Abstract 4056: Preclinical evaluation and safety of CBX-250 in acute myeloid leukemia: A bispecific T cell engager targeting Cathepsin-G peptide/HLA complex

Abstract Acute myeloid leukemia (AML) presents a challenge for targeted immunotherapy due to the scarcity of tumor-specific surface antigens. An emerging strategy involves targeting peptide-HLA (pHLA) complexes derived from intracellular proteins. CG1 (FLLPTGAEA), a peptide derived from the Cathepsin G (CG) leader sequence presented by HLA-A*02:01, is selectively expressed on AML cells. CBX-250 has been developed as a TCR-mimetic (TCRm) bispecific T cell engager (TCE) antibody that binds to CG1/HLA-A*02:01 pHLA on AML cells and CD3 on T cells. We previously demonstrated that CBX-250 elicits in vitro killing of leukemia cells at sub-nanomolar EC50 levels. To further evaluate CBX-250 potency, we assessed a panel of CG1-positive U937 cell lines with varying levels of HLA-A*02:01 in T cell cytotoxicity assays. We observed comparable cytotoxicity across cell lines with target copy numbers ranging from 383-7930/cell, and notably, still observed robust killing of tumor cells with 33 target copies/cell. We also found that CBX-250 induced dose-dependent bystander killing of CG1/pHLA-negative cells only in the presence of CG1/pHLA-positive cells, highlighting the beneficial potency effects of CBX-250 in a tumor microenvironment with heterogenous target expression. To evaluate the safety of CBX-250, we assessed CBX-250-induced cytokine release in healthy donor HLA-A*02:01+ PBMCs and whole blood. We found no induction of cytokines in either PBMCs or whole blood when exposed to CBX-250, suggesting the risk for off-target adverse cytokine release is low. Although platelets likely do not present CG1, they can express high levels of class I HLA on their surface and may be potential targets for cross-reactivity with CBX-250. We found no increases in platelet activation following treatment with CBX-250, supporting the favorable safety profile of CBX-250. We assessed CBX-250 specificity by screening a library of >6,500 human proteins for off-target binding partners and found the only significant interactions were with CD3, one of its targets, and FCGR2A, an interaction which likely occurs through the Fc domain due to the high concentrations of CBX-250 tested. No other interactions were detected, indicating high specificity of CBX-250. To evaluate the pharmacology of CBX-250, serum pharmacokinetics (PK) and anti-drug antibody (ADA) formation were evaluated in naive male cynomolgus monkeys. The PK profile for IV and SC administration were consistent with IgG-like molecules, and the majority of the animals exhibited no ADA formation, suggesting minimal immunogenicity risks for CBX-250. Overall, our data demonstrate that CBX-250 is effective against a wide range of CG1/pHLA target densities and capable of mediating bystander killing. These features, combined with its encouraging safety profile, support clinical development of CBX-250. Phase 1 trial (NCT06994676) is ongoing. Citation Format: Jennifer Helble, Geraldine L. Paulus, Preethi Sankaran, Tanzila Rahman, Sarah Jaffe, Nga Sze Amanda Mak, Chunhua Shi, Jun Yan, Timothy Heffernan, Jeffrey Molldrem, Gheath Al-Atrash, Dmitri Wiederschain, Benjamin Lee. Preclinical evaluation and safety of CBX-250 in acute myeloid leukemia: A bispecific T cell engager targeting Cathepsin-G peptide/HLA complex [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 4056.