CAR T-cells can persist for remarkably different lengths of time depending on how they are built, the cancer they are fighting, and the patient receiving them. In the best-documented cases, engineered T-cells have been found circulating in the blood more than a decade after a single infusion, still keeping leukemia in complete remission. But that kind of durability is not guaranteed, and many patients see their CAR T-cells decline within weeks to months. The gap between those outcomes comes down to a web of interconnected factors, from the molecular architecture of the receptor itself to the health of the patient’s immune system before treatment even begins.
The Decade-Long Cases That Redefined Expectations
The longest-documented persistence of CAR T-cells comes from two patients with chronic lymphocytic leukemia who were treated with CD19-targeting CAR T-cells in 2010. Both achieved complete remission, and researchers continued tracking their engineered cells for over ten years. CAR T-cells remained detectable in the blood throughout that entire period, and deep molecular sequencing confirmed that both patients stayed in sustained remission with no detectable cancer.1PubMed Central. Decade-long leukaemia remissions with persistence of CD4+ CAR T cells2PubMed Central. A decade of CD4+ chimeric antigen receptor T-cell evolution in two chronic lymphocytic leukemia patients: were chronic lymphocytic leukemia cells present? A separate report on a B-cell lymphoma patient also showed CAR T-cell persistence spanning a decade, with the engineered cells remaining functionally active as evidenced by ongoing suppression of normal B-cells.3Nature Medicine. Decade-long persistence of CD19 CAR T cells in B cell lymphomas
These cases are extraordinary but not typical. In many blood cancers treated with CAR T-cells, the engineered cells expand rapidly in the first two weeks, peak around day 10 to 14, and then contract. Some patients retain a low but measurable population for months or years; others lose detectable CAR T-cells within a few weeks. What separates the long-lasting cases from the short-lived ones is the central question driving a huge amount of current research.
How the CAR Itself Is Built Matters Enormously
The receptor bolted onto the T-cell’s surface is not a single monolithic design. One of the most consequential choices in CAR engineering is the costimulatory domain, an internal signaling component that tells the cell what to do after it recognizes a tumor target. The two most widely used options are CD28 and 4-1BB, and they produce strikingly different persistence profiles.
CAR T-cells built with 4-1BB costimulation consistently stick around longer than those built with CD28. In animal models, 4-1BB-costimulated cells survive and expand while CD28-based cells clear quickly.4PubMed Central. 4-1BB costimulation promotes CAR T cell survival through noncanonical NF-κB signaling The difference traces to distinct survival signals: 4-1BB activates a pathway that promotes longer cell life even when the CD28-based CAR is also firing, suggesting these are genuinely separate survival mechanisms rather than one being a weaker version of the other. Researchers have pinpointed part of the CD28 problem to a single amino acid that drives T-cell exhaustion. Swapping that one residue, changing an asparagine to a phenylalanine, rescued CD28-based CAR T-cells and allowed them to persist long enough to maintain durable tumor control in animal models.5The Journal of Clinical Investigation. Single residue in CD28-costimulated CAR-T cells limits long-term persistence and antitumor durability
Beyond the costimulatory domain, tonic signaling is another design-level factor. Some CARs spontaneously cluster on the T-cell surface even when no tumor target is present, sending a constant low-grade activation signal. This background noise pushes the cells toward exhaustion before they ever encounter cancer. The clustering appears to be driven by positively charged patches on the CAR’s antigen-binding surface, and reducing that charge density can dial down the unwanted signaling.6Cell Research. Tuning charge density of chimeric antigen receptor optimizes tonic signaling and CAR-T cell fitness Stronger tonic signaling leads directly to reduced persistence and weaker tumor-killing activity once cells are infused.7PubMed. The persistence and antitumor efficacy of CAR-T cells are modulated by tonic signaling within the CDR
The Starting Material Problem
CAR T-cells are manufactured from the patient’s own blood, which means whatever shape the patient’s immune system is in at the time of collection directly affects the product’s potential. This is a bigger problem than it might seem, because most patients receiving CAR T-cell therapy have already been through multiple rounds of chemotherapy. That prior treatment leaves a mark: patients who have had more chemotherapy show a shift in their T-cell makeup, with fewer naive T-cells and a higher proportion of differentiated effector memory cells.8PubMed Central. Prior chemotherapy deteriorates T-cell quality for CAR T-cell therapy in B-cell non-Hodgkin’s lymphoma Naive and early-memory T-cells are the raw material most likely to produce long-lasting CAR T-cells. When the starting pool has been depleted by prior treatment, the resulting product tilts toward short-lived cells that expand aggressively but burn out fast.
