Application of comparative functional genomics to identify best-fit mouse models to study human cancer

Genetically modified mice have been extensively used for analyzing the molecular events that occur during tumor development. In many, if not all, cases, however, it is uncertain to what extent the mouse models reproduce features observed in the corresponding human conditions. This is due largely to lack of precise methods for direct and comprehensive comparison at the molecular level of the mouse and human tumors. Here we use global gene expression patterns of 68 hepatocellular carcinomas (HCCs) from seven different mouse models and 91 human HCCs from predefined subclasses to obtain direct comparison of the molecular features of mouse and human HCCs. Gene expression patterns in HCCs from Myc, E2f1 and Myc E2f1 transgenic mice were most similar to those of the better survival group of human HCCs, whereas the expression patterns in HCCs from Myc Tgfa transgenic mice and in diethylnitrosamine-induced mouse HCCs were most similar to those of the poorer survival group of human HCCs. Gene expression patterns in HCCs from Acox1(-/-) mice and in ciprofibrate-induced HCCs were least similar to those observed in human HCCs. We conclude that our approach can effectively identify appropriate mouse models to study human cancers.

Vitamin E down-modulates iNOS and NADPH oxidase in c-Myc/TGF-alpha transgenic mouse model of liver cancer

BACKGROUND/AIMS

Co-expression of c-Myc and TGF-alpha in the mouse liver accelerates hepatocarcinogenesis and enhances DNA damage due to chronic oxidative stress. Dietary supplementation with vitamin E (VE) inhibits hepatocarcinogenesis and reduces chromosomal alterations in the same mice. Here we investigated the sources of reactive oxygen species (ROS) production in c-Myc/TGF-alpha transgenic mice.

METHODS

Inducible nitric oxide synthase (iNOS) and NADPH oxidase levels were determined in c-Myc, TGF-alpha and c-Myc/TGF-alpha mice by RT-PCR, western blot analysis and immunohistochemistry.

RESULTS

iNOS and nitrotyrosines levels were higher in the three transgenic lines when compared with wild-type mice. Preneoplastic and neoplastic lesions from c-Myc, TGF-alpha and c-Myc/TGF-alpha transgenic mice displayed upregulation of NADPH oxidase subunits p47-, 67-phox, Rac1, HSP 70, and HO-1. Importantly, dietary supplementation with vitamin E abolished iNOS expression, lowered nitrotyrosines, p47-, p67-phox, and Rac1 levels, and suppressed HSP 70 and HO-1 proteins in c-Myc/TGF-alpha livers.

CONCLUSIONS

The results suggest that iNOS and NADPH oxidase are involved in ROS generation during c-Myc/TGF-alpha hepatocarcinogenesis and are inhibited by VE treatment. The data provide additional evidence for the potential use of VE in treatment of chronic liver diseases and HCC prevention.

Disregulation of E-cadherin in transgenic mouse models of liver cancer

E-cadherin is a cell-cell adhesion molecule that plays a pivotal role in the development and maintenance of cell polarity. Disruption of E-cadherin-mediated adhesion represents a key step toward the invasive phenotype in a variety of solid tumors, including hepatocellular carcinoma (HCC). Here, we investigate whether deregulation of E-cadherin occurs along the multistep process of hepatocarcinogenesis in transgenic mouse models, including c-Myc, E2F1, c-Myc/TGF-alpha and c-Myc/E2F1 mice. Liver tumors from the transgenic mouse lines could be divided into two categories based on E-cadherin levels. Of 28, 20 (71.4%) c-Myc HCCs showed marked reduction of E-cadherin expression when compared with wild-type livers. In contrast, all of c-Myc/TGF-alpha and the majority of E2F1 and c-myc/E2F1 preneoplastic and neoplastic lesions exhibited overexpression of E-cadherin. Downregulation of E-cadherin was associated with promoter hypermethylation in seven of 20 c-Myc HCCs (35%), while no loss of heterozygosity at the E-cadherin locus was detected. Nuclear accumulation of beta-catenin did not correlate with E-cadherin downregulation. Furthermore, c-Myc HCCs with reduced E-cadherin displayed upregulation of hypoxia-inducible factor-1alpha and vascular endothelial growth factor proteins. Importantly, loss of E-cadherin was associated with increased cell proliferation and higher microvessel density in c-Myc tumors. Taken together, these data suggest that loss of E-cadherin might favor tumor progression in relatively more benign HCC from c-Myc transgenic mice by stimulating neoplastic proliferation and angiogenesis under hypoxic conditions.

