Adoptive T-cell therapy targeting a novel non-canonical neoantigen derived from a translated AML1-ETO lncrna shows clinical activity in a first-in-human trial for t(8;21) AML

Abstract Background: Acute myeloid leukemia (AML) with the t(8;21)(q22;q22) translocation, a hallmark of core-binding factor AML (CBF-AML), generates the AML1-ETO (AE) fusion gene. Despite being classified as favorable-risk, this subtype exhibits a high relapse rate and poor outcomes in refractory/relapsed (R/R) patients. While the tumor-specific AML1-ETO fusion is an ideal immunotherapeutic target, its potential to generate clinically relevant neoantigens, particularly from non-canonical translational events, remains underexplored. We sought to identify novel public AML1-ETO-derived neoantigens and evaluate their therapeutic potential in a first-in-human adoptive T-cell therapy trial. Methods: Potential neoantigens from RNA-seq data of t(8;21) AML cell lines and primary patient samples were predicted using the GIANT neoantigen prediction platform. The lead candidate was validated for HLA binding using biolayer interferometry, immunogenicity through tetramer staining and ELISPOT assays with healthy donor and patient cells, and specific cytotoxicity via in vitro LDH release assays. Anti-leukemic activity was assessed in murine xenograft models. A phase 1 clinical trial (NCT06499025) was initiated to evaluate the safety and preliminary efficacy of adoptively transferred autologous or haploidentical neoantigen-specific cytotoxic T-lymphocytes (CTLs) in patients with R/R t(8;21) AML. Results: Typical AML1-ETO transcripts did not generate immunogenic neoantigens. Significantly, we report the discovery of AML1-ETO CRF_f, a novel translated lncRNA transcript we identified and named, characterized by retention of ETO intron 4. Remarkably, this exclusively expressed transcript (found in patient blasts and LSCs but completely absent in normal tissues) produces public, highly immunogenic non-canonical neoantigens that are uniformly detected in all tested samples. Notably, an HLA-A*11:01-restricted neoantigen, Peptide_X, induced robust expansion of specific CD8+ T cells from both healthy donors and patients. These Peptide_X -specific CTLs demonstrated potent, HLA-A*11:01-restricted lysis of AML1-ETO+ target cells and significantly prolonged survival in a murine xenograft model. Three patients with R/R t(8;21) AML received Peptide_X -specific CTLs, whether autologous or haploidentical donor-derived, infusions. The therapy was well-tolerated with no dose-limiting toxicities. In a patient with central nervous system (CNS) leukemia, intrathecal CTL administration led to sustained blast clearance in the cerebrospinal fluid (CSF) for over two months. In a haploidentical recipient, CTL infusion resulted in reduction in bone marrow AML1-ETO transcripts, which correlated with donor T-cell chimerism. A subsequent 10-fold higher CTL dose induced a deeper molecular response and a higher peak chimerism. Single-cell RNA and TCR sequencing of infused CTLs identified clonally expanded, high-avidity Peptide_X-specific TCRs, providing a blueprint for next-generation engineered TCR-T cell therapies. Conclusion: The AML1-ETO fusion locus generates a clinically actionable, non-canonical neoantigen through the translation of a lncRNA transcript. This study uncovers a novel mechanism of neoantigen generation and provides the first preclinical and clinical evidence of safety and anti-leukemic activity for an adoptive T-cell therapy targeting a fusion gene-derived neoantigen in AML, offering a new precision immunotherapy for this disease.