Document 85pG0YB9ej5pO1J35v70orr4a

NOTICE This material may be REVIEWS protected by copyright---------------------------------------------------------------------- ------ law (Title 17 U.S. Code) Occupational Carcinogenesis: The Louisville Experience with Vinyl Chloride-Associated Hepatic Angiosarcoma CHARLES L. DANNAHER, M.D.* CARLO H. TAMBURRO, M.D. LUNG T. YAM, M.D. Louisville, Kentucky From the Divisions of Hematology-Oncology, and Digestive Diseases and Nutrition, Depart ment of Medicine and the Vinyl Chloride Project, Cancer Center, University of Louisville; and the Veterans Administration Medical Cen ter, Louisville, Kentucky. Portions of this work were supported by the National Cancer Institute Contract NO l-CN-55212 and by a Veterans Administration Medical Research Grant Re quests for reprints should be addressed to Dr. Charles L. Dannaher. Manuscript accepted Au gust 21,1980. `Present address; Columbus Hospital Re gional Oncology Center, P.O. Box 5018. Great Falls, MT 59403. Hepatic angiosarcoma in man was first associated with exposure to vinyl chloride in Louisville, Kentucky, where it was identified in 10 persons from a single vinyl chloride polymerization plant; clinical manifestations are summarized herein. Following prolonged expo sure to vinyl chloride, the onset of this disease is insidious and the clinical picture is that of nonspecific hepatic injury with mildly ab normal biochemical liver test results. Carcinoembryonic antigen and alpha fetoprotein are undetectable. Radionuclide and angiographic studies of liver show characteristic but nondiagnostic abnormalities. A definite diagnosis is usually made only by open liver biopsy. Treatment is unsatisfactory but chemotherapy seems to prolong survival. Average survival from diagnosis is about 12 months. Overt liver failure usually occurs only as a preterminal event and was the major cause of death in all of our patients. Preventive measures are now in effect in the plant. This experience illustrates the importance of the clinician in occupationally-related caneer. Most cancers in man result from exposure to environmental factors. The occupational environment may account for as many as 10 percent of these malignancies [1]. We present the course of events and clinical findings of an industrial epidemic of hepatic angiosarcoma, a rare tumor considered to originate from the vascular lining cells of the liver [2,3]. This experience demonstrates the importance of the clinician in recognizing and controlling occupationally-induced cancer. In 1974, several patients in Louisville, Kentucky, were found to have angiosarcoma of the liver [4], All of these patients were men who were chemical workers at a local industrial plant that polymerizes vinyl chloride. Subsequently, additional cases in which the patients had a history of prolonged exposure to vinyl chloride were reported in Louisville [5,6] and elsewhere [7-11], Various aspects of this malig nancy including histology [11], angiographic and radionuclide char acteristics [12], urine [13] and tissue [14] glycosaminoglycan patterns and epidemiologic studies [15] have been reported from Louisville. In addition, interest in this disease has propagated elsewhere, both nationally and internationally [16-18]. However, the total experience of vinyl chloride-associated angiosarcoma in the Louisville area has never been told. With a cohesive summary one begins to appreciate how the initial clinical observation of a patient with hepatic angio sarcoma led to the realization of a possible etiologic association. This, in turn, has cast a profound influence on the health of those working in a major worldwide industry involving the production of polyvinyl chloride. j"! iijlij ^s ** : .t > Mil' m i-MW 't Si | ii i u4 <yfi it Jt: I'r'.fclP m --if m vtrt jt ?=.< 14 mLj.) M u it v.uii-w ii.: /1 ; February 1981 The American Journal of Medicine Volume 70 27S AP00024508 OCCUPATIONAL CARClNOC.ENKSIS-tlANNAHLR F.T AL. This concentration of cases of hepatic angiosarcoma gave us the opportunity to study the clinical manifesta tions of this unusual cancer, In view of concern re J,garding its increasing incidence (7.19-21 we believed a summary of this information would be valuable in detecting and managing future patients. In this paper we present the natural course, diagnosis and treatment of this disease as well as possible preventive measures that we have learned from study of the patients dis covered in the Louisville area. Characteristic features of the disease, including a long period of asymptomatic laboratory abnormalities, the difficulty in diagnosis and the poor response to treatment, are demonstrated by the following illustrative case. CASE REPORT The patient, a 51 year old man who entered the vinyl chloride Industry 28 years ago as a vat cleaner, worked in the vinyl chloride polymerization area for 18 years. During medicnl screening of plant employees in April 1974, he was found to have mildly abnormal serum biochemical liver test results. Liver-spleen scan disclosed no abnormalities, and he was asymptomatic. Because of persistently abnormal biochemical liver tests, he was evaluated further in December 1974. He remained asymptomatic, and findings on physical examination were within normal limits. A hepatic angiogram showed multiple distended sinusoids throughout the liver with pooling of con trast material within those lesions. Transjugular liver biopsy specimens were histologically interpreted as showing chronic inflammatory reaction and fibrosis, anributable to the patient's previously known liver involvement with malaria. He was removed from the vinyl chloride processing area and contin ued active employment without symptoms. His liver profile continued to deteriorate; by September 1975, he began to complain of tiredness and easy fatigability. The findings on physical examination remained within normal limits. Liver scan now demonstrated hepatomegaly with multiple defects throughout the liver. Biopsy specimens demonstrated sinusoidal dilatation and Kupffercell prolifer ation in some of the specimens, but the diagnosis of angiosar coma could not be made. The patient did not return to work. His fatigue increased over the next two months and vague abdominal pain devel oped. In january 1976, gross hepatosplenomegaly with a bruit over the liver and signs of high output cardiac decompensation were detected. Liver-spleen scan and angiogram showed further enlargement of the previously identified lesions. An increase in arterial-venous shunting was demonstrated. At abdominal operation, direct examination revealed that the veins of the portal system were somewhat distended and that the surface of the liver was slightly nodular with some areas of bluish discoloration which were somewhat serpentine in nature. There were multiple areas on the liver in which the tissue was slightly elevated with a central depressed area of a somewhat cyanotic appearance. A strong thrill could be felt when palpating the liver. The hepatic arteries were ligated, and there was complete disappearance of the thrill. Wedge biopsy of the liver demonstrated areas of angiosar- coma among arcus of fibrosis anti sinusoidal dilnlntion uiih hypertrophy and hyperplasia of sinusoidal lining cells. Postoperntively, the cardiac decompensation was corrected and the spleen size was reduced to normal, bul the liver size remained the same. Chemotherapy with doxorubicin, cyclophosphamide and methotrexate was started, and the liver vvas reduced to its normal size but the biochemical liver test results remained abnormal, The reduction in liver size lasted fonr months during which the patient was unable to carry on nor. mal activity without fatigue. Over the next four months there was a gradual increase in the size of the liver with the devel. opmont of ascites and abdominal pain with deteriorating biochemical liver tost results. The patient died in December 1976 with hepatic encephalopathy and gastrointestinal bleeding. Autopsy revealed an enlarged liver with multiple small nodules of angiosarcoma throughout. The tumor was metastatic to the spleen and lungs. BACKGROUND AND METHOD OF STUDY Vinyl chloride is the raw materia! with which the common plastic polyvinyl chloride is made. Vinyl chloride monomer (CHjCHCi) is a gas mode by the reaction of ethylene and chlorine, which, when processed commercially, is polymerized In large vats under increased temperature and pressure to form a white solid. Workers are exposed to gaseous vinyl chloride"- ' by leakage from the manufacturing vats during processing, upon opening the vats after the processing is complete, or, most Intensely, in the cleaning process when the worker enters the vat and chips residue off the interior wall releasing pockets of gas. Vinyl chloride was initially thought to be completely harmless, and there was no limit to the exposure workers re ceived in the 1940s and early 1950s, however, a voluntary re striction of 50 parts per million was eventually adapted. The gas was considered to bo so harmless that it was oven used as a propellant in aerosol sprays [22] and a general anesthetic [23]. It gradually became apparent that prolonged exposure could induce injury in many tissues including skin, blood vasselsartd the reticuloendothelial system [24-26J. Evidence of liver injury. was soon found [27,28], which was associated histologically with changes in the hepalocytes. fibrosis and sinusoidal dila tation [29,30]. In animal experiments, prolonged inhalation of vinyl chloride was shown to induce cancer in rats [31]. Soon thereafter an employee of the Louisville vinyl chloride j polymerization plant died of hepatic angiosarcoma, but the j significance of this was not appreciated until nearly two yean . , later when a second employee also died of this rare turnon^ J Since these workers were cared for by different physicians at | different hospitals, this might still have gone unnoticed were it not for the plant physician, whose system of requesting re-, ports of illnesses and deaths of employees brought the diag noses to his attention. His memory of the earlier patient with " the disease led him to initiate a search for othercases among" former employees and to institute a medical surveillanca program of all present employees which was later supported , by the National Cancer Institute and the University of Louisville. This program involved testing nearly 1,200 employees with a standard automated biochemical profile and radionu clide liver-spleen scan. Workers with persistent biochemical liver abnormalities or with abnormalities of their radionuclide scan were then examined by selective hepatic arteriography, 1 280 February 1981 The American Journal of Medicine iiUiUlVV! IUN.U. t..\KUNlKil,\l;SlS-- DANNAHMK IT AL. hepatic venography and iransjugulnr hepatic biopsy 120.32,33], Patients with hepatic angiosarcoma seen by the staff of the Divisions of Hematology-Oncology and Digestive Diseases and Nutrition, University of Louisville, from January 1,1974, to January 3, 1977, include those accrued through the Vinyl Chloride Project as well as patients referred by physicians in the Louisville area and slscwhcre, and a few patients found in a search of autopsy records and death certificates. A total of 11 patients have been studied. 10 of whom have had heavy industrial exposure to vinyl chloride. The other patient was a woman, referred from Florida, who had a history of pro longed use of hair spray containing vinyl chloride as a pro pellant. Only the 10 patients with occupational exposure to vinyl chloride are Included in this report. In addition, data are derived from employee records of the Louisville vinyl chloride polymerization plant during its 35 year operation to the date of this report. RESULTS Epidemiology. The cohort for this analysis are the employees of the Louisville plant from its opening in 1942 to the end of 1976. The employee population has been extremely stable and during this 35 year period a total of 1,855 persons have been employed. Of these, 1,184 are current employees and 671 are past em ployees. During the 35 year period, 167 members of the cohort died. Sixty-six of these died of heart disease and stroke, 39 of cancer, 19 of accidents and, nine of other disease; in 34 the cause of death could not be determined. Pri mary brain cancer has been related to vinyl chloride exposure [34,35], and five of the 39 deaths from cancer were from this cause. The mortality rate in our popula tion adjusted for age is approximately twice that ex pected [36], Ten cases of hepatic angiosarcoma have been iden tified among our 1,855 employees during the 35 year period. All of these employees worked in areas of the plant with high vinyl chloride exposure. We divided these cases into a retrospective group--cases discovered by review of hospital and pathology records, and a surveillance group--cases discovered prospectively in the surveillance program. In the retrospective group, one case each was diagnosed in 1964,1967,1970,1972, 1973 and 1974. In the surveillance group, three cases were identified in 1974 and one in 1976. No other cases, of angiosarcoma have been identified in the Louisville area. The incidence over the 35 year period in our cohort is 0.54 percent, equivalent to 15.4 cases per 100,000 population peryear. (Predicted annual incidence in the U.S. in 0.0014 cases per 100,000 population [37].] The mean age at the time of diagnosis was 47.6 years, with a range of 36 to 58 years. All patients were white men, reflecting employment patterns at the plant The mean duration of exposure to vinyl chloride was 17.4 years, with a range of 12 to 28 years. The mean duration from first exposure to vinyl chloride to the time of diagnosis of hepatic angiosarcoma was 19.9 years, with a range TABLE I Early Clinical Features Related to Hepatic Angiosarcoma In 10 Patients PlllenU Patients Symptoms (no.) Signs (no.) Pain Fatigue Weakness Weight Loss Anorexia Indigestion Abdominal fullness Melens 5 Hepatomegaly S Splenomegaly 4 Abdominal tenderness 3 Hepatic bruit 3 1 1 1 --------------------1------------- 7 5 3 1 from 12 to 28 years. No differences were noted between the retrospective group and the surveillance group. Three patients had known heavy exposure to alcohol, and none of the patients had previously had viral hep atitis or exposure to any other known hepatotDxins. Clinical Features. The earliest clinical features which could be related to the angiosarcoma are presented in Table L Pain was localized in the right upper abdominal quadrant or in the lower right hemithorax. In two of the patients the pain was acute, but in the others it ranged from two to six months in duration. Fatigue and weak ness had been present from one month to one year. Weight loss was significant, averaging 25 pounds over four and a half months. Hepatomegaly was the most common physical sign. The liver measured from 15 to 24 cm in total vertical span at the right mid-clavicular line by percussion. When splenomegaly was present the spleen was easily palpable 3 to 7 cm below the costal margin. Two patients, both in the surveillance group, had no physical signs, and one of these patients was asymptomatic. Laboratory Findings. Hematologic evaluation at di agnosis showed three patients with normal complete blood counts. Mild anemia was found in five patients, but reticulocyte counts were not determined. Two of the patients had abnormal erythrocyte morphology with target cells and schistocytes. In one patient the anemia was combined with leukocytosis and in three patients with thrombocytopenia. Each patient with thrombocy topenia had splenomegaly. Two other patients had leukocytosis as their only hematologic abnormality. In two patients the bone marrow was examined: one was normal; the other demonstrated erythroid hyper plasia. No patient was without some abnormality in at least one of the standard biochemical liver tests. Most of the values were only minimally abnormal initially. There was no consistent abnormality of any single test result or apparent pattern of groups of tests, although the al kaline phosphatase level was most frequently elevated. Hepatitis B surface antigen and alpha fetoprotein were not found in four patients who were tested, and carcinoembryonic antigen was not present in two pa tients. I'y ; `Pt: iv VM -Wf r rr February 1981 The American Journal of Medicine Volume 70 281 AP0002451O K; i P*:K, -i!r-.ficiiifPii' tit--4* iii. J.. :.-nki .t* i 'fvi tr : $ [( r. II". 3 ' >.r !. ; .rri OCCUPATIONAL CAKUS'tKIUNIiSIS-DANNAHHK I1T Al. T' ts v Figure 1. Liver scan in hepatic angiosarcoma. A, anteroposterior view. Large peripheral defect in the right lobe with other smaller defects. B, right lateral view. - Results of routine roentgenologic studies and gas trointestinal contrast studies were normal in most in stances. However, special roentgenologic studies re vealed many abnormalities. Radionuclide liver scan disclosed abnormalities in all patients; however, the abnormalities were not consistent. Six patients had hepatomegaly. In eight patients a mass was demon strated which usally appeared as a single negative defect greater than 4 cm in diameter in the periphery of the liver. In four patients this was accompanied by one or more additional defects, and two patients had only ev idence of diffuse hepatocellular disease. Typical ab normalities on liver scan are demonstrated in Figure 1. Splenomegaly was demonstrated in five patients by radionuclide scan. Hepatic arteriograms were obtained in six patients and demonstrated enlarged sinusoidal spaces in each of them, with a mass demonstrated in five patients. The masses were supplied by normal-sized hepatic arteries and were characterized by persistent peripheral tumor stains with some degree of central hypovascularity as shown in Figure 2. Hepatic ultrasound was performed in only one patient and demon!' strated a solid mass. Clinical Diagnosis. Methods: Although the signs and symptoms of our patients were fairly nonspecific, there was some finding in almost every patient that would direct one's attention to the liver in an effort to explain the clinical features. Biochemical tests of blood usually revealed nonspecific hepatic injury, but no pattern of abnormality could distinguish angiosarcoma. 'Nek Figure 2. Hepatic arteriogram in hepatic angiosarcoma. A, arterial phase demonstrating dilated sinusoids In both lobes. B,' venous phase demonstrating a large mass in the right lobe with peripheral staining and central radiolucant area. 282 February 1981 The American Journal of Medicine Volume 70 AP00024511 UlA.l't A i iua.aI. L.AitLIi'.wul.M.MV il.\\\Al!l-.R LIT AL. Figure 3. Gross section of an angiosarcomatous liver displaying multifocal solid and hemorrhagic cystic growth pattern with a large ruptured cyst. The radionuclide liver scan suggested involvement of the liver by a malignant process because of the neg ative defect, however, the peripheral location could be mistaken for a rib impression, porta hepatis or other normal anatomic variation. The presence of multiple defects could give the impression of a malignancy metastatic to the liver. Although the liver scan frequently indicated malignancy in the liver, it was not specifically diagnostic of angiosarcoma. The abnormal vascular pattern demonstrated by the hepatic arteriogram with the persistent peripheral tumor stain and central radiolucent area is strongly suggestive of hepatic angiosarcoma. Under unusual circumstances this appearance could mimic benign vascular lesions, liver cell adenomas or hepatic infarction. However, it is the most accurate clinical means of diagnosing hepatic angiosarcoma and should be obtained in all patients In whom the diagnosis is suspected. Further information regarding special roentgenologic studies has been de tailed in the literature [12], Extent of disease: At diagnosis the extent of disease was determined using clinical parameters and infor mation obtained at exploratory laparotomy. In four of the patients disease was confined to one lobe of the liver and in six patients the disease involved more than one lohe of the liver. Direct extension involving the dia phragm was present in two patients and the stomach in one patient. Porta hepatis lymph nodes were involved in one patient. One patient had distant spread to the lung. Pathology. Gross description: The liver is usually enlarged with blunt edges and a variegated brownred-orange color. The surface may contain ona or more purplc-bluo nodules with spongy consistency varying in size from 1 to 4 cm. In two cases these ruptured and caused hemoporitonoum. The cut surface is diffusely involved with cystic areas filled with dark blood with occasional cysts measuring up to 20 cm in diameter (Figure 3). In one case the tumor had a primarily solid appearance. Definitive diagnosis: Tissue for histologic examination was obtained by percutaneous biopsy, transjugular bi opsy, open liver biopsy and autopsy. Although six tissue specimens obtained by percutaneous and transjugular biopsy did identify other liver abnormalities, specimens were insufficient for accurate diagnosis of malignancy. All definitive diagnoses of angiosarcoma in this group of patients were made by tissue obtained at open liver biopsy (seven patients) and at atuopsy (three patients). In our experience, in a patient with biochemical evi dence of hepatic injury with a history of prolonged ex posure to vinyl chloride, open liver biopsy under direct vision is needed for diagnosis. We have no experience with laparoscopy and directed biopsy, which may be associated with less morbidity, but this approach should be tempered with caution in such a vascular tumor with major potential for hemorrhage. Histology: All tumors were primary angiosarcoma of the liver with a wide range of patterns and differentia tions seen between different tumors and within the same tumor (Figures 4 and 5). Areas of extensive hem orrhagic necrosis and areas of enlarged spindle-shaped endothelial cells lining irregular vascular channels appearing as a spongy network of sinusoids were present. Tumor cells occasionally projected into these channels. Large cavities lined by sarcomatous cells contained fresh or altered blood and, at times, organized clot. Three patients had histologic evidence of cirrhosis, two of whom had histories of heavy alcohol intake. One additional patient had fibrosis, and two patients had evidence of toxic hepatitis. The variable histology of angiosarcoma and the difficulties of diagnpsis have been described previously [3,11,38]. Metastatic spread: Nine of the 10 patients have died, and eight of them have had postmortem examination. The single patient without a postmortem had an ex tensive clinical evaluation and exploratory laparotomy two weeks prior to his death. His data have been com bined with the autopsy data regarding extent of disease at death. One patient with extensive hepatic fibrosis considered to be secondary to radiotherapy had mi croscopic angiosarcoma in the minimal remaining liver tissue with no other evidence of disease. Eight patients had extensive disease throughout the liver. Five of these eight patients had regional dissemination with in volvement of diaphragm (three of five], lymph nodes (two of five), abdominal wall (two of five], gallbladder (one of five) and bowel (one of five). Four of these eight patients had distant spread with involvement of lung (four of four], bowel (two of four), adrenals (two of four), lymph nodes (one of four), brain [one of four), bone (one i.. t . ! if *. sVi. I l. h K- v:.-t t ;i.u February 1981 The American Journal of Medicine Volume 70 283 . A wsmammmmmmatmm AP00024512 OCCUPATIONAL C A H CINDCKN KSKS -- 1)A N N'A HICK KT AL. 1 } i."' *. .- ; t. t F. :! i Figure 4. Undifferentiated angiosarcoma with tumor giant cells replacing the usual histologic pattern of liver tissue. Hematoxy/in and eosin stain; magnification X 500, reduced 50 percent. Figure 5. Multifocal minute angiosarcomatous growths ar* conspicuous among liver cell cords in locations distant frorp grossly visable tumors. Hematoxylin and eosin stain; mao. nification X 500, reduced 50 percent. of four), spleen (one of four), pleura (one of four), peri cardium (one of four), myocardium (one of four) and kidney (one of four). Ore patient had distant metastases without evidence of regional spread, and two patients had no evidence of regional or distant spread. Therapy. In no patient was a surgical attempt at re section made. In those patients with disease confined to one lobe of the liver, fibrosis or cirrhosis in the re mainder of the liver was thought to contraindicate sur gery. One patient underwent hepatic artery ligation with no benefit. Radiotherapy to the liver was given to one patient early in his course with minimal tumor found at autopsy one year later, but extensive hepatic fibrosis led lo his death. However, another patient, following tumor recurrence during chemotherapy, received radiotherapy seven months after diagnosis. He was found to have extensive angiosarcoma in the liver-at autopsy two months later, with no apparent response to the radio therapy. Both patients received 5,000 rads in five weeks TABLE II Major Clinical Events Occurring Prior to Terminal Events During the Course of Angiosarcoma Seen in 10 Patients PeBentS ______________ Ereat__________________________(no.) Pneumonia Pleural effusion Staphylococcal septicemia Peripheral platelet destruction Microangiopathic hemolytic anemia Congestive heart failure Renal failure Hypertension Gastrointestinal bleeding Hepatic failure Hemoperltoneum Fever 1 1 1 2 1 3 1 1 1 2 2 1 to a major volume of the liver. Seven patients who re. ceived chemotherapy had some evidence of response by reduction in liver size or tumor size, and one patient continues to respond to chemotherapy 56 months fronr the time of diagnosis. Various drugs have been used individually and in combination including fluorouracil, vincristine, cyclophosphamide, doxorubicin and methotrexate. Clinical Course. Major clinical events during the course of each patient's disease are listed in Tahje IL Congestive heart failure was preceded by a period of high output cardiac volume overload in patients with demonstrated arterial-venous shunts. Microangiopathic hemolytic anemia and platelet destruction in the p*. ripheral blood were present without evidence of dis seminated intravascular coagulopathy or sepsis, and were possibly due to cell damage in the abnormal tumor vasculature (39-42). Symptoms of overt hepatic failure were unusual until very late in the clinical course. The factors contributing to death in these patients are listed in Table III. The major cause of death was the sudden onset of irreversible rapidly progressive hepatic failure. Survival. One patient remains alive 56 months from the time of diagnosis. Survival of the other nine patientsfrom the time of diagnosis ranged from four months to TABLE lit Contributing Causes of Death In Nina Patients with Angiosarcoma Cause Hepatic failure Gastrointestinal bleeding Ruptured tumor with hemoperitdneum Cardiac tamponade Hemothorax Ptl'unit (m.) 9 2 2 .:ti. 1 1 284 February 1981 The American Journal of Medicine Volume 70 AP00024513 (ICCUIWI IU\ \l. It.WNAliKK l.l' Al,, 29 months, with n mcnn survival of 11.7 months. When [7,19-31] have not coinc lo pass. The absence of new measured from the time of first symptoms, the range of ensos, however, should not bo looked on as the end of survival was five months to 34 months, with a moan the story. The Louisville plant is one of tho oldest vinyl duration of 14.9 months. Survival from the lime of di chloride polymerization facilities in the nation, and agnosis was distinctly longer in the six patients who those workers have had the longest potential exposure received chemotherapy (mean 15 months) than in the to the carcinogen. In addition, because of the early lack three patients who did not (mean five months). of concern regarding exposure levels, they also have had Preventive Measures. Shortly after the discovery of tho highest exposure. We have seen that the incubation an epidemic of angiosarcoma in the Louisville vinyl period may range from 12 to 28 years, and it is difficult chloride polymerization plant, new limitations of em to imagine that the recently instituted preventive mea ployee exposure to vinyl chloride were imposed at 1 sures have resulted in the absence of now cases. More ppm. Adherence by industry involved tighter seals on likely, this has resulted from the voluntary reduction in the polymerization vats and conduits which lesson exposure levels or some other alteration in the work leakage into the ambient atmosphere. The vat cleaning environment in the past. Moreover, we know that the process is now performed by workers wearing protec level to which vinyl chloride exposure was voluntarily tive respirators. A record keeping system has been ini reduced is still carcinogenic [45]. We have probably tiated for all workers which relates length of employ seen a dose-response offcct with a large number of cases ment and potential chemical exposure levels in each occurring after exposure to a large dose. A smaller work environment to possible harm from vinyl chloride number of cases resulting from lower exposure levels and other chemicals, so that a worker can be removed will probably continue to develop in the ensuing years re from exposure before he is likely to sustain injury. Since because of the long incubation period. This problem rise this process involved a change in behavior for ail em may become more pressing in other parts of the country rent ployees, and some jobs were more cumbersome and in which workers in vinyl chloride polymerization om lengthy, an education program was designed and im plants, opened in the 1950s and early 1960s, are now `sed plemented in an effort to relieve anxiety related to work nearing the mean incubation period for this disease. acD, in a potentially carcinogenic environment and to pro Mandatory restriction of atmospheric vinyl chloride and mote understanding of good health maintenance. The levels to 1 ppm has only been in effect for five years, and program resulted in better cooperation with the sur it is not known whether lowering exposure levels re the veillance and prevention programs. No new cases of verses the possibility of cancer in those already exposed sIL angiosarcoma have been discovered in the Louisville to higher levels. This must continue to be investigated. ^of plant since 1976 among 1,184 current employees and 504 In the group of workers with high exposure to vinyl h past employees followed at regular intervals for four chloride, close surveillance should be continued in Jiic years. definitely. Although the Louisville experience has re pe rils- COMMENTS vealed the close association between vinyl chloride exposure and hepatic angiosarcoma, the mechanism by nd which vinyl chloride triggers carcinogenesis remains to -ior The Louisville hepatic angiosarcoma experience un be studied through animal model experiments [46]. It ire folds in many ways as a medical detective story. First, is intriguing that angiosarcoma is so frequently associ a fatal incident occurred. A diligent search for clues ated with environmental factors, such as thorium are revealed vinyl chloride ns a potential suspect. Further dioxide [47-49], arsenic [50-52] and vinyl chloride, each :he investigation and intimate cooperation among all those of which results in stimulation of the hepatic sinusoidal itic concerned quickly gathered sufficient evidence to in lining cells [38], Other associated factors include criminate vinyl chloride as the etiologic factor for he hemochromatosis [53-55], copper [56] and estrogens om patic angiosarcoma. The evidence included the occur [57], However, some cases discovered during early nts rence of angiosarcoma in Louisville only in workers childhood are apparently congenital [58]. When faced to exposed to high concentrations of vinyl chloride over with a patient with angiosarcoma without one of these a long period of time, the discovery of similar patients recognized factors, a diligent search for other possible in other plants producing vinyl chloride [8,9], and the environmental factors should be pursued so that pre production of similar hepatic lesions in experimental ventive measures can be undertaken. animals exposed to high concentrations of vinyl chloride Many new chemicals are introduced into our work [43,44]. The detective was an observant clinician who and home environments each year. There is increasing recognized the initial incident, realized its potential concern that environmental factors are responsible for importance, instituted a systematic investigation and many cancers in man. Although sophisticated laboratory initiated a cancer control effort. This entire experience techniques and specialized knowledge are essential in demonstrates the importance of the clinician in con cancer research, the Louisville experience in vinyl trolling occupationally-induced cancer. The early fears of progressively increasing numbers chloride-associated hepatic angiosarcoma demonstrates how a careful clinical observation followed by a well of angiosarcoma cases related to vinyl chloride exposure conceived clinical approach may reveal an environ- February 19S1 The American journal of Medicine Volume 70 285 : vu. FLihlW i ): ;h;A iff'??' : it AP00024514 IKI wm ' !-l : !> I .i occupational CAitciNoiUTNKSis--damnaiIKK et al. mental etiology of neoplasia and result in profound benefit for all of us. ACKNOWLEDGMENT We acknowledge the encouragement and assistance of Laszlo Makk, M.D., John Creech, Jr., M.D., Joseph G. Whelan, Jr., M.D. and Barbara Miller, B.S.N., nnd th cooperation of the B. F. Goodrich Chemical Company Bells Lane, Louisville, Kentucky. The help of Mark Howze, Betty Bailey, Betsi Zimmerman, Lou Wilmot Sr. Margaret LaPorte and Harvey Finkel. M.D. are greatly appreciated. REFERENCES 1. Wyndor EL, Gori GB: Contribution of the environment to cancer incidence: an epidemiologic exorcise. | Natl Cancer Inst 1977; 58: 825-832. 2. Fortwcngier HP, Jones D, Tamburro CH, Espinosa E: Factor VIII content as evidence for endothelial origin of vinyl chloride associated liver angiosarcoma (VCA). Fed Proc 1979; 38: 999. 3. Thomas LB, Popper H: Pathology of angiosarcoma of the liver among vinyl chloride--polwinyl chloride workers. Ann NY Acad Scl 1975; 248: 208-277. 4. Creech JL, Johnson MN: Angiosarcoma of the liver in the manufacture of polyvinyl chloride. J Occupnlional Med 1974; 16:150-151. 5. Block JB: Angiosarcoma of the liver following vinyl chloride exposure. |AMA 1974:229:53-54. 8. Falk H, Creech JL. Heath CW, Johnson MN, Key MM: He patic disease among workers at a vinyl chloride polymer ization plant. JAMA 1974; 230: 59-63. 7. Splrtas R. Kaminski R: Angiosarcoma of the liver in vinyl chloride/polyvinyl chloride workers. | Occupational Med 1978:20:427-429. ' 8. Lee FI, Harry DS: Angiosarcoma of the liver in a vinyl- chloride worker. Lancet 1974:1:1316-1318. 9. Lloyd |W: Angiosarcoma of the liver in vinyl chloride/poly vinyl chloride workers. | Occupational Med 1974; 16: 809. 10. Lloyd jW: Angiosarcoma of the liver in vinyl chloride/poly vinyl chloride workers, j Occupational Mad 1975; 17: 333-334. 11. Makk L, Delmore F. Creech JL Jr, el al.: Clinical and mor- pholegic features of hepatic angiosarcoma in vinyl chloride workers. Cancer 1976; 37:149-163. 12. Whelan )G Jr. Creech JL Jr. Tamburro C: Angiographic and radionuclide characteristics of hepatic angiosarcoma found in vinyl chloride workers. Radiology 1976:118:549-557. 13. Curran KL. Kupchelle CE. Tamburro CH: Urinary glycosaminoglycan patterns in angiosarcoma of the liver. Cancer 1977; 40:3050-3053. 14. Kupchella CE. Tamburro CH: Urinary and tissue glycosaminoglycan patterns in hepatic angiosarcoma. In: Nieburgs HE. ed. Prevention and detection of cancer. Part 1, Vol 1. New York: Marcel Dekker, Inc. 1977,915-926. 15. Wyatt RH, Kotchen }M, Hochstrasser DL, et al.: An epide miologic study of blood screening tests and illness histories among chemical workers involved in the manufacture of polyvinyl chloride. Ann NY Acad Sci 1975; 246:80-87. 16. Selikoff IJ. Hammond EC, edsT Toxicity of vinyl chloridepolyvinyl chloride. Ann NY Acad Sci 1975; 246:4-322. 17. Locker GY, Doroshow JH, Zwelling LA, Chabner BA: The clinical features of hepatic angiosarcoma: a report of four cases and a review of the English literature. Medicine (Baltimore) 1979:58:48-64. 18. Delorme F, Theriault G. Ten cases of angiosarcoma of the liver in Shawinigan, Quebec. | Occupational Med 1978; 20: 338-340. 19. Editorial: Vinyl chloride end cancer. Br Med J 1974; 1: 486-487. 20. Berk PD, Martin JF, Young RS, et alu Vinyl chloride--asso ciated liver disease. Ann Intern Meid 1976; 84:717-731. 21. Wegman DH. Peters JM, Jaeger RJ. Burgess WA. Boden LI: Vinyl chloride: Can the worker be protected: N Engl J Med. 1976; 294: 653-B57. 22. Gay BW Jr, Lonncman WA. Bridbord K, Moran JB; Msa suremonts of vinyl.chloride from aerosol sprays. Ann Mv Acad Sci 1975; 246/266-295. 23. Osier RH. Carr CT. Krantz |C: Narcosis with vinyl chloride Anesthesia. XXVII. 1947; 8:359-361. 24 Cordier JM. et al.: Acroostcolyso et Iesidos cutanecs associees chozdcux ouvriers. affectes au nettoyage d'antoclaves. Cat, Med Travail 1966:4: 14-19. 25. Wilson RH. McCormick WE. Tatum CF, Creoch JL: Occu pational acroosteolysls: report of 31 cases. JAMA 1967- 2fli' 577-581. 26. Vellman C. Lange CE. Juhe S. Stein G. Bachner U: Clinical manifestations and course of vinyl chloride disease Ann NY Acad Sci 1975:246:6-17. 27. Torkelson TR, Oyen F, Rowe VK: The toxicity of vinyl chlo ride as determined by repealed exposure o[ laboratory animals. Am Ind Hyg Assoc 11981; 22:354-361. --1 28. Tribukh SL. Tikhomirova NP. Levina SV, ot al.: Working conditions and measures for their improvement in the production and use of vinyl chloride plastics. Gig Senit 1949; 14:38-44. 29. Popper H. Thomas LB: Alterations of liver and sploen among workers exposed to vinyl chloride. Ann NY Acad Sci 1975246:172-193. 30. Gedigk P, Muller R. Bechtelshcimcr H: Morphology of liver damage among polyvinyl chloride production workers. A report on 51 cases. Ann NY Acad Sci 1975; 246:278-285. 31. Viola PL. Biaotti A. Caputo A: Oncogenic response of rat side. lungs and bones to vinyl chloride. Cancer Res 1971; 31; 516-522. 32. Makk L. Creech JL, Whelan JG. Johnson MN: Liver daraags and angiosarcoma in vinyl chloride workers. A systematic deleclion program. JAMA 1974:230:64-68. 33.. Creoch JL, Makk L: Liver disease among polyvinyl chloride production workers. Ann NY Acad Sci 1975:246:88-94 34. Tabershaw IR. Gaffey W: Mortality study of workers in the manufacture of vinyl chloride and polyvinyl chloride. J Occupational Med 1974:1G: 509-518. 35. Nicholson WJ. Hammond EC. Scidman H, Selikoff IJ: Mor tality experience of a cohort of vinyl chloride-polyvinyl chloride workers. Ann NY Acad Sci 1975; 246:225-230. 36. American Cancer Society: Cancer statistics, 1979. Ca 1979, 29:6-21. 37. Heath CW. Falk H. Creech JL: Characteristics of cases ol angiosarcoma of the liver among vinyl chloride workers in the United States. Ann NY Acad Sci 1975: 246: 231- 236. 38. Popper H, Thomas LB, Tolies NC, Fa|k H, Selikoff IJ: De velopment of hepatic angiosarcoma in man induced by vinyl chloride, thorotrast, and arsenic. Am J Pathol 197* 92: 349-369. 39. Alpert LI, Benisch B: Hemangioendothelioma of the liver associated with microangiopathic hemolytic anemia. An J Med 1970; 49:624-628. 40. Donald D, Dawson AA: Microangiopathic haemolytic anemi* associated with malignant haemangio-endothelioma. J Clin Pathol 1971; 24:456-459. 41. Buntine DW. Lyall IG, Renowden VG: Haemangioendothelial sarcoma of the liver. Med J Aust 1971; 1:201- 203. 285 February 1981 The American Journal of Medicine Volume 70 AP00024515 OCCUPATIONAL CAKCINCJCENKSIS-DANNAHEK ET AU he my, ark not, are l;a:NY Tide. fees Cah !:cu201; 'cal inn . hlo- lory' king she fanit long '975; 'iver *A -a 31: iatk ride 94. : the Je-I lor* inyl 10. 979: s of -- 31- Dftiby 97f iver Am 2JIM Clin oen*01- - 42. Pollard SM, Miliwnrd-Sadlcr GH: Malignant hacmnn- giocndothclioma Involving the liver. ] Clin Pathol 1874; 27; 214-221. 43. Keplinger ML. Goode |W, Gordon DE; Interim results of exposure of rats, hamsters and mice lo vinyl chloride. Ann NY Acad Sci 1975; 246:219-224. 44. Maltoni C, Lcfemine GL; Carcinogenicity bloassays of vinyl chloride. I. Research plan end early results. Environ Res 1974; 7; 387-405. 45. Maltoni C: Predictive carcinogenicity bioassays in industrial oncogenesis. Prog Biochom Pharmacol 1978; 14; 47-56. 46. Tamburro CH; The hepatic role in carcinogenesis end Its early detection--the vinyl chloride model. Yale I Biol Med 1976:51:67-80. 47. da Silva Horta |; Lata effects of thorotrast on liver and spleen and their efferent lymph nodes. Ann NY Acad Sci 1967; 145:676-699. 48. Curry JL, Johnson WG. Fclnborg DH, Updegrove JH: Tho rium Induced hepatic hemangioendothelioma. Am | Roentgen Radium Ther, Nuc Med 1975; 125:671-677. 49. Underwood JCE, Huck P: Thorotrast associated hepatic an giosarcoma with 36 years latency. Cancer 1978; 42: 2010-2612. 50. Roth F: The sequelae ot chronic arsenic poisoning In Moselle Vintners. Cormtm Med Monthly 1957; 2:172-175. 51. Lnnder)]. Stanley Rj. otnl.: Angiosarcoma of liver associated with Fowler's solution (potsssium arsenitc|. Gastroenter ology 1975; 68:1582-1586. 52. Rogolson W, Kin U, Ospina |. Holland |F: Hemangioen- dotholial sarcoma of liver from chronic arsenic intoxication by Fowler's solution. Cancer 1968; 21:514-522. 53. Baker HC, Paget GE, Davson j: Hemangioendothelioma (Kupffer cell sarcoma) of the liver, j Pathol Bacteriol 1958; 72:173-182. 54. Kvvlllken ). Tartow LR; Hemochromatosis and Kuptfer cell sarcoma with unusual localization ofiron. j Pathol Bacteriol 1966:92:571-573. 55. Sussman EB. Nydick 1. Gray G: Hemangioondothclial ost eoma of the liver and hemochromatosis. Arch Pathol 1974; 97:39-42. 1 58. Pimcntal )C, Menezcs P; Liver disease in vineyard sprayors. Gaslroenterology 1977; 72:275-283. 57. Hoch-Ligcti C: Angiosarcoma of the liver associated with diothyXstilbostroI. JAMA 1978; 240:1510-1511. 58. Biumonfeld TA, Fleming ID. Johnson WW: Juvenile hem angioendothelioma of the liver. Cancer 1969; 24: 853- 857. r., it'. m '.w4yi February 2981 The American Journal of Medicine Volume 70 287 l . IK). ;; SBsmmm AP00024516 SCIENTIFIC AMERICANEstablished 1845 August 1979 Volume 241 Number 2 I Bacterial Tests for Potential Carcinogens New short-term tests can identify environmental agents that cause damage to DNA, theprimary event in chemical carcinogenesis. The tests are also valuable for clarifying the mechanism ofDNA damage by Raymond Devoret he fact that physical and chemi beings, whereas bacteria are minute sin Tcal agents in the environment en gle cells at the opposite end of the evo hance the incidence of cancer has lutionary-scale. It may therefore seem become of great concern. It is clear thpaatradoxical that bacteria can serve to an adequate limitation of human expo identify substances that cause cancer. sure to carcinogens would save lives. To Actually, however, there is no paradox. identify potential carcinogens in the en One tends to think of a cancer as a vironment is therefore an urgent task. In tumor that can spread through the body the case of chemical carcinogens that is to form multiple tumor colonies (metas no easy matter. It is estimated that more tases). That is a clinical, macroscopic than 50,000 different man-made chemi view of cancer at a multicellular stage, cals are currently in commercial and in since just one gram of malignant tumor dustrial use; between 500 and 1,000 new already contains a million cancer cells. chemicals are put on the market every Cancer begins at the level of the single year. The standard animal tests for po cell. A cell in an adult tissue evolves in tential carcinogenicity take a long time such a way that it departs from con and cost a great deal of money. formity with the strict physiological Fortunately there is an alternative to rules governing the set of identical cells the classical animal tests. One can take that constitute a tissue; it becomes a advantage of the profound unity of liv unique and distinguishable defect in an ing matter and resort to bacteria as the otherwise monotonous structure. The test organisms. A bacterial assay for car cell begins to divide and a tumor grows. cinogenicity takes a few hours or days Some of the daughter cells may break rather than the two or three years re the tissue barrier, invading adjacent tis quired for an animal test, and it costs sues and usually metastasizing to distant far less. An effective bacterial test has sites. Cancer cells have a great selective been developed by Bruce N. Ames of advantage, since they escape the pro the University of California at Berke grammed fate of most normal cells: to ley, based on the ability of a chemical age and die. For cancer cells the entire to cause mutations in bacteria. More re body is a culture medium in which they cently my colleagues and I at the Centre thrive, ultimately to die with the body National de la Recherche'Scientifique in they kill. Gif-sur-Yvette have devised a group of tests based on a chemical's ability to in physical and chemical agents in the duce the development of a dormant vi A environment cause cancer by damag rus in bacteria. The Ames test and our ing DNA the cell's hereditary material. tests not only provide means of identify DNA damage initiates a complex cellu ing dangerous chemicals but also are lar process that in mammalian cells can powerful new tools for learning to un eventually lead to transformation into derstand the primary events of the carci a cancerous state. Agents that damage nogenic process initiated by chemicals. DNA are therefore potential carcino Cancer is a disease of highly evolved gens. DNA is the hereditary material of multicellular organisms such as human all living cells, and both DNA lesions and the cellular processes that repair them are remarkably similar in bacteria and in human cells; what is detrimental to bacterial DNA is likely to harm hu man DNA. That is the theoretical jus tification for substituting bacteria for mammalian cells in tests to detect dam age to DNA The theoretical justification is sup ported by experimental and practical results: bacterial tests distinguish with more than 90 percent reliability between known carcinogens and known noncar cinogens, and they have identified as po tential carcinogens new chemicals that have subsequently been shown to be carcinogenic in animal tests. Of course, the manifestations of DNA damage are very different in bacteria from the trans formation to the cancerous state that is observed in mammalian cells. In com pensation the bacterial tests are so much faster and less expensive as to finally make comprehensive screening for po tential carcinogens a feasible objective. The carcinogenic potential of physi cal agents, and notably of ionizing radi ations, is much better understood than that of chemical agents. Broad popular awareness that radiations can cause can cer has come only in the era of nuclear weapons and reactors, but the danger actually became apparent soon after the discovery of X rays by Wilhelm Kon rad Rdntgen in 1895. Only four years later it was reported that a technician who checked newly manufactured Xray tubes by fluoroscoping his own hand was afflicted with a skin cancer; he even tually died of it The warning was ig nored, and most of the first generation of radiation therapists died of cancer. The carcinogenic effect of the ultravio- AP00024517 I <. r +; * t, f> I t f The Compact Edition of the Oxford English Dictionary. As an introduction to membership in the Book-of-the-Month Club" YOURS FOR ONLY $1995 (PUBLISHER'S LIST PRICE: $125) You simply agree to buy 4 books within a year. For the price of an average dictionary, you can now treat yourself and your family to the world's best dictionary of the English language. The contents of this two-volume edition are identical to those of the original thirteen-volume set. priced at $495. FEATURES Boxed set of two volumes, fffc" by 13t4" each. All 16.569 pages of the 13-volume original included in 4134 pages of The Compact Edi tion through a photo-reduction process which permits printing of four pages of originaf on one page of compact edition. * Paper is 30-pound Special Dictionary White. Binding is library buckram reinforced and gold-stamped. * pausch & Lomb magnifying glass included in special drawer of slipcase. T by 3?i" lens scientifically designed to make reduced print easily readable. Book critics call the OED a "miracle." Join the Book-of-the-Month Club and this "mira cle" can be yours now. As a Club member, you'll continue to enjoy benefits on the best and most important books published today. Bookstore Quality at Book-Club Savings Vou conveniently shop at home at consider able savings. Whether you're adding up your remarkable savings on the introductory offer, or on books you are offered as a member, these are always true savings...because every Club book is as good as. or better than, those sold in stores. You don't settle lor the altered or inferior editions that some book clubs send their members. Book-Dividends. When you remain a Cub member after the trial period, every book you buy earns Book-Dividend* credits. These entitle vou to choose from a wide variety of significant books at hard to-believe savings of alkast 7(fh BOOK-OF-THE-MONTH CLUB* AMERICA'S BOOKSTORE* AP00024518 let radiation in sunlight has also been known for some time. As long ago as 1905 a French physician named Dubreuilh observed that skin cancers of the back of the neck were particularly prev alent among workers who were exposed to the sun as they tended the vineyards and harvested grapes in the Bordeaux region. Although X rays, gamma rays and other radiations hit all the components of cells in a random manner, it was recognized quite early that the genetic ma- terial must be the most radiation-sensitive cellular target DNA constitutes only a minor fraction of the chemical components of a cell, but direct or indirect damage to it has great impact on the cell's future, whereas damage to pro- teins and other cellular components has much less effect That is because DNA is the memory of the cell. DNA replicates to beget DNA, and so errors in DNA are transmitted from cell generation to cell generation. The double helix of DNA consists of two chains of sugar (deoxyribose) and THREE BACTERIAL TESTS for potential carcinogens reveal DNA damage: the Ames test (lop), the induetest (middle) and the lambda mutatest {bottom). In each case the culture plate at the left Is an un treated control; tester bacteria on tbe center plate were treated with a moderate dose and those at the right with a higher dose. Tbe Ames test shows the extent of reverse mutations in histidine-deficient Sal monella typhimurium that enable the revertant bacteria to prolifer ate. A background of spontaneous mutant colnnles (red stain) Is seen at the left Many more colonies grow when tester bacteria art exposed to 250 nanograms (center) and 750 nanognuns (right) of the potent mutagen (and carcinogen) nitrosoguanidine. In the Induetest (middle) DNA damage is revealed by the induction ot a prophage, a dormant bacterial virus integrated in the DNA of "lysogenic" Escherichia coti; mature phage particles bunt out of the tester bacteria and create plaques on a lawn of Indicator bacteria of strain A (red slain). Here the DNA damage was caused by the antitumor drug mitomycin C, 10 nanograms of it on the center plate, 200 nanagnms at tbe right. In tbe mutatest (bottom) a modified form of the prophage makes plaques on a lawn of . coti at strain 0 (on which nonmutated phage cannot form plaques) when it undergoes mutation In its "operator" regions and can no longer remain dormant Again the treatment was 10 nanograms of mitomycin C (center) and 200 nanognms (right)u AP00024519 3 PROTEIN-DNA CROSS-LINK STRUCTURAL ALTERATIONS arc im- pofed by radiations or by chemical agtals on the double helix of DNA, two chains of sugar and phosphate (roups (helical iaiuh) linked by paired bases: either adenine (A) and ttymine (T), or (uanine (G) and cytosine (C). The alterations can be classified in five categories, examples of which are illustrated: negligible helix distortions (1), as by alkylation of one of the bases; minor distortions (2) caused by hydration or the absence of a base; major dis tortions (J) caused by insertion of an "adduct," Unking of two bases to form a dimer, or crosslinking between the two strands or between a strand and a protein) breaks in a single strand (d) or in both strands (JL Any structural alter ation affects DNA's function as a template for replication, but some negligible alterations are not sensed by a cell as damage to DNA. phosphate groups linked, as by the rungs of a twisted ladder, by paired nitroge nous bases: two purines (adenine and guanine) and two pyrimidines (thymine and cytosine). Adenine always pairs with thymine and guanine with cytosine, but the sequence of the bases along a strand of DNA is variable and carries a particular coded message. Any altera tion, even a slight one, in the structure of the double helix affects the functions of the DNA, one of which is to serve as a template for its own replication. Not every DNA alteration is sensed in the cell as DNA damage. Defined pre cisely, DNA damage is an alteration that constitutes a stumbling block for the replication machinery and hence hampers the replication of DNA, en dangering the survival of the cell. Once incurred, DNA damage calls for repair, which is accomplished by the interplay of at least a score of enzymes whose action is governed by as many genes. Repair is never totally efficient, and so many cells die. Some cells may survive, however, even though the lesions are not totally removed from their DNA, be cause a repair process has bypassed the lesions. Replication then reconstitutes an undamaged double helix, but one bearing a coded message different from the original one. The scars left on the DNA by such a process are mutations. The correlation between radiationinduced cancer and radiation-induced DNA damage has long been apparent to radiation biologists, but it is only recent ly that molecular evidence has been found that DNA damage is a direct cause of cancer. The evidence comes from patients suffering from xeroderma pigmentosum, who are extremely sensi tive to sunlight and, while they are still very young, develop skin cancers of which they may eventually die. James E. Cleaver of the University of California School of Medicine in San Francisco and Dirk Bootsma of Erasmus Universi ty in Rotterdam have demonstrated that xeroderma patients suffer from a wellcharacterized genetic defect: their cells cannot carry out a particular DNArepairing process. At noontime on a sunny day the flow of ultraviolet radiation that reaches the earth is strong enough to generate py rimidine dimers in the DNA of exposed cells by linking two laterally adjacent thymines or a thymine and a cytosine. Most such DNA lesions in skin cells are repaired in normal people by an ex cision process strikingly similar to the "cut and patch" repair process in bac teria exposed to the same radiation. In xeroderma patients, however, the le sions go unrepaired, and their accumu lation appears to bring on cell transfor mation: DNA damage breeds cancer. Appreciation of the carcinogenic role of chemicals in the environment has come slowly, even though instances of cancer caused by occupational exposure have long been observed. As early as 1775 the British physician Sir Percival Pott correlated the incidence of cancer of the scrotum in men who had once been chimney sweeps with the accumu lation of soot in their groin area many years before. Experimental findings on chemical carcinogens date back at least to 1918, when two Japanese investi gators. K. Yamagiwa and K. Ichikawa, showed that skin cancers could be in duced by repeated applications of coal tar to the ear skin of rabbits. One of the chemicals responsible for causing such cancers is benzo[a]pyrene, which is also present in soot, cigarette smoke and charred meat. Progress toward understanding chem ical carcinogenesis has been slow for at least three reasons. First of all, most chemical carcinogens are not biological ly active in their original form, so that testing them in that form does not reveal their carcinogenic nature. Only about a decade ago did it become clear, as a re sult in particular of the investigations of James A. Miller and Elizabeth C. Miller of the University of Wisconsin, that nor mal metabolic processes, which convert food into substances the body can ab sorb and eliminate and convert harmful compounds into harmless ones, trans form environmental chemicals into me tabolites capable of inducing cancer. Those metabolites had to be character ized in order to show just how their par ent substances are threats. A second delaying factor was that the actively carcinogenic metabolites react with various cell components, including RNA and proteins as well as DNA. Since there are a lot of proteins, per forming important functions, it was of ten assumed that damage to proteins might play the major role in the cancer ous transformation initiated by chemi cal carcinogens. The key role of damage to DNA in carcinogenesis by chemicals was recognized only recently, and the results of the bacterial tests have provid ed strong (although indirect) evidence for such a mechanism. Finally, the understanding of chemi cal carcinogenesis has been obscured by the fact that DNA damage, although it is the essential event in the initiation of carcinogenesis, does not usually in itself lead to cancerous transformation; addi tional factors are apparently required to promote the complex chain of cellular events culminating in the transforma tion. DNA-damaging agents in them selves are therefore only potential car cinogens. As I pointed out at the beginning of this la article, only a small fraction of the flood of chemicals reaching the market every year can be tested accurately by means of the standard animal assays. In order to obtain results with statistical significance a great many animals must Y+ AP00024520 Mi Si. > - %: be (or at least should be) exposed to each tested chemical; for that reason alone a comprehensive screening of new chemicals would be unpractical (other than the few, such as food additives, drugs and cosmetics, for which testing is mandated). Even when animal tests are feasible, manufacturers need low-cost, fast tests if they are to identify DNA- damaging substances while new prod ucts are still under development and al ternative ones can still be sought Studies of