T Cell-Engaging Antibody Clears Aging Cells in Mice and Monkeys, With Liver Enzymes as a Built-In Safety Gauge

Scientists have long dreamed of plucking the worn-out cells that accumulate in aging tissues, and a new study moves that dream closer to the clinic. Writing in Aging Cell, a research team based at West China Hospital of Sichuan University and the Kunming Institute of Zoology reports that a bispecific T-cell engager targeting the molecule uPAR can safely eliminate senescent cells in naturally aged mice and in elderly rhesus macaques. Just as importantly, the team identified the therapy’s main danger signal in advance and showed that a routine blood test, serum aminotransferase activity, can track it in real time.

Senescent cells are cells that have permanently stopped dividing but refuse to die. They secrete a toxic cocktail of inflammatory molecules known as the senescence-associated secretory phenotype, or SASP, and even small numbers of transplanted senescent cells can accelerate aging in mice. Conversely, drugs that clear these cells, called senolytics, have extended healthspan and improved age-related phenotypes in animal studies. The landmark dasatinib-plus-quercetin combination, along with fisetin and UBX0101, established proof of concept, but small-molecule senolytics lack precision and have produced disappointing clinical results accompanied by side effects such as diarrhea, anemia, and rash.

Immunotherapy offers a sharper alternative. uPAR, the receptor for urokinase-type plasminogen activator, is broadly expressed on senescent cells and tumor cells but largely absent from healthy young tissue. CAR T cells engineered against uPAR have already reversed senescence-driven pathologies in mice, including liver fibrosis and metabolic dysfunction. Yet the approach carries risks: high doses of uPAR CAR T cells caused hypothermia, weight loss, and elevated cytokines in mice, and CAR T cells, once infused, expand unpredictably from patient to patient, making their toxicity hard to titrate. The team therefore turned to a bispecific T-cell engager, or BiTE, a soluble antibody construct that recruits a patient’s own T cells to target cells and vanishes from the bloodstream within days, allowing dosing to be adjusted or halted at will.

The engineered molecule, named hGFD-CD3, links the GFD domain of human uPA, the natural ligand of uPAR, to an anti-CD3 antibody fragment derived from the clinically tested drug acapatamab, with a non-functional Fc region added to extend its half-life. In laboratory dishes, the engager bound uPAR-positive senescent human lung fibroblasts and CD3-positive T cells simultaneously, killing senescent cells with an IC50 of just 0.015 nanomolar, a potency the authors attribute to the moderate, natural affinity of the ligand-based recognition domain. Young cells and cells treated with a control construct lacking the GFD domain were spared. Treated T cells upregulated the activation markers CD25 and CD69 and released granzyme along with the inflammatory cytokines TNF-alpha and IFN-gamma.

The in vivo story was more complicated, and it is here that the study makes its most consequential discovery. When aged mice received a high dose of the mouse version of the engager, 1 milligram per kilogram, they died within 24 hours; 0.3 milligrams per kilogram caused ruffled fur, anorexia, and a 20- to 30-fold surge in the liver enzymes AST and ALT. Histopathology pointed squarely at the liver. The reason, the team found, is that liver sinusoidal endothelial cells are among the very few non-hematopoietic cells in the body that express uPAR. Unlike the continuous, basement-membrane-backed vessels of the lung and kidney, liver sinusoids are fenestrated and lack a basement membrane, so widespread destruction of their endothelial lining collapses the microcirculation and injures hepatocytes. In uPAR knockout mice, the engager caused no liver damage at all, proving the toxicity was uPAR-dependent. The same hepatotoxicity appeared with uPAR CAR T cells, even in macrophage-deficient NSG mice, overturning the earlier hypothesis that macrophage activation was to blame.

