Document N2MZ7kpKY1D3YpMRXMDRmKqmp

FILE NAME: Oil Industry and American Petroleum Institute (API) DATE: 1955 Oct 1 DOC#: API128 DOCUMENT DESCRIPTION: Interim Report - API Research Project MC-1 - The Kettering Laboratory INTERIM REPORT 'I API RESEARCH PROJECT MC-1 ,!! ll` ij OCTOBER 1, 1955 iii III THE KETTERING LABORATORY in the Department of Preventive Medicine and Industrial Health College of Medicine U N IVERSITY O F C IN C IN N A TI, C IN C IN N A TI, O H IO API 06460 ERRATA for Interim Report API Research Project MC-1 October 1, 1955 Page Number 1 1 3 3 6 7 10 10 10 10 11 11 12 Line Number 1+ 5 3 20 19 18 1 15 16 17 21 2k 2 Suggested Change Insert wscrotal and other" between "of" and "cutaneous" Delete "especially scrotal" Replace "satisfactory" with "one-to-one" Add "in benzene." after "carcinogen" Replace "II" with "VII" Insert "(the contribution to potency due to accelerators)" between "Ma" and "may" Delete after "far" Replace "after each application failed to prevent the" with "accomplished a" Delete "development of tumors. A" Delete "could be accomplished" Insert "(See Table X for experimental details)" after "applied" Replace "III" with "IX" Replace "indicated" with "suggested" API 06461 For Information Only - Not for Publication INTERIM REPORT API RESEARCH PROJECT MC-1 OCTOBER 1, 195 The Kettering Laboratory . in the Department of Preventive Medicine and Industrial Health College of Medicine University of Cincinnati Cincinnati, Ohio API 06462 API RESEARCH PROJECT MC-1 October 1, 1955 The potential hazards deriving from carcinogenic materials in the intermediates and products of the refining of petroleum are being subjected to systematic investigation. The initiation of this work by the sponsors was stimulated in part by historical and recurrent instances of cutaneous, especially scrotal, cancer among wax pressmen. The immediate impetus came, however, with the recognition of the highly increased aromaticity of certain high boiling fractions produced in the newer catalytic processes. Further, pointed evidence from several centers of investigation showed that certain of these fractions possessed the capacity for the induction of cancers in the skin of experimental animals at a considerably higher level than that of the uncracked distillates associated with the problem among the pressmen. Attention was directed in 191+9 to products of the following processes, in which polycyclic aromatics might be synthesized and/or concentrated: STRAIGHT THERMAL CRACKING PROCESSES Non-catalytic cracking Mon-catalytic cracking Viscosity breaking Naphtha reforming Coking Polyforming Steam cracking of- virgin gas oil of catalytic gas oil CATALYTIC CRACKING PROCESSES Houdry Thermofor catalytic cracking Fluid catalytic cracking Cycloversion Catalytic reforming SPECIFIC PRODUCTS Slack waxes and press oils Raw lube stocks from straight run distillation Solvent extracts Acid oils from acid sludge API 06463 With the full cooperation of the research and refining departments of the companies represented on the Medical Advisory Committee, detailed tables of information were assembled in the Laboratory, showing the vari ations in operating conditions and feed stocks of some 200 different units involved in the above processes. From this comparison of available materials it was possible to select a series of about 600 samples to cover the extremes of the ranges of each important variable. Each company then obtained certain analytical data on a selected group of its samples and made this additional information available to responsible personnel in the Laboratory. One item among these data was the absorption spectrum of a fraction representative of the heart-cut of the polycyclic hydrocarbons regarded as potentially carcinogenic. Estimates of the relative potency of each sample were based upon certain characteristics of these spectra. Although this empirical technique proved subsequently to be applicable mainly to residual fractions from catalytic cracking, the approach served a useful purpose. It pointed up one salient characteristic of the group of materials which, when tested biologically, were found to have potencies much higher than those predicted. Almost invariably, these samples^contained high percentages of components boiling below 700F. This observation, coupled with the results of parallel experiments on solutions of 3-methylcholanthrene in certain non-volatile solvents, led to the recognition of the contribution made to the potency of many petroleum products by a group of non-carcinogenic hydro carbons, with boiling points in the range, 5 0 Q - 7 $ 0 F . These materials, termed "accelerators" for reasons of convenience, will be referred to in some detail later. From