- Email: [email protected]
Long term phenobarbital administration does not promote the multiplication of hepatocytes replicating after single cyproterone acetate administration
Life Sciences 76 (2005) 3057 – 3068 www.elsevier.com/locate/lifescie
Long term phenobarbital administration does not promote the multiplication of hepatocytes replicating after single cyproterone acetate administration Virginie Pichard, Nicolas FerryT Biotherapies He´patiques CIC-INSERM 04, CHU Hotel-Dieu 44035, Nantes, France Received 3 September 2004; accepted 2 December 2004
Abstract Cyproterone acetate (CPA) is a synthetic antiandrogenic compound which is widely used in clinic but suspected to be hepatocarcinogenic. CPA is also mitogenic in rat liver. Using genetic labeling of dividing cells, we examined whether hepatocytes dividing in response to acute CPA administration could give rise to preneoplastic foci after administration of a tumor promoter: phenobarbital. CPA was administered orally to rats and dividing hepatocytes were genetically labeled using retroviral vectors carrying the h-galactosidase gene. After labeling rats were given phenobarbital for 10 months and sacrificed. The presence of h-galactosidase labeled hepatocytes as well as preneoplastic hepatocytes was assessed by immunohistochemistry. Genetic labeling of hepatocytes was obtained in all animals. At the end of phenobarbital administration, no hepatic tumors were observed. Preneoplastic foci were not increased in treated animals as compared to control rats. Moreover h-galactosidase positive hepatocytes were never detected in preneoplastic foci. Finally, the size of the h-galactosidase postive clusters was smaller in treated animals as compared to control rats. We conclude that acute CPA administration is not carcinogenic in rat liver and does not initiate preneoplastic hepatocytes capable to give rise to foci after phenobarbital promotion. Therefore the mitogenic property of CPA is distinct from its putative carcinogenic activity. Finally, analysis of the size of h-galactosidase positive cells clusters demonstrate that phenobarbital does not induce hepatocyte proliferation in rats. D 2005 Elsevier Inc. All rights reserved. Keywords: Cell lineage; Genetic labeling; Retrovirus; Carcinogenesis; Phenobarbital
T Corresponding author. Tel.: +33 2 40 08 74 88; fax: +33 2 40 08 75 06. E-mail address: [email protected] (N. Ferry). 0024-3205/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.lfs.2004.12.026
3058
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
Introduction Cyproterone acetate (CPA) is a synthetic progestin with antiandrogenic activity which is widely used in clinic as an oral contraceptive as well as for the treatment of prostate cancer and hirsutism. However, CPA is mitogenic and carcinogenic in the adult rat liver and these effects deserve further attention. When administered to rat, CPA induces rapidly DNA synthesis in hepatocytes resulting in liver hyperplasia (Schulte-Hermann et al., 1980). After CPA administration has stopped, the liver returns to a normal weight through an increase in hepatocyte apoptosis (Bursch et al., 1984; Roberts et al., 1995). Long term studies to evaluate the toxic potential of CPA revealed that this compound was a liver carcinogen when administered to rats at supra therapeutic doses for prolonged periods of time (Schuppler and Gunzel, 1979). It was then proposed that CPA behaved as a tumor promoter, capable of promoting the growth of initiated cells via its mitogenic activity, but devoid of tumor-initiating activity (Schulte-Hermann et al., 1980; Schulte-Hermann et al., 1983a; Schuppler et al., 1983). However, it was subsequently demonstrated that CPA was genotoxic in vivo and in cultured hepatocytes and could initiate preneoplastic lesions in the liver (Deml et al., 1993; Martelli et al., 1995; Neumann et al., 1992). Therefore CPA is now considered as a complete carcinogen. Recently, we used genetic labeling of hepatocytes to decipher the relationship between the immediate mitogenic and tumor initiating activities of CPA administered acutely. We demonstrated that the initiating and mitogenic activities of CPA were directed towards different hepatocyte populations (Auvigne et al., 2002). In that study, we used onco-retroviral vectors carrying the h-galactosidase marker gene to selectively label the hepatocytes dividing upon the effect of acute CPA administration. After labeling, rats were treated with 2-acetyl aminofluorene (2-AAF) to promote the growth of putative CPA-initiated cells. 2-AAF is a carcinogen promoter that is widely used in the initiation/promotion model of hepatocarcinogenesis (Solt et al., 1977). In that previous study, 2-AAF treatment resulted in the appearance of preneoplastic lesions expressing the placental form of glutathion-S-transferase (GST-p). This enzyme is a classical marker of preneoplastic, enzyme-altered hepatocytes (Satoh et al., 1985). We did not observe a preferential appearance of preneoplastic foci in the population of h-galactosidase labeled hepatocytes. However, 2-AAF is a complete carcinogen which can induce liver preneoplastic and neoplastic lesions in rats when given in the diet for long periods of time (Farber, 1956; Reuber, 1964). Therefore, in our previous study a number of preneoplastic foci could have resulted from direct effect of 2-AAF treatment and not from administration of CPA. To resolve this issue and to gain further insights into the carcinogenic potency of CPA in rats, we devised another protocol using phenobarbital as a tumor promoter. Phenobarbital is a tumor promoter compound devoid of complete carcinogenic activity (Hagiwara et al., 1999). Therefore, phenobarbital is able to amplify cells that have been initiated in response to a carcinogen leading to tumors (Peraino et al., 1971; Peraino et al., 1973). In contrast, phenobarbital alone does not cause liver tumors in rats (Anon, 2000; Gould et al., 2001). In the present study, rats were fed CPA for two consecutive days and injected with retroviral vectors containing a h-galactosidase gene to specifically label the population of hepatocytes stimulated by the mitogenic activity of CPA. Moreover, to avoid disappearance of labeled cells due to immune response against the h-galactosidase marker protein, we used the same transgenic h-galactosidase tolerant rats that were used in our previous study.
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
3059
Materials and methods Animals and surgical procedures Sprague Dawley LacZ transgenic rats were obtained as previously described (Charreau et al., 1999). Adult rats from both genders, weighing 160 to 180 g, were used in this study and maintained under a 12hour light/dark illumination cycle. All animals received food and drinking water ad libitum throughout the experimental period. All surgical procedures were conducted on deeply anesthetized animals according to the guidelines of the French Ministry of Agriculture. Cyproterone acetate (CPA) tablets obtained from Schering (Berlin, Germany) were resuspended in distilled water at a concentration of 33 mg/ml and administered for two consecutive days (day 0 and day 1) by force feeding at a daily dose of 200 mg/kg at 10 a.m. each day. Recombinant retroviral vectors (2 ml) were injected into the tail vein twice at 24 and 30 hours after the second CPA administration. Phenobarbital was incorporated at 0.05% (500 ppm) in the diet pellets that were provided by Altromin (Lage, Germany) and continuously administered to rats for 10 months, starting 3 months after retrovirus administration. Five control animals received a single dose of retroviral vector 24 hours after a partial hepatectomy according to the procedure described previously (Aubert et al., 2003). These animals were left untreated for 13 months (rat control 1 and 2) to 15 months (rat control 3 to 5). Liver biopsies (approximately 10 mg) were obtained by section of the tip of the right lateral lobe after median laparotomy and immediately processed for routine histology and immunohistochemistry. Liver biopsies were harvested 7 days after retrovirus administration to check for the presence of transduced hepatocytes. At the time of sacrifice, the whole liver was harvested and processed for immunohistology. Viral vectors We used amphotropic recombinant retroviral vectors containing the Escherichia coli h-galactosidase gene coupled to the nuclear localizing signal (nls-lacZ gene) and produced by the TELCeB6 AF7 producing cell line (Cosset et al., 1995). Cells were cultured at 378C in Dulbecco’s modified Eagle’s medium containing 10% fetal calf serum, 100 U/ml penicillin, and 100 mg/ml streptomycin. Once the cells reached confluency, the medium was harvested and filtered through a 0.45-Am filter. The medium was used immediately after concentration via tangential ultra filtration and titered as previously described (Kitten et al., 1997). The titer of the viral stocks routinely reached 108 transducing units/ml. Histological analysis Morphological analysis was performed on paraffin sections (5Am) after hematoxylin and eosin staining. The presence of h-galactosidase positive cells was assessed by immunohistochemistry on formalin-fixed/paraffin-embedded sections (5 Am). Rabbit anti-h-galactosidase antibodies (Chemicon, Temecula, USA) diluted 1:1000 in PBS containing bovine serum albumin (2% w/v) and Tween 20 (0.1%v/v) were applied overnight at 48C. Positive cells were visualized with biotinylated anti-rabbit immunoglobulin and streptavidin-peroxydase complex using diaminobenzidine or AEC as a chromogenic substrate. Slides were counterstained with hematoxylin. Detection of the placental form
3060
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
of glutathione-S-transferase (GSTp) was carried out in the same way except that the primary antibody (Dako, Trappes, France) was diluted 1:20 and incubated for 2 hours at room temperature. Beta-galactosidase positive hepatocytes were counted at 400 magnification in at least 10 random and non-overlapping fields representing a total of 3000 cells. For each sample, the percentage of positive cells was determined as the ratio between the h-galactosidase positive hepatocytes to the total number of hepatocytes. Foci of GSTp positive cells were counted on 5 different sections of 2 cm2 from different liver lobes at 100 magnification.
