key: cord-1005484-dez1jll8
authors: Mayes, B. A.; Mc Connell, E. E.; Neal, B. H.; Brunner, M. J.; Hamilton, S. B.; peters, A. C.; Ryan, M. J.; Toft, J. D.; Singer, A. W.; Brown, J. F.; Menton, R. G.; Moore, J. A.
title: Comparative Carcinogenicity in Sprague-Dawley Rats of the Polychlorinated Biphenyl Mixtures Aroclors 1016, 1242, 1254, and 1260
date: 1998-01-03
journal: Toxicol Sci
DOI: 10.1093/toxsci/41.1.62
sha: bd80f4399c3a166416f1f826ba8e2e15d6cae2c7
doc_id: 1005484
cord_uid: dez1jll8

A comprehensive chronic toxicity and carcinogenicity study was conducted on a series of Aroclors (1016, 1242, 1254, and 1260). Each Aroclor was assessed at multiple dietary concentrations, ranging from 25 to 200 ppm, for 24 months in male and female Sprague-Dawley rats. Liver toxicity was indicated by elevated serum enzyme activity (AST, ALT, and GGT), elevated serum cholesterol concentration, decreases in hematologic parameters (RBC, Hb, and Hct), hepatocellular hypertrophy, an increased incidence of altered hepatocellular foci, and an increased incidence of hepatocellular neoplasms (primarily adenomas). Liver toxicity was distinctly more severe in females than in males. The incidence of hepatocellular neoplasms was highly sex-dependent (females > males), differed between Aroclor mixtures and, for females, increased with dose and followed the general incidence pattern of Aroclor 1254 > Aroclor 1260 ≈ Aroclor 1242 > Aroclor 1016. A significant response (p < 0.05) in males was seen only for the high dose of Aroclor 1260. A small increase in the incidence of thyroid gland follicular cell adenomas was noted in males for Aroclors 1242, 1254, and 1260, with the incidence being uniform across dose groups and Aroclor mixtures. For females, increased survival relative to controls was observed for all Aroclor treatment groups. A significantly decreased trend in the incidence of mammary gland neoplasms compared to control was also noted for females receiving Aroclors 1242, 1254, and 1260.

Aroclors 1016, 1242, 1254, and 1260 are complex mixtures of polychlorinated biphenyl (PCB) congeners, prepared by the chlorination of biphenyl. For Aroclors 1242, 1254, and 1260, the chlorination reaction was stopped when the weight percent chlorine of the product had reached 42, 54, or 60%, respectively. By contrast, Aroclor 1016 was prepared by the fractional distillation of Aroclor 1242, which excluded the higher boiling (i.e., more highly chlorinated) congeners. Aroclors 1016, 1242, 1254, and 1260 were the most widely used PCBs in the United States, and accounted for 92% of all 1958-1977 production (Monsanto, 1980) . Polychlorinated biphenyls were versatile materials which found use in a variety of applications including plasticizers, printing inks, and heat exchange, dielectric, and hydraulic fluids. The dielectric properties, chemical stability, and noncombustibility of PCB fluids made them a particularly attractive alternative to flammable mineral oils for use in capacitors and indoor transformers.

Concerns in the late-1960s regarding the environmental accumulation of highly chlorinated PCB mixtures resulted, in 1971, in a voluntary cessation of the manufacture of Aroclors 1232 Aroclors , 1248 Aroclors , 1260 Aroclors , 1262 Aroclors , and 1268 ; the introduction of Aroclor 1016; and to a restriction in PCB usage to totally enclosed systems. The Toxic Substances Control Act of 1976 made the manufacture, importation, sale, or use of PCBs illegal after 1978.

The tumorigenicity of PCBs in rodents was first reported in mice by Nagasake et al. (1972) and Ito et al. (1973) , which were followed by reports of tumorigenicity in rats by Kimbrough et al. (1975) , Schaeffer et al. (1984) , Norback and Weltman (1985) , and the National Cancer Institute (NCI) (1978) . In many instances, study features such as single sex representation, single exposure concentration, and restriction of evaluation to only highly chlorinated mixtures (mainly 60% Cl compositions) severely limited the interpretative value of the data set. The studies in rats formed the basis for the Environmental Protection Agency Note. Abbreviations used: Biphenyl, BP; polychlorinated dibenzodioxin, PCDD; polychlorinated dibenzofuran, PCDF; polychlorinated biphenyl, PCB; toxic equivalency, TEQ.° This composition represents the mixture of Aroclor 1254 that was most extensively used commercially (1958) (1959) (1960) (1961) (1962) (1963) (1964) (1965) (1966) (1967) (1968) (1969) (1970) (1971) (1972) (1973) (1974) (1975) (1976) (1977) and is included in the table for comparison.

