Combinatorial Targeting of Multiple Shared Antigens By Adapter-CAR-T Cells (aCAR-Ts) Allows Target Cell Discrimination and Specific Lysis Based on Differential Expression Profiles

Despite tremendous clinical success of chimeric antigen receptor (CAR) expressing T cell (CAR-T) therapies, targeting B-phenotypic antigens in ALL, CLL, NHL or multiple myeloma, there are still major limitations for broader clinical application. CAR-Ts are capable to generate a specific immune response against defined surface-expressed antigens leading to sustained depletion of target antigen expressing tissues e.g. B cells. While B cell function can be substituted by repetitive IgG infusions, prolonged depletion of vitally essential tissues is not compatible with life. In AML for instance, most promising target antigens are expressed along myeloid lineage differentiation, limiting the therapeutic applicability. Therefore, CAR-Ts targeting essential shared antigens mustallow tight regulation of CAR-T function and/or be able to differentiate between cancerous and healthy tissue.

To address these issues, we have developed the adapter CAR-T cell (aCAR-T) system. By splitting antigen recognition and CAR-T activation, introducing adapter molecules (AMs), the system allows precise quantitative (on-/off-switch) as well as qualitative (change and combine target antigens) regulation of CAR-T function. aCAR-Ts are based on the unique properties of a novel scFv targeting a "neo"-epitope-like structure consisting of the endogenous vitamin biotin in the context of a specific linker, referred to as linker-label-epitope (LLE). LLEs can be easily conjugatedtonovel or preexisting AM formats like monoclonal antibodies (mAbs) or mAb fragments in a GMP-compliant manner. We were able to demonstrate that aCAR-Ts allow simultaneous targeting of various antigens ("OR"-gate),preventing antigen evasion by selection of antigen or epitope-loss variants.

In the present study we intended to investigate whether aCAR-Ts are capable to identify and differentiate target cells due toversatileantigenexpression profiles ("AND"-gate). In theory, AMs against different target antigens can be assembled on the surface of a target cell,leading to aCAR-T activation independent of the targeted antigens (surface painting),by binding to the presented LLE-tags. Therefore, combinatorial AMs treatment might allow to translate complex and multiple antigen-dependent targetcellidentification into an aCAR-T activation.

To test this hypothesis, we have generated LLE-AMs against ALL/NHL-(CD10, CD19, CD20, CD22, CD37, CD138, ROR1) and AML-(CD32, CD33, CD38, CD123, CD135, CD305, CLL1) associated antigens. Individual threshold concentrations for aCAR-T activation by different AMs, targeting the model cell lines Nalm6 (ALL), JeKo1 (NHL), HL-60,U973andMolm13 (all AML), have been analyzed. Cutoffswerefound to be between 10 and 100 pg/ml, dependent on target expression and target cell line. Importantly, combinations of 2, 3 or 5 AMs, targeting different antigens expressed on thesame target cell, cause target-cell lysis at concentrations below the activation threshold for single AMs(exemplified for HL-60 in Figure A). Our results clearly demonstrate an additive effect in combining different AMs to hurdle the activation threshold.Moreover, in a JeKo1 CD19 and/or CD20 knock out (KO) antigen-loss model, combinations of AMs targeting CD19, CD20 and ROR1 can differentiate between wild type and antigen -1 (CD19 or CD20 KO) or antigen -2 (CD19 and CD20 KO) variants in mediating target-cell lysis, even though at least one target antigen is expressed. Finally, we found that combinations of CD10, CD19, CD22 and CD138 sufficiently eliminate Nalm-6 BCP-ALL cells, while sparing healthy B cells in co-culture experiments. Similar results were obtained in co-culture experiments of HL-60 AML cells with monocytes, neutrophils as well asCD34-enriched hematopoietic progenitor cells, applying combinations of CD32, CD33, CD38, CD123, CD305 and CLL1. Co-culturing experiments using autologous blasts, monocytes, neutrophils and aCAR-Ts are ongoing.

Together, our results indicate that aCAR-Ts in combination with selected AM combinations might have the ability to identify and specifically eliminate cancer cells based on complex antigen expression profiles. This would have major implications for clinical translation, enabling combinatorial therapy, essential to avoid antigen evasion, and the possibility to spare vitally essential tissue from elimination.

Figure.

View largeDownload slide

Figure.

View largeDownload slide

Close modal