Document 2RVe2e7On1M6gpyy6JQVYdMbL

FILE NAME: Oil Industry and American Petroleum Institute (API) DATE: 1954Jan-June DOC#: API032 DO CUM ENT DESCRIPTION: Letters, Memos, Reports, Financial Reports & Other Relevant Documents from Jan-June, 1954 i- t ; ESSO LABORATORIES Y STANDARD OIL DEVELOPMENT COMPANY P. 0. BCK 51 LINDEN, N. J. January 26, 195b The Members of the Medical Advisory Committee I am attaching a copy of a manuscript of a paper which Dr. A. Wesley Horton wishes to submit to the Journal of Analytical Chemistry in the near future. This paper was presented to the BPA Committee at its meeting in Cincinnati on January 21, 195b. Certain minor modifications were made at that time, and these modifications have been Incorporated into the attached manuscript, the revised copy having been received in my office on January 25, 195b. It was the unanimous agree ment of those of the BPA Committee present in Cincinnati (Dr. L. C. Beard and Dr. C. H. Hlne were absent) that this manuscript was acceptable for publication. I would greatly appreciate it if each of you would review this manuscript in detail and transmit any criticisms or suggestions which you may have concerning it to Mr. Stroop within the next 20 days*. At the end of that time, if no major objections to the publication of this manuscript have been received from any of you, it is the plan to transmit it to the Health Committee of the Board for their approval. As you know, the total tin for clearance of this manuscript is 60 days, at the end of which time Kettering Laboratory can submit this paper for publication, so that anything you can do to expedite comments would be greatly appreciated. Very truly yours, /s/ R. E. Eckardt REE: ilk Attachment R. E. ECKAHDT, M. D. * Note by D. V. Stroop As we have been delayed in transmitting this letter until February 3, vc shall defer reference to Medical and Health Committee of the Board until March 1. Three copies of manuscript are enclosed for each committee mesfeer. None is being sent to any technical advisor. DVS API 05743 AN ANALYTICAL METHOD FOR DETERMINATION OF BENZO j>] PYRENE IN COMPLEX MIXTURES1 Russell Tye, Mary Jane Graf, and A. Wesley Horten Kettering Laboratory, Department of Preventive Medicine and Industrial gealth,College of Medicine, University of Cincinnati, Cincinnati, Ohio PREFATORY SUMMARY This paper describes the isolation and identification of t^e polycyclic hydrocarbon, benzo^ajpyrene, in an oil produced by catalytic cracking of petroleum, together with the analytical method eveloped for the estimation of the concentration of this compound ^ products of refining operations. A selected fraction of a given aample is obtained by the.'use of a standardized chromatography. Two equal portions of the fraction are taken, and one of them is subjected to a catalytic iodination on a column of activated alumina. Spectrophotometric measurement of the difference in absorbance between the iodinated portion and its mate is then used to determine the concen tration of benzo[ajpyrene present. INTRODUCTION A prolonged search for the carcinogenic element in coal tar culminated in the isolation of a pure hydrocarbon carcinogen from coal tar pitch by Hieger (1). Cook and Hewett (2) synthesized benzojjajpyrene and proved its identity with the hydrocarbon from coal tar. In the course of an investigation of the possible car cinogenic properties of high boiling products from petroleum re fining operations, it was observed that a distillate fraction of a catalytically cracked residuum had an ultraviolet absorption i This work was carried out largely as part of the American Petroleum Institute Research Project MC-1. API 05744 - 2 - pectrum which strongly suggested the presence of benzo^aj pyrene in jignii'icant concentration. This hydrocarbon was isolated by the follwin Proceciure ln a sufficient state of purity for positive identification. Fourteen hundred grams of the oil were subjected to a simple vacuum distillation. A fraction, boiling 205 to 275C. at 0.5 mm. pressure, and weighing 175 g. was chromatographed on alumina, and benzojjf]pyrene-enriched fractions selected spectrophotometrically. The chromato graphic procedure was repeated five times, resulting in 6.5 g. of a red semi-solid, the ultraviolet absorption spectrum of which indicated a content of benzo[Vjpyrene between 0,8 and 2.5 g. Fractional adduction by iodine (3) was used to remove perylene and other unidentified com pounds, reducing the weight of the concentrate to 5 g. This was dissolved in benzene and extracted by cold con II centrated H aS0*. The acid was diluted with ice, and the resulting precipitate was dissolved in benzene. An ad ditional series of chromatographies and fractional crystallizations produced 0,116 g. of crude benzo^ajpyrene, having substantially the same ultraviolet absorption as that of an authentic sample. Further recrystallization produced 12 mg. of crystals, m.p. 173-17l|C, (uncor.). The melting point of a mixture with synthetic benzo [ajpyrene was 174.5-175.5C.(uncor.), as compared with that of tne pure material, 1 7 7 .5-178.0C. API 05745 - 3- Ihe method of Goulden and Tipler (i+), as adapted by jailer (5) for the determination of benzo^ajpyrene in soot and city dusts, utilizes chromatography and fluorescence spectroscopy, the final estimation of concentration being made by visual compar ison of the spectrum of the unknown with that of a known sample com posited of benzole]pyrene and suitable impurities. Falk (6) and Wedgwood (7) have used chromatography and ultraviolet spectro photometry to separate and identify polynuclear aromatic hydro carbons from complex mixtures. The method described in this paper attempts to utilize the chromatographic technique not only to concentrate the benzoja]- pyrene but to aid in identifying it by its rate of movement along a carefully standardized column. Secondly, a selection is made by catalytic iodination, and, finally, identification and measurement are accomplished on the basis of the change in the ultraviolet spectrum caused by the iodination. The method is relatively fast, requiring about three hours per analysis, or less if multiple analyses are made simultaneously. CATALYTIC IODINATION Activated alumina catalyses the substitution of iodine for hydrogen of the benzo[a] pyrene molecule, A convenient method of accomplishing the reaction is to pass a solution of benzo[_a]- pyrene and iodine in suitable concentration In benzene through a short column of activated alumina in the manner of elution chroma tography. Under the conditions of the procedure used in this analysis, 6-iodobenzo[ajpyrene is produced in about yield. API 05746 The iodo- compound has been purified by chromatography m and recrystallization. Its melting point was 211-212C. (uncor.); . \ its iodine content (calculated as 33*6# for ^ O ^ l l ^ was 33.1$. The ultraviolet absorption spectrum is shorn in Pig. I; it is noteworthy that the compound does not have any visibly perceptible fluorescence. The position of substitution was established by con version of the iodo derivative to the nitrile by heating with a small excess of Cua (CN)a in a sealed glass tube at 306C. for one and one-half hours. A yield of 97^ of crude product was obtained. This product, when carefully purified, melted at 234.-235*0. (uncor.). 6-Cyanobenzo[aj- pyrene, prepared by the method of Windaus and Raichle (8), melted at 231-232*0. (uncor.), and in a mixture with the first cyano compound, melted at 233-234C.(uncor.). The ultraviolet spectra of the two samples were in agreement and corresponded with the spectrum obtained by Jones (9) API 05747 FIGURE I ULTRAVIOLET ABSORPTION SPECTRUM OF 6-IODOBENZO[a]PYRENE (IN ISOOCTANE) CD - 6- Other activated adsorbents also catalyze the iodlnation 0f benzo[aJpyrene. The pH of the adsorbent appears to be important, tfnder conditions similar to those of the analysis, attapulgus fuller's earth of low activity and of a lower pH than Grade F-20 jlurni*15 (as measured by pH meter on aqueous extracts) produces roughly equal yields of 6-iodo-benzo[a]pyrene and an additional probable iodobenzo[aJpyrene. Activated magnesia, having a pH much higher than Grade F-20 alumina, and activated silica gel, having & pH much lower than the attapulgus fuller's earth, promote only faint reactions. The addition of NaHC09 to the silica gel enhances its catalytic activity markedly, although NaHC09 shows no activity alone. A large increase in the quantity of iodine causes the alumina-catalyzed reaction to produce a mixture similar to that from attapulgus fuller's earth. The ability of aromatic hydrocarbons to form molecular complexes with Lewis acids has been discussed by several authors (10,11,12,13,14) and related by Mulliken (13) and Brown (14) to the ease with which these compounds may be halogenated. Brown found a simple linear relationship between the logarithms of the relative rates of halogenatlon of several methyl benzenes and the logarithms of the relative stabilities of their deeply colored onium-ion complexes with HF-BF9* The solubility of various polycyclic aromatics in cold, concentrated sulfuric acid (3) is doubtless a function of the basic properties of the hydrocarbons. These solutions have vivid colors ranging from yellow to wine-red, suggesting the formation of onium-ion complexes (13,14). Perylene, 3-methylcholanthrene, API 05749 - 7- ,,.,^fa"ipyrcne pyrene, and anthracene can be extracted from benzene th varying degrees of completeness by an equal volume of concentrfted sulfuric acid at 10C. Under the conditions of the analytical pr0Cedure, the same compounds react with iodine to similar relative eJctent3. Benz[ajanthracene can be extracted by sulfuric acid but reacts negligibly with iodine. Chrysene, benzoY]phenanthrene, dibenz[>hjanthracene and phenanthrene cannot be extracted by cold concentrated sulfuric acid, and are not iodlnated by the analytical procedure A more generalized check cn the apparent rule that polycyclic aromatic hydrocarbons not soluble in cold, concen trated sulfuric acid are not substituted by iodine tinder the con ditions of this analytical method, has been made. A typical residuum, resulting from catalytic cracking of petroleum, believed to contain a wide variety of polycyclic aromatics including alkylated and hydrogenated derivatives, was exhaustively extracted by cold con centrated sulfuric acid. Two equal portions of the non-extract were taken, one portion iodlnated, and the difference spectrum measured according to the standard analytical procedure. No apparent reaction had occurred. These data suggest that the iodine substitution occurs through formation of an acid-base reaction Intermediate, probably an onium-ion complex. API 05750 - a - APPARATUS The chromatographic column which is used to prepare a sllitable fraction for the analysis is composed essentially of a reservoir and two separate sections of adsorbent, as shown in Fig. II. The spectrophotometer employed in the development of the method was th0 Qeckman Model CU equipped with 1 cm. cells and a tungsten lamp. is assumed that any comparable instrument could be used equally well* MATERIALS ISOOCTANE, Phillips Petroleum Company, pure grade, re-distilled. BENZENE, C.P. thiophene-free, re-distilled, SOLUTION OF BENZO[6]PYRENE, Edcan Laboratories, C.P., 5*0 mg. in 100.0 ml. of benzene. DEVELOPING SOLVENT, Mixture of 200 ml. of isooctane with 200 ml. of benzene. STANDARD SOLUTION OF IODINE (reagent grade), 1 g. in 10 ml. of benzene. ACTIVATED ALUMINA, Alcoa, Grade F-20, 80-200 mesh. SOLUTION OF SODIUM THIOSULFATE, C.P., approximately 5 g. in 200 ml. of distilled water. SAND, non-porous, white, silica. API 05751 1/2-n 1 l/3-in f 1/2-inr 3 1/8-in. _L r 3C0-ml. reservoir 25-mm. O.D. coarse white sand alumina 20-mm. coarse fritted disk (Corning 39570 HEYXY) 2I4./I4.O Standard Taper Joints coarse white sand alumina 20-mm. coarse fritted disk (Corning 39570 HEYXY) FIGURE II API 05752 10 PROCEDURE . determination of activity of alumina and corresponding volume of developing solvent required for satisfactory chromatographic fractionation. A chromatographic column of the alumina is prepared as shown in Fig. II. Ten milliliters of the developing solvent are poured on to the column followed by 10,05 ml. of the solution of benzo[]pyrene and then by an additional 10C ml. of the solvent. V Each portion of liquid should be allowed to enter the top layer of sand before the next is added. Additional solvent is then added, 10 ml. at a time, until the top of the benzo[a]pyrene band, as indicated by blue-violet fluorescence under ultraviolet light, is approximately l/k In. below the top of the lower section of alumina. At this point, the fluorescent compound will usually be spread over about one-third of the lower section of alumina. With adsor bent of normal activity, this development will require a total volume of about 120 to lsO ml. of the solvent (not including the " two 10 ml. portions used to prewet the column and as solvent for the benzo[a)pyrene). The total volume of solvent thus determined is used as the "standard volume" for the separatory chromatography (Step B below). The use of alumina requiring a "standard volume" of solvent less than lOOml. is to be avoided if possible, as the effectiveness of the separation of benzo[a]pyrene from interfering components is thereby reduced. The only known disadvantage in the use of alumina requiring a "standard volume" greater than 170 ml. is the additional cost in time and material. Alumina of unusually API 05753 j-igh or low activity could probably be used satisfactorily by niodification of the composition of the developing solvent, but should in that case be checked further with pure benzo[&] pyrene to determine whether it has a satisfactory catalytic action in the iodination reaction described in Step C. gp preparation of fraction for iodination. One hundred milligrams of the material to be analysed dissolved in 5 1. of benzene. After the material is completely in solution, 5 nil, of isooctane are added. (If part of the material is insoluble in benzene or if precipitation occurs on addition of Isooctane the sample should receive special procedure - see Modification # The material is then fractionated on a freshly packed chromatographic column exactly as described in Step A above, substituting this solution of the unknown for the solution of benzo- [&]pyrene used to calibrate the alumina. After the "standard volume" of the developing solvent has run through the column, the sections are separated and the material adsorbed on the lower section is eluted by 50 ml, of a solution of 20% ethanol, B0% benzene, by volume. The solvent is removed from the eluate by careful evaporation under a stream of nitrogen, and the residue, representing a concentrate of any benzo [a] pyrene present in the sample, is then ready for Step C. C. Iodination. The residue from Step B is dissolved in 15-.05 ml. of benzene, and two 5 ml. portions, labeled I and II,of the resulting solution, are transferred to two clean 25-ml. flasks, using the - 12 aine pipette for each. Five milliliters of the standard solution 0dine are added to I,and 5 ml. of benzene added to II. Two parailel columns of alumina are prepared, each one inch deep and one inch in diameter, covered with 1/2 in. of sand. Each .. column of adsorbent is wet by 15 ml, of benzene and allowed to rain until the frequency of drops is less than one in ten seconds. Clean, tared, 125-ml, flasks are then placed under the columns, pive milliliters of the standard solution of iodine, diluted with 5 ml. f benzene, are poured on to the first column, 10 ml. of benzene alone poured on to the second. Solutions I and II are then added to the respective columns, and the flasks which had contained them rinsed by 10 ml. of benzene, and the rinsings added to the proper columns. Each is then eluted by 70 ml. of benzene and allowed to drain as before. The eluate from column I, containing the iodine, is then shaken with 200 ml. of the aqueous sodium thiosulfate In a 500-ml. separatory funnel, after which it is washed with two 100 ml. portions of distilled water. The eluate obtained from column II Is used directly in Step D without further handling. The processed fractions, I and II, should be subjected to the spectrophotometric analysis of Step D without delay, since the iodinated sample occasionally proves to be unstable on standing overnight. D. Determination and analysis of difference spectrum. The algebraic difference between the ultraviolet absorption spectra of the final fractions, I and II, is now measured directly by use of a Beckman DU spectrophotometer. Sample I is placed in API 05755 - 13 - the l-cm * c e l1 normally used for solutions, and sample II in the cell used for a blank. Obviously, the cells used must be well matched as to length. The instrument is operated in the normal fashion except that when II has the higher absorbance, the machine fs balanced on I, and the absorbance is recorded as a negative value (see Fig.III). The slit width must be kept moderately low, below 0,1 mm. being desirable. The sensitivity knob should there fore be turned to a point near its extreme counterclockwise position, and the tungsten lamp should be used. The reduced slit widths are necessary because the iodinated solution I usually has a markedly lower level of fluorescence than the solution II. This difference assumes critical importance when compared with the difference in absorption rather than with the usual total absorption. Hence, if the slit width required at 380 mji exceeds 0.15 nun., 10.00 ml. of each solution should be diluted with 50.00 ml. of benzene prior to spectral analysis. Satisfactory analyses may be obtained from difference spectra in which the measured absorbance at 389 or ij.05 m u is 0.02, or greater, so long as each point is determined carefully and the absorbance is plotted on a suitable scale. However, if the absorb ance is less than 0,02, the weights of the solutions should be determined; then they are