'The lot of Aroclor 1254 available for preparing the test article was produced in 1976 using a modified procedure that was in use ca. 1976-1977. Prior to administering Aroclor 1254 in the current study, it was treated by AccuStandard using Preston's procedure (Preston et ai, 1981) (producing Lot No. 122-078) which removed >99% of the PCDFs and about 35% of congener 126 originally present in Lot KI-02-6024. c ND, none detected.

(EPA) classification of PCBs as probable (Group B2) carcinogens (U.S. Environmental Protection Agency, 1988). In studies where PCB mixtures of differing chlorine content have been evaluated (Ito et al., 1973; Nagasaki et al., 1972; Schaeffer et al., 1984) , the data suggest that tumorigenic potential decreases with decreasing chlorination; nevertheless, the EPA chose to regulate all PCBs as equivalent to highly chlorinated mixtures, with a cancer potency factor of 7.7 (mg/kg-day)" 1 (U.S. Environmental Protection Agency, 1988). In 1993, the General Electric Co. initiated a 2-year carcinogenicity study in Sprague-Dawley rats which conformed to EPA testing (U.S. Environmental Protection Agency, 1992) and Good Laboratory Practice (U.S. Environmental Protection Agency, 1992) guidelines. The purpose of this study was to allow direct comparisons of the most widely used Aroclors (1016, 1242, 1254, and 1260 ) in a design that would facilitate evaluation for dose response, chronic toxicity, carcinogenicity, neurotoxicity, and toxicokinetics. The current paper reports the findings associated with the carcinogenicity (24-month exposure) segment of this study. weeks old, and weighing 248-252 g (males) and 198-203 g (females), respectively, at study initiation, were assigned to 11 treatment groups (50 rats/sex/dose group) and one control group (100 rats/sex) in a manner that minimized group mean body weight differences. These groups were augmented by additional animals for evaluations of chronic toxicity and neurotoxicity and the progression and reversibility of pathology, the results of which will be reported separately. Prior to study initiation, animals were certified free of Sendai virus, pneumonia virus of mice, coronavirus/sialodacryoadentis, Kilham rat virus, Toolan's H-l virus, and Mycoplasma pulmonis by collecting serum from 5 rats/sex for evaluation (Microbiological Associates, Inc., Rockville, MD). All animals were acclimated to, and maintained in, a controlled environment of 68-80°F, 40-70% relative humidity and 12-h light/dark cycles. Animals were identified by tail tattoos and individually housed in stainless-steel wire mesh cages. Twice each day animals were evaluated for clinical signs and mortality. Once per month for the first 6 months, and weekly thereafter. animals were examined in detail by palpation for tissue masses or other abnormalities. Ophthalmic examinations were conducted on all animals prior to study initiation and on all surviving control and high-dose group animals from each Aroclor at 24 months. All groups assigned to a single Aroclor mixture were segregated in compound-specific animal treatment rooms. The controls were housed sepa-rately. To minimize the potential for cross-contamination, technical staff members either were restricted to working with a single Aroclor mixture or were required to shower and regarment before entering a different Aroclor animal treatment room. Animal housing and care conformed with the guidelines established by the NIH Guide for the Care and Use of Laboratory Animals, NIH Publication No. 86-23. (Table 1) . PCDD concentrations were 0.6, 0, 20, and 0 ppb for Aroclors 1016, 1242, 1254, and 1260, respectively PCDF analyses showed total concentrations to be within acceptable ranges for Aroclors 1016, 1242, and 1260 (i.e., 0.035, 2 9, and 4.9 ppm, respectively), but elevated for Aroclor 1254 (23 ppm). Accordingly, Aroclor 1254 was treated for PCDF removal by AccuStandard using the procedure of Preston el al. (1981) . This removed >99% of the PCDFs and also reduced the concentration of congener 126 (3,3',4,4',5-PCB) approximately 35%, but still left the concentration of congener 126 three to five times greater, and consequently the TEQ (toxic equivalency) about two times greater (calculated as per Ahlborg (1994) ), than that of ordinary Aroclor 1254. It was subsequently found that this lot of Aroclor 1254 had been made by a modified procedure that was used only in the final years of manufacture, and accounted for under 1 % of the total Aroclor 1254 production for the years 1958-1977.

