Comprehensive profiling of AML and T-cell states across disease stages in acute myeloid leukemia using spectral flow cytometry Open Access

Acute myeloid leukemia (AML) is a heterogeneous hematologic malignancy characterized by clonal expansion of myeloid precursors, with variability in its clinical presentation, molecular landscape, and disease progression.¹ Although AML pathogenesis has been extensively studied at the genomic level,^2-4 the interplay between leukemic and nonleukemic immune cells across different disease stages remains incompletely understood. Recent advances in high-dimensional cytometry and single-cell analyses offer deeper insights into cellular hierarchies and immune microenvironments in AML.^5-8 

To mend this gap, we applied spectral flow cytometry and high-dimensional computational analyses to characterize cellular subsets in peripheral blood mononuclear cells (PBMCs) from patients with AML at 3 disease stages: newly diagnosed (NewDx), remission, and relapsed/refractory (R/R). We profiled major leukemic and immune cell populations, identifying stage-specific enrichment patterns in both myeloid and lymphoid compartments. We also explored T-cell subset differentiation and exhaustion, focusing on their roles in disease progression and immune dysregulation. Finally, we analyzed mutation-specific enrichment patterns in myeloid and lymphoid clusters, revealing associations between genetic alterations and shifts in immune cell populations across disease stages. These findings provide insights into AML's cellular ecosystem and potential therapeutic targets.

This study complies with the Declaration of Helsinki principles. A written informed consent human material use was approved by the institutional review board.

Cryopreserved PBMCs were stained with viability dye, Fc receptor blockers, and antibody panels before analysis on a Cytek Aurora spectral flow cytometer; further details are provided in the supplemental Methods.

We analyzed 88 PBMC samples from 72 patients with AML and 2 healthy donors. This cohort included 41 samples from NewDx patients, 12 samples from patients in remission (paired with NewDx samples), and 33 samples from patients with R/R AML (2 of whom had paired NewDx samples; supplemental Table 1).

To delineate AML hierarchies unbiasedly and identify major non-AML companion cell populations, we used an in-house optimized 32-antibody panel for spectral flow cytometry (referred to here as the AML panel; supplemental Table 2; Figure 1A). Initial clustering analysis identified both myeloid and lymphoid clusters (Figure 1B). Subsequent reclustering resolved these into 7 myeloid and 8 lymphoid clusters (Figure 1C-D), categorized by marker expression (Figure 1E). We then estimated the proportions of each cluster across different disease states (supplemental Figure 1B), expectedly finding enrichment of leukemic clusters (CD34⁺, CD34⁺CD90⁺, CD71⁺, CD38⁺, and CD123⁺) in NewDx patients, with even higher proportions in R/R patients. Conversely, healthy and remission samples were comparable, with remission samples showing enrichment of the CD45^low cluster and a reduction in the B-cell cluster (Figure 1F).

We clustered the samples by myeloid cluster proportions, revealing 6 distinct clusters (CM1-CM6; Figure 1G). NewDx and R/R samples generally clustered similarly, except for the CM6 cluster, which uniquely contained R/R samples enriched in CD123⁺ cells. Remission samples primarily fell within the CM2 cluster, characterized by monocytic cells, whereas the remaining remission samples fell within CM3 and CM4, enriched in CD45^low and CD71⁺ clusters, respectively (Figure 1H). Analysis of overall survival among NewDx samples revealed no significant differences between the myeloid clusters (supplemental Figure 1C).

Similarly, lymphoid cluster proportions allowed us to identify 6 sample clusters (CL1-CL6; Figure 1I). NewDx and R/R samples clustered similarly, with the exception that NewDx samples lacked the CL3 cluster (enriched in naive CD8⁺ T cells). None of the remission samples clustered in CL2, which was enriched in B cells, suggesting some degree of peripheral B-cell aplasia in remission (Figure 1J). Again, no significant differences in overall survival were found among the lymphoid clusters (supplemental Figure 1D).

Lastly, we examined associations between the myeloid and lymphoid clusters, finding associations between CM5 and CL5 (characterized by a higher frequency of CD34⁺ myeloid cells alongside naive CD4⁺ T cells) as well as between CM3 and CL6 (characterized by increased CD45^low myeloid cells along with CD56^dim natural killer [NK] cells; Figure 1K).

