Chronic and initiation/promotion skin bioassays of petroleum refinery streams.

Nine refinery streams were tested in both chronic and initiation/promotion (I/P) skin bioassays. In the chronic bioassay, groups of 50 C3H/HeJ mice received twice weekly applications of 50 microl of test article for at least 2 years. In the initiation phase of the I/P bioassay, groups of CD-1 mice received an initiating dose of 50 microl of test article for 5 consecutive days, followed by promotion with 50 microl of phorbol-12-myristate-13-acetate (0.01% w/v in acetone) for 25 weeks. In the promotion phase of the I/P bioassay, CD-1 mice were initiated with 50 microl of 7,12-dimethylbenzanthracene (0.1% w/v in acetone) or acetone, followed by promotion with 50 microl of test article twice weekly for 25 weeks. The most volatile of the streams, sweetened naphtha, and the least volatile, vacuum residuum, were noncarcinogenic in both assays. Middle distillates, with a boiling range of 150 degrees-370 degrees C, demonstrated carcinogenic activity in the chronic bioassay and acted as promoters but not initiators in the I/P bioassay. Untreated mineral oil streams displayed initiating activity and were carcinogenic in the chronic bioassay, presumably due to the presence of polycyclic aromatic hydrocarbons of requisite size and structure. A highly solvent-refined mineral oil stream lacked initiating activity. These results indicate that the I/P bioassay, which takes 6 months to complete, may be a good qualitative predictor of the results of a chronic bioassay, at least for petroleum streams. Furthermore, the I/P bioassay can provide insight into possible mechanisms of tumor development.

Mouse skin painting has long been used to assess the dermal carcinogenic potential of natural and synthetic materials (1)(2)(3)(4). Compounds derived from petroleum were among the early materials tested by mouse skin painting and were important in demonstrating the carcinogenic potential of compounds such as polycyclic aromatic hydrocarbons (PAHs) of specific structure and elucidating the structure-activity relations now recognized for these compounds.
Early investigators (2,5) painted the skin of mice for long periods of time until or after the appearance of tumors. Modern investigators using this technique apply materials for periods approximating the lifetime of the mouse, particularly since promulgation of the FDA and EPA Good Laboratory Practice Regulations (6,7). It was learned that some materials could cause tumors after a few (or single) applications but only after subsequent longerterm administration of substances that were not necessarily active in the absence of the initial material (8,9). The active materials applied first came to be known as initiators, and those applied subsequently for a longer term were termed promoters. In time, a more or less standardized assay was developed to shorten the time required to identify carcinogens and to discriminate between initiators and promoters (10). This assay, known as the initiation/promotion (I/P) bioassay entails the application of a known or suspected tumor initiator, followed by multiple weekly treatments of a known or suspected tumor promoter.
The I/P bioassay is designed to differentiate substances that produce tumors by genotoxic mechanisms, nongenotoxic mechanisms, or both. Initiation is generally agreed to occur when one or a few cells of a tissue have been exposed to a carcinogen at a level sufficient to cause a precancerous, heritable, somatic, genetic event. A number of characteristics define the initiated cell, including a permanent change in its DNA. Promotion is thought to involve a number of changes in an initiated cell, causing the initiated cell to grow faster than the surrounding normal cells and to develop into a visible neoplasm. Initiating and promoting agents alone do not cause neoplasms to appear, or only rarely so, but in succession elicit both benign and malignant neoplasms.
A number of researchers have investigated the carcinogenicity of various petroleum fractions (11)(12)(13)(14)(15). They have documented the initiating potential and carcinogenicity of streams containing PAHs of characteristic structure and molecular weight. They have also shown that some fractions, which do not contain significant concentrations of PAHs, are tumorigenic in long-term assays and are promoters but not initiators in the I/P bioassay.
The American Petroleum Institute (API) assembled a task force of petroleum company scientists and engineers to identify the various streams produced in petroleum refineries for which carcinogenicity data would be most useful. These were tested chronically, and a subset, reported in this paper, was also tested using an I/P protocol to accomplish the following objectives: 1) to investigate the carcinogenic potential of a number of petroleum streams of varying volatility and composition; 2) to identify the active streams as promoters or initiators as defined by the I/P test protocol; 3) to investigate the correlation between results obtained in chronic studies and in I/P studies to evaluate the I/P bioassay as a possible less expensive and shorter alternative to the lifetime bioassay.

