Identification of the Genetic Mechanism Responsible for Racially-Dimorphic Expression of the Thrombin-Receptor Regulator, Pctp

Genome wide association studies have identified numerous single nucleotide polymorphisms (SNPs) associated with various healthy and pathological phenotypes.  The majority of these SNPs do not fall into protein-coding regions of the genome, leading to the hypothesis that variants in genomic regulatory regions are critical regulators of physiology.  Efforts by the ENCODE project and others to annotate the genome have enabled researchers to better identify and test SNPs for their functional effect on gene expression.  We now report the identification of a SNP responsible for the racially differential expression of phosphatidylcholine transfer p (PCTP), a protein we have previously identified as a regulator of the human platelet thrombin receptor, protease activated receptor 4 (PAR4).  In the Platelet RNA and eXpression 1 (PRAX1) study, we learned that platelets from black subjects were more active in response to signaling through the PAR4 receptor and contained approximately four times more PCTP mRNA and 50% more PC-TP protein as compared to platelets from white subjects.  PCTP levels are significantly associated platelet PAR4 reactivity even after accounting for a racially dimorphic PAR4 polymorphism (Ala120Thr) that alters the receptor’s function. 

We obtained from the 154 PRAX1 subjects genome wide platelet mRNA expression profiles and 4.3 million genotypes.  Using this expression and genotype data, we were able to perform an expression Quantitative Trait Locus (eQTL) analysis to identify single nucleotide polymorphisms associated with the expression level of genes located within 50kb of the variant.  This analysis revealed 16 highly linked SNPs associated with PCTP levels at genome-wide significance (P < 10^-6).

eQTL SNPs can influence gene expression through a variety of mechanisms: (1) Altering the core promoter; (2) Altering the binding site of transcriptional regulators such as a transcription factors (TFs); (3) Altering RNA stability signals such as miRNA binding sites.  Because none of the PCTP eQTL SNPs fall within the annotated core promoter, we reasoned the causative SNPs would fall within predicted transcription factor binding sites or miRNA target sites.  To prioritize the 16 candidate SNPs for functional testing, we annotated each one according to the following criteria: (1) The SNP fell within a predicted binding site for a platelet-expressed TF or miRNA; (2) ChIP-Seq data from megakaryocytes, CD34⁺ hematopoietic stem cells or K562 erythroleukemia cells indicated TF binding in the region of the SNP; (3) The SNP falls in a regulatory region as indicated by epigenetic marks or DNAse hypersensitivity; (4) The allele frequency of the SNP is racially dimorphic, corresponding with PCTP expression.

Using these criteria, we investigated rs2912553, a racially dimorphic SNP located in the first intron of PCTP.   rs2912553 falls within a DNAse hypersensitive genomic locus that contains Lys4 monomethylated histone H3, a marker of enhancers.  Cloning different sized fragments of this region 5’ to a luciferase expression cassette replicated the observed PCTP expression pattern, with the vectors containing the allele most common in black subjects generating 2-8 fold higher luciferase expression than vectors containing the common allele in whites.  In agreement with these results electromobility shift assays indicate that protein complexes have a threefold higher affinity for the common allele in blacks as compared to the common allele in whites (P=0.02), and that these complexes can be interrupted with an anti-GATA1 antibody.  Concordantly, siRNA knockdown of GATA1 expression reduced luciferase activity in both alleles.  Together, these data indicate that the racially dimorphic SNP, rs2912553, causes differential recruitment of a GATA-1 containing transcriptional complex that is responsible for higher PCTP expression in blacks.  This suggests the hypothesis that these genetic variances contribute to the dissimilar thrombotic risk between blacks and whites.  Future studies should address the utility of rs2912553 as a biomarker for diseases or drug effects that differ by race.