concentrated sufficiently to meet this requirement, and their reduced weights adjusted to equal percentages of their respective original weights with benzene (this percentage is used as the value of the factor, V, in the equation below). The difference in absorbance is determined at 380, 309 and L|.05 and at every 2 nyi in the region 1^.10 to 370 mju. The API 05756 - Hi - " Difference Spectrum, (1-1$ - Absorption spectrum of iodinated portion, I absorption spectrum of non-iodinated portion, II ABLIOI BAHCE 0< o< - 0.2 0 o 01 V\| CO 1 o m, -=h I 31+00 3600 3800 1+000 1+200 "AVE LENGTH IN ANGSTROMS ` 1+1+00 FIGURE III - STANDARD DIFFERENCE SPECTRUM 0? BENZ0[a]PYRENE AND IODINATED BENZ0[a]PYRENE " NOTE: ABP = 0.23 in differ ence spectrum obtained from standard analytical procedure on a sample containing 1 mg. of benzojjiJ pyrene, API 05757 T - 15 I f0ll"lnS ^aae-llne technique (1 5 ,16 ) Is applied to the difference I spectrum obtained. A line Is drawn Intersecting the curve at the points corresponding to 1+05 and 380 syj, The distance (In absorbance ^its) along the vertical coordinate at 389 mp from the curve to this line Is designated as aAgp (see Pig. Ill), The percentage of benzo[a]pyrene In the original sample Is calculated from the equation: Percent Benzo [a]pyrene z V, ml. Sample Weight, mg. Abp 0.23 value of 100 ml. is used for the factor, V, unless dilution or concentration of solutions I and II was required to correct for excessive or insufficient absorbance in Step D. A value of 600 ml. Is used in the former case. If it was necessary to reduce the volume, the value of V to be used in the calculation is equal to the percentage determined from the process of concentration. Thus, when 100 mg. of the original material was used in the analysis and such corrections of volume were unnecessary, Percent Benzo [ajpyrene z 1+.35 Agp Modification # 1 - The suspension of 100 mg, of the sample in either 5 ml. of benzene or in 10 ml. of the 50:50 mixture of benzene and isooctane, as the case may be, is poured on to a column of alumina one inch in diameter and one inch deep (comparable to the upper section of the column in Fig, II). The flask in which the suspension was prepared is washed thoroughly with 10 ml. portions of benzene, and the washings added to the column. After this material has passed into the adsorbent the column is eluted with 100 ml, of benzene. API 05758 16 - Host of the solvent Is removed from the eluate by evaporation on a steam bath, the remainder by blowing under a stream of nitrogen. Ihe residue obtained is dissolved in $ ml. of benzene, $ ml. of isooctane are added, and the analysis resumed at the point of interruption (Step A). REPEATABILITY AND INTERFERENCES Quantities of pure benzo[V]pyrene ranging from 0.002$ t o -0.$# were dissolved in several different oils of widely varying type, to which this analytical procedure had been applied pre viously. In each case in which the amount of added benzo[a]pyrene represented a significant increase in the total present, the standard analysis yielded a reasonably precise measure of the quan tity added. The examples, shown in the following table, Indicate that other components of typical petroleum products did not create a serious interference with tne analysis. OIL NUMBER 1 2 3 k 5 BENZO[a]PYRENE FOUND IN ORIGINAL SAMPLE 0.02# 0.38# 0.003# 0,000$# 0.10# CONCENTRATION OF BENZO[a]PYRENE ADDED 0.10# o.$o# o.o$# 0.01# 0.20# TOTAL DETERMINED BY ANALYSIS (to be determined) API 05759 - 1.7 - Data from fifty analyses involving either duplication oT known amounts of added benzo[aJpyrene, indicate a reproducibility #ithin the limits of about - 20# for samples with benzo[a]pyrene c o n c e n tr a ti o n s around 0.01#, 10# for concentrations of 0.10#, and * 5# for concentrations of 0,50#. Many compounds other than benzo [a] pyrene could react with iodine under the conditions of Step C of the analysis. These include anthracene, pyrene, methylcholanthrene, perylene, etc. The majority 0f such compounds are separated from benzo[ajpyrene by the chromato graphic fractionation of"Step B. Of the compounds which accompany benzo[a]pyrene through this separation and the subsequent iodination, the alkyl derivatives of benzo[a]pyrene are perhaps the most im portant potential interferers, if present in any significant concen tration. It is not known how their presence would effect the final analysis. The following experiments were carried out in an effort to obtain some estimate of the extent of such interference, if any. Analysis of a 20# residuum from distillation of the oil from which benzoa]pyrene was isolated indicated a content of 1 .8# of this hydrocarbon, A second sample of this residuum was fraction ated by the procedure of Step B. One-fifth (termed Fraction A in the following table) of the chromatographic fraction containing the benzo [a]pyrene was then subdivided by a more refined chromato graphy into three fractions containing approximately equal portions of benzo [a]pyrene, and these analysed by iodination and difference spectra.. A measured quantity of synthetic benzo V)pyrene was then added to a similar one-fifth portion of the standard chromatographic API 05760 - 18 - fraction from the oil. This sample was then subdivided by a care ful reproduction of the previous chromatography, and the three resulting fractions analysed. Duplications were then made for both experiments. Finally, pure benzo^ajpyrene was processed chromatographlcally in duplicate experiments similar to the preceding ones, and the resulting fractions analyzed. The results of these experi ments are shown in the following table. PERCENTAGE OF TOTAL 3ENZ0ajPYRENE PRESENT IN: SAMPLE_____ Fraction 1 Fraction 2 Fraction 3 Fraction 1+2 + 3 1. Fraction A 29 . 39 31 99 of Cracked Residuum 31 1+2 27 100 2. Fraction A 31 1+2 25 98 + added benzo [Vjpyrene 31+ 1+1 22 97 3. Pure benzo 31 1+7 20 98 ls,j pyrene 28 1+6 26 100 In the cases of samples 1 . and 2., percentages are based on total benzo [aj pyrene indicated oy summation of the weight of syn thetic benzo^ajpyrene added and that determined by independent ordinary analysis of Fraction A. The table shows that the apparent benzo jajpyrene from the oil does not differ chromatographically from synthetic benzoajpyrene. There is some reason, therefore, to believe that alkylbenzoJVJpyrenes are creating no serious interference in the analysis of the oils. API 05761 In the analysis of certain types of materials, such as crude residua, difference spectra are occasionally obtained which lack sharp maxima and minima or well defined inflections and do not furnish satisfactory identification of this hydrocarbon. Hence, the significance of results obtained oy application of the base-line 1 technique to such spectra is somewhat uncertain. Various methods I 0f eliminating the obscuring interference are being examined. I A technique has been devised which, applied to such materi a l ala, provides evidence that the actual concentration of benzo[V]- pyrene is, at most, no higher than that indicated by the standard procedure. Since the values in question have usually been less than 0.003$, and therefore possibly below the range of carcinogenic significance, such evidence may prove very useful. It has been found that a significant proportion of the compounds responsible for the obscuration of the difference spectra may be modified in composition (possibly polymerized) by subjection of a weighed m quantity (approximately 100 mg.) of the sample to what amounts to a flash distillation at atmospheric pressure without reflux. Application of the standard method of analysis to the distillate then yields a much improved difference spectrum. In our exper ience, the value of the concentration of benzo[a]pyrene, thus deter mined, has not proved to be higher than that obtained by the routine method. Indeed, agreement between the values obtained by the two procedures has been rather striking. API 05762 20 SENSITIVITY AND APPLICABILITY Consideration of various factors inherent in the procedure indicates that the limit of sensitivity of the method under ideal con<iitions la about 0.0001 mg. of benzo^aj pyrene. In actual analyses, the presence of other compounds, which modify the shape of the final difference spectra,, limits the practical sensitivity to about 0.002 mg. The method was developed primarily for application to various petroleum products resulting from refining operations. It has been applied to'more than thirty such materials which have been subjected to biological testing for carcinogenic potency. A ll analytical data thus far obtained have been plausible in the light of the observed potencies and the known physical properties of the materials. The use of the method has been extended, in a limited manner, to substances derived from other sources, including coal tars and fractions thereof. No apparent difficulties were encountered with these materials. It seems likely that the analysis may be applied to mixtures extracted by benzene from a variety of additional materials such as soot, carbon black, condensates from atmospheric samples, and other materials which are derived from fossil fuels or the incomplete combustion of organic mixtures. It should be recognized that the results of such analyses would not necessarily indicate the total amount of benzoQiJpyrene in the original sample, since the complete extraction of benzojjfjpyrene from such materials is frequently very difficult to achieve. API 05763 - 21 LITERATURE CITED (1 ) Hieger, I., j. Chem. Soc., 1933. 395. (2) Cook, J. and Hewett, C,, J. Chem. Soc., 1933. 398. (3) Clar,E., ''Aromatisch Kohlenwasserstoffe", Berlin, Springer Verlag, 191+1. (1+) Goulden, P. and Tipler, M., Brit. J. Cancer, , 157 (191+9). () waller, R. E . , Brit. J. Cancer, 6, 8 (1952), (6) Falk, H.L., Steiner, P.E., Goldfein, S., Breslow, A., and Hykes, R., Cancer Research, 11, 310 (1951). (7) Wedgwood, P. and Cooper, R., Analyst, 2.> 1?0 (1953). (8) Windaus, A. and Raichle, K., Ann., 537. 157 (1939). (9) Jones, R. N., J. Am. Chem, Soc., 6j, 2127 (191+5). (10) Benesi, H.A. and Hildebrand, J.H., J. Am. Chem. Soc., 22, 2832 (191+8). (11) Ibid., 21, 2703 (191+9). (12) Mulliken, R.S., J. Am. Chem. Soc., 22, 600 (1950). (13) Ibid., 2k* 811 (1952). (11+) Brown, H.C. and 3rady, J.D., J. Am. Chem. Soc., 2k 3570 (1952). (15) Wright, N., Ind. Eng. Chem., Anal. Ed., 1^, 1 (191+1). (16) Seyfried, W. D. and Hastings, S. H., Ind. Eng. Chem., Anal. Ed., 12_, 298 (191+7). API 05764 API 05767 UNIVERSITY OF CINCINNATI KETTERING LABORATORY ACCOUNT OF FOR AMSRTOAN PSTROT.ETTM TMSTTTIF'E 4th Quartr 1953_________________ SALARIES (Based on Proportion of Time Actually Spent on Project) Direct Salaries....................................................... -6.*3-92*.42..... Indirect Salaries Histopathological Preparation.................. .......... -2.12...00___ Other Services........................................ 7.J.Q68*08.... 13.672.50 MISCELLANEOUS expense Purchase of Animals....... Special Laboratory Supplies ....24.6...SQ...... ....J3.7-5a.4l...... Travel ......................... lj.255.el9... Overhead (Proportion of Heat, Gas, Electricity, Steam, Telephone, General Laboratory Supplies, Postage, Annuities, Pensions, Maintenance, etc...4> Q.95--6.Q..... . 5.977.00 TOTAL . . . . Balance Available for Further W ork at End of.... 3 . L f l S M H & * J 5 2 ------------ 1 2 , 2 1 3 - 5 7 Balance Due Kettering Laboratory at End of..................................................... Receipts .............................................................. .................... Expenditures . A t h - . Q lia i! t . E . . . 1 9 - 5 3 .................. 1 9 , 6 4 9 . 5 0 Balance Available for Further W ork at End of...................................... ............... Balance Due Kettering Laboratory at End of--.kth-QuLat*.--1-953----------- 7,435-93 19,649.50 PI 05768 April 29. 195U I > I U.-J) ( c Dr. Robert Eckardt Standard Oil Development Esso Laboratories p.O. ->ox 51 Linden, lew Jersey Company Dear Doctor Eckardt: There was no opportunity to discuss the man uscript on carcinogenesis in any detail In Cnicago with the subcommittee members. Dave Stroop suggested that I send him the stencils so that he could distribute copies to members of the subcommittee and perhaps also to the general committee. At any rate, 1 expect that you will nave had a chance to Judge from the copy sent to you whether a meeting will be needed to discuss the paper or if you can depend upon an adequate response by coriespondence. 1 believe any date in Kay except the period from Kay 11 to 20 would be OK at this end if you want to call a meeting Sincerely yours, A . Wesley Horton AWH:mJg CC: Mr. David V. Stroop API 05769 Copy From D. V. Stroop For Information of Members of Medical Advisory r-ftfflmlttee and Subcommittee on Carcinogenicity____________ May 6,195^ ESSO LABORATORIES STANDARD OIL DEVELOPMENT CCMPANY P. 0. BOX 51, LINDEN, N. J. May k, 1951* TO: Members of the Subcommittee on Carcinogenicity Gentlemen: I am attaching a copy of a manuscript forwarded to me by Dr. Horton entitled "Carcinogenesis of the Skin: I. Basic Methods Employed in Testing Complex Hydrocarbon Type Tars and Oils, and the Development of a Quantitative Scale to Express Iheir Relative Potencies to Mice". This paper has been written by Dr. Horton and his associate, Dorothy T. Denman. It is his plan, I believe, to submit this paper to Cancer Research for publication, although he has not yet Indicated to me 'just to which journal he plans to submit it for publication. Dr. Horton has indicated that any date in May except the period May 11-20 would be all right for a meeting of the Subcommittee. After having reviewed the manuscript, I believe a meeting of the Subcommittee with Dr. Horton would be advisable, and would suggest that this meeting be held after each of you has had ample time to review the manuscript both personally and possibly with some of your technical people, so that at the time of the meeting with Dr. Horton ve will be in a position to express all suggestions to him. With this in mind I tentatively propose the dates of Friday and Saturday, May 21 and 22, or Thursday and Friday, May 27 and 28 for a meeting with the Kettering group to discuss this manuscript. I would appreciate it if each of you would let me know as soon as possible which of these two dates would be satisfactory so that final arrange ments can be made with the Kettering group. As soon as your wishes are in, I shall send you definite notice of the meeting dates. Sincerely yours, /s/ R. E. Eckardt R. E. ECKARDT, M. D. REE:ilk cc: Members of the Medical Advisory Committee Dr. A. Wesley Horton API 05770 TI m s INFORMATION ONLY NOT FOR P U B LIC A T IO N API RESEARCH PROJECT MC-1 At Kettering Laboratory - University of Cincinnati (l9U6-195k) I Manuscript of paper submitted for review prior to publication. Title: "Carcinogenesis of the Skin, I" Proposed for publication in "Cancer Research" II Status of Hsnuacript m.y 1*, 195U - Received from Kettering Laboratory. May 6, 195^ - Copies mailed to Subcommittee on Carcinogenicity (and to MAC). API 05771 CARCINOGENESIS OP THE SKIN, It Basic Methods Employed in jesting Complex Hydrocarbon-type Tars and Oils, and the Develop ment of a Quantitative Scale to Express Their Relative Potencies to Mice. by A. Wesley Horton and Dorothy T. Denman The investigation of skin cancer in this Laboratory has been concerned with the estimation of the extent of the potential hazard of this disease in certain industries. As in any other aspect of the general field of toxicology, it must be recognized that a specific hazard is the product of several variables, two of which assume dominant importance. These may be termed conven iently the physiological activity of the material or mixture in question, and the conditions of exposure. The latter aspect in volves the practices of personal and industrial hygiene. This paper will be devoted primarily to the development of reproducible methods of determining the first variable, the level of activity, or relative carcinogenic potency, of hydrocarbon-type oils and tars. Mice were ohosen as the primary test species since neo plasms of the skin, comparable in certain respects to those of man In the same tissue, can be reproduced in these animals under suit able experimental conditions. Certain specific materials, which have induced an abnormally high incidence of cancer of the skin of men as a result of repeated exposure in their occupation, have been found capable of producing epitheliomas in the skin of mice. Conversely, in our experience certain inbred mice, such as the C3H strain, have not developed "spontaneous" skin tumors, nor have API 05772 - 2- they reacted by skin tumor formation to repeated experimental application of a number of materials, including water, benzene, white mineral oil, etc,, which are presumed to be non-carcinogenic to the skin of man on the basis of extensive practical experience. The specification, "suitable experimental conditions", in the previous paragraph was used advisedly. It is evident from perusal of the literature, that too often the problem of excessive mortality due to the toxicity of the material under examination, due to as well as that ^ the occurrence of incidental disease among the experimental animals, has been met only by increasing the number of animals on teat. Such an approach ignores the probability that the survivors of these conditions are not at all representative of the population involved with respect to their susceptibility to carcinogenesis. Hence, stress will be laid upon the necessity for standards of selection and care of the experimental animals which will maintain their health at as high a level as possible. EXPERIMENTAL METHODS Most of our work has been carried out with male mice of the C3H strain, purchased from the Jackson Memorial Laboratory, Bar Harbor, Maine, A number of experiments have also been carried out with male mice of the CFW strain obtained from