Chronic Study
The test articles listed in Table 1 were obtained from the API repository (Experimental Pathology Laboratories, Herndon, Virgnia). These materials were among those that had been selected by a task force of the API to represent petroleum streams produced during petroleum refining. These streams have been characterized for inclusion in the EPA listing of materials covered by Toxic Substances Control Act regulations and given Chemical Abstract Service (CAS) identification numbers. Table 1 is a listing of the streams, their CAS numbers, and boiling ranges. Toluene, ACS grade, and benzo[a]pyrene (BaP) were purchased from Aldrich Chemical Co. (Milwaukee, Wisconsin).
Vacuum residuum was applied as a 50% (w/v) solution in toluene, and catalytic cracked clarified oil was applied as a 1% (w/v) solution in toluene. All other test articles were applied neat. We used BaP, a well-known carcinogenic component of crude petroleum, as the positive control. BaP was applied as a 0.05% (w/v) solution in toluene.
Male C3H/HeJ mice are most frequently used in chronic dermal studies of 1-1-~~-Pa-a petroleum-derived materials (11,13,14) and therefore were used in this study. Mice were acquired from Jackson Laboratories (Bar Harbor, Maine). Mice were 7-9 weeks of age at study initiation. We randomized mice upon receipt and double housed them in suspended stainless-steel cages for a quarantine period and then housed them singly before study initiation. Some groups were quarantined for 2 weeks and held for 1 week, others were quarantined for 1 week and held for 2. Purina certified rodent chow #5002 was provided ad libitum. Deionized tap water was provided ad libitum by bottle. Animal rooms were maintained at 24.5 ± 1.7°C (SD) and relative humidity at 40% ± 10% (SD). A 12-hr light/dark cycle was maintained. We randomly divided 600 mice into groups of 50 mice each. Groups of 50 mice are routinely used in chronic skin painting bioassays. In addition to the test groups, one group received no treatment, and one was administered BaP (0.05% w/v in toluene) as the positive control. Because two samples were diluted with toluene, a separate toluene group was included as a vehicle control. We applied the test articles, BaP, and toluene as 50-pl doses twice weekly for a lifetime, or in the case of three samples, 24 months. These three samples showed significantly high tumor incidences and short latencies to median tumor, so the shorter exposure period did not affect the classification of their carcinogenic activity. All mice were sacrificed by ethyl ether inhalation and exsanguinated by puncturing the posterior vena cava.
We applied all materials to the shaved, intrascapular region of the back. Mice were clipped approximately every other week. Materials were applied with a mechanical pipet with disposable syringe and/or pipet tip and allowed to spread evenly without any intervention.
We performed physical examinations for masses weekly. All mice were subjected to a limited necropsy upon death or sacrifice. We removed application-site skin and fixed it in 10% neutral buffered formalin. Skin specimens were embedded in paraf-fin, sectioned, stained with hematoxylin and eosin, and examined microscopically.
Significance of differences between test and control tumor incidences was analyzed by chi square (16). Fisher's exact test (onetailed) was used when comparing low tumor-incidence groups (five or fewer mice with tumors) to the appropriate control. Final effective number (FEN) is the denominator used in calculating tumor incidence. When time to median tumor was under 60 weeks, we used the number of animals alive at that time as the FEN. When time to median tumor was greater than 60 weeks, the number of animals alive at 60 weeks plus any mice dying with tumors before 60 weeks was used as the FEN. Latency was measured as the time in weeks from initiation of dosing to appearance of the first tumor.

Initiation/Promotion Study
The test articles listed in Table 1 were stored refrigerated in 25-ml glass bottles.
Vacuum residuum was applied as a 50% (w/v) solution in toluene, and the catalytic cracked clarified oil was applied as a 1% (w/v) solution in toluene. All other test articles and toluene were applied neat. DMBA, the usual initiator in promotion investigations, was prepared one time only in acetone at a concentration of 0.1% (w/v). We prepared PMA in acetone every 2 weeks at a concentration of 0.01% (w/v) and stored it protected from light at approximately 0°C.
Male CD-1 mice are routinely used as test animals in I/P studies (12,15) and were used in this investigation to permit comparison with previous studies. Male CD-I mice were acquired from Charles River Breeding Laboratories (Portage, Michigan). Mice were 4-5 weeks old upon receipt and 7-9 weeks old at study initiation. They were determined to be free of viral antibodies upon arrival and were quarantined for at least 2 weeks before study initiation. Animals were housed singly at the time of randomization for study initiation and remained housed singly throughout the study in suspended stainless-wire cages. Rodent chow 5001 (Ralston Purina Co., St. Louis, Missouri) and tap water were provided ad libitum by means of an automatic watering system. The animal room temperatures were maintained at 21.3YC ± 1.6 (SD), and relative humidity at 54.5% ± 13.3 (SD). Animal rooms were illuminated with fluorescent lights and maintained on a 12-hr light/ dark cycle.
