PBK as a novel biomarker performed excellent diagnostic and prognostic value in HCC associated with immune infiltration and methylation.

Diagnostic and prognosis of hepatocellular carcinoma (HCC) remain major challenge in clinic. This study aimed to explore a gene signature for diagnosis and prognosis prediction of HCC followed by mechanism investigation. Differentially expressed genes (DEGs) in HCC were screened using TCGA. With specific formula, clinic features of prognosis associated DEGs were calculated to obtained a specific model followed by Kaplan-Meier analysis. Protein-protein interaction (PPI) were predicted using STRING and associations between hub gene and clinic features were analyzed using R software. The hub gene was silenced in HCC cell lines followed by cell behaviors analyses. A prognosis associated 14-gene model was identified in this study which could significantly distinguish samples into high-risk and low-risk groups. PBK, BUB1, NUF2, and CDCA8 were the key nodes involved in the 14 gene-coded PPI with high diagnostic values, and only PBK was an independent risk factor of disease specific survival (DSS) of HCC. Moreover, higher PBK was positively correlated with pathological and histological grades, higher AFP, and infiltrations of Th2, T helper cells and aDC of HCC, but negatively correlated with the killer immune cells. Dysregulated methylation might contribute to the higher expression of PBK and silencing PBK significantly suppressed the proliferation, growth, migration, and invasion of HCC cells. PBK, BUB1, NUF2, and CDCA8 played crucial role in prognosis associated 14-gene model with high diagnostic values. Methylation dysregulation-induced PBK accumulation might promote the development of HCC via modulating immune surveillance.