Document RgjjpaYj43n33YemnEaZk2Va

Induction of Liver Tumors in Sherman Strain Female Rats by Polychlorinated Biphenyl Aredor 1260** Renats D. Kimbrough,' Robert A. Squire,' Ralph E. Linder,' John D. Strandberg,' Richard J. Montali,' and Vlrlyn W. Burse >' SUMMARY--Sherman atrain famala rata (200) wera fad 100 ppm of a polychlorinated biphenyl (Aroelor 1260) for approxi mately 21 months, and 200 famala rats were kept as controls. Tha rati war* kilted whan 23 months old. Twenty-six of 184 experimental animals and 1 of 173 controls had hepatocellular carcinomas. Nona of tha controls but 146 of 184 experimental rats had naoplastic nodules In their livers, end areas of hepa tocellular alteration were noted In 28 of 173 controls and 182 of 184 axparimantai animals. Thus tha polychlorinated bi phenyl Aroelor 1260, when fed in the diet, had a hepatocarclnOfentc affect In these rets. The Incidence ol tumors In other orfant did not differ appreciably between the experimental and control groups.--J Natl Cancer Inst 96: 1453-1459, 1979. Polychlorinated biphenyls (PCB's) have been used over the past 44 years as transformer, capacitor, and cooling fluids in various systems (/\. Since they are ex cellent dielectrics and flame retardants, many new in dustrial uses have been found for them, particularly in tlie past two decades. These compounds are persistent as environmental contaminants ana were first established aa pollutants in 1966 (2). The toxicity of these and re lated compounds has recently been reviewed (7). In a preliminary feeding study in which the toxicities of two PCB mixtures (Aroelor 1254 and Aroelor 1260) were compared (l), a bladder tumor was found in 1 of 10 female rati fed 100 ppm Aroelor 1260 in their diet. This bladder tumor was tentatively classified at the time as a poorly differentiated epidermoid carcinoma. Severe autolysis hampered the microscopic examination (Kim brough RD: Unpublished observation). To establish whether this tumor had developed spontaneously or was induced by the PCB, 200 female rats were fed 100 ppm Aroelor 1260. Another group of 200 females were the controls. This paper reports the results of this study. MATERIALS AND METHODS Four hundred weanling Sherman strain COBS* fe male rats 21-26 days old and weighing 48-97 g were dis tributed into two groups of 200 animals each according to a table of random numbers. Ten animals were housed per cage in conventional humidity, light, and temperature-controlled surroundings. Two hundred rats were fed plain ground Purina Laboratory Chow and 200 were fed the same diet containing 100 ppm Aroelor 1260 (lot No. AK-S; Monsanto Industrial Chemical Co., St. Louis, Mo.). For incorporation of Aroelor 1260 into the diet, 5.2 g was dissolved in ethyl ether; this was added to 100 g cornstarch, and the ether was allowed to evapo rate. The PCB cornstarch mixture was blended with in creasing amounts of ground chow. The diets were pre- (>ared fresh every 10-14 days. Random samples from the inal mixes and samples ol control chow were taken at regular intervals to determine PCB levels, at first bi weekly and later bimonthly. For determination of the PCB levels, the samples were extracted for 2 hours on an automatic shaker with a 3:1 mixture of hexane and isopropanol (5). After filtration of the extract through glass wool, the extracts were washed with a 2% sodium chloride solution and dried with sodium sulfate. Samples were eluted through Florisil (6). lhe 6% fraction was collected and eluted through the silicic acid-celite column of Armour and Burke (7). Samples were analyzed by electron-capture, gas-liquid chromatography, and/or Coulson Conducto metric gas-liquid chromatography. During the first year of the study, three batches of food, both when first received and alter having been in the food cups for a week, were analyzed for aflamxins {8). The combined body weight of rats in each cage was recorded weekly until the rats were 6 months old, bi weekly until 12 