2026 ASH®
Annual Meeting & Exposition December 12-15, 2026 ▪ New Orleans, Louisiana

Paper Details

Abstracts, authors, affiliations, and disclosures.

Saturday, December 12, 2026
04:00 p.m. - 05:30 p.m.
MCCNO - Great Hall BC
Technology and Procedures, Omics Technologies, Pathology, Bioinformatics, Computational Biology, Profiling, Treatment Considerations, Combination Therapy, Biological Therapies, Non-Biological Therapies, Chemotherapy, Artificial Intelligence (AI), Machine Learning, Biological Processes, Genomics, Multi-Systemic Interactions, Diseases, Myeloid Malignancies, Acute Myeloid Malignancies, Research, Fundamental Science, AML, Translational Research, Study Population, Human, Adult, Pediatric
PAPER

381 The OX40/OX40L axis defines a leukemia-specific immune checkpoint that predicts clinical outcomes in acute myeloid leukemia

Nancy Beryl Janet Arthur, PhD, Esther Sui, MS2, Andrew Engeler3, Tiankai Yin, BA4, Anoushka Kadam5, Sharon Heath6, Marianna Ruzinova, MD7, Michael Prinsen, PhD8, Ma Xenia G Illagan, PhD9, Fan Zhang10, Jing Wang, PhD11, Simona Pagliuca, MD PhD12, David H Spencer, MD, PhD13, Daniel C. Link, MD14, Amanda Cashen, MD15, Geoffrey L. Uy, MD16, Stephen Sykes, PhD17, Francesca Ferraro, MD,PhD18
1 Washington University, Saint Louis, Missouri, United States
2 Washington University, Saint Louis, Missouri, United States
3 Washington University, Saint Louis, Missouri, United States
4 Washington University, Saint Louis, Missouri, United States
5 Washington University, Saint Louis, Missouri, United States
6 Washington University, Saint Louis, Missouri, United States
7 Washington University, Saint Louis, Missouri, United States
8 Washington University, Saint Louis, Missouri, United States
9 Washington University, Saint Louis, Missouri, United States
10 Washington University, Saint Louis, Missouri, United States
11 Washington University, Saint Louis, Missouri, United States
12 CHRU de Nancy, Université de Lorraine, Nancy, France
13 Washington University, Saint Louis, Missouri, United States
14 Washington University, Saint Louis, Missouri, United States
15 Washington University, Saint Louis, Missouri, United States
16 Washington University, Saint Louis, Missouri, United States
17 Washington University, Saint Louis, Missouri, United States
18 Washington University, Saint Louis, Missouri, United States

Clinical data from our group and others have demonstrated that reversible T cell states at the time of AML diagnosis are associated with remission duration post chemotherapy, implicating endogenous immunity as a determinant of disease outcome. Despite increasing evidence that endogenous immunity influences AML outcome, the mechanisms establishing immunosuppressive T cell states at diagnosis remain poorly understood.To address this question, we performed functional profiling of T cell responses in 150 newly diagnosed AML and 10 healthy donor (HD) bone marrow (BM) aspirates. Patient-derived T cells were stimulated through the TCR in an ex vivo co-culture in the presence of AML cells; on day 5, the expression of ICOS and OX40, critical costimulatory receptors mediating T cell activation and expansion, was assessed by flow cytometry. Using HD-derived thresholds, T cells from AML patients segregated into three activation states. 37 patients (25%) displayed OX40-positive T cells (OP; OX40 ≥ 20%), 56 (37%) OX40-negative T cells (ON; ICOS ≥ 20%, OX40 < 20%), and 57 (38%) double-negative (DN; ICOS < 20%, OX40 < 20%).OX40-positive immune activation identified patients at highest risk of treatment failure. Median event-free survival was 7.9 months in OP patients compared with 8.4 months in DN patients, whereas median EFS was not reached in ON patients (p = 0.01). Among patients achieving complete remission, median leukemia-free survival was similarly shortest in OP patients (8.5 months versus 13.2 months for DN and not reached for ON; p = 0.025). After adjustment for age, ELN22 genetic risk, and treatment, OX40 upregulation remained independently associated with inferior overall survival (hazard ratio for death 1.8, p < 0.05).We next asked whether these states reflected functional differences in T cell–leukemia interactions. In ex vivo assays, OP samples displayed impaired T cell proliferation and AML cell killing, suggesting a permissive immune environment supporting leukemia persistence.To define the biological basis of this phenomenon, we performed scRNA-seq on sorted T cells from 11 patients (4 ON, 4 OP, 3 DN) at baseline and post-stimulation (~80,000 T cells). While baseline transcriptomes did not markedly differ between immune states, following T-cell activation, ON T cells efficiently transitioned to cytotoxic effectors (GZMB+, GZMA+, PRF1+), while OP T cells upregulated terminal exhaustion programs (TOX). DN T cells occupied an intermediate state, activating survival pathways but failing to induce cytotoxic effectors.High-dimensional secretome profiling (Olink, 1,037 proteins) was performed on cell culture supernatants collected on day 5 derived from OP, ON, and DN. While all groups displayed evidence of TCR activation, with secretion of IFNG, CXCL9, CXCL10, and IL-22, OP marrow displayed selective failure to induce cytotoxic mediators (GZMA, GZMB, and TNF, adjusted p < 0.05), suggesting that OX40 dominance impairs T cell cytotoxic function.Transcriptomic analysis of existing AML single-cell datasets revealed that OX40L is preferentially upregulated in more primitive, stem-like, leukemic cell clones.CRISPR-mediated deletion of OX40L from AML cell lines consistently restored anti-leukemic T cell cytotoxicity across multiple donors and enhanced T cell activation markers.Collectively, these studies demonstrate that functional T cell activation profiling identifies clinically distinct immune states in AML that predict outcomes beyond current genetic risk models. OX40-dominant activation marks an exhausted, non-cytotoxic T cell program driven, at least in part, by leukemia-derived OX40L. These findings identify the OX40/OX40L axis as a targetable leukemia-specific immune checkpoint with the potential to enhance endogenous immune surveillance and improve responses to conventional therapy.

Nancy Beryl Janet Arthur, PhD: No relevant conflicts of interest to declare.