Integrated analysis of epigenomic and genomic changes by DNA methylation dependent mechanisms provides potential novel biomarkers for prostate cancer

Oncotarget 2014 Genomics 6 Explanations View Original
Original Paper (PDF)

Unable to display PDF. Download it here or view on PMC.

Plain-English Explanations
Pages 1-2
Epigenetic Silencing and the Challenge of Identifying Aggressive Prostate Cancer

Epigenetics refers to changes in gene activity that do not alter the DNA sequence itself. One of the most important epigenetic mechanisms in cancer is DNA methylation -- the addition of a methyl chemical group to specific sites in the genome called CpG islands, which are found in the regulatory regions of many genes. When CpG islands near a gene's promoter region become methylated, the gene is typically silenced even though its DNA sequence remains intact.

In prostate cancer, widespread abnormal methylation of tumor suppressor genes is thought to be an early and frequent event in tumor development. The key clinical challenge is distinguishing indolent prostate cancer -- slow-growing tumors that may never cause harm -- from aggressive disease that will progress and metastasize. This distinction drives treatment decisions but is difficult to make reliably from current biomarkers such as PSA or Gleason grade alone.

Methylation-based biomarkers are particularly attractive because DNA methylation patterns are stable, can be detected in small amounts of tissue or biological fluids, and can be measured with highly sensitive quantitative techniques. A methylation biomarker that reliably identifies aggressive prostate cancer could reduce overtreatment of indolent tumors while ensuring patients with dangerous disease receive timely intervention.

Prior studies had identified a handful of methylated genes in prostate cancer -- most notably GSTP1, which is methylated in over 90% of prostate cancers -- but these individual markers lacked the specificity to distinguish low-grade from high-grade disease. A comprehensive, genome-wide discovery approach was needed to identify methylation changes that specifically associate with tumor aggressiveness.

TL;DR: Abnormal DNA methylation silences genes in prostate cancer, but no biomarker reliably separates aggressive from indolent disease, motivating a genome-wide search for methylation changes linked to high Gleason grade.
Pages 2-3
Demethylating Agent Treatment and Genome-Wide CpG Profiling

The study used two established prostate cancer cell lines -- 22Rv1 (androgen-responsive) and DU-145 (androgen-independent) -- to represent different subtypes of prostate cancer. Both cell lines were treated with 5-Aza-2'-deoxycytidine (DAC), a demethylating agent that removes existing methyl groups from CpG sites across the genome, effectively reactivating silenced genes.

After DAC treatment, genome-wide methylation was profiled using Agilent CpG island microarrays containing 237,220 probes covering 27,800 CpG islands. Methylated DNA was captured using methylated DNA immunoprecipitation (MeDIP) -- an antibody-based method that pulls down methylated DNA fragments -- and the enriched DNA was hybridized to the array. Untreated cells provided the baseline methylation comparison.

Probes showing significantly reduced methylation after DAC treatment were identified using established bioinformatics thresholds. In 22Rv1 cells, 11,212 probes corresponding to 4,922 genes showed reduced methylation. In DU-145 cells, the number was substantially higher: 32,511 probes corresponding to 8,008 genes. This difference suggests that DU-145 cells, which represent more aggressive androgen-independent disease, carry broader epigenetic silencing.

Analysis of where the methylated probes mapped within the genome revealed a consistent pattern: approximately 23% of methylation changes were in promoter regions (where gene transcription starts), while roughly 71% were intragenic (within gene bodies). This distribution was similar across both cell lines and is consistent with emerging evidence that gene body methylation may also regulate transcriptional activity.

TL;DR: Two prostate cancer cell lines were treated with a demethylating drug and profiled with a 237,000-probe CpG array, identifying thousands of methylated genes -- more than twice as many in the aggressive androgen-independent line.
Pages 3-4
Gene Expression Profiling and GSEA Pathway Analysis

To identify which of the methylated genes were also functionally silenced, the researchers performed gene expression microarray analysis on both cell lines before and after DAC treatment. Genes that were both methylated (detected by the CpG array) and transcriptionally silenced (detected by expression array) before treatment -- and that showed increased expression after demethylation -- were classified as genes under active methylation-dependent silencing.

Integrating the two datasets identified 670 genes in 22Rv1 and 2,330 genes in DU-145 that met both criteria: reduced methylation after DAC treatment and increased gene expression. These genes represent candidates for methylation-based silencing in prostate cancer -- genes whose activity is suppressed not by mutation but by reversible epigenetic modification.

Gene Set Enrichment Analysis (GSEA) was applied to these candidate gene lists to identify which biological pathways were preferentially silenced by methylation. The analysis identified significant enrichment in pathways related to DNA replication, the ATR-mediated DNA damage stress response, and cell migration. These pathways are directly relevant to cancer progression: uncontrolled replication drives growth, impaired stress responses allow genomic instability, and enhanced migration enables metastasis.

The pathway analysis findings suggest that methylation-dependent silencing in prostate cancer is not random but preferentially targets a coherent set of biological controls. Silencing DNA damage response genes, in particular, would reduce the cell's ability to detect and repair mutations, potentially accelerating the accumulation of additional cancer-promoting genetic changes over time.

