Cancer genetics has changed AML care. Doctors already use mutations in genes like FLT3, IDH1, IDH2, NPM1, and CEBPA to help estimate risk and choose therapies. The 2022 European LeukemiaNet recommendations reflect how deeply genetics now shapes AML diagnosis and treatment planning [2].
But genes are only the recipe card. Proteins are the meal. Metabolites are the crumbs on the counter that reveal who ate at 2 a.m.
This study found that AML subtypes can look clearer when researchers integrate many “omics” layers instead of staring at DNA alone. Some AML samples showed metabolic rewiring linked to MYC and mTOR activity, two pathways that help control growth and cellular fuel use. If cancer cells are the cellular rebels, MYC and mTOR are the managers yelling, “More growth! More supplies! Nobody takes lunch!”
The researchers also found striking mitochondrial protein hyperacetylation in CEBPA-mutant AML. That is a mouthful, so let’s unpack it. Mitochondria are the cell’s energy machinery. Acetylation is a chemical tag that can change how proteins behave. Hyperacetylation means those tags are showing up in unusual abundance, like every machine in the factory suddenly got covered in sticky notes. That could point to a metabolic weakness, though it needs more testing before anyone starts naming it “Operation Sticky Note.”
The Protein Layer Had Plot Twists
One of the more intriguing findings involved NPM1-mutant AML, a common genetic subtype. The team identified a protein-centered subset with unusually high expression of FOXC1 and HOXB8/9. These genes are tied to developmental programs, the kind cells usually use when deciding what to become when they grow up.
AML blasts, being dramatic, often refuse to grow up. They are basically immortal interns blocking the bone marrow hallway.
Previous work has shown that proteomic subtyping can reveal AML groups that genetics alone misses. A 2022 Cancer Cell study used bone marrow biopsies from 252 AML patients and identified proteogenomic subtypes with distinct biology [3]. More recently, studies have connected proteogenomic patterns to drug response and resistance, including work in Cell Reports Medicine showing that protein and phosphoprotein data can improve predictions of how AML samples respond to therapies outside the body [4].
That is the larger promise here: not just labeling AML more elegantly, but making the label useful.
The Drug Resistance Clue
The study also used machine learning to nominate therapy targets across AML subtypes. One candidate was MTA1, which the authors validated as a contributor to resistance to panobinostat, a histone deacetylase inhibitor [1].
This does not mean panobinostat plus an MTA1 strategy is ready for clinic on Monday. Biology loves to humble us, usually before coffee. But it does suggest a practical path: if a leukemia resists a drug, multiomic profiling may help explain why, and maybe reveal how to push back.
Drug resistance remains one of the hardest problems in AML. Even with newer targeted drugs, many responses do not last. Reviews and clinical perspectives keep circling the same painful reality: AML can adapt, relapse, and force patients through exhausting rounds of uncertainty [5]. Better molecular maps could help clinicians avoid treatments unlikely to work and steer patients toward trials that match the leukemia’s actual wiring.
What This Could Mean, If It Holds Up
The most patient-centered version of this future is simple: fewer guesses.
A person newly diagnosed with AML might someday have testing that does not stop at mutation panels. Their care team could see the leukemia’s genetic drivers, protein behavior, fuel preferences, resistance signals, and possible drug vulnerabilities in one integrated report. Not a crystal ball. More like a better flashlight in a very dark basement.
There are still hurdles. Multiomic testing is technically complex, expensive, and not yet routine. Findings from 173 treatment-naive patients need validation in larger, diverse groups. Doctors also need proof that acting on these data improves survival, quality of life, or both. A beautiful molecular map is nice. A patient getting more good days because of it is the actual win.
Still, this study nudges AML research in a useful direction. It treats leukemia less like a single villain and more like a messy, adaptive system. Annoying? Yes. Scientifically rich? Also yes. Cancer biology is rude like that.
References
-
Chu SCA, Hsiao Y, Wang C, et al. Integrated proteogenomic and metabolomic profiling of acute myeloid leukemias to identify molecular subtypes and associated therapy targets. Nature Cancer. 2026. https://doi.org/10.1038/s43018-026-01175-6
-
Dohner H, Wei AH, Appelbaum FR, et al. Diagnosis and management of AML in adults: 2022 recommendations from an international expert panel on behalf of the ELN. Blood. 2022;140(12):1345-1377. https://doi.org/10.1182/blood.2022016867
-
Jayavelu AK, Wolf S, Buettner F, et al. The proteogenomic subtypes of acute myeloid leukemia. Cancer Cell. 2022;40(3):301-317.e12. https://doi.org/10.1016/j.ccell.2022.02.006
-
Pino JC, Posso C, Joshi SK, et al. Mapping the proteogenomic landscape enables prediction of drug response in acute myeloid leukemia. Cell Reports Medicine. 2024;5(1):101359. https://doi.org/10.1016/j.xcrm.2023.101359
-
Leppa AM, Grimes K, Jeong H, et al. Single-cell multiomics analysis reveals dynamic clonal evolution and targetable phenotypes in acute myeloid leukemia with complex karyotype. Nature Genetics. 2024;56:2790-2803. https://doi.org/10.1038/s41588-024-01999-x
Disclaimer: The image accompanying this article is for illustrative purposes only and does not depict actual experimental results, data, or biological mechanisms.