months, and monthly thereafter. Indi vidual weights were recorded only at the onset of the ex- erimem and at death or when the animals were killed, 'he rats were observed briefly each day, and those ex hibiting debilitating signs or large tumors were removed from the group cage and housed individually. At each weighing the animals were examined individually and abnormalities were noted. Food consumption was meas ured on all rats during the first 2 weeks of the experi ment, and then during weeks 5, 8, II, and 20, and every 12 weeks thereafter. The dietary exposure of the expertmental group to Aroelor 1260 was discontinued 6 weeks before they were killed. Autopsies were performed on all that died. When the animals were 23 months old, they were anesthetized and their venae cavae were severed. Tissues were fixed in 10% buffered formalin and stained with hematoxylin and eosin. Selected tissue sections were stained with periodic acid-Schiff, Wilder's reticu lum, azure eosin, and Masson's uichrome. In addition to tissue masses, the following organs were studied micro* scopicaily: brain, pituitary, thyroid, parathyroid, tra* i Received May 20. 1975; accepted July 24, 1975. * Supported in pan by Public Health Sen ice contract N01 Cl'IJMO and .NOl CIM3288 from the Division ol Cancer Cauae and Prevention, National Cancer Institute. Toxicology Branch. Center for Disease Control. Public Health Service, U.S. Department ol Health. Education, and Welfare, At lanta, Ca. 30539. * National Cancer Institute. National Institutes of Health, Public Health Service, U.S. Department of Health, Education, and Welfare, Beihesda, Md. 20014. * Environmental Protection Agency (EPA), Research Triangle Park. N. C. 27711. " The Johns Hopkins University, School of Medicine, Baltimore. Md., 21205. ' W'e thank Mr. Martin Goldstein, Federal Drue Administration (FDA). New Orleans District, for the fret! analysis for iflatoxinv and Mr. Sol Cohen. FDA, Atlanta District, for helpful suggestions. Mr. Richard Moore (EPA) assisted with animal care and autopsies. The Sherman rats arc descendants of the Sherman strain devel oped at Colombia University from the Osborne Mendel strain. They have !>eeti randomly bred in our colony at the Center for Disease Control since 1950. They were cesarean obtained and barrier sus tained (COBS) in I960. In conventional rats, the incidence of murine pneumonia was high. Since they have been COBS, this is no longer a problem. JOURNAL OF THE NATIONAL CANCER INSTITUTE, VOL. 55. NO. 6, DECEMBER 1975 HONS 081.983 1453 1454 KIMftftOUCH ET Al_ 'chea, esophagus, King, stomach, kidney, heart, liver, spleen, adrenal, ovary, uterus, and urinary bladder. Stu dent's /test was used for comparison of mean body weights and mean weight gain. RESULTS The PCB concentrations in the experimental diet ranged usually from 70 to 1Q7 ppm. Occasionally, lower values were found at the beginning of the study. The orcunenre of low PCB concentrations in the experimen tal diet was later resolved by improvement of the extrac tion method. PCB'# in the control diets were less than 0.1 ppm and usually below the limit of detection. Aflatoxins were not detected in cither diet. The sensitivity levels were 2.5 ppb' aflatoxins B, and Gt and 0.1 ppb aflatoxins B, and G#. A study in which aflatoxin B, was added to a portion of the feed showed 94% recovery. No definite dose-related signs of toxicity were ob served in the test animals. Comparative curves of bodyweight gain and food consumption on the basis of body' weight, as well as PCB intake of the test group, are given in text-figures 1 and 2. A slight decline in the rate of weight gain in the test group compared to that in the coiurul group began about 5 months after onset of the experiment. Mean final body weights were 420 g ($d--72 g, ax-5.4 g) for the control group and S92 g (so -62 g. se-4.6 g) in the lest group: the difference was statisti cally significant (P<0.00i). Food consumption (text-fig. 2) was comparable