All test materials were administered in the feed (Purina Rodent Chow Meal No. 5002) fed ad libitum. Each lot of the basal diet was confirmed by the manufacturer to contain <0.15 ppm of PCB. Dosed diets were prepared by dissolving Aroclor in hexane and combining with a small amount of diet (the "premix"). The hexane was evaporated, and the premix was subsequently blended (Patterson-Kelley Cross Flow Blender) with an appropriate amount of basal diet to achieve the desired final PCB concentrations. Control feed was handled in a similar manner, without the addition of test material. Periodically during the study, samples were taken to confirm that diet preparations were achieving target PCB concentrations, were homogenous, and were not cross-contaminated. Randomly selected samples of diet were also periodically submitted to an independent laboratory (Northeast Analytical Laboratories, Schenectady. NY) for referee PCB concentration analyses.

Feed consumption was monitored "continuously" throughout the study by determining the quantity of feed consumed at weekly intervals. Water was supplied ad libitum from a municipal water supply and was analyzed periodically for microbes and other contaminants, including PCBs.

In addition to evaluations for mortality, clinical signs, and palpation for tissue masses described previously, general toxicity was monitored using weekly body weight and feed consumption data.

At 24 months, urine was collected from 10 nonfasted animals/sex/group over a period of approximately 18 h for measurement and evaluation of specific gravity, volume, color, appearance, and microscopic characteristics; and qualitative dipstick determination (Rapignost, Behringwerk AG, Marburg, Germany) of pH, protein, glucose, blood, ketones, bilirubin, and urobilinogen. Subsequently, these animals were anesthetized by CO 2 inhalation and blood was collected from the retro-orbital sinus. Serum chemistry evaluations included aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), y-glutamyl transpeptidase (GGT), lactate dehydrogenase (LD), urea nitrogen (BUN), creatinine, creatine kinase (CK), total protein, albumin, sodium, potassium, chloride, calcium, inorganic phosphate, total bilirubin, cholesterol, and glucose. Hematology evaluations included red blood cell count (RBC), white blood cell count (WBC), differential count, platelet count, hematocrit (Hct), hemoglobin concentration (Hb), mean cell hemoglobin (MCH), mean cell volume (MCV), and mean cell hemoglobin concentration (MCHC).

Pathology. At the end of the study (24 months), the surviving rats were killed by CO 2 asphyxiation and examined for macroscopic lesions. Selected organs were removed and weighed. Tissue samples were fixed in 10% neutral-buffered Formalin, except eyes and testes which were preserved in Bouin's solution. The following tissues were fixed (and weighed (*)): adrenals (*), aorta, bone marrow, brain (•), cecum, colon, duodenum, epididy- mides (•), esophagus, eyes, femoris muscle, Harderian glands, heart (*), lleum, jejunum, kidneys (•), exorbital lacrimal glands, liver (•), lungs (*), mammary gland, mesentenc lymph nodes, optic nerves, ovaries (•), pancreas, parathyroids, pituitary (*), prostate (*), rectum, sciatic nerve, seminal vesicles (*), skin, sublingual and submaxillary (•) salivary glands, spinal cord, spleen (*), sternum, stomach, testes (•), thymus (•), tongue, thyroid (* with parathyroids), trachea, urinary bladder, uterus (*), vagina, and gross lesions and tissue masses. Histopathologic evaluations of paraffin-embedded and hematoxyhn-and eosin-stained sections were performed on the above organs/tissues for all animals from Aroclor high-dose groups and controls, and on all animals that died prior to 24 months. Evaluations for all remaining animals included the brain, liver, mammary gland, and gross lesions of both sexes, and the thyroid gland for males. Upon completion of the study, all liver (both sexes) and thyroid gland (males) suspect neoplastic lesions were reviewed by a group of independent pathologists (Pathology Working Group) (Hardisty el al., 1996) without knowledge of exposure groups, utilizing the diagnostic criteria for proliferative hepatocellular and thyroid lesions in rats currently proposed by the Society of Toxicologic Pathologists (Botts et al., 1991; Goodman et al, 1994) .

Statistics. The primary objective of the statistical analysis was to determine whether the administration of Aroclor test substances in diet resulted in increased tumor occurrence. The number of neoplasms in Aroclor-dosed groups were compared to those observed in the Control group using Fisher's exact test (Haseman, 1984) . Tests for linear trend were also conducted for tumor response data. Nonneoplastic histomorphologic lesions were not analyzed for statistical differences. Additional analyses were conducted to determine the effects of each Aroclor test substance on mortality, body weight, body weight gains, feed consumption, clinical pathology, and organ weights.