Results Our transgenic rats harbored a LacZ gene encoding the cytoplasmic form of h-galactosidase under the transcriptional control of the long terminal repeat of the human immunodeficiency virus. Although there is no detectable expression of h-galactosidase in most tissues these animals are tolerant for hgalactosidase making them particularly suitable for long term studies of h-galactosidase expressing cells without immune elimination of transduced cells (Aubert et al., 2002; Menoret et al., 2002). Moreover, in case the endogenous cytoplasmic h-galactosidase would be reexpressed, it could easily been distinguished by immunohistochemistry from the nuclear h-galactosidase expressed from the retroviral vector. Hepatocyte labeling after CPA administration Animals were given CPA and received recombinant retroviral vectors according to the schedule described in Material and Methods. The h-galactosidase retroviral vector (108 infectious viral particles/ ml) was injected twice at 24 and 30 hours after the second administration of CPA since we previously documented that this delay corresponded to the peak of mitoses in the liver after acute CPA administration (Auvigne et al., 2002). Because retroviral vectors are only able to infect, and hence to Table 1 Immunohistochemical analysis of rat liver Rat number
% h-gal at day 7
% h-gal at sacrifice
GSTp foci/cm2
1 2 3 4 5 mean Control Control Control Control Control mean
1.2% 3.3% 1.6% 3.1% 0.6% 1.96 F 1.18%
0.3% 1.1 1% 1.4% 0.5% 0.86 F 0.45%
1.01 1.77 1.28 1.7 1.5 1.45 F 03 1.77 2.16 3.6 3 3.2 2.75 F 0.76
1 2 3 4 5
Control animals were injected with retroviral vectors 24 hours after a partial hepatectomy and did not receive CPA or phenobarbital thereafter.
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
3061
Fig. 1. Immunohistochemical detection of h-galactosidase in liver sections. Liver fragments were embedded in paraffin and hgalactosidase positive cells were revealed by immunohistochemistry using anti-h-galactosidase antibodies. Positive cells in liver biopsies were mainly present around the portal space (PS) and appeared with a brown nucleus. Hematoxylin counterstained (original magnification 400).
transduce, actively dividing cells, it was of importance to deliver the virus at the time of maximal proliferation to optimize gene transfer to hepatocytes. The efficacy of hepatocyte labeling was evaluated in a liver biopsy harvested 7 days after virus injection. Nuclear expression of h-galactosidase was present in 1.96 F 1.18% of hepatocytes (Table 1, Fig. 1). This figure is in keeping with our previous study of gene marking after CPA administration (Auvigne et al., 2002). Also, as previously observed, the positive hepatocytes in the day 7 biopsy were either isolated or gathered in small clusters of 2 to 3 cells, preferentially located in periportal and mediolobular areas. After the genetic labeling, animals were allowed to recover for 3 months and thereafter were given phenobarbital in the diet continuously for 10 months. Three animals were sacrificed at the end of the recovery period. In these animals, the proportion
Fig. 2. Hypertrophy of hepatocytes after phenobarbital administration. In this animal (rat #2) a moderate (grade 3) hepatocyte hepertrophy is present around the central vein (CV) at the time of sacrifice. Hepatocyte surrounding the portal space have a normal size. Hematoxylin and eosin staining (original magnification 400).
3062
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
Fig. 3. Immunohistochemical detection of preneoplastic foci. As a marker of preneoplastic hepatocytes, GSTp expression was revealed on paraffin liver sections by immunohistochemistry. Preneoplastic cells were detected in small clusters in the liver parenchyma in both control (panel A) and experimental animals (panel B). Hematoxylin counterstained (original magnification 200).
of h-galactosidase expressing hepatocytes (1.78 F 0.64%) was not significantly different from the one observed at day 7 after virus delivery (1.96 F 1.18%; p=0.082 using the two sided Student’s t test). Immunohistochemical analysis of treated rat liver After completion of phenobarbital administration, animals were sacrificed and their livers were analyzed by immunohistochemistry. We observed hypertrophy of hepatocytes around the central vein (Fig. 2) as previously noticed after long term administration of phenobarbital (Anon, 2000; Berman et al., 1983; Gould et al., 2001). Mild (grade 2) to moderate (grade 3) centrilobular hypertrophy was present in all animals. This indicated that animals correctly received phenobarbital. In all animals, foci of GSTp positive cells were observed (Table 1, Fig. 3) and we never detected liver tumors. The mean Table 2 Quantification by size of GSTp foci Rat number
b20 cells
20 to 100 cells
N100 cells
1 2 3 4 5 mean Control Control Control Control Control mean
37.5% 20% 33.3% 16.6% 23.1% 26.1% 33.3% 31.6% 28.6% 50.1% 33.3% 35.4%
25% 40% 36.4% 66.7% 53.9% 44.4% 33.3% 44.4% 42.8% 33.3% 44.4% 39.6%
37.5% 40% 27.3% 16.6% 23.1% 28.9% 33.3% 22.2% 28.6% 16.6% 22.3% 24.6%
1 2 3 4 5
Control animals of the same age did not receive CPA or phenobarbital.
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
3063
number of GSTp foci was 1.45 F 0.3 foci/cm2. This value was in the same range than the one found in control animals that did not receive CPA or phenobarbital (2.75 foci/cm2). As shown in Table 2 the size distribution of GSTp foci was similar in treated rats and in control untreated animals (p N 0.1 for all groups using the Student’s t test). Interestingly, we did not detect GSTp foci in animals sacrificed at the end of the recovery period and that did not receive phenobarbital (data not shown) indicating that CPA alone did not trigger the development of GSTp foci. All these values were in sharp contrast with the previous values that we obtained using promotion with 2-AAF after CPA initiation (i.e. 38.9 GSTpfoci/ cm2). We next assayed the presence of h-galactosidase positive cells on serial sections. We observed the presence of 0.86 F 0.45% h-galactosidase positive hepatocytes outside the GSTp foci. This value was not significantly different from the proportion of positive cells present in the day 7 biopsy (p = 0.09 using the Student’s t test). Moreover, the h-galactosidase positive cells gathered in small clusters and were mainly located in the periportal and mediolobular areas. We never observed foci of GSTp positive cells that also expressed h-galactosidase indicating that preneoplastic foci never arose from hgalactosidase positive cells. Mitogenic effect of phenobarbital Finally, we analyzed the size of h-galactosidase positive clusters to evaluate the mitogenic effect of phenobarbital. We found clusters of up to 46 h-galactosidase positive hepatocytes in the liver sections (Fig. 4). The clusters were divided in three groups according to their size (i.e. b3 cells, from 4 to 6 cells and N6 cells). The vast majority of the positive cells were present in the small clusters and very large clusters were rarely seen in the liver sections (Table 2). The values were compared to those obtained in five control animals. In these animals, hepatocytes were labeled 24 hours after partial hepatectomy and they were sacrificed 12 to 15 months after genetic labeling without treatment with CPA or phenobarbital. Therefore, the cluster size in these animals reflects the normal renewal of hepatocytes in the ageing liver without pharmacological treatment. As shown in Table 2, the distribution of the clusters in treated
Fig. 4. Imunohistochemical detection of a large h-galactosidase cluster. h-galactosidase positive cells were revealed by immunohistochemistry using anti-h-galactosidase antibodies. Hematoxylin counterstained (original magnification 200).
3064
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
animals was shifted towards smaller sizes as compared to control rats. The proportion of large clusters was significatively higher in control animals (Student’s t test, p b 0.005 for all groups) indicating a lower number of cell division in phenobarbital treated animals.
