Innate & Adaptive Determinants of Post-CAR Cytokine Release Syndrome

About This Grant

ABSTRACT Advances in cell therapies have drastically changed treatment algorithms for hematologic malignancies, broadening curative approaches from the previously singular option of allogeneic hematopoietic cell transplantation to now include the highly effective chimeric antigen receptor (CAR) T cell therapies. Despite the promise of CAR T cell therapies, which harness the patient’s own T cell compartment upon the direction of a transduced antigen-specific receptor, they are still accompanied by significant morbidity. Referred to as “cytokine release syndrome” (CRS) due to rapid increases in multiple cytokines, including IL-6, IFN-γ, IL-2, and TNF, the disordered inflammatory response accompanying CAR T-cell therapies manifests clinically as fever, hypotension, hypoxia, and dyspnea. Overall occurrence of CRS is 93% following CAR T-cell administration (23% severe, with potential for fatal outcome), resulting in use of costly antibody therapeutics and intensive monitoring and translating to a financial burden on the healthcare system. Despite its frequency, CRS lacks a widely accepted definition, a clear molecular mechanism, and reliable markers predictive of its occurrence. Most critically, it is known that the CAR itself is not sufficient to elicit CRS. Our central hypothesis is that the origins of CRS are both effector cell- and environment-intrinsic: critical differences between the CAR effector lymphocytes are what elicit CRS from myeloid cells that have become susceptible in certain patients and that absence of either the stimulus or the response mitigates the initiation and propagation of CRS. Natural killer (NK) cells endowed with the same CAR as T cells do not elicit CRS, despite being capable of achieving complete tumor remission. This crucial difference between the two effector lymphocytes is further highlighted by our own findings that NK-like T cells (TKLR), a lymphocyte population demonstrating features of each, also fail to cause CRS in a mouse model for CAR therapy despite robust anti-tumor efficacy. The overlapping spectrum of these three populations with different CRS outcomes permits a more focused dissection in Aim 1 of the features initiating, propagating or suppressing CRS, including differences in response kinetics. Using in vitro models, we will test combinations of cellular components for their ability to elicit cytokine production and transcriptional changes leading to CRS. CRS-associated cytokine and molecular candidates will then be validated in a murine xenograft CRS model. In addition, the toxicity profile of CAR T cell therapy ranges from no CRS to severe CRS, even when treating the same disease with the same CAR T product. Thus, patient-specific factors, particularly the myeloid cell population critical for the pathological feedback loop of CRS, will be analyzed. In Aim 2, we will identify the pre-treatment risk signatures (phenotypic functional, and transcriptional) in the myeloid cell population in matched CAR T recipients who did and did not develop CRS. This comprehensive profiling will reveal novel points of intervention and prevention for CRS and ultimately lead CAR therapies to greater safety.