AUTHORS
Yuki Nishida, MD, PhD, Edward Ayoub, PhD2, Yoonkyu Lee, PhD3, David Lehotzky, PhD4, Gaddy Getz, PhD5, Michael Andreeff, MD, PhD6
AFFILIATIONS
1 Section of Molecular Hematology and Therapy, Houston, Texas, United States
2 Section of Molecular Hematology and Therapy, Houston, Texas, United States
3 Section of Molecular Hematology and Therapy, Houston, Texas, United States
4 Broad Institute of MIT and Harvard, Boston, Massachusetts, United States
5 Broad Institute of MIT and Harvard, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States, Krantz Family Center for Cancer Research, Massachusetts General Hospital, Boston, Massachusetts, United States
6 Section of Molecular Hematology and Therapy, Houston, Texas, United States
ABSTRACT
Background:
TP53 mutations account for approximately 10% of de novo and 30% of relapsed/refractory acute myeloid leukemia (AML), with the most devastating outcome and no approved targeted therapeutics are available. Recent studies have shown the enrichment of immature and erythroid populations in TP53 mutated (TP53mut) AML especially after venetoclax (Ven)-based therapies (Rodriguez-Meira, 2023; Daver, 2025; Waclawiczek, 2026), however, the underlying mechanisms of the simultaneous enrichment of immature and erythroid populations have not been well established.
Methods:
We performed single-cell RNA sequencing (scRNAseq) in normal bone marrow (NBM) (N = 3) and paired, pre- and post-treatment TP53 wild-type (N = 7) and TP53mut (N = 19) AML samples (N = 55 in total). Copy number variations (CNV) were annotated using numbat (Gao, 2023) and CopyKAT (Gao, 2021) algorithm. Pseudotime and LSC17 (Ng, 2016) scores were also determined. We employed univariate linear regression using collecTRI regulatory network (Müller-dot, 2023) to compute the activity of master transcription factors (TFs) for immaturity, erythroid and myeloid differentiation. To infer the activity of MECOM in TP53mut AML megakaryo-erythroid/mast cell progenitors (MEP) over NBM samples, we computed MECOM footprint, where the expression levels of MECOM’s downstream targets in NBM counterparts were anchored. External datasets (Waclawiczek, 2026, Andricovich, 2025) were used for further analyses.
Results:
TP53mut AMLs at diagnosis had higher frequencies of CNV+ cells (45.6% vs 58.5% in TP53wt vs TP53mut, respectively). Strikingly, pseudotime analysis revealed CNV+ terminally differentiated erythroid cells and monocytes in TP53mut AMLs, while CNV+ cells were restricted in immature components in TP53wt AMLs. On the other hand, LSC17 scores were higher in TP53mut CNV+ cells than NBM and TP53wt CNV+ cells (0.023, -0.007 and -0.025 for TP53mut, NBM and TP53wt samples, respectively, Kruskal-Wallis p = 0.033), suggesting that malignant clones can be terminally differentiated into erythroid and myeloid lineages while retaining immaturity in TP53mut AMLs. The master TFs for erythroid and myeloid lineages demonstrated comparable expression trajectories from immature toward mature stages between NBM and TP53mut AMLs. We identified that MECOM TF activity was highest in TP53mut AML CNV+ cells compared to NBM and TP53wt CNV+ cells (1.7, 1.4 and -0.1 for TP53mut, TP53wt and NBM, respectively, FDR p = 0.006 for TP53mut vs NBM), supporting the maintained immaturity in TP53mut CNV+ cells. The proportions of GATA1+ immature CNV+ cells were increased in TP53mut compared to TP53wt, suggesting an erythroid priming in TP53mut immature CNV+ cells (46.0%, 3.3% and 22.7% for NBM, TP53wt and TP53mut AML samples, respectively, Mann-Whitney (MW) p = 0.057). On the other hand, TP53mut AML CNV+ MEP exhibited retained GATA2 levels compared to NBM MEP (32.6% vs 90.3% for NBM vs TP53mut AML, respectively, MW p = 0.063). MECOM footprint, another readout for immaturity, was sustained in TP53mut CNV+ compared to NBM MEP and later-stage erythroid cells. GATA1 onset was increased post-Ven/HMA in available samples (N = 3, 0 to 10%, 20 to 71% and 89 to 97%). The delayed MECOM silencing persisted in post-treatment samples in our and the Waclawiczek dataset, suggesting that desynchronized erythroid differentiation with increased immaturity was sustained post-Ven/HMA. Data suggest an early onset of GATA1 and delayed silencing of GATA2 and MECOM in TP53mut AMLs, suggesting an erythroid-prone trajectory with retained immaturity. Lastly, we asked whether targeting BCL-XL overcomes desynchronized erythroid populations in a publicly available murine erythroid leukemia dataset (Andricovich), based on increased levels of BCL2L1 at late-stage erythroid cells in our and Waclawiczek datasets. BCL-XL inhibition effectively depleted late-stage but enriched immature-stage erythroid cells in mice, suggesting a potential limitation of BCL-XL inhibition in TP53mut AML with erythroid differentiation.
Conclusions:
TP53mut AMLs contain CNV+ differentiated erythroid and myeloid cells while retaining immaturity. Early onset of GATA1 and delayed silencing of GATA2 and MECOM suggest desynchronization of differentiation, especially toward the erythroid lineage. Our analyses propose TF expression and activities underlying the paradoxical enrichment of immature and erythroid populations pre- and post-treatment in TP53mut AMLs.
DISCLOSURES
Yuki Nishida, MD, PhD: Sellas Biosciences - Research Funding; Kintor Pharmaceuticals - Research Funding