Chromosome-wide mapping of DNA methylation patterns in normal and malignant prostate cells reveals pervasive methylation of gene-associated and conserved intergenic sequences

Abstract

Background: DNA methylation has been linked to genome regulation and dysregulation in health and disease respectively, and methods for characterizing genomic DNA methylation patterns are rapidly emerging. We have developed/refined methods for enrichment of methylated genomic fragments using the methyl-binding domain of the human MBD2 protein (MBD2-MBD) followed by analysis with high-density tiling microarrays. This MBD-chip approach was used to characterize DNA methylation patterns across all non-repetitive sequences of human chromosomes 21 and 22 at high-resolution in normal and malignant prostate cells.

Results: Examining this data using computational methods that were designed specifically for DNA methylation tiling array data revealed widespread methylation of both gene promoter and non-promoter regions in cancer and normal cells. In addition to identifying several novel cancer hypermethylated 5' gene upstream regions that mediated epigenetic gene silencing, we also found several hypermethylated 3' gene downstream, intragenic and intergenic regions. The hypermethylated intragenic regions were highly enriched for overlap with intron-exon boundaries, suggesting a possible role in regulation of alternative transcriptional start sites, exon usage and/or splicing. The hypermethylated intergenic regions showed significant enrichment for conservation across vertebrate species. A sampling of these newly identified promoter (ADAMTS1 and SCARF2 genes) and non-promoter (downstream or within DSCR9, C21orf57 and HLCS genes) hypermethylated regions were effective in distinguishing malignant from normal prostate tissues and/or cell lines.

Conclusions: Comparison of chromosome-wide DNA methylation patterns in normal and malignant prostate cells revealed significant methylation of gene-proximal and conserved intergenic sequences. Such analyses can be easily extended for genome-wide methylation analysis in health and disease.

Figure 1

Overview and pre-microarray performance of MBD-chip. A, Overview of MBD-chip. Genomic DNA is: i) fragmented (in this case with restriction enzymes), ii) enriched for methylated DNA using MBD2-MBD-magnetic beads, and iii) amplified, fragmented, labeled and hybridized to tiling microarrays. Comparison with a total input fraction allows identification of methylated regions. B, Degree of MBD2-MBD enrichment is non-linearly proportional to the number of methylated CpGs. WBC DNA was methylated at 0, 6, 10 or 37 CpG sites within an R.AluI restriction fragment within the GSTP1 promoter by treatment with M.HpaII +/- M.HhaI or with M.SssI or no enzymes. The degree of MBD2-MBD enrichment compared to mock (no MBD control), as measured by qPCR, was related to the number of methylated CpGs. ND, not detectable. C, The MBD-chip process enriches DNA with high density of methylated-CpGs. DNA from LNCaP and PrEC cells was completely methylated with M.SssI or left untreated. MBD2-MBD enrichment at regions in HBB and GSTP1 promoters, as examined by qPCR, are shown. The CpG density (Low, indicates < 5 CpGs per kbp and high indicates > 20 CpGs per kbp) and known degree of methylation (as determined by the Infinium 27K DNA methylation platform for HBB (unpublished data; Yegnasubramanian S and Haffner MC, 2011) and by bisulfite sequencing for GSTP1 [9]) are indicated. Schematics of each region are annotated with position of CpGs (vertical hashes), transcriptional start sites (yellow arrow), and amplification primers (red arrows). The start and end positions are with respect to transcriptional start sites.

