Morishita and colleagues1 reported an interesting study investigating the expression of the cell cycle inhibitor p18(INK4c). The authors demonstrated that loss of p18 expression occurred especially in lower-differentiated hepatocellular carcinomas (HCCs) and correlated with an unfavorable prognosis. A decreased expression in advanced tumor stages and a very similar prognostic role has been demonstrated in previous studies for p27(KIP1), but not for p21(CIP1/WAF1).2, 3 The association of a lack in p18 and p27 expression in human liver tumors reflects data gained in basic animal studies. In contrast to p21−/− mice, p18−/− and p27−/− mice display increased body size and develop the same tumors at a higher age.4 The induction patterns of p18 and p27 messenger RNA in mouse embryonic development are strikingly similar.5 We recently performed a study using knockout mice for p18, p21, p27, or p18/p21 and p18/p27 in combination.6 In partial hepatectomy experiments, we demonstrated that the effect of a p18 knockout on hepatocyte cell cycle progression is similar to a p27 knockout, whereas it differs from a p21 knockout. The present study by Morishita and colleagues therefore highlights the relevance of observations made in the partial hepatectomy model for liver carcinogenesis in humans.
However, the authors suggest in their article that the loss of p18 in the development of HCC is mediated through an upregulation of cyclin-dependent kinase 4 (CDK4) activity. In our study, we showed that loss of p18 expression alone did not influence CDK4 activation after partial hepatectomy, whereas lack of p21 lead to an earlier activation of CDK4. This phenotype was enhanced in p18/p21 double knockout animals. Moreover, combined p18/p27 knockout mice displayed increased amounts of hepatocytes entering S phase after partial hepatectomy compared with the respective single knockouts.6 Because p21 and p27 expression are frequently downregulated in HCC,7 it remains unclear if the effect on CDK4 activity observed by Morishita and colleagues might be caused by a simultaneous loss of p18 and p21 or p27 expression. Although p18 single mutant mice develop liver tumors in a model of chemical carcinogenesis,8 these tumors were not HCCs but hemangiosarcomas from the hepatic sinusoidal endothelial cells, suggesting that a single lack of p18 might not be sufficient for increased carcinogenesis in hepatocytes. Additionally, it is unclear why the authors observed changes in CDK4 but not CDK6 activity depending on the p18 status, whereas it was shown that p18 in vivo preferentially associates with CDK6.9 Therefore, further studies should be conducted evaluating the expression status of p18, p21, and p27 in parallel to examine their functional collaboration in HCC development. Because mutations or increased promoter methylation appear not to be involved in the downregulation of p18 expression in HCC, an additional analysis of p21 and p27 expression might even reveal a functional dependency between these proteins with regard to their expression during liver carcinogenesis, as suggested previously for p16(INK4a) and p27.10 A thorough understanding of the role of CDK inhibitors in hepatic cell cycle regulation may ultimately provide new insights into molecular hepatocarcinogenesis and may uncover new targets for therapeutic approaches.