Immunologic and Brain Bases of Neurotoxicity Associated with CAR T-Cell Therapy

Introduction: CD19-directed chimeric antigen receptor (CAR)-T cell therapy has revolutionized the treatment for B-cell acute lymphoblastic leukemia (ALL), inducing complete remission in 70-90% of highly refractory patients. Unfortunately, CAR-T cell therapy is associated with immune effector cell-associated neurotoxicity syndrome (ICANS) in 40% of patients, which presents as confusion, memory impairment, aphasia, altered level of consciousness and/or delirium. The neurobiological and immune bases of ICANS remain poorly understood, and identifying them will help develop strategies for early intervention to mitigate ICANS. Methods: From an ongoing prospective, longitudinal study, we report on the first 7 patients with relapsed ALL (age 5-19 years, 5 males) who underwent comprehensive assessments at baseline before lympho-depleting (LD) chemotherapy (fludarabine and cyclophosphamide) and infusion of CD19-directed CAR T-cells, and post-infusion, on days 10 and 28. These assessments included multimodal magnetic resonance imaging (MRI) of the brain (anatomical, perfusion, diffusion tensor imaging and MR spectroscopy), neuropsychological testing, serum measures of blood-brain barrier permeability and cytokines, and immune cell profiles in peripheral blood using mass cytometry by time-of-flight (CyTOF). We also acquired multimodal MRI and neuropsychological testing in 9 healthy controls (age 5-20 years, 7 males), once, at baseline. MRI data were acquired on a 3T Philips MR scanner using a 32 channel head coil without the use of sedatives or contrast agents and were processed blind to the diagnosis and order of data acquisition. Results: In our cohort of 7 patients, 5 developed cytokine release syndrome and 4 developed ICANS. Brain imaging findings are demonstrated in figures 1 (cortical thickness) and 2 (blood perfusion). Prior to CAR T-cell infusion patients relative to healthy controls had thinner cortex bilaterally across large portions of the brain. By post-infusion day 10, cortical thickness decreased in the postcentral gyrus (PoG) and posterior temporal lobe (pTL). Subsequently, by post-infusion day 28, cortex thickened in the PoG but continued thinning in the pTL. Patients relative to controls had lower regional cerebral blood flow (rCBF) in the thalamus, insular cortex, and cingulate gyrus, and higher rCBF in the genu of the corpus callosum (gCC). By post-infusion day 10, rCBF increased in the thalamus and decreased in the putamen and gCC. By post-infusion day 28, rCBF in the thalamus decreased and reverted towards baseline values, whereas rCBF in gCC continued to decrease away from baseline values. CyTOF analyses, as expected, showed expansion of monocytes and myeloid-derived suppressor cells by post-infusion day 10. Furthermore, in these immune cells, phosphorylation of key signaling proteins (e.g. extracellular-signal regulated kinase, ERK1/2) increased in response to external stimulation. Conclusions: Patients had widespread brain abnormalities at baseline compared to controls, likely because of prior therapies and/or underlying ALL. Post LD chemotherapy and CAR T-cell infusion, several brain abnormalities worsened by day 10 and then reverted toward baseline values by day 28. These temporal changes are generally consistent with clinically reported development of ICANS by day 10, which largely resolves by day 28. However, in certain regions brain measures progressively deviated from baseline and control values through day 28, suggesting some abnormalities associated with CAR T therapy may persist. Preliminary immune cell profiling implicates monocytes as a possible source of elevated cytokines following CAR T-cell infusion. Analysis for correlation of these findings with symptomatic ICANS and neuropsychological measures is ongoing. Patients on this study will also be followed at 6 months and 1 year to evaluate for long term neurological effects of CAR T-cell therapy. Figure 1: Cortical Thickness abnormalities in patients relative to healthy controls (1st panel); within-patient changes post-infusion (2nd, 3rd, & 4th panel). Blue & violet show decreases and orange & red show increases in cortical thickness. Figure 2: Blood perfusion abnormalities in patients relative to healthy controls (1st panel); within patient changes post infusion (2nd, 3rd, & 4th panel). Blue & violet show decreases and orange & red show increases in regional cerebral blood flow. Disclosures Pulsipher: Novartis: Consultancy, Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Jazz: Other: Education for employees; Adaptive: Membership on an entity's Board of Directors or advisory committees, Research Funding; Miltenyi: Research Funding; Bellicum: Consultancy; Amgen: Other: Lecture; CSL Behring: Membership on an entity's Board of Directors or advisory committees; Medac: Honoraria. Wayne:Spectrum Pharmaceuticals: Consultancy, Research Funding; AbbVie: Consultancy; Kite, a Gilead Company: Consultancy, Research Funding; Servier: Consultancy.