the epidemiology of human cancers have provided much informa tion on environmental carcinogens, but such studies are of little immediate val ue in detecting new potential carcino gens because most human cancers ap pear only some 20 or 30 years after the exposure that gives rise to them. Epide miological analysis has most frequently been successful in detecting a chemical to which an identifiable subpopulation, such as workers in a particular industry, are exposed and because of which they show a high incidence of a particular type of cancer. For some widely distrib uted carcinogens there are no clearly identifiable subpopulations for statisti cal analysis. There is no doubt, then, that the pres ent situation calls for simple, fast, inex pensive methods of detecting potential carcinogens, which is to say DNA-dam aging agents. One possibility is to mea sure DNA damage directly in mam malian cells by determining biochemi cally the incidence of particular forms of DNA damage In cells exposed to a chemical. Such studies are being -done by molecular biologists, and their re sults provide valuable standards of ref erence. For screening purposes, how ever, it is cheaper and faster to detect and measure the extent of DNA damage by scoring its manifestations in bacteria. One of the great advantages of assays done with bacteria is the enormous bio logical amplification implicit in bacteri al manipulations. It is easy to grow as many as a billion (10*) bacteria per mil liliter of culture medium. A mutational event such as a change in a single base pair in the bacterial DNA, which is im possible to detect by standard biochemi cal methods, will be repealed as a mu tant bacterium. That single bacterium can be selected from among 10* cells because its daughter cells, and only they, will proliferate and form a colony visi ble to the unaided eye on an agar nutri ent plate. Since a colony consists of about a million (10*) bacteria, a rare sin gle mutational event with a probability of, say, one in 100 million (a probability of 10-*) would thus be amplified by a factor of 100 trillion (10M). which inhabits the colon of a number of mammals including human beings, is genetically programmed to divide and (under optimal conditions for growth) form two daughter bacteria in 30 min utes. When the DNA of a bacterium is damaged, the bacterium may need more than two hours to resume its cycle of division (if it is not killed by the dam age). During that time there is a se quence of cellular events. Immediately after the primary DNA lesions are incurred a first round of re pair is effected: an excision repair, in which the damaged segment is cut out of the DNA and replaced by an undamaged DNA sequence. Some residual le sions will remain, however. If they are bulky or clustered, the residual lesions hamper the DNA-replication machin ery, and replication stops abruptly. Un less the arrest of replication is transitory 'v < The series of cellular events that leads to mutation, and to several other XERODERMA PIGMENTOSUM is a genetic disease whose victims develop slrin cancer as a result of normal exposure to sunlight, typically on uncovered parts of the body, as Is shown la the photograph at the top. These patients lack a DNA-repahr process that In normal Individuals manifestations of DNA damage in bacremoves thymine dimers, the lesions created by the ultraviolet radiation present In sunlight (ft teria, has recently been somewhat clar In the normal repair process the bases in the lesion areas are first excised (% Then the excised ified. The bacterium Escherichia coli, stretches ol DNA are resynthesized and the new segments are ligated to undamaged stretches (JJl i AP00024521 f CAUSAL EVENTS CELLULAR PATHWAY X OBSERVED ^ PHENOMENA CLEAVAGE OF prophage Repressor ULTRAVIOLET RADIATION t r DNA LESIONS AFLATOXINB, IMMEDIATE EXCISION, RECOMBINATION RESIDUAL LESIONS UNSCHEDULED ARREST OF DNA REPLICATION INDUCIBLE ERRORPRONE REPAIR /-'CELL vn,SURVIVM.' rrrr------------- "t -V'~ - m .MUTAGENESIS ->m ! INHIBITION OF CELL PARTITION CELL T ELONGATION. SEQUENCE OF CELLULAR EVENTS follows DNA damage im posed on . coll by ultraviolet radiation or by the carcinogen aliatoxin Bi. Immediate repair processes mend most of the lesions but leave residual damage, causing an unscheduled arrest of DNA repli cation. This threat to cell survival activates the RecA protein to be come a protein-cleaving enzyme and also induces the synthesis of more RecA. Various forms of the protein are Involved in three proc esses that give rise to the four observed phenomena shown at right SUSPECTED CARCINOGEN ON FILTER PAPER I MUTAGENESIS is detected in the Ames test by mixing an extract of rat liver (which supplies mammalian metabolic functions) with tester bacteria (which cannol grow because a mutation makes them unable to manufacture histidine, a necessary nutrient) and plating the mixture on an agar medium so that a thin layer of bacteria covers the medium evenly, as is shown on a microscopic scale (/X, In this "spot assay" a dose of the chemical to be tested is placed on a disk of filter paper on the tester bacteria. After two days most of the hit - bac teria bare died for lack of histidine (A but DNA damage caused by the chemical diffusing out from tbe disk has given rise to mutations, some of which result in reversion of the his- mutation. The hislidinemaking revertant bacteria proliferate, forming visible colonies O). T* TESTER BACTERIA , 1 SUSPECTED CARCINOGEN ON FILTER PAPER RAT-UVER I INDICATOR EXTRACT BACTERIA W INDUCTION of a dormant bacterial virus, prophage lambda, is de tected in the inductest Lysogenic tester bacteria are mixed with ratUver extract and then with Indicator bacteria. Tbe mixture is plated; the medium is covered with a thin layer of indicator bacteria inter spersed with a few lysogenic bacteria (/). After t day most of the plate Is covered by a thick lawn of indicator bacteria (2). Where the chemi cal that is being tested has diffused from tbe filter-paper disk tbe DNA damage it causes leads to Ihc induction of mature lambda phage. The phage particles bunt out of the lysogenic cells and kill indica tor bacteria in the vicinity, making visible plaques on the lawn (J). AP00024522 the cell's survival is at risk. As Miroslav histidine. Obviously the spontaneous sources of potential carcinogens has re Radman of the University of Brussels rate of reversion, which is ordinarily sulted in its current application in more and Evelyn M. Wilkin of Rutgers Uni very low, will be considerably enhanced than 2,000 governmental, industrial and versity first suggested, the following cel if the his bacteria are exposed to a academic laboratories throughout the lular adaptive mechanisms come into chemical that induces mutations. This is world; it is estimated that 2,600 chemi play to cope with the emergency: the theoretical basis of the Ames test. cals have been subjected to the test. 1. A second round of repair is in Actually Ames and his colleagues had Ames and his collaborator Joyce Mc duced. This inducible and error-prone to introduce three important modifi Cann have themselves validated the repair (nicknamed "SOS repair") tends to restore the structure of the DNA even though there are errors in the coded cations into the original his- strain to make it a sensitive and versatile tester bacterium. Bacteria such as coll and procedure by testing more than 300 chemicals that were previously report ed, on the basis of animal experiments, message; indeed, this adaptive response S. typhimurium have a rather imperme to be either carcinogens or noncarcino may be successful partly because it does able envelope that reduces or even pre gens. About 90 percent of the reputed no) "bother" to follow the base-pairing vents the penetration of many chemicals carcinogens turned out to be mutagenic rules of normal DNA replication. At into the cell. (This bacterial armor has and about the same proportion of the any rate, the price paid for cell survival evolved because the bacteria must usu noncarcinogens were negative for muta appears often to be mutagenesis. ally survive in a hostile environment genicity. Other mutagenicity tests have 2. Cell partition ceases. The elonga such as an intestine or a sewer.) Ames since been devised with coli as the tion of the cell that ordinarily precedes and his colleagues overcame the enve tester bacteria, and their efficiency is division is protracted, and the ceils may lope barrier by introducing a mutation about as high as that of the original Sal form filaments. The elongation may be that gives rise to defects in the envelope. monella test. adaptive in that it facilitates recombina They went on to make the strain more Two impressive accomplishments of tion between the two sets of damaged sensitive to DNA-damaging agents by mutagenicity tests can be mentioned to chromosomes in the cell, the intact seg eliminating its capacity for excision re give an idea of their value. In Japan the ments of each combining to yield an in pair, so that most of the primary lesions chemical furyl furamide, known as AF- tact chromosome. remain unhealed. And they introduced 2, was added to a broad range of com 3. If a prophage, or dormant bacterial into the bacterium a plasmid, a foreign mon food products for some years as an virus, is present in the cell, it is induced genetic element that makes DNA repli antibacterial agent. It had not shown to develop into a large number of ma cation more error-prone. By means of any carcinogenic activity in standard ture progeny phage that burst out of the these three modifications a strain was tests on rats in 1962 or on mice in 1971. cell. This adaptive response evolved by constructed in which just a few mole Then in 1973 T. Sugimura and his col the prophage ensures its survival when cules of a carcinogen are able to create leagues at the National Cancer Centre in a host cell appears to be doomed to die: DNA lesions, each of which is likely Tokyo found that AF-2 was highly mu the rats leave the sinking ship. to engender a mutation; of those muta tagenic in bacteria; they could easily Each of these adaptive responses is in tions, some will be such that the internal demonstrate the mutagenic activity of part promoted by a multifunctional pro supply of histidine is restored. the additive contained in just one slice of tein called the RecA protein (for "re The real breakthrough, and the one fish sausage! The discovery prompted a combination," since defects in the pro that made the Salmonella test truly effec new round of more thorough animal tein impair recombination in general). tive, was Ames's idea of mixing the test test* for carcinogenicity, which showed The RecA protein appears to be re er bacteria with an extract of rat liver that AF-2 was indeed a carcinogen. It quired in coli not only for recombina and thereby subjecting the tested chemi was withdrawn from the market If it tion but also for the responses listed cal to mammalian metabolic processes. were not for the bacterial test, AF-2-- above. The RecA protein is activated As I pointed out above, it is usually not which had passed two approved animal and its synthesis dramatically increased the original form of a chemical carcino tests and been declared negative for car when DNA replication is blocked. gen that is active but rather one of its cinogenicity--would presumably still be metabolites. Because the liver is the a component of fish sausage and other Of the various bacterial manifes body's major metabolic factory the en Japanese food products. tations of DNA damage, Ames zymes of a rat-liver extract should con In 1973 Ames and his colleagues re chose mutagenesis as the basis of hivsert the chemical being tested into me ported that 89 percent of the major class pioneering work to develop a test for tabolites that react with DNA--if there of hair dyes sold on the U.S. market potential carcinogens. His Salmonella- are any. contained mutagenic compounds. Since mammalian-liver assay, known general In practice the Ames test is usually then the cosmetic industry has modified ly as the Ames test, is currently the stan done by adding the chemical to be ex the composition of most hair dyes. It is dard test and by far the most widely amined to his- tester bacteria immersed estimated that several tens of millions of used. The tester organism is a strain of in a rat-liver extract and plating the mix people dye their hair in the U.S. alone, Salmonella typhimurium, 'another colon ture on a solid nutrient medium devoid which suggests that the discontinued bacterium, bearing a mutation (his-) of histidine. (For demonstration purpos components had presented a consider that renders it unable to manufacture es the chemical can Instead be spotted able risk. one of the enzymes required for the syn on a disk of filter paper, which is then In spite of all their advantages the mu thesis of the amino acid histidine, a nec placed on a medium on which the bacte tagenicity tests have some technical and essary component of proteins. As a re ria, mixed with the liver preparation, even theoretical limitations. Since muta sult of the mutation the bacterium is have previously been plated.) After two tions are revealed in the Ames assay by a unable to grow in a mineral nutrient me days of incubation any cells that have restoration of enzyme activity, any mu dium unless the medium is supplement undergone the reversion mutation will tation that does not happen to recon ed with an external supply of histidine. give rise to revertant colonics. The num struct the precise DNA sequence that On very rare occasions a his- muta ber of such colonies per mole of the test codes for the histidine-making enzyme tion undergoes reversion: a back muta ed chemical provides a quantitative esti is not observed. To take one example, tion restores the DNA's normal coding mate of the mutagenic potency of the the antitumor drug bleomycin, which sequence for the needed enzyme and chemical. does its therapeutic work by damag thereby restores the internal supply of ing the DNA of tumor cells (as do Thistidine. The reversion can be scored because only the revertant bacteria he simplicity, sensitivity and accura about half of all antitumor agents), fails cy of the Salmonella test for screen to induce the mutation that is scored in form colonies on a medium that lacks ing large numbers of environmentalthe Salmonella test. A false-negative re- i i AP00024523  ti, i kponse of this kind is a technical prob mutations. A mutagen can have strong proteins forming the virus particle, and r- i lem that can be remedied by substituting genetic effects on a biological popula a progeny of mature phage develops, other tester bacteria or a complementa tion by altering the information encod bursting out of the host cell. The process ry short-term test ed in the DNA of individuals' germ cells is called lysogenic induction. Under nor False-positive responses are more sig but nonetheless fail to cause cancer be mal conditions the dormant state of pro nificant from a theoretical point of view cause it does no real DNA damage to phage lambda is maintained by a repres because they may raise some doubts the somatic cells: the cells of the rest of sor. a protein that lies on the DNA'* about the validity of mutagenicity tests the body. Not all genetic toxic sub "operator" regions and blocks the oper for identifying potential carcinogens. stances (chemicals that affect DNA) are ation of the lambda genes (except the The point is that some chemical reac potential carcinogens, whereas chemi gene that directs the synthesis of repres tions with DN A are highly mutagenic in cal carcinogens--because they damage sor to keep the prophage dormant). Jef bacterial and mammalian cells without, DNA--are all indirect mutagens. In oth frey W. Roberts and Christine W. Rob as far as one can tell, being carcinogenic. er words, mutagenic activity in bac erts of Cornell University discovered Among these are the incorporation of teria is correlated with carcinogenesis that the induction of prophage lambda an analogue of one of the nitrogenous in mammals primarily through indirect results from the cleavage of the lambda bases and the methylation of certain mutagenesis, which results from DNA repressor; together with Nancy Craig sites on the bases. Such reactions cause damage. they have recently shown that relatively negligible alterations in DNA structure pure repressor can be cleaved when it is that are not sensed in the cell as DNA damage; DNA replication proceeds on schedule and the new DNA carries a In 1953, before the structure of DNA mixed with an activated form of the had even been established, Andrd RecA protein. Lwoff of the Pasteur Institute in Paris By triggering the activation of the readable--although wrong--coded mes anticipated that "inducible lysogenic RecA protein into the form that can sage. This process is termed direct mu bacteria might become a good test for cleave a viral repressor, DNA damage tagenesis, and in scoring for mutations carcinogenic and perhaps anticarcino- results in the induction of a dormant vi it should be clearly distinguished from genic activity." A bacterium is said to be rus. The induction of prophage lambda the more frequent indirect mutagenesis. lysogenic when it carries, in a dormant can therefore serve as a test for DNA In the latter process, described above, state, the DNA of a "temperate" bacte damage. Even before the mechanism of DNA damage brings about a transient rial virus, which in this dormant state is prophage development was understood arrest in replication; replication is re called a prophage. One such temperate at a molecular level, lysogenic induc sumed with the help of the RecA protein virus is phage lambda, which becomes tion was applied in the pharmaceutical on a template that carries lesions, caus a prophage when it is integrated into industry to identify prospective antibi ing mutagenesis of the newly formed the DNA of certain lysogenic strains otics and antitumor drugs. When in strand. of coll. duction tests were done with ordina Since it is DNA damage that appears When lysogenic E. coll bacteria are ry strains of lysogenic coli. however, to initiate the cancerous transformation subjected to any treatment that halts they did not give a positive response of a mammalian cell, a chemical is a DNA replication, the prophage is in for such known carcinogens as ben- potential carcinogen if it is a DNA-dam- duced: its DNA loops out of the bacteri zo[a]pyrene. aging agent, not simply because it causes al DNA and also directs the synthesis of Patrice Moreau, Adriana Bailone and SPOT ASSAYS visualize the efficacy of bacteria] tests qualitatively, although quantitative assays are preferable {set illustration on page 41% la the Ames test (left) the tested chemical was ethyl methane sulfonate, an alkylating agent and potent mutagen. A dense halo of revertant typhimurium colonies is seen around the disk from which the mutagen diffused. (Close to the disk there is a zone in which a toxic concentration of the chemical killed all bacteria.) The larger colonies are muiant bacteria that arose spontaneously without exposure to the chemical. In the inductest (right) the tested chemical was afiatoxin B,, a potent carcinogen of the liver. The afiatoxin doses placed on the four test disks were (clockwise from top right) 0, 20, 200 and 2,000 nanograms. In the background a few of the lysogenic coli tester bacteria have spontaneously produced mature phage that have killed the nearby indicator bacteria, producing plaques at random locations. AP00024524 T 1 therefore set out five years ago to reno vate the lambda-induction test to make it capable of identifying all kinds of DNA-damaging agents. We reasoned that because the molecular mechanism of lysogenic induction was understood better than that of mutagenesis, a pro phage-induction test should provide more insights than mutagenicity tests could into the precise effect of chemical carcinogens on DNA and should pro vide a supplementary tool for screening as well. Following Ames's lead, we construct ed new tester bacteria that are perme able to chemicals and deficient in exci sion-repair enzymes, and we assayed potential carcinogens in the presence of a rat-liver metabolizing mixture. The prophage-induction test, or inductest, has turned out to discriminate effective ly between carcinogens and noncarcin ogens in our laboratory and in several others. zy*rDvNaAii-POLYMERIZING COMPLEX NEGLIGIBLE DISTORTION ^4^ MUTATION REPLICATION ERROR 4 REPRESSOR" J2- ' OPERATOR - --vSc* The inductest has certain advantages over mutagenicity tests. Mutations can be scored only in bacteria that surD IRECT MUTATION is a result of DNA replication on a minor distortion of a DNA strand. vive exposure to a chemical treatment This diagram and the one below show the replication of only one strand of bacterial DNA (blue) carrying a piece of viral DNA, prophage lambda (red), which is kept in a dormant state that is often toxic as well as mutagenic, by a repressor protein that blocks an operator region. The DNA strand is replicated by a and no more than one bacterium out of polymerizing complex (I). A negligible distortion of the bacterial DNA (2) does not abstract 1,000 is likely to be detected as a histi replication, bnt it does give rise to an error in replication, so that the newly formed DNA strand dine revertant In contrast, lysogenic in {light him) carries a mutation (I). The negligible distortion is not sensed as DNA damage, RecA duction can be observed, when it takes protein is aot induced, repressor stays in place and prophage is replicated ia dormant state 5) place, in the bulk of a cell population. Prophage induction is a mass effect, largely independent of whether or not cells would survive the chemical treat ment; a cell that is induced to produce a progeny of phage would die in any case. The inductest can therefore continue to give a positive response for a highly tox A/--------------------XDNA DAMAGE V ic potential carcinogen at dose levels that would kill the tester bacteria in a mutagenicity test The fact that in the inductest most cells undergo a dramatic change led us to design a biochemical assay of lyso genic induction. Sankar Adhya, Max well E. Gottesman and Asis Das of the National Cancer Institute in the U.S. constructed a bacterial strain in which the gene for the enzyme galactokinase is hooked up to the lambda DNA in such a way that the lambda repressor blocks the synthesis of the dnzyme. Alain Le vine, Moreau, Steven Sedgwick and I demonstrated that when a DNA-dam aging chemical activates the RecA protein in this strain, the repressor is cleaved and galactokinase is synthe sized, and the amount of enzyme activi A,-------------------------- yj)- PHAGE DNA ty that can be detected reveals the extent of DNA damage. This biochemical as -mm--------------- m S 4 ..A say may be the shortest of the short term tests; it takes only half a day to test for a potential carcinogen. Moreau and I have also constructed strains, containing a new form of pro phage lambda, in which one can identify chemicals that induce mutagenesis with out damaging DNA. This lambda mu- INDIRECT MUTATION results from the events that foUow DNA damage (i), which blocks DNA replication and produces an "SOS" signal (2). A large amount of RecA protein (onutgt) is synthesized (A The RecA somehow facilitates DNA replication on a damaged template, ap parently by temporarily modifying the polymerizing complex (A This replication is highly. mutagenic not only at the lesion site bnt also elsewhere (5). Activated RecA cleaves the repres sor (4, right), freeing the operator region of the prophage DNA. The prophage is thereby in duced to form phage DNA (5), which ia turn forms virus particles that burst out of bacterium. AP00024525 ENZYME-INDUCTION TEST is biochemical counterpart of the inductcst. The lent, %alK, for synthesis of galactokinase is hooked up to propha|t-lambda DNA in such a way that eniyme synthesis is 5 DNA MESSENGER RNA 2^? GALACTOKINASE^O0 blocked by lambda repressor. When DNA is damaged, activated RecA protein cleaves repressor. Liberated galK DNA (t) is transcribed into messenger RNA (5), which is translated into galactokinase (6\ tatest, as we call it, reveals whether a ses. Several independent and comple ing standards for chemical carcinogens. chemical gives rise to mutations directly mentary determinations of a chemical's. The experts who set the standards for (that is, without causing DNA dam DNA-damaging capacity can provide radiation safety have proceeded from age and therefore without inducing pro a balanced and hence more compelling two basic facts. One is that mankind is phage) or indirectly (by damaging the assessment of carcinogenic risk. exposed to a natural background of ra DNA and producing active virus). The pronouncement that a particular diations whose biological effect is not Although neither the inductest nor the chemical is a potential carcinogen has perceptible (even though in principle lambda mutatest has yet been validated obvious social, economic and even po there is no threshold for the effects of as extensively as the Ames mutagenicity litical implications. The chemical may ionizing radiations). The other is that test, the two appear to be valuable com present a risk to a very large population, biological damage is proportional to the plementary assays, in particular for de as was the case with the food additive energy released within the living system termining the nature of DNA damage AF-2 in Japan. That situation was dealt being considered, and that this linear and for establishing a correlation be with rather easily: AF-2 could be re relation is more or less independent of tween potency in producing DNA dam placed by an innocuous substitute. It the particular source of radiation. The age in bacteria and potency in causing was dispensable, and so it was simply experts assumed that for the known cancer in mammals. Perhaps more im banned from the market. Some chemi sources an absorbed dose only a few portant in the long run, a biological sys cal carcinogens cannot easily be dis times higher than the dose delivered by tem comprising both a cell and a dor pensed with and cannot be banned; background radiation, and less than the mant virus should be a valuable tool for particular problems are raised by sub level that would cause a doubling of the analyzing the pathway of biochemical stances as different as cigarette smoke, mutation rate in a mammalian popula changes, genetic or nongenetic (cleav motor-vehicle exhaust fumes and anti tion, would constitute a negligible risk; age of repressor, for example), that are tumor drugs. For most. indispensable in general such doses were adopted as triggered when cells are exposed to car carcinogens it is clearly necessary to re being permissible for human exposure. cinogens. duce human exposure to the safest lev In the case of newly devised or new els possible. That calls for the adoption ly discovered radioactive materials for Given that man is exposed to a flood of a broadly accepted set of safety stan which there were no accumulated ex of chemicals, some of which surely dards to be implemented by stringent perimental data, permissible values were increase the risk of cancer, there is anregulations. calculated on the basis of the linear rela urgent need to identify potential carcin We stand today, with regard to chemi tion between biological damage and the ogens and lower human exposure to cal carcinogens, about where we were energy released within the organism. them to a level of negligible risk and with regard to ionizing radiations three keep it there. In such an effort bacterial decades ago. The advent of nuclear Stests can play a major role. In order to power reactors made it necessary to pro-. ince chemical pollution is worldwide, an international committee of ex ensure broad acceptance of uniform tect workers in the industry from direct perts should propose protective regu standards, some investigators have sources of radiation and to protect ev lations for adoption by individual gov strongly favored the adoption of a sin eryone from exposure to radioactive ernments. Some principles should gain gle, standardized mutagenicity test. Of effluents and solid wastes. An interna broad acceptance rather easily: for ex ten, however, it is advisable to adapt a tional agency sponsored by the United ample, all unnecessary potential carcin screening effort to the particular physi Nations worked out a set of standards ogens should be banned. A list of carcin cal and chemical properties of the sub and regulations designed to minimize ogens that cannot be dispensed with but stance that is under study. It would human exposure to various sources of to which exposure should be strictly lim therefore seem to be sensible to subject radiation. The regulations were broad ited will also have to be drawn up. It each chemical to a battery of tests. Any ly accepted and implemented, and they may be difficult, however, to reach a test has its foibles, which give rise to have stood the test of time. They should consensus on such a list and on the per false-negative or false-positive respon provide a valuable reference for evolv missible environmental concentrations AP00024526 ol each chemical. One would like to be effective in curing cancer without also ter and our food, in the products we.' - gin by defining permissible concentra causing cancer. manufacture and consume--constitutes tions of indispensable carcinogens in relation to a "natural" background. The trouble is that cultural differences, so Bacterial tests for potential carcino gens should help to answer two ques tions now being asked by industry, a biological risk? These two questions will be answered satisfactorily, how- ever, only if a third question is asked: cial habits and other factors have con government and people at large. What What is the mechanism whereby each siderably distorted notions of what is a chemicals are DNA-damaging agents? chemical affects DNA? Only basic natural background level of negligible How much of each--in our air, our wa knowledge can breed safety. risk. (For geographical, economic and religious reasons the carcinogenic back ground is surely lower in Salt Lake City la than it is in New York.) If a negligible'-risk level can be agreed on, the next step would be to estimate, for each indispensable carcinogen, the concentrations that should be allowed for workers in the industry and (at a much lower level) for the general popu 2a A lation. Can bacterial tests help in setting such levels? Yes, if there is proportion ality between potency in causing DNA damage in bacteria and potency in caus 3a ing cancer in mammals. Matthew S. --A- Meselson and Kenneth Russell of Har vard University have attempted to dem onstrate a direct relation between muta genic potency in the Salmonella test and 'IT carcinogenic potency in laboratory ani mals. For 10 of the 14 known chemical -A- carcinogens they considered there was a correlation. The many exceptions to the rule of correlation were to be expected because o in a mammal the extent of DNA dam age in a target cell depends on factors (which may be different in bacteria) such as the specific metabolism of car cinogens in the organ involved, the cells' permeability to particular metabolites and tendency to accumulate them and the extent of residual DNA damage af ter the cellular repair mechanisms have 1b n had their effect Each of these variables must be determined for each individu al carcinogen if one is to set valid levels for human exposure. Each chemical- in contrast to individual sources of ra diation--requires a thorough specific study to determine its carcinogenic po tency. Bacterial tests have good indica tive value. Even though establishing regulations. 3b #02 will be much more difficult for chemical carcinogens than it was for radiations, the regulations should be established soon. It is wiser and safpr to have safety 0*1* regulations than not to have them. A large part of the human population con tracts cancer; one person in five dies of 4b O-* -yf;-* it. It is important to determine the pre cise mechanism by which each DNAdamaging agent acts on DNA, not only "O' to prevent cancer but also to develop better ways of curing it. About half of DIRECT AND INDIRECT MUTAGENESIS can be disUngulshed from each other by the all the drugs administered in an effort mutatest, based on mechanisms diagrammed here. The repressor binds to a slightly modified to arrest cancer are DNA-damaging agents, and so a significant number of people are being regularly exposed, for therapeutic purposes, to such agents. By establishing, through studies in bacte ria, the mechanism whereby antitumor agents affect DNA, investigators may be able to help design more drugs that are operator (la, lb). A negligible alteration of operator DNA (la) results, on replication, in a muta tion (2a) that prevents further binding of the repressor (Jo). The prophage develops into active phage particles (4a) that are selectively revealed by strain-# indicator bacteria, which are not sensitive to nonmutated phage; phage development independent of RecA-proteia activation reveals direct mutagenesis. DNA damage (lb), on the other hand, leads on replication (2b, 3b) to RecA-dependent indirect mutagenesis, which is detected with strain-# indicator bacteria. The concomitant event, prophage induction (3b, 4b) resulting from activation of RecA, is de tected with indicator bacteria of strain A, not with strain-# bacteria. Mutagenesis can be diag nosed as indirect when it occurs along with prophage induction: both are dependent on RecA. L AP00024527 DENSITY (PARTICLES PER CUBIC CENTIMETER) AP00024528 THE LANCET, SEPTEMBER 14, 1974 629 THE LANCET Does this Chemical Cause Cancer in Man? Every year some thousands of new chemical compounds are synthesised and brought into use in industry, and some of these inevitably escape into the environment as contaminants of food, air, water, and consumer products. In addition, in 1972 the Depart ment of Health received 132 applications to permit clinical trials of new drugs. We need to know whether any of the new chemicals to which we are exposed will cause cancer, for we cannot afford to release another product which in 30 years' time will cause, in England alone, 30,000 deaths a year--the lung-cancer death-rate at the moment. The expectation of life for a man aged 40 has been static since 1948. All advances in medical care and social conditions have been wiped out by the combined effects of increas ing lung cancerand coronary heart-disease. Ignorance, normal commercial promotion, and the willing con sumer led our society 50 years 2go into a pathway of development that was damaging to health, and now we do not know what to do about cigarettes or coronary heart-disease. Are we any better these days at excluding carcinogenic factors from our environment) The conventional test of carcinogenesis involves feeding or skin application or injection of the chemical in doses starting from the maximum tolerable amount for about two years in several species (preferably including one non-rodent as well as rats and mice). The procedure is so expensive and takes so long that we do not have the resources and skilled manpower to test every new chemical. For every test, not only do hundreds of animals have to be fed and looked after, but clinical, chemical, and post-mortem examinations must also be carried out, together with histological examination of many tissues from each animal and its controls. In addition, the conventional procedure is under criticism, because the diet, strain of animal, and housing conditions affect the result. Should one use animals that are very susceptible and get spon taneous tumours, with the risk of false positives ? Should one " pair-feed " theanimals in order to prevent the high tumour incidence in animals that over-cat ? Roe and Tucker 1 suggest that we cannot ignore the fact that animals under test live under sex-free no-exercise conditions, with unknown virus con tamination. The refinement of animal test procedures is one suggested way out of these dilemmas. Another suggestion is that we now know enough about chemical carcinogenesis to do rapid tests employing tissue-culture and bacterial systems. * An attack on v.v. cellular d.njl (which is associated with carcino genesis) is recognised by effects on cell trans formation s or by the production pf mutations in bacteria.* The advantage of such systems is that they are quick and simple enough to be applied to all new chemicals as they enter the environment. The disadvantage is that they commonly give falsepositive results with chemicals which have been safely used for many years, and in some instances they give false-negative results as well. The same criticisms apply to animal tests. For instance, Thorpe 4 and others have reported an increased incidence of liver tumours in mice fed on phenobarbitone. But an excellent study in human beings who had taken phenobarbitone (and also other anticonvulsants) for many years shows, not only that there is no increase of liver cancer, but that there is an overall deficit of cancer of about a third in the population taking anticonvulsants (brain tumours excepted).* Clemmesen et al., who did this work, attribute the deficit to life in the epilepsy colony, but the metabolic effects of phenobarbitone are at least as likely an explanation. It is clear enough that inducers of microsomal-enzyme synthesis, such as phenobarbitone and D.D.T., do cause liver tumours in certain strains of mice on certain diets. But phenobarbitone does not cause tumours in rats or in man. Could these inducers cause increased meta bolism of endogenous or dietary chemicals, perhaps steroids, which lead to rumours in mice while in man and rats the same enzymes lead to an overall detoxi cation effect ? The predominant effect may well depend on the exact chemical environment. The common-sense answer seems to be to use a selection of the quick methods to screen all new compounds entering the human environment. Those that give a " positive " in the screen and those that arc suspect by reason of chemical structure, those that are to be widely used, and those where industrial exposure is likely to brc heavy, should also be tested in animals before use. In this way we can expect to detect major carcinogens (i.e., those that cause a high proportion of cancers in many species, like the nitrosamincs, aflatoxins, and benzo-a-pyrene). But1 2 3 4 5 1. Roe, F. J. C., Tucker, M. J. Prec. Eurep. Soe. Study Drug Tex. 1973, !5, 171. (lixeerpta mtd. int. Certgr. Sir. 311.) 2. Majtec, P, N\ *Vjtun, 1974, 245, 759. 3. Freeman, A. ii.* Wei^burrtcr, E. K., Weiiburjcr, J, H., VCuIfcrd, R. G., Maryjak, J. M., Huebncr, R. J. J. notn. Comer Jmt. 1973, ft, 799. 4. Thorpe, F. FJCtrmet. To*. 1973. 11. 4JX 5. Clcmaaen,J., ltdcnkon* V, f., Plata, C. M Lanett, 1974, i, 70J. ( i j i AP00024529 <530 THE LANCET, SEPTEMBER 14, 1974 after that we have to depend on long-term follow up of human beings exposed to the hazard. We can expect to detect a carcinogen that produces cancers in 1 in 20 heavily exposed animals. But what- if it prdduccs a cancer in 1 in 50,000 persons who are exposed to a low concentration of the chemical (chloramphenicol agranulocytosis seems to be a risk of this order of magnitude) ? The genetic variability of response in the human population is likely to be fat wider than can possibly be tested in animals. We badly need to use the records of men exposed to chemicais at work. We need to follow the cancer history of patients in relation to drug intake. If cigarettes were today being introduced for the first time, it seems doubtful that we would detect the major carcinogenic hazard. For that we should have to depend on follow-up of smokers and nonsmokers. We cannot have new products without risk, but it is irresponsible to permit new products without assessment of their effects. In devising a monitoring system, we should work back to trying to find out what existing factors in our environment cause the 50% of cancers that are thought to be due to environ mental chemicals.* To avoid the sort of mistakes that permitted widespread use of cigarettes 50 years ago, we need record linkage. How can we have record linkage without sacrifice of privacy and its attendant political dangers ? Perhaps we should start thinking of voluntary record linkage. Could industrial exposure, family food and shopping habits, medica tion, and patterns of illness be followed at least in a volunteer population ? Doll persuaded doctors to give information about smoking habits which enabled him to sort out the effect on lung cancer of smoking and stopping smoking. The Medical Research Council Epidemiology Unit has studies on these lines--for instance, to test the dietary-fibre hypothesis of intestinal disease. We need more epidemiology linked with laboratory facilities. More mice will not solve the problem. Something Rotten in Scotland, and Elsewhere Six years ago a socio-dental study suggested that, in knowledge of oral health, the intelligent professional Englishman scored rather lower than the nine-yearold Californian schoolgirl.67 *A9 survey in England and Wales showed widespread neglect of teeth *; and now follows a report from Scotland * showing an even 6. Doll, R-, Vodupila. I. (editors). Hose Environment Interactions in the Etiology of Cancer in Man. W.H.O./IntcmationaJ Agency for Research on Cancer. Lyon, 1973. 7. Butman, S., Richards* N. D., Slack, G. L., XTilTcocki, A. J. Demand end Seed for Dental Care: a Socio*dtntal Study. London, 196S. t. Gray, P. G., Todd, J. E., Slack, G. L. Butman, J. S. Adult Dental Health in England and Wales in 1963. K.M. Stationery Office, 1969. 9. Adult Dental Health in Scotland. By J. E. TOGO and A. WHtl* worth. H.M_ Stationery Office, 1974. 2.95. worse state of affairs. 44% of Scots over the age of sixteen have lost all their natural teeth; over half go to the dentist only when their teeth hurt; and half again, if they had an aching tooth, would opt for extraction rather than filling. Almost 1 in 5 of dentate adults said they did not usually clean their teeth as often as once a day. Findings like these ought to tell us something about dental priorities for the National Health Service, and there are three obvious ones. Firstly, prevention of the massive amount of untreated dcptal decay can be brought about only bv water fluoridation: con tinuation of the existing system is unlikely to produce substantial changes in community dental health; people will go on losing teeth, and they will go on having untreated decayed teeth. Fluoridation in the major population centres would reduce the incidence of new decay in all sections of the com munity. Secondly, " we need more dental-health education ". This, is a cliche: what is needed is education based on sound principles rather than endless unevaluated repetition of old (and now disproved) saws about apples and carrots. Such education should be incorporated in school curricula. Thirdly, at a time wheD dental science is advancing rapidly, the dental profession must be kept up to date by postgraduate education. The main dental diseases, decay and periodontal disease, ate now well enough understood to be completely prevented (in cooperative individuals), yet practically no-one in the.Scottish survey was free of disease: perhaps most telling of all was the observation that little difference existed in mean tooth loss and in perio dontal disease between regular and irregular atten dee. This indicates that people who go regularly to the dentist are still losing teeth. What sort of care are they getting ? The above are measures which can be taken within the existing system; but they do not go to the heart of the problem. Perhaps more than any other branch ofthe National Health Service, dentistry is in a critical state, and its predicament stems partly from the system of payment: dentists are paid by item of ser vice rather than by capitation fee, as general medical practitioners are. This method causes grave problems in inflationary times. For example, the fee for a full upper and lower deptutc, at 16, has not changed substantially for several years, whereas in some cases the laboratory fee now exceeds 16, so that the dentist actually makes a loss. Small wonder that there is how hardly a dentist in the West End of London who docs any N.H.S. work at all. This discrepancy could doubtless be remedied; but the chief defect in the itcm-of-scrvice system is that it emphasises episodic rather than continuous care. The National Health Service gives no encouragement at all to the dentist and patient who, between them, wish to prevent disease and preserve teeth. APOOC 'the lancet, September 14, 1974 629 THE LANCET Does this Chemical Cause Cancer in Man? Every year some thousands of new chemical compounds arc synthesised and brought into use in industry, and some of these inevitably escape into the environment as contaminants of food, air, water, and consumer products. In addition, in 1972 the Depart ment of Health received 132 applications to permit clinical trials of new drugs. We need to know whether any of the new chemicals to which we are exposed will cause cancer, for we cannot afford to release another product which in 30 years' time will cause, in England alone, 30,000 deaths a year--the lung-cancer death-rate at the moment. The expectation of life for a man aged 40 has been static since 194S. All advances in medical core and social conditions have been wiped our by the combined effects of increas inglung cancer and coronary heart-disease. Ignorance, normal commercial promotion, and the willing con sumer led our society 50 years ago into a pathway of development that was damaging to health, and now we do not know what to do about cigarettes or coronary heart-disease. Are we any better these days at excluding carcinogenic factors from our environment ? The conventional test of carcinogenesis involves feeding or skin application or injection of the chemical in doses starting from the maximum toierable amount for about two years in several species (preferably including one non-rodent as well as rats and mice). The procedure is so expensive and takes so long that we do not have the resources and skilled manpower to test every new chemical. For every test, not only do hundreds of animals have to be fed and looked alter, but clinical, chemical, and post-mortem examinations must also be carried out, together with histological examination of many tissues from each animal and its controls. In addition, the conventional procedure is under criticism, because the diet, strain of animal, and housing conditions affect the result. Should one use animals that arc very susceptible and get spon taneous tumours, with the risk of false positives ? Should one'1 pair-feed ''theanimalsinorderroprcvcnt the high tumour incidence in animals that over-cat ? Roe and Tucker1 suggest that we cannot ignore the fact that animals under test live under sex-free no-exercise conditions, with unknown virus con tamination. The refinement of animal test procedures is one suggested way out of these dilemmas. Another suggestion is that we now know enough about chemical carcinogenesis to do rapid tests employing tissue-culture and bacterial systems.2 3A4n5 attack on VS cellular d.n.a. (which is associated with carcino genesis) is recognised by effects on cell trans formation* or by the production ef mutations in'' bacteria.1 The advantage of such systems is that they are quick and simple enough to be applied to all new chemicals as they enter the environment. The disadvantage is that they commonly give false positive results with chemicals which have been safely used for many years, and in some instances they give false-negative results as wclL The same criticisms apply to animal tests. For instance, Thorpe * and others have reported an increased incidence of liver tumours in mice fed on phenobarbitone. But an excellent study in human beings who had taken phenobarbitone (and also other anticonvulsants) for many years shows, not only that there is no increase of liver cancer, but that there is an overall deficit of cancer of about a third in the population taking anticonvulsants (brain tumours excepted).* Clemmesen et al., who did this work, attribute the deficit to life in the epilepsy colony, but the metabolic effects of phenobarbitone arc at least, as likely an explanation. It is dear enough that inducers of microsomal-enzyme synthesis, such as phenobarbitone and D.D.T., do cause liver tumours in certain strains, of mice on certain diets. But phenobarbitone does not cause tumours in rats or in man. Could these inducers cause increased meta bolism of endogenous or dietary chemicals, perhaps steroids, which lead to tumours in mice while in man and rats the same enzymes lead to an overall detoxi cation effect ? The predominant effect may well depend on the exact chemical environment. The common-sense answer seems to be to use a selection of the quick methods to screen all new compounds entering the human environment. Those that give a " positive " in the screen and those that are suspect by reason ofchemical structure, those that are to be widely used, and those where industrial exposure is likely to be heavy, should also be tested in animals before use. In this way we can expea to detea major carcinogens (i.c., those that cause a high proportion of cancers in many species, like the nitrosamincs, aflntoxins, and benzo-a-pyrene). But I. Roe. F. J. C,, Tucker. M. J. Free. Eurtp. Stc. Sluij Ornr Jx. 1973, 15, 171. (ilxcirpta mtJ. int. Congr. Srr, 311.) 2* Magee* P. N. 1974, 249, 759. 3. Freeman, A. Wcnbutgcr, E. K.* Vciaburjer, J, H. VCoIford, R. G., Maryjak* J* M,, Hucbncr, K. J. J. nam. Conor Jntt. 1973, 61, 799. 4. Thorpe, F. W Conntf. To*. 1973, t (, 03. 5. Gemmcacn,J., I'reJcnkson, V. F., Plum, C. M. 1974,1,705. AP00024531 630 THE LANCET, SEPTEMBER 14, 1974 after that we have to depend on long-term follow up of human beings exposed to the hazard. We can expect' to detect a carcinogen that produces cancers in 1 in 20 heavily exposed animals. But what if it prdduccs a cancer in 1 in 50,000 persons who are exposed to a low concentration of the chemical (chloramphenicol agranulocytosis seems to be a risk of this order of magnitude) ? The genetic variability of response in the human population is likely to be far wider than can possibly be tested in animals. We badly need to use the records of men exposed to chemicals at work. We need to follow the cancer history of patients in relation to drug intake. If cigarettes were today being introduced for the first time, it seems doubtful that we would detect the major carcinogenic hazard. For that we should have to depend on follow-up of smokers and nonsmokers. We cannot have new products without risk, but it is irresponsible to permit new products without assessment of their effects. In devising a monitoring system, we should work back to trying to find out what existing factors in our environment cause the 50% of cancers that are thought to be due to environ mental chemicals. To avoid the sort of mistakes that permitted widespread use of cigarettes 50 years ago. We need record linkage. How can we have record linkage without sacrifice of privacy and its attendant political dangers ? Perhaps we should start thinking of voluntary record linkage. Could industrial exposure, family food and shopping habits, medica tion, and patterns of illness be followed at least in a volunteer population ? Doll persuaded doctors to give information about smoking habits which enabled him to sort out the effect on lung cancer of smoking and stopping smoking. The Medical Research Council Epidemiology Unit has studies on these lines--for instance, to test the dietary-fibre hypothesis of intestinal disease. ,We need more epidemiology linked with laboratory facilities. Mote mice will not solve the problem. Something Rotten in Scotland, and Elsewhere Six years ago a socio-dental study suggested that, in knowledge of oral health, the intelligent professional Englishman scored rather lower than the nine-yearold Californian schoolgirl.67 *A9 survey in England and Wales showed widespread neglect of teeth *; and now follows a report from Scotland * showing an even 6. Doll, R., Vodopiia, 1. (editors). Host Environment Interactions in the Etiology ot* Cancer in Man. V7.K.O./lnternationaJ Agency for Research on Cancer. Lyon, 2973. 7. Bui man, J. S., Richards', N. D., Slack, G. L., ViHeoeks, A. J. Demand and Need for Dental Care: a Socio-dental Study. London, 1963. S. Gray, P. ., Todd, J. E,, Slack, G. L., Butman, J. S. Adult Dental Health in England and Wales in 1968. H.M. Stationery Office, 2969, 9. Adult Dental Health in Scotland. By J. E. Todd and A. WHIT WORTH. H.M. Stationery Office, 1974. 2.99. worse state of affairs. 44% of Scots over the age of sixteen have lost all their natural teeth; over half go to the dentist only when their teeth hurt; and half again, if they had an aching tooth, would opt for extraction rather than filling. Almost 1 in 5 of dentate adults said they did not usually dean their teeth as often as once a day. Findings like these ought to tell us something about dental priorities for the National Health Service, and there arc three obvious ones. Firstly, prevention of the massive amount of untreated dcptal decay can be brought about only by water fluoridation: con tinuation of the existing system is unlikely to produce substantial changes m community dental health; people will go on losing teeth, and they will go on having untreated decayed teeth. Fluoridation in the major population centres would reduce the incidence of new decay in all sections of the com munity. Secondly, " we need more dental-health education ". This' is a cliche: what is needed is education based on sound principles father than endless unevaiuated repetition of old (and now disproved) saws about apples and carrots. Such education should be incorporated in school curricula. Thirdly, at a time when dental science is advancing rapidly, the dejital profession must be kept up to date by postgraduate education. The main dental diseases, decay and periodontal disease, are now well enough understood to be completely prevented (in cooperative individuals), yet practically no-one in the Scottish survey was free of disease: perhaps most telling of all was the observation that little difference existed in mean tooth loss and in perio dontal disease between regular and irregular atteaders. This indicates that people who go regularly to the dentist arc still losing teeth. What sort of cafe are they getting ? The above are measures which can be taken within rite existing system; but they do not go to the heart of the problem. Perhaps more than any other branch ofthe National Health Service, dentistry is in a critical state, and its predicament stems partly from the system of payment: dentists are paid by item of ser vice rather than by capitation fee, as general medical practitioners are. This method causes grave problems in inflationary times. For example, the fee for a full upper and lower deptum, at 16, has not changed substantially for several years, whereas in some cases the laboratory fee now exceeds 16, so that the dentist actually makes a loss. Small wonder that there is now hardly a dentist in the West End of London who docs any N.H.S. work at all. This discrepancy could doubtless be remedied; but the chief defect in the itcm-of-servicc system is that it emphasises episodic rather than continuous care. The National Health Service gives no encouragement at all to the demist and patient who, between them, wish to prevent disease and preserve teeth. APOOC * THH LANCET, SEPTEMBER 14, 1974 629 THE LANCET Does this Chemical Cause Cancer in Man? Every year some thousands of new chemical compounds arc synthesised and brought into use in industry, and some of these inevitably escape into the environment as contaminants of food, air, water, and consumer products. In addition, in 1972 the Depart ment of Health received 132 applications to permit clinical trials ofnew drugs. We need to know whether any of the new chemicals to which we are exposed will cause cancer, for we cannot afford to release another product which in 30 years' rime will cause, in England alone, 30,000 deaths a year--the lung-cancer death-rate at the moment. The expectation of life for a man aged 40 has been static since 194S. All advances in medical care and social conditions have been wiped out by the combined effects of increas inglung cancerand coronaryheart-disease. Ignorance, normal commercial promotion, and the willing con sumer led our society 50 years ago into a pathway of development that was damaging to health, and now we do not know what to do about cigarettes or coronary heart-disease. Are we any better these days at excluding carcinogenic factors from our environment ? The conventional test of carcinogenesis involves feeding or skin application or injection of the chemical in doses starting from the maximum tolerable amount for about two years in several species (preferably Including one non-rodent as well as rats and mice). The procedure is so expensive and takes so long that we do not have the resources and skilled manpower to test every new chemical. For every test, not only do hundreds of animals have to be fed and looked after, but clinical, chemical, and post-mortem examinations must also be carried out, together with histological examination of many tissues from each animal and Its controls. In addition, the conventional procedure is under criticism, because the diet, strain of animal, and housing conditions affect the result. Should one use animals that are very susceptible and get spon taneous tumours, with the risk of false positives ? Shouldone" pair-feed "the animals in orderto prevent the high tumour incidence in animals that over-cat ? Roe and Tucker 1 suggest that we cannot ignore the fact that animals under test live under sex-free no-exerdse conditions, with unknown virus con tamination. The refinement of animal test procedures is one suggested way out of these dilemmas. Another suggestion is that we now know enough about chemical carcinogenesis to do rapid tests employing tissue-culture and bacterial systems.