in both groups. Mean weight gain was 550 g (sd--70 g, se --5.3 g) and 323 g (sd - 60 g, se-4.5 g) for the control and test groups, respectively; the difference in weight gain was also statistically signifi cant (P<0.001). PCB intake declined from 11.6 mg/kg/day during the first week of exposure to 6.1 mg/kg/day at 3 months of exposure and to 4.3 mg/kg/day at 20 months (text-fig. 2). Pathologic Finding* Control rats (173) and experimental animals (184) were examined grossly and microscopically. The experi mental group mcluued 5 rats that were killed 1-2 months before the final kill. The remaining animals l'EXT`F,clmE 2.--Food intake of control and experimental rau and consumption of Aroclor 1260 by experimental rats. were not included because of improper tissue fixation or because they died early in the experiment. A consistent difference in the appearance of the liven was observed between the experimental and control groups. Almost all (170/184) liven of the experimental animals had from a few to multiple elevated tan nod ules on the surfaces; additional nodules were usually seen on sectioning. These nodules varied from 0.1 to several cm in diameter, and in some rats replaced almost the entire liver. In contrast, the liver of only one con trol showed gross abnormalities and was markedly en larged, nodular, tan, and firm. In addition, a variety of tumors of other organs was observed in both the experi mental and the control groups. The incidence and type of tumors are given in table 1. Histologic examination showed that 26 experimental animals and 1 control with enlarged, nodular livers had hepatocellular carcinomas. The additional 144 experi mental rats with gross liver nodularity had hepatocellu lar nodules characteristic of neoplastic nodules [syn onym, "hyperplastic nodules" (?)]. A recent workshop sponsored by The National Cancer Institute* recom mended the term "neoplastic nodules" for these lesions as a more accurate indication of their biologic signifi cance. No nodules were in control animals. The hepatocellular carcinomas were well-differen tiated trabecular types (fi$s. 1-3), except for three in the experimental animals which had a glandular, papillary pattern (fig. 4). The trabecular tumors showed severe disruption of normal liver architecture and were usually easily recognized at low magnification. Liver plates, two or more cells th. k in some areas, were arranged in hap hazard linear, I ranching, or pseudoglandular patterns. Different patterns were usually present in the same tumor. Blunt-ended plates, sinusoidal ectasia, and congestion were frequent. The hepatocytes varied from a normal appearance to enlarged, acidophilic, or dif fusely basophilic cells with large, hyperchromatic nuclei and prominent nucleoli. The cytoplasm often contained eosinophilic inclusions within vacuoles. Mitotic figures were sometimes present. Foci of coagulative necrosis were occasionally observed in cancerous areas, but there Trxr hu hi: I.--Average hotly weight o! control nit and those con suming Anxlor 1260. Rat Liver Tumor Workshop at Silver Spring. Mt!.. Dee. M-11. 1974: sponsored by Carcinogenesis, Division of Cancer Cause ami Prevention, National Cancer institute. HONS 081984 HEFATOCARClNOGENtCITY BY AROCLOR 1200 1455 Tabus 1 ,--Incidtnct and type of liver Ittiont and luiHom of other orpart* exaviineil hiefoloyirally Orem or tissue Type of lesion Incidence ------Control* Experi mental Liver Thyroid (lend Adrenal (land Pituitary (land Uterus Urinary bladder Mammary aland Bauvary (land Lung Adinoae tiaaue Brain Ovary Hematopulctio yatem Kidney Thymus Parathyroid SkSland Hepatocellular carcinoma Neoplastic noduiea Pori or areas of cyto plasmic alteration Parafollicular cell tumor Pheochromocytoma Chromophobe adenoma Carcinoma Endometrial polyp Adenocarcinoma Sarcoma of endometnal atroma Transitional cell papilloma Fibroadenoma Adenocarcinoma Fibrosarcoma Adenoma Lipoma Glioma Granulosa theca cell tumor Papillary adenoma