MiR-92a Family: A Novel Diagnostic Biomarker and Potential Therapeutic Target in Human Cancers

Abstract

Purpose: This study tried to explore whether members of miR-92a family contribute to early diagnosis and prognosis for human cancers and how they work. Methods: Integrated meta-analysis retrieved from public repositories was employed to assess the clinical roles of the miR-92a family for cancer diagnosis and prognosis. Expression level of miR-92a was detected by the TCGA database and was confirmed by non-small-cell lung cancer (NSCLC) tissues. Targets of miR-92a were predicted using starbase, and validated by dual luciferase assay. Correlation between miR-92a and the target gene was assessed by linkedOmics while expression of the target gene and its role in cancer prognosis were analyzed with UALCAN and Gepia. Results: We recognized the miR-92a family could serve as a potential diagnostic biomarker with a pooled sensitivity of 0.85 [0.81-0.88] and specificity of 0.86 [0.83-0.90]. The overall hazard ratio (HR) was 2.26 [95% CI: 1.70-3.00] for high expression groups compared to low expression groups. Expression of miR-92a was identified to be upregulated in NSCLC, especially in lung squamous cell carcinoma (LUSC). Results from starbase and dual luciferase assay indicated the regulator of G-protein signaling 3 (RGS3) was a direct target of miR-92a. Statistical negative correlation was found for the expression of miR-92a and RGS3. In addition, expression of RGS3 was downregulated in NSCLC and patients with the high expression had a poor prognosis (HR = 1.3) for LUSC patients. However, results were to the contrary for lung adenocarcinoma (HR = 0.7). Conclusion: This study revealed that miR-92a family could be ideal biomarkers for cancer diagnosis and prognosis, which might function through targeting RGS3.

Keywords: RGS3; diagnosis; miR-92a; neoplasms; prognosis.

Figure 1

Flow diagram of study selection for cancer diagnostic and prognostic meta-analysis of miR-92a family.

Figure 2

Forest plots of sensitivity and specificity for the cancer diagnosis of miR-92a family. Both the sensitivity and specificity of each study were showed by each square with its 95% confidence interval showed by the error bars.

Figure 3

Forest plot of the prognostic meta-analysis of the association between miR-92a family and the risk of cancers. HR, hazard radio; 95% CI, 95% confidence interval.

Figure 4

miR-92a is upregulated in NSCLC. (A,B) Expression of miR-92a in lung adenocarcinoma (A) and lung squamous cell carcinoma (B), values are expressed as Transcript per million, ***p <.001, t-test; (C) Expression of miR-92a in non-small-cell lung cancer tissues, normalized to the mean of control group and expressed as 2–ΔΔCt, *p < 0.05, t-test.

Figure 5

RGS3 is targeted by miR-92a. (A) Putative binding sites in the RGS3 3'UTR for miR-92a by starbase. (B) miR-92a targets the RGS3 3'-UTR. Luciferase activity of the RGS3 reporter with a wild or mutated miR-92a binding site was measured at 48 h post-transcription with miR-92a mimics or negative control, *p < 0.05, ***p < 0.001, Dunnett-t-test. (C–F) Expression of miR-92a was negatively correlated with the level of RGS3 in NSCLC from TCGA, r value and p-value were determined using spearman correlation analysis.

Figure 6

Expression level of RGS3 and its prognostic role in NSCLC. (A,B) Expression of RGS3 was downregulated in lung adenocarcinoma and lung squamous cell carcinoma, values are expressed as Transcript per million, **p < 0.01, ***p < 0.001, t-test; (C,D) The association between expression of RGS3 and the risk of lung adenocarcinoma (C) and lung squamous cell carcinoma (D). HR, hazard radio.