Inactivation of Cyclin E1 acts pro-apoptotic and anti-proliferative in primary hepatoma cells and protects from hepatocarcinogenesis in mice

Background & Aims: E-type cyclins (CcnE1, CcnE2) are regulatory subunits of Cyclin-dependent kinase 2 (Cdk2). There is evidence that CcnE/Cdk2 complexes control the transition of quiescent cells into the cell cycle. Surprisingly, genetic deletion of CcnE1, CcnE2 or Cdk2 does not inhibit cell proliferation during embryonic development or liver regeneration. The aim of the present study was to evaluate if E-type cyclins or Cdk2 may have a specific role for hepatocarcinogenesis. Methods: We used genetically modified mice with constitutive deletion of Cyclin E1 (CcnE1-/-) or Cyclin E2 (CcnE2-/-) and conditional knockout mice with hepatocyte-specific inactivation of Cdk2 (Cdk2Δhepa), respectively. To induce hepatocellular carcinoma (HCC), mice were treated with a single dose of Diethylnitrosamine (DEN) at the age of two weeks and analysed for tumor initiation and progression after 24 and 40 weeks. For in vitro studies, we generated immortalized Cyclin E1-floxed hepatoma cells from DEN-treated CcnE1f/f mice. In these cells, acute CcnE1 deletion was performed and visualized by infection with a cre/EGFP expressing adenovirus. Results: Twenty-four weeks after DEN treatment 40% of wildtyp (WT) and CcnE2-/- mice developed few small tumor nodules, whereas all CcnE1-deficient animals revealed normal liver morphology lacking visible tumor nodules. At the age of 40 weeks, 100% of DEN-treated WT and CcnE2-/- mice developed HCC with an average of 35 – 40 nodules per liver. In sharp contrast, HCC incidence in CcnE1-/- animals was only 50%, and the overall number of tumor nodules was tenfold reduced in comparison to WT and CcnE2-/- mice. Interestingly, Cdk2Δhepa mice and Cdk2ΔhepaCcnE2-/- double mutants were also strongly protected from DEN-mediated hepatocarcinogenesis suggesting that the oncogenic potential of CcnE1 is Cdk2-dependent. A similar anti-tumorigenic effect of CcnE1 ablation was observed in a second HCC model using mice with concomitant hepatic over-expression of the proto-oncogene c-myc indicating that the observed effects are model-independent. Finally, we isolated primary, CcnE1-floxed hepatoma cells from DEN-treated CcnE1f/f mice and inactivated CcnE1 in vitro using EGFP-labeled, cre-expressing adenoviruses. Confocal live cell imaging revealed that acute CcnE1 inhibition in transformed hepatoma cells effectively results in apoptosis within 72 hours. In addition, CcnE1-depleted hepatoma cells showed strong activation of caspase-3, and de-regulation of S-phase and cytokinesis related genes and proteins, respectively. Conclusion: Our data demonstrate that CcnE1-but not its homologue CcnE2, is essential for proliferation and survival of immortalized hepatoma cells in vivo and in vitro. This requirement of CcnE1 seems to be dependent on the availability of Cdk2. We hypothesize that hepatoma cells get addicted to high CcnE1 expression levels during the transformation process. Thus, acute inhibition of CcnE1 during HCC progression could be a novel therapeutic option. Corresponding author: Liedtke, Christian E-Mail: cliedtke@ukaachen.de