21-23 In contrast to cytoplasmic viral sensor (RIG-I, MDA5, and LGP2) and modulator (ISG15 and USP18) expression, the adaptor molecule (IPS-1) expression was significantly lower in IL28B minor patients. Moreover, western blotting further confirmed IPS-1 protein downregulation in IL28B minor patients by revealing decreased protein levels. Because IPS-1 is one of the main target molecules of HCV evasion,9, 18 transcriptional and translational IPS-1 expression are probably suppressed by HCV with resistant phenotype, which may be more adaptive

in IL28B minor patients than in IL28B major patients. When we analyzed the proportion of full-length or cleaved IPS-1 to the total IPS-1 protein in a subgroup of IL28B minor patients, cleaved IPS-1 product was less dominant in SVR than in NVR, whereas uncleaved full-length IPS-1 protein was more dominant in SVR than in NVR. Therefore, the ability of HCV to evade host innate immunity by GSK3235025 cleaving IPS-1 protein and/or host capability of protection from IPS-1 cleavage is probably responsible for the variable treatment responses in IL28B minor patients. Our results indicated a close association between IL28B minor patients with higher Ruxolitinib γ-GTP level and higher frequency of HCV core double mutants, which are known factors for NVR. In contrast, no significant association

was observed between IL28B genotype and age, gender, or liver fibrosis, which are also known to be unfavorable factors for virological response to PEG-IFNα/RBV. Therefore, certain factors other either than the IL28B genotype may independently influence virological

response. To elucidate whether gene expression involving innate immunity independently associates with a virological response from the IL28B genotype, we performed further analysis in a subgroup and conducted a multivariate regression and ROC analyses. Our multivariate and ROC analyses demonstrate that higher expressions of RIG-I and ISG15 as well as a higher ratio of RIG-I/IPS-1 are independently associated with NVR, and quantification of these values is more useful in predicting final virological response to PEG-IFNα/RBV than determination of IL28B genotype in each individual patients. However, the SVR rates in our patients were similar among IL28B genotypes, which suggests more SVR patients with the IL28B minor allele were included in the present study than those in the general CH-C population. Hence, our data did not necessarily exclude the possibility of the IL28B genotype in predicting NVR, although our multivariate analysis could not identify the IL28B minor allele as an independent factor for NVR. Interestingly, an association between IL28B genotype and expressions of RIG-I and ISG15 as well as RIG-I/IPS-1 expression ratio is still observed even in patients with the same subgroup of virological response (Fig. 3).