Granuloeytio leukemia Lymphoma Hemangioma Thymoma Adenoma Fibroma 1/173 0/173 28/173 37/160 1/173 41/153 0/153 18/140 0/149 3/149 1/167 17/1735/173* 1/173* 2/173 0/173* 0/173 5/149 1/149 1/173 0/173 0/173 1/173* 0/173 0/173* 20/184 144/184 182/184 18/166 1/167 28/139 1/139 25/163 2/163 7/163 0/169 13/184* 1/184* 0/184* 2/184 2/184* 2/184 0/163 2/163 0/184 2/184 1/184 0/184* 2/184 1/184* Inridvnn bund on erotic detection with mkrotoopie confirmation. was no fibrosis or oilier evidence of chronic degenerative changes. Periodic acid-Schiff (PAS) without diastase stained carcinomas less intensely than uninvolved liver, which suggested a decrease in glycogen. Most carcinomas were more basophilic than the normal liver with azure eosin stain. No definite intravascular invasion or metas- tasei were found. Neoplastic nodules were generally spherical and well demarcated, and occupied areas equal to those of several liver lobules (fig. 5). The cells In these nodules were gen erally enlarged, and the cytoplasm was either groundglais-apjiearmg, diffusely basophilic, or clear. Enlarged nypetchrotnauc nuclei, double nuclei, and mitotic fig ures were often present. The cytoplasm frequently con tained inclusions similar to those in the carcinomas, ex cept they were larger and appeared as whorled, concentric lamellae. In previous studies ($) these forma tions were shown by electron microscopy to represent ag gregates of smooth endoplasmic reticulum. Tnc normal liver architecture was absent within nodules, and cells were in sheets or irregular plates. Portal areas and cen tral veins were absent and sinusoids were dilated in some areas. At the )>eriphery of the nodules, the sur rounding liver plates were tangentially arranged and narrowed, due to compression (figs. 6, 7). Nodules varied in PAS positivity, but always differed from surrounding liver. Most nodules were more eosinophilic, and a few were more basophilic than the normal liver with azure eosin stain. In 182 treated and 28 control animals, there were also foci or areas of hepatocytes with altered cytoplasm (figs. 8-10). In controls, these were mostly collections of ceils with water-clear cytoplasm. In treated animals, most af fected cells were enlarged and had eosinophilic, groundglass-appearing cytoplasm or were dilfusely basophilic and smaller than normal cells, in basophilic areas, sinu soids were dilated and liver plates somewhat tortuous. In general, the cells in these areas were like those in neoplastic nodules, but there were no architectural alter ations, and plates of involved liver cells merged with the surrounding liver. No unusual features were noted in tumors of other or gans, and there were no apparent differences in inci dence between experimental and control animals. Parafollicular thyroid tumors varied from small circum scribed nodules of pale, oval-to-spindle cells to masses that obliterated the gland and invaded the capsule. They were similar to those described by Boorman et al. (JO). Other frequent tumors included pituitary chromo phobe adenomas, mammary fibroadenomas, and endome trial polyps. Endometrial stromal sarcomas were also present in 10 animals. The only bladder tumor observed was in a control an imal. Therefore, the occurrence of the bladder tumor in the previous experiment was apparently unrelated to the ingestion of Aroclor 1260. A few rats in the experimental group also showed areas of adenofibrosis (synonym, "cholaneiofibrosM") of the liver, a lesion described previously following the ad ministration of Aroclor 1260 (4). DISCUSSION The livers of treated animals showed neoplastic le sions in 170 of the 184 examined, and in only 1 of 173 controls. Although only 26 of the lesions in treated rats were clearly carcinomas according to traditional histo logic criteria, neoplastic nodules are part of the spec trum of response to hepatocarcinogens and must be in cluded in the evaluation of tumorigenesia. In the past, these liver lesions have been