TL;DR: Combining methylation and expression data identified hundreds of genes silenced by methylation, and pathway analysis revealed that DNA replication control, DNA damage response, and cell migration pathways are preferentially targeted.
Pages 4-5
Identifying Methylation Changes That Predict Gleason Grade in Patient Tumors

To determine which of the cell line-derived methylation candidates were relevant in actual patient tumors, the 670 and 2,330 gene lists were cross-referenced against methylation data from prostate tumor tissue samples. Genes showing higher methylation in Gleason score 8 (aggressive) tumors compared to Gleason score 6 (low-grade) tumors were prioritized as potential biomarkers of aggressiveness.

This analysis identified 117 genes from the 22Rv1 list and 835 genes from the DU-145 list that showed significantly higher methylation in high-grade versus low-grade patient tumors. The fact that the androgen-independent DU-145 cell line contributed more than seven times as many validated candidates may reflect its closer biological resemblance to the most clinically aggressive forms of prostate cancer.

Several genes that had been previously identified as prostate cancer methylation markers in the literature were confirmed in this dataset, including HOXD3, TBX15, TBX3, TGFB2, and DSC3. The recovery of these known markers validated the methodology and increased confidence that the novel candidates emerging from the same analysis were genuine disease-related findings rather than technical artifacts.

Two genes with no prior characterization in prostate cancer were selected for detailed follow-up validation: ACTA1, encoding alpha skeletal muscle actin, and B4GALNT1, encoding a glycosyltransferase enzyme involved in glycolipid synthesis. Both genes showed consistent methylation differences across multiple comparison groups within the discovery dataset and were taken forward for independent validation in three separate patient cohorts.

TL;DR: Cross-referencing cell line data with patient tumor methylation identified 117-835 grade-associated genes per cell line, and two novel biomarker candidates -- ACTA1 and B4GALNT1 -- were selected for independent validation.
Pages 5-6
Independent Validation of ACTA1 and B4GALNT1 Across Three Patient Cohorts

Validation was performed using two University Health Network (UHN) prostate cancer patient cohorts and the publicly available TCGA (The Cancer Genome Atlas) prostate cancer dataset containing 248 patients. In the TCGA data, ACTA1 methylation was significantly higher in Gleason score 8 or above tumors compared to Gleason 7 tumors (p = 0.035). B4GALNT1 methylation showed an even stronger association in the same comparison (p = 0.027).

The associations extended across multiple Gleason score groupings: both markers showed progressive increases in methylation from lower to higher Gleason grades, consistent with them being markers of tumor aggressiveness rather than simply cancer presence. This graduated relationship is clinically meaningful because it suggests the markers could help stratify patients within the intermediate-risk group, where treatment decisions are most uncertain.

Quantitative methylation was further confirmed using MethyLight, a highly sensitive real-time PCR-based technique that measures the percentage of methylated reference (PMR) at a specific locus. For ACTA1, mean PMR was 5.27 in high-grade tumors versus 1.44 in low-grade tumors (p = 0.040). For B4GALNT1, mean PMR was 8.23 in high-grade versus 3.11 in low-grade tumors (p = 0.005), confirming the array-based findings with an orthogonal technology.

The consistency of results across three independent cohorts -- representing different patient populations, different DNA extraction and processing methods, and different measurement platforms -- provides strong evidence that ACTA1 and B4GALNT1 methylation reflects a real biological difference between aggressive and indolent prostate cancer, not a technical or population-specific artifact.

TL;DR: ACTA1 and B4GALNT1 methylation were validated as markers of high Gleason grade prostate cancer across three independent cohorts using both microarray and quantitative MethyLight confirmation, with p-values of 0.005-0.040.
Pages 6-7
Implications for Prostate Cancer Risk Stratification and Epigenetic Biomarker Development

This study demonstrates that an integrated approach -- combining genome-wide methylation discovery in cell lines with validation across independent patient cohorts -- can identify novel methylation biomarkers for prostate cancer aggressiveness. The pipeline moves efficiently from broad discovery to targeted confirmation, with the GSEA pathway analysis providing a biological framework for understanding why specific genes are methylated in aggressive disease.

The identification of ACTA1 and B4GALNT1 as novel markers is clinically significant because these genes had not been previously linked to prostate cancer methylation. ACTA1 encodes a cytoskeletal protein whose silencing may affect cellular mechanics and invasion, while B4GALNT1 is involved in glycolipid synthesis pathways relevant to cell signaling. Their discovery expands the known repertoire of prostate cancer epigenetic alterations and provides new targets for mechanistic investigation.

A key practical advantage of methylation-based biomarkers is that DNA methylation is highly stable and can be measured from small biopsy samples, serum, urine, or other minimally invasive sources. This opens the possibility of developing blood- or urine-based tests for high-grade prostate cancer, which would reduce the need for repeat biopsies and provide a more objective risk assessment than current PSA-based monitoring.

The authors emphasize that validation in larger prospective cohorts is required before clinical implementation. However, the methodology -- using demethylating agents to unmask silenced genes, then cross-validating against tumor grade in independent datasets -- offers a generalizable template for epigenomic biomarker discovery that could be applied across other cancer types where distinguishing aggressive from indolent disease represents a major unmet need.

TL;DR: The integrated epigenomic discovery pipeline identified clinically validated methylation biomarkers for aggressive prostate cancer, with ACTA1 and B4GALNT1 showing promise as objective, stable markers that could eventually be measured from minimally invasive samples.
Citation: Open Access, . Available at: PMC4202166.