Introduction: Lymphodepletion (LD) is an integral component of anti-CD19 chimeric antigen receptor T cell (CART19) immunotherapy, creating the proper environment and cytokine milieu for CART engraftment and function. We previously demonstrated that bendamustine (Benda) LD is as effective as standard fludarabine and cyclophosphamide (Flu/Cy) LD prior to 4-1BB-costimulated tisagenlecleucel but is characterized by significantly reduced cytokine-release syndrome (CRS), neurotoxicity (ICANS), and hematological toxicities. However, whether Benda LD is also effective and safe in CD28 costimulated CART19, and the mechanism for the reduced toxicities associated with Benda LD remain unknown. Methods: We retrospectively evaluated the outcomes of 59 consecutive non-Hodgkin lymphoma patients (pts), including large B cell lymphomas (LBCL, n=48) and follicular lymphoma (FL, n=11), treated with commercial CD28 costimulated axicabtagene ciloleucel (axi-cel) at the University of Pennsylvania between January 2018 and March 2023. Data cut-off was June 30 th, 2023. LD regimens were Flu 30 mg/m 2/day and Cy 500 mg/m 2/day x 3 days or Benda 90 mg/m 2/day x 2 days. LD choice was at the treating physician's discretion, partially determined by Flu shortage. Pts were evaluated for 3-month response (Lugano criteria), survival, and toxicities (ASTCT, CTCAE) in the 30 days after axi-cel infusion. For hematological toxicities, we analyzed lymphocyte, neutrophil, platelet, and hemoglobin levels from the first day of LD to 30 days after axi-cel infusion. For long-term hematological toxicities, we retrieved blood values at 6 months in pts with ongoing remission. Moreover, we analyzed serum samples collected pre- and post-LD (CART19 infusion day) for changes in circulating cytokines (Luminex) and metabolites (Liquid Chromatography/Mass Spectrometry) in a second cohort of 32 pts undergoing CART19. Results: Of 59 pts, 37 received Flu/Cy while 22 Benda LD. The two groups were balanced for most clinical features. Responses with Flu/Cy and Benda LD were similar, with complete remission rates of 51.4% and 50.0%, respectively. Progression-free and overall survivals were similar in the two LD groups. However, any-grade CRS occurred in 91.9% of pts who received Flu/Cy versus 72.7% of pts who received Benda LD; any-grade ICANS after Flu/Cy occurred in 43.2% and after Benda LD in 18.2% of pts ( Fig.1A). These results were confirmed in both LBCL and FL cohorts. Both LD regimens induced profound lymphopenia by the time of axi-cel infusion. However, Flu/Cy LD induced a deeper reduction in median lymphocyte counts (Flu/Cy: 0.00x10 9/L; Benda, 0.20x10 9/L). Additionally, Flu/Cy was associated with a higher incidence of neutropenia of grade ≥3 (100.0% vs 54.5%). Consequently, Flu/Cy pts were more likely to develop infections (78.4% vs. 27.3%), and neutropenic fever (78.4% vs. 13.6%) than Benda receiving pts. By evaluating long-term hematological toxicities, 75.0% of Flu/Cy pts did not have lymphocyte recovery to normal values at 6 months compared to 37.5% with Benda LD. Mechanistically, both Flu/Cy (n=7) and Benda (n=25) LD regimens induced an increase in the levels of cytokines facilitating CART proliferation (IL2, IL7, IL15). Overall, Flu/Cy LD pts had a higher increase in inflammatory cytokine levels compared to Benda LD pts, including a higher increase in the levels of cytokines previously associated with CRS and ICANS (IL15, IL8, and MCP1, Fig.1B). Moreover, after LD, several indispensable anabolic metabolites, including carnitine esters, were decreased in pts receiving Flu/Cy. We also found that NADH and nicotinamide riboside, critical elements for cellular redox balance, were reduced in pts exposed to Flu/Cy LD. We speculate that the higher cytotoxicity/cytopenias, induced by Flu/Cy LD, causes a faster utilization of the above-mentioned metabolites in order to reconstitute the hematopoietic cell compartment to physiological levels, resulting in a higher patient susceptibility to additional metabolic stress. Conclusions: In this retrospective comparison,Benda LD is a safer alternative to Flu/Cy for CART19 immunotherapy with similar efficacy and significantly reduced toxicities. Our data can serve as guidance to clinicians for the selection of the best regimen for lymphodepletion. Our conclusions are supported by our mechanistic studies showing reduced inflammatory cytokines after Benda LD.
Read full abstract