for further research (Wynder and Graham, 1951; Breslow et al, 1954; Menck and Hender son, 1976; Blot et al, 1978). Instead, inferences are often drawn from occu pational studies in choosing a prudent course of action. Roswell Park Memorial Institute, a major cancer treatment and research Center located in Buffalo, New York, gathered lifetime occupa References tional histories on all incoming patients before a diagnosis was made. These data were the sub ject of a comprehensive analysis (Houten et al, 1977a) and several specialized reports (Viadana et al, 1976; Houten et al, 1977b; Brass et al, 1978; DecouflC, 1979). Some previously unsus pected relationships were identified (e.g., dairy farmers with bladder cancer; hairdressers with cancers of the ovary and corpus uteri and lymphomas) and remain to be confirmed. Other findings were consistent with previous reports (e.g., leather workers with bladder cancer), demonstrating the value of such a data source. CONCLUSION Occupational hazards have implications beyond the work force because many chemical and physical agents are not confined to a plant Acheson ED: Nasal cancer in the furniture and boat and shoe manufacturing industries. Prev Med 5:295--315. 1976. Acheson ED, Cowdell RH. Rang E: Adenocarcinoma of the nasal cavity and sinuses in England and Wales. Br J Ind Med 29:21-30. 1972. Aksoy M, Erdem S, Dined G: Leukemia in shoeworkers exposed chronically to benzene. Blood 44:837-841, 1974. Alexander V, Leffingwell S, Lloyd |W, et al: Brain cancer in petrole um workers--a case series report. Am | Ind Med 1:115-123. 1980. Andjelkovich D. Taulbee J, Symons M: Mortality experience of a cohort of rubber workers, 1964-1973. | Occup Med 18.387-394, 1976. Andjelkovich, D, Taulbee J, Symons M, et al: Mortalitv of rubber workers with reference to work experience. I Occup Med 19:397- 405. 1977. Annual Report of the Chief Inspector of factories and Workshops for the Year 1932. London, HMSO, 1933, pp. 103-104. Anthony HM, Thomas CM: Tumors of the urinary bladder: an analy sis of (he occupations of 1,030 patients in Leeds, England. |NO 45:879-895, 1970. Axelson O, Dahlgren E, Jansson CD. et al: Arsenic exposure and mortality: a case-referem study from a Swedish copper smelter. Br J Ind Med 35:8-15. 1978. Axelson O. Sjoberg A: Cancer incidence and exposure to iron oxide dust. J Occup Med 21:419-422, 1979. 332 -- THE CAUSES OF CANCER Baetier AM: Pulmonary carcinoma in chromate workers. I. A review of the literature and report of cases. AMA Arch Ind Hyg Occup Med 2:487-504, 1950a. Baetier AM: Pulmonary carcinoma in chromate workers. II. inci dence on basis of hospital records. AMA Arch Ind Hyg Occup Med 2:505-516, 19S0b. Bahn AK, Rosenwaike I, Herrmann N, et al: Melanoma after expo sure to PC8's. N Engl | Med 295:450, 1976. Ball CB: Tar cancer. Ir J Med Sci 80:85-87, 1885. Bell): Paraffin epithelioma of the scrotum. Edinburgh Med J 22:135- 137. 1876. Birfstrup PL, Case RAM: Carcinoma of the lung in workmen in the bichromates-producing industry in Great Britain. 6r ) Ind Med 13260-264, 1956. Bingham E, Horton AW, Tye R: The carcinogenic potency of certain oils. Arch Environ Health 10:449--451, 1965. Blot W), Harrington JM, Toledo A, et al: Lung cancer after employ ment m shipyards during World War It. N Eng) / Med 299:620624. 1978 Boyd |T. Doll R. Faulds IS, et al: Cancer ot the lung in iron ore lhaematite] miners. Sr J Ind Med 27:97-105, 1970. Bres/ow L, Hoaglm t, Rasmussen G. etal: Occupationsandcigarette smoking as factors in lung cancer. Am | Public Health 44:171181, 1954. Brinton HP. Frasier S. Koven AL: Morbidity and mortality experi ence among chromate workers. Public Health Rep 67:835-847, 1952. Brinton LA, Blot WJ. Stone BJ, et al: A death certificate analysis of nasal cancer among furniture workers in North Carolina. Cancer Res 37:3473-34 74, 1977. Bross ID). Viadana E, Houten l: Occupational cancer in men ex posed to dust and other environmental hazards. Arch Environ Health 33:300-307, 1978. Butlm HT: Three lectures on cancer of the scrotum in chimney sweeps and others. Lecture II). Tar and paraffin cancer. Sr Med ] ii:66-71, 1892. Case RAM. Hosker ME: Tumour of the urinary bladder as an occupa tional disease in the rubber industry in England and Wales. Br | Prev Soc Med 8:39-50, 1954. Case RAM, Hosker ME, McDonald DB. et al: Tumours of the urinary bladder in workmen engaged in the manufacture and use of certain dyestuff intermediates in the British chemical industry. Part I. The role of aniline, benzidine. alpha-naphthylamine and betanaphthyfamine. Br J Ind Med 11:75-104, 1954. Case RAM. Pearson JT: Tumours of the urinary bladder in workmen engaged in the manufacture and use of certain dyestuff inter mediates in the British chemical industry. Part II. Further consider ation of the role of aniline and of the manufacture of auramine and magenta (fuchsine) as possible causative agents. Br J Ind Med 1 1.213-216. 1954. Christian HA: Cancer of the lung in employees of a public utility.) Occup Med 4:133-139, 1962. Cole P, Hoover R, friedell GH: Occupation and cancer of the lower urinary tract. Cancer 29:1250-1260, 1972. Collins )F, Redmond Of: The use of retirees to evaluate occupational hazards. J Occup Med 18:S95-602, 1976. Cooper WC. Gaffey WR: Mortality of lead workers, j Occup Med 17:100-107, 1975. Court Brown WM. Doll R: Expectation of life and mortality from cancer among British radiologists. Br Med I ii t81-107, 1956. Creech Jl )r, Johnson MN: Angiosarcoma of liver in the manufacture of polyvinyl chloride. J Occup Med 16:150-151, 1974. Cruickshank CND. Gourevitch A: Skin cancer of the hand and forearm. Br) Ind Med 9:74-79. 19S2. Cruickshank CND. Squire ]R: Skin cancer in the engineering in dustry from the use of mineral oil. BrJ Ind Med 7:1-11, 1950. Dalager NA. Mason TJ, Fraumenr )F Jr, et al: Cancer mortality among workers exposed to zinc chromate paints. J Occup Med 22:25-29, 1980. Davies JM: Bladder tumors in the electric-cable industry. Lancet ii: 143-146, 1965. Davies JM: Lung cancer mortality of workers making chrome pig ments. Lancet i:384, 1978. Decoufle P: Further analysis of cancer mortality patterns among workers exposed to cutting oil mists. (NCI 61:1025-1030, 1978. Oecoufle P: Cancer risks associated with employment in the leather and leather products industry. Arch Environ Health 34:33-37, 1979. Ocouffe P. Lloyd JW, Saivin LG: Causes of death among construc tion machinery operators, J Occup Med 19:123-128, 1977 Decoufle P. Wood DJ: Mortality patterns among workers in a gray iron foundry. Am J Epidemiol 109:667-675, 1979. DeFonso LR, Kelton SC: Lung cancer following exposure to chloromethyl methyl ether. An epidemiological study. Arch Environ Health 31:125-130, 1976. DeVilliers AJ, Windish JP: Lung cancer in a fluorspar mining com munity. I. Radiation, dust and mortality experience. Br j Ind Med 21:94-109, 1964. Doll R: The causes of death among gasworfcers with special refer ence to cancer of the lung. 8r J Ind Med 9:180-185, 1952. Doll R: Mortality from lung cancer in asbestos workers. Br J Ind Med 12:81-86, 1955. Ooil R: Cancer of the lung and nose in nickel workers. Br | Ind Med 15:217-223, 1958. Doll R, Fisher REW, Gammon E), et al: Mortality of gasworfcers with special reference to cancers of the lung and bladder, chronic bronchitis, and pneumoconiosis. Br J Ind Med 22:1-12. 1965. Doll R, Mathews JD, Morgan LG: Cancers of the lung and nasal sinuses in nickel workers: a reassessment of the period of risk. Br) Ind Med 34:102-105, 1977. Doll R, Morgan LG, Speizer FE: Cancers of the lung and nasal sinuses in nickel workers. Br ) Cancer 24:623-632, 1970. Doll R, Vessey MP, Beasley RWR. et al: Mortality of gasworkers -- final report of a prospective study. Br) Ind Med 29:394-406. 1972. Dunham LJ, Rabson AS, Stewart HL, et al: Rates, interview, and pathology study of cancer of the urinary bladder in New Orleans. Louisiana. JNCI 41:683-709, 1968. Dunn JE. Weir JM: A prospective study of mortality of several occu pational groups. Special emphasis on lung cancer. Arch Environ Health 17:71-76, 1968. Eckardt RE: Annals of industry. Noncasualties of the work place. J OCCUP Med 16:472-477, 1974. Edge JR: Asbestos related disease in Barrow-in-Fumess. Environ Res 11:244-247, 1976. Egan B, Waxweiler R|. Blade L, et al: A preliminary report of mortali ty patterns among foundry workers. J Environ Pathol Toxicol 2:259-272, 1979. Eimes PC. Simpson MJC: Insulation workers in Belfast. 3. Mortality 1940-66. Br J Ind Med 28:226-236, 1971. Enterline PE: Respiratory cancer among chromate workers. J Occup Med 16:523-526, 1974. Enterline P, Decoufle P, Henderson V: Respiratory cancer in relation to occupational exposures among retired asbestos workers. Br J Ind Med 30:162-166, 1973. Enterline PE, Kendrick MA: Asbestos-dust exposures at various levels and mortality. Arch Environ Health 15:181-186, 1967. Faulds JS, Stewart MJ; Carcinoma of the lung in haematite miners. J Pathol 8acieriol 72.353-366, 1956. Ferguson RS, et al: Symposium on anilin tumors of the bladder. J Urol 31:121-171, 1934. Figueroa WC, Raszkowski R, Weiss W: Lung cancer in chioromethyl methyl ether workers. N Engl ) Med 288:1096-1097, 1973. Fletcher DE: A mortality study of shipyard workers with pleural plaques. Br J Ind Med 29:142-145, 1972. Fox AJ, Collier PF. A survey of occupational cancer in the rubber and cablemaking industries: analysis of deaths occurring in 1972-74. Br j Ind Med 33:249-264, 1976. Galy P, Touraine R, Brune J. et al: Les cancers broncho-pulmonaires de ('intoxication arsenicale chronique chez les viticulture du Beaujolais Lyon Med 210:735-744, 1963. Gibbs GW, Horowitz I: Lung cancer mortality in aluminum reduc tion plant workers. J Occup Med 21:347-353, 1979 Gibson E5, Martin RH, Lockington JN: Lung cancer mortality in a steel foundry. J Occup Med 19:807-812, 1977. Girard R. Tolot F, Bourret J: Emopatie maiigne e benzolismo. Med lav 62:71-76, 1971. Goldwater LJ, Rosso A), Kleinfeld M: Bladder tumors in a coal tar dye plant. Arch Environ Health 11:814--817, 1965. Gottlieb MS, Pickle LW, Blot WJ, et al: Lung cancer in Louisiana: Death certificate analysis. JNO 63:1131-1137, 1979. Greenberg M; A proportional mortality study of a group of newspa per workers. Br J Ind Med 29:15-20, 1972. Greenberg M, Lloyd Davies TA: Mesothelioma register 1967-68. Br J Ind Med 31:91-104, 1974. Greene MH, Brinton LA. Fraumeni )F Jr, et al: Familial and sporadic i URL 0404 iMruc- a gfay chloroEnviron g comid Med 11 refer2. id Med id Med rs with hronic 965i nasal sk. Br | i nasal kers -- 4-406, v. and rfeam, occunviton lace. J Jr Res ortatioxicol trtalrty 3ccup 'lation i. Br ] levels ierv J Ider. ] lethyl leural r and 2-74. laires rs du fdoc- in a Med al tar iana: vspa- 8. Br radic OCCUPATION--333 Hodgkin's disease associated with occupational wood exposure. Lancet ri:626-627, 1978. Grufferman S. Duong T. Cole P: Occupation and Hodgkin's disease. JNCI 57:1193-1195, 1976. Guratnick L. Mortality by Occupation and Cause of Dealth among Men 20-64 Years of Age: United States, 1950. Vital Statistics. Special Reports. Vol. 53, No. 3. Washington. D.C.. USCPO, 1963. Hammond EC. Selikoff I), Lawther PL. et ah Inhalation of benz pyrene and cancer in man. Ann NY Acad Sci 271:116-124. 1976. Hanis NM, Stavraky KM, Fowler |L: Cancer mortality in oil refinery workers. | Occup Med 21:167-174, 1979. Harting FH, Hesse W: Der Lungenkrebs, die Bergkrankheit in den Schneeberger Gruben. Vrtljhrssch Gerichtl Med 30:296-309; 31:102-132, 313-337, 1879. Masterlik RJ, Finkel A). Miller CE: The cancer hazards of industrial and accidental exposure to radioactive isotopes. Ann NY Acad Sci 114:832-837, 1964. Hayes RB, Lilienfeld AM, Snell IM: Mortality in chromium chemical production workers: a prospective study. Int | Epidemiol 8:365-- 374, 1979. Hendricks NV. Berry CM. Lione IG. et al: Cancer of the scrotum in wax pressmen. I. Epidemiology. AMA Arch Ind Health 19:524-- 529, 1959. Henry SA: Cancer of the Scrotum in Relation to Occupation. Lon don, Oxford University Press. 1946. Henry SA: Occupational cutaneous cancer attributable to certain chemicals in industry. Br Med Bull 4:389-401, 1947. Henry SA. Kennaway NM, Kennaway EL: The incidence of cancer of the bladder and prostate m certain occupations. | Hyg (Cam bridge) 31:125-137, 1931. Hill AB. Faning EL: Studies in the incidence of cancer in a factory handling inorganic compounds of arsenic. I. Mortality experience m the factory. Br | Ind Med 5:1-6, 1948. Hogstedt C. Rohlen O, Bemdtsson BS, et ai: A cohort study of mortality and cancer incidence in ethylene oxide production workers. Br | Ind Med 36:276-280, 1979. Houten L. Bros5 IDJ. Viadana E, et al: A Retrospective Survey of Cancer in Relation to Occupation. DHEW (NHDSH) Pub No. 77-178. Washington, D.C., USCPO, 1977a. Houten L, Bross ID), Viadana E: Occupational cancers in men exposed to metals. Adv Exp Med Biol 91:93-102. 1977b. Hueper WC: Occupational and Environmental Cancers of the Re spiratory System. New York, Springer-Verlag, 1966, pp. 59-64a, J 27--131b. Hunter D: The Diseases of Occupations. Fourth edition. London, The English Universities Press, 1969, pp. 896-900. Infante PF, Wagoner JK. Rinsky RA, et al: Leukemia in benzene workers. Lancet ii:76-78, 1977. Jacob C, Anspach M: Pulmonary neoplasia among Dresden asbestos workers. Ann NY Acad Sci 132:536-548, 1965. Jarvhoim B, Lillienberg L, Sallsten G. et al: Cancer morbidity among men exposed to oil mist in the metal industry. J Occup Med 23:333-337, 1981. Kawai M, Amamoto H, Harada K: Epidemiologic study of occupa tional lung cancer. Arch Environ Health 14:859-864, 1967. Kennaway EL: The anatomical distribution of the occupational can cers. J Ind Hyg 7:69-93. 1925. Kennaway EL, Kennaway NM: A further study of the incidence of cancer of the lung and larynx. Br) Cancer 1:260-298. 1947. Kennaway NM, Kennaway EL: A study of the incidence of cancer of the lung and larynx. J Hyg (Cambridge) 36:236-267. 1936. Kipling MD: Oil and die skin. In: Annual Report of HM Chief Inspector of Factories, 1967. London, HMSO, 1968, pp. 10S119. Kipling MD, Waldron HA: Polycyclic aromatic hydrocarbons in mineral oil, tar, and pitch, excluding petroleum pitch. Prev Med 5:262-27B, 1976. Kipling MD. Waterhouse JAH: Cadmium and prostatic carcinoma. Lancet 1:730-731, 1967. Knox |F, Holmes 5, Doll R, et al: Mortality from lung cancer and other causes among workers in an asbestos textile factory. Br j Ind Med 25:293-303, 196B. Koskela RS, Hemberg S, Karava R, et al: A mortality study of foundry workers. Scand J Work Environ Health 2 (Supplement 11:73-89, 1976. Kuratsune M, Tokudome S, Shirakusa T, et al: Occupational lung cancer among copper smelters. Int) Cancer 13:552-558, 1974. Kuroda S, Kawahata K: Uber die gewerbliche Entsiehung des Lungenkrebses bei Generatorgasarbeitem. Z Krebsforsch 45:36-39, 1936. langard S, Norseth T: A cohort study of bronchial carcinomas in workers producing chromate pigments. Br J Ind Med 32:62^S, 1975. Lee AM, Fraumeni JF Jr: Arsenic and respiratory cancer in man: An occupational study. JNCI 42:1045-1052. 1969. Lemen RA, Johnson WM, Wagoner |K, Archer VE, et al: Cytologic observations and cancer incidence following exposure to BCME. Ann NY Acad Sci 271:71-80. 1976. Lemen RA, Lee JS, Wagoner JK, et al: Cancer mortality among cadmium production workers. Ann NY Acad Sci 271:273-279, 1976. Li FP, Fraumeni JF Jr, Mantel N, et al: Cancer mortality among chemists. JNCI 43:1159-1164. 1969. Ueben J: An epidemiologic study of occupational bladder cancer. AcM Un IntCancr 19:749-750. 1963. Itetoen J: Malignancies in asbestos workers. Arch Environ Health 13:619-621, 1966. Lloyd JW: Long-term mortality study of steelworkers. V. Respiratory cancer in coke plant workers. J Occup Med 13:53-66. 1971. Lloyd |W, Ciocco A: Long-term mortality study of steelworkers. I. Methodology. J Occup Med 11:299-310. 1969. Ltovd JW. Decoufle P, Saivin LG: Unusual mortality experience of printing pressmen. J Occup Med 19:543-550, 1977. Lorenz 6: Radioactivity and lung cancer: A critical review of lung cancer in the miners of Schneeberg and Joachimsthal. JNCI 5:115. 1944. Lueke AW: Epithelioma in carbon workers. Cleveland Med J 6.199-- 202, 1907. Lundin FE Jr. Lloyd JW, Smith EM, et al: Mortality of uranium miners in relation to radiation exposure, hard-rock mining and cigarette smoking --1950 through September 1967. Health Physics 16:571-578, 1969. Lundin FE Jr, Wagoner JK, Archer VE: Radon daughter exposure and respiratory cancer -- quantitative and temporal aspects. NIOSHNIEHS Joint Monograph No. 1. Springfield, Virginia, National Technical Information Service, 1971. Lynch J, Hanis NM, Bird MG, et al: An association of upper respira tory cancer with exposure to diethyl sulfate. J Occup Med 21:333-341, 1979. Mabuchi K, Lilienfeld AM. Snell LM: Cancer and occupational expo sure to arsenic: A study of pesticide workers. Prev Med 9:51-77, 1980. Machle W, Gregorius F: Cancer of the respiratory system in the United Stacs chromate-producing industry. Public Health Rep 63:1114-1127, 1948. Mancuso TF: Occupational cancer and other health hazards in a chromate plant: A medical appraisal. II. Clinical and toxicologic aspects. Ind Med Surg 20:393-407, 1951. Mancuso TF: Mortality study of beryllium workers' occupational lung cancer. Environ Res 21:48-55, 1980. Mancuso TF, Coulter EJ: Methodology in industrial health studies. The cohort approach, with special reference to an asbestos com pany. Arch Environ Health 6:210-226, 1963. Mancuso TF, El-Attar AA: Cohort study of workers exposed to betanaphthylamine and benzidine. J Occup Med 9:277-285. 1967. Mancuso TF, Hueper WC: Occupational cancer and other health hazards in a chromate plant: A medical appraisal. I. Lung cancers in chromate workers. Ind Med Surg 20:358-363. 1951. Manouvriez A: Maladies et hygiene des ouvriers cravaillant a la fabrication des agglomeris de houille et de brai. Ann Hyg Pub! Med Leg 45:459-482, 1876. Martischnig KM, Newell D|, Barnsley WC, et al: Unsuspected expo sure to asbestos and bronchogenic carcinoma. Br Med J i:746749, 1977. Martland HS: The occurrence of malignancy in radioactive persons. Am J Cancer 15:2435-2516, 1931. Mastromatieo E. Nickel: A review of its occupational health aspects. J Occup Nted 9:127-136, 1967. Matanoski GM. Seftser R, Sartwell PE. et al: The current mortality rates of radiologists and other physician specialists: specific causes of death. Am J Epidemiol 101:199-210, 1975. Mazumdar $, Redmond C, Sollecrto W, et al: An epidemiologic i t c i ( c fcc JD c c 2 e-C>* CO $ 3 334 -- THE CAUSES OF CANCER URL 04047 study of exposure to coal tar pitch volatiles among coke oven workers. 1 Air Pollut Control Assoc 2S:382-369. 1975. McDonald JC. Liddell FDK, Cibbs CW, et al. Dust exposure and mortality in chrysotile mining, 1910-75. Br J Ind Med 37:11-24, 1980. McLaughlin AtC. Harding HE: Pneumoconiosis and other causes of death m iron and steel foundry workers. AMA Arch ind Health 14.350-378, 1956. McMichael A|, Spirtas R, Gamble |F. et al: Mortality among rubber workers: Relationship to specific jobs. |Occup Med 18:178-185, 1976. McMichael AJ, Spirtas R. Kupper LI: An epidemiologic study of mortality within a cohort of rubber workers. 1964-72. J Occup Med 16:458-464. 1974. Melick WF, Escue HM. Naryka )). eS aV. The first reported cases of human bladder tumors due to a new carcinogen -- renylamine. ] Urol 74:760-766, 1955. Melick WF, Naryka JJ. Kelly RE: Bladder cancer due to exposure to paraammobiphenyl: A 17-year followup. I Urol 106:220-226, 1971. Menck HR, Henderson SE: Occupational differences in rates of lung cancer, i Occup Med 18:797-801, 1976. Merewether ERA: Asbestosis and carcinoma of the lung, in: Annual Report of the Chief Inspector of Factories for the Year 1947. London. HMSO. 1949, pp. 79-81. Michel-Briand C. Simonin M: Cancers broncho-pulmonaires survenus chez deux salaries occupes a un poste de travail dans le meme atelier de chramage electrolytique. Arch Mai Prof 38:1001-1013. 1977. Milham S: Occupational Mortality in Washington State, 1950-1971. DHEW (NIOSHI Pub Nos. 76-175-A.B.C. Washington, D.C., USGPO. 1976. Milham S Jr, Hesser )E: Hodgkin's disease in woodworkers. Lancet ii: 136-137, 1967. Milham 5 Jr. Strong T: Human arsenic exposure in relation to a copper smelter. Environ Res 7:176-182, 1974. Monson RR, Fine LJ: Cancer mortality and morbidity among rubber workers. )NCi 61:1047-1053, 1978. Monson RR, Nakano KK: Mortality among rubber workers. I. White male union employees in Akron. Ohio. Am | Epidemiol 103:284296. 1976. Morgan JG: Some observations on the incidence of respiratory can cer in nickel workers. Br J Ind Med 15:224-234, 1958. Morrison SL: Occupational mortality in Scotland. Br J Ind Med 14:130-132, 1957. Moss E, Scott TS, Athetley GRC: Mortality of newspaper workers from lung cancer and bronchitis 1952-66. Br) Ind Med 29:1-14, 1972. Newhouse ML: A study of the mortality of workers in an asbestos factory. Br J Ind Med 26:294-301, 1969. Newhouse ML, Berry G. Wagner JC, et al: A study of the mortality of female asbestos workers. Br J Ind Med 29:134-141, 1972. Newhouse ML, Thompson H: Mesothelioma of pleura and peritone* um following exposure to asbestos in the London area. Br J Ind Med 22:261-269, 1965. Newman JA, Archer VE, Saccomanno G, al: Histologic types of bronchogenic carcinoma among members of copper-mining and smelting communities. Ann NY Acad Sci 271:260-268, 1976. O'BergMT: Epidemiologic study of workers exposed to acrylonitrile. J Occup Med 22:245-252, 1980. Ohn CR, Ahlbom A: The cancer mortality among Swedish chemists graduated during three decades. Environ Res 22:154-161, 1980. Ott MC, Holder BB, Gordon HL: Respiratory cancer and occupa tional exposure to arsenical*. Arch Environ Health 29:250-255, 1974. Paris JA: Pharmacologia. Fifth edition. London, W. Phillips, 1822, Volume 2, pp. 88-89. Passey RD: Experimental soot cancer. Br Med J ii: 1T T2--11T 3, 1922. Pasternack SS, Shore RE, Albert RE: Occupational exposure to chloromethy! ethers. A retrospective cohort mortality study (1948-1972). J Occup Med 19:741-746, 1977. Pedersen E, Hogetveit AC. Andersen A; Cancer of respiratory organs among workers at a nickel refinery in Norway. Int | Cancer 12:3241, 1973. Peller S: Lung cancer among mine workers in Joachimsthal. Hum Biol 11:130-143, 1939. Petersen GR. Milham S Jr: Hodgkin's disease mortality and occupa tional exposure to wood. JNCI 53:957-958, 1974. PetO ), Doll R, Howard SV. et al: A mortality study among workers in an English asbestos factory. Br J Ind Med 34:169-173. 1977. Pinto SS, Henderson V, Enterline PE: Mortality experience of arse nic-exposed workers. Arch Environ Health 33:32S-331, 1978. Polednak AP, Stehney AF, Rowland RE: Mortality among women First employed before 1930 in the U.S. radium dial-painting in dustry. Am ) Epidemiol 107:179-195, 1978. Pott P: Cancer scroti, fn: Chirurgical Observations. London, Hawes, Clarke, and Collins. 177S, pp. 63-68. Potts CL: Cadmium proteinuria. The health of battery workers ex posed to cadmium oxide dust. Ann Occup Hyg 8:55-61, 1965. Redmond CK, Ciocco A, Lloyd JW, et al: Long-term mortality study of steelworkers Vi. Mortality from malignant neoplasms among coke oven workers. J Occup Med 14 621-629, 1972. Redmond CK, Strobino BR, Cypess RH: Cancer experience among coke by-product workers. Ann NY Acad Sci 271:102-115, 1976. The Registrar General's Decennial Supplement. England and Wales, 195). Part II, Vol. 2. Occupational Mortality Tables. London. HMSO, 1958. The Registrar General's Decennial Supplement, England and Wales, 1961. Occupational Mortality Tables. London. HMSO, 1971. The Registrar General's Decennial Supplement, England and Wales. 1970-72. Occupational Mortality Series DS No, 1 London, HMSO, 1978. Rehn L: Blasengeschwulste bei Fuchsm-Arbeitem. Arch Klin Chir SO.588-600, 1895. Reid DD, Suck C: Cancer in coking plant workers. 9r) Ind Med 13:265-269, 1956. Rencher AC. Carter MW, McKee OW: A retrospective epidemiologi cal study of mortality at a large western copper smelter. J Occup Med 19:7S4-758, t977. Robinson C, Lemen R, Wagoner JK: Mortality patterns, 1940-1975, among workers employed in an asbestos textile friction and pack ing products manufacturing facility. In Lemen R, Dement )M teds): Dusts and Disease. Park Forest South. Illinois, Pathotox Publish ers, 1979, pp. 131-143. Rockette HE: Cause-specific mortality of coal miners. I Occup Med 19:795-801, 1977. Roth F: Uber den Bronchialktebs arsengeschadigter Winzer. Vir chows Archiv 331:119-137, 1958. Royle H: Toxicity of chromic acid in the chromium plating industry (1). Environ Res 10:39-53, 1975. Saccomanno G, Archer VE, Auerbach O, et al; Histologic types of lung cancer among uranium miners. Cancer 27:515-523, 1971. Sakabe H, Matsushita H, Koshi S: Cancer among benzoyl chloride manufacturing workers. Ann NY Acad Sci 271:67-70, 1976. Schneiderman MA, Decoufld P, Brown CC. Thresholds for environ mental cancer: Biological and statistical considerations. Ann NY Acad Sci 329:92-130, 1979. Scott A: On the occupation cancer of the paraffin and oil workers of the Scottish shale oil industry. Br Med J ii: 110S--1109. 1922. Scon T5: The incidence of bladder tumours in a dyestuffs factory. Br J Ind Med 9 127-132, 19S2. Selikoff IJ, Churg J, Hammond EC: Asbestos exposure and neoplasia. JAMA 188:22-26, 1964. Selikoff IJ, Churg J, Hammond EC: Relation between exposure to asbestos and mesothelioma. N Engl J Med 272:560-565, 1965. Selikoff IJ, Hammond EC, Churg): Asbestos exposure, smoking, and neoplasia. JAMA 204.106-112, 1968. Selikoff IJ, Hammond EC, Churg |: Carcinogenicity of amosite asbes tos. Arch Environ Health 25:183-186, 1972. Selikoff IJ, Hammond EC, Seidman H: Cancer risk of insulation workers in the United States, in: Biological Effects of Asbestos. Lyon, International Agency for Research on Cancer, 1973, pp. 20^-216. Selikoff IJ, Seidman H. Hammond EC. Mortality effects of cigarette smoking among amosite asbestos factory workers. JNCI 65:507513, 1980. Seltser R, Sartwell PE: The influence of occupational exposure to radiation on the mortality of American radiologies and other medical specialists. Am J Epidemiol 81:2-22, 1965. OCCUPATION -- 335 Sexton RJ: The hazards to health in the hydrogenation of coal. IV. The control program and the clinical effects. Arch EnvTron Health 1-208-231, 1960. Shettigara FT, Morgan RW: Asbestos, smoking, and laryngeal carci noma. Arch Environ ^Health 30:517-519, 1975. Southam AH, Wilson, SR: Cancer of the scrotum: the etiology, clinical features, and treatment of the disease. Br Med ] ii:971973, 1922. Spirtas R, Kaminski R: Angiosarcoma of the liver in vinyl chlo ride/polyvinyl chloride workers. 1977 update of the NOSH regis ter. I Occup Med 20:427-429, 1978. Steil PM, McGill T: Asbestos and laryngeal carcinoma. Lancet ii:416-417, 1973. Sunderman FW |r: A review of the carcinogenicities of nickel, chro mium and arsenic compounds in man and animals. Prev Med 5.279-294. 1976. Tabershaw IR, Copper WC. Baizer |l: A labor-management occupa tional health service in a construction industry. Arch Environ Health 21:784-786, 1970. Theriault G, Goulet L: A mortality study of oil refinery workers. | Occup Med 21:367-370. 1979. Thomas TL. Decoufl* P, Moure-Eraso R: Mortality among workers employed in petroleum refining and petrochemical plants. J Occup Med 22:97-103, 1980. Tokudome S, Kuratsune M: A cohort study on mortal ity from cancer and other causes among workers at a metal refinery Ini 1 Cancer 17:310-317. 1976. Tola S, Koskeia RS, Hernberg S, et al: Lung cancer mortality among iron foundry workers.) Occup Med 21:753-760. 1979. Tomatis L. Agthe C. Bartsch H, et al: Evaluation of the carcinogeni city of chemicals: A review of the monograph program of the International Agency for Research on Cancer (1971-19771. Can cer Res 38:877-885, 1978. Tsuchiya K, Okubo T. Ishizu 5: An epidemiological Study of occupa tional bladder tumours in the dye industry of Japan. Br J Ind Med 32:203-209, 1975. Turner HM, Grace HG: An investigation into cancer mortality among males in certain Sheffield trades. | Hyg (London) 38 90103. 1938. Versluys |J: Cancer and occupation in the Netherlands. Br) Cancer 3:161-185. 1949. Viadana E, Bross ID): Leukemia and occupations. Prev Med 1:513- 521. 1972. Viadana E, Bross ID), Houten L: Cancer experience of men exposed to inhalation of chemicals or to combustion products. J Occup Med 18:787-792. 1976. Vigliani EC: Leukemia associated with benzene exposure. Ann NY Acad Sci 271:143-151. 1976 Vigliani EC, Barsotti M: Environmental tumors of the bladder in some Italian dye-stuff factories. Med Lav 52:241-250, 1961. Volkmann R: Ueber Theer-, Paraffin und Russkrebs (Schomstein- fegerkrebs). In. Beitrage zur Chirurgie. Leipzig. Druck und Verlag von Breitkopf und Hartel, 1875. pp. 370-381. Wada S. Miyanishi M, Nishimoto Y. et al: Mustard gas as a cause of respiratory neoplasia in man. Lancet i:1161-1163, 1968. Wagner )C. Sleggs CA, Marchand P: Diffuse pleural mesothelioma and asbestos exposure in North Western Cape Province. Br J Ind Med 17:260-271, 1960. Wagoner |K, Archer VE, Lundin FE Jr, et al: Radiation as the cause of lung cancer among uranium miners. N Engl j Med 273:181-188, 1965Wagoner IK, Infante PF, Bayliss DL: Beryllium: An etioiogic agent in the induction of lung cancer, non-neopiastic respiratory disease and heart disease among industrially exposed workers. Environ Res 21:15-34, 1980. Wagoner |K. Miller RW, lundin FE |r. et al: Unusual cancer mortality among a group of underground metal miners. N Engl J Med 269:284-289. 1963. Warren 5: Longevity and causes of death from irradiation in physi cians. JAMA 162:464-468. 1056. Warren 5. lombard OM: New data on the effects of ionizing radia tion on radiologists. Arch Environ Health 13415--421. 1966. Waterhouse JAH: lung cancer and gastro-intestinal cancer m miner al oil workers. Ann Occup Hyg 15 43-44, 1972. Waxweifer R|, Stringer W, Wagoner ]K. et al: Neoplastic risk among workers exposed to vinyl chloride. Ann NY Acad Sci 271 40--48. 1976. Weil CS, Smyth HF Jr, Nale TW: Quest for a suspected industrial carcinogen. AMA Arch Ind Hyg Occup Med 5:535-547, 1952. Williams RR, Stegens NL, Coldsmith JR: Associations of cancer she and type with occupation and industry horn the Third National Cancer Survey Interview. JNCI 59:1147-1185. 1977. Wynder El, Graham EA: Etioiogic factors in bronchogenic carcino ma with special reference to industrial exposures. AMA Arch Ind Hyg Occup Med 4:221-235, 1951. Wynder EL. Onderdonk J, Mantel N: An epidemiological investiga tion of cancer of the bladder. Cancer 16:1388-1407. 1963. Zavon MR. Hoegg U. Bingham E: Benzidine exposure as a cause of bladder tumors. Arch Environ Health 27:1-7, 1973. iy n ai&Hk .CANCER EPIDEMIOLOGY and PREVENTION DAVID SCHOTTENFELD, M.D. Chief of Epidemiology and Preventive Medicine Memorial Sloan-Kettering Cancer Center New York, New York JOSEPH F. FRAUMENI, JR., M.D. Chief, Environmental Epidemiology Branch National Cancer Institute Naiional Institutes of Health Bethesda, Maryland K y /f . URL 0UW9 U'W OF CONN. M 24 1982 HEALTH CENTER UtiRARY 1982 W. B. SAUNDERS COMPANY Philadelphia London Toronto Mexico CHy ftio de Janeiro Sydney Tokyo CHAPTER TWO Causation and Causal Inference Kenneth J. Rothman URL 04050 CAUSATION logic. For the empiricists, causation was an The Evolution of the Concept of Causation inferred relation between two events that were repeatedly observed to occur in the same time The understanding of causation has devel oped in parallel with the growth of science. Aristotle, whose scientific views dominated Western thought for centuries, described the following four types of cause for each object; (1) sequence. To infer causation required observa tion, prediction, and additional observation to verify the prediction. This profound change in scientific thought characterized the rise of mod ern science. Empiricism was not without its critics. Fore material cause, the matter of the object, (2) formal cause, the pattern imposed on the mat ter, (3) efficient cause, the force producing the object, and (4) final cause, the purpose of the object. Aristotle's description of causes did little to enhance understanding, at least when judged by today's standards. Coupled with the weight of his authority, Aristotle's views probably re pressed the further development of philosophic inquiry about causation for many centuries. most among them was David Hume, who ar gued that the mere temporal conjunction of two events cannot impart to an observer any knowl edge about a causal connection between them. Hume revealed the process of inductive in ference to be a faulty process: without direct evidence for a causal connection, which was impossible, induction could lead to incorrect inferences. Inductive logic, therefore, was not really a "logic" at all. With this simple but During these centuries, the prevailing doc profound skepticism, Hume shook the new trine of scientific philosophy was rationalism, foundation of empiric science, which regarded reason as the source of all Ever since Hume, defenders of empiricism i knowledge of the physical world. Rationalism have attempted to answer his criticism. The first flourished in an era when religion overshad owed science, but as modern science took root in the seventeenth century, empiricism replaced rationalism as the accepted philoso phy of science. Empiricism substituted observa tion for reason, or rather combined observation with reason in a new methodology, inductive defender was Kant, whose answer to Hume was imbedded in his concept of the synthetic a prior/. He distinguished between two types of judgments: ana/ytic, which were logical truisms unrelated to empiric judgments, and synthetic, which were judgments that extended knowl edge. Kant believed that some synthetic judg- 15 J 16 -- BASIC CONCEPTS URL 04051 ments are made without reference to experi ence. For him, the explanation of the existence of these synthetic a priori judgments was the central issue in metaphysics. Causation was one of Kant's synthetic a priori judgments: that events had causes was known on the basis of synthetic a priori judgment, and induction was merely the process of identifying the actual cause. Though Hume died before Kant's arguments were published, Reichenbach (1951) supplies us with the hypothetical response that Hume might have given to Kant: How does it help us to know there is a cause if we wish to know what the cause is? It is true, if we knew there were no cause it would be nonsensical to search for one; but that is not our situation. We do not know whether there is a cause; in this situation we make inductive inferences, based on observation, and conclude, say, that the moon is the cause of the tides. This inductive inference is what I question; and it would remain just as questionable if you could prove the general statement that there is a cause. Reichenbach himself proposed a different solution to "Hume's problem" (as Kant de scribed it), a solution he contended could not have been proposed before the development of probability theory and quantum mechanics. He argued that Heisenberg's principle of uncer tainty, by which individual atomic phenomena are explainable only in probabilistic terms, serves as a proof that a concept of strict causal ity is not tenable. According to his view, natural laws are, in effect, only idealized statistical laws, and causality is not an ultimate principle, but only a substitute for statistical regularity. He proposed that ... the idea of a strict causality is to be abandoned, and that the laws of probability take over the place once occupied by the law of causality. Reichenbach concluded that scientific phi losophy has led us to the truth that empiric knowledge can never be certain. Hume's ob jection to the use of inductive logic was the inconsistency of a "logic" that is uncertain. For Hume, such a methodology was invalid; for Reichenbach, it is the best we can achieve, which validates it. Carnap (1962) took Reichenbach's argument further by attempting to for malize inductive logic with a set of probability laws quantifying the degree of confirmation of a hypothesis. A different answer to "Hume's problem" was proposed by Popper (1968), who rejected the view that the validity of scientific statements can be quantified by probabilities. He thought such a compromise position about the verifiabil ity of a causa! proposition to be only a sidestep ping of Hume's criticism of inductive fogic. Popper found that the inadmissibility of induc tive logic leads to a contradiction with the basic premise of empiric science only if it is assumed that all empiric scientific statements are conclu sively verifiable. Popper maintained that scien tific statements are conclusively decidable only in the negative -- they can be falsified -- but no scientific statement can be conclusively veri fied. Falsification, unlike verification, requires no inductive inference, only deductive logic. Popper's view coincides with the historical view of science; widely accepted theories have been subsequently replaced, after falsification, with new theories. No conclusive verification has ever occurred. Although Popper resolved Hume's problem by arguing that scientific progress does not involve inductive logic, whereas Reichenbach and Carnap attempted to modify and formalize that logic, all would agree that scientific knowl edge is uncertain. For Reichenbach and Carnap the nature of the uncertainty was that of proba bility; for Popper, it was that of tentativeness. Popper emphasized that the utility or content of a scientific statement rests on the degree of falsifiability of the statement: statements such as "Cod is one" or "Levitation is possible for those who try hard enough" have no scientific con tent, because neither is falsifiable. Popper's emphasis on the falsifiability of a theory has implications for current thought in carcinogene sis, as we shall see; the multi-stage theory, which is much discussed, is so general a state ment that it cannot be falsified; it therefore has little scientific content. Specific variants of the theory, however, such as Knudson's (1971) two-mutation model, are more easily falsifiable and therefore of greater scientific interest. Epidemiologic Concepts of Causation Philosophers of science have improved our understanding of the process of causal in ference, but there remains the need, at least in epidemiology, to formulate a general and co herent model of causation to facilitate the con ceptualization of epidemiologic problems. A first step is to distinguish conceptually between sufficient causes and components of sufficient causes (Rothman, 1976). A sufficient cause is one that by definition inevitably produces the CAUSATION AND CAUSAL INFERENCE --17 effect. Each occurrence of the effect is pre sumed to result from a sufficient cause. Pre cisely because a sufficient cause implies inevi tability of the effect, sufficient causes are difficult to identify in their entirety. Components of sufficient causes, however, imply by their presence only an increased probability of the effect; the actual outcome depends on the occurrence of the complementary component causes that are needed to complete a sufficient cause. The causes that epidemiologists study as causal risk factors are component causes. Though such causes are not "complete," block ing the action of a component cause will render an otherwise sufficient cause insufficient, and thereby prevent the effect, at least through one mechanism. It is still possible that the effect will occur anyway through another mechanism, i.e., a different sufficient cause, which does not include as a component cause the factor that has been blocked. Consider lung cancer as an example. Smok ing is a causal risk factor, but it is not a sufficient cause, because most smokers (more than 90 per cent) never get lung cancer. Most cases of lung cancer, however, are attributable to a causal mechanism that involves smoking as a compo nent cause. By preventing smoking, or by mak ing a nontoxic cigarette, most cases of lung cancer could be prevented. A few cases of lung cancer would still occur, though, because there are other sufficient causes of lung cancer that do not involve smoking, as evinced by the occur rence of lung cancer among nonsmokers. Some component causes of a sufficient cause may be passive components, those that contrib ute appropriate conditions for the other compo nent causes to act. Genetic characteristics that are risk factors fall into this category of compo nent causes. Nearly every complete cause has components that are genetic, and these can be thought of as susceptibility factors. Nongenetic component causes also may contribute to the effect in a similar way by providing the right conditions for other component causes to act. Ingested nitrite, for example, may cause cancer if it is chemically converted to nitrosamine in the stomach. The reaction is dependent on pH, with the highest yield occurring at a pH of 3.4, and lesser yields at a lower or higher pH (Mirvish, 1970). Very little conversion occurs below a pH of 1.5 or above a pH of 5.0. The pH of stomach contents depends on the type and quantity of food consumed; it is usually much lower than 3.4 when the stomach is empty, but may be near 3.4 after a large meal. Sufficient alkalinity of the gastric contents may be thought of as a passive component cause of cancer, permitting dietary nitrite to be converted to an active carcinogen, which may then induce can cer. Generally speaking, only active component causes of cancer are considered carcinogens. Passive component causes create the conditions in which active causes produce the transitions between stages; their passive role corresponds to that of a susceptibility factor, but not to what is usually meant by a carcinogen. This distinc tion between carcinogenic agents and other component causes of cancer is obviously not conceptually rigorous, but rather only a distinc tion that derives from common usage. CARCINOGENESIS The Multistage Theory of Carcinogenesis While some component causes thus establish a set of susceptible conditions, others actively play a role in converting the system from one stage to the next along a causal sequence. The precise number of stages in a causal path can not be determined meaningfully, because the causal sequence can always be described in finer detail, and the designation of stages is ultimately an arbitrary classification. Even a single abrupt transition event could be dissected into many graduations. Often, however, the transitions in a causal path appear to be reason ably categorized into one or more distinct events. Carcinogenesis has been demonstrated to have at least two and possibly many such transitions; carcinogenic agents are considered initiators if they act on an early stage in the sequence, and promoters if they act on a later stage. (Some scientists restrict the use of the term carcinogen to include only initiators and not promoters, but few epidemiologists adhere to this restriction.) A formalization of the implications of sequen tially acting component causes as applied to carcinogenesis has been discussed in terms of the multi-stage theory of carcinogenesis. Under this theory, the development of a cancer is the end stage of a sequence of stochastic transition al cellular events, for which the transition prob abilities are usually assumed to be constant over time. While theoreticians have not reached any consensus as to which form or forms of such a theory might actually apply in carcinogenesis, the consequences of various forms of the gener al theory have been examined in detail and compared with epidemiologic evidence. (For a URL 040' <.n r-j 18 -- BASIC CONCEPTS review, see Peto, 1977.) Such comparisons of theory and observations are difficult to interpret, because time trends in incidence, inaccuracies in reporting, and multiple etiologies, among other concerns, all work to distort the observa tions predicted by the theory. The mathematical assumptions and approximations for some forms of the theory have also been misunder stood (Moolgavkar, 1978). Despite these reservations, many scientists have been attracted to the multi-stage