Myocardial Infarction-Associated SNP at 6p24 Interferes With MEF2 Binding and Associates With PHACTR1 Expression Levels in Human Coronary Arteries

Abstract

Objective: Coronary artery disease (CAD), including myocardial infarction (MI), is the main cause of death in the world. Genome-wide association studies have identified dozens of single nucleotide polymorphisms (SNPs) associated with CAD/MI. One of the most robust CAD/MI genetic associations is with intronic SNPs in the gene PHACTR1 on chromosome 6p24. How these PHACTR1 SNPs influence CAD/MI risk, and whether PHACTR1 itself is the causal gene at the locus, is currently unknown.

Approach and results: Using genetic fine-mapping and DNA resequencing experiments, we prioritized an intronic SNP (rs9349379) in PHACTR1 as causal variant. We showed that this variant is an expression quantitative trait locus for PHACTR1 expression in human coronary arteries. Experiments in endothelial cell extracts confirmed that alleles at rs9349379 are differentially bound by the transcription factors myocyte enhancer factor-2. We engineered a deletion of this myocyte enhancer factor-2-binding site using CRISPR/Cas9 genome-editing methodology. Heterozygous endothelial cells carrying this deletion express 35% less PHACTR1. Finally, we found no evidence that PHACTR1 expression levels are induced when stimulating human endothelial cells with vascular endothelial growth factor, tumor necrosis factor-α, or shear stress.

Conclusions: Our results establish a link between intronic SNPs in PHACTR1, myocyte enhancer factor-2 binding, and transcriptional functions at the locus, PHACTR1 expression levels in coronary arteries and CAD/MI risk. Because PHACTR1 SNPs are not associated with the traditional risk factors for CAD/MI (eg, blood lipids or pressure, diabetes mellitus), our results suggest that PHACTR1 may influence CAD/MI risk through as yet unknown mechanisms in the vascular endothelium.

Keywords: coronary artery disease; genetic association studies; myocardial infarction.

Figure 1

Association results between 387 genotyped or imputed DNA markers in PHACTR1 and myocardial infarction status in 3,172 French Canadians from the Montreal Heart Institute Biobank. rs9349379 has the strongest association signal (P=8.4×10⁻⁶). We used the LocusZoom tool to plot association results.

Figure 2

PHACTR1 gene expression in human tissues. (A) PHACTR1 expression levels were measured by quantitative PCR (experiment done in quadruplicates). Results were normalized on the housekeeping gene HPRT and calibrated on the expression in the heart (post-infarct). Error bars represent standard deviations. (B) rs9349379 is an expression quantitative trait locus (eQTL) for PHACTR1 in human right coronary arteries (total n=25). The dashed line represents the best-fit regression line for PHACTR1 expression levels.

Figure 3

PHACTR1-rs9349379 is differentially bound by MEF2. Electrophoretic mobility shift assays (EMSA) with human umbilical vein endothelial cell (HUVEC) nuclear extract. Probe C contains the canonical MEF2 binding site and acts as positive control. Probes A and G differ by their respective allele at rs9349379 : the G-allele disrupts the MEF2 consensus binding site. Only Probe A shifts in the presence of nuclear extract (arrow, left panel), and the complex supershifts when an antibody against MEF2 is added (arrowhead, left panel). The binding between Probe A and MEF2 is specific: excess unlabeled (cold) Probe A, but not unlabeled Probe G, competes and disrupts the interaction between labeled Probe A and MEF2 (star, right panel).

Figure 4

A CRISPR/Cas9-induced deletion that encompasses rs9349379 and the MEF2 binding site reduces PHACTR1 expression in endothelial cells. (A) Schematic representation of the PHACTR1 locus. The star (bold, underline) corresponds to rs9349379 in intron 3 of PHACTR1 and the box highlights the DNA sequence targeted by the CRISPR guide RNA. We isolated a heterozygous clone that carries a 34-bp deletion (KO allele) that removes rs9349379 (bold) and most of the MEF2 binding site (underline). (B) Human endothelial cells that are heterozygous for the PHACTR1 34-bp deletion (HET) express 35% less PHACTR1 than wild-type cells (WT)(t-test P=0.03). Data shown is mean ± standard deviations. The experiment was done in triplicate.

Figure 5

Different endothelial cell stimuli do not induce PHACTR1 expression levels as measured by quantitative PCR. (A) Serum-starved HUVEC were treated with VEGF (20ng/mL) for five minutes before RNA was extracted and PHACTR1 transcript levels measured. See Supplemental Figure III for the positive control of this experiment. (B) HUVEC were treated with TNFα (10ng/mL) for 16 hours. RNA was extracted and the levels of PHACTR1 and NFKB1 (positive control) were quantified by quantitative PCR. (C) HUVEC were treated with no, low or high shear stress. RNA was extracted six hours after the beginning of the experiments, and PHACTR1 and KLF2 (positive control) expression levels were measured by quantitative PCR. The same results were observed after 24 hours of shear stress. For all panels, data shown is mean ± standard deviations. The VEGF and TNFα experiments were done in triplicates; we had four replicates for the shear stress experiment. All comparisons are non-significant (t-test P>0.05), except for NFKB1 without and with TNFα treatment (P=7.8×10⁻⁵) and for KLF2 without and with high shear stress (P=0.0012).