DNMT3A-driven Clonal Expansion of Hematopoiesis Influences Autoimmunity Mediated by Distinct Immune Cell Populations: A Mendelian Randomization Study

INTRODUCTION: The causal role of genetic liability to DNMT3A-driven clonal hematopoiesis of indeterminate potential (CHIP) in autoimmune diseases (AIDs) remains unclear. We aimed to assess its associations with 14 AIDs and to explore whether circulating immune phenotypes mediate these associations. METHODS: We conducted a two-sample Mendelian randomization (MR) study of genetic liability to DNMT3A-driven CHIP and 14 AIDs. Two-step MR was then used to evaluate mediation by 731 circulating immune phenotypes. The Pleiotropic Clustering framework for Mendelian randomization (PCMR) was applied as a key sensitivity analysis to address correlated horizontal pleiotropy. RESULTS: A significant causal relationship was identified between genetic liability to DNMT3Adriven CHIP and systemic lupus erythematosus (SLE; OR = 348.020, p = 0.016, FDR q = 0.045), polymyositis (PM; OR = 2.28E-06, p = 0.012, FDR q = 0.043), dermatomyositis (DM; OR = 5.64E-05, p = 0.012, FDR q = 0.043), and psoriasis (PsO; OR = 0.091, p = 0.007, FDR q = 0.043). The large effect estimate for SLE should be interpreted cautiously, as it reflects the scale of genetically predicted liability and was accompanied by a wide confidence interval. Further, HLA DR+ CD4+ T cell %T cell (15.1% mediation) and CD25 on CD24+ CD27+ B cell (9.73% mediation) mediated the relationship between genetic liability to DNMT3A-driven CHIP and SLE. HLA DR+ CD4+ T cell Absolute Count mediated the association between genetic liability to DNMT3A-driven CHIP and PsO, with a mediated proportion of 27.5%. DISCUSSION: These findings support disease-specific rather than uniform immune effects of DNMT3A-driven CHIP liability. The large SLE estimate should be interpreted cautiously because it reflects genetically predicted liability and has a wide confidence interval. In addition, the exposure was not a direct somatic mutation burden; some GWAS datasets were relatively small, and all participants were of European ancestry. CONCLUSION: Genetic liability to DNMT3A-driven CHIP shows heterogeneous associations with AIDs. Selected circulating immune phenotypes may mediate the associations with SLE and PsO, whereas no circulating immune-cell mediator was identified for PM or DM. Further mechanistic and longitudinal validation is needed.