Two-Stage Liver Carcinogenesis in the Mouse*

SYMPOSIUM Agents and Processes in Chemical Carcinogenesis TOXICOLOGIC PATHOLOGY ISSN:0192-6233 Volume 15, Number 2, 1987 Copyright 0 1987 by the Society of Toxicologic Pathologists Two-Stage Liver Carcinogenesis in the Mouse* GIUSEPPE DELLA PORTA, TOMMASO A. DRAGANI, AND GIACOMO MANENTI Dirisioiie di OncoIogia Sperirnentale A, Istittito Nazionaleper lo Studio e la Cirra dei Tirnrori, Via G. Venezian 1-20133 Milano, Italy ABSTRACT A two-stage protocol for studying liver carcinogenesis was applied to the mouse. The protocol includes the treatment of 7-day-old mice with a single low dose of an initiating agent (diethylnitrosamine, NDEA), promotion starting after weaning and lasting about 20 weeks, and histologic analysis, at 30 weeks of age, of hepatocellular nodules on H&E stained sections. A stereologic analysis of results allows the evaluation of nodule frequency and size. Using this protocol in and mice, we have identified the promoting activity of the phenobarbital-like enzyme inducer, 1,4-bis[2-(3,5-dichloropyridyloxy)]benzene (TCPOBOP), which was found to be a strong hyperplaseogenic agent for mouse liver. These studies also indicated that the different susceptibility to hepatocarcinogenesis in and mice may be related to a higher susceptibility of than initiated liver cells to growth stimulation. A long-term study showed that mice have a low incidence of spontaneous liver tumors but are susceptible to chemical hepatocarcinogenesis and, therefore, they may be an alternative model to mice in carcinogenesis bioassays. INTRODUCTION The two-stage nature of tumor development was demonstrated in the skin of rodents by applying a single subthreshold dose of an initiating carcinogen (DMBA) followed by the continuous treatment with the promoter croton oil (2, 3). The single subthreshold dose of the initiating carcinogen or the continuous treatment with the promoter alone were not able to induce tumors, but the combined treatment did (2,3). In the last two decades it has become clear that the two-stage phenomenon in the development of neoplasms is not unique to the skin. Multi-stage induction of tumors was observed in other organs and tissues such as liver, lungs, mammary gland, and urinary bladder, and different tumor promoters have been identified (1, 10, 11, 12, 19, 22). The liver model of carcinogenesis seems to be particularly useful for studying the different phases of tumor development. Since up to 80% of liver mass is constituted by a single cell type, the hepatocyte and the organ is quite well characterized biochemically in its many functions. Moreover, the early phases of multi-stage hepatocarcinogenesis are characterized by preneoplastic lesions, which can be induced in a few months. Different systems are available to induce preneoplastic lesions in a short time or tumors after a longer period * Presented at the Third Sardinian International Symposium, October 6-9, 1985 in Cagliari, Italy. of observation. The systems consist of a single treatment with an initiating carcinogen, usually NDEA, during or before liver cell replication. Then a promoting agent such as phenobarbital (PB), or a choline deficient diet, or a toxic environment (AAFdiet) that inhibits proliferation of normal hepatocytes is administered, which enables identification ofenzyme-altered foci that are considered to be preneoplastic lesions (9, 12). GENETIC SUSCEPTIBILITY TO HEPATOCARCINOGENESIS These systems have been studied mostly in the rat, whereas the mouse has been used in a few studies only. This species seems particularly useful to investigate the genetic mechanisms controlling tumor development because murine strains are available with a high or low incidence of spontaneous liver tumors. As shown in Table I, murine strains like C57BLl 65 and SWR do not develop liver tumors, even after 2 years of life. BALB/c mice have a low incidence of liver tumors, whereas C3Hf mice have a high incidence (4, 5). To better define the mouse model of hepatocarcinogenesis, we carried out a long-term experiment to compare the spontaneous tumor incidence in two related hybrids: and mice, both derived from the F1 cross between C57BL/6J females and either C3Hf or BALB/c males. mice are com- 229