Survival analysis assessments utilized Tarone 

Group mean body weights for males in all treatment groups of Aroclors 1016, 1242, and 1260 were not statistically different from those of control (Figs. 1 -4) . Dose-related significant decreases in group mean body weights were recorded for males in the 50 and 100 ppm exposure groups of Aroclor 1254 beginning at 12 and 1 months, respectively. The decreases from control at 24 months were approximately 12 and 18%, respectively.

Group mean body weights for females in all treatment groups of Aroclors 1016 and 1260 were not statistically different from those of control. Dose-related significant decreases in group mean body weights were recorded for the 100 ppm Aroclor 1242 group beginning at approximately 4 months. Although the weight decrease continued for the duration of the study, it was not significant (p < 0.05) after 17 months. Dose-related significant decreases in group mean body weights for the 25, 50, and 100 ppm Aroclor 1254 groups were evident beginning at approximately 9, 3, and 1 months, respectively. At 24 months, the differences from control were 10% for the 100 ppm Aroclor 1242 group, and 15, 15, and 28% for the 25, 50, and 100 ppm Aroclor 1254 groups, respectively. There was no treatment-related effect on feed consumption or evidence of poor palatability of the Aroclor-dosed diets. Feed consumption rates (g/kg/day) were greater for females than males, but were similar for all groups within each sex. The slightly greater feed consumption rate for females resulted in a correspondingly greater PCB exposure for females in all dose groups (Table 2) .

Clinical signs and mortality. No patterns of abnormal behavior or overt signs of toxicity were observed in any treatment group. Palpations for masses, and ophthalmic examinations, did not reveal any patterns attributable to treatment.

Mortality for PCB-exposed males was similar to controls, with the exception of significant increases for the 100 ppm Aroclor 1016 and 50 ppm Aroclor 1254 groups. In no instance was the increased mortality clearly attribut- 1242, or 1260. A slight dose-related decrease in MCHC was calculated for groups receiving Aroclor 1254, correlating with the slight decrease in Hb and Hct for the 50 and 100 ppm Aroclor 1254 groups (Table 4) . For females, with the exception of the group receiving 50 ppm Aroclor 1242, decreases in mean RBC count were measured, generally accompanied by concomitant decreases in Hb concentration and Hct (Table 4 ). These reductions resulted in associated decreases in the calculated parameters (MCV, MCH, and MCHC). The greatest reductions were associated with Aroclor 1254, which also produced slightly increased reticulocyte counts for the 50 and 100 ppm dose groups.

No treatment-related effects occurred for ALP, CK, LDH, BUN, creatinine, Na + , K + , C\~, Ca 2+ , phosphate, glucose, albumin, total protein, WBCs, or any of the measured urinalysis parameters.

Organ weights. Evaluation of male group mean organ weight data (normalized as organ-to-brain weight ratios to eliminate variability from age-and treatment-related body weight effects) showed treatment-related increases for all Aroclors in liver weight ratios, and in thyroid weight ratios for Aroclors 1242, 1254, and 1260 (Table 5 ). The magnitude of the effects were generally small and were not clearly dose-related. FIG. 11 . Kaplan-Meier survival curves: Aroclor 1254 females able to exposure to PCBs. Survival for PCB-exposed females was greater than the controls for all Aroclor groups, and was significantly increased compared to control for all Aroclor 1016 and Aroclor 1242 groups and for the 25 ppm Aroclor 1254 and 100 ppm Aroclor 1260 dose groups .

Serum chemistry, hematology, and urinalysis. In males, no treatment-related effects on serum chemistry were noted for any groups receiving Aroclors 1016 or 1242. Males receiving Aroclor 1254 had slight increases in serum cholesterol concentrations and males receiving 100 ppm Aroclor 1260 had increased GGT that appeared to be treatment-related (Table 3) .

In females, no treatment-related effects on serum chemistry were noted for Aroclor 1016. Dose-related increases in AST, ALT, and GGT were measured for Aroclor 1254, and elevations in AST and GGT were also noted for the 100 ppm dose groups of Aroclors 1242 and 1260. Additionally, treatment-related increases in cholesterol were measured for Aroclors 1242, 1254, and 1260 (Table 3) . The increases in total bilirubin measured for Aroclor 1254 were small and were considered equivocally related to treatment.

No treatment-related effects on hematology for males were measured for any groups exposed to Aroclor 1016, GGT (IU/L) 1 ± 3 2 ± 2 1 ± 3 2 ± 3 0 ± 0 1 ± 1 4 ± 6 2 ± 1 3+2 2 ± 2 4 ± 4 11 ± 12" 1 ± 1 0 ± 1 2 ± 2 2+4 2 ± 3 27 ± 41 26 ± 38 62 ± 51' 57 ± 49* 6 ± 8 8 ± 6' 29 ± 43 (mg/dL) Note. Abbreviations used: Aspartate aminotransferase, AST; alanine aminotransferase, ALT; y-glutamyl transpeptidase, GGT. " Mean ± SD. * n = 10, unless otherwise noted. c n = 9.