theory because of its generality and the close corre spondence between age-incidence curves pre dicted from the theory and those actually ob served (Cook et al, 1969), The utility of the theory is questionable, however, in view of the difficulty with which the theory might be fal sified. Indeed, the multi-stage theory in a gener al form could be used, as I have already indicat ed, as a general model of causation to apply to all etiologic situations. What makes the theory interesting and gives it scientific content are the restrictions that are incorporated into it for any given situation. For example, Moolgavkar and Venzon (1979) have recently shown that a two-stage model can generate age-incidence curves similar in shape to those generated from more complicated multi-stage models, thus ex tending the applicability of Knudson's (1971) two-stage mode! for embryonal tumors to in clude adult tumors. They argue that biologic interpretation (i.e., initiation and promotion) can be given to no more than two distinct stages, making their two-stage model more plausible than multi-stage alternatives. Induction Period and Latent Period Some scientists use the terms induction period and latent period interchangeably (in cubation period is also used), but the time interval between exposure to an etiologic agent and appearance of disease symptoms can be meaningfully partitioned in a way that suggests a distinction between these terms. The Oxford English Dictionary (1971) defines latent as "hid den," whereas induction is defined as the "act of causing or producing." Conceptually, then, the induction period is the time interval during which a sufficient cause is completed. In refer ence to a particular component cause, the induction period would begin at the time of action of the component cause and end when all the other component causes that constitute a sufficient cause have completed their action to initiate the disease. At the time of disease initiation, when the induction period ends, the latent period begins. In principle, the latent period could be reduced to near zero with more accurate means to detect presymptomatic dis ease; the length of the latent period is simply the time it takes for a disease, once initiated (by a sufficient cause), to be detected. The latent period is conceptually part of the causal process only to the extent that during the latent period disease may "spontaneously" dis appear or fail to progress, in which case clinical disease would not occur. Logically, however, such alterations in the progression of disease should be considered part of the causal process of clinical disease, thus implying that where the natural history of disease is not that of inevitable progression, the latent period is subsumed within the induction period, which accordingly would end not at the initiation of preclinical disease, but at the start of symptoms. For dis eases that, once initiated, progress inevitably to become clinically apparent, however, the latent period would follow the induction period. It is impossible, of course, to determine when the induction period ends and the latent period begins for any individual case, though this can be estimated for large groups of people. The conceptual distinction between induction period and latent period is worth maintaining, because these two time periods define the inter vals during which primary prevention (i.e., blocking a sufficient cause) and secondary prevention^i.e., preclinical detection and treat ment) would be applicable. For exposures that are continuous or intermit tent, there remains the difficulty in determining when the exposure acts etiologically. Custo marily, epidemiologists measure induction periods from the time of first exposure, because the relevant time of action is unknown, whereas the time of first exposure can usually be deter mined. A better practice, however, is to choose an arbitrary accumulation of exposure, and determine the time at which that level is reached. It is possible to vary the estimates for the sufficient dose of exposure and learn from the epidemiologic analysis which estimate is most appropriate. This procedure can also in corporate variations in exposure dose with time. It re known that some carcinogens have lengthy induction periods for certain cancers; an example is the occurrence of adenocarcino ma of the vagina 15 to 30 years after exposure of female fetuses to diethylstilbestrol. It appears probable, on the basis of the age distribution of cases, that a sufficient cause is completed dur- , the atent nore dis. the by a the l the dtslical jver, ease cess ? the able ned ngly lical disy to tent It is the riod can The tion ing, iteri.e., preeat- mil ling stonon :use reas ter>ose and I is for "om ? is invith ave ers; noure jars i of iur- I CAUSATION AND CAUSAL INFERENCE--19 ing puberty or adolescence for women who develop this disease; therefore the time before puberty or adolescence would be the induction period, and the interval following that period until symptoms appear would be the latent period. Frequently some types of cancer are de scribed as diseases of lengthy induction period, the implication being that the action of any carcinogen causing that cancer precedes the development of the tumor by a long interval. The disease itself, however, should not be clas sified as having a lengthy induction period; the induction period refers to the interval between the action of a specific component cause and the disease initiation. Because a component cause might act at any "stage" during carcin ogenesis, the length of the induction period for a specific component or carcinogen would de pend on the point during the causal sequence at which the carcinogen acted. Carcinogens act ing during a late stage (promoters) would have a short induction period, even if the complete causal chain involves a long sequence of stages requiring many years or decades to complete. From animal experiments, it appears that es trogens act as late-stage carcinogens; this find ing has been confirmed from epidemiologic research on replacement estrogen therapy and endometrial cancer, which indicates that es trogens have relatively brief (3 to 10 year) induction and latent periods before the occur rence of endometrial cancer (Jick et al, 1979). In addition, it has been found that the risk for endometrial cancer increases with the length of time estrogens have been used, but drops quick ly if estrogen therapy is discontinued, a pattern that also agrees with the results from animal studies of promoters. The Implications of Epidemiologic Observations One prediction from certain multi-stage mod els is that each of two carcinogens acting at different stages in the causal process will have effects that are, under some conditions, propor tional to the effect of the other; i.e., the effects are "multiplicative." Asbestos and smoking seem to interact in exactly this way in causing lung cancer (Saracci, 1977); the implication, given the appropriateness of the multi-stage model, is that asbestos and smoking act at different stages. Smoking and alcohol also show an interaction that may be multiplicative in causing cancer of the mouth and pharynx (Roth man and Keller, 1972). Some epidemiologic evidence is suggestive of a single cause affecting more than one stage of the carcinogenic process. One example, de scribed by Peto (1977), is smoking and lung cancer. Ex-smokers maintain a constant inci dence of lung cancer from the time of cessation, whereas continuing smokers show a steadily increasing incidence, According to Peto, this pattern suggests that smoking affects the penul timate stage of carcinogenesis. Other data, however, suggest that new smokers do not exhibit a change in incidence for many years, indicating an effect on an early stage. Of course, other interpretations might be given to these findings, but the implication that smoking af fects both an early and a late stage in the carcinogenic process for lung cancer is both interesting and plausible. Land and Norman (1978), analyzing the dis tribution of induction periods for various can cers after exposure to ionizing radiation, con cluded that, if a two-stage model is assumed, radiation affects only the first stage for lung and breast cancer (resulting in a typically long in duction period), but the second stage for chron ic granulocytic leukemia (resulting in a short induction period). CAUSAL INFERENCE The Process of Scientific Inference Having examined the theoretical basis for scientific inference and a conceptual model of causation that can be modified as needed to describe carcinogenesis, it is appropriate to consider in more specific and practical terms how scientific inference proceeds. The process of generalizing beyond a set of observations requires a judgment about what features of the observations may be extrapolated to a target zone (in epidemiologic studies, to a target pop ulation). Such judgments require an under standing of which conditions are relevant and which are irrelevant to the generalization. For example, from a study of smoking and lung cancer in men, one might generalize the results to a target population of women. To do so presumes that being male is irrelevant to the carcinogenic action that smoking has on lung tissue, a judgment based on knowledge about the likely mechanism of carcinogenesis and the biologic similarity between male and female lungs. On the other hand, a study of diet and myocardial infarction in men might not be considered generalizable to women, because wow iyn 20 -- BASIC CONCEPTS URL U4G55 physiologic differences between the sexes may play a role in the causal process. The validity of a generalization is ultimately a matter of in formed judgment. Some epidemiologists have taught that gener alization from a study group depends on the study group being a representative subgroup of the target population, in the sense of a sample (the confusion on this point is deeply rooted, as indicated by the use of the term sample-size to refer to the number of subjects in a study). Were this the case, generalization would be limited literally to those individuals who might have been included as study subjects. This miscon ception has influenced the design of many epidemiologic studies. If the notion were valid, there would be no application to humans of any results obtained from animal research. In addi tion, every population would require its own set of epidemiologic studies, and these studies would have to be repeated for every new gener ation! The tendency to use "representative" study groups probably derives from surveys for which the inferential process is purely statistical (de scriptive) and not scientific (synthetic). Social scientists often rely on statistical inference be cause the decision as to what is relevant and irrelevant for generalization is more difficult in the social sciences, and populations are consid erably more diverse in sociologic phenomena than in biologic phenomena. In the biologic sci ences, however, investigators conduct experi ments using animals with characteristics select ed to enhance the validity of the experimental work, rather than to represent the target popula tion. Epidemiologic study designs are usually stronger if the subject selection is guided by the need for making a valid comparison, which may call for severe restriction of admissible subjects to a narrow- range of characteristics, rather than a futile attempt to make the subjects representative, in a sampling sense, of the po tential target populations. The preceding discussion implies that to achieve the goal of valid generalization, a pre liminary requirement is to reach a valid inter pretation of the observations at hand. In the domain of nonexperimental research, the cen tral issue in attaining internal validity in a study is the question of confounding bias. Simply expressed, the problem is to separate the effect under study from the effects of extraneous fac tors associated with the study factor. In ex periments, confounding can be prevented by allocating the treatment (using methods such as randomization and blocking) so that there is no association between the treatment and the ex traneous factors. In nonexperimental research, the epidemiologist must rely on judicious sub ject selection and data analysis to prevent or remove confounding. Practical Causal Inference Though science does not admit conclusive inferences that one event causes another, in epidemiology the exigencies of public health problems require action despite imperfect knowledge. Consequently many people have attempted to codify the basis for making practi cal decisions about cause and effect. Among those who have examined the aspects of an observed association that lead us to believe it is causal is Sir Austin Bradford Hill 0 965). He suggested the following criteria: (1) Strength, (2) Consistency, (3) Specificity, (4) Temporality, (5) Biologic gradient, (6) Plausibility, (7) Coher ence, (8) Experimental evidence, and (9) Analo gy. Strength. By strength of association, Hill refers to the magnitude of the ratio of in cidences. His argument is essentially that strong associations are likely to be causal, because if they were due to confounding or some other bias, the biasing association would have to be even stronger, and would therefore presumably be more evident. Weak associations are more likely to be explained by biases, but the fact that an association is weak does not rule out a causal connection. It has been pointed out that the strength of an association is not a biologi cally consistent feature, but rather a characteris tic that depends on the relative prevalence of other causes (Rothman, 1976). Consistency. Consistency refers to the re peated observation of an association in different populations under different circumstances. Lack of consistency, however, does not rule out a causa! association, because some effects are produced by their causes only under unusual circumstances. Specificity. The criterion of specificity re quires that a cause should lead to a single, but not multiple, effects. This argument has often been advanced, especially by those seeking to exonerate smoking as a cause of lung cancer. Causes of a given effect, however, cannot be expected to be without other effects on any logical grounds (Sartweli, 1960). In fact, every day experience teaches us repeatedly that single events may have many effects. Temporality. Temporality refers to the ne cessity that the cause precede the effect in time. exrch, ubt or .ive in ilth ect ave ctiong an it is He (2) (5) erlo- -iill in>ng ? if ier be 3ly Dre lat a nat isof re?nt ck a ire jal re>ut en to er, be ny y* ;le teie. CAUSATION AND CAUSAL INFERENCE -- 21 This is an inviolable condition, which stems from the very concept of a cause. 8/0/09/c Gradient. Biologic gradient re fers to the presence of a dose-response curve. If the response is taken as an epidemiologic meas ure of effect, measured as a function of compar ative disease incidence, then this condition will ordinarily be met. Some causal associations, however, show no apparent trend of effect with dose; an example is the association between diethylstilbestrol and adenocarcinoma of the vagina. A possible explanation is that the doses of diethylstilbestrol administered were all suffi ciently great to produce the maximum effect, with actual development of disease depending on other component causes. Associations that do show a dose-response trend are not neces sarily causal; confounding can result in such a trend between a noncausal risk factor and dis ease. Plausibility. Plausibility refers to biologic plausibility, an important criterion, but one that may be difficult to judge. Sartwell 0960), em phasizing this point, cited the remarks of Cheever in 1861 (this statement was also noted by Hill): It could be no more ridiculous for the stranger who passed the night in the steerage of an emigrant ship to ascribe the typhus, which he there contracted, to the vermin with which bodies of the sick might be infested. An adequate cause, one reasonable in itself, must correct the coincidences of simple experi ence. Coherence. Taken from the Surgeon Gen eral's report on smoking and health (USDHEW, 1964), the term coherence implies that a cause and effect interpretation for an association does not conflict with what is known of the natural history and biology of the disease. The exam ples Hill gives for coherence, such as the histo pathologic effect of smoking on bronchial epi thelium (in reference to the association between smoking and lung cancer) or the difference in lung cancer incidence by sex, could reasonably be considered as examples of plausibility as well as coherence; the distinction appears to be fine. Hill emphasizes that the absence of coher ent information, as distinguished, apparently, from the presence of conflicting information, should not be taken as evidence against an association being considered as causal. Experimental Evidence. Such evidence is often telling but seldom available for human populations. Analogy. The criterion of analogy seems to be handicapped by the inventive imagination of scientists who can find analogies everywhere. Nevertheless, the simple analogies that Hill offers -- if one drug can cause birth defects, perhaps another one can also -- could conceiv ably enhance the credibility that an association is causal. As is evident, these nine criteria offered by Hill to judge whether an association is causal are saddled with reservations and exceptions; some may be wrong {specificity) or occasionally irrelevant {experimental evidence and perhaps analogy). Hill admits that None of my nine viewpoints can bring indisputable evidence for or against the cause-and-effect hypothe sis and none can be required as a sine qua non. In describing the inadequacy of these criteria, Hill goes too far. The fourth criterion, tem porality of an association, is a sine qua non: if the "cause" does not precede the effect, that indeed is indisputable evidence that the associ ation is not causal. Other than this one condi tion, which is part of the concept of causation, there are no reliable criteria for determining whether an association is causal. Nevertheless, Hill's list is a helpful guide to practical causai inference, if the criterion of specificity, which is clearly wrong, is deleted. No rule exists, howev er, to weigh the extent to which the evidence conforms to these criteria, nor is there any formal method by which one criterion can be weighed relative to the others. Ultimately, caus al inference is an informal, rather than a rigor ous, judgment. The failure of some to recognize the theoreti cal impossibility of "proving" the causal nature of an association has led to fruitless debates pitting skeptics who await such proof against scientists who are persuaded to make a practi cal inference on the basis of existing evidence. The responsibility that scientists have for mak ing causal judgments was Hill's (1965) final emphasis in his discussion of causation: All scientific work is incomplete -- whether it be observational or experimental. All scientific work is liable to be upset or modified by advancing knowl edge. That does not confer upon us a freedom to ignore the knowledge we already have, or to post pone the action that it appears to demand at a given time. References Carnap R; Logical Foundations erf Probability 2nd ad. University of Chicago Press. 1962. Cook PJ, Doll R. Fellingham SA: A mathematical model for the age distribution of cancer in man. Int I Cancer 4:92-112. 1969. URL 04056 22 -- BASIC CONCEPTS Mill AB: The environment and disease association or causanon' Proc Koval Soe Med 58.295-300. 1965. lick H, Watkins KN, Hunter |R. et a). Replacement estrogens and endometrial cancer. N Engl I Med 300:216-222, 1979. Knudson AG' Mutation and cancer: statistical study of retinoblas toma. Proc Nall Acad So USA 86:820-623, 1971Land Cl. Norman it: The latent periods of radiogenic cancers occurring among Japanese A-bomb survivors. In International Atomic Energy Agency Symposium on the Late Biological Effects Of lomzmg Radiation. Hiroshima. Radiation Effects Research Foundation; and Washington, D.C.. Medical Follow-up Agency, National Academy of Sciences. 1976. Mirvish SS: Kinetics of dimethyl mttosation n relation to nitrosamine carcinogenesis. | Natl Cancer Inst *4:633-639, 1970 Moolgavkar SH: The multistage theory of carcinogenesis and the age distribution of cancer in man. J Nail Cancer Inst 61:49-52, 1978 Mootgavkar SH. Venzon Dl Two-eveni models for carcinogene sis incidence curves for childhood and adult tumors. Math Bioscience 47-55-77, 1979. Oxford English Dictionary. Oxford University Press, 1971 Peto R: Epidemiology, multistage models, and short-term mutageni city tests. In Hiatt H et al tedsi. Origins of Human Cancer. New York. Cold Spring Harbor Publications, 1977. Popper KR: The Logic of Scientific Discovery. New York. Harper and Row. Inc., 1966. Reichenbach H. The Rise of Scientific Philosophy. Berkeley. Univer sity of California Press. 1951 Rothman K}: Causes Am | Epidemiol 104:SB?-S92.1976. Rothman KJ, Keller AZ The effect of foml exposure to alcohol and tobacco on risk of cancer of the mouth and pharynx. ) Chronic Dis 25:711-716,1972. Saracci R: Asbestos and lung cancer: an analysis of the epidemiolog ical evidence on the asbestos-smoking interaction. Ini J Cancer 20:323-331,1977. Sartwell P: On the methodology of investigations of etiologic factors in chronic diseases -- further comments. J Chronic Dis 11:61-63. I960 U.S. Dept, of Health, Education, and Welfare: Smoking and Health Report of the Advisory Committee to (he Surgeon General Washington. D.C.. Public Health Service Publication No 1103. 1964. URL 04057 URL 04058 CANCER AND THE ENVIRONMENT An Academic Review of the Environmental Determinants of Cancer Relevant to Prevention Edited by: H.B. Demopoulos, M.D. New York University Medical Center New York, New York M.A. Mehlman, Ph.D. American College of Toxicology/ Mobil Oil Corp. Princeton, New Jersey This Symposium was held in cooperation with the AMERICAN CANCER SOCIETY, INC. Symposium Committee: Dr. Harry B. Demopoulos, New York University Medical Center Dr. Benjamin Van Duuren, New York University Medical Center Dr. Vernon Young, Massachusetts Institute of Technology Dr. Marvin Kuschner, Dean, Health Sciences Center, State University of New York at Stony Brook Dr. John Higginson, International Agency for Research on Cancer, Lyons, France Dr. Joseph Cimino, President, New York Medical College; Former Commissioner of Health, New York City *!* URL INTRODUCTION: DETERMINANTS OF CANCER RELEVANT TO PREVENTION, IN THE WAR ON CANCER Joseph A. Cimino President, New York Medical College Harry B. Demopoulos New York University Medical Center We are gathered here today because we don't know all the questions, let alone the answers when it comes to environmental determinants of cancer relevant to prevention. By the end of this three-day international symposium, l hope we will have outlined one possible course and direction in this war on cancer. 1. We should be able to identify the cancer culprits that are amenable to successful intervention. 2. We must pay more attention to the factor of multiplicity and syner gistic causes. 3. We must be more concerned with nutrition's role in cancer pre vention. 4. We must once and for all put into their proper place the role that urban environments and occupational exposures play in the total cancer picture. 5. We must devote more energies and monies to research on life styles and know how that offers us one of our greatest prevention opportu nities. *- 6. We must correlate intervention with the most effective cost-benefit relationships and know how to maximize the number of lives that can be saved. Given a few minutes, most of you could easily describe numerous major cancer causing agents. As a group, I am sure that we could come up with the major culprits. I suggest that the day is here for us to say we've wasted too much time and money going after too many suspected culprits and our war on cancer is beginning to look a lot like another Viet Nam. Too much money. Too much manpower. Too much government intervention. Too much shooting in the dark. Too much waste. Too much time being devoted to snipers. The observation from where I sit is that we have wasted a decade by concentrating on viruses and hoping to develop a miracle cure with a miracle vaccine, while letting the big cancer killers get through ourfront lines. Granted, significant victories have been achieved in the medical arena: improved surgical techniques, better utilization of radiation, chemotherapy, the develop- 2 CIMtNO AND DEMOPOULOS ment of immunotherapy, and most important an aggressive and somewhat successful promotion by the American Cancer Society of the benefits of early detection. To be objective, with Proposition 13, national health care debates and soaring medical costs, we know that cancer research money will no longer flow as freely as before. Like Viet Nam it has been a long war and we have yet to come close to achieving victory, but we can't pull out of this one. The military objective of this conference should be, and must be, to marshall our forces and attack the major culprits, give them a priority rating and go after the big ones. In fighting the war, let us recognize the fact that it is counterproductive to use your major resources hunting down snipers. Let us spend our cancer war dollars where they will produce some significant casuaiities among the cancer culprits. A personal experience that vividly illustrates some of the problems we all face in waging this war on cancer concerns the control of air pollution in New York City. It is pertinent because it focuses on the people and forces that seem to get involved: environmentalists; city, state and federal regulatory agencies; private industry; physicians; scientists; the general public and the news media. During my tenure as Commissioner of Health of New York City, the big debate was raging over air pollution as a major cause of disease in urban America and a possible contributing factor to lung cancer. It escalated to a point where many Americans were convinced that by eliminating this "snip er,"a major health hazard would disappear. The marker chosen for control of air pollution was sulfur dioxide and although no major voice was blaming sulfur dioxide as a cancer causing agent directly, it was blamed for about everything else. At one point the citizens of New York probably believed that the City would be perfed if only S02 could be reduced. Hundredss of millions of dollars have been spent in New York City to burn low sulfur fuel. And I suggest to you that to this day I have yet to be convinced by any evidence that sulfur dioxide at levels reached in our urban setting is a major health hazard. Granted, on a clear day you can see the Twin Towers of the new World Trade Center but we went after S02 for the wrong reason. In my view the amount of money spent on this one battle may have contributed to New York City's financial plight and represents a colossal cost-benefit error. Because of the mandatory use of low sulfur fuel, Con Edison had to convert, and the bi II was paid by homeowners, renters, industry and taxpayers. Con Ed wasn't the only victim. City facilities, of course, had to be con verted as well as apartment house incinerators, which naturally raised rents. Whether it be the campaign for cleaner air or the fight to ban saccharin, we must stop wasting money going after suspected snipers and concentrate on the major determinants of cancer. The regulation of sulfur dioxide was one of the most expensive health bullets ever fired. The scenario on sulfur dioxide is typical of many of the same controversies over cancer causing agents and the disruptive debates now going on. First, a regulatory agency or environmental group raises a public issue about a suspected health hazard. This is followed by some quickly gathered URL 04060 AND DEMOPOULOS and somewhat benefits of early are debates and t will no longer d we Have yet to >ne. The military II our forces and *rthe big ones. Derproductive to our cancer war nongthe cancer oroblemswe all ollution in New orces that seem atory agencies; ie news media. rk City, the big sease in urban escalated to a ting this "snip- jr dioxide and r causing agent the citizens of mly S02 could New York City have yet to be d in our urban n see the Twin for die wrong Atle may have nte a colossal adison had to id taxpayers, to be conrents. :harin, we ate on the one of the Controversies {on. i public issue ckly gathered ENVIRONMENTAL DETERMINANTS Of CANCER 3 scientific studies -- and I use that terminology loosely -- which are quoted as Gospel. Next, an activist group clamors for immediate change. The news media highlights and sensationalizes, and then we have the crisis period. Conflicting reports. Confusing counter charges. Shoddy scientific data. And finally mandated change is forced upon us, or a product is withdrawn. I have not mentioned an important step that unfortunately is a typical flaw in our war on cancer. No one ever stops to make any real attempt to quantify the costs or effects of such a ban. It is time we apply a tourniquet to this kind of thinking and stem the flow of wasted effort and money. If the proper effort was expended in determining the cost benefit relation ship and if the best scientific data available were utilized, the conversion to low sulfur fuel in New York City would have never come about in the crisis oriented manner in which it did. There must be objective evaluation ofthe risks and benefits involved and a defense against outbursts of emotionalism. More emphasis must be placed on rational decision-making concerning the prevention of cancer. There is little doubt that Washington continues to give the war on cancer a stamp of highest priority. However, what is most disturbing is the accusatory climate this priority has taken. Much controversy swirls around maximum tolerated levels or the pinpointing of thresholds. Things get so confusing that the EPA, OSHA, NIOSH, FDA and CPSC aren't too sure of what the other is doing. In any war, be it Viet Nam or cancer, the allies should have some idea of what each other is doing. It may be that most of the foods we eat, the chemicals we use, the materials we manufacture will all be illegal by current or proposed Government regula tions. And perhaps that's not such a bad idea, for then everyone in Washington will have to return to reality, change the ground rules and step back and take a good long took at costs and effects. My remarks are not anti-Washington or anti-Govemment. Quite the con trary. We in science and education must be more aggressive in communicating with federal agencies. All too often we seem to sit back and react, usually negatively, to Government findings and regulations instead of being involved before such findings are made public policy. I maintain that in the past many of us went along with Government to achieve our own personal and professional goals without asking a word about cost-benefit relationships. We too are to blame for some of the past sins as we strived to keep research monies flowing into our laboratories and institutions. It is time for us professionals to make the commitment that the war on cancer warrants. The causes of cancer must be put into perspective. Normal life would not be possible if we attempted to eliminate all agents that had the potential for harm, such as Vitamins A and D, essential elements like copper and iron, sodium chloride, etc. The thought that the potentially harmful chemicals that occur naturally in food and in the air we breathe (oxygen can be toxic) are somewhat safe, while comparatively low levels of food additives or common air pollutants are dangerous, is naive. URL 04062 4 CIMINO AND DCMOPOULOS Let us not perpetuate the tactic of going after the snipers, but rather concentrate efforts on finding out how to intervene in the major lethal cancers. Let's move viruses to a side burner and concentrate on prevention. As a physician and educator I am deeply concerned that the funds for the war on cancer continue to short-change those projects that deal with cancer preven tion. We should agree that there is no absolute safety in human life and deal with the unreasonable risks to human health. Let's make a commitment that we can prevent cancer via better directed research. Let's refrain from the ir responsible misuse of cancer statistics, and agree that we have wasted millions and millions of dollars and man years in past efforts and concentrate on those areas, ^uch as life style and diet, that can help prevent cancer now. Let's make a more objective evaluation of the risks and benefits involved and relate our efforts to the cost/benefit concept. Let's start working with, rather than against, governmental agencies. Finally, to repeat my major point, let's give prevention the major status it so richly deserves. It is not necessary to know the exact cause and pathogenesis of a disease in order to prevent its occurrenceor progression. By intervening at some point in the development of a disease, at the host, vector, or agent level, we may alter its course. Examples in the history of public health are numerous: quarantines; sanitary water supplies; protection against food spoilage; the first vaccines, as against smallpox; pest and vermin control; swamp drainage; incest taboos; vitamin supplements; hypertension control. As a result of the efforts of numerous basic scientists, clinicians and epidemiologists, we have begun to understand the major causes of cancer. That most human cancers are caused predominantly by environmental factors is hardly disputed in present times. For the purposes of prevention, which in the long run is our most effective weapon in the War on Cancer, it is of central importance to describe what is meant by the term, environmental, since it is this term which suggests that most human cancers might be prevented or markedly delayed. An appropriate description perhaps is that given by Dr. John Higginson who helped to popularize the concept of "Environmental Cancer" starting in the 1950's: "...when I used the term environment in those days, I was considering the total environment, cultural as well as chemi cal. By cultural, I meant mode of life...I've checked it in every dictionary and every dictionary gives the same: Environment is what surrounds people and impinges on them. The air you breathe, the culture you live in, the agricultural habits of your community, the social-cultural habits, the social pressures, the physical chemicals with which you come in contact, the diet, and so on. A lot of confusion has arisen in later days because most people have not gone back to the early literature, but have used the word environment purely to mean chemicals" (Sci ence, 1979). Aside from describing what academic scientists mean when referring to the term, "Environmental Cancer", it is relevant for the purposes of launching some prevention efforts now, to identify specific predominant causes of human cancers, if at ad possible. Many situations may be associated with an increased 1NO AND DEMOROUIOS snipers, but rather najor lethal cancers, n prevention. As a jnds for the war on vith cancer preven>uman life and deal ommitment that we efrain from the irjve wasted millions xicentrate on those er now. Let's make Ived and relate our rather than against, et's give prevention iow the exact cause nee or progression, isease, at the host, he history of public protection against nd vermin control; ftension control. *ts, clinicians and uses of cancer, by environmental rs of prevention, ron Cancer, it is of nronmenta), since it be prevented or : given by Dr. John omental Cancer" iys, I was as chemiit in every onment is t air you feof your Kie diet, Because have Jhen referring to Rs of launching auses of human ith an increased ENVIRONMENTAL DETERMINANTS Of CANCER s cancer risk, but a comprehensive listing of all of these is probably not compati ble with prevention because the length of such a compendium of contributing circumstances would be too long. The concept of predominant causes or factors, on the other hand, is useful for the purpose of initiating reasonable prevention efforts at the present time. The definition of a predominant cause is that in the absence of such a factor, the disease would not occur, or would be markedly less frequent. An example is smoking high-tar cigarettes and lung cancer, in both sexes, and all races. The analogy of tuberculosis and its predominant cause is also helpful in delineating the concept. In this disease, the predominant factor is the tubercle bacillus and the size of the infectious innoculum. Associated factors are crowding, malnutrition, lower socio-econ omic status, predisposition by diabetes mellitus or steroid therapy, inclement weather, and so on. The control of tuberculosis is most effective when aim is taken at the tubercle bacillus and the size of the innoculum; the associated societal factors must be alleviated for their own humanitarian reasons. Any academic contributions that are made for the purposes of cancer prevention among the public must be clear cut so as not to confound individu als into apathy and inaction. Prevention messages, carried out on a massive scale, must be relatively free of confusion and therefore, the concept of specific, predominant causes has value since the potential list may be shorter than a compendium of all of the possible contributing circumstances. The term, predominant causes, has been conceptually advanced, despite the fact that cancer is a multifactorial set of diseases. It is necessary, in order to have a high likelihood of developing cancer, to have multiple factors and circumstances operating conjointly; however, avoiding, or greatly decreasing the risk of developing cancer may only require the control of one or two predominant factors out of the whole constellation of contributing causes and circumstances. Some of the predominant causes do fit together, such as excess use of high-tar cigarettes and heavy alcohol consumption; in such cases it is reason able to refer to them as one set of predominant factors, especially since the two are very often used at the same time. Similarly, dietary excesses appear to constitute a constellation of sub-factors that are in co-use, and therefore comprise a second set of predominant factors that could be termed, disordered nutrition. Described in this manner, such sets of predominant factors would generally not have substantial cross-over in terms of the major types of cancers caused, and can therefore be considered, to a large extent, as separate from one another. For example, smoking high-tar cigarettes with concomitant heavy alcohol consumption constitutes one predominant cause responsible for ap proximately 35% of all cancer deaths in the United States (Hammond and Seidman, 1980), the cancers thought to be induced include nearly 90% of epithelial primaries in the lunesj-noufh, larynx.esophagus, and liver, as wefl as approximately 4U%ofurinary bladder cancers. Disordered nutrition, which is 'comprised of several factors including a low-fibre diet with a high-fat intake, excess calories, plus obesity, is another predominant cause which is thought by some to account for nearly 45% of all cancer deaths ki the U.5.A. (Hammond and Seidman, 1980; Upton, 1979; Nutrition and Cancer, 1977); these cancers URL 04064 6 CIMINO AND DEMOPOULOS include between 80-90% of the epithelial primaries in the colon, rectum, breast, ovaries, and endometrium. These two categories: (1) high-tar cigarettes/ alcoholism and (2) disordered nutrition may have some validity as separate, predominant causes since the types of cancers found are different. While most individuals who smoke high-tar cigarettes and drink heavily have poor nutri tion, especially nutritional deficiences, their dietary patterns are not generally associated with excess calories and obesity. Most heavy smokers/drinkers have poor appetities and tend to be thin. Many of these individuals undergo major weight gains and uncontrolled appetites upon cessation of smoking and will use this to reinforce the smoking habit. The two types of lifestyles: (1) smoking and drinking to excess, and {2) eating to excess with a predominant high-fat, low-fibre diet, seem to be somewhat segregated and are associated with different types of cancers. This is undoubtedly an over-simplification, espe cially since heavy smokers/drinkers tend to be malnourished, and therefore also have "disordered nutrition"; however, this is of a different type from the disordered nutrition that is linked to the causation of cancers of colon, rectum, breast, ovaries, and endometrium. The area of nutritional deficiences is probably applicable to both types of predominant causes discussed so far, i.e., (1) excess smoking/drinking and (2) low-fibre diet with excess calories/fat ingestion/obesity. Deficiences of vitamin A, other retinoids, antioxidants, selenium etc. can occur as a result of poor appetites, as occurs in heavy smokers/drinkers, or as a result of consuming excess quantities of only a few types of foods. Obese individuals can and often do have nutritional deficiences, especially of antioxidants. At a symposium sponsored by the American Health Foundation, in 1979, a group of medical scientists concurred that the dramatic decline in stomach cancers, from the major lethal cancer to its present comparative rareness, is due largely to the greater availability of low-cost, fresh produce, made available by an efficient infrastructure of rapid trucking and refrigerated freight carriers; the net result being the greater consumption of more varied, fresh produce with, presum ably, enhanced intake of different trace nutrients and antioxidants (Preventive Medicine, Vol. 9,1980). The value of attempting to segregate the predominant causes, and the types of related cancers, centers on the effectiveness of prevention-education and, to some extent, targeted early detection programs. Public health messages must be clear and specific. If predominant causes can be identified, in general, with specific types of major, lethal cancers, then public education for preven tion and targeted early detection will be easier. The overall purpose in trying to simplify the predominant causes of cancer is to inititate prevention education now. As further research is performed to verify the major causes of cancer, new and more appropriate public health education can be developed. Several symposia (Nutrition and Cancer, 1977; Preventive Medicine, Vol. 9, 1980) and individuals like J. Higginson (Science, 1979) and A. Upton (Upton, 1979) have developed relatively short lists of specific predominant causes, quite independently of one another. The concordance among these observers is good; Table 1 reproduces these with a relative percent contribu- tND DEMOPOULOS colon, rectum, h-tar cigarettes/ iity as separate, ent. While most lave poor nutrire not generally *s/drinkers have undergo major loking and will es: (1) smoking ninant high-fat, ssociated with ification, espe- and therefore t type from the colon, rectum, o both types of linking and (2) nces of vitamin result of poor of consuming * can and often a symposium up of medical cers, from the largely to the by an efficient the net result with, presum es (Preventive uses, and the on-education jth messages Kin general, for prevenB in trying to n education Jf cancer, new tedicine, Vol. fd A. Upton predominant among these ent contribu- ENVIRONMENTAL DETERMINANTS OF CANCER 7 TABLE 1. Predominant Sets of Factors in Developing Cancer** Estimated Relative Percentage 35%ofcancerdeathsarepredominantlydueto smoking high-tar cigarettes and consuming excess amounts of distilled liguor (this in cludes 90% of the cancers of the lungs, mouth, larynx, esophagus and liver, plus 40% of cancers'of the urinary bladder) 45% are predominantly caused by disordered nutrition, with the following subcategories: a) excess calories b) excess fat ingestion, including saturated and unsaturated