High Prevalence of PPM1D Mutations in Therapy-Related AML/MDS Is Due to Context-Specific Clonal Hematopoiesis

Abstract Clonal hematopoiesis (CH) is a condition in which individual hematopoietic stem cells (HSCs) acquire a fitness advantage and contribute disproportionately to peripheral blood production. Somatic mutations in around 20 genes are recurrently associated with CH, with truncating mutations in PPM1D being among the most common. PPM1D encodes the WIP1 protein (Wildtype p53-Induced Phosphatase 1), which is upregulated by p53 during DNA damage and acts homeostatically to downregulate the DNA damage response (DDR). More recently, PPM1D mutations have been observed in the blood of individuals who had previously been exposed to chemotherapeutic agents for prior malignancies. The effect of PPM1D truncating mutations on hematopoiesis remains unclear, as does the precise mechanism by which PPM1D mutations confer a fitness advantage to hematopoietic stem and progenitor cells in CH. To explore the prevalence and features of PPM1D mutations in a focused subset of patients with prior chemotherapy exposure, we screened for PPM1D mutations in 156 patients with t-AML (n=77) and t-MDS (n=79) by targeted-capture deep sequencing of PPM1D and 295 cancer genes. Truncating mutations in PPM1D were detected in 31/156 cases (20%) with a mean VAF of 0.105 (range 0.02-0.48), making it the second most commonly mutated gene in t-AML/MDS after TP53 (29%). In contrast, PPM1D mutations were detected in only 0.5% of a matched de novo MDS/AML cohort, confirming a specific enrichment in therapy-related cases (Fig 1a). Notably, PPM1D-mutant CH was specifically associated with prior exposure to platinum agents (p=0.004) and etoposide (p=0.021). To investigate the mechanisms behind this clinical association, we created PPM1D truncating mutations in multiple cell lines using CRISPR-Cas9. The truncated WIP1 protein was highly stabilized, leading to a net hyperactive effect on the dephosphorylation of DDR pathway members including phospho-p53 and γH2AX. We next asked whether this translates to chemoresistance, by comparing the sensitivities of PPM1D-mutant and wild-type (WT) cells to various agents. PPM1D mutants were resistant to DNA-damaging agents such as cisplatin, etoposide and doxorubicin, but not to vincristine, a microtubule inhibitor, indicating enhanced survival in specific contexts. We then asked whether this conferred an advantage, by competing PPM1D-mutant and WT cells in vitro with multiple exposures. PPM1D mutants did not have an advantage at baseline or in the context of vincristine, but significantly outcompeted their WT counterparts when exposed to cisplatin, etoposide, and doxorubicin. Annexin V staining revealed that PPM1D mutants exhibit diminished apoptosis with DDR-inducing agents, explaining most of their competitive advantage. Co-treatment with a PPM1D inhibitor reversed this gain and may have clinical implications. To understand the in vivo parameters that impact PPM1D-mutant cell fitness, we next generated a novel Ppm1d truncated knock-in mouse model. Characterization of baseline hematopoiesis in the Ppm1d mutant mouse revealed no appreciable differences in lineage composition or proportion of hematopoietic progenitors. To analyze the factors that impact clonal evolution, we competed Ppm1d-mutant and WT cells in specific proportions in bone marrow transplantation. In the context of stress from transplantation alone, Ppm1d-mutant and WT HSCs remained at the same proportions as the initial transplant. Similarly, serial transplantation did not alter their relative contributions to peripheral blood production. In contrast, competitively transplanted mice treated with cisplatin demonstrated a striking selection for Ppm1d mutants in both the peripheral blood (Fig 1b) and LT-HSCs, suggesting that Ppm1d-mutant clones achieve greater fitness and gain a selective advantage under specific extrinsic stressors. This differs from CH-associated mutations in DNMT3A and TET2, where intrinsic characteristics such as enhanced self-renewal appear sufficient to drive clonal expansion. Broadly, this study highlights the significance of context-specificity underlying CH of different somatic mutations. Furthermore, these findings establish the prognostic significance of CH in the development of t-AML/MDS and demonstrate the importance of understanding specific treatment <-> mutation interactions to inform choices of therapeutic interventions in patients with primary cancers. Disclosures Vassiliou: Celgene: Research Funding; KYMAB: Consultancy, Equity Ownership.