~ ~ ~~ 230 DELLA PORTA ET AL TOXICOLOGIC PATHOLOGY TABLE 1.-Incidence of spontaneous liver tumors in different murine strains after observation until 120 weeks of age. Mice with liver tumors Strain or hybrid Sex mice No. (Yo) CS 7BL/6Ja M 50 F 50 C3HP M 30 20 67 SWR" M 45 F 47 F 63 15 24 BALB/cb M 173 16 9 F 167 3 2 From reference (4). From reference (5). monly used in carcinogenesis bioassays carried out under the National Toxicology Program sponsorship in the United States. As shown in Table 11, and as reported by other authors (20), the hybrid inherited from the C3Hfparental generation a relatively high incidence of spontaneous liver tumors. Therefore, the high incidence of spontandous liver tumors acts as a dominating trait. In contrast, mice showed a low incidence of spontaneous liver tumors (8). To determine the susceptibility of the two hybrids to a weak multipotential carcinogen, urethan, we carried out the experiment shown in Table 11. Urethan is an effective liver carcinogen when administered during infancy even for short periods oftime. Urethan treatment induced a high percentage of liver tumors in both hybrids. However, mice survived longer than mice. For this reason the incidence of liver tumors seems higher in mice, but a statistical analysis, which takes the survival into consideration, showed that mice were more susceptible than mice. Our study showed that mice have a low incidence of spontaneous liver tumors, but they are susceptible to chemical hepatocarcinogenesis and, therefore, they may be an alternative to in carcinogenesis bioassays (8). TWO-STAGE HEPATOCARCINOGENESIS In studying mouse hepatocarcinogenesis, it must be taken into account that this model differs in some aspects from the skin model. A single low dose of an initiating agent does not result in tumor development in the skin even after a prolonged period of observation; whereas a single dose of diethylnitrosamine at 1.25 mg/kg induced hepatocellular carcinomas in 15-day-old mice within 7O'weeks after treatment (18). Therefore, a progression of neo- TABLE 11.-Hepatocellular tumors in untreated or urethan-treated and mice." Hybrid Sex mice Treatment M 50 - M 48 - F 50 - F 49 - M 41 urethan M 46 urethan F 51 urethan F 50 urethan Mice with liver tumors No. Ratiob 13 1.8 2 0.3 3-26 2.6' 36 0.1 18 3.1= 22 0.2 From reference (8). Observedlexpected ratio for tumor incidence comparing the two hybrids, allowing for survival. =p < 0.01, two-tailed, between the two hybrids. plastic lesions must be operating in the mouse liver, which may be due to the presence of endogenous promoters of hepatocarcinogenesis. Some hormones have been found to promote liver carcinogenesis, and they may account for the well known sex difference in the susceptibility to hepatocarcinogenesis. Our studies were started on two-stage hepatocarcinogenesis choosing a protocol in which a single low dose of initiating agent is administered to 7- day-old mice. After an observation period of 3-6 months, the animals are sacrificed and the number and size of hepatocellular nodules are measured in hematoxylin and eosin (H&E) stained liver sections. The initiation treatment given during infancy (period of physiologic liver cell replication) reduces at minimum the manipulation of the system, and avoids stressing the animals with partial hepatectomy or treatment with necrogenic doses of toxins, as it is commonly done with the adult rat. The analysis of hepatocellular nodules on H&E slides was chosen because it was considered more reliable and had a clearer biologic significance than the analysis of enzyme altered foci, Noteworthy, the best enzymatic marker for preneoplastic foci in the rat liver (gamma-glutamyltranspeptidase) (9, 12) is a poor marker for mouse hepatocellular nodules (1 7). A stereological method for the analysis of data allows the calculation of nodule frequency (number of nodules/cm3 of liver) and size (mean volume of nodules) (I 5). In our study, as well as in others, most hepato: cellular nodules induced by treatment withsthe initiating agent alone (NDEA) were basophilic and may be scored starting from 14 weeks of age. In male mice, nodule frequency and size progressively increased from 14 to 24 weeks of age (Table 111).

