Discovery and Validation of Novel microRNA Panel for Non-Invasive Prediction of Prostate Cancer

Abstract

Background: Early diagnosis remains a challenge for prostate cancer (PCa) due to molecular heterogeneity. The purpose of our study was to explore the diagnostic potential of microRNA (miRNA) in both tissue and serum that may aid in the precise and early clinical diagnosis of PCa.

Materials and methods: The miRNA expression pattern analysis was carried out in 250 subjects (discovery and validation cohort). The Discovery Cohort included the control (n = 30) and PCa (n = 35) subjects, while the Validation Cohort included the healthy control (n = 60), benign prostate hyperplasia (BPH) (n = 55), PCa (n = 50), and castration-resistant PCa (CRPC) (n = 20) patients. The expression analysis of tissue (Discovery Cohort) and serum (Validation Cohort) was carried out by quantitative polymerase chain reaction (qPCR). The diagnostic biomarker potential was evaluated using receiver operating characteristics (ROC). Bioinformatic tools were used to explore and analyze miRNA target genes.

Results: MiRNA 4510 and miRNA 183 were significantly (p<0.001) upregulated and miRNA 329 was significantly (p<0.0001) downregulated in both PCa tissue and serum. ROC curve analysis showed excellent non-invasive biomarker potential of miRNA 4510 in both PCa (area under the curve (AUC) 0.984; p<0.001) and CRPC (AUC 0.944; p<0.001). The panel of serum miRNAs (miRNA 183 and miRNA 4510) designed for PCa had significant and greater AUC with both 100% sensitivity and specificity. Computational analysis shows that the maximum number of target genes are transcription factors that regulate oncogenes and tumor suppressors.

Conclusion: Based on ROC curve analysis, miRNAs 4510, 329, and 711 were identified as potential non-invasive diagnostic biomarkers in the early detection of PCa. Our findings imply that a panel of miRNAs 183 and 4510 has high specificity for distinguishing PCa from healthy controls and providing therapeutic targets for better and earlier PCa therapy.

Keywords: castration-resistant prostate cancer; diagnostic biomarker; microrna; non-invasive; panel mirna; prostate cancer.

Figure 1. Flow chart representing the study design.

BPH: benign prostate hyperplasia; CRPC: castration-resistant prostate cancer; miRNA: microRNA; ROC: receiver operating characteristic

Figure 2. The relative expression of (A) all miRNA in prostate tissue as compared with control and the expression of (B) miRNA 183, (C) miRNA 4510, (D) miRNA 711, and (E) miRNA 329 in serum.

The graph is presented in the box plot with significant values marked according to p-value: * p<0.05; ** p<0.01; *** p<0.001 miRNA: microRNA; BPH: benign prostate hyperplasia; CRPC: castration-resistant prostate cancer

Figure 3. The ROC curve analysis of (A) miRNA 183, (B) miRNA 4510, (C) miRNA 711, and (D) miRNA 329 in tissue. The AUC is marked with significant values marked according to the p-value.

* p<0.05; ** p<0.01; *** p<0.001 miRNA: microRNA; AUC: area under the curve

Figure 4. The ROC curve analysis of (A) miRNA 183, (B) miRNA 4510, (C) miRNA 711, and (D) miRNA 329 in serum The AUC is marked with significant values marked according to the p-value.

* p<0.05; ** p<0.01; *** p<0.001 miRNA: microRNA; BPH: benign prostate hyperplasia; PCa: prostate cancer; CRPC: castration-resistant prostate cancer; AUC: area under the curve

Figure 5. The ROC curve analysis of panel miRNA, miRNA 183 and miRNA 4510, in serum PCa group. The AUC is marked with significant values marked according to the p-value.

*** p<0.001 miRNA: microRNA; AUC: area under the curve

Figure 6. The ROC curve analysis of (A) panel miRNA 4510 and miRNA 329, (B) panel miRNA 183, miRNA 711, and miRNA 329 in CRPC.

miRNA: microRNA; CRPC: castration-resistant prostate cancer; ROC: receiver operating characteristic

Figure 7. The Venn Diagram of overlapping target genes of (a) miRNA 183, (b) miRNA 4510, (c) miRNA 711, and (d) miRNA 329

Figure 8. The KEGG and REACTOME pathway analysis of overlapped target genes with significant p values (log values). The target genes are selected on the basis of p value <0.05.

KEGG: Kyoto Encyclopedia of Genes and Genomes Genome

Figure 9. The visual representation of miRNA-Target genes by Cytoscape 3.9.1*

*Cytoscape Consortium, San Diego, California, United States