Renal cell tumors (RCTs) are the most lethal of the common urological cancers. The widespread use of imaging entailed an increased detection of small renal masses, emphasizing the need for accurate distinction between benign and malignant RCTs, which is critical for adequate therapeutic management. Histone methylation has been implicated in renal tumorigenesis, but its potential clinical value as RCT biomarker remains mostly unexplored. Hence, the main goal of this study was to identify differentially expressed histone methyltransferases (HMTs) and histone demethylases (HDMs) that might prove useful for RCT diagnosis and prognostication, emphasizing the discrimination between oncocytoma (a benign tumor) and renal cell carcinoma (RCC), especially the chromophobe subtype (chRCC). We found that the expression levels of 3 genes--SMYD2, SETD3, and NO66--was significantly altered in a set of RCTs, which was further validated in a large independent cohort. Higher expression levels were found in RCTs compared to normal renal tissues (RNTs) and in chRCCs comparatively to oncocytomas. SMYD2 and SETD3 mRNA levels correlated with protein expression assessed by immunohistochemistry. SMYD2 transcript levels discriminated RCTs from RNT, with 82.1% sensitivity and 100% specificity [area under curve (AUC) = 0.959], and distinguished chRCCs from oncocytomas, with 71.0% sensitivity and 73.3% specificity (AUC = 0.784). Low expression levels of SMYD2, SETD3, and NO66 were significantly associated with shorter disease-specific and disease-free survival, especially in patients with non-organ confined tumors. We conclude that expression of selected HMTs and HDMs might constitute novel biomarkers to assist in RCT diagnosis and assessment of tumor aggressiveness.

Background: The rising incidence of renal cell carcinomas (RCC) constitutes a significant challenge owing to risk of overtreatment. Because aberrant microRNA (miR) promoter methylation contributes to cancer development, we investigated whether altered miR-30a-5p expression associates with DNA promoter methylation and evaluated the usefulness as clear cell RCC (ccRCC) diagnostic and prognostic markers. Methods: Genome-wide methylome and RNA sequencing data from a set of ccRCC and normal tissue samples from The Cancer Genome Atlas (TCGA) database were integrated to identify candidate CpG loci involved in cancer onset. MiR-30a-5p expression and promoter methylation were quantitatively assessed by PCR in a tissue set (Cohort #1) and urine sets (Cohorts #2 and 3) from IPOPorto and Homburg University Hospital. Non-parametric tests were used for comparing continuous variables. MiR-30a-5p promoter methylation (miR-30a-5p me) performance as diagnostic (receiver operator characteristics [ROC]-validity estimates) and prognostic [metastasis-free (MFS) and disease-specific survival (DSS)] biomarker was further validated in urine samples from ccRCC patients by Kaplan Meier curves (with log rank) and both univariable and multivariable analysis. Results: Two significant hypermethylated CpG loci in TCGA ccRCC samples, correlating with miR-30a-5p transcriptional downregulation, were disclosed. MiR-30a-5p me in ccRCC tissues was confirmed in an independent patient's cohort of IPOPorto and associated with shorter time to relapse. In urine samples, miR-30a-5p me levels identified cancer both in testing and validation cohorts, with 83% sensitivity/53% specificity and 63% sensitivity/67% specificity, respectively. Moreover, higher miR-30a-5p me levels independently predicted metastatic dissemination and survival.

Increasing detection of small renal masses by imaging techniques entails the need for accurate discrimination between benign and malignant renal cell tumors (RCTs) as well as among malignant RCTs, owing to differential risk of progression through metastization. Although histone methylation has been implicated in renal tumorigenesis, its potential as biomarker for renal cell carcinoma (RCC) progression remains largely unexplored. Thus, we aimed to characterize the differential expression of histone methyltransferases (HMTs) and histone demethylases (HDMs) in RCTs to assess their potential as metastasis biomarkers. We found that SETDB2 and RIOX2 (encoding for an HMT and an HDM, respectively) expression levels was significantly altered in RCTs; these genes were further selected for validation by quantitative RT-PCR in 160 RCTs. Moreover, SETDB2, RIOX2, and three genes encoding for enzymes involved in histone methylation (NO66, SETD3, and SMYD2), previously reported by our group, were quantified (RT-PCR) in an independent series of 62 clear cell renal cell carcinoma (ccRCC) to assess its potential role in ccRCC metastasis development. Additional validation was performed using TCGA dataset. SETDB2 and RIOX2 transcripts were overexpressed in RCTs compared to renal normal tissues (RNTs) and in oncocytomas vs. RCCs, with ccRCC and papillary renal cell carcinoma (pRCC) displaying the lowest levels. Low SETDB2 expression levels and higher stage independently predicted shorter disease-free survival. In our 62 ccRCC cohort, significantly higher RIOX2, but not SETDB2, expression levels were depicted in cases that developed metastasis during follow-up. These findings were not apparent in TCGA dataset. We concluded that SETDB2 and RIOX2 might be involved in renal tumorigenesis and RCC progression, especially in metastatic spread. Moreover, SETDB2 expression levels might independently discriminate among RCC subgroups with distinct outcome, whereas higher RIOX2 transcript levels might identify ccRCC cases with more propensity to endure metastatic dissemination.

Prostate carcinomas harboring 8q gains are associated with poor clinical outcome, but the target genes of this genomic alteration remain to be unveiled. In this study, we aimed to identify potential 8q target genes associated with clinically aggressive prostate cancer (PCa) using fluorescence in situ hybridization (FISH), genome-wide mRNA expression, and protein expression analyses. Using FISH, we first characterized the relative copy number of 8q (assessed with MYC flanking probes) of a series of 50 radical prostatectomy specimens, with available global gene expression data and typed for E26 transformation specific (ETS) rearrangements, and then compared the gene expression profile of PCa subsets with and without 8q24 gain using Significance Analysis of Microarrays. In the subset of tumors with ERG fusion genes (ERG+), five genes were identified as significantly overexpressed (false discovery rate [FDR], ≤ 5%) in tumors with relative 8q24 gain, namely VN1R1, ZNF417, CDON, IKZF2, and NCOA2. Of these, only NCOA2 is located in 8q (8q13.3), showing a statistically higher mRNA expression in the subgroup with relative 8q gain, both in the ERG+ subgroup and in the whole series (P = 0.000152 and P = 0.008, respectively). Combining all the cases with NCOA2 overexpression, either at the mRNA or at the protein level, we identified a group of tumors with NCOA2 copy-number increase, independently of ETS status and relative 8q24 gain. Furthermore, for the first time, we detected a structural rearrangement involving NCOA2 in PCa. These findings warrant further studies with larger series to evaluate if NCOA2 relative copy-number gain presents prognostic value independently of the well-established poor prognosis associated with MYC relative copy-number gain.