Carworth Farms, but the results were not as reproducible as those derived from the C3H strain. Although the difference in genetic homogeneity is undoubtedly a factor, the fact that the C3H mice are more docile probably contributes to the reproducibility of experiments in which they are employed. The frequent occurrence of bites and scratches API 05773 - 3- of the skin of the CFW mice, inflicted by their companions, probably results in occasional direct intraepithelial introduction of car cinogens. Added to this complication is the possibility that wound-healing may play an accelerating role in carcinogenesis (1 ), The same trouble has been experienced in recent experiments with Swiss males, although the females seem to be more congenial. The mice were received at six to eijht weeks of age and placed under observation for approximately four weeks before use in any experiment. To control infestation of the animals with mites or lice (which when uncontrolled affects adversely their health), the mice were dusted routinely with Aramite1 by the sup plier prior to shipment and then twice again with the same miticide during the period of observation in this laboratory. Any group in which there was evidence of poor health during this period was withheld from use in tests in which quantitative results were needed* In all experiments described herein, the mice were fed on Purina Laboratory Chow and water without restriction. At six months of age, C3H mice fed on this diet and subjected to no adverse I experimental procedure reach an average weight of about 29-31 g., while corresponding CPVr mice weigh about 33-35 g* Since growth of both the hosts and.their tumors might vary with the temperature of their surroundings, an effort has been made to maintain a relatively constant temperature of 75-2 in the animal rooms. For a giver experiment, a group of 20 to 30 mice was divided into three to five cages. The animals were assigned in dividual numbers for identification and marked by clipping their toes. Aramite - 15W, U. S. Rubber Company, Naugatuck, Conn. API 05774 - k - The exposure of the mice to any given material involved re p ea te d applications upon the interscapular region. In preparation f0r the application of dosages of 100 mg. of a material, the fur was left intact, but when the dosage was to be less than 100 mg., the fur was removed by means of electric clippers. A small camel's hair brush cf a selected size to deliver the desired dosage was used to apply the material, which was then simply allowed to spread according to its individual physical characteristics. Applications were repeated in accordance with a predetermined schedule (one, two, or three times each week) until the animals died. Care was taken to keep the dosage and the location of each application as constant as possible throughout the experiment. Normal growth of the mice could not be maintained in testing certain of the materials, particularly those of high potency, when the experiment involved three 100 mg. applications each week. Hence, when there was some physicochemical or biological evidence that a given sample was likely to have a relatively high carcino genic potency to the mice (equivalent to that of a solution of 0.15 percent, or greater, of methylcholanthrene in benzene), the material was applied at a lower frequency, usually once each week. In the case of materials of lower potency, more frequent applications were required to obtain sufficient tumors. Hence, when toxicity of such samples proved a serious problem at the 100 mg. level, the dose was reduced, in some cases to as low as 5 nig. per application. API 05775 - 5 - Experience has shown that, even when the mortality among fln experimental group was low, the time of appearance of tumors was gometimes delayed seriously by conditions of variable or uncertain origin which also interrupted the normal growth of the animals. As a measure of such conditions, therefore, the mice were weighed at frequent intervals throughout the experiments. Since early observations gave evidence that interruptions of growth were usually the result of intercurrent infections of low severity affecting most of the animals in a given cage, the weighing of individual animals was discontinued in favcr of determining the average weight of the survivors in each group. Various precautions, such as regular cleaning and sterilization of cages, feeders and water bottles, decontamination (with a bacteriostatic agent) of gloves used to handle animals from one cage before proceeding to another, and isolation of sick mice, were taken to minimize infec tions and to prevent their spread from one group to another. Throughout each experiment, the animals were examined at least once each week by one of us (D.T.D.), and the presence and time of appearance of any abnormal gross changes in the skin were recorded. Epilation and crusting of small areas were frequently noted and, in the early stages of tests on solutions of greater than 0.2 percent of methylcholanthrene or benzopyrene in benzene, large swellings, of the type previously described by Cramer (2), and small ulcerated areas, were occasionally observed. In response to the application of solutions of these synthetic carcinogens, Papillomas eventually developed on the skin of practically all of API 05776 - 6- th0 s u r v i v i n g animals except i n the case o f the lowest level of c o n c e n tr a tio n tested ( 0 ,0 1 percent). In general, the induced neo plasms progressed through the familiar pattern from papilloma to squamous cell carcinoma, with gross evidence o f invasion. In the f i r s t teat of a given material, sections of any epidermal tumors produced were prepared for microscopic examination. The date of which each typical papilloma reached the arbitrary size of 1 mm, in diameter and elevation was recorded as the "time of appearance" of the papilloma. In addition, the date on which it was grossly apparent that the turner was invading the subcutaneous tissues was noted (rolled border, generally accom panied by cratering necrosis of the center). When a mouse developed such an apparently malignant tumor, it was killed. Usually, the others were maintained on test until death from "natural causes" occurred. The data, thus obtained, were recorded on graphs such as those shown in Figures 1 through i;. As experience was gained In associating the physical and chemical characteristics of various types of materials with their relative toxicity and with their ability to induce tumors of the skin in C3H mice, certain criteria were developed for selecting optimum conditions for any given biological test. As a general practice, a schedule of application was chosen which would lead to induction of papillomas in an average time of not less than 12 weeks. The restriction was based on repeated observations that this average time was unduly variable in experiments in which the severity of exposure resulted in more rapid induction of tumors. API 05777 - 7- on the other hand, experimental conditions which involved very long periods of exposure before tumors were induced proved to be even more unsatisfactory than those associated with very short latent periods, because cf the increased possibility of inter current infections or other unidentified interferences with the normal growth of the mice. Even when these interruptions resulted in little or no loss of animals, the irregular effects on the growth of the induced tumors made it very difficult to interpret the results of the experiment. Therefore, as a general rule the frequency and severity of the applications were scheduled to produce an average latent period as close to 15 weeks as possible. Vhen the materials were low in potency, this obviously meant the use of as severe conditions as were compatible with the normal growth and survival of the mice. INTERPRETATION OF DATA OBTAINED In general, under these experimental conditions one of two results has been obtained in response to the applications of any one of a series of materials, l.e., most of the animals which lived long enough developed tumors of the skin at the site of the application, or none of them did so. The outcome of the experiments, therefore, has been essentially unequivocal, and the relative potencies of the different active materials could be expressed in terms of the relative rates of induction of tumors. A p l 05778 -- 8-- The calculation of the "mean time of appearance of tumors" involved the determination, by the method of least squares, of the linear function which best fitted the plot of the cumulative frequen cy of tumor response (in probits) versus the time (in weeks) after the first application (shown on the right side of Figures 1 through li). The latter coordinate is equivalent to cumulative dosage. It is to be noted that, for these experiments involving repeated applications of carcinogenic materials to the skin, the response varied linearly with the dosage itself, and not with the logarithm of dosage. It is assumed that the exponentially decreasing rate of appearance of tumors usually observed following single or limited numbers of exposures to carcinogens (3 ) was not seen in these ex periments because the more resistant animals in any given group were given a larger effective dosage than were the more susceptible. Calculation of Cumulative Frequencies of Response The convention is adopted (after Bliss, l|.) that, at the midpoint of the period between two given obser vations, the number of mice bearing tumors is equal to the average of the numbers at the times of the obser vations. At the midpoint of each period during which one or more mice develop their first papilloma, a calcu lation is made of the cumulative percentage of the "effective group of mice" which bear tumors on that date. The "effective group" at a given time is defined as the sum of the number of survivors plus the number of API 05779 - 9mice which died previously after developing tumors of the skin, except that after the appearance of a tumor in the ''average" mouse this sum is held constant. The "average" mouse is designated by the median in a simple arithmetic array of the times of appearance of the first tumor in each animal. As an example, the calculations used to obtain the cumulative frequencies of tumor response from the data of Figure 3 are shown in the following table: API 05780  X A Pitt 1. TIME* OF OBSERVATION OF NEW POSITIVES MIDPOINT OF PERIOD BETWEEN OBSERVATIONS, X NUMBER OF NEW POSITIVES, R TOTAL NUMBER OF MICE BEARING TUMORS,** R* + ZR CUMULATIVE RESPONSE, ZR + (ZR + R') 2 EFFECTIVE NUMBER OF MICE PREQUEN( OF RESPONSE (probit 19Ls5. 20 1 0.5 18 3.08 1 2iiS 2h 1 1.5 18 3.62 2 2k*l 25 1 31*1 3k 2 Usi 36 3 2.5 18 2i2i 3 0 li-.o 16 k.33 1 5 6.5 16 U.76 8 Usi 38 2 9.0 10 16 iili 30^5 39 2 12 U2 1 13 U6.$ k7 2 15 * Number of weeks after first application. 11.0 16 Zdil 12.5 16 l*ll 11*. 0 16 6i l l previou^oBserva^ioS with tumors at the API 05781 - 11 calculation of Mean Time of Appearance of Tumors, x ; 1 . Assume that the tumor response in probits, r, is a linear function of the time, x, or r = mx + b 2. Determine constants, m and by method of least squares (5). Normal equations: Zrx = mZx2 + bZx r - mix + 9b (Values of x in weeks and r in probits are under lined in proceeding table) IJ4.85.67. = 10,-688.25m + 300.5b I4.2.28 = 300.5m 9 b b - 0.960 m = 0.112 a slope of dose-response relationship .*. r = .112x + .960 3. x : time of 5-probit response = 3 6 .1 weeks s, standard deviation of mean, z x minus time of lj.-probit response 8.9 weeks. CLASSIFICATION OF EXPERIMENTAL RESULTS ON THE BASIS OF THE HEALTH OF THE ANIMALS In the appraisal of data derived from these experiments, consideration has been given to the following apparent relation ships between the average growth and survival of the mice and the mean "time of appearance" of their epidermal tumors. API 05782 12 1. If, in the course of an experiment, an intercurrent i n f e c t i o n overtakes a group of mice at about the time when pap illomas would be expected to appear (on the basis of other compar able experiments),, there is usually a delay in the mean time of appearance of tumors. The infection need not be lethal to any of the mice, a change in sign of the slope of the average growth curve frequently being an indication of such a condition (see Figure 1). It is believed that, in many such instances, the 0lce would probably be unaffected were their resistance not diminished somewhat by systemic toxic effects induced by the hydrocarbons being applied upon their skin. Thus, untreated con trols may not show any symptoms of the infection. Not infrequently, as the animals recover from such dis turbances, tumors grow to measurable size in a large proportion of the animals in a relatively short time. The slope of the plot of the frequency of response versus time is then exceedingly steep. In other words, the standard deviation of the mean time of appear ance of tumors in the group is much smaller than normal (see Figure 2). Thus, it could be assumed mistakenly, If It were not for the evidence from the weight curves of the presence of uncon trolled variables, that such an experiment had defined the statistical limits of the mean better than another in which such disturbances had not occurred. ApI 05783 2. In some experiments the cumulative toxic effects of the material applied, or chronic infections, or a combination of both, result in prolonged inhibition of the growth of the mice. 4 number of examples have been observed in which the average weights have increased for a few weeks, but became stationary at a point 10 percent or more below that of adult controls, and then, in many instances, gradually declined (see Figure 3). Thus, whether or n0t these conditions influence the timing of the initial changes resulting in the neoplasms they may retard the growth of the in duced tumors to visibly discernible size and, accordingly, lengthen the apparent average period of exposure required for the induction of tumors by the material under test. In such experiments, the standard deviation of this mean time is frequently much greater than normal. This somewhat extended discussion of the apparent relation ships between the irregularities in growth due to the toxicity of applied materials and to intercurrent infections, and the rate of induction of tumors, does not relate to the originality of this observation but rather from our recognition of its importance in experiments in which various materials are being compared. Occasion al mention of the phenomenon has appeared in the literature on experimental carcinogenesis (i.e., 6,7 ,8) and, in the field of chemotherapy of cancer, recognition has been given to the necessity of considering the systemic toxicity of a material in evaluating the specificity of its apparent "anticarcinogenic" effects (i.e., 9 ) API 05784 - Ik - ^ t h e r , the relationships between general health and susceptibility spontaneous" tumors in man are currently receiving attention, iience, is believed that, in the experimental assay of the relative p oten cies of various materials or of the susceptibility of various sp ecies or strains of one species to a given material, information on the health and survival of the groups of animals being compared Is essential to any quantitative interpretation of the data. The results of all experiments described herein have been examined, therefore, in the light of these relationships. Each one has been classified according to the estimated significance of such effects in the following manner. Class A. Those characterized by normal growth and'survival of the mice at least until the time when most of the animals have developed papillomas (see Fig.ij.). Class 5. Those in which all or most of the mice lived long enough to develop tumors, but in which there was evidence (from average weights) of intercurrent infection at times prior to or during the period when the new growths were appearing. So long as recovery from the infection was prompt, and no obvious skewing of the distribution of the times of appearance of tumors occurred, it was felt that reliable estimates of the mean time of appearance could reasonably be made (see Figure 1). However, when no tumors had appeared prior to the onset of an infection, and when with recovery tumors developed API 05785 - 15 - somewhat precipitously, the reliability of the apparent mean time of appearance was not determin able, and therefore such tests were classified X (see Figure 2), Class X. Those characterized by poor growth (as discussed in Section 2 above) or high rate of mortality. The confidence'limits of the apparent mean time for the induction of tumors could not be calculated by stand ard -statistical methods, because of inability to assign proper weight to the effects of the uncon trolled variables. Several types of moderately toxic materials failed to y ie ld experiments of the Class A category. On the other hand, experimental results of the frankly unsatisfactory type (Class X) were usually found to be avoidable through the exercise of appro priate care in the initial choice and the subsequent handling of the mice. REFERENCE STANDARDS To provide a background of reference which might permit a more nearly quantitative interpretation of the mean times of appearance of tumors induced in C3H mice by certain complex oils and tars under various conditions of exposure, a number of experi ments have been carried out over the past five years employing solutions of the synthetic carcinogens, 3-methylcholanthrene, benzo [b3 pyrene, and 7 ,12 -dimethylbenz [a-]anthracene, in various solvents. It has been found that one of the essential requirements API 05786 - 16 0T obtaining reproducible and logically consistent results is the maintenance of the animals in a good state of health as measured growth curves. Thus, in Table II, the results of Experiments P05 show the precision of the determination of the mean time, x, when Class B experiments are involved. Similarly, it has been found that even in the case of a complex material, so long as the conditions of the experiments permitted reasonably normal growth 0f the animals, the repeatability of the determination of the mean time, x, has been quite satisfactory(within 10$). To determine the effect of variations in the level of concentration and in the frequency of exposure, comparable experiments were conducted with other solutions of methylcholanthrene in benzene, and the results are shown in Table II. The purity of the polycyclic hydrocarbon is obviously an important consideration in obtaining consistent results. The usual chemical criteria are necessary, but not always sufficient, requirements. In one case in particular, the methylcholanthrene, as received, had a sharp melting point, 176-7C., and no indication was apparent by its ultraviolet absorption spectrum of any contamination. Yet, on repeated bioassay, it proved to be excessively irritating and much more rapidly carcinogenic than the normal material. Un fortunately the quantity available did not permit deter mination of a satisfactory procedure for purification. Since no further sample with such unusual potency has been obtained, bioassay remains the only way of acertaining that the methylcholanthrene is free of the material responsible for this activity. API 05787 ) ) TABLE IX EXPERIMENTS INVOLVING THE APPLICATION OP SOLUTIONS OP METHYLCHOLANTHRENE IN OEN'ACNE UPON THE SKIN OF C3H MICE EXPERIMENT NUMBER CONCENTRATION OF METHYLCHOL ANTHRENE IN SOLUTION (% by weight) DOSAGE OF SOLUTION PER APPLICATION (mg.) NUMBER OF APPLI CATIONS PER WEEK DOSAGE OF METHYLCHOLANTHRENE PER UNIT AREA1 PER WEEK , (mg.)