Classification and prediction of survival in hepatocellular carcinoma by gene expression profiling

We analyzed global gene expression patterns of 91 human hepatocellular carcinomas (HCCs) to define the molecular characteristics of the tumors and to test the prognostic value of the expression profiles. Unsupervised classification methods revealed two distinctive subclasses of HCC that are highly associated with patient survival. This association was validated via 5 independent supervised learning methods. We also identified the genes most strongly associated with survival by using the Cox proportional hazards survival analysis. This approach identified a limited number of genes that accurately predicted the length of survival and provides new molecular insight into the pathogenesis of HCC. Tumors from the low survival subclass have strong cell proliferation and antiapoptosis gene expression signatures. In addition, the low survival subclass displayed higher expression of genes involved in ubiquitination and histone modification, suggesting an etiological involvement of these processes in accelerating the progression of HCC. In conclusion, the biological differences identified in the HCC subclasses should provide an attractive source for the development of therapeutic targets (e.g., HIF1a) for selective treatment of HCC patients. Supplementary material for this article can be found on the HEPATOLOGY Web site (http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html)

Polygenic control of hepatocarcinogenesis in Copenhagen x F344 rats

Cop and CFF1 rats exhibit resistance to hepatocarcinogenesis, associated with high rates of remodeling of neoplastic lesions. We have mapped hepatocarcinogenesis susceptibility, resistance and remodeling loci affecting the number, volume and volume fraction of neoplastic nodules induced by the "resistant hepatocyte" model in male CFF2 rats. Three loci in significant linkage with the number or volume of nonremodeling lesions were identified on chromosomes 1, 4 and 18. Suggestive linkage with number or volume fraction of total, nonremodeling or remodeling lesions was found for 7 loci on chromosomes 1, 2, 13, 14 and 15. All of these loci showed significant allele-specific effects on the phenotypic traits. We also detected by analysis of variance 19 2-way interactions inducing phenotypic effects not predictable on the basis of the sum of separate effects. These novel epistatic loci were in significant linkage with the number and/or volume of total, nonremodeling or remodeling nodules. These data indicate that susceptibility to hepatocarcinogenesis in Cop rats is controlled by a complex array of genes with several gene-gene interactions and that different genetic mechanisms control remodeling and nonremodeling liver nodules. Frequent deregulation in human liver cancer of genes positioned in chromosomal segments syntenic to rat susceptibility/resistance loci suggests some similarities between the genetic mechanisms involved in hepatocarcinogenesis in rats and humans.

Disruption of beta-catenin pathway or genomic instability define two distinct categories of liver cancer in transgenic mice

BACKGROUND & AIMS

Human liver cancer can be divided into 2 categories that are characterized by activation of beta-catenin and genomic instability. Here we investigate whether similar categories exist among 5 transgenic models of liver cancer, including c-myc, transforming growth factor-alpha, E2F-1, c-myc/transforming growth factor-alpha, and c-myc/E2F-1 mice.

METHODS

The random amplified polymorphic DNA method was used to assess the overall genomic instability, and chromosomal loci affected by genomic alterations were determined by microsatellite analysis. beta-Catenin mutations and deletions were analyzed by polymerase chain reaction and sequencing screening. Cellular localization of beta-catenin and expression of alpha-fetoprotein, a prognostic marker of hepatocellular carcinoma, were investigated by immunohistochemistry.

RESULTS

Liver tumors from the transgenic mice could be divided into 2 broad categories characterized by extensive genomic instability (exemplified by the c-myc/transforming growth factor-alpha mouse) and activation of beta-catenin (exemplified by the c-myc/E2F-1 mouse). The c-myc/transforming growth factor-alpha tumors displayed extensive genomic instability with recurrent loss of heterozygosity at chromosomes 1, 2, 4, 6, 7, 9, 12, 14, and X and a low rate of beta-catenin activation. The genomic instability was evident from the early dysplastic stage and occurred concomitantly with increased expression of alpha-fetoprotein. The c-myc/E2F-1 tumors were characterized by a high frequency of beta-catenin activation in the presence of a relatively stable genome and low alpha-fetoprotein levels.