The specific composition of the infused product turns out to matter in a granular way. Formulating products with a defined ratio of CD4 and CD8 T-cells, rather than using whatever mix happens to come from the patient, dramatically improved potency in preclinical work. A balanced 1:1 ratio of CD4 and CD8 CAR T-cells resulted in complete tumor eradication at doses that were ineffective when either subset was used alone, with the CD4 cells appearing to sustain the CD8 cells through cytokine support.9PubMed Central. Chimeric antigen receptor-modified T cells derived from defined CD8+ and CD4+ subsets confer superior antitumor reactivity in vivo Modeling work has confirmed that CD8 CAR T-cells expand faster than CD4 cells in the body, but that the interplay between the two populations is what sustains the overall response over time.10PubMed Central. Development of a multiscale mechanistic modeling framework integrating differential cellular kinetics of CAR T-cell subsets and immunophenotypes in cancer patients
Within those broad categories, the presence of stem-like memory T-cells in the infused product is an emerging predictor of success. Products enriched in these progenitor-like cells are associated with better disease control because they can self-renew and generate fresh waves of functional effector cells. When these stem-like cells are scarce, as was observed in one clinical trial of Lewis Y-targeting CAR T-cells, the infused cells show poor persistence in patients.11PubMed. TSTEM-like CAR-T cells exhibit improved persistence and tumor control compared with conventional CAR-T cells in preclinical models
Exhaustion Is the Central Threat to Persistence
T-cell exhaustion is the single biggest biological obstacle to long-lived CAR T-cells. When T-cells are continuously stimulated, whether by ongoing tumor exposure or tonic signaling from the CAR itself, they progressively lose function. They produce fewer cancer-killing molecules, proliferate less, and eventually enter a state where they are alive but largely inert. This is not just a temporary dip. Exhaustion is driven by deep epigenetic changes, alterations to how genes are turned on and off, that become increasingly locked in over time.12PubMed Central. Transient rest restores functionality in exhausted CAR-T cells through epigenetic remodeling
One encouraging finding is that exhaustion can be at least partially reversed if caught early enough. In animal and lab models, giving CAR T-cells a transient rest, a brief pause from receptor signaling, triggers epigenetic remodeling that restores some functional capacity. The caveat is that the longer exhaustion has been in progress, the more the epigenetic changes become permanent. There seems to be a window during which intervention can rescue the cells, and a point of no return after which it cannot.
The Patient’s Body Can Reject CAR T-Cells
An underappreciated factor in persistence is the patient’s own immune system attacking the CAR T-cells. Most commercially available CAR T-cell products use an antibody-derived targeting fragment that comes from mouse proteins. The human immune system can recognize these as foreign and mount an antibody response against the CAR itself. When that happens, the patient’s immune system clears the therapeutic cells, raising the chance of relapse and reducing the effectiveness of any re-dosing attempts.13PubMed Central. Immunogenicity of CAR-T Cell Therapeutics: Evidence, Mechanism and Mitigation
Humanized CARs, in which the mouse-derived portions are re-engineered to look more like human proteins, are one attempt to solve this. In one trial testing humanized CD19-targeting CAR T-cells, there was no significant increase in anti-mouse antibodies after infusion, suggesting the humanized design successfully evaded this particular immune response.14Journal for ImmunoTherapy of Cancer. Safety and efficacy of autologous and allogeneic humanized CD19-targeted CAR-T cell therapy for patients with relapsed/refractory B-ALL Whether this translates to meaningfully longer persistence in large patient populations is still being studied.
