Cyclin E1 expression level determines the anti-proliferative response of the pharmacological Cdk2 inhibitor Roscovitine in hepatoma cells and in the liver

Background & Aims: The cyclin-dependent kinase inhibitor Roscovitine (Rosc) has been regarded as a promising anti-cancer agent. However, the anti-proliferative effects of Rosc in hepatoma cells and in the liver are currently unknown and were investigated in the present study to elucidate if Rosc could be a therapeutic option for the treatment of HCC.

Methods: The anti-proliferative and pro-apoptotic properties of Rosc were analyzed in murine hepatoma cells and primary hepatocytes in vitro and in mice following partial hepatectomy (PH) in vivo. Potential off-target effects of Rosc were investigated in mice with ubiquitous deletion of Cyclin E1 (CcnE1-/-), Cyclin E2 (CcnE2-/-) or hepatocyte-specific inactivation of Cdk2 (Cdk2Δhepa), respectively.

Results: In vitro, cell cycle activity and cell viability were dramatically reduced by Rosc in Hepa1 – 6 hepatoma cells and in primary hepatocytes. In vivo, we subjected wildtype (WT) mice to PH and applied Rosc 28 and 40 hours (h) after surgery. Rosc treatment significantly diminished DNA-synthesis and completely blocked both Cdk2 and Cdk1 kinase activities in liver, 48h after PH. However, this inhibition was transient; delayed DNA-synthesis, full restoration of Cdk1 kinase activity and slight activation of Cdk2 occurred in Rosc-treated mice 72h post-surgery. We previously demonstrated that genetic deletion of Cdk2 had no effect on liver regeneration. Intriguingly, Rosc application after PH inhibited DNA-synthesis in WT and Cdk2Δhepa livers to the same extend. Thus, the inhibitory effect of Rosc in liver is not mediated via Cdk2 but most likely through inhibition on both Cdk2 and Cdk1. To further evaluate the role of E-Cyclins for the inhibitory effect of Rosc, we applied this drug to CcnE2-/- and CcnE1-/- mice after PH. Ablation of Cyclin E1 accentuated the inhibitory effects of Rosc. In contrast, Rosc treatment induced only a slight reduction of DNA-synthesis in CcnE2-/- mice after PH. Since genetic inactivation of Cyclin E2 results in over-expression of CcnE1, we hypothesized that cell cycle inhibition by Rosc can be in part rescued by high expression level of Cyclin E1. We therefore investigated the anti-proliferative effects of Rosc in hepatoma cells with different Cyclin E1 expression levels due to adenoviral over expression or conditional gene ablation. Interestingly, we found a strong synergistic cell cycle inhibition by Cyclin E1 ablation and Rosc treatment in vitro. In contrast, adenovirus-mediated Cyclin E1 over-expression partially rescued the cell cycle arrest by Rosc.

Conclusion: Rosc can inhibit proliferation of healthy regenerating hepatocytes and immortalized hepatoma cells most likely through blocking of Cdk1 and Cdk2, as initially suggested. The inhibitory effect of Rosc inversely correlates with endogenous levels of Cyclin E1 expression. This suggests that only HCC patients with moderate hepatic Cyclin E1 expression may benefit from a potential Rosc therapy.

Corresponding author: Nevzorova, Yulia A.

E-Mail: [e-mail]