* An attack on VV cellular D.N.A. (which is associated with carcino genesis) is recognised by effects on cell trans-^ formation * or by the production pf mutations in ^ bacteria.' The advantage of such systems is that they are quick and simple enough to be applied to all new chemicals as they enter the environment. The disadvantage is that they commonly give false positive results with chemicals which have been safely used for many years, and in some instances they give false-negative results as well. ' The same criticisms apply to animal tests. For instance, Thorpe4 and others have reported an increased incidence of liver tumours in mice fed on phenobarbitone. But an excellent study in human beings who had taken phenobarbitone (and also other anticonvulsants) for many years shows, not only that there is no increase of liver cancer, but that there is an overall deficit of cancer of about a third in the population taking anticonvulsants (brain tumours excepted).5 Clemmesen et al., who did this work, attribute the deficit to life in the epilepsy colony, but the metabolic effects of phenobarbitone arc at least as likely an explanation. It is clear enough that inducers of microsomal-enzyme synthesis, such as' phenobarbitone and D.D.T., do cause liver tumours in certain strains of mice on certain diets. But phenobarbitone does not cause tumours in rats or in man. Could these inducers cause increased meta bolism of endogenous or dietary chemicals, perhaps steroids, which lead to tumours in mice while in man and rats the same enzymes lead to an overall detoxi cation effect ? The predominant effect may well depend on the exact chemical environment. The common-sense answer seems to be to use a selection of the quick methods to screen all new compounds entering the human environment. Those that give a " positive " in the screen and those that are suspect by reason of chemical structure, those that axe to be widely used, and those where industrial exposure is likely to b'e heavy, should also be tested in animals before use. In this way we can expect to detect major carcinogens (i.e., those that cause a high proportion of cancers in many species, like the nitrosamincs, aflatoxins, and bcuzo-a-pyrene). But 1. Rot. F. 1. C.t Tucker, M. J. Prut. Eurcp. Set. Study Druy Tux. 1973, IS, 171. (Cxtcrpla mud. ini. Ctitjr. See. 11U 2. Mu. P. N\ 1074, 249. 750. J. Freeman, A. Ii., \i euburwr. li. K... Vciiburgcr, J. It., Wolford, R. G., iUnjj'r. J. M,, Huebner, R. 17. nun. Center Imt. 1973, SI, 799. 4. Thorpe, F. Fd Coma. Tux. 1975, 11, 433. 5. Clcmmcscn, J., rrciicrikson, V. F., Tiuni, C. AL oowi, 1974, i, 70S. I I I t I t I AP00024533 * 630 THE LANCET, SEPTEMBER 14, 1974 after that we hive to depend on long-term follow worse state of affairs. 44% of Scots over the age of up of human beings exposed to the hazard. We can sixteen have lost all their natural teeth; over half go expea'to detect a carcinogen that produces cancers to the dentist only when their teeth hurt; and half in 1 in 20 heavily exposed animals. But what if it again, if they had an aching tooth, would opt for prdduccs a cancer in 1 in 50,000 persons who are extraction rather than filling. Almost 1 in 5 of exposed to a low concentration of the chemical dentate adults said they did not usually clean their (chloramphenicol agranulocytosis seems to be a teeth as often as once a day. risk of' this order of magnitude) ? The genetic variability of response in the human population is likely to be far wider than can possibly be tested in animals. We badly need to use the records of men exposed to chcmicais at work. We need to follow the cancer history of patients in relation to drug intake. If cigarettes were today being introduced for the first time, it seems doubtful that we would detect the major carcinogenic hazard. For that we should have to depend on follow-up of smokers and-non- smokers. We cannot have new products without risk, but it is irresponsible to permit new products without assessment of their effects. In devising a monitoring system, we should work back to trying to find out what existing factors in our environment cause the 50% of cancers that are thought to be due to environ mental chemicals.* To avoid the Sort of mistakes that permitted widespread use of cigarettes 50 years ago, we need record linkage. How can we have record linkage v/ithout sacrifice of privacy and its attendant political dangers ? Perhaps we should start thinking of voluntary record linkage. Could industrial exposure, family food and shopping habits, medica tion, and patterns of illness be followed at least in a volunteer population ? Doll persuaded doctors to give information about smoking habits which enabled him 'to sort out the effect on lung cancer of smoking and stopping smoking. The Medical Research Council Epidemiology Unit has studies on these lines--for instance, to test the dietary-fibre hypothesis of intestinal disease. Wc need more epidemiology Findings like these ought to tell us something about dental priorities for the National Health Service, and there are three obvious ones. Firstly, prevention of the massive amount of untreated dental decay can be brought about only by water fluoridation: con tinuation of the existing system is unlikely to produce substantial changes in community dental health; people will go on losing teeth, and they will go on having untreated decayed teeth. Fluoridation in the major population centres would reduce the incidence of new decay in all sections of the com munity. Secondly, " we need more dental-health education". This is a cliche: what is needed is education based on sound principles rather than endless unevaluated repetition of old (and now disproved) saws about, apples and. carrots. Such education should be incorporated in school curricula. Thirdly, at a time wheo dental science is advancing rapidly, the dcptal profession must be kept up to date by postgraduate education. The main dental diseases, decay and periodontal disease, are now well enough understood to be completely prevented (in cooperative individuals), yet practically no-one in the Scottish survey was free of disease: perhaps most telling of all was the observation that little difference existed in mean tooth loss and in perio dontal disease between regular and irregular atten dee. This indicates that people who go regularly to the dentist are still losing teeth. What sort of care are they getting ? linked with laboratory facilities. More mice will not The above are measures which can be taken within . solve the problem. the existing system; but they do not go to the heart of the problem. Perhaps more than any other branch ofthe National Health Service, dentistry is in a critical Something Rotten in Scotland, and Elsewhere state, and its predicament stems partly from the system of payment: dentists are paid by item of ser vice rather than by capitation fee, as general medical Six years ago a socio-dental study suggested that, in knowledge of oral health, the intelligent professional practitioners are. This method causes grave problems in infiationary times. For example, the fee for a full Englishman scored rather lower than the nine-yearold Californian schoolgirl.* A survey in England and upper and lower dejjtup:, at 16, has nor changed substantially for several years, whereas in some eases Wales showed widespread neglect of teeth j and now the laboratory fee now exceeds 16, so that the follows a report from Scotland* showing an even* 7 * d9entist actually makes a loss. Small wonder that there Dell, R., Vodcpija, X. (editor*). Host Environment Interactions in the Edolopy of Cancer in Man. W.H.O,/International Agency for Research oa Cancer. Lyon* 1973. 7. Bulnun, J. S-, Richards; K. D., Slack, G. L., TPitlcocks, A. J. Demindand Need for Dental Care: a Socio-dental Study. London, 1969. Gray, P. G., Todd, J. E., Slack, G. L., Buiman, J. S. Adult Dental Health in Enclind and Wales in 1968. H.M. Stationery Office. 19*9. 9. Adult Dental Health in Scotland. By J. E. TODD and A. Whit* worth. HJ4. Stationery Office, 1974. 2.95. is now hardly a dentist in the West End of London who does any N.H.S. work at all. This discrepancy could doubtless be remedied; but the chief defect in the itcm-of-servicc system is that it emphasises episodic rather than continuous care. The National Health Service gives no encouragement at all to the demist and patient who, between them, wish to prevent disease and preserve teeth. > AP00024534 f' za if 1 'or wi- ly ity In ng >se ay ns of : ja, ify ip- ti- im '*> .Is, il- ig rte id n: g Ic id Hi ly ic V id le id SO d n .d ic si *s s' e > B .> 7*,f.^.Vi :t- THE tANCEt, SEPTEMBER 14, 1974 629 THE LANCET Does this Chemical Cause Cancer in Man? Every year some thousands of new chemical compounds are synthesised and brought into use in industry, and some of these inevitably escape into the environment as contaminants of food, air, water, and consumer products. In addition, in 1972 the Depart ment of Health received 132 applications to permit clinical trials ofnew drugs. We need to know whether any of the new chemicals to which we are exposed will cause cancer, for we cannot afford to release another product which in 30 years' time will cause,in England alone, 30,000 deaths a year--the lung-cancer death-rate at the moment. The expectation of life for a man aged 40 has been static since 1948. All advances in medical care and social conditions have been wiped out by the combined effects of increas inglungcancerand coronary heart-disease. Ignorance, normal commercial promotion, and the willing con sumer led our society 50 years ago into a pathway of development that was damaging to health, and now we do not know what to do about cigarettes or coronary heart-disease. Are we any better these days at excluding carcinogenic factors from our environment ? The conventional test of carcinogenesis involves feeding or skin application or injection of the chemical in doses starting ffcom the maximum tolerable amount for about two years in several species (preferably including one non-rodent as well as rats and mice). The procedure is so expensive and takes so long that we do not have the resources and skilled manpower to test every new chemical. For every test, not only do hundreds of animals have to be fed and looked after, but clinical, chemical, and post-mortem examinations must also be carried out, together with histological examination of many tissues from each animal and its controls. In addition, the conventional procedure is under criticism, bemuse the diet, strain of animal, and housing conditions affect the result. Should one use animals that are very susceptible and get spon taneous tumours, with the risk of false positives ? Should one " pair-feed " theanimals inorder to prevent the high tumour incidence in animals that over-cat ? Roe and Tucker 1 suggest that we cannot ignore the fact that animals under test live under sex-free no-exercise conditions, with unknown virus con tamination. The refinement of animal test procedures is one suggested way out of these dilemmas. Another suggestion is that we now know enough about chemical carcinogenesis to do rapid tests employing tissue-culture .and bacterial systems.1 2 A3n attack on cellular d.njl (which is associated with carcino genesis) is recognised by effects on cell transformation * or by the production of mutations in bacteria.1 The advantage of such Systems is that they are quick and simple enough to be applied to all new chemicals as they enter the environment. The disadvantage is that they commonly give false positive results with chemicals which have been safely used for many years, and in some instances they give false-negative results as well. The same criticisms apply to animal tests. For instance, Thorpe4 5and others have reported an increased incidence of liver tumours in mice fed on phenobarbitone. But an excellent study in human beings who bad taken phenobarbitone (and also other anticonvulsants) for many years shows, not only that there is no increase of liver cancer, but that there is an overall deficit of cancer of about a third in the population taking anticonvulsants (brain tumours excepted).1 Clemmesen et al., who did this work, attribute the deficit to life in the epilepsy colony, but the metabolic effects of phenobarbitone are at least as likely an explanation. It is clear enough that inducers of microsomal-enzyme synthesis, such as ' phenobarbitone and d.d.t., do cause liver tumours in iw nm strains of mice on certain diets. But phenobarbitone does not cause tumours in rats or m man. Could these inducers cause increased meta bolism of endogenous or dietary chemicals, perhaps steroids, which lead to tumours in mice while in man and rats the same enzymes lead to an overall detoxi cation effect? The predominant effect may well depend on the exact chemical environment. The common-sense answer seems to be to use a selection of the quick methods to screen all new compounds entering the human environment. Those that give a " positive " in the screen and those that are suspect by reason of chemical structure, those that are to be widely used, and those where industrial exposure is likely to be heavy, should also be tested in animak before use. In this way wc can expect to detect major carcinogens (i.e., those that cause a high proportion of cancers in many species, like the nitrosamincs, aflatoxins, and benzo-a-pyrenc). But 1. Roe, F. J. C,, Tucker. M. J. Peer. Eure*. See. Sniy Brut Tex. 1973, It, 171. (Exarfta mti. inf. Cmfr. Sir. 311.) 2. Magee, P. N. Nano, 1974, 249. 759. 3. Freemen, A. R, Weiiburger, E. K-, Weieburger, J. H-, Wolford, R_ G., Maryjak, J. M., Huebner, R. J.J. nan. Canon but. 1973, SI,799. 4. Thorpe, F. Fd Catmtt. Tax. 1973, 11,433. 5. Cleminefea, J., Frederikson. V. F.,Plum, C. M. Lanut, 1974, i, 70S. I I AP00024535 630 THE LANCET, SEPTEMBER 14, 1974 after that we have to depend on long-term follow* up of human beings exposed to the hazard. We can expea to detea a carcinogen that produces cancers in 1 in 20 heavily exposed animals. But what if it produces a cancer in 1 in 50,000 persons who arc exposed'to a low concentration of the chemical (chloramphenicol agranulocytosis seems to be a risk of this order of magnitude) ? The genetic variability of response in the human population is likely to be far wider than can possibly be tested in animals. We badly need to use the records of men exposed to chemicals at work. We need to follow the cancer history of patients in relation to drug intake. If cigarettes were today being introduced for the first time, it seems doubtful that we would detox the major carcinogenic hazard. For that we should have to depend on follow-up of smokers and nonsmokeis. We cannot have new produas without risk, but it is irresponsible to permit new products without assessment of their effects. In devising a monitoring system, we should work back to trying to find out what existing factors in our environment cause the 50% of cancers that are thought to be due to environ mental chemicals.* To avoid the sort of mistakes rhar permitted widespread use of cigarettes 50 years ago, we need record linkage. How can we have record linkage without sacrifice of privacy and its attendant political dangers ? Perhaps we should start thinking of voluntary record linkage. Could industrial exposure, family food and shopping habits, medica tion, and patterns of illness be followed at least in a volunteer population? Doll persuaded doaors to give information about smoking habits which enabled him to sort out the effea on lung cancer of smoking and stopping smoking. The Medical Research Council Epidemiology Unit has studies on these lines--for instance, to test the dietary-fibre hypothesis of intestinal disease. We need more epidemiology linked with laboratory facilities. More mice will not solve the problem. Something Rotten in Scotland, and Elsewhere Six years ago a socio-dental study suggested that, in knowledge of oral health, the intelligent professional Englishman scored rather Iowa than the nine-yearold Californian schoolgirl.67 *A9 survey in England and Wales showed widespread neglea of teeth *; and now follows a report from Scotland* showing an even 6. Doll. R., Vodopija, L (aditon). Host Environment Intcxrctions in tbo Etiology of Cancer in Man. WXLO^Intemationol Agency for Research on Cancer. Lyon. 1973. 7. Buinun. J. s, Richards; N. D., Slack, G. L., Willcocks, A. J. Demand and NerdforDenial Cue: a Sodo-dental Study. Loudon, IMS. t. Gray, P. G., Todd,E., Slack, G. I_, Bulman, J. S. Adult Dental Health la Entland and Wales la 1966. H.M. Stationery Office, 1999. ' 9. Adult Dental Health in Scotland. By J. B. Torn and A. Wan- vkwth. HAL Stationery Office, 1974. 2.95. worse state of affairs. 44% of Scots ova the age of sixteen have lost all their natural teeth; ova half go to the dentist only when their teeth hurt; and half again, if they had an aching tooth, would opt for extraction rather than filling. Almost 1 in 5 of dentate adults said they did not usually dean their teeth as often as once a day. Findings like these ought to tell us something about dental priorities for the National Health Service, and there are three obvious ones. Firstly, prevention of the massive amount of untreated dental decay can be brought about only by water fluoridation: con tinuation of the existing system is unlikely to produce substantial changes in community dental health; people will go on losing teeth, and they will go on having untreated decayed teeth. Fluoridation in the major population centres would reduce the incidence of new decay in all sections of the com munity. Secondly, " we need more dental-health education". This is a cliche: what is needed is education based on sound principles ratha than endless unevaluated repetition of old (and now disproved) saws about apples and carrots. Such education should be incorporated in school curricula. Thirdly, at a time when dental science is advancing rapidly, the dental profession must be kept up to date by postgraduate education. The main dental diseases, decay and periodontal disease, are now well enough understood to be completely prevented (in cooperative individuals), yet practically no-one in the Scottish survey was free of disease: pahaps most telling of all was the observation that little difference existed in mean tooth loss and in perio dontal disease between regular and irregular attenders. This indicates that people who go regularly to the dentist ate still losing teeth. What sort of cate are they getting ? The above are measures which can be taken within the existing system; but they do not go to the heart of the problem. Pahaps more than any other branch ofthe National Health Service, dentistry is in a critical state, and its predicament stems partly from the system of payment: dentists are paid by item of ser vice rather than by capitation fee, as general medical practitioners arc. This method causes grave problems in inflationary times. For example, the fee for a full upper and Iowa denture, at 16, has not changed substantially for several years, whereas in some cases the laboratory .fee now exceeds 16, so that the dentist actually makes a loss. Small wonda that there is now hardly a dentist in the West End of London who does any N.H.S. work at all. This discrepancy could doubtless be remedied; but the chief defea in the item-of-service system is that it emphasises episodic ratha titan continuous care. The National Health Service gives so encouragement at all to the dentist and patient who, between them, wish to prevent disease and preserve teeth. | AP00024536 mm comavirus RNA synthesis was indepen dent of DNA synthesis (5). Furthermore, studies with actinomycin D had shown that neither synthesis nor expression of Viruses, Polymerases, and Cancer cellular DNA was involved in viral RNA synthesis (6). These results suggested that Mengovirus might make a cy ' David Baltimore toplasmic RNA-dependent RNA syn thesis system. The concept that viruses induce synthesis of their own enzymes had strong precedents in bacteriophage systems--Seymour Cohen's work had The study of biology is partly an exer infection with paralysis while the other shown decisively that new virus-speci cise in natural esthetics. We derive causes the common cold. One class of fied enzymes were found in infected bac much of our pleasure as biologists from RNA viruses, however, does have a teria (7). . the continuing realization of how eco unique symptom associated with it: the We approached the problem of the nomical, elegant, and intelligent are the biochemically defined group of viruses virus' effect on intracellular RNA syn accidents of evolution that have been called the retroviruses are the only RNA thesis as a question in enzymology. Wc maintained by selection. A virologist is viruses known to cause cancer. For a first showed that the nuclei from Men- among the luckiest of biologists because virologist interested in cancer, the prob govirus-infected cells were greatly re he can see into his chosen pet down to lem is first to understand the molecular duced in their ability to carry out cell- the details of all of its molecules. The biology of retroviruses and then to under free DNA-dependent RNA synthesis virologist sees how an extreme parasite stand how they cause the disease. compared to nuclei of uninfected cells functions using just the most fundamen In what follows, I will review my per (8) . Later, we showed that cytoplasmic tal aspects of biological behavior. sonal involvement in uncovering the dif extracts of Mengovirus- or poliovirus- A virus is a form of life with very ferent genetic systems of RNA viruses, a infected cells contained an RNA syn simple requirements (/). The basic needs story which leads to the recognition of thesis activity not evident in uninfected of a virus are a nucleic acid to be trans the unique style of retroviruses. I will cells and not inhibited by actinomycin D mitted from generation to generation (the then consider what is known about the (9) . When we learned that the system genome) and a messenger RNA to direct relationship between the biochemistry of made RNA of the size and structure of the synthesis of viral proteins. The criti retroviruses and their ability to be onco virion RNA (10). it became clear that it \ cal viral proteins that the messenger genic. represented the postulated viral RNA- RNA must encode are those that coat the As I tell my story I will mention a few dependent RNA synthesis system. genome and those that help replicate the of the many co-workers, teachers, and While there has been extensive further genome. One of the great surprises of students who have influenced my think analysis of crude cytoplasmic systems modem virology has been the discovery ing or contributed their labors and ideas (II. 12) and impressive enrichment of the of the variety of genetic systems that to the products of my laboratory. To RNA synthesis system has been viruses have evolved to satisfy their mention all of the people to whom I am achieved (13). no pure enzyme able to needs. Among the animal viruses, at indebted would make too long a list; I make picomavirus RNA has ever been least six totally different solutions to the can only say that the honors I receive are isolated, so the detailed mechanism of basic requirements of a virus have been in large measure testaments to their ac viral RNA synthesis still remains ob found Q). complishments. scure. Most of our knowledge of how If we look back to virology books of 15 picomavirus RNA is made has come years ago, we find no appreciation yet from studies on the virus-specific RNA for the variety of viral genetic systems Picomariruses molecules in infected cells and their ki used by RNA viruses (J). Since then, the netics of labeling by radioactive pre various systems have come into focus, My work on the genetic systems of cursors. Such research has been carried the last to be recognized being that of the RNA viruses dates back to my graduate out in many laboratories (II, I2)\ my retroviruses (RNA tumor viruses). As school days. As part of my introduction work in this area was done in association each new genetic system was discov to animal virology during a Cold Spring with James Darnell and Marc Girard. ered, it was often the identification of an Harbor course, I heard Richard Franklin Together we found and studied the rela RNA or DNA polymerase that could be describe his then-recent experiments us tions between the poliovirus double- responsible for the synthesis of virus- ing autoradiography to show that Men- stranded RNA, the poliovirus replicative specific nucleic acids that gave the most govirus, a picomavirus and a close rela intermediate, and the poliovirus replica convincing evidence for the existence of tive of poliovirus, could shut off the nu tion complex (14). Girard's precise in the new system. clear synthesis of cellular RNA early I Now that. the life-styles of different after infection and could later induce CopyrightO 197(5 by the Nobel Foundation. types of viruses have been delineated we new RNA synthesis in the cytoplasm The author is professor of microbiology at the Massachusetts Institute of Technology, Cambridge ' can ask what relation there is between a which appeared to represent synthesis of 02139. This article is the lecture he delivered Tn virus' multiplication cycle and the dis viral RNA (4). I decided to go to the Stockholm. Sweden, on 12 December 1975 when he received the Nobel Prize for Physiology or Medicine, ease it causes. In general, this question has no simple answer because disease Rockefeller University as a graduate stu a prize he shared with Renato Dulbecco and Howard M. Tcmin. The article is published here with the dent with Richard Franklin in order to permission of the Nobel Foundation and will also be symptoms do not correlate with the bio chemical class of the virus. For instance, work on the system he had developed. included in the complete volume of Les Prtx Nobel tn 1975 as well as in tbe series Nobel Lecture* (in Before I began to study how picoma English) published by the Elsevier Publishing Com- both poliovirus and rhinovirus are picornavimses, but one causes an intestinal virus RNA was made, it was already pany, Amsterdam and New York. Dulbcdbo's lec ture appeared in the 30 April issue, and Tentin's known from the work of Simon that pi- lecture will appear in a laterissue. }z ____ ____ . ^^BCIENCE. VOL. W2 AP00024537 more controls to abate pollution, or In what form the nondegradation deci Summary whether they lessen the impetus to sion will ultimately manifest itself is diffi change in order to accommodate the stat cult to say. In compliance with the court It is clear that natural resource devel us quo--will substantially affect us all. decision, the EPA has issued regulations opment and industrial expansion must be In discussing the pervasive effect of establishing the manner in which areas of done so as to produce the least impact on the Clean Air Act, 1 have limited myself the country having "cleaner than clean" our physical environment. Our energy to one section of the act and to three air will be preserved (11), but these regu requirements, along with the exigencies requirements only in that section--that lations are presently being challenged in of life in modern society, demand adjust emission limitations, land use controls, the court by the Sierra Club and other ment and balance between pollution con and transportation controls be effec environmental groups on the one hand trol and development. That adjustment tuated within a state to implement am and by the American Petroleum Insti will be made based largely on the politi bient air quality standards established by tute, various oil companies, and the state cal interests reflected in our commu the statute. I have not discussed the of New Mexico on the other hand (12). nities, and where the pendulum will stop provisions of the statute relating to the In their most extreme form, non has yet to be determined. siting of new stationary sources of air degradation principles could operate so pollution--this alone a significant land as to prevent industrial development of 1 it References and Notes use control--nor have I discussed the any consequence in those areas of the 1. United Slates Code, voi. 42, sect. 1857c-1 provisions relating to the control of auto country (mostly the western states) mobile pollution through the devel where the air is now superior to that (1970), amending ibid., sect. 1857c (Suppl. 1975). 2. __ 1857bet stq. (1970), amend- opment of improved internal combustion required by the national ambient air qual ingftirf..sect. 1857 et seq. (Suppl. 1975). 3. Fed. Reporter 2nd 508.743 (3rd circuit, 1975). engines. The discussion of these provi ity standards; new towns and cities in sions would only reinforce the point I such areas might also be difficult to estab 4. Ibid. 481,162 (6th circuit, 1974). 5. Indiana and Michigan Electric Co. v. EPA. ibid. 509,839 (7th circuit, 1975); Union Electric have attempted to make, that the Clean lish, as might be recreational areas, Air Act necessarily does more than clean roads, and other activities that would Company v. EPA, ibid. 5IS, 206 (8th circuit, 1975). 6. Stale plans are published in the Federal Regis the air: it relocates industry, it changes draw automobiles. A consequence of centers of population, it alters life-styles thus freezing at their present level of ter; each plan also appears in the Code ofFeder al Regulations. See Code of Federal Regu lations, vol. 40, sect. 52 et sea. (1975). Follow and living patterns, and it touches im development these "cleaner than cleap" ing are two examples of provisions having the effect noted in the text: mediately a broad range of interests, areas might be the confining of industry from the esthetic to the economic. and population' to those areas of the "No later than February 15, 1974, the City of Fairbanks shall prohibit the idling of all motor vehicles withio the City unless attended by a country which are now "dirty air" areas. Thus, we have another example of an licensed driver. . (Alaska's plan, ibid., sect. 52.87). "Beginning June I, 1975, the State of Mary Cleaner than Clean Air environmental statute's impact going far beyond mere land and water. But there is one other requirement of land snail . . . establish a computer aided carpool matching system which is conveniently available at least to all employees of employers within the region who employ 100 or more em ployees" (Maryland's plan, ibid., sect. 52.1104). the Clean Air Act that should be men 7. United States Code, vol. 42, sect. 1857c-!0 and tioned. Although this requirement is diffi Other Land Use Control Authority I857f-6f (1974), amending ibid., sect. 1857 et seq. (1970) (Suppl. 1975). cult to find explicitly set forth in the 8. _ ibid., sect. 1857(bXO (1970) (Suppl. statute, the courts have found it to be in In addition to administering the Clean 1975). 9. Fed. Suppl. 544, 255 (District of Columbia Dis the statute implicitly, and it may have as Air Act, the EPA administers the Feder trict Court, 1972). 10. U.5. Rep. 412,541 (1973). profound an effect as any I have men al Water Pollution Control Act and a 11. Code of Federal Regulations, vol. 40, sect. tioned. The District Court for the Dis substantial number of other statutes. The 52.21(1975). 12. Sierra Club v. EPA, civil No. 74-2063 (District trict of Columbia, in a suit entitled Sierra Federal Water Pollution Control Act af of Columbia Court of Appeals, 1974); Sierra Club v. Ruckelshaus (9). held that in fects matters other than clean water just Club v. EPA, civil No. 74-2079 (District of Co lumbia Court of Appeals. 1974); Public Service those portions of the nation where the air as fully and pervasively as the Clean Air is already of a quality higher than that Act affects matters other than dean air; Company of Colorado v. EPA, civil No. 75-1368 (District of Columbia Court of Appeals, 1975); Utah Power and Light Company v. EPA, civil required by the national ambient air qual one author commenting on one provision ity standards, the air may not be per only of the act, section 208, observed No. 75-1369 (District of Columbia Court of Ap peals, 1975); State of New Mexico ex rei. New Mexico Environmental Improvement Agency v. mitted to deteriorate to the level of the that "it is difficult to imagine how a national ambient air quality standards. successful water pollution control strate EPA, civil No. 75-1370 (District of Columbia Court of Appeals, 1975); Pacific Coal Gasifica tion Company v. EPA, civil No. 75-1371 (Dis This is called the principle of non gy of this magnitude could be effective trict of Columbia Court of Appeals. 1975); Utah International Inc. v. EPA, civil No. 75-1372 deterioration or nondegradation. The without extensive reforms in the state holding of the District Court for the Dis land use guidance system, or apart from (District of Columbia Court of Appeals. 1975); Indiana-Kentucky Electric Corporation v. EPA, civil No. 75-1575 (District of Columbia Court of trict of Columbia was affirmed per curi other conservation, social, and econom am, that is, unanimously, and without an ic objectives'* (13). Clearly, the growth Appeals, 1975): Dayton Power and Light Com- Kv. EPA, civil No. 75-1663 (District of Coa Court of Appeals, 1975); Buckeye Power opinion, by the Court of Appeals for the of almost any metropolitan area could be District of Columbia. The matter was determined through the control of the Inc. v. EPA, civil No. 75-1664 (District of Co lumbia Court of Appeals. 1975); American Petro leum Institute v. EPA, civil No. 75-1665 (Dis then taken to the Supreme Court, where size and location of sewage treatment trict of Columbia Court of Appeals, 1975); Ala bama Power Company v. EPA, civil No. 75-1666 the justices were divided equally, four to plants and restrictions on hookups to four {10), on the question (Justice Powell these plants. It follows then that the (District of Columbia Court of Appeals, 1975); Mootana Power Company v. EPA, civil No. 75- 1763 (District of Columbia Court of Appeals, having taken no part in the consideration agency that administers statutes of such of the case). A consequence of an even comprehensive scope is itself an agency 1975); Salt River Project Agricultural Improve ment A Power District v. EPA, civil No. 75-1764 (District ofColumbia Court ofAppeals, 1975). division by the Supreme Court is the affirmation of the decision which it is whose policies, like the statutes it admin isters, have an impact going far beyond 13. w. K. Reilly, io Federal Environmental Law, E. L. Doten and T. G. T. Guilbert, Eds. (West, St-Eaul, Minn., 1974), p. 1431. reviewing; the nondeterioration prin the area suggested by the name of the 14. I thank Martin Green, special assistant to the a Assistant Attorney General, for his help m pre- ciple, therefore, is today the law. ' agency. * paring the original text ofthese remarks. ggfci': .4 -r ------- 14 MAY 1976 631 - AP00024538 vitro analysis of RNA synthesis capped this whole series of experiments (15). A crucial part of the viral genetic sys tem is the manner of translation of the viral messenger RNA. While working on viral maturation, my first graduate stu dent, Michael Jacobson, and I began to realize that proteolytic cleavage was an important part of the process (16). Our work led us to suggest that the whole 7500-nucleotide viral genome might be translated into a single continuous poly peptide that we have called a polyprotein (17.18). Recently, this work culminated in the demonstration that poliovirus RNA can be translated into this contin uous polyprotein in a cell-free system and that some of the cleavages that make the polyprotein into the functional pro teins appear to occur in extracts of unin fected cells (19). The demonstration that the poliovirus genome RNA is the messenger RNA for the synthesis of viral proteins, coupled with the demonstration of the infectivity of viral RNA (20), implies that the polio virus genetic system is very simple. The existing evidence confirms this sim plicity. As seen diagrammaticaJly in Fig. 1, it appears that the incoming "plus" strand of RNA synthesizes a "minus" strand which in turn synthesizes a series of plus strands. This diagrammatic sim plicity of poliovirus replication hides a fair amount of as yet undeciphered com plexity, as shown by the work of Eckard Wimmer and his colleagues as well as by work in my laboratory. For instance, the 3'-ends of the virion RNA and messen ger RNA arc both polyadenylate [poly(A)] and the 5'-end of the minus strand is polyuridylate [poly(U>], so we assume that they are templates for each other (27). But these homopolymer stretches have very variable lengths even in the progeny of a cloned virus; what then determines their length? The poly(A) serves some obscure but necessary func tion in the life cycle of the virus (22); what is this function? There is no triphos phate 5'-terminus, either free or capped, on the virion RNA or messenger RNA (23. 24)\ how then is the synthesis of these molecules initiated? The 5'-end of the virion RNA and messenger RNA are different (24); what does this mean? Vesicular Stomatitis Virus Most of the work in my laboratory until 1969 centered on poliovirus. We had assumed that all RNA viruses would be similar in their basic molecular biol ogy, but during the 1960*s results emerg ing from various laboratories implied m may im Th Poliovirus Genetic System Fig. 1. Schematic representation of polio virus-specific RNA synthesis in the cytoplasm ofinfected cells. that poliovirus, rather than being a mod el for all RNA viruses, used one out of a collection of different viral genetic sys tems. Probably the first dramatic demon stration of the variety in RNA viruses came from next door to Richard Frank lin's laboratory at the Rockefeller Insti tute, where Peter Gomatos and Igor Tamm found that reovirus has doublcstranded RNA as its genome (25). The peculiarity of reovirus was underscored by the demonstration later that an RNA polymerase in the virion of reovirus is able to assymetrically transcribe the double-stranded RNA (2d). This was the first virion-bound RNA-dependent RNX polymerase ever found and followed af ter the finding of the first nucleic acid polymerase in a virion--the DNA-dependent RNA polymerase found by Kates and McAuslan (27) and Munyon et al. Q8) in virions of vaccinia virus. Another observation that suggested there were profound differences among the RNA viruses was the finding that in cells infected by the paramyxoviruses, Newcastle disease virus, or Sendai vi rus, much of the virus-specific RNA was complementary to the virion RNA (29). This result was in sharp contrast to what was found in poliovirus-infected cells, where most of the virus-specific RNA was of the same polarity as virion RNA an. We branched away from concentra tion only on poliovirus to include the study of vesicular stomatitis virus (VSV) because of the lucky circumstances that brought Alice Huang to my laboratory. --4r, She joined me first at the Salk Institute and then we both came to MIT in 1968. Alice had studied VSV as her graduate work with Robert R. Wagner at Johns Hopkins. We decided that the tech niques developed for studying poliovirus should be applied to VSV, hoping that the peculiar ability of VSV to spawn and then be inhibited by short, defective par ticles could be understood at the molecu lar level. A graduate student, Martha Stampfer, joined in this work and togeth er we found that wc had bitten off an enormous problem because VSV in duced synthesis of so many species of RNA. Only three species of RNA are seen in poliovirus-infected cells, but we found at least nine RNA's in VSV-infected cells, and one of these RNA's was clearly heterogeneous GO)--later work showed it had four components (31). In our description of this work we said that nine RNA species "seems exorbitant" GO) but we soon realized that each RNA had its place in the cycle of growth and growth inhibition ofVSV. As we were beginning to unravel the multiple RNA's of VSV, Schaffer et al. G2) published a paper showing that the major VSV-induced RNA's in infected cells, like those induced by Sendai and Newcastle disease viruses, were com plementary in base sequence to the vi rion RNA. We confirmed and extended that observation, showing that the virusspecific RNA recovered from the poly ribosomes of infected cells (the viral mes senger RNA) was all complementary to virion RNA G-3). This finding presented a pregnant paradox: if these viruses were like poliovirus and induced a new po lymerase in the infected cell, how could a virus that carried as its genome the strand of RNA complementary to mes senger RNA ever start an infection? There seemed two possible solutions: the RNA came into the cells and was copied by a cellular enzyme to make the messenger RNA to initiate the infection cycle, or the RNA came into the cell carrying an RNA polymerase with it. Because no convincing evidence for RNA-to-RNA transcription existed (or exists) for any uninfected cell, the possi bility of a polymerase with the incoming RNA seemed attractive. This possibility implied that there might be polymerase activity demonstrable in disrupted vi rions of VSV. Almost as soon as the power of this reasoning was clear to us, we had shown the existence of the virion RNA polymerase (24). Tlie demonstra tion of this enzyme was the piece of evidence that led to the realization that there is a huge cls of viruses, now called negative strand viruses G5), that " AP00024539 f all carry the strand of RNA com plementary to the messenger RNA as their genome and that carry an RNA polymerase able to copy the genome RNA to form multiple messenger RNA's. Retroviruses The discovery of a virion polymerase in VSV led us to search for such en zymes in other viruses. Because New castle disease virus made a lot of comple mentary RNA after infection, it seemed a logical candidate; after an initial fail ure {34), we found activity in virions of that virus (36). But a more exciting possibility occurred to me; maybe by looking for a virion polymerase, light could be shed on the puzzle of how RNA tumor viruses multiply. In his Nobel lecture, Howard Temin has outlined how he was led to postulate a DNA intermediate in the growth of RNA tumor viruses (37). Although his logic was persuasive, and seems in retro spect to have been flawless, in 1970 there were few advocates and many skeptics. Luckily, I had no experience in the field and so no axe to grind--I also had enormous respect for Howard dating back to my high school days when he had been the guru of the summer school I attended at the Jackson Laboratory in Maine. So I decided to hedge my bets--I would look for either an RNA or a DNA polymerase in virions of RNA tumor vi ruses. To make the foray into tumor virolo gy, 1 needed some virus. Peter Vogt first sent me some Rous sarcoma virus and, although 1 later used it to good advan tage, I initially assayed for an RNA po lymerase in this viral preparation and failed to find anything. Then George Todaro helped me utilize the resources of the Special Virus Cancer Program of the National Cancer Institute to get some Rauscher mouse leukemia virus. Using that virus preparation I set out to look for a DNA polymerase activity. With little difficulty, I was able to demonstrate that virions of Rauscher virus contained a ribonuclease-sensitive DNA polym erase activity, and, after confirming the results with Rous sarcoma virus, I knew that the machinery for making a DNA copy of the RNA genome was wrapped up inside the virions of RNA tumorvi rus es (38). Simultaneously with my work. Temin and Mizutani discovered the DNA synthesis activity in Rous sarcoma virus (39). Production of the Integrated Viral Genome Expression of the Integrated Wal Genome in a Productive Infection Fig. 2. The life cycle of an RNA tumor virus. On the basis of present knowledge (42), the life cycle ofan RNA tumor virus can be separated into two parts. In the first part the virion attaches to the cell and somehow allows its RNA along with reverse transcriptase to get into the cell's cytoplasm. There the reverse transcriptase causes the synthesis of a DNA copy of the viral RNA. A fraction of the DNA can be recovered as closed, circular DNA (43), and it is presumably that form which integrates into the cellular DNA. Once the proviral DNA is integrated into cellular DNA it can then be expressed by the noimal process of transcription. The two types of product which have been characterized are new virion RNA and messenger RNA. Much of the messenger RNA which specifies the sequence of viral protein is of the same length as the virion RNA. but there may also be shorter messenger RNA's (48). The virus specific proteins have two known functions: one is to transform cells, which occurs when, for instance, a sarcoma virus infects a fibroblast; the second is to provide the protein for new virion production. The transforming protein is shown here as acting at the cell surface, but that is only a hypothesis; v, viral; Tranxf., transforming; Rev. Tptase, reverse transcriptase. 634 Biochemistry of Reverse Transcriptase ` Once the DNA polymerase activity had been demonstrated in the virions of what we now call retroviruses, many laboratories began to study the enzyme. A correspondent of Nature dubbed the enzyme "reverse transcriptase" (40) and this romantic name has become common parlance. About 2 years after its dis covery, Howard Temin and I reviewed the literature on the enzyme (41). That review and later compendia (42) make a detailed rehash of the biochemistry of retroviruses superfluous. So. I will only present a brief sketch of the picture we have today of how a retrovirus multiplies and how the reverse transcriptase func tions. I will not attempt to credit all of the people who have helped to clarify this picture. There are two separate time periods that can be distinguished after infection of a cell by a retrovirus (Fig. 2). The first period, which lasts a few hours, involves reverse transcription and establishment of the DNA provirus as an integrated part of the cellular DNA; the second period involves the expression of the integrated genome and synthesis of prog eny virions. Analysis of the first period of retro virus growth has focused on the types of virus-specific DNA molecules that are produced. One important type of DNA that has been found is a closed circular duplex DNA of about 5.7 x 10" daltons (43). This DNA can transfect cells with one-hit kinetics (44) and therefore con tains the total viral genetic information. Other DNA's that may be on the way to becoming the closed circular form are al so evident (45). It has been hard to get de finitive evidence as to what DNA form in tegrates but presumably it is the circular duplex DNA. Whatever the form that in tegrates, the evidence is quite good for acquisition of proviral DNA by the chromosomes of infected cells (46, 47). The second period in a productive ret rovirus growth cycle starts when the in tegrated genome begins making viral RNA (48). Synthesis of viral proteins and progeny virions ensues, and the cell ever after continues to make' viral pro ducts except for variations imposed by the cell cycle (49). Among the viral pro teins made in the infected cell may be a product that changes the growth proper-1 ties of the cell (50); in such a case the ret rovirus becomes a tumor virus. The second period of the infection cycle can be dissociated from the first in a number of experimental systems. For instance, mammalian cells infected by avian viruses can gain viral DNA but not SCIENCE. VOL. 192 , AP00024540 express it {46). Also, cells can have viral genomes that they inherited from their ancestors, and such genomes are generally not being transcribed. Nonexpressed genomes can be activated: bro- DNA formally takes the place of the "mimis" strand in the picomavirus genetic system--but the DNA is probably the clue to why retroviruses are the only ones able to cause cancer. The DNA pro- play some important role in the life of the animal and so are not dispensable. This explanation is basically the protovims hypothesis put forward by Howard Te- min (37). v modeoxyuridine and iododeoxyuridine, for instance, stimulate the expression of inherited, silent viral genomes (5/). The exact mechanism of activation of the genome for transcription, initiation of the transcript, and termination of vides the necessary stability to the viruscell interaction so that a viral gene product can permanently change the growth properties of an infected cell. Equally signiiicant, the DNA stage is probably important to the ease with which retro- 2) They are genes inserted into the chromosome of some ancestral animal by a retrovirus infection of the germ cells of that animal. Because once the provirus is integrated it remains stably associated with the chromosome, the viral transcription are obscure, as are any processing events of the initial transcript which may occur. It is evident that the key piece of ma- viruses carry out genetic recombination (55); it is quite possible that the recombination system can bring together host cell genetic information with viral infor- genes are inherited by progeny of the original infected animal. There is one force that can eliminate such genes from a species, the'slow but inexorable pro- chinery provided by the virus for this- mation and that in this way nonunique life cycle is the reverse transcrip- oncogenic retroviruses become oncogentase. Purified reverse transcriptase has ic (56). the properties of most DNA polymer- So the inclusion of a proviral stage in ases: it is a primer-dependent enzyme the retrovirus life cycle may provide critithat makes DNA in a 5' - 3' direction cal capabilities toward the development cess of mutation. As part of this ex planationof inherited viral genomes, therefore, it has been suggested that the viral genes have some positive influence on the life of the animal and so are maintained by positive selection. This ex using deoxyribonucleoside triphosphates of an oncogenic potential. But the actual planation is closely related to the viroas substrates and taking the direction of transformation of cells by retroviruses is gene-oncogene hypothesis (61). a template for determining the base sequence of the product. The enzyme differs from normal cellular DNA polymerases by having a unique polypeptide structure, having an associated ribonudeaseH activity, and being able to make a highly selective process; each type of oncogenic virus transforms a very limited range of cell types (57). If we assume that all transforming genes of viruses are like those of Rous sarcoma virus, genes that are not necessary to the growth of 3) The previous explanation can be modified by the exclusion of any positive role for viral genes in the life ofthe infected animal. There are a number of reasons why positive selection may be an unnecessary attribute to postulate. For copies of RNA templates as readily as DNA templates (41). Genetics has shown us that the avian leukosis viral enzyme, at least, is encoded by viral RNA and needed only in the first period of the infection cycle (52). The primer dependence of the reverse transcriptase means that the enzyme can only elongate nucleic acid molecules, it cannot initiate DNA synthesis de novo. How then does the enzyme initiate the copying of viral RNA? The answer is that the genome RNA has attached to it a primer RNA molecule which is, in the the virus (50, 58), then we can postulate that each type of transforming virus makes a specific type of transforming protein. Such a protein, by this model, would not be critically involved in the multiplication cycle of the virus. Isolation of such transforming proteins and elucidation of their mechanism of action is the present challenge of cancer virology. Not only will such work help us to understand carcinogenesis, it may also be important to the study of developmental biology because of the intimate relationship between the differentiated state of one thing, mutations are rare events and totally inactivating mutations are much rarer. Also, the virus can be genetically invigorated by becoming a replicating vi res in the body of the animal and then re infecting the germ line. When the vims starts multiplying as an independent entity, the burden of debilitating mutations it might have accumulated could be purged if a sufficient number of generations intervened between the activation of the latent provirus and its reintegration into the germ line. The rein tegration might even replace the original case of avian leukosis viruses, cellular tryptophan transfer RNA (53). The avian leukosis virus reverse transcriptase has a high-affinity binding site for that transfer RNA which the enzyme presumably uses for precise initiation of reverse tranxcription (54). Reiroviiuses and Cancer The last 15 years of research in animal virology has allowed us to see the diversity of genetic systems used by the vanous types of RNA viruses and has most recently shown us how distinct the retroviruses are from the others. Rather than using an entirely RNA-dependcnt replication and transcription machinery, the retroviruses have included the DNA provirus in their life cycle. Having a DNA intermediate does not make their mode o growth especially complicated--the cells and the type of virus able to transform them. Another implication of the occurrence of a proviral stage in the life cycle of retroviruses is that cells can harbor such viruses as genetically silent DNA molecules. In fact, in most, if not all, animal species, the normal cells of the body have DNA related to one or more types of retroviruses (59). They receive that DNA by inheritance, not infection, and in favorable cases it can be as precisely located in the chromosomes as any gene (60). What is the significance of these genes that look like viruses? There have been three types of expianations for virus-related genes that are inherited in the germ line of so many animal species: I) They are an aspect ofthe normal genetic complement of the animal and they are virus-related because they are the progenitors of retroviruses. These genes provirus (62). If the viral genes were not transcribed in most cells that have the viral genome, as appears to be the case, the negative effect of having one or a few integrated genomes would be slight and probably insufficient to cause a serious negative selection against animals with inherited proviruses, I would argue that the third explanation above is the one most likely to be correct. It is an explanation that maintains the separation of viral activities and cellular activities and does not require the ad hoc postulation of beneficial properties of viral products. It treats retroviruses like any other virus, as an entity with its own life-style and its own accommodation with evolution, In summary, I have tried here to develop the view of retroviruses as one of a number of solutions to the problem of creating a virus. Each virus directs synthesis of two critical classes of proteins: M MAY 1976 ***'' 635 ;.L-.- .. AP00024541 proteins for replication and proteins for constructing the virus particle. By encod ing the reverse transcriptase, retro viruses have evolved the ability to in tegrate themselves into the cell chromo some as a provirus. This is a very sheltered environment in which to live; only mutation interferes with the contin ual transmission of the virus to the proge ny of an animal that is infected in its germ cells. In this context, the ability of some retroviruses to cause cancer is a gratuitous one. But it is today the most challenging and important attribute of these retroviruses and the one that will dominate future research efforts in this area. RefarencH and Notes 1. T.Diencr.Adv. VlmiRes. 17,295(1972). 2. D. Baltimore.Bacterid. Rev. 35.235 (1971). 3. S. E. Luria. in The Viruses. F. M. Burnet and W. M. Stanley, Eds. (Academic Press, New York, 1959), vol. l.pp. 549-56*. 4. R. M. Franklin and J. Rosner, Blochlm. Biophys. Acta 55.240 (1962). 5. E. Simon. Virology 13,10511961). 6. E. Reich. R. M. rranklin, A. J. Shitkin, E. L. Tatum, Science 134,556(1961). 7. S.S. Cohen.Fed. Froc. 20,641(1961). 8. D. Baltimore and R. M. Franklin, Proe. Nad. Acad. Sci. U.SA. 48, 1383 (1962); R. M. Frank* lin and D. Baltimore, Cold Spring Harbor Symp. uaw.Biol.n, 175(1962). g. Baltimore and R. M. Franklin. Biochem. Biophys. Res. Ccmmun. 9. 388 (1962); J. Biol. Chem. 238. 3395 (1963); Cold Spring Harbor Symp. Quant. Biol. 28. 105 (1963); H. J. Eggers, I. Tamm, Proc. Nad. Acad. Scl. U.SA. 49,843(1963). 10. D. Baltimore. Proc. Nad. Acad. Sci. U.SA. 51, 450(1964). 11. -- in The Biochemistry of Viruses. H. B. Levy, Ed. (Dekker. New York. 1969). pp. 101176. 12. L. Levintow, in Comprehensive Virology, H. Fraenkel-Conrat and R. R. Wagner, Eds. (Pie* num. New York, 1974), vol. 2, pp, 109-169. 13. R. E. Lundquist, E. Ehrenfeld, J. E. Maizel, Proc. Natl. Acad. Sci. U.SA. 71,4773(1974). 14. D. Baltimore, Y. Becker, J. E. Darnell, Science 43, 1034 (1964); D. Baltimore, J. Mol. Biot. 18, 421 (1966); --------- - M. Girard, J. Darnell, Virology 29. 179 (1966); D. Baltimore and M. Girard, Proc. Nad. Acad. Scl. U.SA. 56, 741 (1966); M. Girard. D. Baltimore. J. E. Darnell J. Mol. Bioi, 24 , 59 (1967); D. Baltimore, ibid. 32.359(1968). 15. M. Girard.7. Virol. 3,376(1969). 16. M. Jacobson and D. Baltimore,y. Mol. Bid. 33. 3690966). 17. ____ _ Proc. Nad. Acad. Sci. U.SA. 61, 77 (1968); M. F. Jacobson. J. Asso, D. Baltimore, y. Mol. Biol. 49,657 (1970). 18. D. Baltimore, Harvey Lectures (Academic Press, New York.in press). 19. L. Villa-Komaroff, N. Guttman, D. Baltimore. H. F. Lodish, Proc. Nad. Acad. Sci. USA. 72, 4157(1975). 20. J. S. Colter, H. H. Bird, A. W. Mayer, R. A. Brown, Virology 4,522 (1957); H, E. Alexander. G. Koch, 1. M. Mountain, K. Sprunt, O. Van Damme,ibid. 5, 172(1958). 21. Y. Yogo and E. Wimmer, Proc. Nad. Acad. Sci. U.SA. 69, 1827 (1972); Y. Yogo, M. Ten*, E. Wimmer, Biochem. Biophys. Res. Commun. 61, 1101 (1974); D. H. Spector and D. Balti more, /. Virol. 15, 1418 (1975); Virology 67, 498(1975). 22. D. H. Spector and D. Baltimore, Proc. Natl. Acad. Scl. U.S.A. 71,2983(1974). 23. E. Wimmer.7. Mot. Bid. 68.537(1972). 24. A. Nomoto, Y. F. Lee, E. Wimmer,Proc. Nad. Acad. Sci. U.SA. 73,375 (1976); M. J. Hewlett, J. K. Rose, D. Baltimore.ibld..p. 327. 25. P. J. Gomatos and I. Tamm, fif'd: 49.707 (1963). 26. A. J. Shatkin and J. D. Sipe, Ibid. 61. 1462 (1968); J. Borsa and A. F. Graham, Biochem. Biophys. Res. Commun. 33,895 (1968). 27. J. R. Kates and B. R. McAuslan. Proc. Nad. Acad. Sci. U.S.A. 58.1134 (1967). 28. W. Munyon, E. Paolett'i, J. T. Grace, ibid.. p. 2280. 29. D. Kingsbury, J. Mol. Biol. 18. 204 (1966); M. A. Bratt and W. S. Robinson, ibid. 23,1 (1967). 30. M. Stampfer. D. Baltimore, A. S. Huang, J. . Vird. 4,154 (1969). 31. T. G. Morrison, M. Stampfer, H. F. Lodish, D. Baltimore, in Negative Strand Viruses, B. W. J. Mahy and R. D. Barry, Eds. (Academic Press, New York, 1975). vol. 1. pp. 293-300; J. K. Rose and D. Koipe, J. Vird. 15,994 (1975). 32. F. L. Schaffer, A. J, Kackett. M. E. Soeigd. Biochem. Biophys. Res. Commun. 31, 685 0968). 33. A. S. Huang, D. Baltimore, M. Stampfer, Virolo gy 42,946(1970). 