Epigenetic silencing of MEIS2 in prostate cancer recurrence

Abstract

Background: Current diagnostic and prognostic tools for prostate cancer (PC) are suboptimal, resulting in overdiagnosis and overtreatment of clinically insignificant tumors. Thus, to improve the management of PC, novel biomarkers are urgently needed. Results: In this study, we integrated genome-wide methylome (Illumina 450K DNA methylation array (450K)) and RNA sequencing (RNAseq) data performed in a discovery set of 27 PC and 15 adjacent normal (AN) prostate tissue samples to identify candidate driver genes involved in PC development and/or progression. We found significant enrichment for homeobox genes among the most aberrantly methylated and transcriptionally dysregulated genes in PC. Specifically, homeobox gene MEIS2 (Myeloid Ecotropic viral Insertion Site 2) was significantly hypermethylated (p < 0.0001, Mann-Whitney test) and transcriptionally downregulated (p < 0.0001, Mann-Whitney test) in PC compared to non-malignant prostate tissue in our discovery sample set, which was also confirmed in an independent validation set including > 500 PC and AN tissue samples in total (TCGA cohort analyzed by 450K and RNAseq). Furthermore, in three independent radical prostatectomy (RP) cohorts (n > 700 patients in total), low MEIS2 transcriptional expression was significantly associated with poor biochemical recurrence (BCR) free survival (p = 0.0084, 0.0001, and 0.0191, respectively; log-rank test). Next, we analyzed another RP cohort consisting of > 200 PC, AN, and benign prostatic hyperplasia (BPH) samples by quantitative methylation-specific PCR (qMSP) and found that MEIS2 was significantly hypermethylated (p < 0.0001, Mann-Whitney test) in PC compared to non-malignant prostate tissue samples (AN and BPH) with an AUC > 0.84. Moreover, in this cohort, aberrant MEIS2 hypermethylation was significantly associated with post-operative BCR (p = 0.0068, log-rank test), which was subsequently confirmed (p = 0.0067; log-rank test) in the independent TCGA validation cohort (497 RP patients; 450K data). Conclusions: To the best of our knowledge, this is the first study to investigate, demonstrate, and independently validate a prognostic biomarker potential for MEIS2 at the transcriptional expression level and at the DNA methylation level in PC.

MEIS-mediated suppression of human prostate cancer growth and metastasis through HOXB13-dependent regulation of proteoglycans

“…We originally demonstrated that increased mRNA expression of MEIS1 and MEIS2 in PrCa is correlated with significantly longer overall survival in a large cohort of watchful waiting patients with mid-range Gleason scores ( Chen et al, 2012 ). More recently, we and others demonstrated that patients harboring MEIS-positive tumors have a significantly favorable outcome; there is a step-wise decrease in both MEIS1 and MEIS2 expression as tumors progress to metastatic ( Bhanvadia et al, 2018 ; Jeong et al, 2017 ; Nørgaard et al, 2019 ). These correlative findings provide support to a tumor-suppressive role for MEIS1 and MEIS2 in PrCa.…”

Advances in Prognostic Methylation Biomarkers for Prostate Cancer

“…We refer readers to the original studies for the full lists of methylation markers. We focus on those studies that used measures of disease risk to identify potential biomarkers (for example, comparing methylation of patients of different GS, or different survival outcomes), rather than studies that compared methylation differences between benign and tumour tissue to identify disease-specific biomarkers, with assessment of their prognostic value as only a secondary step [ 113 , 114 , 115 , 116 , 117 , 118 , 119 ].…”

Advances in Prognostic Methylation Biomarkers for Prostate Cancer

“…In this section, we summarise genome-wide prognostic methylation biomarker discovery studies in primary PCa tumours ( Table 2). We focus on those studies that used measures of disease risk to identify potential biomarkers (for example, comparing methylation of patients of different GS, or different survival outcomes), rather than studies that compared methylation differences between benign and tumour tissue to identify diseasespecific biomarkers, with assessment of their prognostic value as only a secondary step (for example, [113][114][115][116][117][118][119]). Low GS vs High GS: Cohort 1 -U: • No recurrence -6.5 ± 1.6…”