Common variants in the human platelet PAR4 thrombin receptor alter platelet function and differ by race

Abstract

Human platelets express 2 thrombin receptors: protease-activated receptor (PAR)-1 and PAR4. Recently, we reported 3.7-fold increased PAR4-mediated aggregation kinetics in platelets from black subjects compared with white subjects. We now show that platelets from blacks (n = 70) express 14% more PAR4 protein than those from whites (n = 84), but this difference is not associated with platelet PAR4 function. Quantitative trait locus analysis identified 3 common single nucleotide polymorphisms in the PAR4 gene (F2RL3) associated with PAR4-induced platelet aggregation. Among these single nucleotide polymorphisms, rs773902 determines whether residue 120 in transmembrane domain 2 is an alanine (Ala) or threonine (Thr). Compared with the Ala120 variant, Thr120 was more common in black subjects than in white subjects (63% vs 19%), was associated with higher PAR4-induced human platelet aggregation and Ca2+ flux, and generated greater inositol 1,4,5-triphosphate in transfected cells. A second, less frequent F2RL3 variant, Phe296Val, was only observed in blacks and abolished the enhanced PAR4-induced platelet aggregation and 1,4,5-triphosphate generation associated with PAR4-Thr120. PAR4 genotype did not affect vorapaxar inhibition of platelet PAR1 function, but a strong pharmacogenetic effect was observed with the PAR4-specific antagonist YD-3 [1-benzyl-3(ethoxycarbonylphenyl)-indazole]. These findings may have an important pharmacogenetic effect on the development of new PAR antagonists.

© 2014 by The American Society of Hematology.

Figures

Figure 1

SNPs in F2RL3 associate with PAR4 reactivity. (A) PAR4 ARS values of PRAX1 subjects by self-identified race. P = 6.8 × 10−9, 2-tailed t test. (B) Microarray analysis of F2RL3 gene expression. Values are normalized and log2 transformed. P = 0.043, 2-tailed t test. (C) PAR4 protein levels normalized to GAPDH. P = .0427, 1-tailed t test. (D) Correlation analysis of platelet PAR4 reactivity with PAR4 protein expression level.

Figure 2

Association of racially dimorphic PAR4 variants with platelet PAR4 function. (A) Manhattan plot showing association of SNPs in the F2RL3 gene with platelet PAR4-AP reactivity. Each circle represents a SNP. The circles for rs773903 and rs773904 overlap. Y-axis is –log10 of the P values for association, controlling for age, race, and sex; x-axis, chromosomal location. (B) Schematic of F2RL3 and LD plot of SNPs in F2RL3. Nonsynonymous SNPs coding changes are shown above the schematic. Red rs numbers indicate SNPs identified in quantitative trait locus analysis; blue rs number indicates a less common SNP observed only in black subjects. (C) Genotype and allele frequency of rs773902 in black and white PRAX1 subjects. (D) Worldwide allele frequency of rs773902 in HGDP data set.

Figure 3

PAR4 variants and platelet function. Platelet PAR4 reactivity by rs773902 genotype for (A) all participants and (B) each race individually. P = 9.15 × 10−16, association of PAR4 reactivity with rs773902 genotype by linear regression. (C) Platelet ARS of PRAX1 subjects for AA, adenosine 5′-diphosphate, collagen-related peptide, α-CD9 antibody (CD9), PAR1-AP (PAR1), and PAR4-AP (PAR4), segregated by rs773902 genotype. One-way ANOVA showed no association between rs773902 and any agonist except PAR4-AP (PAR4-AP [A] is shown again for ease of comparison). (D) Platelet calcium flux from subjects genotyped for rs773902 (cohort 2). Platelets were treated with 50 μM PAR4-AP. n = 10 GG; n = 5 AA/AG. P = .03 1-tailed t test. (E) Platelet PAR4 reactivity by rs773902 and rs2227346 genotypes among self-identified black subjects in PRAX1. P = 1.75x10−8, association of PAR4 reactivity with rs2227346 genotype partial F-test after controlling for rs773902 genotype. For all panels, the box represents the interquartile range, the horizontal line in box is the median, and the whiskers represent 1.5 times interquartile range.

Figure 4

Functional differences in IP3 generation. (A) Representative flow cytometry tracings for surface FLAG-PAR4 expression. 120A-296V indicates the expression construct for PAR4-Ala120-Val296, 120A-296F for PAR4-Ala120-Phe296, 120T-296F for PAR4-Thr120-Phe296, and 120T-296V for PAR4-Thr120-Val296. Blue, α-FLAG; red, IgG control. (B) IP3 generation in 293 cells transfected with FLAG-PAR4 variants or control stimulated with 1 mM PAR4-AP. IP3 levels were normalized to cell number and PAR4 surface expression. For each independent experiment, all data points were calculated relative to the IP3 quantified in the PAR4-Thr120-Phe296 expressing cells after 30 seconds of treatment (ie, the maximum value measured in all experiments [percentage of maximum IP3 level]). n = 3. P value for 120T-296F vs 120A-296F at 30s = 0.01, 2-tailed t test.

Figure 5

PAR4 variants are differentially susceptible to pharmacological inhibition. (A-C) Platelets from donors genotyped for rs773902 were washed, incubated with the indicated concentrations of vorapaxar (VPX) (A-B) or YD-3 (C), and then stimulated with (A) 5 μM PAR1-AP or (B-C) 100 μM PAR4-AP. Aggregation was measured using light transmission aggregometry. n = 3 each genotype. Horizontal lines indicate mean. P = .02.