fats, as well as cholesterol c) obesity, of a magnitude of 40 lbs. or more overweight, for the average individual d) nutritional deficiencies, especially dietary fiber and retinoids (the latter includes vitamin A deficiency: however, only the recommended daily allowance is needed; excesses are extremely dangerous) (the cancers that relate to disordered nutrition are most of the cancers of the colon, rectum, stomach, breast, and many of the cancers of the ovaries and endometrium) 5% or less of the present deaths are due pre dominantly to occupational exposures that occurred in the past when the dangers were not known (the cancers that are generally included represent a small proportion of the malignancies of the urinary bladder, lung, nasopharynx, stomach, hemoiymphatic system, liver, bones and skin) 3% are predominantly due to ionizing radiation; about Vi of these are due to "background", or naturally occurring radiation, and Vs to medical radiation 2 % are due predominantly to pre-existing, benign medical disorders such as chronic fibrocystic disease of the breast, ulcerative colitis, regional enteritis, and chronic atrophic gastritis Approximate Number of Such Cancer Deaths* 3 50 cancer deaths per day 450 cancer deaths per day 50 cancer deaths per day 30 cancer deaths per day 20 cancer deaths per day 1% art due predominantly to the administration of prescribed pharmaceutical agents such as cancer chemotherapeutics, estrogens for the menopause, and possibly some anti-hypertensive drugs 10 cancer deaths per day 'Approximately 1.000 individuals in the United Statesdieevery day from cancer, i.e., about 365,000/year. "(Editor's Inveiview, 1979: Hammond and Seidman, 1960; Upton, 1979; Wynder, 1980). U R L 04065 I CIMINO AND DEMOPOULOS URL 04066 tion, in a highly tentative manner. These percentages represent working figures and are only estimates to be probed and tested through further research. Since these sets of predominant causes in Table 1 are largely segregated, it is valid to add them separately to develop an idea of approximately how much of the cancer burden might be preventable, or put off to a later time in life. It is fair to say that the preponderant causes of most of the lethal cancers in the U.S.A. are known, and present varying degrees of difficulty in terms of public health education. A comparatively small relative percent of cancers have unknown predominant causes, about 10%. While public health education in cancer prevention requires clarity and simplification, academic standards require lull and more complex considera tions. Therefore, tabular estimates of the relative percent contribution by various predominant causes have limited academic value and must be mod ulated by some scientific input in order to avoid mistaken simplifications. For example, while low-fibre diets are thought to be a major factor in the develop ment of colon/rectal cancer, the simplistic corollary that indiscriminate high- fibre intakes would offer prevention requires an answer to the question of "...what type of fiber?". It is entirely possible to experimentally enhance mucosal contact with carcinogens adsorbed on certain types of fibres. Rather than focusing on one aspect of disordered nutrition, it would be perhaps more relevant to realize that low-fibre diets are generally also high fat and lack sufficient fresh produce. The fibre content relates to enteric transit time, and possible also to the adsorptive capacity of the fibre. Transit time may be important since it may affect the duration of exposure of the mucosal epi thelium to enteric carcinogens and "promoters"; however, copious fluid intake, high carbohydrate diets and physical activity can also decrease enteric transit time. Therefore, it is possible to see how lifestyle and diet interface and why the term disordered nutrition is a constellation of factors that includes obesity, with its presumed relative inactivity, together with excess calories, high fat intake and low fibre. Against this type of background, a number of academicians decided that the complexities involved in the development of the major lethal cancers had to be presented to the relevant segments of society, and to the public at large through the media. The Cancer Symposium was under the co-sponsorship of the New York Academy of Sciences and the Cancer Symposium Committee, and was held in cooperation with die American Cancer Society. The Speakers and Chairpersons were selected by die Cancer Symposium Committee and reviewed by the New York Academy of Sciences and the American Cancer Society. These Speakers and Chairpersons reported on work that was sup ported by peer-reviewed grants/contracts from the N.S.F., W.H.O., the American Cancer Society, the National Research Council, and granting sources in several countries. The Cancer Symposium was organized into sequential sessions that at tempted to answer die following questions: How do chemical/physical agents in the personal and general environment cause normal cells to become cancer cells? Are there clues in existing human data that can direct the basic > DEMOPOUIOS rking figures sarch. egregated, it / how much e in life. It is ncers in the ns of public incers have clarity and < consideraribution by jst be modcations. For ne developlinate highquestion of ly enhance >res. Rather rhaps more it and lack t time, and ie may be jcosal epiaious fluid ase enteric efface and it includes s calories, *cided that incers had lie at large isorship of ommittee, ? Speakers nittee and in Cancer was sup.H.O., the 1 granting ns that at- d general foe basic ENVIRONMENTAL DCTBLMINANTS OF CANCER 9 researchers to the causes ofcancer? For example, are there groups of people who follow particular lifestyles that either predispose, or protect them? Are there areas of our country where cancer rates are particularly high, and if so are there consistent factors when scien tific controls are studied? What are the causes of the lethal cancers, as shown by human epidemiology and animal experiments? What are the contributions of high-tar versus low-tar cigarettes, excessive alcohol consumption, general pollution, disordered nu trition, nitrites/nitrates, saccharin/cyclamate, administration of pharmaceuticalsubstances? Are there ways to scientifically extrapolate data, since there is a real need to protect workers, communities and consumers right now? Are there systems available now that can be logically used to help determine risk? If exposures have occured, is it possible to intercede during the well-known lag phase (between exposures and development of cancer) with chemicals that might act as "antidotes"? For example, can anything be done to prevent the progression of asbestos-related diseases among the individuals who were exposed 15-30 years ago, at a time when the dangers of heavy exposures were not clear? The presentations did provide significant portions of foe answers to the above questions and additionally offered data on foe following points which are discussed and summarized following the Conference Conclusions; these include foe questions: Is there a cancer epidemic involving most anatomic sites, or is it just lung cancer that is epidemic? Is there a reasonable biologic basis for practical thresholds to many of the predominant causes of cancer? Is synergism a significant factor, since most humans have multiple exposures? Is there a way of approching this question on a practical basis? is it likely that industrial and general pollution of air and water will , lead to major increases in cancer rates (per 100,000) in foe future, in view of increases in industrial production and in population? What are foe percent relative contributions of widely publicized factors such as general community air pollution, general water pollution and saccharin/cyclamates? ACKNOWLEDGEMENTS ft is imperative to understand that foe advances in the War on Cancer have come about by foe concerted efforts of basic scientists, clinicians, biostatisti cians, epidemiologists, lay lobbyists, legislators, and administrators in many parts of the world. It is hoped that their efforts will continue, particularly ft CnMfri Tnxiiul Vol. 19. pp $39 lo $48 1981 Primed in Greet Britain All righit reserved 001$-42*4/81 /05>V-IOt02.00/tl Copyright O 1981 Pergemon Press Ltd RETHINKING THE ENVIRONMENTAL CAUSATION OF HUMAN CANCER J. Higginson InternationaI Agency for Research on Cancer, till cours Albert Thomas, 69372 Lyon Cedex 2, France {Received IV December IVHO) Summary--Many of (he hypotheses regarding the aetiology of cancer in man were suggested by the early 1950s and much recent effort has been directed to defining causes and mechanisms in more detail through more sophisticated epidemiological and laboratory approaches. Although there are many factors involved in its aetiology cancer may be prevented by controlling a single predominant factor. Epidemiology has contributed considerably to our understanding not only of the aetiologies! factors but also of the possible mechanisms involved, especially those associated with lifestyle. For cancers believed to be related lo the latter, notably those of the gastro-intestinal tract and endocrine-dependeni organs, there is evidence for the involvement erf cultural, behavioural and dietary factors rather than direct exposures to carcinogens alone. Future efforts should be directed towards a better understanding of factors modulating carcinogenesis since it seems improbable at present that the initiating factors for many cancers can be identified. Introduction It is a great pleasure to contribute to this issue dedicated to Dr Leon Golberg. I have known Leon Golbcrg for many years, and admire not only his 'scientific work but also his erudite and objective studies on the complex interactions between toxi cology and human health and welfare. Present concepts of the aetiology of human cancer have developed gradually, reflecting new ideas and hy potheses arising from both epidemiological and labor atory studies. Thus, the isolation of pure carcinogenic polycyclic aromatic hydrocarbons was essentially dependent on the original observations of Pertivall Pott. The later demonstration of other occupational hazards provided the foundations of modem chemical carcinogenesis. The recognition of oncogenic animal viruses and the development of pure strains of mice as our understanding or inherited susceptibility has in creased has had considerable impact on aetiological theories. Willis (1948). while drawing attention to the possible environmental background of many human cancers, largely emphasized industrial chemical car cinogens. It is only comparatively recently that the role of environmental factors in human cancer in a wider sense has been accepted. Moreover it appears that progress in environmental carcinogenesis has been as much dependent on the confirmation or ex clusion of hypotheses through more sophisticated investigations as on the promulgation of completely new theories. In 1950, a group of distinguished experimentalists and laboratory workers meeting in Oxford (Clemmesen. 1950> concluded that geographical variations in cancer incidence predominantly reflected environ mental influences. In that term, however, they in cluded all exogenous factors which impinge on man. i.e. the dietary, social and cultural environments as well as discrete chemical carcinogens. Stimulated by the Oxford report. George Octtle and myseir decided to collect accurate statistics in Southern Africa to pro vide a baseline of canoer incidence in a newly urba nized population in an attempt to evaluate the rela tive aetiological importance of lifestyle, diet and industrialization. The results of this survey (Higginson & Oettle. 1960f in combination with the migrant studies of Kennaway (1944), Haenszel (1961) and Haen&zei & Kurihara (1968), and the reports of others (Boyland, 1969; Doll 1967), strongly suggested that most human cancers were related to environmental factors and not to racial or hereditary factors (Higgin son, 1960). Further, the survey suggested that "way of life" was of major significance for many cancers, es pecially those of the gastro-intestinal tract and endo crine-dependent organs. These estimates were later extended using a wider range of populations (Higgin son, 1969). Carcinogenesis in man Definition of cause The massive increase in the production and use of synthetic chemicals in industry, agriculture and medi cine since 1940 caused renewed interest in the poten tial carcinogenic effects of chemical pollutants in the human environment (Carson, 1962). Since the concept of a discrete carcinogenic chemical, physical agent or virus as a direct cause of cancer was widely compre hended. the possibilities for prevention through simple legislative action were readily accepted. In contrast, there was a tendency to neglect the inherent complexities of chemical carcinogenesis, although such complexities and the many associated modulat ing factors involved had already been demonstrated by many workers (Berenblum, 1978; Clemmesen. 1950; Miller. E. C,, 1978). Moreover, for many cancers in humans, the data were inconsistent with an aeti ology based only on simple chemical exposures (Hig ginson & Oeitle. 1960). In addition, an increasing number of parameters related to life-style were ident ified and associated with increases or decreases in A 540 J. Higginson cancer incidence. Such parameters, now called "card- ) ]t should be emphasized that no epidemiology nor nogenic risk factors", included absence of fibre in the C laboratory study, whether relating to discrete carcino- diet, age at first marriage, etc., and could not readily f gens or lifestyle factors, can now be completely con- be defined as carcinogens in the classical sense. Both J elusive since all results (positive or negative) are laboratory and epidemiological studies, however, sug- dependent on statistical probabilities (IARC Working gested that the role of such factors including their Group, 1979a: Vesell, 1980). possible relationship to individual susceptibility should be explicable in more objective biological terms through a-betier understanding of the multi Hypotheses on the causation of human cancer stage and molecular basis of carcinogenesis Prior to 1950 the cause of very few human cancers The multifactorial nature of carcinogenesis tends had been firmly established, but considerable data are theoretically to complicate the definition of cause, now available. Most cancers can be described as since, in addition to the role of complex modulating tumours of well-defined environmental origin, or factors, combinations of carcinogens may show syner those for which the environmental background can gistic (Hammond Sc Selikoff. 1973) or inhibitory only be deduced. For a smaller group no satisfactory effects (Miller. Miller. Brown Sc MacDonald. 1958). In causal hypotheses are available. Estimates of the pro practice, however, a single factor may be so predomi portion of cancers in each aetiological group in differ nant as to be regarded as the practical cause in the ent communities are reported elsewhere (Higginson, public health sense, in that in its absence a significant Sc Muir, 1979). proportion of related cancers would not arise. This in no way excludes a role for modulating factors (Cole & Cowers caused by defined exogenous factors Merletti, 1980) or necessarily implies understanding of The majority of tumours in this group are epithelial the basic mechanisms involved. Thus, although the cancers of the skin, respiratory and upper digestive action of cigarette smoking may be multiplied many systems, liver and bladder. In addition, the causes of a times by asbestos exposure, the former is the practical small number of cancers of the endocrine-dependent cause of 85% of lung cancers in males whether or not organs, haemopoietic system, bone and soft tissues exposed to asbestos (Hammond & Seidman. 1980; have also been identified. Hammond & Selikoff. 1973). In contrast, asbestos Personal habits are by far the most important would be the predominant cause of mesothelioma in-^^timuli established, especially cigarette smoking, exposed shipyard workers. Although oestrogens are r'Studies in all countries emphasize the overwhelming believed to be promoters, they can be regarded as the , role of this habit which causes between 25-35% of all practical cause or a significant proportion of endo- cancers in males in North America. Europe and methai cancers in parts of the United States (Gus^" ' Japan (Hammond & Seidman. 1980). The proportion berg. 1980). Thus, in considering aetiology, it is im-* in females is small but increasing rapidly. Cigarette portant not to confuse multifactorial mechanisms. i.e..' smoking is not only carcinogenic per sc but also relative role of initiator versus promoter in cigarettes^' enhances multiplicatively the effects of such other individual susceptibility, etc., with/practical causy' es- /factors as asbestos and alcohol (Hammond & Selikoff. pecially since the latter may have more immediate 1973; Tuyns. 1978). In most countries this habit is health relevance. Routine epidemiology and long predominantly responsible for the increases in cancer term experimental testing essentially demonstrate the incidence reported since 1950. Excess alcohol con overall impact of numerous events, and usually the sumption, sunbathing and in Asia betel-quid chewing effect of individual modulating factors cannot be esti are other carcinogenic habits. A much smaller part of mated. the cancer burden even in industrialized countries is Role of epidemiology in evaluating aetiological h\-potheses related to occupational (probably teyi than and iatrogenic and radiation exposures. It is estimated that approximately 39 of the 537 chemicals and 1/ While defined aetiological hypotheses in human related industrial processes evaluated in the IARC cancer have been largely developed through case-his Monographs Series are probable or definite human tory epidemiological studies, it is less widely appreci carcinogens, of which the most important of indus ated that epidemiology may also contribute to the trial origin is asbestos, especially in association with analysis of possible mechanisms for those environ cigarette smoking However, the list is not definitive mental cancers of uncertain aetiology, through deduc and further iatrogenic and occupational hazards will tions made from geographical and temporal vari probably be detected in the future. ations in incidence and associated migrant studies Although less well-documented, there is consider- | (Clemmesen. 1950; Doll. 1967; Higginson. 1969; Hig able evidence in Africa and Asia that primary liver 'y, ginson & Muir. 1979; Higginson Sc Oetile. 1960; cancer arises in hepatitis B virus carriers who are ( Wynder & Gori, 1977). In this context the study of exposed to aflatoxin (Larouzi, Blumberg London, \ low-risk populations is particularly important in Lustbader, Sankale & Payet, <4977; Unsell Sc Peers, ) allowing the evaluation of those environmental 1977). '-S factors especially related to lifestyle (Enstrom, 1980; Higginson Sc Oettli, I960; Lyon, Gardner Sc West, Cancers of probable environmental origin 1980; Phillips, Kuzma Sc Lou. 1980). Epidemiology is This group comprises tumours of the gastrointesti of course complemented by laboratory studies which nal tract, e.g. stomach and large intestine, and of the may not only suggest new aetiological hypotheses for endocrine-dependent organs, e.g. breast, body and testing in man but also the nature of the mechanisms cervix of the uterus, and ovary, etc. Although the defi involved. nitive stimuli have not been identified, the most URL 04069 iogy nor carcino- e)y conive) are forking cancers data are ibed as igin. or jnd can sfaciory the pron differgginson. oithelia! digestive ses of a indent tissues pouant noting, telming 0 of all pe and portion igarette Jt also other elikoff, abit is cancer >1 conhewing part of tries is o) and (mated s and IARC iX -luman indusn with inhive is will isiderliver 10 are ndon. Peers, ' t ntesti- of the and : defi* most Environmental cancer causation 541 rational interpretation of the available epidemiologi cal data would indicate a direct or indirect associ ation with environmental factors (Higginson & Muir, 1976). While it is probable that further strong exogen ous agents will eventually be identified for other tumours, e.g. oesophagus in parts of China, Africa and Iran, where precancerous mucosal lesions are found in most adults (Crespi, Munoz, Grassi, Aramesh, Amiri, Mojtabai & Casale, 1979), this would appear less likely for most cancers in this group and alternative aetiologies! explanations must be sought. On one hand, it has been suggested that such cancers largely reflect additive exposures to multiple mutagens or carcinogens within the general environ ment predominantly of industrial origin (Epstejn, 1978). Others believe that most of these cancers are predominantly related to exogenous and/or endogen ous factors inherent in life-styje modulating cells already initiated by undetermined stimuli, and are not dependent on exogenous initiators alone (HigginsoIT & Oettle, 1960; Wynder & Gori, 1977), as suggested by much recent experimental work (Slaga, 19%0y The role of ambieni environmental pollution The effects of ambient environmental pollution on several diseases, especially respiratory, are recognized. The fact that exposure to high doses of certain chemi cals causes cancer in man and the synergism existing be'ween several toxic agents have led to concern about the significance of multiple low exposures to chemi cals in the general environment. Chemicals of natural and synthetic origin with biological activity have been and are ubiquitous in the environment in all coun tries. They include carcinogens and mutagens, pro moters. enhancers, etc. Polycyclic aromatic hydro carbons, nitrosamines. mycotoxins, flavonoids and other suspected carcinogens have been demonstrated in air, water (Kraybill, 1976; Wilkins, Reiches & Kruse, 1979), food, and alcoholic beverages (Rose, 1977). Many individuals have been and are exposed to many of the animal carcinogens reviewed in the IARC Monographs Series (Supplement 1; IARC Working Group 1979a). The problem of their evaluation at very low levels of exposure is illustrated by the nitro samines which are not only widely distributed in the environment, but may be demonstrated in body tissues and fluids (Walker, Castegnaro, Griciule & Lyle. 1978; Yamamoto. Yamada & Tanimura, 1980). While many are carcinogenic to several animal spe cies, to dale there is no Arm evidence of carcinogenic activity in man (Tannenbaum & Young,_1980)^ Attempts to evaluate the specific carcinogenic potential to man of an individual chemical at low dose among the myriad chemicals present in the en vironment poses almost insoluble logistic and techni cal problems for the epidemiologist, especially in the presence of powerful confounding variables such a$y cigarettes. The difficulties are even greater than those of determining the effects of low doses of ionizing radiation, a recognized human and animal carcinogen where controversy still exists (Land, 1980), although in this case the "target cell" dose can be estimated. ^ The value of animal experiments in identifying potential human carcinogens is well recognized, but there is a tendency to concentrate on the statistical limitations of epidemiological methods rather than the equally great limitations of biological extrapola tion between species. Such difficulties cannot be over come simply by theoretical mathematical models ex trapolating from animals since many of the pertinent parameters relating to carcinogen metabolism, inhi bition, etc., cannot be measured (Coulston, 1979k nor can the dose at the target cell, although newer tech nology, i.e. measurement of DNA adducts, may per mit this possibility in the future (Rajewsky, 1979). Within limits, however, epidemiology, by compar ing cancer patterns in different environments, may help to evaluate the additional impact of the overall burden of chemicals in a specific environment, i.e. the sum of total increased risks. Such an approach has been attempted (Higginson, 1979; Lyon er al. 1980; Royal Society Study Group, 1978; Wynder & Gori. 1977) and further details may be obtained from these reports. In brief, no consistent relationships have been observed between total cancer patterns or between . organ sites, and indices of probable ambient environmental pollution such as industrialization and urban ization. The most intensive studies have been in relation to air pollution, where no significant effect can be demonstrated if correction is made for such variables as cigarette smoking and occupational ex-!' posures (Cederlof, Doll, Fowler. Friberg. Nelson & Vouk, 1978; Goldsmith, 1980; Haenszel. Loveland & Sirken. 1962; Haenszel & Taeuber, 1964; Hammond & Garfinkel, 1980). Such studies do . however, permit partial evaluation of the additive effects of occupa tional exposures in both urban and non-urban en vironments (Hammond &. Garfinkel, 1980). The'; above observations are supported in a recent report indicating no significant differences in cancer inci- / dence between urban and rural Mormons (Lyon et al. I 1980). Thus it appears that only a very small part of-' the total cancer burden can be directly related to industrialization in a general sense, and alternative aetiological explanations must be considered for the majority of tumours of environmental origin. These observations also have obvious relevance to evaluating the existence of `no-effect' exposure levels for identified carcinogens or reversibility in carcinoge nesis. Lifestyle In addition to such dearly defined habits as cigar ette smoking, alcohol ingestion, sunbathing and occu pation, this term covers the total cultural, behavioural and dietary environment i.e. all exogenous factors inherent in daily life. Lifestyle includes such carcino genic risk factors as behavioural patterns, e.g. age at first marriage and pregnancy; such physiological par ameters as age at menarche or menopause; and such dietary patterns as quantity and quality of dietary fat, fibre etc. While the recognition of lifestyle' factors in cancer is not recent (CJemmesen. 1950; Higginson & Pettie. I960; Wffiiv 1948; Wynder & Gori. 1977). it has been neglected due to its poorly defined nature. Diet should be considered as a highly complex chemical mixture. It not only includes preformed car cinogens and carcinogen precursors but also enhanc ing and inhibiting factors (Wallenberg, 1979; Wynder, Hoffmann, McCoy, Cohen & Reddy, 1978). Further, while diet might be directly related to stomach or i'~' r~ 9 o ffl 19 S-- URL 04071 $42 J. HlCHilNSON large intestine cancer through exogenous carcinogens or endogenous carcinogen formation e.g. niirosamines (Tannenbaum Sl Young, 1980), it may also affect tumour development of the breast for example in a non-specific manner through variations in fat and calorie intake (Miller, 1980; Newbeme Sl McConnell, 1980; Wynder Sl Gori, 1977). Further, diet is a determinant in menarche. height, obesity, etc., risk factors for cancers of the breast and endometrium (Hill, Wynder. Barbaczewski, Heiman, Hill. Sporan- gisa Sl Husk'isson, 1980). The nature of such nonspecific effects is uncertain but they may operate through the hypophyseal axis. Moreover, numerous studies confirm the role of diet in modulating the induction of enzymes of potential significance in car cinogenesis (Cl BA Foundation Symposium 76, 1980; Conney. Pantuck, Pantuck. Buening, Jerina. Fortner. Alvares, Anderson & Kappas. 1978; Kaiamegham, Krishnawamy. Krishnamurthy & Bhargava. 1979; Newbeme & McConnell, 1980; Vesell, 1980). Calorie and protein reduction have been associated with inhi bition of tumours of the breast, intestine and skin in animals and a high-fat diet- has been associated with increased cancer in humans and animals possibly due to late-stage effects (Gori, 1978; Hirayama. 1979; Hoehn & Carroll, 1979; Newbeme. Weigert & Kula, 1979) . The type of carbohydrate has been shown to be important under certain experimental conditions, but in man so far only for fibre is the evidence reasonably strong, notably for cancers of the colon and rectum. While the role of vitamin A deficiency has yet to be demonstrated in man. some studies report that in creased vitamin A protects against lung cancer (Wald. Idle. Boreham Sl Bailey. 1980). Earlier reports that malnutrition was a major factor in liver cancer in Africa have noi stood up to further examination (Higginson. 1963). Further, diet also modifies the nature and presence of mutagens in the faeces, some of which have been causally associated with cancer of the large intestine (Dion, Bright-See. Furrer. Eng Sl Bruce, 1980) . Mutagens may arise from normal nutrients during cooking (Kawachi. Nagao. Yahagi, Takahashi. Sugimura. Takayama. Kosuge Sl Shudo, 1979). Reports relating gastric cancer to nitrate ingestion leading to endogenous formation of niirosamines remain to be confirmed (Tannenbaum & Young. 1980). Considerable data have accumulated, indicating that risk factors arising from behaviour and diet may also be associated with biochemical and metabolic variations in the host, especially hormonal variations which have possible relevance to buman carcino genesis (de Waard. 1979; Emster. Sacks, Setvin Sl Petrakis. 1979; Gusberg. 1980; Moolgavkar, Day Sl Stevens, 1980; Ross, Paganini-Hill, Gerkins, Mack, PTeffer. Arthur & Henderson. 1980; Trichopoulos. Cole, Brown, Goldman Sl MacMahon, 1980; Vesell. 1980). The promoting role of exogenous oestrogens in endometrial cancer (1ARC Working Group, 1979b) and the inhibitory effect of certain contraceptives in endometrial and ovarian cancers also support the possibility that endogenous hormonal factors may be implicated (Cole. 1980; Hulka, Fowler, Kaufman. Crimson. Greenberg. Hogue, Berger Sl Pulliam. 1980; Smith. Prentice, Thompson Sl Herrmann, 1975). Cancer incidence has been shown to be modified by socio-economic gradients (Office of Population Cen suses and Surveys. 1978) which are closely associated with diet, behaviour and cultural variations (Morris. 1979). Further, while epidemiological studies on cancer and occupation are classically associated only with the identification of discrete carcinogens, they may be equally informative in evaluating lifestyle and socio-economic factors (Morris. 1979; Office of Popu lation Censuses and Surveys, 1978). Since individual occupations are recruited from specific segments of the community, the health patterns in such occupa tions reflect the local community environment and socio-economic background. Fox Sl Adelsiein (1978) calculated that most of the differences' (88%) in cancer patterns between occupational groups are probably due to lifestyle variations and not lo workplace ex posures. This is further illustrated by studies on breast and cervical cancers (Devesa Sl Diamond. 1980; Pell, O'Berg & Karrh, 1978; Trichopoulos, MacMahon & Brown, 1980). Davies. Edmundson, Raffonelli. Cassadi Sl Morgrade (1972) report that storage of chlorinated hydrocarbons also correlates with socio economic gradient Distinction should be made accordingly between tumours due lo industrialization per se and cancers occurring within a highly devel oped society. In contrast to the delay in accepting the role of lifestyle factors in cancer, the concept was accepted more rapidly in the case of cardiovascular disease, and led to considerable publicity on the benefits of dietary and behavioural changes. Significance of lime trends The process of establishing the relative importance of ambient pollution and lifestyle factors has been complicated by conflicting interpretations of cancer time trends. The evaluation of such trends is hindered by registration artefacts resulting from variations in population boundaries, census, diagnostic criteria, etc., as well as confounding variables such as cigarette smoking and social habits. The morbidity data in the US from 1947 to 1970 have been analysed by Devesa & Silverman (1978) who showed that apart from the increase in tobaccoand alcohol-related cancers, other cancers were tending to decrease, trends supported by the mortality data. Changes since 1970 are especially difficult to interpret (Pollack & Horm, 1980), due to registration arte- facts, etc., which have often been ignored. However, the available morbidity and mortality data are in genera! consistent with earlier observations, and do not suggest real increases in cancers apart from those associated with cultural habits, e.g. tobacco and sunlight. The modest increases in cancers of the breast. prostate and bladder are probably artefactual. e.g increased histological diagnosis, or changes in lifestyle factors. The latter are also probably responsible for the decrease in gastric and cervical cancer. These interpretations do not imply that localized or general chemical exposures have been completely without effect on general cancer patterns. However, they pro vide no support for the existence of a new general cancer 'epidemic' apart from the effects of cigarette smoking, alcoholism, and to a much lesser degree asbestos in certain limited population groups. tf t.^ - ^ V j , , y "/, * -'W Rl on Censociated (Morns, dies on ted only ns, they tyle and jf Popudividual nents of occupaent and n (1978) n cancer jrobably >Iace exn breast 80; Pell, ahon & affonelli. 'rage of h socio; made ilization y devel- role of accepted disease, nefits of oortance ias been cancernindered itions in criteria, cigarette to 1970 n (1978) tobacco:re tendnortality t/ p' f/ */ * to interion arteJowever. e in gen- 1 do not m those and sun- . e breast. 1. e.g. in- y, lifestyle ^ sible for v\ r. These ^ r general without hey progeneral cigarette r degree Environmental cancer causation 543 It has been argued that the increased production of industrial organic chemicals is too recent to permit evaluation of the long-term effects. However, con siderable quantities were already in fact produced by 1950 (Davis &. Magee, 1979). It should be emphasized that increased production, however, cannot automati cally be equated with a comparable increase in car cinogen exposures in man (Lenihan & Fletcher, 1976; Morris, 1975) since public health and industrial con trols affecting both ambient and point-source ex posures have been implemented in many countries. From the microbiological angle it would be interest ing to know if there has been any increase in back ground mutational rates which could be attributed to variations in the chemical environment over the last 30 yr. The existence of localized areas of unusual cancer incidence, or marked changes in temporal trends requires the continual support of adequate sur veillance and monitoring mechanisms. While attention has been concentrated on changes in chemical exposures, it is frequently forgotten that comparable changes in both dietary and behavioural patterns have occurred in many countries, largely as a result of efforts to control cardiovascular disease, and the effects of these changes on cancer patterns remain to be determined. In the United States, for example, between 1963 and 1977, there was a fall in the per capita consumption of tobacco of 29%, of eggs 15% of milk and cream 23% of butter 36% and of animal fats and oils 47% whereas consumption of vegetable fats and oils increased by 58% (McQuade. 1980). Furthermore, between 1910 and 1976, the consump tion of carbohydrates fell by ?1% of which the contri bution from starches fell by 45% (Brewster & Jacob son. 1978). If such changes are as important in the modification of human cancer patterns as migrant and geographical pathology studies suggest, their effects should be observed during the next few decades, and appropriate action initiated. However, the literature on diet in animal and human cancer indicates many inconsistencies in establishing clearcut relationships (Higginson & Muir, 1979). In conclusion, although there is a widespread belief in the importance of dietary factors, and while their role in promotion, enzyme induction, etc., is wellestablished in experimental animals, their role in man is much less clear. The difficulties of evaluating the significance of modest dietary changes 'as seen in man in the context of other modulating factors compared with the very marked changes seen in isolated animal experimentation increases the difficulty of extrapola tion. This is well illustrated by Hoehn & Carroll (1979) who demonstrated the effects of simple carbohydrate changes on DM BA-induced tumours in rats. Cancers of unknown aetiology Only a few discrete causal factors have been deter mined for most tumours of children, bone and soft tissues and haemopoietic system, the aetiology of the majority being unknown. Most tumours show very modest geographical and temporal variations in inci dence, and their relationship to environmental factors cannot accordingly be easily evaluated. It is templing to regard such cancers as representing a background incidence in man similar to 'spontaneous' tumours in animals which, however, are often related to inherited factors or endogenous viruses. The implications of epidemiology to the study oj car cinogenic mechanisms The multifactorial nature of human canoer has long been recognized. As the influence of other modulating factors in the early and late stages of carcinogenesis is increasingly recognized (Berenblum, 1979; Higginson. 1980; Miller & Miller, 1979; Wynder ex al. 1978), it seems less appropriate to consider carcinogenesis in man only in terms of the classical two-stage skin model. Contrary to recent reports, human cancer has not been considered in terms of single causes by most epidemiologists (Table 1). Kennaway (1950) empha sized the modifying effect of diet. The interplay of several factors was strongly supported by numerous studies such as on alcohol, asbestos and cigarette smoking (Hammond & Selikoff, 1973; Tuyns. 1978; Wynder A Gori, 1977), hepatitis virus in liver cancer (Higginson, 1963k and the interaction of multiple car cinogenic risk factors in cancers of the endocrinedependent systems (Miller. A. B,, 1978). Adenocarci noma of the vagina due to diethylstilboestrol is also suspected to involve a second factor. Further, the fact that latent carcinoma of the prostate is equally preva lent in most races, whereas the incidence of invasive carcinoma varies widely, strongly suggests an inter play of initiating and promoting factors (Breslow, Chan, Dhom, Drury, Franks, Gellei, Lee, Lundberg, Sparke, Sternby ft Tulinius, 1977). Figures 1 to 3 are diagrammatic representations of the carcinogenic process. While division into early, initiation, and late phases is probably an over-simpli fication in view of the probable overlaps, the diagram provides a useful base for discussion. It does illus trate moreover that clinical cancer does not represent a single mechanism but a series of events, only a few Table 1. Human cancers with evidence <4 multistage and multifactorial origin Site/tumour type Aetiological factors Lung Oesophagus Liver Stomach and large intestine Breast and endometrium Burkitt's lymphoma Nasopharyngeal carcinoma Smoking, asbestos, ionizing radiation Alcohol, smoking Hepatitis B, mycotoxins Endogenous carcinogen formation, non-spe cific dietary factors Behaviour, diet, hormonal promotion EBV, malaria, factor X Host factor, EBV, factor X URL 04G72 i URL 04073 a54 J. Higginson MOD*viAG CAfte<NMftNS'S Ufttnime <v avmc) (II SAftLV STA&( 1 r* (21 INITIATION Irn Ol tATf STAGt r* 1 lNCOM*Llf* 0* ^ Pli VI Hi( ePWj|T| ^ o AM vff MODULATING (ACTORS ChGISTiO. A*$OPFTiQ* ACTiVATlON'INACTIVATION tHZrm INDUCTION fTCSpecific "ICC*TOAS MUTAGC NCStS ONA K#lGE*C TrC tHClSiCNCCfrrOAS CTC gncnttn mocrcskm MIOMOTION INHIBITION iNHANCCWCNT OlPACMNTlATlON NUTftlTlONAl STATUS IMMUNITY ITC Fig. I. A simplified diagram showing the hypothetical stages in carcinogenesis and some of (he exogen ous and endogenous factors possibly involved. of which have adequately been studied and measured (Vesell, 1980). Although order is essential to the con* oept of cancer development and progression, there is good experimental evidence that many steps are at least partially reversible (Farber & Cameron, 1980; Singer. 1979). It seems a reasonable possibility that individual susceptibility and the contribution of lifestyle as indi cated by carcinogenic risk factors may eventually be explicable in terms of reversible and irreversible bio chemical events and feedback mechanisms operating in both early and late-stage carcinogenesis. Further, there may be both qualitative and quantitative differ ences between the effects of modulating factors in cases where discrete carcinogens providing a genotoxic stimulus have been identified (Fig. 2\ and life style' cancers where both extrinsic and intrinsic car cinogenic initiating stimuli may be less important (Fig. 3). The latter illustrates the difficulty of extrapo lation from experimental systems, especially for cancers where the modulating factors have yet to be clarified. The importance of enzyme activation and inactiva tion of procardnogens and carcinogens (C1BA Foun dation, 1980) is widely recognized. Further, variations in enzyme induction, whether genetically or environ mentally determined, may affect both individual susceptibility to carcinogens (Vesell, 1980), and the inddenoe of certain endocrine-dependent tumours in which subtle variations in hormonal metabolism may be important (Cole, Brown St MacMahon, 1976k In creased extra-glandular formation of oestrogens occurs largely in fat tissues and may explain the re lationship between obesity, overeating and endo metrial cancer. Endogenous caTdnogen formation from carcinogen precursors may be controlled to a considerable degree by environmental factors, by mechanisms including altered enzyme metabolism Epidemiological studies are only now being devel oped and will require much more intensive laboratory backup. Initiating events are easier to explain in terms of genotoxic than non-genotoxic stimuli, since the latter are poorly understood. They may involve gene unmasking, modification of DNA repair, epigenetic changes, etc. (Kroes, 1979), all of which can be envir onmentally modified and have important implications in terms of the completeness and the reversibility of initiation. It is thus possible to assume that in man initiation is common but often incomplete or lethal cancer de velopment being dependent on events in the later stages as well as during initiation (Rajewsky, 1979). This is obviously difficult to demonstrate, but strains of rats with a high frequency of spontaneous hepa tomas are regarded as being more susceptible to the promoting action of phenobarbitone than low-inci dence strains (Periino, Staffeidl, Haugen, Lombard, Stevens & Fry, 1980). Similarly, it has been found that initiated but not transformed cells are present in highskin-cancer strains of mice but not in those of low incidence (S. Yuspa, U, Lichei, D Morgan St H. Hen nings, unpublished data, 1980). In man, the possibility of incomplete initiation is supported by pathological studies on the prostate usd breast (Farber St Cameron, 1990k etc. Whether such initiation is due to endogenous agents alone is a matter of speculation. Archer (1980) has suggested that background radi- JMIYVTMC 1 MtTUTlOM _L 'S|N LATl (TAW griwi CMLTITMI WTUDGN iATI Fig. 2. The predominant effect of a strong discrete initiator (e.g an alkylating agent) is illustrated in comparison with (he slight effects of modulating factors. Fig 3. The relatively important role of`lifestyle' modulat ing factors is illustrated in a situation where the role of a direct initiating stimulus may be weak. 1 I i 1 nation d to a rs, by x>lism. develiratory rms of : latter gene genetic envir;ations ility of tiation xt de: later 1979). strains hepato the *-indnbard, d that i highof low . Hensibility logical r & due to lation. . radi- idulatle of a Environmental cancer causation 545 ation is a factor of significance in many cancers, but The implication of a discrete chemical carcinogen the failure to demonstrate any relation with ambient or of a defined cultural habit allows a relatively chemical pollution makes it more difficult to impli straightforward approach to cancer prevention. How cate the latter although obviously it cannot be ex ever, the situation has not been so simple in the case cluded. of cancers thought to be related to lifestyle, where The many factors that may be involved in late-stage there have been only very general indications of the carcinogenesis have been reviewed (Slaga, 1980). But factors involved and understanding of the underlying c\en for such extensively studied carcinogenic agents biological mechanisms has been imperfect. It is in man as cigarettes, the relative importance of in becoming increasingly clear that combinations of 1 itiation and promotion is not understood. However, modulating endogenous and exogenous factors may the fact that the incidence of lung cancer 'freezes' after explain the influence of lifestyle based on identifiable 1 cessation of smoking suggests that promoting factors biochemical mechanisms. This may eventually allow are important in later development. In addition to the possibility of tumour prevention by interference reactions at the target cell, changes in the host may with other stages of carcinogenesis through chemo- also be pertinent, e.g. immunological factors (Kinlen, prevention (Sporn & Newton, 1979), use of cancer in Eastwood, Kerr, Moorhead, Oliver, Robinson, de hibitors and dietary changes. The process of attribut Wardener & Wing. 1980). Conney et al. (1978) have ing many cancers to lifestyle factors has the benefits of demonstrated the role of diet in carcinogen metab directing research attention to factors beyond discrete olism in man and animals. The slow increases in carcinogens and of arousing intellectual curiosity. Un breast and prostatic cancers in migrants (Haensze) & fortunately. at present it does not provide a basis for Kurihara. 