Strain dependent effects of sex hormones on hepatocarcinogenesis in mice

Transcription

1 Carcinogenesls vol.17 no.2 pp , 1996 ACCELERATED PAPER Strain dependent effects of sex hormones on hepatocarcinogenesis in mice Therese M.Poole and Norman R.Drinkwater 1 McArdle Laboratory for Cancer Research, University of Wisconsin Medical School, 00 University Avenue, Madison, WI 53706, USA 'To whom correspondence should be addressed In order to study the interaction of genetic and hormonal factors during murine hepatocarcinogenesis, we compared the number of liver tumors induced by treatment of 12- day-old mice with AyV-dlethylnitrosamine (DEN) (0.05 (imol/g body wt) in intact mice and animals gonadectomized at 8 weeks of age from the three inbred strains, C3H/HeJ (C3H), (B6), and (BR). At 50 weeks of age, the mean liver tumor multiplicity in intact BR females was 28 ± 13, while that for intact female C3H and B6 mice was 1.4 ± 4.7 and 0.5 ± 1.0, respectively. In ovariectomized mice, the yield of liver tumors was ~8-fold higher than in intact C3H (103 ± 7.5) and B6 (4.1 ± 6.6) females. Only a slight increase (35 ± ) was seen in ovariectomized BR females compared to intact BR females. Castration resulted in lower mean tumor multiplicities at weeks of age in the males of all three strains. Intact male C3H, B6, and BR mice had mean liver tumor multiplicities of 61 ± 34, 7.4 ± 13, and 26 ± 18, respectively, while the mean tumor multiplicities in castrated C3H, B6, and BR mice were ±, 0.5 ± 0.9, and 6.1^ ± 10 tumors per mouse, respectively. The apparent rate of growth of glucose-6-phosphatase-deficient, preneoplastic foci in DEN-treated BR females was significantly higher than in B6 females. The growth rates of hepatic foci in BR and B6 males were similar but foci in BR males were 5-fold more numerous than in B6 males. The high sensitivity of BR females may be due, at least in part, to the failure of ovarian hormones to inhibit the growth of preneoplastic foci and the subsequent development of liver tumors. Since BR males had a larger number of hepatic foci, it is likely that androgens increase the rate of focus formation in BR males. Introduction The hormonal environment of the host affects the development of many types of tumors. For example, in humans and rodents, the growth of mammary cancer is greatly increased by the presence of estrogen and prolactin (1-6) and the growth of prostatic cancer is promoted by androgens (7,8). Although the liver is not a reproductive organ, the sensitivity of inbred mice to hepatocarcinogenesis is also affected by the hormonal status of the host (9-17). In addition to being markedly more sensitive to spontaneous liver tumor induction, male mice are significantly more susceptible than females to the development of liver tumors induced by perinatal treatment with carcinogens Abbreviations: B6, ; BR, ; C3H, C3H/HeJ; DEN, Nfldiethylnitrosamine; T,j, apparent doubling time. Oxford University Press (9-12). This difference in sensitivity has been shown to be independent of the type of carcinogen used and occurs over a wide range of doses (10-). The sexual dimorphism in murine hepatocarcinogenesis results from the contrasting effects of androgens and ovarian hormones on liver tumor induction (9,11). Thus, the castration of male mice results in a decrease in mean liver tumor multiplicity (9,12,13,) while the yield of tumors increases following the ovariectomy of females (9, 17). Previous studies have shown that the removal of circulating sex hormones by gonadectomy of B6C3Fj mice resulted in a similar tumor incidence for males and females (9). We demonstrated previously that promotion of liver tumor induction by testosterone in mice requires the presence of a functional androgen receptor by comparing hepatocarcinogenesis in Testicular feminization (Tfin) mice, which are mutant for the X-linked androgen receptor gene, to that in wild type (B6*) male mice (18,19). The analysis of androgen receptor expression in tumors from Tfm/+ mosaic female mice treated with testosterone revealed that the androgen receptor need not be present in the target cell in order for a tumor to develop (19). The inhibition of hepatocarcinogenesis in female mice is likely to be the result of direct or indirect effects of estrogen on the development of liver tumors. Lee et al. (20) demonstrated that treatment of C3H, B6, and BALB/c male mice with high doses of estrogen inhibited liver tumor induction. Studies from Yamamoto and coworkers (17) provide more direct evidence; chronic treatment of ovariectomized