Figure 2

Methylation of 5' gene upstream regions. A, Methylated regions in LNCaP (left) and PrEC (middle) cells, and hypermethylated (right) regions in LNCaP vs. PrEC cells are significantly enriched within 2 kbp upstream of transcriptional start sites. The expected probability distribution for (hyper)methylated regions to overlap with 5' gene upstream regions is shown (gray bars and blue line). The red line indicates the observed fraction of base pairs overlapping 5' gene upstream regions in our actual dataset. B, DNA methylation signals (smoothed adjusted log₂(M/T)) surrounding a representative 5' gene upstream region hypermethylated in LNCaP compared to PrEC. Annotations include chromosome coordinates (top), CpG density (number of CpGs in sliding 250 bp windows), Refseq genes, and CpG islands. The box indicates a region that was verified by bisulfite sequencing. C, Bisulfite verification of a hypermethylated region (boxed region from panel (B)) upstream of ADAMTS1. Circles represent position of CpGs. In the top line for each cell line the color of each circle represents the fraction of sequenced alleles that were methylated at that CpG according to the color scale (bottom). Each subsequent line represents the methylation pattern for each sequenced clone; black and white circles indicate methylated and unmethylated CpGs respectively. D, AZAdC induces re-expression of ADAMTS1 in LNCaP cells. Expression of ADAMTS1 with respect to that of GAPDH was measured by real time RT-PCR in LNCaP cells treated with vehicle (DMSO) or 1 μM AZAdC for 3 or 7 days.

Figure 3

Methylation of 3' downstream regions. A, Methylated regions in LNCaP (left) and hypermethylated regions in LNCaP (right) are enriched for sequences within 2 kbp downstream of gene transcriptional termination sites. Conventions are the same as for Fig. 2A. B, DNA methylation signals at a representative 3' downstream gene hypermethylated region in LNCaP cells compared to the PrEC cells. Conventions are the same as for Figure. 2B. C, Bisulfite verification of a hypermethyalted region (boxed region from panel (B)) downstream of the DSCR9 and DSCR3 genes. Conventions are the same as for Figure. 2C.

Figure 4

Methylation of introns, exons and intron-exon junctions. A, LNCaP (left) and PrEC (right) cells show significant enrichment for DNA methylation at intron sequences. Conventions are the same as for Figure. 2A. B, The average smoothed adjusted log₂(M/T) across all short (842 - 2,715 bp) and long (2715 - 11,673 bp) introns was plotted with respect to the relative intron position (as a percentage of intron length), showing generally increased average signal towards the ends of introns in LNCaP (left) and PrEC (right) cells. C - D, Identified methylated regions in LNCaP (left) and PrEC (middle) cells, and hypermethylated (right) regions in LNCaP vs. PrEC cells, were highly enriched for overlap with exons and intron-exon junctions. Conventions are the same as for Figure 2A.

Figure 5

Methylation of distal intergenic regions. A, Methylated and hypermethylated distal intergenic regions were highly enriched for overlap with sequences with a high degree of conservation in mammalian and vertebrate species, as indicated by phastCons scores > 0.8. B, Conserved transcription factor binding sites are highly enriched in the methylated and hypermethylated regions. A - B, Conventions are the same as for Figure 2A. C, DNA methylation signals surrounding a hypermethylated region that was not near any known genes in the hg15 genome build. In addition to the annotations described in Figure 2B, the phastCons scores, representing the degree of conservation across 28 vertebrate and mammalian species, is shown. D, Bisulfite sequencing verification of the boxed region from (C), following the same conventions at Figure 2C. Due to space limitations, only the summary schematics are shown. E, Normalized expression of psiTPTE22, a newly annotated pseudogene in hg18 that arises just downstream of the hypermethylated region shown in Figure 5C-D, in PrEC, DMSO (Ctl) and 1 μM AZAdC treated LNCaP cells.

Figure 6

Hypermethylated regions in LNCaP compared to PrEC cells can serve as biomarkers for prostate cancer. A, DNA methylation at representative regions identified as hypermethylated in LNCaP cells compared to PrEC cells was measured in prostate cell lines using the COMPARE-MS assay as described previously [9]. The extent of methylation at each region is color scaled from white to red as shown, with white representing absence of detectable methylation, and red representing nearly complete methylation of all input copies. M.SssI-treated, completely-methylated, WBC DNA served as a positive control. All of the prostate cancer cell lines showed a high degree of methylation at multiple regions, while the PrEC normal prostate cells did not show any detectable methylation at these regions as expected. B, COMPARE-MS analysis of DNA methylation at three of the five regions from (A) showed significant methylation of at least one of the three regions in every tumor sample with very low or undetectable methylation in the matched normal tissue samples.