1968) show that environmental differences the immediate and effective control of many cancers may take several generations to become apparent, of environmental origin. and the possibility of sex-linked enzyme imprinting in such cancers in humans should not be ignored. Such a mechanism has been postulated for breast cancers in rats (Mori. Nagasawa & Bern. 1979). REFERENCES Individual susceptibility It is not easy to differentiate between environmen Archer, V. E. (1980). Cancer and anencephalu in man as sociated with background radiation. Med- Hypotheses 6, 81. tally and genetically determined individual susceptibi Berenbium. I. (1978). Historical perspectives. In Carcino lity (Vesell, 1980). Previous attempts to relate lung genesis--A Comprehensive Survey. Vol. 2. Mechanisms of cancer susceptibility to the induction of aryl hydro carbon hydroxylase have not been confirmed. Apart from such clear-cut hereditary syndromes as xeroderma pigmentosum and ataxia-telangiectasia, etc., few cancers can be directly attributed to genetic factors. On the other hand, genetic polymorphism may have considerable significance both in relation to Tumor Promotion and Cotarcinogencsis. Edited by T. J Slaga. A. Sivak A R. K. Boutwcll. p I. Raven Press. New York. Berenbium, I. (1979) Theoretical and practical aspects of the two-stage mechanism of carcinogenesis. In Carcino gens. Identification and Mechanisms of Action. Edited by A- C. Griffin & C. R. Shaw, p 25. Raven Press, New York. enzyme induction and host reactions to exogenous Boyland, E. (1969). Correlation of experimental carcino modulating factors (del Villano, Miller, Schacter & Tischfield, 1980; Schull, 1979). Migrant studies, how ever. emphasize the overwhelming impact of environ*mental changes. genesis and cancer in man. Prog. exp. Tumor Res. It, 222. Breslow, N,, Chan, C. W.. Dhom, G., Drury, R. A. B., Franks, L. M.. Geliei. B, Lee. Y. S.. Lundberg, S_ Sparke. B,, Steraby. N. H. A Tulimus. H. (1977). Latent Conclusions For nearly 50% of environmentally influenced cancers among men in North America and Europe carcinoma of the prostate at autopsy in seven areas. Ini. J. Cancer 20, 680. Brewster, L. A Jacobson. M. F. (1978). The Changing American Diet. Center for Science in the Public Interest. ) the main aetiological factors have been identified but Washington. the percentage is much lower for women. Many of the Carson. R. (1962). Silem Spring. Houghton Mifflin Co.. causes had already been suggested in the fifties. Since Boston. then, much laboratory and epidemiological research has been directed towards investigating the nature of and the mechanisms involved in cancer in man. Increasingly sophisticated technology has been used in attempts to identify the initiation stimuli and the many modulating factors involved. Studies of the pro Cederlof, R_ Doll R., Fowler, B.. Friberg. L-, Nelson, N. A Vouk, V. (1978). Air pollution and cancer: risk assess ment of methodology and epidemiological evidence Encir. Hlth Perspect. 22, 1. CIBA Foundation Symposium 76 (1980) Environmental Chemicals, Enzyme Function and Human Disease. Excerpta Medica. Amsterdam. Oxford. New York. gression of cancer and the multi-stage nature of carci Clemmescn. J. (Ed.) (1950). Sympo.stum an Geographical nogenesis make it increasingly evident that data from Pathology and Demographi of Cant er. Oxford. England e-v.h species are complementary and that an interdis Council for International Organizations of Medical ciplinary approach is important. Clearly exposure to high levels of carcinogens must be controlled, but the" view- that multiple exposure to very low levels of mutagens and carcinogens (general environmental pollution) may have relatively little effect on the total cancer burden indicates the importance of alternative Sciences, Paris. Cole, P. (1980). Major aspects of the epidemiology of breast cancer. Cower. N.Y 46. 865. .Cole. P_ Brown. J. B. A MacMahon. B. (1976) Oestrogen profiles of parous and nulliparous women. Lancet II. \ 596. | Cole. P. A Merletti, F (1980) Chemical agents and occu approaches to cancer prevention and control. ' pational cancer. J. encir. Path. Toxicol. 3, 399. URL 04074 546 J. HIGGINSON Conney, A. J.. Pantuck. E. J_ Pantuck. C. B.. Buening. M.. Jenna. D. M, Fortner. J. G.. Alvares, A. P_ Anderson. K. E. A Kappas. A. (I97B). Role of environment and diet in the regulation of human drug metabolism. In The Induction of Drug Metabolism. Edited by R. W. Eastbrook & E. Lindenbaub. p. 563. F. K. Schatttuer Verlag. Stuttgart. New York. Coulston. F. (Ed ) (1979). Regulatory Aspects of Carcino genesis and Food Additives: The Delaney Clause. Aca demic Press. New York. Crespi. M.. Munoz. N- Grassi. A.. Aramesh. B.. Amiri. G., Mojtabai. A. St Casalc, V. (1979). Oesophageal lesions in Northern Iran: a prcmalignant condition? Lancet (I, 217. Davies. J. E,, Edmundson. W. F, Raffonelli. A. Cassadi. J. C. A Morgradc. C. (1972). The role of social class in human pesticide control. Am. J. Epidemiol. 96. 334. Davis. D. L. A Magee. B. H. (1979). Cancer and industrial chemical production. Science, N.Y. 206, 1356. Del Villano. B. C* Miller. S. L Sdiacter. L. P. A Tischfield. J. A. (1980). Elevated superoxide dismutase in black alcoholics. Science. N.Y. 207. 991. Devesa. S. S. A Diamond. E. L (I960). Association of breast cancer and cervical cancer incidences with income and education among whites and blacks. J. natn. Cancer Inst. 65, SI5. Devesa. S- S. A Silverman. D. T. (1978). Cancer incidence and mortality trends in the United Stales: 1935-74. J. natn. Cancer Inst. 60, 545. de Waard. F. (1979). Premenopausal and postmenopausal breast cancer . One disease or two? J. num. Cancer Inst. 63.549. Dion. P. W_ Bright-See. E. B- Furrer. R,, Eng. V. W. S. & Bruce. W. R. (I960). The effects of dietary fat. ascorbic acid, and alpha-tocopherol on fecal mutagens. In Oiical Investigations. Proceedings of AACR and ASCO. p. 171. AACR Abstracts. Doll. R. (1967). Prevention of Cancer: Pointers from Epi demiology. Nuffield Provincial Hospitals Trust. London. Enstrom. J. E. (I960) Health and dietary practices and cancer mortality among California Mormons. In Ban bury Report 4: Canrcr Incidence in Defined Populations. Cold Spring Harbor Laboratory, p. 69. Cold Spring Harbor. New York. Epstein. 5- S. (1978). The Politics of Cancer. Sierra Club Books. San Francisco. Ernster. V. L.. Sacks. S. T.. Selvin. S. A Petrakis. N. L. (1979). Cancer incidence by marital status: U.S. Third National Cancer Survey. J. natn. Cancer Inst. 63. 567. Farber. E. A Cameron. R. (1980). The sequential analysis of cancer development. Adr. Cancer Res. 31. 125. Fox. A. J. & Adelstein. A. M. (1978). Occupational mor tality; Work or way of life? J. Epidem. commun. Hlth 32. 73. Goldsmith. J. R. (1960). The "urban factor- in cancer: Smoking, industrial exposures, and air pollution as poss ible explanations. J. encir. Path. Toxicol. 3, 205. Gori. G. B (1976). Diet and nutrition in cancer causation. Nutr. Cancer I, 5. Gusberg. S. B (1980) The changing nature of endometrial cancer. New Engl. J Med. 302, 729. HaenszeL W. (1961), Cancer mortality among the foreign born m the United States. J. asm. Cancer Inst. 26. 37. HaenszeL W, A Kurihara M. (1968) Studies of Japanese migrants. I. Mortality from cancer and other diseases among Japanese in the United States. J. natn. Cancer Inst 40, 43. HaenszeL W_ Loveland. D. B. A Sirkcn. M. G. (1962). Lung cancer mortality as related to residence and smok ing histories. I. White males. J natn. Cancer Inst. 28, 947. HaenszeL W. A Taeuber. K. E. (1964). Lung-cancer mor tality as related to residence and smoking histories. II. White females. J. natn. Cancer Inst- 32, 803. Hammond. E. C. A GarfinkeL L. (1980). General air pollu tion and cancer in the United States. Prec. Med. 9, 206. Hammond. E. C. A Seidman. H- (1980). Smoking and cancer in the United States. Prec. Med. 9, 169. Hammond. E. C. A Selikoff. I. J. (1973). Relation of cigar ette smoking to risk of death of asbestos-associated dis ease among insulation workers in the United States. In Biological Effects of Asbestos. Edited by P. BogovskL J. C. Gilson, V. Timbrel! A J. C Wagner. 1ARC Scient. Pub!, no. 8. p. 312. International Agency for Research on Cancer. Lyon. Higginson. J. (I960). Population studies in cancer. Acta Vn. int. Cancr. 16, 1667. Higginson. J. (1963). The geographical pathology of primary liver cancer. Cancer Res. 23, 1624. Higginson. J. (1969). Present trends in cancer epidemi ology In Proceedings of the Eighth Canadian Cancer Conference, Honey Harbour. Ontario, 196b. p. 40 Pergamon Press. Canada. Higginsoa J. (1979). Perspectives and future developments in research on environmental carcinogenesis. In Carcino gens: Identification and Mechanisms of Action Edited by A. C. Griffin & C. R. Shaw. p. 187. Raven Press. New York. Higginson, J. (1980). Implications for future studies in humans. In Environmental Chemicals. Enzyme Function and Human Disease. Ciba Foundation Symposium 76. p. 349. Excerpta Medica. Amsterdam. Higginson. J. & Muir. C. S. (1976V The role of epidemi ology in elucidating the importance of environmental factors in human cancer. Cancer Detect. Prev. I, 79. Higginson. J. A Muir, C. S. (1979). Environmental carcino genesis: Misconceptions and limitations to cancer con trol. J. natn. Cancer Jnsl. 63, 1291. Higginson. J. A Oeltle. A. G. (I960). Cancer incidence in the Bantu and "Cape Colored- races of South Africa: Report or a cancer survey in the Transvaal (1953-55). J. natn. Cancer Inst. 24, 589. HilL P- Wynder. E. L.. Garbaczewski. L,, Hetman, P_ Hill. M. Sporangisa. J. A Huskisson. J. (1980V Diet and menarche in different ethnic groups. ur. J. Cancer 16. 519. Hirayama. T. (1979). Diet and cancer. Nutr. Cancer 1, 67. Hoehn, S. K. A Carroll K. K. (1979). Effects of dietary carbohydrate on the incidence of mammary tumors induced in rats by 7.l2-dimethy)ben4a)anthraccne. Nutr. Cancer 1, 27. Huika. B. S,, Fowler. W. C- Kaufman. D. G.. Grimson. R. C- Greenberg. B. G- Hogue. C J. R.. Berger. G. S. A Pulliam, C. C. (1980). Estrogen and endometrial cancer: Cases and two control groups from North Carolina. Am. J. Ohstet. Gynec. 137. 92. IARC Working Group (1979a). IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans. Chemicals and Industrial Processes Associated nith Cancer in Humans. Vols 1 to 20. Supplement I. International Agency for Research on Cancer. Lyon. IARC Working Group (I979bl IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans. Sex Hormones {II). Voi. 21. International Agency for Research on Cancer, Lyon. Kalamegham. R- Krishnaswamy, K,, Krishnamurthy. S. A Bhargava. R. N. K, (1979). Metabolism of drugs and carcinogens in man; Antipyrine elimination as an indi`ttlor. Clin. Phormoc. Ther. 25, 67. Kawachi. T.. Nagao. M- Yahagi. T- Takahashi. Y.. Sugimura. T,, Takayama, S. Kosuge, T A Shudo. T. (1979). Mutagens and carcinogens m food. In Advances in Medi cal Oncology Research and Education. Volume I. Carcino genesis. Edited by G. P. Margison. p. 199. Pergamon Press. Oxford and New York. URL 0407: cn 19 ts II. DOllu- . 206. : and cigar* d dis cs. In ovski. idem. ch on tf Vn. *y or demi* 'ancer 3erga- ments rcinoed by New ies in union 76. p. ideminental 9. treinor con- nce in Africa: 55). J- _ Hill. m and rrr 16, 1.67. iieiary umors Nuir. imson. jS& ancer; ia Am. on the cals to oeiated nent 1. on. on the cals to ational y. SA gs and n indi- .. Sugi(1979). i Mediarcinogamon 1 Environmental cancer causation 547 Kennaway. E. L (1944). Cancer of the liver in the Negro in Africa and in America. Cancer Res. 4, 571. Kennaway. E. L (1950) Forms of cancer in man suitable for investigation. In Symposium on Geographical Pathol ogy and Demography of Cant er. Edited by J. Clemmesen. p 122 Council Tor International Organizations of Medi cal Sciences. Paris kinlen. L. Eastwood. J. B.. Kerr. D. N. S.. Moorhead. J. F.. Oliver. D. A.. Robinson, B H B.. de Wardener. H. E & Wing. A. J. (1980). Cancer in patients receiving dialysis, fir Med J 280. 1401 Kraybill. H F (1978) Origin, classification and distribu tion of chemicals in drinking water with an assessment of their carcinogenic potential. In Water Chlorination: Environmental Impact and Health Effects. Edited by R. L. Jolley, p. 211. Ann Arbor Science Publishers. Ann Arbor. Michigan. Kritchevsky. D. (1980). Steroids, fiber and related factors in carcinogenesis J. emir Path. Toxicol. 3, 305. Kroes. R. (1979) Animal data, interpretation and conse quences In Environmental Carcinogenesis: Occurrence. Risk Evaluation and Mechanisms. Edited by P. Emmelot & E Knek. p. 287. Elsevier 'North-Holland Biomedical Press. Amsterdam Land. C E (1980) Estimating cancer risks from low doses of ionizing radiation Science. H.Y. 209, 1197. Larouze. B.. Blumberg. B S,, London. W. T , Lustbader. E. D.. Sankale. M. & Payet. M. (1977). Forecasting the development of primary hepatocellular carcinoma by the use of risk factors. Studies in West Africa. J. natn Cancer Inst. 58. 1557. Lenihan. J. & Fletcher. W. (Eds.) (1976) Environment and Man. Blackie. Glasgow and London. Lmsell. C. A. & Peers. F. C (1977). Field studies on liver cell cancer. In Origins of Human Cancer Edited by H H. Hiatt. J. D Watson & J. A. Wtnsten. p. 549. Cold Spring Harbor Laboratory. Lyon. J. L,, Gardner. J. W & West. D W (1980) Cancer risk and lifestyle: cancer among Mormons from 1967 to 1975 In Banhur\ Report 4. Cancer Incidence in Defined Populations, p 3. Cold Spring Harbor Laboratory, New York. McQuade. W (1980) Good news from the House on Lin coln Street (The Framingham Heart Study is helping the nation win the long struggle against coronary heart dis ease) Fortune, p. 86. 14 January Miller. A B (1978) An overview of hormone-associated cancers Cumer Res. 38, 3985. Miller. A B (1980) Nutrition and cancer. Prev. Med. 9, >89. Miller. E. C. (1978). Some current perspectives on chemical carcinogenesis m humans and experimental animals: Presidential address. Cancer Res. 38, 1479. Miller. E C.. Miller. J A.. Brown. R. R & MacDonald. J C. (1958). On the protective action of certain poly cyclic aromatic hydrocarbons against carcinogenesis by ammoazo dyes and 2-acetylaminofluorene. Cancer Res. 18, 464 Miller. J. A. A Miller. E. C. (1979). Perspectives on the metabolism of chemical carcinogens. In Environmental Carcinogenesis Edited by P. EmzneJor A E K/iek. Elsevier North-Holland Biomedical Press Amsterdam. Mooigavkar. S. H., Day. N. E. A Stevens, R. G (I980|. Two-stage model for carcinogenesis: Epidemiology of breast cancer in females. J natn. Cancer Inst. 65, 559. 'Ion. T.. Nagasawa. H & Bern. H A. (19791. Long-term effects of perinatal exposure to hormones on norma) and neoplastic mammary growth of rodents: A review. J. emir Path. Toxicol. 3, 191. Morns. J N. (1975) Uses of Epidcmiologv. Churchill Liv ingstone. Medical Division of Longman Group Limited. Morris. J N.(1979). Social inequalities undiminished Lan cet 1, 87 Newberne. P. L M i McConnelL R G (1980) Nutrient deficiencies in cancer causation. J. envir. Path. Toxicol. 3, 323. Newberne. P. M-. Weigert. J. i Kula, N. (1979). Effects of dietary fat on hepatic mixed-function oxidases and hepa tocellular carcinoma induced by aflatoxin Bl in rats. Cancer Res. 39, 3986. Office of Population Censuses and Surveys--Occupational Mortality (1978) The Registrar General's decennial sup plement for England and Wales. 1970-72, series DS No. 1. London. Her Majesty's Stationery Office. Pell. S.. O'Berg. M. T. i Karrh. B W. (1978). Cancer epi demiologic surveillance in the Du Pont Company. J. occup. Med. 20. 725. Peraino. C,, Staffeldt, E. F.. Haugen. D. A, Lombard. L. S,, Stevens. F. J i Fry. R J. M. (1980) Effects of varying the dietary concentration of phenobarbital on its en hancement of 2-aceiylaminofiuorene-induced hepatic tumorigenesis Cancer Res. 40, 3268. Phillips. R. L. Kuzma. J. W. & Lotz, T. M. (1980). Cancer mortality among comparable members versus non members of the Seventh-day Adventist Church. In Bunhury Report 4.- Cancer Incidenc e in Defined Populations p. 93 Cold Spring Harbor Laboratory. New York. Pollack. E S. & Horm, J. W (1980). Trends in cancer incidence and mortality in the United Stales, 1969-1976. J. natn. Cancer Inst. 61. 1091. Rajewsky. M F. (1979). Specificity of DNA damage in chemical carcinogenesis. In Molecular and Cellular Aspects of Carcinogen Screening Tests. Edited by R. Montesano. H Bartsch & L. Tomatis. p. 41. IARC Scient Pub), no. 27. International Agency for Research on Cancer. Lyon. Rose, A. H. (Ed.| (1977). Alcoholic Beverages (Economic Microbiology. Volume I). Academic Press, London. New York. San Francisco. Ross. R. K., Paganini-Hill. A.. Gerkins. V. R . Mack. T. M, Pfeffer. R . Arthur. M. & Henderson. B E. (1980) A case-control study of menopausal estrogen therapy and breast cancer. J. Am. Cancer Sot. 243, 1635. Royal Society Study Group (1978). Long-term Toxic Effects. Final report of a Royal Society study group p. 5. Royal Society. 6 Carlton House Terrace. London SW1Y 5AG, July 1978. Schuli, W. J. (1979). Genetic structure of human popula tions. J. envir. Path. Toxicol. 2. 1305. Singer, B- (1979). N-Nitro*o alkylating agents: Formation and persistence of alkyl derivatives in mammalian nu cleic acids as contributing factors in carcinogenesis. J. natn. Cancer Inst. 62. 1329. Slaga. T. J. fEd I 11980). Cocuninogencsis--A Comprehen sive Survey. Volume 5: Modifiers of Chemical Carcino genesis--An Approach to the Biochemical Mechanism and Cancer Prevention. Raven Press, New York. Smith, D. G. Prentice. R.. Thompson, D. J. A Herrmann. W. L. (1975). Association of exogenous estrogen and endometrial carcinoma. New Engl J Med. 293, 1164 Sporr., M. B & Newton. D L. (1979). Chemoprevention of cancer with retinoids. Fedn Prot. Fedn Am. Snc\ exp. Biol. 38, 2528. Tannenbaum. S. R. A Young. V. R. (1980) Endogenous nitrite formation in man. J. envir. Path Toxicol. 3, 357. Trichopoulos, D., Cole. P.. Brown. J. B.. Goldman. M B. & MacMahon. B. (1980). Estrogen profiles of primiparous and nulliparous women in Athens. Greece J. natn. Cancer Inst. 65, 43 Trichopoulos, D-. MacMahon. B. A Brown. J (1980) Socioeconomic status, urine estrogens and breast cancer risk J. natn. Cancer Inst. 64. 7}3. Tuyns. A. J. (1978) Alcohol and canceT. Alcohol Health and Research World 2. 20 Vesell. E. S. (1980). Genetic and environmental factors af fecting the metabolism of carcinogens. In Molecular and UfiL 04076 i I 54S J. Higoinson Cellular Aspects of Carcinogen Screening Tests. Edited by R. Montetano. H. Bartsch ft L. Tomatii. p. 23. IARC Scient. Publ. no. 27. Internationa! Agency for Research on Cancer. Lyon. Wald. N- Idle. M.. Boreham. J ft Bailey. A. <19801. Low erum-vitamin-A and subsequent risk of cancer. Lancet 11,813. Walker. E. A- Castegnaro. M,, Griciute. L. ft Lyle. R. E. (Eds.) (1978). Environmental Aspects of N-Nifroso Com pounds. IARC Scient. Publ. no. 19. International Agency for Research on Cancer. Lyon. Wattenberg. L. W. (1979). Inhibitors of chemical carcinogens. In Eneirtmmemal Carcinogenesis--Occurrence. Risk Evaluation and Mechanisms. Edited by P. Emmelot ft E. Kriek p 241 Elsevier/North-Holland Biomedical Press. Wilkins. J. R. Reiches. N. A. ft Kruse. C. W. (1979). Or ganic chemical contaminants in drinking water and cancer. Am. J. Epidemiol. 110. 420. Willis, R. A. (1948). Pathology of Tumours. Butterworth ft Co. Ltd, London. Wynder. E. L, ft Cork G. B. (1977). Contribution of the environment to cancer incidence: An epidemiologic exer cise. J. natn. Cancer Inst. SB. 825. Wynder. E, L-. Hoffmann. D_ McCoy, D.. Cohen. L. ft Reddy. B- (1976k Tumor promotion and cocarcinogenesis as related to man and his environment. In Carcinogenesis--A Comprehensive Survey. Volume 2. Edited by T. I. Slaga, A. Sivak ft R. K. Boutweli. Raven Press. New York. Yamamoto. Yamada, T. ft Tanimura. A, (1980). Vola tile nitrosamines in human blood before and after inges tion of a meal containing high concentrations of nitrate and secondary amines. Fd Cosmet. Toxicol. 18. 297. URL 04077 Carcinogenicity of Epoxides, Lactones, and Peroxy Compounds '*1 * * B. L. VAN DUUREN, N. NELSON, L. ORRIS, E. D. PALMES, and F. L. SCHMITT, Institute of Industrial Medicine, New York University Medical Center, New York, New York SUMMARY Attention is focused on epoxides, lac tones, hydroperoxides, and peroxides as carcinogens, as possible carcinogenic intermediates in the metabolism of aromatic hydrocarbons and as environ mental carcinogens. Available Infor mation about the carcinogenicity of these compounds is reviewed. To as certain their carcinogenic activity, 14 compounds in acetone or benzene solu tions were tested by skin painting on mice. Five epoxides, styrene oxide, l-ethyleneoxy-S,4-epoxycyclohexane, l,2-eporybutene-3, and dl- and meso1,2,3,4-diepoxybutane and one hydro peroxide, 1-bydroperoxy-l-vinylcyclohexeoe-3 showed carcinogenic activity. Some aspects of the relationship between structure and carcinogenic activity are discussed.--J Nat Cancer Inst 31: 41~55, 1963. MUCH ATTENTION has been devoted to the study of the carcinoge nicity, metabolism, and mode of action of many aromatic hydrocarbons, heterocyclics, amines, and azo dye9. Several' other types of compounds, mostly nonaromatic in structure, have been investigated to a limited extent, some of which showed carcinogenic activity, e.g., 0-propiolactone (/, B), 1,2,3,4-diepoxybutane (3, 4)t ^d 6-0-hydroperoxy-A4cholestene-3one (5). Several features of these compounds suggested that further inquiry into their carcinogenic activity was merited. Chemical reactivity.--These compounds are all highly reactive chem ically, in contrast to the relatively unreactive aromatic hydrocarbon carcinogens, and probably undergo a variety of chemical reactions with tissue constituents. Thus the crosslinking abilities of epoxides are well known (6); 0-lactones react with amineB and other functional groups such as the sulfhydryl group of cysteine. Thus, S-2-carboxyethylcysteine, I, is obtained (7) by reaction of cysteine with 0-propiolactone. HOOC (CHs),S CH,CH COOH Ira, I * Received January 2,1963. 1 This work vm aided by grant from tlie American Cancer Society, Inc., and by grant C-&MS from tbe National Cancer Institute, National Institutes ol Health, Public Health Service. Tbe infrared spectrophotometer and vapor phase tactometer were purchased through field investigation (rant CB-9713 from tbe National Cancer Institute, 41 683409--6; 3 ..ue / VAN PUUJREN T JX. This compound has been prepared earlier by other methods (8, 9). Hydroperoxides and peroxides can be expected readily to bring about oxidation reactions. The epoxides, like the various nitrogen and sulfur mustards, exhibit radiomimetic properties, e.g., growth-inhibitory, mutagenic, and cytotoxic activity (10-13). These radiomimetic agents all have one property in common: They are alkylating agents that undergo reactions with suitable nucleophilic groups via carbonium ions (14)- The types of reactions | which the carbonium ions can undergo with anions, e.g., 0-P-, Cl, etc., and bases, e.g., RNH:, R2S, etc., have been summarized and discussed by Ross (14). The effects of a number of alkylating agents in biological systems have been investigated (12, 15-17) extensively. Relationship to aromatic hydrocarbon metabolism.--An understanding of the necessary structural features required for carcinogenicity of the nonaromatic materials may be enlightening with respect to the mech anism of chemical carcinogenesis by aromatic hydrocarbons. For example, Boyland (18) has suggested that a phenanthrene-epoxide, II, may be an intermediate in the metabolism of phenanthrene, and Haddow has suggested that epoxides may be the proximal carcinogens (10). Epoxides of this type have not yet been synthesized, but if they were obtainable would probably be highly reactive. Although unrelated in its biological activity, tbe insecticide heptachlor, III, is metabolized to the epoxide, IV, in dogs and rats (19). The epoxide, CARCINOGENIC OXYGEN COMPOUNDS 43 extractable portion of photooxidized benzo(aJpyrene is carcinogenic to mouse skin. The products of the photooxidation were not identified, but they probably contained peroxides, phenols, and quinones (). Significance as environmental carcinogens.--Kotin (23) has suggested that epoxides and peroxy compounds, formed by the oxidation of olefins, may be carcinogenic air pollutants and has produced skin tumors in mice with atmospheric extracts free of aromatic hydrocarbons. It was suggested that the activity of this material was due to the presence of epoxides, but their presence was not confirmed experimentally. We have suggested earlier that compounds such as diketene (methyl- ene-0-propioIactone), hydroperoxides, and peroxides may be components of cigarette Bmoke (4)- Diketene, for example, can conceivably be formed by dimerization of ketene, itself a pjTolysis product of acetone. Also, hydroperoxides and peroxides can be formed by autoxidation of olefinic hydrocarbons in cigarette smoke. The search for, and identifi cation of, environmental carcinogens of these types will become worth while after their biological activity has been examined more extensively. The carcinogenicity of hydroperoxides, peroxides, and epoxides known to be present (5, 6) in autoxidized fats and oils are of obvious importance because of extensive use of such materials for human consumption. Stereochemical considerations.--Some metabolites of aromatic hydro carbons have been isolated in optically active forms (7, 8). But the stereochemistry of carcinogens has received little or no attention, though in some instances mixtures of stereoisomers were subjected to biological testing (S). The biological testing of pure isomers thus becomes of considerable importance, e.g., the racemic- and rreso-l,2,3,4-diepoxybu- tanes, V and VI. OH I^ CH, {c / Vc /l H! \01 OHO in, r 1/ H,C \JCL 1 H V VI Biological testing.--Tumors were produced (10, 9) in rats, mice, and hamsters with bifucctional aliphatic and aromatic nitrogen mustards, e.g., VII and VUI. N(CH,CH*CI)i vn vm Walpole and co-workers (3, 13, SO) showed that some bifunctional epoxides exhibiting radiomimetic properties, like the nitrogen mustards, are also carcinogenic. Since then, several other compounds of these and related types have shown carcinogenic activity. These experiments and the results on other known carcinogenic epoxides, lactones, and peroxy compounds are listed in tables 1 and 2. YOU 81, NO. 1, JULY 1963 44 VAN DUUREN ET AL. Table 1.--Known carcinogenic epoxides Compound Method o( testing 7,t2-Dlmetbyl-7,l2-epoxybem(al intbreceoe, X $' Mouse skin References (SI) <3H* 1,2,3,4-Dkpoiybatane, xi O0 C^r--CH--Cfl-- CH* l-Ethylenory-3,Apoiycyclohexane, XII Cflr-^CH--'y/ l.S-Epoxy-a-stearuylosypropene, XIII O C^CH-CHr-0--C--(CBt)it- CH, $ l^-Epory-3-lauroyloxypropane, XIV O cllj---CH--CH*--0--C--(CHi) it--CHj A l,2-Epoxy-3-eeproyloiypropao, XV O C^r--CH--CBj--O--C--(CB*)--CH* & ),3-Epoxyprop&ne, XVI 0 C^*---CH--CH* p-Hydroiydipbenylaminetlycldyl ether, XVII C^r--CH--CHr--O<>*<3 Bats, lntnperltoceel Injection (?) Bats end mice, skin and Id- (?.S0) traperltoneal injection Beta, subcutaneous Injection vn Beta, subcuteneoua injection (> Beta, subcuteneoua Injection (IS) Bats, subcutaneous injection vn Bets, subcutaneous injection (IS) Dickens and Jones (7) reported a low and possibly questionable order of carcinogenicity for a-carboxy-0-pbenyl-0-propiolactone, a,a-diphenyl0-propiolactone, and S-2-carboxyethyl-L-cysteine. These materials were tested subcutaneously in rats. Hine and co-workers (88) produced sarcomas in rats and skin cancer in mice with certain epoxy resins. The following materials were tested by various methods and gave nega tive results: l,2-epoxy-3-oleyloxypropane (18); l,2-epoxy-3-(2,4-dinitro- phenyloxy)propane (18); ethylene oxide (18); 7-valerolactone (84); 2,4-cholestadiene peroxide (35); and 7,12-dimethyI-7,12-peroxybenz[aJanthracene (86). Biological testing was in some instances carried out by JOURNAL OP THE NATIONAL CANCER INSTITUTE CARCINOGENIC OXYGEN COMPOUNDS 45 Table 2.--Known carcinogenic lactones and hydroperoxides Compound 0-Propiolaetone, XVIII Patulin, XIX HC--C=0 Hi 0------j=0 Method of testing Reference Meuse skin Rats, subcutaneous Injection '6F Rata, subcutaneous Injection (7) PeoJdUlc Kid. XX OCHi BO AHiC vHiC^ Meibyl protoanemonin, XXI ClH,C-CH= 4-Hes-3-eoolectooe, XXII H,C--CHr Rata, subcutaneous Injection (7) Rats, subcutaneous Injection (7) Rats, subcutaneous Injection <n VSea 4<noUctonc, Xalu a.HiC--CH= feMfydroperoty-Aicbolostcne^cn*, XXIV 'jH,( 0 Rats, snbcutineoos laJectioB (7) Mloe, subcutaneous Injection (f) subcutaneous injection in aracbis oil solutions in Tats. Walpole (IS) points out that bin method is unsatisfactory because arachis oil itself gave rise to late tumors in rats by subcutaneous injection. Walpole also observed that impure preparations of l-ethylenoxy-3,4-epoxycyclohexane gave higher tumor yields than purified preparations (/S). The paucity of information that could serve to establish relationships between structure and carcinogenicity indicated the need for biological VOL. 31, NO. 1, JULY 1963 / 46 VAN DtJUREN T AL. testing of pure compounds and, where necessary, of known stereochemical configuration. In the present work we attempted to select for biological testing, from available information, a variety of compounds of the types under investi gation. Commercially available compounds were subjected to careful purification, and purity was checked by the usual procedures including infrared spec troscopy; volatile materials were also checked by vapor phase chromatography. Compounds not commercially available were synthe sized in this laboratory. The results obtained in the initial phases of this work are presented here. MATERIALS AND METHODS Preparation and Purification of Compounds for Biological Testing Hexaepoxysgualene.--The synthesis, purification, and characterization have been described elsewhere (87). 9,10-Epoxystearic acid.--Oleic acid was oxidized with perbenzoic acid (38) and purified by crystallization from acetone, mp 59 [reported mp 59.5-59.8 (38)1. 9,10,t,18-Diepoxystearic acid.--Linoleic acid was oxidized with pera cetic acid (89) and purified by crystallization from acetone, mp 77 [reported mp 78 (85)1. Styrene oxide.--Commercial quality (Matheson, Coleman & Bell, East Rutherford, N.J.) styrene oxide, was "purified by vacuum distillation, bp 105/1 mm [reported bp $7-88723 mm; n" 1.5319, reported n" 1.5331 (40)]. The homogeneity of the material was determined by vapor phase chromatography; only one peak was observed with a 2 m polypropylene glycol column at 175 and 20 psi; retention time 13.6 minutes. l-Ethylenoxy-8,4-epoxycyclohexane.--Commercial quality (K & K Laboratories, Jamaica, N.Y.) material was purified by vacuum distillation, bp 134 to 135719 Him [reported bp 110-113720 mm (/)], n" 1.4747. The vapor phase chromatogram on a 2 m silicone column, 130, 5 psi, showed a peak with retention time 10.5 minutes. Two minor impurities, retention time 9.0 and 11.9 minutes, could not be removed by redistillation. l,-Epoxybutene-8.--Commercial quality (K & K Laboratories) 1,2epoxybutene-3 was purified by distillation, bp 63.5 to 64.0 at atmospheric pressure [reported bp 65-65.87739 mra (4^)]- The vapor phase chroma togram on a 2 m polypropyleneglycol column, 100, 20 psi, showed one peak only, retention time 2.4 minutes. dl~l,2,8,4-Diepoxybutane,--efs-2-Butene-l,4-diol (General Aniline and Film Corp,, New York, N.Y.), purified by chromatography on acidwashed alumina followed by distillation, was brominated (43) to give 1,4dihydroxy-2,3-dibromobutane; this product was purified by crystalli zation from benzene, mp 87.5 to 88 [reported mp 87 (44)], yield 74 percent. This compound was converted to the epoxide by treatment JOURNAL OF THE NATIONAL CANCER INSTITUTE URL 04083 CARCINOGENIC OXYGEN COMPOUNDS 47 with potassium hydroxide in ether (-5) to give the df-epoxide, yield 75 percent. The product was purified by chromatography on activated alumina and eluted with petroleum ether (bp 30-60)-ether (1:1) to remove halogen-containing impurities. The epoxide was distilled, bp 144/ 760 mm, n? 1.4289 [reported bp 1407760 mm (45)]- The vapor phase chromatogram on a 2 m polypropyleneglycol column, 125, 20 psi gave one peak only, retention time 14.0 minutes. Infrared absorption (liquid film): 3070, 3010, 2942 (C-H), 1255, 1245 (doublet), 917, and 840 cm'1 (<^--p). Bands at 1449, 1198, 970, and 780 cm-' were characteristic of this isomer and were absent in the infrared spectrum of*the meso-isomer. To confirm the identification of the df-epoxide, d2-l,4-dipiperidino-2,3- dihydroxybutane hydrochloride was prepared from it (45), mp 263.5 to 264.5* [reported mp 258-259* W5)]. Analysis: Calculated for CuH,Cl2NaOa: C, 51.11; H, 9.19. Found: C, 51.03; H, 9.15. The free base was liberated from the hydrochloride and crystallized from ether, mp 62 to 63* [reported mp 62* (45)]. me80-i,8,8,4-Diepoxybutane.--Erythritol was converted to meso-1,4- dibromo-2,3-dihydroxybutane by treatment with 47 percent hydrobromic acid; the dibromide was crystallized from chloroform, mp 137, yield 10 percent. The dibromide was converted to 7neso-l,2,3,4-diepoxybutane by treatment with potassium hydroxide in ether (45)', yield 65 percent, bp 140/760 mm, Ed 1.4272 [reported bp 37/ll mm (-47)). The purity of the material was established by vapor phase chromatography. It showed, under the conditions specified for the df-isomer, only one peak with a retention time of 12.7 minutes. Infrared absorption (liquid film): 3070, 3005, 2930 (C-H), 1260, 1250 (doublet), 915, and 845 cm"1 Bands at 1030, 1000, and 731 cm*1 were characteristic of tlus isomer and were absent in the infrared spectrum of the dMsomer. The meao-diepoxide was also prepared, in higher over-all yield, from butadiene via trans-1,4-dibromo-2-butene (48), irarw-1,4-diacetoxy-2butene (47), <ran-l,4-dihydroxy-2-butene (47), and meso-1,4-dihydroxy2,3-dibromobutane (47). The derivative roeso-l,4-dipiperidino-2,3-dihydroxybutane was pre pared as preriouly described (45), mp 108 [reported mp 106* (45)\. Analysis: Calculated for CuHjaNaOs: C, 65.68; H, 11.03. Found: C, 65.38; H, 10.97. Diketene.--Commercial quality diketene (K & K Laboratories) was purified by distillation, bp 126/760 mm [reported bp 125-1287760 mm (49)]. The vapor phase chromatogram on a 2 m silicone column, 40, 2.9 psi, Bhowed a single peak with a retention time of 9.7 minutes. y-Butyrolactone.--Commercial quality -y-butyrolactone (K & K Labora tories) was purified by vacuum 'distillation, bp 37*/1.0 mm [reported bp 89712 mm (50); n" 1.4328, reported n" 1.4348 (51)]. The vapor phase chromatogram on a 2 m polypropyleneglycol column, 175*, 20 psi, showed a angle peak with a retention time of 9.6 minutes. VOL. 31, NO. 1, JULY 1963 / / 48 VAN DUUREN BT AL. 1 -Vinylcydohezene-3.--Commercial quality (K & K Laboratories) material was purified by removal of autoxidation products with aqueous ferrous sulfate followed by distillation in a nitrogen atmosphere, bp 1287760 mm, n? 1.4593 (reported n" 1.4629 (52)1 The material gave one peak in the vapor phase chromatogram, 2 m silicone column, 100, 5 psi, retention time 3.17 minutes. 1-Hydroperoxy-1-vinylcyclohexene-3.--This material was prepared by the previously described procedure (53) as a 0.5 percent solution in 1 -vinylcyclohexene-3. Lauroyl peroxide.--Commercial quality (Thompson Chemical Co., Pawtucket, R.I.) material was purified by repeated crystallization from chloroform-petroleum ether (bp 30 to 60) at 4, mp 49 [reported mp 49 (54)}. Benzoyl peroxide.--Commercial quality (Matheson, Coleman & Bell) material was purified by repeated crystallization from chloroformpetroleum ether (bp 30 to 60) at 4, mp 106.6 to 107.6 [reported mp 104-106 (55)1 Solvents.--Reagent grade benzene and acetone were used as solvents. Biological testing methods.--Male Swis3-Millerton mice, 30 to a group, were used. The animals were approximately 8 weeks old at the beginning of the experiments and were painted 3 times weekly with an artist's watercolor brush, which delivered approximately 100 mg of solution per application. The few exceptions to these conditions are noted in tables 3 and 4. The entire backs of the animals were painted and the hair clipped when necessary. The concentrations and solvents used were determined by the toxicity and solubility of the materials. Four types of control groups were included in the test: 1) groups that received 3 paintings per week of 100 mg of benzene only; 2) groups that received 3 paintings per week of 100 mg of acetone only; 3) positive control groups that received 100 mg of a bonzo[a]pyrene solution in benzene or acetone containing 100 ppm (0.01%) of the carcinogen; 4) groups receiving no treatment. All tumors were excised at death and confirmed microscopically. RESULTS The tumor responses obtained with 14 compounds are given in tables 3 and 4. The median survival time shown in the tables gives an indica tion of longevity of the group. The total tumor and malignant tumor indexes are values for relative tumorigenic activity defined as 10,000 times the reciprocals of the computed time in days to produce tumors in 50 percent of the mice. The nominal or computed times to produce tumors in 50 percent of the mice were calculated by a `life-table" analysis which takes into account both the incidence of tumors and survival of the mice; this procedure is explained in an earlier publication (1). As an example, if the computed time to the presence of tumors, benign or malignant, in JOURNAL OF THE NATIONAL CANCER INSTITUTE waoff URL 04085 CARCINOGENIC OXYGEN COMPOUNDS 49 Table 3.--'Mouse skin painted with epoxides* Material Concen Cumulative Tumor tadexf tration Median No. of mice solvent time Tumor Cuoer cant Hexacpoxyaqualene, XXV rcm o o 1% 253 1 0 <10 <10 Ace o "j tone -bH(CH)r-d----fcH--CHr-l" l_CH CHt Ji ,lQ-poxy*teerlc odd, XXVI O CHi(CHi)s--C^----"CH--(CHi)rCO OH 1%Ace 165 3 0 23 <10 tone V,i0,12,13-Diepory?tearlcecld, XXVn O0 1% 196 4 0 20 <10 Ace tone CHifCHiJr--C^--Cfi--CHi--C^--CH--(OHi)rCOOH Styrene oxide, XXVIU cC^CH-^ ^ 10% 431 3 1 14 <10 Bensene l-Etbyleneoxy-a.tepoxycyclobexuie, XXIX C^r--CH-^ 10% lit 14 9 24 22 Bentene l,3-Epoxybutane-3, XXX O C^t---CH--CH=*CHi dM^.i-Dlepoxybutene, V No sol 237 3 1 10 <10 vent 10% 78 2 1 26 16 Ace tone wro-lAS.VDkporybQUoe, VX 10% 154 4 23 Ace tone "ThtrtT male Swiss-MJUerton mice painted 3 times weekly, 100 me per painting. t Calculate* from tine (or total tumor ladoetlon or/and cancer induction In 50 percent of animals adjusted for ofrtered survival ($c text). 50 percent of the mice were 100 days and cancers in 50 percent of the mice were 200 days, this substance would have a "total tumor index" of 100 and a "malignant tumor index" of 50. Thus with substances of low tumorigenic activity, the computed time to 50 percent incidence will be longer than the lifetime of the mice. At the concentration levels of l-vinylcyclohexene-3 and its hydroperox ide used there was extensive skin damage and slight to moderate skin damage with lauroyi peroxide, <#-l,2,3,4-diepoxybutane, and styrene oxide, but none with any of the other compounds. Both H~ and meso1,2,3,4-diepoxybutane showed pronounced toxicity. The notable toxicity of these materials in acetone has been reported by earlier workers (5, IS) and the high mortality may account for the negative carcinogenicity re- VOL. *1, NO. 1, JULY 1963 50 VAN DUUREN ET AL. Table 4.--Mouse 6kin painted with lactones, hydroperoxides, and peroxides* Manorial LACTONES Diketene, t XXXI CHi--C=C Ht o-i--oJ 'j-ButyroUctone, XXXII CHi :Ej O Concentration and solvent Cumulative No. Median of mice survival time Tumor Cancer Tumor fudext Total Malig nant 10% Acetone 61 0 0 <10 <10 lO%Benteu 203 2 I <10 <i0 HYDROPEROXIDES l-VliJyleyclobexww'3 (control). ) XXXUI <0 HCHt 50% Bem-eo* 076 l-Hydroperory-l-rinylcydobexene- 0j% Beni*o^l-vinyl- a, xxxrv cyclohexeoe-i (VU Same D Sam* <zx CH=CHt OOH PEROXIDES _ Lauroyl peroxide, ZXXv' [ O" CH(CHi)i^--0----_|j Benzoyl peroxide, XXXVI O ^-0 6% Benzene 6% Bentene 166 403 437 303 6- 1 8a 62 10 10 la 10 38 IS 17 <10 <10 <10 <10 <10 URL 04087 Thirty male Swlss-MUJerton mioe painted 3 times weekly, 100 mg per painting, except where otfeerwizo Indicated. fCaleulated from estimated time for total tumor Induction and/or cancer Induction to 60 percent of animals adjusted for observed survival (in text), jpifty CAFi, Jackson Laboratory female mice. 1Forty-five mg per pelntlnaNineteen animals. suits obtained in the mouse skin experiments byHendry and co-workers (5). Moreover, as 10 percent solutions In benzene the materials were very toxic, and a 50 percent or higher mortality resulted within hours of pain ting. But 10 percent solutions in acetone were less toxic; in a group of 30 animals receiving a 10 percent solution of the (//-isomer 3 times weekly the median survival time was 78 days. The results obtained in the control experiments are shown in table 5. DISCUSSION During the period in which the compounds listed in tables 3 and 4 were tested, an extensive series of control groups were maintained. The results of these experiments (table 5) merit discussion. Of the 150 ani- JOURNAL OF THE NATIONAL CANCER INSTITUTE CARCINOGENIC OXYGEN COMPOUNDS 51 Table 5.--Painting of mouse skin, controls Treatment t Number of animals $ Median survival time Cumulative Noof mice Tumor Cancer Tumor index* Total Malig nant Benzene Acetone Benrolalorrene. 100 ppm, acetone Benzo[o}pyreoe, 100 ppm, benzene No treatnent 30 30 30 60 30 3011 30 30 30 30!