Identification of miRNAs and Their Target Genes Associated with Sunitinib Resistance in Clear Cell Renal Cell Carcinoma Patients

Abstract

Sunitinib has greatly improved the survival of clear cell renal cell carcinoma (ccRCC) patients in recent years. However, 20-30% of treated patients do not respond. To identify miRNAs and genes associated with a response, comparisons were made between biopsies from responder and non-responder ccRCC patients. Using integrated transcriptomic analyses, we identified 37 miRNAs and 60 respective target genes, which were significantly associated with the NF-kappa B, PI3K-Akt and MAPK pathways. We validated expression of the miRNAs (miR-223, miR-155, miR-200b, miR-130b) and target genes (FLT1, PRDM1 and SAV1) in 35 ccRCC patients. High levels of miR-223 and low levels of FLT1, SAV1 and PRDM1 were associated with worse overall survival (OS), and combined miR-223 + SAV1 levels distinguished responders from non-responders (AUC = 0.92). Using immunohistochemical staining of 170 ccRCC patients, VEGFR1 (FLT1) expression was associated with treatment response, histological grade and RECIST (Response Evaluation Criteria in Solid Tumors) score, whereas SAV1 and BLIMP1 (PRDM1) were associated with metachronous metastatic disease. Using in situ hybridisation (ISH) to detect miR-155 we observed higher tumoural cell expression in non-responders, and non-tumoural cell expression with increased histological grade. In summary, our preliminary analysis using integrated miRNA-target gene analyses identified several novel biomarkers in ccRCC patients that surely warrant further investigation.

Keywords: miRNA; pathway analysis; renal cancer; resistance; sunitinib; transcriptome.

Figure 1

Schematic diagram of the workflow used in this study.

Figure 2

Heatmap of unsupervised cluster analyses depicting expression of (A) mature miRNAs, (B) pre-miRNAs, (C) snoRNAs and scaRNAs, (D) lncRNA and (E) coding genes in ccRCC cases. The dendrogram at the side shows the distribution of the RNAs, and at the top the relationship between patient samples (blue responder and red non-responder) is shown.

Figure 3

String visualisation network of miRNA–target gene interactions associated with sunitinib resistance in ccRCC patients.

Figure 4

Gene ontology and pathway mapping of miRNA targeted genes. Terms are functionally grouped based on shared genes (kappa score) and are shown in different colours. The node size represents the degree of significance.

Figure 5

Box and whisker plots of levels of differentially expressed miRNAs measured by qRT-PCR in NR and R ccRCC cases. (A) miR-17-3p; (B) miR-99a-5p; (C) miR-223-3p; (D) miR-155; (E) miR-484; (F) miR-200b-3p; (G) miR-200c-3p; (H) miR-150-5p; (I) miR-130b-3p. Significant differences (p < 0.05) are denoted by asterisks (*).

Figure 6

Box and whisker plots of levels of differentially expressed genes measured by qRT-PCR in NR and R ccRCC cases. (A) CD274; (B) EPAS1; (C) VEGFA; (D) FLT1; (E) ZEB1; (F) LRP6; (G) PTBP2; (H) PRDM1; (I) SAV1. Significant differences (p < 0.05) are denoted by asterisks (*).

Figure 7

Kaplan–Meier survival curves in univariate analysis of expression levels of (A) miR-223-3p, (B) PRDM1, (C) FLT1 and (D) SAV1 as a function of overall survival (OS) in months.

Figure 8

Examples of miR-155 expression detection by ISH in ccRCC cases demonstrating (A) positive expression in tumour cells, (B) positive expression in non-tumour cells and (C) negative expression.

Figure 9

Schematic summary of main findings in this study.