An aged mouse finishes its last lap on a running wheel, and within hours something unexpected happens: its liver floods the bloodstream with an enzyme that will travel to the brain and strengthen the very circuits responsible for forming new memories. That enzyme, along with a protein released by platelets, now sits at the center of a body of research that traces, for the first time, a concrete molecular route from exercising muscles to the hippocampus, the brain region most vulnerable to aging and Alzheimer’s disease.
The work, led by teams at the University of California, San Francisco and other institutions and published across Science, Nature Aging, and most recently Cell, offers the clearest explanation yet for a pattern that older adults and their doctors have long observed: a brisk walk or a short cycling session can sharpen recall hours or even days later, well after the muscles have cooled down.
A liver enzyme that reaches the aging brain
The strongest thread in this research centers on GPLD1, a liver-derived enzyme whose blood levels rise after sustained physical activity in aged mice. In a May 2025 study published in Cell, researchers led by Saul Bhatt and Saul Villeda at UCSF showed that manipulating GPLD1 directly reversed memory deficits in both aging and Alzheimer’s-model mice by acting on a specific vascular target in the hippocampus. The protein did not need to stimulate neurons directly. Instead, it remodeled blood vessel function, improving the infrastructure that supports memory formation.
That finding built on an earlier experiment, published in Science in 2020 by many of the same researchers, which demonstrated that the benefit could be transferred between animals. Plasma collected from exercised aged mice and infused into sedentary aged peers improved hippocampal neurogenesis and cognition in the recipients. GPLD1 was identified as a key factor in that plasma. The National Institutes of Health highlighted the result, noting that liver-to-brain signaling may be a central mechanism behind exercise-related cognitive protection in aging.
The plasma transfer experiment is especially telling because it rules out explanations tied to body temperature, heart rate, or stress hormones during exercise itself. The cognitive benefit traveled in the blood.
Platelets carry a second signal
GPLD1 is not working alone. A separate line of research has identified platelet factor 4 (PF4, also called CXCL4) as another exercise-induced protein that reaches the hippocampus. In a 2023 study published in Nature Aging, a team led by Odette Leiter and Tara Walker showed that PF4 improved hippocampal-dependent learning and memory in aged mice while reducing neuroinflammation.
A companion study, also in Nature Aging, linked the longevity factor klotho to elevated PF4 levels and demonstrated that blocking PF4 eliminated the cognitive gains entirely. That result established necessity, not just correlation: without PF4, the memory improvements disappeared.
Together, the GPLD1 and PF4 findings point to at least two distinct peripheral routes, one from the liver and one from platelets, that converge on the hippocampus after exercise.
A broader family of exercise signals
These two proteins fit within a larger catalog of molecules that the body releases during physical activity, collectively known as exerkines. Cathepsin B, which originates in muscle, and the irisin/FNDC5 pathway have both been shown to influence hippocampal function through separate receptor mechanisms, according to a review in Nature Reviews Endocrinology.
A more recent synthesis of peripheral-to-hippocampus crosstalk, published in Life Sciences, confirms that muscle, liver, fat tissue, and platelets all contribute circulating factors with documented effects on brain structure and cognition in animal models. The picture that emerges is of the body functioning as a distributed endocrine system during exercise, with multiple organs sending chemical messages to the brain simultaneously.
The gap between mice and humans
All of the vascular and neurogenesis evidence described above comes from mouse experiments. No published study has yet correlated post-exercise changes in GPLD1 or PF4 blood levels with memory performance over time in people. (The original 2020 Science paper did measure higher GPLD1 levels in physically active older adults compared to sedentary peers, but it did not track whether those levels predicted cognitive outcomes.)
Whether these proteins cross the human blood-brain barrier in meaningful concentrations, and whether they act on the same vascular targets identified in mice, has not been tested directly.
The closest human data involves a different mechanism. A clinical study of older adults and patients with amnestic mild cognitive impairment found that acute exercise enhanced memory consolidation through a locus coeruleus-norepinephrine pathway. That noradrenergic effect, documented in a 2012 study and supported by subsequent work, is real and measurable in people, but it operates on a much shorter timescale than the durable circulating-factor transfer seen in the mouse plasma experiments. Whether the two mechanisms work in tandem or whether one dominates in humans remains an open question.
Unanswered questions about dose, timing, and individual differences
Researchers have not established whether a single bout of moderate activity in an older adult produces enough GPLD1 or PF4 to trigger hippocampal remodeling, or whether repeated sessions are required to build a threshold concentration. The mouse studies used sustained exercise regimens or direct protein infusions, neither of which maps neatly onto a 30-minute walk.
There is also uncertainty about how age, sex, and common health conditions may alter these pathways. Most of the animal work has been done in relatively healthy aged mice. Whether diabetes, cardiovascular disease, or chronic inflammation would blunt the liver and platelet responses to exercise is unknown. The same goes for medications commonly prescribed to older adults: statins, blood thinners, anti-inflammatory drugs, and others could theoretically interfere with GPLD1 or PF4 signaling, but no one has checked.
Exercise type matters too, and the current research does not resolve it. The mouse studies relied on voluntary wheel running, which is roughly analogous to moderate aerobic activity. Whether resistance training, high-intensity intervals, or low-intensity movement like yoga would trigger the same liver and platelet responses is simply not known.
How peripheral signaling reshapes the case for exercise in aging
For readers tracking headlines about “exercise pills” or “memory-boosting proteins,” the distinction between mechanistic excitement and clinical readiness is important. The identification of liver- and platelet-derived factors gives scientists concrete drug targets and potential biomarkers for future clinical trials. It does not mean that purified GPLD1 or PF4 infusions are safe, effective, or appropriate for humans. Any such therapy would need rigorous testing for off-target effects on coagulation, immunity, and peripheral organs before it could be considered.
In practical terms, the current evidence most strongly supports recommendations that were already in place: regular physical activity is associated with better cognitive aging. What these studies add is a plausible molecular explanation for part of that benefit, and a reframing of how the body works during exercise. The brain does not improve in isolation. The liver, platelets, and muscles act as signaling organs, releasing proteins that reach the hippocampus and influence how memories are formed and preserved.
As human data accumulate, clinicians may eventually personalize exercise prescriptions based on measurable changes in circulating exerkines, or combine physical activity with drugs that amplify beneficial liver-to-brain and platelet-to-brain signaling. As of June 2026, that future remains speculative. The safest conclusion is also the most actionable: moving the body sends powerful chemical messages to the aging brain, and scientists are now reading those messages at the molecular level for the first time.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
