1:Hemato-Oncology Program. CIMA Universidad de Navarra. IdiSNA. Pamplona, 31008. Spain; 2:Computational Biology Program. CIMA Universidad de Navarra. IdiSNA. Pamplona, 31008. Spain; 3:Hematology and Hemotherapy Department. Clínica Universidad de Navarra. Pamplona, 31008. Spain; 4:Hematology Service, Complejo Hospitalario de Navarra. IdiSNA. Pamplona, 31008, Spain.; 5:Cell Therapy Area. Clínica Universidad de Navarra. IdiSNA. Pamplona, 31008. Spain.; 6:Immunology and Immunotherapy Department, Clínica Universidad de Navarra. Pamplona, 31008. Spain.; 7:Department of Hematology, Hospital Clinic, IDIBAPS, Barcelona, 08036, Spain; 8:Department of Immunotherapy, Hospital Clinic, IDIBAPS, Barcelona, 08036, Spain; 9:Immunology and Immunotherapy Program. CIMA Universidad de Navarra. IdiSNA. Pamplona, 31008. Spain.; 10:Centro de Investigación Biomédica en Red de Cáncer (CIBERONC). Madrid, 28029. Spain.

Recent studies have shown that control of CAR expression influences CAR-T efficacy. Therefore, we hypothesized that CAR density directly affects CAR-T cell function. In this study we characterized the in vitro and in vivo antitumoral efficacy of FACS-isolated subpopulations of CAR-T cells with different CAR densities targeting BCMA. CARHigh-T cells presented increased basal activation together with more differentiated phenotypes, increased cytotoxicity and cytokine production against multiple myeloma cell lines. This effect was also observed in the infusion products of a clinical trial, where products enriched in CARHigh-T cells presented increased cytotoxicity. After antigen-driven activation, higher presence of terminally differentiated effector cells was observed in CARHigh-T cells, along with increased exhaustion. In vivo, CARLow-T cells presented increased persistence, suggesting that higher CAR levels could reduce long-term efficacy. Transcriptomic analysis revealed completely different expression profiles, with increased tonic signal in CARHigh-T cells. We generated a gene signature associated to CAR density to annotate CARHigh-T cells in scRNA-seq analysis to further infer regulatory dissimilarities driven by different CAR densities, using novel computational methods. Finally, to evaluate the impact of CAR density in the clinical outcome of CAR-T therapies, we applied the inferred gene signature to score infusion products from a clinical trial, observing absence of complete responders in those products enriched on CARHigh signature.

Our data demonstrate that CAR density plays important roles in CAR-T activity with impact on clinical outcome. Comprehension of regulatory mechanisms driven by CAR densities at the single cell level offer an important tool for the development of improved therapies.