Rare and low-frequency coding variants in CXCR2 and other genes are associated with hematological traits

Abstract

Hematological traits are important clinical parameters. To test the effects of rare and low-frequency coding variants on hematological traits, we analyzed hemoglobin concentration, hematocrit levels, white blood cell (WBC) counts and platelet counts in 31,340 individuals genotyped on an exome array. We identified several missense variants in CXCR2 associated with reduced WBC count (gene-based P = 2.6 × 10(-13)). In a separate family-based resequencing study, we identified a CXCR2 frameshift mutation in a pedigree with congenital neutropenia that abolished ligand-induced CXCR2 signal transduction and chemotaxis. We also identified missense or splice-site variants in key hematopoiesis regulators (EPO, TFR2, HBB, TUBB1 and SH2B3) associated with blood cell traits. Finally, we were able to detect associations between a rare somatic JAK2 mutation (encoding p.Val617Phe) and platelet count (P = 3.9 × 10(-22)) as well as hemoglobin concentration (P = 0.002), hematocrit levels (P = 9.5 × 10(-7)) and WBC count (P = 3.1 × 10(-5)). In conclusion, exome arrays complement genome-wide association studies in identifying new variants that contribute to complex human traits.

Figure 1

Association results in the meta-analysis MHI+WHI (N=24,814) for (A) hematocrit at the TFR2-EPO locus on chromosome 7 and (B) platelet count at the SH2B3 locus on chromosome 12. The x-axis corresponds to genomic coordinates. (A) For TFR2-EPO, we provide association results for each variant in the region before (circles) and after (stars) conditioning on the GWAS sentinel SNP at the locus (rs7385804). (B) Similarly, for SH2B3 we provide association results without (circles) and with (stars) adjustment for genotypes at the GWAS sentinel SNP rs3184504. In the GWAS row, we represent all markers in linkage disequilibrium (LD) with the GWAS sentinel SNP (marked by an inverted triangle) based on data from European populations in the 1000 Genomes Project: the size of the circle is proportional to the strength of LD. We added a vertical line between the GWAS and ExomeChip rows if GWAS LD proxies were also present on the ExomeChip array. On the ExomeChip row, all markers analyzed in the region are represented as triangles (intronic, intergenic and synonymous variants) or diamonds (missense, nonsense and splice site variants). Vertical lines between the ExomeChip and Freq 0.5 rows represent allele frequencies in the meta-analysis MHI+WHI.

Figure 2

Rare and low-frequency missense variants in CXCR2 are associated with lower white blood cell (WBC) count. The symbol legend is in the upper left corner of the figure. A vertical line represents the mean WBC count for each study. The middle of each color-coded circle (black=MHI, white=WHI, grey=SHIP) corresponds to the mean WBC count for individuals that carry the corresponding missense variants and the size of the circles is correlated with allele frequency. The three white circles in the lower left corner of the figure are provided as references and represent variants with minor allele frequency of 0.01%, 0.1% and 0.5%. For rs2228413 (p.Arg294Gln), there is a participant from MHI that is homozygote for the rare allele (star).

Figure 3

Functional characterization of the CXCR2fs mutant receptor. (A) Intracellular ERK1/2 signaling in HeLa cells transfected with wild type or mutant CXCR2 receptor. HeLa cells transfected with CXCR2wt-YFP, CXCR2fs-YFP or YFP alone were stimulated with CXCL8 (100 ng/mL) and lysates collected over the time course indicated to assess the activation of the ERK1/2 pathway. Peak phosphorylation following stimulation of CXCR2wt occurred at 5 min with significant attenuation of activation by 15 min. No activation over baseline was detected in cells expressing CXCR2fs. ERK1/2 activation at 5 min following stimulation of endogenous CXCR4 with CXCL12 (100 ng/mL) was used as a positive control (+). Protein abundance of CXCR2wt-YFP (wt), CXCR2fs-YFP (fs) and the YPF control (Y) proteins in transfected cell lines are shown (right panel). (B) Chemotaxis assay results for HeLa cells expressing CXCR2wt-YFP or CXCR2fs-YFP. Transiently transfected HeLa cells were seeded in the top chamber of modified Boyden chambers and CXCL8 (100 ng/mL) was added to media in the bottom chamber. The chemotactic index (migration in presence of CXCL8/migration in absence of CXCL8) was calculated from the results of assays performed in triplicate. Robust migration was observed in cells expressing the wild type CXCR2 fusion protein (gray bars), but not in those expressing the frameshift mutant (open bars) or in untransfected cells (black bars). Overexpression of each construct was confirmed by detecting YFP fluorescence (expressed as relative fluorescence units; RFU) in aliquots of transfectants (right panel). *P<0.02, Student’s t-test, n.s., non-significant.