Of the 88 PBMC samples analyzed in the AML panel, we conducted further staining on 70 samples from 54 patients with AML and 2 healthy donors using Human T Cell Differentiation and Exhaustion panel (BioLegend), a 27-color panel designed to characterize T-cell subsets (supplemental Table 3). These 70 samples included 35 from NewDx patients, 12 from patients in remission (paired with NewDx samples), and 21 from R/R patients with AML (2 of whom had paired NewDx samples; Figure 2A; supplemental Table 4).

Initial clustering of T cells based on canonical markers (CD4, CD8, TCRVα7.2, and TCRγ/δ) identified major T-cell populations, including CD4⁺ T cells, CD8⁺ T cells, mucosal associate invariant T cells, γδT cells, double negative T cells, and double positive T cells (Figure 2B). Further subclustering revealed 7 CD4⁺ T-cell clusters and 9 CD8⁺ T-cell clusters, each categorized by marker expression (Figure 2C-E; supplemental Table 5).

We then examined the proportions of each cluster across disease states (supplemental Figure 2B). R/R samples showed enrichment of a CD8⁺ T-cell cluster with markers of terminal effector and senescent state (CD57⁺/KLRG1⁺/PD-1⁺/CTL4⁺/LAG3⁺; termed CD8⁺ T.Term/senL) compared to NewDx and remission samples, whereas R/R and remission samples had lower proportions of naive CD8⁺ T cells (CD45RA⁺/CCR7⁺/CD95^–; termed CD8⁺ T.n) compared to NewDx patients (Figure 2F). These observations are consistent with mean fluorescence intensity comparisons across disease states for both CD8⁺ and CD4⁺ T cells (supplemental Figure 3; supplemental Table 6). Subsequent analysis of sample clustering by T-cell cluster proportions identified 9 distinct clusters (CT1-CT9; Figure 2G). CT2, enriched in CD8⁺ T.Term/senL cells, was predominantly observed in R/R samples, whereas CT1, enriched in CD4⁺ T.n (CD45RA⁺/CCR7⁺/CD95^–) and CD4⁺ T effector memory (T.em) cells (CD45RA^–/CCR7^–/CD95⁺), was more frequent in NewDx patients (Figure 2H). Survival analysis among NewDx samples revealed no significant differences between T-cell clusters (supplemental Figure 2C). Crossanalysis was also conducted between myeloid clusters identified in the AML panel and T-cell clusters (supplemental Figure 2D), as well as between lymphoid clusters and T-cell clusters (supplemental Figure 2E).

In our cohort of AML samples, the most frequently mutated genes were NRAS (27%), DNMT3A (22%), and TP53 (21%), alongside several other genes commonly mutated in AML. Comutations frequently involved DNMT3A, NPM1, IDH2, and FLT3 (Figure 3A). To explore the impact of mutation status on cell cluster enrichment, we examined cluster distributions across disease stages by mutation status (Figure 3B). Samples with TP53 mutations demonstrated an increased proportion of CD71⁺ AML cells at both diagnosis and relapse (Figure 3B-C), suggestive of erythroid differentiation, and exhibited higher levels of CD8⁺ memory T cells at relapse (Figure 3D).

Distinct enrichment patterns were associated with certain AML mutations. For example, patients with DNMT3A, NPM1, IDH2, and FLT3 mutations had reduced levels of CD34⁺ and CD34⁺CD90⁺ AML clusters at diagnosis (Figure 3B,E). At relapse, these samples showed increased proportions of CD123⁺ AML cells (Figure 3F), and there was an increased CD45^low cells during remission (Figure 3G). In addition, R/R NPM1-mutated samples had increased B cells and reduced memory T cells (Figure 3H-J). At remission, NPM1-mutated samples had less CD56^bright NK cells, whereas KRAS/NRAS-mutated samples showed reduced CD56^dim NK cells (Figure 3K-L).

Samples stained with the T-cell panel had a mutation distribution similar to that of the AML panel cohort (Figure 3M). We further investigated the effect of mutation status on T-cell cluster enrichment across disease stages (Figure 3N). In KRAS/NRAS-mutated samples, we observed a reduction in CD4⁺ T.n cells in the R/R stage and an enrichment of CD4⁺ T.em cells during remission (Figure 3O-P). In TP53-mutated samples, CD4⁺ T.n cells were enriched at the NewDx stage, and CD8⁺T.em.CD73⁺ cells were enriched at the R/R stage (Figure 3Q-R). Notably, RUNX1-mutated samples had fewer effector CD8⁺T cells and an increase in CD4⁺T_reg at the R/R stage (Figure 3S-V).

