We appreciate the comments of Luedde et al.1 We read the interesting article reported by Luedde et al.2 In the partial hepatectomy (PH) experiments using knockout mice for various cyclin-dependent kinase (Cdk) inhibitors, the authors have shown that a loss of p18INK4C alone did not influence Cdk4 activity, but the lack of p21CIP1/WAF1 led to Cdk 4 activation. On the other hand, we have reported that the kinase activity of Cdk4 was higher in p18INK4C-negative hepatocellular carcinomas (HCCs) than in p18INK4C-positive ones.3 Based on the results, we concluded that p18INK4C expression might play an important role in the development of HCC through the upregulation of Cdk4 activity. These two reports appear to be contradictory. However, our experiments are different from the study of Leudde and colleagues. The reasons are as follows. Liver regeneration after PH reflects the proliferation of normal hepatocytes, and the process is completed after one week. Thus, changes induced by liver regeneration after PH are quite short, and are reversible. On the other hand, proliferation of HCC is a chronic and continuous process in transformed cells (cancer cells), and it is irreversible. Therefore, although the kinetic changes of cell cycle-related molecules in liver regeneration after PH provide important hints for the mechanism of hepatocarcinogenesis, they may not be directly applied to the proliferation of cancer cells. To solve this problem, it may be necessary to study Cdk 4 activity in carcinogen-induced HCC of p18INK4C knockout animals.
It has been shown that the expression of p21CIP1/WAF1, p27 KIP1 and p57KIP2 were frequently downregulated in HCC.4, 5 Therefore, we agree with the suggestion that the effect on Cdk4 activity in HCC may be caused by a simultaneous loss of p18INK4C, p21CIP1/WAF1 and/or p27KIP2. In the future, it will be needed to determine what Cdk inhibitors influence Cdk4 activation in HCC. In addition, we would like to address the following point. Luedde et al. have stated that a loss of p18INK4C expression alone did not influence Cdk4 activation after PH. However, a clear band corresponding to the retinoblastoma protein phosphorylated at Ser-780 was visible 36 and 48 hours after PH in p18INK4C knockout animals (Fig. 5C in the text of Luedde et al.2). This result suggests that a loss of p18INK4C alone may also influence Cdk4 activation in liver regeneration after PH. We previously reported that the overexpression of Cdk4 was detected in HCC of Long Evans cinnamon rats and human.6, 7 In addition, Pascale et al.8 have also demonstrated that the overexpression of cyclin D1/Cdk4 complex occurred in chemically induced HCC of Fischer 344 rats. Conversely, it has been shown that Cdk6 protein was not increased in HCC.4, 5 Therefore, we assume that the amount of Cdk4 (cyclin D/Cdk4) may be higher than that of Cdk 6 (cyclin D1/Cdk6), and p18INK4C may dominantly interact with Cdk4 rather than with Cdk6 in p18INK4C-positive HCCs, suggesting that p18INK4C in HCC may contribute only to the up-regulation of Cdk4. Finally, expression levels of p18INK4C, p21CIP1/WAF1, p27KIP1, p57KIP2 and other Cdk inhibitors should be evaluated to examine their functional collaboration in the development of HCC.