female mice with estradiol, but not progesterone, decreased the yield of tumors to that observed for intact females. There is a broad range of susceptibility to hepatocarcinogenesis among the males of different strains of mice (21). When treated with a single perinatal dose of a carcinogen, sensitive C3H/HeJ (C3H) male mice are times more susceptible to liver tumor induction than resistant B6 males (22). In contrast, the inhibitory effect of ovarian hormones results in a narrow range of sensitivity among the females of the same strains; C3H females are only 3 times as susceptible as B6 females and are just 3% as sensitive as C3H males (22). The (BR) mouse strain is unique because, compared to males of other strains of mice, BR males are intermediate in susceptibility to hepatocarcinogenesis, while BR females are extremely susceptible, with a 20- to 40-fold higher mean tumor multiplicity than the females of any other strain (21). Although the BR females are much more sensitive than the females of other strains, they are still relatively resistant when compared to the BR males. The basis for the extreme sensitivity of the BR females is unknown and was unexpected given the intermediate sensitivity of the males. Determining the biological basis for the unusual tumor sensitivity phenotype of the BR females will help gain further insight into the mechanism(s) by which the growth of liver tumors is influenced by sex hormones in inbred mice. In order to determine the strain-dependent effects of sex hormones on liver tumor induction, we compared tumor 191

2 TJVtPoole and N.R-Drinkwater multiplicity in intact and gonadectomized BR, B6, and C3H male and female mice treated with DEN. Since histochemicallyaltered liver foci are thought to be precursors of liver tumors (23), we also studied the growth of glucose-6-phosphatase (G6Pase)-deficient foci in order to evaluate the effects of sex hormones on the promotion phase of hepatocarcinogenesis. In the experiments reported here, we demonstrate that ovarian hormones inhibited hepatocarcinogenesis in C3H and B6 female mice but not in BR females. In contrast, liver tumor induction was promoted by androgens in all three strains. We observed a difference in the rate of growth of hepatic foci between the BR and B6 females that could account for the increased sensitivity of the BR females to hepatocarcinogenesis. BR males are more susceptible to hepatocarcinogenesis than B6 males despite exhibiting similar growth rates in hepatic foci. This difference in sensitivity may result from a higher rate of focus formation in the livers of the BR males. Unexpectedly, the growth rates of preneoplastic lesions were similar in intact BR males and females indicating that androgens may not influence the growth rate of preneoplastic foci, but rather serve to increase the rate of formation of preneoplastic foci in BR mice. Materials and methods Mice Mice were bred in our laboratory from stocks of, and C3H/HeJ mice purchased from the Jackson Laboratory (Bar Harbor, ME). The mice were housed in plastic cages on corncob bedding (Bed O'Cobs, Anderson Cob Division, Maumee, OH) and fed Wayne Breeder Blox (Continental Grain Co., Chicago, IL). Food and acidified tap water were available ad libitum. Mice were inspected daily and weighed monthly. Tumor induction study Twelve-day-old male and female, and C3H/HeJ mice were injected i.p. with DEN (0.05 nmol/g body wt) (Eastman Kodak Co., Rochester, NY) dissolved in sterile trioctanoin (0.01 rnl/g) (Pfaltz and Bauer, Inc., Stamford, CT). One group of mice from each of the three strains was gonadectomized, as described below, at 8 weeks of age. All intact and castrated males were sacrificed at weeks and all intact and ovariectomized females were sacrificed at 50 weeks of age. Mice were euthanized by CO2 asphyxiation and the livers were removed and weighed. All tumors >2 mm on the surface of the liver were counted. A random sample of tumors from each group were excised, fixed in phosphate buffered formalin and later examined microscopically in thin sections stained with hematoxylin and eosin. Gonadectomy Gonadectomies were performed as described previously (). Sterile gloves and instruments were used throughout the procedure. Eight-week-old mice were anesthetized using a mixture of Ketaset (0.04 mg/g body wt) (Aveco, Fort Dodge, LA) and PromAce (0.025 mg/g body wt) (Aveco, New York, NY). The skin was shaved, washed with Betadine (Purdue Frederick, Norwalk, CT) and rinsed with 70% ethanol. A single incision was made in the skin, and a