The Cytokine Environment After Infusion
What is happening inside the patient’s body after CAR T-cells are infused matters independently of the product itself. One key variable is IL-15, a cytokine that supports T-cell survival and proliferation. Patients with higher levels of IL-15 in their blood after conditioning chemotherapy show substantially greater CAR T-cell expansion and persistence. The peak IL-15 concentration was the strongest independent predictor of overall CAR T-cell exposure over the first 90 days, outperforming pre-treatment levels or other time-point measurements.15Blood. High IL-15 Serum Concentrations Are Associated with Response to CD19 CAR T-Cell Therapy and Robust In Vivo CAR T-Cell Kinetics This is one reason that the conditioning chemotherapy given before CAR T-cell infusion is considered so important: it creates a hospitable environment by depleting competing immune cells and triggering a surge in homeostatic cytokines like IL-15.
Why Solid Tumors Are a Different Problem
Almost everything discussed so far applies primarily to blood cancers, where CAR T-cells have had their greatest success. In solid tumors, persistence faces an entirely additional layer of challenges. The tumor microenvironment in solid cancers is actively immunosuppressive: it contains molecules that shut down T-cell activity, physical barriers that limit T-cell infiltration, and metabolic conditions like low oxygen and scarce nutrients that are hostile to T-cell survival.16PubMed Central. Advancing CAR-T Therapy for Solid Tumors: From Barriers to Clinical Progress
Antigen heterogeneity compounds the problem. In many solid tumors, not every cancer cell expresses the target that the CAR T-cell is looking for. The cells that do express the target get killed, but the cells that do not survive and regrow. Meanwhile, without persistent antigen stimulation, the CAR T-cells that were supposed to provide ongoing surveillance lose their reason to stick around. This dynamic has played out in multiple myeloma trials targeting BCMA: while some patients achieve sustained responses lasting over a year, most eventually relapse, which may relate to loss of CAR T-cells, loss of antigen on the tumor surface, or the immunosuppressive microenvironment impairing T-cell function.17The Lancet Haematology. Immunotherapy with chimeric antigen receptor T cells in multiple myeloma
Engineering Strategies to Extend Persistence
Given these challenges, researchers are pursuing multiple engineering approaches to make CAR T-cells last longer. One of the most promising is building CAR T-cells that secrete their own supportive cytokines, sometimes called armored CARs. These cells are engineered to produce molecules like IL-12, CD40L, or 4-1BBL that help them survive in hostile environments and resist immunosuppressive signals from the tumor.18PubMed Central. Armored CAR T-cells: utilizing cytokines and pro-inflammatory ligands to enhance CAR T-cell anti-tumour efficacy
IL-15 co-expression has shown particular promise. CAR T-cells engineered to produce IL-15 alongside their tumor-targeting receptor showed significantly greater expansion in patients with solid cancers, achieving a disease control rate of about two-thirds and an antitumor response rate of a third in one cohort.19PubMed Central. Interleukin-15-armored GPC3-CAR T cells for patients with solid cancers In a glioma model, IL-15-expressing CAR T-cells had greater proliferative capacity, survived longer, and produced more cancer-killing molecules than standard CAR T-cells, with no sign of uncontrolled growth in the absence of tumor targets.20Cancer Immunology Research. Transgenic Expression of IL15 Improves Antiglioma Activity of IL13Rα2-CAR T Cells but Results in Antigen Loss Variants
Another avenue is combining CAR T-cell therapy with checkpoint inhibitors, the antibody drugs that block immune-suppressing signals like PD-1. In preclinical models, PD-1 blocking antibodies rescued exhausted CAR T-cells, restoring their ability to accumulate, kill tumor cells, and produce inflammatory cytokines. The marked tumor reduction following antibody treatment confirmed that PD-1 signaling is a real mechanism of CAR T-cell inhibition and that blocking it can reinvigorate cells that have started to lose steam.21The Journal of Clinical Investigation. Human CAR T cells with cell-intrinsic PD-1 checkpoint blockade resist tumor-mediated inhibition
Gene editing is adding yet another dimension. Disrupting TET2, a gene that regulates some of the epigenetic changes associated with exhaustion, has been shown to restrain the progression toward terminal exhaustion and enhance antitumor responses. Researchers have developed clinically actionable versions of this approach by knocking in a safety switch alongside the CAR at a specific genomic location, disrupting TET2 in the process.22PubMed Central. TET2 regulates early and late transitions in exhausted CD8(+) T cell differentiation and limits CAR T cell function The decade-long remission cases mentioned earlier turned out to involve an accidental disruption of TET2 in the dominant CAR T-cell clone, suggesting this gene’s role in persistence may have been hiding in plain sight for years.
