Metabolomic dynamics of cytokine release syndrome in CD19 CAR-T therapy

Abstract Purpose: Cytokine release syndrome (CRS) after CD19 CAR-T therapy poses a significant risk of morbidity and stands as a formidable obstacle to the success of CD19 CAR-T therapy. Despite the well-established influence of metabolism on immune responses against infectious diseases, its role in CRS during CAR-T therapy is not fully explored. Methods: We performed untargeted LC-MS-based metabolomic profiling on sequential serum samples from 36 acute lymphoblastic leukemia (ALL) patients at distinct post-infusion timepoints: pre-CRS (CRS_before), CRS peak (CRS_middle), and CRS recovery (CRS_after). A total of 108 samples were analyzed. Results: Differentially expressed metabolites (DEMs) were firstly investigated across different timepoints. At CRS_middle (vs. CRS_before), 52 metabolites were upregulated and 12 downregulated, predominantly enriched in caffeine metabolism and ascorbate and aldarate metabolism. Post-CRS recovery (CRS_after vs. CRS_middle) showed 20 upregulated and 55 downregulated metabolites. Upregulated species included amino acids (lysine, L-threonine) and lipids (glycerophosphocholine, Lyso PE). Co-expression clustering revealed metabolites enriched in glycine, serine, leucine, and isoleucine biosynthesis, phenylalanine metabolism, and pantothenate and CoA biosynthesis gradually decreased over time. Comparing CRS grade 1 vs. 2: 15, 44, and 38 DEMs were identified at CRS_before, CRS_middle, and CRS_after, respectively. Receiver operating characteristic (ROC) analysis identified six plasma metabolites distinguishing CRS grade 3 from grade 1 patients with strong discriminatory power: upregulated metabolisms ( 2',3'-Dideoxy-3'-fluorouridine, PC (24:0/20:3(6,8,11)-OH(5), glutaric acid) and downregulated metabolisms (Acetamide, 1-Amino-2-sulfanylethanesulfinic acid, Tyr-Gly-Gly-Trp-Leu). Conclusion: This study provides comprehensive insights into the dynamic remodeling of the plasma metabolome during CD19 CAR-T therapy and reveals metabolic signatures associated with CRS severity in ALL patients. These findings enhance our understanding of metabolic dynamics in CAR-T therapy and may inform strategies for predicting and managing CRS-related complications.