________ MEAN TIME OF APPEARANCE OF TUMORS, x (weeks) CLASS CAT 201 0.012 100 3 0.036 72.6 X 202 0.01^ 100 3 .0.132 30.1|3.8* B 203 0.086 100 3 0.258 29.9*2.2 B 203 0.086 100 3 0.258 26.3-3.1* B 20k 0.172 100 3 0.516 15.7*1.6 B 20k 0.172 20 3 0.516 16 .14.il. 1 B I 20k 0.172 20 3 0.516 16.611.2 B 20k 0.172 20 3 0.516 15.5+1.1 B 20$ 0.3U5 100 3 1.035 1 1 .01 . 9 B 205 0.3U5 100 2 0.69 15.1*1.5 B 205 0.31*5 100 1 0.31*5 22.I|i2 .7 B 205 0.31*5 100 3 1.035 11*.7 X 206 0.69 100 1 0.69 17.8 X 206 0.69 100 3 1.98 6.7l*.l B 1 Area of skin covered by 100 mg. of solution of methylcholanthrene in benzene. 3 $% fiducial limits. I API 05788 ) ) ) EXPERIMENTS INVOLVING THE APPLICATION OP SOLUTIONS OP METHYLCHOLANTHRENE IN BENZENE UPON THE SKIN OP C3H MICE EXPERIMENT NUMBER CONCENTRATION OP METHYLCHOL ANTHRENE IN SOLUTION (5? by weight) DOSAGE OP SOLUTION PER APPLICATION (mg.) NUMBER OP APPLI CATIONS PER WEEK DOSAGE OP METHYLCHOLANTHRENE PER UNIT AREA1 PER WEEK (mg.) MEAN TIME OP APPEARANCE OP TUMORS, 5c (weeks) CLASSIPI CATION 207 1.15 100 3 209 0.517 20 1 209 0.517 100 1 209 0.517 100 3 219 0.115 100 3 220 1.031* 100 1 238 0.23 20 2 238 0.23 100 2 21*3 0.15 100 3 21*3 0.15 20 3 268 0.287 20 2 3.1*5 0.517 0.517 1.551 0.31*5 1.031* 0.1*6 0.1*6 0.1*5 0.1*5 o.57l* 8.9 X 18.3 1 1.9* B 16 .22.7 B 10.1* X 21*.0 X 9.7*1.6 B 18.7*2.3 B 1 6 .1*2.0 B 17.92.o A 18.22 .1 B 17.6 X 1 Area of skin covered by 100 mg. of solution of methylcholanthrene in benzene. a 5% fiducial limits. ll it--Jill* >111 Mh API 05789 TI - 19 In Table II it is apparent that variation in the dosage 0f solution per application from 20 to 100 mg., other factors remaining constant, had no effect on the rate of Induction of tumors tjy methylcholanthrene in benzene. Similar results have been observed when solutions of benzopyrene in benzene were used as the carcinogenic stimulus. This constancy of potency with variation in the quantity 0f material applied (or area of skin exposed) also holds for some complex oils and tars, but for others there is significant change of potency with changes in the quantity used for each application. The chemical basis for these differences will be discussed in a later paper. . The values in column 5* Table II, were calculated from the concentration of methylcholanthrene in the solutions and the number of applications per week. It has been determined experi mentally that variations in the concentration of the carcinogen did not affect the area covered by a given weight of the solutions. It was further assumed that the dosage per unit area was not slgnlfi- 0 cantly different when 20 or 100 mg. of the same solution were and applied. Comparison of the data in columns 5X& shows a consistent inverse relationship between the dosage of carcinogen per unit area per week and the mean time of appearance of tumors. The data (omitting Class X experiments) are plotted in Figure 5 together with the curve of the corresponding hyperbolic function (calculated by the method of least squares), {* - 3.7 )(d + 0.1 0 ) = 8.2 8 2 or. x = + 3.7 (d + 0.10 ) ***,? API 05790 - 20 - j{. is interesting to note the similarity between this equation flnd that calculated from theoretical considerations of the mechanism 0f carcinogenesis by Iversen and Arley (10), taT5 s + t , where k and T are constants.. P kc0 -j>he constant, 0.10 , in the experimental equation might be inter preted as a correction required by the definite, though very small, probability of the ''spontaneous" excitation of a cell of the dorsal skin of a C3H mouse to the rate of proliferation necessary for the development of a new growth. Similar experiments have been carried out with solutions of two other commonly used carcinogenic hydrocarbons, benzopyrene and 7,12-dimethylbenzanthracene. The results are shown in Table III, together with some data on solutions of methylcholanthrene in sol vents other than benzene. As is evident from the data, it has been difficult to obtain satisfactory experiments withbenzopyrene on C3H mice. The growth of the animals has tended to be extremely erratic. Even in tests in which the average weight eventually reached 30 g., the curve was, more often than not, Interrupted at critical times by sharp downward breaks. In such experiments, recovery from these periods of poor health was sometimes accompanied by the appearance of tumors in a large percentage of the animals in a relatively short time (relationship exemplified in Figure 2). It will also be noted that, even at high levels of concentration, the mean time of appearance of tumors was in no case less than 20 weeks. API 05791 ) ) ) EXPERIMENTS INVOLVING THE APPLICATION OF SOLUTIONS OF SYNTHETIC CARCINOGENS IN VARIOUS SOLVENTS UPON THE SKIN OF C3H MICE EXPERI MENT CARCIN NUMBER OGEN1 SOLVENT CONCENTRATION OF CARCINOGEN IN SOLUTION {% by weight) DOSAGE OF SOLUTION PER APPLI CATION (mg.) NUMBER OF APPLI CATIONS PER WEEK n w c! O WHWO SHOW M > . J 1 Pi d MEAN TIME OF APPEARANCE OF TUMOR S , x (weeks) RELATIVE CARCINO GENIC CLASSIFI POTENCY, CATION PMC 215 DMBA Benzene 0.115 100 3 0.31*5 15.63.33 B 0.19 211* 21*1* DM BA BP Benzene Benzene 0.31*5 100 0.172 100 3 ' 1.035 7.1*11.1 B 0.62 3 0.516 25. H 2 .5 B 0.10 21*1* BP Benzene 0.172 20 3 252 BP Benzene 0.322 20 1 252 BP Benzene 0.322 20 3 252 BP Benzene 0.322 100 3 282 BP Benzene 0.092 20 3 0.516 0.322 0.966 0.966 0.276 ^27.3 ^35.0 V23.0 ^20.1 X (0.09) i A) X (0.18) X (0.11) X (0.13) X 261* BP White oil 0.175 20 3 0.525 ^25.1* X (0.10) 270 BP White oil 0.35 20 3 1.05 ^ 21.5 X (0.12) 285 BP Cotton- 0.50 20 3 seed oil 286 BP Cotton- 1.00 20 3 seed oil 281 MC Cottonseed 0.01*2 100 3 1.50 3.00 0.126 V20.1* ^ 1*1.0 X B(X) B (0.13) 0.0i*7 1 Carcinogens are^, 12-dimethylbenz [al anthracene,DMBA;benzo fa] pyrene, EP;3-methylcholanthrene MC< a Area of skin covered by 100 mg. of solution of methylcholanthrene in benzene. 3 % fiducial limits. API 05792 - 22 - ffSE OF REFERENCE STANDARDS AS A SCALE OF RELATIVE POTENCY OF COMPLEX TARS AND OILS Various methods of expressing the relative potencies of the pure polycyclic carcinogens and complex materials have been reviewed by Badger (11). Many of these are simply rough classifi cations based upon a particular investigator's personal experience. The Iball index (12 ) and a recent adaptation (13) attempted a more quantitative approach, involving the use of the reciprocal of the mean time of appearance of tumors. Such methods are useful in com paring results obtained in a given laboratory using a standardized and inflexible schedule of application of various materials. How ever, experimental techniques differ significantly from one research organization to the next, thus rendering the comparison of data on this basis rather difficult. Further, as was pointed out previously in this article, it was sometimes necessary in the current investigation to change either or both the frequency and severity of exposure to obtain satisfactory conditions for the measurement of the potency of a sample while minimizing its toxic side effects. Hence, a scale of relative potencies which would permit such flexibility in experimental design was needed. The use of the data on relative rates of induction of tumors by the pure carcinogens, methylcholanthrene and benzopyrene, as external standards of reference, provided a practical solution to this problem. The latter is one of the important contributors to the potency of cracked residua (1 4 ); the former has not been identified in complex tars and oils but might be regarded as a API 05793 - 23 - rep r e s e n t a t i ve of the Ij.-ringed carcinogens supposed to be present (1?,lo)* Thus, the mean time of appearance of tumors induced by some complex tar or oil may be translated into a value of Relative Carcinogenic Potency, PMC ( potency as compared to methylchol anthrene), by use of Figure 5. The dosage of methylcholanthrene corresponding to the given mean time is either read off the graph or calculated from the equation, d = 8.2 * " 3.7 - 0.10 Then, the value of the relative potency, PM C , for the complex material in question is PMC - d for an experiment = 1 , for an experiment involving one application each week, involving two applications each week, = d > for an experiment involving three applications each week; 3 Thus, by comparing the results of tests on complex mater ials with those involving methylcholanthrene as an external standard of reference, one may obtain the relative carcinogenic potencies of the different materials, even though the experiments may have in volved differing frequencies of applications. It should be noted, however, that the description of the potency of a complex oil can not be considered complete until it has been tested at two apprec iably different dosages, i.e., 20 mg. and 100 mg. per application, unless negative results were obtained at the higher level of dosage In a test in which the mice survived and grew at a normal rate. API 05794 - 2k - Thusf a S iven ma7 b much more potent than another under gevere conditions of exposure, but the former may actually have a lower potency than the latter if the comparison is carried out under mild conditions. Chemical research on such oils has led to an understanding of this apparent paradox, and the matter will be dealt with in a later publication. The technique of relating results on complex materials to those on standard synthetic carcinogens should facilitate compar ison of results from different laboratories. The synthetic com pound chosen should, of course, be comparable to the type of carcin ogens supposed to be present in the complex mixtures. Under appro priate conditions of dosage and exposure, it must be capable of inducing tumors at least as rapidly as the most potent of the samples to be tested. Thus, benzopyrene has not proved to be a suit able standard for work with C3H mice, since it will not induce tumors sufficiently rapidly even at high levels of concentration. 7,12 -Dimethylbenz{alanthracene on the other hand, is very sat isfactory from this standpoint, but has the disadvantage of requir ing special care in handling to prevent its oxidation to the non carcinogenic 7 ,12-photooxide (7)* It should be reemphasized that this consideration of a quantitative scale of potencies applies only to the results of experiments which may be classified as satisfactory (A and B by our criteria). Those in our Class X have only a limited value. If positive results are obtained, the material may be classified qualitatively as carcinogenic; if survival is good and most of the mice eventually develop tumors (X classification based on certain irregularities in growth) the estimation of the apparent potency may be useful as a minimum value. ^ ----------------- - API 05795 - 25 - RELATIVE POTENCIES OF POLYCYCLIC AROMATIC COMPOUNDS The data available to the writers on experiments involving repeated applications of solutions of various polycyclic carcinogens in solvents such as benzene, acetone, and white mineral oil of high viscosity, upon the skin of mice, indicate that the average rate 0f induction of papillomas is independent of the quantity of sol ution used in each application so long as it exceeds 10 mg. The controlling variables are the frequency of application and the dosage of carcinogen per unit area of the skin. The latter will depend upon the concentration of carcinogen in the solution applied and the relative spreading coefficient of the solution. Solutions of polycyclic hydrocarbons in different solvents have different capacities for spreading over the skin of a mouse because of variations in such physical properties as volatility, interfacial tension, and viscosity. Since much of the previous experimental testing of polycyclic compounds for their carcinogen icity to the skin has been carried out with solutions in benzene, this solvent was used for the reference standards described herein. An intensive investigation of the effects of variation of the com position of the solvent has also been carried out and will be described in a later paper. By comparing data on the rate of induction of tumors by solutions of known concentration of various aromatic hydrocarbons in benzene under suitable experimental conditions, with such data as those shown in Table II and Figure 5 it should be possible to arrive at logical estimates of the relative carcinogenic potencies API 05796 - 26 - f the different compounds. By inference, the relative potency of ^thylcholanthrene on the PMC scale would be 100, Corresponding numerical values for benzopyrene and 7,12-dime thy lbenzanthracene may then be derived from the data of Table III, Thus, since 0.345 percent 7, 12-dimethylbenzanthracene produced tumors at the same rate as 0,62 percent methylcholanthrene under comparable conditions, tbe relative potency of this dimethyl- derivative is p = J L x 0,62 = 100 M0 0.3W prom Experiment 215, the relative potency, PMC, for 7,13-dimethyl- benzanthracene is 165.' The higher value is preferred at the present time, since, if autoxidation of the hydrocarbon caused any reduction In activity, it might reasonably be assumed that the effect would have been greater in Experiment 215. Similarly, the relative potency of benzopyrene on this scale is 57 - 65 (Experiments 244, 252, and 282), The data from Experiment 252 Involving three applications each week cannot be used since apparently the maximum effective dosage of this carcin ogen had been reached at a lower level. Thus, as a general rule the determination of the relative potency of. any given compound should be based upon the results of experiments at at least two different concentrations, which yielded significantly different mean times of appearance of tumors under otherwise comparable conditions. At least one and preferably both of these should meet our classification A or B from the standpoint of growth and survival. API 05797 - 27 - LITERATURE CITED (1) Pullinger, B. D., J. Path. Bact., 1 301 (1943) (2 ) Cramer, W. and Stowell, R. E., J. Nat. Cancer Inst., 2j 3&9 (1942). Bryan, W.R. and Shimkin, M.B., J.Nat.Cancer Inst., 1, 807 (1941), !)) Bliss (t) Whittaker, E. and Robinson, G., The Calculus of Observation, (5) London, Blackie and Sons, Ltd., 1924 P. 209. ( 6) Watson, A. P. and ellanby, E., Brit. J. Exp. Path,, 11,267 4930). ( 7) Bradbury, J. T., Bachmann, W.E., and Lewisohn, M.G., Cancer Research, 1L, 685 (144D. ( 8) Cook, J.W. and Kennaway, E.L. (9) (10) Iversen, S. and Arley, N., Acta Pathol. Microbiol. Scand., 21, 1 (1950). (ID Badger, G.M., Brit. J. Uancer, 2, 309 (1948). (12) Iball, J., Am. J. Cancer, 188 (1939). (13) Blending, P.H., King, W.H., Priestley, W., and Rehner, J., Arch, Ind, Hyg. Oco. Med., 335 (1951). (14) Tye, R., Graf, M.J., and Horton, A.W., (Anal, Chem., 1954). (15) Berenblum, I. and Schoental, R., Brit. J. Cancer, 1, 157 (1947). (16) Fischer, H.G.M., Priestley, W., Eby, L.T., Wanless, G.G., and Rehner, J., Arch. Ind. Hyg. Occ. Med., 315 (1 9 5 D . API 05798 "7 EXPERIMENTS INVOLVING THE APPLICATION OP MATERIALS UPON THE SKIN OF C3H MICE LEGEND Symbols used In connection with average weight curves; ^ 7 Time of death (from disease) of i t h tumor-free mouse. L Time of death (from disease) of or T Time mouse killed. i th tumor-bearing mouse * Painting discontinued for 1 week at each Indicated time. Gross and microscopic pathology: Time of appearance of first papilloma In lth mouse. When this symbol Is first used for a given mouse after its death, the presence of non-invading carcinoma (intra-epithelial) or small areas of benign neoplasm was determinable only by histopathology. Time of appearance of gross changes indicative of malig nancy In tumor of ith mouse. Diagnosis confirmed except in cases where no tissue section available. When this symbol is first used for a given mouse after its death, the invasive malignancy of the tumor was determinable only by histopathology. No tissue section available for histopathology because of extensive post-mortem decomposition or cannibalism. API 05799 FIGURE 1 EXPERIMENT 2 2 0 ,SOLUTION OF METHYLCHOLANTHRENE, 1.03U- PERCENT IN SENZENE i ) ) PIUURK 2 EXPERIMENT 2 6 8 , SOLUTION O P METHYLCHOLAUTHRENE, 0 . 