the series of samples which had been collected and analyzed in this manner, approximately 120 were subjected to tests of their relative carcinogenic potency when applied upon the skin of C3H mice. This inbred ,tr ,n has proved particularly useful, in that the results of tests made , differentiate non-carcinogenic from carcinogenic materials by means of ication of these materials upon the skin, have shown a satisfactory PP1 .rrelation thus far with the results of prolonged contact of human skin th o se instances (admittedly limited in number) In which occupational ,,0sure has been documented properly. Since the 120 samples represented *ide range of potencies for the skin, certain modifications of the time- i-oDored experimental procedure (for the application of such materials on s k i n of animals) were developed in the course of the research. As ^scribed in an article which has been circulated and will appear soon in .a3Cer Research, these improvements have made this experimental method more flexible and more readily subject to quantitative interpretation. The relative potencies of these samples are shown in Table I. jbe scale on which they are expressed is based upon a series of parallel tests on solutions of 3-ethylcholanthrene in. benzene. The conventionalized expression (P^q ) for the capacity of a material to induce cancer in the skin of the mouse, simply indicates that the oil in question induces tumors h the skin of mice at the same rate, under"1'comparable experimental conditions, is a solution of that level of concentration, in percent by weight, of the . '7r.thetic carcinogen. It may be pertinent at this point to glance at the relative potencies :f certain products of another industry which has experienced an increased bcidence of skin cancer among specific groups of its employees. Crude *Ml tars from by-product coke ovens have shown potencies in the range, *08 to O.I4., there being some evidence that the majority of these would be luded in the 0 . 1 0 to 0 . 1 5 part of this range. Thus, it will be noted t some uncracked distillates from petroleum (e.g., A P I-79 and 105) have roxiamtely the same potency for mice as these coal tars. The average API 06465 - 1+ - atlon of the exposure of workmen who developed cancer of the skin in s3Ciation with contact with tar was comparable to that of wax pressmen, as _,ut 15 to 20 years, a*" . It is apparent from the data in Table I that thermal cracking of .fttalytic gas oils, as it is usually practiced, does not reduce their carcinoz0nic potency. Indeed, when one notes the relative increase in the content benzo [a)pyrene (BP) which often results from this process, it might be 9Xpected that potency would increase significantly, were it not for t h e fact .v,at t h e probable accompanying destruction of long-chain accelerators 3Ce r a t e s in the opposite direction. Such correlations between the relative potencies of different oils jfld their chemical composition have been among the major aims of this inves tigation. It was recognized from the outset that the needs of such a dynamic industry would not be met by determining and listing the potencies of samples s e le c te d more or less arbitrarily at any particular time. The changes in refining methods which have accompanied the competitive technological developments of the past decade have had a profound effect upon the composition :f many fuels and other petroleum products. --Since .continued biological testing is time-consuming and costly, the need for chemical-analytical methods .'or the estimation of carcinogenic potency of complex mixtures has been evident. To these ends an intensive program of fractionation and analysis rf selected carcinogenic oils was initiated in 19^4-9 13a its early stages his work yielded suggestive evidence of a relationship between the carcinolQic potency of cracked residua for mice and the intensity of a band in toe absorption spectra of certain distillates prepared from these oils in toe laboratory. The most intensive w.ork has, been carried out with a sample 8lurry from Fluid Catalytic Cracking, API-8. Spectrographic evidence of API 06466 " 0- rr-:senc6 in this oil of the well known carcinogenic hydrocarbon, oenzo V)- ^ ne : led eventually to 0 '' ' ,;c sample of the pure u ** e *^i particular family its isolation and its identification compound. Further evidence that the of carcinogens in many residual oils with a synconcentration was of such a ,aV0X as to contribute to their carcinogenic potency to a major extent .,,^nished the stimulus for the development of a generally applicable method analysis for this family of compounds, The method has oeen published .