DNA methylation of neural EGFL like1 (NELL1) is associated with advanced disease and the metastatic state of renal cell cancer patients.

Abstract

Recent studies have shown that NELL1 expression is silenced epigenetically in human renal cell cancer (RCC) tissues and in RCC cell lines. However, it remains unknown whether NELL1 promoter methylation observed in clinical specimens might be associated with the clinicopathology or survival of patients with RCC. We analyzed NELL1 DNA methylation in tissues from patients with RCC and in adjacent normal renal tissues. In addition, we evaluated NELL1 methylation in cell lines derived from different urogenital tumors (prostate cancer, urothelial cancer and RCC). We performed regression analyses to determine whether NELL1 methylation is associated with clinicopathological parameters and recurrence‑free survival (RFS). This cross‑sectional study included 98patients with RCC and 63 paired tumor and adjacent normal tissue samples. We analyzed a locus in the intron 1 region of NELL1 with pyrosequencing. We performed insilico analysis of NELL1 methylation in the TCGA Kidney Renal Clear Cell Carcinoma (KIRC) data set (n=284patients), which served as a validation study. Statistical analyses were performed with the two‑sided paired t‑test for paired tumor and adjacent normal samples. We used logistic regression for subgroup comparisons and Cox regression for RFS comparisons. The mean methylation level was 6.8% higher in RCC tissues compared to paired adjacent normal tissues (paired t‑test, P<0.001). Methylation levels in RCC were associated with advanced disease (P=0.002), the presence of distant metastases (P=0.004), and shorter RFS (P=0.035, HR: 4.15). Insilico validation with TCGA KIRC data for adjacent loci also demonstrated that high relative methylation levels were associated with adverse clinicopathology and shortened RFS. Our results suggest that NELL1 methylation contributes to RCC disease progression. This finding could provide a clinical marker to complement recent functional analyses in tumor models.