Role of cyclin E in the pathogenesis of hepatocellular carcinoma
Abstract
Hepatocellular carcinoma (HCC), or primary liver cancer, is the fifth most common cancer worldwide and the third most common cause of cancer mortality (El-Serag 2012). The development of HCC is thought to be a multi-staged process that involves several risk factors including the chronic hepatitis B and C infection, carcinogen exposure, metabolic disease, excessive alcohol consumption and male gender. Accumulation of genetic and epigenetic alterations with DNA-damaged hepatocytes can also contribute to the molecular pathogenesis of HCC. A better understanding of molecular mechanisms associated with HCC could ultimately improve our current strategies for screening and targeted therapy of this disease.
Array comparative genomic hybridisation studies in murine diethylnitrosamine (DEN)-induced HCC have identified ccne1 (cyclin E) as a candidate gene associated in accelerated liver carcinogenesis (Teoh et al. 2008). In order to investigate the role of cyclin E and its impact on hepatocyte cell cycle regulation in early HCC development, we employed a well-known and highly reproducible rodent model of DEN-induced hepatocarcinogenesis. C57BL/6J male mice were injected with DEN (10 mg/kg i.p.), age day 12-15. In this model, male animals develop hepatocyte dysplasia at 6 mths and HCC in >90%. Transcript and protein expression of cyclin E was evident as early as 6 mths in dysplastic nodules (DNs), and significantly increased in HCCs. In contrast, there was little or undetectable cyclin E in normal liver and liver surrounding HCCs at all timepoints. Glutathione S transferase-pi form and cyclin E expression co-localised in DNs from liver in mice at 6 and 9 mths. Cyclin E/cdk2 kinase activity was also significantly upregulated in DNs, while increased proliferative activity by cyclin D1 and proliferating cell nuclear antigen (PCNA) in HCCs was observed at 9 mths, a timepoint where there was maximal p53 and p21 tumour suppressor expression. Aberrant cyclin E protein expression, including low molecular weight (LMW) isoforms were detected in HCCs and liver surrounding HCCs. Interestingly, sequencing analyses of p53 revealed a 1093-1361 nucleotide deletion in up to 90% of DNs, causing dysfunctional p53 nuclear localisation and export signalling. Because p53 directly signals to p21, co-immunoprecipitation studies were performed and revealed preferential binding of p21 to cyclin D1, rather than cyclin E, thereby allowing "escape" from the G1/S checkpoint.
To directly test whether cyclin E regulates p53 expression in HCCs derived from DEN-treated male mice at 9 mths, we conducted cyclin E knockdown in primary HCC cells. This strategy resulted in increased p53 and p21 expression, as well as significant diminution of Bcl-xL, the p53-induced anti-apoptotic marker. Cell viability tetrazolium/formazan assay was significantly impaired in cyclin E RNAi targeted primary HCC cells. Conversely, chemical inhibition of p53 by pfithrin-α, augmented cyclin E, PCNA and Bcl-xL protein expression whilst cell viability was restored following co-treatment with MG-262, a 26S proteasome inhibitor. In contrast, overexpressing cyclin E in naïve primary hepatocytes enhanced PCNA expression, increased hepatocyte viability, downregulated p53 and its downstream signalling intermediate, p21.
We next determined whether miRNA-34, a co-regulator of cyclin E and p53, was instrumental in the reciprocity between cyclin E and p53 as key "drivers"of hepatocarcinogenesis. Dysplastic liver and HCCs obtained from DEN-injected male mice were assayed for miR-34a,b,c. miR-34a and c were significantly upregulated in HCCs and dysplastic liver compared with normal liver. Similar trends were noted for miR-34a,b,c in human hepatitis C-related HCCs when compared with normal human liver. Importantly, this was associated with significantly enhanced cyclin E and p53 mRNA expression in human HCCs compared to normal and cirrhotic liver.
In this murine model, there was disproportional and upregulation of functionally active cyclin E, miR-34a,c in DNs and early HCCs with congruent loss of p53 function associated with cell cycle checkpoint failure, diminished apoptosis and increased proliferative drive. In human HCV-related HCCs, miR-34a, p53 and cyclin E transcript levels were universally upregulated. When we performed in vitro experiments in murine primary hepatocytes and primary HCC cells, knocking down or overexpressing cyclin E did not affect miR-34 expression. However, stabilising p53 with MG-262 enhanced miR-34a,c expression (though not significant), whilst inhibiting p53 using pfithrin-α significantly reduced miR-34a,c. miR-34 may provide a plausible link to increased cyclin E expression, activity and increased proliferative drive in dysplastic and neoplastic liver in mice and humans.
Gender disparity in human HCC is well described, with a strong male predominance. However, the role of sex hormones in hepatocarcinogenesis remains poorly defined. In order to determine if there are gender differences in the expression of cyclin E, effects on cell cycle regulators and tumour suppressors, dysplastic liver and HCCs were studied in DEN-treated C57BL/6J female mice. These mice displayed a significant reduction in dysplastic hepatocytes compared with intact DEN-treated males at 3, 6 mths, while HCC incidence, number and size of tumours were significantly diminished in females at up to 15 mths. In carcinogen-treated female mice, cyclin E (native and LMW isoforms) protein expression and kinase activity were reduced compared to males at 6-12 mths, with concomitant reduction in hepatocyte proliferation by PCNA and cyclin D1 expression. Unlike male mice, G1/S checkpoint is evident by robust p53-mediated apoptosis.
To ascertain if these differences are attributable to the effects of oestradiol/progesterone (E/P) and/or testosterone, we conducted hormonal manipulation studies using the same carcinogen-model by performing ovariectomy in female animals and orchidectomy in male mice, in which some animals received E/P or testosterone supplementation. Castration of DEN-injected male mice resulted in a loss of cyclin E LMW isoforms compared to intact males, diminution of cyclin E kinase activity and phospho-retinoblastoma expression. There was also induction of p53-mediated apoptosis in dysplastic hepatocytes, leading to a reduction in number of DNs by 6 mths. These anti-proliferative and pro-apoptotic effects were magnified by E/P replacement in castrated DEN-treated males. In contrast, testosterone-replacement in ovariectomised-female mice exhibited accelerated hepatocarcinogenesis compared to intact female DEN-treated animals and displayed LMW cyclin E isoforms similar to those detected in DEN-injected intact males.
In further analyses, there was increased oestrogen receptor-α (ERα) transcript and protein expression in HCCs derived from DEN-injected intact male mice, and in dysplastic liver from castrated male mice replaced with E/P. E/P replacement and testosterone withdrawal were associated with ERα expression, the loss of cyclin E LMW isoforms, intact cdk2 expression, functional G1/S checkpoint control and induction of p53-mediated apoptotic cell death in preneoplastic hepatocytes. Further, oestrogen (E2) stimulation had varying effects on cell cycle regulation and viability in primary hepatocytes and HCC cells. As there was little to no significant correlation between ERα and its downstream target, c-myc transcript levels, we propose that E2/ERα signalling may be operative via other pathways to subsequently activate p53. These findings open up tantalising avenues to further explore the inter-regulatory signalling pathways between E2/ERα, cell cycle regulators and the tumour suppressor, p53.