Germline and Somatic Mutations in DNA Methyltransferase 3A (DNMT3A) Predispose to Pulmonary Arterial Hypertension (PAH) in Humans and Mice: Implications for Associated PAH

BackgroundMutations are found in 10-20% of idiopathic PAH (IPAH) patients, but none are consistently identified in connective tissue disease-associated PAH (APAH), which accounts for [~]45% of PAH cases. TET2 mutations, a cause of clonal hematopoiesis of indeterminant potential (CHIP), predispose to an inflammatory type of PAH. We now examine mutations in another CHIP gene, DNMT3A, in PAH. MethodsWe assessed DNMT3A mutation prevalence in PAH Biobank subjects as compared with controls, first using whole exome sequencing (WES)-derived CHIP calls in 1832 PAH Biobank patients versus 7509 age-and sex-matched gnomAD controls. We then performed deep, targeted panel sequencing of CHIP genes on a subset of 710 PAH Biobank patients and compared the prevalence of DNMT3A mutations therein to an independent pooled control cohort (N = 3645). In another cohort of 80 PAH patients and 41 controls, DNMT3A mRNA expression was studied in peripheral blood mononuclear cells (PBMCs). Finally, we evaluated the development of PAH in a conditional, hematopoietic, Dnmt3a knockout mouse model. ResultsDNMT3A mutations were more frequent in PAH cases versus control subjects in the WES dataset (OR 2.60, 95% CI: 1.71-4.27). Among PAH patients, 33 had DNMT3A variants, most of whom had APAH (21/33). While 21/33 had somatic mutations (female:male 17:4), germline variants occurred in 12/33 (female:male 11:1). Hemodynamics were comparable with and without DNMT3A mutations (mPAP=58{+/-}21 vs. 52{+/-}18 mmHg); however, patients with DNMT3A mutations were unresponsive to acute vasodilator testing. Targeted panel sequencing identified that 14.6% of PAH patients had CHIP mutations (104/710), with DNMT3A accounting for 49/104. There was a significant association between all CHIP mutations and PAH in analyses adjusted for age and sex (OR 1.40, 95% CI: 1.09-1.80), though DNMT3A CHIP alone was not significantly enriched (OR:1.15, 0.82-1.61). DNMT3A expression was reduced in patient-derived versus control PAH-PBMCs. Spontaneous PAH developed in Dnmt3a-/- mice, and it was exacerbated by 3 weeks of hypoxia. Dnmt3a-/- mice had increased lung macrophages and elevated plasma IL-13. The IL-1{beta} antibody canakinumab attenuated PAH in Dnmt3a-/- mice. ConclusionsGermline and acquired DNMT3A variants predispose to PAH in humans. DNMT3A mRNA expression is reduced in human PAH PBMCs. Hematopoietic depletion of Dnmt3a causes inflammatory PAH in mice. DNMT3A is a novel APAH gene and may be a biomarker and therapeutic target.