Proteogenomic analysis of NPM1-mutated AML reveals clinical heterogeneity related to differentiation states and mitochondrial metabolism

Abstract Introduction: Acute myeloid leukemia (AML) is a clonal malignancy characterized by specific genetic alterations that impact cell differentiation and survival programs. Only few studies have so far explored proteomic and proteogenomic disease characteristics that furthermore contribute to AML pathogenesis. Methods: We conducted a large-scale proteomic analysis on bone marrow blasts from 284 patients with newly diagnosed NPM1-mutated AML enrolled in the phase III AMLSG 09-09 trial NCT00893399, comparing intensive chemotherapy plus all-trans retinoic acid with or without gemtuzumab ozogamicin. We quantified the expression of 7,223 proteins. Additionally, targeted DNA sequencing of 263 genes frequently mutated in myeloid malignancies was performed as published recently (Cocciardi et al., Hemasphere 2025). Results: Applying previously defined gene signatures that distinguish immature leukemic stem cell (LSC)-like states from more differentiated granulocyte-macrophage progenitor (GMP)-like promonocyte-like, and monocyte-like leukemic blasts, we identified a robust differentiation gradient as a principal source of proteomic variance. This gradient was significantly associated with genetic alterations, particularly an enrichment of FLT3-ITD (adjusted p-value 0.001) in immature disease, and the absence of DNMT3A (adjusted p-value 0.001), NRAS (adjusted p-value 0.026) and PTPN11 (adjusted p-value 0.001) mutations. Immature AML strongly correlated with BCL2 protein expression (Spearman's rho = 0.7), consistent with previous evidence of sensitivity to venetoclax, a clinically approved BCL2 inhibitor. Building on our prior observation linking venetoclax sensitivity to a mitochondrial protein-rich AML subtype, we applied the previously described mitochondrial gene signature and observed a positive correlation with immature NPM1-mutated AML (Spearman's rho = 0.48) and a negative correlation with more committed differentiation states (Spearman's rho=-0.78). Interestingly, although the Mito-score generally declined with increased differentiation, a modest increase was observed in highly differentiated, monoblast-like AML. Functional enrichment analysis using mitochondrial gene sets from the Broad Institute's MitoCarta database revealed distinct mitochondrial functional states between AML with either immature or differentiated blast characteristics. Immature AML was enriched for mitochondrial DNA translation and mitochondrial central dogma terms, whereas AML characterized by more differentiated blasts was enriched for terms related to mitochondrial metabolism, suggesting both quantitative and qualitative differences in mitochondrial biology within NPM1-mutated AML. Clinically, a high Mito-score (top quartile) was significantly associated with worse event-free survival (EFS; log-rank p-value = 0.026, HR = 4.46) and overall survival (OS; log-rank p-value = 0.03, HR = 4.74) in the entire cohort, as well as in genetically favorable-risk patients characterized by NPM1 mutations in the absence of FLT3-ITD (EFS log-rank p-value = 0.021, HR = 5.4; OS log-rank p-value = 0.065, HR = 5.24). Importantly, the prognostic value of a high Mito-score remained significant in a multivariate Cox regression model controlling for treatment arm, gender, LDH levels, and white blood cell count (multivariate p-value = 0.017), with borderline significance when accounting for age (multivariate p-value = 0.05). Conclusion: In summary, our comprehensive proteogenomic analysis of NPM1-mutated AML within a prospective clinical trial identifies differentiation trajectories as a key driver of disease heterogeneity. Immature NPM1-mutated AML was characterized by distinct genetic profiles including FLT3-ITD, high BCL2 expression, and a specific mitochondrial protein expression signature (high Mito-score). Clinically, a high Mito-score predicted poorer outcomes in not only the entire cohort but also the favorable NPM1-mutated 2022 ELN subgroup, establishing a link between differentiation states and mitochondrial biology as a relevant feature impacting patient outcome upon intensive chemotherapy treatment.