Abstract 4369178: Causal Proteomic Consequences of Clonal Hematopoiesis: Integrated Human Genetic and Murine Validation

Abstract

Background: Clonal hematopoiesis of indeterminate potential (CHIP), the age-related clonal expansion of hematopoietic stem cells with preleukemic driver mutations, is a novel risk factor for cardiovascular diseases (CVD). While associations between CHIP and plasma proteins are emerging, the causal nature of these relationships remains undetermined. Hypothesis: We aimed to identify specific CHIP driver mutations that causally alter plasma protein levels using human genetics and validate key findings in experimental models. Approach: We analyzed 61,833 participants from the NHLBI TOPMed Program (SomaScan for proteomics measurement) and UK Biobank (Olink for proteomics measurement) with paired DNA sequencing and proteomics. Associations between CHIP variables (composite or drivers DNMT3A , TET2 , ASXL1 ) and proteins were analyzed separately by platform. Bi-directional Mendelian Randomization (MR) was performed to assess causal relations between CHIP, DNMT3A , or TET2 and protein levels. Proteins implicated by MR for TET2 (LCN2, MPO, FLT3LG) were selected for validation via ELISA in 8-9 week-old hematopoietic Tet2 -/- vs wild-type (WT) mice. Results: MR analyses identified multiple instances where CHIP likely causes proteomic changes, with less evidence for reverse causality. In TOPMed (SomaScan), we identified 9 causal pairs (FDR<0.05) among 35 (24 examined) significant CHIP-protein pairs, with TET2 causally increasing MPO levels being the strongest (beta [SE]:0.022[0.008]; P=3.6x10 -3 ). In UK Biobank (Olink), among 473 (318 examined) significant CHIP-protein pairs, 121 were causal, with TET2 showing the strongest effects on increased LCN2 (beta[SE]:0.056[0.008]; P=4.3x10 -11 ) and decreased FLT3LG (beta[SE]:-0.089[0.015]; P=1.1x10 -9 ). The causal effect of TET2 on LCN2 was consistent across platforms. Murine experiments corroborated these findings: hematopoietic Tet2 -/- mice exhibited significantly increased plasma MPO and LCN2 levels compared to WT controls. Plasma FLT3LG levels were not significantly different, though a decrease in Tet2 -/- mice was directionally consistent with human MR. Conclusions: This study provides robust human genetic and experimental evidence for causal effects of CHIP, particularly TET2 mutations, on the plasma proteome. This validation of MR-identified proteomic changes strengthens the causal link and offers potential mechanistic insights into how CHIP may influence downstream outcomes, including CVD.