Low expression of bestrophin-2 is associated with poor prognosis in colon cancer

Abstract

Objectives: The purpose of this research was to confirm the prognostic value of bestrophin-2 (BEST2), one of the hub genes in colon cancer, via bioinformatics analysis and validation in public databases and immunohistochemistry detection.

Methods: The GEO2R online tool and Venn diagram software were utilized to identify differentially expressed genes (DEGs) from expression profiles, including GSE20916, GSE44861 and GSE74602, from the Gene Expression Omnibus (GEO). The overall survival (OS) and disease-free survival (DFS) of colon cancer patients from The Cancer Genome Atlas (TCGA) were analyzed through Kaplan-Meier survival curves. Verification of the significance of BEST2 in colon cancer was based on TCGA, Genotype Tissue Expression (GTEx) and 10 datasets from GEO. BEST2 expression was detected with immunohistochemistry (IHC) in 330 colon tissue samples on microarrays including 165 colon cancerand 165 adjacent normal tissues. For further validation, comprehensive analysis from tissue microarrays and multiple datasets was performed by the summarizing of receiver operating characteristic (SROC) curves and the standard mean differences (SMDs). BEST2 expression in various kinds of colon cancer tissues and cell lines in the context of pancancer was obtained from the Expression Atlas database. The CBioPortal database was queried to identify BEST2 gene alterations and mutation status in colon cancer. Correlated genes (CEGs) with BEST2 and DEGs from public database data were assembled for functional and pathway enrichment analysis.

Results: We identified 85 DEGs from the three datasets and screened out BEST2 as a prognostic predictor via the TCGA database. Colon cancer patients with high expression of BEST2 had better survival than patients with low BEST2 (HR = 0.5, P = 0.006) as shown in Kaplan-Meier survival curves in GEPIA. In all, 1463 colon cancer tissues and 1023 colon normal tissues were gathered via public databases as well as in-house tissue microarrays. The comprehensiveexpression analysis suggested low-expression of BEST2 in colon cancer (SMD = -2.48, 95% CI [-3.15- -1.80]) and the notable efficacy of BEST2 expression in differentiating colon cancer from noncancer samples (AUC = 0.97). Gene alteration status of BEST2 occurred in 5% of colon cancer cases, mostly missense mutations and deep deletions. Genes positively correlated with BEST2 and DEGs primarily aggregated in pathways such as anion absorption, digestive juice secretion, cAMP signaling and so on (P < 0.05).

Conclusion: Ampleevidencesupportsthe role of BEST2 in distinguishing colon cancer from normal tissues in this research. Low expression of BEST2 is correlated with a shorter OS, which implies that BEST2 can be employed as a potential biomarker and therapeutictarget in colon cancer.

Keywords: BEST2; Bioinformatics analysis; Colon cancer; Microarrays; Prognosis; Validation.