Discussion More than 20 years ago, it was recognized that CPA produced liver tumors in rats when given at pharmacological doses (Schuppler and Gunzel, 1979). Because this drug is widely used in clinic, many studies have adressed its carcinogenic potential in human and most of the studies have come to the agreement that the liver cancer risk associated with the use of CPA is low, if it exists (Hinkel et al., 1996; Kasper, 2001; Kattan et al., 1994; Laron and Kauli, 2000; Rabe et al., 1996). Nevertheless, progestogens are still classified as potentially carcinogenic in human by the International Agency for Research on Cancer (IARC) and it was reported that in human liver slices, CPA induced DNA adducts (Baumann et al., 1996). In the rat, CPA was long considered as a tumor promoter capable of inducing the growth of initiated cells through its mitogenic activity (Schulte-Hermann et al., 1983b; Schuppler et al., 1983) but devoid of tumor-initiating potency (Schuppler et al., 1983). It was subsequently demonstrated that CPA was also genotoxic (Deml et al., 1993; Krebs et al., 1998; Neumann et al., 1992) and it was suggested that CPA was a complete carcinogen, and that its promoter activity was linked to its mitogenic activity when administered for long period of time (Krebs et al., 1998). However, the relationship between genotoxic and mitogenic activity of CPA after acute administration were less clear. We previously reported that the mitogenic and tumor initating activities of CPA administered during a short period of time may target different cell populations (Auvigne et al., 2002). However, in all previous experimental studies dealing with CPA initiating activity, CPA administration was systematically followed by administration of strong tumor promoters such as Clophen A50 (Deml et al., 1993) or 2-AAF (Auvigne et al., 2002; Martelli et al., 1995) which are complete carcinogenic compounds. Therefore, the appearance of enzyme-altered foci in this setting could result from the concomittant administration of these carcinogens and when CPA was administered alone, no preneoplastic foci were observed in the liver (Krebs et al., 1998). To circumvent this drawback, we performed here a lineage analysis using phenobarbital, a widely used liver tumor promoter in rats (Peraino et al., 1971; Peraino et al., 1973) devoid of any genotoxic activity. We used genetic labeling to study the fate of hepatocytes dividing in response to acute CPA administration. Using h-galactosidase retroviral vectors that are only able to transduce dividing cells, labeling was specifically restricted to hepatocytes replicating in response to the acute administration of CPA. Using this approach, most of dividing hepatocytes were labeled. Indeed, we previously documented that after CPA administration the mitotic index found in hepatocytes was 1.03 F 0.08% (Pichard et al., 2001). The labeling index that we obtained in the present study (1.96 F 1.18%) is in the same range demonstrating the few, if any, replicating hepatocytes escaped retroviral infection. This indicated that we were able to follow the fate of the majority of the population of hepatocytes that are sensitive to the mitogenic effect of CPA. We did not observe the appearance of preneoplastic foci up to three months after CPA administration indicating that acute administration of CPA did not promote the growth of putatively initiated hepatocytes. After 10 months of phenobarbital promotion, the mean number of preneoplastic foci (1.45 F 0.3 foci/cm2) that we observed in treated rats was similar to the one observed in control non treated animals but clearly lower that the number of preneoplastic foci that we
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
3065
previously detected after CPA administration followed by 2-AAF treatment (38.9 F 5.45 foci/cm2) (Auvigne et al., 2002). This indicates that the results of previous studies using strong promoters were flawed because GSTp foci might have been induced by the complete carcinogenic potency of these promoters. It is therefore critical to compare various promoters in carcinogenic studies and our present findings support the previous demonstration that developemnt of preneoplastic foci is affected differently by various tumor promoting agents (Dragan and Pitot, 1992; Ledda-Columbano et al., 1989). The preneoplastic foci that we observed here did not preferentially expressed h-galactosidase. This confirmed that the hepatocytes dividing in response to acute CPA administration and that were labeled after retrovirus administration were not initiated cells committed to preneoplastic transformation. Indeed, if the mitogenic effect of CPA was linked to initiating activity we should have observed the generation of preneoplastic foci from labeled hepatocytes and hence most of the GSTp positive foci should also express h-galactosidase. We cannot exclude that the initiated hepatocytes were eliminated by the wave of apoptosis that follows withdrawal of CPA (Bursch et al., 1990). However this is unlikely because we and others previously demonstrated that the wave of apoptosis did not target specifically those hepatocytes that divided in response to acute CPA administration (Auvigne et al., 2002; Roberts et al., 1995). Finally, we also observed that phenobarbital was not mitogenic in our experimental animals. We previously documented that retroviral genetic labeling is a powerful tool to study the fate as well as the clonal expansion of normal hepatocytes in the liver. When clusters of h-galactosidase positive hepatocytes were divided in three groups (i.e. b3 cells, from 4 to 6 cells and N6 cells) we observed that the proportion of small clusters was significantly higher in the treated group than in the control rats (Table 3). If phenobarbital had triggered cell proliferation the number of small clusters should be largely lower than in control animals after 10 months of phenobarbital administration. Conversely, the number of large clusters should be higher in phenobarbital treated rats than in control animals. This is clearly not the case and therefore we believe that phenobarbital was actually unable to increase hepatocyte division in our conditions. The observation that phenobarbital is not mitogenic in rat liver may seem highly surprising since this compound has widely been described as mitogenic. However, it is noteworthy that most of the studies have used phenobarbital in conjunction with other carcinogenic compounds and at Table 3 Quantification by size of clusters of h-galactosidase positive hepatocytes Rat number
b3 cells
4 to 6 cells
N6 cells
1 2 3 4 5 mean Control Control Control Control Control mean
96.2% 95.5% 97.7% 95.3% 96.7% 96.3 F 0.93% 85.4% 73.2% 76.3% 68.6% 86.2% 77.94 F 7.7%
1.9% 3% 0% 3.1% 0% 1.6 F 1.53% 6.1% 8.2% 9.3% 17.1% 10% 10.14 F 4.2%
1.9% 1.5% 2.3% 1.6% 3.3% 2.12 F 0.73% 8.5% 18.6% 14.4% 14.3% 3.8% 11.92 F 5.8%
1 2 3 4 2
Control animals were injected with retroviral vectors 24 hours after a partial hepatectomy and were left without further treatment for 12–15 months.
3066
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
various dose levels. In contrast it has also been reported that phenobarbital, when given alone, did not increase significantly mitotic index in partially hepatectomized animals (Hashimoto et al., 1999) and administration of phenobarbital alone was able to induce liver atrophy and reduced hepatic regenerative response in a previous study (Rozga et al., 1991). Furthermore, when measuring hyperplasia through DNA synthesis, polyploidization may be confounded with true hyperplasia since phenobarbital is a potent inducer of polyploidisation (Martin et al., 2001) and accurate methods to assess the actual increase in hepatocyte number are required (Carthew et al., 1998). Our genetic labeling technics does not allow to detect such polyploidisation. Therefore although we cannot exclude that it actually increases the ploidy status of hepatocytes we believe that in our conditions (i.e. long term administration at 500 ppm in drinking water) phenobarbital does not increase the number of hepatocytes. Although the precise mechanism underlying the effect of tumor promoters is not fully understood, the importance of proteins involved in growth regulation has been reported. Indeed, phenobarbital was shown to induce expression of EGF as well as HGF receptors in centrilobuler hepatocytes (Orton et al., 1996). Beta catenin mutations were also implicated together with EGF receptor and in c-myc/TGF-alpha transgenic mice phenobarbital treatment selected for a cell population displaying activation of betacatenin and resulted in liver tumors (Calvisi et al., 2004). More recently, the nuclear receptor constitutive/active androstane receptor (CAR), also has been described as a target of phenobarbital. Indeed, mice lacking CAR are much less susceptible to liver tumor promotion by phenobarbital (Yamamoto et al., 2004). Although these previous studies were carried out in mice, the absence of promotion that we observed in our present study could be attributable to lack of, or decreased, expression of any of these potential targets. However, we believe that it is not the case since we observed obvious centrilobular hepatocyte hypertrophy indicating that our rat strain was sensitive to phenobarbital. In conclusion, our present data indicate that the mitogenic activity of CPA is not a tumor initiating activity and suggest that CPA is not carcinogenic after acute administration. Acknowledgements We wish to thank Be´atrice Sacre´-Salem for help with histological analysis and Je´roˆme Gournay for help with statistical analysis. This work was supported by grants from the Association Franc¸aise contre les Myopathies and the Ligue De´partementale contre le Cancer de Loire Atlantique. References Anon, 2000. NTP Toxicology and Carcinogenesis Studies of Primidone (CAS No. 125-33-7) in F344/N Rats and B6C3F1 Mice (Feed Studies). Natl. Toxicol. Program. Tech. Rep. Ser. 476, 1 – 290. Aubert, D., Menoret, S., Chiari, E., Pichard, V., Durand, S., Tesson, L., Moullier, P., Anegon, I., Ferry, N., 2002. Cytotoxic immune response blunts long-term transgene expression after efficient retroviral-mediated hepatic gene transfer in rat. Mol. Ther. 5, 388 – 396. Aubert, D., Pichard, V., Durand, S., Moullier, P., Ferry, N., 2003. Cytotoxic immune response after retroviral-mediated hepatic gene transfer in rat does not preclude expression from adeno-associated virus 1 transduced muscles. Hum. Gene Ther. 14, 473 – 481. Auvigne, I., Pichard, V., Aubert, D., Robillard, N., Ferry, N., 2002. In vivo cell lineage analysis in cyproterone acetate-treated rat liver using genetic labeling of hepatocytes. Hepatology 35, 281 – 288.
