Pre-existing biochemical entities predict severity of cytokine release syndrome in B-cell lymphoma patients treated with CD19-directed CAR T-cells

Abstract

Abstract Potentially severe immune-related adverse events (irAEs) develop in relapsed/refractory large B cell lymphoma (R/R LBCL) patients during treatment with anti-CD19 CAR T-cell therapy. Therefore, prediction and better management of toxicities is critically required to improve patient outcomes. Cytokine release syndrome (CRS), one of the most notable acute irAEs, is the result of an inflammatory response of activated CAR T-cells and myeloid cells. Interestingly, high levels of glucose and low levels of certain amino acids are associated with pro-inflammatory immune activation in infectious and autoimmune diseases, but such associations have not been investigated in CAR T-cell therapy. Therefore, we employed mass spectrometry-based untargeted and targeted metabolomics analysis to identify pre-treatment host biological metabolites that predict rapid onset and severe CRS in 41 R/R LBCL patients treated with Axicabtagene Ciloleucel or Tisagenlecleucel. Analysis of plasma metabolites revealed significant associations of CRS with several metabolic classes, including carbohydrates, lipid, amino acids, dipeptides and nucleotides. Interestingly, high levels of the carbohydrates, glucose, mannose and 1,5-anhydroglucitol were observed in patients who experienced more rapid onset of CRS. In contrast, patients with greater severity of CRS had lower plasma abundance of the amino acids, glycine, proline and glutamate. Our data suggest that specific pre-existing plasma metabolites have the potential to serve as predictive biomarkers of CRS in patients undergoing anti-CD19 CAR T-cell therapy, and may be used for risk stratification and clinical management of irAEs in advance of treatment.