Novel risk genes and mechanisms implicated by exome sequencing of 2572 individuals with pulmonary arterial hypertension

Abstract

Background: Group 1 pulmonary arterial hypertension (PAH) is a rare disease with high mortality despite recent therapeutic advances. Pathogenic remodeling of pulmonary arterioles leads to increased pulmonary pressures, right ventricular hypertrophy, and heart failure. Mutations in bone morphogenetic protein receptor type 2 and other risk genes predispose to disease, but the vast majority of non-familial cases remain genetically undefined.

Methods: To identify new risk genes, we performed exome sequencing in a large cohort from the National Biological Sample and Data Repository for PAH (PAH Biobank, n = 2572). We then carried out rare deleterious variant identification followed by case-control gene-based association analyses. To control for population structure, only unrelated European cases (n = 1832) and controls (n = 12,771) were used in association tests. Empirical p values were determined by permutation analyses, and the threshold for significance defined by Bonferroni's correction for multiple testing.

Results: Tissue kallikrein 1 (KLK1) and gamma glutamyl carboxylase (GGCX) were identified as new candidate risk genes for idiopathic PAH (IPAH) with genome-wide significance. We note that variant carriers had later mean age of onset and relatively moderate disease phenotypes compared to bone morphogenetic receptor type 2 variant carriers. We also confirmed the genome-wide association of recently reported growth differentiation factor (GDF2) with IPAH and further implicate T-box 4 (TBX4) with child-onset PAH.

Conclusions: We report robust association of novel genes KLK1 and GGCX with IPAH, accounting for ~ 0.4% and 0.9% of PAH Biobank cases, respectively. Both genes play important roles in vascular hemodynamics and inflammation but have not been implicated in PAH previously. These data suggest new genes, pathogenic mechanisms, and therapeutic targets for this lethal vasculopathy.

Keywords: Case-control association testing; Exome sequencing; Genetics; Pulmonary arterial hypertension.

Fig. 1

Contribution of known PAH risk genes in the PAH Biobank cohort (n = 2572 cases). BMPR2, ACVRL1/ENG, TBX4. Other established risk genes included in the analysis: BMPR1A, BMPR1B, CAV1, EIF2AK4, KCNK3, SMAD4, and SMAD9. Newly validated risk genes: ABCC8, ATP13A1, GDF2, KCN5A, KLF2, SMAD1, and SOX17

Fig. 2

Age of disease onset for PAH Biobank cases with rare deleterious variants in known PAH risk genes. a Box plots showing median, interquartile range, and min/max values for age of disease onset (i.e., age at diagnostic right heart catheterization). The number of cases carrying variants for each gene is given above each box plot. Genes represented by less than four cases are not shown. b Histogram plots showing age-of-onset distributions for the whole cohort (n = 2572), BMPR2 (n = 180), or TBX4 (n = 23) variant carriers. Red vertical lines indicate the group means. BMPR2 carriers had a younger mean age of onset (mean = 37 years, SD = 15; Mann-Whitney U test, p = 1.1E−15) but no enrichment of child-onset cases (binomial test p = 1, RR = 0.93) compared to the whole cohort, whereas TBX4 carriers had a younger mean age of onset (mean = 29 years, SD = 25; Mann-Whitney U test, p = 0.001) and significant enrichment of child-onset cases (binomial test p = 6.5E−08, RR = 12.3) compared to the whole cohort

Fig. 3

Gene-based association analysis using 1832 European cases from all PAH subclasses and 12,771 European controls. a Results of a binomial test confined to rare LGD and D-Mis (REVEL variable threshold) variants in 20,000 protein-coding genes. Horizontal gray line indicates the Bonferroni-corrected threshold for significance. b Complete list of top association genes (p ≤ 0.001)

Fig. 4

Gene-based association analysis using 812 European IPAH cases and 12,771 European controls. a Results of a binomial test confined to rare LGD and D-Mis (REVEL variable threshold) variants in 20,000 protein-coding genes. Horizontal gray line indicates the Bonferroni-corrected threshold for significance. b Complete list of top association genes (p ≤ 0.001)

Fig. 5

Locations of rare, predicted deleterious variants in KLK1 (a) and GGCX (b) across the PAH Biobank cohort (n = 2572 cases). Locations are provided within the two-dimensional protein structures. The numbers of variants at each amino acid position are indicated along the y-axes. The vertical gray lines indicate exon borders. D-MIS, predicted damaging missense; LGD, likely gene disrupting (stop-gain, frameshift, splicing)