2 6 7 PERCENT IN iiENZENE AVERAGE WEIGHT OP SURVIVORS (grams) ) ) ) AVERAGE WEIGHT OP SURVIVORS (grams) ) ) ) ) ) VOR INFORMATION ONLY NOT FOR PUBLICATION API RESEARCH PROJECT MC-l At Kettering Laboratory - University of Cincinnati (19^6-195^4-) Manuscript of paper submitted for review prior to publication. Title: "Carcinogenesis of the Skin, I" Proposed for publication in "Cancer Research" Status of Manuscript May h, 1951*- - Received from Kettering Laboratory. May 6, 1951* - Copies mailed to Subcommittee on Carcinogenicity (and to MAC). CARCINOGENESIS OF THE SKIN* I. Basic Methods Employed in Testing Complex Hydrocarbon-type Tars and Oils, and the Develop ment of a Quantitative Scale to Express Their Relative Potencies to Mice, by A. Wesley Horton and Dorothy T. Denman The investigation of skin cancer in this Laboratory has been concerned with the estimation of the extent of the potential hazard of this disease in certain industries. As in any other aspect of the general field of toxicology, it must be recognized that a specific hazard is the product of several variables, two of which assume dominant importance. These may be termed conven iently the physiological activity of the material or mixture in question, and the conditions of exposure. The latter aspect in volves the practices of personal and industrial hygiene. This paper will be devoted primarily to the development of reproducible methods of determining the first variable, the level of activity, or relative carcinogenic potency, of hydrocarbon-type oils and tars. Mice were chosen as the primary test species since neo plasms of the skin, comparable in certain respects to those of man in the same tissue, can be reproduced in these animals under suit able experimental conditions. Certain specific materials, which have induced an abnormally high incidence of cancer of the skin of men as a result of repeated exposure in their occupation, have been found capable of producing epitheliomas in the skin of mice. Conversely, in our experience certain Inbred mice, such as the C3H strain, have not developed "spontaneous" skin tumors, nor have API 05806 T - 2 - they reacted by skin tumor formation to repeated experimental application of a number of materials, including water, benzene, white mineral oil, etc., which are presumed to be non-carcinogenic to the skin of man on the basis of extensive practical experience. The specification, "suitable experimental conditions", in the previous paragraph was used advisedly. It is evident from perusal of the literature, that too often the problem of excessive mortality due to the toxicity of the material under examination, due to as well as that * the occurrence of incidental disease among the experimental animals, has been met only by increasing the number of animals on test. Such an approach ignores the probability that the survivors of these conditions are not at all representative of the population involved with respect to their susceptibility to carcinogenesis. Hence, stress will be laid upon the necessity for standards of selection and care of the experimental animals which will maintain their health at as high a level as possible. EXPERIMENTAL METHODS Most of our work has been carried out with male mice of the C3H strain, purchased from the Jackson Memorial Laboratory, Bar Harbor, Maine, A number of experiments have also been carried out with male mice of the CFW strain obtained from Carworth Farms, but the results were not as reproducible as those derived from the C3H strain. Although the difference in genetic homogeneity is undoubtedly a factor, the fact that the C3H mice are more docile probably contributes to the reproducibility of experiments in which they are employed. The frequent occurrence of bites and scratches API 05807 - 3- 0f the akin of the CFW mice, inflicted by their companions, probably results in occasional direct intraepithelial introduction of car cinogens. Added to this complication is the possibility that wound-healing may play an accelerating role in carcinogenesis (1), iphe same trouble has been experienced in recent experiments with gwiss males, although the females seem to be more congenial. The mice were received at six to eijht weeks of age and placed under observation for approximately four weeks before use in any experiment. To control infestation of the animals with mites or lice (which when uncontrolled affects adversely their health), the mice were dusted routinely with Aramite1 by the sup plier prior to shipment and then twice again with the same miticide during the period of obiervation in this laboratory. Any group in which there was evidence of poor health during this period was withheld from use in tests in which quantitative results were needed. In all experiments described herein, the mice were fed on Purina Laboratory Chow and water without restriction. At six months of age, C3H mice fed on this diet and subjected to no adverse experimental procedure reach an average weight of about 29-31 g. hile corresponding CP*" mice weigh about 33-35 g Since growth of both the hosts and their tumors might vary with the temperature of their surroundings an effort has been made to maintain a relatively constant temperature of 75-2 in the animal rooms. For a giver experiment, a group of 20 to 30 mice was divided into three to five cages. The animals were assigned in dividual numbers for identification and marked by clipping their toes. 1 Aramite - 15W, U. S. Rubber Company, Naugatuck, Conn. API 05808 - 4- The exposure of the mice to any given material involved repeated applications upon the interscapular region. In preparation for the application of dosages of 100 mg. of a material, the fur wa3 left intact, but when the dosage was to be less than 100 mg., the fur was removed by means of electric clippers. A small camel's hair brush of a selected size to deliver the desired dosage was used to apply the material, which was then simply allowed to spread according to its individual physical characteristics. Applications were repeated in accordance with a predetermined schedule (one, two, or three times each week) until the animals died. Care was taken to keep the dosage and the location of each application as constant as possible throughout the experiment. Normal growth cf the mice could not be maintained in testing certain of the materials, particularly those of high potency, when the experiment involved three 100 mg. applications each week. Hence, when there was some physicochemical or biological evidence that a given sample was likely to have a relatively high carcino genic potency to the mice (equivalent to that of a solution of 0.15 percent, or greater, of methylcholanthrene in benzene), the material was applied at a lower frequency, usually once each week. In the case of materials of lower potency, more frequent applications were required to obtain sufficient tumors. Hence, when toxicity of such samples proved a serious problem at the 100 mg. level, the dose was reduced, in some cases to as low as 5 mg. per application. API 05809 - 5- Experience has shown that, even when the mortality among an experimental group was low, the time of appearance of tumors was sometimes delayed seriously by conditions of variable or uncertain o r i g i n which also interrupted the normal growth of the animals. As a measure of such conditions, therefore, the mice were weighed at frequent intervals throughout the experiments. Since early observations gave evidence that interruptions of growth were usually the result of intercurrent Infections of low severity affecting most of the animals in a given cage, the weighing of individual animals was discontinued in favor of determining the average weight of the survivors in each group. Various precautions, such as regular cleaning and sterilization of cages, feeders and water bottles, decontamination (with a bacteriostatic agent) of gloves used to handle animals from one cage before proceeding to another, and isolation of sick mice, were taken to minimize infec tions and to prevent their spread from one group to another. Throughout each experiment, the animals were examined at least once each week by one of us (D.T.D.) and the presence and time of appearance of any abnormal gross changes in the skin were recorded. Epilation and crusting of small areas were frequently noted and, in the early stages of tests on solutions of greater than 0.2 percent of methylcholanthrene or benzopyrene in benzene, large swellings, of the type previously described by Cramer (2), and small ulcerated areas, were occasionally observed. In response to the application of solutions of these synthetic carcinogens, papillomas eventually developed on the skin of practically all of API 05810 T -- 6 -- th0 surviving animals except in the case of the lowest level of concentration tested (0.01 percent). In general, the induced neo plasms progressed through the familiar pattern from papilloma to squamous cell carcinoma, with gross evidence of invasion. In the first test of a given material, sections of any epidermal tumors produced were prepared for microscopic examination. The date of which each typical papilloma reached the arbitrary size of 1 mm, in diameter and elevation was recorded as the "time of appearance" of the papilloma. In addition, the date 0n which It was grossly apparent that the tumor was invading the subcutaneous tissues wai noted (rolled border, generally accom panied by cratering necrosis of the center), When a mouse developed such an apparently malignant tumor, it was killed. Usually, the others were maintained on test until death from "natural causes" occurred. The data, thus obtained, were recorded on graphs such as those shown in Figures 1 through i;. As experience was gained in associating the physical and chemical characteristics of various types of materials with their relative toxicity and with their ability to induce tumors of the skin in C3H mice, certain criteria were developed for selecting optimum conditions for any given biological test. As a general practice, a schedule of application was chosen which would lead to Induction of papillomas in an average time of not less than 12 weeks. The restriction was based on repeated observations that this average `time was unduly variable In experiments in which the severity of exposure resulted In more rapid induction of tumors. API 05811 - 7- ,r the other hand, experimental conditions which involved very long periods of exposure before tumors were induced proved to be even more unsatisfactory than those associated with very short latent periods, because of the increased possibility of inter current infections or other unidentified interferences with the normal growth of the mice. Even when these interruptions resulted in little or no loss of animals, the irregular effects on the growth of the induced tumors made it very difficult to interpret the results of the experiment. Therefore, as a general rule the frequency and severity of the applications were scheduled to produce an average latent period as close to 15 weeks as possible. Vlhen the materials were low in potency, this obviously meant the use of as severe conditions as were compatible with the normal growth and survival of the mice. INTERPRETATION OF DATA OBTAINED In general, under these experimental conditions one of two results has been obtained in response to the applications of any one of a series of materials, i.e., most of the animals which lived long enough developed tumors of the skin at the site of the application, or none of them did so. The outcome of the experiments, therefore, has been essentially unequivocal, and the relative potencies of the different active materials could be expressed in terms of the relative rates of induction of tumors. API 05812 The calculation of the "mean time of appearance of tumors" involved the determination, by the method of least squares, of the l i n e a r function which best fitted the plot of the cumulative frequen cy of tumor response (in probits) versus the time (in weeks) after the first application (shown on the right side of Figures 1 through jj.), The latter coordinate is equivalent to cumulative dosage. It is to be noted that, for these experiments involving repeated applications of carcinogenic materials to the skin, the response varied linearly with the dosage itself, and not with the logarithm of dosage. It is assumed that the exponentially decreasing rate of appearance of tumors usually observed following single or limited numbers of exposures to carcinogens (3) was not seen in these ex periments because the more resistant animals in any given group were given a larger effective dosage than were the more susceptible. Calculation of Cumulative Frequencies of Response The convention is adopted (after Bliss, I4.) that, at the midpoint of the period between two given obser vations, the number of mice bearing tumors is equal to the average of the numbers at the times of the obser vations. At the midpoint of each period during which one or more mice develop their first papilloma, a calcu lation is made of the cumulative percentage of the "effective group of mice" which bear tumors on that date. The "effective group" at a given time is defined as the sum of the number of survivors plus the number of API 05813 - 9mice which died previously after developing tumors of the skin, except that after the appearance of a tumor in the "average" mouse this sum is held constant. The "average" mouse is designated by the median in a simple arithmetic array of the times of appearance of the first tumor in each animal. As an example, the calculations used to obtain the cumulative frequencies of tumor response from the data of Figure 3 are shown in the following table: ( TIME* OP OBSERVATION OP NEW POSITIVES * MIDPOINT OP PERIOD BETWEEN OBSERVATIONS, X A NUMBER OP NEW POSITIVES, R' - T A .B L .b l X TOTAL NUMBER OP MICE BEARING TUMORS,, R + ZR** CUMULATIVE RESPONSE, 2R + (IR + R ) EFFECTIVE NUMBER OF MICE PREQUELU OP RESPONSE (probit 0.5 18 3.08 20 1 JiJS 2k 1 25 1 33x5 3k 2 35^5 36 3 32x5 38 2 1 1.5 2 * 2.5 3 if.o 5 6.5 8 9.0 10 18 3.62 18 3x2i 0 16 If. ,33 1 16 kx!6 16 5.16 API 05815 38.5 11.0 16 5A2 39 2 12 t-S 12.5 16 5.78 k2 1 13 lU.o 16 xi5 h7 2 15 * Number of weeks after first application. * * Z R t|,g0the of ptroetvailounsumboteSrse>?a?iS8.with tumor3 T Calculation of Mean Tima of Appearance of Tumors, x ; Assume that the tumor response in probits, r, is a linear function of the time, x, or r s mx + b Determine constants, m and by method of least squares (5). Normal equations: Zrx s mEx2 + bx r - mix + 96 (Values of x in weeks and r in probits are under lined in preceeding table) 11*85.67- = 104-688,25m + 300,5b . 1*2.28 = 300,5m + 9 b b s 0.960 m = 0 ,112 s slope of dose-response relationship .*. r = .112x + ,960 x = time of 5-prkit response = 36.1 weeks s, standard deviation of mean, s x minus time of l*-probit response s 8.9 weeks. CLASSIFICATION OF EXPERIMENTAL RESULTS ON THE BASIS OF THE HEALTH OF THE ANIMALS In the appraisal of data derived from these experiments, consideration has been given to the following apparent relation ships between the average growth and survival of the mice and the mean "time of appearance" of their epidermal tumors. J API 05816 < - 12 -l. If, in the course of an experiment, an intercurrent I infection overtakes a group of mice at about the time when pap illomas would be expected to appear (on the basis of other compar able experiments), there is usually a delay in the mean time of appearance of tumors. The infection need not be lethal to any of the mice, a change in sign of the slope of the average growth curve frequently being an indication of such a condition (see Figure 1). It is believed that, in many such instances, the mice would probably be unaffected were their resistance not diminished somewhat by/systemic toxic effects induced by the hydrocarbons being applied upon their skin. Thus, untreated con trols may not show any symptoms of the infection. Not infrequently, as the animals recover from such dis turbances, tumors grow to measurable size in a large proportion of the animals in a relatively short time. The slope of the plot of the frequency of response versus time is then exceedingly steep. In other words, the standard deviation of the mean time of appear ance of tumors in the group is much smaller than normal (see Figure 2). Thus, it could be assumed mistakenly, if it were not for the evidence from the weight curves of the presence of uncon trolled variables, that such an experiment had defined the statistical limits of the mean better than another in which such disturbances had not occurred. API 05817 - 13 - 2. In some experiments the cumulative toxic effects of the material applied, or chronic infections, or a combination of both, result in prolonged inhibition of the growth of the mice. number of examples have been observed in which the average weights have increased for a few weeks, but became stationary at a point 10 percent or more below that of adult controls, and then, in many instances, gradually declined (see Figure 3). Thus, whether or not these conditions influence the timing of the initial changes resulting in the neoplasms they may retard the growth of the in duced tumors to visibly discernible size and, accordingly, lengthen the apparent average period of exposure required for the induction of tumors by the material under test. In such experiments, the standard deviation of this mean time is frequently much greater than normal. This somewhat extended discussion of the apparent relation ships between the irregularities in growth due to the toxicity of applied materials and to intercurrent infections, and the rate of induction of tumors, does not relate to the originality of this observation but rather from our recognition of its importance in experiments in which various materials are being compared. Occasion al mention of the phenomenon has appeared in the literature on experimental carcinogenesis (i.e., 6,7 ,8) and, In the field of chemotherapy of cancer, recognition has been given to the necessity of considering the systemic toxicity of a material in evaluating the specificity of its apparent "anticarcinogenic" effects (i.e., 