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
3067
Baumann, A., Kerdar, R.S., Cramer, P., Feser, W., Blode, H., Salomon, A., Kuhnz, W., 1996. Use of rat and human liver slices for the detection of steroid hormone-induced DNA-adducts in vitro by means of the (32)P-postlabeling technique. Pharmacol. Toxicol. 78, 214 – 223. Berman, J.S., Hiley, C.R., Wilson, A.C., 1983. Comparison of the effects of the hypolipidaemic agents ICI53072 and clofibrate with those of phenobarbitone on liver size, blood flow and DNA content in the rat. Br. J. Pharmacol. 78, 533 – 541. Bursch, W., Lauer, B., Timmermann-Trosiener, I., Barthel, G., Schuppler, J., Schulte-Hermann, R., 1984. Controlled death (apoptosis) of normal and putative preneoplastic cells in rat liver following withdrawal of tumor promoters. Carcinogenesis 5, 453 – 458. Bursch, W., Paffe, S., Putz, B., Barthel, G., Schulte-Hermann, R., 1990. Determination of the length of the histological stages of apoptosis in normal liver and in altered hepatic foci of rats. Carcinogenesis 11, 847 – 853. Calvisi, D.F., Ladu, S., Factor, V.M., Thorgeirsson, S.S., 2004. Activation of beta-catenin provides proliferative and invasive advantages in c-myc/TGF-alpha hepatocarcinogenesis promoted by phenobarbital. Carcinogenesis 25, 901 – 908 (Epub 2004 Jan 2023). Carthew, P., Edwards, R.E., Nolan, B.M., 1998. The quantitative distinction of hyperplasia from hypertrophy in hepatomegaly induced in the rat liver by phenobarbital. Toxicol. Sci. 44, 46 – 51. Charreau, B., Menoret, S., Tesson, L., Azimzadeh, A., Audet, M., Wolf, P., Marquet, R., Verbakel, C., Ijzermans, J., Cowan, P., Pearse, M., d’Apice, A., Soulillou, J.P., Anegon, I., 1999. Protection against hyperacute xenograft rejection of transgenic rat hearts expressing human decay accelerating factor (DAF) transplanted into primates. Mol. Med. 5, 617 – 630. Cosset, F.L., Takeuchi, Y., Battini, J.L., Weiss, R.A., Collins, M.K.L., 1995. High-titer packaging cells producing recombinant retroviruses resistant to human serum. J. Virol. 69, 7430 – 7436. Deml, E., Schwarz, L.R., Oesterle, D., 1993. Initiation of enzyme-altered foci by the synthetic steroid cyproterone acetate in rat liver foci bioassay. Carcinogenesis 14, 1229 – 1231. Dragan, Y.P., Pitot, H.C., 1992. The role of the stages of initiation and promotion in phenotypic diversity during hepatocarcinogenesis in the rat. Carcinogenesis 13, 739 – 750. Farber, E., 1956. Similarities in the sequence of early histological changes induced in the liver of the rat by ethionine, 2acetylaminofluorene and 3V-methyl-4-dimethylaminoazobenzene. Cancer Res. 16, 142 – 148. Gould, S., Sidaway, J., Sansom, N., Betton, G., Orton, T., 2001. Phenobarbitone-induced liver response in wild type and in p53 deficient mice. Toxicol. Lett. 122, 131 – 140. Hagiwara, A., Miyata, E., Tamano, S., Sano, M., Masuda, C., Funae, Y., Ito, N., Fukushima, S., Shirai, T., 1999. Noncarcinogenicity, but dose-related increase in preneoplastic hepatocellular lesions, in a two-year feeding study of phenobarbital sodium in male F344 rats. Food Chem. Toxicol. 37, 869 – 879. Hashimoto, M., Kothary, P.C., Raper, S.E., 1999. Phenobarbital in comparison with carbon tetrachloride and phenobarbitalinduced cirrhosis in rat liver regeneration. J. Surg. Res. 81, 164 – 169. Hinkel, A., Berges, R.R., Pannek, J., Schulze, H., Senge, T., 1996. Cyproterone acetate in the treatment of advanced prostatic cancer: retrospective analysis of liver toxicity in the long-term follow-up of 89 patients. Eur. Urol. 30, 464 – 470. Kasper, P., 2001. Cyproterone acetate: a genotoxic carcinogen? Pharmacol. Toxicol. 88, 223 – 231. Kattan, J., Spatz, A., Culine, S., Terrier-Lacombe, M.J., Elias, D., Droz, J.P., 1994. Hepatocellular carcinoma during hormonotherapy for prostatic cancer. Am. J. Clin. Oncol. 17, 390 – 392. Kitten, O., Cosset, F.L., Ferry, N., 1997. Highly efficient retrovirus-mediated gene transfer into rat hepatocytes in vivo. Hum. Gene Ther. 8, 1491 – 1494. Krebs, O., Schafer, B., Wolff, T., Oesterle, D., Deml, E., Sund, M., Favor, J., 1998. The DNA damaging drug cyproterone acetate causes gene mutations and induces glutathione-S-transferase P in the liver of female Big Blue transgenic F344 rats. Carcinogenesis 19, 241 – 245. Laron, Z., Kauli, R., 2000. Experience with cyproterone acetate in the treatment of precocious puberty. J. Pediatr. Endocrinol. Metab. 13, 805 – 810. Ledda-Columbano, G.M., Columbano, A., Curto, M., Ennas, M.G., Coni, P., Sarma, D.S., Pani, P., 1989. Further evidence that mitogen-induced cell proliferation does not support the formation of enzyme-altered islands in rat liver by carcinogens. Carcinogenesis 10, 847 – 850. Martelli, A., Mattioli, F., Fazio, S., Andrae, U., Brambilla, G., 1995. DNA repair synthesis and DNA fragmentation in primary cultures of human and rat hepatocytes exposed to cyproterone acetate. Carcinogenesis 16, 1265 – 1269. Martin, N.C., McGregor, A.H., Sansom, N., Gould, S., Harrison, D.J., 2001. Phenobarbitone-induced ploidy changes in liver occur independently of p53. Toxicol. Lett. 119, 109 – 115.
3068
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
Menoret, S., Aubert, D., Tesson, L., Braudeau, C., Pichard, V., Ferry, N., Anegon, I., 2002. lacZ Transgenic Rats Tolerant for beta-Galactosidase: Recipients for Gene Transfer Studies Using lacZ as a Reporter Gene. Hum. Gene Ther. 13, 1383 – 1390. Neumann, I., Thierau, D., Andrae, U., Greim, H., Schwarz, L.R., 1992. Cyproterone acetate induces DNA damage in cultured rat hepatocytes and preferentially stimulates DNA synthesis in gamma-glutamyltranspeptidase-positive cells. Carcinogenesis 13, 373 – 378. Orton, T.C., Doughty, S.E., Kalinowski, A.E., Lord, P.G., Wadsworth, P.F., 1996. Expression of growth factors and growth factor receptors in the liver of C57BL/10J mice following administration of phenobarbitone. Carcinogenesis 17, 973 – 981. Peraino, C., Fry, R.J., Staffeldt, E., 1971. Reduction and enhancement by phenobarbital of hepatocarcinogenesis induced in the rat by 2-acetylaminofluorene. Cancer Res. 31, 1506 – 1512. Peraino, C., Fry, R.J., Staffeldt, E., Kisieleski, W.E., 1973. Effects of varying the exposure to phenobarbital on its enhancement of 2-acetylaminofluorene-induced hepatic tumorigenesis in the rat. Cancer Res. 33, 2701 – 2705. Pichard, V., Aubert, D., Ferry, N., 2001. Efficient retroviral gene transfer to the liver in vivo using nonpolypeptidic mitogens. Biochem. Biophys. Res. Commun. 286, 929 – 935. Rabe, T., Feldmann, K., Heinemann, L., Runnebaum, B., 1996. Cyproterone acetate: is it hepato- or genotoxic? Drug Saf. 14, 25 – 38. Reuber, M., 1964. Development of preneoplastic and neoplastic lesions of the liver in male rats given 0.025 percent N-2fluorenyldiacetamide. J. Intl. Canc. Instit. 34, 697 – 723. Roberts, R.A., Soames, A.R., Gill, J.H., James, N.H., Wheeldon, E.B., 1995. Non-genotoxic hepatocarcinogens stimulate DNA synthesis and their withdrawal induces apoptosis, but in different hepatocyte populations. Carcinogenesis 16, 1693 – 1698. Rozga, J., Foss, A., Alumets, J., Ahren, B., Jeppsson, B., Bengmark, S., 1991. Liver cirrhosis in rats: regeneration and assessment of the role phenobarbital. J. Surg. Res. 51, 329 – 335. Satoh, K., Kitahara, A., Soma, Y., Inaba, Y., Hatayama, I., Sato, K., 1985. Purification, induction, and distribution of placental glutathione transferase: a new marker enzyme for preneoplastic cells in the rat chemical hepatocarcinogenesis. Proc. Natl. Acad. Sci. USA 82, 3964 – 3968. Schulte-Hermann, R., Hoffman, V., Parzefall, W., Kallenbach, M., Gerhardt, A., Schuppler, J., 1980. Adaptive responses of rat liver to the gestagen and anti-androgen cyproterone acetate and other inducers. II. Induction of growth. Chem. Biol. Interact. 31, 287 – 300. Schulte-Hermann, R., Schuppler, J., Timmermann-Trosiener, I., Ohde, G., Bursch, W., Berger, H., 1983a. The role of growth of normal and preneoplastic cell populations for tumor promotion in rat liver. Environ Health Perspect. 50, 185 – 194. Schulte-Hermann, R., Timmermann-Trosiener, I., Schuppler, J., 1983b. Promotion of spontaneous preneoplastic cells in rat liver as a possible explanation of tumor production by nonmutagenic compounds. Cancer Res. 43, 839 – 844. Schuppler, J., Damme, J., Schulte-Hermann, R., 1983. Assay of some endogenous and synthetic sex steroids for tumor-initiating activity in rat liver using the Solt-Farber system. Carcinogenesis 4, 239 – 241. Schuppler, J., Gunzel, P., 1979. Liver tumors and steroid hormones in rats and mice. Arch. Toxicol., 181 – 195. Solt, D., Medline, A., Farber, E., 1977. Rapid ermergence of carcinogen-induced hyperplastic lesions in a new model for the sequential analysis of liver carcinogenesis. Am. J. Pathol. 88, 595 – 610. Yamamoto, Y., Moore, R., Goldsworthy, T.L., Negishi, M., Maronpot, R.R., 2004. The orphan nuclear receptor constitutive active/androstane receptor is essential for liver tumor promotion by phenobarbital in mice. Cancer Res. 64, 7197 – 7200.