In the initiation phase, we randomly divided 360 mice into 12 groups of 30 each. This number of mice per group has been shown to provide satisfactory sensitivity and selectivity in I/P studies (12,15). In addition to the nine test articles, acetone was used as a negative control, and DMBA was the positive control. Because two samples were diluted in toluene, a separate toluene-initiated group was included as a vehicle control. The test articles, acetone, and toluene were applied as five consecutive daily applications of 50 pl. In the positive control group, 50 pl of DMBA (0.1% w/v in acetone) was applied once on the last day of the dosing week. After a 2-week rest period, 50 pl of PMA promoter (0.01% w/v in acetone) was applied to each animal twice weekly for 25 weeks. Physical examinations including observations of dermal masses were performed weekly on all surviving mice until study completion. We sacrificed all mice by intraperitioneal injection of nembutal the week of PMA treatment termination, subjected mice to gross necropsy, and took skin sections for histopathological examination.
In the promotion phase, we randomly divided 630 mice into 21 groups of 30 mice each. Test article groups were the same as for initiation except that neither acetone nor PMA was tested as promoters. A negative control was added (sham-handled), which received no treatments after initiation. All mice were initiated with 50 pl of DMBA (0.1% w/v in acetone). In addition, each mouse in an identical series of test article groups of 30 mice each, except the sham-handled group, was treated with a single 50-pl dose of acetone as a negative initiating agent. After a 2-week rest period, the mice in each DMBA or acetone-initiated group received 50 pl of the test articles or toluene twice a week for Volume 102, Number 1, January 1994 25 weeks (50 applications). We sacrificed all mice by injection of intraperitoneal nembutal, subjected mice to gross necropsy, and took skin sections for histopathological examination. The procedures for skin preparation, dosing, and examinations were the same as those followed in the chronic study.
The outcome measures were 1) time to appearance of first tumor, 2) number of animals with clinically observed tumors, and 3) number of animals with histologically confirmed tumors. Statistical significance of treatment effect in the initiation study was assessed by comparison with the acetone-initiated group. In the promotion study, each sample had its own DMBAand acetone-initiated groups and comparisons were made between them. We used two different statistical methods in the analysis of these data. For the time to first tumor data, a product-limit survival analysis (17) was used to contrast the survival distributions for the desired pairwise group comparisons. To analyze the number of animals with tumors, a one-tailed Fisher's exact test (18) was performed to determine whether the incidence of histologically confirmed tumors in the sample group was higher than its respective control. The type I error rate used for all statistical comparisons was 5%.

Results
Irritation and inflammation at the dosing site were common in the exposed groups, with the most severe signs noted in mice exposed to light catalytic, cracked distillate, hydrodesulfurized middle distillate, and the mixture of 50% straight run, middle distillate/50% light catalytic, cracked distillate. A number of skin tumor types were noted: benign neoplasms classified as fibromas, papillomas, keratoacanthomas, and hemangiomas, and malignant neoplasms including squamous cell carcinomas, fibrosarcomas, and malignant melanomas. The incidence of keratoacanthomas, hemangiomas, and malignant melanomas was extremely low. Only one mouse in each of two different groups developed hemangiomas and only one a keratoacanthoma. Two mice in one group had malignant melanomas. At the time of micropathological examination, most of the tumors were malignant. This is not inconsistent with expectations based on a two-stage model of carcinogenesis suggesting that initiated (first stage) cells have an increasing probability of undergoing a second stage (malignant) change as the mouse ages. The results are given in Table 2.
The samples in this study represent complex mixtures derived from distillation of petroleum and different refining treatments. The most volatile, sweetened naph- Chi-square test, different from sham-handled (p < 0.05). Fisher's exact test, one-tailed, not higher than sham-handled (p < 0.05). +Fisher's exact test, one-tailed, not higher than toluene control (p < 0.05). tha, does not demonstrate significant tumorigenic potential. This is also true for vacuum residuum, a nonvolatile fraction containing high molecular weight compounds including PAHs. Since this stream was diluted 50% in toluene, statistical comparison was made with the toluene solvent control rather than acetone.