34. D. Baltimore, A. S. Huang, M. Stampfer, Proc. Natl. Acad. Sci. U.SA. 66.572(1970). 35. B. W. J. Mahy and R. D. Barry, Eds.. Negative Strand Viruses (Academic Press, New York, 1975), vols. 1 and 2. 36. A. S. Huang, D. Baltimore, M. A. Bratt, J. Virol. 7,389(1971). 37. H.Temin.5ci>ftcr, in press. 38. D. Baltimore, Nature {London} 226,1209(1970). 39. H.TeminandS. Mizutam.fW.. p. 1211. 40. Anonymous, ibid. 228,609(1970). 41. H. Temin and D. Baltimore, Adv. Vitus Res. 17, 129(1972). 42. Cold Spring Harbor Symp. Quant., BioI. 39 (1975); J. Took, Ed., The Molecular Biology of Tumor Viruses (Cold Spring Harbor Laborato ry, Cold Spring Harbor, N.Y,, 1973); J. M. Bishop and H. E. Varmos, in Cancer, F. F. Becker, Ed. (Plenum, New York, 1975), pp. 3- 43. H. E. Yarmus, R. V. Guntaka, W. J. w. Fan, S. Heasley, J. M. Bishop, Proc. Natl. Acad. Sci U.SA. 71, 3874(1974); A. M. Gianni, D. Smot* kin, R. A. Weinberg, ibid. 72,447(1975). 44. D. Smotkin, A. M. Gianni, S. Rozenblatt, R. A Weinberg,ibid. 72,4910 0 975). 45. A. M. Gianni and R. A. Weinberg, Nature (Lon don) 255.646 (1975). 46. H. E. Vannus, P. K. Vogt. J. M. Bishop. Proc. Natl. Acad. Set. U.SA. 70,3067(1973). 47. P. D. Markham and M. A. Baluda.y. Virol. 12, 721 (1973). 48. H. Fan and D. Baltimore, J. Idol. Bid. 80. 93 (1973). 49. J. A. Leong, W. Levinson. J. M. Bishop, Virolo gy 47, 133*1972); M. P. Paskind, R. A. Wein berg. D. Baltimore.ibid. 67,242 (1975). 50. G. S. Martin. Nature (London) 227,1021 (1970). 51. D. R. Lowy, W. P. Rowe, N. Teich, J. W. Hartley, defence 174,155(1971). 52. I. M. Verma. W. S. Mason, S. D. Drost, D. Baltimore. Nature (London) 251,27 (1974). 53. `J. E. Dahlberg, R. C. Sawyer, J. M. Taylor. A. J. Faras. W. E. Levinson. H. M. Goodman, J. M. Bishop,7. Virol. 13,1126(1974). 54. A. Panel, W. A. Haseltine, D. Baltimore, G. Peters, F. Harada, J. E. Dahlberg, Proc. Natl. Acad. Sci. U.SA. 72,2535(1975). 55. J. Wyke, J. G. Bell, J. A. Beamund. Cold Spring Harbor Symp. Quant. Biol. 39, 897 (1975); J. Wyke.Rev. Cancer 417,91 (1975). 56. R. A. Weiss, in Possible Epitomes in Eu karyotes, 4th Lepedt Symposium, L. G. Sylvestri, Ed. (North-Holland, Amsterdam. 1973), pp. 130-141; D. Stehelin, R. V. Guntaka, H. E. Vannus, J. M. Bishop, J. Mol. Bid. 101, 349 (1976); D. Stehelin, H. E. Vannus, J. M. Bishop, E. K. Vogt, Nature (London) 260, 170(1976). 57. L. Gross, Oncogenic Viruses (Pergamon, New York,ed.2,1972). 58. H. Hanafusa, in Cancer, F. F. Becker, Ed. (Ple num, New York, 1975), vol. 2, pp. 49-90. 59. G. J. Todaro, R. E. Beneviste. R. Callahan, R. R. Lieber. C. Sherr, Cold Spring Harbor Symp. Quant. Biol. 39, 1159 (19/5); R. E. Beneviste andG. J. Todaro, Proc. Nad. Acad. Set. U.S.A. 72,4090(1975). 60. S. K. Chattopadhyay, D. R. Lowy, N. M. Teich. A. S. Levine, W. P. Rowe, Cold Spring Harbor Symp. Quant. Bid. 39, 1085 (1975); S. K. Chattopadhyay, W. P. Rowe, N. M. Teich, D, R. Lowy, Proc. Nad. Acad. Sci. U.SA. 72, 906(1975). 61. R. J. Huebner and G. J. Todaro, Proc. Natl. Acad. Scl. U.S.A. 64, 1087 (1969); G. J. Todaro and R. J. Huebner, ibid. 60, 1009 (1972). 62. H. M.Temin, Vlrdogy 13,159(1961). H1W AMP COMMENT Anatomy of a Decision: How the Nation Declared War on Swine Flu Late on the afternoon of 24 March, President Gerald R. Ford appeared be fore White House reporters to discuss "a subject of vast importance to all Ameri cans"--the appearance of a new strain of flu. A month earlier, scientists had discovered that Army recruits at Fort Dix, New Jersey, were infected by "swine flu virus." One of them had died. The last time an outbreak of swine flu appeared in the United States was in 636 1918-19 when, the President told re porters, "a widespread and very dead ly flu epidemic" killed 548,000 Ameri cans of all ages. The 1918 epidemic was really a pandemic; in successive waves, swine flu swept around the world, leav ing 20 million persons dead. It is said that the pandemic killed more individuals in a short period of time than any other catas trophe in history. After consulting with his top health advisers and virologists from throughout the country. Ford was concerned about the "very real possibility" of a danger ous epidemic in the United States next fall and winter. To head off this threat, the President announced that he would ask Congress to appropriate $135 mil lion right away to buy enough vaccine to inoculate "every man, woman, and child in the United States." This would be the largest immuniza tion campaign ever launched in this coun try, far more ambitious even than the polio immunization drives of the 1960's. Then, about 100 million Americans re ceived polio vaccine during a period of a year and a half. Ford was launching a campaign to vaccinate twice as mahy citizens in a matter of months. Officials said that all his health advisers thought ' SCIENCE. VOL. 192 AP00024542 1 i. AtU -A ing. The demonstration that viral DNA is integrated in the cells, in conjunction with the finding that the provirus is tran scribed into messenger RNA (6) hundreds of generations after the establishment of a transformed clone, made the hit-and- From the Molecular Biology of run hypothesis unlikely and supported a continuing role of viral gene functions in Oncogenic DNA Viruses to Cancer determining transformation. This possi bility was later supported by observa tions with abortively transformed cells, Renato Dulbecco which behave as transformed only for several generations after infection, but then return to nbrmai (7). When they are back to normal these cells no longer contain the viral DNA (8). , Oncogenic viruses, able to elicit tumor Integration: The Provirus formation in animals, have been on the The viral genes that remain unex pressed in the transformed cells, such as scientific scene for many years. After the early discovery of Ellerman and Bang at the beginning of this century, Peyton Let me start with a brief review of our those for capsid protein in SV40-transwork on the molecular events in transfor- formed cells, were also interesting, al malion. The first results, crucial for fu though in a different way. In fact their Rous opened up the field in the next ture developments, showed that poly expression could be renewed in hetero- decade and in prophetic words gave a oma virus could be assayed in certain karyons formed by fusing transformed good hint of things to come. However, cell cultures (/), which we call per cells with permissive cells (9), a result these discoveries were soon forgotten, missive, and could induce a cancer-like that gave the first evidence that the viral and only after a long eclipse was interest state in other cultures (2) in which the functions are under the control of cellu in oncogenic viruses revived in the virus does not grow, which we call non- lar functions. The provirus thus became 1950's. My involvement in this field be permissive. The induction of the cancer a tool for studying regulation of DNA gan at that time, when Rubin and Temin like state in vitro was called transforma transcription in animal cells. Sub worked in my laboratory with the Rous tion. We were able to show that the virus sequently, the presence of giant RNA's sarcoma virus. When polyoma virus, a contains DNA (7), and within a few containing viral sequences in the nucleus new oncogenic virus with different prop erties, was isolated in 1958, I jumped at the new opportunity and started working years we gave the first evidence of its cyclic, or circular, shape (4), which is important for two critical biological of transformed or tylically infected cells (10) raised the question of the initiation and termination signals for transcription with it. Within a short time poiyoma events: DNA replication and integration. in animal cells, as well as the question of virus became the main interest of my In integration, which we discovered a processing of nuclear RNA precursors of laboratory, to be joined, a few years few years later with the virus SV40 (J), messenger RNA, questions that are still later, by SV40, another papovavirils. It the viral DNA becomes a provirus, that largely unresolved. became clear fairly soon that the molecu is, it establishes permanent, covalent lar biology of these viruses could be bonds with the cellular DNA. The cyclic worked out, and 1 set out to find the configuration explains how a complete Viral Functions in Transformation molecular basis of cancer induction. The molecule of the SV40 DNA can be in results that I and a number of brilliant tegrated without losses. In the meantime efforts were directed young collaborators have obtained dur Integration is one of the key events in at identifying the viral genes transcribed ing the following 15 years have brought virus-induced cell transformation. It ex in the transformed cells. It was estab us close to that goal. I will review the plained the persistence of the trans lished that in lytic infection with SV40 most interesting steps of our work and formed state in the cell clone deriving the whole viral DNA is transcribed in will then ask some questions concerning the nature of cancer and about per spectives for prevention and treatment. I stress the relevance of my work for can from a transformed ceil, since the pro virus replicates with the cellular DNA. It also permitted us to resolve one of the main questions about the role of viruses two nearly equal parts--one early, be fore the inception of replication of the viral DNA, the other late, after DNA replication has begun--and that the early cer research because I believe that sci in transformation. It was known at the RNA is also present in transformed cells ence must be useful to man. time that papovaviruses leave their foot (6). Subsequently, the early and the late f i -----:-------------- :---------------------------------- -- Copyright Cl976 by the Nobel Foundation. The author la deputy director for research at the Imperi al Cancer Research Fund Laboratories. Lincoln's Ion Fields, London, W.C.2, and fellow of the Salk Institute, San Diego, California. This article is the lecture he delivered in Stockholm, Sweden on 12 December 1975, when he received the Nobel Prize in Physiology or Medicine, a prize be shared with David Baltimore and Howard M. Temin. The article is published here with the permission of the Nobel Foundation and win alto be included in the complete volume of Let Prlx Nobtl fit 1971 as well as in the series Nobtl Leeturei (in English) published by the Elsevier Publishing Company, Amsterdam and New York. The lecturet by Baltimore and Temin will appear in later issues. JO APRIL 1976 y prints in the ceils of the cancers they induce and those they transform in vitro, in the form of characteristic antigens. However, it was not known whether the antigens were expressed by viral genes or by derepressed cellular genes. Hence, it was uncertain whether cells were trans formed by the expression of viral genes persisting in the ceils or, alternatively, whether the virus altered the cells by a hit-and-run mechanism, changing the ex pression of cellular genes and then leav messengers were found to be transcribed from different DNA strands (II), an ob servation that facilitated further charac terization of the viral transcripts. Later work in other laboratories with specific fragments produced by restriction en donucleases confirmed and refined these findings, and the results were extended to adenoviruses by showing that a segment of the early part of that DNA is always present and transcribed in trans formed cells (12). -y ' 'r~ I . 437 .. . * -a x AP00024543 -5 These facts suggested that some early viral function is essential for maintaining the transformed state but they could also be interpreted differently: for instance, transformation might be caused by the mere presence of the viral DNA in the cellular DNA. the persistent viral func tions being perhaps required for estab lishing and maintaining integration. Attempts were made to solve the di lemma by isolating temperature-sensi tive mutants affecting either initiation or maintenance of transformation. Many transformation mutants were found, with mutations all clustered in a segment of the early region of the viral DNA. desig nated as the A gene, but they were all initiation mutants {13). These mutations prevent the onset of transformation at high but not at low temperature, and cells transformed at low temperature re main transformed at high temperature. It was not possible to find clear-cut main tenance mutants, that is. mutants ca pable of causing a complete reversion of the phenotype when cells transformed at low temperature were shifted to high temperature. However, careful observa tion later showed that the initiation mu tants were also partial maintenance mu tants, since the cells they transform un dergo a partial reversion of phenotype at high temperature (14). This result shows that the viral genes play a continuing role in transformation; however, the failure to obtain complete maintenance mutants suggests that the relation between viral gene expression and cell phenotype is complex. Search for the Viral Transforming Protein Further progress in this subject has been achieved by studying the proteins specified by the early region of the viral DNA. This work has centered around the so-called T antigen (15) present in the nucleus of cells infected or transformed by SV40; the synthesis and properties of this antigen are affected by mutations of the A gene (16). In nonpermissive trans formed cells the antigen is a protein with molecular weight of about 94,000 daltons (17), which binds firmly" to doublestranded DNA but without much speci ficity (18). That the T antigen is specified by the viral DNA is strongly suggested by its in vitro synthesis by a wheat germ extract primed with various messengers (19), especially since the size of the pro duct depends on the nature of the messen ger. Thus, when the messenger was viral RNA made in vitro by transcribing SV40 DNA with Escherichia coli RNA polym erase, an antigenic protein of about 62.000 daltons was synthesized; but when mes senger RNA extracted from infected cells was used, the protein synthesized was, like the T antigen of transformed cells, of about 94.000 daltons. The dis crepancy of the two molecular weights makes it very unlikely that the T antigen is a cellular protein modified by a viral function, because two different proteins would have to be modified in the same extract depending on the messenger used. In contrast, the synthesis of a shorter polypeptide chain with the arti ficial messenger may be justified by the absence of accessory signals, such as the special nucleotide sequence present at the 5'-end, known as "cap," polyad enylate at the 3'-end, and possibly other modifications. Further definition of these findings awaits peptide maps of the vari ous products. Since the early, transforming, part of the SV40 genome can specify proteins of a molecular weight of about 100.000 dal tons altogether, the T antigen is likely to be its sole product and. therefore, to be the transforming protein. However, the same protein must also initiate viral DNA replication, which cannot begin at high temperature in cells infected by mu tants of the A gene. The different func tions in transformation and lytic infec tion could be performed by different do mains of the same protein, or could re sult from modifications (such as phosphorylation and glycosylation) or from processing. Processing of SV40 T antigen seems to occur in lyrically infect ed cells which contain a smaller T anti gen of about 84,000 daltons; this smaller size contrasts with the regular size (94,000 daltons) of the antigen specified in vitro by messenger RNA extracted from the same cells (17). Whether the two forms of the antigen have different roles in transformation and DNA replica tion remains to be established. Since the transforming protein should control both initiation and maintenance of transformation, the partial reversion of the phenotype of cells transformed by A mutants when shifted to high temper ature may be explained by a decreased requirement for the transforming protein once transformation has taken place, which in turn could result from a positive feedback stabilizing the transformed state. For instance, unstable protein monomers specified by the mutated gene might form self-stabilizing oligomers (20), or the transforming protein might generate changes that tend to favor the transformed state. An example of the latter model is the 0-galactcsidase induc tion in E. coli which is maintained by inducer concentrations much smaller than that required for initiating induc tion. because inducer is pumped into the cells by the induced permease (2/). I wonder whether a certain degree of self stabilization of the state of gene ex pression is a general property of animal cells which has developed for main taining differentiation. Cellular Events in Transformation I now turn to cellular events participat ing in transformation, which will be the main problem after the remaining ques tions on the role of the virus have been answered. Among the cellular events are functional changes and mutations. Some functional changes, which affect many cellular properties, are associated with the shift of resting cells to a growing state after infection with polyoma virus or SV40 (22); other changes observed in transformed cells and in cancer cells in general consist of the reexpression of cellular genes normally expressed in a preceding state of differentiation, in fetal life (23). These functional changes might be caused by the binding of transforming proteins to DNA; if so, they may be mediated by an alteration of transcrip tion of the cellular DNA. However, we do not know whether the transcription pattern changes, because experiments based on competition hybridization have given ambiguous results. Perhaps the methodology is not good enough. Clon ing of cellular DNA fragments in phages or plasmids may afford the necessary probes for carrying out significant experi ments. In order to understand further how the virus deregulates cellular growth we would need detailed knowledge of the mechanisms of growth regulation in ani mal cells, which is now lacking. How ever, certain useful ideas about growth regulation are now available, and can be used to draw inferences about the action of the virus. Thus it seems clear that with a given cell type, growth regulation in volves a complex chain of events, begin ning with extracellular regulators of many kinds, probably interacting with the cell plasma membrane. Cytoplasmic mediators then appear to transmit regu-' latory signals from the plasma membrane to the nucleus, where they perhaps con trol DNA-binding proteins similar to the transforming protein of papovaviruses. The complexity of growth regulation in creases markedly when different cell 43* '' SCIENCE, VOL. m AP00024544 types are considered, since they seem to recognize different sets of extracellular regulators and may have different media tors and DNA-binding proteins. Proceeding from this general picture it would be tempting to propose that the viral transforming protein replaces one of the normal nuclear regulatory proteins of the cell and. being unaffected by the mediators that control the normal pro Role of Cellular Mutations I will now consider the other cellular events important in viral transformation: cellular mutations. Several results suggest that cellular mutations may be needed for obtaining the full state of transformation with papovaviruses. Thus, after infection primary cultures generate clones with various degrees of from the various points 1 have discussed so far. One deduction, deriving from the persistence of the viral DNA in the cells, is that we can test whether a given DNA virus is a possible agent of human cancer by looking for its DNA in the cancer cells. I think that much more extensive surveys than those carried out so far are warranted, bul they should have a sensitivity sufficient to detect fragments of , tein, keeps growth-related transcription transformation, some of which appear to viral DNA of about 1 million daltons, going, bypassing the signals of the undergo full transformation in steps (26) which is within the reach of modem tech- plasma membrane. If so, however, the that may correspond to the occurrence noiogy, even with the most difficult virus- transformed state should be dominant over the normal state in cell hybrids, whereas the contrary is usually true Q4). On the other hand, the dominance of the of cellular mutations. Cells that achieve full tra isformation immediately, as is common with permanent lines, may have already undergone similar mutations be- es. A positive finding would be signifi- cant because DNA viruses do not appear to exist in widespread endogenous forms. . in- normal state could be explained if the fore infection. Some cellular mutations Another deduction is that somatic mutransformed cells had a changed surface, occurring in transformed cells may even tations are one of the fundamental in- unable to respond to regulatory signals. be virus-induced, because in the early gredients of cancer, although they apSuch a change could result from the reex-. stages of transformation by papovavi- pear to require the occurrence of several pression of fetal functions to make the ruses cells of primary cultures have other events not yet understood. The transformed cells anachronistic, that is, frequent chromatid breaks (26). Con- role of mutations in tum suggests that belonging to a stage of differentiation versely, cells fully transformed by SV40 the incidence of cancer in man could be inappropriate to that of the organism which contains them. The cells with an anachronistic surface, being insensitive to the growth regulators which operate can revert to a relatively normal phenotype although they still contain normal viral DNA and T antigen (27). It is conceivable that these mutations are rever- reduced by identifying as many promutagens as possible, and by eliminating them from the environment. One important feature of this approach to cancer on adult cells in the adult organisms, sions of mutations of the former kind, prevention is that it can be started now, would grow without control. A striking which enhance the transformed state of since these substances can be identified support of the role ofcell anachronism in cancer has been obtained with teratoma, a tumor originating when cells from an the cells. Stepwise transformation may occur not only with viruses. Thus 1 have observed it in primary cultures exposed with simple bacterial tests suitable for mass screening {30). The feasibility of prevention is shown by the fact that the early embryo are transplanted to an adult to a chemical carcinogen. In this expert- promutagens already identified in a pre- environment. When, after many trans ment fully transformed cells evolved liminary screening, such as tobacco or plants. cells of this tumor are introduced from the normal cells, which have limit- some hair dyes, are inessential for hu- back into a blastocyst (an early embryo), ed life. The normal cells first generate man life (31). they return to normal (25), presumably ceils with unlimited life but unable to However, it is practically difficult to because the internal growth control of form colonies in agar, then cells with achieve a substantial reduction of the use the cells becomes again matched by the progressively increasing colony-forming of these substances, as shown by the environmental regulators of the recipient efficiency in agar, and finally cells that example of tobacco. According to epide- embryo. In this model a hybrid cell reach 100 percent efficiency. miological evidence, tobacco smoke is formed by fusing a transformed and a normal cell may be untransformed if the normal partner contributes normal sur All these observations show the impor- the agent of human lung cancer, which in tant role of cellular mutations in cell Britain is responsible for one in eight of transformation induced by different all male deaths (52). Yet only mild sanc- face components which respond to the agents. This conclusion is reinforced and tions have been imposed on tobacco normal extracellular regulators. For this generalized by additional findings, such products, such as a vague health warning result to be possible, anachronistic tran as (i) the experimental enhancement of on cigarette packets which sounds rather scription after cell fusion should not be initiated on the DNA deriving from the normal parent. The virological studies suggest that this may well be the case, the transforming activity of viruses by mutagenic agents (23); (ii) the elevated cancer frequency in some genetic diseases; and (iii) the evidence that most like an official endorsement. Any limitation on the use of tobacco is left to the individual, although it is clear that the individual cannot easily exercise volun- since the initiation of transformation seems to require much more trans forming protein than its maintenance. It would be important to recognize the developmental period in which the carcinogens are promutagens, that is, generate mutagenic substances when acted upon by normal metabolism (29). Most of the carcinogens themselves must be activated by metabolism in a tary restraint in the face of very effective advertisements, especially as one does not usually appreciate the danger of a cumulative action over a long period of time. anachronistic genes of transformed or cancer cells arc normally expressed, not only for understanding but possibly also for controlling cancer. In fact, if the growth regulators specific for the periods similar way in order to induce cancer. Prospects for Cancer Prevention The lax attitude of governments to ward tobacco probably also derives from the difficulty of appreciating epidemiological evidence, especially since this evidence is contradicted from time to expressed in cancer cells could be identi fied, they could be used for halting the owth of the cancer cells. I now turn to some general deductions time by single-minded individuals who concerning the etiology and possible pre- use incomplete or even erroneous analyvention of human cancer which derive ses of the data and whose views are APRIL 1976 - -J: -2T- r~ 439 ' - AP00024545 t magnified out of all proportion by the media. However, the recent recognition that tobacco smoke contains promuta- gens contributes direct experimental evi dence on the dangers of tobacco smoke, evidence on which there cannot be any equivocation. I, therefore, call on govemments to act toward severely dis couraging tobacco consumption, and to act now because it will be at least 30 years before their action has its full ef fect. Although tobacco smoke is a striking example of an environmental carcino gen, many others are known and prob ably many more remain to be identified. Identification by conventional tests is dif ficult because they are costly and labori ous, but they can now be replaced by the bacterial tests for promutagens. Since the tests are easy and inexpensive it should be possible to investigate many normal constituents of the environment, and every new compound before it is offered to the public. The feasibility of such a program is borne out by the find ing that most of the commonly available substances are not promutagens (31). Given the strong correlation between mu tagenicity and carcinogenicity (29), any promutagen is suspect and, if at'all pos sible, should be withdrawn. In fact, this is precisely the attitude that scientists have taken for themselves concerning the experiments in genetic engineering, which carry the theoretical possibility of creating new viruslike molecules endowed with carcinogenic ac tivity. Although the danger is only hypo thetical. experiments that might be very useful for science and society have been postponed until they can be carried out under the strictest safeguards (33). Gov ernments have accepted this position and are eager to impose severe restric tions on the performance of these experi ments. While I fully approve of their concern, I cannot help noticing that they follow a double standard: if there is any doubt you must discourage experiments. but if there is any doubt you cannot discourage cigarettes. Biologists and Society This discussion about cancer pre vention is a development of the experi mental results obtained in the field of oncogenic viruses, but it is also strongly influenced by the new social conscience of many scientists. Historically, science and society have gone separate ways, although society has provided the funds for science to grow and in return science has given society all the material things it enjoys. In recent years, however, the separation between science and society has become excessive, and the con sequences are felt especially by biolo gists. Thus, while we spend our life ask ing questions about the nature of cancer and ways to prevent or cure it. society merrily produces oncogenic substances and permeates the environment with them. Society does not seem prepared to accept the sacrifices required for effec tive prevention of cancer. The situation is clearly unacceptable, and we biolo gists would like to see it corrected. We have ourselves begun to put our house in order, by banning some experiments that may contain a risk for mankind. We would like to see society take a similar attitude, abandoning selfish practices that are dangerous for society itself. We would also like to see a new cooperation of science and society for the benefit of all mankind and hope that the dominant forces in society will recognize that this is a necessity. Rcfcrtnw and Notes 1. R. Dulbecco and C. Freeman. Virology I, 396 (1959). 2. M. Vogt end R. Dulbecco, Proc. Natl. Acad. Sci. U.SA. 46. 365 (I960): R. Dulbecco and M. Vogt,iWd..p. 1617. 3. i. D. Smith. G. Freemen. M. Vogt. R. Dul becco, Virology 12. 185 (1960). 4. R. Dulbecco and M. Vogt, Proc. Nad. Acad. Sci. U.SA. SO. 236(1963). 5. J. Sambrook, H. Wcstphal, P. R. Srinivasan, R. Dulbecco. 60.1288(1968). 6. K. Oda and R. Dulbecco. ibid.. ?. 525. 7. M. Stoker, Nature (London! 218. 234 (I*.8). 8. P. Berg and M. Stoker, personal communica tion. 9. H. Koprowski, F. C. Jensen, Z. Steplewski, Proc. Natl. Acad. Sci. V.S.A. S8, 127 (1967); J. F. Wakins and R. Dulbecco, ibid., p. 1396. 10. S. Toncfawa, G. Walter, A. Bemardlni, R. Dul becco. Cold Spring Harbor Symp, Quant. Biot, 35.823 (1970). 11. D. M. Undstrom and R. Dulbecco, Proc. Natl. Acad. Sci. U.SA. 69. 1517 0972). 12. P. A. Sharp, U. Pcttcnon, J. Sambrook, J. Mol. Biol. 84. 709 (1974); P. A. Sharp, P. H. Gam- more. S. J. Flint, Cold Spring Harbor Symp. Quant. Biol. 39, 457(1974). 13. M. Fried, Proc. Natl. Acad. Sci. U.SA. 53,486 (196.5): W. Eckhart, Virology 38, 120 (1969); P. Tegtmeyer and H. L. Ozer, J. Virol. 8, 516 (1971). 14. R. G. Martin. J. Y. Chou. J. Avila, R. Sara], Cold Spring Harbor Symp. Quant. Biot. 39, 17 (1974); J. S. Butel. J. S. Braege, C. A. Noonan, ibid., p. 25; G. Kimura ana A. Itagaki, Proc. Natl. Acad. Sci. USA. 72, 673 (1975). 15. P. H. Black, P. W. Rowe, H. C. Turner, R. J. Huboer. Proc. Natl. Acad. Sci. U.SA. 50. 1148 (1963). . 16. P. Tegtmeyer. Cold Spring Harbor Symp. Quant. Biol. 39, 9 (1974); M. Oxman. K K. Taketnoto, W. Eckhart, Virology 49 , 675 (1972); D. Paulin and F. Cuzin, J. Virol. IS, 393 (1975). 17. R. B. Carroll, personal communication. 18. , L. Hager. R. Dulbecco, Proc. Natl. Acad. Set. US A. 71, 3754 (1974); D. Jessel,). Hudson, T. Landau, D. Tenen, D. M. Livingston, ibid. 72, I960 (1975); S. 1 Reed. J. Ferguson. R. W. Davis, G. R. Stark, ibid., p. 1605. 19. A. E. Smith. S. T. Bayley, T. Wheeler, W. F. Mangel. In In vivo Transcription and Translotion of Viral Genomes, A. Haenni and J. Beaud. Eds. (Institut National de la Sant* ei de la Recherche Mddicale, Paris, 1975). 20. R. Dulbecco, Proc. R. Soc. London Ser. B. 189, I (1975). 21. A. Novick and M. Weiner. Proc. Natl. Acad. Sci. USA. 43, 553 (1957). 22. R. Dulbecco, L. H. Hartwell, M. Vogt, ibid. S3, 403 (1965); L. H. Hartwell, M. Vogt, R. Dul becco, Virology 27. 262 0963). 23. J. H. Coggin, J. Immunol. 105. 524 (1970). 24. F. Wiener. G. Klein, H. Harris, J. Cell Sci. 8, 681 (1971). 25. B. Mints and K. Illmensee, Proc. Natl. Acad. Sci. U.SA. 72, 3585 (1975); V. E. Papaionnou, M. W. McBumey, R. L. Gardner, M. J. Evans, Nature (London) 2S8. 70 (197S). 26. M. Vogt and R. Dulbecco, Cold Spring Harbor Symp. Quant. Biol. 27. 367 (1962). 27. R. E. Pollack, H. Green, G. J. Todaro, Proc. Natl. Acad. Sci. U.SA. 60, 126 (1968); H. C. Renaer and C. Basiiico, ibid. 69,109 (1972). 28. H. F. Stick, R. H. C. San, Y. Kawazoe, Nature tLondon.) 229.416(1971). 29. J. McCann. E. Choi, E. Yamasaki, B- N. Ames, Proe. Natl. Acad. Sci. U.SA., in press. 30. B. N. Ames, F. D. Lee, W. E. Durston. ibid. 70, 782 (1973). 31. L. D. Kier. E. Yamasaki. B. N. Amet. ibid. 71, 4159 (1974); B. N. Ames. H. O. Kammen. E. Yamasaki, ibid. 72. 2423 (1975). 32. R. Doll. /. Roy. Stat. Soc. Ser. A 134, 133 (1971). 33- P. Bag, D- Baltimore, S. Brenner, R. 0. RobItn, M. F- Singer, Proc. Natl. Acad. Set. U.S.A. 72. 1981 (1975). .. . ; - y- yjj r si1** . s.-v ' " VOL. 192 f AP00024546 How the Liver Metabolizes Foreign Substances Among the most significant of the liver's chemical transformations are the inactivation of drugs, the detoxification of environmental pollutants and the activation of chemicals that can cause cancer by Attall&h Kappas and Alvito P. Alvares he intensity and duration of the ters the liver through the large portal Taction of most drugs is determined in large part by their rate of me vein, which ramifies into fine channels through which the blood perfuses slow tabolism. If nothing else happened lytoamong the liver cells. Here nutrients a drug after it entered the body and and other foreign substances are re reached its target organ, for example, it moved, metabolized, in some cases might continue to act indefinitely. Some stared and then released into the gen thing does happen, however: most drugs eral circulation Amino adds, for exam are transformed into inactive substances ple, are made into proteins and other and then excreted. The biotransfonna- nitrogenous compounds; glucose is con tion can occur in any of several tissues verted into glycogen and stored, to be and organs. Some drugs are transformed converted back into glucose and released chemically in the intestine, some in the as required. And drugs and other toxic lung, the lddney or the skin. By far the substances are detoxified. Not every greatest number of these chemical re thing is metabolized on the first passage actions are carried out in the liver, which of blood through the liver, of course; metabolizes not only drugs but also most drugs, for example, are given in doses of the other foreign chemicals to which such that a sufficient amount of the drug the body is exposed. Biotransformation moves through the liver to its site of ac in the liver is therefore a critical factor tion and is transformed later, on return not only in drug therapy but also in de visits to the liver [see illustration on page fending the body against the toxic effects 25). The liver also produces bile, which of a wide variety of environmental chem is a secretion that aids in the digestion icals such as insecticides, herbicides, of fats when it is released into the small dyes, food preservatives and a number intestine and is also a vehicle for the of substances that are suspected of in excretion of transformed substances and ducing cancer. The central step in the other waste products of metabolism. metabolism of most of these agents in The biotransformation of drugs and volves an oxidation reaction mediated by other foreign compounds in the liver is a complex of enzymes that has come un accomplished by several remarkable en der intensive study in recent years in our zyme systems that can metabolize a wide laboratory at the Rockefeller University variety of structurally unrelated drugs, Hospital and in other laboratories. toxic agents and environmental pollu tants, which enter the body primarily nPhe liver is the largest organ in the through ingestion but also through the *- body (itweighs about three pounds in lungs and die skin. The enzyme systems an adult) and has diverse functions. It are built into the membranes of the en serves, first of all, as the primary receiv doplasmic reticulum of the liver cells, a ing depot, chemical-processing plant and network of interconnected channels that distribution center for almost everything is present in the cytoplasm of most ani that enters die body through die walls mal cells. There are two kinds of endo of the alimentary canaL All the blood plasmic reticulum, rough and smooth, that has absorbed digested food and and they differ in both form and func other substances from the intestines en tion. The surfaces of the rough mem branes are studded with ribosomes, small granules that translate the genetic code into the sequences of amino acids that constitute proteins. The smooth mem branes have no ribosomes. In the liver a major function of both kinds of mem brane is to assemble the enzymatic com plexes that transform foreign substances and then to serve as the site of those transformations. Unlike some other cellu lar subsystems, the endoplasmic reticu lum cannot be separated from cells as an intact structure. If liver cells are ho mogenized and then centrifuged, the tubular reticulum breaks up and bits of the membranes are sealed off to form the tiny vesicles, or sacs, called microsomes. The microsomal fraction thus ob tained from liver cells is a convenient natural source of enzymes for laboratory studies of liver-cell metabolism. T'krugs and other foreign compounds are metabolized in the liver by a rather small number of reactions: oxida tions, reductions, hydrolyses and con jugations. Their essential effect is to convert lipophilic, or fat-soluble, com pounds into hydrophilic, or water-solu ble, ones. The hydrophilic compounds are the more readily removed from the blood by the kidneys and excreted. Oxidation accounts for most of the transformations, largely because there are so many different ways in which a compound can be oxidized [see illustra tion on pages 28 and 29], The alkyl side chains of barbiturates and some other drags, for example, are oxidized to form alcohols. In the case of compounds in corporating aromatic rings, including polycyclic hydrocarbons (such as those in cigarette smoke) and many drugs, a hydroxyl group is inserted into the ring. AP00024547 omplished as a result of the estab lishment of certain new rural practioes. be first of these is related, to house hold economics at the production-team level. Individual household members re ceive a snare (in money and produce) of the annualWmings of their team; the share is proportional to the number of work points eahh has earned during tire year. For exampit a 70-year-old grand mother in a suburban Shanghai com mune told me that she and nine of the other 10 members of the household worked regularly in therSelds. The only nonproducer was a five-; The family received a monthly advance of the equivalent of $80 su'nst '' joint annual earnings. In 1973 share of the team's '-------- after operating expenses and taxes were deducted was $2,ly0, or about $200 per person. Withholding for taxes and exper and to provide funds for reinvestment absorbs about half of the gross income' of a commune. For example, the man ager of the suburban Kainbow Bridge Commune in Shanghai told us that 30 percent of the commune's gross income was required to meet the cost of its agri cultural endeavors. Another 11 percent want into a fund for capital reinvest ment, 4 percent was paid to the govern ment as tax and 3 percent went into a welfare fund for the support of new mothers and the elderly. That left 52 percent of the gross income for distribu tion to individuals in proportion to the work points earned. Such a work-point system simulta neously accomplishes two objectives. It provides a strong incentive for increased overall output in order to maxunizo the amount that is divided among the indi viduals in a household. At the same time it encourages householders and produc tion teams alike to keep their numbers from increasing, so that the maximum earnings are divided into the fewest pos sible shares. The latter constraint, of course, places the disadvantages of pop ulation growth squarely before each in dividual. This economic consideration must generate substantial social pressure in favor of small families. Other growth-limiting factors are at work. As one example, it is said that im proved health care has greatly reduced infant mortality, thus undercutting the validity of the traditional view that many births are needed to ensure the survival of even one offspring to maturity. An other example is the current program of providing care for the aged, one of the objectives of the 3 percent welfare with holding reported by the Shanghai com mune. We visited one such residence for LARGE ANIMALS SHEEP AND GOATS PIGS 100 150 200 LIVESTOCK (THOUSANDS) LIVESTOCK IN CHINA are compared according to categories; eatimatea, made by the TJJS. Department of Agrienlture, are for 1949 (gray) and 1972 (color). Sheep and goata are largely confined to grasslands end moontaint in the west and northwest. That it also time of range cattle, lumped together here with largo draft animals: horses, moles, donkeys and water buffaloes. Pigs are by far the most nnmerons of all the domestic mammals in China. the elderly, maintained by a commune near future our group was inclined to ac near Peking. Normally, we were told, cept the government's assertion that self- sach family cared for its own elders. sufficiency was achieved in 1971, when Nonetheless, some individuals inevitably some 250 million metric tons of rice, lost (heir families or became separated wheat and other major foodstuffs were frorikthem; that was the case with the 75 harvested. Furthermore, given a contin residents we saw. They were encouraged uation of the present aggressive and co to engage in handicrafts; meanwhile ordinated effort it seems that China will they were, housed, clothed and fed at be able to achieve substantial increases commune expense, and a clinic attached in agricultural production over the next to the residents looked after their health decade. Whether the increase for any needs. Custodial care of this kind, if it particular crop will be as little as 20 per were widely practiced, would diminish cent or as much as 50 percent will prob the validity of another traditional view ably prove to be a function of present that favors multiple berths: the view that yields. For example, rice yields are al one needs to have maty children in or ready high, so that an increase of 50 per der to ensure one's welfare in old age. cent in tire annual harvest will be much A third growth-limiting factor that harder to achieve than a 50 percent in may even now be undergoing a crucial crease in the maize or sorghum harvest test is the continued existenckof private landholdings throughout rural China. /'''hina nonetheless faces serious long- We were told that when the commune ^ term problems. If the country's re system was established in the\ late markable agricultural advance is to con 1950"s, between 5 and 7 percent ofSdte tinue, two of these problems must soon arable land in each commune was sfet be resolved. The first problem lies in the aside for private use. The land was par-\ current Chinese policy that emphasizes celed out at a fixed rate: each adult in a applied and decentralized agricultural household was assigned one 150th of rch for the sake of immediate in a hectare. The household children re creases in food production. Although the ceived private land too, but only the two short-term benefits of this policy are ob eldest children in each household were viously important, the policy must be eligible. If a household had more than complemented in the near future by sim two children at the time of the land ilar empntyis on more basic scientific in- assignment, or if more children arrived vestigation\ later, those children received no share. The second problem has to do with In such a family there would simply be the government's remarkably compre less private land per person. In its early hensive efforts tchretard the growth of days the allocation system probably had China's popuktionXOnly if this policy relatively little effect on rural family meets with success wnl the Chinese peo planning. Today, nearly a generation ple continue to receive urn benefits of in later, the system must act as a powerful creased agricultural production and (he social force favoring small families. accompanying advances in giving stan In summary, with respect to agricul dards that so many of Chinas people tural production in China now and in the now seem to expect. AP00024548 ENDOPLASMIC RETICULUM of a rat lirer call it enlarged 18,000 diameters in this electron micrograph made by Edward S. Reynolds of the Harvard Medieal School. Most of the biotransformitiom of foreign substances take place in this system of mem* branous tubules. The "rough** endoplasmic reticulum is covered with ribosomes, the structures in which proteins are synthesized; It appears here as fairly linear double membranes studded with black dots. The "smooth'* endoplasmic reticulum lacks ribosomes and forms a more branching, tabular network; there are patches of ft between the mitochondria, the fine-grained gray objects, at right. SMOOTH ENDOPLASMIC RETICULUM proliferates when the synthesis of tbs enzyme systems it contains is stimulated by drugs. This micrograph, alto made by Reynolds and reproduced at the same scale as the one at the top of the page, is of a cell from the liver of a rat that was treated with phenobarbitaL The drug stimu lates the synthesis of cytochrome P-450, an enryme located in the endoplasmic reticulum membranes. The cytoplasm of the cell bss expanded because of a striking Increase in the smooth endoplas mic reticulum membranes, which now fill most of the space be tween the more dispersed mitochondria, Enzyme induction by ono drag can significantly affect the metabolism, and thus the ac tivity, of other drugs that are metabolised by the same enzyme*. AP00024549 In other cases alkyl groups are removed from either nitrogen or oxygen atoms, amino groups are removed or sulfoxides are formed. Reduction and hydrolysis are also catalyzed by liver enzymes, but these reactions are less common than oxidation. ' Conjugation of a chemical is combina tion with some natural constituent of the body such as the glucose derivative glu curonic add, the amino add glycine or the tripeptide glutathione. In the pres ence of the appropriate enzyme these natural agents can combine readily with compounds that have carboxyl (COOH), sulfhydryl (SH), amino (NHa) or hy droxyl (OH) groups. Some drugs have these groups when they are in their ac tive form and are handled in the liver by conjugation; for most drugs, however, conjugation is a second step that comes after metabolism by oxidation, reduction or hydrolysis. Almost without exception the conjugated compound is devoid of pharmacological (or any biological) ac tivity. Clearly the enzyme systems that cata lyze these reactions were not invented by the mammalian organism in order to cope specifically with drugs or novel pollutants. Their basic physiological role has presumably been to metabolize en dogenous substrates: substances nor mally present in the body. For exam-ple, liver microsomal enzymes oxidize steroid hormones, cholesterol and fatty acids. The products of these oxidations may then be conjugated (with glucuron ic add, for instance) and excreted. Bili- t ... t- LTVER is a primary site of metabolism of substances entering the body tbrongh the alimentary tract. All the fine blood Teasels that absorb nutrients and other substances through the wall of the intestines come together and enter the liver through the portal vein. The liver's supply of oxygenated blood from the heart enters through the hepatic artery. After patting tbrongh the ainnaoida of the liver (see illustration on pagt tf) the blood is collected by the hepatic veins, which feed inte the vena cava. The liver secretes bile, which is collected by the bile duct, stored in the gall bladder and emptied into the duodenum, the first segment of the small intestine. AP00024550 r.. r * T-nVtin a product of the oxidation of heme, the red pigment of hemoglobin, is an example. It is nonnally prepared for excretion by glucuronide conjugation. The rate of formation of glucuronides is generally low in newborn infants, how ever, because of a deficiency of the en zyme glucuronyl transferase. As a result bilirubin may not be conjugated and excreted at an adequate rate. Excessive amounts of it can then accumulate and cause grave damage to the brain, a con dition called kemicterus. It has been apparent for some time that the human fetus and die newborn infant are far more sensitive than adults to many drugs. A number of drugs can pass across the placenta, so that obste tricians need to exert care in administer ing them to an expectant mother. Bar biturates or morphine given to a woman during childbirth can be stored in the infant's tissues and cause respiratory de pression and occasionally death. The ex planation for the sensitivity of infants to drugs has emerged from a number of re ports in recent years on the maturation of the capacity to metabolize and con jugate drugs. These studies make it clear that die capacity to oxidize and conju gate is negligible in the mammalian fetus and newborn animal and increases after birth at a rate that varies with the spe cies, the type of reaction and the drug. Impaired drug metabolism increases the intensity and duration of drug action. For example, newborn mice treated with a dose of the hypnotic drug hexobarbital equivalent to 10 milligrams per kilogram of body weight sleep more than six hours, whereas adult mice given 10 times as large a dose sleep less than an hour. 11 'he adverse effect of drugs on new^ bora infants as a result of inefficient metabolic conversion is illustrated by the striking "gray-baby syndrome" in infants that may come a few days after treat ment with the antibiotic chlorampheni col: abdominal disterkion, respiratory difficulty, cyanosis (blue skin color as a result of insufficient oxygenation of the blood) and shock. The condition is ap parently the result of deficient metabo lism of the antibiotic and deficient con jugation with glucuronic arid. In adults about 90 percent of the antibiotic is ex creted in the urine in the form of con jugated metabolites within 24 hours after it has been orally administered; only a small amount is excreted un changed. In comparison a 10-day-old infant in one study excreted less than 50 percent of the drug in 24 hours. Rates of drug metabolism are very dif GENERAL CIRGII1 ATIDN INTESTINE PATHWAY of a drug that la tranifonned in the liver la thown schematically. The drug (colored arrows) enters the liver through the large portal vein, passes into the general circulation, has Its effect on the target organ and eventually returns to the liver. On each passage through the liver a fraction of the drug is converted, usually into inactive metabo lites {black arrows). The metabolites may he carried by the bile into the Intestines for excretion or may pats through the circulation to the kidneys, to be excreted into the mine. ferent in different species, and the effec tive dosage varies accordingly. The anti inflammatory agent phenylbutazone is metabolized slowly in man; its half-life in die plasma averages about three days. In the horse, the dog, the rabbit, the rat and the guinea pig, however, the drug is metabolized much more quickly and the half-life ranges from three to six hours. (Knowledge of the rapid rate of metabo lism of phenylbutazone in the horse is of practical importance because the drug is administered to treat arthritic conditions in racehorses.) A dose of hexobarbital (adjusted for the body weight of the ani mal) that makes mice sleep for an aver age of 12 minutes puts rabbits to sleep for 49 minutes, rats for 90 minutes and dogs for 315 minutes. When the enzy matic oxidation of hexobarbital by microsomes from the liver of these animals was measured, the fastest oxidation was carried out, as expected, by microsomes from the mouse; the rates of oxidation by microsomes from the rabbit, the rat and the dog were proportionately lower. Dif ferences in the rates and patterns of drug metabolism in the various animals seem to explain most of the species differences in effect, but there may also be differ ences in the distribution of the drug, the response of the target tissues and excre tion. Even among human patients there are marked individual variations in the metabolism of drugs that are bandied primarily by microsomal enzymes. The variability causes some patients to me tabolize a drug so quickly that thera peutically effective blood and tissue lev els are difficult to achieve and others to metabolize the drug so slowly that they suffer toxic effects. It can therefore be difficult for the physician to predict just what dosage of some drugs will provide a safe and therapeutic effect in an in dividual patient. Very large individual differences have been noted in the me tabolism of the ooumaxm-derivative an ticoagulant drug Dicumarol; the halflife can vary from seven to 74 hours. That makes it hard to predict how much of the drug will provide the desired an ticoagulant effect in a patient Marked variations have also been observed in the metabolism in man of phenylbuta zone and of diphenylhydantoin, a drug that is administered to control epileptic seizures. Genetic factors may play an impor- AP00024551 -S' tant role in the metabolism of drugs, as Elliot S. Vesell of the Pennsylvania State University College of Medicine has shown. The large individual differences in Dicumarol half-life persist to some ex tent when fraternal twins are compared but almost disappear in identical twins, and similar results have been reported for other drugs studied in twins. The antituberculosis drug isoniazid has been the subject of a number of investigations covering different populations. Isoniazid is metabolized in man primarily by an acetylation reaction catalyzed by the enzyme N-acetyl transferase. Soon after the drug was introduced it became apparent that individuals vary in their ability to acetylate it. The distribution of the acetylation rates for a group of subjects tends to be blmoda.1, that is, plotting the rate against the number of patients exhibiting each rate yields a curve with two peaks. Apparently there are two classes of individuals: those who acetylate isoniazid rapidly and excrete the drug primarily as acetylisoniazid and those who acetylate and excrete the drug more slowly and excrete more of it In an unchanged form. The frequency of rapid inactivators is about 90 percent in a population of Eskimos or of Japanese, whereas among both whites and blacks in North America there are about as many slow inactivators as there are rapid ones. A clear relation has now been estab lished between this variation in acety lation rate and the incidence of isoniazid toxicity. Several studies have shown that the slow inactivators are more suscep complex of protein and heme, the iron- tible to a toxic effect of isoniazid: a dis containing ring structure that is the oxy order of the peripheral nerves caused by gen-binding component of hemoglobin. a specific vitamin B deficiency resulting Like hemoglobin, the various cyto from an interaction of the drug and die chromes serve to bind oxygen, which vitamin. they deliver to their substrates in such processes as cell respiration. Cytochrome r Phe first indication that microsomal . P-450 gets its designation from the fact enzymes were special kinds of com- that in the reduced form it binds carbon plexes was developed some 20 years ago monoxide and then absorbs light most by Gerald C. Mueller and James A. Mil intensely at a wavelength of 450 nano ler at the University of Wisconsin. They meters. The amplitude of the absorbance discovered that compounds known as peak is the basis of quantitative studies aminoazo dyes could be oxidized (spe of the enzyme. In the mixed-function cifically, N-demethylated) by liver mi- oxidases cytochrome P-450 serves as the crosomes and that the oxidation required molecular oxygen (03) and the coenzyme NADPH (reduced nicotinamide-adenine dinucleotide phosphate). Soon afterward Bernard B. Brodies group at the Na tional Heart and Lung Institute showed that there was a similar requirement for the oxidative metabolism of a number of drugs. In 1957 Howard S. Mason of the University of Oregon Medical School terminal oxidase: it accepts electrons passed along by several intermediates, binds oxygen and then delivers the oxy gen to oxidize its substrate and (usually) produce water [see illustration on page 30], In addition to NADPH the system includes the enzyme cytochrome P-450 reductase. Another heme protein, cyto chrome bB, is present in liver microsomes and may participate in drug oxidations, proposed that such oxidations are cata lyzed by a class of `mixed-function oxi dases": enzyme complexes that require but its precise function is not dear. Nor is it known just how the various components of the mixed-function oxi oxygen and NADPH, are nonspecific (they catalyze die oxidation of different kinds of compounds) and catalyze die dase complex are arrayed in the tubular membranes of the endoplasmic reticu lum. It appears that the reductase and consumption of a molecule of oxygen for each molecule of the drug or other sub strate, with one atom of oxygen appear ing in the metabolized substrate and the other atom usually combining with two cytochrome bs are on the outside of the membranes, with cytochrome P-450 in the deeper layers. The enzyme glucuronyl transferase, which catalyzes the most important conjugation reaction, is also hydrogen atoms to form water. The key enzyme of these oxidases is cytochrome P-450. A cytochrome is a present within the membranes. The drug or other fat-soluble compound to be transformed is presumably bound and BRANCH OF PORTAL VEIN BRANCH OF HEPATIC ARTERY SINUSOIDS CENTRAL VEIN BILE DUCTULE LTVER LOBULE, the functional unit of liver tissae, is defined by branehei of the portal vein, the hepatic artery, the bile duct and lympbatie vends, which run totether through the tissue, outlining the lobules. (The lymphaties are sot shown here.) The venous blood, with its nutrients and other ingested substances, and the oxygenated arterial blood enter the fine sinusoids and perfuse the BILE CANALICULUS Krer cells, which carry out the metabolic functions discussed in the text. The blood drains into a central vein in each lobnle and thence to sublobnlar and hepatic veins. The liver ceils also secrete bile, which is collected by bile canaliculi that feed into bile duct ules and eventually into the main bile duct. This two-dimension al view of part of a lohule Is highly simplified and diagrammatic. AP00024552 metabolized by the mixed-function oxi date system that contains cytochrome P-450 and then is often converted into a highly water-soluble compound by conjugation with glucuronic add; die transformed product passes into the lu men, or central channel, of the tubular membranes and is excreted from the cell into the bile or the bloodstream. The cytochrome P-450 mixed-funo- tion oxidase system has oome to be rec ognized as having a central role in the body's defense against chemical agents, whether they are normal body constitu ents or are introduced from the environ ment. It is now clear that the system is responsible for the detoxification of many of the potentially harmful environ mental pollutants. The system is highly indudble, that is, its activity can be greatly increased by exposure to a wide variety of environmental agents and drugs that act as substrates for the sys tem. Such chemicals stimulate the syn thesis of cytochrome P-450 and other components of the complex. More than 200 steroid hormones, drugs, insecti cides, carcinogens and other foreign chemicals are now known to stimulate drug metabolism in experimental ani mals, and many of them have been shown to do die same tiling in man. On the other hand, some substances (such as lead and other heavy metals) inhibit the mixed-function oxidase system, al though they are fewer In number and less diverse than the inducers. The P-450 system is also the site of much competi tive interaction among drugs and other chemicals that are undergoing transfor mation. NUCIEI, UNBROKEN CELLS AND DEBRIS CELL MEMBRANES AND MfrOCHONDRIA LIVER TISSUE HOMOGENIZEO HOMOGENATE CENTRIFUGED 10 MINUTES AT 600 g SUPERNATANT I CENTRIFUGED 20 MINUTES AT 8.000 g \|/ SUPERNATANT II rPhe consequences of this inducibility, inhibition and competition have im portant implications in drug therapy. Pa tients are often given several drugs at the same time. Certain combinations can have unpredictable and often undesir able effects if one drug inhibits or stim ulates the metabolism of another or competes with it. For example, phenyl butazone, the coumarin anticoagulants and chloramphenicol compete with the metabolic inactivation of tolbutamide, a drug given to reduce the blood-sugar level in diabetics; the competition can lead to excessive tolbutamide activity and thus to serious hypoglycemia, or low blood sugar. Phenobarbital is a prime example of a drug that has a different effect: enzyme induction. When rats are treated with phenobarbital, there can be a three- to fourfold increase in the microsomal con tent of cytochrome P-450 and a two- MICROSOMES SUPERNATANT II CENTRIFUGED 00 MINUTES AT 100,000 : \/ LTVER MICROSOMES, the membrane etroetures that contain most o the liver enzymes encaged in detoxification, are isolated by epiuning homogenised liver tissue st successive!