CONCLUSIONS

We have identified 2 prototype experimental models, i.e., c-myc/transforming growth factor-alpha and c-myc/E2F-1 mice, for the 2 categories of human hepatocellular carcinoma characterized by genomic instability and beta-catenin activation, respectively. These mouse models will assist in the elucidation of the molecular basis of human hepatocellular carcinoma.

Activation of beta-catenin provides proliferative and invasive advantages in c-myc/TGF-alpha hepatocarcinogenesis promoted by phenobarbital

Previously, we have found that phenobarbital (PB) enhanced cell survival and facilitated tumor growth in our c-myc/transforming growth factor (TGF)-alpha transgenic mouse model of liver cancer. Given that PB selectively promoted initiated cells harboring beta-catenin mutations during chemically induced hepatocarcinogenesis and that Wnt/beta-catenin signaling is involved in both anti-apoptotic and proliferative processes, we now have extended our analysis to investigate whether promotion by PB affects the occurrence of beta-catenin mutations in c-myc/TGF-alpha-driven tumors. The frequency of beta-catenin activation as judged by somatic mutations and/or nuclear localization was significantly increased in hepatocellular carcinomas (HCCs) from c-myc/TGF-alpha mice treated with PB (15/28; 53.6%) as compared with that in control HCCs (2/28; 7.1%). Furthermore, an intact beta-catenin locus was detected in all neoplasms following PB treatment, whereas 57.1% (16/28) of malignant tumors from c-myc/TGF-alpha untreated mice displayed loss of heterozygosity at the beta-catenin locus. Strikingly, in the majority of PB-treated HCCs beta-catenin nuclear localization was limited to small cells with high nuclear/cytoplasmic ratio forming an invasion front (NAinv). beta-Catenin NAinv cells showed cytoplasmic redistribution of E-cadherin associated with intense mucin 1 and matrilysin immunostaining, suggesting their invasive phenotype. All beta-catenin-positive HCCs displayed increased proliferation and tumor size, but no difference in the apoptotic rate when compared with beta-catenin negative tumors. These findings show that PB treatment positively selects for a cell population displaying activation of beta-catenin in c-myc/TGF-alpha HCCs. beta-Catenin activation confers additional growth and invasive advantages in a model of liver cancer already accelerated by synergistic activity of the c-myc and TGF-alpha transgenes.

Reversal of liver fibrosis in aryl hydrocarbon receptor null mice by dietary vitamin A depletion

Aryl hydrocarbon receptor (AHR)-null mice display a liver fibrosis phenotype that is associated with a concomitant increase in liver retinoid concentration, tissue transglutaminase type II (TGaseII) activity, transforming growth factor beta (TGF beta) overexpression, and accumulation of collagen. To test the hypothesis that this phenotype might be triggered by the observed increase in liver retinoid content, we induced the condition of retinoid depletion by feeding AHR-null mice a vitamin A- deficient diet with the purpose to reverse the phenotype. Liver retinoid content decreased sharply within the first few weeks on the retinoid-deficient diet. Analysis of TGF beta 1, TGF beta 2, and TGF beta 3 expression revealed a reduction to control levels in the AHR -/- mice accompanied by parallel changes in TGaseII protein levels. In addition, we observed an increase in the TGF beta receptors, TGF beta RI and TGF beta RII, as well as in Smad4, and their reduction to wild-type mouse liver levels in AHR -/- mice fed the retinoid-deficient diet. Reduction of peroxisomal proliferator-activated receptor gamma (PPAR gamma) messenger RNA (mRNA) and protein levels in AHR -/- mice was consistent with the presence of hepatic stellate cell (HSC) activation and liver fibrosis. Vitamin A deficiency normalized PPAR gamma expression in AHR -/- mice. In conclusion, livers from AHR -/- mice fed the vitamin A-deficient diet showed a decrease in collagen deposition, consistent with the absence of liver fibrosis.