BaP, a strong dermal carcinogen, produced tumors in virtually all exposed animals in this study at a concentration of 0.05%. The only petroleum stream that approached it in activity was catalytic, cracked clarified oil, which achieved 100% incidence at a concentration of 1%. Solvent-refined heavy naphthenic distillate lacked carcinogenic activity in the chronic study.
Except for light paraffinic distillate, the rest of the samples can be characterized as middle distillates with boiling ranges in the region of 150°-370°C. These streams appear to be tumorigenic after chronic dermal application under the conditions of this experiment. The skins of mice chronically exposed to the middle distillate samples evidenced severe irritation with inflammation, some erosion, and scab formation.
The tumors seen in the I/P study were overwhelmingly benign squamous cell papillomas and keratoacanthomas. The only malignant tumors, squamous cell carcinomas, occurred in two mice of the group promoted with light catalytic cracked distillate. Table 3 details the tumor incidence and latency after initiation and promotion phases of this assay.
The only samples manifesting initiating activity, as shown in Table 3, were 1% catalytic cracked clarified oil and light paraffinic distillate. The catalytic cracked clari-fied oil was almost as potent as 0.1% DMBA. catalytic cracked clarified oil does not appear to have promoting potential, although this lack of activity may be a result of the low concentration tested (1%). Light paraffinic distillate appears to have promoting as well as initiating potency.
The middle distillate samples (hydrodesulfurized kerosene, hydrodesulfurized middle distillate, light catalytic cracked distillate, and the 50/50 mixture of straight run middle distillate/light catalytic cracked distillate) acted as pure promoters in the I/P assay with no significant initiating potential. Lastly, a heterogeneous trio of samples, sweetened naphtha, vacuum residuum, and solvent-refined heavy naphthenic distillate, failed to demonstrate initiating or promoting activity.
In the initiating phase, there was some inverse correlation between tumor incidence and latency with exceptions such as vacuum residuum, which had a low incidence and low latency. This did not appear to be the case in the promoting phase, however, where the latencies were quite similar for all tested groups developing tumors.
As noted previously, the only tumors occurring in chronic and I/P studies at higher than incidental rates were squamous cell carcinomas, squamous cell papillomas, fibrosarcomas, and keratoacanthomas. Table 4 lists the incidence rates of these tumors in both studies. The bulk of the neoplasms in the chronic study were squamous cell carcinomas with a significant representation of the fibrosarcomas and a relatively small incidence of squamous cell papillomas. As   bNumber of malignant tumors in parentheses.
**Significantly different from acetone control (p < 0.05) using Kaplan-Meier test. study were benign squamous cell papilloover, analysis of the stream revealed no mas and keratoacanthomas, with the known carcinogenic components. The lack exception of two mice with squamous cell of activity may be due to one or both of carcinomas in the groups exposed to light these factors. Vacuum residuum is the least catalytic cracked distillate. Aside from the volatile of the streams and does not begin fibrosarcomas, these are tumors of epiderto distill below 350'C. No analytical data mal origin that might be expected to arise were available on this stream, but it is from dermal treatment with a tumorigen. expected to contain some high molecular weight PAHs. These compounds may not Discussion be carcinogenic, explaining the innocuous Three streams in this study did not show character of the stream. Alternatively, any statistically significant carcinogenic activity active materials in vacuum residuum, if chronically or in I/P tests: sweetened naph-they exist, may not be available to the skin tha, solvent-refined heavy naphthenic dis-cells at risk for carcinogenic transformatillate, and vacuum residuum. Sweetened tion. Lewis (14) has reported similar results naphtha, a volatile fraction boiling in the from the nonvolatile asphalt fraction of range of 370-128"C, was made up mostly crude oil. Solvent-refined heavy naphof paraffinic hydrocarbons, some naphthenic distillate is a mineral oil that has thenes, and a low concentration of aromat-been extracted to remove any carcinogenic ics. The low boiling range of this stream PAHs. Kane (20) have reported that time on the skin after application. More-intensive solvent refining is effective in  eliminating the carcinogenic potency of mineral oils. [ARC has affirmed that judgment with the statement, "There is no evidence that severely solvent-refined oils are carcinogenic to experimental animals" (21: 151). Two streams, catalytic cracked clarified oil and light paraffinic distillate, exhibited carcinogenic activity in the chronic study and were classified as initiators in the I/P study. Catalytic cracked clarified oil is a residual fraction from distillation of the products of a catalytic cracking process. It is likely to contain active PAHs, some of which may be heterocyclic. The lack of activity of this stream as a promoter may be a result of its dilution (1%) rather than any intrinsic property of the stream itself.
