Radia M. M. Khan

BACKGROUND: Limited durability of clinical responses remains a major challenge in chimeric antigen receptor (CAR)-T therapy. CAR-enhancers (CAR-Es), which fuse tumor antigens to interleukin (IL)-2 muteins, provide a targeted strategy to enhance CAR-T persistence and function. It remained unclear whether CAR-Es are effective across distinct tumor contexts, when using patient-derived T cells, or in preventing exhaustion and sustaining persistence. It was also unknown whether CAR-Es can selectively expand CAR-Ts in humanized mice with pre-existing T cells, and to what extent their efficacy depends on IL-2Rβγ vs IL-2Rα engagement. While IL-2Rα (CD25) has been classically linked to potent antitumor responses and memory formation, it also drives IL-2-associated toxicities, including vascular leak and preferential regulatory T cell expansion. METHODS: We systematically dissected CAR-E signaling requirements by engineering IL-2 variants with selective receptor affinities. Multiple CAR-E constructs were developed and tested across a range of in vitro and in vivo models. RESULTS: We demonstrate that CAR-E activity is entirely independent of IL-2Rα and critically dependent on IL-2Rβγ signaling. A next-generation IL-2Rα-sparing CAR-E maintained full potency, driving robust CAR-T expansion, persistence, and tumor clearance, even at low doses and when using CAR-T cells derived from previously treated multiple myeloma patients. These CAR-T cells not only resisted exhaustion but also re-expanded months later to eradicate tumor rechallenges. In humanized mice with pre-established T cells, CAR-Es selectively expanded CAR-Ts to dominate the circulating T-cell pool. CAR-E exerted a dominant influence on CAR-T fate, overriding tumor-derived cues and enforcing consistent phenotypes across diverse preclinical models. CONCLUSIONS: These findings nominate a lead B-cell maturation antigen (BCMA)-IL-2 CAR-E candidate with strong translational potential for clinical development and establish IL-2Rβγ as a key driver of CAR-E activity. The results also identify IL-2Rα as dispensable and provide a mechanistic framework for designing safer, IL-2Rα-sparing CAR-Es.

Abstract BCMA-directed CAR T cell therapies have improved outcomes for multiple myeloma (MM), and several other targeted CAR T cell therapies are in development. However, one of the main hurdles for further progress is the limited availability of tumor-specific cell surface targets. To take advantage of a very large number of tumor-specific intracellular targets, we here describe the development of T cell receptor (TCR)-like antibodies that can recognize intracellular targets. Unlike conventional CAR T cells, CAR T cells utilizing TCR-like antibodies can recognize intracellular tumor antigens presented in the context of MHC molecules, mimicking TCR, and significantly expanding the range of potential cancer targets. We have previously identified MZB1 (marginal zone B and B1 cell-specific protein) as a tumor-associated intracellular antigen in MM and other B-cell malignancies and predicted putative MZB1 peptides that can be presented in the context of HLA-A*02:01, the most frequent HLA allele worldwide. Using a human scFv antibody yeast surface display library, we identified scFv clones specific to the MZB1/HLA-A2 complex. After confirming MZB1- and HLA-A2-specific binding, we generated MZB1 CAR T cells using the MZB1-A2 scFv. MZB1 CAR T cells exhibit significant activity and specificity against MZB1 and HLA-A2 positive MM cell lines and primary MM cells, but not against HLA-A2neg and/or MZB1neg cells, in a dose-dependent manner in vitro. In vitro killing assays showed MZB1 CAR T cells had significant cytotoxicity against MZB1+ and HLA-A2+ cells and secreted significantly higher levels of IFN-γ, sFas ligand, and granzyme A/B compared to untransduced T cells when co-cultured with MZB1+ HLA-A2+ MM cells. Importantly, we also identified high expression of MZB1 in Waldenström Macroglobulinemia (WM) cells from both MYD88 mutated and non-mutated WM patient cells and MZB1 CAR-T cells showed significant cytotoxicity against MZB1+ A2+ WM cells both in vitro and in vivo, including MZB1+ CD19+ A2+ primary WM cells. Furthermore, we observed cross-reactivity of our antibody as well as MZB1 CAR T cell product, with other HLA-A alleles, such as HLA-A*24:02 or HLA-A*23:01, which are more common in African Americans and Asian Americans, respectively, suggesting potential broader applicability of our antibody across patient groups. To further increase persistence and expansion of our CAR T cells, we took advantage of our recently developed novel CAR enhancer (CAR-E) molecule comprising of a BCMA antigen fused to a low-affinity interleukin 2 (IL-2) able to induce IL-2 signaling upon binding to the BCMA CAR T cells leading to increased CAR T cell proliferation, antitumor efficacy and development of memory phenotype (Rakhshandehroo et al., Nature Biotechnology, 2025).We generated bispecific MZB1/BCMA CAR T cells that express a modified BCMA CAR to serve as a binder for the CAR-E but would not be associated with known toxicities of BCMA as a CAR target. This construct allows CAR-T cells to expand upon CAR-E treatment in vivo after CAR T infusion. To test the potential expansion and cytotoxic capacity of this MZB1/BCMA CAR E, we administered a very low number (0.25 x 106 CAR T cells per mice – usual is 10 x 106) of either MZB1 or bispecific MZB1/BCMA CAR E T cells. We administered CAR enhancer protein the following day to induce expansion of the CAR T cells in vivo. MZB1/BCMA CAR E T cells receiving CAR-E treatment showed superior clearance of all MM tumor cells in vivo, and high levels of circulating CAR T cells detected beyond 2 months from injection compared to BCMA CAR T cell alone or MZB1 CAR T cell alone were found. These cells were enriched in CD8+ T cells and displayed an acquired memory phenotype. To further evaluate the long-term activity of persistent low level CAR T cells, a second rechallenge with CAR E infusion was carried out at day 80 after CAR T cell administration: only MZB1/BCMA CAR E T cells in the mice that received CAR-E treatment retained the ability to re-expand upon restimulation, potentially confirming our ability to expand CAR T cells in vivo even at a late stage. In conclusion, this study provides the rationale to evaluate MZB1 TCR-like CAR T cell therapy and also report development of a novel CAR E as a therapeutic approach in MM, WM and other B-cell malignancies and provide the logic and framework for similar therapies to be developed against other intracellular antigens in MM and other cancers.