9 ) API 05818 - 34 - Further, the relationships between general health and susceptibility ^"spontaneous" tumors in man are currently receiving attention. Hence, it is believed that, in the experimental assay of the relative potencies of various materials or of the susceptibility of various gpecies or strains of one species to a given material, information on the health and survival of the groups of animals being compared is essential to any quantitative interpretation of the data. The results of all experiments described herein have been fjp examined, therefore, in the light of these relationships. Each one has been classified according to the estimated significance of such effects in the following manner. Class A. Those characterized by normal growth and survival of the mice at least until the time when most of the animals have developed papillomas (see Fig.l).). Class B. Those in which all or most of the mice lived long enough to develop tumors, but in which there was evidence (from average weights) of intercurrent infection at times prior to or during the period when the new growths were appearing. So long as recovery from the infection was prompt, and no obvious skewing of the distribution of the times of appearance of tumors occurred, it was felt that reliable estimates of the mean time of appearance could reasonably be made (see Figure 1). However, when no tumors had appeared prior to the onset of an infection, and when with recovery tumors developed API 05819 - 15 - somewhat precipitously, the reliability of the apparent mean time of appearance was not determin able, and therefore such tests were classified X (see Figure 2), Class X. Those characterized by poor growth (as discussed in Section 2 above) or high rate of mortality. The confidence limits of the apparent mean time for the induction of tumors could not be calculated by stand ard .statistical methods, because of inability to assign proper weight to the effects of the uncon trolled variables. Several types of moderately toxic materials failed to yield experiments of the Class A category. On the other hand, experimental results of the frankly unsatisfactory type (Class X) were usually found to be avoidable through the exercise of appro priate care in the initial choice and the subsequent handling of the mice. 4 REFERENCE STANDARDS j To provide a background of reference which might permit ! a more nearly quantitative interpretation of the mean times of I appearance of tumors induced in C3H mice by certain complex oils and tars under various conditions of exposure, a number of experi ments have been carried out over the past five years employing solutions of the synthetic carcinogens, 3-methylcholanthrene, benzo jY)pyrene, and 7 ,12 -dimethylbenz [}anthracene, in various solvents. It has been found that one of the essential requirements API 05820 f - 16 0l> obtaining reproducible and logically consistent results is the maintenance of the animals in a good state of health as measured by growth curves. Thus, in Table II, the results of Experiments 20$ show the precision of the determination of the mean time, x, when Class B experiments are involved. Similarly, it has been found that even in the case of a complex material, so long as the conditions of the experiments permitted reasonably normal growth of the animals, the repeatability of the determination of the mean time, x, has been quite satisfactory (within 10%). To determine the effect of variations in the level of concentration and in the. frequency of exposure, comparable experiments were conducted with other solutions of methylcholanthrene in benzene, and the results are shown in Table II. The purity of the polycyclic hydrocarbon is obviously an important consideration in obtaining consistent results. The usual chemical criteria are necessary, but not always sufficient, requirements. In one case in particular, the methylcholanthrene, as received, had a sharp melting point, 176-7 c ., and no indication was apparent by its ultraviolet absorption spectrum of any contamination. Yet, on repeated bioassay, it proved to be excessively irritating and much more rapidly carcinogenic than the normal material. Un fortunately the quantity available did not permit deter mination of a satisfactory procedure for purification. Since no further sample with such unusual potency has been obtained, bioassay remains the only way of acertaining that the methylcholanthrene is free of the material responsible for this activity. API 05821 # TABLE II EXPERIMENTS INVOLVING THE APPLICATION OP SOLUTIONS OP METHYLCHOLANTHRENE IN oENZENl UPON THE SKIN OP C3H MICE EXPERIMENT NUMBER CONCENTRATION OP METHYLCHOL- ANTHRENE IN SOLUTION A % by weight) DOSAGE OP SOLUTION PER APPLICATION (raP?)... NUMBER OP APPLICATIONS PER WEEK DOSAGE OF METHYLCHOLANTHRENE PER UNIT AREA1 PER WEEK (mg.) MEAN TIME OF APPEARANCE OF TUMORS, x. (weeks) CLASS CAT 201 0.012 100 3 0.036 72.6 X 202 0.01*1* 100 3 O .132 30.1*3 .8* B 203 0.086 100 3 v0.2$8 29.9 2.2 B 203 0.086 100 3 0.2$8 26.3*3.4 B 20^ 0.172 100 3 0.$l6 I 1$.7*1-6 B 201* 0.172 20 3 0.$16 I6 .1*l.l B 201* 0.172 20 3 0.$16 16.61 .2 B 204 0.172 20 3 0.$16 1$.$1 .1 B 205 0.34S 100 3 1.03S 1 1 .0*1 .9 B 20$ 0.34S 100 2 0.69 IS.1*1.S B 20$ 0.34S 100 1 0.34S 22.I*i2 .7 B 20$ 0.34S 100 3 1 .03$ 14.7 X 206 0.69 100 1 O .69 17.8 X 206 0.69 100 3 1.98 6.7*4.! B 1 Area of skin covered by 100 mg. of solution of methylcholanthrene in benzene. $; fiducial limita. API 05822 TABLE II (page 2) EXPERIMENTS INVOLVING THE APPLICATION OF SOLUTIONS OF MGTHYLCHOLANTHRENE IN BENZENE UPON THE SKIN OF C3H MICE EXPERIMENT NUMBER CONCENTRATION OF METHYLCHOL- ANTHRENE IN SOLUTION by weight) DOSAGE OF SOLUTION PER APPLICATION (mg.) NUMBER OF APPLI CATIONS PER WEEK DOSAGE OF METHYLCHOLANTHKENE PER UNIT AREA1 PER WEEK (mg.) MEAN TIME OF APPEARANCE OF TUMORS, 5c (weeks) CLASSIFI CATION 207 1.15 100 3 3.U5 8.9 X 209 0.517 20 1 '0.517 10.311.9* B 209 0.517 100 1 0.517 16.22.7 B 209 0.517 100 3 1.551 1 0 .1+ X 1 219 0.115 100 3 0.31+5 21+.0 X CD 220 1.03U 100 1 1.03U 9.71.6 B 230 0.23 20 2 0.1+6 10.7-2.3 B 230 0.23 100 2 0.1*6 1 6 .1+1 2 .0 B 21*3 0.15 21+3 0.15 100 3 20 3 o.i+5 0.1+5 17.912.0 A 10.22 .1 B 260 0.207 20 2 0.57U 17.6 X 1 Area of skin covered by 100 mg. of solution of methylchulanthrene in benzene. 5# fiducial limits. API 05823 **" " "* ........... S i l l n 1 ........... I'M-- "V ' t. - 19 - In Table II It Is apparent that variation In the dosage of solution per application from 20 to 100 mg., other factors remaining constant, had no effect on the rate of Induction of tumors by methylcholanthrene In benzene. Similar results have been observed when solutions of benzopyrene In benzene were used as the carcinogenic stimulus. This constancy of potency with variation in the quantity of material applied (or area of skin exposed) also holds for some complex oils and tars, but for others there is significant change of potency with changes in the quantity used for each application. The chemical basis for these differences will be discussed In a later paper, . . The values In column %, Table II, were calculated from the concentration of methylcholanthrene in the solutions and the number of applications per week. It has been determined experi mentally that variations in the concentration of the carcinogen did not affect the area covered by a given weight of the solutions. It was further assumed that the dosage per unit area was not signifi cantly different when 20 or 100 mg. of the same solution were and applied. Comparison of the data in columns 5X& shows a consistent Inverse relationship between the dosage of carcinogen per unit area per week and the mean time of appearance of tumors. The data (omitting Class X experiments) are plotted In Figure 5 together with the curve of the corresponding hyperbolic function (calculated \ by the method of least squares), (* - 3.7) (d + 0.10) = 8 . 2 _ ,, or, 8 2 x = ---- :----- + 3.7 . (d + 0.10 ) I API 05824 ( - 20 - It is interesting to note the similarity between this equation and that calculated from theoretical considerations of the mechanism of carcinogenesis by Iversen and Arley (10), -i-- + T, where k and T are constants. kc0 The constant, 0.]0, in the experimental equation might be inter preted as a correction required by the definite, though very small, probability of the "spontaneous" excitation of a cell of the dorsal skin of a C3H mouse to the rate of proliferation necessary for the development of a new growth. Similar experiments have been carried out with solutions of two other commonly used carcinogenic hydrocarbons, benzopyrene and 7,12-dimethylbenzanthracene. The results are shown in Table III, together with some data on solutions of methylcholanthrene in sol vents other than benzene. As is evident from the data, it has been difficult to obtain satisfactory experiments with benzopyrene on C3H mice. The growth of the animals has tended to be extremely erratic. Even in tests in which the average weight eventually reached 30 g., the curve was, more often than not, interrupted at critical times by sharp downward breaks. In such experiments, recovery from these periods of poor health was sometimes accompanied by the appearance of tumors in a large percentage of the animals in a relatively short time (relationship exemplified in Figure 2). It will also be noted that, even at high levels of concentration, the mean time of appearance of tumors was in no case less than 20 weeks. API 05825 ) ) ) EXPERIMENTS INVOLVING THE APPLICATION OP SOLUTIONS OP SYNTHETIC CARCINOGENS IN VARIOUS SOLVENTS UPON THE SKIN OP C3H MICE DOSAGE OF CARCINOGEN PER UNIT AREA2 PER WEEK (mg.) CONCENTRATION OF CARCINOGEN IN SOLUTION {% by weight) EXPERI MENT CARCIN NUMBER OGEN1 SOLVENT 21$ DMBA Benzene 0.11$ DOSAGE OF SOLUTION PER APPLI CATION (mg.) NUMBER OP APPLI CATIONS PER WEEK 100 3 0.34$ MEAN TIME OF APPEARANCE OP TUMCRS, x (weeks) 1$.63.33 RELATIVE CARCINO GENIC CLASSIFI POTENCY, CATION P MC B 0.19 214 DMBA Benzene 0.34$ 100 244 BP Benzene 0.172 100 244 BP Benzene 0.172 20 2$2 ^ BP Benzene 0.322 20 2$2 BP Benzene 0.322 20 2$2 BP Benzene 0.322 100 282 BP Benzene 0.092 20 264 BP White oil 0.17 $ 20 3 ' 1 .03$ 7.4*1.! 3 0.$16 2$.1 *2 .$ 3 0.$16 ^27.3 1 0.322 ^3$.0 3 0.966 V 23.0 3 0.966 ^ 20.1 3 0.276 3 0. $2$ ^ 2$. 4 B 0.62 B 0.10 X (0.09) X (0.1 8 ) X (0.1 1 ) X (0.13) X X (0.10) 270 BP White oil 0.3$ 20 3 1.0$ ^21.$ X (0.12) 28$ BP Cotton- o.$o 20 3 l.$0 X seed oil 286 BP Cotton- 1.00 20 3 seed oil 281 MC Cottonseed 0.042 100 3 3.00 0.126 /v/20.4 ^41.0 B(X) B (0.13) 0.047 1 Carcinogens are 7,12-dimethylbenz (a] anthracene,DMBAjbenzofa] pyrene, EP;3jnethylchlanthrene, MC. 2 Area of skin covered by 100 mg. of solution of methylcholanthrene in benzene. 3 fiducial limits. I API 05826 - 2.2 - ttsS OF R3FSP.EHCE STANDARDS AS A SCALE OF RELATIVE POTENCY OF CQMPT.mr TARS AND OILS Various methods of expressing the relative potencies of the pure polycyclic carcinogens and complex materials have been reviewed by Badger (11). Many of these are simply rough classifi cations based upon a particular investigator *s personal experience. The Iball index (12) and a recent adaptation (13) attempted a more quantitative approach, involving the use of the reciprocal of the mean time of appearance of tumors. Such methods are useful in com paring results obtained in a given laboratory using a standardized and inflexible schedule of application of various materials. How ever, experimental techniques differ significantly from one research organization to the next, thus rendering the comparison of data on this basis rather difficult. Further, as was pointed out previously in this article, it was sometimes necessary in the current investigation to change either or both the frequency and severity of exposure to obtain satisfactory conditions for the measurement of the potency of a sample while minimizing its toxic side effects. Hence, a scale of relative potencies which would permit such flexibility in experimental design was needed. The use of the data on relative rates of induction of tumors by the pure carcinogens, methylcholanthrene and benzopyrene, as external standards of reference, provided a practical solution to this problem. The latter is one of the important contributors to the potency of cracked residua (11+); the former has not been identified in complex tars and oils but might be regarded as a API 05827 Tj - 23 - | representative of the l|-ringed carcinogens supposed to be present (1$,16). Thus, the mean time of appearance of tumors induced by I some complex tar or oil may be translated into a value of Relative ! carcinogenic Potency, PMC ( potency as compared to methylchol- anthrene), by use of Figure 5 The dosage of methylcholanthrene j corresponding to the given mean time is either read off the graph or calculated from the equation, d = - J b i __ - 0.10 . x - 3.7 Then, the value of the relative potency, PM(,, for the complex material in question is PMC r d, for an experiment involving one application each week, s , for an experiment involving two applications each week, 2 = d, for an experiment involving three applications each week; 3 Thus, by comparing the results of tests on complex mater ials with those involving methylcholanthrene as an external standard of reference, one may obtain the relative carcinogenic potencies of the different materials, even though the experiments may have in volved differing frequencies of applications. It should be noted, however, that the description of the potency of a complex oil can not be considered complete until it has been tested at two apprec iably different dosages, i.e., 20 mg. and 100 mg. per application, unless negative results were obtained at the higher level of dosage in a test in which the mice survived and grew at a normal rate. - 2k - Thus, a given oil may be much more potent than another under s ev ere conditions of exposure, but the former may actually have a lower potency than the latter if the comparison is carried out I under mild conditions. Chemical research on such oils has led to an understanding of this apparent paradox, and the matter will be dealt with in a later publication. The technique of relating results on complex materials to those on standard synthetic carcinogens should facilitate compar ison of results from different laboratories. The synthetic com pound chosen should, of course, be comparable to the type of carcin ogens supposed to be present in the complex mixtures. Under appro priate conditions of dosage and exposure, it must be capable of inducing tumors at least as rapidly as the most potent of the samples to be tested. Thus, benzopyrene has not proved to be a suit able standard for work with C3H mice, since it will not induce tumors sufficiently rapidly even at high levels of concentration. 7,12 -Dimethylbenz{a3 anthracene on the other hand, is very sat isfactory from this standpoint, but has the disadvantage of requir ing special care in handling to prevent its oxidation to the non I carcinogenic 7 ,12-photooxide (7) It should be reemphasized that this consideration of a quantitative scale of potencies applies only to the results of experiments which may be classified as satisfactory (A and B by our criteria). Those in our Class X have only a limited value. If positive results are obtained, the material may be classified qualitatively as carcinogenic; if survival is good and most of the mice eventually develop tumors (X classification based on certain irregularities in growth) the estimation of the apparent potency may be useful as a minimum value. " API 05829 - 25 - RELATIVE POTENCIES OF POLYCYCLIC AROMATIC COMPOUNDS The data available to the writers on experiments involving repeated applications of solutions of various polycyclic carcinogens in solvents such as benzene, acetone, and white mineral oil of high viscosity, upon the skin of mice, indicate that the average rate of induction of papillomas is independent of the quantity of sol ution used in each application so long as it exceeds 10 mg. The controlling variables are the frequency of application and the dosage of carcinogen per unit area of the skin. The latter will depend upon the concentration of carcinogen in the solution applied and the relative spreading coefficient of the solution. Solutions of polycyclic hydrocarbons in different solvents have different capacities for spreading over the skin of a mouse because of variations in such physical properties as volatility, interfacial tension, and viscosity. Since much of the previous experimental testing of polycyclic compounds for their carcinogen icity to the skin has been carried out with solutions in benzene, this solvent was used for the reference standards described herein. An intensive investigation of the effects of variation of the com position of the solvent has also been carried out and will be described in a later paper. By comparing data on the rate of induction of tumors by solutions of known concentration of various aromatic hydrocarbons in benzene under suitable experimental conditions, with such data as those shown in Table II and Figure 5 it should be possible to arrive at logical estimates of the relative carcinogenic potencies API 05830 - 26 - of the different compounds. By inference, the relative potency of methylcholanthrene on the scale would be 100, Corresponding numerical values for benzopyrene and 7,12-dimethylbenzanthracene may then be derived from the data of Table III, Thus, since 0,3^5 percent 7, 12.dimethylbenzanthracene produced tumors at the same rate as 0,62 percent methylcholanthrene under comparable conditions, the relative potency of this dimethyl- derivative is ?Mr , _12_ X 0.62 : 160 M0 0.3W prom Experiment 215, the relative potency, PMC, for 7, 12-dimethyl- benzanthracene is 1-65 T&e higher value is preferred at the present time, since, if autoxidation of the hydrocarbon caused any reduction in activity, it might reasonably be assumed that the effect would have been greater in Experiment 215. Similarly, the