interpreted in various ways, and different names have been used to report them: nodular hyperplasia, hyperplastic nodules (11), hepatomas (12, ])), and hepatic nodules (14). Simi lar ambiguity exists in the classification of human liver lesions of this type (15). Several studies with known carcinogens have demon strated the development of liver nodules, indistin guishable from the neoplastic nodules in this study, be fore the appearance of carcinoma (11, 16-19). In a recent review (9), the biology and significance of nodules and their relationship to hepatocellular carcinoma were thoroughly discussed. In our study, Aroclor 1260 in duced a spectrum of nodules and cancers in treated ani mals as outlined by the Rat Liver Tumor Workshop.* A few Japanese polychlorinated biphenyls and the American product, Aroclor 1254, have been shown to in duce neoplastic liver lesions in rodents. In a study with male BALB/cJ mice, nodules were observed in the liver after II months' exposure to Aroclor 1254 (19). The Japanese PCB's Kanechlor 400 and 500 have reportedly produced liver tumors in female Donryu rats and male dd mice, respectively (20, 21). PCB's have a promoting effect on liver tumor induction by benzene hexachloride (21), whereas they protected male Sprague-Dawley rats from the tumor-inducing effect of three hepatocar cinogens: 3'-methyl-4-dimethylaminozobenzene, N-2-fluorenylacetamide, and diethylnitrosaminc (22). The au thors suggested that this protective effect may be due to the induction of microsomal enzymes by PCB's. The ani mals treated only for 20 weeks probably did not receive MOWS 081985 1456 KIMBROUGH ET AL. the material long enough for the PCB's to induce tu mors. It is not known whether mixtures of polychlorinated biphenyls that are primarily composed of isomers with four or less chlorines on the ring would also induce liver tumors, or whether only those mixtuies containing an appreciable proportion of pentachlorobiphenyl, hex* achlovobiphenyl. and hepiachlorobiphenyl have this ef fect. Kanechlor 400 contains 3% dichiorobiphenyJ, 32.8% trichlorobiphenyl, 43.8% tetrachlorobiphenyl, 15.8% pentachlorobiphenyl, and 4.6% hexachloro* biphenyl. Kanechlor 500 contains 5% trichlorobiphenyl. 26% tetrachlorobiphenyl, 55% pentachlorobiphenyl, and 13% hexachlorobiphenyl (21). Chlorinated dibenzofuran was found in one Kanechlor 400 sample (2J). The composition of the Aroclor 1260 studied by us is not known. Sissons and Welti (24) analyzed Aroclor 1260 manufactured by Monsanto Chemicals Ltd. and com eluded that hexachlorobiphenyl and heptachloro* biphenyl were major constituents. REFERENCES (/) Broadhurst MG: Use and replaceability of polychlorinated biphenyls. Environ Health Pertpeci 1:81-101:, 1972 (2) Jensen $: A new chemical hazard. New Sci 52:612, 1966 (}) Kimmovch RD: The toxicity ot polychlorinated polycyclic compounds end related chemicals. CRC Grit Rev Toxicol 2:445-498. 1974 (4) Kimmoucm RD, LiNora RE, Gaines IB: Morphological changes In livers of rats fed polychlorinated biphenyls. Arch Environ Health 25:354-364. 1972 (J) Dawsey LH, Barthel WF, Schutzmann RL, et al: Southhaven Region Manual. U.S. Department of Agriculture. Agriculture Research Service. Plant Pesticide Control Divi sion, June 1965, pp 50-59 (4) Mills PA: Detection and semicjuamftaiive estimation of chlorinated organic pesticide residues in foods by paper chromatography, J Assoc Off Agrl Chcm 12:734-740, 1959 (7) Armour J, Burke J: Method for separating polychlorinated biphenyls for DDT and it* analogs, j Assoc Off Anal CUenj 59:701. 1970 (8) Poxa WA Jr, Cvculi.u AF, Lee LS: Determination ol aft*. toxins In mixed feeds. In Third International Congress of Food, Science and Technology (5.O.5./70). W ashington, D.C., 1970, pp 705-711 (9) Parser E: Hyperplastic liver nodules. In Methods in Cancer Restarch, vo! 7, chapi Mil. New York and Loudon /Ata. demic Press, 10,3. pp 345-373 ' (10) Boorman CA, Van Boord M, Hollanoer CF: Naturally occur ring medullary thyroid carcinoma in the rat. Arch Pathol 94:35-41, 1972 (11) trntiN S. Ito N. Merkow l. et al: Cellular analysis of liver carcinogenesis; The induction of large hyperplastic nodules in the liver with 2 /luorenylacetamide or ethiomne and some aspects of their morphology umf iiIycokcu metabolism. Cancer Res 27.1702-1711. 