Lookup NU author(s): Dr Sarah Rice,
Dr Tony Sorial,
Dr Colin Shepherd,
Dr David Almarza Gomez,
Professor David Deehan,
Dr Louise Reynard,
Professor John Loughlin
Full text for this publication is not currently held within this repository. Alternative links are provided below where available.
Purpose: The vast majority of OA genetic risk-conferring alleles mediate their effect by modulating the expression of a target gene. These are known as expression quantitative trait loci (eQTLs). DNA methylation at CpG dinucleotidesis an epigenetic process used by the cell to regulate gene expression and can amplify or attenuate the impact of an eQTL. Therefore, identifying CpGs where methylation correlates with a genetic association signal offers insight into the mechanism by which the risk at that signal operates. Such CpGs are known as methylation quantitative trait loci (mQTLs). Furthermore, if mQTL CpGs cluster in or close to a particular gene, their mapping can prioritize that gene for further investigation as a plausible functional candidate for an association signal. As a means of gene prioritization for subsequent functional analysis, we assessed whether mQTLs were operating at 18 recently identified novel OA association signals.Methods: Using the Infinium HumanMethylation450 BeadChip (Illumina), genome-wide methylation profiling of CpGs was carried out on the bisulphiteconverted cartilage DNA of 87 patients who had undergone hip or knee joint arthroplasty. Genotyping was performed using the HumanOmniExpress Bead Chip (Illumina). Genotype at the SNP that had originally identified the association signal was correlated with CpG methylation levels by logistic regression, with P-values corrected for multiple testing. For each locus, we investigated CpGs positioned up to 1Mb up- and downstream of the association SNP. Quantitative PCR (qPCR), RNA-sequencing, and in-silico analyses were performed to investigate gene expression in a range of primary and transformed cell types. Genomic DNA and RNA were extracted from the intact hip or knee cartilage of OA patients following arthroplasty and allelic expression imbalance (AEI) analysis was performed by pyrosequencing. CRISPR/Cas9 genome editing was undertaken on the chondrocyte cell line TC28a2.Results: Four of the 18 novel OA risk signals had mQTLs (corrected P<0.05): rs10471753 (one CpG), rs11780978 (10 CpGs), rs4764133 (one CpG), and rs6516886 (six CpGs). The rs11780978 signal (G>A, risk allele frequency 0.39) was particularly striking, with the majority of the 10 positive CpGs showing a strong correlation between methylation and genotype (corrected P<1x10−10). These 10 CpGs are all located within the gene body of PLEC, nine intronic and one exonic. Six of the intronic CpGs reside within a 1kb interval, predicted to have regulatory activity. In-silico analysis of the rs11780978 signal using data from non-joint tissues highlighted eQTLs operating on PLEC and on three flanking genes: GRINA, PARP10 and SPATC1. CRISPR/Cas9 deletion of the cluster of six CpGs resulted in a significant reduction in the expression of PLEC(P<0.001) but no effect on the flanking genes. qPCR revealed abundant expression of PLEC in cartilage whilst AEI revealed that a PLEC eQTL operates in OA cartilage, with the risk A allele of rs11780978 being significantly associated with decreased PLEC expression (P = 0.02). RNA-seq highlighted a significant increase in the expression of PLEC in the hip cartilage of OA patients versus hip cartilage from age-matched control individuals (P = 0.01).Conclusions: Our study has identified four novel mQTLs correlating with OA susceptibility loci and has prioritized PLEC as a functional candidate at one of these loci. PLEC encodes plectin, a cytoskeletal protein that maintains tissue integrity by regulating signalling from the cytoplasm to the nucleus. Plectin enables cells to respond to external mechanical stimuli, such as those experienced by chondrocytes during joint movement. We hypothesise that in a joint predisposed to OA PLEC expression increases in order to combat an aberrant bio-mechanical environment. However, carriage of the lower expressing A allele of rs11780978 could hinder this response and contribute to an enhanced risk of disease development. Our analysis highlights the power and utility of mapping mQTLs in the context of OA genetic susceptibility.
Author(s): Rice SJ, Sorial AK, Aubourg G, Shepherd C, Tselepi M, Almarza D, Skelton AJ, Deehan D, Reynard LN, Loughlin J
Publication type: Conference Proceedings (inc. Abstract)
Publication status: Published
Conference Name: 2018 OARSI World Congress on Osteoarthritis Promoting Clinical and Basic Research in Osteoarthritis
Year of Conference: 2018
Print publication date: 01/04/2018
Acceptance date: 28/01/2018