Vol. 15, No. 2, 1987 LIVER CARCINOGENESIS 23 1 TABLE 111.-Hepatocellular nodules in male mice treated once with NDEA (15 mdkg PO) at 1 week of age, and observed until 9-24 weeks of age. Mean volume" Age mice ~o./cm3a (mm3 x lo3) 9 8 14 8 122 f 20 5+2 19 10 343 f 51 15 & 3 24 10 654 + 86 76 f 8 a Values represent mean? SE. Therefore, an observation period of about 20 weeks of age is advisable when using this model. In a first series of experiments we used PB and 1,4-bis[2-(3,5-dichloropyridiloxy)]benzene (TCPO- BOP) as promoters. PB is a well known promoter of rat hepatocarcinogenesis and in some experiments it was shown to promote mouse hepatocarcinogenesis as well (1 2, 16,2 1). PB stimulates a number ofenzyme activities involved in drug metabolism and also DNA synthesis in the liver, although the relevance of these effects on the promotion of hepatocarcinogenesis is still not clear (12). A potent PB-like agonist, with regard to induction of drug metabolism, has been recently described by Poland et a1 (13, 14). TCPOBOP is 650-fold more potent in mice than PB, but it is ineffective in rats. The compound has not been studied for promoting activity on hepatocarcinogenesis. In preliminary experiments, we confirmed the PBlike pattern of enzyme activity induced by TCPO- BOP in mouse liver, and we also observed a strong hyperplastic activity. As shown in Table IVY the P-450 associated enzyme activities p-nitroanisoleo-demethylase and aminopyrine-n-demethylase were induced by both PB and TCPOBOP; whereas, the P-448 associated UDPG-transferase activity (pnitrophenol as substrate) was not induced. The liver DNA content was almost doubled by TCPOBOP treatment (Table IV) (7). The results of the two-stage hepatocarcinogenesis experiment reported in Table V show that NDEA treatment of 7-day-old mice, followed by vehicle, resulted in about 400 nodules/cm3 of liver at 28 weeks of age in female mice. PB treatment after NDEA caused centrilobular hypertrophy with cells showing enlarged and acidophilic cytoplasm, resulted in a slight decrease of the number of nodules/cm3, and did not significantly change their mean volume with respect to controls. PB did not promote liver nodules probably due to the short period of administration or to an insufficient dose. Mice treated with TCPOBOP after NDEA had fewer nodules/ cm3 than controls, but the mean volume of nodules was higher. Most of the hepatocytes, throughout the lobules, were enlarged and acidophilic. Also, most of the nodules consisted of large acidophilic hepatocytes or ofclear cells. Therefore, TCPOBOP treatment stimulated the growth of hepatocellular nodules that were larger than those in controls. The reduced number of nodules/cm3 of liver in the TCPOBOP-treated group was due to nodule dilution caused by liver enlargement and to a confluence of small nodules in large ones (7). We have therefore identified TCPOBOP as a new promoter of mouse hepatocarcinogenesis. It may represent a valuable tool for the study of the mechanisms involved in the promotion phase of hepatocarcinogenesis and for the identification of a possible receptor that mediates its biologic activity. As already mentioned, the spontaneous incidence of hepatocellular tumors is high in mice and low in mice. To investigate whether the different susceptibility to hepatocarcinogenesis is due to a different susceptibility to initiation or promotion phases of hepatocarcinogenesis, both hybrids were treated with a single dose of NDEA at 1 week of age followed by TCPOBOP or vehicle and sacrificed at 30 weeks of age. The frequency of hepatocellular nodules was similar in the two hybrids. However, the mean volume of nodules was about 10-fold greater in than male mice receiving NDEA plus vehicle. The treatment with TCPOBOP after NDEA increased the mean volume ofnodules 5-8-fold in both hybrids, with mice still having a value about 7-fold greater than TABLE 1V.-Effects of 4-week treatment with PB and TCPOBOP on microsomal enzyme activities in female mice..? p-nitroanisole- Aminopyrine- UDPG- Liver DNA Treatment OrdemethylaSeb N-demethylaseb transferaseb (me) Vehicle 2.9 + 0.2 11.5 + 0.2 6.5 & 1.3 4.9 f 0.5 PB 10.8 f 0.7' 43.6 & 1.8' 8.9 & 0.9 6.0 f 0.6 TCPOBOP 14.5 & 0.4' 66.6 & 3.3' 8.1 k 0.3 9.2 i- 1.1' From reference (7). Values are mean,k SE, 6 animals per group. Units represent nmol productlmidmg protein. cp < 0.01 when compared with vehicle-treated group.