This study highlights dynamic shifts in myeloid and lymphoid populations at different stages of AML. At diagnosis and in R/R patients, we observed an enrichment of leukemic CD34⁺ and CD123⁺ cells, whereas remission samples showed increased monocytic and CD45^low populations. T-cell profiling revealed an enrichment of CD8⁺ T.Term/SenL cells in R/R patients, suggesting that T-cell senescence contributes to relapse and resistance, because these senescent cells are associated with impaired cytotoxic function and poor survival outcomes.⁹ 

Our findings align with those of Mazziotta et al,⁵ who used flow cytometry and single-cell RNA sequencing to link CD8⁺ T-cell senescence to therapy resistance in AML. Herbrich et al,¹⁰ using cytometry by time of flight (CyTOF), also identified T-cell dynamics as predictive of treatment response, reinforcing the importance of T-cell compartment shifts in shaping therapeutic outcomes. Badar et al¹¹ used CyTOF to profile T-cell clusters, showing strong concordance between bone marrow and peripheral blood, supporting PBMCs for immune profiling. Although their study pooled samples across disease stages, our study specifically examined TP53-associated differences in peripheral blood at defined disease states using spectral flow cytometry. Both studies found increased CD8⁺ T cells in TP53-mutant AML, with our unique finding of enrichment at the R/R stage marked by a CD8⁺T.em.CD73⁺ phenotype.

These results support the potential of immune profiling for predicting disease progression and resistance, offering insights for targeted therapies in AML.

This study was supported by a grant from the Ladies Leukemia League (H.A.A.). P.K.R. received salary support from the Lymphoma Research Foundation. This research was performed in the Flow Cytometry and Cellular Imaging Core Facility, which is supported in part by the National Institutes of Health, National Cancer Institute through MD Anderson's Cancer Center Support Grant P30 CA016672.

The funding agencies had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Contribution: M.Y.Y., H.A.A., and P.K.R. had full access to all the data in the study and take responsibility for the integrity of the data and verification of the accuracy of the data analysis; H.A.A. and P.K.R. contributed to conception and design; M.Y.Y., J.L.R., H.A.A., and P.K.R. contributed to collection, assembly, analysis, and interpretation of data; and all authors contributed to manuscript writing, provided final approval of the manuscript, and are accountable for all aspects of the work.

Conflict-of-interest disclosure: F.S. reports consulting for AstraZeneca, Amgen, Revolution Medicines, Novartis, BridgeBio, BeiGene, BerGenBio, Guardant Health, Calithera Biosciences, Tango Therapeutics, Hookipa Pharma, Novocure, Merck Sharp & Dohme, and Roche; grant or research support from Amgen, Mirati Therapeutics, Revolution Medicines, Pfizer, Novartis, and Merck & Co; having stocks in BioNTech and Moderna; and honoraria from the European Society for Medical Oncology (ESMO), Japanese Lung Cancer Society, Medscape, Intellisphere, Visiting Speakers Program in Oncology (VSPO) McGill Universite de Montreal, RV Mais Promocao Eventos, MJH Life Sciences, IDEOlogy Health, Moving Innovation & Technology (MI&T), Physicians' Education Resource (PER), CURIO, DAVA Oncology, the American Association for Cancer Research, and the International Association for the Study of Lung Cancer. H.A.A. reports honoraria from Illumina and Alamar Biotechnology; in-kind support from Illumina; research support from Genentech, Enzyme by Design, GlaxoSmithKline, Blueprint Medicines, Ascentage, and Illumina; advisory board fees from Cogent Biosciences; and consultancy fees from Molecular Partners. The remaining authors declare no competing financial interests.

Correspondence: Hussein A. Abbas, Department of Leukemia, The University of Texas MD Anderson Cancer Center, 1500 Holcombe Blvd, Houston, TX 77030; email: [email protected]; and Patrick K. Reville, Department of Leukemia, The University of Texas MD Anderson Cancer Center, 1500 Holcombe Blvd, Houston, TX 77030; email: [email protected].