Document resource
Background Chimeric antigen receptor (CAR) T cell therapy has revolutionized treatment for hematologic malignancies, but response rates vary significantly among patients, even when using standardized products. While tumor-intrinsic and clinical factors have been studied extensively, less is known about how donor-intrinsic T cell biology, particularly proteomic variation, influences therapeutic durability.Methods We performed a longitudinal killing assay using CD8+ CD19-targeted CAR-T cells generated from peripheral blood mononuclear cells (PBMCs) of 100 healthy donors. CAR-T cells were co-cultured with Nalm6 leukemia cells at a 4:1 effector-to-target ratio and imaged every 24 hours via fluorescence microscopy. Every 48 hours, half of the co-culture was transferred onto fresh tumor cells, and re-plating continued for over a month until T cell exhaustion ( figure 1A). Donors were stratified by persistence metrics, and both naïve CD8+ and resulting CAR-T cells were analyzed by mass spectrometry-based proteomics (7400 IDs detected). Protein abundance was correlated with persistence using Spearman correlation.Results Despite identical engineering and stimulation conditions, donors exhibited reproducible variation in CAR-T persistence, ranging from 2 to 32 days ( figure 1B). Proteomic profiles from both naïve and CAR-T cells reflected this divergence. High-persistence donors showed enrichment in stress resilience and biosynthetic readiness pathways, including heat shock responses (FDR = 2.65×10- 1 0 in naïve) and rRNA processing (FDR = 2.36×10- 1 4 in CAR-T), suggesting that durable function may rely on cellular stress management and protein synthesis capacity (figure 2). Low-persistence donors showed enrichment in mitochondrial degradation (FDR = 4.51×10- 1 0 in naïve cells) and cytoskeletal signaling (FDR = 1.03×10- 8 in CAR-T), indicating a predisposition toward metabolic strain and dysregulation. Notably, these patterns were evident in naïve T cells prior to CAR transduction - heat shock pathways (FDR=2.65×10- 1 0) and mitochondrial turnover (FDR = 4.51×10- 1 0) were already significantly enriched - indicating that persistence capacity may be pre-encoded in the donor’s immune baseline. We identified several candidate regulators of persistence, including HSPA1B (R = +0.71 in naïve), FH (R = -0.74 in naïve), DDX49 (R = +0.76 in CAR-T), and CNN2 (R = -0.73 in CAR-T).Conclusions This is the first study to systematically link proteomic programs in both naïve and CAR-T cells to longitudinal persistence, rather than early cytotoxicity alone. Our findings demonstrate that individual differences in T cell proteomic architecture—present before engineering and retained afterward—contribute to variability in CAR-T cell function. Profiling these signatures offers a path to predict product performance and identify novel engineering targets to improve the durability of cell-based immunotherapies.Abstract 392 Figure 1Persistent killing assay schematic and results. A) CD8+ CD19-targeting 28ζ CAR-T cells were co-cultured with Nalm6 cells (4:1 ratio), imaged every 24h, and re-plated every 48h until exhaustion. B) Representative results show persistence ranging from 2 to 32 daysAbstract 392 Figure 2Pathway enrichment analysis of proteins correlated with CAR-T cell persistence. Proteomic-derived protein expressions in naïve CD8+ and CAR-T cells were correlated with killing persistence. Pathway enrichment of strongly correlated proteins (|R| > 0.7) reveal mechanisms linked to persistence variability