CAR-T cell therapy has transformed the treatment landscape for certain blood cancers. The approach involves harvesting a patient’s own T cells, genetically engineering them to express a synthetic receptor that recognizes CD19, a protein found on the surface of most B cell malignancies, and reinfusing them after lymphodepleting chemotherapy. Despite dramatic remission rates in clinical trials, the therapy does not work for everyone, and even among initial responders, relapse remains common. Immunologists have attributed this heterogeneity to factors such as tumor burden, prior treatment lines, T cell fitness, and the immunosuppressive tumor microenvironment. The new study adds a previously underappreciated variable to that list: the metabolic products of drug-resistant bacteria colonizing the gut and other mucosal surfaces.

Carbapenem-resistant Klebsiella pneumoniae, or CRKP, is one of the most feared pathogens in modern hospitals, classified by the World Health Organization as a critical-priority pathogen for which new treatments are urgently needed. Patients with hematologic malignancies who have undergone intensive chemotherapy, stem cell transplantation, or prolonged antibiotic exposure are particularly susceptible to CRKP colonization, in which the bacterium establishes itself in the body without necessarily causing an overt bloodstream infection. The research team, led by corresponding authors Xiaojian Zhu and Yi Xiao, focused on 1,5-pentanediamine, a diamine metabolite produced by several members of the Enterobacteriaceae family, including Klebsiella species, through the decarboxylation of lysine.

Using liquid chromatography–tandem mass spectrometry, an analytical technique capable of detecting and quantifying small molecules with high sensitivity and specificity, the researchers measured serum PDA concentrations in 30 CRKP-colonized hematology patients who had no documented bloodstream infection at the time of sampling. The metabolite was detectable in the serum of all 30 patients. The authors are careful to note an important caveat: because the study lacked a non-colonized comparator cohort, these findings demonstrate that PDA is present in the circulation of colonized patients but do not prove that CRKP is the exclusive source of the metabolite.

With detectable PDA levels established in the clinical population, the team turned to the central question of the study: what happens to CAR-T cells when they are exposed to this metabolite? Healthy-donor-derived CD19 CAR-T cells were cultured with PDA at concentrations ranging from 0 to 12 millimolar, with 9 millimolar used for most functional assays. These millimolar concentrations reflect the acute exposure levels achievable in vitro and are considerably higher than the trace serum levels measured in patients, a point the researchers acknowledge when discussing the physiological relevance of their findings.

The results were striking. PDA exposure reduced the metabolic activity of CAR-T cells, as measured by assays of cellular respiration and energy production, and increased apoptosis, the programmed cell death pathway that determines how long engineered T cells survive in circulation. Since CAR-T persistence correlates strongly with durable clinical responses, any insult that shortens the lifespan of these cells could directly undermine therapeutic efficacy. Beyond survival, PDA-treated cells showed a shift in their immunological identity. The metabolite upregulated both activation markers and checkpoint-associated inhibitory molecules—the same brakes that tumors exploit to disable T cells—and altered the balance between CD4 helper and CD8 cytotoxic subsets. Most tellingly, the proportion of regulatory T cells, an immunosuppressive population that dampens antitumor immunity, increased in the presence of PDA.

Functional testing reinforced the picture of a compromised therapeutic product. When PDA-treated CAR-T cells were confronted with NALM-6 cells, a well-established B cell leukemia line used as a standard CD19-positive target, their killing capacity dropped significantly. The cells also produced lower amounts of inflammatory cytokines such as interferon-gamma, which recruits and activates other arms of the immune system, and released reduced levels of perforin and granzyme B, the cytotoxic molecules that CAR-T cells use to punch holes in tumor cells and trigger their self-destruction. Intriguingly, one measure of immune engagement was spared: CD107a degranulation, a marker of the physical process by which T cells release their toxic granules, remained intact. This dissociation—cells that can still fire their weapons but do so with less lethality and less inflammatory support—suggests that PDA does not simply shut CAR-T cells down but pushes them into a subtle, dysfunctional state.

To understand the molecular basis of this dysfunction, the researchers performed RNA sequencing on PDA-exposed CAR-T cells, a technique that catalogs the activity of thousands of genes simultaneously. The transcriptomic profiles revealed enrichment of pathways governing the cell cycle, apoptosis, and stress responses, alongside a suppression of immune signaling pathways. The gene-expression signature bore hallmarks of T cell exhaustion, the hypo-responsive state familiar from chronic viral infections and tumors. Quantitative reverse-transcription PCR confirmed key transcriptional changes at the individual gene level.

One of the most clinically consequential findings involved immune checkpoint blockade. Because PDA upregulated checkpoint-associated inhibitory markers, the researchers tested whether blocking PD-1, the receptor targeted by some of the most widely used cancer immunotherapies, could rescue the metabolite-impaired cells. Under the conditions tested, PD-1 blockade alone failed to restore CAR-T function. This result implies that the damage inflicted by the metabolite extends beyond a single checkpoint axis and may involve broader metabolic and transcriptional reprogramming that checkpoint inhibitors cannot readily reverse.

The authors are explicit about the limitations of their work. The experiments relied on acute exposure of healthy-donor-derived CAR-T cells to millimolar PDA concentrations in vitro, whereas patients are likely exposed to lower metabolite levels over longer periods, in a body shaped by infection, inflammation, and prior therapies. Serum PDA was measured in only a single cohort without controls, and the killing assays used a single target-cell line. Validation in chronic low-dose exposure models, controlled clinical cohorts comparing colonized and non-colonized patients, patient-derived CAR-T cells, and additional tumor targets will be essential before these findings can inform clinical practice.

Even with those caveats, the study opens an unexpected frontier at the intersection of microbiology, metabolism, and cellular immunotherapy. If drug-resistant bacterial colonization can chemically undermine engineered immune cells, then screening patients for CRKP colonization, quantifying bacterial metabolites before cell infusion, or intervening with decolonization strategies, adsorbents, or metabolic inhibitors might one day become part of standard CAR-T preparation. The work also carries broader implications for the growing recognition that microbiota-derived metabolites—molecules once dismissed as inert waste products of bacterial metabolism—can act as systemic immunomodulators with the power to shape the success or failure of cutting-edge cancer treatments.