Light paraffinic distillate is a mineral oil of relatively low viscosity with a high enough boiling range that it may contain low concentrations of PAHs, which may be responsible for its activity. The initiating activity of this stream is weak compared to that of the DMBA-positive control and catalytic clarified oil, reflecting the weak carcinogenicity relative to these materials in the chronic study.
The middle distillate samples (hydrodesulfurized kerosene, hydrodesulfurized middle distillate, light catalytic cracked distillate, and the 50/50 mixture of straight run middle distillate and light catalytic cracked distillate) were all consistent in their activities. They all led to tumor for-n after dermal exposure to carcinogens. However, Biles et al. (11) used C3H mice in their study of chronic dermal application of middle distillate samples and identified only one fibrosarcoma in 490 mice in the 10 groups tested. It appears that the C3H strain is not consistently susceptible to the development of fibrosarcomas. In any event, the appearance of fibrosarcomas in the chronic study does not affect the good correlation of the I/P study with the chronic investigation as far as tumorigenic activity is concerned. Table 5 summarizes the activity of the various petroleum streams tested in this investigation in both chronic and I/P studies. The results of the chronic bioassay in this investigation are consistent with previous studies (11,13,14), which have shown the following: 1) the most volatile streams are inactive, 2) the least volatile residues (vacuum residuum) are also inactive, 3) untreated distillates (light paraffinic distillate in this study) are active but may be made inactive by suitable treatment to remove PAHs (solvent-refined heavy naphthenic distillate in this study), 4) middle distillates show carcinogenic activity in chronic dermal studies.
The I/P investigations carried out in this study demonstrate convincingly that the middle distillates with carcinogenic activity in the chronic study act as promoters and not as initiators. This too supports results of previous investigations (12,15).
The activity of the two streams showing initiation potential, 1% catalytic cracked clarified oil, and light paraffinic distillate, may result from the presence of active PAHs and/or polyaromatic heterocycles in the case of catalytic cracked clarified oil.
The I/P portion of the investigation demonstrated perfect correlation with the chronic phase in that all streams negative in the chronic bioassay were inactive as initiators and promoters, and all streams positive in the chronic bioassay were active in either the initiation phase or promotion phase or both. This indicates, at least for petroleum streams, that the I/P bioassay may be a cost-effective substitute for the chronic test. Results would be available in 6 months rather than 2 years, providing earlier information at a lower cost. It would also furnish additional information with regard to the nature of the carcinogenic activity, which might be useful in assessing hazard.
The significance of the distinction between initiation and promotion in regard to risk assessment of petroleum streams remains to be clarified, and an API consortium is currently sponsoring research in the field. Over a number of years, hypotheses have been put forth regarding the development of at least some cancers as a two-stage process (24)(25)(26)(27). These authors and their co-workers have demonstrated that the time course of tumor incidence can be successfully modeled using this approach for naturally occurring tumors in humans (27) or for chemically induced cancers in experimental animals (24,28). One element of great significance in the model is cellular proliferation, with critical importance associated with the particular cell stage affected and the schedule of proliferation. If initiation in the I/P bioassay is considered to correspond to initiation in the two-stage model and promotion to providing a proliferative advantage to the initiated cells, skin carcinogenic activity by promoters can be examined using this model.
McKee et al. (15) have documented that skin irritation and hyperplasia are caused by repeated application of middle distillates to mouse skin, and Skisak (23) has shown that elimination of these effects can reduce or eliminate tumor development after middle distillate exposure. If compensatory hyperplasia provides a proliferative advantage to initiated cells, the dermal carcinogenic activity of middle distillates can be visualized as promotion of cells initiated by exposure to background environmental mutagens. One consequence of application of the two-stage model is that the results of a chronic bioassay of a promoter cannot be applied uncritically to risk assessment (25). Because cellular proliferation is a function of the dermal application schedule, which is different in the experimental situation than in human experience, direct extrapolation will probably lead to erroneous conclusions. If all your children deserved to go to college, but you could only send one, how would you choose?
A better question is why should you have to choose? Why shouldn't every bright, hard-working kid get a chance to fulfill his or her dreams?
That's why your donation to the United Negro College Fund is so important. As the most experienced source in providing general operating support for historically black colleges, we've enabled thousands of deserving students to get a quality education.