small incision was made in the peritoneum close to the gonad. In females, both ovaries and part of the uterine horn were tied off with 5-0 silk suture (Medix, Madison, WT) and removed. In males, the testes and the vas deferens were tied off and removed. The incision in the peritoneum was closed with a suture and the skin was closed with two wound clips (VWR, Chicago, IL). Analysis of G6Pase-deficient foci Twelve-day-old male and female and mice were injected i.p. with DEN (0.05 pmol/g body wt) dissolved in sterile trioctanoin. Groups of animals were sacrificed by COj asphyxiation at,,, and 40 weeks of age. The livers were removed quickly and weighed. Four liver sections were taken from the two largest lobes, placed on filter papet, and frozen on solid CO2. Thin cryostat sections were stained for glucose-6-phosphatase (C-6Pase) and analyzed as described previously (25). G6Pase-deficent foci were quantified by microscopic analysis with a calibrated reticule. The total area of the liver section was determined with an electronic planimeter (Keuffel and ESSCT, Hoboken, NJ). The mean number of cells per lesion was estimated from the ratio of the volume fraction of the hepatic foci to the number of foci per cm 3 of liver. The value 5X10"" 9 cm 3 was used as the mean volume of a single hepatocyte (25). 192 Table I. Effect of gonadectomy on liver tumor induction in male and female C3H/HeJ, and mice treated perinatally with DEN 8 Strain Females C3H/HeJ Males C3H/HeJ Intact Number of mice Mean liver tumor mult. c 1.4 (4.7) 0.5 (1.0) 28(13) 66(34) 7.4(13) 26(18) Gonadectomized b Number of mice Mean liver tumor mult. 10(8) 4.1 (6.6) 35 () () 0.5 (0.9) 6.1 (10) "Mice were treated at 12 days of age with iv.w-diethylnitrosamine (DEN) (0.05 imol/g body wt). The number of tumors was determined at 50 weeks in females and weeks in males. b Gonadectomies were performed at 8 weeks of age as described in the Materials and methods section. c Values are mean (SD). Statistical analyses The yield of tumors from each group was compared to the others using the Wilcoxon rank sum test (26). The same test was used for comparison of the treatment groups for the number and average size of G6Pase-deficient hepatic foci. The frequencies of hepatic adenomas and carcinomas among the groups were compared statistically with Fisher's exact test (27). Results Effects of gonadectomy on liver tumor induction The mean tumor multiplicity at 50 weeks of age in intact BR females (28 ±13) was 20 and 56 times greater than that in intact C3H (1.4 ± 4.7) and B6 females (0.5 ± 1.0), respectively (Table I). Ovariectomy resulted in a large increase in the mean liver tumor multiplicity in C3H and B6 females but not in the BR females (Table I). An ~8-fold increase in mean tumor multiplicity was observed in both B6 and C3H ovariectomized female mice as compared to the intact females of the same strain (P < 4X10" 6, two-sided Wilcoxon rank sum test). The mean tumor multiplicity in BR females after ovariectomy was only 25% higher than that in intact BR females (P < 0.02). These observations indicate that BR females are less responsive to the inhibitory effects of ovarian hormones on hepatocarcinogenesis than the other two strains. Intact -week-old C3H male mice had a mean tumor multiplicity of 66 ± 34 which was ~9-fold higher than that in intact B6 male mice (7.4 ± 13) (P < 6XIO- 9 ) and -2.5 times greater than that in the intact BR males (26 ± 18) (P < 2X10" 6 ) (Table I). A decrease in the mean tumor multiplicity was observed following castration of the male mice of all three strains, although the magnitude of the decrease varied among the strains. The mean tumor multiplicity of castrated C3H males decreased to -33% of the mean value for the intact C3H males while the mean tumor multiplicity for B6 and BR castrated males decreased to 6.7% and 25% of mean for the intact males, respectively. We sacrificed female mice at an older age than males in the above experiment because of the delayed development of liver tumors in females (28). At weeks of age (see experiment in Table H), the mean numbers of liver tumors observed in intact female B6, C3H, and BR mice were 0.19 ± 0.39 ( mice), 0.13 ± 0.33 ( mice), and 0.53 ± 1.1 ( mice),