| 30 30 30 30 30 30 30 30fl 30 B 30 23 29 60 264 262 412 292 240 652 330 134 211 378 240 259 351 348 370 175 342 730 624 217 112 253 345 20 10 <10 50 19 <10 2 0 <10 <10 2 15 <10 <10 20 10 <10 0 0 <10 <10 40 14 <10 20 18 <10 16 7 50 32 24 20 34 29 25 11 45 27 18 11 45 36 16 7 10 6 42 27 24 21 23 13 36 26 2 is 14 <10 0 0 <10 <10 0 0 <10 <10 0 0 <10 <10 50 20 <10 1 0 <10 <10 4 0 <10 <10 1 0 <10 <10 * Calculated from estimated time for total ramor induction and eaaoer induction Is SO percent of animal* adtuned for observed sarvtal (in text). t Time times weekly, 100 ms per painting. t Male, Swlu-MlUerton mioe, except where Indicated otherwise. ISertext. II /acfcaoa Laboratory, Bar Harbor, Maine, female CA.Fi mioe. mals treated with benzene, 6.7 percent developed tumors. One of 11 tumors* was a cancer; the remainder were papillomas. The cancer appeared on the left front leg, remote from the site of application, in an animal from a group in which much biting occurred. The tumor appeared in the area of the bites. The 120 acetone-control animals showed a 6 percent incidence of papillomas and no cancers. In the 3 positive control groups, in which benzo[c]pjTene with benzene as solvent was used, 49 of 90 animals developed tumors--26 were cancers. In the 4 positive control groups in which acetone was used as solvent, 83 of 120 animals developed tumors--49 were cancers. It should be noted that there were considerable variations in gross tumor incidence and survival time for the positive controls. Neverthe less the tumor indexes, both total and malignant, showed good agreement, e.g., in the benzopyrene (100 ppm, benzene) group the number of animals with cancer ranged from 6 to 13 in groups of 30 animals. The malignant tumor indexes in these groups were 21 to 27. In the untreated control series 267 animals were used; five percent developed tumors, 1 of which was a squamous cell cancer. The median survival time ranged from 112 to 643 days. In *0 these experiments tbe papillomas were of the benign squamous cell type sad tbe cancers were predomi nantly squamous cell carcinomas. In this "Discussion," turnon refer to total tumors, Lt., Including benign papillomas and squamous cell cancers. Many animals bore multiple tumors. VOL. 31, NO. 1, JULY 1963 / URL 04089 52 VAN DUUBEN* ET AL. During part of the test period ectromelia was present in the colony, in experimental and control groups. Subsequently all mice were success fully vaccinated against this disease. During the past year the background incidence of tumors in the un treated control groups in which Swiss-Millerton male mice were used was 1.4 percent. These tumors were all squamous papillomas. Tumors were not observed, however, in the Jackson Laboratory CAFt control female mice nor in 1 experimental group painted with diketene (table 4). A 2 percent benign skin-tumor incidence was reported by Jackson Labora tories for the similar strain C3H mice (66). In the present work l-ethyleneoxy-3,4-epoxycyclohexane exhibited pronounced activity, and this result confirms the findings of Walpole (15). Tests on the two isomers of 1,2,3,4-diepoxybutane were unsatisfactory because of poor survival of the mice. One cancer was observed in the group receiving dMsomer. Only 4 animals survived for longer than 8 months and after 12 months there were no survivors. In the group receiving meso-isomer 10 animals survived more than 8 months; 4 cancers were observed. New tests are currently under way with these 2 compounds. Included in the series of epoxides tested in the present work are 3 monofunctional epoxides, 1 of which, l,2-epoxybutene-3, showed a low order of activity. Five diepoxy compounds and hexaepoxysqualene were examined and 3 of these 6 showed activity. From these and earlier results (5, IS, SO) it appears that a vicinal diepoxide structure, such as occurs in the isomers of diepoxybutane, is not required for carcinogenicity in this series; in the most potent compound, l-ethyleneoxy-3,4-epoxycyclobexane, the two epoxide functions are separated from each other by ring carbon atoms. It does appear, however, that specific structural features in addition to one or more epoxide functions are required for carcinogenicity, and this structural specificity may be associated with physical properties as well as chemical reactivity. Although the present work has been restricted to date to oxygenated functions, carcinogenicity has been observed earlier in the related ethyleneimines (IS), e.g., stearoylethyleneimine, IX, CH,(CH,)wCO--N IX In this series, also, bifunctionality was not a prerequisite for carcino genicity. From the results obtained in this work and those summarized in tables 1 and 2, it can be concluded tentatively that strained lactone rings may favor carcinogenicity. 7-Butyrolactone, the most stable lactone in the series, has a low order of activity. Dickens and Jones (7) found no activity for this material by subcutaneous injection in rats. In the strained ring systems, 0-propiolactone and the unsaturated 5-membered-ring lactones JOURNAL OF THE NATIONAL CANCER INSTITUTE CARCINOGENIC OXYGEN COMPOUNDS 53 (table 1), more pronounced activity is observed. Reactivity and ring strain obviously cannot be the only requirements for carcinogenicity, as evidenced by the complete lack of activity observed for diketene, a highly reactive compound. Fieser's reported carcinogenicity for 6-hydroperoxy-A*-cholestene-3-one (5) prompted an examination of other hydroperoxides. The hydro peroxide of butadiene dimer (l-vinylcyclohexene-3) showed definite carcinogenic activity in 2 experiments. The hydrocarbon itself, in which the hydroperoxide was dissolved, gave 1 cancer on prolonged painting. The possibility was not excluded that the hydrocarbon contained a minute amount of the hydroperoxide formed by autoxidation, especially since the hydrocarbon is readily autoxidized in air (S3). This possibility is currently being examined in a new test that utilizes the hydrocarbon freed from autoxidation products by treatment with ferrous sulfate before use. Neither of the 2 peroxides showed carcinogenic activity; 2,4-cholestadiene peroxide (S3) and 7,12-dimethyl-7,12-peroxybenz[a]anthracene, (SI, 36) also did not show any activity when tested subcutaneously in mice. But the corresponding epoxide 7,12-dimethyl-7,12-epoxybenz[<z]anthracene did elicit skin tumors in mice 1(32), table 1']. The present work has been extended to include many structures con taining the functional groups of interest. These results may facilitate a re-examination of the structural types of importance and a more detailed discussion is deferred until then. REFERENCES (I) Pallies, E. D., Orris, L., and Nelson, N.; Skin irritation and skin tumor production by beta propiolactone (BPL). Amer 2nd Hyg Ass J 23: 257-264, 1962. () Roe, F. J. C., and Glendenning, O. M.: The carcinogenicity of ^-propiolactone for mouse skin. Brit J Cancer 10: 357-362, 1956. (3) Hendry, J. A., Homer, R. F., Rose, F. L., and Walpole, A. L.: Cytotoxic agents. II. Bis-epoxides and related compounds. Brit J Pharmacol 6: 235-255, 1951. (4) McCammos, C. J., Kotin, P., and Falk, H. L.: The carcinogenic potency of certain diepoxides. Proc Amer Ass Cancer Res 2: 229-230, 1957. (5) Fxeseb, L. F., Gbeene, T. W., Biscbofp, F., Lopez, G., and Rupp, J. J.: A carcinogenic oxidation product of cholesterol. J Amer Chem Soc 77: 39283929, 1955. (6) Fearnlt, C., and Spearman, J. B.: Cross-linkage formation in keratin. Nature (Lond) 166: 743-744, 1950. (7) Dickens, F.r and Jones, H. E. H.: Carcinogenic activity of a series of reactive lactones and related substances. Brit J Cancer 51: 85-100, 1961. (5) Armstrong, M. D., and Lewis, J. D.: Thioether derivatives of cysteine and homocysteine. J Org Chem 16 : 749-753, 1951. (3) Maualis, P., McHale, D., and Green, J.: Some derivatives of S-carboxyalkylcysteines and -homocysteines. J Chem Soc 2906-2914, 1960. (10) Hadoow, A.: Chemical carcinogens and their mode of action. Brit Med Bull 14: 79-92, 1958. (II) Auerbach, C.: Mutagenic effects of alkylating agents. Aon NY Acad Sci 68: 731-749, 1958. VOL. 81, NO. 1, JULY 1968 / 54 VAN DUUftEN ET AX. (It) Stacey, K. A., Cobb, M., Cocsens, S. F., and Alexander, P.: The reactions of the radiomimetic alkylating agents with macromoleculee, in vitro. Ann NY Acad Sci 68: 682-701, 1958. (13) Walpole, A. L.: Carcinogenio action of alkylating agents. Ann NY Acad Sci 68: 750-761, 1958. (14) Ross, W. C. J.: In vitro reactions of biological alkylating agents. Ann NY Acad Sci 68: 669-681, 1958. (15) Trams, E. G., Nadkarni, M. V., and Smith, P. K.: On the mechanism of action of the alkylating agents. I. Interaction of alkylating agents with nucleic acids. Cancer Res 21: 660-666, 1961. (15) : On the mechanism of action of alkylating agents. II. Effects of nitrogen mustard, Myleran and X-radiation on nucleic acid biosynthesis. Cancer Res 21: 567-570, 1961. (17) Lawlet, P. D., and Brooks, P.: The possible correlation of the alkylation of DNA and subsequent degradation of alkylated DNA with the cytotoxic, mutagenic and carcinogenic properties of the alkylating agents. Brit Emp Cancer Camp Ann Report part II, 8-13, 1960. (18) Boyland, E., and Wolp, G. Metabolism of polycyclic oompounds. Biocbem J 47: 64-69, 1950. (19) Williams, R. T.: Detoxication Mechanisms. New York, J. Wiley & Sons, 1959, p. 270. (SO) --------------Ibid; p. 232. (1) Allsopp, C. B,: A water soluble carcinogen derived from 3;4-benzpyrene. Nature (Lond) 150: 548-549, 1942. (53) Badger, G. M.: Structures and Reactions of the Aromatic Compounds, Photooxidation and Photoniimerization. London, Cambridge Univ. Press, 1964, chapt. 9, p. 364. (23) Kotin, P., and Falk, H. L.: Production of tumors in C57BL mice with atmoephere-extracted aliphatic hydrocarbons. Proc Amer Asa Cancer Res 2: (No. 1) 30, 1955. (54) Van Duuren, B. L.: Some Aspects of the Chemistry of Tobacco Smoke. In Tobacco and Health. Springfield, 111., Charles C Thomas, 1962, chapt. III. (So) Dugan, L. R., Beadle, B. W., and Herrick, A. S.: An infrared study of autoxidized methyl linoleate. J Amer Oil Chem Soc 26 : 681--685, 1949. (S6) Pbivett, O. S., Lundbebg, W. O., Khan, N. A., Tolberg, W. E., and Wheeler, D. H.: Structure of hydroperoxides obtained from autoxidized methyl linoleate. J Amer Oil Chem Soc 30: 61-66, 1953. (7) Booth, J., and Boyland, E,: Metabolism of polycyclic compounds. Biochem J 44 : 361-365, 1.949. (28) Boyland, E., and Levi, A. A.: Metabolism of polycyclic compounds. II. Production of dibydroxydihydroanthracene glycuronie acid from anthracene. * Biochem J 30: 728-731, 1936. (9) Boyland, E., and Horning, E. S.: The induction of tumours with nitrogen mustards. Brit J Cancer 3: 118-123, 1949. (SO) Rose, F. L.r Hendry, J. A., and Walpole, A. L.: New cytotoxic agents with tumour-inhibitory activity. Nature (Lond) 165 : 983-996, 1950. (37) Badger, G. M., Cook, J. W., Hewett, C. L., Kennawat, E. L., Kennaway, N. M., Martin, R. H., and Robinson, A. M.: The production of cancer by pure hydrocarbons. Proc Roy Soc, ser B 129: 439-467, 1940. (S) Salaman, M. H.: The use of cocarcinogens in the study of carcinogenesis. In Ciba Foundation Symposium on Carcinogenesis (Wolstenholme, G. E. W., and O'Conner, M., eds.). Boston, Little, Brown & Co., 1959, pp. 70-82. (S3) Hike, C. H., Guzman, R. J., Coursey, M. M., Welunoton, J. S., and Anderson, H. H.: An investigation of the oncogenic activity of two representative epoxy resins. Cancer Res 18: 20-26, 1958. JOURNAL OF THE NATIONAL CANCER INSTITUTE URL 04091 CARCINOGENIC OXYGEN COMPOUNDS 55 (34) Deichmann, W. B., Hiaose, R., and Witheruf, 6.: Observations on the effects of gamma-v&lerolactone upon experimental animals. J Ind Hyg and Toxicol 27: 263-268, 1945. (35) Butenandt, A., and Kvdssus, H.: A*<-Cholestadiene and its photochemical transformation. Z Physiol Chem 253: 1-111, 1938. (36) Coox, J. W., and Martin, R. H.: Polycyclic aromatic hydrocarbons, part XXIV. J Chem Soc 1125-1127, 1940. (87) Van Dttuben, B. L., and Schmitt, F. L.: Epoxidation and cyclization of squa- lene. J Org Chem 25: 1761-1765, 1960. (SS) Swern, D., Pi nolet, F. W., and Scalan, J. T.: Epoxidation of oleic acid, methyl oleate and oleyl alcohol with perbenzoic acid. J Amer Chem Soc 66: 19251927, 1944. (39) Swern, D., and Dickel, G. B.: Chemistry of epoxy compounds. XV. Epoxida tion of linoleic acid with peracetic and performic acid. J Amer Chem Soe 76: 1957-1958, 1954. (40) Goltjubic, C., and Cottle, D. L.: The reaction of styrene oxide with methylmagnesium iodine. J Amer Chem Soc 61: 996-1000, 1939. (41) Everett, J. L., and Kon, G. A. R.: The preparation of some crtotoxic epoxides. J Chem Soc 3131-3135, 1950. (42) Kadesch, R. G.: Reaction of 3,4-epoxy-1-butene with methanol. Direction and mechanism of ring opening. J Amer Chem Soc 68: 41-45, 1946. (4$) Yalxttb, A.: New compounds related to ethyleuic cis-tram isomerism: 2-butene* 1,4-dioL Ann Chim (12), 3: 644-678, 1948. (44) Fxbhbein, L., and Gallaoban, J. A.: The preparation of cis and lrans-1, 4dinitroxy-2-butene. J Amer Chem Soc 78: 1216-1220, 1956. (45) Bxxcb, W. F.: Some reactions of 1:2:3:4-diepoxybutane. J Chem Soc 24832487, 1951. (49) PoiuceREB, R., and Reindel, W.: On the oxides of isoprene and butadiene. Chem Ber 66B: 335-339, 1933. (47) Prevobt, C.: A closed reaction cycle of some erythrine derivatives. C R Acad Sci (Par) 183: 1292-1295, 1926. (48) Srantz, E. M.: Synthesis of compounds related to vitamin A from hydroxymethylene-A-ionone. J Amer Chem Soc 68: 2553-2557, 1946. (4ff) Hurd, C. D., and Williams, J. W.: Ketene and acetylketene. J Amer Chem Boo 58: 962-968, 1936. (50) Sircar, 6. S. G.: The influence of groups and associated rings on the stability of certain heterocyclic systems. IV. The substituted butyro- and valerolactones. J Chem Soc 898-903, 1928. (51) McKxnlet, C., and Copes, J. P.: Some properties of purer-butyrolactone. J Amer Chem Soc 72: 5331, 1950. (/*) Abchan, O.: On two homologs of diprene. Chem Ber 57B: 1959-1962, 1924. (58) Brill, W. F.: Preparation of hydroperoxide by the autoxidation of 4-vinyleydohexene. J Org Chem 24: 257-259,1959. (64) Beattoet, C.: Synthesis of new derivatives of 2-hydroxy-l,4-naphthoquinoiie. Helv Chim Acta 30: 1804-1807, 1947. (65) DxTab, D. F., and Carfino, L. A.: A new method for the preparation of diacyl peroxides. J Amer Chem Soc 77: 6370-6371, 1955. (66) Ruuel, E.: Personal communication. VOL. SI, NO, 1, JULY 1963 Mtilllrbi iu\a. .iW V**.- .' .-> Carcinogenicity of Epoxides, Lactones, and Peroxy Compounds, IV. Tumor Response in Epithelial and Connective Tissue in Mice and RatsLS 8. L VAN OUUREN, L LANGSETH, L. ORRIS, G. TEEBOR,3 N. NELSON, and M. KUSCHNER,3 institute of En vironmental Medicine and Department of Pathology, New York University Medical Center. New York, New York 10016 SUMMARY--Hie carcinogenic potencies of a series of alkylating and peroxy compounds were compared by the use of various routes of adminis tration in mice and rats. Twelve of these compounds, earlier tested on mouse skin, were tested by one or more of the following routes: subcutaneous injection in mice and rats and infrogastne feeding in rats. These experi ments showed that those compounds lumorigenic to mouse shin are also active by other routes. These compounds ares -propiolaetone, 0-butyrolaetone, dJ-diepoxybutane, glycidaldehyde, end 1,2,5,6-dicpoxyhexane. Those compounds inactive on mouse skin are weakly active or inactive by other routes. These compounds are: cumene hydroperoxide, 9,10-epoxystearie acid, diketene, 2^/Ltrimethylhydroxypentenoic acid-0-Iactone, cyclohexenehydroperoxide, and lauroyl peroxide. -Propiolactone and 0-butyrolacfone are mouse-skin carcinogens and induce gastric cancers when fed repeatedly to rats/ however, 3 epoxides carcinogenic to mouse skin (glycidaldehyde, d,Ldicpoxybufane, and 1-efhyIeneoxy-3,4-epoxy cyclohexane) do not induce gastric cancers in rats. This finding is related to the rapid acid-catalyzed hydrolysis of epoxides in the rat stomach. The compounds tested induced tumors mainly at the site of application and did not give any significant incidence of tumors at sites distant from the site of application. These experiments indicate that connective tissue of rats is not unusually sensitive to tumor induction by these compounds and that subcutaneous injection in the rat is a useful system for the evaluation of potential carcinogens.--J Nat Cancer Inst 37: 825--838,1966. CARCINOGENIC ALKYLATING agents of the epoxide and /3-lactone types have a low order of potency when compared to the classical aromatic hydrocarbon and heterocyclic carcinogens, and tumorigenic dosages of these alkylating agents are frequently close to their toxic dosages (/, 2). With a few exceptions, the assay of these compounds has 1 Received June 6, 1966. * Supported by Public Health Service grants CA-05946 and CA-06989 from the National Cancer institute and Public Health Service grant ES-00014 from the Bureau of State Services. * Department of Pathology. 825 826 VAN DUUREN ET AL. been carried out by skin application in mice, and for rats. Materials for gastric intubation in rats species specificity and tissue specificity have not were in solution in tricaprylin and administered been explored. Also, tumors distant from the site with a #8 rubber catheter and syringe. Some of application have usually not been observed. materials were tested by single intubations and In view of our continuing interest in the carcino others by repeated weekly intubations. Animals genicity and mechanism of action of epoxides, were examined for palpable tumors at regular lactones, and peroxy compounds, a selected group intervals, and their condition was recorded once of materials, including some that are carcinogenic a month. They were treated and observed for and others that are inactive on mouse skin, were their lifespan, or until the predetermined end of tested. The compounds were tested by subcuta the experiment as noted under Results; thn^ neous injection in mice and rats at various dosages having massive tumors were killed. Whenever pov and by single and repeated intragastric feeding in sible, animals were completely autopsied at death rats. The results of these tests are described in the and their tumors excised. All diagnoses were con present report and are compared with earlier results firmed histologically. Control groups included ani dealing with bioassay of these materials by skin mals treated with, tricaprylin only, as well as groups application in mice and other routes of administra receiving no treatment. tion in mice and rats. The usefulness of these various bioassay procedures is discussed here, par ticularly since considerable discussion has been RESULTS related to the reliability of subcutaneous injection, Tables 1, 2, and 3 show the results of a series oi especially in rats, for bioassay of potential carcino tests in which mice and rats received subcutaneous gens. injections, at various dosages and intragastric feed ing, of 12 compounds earlier tested on mouse skin MATERIALS AND METHODS (1, 2). The duration of all experiments is indicated in the tables; the times to first tumor appearance Chemicals.---The preparation and purification of are given in table 1. These are the approximate all compounds used for bioassay in this study were times to the appearance of palpable masses. described in detail in earlier papers in this series (A 2). Animals.--The rats were female Eastern Sprague- Subcutaneous Injection in Mice Dawley (Blue Spruce Farms, Altamont, N.Y.); In the first series of experiments (table I), mice. they were at least 6 weeks old and weighed 120-125 50 to a group, were given weekly subcutaneous g when tested. The mice were of an ICR/Ha injections of 0.1 mg of the compounds. The me Swiss strain (Millerton Research Farms, Millerton, dian cumulative dose per animal ranged between N.Y.); 8-week-old females were used throughout. 6.5 and 7.8 mg and the median survival with most All mice were vaccinated against ectromelia at age test compounds was 15 months or more. A few be r 6 weeks. Mice and rats were fed Purina Labora tory Chow and water ad libitum. The animal rooms nign tumors were observed. Animals treated with tricaprylin only bore local injection granulomas. were maintained at 22-24.5 C. All compounds The tumor response of the mice to the car were in solution in tricaprylin (Eastman Kodak cinogen was low in most cases. rf.l-Dtepoxybu- Company, Rochester, N.Y., except where noted in tane was the most active and its injection resulted tables). Where there were solubility limitations, in 5 animals with fibrosarcomas and 2 wiih solutions were wanned to 37 C; otherwise, they adenocarcinomas at the site of injection. Lauroyl were administered at room temperature. Animals peroxide showed some activity, causing 3 local were given subcutaneous injections once weekly in fibrosarcomas. /3-Butyrolactone induced 4 local the left axillary area; a %-inch, 26-gauge needle malignant tumors (2 fibrosarcomas and 2 ana was used for injections in mice and a X-inch, 23- plastic sarcomas). All other compounds in this gauge needle for rats. The compounds were dis group induced 1-3 local malignant tumors m solved in 0.05 ml tricaprylin for mice and 0.1 ml each group. JOURNAL OF THE NATIONAL CANCER INSTITUTE r- s o CARCINOGENIC EPOXIDES, LACTONES, AND PEROXIDES 827 Table 1.--Subcutaneous injection in mice Compound Dose, mg In 0.03 ml trlcaprylin sntmnle Median time (days) Duration Month* to Snt (daa) tumor Animal* with turnon at Injection site Benign Malignant* * 3 Adenocarcinomas 1.1 30 338 401 8 None.................... 5 Fibrosarcomas 0.1 30 1 Anaplastic sarcoma 1.1 30 387 533 1# Non*.................... 3 Fibrosarcomas I Sqoamons celt carcinoma 0.1 10 30 301 530 _ None.................... None ' 3.3 2,2,4-Trimetbylbydroxy. pentenoic acid-fS-tactane. 10 10 30 473 530 30 2S3 490 30 474 534 1 Squamoo* cell carcinoma 1# 1 Papillary tumor 3 Fibrosarcoma# 1 Squamous cell carcinoma 1 Undifferentiated sarcoma 3 Adenocarcinomas 3 Anaplastic sarcomas 2 Fibrosarcomas s None.................... 15 FIbroearcomaa 2 Squamous cell carcinomas 1 Adenocarcinoma None.................... Non* 10 (0.1 ml vehicle) 30 331 533 None.................... None 3.3 0-Propiolaetciie....................... 0.73 30 37# 533 30 270 503 18 None.................... 1 Fibrosarcoma 7 3 Squamous 9 Fibrosarcomas papilloma*. 3 Adenocarcinomas G Squamous cell carcinomas 4- URL 04095 :;,D-Dlinetb;lbem[4) 4. anthracene. on 0.01 so 141 30 184 0.001 30 332 > IT* 0.03 ml ___ ** ^-*Ail sdanocarctaom&s were d breast origin. 90 30 60 30 53J 484 431 k # Some of these same compounds were subt*quently tested in mice at doses from 10-100 tunes those used in the first series. These dosages **ere as high as possible and were limited only V toxicity and skin effects. In this second series, *bo given in table 1, 30 mice per group received from 1.1-10 mg/injection; the exact doses are liven in table 1. Tumorigenic activity increased higher dosages with most of the compounds viuch showed activity at lower dosages. Several compounds which were weakly active or inactive ** lower dosages were markedly active at higher 153 4 None.................... 28 Fibrosarcomas 4 Squamous cell carcinomas 251 C None.................... 20 Ffbroearoomast l Squamous carcinoma 530 13 None.................... 4 Fibrosarcomas 4 Undifferentiated sarcomas 1 Reticulum cell sarcoma 5S1 ___ 533 __ None.................... None 599 ___ None fOne animal bod metastasis to the liver (flg. 6). dosages. For example, /S-butyrolactone admin istered at 0.1 mg/injection resulted in an 8% tumor response, but at the 10 mg/injection level exhibited a 60% incidence of local malignant tumors. Similarly, glycidaldehyde, which gave a 6% tumor response at the 0.1 mg/injection level, showed a 23% incidence of local tumors at the 3.3 mg/injection level. A local squamous cell carcinoma from this group is shown in figure l. In the lauroyl peroxide-treated group, 3 of 50 animals had malignant tumors at a low dose, but no tumors were induced at the higher dose ?0L. 37, NO. 6, DECEMBER 1966 828 VAN DUUREN ET AL. level, possibly due to the toxicity and hence lower survival at this dosage, as shown in table 1. With 0-propiolactone, 3.3 mg/animal was first used but severe chronic skin damage at the site of injection necessitated a decreased dose. A satisfactory dose was obtained at 0.73 mg/mouse/ injection. At this level 18 local malignant tumors were induced. Cumene hydroperoxide induced 1 fibrosarcoma at the injection site; cyciohexenehydroperoxide, diketene, 2,2,4-trimethylhydroxypentenoic acid-/3lactone, and the vehicle control group did not cause any tumors, benign or malignant, at the injection site. As a positive control experiment, 7,12-dimethylbenz[fl]anthracene (DMBA) was subcutaneously injected in mice and tested at 3 dosage levels to establish a level which gave a tumor response corresponding to that obtained with the weakly active epoxides, lactones, etc. The results (table 1) indicate that a suitable dose in mice is 1 /ig/ injection, once weekly. Subcutaneous Injection in Rats In a companion experiment, the same series of compounds were tested at a weekly dose of I mg/animal by subcutaneous injection in rats. The results of these tests are given in table 2. At this level, significant activity was observed in animals receiving d,/-diepoxybutane, in which 10 local malignant tumors were induced, with a median cumulative dose of 67 mg. This compound is currently on test at a dose of 2 mg/injection After I year of treatment, 7 of the 20 rats bort malignant tumors; the first tumor was recorder after 10 months of testing. 1,2,5,6-Diepoxyhexane, 2,2,4-trimethylhydroxy pentenoic acid-0-lactone, and giycidaldehydc eacf induced I malignant tumor at the injection site whereas 9,10-epoxystearic acid, -butyrolactonc and cyciohexenehydroperoxide did not induce any /3-Butyrolactone is currently on test at a dose c 100 mg/injection in 20 rats. At this dose, 7 loci malignant tumors, one of which is shown in figur 2, have been confirmed to date. These prelim nary data indicate that -butyrolactone is sarct magenic at higher dosages in rats. In the nt treatment control group, 1 animal had fibroadenoma of the breast. In the group reeeivir only tricaprylin, 1 animal had an adenocarcinon of the breast. Four rats in treatment groups bo local benign tumors, skin fibroadenomas; 3 of the tumor-bearing animals were treated with glycid; dehyde and 1 was treated with 2,2,4-trimethylh droxypentenoic acid-3-lactone. Table 2.--Subcutaneous injection In rata* Compound Median survival time (days) Duration of test (days) Anirrmla with malignant tumors a< injection sitef d,{-Diepoxybutane.......................... 1,2,5,6-Diepoxyhexane.................. 9,10-Epoxyateario acid.................. Glycidaldehyde............................... 0-Butyrolactone.............................. 2,2,4-TrLmetityliiydroxypentanoio lactone. Cyciohexenehydroperoxide........... 7,12-Dimethylbenzlalanthracene. Tricaprylin.. No treatment 471 >552 >551 545 >559 >564 543 >158 >565 554 550 1 Adenocarcinoma 9 Fibrosarcomas 552 1 Fibrosarcoma 551 None 553 1 Fibrosarcoma 559 None 564 1 Adenocarcinoma 565 None 153 20 Fibrosarcomas} 6 Rhabdomyosarcomas} 2 Undifferentiated sarcomas 1 Reticulum cell sarcoma 1 Adenocarcinoma 565 1 Adenocarcinoma 563 None 20 i-- -fc> O co Sprague-Dawley female rats, 60 per group, were tires weekly injeetIons oil mg of compound la0.1 m] trlcsprylla. (All adenocarcinomas were of breast origin JTwo animals bad metastasis to the lung; I also. |0ne animal bad metastasis to ttao hug. had a Abroad* JOURNAL OF THE NATIONAL CANCER INSTTT Compound CARCINOGENIC EPOXIDES, LACTONES, AND PEROXIDES Table 3.--Gastric feeding in rats* Number of animals Dose, mg/ml, tricaprylin Median Duration survival of test time (days) (days} Animals with gastric tumors 829 Animals with other tumors so to ia o 0-Propiolactone........... d,f'l>iepoxybutane.... 0-Butyrolaotone.......... Glycidaldehyde........... l-Ethyleneoxy-3,4- epoxycyclonexane. 7,12-Dimethylbens(aJ- anthracenef. Tricaprylin J................. 5 10/0.5, once weekly. 5 5/0.5, once weekly. 5 100/0.5, once weekly. 5 33/0.5, once weekly. S 100/0.5, once weekly. 6 50/1, single treatment. 5 0.5, once weekly... 426 342 426 329 500 357 525 487 3 Squamous cell Nona carcinomas of stomach. 363 None........................ None 492 3 Squamous cell carcinomas of stomacht- 492 None.................... None None 538 None........................ None 385 None........................ 2 Fibroadenomas 2 Adenocarcino mas of breast 1 Lymphoma 537 None........................ None Female Sprague-Dawley rata, age 51-57 days at beglnntnjoftreatment, were used exwpt wbere noted. (One of there animals also b*d metastasis to tha llwr. (Aolmals were 10C-12S days old at time of first treatment. {Obtained from E & E Laboratories, Jamaica, N Y. In the results listed in tables 1 and 2, the median survival times were good except in those experi ments where high tumor yields or toxicity influ enced the survival time. Jntragastrie Feeding in Rate A selected series of compounds were tested in rats by single or repeated gastric feedings. The results of some of these experiments are listed in table 3. DMBA was included in this series as a positive control; after a single feeding it induced several mammary tumors (table 3) but no stomach tumors. The other 5 compounds listed in table 3 were tested by repeated gastric feedings of rats beginning at 51-57 days of age, 5 animals per group. Of these, -propiolactone and -butyrolactone induced malignant stomach tumors. Of the 5 rats fed /3-propioIactone, 3 had squamous carcinomas of the stomach after about 1 year. Of the 5 rats fed /3-butyrolactone, 3 developed squamous cell car cinomas of the forestomach with marked basal cell hyperplasia and moderate atypia of the squamous epithelium of the stomach (fig. 3). One of these animals also had isolated metastasis to the liver (fig. 4) which was found when the rat was killed at 385 days, d, /-Diepoxybutane, glycidaldehyde, and l-ethyleneoxy-3, 4-epoxycyclohexane did not induce tumors in any animals. Tumors in Control Croups The incidence of tumors distant from the site of weekly injections of tricaprylin in mice and rats is given in table 4. Also included in this table is the tumor incidence in no-treatment control groups for both mice and rats. These are the same control groups listed in detail in tables 1 and 2. Both groups of mice had a substantial number of lym phomas, lung adenomas, and some reticulum cell sarcomas. Several tumors were also noted in the rat control groups. An unusual type of tumor observed in die no-treatment group of rats was an adenocarcinoma of the uterus. Tumors at Distant Sites in Test Groups Mice.--In experiments in which mice were given subcutaneous injections, tumors were observed at sites other than those of injection. Various such tumor types were observed, most likely due to the longevity of the test and control groups. In mouse test groups, lymphomas, lung adenomas, and retic ulum cell sarcomas of lymph nodes occurred as frequently as in the control groups listed in table 4. VOL. 37, NO. 6, DECEMBER 1966 / / 830 VAN DUUREN ET AL. Table 4.--Distant tumora in tricaprylin-control groups and tumors in no-treatment groups Treatment Duration of test (days) Number of animals Number of animals with tumors 532 519-599 565 563 110 8 Lung adenomas 7 Lymphomas 1 Breast adenosis 2 Reticulum cell sarcomas of lymph nodes. 89 8 Lymphomas 3 Lung adenomas 3 Reticulum cell sarcomas of lymph nodes 1 Malignant hemangiopericytoma 50 1 Reticulum cell sarcoma of lung 1 Lymphoma 1 Fibroadenoma 50 1 Lymphoma 1 Fibroadenoma URL 04098 In addition, other tumor types were observed. Thus injection of glycidaldehyde resulted in l ani mal with an adenocarcinoma of the breast; * in the group given epoxystearic acid, 1 mouse bore an anaplastic sarcoma on the right-side flank; in the group given 2,2,4-trimethylhydroxypentenoic acid0-lactone, 1 mouse bore a hepatoma and 1 an adenocarcinoma of the breast; in the group given lauroyl peroxide, 1 animal had an intraabdominal sarcoma with metastasis to the pancreas and stom ach. The cyclohexenehydroperoxide-treated group included 1 animal with breast tumors, an adeno carcinoma and a fibroadenoma; in the cumene hydroperoxide-treated group, 1 mouse had an adenocarcinoma of the breast. Rais.--In rats treated subcutaneously with the various compounds listed in table 2, some tumors appeared at sites distant from the application. The tumor types and frequency of incidence were not significantly different from those noticed in un treated and tricaprylin control groups (table 4). The occurrence of spontaneous mammary tumors in Sprague-Dawley female rats has been reported earlier (3). * Breast tumors on the side of Injection were classified as local tumors. Breast tumors on the opposite side were included under distant tumors. DISCUSSION The malignant tumor incidence in rats and mice treated with a series of epoxides, lactones, hydro peroxides, and peroxides by various routes of ad ministration is compared in table 5 and discussed below. Comparison of Various Routes of Administration in Mice and Rats The reliability of testing of potential carcinogens by subcutaneous injection in mice and particularly in rats has been discussed considerably in recent years. Clayson (5), for example, concluded that sarcomas in the rat by themselves cannot be regarded as evidence for the chemical induction of cancer, since substances normally occurring in the body, such a9 sodium chloride and glucose have been associated with sarcomas on sul> cutaneous injection in the rat. The Food Pro tection Committee of the Food and Nutrition Board, National Academy of Sciences (U.S.), raises the same issue and points out that trauma and regeneration are produced by each injection and may be significant if repeated injections a** necessary to produce the sarcomas (S). The same group concluded that this test is of limited JOURNAL OF TRE NATIONAL CANCER INSTITUT* URL 04099 CARCINOGENIC EPOXIDES, LACTONES, AND PEROXIDES v 831 Table 5.---Comparison of number of animals with malignant tumors with the use of different test systems Compound Skin application in mice* Subcutaneous injection in znicef Low dose High dose Subcutaneous injection In ratsf Intragastrie feeding in ratsf -Propiolactone..................................... djf-Diepoxybutane............................... Glvcidaldebyde............................... .. 0- 6utyrolactone................................. 11/30$ 6/30 8/30 21/30 7/50 3/50 4/50 18/30 5/30 7/30 18/30 1,2,5,6-Diepoxyhexane......................... 1- Ethyleneoxy-3,4-epoxycycIohexane.... Cyclohexenehvdroperoxide.. -.................... Lauroyi peroxide......................................... Cumene hydroperoxide............................. 9,10-Epoxystearic acid............................. . 2,2,4-Trimethylhydroxypeatenoic acid0-lactoiie................................................... Diketene.......................... ............................. 0/30 (4 Papillomas) 9/30 1/30 0/30 (1 Papilloma) 0/30 0/30 (3 Papillomas) 0/30 0/50 2/50 ---- 1/50 3/50 -- 0/50 __ . 3/30 -- 0/30 0/30 1/30 0/30 0/30 0/30 13/205 10/50 7/205 1/50 0/50 (Low dose) 7/20 (High dose) $ 1/50 -- 0/50 -- -- 0/50 1/50 3/5 0/5 0/5 3/5 "References (/) end (f> eicepi where noted. tThUitady. IReference (4). 5Tests currently under way. value and dubious Interpretation. The opposite view is held by Hueper and Conway, who regard such subcutaneous tests valid and significant (7). It is therefore important to examine the present findings against this background of conflicting viewpoints. In the tests carried out in this lab oratory, a series of compounds of known and constant purity were examined by the various routes under discussion, and adequate solvent and no-treatment controls were included. In these experiments no local benign or malignant tumors were observed in any of the 110 mice of the control groups treated once weekly with tricaprylin. In the group of 100 rats, treated similarly, 1 animal bore an adenocarcinoma of the breast which may be ascribed to the spon taneous background incidence of tumors in rats. These findings suggest that trauma and regenera tion associated with the injections were not effective in inducing tumors under the conditions used in our laboratory in either mice or rats, contrary to some of the viewpoints cited above. A similar conclusion was drawn by Dickens and Jones (8) who also used both mice and rats, although in small groups. In their work the medium was arachis oil. Walpole et at. (9) earlier had observed tumor induction with arachis oil (peanut oil) alone, but the possibility of con tamination of arachis oil by the recently discovered aflatoxin cannot be excluded {10). Hieger (11) found that, in control tests with 1,122 C57 mice given subcutaneous injections of a variety of vegetable oils, only 5 animals bore sarcomas. Sodium chloride, glucose, and related sugars, when given repeatedly at high concentration by subcutaneous injections in mice or rats, induce sarcomas at the injection site {12-14). These experiments are frequently quoted to indicate that the subcutaneous tissue of the rat is too sensitive for carcinogenicity assay. An exam ination of the results indicates that daily injections with 25% glucose for 250 days resulted in 5 of 18 mice with sarcomas (13). In rats given daily injections of 25% sodium chloride for 150-250 days, 2 of 39 animals bore sarcomas (14). In another experiment using the same regimen with 25% glucose, 11 of 46 rats bore sarcomas (12). These results do not support the contention that rat connective tissue is more sensitive than that of the mouse (6). Further, these data indicate VOL. 37, NO. 6, DECEMBER 1966 / 832 VAN DUUREN ET AL. clearly that repeated injections of hypertonic only and even in several groups treated with chem solutions in mice and rats are sarcomagenic and icals in tricaprylin. Thus the single fibrosarcomas that this phenomenon may well be one of physical obtained with 1,2,5,6-diepoxyhexane and gly- rather than chemical carcinogenesis. This dis cidaldehyde by subcutaneous injection in 50 rats tinction should be made clear when conclusions may be significant, since these materials also were are drawn about the findings of these workers. tumorigenic in mice by skin application and The induction of local sarcomas by hypertonic subcutaneous injection. On the basis of these solutions is of interest, and, surprisingly, has not results, the carcinogenic activity of the 12 com been reexamined since the findings of the Japanese pounds tested by various routes can be ordered workers almost 30 years ago {12-14). as follows: moderately active--0-propiolactone, Of the 12 compounds tested in mice and rats d,f-diepoxybutane, 0-butyTolactone, and glycid- by various routes, 6 induced malignant tumors aldehyde; weakly active--I-ethylenoxy-3,4-epoxy- on mouse skin (/, 2), 8 caused malignant tumors cyclohexane, 1,2,5,6-diepoxyhexane, cyclohexene- in mice by subcutaneous injection, and 6 induced hydroperoxide, lauroyl peroxide, 2,2,4-trimethyl- one or more malignant tumors in rats by subcu hydroxypentenoic acid-0-lactone, and cumene taneous injection. The connective tissue of rats, hydroperoxide; and inactive--9,10-epoxystearic at least the strain used in this work, does not acid and diketene. appear to be remarkably sensitive to repeated injections. These animals, however, gave the expected high tumor response in the positive Sites of Tumor Induction control group treated with DMBA. A significant feature of tumor induction with When the results of the various test methods the compounds under discussion is that tumors are compared for each compound, clearly the are usually induced by these compounds at the various test methods correlate well. Compounds site of application, whether this be skin, connective inactive on mouse skin, oiz.t cumene hydro tissue, or stomach tissue. Malignant tumors at peroxide, 9,1 Q-epoxystearic acid, diketene, and distant sites were induced in only a few instances. 2,2,4-trimethyI-3-hydroxypentenoic aeid-0-lactone, The incidence of local malignant tumors is partic induced 0-1 tumor by subcutaneous injection into ularly noticeable with 0-propiolactone and 0- mice and rats. The first 5 compounds listed in butyrolactone, both of which are highly reactive table 5 caused tumors on mouse skin (i, 2) and chemically; nevertheless, both gave a 60% yield subcutaneous tumors in mice and rats. Of these, of malignant stomach tumors when fed to rats. By 1,2,5,6-diepoxyhexane induced only benign tumors contrast, aromatic hydrocarbons such as 3-methyl- on mouse skin (2). All other tumors obtained with cholanthrene, when fed repeatedly to rats, did not 1,2,5,6-diepoxyhexane and with 4 other compounds induce gastric tumors (IS); in our experiment (0-propiolactone, d,/-diepoxybutane, glycidalde- a single feeding with DMBA caused 3 mammary hyde, and 0-butyrolactone) were malignant. That tumors in a group of 6 rats, but no stomach 9rt malignant skin tumors were not induced by tumors. The induction of mammary tumors in 1,2,5j6-dLepoxyhcxane on mouse skin may be rats fed DMBA was described recently by Huggins related to the severe corrosive effects of this com and Yang (IS). Hueper and Conway (7) indicate pound on skin and to its toxicity. that carcinogenesis in the stomach is related to In the discussion of these results, there is some the degree of absorption of the chemical by the question as to the validity of a test in which only gastric tissues; metabolic patterns are also sig 1 or 2 of 50 animal* bore malignant tumors at nificant. the site of application. These tumor responses The induction of stomach cancers by the < are below the accepted percentages if one uses 0- lactones is of interest, since d,l~diepoxybutane a procedure outlined earlier (5). On the other 1- ethyleneoxy-3,4-cpoxycyclohexar.e, and glycidal hand, these results should probably not be dis dehyde are carcinogenic by other routes and regarded in view of the absence of local sarcomas like the 0-lactones, are alkylating agents which ar in most control groups treated with tricaprylin expected to act locally, but they gave no gastri' JOURNAL OP THE NATIONAL CANCER. TNSTTTUT C 3D 2 URL 04101 CARCINOGENIC EPOXIDES, LACTONES, AND PEROXIDES 833 cancers. Tliis Is probably due to the difference in the acid-catalyzed hydrolysis of lactones and epoxides. Ross (IT) has shown that the hydrolysis of epoxides is catalyzed by acid or base. The 3 epoxides, d,/-diepoxybutane, l-ethyIeneoxy-3,4epoxycyclohexane, and glycidaldehyde, are there fore expected to be hydrolyzed rapidly in the acid pH environment of the rat stomach (18, 19). On the other hand, the hydrolysis of 0-propiolactone and /S-butyrolactone, although relatively fast, is independent ofpH. in die range ofpH 1-7 (20, 21) so that these 2 materials may remain in the stom ach long enough to exert a carcinogenic effect. Other alkylating agents such as ^V-nitroso^Valkylurethans (22) also Induce malignant stomach tumors when fed to rata. The afiatoxins of moldy peanuts are also implicated in the induction of ad enocarcinomas of the glandular stomach when fed to rats (23). Dickens and Jones (8) and Walpole et d. (9) reported earlier on sarcoma induction in mice and rats by subcutaneous injection of 0-propiolactone; Hendry et al. described the carcino genicity of diepoxybutane (stereochemistry not specified) when injected intraperitoneally in rats m In our study, subcutaneous injections into mice of d,/-diepoxybutane, 1,2,5,6-diepoxyhexane, &propiolactone, and /3-butyrolactone all induced squamous cell carcinomas in one or more animals, and fibrosarcomas in larger numbers of animals. This probably results from spreading of material injected subcutaneously into the skin. With propiolactone, the ratio of squamous carcinomas to fibrosarcomas was particularly high. In a recent report from this laboratory, the chemical reactivities of a series of these compounds were compared with their biological activities (25). Further studies are under way on the mode of action, chemical reactivity, and carcinogenic ictivity of the compound types used in this work. iEFERENCES (/) Van Duumn, B. L., Nelson, N., Orris, L., Palmes, E. D., and Schmitt, F. L.: Carcinogenicity of epoxides, lactones, and peroxy compounds. J Nat Cancer Inst 31: 41-55, 1963. (2) Van Duvmh, B. L., Orris, L., and Nelson, N.: Carcinogenicity of epoxides, lactones, and peroxy compounds. Part II. J Nat Cancer Inst 35: 707- 717, 1965. (3) Noble, R. L., and Cirrrs, J. H.: Mammary tumors of the rat: A review. Cancer Res 19: 1125-1139, 1959. (#) Palmes, E. D., Orris, L., and Nelson, N.: Skin irrita tion and tumor production by beta propiolac tone (BPL). Amer Industr Hyg Assoc J 23: 257- 264, 1962. (J) Clayson, D. B.: Chemical Carcinogenesis. Boston, Little, Brown & Co., 1962, p 57. (6) Food Protection Committee: Problems in the evalu ation of carcinogenic hazard from use of food addi tives. Publication No. 749. Washington, D.C., Nat Acad Sci, 1960. (7) Hueper, W. CL, and Conway, W. D.: Chemical Car cinogenesis and Cancers. Springfield, 111., CL C Tbcgnas, 1964, pp 517-533. (3) Dickens, F., and Jones, H. E. H.: Further studies on the carcinogenic action of certain lactones and re lated substances in the rat and mouse. BritJ Cancer 19: 392-403, 1965. (9) Walpole, A. L., Roberts, D. C., Ross, F. L., Hendry, J. A., and Homer, R. F.: Cytotoxic agents. IV, The carcinogenic actions of some monofunctional ethyleneimine derivatives. Brit J Pharmacol 9: 306-323, 1954. (10) Wooan, G. N.: Experimental toxicity and carcino genicity of aflatoxins. In Mycotoxins in Foodstuffs (Wogaa, G. N., ed.). Cambridge, Mass., M. I. T. Press, 1965, pp 163-173. (//) Hisger, I.: Carcinogenesis. London and New York, Academic Press Inc., 1961, pp 67-68. (12) Nohiyama, Y.: Experimental production of sarcomas in rats by repeated injections of glucose solutions. Gann 32: 85-98, 1938. (13) Taxizawa, N.: Experimental induction of sarcomas in the mouse by injection of glucose, fructose and galactose. A contribution on the question of histo genesis of fibroplastic sarcomas. Gann 34: 1-5, 194a (14) Toxoro, Y.: On the artificial production of sarcomas in white rats with concentrated salt solutions. Gann 34: 149-155, 1940. (15) Shay, H., Aecerter, E. A., Gruenstein, M., and Komarov, S. A.: Development of adenocarcinoma of the breast in the Wistar rat following the gastric instillation of methylcholanthreae. J Nat Cancer Inst 10: 255-266, 1949. (16) Huggins, C., and Yang, N. C.: Induction and extinc tion of mammary cancer. A striking effect of hydro carbon permits analysis of mechanisms of causes and cure of breast cancer. Science 137: 257-262, 1962. (17) Ross, W. C. J.: The reactions of certain epoxides in aqueous solutions. J Chem Soc: 2257-2272, 195a OL. 37, NO. 6, DECEMBER 1966 / / 834 VAN DUUREN ET At, (/5) Roe, J. H., and Dyer, H. M.: Relation of nutrition to gastric function. I. An experimental method. Proc Soc Exp Biol Med 41: 603-606, 1939. {19) Friedman, M. H. F.: Histamine ineffective in the rat as a gastric secretory stimulant. Proc Soc Exp Biol Med 54: 42-44, 1943. {20) Lono, F. A., and Purchase, M-: The kinetics of hy drolysis of 0-propiolactone in acid, neutral and basic solutions. J Amer Chem Soc 72: 3267-3273, 1950. {21) Olson, A. R-, and Miller, R. J.: The mechanism of the aqueous hydrolysis of 0-butyrolactone. J Amer Chem Soc 60: 2687-2692, 1938. (22) Schosntal, R.: Induction of tumours of the stomach in rats and mice by N-nitroso-N-aikylurethanes. Nature (London) 199: 190, 1963. {23) Butler, W. H., and Barnes, J. M.: Carcinoma of the glandular stomach in rats given diets containing aflatoxia. Nature (London) 209: 90, 1966. (2<) Hendry, J. A,, Homer, R. F-, Rose, F. L., and Wal pole, A. L.: Cytotoxic agents. II. Bis-epoxides and related compounds. Brit J Pharmacol 6: 235-255, 1951. (25) Van Duuren, B. and Goldschmidt, B. M.: Car cinogenicity of epoxides, (actones and peroxy coea- pounds. III. Biological activity and chemical re activity. J Med Chem 9: 77-79, 1966. URL 04103 & Ticuse 1.