Crucially, that toxicity proved monitorable and avoidable. At 0.1 milligrams per kilogram, a dose that left liver enzymes and tissues unremarkable, weekly mGFD-CD3 treatment for three weeks significantly reduced uPAR, p16INK4a, and p21 expression across tissues of 22-month-old mice, decreased senescence-associated beta-galactosidase staining in fat, liver, and lung, and suppressed SASP gene expression in liver RNA-sequencing analysis. The engager physically bridged T cells to senescent cells in tissue sections, activated peripheral T cells, and reduced uPAR-positive macrophages. Cytokine rises were modest, roughly twofold, far below cytokine-storm thresholds, and returned to baseline after treatment stopped.

The long-term benefits in mice were striking. Three months after starting treatment, circulating SASP factors had fallen, liver and lung fibrosis had eased, glomerulosclerosis was attenuated, and bone remodeling shifted favorably, with fewer osteoclasts and more active osteoblasts. After six months, treated mice showed better grip strength, treadmill endurance, walking speed, and hanging endurance, along with denser skeletal muscle fibers, less trabecular bone loss, a higher lean-mass fraction, healthier subcutaneous fat distribution, lower fasting glucose, and improved glucose tolerance and insulin sensitivity. Serum transaminases remained within normal ranges throughout, and no tissue showed pathological abnormalities.

The team then crossed the species barrier. In 24-year-old rhesus macaques, roughly equivalent to a 70-year-old human, a dose-escalation trial showed no abnormalities at 10 or 20 micrograms per kilogram, while 40 micrograms per kilogram produced anorexia and a 50 to 100 percent rise in AST and ALT, though still within the normal reference range. At the selected 20-microgram-per-kilogram dose given every three days, two of three macaques tolerated treatment without clinical signs; one showed transient anorexia and transaminase elevations that peaked at 108 U/L for AST and 134 U/L for ALT and cleared within three days. Temporary drops in white blood cells, neutrophils, lymphocytes, and blood pressure all rebounded by day three, a pattern consistent with T-cell migration into tissues rather than destructive depletion, and these fluctuations diminished with repeated dosing.

Over 90 days of follow-up, senescent cell burden in subcutaneous fat, measured by beta-galactosidase staining, fell by day 42, and uPAR, p16INK4a, and p21 expression dropped in fat, muscle, and skin before drifting back up by day 90 as senescent cells slowly re-accumulated. Serum SASP levels fell in parallel, and immune-cell gene signatures declined, indicating reduced tissue inflammation. The macaque with the heaviest senescent burden was the one with the strongest transient enzyme and cytokine responses, reinforcing the authors’ proposal that liver function tests can serve as a practical, biomarker-guided safety gauge for uPAR-targeted senolytics. Because senescent cells are rare, non-proliferative, and take months to return, the team argues that intermittent, low-dose regimens, potentially as infrequent as twice yearly, could impose minimal treatment burden. They caution that for preventive anti-aging use the risk-benefit balance demands careful evaluation, and suggest that diseases with clear unmet need, such as chronic kidney disease, Alzheimer’s disease, and liver fibrosis, may be the most sensible first indications for clinical translation.

Subject of Research: Senolytic immunotherapy targeting uPAR with a bispecific T-cell engager in aged mice and non-human primates

Article Title: uPAR‐Targeting T Cell Engager Exerts Senolytic Effects in Mice and Non‐Human Primates With Serum Aminotransferase Activity as a Safety Monitor

Article References: Deng, J., Zeng, X., Guo, J., Li, L., Zou, S., Liu, N., He, S., Yan, L., Wang, Q., Zhu, Y., Wang, B., Yang, D., Lv, L., & Zhao, X. (2026). uPAR ‐Targeting T Cell Engager Exerts Senolytic Effects in Mice and Non‐Human Primates With Serum Aminotransferase Activity as a Safety Monitor. Aging Cell, 25(10), Article e70742. https://doi.org/10.1111/acel.70742

Image Credits: AI Generated

DOI: 10.1111/acel.70742

Keywords: senolytics, cellular senescence, uPAR, T-cell engager, BiTE, aging, hepatotoxicity, aminotransferases, CAR T cells, rhesus macaques, SASP, liver sinusoidal endothelial cells