Chimeric antigen receptors have different downstream signal transduction compared to T cell receptors that may be tunable for optimal efficacy Title: Chimeric antigen receptors have different downstream signal transduction compared to T cell receptors that may be tunable for optimal efficacy Authors: Katherine J Carpenter, PhD 1; Jeremy Bjelajac 1,2; Ashley Wilson, PhD 1; Joshua Gustafson, PhD 1; James Matthaei, PhD 1,3; Michael Jensen, MD 1,4,5; Stephen EP Smith, PhD 1,5 Affiliations: 1 Seattle Children’s Research Institute 2 Stanford University School of Medicine 3 Sonoma Biotherapeutics 4 Fred Hutchinson Cancer Research Center 5 University of Washington School of Medicine Abstract: Chimeric antigen receptors (CARs) are synthetic constructs containing extracellular antigen-specific scFv (single-chain variable immunoglobulin) fragments linked to intracellular signaling domains. These intracellular domains most commonly include CD3z subunits and either CD28 or 4-1BB domains. This complex differs drastically from the endogenous T cell receptor (TCR), which is comprised of 2 TCR chains and 6 intracellular CD3 chains. TCR signaling is a well characterized molecular circuit, but the CAR signal transduction network (STN) has not been well studied. Another unique aspect of CAR design is antigen binding affinity. CARs have been designed to behind strongly to antigen (Kd 106-109 M-1) compared to endogenous TCRs (104-106 M-1). Despite their different components, when CARs are activated by antigen they are able to stimulate potent anti-tumor T cell responses. CAR T cell therapy has been shown to be effective at killing tumors in vivo, particularly in the case of B cell lymphomas and leukemias. These CARs are the epitome of personalized medicine as they are created specifically for each patient’s individual needs. However, as with many other immuno-oncology approaches, CARs have been shown to cause negative side effects such as cytokine release syndrome (CRS) and neurotoxicity (NT). We hypothesize differential STNs downstream of CAR activation promote either optimal (tumor clearance) or negative (CRS/NT) outcomes in patients. We stimulated human clinical-grade anti-CD19 CAR T cells with antigen presenting cells expressing either anti-CD3 (TCR stimulation) or CD19 (CAR stimulation). We then utilized Quantitative Multiplex co-Immunoprecipitation (QMI) to analyze protein-protein interactions downstream of either TCR or CAR engagement. Using weighted correlation network analysis (CNA), we found 2 distinct modules correlating with CD3 or CD19 stimulation. We also used anti-fluorescein CARs with different ligand affinities (low, medium, or high) to assess maximum responses. We found, surprisingly, that medium affinity anti-fluorescein CARs had the highest response intensity compared to low and high affinity CARs. Lastly, we produced CARs from 3 different healthy donors and stimulated them with CD19. Through Hierarchical Clustering by Principal Components (HCPC) we saw distinct clustering of Donor 3 compared to Donor 1 and 2. This leads us to believe there is detectable inter-individual differences in STNs between CARs from different donors. Using these results, we believe CAR functionality could be optimized by tuning scFv affinity and modifying downstream STNs to promote clearance of tumors potentially without causing negative side effects.