Metformin Treatment Abrogates the Competitive Advantage of Dnmt3a R878H HSPCs By Enhancing DNA Methylation Activity

Clonal hematopoiesis of indeterminate potential (CHIP) refers to the clonal expansion of hematopoietic stem and progenitor cells (HSPCs) carrying leukemogenic mutations in individuals without evidence of a blood disorder. CHIP carriers are at an increased risk of blood cancers and cardiovascular diseases. Targeting the clonal expansion of mutant HSPCs has the potential to mitigate CHIP-related illnesses. The most commonly mutated gene in CHIP is DNMT3A, which encodes an enzyme that catalyzes the methyl transfer from S-adenosylmethionine (SAM) to cytosines in DNA. The heterozygous DNMT3A hotspot mutation at R882 exerts a dominant negative effect on the wild-type (WT) allele, resulting in decreased DNA methylation activity. Prior studies have shown that murine HSPCs carrying mutant Dnmt3a R878H (equivalent to R882 in humans) have an advantage over WT cells in competition assays. Here, we aimed to first identify differences in the functional properties between Dnmt3a-mutated and WT HSPCs. Given the crosstalk between mitochondrial metabolism and DNA methylation, we investigated potential differences in mitochondrial activity and found that the level of mitochondrial respiration was higher in Dnmt3aR878H/+ HSPCs than in WT cells. To determine if this translated to differential dependencies, we silenced the expression of critical subunits in the each of five electron transport chain (ETC) complexes. Knockdown of NDUFV1 in Complex I was the most effective in reducing the competitive advantage of Dnmt3aR878H/+ cells in vitro. To explore translational relevance of this finding, we tested the impact of metformin, a pharmacologic inhibitor of Complex I, on Dnmt3a-mutated HSPCs. Treatment with metformin at clinically relevant concentrations suppressed the competitive advantage of mutant cells in vitro. Importantly, this effect was rescued by expression of NDI1, a metformin-resistant yeast analog of Complex I. To extend these findings, we conducted a competitive repopulation assay between CD45.2+ Dnmt3aR878H/+ and CD45.1+ Dnmt3a+/+ HSPCs. Metformin treatment abrogated the in vivo competitive advantage of mutant cells over an 8-month period based on peripheral blood chimerism. To decipher the mechanism of action, we performed transcriptome and metabolome profiling of mutant HSPCs treated with or without metformin for 1 month. GSEA analysis of the transcriptome data revealed a downregulation of genes associated with stemness in metformin-treated cells. Intriguingly, we found that metformin increased the expression of genes involved in one-carbon metabolism, which generates SAM. Consistent with these findings, metabolomics analysis showed that the levels of several key metabolites in one-carbon metabolism, including SAM, were higher in metformin-treated cells compared with controls. Genetic and pharmacologic inhibition of serine hydroxymethyltransferase 2 (SHMT2), a key enzyme in one-carbon metabolism, as well as the addition of exogeneous S-adenosylhomocysteine (SAH), an inhibitor of SAM-dependent methyltransferases, effectively rescued the effect of metformin on mutant cells in vitro. Based on the above findings, we hypothesized that metformin treatment might enhance the residual DNA methylation activity found in Dnmt3aR878H/+ mutated cells by increasing SAM concentration. In support of this hypothesis, we found that exogeneous SAM reproduced the effect of metformin, and HSPCs with homozygous Dnmt3aR878H mutations were insensitive to metformin. Furthermore, we performed DNA methylome profiling of untreated and metformin-treated mutant and WT HSPCs using RRBS. This analysis revealed that the hypomethylated regions (HMRs) in untreated Dnmt3aR878H/+ HSPCs relative to WT controls were enriched for regions of repressed chromatin with H3K27 methylation. Importantly, metformin treatment increased the level of DNA methylation at these regions in mutated HSPCs. In summary, our findings demonstrate that Dnmt3aR878H/+ HSPCs depend on ETC activity and particularly complex I activity to maintain their competitive advantage over WT cells. We propose a model in which inhibition of complex I activity with metformin increases one-carbon metabolism and SAM level, thereby restoring DNA methylation at HMRs and the differentiation of mutant HSPCs. Our findings have important implications for the development of preventive interventions against DNMT3A-mutated CHIP.