" Statistically significantly different from control atps 0.05. ' Statistically significantly different from control at p = 0.01.

For females, an increase in liver-to-brain weight ratio was similarly apparent for all Aroclors, and was marginally greater than that observed in males. The increase was greatest for Aroclor 1254, followed by Aroclor 1260, Aroclor 1242, and Aroclor 1016 (Table 5) . No effects were noted for female thyroid-to-brain weight ratios.

Macroscopic pathology. Treatment-related increases in macroscopically visable lesions were restricted to the liver and thyroid gland for males and to the liver for females. In both sexes, liver abnormalities included enlargement, discoloration (generalized darkening), and foci (local, circumscribed color change), nodules (<0.5 cm) and masses (£=0.5 cm). In general, the incidence of liver lesions was greatest for females, but in both sexes the trend approximately followed a dose-response pattern, with the greatest incidences found in Aroclor 1254, followed, in order, by Aroclor 1260, Aroclor 1242, and Aroclor 1016. The thyroid glands of male animals receiving Aroclors 1242, 1254, and 1260 were enlarged in a non-dose-related manner. Treatment-related de-creases in mammary gland masses compared to control were recorded for females in one or more groups from each of the four Aroclors evaluated. The reduced incidences of mammary gland masses were not dose-related.

Histopathology (nonneoplastic). The liver was the primary target organ in both males and females. The incidences of nonneoplastic lesions is presented in Table 6 . Hepatocellular hypertrophy and hepatic foci, the principal findings, were observed for all Aroclor treatment groups in both males and females. The hepatocellular hypertrophy was most evident in centrilobular regions, with gradual attenuation as the portal triads were approached. Hepatocellular hypertrophy increased in severity in a dose-response manner for each Aroclor. The response for Aroclors 1016, 1242, and 1260 was minimal to mild, and for Aroclor 1254 mild to moderate. The median lobe of the liver was generally slightly more severely affected than the remaining lobes. Hepatocellular hypertrophy was not observed in controls.

For all Aroclors, increased incidences of hepatic foci were 33.4 ± 1.1 33.1 ± 1.1 33.3 ± 0.9 33.3 ± 0.7 33.0 ± 0.8 33.1 ± 1.0 32.3 ± 0.9" 32.1 ± 0.8" 31.9 ± 0.4' 32.5 ± 0.7" 32.8 ± 0.3 32.5 ± 1.0 33.6 ± 0.5 33.5 ± 1.1 33.8 ± 1.0 32.9 ± 0.7 33.9 ± 1.0 34.1 ± 0.9 34 0 ± 0.7 33.3 ± 0.9 33.9 ± 0.7 34.0 ± 1.1 33.6 ± 0.5 33.9 ± 1.0

6.6 ± 13.4 4.9 ± 6.8 2.0 ± 1.4 2.7 ± 1.4 1.9 ± 1.1 2.7 ± 3.1 2.2 ± 1.1 2.4 ± 1.0 2.7 ± 0.9 1.9 ± 0.5 2.0 ± 1.8 2.1 ± 0.9 1.7 ± 0.4 2.9 ± 2.0 3.3 ± 3.1 1.9 ± 0 7 2.9 ± 3.4 3.8 ± 5.6 1.9 ± 0.8 3.0 ± 0.9* 4.0 ± 1.5' 2.5 i 1.7 1.8 ± 0.8 5.5 ± 11.6 RBC (10-*/mL) 6.92 ± 1.53 6.66 ± 1.62 7.31 ± 1.15 6.97 ± 1.20 7.22 ± 0.70 7.26 ± 1.40 7.05 ± 0.75 6.65 ± 0.86 6.82 ± 0.61 6.85 ± 1.00 6.80 ± 0.92 6.96 ± 0.64 7.36 ± 0.81 6.37 ± 1.05" 5.94 ± 1.54" 6.56 ± 0.40" 6.54 ± 1.43 6.30 ± 1.45 6.77 ± 0.55 6.19 + 0.66' 5.95 ± 0.66' 6.30 ± 0.97" 6.30 ± 1.04" 5.76 ± 1.84" Note. Abbreviations used: Hemoglobin concentration, Hb; hematocnt, Hct; mean cell volume, MCV; mean cell hemoglobin, MCH; mean cell hemoglobin concentration, MCHC; red blood cell count, RBC.