232 DELLA PORTA ET AL TOXICOLOGIC PATHOLOGY mice. No differences in mean volume of nodules were seen in the females ofboth hybrids after NDEA and vehicle, whereas upon administration ofndea and TCPOBOP, the mean volume of nodules was 25-fold greater in than 'in females. TCPOBOP alone induced the same level of DNA synthesis and of aminopyrine-n-demethylase in the liver of both hybrids. Therefore, the different susceptibility to hepatocarcinogenesis in and mice may be related to a higher susceptibility of than -initiated liver cells to growth stimulation. ALTERATIONS IN GEN EXPRESSION IN MOUSE LIVER TUMORS The expression of endogenous retroviral sequences and ofthree cellular oncogenes were studied in three hepatocellular adenomas and in four carcinomas induced in male mice by a single dose of NDEA, in five spontaneous carcinomas in male C3Hf mice, and in the livers of normal agematched control mice. In all of these adenomas and carcinomas there was increased expression of Moloney murine leukemia virus (M-MuLV)- and intracisternal A particle (1AP)-related sequences. The retrovirus-like VL30 sequence was expressed at significant levels in the normal liver of these mice and increased expression of this sequence was found in only four of the 12 tumors examined. Expression of endogenous mouse mammary tumor virus (MMTV)-related sequences was not detected in the normal livers nor in any of the liver tumors. With respect to cellular oncogenes, increased expression of c-niyc was seen in all of the tumors. Two out of five normal liver samples and all of the tumors of the C3Hf mice displayed significant expression of c-iizyc. There was a slight increase in expression of c-ha-ras in some ofthe tumors. Increased expression of c-fos was found in only one of the 12 tumors. Taken together, these studies indicate that both carcinogen-induced and spontaneous liver tumor formation in mice is associated with abnormalities in the expression of endogenous retrovirus-related DNA sequences and also specific cellular oncogenes (6). In conclusion, the mouse liver system may represent a good model to study the different phases of carcinogenesis. The sensitivity of infant liver to carcinogens may enable us to study the hepatocarcinogenicity of chemicals in a medium-term (20-30 weeks) in viw carcinogenesis assay. The studies on the genetic differences in the susceptibility to hepatocarcinogenesis between murine strains may lead to insights into the underlying mechanisms responsible for these differences. The understanding of these mechanisms may be of generic value, not limited TABLE V.-Hepatocellular nodules in female mice after various treatments." Hepatocellular nodules Treatment after Volume NDEA mice No./cm3 (mm3 x 10') Vehicle 7 419 f: 51 9k2 PB 10 311 +. 4gb 12 k 3 TCPOBOP 10 183 f 19"d 374 +. 124'~' From reference (7). Mice were treated once with NDEA (15 mg/kg PO) at 7 days of age and then with PB (500 ppm in the drinking water continuously) and TCPOBOP (3 mg/kg ip every week) starting from 8 weeks ofage, for 20 weeks; values represent mean k SE. *The single hepatocellular carcinoma observed (14 mm diameter) was excluded from tabulation. Including 11 nodules of 2.4-6.4 mm in diameter. dp < 0.01 with respect to the vehicle-treated group. to the particular mouse system, and may allow the identification of factors affecting tumor development. REFERENCES 1. Armuth V and Berenblum I (1974). Promotion of mammary carcinogenesis and leukemogenic action by phorbol in virgin female Wistar rats. Cancer Res. 34: 2704-2707. 