Long term phenobarbital administration does not promote the multiplication of hepatocytes replicating after single cyproterone acetate administration Virginie Pichard, Nicolas FerryT Biotherapies He´patiques CIC-INSERM 04, CHU Hotel-Dieu 44035, Nantes, France Received 3 September 2004; accepted 2 December 2004
Abstract Cyproterone acetate (CPA) is a synthetic antiandrogenic compound which is widely used in clinic but suspected to be hepatocarcinogenic. CPA is also mitogenic in rat liver. Using genetic labeling of dividing cells, we examined whether hepatocytes dividing in response to acute CPA administration could give rise to preneoplastic foci after administration of a tumor promoter: phenobarbital. CPA was administered orally to rats and dividing hepatocytes were genetically labeled using retroviral vectors carrying the h-galactosidase gene. After labeling rats were given phenobarbital for 10 months and sacrificed. The presence of h-galactosidase labeled hepatocytes as well as preneoplastic hepatocytes was assessed by immunohistochemistry. Genetic labeling of hepatocytes was obtained in all animals. At the end of phenobarbital administration, no hepatic tumors were observed. Preneoplastic foci were not increased in treated animals as compared to control rats. Moreover h-galactosidase positive hepatocytes were never detected in preneoplastic foci. Finally, the size of the h-galactosidase postive clusters was smaller in treated animals as compared to control rats. We conclude that acute CPA administration is not carcinogenic in rat liver and does not initiate preneoplastic hepatocytes capable to give rise to foci after phenobarbital promotion. Therefore the mitogenic property of CPA is distinct from its putative carcinogenic activity. Finally, analysis of the size of h-galactosidase positive cells clusters demonstrate that phenobarbital does not induce hepatocyte proliferation in rats. D 2005 Elsevier Inc. All rights reserved. Keywords: Cell lineage; Genetic labeling; Retrovirus; Carcinogenesis; Phenobarbital
T Corresponding author. Tel.: +33 2 40 08 74 88; fax: +33 2 40 08 75 06. E-mail address: [email protected] (N. Ferry). 0024-3205/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.lfs.2004.12.026
3058
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
Introduction Cyproterone acetate (CPA) is a synthetic progestin with antiandrogenic activity which is widely used in clinic as an oral contraceptive as well as for the treatment of prostate cancer and hirsutism. However, CPA is mitogenic and carcinogenic in the adult rat liver and these effects deserve further attention. When administered to rat, CPA induces rapidly DNA synthesis in hepatocytes resulting in liver hyperplasia (Schulte-Hermann et al., 1980). After CPA administration has stopped, the liver returns to a normal weight through an increase in hepatocyte apoptosis (Bursch et al., 1984; Roberts et al., 1995). Long term studies to evaluate the toxic potential of CPA revealed that this compound was a liver carcinogen when administered to rats at supra therapeutic doses for prolonged periods of time (Schuppler and Gunzel, 1979). It was then proposed that CPA behaved as a tumor promoter, capable of promoting the growth of initiated cells via its mitogenic activity, but devoid of tumor-initiating activity (Schulte-Hermann et al., 1980; Schulte-Hermann et al., 1983a; Schuppler et al., 1983). However, it was subsequently demonstrated that CPA was genotoxic in vivo and in cultured hepatocytes and could initiate preneoplastic lesions in the liver (Deml et al., 1993; Martelli et al., 1995; Neumann et al., 1992). Therefore CPA is now considered as a complete carcinogen. Recently, we used genetic labeling of hepatocytes to decipher the relationship between the immediate mitogenic and tumor initiating activities of CPA administered acutely. We demonstrated that the initiating and mitogenic activities of CPA were directed towards different hepatocyte populations (Auvigne et al., 2002). In that study, we used onco-retroviral vectors carrying the h-galactosidase marker gene to selectively label the hepatocytes dividing upon the effect of acute CPA administration. After labeling, rats were treated with 2-acetyl aminofluorene (2-AAF) to promote the growth of putative CPA-initiated cells. 2-AAF is a carcinogen promoter that is widely used in the initiation/promotion model of hepatocarcinogenesis (Solt et al., 1977). In that previous study, 2-AAF treatment resulted in the appearance of preneoplastic lesions expressing the placental form of glutathion-S-transferase (GST-p). This enzyme is a classical marker of preneoplastic, enzyme-altered hepatocytes (Satoh et al., 1985). We did not observe a preferential appearance of preneoplastic foci in the population of h-galactosidase labeled hepatocytes. However, 2-AAF is a complete carcinogen which can induce liver preneoplastic and neoplastic lesions in rats when given in the diet for long periods of time (Farber, 1956; Reuber, 1964). Therefore, in our previous study a number of preneoplastic foci could have resulted from direct effect of 2-AAF treatment and not from administration of CPA. To resolve this issue and to gain further insights into the carcinogenic potency of CPA in rats, we devised another protocol using phenobarbital as a tumor promoter. Phenobarbital is a tumor promoter compound devoid of complete carcinogenic activity (Hagiwara et al., 1999). Therefore, phenobarbital is able to amplify cells that have been initiated in response to a carcinogen leading to tumors (Peraino et al., 1971; Peraino et al., 1973). In contrast, phenobarbital alone does not cause liver tumors in rats (Anon, 2000; Gould et al., 2001). In the present study, rats were fed CPA for two consecutive days and injected with retroviral vectors containing a h-galactosidase gene to specifically label the population of hepatocytes stimulated by the mitogenic activity of CPA. Moreover, to avoid disappearance of labeled cells due to immune response against the h-galactosidase marker protein, we used the same transgenic h-galactosidase tolerant rats that were used in our previous study.
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
3059
Materials and methods Animals and surgical procedures Sprague Dawley LacZ transgenic rats were obtained as previously described (Charreau et al., 1999). Adult rats from both genders, weighing 160 to 180 g, were used in this study and maintained under a 12hour light/dark illumination cycle. All animals received food and drinking water ad libitum throughout the experimental period. All surgical procedures were conducted on deeply anesthetized animals according to the guidelines of the French Ministry of Agriculture. Cyproterone acetate (CPA) tablets obtained from Schering (Berlin, Germany) were resuspended in distilled water at a concentration of 33 mg/ml and administered for two consecutive days (day 0 and day 1) by force feeding at a daily dose of 200 mg/kg at 10 a.m. each day. Recombinant retroviral vectors (2 ml) were injected into the tail vein twice at 24 and 30 hours after the second CPA administration. Phenobarbital was incorporated at 0.05% (500 ppm) in the diet pellets that were provided by Altromin (Lage, Germany) and continuously administered to rats for 10 months, starting 3 months after retrovirus administration. Five control animals received a single dose of retroviral vector 24 hours after a partial hepatectomy according to the procedure described previously (Aubert et al., 2003). These animals were left untreated for 13 months (rat control 1 and 2) to 15 months (rat control 3 to 5). Liver biopsies (approximately 10 mg) were obtained by section of the tip of the right lateral lobe after median laparotomy and immediately processed for routine histology and immunohistochemistry. Liver biopsies were harvested 7 days after retrovirus administration to check for the presence of transduced hepatocytes. At the time of sacrifice, the whole liver was harvested and processed for immunohistology. Viral vectors We used amphotropic recombinant retroviral vectors containing the Escherichia coli h-galactosidase gene coupled to the nuclear localizing signal (nls-lacZ gene) and produced by the TELCeB6 AF7 producing cell line (Cosset et al., 1995). Cells were cultured at 378C in Dulbecco’s modified Eagle’s medium containing 10% fetal calf serum, 100 U/ml penicillin, and 100 mg/ml streptomycin. Once the cells reached confluency, the medium was harvested and filtered through a 0.45-Am filter. The medium was used immediately after concentration via tangential ultra filtration and titered as previously described (Kitten et al., 1997). The titer of the viral stocks routinely reached 108 transducing units/ml. Histological analysis Morphological analysis was performed on paraffin sections (5Am) after hematoxylin and eosin staining. The presence of h-galactosidase positive cells was assessed by immunohistochemistry on formalin-fixed/paraffin-embedded sections (5 Am). Rabbit anti-h-galactosidase antibodies (Chemicon, Temecula, USA) diluted 1:1000 in PBS containing bovine serum albumin (2% w/v) and Tween 20 (0.1%v/v) were applied overnight at 48C. Positive cells were visualized with biotinylated anti-rabbit immunoglobulin and streptavidin-peroxydase complex using diaminobenzidine or AEC as a chromogenic substrate. Slides were counterstained with hematoxylin. Detection of the placental form
3060
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
of glutathione-S-transferase (GSTp) was carried out in the same way except that the primary antibody (Dako, Trappes, France) was diluted 1:20 and incubated for 2 hours at room temperature. Beta-galactosidase positive hepatocytes were counted at 400 magnification in at least 10 random and non-overlapping fields representing a total of 3000 cells. For each sample, the percentage of positive cells was determined as the ratio between the h-galactosidase positive hepatocytes to the total number of hepatocytes. Foci of GSTp positive cells were counted on 5 different sections of 2 cm2 from different liver lobes at 100 magnification.