} higher speeds,'which produce successive!} higher gravity (g) forces, in a centrifuge. Homogenised liver tissue is centrifuged for 10 minutes at <00 g; a pellet of dense material, primaril} whole cells, cell debris and cell nuclei, collects at the bottom of the tube. The supernatant, or liquid portion, it centrifuged more strongly, isolating less dense structures such as mitochondria and pieces of membrane. When supernatant II is spun at very high g forces in an nltrseentrifuge, the microtomes are separated from the cytosol, or cell fluid. AP00024553 told Increase in the reductase, and drugs such as methadone and ethylmorphine are metabolized three or four times as fast In accordance with these experi mental observations chronic adminis tration of phenobarbital to patients decreases the effects of many drugs by hastening their inactivation. Sedative doses of the drug reduce the concentra tion in the blood plasma of phenylbuta zone, the analgesic antipyrine and the coumarin anticoagulants; it also de creases their pharmacological actions. Antiooagulant therapy is particularly sensitive to dosage. A patient who is sat isfactorily maintained on a given coumarin-drug dosage while he is being given phenobarbital as a sedative may hemor rhage when the phenobarbital is with drawn because the P-450 system is no longer being induced and the antlcoagu- 4 BIOTRANSFORMATIONS carried out in the liver include oxida tion, reduction, hydrolysis and conjugation, examples of which are given here. The molecular aites of each reaction are indicated in color. In a few cases transformation eonverta an inactive iorm of a drug (such as Prontosil) into an active form (sulfanilamide), hut most of the reactions lead to inactivation. Conjugation with a nat- AP00024554 iV lanl is therefore being inactivated more cytochrome P-450. The habitual con constitute another class of environmen slowly.' jT^V sumption of alcohol therefore stimulates tal pollutants that have been shown ; Alcohol is converted into acetalde the metabolism of a wide variety of to induce microsomal enzymes. The hyde largely in the liver, perhaps to drugs. This helps to explain why heavy PCB's are lubricants, heat-exchange some extent by the mixed-function oxi drinkers are less affected than other peo fluids, insulators, plasticizers for paints dase system: heavy drinkers are found ple by barbiturates and other sedatives-- and plastic compounds and a major com to have an increased concentration of when they are sober. A single very large ponent of the lens-immersion oil used In dose of alcohol taken together with an microscopy. Whereas some of the con TRANSFORMED PRODUCTS other drug, on the other hand, inhibits sumer-product applications have recent the drug's metabolism, presumably by ly been curtailed, the immersion oils are . competing with the drug for the ap still handled daily by many laboratory propriate enzymes. This effect, in addi workers. PCB's have been found in the AZO tion to the depressant effect of alcohol tissues of numerous bird and fish species on the central nervous system, helps to and in human fat and milk, although the NHt explain the enhanced sensitivity to bar- route of entry into the human body has biturates and other sedatives of a person not been accurately determined. In re who has been drinking heavily. The syn cent experiments we have been able to ergistic actions of alcohol and sedatives show that the application of pure PCB's in the brain can cause death. or of microscope immersion oil to the The indudbility of microsomal en skin of experimental animals in very NITRO REDUCTION KH, zymes by drugs suggests a form of ther small amounts (one microliter) causes a apy for certain conditions in which marked increase in mixed-function oxi normal body constituents ordinarily me dase activity and reduces the pharma tabolized by such enzymes are present cological effect of zoxazolamine, a mus in excessive.amounts. Long-term admin cle relaxant, and of hexobarbital in the HDC | istration of phenobarbital, for example, live animal. These findings suggest that can lower the concentration of bilirubin, trivial skin exposure to chemicals can HOCttjCHNHCCHOf the pigment that produces jaundice, in have significant and perhaps harmful the blood of patients with chronic ob biological effects in man. struction of bile flow in the liver. The A number of chemicals to which hu excessive bilirubin levels that are nor man beings are regularly exposed have mally observed in infants after birth can been identified as chemical carcinogens, also be markedly reduced if die mother that is, they cause cancers when they is given a small dose of phenobarbital are applied to the skin of experimental for a number of days before delivery. animals or otherwise administered to Presumably the drug crosses the placen them. Clearly a factor that inhibits or ta and stimulates the conjugating en stimulates the metabolism of such com zyme system that is ordinarily slow to pounds may affect the development of develop in the fetus and the newborn human cancers. Benzpyrene, benzan + MOCHjOVaCfH,), infant thracene and similar polycyclic aromat ic hydrocarbons are among the most TJalogenated hydrocarbon insecticides ubiquitous carcinogens: they sure present such as DDT are potent stimulators in tobacco smoke, in polluted city air of drug and steroid sex-hormone me and in charcoal-broiled and smoked tabolism in mammals and in birds; the foods. The polycyclic hydrocarbons are breakdown of sex hormones explains in metabolized by a mixed-function oxidase part the devastating effects of DDT on enzyme that in this case is called aryl reproduction in some bird populations. hydrocarbon hydroxylase because of the OWi The minimum exposure to DDT that particular oxidation It catalyzes. Like will stimulate the metabolism of pento other microsomal oxidations, this one barbital and decrease its hypnotic action . requires NADPH and molecular oxygen, in experimental animals is one that re but the terminal oxidase of the system <*0. sults in concentrations of from 10 to 15 induced by the polycyclic hydrocarbons micrograms of DDT per gram of fat; is somewhat different from the oxidase that is a level commonly found in human induced by drugs. The catalytic proper MHCOCH, fat tissues. Among the other insecticides ties of the cytochrome are changed^ as that induce microsomal enzymes in ex are its spectral properties: the absorb perimental animals are chlordane, aldrin ance maximum of the complex of carbon and dieldrin. It is interesting that pipero- monoxide with the reduced cytochrome nyl butoxide, a synergist that was added is at 448 nanometers rather than at 450, so^h, -|- iv) to insecticides to Inhibit the enzymatic and so the enzyme is designated cyto defenses insects had developed against chrome P-448. Apart from the polycyclic DDT and its chemical relatives, also in hydrocarbons, only one other class of nrtl rahitance inch at glucuronic acid often hibits the activity of the microsomal en compounds has so far been noted to in follows metabolism of drags and facilitates zymes in the mammalian liver. duce the formation of cytochrome P-448: excretion. Not all the byproducts are shown. The polychlorinated biphenyls (PCB's) we have found dial the PCB's induce AP00024555 -'/KCfT- J^ ^ .V^'T*^r. r - / > * ' -- r- % u- V- itf *. some of the newly identified cytochrome along with cytochrome P-450. This sug gests the possibility, for which there is developing evidence in animals, that the PCB's too may have carcinogenic prop erties. The aryl hydrocarbon hydroxylase system has been the subject of intensive investigation as a possible link in the causation of some cancers. The reason is that rather than detoxifying its polycy clic-hydrocarbon substrates it seems to make some of them more toxic: interme diates of polycyclic-hydrocarbon metab olism such as epoxides are more active in the malignant transformation of tis sue-cultured cells than the parent prod ucts are. Moreover, the enzyme system for carcinogen metabolism is available at many sites in the body that are ex posed to polycyclic hydrocarbons and is induced not only in the liver but also in the gastrointestinal tract, the kidneys, the skin and the lungs. Cigarette smok ing markedly induces aryl hydrocarbon hydroxylase activity even in the human placenta, as Allan H. Conney, Richard M. Welch and their colleagues showed at the Wellcome Research Laborato ries. Little or no such enzyme activity was found in placentas from nonsmokers. We have investigated the cytochrome P-448 system in human skin and have found that there is marked variability in the hydroxylase activity of different skin samples and that incubation of skin in tissue culture with the polycyclic hydro carbon benzanthracene induces more enzyme activity. A skin biopsy is easy to do, so that assaying skin aryl hydrox ylase in die presence or absence of polycyclic hydrocarbons may provide a convenient test of individual differ ences in the capacity to metabolize cer tain environmental carcinogens. The possibility that induction of aryl hydro carbon hydroxylase is of considerable significance in chemical carcinogenesis is suggested by file results of some re cent experiments performed by Gott fried KeUermann and his associates at the University of Texas, They found that the indudbility of such hydroxylase activity in lymphocytes is significantly greater in cigarette smokers who have lung cancer than it is in healthy nonsmokers. OXIDATION OF A DRUG by tbe enzyme cytochrome P-450 i visualized here as a sequen tial process. The enzyme (1) is a complex of a protein and the oxygen-binding compound heme, which contains an iron atom that it initially in the ferric (Fe+ + +) form (open circle). The cytochrome binds the drug (2). Then (S) the enzyme cytochrome P-450 reduc tase, utilizing the coenzyme NADPH, reduces the iron of the heme to the ferrous (Fe+ +) form (black dot), in which it can hind a molecule of oxygen (4). It supplies one atom oi oxygen to oxidize the drug and one generally to form water, in the process reverting to its oxidized form (5). The drug, oxidized and in most cases inactive, it thereupon released (6). Tt is clear that the liver's mixed-function oxidase system is implicated in a number of processes affecting human health. The ability of a drag or other foreign substance to stimulate the me tabolism of another drag by the system may explain some of the adverse drug reactions observed in clinical practice. Drug interactions have been well docu mented in anticoagulant therapy; more research is required to explore other drug interactions, since most patients are given several drugs at the same time. The ability of environmental pollutants to modify drug action is now under ac tive investigation. It is clear that insecti cides, for example, stimulate drug-me tabolizing enzymes and that heavymetal substances such as lead and methyl mercury inhibit the enzymes; the clinical significance of these effects may be appreciable in populations that are exposed occupationally to such agents. Finally, recently acquired evidence seems to indicate that the induction of aryl hydrocarbon hydroxylase in human tissues by polycyclic hydrocarbons may play a significant role in chemical car cinogenesis. Research on this biological action of carcinogens may provide im portant leads toward predicting tbe sus ceptibility of individuals to certain kinds of chemically induced cancer. i 'r vyrr'.V^Ajrr AP00024556 SLAVERY IN ANTS V"' Certain species of ants raid the nests of other species for ants to work in their own nest. Some raiding species have become so specialized that they are no longer capable of feeding themselves J, P ,-->-rv.u w y '/' by Edward O. Wilson .- -v.-- ',> .- I't-lin i!iw i!WW !>: 'tnw w ifftsw iy w r ?ngrwr-- he institution of slavery is not When Polyergus ants launch a raid, Tunique to human societies. No however, they are completely transfewer than 35 species of ants, formed. They swarm out of the nest in constituting six independently evolvaed'solid phalanx and march swiftly and Polyergus ants. The workers launch into the round of tasks for which they have been genetically programmed, with the slave makers being the incidental bene- groups, depend at least to some extent directly to a nest of the slave species, ficiaries. Since the slaves are members of on slave labor for their existence. The They destroy the resisting defenders by the worker caste, they cannot reproduce, techniques by which they raid other ant colonies to strengthen their labor force rank among the most sophisticated be havior patterns found anywhere in the insect world. Most of (he slave-making puncturing their bodies and then seize and carry off the cocoons containing the pupae of worker ants. When the captured pupae hatch, the workers that emerge accept their cap- In order to maintain an adequate labor force, the slave-making ants must periodically conduct additional raids. It is a remarkable fait that ants of slave-making species are found only in ant species are so specialized as raiders tors as sisters; they make no distinction cold climates. Although the vast majorthat they starve to death if they are de between their genetic siblings and the ity of ants live in the Tropics and the prived of their slaves. Together they dis play an evolutionary descent that begins with casual raiding by otherwise freeliving colonies, passes through the de velopment of full-blown warrior societies and ends with a degeneration so ad vanced that the workers can no longer even conduct raids. Slavery in ants differs from slavery in human societies in one key respect: the ant slaves are always members of other completely free-living species that them selves do not take slaves. In this regard the ant slaves perhaps more closely re semble domestic, animals--except that tile slaves are not allowed to reproduce and they are equal or superior to their captors In social organization. 1 nphe famous Amazon ants of the genus Polyergus are excellent examples of advanced slave makers. The workers are strongly specialized for fighting. Their mandibles, which are shaped like minia ture sabers, are ideally suited for punc turing the bodies of other ants but are poorly suited for any of the routine tasks that occupy ordinary ant workers. In deed, when Polyergus ants are in their home nest their only activities are beg ging food from their slaves and cleaning themselves ("burnishing their ruddy ar mor," as the entomologist William Mor ton Wheeler once put it). RAID BY SLAVE-MAKING AMAZON ANTS of the species Polyergus rufesemu (light color) against a colony of the slave species Formica fusca (dark color) is depicted. The fusca ants mslce their nest in dry soil under s stone. The raiding Amazon ants kill resisting AP00024557 -s 1*?.' i PHENOBARBETAL induces cytochrome P-450 and that stimulate! than that of microtomes from an untreated rat (black curve). the metabolitm of drugs oxidised by that enxyme. The curvet (ie/t) More oi the test drug ethylmorphlne it therefore metabolised (cen 1 show the light-absorhance spectrum, with t peak at 450 nanome ter) by liver tissue from a phenobarhital-treated rat (colored bar) ters, characteristic of reduced cytochrome P450 complexed with than by the same amount of tisane from an untreated rat (gray). r carbon monoxide, the usual means of identifying and quantifying Similarly, phenobarbital speeds np metabolism of the drag antlpy- , the enxyme. In a rat treated with phenobarbital the cytochrome P- rine in man (right): antipyrine level In plasma falls faster after 450 content of microtomes (colored curve) is suhitantiaily higher phenobarbital treatment (colored curve) than ordinarily (black). i i I CARCINOGENS induce a different specie! of cytochrome, cyto chrome P-448. The curvet (left) compare the absorbance of micro tomes from the liver of a rat treated with the carcinogen methyleholanthrene (color) with the absorbance of control microsomet (black); the induced cytochrome hat a peak at 448 nanometers. Methylcbolanthrene enhances metabolism of another carcinogen, bentpyrene (center) i the microsomet of treated rats produce more metabolite (color) than control microtomes (gray) do. Similarly, treating human skin in tissue culture with the carcinogen henxanthrteene stimulates skin-cell metabolism of bentpyrene (right). ' -M -'-l AP00024558 s Geifand. G. A. Galau. R. J. Britten. . H. Davidson. Chromosoma 51. 225 (1975); R. C. Ansecu-. E. H. Davidson. R. J. Britten. Cell 6. 290975). 25. R. A. Firtel. A. Jacobson. J. Tuchman. H. F. Lodish. Genetics 78,3550974); R. A. Find and K.Kmdle.Ce//5.401 (1975). 26. E. Pays and A. Ronsse.Bioc/iem. Biophvs. Res. Commun. 62.862 (1975). 27. C. W. Schmid and P. L. Deininger, Celt 6. 345 (1975). 28. E. H. Davidson. B. R. Hough, W. H. Klein. R. J. Britten.ibid. 4.217(1975). 29* D. C. Rcanny.7. Theor. Biol. 49,461 (1975). 30. J. S. Semancik and J. L. M. C. Geelen. Nature tLondon) 256.753 0975). 31. F. Crick, ibid. 234, 25 0971); C. D. Laird. Chromosoma 32. 378(1971); H. O. C&llan. Proc. R. Soc. London Ser. B 181. 19 (1972); R. Kavenoff and B. H. Zitnm. Chromosoma 41.1 (1973); K. Brasch and G. Setlerfield. Exp. Cell Res. S3, 175 (1974); H. G. Callan. Cold Spring Harbor Symp. Quant. Biol. 38. 195 0974); D. E. Comings. in Chromosomes and Cancer. J. German, Ed. (Wiley. New York, 1974). pp. 95-133; D. E. Coinings and T. A Okada. Cold Spring Harbor Symp. Quant. Biol. 38. 145 (1974); J. H. Taylor, M. Wu, L. C. Erickson, ibid., p. 225. 32. D. A. Wilson and C. A. Thomas, Jr.. J. Mol. Biol. 84. 115 (1974); R. D. Wesley, Proc. Natl. Acad. Sci. V.S.A. 72.678 (1975). 33. C. F. Garon. K. W. Beny. J. A. Rose. Proc. Natl. Acad. ScL U.S.A. 69,2391 (1972); J. Wolfson and D. Dressier, ibid., p. 3054; R, j. Roberts, J. R. Arrand. W. Keller, ibid. 71.3829 (1974); P. H. Steenbergh. J. S. Sussenbach. R. J. Roberts, H. S. Jansr,/. Virol. 15, 268(1975). 34. T. Cavalier-Smith. Nature (London) 250, 467 0974). 35. A. J. Bateman.ibid. 253.379(1975). 36. H. Traut and W. Scheid. Mutat. Res. 17, 335 0973); W. Scheid and H. Traut, ibid. 25. 321 (1974); Can. J. Genet. Cytol. 17, 129(1975). 37. 0. K. Haapala and M.-O. Soyer. Herediias 80. 1850975). 38. E. J. DuPraw. DNA and Chromosomes (Holt. , Rinehart and Winston, New York. 1970). 39. M. J. D. White. Ammo/ Cytology and Evolution (Cambridge Univ. Press. Cambridge. 1973); D. E. Comings and T. A. Okada. Cold Spring Har bor Symp. Quant. Biot. 38. !45 0974). 40. E. G. Diacumakos, S. Holland. P. Pccora, Na ture (London) 232, 33 (1971); P. M. Kraemer. D. F. Petersen. M. A. Van Dilta. Science 174, 714 0971); L. K. Schneider. J. Cel! Biol. 56. 608 (1973): T. Ashley and E. B. Wagenaar. Can. J. Genet. Cytol. 16. 61 (1974); H. M. Golomb and G. F. Bahr. Exp. Cell Res. 84 . 79 (1974); J. Y. Jaffr&y and A. Geneix. Humangenetik 25. 119 (1974); S. Takayami, Exp. Cell Res. 91, 408 (1975). 41. G. D. Lauer and L. C. Klotr. J. Mol. Biol. 95. 309(1975). 42. A. B. Burdick. B. L. Libbus, V. Athanassiou, A. Falek, Can. J. Genet. Cytol. 16,481 (1974). 43. H. Fraenkel'Conrat and R. R. Wagner. Eds., Comprehensive Virology (Plenum, New York, 1974), vol. l.pp. 39-40. 44. K. W. Buck and G. F. Kempson-Jones./. Gen. Virol. 18.223 0973). 45. L. V. Crawford et al.Jbld. 4.37 (1969). 46. G. Askai, C. Christiansen. H. Emo. J. Gen. Microbiol. 75,283 (1973). 47. E. N. Sokurova. Mlkroblologlya 41. 1020(1973). 48. B. A. Whitton. N. G. Can. 1. W. Craig. Proto plasma 71, 325 (1971). DNA/G vanes with growth rate; value given is the lowest extreme. 49. P. R. Whitfield. Aust. J. Biol. Set. 6. 234 (1953). Lowest of four determinations; mean - 2.7 x 10"* pg. 50. T. Iwamura and S. Kuwashima. Blochim. Biophys.Acta 174.330 0 969). 51. A. E. Mirsky and H. Ris.7. Gen. Physiol. 34. 4510951). 52. J, E. Sulston and S. Brenner. Genetics 77, 95 (1974). 53. E. Jost and M. Mameli. Chromosoma 37. "Ol (1972). 54. C. K. Leach and J. M. Ashworth. J. Mot. Biol 68.35 0972). 55. R. Hincgardner and D. E. Rosen. Am. Nat. 106. 621 0972). 56. R. Hihcgardncr. Comp. Biochem. Physiol. A 47. 447(1974). 57. , Comp. Biochem. Physiol. B 49, 219 (1974). 58. E. L. Rheinsmith. R. Hinegardner. K. Bach* mann./biV. 48,343 0974). 59. H. S. Shapiro, in Handbook ofBiochemistry. H. A. Sober. Ed. (Chemical Rubber Company, Cleveland, Ohio, 1970), p. H-107. 60. H. Swift and R. K\t\nfe\6.Physiol. Zool. 26.301 (1953); B. Daneholt and J.-E. Edstrfim. Cy togenetics 6. 350 (1967); E. M. Rasch. in In troduction to Quantitative Cytochemistry, G. L. Wied and G. r. Bahr, Eds. (Academic Press, New York, 1970), vol. 2, pp. 357-397; J. M. Lagowski, M.-Y} W. Yu. H. S. Forrest, C. D. Laird. Chromosoma 43. 349(1973); A, Li'ma-deFaria. T. GustafTson. H. Jaworska. Heredltas 73, 119 (1973); E. M. Rasch, Chromosoma 45. 1 0974). 61. A. von Hofsten. Physiol. Plant. 16 . 709(1963); K. Esser and R. Kuenen. Generics of Fungi (Springer-Verlag. New York. 1967); J. N. Bicknell and H. C. Douglas. J. Bacteriol. 101. 505 (1970); C. Christiansen. A. L. Bak, A. Stenderup. G. Christiansen. Nature (London) New Biol. 231, 176 (1971); R. A. Firtel and J. Bonner, J. Mol. Biot. 66, 339 (1972); J. Mohberg, K. L. Babcock. F. B. Haugli. H. P. Rusch. Dev. Biol. 34. 228 (1973); S. K. Dutta. Nucleic Acids Res. 1, 1411 (1974); R. L. Dusenbery, Biochim. Biophys. Acta 378, 363 (1975). 62. We thank K. H. Thompson for statistical analy sis of data and R. 3. Setlowand H. H. Smith for helpful comments on the manuscript. Supported by the Energy Research and Development Ad ministration. NEWS AND COMMENT Cancer Institute; Expert Charges Neglect of Carcinogenesis Studies A prominent scientist has resigned from directorship of a key program in the National Cancer Institute for reasons which, if well founded, could provoke a serious perturbation in the agency's affairs. The scientist, Umberto Saffiotti, heads the NCI's program in chemical carcinogenesis, a subject whose impor tance has been increasingly acknowl edged by cancer epidemiologists, by regulatory agencies, and in Congress. A threatened fragmentation of the pro gram which could prolong the public's exposure to carcinogens is one of the reasons behind his decision to quit. NCI director Frank Rauscher pays tribute to Saffiolti's scientific expertise but regards the issue of his resignation as the result of a difference in approach to the management of certain programs un der his control, which Rauscher believes could have been pushed ahead faster, Saffiotti believes that the carcinogen esis program has long been denied the 7 MAY 1976 manpower necessary to keep pace with its growing responsibilities. It has only 20 percent more staff than in 1968. but al most 8 times the amount of money to ad minister. This year the program received only 3 of the 79 new staff positions as signed to the NCI, although Rauscher told the House Appropriations Commit tee that he was giving Saffiotti the highest priority possible. With a current budget of $47 million, the program conducts basic research on chemical carcinogene sis as well as developing bioassay tests for carcinogens. Saffiotti also feels that he and other col leagues with relevant expertise have been excluded from a series of decisions on chemical carcinogenesis, the most re cent being the announcement of a Na tional Clearinghouse on Environmental Carcinogenesis, on which he says he was not consulted until a late stage. The final straw for Saffiotti was a recent decision to split away from his program the re sponsibility for developing bioassay tests for chemical carcinogens. The move will, in his view, compromise the scientif ic credibility of the tests, delay their being put into action, and increase the time that people will be exposed to the chemicals the tests may show to be car cinogenic. Although he has been asked to remain as director for the research part of the carcinogenesis program, Saffiotti has chosen to resign altogether from the pro gram management, lest he seem by staying to concur with the decision on the bioassay tests. He plans to take up full time research in his laboratory at the NCI. "I am glad to call it quits rather than endorse a mode of operation 1 dis agree with." he told Science in an inter view last week before announcing his res ignation. Saffiotti adds that a fundamental rea son for resigning is his belief that active scientists have very little voice in setting policy or priorities in his division of the NCI, and that the division is being run by managers with the help of scientists rath er than the other way around. Because of the growth of successive layers of bu reaucracy. whose actions are not ac countable to detailed peer review by sci entists. Saffiotti says, "There seems to be a growing gap between the top policy making decisions of the institute and the 529 AP00024559 expertise which is needed to make these complicated value judgments." Saffiotti's resignation is likely to be taken seriously because he is well known outside the NCI. He has been an active researcher in chemical carcinogenesis for 20 years and has headed the NCI pro gram since 1968. Evaluations of data by his program staff have played an impor tant part in regulatory decisions banning various chemicals, such as the pesticides aldrin and dieldrin. As the NCI's leading expert on chemical carcinogenesis, he has given frequent testimony before con gressional committees and is a key figure in the field. NCI director Rauscher calls Saffiotti a "superb scientist" but ascribes the rea sons for his resignation to a difference in managerial approach. In a briefconversa tion before going out of town. Rauscher told Science he had been unable to as sign as many staff positions as he would have liked to the carcinogenesis program but that Saffiotti could have reassigned his own staff to the bioassay area. The NCI position has been discussed in great er detail by deputy director Guy Newell and by James Peters, director of the divi sion of cancer cause and prevention, to which the carcinogenesis program be longs. Part of Saffiotti's unhappiness about the role of scientists in NCI policy-mak ing relates to the fact that Peters, as divi sion director, has to make decisions af fecting chemical carcinogenesis research even though he has no scientific research experience. Peters, a veterinarian by background, notes that he has a master of public health degree in epidemiology and w orked for 5 years as Saffiotti's dep uty in the carcinogenesis program. Asked if he has the expertise to make de cisions about carcinogenesis research, Peters says that he feels perfectly quali fied to do so: "While I don't have the bench-side experience that Saffiotti has, I am in a better position to make deci sions because from my position as divi sion director I have an overall view of the program's needs within the division and of its relationship to other agencies." The specific issues which prompted Saffiotti to resign axe hard to assess be cause he and the NCI top management are to some extent talking past each oth er. Saffiotti discusses the issues in terms of their impact on research, Rauscher and Peters in terms of managerial and public relations aspects. A principal point of disagreement concerns the man agement of bioassay tests for carcino gens, both in vivo tests in which a chem ical is fed to animals for an extended peri- 530 od, and in vitro tests in which cell cultures are the test medium. The in vi tro tests, still in the process of devel opment, promise to be very much quick er and cheaper than the $100.000-per- chemical tests in animals. What triggered Saffiotti's resignation was a decision to split the validation of the in vitro tests from the basic research, leaving the research in Saffiotti's pro gram and moving the validation to Pe ters's office. The project will not receive any more staff in Peters's office, but. ac cording to Rauscher. it will be "in an or ganizational position of higher visibility and emphasis." Rauscher says he has promised Congress that the NCI would press ahead with the in vitro bioassays. "Umberto feels this has been done. I feel that more could be done. Some of the ad vice I have been getting from some people is that he has not been moving fast enough," Rauscher says. Saffiotti's position is that he has been moving ahead as fast as he could, but that the tests are at an extremely delicate stage where the state of the art varies from one laboratory to another and ex pertise at the bench counts for almost ev erything. To divorce research from de velopment at this stage will, in his opin ion, crucially delay the validation of the tests. "All the experts in the field are saying, 'Go full steam ahead in devel oping the tests, but don't stop to test chemicals because you will get a lot of test data you can't interpret properly.' " Saffiotti observes. Peters and Newell say that there is no question at all about Saffiotli's manage rial ability, but that they believed the tests could be developed more quickly. Asked the basis for disagreeing with Saf fiotti's scientific judgment, Peters says he regards the decision as managerial .not scientific, but that outside scientists were consulted. The 'discussions took place on an informal basis over a long pe riod of time until the decision just emerged. Peters says. A member of the NCI advisory board says that the deci sion to split off validation of the in vitro tests "ought to be made with a good deal of advice from scientific people who are not members ofthe NCI. That is what ad visory committees are for." Rauscher. Newell, and Peters also seem to hold Saffiotti to blame, though without dirpctly saying so, for the bot tleneck that has arisen in assessing and disseminating the results of in vivo tests. Some 200 chemicals have been put through animal feeding tests, but the re sults have not yet been published. Saf fiotti says that he warned of the bot tleneck 3 years ago, at the time the tests were started, but never received the staff he needed to get the results out. He was unable to transfer staff from elsewhere in his program without seriously damaging other program areas, which themselves are of high priority in interpreting the re sults of carcinogenicity tests and their relevance to man. Newell and Peters indirectly criticize Saffiotti for a reluctance to delegate tasks to contractors. "I feel these things can be done without direct control by active scientists," remarks Newell of Saffiotti's general approach to management. "I be lieve that you can go out and buy exper tise and you can get good expertise. With the in vivo tests the protocols are so well developed that it is sort of cookbook," Peters says. Saffiotti rejects this attitude, believing that unless the contractors are rigorously supervised, the tests will not be done properly, and a sloppy test is a total waste of investment. "There is no such thing as cookbook stuff at any stage of the way," says Saffiotti. As the Kennedy hearings on commercial laboratory operations have shown (page 531, this issue), it is crucial to set high standards for contrac tor performance, Saffiotti contends. "In fact, it is almost axiomatic that if bioassays are used to claim a negative re sult for regulatory purposes, the sloppier the test the `safer' will the product ap pear--unless one investigates the ade quacy of the procedures used." In a letter of resignation sent on 23 April. Saffiotti ascribes the backlog in publishing the test results to a "tragic policy" by the NCI of failing to provide his bioassay team with sufficient staff. Reassignment and disruptions in his pres ent staff appear likely "to lead to further delays in the publication of well-docu- SCIENCE. VOL. 192 I AP00024560 merited evidence of cancer hazards." he says in his letter. "I cannot accept any longer a situation which in fact deprives the regulatory agencies, industry, labor, consumers and the scientific community of data of urgent public health value: it is people who are now exposed to toxic agents and who are not protected be cause the necessary support was not pro vided in time." Sloppy Tests Saffiotti considers that he has had ``at best only a few opportunities to discuss and participate in major policy decisions at the Institute level." Rauscher. he says, has received advice from his pro gram "at second or third hand." The problem, as he sees it, is layers of bu reaucracy separating the NCI director from his scientific experts. "There is a huge bureaucracy in the front office-- some 200 people, few of whom are active scientists--and our own division director has built up a small layer of people in his front office," Saffiotti says. "Essentially the direct involvement in research stops at our level and all the rest is bureau cratic overhead. This may be the inevi table result of the very rapid growth of the NCI's budget, but the fact is that it is so." The observation is serious, if true, not least because one of the principal argu ments for giving the cancer institute greater autonomy within the National In stitutes of Health was to free it from en cumbering layers of bureaucracy. Raus cher, however, says he has never re fused to see Saffiotti. Peters observes that he is the only person between Saffiotti and the NCI director, and that he has a larger program but smaller staff than any other division in the NCI. One apparent example of the exclusion of experts from their proper role is that of a proposed interagency committee on the assessment of carcinogens. Peters has nominated himself as the NCI mem ber. Asked if Saffiotti wouldn't be better qualified. Peters says no, because the committee is a broad-based policy group. But the chairman of the committee, Roy Albert of the Environmental Protection Agency, says that the committee will be a technical group making scientific judgments. Asked why Peters was a member. Albert said. "It does look pecu liar that he is on it without being an expert in carcinogenesis or having that background, but 1 view it as only on the basis of setting up a channel to the NCI's experts such as [H.F.] Kraybill and Saffiotti." Peters has designated Kraybill as an alternate on the commit tee. The bureaucracy's actions are not al ways accountable to peer review by out side scientists, Saffiotti believes. Forexample, the division of cancer cause and ptevention has disbanded the peer re view group it used to have. Thus al though there are peer review groups at the program level, the decisions taken at divisional level, such as on allocation be tween the three programs (carcinogen esis. virology, and epidemiology) are not subject to direct peer review. Peters says the divisional level peer review group was disbanded because it did not make sense to have experts in one program area reviewing work done in another. Ac cording to Rauscher. however, "Active researchers have probably never had a greater input to the NCI--we have 62 ad visory committees and I don't know of a single recommendation from them over which I have control which I did not im plement." An attempt to test the opinion of scien tists in the chemical carcinogenesis field and on the NCI advisory board produced general praise of Saffiotti's program-- "He has done as well as anybody could possibly do." says one of the outside ad visers to his program. There seems to be a general reluctance to comment on the specifics of the issue, about which most people expressed ignorance. Members of the NCI advisory board were unaware even of the way in which the NCI's 79 staff positions had been distributed. "In ternal NCI affairs are very complex, and outsiders comment on them at their peril because there are so many political is sues involved," observes Nobel laureate Howard Temin of the University of Wis consin. Saffiotti's complaints may or may not be fully justified, but his resignation has at the least drawn attention to what he re gards as a critical shortage of support for the carcinogenesis program. --Nicholas Wade Clinical Labs: Bills Aimed at Correcting "Massive" Problems In medical practice today, physicians and patients rely increasingly on the re sults of laboratory tests to tell them what is wrong. And what is wrong, it seems, in an appalling number of cases, is the test itself. It is almost impossible to get accu rate data on the quality of work per formed in the 64,000 to 94,000 clinical laboratories in the United States (no one is sure exactly how many there are), but the Senate has evidence that "from 7 to 26 percent of all lab tests may be in er ror."* Considering the number oflabora tory tests performed in any year, that is a lot oferror. 7 MAY 1976 According to figures from the Senate subcommittee on health, chaired by Ed ward M. Kennedy (D-Mass.), 4k billion lab tests were conducted in this country in 1975; that comes out to 12 million a day. By the end of the year, the nation's total bill for laboratory tests hit $12 bil lion, which is roughly equal to 10 percent of the entire cost of health care for that year. Reports by the Senate's Special Committee on Aging, which conducted an investigation of Medicare fraud in the * Fraud nd abuse among clinical laboratories," a staff report prepared for the U.S. Senate's Special Committee on Aging, 1976. lab business, the Senate's subcommittee on health, and others reveal that prob lems with clinical laboratories are "mas sive and widespread." As Kennedy has said. "Clinical laboratories perform in adequately and constitute a threat to the public health." In the Senate, Kennedy and Jacob K. Javits (R-N.Y.) have sought a reme dy through the Clinical Laboratories Im provement Act of 1975, which is likely to go to the floor for action soon. Represen tative Paul G. Rogers (D-Fla.) has in troduced similar, though not identical, legislation in the House. And the Admin istration, which is perfectly willing to concede there is a problem, is trying to block legislative action on the grounds that there already exist adequate powers to regulate clinical laboratories and that the government is now going to use them. But Congress is not going to hold its breath until that happens. Nine years ago. Senator Javits in- 55! AP00024561 NAE Elects 104 New Members The National Academy of Engineering, established to share the responsi bility given the National Academy of Sciences under its congressional char ter to examine questions of science and technology at the request of the fed eral government, has elected 104 new members. This addition brings the to tal membership to 685. The new members are as follows: H. Norman Abramson, Southwest Re search Institute: Harold M. Agncw, Los Alamos Scientific Laboratory; Clarence R. Allen, California Institute of Tech nology; Alfredo H. -S. Ang, University of Illinois. Urbana-Champaign; Horace S. Bealtic, IBM Corporation; Daniel Berg, Westinghouse Electric Corporation; Don ald J. Blickwcde, Bethlehem Steel Corpo ration; John E. Breen, University of Tex as. Austin; Charles M. Brinckerhofif. Consulting Engineer. New York; Fred erick P. Brooks, Jr., University of North Carolina; Donald B. Broughton, UOP Process Division; Bernard Budiansky, Harvard University. Joseph E. Burke, General Electric Re search and Development Center; Marvin Camras, 1IT Research Institute; Dayton H. Clewell, Mobil Oil Corporation: Ju lian D. Cole, University of California, Los Angeles: John W. Collman, West inghouse Research Laboratories; Frank lin S. Cooper, Haskins Laboratories; F. J. Corbato, Massachusetts Institute of Technology; Ruth M. Davis, Department of Commerce; Anthony J. DeMaria, United Technologies Research Center; Ira Dyer, Massachusetts Institute of W. Johnson. University of California. Berkeley; Robert L. Johnson, McDonnell Douglas Astronautics Com-. party; Donald J. Jordan, Glastonbury, Connecticut: Joseph H. Keenan, Massa chusetts Institute of Technology; Robert W. Keyes, IBM T. J. Watson Research Center: Lee A. Kilgore, Consulting Engi neer. Pennsylvania; Gordon S. Kino, Stanford University; Leon Lapidus, Princeton University; Milton Levenson. Electric Power Research Institute; Jo seph T. Ling, 3M Company; Ray K. Llnsley, Hydrocomp, inc.; John P. Longwell, Exxon Research and Engineering Com pany: Bruce T. Lundin, NASA Lewis Re search Center; John D. Mackenzie, Uni versity of California. Los Angeles; En rique A. J. Marcatili, Bell Laboratories; Hans M. Mark, NASA Ames Research Center; Sidney Metzger, Communica tions Satellite Corporation. Herbert L. Misch, Ford Motor Compa ny; James K. Mitchell, University of California. Berkeley: Gordon E. Moore, Intel Corporation: Ben Morccll, Pitts burgh. Pennsylvania: Richard S. Morse, Massachusetts Institute of Technology Development Foundation. Inc.; Albert Technology; Milton C. Edlund, Virginia Polytechnic Institute and State Universi ty; Lloyd E. Elkins, Amoco Production Company; Martin A. Elliott, Energy Con sultant. Texas; Richard S. Engelbrecht, University of Illinois. UrbanaChampaign; Elliott M. Estes, General Motors Corporation; Joseph Fcinstein, Varian Associates; Steven J. Fenves, Carnegie-Mellon University; Michael Field, Metcut Research Associates Inc. MertonC. Flemings, Massachusetts In stitute of Technology; Charles F. Fo garty, Texasgulf Inc.; Gerard F. Fox, Howard Needles Tatnmen & Bergendoff: Alfred M. Freudenthal, George Washington University; King-Sun Fu, Purdue University; Douglas W. Fuerstenau. University of California, Berke ley; Robert A. Fuhrman, Lockheed Mis siles and Space Company, Inc.; Solomon W. Golomb, University of Southern Cali fornia: John B. Goodenough, Lincoln Laboratory; Robert C. Gooding, NavaJ Sea Systems Command; Arthur G. Han G. Mumma, Short Hills, New Jersey; Pe ter Murray, Westinghouse Advanced Re actors Division; Eugene F. O'Neill, Bell Laboratories; Henry J. Ongerth, Califor nia State Department of Health: Jack S. Parker. General Electric Company: Nor man F. Parker, Varian Associates; Thomas H.Pigford. University ofCalifor-' nia. Berkeley; Egor P. Popov, University of California. Berkeley: Jacob Rabinow, Department of Commerce; Eric Reissner, C niversity of California, San Diego; James B. Reswick. Rancho Los Amigos Hospital; Allen S. Russell, Aluminum Company of America; Robert S. Schechter, University of Texas, Austin; Rein hardt Schuhmann, Jr., Purdue Universi ty. David Sleplan, Bell Laboratories; Wil liam P. Slichter, Bell Laboratories; Ar thur C. Stern, University of North Caro lina; Archie W. Straiton, University of Texas. Austin; Morgan C. Sze, Lummus Company; Harold A. Thomas, Jr., Har vard University; Chang-Lin Tien, Uni sen, Purdue University: Edwin L. Hard er, Pittsburgh. Pennsylvania; Milton Har ris, Washington. D.C.; Herman A. Halts, Massachusetts Institute of Technology; Arthur Hauspurg, Consolidated Edison Company of New York Inc.; Heinz Heinemann, Mobil Research & Devel opment Corporation; Joseph M. Hendrie, Brookhaven National Laboratory; Abra ham Hertzberg, University of Washing ton. Wllmot N. Hess, National Oceanic and versity of California. Berkeley; Milton Van Dyke, Stanford University; Henning E. Von Gierke', Aerospace Medical Re search Laboratory'; John B. Wachtman, Jr., Department of Commerce: William M. Webster, RCA Laboratories; Jo hannes Weertman, Northwestern Univer sity; Roy F. Weston, Roy F. Weston, Inc.: RichardT. Whitcomb, NASA Lang ley Research Center; J. Ernest Wilkins, Jr., Howard University; Amnon Yariv, California Institute of Technology; Al Atmospheric Administration; William C. Hittinger, RCA Corporation; Charles H. Holley, General Electric Company; Joe fred A. Yee, Alfred A. Yee & Associates, Inc. 532 troduced the first piece of legislation cov ering the operation of clinical labs--the Clinical Laboratories Improvement Act of 1967. That law. intended to guarantee high quality lab work, authorized the Center for Disease Control (CDC) in At lanta, Georgia, to test the proficiency of clinical labs; but its authority was re stricted to those laboratories that engage in interstate commerce, estimated to be a mere 6 percent of the total. CDC officials believe that those 900-odd labs that come under their purview represent the best in the country, but the Center's own analy ses of the labs' performance indicates that all is no( well, even among the best. As part of its quality monitoring pro gram, CDC sends sample specimens of various sorts to laboratories to see if they are identified correctly. Sometimes the laboratories know they are being test ed, sometimes not. Among recent dis couraging findings are these: Thirty-one percent of labs, which knew they were being tested, failed to identify sickled red blood cells. Leukemia was mistakenly diagnosed by more than 10 percent of labs in one test. Mononucleosis is incorrectly diag nosed as much as one-third of the time. Seventeen of 22 laboratories cor rectly identified drugs in urine samples when they knew they were being tested, but 16 of those 22 labs missed the drug identification 60 percent of the time in a repeat quiz when they received urine la beled as patient specimens rather than as CDC test material. Five to 12 percent of the time, labo ratories "find something" when CDC sends them slides that contain nothing of consequence. Sometimes the errors are trivial, but sometimes they are of more than passing consequence. Even people in the labora tory business acknowledge problems of error and some capitalize on them, as is shown in an advertisement in the November-December 1975 issue of Cadence (the journal of the American Society for Medical Technology), which is devoted to the issues of achieving quality assur ance by legislation. A two-page ad shows a dramatic sketch of a young woman lying in a hospital bed, her husband, head bowed, staring out of the window. The ad, in bold letters, reads, "She searched her conscience, conferred with her hus band and went through with the abor tion. But there was no fetus there." The ad is for a particular pregnancy test that is claimed to meet a "high standard of ac curacy." The Senate has reached the conclusion that the reason for all these sorts of prob lems is that clinical laboratories and labo- SCIENCE. VOL. 192 / t AP00024562 "That letter disqualified them," he said, "If we'd tried to appoint them we'd have had resignations from half the com mittee." The episode has led some participants to question the adequacy of the Acad emy's procedures for uncovering bias among prospective committee members. The list of names from which the com mittee was chosen was generated primar ily by the Academy staff with help from relevant consultants, and Hastings add ed some names of his own. Then Has tings, after analyzing a list of the fields of expertise needed and the potential can didates from those fields, indicated whom he wanted as committee mem bers. Hastings told Science. "1 person ally wasn't sensitive to any stands taken by these people." Hastings' recommendations then had to gain the approval of other key figures in the Academy. But it was only after the committee members were appointed that they were asked to fill out bias state ments indicating, among other things, any views they might have expressed publicly on the issues to be considered by the committee. By that time, it would have been embarrassing to ask anyone to withdraw, though Academy officials say they have done so on occasion in the past. Some Academy representatives are suggesting that appointments should be made conditional upon review of the bias statement, but others consider it presumptious to ask scientists to reveal their stockholdings, commercial affilia tions, grant support, and other such mat ters if they are not sure they will actually be appointed. And for every scientist who wants to tighten up the bias proce dure, there is another who wants to weaken it. Schwan, for example, consid ers it "an awful thing" for the Academy to ask what stands he has taken on an is sue. "It intimidates my freedom of ex pression," belaid. "Where's the border line between such things and what hap pens in Russia?"--Philip M. Boffey Chemical Carcinogens: Industry Adopts Controversial "Quick" Tests how to evaluate potentially harmful sub stances in accord with the pending Toxic Substances Control Act. Ames reports that 60 or 70 major companies, including such giants as American Cyanamid, Inc., Merck Sharp & Dohrae. and Dupont have asked him to supply them with the strains of salmonella bacteria he uses in For generations, industry has been in troducing new chemicals into the envi ronment in staggering numbers without really knowing whether they might be hazardous. And the public, assuming there was nothing to be done, or not thinking about it, has passively tolerated the situation. But then the environmental movement came along, as did the calcu lations by epidemiologists that a large proportion ofall cancers are environmen tally caused. As a result, there has been growing pressure to force industry to evaluate new chemicals before they are released, on the theory that safety should be tested in the laboratory and not in the environment. And there is a good chance that Congress this year will pass the Toxic Substances Control Act that would mandate premarket testing (Science, 13 February). The major impediment to premarket testing has been the lack of a test system that is reliable, fast, and cheap. How ever, during the past few years some progress has been made in that area, largely because of the leadership of bio chemist Bruce Ames of the University of California, Berkeley. Ames developed a simple system for taking a quick look at the mutagenic, and by implication, car cinogenic, properties of chemicals. No one yet is sure just how. reliable a predictor the Ames test--a bacterial sys tem--is, but it is generally thought to be the best available of its type. In view of the probable passage of the Toxic Sub- stances Control Act. the chemical in dustry has begun, during the past year or so, to use the Ames, and other quick tests, on its own in order to get some idea of the safety of new products. In the process, industry itself may help to answer questions about the value of vari ous types of screening systems by gener ating sufficient volume of data on which to base scientific judgments. At present, the only officially recog nized way to test a chemical for carcino genicity is to see whether it causes can cer in laboratory animals, which takes 2 to 3 years and costs about $100,000 per chemical. Citing the time and money in volved, industries have been notoriously reluctant to routinely screen new prod ucts in animals. On the other hand, they hesitate to invest huge sums of money in the development of new products with out knowing whether those products will later be banned as carcinogens. There fore, industries have seized on a variety of quick and inexpensive tests that, they hope, will tell them whether substances are carcinogens. This has led to a curious situation in which industries are implicit ly endorsing the tests at the same time that scientists and legislators deliberate over whether companies should be forced to use them. The extensive use by industries of these quick tests is hailed by many scien tists as a change in the tradition of wan ton release ofchemicals into the environ ment even while debate continues on his system. Numerous other firms do not test their products themselves but send them to commercial laboratories for test ing. Companies are reluctant to discuss their uses of the quick tests, but Ames re lates one story told to him by investiga tors at American Cyanamid's agricultural division. It seems that American Cyanamid found what looked like a promising new pesticide that turned out to be highly mutagenic when tested in Ames' bacte rial strains. Not willing tojust drop their new product the investigators of Ameri can Cyanamid took a second look and found that this mutagenic effect was due, not to the primary chemical but to an impurity in the pesticide. Now, Ames reports, the company has removed the mutagen from the pesticide and has deci ded that it is worthwhile to go the full route with the purified product by testing it in animals. The Ames test is based on the pre sumption that many cancers are related to mutations or some sort of damage to the DNA of a cell and, therefore, that agents that are mutagenic are likely to be carcinogenic as well. After searching through innumerable bacterial strains, Ames hit upon some mutants of Salmo nella typhimurium. that have lost the ability to make the amino acid histidine. Consequently, in a histidine-free culture medium, these bacteria cannot grow. What Ames has shown is that, when these bacteria are exposed to mutagenic chemicals, they undergo additional muta 1S JUKE 1976 nts AP00024563 i). tions that can have the effect of repairing At present, no other assay is as widely companies that may be marketing toxic the original defect. What it amounts to is used as the Ames test, but each is being substances. ,l that in the presence ofa mutagenic chem studied extensively. Litton Bionetics, like most dthercom- ical, the bacteria begin to grow again, Commercial laboratories report that mercial laboratories, offers its clients a forming colonies that show up as white the recent interest by industries in quick range of quick tests, including the Ames spots. A particularly handy feature of the tests has provided them with a substan test, which is almost always performed: test is that powerful mutagens will cause tial increase in business. David Brusick first. In many instances, a firm is told a larger number of bacteria to revert than of Litton Bionetics in Kensington, Mary that its chemical is mutagenic in the Ames will less potent ones, thereby providing land, says that his linn has had some con test and will then request other quick Jfc'-.: at least some indication of how poten tracts to screen chemicals for the past tests of the chemical before deciding tially hazardous a suspect chemical may 2 years but that the vast majority of its what course of action to take. The cost be. The test is cheap--it costs only $200 clients have been signed up in the past of a whole battery of tests is less than per chemical--and fast--it can be com few months. Now Litton Bionetics does one-tenth of the cost of a cancer test pleted in 3 days. the tests for about 50 companies. Clients in which laboratory animals are used. Other quick assays for mutagenic include pharmaceutical companies, man The widespread use of the Ames test is chemicals are based on yeast, fruit flies, ufacturers of agricultural chemicals, pro a tribute to the decade of work put in by and mammalian cells grown in culture. ducers of pigments and dyes, and other Ames and his associates. They have Medical Devices Law Is on the Books at Last Congress has finally given the Food and Drug Adminis tration explicit authority to regulate medical devices in a law signed by the President at the end of May. The event concluded 15 years of intermittent congressional efforts to fill a regulatory gap that was becoming ever more evident with leaping advances in medical technology. Efforts to pass a devices law began in earnest following a 1970 report by a study group at the Department of Health, Education, and Welfare, which revealed that medical de vices had been implicated in 10,000 injuries and 731 deaths between 1963 and 1969. Most of the deaths resulted from malfunctioning heart valves and pacemakers. The new amendments to the Food, Drug and Cosmetic Act for the first time empower FDA to review and approve high-risk devices before they go on the market. The law puts ail medical devices from tongue depressors to artificial organs into three categories. Classification, to be done by outside panels appointed by the secretary of HEW, will be made according to the potential danger of each device and the availability of information sufficient to formulate safe standards governing its design, manufacture, and use. Devices put in class III, the most stringent category, will require FDA clearance before they are marketed. This ap plies to devices that are deemed to be life supporting or life sustaining or are implanted in the body. Class II devices must conform to standards to be promulgated either by groups from outside the government or by the government. Class I devices are subject to "general controls," which means they basically won't be regulated any more than they are now. This is equivalent to the "generally recog nized as safe" designation for food additives. Until now, the only explicit statutory authority the FDA has had to regulate devices has come from the 1938 drug law which permits the agency to take action against any de vice found to be "misbranded" or "adulterated." In 1969 the concept that devices could be regulated as drugs within the law was elaborated by a Supreme Court decision which ruled that Bacto-Unidisk, a paper disk used for testing bac terial sensitivity to drugs, should be classified as a drug. But since then, only a handful of devices, such as copper IUD's and soft contact lenses, have been regulated as drugs. According to a lawyer on the staff of Senator Gaylord Nelson (D-Wis.), who is largely responsible for strengthen ing the bill from its earlier versions, what the new law does is shift the burden of proof that a device is safe and effec tive from the FDA (which only had the power to intervene after a device was on the market) to the manufacturer. The law requires that every "new" device--that is, every one that is introduced after passage of the law and is not "sub stantially equivalent" to something already in use--must be automatically put in class III. From there, panels have 6 months to decide whether to approve it and whether to re classify it in class I or II. All "old" devices--those already on the market--that are implantable or life sustaining also go into class III. Their manufacturers are given 3 years from the date of the law's enactment to get marketing ap proval. Devices now covered by new drug applications would also probably go into class III. Passage of the law has taken a remarkably long time con sidering the fact that some sort of legislation has been wide ly thought to be not only desirable but inevitable. Even de vice manufacturers have supported it as being far prefer able to alternative and even more stringent regulatory procedures. Their main complaint about the new law, ac cording to a spokesman from the Pharmaceutical Manufac turers' Association, is that class III is unnecessarily broad and that the restrictions in this class will impede the flow of new devices onto the market. As for consumer advocates, the chief problem, accord ing to attorney Anita Johnson of Ralph Nader's Health Re search Group, is that the major classification decisions are to be made by committees of nongovernment personnel. Johnson believes outsiders are more lax and subject to con flicts of interest, and that the only way to ensure account ability is to keep all responsibility on the backs of public servants. "We must decide whether we want outsiders to be making basic public health decisions," she says. John son also calls the standards-setting procedures "a Rube Goldberg machine" that offers numerous opportunities for industry interests to create obstructions and delays. The new law covers about 12,000 devices, products of a more than S3-biIlion-a-year industry. According to an FDA official, about 10 percent of all devices would go into the premarketing approval category and half would be allowed to stay under general controls.