" Mean ± SD. b n = 10, unless otherwise noted. c n = 9. " Statistically significantly different from control at p s 0.05. * Statistically significantly different from control at p s 0.01. identified more frequently in females than in males, and occurred at a generally similar incidence for Aroclors 1242, 1254, and 1260, and at a slightly reduced incidence for Aroclor 1016. Eosinophilic foci were the most prevalent form, with clear cell, mixed cell, and basophilic foci also being increased, but to a much lesser extent. The occurrences of hepatocyte vacuolization and bile duct hyperplasia were also increased in males and females dosed with Aroclor 1242, 1254, or 1260, and in females only for Aroclor 1016. Females also showed an increased incidence of liver pigment, especially in animals receiving Aroclor 1242 or 1254. The pigment, which was gray/brown in color, was localized in the sinusoidal lining cells and generally filled the cytoplasm of affected cells.

The incidence of follicular cell hyperplasia in the thyroid gland (generally minimal or mild) was increased in a non dose-related pattern for males exposed to Aroclors 1242, 1254, and 1260 (Table 7) .

Histopathology (neoplastic). Sex-dependent significant increases in tumor responses were noted for liver (hepatocellular adenoma and carcinoma, hepatocholangioma, and hepatocholangiocarcinoma), and thyroid gland (follicular cell adenoma), while significant decreases were measured for mammary gland (fibroadenoma, adenoma, and adenocarcinoma) (Tables 7-10). For males, the response in liver was restricted to a single dose group, 100 ppm Aroclor 1260. Slight, non-dose-related, increases in the incidence of thyroid gland follicular cell adenomas were measured for males receiving Aroclors 1242, 1254, and 1260.

For females, a significant and generally dose-related increase in the incidence of hepatic neoplasms (primarily adenomas) was measured for all treatment groups, except 50 ppm Aroclor 1016. The magnitude of the increases was greatest for Aroclor 1254, followed by Aroclor 1260 » Aroclor 1242 > Aroclor 1016. The number of animals with multiple hepatocellular adenomas increased in a dose-re- sponse manner for all Aroclors, and was most pronounced for the 50 and 100 ppm dose groups of Aroclor 1254 and the 100 ppm dose group of Aroclor 1260, which were also the groups with the greatest total incidences of hepatocellular neoplastic response. With the exception of one female in the 100 ppm Aroclor 1260 dose group, the liver neoplasms did not metastasize. No increases in thyroid gland neoplasms were noted for females.

A significantly decreased trend in the incidence for spontaneously occurring mammary gland tumors in females was measured for Aroclors 1242, 1254, and 1260 (Table 10 ).

The current study is the most comprehensive chronic toxicity and oncogenicity evaluation of PCBs ever reported. Previously, the tumorigenic potency of PCBs was estimated by the EPA to have a cancer slope factor of 7.7 (mg/ kg • day)" 1 , based upon the findings with Aroclor 1260 in a 1985 study by Norback and Weltman (U.S. Environmental Protection Agency, 1988) . The upper-bound cancer slope factors calculated from the current, and more complete, data set range from 0.07 to 1.5 (mg/kg-day)" 1 (U.S. Environmental Protection Agency, 1996) , a clear indication that the carcinogenic potency of PCBs is less than previously assumed.

The exposure of male and female Sprague-Dawley rats to Aroclors 1016, 1242, 1254, and 1260 for 2 years did not result in any changes in appearance or behavior that were suggestive of systemic toxicity. In chronic studies, decreases in body weight gain patterns and/or increased mortality are predictive indicators of systemic toxicity. In the current study biologically meaningful decreases in body weight gains were confined to only groups receiving Aroclor 1254, especially females. The absence of a morbid response is further reinforced by the lack of treatmentrelated mortality in males and the increased survivability exhibited for all groups of females receiving Aroclors.

In this and previous chronic studies of PCBs, the principal target organ for toxicity in rats is the liver. Significant dose-related increases in group mean liver weights were measured for females (24-120%), while for males the increases were small (s;20%) and were not dose-related. These weight increases correlated with hepatocellular hypertrophy which, in turn, was associated with induction of mixed-function oxidases, a well-known consequence of PCB exposure in rats, and which has been confirmed for the current study .