3 Hormonal regulation of hepatocarcinogenesis Table II. Development of G6Pase-deficient foci in male and female BR and B6 mice treated with DEN 1 Strain Sex Age (weeks) Number of mice Number of foci/liver b Mean focus diameter (mm) c Fraction of liver volume occupied by foci M M (220) 190(180) 0(130) 330 (260) 38(54) 99(120) 270 (380) 190 (200) 430 (340) 680(410) 770 (230) 93 (130) 110(0) 62 (70) 205 (0) (0.003) (0.007) (0.0) 0. (0.111) (0.0002) (0.0005) (0.002) (0.003) (0.005) (0.078) 0.17 (0.087) (0.0005) (0.006) (0.009) (0.059) Twelve-day-old mice were injected i p. with DEN (0.05 imol/g body wt). Hepatic foci were analyzed a:,,, and 40 weeks of age. b Values are mean (SD). TTie mean focus diameter was determined by pooling the data from all of the animals in the group and considering the liver section as one composite liver. These data were truncated to exclude all foci <0.135 mm in diameter. respectively. The dose of DEN used in this study was within the approximately linear range of the dose-response for liver tumor induction; treatment of intact B6, C3H, or BR mice of either sex with 0.1 (imol DEN/g body wt resulted in approximately twice the yields of liver tumors shown in Table I (Poole and Drinkwater, unpublished observations). Histological analysis of 6 formalin fixed tumors randomly chosen from each group of animals in the gonadectomy experiment demonstrated that 127 (77%) of the tumors were hepatocellular adenomas, 25 (%) were hepatocellular carcinomas (HCC), and (8%) were mixed adenomas and carcinomas. The distribution of each type of lesion was similar among the strains. However, gonadectomy did appear to increase the incidence of hepatocellular carcinomas that formed in the liver. Of 39 HCC, 34 (87%) were found in gonadectomized animals (P < 0.0, Fisher's Exact Test). Induction of G6Pase-deficient foci The number, size, and volume fraction of G6Pase-deficient foci in the livers of male and female BR and B6 mice treated with DEN are shown in Table II. G6Pase-deficient foci were ~2-fold more numerous in the BR females (190 ± 180 at weeks) than in the B6 females (99 ± 120) (P < 0.05) at all comparable times except weeks. We observed a 3-fold increase in focus diameter over time in both BR and B6 female mice from to 40 weeks of age. However, at each time point, the hepatic foci were significantly larger (P < 0.02) in the BR females than in the B6 females. Although the percent liver volume containing G6Pase-deficient foci increased over time in both strains, the volume fraction occupied by foci was significantly greater (P < 0.002) in BR females than in B6 females at all timepoints. At weeks of age the percentage of liver volume occupied by G6Pase-deficient foci in BR females was 20-fold greater than that in B6 female livers and by 40 weeks the value was 75-fold higher in BR females ( ± 11%) compared to B6 females (0.2 ± 0.3%). Analysis of the change in the number of cells per focus over time revealed that the growth rate of hepatic foci in the BR females was higher than that in B6 females (Figure 1). The apparent doubling time, T d, for foci was significantly shorter in the BR 25 Age (wk) Fig. 1. Growth rate of hepatic foci in B6 females (filled circles), BR females (filled squares), B6 males (open circles), and BR males (open squares). The solid and dashed lines represent the regression curves for the focus growth in females and males, respectively. The vertical bars represent the standard deviation of the value. As noted in the text, mice were treated with A/^V-diethylnitrosamine at 12 days of age and hepatic foci were analyzed for glucose-6-phosphatase activity at,,, and 40 weeks of age. females (4.6 ± 2.1 weeks) than in the B6 females (10.2 ± 2.3 weeks). The number, size, and volume fraction of G6Pase-deficient foci in the livers of the male mice are also presented in Table II. Hepatic foci could not be quantitated at the 40 week timepoint in the BR males because nearly 100% of the liver stained negatively for G6Pase. The number of hepatic foci was significantly greater in BR males (770 ± 230 at weeks) than in B6 males (205 ± 0) at all timepoints (P < 2X10" 5 ). Analysis of the size of these preneoplastic foci revealed that they increased in diameter in both BR and B6 male mice. The foci in BR males at weeks were ~3-fold larger than those in B6 males (P < ), but by weeks this difference 193