--C.H. 1216 SCIENCE. VOL. 192 AP00024564 shown that 90 percent, or 156, of 174 known' carcinogens cause mutations in their bacterial strains. By contrast, few of the 109 "noncarcinogens" that they have tested are mutagens. Because a wide variety of classes of chemical car cinogens are mutagens in his system, Ames argues that this bacterial test sys tem provides a reliable screen for poten tially harmful chemicals. As an example of the usefulness of quick tests for carcinogenicity, Ames tells the story of a preservative (furofuramide) that was extensively used in Ja pan. It was tested in animals and found not to cause cancer. The chemical, how ever, did cause mutations in bacteria. It was subsequently retested in animals, found to be carcinogenic, and banned. Now, Ames reports, bacterial mutageni city tests are extensively used in Japan. In particular, the Japanese require that all pesticides be shown not to cause mu tations in bacteria. The Japanese are also taking very seriously the finding that hair dyes are potent bacterial mutagens and are currently trying to develop hair dyes that do not have this drawback, ac cording to Ames. Despite these arguments in favor of the Ames and similar tests, it is by no means clear how results of these tests are to be interpreted. One problem is that, with the data obtained so far, the logic behind the test evaluations goes the wrong way. Ames can say that 90 per cent of all tested carcinogens are muta gens in his bacterial strains but he cannot say what the probability is that a chem ical that is a mutagen will turn out to be a carcinogen. Some investigators believe that this difficulty can be partially remedied by the use of more than one test system. A positive result in several quick tests might carry more weight than a positive result in the Ames test alone. Mamma lian cell systems are of particular interest to some investigators who believe it would be intellectually more satisfying to detect DNA damage to these cells- or transformation of them into tumor cells than to detect mutations in bacteria. Mammalian cell test systems are not yet extensively used, however, because these tests cost 5 to 10 times more than the Ames test and neither they nor the other quick tests have the data base or the sensitivity of the Ames test. Legislators and officials at the Nation al Cancer Institute have expressed inter est in the quick tests but have hesitated to endorse any of them because of the difficulties in interpreting results. Yet rapid and inexpensive screens for harm ful chemicals are needed if the Toxic Substances Control Act is to be economi cally feasible. The ultimate solution to the problem with the quick tests lies in obtaining more data. Then it will be pos sible to correlate results from quick tests with results from animal tests for carcino gens in a statistically convincing manner. By their recent extensive use of the quick tests, industries seem to be making a sub stantial contribution to the data base that must exist for these tests to be evaluated.--Gina Bari Kolata Nuclear Testing: U.S.-Soviet Treaties Viewed with Doubts and Misgivings Control of nuclear arms is a matter about which the public, aware of the gen erally frustrating history of arms negotia tions, has learned to lower its ex pectations. Accordingly, the treaties ne gotiated by U.S. and Soviet officials to limit underground weapons tests and peaceful nuclear explosions have not been awaited with much excitement or anticipation. But, now that the Ford Ad ministration is finally ready to submit them to the Senate for ratification, these treaties give rise to so many doubts and objections that many senators will find it a close question whether they are margi nally better than nothing or whether they are actually worse than nothing. The treaties now subject to ratification go back to the Moscow summit of July 1974, Richard Nixon's last hurrah before he was forced to resign over Watergate. With the strategic arms limitations talks still at an impasse, the President seized the opportunity to sign a treaty limiting underground tests to a certain maximum yield or "threshold." Had the treaty banned all underground tests or even all tests susceptible to It JVHB 1976 unambiguous "verification" by seismic monitoring, it would have been ap plauded by the private arms control com munity that is made up in good part of groups such as the Federation of Ameri can Scientists (FAS) and the Arms Con trol Association (ACA). (A number of former government officials with respon sibilities for arms control are active in both the FAS and the ACA.) But the treaty Nixon brought back from Moscow was instantly put down as a mockery by many of these arms controllers, who wanted a treaty that would effectively discourage further weapons devel opment and inhibit nuclear proliferation. By prohibiting testing in the atmo sphere, in outer space, and under water, the Limited Test Ban Treaty (LTBT) of 1963 had stopped the dangerous radio active contamination of the world envi ronment. But the United States and the Soviet Union had simply moved their am bitious programs of testing underground, and, consequently, the LTBT had con strained the arms race little if at all. The Threshold Test Ban Treaty (TTBT) signed by President Nixon and General Secretary L. I. Brezhnev in 1974 clearly would not do much to constrain the arms race, either. The threshold, set at 150 kilotons, would allow the testing of weapons 10 times more powerful than the one that destroyed Hiroshima. Also, this threshold would bear no relation to verification capabilities, which have been improved to the point that even ex plosions at yields of 10 kilotons or less may not escape detection. Furthermore, many arms controllers objected strongly to the fact that, per missive as it was, the treaty would not take effect until 31 March 1976, thus leav ing time for each of the two superpowers to carry out a series of tests at high yields. They objected, too, to a glaringly evident loophole in the TTBT--"peace ful" explosions, of whatever magnitude, were not covered even though such explosions can be indistinguishable from weapons tests. Administration officials acknowledged that plugging this loophole was priority business. Assurances were given that the TTBT would not be submitted for ratifi cation unless it was accompanied by an agreement on peaceful nuclear ex plosions, or "PNE's." And, sure enough, after a year and a halfofhard ne gotiations, a PNE treaty (PNET) was signed on 28 May, in ceremonies con ducted simultaneously in the White House and the Kremlin. After its fashion, the PNET does ad dress some of the concerns of the arms AP00024565 controllers. No individual PNE would be allowed to exceed 150 kilotons, which is to say that the threshold for weapons tests would also apply to PNE's. And, in the event either nation should plan a si multaneous detonation of two or more peaceful devices with an aggregate yield greater than 150 kilotons, the other could have observers present to monitor the event on-site. With the yield limited to 150 kilotons, no PNE would produce militarily useful information not otherwise obtainable through weapons tests allowable under the TTBT. Also, properly equipped ob servers present for simultaneous multiple explosions could verify that no individual explosion has exceeded the threshold, something which could not be accomplished by seismic monitoring at a distance. The negotiations for the PNET began in late 1974 and were not successfully concluded until this past April. The United States would have readily agreed to ban PNE's altogether. After spending more than $160 million on earth-moving (cratering) and gas-stimulation experi ments with PNE's, the U.S. government had pretty well decided that their prom ise was illusory. But the Soviets, who had experimented with a number of dif ferent PNE applications, professed an es pecially keen interest in pursuing grandi ose plans to use PNE's for canal build ing. There might be some simultaneous multiple explosions having an aggregate yield of several megatons, with some in dividual explosions of up to 400 kilotons or greater. In the hope that Ihey would be able to continue conducting PNE's without restrictions, the Soviets had agreed in principle at the Moscow sum mit to allow some PNE's to be wit nessed by on-site observers, subject to such terms and conditions as the nego tiators might later arrive at. And this was, in fact, a significant concession in light of the traditional Russian aversion to the idea of on-site inspections. But, by the time negotiations for the PNET got under way it had become ap parent that the Senate would probably re ject the TTBT unless PNE's as well as weapons tests were made subject to the 150-kiloton threshold, with adequate pro visions for verification. The on-site observer provisions ulti mately agreed to testify to the delicate balance that had to be struck to satisfy both the U.S. negotiators' insistence on effective verification and the Soviets' in sistence on minimizing any American in trusion. For instance, American observ ers witnessing a Soviet PNE would bring two identical sets of yield monitoring I21S equipment and allow the Soviets to choose the set actually to be used; the other set would be turned over temporar ily to the Soviets. This, together with other conditions meticulously spelled out in the treaty, would make any unauthor ized snooping hard to get away with. The nation carrying out PNE's would be required to provide the other party prehensive test ban or harm, them, one person's speculation may be as good as another's. The treaties would establish a threshold from which negotiators could work downward. But many arms con trollers believe that, if the treaties are for mally ratified, the U.S. and Soviet gov ernments will not resume serious test ban negotiations during the TTBT's with a substantial amount of geologic, hydrologic, and other data bearing on the interpretation of seismic signals and the measurement of explosive yields. In general, the larger the PNE, the more data required. Such data would be essen tial to verification because, unlike weap initial 5-year term (renewal is automatic unless either party withdraws). The treaty does call for negotiations looking toward a comprehensive test ban, but, itj light of its other provisions, this has the ring of an empty promise. There is even the very real likelihood ons tests conducted solely at designated test sites (about which there would be in formation exchanges), PNE's might be carried out under widely varying con ditions. The question whether the TTBT and the PNET represent a gain or a loss for the cause of arms control must be consid ered from the standpoint of weapons de velopment, nuclear proliferation, and the chances for a comprehensive test ban. The TTBT would impose some con straints on weapons development, but they would be clearly marginal, espe cially in light of the numerous kinds of high-yield weapons already in the U.S. and Soviet inventories and ofall the test ing done between July 1974 and the end of March 1976. During this period, the United States conducted 12 announced tests at yields over 200 kilotons; a num ber of these tests were at yields of up to a megaton. The Soviets also detonated a number of high-yield devices, including several in the multimegaton range. From the standpoint of nuclear prolif eration, the 150-kiloton threshold does indeed seem to make light of the super powers' solemn obligation under the Nonproliferation Treaty of 1970 to seek arms reductions and a comprehensive that, if the Soviets try to employ PNE's in canal building, they will violate the 1963 test ban treaty. Expert opinion holds that some "venting" of radioactive debris would inevitably occur and that part of this material probably would drift across international boundaries. Yet, despite all doubts and migivings, if the TTBT and the PNET are sent to the Senate floor this year and brought to a vote, they are likely to be ratified. This is so because even such senators as Ed ward Kennedy, sponsor of a resolution calling for a comprehensive test ban, are afraid that Senate rejection of the treaties might undermine hopes for fur ther arms control agreements. On the other hand, there may be a better than even chance that in this elec tion year time will run out before the Senate brings itself to act. Indeed, sena tors who regard the treaties dubiously may contrive to make this happen. Iftime does run out this year, the question of ratification will go over until 1977 when a newly elected President will have to re view it. If the president should be some one like Jimmy Carter, who favors an immediate moratorium on testing, the treaties' fate will be in serious doubt. --Luther J. Carter test ban. Furthermore, the PNET in a sense legitimizes PNE's, although it does underscore the fact that a PNE and a weapons test can be indistinguishable. Arms controllers can only hope that, as some people in the Ford Administra tion are now suggesting, the PNET'S on site observer and data requirements may discourage the Soviets from undertaking a major program of PNE's. But, if such a PNE program is carried out by the So viets, this could encourage some non nuclear nations to develop nuclear de vices, either out of a genuine interest in PNE technology or from a recognition that PNE's offer a convenient mask for a fledgling nuclear weapons program. As for whether the TTBT and the PNET would improve chances for a com Clarification: The article, "Pesticide*: Three EPA attorneys quit and hoist a warning flag" (19 March) referred to aklrin and dieldrin and heptachlor and chiordane as "two pairs of compounds found to be potent carcinogens." Most uses of- both aldrin/ dieldrin and neptachlor/chlordane were in fact ordered suspended by the administrator of the Environmental Protection Agency as an "imminent hazard" to human health and the environment. But some readers may take the words "potent carcinogen" to mean that a compound has been de termined by federal authorities to be carcinogenic in laboratory animals at relatively low dose levels. No such finding has been made with respect to hepta- chlor/chlordane. In the ease of these compounds, the administrator overruled the decision of the administrative law judge, who held that to suspend their use as an immi nent hazard was notjustified. He felt "hesitantly un witting at this time to find that heptachlor/chlordane are conclusively carcinogens in laboratory ani mals." The adoiinistrator concluded, however, that heptachlor is a carcinogen in both the rat and the mouse. Weighing the risks of continued use of heptachlor/chlordane against the benefits, he decided that most uses should oe suspended. His ruling is oow under review by the U.S. Circuit Court of Appeals for the District of Columbia.--L.J.C. SCIENCE, VOL. 192 1 AP00024566 r mi AM B III CANEstablished (S4J March 1976 Volume 2.14 Number 1 The Metabolism of Alcohol Overconsumption ofalcohol can cause cirrhosis and death not only because alcoholism promotes malnutrition but also because alcohol and its products disturb liver metabolism and damage the liver cells by Charles S. Lieber or all the attention being directed to recognized that heavy ingestion of alcohol F ward heroin, cocaine and marijuana, the favorite mood-altering drug in is associated with diseases of several tissues, notably those of the liver. Until reccnlly, the U.S., as it is in almost every humahnowever, such disease was attributed not to society, is alcohol. Its psychic effects, both alcohol itselfbut to the malnutrition that is pleasant and unpleasant, arc well enough so often an accompaniment of excessive known. What is less well known is that alco drinking. Alcohol was considered not a hol, in different quantities for different peo drug but a special kind of food with interest ple, is a toxic drug: its overconsumption ing psychotropic effects, and one that was taxes the body's economy, produces patho metabolized by the body like any other en logical changes in liver tissue and function ergy-rich food. As recently as 1949 the dis and can cause disability and death. As the tinguished physiologist Charles H. Best and "idence of alcoholism has risen in the U.S. his colleagues wrote that alcohol's metabol >ulation, so has the incidence of cirrhosis ic contribution was simply to supply calo ui the liver, which in 1974 climbed past rics and that "there is no more evidence of a both arteriosclerosis and influenza and specific toxic effect of pure ethyl alcohol pneumonia to rank seventh among the lead upon liver cells than there is for one due to ing causes of death; in some urban areas sugar." Perhaps it was wishful thinking on (including New York City) it is actually the the part of people in general and physicians third most frequent cause of death between in particular that installed as accepted fact the ages of 25 and 65. Such a disease de the concept that alcohol lacked intrinsic serves intensive study. It is only in recent toxicity. years that the changes in liver tissue and It was a rather nai'vc belief. Ethanol has liver function have been directly associated little in common with other energy-rich with specific steps in the metabolism of al compounds. Carbohydrates and fats can be cohol, for the first time giving some hope synthesized in the body as well as ingested that rational methods can be developed in the diet, but alcohol is essentially foreign for preventing and treating alcoholic liver to the body. As do carbohydrates and fats, disease. alcohol has a high caloric value and is readi What we call alcohol is of course ethyl ly absorbed from the gastrointestinal tract; alcohol, or ethanol (CHaCH-jOll). On the unlike them, however, it is not effectively earth ethanol is probably almost as old as stared in the tissues. Moreover, very little lire itself. Whereas for man it is something alcohol can he disposed of through the to be ingested, it is a waste product for lungs or the kidneys, so that the body can yeasts, the microorganisms that manufac get rid ofalcohol only by oxidizing it. Again ture it. Ethanol is what is left over after unlike flit and most carbohydrates, which fermentation, the process by which yeasts, can be oxidized by almost all tissues, alco through the agency of their particular en hol must be oxidized in the liver, the organ zymes, derive energy from various vegeta that contains the bulk of the enzymes neces ble sugars. Man probably encountered etha sary for initiating the process. The organ- nol during prehistoric times in the form of specificity of alcohol explains to a large ex naturally fermented fruit juice (wine), hon tent the concentration of so many of alco ey (mead) and malted grain (beer), hol's deleterious effects in the liver, which is At least since the anatomical studies of the body's main chemical plant, the pri Vcsalius in the 16th century it lias been mary site of metabolic processes ranging from the synthesis of proteins to the detoxi fication of drugs and other foreign sub stances. The purpose of this article is to relate how our studies at the Bronx Veterans Ad ministration Hospital and the Mount Sinai School of Medicine of the City University of New York demonstrated that alcohol's tox icity for the liver is independent of poor diet and have shown how specific events in the metabolism of alcohol are linked to specific biochemical and structural changes in the liver and other tissues. There are reasons enough, both socio economic and physiological, for on al coholic to suffer from malnutrition. The al coholic spends his time, money and energy drinking and may tend to disregard such routine tasks as the preparation of mails and even eating. Unlike other drugs, alco hol has a high caloric value: 7.1 calories per gram, so that a pint of 86-proof spirits (not an unusual daily intake for an alcoholic) represents about half of the daily caloric requirement. As a result the appetite for food may be diminished. Ethanol calorics, however, are "empty calories," quite unas sociated with nutritive proteins, minerals and vitamins. Moreover, alcohol directly enhances malnutrition. By causing inflam mation of the stomach, pancreas and intes tine it can impair the digestion of food and absorption of nutrients into the blood; mal nutrition can in turn cause the intestine to function still less effectively. Finally, etha nol and its primary conversion product, acetaldehyde, can interfere with the activa tion of vitamins by liver cells. As a result of all these effects malnutrition is common In alcoholics, and malnutrition alone docs un questionably impair liver function, as is ob served in experimental animals that arc fed severely deficient diets. It was therefore quite natural to speculate 25 t t i i AP00024567 e' ' \* '" , v v .-- V. <' <* <\ . i *> , v I ,s, -, > " {? J > - "i'o * *- . /'*:. > pi',.*; - I h'-t-y Wr. . A.. . . ' --C' ' ' ; o * tVA'f ,y: .JN-*iNc:> J's ,.-0 ,v' *1. 0;- .ft.-41- ......... ; 'iC - *&"? 'y '> Vjy >. & ' .X' VL- FATTY LIVER was produced in laboratory rats by alcohol in spite of an adequate diet. Liver-ttssuo sections from two rats arc enlarged 240 diameters in these photomicrographs. One rat (/c/f) was fed A liquid diet without alcohol. The other (right) had a liquid diet containing ethanol as 35 percent of total calories. After 24 days the micrograph of the ethanol-fed rat's liver shows numerous globule fat droplets. l ntr&'&b MICROS1KUCTUUAL CHANGES flrc revealed in electron micro graphs made by Oscar A. Ison and the author, in which part of the liver cell is enlarged 16.000 diameters. In the control liver {left) the raj* organelles with infolding membranes (cristae) are normal mito- uominn; between thorn the parallel arrays of ribosome-studded rough endoplasmic reticulum ore seen, along with some vesicular (sac-shaped) smooth endoplasmic reticulum. In the liver of the etha nol-fed rat (right) the milochondtia arc swollen and distorted; in some mitochondria the external membrane mid cristae urc disrupted. There is marked proliferation of the smooth endoplasmic reticulum. V, AP00024568 that malnutrition might tv the cause of al- alcohol. With this method we produced fat cell [sec bottom illustration on opposite C'o/iolic liver disease even if one recognized ty liver in rats in spite of an adequate arid patre). that alcohol was not a food. The fact re fully controlled diet. The changes in the Although our alcohol-fed rats developed mains that in medical practice one encoun liver were comparable to those produced in a fatty liver, they did not go on to incur the ters alcoholics who develop liver disease in man. both on gross examination and when more severe forms of alcoholic liver injury- spile of what is apparently an adequate diet. tissue was viewed with the light microscope seen in human beings. The first of these In the late 1950's I therefore began to won [sec top illustration on opposite page]. The stages is alcoholic hepatitis, in svhich de der whether alcohol might exert direct toxic ultrastructural changes seen with the elec creased livcr-ccll function leads to the death effects on the liver. Such a finding would tron microscope were also similar, includ of cells, inflammation and a mortality rate clearly be of therapeutic as well as theoreti ing enlargement and disruption of the mito of from 10 to 30 percent. The final stage is cal interest. Belief in the exclusively nutri chondria and proliferation of the smooth cirrhosis, when fibrous scars disrupt the tional theory was leading many physicians endoplasmic reticulum, part of the sys normal architecture of the liver and give to advise alcoholic patients that normal liv tem of involuted membranes within the rise to a number of potentially fatal compli- er function could be maintained in spite of heavy alcohol consumption as long as the diet was adequate. If that was not true, it would be good to know it. The importance of the issue, along with our confidence that even heavy consump tion ofalcohol carries a negligible risk ifit is only for a brief period and under supervised conditions, led to a decision to study the effect of ethanol directly in volunteers. We fed the volunteers a nutritionally optimal low-fat diet, in which protein accounted for 25 percent of the calories, or two and a hair times the recommended percentage. That meant beef for breakfast, cheese and fish for lunch and meat or fowl for dinner, with supplementary vitamin and mineral cap sules. The volunteers also had six drinks a day, a total of about 10 ounces of 86-proof alcohol. On that regimen there was a pro gressive rise in liver fat; routine thin-needle biopsy revealed significant fat accumulation after a few days, and there was an average eightfold increase over the 18-day course of the experiment. Here was the characteristic "fatty liver," the first (and typically revers ible) stage ofliver disease. There were strik ing changes in the ultrastructure of theliver cell; the mitochondria, the energy-convert ing organelles, were enlarged and distorted, and the smooth membranes of tire endo plasmic reticulum, the site of enzymes asso ciated with the metabolism of alcohol and other substances, proliferated. These liver changes were brought about by a rather moderate ingestion of ethanol, which did not result in any clinical signs of intoxica tion; the blood levels peaked at from 80 lo 90 milligrams of alcohol per 100 milliliters of blood, which is below the levels (100 or 150 milligrams) that constitute evidence of excessive drinking in riiost states. Drunken ness is no prerequisite for the development of liver damage. , In order to verify these effects and study their mechanisms in detail we needed lo reproduce them in an experimental animal. The laboratory rat is a common subject for Such experiments. Ethanol is usually fed to rats in their drinking water, however, and under these circumstances the animals gen erally refuse to take enough ethanol to de velop liver injury. Eeonore M. DcCarli and I countered the rat's aversion to alcohol by developing a new technique; we incorporat ed the ethanol in a nutritionally adequate but totally liquid diet, so that in order to eat or drink the animals also had to take the CIRRHOSIS or I.IVI'U was produced in a baboon after four years of heavy alcohol coitsunip. (ion. This section of Site baboon's liver, cnlnrcvd 70 diameters, shows the broad strands of fi brous connective tissue, like scar tissue, that have disrupted the once orderly arrays of liter cells, scurcealina nodules of irrceiihuly arraniied cells uud heavily concentrated clntiulc* of fat. ?.7_ AP00024569 ETHANOL HH II H--- C-----------C-------OH ' -J I : .H H ACETALDEHYDE H Ir h--c--c-- o II .H H H, ACETIC ACID H I -> H------ C-------C=0 II H OH -> CO, 4- h,o OXIDATION OF ETHYL ALCOHOL (ethanol) is accomplished in icveral stages! In- the liver cells two hydrogen atoms Are removed from each molecule of ethanol to form acetaldehyde. The acetalde hyde is normally oxidized, primarily in the liver, to form acetic acid (as acetates), which is eventually broken down Into carbon dioxide and water. The illustration omits tho various enzymes and cofactors. BLOODSTREAM'------ > '' ' ; ' LIVER CELL . .. . "V ____________1 _ _ _ 1 -ETHANOL--------- ^ - K DRUGS I NAD _ NADH ADH SYSTEM NADPH fcA MEOS J- NADP + H,0 S*. ACETALDEHYDE 4? --______........... J ALPHA-GLYCERO-.. PHOSPHATE Y PYRUVATE <rr LACTATE V GLUCOSE FATTY LIVER i LIPO- RUG ACETALDEHYDE PROTEINS ____________ ,______________jI______ OULLIIT1ECSO^ , METABOLISM OF ETHANOL in the liver cell by the nlcohol-dclydrogcnasc (AIMI) system nnd the microsomal ethanol-oxidizing yilcm (AtMtS) i* represented schematically, with the two ethanol athways nnd the movements of their products shown in color. Psitlirays that operate in the absence of ethanol urc shown in black and hose whose netivity is decreased by ethanol are shown by broken incs. In tho primary clhtmol pathway alcohol dehydrogenase eataS' he removal of hydrogen atoms (if) through reduction of a co '2 ' V- BLOODSTREAM - factor, nicotinamide adenine dinuctcotidc (\AD). The excess hydro gen is shunted into various processes. For example, it supplies the energy-producing electron-transport system, supplanting fats, the normal fuel. Hydrogen also contributes (o the synthesis of excess tri glycerides, or fats. A secondary pathway that comes Into play nl high alcohol levels utilizes the microsomal system, which nlso metabo lizes certain drugs and other foreign compounds. Here the cofactor is reduced nicotinamide adenine dinurlcotidc phosphate (.SADI'll). I ! AP00024570 cations. W{ could imagine two reasons for (he rats' (iii/nrc to develop hepatitis and cirrhosis.'liven with our liquid diet their alco hol intake did not exceed J6 peteent of the a! calorics, which corresponds to modcrconsumption in man. Moreover, it takes irom five to 25 years ofstcady drinking for a human being to develop cirrhosis, and a rat lives'for only about two years. We turned to the baboon, which is longlived and more closely related to man, and which can tolcralc liquid diets containing alcohol as 50 percent of the total calories, about the same as an alcoholic human be ing. At that level the baboons were visibly intoxicated, and when alcohol was tempo rarily withdrawn, t hey gave signs ofphysical dependence, such as tremors and seizures. At the Laboratory for Experimental Medi cine'and Surgery in Primates we put 16 ba boons on the ethanol-rich diet and 16 oth ers on an alcohol-free control diet with the same caloric content. Whereas the liver remained normal in the control animals, all 16 animals given ethanol soon showed ex cessive fat accumulation in the organ; five of them showed typical alcoholic hepatitis, and six of the animals kept on the ethanol diet developed cirrhosis after'between two ..and four yrs. The rat and baboon experi ments demonstrated clearly that severe liv'.er injury can be produced by prolonged heavy'ingestion of alcohol even when the diet is good. They also gave us for the first ' lime experimental animal models that re. producc all the liver lesions seen in the alco holic human being. The models have en ded us to find biochemical explanations or the early stages of alcoholic liver disease (fatty liver and the metabolic disturbances associated with it) and for some of the per sistent lesiQns that seem to lead to the more Severe stages: hepatitis and cirrhosis. The first step in alcohol's primary meta bolic pathway is catalyzed by the en LIVER FAT comes from three sources: from the diet, by way of the intestines; in the form of zyme alcohol dehydrogenase. (In spite ofrfee fatty acids mobilized from storage in adipose tissue, and through synthesis within the liver its name; alcohol dehydrogenase serves as itself. A fatty liver can develop if there is an excessive supply of fat or if disposal of fat from the a generalized remover of hydrogen atoms liver (by oxidation to carbon dioxide or hy secretion into the blood as lipoproteins) is impaired. from-various compounds, including some steroids; it was therefore probably present in the liver of the prehistoric men who first effects. One such pathway is the process tion of acute alcoholism, but it is often over sampled alcohol,-available to take on what whereby amino acids (derived from the looked. When an intoxicated person is has since become a major function.) Alco- breakdown in the liver of proteins) are con brought into a hospital emergency room, it hoi dehydrogenase catalyzes the transfer of verted, with pyruvate as an intermediate, is important to test for hypoglycemia, since hydrogen atoms from clhanol to a cofactor, into glucose. In the presence of excess hy crucial organs, including the brain, can be nicotinamide adenine diuuclcotide (NAD), drogen ions the process is turned in a differ critically affected by a lack of sugar; some converting the ethanol into acetaldehyde. ent direction: the pyruvate is reduced to deaths of "unknown origin" in alcoholics The acetaldehyde is then oxidized, primari lactate instead of being converted into glu may be attributable to the condition. ly in the liver, to form acetate, which is cose. Blood sugar is derived primarily from The lactate buildup from excess hydro eventually converted into carbon dioxide three sources: gluconeogcncsis, or synthesis gen lias other (fleets. The lactate moves into and water. A number of the metabolic ef from amino acids in the liver, breakdown of the blood, bringing on lactic acidosis. In the fects of alcohol are directly linked to the glycogen stored in the tissues and conver kidney it interferes with the excretion of two first products of its oxidation: hydrogen sion ofcarbohydrates in the diet. If the alco uric acid. A high uric acid level in the blood and acetaldehyde. holic person has been drinking and not eat (hyperuricemia) exacerbates gout, so that The excess hydrogen from alcohol unbal ing, there are no dietary carbohydrates and tile process 1 have described may explain ances the liver cell's chemistry. In order to glycogen may lie used up; if glueoncogene- the ancient clinical observation that exces live the cell must get rid of the hydrogen, sis is then blocked by the diversion of pyr sive drinking can trigger or aggravate gout and it docs so by shunting hydrogen ions uvate to lactate, the level of sugar in the attacks. iiito one or more of several pathways depen- blood will he lowered. Low blood sugar, or There are otlu-r ways in which the liver Icnl on them, sometimes with deleterious hypoglycemia, is known to be a complica cell can rid itself of alcohol's excess hydro- i l i f * if ;! f * 29 AP00024571 1.5 a* "DOU- | "5 3 8ai5cc 9 8S 3l< feg S3 CONTROLS ALCOHOL-PRETREATED ALCOHOL-FED RATS develop ft tendency toward hyperlipemia, or high blood fat. Animals were fed cither the control diet or the alcohol-containing diet of the same caloric value for about a month. Then, after fasting, the enimals were given a `'load" of one or the other diet. The rats' fasting blood-lipid levels had been about the same. In response to the load the animals that had been chronic alcohol consumers (right) developed markedly higher blood-lipid lev els, whether the load contained alcohol (colored bars) or did not contain alcohol (gray bars). 100' 'h \ 'X 50 'X V. i i\ t . i ii i Xi i i N * 10 0.. ! 11 1 l 1 1 l 1 l111 i 6 .8 10 '. 12 14 16 18 TIME (HOURS AFTER DRUG ADMINISTRATION) \ \ 20 22 24 ALCOIIQL ACTIVATES the enzyme systems of the smooth endoplasmic reticulum, stepping up the metabolism of certain drugs. The curves show the dis-nppenrnneo from the Mood of a dose or the tranquilizer meprobamate'given to volunteers before (bluet) and after (color) a m of alcohol ingestion. Alcohol activation cuts the drug's half-life (broken lines) in half. gen. Several of them involve the formation of lipids, or fut. The hydrogen can he shunt- ed directly info the synthesis of a/pha-gtycemphosphate and fatty acids. Those arc the two precursors of the triglycerides, and tri glycerides arc the lipids that accumulate in the alcoholic fatly liver. The main mechanism for disposing of hydrogen is more indirect, but it has a similar result. The hydrogen is transferred to the mitochondria, the cell organdies that produce the energy for liver functions. Normally it is fat that is oxidized--in effect burned--in the mito chondrial citric acid cycle to produce usable energy in the form of phosphate ions in an energized state. The plentiful hydrogen from alcohol, however, provides an alter nate fuel that is oxidized instead of the hydrogen from fat. In promoting the oxidalion of alcohol the substitution exacts a price: "the lipids accumulate, again leading to fatty liver. If alcohol is ingested along with a diet containing fat, the fat of dietary origin accumulates in the liver; even when alcohol is taken with a low-fat diet, fat made in the liver itself is deposited there. In addi tion, when alcohol is ingested in very large quantities, it can trigger hormonal discharges that mobilize fat from stores of adipose tissue and move it toward the liver [see illustration on preceding page], What can the liver do with the nccumulating fat? One possibility is for it to be secreted into the bloodstream, which deliv ers blood lipids to provide fuel for peripher al tissues such as muscle and deposits extra supplies-for storage in adipose tissue. The secretion is complicated by the fact that lipids arc not soluble in water; they have to be made soluble by being wrapped in a thin coat of protein to form lipoproteins. The assembly oflipoproteins is carried out in the liver in the smooth membranes of the endoplasmic reticulum. I have alluded to the proliferation of the smooth endoplasmic re ticulum that comes with heavy alcohol con sumption in both rats and human beings, which I observed with Oscar A. Iscri, Ber nard 1'.. Lane and Emanuel Rubin. The pro liferation is rellected, Jean-Gil Joly, Law rence Fcinman and I found, in the increased activity of certain enzymes in the smooth reticulum, which enlarges the liver's capaci ty for secreting lipoproteins. A liver that has thus adapted lo alcohol after being con ditioned by a period of heavy alcohol intake will respond with exaggerated secretion of lipoproteins even after a normal meal is eat en, producing hyperlipemia, or an abnormally high level of fat in the blood. Enrique Bamon.i and 1 observed that effect in rats, and William H. Pcrlow, Stephen A. Borowsky and I have now verified it in man. The effect is of particular significance in people who have underlying abnormalities of ci ther lipid or carbohydrate metabolism and therefore n propensity for elevated bloodlipid levels. Hyperlipemia is a major predis posing factor for heart attacks. Wlial is usu ally overlooked is that in individuals with preexisting hyperlipemia alcohol is proha- i , , ! ' , 1 | i j ! | : i . . j ! 30 .. AP00024572 0 S 10 15. 20 25 30 35 40 45 0 5 10 15 20 25 30 ' : TIME (DAYS) TIME (DAYS) ALCOHOL CALORIES are apparently not fully equivalent to othCr calories. When 8 basic 2,200-calorie diet was supplemented by 2,000 calories of alcohol (Itfi) and of chocolate (rixht), the gain in weight caused by the alcohol was much less marked and more erratic. 'i bly the single- most important aggravating creased from 16 hours during a control pe factor that is amenable to correction. riod to eight hours after a month of alcohol Still another way for the liver to dispose consumption. of excess fat is to convert some of it into the Anesthesiologists had known for many water-soluble breakdown products called years that larger doses of sedatives are re ketone bodies and secrete them into the quired to achieve a given effect in alcoholics bloodstream. In some susceptible people than in other people. That drug tolerance this response may be exaggerated, resulting was attributed to adaptation in the central in an elevation of ketone bodies in the blood nervous system: increased resistance to sed that mimics the condition known as keto- atives by the brain of the alcoholic. Our ` acidosis in diabetic patients. . finding pointed to a metabolic adaptation as well: increased capacity of the alcoholic liv The conversion of lipids into lipoproteins er to inactivate and excrete sedatives and is just one of a number of functions of other compounds detoxified by the endo the endoplasmic reticulum of liver cellsp, lasmic reticulum. At the adaptive stage of which also inactivates a wide variety of their disease (which, as vve shall see, does ' drugs and other foreign substances and con nol continue indefinitely) alcoholics there verts them into water-soluble products that fore require larger doses of many drugs. can be excreted [sec ``How ihe Liver Me That is true, however, only when the indi tabolizes Foreign Substances," by Attallah vidual is sober. When the alcoholic has been Kappas and Alvilo P. Alvares; SciENTinc drinking, the effect is quite the opposite. American, June, 1975], I therefore won- The reason, we found, is that one of the .dered whether the proliferation of smooth drugs the smooth reticulum metabolizes is reticulum after chronic alcohol consump alcohol itself, by way of an accessory path tion would be reflected in an increased ca way that supplements the basic alcohol- pacity of Ihe alcoholic liver to metabolize dehydrogenase system. various drugs. That turned out to be the DeCarii and I demonstrated the accesso case. After repeated administration bf alco ry pathway by spinning liver tissue in the hol (but before the evolution of severe liver ulIracentrifuge and isolating the endo injury) ihe enzymes of the smooth rclicu- plasmic reticulum as what is tailed the mi lum that inactivate tranquilizers, anticoag crosomal fraction. We then found that a ulants and other drugs and that detoxify preparation of the microsomal fraction certain food additives, canccr-causing sub would oxidize clhanol. Wc called the acces stances and insecticides do increase their sory pathway the microsomal ethanol-oxi activity, enhancing the body's capacity to dizing system, and KoirTcschkc, Kunihiko rid itself of these compounds. For example, Ohnislii and I were able to obtain it in a Prcm S. Misra and I gave Ihe tranquilizer semipurilied form. We have found that this meprobamate to volunteers and measured microsomal ethanol pathway conics into its rale of disappearance from the blood operation after the blood alcohol reaches a stream. The time for the blood level of the certain level. The alcohol (bus enters inlo drug to full to half its original value de competition with other drugs whose metab olism shares some elements of the microsomal system, thereby slowing down the me- tabolism of those drugs and enhancing their effect. That is why simultaneous drinking and taking of tranquilizers is particularly dangerous: ihe alcohol can accentuate the action of the drug not only because the ef fect of the two drugs on the brain may be additive but also because the presence of alcohol can interfere with Ihe liver's capacity to inactivate 1 he drug, so that at a given dosage more ofthe drug remains active for a longer time. Along with the rest of the microsomal enzyme systems, the ethanol-oxidizing system adapts to heavy alcohol consumption by increasing its activity, thus contributing to alcohol tolerance. The alcoholic's ability to drink more than most nonalcoholics is primarily due to brain tolerance, a progressive decrease in the effectiveness of alcohol's action on the brain. The alcoholic also devcl- ops an increased capacity to metabolize alcoho!, not only through me microsomal system but also through the alcohol-deitydrogenasc pathway and perhaps a third pathway that depends on the enzyme cat alase. After heavy or prolonged drinking, however, this adaptation can be offset by progressive liver injury', so that the livcr's overall ability to handle alcohol rcmains about the same or even decreases. The microsomal changes I have de scribed arc beneficial in that they help to rid the lit er of fat and speed the detoxification of many drugs, food additives and other foreign compounds. Hyperlipemia is one undesirable concomitant of Ihe adaptive process, and there are others. Some foreign substances are activated rather titan inacti vated by the conversions they undergo in 31 ,i I : j . jj -! J ; ,\ 1 , ; [ | f j i 1' ! t j t j * * ; J AP00024573 0 .1 2 34 S 6 TIME (HOURS AFTER ALCOHOL INFUSION) 7B ACETALDEHYDE.LEVEL IN BLOOD is higher in alcoholics than in nonalcoholic individ uals. Identical quantities of pure ethanol were Infused Into the blood of alcoholic and nonalco holic volunteers, achieving similar blood-alcohol levels. The acetaldehyde plateau was higher in the alcoholics {colored runt) than in the nonalcoholics (black). Apparently the alcoholics metabolized acetaldehyde less effectively, perhaps because of ethanol-induced liver damage. the endoplasmic reticulum. Certain poten tially cancer-causing substances become tarcinogenic only after activation by the microsomes, and other substances become o ' *o the liver itself only after such acti For example, exposure to carbon etmchloride is known to cause liver dam age, but the familiar dry-cleaning agent is larmlcss to the liver until it is activated by he liver-cell endoplasmic reticulum. Yssu;hi Hasumura and I found that the increase n microsomal activity induced by alcotol enhances (lie toxicity of carbon tetrathloridc: rats chronically treated with aljohol were much more susceptible to the oxic effects of carbon tetrachloride than patched control animals. This effect pre- tumably explains the clinical observation hat alcoholics are particularly susceptible o carbon tetrachloride poisoning in dry.leaning plants that still use the compound. The enhanced susceptibility of alcoholics .cry probably extends to a number of other breign compounds that are harmless to ttost people but that may become toxic in he alcoholic as an undesirable side effect of ncreased microsomal activity. Another side effect is energy waste. Mivosomal activity requires energy. Morewer, it is a peculiarity of the various microomal oxidations that they produce heat Vithout conserving chemical energy. This lonkl conceivably contribute to the im paired growth we observe in animals that !re fed ethanol, since the production of heal eyond whul is required to maintain body .mperature represents a waste of energy. Such waste may explain at least in part my observation with Romano C. Pirola that the addition ofethanol to the diet results in less weight gain than the addition of the same number of calories from other sources [see illustration on preceding page]. That is, with respect to body weight ethanol calories ap parently do not fully "count," at least in alcoholics and when the alcohol intake is large. Apart from the metabolic complications ** of excess hydrogen and of the adaptive changes in microsomal activity, heavy alco hol consumption has direct toxic effects on liver tissue. It appears that a central role in the toxicity of alcohol may be played by acetaldehyde, a product of alcohol's several metabolic pathways that is extremely re active and that affects most tissues in the body. Most of the acetaldehyde is convened into acetate by the liver mitochondria, but some of it escapes into the bloodstream. Both the alcohol dehydrogenase and the microsomal ethanol-oxidizing pathways be come saturated when the liver is loaded with a large amount of alcohol, so that the acetaldehyde level in the blood reaches a plateau and slays there until the alcohol level drops, at which point the microsomal pathway apparently cuts out. Mark A. Korsten, Sliohei Matsuzaki, Feinmnn and 1 found that the acetaldehyde plateau is sig nificantly higher in alcoholics than it is in nonalcoholic*, even when the same amount of alcohol has been given to huh groups and the same blood level of alcohol has been attained f.stv illuUraiion ui h'ft]. This abnor mally high acetaldehyde level in tdcoheta cou/cJ rc-Mt/f from faster metabolism ofaleo lio! to form acetaldehyde (because of the adaptive increase in microsomal activity) or from slower disposition of the acetaldehyde (because of impaired acetaldehyde metabo lism). Probably both factors arc involved, at least in the beginning. 1 have mentioned the striking alterations in the mitochondria re vealed by the electron microscope even in llie early stages of heavy alcohol consump tion. Hasaniura and I isolated the damaged mitochondria and found that their capacity for metabolizing acetaldehyde to acetate is reduced. Acetaldehyde itself may be re sponsible for part of the decrease in mito chondrial function: Arthur I. Cederbaum, Rubin and 1 sbosvcd that it has a toxic effect on the organelles. The alcoholic may there fore be the victim of a vicious circle: a high acetaldehyde level impairs mitochondrial function in the liver, acetaldehyde metabo lism is decreased, more acetaldehyde accu mulates and causes further liver damage. The acetaldehyde affects not only the Ihcr bul also other tissues. For example, Marcus A. Rothschild and his colleagues at the Manhattan Veterans Administration Hos pital have reported that acetaldehyde levels no higher than sve have measured in the blood of alcoholics can inhibit the synthesis of proteins in heart muscle. Such an effect could in part explain the impaired cardiac function that is common in alcoholics. Ac etaldehyde may affect the functioning of other muscles too. Acetaldehyde has striking effects on the brain. Several investigators have suggested that acetaldehyde rather than alcohol itself may be responsible for the development of the dependence that, along with tolerance, characterizes alcohol addiction. Depen dence manifests itself by a slate of extreme discomfort, often accompanied by physio logical disturbances such as tremors and seizures, produced by withdrawal of a drug. Several mechanisms mediated by acetalde hyde have been proposed to explain depen dence but none has yet been confirmed. One proposal begins with the fact that certain amine neurotransmitters, which transmit nerve impulses from one cell to another in the brain, arc inactivated by the enzyme monoamine oxidase lo form an aldehyde that is then converted into an acid. The conversion to acid requires an enzyme that is also active in the metabolism of acetal dehyde. Virginia E. Davis and Michael J. Walsh of the Houston Veterans Adminis tration Hospital have suggested that if ac etaldehyde is present in the brain, it may compete for the enzyme: unmetabolized neurotransmitter aldehydes may therefore accumulate; the aldehydes may then com bine with the ncurotransniittcr, forming compounds that are startlingly similar to certain morphine derivatives known lor (heir ability to piomote dependence. Another possibility, pointed out by Ger- APOOC aid Cohen nf Mount Sinai, is that the neelaldchydc coinhipcs; directly with amines to form jsoquinolmc derivatives, potent psy- choitclive compounds that could play a role in developing dependence. Alcohol depen- nce is probably determined by a number -> factors acting in conceit. It is at least possible, however, that acetaldehyde is in volved in the predisposition to alcoholism: a predisposed .person may have deficiencies of acetaldehyde metabolism that make for a higher blood acetaldehyde level and a pro pensity to dependence because of the ef fect of the higher acetaldehyde level on the brain. If heavy drinking continues, reversible fatty liver evolves, in most people, toward more severe and irreversible liver disease: hepatitis and then cirrhosis. Just why and how the liver may lose its ability to adapt to the alcohol load is not clearly established. Even in the fatty-liver stage there are some indications that more severe lesions are de veloping. There may be a ballooning of the liver cells, which is usually allributcd to fill accumulation. Ilaraonu, Maria-Anna Leo, llorowsky and I have noted that proteins, of which the liver is the major synthesizer, also pile up in the cells, whose ability to export proteins is somehow depressed by large amounts of alcohol. The engorgement of liver cells with fat and protein interferes with their normal functioning. So docs the reduced energy production to which l have alluded. For all these reasons some liver cells may die, and the necrosis can trigger an inflammatory process that is characteris tic of the stage called alcoholic hepatitis. The acute reduction in liver-cell function at this stage is enough to cause death in some patients. Cell necrosis and inflammation in turn promote the next stage: fibrosis, or the de velopment of scar (issue, the hallmark of cirrhosis. The fibrous connective-tissue bar riers between groups of liver cells interfere with the flow of blood to and from the cells, further decreasing liver function. Partially obstructed in the liver, blood backs up, in creasing the pressure in the porta/ system, which brings blood to the liver from the intestines. Abnormal channels may there fore develop in the venous system so that some of the blood can bypass the circulato ry block cslablished by (he cirrhotic liver. Sonic newly overloaded veins (notably in the esophagus) may become distended, and such veins, called varices, can rupture and hemorrhage. Bleeding varices are a major cause of death in cirrhosis. Moreover, un der high pressure plasma leaks out of the portal-system blood vessels. Extra lymph is also formed, and it leaks out ofthe lymphat ic vessels. Both sources contribute to as cites, the accumulation of fluid in the ab dominal cavity. A final complication stems from the liver's inability to clear from the blood ammonia and other nitrogenous com pounds produced by bacteria in the intes tines. As such compounds accumulate they act on the brain and may cause functional disturbances, hepatic coma and death. ALCOHOL FAT V / HYDROGEN ACETALDEHYDE HYDROGEN REPLACES,. ' -. FAT as fuel; and ..'pc FAT ACCUMULATES' . 'Y .V FATTY LIVER I/ /' . ` -v `-'.V " ' V' : -s HEPATITIS V CIRRHOSIS /\ CELL DEATH AND INFLAMMATION TOXIC EFFECTS ON LIVER CELLS DECREASED VITAMIN ACTIVATION DECREASED LIVER CELL FUNCTION OBSTRUCTION OF BLOOD FLOW Y . . / LACTA(E-dl> GLUCOSE KETOSIS V EXCESS LIPIDS :.-: l `' HIGH AMMONIA HIGH BILIRUBIN (JAUNDICE) fi ! ti ; iu i ti < HEPATIC COMA T ASCITES (FLUIOS IN ABDOMINAL CAVITY) HIGH LOW BLOOD LACTIC ACID SUGAR Nf V HIGH ACETALDEHYDE PORTAL HYPERTENSION HIGH URIC ACID LOW VlTAMiN LEVEL ^ \ '.j VA RUPTURED VARICES (HEMORRHAGE OF DISTENDED VEINS) KIDNEY MALFUNCTION GOUT MALFUNCTION OF HEART AND OTHER MUSCLE COMPLICATIONS of excessive alcohol consumption stem largely from excess hydrogen and from acetaldehyde. Hydrogen produces fatty liver and hyperlipemia, high blood luetic ncid and low blood suitor. The accumulation of fat, the e/Tcet of acetaldehyde on liver cells and other factors ns yet uuLnowii lend to alcoholic hepatitis*. The r step is cirrhosis. The consequent impairment of liver function disturbs blood chemistry, notably causing a high ammonia level, which can lend (o coma and death. Cirrhosis also distorts liscr slructurc, inhibiting blood How. High pressure in vessels supplying the liver may cause ruptured varices mid accumulation of fluid in nbdom. ittal cavity. There are individual differences in response to alcohol; ill piwticular, nut nil heavy drinkers develop hepatitis and cirrhosis. 