Significant liver toxicity was restricted almost exclusively to females. It was manifested by increases in cellular enzyme leakage associated with membrane disruption (AST, ALT), bile duct proliferation (GGT), altered lipid metabolism (cholesterol), and by hematologic changes (decreases in RBC count, Hb concentration, and Hct), which was presumably the result of decreased erythropoietinogen production by the liver, or possibly, of decreased hepatic (6) 7(14) 14 (28) (56) Basophilic focus (%) 7(7) 3(6) 5(10) 2(4) 2(4) 2(4) 3(6) 2(4) 2(4) 3(6) 3(6) 1 (2) 5(5) 12 (24) (24) 18 (36) retinoid concentrations (Chen et at, 1992; Sauberlich and Hodges, 1997) . The hematologic changes were small and were generally not sufficient to induce an increased release of reticulocytes.

The current study clearly indicates that the hepatic tumor response to PCBs in Sprague-Dawley rats is sexdependent. The tumorigenic response in females was Aroclor-and dose-dependent, while the response in males was Follicular cell hyperplasia (%) 1 CD 1 (2) 3(6) 1 (2) 4(8) 3(6) 12 (24) 6(12) 6(12) 6(12) 9(18) 5(10) Follicular cell adenoma (%) 1 (1) 3(6) 2(4) 0(0) 5 (10)' 5 (10)' 6(12/ 4(8)' 5(10r 6(12)-4(8)' 3(6) Follicular cell carcinoma (%) 1 (1) 1 (2) 1 (2) 1 (2) 2(4) 1 (2) 1 (2) 3(6) 1 (2) 1 (2) 1 (2) 1 (2) Total males with neoplasms* (%) 2(2) 4(8) 3 (6) 4(4) 1 (2) 1 (2) 2(4) 1 (2) 3(6) 2(4) 2(4) 6(12) 2(4) 5(10) 7 (14)' 3(3) 1 (2) 1 (2) 2(4) 1 (2) 1 (2) 2(4) 2(4) 0(0) 1 (2) 1 (2) 3(6) 0(0) 0(0) 1 (2) 0(0) 0(0) 2(4) 0(0) 4(8/ 6(12/ 1 (2) 1 (2) 5 (10/ Hepatocholangioma (%) (6) (Keenan et al., 1995) . The morphologic appearance of the thyroid tumors were characteristic of those that develop as a secondary response to chronic overstimulation by thyroid-stimulating hormone (TSH). This phenomonon is more common in males than female rats due to higher circulating levels of TSH in males (Capen 1996) . Oversecretion of TSH results when hepatic enzyme induction accelerates the metabolism of T3 and T4, with subsequent loss of feedback inhibition within the hypothalamic-pituitary-thyroid axis (Curran and De-Groot, 1991) . PCBs are very effective hepatic mixed-function oxidase enzyme inducers, and presumably produce thyroid tumors in male rats through hypersecretion of TSH as described above.

The incidence of spontaneous mammary tumors in untreated female Sprague-Dawley rats is high, 54% in the current study with even higher rates (67-94%) being reported by others (Keenan etal., 1995; Kociba et al., 1978) . In a study by Kimbrough et al. (1975) , administration of Aroclor 1260 to Sherman rats resulted in a reduced incidence of mammary tumors. The results of the current study are consistent with this finding because the rate of spontaneous mammary tumors in PCB-exposed females was decreased. The decreases were most consistently observed for groups receiving Aroclors 1242 and 1260, although decreases were also observed for some groups receiving Aroclors 1016 and 1254. The decreased tumor rates may be the result of an alteration in the 2-hydroxylation versus 16a-hydroxylation ratio of estrogen (Bradlow et al., 1995; Fishman et al., 1995) , which has been demonstrated for the current study (Amaratunga et al., 1997) . In summary, the results of this study indicate that PCBs have sex-dependent tumor-suppressing and tumor-enhancing activities in Sprague-Dawley rats. Tumor repression occurred in the mammary glands of the females. Tumor enhancement occurred in the thyroid gland of the males and the liver of both sexes, but primarily in the liver of the females. The tumors that were produced were mostly benign and did not curtail the natural life span of the animals. The study also shows that PCBs differ in tumor-inducing potency, and supports the use of PCBspecific cancer slope factors for scientifically based risk assessments.