2. Berenblum I and Shubik P (1947). A new quantitative approach to the study of stages of chemical carcinogenesis in the mouse's skin. 3r. J. Cattcer I: 383-391. 3. Boutwell RK (1964). Some biological aspects of skin carcinogenesis. Progr. Exp. Tirrrror Res. 4: 207-250. 4. Della Porta G, Capitano J, Parmi L, and Colnaghi MI (1967). Cancerogenesi da uretano in topi neonati, lattanti e adulti dei ceppi C57BL, C3H, BC3F1, C3Hf e SWR. Ttritiori 53: 81-102, 5. Dragani TA (1979). Analysis of tumor incidence in BALB/c mice used as controls in carcinogenicity experiments. Tirrrrori 65: 665-675. 6. Dragani TA, Manenti G, Della Porta G, Gattoni-Celli S, and Weinstein IB (1986). Expression of retroviral sequences and oncogenes in munne hepatocellular tumors. Cancer Res. 46: 19 15-1 9 19. 7. Dragani TA, Manenti G, Galliani G, and Della Porta G (1985). Promoting effects of 174-bis[2-(3,5-dichloropyridy1oxy)lbenzene in mouse hepatocarcinogenesis. Carcitiogenesis 6: 225-228. 8. Dragani TA, Sozzi G, and Della Porta G (1984). Spontaneous and urethan-induced tumor incidence in F1 versus Fl mice. Tiinrori 70: 485-490. 9. Farber E (1980). The sequential analysis of liver cancer induction. Biochim. Bioplrys. Acta 605: 149-166. 10. Hicks RM, Wakefield JSt J, and Chowaniec J (1973). Cocarcinogenic action of saccharin in the chemical induction of bladder cancer. Nature 243: 347-349. 11. Nadina D (1976). Preneoplastic lesions in munne mammary cancer. Catrcer Res. 36: 2589-2595. 12. Pitot HC and Sirica AE (1980). The stages of initia-

Vol. 15, No. 2, 1987 LIVER CARCINOGENESIS 233 tion and promotion in hepatocarcinogenesis. Biocliim. Biophys. Acta 605: 19 1-2 15. 13. Poland A, Mak I, and Glover E (1981). Species differences in responsiveness to 1,4-bis[2-(3,5-dichloropyridyloxy)]benzene, a potent phenobarbital-like inducer of microsomal monooxygenase activity. Mol. Pharntacol. 20: 442-450. 14. Poland A, Mak I, Glover E, Boatman RJ, Ebetino F, and Kende AS (1980). 1,4-Bis[2-(3,5-dichloropyridyloxy)]benzene, a potent phenobarbital-like inducer of microsomal monooxygenase activity. Mol. Pharinacol. 18: 57 1-580. 15. Pugh TD, King JH, Koen H, Nychka D, Chover J, Wahba G, He Y, and Goldfarb S (1983). Reliable stereological method for estimating the number of microscopic hepatocellular foci from their transections. Cawer Res. 43: 1261-1268. 16. Uchida E and Hirono I(1979). Effect ofphenobarbital on induction of liver and lung tumors by dimethylnitrosamine in newborn mice. Ganit 70: 639-644. 17. Vesselinovitch SD, Hacker HJ, and Bannasch P (1985). Histochemical characterization of focal hepatic lesions induced by single diethylnitrosamine treatment in infant mice. Cancer Res. 45: 2774-2780. 18. Vesselinovitch SD and Mihailovich N (1983). Kinetics of diethylnitrosamine hepatocarcinogenesis in the infant mouse. Cancer Res. 43: 4253-4259. 19. Wang CY, Hayashida S, Pamucku AM, and Bryan GT (1976). Enhancing effect of allopurinol on the induction of bladder cancer in rats by N-[4-(5-nitro- 2-furyl)-2-thiazolyl]formamide. Cancer Res. 36: 1551-1 555. 20. Ward JM, Goodman DG, Squire RA, Chu KC, and Linhart MS (1979). Neoplastic and non neoplastic lesions in aging (C57BL/6N x C3H/HeN)Fl (F1) mice. J. Natl. Cancer Itzsf. 63: 849-854. 21. Ward JM, Rice JM, Creasia D, Lynch P, and Rigs C (1983). Dissimilar patterns of promotion by di(2- ethylhexy1)phthalate and phenobarbital of hepatocellular neoplasia initiated by diethylnitrosamine in F1 mice. Carcirtogettesis 4: 1021-1029. 22. Witschi H, Williamson D, and Lock S (1977). Enhancement of urethan tumorigenesis in mouse lung by butylated hydroxytoluene. J. Natl. Cancer Inst. 58: 301-305.