33 AP00024575 Looking at Genes in the Workplace Screening for individual susceptibilities and monitoring of the workplace promise to stir up a host of legal, ethical, and scientific questions The Office of Technology Assessment (OTA) raised some eyebrows at a con gressional hearing last month when it said that at least 59 major corporations may be planning to inaugurate some kind of genetic testing of employees in the foreseeable future. This apparent surge of interest is occurring despite the fact that very few tests have been developed to screen employees for susceptibilities to certain chemicals, and their useful ness so far is questionable. Moreover, the whole idea of genetic screening is looked on with great suspicion by labor unions. The OTA's testimony was part of a report of preliminary findings from a study of the problems and promises posed by two types of genetic testing. One is biochemical genetic testing of individual workers for certain genetic traits; the other is cytogenetic monitor ing, which entails testing groups of work ers for aberrations in chromosomes that might occur from exposure to chemicals. The OTA said only six companies are currently engaged in such testing, down from 17 who have done it in the past 10 years. The dip mainly reflects a decline in testing for the sickle cell trait among blacks. The OTA presented its figures at hearings held by a subcommittee of the House Science and Technology Commit tee chaired by Representative Albert Gore, Jr. (D-Tenn.), who predicted that genetic testing "is likely to become the source of legal concerns very soon." Indeed, although the field is in its infancy it raises a host of concerns relat ing to confidentiality, the use of human subjects in research, discrimination in the workplace, and the proper use of findings--particularly when the implica tions of such findings are not known. ^The subject is especially sensitive beguse of its rancorous setting in the rid oflabor-management relations. In*sing interest in the subject is evi\ed by the OTA study as well as a Ynment-funded project at the Has- |lnstitute, which is examining the I implications. \two types of testing at issue are pes confused in the rhetoric of put Ihey pose somewhat different ftV Biochemical genetic screening Veptible" workers is the more politically controversial of the two since it is aimed at pinpointing individuals who may be unsuitable for certain types of jobs, but it falls into a time-honored tradition. In the past, makers of tar and creosote denied employment to fair skinned and freckled people because of their higher susceptibility to skin cancer. The railroads for years have been screening out applicants whose x-rays show a potential for back problems. More recently, companies have been grappling with the difficult problem of how to prevent pregnant women from being exposed to chemicals that might cause birth defects. Genetic screening is at least as problematic because it could lead to discrimination against certain ra cial and ethnic groups, and because in most cases it is impossible to predict when a person with a given trait will suffer from exposure to a given substance. There are probably thousands of ge netic deficiencies that render individuals unusually vulnerable to certain chemi cals, but only a few are known presently and several, such as the sickle cell trait, occur primarily to members of minority groups. The sickle cell trait, which some believe could interfere with the blood's oxygen-carrying capacity in a person ex posed to oxidizing chemicals, is proba bly not a very useful one to know about because there is no solid evidence that it makes any difference in the workplace. Indeed, the Air Force Academy recently reversed a 10-year-o!d policy of barring blacks with the trait from pilot training because no one could demonstrate that a low-oxygen environment would precipi tate a medical crisis. There are several other traits, though, that give cause for concern. One--which also predominates among minorities--is called G6PD or glucose-6-phosphate de hydrogenase deficiency. This occurs in 10 percent of American black males and a higher percentage of Mediterranean Jews. It can predispose carriers to a hemolytic crisis causing anemia from lack of oxygen resulting from exposure lo certain chemicals such as napthalene. Another much-talked-about deficiency is that for alpha-1 antitrypsin (AAT) which can predispose individuals to lung disor ders and emphysema from exposure to lung irritants. The severe form of the deficiency, which only affects one per son in 2000 or 3000, is clearly linked to emphysema. But among those who carry the heterozygous trait--2 to 4 percent of the population--no ill effects can be reli ably predicted. For a while durijtg the 1970's, Dow Chemical Company and Du Pont Corpo ration did limited testing for AAT and .G6PD deficiencies in employees as signed to work with cyanogenic com pounds. However, according to Du Pont's medical director Bruce Karrh, tests were terminated because they did not supply any additional useful informa tion. The only company Science could learn of (the OTA poll is anonymous) that currently does any genetic screening is Du Pont, which screens blacks for the sickle cell trait. This program was begun in 1972 at the request of black employees at a time when sickle cell anemia was getting a lot of publicity and screening for the trait was regarded as a public service. Although the information is not used for employment decisions but only, said Karrh, for the "information and edification" of black employees, the company came in for a good bit of criti cism following a series of articles in the New York Times in early 1980. There appears to be very little support at present for biochemical genetic screening within the scientific communi ty. The tests are regarded as arbitrary and, although valid, not very predictive. Says Gilbert Omenn of the University of Washington. "The extent of debate is way out of proportion to the scientific knowledge and any test applications at this time." This being so, the political objections to such screening begin to appear compelling. Union officials such as Anthony Mazzochi of the Oil, Chemi cal and Atomic Workers Union have decried such testing as having devastat ing potential for discriminating against certain groups of workers and creating, in effect, leper colonies of susceptible workers who may be assigned to lowpayingjobs because they are regarded as genetically unfit. Although companies are by no means rushing into the business of genetic screening, many critics believe they are attracted to it as an alternative to clean ing up the workplace. If genetically hy- 0036-8075/82/0723-0336S01.00/D Copyright 1982 AAAS SCIENCE. VOL. 217. 21JULV 1982 AP00024576 scientists who have already been subject to the audit say zealous and inexperi enced auditors may not be painting a fair portrait. At the University of Wisconsin at Madison, one indignant scientist said the audit team seemed preoccupied with equipment such as minicomputers and kept asking naive questions about why they were needed. There were no scien tists on the team, and answers provided seemed to go over the heads of the auditors. The scientist also complained that audits seemed to be held when the \ principal investigator was absent, a situ- j ation in which the person responsible for the equipment is not in a position to defend or explain apparent discrepan cies.--William J. Broad FDA to Reexamine Bendectin Data The Food and Drug Administration (FDA) has decided to take another look at all studies to date that might bear on the question of whether Bendectin causes birth defects. Bendectin is the only drug specifically approved for nausea and vomiting of pregnancy and it is taken by about 25 percent of all pregnant women in this country. The FDA last looked at Bendectin studies in September of 1980 at which time its panel of experts examined data from animal studies and 13 epidemiological studies and concluded that there is no demonstrated relation between the drug and birth defects. However, it is impossible to prove that any drug is harmless. The panel recognized a "residual uncertainty" about Bendectin's safety during pregnancy {Science, 31 October 1980. p. 518). Despite the FDA's conclusions, a growing number of parents arc convinced that Bendectin caused birth defects in their children. More than 100 lawsuits have been filed against Merreli-Dow, Bendectin's manufacturer, although the plaintiffs lost in the one case that did go to trial. Recently, public attention has been focused again on Bendectin as a result of newspaper stories indicating that there is new evidence which shows a strong link between Bendectin and birth defects. The FDA's decision to reexamine all the Bendectin data was prompted by a meeting on 8 April between FDA commissioner Arthur Hull Hayes, Surgeon General C. Everett Koop, Representative Doug Walgren (D-Pa.). and Harry Meyer, director of the National Center for Drugs and Biologies. Susan McFalis, a member of Walgren's staff, says (he congressman is concerned by some new data that he believes may implicate Bendectin in birth defects. There is a rat study showing that Bendectin may cause diaphrag matic hernias, a potentially fatal defect in which the stomach and other organs get into the lung cavity through a hole in the diaphragm. There is a monkey study showing that Bendectin may cause ventricular septal defect, a hole between the chambers of the heart. And there is evidence from a new in vitro test for teratogens that Bendectin may cause birth defects. "We have been concerned for some time because we have seen both these defects [diaphrag matic hernias and ventricular septal defects] in reports from physicians and patients," McFalis says. Ann Wilk, a reviewing pharmacologist at the FDA, notes the very preliminary nature of the new Bendectin studies. The rat study, done by Reimer Roll of Bundesgesundheitsamt in West Berlin, is still unpublished and the FDA has only a summary statement and reams of raw data which the agency is now having translated into English. Roll appar ently sees a slight incidence of diaphragmatic hernias at very high doses of Bendectin with no dose-response effect. But, says Wilk, the FDA cannot yet say anything about his methods. In the meantime. Roll is repeating his study and the FDA is repeating it "with pme modifications." ac cording to Wilk. The monkey study, conducted by Andrew Hendrickx of the Primate Research Center at the University of California at Davis, also is unpublished except in abstract form. Hendrickx. however, notified the FDA of his results in May of 1981. He gave 12 cynomolgus monkeys 10 to 20 times the normal human dose of Bendectin throughout the major period when organs are developing. Two of the monkeys aborted their fetuses. He then examined seven of the remaining fetuses about 2 months prior to term. In four of the seven, he saw an intraventricular septal defect, but it is not clear what this finding means because fetal monkeys normally have a hole in the septum earlier during develop ment. When Hendrickx examined the three monkey babies that were carried to full-term, he found that they were completely normal. Wilk asks of the septal holes that Hendrickx found in the four fetuses, "Was this a delay in development or would it persist until birth?" Merreli-Dow is now funding Hendrickx in a much larger study, which will take IV: years and will be double-blind. There will be four groups of monkeys, 20 pregnancies in each group, and three doses of Bendectin. In the meantime, another small-scale monkey study, involving nine animals and conducted by Harold McClure of Yerkes Primate Research Center, showed no effect of Bendectin on rhesus monkey fetuses. After hearing of these animal studies, the FDA requested that researchers conducting epidemiological studies, in cluding Boston University's Drug Epidemiology Unit and the Centers for Disease Control, look for an association between Bendectin and heart defects or diaphragmatic hernias. In these studies none has been found as yet. The in vitro study that McFalis referred to was conduct ed by John Hassell of the National Institute of Dental Research. He added Bendectin to cultured cells that nor mally would develop into cells resembling cartilage. The added Bendectin prevented this development. Hassell and Wilk concur that it is difficult to evaluate this study because no in vitro test has yet been validated--it has not been shown that these tests can reliably distinguish known teratogens from substances known not to harm fetuses. At the present time, (he FDA is considering a labeling change for Bendectin to reflect the possibility that the recent animal studies may turn out to demonstrate terato genicity. Bendectin's label now says there is no evidence of teratogenicity in animal tests. But the agency scientists feel that, in the final analysis, their best data will be from continued epidemiological monitoring of Bendectin. And, as yet, there is no persuasive evidence from human studies that the drug causes birth defects.--Gina Kolata SCIENCE. VOL. 217. 23 JULY I9g2 003*-8075/82723-0335S0l.00/0 Copyright 1982 AAAS 335 AP00024577 persensitive workers can be identified and removed from some plants, the argu ment goes, companies may be able to press successfully for less stringent oc cupational exposure standards. Some observers, including toxicologist Samuel Epstein of the University of Illinois, con tend that industry is putting a big empha sis on "weeding out the susceptiblcs" rather than taking the steps necessary' to clean up the environment entirely. Oth ers argue that the latter is not always technically feasible. Companies nowa days are getting very interested in pre ventive medicine and are trying to edu cate workers on the dangers of alcohol, smoking, overeating, and lack of exer cise. But some see even this as evading their real responsibilities and trying to "blame the victim." as Epstein puts it. If such plainly positive activities on the part of management are looked on with suspicion, then a very extensive set of legal and ethical safeguards indeed will have to be designed to make genetic screening acceptable. Cytogenetic monitoring, on the other hand, is not as controversial politically but probably more so scientifically. This involves monitoring a group of workers over time to test for increases in chromo some aberrations. What may be the earli est such program was carried on for about 10 years at Dow under the supervi sion of then medical director Jack Kilian who is now ai the University of Texas. The program proceeded smoothly until Kilian's group found what it regarded as significantly raised levels of chromo somal aberrations among workers ex posed to benzene and epichlorohydrin. The results of the studies are very con troversial and Dow. which had doubts about their validity, went back to a "re search mode." according to its present medical director. Now it is following groups of workers who have not been exposed to chemicals to see what effect other factors, such as age. smoking, and time of year, have on the rate of chromo some aberrations. Another instance of industrial involve ment in cytogenetic monitoring is a re search program at Johnson & Johnson. Researchers have been monitoring the effects of ethylene oxide, a sterilant gas. on workers at three plants with three different levels of exposure, matched with three control groups. After 6 months they found that employees in the plant with the highest exposure had sig nificantly higher incidence of a chromo somal abnormality known as SCE than the control group, according to a prelimi nary report by the company. They there upon discontinued the use of ethylene 23 JULY 1982 oxide at that plant. The company is also doing chromosome studies on individ uals before they are placed in jobs in volving potential exposure to ethylene oxide. Cytogenetic monitoring is of dubious value in determining an individual's sus ceptibility to cancer. All that can be said is that some substances that cause can cer cause aberrations, and. therefore, a group with high numbers of aberrations may be more susceptible. But an individ ual from that group who develops cancer may not be one with significant chromo some damage. The only good epidemio logical data linking chromosome aberra tions to cancer comes from Hiroshima and Nagasaki survivors. They were ex posed to massive radiation doses, which may have caused chromosome damage and other changes on a scale far greater than that caused by exposure to chemi cals. so the Japanese survivors may not tell us much about the effect of lower levels or cell damage. Nonetheless, some scientists, includ ing Kilian. believe there is no reason to delay in introducing cytogenetic moni toring into the workplace since the asso ciation between chromosome damage and cancer has been established beyond question. Margery Shaw of the Universi ty of Texas concurs. She believes enough close correlations already have been found--between mutagenicity and carcinogenicity; between carcinogens and substances that disrupt DNA; between chromosome breakage and SCE's, for example--that chromosome monitoring can be used as an indicator of potential health hazards much as badges are used to monitor workers' exposures to radiation. Marvin Legator, also of the University of Texas, contends that cyto genetic monitoring, in combination with other tests for abnormalities in semen and mutagens in urine, "can serve as an advance warning system" of hazardous chemicals in an environment. But. he says, progress has been stymied because industry does not want to use the tests until they are validated and industry is the only party in a position to validate them. Kilian asserts that industries do not want to get involved because cytoge netic monitoring will open up new possi bilities for getting sued by workers. Other scientists believe there is still too much uncertainty surrounding the technology. Richard Albertini of the University of Vermont says "the big problem with a monitoring system is that we don't have a disease outcome" and thus it is impossible to meaningfully in terpret ihe results of any tests. J. Grant t Brewen. cytogeneticist at Allied Chemi cal Corporation, puts it more succinctly: there is "not a shred of evidence." he says, that directly links chromosome damage to any disease. Brewen has an other problem with cytogenetic monitor ing: he says that counting structural ab errations in chromosomes is not particu larly useful because to achieve substan tial positive results would require very high levels of exposure, high enough to manifest themselves in systemic toxici ty. He believes there is more promise in looking at SCE's. which are numerically more sensitive. However, he warns that while chromosome changes can indicate effects of chemicals, the absence of such changes do not necessarily mean there has been no effect on the organism. Therefore, he says, cytogenetic monitor ing could lead to a "false sense of securi ty-" Most observers seem to agree that neither biochemical generic screening nor cytogenetic monitoring is going to be widely adopted any time soon. Although the OTA reports 59 companies are think ing about it. the uncertainty is too great and the potential usefulness too ques tionable for many to actually begin pro-' grams in the near future. Unions have a lot more things on their minds which seem to them more important than genet ic screening. And to industries, opening up new programs of genetic testing may just seem like asking for trouble, --Constance Holden 337 AP00024578 scientists who have already been subject to the audit say zealous and inexperi enced auditors may not be painting a fair portrait. At the University of Wisconsin at Madison, one indignant scientist said the audit team seemed preoccupied with equipment such as minicomputers and kept asking naive questions about why they were needed. There were no scien tists on the team, and answers provided seemed to go over the heads of the auditors. The scientist also complained t that audits seemed to be held when the \ principal investigator was absent, a situ- J ation in which the person responsible for the equipment is not in a position to defend or explain apparent discrepan cies.--William J. Broad J FDA to Reexamine Bendectin Data The Food and Drug Administration (FDA) has decided to take another look at all studies to date that might bear on the question of whether Bendectin causes birth defects. Bendectin is the only drug specifically approved for nausea and vomiting of pregnancy and it is taken by about 25 percent of all pregnant women in this country. The FDA last looked at Bendectin studies in September of 1980 at which time its panel of experts examined data from animal studies and 13 epidemiological studies and concluded that there is no demonstrated relation between the drug and birth defects. However, it is impossible to prove that any drug is harmless. The panel recognized a "residual uncertainty" about Bendectin's safety during pregnancy (Science, 31 October 1980. p. 518). Despite the FDA's conclusions, a growing number of parents are convinced that Bendectin caused birth defects in their children. More than 100 lawsuits have been filed against Merrell-Dow, Bendectin's manufacturer, although the plaintiffs lost in the one case that did go to trial. Recently, public attention has been focused again on Bendectin as a result of newspaper stories indicating that there is new evidence which shows a strong link between Bendectin and birth defects. The FDA's decision to reexamine all the Bendectin data was prompted by a meeting on 8 April between FDA commissioner Arthur Hull Hayes. Surgeon General C. Everett Koop. Representative Doug Walgren (D-Pa.). and Harry Meyer, director of the National Center for Drugs and Biologies. Susan McFalls, a member of Walgren's staff, says the congressman is concerned by some new data that he believes may implicate Bendectin in birth defects. There is a rat study showing that Bendectin may cause diaphrag matic hernias, a potentially fatal defect in which the stomach and other organs get into the lung cavity through a hole in the diaphragm. There is a monkey study showing that Bendectin may cause ventricular septal defect, a hole between the chambers of the heart. And there is evidence from a new in vitro test for teratogens that Bendectin may cause birth defects. "We have been concerned for some time because we have seen both these defects [diaphrag matic hernias and ventricular septal defects] in reports from physicians and patients." McFalls says. Ann Wilk, a reviewing pharmacologist at the FDA, notes the very preliminary nature of the new Bendectin studies. The rat study, done by Reimer Roll of Bundesgesundheitsamt in West Berlin, is still unpublished and the FDA has only a summary statement and reams of raw data which the agency is now having translated into English. Roll appar ently sees a slight incidence of diaphragmatic hernias at very high doses of Bendectin with no dose-response effect. But, says Wilk. the FDA cannot yet say anything about his methods. In the meantime. Roll is repeating his study and the FDA is repeating it "with ^ome modifications." ac cording to Wilk. The monkey study, conducted by Andrew Hendrickx of the Primate Research Center at the University of California at Davis, also is unpublished except in abstract form. Hendrickx. however, notified the FDA of his results in May of 1981. He gave 12 cynomolgus monkeys 10 to 20 times the normal human dose of Bendectin throughout the major period when organs are developing. Two of the monkeys aborted their fetuses. He then examined seven of the remaining fetuses about 2 months prior to term. In four of the seven, he saw an intraventricular septal defect, but it is not clear what this finding means because fetat monkeys normally have a hole in the septum earlier during develop ment. When Hendrickx examined the three monkey babies that were carried to full-term, he found that they were completely nbrmal. Wilk asks of the septal holes that Hendrickx found in the four fetuses. "Was this a delay in development or would it persist until birth?" Merrell-Dow is now funding Hendrickx in a much larger study, which will take 1'h years and will be double-blind. There will be four groups of monkeys, 20 pregnancies in each group, and three doses of Bendectin. In the meantime, another small-scale monkey study, involving nine animals and conducted by Harold McClure of Yerkes Primate Research Center, showed no effect of Bendectin on rhesus monkey fetuses. After hearing of these animal studies, the FDA requested that researchers conducting epidemiological studies, in cluding Boston University's Drug Epidemiology Unit and the Centers for Disease Control, look for an association between Bendectin and heart defects or diaphragmatic hernias. In these studies none has been found as yet. The in vitro study that McFalls referred to was conduct ed by John Hassell of the National Institute of Dental Research. He added Bendectin to cultured cells that nor mally would develop into cells resembling cartilage. The added Bendectin prevented this development. Hassell and Wilk concur that it is difficult to evaluate this study because no in vitro test has yet been validated--it has not been shown that these tests can reliably distinguish known teratogens from substances known not to harm fetuses. At the present time, the FDA is considering a labeling change for B.endectin to reflect the possibility that the recent animal studies may turn out to demonstrate terato genicity. Bendectin's label now says there is no evidence of teratogenicity in animal tests. But the agency scientists feel that, in the final analysis, their best data will be from continued epidemiological monitoring of Bendectin. And. as yet, there is no persuasive evidence from human studies that the drug causes birth defects.--Gina Kolata SCIENCE. VOL. 217. 23 JULY 1982 0O36#)75/82/D723-0335S01.00/0 Copyright 1982 AAAS 335 AP00024577 British Out of 11ASA, Americans May Stay In Proponents of continued American participation in the International Insti tute tor Applied Systems Analysis (NASA) are plugging away at finding an alternative to the U.S. government sponsorship that runs out at the end of the year. They are encouraged by the possibility that the American Academy of Arts and Sciences will agree to replace the National Academy of Sci ences (NAS) as the U.S. "national member organization" and are seek ing to raise money from industry and foundations to pay the U.S. way. The Reagan Administration last year decided to discontinue federal funding for IIASA (Science, 2 April 1982, p. 35), an East-West think tank established near Vienna in the early 1970's in the early glow of detente. When the withdrawal of government support became likely, an IIASA-U.S. Planning Group was formed in this country by those favoring a continued American presence at the institute. The group set out to secure alterna tive sources of funding and a replace ment for the quasi-public NAS, which is bowing out as national member organization. Approaches to the American Acad emy were made. Although those in volved in the efforts decline to discuss developments until the academy's council has acted, hopes appear to be high that the academy will consent to become the U.S. member. Founded in 1790, the American Academy is an honorary society with about 2300 members elected from among scholars and national leaders In the sciences, arts, and humanities. The organization, which is headquar tered in Boston, publishes the quarter ly Daedalus and operates an interdis ciplinary study center. Results of the planning group's ef forts to raise funds are also currently under wraps. Here too, however, there seems to be optimism that the campaign, headed by former NASA administrator T. Keith Glennan, will be productive in enlisting support from industry and foundations. Chairman of the planning group is Charles Maechling, Jr., a Washington attorney who negotiated the original U.S. member ship in IIASA. IIASA recently received another blow when Britain's Royal Society confirmed that Britain would withdraw from IIASA at the end of the year. The British departure will not impart as serious a financial jolt as that of the United States. Britain is one of 15 "Class B" IIASA members who con tribute about $350,000 a year each, while the United States and the Soviet Union have each been contributing some $2.3 million a year to the insti tute's operating budget of about $10 million. The official reason given for the British action was a change in re search policy at the Department of Environment which has been the source of funds for MASA member ship. There also were reports that influential members and staff of the Royal Society were critical of IIASA's research plan in particular, and less than enthusiastic about applied sys tems analysis in general. IIASA is said to be carrying on with assurances of continued backing from the Soviet Union and other member countries, is cutting administrativeoverhead and research plans to con form to the reduced budget and. figu ratively, leaving the gate open to make it easy for the strays to return to the fold.--John Walsh Using Experience to Calculate Nuclear Risks A new and provocative estimate* of the risks of nuclear plant accidents came to light in early July, the first such analysis to base its findings on the actual record of a decade of indus try performance. Earlier work has re lied more on probabilistic calculations than on experience. The new report, called the Accident Sequence Precursor (ASP) study, was written for the Nuclear Regulatory Commission (NRC) by the consulting firm, Science Applications, Inc. An antinuclear lobby, Critical Mass, re leased a draft of the report over the 4th of July weekend, citing it as evi-' dence that officials have drastically "Precursors to Potential Severe Core Damage Accidents: 1969-1979. a Status Report." by J. W. Minarick and C. A. Kukleika, published by the Nuclear Regulatory Commission (NUREG-CR 24971. June 1982. increased their estimates of the'risk of an accident. The report's chief finding is that during the 1970's, the chances for having a Three Mile Island accident were one per 1000 years of reactor operation. At Three Mile Island, the fuel core was partially melted. This contrasts with the official risk estimate published by the NRC in 1974, which said that a fuel core meltdown would occur only once in 20,000 years of operation. The^ second figure comes from the Reactor Safety Study (WASH-1400), prepared for the NRC by Norman Rasmussen of MIT. The comparison suggests that Ras mussen's work underestimated the real risk by a factor of 20. However, one industry spokesman, David Rossin of the Nuclear Safety Analysis Center, points out that WASH-1400 did not try to estimate the risk of the kind of accident that occurred at Three Mile Island. WASH-1400 looked instead at the chances for a total meltdown, a more severe and presumably rarer event. Rossin says it is wrong to compare WASH-1400 with ASP. One of the reasons ASP was com missioned was that the NRC was criti cized in 1978 for giving so much at tention to WASH-1400's probabilistic theory when a better source of infor mation--actual operating experi ence--was available. One review committee said the NRC ought to make a broad survey of nuclear plant operating records and use these to draw up risk estimates based on gen eral experience. The ASP report attempts to do this. The authors looked through 19,400 "licensee event reports" sent to the NRC between 1969 and 1979. After sifting through them, the authors iden tified 169 as "precursors of accident sequences," and 52 as "significant events." Drawing on these cases, they calculated the frequency with which certain accident sequences might occur and the frequency with which safety systems would fail. They concluded that severe core damage of the kind seen at Three Mile Island would occur between 1.7 and 4.5 times per 1000 years of reactor opera tion. Thus, if 1000 reactors were in operation, there would be at least one severe accident a year. At present there are only 74 commercial reactors in the United States. 338 0036-8075/82/0723-0338S01.00/0 Copyrighl 1982 AAAS SCIENCE. VOL. 217. 23 JULY 1982 AP00024579 efficient, and the morphology of largeKale jets argues strongly for nonrelativ' istic motion of the jets. If these jets are ' only mildly supersonic and dominated by nonrelativistic plasma, then the turbu lent boundary layer that surrounds the jet may pro^de the required' particle acceleration, magnetic-field amplifica tion, and transport of metal-enriched gas far from the interstellar medium. Slow jets ease the total energy require ments for the source, but not beyond the limits of credibility. Compact jets are also inefficient, and arguments can be made for their relativ istic or nonrelativistic motion. At pres ent there is no clear-cut resolution of this issue, but it is fair to say that a consistent picture could be constructed wherein all jets are nonrelativistic. Observations of jet phenomena in extragalactic radio sources seem to require modifications of earlier models for the central energy source, and. although these have yet to be done in detail, models of the interac tion of jets with their environment sup ply encouraging agreement with obser vations. References end Notes 1. A. Pacholczyk. Radio Astrophysics (Freeman, San Francisco. 1970), pp. 77-99. 2. D. De Yount and W. I. Axford, Nature {Lon don) 216. 129 (1967). 3. K. Kellcrmann and I. Pauliny-Toth, Astrophys. Lett. 8. 153 (1971). 4. A. Willis, R. Strom. A. Wilson. Nature {Lon don) 250. 625 (19741. 5. G. Miley. Annu. Rev. Astron. Astrophxs. 18, 165(1980). 6. F. Owen, P. Hardee, R. Bigncll. Astrophxs. J. 139. LI! 11980). 7. E. Schrcicr. J. Bums. E. Feigeison. ibid. 251. 523 (1981). 8. M. Rees. Nature tLondon] 229. 312 (19713. 9. R. Blandford and M. Rce*. Mon. Not. R. As* iron. Soc. 169. 395 (1974). 10. R. Blandford and A. Konigl. Astrophys. Lett. 20. 13 (1979). M. J. Bums et al.. Astrophxs. J.. in press. 12. D. Dc Young. Ann. N.Y. Acad. Set. 302, 669 (1977). 13. J. Eilck. Astrophxs. J. 254.472 (1982). 14. G. Miley. in Astrophyskal Jets. A. Ferrari and A. Pacholczyk, Eds. (Rcidcl, Boston, 1983), pp. 99-112. 15. D. De Young, Nature tLondon) 293. 43 (1981). 16. Astrophxs. J. 241.8) (1980). 17. R. Blandford and J. Ostnker. ibid. 221. L29 (1978). 18. K. Kellermann and I. Pauliny-Toth. Annu. Rev. Astron. Astrophxs. 19. 373 (1981). 19. T. Muxlow. personal communication. 20. S^t'nwin et al.. Astrophxs. J. 271. 536 (1983). 21. A. Marschcr and 3. Scott. Publ. Astron. Soc. $ac. 92. 127 (1980). 22. M. Norman. K. Winkler. L. Smarr. in Astro* physical JAfs. A. Ferrari and A. Pacholczyk. Eds. (Reidel. Boston. 1983). pp. 227-252. 23. D. De Young. Annu. Rev. Astron. Astrophys. 14. 447 (1976). 24. The National Optical Astronomy Observatories are operated by the Association of Universities for Research in Astronomy. Inc., under contract with the National Science Foundation. were adequate for present purposes without modification (4). Since 1977 numerous, bioassays of chemicals have been reported from in Criteria for Evidence of vivo studies; many in vitro tests have been undertaken and new in vitro tests have been devised. As a result much new Chemical Carcinogenicity research on the mechanisms of action of chemical carcinogens has been reported Interdisciplinary Panel on Carcinogenicity and suggestions put forward for classify ing carcinogens; the influence of biomet rics has been prominent in the develop ment of a variety of mathematical mod els for risk assessment: and there have Criteria for assessing the evidence for larly concerned with those areas that been controversies about the overall the carcinogenicity of chemicals have have been most affected by advances in quantification of risk from chemical car been described in several documents in knowledge. cinogens to the human population. These cluding a 1977 report of the Subcommit The panel expresses its concurrence factors have been influential in determin tee on Environmental Carcinogenesis of with the philosophy that was well ex ing the priorities for the present delibera the National Cancer Advisory Board pressed by the NCAB subcommittee (/): tions. (NCAB) (/) and the preambles to the monograph series of the International Agency for Research on Cancer on the evaluation of the carcinogenic risk of chemicals to hamans (2). The report of the subcommittee of the NCAB recom mended that "The general criteria should be reviewed on a continuing basis The criteria that are described are general guidelines and not rigid, universal criteria. The complexity of the problem dictates that the evaluation of potential human hazard's ofa given agent must be individualized in terms of the chemical and metabolic aspects of that agent, its intended usds). the data available it the time the decision must be made, and other factors pertinent to the case under consider Evidence Derived from Human Studies Clusters of cases of a specific type of cancer associated with a particular expo sure suggest that certain chemicals or combinations of chemicals may be carci and revised in the light of new knowl edge." The Interdisciplinary Panel on Chemi ation. Each case must be considered on its own and the criteria appropriate for one agent may not necessarily apply to another. nogenic to man. This has been the basis for much of our current knowledge of occupational cancer occurrence. The cal Carcinogenicity was convened by The panel also agreed that the general problems that confront the epidemiolo Philippe Shubik at the request of the ized definitions and significance of be gist in such situations include the reali American Industrial Health Council to nign and malignant neoplasm presented ty of the excess cancer incidence, the reevaluate the criteria for assessing the in the NCAB subcommittee's report difficulty of estimating exposure, and evidence for the carcinogenicity of chemicals as a result of recent increases in the quantity of data and research in carcinogenesis (J). The panel met on 31 October and I November 1983 and 24 and 25 February 1984 in Washington, D.C. The report that follows is particu The members of the panel were: Philippe Shubik. Green College. Oxford University. chairman: Arnold Brown. University of wisconsin-Madison; Charles Brown. National Cancer Institute. Scthesda. Maryland; J. R. P. Cabral, International Agency for Research on Cancer: David Clayson. Bureau of Chemical Safety. Health Protection Branch. Health, and Welfare. Canada: tan Higgins. University of Michigan; Wayne Levin. Holfmann-La Roche. Inc.; Peter Magee. Ftls Research Institute. Temple University. School of Medicine; Mortimer L. Mendelsohn. Lawrence Livermore National Laboratory: Robert A. Squire. Johns Hopkins University. School of Medicine: Bruce K. Bernard. Scientific Research Associates. Inc., rapporteur. Requests for reprints should be addressed to Dr. Philippe Shubik. Green College. RadclifTe Observatory. Oxford 0X2 6HC. England. M2 SCIENCE. VOL. 225 AP00024580 ihc exclusion or confounding factors (5). Potential carcinogenicity may be sug gested by descriptive (hypothesis gener ating) epidemiological studies that ex gens. Personal sampling has been used increasingly. For substances where por table sampling instruments are available, this provides a more relevant approach to the test compound. Determination that the incidence of neoplasms in creases as the result of exposure to the test compound requires a full biological, plore the relationship between exposure to individual exposure, We look forward pathological, and statistical evaluation. to some factor or factors and the devel to the time when such personal sampling Statistics assist in evaluating the biologi opment of cancer in general or canter at can be coupled effectively to health out cal conclusion, but a biological conclu specific sites. Delcriptive studies are usually classi- fied'as exploring relationships with re spect to time, place, or personal charac comes. Human data provide the only direct evidence that a chemical produces can cer in man. Negative epidemiologic re sion is not determined by the statistical results (6). Certain data derived from animal stud ies. although suggestive of carcinogenic teristics. Observations relate incidence sults cannot ensure complete absence of risk, may require additional study before of, or mortality from, cancer in general carcinogenic hazard from a chemical ora extrapolation to conditions of human ex or cancer at specific sites to the introduc process. Negative epidemiological re posure ,can be made. For example, there tion. increasing use, and removal of po -----------------5------------------------- tentially carcinogenic exposures; to dif ferences between geographical loca Summary. The Interdisciplinary Panel on Carcinogenicity reviewed and reevaluat tions: and to personal exposures at ed criteria for assessing evidence of carcinogenicity of chemical substances. The work, from hobbies, leisure activities, panel reviewed criteria applicable to data derived irom human epidemiological studies food, water, and behavior. A primary and from both in vivo and in vitro laboratory studies. A critical appraisal of all Ihese problem is posed by difficulty in exclud sources of information led to the conclusion that the characterization of human risk ing confounding factors. always requires interdisciplinary evaluation of the entire array of data on a case-by Analytical studies (hypothesis test case basis. Animal studies, whenever possible, should be augmented by studies of ing). in contrast, are carefully designed mechanisms, metabolism, and pharmacodynamics. Such studies may assist in to explore hypothesized relationships assessing risk to man. Recognizing the utility of such data should point the way for and again two types are recognized; (i) better assessment in the future. case comparison studies in which the past history of exposure and personal characteristics of persons with cancer suits may indicate that the duration from are different tumor end points in bioas are contrasted with the same characteris the first exposure was too short, the says conducted in vivo, some of which tics of those who do not have cancer and sample size was too small, the dose was are enhancements of the background in (ii) cohort studies in which the personal too low, or the substance to which peo cidence of neoplasms seen in the untreat histories and exposures of a sample of ple were exposed was not a carcinogen. ed control animals. It is not uncommon the population are documented and the In epidemiology, as in other disciplines, for some of the control rodents in such sample is followed over the years to it is impossible to prove a negative. The studies to develop incidences of even 100 determine the rate of occurrence of can lowest degree of risk that is likely to be percent of neoplasms of a particular cer. This rate is then related to the directly detected by epidemiological kind. If control animals develop, say, 50 categories of exposure determined at the means can be estimated, but risks below percent of a certain kind of neoplasm, outset. The latency of cancer introduces this can only be inferred by extrapolation and this incidence is significantly in difficulties into both types of study. from data obtained at higher exposures. creased in treated animals, or if there is a A helpful type of study is often possi Negative results thus indicate the limits decreased latency period for the occur ble in industries where good occupation within which a specific type of exposure rence of such tumors, this is usually al hygiene records have been kept. Here* will not affect the incidence of cancer in. classified as an example of carcinogene it may be possible to define a group of man. sis. In other studies the end point may be employees whose exposures can be esti Because of their central role in the represented by the occurrence of neo mated at some time in the past. By identification of human risk, epidemio plasms that are not seen at all or only relating categories of past exposure to logical studies are indispensable and re rarely seen in the controls. Instances of subsequent mortality the waiting period quire substantial expansion. Strengths decreased latency or increased incidence needed in current cohort studies can be and weaknesses of such studies, as is the of neoplasms in animals with a high eliminated. Such studies are usually re case with experimental data, should be incidence in controls requires full evalu ferred to as historical mortality studies. evaluated individually for robustness ation using a high level of statistical An important factor in assessing carci and weight. significance and. if possible, an analysis nogenicity in man is (he demonstration of the incidence in historical controls. of dose response. Whenever possible, The potentially large variation in the the dose estimated from level and dura Evidence from Long-Term Bioassays incidence of these neoplasms adds to the tion of exposure should be related to uncertainty in the evaluation of their cancer incidence or mortality. The carcinogenicity of a substance in significance. One of the main problems ofepidemio animals is established when administra The interpretation of these different logic studies has been the difficulty of tion in adequately designed and conduct patterns of tumor development in test measuring exposure and estimating ed experiments results in an increase in animals should recognize the likelihood dose. Often studies of occupational ex the incidence of one or more types of that they represent different mechanisms posure have been based on qualitative malignant (or, where appropriate, a com and may have different significance to assessments such as high, medium, or bination of benign and malignant) neo human risk. Each experiment must be low. Major improvements could be real plasms in treated animals as compared assessed according to the nature of the ized through better documentation of to untreated animals maintained under end points in the test animals, and differ human exposure to potential carcino- identical conditions except for exposure ent weights of evidence may be accorded ` 17 AUGUST 1984 ' 687 i I AP00024581 for purposes of extrapolation to humans. avoidance of confounding factors. The points and. when used in baUeries. are Instances in which the only manifesta strength of the data depends in part upon effective for identifying genetic effects of tion of carcinogenesis is a decreased its robustness. Robustness in turn is a chemical toxins. It is less clear how well latent period or an increased number of function of scientific design, execution, these tests identify chemical carcinogens neoplasms with a high incidence in con analyses, and interpretation of the study. or how they should be used in the ab trols may require validation in further An important clement is the rigorous sence of corroborative data on carcino experiments before being extrapolated to adherence to codes of good laboratory genicity. humans. Some of these enhancements practices (<3) and suitable quality assur The initially high correlation observed ihay be a result of relatively nonspecific ance procedures. between genetic change and carcinoge 'effects such as changed caloric intake, nicity has decreased with the enlarge the use of high doses leading to exces ment of the set of chemicals tested and sive cytotoxicity and cell proliferation, Evidence from Short-Term Tests with the separation of test development or other experimental conditions, which fronvtest deployment. Estimates of cor may not be relevant to human exposure. Early markers of neoplasia. Many in relations between findings in such tests Special bioassays of chemicals are vestigators have long hoped to find early an<^ determination of carcinogenicity in sometimes undertaken to demonstrate tissue changes that would predict carci vivo varies) depending on the chemical effects such as cocarcinogenesis and pro nogenicity, but this has so far eluded class, test type, and laboratory. At pres motion (7). These terms were originally discovery. A recent approach is the ef ent. the overall performance of short meant to describe specific experimental fort to correlate enzymc-aitcrcd foci in term tests, as judged by the proportion of conditions but are now widely used to the liver with eventual hepatoma occur correct results for chemicals classified describe modifications of the effects of a rence. So far. the correlation has been by carcinogenesis bioassay, is in the variety of carcinogens. There is no doubt too inconstant for this test alone to be a range of 50 to 70 percent. Although often that when the mechanisms of carcino satisfactory basis for classifying com significantly better than chance, these genesis are fully understood the role of pounds as carcinogens. A variety of oth results are not adequate to allow reliance certain modifying factors will prove to be er short-term tests for identification of on short-term tests alone in the determi of considerable practical importance. preneoplastic lesions or markers of neo nation of carcinogenicity. However, Presently, the relevance of these factors plastic transformations have been sug thousands of chemicals have yielded in human carcinogenesis cannot be de gested. but they are still at the research positive results in short-t'erm tests and termined, and for the moment such bio- stage; the panel expects that this area require further analysis. This has created assays can only be used as ancillary will receive increasing attention (9). a gnawing problem for scientists, regula evidence in assessing data for potential Tests for genetic alterations. Begin tors, industrialists, and the public. Pend human hazard. ning over a decade ago, in vitro tests for ing developments that may clarify these The statistical evaluation of carcino genetic changes were developed and rap issues arc: (i) continued improvements in genesis results makes use of experimen idly applied to the practical problem of the assays, particularly in regard to stan tal variables. These may include the time carcinogen identification. This approach dardization and metabolic activation: (ii) for observation of neoplasm (latent peri was spurred on by the belief that genetic expanding the chemical classes to which od), incidences of neoplasms, and time alteration in somatic cells is closely such assays respond: (iii) analyses of to death from neoplasms and other linked to one or more of the stages of large data bases of results from short causes. The determination of time to carcinogenesis, and by the early results term tests with carcinogens and noncar neoplasm often requires an experimental which showed that the coupling of meta cinogens so that patterns of response and design in which subgroups of animals are bolic activation to relatively simple bac validation of performance can be estab killed and examined at frequent intervals terial assays for mutation gave results lished: (iv) extensions of assay methods during the experiment in order to identi highly correlated with the carcinogenic to easily obtained human samples for fy the onset of neoplasia. The exceptions ity of certain groups of chemicals. coupling lo epidemiological studies; and to this added requirement are the in Approximately 100 of these tests are (v) clarification of those mechanisms of stances in which neoplasms, such as now available. They involve many orga carcinogenesis that these tests are in those in the skin, can be diagnosed clini nisms ranging from prokaryotes to hu tended to simulate. cally from the time of their first appear man cells and can be performed under Tests for transformation in vitro. ance. However, most experimentally in various conditions ranging from studies Changes in growth, colony formation, duced or enhanced animal neoplasms are of isolated DNA to observation of cells and colony morphology in culture have not rapidly fatal, and clinical observa in vitro and in vivo. The tests can be been related to exposure to carcinogens tions or terminal autopsies do not gener grouped into three general categories and to lumorgenicity when the affected ally provide a valid basis for estimating based upon their end point: cells are transplanted to animals. The lime to onset of neoplasia. 1) Tests for DNA damage including use of cell transformation has become a In most cases, neoplasms of different adduct formation, strand breakage, pro major research tool for those studying cellular origin appear to originate inde phage induction, and DNA repair. mechanisms of carcinogens by viral, pendently. Consequently, it is reason 2) Tests for mutagenicity, including chemical, and physical agents. The able for purposes of statistical analysis to forward and reverse mutation as evi methods have also been applied to the evaluate tumors of different cellular ori denced by alterations of DNA, gene identification of chemical carcinogens gins separately. products, or cellular behavior. but with less success and reportedly In all instances the robustness of the 3) Tests of chromosomal effects in poor predictivity. However, 'in view of data must be ensured; This may vary not cluding aneuploidy, structural aberra the potential inherent in these methods only with the particular model chosen for tions. micronuclei, and sister chromatid to elucidate some aspects of cancer cau the study, but also with practical condi exchange. sation. continued research and valida tions such as experience and training of In the main these tests are effective at tion of these methods is to be encour personnel, adequacy of facilities, and measuring their intended genetic end aged. 684 SCIENCE. VOL. 223 I AP00024582 Metabolism and Pharmacokinetics tigated method for modifying enzyme identification and provide logical ap activity. It can either enhance the forma proaches to methods of cancer preven Administration of a chemical to an tion rate of the active metabolite or stim tion. A working group from the Interna animal is usually followed by its absorp ulate other pathways that produce inac tional Agency for Research on Cancer tion and distribution throughout the ani tive metabolites. The magnitude of in (1ARC) (//) concluded that, at present, mal's body. In some cases the physical duction varies with the inducing agent, no classification of carcinogens accord properties of the chemical, such as lack the substrate used to assay the activity of ing to mechanisms could be exhaustive of solubility or the presence of an ionized the enzymes and the species, strain, and or definitive. However, elucidation of format normal physiological pH, may sex of animat. The biological conse mechanisms has value in the identifica lead cither to the substance not being quences of enzyme induction on the car tion and evaluation of specific carcino absorbed from the gastrointestinal tract cinogenic response depend not only on gens [12). or to its localization at the site of admin rates and pathways of metabolism but Our panel concurs with the lARC's istration. Once in the body, the chemical also on disposition and clearance rates of conclusion: in certain situations, the par will usually be metabolized by a number active metabolites. ticular mechanisms of action of some of enzymes that are present in different The situation becomes more compli carcinogens provide guidelines for pre tissues. Overall, the beneficial effect of cated when you consider the complex ventive measures. These situations in metabolism is to solubilize the chemical nature of the mixed function oxidase clude carcinogenic effects directly relat and thus facilitate its excretion from the system involved in the transformation of ed to the hormonal changes caused by body [10). many chemical carcinogens to active certain compounds and carcinogenic ef The pharmacokinetics of chemical car metabolites. For example, multiple iso fects in the bladder caused by com cinogens is of special importance be zymes of cytochrome P-450, that is. pounds that induce bladder calculi. cause at elevated doses the body may be forms with different amino acid se These carcinogenic effects can be dealt unable to clear the chemical as rapidly as quences, exist in a given tissue, each with differently from those of com it is administered so that, after a lag, the with different, but overlapping, substrate pounds that induce cancer without the chemical may attain toxic concentra capacities. Many are regulated indepen apparent intervention of other physiolog tions. Pharmacokinetic studies are also dently, so various factors lead to selec ical or pathological factors. important because they may determine tive increases or decreases in specific In spite of the inability to.derive a the concentration of a carcinogen or its isozymes. The amount of each isozyme generic classification of carcinogens, metabolites at particular sites. As in oth present in different individuals is an im case-by-case scientific evaluations may er enzyme systems, metabolic enzymes portant factorto consider in the variabili be reached for individual substances become saturated as the amount of sub ty of human response to chemical carcin Chemical carcinogens can, in principle strate increases. In such a situation, en ogens. be divided into two categories. One cate zymes with a low affinity for the sub These parameters limit extrapolation gory gives nonthreshold dose responses strate may become more involved in the of metabolic data from test species to the is stochastic in mechanism, and may metabolic process. Overall, such an ef human situation. Comprehensive meta have some probability of producing car fect will appear as a discontinuity in a bolic studies in different species can. cinogenic effects at any dose. The sec ' plot of the percentage of an agent con however, provide valuable data for com ond category gives threshold dose re verted to a specific metabolite versus the parison with that data obtained from sponses and, theoretically, has a no- administered dose of the agent. Such a human studies. Once the metabolic pro-' effect level. A few chemicals can be discontinuity has been called a metabolic file of a particular carcinogen is estab placed, provisionally, in one category or threshold and is important in assessing lished from animal studies, human tis the other: but for the bulk <5f chemical the effects of the high doses given in sues can be tested in vitro to determine if carcinogens, we are currently unable to animal studies relative to the lower doses the same metabolites are formed. discern in which compartment they fall. to which humans are normally exposed. In most in vitro tests the activation By dealing with chemicals case-by-case With many carcinogens, metabolism system determines the outcome. Numer and by studying mechanism, we can look leads to the formation of reactive species ous factors influence the metabolism of forward to doing better than this. as well as more readily excreted meta foreign chemicals in vivo, and results bolic prpducts. Reactive metabolites are from short-term tests are often influ formed from most major classes of enced by the selection and preparation of Extrapolation from Experimental Data chemical carcinogens, such as jY-nitroso the metabolizing systems. Consequent compounds, aromatic amines, polycyclic ly, no single system in vitro can be Animal experiments are commonly aromatic hydrocarbons, and aflatoxins. totally satisfactory or can exactly mimic conducted with higher exposure levels In each case the reactive metabolite is pharmacokinetic parameters that influ than those normally encountered by hu capable of reacting with macromol ence the fate and disposition of the test mans. Therefore, the first step in a quan ecules. including DNA, the target that is chemical in vivo. Nevertheless, when titative evaluation is to extrapolate the assumed to be critical for initiating carci used appropriately, in vitro metabolic results from high doses to effects from nogenesis. The chemical structures of activation systems can provide valuable doses corresponding to human exposure. the DNA adducts derived from these insight about the generation of potential The second step is to extrapolate across interactions have been established in ly carcinogenic intermediates. species from animal experimentation to many cases and lend support to the sus the human situation. Both steps .entail pected structure of the metabolite. large uncertainties (13). The activities of metabolic enzymes The Mechanism of Carcinogenesis may b'e markedly affected by the physio The first step, extrapolation from high to low doses, depends upon the pre logical and pathological condition of the The classification of carcinogens on sumed dose-response relationship. A host and by environmental factors. En the basis of their mechanisms of action number of mathematical models have zyme induction has been the most inves would greatly accelerate carcinogen been proposed for this extrapolation. 