Antigen heterogeneity and the emergence of antigen-negative tumor cells are major impediments to long-term efficacy of chimeric antigen receptor T-cell therapies in B-lineage malignancies. Traditional single-antigen CAR designs can exert strong selective pressure, favoring outgrowth of tumor cells that lack the targeted antigen and contributing to relapse. Addressing pre-existing heterogeneity across B-cell antigens is therefore a priority for improving durable responses.
The study evaluates the Adapter CAR-T (AdCAR-T) approach, which decouples antigen recognition from the engineered T-cell effector by using biotinylated adapter molecules (AMs). These AMs bind defined surface antigens on tumor cells and are recognized by AdCAR-T cells, enabling flexible and modular redirection of T-cell cytotoxicity to different targets without reengineering the T-cell receptor itself.
Adapter molecules tested in this work were generated from three antibody sources: an in-house-produced tafasitamab biosimilar directed at CD19, commercial rituximab directed at CD20, and an in-house-produced daratumumab biosimilar directed at CD38. Each AM mediated antigen-specific AdCAR-T cell cytotoxicity in the experiments reported.
Using a defined heterogeneous Burkitt lymphoma model, the investigators examined how targeting single antigens versus combinations affected control of mixed tumor populations. When AdCAR-T cells were redirected with a single AM, selective pressure led to emergence or selection of antigen-negative tumor populations within the model, demonstrating the limitation of monovalent targeting in heterogeneous settings.
In contrast, simultaneous application of AMs against CD19, CD20, and CD38 effectively controlled the antigen-heterogeneous tumor population in vitro. Combinatorial redirection prevented the outgrowth of tumor subsets lacking any one of the targeted antigens, indicating broader coverage of the B-lineage antigenic landscape than single-antigen approaches.
The study reports that combinatorial targeting of CD19, CD20, and CD38 via AdCAR-T plus corresponding AMs induced sustained tumor control in an in vivo model. Sustained control in vivo suggests that simultaneous multi-antigen targeting can translate from controlled in vitro settings into improved antitumor activity in an organismal context.
Specific quantitative outcomes, such as survival curves, tumor burden measurements, dosing regimens, or statistical analyses, were not detailed in the abstract and therefore are not reported here.
Exposure to the CD38-directed adapter molecule resulted in selective loss of the CD38-positive AdCAR-T cell subset. The authors interpret this observation as consistent with T-cell fratricide, a recognized risk when targeting antigens that are also expressed on effector T cells. Despite this loss, a surviving CD38-low AdCAR-T population retained cytotoxic activity against tumor targets.
The persistence of cytotoxic function in the CD38-low fraction suggests that complete elimination of all AM-directed T cells did not occur, and that residual T-cell populations can contribute to tumor control. Details on the frequency of fratricide, kinetics of population changes, or functional assays quantifying residual activity were not provided in the abstract.
These results position combinatorial AdCAR-T targeting as a flexible pan-B-lineage strategy to mitigate pre-existing antigen heterogeneity in B-cell malignancies. By using antibody-derived AM combinations, the platform allows modular, simultaneous engagement of multiple canonical B-lineage antigens (CD19, CD20, CD38) without needing to generate separate CAR constructs for each target.
The finding that CD38-directed AMs can induce fratricide highlights a practical consideration for selecting antigen combinations and designing adapter dosing schedules. Optimizing AM combinations, sequences, and concentrations could reduce self-targeting while maintaining broad tumor coverage.
The authors conclude that the data support further development of antibody-derived AM combinations for B-cell malignancies. Translation to clinical settings will require detailed characterization of safety, dosing, potential fratricide mitigation, and demonstration of durable benefit in more extensive preclinical and clinical studies.
Research support was provided to University Children’s Hospital Tuebingen by Miltenyi Biotec GmbH under a collaborative research agreement. One author (CMS) receives research funding from Miltenyi Biotec unrelated to the present work and is a co-inventor on a patent concerning Adapter CAR technology (WO2018078066A1). The other authors declared no competing interests in the abstract. Specific grant numbers or detailed funder roles beyond what is reported in the abstract were not provided.