-n Analytical Chemistry, 27, 2I4.8 (1955). The content of benzo[a] pyrene has been determined in each of the which have been tested by application on the skin of mice. The results ire shown on Table I. With certain striking exceptions, it will be noted that ->j5re is a rough correlation between the content of 5? and the relative -etency of the residual products from any given process. One of the exceptions 3 sample API-71 from Thermofor Catalytic Cracking, A detailed examination this oil, and of certain of the distillates (sidestreams) which have been nested, has oeen made in order to learn more about other classes of corn- rounds which contrioute significantly to the carcinogenic potency of these oils In parallel with the program of frae'tionation of samples obtained from refining operations, solutions of synthetic carcinogens in various solvents and in mixtures of known composition have been investigated, knerally, it has been found that the replacement of benzene by progressively Us3 volatile solvents has resulted in progressive retardation of the rate at which a given concentration of carcinogen would induce tumors. Comparison :? the relative areas of skin wetted by given quantities of the solution h benzene with those involved in contact with the less volatile solvents aas suggested a very plausible explanation for the slower induction of tumors by the latter. It could be shown that, so long as the dosage in ^iligrams of the carcinogenic solute per unit area of skin was kept constant, API 06467 6 i - - ,>e rate of Induction of tumors was unchanged, f'or example, since solutions methylcholanthrene (MC) in amylbenzene spread about twice as far over the 3icin as solutions In benzene, It Is necessary to use almost 0.2 percent MC .0 the alkylbenzene to obtain the same dose per unit area and rate of carcino t,enesis as were obtained with 0.1 oercent MC in benzene. Extension of this work to Include solutions of MC in the detergent intermediate, dodecylbenzene, yielded results which were surprising at first, v,ut subsequently rewarding in that they provided the explanation for the ;0tency of oils such as API-71 Solutions of MC in this 18-carbon solvent 3pread more widely upon the skin then did those in amylbenzene, and yet induced tumors much more rapidly than did the corresponding solutions in benzene. The presence of fractions with comparable "accelerating,,properties in a number of petroleum oils was demonstrated subsequently. It soon became parent that these materials are derived from the crude petroleum, that they may increase in concentration to some extent under the conditions of catalytic cracking, and that they may eventually be reduced significantly by very drastic cracking. Further investigation of solutions of BP in various pure solvents j (Table II) has shown that this property of acceleration is shared by at least some members of almost all classes of hydrocarbons containing 10 to 20 carbon atoms per molecule. The shape of the molecule, as well as its overall length, seems to be an important characteristic, since 7-cyclohexyltridecane (Experiment API-279) proved to be rather weak as an accelerator. The finding that normal paraffins, such as cetane, may contribute signif icantly to the carcinogenic potency of a mixture, although they themselves fc'e non-carcinogenic, has thrown some light, it seems, upon the apparent J^radox involved in the occurrence of scrotal cancer, under certain eir- 'Jnstances, among men engaged in wax-pressing operations, despite the low ^ooaticity of the distillates which enter into these operations. API 06468 - 7 - The current systematic study of the properties of distillates otn a variety of cracked and straight run oils represents an effort to ^gntitste the knowledge which has been accumulated on accelerators. The ,0al *3 an analytical method by which the factor introduced into the potency {(,uation by these fractions in any given oil may be determined. The knowledge accumulated thus far from numerous experiments on jolutions of certain polycyclic hydrocarbons in solvents of well defined imposition suggests that a realistic technique for estimating the potency 0f any given oil on the basis of its physical and chemical characteristics i8y involve a summation of four separate analyses. In algebraic terms MC s (Cx + Ca + Ca ) jbe summation of Cx, Ca, and Ca represents the effective potency contributed by three major categories of polycyclic carcinogens, expressed in methylcholanthrene equivalents. Included in these three classes are (1) the benzo pyrenes previously discussed, (2) carcinogenic substances extractable by maleic anhydride (probably derivatives of benzanthracene), and (3) carcino genic substances insoluble in concentrated sulfuric- acid (sometimes referred