4 T.M.Poole and N.R.Drinkwater was not significant {P < 0.39). The percentage of the liver occupied by foci increased in both strains over time, but the percentage in BR males was significantly higher than that in the B6 males at all timepoints (P < 10~ 5 ). By determining the change in the number of cells per focus over time in the livers of the males, we observed that the growth rate of lesions was similar in BR and B6 males (Figure 1). The apparent doubling time, T d, for focal growth in BR males was 3.3 ± 0.8 weeks and in B6 males it was 3.5 ± 0.5 weeks. Strikingly, the rate of growth of hepatic foci in the BR males was similar to that in the BR females. Discussion Although we observed an 8-fold increase in tumor multiplicity in both ovariectomized C3H and B6 females compared to their intact counterparts, the average number of liver tumors in ovariectomized BR females increased only 25% over that in the intact BR females. Thus, ovarian hormones strongly inhibited hepatocarcinogenesis in C3H and B6 female mice, consistent with previous studies in B6C3F, females (9,12), while only weakly suppressing the development of liver tumors in BR females. Therefore, BR mice must carry sensitivity genes that overcome the inhibitory effect of ovarian hormones. Interestingly, the tumor multiplicity in ovariectomized C3H and B6 females was still much lower than that in intact or ovariectomized BR females. If the genetic background were the sole determinant of the difference in sensitivity among the strains, one would expect that C3H would be the most sensitive of the ovariectomized animals since that is what is observed in both castrated and intact males. However, the gonadectomies were not performed until 8 weeks of age and inhibitory effects of ovarian hormones prior to that time could account for the low tumor yield in ovariectomized C3H and B6 females. Previous studies support this conclusion; when groups of B6C3F] females were ovariectomized at various ages and the tumor multiplicity for each group determined at the same endpoint, a differential inhibition of hepatocellular adenomas was observed. Ovariectomy of females at later timepoints resulted in fewer tumors than in those mice in which the surgery was performed earlier (28), indicating that the degree to which tumor formation was suppressed correlates with the length of time that the ovarian hormones were present in the body. Our studies do not address the mechanisms by which ovarian hormones inhibit hepatocarcinogenesis in females of typical inbred mouse strains. It is likely that the relevant ovarian hormone is estrogen, based on the observation that this hormone inhibits liver tumor induction when administered chronically to male (20) or ovariectomized female (17) mice. The inhibition of liver tumor development could result from the direct effects of estrogen on the target hepatocyte or from an indirect mechanism involving the estrogen-dependent regulation of a hormone or growth factor produced by another tissue. We have shown that the amounts and affinities of estrogen receptor in the livers of B6 and BR female mice are similar (29). Yamamoto et al. (17) have argued that the effects of estrogen on liver tumor induction are indirect, based on the observation that subcutaneous, but not intrasplenic, implantation of slowrelease estrogen pellets inhibited tumor development in ovariectomized female mice. This question could be addressed more straightforwardly by an approach analogous to that used to demonstrate that the promotion of hepatocarcinogenesis by androgens is not cell-autonomous (19). Mice mosaic for the expression of functional estrogen receptor could be constructed by preparing chimeric embryos from wild type mice and mice carrying an insertional mutation in the estrogen receptor locus (30). We observed a significantly higher net growth rate of G6Pase-deficient foci in the BR females compared to the B6 females indicating that the apparent rate of growth of preneoplastic foci is higher in BR females. We have previously shown that hepatic foci grow at similar low rates in C3H and B6 females (25). The difference in the rate of growth of foci between BR and B6 females may reflect the effect of the ovarian hormones since tumor multiplicity increases to the same degree in B6 and C3H females following ovariectomy. Thus, BR females are deficient in their ability to effectively suppress the growth of preneoplastic lesions, leading to enhanced tumor development. One test of this model would be to compare the growth rate and number of G6Pase-deficient foci in intact and ovariectomized B6, C3H, and BR females. If the hypothesis is correct, then one would expect to see an increased growth rate of hepatic foci in ovariectomized B6 and C3H females, but not in BR females, compared to the intact mice. Goldfarb and Pugh have shown previously that ovariectomy accelerated the growth of preneoplastic and neoplastic hepatocellular lesions in B6C3F, mice (). Tumor multiplicity was lower in castrated male C3H, B6 and BR mice compared to the intact