Clonal analysis shows that clones reliant on Dnmt3a for HSC retention are primed to resist Npm1c reprogramming—driving a Gata1-lineage bias. In contrast, clones maintaining HSCs even without Dnmt3a undergo more pronounced Npm1c-driven reprogramming.

Gene expression analysis shows both Npm1c-only and Dnmt3a/Npm1c HSCs upregulate Npm1c signature genes. However, Npm1c-only cells activate Myc, E2F, and mTOR/PI3K programs, whereas the double mutants suppress these pathways, shifting fate bias.

Our sequential mutagenesis experiments reveal that combining Dnmt3a‑R878H and Npm1c mutations produces a synergistic effect—with over 60% of clones retaining HSC identity, far exceeding the outcome of either mutation alone.

Since most AML cases with Npm1c mutations are preceded by somatic mutations in epigenetic regulators like DNMT3A, we employed STRACK to track stem cell clones and assess mutational synergy

Interestingly, our tdTom+ vs. tdTom– Npm1c gene signatures mirror the mature vs. primitive signatures from patient GSEA—pointing to cell-of-origin as a key driver of heterogeneity in AML!

Using above mentioned Flt3-Cre system we then can separate HSCs by intrinsic fitness and found that low-fitness (tdTom⁺) HSCs, when mutated with Npm1c, shift into a primitive, low-output state with heightened Hox/Pbx/Meis activation, in line to our ex-vivo data

Since Flt3 marks low‐fitness, high‑output HSCs, we crossed Flt3‑Cre; LSL‑tdTom reporter to Npm1c models to validate in vivo, the distinct reprogramming of HSCs sub-compartment by Npm1c mutation
pmc.ncbi.nlm.nih.gov/articles/PMC...
www.nature.com/articles/s41...

Even more intriguingly, sister clone analysis revealed that Npm1c reprograms HSC fate in a heritable manner: clones with high output in WT become more primitive and differentiation-blocked upon mutation, while low-output clones mature.

Remarkably, by day 27, Npm1c mutant HSCs not only expanded robustly with near-perfect clonality but also exhibited a striking HSC-like state.

We next wondered if our ex vivo expansion cultures could capture mutation-specific reprogramming. Given that Npm1c is known to massively upregulate stemness program—we set out to test its impact at clonal resolution using our STRACK system.
www.science.org/doi/10.1126/...

Intriguingly, R878H HSCs and MPPs display reduced expression of early response genes, suggesting dampened inflammatory activation may underlie their competitive expansion.

Dnmt3a‑R878H mutant HSC clones expand more robustly than their WT sisters. While overall behaviors appear similar, most R878H clones gain extra HSCs—reprogramming differentiation‑biased cells to favor self‑renewal.

Next, To investigate how different cancer driver mutations influence stem cell fates, we performed STRACK on E-SLAM HSCs from a mouse model carrying conditional knockin of Dnmt3a-R878H mutation (R878H) that can be activated using cre recombinases

Our ex vivo cultures recapitulate not only the functional and transcriptional heterogeneity but also showed long-term maintenance and propagation of these programs. Providing unparalleled clonal resolution to explore how cancer driver mutations further shape these dynamics.

Despite starting from a pure HSC pool, we observed a gradual loss of clonality, Our fate heatmaps revealed a spectrum of behavior patterns: some clones initiate early, producing committed progeny yet fading over time, while others emerge later and persist with robust self-renewal

Labeled HSC clones tracked over 27 days showed a spectrum of differentiation and self-renewal behavior

But can long‐term ex vivo HSC cultures truly mimic native stem cell heterogeneity? To establish this, we first labeled and tracked WT HSCs and their progeny for 27 days to understand the dynamic of expanded culture.

To address these discrepancies we developed STRACK which combines a genetic barcoding tool allowing state-fate mapping, ex-vivo expansion culture and transgenic mice containing floxed AML driver mutation
www.science.org/doi/10.1126/...
www.nature.com/articles/s41...
www.nature.com/articles/s41...