Construction of miRNA-mRNA network and a nomogram model of prognostic analysis for prostate cancer

Abstract

Background: Dysregulated genetic factors correlate with carcinoma progression. However, the hub miRNAs-mRNAs related to biochemical recurrence in prostate cancer remain unclear. We aim to identify potential miRNA-mRNA regulatory network and hub genes in prostate cancer.

Methods: Datasets of gene expression microarray were downloaded from Gene Expression Omnibus (GEO) database for Robust Rank Aggregation (RRA), targeted gene prediction, gene function and signal pathway enrichment analyses, miRNA-mRNA regulatory network construction, core network screening, as well as validation and survival analysis were carried out by using exogenous data.

Results: Prostate cancer-related differentially expressed genes were mostly related to actin filament regulation. Moreover, the cGMP-PKG signaling pathway might play a role in prostate cancer progression. As the core of microRNAs, hsa-miR-106b-5p, hsa-miR-17-5p and hsa-miR-183-5p were matched to hub genes (such as TMEM100, FRMD6, NBL1 and STARD4). The expression levels of hub genes in prostate cancer tissues were significantly lower than normal and closely related to prognosis of patients. The ridge regression model was applied to establish a risk score system. Both risk score and Gleason were used to establish a nomogram. Nomogram predicted the area under the [receiver operating characteristic (ROC)] curve (AUC) of biochemical recurrence at 1-, 3-, and 5-year of 0.713, 0.732 and 0.753, respectively.

Conclusions: Hub genes were closely related to prostate cancer development and progression, which might become biomarkers for diagnosis and prognosis. This novel nomogram established could be applied to clinical prediction.

Keywords: Prostate cancer; bioinformatics; microRNA-mRNA (miRNA-mRNA); nomogram; prognosis.

Figure 1

Heatmaps of DEMs and DEGs in prostate cancer tissues by RRA analysis. (A) Heatmap of DEMs; (B) Heatmap of DEGs; Green color represents down-regulation; Red color represents down-regulation. DEM, differentially expressed miRNA; DEG, differentially expressed gene; RRA, Robust Rank Aggregation.

Figure 2

Construction of miRNA-mRNA regulatory network and core network. (A) miRNA-mRNA regulatory network; (B) Core miRNA-mRNA network.

Figure 3

GO and KEGG enrichment analyses of co-DEGs in prostate cancer. (A) Co-DEGs GO analysis; (B) Co-DEGs KEGG analysis; (C) enrichment relationship between co-DEGs and KEGG pathway. GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genome; co-DEGs, common differentially expressed genes.

Figure 4

Differential expression and survival analyses of the hub targeted genes in prostate cancer obtained from TCGA and GTEx databases. *, P<0.05. TCGA, The Cancer Genome Atlas; GTEx, Genotype-Tissue Expression; PRAD, prostate adenocarcinoma; num, number; T, tumor; N, normal.

Figure 5

Independent predictive significance of a risk score in the TCGA and GSE70769 dataset, respectively. (A) Univariate and multivariate Cox regression analyses of clinic-pathological features and risk score in both sets. (B) Nomogram based on risk score and Gleason to predict BCRFS in the TCGA dataset at 1-, 3-, and 5-year. (C) Nomogram calibration plots predicting 1-, 3-, and 5-year BCRFS to evaluate the accuracy of candidate nomogram in both sets. (D) Time-dependent BCRFS ROC plots for TCGA and GSE70769 cohort. (E) Kaplan-Meier survival curves for biochemical relapse-free based on risk score in both sets. TCGA, The Cancer Genome Atlas; BCRFS, biochemical relapse-free survival; ROC, receiver operating characteristic; pN, pathological N staging; pT, pathological T staging; cM, clinical M staging; cT, clinical T staging; AUC, area under the curve.