to as low-boiling carcinogenic compounds). The factor Ma may vary, apparently, from 1 to 8, and thus in certain instances, such as catalytic cycle gas oils and light lube distillates, may provide the major contribution to potency. Hence, the quantitation of the effects of accelerators becomes a matter of consideraole practical importance. The available evidence indicates that a significant fraction of the carcinogens may be extracted from catalytically cracked oils by maleic Qhydride. In the case of API-8, approximately 40 percent of the carcinogenic aterial (in terms of methylcholanthrene equivalents) may be separated from the oil by this technique. Experiments 8-2a and 8-2b have shown that there API 06469 - 8 - no significant difference in potency between the fractions obtained by cCessive extractions of API-8 with maleic anhydride, even though the mass jpectra of the different fractions indicate that there are gross differences jD their chemical composition. Similarly, when such an extraction is applied to a cracked sidejtream, such as API-89, a small amount of carcinogenic material may be gxtracted. Because of the uncertainty about the multiplication factor due {0 accelerators in API-89, it cannot be said yet what portion of the carcinogen tere removed by maleic anhydride. However, separation of the extract from pi-89 hy chromatography yielded valuable information to the effect that the large proportion of the fraction which was desorbed more easily than benz- inthracenes, approximately 90 percent of the total Diels-Alder adduct, contained practically none of the carcinogenic materials. Further, to resolve some of these questions a series of chromatographic fractionations are being applied to a series of straight-run and cracked.distillates, similar in 1 boiling range to API-89 (600 to 73>0F.) to produce accelerator-free concen trates of carcinogens, which may then be tested directly for their relative potency. This should then provide values for"~the summation (Cj + C a + Ca) In the potency equation. A parallel series of extractions of the oils by aleic anhydride will furnish quantitative information regarding the relative contribution of this class of carcinogens to the effective sum. Some evidence has been collected which indicates that these compounds extracted by maleic anhydride contribute relatively less to the potency of sidestreams than to that of residual oils. From the practical standpoint of the occupational hazard of skin tocer, perhaps the most important oils are the distillates obtained in the p*nge 550 to 800F. Such materials as wax distillates, diesel fuels, and t*se stocks for cutting oils are found within this range. It Is probable ^at the first two major classes of carcinogenic substances, ^ Ct and Ca yj the potency equation, may contribute relatively little to these distillates ?0th of these classes would be extracted in fairly large part by concentrated 3iilfuric acid. It has been noted previously that the raffinate, API-8-20, from such an extraction of the lower boiling carcinogenic distillates of pi-8 contains most of the carcinogens of these fractions. A large batch of polycyclic aromatics concentrated from API-8 by chromatography has been subjected to careful distillation through a column having a theoretical efficiency of 80 plates. Selected fractions are being tested under the code number 8 - 2 6 . The fractions which were distilled below 720F. showed at most a very low order of carcinogenic potency. The behavior of the fractions between 75 and 710F. suggested the presence of a trace of carcinogen in an active accelerating medium. This was confirmed by experiment 8-26-2, which demon strated that the accelerators in these fractions had a multiplication factor, with respect to potency, of about two. The alkylanthracenes removed during the same chromatography were tested for potential acceleration in experiment 8-26-3. Their lack of significant activity~in this respect is compatible with the general idea that more alkylation of such a nucleus would be re quired then would be found in a cut of this boiling range. The major components of the carcinogenic fractions boiling between 720 and 76>0F. are derivatives of pyrene, benzfluorene, and fluoranthene, together with a substantial amount of non-hydrocarbons, particularly der ivatives of carbazole. A separation of these classes is being carried out in order to determine their relative contribution to the potency of traight-run and cracked distillates. API 06471 1 10 This report thus far, has concerned Itself with the problem of p3 relationships between chemical composition and carcinogenic potency. ,3nsiderations of industrial hygiene involved in the general problem have ;go