relative potency of benzopyrene on this scale is 57 - 65 (Experiments kh, 252, and 282), The data from Experiment 252 involving three applications each week cannot be used since apparently the maximum effective dosage of this carcin ogen had been reached at a lower level. Thus, as a general rule the determination of the relative potency of any given compound should be based upon the results of experiments at at least two different concentrations, which yielded significantly different mean times of appearance of tumors under otherwise comparable conditions. At least one and preferably both of these should meet our classification A or B from the standpoint of growth and survival. API 05831 - 27 - LITERATURE CITED (1) Pullinger, B. D . , J. Path. Bact., 301 (191+3). (2) Cramer, W. and Stowell, R. E., J. Nat. Cancer Inst., Zj 369 (X9U2). (3) Bryan, W.R. and Shimkin, M.B., J.Nat.Cancer Inst,, 1, 807 (19U-1) (U) Bliss (5) Whittaker, E. and Robinson, G., The Calculus of Observation, London, Blackie and Sons, Ltd,, 1921+, p. 209. (6) Watson, A. P, and mellanby, E, Brit. J. Exp. Path,, 11,267 0-930)* 4 (7) Bradbury, J. T., Bachmann, W.E., and Lewisohn, M.G., Cancer Research, 1^ 685 (191+1). (8) Cook, J.W. and Ksnnaway, E.L. (9) (10) Iversen, S. and Arley, N., Acta Pathol. Microbiol. Scand., 21, 1 (1950). (11) Badger, G.M., Brit. J. dancer, 2, 309 (191+8). (12) Iball, J., Am. J. Cancer, 188 (1939). (13) Blanding, P.H., King, W.H., Priestley, W., and Rehner, J., Arch, Ind. Hyg. Occ. Med., 1^, 335 (1951). (114.) Tye, R., Graf, M.J., and Horton, A.W., (Anal. Chem., 1951+) (15) Berenblum, I, and Schoental, R., Brit. J. Cancer, 1, 157 (191+7). (16) Fischer, H.G.M., Priestley, W., Eby, L.T., Wanless, G.G., and Rehner, J., Arch. Ind. Hyg. Occ. Med., j, 315 (1951). API 05832 EXPERIMENTS INVOLVING THE APPLICATION OP MATERIALS UPON THE SKIN OF C3H MICE LEGEND Symbols used In connection with average weight curves: ^ 7 Time of death (from disease) of Lt h tumor-free mouse, i Time of death (from disease) of i th tumor-bearing mouse. ^ o r T Time mouse killed. * Painting discontinued for 1 week at each indicated time. Gross and microscopic pathology; m Time of appearance of first papilloma in Lth mouse. When this symbol is first used for a given mouse after its death, the presence of non-invading carcinoma (intra-epithelial) or small areas of benign neoplasm was determinable only by hlstopathology, 0 Time of appearance of gross changes indicative of malig nancy in tumor of ith mouse. Diagnosis confirmed except in cases where no tissue section available. When this symbol is first used for a given mouse after its death, the Invasive malignancy of the tumor was determinable only by hlstopathology. fn o r ^ No tissue section available for hlstopathology because of extensive post-mortem decomposition or cannibalism. API 05833 \ ) ) ) 9E8S0 IdV I (. < i ) ) FIGURE 5 EXPERIMENTS INVOLVING THE APPLICATION OF' SOLUTIONS OP METHYLCHOLk NTHKENE IN bEi-UENE  Copy From D. V. Stroop to Recipients of the Kettering Manuscript With Copies ( of Summary for Attachment to Manuscript Which Had Been Fully Prepared for Shipment Prior to Receipt of Summary_______________________ May 10, 195I+ UNIVERSITY OF CINCINNATI The Kettering Laboratory College of Medicine - Eden Avenue Cincinnati 19, Ohio May 8, 195^ Mr. D. V. Stroop, Director Technical Service Division American Petroleum Institute 50 West 50th Street Nev York 17, Nev York Dear Mr. Stroop: Attached are the stencils of the summary of the paper on "Carcinogenesis of the Skin". These pages should be inserted in the manuscript just before the bibliography. Thanks again for your cooperation in these matters. Sincerely yours, /s/ Wes A. Wesley Horton AMH:mjg cc: Dr. R. E. Eckardt Enc. - 2? SUMMARY Repeated applications of solutions of carcinogenic hydro carbons upon the dorsal skin of male mice of the C3H strain under suitable experimental conditions result in the induction of epithel iomas in essentially all of the exposed animals. The average period of exposure (x) to solutions of 3-methyleholanthrene in benzene required to induce papillomas of the skin of these mice varies inversely with the dosage (d) of the carcinogen per unit area per week, according to the following relationship: d + 0.10 For experiments involving combinations of synthetic carcinogens with non-accelerating solvents, the mean time of induction of tumors is independent of the total area of skin exposed, so long as a certain minimum is exceeded (about 10 sq. mm.). The good health of the mice, as measured by their growth, has proved to be an essential factor in obtaining reproducible results. A quantitative scale for expressing the relative carcino genic potencies of various synthetic carcinogens and complex tars and oils has been derived, using the experiments on solutions of methylcholanthrene in benzene as reference standards. Utilizing the relationship between the dosage of methylcholanthrene and the rate of induction of tumors shown above, a value for the relative potency, Pj^q is obtained. The numerical value of the of a tar or oil is equal to the level of concentration, in percent by weight, of the solution of methylcholanthrene in benzene which would induce papillomas at the same rate as the material in question under API 05842 - 28 comparable experimental conditions. The potenc ies, PMC, of pure compounds, relative to that of methylcholanthrene, taken as 100, may be derived in a similar manner. Prom the data in this paper, the relative potencies of 7*12-dimethylbenz[a] ) anthracene and benzo [a]pyrene are 180 and 60, respectively. API 05843 1 T I file: Omlversity of Cincinnati Expenditure* toy 2^, 1951 Dr. Robert A. Kebo* University of Cincinnati 13 Kettering Laboratory Cincinnati 19, Ohio Dear Dr. Ketaoe: A 2hla ill aefcnoeledg* the receipt of the invoice covering the Assrieaa Petrleo Inatitute'a abare of eigen** for iaveetlgmtiom of fluorine aai fluorine conpwmita for the year eH eg Anee JO, 1995 He plan to iaeoe a chao in w a n t of the lavle* on or about Juaa 2fc. Tory truly your*, DVS:> ee: Mr. laeey Walher 1 rH(UTTERING UIORATORT cottESEOP MEDICINE-- EDU AVENUE CINCINNATI It. OHIO UNIVERSITY OF CINCINNATI DEPARTMENT OP PREVENTIVE MEDICINE AND INDUSTRIAL HEALTH May 17, 1954 CAILE ADDRESS. KETLAS, CINCINNATI TELEPHONE: CAptTOL I4U - `\ *;^z s Vs. sfc-(. \ UNIVERSITY OF CINCINAI OEPARTMENT OF PREVENTIVE MEDICINE AND TH KETTERINC LABORATORY colLE5EOF HtDICINC-- KOCM AVtNUC CINCINNATI 19. OHIO May 25, 1954 CAELE ADORESE: KETLAI.CINCINNATI TELEPHONE: CAPITOL III! Mr. D. V. Stroop, Director Department of Technical Services American Petroleum Institute 5N0ewWYesotrk50CthitySt2r0eet . Dear Mr. Stroop: I am sending you herewith, for your information, a statement of expen ditures made on behalf of the Am erican Petroleum Institute for the first quarter of 1954. I believe that this statement will be self-explanatory, but if you have any questions or comments, Doctor Kehoe would appreciate your bringing them to his attention. Very truly yours, E. R. Fortlage, Secretary to Dr. Kehoe ef Enc. API 05848 UNIVERSITY OF CINCINNATI KETTERING LABORATORY ACCOUNT OF AMERICAN PETROLEUM INSTITUTE FOR 1 st Quarter 1954_____________ SALARIES (Based on Proportion of Tim e Actually Spent on Project) Direct Sa la rie s................................................ Indirect Salaries Histopathological Preparation___ ___ 4 S S Q ...8 .3 ....... 2 4 2 .0 0 Other Services................................................. 5 3 -7 -.-9 9 ____ 11 Z 7 .Q .-8 2 MISCELLANEOUS EXPENSE Purchase of Anim als.........................................................1 9 5 * 8 0 . ........ Special Laboratory Supplies........................................... ...... 2 9 8 9 0 ____ Travel ___________ _________________________________________ . 7 8 4 * 4 9 - ____ Overhead (Proportion of Heat, Gas, Electricity, Steam, Telephone, General Laboratory Supplies, Postage, Annuities, Pensions, Maintenance, e tc...Jc L Q 4 *2 4 ____ - 3 J.~ 5 2 '5 .4 3 Balance Available for Further W ork TOTAL ......... at End of_____________________________________ Balance Due Kettering Laboratory at End of-- 1955------------------- Receipts Su b Aval lab 1 .................... ' Expenditures .... l& t -- ftu & r .t f lP ...195-4_.*--*...*--a _ 7,*35.93 58.283.00 30,847.07 1 5 ,O S *.25 Balance Available for Further W ork at End of-- 15,752.82 Balance Due Kettering Laboratory *t End of____________ ________________ FORM 1230 KIT.LAB.SM-- H-S* 15,09*.25 API 05849 ( Mimeographed Copies Prom D. V. Stroop To Those Indicated In L e t t e r . _________________________________June 9. 1954 ESSO LABORATORIES Standard Oil Development Company P. 0. Box 51 Linden, N. J. Medical Research Division flobert E. Eckardt, PH.D., M.D. Director May 28, 1954 Mr. D. V. Stroop American Petroleum Institute 50 West 50th Street New York 17, New York Dear Mr. Stroop: I am attaching the minutes and attachments to the minutes of the April 29th meeting of the Committee on the Carcinogenic Action of Mineral Oils of the Medical Research Council of Great Britain. This was forwarded to me by Dr. Newqulst who received it from Colonel Auld. It would be my suggestion that this material be duplicated for circulation to the members of the Subcommittee on Carcinogenicity and to the members of the Medical Advisory Committee. In addition, I would appreciate it if a copy of this material is forwarded to Dr. Horton at the Kettering Laboratory. REE:ilk Ends. Sincerely yours, /s/ R. E. Eokardt R. E. ECKARDT, M. D. I $ MEDICAL RESEARCH COUNCIL For Specified (Normal Committee') Circulation M R C .54/277 CAMO.Ag.15 Committee on the Carcinogenic Action of Mineral Oils AGENDA for the Fifteenth Meeting to be held at 38 Old Queen Street, Westminster, S.W.l on Thursday, 29th April, 1954 at 4.00 p.m. 1. Minutes of the last meeting. (MRC.53/621 - already circulated) 2. Report by Professor J. W. Cook. (CAMO.54/1 - Enclosure A) 3. Report by Professor P. Morton. (To be circulated) 4. Other reports. 5. Future plans. (CAM0.5V2 - Enclosure B) (CAM0.5V3 - Enclosure C) 6. Any other business. 7. Date of next meeting. API 05851 f f MEDICAL RESEARCH COUNCIL For Specified (Normal Committee) Circulation .1 MRC.54/275 CAMO.54/1 Committee on the Carcinogenic Action of Mineral Oils PROGRESS REPORT NO. X from Dr. W. Carruthers and Professor J. W. Cook, Glasgow University We have received from Professor Morton two of the fractions prepared at Birmingham by distillation of the Kuwait No. 1 side stream. These are the fractions designated KiiSob (b.p. 357i-36oo) and fySgb (b.p. 377^-380) of which we had 3 75J! . and 6 .32^ . respectively. Fraction KUSgb was a mobile orange-yellow oil; KhSrb contained some waxy material which was filtered off before pro- 0 ceeding further. 'Aromatic- extracts' of these fractions were prepared by shaking them with 5# aqueous acetone as previously arranged. This yielded 940 c.c. of a dark coloured extract from K^Sob (25$), and 1430 c.c. from K i ^ b (23#)* These were washed with dilute acid and alkali (which removed only a trace of material) and the absorption spectra of distilled samples examined. The spectra showed the characteristics noted in previous Kuwait fractions. Further work has been entirely confined to these extracts. The almost colourless rafinnates have been stored. It is, perhaps, of interest to note that the separation with acetone into 'aromatic' and 'non-aromatic' fractions is by no means clear cut. Chromatography of the 'aromatic' part of i Q ^ b on silica gel showed that it contained about 25# of material of refrac tive index below 1.5000 and thus to be regarded as largely non aromatic in character. Similarly the 'non-aromatic' part was shown to contain 20# of material with refractive index above 1 .5000. The extracts were heated with maleic anhydride at 100 for several hours. This treatment was designed to remove anthracene type compounds from those of other structure present in the oil through their alkali soluble maleic anhydride adducts. 'Anthracene fractions' were then recovered from the separated adducts by distillation over sodium hydroxide. Thus, from the aromatic extract of K4S2b 3*8 g. of an 'anthracene fraction' was obtained, b.p. 175-1800/0.4 mm. This deposited a small amount of crystalline material which, after purification, formed pale yellow plates, m.p. 2350. The ultra-violet absorption spectrum showed it was a derivative of anthracene and elementary analysis suggested the formula C15H24. The quinone melted at 163-1650. It was identified as 2 :7-dimethyIanthracene by mixed melting point determination (literature m.p. 2410 cor.; quinone, m.p. 170. cor.). The remaining oily part of the 'anthracene fraction' was chromatographed, but no further crystals were obtained. The fractions showed well defined anthracene-type spectra. API 05852  - 2- A The adduct from Ki^Sgb similarly yielded 11 g. of a dark oil b.p. 190-2100/ 2 mm* From this three crystalline hydrocarbons were isolated by a combination of low temperature crystallisation and chromatography. The first formed pale yellow plates, m.p. 3 4 4 2450. Its absorption spectrum confirmed that it was a derivative of anthracene, and analysis indicated a molecular formula C]_6Hi 4, Ci7Hi6> C18H16 or CigHi8 Oxidation with chromic acid in acetic acid yielded a quinone, m.p. 236-237 Comparison of the compound with the known di- and tri-methylanthracenes suggests that it may be.2 :3 ;6- trimethylanfehracene (literature m.p, 255 orj quinone, m.p. 24o cor.). We hope to confirm this by mixed melting point. Another compound formed colourless blades, m.p. 169-1700. The absorption spectrum again indicated an anthracene structure. The quinone, obtained as above, had m.p. 169-170. Analyses of the two compounds suggested a trimethylanthracene, or its equivalent, as the most likely structure. The third hydrocarbon from the 'anthracene fraction' of K4Sgb crystallised as pale yellow blades, m.p. 224-225. Elementary analysis is in agreement with the formulae C]_gH]_g or C19H18, and the absorption spectrum confirmed the anthracene structure. Oxidation yielded a quinone m.p. 225-226. Neither of these hydrocarbons corresponds to any known anthracene derivative. Further efforts are being made to identify them. The great bulk of the aromatic extract in each case remained unattacked by the maleic anhydride and was recovered. It was shown by its absorption spectrum to be free of anthracene com pounds. It was treated with picric acid in ethanol, and thereby separated into a picrate-forming part and a (greater) part which did not form a crystalline picrate under these conditions. By this means about 756 of Ki;S2b (30# of the aromatic extract) and 5# of Kj+Sgb (25# of the extract) have been concentrated through their crystalline picrates. Present work is centred on the further subdivision of these picrate forming fractions by chromatography, fractional distillation and other means. The non-picrate-forming parts of the aromatic extracts have been set aside for the time being. Mention was made in a previous report of experiments being done with the residue K5 from distillation of the original Kuwait crude. These experiments were pursued at some length, but gave very unpromising results. Some idea of the complexity of this material can be gathered from the fact that small fractions prepared by distillation, extraction and chromatography, and representing 1 part in io5 of the original material gave absorption spectra which were completely devoid of any characteristic features. In view of this, and since this material was found in any case to be non-carcinogenic, further chemical investigation of it has been discontinued. API 05853 r  I MEDICAL RESEARCH COUNCIL pnv Specified (Normal Tv^mrnlttee) Circulation MRC.54/284 CAM0.5V2 Committee on the Carcinogenic Action of Mineral Oils STATISTICAL ANALYSIS OF 1952-3 EXPERIMENTS AT BIRMINGHAM by Dr. J. 0. Irwin Experiments with 21 treatments were carried out with both mice and rabbits. The treatments consisted of applications with the 15 sub-fractions separated under the supervision of Professor Morton and shown in Table 1 of the first report of the Committee (CAMO.53/1) and the 6 further substances:- 9:10-dimethyl-1 :2-benzanthracene, 0.2# and 0.05# in liquid paraffin (D2 and D) Methylcolanthrene 0.2# and 0.05# in liquid paraffin (M2 and M|) Kuwait fraction 4, neat and diluted 4 times with liquid paraffin (K4 ) The 15 were sub-fractions of Kuwait 4, Lagunillas 4 and Oklahoma 4 and are designated fy.l* k 4 . 2 `*,k 4 . 5 ; L4.1' l 4 . 2 ,,,l 4.5j 4.l' 4.2**,04.5 Experiments with Mice The 21 specimens were each tested on 50 mice; the results are shown in Table 1. They show that the mice responded only very slightly to the oil fractions in coapsrlson with their response to the carcinogenic hydrocarbons. However, the higher Lagunillas sub fractions show a significant increase in carcinogenicity over the lower. In general the fractions show a very slightly lower average carcinogenicity than that obtained in the previous year, when the average values were:- Expectation of Tumourless Life (limited to 52 weeks) Date when 10# of survivors had tumours Total number of Tumours *4 50.2 + 1.3 not reached 2 L4 49.O + 1.7 31.8 + 1.2 3 O4 49.9 + 1 . 