Relevance of chromosomal region 2q37.3 to molecular determinants of airway function

Abstract

Background: Recent Genome-Wide Association Study (GWAS) meta-analyses have identified a number of significant association signals for pulmonary function, one of which is SNP rs12477314(C>T; MAF=0.14), which maps in an intergenic region on 2q37.3 flanked by two oppositely transcribed genes - HDAC4 and TWIST2, and a lincRNA (FLJ43879). The aim of this study is to investigate the genetic and functional relevance of the association of this common variant with pulmonary function (FEV1/FVC) and its potential contribution to pulmonary disease, in particular COPD. Methods: The methodology applied in this project consisted of a stepwise approach aimed at identifying the underlying mechanisms of the reported association of locus tagged by rs12477314 with lung function. Transcription isoforms of neighbouring protein coding genes HDAC4 and TWIST2 expressed from a number of cell types relevant to COPD and lung disease were investigated using 3’Rapid amplification of cDNA ends (RACE). The possibility that the association of rs12477314 with lung function is driven by neighbouring genes was also explored by evaluating the effect of inflammatory mediators typically expressed in lung disease (IL6, IL8, IL1β, TNFα) on HDAC4 and TWIST2, in an alveolar epithelial cell line using quantitative polymerase chain reaction (qPCR). To gain further insights into the mechanisms underlying the GWAS signal tagged by sentinel SNP rs12477314, linkage disequilibrium, expression and methylation quantitative trait loci, histone methylation signatures, association of proxy SNPs with GWAS signals, and topologically associating domains (TADs) were explored using publicly available sources. Bioinformatic analysis and resultant data suggestive of epigenetic regulation at locus tagged by rs12477314 was followed up by investigating the functionality of putative enhancer/s and their potential interaction with CpG islands in the intergenic region encompassing the linkage disequilibrium (LD) block (r2>0.5) of rs12477314. This investigation involved targeted genomic deletion/s using the CRISPR/Cas9 nuclease technique in an airway epithelial cell line, and RNA-Seq of resultant CRISPR deletion clones as compared to wild type cells with the aim of exploring the functionality of the genomic regions in proximity of rs12477314, through generation of differentially expressed genes’ (DEG) datasets and resultant DEG data analysis. Results: Investigations into transcription isoforms of HDAC4 and TWIST2, and the latter genes’ response to inflammatory mediators did not provide any conclusive evidence of a direct role of these genes supporting the association of genomic region tagged by rs12477314 with lung function. Bio-informatic analysis, however, revealed that the intergenic region in proximity of rs12477314 is enriched for H3K4me1 histone methylation marker suggestive of active enhancer regions. The putative enhancer regions also overlapped with the presence of methylation quantitative trait signals, mainly cg16640358 and cg19405162, of SNPs in LD (r2>0.5) with the sentinel SNP. Bio-informatic analysis also revealed genome wide associations with cardiovascular disease related traits within the LD block of sentinel SNP. Two independent targeted deletions of the putative enhancer region/s using CRISPR/Cas9, and DEG analysis of the deletion clones as compared to wild type cells, resulted in significant differential gene expression (672 downregulated, 331 upregulated genes for a 5Kb deletion, and 847 downregulated, 604 upregulated genes for the other 15kb deletion), meeting the statistical thresholds of q (pFDR) <0.05 and log2 fold change of <-1, >1. The two genomic deletions showed a notable similarity in terms of both identity of DEGs and also in direction of fold change; 338 common DEGs, 89% of which showed the same direction of fold change. Our results suggest that the deleted regions in proximity of sentinel SNP form part of a wider enhancer system, and probably are co-acting enhancers regulating overlapping gene expression patterns. The genes immediately proximal to the deleted regions, i.e. HDAC4, FLJ43879 and TWIST2 were not differentially expressed, with transcription factor PER2, a component of the molecular circadian clock, located on 2q37.3, being the closest DEG to the putative enhancers represented by the deleted regions. We suggest PER2 as a candidate enhancer target gene which could explain at least in part, the extensive dysregulation observed in our DEG datasets. The DEG datasets (q<0.05 and log2 fold change of <-1, >1) from the two genomic deletions enrich for similar canonical pathways including ‘Hepatic fibrosis/Hepatic stellate cell activation’ ‘Atherosclerosis signaling’, ‘Axonal guidance signaling’, Th1 and Th2 activation pathway’ and ‘Th1 pathway’. ‘Airway pathology in Chronic Obstructive Pulmonary Disease’ also resulted as a significant canonical pathway in our DEG analysis, with an evident pro-inflammatory phenotype emerging from our analysis of the most significant canonical pathways. The direction of fold change of a significant number of DEGs, including CXCL2, CXCL8, IL6, FGG, FGF2, CXCL1, CXCL3, IL1α, IL1β, IL10Rα, IL27Rα and COX2/PGE2 in a deactivated PPAR signaling pathway all correlate with a pro-inflammatory phenotype. The latter is also supported by a number of pro-inflammatory cytokines and growth factors (IL1α, IL1β, IL6, TGFα, TGFβ1) identified as upstream regulators. Regulator effects analysis of the two DEG datasets resulted in similar downstream biological processes with the highest consistency score networks including ‘concentration of prostaglandin E2’, ‘chemotaxis’, ‘accumulation of neutrophils’, ‘recruitment of cells’, ‘activation of granulocytes’ and ‘metabolism of eicosanoid’, all of which are synonymous with a pro-inflammatory phenotype. Our results support a model where it is hypothesized that enhancer dysregulation in the genomic region tagged by rs12477314, hypothetically brought about by ageing induced differential methylation result in a pro-inflammatory transcriptome underlying biological pathways typical of ageing associated diseases. Conclusions: Our results suggest that the genomic region tagged by sentinel SNP rs12477314 constitutes a regulatory genomic region responsible for a shared genetic aetiology for biological pathways conducive to a systemic inflammatory phenotype resulting in distinct pathologies such as COPD and atherosclerosis.