1967 (12) VtssELtnovitch SD. Mihailovich N: The induction of benign ami malignant liver tumors by orethan in newlwrn rats. Cancer Res 28:881-887, 1968 (I)) Eowahus JE. White J: Paihologic changes, with special refer ence to pigmentation and classification of hepatic tumor* in rats fed p-dimc-thylaminoaiobeiizene (butler yellow). I Natl Cancer insi 2:157-183. 194) (ff) Poiter H, Stirsrirl SS, Oser BL, el al: The carcinogenic effect of aramite in rats. Cancer 13:1035-1046, I960 (I!) Mavs ET. Christoi-hi.rson WM. Barrows CH: Focal nodular hyperplasia o! the liver, Am J Clin Pathol 61:735-746. 1974 (IS) Klavihs IV, Kinney ID, Kaueman N: Hepatic nodules resembling turnon in raw after administration and with drawal of ethionine. Proc Soc E*p Biol Metl 89,540-543. 1965 (17) Popper H, d* la Hlerc.a J, Ycmnick C: Tumors due to pro longed eihionine feeding. Science 1)8:80-82, 1953 (18) Wachstein M. Mcjscl E: Enzymatic histochemistry of ethio nine induced liver cirrhosis and hepatoma. I Histochem Cytochtm 7:189-201, 1959 (19) Kimiirouch RD, Linder RE: Induction of adenofibrosis and hepatomas of the liver in BALB/cJ mice by polychlorinated thjihenyls (Aroclor 1254). J Natl Cancer Inst 53:547-552. (20) Kimura NT, Baba T: Neoplastic changes in the rat liver induced by polychlorinated biphenyl. Gann 64:105-109, 1979 (21) Ito N, N'acasaki H, Aria M, et al: Histopathologic studies on liver tiiniorigenesis induced in mice by technical poly chlorinated biphenyls and its promoting effect on liver tumors induced bv benzene hexacltloride. 1 Natl Cancer Inst 51.1037-1646. 1973 (22) Makiura S, Aoe H, Sicihara S. et al: Inhibitory effect of poly chlorinated biphenyls on liver tumnrigenesis in rats treated with 3'mcth)l 4-dimeihviaminoaznl>en/ene. Ar-2-ffuorenylacetamide, and diethylnitrosamine. J Nail Cancer Inst 53:1253-1257. 1974 (2J) Roach JA, Pomerastz JH: The findings of chlorinated dibrn- zofnran* in a Japanese polychlorinated biphenyl sample. Bull Environ Contam Toxicol 12:338-341, 1974 (24) Sissons D, Welti D: Structural identification of polychlori nated biphenyls in commercial mixtuies by gas-liquid chromatography, nuclear magnetic resonance and mass spectrometry. J Chroraalogr 60:15-32. 1971 MOWS 0B19tt6 1 Ftcuax 1.--Focus of hepatocellular carcinoma in rat fed Arocior 1260. Tumor cell* are in sheets. irregular nests, and cords. Hematoxylin and eoain (H It E). x 45 Fictiae 2.--Focus oi Itcpaincrllolai carcinoma in rat fed Aiv1 r 1200. A/ofe plates and nests of cells, two or mote cells in thickness, enveloped by lining tells. It It t. X I ).*> ^ * Fiutitu. 3.--Focus of hepatocellular tauiiiutiia in rat fed Arocior 1200 showing nests of cells in psvuduaciuar patterns II a I x aim Ftctiac 4.--Focus of hepatocellular carcinoma with glandular pattern in rat fed Arocior 1260. H lc E. X 105 HONS 081987 UMBROUCH KT AL. y% 1457 1 Fiouar S.--Focus of hepatocellular carcinoma in rat fed Aroclor 1260. Nolt thick cell plain, hypvrchiomatic nuclei, and proiim" nucleoli. H It E. x 900 Kir.imr. 0.--Ncnplaiiie nodule iti rat led Aroclor 12G0. Periphery la shnrply demarcated from surrounding parenchyma. H It l. x Firsuar: 7.--Edge of neoplastic nodule in figure 6. Surrounding liver plates arc compressed and tangentially arranged around nodi. Nott eoiinuphillc, lamellar bodies In cytoplatm o( nodule cells. H It . x 590 Fiouat 8.--Area of elesr-cell cytoplasmic alteration. H It E. X 185 1458 KIMBROUGH Cl \ MOMS 081988 1'ir.imi 9,--Am o( eosinophilic cytoplasmic alteration having the appearance of ground glaaa. H It E. x 185 I'lcuat 10.--Are* of baaophillc cytoplasmic alteration. H fc E. X 105 KlMDROUGH ET AL. MONS 081989 1459