been investigated in a number of ways. Experimental work on the pos- .Dility of reducing the hazard by washing the materials from the skin, I by the application of so-called barrier creams, has been carried out. jenerally speaking, it has been found that the development of tumors could -e retarded significantly by washing the skin with soap and water at an Iappropriate interval after the application of a carcinogenic material, Ijo long as the frequency of application did not exceed once per week. For example, in an experiment in which the oil API-Q was removed by washing within 10 minutes after its application, no tumors developed during the entire |life span of the mice. However, in most of the experiments involving three applications of a material per week, the washing of the skin with an appar ently effective detergent after each application failed to prevent the ^development of tumors. A significant retardation of the rate at which the tumors developed could be accomplished, but, as a general rule, half or . aore of the animals developed neoplastic growths. The use of barrier creams prior to the application of a carcinogenic oil had no significant effect upon the outcome of the experiment. The use of various hydrocarbon solvents to rinse off the excess of oil prior to washing the skin with a detergent has been referred to in some detail in a previous report and in verbal discussions. Briefly to summarize these observations, these procedures in some instances, have augmented the carcinogenic effect instead of contributing to the efficiency of the cleansing the skin by the detergent and water. Solvents containing accelerators have proved to be particularly disadvantageous in this respect. In short, it appears that such rinsing with solvents does not improve the efficiency  - 11 - f the washing, and may accelerate the induction of experimental cancers. As ' measure of occupational hygiene, such a procedure does not appear to be ^girable. In the course of the general experimental program, a number of 0thah questions raised withing the Laboratory and by representatives of the |jjedical Advisory Committee have been answered. For example, numerous experiJ0nts have shown that if the applications of a carcinogenic oil from catalytic .yacking be discontinued at the time the first evidence of neoplastic growth 3f the skin of the mouse appears, the likelihood of the further development of the lesion into a malignant tumor is the same as it would be if the applications had been continued. On the other hand, if the applications upon the skin of all of the numbers of a group of mice terminated at the time of the earliest appearance of a papilloma on the skin of the first member of its group, widely variant results are obtained in different experi ments. In the case of a test with API-71 the majority of the animals subjected to no further applications went on to develop tumors at almost the same rate as that which characterized a parallel group which was the i,subject of uninterrupted applications. On the other hand, in an experiment Iwith A P I-U 3 -1 , very few additional animals developed tumors after the interruption of the applications until at a much later time, after a second 4period during which the carcinogenic oil was applied. The action of accelerators upon skin previously subjected to contact with carcinogenic materials, has also been investigated. A number of experi ments referred to in Table III have demonstrated that accelerators such as dodecylbenzene and dodecane are capable of eliciting tumors in skin previously subjected to single applications of catalytically cracked oil. Other experiments have shown that contact with the accelerating *lkylbenzenes may so alter the skin as to augment the effects of its subsequen contact with carcinogenic oils. The nature of this alteration is speculative, API 06473 - 12 5ut the phenomenon, of itself, suggests certain approaches in investigating ,0 mechanisms of acceleration. That these are multiple is indicated by the judication (Experiments 600 and 620) that dodecane seems to lack the capacity , 0 condition the skin in this manner. on the Kettering Laboratory, College of Medicine, niversity of Cincinnati, Cincinnati, Ohio Investigative Team: Bernard H. Braun, Ph. D. Frank P. Cleveland, M. D. Ralph T. Denham, B. S. Dorothy Templeton Denman, B. S. Frank R. Dutra, M. D. Mary Jane Graf, B. S. Francis F. Keyroth, ph. D., M. D. A. WeBley Horton, Ph. D. John J. Phair, M. D. Fred Shaffer, M. S. Ruth Pierle Trossett, Ph. D. Russell Tye, M. S. Waldo J. Younker, M. S. Helen Plagge Wittrock, B. S. Otto Bufe Patricia Clapsaddle Lenore Hull Irvin Rapien Effie West . Report: A. Wesley Horton, Ph. D. Date; October 1. 1955