males, indicating that androgens promoted the development of liver tumors in all three strains. The difference in sensitivity to hepatocarcinogenesis between BR and B6 males cannot be explained simply by a difference in the net growth rate of hepatic foci because the growth rate of preneoplastic foci was found to be similar in BR and B6 males. The number of hepatic foci at each timepoint was significantly higher in BR males than in B6 males suggesting that the difference in sensitivity between the strains lies in the process of focus formation. The disparity in the number of observed preneoplastic foci in the livers of the males of the BR and B6 strains could be caused by a difference in the level of initiation. If the level of DNA damage following carcinogen treatment, or the efficiency with which such damage yielded fixed mutations, were higher in BR males, then a greater number of cells could become preneoplastic. However, DNA damage following DEN treatment has been shown to be similar in C3H and B6 mice, yet the livers from DEN-treated C3H male mice contained many more foci than did the livers from DEN-treated B6 males (22). In addition, BR males are susceptible to both spontaneous (31) and yv-ethyl-jv-nitrosourea-induced hepatocarcinogenesis (21), which argues against a carcinogen-dependent sensitivity. Although the multistage model for carcinogenesis, which divides the process into initiation, promotion, and progression phases, was first developed to describe tumor induction in mouse skin (), this model has also been shown to apply broadly to hepatocarcinogenesis (23). Boutwell () has proposed that the promotion stage of skin carcinogenesis can be subdivided further into a conversion step that provides the initiated cell with the capacity to proliferate rapidly and respond to promoting agents and a propagation stage, which is characterized by the clonal proliferation of the preneoplastic cell. The conversion-propagation model for promotion may also be applied to hepatocarcinogenesis. Thus, even though the rate of growth of preneoplastic foci after DEN treatment is similar in B6 and BR males, a difference in the ability to become responsive to promotional stimuli (conversion) could 194

5 Hormonal regulation of hepatocardnogenesis account for the difference in the number of hepatic foci. Furthermore, androgens may affect this conversion step, since, despite the similar growth rate of preneoplastic lesions in BR males and females, the livers of BR males contain many more foci than those of BR females. Our previous observation that a partial hepatectomy acted as a promoter of liver tumor induction in B6 males but not in C3H males provides evidence consistent with this model (33). Spontaneous conversion could occur more readily in C3H males than in B6 males, while an exogenous stimulus may be required for tumor formation in B6 mice. The observation in the present study that the number of foci increased in an age-dependent manner in BR male mice is also consistent with this model. Previously, we proposed that androgens strongly interacted with the Hepatocarcinogen sensitivity (Hcs) gene identified as the major determinant of the difference in sensitivity between C3H and B6 male mice (22). Therefore, it was expected that tumor multiplicity in castrated C3H and B6 male mice would be more similar, as had been observed for the female mice. In this study, however, we demonstrated that castrated C3H male mice still were significantly more sensitive to liver tumor induction than castrated B6 male mice. Furthermore, B6 mice which carry resistance alleles at the Hcs locus are still susceptible to the promoting effects of androgens. The genetic effects on the sensitivity of C3H are complex and several other Hcs genes have now been identified (34,35); it is likely that each Hcs gene will affect a distinct stage of the carcinogenic process. It was somewhat surprising that the BR mice are sensitive to tumor induction since these mice are closely related to B6. The BR and B6 strains were generated from two siblings in a single cross (36) so 25% of their genes are identical by descent (37). In genetic mapping experiments, we have demonstrated that two loci are the major genetic factors contributing to the sensitivity of both the BR females and males (Poole and Drinkwater, submitted for publication). These gene mapping studies strongly suggest that the same underlying pathway is responsible for the sensitivity of both BR males and females to hepatocarcinogenesis. It is possible that the products of these