Systematic characterization of cross-source miRNA biomarkers in prostate cancer with computational-experimental integrated analysis.

Purpose: Prostate cancer (PCa) is occult and remains largely incurable once it metastasizes. Our research aims to identify the key miRNAs and construct miRNA-mRNA networks for PCa.

Methods: The microarray dataset GSE112264, consisting of 1,591 male serum samples, and tissue miRNA data from TCGA, including 497 prostate cancer and 52 normal samples, were included in the analysis. Differentially expressed miRNAs (DE-miRNAs) were detected, and miRTarBase was used to predict the common target genes. Then, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed for the target genes. The protein-protein interaction (PPI) network, which revealed the top 10 hub genes, was constructed using the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) and Cytoscape. The expression of the potential hub genes was examined using the UALCAN database. Finally, GSE112264, TCGA datasets, and clinical samples were used to verify the consistency of miRNA expressions in serum and tissue.

Results: A total of 948 target genes of the two overlapped downregulated miRNAs (miR-146a-3p and miR-136-3p) were predicted. Functional enrichment analysis indicated that significant DE-miRNAs were related to PCa-related pathways, such as protein binding, the mammalian target of rapamycin (mTOR) signaling pathway, and porphyrin and chlorophyll metabolisms. Four hub genes were identified from the PPI network, namely, NSF, HIST2H2BE, IGF2R, and CADM1, and verified to be aberrantly expressed in the UALCAN database. Experiment results indicated that only miR-136-3p was markedly reduced in both serum and tissue.

Conclusion: In this study, we established the miRNA-mRNA network, offering potential PCa targets.