Toxic equivalency factors for dioxin-like PCBs

Estrogen metabolism by hepatic microsomes from rats chronically dosed with Aroclors 1016, 1242, 1254, and 1260

Proliferative lesions of the thyroid and parathyroid glands, E-3. Guides for Toxicologic Pathology

Indole-3-carbinol-A novel approach to breast cancer prevention

Hormonal imbalances and mechanisms of chemical injury of the thyroid gland

Polychlorinated and polybrominated biphenyl congeners and retinoid levels in rat tissues: Structure-activity relationships

The effects of hepatic enzymeinducing drugs on thyroid hormones and the thyroid gland

Biochemical measurements on hepatic tissues from SD rats fed Aroclors 1016, 1242, 1254, and 1260

The role of estrogen in mammary carcinogenesis

Proliferative and selected other lesions in the liver of rats, GI-5. Guides for Toxicologic Pathology

An Assessment of the Chronic Toxicity and Oncogenicity ofAroclor-1016, Aroclor-1242, Aroclor-1254, and Aroclor-1260 Administered in Diet to Rats: Pathology Peer Review and Pathology Working Group Review of Neoplastic Lesions in the Liver and Thyroid Gland Experimental Pathology Laboratories

Statistical issues in the design, analysis and interpretation of animal carcinogenicity studies

Histopathologic studies on liver tumorigenesis induced in mice by technical polychlorinated biphenyls and its promoting effect on liver tumors induced by benzene hexachloride

Diet, overfeeding, and moderate dietary restriction in control Sprague-Dawley rats. I. Effects on spontaneous neoplasms

Induction of liver tumors in Sherman strain female rats by polychlorinated biphenyl Aroclor 1260

Results of a two-year chronic toxicity and oncogenicity study of 2,3,7,8-tetrachlorodibenzo-p-dioxin in rats

Polychlorinated Biphenyls (PCB's): A Report on Uses, Environmental and Health Effects and Disposal

A comparison of liver tumor diagnoses from seven PCB studies in rats

Bioassay of Aroclor 1254 for possible carcinogenicity. CAS No. 27323-18-8. NCI Carcinogen

Polychlorinated biphenyl induction of hepatocellular carcinoma in the Sprague-Dawley rat

Promoting effects of polychlorinated biphenyls (Aroclor 1254) and polychlorinated dibenzofuran-free Aroclor 1254 on diethylnitrosamine-induced tumorigenesis in the rat

Vitamin A deficiency and anemia

Pathology of chronic polychlorinated biphenyl (PCB) feeding in rats

Jr (1997) Tissue levels and tumor responses of SD rats fed Aroclors 1016, 1242, 1254, and 1260 suggest two processes critical to hepatotumorigenesis

Drinking Water Criteria Document for Polychlorinated Biphenyls (PCBs) EPA

PCBs: Cancer Dose-Response Assessment and Application to Environmental Mixtures

Health Effects Testing Guidelines: Oncogenicity. Code of Federal Regulations

Good Laboratory Practice Standards. Code of Federal Regulations

Total animals with neoplasms (%) 7(7) 2(4) 2(4) 4(8) 1 (2) 4(8) 4(8) 4(8) 6(12) 3(6) 6(12) 10(20/ nonsignificant for all groups, except the group receiving the highest dose of Aroclor 1260 (100 ppm). The overwhelming majority of the tumors induced were benign hepatocellular adenomas (80%), with smaller numbers of hepatocellular carcinomas (11%), hepatocholangiomas (8%), and hepatocholangiocarcinomas (<1%). These neoplasms did not adversely impact treatment-group survival rates because, for all Aroclor-exposed groups of females, survival was greater for treatment groups than for controls. For females, the tumor incidence differed between Aroclor mixtures in a manner that paralleled the differences between Aroclors in total TEQ (calculated as per Ahlborg (1994) ), i.e., Aroclor 1254 > Aroclor 1260 » Aroclor 1242 > Aroclor 1016 (Table 1) . A more complete description of the correlation of hepatic TEQ and PCB accumulation with tumor incidence for males and females has been reported by . The composition of Aroclor 1254 used in the current study contained twice the TEQ of the product that was generally used in commercial applications. Therefore, the response for females in this study overrepresents the response that would have been expected for ordinary Aroclor 1254.The liver tumor response rate for females receiving Aroclor 1254 in the current study (38, 56, and 56% for the 25, 50, and 100 ppm dose groups, respectively) was also much greater than that observed in a study conducted by the NCI (1978), wherein liver tumor incidence rates of 4, 8, and 4% were recorded at dietary doses of 25, 50, and 100 ppm, respectively (Moore et al., 1994) . The reason for the disparate tumor response rates is unknown, but a differential sensitivity between the rat strains used in the two studies (Fisher 344 vs Sprague-Dawley) is clearly a possibility. Another difference between the studies may be the composition of Aroclor 1254 used, i.e., lot KB01-604, produced.in 1971 , versus lot KI-02-6024, produced in 1976 and subsequently modified (see Table 1 ).A slight, non-dose-related, increase in the incidence of thyroid gland follicular cell adenomas was observed in males receiving Aroclors 1242, 1254, and 1260. The background incidence of this tumor type in male Sprague-