--Squamous cell carcinoma of mouse skin at site of subcutaneous injection of glycidaklehyde given once weekly. X 400 VAN DUUREN, LANGSETH, ORRIS, TEEBOR, NELSON, AND KUSCHNER / 835 125 JOURNAL OF THE NATIONAL CANCER INSTITUTE, VOL. 37 Figure 2.-- Fibrosarcoma in a rat at the site of subcutaneous injection of 0-butyxolactone given once weekly. X 300 836 VAN DUUREN, LANGSETH, ORRIS, TEEBOR, NELSON, AND KUSCHR s o iM n a n Ficure 3a and b.--Squamous cell carcinoma of rat stomach after gastric intubation of /i-butyrolactonc given once weekly. Tumor infiltrating entire thickness of forestocnaeh wall; lumen at lop. (a) X 30, (b) X 400 Van* duuren, lancseth, orris, teebor, nelson, and kcschner / JOURNAL OF THE NATIONAL C.ANCER INSTITUTE, VOL. 37 Date of reoafet 10/24/83 ^ needed after CALL NO. UNIROYAL CHEMICAL LIBRARY ELM STREET NAUGATUCK, CT 06770 A REQUEST Requester's order no. OCT 2 7 983 For use of p. Dovell Status Book author OR periodical title, vol. and date Radiation Research, Supplement Vol. 3, Dept' 1963 Book Title, edition, place, year senes: OR periodical article author, title, pages. O This edition only Kotin^_P_, H.L. "Organic Peroxides, Hydrogen Pe^cCcTde, Epoxrd&s, & Neoplasia." ^p. 193 Verified m: ORTltem-cited in--------- " ISBN, OR ISSN, or LC card, or OCLC. or other number if known .SerialS List. If non-circulating, & cost does not exceed $ ______ please supply Microfilm Hard copy Health Center Library University of Connecticut Farmington, CT 06032 Request complies with 108(g)(2) Guidelines (CCG) GJ other provisions of copyright law (CCL) AUTHORIZED BY: (full name) TITLE__________________ Ann Harmon Librarian Request for LOAN or B PHOTOCOPY According to the A L.A. Interlibrary Loan Code REPORTS: Checked by SENT BY: Library rate ____________ Charges SInsured for? Date sent___________ ____________________ DUE________________________________ RESTRICTIONS: 0 For use >n library only Copying not permitted 0_____________ NOT SENT BECAUSE: Non Circulating In use Not Owned C 33 r~ Estimated Cost of: Microfilm Hard copy - o --j BORROWING LIBRARY RECORD:* *t ' Date received_________________________ -- Date returned_________________________ By Libraryrate D____ i--- Postage enclosed?_______ Insured for?--------------------------- RENEWALS: No renewals Requested on------------------------------------------------- - Renewed to____________________________ (or period of renewal) Vaf Ine ,fce'v">| tibrtty iss^mts itSBOMO'iiiy 'o' neWiCifw 0/ nor itc B REPORT | | CALL*N0. i. | 10/24/83 Not needed after r _' -. / _ 1 .. . s fxc' * ' Requester's order no. . :............... - ^ I - for use of p . Dovell Status : Book author; OR periodical title, vol. and date i ' Radiation Research, Supplement Vol. * 2, Oept. 1961 1 Book Title, edition, place, year series: OR periodical article author, title, pages. This edition only ,j Kotin, P & ialk, H.L, "Organic Peroxides, Hydrogen Peroxide, Epoxides, & Neoplasia." pp. 193 ff. : Verified in: OR: item cited in ISBN, OR ISSN, or 1C card, or OCLC, or other number if known Vc*4.ry Seri al s Hist, i If non-circulating. J> cost does not exceed $_____________ please supply Microfilm Hard copy Health Center Library 1 . University of Connecticut ^ Farmington, CT 06032 / Request complies with , 108(g)(2) Guidelines (COG) u=other provisions ol copyright law (CCL) AUTHORISED BY: (full name)____ -'am'c.i TlTI P Librarian IMLC-------------------------------------------------------------- Request lor D LOAN or PHOTOCOPY According to the A L A. Interlibrary Loan Code REPORTS: Checked by SENT BY: Library rate ____________ Charges $Insured for$ Date sent_______________________________ DUE_________________________________ RESTRICTIONS: for use in library only Copying not permitted _____________ NOT SENT BECAUSE: 'n u" Not O'wned Non Circulating Requestot__________________ C Estimated Cost ol: Q Microfilm___ Hard copy __ BORROWING LIBRARY RECORD Date received_____________ I-- Date returned_______________ >-- By Library rate _____ Postage enclosed $,Insured for S - ' o CO RENEWALS: No renewals Requested on! Renewed to--------------------------------------------------------- (or period of renewal) Alois ine lece .'iy t'Ori'r iSSL'i'fs 'tsconuMty tor nohhcjuor' ol non e pl Management & Technical Information Services/Library / --'J'^! / f' ` v- - -< UNIROYAL CHEMICAL Oivision of UNIROYal, Inc Eim Street Naugatuck. Connecticut 06770 Tel 203-723-3252 UNIROYAL MEMO FROM % Patricia Ann Harmon URL 04109' \J> ./Yvo ov^S;- Lfcc r radiation research supplement 3, 193-211 (1963) Orgonic Peroxides, Hydrogen Peroxide, Epoxides, and Neoplasia1 k PAUL KOTIN and HANS L. FALK2 University oj Southern California School oj Medicine, Los Angeles, California S * | INTRODUCTION In the more than the half-century since the recognition of the carcinogenic properties of ionizing radiation (I-R), there have been numerous suggestions as to the mechanisms concerned with the induction of neoplasia. Currently there is no universally accepted concept; however, the physical properties of I-R, par ticularly in relation to the reactions occurring in both water and tissues, suggest a possible avenue by which the carcinogenic effect may be mediated. The production of ions (H*, OH-) and inorganic radicals (OH, H02) with hydrogen peroxide formation in aqueous systems in response to I-R is well known. A similar reaction sequence within the aqueous milieu of the cell apparently occurs. Controversy exists, however, as to which of the reaction products, if indeed any, is responsible for the radiation effects. Organic peroxides have been considered by some to be responsible for the cellular effect seen after radiation, whereas others, of whom perhaps Warburg is the most resolute, maintain that hydrogen peroxide is the active intracellular product. Warburg (1) has recently stated that the direct action of X-rays on cell metabolism, in this instance cancer cells, was due to the hydrogen peroxide produced rather than to organic peroxides, intermediate hydrogen atoms, or OH radicals. j On the assumption that organic peroxides are produced after exposure to I-R, a preliminary study was undertaken to determine the carcinogenicity of organic peroxides. This report presents data obtained at the unsophisticated level of routine bioassay of a broad spectrum of these compounds. We wished to determine whether information and research leads could be obtained in a manner analogous to that ff' This work was supported by a grant from the National Cancer Institute, National Institutes of Health. From the Department of Pathology' of the University of Southern California School of Medicine and the Pathology Department, Los Angeles County Hospital-, Los Angeles, Cali fornia. ,. 'Current address: Carcinogenesis Studies Branch, National Cancer Institute, Bethesda 14, Man-land. 193 ,, ourrroCOP'f COWWES V*W <-* , '*&* >.. <-*' - 194 KOTIN AND FALK which has served so brilliantly the studies of mechanisms of action of carcinogenic hydrocarbons, azo dyes, and aromatic amines. The utilization of bioassav data served as a basis for the studies of Pullman and Pullman (2), Daudel and Daudel (5), and Daudel et al. (4) aimed at relating chemical structure to carcinogenic properties and potency. Of additional concern to us is the growing general environ mental occurrence of the class of compounds of primary interest to this symposium. Urban air pollution, food preparation, processing, and storage, and the increasing production and use of synthetic polymers are all potential sources of human ex posure to peroxides. The established carcinogenic properties of I-R are of particular interest in that peroxide formation suggests an in vivo mode of action. In a discussion of carcino genesis by I-R, Furth and Lorenz (5) concluded: "The induction of neoplasms by ionizing irradiation is a fascinating chapter in the history of science. It is full of problems of increasing practical importance awaiting solution. As a tool in cancer research, it is unequaled." Pertinent to this are the observations that similarities exist between the biochemistry of the response to I-R and the bio chemistry of aging {6-8). Both biological situations have additionally in common an increased incidence of neoplastic disease. The studies of Harman (5) have been particularly pertinent, and his experimental data describe a prolongation of the normal life span and inhibition of spontaneous cancer in mice by antioxidants. Finally, environmental carcinogenic stimuli have been shown to be capable of acting in a combined manner. Experimentally, it has been demonstrated that I-R has an enhancing effect on the leukemogenic action of methylcholanthrene \W), a cocarcinogenic effect on the production of skin cancer by benzo(a)pyrene (III and benzo (a) pyrene and dibenzi a, A) anthracene {12), a synergistic action in association with estrogenic hormones in the induction of lymphosarcomas in mice (13), and an additive effect in the carcinogenic response after the application of cigarette tar to the skin of mice (14). Prior to discussion of certain factors con cerned with the in vivo formation and action of peroxides, a general assessment of their enviromental status and role in carcinogenesis will be undertaken. ENVIRONMENTAL SOURCES Gastrointestinal Environment Among the earliest studies on the causative factors in human cancer were those on heated fats in an effort to explain the pathogenesis and high incidence of cancer of the gastrointestinal tract. The chemical relationship between sterols and bile acids on one hand and 3-methylcholanthrene on the other was recognized at the outset of these studies, and a diligent search was instituted for 3-methvlcholanthrenc and associated compounds, possibly resulting from the application of heat to the sterols present in fatty materials. Although many investigators produced sarcomas after the injection of subfractions of different heated fats, none succeeded dciuom PEROXIDES, EPOXIDES AND NEOPLASIA 195 in isolating the proximate carcinogenic agent. Hieger (15), who has performed perhaps the most detailed of these studies particularly utilizing cholesterol and its derivatives, ultimately had to concede that the proximate carcinogenic compound or compounds were most elusive. During the heating of lipids, dehydrogenation occurs with the production of polycyclic aromatic hydrocarbons (PAH) pro ceeding at high temperatures. However, the yield of carcinogenic hydrocarbons under usual circumstances is at best minimal, and those theoretically capable of formation are primarily of weak potency, e.g., chrysene {16). Davies and Wilmshurst (17) have reported the formation of benzo(a)pyrene from starch at 370 to 390C. The low temperature of formation is of pertinence here. Fieser (18) and BischofF (75) investigated the production of peroxides from cholesterol and more unsaturated sterols with these serving as the sources of the proximate carcinogenic compounds. Fieser (18) advanced the concept that with the in vivo peroxidation of cholesterol an endogenously initiated carcinogenic mechanism of considerable significance existed. It is wholly conceivable that the elusive compound or com pounds of Hieger which disappear on successive fractionations are peroxides. Data from numerous studies are either inconclusive or negative. Studies are now in progress on epoxy derivatives of corticosteroids, and evidence of weak carcino genicity is being obtained. The role of heated fats and allied compounds as car cinogens has been reviewed by Arffmann (20). A ubiquitous source of entry of peroxidic compounds into the intact host is through dietary fats. Further, there is ample evidence that polyunsaturated fatty acids in the depot fats of animals can peroxidize in vivo. Peroxidation can be demonstrated by analytical studies on the fat and by the typical yellow-brown discoloration which has led to the term "yellow fat disease" (21). Although per haps superficially unrelated to carcinogenesis, it has been suggested that in vivo peroxidation of lipids, particularly in nonadipose tissue, may result in symptoms of vitamin E deficiency. The antioxidant effect of vitamin E will be discussed. Anti oxidants have the capability of preventing in vivo peroxidation, and tocopherols serve as an excellent example of these compounds. Of major concern to us is the observation that this effect is not limited exclusively to fatty tissue but also to the lipid-containing units within the cells, as, for example, mitochondria. It is reason able to assume that other lipid structures, including those occurring in the mem branes of the cell walls, the microsomes, and the lvsosomes, are susceptible to dietary influence, and these structures also, as a result, may undergo peroxidation (22). In addition to in vivo peroxidation, the lipids in dietary substances may also undergo peroxidation. This exogenous process may be by far the most signifi cant source of lipoperoxides. Heating in the presence of air. as well as the radiation process, is capable of forming peroxides in fats. On a practical level has been the report that highly peroxidized oils and fats present in fish meal induce hepa tomas when fed to trout (23). The toxicity and carcinogenic effect of lipoperoxides 196 KOTIN AND FALK TABLE I Tumobigenic Response after Exposure to Lipoperoxides Compound dose (mM) No. of mice and Strain Appearance of first tumor Tumor-bearing animals S 2 S-*. fe S6cuan.eo!u Malignant Pulmonary sarcomas lymphomas adenomas No. % No. % No. % Epoxydated soya oil Butyl 9,10-epoxy stearate Allyl 9,10-epoxy stearate 0.5 1.7 50 C57B1 50 C57B1 1.0 30 C57B1 24 17 20 1 5 2 10 2 10 31 3 10 2 6 3 are in need of intensive study, particularly as the data obtained may well also apply to peroxides formed after exposure to I-R. The studies of Glavind and Tryding (4) are pertinent to this. Their findings indicate that peroxidized lipids when ingested are hydrolyzed and absorbed through the intestine and that after absorption the peroxidic nature of the compound is lost, indicating perhaps that the peroxide is destroyed in the intestinal mucosa. The relation of the destruction of the lipoxy material to development of neoplasms of the gastrointestinal tract is unknown. Lipoperoxides may undergo further reaction leading to epoxidated fats with potential carcinogenic properties (Table I). Bioassav data in this and subsequent tables frequently report findings from experiments still in progress, and quantitative interpretation of all the data is impossible at the time of this symposium. The method of application is given on the table for each experi ment, and except where noted C57 Black mice were used in all studies. A dose of 0.5 raM was used for the testing of pure compounds except when otherwise designated. Epidemiologic studies on groups with both high and low gastrointestinal cancer rates, particularly in relation to dietary lipids, have been unrewarding. The real and progressive decrease in gastric cancer being reported from other than high endemic areas is resulting in an increased effort to identify variations in dietary characteristics. Although this phase of the symposium is concerned primarily with the carcino genic properties of peroxides, data from bioassav experiments with epoxides arc also being reported. The theoretical possibility that the carcinogenic properties of peroxides may be mediated through the formation of epoxides must be con sidered. since epoxides arc known to be readily formed from peroxides, hydrogen peroxide, and peracids. Inasmuch as considerably more information is available relative to the mode of action and carcinogenic properties of epoxides, the findr would appear to be pertinent. URL 0411 oa Pprimmy part in t traps h injected line svn Pro Aerosol Angeli PEROXIDES, EPOXIDES AND XEOPLASIA 197 Fir.. 1. Postulated photochemical reactions of carcinogenic significance occurring in polluted air Respiratory Environment Peroxides have been identified in the polluted atmosphere of several large Ameri can cities, most particularly, however, in Los Angeles. The photochemical reac tions between the oxides of nitrogen and the unsaturated hydrocarbons resulting in the formation of peroxides have been well established (Fig. li. The exhaust products of gasoline engines and industrial effluents have been shown to be the primary source of the oxides of nitrogen and unsaturated hydrocarbons taking part in this reaction. Samples of Los Angeles atmosphere collected in liquid nitrogen traps have been shown to be carcinogenic when either painted on the skin or injected subcutaneously in mice (Table II and Fig. 2). Aerosols of ozonized gaso line synthetically reproduced in inhalation chambers have resulted in the produc- TABLE II Production of Skin Tumors in Mice with Oxidation Products of Aliphatic Hydrocarbons after Painting Three Times a Week Coin pound So. of mice and strain Aerosol condensate, Los Angeles atmosphere 50 C57B1 A ppearance of first tumor vlft *s a % 14 35 Tumor-bearing animats Skin Malignant lymphomas Pulmonary adenomas So. % So. % No. % 20 j -- -- ! -- 7- *#* .M 198 KOTIN AND FALK Fig. 2. Photomicrograph of carcinoma in mouse induced by skin painting with polycyclic aromatic hydrocarbon-free condensate of smog. tion of pulmonary neoplasms in strain A and Co7 Black mice. Strain A mire arc characterized by a high spontaneous incidence of pulmonary neoplasms and a marked susceptibility to tumor induction bv carcinogenic agents. This property permits the implementation of a modified time-tumor-dosc study as shown in Figs. 3 and 4. Co7 Black mice, in contrast, have a less than l<~c spontaneous in cidence rate of pulmonary neoplasms and extreme resistance to pulmonarv tumor induction. The results of inhalation experiments with this strain are shown in Table III. The end state of the complex reaction in the atmosphere is the formation of aldehydes and organic acids. The aldehydes in particular resulting from partial combustion of gasoline have been found to be responsible for the major proportion of the irritant effect of Los Angeles smog. In the presence of atomic oxygen these aldehydes can give rise to hydroperoxides when moisture is available either exogenously or endogenously. Another group of compound* formed fvum the reac tion products of vehicular exhausts are epoxides which, though far less irritating than aldehydes, have been shown to be carcinogenic (25, 201. The formation of epoxides in the atmosphere may or may not require the intermediate formation of PEROXIDES, EPOXIDES AND NEOPLASIA pollam P/P WASHCO P/P CCAfTPOL 199 WKS Fig. 3. Percentage of pulmonary tumor-bearing mice (strain A) removed from synthetic smog chambers at planned intervals. URL 04116 Wertcs Fig. 4. Percentage of multiple pulmonary tumor-bearing mice (strain A) removed from syn thetic smog chambers at planned intervals. W 200 KOTIN AND FALK TABLE III Rate or Induction or Pulmonary Tumors in C57 Black Mice after Exposure to an Atmosphere or Ozonized Gasoline Washed air chamber Smog chamber Total Dumber of tumor-bearing animals Number of survivors at time of first tumor Per cent tumor-bearing animals 6 367 1.6 15 155 9.6 TABLE IV Expoxides Currently Being Bioassayed for Carcinogexicitv : Significance of Compounds Relates to Theoretical Formation in Urban Polluted Air 1.2-Epoxv-3-methoxypropane 1.2-Epoxy-3-ethoxypropane 1.2-Epoxy-3-propoxypropane 1.2-Epoxy -3 -(1-methy let hoxy)propane 1.2-Epoxy-3-butoxypropane 1.2-Epoxy-3-(2-methyl propoxy) propane 1.2-Diisobutylene oxide Triisobutylene oxide Vinylcyclohexene monepoxide peracids; however, chromatographic analysis of smog has indicated that peracids are indeed present in addition to a large variety of other peroxides (P. Kocin, H. L. Falk, and N. Hellinger, unpublished data). A listing of epoxides theoretically present in polluted air and currently under bioassay is given in Table IV. Subse quent reaction between aldehydes and peroxides may result in the formation of the potent chemical agent bishydroxymethyl peroxide (Fig. 5). The formation of re lated compounds may begin with the lower aldehydes and ozone or other oxidants with the bishydroxymethyl peroxide having been shown to possess carcinolytic properties. Although likely, it is yet to be determined whether this carcinolytic agent in common with numerous others possesses carcinogenic potency as well. Ozone, a major constituent (35 to 40%) of the oxidant material in the urban atmosphere in the Los Angeles area {27), has now been classified as a radiomimetic gas {28, 29). It would therefore appear to be of double significance, first as an integral reagent in the photochemical reactivity in the Los Angeles atmos phere, and second as a mutagenic, radiomimetic agent on its own. The toxicity of this agent has heretofore precluded any detailed carcinogenic studies. The expanding synthesis and commercial use of polymers have created in a small group of workers a theoretical occupational hazard incidental to the use of benzoyl peroxide, for example, as a major compound in the polymerization of monomers. R- MORI C< C'yclohe.v Styrene e Octylene UodeceiuHexadecu Epichlori C'oudens: of ey with 1 idene d epoxy i Epoxide bioassax As St; consider PEROXIDES, EPOXIDES AND NEOPLASIA NO? hv --- NO 0 0 + 02 201 /0-0 ch2-ch2 03 -- CH2 HgyO 1 -HO-CH2-0-0-CH2-OH **/* lJ' H-,0 CKj R - C- CH2 + 0- q3 + R * C - CH2 -ft- - V0 R-C-CH, '0' 2 Fig. 5. Anticipated reactions leading to the formation of bishydroxymethyl peroxide in pol luted air. TABLE V Tumorigexic Response after Exposure to Epoxides of Industrial Significance Compound t Total dose (nil/) A pptarancc oj first tumor So. of mice strain c 5 -E V) vi = Tumor-bearing animals Skin Malignant Pulmonary lymphomas adenomas No. % No. % No. % Cyclohexene epoxide Styrene epoxide Octylene epoxide Dodecene epoxide Hexadecene epoxide Epichlorohvdrin Condensation product of epichlorohydrin with 4,4'-isopropyl* , idene diphenol (8.1% ; epoxv 0) * 20 20 20 20 25 5 900 30 C,H 30 CjH 30 CjH 30 CjH 30 CjH 30 CjH 50 C57B1 7 11 7 10 9 3 8 22 -- -- --1 5 19 -- -- 3 16 -- -- 26 -- -- 6 23 -- -- 21 -- -- 4 19 -- -- 30 -- -- 7 23 3 30 3 3 10 3 46 1 2 3 6 Epoxides similarly have broad industrial use, and certain ones of these have been bioassaved for carcinogenicity (Table V). 1 As stated previously, our interest in the carcinogenic response has caused us to consider the possibility that the carcinogenic effect of peroxides might be mediated 202 KOTIX AND FALK ,' TABLE VI Tumorigenic Response after Exposure to Selected Diepoxides A ppearance of first tumor Tumor-bearing animals Time (months) Compound Total dose (mM\ .Vo. oj mice and strain Malig Pulmon r Skin nant ary lymph adeno omas mas 1,2,3, 4-Diepoxybutane 1,2,3,4-Diepoxy butane l,2,3,4-Diepoxy-2-methy!butane 1,2,4,5-Diepoxv pentane 1,2,5,6-Diepoxyhexane Diepoxvlimonene Vinyl cyclohexene diepoxide Di(2,3-epoxy)propvl ether Di(2,3-epoxy)propyl ether Diepoxydihydromvrcene Methyl diepoxydiallyl-acetate 1,2,7,8-Diepoxvoctane 1,2,8.9-Diepoxvnonane 1,2,9,10 - Diepoxy decane 1,3-Bis (2,3-epoxy propoxy)benzene 1,3-Bis i'2,3-epoxy propoxyj benzene 1,4-Bis(2,3-epoxvpropoxy)benzenc 1 1 0.2 0.2 1 1 0.5 0.75 0.25 0.75 0.5 0.75 0.0 0.75 0.75 0.25 1 20 C57B1 20 C57BI 10 C57BI 10 C57BI 20 C57B1 20 C57B1 20 C57B1 20 C57BI 20 C57B1 10 C57B1 10 C57B1 10 C57B1 10 C57B1 10 C57BI 20 C57B1 20 C57BI 20 C571J1 5 5 10 12 18 14 5 6 12 9 18 8 No. * No. % No. % 20 7 35 2 10 __ 20 5 25 1 5 __ 10 -- 1 10 __ __ -- __ __ __ _ 20 -- 2 14 __ -- 17 6 6 35 __ __ 16 6 4 25 __ 10 4 40 __ __ 8-- ---- -- 8 12 1 12 -- -- 5 20 1 20 __ -- -- ---- ---- -- 6 16 1 16 -- -- 7 -- - 1 14 -- 14 7 -- _ -- -- -- ---- ---- -- ---- ---- through the formation of epoxides. Because of the plethora of knowledge relating to the carcinogenic mechanisms of epoxides--e.g., crosslmking. mutagenicity--we are intentionally including data from experiments devoted to bioassay of epoxides. The carcinogenicity of certain selected diepoxides is given in Table VI. FACTORS MODIFYING THE PEROXIDE EFFECT The assessment of host resistance factors is crucial to the study of mechanisms of carcinogenesis. Compounds possessing an active SIT group somewhere in the molecule may act as protective agents against the potential adverse effects of peroxides by sparing the essential SH groups in the cells for physiologic activity, Vitamin E, on the basis of its antioxidant properties, lias also been found to be a significant protective agent. Particularly important is the observation that a protein-bound vitamin E complex, presumably responsible for protection, exists (30). It is also of interest that different tissues require different lengths of time to acquire protection against radiation after vitamin E has been restored to the . .et, .iur In .ay iniiObi .esti ?ro> lint T1 on. dec T1 <--_ eid ^ st CO 1 1 >\\ 10 Ci .lism ` duce .icce; ____ i ____ ' i animals Pulmon ary adeno mas Xo. % relatin ity--w ioxider lanisrr. in th ects c vity. d to b. that exisi of tim 1 to th PEROXIDES, EPOXIDES AND NEOPLASIA 203 diet, and fluctuations in the supply of the vitamin may enhance susceptibility to injury. Iron and many iron-containing compounds are of great importance, since they may exert a protective effect under circumstances when the complexity of an or ganic peroxide precludes the protective action of catalase and peroxidase. Hemo globin and cytochromes are active, although when compared to catalase their destructive action on simple peroxides is insignificant, In the case of complex peroxides, when catalase is inactive, the action of iron-containing compounds as sumes great significance. The action of siderophilin and chelating agents is related to the availability of iron, the local oxygen tension being the controlling factor in the intensity of the effect {31, 32). RAD10MIMETIC ACTION OF PEROXIDES The biological significance of hydrogen peroxide which is presumably formed incidental to the action of I-R on cells is as yet not fully determined. A limited role is strongly suggested when one recognizes that the yield is insignificant when contrasted with the far greater amounts produced incidental to the action of certain riboflavin-containing enzymes {33). Furthermore, of all peroxides, hydro gen peroxide is the most efficiently destroyed, particularly by catalase and formic acid oxidase action. The destructive effect of catalase on various peroxides is shown in Table VII. The relation of stability in the presence of catalase to carcino genesis is unknown; however, we feel it merits further study. Current concepts suggest a possible primary role for organic peroxides in mecha nisms of carcinogenesis by radiation. Peroxides formed have been shown to pro duce cell injury at a level consistent with cellular survival, Further, the growing acceptance that cancerization of the cell requires alteration in the genetic material TABLE VII Susceptibility to Destruction by Catalase* Compound Destruction Hydrogen peroxide Carbamine peroxide Histidine peroxide Methyl hydroperoxide Ethyl hydroperoxide 1-Hydroxvethyl hydroperoxide Bishydroxv methyl peroxide terf-Butyl peroxide Caprylic peracid Yes Yes Yes Yes Yes Y es No . No No From Weitzel el al. {34) and Maehlv (35). / 204 KOTIN' AND FALK is compatible with the mutagenic action described for certain organic peroxides (36). Dickey et al. (37) induced mutation in Neurospora with organic peroxides and suggested that in the mixture of formaldehyde and hydrogen peroxide the active agent might be bishydroxymethyl peroxide. Altenburg (SS) used tert-butyl hydroperoxide for the production of mutations in Drosophila. Chromosome break age was described by Loveless (3P), also using tert-butyl peroxide. The mutagenic effects of organic peroxides are critical to any discussion of the mechanism of the carcinogenic action of I-R. A strict parallelism between carcinogenicity and mutagenicity limited specifi cally to I-R has been emphasized by Latarjet {36), The studies on the mutagenicity of peroxides referred to previously suggest a species-specific effect. The data of Chevallier and Luzatti (40) additionally suggest a site-specific effect within the genetic material. Data obtained from bioassay of certain organic peroxides are shown in Table VIII. We have bioassayed numerous mutagenic agents for carcinogenic potency, and in our hands there is an inconstant relationship between the two properties. Muta genic action of such compounds as formaldehyde, ethylene oxide, and propylene oxide has not been found to be associated with carcinogenicity, but technical limitations incidental to the study of these compounds are obvious. Of interest is the mutagenicity of the organic peroxides listed in Table IX. Data on associated TABLE VIII Tumorigenic Response after Exposure to Selected Organic Peroxides Appearance of first tumor Tumor-bearing animals | Tim e {months) No. of survivors Compound Dose .Vo. of mice UA/J and strain Acetyl peroxide Lauroyl peroxide Sodium peracetate Diisopropyl benzene hydro- peroxide p-Menthane hydroperoxide Cumene hydroperoxide m- and p-/erf-Biitvl isopropyl ben- zene hydroperoxide Methyl elhvl ketone peroxide Di-terl-butyl peroxide ferl-Butvl perbenz.oate 00 40 80 50 90 50 50 40 100 40 50 C57B1 50 C57B1 50 C57BI 50 C57B1 50 C57B1 50 C57BJ 50 C57B1 50 C57B1 50 C57B1 50 C57B1 12 12 14 10 11 14 13 15 19 13 Subcutan eous sarcomas MoliKnant lymph omas Pulmon ary adeno mas No. 9c No. 1 No. % 38 __ __ 4 10 1 3 43 -- -- 2 5---- 37 -- -- 3 4G -- -- 2 8 --- 436 4(i -- _ 0 13 2 4 38 3 11 29 -- -- 44 -- -- 2*2 50 3 7 34 1 3 3 9 3 35 -- -- 7 20 3 47 -- 10 21 -- -- PEROXIDES, EPOXIDES AND NEOPLASIA TABLE IX Orgastc Peroxides with Mutagenic Properties Compound Ref Hydrogen peroxide Bishydroxvmethyl peroxide Hvdroxymethyl er(-butvl peroxide Diisopropyl ether peroxide ferf-Butvt peroxide Succinic peroxide Cumene peroxide Thj mine peroxide Pcroxy fatty acids Ethylene oxide (37, i2) (31, 57) (37) (57) (37-59) (40) (40) (40) (45) (39) 205 mutagenicity and carcinogenicity are least satisfactory' relation to carcinogenic hydrocarbons. It is true that 3-methvIcholanthrene, dibenzfa,Inanthracene, and 7,12-dimethylbenz(a)anthracene were tested by Barratt and Tatum (4D and found to be mutagenic for Neurospora; however, their mutagenic capability was at a significantly lower level than that observed for nitrogen mustard or radiation. Although slightly remote from the immediate interests of this symposium, cer tain parallelisms between carcinogenicity and mutagenicity are worthy of denota tion. Alh'lisothioeyanate has been shown to be mutagenic [44). We are currently reporting two cases of lung cancer in workers exposed to Fiberglas manufacture and synthesis (45). /3-Propiolactone was shown to be carcinogenic by Roe and Glendenning (46) and by us (47). The mutagenicity of this compound for A'eurospora was reported by Smith and Srb (45). Extensive data exist describing the mutagenicity of mustard gas and nitrogen mustards. Alkylating agents, particu larly the mono- and diepoxides, similarly show mutagenicity and carcinogenicity (49). In summary, we do not feel that the correlation as originally claimed for PAH but later questioned by Demerec (50) necessarily precludes a combination of mutagenicity and carcinogenicity for these compounds. Rather our own data sug gest that, in the instance of PAH, factors affecting biologic availability may be critical in elucidating mutagenicity. These factors include the vehicle of the carci nogen, its stability, sites of elective anatomic localization, and local and hepatic metabolic degradation. Similar reasoning may be applied to the lack of data sug gesting mutagenicity for carcinogens belonging to the group of azo dyes and aromatic amine carcinogens. SUMMARY AND CONCLUSIONS The preliminary data being reported support the suspicion that organic peroxides mav possess carcinogenic properties. This suspicion in large measure relates to the acknowledged mutagenicity of certain peroxides. 206 KOTIN AND FALK As a possible mechanism of action it is suggested that peroxides may act by catalyzing the depolymerization of deoxyribonucleic (DNA) and ribonucleic acid (RNA). This radiomimetic property may provide a fundamental bridge between the carcinogenicity of I-R and that of organic peroxides when used in pure form. Emphasis has also been directed to the carcinogenic properties of epoxides. Pertinence to the topic of the symposium exists in two areas. First, organic peroxides may lead to the formation of epoxides. Second, the established carcinogenicity of epoxides has resulted in the accumulation of a large body of information relating to the mechanism of their action. Oi particular significance in this area is their capability of acting as crosslinking agents. The possibility of utilizing the established carcinolytic properties of certain carcinogenic agents must be considered in the study of the carcinogenic properties of organic peroxides. Bishydroxymethyl peroxide is an example of a compound possessing carcinolytic activity (34* in need of bioassay for carcinogenicity. REFERENCES 1. 0. Warburg, .Yen? -Methods oj Ceil Phyriology. Intcrscience Press, Germany, 1962. i. A. Pullman and B. Pullman, Electronic structure and carcinogenic activity of aromatic molecules. Advances in Cancer Research 3, 117-167 (1955). 3. P. Dauull and R. Dacdel, On tlie physico-chemical theory of the mechanism of action of carcinogenic substances. Ada Unto intern, contra Cancrum 6, 20-26 (1948). P. Daudvsl, R. Daudel, and N. P. Buv-Hoi, The problem of predicting the carcinogenic property of chemical substances. Acta Vnio Intern, contra Cancrum 7, 91-115 (1950). 5. J. Furth and E. Lorenz, Carcinogenesis by ionizing radiations. In Radiation Biology (A. Hollaetuier, e<l.), Yol. I. Part II, pp. 1145-1201, McGraw-Hill Book Company, New York, 1954. 3. A. C. Upton. Ionizing radiation and aging process: Rev iew. J. Gerontol. 12, 30G-313 (1957). 7. G. J. Neary, Ageing and radiation. Xoturc 187, 10-17 (19G0), 8. B. L. Stueiiler, Origin and comparison of the effects of time and high energy radiations on living systems. Quart. liev. Biol. 34, 117-142 (1959). 9. D. Harman, Prolongation of the normal lifespan and inhibition of spontaneous caucer by antioxidants. J. Gerontol. lf>, 247-254 (1961). 10. J. Furth and M. D. Boon. Enhancement of leukeinogcnic action of methylcholanthrone by pre-irradiation with X-rays. Science 98, 13S-139 (1943). 11. J, D, Mottram, Production of epithelial tumors by irradiation of a precancerous skin lesion. Am. J. Cancer 30, 746-748 (1937). IS. W. V. Mayneoro and L. D. Parsons, The effect of X-irradiation on tumor pioduction hv a chemical compound in mice, and the associated blood changes. J. Pathol. Bacterid. 45, 35-48 (1937). 13. A, Tyiksch r aum, J. R. Shapiro, anil H. W. Mixer, Svnerstistic action of estrogenic hormone and x-rays m inducing thymic hmpliosarfomn of mice. Pmc. Snc. Eiptl. Biol Med. 72, 632-634 (1949). 1/,. E. V. Cowpry, A. Croninger, S. Solaric, and V. Suntzkff, Combined action of cigarette tar and beta radiation on mice. Ca> r 14. 344-352 (1961). 15. I. Hikgkh, Carcinogenesis by clioi- -toI. Bril. J. Cancer 13, 439-451 (1939). 1C. H. L. Falk. S. Goldfein, and P. E. oteineu. The products of pyrolysis of cholesterol at 360 Cand their relation to carcinogens. Cmircr Rvscn'rh 9, 438-447 (1949). URL 04123 R J. .!. A . J. >.;. F G ;A 1; 17. l h. ! at. . PEROXIDES, EPOXIDES AXD NEOPLASIA 207 17. W. Davies and J. R. Wilms hvbrt. Carcinogens formed in the heating of foodstuffs. Forma tion of 3,4-benzopyrene from starch at 370-390C, Brit. J. Cancer 14, 295-299 (I960). IS. L. F. Fieser, Some aspects of the chemistry and biochemistry of cholesterol. Science 119, 710-716 (1954). 19. F. Bischoff, Carcinogenic activity of cholesterol oxidation products and sesame oil. J. Natl. Cancer Inst. 19, 977-987 (1957). 20. E. Arffmann, Heated fats and allied compounds as carcinogens: A critical review of experi mental results. J. Sail. Cancer lust. 25, 893-926 (I960). 21. L. J. Machli.v, Dietary factors influencing the susceptibility of animal fats to peroxidation. Presented at the Symposium on Foods, Oxidative Deterioriation of Food Lipids. Oregon State University. September 11-13, 1961. 22. G. J. Marco, L. J. Machlin, E. Emory, and R. S. Gordon, Dietary effects of fats upon fatty acid composition of the mitochondria. Arch. Biochem. Biophys. 94, 115-119 (1961). 23. E. M. Wood and C. P. Larson, Hepatic carcinoma in rainbow trout. A.MA. Arch. Pathol. 71,271-479 (1961). 21.. J. Glavind and X. Trydinc, On the digestion and absorption of lipoperoxides. Acta Physiol. Scam!. 49, 97-102 (1960). 25. J. A. Hf.ndry. R. F. Homer, F. L. Rose, and A. L. Waltole, Cytotoxic agents; II. Bis-epoxides and related compounds. Brit. J. Pharmacol. 6, 235-255 (1951). 20. C. J. McCammon, P. Kotin, and H. L. Falk, The cancerogenic potency of certain diepoxides. Proc. Am. -issoc. Cancer Research 2, 229-230 (1957). 27. F. E. Littman and C. W. Marynowski, Identification of ozone in the Los Angeles atmos phere. Anal. Chem. 28, 819-825 (1956). 25. R. Brinkman and H. B. Lamberts, Ozone as a possible radiomimetie gas. Nature 181, 1202- 1203 (1958). 29. R. H. Fetner. Chromosome breakage in Vicia faba by ozone. Nature 181, 504-505 (1958). 30. J. G. Bieri and A. A Anderson, Peroxidation of lipids in tissue homogenates as related to . vitamin E. Arch. Biochem. Biophys. 90, 105-110 (1960). 31. A. A. Barber, Inhibition of lipid peroxide formation by vertebrate blood serum. .4reft. Biochem. Biophys. 92, 38-43 (1961). 32. J. A. Blain and T. Barr, Destruction of linoleate hydroperoxide by sova extracts. Nature 190, 53S-539 (1961). 33. F. Portwich and H. Aebi, Determination of the peroxide formation of animal tissue by means of peroxidative oxidation. Helv. Physiol. Pharmacol. Acta 18, 312-327 (1960). Si. G. Weitzel, E. Bcddecke, and F. Schneider. Cytostatic effect of organic peroxide compounds -j on ascites tumor cells. Z. physiol. Chem. 323, 211-235 (1961). 35. A. C. Maehly, Complexes of bacterial catalase with peroxide, azide, CO. Biochim. el i Biophys. Acta 54, 132-144 (1951). 36. R. Latarjet, Effects of radiation and peroxides on viral and bacterial functions linked to i DXA specificity. Ciba Foundation Symposium, Ionizing Radiation and Cell Metabolism S pp.275-308 (1956). 87. F. H. Dickey, G. H. Cleland, and C. Lotz. The role of organic peroxides in the induction of mutations. Proc. Natl.Acad.Sci. U. S.35, 581-586 (1949). 3S. L. S. Altenberg. The production of mutations in Drosophila by tertiary butyl hydroperox ide. Proc. Natl. Acad. Sci. L . S. 40, 1037-1040 (1940). 39. A. Loveless. Qualitative aspects of the chemistry and biology of radiomimetie (mutagenic) substances. .Vciture 167, 338-342 (1951). 2.0. M. R. Chevallier and D. Lvzatti, Action mutagene specifique de trois peroxydes organiques 4 sur les mutations reverses dedeux loci deE. coliT-9-13. Compt. rend. acad. sci. 250, 1572' 1574(1960). I i i t \ 1 208 DISCUSSION: PEROXIDES AND NEOPLASIA 41. R- W. Barratt and E. L. Tatum, An evaluation of some carcinogens as mutagens. Cancer Research 11, 234 (1951). 42. F. H. Sobels, Organic peroxides and mutagenic effects in Drosophila. Nature 177, 979-982 (1956). 43. 0. Wyss, Peroxides and mutations. II. Paper presented to Symposium on Implications of Organic Peroxides in Radiobiology, Argonne, Illinois, May 7-9, 1962. 44. C. Auerbach and J. M. Robson, Production of mutations by allyl-isothiocvanate. Nature 154,81 (1944). 43. D. Herman and P. Kotin, Lung cancer associated with the inhalation of Fiberglas plastic dust: A report of two cases. In preparation. 46. F. J. C. Roe and 0. M. Glendexning, The carcinogenicity of beta-propiolactone for mouseskin. Brit. J. Cancer 10, 357-362 (1956). 47. P. Kotin and H. L. Falk, Progress Report to Public Health Service (1959). 48. H. H. Smith and A. M. Srb, Induction of mutations with beta-propiolactone. Science 114, 490-492(1951). 49. W. C. J. Ross, Biological action of X-rays, nitrogen mustards, diepoxides and peroxides. Vat it re 165, 808-810 (1950). 