4 46.5 + 4.3 2 but the difference might well be due to the mice. As m 1951-2 the order of carcinogenicity Is L>K>0; however, L is more distinctly differentiated from K and 0 in the 1952-3 than in the 1951-2 results API 05854 t t - 2 - b for mice. Whild Kiio appears to be the most carcinogenic Kuwait sub-fraction, L ana 0 show a slight increase of concentration in the higher boiling fractions. From the data of Table 1, estimates may be made of the relative carcinogenic potencies of D and M. These are as follows: Estimated Potency (D/'M) (1) From E.T.L. (2) From 10# dates (3) From total number of tumours 10.5 6.5 4.8 The errors of these estimates cannot be very precisely calculated, but the fiducial limits (P - 0.95) are about 50-200# for (1) and (2) and about 25-400# for (3). Thus we can conclude that, Judged by the mice used, D is about 7 times as potent as M. Experiment with Rabbits The design of this experiment is shown in Table 2. It is of the same type as that employed in 1951-2, a Youden square. There are 21 groups of 5 rabbits. In each group, each rabbit was painted with the same 5 fractions, one on each of the sites I, II, III, IV, and VII (see diagram in Appendix I to the First Report of the Committee*). Each of the 21 sub-fractions or standards was painted on each site in one and only one group. Each fraction was used in exactly 5 groupa and every pair of fractions was used in Just 1 group. The -percentage of each group of 5 animals which got tumours on each site was calculated. The method of analysis was to transform percentages to degrees by means of the angular transformation**(which approximately equalises variance at different levels) and then to perform an analysis of variance of the results. Table 3 shows the mean values of the transformed variables and the probabilities of getting a tumour, estimated for each fraction and standard. The general order of carcinogenicity is again L>K>0. The 5 sub-fractions 0 show no significant differences among themselves the K sub-fractions show a maximum at K4 0 } the L sub-fractions . increase in carcinogenicity up to I o . `The residue is in each case less carcinogenic than the original Ka, the value for which is in agreement with that obtained in 1951-2. The probability value obtained for in 1951-2 was 0.473 and none of the sub-fractions * Sites I and II, Ears, outer surface near base. Sites III and IV, Lateral Thoracic. Site VII Interscapular. . ** The-actual transformation used was y * sin-l^||^j. where r is the .number of positive responses out of n. In this case n5 and r is the number of anlaals out 5 which got tumours. This is an improve ment on ysin"l ( u s e d for the 1951-2 data) when n.is small, because it removes bias in the mean, and more nearly equalises variance at"different levels. A.PI 05855 * - 3- B in 1952-3 are significantly more carcinogenic than this. tractions on the other hand suggest a slight increase of carcinogenicity compared with O4 in 1951-2 . The 0 sub From the data of Table 3/ the standard D can be estimated to be 9 times as potent as M with fiducial limits (P = 0.95) of 25- HQ0%. This is in agreement with the value found from mice. The analysis of variance is shown in Table 4. There are significant differences in response between the different sites of application, and between the different groups of rabbits. Table 5 shows the number of tumours out of 105 possible on the 5 different sites. The results agree with those of 1951-2 in showing that the greatest number of tumours is on site VII (middle of the back nearest the ears) and that sites I and II (ears) come next in order. CowHgftrlaon of Results with Mice and Rabbits The relative potencies of the standards D and M do not differ significantly in.-mice and in rabbits. The sub-fractions on the other hand are much less potent in mice than in rabbits, relative to the two standards. For example, L4 0 appears to be very nearly equal in potency to m , or say D]/j 4 , 5 Judged by the mice experi ments, but in the rabbit experiments it is about equal in potency to Di- Nevertheless the selves seems much the same by Lj, ,, L4 2 L4 .3* l 4.4* than these in t h e 'mice and experiments. relation between the sub-fractions them in mice and rabbits. This is exemplified though L4 = appears a little more potent definitely less potent in the rabbit As Judged by the limits of error of the potency determina tions of M and D made above, 20 rabbits painted on 5 sites give about the same amount of information as 50 mice. This might suggest that it was more profitable to use mice, but the response of the mice to the sub-fractions is too slight to make differences of carcinogenic effect easily ascertainable. For example, the difference between L4 o and L4 4 in rabbits is about 2.4 times its standard error, in midi thereHii no significant differentiation at all. API 05856  I - 4- B Table 1 - Experiments with mice at Birmingham 1952-3 Two pure carcinogens and one crude (at two dilutions each) and fifteen derived fractions 1 Standard, crude or fraction Expectation of Tumourless Life (limited to 52 weeks) Mo mJ d2 D *4 *4.1 K4.1 *4.2 K4.3 *4.4 *4.5 L4.i l 4.2 L4.3 l 4.4 l 4.5 4.1 4.2 4.3 4.4 33.7 + 1.8 48.3 + 1.4 13.9 + 0.4 24.8 + 1.6 52 52 52 51.4 + 0.6 49.4 + 1.3 52 52 52 50.0 + 1.2 47.3 2.2 48.0 + 2.2 44.6 + 2.5 52 52 51.8 + 0.3 51.4 + 0.6 51.3 0.7 -- -- - -- -- Date when 10# of survivors had txacrars and standard error Total number of tumours 20.9 + 0.9 23 37*1+0.7 6 11.4 + 0.8 38 17.1 + 0.6 25 not reached 0 not reached 0 not reached 0 not reached 1 44.9 + 1.3 4 not reached 0 not reached 0 not reached 0 40.5 + 8.9 3 37.4 + 4.3 3 34.1 + 0.8 3 23.9 + 6.4 7 not reached 0 not reached 0 not reached 1 not reached 1 not reached 1 1 in O 1 API 05857  * -5 - B Table 2 - Design of experiment on 21 fractions (1952-3 teats) Rabbits 1 -5 6-10 11 - 15 16 - 20 21 - 25 26 - 30 31 - 35 36 - 40 41 - 45 46 - 50 51 - 55 56 - 60 61 - 65 66 - 70 71 - 75 76 - 80 81 - 85 86 - 90 91 - 96 97 - 100 101 - 105 * 4 .1 * 4 .2 * 4 .3 *4.4 *4.5 = * - * = u = m = S a = L4 . 3 b = K4 . 2 c = D2 d = L4 . 5 e = 04 i 5 <D II in 0 I II III . ... 11 IV d P n u e q 0 a f r P b S s q h t r dc 1 u 3 e J a t f k b u S 1 m dc a b h 1 n 0 e f dc J k P S e 1 q h f m r 1 S n 3 J h 0 t k 1 P u 1 i q a m k r b n 1 s c 0 m t t Key to fractions a - u 4.1 " 0 O4 . 2 t O4 . 3 r O4 . 4 = J L4 . I f L4 . 2 - 8 1*4.3 * a L4 . 4 = 1 L4 . 5 = d f = L4 . 1 g = *4.5 h = Mg i J = O4 . 4 k - K4 t l 1 = l 4.4 m = K4 . 4 n = K4 0 " 4.1 V II a b dc e f S h 1 J k 1 m n 0 P q r s t u Mg = h Mi = P Da * c Di - q *4 = n *4 .J. = 1 P - Ri q * Dt r = 4.3 s = l4.2 t = 04.2 u = k4.3 API 05858 * -6 Table 3 - Experiments with rabbits at Birmingham 1952-3 Two pure carcinogens and one crude (at two dilutions each) and fifteen derived fractions Mean Values corrected for group differences Fraction or Standard Transformed Variable () Probability of tumour l Total number of Tumours (out of 25 possible) K4.1 *4.2 k 4.3 K4.4 k4 .5 L4.2 L4.3 l 4.4 l 4.5 i--1 36.52 44.79 30.19 30.41 21.77 ' 32.48 37.57 49-79 54.43 29.60 .332 .496 .216 .220 .158 .257 353 521 .686 .206 4.1 4.2 4.3 04.4 4.5 23.55 20.27 26.61 25.21 21.13 .109 .063 .156 .134 .074 2 d2 *4 H U S.E. one mean 23.55 11.92 60.62 34.63 37.00 19.76 4.96 .109 0 .798 .296 .342 .056 Difference between two means 7 .0 1 i1 I t 8 13 7 7 2 7 9 14 17 5 2 3 3 3 2 3 0 20 7 9 2  -7 - B Table 4 - Experiment with rabbits on fifteen sub fractions and six "standards" (.1952-3) Analysis of variance on angular transformation of proportion of tumours out of 5 Sums of Degrees of Squares Freedom Mean Square Variance Ratio Sites 4020.09 4 groups of rabbits 4917.58 20 Fractions 15358.24 20 Error 6678.57 60 Total 30974.47 104 1005.02 245.88 767.91 I I I .31 9.03 ( P < 0 .001) 2.21 (.05<* P (.01) 6.90 (P<0.001) i 1 Table 5 - Experiment with rabbits on fifteen sub fractions and six "standards" (1952-3) Number of tumours out of 105 on the five l1 different sites Site I II III IV , VII Number of tumours 25 24 23 22 51 API 05860 r 3 i MEDICAL RESEARCH COUNCIL Specified (Normal ittee) Circulation MRC.5V286 CAM0.5V3. Committee on the Carcinogenic Action of Mineral Oils PROGRESS REPORT NO. X. (APRIL 1954) by D. L. Woodhouse and J. R. Squire TESTS CARRIED OUT BY D. L. WOODHOUSE 1953-54 The tests have been carried out satisfactorily for 32 weeks, and the observations of tumours, etc., are recorded in the Tables for the period of 30 complete weeks of applications. The survival of the oiice has been somewhat better than in previous years and the weekly weighing has demonstrated that their health has been maintained throughout without any period of epidemic infections. Two rabbits were lost through accidental injury. The fractions have been warmed to 40C for each application to liquify the waxy constituents. This mild treatment does not appear to have affected the properties or appearance. Results. (a) Mice. The mice as usual reacted with many tumours to the 9:10diraethy1-1:2-benzanthracene (DBA) at both concentration levels. Pew tumours are being obtained with the other fractions, Ki^SgA being the most potent at present. (b) Rabbits. All sites have reacted somewhat earlier than in last year's tests of. 1953-54 1952-53 D(2) at 30 weeks sites affected 8/25 sites affected 1/25 :<i*4 at 30 weeks 5/25 1/25 However, many of the fractions prepared by Professor Morton show even greater potencies, and the results have therefore been examined in some detail in that they may help to guide planning for 1954-55 tests. None of the conclusions have as yet been' checked for statistical significance by Dr. Irwin. As these 30 weeks results are preliminary, one may consider the Kjj.S^-8 -K]|St -8A series together. These show that the main tumour- producing compound (s) is concentrated In fractions 4-8 inclusive, the bulk (about 4556) being in fraction 6 and 7 (wnieh make up together about one third of the total volume distilled). API 05861 I - 2 C Solvent extraction with acetone has resulted in a much greater concentration in the aromatic fractions in fractions 4 - 8 (K43-.A - 3A have only similar potencies to LSi.a), The estimated amount present agrees closely with the proportion of aromatics obtained normally, the exceptions being ItyS^A (about double the potency expected) and K4S7A (about half the potency expected). Otherwise there is no need to postulate very marked accelerator or inhibitor action interfering with the biological assay. The potency of the pooled paraffinic residue is low, but unfortunately not zero. In view of the large volume of the fractions made up by this residue (about 80#), it may be that appreciable amounts of carcinogen have been left in this residue. The concentration of carcinogen in K4S6A and K4S8A leads to a potency of 2^ x li x the potency of 0.2# 9 :10-aimethyl-l12- benzanthracene (D(2) ) respectively. The potencies of K4S4A, K4S5A, and K4S7A are similar to that of this standard. Plotting # sites bearing tumours as probits against lOg (time in weeks) the K4S5A graph runs parallel to the D(2) standard. This is interesting since most of the Kuwait fractions so far tested have shown a divergence of slope in this kind of plotting, making precise potency grading difficult. API 05362 t r r a c t ion K ^ 4 Table 1 Mouse Experiment 1Q53-54 (after 30 weeks) v'eeks 20 21 22 23 24 25 26 27 28 29 30 Survivors 38 32 31 30 28 28 27 27 26 24 24 Survivors with tumours 1 1 2 2 2 1 1 1 0 % Survivors New Total dead with tumours Tumours with tumours 3.2 1 3.3 7.1 1 7.1 7.4 3.7 1 3.8 1 4.2 1 - 2 > r fr a c tio n Kij.SlA 20 34 - 21 32 22 31 1 3 .2 1 23 30 1 24 28 2 3 .3 7 .1 1 25 28 2 7 .1 26 27 2 7 .4 27 27 1 3.7 1 28 26 1 3 .8 1 29 24 1 4 .2 1 30 24 0 2 v< Fraction fyS^A 5 45 6 43 7 43 8 43 9 43 1 10 43 1 11 43 1 12 43 1 13 43 1 14 43 1 15 43 2 16 42 2 17 42 2 18 40 2 19 40 2 20 40 2 21 40 2 22 40 2 2 .3 2 .3 2 .3 2 .3 2 .3 2 .3 4 .6 4 .7 4 .7 5 .0 5 .0 5 .0 5 -0 5 .0 1 1 ii i 1 API 05863 C Mouse experiment (Contd.) ;;eeks Survivors Survivors zpciction K4S4A ^23 24 -- w40 25 40 26 40 27 40 28 40 29 30 338! with tumours (Contd. ) 2 2 2 2 2 2 2 3 % Survivors with tumours 5.0 5-0 5.0 5.0 5.0 5.0 5.1 7.9 New Total dead Tumours with tumours 1 Fraction K4S 5A 20 42 21 42 22 42 23 4l 24 39 1 25 39 1 26 38 1 27 38 1 28 38 1 29 38 1 30 38 1 2.5 1 2.5 2.6 2.6 2.6 2.6 2.6 Fraction K4S6A 18 42 19 41 20 41 1 21 41 1 22 40 1 23 40 1 24 39 1 25 39 1 26 38 1 27 37 1 28 26 1 29 36 1 30 36 1 2.4 1 2.4 2.5 2.5^ 2.56 2.56 2.63 2.7 2.77 2.77 2.77 Fraction K4S7A 20 40 21 40 22 40 23 40 1 24 40 1 2.5 1 2.5 API 05864 r T ~D - Mouse experiments (Conta.) Weeks "Fraction 25 26 27 28 29 30 Survivors Survivors KS7 A 40 40 38 34 30 27 (Contd. with ) 1 1 1 1 0 0 tumours % Survivors with tumours 2.5 2.5 2.63 2.9 New Total dead Tumours with tumours 1 1 Fraction K 4S 8A 20 37 21 37 22 37 1 23 36 2 24 35 3 25 35 .'3 26 33 3 27 33 3 28 31 5 29 30 5 30 30 8 2.7 1 5.5 1 8.6 1 8.6 9.9 9.9 16.1 2 16.6 26.6 3 Standard D(2l 5 48 6 43 f7 41 8 41 9 41 10 40 11 40 12 40 13 40 14 40 15 40 16 40 17 40 18 40 19 38 20 28 21 27 22 0 X 2.4 1 k 9.76 3 13 30.8 9 17 42.5 4 23 57.5 6 26 6 5 .O 3 26 65.0 26 65.0 26 65.0 26 65.O 39 97.5 13 37 97.4 2 28 100.0 1 12 27 100.0 13 0 40 Standard D() 6 48 7 8 9 48 48 48 API 05865 1 2 .1 1 -6 - Mouse experiment (Contd.) Weeks Survivors Standard D(j) (Contd.) -- TO---- ---- 48 11 48 12 48 13 48 14 48 15 48 16 48 17 48 18 47 19 47 20 40 21 40 22 39 23 39 24 36 25 24 26 24 27 24 28 24 29 24 30 23 31 5 Survivors with tumours 4 7 11 19 19 20 20 23 31 31 28 28 33 -33 ' 30 18 18 22 22 22 22 3 % Survivors with tumours 8.3 14.6 22.9 39.6 39.6 40.7 40.7 47.9 65.9 65.9 70.0 70.0 84.6 84.6 83.3 75.0 75.0 91.7 91.7 91.7 95.6 60.0 c New Total dead tumours with tumours 3 3 4 8 1 3 8 3 6 1 7 5 7 7 10 22 22 4 22 22 22 22 43 II API 05866 1 ) ) y Table 2 Rabbit Experiment !Q53-5fr Site and Time of Appearance of Tumours (after 30 week3) K4SIA Weeks Site KJ|S2A Weeks Site K4S3A K 4S4A K4S5A K4S6A K 4S 7A Weeks Site Weeks Site Weeks Site Weeks Site Weeks Site K 4S8A Weeks Site 20 II 18 III 27 VII 1 tumour 13 III 13 II XI VII 10 VII 11 III 20 IV 22 I ?0 III 4 tumours 25 VII 27 I 30 III 4 tumours 13 IV 13 III 12 IV 11 VII 11 IV 17 II 13 VII 13 III 17 III 11 IV 20 VII 18 I . 17 IV 17 IV 11 VII 20 VII 16 II 17 IV 17 IV 11 VII 22 I 20 I 18 II 18 III 13 III 1 25 VII 20 III 18 II 20 I 17 II 1 28 I 22 III 20 8 tumours 22 IV 20 I 25 IV I 27 I 18 III 20 III 27 I 20 III 29 I 20 VII 11 tumours 22 II 30 III 10 tumours 22 VII 27 VII 28 I >* 22 IV 28 I d 13 tumoui 22 VII U001 27 II CT 15 tumours O *1 mm Table 2 (Contdi)j K 4S 1 Weeks Site 25 1 1 tumour K^S2 Weeks Site 18 1 20 VII 22 I 22 II 4 tumours Rabbit Experiment 1Q53-54 Site and Time of Appearance of Tumours (after 30 weeks) K.S3 Weeks Site Kij.s4 Weeks Site Ku S5 K^S6 K1*S7 K 4S8 Weeks Site Weeks Site Weeks Site Weeks Site 18 II 25 III 2 tumours 11 I 13 VII 25 IV 3 tumours 27 III 30 I 2 tumours 18 III 20 VII 22 II 25 VII 4 tumours 14 II 17 III 18 VII 18 VII 4 tumours 10 IV 11 II 30 VII 1 3 tumours WeekDs(2)Site WeeksD(iS)ite WeeksK44Site WeeksK lS^iUt)e WeeksKfcSSiPte 1 Ou 17 I 22 VII 10 II 20 VII 1 0 tumours 17 II 18 II 27 VII 30____ II_ 13 VII 18 I 5___ I i _ 2 tumours 3 tumours 18 VII 27 H I * 22 I 22 II 27 III 5 tumours 25 IV 27 VII 8 tumours C API 05868 ) ) -4 'ruble 3 Approximate concentrations and amounts of carcinogen in oil fractions (expressed as 9 :lQ-dimethyl-i :2'-benzanttiracene equivalents) deduced from Dr. Woodhouse's 3 week results of rabbit testa 1953-54. STILL FRACTIONS SOLVENT EXTRACTS I Fraction II Boiling range C K 4S 1 K4S2 k us 3 K 4S 4 350-355 355-360 360-365 365-370 Kis 5 kus6 K 4S 7 V8 Total 370-375 375-380 38O -385 385-390 III Carcinogen concn. g/1 . IV Volume 1. 0.3 8.7 0.6 12.2 0.3 16.4 0.5 15*3 0.3 23.0 0.6 18.1 0.6 27.6 0.5 19.3, 140.7 V Carcinogen amount, g. 2.6 7*3 4*9 7*6 6.9 10.9 16.6 -2zl 66.3 VI Fraction KlfS1A K 4S 2A K 4S3A k^ a Kl|S^A k 4s 6a K^S^A K^Sq A VII Carcinogen concn. g/1 . 0.6 0.6 VIII Volume 1. 2.8 2*3 0.3 3*1 2.0 3*3 2.0 4.6 5*0 3*4 2.0 4.4 3*0 _3_a 27.2 IX Carcinogen amount, g. 1.7 1.4 0.9 6.3 9*2 17*0 ' 8.8 9-9 55.2 - Notes (a ) If assay were perfect and all the carcinogen were transferred to solvent extracts, columns V and IX should be identical. In fact, the correlation coefficient between them r* 0.61, and since n-7 , the probable significance of this is 0.10-0.05. (b) Recovery by totalling solvent extracts (55-2 g.) is 82$. Slight carcinogenicity was found in pooled paraffin residue (about 0.3 g./l. in 113*51 = 34 g.). If this is > added to the calculation recovery is well over 100$. In general, the results in 3 fractions with low potencies should be treated with reserve until the tests are complete. o Ul 00 O) o VO &j u i j msmm file: Medical M isery Conitte - UHIV. Cf CIHCIHHAII Expenditures June 3, 195* Or. Robert a. jfiehoe (diversity Cincinnati an lattario* laboratory ''rrll'tl* of Ma11r1na Cincinnati 19, Chlo Daar Doctor Knboa: Puraunot to uadarstaodlag ranchad by onr lattare of October 2d and Invaiar 5, 19*T, a a ancloaia* tha laatltute'a check la tha m a m of #2,900 papable to tha Univaraity or Cladaaatl an car oootrlbatlan tenari tha far tha tvelvo-ooath period aadlag M a JD, 1959 Vary truly yoara, DVJJn Eacloaure et Oacaga M. mandara, H. D. taoay Valbar NEMO TO NB. HAUSSE: fwi*t cpaadltura la to ba cbar^ri. a lla n t tba Medical Resaarch fund. API 05870 F ila : Medicati dvdjory C oanittao - BIY. Qt CmCUCUTl xpoditure 8 June 23, 1954 Dr. Robart A. Habt Univanity of Cincinnati S Cattarli* Laboratory Collosa at teAlclno Ciacinnati 19, etilo Puriwit to tba aetloo taten at tba Modica! AdTlaory Coaadttaa aaatln m ili ptite ir JO, 1993 ** aaaloalng tba Inatituta'a aback la tba noaat at $9000 pagatola to tba orbar at tba Oalaanity at Cincinnati aa sur eootrtbu- tlon tonar* tba akla pbyolology yrojnet far tba taalaa-aoath period andia< tena JO 1999 Vary truly yuun, Wtoja ce: Oaorgn M. Sanaban, M. D. Lacay Malter MEMO TO MR. WALKER: xpandltun la to ba cbnrgnd agaloat tba Modica! Research fund. API 05871 ,u .* X, ' ^ NIVERS,TY 0F CINCINNATI L DEPARTMENT OF PREVENTIVE MEDICINE AND INDUSTRIAL HEALTH THtKETTERING UNORATORY f COLLEGE OF MEDICINE-- COEN AVENUE CINCINNATI I. OHIO June 26, 1954 CARLE ADDRESS: XETLAS. CINCINNATI TELEPHONE: CAPITOL MIA Mr. D. V. Stroop, D irector Department of Technical Services American Petroleum Institute 50 West 50th Street New York City 20 Dear Mr. Stroop: AI hmeerreiwcaitnh PacektrnoolweulemdgIenswtiittuhteth: anks receipt of the following checks from $5,000. 00 - in support of the investigation of the physiology of the skin for the year ending June 30, 1955; $2, 500.00 - in support of the investigation of fluorides for the year ending June 30, 1955. Very truly yours, ^ - E-. /RC.__F_o_r_t_la_g__e,_ ^_c__r_e_t_a_r_y__to_ Dr. Kehoe ef API 05872 r