Phenotypic reversion of rat neoplastic liver nodules is under genetic control

Low DNA synthesis and high redifferentiation (remodeling) characterize neoplastic nodules induced by chemical carcinogens in hybrid BFF1 rats, generated by crossing the susceptible F344 and resistant BN strains. We performed whole-genome scanning of BFF2 rats to identify loci controlling remodeling of nodules induced, 32 weeks after initiation with diethylnitrosamine, by the RH protocol. Remodeling nodules were identified as areas lacking uniformity of GST-P immunostaining and with irregular margins. Two loci in suggestive linkage with the percentage of remodeling nodules were identified on chromosomes 7 and 1 (LOD scores 3.85 and 2.9 at D7Rat25 and D1Mgh14). Significant dosage-negative effect of the B allele on remodeling and additive interaction between these loci were found. Significant epistatic interactions, showing a recessive, remodeling-enhancing effect of B alleles, occurred between D1Mit3 and D11Rat11 (corrected p = 0.0013) and between D6Rat14 and D8Rat46 (corrected p = 0.028). These data show that remodeling of neoplastic nodules during rat hepatocarcinogenesis is under genetic control. Loci affecting remodeling map to chromosomal regions syntenic to chromosomal segments of human HCC showing structural abnormalities.

Dysregulation of DNA repair pathways in a transforming growth factor alpha/c-myc transgenic mouse model of accelerated hepatocarcinogenesis

Previous work from our laboratory has implicated oxidative DNA damage and genetic instability in the etiology of transforming growth factor-alpha (TGFalpha)/c-myc-associated hepatocarcinogenesis. In contrast, oxidative DNA damage was lower in c-myc single-transgenic mice, consistent with less chromosomal damage and with later and more benign tumor formation. We examined whether defects in the DNA repair pathways contribute to the acceleration of liver cancer in TGFalpha/c-myc mice. A cDNA expression array containing 140 known genes and multiplex RT-PCR were used to compare the basal levels of expression of DNA repair genes at the dysplastic stage. Thirty-five percent (8/23) and 43% (10/23) of DNA repair genes were constitutively up-regulated in 10-week-old TGFalpha/c-myc and c-myc transgenic livers, respectively, compared with wild-type controls. The commonly up-regulated genes were OGG1 and NTH1 of base excision repair; ERCC5, RAD23A, and RAD23B of nucleotide excision repair; and RAD50, RAD52, and RAD54 involved in DNA strand break repair. Additional treatment with a peroxisome proliferator, Wy-14,643, known to increase the level of oxidants in the liver, failed to induce a further increase in the expression level of DNA repair enzymes in TGFalpha/c-myc but not in c-myc or wild-type livers. Moreover, expression of several genes, including Ku80, PMS2, and ATM, was decreased in TGFalpha/c-myc livers, suggesting a fault or inefficient activation of the DNA repair pathway upon induction of oxidative stress. Together, the results show that DNA damage response is attenuated in TGFalpha/c-myc mice, creating a condition that may contribute to acceleration of liver cancer in this model.

Chromosome mapping of multiple loci affecting the genetic predisposition to rat liver carcinogenesis

Previous studies on (BNxF344)F1 (BFF1) rat model of genetic predisposition to hepatocarcinogenesis led to the identification, in BFF1xF344 backcross progeny, of two hepatocarcinogenesis susceptibility (Hcs) and three resistance (Hcr) loci affecting the progression of neoplastic liver nodules. To evaluate the presence of other hepatocarcinogenesis-related loci in the BFF1 genome, nodule induction by resistant hepatocyte model in 116 male BFF2 rats 32 weeks after initiation with diethylnitrosamine was subjected to quantitative trait loci analysis. The rats were typed with 179 genetic markers, and linkage analysis identified three loci on chromosomes 1, 16, and 6, in significant linkage with nodule mean volume (V), volume fraction, and number, respectively, and two loci on chromosomes 4 and 8 in suggestive linkage with V. These loci were differently positioned with respect to Hcs and Hcr loci mapped previously in backcross rats. On the basis of phenotypic and allele distribution patterns of BFF2 rats, loci on chromosomes 1 and 16 were identified as Hcs3 and Hcs4, and loci on chromosomes 4, 8, and 6 as Hcr4, Hcr5, and Hcr6. Additive interactions occurred between Hcs3 and Hcs4, and Hcr4 and a locus on chromosome 3 with less than suggestive linkage with V. All of the loci were in chromosomal regions syntenic to mouse and/or human chromosomal segments showing allelic gain or loss in hepatocellular carcinomas. These data indicate that inheritance of predisposition to rat liver tumor is characterized by the interplay of several genetic factors and suggest some possible mechanisms of polygenic control of human liver cancer.

