EIF5A2 acts as a potential marker for prognosis and immunity in human cancers.

Abstract

EIF5A2, a key member of the EIF family, has not been extensively studied regarding its role and mechanism in pan-cancer. TCGA and GTEx analyses were performed to investigate the differential expression of EIF5A2 in tumors and normal tissues. cBioportal was used to investigate the gene alterations of EIF5A2 in tumors. We used Cox regression and Kaplan-Meier analyses to discuss the impact of EIF5A2 expression on prognosis. Quantitative real-time PCR (qRT-PCR) was used to detect EIF5A2 mRNA levels in 12 cases of fresh liver hepatocellular carcinoma (LIHC) and corresponding adjacent non-tumor tissues. Immunohistochemistry (IHC) was employed to evaluate the expression levels of EIF5A2 in 284 cases of LIHC as well as in adjacent non-tumorous tissues. Furthermore, the study explored the association between EIF5A2 expression and various clinicopathological parameters, along with its prognostic implications. Spearman correlation analysis was used to estimate the relationship between EIF5A2 expression and tumor mutation burden (TMB), microsatellite instability (MSI), and immunologic features. "pRRophetic"R package was utilized to obtain the sensitivity of common drugs. The Gene Set Enrichment Analysis (GSEA) was used to study the functional enrichment analysis of EIF5A2-related genes. EIF5A2 was overexpressed in most tumors using TCGA and GTEx databases. Cox regression analysis demonstrated that high EIF5A2 expression was associated with unfavorable overall survival (OS) and disease-specific survival (DSS) in head and neck squamous cell carcinoma (HNSC), brain lower grade glioma (LGG), and LIHC. The frequency of EIF5A2 gene alterations was the highest in lung squamous cell carcinoma (LUSC). EIF5A2 expression was associated with TMB, MSI, and immune cell infiltration in some tumors. We performed IHC and qRT-PCR to evaluate EIF5A2 expression in HCC and normal tissues, and found upregulation of EIF5A2 expression at the mRNA and protein levels in LIHC. There was a correlation between EIF5A2 expression and tumor size, tumor grade, and TNM stage in LIHC. Kaplan-Meier survival suggested that the overexpression EIF5A2 group had unfavorable outcomes in LIHC. EIF5A2 expression was correlated with immune cell infiltration in LIHC. The high EIF5A2 expression group was more sensitive to cisplatin, crizontinib, gemcitabine, and nilotinib in LIHC. High EIF5A2 expression was associated with several pathways, including cell cycle, proteasome, DNA replication, primary immunodeficiency and oocyte meiosis. EIF5A2 may serve as a potential prognostic marker and a latent focus for cancer immunological treatment.