Results Our transgenic rats harbored a LacZ gene encoding the cytoplasmic form of h-galactosidase under the transcriptional control of the long terminal repeat of the human immunodeficiency virus. Although there is no detectable expression of h-galactosidase in most tissues these animals are tolerant for hgalactosidase making them particularly suitable for long term studies of h-galactosidase expressing cells without immune elimination of transduced cells (Aubert et al., 2002; Menoret et al., 2002). Moreover, in case the endogenous cytoplasmic h-galactosidase would be reexpressed, it could easily been distinguished by immunohistochemistry from the nuclear h-galactosidase expressed from the retroviral vector. Hepatocyte labeling after CPA administration Animals were given CPA and received recombinant retroviral vectors according to the schedule described in Material and Methods. The h-galactosidase retroviral vector (108 infectious viral particles/ ml) was injected twice at 24 and 30 hours after the second administration of CPA since we previously documented that this delay corresponded to the peak of mitoses in the liver after acute CPA administration (Auvigne et al., 2002). Because retroviral vectors are only able to infect, and hence to Table 1 Immunohistochemical analysis of rat liver Rat number
% h-gal at day 7
% h-gal at sacrifice
GSTp foci/cm2
1 2 3 4 5 mean Control Control Control Control Control mean
1.2% 3.3% 1.6% 3.1% 0.6% 1.96 F 1.18%
0.3% 1.1 1% 1.4% 0.5% 0.86 F 0.45%
1.01 1.77 1.28 1.7 1.5 1.45 F 03 1.77 2.16 3.6 3 3.2 2.75 F 0.76
1 2 3 4 5
Control animals were injected with retroviral vectors 24 hours after a partial hepatectomy and did not receive CPA or phenobarbital thereafter.
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
3061
Fig. 1. Immunohistochemical detection of h-galactosidase in liver sections. Liver fragments were embedded in paraffin and hgalactosidase positive cells were revealed by immunohistochemistry using anti-h-galactosidase antibodies. Positive cells in liver biopsies were mainly present around the portal space (PS) and appeared with a brown nucleus. Hematoxylin counterstained (original magnification 400).
transduce, actively dividing cells, it was of importance to deliver the virus at the time of maximal proliferation to optimize gene transfer to hepatocytes. The efficacy of hepatocyte labeling was evaluated in a liver biopsy harvested 7 days after virus injection. Nuclear expression of h-galactosidase was present in 1.96 F 1.18% of hepatocytes (Table 1, Fig. 1). This figure is in keeping with our previous study of gene marking after CPA administration (Auvigne et al., 2002). Also, as previously observed, the positive hepatocytes in the day 7 biopsy were either isolated or gathered in small clusters of 2 to 3 cells, preferentially located in periportal and mediolobular areas. After the genetic labeling, animals were allowed to recover for 3 months and thereafter were given phenobarbital in the diet continuously for 10 months. Three animals were sacrificed at the end of the recovery period. In these animals, the proportion
Fig. 2. Hypertrophy of hepatocytes after phenobarbital administration. In this animal (rat #2) a moderate (grade 3) hepatocyte hepertrophy is present around the central vein (CV) at the time of sacrifice. Hepatocyte surrounding the portal space have a normal size. Hematoxylin and eosin staining (original magnification 400).
3062
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
Fig. 3. Immunohistochemical detection of preneoplastic foci. As a marker of preneoplastic hepatocytes, GSTp expression was revealed on paraffin liver sections by immunohistochemistry. Preneoplastic cells were detected in small clusters in the liver parenchyma in both control (panel A) and experimental animals (panel B). Hematoxylin counterstained (original magnification 200).
of h-galactosidase expressing hepatocytes (1.78 F 0.64%) was not significantly different from the one observed at day 7 after virus delivery (1.96 F 1.18%; p=0.082 using the two sided Student’s t test). Immunohistochemical analysis of treated rat liver After completion of phenobarbital administration, animals were sacrificed and their livers were analyzed by immunohistochemistry. We observed hypertrophy of hepatocytes around the central vein (Fig. 2) as previously noticed after long term administration of phenobarbital (Anon, 2000; Berman et al., 1983; Gould et al., 2001). Mild (grade 2) to moderate (grade 3) centrilobular hypertrophy was present in all animals. This indicated that animals correctly received phenobarbital. In all animals, foci of GSTp positive cells were observed (Table 1, Fig. 3) and we never detected liver tumors. The mean Table 2 Quantification by size of GSTp foci Rat number
b20 cells
20 to 100 cells
N100 cells
1 2 3 4 5 mean Control Control Control Control Control mean
37.5% 20% 33.3% 16.6% 23.1% 26.1% 33.3% 31.6% 28.6% 50.1% 33.3% 35.4%
25% 40% 36.4% 66.7% 53.9% 44.4% 33.3% 44.4% 42.8% 33.3% 44.4% 39.6%
37.5% 40% 27.3% 16.6% 23.1% 28.9% 33.3% 22.2% 28.6% 16.6% 22.3% 24.6%
1 2 3 4 5
Control animals of the same age did not receive CPA or phenobarbital.
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
3063
number of GSTp foci was 1.45 F 0.3 foci/cm2. This value was in the same range than the one found in control animals that did not receive CPA or phenobarbital (2.75 foci/cm2). As shown in Table 2 the size distribution of GSTp foci was similar in treated rats and in control untreated animals (p N 0.1 for all groups using the Student’s t test). Interestingly, we did not detect GSTp foci in animals sacrificed at the end of the recovery period and that did not receive phenobarbital (data not shown) indicating that CPA alone did not trigger the development of GSTp foci. All these values were in sharp contrast with the previous values that we obtained using promotion with 2-AAF after CPA initiation (i.e. 38.9 GSTpfoci/ cm2). We next assayed the presence of h-galactosidase positive cells on serial sections. We observed the presence of 0.86 F 0.45% h-galactosidase positive hepatocytes outside the GSTp foci. This value was not significantly different from the proportion of positive cells present in the day 7 biopsy (p = 0.09 using the Student’s t test). Moreover, the h-galactosidase positive cells gathered in small clusters and were mainly located in the periportal and mediolobular areas. We never observed foci of GSTp positive cells that also expressed h-galactosidase indicating that preneoplastic foci never arose from hgalactosidase positive cells. Mitogenic effect of phenobarbital Finally, we analyzed the size of h-galactosidase positive clusters to evaluate the mitogenic effect of phenobarbital. We found clusters of up to 46 h-galactosidase positive hepatocytes in the liver sections (Fig. 4). The clusters were divided in three groups according to their size (i.e. b3 cells, from 4 to 6 cells and N6 cells). The vast majority of the positive cells were present in the small clusters and very large clusters were rarely seen in the liver sections (Table 2). The values were compared to those obtained in five control animals. In these animals, hepatocytes were labeled 24 hours after partial hepatectomy and they were sacrificed 12 to 15 months after genetic labeling without treatment with CPA or phenobarbital. Therefore, the cluster size in these animals reflects the normal renewal of hepatocytes in the ageing liver without pharmacological treatment. As shown in Table 2, the distribution of the clusters in treated
Fig. 4. Imunohistochemical detection of a large h-galactosidase cluster. h-galactosidase positive cells were revealed by immunohistochemistry using anti-h-galactosidase antibodies. Hematoxylin counterstained (original magnification 200).
3064
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
animals was shifted towards smaller sizes as compared to control rats. The proportion of large clusters was significatively higher in control animals (Student’s t test, p b 0.005 for all groups) indicating a lower number of cell division in phenobarbital treated animals.
