A panel of three serum microRNAs as a potential diagnostic biomarker for renal cell carcinoma

Abstract

Hitherto there is no praiseworthy noninvasive methods in the early diagnosis of renal cell carcinoma (RCC). MicroRNAs (miRNAs) could be utilized as molecular markers for diverse malignancies. In this study, we aim to discern potential miRNAs as markers for screening RCC. We employed quantitative reverse transcription-polymerase chain reaction (RT-qPCR) to detect expression levels of candidate miRNAs in serum specimens of 108 RCC patients and 112 health volunteers. Diagnostic values of miRNAs were appraised, and panel was constructed by dint of receiver operating characteristic curves, the area under the ROC curve and backward stepwise logistic regression analysis. Moreover, we capitalized on bioinformatics analysis for exploration of miRNAs biological functions. The expression of five miRNAs (miR-30c-5p, miR-142-3p, miR-206, miR-223-3p, miR-200c-5p) were markedly alteration in serum specimens of RCC patients and health subjects. A three-miRNA panel combining miR-30c-5p, miR-142-3p and miR-206 was constructed and could discriminate RCC patients and healthy subjects satisfactorily with 0.872 (0.811-0.919, P < 0.001) AUC, 81.25% sensitivity and 86.90% specificity. ATF3 and MYC seem to be potential targets of the three-miRNA panel. The novel miRNA-based panel may perform as potential noninvasive markers to discriminate RCC patients and healthy subjects in advance.

Keywords: Bioinformatics; Biomarker; MicroRNAs; Renal cell carcinoma; Serum.

Fig. 1

In screening phase, through reviewing literatures and screened in ENCORI database with 517 RCC patients and 71HCs (P < 0.01 and | logFC | > 1), 10 abnormally expressed miRNAs highly associated with RCC were found out. Filter criteria: p-value < 0.01 and | logFC | > 1. *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 2

The recapitulation of this study. RCC = Renal cell carcinoma; HC = Healthy control; ROC = receiver operating characteristic.

Fig. 3

The relative expression levels of ten candidate miRNAs in training phase. Filter criteria: p-value < 0.05. *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 4

Serum expression levels and receiver operating characteristic (ROC) curve of five chosen miRNAs in validation phase with 80 RCC patients and 84 HCs. (C) miR-142-3p and (G) miR-223-3p were significant upregulation in serum of RCC patients. (A) miR-30c-5p, (E) miR-206 and (I) miR-200c-5p were significant downregulation in serum of RCC patients. Respectively, AUCs of (B) miR-30c-5p, (D) miR-142-3p, (F) miR-206, (H) miR-223-3p, (J) miR-200c-5p were 0.769(P < 0.001), 0.716(P < 0.001), 0.685(P < 0.001), 0.685(P < 0.001), 0.639(P = 0.001). *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 5

The ROC curves analysis for the three-miRNA panel (miR-30c-5p, miR-142-3p and miR-206). The area under the ROC curves (AUC) of three-miRNA panel was 0.872 (95% CI:0.811–0.919, P < 0.001).

Fig. 6

Putative target genes of the three-miRNA panel were detected. (A) 259 genes targeted by two or three miRNAs were predicted by miRWalk 3.0. The expression analysis of 14 target genes was carried out through GEPIA. (B,C) ATF3 and MYC were differentially expressed in 523 kidney renal clear cell carcinoma (KIRC) contrasted to 100 normal controls with |log2Fold Change| cutoff > 1.5, P < 0.01, thereof MYC upregulated and ATF3 downregulated. Kaplan-Meier survival analysis of 14 target genes was conducted using ENCORI database. (D–F) Lower expression levels of GPATCH2L, SP1 and STRN were significantly associated with poor prognosis in kidney ccRCC patients (Log-Rank p < 0.05). *P < 0.01.

Fig. 7

GO functional annotation and KEGG pathway enrichment analysis of the target genes of miR-30c-5p, miR-142-3p, miR-206. The top ten (A) Biological process analysis, (B) cellular component analysis, (C) molecular function analysis, (D) KEGG pathway enrichment analysis are shown.