Cell cycle deregulation in liver lesions of rats with and without genetic predisposition to hepatocarcinogenesis

Preneoplastic and neoplastic hepatocytes undergo c-Myc up-regulation and overgrowth in rats genetically susceptible to hepatocarcinogenesis, but not in resistant rats. Because c-Myc regulates the pRb-E2F pathway, we evaluated cell cycle gene expression in neoplastic nodules and hepatocellular carcinomas (HCCs), induced by initiation/selection (IS) protocols 40 and 70 weeks after diethylnitrosamine treatment, in susceptible Fisher 344 (F344) rats, and resistant Wistar and Brown Norway (BN) rats. No interstrain differences in gene expression occurred in normal liver. Overexpression of c-myc, Cyclins D1, E, and A, and E2F1 genes, at messenger RNA (mRNA) and protein levels, rise in Cyclin D1-CDK4, Cyclin E-CDK2, and E2F1-DP1 complexes, and pRb hyperphosphorylation occurred in nodules and HCCs of F344 rats. Expression of Cdk4, Cdk2, p16(INK4A), and p27(KIP1) did not change. In nodules and/or HCCs of Wistar and BN rats, low or no increases in c-myc, Cyclins D1, E, and A, and E2F1 expression, and Cyclin-CDKs complex formation were associated with p16(INK4A) overexpression and pRb hypophosphorylation. In conclusion, these results suggest deregulation of G1 and S phases in liver lesions of susceptible rats and block of G1-S transition in lesions of resistant strains, which explains their low progression capacity.

Frequent loss of heterozygosity at the Hcr1 (hepatocarcinogenesis resistance) locus on chromosome 10 in primary hepatocellular carcinomas from LFF1 rat strain

Hepatocarcinogenesis sensitivity (Hcs1, 2) and resistance (Hcr1-3) loci have been identified by linkage analysis on rat chromosomes 7 and 1, and 10, 4, and 8, respectively. Cytogenetic studies documented deletions on chromosomes 3 and 6 of neoplastic rat hepatocytes. Hepatocellular carcinomas (HCCs) were produced in F1 hybrid rats between Long-Evans (LE) and Fisher 344 (F344) rats. Scanning of the above chromosomes for loss of heterozygosity (LOH) showed allelic imbalance (AI) at multiple regions on chromosomes 6, 7, and 10q. Detailed deletion mapping of chromosome 10 localized a putative suppressor Hcr1 gene to within a 3.2-cM interval flanked by D10Rat51 and D10Rat121. Two other distinct regions with frequent AIs were found inside the Hcr1 locus, at marker loci including DNaseI and Mrp genes, and in a segment including 4 consecutive markers (D10Rat64, D10Rat182, D10Rat113, D10Rat216). In 40% of HCCs, AI was seen at the p53 locus. AI on chromosome 7 occurred at the Hcs1 locus, where is located c-myc, which is amplified in HCCs, suggesting allelic gain. Most AIs occurred in poorly/moderately differentiated carcinomas, and a few events were seen in well-differentiated tumors on chromosomes 7 and 10. These data suggest that alteration of a cluster of oncosuppressor genes on 10q is important for HCC progression. The existence of AI on segments of rat chromosomes 6, 7, and 10, syntenic to chromosomal segments of human HCCs where chromosomal gains or deletions occur, suggests a commonality of some molecular events in the pathogenesis of HCCs in rats and humans. Our map provides information toward cloning putative oncosuppressor genes associated with this carcinoma.

5'-Methylthioadenosine administration prevents lipid peroxidation and fibrogenesis induced in rat liver by carbon-tetrachloride intoxication

BACKGROUND

5'-Methylthioadenosine (MTA), a product of S-adenosylmethionine (SAM) catabolism, could undergo oxidation by mono-oxygenases and auto-oxidation. MTA and SAM effects on oxidative liver injury were evaluated in CCl4-treated rats.

METHODS

Male Wistar rats were killed 1-48 h after poisoning with a single intraperitoneal CCl4 dose (0.15 ml/100 g) or with the same dose twice a week for 14 weeks. Daily doses of MTA or SAM (384 micromol/kg), started 1 week before acute CCl4 administration or with chronic treatment, were continued up to the time of sacrifice.