genes affect the growth rate of hepatic tumors since the rate of growth of preneoplastic foci is nearly identical in BR males and females. In addition, ovariectomized and intact BR female mice and castrated BR male mice, sacrificed at the same timepoint had similar mean tumor multiplicities (data not shown), which supports the hypothesis that a similar tumor induction pathway is followed in the absence of sex hormones. It is unclear whether our studies of murine hepatocarcinogenesis reflect the mechanisms of liver tumor induction that occur in humans. Analysis of the factors influencing hepatocarcinogenesis in humans is complicated by the association of Hepatitis B Virus (HB V) infection with the development of hepatocellular carcinomas (38). Men develop liver cancer two to five times more often than women, but die higher incidence of this cancer in males may result from die fact diat more men are chronically infected with HBV dian women (39). Several reports provide evidence diat bodi anabolic steroids and exogenous estrogens in die form of oral contraceptives are associated widi a high risk for die development of hepatocellular adenomas (40-43). The molecular basis for die modulation of liver tumor development by ovarian hormones is unknown. The BR strain provides a unique system in which to identify padiways which overcome die inhibitory effect of ovarian hormones on liver tumor development. Determining die genetic basis for die sensitivity of die BR females will provide a foundation for understanding the pathway(s) responsible for die resistance to die suppressing effects of ovarian hormones in BR mice and may give clues as to how ovarian hormones inhibit tumor development in otiier strains. Acknowledgements The authors would like to thank Mary Winkler and Dona] Kaehler for assistance with the animal work, Paul Merline and Jane Weeks for preparation of the tissue sections, Dr Gang-Hong Lee for the histological analysis, Teresa Chiaverotti for helpful discussions, and Susan Schadewald and Rey Carabeo for critical comments on the manuscript. This work was supported by Public Health Service Grants CA284, CA07175, CA References l.henderson.b.e., Rossjl. and Bemstein,L. (1988) Estrogens as a cause of human cancer. Cancer Res., 48, sborne,K.C. and Lippman,M.E. (1978) Human breast cancer in tissue culture: the effects of hormones. In McGuire.W.L. (ed.), Breast Cancer. Plenum Medical Book Company, New York, pp MeitesJ. (1972) The relation of estrogen and prolactin to mammary tumorigenesis in the rat. In Dao,T.L. (ed.), Estrogen Target Tissues and Neoplasia. University of Chicago Press, Chicago, pp Sinha,D., Cooper.D. and Dao.T.L. (1973) The nature of estrogen and prolactin effect on mammary tumongenesis. Cancer Res., 33, SchardeinJ.L. (1980) Studies of the components of an oral contraceptive agent in albino rats. I. Estrogenic component J. ToxicoL Em. Health, 6, SchardeinJ.L. (1980) Studies of the components of an oral contraceptive agent in albino rats. II. Progestogenic component and comparison of effects of the components and the combined agent. /. Toxicol. Env. Health, 6, Wilding,G. (1992) The importance of steroid hormones in prostate cancer. Cancer Surveys,, Webber,M.M. (1981) Polypeptide hormones and the prostate. In The Prostatic Cell: Structure and Function Part B. Alan R.Liss Inc., New York, pp Vesselinovitch,S.D. and Mihailovich,N. (1967) The effect of gonadectomy on the development of hepatomas induced by urethan. Cancer Res., 27, Vesselinovitch,S.D., Rao.K.V.N., Mihailovich.N., RiceJ.M. and Lombard, L.S. (1974) Development of broad spectrum of tumors by ethylnitrosourea in mice and the modifying role of age, sex, and strain. Cancer Res., 34, ll.vesselinovitch.s.d. (1969) The sex-dependent difference in the development of liver tumors in mice administered dimethylnitrosamine. Cancer Res., 29, Vesselinovitch,S.D., Itze,L. Mihailovich,N. and Rao,K_V.N. (1980) Modifying role of partial hepatectomy and gonadectomy in ethylnitrosourea-induced hepatocarcinogenesis. Cancer Res., 40, Toh,Y.C. (1981) Effect of neonatal castration on liver tumor induction by Af-2-fluorenylacelamide in suckling BALB/c mice. Carcinogenesis, 2, Vesselinovitch,S.D., Mihailovich^J. and Pietra,G. (1967) The prenatal exposure of mice to urethan and the consequent development of tumors in various tissues. Cancer Res., 27, Yamamoto,R., Iishi.H., Tatsutajvl., Tsuji,M. and Terada,N. (1991) Roles of ovaries and testes in hepatocellular tumorigenesis induced in mice by 3'-methyl-4-dimethylaminoazobenzene. Int. J. Cancer, 