17 AUGUST 1984 1 > i I i AP00024583 each being necessarily simplistic. These .'. dose-response models are quantitative` ly similar to one another in the range of experimentally observable response rates, that is. 10 to 100 percent, yet they may yield substantially different esti mates at lower, unobservableTesponse rates. Estimates for low-dose responses , that differ by three to four orders of magnitude are not uncommon. This un certainty is the single most important limitation of the models. Current knowl edge does not yet allow the selection of one particular model, and experimental data from animal bioassays are not suffi cient to discriminate among the compet ing models. Cancer can occur through means other than exposure to a specific chemical. The manner in which this "spontane ous" response is incorporated into a dose-response model is another major source of uncertainty in extrapolating from high to low doses. Two methods have been proposed: one method as sumes that the process leading to sponta neous cancer is independent of the pro cess induced by exposure to the suspect agent; the other, that the two processes are identical. As with the different math ematical models for dose response, these two methods yield dose-response curves that are statistically indistinguishable in the observable response range but yield substantially different extrapolations. Other sources of uncertainty in extrapo lating from high to low doses include the possible existence of thresholds, deter mination of the effective dose at the site of action (compared to the administered dose), and the mechanism of carcinogen ic action. The second step in extrapolation, from the laboratory to the outside world, in volves additional sources of uncertainty. The route of exposure for the laboratory animal is often different from that for exposed humans, and the responses of the two species to the carcinogenic insult may be substantially different. Experi mental animals are often genetically ho mogeneous and share nearly identical environmental conditions; however, they do not always agree, qualitatively and quantitatively, in their carcinogenic response. In addition, they may differ in the site of their carcinogenic response. Humans are a genetically outbred spe cies living under widely diverse environ ments; they are exposed to a large varie ty of carcinogens and noncarcinogcnic modifying factors that may enhance or 'even inhibit the agent in question. There are unknown biases involved in extrapo lating dose response obtained under ho mogeneous experimental conditions to heterogeneous environmental condi tions. Different age-related patterns of expo sure may have a substantial effect on risk. Experimental animals are usually exposed to a near-constant level for most of their lifetime, whereas human expo sure patterns may vary widely from day to day. The results of extrapolating from one exposure pattern to another are de pendent upon assumptions about the mechanism of carcinogenic action, thus providing another degree of uncertainty to extrapolation. Further development of biometric models should be encouraged in order to provide statistical analyses for evalua tion along with all relevant scientific data. Pending elucidation of the mecha nisms of cancer, statistical estimations of the relation between exposure and re sponse will be most helpful when the models incorporate pharmacokinetic data and the time between exposure and tumor development, distinguish between the administered doses and target doses, and correct for the duration of exposure and competing risks. The most probable estimates should always be presented together with the confidence limits of the estimates. Enough data should be pre sented to show how well the estimates fit the experimental data; and assumptions incorporated in the model, together with any uncertainties, should be clearly stat ed. The Overall Assessment Process Chemical carcinogenesis is a rapidly moving field, and great quantities of data have been accumulated 'during the past decade. Even though an individual ex periment may yield only suggestive in formation, this information may be of considerable importance when consid ered together with other data (14). Clearly, when the primary source of data comes from epidemiological studies in man, it may be possible to evaluate a chemical and institute scientifically based preventive measures. However, even in the instances where data are available from humans, the data must be. supplemented with information from other sources before a conclusion can be reached. For example, toxicological evaluation of carcinogenicity has classically relied upon long-term in vivo studies as the primary source of data. Such studies have been performed in a routine man ner, and evaluations have followed pre determined formulas. This rote method is rapidly giving way to evaluations that take into account findings from in vitro tests, metabolism studies, and biometric analyses as well as any other available information. One of these methods alone cannot produce a reliable estimate of a chemical's risk to man. but taken togeth er they provide an estimate with a high level of confidence. Carcinogens act via different mecha nisms, which results in their having dif ferent magnitudes of risk to man. Even though there is no basis for the exact extrapolation of risk from experimental anjmal to man, ament advances, if ex ploited to t^ie fullest, can provide a basis for distinguishing the degrees of risk from different carcinogens. The scien tific criteria should be reviewed often, and scientific advances should be fully adopted. The scientific characterization of hu man risks from carcinogens involves the evaluation and integration of data from many disciplines. It requires scientific impartiality to review all appropriate data, both negative and positive, includ ing statistical estimations of low-dose response. Quantitative characterization of human risk requires scientific experi ence and judgment in order to weigh the evidence. Because of the strengths and weaknesses of the data to be evaluated in the assessment of human risk and the complexity of the problem, case-by-case analysis is most appropriate (IS). References and Notts 1. Subcommittee on Environmental Carcinogene sis, National Cancer Advisory Board. /. Nad. Cancer Inst. 58. 461 (1977). 2. International Agency for Research- on Cancer, Polynuclear Arnmaiic Compounds, part I. Chemical. Environmental, and Experimental Daia (IARC Monograph Series. International Agency for Research on Cancer. Lyon. France. 1983). vol. 32. pp. 13-31. 3. The panel acknowledges support of the work shop by the American Industrial Health Coun cil. 4. In (/). p. 461. column 2: "This subcommittee has found it useful to state generalized defini tions of malignant and benign neoplasms, recog nizing that in practice the diagnosis of a particu lar neoplasm is an operational one based on convention, experience, and experimental data. "A malignant neoplasm is composed of a population of cells displaying progressive growth and varying degrees of autonomy and cellular atypia. It displays, or it has the capacity for. invasion of normal tissues, metastasis, and causing death to the host. Benign neoplasms are a less autonomous population of cells, exhibit little or no cellular atypia or invasion or normal tisfiues. and do not metastasize. In particular cases, however, benign neoplasms may endan ger the life of the host by a variety of mecha nisms, including hemorrhage, encroachment on a vital organ, or unregulated hormone produc tion. The cytologic and histologic criteria uti lized in determining whether a lesion is benign or malignant differ depending upon the tissue in which the neoplasm arises. Evaluation of wheth er a specific lesion is benign nr malignant should, therefore, follow standard criteria used by experimental oncologists and pathologists with emphasis on correlation of the histopatho logic pattern with the biologic behavior of the lesion or type of lesion. In equivocal cases, the diagnosis ora specific lesion may require a panel of experts, recognizing that they may not always agree. "Depending upon the particular cafe, benign 686 SCIENCE. VOL. 225 AP00024584 t neopla^rm may represent a Mage in (he evolu* 9. For additional information on evidence from Cion of a malignant neoplasm and in other cases short-term tests, see H. Bartsch et a!., Mutat. *'' Cancer 7ns,. . 97711980): RLFalk they may be 'end points' which do not readily undergo transition to malignant neoplasms." Res. 76. 1 0980); C. Heidclbcrger et al.. ibid. 114. 283 (1983); International Commission for Fed. Proc.^FeHdo. cAl.mE.nvSiorocn. ,Ehxfpe.ahBl.ioPl rn39prc7t6. in inis repon, we have used the term "neo Protection against Environmental Mutagens and G' Pocl- N- L Kaplan. M. W. plasm'* freely to describe both benign and malig nant neoplasms. Carcinogens, ibid. 99, 73 (1982): W. K.. Lutz. ibid. 65, 28911979); M. F. Rajewsky, Specificity Foundation, 7V 1032 Rcicvanct ,l983>: Nutrition of Mo,,,, u.tr 5. For further information on evidence derived from human studies, see R. Doll and R. Peto, J. Natl. Cancer Inst. 66. 1191 11981): J. Higginson, of DNA DttmfigC in Chemical Carcinogenesis. in Molecular and Cellular A xpects of Carcino gen Srrermnj? Tests (1ARC Scientific Publica Htpmoma to Human Cnrcinoptnic Kirk (Nutri tion Foundation, Washington. D.C loan- p w ,Rc"J' V', S-T- A- M' Scfcu-wm!": C food Cosmei. Toxicol. 19. 539 (1981); G. 3. tion No. 27. international Agency for Research ^ 7?n?|b4'nf' i 'hnn*- A"k Toxicol. Suppl. Hutchinson, in Cancer Epidemiology and Pre lation, D. Scholtenfeld and J. F. Fraumeni. on Cancer. Lyon. France. 1980), pp. 41-34: A. C. Upton. D. B. Clayson. J. D. Jansen. H. 3. 79 (I9B0 ; C. Brown and J. Social SIAM Rer. 15. 151 (1983). 01' ds. (Saunders, Philadelphia. 1982), pp. 3-14: Rosenkranz, G. Williams. Afituu. Res. 133, I 14. For added information on the overall assess L. Tomatis. N. E. Breslow. H. Bartsch. ibid., (1984). ment process, see E. Farber, Am. J. Pathol. pp. 44-73: A. Lilienfeld andD. Lilienfeld.Foun- 10. For further details on metabolism and pharma 106. 269 (1982): Food Safety Council. Proposed dations of Epidemiology (Oxford Univ. Press. cokinetics. see J. Caldwell et al.. Food Tech Systemfor Food Safety Assessment\TooqSafe New Vorx, 1976). pp. 289-321: R. M. Madure no!.. in press: E. C. Miller and 3. A. Miller, in ty Council. Washington, D.C., 1980); National and B. MacMahon, Epidemiol. Rev. 2. 19 0980). The Metabolism of Chemical Carcinogens to Research Council. Commission on Life Sci 6. For more detailed discussion of evidence from Reactive Electrophiles and Their Possible ence^. Committee on Environmental Mutagens, long-term bioassays, see D. B. CUyson. D. Mechanism of Action in Carcinogenesis, in and Board of Toxicology and Environmental Krewski. I. C. Munro, Reg. Toxicol. Pharma Chemical Carcinogens. C. E. Searle. Ed. (ACS Health Hazards. Identifying and Estimating the col. 3. 329 (1983); Guidelines for Carcinogen Monograph No. 173. American Chemical Socie Genetic Intpatt of Chemical Mutagens (Nation Bioassay in Small Rodents, J. M. Sontag. N. P. ty. Washington. D.C.. 1976). pp. 737-762; E. C. al Academy Press. Washington, D.C.. 1982): R. Page. U. Saffioui. Eds. (National Cancer Insti Miller and J. A. Miller. Cancer 47. 2327 (1981): A. Squire. Science 214. 877 (1981): li.S Food tute Carcinogenesis Technical Report Series P. E. Thomas. L. M. Rcik. D. E. Ryan. W. and Drug Administration. Toxicological Princi No. i, N1H76-801. Department of Health. Edu- Levin. J. Bud. Chem. 258. 4590 (1983): R. H. ples for the Safety Assessment of Direct Food * cation, and Welfare. Washington. D.C.. 1976): Reitz. J. F. Quasi. A. M. Schumann. P. G. Additives and Color Additives used In Food R. E. Tarone, K. C. Chu. J. M. Ward. J. Nail. Watanabe. P. J. Gchring. Arch. Toxicol. Snppl. (U.S. Food and Drug Administration. Bureau of Cancer Inst. 66. 1175 (1981): J. H. Weisburgcr 3. 7911980). Foods. Washington. D.C.. 1982). and G. M. Williams. Science 214. 401 (1981). It. Joint working group from the International 15. In today's phraseology, this repon has been 7. Cocarcinogenesis was originally used as a term Agency for Research on Cancer. International concerned with scientific risk assessment and to describe enhancement of carcinogenesis Programme on Chemical Safety, and the Com hot with risk management. The process of risk when a carcinogen-and a noncarcinogen were mission for European Communities. Approodi- management, by definition, begins after risk applied together [R. D. Sail and M. J. Shear. J. rs to Classifying Chemical Carcinogens Accord assessment has determined that a risk to a Natl. Cancer Inst. 1. 45 (1940)]. Promotion, on ing to Mechanism of Action (IARC Internal human population exists. Whereas 'assess the other hand, described enhancement when Technical Report No. 83/001. Internationa! ment'' deals with biological significance, "man such agents were applied sequentially (1. Ber- Agency for Research on Cancer, Lyon. France, agement" deals with the possible alternative enblum and P. ShubiK. Br. J. Cancer I. 383 19831, regulatory actions. Included in risk management (1947)]. 12. For more details on the mechanism of carcino may be evaluations of costs, feasibilities, risk- B. Food and Drug Administration. Non-Clinical genesis. see U. H. Ehltng et al.. Mutat. Res. benefit ratios, availability of replacement sub Laboratory Studies Good Laboratory Practice 123.281 (1983): P.). Gehnng and G. E. Blau. J. stances or processes, and the lev$l of risk that is Regulations (43 Fed. Regis. 59986 09/8)1; Envi Environ. Pathol. Toxicol, I. 163 (1978): E. C. acceptable to the society in question. Manage ronmental Protection Agency. Toxic Substances Miller and i. A. Miller. Cancer (Philadelphia) ment of risks is a political, social, and economic Control. Laboratory Practice Standards (48 47. 1055 (1981), issue. Scientists acting as scientists hav< a role Fed. Regis. 5)922 (1983)]; Laboratory Practice 13. For further information on extrapolation from in (his phase, but it is limited to ensuring that the Standards. Pesticide Programs (48 Fed. Regis. experimental data, see K. S. Cramp. D. G. biological meaning of the risk is understood 53946(1983)]. Hoel. C. H. Langley, R. Peto, Cancer Res. 36. throughout the process RESEARCH ARTICLE of an untranslated 5' exon, thus facilitat ing c-myc gene expression at the transla tional level (8). release ofa transcription al repressor (9). a differential usage of r Antibodies to Human c-myc Oncogene promoters in normal and malignant cells or somatic mutations occurring al a high Product: Evidence of an Evolutionary level as a result of its proximity to the immunoglobulin locus (/0). Conserved Protein Induced During The protein product of the c-myc gene is. most likely, responsible for c-myc Cell Proliferation oncogenicity, and some information on potential properties of the c-myc protein has been obtained from studies with the HSkan'Persson, Lothar Hennlghausen, Rebecca Taub William DeGrado, Philip Lcder myelocytomatosis viruses. The transfor mation specific protein from MC29-type viruses is synthesized as part of a poly- protein with a molecular weight of 110,000 (110K). in which v-myc protein is fused to the gag protein (//). The The viral oncogene v-myc, which is noglobulin heavy chain locus, t(8;!4), or p| |0K*a*'n,'e polyprolein is found largely harbored by avian myelocytomatosis vi the immunoglobulin light chain loci. in the cell nucleus, binds to double- ruses. is derived from a cellular gene (c- 1(2:8) and t(8;22) (5). The molecular stranded DNA. and at least a fraction of myc) found in all vertebrates (/). The mechanism by which c-wyr oncogenicity the protein is associated with chromatin cellular myc gene contains three exon can occur, however, remains obscure (II). Two other members of the myelo- sequences transcribed from two promot although several mechanisms regarding ers located either just 5' of or just within the first exon (2-4). Considerable inter est has been shown in the c-myc gene since Burkitt lymphoma cells show translocations that have brought the cmyegene in close proximity to the immu- its activation have been proposed. These include a transcriptional activation of the c-myo gene resulting from the chromo somal translocation (6). a deregulation of the c-myc gene allowing constilutively high levels of expression (7), a removal Hdkan Persson. Lothar Hennighnuscn. Rcbccca Taub. and Philip Leder are associated with the Department of Genetic. Harvard Medical School, Boston. Massachusetts 02115; William DeGrado is associated with the Central Research and Develop ment Department. Experimental Station E328/229, E. I. DuPont deNemours and Company. Wilming ton. Delaware 19898. 17 AUGUST 19B4 087 AP00024585 1. REPORT NO. E'PA-450/5-83-006 TECHNICAL REPORT OATA frlcase read Instruction on the reverse before completing/ 2. 3. RECIPIENT'S ACCESSION NO. 4. TITLE AND SUBTITLE Review and Evaluation of the Evidence for Cancer Associated with Air Pollution 5. REPORT OATE November 1983 6. PERFORMING ORGANIZATION CODE 7. AUTHOR(S) I.C.T. Nisbet, M.A. Schneiderman, N.J. Karch, D.M. Siegel 9. PERFORMING ORGANIZATION NAME AND AOORESS Clement Associates, Inc. 1515 Wilson Boulevard Arlington, Va. 22209 8. PERFORMING ORGANIZATION REPORT NO. 10. PROGRAM ELEMENf NO. 11. contract/gRan^ NO. EPA Contract No. 68-02-3396 12. SPONSORING AGENCY NAME ANO AOORESS Pollutant Assessment Branch Office of Air Quality Planning and Standards U.S. Environmental Protection Agency Research Triangle Park, N. C. 27711 19. SUPPLEMENTARY NOTES 13. TYPE OF REPORT AND PERIOD COVERED Draft 14. SPONSORING AGENCY CODE This draft report is a comprehensive summary and compilation of the scientific evidence related to the hypothesis that cancer rates in human populations are associated with their exposure to pollutants present in the ambient air. Critical jComments on the strength and weaknesses of the studies are presented and general methodological problems in the conduct and interpretation of the studies are discussed. No overall judgments about the weight of the entire body of scientific evidence are presented. This draft Is being circulated for technical review and comment. H fU l lJIWWI IIIM, U R IA H ........ i. DESCRIPTORS Air Pollution/Cancer: Scientific Evidence KEY WORDS AND DOCUMENT ANALYSIS b. 10ENTIP16RS/OPEN ENDED TERMS Air Pollution/Cancer c. COSATI r-'icld/Group 8. O'STRIBUTION STATEMENT Unlimited 19. SECURITY CLASS (This Report! 20. SECURITY CLASS /This page) :PA For* 2220-1 4-77) prcvioui edition is peouETC 21. NO. 22, PRICE AP00024586 and case reports of mesotheliomas among residents in neighbor hoods surrounding asbestos mines and factories. However, these studies yielded no specific evidence for exposure other than location of residence. Environmental exposure to asbestos also results from other activities (e.g., wearing out of brake linings in automobiles), in one study of urban dwellers, nearly all (961) had asbestos fibers in their lungs (Churg and Warnock 1977). This suggests that asbestos from diverse sources, parti cularly airborne asbestos, may be an important problem for additional study. 3. Vinyl Chloride Cases of the. rare cancer, angiosarcoma of the liver (ASL), have been reported among individuals living near vinyl chloride fabrication, or polymerization, plants. Brady et al. (1977) studied the cases of ASL reported to the Tumor Registry of the Cancer Control Board of the New York State Department of Health during the years 1958-1975. For each of these cases a matched control with an internal malignant tumor other than primary liver cancer was selected from the registry. Cases and controls were matched on age (same 5-year age group), race, sex, county of residence, and vital status. Relatives of both the subjects and matched controls were interviewed in order to obtain information on potential exposure to vinyl chloride (VC), arsenic (Ab), or thorium oxide (Th02), as well as medical, familial, residential, and occupational histories. Of the 26 cases of ASL diagnosed .durina^l958-1975", 3 had direct exposure 11-40 AP00024587 DRAFT to VC, As, or Th02 (p<0.02). Of the remaining 19, 5 lived within one mile of a VC fabrication or polymerization plant. Although this is suggestive of an association, no statistical test of the possibility of this finding arising by chance was reported. Due to the small number of cases and the lack of monitoring data directly demonstrating exposure, no firm con clusions are possible. Infante (1976) studied the mortality patterns of residents of four Ohio communities with polyvinyl chloride (PVC) production facilities. Using data for the Ohio white population as the standard, SMRs were calculated for central nervous system (CNS) cancer, leukemia and aleukemia, and lymphomas. He found that in these four communities the number of observed CNS cancers for both sexes combined during 1958-1973 was significantly greater than that expected (38 observed vs. 24.07 expected p<0.001). SMRs were also calculated for each of the counties excluding the areas surrounding the PVC facilities, but no significant excesses were found. This study was reviewed by Air Products and Chemicals (1980), who commented that interpretation of this study is complicated by the fact that (1) the increase in CNS tumors was observed primarily in males, and (2) most of the excess occurred in one part of the study area (Painesville). They argued that these factors seriously challenge any conclusions of association of vinyl chloride with community cancer risks. i. 11-41 11 AP00024588 Infante (1976) has also been criticized by the Society of the Plastics Industry (1980) for including North F,idgeville in the study group while not including other cities located as close as North Ridgeville or closer to the PVC facilities (e.g., Mentor, Ohio). If North Ridgeville is excluded from the study group, the excess in CNS tumors remains significant (p<0.05, one-sided test), however. __| 4. Petrochemical and Other Chemical Emissions A number of studies have indicated that workers exposed to a wide range of industrial chemicals are at increased risk of developing cancer (Althouse et al. 1980). An increased risk of bladder cancer has been reported among workers exposed to benzidine (Case et al. 1954) and paints (Cole et al. 1972). Exposure to polycyclic aromatic hydrocarbons (found in crude petroleum, catalytically cracked oils, soot, and other pyrolysis products) has been associated with increased incidence of cutaneous and pulmonary cancers in workers (Doll et al. 1972, Lloyd 1971, Hammond et al. 1976, Fraumeni 1975). Blot et al. (1977) studied cancer mortality patterns for 1950-1969 in the U.S. counties where the petroleum and petro chemical industries are most heavily concentrated. Using methods similar to those of Blot and Fraumeni (1975) described above, it was found that male residents of these counties experienced significantly higher rates' for cancers of the lung, nasal cavity 11-42 AP00024589 REFERENCES ?AIR PRODUCTS AND CHEMICALS CORP. 1980. Comments on Environmental Protection Agency Proposed Rulemaking: National Emission Standards or Identifying, Assessing and Regulating Airborne Substances Posing a Risk of Cancer (44 FR 58642). OAQPS Doc. No. 79-14 > 'AIR PRODUCTS AND CHEMICALS CORP. 1981. Posthearing Comments on Environmental Protection Agency Proposed Rulemaking: National Emission Standards for Identifying, Assessing and Regulating Airborne Substances Posing a Risk of Cancer (44 FR 58642). OAQPS Doc. No. 79-14 ALBERT, R.E., PASTERNACK, B.S., SHORE, R.E., LIPPMANN, N.N., and FERRIS, B. 1975. Mortality patterns among workers exposed to chloromethyl ethers--a preliminary report. Environ. Health. Perspect. 11:209-214 ALTHOUSE, R., HUFF, J., TOMATIS, L., and WILBOURN, J. 1980. An evaluation of chemicals and industrial processes asso ciated with cancer in humans based on human and animal data. IARC Monographs Vols. 1-20. Cancer Res. 40:1-12 AMERICAN CANCER SOCIETY. 1980. Cancer Facts and Figures, 1980. New York AMERICAN INDUSTRIAL HEALTH COUNCIL (AIHC). 1980. Comments on Environmental Protection Agency Proposed Rulemaking: National Emission Standards for Identifying, Assessing and Regulating Airborne Substances Posing a Risk of Cancer (44 FR 58642). OAQPS Doc. NO. 79-14 AMERICAN INDUSTRIAL HEALTH COUNCIL (AIHC). 1981. Posthearing Comments on Environmental Protection Agency Proposed Rulemaking: National Emission Standards for Identifying, Assessing and Regulating Airborne Substances Posing a Risk of Cancer (44 FR 58642). OAQPS Doc. No. 79-14 AMES, B.N., DURSTON, W.E., YAMASAKI, E., and LEE, F.D. 1973. Carcinogens are mutagens: A simple test system combining liver homogenates for activation and bacteria for detection. Proc. Natl. Acad. Sci. USA 70:2281-2285 AMES, B.N., MCCANN, J., and YAMASAKI, E. 1975. Methods for detecting carcinogens and mutagens with the Salmonella/ mammalian-microsome mutagenicity test. Mutat. Res. 31: 347-364 1 AP00024590 GGT is a highly sensitive indicator of liver disease, regardless of the disorder's origin. Gamma-Glutamyltransferase in the Laboratory Evaluation of Liver Disease Leslie M. Shaw, Ph.D., Associate Prof, of Pathology, U. of Pennsylvania School of Medicine, Philadelphia KEY POINTS 1. The biochemistry, physiology, tissue distribution, and assay meth ods for gamma-glutamyltransferase (GGT) are discussed in detail. 2. GGT is a highly sensitive indi cator of liver disease, regardless of whether the origin is inflammatory, apace-occupying, or obstructive. It is a useful screening test for alco holism, particularly because GGT levels are not elevated in social drinkers. 3. GGT is not elevated after infan cy, in bone or muscle disorder, or during pregnancy, as are other en zyme tests. Whereas the transaminases, ala nine aminotransferase (ALT) and aspartate aminotransferase (AST), and alkaline phosphate (AP) could be termed the "classical" liver en zymes, gamma-glutamyltransferase (GGT) is a relative newcom er, particularly in clinical practice in the United States. Much of the early clinical work on this interest ing enzyme has been done in Eu rope.1 In fact, in most European countries, requests for GGT as an integral parameter of the hepatic profile are at least as frequent as those for ALT and AP.8 In recent years, the use of GGT in the United States has been grow ing at a very rapid pace, in fact, much faster than for any other liver test. Most likely, the reasons for this recognition of GGT are its specificity for liver diseases, its reputation as the most sensitive enzyme test of liver disease, and its growing use in the laboratory as sessment of the chronic alcoholic. This article attempts to interpret some of the recent studies that have stimulated interest in GGT. Structure and Function of GGT Before we discuss the use of GGT as an indicator of chronic FEBRUARY 1985 MEDICAL TIMES 65 AP00024591 alcohol consumption and in the detection and differentiation of liver disease, it will be of interest to review briefly the biochemistry, physiological role, and tissue dis tribution of this enzyme. GGT is widely distributed in human tissues. It is present in very high concentrations in the brush border of the proximal tubules of the kidney. In fact, the kidney is the human tissue richest in GGT. The content of GGT is second and third highest in human pancreas and liver, respectively. At the cel lular level, GGT is located primari ly on the outside surface of cell membranes (Figure l).4 The pri mary localization sites in the nor mal human liver are the canalicular portion of the hepatocyte mem brane and, to a lesser extent, the plasma membrane of epithelial cells lining the bile ducts.5 What are the location and origin of elevated GGT in various liver diseases? If we take alcoholic fatty liver as an example, we start to see spreading of GGT to parts of the hepatocyte membrane beyond the canalicular region.6 The enzyme is still located on the outside surface of the membrane, but in much higher concentrations and spread out over the entire membrane sur face. In cholestatic liver disease, where one primarily sees prolifer ating bile ducts, high concentra tions can be found in the membrane fraction of proliferating ducts.7 A more detailed look at the structure of GGT reveals that the enzyme is composed of two sub units, a smaller and a larger one.8 One end of the large subunit has a sequence of approximately 40 hy 66 drophobic amino acids which are probably responsible for the tight binding of the enzyme to the lipid matrix of the plasma membrane.8 The remainder of the molecule is coated with carbohydrate.10 So one can think of GGT as a plasma membrane-bound glycoprotein with the hydrophobic part of the molecule serving as an anchor to the lipid membrane and the hydro philic sugar groups facilitating in teraction of the enzyme with its plasma environment. What is the role of this enzyme in human metabolism? To the best of our knowledge, the primary physiological role of GGT is the hydrolysis of glutathione.11'1* The products of the hydrolysis reaction are L-glutamate and the dipeptide, cysteinylglycine. Cysteinylglycine is further broken down into its constituent amino acids. The three amino acids L-glutamate, cysteine, and glycine now become available for reutilization. One of the hall mark aspects of intermediary me tabolism is that the chemical con stituents of the human body, as well as of all living organisms, undergo continuous synthesis and breakdown. We do not always un derstand why that is the case,.and we certainly do not understand why it is important for glutathione to undergo synthesis and break down continuously. But, neverthe less, it does occur, and the key enzyme in this catabolic pathway is GGT. It should be emphasized that sconjugates of glutathione are also natural substrates for GGT. These conjugates are produced in the body through the glutathione-s- transferase pathway after exposure to drugs or environmental toxins. By catalyzing the hydrolysis of these s-conjugated compounds, GGT plays a role in one of the pathway^ for the elimination of some drugs and toxins referred to as the mercapturic acid pathway.13 Test Performance and Prelaboratory Factors In interpreting a laboratory re sult (normal or abnormal), one has to be concerned with three main factors: the technical aspects of the test and the extent that they might influence the test result, the prela boratory or nondisease factors that might produce an abnormal result, and the possible pathological caus es of an abnormal result. Let us briefly discuss the first two factors before we consider the interpretation of abnormal results. As far as technical performance is concerned, the most important pa rameters to consider are tempera ture and pH control, the choice of substrate, the concentration and purity of the substrates, buffer se lection, selection of the wavelength for monitoring the reaction, and, of course, the standardization and performance of the instrument with which the test is performed. The International Federation of Clinical Chemistry (IFCC) has re cently proposed a reference meth od for the assay of GGT which has several advantages.14 The reaction scheme for this assay is shown in Figure 2. This in vitro assay takes advantage of the ability of GGT to. transfer the y-glutamyl moiety of the donor substrate to a variety of acceptor substrates, including MEDICAL TIMES VOL. 113, NO.2 / AP00024592 Blood Plasma * 4 H 'I n ,.!qfl >1 A !fJi !i ti 1. Cell Membrane Location of GGT--This schematic presenta tion indicates the location of GGT on the surface of the cell membrane. At this membrane site, GGT catalyzes the hydrolysis of glutathione (GSH) and glutathione conjugates. water, a variety of amino acids and small peptides, as well as the donor substrate itself. Both the donor and the acceptor substrate are nonphysiological molecules that were se lected to achieve optimal test re sults. The L-y-glutamyl-3-carboxy-4-nitroanilide (the donor substrate) is a colorless compound. The rate of formation of the yel low-colored product (5-amino-2nitrobenzoate) of GGT's catalytic action on the donor substrate is determined by monitoring the pro duction of the yellow color at 410 nm. The reaction is carried out at the optimal pH of 7.9 and at 30C. To my knowledge, this is the simplest and least expensive en zyme assay in the world. It consists of only two reagents, L-yglutamyl-3-carboxy-4-nitroanilide and glycylglycine, in addition to water. Glycylglycine, the acceptor substrate, plays a dual role in the assay since it also serves as the reaction buffer. An important fea ture of the donor substrate in the IFCC method is the presence in the nitroanilide ring of a carboxyl group. This facilitates the water solubility of the donor substrate.14 The noncarboxylated donor sub-. strate is only poorly soluble in water and requires prior heating or the use of dilute HC1 for solubiliza- FEBRUARY 1985 MEDICAL TIMES 67 AP00024593 tion. The carboxylated substrate is considerably more stable in aque ous solution than the noncarboxylated substrate. Another point of interest is that most reference methods are not always easily adaptable to the rou tine laboratory. In this case, we have an example of a reference method that is probably simpler to perform than most GGT routine methods currently in use. The second area of consideration is to determine what effects non disease or prelaboratory factors, such as serum sample storage time and temperature, alcohol consump tion, medications, diet, exercise, age, and sex, have on GGT activity in patients' sera. Serum GGT activity is quite sta ble. Activity remains constant for five days at 4C and for at least seven months at --70C. Thus, if samples require storage for more than one working day, no error will be introduced by storage in a re frigerator or freezer until the assay can be performed. The most noted clinical feature of GGT is its sensitivity to alcohol consumption. There is now a large convincing body of evidence de scribing the response of GGT to various types and doses of alcohol. 1B~18 The key question here is, "How much alcohol must an individual consume before one can observe an interpretable elevation of GGT?" Because of the para mount importance of this question in the face of the magnitude of the nation's alcoholism problem, we will later give this question some special consideration. Another important and wellknown nondisease parameter is the type of medication. GGT activity in serum is elevated by medica tions, particularly the anticonvul sants which produce elevations up to about four times normal.19 Small increases in serum GGT activity above the reference range are seen in some people who are grossly obese. No significant change in GGT activity occurs after strenu ous exercise. A recent study on reference ranges of GGT in children makes an important point concerning the influence of age on the interpreta tion of hepatic enzyme levels.20 These authors demonstrated that the immature liver, which has higher concentrations of GGT than the normal adult liver, pours out more GGT into the infant's blood stream, so that in the first six months of life we see much higher serum levels compared with older age groups. But the important point thaf these investigators make is the fact that after six months of age the reference interval for GGT does not vary much with age in young, growing children. In con trast, alkaline phosphatase does not exhibit that degree of stability in this age group, due to out put of bone osteoblastic alkaline phosphatase.21 As a result, we see higher levels throughout childhood than we see in adulthood. Even more compromising is the spike of alkaline phosphatase at puberty which may vary from individual to individual in intensity and time of onset. It is, therefore, difficult to develop reference ranges in this age group that allow maximum sensitivity in detecting liver dis ease. The conclusion of this work is that after six months of age, GGT is a much more reliable indi cator of liver disease in children than alkaline phosphatase. The serum levels of GGT in the normal human are somewhat sexdependent, with males having higher levels than females.22 One should, therefore, have access to male and female reference ranges. \ \ L--y-glutamyl-3-carboxy-4-nitroanilide (y-glutamyl donor substrate) S-atnino-2-nitrobenzoate l GGT + + glycylglycine (acceptor substrate) -y-glutamyl-glycylglycine 2. GGT Assay--The reaction scheme for GGT assay as recently proposed by the IFCC. Interpretation of Elevated Results There are a variety of clinical and subclinical conditions that can lead to GGT levels elevated above the individual subject's own refer ence value as well as above the reference range of a healthy popu lation. As far as the IFCC reference method is concerned, well-defined 68 MEDICAL TIMES VOL. 113, NO.2 AP00024594 reference intervals for male and female populations will soon be forthcoming. Due to procedural optimization, these intervals will be somewhat higher than the ones currendy in use. Since one of the prime uses of GGT is in the recognition and mon itoring of the chronic alcoholic, some of the most informative clini cal studies on this subject will be reviewed. A very interesting epide miologic study took place in Malmo, Sweden, and was reported in Preventive Medicine. In con junction with a department of pre ventive medicine, this group looked at two populations of male birth cohorts, one bom in 1926, the other in 1927. Total participa tion was 3,050, which represented 76% of all males born at that time in that city. A comprehensive clini cal screen, including a detailed his tory and a comprehensive bio chemical profile, was performed on each subject. The main focus was to deter mine what caused elevated levels of GGT in these individuals. The experimental finding was that about 10% of these people had abnormally high values of GGT activity. The key finding of the study was that 72% of the subjects with abnormal GGT results were heavy, regular consumers of alco hol; 9% were chronic consumers of drugs such as anticonvulsants, antiphlogistics, and mixed analge sics containing phenobarbital. Or ganic liver disease was present in 2.7% of the subjects with abnormal GGT results; 1.8% had a diagnosis of chronic hepatitis and 0.9% had biliary cirrhosis. There were also other laboratory abnormalities: 41% of the individuals with an elevated GGT also had elevated ALT, 40% had elevated HDL cho lesterol, and 34% had elevated AST. Interestingly, only 6% had detectable levels of alcohol and 1% had abnormal bilirubin levels. Since this was the only biochemi cal abnormality in about half of the people with elevated GGT, the bot tom line of this extensive study appears to be the sensitivity of GGT as a biochemical indicator of heavy drinking and possibly of early covert liver disease. Numer ous other publications exist that tend to support the evidence provided by the Malmo group. It would go beyond the scope of this article to report on them in great detail. Instead, let us briefly dis cuss the conclusions drawn in three epidemiological studies. One was published in Clinical Chemistry by a French group.24 These authors tested GGT, iron, mean cell volume (MCV), mean cell hemoglobin concentration (MCHC), urea, uric acid, and AIT in an unselected population of 3,930 men. The self-reported alco hol consumption of these people was compared with their levels of the biological parameters by princi pal components and multiple re gression analysis. The authors were able to show that the three variables that correlated most sig nificantly with alcohol consump tion were GGT levels, MCV, and the use of tobacco. These three parameters contribute as much in formation as the whole set of ten variables examined, according to these workers, and they came to The Author Leslie M. Shew, Ph.D. Dr. Shaw is Associate Professor of Pa thology at the University of Pennsylvania School of Medicine. After receiving his Ph.D. in Biochemistry from SUNY Up state Medical Center, Sprague, New York, he completed postgraduate training at Johns Hopkins University and at the Hospi tal of the University of Pennsylvania. Dr. Shaw also serves as Director of the Toxi cology and Special Chemistry Section, William Pepper Laboratory, in the Depart ment of Pathology at the University of Pennsylvania. the conclusion that particularly GGT levels "can be considered a screening test." However, corre lation coefficients which served as an index of correlation, with the quantity of alcohol consumed were only between 0.34 and 0.43 when the previously mentioned three variables were combined. The authors, therefore, recommended that the search for discriminative biological indices of the quantity of alcohol consumption needs to continue. Chick et al.,25 in Edinburgh, Scotland, came to similar conclu sions as the French group. They assessed the usefulness of GGT and MCV determinations in 266 company executives and 222 man ual workers of two alcohol produc tion firms. The correlations with admitted drinking were significant ly higher than the French study. A man with an MCV of over 98 and a FEBRUARY 1985 MEDICAL TIMES . 69 AP00024595 GGT level above 50 U/L had a 62% chance of admitting to drink ing over 450 gm of alcohol per week. In their conclusions, these workers stated: "In those patients in whom there is an alcoholinduced elevation of one of these tests, cessation or reduction of drinking leads to a fall in the next three to six weeks, which provides the clinician with a useful indica tion of treatment progress.28-28 Pa tients whose goal is reduction of drinking rather than abstinence find serial test results an aid to monitor ing their consumption." In another study, a group of investigators analyzed the relation ship between alcohol consumption, GGT levels, MCV, and tobacco consumption.29 In this study, 455 apparently healthy outpatients were asked to report their daily alcohol consumption. Multiple regression analysis confirmed the superiority of GGT levels over MCV as a laboratory marker of alcohol in take, as well as the advantage of using them together. On this basis, the authors proposed a "rough esti mation of alcohol consumption as a first step in mass screening of heavy drinkers." Using 80 gm of pure alcohol a day (equivalent to one liter of wine) as a cut-off point, the authors were able to identify correctly three out of four "heavy" drinkers. As in the previously cited studies, no allowance was made for the possibility of underreporting alcohol intake in the subjects with elevated GGT levels. Thus, of all laboratory tests studied, the level of GGT has been shown to be the most sensitive indicator of chronic alcohol consumption. 70 Another important factor to con sider is how to determine the effect on GGT activity of "normal" so cial drinking in healthy subjects without any indication of liver dis ease. We have pursued this ques tion in our laboratory and have conducted a study with 22 normal subjects, none of whom were regu lar consumers of large amounts of alcohol. In a one-evening drinking session, these individuals con sumed several drinks of the alco holic beverages of their choice over a period of approximately four hours. A baseline value was taken the day before the consumption. Postconsumption levels were deter mined after one, two, and four days, respectively. What we found was that there was no significant increase in GGT activity in the serum of any of these volunteers after we had subjected the data to analysis of variance. Even when compared with each individual's own baseline value, no significant increase was observable. The same was found for the other enzymes, ALT, AST, AP, and glutamate de hydrogenase. In another study conducted by Belfrage and co-workers,15 eight healthy young medical and dental students were subjected to a uni form alcohol regimen after a peri od of abstinence from alcohol bev erages of several weeks. Every subject had one 16 ounce can of beer at noon, 6 p.m., 8 p.m., and 10 p.m. every day for five weeks. That amounts to about 60 to 70 gm of alcohol per day. There were slight variations in grams of alco hol per kilogram of body weight due to the varying weight of the MEDICAL TIMES - VOL.113, NO.2 students. It was found that after one week of drinking, there was no significant change of the GGT level when compared with the predrink ing control value. The same ap plied to'the transaminases, AP, bilirubin, and bromosulfophthalein (BSP) retention. At the end of the study, after five weeks of consum ing four cans of beer a day, there was a small but statistically signifi cant increase in the GGT value in each individual. The range of in crease was between 11 and 44 U/L when compared with each person's own predrinking reference value. All of the postdrinking GGT val ues, however, were within the ref erence range of a healthy popula tion. In no case was there an eleva tion above the upper limit of normal. In another clinical study, Stat- land and colleagues studied the ef fects in nine healthy subjects con suming 75 gm of alcohol per kilo gram of body weight on each of three successive evenings.16 Again, predrinking and postdrinking val ues were determined. Maximal changes in GGT activity were found at 60 hours, varying between 0.1 and 3.5 U/L, with an average increase of 1.2 U/L. These increases were certainly statistically significant because they occurred in all subjects. How ever, they were very small, and it may be difficult to reproduce these data in the average routine labora tory. Again, all increases were within the normal range. What is the "hard core" of infor mation that ties all these studies together? What conclusions can we make? We conclude from these AP00024596 studies that normal social drinicing is not likely to produce an abnor mal GGT level in an otherwise healthy person. Only from an ex tended period of chronic drinking (e.g., at least five weeks) is one likely to see abnormal GGT val ues. We must realize the danger that is always involved if one ex trapolates from studies on a few isolated populations to the whole universe. But keeping this reserva tion in mind, it is reasonable to draw a conclusion similar to that of the author of one article80 on this subject: "It has been repeated ly demonstrated that in the chronic alcoholic, serum GGT levels fall upon withdrawal, and rise with reexposure to alcohol.31-33 On the basis of these observations, the GGT test is now being used by alcohol treatment centers in docu menting the success of therapy and in identifying patients who relapse following therapy. Clearly, the GGT test--by providing an ad vance warning of alcohol related liver damage (alteration), and by enabling newly discovered alcohol ic patients to admit their problem, seek help, and thereby arrest their disease--could turn out to be among the most important contri butions to the practice of preven tive medicine." Let us now turn to the question of whether the measurement of serum GGT contributes informa tion to other diagnostic aspects of liver disease. One aspect is the high sensitivity of GGT as a gener al liver screening test, functioning as an indicator of intoxication in people whose livers are otherwise normal. One case that happened in our laboratory points this out quite clearly. A man in his middle thir ties presented to the emergency room of our hospital. He indicated that he was forced to swallow an organic liquid and did not know what it was. Examination of the gastric contents revealed that this man had consumed a kerosene-like petroleum fraction. The functions of the kidneys, pancreas, and liver were monitored. Interestingly, in the first 24 hours following the incident, there was a sevenfold rise in amylase, indicating a pancreati tis-like reaction. After one day, amylase rapidly returned to nor mal, followed two or three days later by GGT elevation of ten times the upper limit of normal. There was no clinical indication of jaun dice, and bilirubin and the other liver enzymes were only inildly elevated. Total bilirubin was 1.5 mg/dl, and the transaminases were less than one and a half times the upper limit of normal. During days four and five following the epi sode, the GGT level rapidly re turned to normal, indicating that after some intermittent involve ment, the liver had fully recovered. Another possible application of GGT levels is in the differentiation of neonatal hepatitis from biliary atresia where levels are extremely high. In a recently published study, the authors followed seven children who had presented with the suspi cion of biliary atresia.34 A series of laboratory tests were performed, including a full liver panel and ai-antitrypsin in order to rule out di-antitrypsin deficiency. The av erage value of GGT was about 760 U/L compared with a value of 183 in a group of children with neona tal hepatitis. Using 300 U/L as a cutoff point below which the diag nosis would be neonatal hepatitis and abi^ve which it would be bili ary atresia, 16 out of the 17 chil dren were correctly classified. Clearly, this is a preliminary find ing, yet it may turn out to make a contribution to the differential di agnosis of these two childhood liver disorders. This is a very im portant differential diagnosis, since treatment of biliary atresia may include surgery, whereas surgery is contraindicated in children with neonatal hepatitis. Interestingly, in the two cases with arantitrypsin deficiency, GGT activities were 869 and 2,700, both above the average value for children with bil iary atresia. In these cases, the differentiation was made on the basis of the levels of otrantitrypsin (low in children with aj-antitrypsin deficiency, normal in children with biliary atresia). It appears that the lack of the trypsin inhibitor has caused increased proteolytic cleav age of GGT from the external sur face of cell membranes and an increase in activity in serum. A study on the diagnostic utility of GGT in various forms of liver disease and the behavior of GGT in comparison with other liver param eters was published by Lum and Gambino in 1972.35 The paper by E. and F.W. Schmidt is another study of clinical utility of serum GGT measurements.36 If we extract the key information from these two publications, we can summarize the diagnostic contributions of serum GGT. GGT is a highly sensitive indica- FEBRUARY 1985 MEDICAL TIMES . 71 AP00024597 tor of liver disease, regardless of whether the disorder involves in flammation of the liver, a spaceoccupying lesion, or obstruction of the biliary tract. It is usually most significantly elevated by obstructive liver dis ease. It is superior to alkaline phos phatase and leucine aminopepti- dase, both in terms of diagnostic sensitivity and specificity. In the absence of other laborato ry evidence, GGT elevation is a sensitive screening test for occult alcoholism. GGT has good specificity for the liver compared with AP and AST, since it is not elevated in bone disorders or during pregnancy, as is AP, nor is GGT elevated in skeletal muscle disease, as is AST. GGT elevation is a sensitive and inexpensive assay which is useful as an indirect indicator of the tissue source of AP. The test also helps to differenti ate mechanical and viral chole stasis from drug-induced chole stasis. In the first group, GGT and AP are elevated approximately to an equal degree. However, in drug-induced cholestasis, GGT and AP are disproportionately ele vated, with the former being much more significantly elevated above normal than the latter. GGT is also useful in anicteric patients with cancer in whom ele vated GGT levels indicate liver metastases. 4 Reference* 1. Schmidt E, Schmidt FW: Enzym-diagnostik von Lebereckxankungen in der Praxis. Diagnostic 10:348351, 1977. 2. Schmidt & Strategy and evaluation of enzyme determination* in serum in diseases of the liver and the biliary system. Im Evaluation of Liver Function: A Multifaceted Approach to Clinical Diagnosis, Demers LM, Shaw LM (Eds), pp 79-101. Baltimore, Urban and Scbwarzenberg, 1978. 3. Diagnostic Market, I.M.S. America, Ltd., 1979. 4. Marathe GV. Damle NS, et al: Localization of `Y-glutamyltranspeptidase in lymphoid cells. FEBS Lett 115:273-277, 1980. 5. Tanaka M: A histochemical study on the activity of y-glutamyltruspeptidase in liver disease. Acta Path Jap 24:651-655, 1974. 6. Weitholz H, Colombo JP: Das verbalten der y-glutamyltranspeptidase und anderer leberenzyme in plasma wahrena der alkohol-entiziehungskur. Schweiz Med Wocbenschr 106:981, 1976. 7. Huaeby NE: The activity of y-glutamyltransfense after bile duet ligation in guinea pig. Clin Chim Acta 88:385, 1978. 8. Tate, SS, Ross ME: Human kidney y-glutarayl* transpeptidase. Catalytic properties, subunit structure, and localization of the y-glutamyl binding site on the light subunit. J Biol Chem 252:M2, 1977. 9. Hughey RP, Curthoys NP: Comparison of the size and physical properties of `y-glutamyltransfense purified from rat kidney following solubilization with paj)rin or with Triton X-100. J Biol Chem 251:7863, 10. Melster A., Taut SS, Hess LL: Metnbnne-boviTid -y-giutamyltnukspeptidase. Is: Enzymes of Biological Membranes, Martonosi AN (Ed), vol 3, pp 315-347. New York, Plenum, 1976. 11. Elce JS, Broxmeyer B: y-ghitamyltransferese of rat kidney: Simultaneous assay of the hydrolysb and transfer reactions with [glutamate-t4Cj glutathione. Biochem J 153:223-232, 1976. 12. Shaw LM, Newman DA: Hydrolysis of glutathi one by human liver y-glutamyitransferase. Clin Chem 25:75-79, 1979. 13. Chasseaud LF: Conjugation with glutathione and mercapturic acid excretion. Is: Glutathione: Metab olism and Function, Arias IM, Jakoby WB (Eds), pp 77-114. New York, Raven Press, 1976. 14. IFCC Methods for Measurement of Enzymes, Part 4. IFCC Method for y-glutamyltransferase [(ygulamyl) - |>e^tide: Amino acid `y-glutamyltransferase, 15. Belfrage P, Berg B,etal: Prolonged administra tion of ethanol to young healthy volunteers: Effects on biochemical, morphological, and neurophysiological parameters. Acta Med Scand 552 (Suppl): 1-44. 1973. 16. Freer EKE, Statland BE: The effects of ethanol (0.75 gm/kg body weight) on the activities of selected enzymes in sera of healthy young adults. 1. Intermedi ate-term effects. Clin Chem 23:830-434. 1977. 17. freer DE, Statland BE: Effects of ethanol (0.75 gm/kg body weight) on the activities of selected en zymei in sera of healthy young adults. 2. Interindividual variations in response of *ygluumyltnnsferase to repeated ethanol challenges. Clin Chem 23:20992102. 1977. 18. Shaw LM, Marsh EM: The effect of one evening of social drinking on serum y-glutimyltransferase ac tivity. Clin Chem 27 (Abstr): 1036, 1981. 19. Heipera R, Eickhoff K, Poser W: Anticonvul sant therapy and serum y-glutimyltransferase. Klin Wocbenschr 56:9211-928, 1978. 20. Knight JA, Haymond RE: y-glutamyltransferase and alkaline phosphatase activities compared In serum of normal children and children with liver disease. Clin Chem 27:48-51, 1981. 21. McComb RB, Bowen BN, Posen S: Alkaline Phosphatase, pp 531-536. New York, Plenum, 1979. 22. Rosalki SB: Gamma-glutamyltranspeptidase. Adv Clin Chem 17:53. 1975. 23. Kristenson H, Trell E, et al: Serum y-glutamyltransferase: Statistical distribution in a middle-aged male population and evaluation of alcohol habits in individuals with elevated levels. Prev Med 9:108-119, 1980. 24. Bagrel A, d'Houtand A, et ah Relations be tween reported alcohol consumption and certain bio logical variables in an unselected population. Clin Chem 25:1242-1246, 1979. 25. Chick J, Krietman N, Plant M: Mean cell volume and -y-gluiamykranspeptidasc as markers of drinking in working men. lancet 6:1249-1251. 1981. 26. Lamy J, Baglin MC. et al: Emploi de la mesure de la gamma-glutamyltriaspcxxklase aerique pour coo- txoler le suoces des cures ot desintoxication aotialcoolique. CUn Chim Acta 60:103-107,1975. 27. Buffet C L'augmentatkm de volume globulairc moyen (V.G.M.) test de depisage et de surveillance de I'alcoolisme ebrottique. Rev Ale 22:15-32. 1976. 28. Wadstein I, Skude G: Changes in amylase, hepatic enzymes, and bilirubin in serum upon initiation of alcohol abstinence. Acta Med Scand 205:313-316, 1979. 29. Papoz L, Warnet JM, et al: Alcohol consump tion in a healthy population. Relationship to y-glutam- yltransferase activity and mean corpuscular volume. JAMA 245:1748-51, 1981. 30. Bates HM: A sober look at GC3TP and alcohol ism. Lab Management 19:17-19, 1981. 31. Weitboltz H, Colombo JP: Das verhalten der gamma-glutamyltranspeptidase and anderer leberen- zyme, in plasma wahrend der alcohol entziehungskur. Schweiz Med Wocbenschr 106981-988, 1976. 32. Shaw S, WarnerTM,etal: Detection of alcohol ism relapse: Comparative diagnostic value of MCV, GGTP, and AANB. Alcoholism Clin Exp Res 3:297- 301, 1979. 33. Lamy J, Baglin MC, et al: Gamma- glutamyltranspeptidase serique et alcoolisme. Diagnos tic et controle du sevrage. Nouv Presse Med 4:447- 453, 1975. 34. Wright K, Christie DL: Use of*y-glutamyltrans- peptidase in the diagnosis of biliary atresia. Am J DLs Child 135:134-136, 1981. 35. l-mn G, Gambino SR: Serum y-glutamyltranspeptidase activity as an indicator of disease of liver, pancreas, or bone. Clin Chem 18:358-362, 1972. 36. Schmidt E, Schmidt FW, -y-glutarnyhianjpepti dase. Dent Med Wocbenschr 98:1572^78, 1973. AMERICAN LUNG ASSOCIATION The Christmas Seal People 72 MEDICAL TIMES VOL. 113, NO.2 AP00024598