Discussion More than 20 years ago, it was recognized that CPA produced liver tumors in rats when given at pharmacological doses (Schuppler and Gunzel, 1979). Because this drug is widely used in clinic, many studies have adressed its carcinogenic potential in human and most of the studies have come to the agreement that the liver cancer risk associated with the use of CPA is low, if it exists (Hinkel et al., 1996; Kasper, 2001; Kattan et al., 1994; Laron and Kauli, 2000; Rabe et al., 1996). Nevertheless, progestogens are still classified as potentially carcinogenic in human by the International Agency for Research on Cancer (IARC) and it was reported that in human liver slices, CPA induced DNA adducts (Baumann et al., 1996). In the rat, CPA was long considered as a tumor promoter capable of inducing the growth of initiated cells through its mitogenic activity (Schulte-Hermann et al., 1983b; Schuppler et al., 1983) but devoid of tumor-initiating potency (Schuppler et al., 1983). It was subsequently demonstrated that CPA was also genotoxic (Deml et al., 1993; Krebs et al., 1998; Neumann et al., 1992) and it was suggested that CPA was a complete carcinogen, and that its promoter activity was linked to its mitogenic activity when administered for long period of time (Krebs et al., 1998). However, the relationship between genotoxic and mitogenic activity of CPA after acute administration were less clear. We previously reported that the mitogenic and tumor initating activities of CPA administered during a short period of time may target different cell populations (Auvigne et al., 2002). However, in all previous experimental studies dealing with CPA initiating activity, CPA administration was systematically followed by administration of strong tumor promoters such as Clophen A50 (Deml et al., 1993) or 2-AAF (Auvigne et al., 2002; Martelli et al., 1995) which are complete carcinogenic compounds. Therefore, the appearance of enzyme-altered foci in this setting could result from the concomittant administration of these carcinogens and when CPA was administered alone, no preneoplastic foci were observed in the liver (Krebs et al., 1998). To circumvent this drawback, we performed here a lineage analysis using phenobarbital, a widely used liver tumor promoter in rats (Peraino et al., 1971; Peraino et al., 1973) devoid of any genotoxic activity. We used genetic labeling to study the fate of hepatocytes dividing in response to acute CPA administration. Using h-galactosidase retroviral vectors that are only able to transduce dividing cells, labeling was specifically restricted to hepatocytes replicating in response to the acute administration of CPA. Using this approach, most of dividing hepatocytes were labeled. Indeed, we previously documented that after CPA administration the mitotic index found in hepatocytes was 1.03 F 0.08% (Pichard et al., 2001). The labeling index that we obtained in the present study (1.96 F 1.18%) is in the same range demonstrating the few, if any, replicating hepatocytes escaped retroviral infection. This indicated that we were able to follow the fate of the majority of the population of hepatocytes that are sensitive to the mitogenic effect of CPA. We did not observe the appearance of preneoplastic foci up to three months after CPA administration indicating that acute administration of CPA did not promote the growth of putatively initiated hepatocytes. After 10 months of phenobarbital promotion, the mean number of preneoplastic foci (1.45 F 0.3 foci/cm2) that we observed in treated rats was similar to the one observed in control non treated animals but clearly lower that the number of preneoplastic foci that we
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
3065
previously detected after CPA administration followed by 2-AAF treatment (38.9 F 5.45 foci/cm2) (Auvigne et al., 2002). This indicates that the results of previous studies using strong promoters were flawed because GSTp foci might have been induced by the complete carcinogenic potency of these promoters. It is therefore critical to compare various promoters in carcinogenic studies and our present findings support the previous demonstration that developemnt of preneoplastic foci is affected differently by various tumor promoting agents (Dragan and Pitot, 1992; Ledda-Columbano et al., 1989). The preneoplastic foci that we observed here did not preferentially expressed h-galactosidase. This confirmed that the hepatocytes dividing in response to acute CPA administration and that were labeled after retrovirus administration were not initiated cells committed to preneoplastic transformation. Indeed, if the mitogenic effect of CPA was linked to initiating activity we should have observed the generation of preneoplastic foci from labeled hepatocytes and hence most of the GSTp positive foci should also express h-galactosidase. We cannot exclude that the initiated hepatocytes were eliminated by the wave of apoptosis that follows withdrawal of CPA (Bursch et al., 1990). However this is unlikely because we and others previously demonstrated that the wave of apoptosis did not target specifically those hepatocytes that divided in response to acute CPA administration (Auvigne et al., 2002; Roberts et al., 1995). Finally, we also observed that phenobarbital was not mitogenic in our experimental animals. We previously documented that retroviral genetic labeling is a powerful tool to study the fate as well as the clonal expansion of normal hepatocytes in the liver. When clusters of h-galactosidase positive hepatocytes were divided in three groups (i.e. b3 cells, from 4 to 6 cells and N6 cells) we observed that the proportion of small clusters was significantly higher in the treated group than in the control rats (Table 3). If phenobarbital had triggered cell proliferation the number of small clusters should be largely lower than in control animals after 10 months of phenobarbital administration. Conversely, the number of large clusters should be higher in phenobarbital treated rats than in control animals. This is clearly not the case and therefore we believe that phenobarbital was actually unable to increase hepatocyte division in our conditions. The observation that phenobarbital is not mitogenic in rat liver may seem highly surprising since this compound has widely been described as mitogenic. However, it is noteworthy that most of the studies have used phenobarbital in conjunction with other carcinogenic compounds and at Table 3 Quantification by size of clusters of h-galactosidase positive hepatocytes Rat number
b3 cells
4 to 6 cells
N6 cells
1 2 3 4 5 mean Control Control Control Control Control mean
96.2% 95.5% 97.7% 95.3% 96.7% 96.3 F 0.93% 85.4% 73.2% 76.3% 68.6% 86.2% 77.94 F 7.7%
1.9% 3% 0% 3.1% 0% 1.6 F 1.53% 6.1% 8.2% 9.3% 17.1% 10% 10.14 F 4.2%
1.9% 1.5% 2.3% 1.6% 3.3% 2.12 F 0.73% 8.5% 18.6% 14.4% 14.3% 3.8% 11.92 F 5.8%
1 2 3 4 2
Control animals were injected with retroviral vectors 24 hours after a partial hepatectomy and were left without further treatment for 12–15 months.
3066
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
various dose levels. In contrast it has also been reported that phenobarbital, when given alone, did not increase significantly mitotic index in partially hepatectomized animals (Hashimoto et al., 1999) and administration of phenobarbital alone was able to induce liver atrophy and reduced hepatic regenerative response in a previous study (Rozga et al., 1991). Furthermore, when measuring hyperplasia through DNA synthesis, polyploidization may be confounded with true hyperplasia since phenobarbital is a potent inducer of polyploidisation (Martin et al., 2001) and accurate methods to assess the actual increase in hepatocyte number are required (Carthew et al., 1998). Our genetic labeling technics does not allow to detect such polyploidisation. Therefore although we cannot exclude that it actually increases the ploidy status of hepatocytes we believe that in our conditions (i.e. long term administration at 500 ppm in drinking water) phenobarbital does not increase the number of hepatocytes. Although the precise mechanism underlying the effect of tumor promoters is not fully understood, the importance of proteins involved in growth regulation has been reported. Indeed, phenobarbital was shown to induce expression of EGF as well as HGF receptors in centrilobuler hepatocytes (Orton et al., 1996). Beta catenin mutations were also implicated together with EGF receptor and in c-myc/TGF-alpha transgenic mice phenobarbital treatment selected for a cell population displaying activation of betacatenin and resulted in liver tumors (Calvisi et al., 2004). More recently, the nuclear receptor constitutive/active androstane receptor (CAR), also has been described as a target of phenobarbital. Indeed, mice lacking CAR are much less susceptible to liver tumor promotion by phenobarbital (Yamamoto et al., 2004). Although these previous studies were carried out in mice, the absence of promotion that we observed in our present study could be attributable to lack of, or decreased, expression of any of these potential targets. However, we believe that it is not the case since we observed obvious centrilobular hepatocyte hypertrophy indicating that our rat strain was sensitive to phenobarbital. In conclusion, our present data indicate that the mitogenic activity of CPA is not a tumor initiating activity and suggest that CPA is not carcinogenic after acute administration. Acknowledgements We wish to thank Be´atrice Sacre´-Salem for help with histological analysis and Je´roˆme Gournay for help with statistical analysis. This work was supported by grants from the Association Franc¸aise contre les Myopathies and the Ligue De´partementale contre le Cancer de Loire Atlantique. References Anon, 2000. NTP Toxicology and Carcinogenesis Studies of Primidone (CAS No. 125-33-7) in F344/N Rats and B6C3F1 Mice (Feed Studies). Natl. Toxicol. Program. Tech. Rep. Ser. 476, 1 – 290. Aubert, D., Menoret, S., Chiari, E., Pichard, V., Durand, S., Tesson, L., Moullier, P., Anegon, I., Ferry, N., 2002. Cytotoxic immune response blunts long-term transgene expression after efficient retroviral-mediated hepatic gene transfer in rat. Mol. Ther. 5, 388 – 396. Aubert, D., Pichard, V., Durand, S., Moullier, P., Ferry, N., 2003. Cytotoxic immune response after retroviral-mediated hepatic gene transfer in rat does not preclude expression from adeno-associated virus 1 transduced muscles. Hum. Gene Ther. 14, 473 – 481. Auvigne, I., Pichard, V., Aubert, D., Robillard, N., Ferry, N., 2002. In vivo cell lineage analysis in cyproterone acetate-treated rat liver using genetic labeling of hepatocytes. Hepatology 35, 281 – 288.
