A newly discovered pathway linking liver function to brain health offers a potential explanation for how exercise protects against cognitive decline, including Alzheimer’s disease. Researchers at the University of California, San Francisco (UCSF) have identified a liver-derived enzyme, glycosylphosphatidylinositol (GPI)-specific phospholipase D1, or GPLD1, that appears to repair damage to the brain’s blood vessels and improve memory in aging mice, according to a study published in the journal Cell.
The findings, announced February 20, 2026, suggest that the benefits of physical activity on brain health may stem from a repair process occurring at the blood-brain barrier (BBB) – the protective shield surrounding the brain – rather than directly within neurons themselves. As people age, the BBB can become increasingly porous, allowing harmful substances to enter brain tissue and triggering inflammation linked to cognitive decline and Alzheimer’s disease.
The UCSF team initially discovered six years ago that exercising mice exhibited cognitive improvements that could be transferred to sedentary mice through blood plasma. This led them to focus on GPLD1, an enzyme released by the liver during exercise. While GPLD1 itself cannot cross into the brain, researchers found it influences another protein, TNAP, which accumulates on the surface of brain blood vessels with age. GPLD1 effectively trims away this buildup of TNAP, restoring the integrity of the BBB.
“This discovery shows just how relevant the body is for understanding how the brain declines with age,” said Dr. Saul Villeda, a researcher at UCSF. “It may open new therapeutic possibilities beyond the traditional strategies that focus almost exclusively on the brain.”
Experiments revealed that older mice engineered to have elevated levels of TNAP on their brain vessels experienced memory impairments, mirroring those seen in aging. Conversely, administering GPLD1 to older mice reduced BBB leakiness, decreased brain inflammation and improved performance on memory tasks. The researchers also tested a compound that lowered TNAP buildup without entering the brain, achieving similar positive results in vessel tightness and memory function.
Further investigation showed that boosting GPLD1 levels in mice bred to develop Alzheimer’s-like plaques led to a reduction in those deposits in the hippocampus, a brain region critical for memory. Human brain samples from older adults with Alzheimer’s disease also exhibited higher levels of TNAP buildup on vessels, suggesting a similar process may occur in humans.
While the research points to a promising therapeutic target, scientists caution that GPLD1 plays roles in other tissues, and long-term disruption of its function could have unintended consequences. UCSF researchers are now focused on testing the safety and efficacy of targeting the TNAP buildup in humans, as well as identifying other potential targets of GPLD1. The team plans to begin safety trials within the next year.