49, Goldfarb,S. and Pugh.T.D. (1990) Ovariectomy accelerates the growth of microscopic hepatocellular neoplasms in the mouse: possible association with whole body growth and fat depostion. Cancer Res., 50, Yamamoto,R., Tatsuta>I. and Teradajvl. (1993) Suppression by oestrogen of hepatocellular tumorigenesis induced in mice by 3'-methyl-4- dimethylaminoazobenzene. Br. J. Cancer, 68, Kemp,CJ. and Drinkwater,N.R. (1990) The androgen receptor and liver tumor development in mice. In StevensonJD.E., PoppJ.A., WarxLJ.M., McClain,R.M., Slaga,TJ.,and Pitot,H.C. (eds), Mouse Liver Carcinogenesis: Mechanisms and Species Comparisons. Wiley-Liss New York, pp Kemp,CJ., Leary.C.N. and Drinkwater.N.R. (1989) Promotion of murine hepatocarcinogenesis by testosterone is androgen receptor-dependent but 195

6 TM.Poole and N.R.Drinkwater not cell autonomous. Proc. Natl Acad. Sci. USA, 86, O.Lee,G.-H., Nomura,K. and Kitagawa,T. (1989) Comparative study of diethylnitrosamine-initiated two-stage hepatocarcinogenesis in C3H, C57BL and BALB mice promoted by various hepatopromoters. Carcinogenesis, 10, Kemp,CJ. and Drinkwater,N.R. (1989) Genetic variation in liver tumor susceptibility, plasma testosterone levels, and androgen receptor binding in six inbred strains of mice. Cancer Res., 49, Drinkwater,N.R. and GinslerJJ. (1986) Genetic control of hepatocarcinogenesis in and C3H/HeJ inbred mice. Carcinogenesis, 7, Pitot,H.C. and Sirica,A.E. (1980) The stages of initiation and promotion in hepatocarcinogenesis. Biochem. Biophys. Ada, 605, Waynforth,H.B. (1980) Experimental and Surgical Technique in the Rat. Academic Press, New York, pp Hanigan,M.H., Kemp.CJ., GinslerJJ. and Drinkwater.N.R. (1988) Rapid growth of preneoplastic lesions in hepatocarcinogen-sensitive C3H/HeJ male mice relative to male mice. Carcinogenesis, 9, Kendall,M.G. and StewarU- (1973) 77ie Advanced Theory of Statistics. Hafner, New York, vol Sokal,R.R. and Rohlf.FJ. (1995) Biometry. Freeman, New York. 28.Vesselinovitch,S.D. (1990) Perinatal mouse liver carcinogenesis as a sensitive carcinogenesis model and the role of the sex hormonal environment in tumor development. In Stevenson.D.E., PoppJ.A., WardJ.M., McClain.R.M., Slaga,TJ. and Pitot,H.C. (eds), Mouse Uver Carcinogenesis: Mechanisms and Species Comparisons. Wiley-Liss New York, pp Poole.T.M. and Drinkwater,N.R. (1995) Hormonal and genetic interactions in murine hepatocarcinogenesis. In McClain.R.M., Slaga.TJ., LeBoeuf.R. and Pitot.H. (eds), Growth Factors and Tumor Promotion. Wiley-Liss, New York, pp Lubahn,D.B., MoyerJ.S., Golding.T.S., CouseJ.F, Korach.K.S. and Smithies.O. (1993) Alteration of reproductive function but not prenatal sexual development after insertional disruption of the mouse estrogen receptor gene. Proc. Natl Acad. Sci. USA, 90, StorerJ.B. (1966) Longevity and gross pathology at death in 22 inbred mouse strains. J. Gerentol., 21, Boutwell.R.K. (1964) Some biological aspects of skin carcinogenesis. In Karger.A. (ed.). Progress in Experimental Tumor Research, vol. 4, pp Hanigan,M.H., Winkler.M.L. and Drinkwater.N.R. (1990) Partial hepatectomy is a promoter of hepatocarcinogenesis in male mice but not in C3H/HeJ male mice. Carcinogenesis, 11, Gariboldi,M., Manenti.G., Canzian.F, Falvella,F.S., Pierotti.M.A., Delia Porta,G., Binelli.G. and Dragani.T. (1993) Chromosome mapping of murine susceptibility loci to liver carcinogenesis. Cancer Res., 53, Manenti,G., Binelli.G., Gariboldi.M., Canzian,F., DeGregorio,L., Falvella,F.S., Dragani.T.A. and Pierotti.M.A. (1994) Multiple loci affect genetic predisposition to hepatocarcinogenesis in mice. Cenomics, 23, Festing.M.F.W. (1979) Inbred Strains in Biomedical Research. Oxford University Press, New York. 37.Htch,W.M. and Atchley.W.R. (1985) Evolution in inbred strains of mice appears rapid. Science, 228, Wilkinson.M.L. (1987) Sex steroids and primary hepatocellular carcinoma. Anticancer Res., 7, Taylor.W. (1987) Risk factors associated with the use of sex hormones. Anticancer Res., 7, BaumJ.K., BooksteinJJ., Holtz,F. and Klein.E.W. (1973) Possible association between hepatomas and oral contraceptives. Lancet, 2, I.Contostavlos,D.L. (1973) Benign hepatomas and oral contraceptives. Lancet, 2, Edmondson.H.A., Reynolds.T.B., Henderson,B. and Benton.B. (1977) Regression of liver cell adenomas associated with oral contraceptives. Ann Internal Med, 86, YagerJ.D., ZurloJ. and Ni,N. (1991) Sex hormones and tumor promotion in liver. Proc. Soc. Expt. BioL Med, 198, Received on October 20, 1995; revised on November. 1995; accepted on November 29,