50. M. Demerec, Mutations induced by carcinogens. Brit. J. Cancer 2, 114-117 (1948). DISCUSSION Latarjet: How many animals on that slide? (Kotin and Falk, Table VIII) Kotin : Fifty animals to an experiment, fifty C-57 black mice. Epstein: Do you have aerosol experiments? Kotin: Yes. The only thing we have been able to produce with aerosols thus far has been squamous metaplasia of the tracheobronchial epithelium and an increased tumor yield in A strain mice. No effect at all on C-57 black mice. C- F. Decker: What age animals do you use? Kotin: The animals are exposed at 6 weeks of age. Milas: Were any animals tested with more than one dose? Kotin: No, all at one dose. The concentration of total oxidant was always 0.5 to 0.75 ppm, measured as oxidant. A Voice: Did you pick these substances on the basis of their mutagenicity? Kotin: Actually they are classified on another table on the basis of mutagenicity. We at tempted to analyze the data in terms of going back to the literature of Dr. Sobels and the like, and to see whether they were related to the mutagenic compounds at all. We find no constant correlation between mutagenicity and carcinogenicity, although more of tlie carcinogenic com pounds have been reported as being mutagenic than not. That is all that we can say. There is no statistical analysis of the data. This is a crude bioassav, and I think I should indicate that, my invitation to be here is pre sumably predicated on the fact that we are the only ones foolish enough to waste this much time bioassaying these compounds. (Laughter) Wallinc: Do you have comparable data on a series like aldehydes, ketones, etc.? Kotin: I should mention that methyl ethyl ketone peroxide is another peroxide that has just produced its first tumor in 11 months. We have bioassayed formaldehyde, acetaldehyde, and propionaidohyde, with no tumors, in incomplete experiments. The reason I say that some of these experiments are not finished is that we do not report any carcinogenicity without doing the experiment twice. Milas: Would you care lo comment on the early work of Koch and Maisin,1 who found that peroxides cure cancer? 1 Cumpt. rend soc. hiol. 120, 106 (1935) ; see also Maeat, Z. Krcb*forch. 27, 37S 11928). URL 04125 .`01 ad: mai 'i pap pub wh om in i hen the ONI lh: KI do (IV 111: .`ns. Cancer 77, 979-9S2 lications of ite. Nature glas plastic for mouse- ience 114, peroxides. .r has bp-u yield iii A 0.75 p, ty. We id the i. ;0 const genic c v. The . t - iere is much e that ehvde 1 it son, ; iiout d' . a found it J28). DISCUSSION: PEROXIDES AND NEOPLASIA 209 Kotin: Yes. I think one of the truisms in carcinogenesis is that at a given concentration a carcinolvtic agent may be carcinogenic, and a carcinogenic agent may be carcinolvtic. There is nothing incompatible with those findings. Rather, I think this demonstration of carciuogenicitv is not incompatible with this effect. Along a different line, we assayed carcinogenic compounds for 2 years by attempting to see how much carcinolvtic activity they had. This was a blind alley, very regrettably. Harris: I should like to extend the comments vou made on heated fats: I think it was Roffo5 who first reported a carcinogenic effect of heated fats. People forget that Roffo heated his fats to about 300C, which would crack the fats. This is not the wav fats are treated when used for human food. The temperature is 350'F. Studies done with fats that have been heated at 3503F have not shown this carcinogenic effect. About 12 years ago. because of this work, we did some studies with heated fats. We found that if we actually abused a fat by heating it for quite a long time, and then fed it in the diets of rats, in about 3 weeks they were dead. We became quite concerned about this. We analyzed the diet that had been fed during this period, and we found that essentially all the fat-soluble factors had been destroyed while standing in the refrigerator during this period. All the toco pherol, all the vitamin A, all the vitamin K, and some fair amount of some of the B complex factors were destroyed. We repeated this work with a diet newly mixed every day, and the `'toxicity" was much milder; the animals developed a mild diarrhea but recovered quite well. One of the effects of heated fats is the destruction of nutrients in the food with which it is mixed. Since heated fats may be quite well peroxidized, we wondered if part of the toxicity we had observed with heated fats was due to peroxides that were absorbed into the tissues and that destroyed nutrients in the tissues. As I mentioned yesterday, this was why we did a study on vitamin A-deficient diets with animals. We found that rats on a vitamin A-deficient diet developed the deficiency about 25 to 30% sooner with heated fat than with ordinary fat that had not been heated. An analysis of the tissue showed that much of the vitamin A had been destroved in the tissues. This then made us wonder whether the peroxide caused this destruction. As I mentioned yesterday, the evidence indicates that apparently the peroxides are not absorbed. Thus it was some chemically active material in the heated fat which was absorbed, that caused the destruction in the tissues. 1 You will notice that people who are connected with academic institutions often publish papers on heated fats that are quite alarming, whereas those connected with the food industry publish papers that appear to indicate that heated fats are innocent. I think the truth is some where in between. In academic studies, fats have been overtreated and abused in order to bring out an academic point, whereas in the food industry studies are designed to see what happens in the practical use of normally processed fats. 4 This problem is not so alarming as we think. It seems that the only important problem in heated or oxidized fats is the destruction of tocopherol in the fat before it reaches the palate of the consumer. For instance, tocopherol is destroyed more rapidly than the man-made anti oxidants, BHA and BHT, that are added to fats that are to be used in frying. ' In studies that we have made, maybe as much as 95% of the tocopherol was destroyed in the f ordinary heating of these food fats, whereas the BHA and BHT survived. Because the fats do not become oxidized, the food industry feels it has solved the problem. But it should be realized that none of the man-made antioxidants completely replaces vitamin E in the tissues. Vitamin E has the ability to get into tissue cells and function as an antioxidant. These other antioxidants do this incompletely, if at all. A fat that has been treated this way is not protected by anti oxidants in terms of human tissues. Therefore the stress from a greater use of these oxidized fats may cause deficiency. The compounds in fats that cause toxicity do not form urea adducts. Ap- Bol. In*t. Med. Exptl. Estud. Cancer 15, 539 (1938). s I I 210 DISCUSSION': PEROXIDES AND NEOPLASIA parently they are branched-chain compounds, and most of them are peroxidic. Their importance is mostly related to the side reactions which occur in contact with food. Kraybill: In agreement with what Dr. Harris has just said, I should like to report again the work that Da Costa and Levenson did some years ago* in which they produced sterility in rats. This is primarily due, we think, to the radiation end products which have an effect on vitamin E. It can be corrected by the administration of vitamin E. This was found later also in subse quent experiments in which dogs were fed various types of irradiated foods. So it has become quite essential to give the fat-soluble vitamins, E, A, and K, orally to the rats, not simultane ously with the ration. If one leaves the vitamin mixture in contact with the irradiated ration, in a couple of days the vitamin is destroyed. Later, about 1958, it was also Qoted that in feeding irradiated foods, like beef, there was a destruction of vitamin K. Particularly if one leaves menadione out of the diet, with certain types of diets one produces hemorrhagic diathesis in rats. The effect is more marked as the radiation dose used for processing food is increased. > In the case of vitamin K destruction and hemorrhagic diathesis, there is also a vitamin A- vitamin E relationship, and an influence here of estrogens and certain bile acids. If you ad minister a ten- or twelve-carbon aldehyde to the animal, you will get a hemorrhagic syndrome confirming the influence of carbonyls. If you take soybean oil or cod liver oil and autoxidize them, you can similarly get an effect on vitamin K. Again in agreement with what was said, Dr. Mead at California found' that fats with peroxide values of 200 were somewhat toxic, and at 300 or 350 they were not only toxic but lethal in effect. Last fall, in a meeting in Brussels, Dr. Lang from Germany reported* that autoxidized fats would produce certain pigmented lesions in the liver of the rat. There are thus many effects of oxidation of vitamins and formation of these degradative products on physiological action. Epstein: I should like to ask Dr. Harris what evidence there is that BHA and BHT are not taken up by cells. Harris: BHA aiad BHT are among those compounds that are absorbed into the blood stream and exert an antioxidant effect extracellularly. They do not completely replace vitamin E. I think the reason is that they are rejected in part, or entirely, by certain cells, so that intro- celluJarlv they are not effective. Epstein : What evidence is there that they are not taken up by cells? Harris: Because they do not replace vitamin E. Epstein : That is not the answer to the question. Surely if they were taken up by cells in the absence of vitamin E, that would mean that the uptake would exert an opportunity for pro tection. Harris: I don't know the answer to that. I am saying that they do not replace vitamin E in the diets of animals, which indicates that there has been a rejection somewhere. Epstein: They may be exerting a different kind of antioxidant effect. Harris. Yes. but you get the vitamin E syndrome pathology even in their presence. Epstein: Dr. Kotin, is there any evidence that these aliphatic oxidation products are ob tained in tobacco distillates? That is the first question. The second question: Is there any evi dence that oxidation products of polycyclic hydrocarbons are found in the atmosphere, and, if they are found, are they carcinogenic? Ivotin: In answer to the first question, I know of uone. Secondly, what we did was to expose * V. S. Army Med. A'utr. Lab. Repl. 89 (19.51). 4 Federation Proc. 15, 918 (1956). * FAO/UN Technical Meeting on the Evaluation of the Wholesomeness of Irradiated Foods (Brussels, October 1961). <o. a to a:- . iracc ai the G 6i V iped h n aD' ii our; t' t sm <OTr 1 -nt p \ ed oi t bas r dily l mb Sta' r~ e an 3D 1-- virt o pare -U -se p ro tie oi Seco it tha eupa adily ! the ori. ith a> j nd hi- G. ( 1 'lllS? Koi other i nonim: Life I hmg r Bei i ui oil- In-con Kit- I rtaftce in the n rats, itamin subseecome iltaneratiou, was n certain as the nin Aou alldrorncoxidize eroxidu thal iu zed fo lects ion. are n- ? i stre.' i lin E I it int .- Is in for i nin 1 i are any e, an- i o ex ed I' -i? DISCUSSION: PEROXIDES AND NEOPLASIA 211 to an oxidative atmosphere the crystalline carcinogens, 3,4-benzupyrene and 1.2,5.6-dibenzanthracene, both in crystalline form and adsorbed on soot, and bioassayed their carcinogenicity, and the property was attenuated. G. Cohen: Would you comment on the problem of carcinogenesis in man versus carcino genesis in the mouse? You gave one or two examples of workers exposed to peroxide who de veloped cancer subsequently, with the implication that this was perhaps the cause. Have there been any statistical studies of people exposed over a long period of time, and would you include in your answer a comment about the incidence of cancer in smog-bound Los Angeles as compared to a smog-free city? Kotin: I have a habit of answering the last question first. According to the best estimates, the latent period for the induction of pulmonary neoplasms is one and one-half to two decades, based on population exposed to cigarette smoke. Here you do have a restricted population, and the basis for a carcinogenic effect can be properly anticipated. Such an effect would not be readily manifest in a migratory population such as Los Angeles represents. I might say, though, that a five-year study supported by the American Cancer Society and the State Department of Health is now in its third year, and the findings in Los Angeles indi cate an increased risk on the part of residents over and above that one would normally anticipate by virtue of their living in an urban area. Living in an urban area gives you an increased risk. Apparently there is an enhancement of this risk, but you do have this catch: How many of these people came from breathing the polluted atmosphere of Chicago or New York at the time of the induction of the neoplasm? I don't know. Secondly, it was not a peroxide, it was Plexiglas and isothiocyanate that we are reporting. All that we are reporting is that we have the induction of two pulmonary' neoplasms in persons occupationally exposed to these products. The product is present in the neoplasm. The two most readily available pathogenic environmental factors have been ruled out. So in the manner of the original cases of asbestosis which were reported as the appearance of lung cancer in persons with asbestos bodies, we are reporting these as two cases of human lung cancer, Plexiglas fibers, and history of exposure to Fiberglas. G. Cohen : You also suggested benzoyl peroxides in industrial sites. Is there any evidence on this? Kotin: I put a good healthy "theoretically" in front of that. The only evidence is that, all other things being equal, industrial populations have a higher incidence of lung cancer than nonindustrial. I should say industrial policy holders according to the work of the Metropolitan Life Insurance Company. They have a significantly greater liability to the development of lung cancer than ordinary policy holders. Bernheim: Has anybody bioassayed haze? As I understand it, haze is due to the distillation of oils and waxes from the leaves of trees by the hot sun, and once in the atmosphere these become highly peroxidi2ed. ^ Kotin: To my knowledge, it hasn't. 5 A t i i. < URL 04128 Behaviour of plastics and rubber foams in fire Keith Paul The behaviour of piasiics and rubber foams when burned depends on man> factors including the type of material used, its form, any interactions with other materials used and the design of the product. It also depends on the many environmental and human factors affecting the actual fire in which the product is burned. The majority of plastics and rubbers can be prepared as foams while many individual polymeric foams can be pre pared with a range of properties. Polyurethane foams are well known for their wide range of properties, examples of which are indicated below. low and medium density, flexible foams for upholstery and bedding, etc; medium and high densiry flexible foams for packaging and impact-absorbing applications; low density, rigid foams for insulation purposes; medium density, rigid foams for insulation and structural purposes; nigh density foams for structural purposes: Integra; skm, medium density, semi-flexible materials for impaci-absorbing uses e.g. crash pads; integral skin, medium high density rigid materials for structural mouldings: micro-cellular and non-foamed materials are also available, The complexity becomes even more apparent when the foams are combined with additives ranging from glass fibre reinforcements to flame retardants. With the exception of the integral skin foams, few foams are used on their own but are invariably used in combination with other materials. Examples include laminated, foam core building panels and fabric-covered, foam-filled upholstered furniture. Other applications include tn situ generated foams which may be injected into cavities or which may be sprayed onto structures and then coated. Polyurethanes probably have the widest range of uses but other foamed polymers with a range of properties and applications can also be produced. Many foams are chemically similar to their unfoamed polymer bases but differ markedly in density and consequently in thermal properties. It is proposed to consider the effect of these physical changes on the burning behaviour of polymers on a genera! basis and then to consider the burning behaviour of typical foam-based products selected to illustrate the many types of product and behaviour encountered. In some products, residua) blowing agents or the breakdown products from blowing agents may be present although the quantities are likely to be insignificant. Or Paul is head of the fire study section at the RubDer and Plastics Research Association of Great Bnta;" {RAPRA' Shswoury, Shrewsbury, Shroosn/e SY4 4NR. c: <* r~>,rfr-r si*'*' /V- 6J Effect of cellular structure on fire performance Temperature/time curves are given for test walls without insulation and for walls insulated with 25mm of poly urethane foam and 50mm expanded polysryrene flameretarded additive, both protected with 12.5mm of plaster board (Fig 1). The plasterboard did not fail and at no time did the foams become involved in the fires. The increased temperatures and faster burn rates were due to the insula tion preventing heat loss from the fire. Thomas and Bullerr have studied this theoretically and have shown that: tg tr (KtCr where /g=fire growth time; K = thermal conductivity; q = specific gravity; C- specific heat capacity. The index m is typically 0.25 to0.125, This implies that a fire which takes 20-30 minutes to full involvement in a room of brick/plaster construction would take about one third of that time in one insulated with a non-combustible lining of low1 conductivity. Note that the type of insulation does not matter. Typical values of K^C are given in Table ! lor various wall lining materials. It has been shown that some polymeric materials have an uT' exponential rise in burning rate with temperature and that for such very rapid fires, the effect of K.pC becomes of secondary importance/ A major consequence of this is the higher temperatures reached in fires and the shorter times to reach flash-over conditions. Tne lane: have been defined by various authors but are typically ceiling tempera'-ures cf 400-600''C corresponding to a heat flux of 2V./cm* on the floor.2 The rate of burning of a material increases with iricreasing Fig 1 E^ec.i of wall irsu!ation or, ceiling Temp^raU'e CO-ner tfeS'. Ceiling Temperature lcC; URL 0412 Sate of burring (mm/mm) Fig 2 Relationship between rate of burning and temperature (small specimens, vertical downwards test)' temperature and the rate of flame spread on a continuous surface approaches infinity as the surface temperature approaches the ignition temperature of the surface material (Fig 2).1 The effect of insulation on fire growth is therefore considerable even when the insulating material is not actually burning. With foamed materials, the insulation effect is actually built into the polymer system and conse quently heat is concentrated at the polymer surface. The net result is that cellular materials are more easily ignited and burn more rapidly than similar non-foamed materials. The `edge effect' of the chin foam walls may also be significant (see Table 2). Rigid polymer foams are used for insulation and struc tural applications because of their very low density and conductivity or because of their light weight per unit volume. Cellular upholstery materials attain their desirable seating properties because of their cellular nature. Consequently, there is no possible way of reducing the insulation effect of foams and it is necessary to consider carefully all aspects of material design and end-use hazard to ensure their proper and safe use. Fira and causaa of fire casualties Fire is the oxidative destruction of a material and is charac terised by the evolution of heat and light. When heated, an organic solid will break down to produce a gaseous fuel w hich mixes with air (oxygen) and burns as a flame or glow. Heat from the combustion causes further breakdown of the material and the process continues providing that fuel, heat and oxygen are present. Removal of one or more compon ents causes the flame or glow to cease. In 19^6, a major research project was begun in the Strath clyde area of Scotland4 to determine the physical causes of death in fire and the circumstances leading to the fatal fire. Attention was also paid to factors such as the existence of cardiovascular and lung diseases, age, drugs, alcohol, suicide, homicide, etc. The study showed that more than half of the fatalities (54 per cent) had received a fatal exposure to carbon monoxide and that the majority of the remainder involved carbon monoxide in conjunction with burns or cardiovascular or lung diseases etc, with a minor proportion (about 10 per cent) involving other unidentified factors. Fire gases are the major contributor to fire deaths and it is tempting to relate fire deaths directly to the toxic gases produced by a burning material. However, toxicity involves concentration, exposure time and temperature, all of which vary during a fire. Carbon monoxide, which is produced by all burning organic materials, is the main toxic component of the smoke and alone accounts for more than half of the fire deaths. Oxygen depletion can also be significant. People will not willingly inhale fire gases and it is therefore logical to infer that they do so either because they are trapped by the fire itself or become incapacitated or disorientated and thence unable to escape. Such effects may be due to smoke and fire gases causing loss of visibility, irritancy or narcotic effects. The role of minor concen trations of other gases at sub-lethal levels is receiving considerable study to determine possible incapacitating effects but the main problem is the loss of visibility and disorientation preventing escape. Trapped persons may sub sequently be overcome by toxic gases. This is indicated by the number of persons who are overcome by smoke and have to be rescued by the fire brigade. Concern has been expressed in the UK by various fire and legislative authorities about the Increasing number and changing pattern of fires. Suspicion has been growing, especially among firemen, that modern, upholstered furniture and plastics foams are relatively easy to ignite and once burning will produce intense fires and large quantities of dense smoke, This concern resulted in a number of major research programmes to investigate fire hazards.3,5 The results of these programmes were in general agreement and provided detailed and comprehensive background inform ation about the fire behaviour of modern materials. The UK Home Office combined these data with surveys of opinions of fire officers and analyses of fire statistics in its recent report about the fire risks of new materials.6 Furni ture manufacturers were recommended to improve the fire performance of their products, giving top priority to ignition resistance to small sources, to the rate of fire Table 1 Typical physical properties of some common materials at 20:C; Thermal Specific Thermal con- heat ductivity Density capacity * inertia KpC Material (W m/K( (kg/'rrv) Ukg. K) (W''mvK)}sec Brick Chipboard Plasterboard Fibre insulation board Polystyrene foam Air 0.8 0.15 0.16 0.053 0.034 0.024 2600 800 950 240 20 1.3 800 1250 840 1250 1500 1000 1700 x 101 150 x I0J 130 x 10' 15 x 10J l x 10' 0.03 x I03 Table 2 Effect of density on rate of burning (o BS4735 for standard polyelher foams' Density of material Rate of burning (kg m;) (mm/min) 15 22 32 10"' (aolldl *D,C no. ft 3.7 2.5 1.6 0* URL 04130 growth, and to reducing the rate of production of smoke and toxic products. These recommendations are perfectly logical because without ignition there can be no fire while a slow burning fire would produce lower concentrations of smoke etc, thereby giving greater time for escape and for first aid fire-fighting. This situation applies to items ignited in the early stages of a fire but once the fire is fully developed, the situation changes and the hot smoke and toxic gases produced become more dependent on the total materials involved and the ventilation, building design, use, etc. Punderson' has described how the use of toxicity tests to rank or rate materials with respect to fire safety has not proved to be a fruitful concept because of the fundamental nature of fire. He concludes that it is impossible to specify materials of zero or acceptable fire toxicity because all burning organic materials produce toxic gases and the degree of toxicity, whether lethal or harmless depends on concentration, and duration of exposure. The wide range of rates of burning recorded for materials greatly exceeds the maximum differences of the specific toxicity of their fire gases and control of ignition and burn rate offers the best means of reducing toxic hazards. Smoke end gases from fires Although many of the chemical compounds found in fire smoke are undeniably highly toxic, there are a number of factors which must be taken into account when considering toxicity: firstly, the nature of the compound in question; secondly, the concentration of the compound; thirdly, the duration of exposure; and fourthly, other compounds present as well as heat and oxygen depletion. Factors 1 and 4 are directly related to the materials burned and to oxygen availability. Factor 2 depends largely on the rate of burning, the smoke catchment volume, degree of ventilation and the amount of material present. The duration of exposure (factor 3) depends on the possibilities of escape which are partly determined by human and en vironmental factors and partly by the density and irritancy of the smoke restricting visibility, factors which relate to the type of material and the rate of burning. Hydrocarbon polymers containing only hydrogen and carbon will burn to completion to produce carbon :de and water. Polymers containing sulphur will adr produce sulphur dioxide and those containing nn. , nitrogen oxides. Halides will be released as hydrogen halides, Complete combustion rarely occurs in natural fires and the products of combustion normally comprise a mixture of pyrolysis, partial oxidation and complete Table 3 Typical analysis of smoke from polyurethane foam' Approximate percentage Cases produced by weight Comments Carbon monoxide Carbon dioxide Hydrogen cyanide Isocyanates Acetonitrile Acrylonitrile Benzonitnle Pyridenc Ammonia Nitrogen oxides Aliphatic hydrocarbons Aromatic hydrocarbons 12-30 60-80 1-3 1 1 1 Mainly above 300C, increased without oxygen Mainly above 300eC Mainly above 600C Mainly 200-300*C Decompose at 800C Mainly during pyrolysis at ~00-1000 C Mainly 400-600C Mainly above 800C Above 300C Above 300CC, increases with higher temperature Table 4 Hydrogen cyanide from polymers at SOOeC' Hydrogen cyanide Material (mg. g) Polyurethane flexible foam Polyurethane flame-retarded flexible foam Polyurethane rigid foam Polyurethane flame-retarded rigid foam Polyisocyanurate foam Nylon Polyacrylonitrile Wool Dried sardines 0.1- 9.0 1.7 3 3.5-17] 4.3- 8.0 9.1-20.' 1.6 20.4 6^.7 5.3 oxidation products. Low temperature pyrolysis (less than 400C) gives a number of complex degradation products which directly relate to the basic polymer structure. Higher temperatures (400-700C) produce a complex mixture of degradation and partial oxidation products including carbon monoxide (in large quantities), aldehydes etc. High temperatures (>800*0) combined with high oxygen avail ability tend to produce the simple products of complete combustion. In actual fires, products from each of these three zones are present, their relative concentrations depending on both material and fire parameters.8 Typical breakdown products from polyurethane foams are shown in Table 3.' Polyurethane foams in fire are invariably linked to the production of hydrogen cyanide but the assumption that removing polyurethane from fires would eliminate hydrogen cyanide is wrong as it is produced by many other materials (see Table 4).10 Despite the prevalence of hydrogen cyanide in smoke, however, the main toxic hazard in fires is carbon monoxide which is produced in ail fires.4 Animal toxicity tests using pyrolysis products from polyurethane foams have shown that hydrogen cyanide gave a statistically significant increase in toxicity over that due to carbon monoxide but that the increase in actual hazard was only marginal.11 Much has been said about super-toxicants but the only proven instance of a super-toxicant is related to the use of trimethylol propane polyols in polyisocyanurate (P1R) and polyurethane (PUR) foams. The formation of trimethylol propane phosphate in very small concentrations was demonstrated under laboratory conditions rather than in actual fires and the result of this work was the reformulation of rigid foams eliminating the use of trimethyl propane and thereby the possibility of its phosphate occurring in fire.': Behaviour of plastics and rubber foams in fire Rigid foams, building and insulation applications A number of rigid foams can be formulated to give a relatively high performance in specified standard tests, for example, polyisocyanurate and phenolic foam boards will give a Class 1 spread of flame rating to BS476, Part 7. As with all organic materials, polymer-based foams will burn if the fire conditions are sufficiently severe. In actual fires, Table 5 Comer tests, unprotected foam - 75kg ertb1 Polyisocyanurate Parameter foam Time to flashover (mins) Time to maximum temperature (mins) Maximum temperature (*C) Maximum smoke density (OD/m) Smoke volume (m5) Ceiling temperature* (*C) Smoke density* (OD/m) Carbon monoxide* (per cent) Carbon dioxide* (per cent) Oxygen" (per cent) Aflr 10 mins 1.5 3 600 >4.5 >4000 875 1.6 0.33 19 6.7 Phenolic foam -- 14 637 1.4 1100 780 0.6 0.16 16 7.9 URL 04131 Heat release rate (W'cm'/s) 14kg wood crib produced rapid flame spread with polyiso cyanurate foam, conditions approaching flashover. Polystyrene (PS), phenolic foams (PF) and furane foams (FF) gave progressively less severe fires with animal mortality data indicating that phenolic foam gave the worst result (Table 6), possibly due to a smouldering reaction which gave high amounts of carbon monoxide. A comparison of bulk materials is given in Table 7 which shows the effect of different material types and of different material foams. As expected, foamed polyurethanes and polystyrene produce hot smoky fires with phenolic foams, wood sticks and wood wool producing similar fire temperatures but less rapid burning and less smoke.14 Although important from a storage point of view, these differences are less important in practice because foam insulation is rarely used in its natural state but is almost invariably protected from mechanical damage. The fire behaviour then becomes intimately linked with that of the composite. Large fire tests with stacks of crosslinked polyethylene and ethylene-vinyl acetate foams produced lower smoke than other plastics foams. Fire gases comprised carbon monoxide and carbon dioxide and a wide range of hydrocarbons which produced an irritant effect. A major hazard was the formation of a pool of burning polymer melt.23 The latter effect has also been observed with other thermoplastics materials. URL M l Fig 3 Relationship between heat release rate and incident heat flux for polyurethane foams14 heat flux values of more than 6-10W/cnr can occur whereas the maximum heat flux in BS476, Part 7 is only 3.7W/cm2. BS476, Pan 6, the fire propagation box, exposes samples to 5W/cm:.15 Figure 3 shows that an increase in heat flux can completely alter the ranking and behaviour of materials.14 In RAPRA corner tests, a Class 1 polyisocyanurate foam burned rapidly when exposed to a heat flux of more than 7W/cm2 whereas a phenolic foam, also Class 1, merely charred under similar conditions (Table 5). Corner tests13 with unprotected insulating foam with a 2.5m corner and a Fire performance of polymeric foams in building applications A major part of this section deals with panels comprising surfaced polyurethane boards but the principles indicated also apply to boards based on other foams. Ignition tests using sources such as matches, crumpled newspaper and fibreboard cribs failed to cause sustained flaming of any of the board surfaces tested. Although these sources are minor, they represent ignition sources likely to occur on building sites etc, and show that the boards concerned are not likely to be ignited accidentally. Tests to BS476, Part 6 indicate the different performance when the panels were exposed to a more intense heat source (Table 8). Intermediate scale tests involved the use of a lm3 Table 6 Comer tests, unprotected foams - |4kg crib 1$ Parameter Soft wood Flame spread (ml Maximum temperature (C) Time to reach OD m = 0.26 (mins) Maximum carbon monoxide (per cent) Maximum carbon dioxide (per cent) Minimun oxygen (per cent) Maximum hydrocarbons (ppm) Maximum hydrogen cyanide (ppm) Maximum nitrogen oxides (ppm) Rat mortaJity (rank order A = worst) 1 220 CD 0.1 5.8 12.8 <100 0 5 E Table 7 Burning characteristics of materials - small room lest ` Material Newsprint Wood - pine Wood - oak Wood Phenolic foam Phenolic foam Polystyrene foam Flame-retarded polyurethane foam Flame-retarded polyurethane foam Polyisocyanurate foam Form Folded Sticks Sticks Woo! Granules Board Blocks Blocks Blocks Blocks Maximum temperature CO 23"? 748 130 729 82 770 1100 880 850 77 I P[R 2.5 560 4 1.6 5.9 10.5 1900 380 45 B PF 0.5-1 340 g 1.2 7.3 11.8 1000 50 12 A PS 1.5-2.5 230 5 0.4 6.1 11.4 400 0 0 C Time to maximum temperature (minsl 4 7 7 4 7.J 75 6.5 4.5 35 55 Rate of temperature rise (~C min) IS"1 187 30 '60 13 148 458 419 644 482 Maximum smoke density (OD/m) 0.4 0.8 -- 2.7 _ 1.2 )8.: 15." 15.2 15 5 Maximum smoke rate (mJ min of ODm = l) 0.8 55 -- 0.8 _ 36 139 225 165 130 FF 1 300 02 5.6 13.2 <100 130 130 D Total smoke <m' of OD- m = 1) 8 38 -- 38 _ 9 |4<2 503 482 322 ^ Table 8 Fire propagation indices of panels containing polyurethane foams' Indices BS476. Par: 6 Surface Core i I None None AcTylic paint, aluminium foil Hardbo&rd Plywood Plasterboard Plasterboard PUR PIR PUR PUR PUR PUR PIR 24 11 16 7.9 6.0 3.9 4.2 32 19 4.1 29.0 25.2 85 9.0 ventilated steel box.1 Measurements of smoke, temperature, carbon monoxide and hydrogen cyanide were made during a series of tests in which foam-based panels were exposed to a 3kg, fast burning crib on the floor of the box. All panels consisted of a 25mm foam slab (except the penultimate which was 16mm thick) bonded to the surface shown. Fire retarded polyurethane foam was used for all the tests except the ultimate which used polyisocyanurate foam. Table 9 shows that when the polyurethane and polyiso cyanurate foam boards were protected by 9.5mm plaster board, the concentration of carbon monoxide was similar to that of the ignition source alone and that hydrogen cyanide concentrations remained less than 20ppm. in contrast, carbon monoxide concentrations of 2.5-5.5 per cent and hydrogen cyanide concentrations of 150-250ppm were recorded when the polyurethane foam was used behind flammable coverings. The painted aluminium foil did not thermally insulate the foam and actually melted during the test, thereby enabling the foam to burn, producing similar atmosphere to that with the wood/foam panels.1 Corner tests with a 16kg crib placed 50mm from the test walls showed that 12.5mm plaster offered considerable protection to the 25mm extruded polystyrene foam boards.1* Even when the plaster was partly removed and the foam directly exposed to the fire source, the resultant fire was limited in area and went out when the fire source was extinguished. The use of 9.5mm plasterboard offered less protection than the 12mm plaster but even so the protection was considerable. Tests showed non-burning adhesives should be used to minimise the extern of burning behind the plaster. In tests using a much larger crib (50kg), the plaster again offered considerable protection but collapsed as the thermoplastic foam melted. In these tests, the plasterboard was bonded to the foam which was in turn bonded to the wall. The foam melted leaving the plaster unsupported which then collapsed giving rapid flashover and a very smoky fire of brief duration. The latter was due to the low Table 10 Penetration of polyurethane foam ceilings, crib test Penetration Smoke Penetration of adjacent (OD- m = 1.4 in Type of ceiling of ceiling room adjacent room) Plasterboard 02mm) + 50mm glass fibre Fibre insulating board Polyurethane Polyisocyanurate 29.5 21.5 6.75 11 None 32.5 8.5 14.75 38 26 10 15 Table 11 Penetration of polyurethane foam ceilings, high severity tire test" Penetration Failure of Full fire of adjacent ceiling in loft room Type of ceiling (mins) (mms) (mins) Plasterboard (9mm) * skin coat Polyisocyanurate (12mm) + 12mm plasterboard + ceiling finish Polyisocyanurate (12mm) f ceiling finish + 50mm mineral wool inadequate laying Polyisocyanurate (12mm) + ceiling finish + 50mm mineral wool good laying 21 25 10.5 20 21 25.25 11.25 16 None None 13 None density of the foam and the small mass of material actually present. A combination of mechanical fixing of the plaster to the wall with adhesive bonding should increase the time to failure.1 Corner tests involving a 50kg crib with plaster board (9mm) covered with polyisocyanurate and poly styrene foams have been discussed earlier and their temperatureritme profiles are shown in Fig 1. In these tests, the plasterboard was mechanically fixed to the wall (Fig 1).` The use of polyurethane foam insulation panels in ceilings has been questioned following fires in residential buildings. Fire investigations and large-scale tests at the Fire Research Station16 concluded that the use of polyurethane ceilings could lead to increased fire hazard because of easier penetration of the roof space and adjacent rooms. The rate of burning of the roof itself was independent of the ceiling, being related to the roof timbers and sarking (the weather barrier between the tiies and the roof - often of bitumenised paper). Ceiling penetration was not observed with plasterboard ceilings and although penetration occurred with fibre insulation board ceilings, it was onlyafter 30 minutes compared to 8.5-J5 minutes for poly urethane and polyisocyanurate boards (Tables 10 and 11). With hindsight, it is easy to point out the poorer perfor- URL 04133 Table 9 Burning characteristics of panels coniainieg polyurethane foam Maximum core Maximum surface Test no. 1 Panel surface None , Acrylic * painted temperature rc> _ - temperature <C) 725 2.5 Maximum smoke 0.66 3 ) Aluminum I foil 780 725 >1.7 3 Hardboard 64 900 775 8 >1.7 4 Plywood 15 8 BOO 800 10 >1.7 5 Plasterboard (9mm) 12 11 90 325 )0 0.`1 33 3 6 Plasterboard (12mm) 100 600 0.83 43 3 Maximum CO (per cent) 0.4 3 1.0 6.25 2.5 8.5 5.6 11 0.4 7 1.4 3 Maximum CO; (per cent) 6.0 6 S.O 7 15.6 9.5 15.6 12 12.8 5 7.5 35 Minimum O; (per cent) n.4 3.5 12.0 6 2.75 8.5 1.25 12 7.0 5 13.5 < Foam maximum HCN (ppm) 12 3.5 150 12 150 6.5 250 ]l 5 6 20 45 Table 12 Poiystyraae foam catlings In Arc1' Thickness Adhesive (nun) Finish Five dabs 9 None All over 9 None Ail over 9 Emulsion paint All over 9 Gloss paint AJl over 9 Ceiling paper All over 9 Plaster Fire growth Rapid fire spread, htiming debris Limited spread, no burning debris Limited spread, no burning debris Rapid fire, burning debris Rapid fire, burning debris Very limited, rate determined by failure of plaster mance of the synthetic materials but this type of application is controlled by building regulations which specify tests and performance levels for the products used in buildings. In the case of ceilings, fire penetration tests are not required if the space above is not occupied. In addition, polyisocyanurate foam boards can comply with the Class 1 spread of flame rating (BS476, Part 7). This is a good example of the development of new materials and products outstripping that of tests and test specifications. Although these short comings could be and were shown by ad hoc tests, the existence of regulations would tend to discourage this type of research. The performance of other materials would be one of pure speculation in the absence of experimental data. Polystyrene foam would be expected to melt causing rapid penetration and while other foams such as phenolic materials would not melt, shrinkage cracks may occur. In the absence of experi mental data, the only satisfactory method of decreasing the penetration risk of foam ceilings is by protecting the foam by the use of materials such as gypsum plaster (Table 11). Polystyrene foam is widely used for ceiling insulation, as insulation boards, in suspended level ceilings and as `do-ityourself tiles. The total amount of polystyrene present in a given room is relatively low because of the low density of the tiles and even if burned to completion, the smoke and toxic gases produced would be relatively low. A large number of tests based on simulated fat frying fires (using a 6.5kg wood Table 13 Comparison of traditional and stKl/polynrrthiK foam panel buildings'' Traditional Steel/PUR 5ieeI/PUR mineral wool panels * cavity Parameter ABC Minimum visibility (m) Maximum carbon monoxide tppm) Maximum hvdrogen esamde (ppm) Nitrogen oxides (ppm) 3.5 200 Trace . 