How CAR T-Cell Persistence Is Measured
Tracking how long CAR T-cells last requires sensitive assays, and the choice of method affects what you can detect. Flow cytometry, which identifies cells by their surface markers, can detect CAR T-cells down to about one in a thousand cells. Quantitative PCR, which looks for the CAR gene in a blood sample, offers roughly similar sensitivity. But for long-term monitoring, when CAR T-cells may be present at very low levels, digital PCR offers about a tenfold improvement in sensitivity, detecting as few as one CAR T-cell per ten thousand cells.23PubMed. Digital PCR Improves Sensitivity and Quantification in Monitoring CAR-T Cells in B Cell Lymphoma Patients This higher sensitivity is especially valuable during the early phase before expansion and during long-term follow-up when cell numbers are at their lowest.24PubMed. Bioanalytical Methods for Characterization of CAR-T Cellular Kinetics: Comparison of PCR Assays and Matrices
This matters practically because a patient might have “undetectable” CAR T-cells by one method while still harboring a functional population that a more sensitive test would pick up. As monitoring technology improves, our understanding of how long CAR T-cells truly persist is likely to shift. Some cases of apparent CAR T-cell disappearance may turn out to be cases of CAR T-cells dropping below a detection threshold rather than actually vanishing.
When Persistence Is Too Much of a Good Thing
Long-lasting CAR T-cells are not an unqualified benefit. When CD19-targeting CAR T-cells persist for years, they continue doing exactly what they were designed to do: kill cells expressing CD19. Normal B-cells express CD19, so persistent CAR T-cells wipe them out along with cancer cells. This means ongoing B-cell aplasia and severe hypogammaglobulinemia, conditions where the body cannot produce adequate antibodies. In the decade-long persistence cases, this required monthly intravenous immunoglobulin replacement therapy to prevent life-threatening infections.3Nature Medicine. Decade-long persistence of CD19 CAR T cells in B cell lymphomas
For a patient whose alternative was dying of leukemia or lymphoma, monthly infusions of immunoglobulin are a manageable trade-off. But as CAR T-cell therapy expands to earlier lines of treatment and less aggressive diseases, the calculus changes. Lifelong immunoglobulin dependence and increased susceptibility to infections represent a real cost. This is one reason researchers are interested in building CAR T-cells with controllable off-switches or limited-lifespan designs for situations where indefinite persistence is not desirable. The ideal scenario is not simply “longer is better” but rather persistence calibrated to the clinical need: long enough to eliminate the cancer and provide surveillance against relapse, but with an exit strategy when the job is done.
Manufacturing as a Persistence Variable
The manufacturing process itself is increasingly understood as a biological variable rather than just a logistical step. How the CAR gene is delivered into the T-cell, which cytokines are used during cell expansion in the lab, how long the cells are cultured, and what stresses they experience during production all leave an epigenetic imprint that shapes how the cells will behave after infusion. A recent framework reframes manufacturing through three layers: the genomic consequences of gene insertion, the clonal fitness of individual cell populations, and the epigenetic programming imprinted by culture conditions.25Mary Ann Liebert, Inc.. CAR-T Cell Therapy: Manufacturing Platforms and Clinical Consequences Shorter manufacturing protocols that minimize time in culture tend to preserve more of the desirable naive and stem-like memory phenotypes. Longer, more intensive expansion can yield higher cell numbers but at the cost of pushing cells toward differentiation and exhaustion before they even reach the patient.
This is an area where there is real tension in the field. Getting enough cells to meet a minimum dose often requires extended culture, but each additional day in the incubator risks degrading the product’s long-term potential. Companies and academic centers are actively experimenting with rapid manufacturing protocols, novel culture media, and automated closed systems to try to thread this needle. The fact that two patients treated over a decade ago with relatively primitive manufacturing methods still have functional CAR T-cells is a reminder that the biology of the starting cells can sometimes overcome suboptimal production conditions, but relying on that luck is not a scalable strategy.