Delayed Trp53 Activation Protects Dnmt3a-Mutant Haematopoietic Stem Cells from Inflammatory Attrition

Hematopoietic stem cells (HSCs) accumulate somatic mutations with time, some of which confer a fitness advantage that leads to clonal haematopoiesis (CH). Mutations in DNMT3A, especially at hotspot R882, are highly prevalent in CH and increase the risk of acute myeloid leukaemia (AML). The dominant negative effect of the DNMT3A R882 variant results in global DNA hypomethylation, but it is not clear how this relates to the fitness advantage of mutant HSCs, or their propensity for AML transformation. Whilst DNMT3A CH is common, its progression to myeloid neoplasms (MN) is relatively rare. This heterogeneity suggests that external factors contribute significantly to disease development, with previous experimental evidence demonstrating that DNMT3A-mutant HSCs manifest an advantage primarily under conditions of stress.

We performed single-cell RNA-Seq and ATAC-Seq on murine HSCs from a conditional DNMT3A R882H knock-in model (Dnmt3a^R882H/+ or mut) and found abnormal activation of Interferon (INF) and p38-MAPKα pathways at steady state. This led us to investigate the functional response of Dnmt3a^R882H/+ HSCs to IL-1β-induced emergency granulopoiesis. In vitro, single mut long-term HSCs (LT-HSCs) exposed to IL-1β exited G0 earlier than WT and gave rise to larger single-cell-derived colonies at 10 days. These, showed a greater retention of immature (c-kit⁺) cells compared to WT and, despite their increased proliferation, mut LT-HSCs actually retained higher clonogenic efficiency in replating assays. Similarly, in vivo IL-1β stimulation caused a more sustained exit from quiescence of mut LT-HSCs at 24h.

We next performed RNA-Seq on LT-HSCs over a time-course of sustained ex vivo IL-1β stimulation. At 12h we observed upregulation of genes associated with nuclear replication fork and DNA damage sensing in mut HSCs. This was confirmed by immunofluorescence, with the accumulation of nuclear phospho-RPA2 foci under the same conditions. Interestingly, and consistent with our in vivo and in vitro phenotype, mut HSCs upregulated genes associated with G1 to S cell cycle progression at 24h, highlighting an aberrant response to replicative stress at the transcriptomic level. This response is mediated by the TP53–p21–DREAM–E2F/CHR axis, and we observed a lack of repression of DREAM targets in mut LT-HSCs exposed to IL-1β, demonstrating dysregulation of the TP53-DREAM pathway in mutant LT-HSCs.

We then studied the response of mutant progenitors (HSPCs) to irradiation, a stimulus known to elicit a strong DNA-damage/Trp53 response. CUT&RUN at 2h post-irradiation showed reduced Trp53 binding to targets in mutant cells, but RNA-Seq at 6h indicated similar expression of these Trp53 targets between mut and WT, overall implying a delay in Trp53 activation rather than complete abrogation of its function in mutant HSPCs. Potentially secondary to this delay, we observed uniform deregulation of DNA repair genes in mutant HSPCs. Consistent with these transcriptional changes, we observed accumulation of nuclear γH2AX (a surrogate of double strand breaks - DSBs) in mut cells, that persisted at 24h post-irradiation. Mutant HSPCs however, exhibited higher tolerance to DSBs, retaining their proliferative advantage and showing no increase in apoptosis, when single LT-HSCs were irradiated and cultured with IL-1β. Importantly, mut LT-HSCs showed a marked advantage over WT cells, following ex vivo irradiation and competitive transplant into myeloablated congenic recipients, underscoring their higher resilience under stress.

Our studies discover a mechanistic link between DNMT3A-R882H and TP53 that does not increase levels of genomic instability, a characteristic phenotype of biallelic TP53 loss. By analysing a large cohort (3,323) of MN patients, we find that DNMT3A R882 phenocopies monoallelic TP53 alteration, but that both differ significantly from multi-hit TP53 MN, with only the latter associated with marked genomic instability and inferior outcomes.

In conclusion, our findings highlight a complex interplay between DNMT3A R882 and TP53 in HSCs, whereby a delayed Trp53 activation under stress conditions such as emergency granulopoiesis and radiation exposure, enhances resilience of Dnmt3a^R882H/+ HSCs to these stressors. This phenotype contributes to clonal dominance, disease progression and therapy resistance.