Author(s)
Joanna Kam
Lamont R Jones
Kang Mei Chen
Josena Stephen
Laura Garcia-Rodriguez
Indrani Datta
George Divine
Maria J. Worsham
Affiliation(s)
Abstract:
Identify miRNA contribution to keloid pathogenesis. Prospective Cohort. Sixteen fresh keloid with paired normal fresh tissue samples, from the head and neck area, were analyzed using the Illuminal Methylation EPIC Chip®. The platform analyzes 880k CpGs of which 842,625 were carried forward after quality control and data cleanup. Of the 842,625 CpGs, 6763 fell into genomic region of microRNAs and were retained for statistical analysis. A paired t-test was used among keloid vs normal to calculate raw p-values and then FDR corrected p-values on the 6763 CpGs. There were 2 pairs of duplicate samples from the same patient which were resolved by taking the average of methylated and unmethylated intensities for each pair which resulted in 15 paired samples for analysis. Statistical analyses was performed with conversion of methylated and unmethylated intensities to M-values (formula from Du et.al, Comparison of Beta-value and M-value methods for quantifying methylation levels by microarray analysis. BMC Bioinformatics 2010). Statistically significant results were divided into 3 tiers (tier1 with adaptive FDR <=0.05, tier2 with tier1 threshold along with keloid vs normal BetaRatio >=2 or keloid vs normal BetaRatio <=0.50, tier3 with tier 1 & tier2 threshold along with absolute difference keloid minus normal AvgBeta =0.2). Hyper and hypo status of a CpG site is based on keloid vs normal BetaRatio<1 (i.e hypo) and keloid vs normal BetaRatio>=1 (i.e hyper). There were 1836 CpGs in Tier1 at FDR <=0.05 level, which were in genomic region(s) of 649 microRNA genes. Three of the microRNAs were in Tier2 and none made it to Tier3. Because miRNAs perform their important functions via their target genes, the Tier1 miRNAs were used to extract coding gene (mRNA) targets in Ingenuity’s IPA knowledgebase. This IPA resource draws from online databases such as TargetScan, TarBase, miRecords, and Ingenuity Expert Findings for prediction of binding sites and target genes of miRNAs. The degree of confidence that a target gene is associated with a miRNA is characterized in these databases as either “high” (predicted) or “experimentally observed” (validated). There were 649 Tier1 microRNAs and 3 Tier2 miRNA (miRNA 548W, 548N and127`3E. None of the Tier2 miRNAs had target genes in IPA. Among the 649 Tier1 miRNAs, 5 miRNAs: mir-147a, 203a-3p, 292-3p, 486-5p, and 499-5p have 12 coding genes (mRNA) targets: VEGFA, ABL1, RUNX2, SOCS3, TP63, AR, AFF3, ARID4B, FOXO1, PTEN, TWF1 and SOX6, which were indicated as experimentally validated in IPA. Characterization of biological pathways enriched by these target genes to support the 5 miRNAs as candidate miRNAs in keloid pathogenesis, is underway.