6 2.5 100 <5 but peaks of 8& P 8 2.5 100 <5 but peaks of 8 & 17 8 Table 14 Comparison of structural foams iSPl/L'L study)' < ABCD UL94 Oxygen index (per cent) Ignition temperature (flash) (C) Ignition temperature (self) (*C) Ignition temperature (OSU) (C) Smoke density - ASTM E662 <Dm) Smoke obscuration - ASTM D2843 (per cent) Rank order Full-scale room test Smoke production Carbon monoxide VO 5V 38 425 560 370 448 82 79 1 A A VI JV 29 414 475 1310 732 346 VO 5V 27 395 470 2480 750 791 VO JV 27 370 400 2870 752 577 94 96 95 234 EBC EDB E VO 5V 33 422 560 530 389 86 69 5 D C Table 15 Ignition characteristics of structural plastics foams and traditional materials used in furniture applications' Minimum Ignition time (mins) at heat specified heat flux Material (W Zero') 2.5W/cmJ 3.75W/cmJ 5W/cmJ Chipboard Plywood Hardboard Polyurethane foam Flame-retarded polyurethane foam Polystyrene foam Polypropylene 3.0 2.2 1.9 1.6 1.3 1.6 1.1 Chipboard 2.9 Plywood 2.9 Hardboard 2.2 Polyurethane foam 2.5 Flame-retarded polyurethane foam 2.2 Polystyrene foam 2.2 Polypropylene 14 * Mm heat Hun 10 cause sustained burnmi Pilot flame ignition 6.2 7.6 4.8 7.6 4.8 3.9 2.5 4.5 3.8 3.5 2.2 7.5 5.4 43 _ -- 110 14.0 3.2 3.4 3.3 Self-ignition 6.0 64 5.7 3.2 2.0 2.7 3.0 3.9 4.2 4.1 2.6 16.5 3.9 9.2 4.3 6.3 3.5 2.8 2.9 2.7 crib) and a full-scale room burn (200kg wood crib), proved that the amount of foam burned depended largely on the method of fixing of the tiles and boards1' (see Table 12). The use of 50mm foam boards instead of 9mm tiles produced a rain of burning molten polymer. Flame retarding additives did not significantly affect the results of 9mm thick tiles but resulted in a non-burning rain of molten polymer with the thicker (50mm) materials. With suspended level ceilings in which sheets of material softening below 120C to BS2782, Part 102C are loosely supported on an inverted T-frame, flame-retarded polystyrene foam slabs fell to the floor without igniting. Plaster offered considerable protection to the polystyrene foam but could only be effective as long as it remained in position. As poly styrene foam is a thermoplastic, the plaster should be mechanically fixed e.g. by nailing to the ceiling structure and should remain intact during the fire if it is to be fully effective. Steel/PU foam building panels. Full-scale tests of buildings constructed from polyurethane foam-filled steel panels and of traditional buildings, were carried out in buildings measuring 30ft x 15ft with a sloping room 17.5ft x 15ft high.20 The traditional building was insulated with mineral wool covered with fibre insulating board treated to give Class 1 spread of flame (BS476, Part 7) performance. The fire source consisted of timber cribs of 635kg giving a maximum temperature of 740C. The roof of the traditional building burned fiercely, causing burning bitumen to spread fire, but although there was damage, this did not occur with the polyurethane foam building. Fire did not penetrate the walls of the panels in which the steel facings were in direct contact with both sides of the foam. In contrast, fire spread along the wall comprising 25mm polyurethane foam with a 25mm cavity. It was concluded by the Fire Research Station that the use of polyurethane foam cored panels did not add to the severity of the fire and did not produce any serious toxic hazard within the building (Table 13). The Fire Offices Committee accepted that steel-faced panels, completely filled with polyurethane foam did not increase the risk above that of conventional steel cladding and should not attract increased premiums,20 The data show that polymeric foams do not constitute a hazard provided URL 04134 Table It Burning behaviour of three-piece suites upholstered with flexible poivether foam with different structures Total Maximum Maximum smoke smoke smoke volume Cover Structure temperature <C) densitv (OD m) (OD/m = l m` x I0:) Cellulosic Plasticised PVC Wood Polystyrene foam Polyurethane flexible foam Wood Polyurethane rigid foam 540090 590-640 500-*700 4SO 530 1.8-1.9 2.9->3 1 9-2.5 2.9 >3 5-9 >24->30 7-1 1 17 >29 Maximum carbon monoxide (per cent) .33-.35 .85-1 7 .35- 4 0.37 13 Maximum carbon dioxide (per cent) 17-20 15-18 I3-J5 6 12 they are adequately protected from fire. The protection must obviously remain in position and must resist the heat flux values imposed by a real fire. A major problem exists for specifiers in that no standard test exists which imposes heat flux values greater than 5W/cm: and ad hoc or fullscale tests are often needed for critical applications. Corner tests are increasingly used to assess the performance of wall linings, building panels etc, and are now being considered for standardisation. also measured smoke and carbon monoxide. Five structural foam materials were used but these were not identified except by their performance in standard tests. It was concluded that small-scale tests can identify structural foam materials which ignite readily (\JL 94, Setchkin), flame spread (radiant panel, OSU calorimeter) and to the rank order of smoke release (XP2, NBS and especially the OSU calorimeter). An important conclusion was that structural foam materials could be prepared which resisted the internal ignition sources (see Table 14). URL 04135 Fire performance of structural or integral skin foams Structural foam consists of a low-density, closed-cell, cellular core contained within an integral polymer skin. The resulting structure gives a high stiffness to weight ratio. Structural foams can be produced using chemical blowing agents, or by gases such as nitrogen, pentane, fluorinated compounds etc. They may be single thermoplastic materials, layered thermoplastic materials made using the sandwich process or, alternatively, may be a crosslinked material made using a casting or reaction moulding process. The latter is usually a polyurethane but thermoplastic foams may be prepared from many polymers including polyacetal, ABS (ter-polymer of acrylonitrile, butadiene and styrene), polyamide, polycarbonate, polyester, polyethylene, modified polyphenoline oxide, polypropylene and poly styrene. The burning behaviour of structural foams from these polymers and especially the smoke and toxic gases produced will obviously vary enormously. Little published inform ation is available concerning their burning characteristics but theoretical considerations would indicate that they would be expected to burn faster than the unfoamed polymer bases. Structural foams are increasingly used as replacements for metal cases for business machines, computers etc. A major project was carried out by the Society of the Plastics Industry and the Underwriters Laboratory in the USA to determine the behaviour of structural foams in fire using standard, small test methods.21 An important aspect of the work was the use of `internal' gas ignition sources representing burning electronic components inside the cabinets. The tests concentrated on ignition and burning but Table 17 Comparison of fully furnished office burns' Parameter Wood-based office Maximum temperature (C) Time to maximum temperature (mins) Time to OD/m (mins) Time w OD/m (mins) Maximum CO (per cent) Maximum CO, (per cent) Minimum 03 (per cent) Maximum NO, (ppm) Maximum HCN (ppm) 600 123 5 8.5 0.05 2.25 20 40 3 Structural foambased office 775 11 3.5 3.5 3 12.5 0.5 50 300 Structural foam furniture. A comparison of the ignition resistance of structural foam materials used in furniture is shown in Table 14 and 15.' Two main structural furniture applications exist: first, as the main frame or shell of upholstered furniture and, second, as simulated wood in cabinet furniture. The latter application is similar lo that of materials-handling and machinery covers mentioned above, but, in contrast, plastics foams are used to replace wood rather than metal. Upholstered furniture. When used as chair shells, plastics foams are not structural foams within the definition given above but are, strictly speaking, foams used for structural purposes. They do not have an integral skin and correspond more closely to higher density versions of the insulation foams described previously. When used as a chair shell, polyurethane and polystyrene foams permit the use of complex, curved designs without the complicated construction needed to produce a similar design with traditional materials. The ignition and early burning characteristics of upholstered shell chairs will be largely determined by the covering fabric and soft upholstery which are discussed in the next section. The foam core would not be expected to burn until the later stages of a fire. If ignited, both polyurethane and polystyrene foam shells would be expected to make a major contribution to the heat, smoke and toxic fume hazards, but this would depend on the ignition resistance of the foam used. A polystyrene foam shell would melt while the polyurethane shell would probably not. Both shell materials would be expected to produce a greater fire hazard than the alternative wooden structure although it would be unlikely to be involved in the very early stages of the fire (Table 16). Full-scale tests with structural foam. Ignition tests on moulded and rigid foam furniture showed that it was easier to ignite and burned more readily than traditional furniture producing burning molten drips and dense smoke (Table 17). Certain items were ignited by a match while other items, particularly the flame-retarded polyurethane foams proved relatively difficult40 ignite. In a full-scale fire test, an `all Total vriokelm3OD'm = 1| Total smpkeim'OO'm - 1) URL 04136 Fig 4a Correlation between smoke formation from full-scale tests and results calculated from N8$ i flaming] data (1 * cotton/standard poivether; 2 = acrviic,standard poiyether 3 = polypropylene/standard polyether; 4 = polypropylene/standard polyether; 5 = flame-retarded PVC standard polyether; 6 = flame-retarded PVC,'standard polyether; 7 = PVC,'standard polyetherl1 plastics' office suite including polyurethane, polystyrene, and PVC structural foams gave a fast burning fire with large quantities of smoke. The smoke temperature did not remain high for long, but the emission of dense smoke continued from small floor fires after the temperature had dropped. In contrast, a traditional `wooden' office generated more heat but burned slowly over a long period with less smoke especially towards the end of the fire.3 A true assessment cannot be made, especially in view of the very wide range of improved materials now available, but the limited test data indicate the poor performance obtained from ill-chosen materials. It is now possible to obtain structural foams which will resist the No. 7 crib of DoE/PSA/FRIO and there is an obvious need to investigate the fire performance of structural foam materials in order to avoid the fire situation illustrated by the office test discussed above. In contrast to the many thermoplastic structural foams, rigid polyurethane foams do not melt and as such come closer to the Home Office recommendation6 that structural foam furniture should behave in a similar manner to wood. The criteria cannot be applied in detail because of the Fig 4b Correlation between smoke formation from full-scale tests and results calculated from XP2 (modified ASTM D2834i data1 inherent differences between wood and plastics but the idea is that the ignition resistance of structural furniture should be increased and the burn rate, smoke and toxic gas formation of current products should be reduced. Many timber products and panel furniture are resistant to fairly large ignition sources and the Home Office suggestion6 that large areas of panel furniture should comply with flame spread ratings may be obtainable with carefully formulated structural foams. Fire performance of polymeric foams in upholstered furniture and bedding The burning characteristics of some modern upholstered furniture have been identified as relatively easy ignition (estimated as 90 per cent by match flame, 20 per cent by cigarette) and rapid burning, the latter causing dense smoke, intense heat and toxic combustion products. An extreme case is represented by a chair being ignited by a match and smoke filling an open room (34m3) within 4-5 mins with heat and toxic fumes reaching lethal levels shortly afterwards.3 These characteristics are often considered to be directly due to the introduction of polyurethane foam and while it is an important factor, it is the combustion of cover and core which is important (Table 18a). If many of the traditional fabrics such as woven wool and wool moquettes were used with polyurethane foam, the incidence of match and Table IS Effect of alternative fabrics and alternative fillings Material [gnitability* - model upholstered chair test Time to Maximum maximum temperature temperature (C) <mins) Table 18a Standard poiyether (SPE) foam with different fabrics l Acrylic, cotton, SPE IP 2 Polypropylene/SPE IP 3 Viscosc/SPE IP ISO 283 190 4 3 8 Table iSb Traditional fabrics 4 Wool/SPE 5 Wool/cotion wadd/SPE 6 PYC.'cotton/ SPE 2P 3/4P 6P 125 38 280 4 3.5 9 Table 18c Viscose fabric with alternative fillings 7 SPE 8 Flame-retarded polyurethane foam-f- 9 Polyester fibre 10 Latex foamt 11 Cotton wadd. SPE 12 FR cotton ini. SPE To as 5*52 P > .,greiit. F = f,l igirrue T Estimated aJue IF IP IP IF IF IF 1jquals 'imf 190 170 250 170 324 280 u:ct -c fin 87 4.5 12 11.5 6.5 Time (OD'm = 0.2 mins) 1.25 1.25 8.5 3.5 -- 2.5 8.5 16 3 3 6.5 8.5 Maximum smoke density <OD m) 0.7 0.5 0.3 0.4 0.1 !9 0.3 0.5 1.2 >13 1.1 0.3 Total volume smoke <m') 307 114 71 132 4 500 71 107 177 >200 314 75 Table J9 Comparison of heat, smoke and gas data for fulls furnished room tests Parameter Maximum temperature (BC) Time to maximum temperature (mins) Time to 0.32 OD/m (mins) Time to 0 64 OD m (mins) Maximum CO (per cent) Maximum CO; (per cent) Minimum 0; (per cent) Maximum NO* (.ppm) Maximum HCN (ppm) * Flbnc and main uphoisrerv Wool/cotlon* fibre and wadding 620 25-56 15-16 24-40 0.4-0.7 5-10 8-13 0-408-50 Cotton'wool nylon latex 710-790 20-38 1-5 2.25-6 0.3-0.7 n-i6 2.5-7.5 0-10 0-10 Cotton' standard polyurethane 775 3-7.5 1.75-3 2.5-5.5 0.2-1.8 7.5-18 4.5-12.5 60 40- B0 Flame-retarded PVC/standard polyurethane 735 47 6 8.5 33 15 1 40 100 Flame-retarded PVC.'riame- retarded polyurethane 750 12 6 13 0.5 8 11.S 60 500 Flame-retarded cotton Rame retarded polyurethane 680 117 10 47 2.5 11 2.5 100 56 URL 04137 cigarette ignition would largely be eliminated (Table 18b). The frequently suggested solution of banning polyurethane foam or of using flame-retarded polyurethane foam will not and cannot be effective in preventing ignition, because it is the combination of modern easily ignitable fabrics with foams which causes the problem. The effect of alternative cores with a viscose fabric is shown in Table 18c. All composites will ignite with a match and all except 8, 9 and 12 with a cigarette. In terms of increasing escape time and low smoke production, clearly only 8 and 12 show any improve ment. Non-urethane alternatives 9 and 10 show a major increase in smoke production. The purpose of this table is to illustrate that removal of polyurethane foam cannot alone solve the fire problem as it can only be solved by setting meaningful standards in terms of identified hazards i.e. ignition, heat/temperature rise, smoke and toxic fumes. Table 18b shows data of chairs containing polyurethane which pass both cigarette and match test (BS5852, Part 1) and are considerable improvements on many current fabric/foam combinations. Current furniture controls are aimed at increasing the ignition resistance of composites which indirectly may or may not reduce the rate of burning. A significant decrease in burn rate should, providing care is taken with material selection, reduce the rate of production of heat, smoke and toxic fumes and thereby increase the possibility of escape and of fire-fighting. By careful selection of materials, it is possible to produce furniture of similar burning characteristics to traditional furniture. At present there is not a wide variety of flame resistant materials but this should improve. An important recent development is the fire retarded barrier foams based on polyurethanes and polychloroprenes. These foams, which can individually attain a Class 1 spread of flame to BS476, Part 7 are used as barriers around standard polyether foams to impart increased ignition resistance (often greater than the 126g, No. 7 wood crib of BS5852, Part 2) and to reduce significantly the rate of burning and smoke and toxic gas production (Table 19). It has been shown that the maximum amount of smoke produced from upholstered composites can be approximated from the sum of the smoke produced by the individual components. Good correlations have been obtained between full-scale and model furniture tests using the XP2 and the NBS smoke tests, the latter being operated in the flaming mode (see Figs 4a and 4b). The use of highly flame-retarded fabrics and barrier foams can restrict the amount of foam burned22 and in certain fires, the theoretical amount of smoke and loxic gas may not be produced. While small-scale tests can, in this case, be used to indicate the maximum amount of smoke produced, they are not able to indicate the rate of burning. A similar situation exists with mattresses and beds where once again, the role of the cover, in this case the blankets, is all important. A direct change from a polyurethane foam mattress to a rubber latex mattress in a fully furnished room test resulted in complete destruction of the room furniture compared to destruction of only the bed in the polyurethane mattress test. Table 20 shows that it is possible to obtain more satisfactory characteristics by the careful selection of polyurethane foams, covers etc.3 An important advantage of polyurethane foams and especially the high resilience of fire retarded polyurethane foams is that they do not readily smoulder. In contrast, many traditional materials are veryprone to smouldering. It is natural to compare the performance of new with traditional products but in devising a safer environment, it is necessary to recognise the shortcomings of existing products and to overcome both traditional and new hazards. Fire performance of polymeric foams in transport Fire requirements vary between applications but the US test FMVS5 302 is widely used in the automobile and road transport industries and is applied to internally used materials. A strip of a material or a component is subjected to a small flame and the horizontal rate of burning measured. Specifications exist for maximum rates of burning of 100mm/min, 80mm/min and non-buming in the test. Many of the applications for foams in transport are similar to those discussed previously; they include sound insulation, rigid structural foams, flexible structural foam crash pads and upholstered seating. Discussion and conclusion In its Report on the Fire Risks of New Materials/ the Home Office Fire Department Technical Sub-Committee said `There is no case at present for a general ban on the use of any particular material'. This statement should be applied to polymeric foams for two reasons. First, a ban is not effective as it merely encourages the use of alternative materials, some of which may have even worse fire behaviour than the material banned. Secondly, foams are rarely used on their own but are almost invariably part of a composite and it is the fire performance of the composite which is important. Providing the selection of materials and product design are properly carried out, then products con taining polymeric foams can be produced with satisfactoryfire performances. The only way to reduce the fire hazard of a product is by design based on identifying the risks, and the critical factors in product performance. All product designs are com promises in which the designer has to determine and trade off the various advantages and disadvantages of his product. Many factors are involved e.g. mechanical, t able 20 Comparison of burning characteristics of furnished bedrooms, showing effect of blanket and mattress' Blanket type Acrylic Wool Wool Wool Flame-retarded Mattress type Polyurethane Polyurethane Interior spring polyurethane Maximum temperature (C) Time io maximum temperature (mins) Time to 50 per cent obscuration (mins) Time to 100 per cent obscuration (mins) Maximum CO (per cent) Maximum CO- (per cent) Maximum HCN (ppm) Maximum NO, (ppm) 690 6.3 5 6.5 0.14 6.0 30 50 570 6.5 4 8.25 0.13 3.2 20 14 100 80 8 58 0.03 2.0 1 20 445 24 23 25 0.03 2.4 <30 17 Acrylic Flame-retarded polyurethane 500 8 2.5 8.25 0.13 3.2 <30 15 Acrylic Latex 730 27 5 6.5 0.2 6.3 40 80 URL 04138 comfort, aesthetic etc, as well as fire and it is only by meeting a properly evolved design standard which considers all aspects of the use and potential hazard that a satisfactory product will be produced. Fire tests and test data are only two of a number of factors to be considered in assessing fire hazard which is discussed in BSDD64, 'Guidelines for the development and presentation of fire tests and their use in hazard assessment'. The fire performance of polymeric foams varies with material type and composition. Polymeric foams do not necessarily constitute a hazard in buildings and furniture provided proper consideration is given to materials selection, design, construction, and fire protection of such products. This statement is applicable to all products whether or not they contain polymeric foams and a minimum, acceptable level of hazard can only be obtained by correctly identifying the hazard and by hazard-based performance specifications. Effective control of fire hazard is achieved primarily by increasing ignition resistance (without ignition there can be no fire) and by decreasing burn rate. The latter has a direct influence on heat, smoke and toxicity and attempts to control smoke and toxicity cannot succeed without control of burn rate. Cellular plastics are, by definition, low density products based on polymers and because of their insulation character istics will influence a fire even if they are not burned. Cellular plastics are, from a theoretical point of view, easier to ignite and will bum faster than unfoamed products, but as foams frequently form part of a composite for reasons of wear resistance or aesthetics, it is essential to design the composite to meet relevant safety standards. Ignition and initial burning characteristics frequently depend on the outer covering material. Structural foams are rarely covered but the indications are that material developments can reduce the fire hazards of these materials, but little information has been published. Plastics and rubber foams offer an extremely wide range of very useful and desirable properties and products and, provided all aspects of material selection, design and end-use hazard are carefully assessed, products based on these foams should not present an unacceptable hazard. References 1 RAPRA research projects 2 Thomas, P. H.. & Bullen. M. L., BRE. IP 19. "9, Carlton, Herts: Building Research Establishment 3 Palmer. K. N., Taylor, W., & Paul. K. T.. BRE CP 18. 74, CP 3'75 and CP 21.-'6, Garston, Herts: Building Research Establishment 4 Harland, W. A.. <fc W'ooley, W. D., BRE/IP 18/79, Garston, Herts: Building Research Establishment 5 W ilson. W. J . `Flexible polyurethane foams, uses and misuses', J. Fire & Flam.. !976. 7. 112 6 Home Office Fire Department Technical Sub-Committee Report on Fire Risks of New Materials. London. HMSO, 1978 ~ Punderson. J. O., Fire and Materials. 1981. 5, (1) 8 Woolley, W. D., et al. Interjlam. 1982 9 Beetlestone, K., Hazardous environments', iVot 111. Shrewsbury, Shropshire: RAPRA, 19"'6; Taylor, W., &. Scon. K. A.. Fire performance and testing of plastics'. London: Plastics and Rubber Institute, 19" 5 10 Ashida, K., et al. J. Cell. Plast., July Aug !9"4 11 EJgmton, J. A. G., & Lynch, R. D.. BRE FR Note 1077. Garston, Herts Budding Research Establishment 12 Woolley, W. D.. & Fandell, P. J., BRE. FR Note 1060, Garston, Herts: Budding Research Establishment 13 Fire Surveyor. April 1980 14 Rowan. J. W., & Lyons, J. W., J Cell. Plast., Jan- Feb 1978 15 v an Leer (UK) Lid, Technical Data Sheet No 1, May 1979 16 Ames. S. A., St Fandell, P. J., BRE CP 3 80. Garston, Herts: Building Research Establishment 17 DOW' Chemicals Europe, RAPRA tests. `Walls containing styroliner tB\ K. T. Paul. !9'6 18 Morns. W. A.. A Hopkinson, J. S., BRE CP 73 76. Garston, Herts: Building Research Establishment 19 EPS Association. RAPRA tests, 'Expanded polystyrene ceiling tiles in lire'. K. T Paul. I9'3 20 'Rigid urethane t'oam, full scale fire tesis fcs FRS\ ICI Ref 112/1 21 Nelson, G. L . J Ceil Plast., 1982 22 Du Pont iLk) Ltd. RAPRA tests, `Fire testtng of upholstered composites containing 'onar ir.terhner*'. k. T Paul. !9*0 23 BNL L:J. Beiuv.our of plasiccoie arid e-oco'e :n tires'. 1^81 k URL 04139 THE DUBIOUS WAR BUILDING MATERIALS/STEVEN FLAX Fire horrifies. Leaping flames and billowing smoke, victims carried chok ing on stretchers or lifeless in body bags--the sights arouse a deep-seated urge to control the menace. Major disasters like the hotel fire below can also be a marketing windfall for companies seeking to fight competing products by playing on the fear of fire. This is the story of a company that has done so with an extraordinary outpouring of half-truths and misinformation. Its target: plastic pipe. Allied Tube & Conduit Corp. of Harvey, Illinois, is the largest U.S. producer of rigid metal pipe used to contain and pro* tect electrical wiring. Plastic conduit, mainly made of polyvinyl chloride (PVC), has made heavy inroads, partly because it's easier to install. According to Predicasts, a market-re search firm, its share of the conduit market widened from an estimated 9% in the late 1960s to 54% in 1980. Meanwhile, steel con duit's share fell, from nearly 50% to 32% of a business worth hundreds of millions a year. But Allied, a privately owned company with 1982 sales of about $300 million, has not been content to fight back with conventional salesmanship. Since the late Seventies, Allied has run a campaign to publicize the supposedly unusu al fire hazards of PVC, pitching it at consum ers, contractors, legislators, and officials who write building codes. To give its effort the appearance of impartiality, Allied has set up a not-for-profit organization, the Founda tion for Fire Safety. Supposedly the founda tion, with offices in Rosslyn, Virginia, is dedi cated to disseminating impartial scientific information. But with about 75% of its $750,000 budget coming from Allied, it has often served as a vehicle for anti-PVC propa ganda. The foundation's officials frequently travel the country citing plastics as contrib uting to some of the deaths in prominent fires, even though there's no proof that their combustion products were responsible. Allied has also lobbied heavily against the inclusion in the National Electrical Code, which serves as a guide for local codes, of a new, flexible type of PVC conduit. The new conduit--that's a sample above--is even cheaper to install. Allied's efforts have prompted the Carlon division of New York-based Indian Head Inc.--in turn con trolled by Curagao-based Thyssen-Bomemisza N.V,--to sue it for restraint of trade. The Society of the Plastics Industry, which has long retained the public relations firm of Hill & Knowlton, has also redoubled its ef forts to present its side of the controversy. Carlon, with 1982 sales of $150 million, intro duced the flexible conduit in 1980; it's also the leading maker of rigid PVC conduit. Oth er makers include CertainTeed Corp. of Val ley Forge, Pennsylvania, and Robintech Inc. of Fort Worth, both publicly owned. Leading suppliers of PVC are B.F. Goodrich, Tenneco, and Du Pont. LLIED SAYS IT HAS "a moral and legal responsibility" to op pose products that constitute "an inherent, immediate, and sub stantial danger to the public." As serts Theodore H. Krengel, 57, Allied's founder, president, and controlling share holder "The plastics produced now kilL" Al lied's awareness that plastics are hazardous goes back to 1972, he says. MWe began to hear more and more about the problems of toxicity, flammability, and smoke of plas tics." Buildings, warns Krengel, are becom ing "plastic bombs--they go up in a matter of seconds." Allied's general counsel, John Uson, adds ominously: "Any company that knowingly puts a harmful product into the stream of commerce is liable for punitive damages." Allied is not the only campaigner against plastic pipe. Carol Bellamy. New York City's Research Associate Ford S. Worthy M FORTUNE FEBRUARY 7.1983 PHOTOGRAPH RV Gahr THOMPSON HSI PM lA IS P N III mam**itaiMMUittafii m itftiiifli URL 04140 ON PLASTIC PIPE {politically ambitious city council president, ^ made a fuss last year about new PVC conduit ; installed in subways. At her insistence $2 l(j million is being spent to rip out some and re place it with metal, though most of the sub* (f. way system's PVC will be left where it is. Plastics' growing use in plumbing, primarily y for water and drain pipes, has also come in for criticism. Allied's crusade against PVC g has been especially vehement, however. A significant proportion of its sales are threat3 ened. The last straw may have been Carlon's flexible conduit, aimed at a $190-miliion-ayear market that steel had all to itself. "I wouldn't deny we have a commercial in terest," says Allied's President Krengel "It's a big market. If you include cable, con duit, ducts, and pipe, you're talking in the multibillions of dollars." If the competition gets much worse, Krengel adds, Allied might switch to plastics--"ifwe could find one that is proven safe." Krengel goes on to declare: "I'm not going to make any Three Mile Islands or any Love Canals. I'm not going to make anything where I can't sleep nights be cause we've made a product that's unsafe. I don't want that on my conscience/' Krengel could be speaking his convictions, of course. But doubt is fanned by a curious fact recently discovered by Fortune: Allied itself has marketed PVC conduit on and off in recent years. In fact, it's still selling PVC- coated products. PVC is hazardous when it bums, as the plastics industry admits. So is everything else that bums. PVC may give off, among other things, hydrogen chloride, a corrosive gas that is lethal in high concentrations. But burning natural materials also give off a host of dangerous substances, such as carbon monoxide and acrolein. When wool carpeting or upholstery catches fire, it can produce deadly cyanide gas. Furthermore, a govern ment-sponsored study has shown that Doug las fir, widely used in construction, is every' bit as hazardous when it bums as PVC. Still, wouldn't it make sense to ban plas tics and use just noncombustible metal for conduit? The answer, according to experts not financed substantially by either side in this dispute, is a surprising no. There is much more to fire hazards than the toxicity of a material's combustion products. "You may be making the situation worse by having metal conduits," says Dr. Edward Radford of the University of Pittsburgh's Graduate School of Public Health. "Many fires result from electrical short circuits. One common type occurs as a result of improperly ground ed metal conduits." Adds Radford, a leading authority on what kills in fires: "There is no evidence that PVC plays a major role in whether an individual dies in a fire." Since no material is hazard free, the real issue is whether PVC is more hazardous than others. Here again the answer, accord ing to impartial authorities, seems to be no. According to Jack Snell, director of the Cen ter for Fire Research at the National Bureau of Standards, PVC conduit would probably be among the least of the worries in most fires. "Plastics would not create a significant additional hazard to life," says Snell. "Typi cally it's the contents of a building--in con trast to construction, plumbing, and electri cal materials--that represent the largest fuel load. You would need a large fire before the conduit became involved, and by that time the burning contents of the room would Neptune stood helpless osflames ravaged the MGM Grand Hotel in Las Vegas on Novem ber21, 1980, leaving85dead. Allied Tube 4 Conduit has cited this and otherfires to publi cise the supposedly extraordinary hazards ofthe lethalproducts given offwhen plasticpipe catchesfire. Thefinal report on thefire, however, did not Nameplasticsforany ofthe deaths. 5^.-- URL 04141 Caught with hit funding showing, Michael Olsen (right), then managing directorofthe AHiedsupported Foundation for Fire Safety, admitted in an interviewfor TVthat the bulk ofhis support camefrom the metal industry. At left on this April 1982 broadcast on Houston's KHOU-TVis re porter Roger Lmdberg, interviewing Olsen at thefoundation's headquarters. have caused the occupants to succumb." Snell's opinion is tersely endorsed by Ir win Benjamin, an eminent lire researcher and now a private consultant after many years at the Bureau of Standards. Says Ben jamin, "If the public has been terrified about plastic conduit, that's completely ridiculous." Surprisingly, these opinions are echoed even by Merritt Birky, director of research at the Allied-supported Foundation for Fire Safety. "Plastic conduits play little role in an ordinary hotel fire," Birky conceded in a re cent interview, "It is unlikely," he adds, "that plastic conduit played any role in, for example, the fire at the Westchase Hilton in Houston." He refers to a March 1982 fire that took 12 lives. A toxicologist with a Ph.D. in chemistry from the University of Virginia, Birky until January 1982 headed combustion toxicity studies at the National Bureau of Standards. He has generally been careful in his public utterances. Yet some of them, both before and after he changed jobs, have been used by Allied to give spurious scientific credibility to its campaign against PVC. One of Allied's chief exhibits is the No vember 1980 fire at the MGM Grand Hotel in Las Vegas, which killed 85 people and in jured hundreds of others. The cause of the fire, it should be noted, was a short circuit in the hotel's metal conduit. But burning PVC plumbing, Allied asserts, has been implicated in some of the deaths. The company has seized on a statement made by Birky while be was still with the Bureau of Standards and involved in the investigation of the blaze. The carbon monoxide levels in the victims' blood, Birky speculated at that time, were too low to have caused all the deaths. This implied that something else caused them, and Allied has repeatedly pointed to plastics ever since. In a September 1982 news release headlined "Plastics in Con struction Add to Fire's Tombstones," the company's public affairs vice president, Laurence Zoeller, declared: "There is mounting evidence that decomposing plas tics contributed to the vast majority of fatali ties in such recent tragic fires as at the MGM Grand Hotel." This release did not mention that in No vember 1981 the National Fire Protection Association, an investigative and code-writ ing body, had issued its final report on the fire. It said that 79 out of the 85 victims died primarily from a combination of smoke inha lation and carbon monoxide or smoke inhala tion only. Bums and smoke inhalation killed three more, while the remaining three each died of a different cause: burns, skull frac ture, and heart failure. As Birky emphasizes, the report's findings are vague. But they are a far cry from implicating plastics. Birky may nevertheless have done a bit of out-and-out propagandizing, though he vigor ously denies it. He had taken up his post at the Foundation for Fire Safety when he was interviewed by reporter Rolando Santos of KPIX-TV in San Francisco. Santos reported last March 23: "The leading researcher in the study of the deaths at the MGM, Dr. Merritt Birky, told me in these cases the tox ic gases had to come from the plastics in the room, probably cyanide." Birky claims that Santos misquoted him. Santos says that he and Birky had a detailed conversation prior to the broadcast, and that Birky said exactly what was reported. Whatever Birky may have said about cya nide, Allied brandished this scare word be fore the final report on the MGM fire was out. In one May 1981 advertisement in the trade press, the company said: "The Clark County Coroner reported that five victims, who were among the first autopsied, showed evidence of cyanide in their bodies. Cyanide is produced from burning plastic pipe fre quently used for drain, vent, and waste dis posal, as well as plastic that is commonly used for decorative wall coverings." Allied omitted to say that cyanide is also given off when other synthetic materials bum. In any event, the presence of cyanide is no proof that it killed anyone. The final investigation report on the MGM fire never mentioned synthetic materials--plastic pipe or whatev er--as the primary cause of any death. One of the Foundation for Fire Safety's of ficials has resorted to innuendo in discussing the Westchase Hilton fire in Houston. Birky, as noted, says it's unlikely that plastic con duit caused any deaths there. Yet Michael Ol sen, then the managing director of the foun dation, got on KHOU-TV in Houston a month after the fire, offering his "prelimi nary report" that "toxic gases in addition to carbon monoxide must be considered as causes of death." The TV reporter told his audience, "The Fire Safety Foundation be lieves some of those deadly gases can come from plastics." S A RULE, the foundation's spokesmen have not volunteered the source of their funds during such interviews. Unfortunately for Olsen, KHOU-TV's reporter Roger Undberg was one of the few to ask him where the money came from. Olsen ad mitted, listeners were told by another re porter, that "he was funded largely by the metal industry." Olsen insisted that his re port was unbiased. KHOU-TV then present ed Dale Everitt of the Houston Fire Depart ment, who dismissed Olsen's views. Said Everitt, "When you have a group like this, 1 think they're going to be interested in keep ing those funds coming in." Olsen has since 70 POST1 tkif renoi iadv 7 ioo7 left the foundation. Conveniently for Allied, some professional firefighters have strong opinions against plastics and will state them publicly. Shortly after the Westchase Hilton fire. Andrew Casper, then chief of the San Francisco Fire Department, appeared on KPIX-TV. Casper sketched a simple causal sequence in the MGM Grand fire: "More plastics, more fires, more deaths caused by the inhalation of toxic fumes." Later he went on the payroll of the Foundation for Fire Safety, which he has since left. URL 0414 HILE AN EX-FIRE CHIEF has credentials to speak out. Wthis can hardly be said for another PVC opponent fi nanced by Allied. Deborah Wallace is president of the Public Interest Scientific Consulting Service, another orga nization mainly backed by the company, to the tune of $17,000 a quarter. With a Ph.D. in environmental physiology, Wallace has little expertise in deaths from fires. Yet she trav els the country preaching the dangers of PVC and calling herself "an expert on fire toxicology." She also made a pretrial deposition in 1980 as an "expert witness" in the litigation aris ing from a May 1977 fire at the Beverly Hills Supper Club in Southgate, Kentucky, in which 165 died. Later she testified that au topsy reports, hospital admission records of H you d kte to urtoa nvxe about Jack Dantf's Whiskey, drop us a bM. survivors, and questionnaires strongly sug gested that PVC was responsible for the deaths and injuries at the fire. On this point she was disputed during the trial by the med THESE MEN can cell exactly what's happening inside every barrel in a Jack Daniels warehouse. ical examiner of St. Louis County. The Foundation for Fire Safety has lob bied repeatedly. Several states have been considering bills that would require that all In the heat of summer the whiskey is expanding into the charred inner wood of the barrel. Come materials be tested for their combustion tox icity before they can be considered for ap Halloween, it s starting to cool. And inching proval in building codes. Since metal conduit can't bum, it would pass such tests; PVC might not. Unfortunately, satisfactory test ing procedures don't yet exist. Yet on May 6, 1982, Birky appeared before the fire safety subcommittee of the New York State senate finance committee and declared that they do. An inhalation test protocol that he helped de velop at the National Bureau of Standards, he said, is able to prove that one material is more toxic than another. One reputable authority disputes Birky. "There's no correlation between the test method and what happens in a real fire," says its way back toward the center. Over the aging period, this gende drculadon of whiskey is going on constantly. Of course, it can't be perceived by the human eye. But after a sip of Jack Daniels, we believe you 11 recognize CHARCOAL MELLOWED 6 DROP 6 BY DROP Irving Einhom, an adjunct professor at the University of Utah, who has published 100 its importance. papers on combustion toxicology. "Yet Birky is going around the country trying to ban ma terials based on incomplete tests." The Bu Tennessee Whiskey 90 Proof Distilled md Bottled by Jack Daniel Distillery reau of Standards adds that even if its test Lem Motto* Prop., Inc., Lynchburg (Pop. 361). Tennessee 37352 Placed in the National Register of Historic Places by the United States Government FEBRUARY 7,1983 FORTUNE 71 i Ml M IX Reserve space now in the new Fortune Spring Gift Guide for Executives--a special adver tising section in the issue of May 15,1963, delivered May 2. Closing February 9. Ybur gift suggestions wilt go straight to the heart of the busi ness and personal gift-buying market--the affluent readers of Fortune Worldwide. Just when they're looking for ideas for wed* dings. Mother's Day, Father's Day, birthdays, anniversaries, all summer's gala occasions. What's more, you'll get-- Exclusivity in your category among four-to-a-page partici pants. Credit toward space and fre quency discounts in Fortune. Free art and production. Your low $10,000 space cost includes full-color photography, type, everything except your own copy up to 30 words. A $2,000 savings if you combine your Spring Guide order with an insertion in the Holiday Gift Guide scheduled for Fortune's issue of November 28,1983. Open up a whole new hotline to summer sales with Fortune's Spring Gift Guide. But don't delay--it closes February 9! Call your Fortune representative today. Or get in touch with: Bruce A. McNaughton Associate Consumer Advertising Manager Fortune Time & Life Building Rockefeller Center New York 10020 212/307-4810 74 FORTUNE FEBRUARY 7. 1983 protocol were ready to be used, which it fire hazards it professes to be concerned isn't, it alone would not justify banning some about. According to Lison, Krengel got out of materials from building codes. the business because this time he really The plastics industry and Carlon have found the material hazardous. "Ted was charged Allied with using questionable tac amazed by the horrors of the MGM fire, and tics within the National Fire Protection Asso then the others," says Lison. ciation. Prior to an NFPA vote in May 1980 Allied's suppliers remember a completely that would have recommended the inclusion different story. When Allied concluded its re of Carlon's flexible PVC conduit in the Na lationship with Robimech, Pat Madormo, tional Electrical Code, they say, Allied and Robintech's executive vice president, met other companies and trade organizations in Bemie Auerbach, Allied's product develop the steel industry purchased scores of (and ment manager, and asked why Allied was perhaps as many as 100) $50 memberships calling it quits. According to Madormo. Auer in the organization. When the vote came up, bach told him that Allied felt that PVC was a these new members helped to defeat Car- different market from those it was used to. Ion's conduit by 394 to 390. "The PVC market got pretty tough around then." says Madormo, "and there wasn't EVERAL NFPA MEMBERS were enough margin for them to buy from Robin- outraged by the way the outcome tech and resell." Swas achieved. One, Nathaniel Add- Allied returned to the PVC business for a ieman. a registered fire protection third time in November 1981. It acquired El- engineer with Boeing Co., was par cen Metal Products Co. of Franklin Park. Illi ticularly incensed when he addressendoist,hoene of whose product lines is PVC coat group after the vote: "I had occasion teodh.avAe current product, called Strut, is a lunch with a gentleman who is a salesman for hanging PVC-coated pipe holder. If the big one of the steel companies and he didn't hotel fires convinced Krengel he should get know why he was here. He was going to go out of the business, why is Allied selling to a meeting at 12:30 to find out why he was something like this? "Look," says Lison, "we here and be told how to vote. This is what's don't know why people die from fires in this happening here today. We cannot let the country." But wouldn't PVC be dangerous in NFPA degenerate to that type of thing." this sort of application? "We don't have a Before and after Allied helped vote Car- corner on the wisdom of the world," Lison Ion's conduit out of the electrical code, it was says, "and Factory Mutual (a testing organi selling PVC products. Even though Allied zation that does inspections for the insur President Krengel claims that he was aware ance industry] kept telling us that this stuff is of plastics' hazards as far back as 1972, his fine, that it'll be included in all the codes. You company sold PVC conduit during the mid- know, for lots of applications PVC is better Seventies through distributors. Asked how than our stuff." he could sell a product that he considers haz Despite this strange admission, Allied's ardous, Krengel says, "Maybe we began to war on PVC has continued without letup. realize how bad it was later. Anyway, as soon Last September the newsletter of the Ameri as we began to realize how bad it was, we got can Council on Science and Health, a con out of the business." sumer education group, published an article Allied's general counsel, John Uson, says headlined "The Merchandising of Fear" on Allied got out of the PVC business for a dif some less than savory aspects of AUied's ferent reason: it lost its supplier. He adds, campaign. Zoeller's quick reply was morally "Our suppliers could not keep us price com lofty: "Although Allied obviously has an in petitive. There wasn't enough profit margin terest in this issue, we had every opportuni in 1975." Allied got back into the PVC mar ty to enter the PVC market and chose not to ket during the summer of 1980. Krengel and on safety grounds. We believe this issue con other Allied officers worked out a deal in cerns human life and safety, and therefore which Robimech would produce rigid plastic transcends the commercial interest of any conduit for Allied on a private label basis. company or industry. We believe our record Krengel now says be wasn't aware that Al shows this concern." lied was in the PVC business a second time. Asked if subordinates had put Allied back Allied's record is something rather differ into PVC without his knowing about it, he re ent--a rare well-documented example of un sponds: "Look, it was a nothing kind of thing. fair tactics in the marketplace. Companies It was incidental and momentary. When we battling new products often play rough, with knew what was going on we backed off." high-pressure salesmanship, aggressive pric Actually Allied sold PVC conduit from ing, antagonistic ads. But a few go further, in about September 1980 to March 1981. It did ways that sidestep notice. Allied has called so, moreover, without warning users of the attention to itself by overreaching. D