Integrative Genomic Analysis Reveals Novel Methylation Biomarkers And Potential Genetic Risk Mediators For Psoriasis In Skin Tissue

Introduction:Psoriasis is a chronic inflammatory immune-mediated skin disease, which is characterized by red, scaly skin patches that are usually found on the scalp, elbows, and knees. It affects approximately 0.1-4% of the people worldwide depending on geographical area and ethnic diversity. The prevalence of Caucasian population is higher that of Asian. In China, the prevalence of psoriasis is approximately 0.123% with the taotal patient number about 10 million. The distribution trend of China presented that the prevalence of urban and north areas is higher that in rural and south arears, the prevalence of man is higher than tha of women. The etiology of psoriasis is incompletely understood, but clear evidence suggests both genetic and environmental factors are involved. It is estimated that 60～90% of disease variance is due to genetic component. However, currently estimation attributes only～13% of the disease heritability to identified genetic variants, indicating additional genetic, epigenetic or environmental factors to disease manifestation.Epigenetic variations of DNA, particularly the CpG DNA methylation (DNAm) and histone modifications, are an important source of regulation in human genome. It appears compelling that CpG methylation could account for some of the "missing" heritability not explained by sequence variation. DNAm is an epigenetic mechanism that regulates gene expression by directly blocking DNA-transcription factor (TF) binding or by recruiting methyl-binding proteins (MBDs). These MBDs, coordinating with other members of the epigenetic machinery such as histone deacetylases (HDAC) and histone methyltransferases, result in chromatin reconfiguration and gene silencing. Up to date, DNAm has been extensively studied in various biological processes and for its roles in cancers and several other diseases.There have been a few reports of altered methylation within promoters of target genes in diseased skin. For example, the promoter regions of the p15 and p21 genes are hypomethylated in psoriasis. However, SHP-1 gene promoter is reported to be demethylated in psoriatic skin, but not in skin from atopic dermatitis patients or health controls. Although these studies have profiled the methylation status of psoriasis, they have either been limited in the number of samples or genes assayed, focused on candidate gene promoters, or lacked expression studies that allow the potential function of DNA methylation alterations to be determined.Object:In order to better characterize DNA methylation associated with psoriasis in genome level, and identify susceptible genes in Chinese Han population by using comprehensive whole genome methylation and expression study of psoriasis.Methods:Whole genome methylation level of 114 psoriatic skins (PP),41 adjacent non-lesional skins (PN) and 62 normal controls (NN) was measured by IlluminaHuman 450K chip. Locus by locus DNA methylation analysis were performed using the non-parametric Wilcoxon rank-sum test, and Benjamini-Hochberg multiple comparisons correction was performed with Illumina Methylation Analyzer (IMA) package in R. Correlation between gene expression and DNA methylation for each gene was measured using the Spearman correlation coefficient. Methylation level validation in 20 PP vs 20 NN was performed by sequenom MassArray system.Results:An unpaired locus-by-locus analysis revealed 286 differentially methylated sites, corresponding to 188 unique mapped genes, in lesional skins compared to non-lesional and normal skins. Further integration with whole-transcriptome sequencing data showed the effects of methylation on expression of 36 potential drives of psoriasis. Genes presented reverse correlation between methylation and expression included AIM2, TGFBR3, PHYHD1, PHOB, FOLR1, ZC3H12A, LYPD1 and ZBP1 genes.1. We found 1638 differentially methylated sites between PP and PN. Eight hundred and fifty four of which were significantly hypermethylated while six hundredand eight four were hypomethylated in PP compared to PN2. CpG methylation in PP versus NN differentiated at 434 sites. There were 286 sites, corresponding to 188 unique annotated genes, overlapped with findings discovered from PP versus PN comparison.3. We found noticeable location-specific distribution trends in RefSeq genes and CGI-assocaited region. Nearly sixty percent of differential methylated sites were located on open sea.42.2% of the DMS were in gene body.4. To view the enrichment of 188 genes presenting significantly methylated, we carried out DAVID functional analysis. The differentially hypermethylated set of genes was significantly enriched in functional categories including protein phosphorylation, and regulation of transcription, while the hypomethylated gene sets were not passed the enrichment threshold (Benjamini-Hochberg [BH] adjusted P< 0.05).5. Our Infinium IlluminaHuman450K array covered 776 of the 1108 sites presented differential methylation in Roberson’s study and 456 (58.7%) of these sites were differentially methylated with Benjamini-Hochberg (BH) corrected P< 0.01 in our study, which was more than expected by chance (Chi-square= 276.5, P< 2.2×10-16). We also compare methylation trend of DMS from both study. Among 456 DMS, only one site, cg0002599, showed opposite methylation trend, suggesting high consistency between two findings.6. To examine the extent to which DNA methylation affect gene expression, we performed transcritpome sequencing of 20 paired PP/PN and 20 NN samples. We were able to look at expression level for 178 out of 188 unique differentially methylated genes. Using a BH-adjusted P-value cutoff of 0.01, we found 44 differntially methylated genes were differentially expressed in PP compared with PN or NN samples. Of these,17 genes (9.6%) were hypomethylated and up-regulated, while 19 gene (10.7%) were hypermethylated and down-regulated, suggesting that abnormal DNA methylation might have functional effect on one-fifth of the relevant genes. The genes showed negative correlation of expression and DNA methylation included AIM2, TGFBR3, PHYHD1, PHOB, FOLR1, ZC3H12A, LYPD1 and ZBP1 genes.7. We constrcuted a statistical model to evaluate the relationship among SNP, methylation and disease status. By performing CIT test, we found three CpG potentially mediate genetic risk for psoriasis.Conclusion:We found 286 differentially methylated sites, corresponding to 188 unique mapped genes, in lesional skins compared to non-lesional and normal skins. Further integration with whole-transcriptome sequencing data showed the effects of methylation on expression of 36 potential drives of psoriasis. Genes presented reverse correlation between methylation and expression included AIM2, TGFBR3, PHYHD1, PHOB, FOLR1, ZC3H12A, LYPD1 and ZBP1 genes. By performing CIT test, we found three CpG potentially mediate genetic risk for psoriasis.