Microbiota-derived 1,5-pentanediamine attenuates CD19 CAR-T function

Abstract

Abstract Background: Previous studies indicate that bacteremia may compromise CAR-T cell therapy efficacy. However, it remains unclear whether antibiotic-resistant gut bacteria can remotely regulate CAR-T cell function in the absence of bacteremia. Our clinical data show that patients colonized or infected with carbapenem-resistant organisms (CRO) exhibit significantly lower progression-free survival (PFS) and overall survival (OS) after CAR-T therapy, implicating microbiota-derived factors in CAR-T cell suppression. We selected Carbapenem-resistant Klebsiella pneumoniae (CRKP), a prevalent bloodstream pathogen with high mortality and rising resistance, as the model strain. Untargeted metabolomics revealed elevated levels of 1,5-pentanediamine (PDA) in CRKP secretions, suggesting that this metabolite might systemically inhibit CAR-T cells independently of bacteremia. Objective: This study evaluates the immunosuppressive effects of CRKP-secreted PDA on CD19 CAR-T cells, focusing on proliferation, apoptosis, activation/exhaustion markers, and effector functions. We also investigate the underlying mechanisms via RNA sequencing (RNA-seq) and measure serum PDA levels in patients with Carbapenem-resistant Enterobacterales (CRE) colonization but no bacteremia, exploring PDA's role as a microbial-derived systemic immune regulator. Methods: Human CD19 CAR-T cells were pretreated with PDA (3–12 mM, 24–72 h), with untreated cells serving as controls. Proliferation, apoptosis, activation/exhaustion markers (CD25, CD69, PD-1, TIM-3, LAG-3), and Treg proportions were assessed. CAR-T cells were co-cultured with NALM-6 cells to measure cytotoxicity and CD107a degranulation, while cytokines were quantified using a Th1/Th2 assay kit. RNA-seq compared gene expression in PDA-treated (9 mM, 48 h) and control CAR-T cells. Serum PDA levels in CRE-colonized, non-bacteremic patients were measured via LC-MS/MS. All experiments included matched controls, and statistical analyses were performed at predefined intervals. Results: PDA suppressed CAR-T cell function across multiple time points and concentrations. At 48 hours, the proliferation IC₅₀ was 7.40 mM (95% CI: 6.43–8.34 mM). Under primary analysis conditions (9 mM, 48 h), proliferation decreased by 71.00% (P<0.01), and apoptosis rose from 7.19% to 22.40% (P<0.05). Co-culture with NALM-6 cells (72 h, E:T 3:1) reduced cytotoxicity from 79.30% to 18.11% (P<0.001), while CD107a expression increased from 61.65% to 74.15% (P<0.01). Activation markers CD25 and CD69 increased 1.03- and 3.63-fold (P<0.05), respectively. Exhaustion markers rose significantly: PD-1⁺ (17.10% to 26.63%; P<0.05), TIM-3⁺ (17.36% to 38.62%; P<0.01), and LAG-3⁺ (10.67% to 18.55%; P<0.05). Treg proportions increased from 12.38% to 26.27% (P<0.01). Cytokine analysis showed reductions in IFN-γ (83.20%; P<0.01), TNF-α (32.80%; P<0.05), and IL-2 (51.40%; P<0.01), but no significant changes in IL-4, IL-6, or IL-10 (P>0.05). RNA-seq identified upregulated immunosuppression/exhaustion genes (SLC7A5, SOCS1, TNFRSF9, TIGAR, CD276) and downregulated cytotoxicity (GZMK) and antigen presentation (HLA-E) genes. PDA was detected in the serum of CRE-colonized patients (0.02–0.19 μM). Conclusion: CRE colonization systemically impairs CAR-T cell function via PDA, even in the absence of bacteremia, by elevating Treg proportions, reducing effector cytokines, and sustaining exhaustion/activation markers. PDA may drive terminal exhaustion through chronic overstimulation and disrupted immune feedback. RNA-seq links PDA to metabolic and checkpoint dysregulation, explaining T cell dysfunction. Clinically detectable PDA supports its role as a microbial-derived immune regulator, providing a mechanistic basis for reduced CAR-T efficacy in CRE-colonized patients and a potential therapeutic target.