Somatic mutation proﬁles of 537 ccRCC patients downloaded from TCGA database were analyzed and visualized using R package “maftools”. A waterfall plot was performed to exhibit the detailed mutation information in each sample. The majority of Variant Classification, Variant Type, SNV Class were missense mutations, SNP, C>T, respectively (Figure 1a i-iii). Counting each sample separately, the median and maximum of mutations in the TCGA-ccRCC cohort were 47 and 1611, respectively (Figure 1a iv, Figure 1b). In addition, we exhibited the number of each variant classiﬁcation in the diﬀerent sample using box plots (Figure.1a v). The top 10 mutated genes in the 336 ccRCC patients were VHL (49%), PBRM1 (42%), TTN (18%), SETD2 (12%), BAP1 (10%), MUC16 (7%), MTOR (7%), KDM5C (6%), HMCN1 (5%), DNAH9 (5%) (Figure 1a vi, Figure 1b, Figure 1d), while the PBRM1 and VHL have the highest correlation (Figure 1c).
Correlation of TMB with prognosis and clinical features
After the clinical data of ccRCC patients were collected from TCGA, we selected 336 patients with complete information. Based on the median TMB value (1.29), we divided 336 patients into high-and low-TMB groups (n=163 VS n=173). As showed in the KM curve, the high-TMB group had signiﬁcantly poor OS outcomes (P=0.007) (Figure 2h). Besides, the correlation of TMB with clinical features showed that only age and grade were signiﬁcantly associated with TMB (p<0.05) (Figure 2a, f), the other clinical characteristics showed no significantly difference (Figure 2b, c, d, e, g).
TMB-Related differentially expressed genes and Functional enrichment analysis
According to the DEGs expression analysis between high-TMB group and low-TMB group, a total of 120 DEGs with |log2FC|> 1 and FDR < 0.05 was identiﬁed. There are 44 up-regulated genes and 76 down-regulated genes in the high-TMB group, compared to the low-TMB group. The heatmap and volcano plot showed the top 40 DEGs ranked in the order of FDR and all DEGs, respectively (Figure3a, b). According to GO and KEGG analysis, DEGs were enriched in the function enrichment of apical part of cell, apical plasma membrane, axon terminus and the pathway of Calcium signaling pathway, Vibrio cholerae infection and Taurine and hypotaurine metabolism (Figure 3c, d).
Comparison of gene expression proﬁles between two tmb groups
We further identiﬁed the TMB-related immune genes through the intersection of 1793 immune related genes and 120 DEGs for subsequent analysis (Figure 4a). Then, we selected 8 TMB-related immune genes including CRP, IGHA2, IGLC3, IL6, LBP, LCN1, PAEP, SLIT2. We extracted the mRNA expression of each TMB-related immune genes to draw the box plot, heatmap and correlation heatmap (Figure 4b, c, d). The results revealed that IGHA2, IGLC3, PAEP were increased, while CRP, LCN1 and SLIT2 decreased, IL6 and LBP resting in ccRCC tissues, compared to the normal tissues (Figure 4b, c). Importantly, PAEP and LCN1 had the highest correlation score (0.96) (Figure 4d).in addition, the ccRCC patents with higher risk scores, composing by these genes, had poor outcomes (Figure 4e). Finally, we selected the PAEP which is a never reported gene in ccRCC for the further research.
PAEP was overexpression and associated with shorter survival in ccRCC patents
Considering PAEP overexpression in ccRCC, we further investigated whether it contributed to ccRCC progression. Thus, we chose the 786-0 cells which had the highest levels of PAEP for further investigation (Figure 5a, b). 786-0 cells were transfected with short hairpin RNAs (shRNAs) targeting to PAEP, which signiﬁcantly silenced the PAEP expression (Figure7a). PAEP was significantly overexpressed in the TCGA-ccRCC cohort, compared to the normal group (Figure 5f), which was further validated in ccRCC cells and ccRCC tissues by Western blot analysis and RT-qPCR (Figure 5c-e). Moreover, analysis of clinical characteristics indicated that PAEP overexpression was significantly correlated with Primary tumor size, TNM stage and Fuhrman grade (Table 1). Importantly, Kaplan-Meier analysis revealed that ccRCC patients with high PAEP expression had poor OS and DFS in the TCGA cohorts and further confirmed in our cohort (without DFS result) (Figure5g-i), indicating that PAEP upregulation was potentially related to the ccRCC progression. To evaluate the diagnostic value of PAEP in ccRCC, we constructed ROC curve and calculated the AUC (Figure5j). The ROC curve of PAEP indicated it was a medium diagnostic tool with an AUC is 0.764. We performed GSEA to reveal potential molecular mechanism of PAEP regulating the progression of ccRCC. The results showed that all the signiﬁcant pathways were enrichment in the PAEP overexpression group related to various tumorigenesis-related characteristics, containing MAPK signaling pathway, Renal Cell Carcinoma, TGF-β signaling pathway, ubiquitin mediated proteolysis and Wnt signaling pathway, etc (Figure 5k). These results provide new clues for exploring the molecular mechanism of ccRCC in the future. In conclusion, PAEP serves as an important oncogene and is associated with a poor clinical outcome of ccRCC.
PAEP promotes proliferation, migration, and invasion of ccRCC cells
Then, we performed the transwell assays and colony formation assays to evaluate invasion, migration and proliferation in response to PAEP knockdown in 786-0 cells. Transwell assays showed that PAEP knockdown remarkably reduced cell invasion and migration in 786-0 cells, compared to Mock groups and sh Con groups (p < 0.05; Figure6a, b), Meanwhile, colony formation assays revealed that PAEP knockdown significantly reduced cell colonies in 786-0 cells compared with Mock groups and sh Con groups (p < 0.05; Figure6c, d). These results showed that PAEP participates in proliferation of 786-0 cells. Taken together, our results suggested that PAEP promoted the progression of ccRCC.
PAEP induces activation of the PI3K/Akt/NF-κ B pathway
Previous studies showed that PAEP contributed to the activation of the PI3K/Akt/NF-κ B signaling pathway. It could be inferred that PAEP played an important role in the activation of the PI3K/Akt/NF-κ B signaling pathway. Therefore, we performed western blotting assays to assess changed genes involved in the PI3K/Akt/NF-κ B signaling pathway in 786-0 cells. The results shown that PAEP knockdown inhibited PI3K/Akt/NF-κ B activation (Figure7a, b).