RESULTS

Acute and chronic CCl4 intoxication decreased MTA and, to a lesser extent, SAM and reduced glutathione (GSH) liver levels. MTA administration increased liver MTA without affecting SAM and GSH. SAM treatment caused complete/partial recovery of these compounds. MTA and, to a lesser extent, SAM prevented an increase in liver phospholipid hydroperoxides in acutely and chronically intoxicated rats and in prolyl hydroxylase activity and trichrome-positive areas in chronically treated rats. MTA prevented upregulation of Tgf-beta1, Collagen-alpha1 (I) and Tgf-alpha genes in liver of chronically intoxicated rats, and TGF-beta1-induced transdifferentiation to myofibroblasts and growth stimulation by platelet-derived growth factor-b of stellate cells in vitro.

CONCLUSIONS

MTA and SAM protect against oxidative liver injury through partially different mechanisms.

Activation of beta-catenin during hepatocarcinogenesis in transgenic mouse models: relationship to phenotype and tumor grade

Mutations affecting phosphorylation sites in the beta-catenin gene have been implicated in the development of human and rodent hepatocellular carcinomas (HCCs). To further investigate the involvement of this gene in hepatocarcinogenesis, we used several transgenic mouse models of hepatic tumors induced by overexpression of c-myc in the liver either alone or in combination with transforming growth factor (TGF) alpha or TGF-beta1. Activation of beta-catenin, as judged by the presence of mutations and/or nuclear translocation of the protein, was most frequent in liver tumors from c-myc (4/17; 23.5%) and c-myc/TGF-beta1 (6/18; 33.3%) transgenic mice. However, it was very rare in faster growing and histologically more aggressive HCCs developed in c-myc/TGF-alpha mice (1/20; 5%). Administration of diethylnitrosamine, phenobarbital, or 2-amino-3,8-diethylimidazo[4,5-f]quinoxaline did not significantly affect the occurrence of beta-catenin mutations. Notably, nuclear accumulation of beta-catenin was observed only in adenomas and highly differentiated carcinomas with eosinophilic phenotype. Furthermore, preneoplastic lesions with eosinophilic phenotype frequently displayed focal nuclear positivity, colocalized with areas of high proliferation. In contrast, basophilic and clear-cell foci, as well as pseudo-glandular and poorly differentiated HCCs, exhibited a normal or reduced membranous immunoreactivity for beta-catenin. These studies suggest that nuclear translocation of beta-catenin and activation of Wingless/Wnt signaling may represent an early event in liver carcinogenesis, providing a growth advantage in a subset of hepatic tumors with a more differentiated phenotype.

Implication of Bcl-2 family genes in basal and D-amphetamine-induced apoptosis in preneoplastic and neoplastic rat liver lesions

Molecular mechanisms of basal and D-amphetamine (AMPH)-induced apoptosis were studied in rat liver nodules, 12 (N12) and 30 (N30) weeks after initiation, and in hepatocellular carcinoma (HCC) induced by diethylnitrosamine in rats subjected to resistant hepatocyte model. Basal apoptosis in hematoxylin/eosin- and propidium iodide-stained sections was higher in nodules and HCC than in normal livers. It sharply increased in all tissues 4 hours after AMPH treatment (10 mg/kg), and declined to basal levels at 8 to 12 hours in liver and N12, but remained high up to 18 hours in N30 and HCC. c-myc, Tgf-alpha, p53, and Bcl-X(S) messenger RNA (mRNA) levels were higher, and Bcl-2 mRNA was lower in N12 and/or N30 and HCC than in normal liver. Four hours after AMPH injection, increase in c-myc and decreases in Bcl-2 and Bcl-X(L) mRNAs occurred in all tissues, whereas p53, Bax, and Bcl-X(S) mRNAs increased in N30 and HCC. These changes disappeared in liver and N12 at 18 hours, but persisted in N30 and HCC. c-Myc, P53, Bcl-2, and Bax proteins in normal liver and HCC +/- AMPH showed similar patterns. Tgf-beta1, Tgf-beta-RIII, CD95, and CD95L mRNA levels underwent slight or no changes in any tissue +/- AMPH. Basal Hsp27 expression was high in nodules and HCC, and was stimulated by AMPH in liver and N12, but not in N30 and HCC. These data suggest a role of dysregulation of Bcl-2 family genes and, at least in atypical lesions, of p53 overexpression, in basal and AMPH-induced apoptosis in nodules and HCCs. Hsp27 does not appear to sufficiently protect atypical lesions against apoptosis.