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
3067
Baumann, A., Kerdar, R.S., Cramer, P., Feser, W., Blode, H., Salomon, A., Kuhnz, W., 1996. Use of rat and human liver slices for the detection of steroid hormone-induced DNA-adducts in vitro by means of the (32)P-postlabeling technique. Pharmacol. Toxicol. 78, 214 – 223. Berman, J.S., Hiley, C.R., Wilson, A.C., 1983. Comparison of the effects of the hypolipidaemic agents ICI53072 and clofibrate with those of phenobarbitone on liver size, blood flow and DNA content in the rat. Br. J. Pharmacol. 78, 533 – 541. Bursch, W., Lauer, B., Timmermann-Trosiener, I., Barthel, G., Schuppler, J., Schulte-Hermann, R., 1984. Controlled death (apoptosis) of normal and putative preneoplastic cells in rat liver following withdrawal of tumor promoters. Carcinogenesis 5, 453 – 458. Bursch, W., Paffe, S., Putz, B., Barthel, G., Schulte-Hermann, R., 1990. Determination of the length of the histological stages of apoptosis in normal liver and in altered hepatic foci of rats. Carcinogenesis 11, 847 – 853. Calvisi, D.F., Ladu, S., Factor, V.M., Thorgeirsson, S.S., 2004. Activation of beta-catenin provides proliferative and invasive advantages in c-myc/TGF-alpha hepatocarcinogenesis promoted by phenobarbital. Carcinogenesis 25, 901 – 908 (Epub 2004 Jan 2023). Carthew, P., Edwards, R.E., Nolan, B.M., 1998. The quantitative distinction of hyperplasia from hypertrophy in hepatomegaly induced in the rat liver by phenobarbital. Toxicol. Sci. 44, 46 – 51. Charreau, B., Menoret, S., Tesson, L., Azimzadeh, A., Audet, M., Wolf, P., Marquet, R., Verbakel, C., Ijzermans, J., Cowan, P., Pearse, M., d’Apice, A., Soulillou, J.P., Anegon, I., 1999. Protection against hyperacute xenograft rejection of transgenic rat hearts expressing human decay accelerating factor (DAF) transplanted into primates. Mol. Med. 5, 617 – 630. Cosset, F.L., Takeuchi, Y., Battini, J.L., Weiss, R.A., Collins, M.K.L., 1995. High-titer packaging cells producing recombinant retroviruses resistant to human serum. J. Virol. 69, 7430 – 7436. Deml, E., Schwarz, L.R., Oesterle, D., 1993. Initiation of enzyme-altered foci by the synthetic steroid cyproterone acetate in rat liver foci bioassay. Carcinogenesis 14, 1229 – 1231. Dragan, Y.P., Pitot, H.C., 1992. The role of the stages of initiation and promotion in phenotypic diversity during hepatocarcinogenesis in the rat. Carcinogenesis 13, 739 – 750. Farber, E., 1956. Similarities in the sequence of early histological changes induced in the liver of the rat by ethionine, 2acetylaminofluorene and 3V-methyl-4-dimethylaminoazobenzene. Cancer Res. 16, 142 – 148. Gould, S., Sidaway, J., Sansom, N., Betton, G., Orton, T., 2001. Phenobarbitone-induced liver response in wild type and in p53 deficient mice. Toxicol. Lett. 122, 131 – 140. Hagiwara, A., Miyata, E., Tamano, S., Sano, M., Masuda, C., Funae, Y., Ito, N., Fukushima, S., Shirai, T., 1999. Noncarcinogenicity, but dose-related increase in preneoplastic hepatocellular lesions, in a two-year feeding study of phenobarbital sodium in male F344 rats. Food Chem. Toxicol. 37, 869 – 879. Hashimoto, M., Kothary, P.C., Raper, S.E., 1999. Phenobarbital in comparison with carbon tetrachloride and phenobarbitalinduced cirrhosis in rat liver regeneration. J. Surg. Res. 81, 164 – 169. Hinkel, A., Berges, R.R., Pannek, J., Schulze, H., Senge, T., 1996. Cyproterone acetate in the treatment of advanced prostatic cancer: retrospective analysis of liver toxicity in the long-term follow-up of 89 patients. Eur. Urol. 30, 464 – 470. Kasper, P., 2001. Cyproterone acetate: a genotoxic carcinogen? Pharmacol. Toxicol. 88, 223 – 231. Kattan, J., Spatz, A., Culine, S., Terrier-Lacombe, M.J., Elias, D., Droz, J.P., 1994. Hepatocellular carcinoma during hormonotherapy for prostatic cancer. Am. J. Clin. Oncol. 17, 390 – 392. Kitten, O., Cosset, F.L., Ferry, N., 1997. Highly efficient retrovirus-mediated gene transfer into rat hepatocytes in vivo. Hum. Gene Ther. 8, 1491 – 1494. Krebs, O., Schafer, B., Wolff, T., Oesterle, D., Deml, E., Sund, M., Favor, J., 1998. The DNA damaging drug cyproterone acetate causes gene mutations and induces glutathione-S-transferase P in the liver of female Big Blue transgenic F344 rats. Carcinogenesis 19, 241 – 245. Laron, Z., Kauli, R., 2000. Experience with cyproterone acetate in the treatment of precocious puberty. J. Pediatr. Endocrinol. Metab. 13, 805 – 810. Ledda-Columbano, G.M., Columbano, A., Curto, M., Ennas, M.G., Coni, P., Sarma, D.S., Pani, P., 1989. Further evidence that mitogen-induced cell proliferation does not support the formation of enzyme-altered islands in rat liver by carcinogens. Carcinogenesis 10, 847 – 850. Martelli, A., Mattioli, F., Fazio, S., Andrae, U., Brambilla, G., 1995. DNA repair synthesis and DNA fragmentation in primary cultures of human and rat hepatocytes exposed to cyproterone acetate. Carcinogenesis 16, 1265 – 1269. Martin, N.C., McGregor, A.H., Sansom, N., Gould, S., Harrison, D.J., 2001. Phenobarbitone-induced ploidy changes in liver occur independently of p53. Toxicol. Lett. 119, 109 – 115.
3068
V. Pichard, N. Ferry / Life Sciences 76 (2005) 3057–3068
Menoret, S., Aubert, D., Tesson, L., Braudeau, C., Pichard, V., Ferry, N., Anegon, I., 2002. lacZ Transgenic Rats Tolerant for beta-Galactosidase: Recipients for Gene Transfer Studies Using lacZ as a Reporter Gene. Hum. Gene Ther. 13, 1383 – 1390. Neumann, I., Thierau, D., Andrae, U., Greim, H., Schwarz, L.R., 1992. Cyproterone acetate induces DNA damage in cultured rat hepatocytes and preferentially stimulates DNA synthesis in gamma-glutamyltranspeptidase-positive cells. Carcinogenesis 13, 373 – 378. Orton, T.C., Doughty, S.E., Kalinowski, A.E., Lord, P.G., Wadsworth, P.F., 1996. Expression of growth factors and growth factor receptors in the liver of C57BL/10J mice following administration of phenobarbitone. Carcinogenesis 17, 973 – 981. Peraino, C., Fry, R.J., Staffeldt, E., 1971. Reduction and enhancement by phenobarbital of hepatocarcinogenesis induced in the rat by 2-acetylaminofluorene. Cancer Res. 31, 1506 – 1512. Peraino, C., Fry, R.J., Staffeldt, E., Kisieleski, W.E., 1973. Effects of varying the exposure to phenobarbital on its enhancement of 2-acetylaminofluorene-induced hepatic tumorigenesis in the rat. Cancer Res. 33, 2701 – 2705. Pichard, V., Aubert, D., Ferry, N., 2001. Efficient retroviral gene transfer to the liver in vivo using nonpolypeptidic mitogens. Biochem. Biophys. Res. Commun. 286, 929 – 935. Rabe, T., Feldmann, K., Heinemann, L., Runnebaum, B., 1996. Cyproterone acetate: is it hepato- or genotoxic? Drug Saf. 14, 25 – 38. Reuber, M., 1964. Development of preneoplastic and neoplastic lesions of the liver in male rats given 0.025 percent N-2fluorenyldiacetamide. J. Intl. Canc. Instit. 34, 697 – 723. Roberts, R.A., Soames, A.R., Gill, J.H., James, N.H., Wheeldon, E.B., 1995. Non-genotoxic hepatocarcinogens stimulate DNA synthesis and their withdrawal induces apoptosis, but in different hepatocyte populations. Carcinogenesis 16, 1693 – 1698. Rozga, J., Foss, A., Alumets, J., Ahren, B., Jeppsson, B., Bengmark, S., 1991. Liver cirrhosis in rats: regeneration and assessment of the role phenobarbital. J. Surg. Res. 51, 329 – 335. Satoh, K., Kitahara, A., Soma, Y., Inaba, Y., Hatayama, I., Sato, K., 1985. Purification, induction, and distribution of placental glutathione transferase: a new marker enzyme for preneoplastic cells in the rat chemical hepatocarcinogenesis. Proc. Natl. Acad. Sci. USA 82, 3964 – 3968. Schulte-Hermann, R., Hoffman, V., Parzefall, W., Kallenbach, M., Gerhardt, A., Schuppler, J., 1980. Adaptive responses of rat liver to the gestagen and anti-androgen cyproterone acetate and other inducers. II. Induction of growth. Chem. Biol. Interact. 31, 287 – 300. Schulte-Hermann, R., Schuppler, J., Timmermann-Trosiener, I., Ohde, G., Bursch, W., Berger, H., 1983a. The role of growth of normal and preneoplastic cell populations for tumor promotion in rat liver. Environ Health Perspect. 50, 185 – 194. Schulte-Hermann, R., Timmermann-Trosiener, I., Schuppler, J., 1983b. Promotion of spontaneous preneoplastic cells in rat liver as a possible explanation of tumor production by nonmutagenic compounds. Cancer Res. 43, 839 – 844. Schuppler, J., Damme, J., Schulte-Hermann, R., 1983. Assay of some endogenous and synthetic sex steroids for tumor-initiating activity in rat liver using the Solt-Farber system. Carcinogenesis 4, 239 – 241. Schuppler, J., Gunzel, P., 1979. Liver tumors and steroid hormones in rats and mice. Arch. Toxicol., 181 – 195. Solt, D., Medline, A., Farber, E., 1977. Rapid ermergence of carcinogen-induced hyperplastic lesions in a new model for the sequential analysis of liver carcinogenesis. Am. J. Pathol. 88, 595 – 610. Yamamoto, Y., Moore, R., Goldsworthy, T.L., Negishi, M., Maronpot, R.R., 2004. The orphan nuclear receptor constitutive active/androstane receptor is essential for liver tumor promotion by phenobarbital in mice. Cancer Res. 64, 7197 – 7200.