Fig 2: Liver specific PPI network of KPNA2 and the correlation of KPNA2 with immune response.(A) The PPI network of KPNA2 in liver. The blue dots genes participated in the pathways in (B). (B) The KEGG pathways that the genes in KPNA2 PPI network were enriched. (C–G) The significant positive correlations of KPNA2 expression with infiltration of monocytic lineage cells, T cells, B lineage cells, CD8+ T cells, and myeloid dendritic cells, respectively. (H–J) No significant correlation between KPNA2 expression and infiltration of cytotoxic lymphocytes, neutrophils, and NK cells. (K) The significant negative correlation of KPNA2 with infiltration of endothelial cells. (L) No significant correlation between KPNA2 expression and infiltration of fibroblasts. The red node in (A) represented KPNA2 protein, the black and blue nodes were the proteins which had liver-specific PPIs with KPNA2. The blue nodes also indicated the proteins significantly enriched in the KEGG pathways in (B). PPI, protein-protein interaction. Spearman correlation analysis was used and p < 0.05 was considered significant.

Fig 3: Construction and evaluation of HCC DFS risk model with KPNA2 data and clinical features.(A) Tumor stage, KPNA2 CNV, and cg14898140 methylation were independent prognostic factors for HCC DFS and they were selected for the risk model construction. (B) According to the DFS risk model, the HCC patients with high risk scores presented shorter disease-free time than the ones with low risk scores. (C) The DFS risk model could discriminate the 1-year, 3-year, and 5-year DFS status of the HCC patients effectively. Multivariable Cox regression analysis was used for the risk model construction. Kaplan-Meier survival analysis with log rank test was used for the survival comparison between high and low risk groups. ROC analysis was used to evaluated the effectiveness of the DFS risk model in DFS status prediction. DFS, disease-free survival; CNV, copy number variation; ROC, Receiver operating characteristic. For all the analyses, p < 0.05 was considered significant.

Fig 4: Expression differences of KPNA2 various transcripts between HCC and the paired normal liver controls.(A–B) ENST00000330459 and ENST00000537025 were higher expressed in HCC than in the controls. (C–E) No significant expression differences of ENST00000579754, ENST00000584026, and ENST00000583269 between HCC and the controls. (F–G) ENST00000583392 and ENST00000582898 were higher expressed in HCC than the controls. Paired samples Wilcoxon test was used for expression comparisons and p < 0.05 was considered significant.

Fig 5: Correlations of KPNA2 CNV and KPNA2 methylation with KPNA2 expression in HCC.(A) Significant positive correlation between KPNA2 CNV and KPNA2 expression. (B–E) Significant negative correlations between KPNA2 expression and methylation level of cg23206777, cg22429852, cg21018429, and cg21820889, respectively. (F) Higher methylation of cg23206777 on KPNA2 gene in HCC tumors than in the normal tissues. (G–H) No significant difference of methylation status of cg21018429 and cg21820889 on KPNA2 gene between HCC tumors and normal livers, respectively. (I) Lower methylation of cg22429852 on KPNA2 gene in HCC tumors than in the normal tissues. For (A–E), x-axis represented the relative expression of KPNA2 [log2(TMM+0.001)] in HCC samples while the y-axis indicated CNV or methylation level (beta value) of the CpG sites. For (F–I), x-axis represented samples of different groups and y-axis represented methylation value of KPNA2 gene. Spearman’s correlation analysis and Wilcoxon test were used for the analyses and p < 0.05 was considered significant. CNV, copy number variation; TMM, trimmed mean of M-values.