Alzheimer's disease is often framed as a problem of plaques and tangles. That framing misses a quieter, earlier process: the loss of new neurons in the hippocampus. A preprint posted to bioRxiv last month (not yet peer reviewed) suggests a peptide called P21 can restore neurogenesis in a mouse model of accelerated brain aging. The study lands at a moment when GLP-1 receptor agonists are being tested for dementia, raising a question about whether these two classes of compounds might one day overlap.
P21 is a small peptide derived from the neurotrophic factor CNTF. It was designed to cross the blood-brain barrier and activate the same pathways that support neuron survival and growth. The new preprint, from a group at the University of Zagreb, used senescence-accelerated mouse prone 8 (SAMP8) mice, a strain that shows rapid cognitive decline and reduced hippocampal neurogenesis by six months of age. Mice received daily intraperitoneal injections of P21 for 30 days. At the end, the treated mice had roughly twice the number of doublecortin-positive neurons in the dentate gyrus compared to saline controls (n=10 per group).
The researchers also ran a Morris water maze test. P21-treated mice found the hidden platform faster and spent more time in the target quadrant during the probe trial. The effect size was moderate but consistent. The preprint's figures show a clear separation between the P21 and saline groups, though the authors note the sample size limits generalizability. Side-effect and adverse-event data for many peptides is sparse. Absence of reported harm does not equate to absence of risk.
Where the neurogenesis hypothesis stands now
For two decades, the idea that adult hippocampal neurogenesis matters for Alzheimer's prevention has cycled through enthusiasm and skepticism. Early postmortem studies found fewer immature neurons in the hippocampi of people with Alzheimer's. Then a 2018 paper in Nature argued that adult neurogenesis drops to undetectable levels in humans after childhood. That paper was contested. A 2019 study using better tissue preservation found thousands of immature neurons in the dentate gyrus of people in their 90s, even those with mild cognitive impairment. The field remains unsettled.
What is clearer is that reduced neurogenesis correlates with cognitive decline, not just in Alzheimer's but in normal aging. The SAMP8 mouse used in the P21 preprint shows a 40 to 60 percent drop in new neurons by six months. Restoring that pool, even partially, might buy time. The preprint does not claim P21 reverses pathology. It claims P21 boosts a process that normally declines. That distinction matters.
P21's mechanism and its cousins
P21 is a 21-amino-acid fragment of ciliary neurotrophic factor. It binds to the CNTF receptor complex and activates STAT3 and AKT signaling. The result is increased expression of brain-derived neurotrophic factor and synaptic proteins. Unlike full-length CNTF, P21 does not cause the severe weight loss or immune activation that doomed earlier CNTF trials. A 2010 study by Chohan et al. showed P21 improved cognition in a mouse model of Down syndrome. Since then, a handful of papers have reported pro-cognitive effects in normal aging, traumatic brain injury, and amyloid-infused rats.
P21 is not the only peptide in this space. Semax, a Russian nootropic peptide, has been shown to increase BDNF and improve attention in small human trials. Selank, its anxiolytic cousin, modulates GABA and may indirectly support neurogenesis by reducing stress-induced cortisol. Pinealon, a tripeptide, appears to protect neurons from oxidative damage. Dihexa, a small molecule, was developed to enhance hepatocyte growth factor signaling and showed remarkable potency in a 2012 rodent study, but human data are absent. NAD+ precursors like nicotinamide riboside are often discussed alongside these peptides because they support mitochondrial health, which is required for neuronal differentiation. None of these compounds have completed large phase 3 trials for Alzheimer's.
The P21 preprint adds a new data point: a direct count of new neurons after chronic treatment. The authors used immunohistochemistry for doublecortin and Ki-67, standard markers. They also measured synaptic density in the CA1 region. P21 increased synaptophysin puncta by about 35 percent. That is a large effect for a 30-day intervention. The cost of custom peptide synthesis for research-grade P21 runs around $48 per milligram, though bulk pricing lowers that considerably. No pharmaceutical company has taken P21 into clinical development.
The GLP-1 connection
GLP-1 receptor agonists like semaglutide are now being studied in Alzheimer's. The EVOKE and EVOKE Plus trials are testing oral semaglutide in people with early Alzheimer's. The rationale is partly metabolic: insulin resistance in the brain is a feature of Alzheimer's, and GLP-1 drugs improve insulin signaling. But GLP-1 receptors are also expressed on neural stem cells in the dentate gyrus. A 2020 study by During et al. showed that liraglutide increased hippocampal neurogenesis in mice by about 25 percent. The effect was dependent on the GLP-1 receptor and involved the same STAT3 pathway that P21 activates.
This overlap raises a question: if GLP-1 drugs already boost neurogenesis, what does P21 add? The preprint does not address that directly, but the effect sizes hint at an answer. The 25 percent increase seen with liraglutide is meaningful but modest. The doubling of new neurons reported in the P21 preprint is larger, though cross-study comparisons are fraught. P21 also acts directly on the CNTF receptor, which is more selectively expressed in the nervous system. GLP-1 receptors are everywhere: pancreas, gut, heart. That broad expression likely contributes to the nausea and weight loss that accompany GLP-1 drugs. P21's narrower target might mean fewer off-target effects, but that remains untested in humans.
The preprint's authors suggest P21 could be used as an adjunct to GLP-1 therapy, not a replacement. The idea is that GLP-1 drugs address metabolic dysfunction while P21 directly stimulates neurogenesis. There is no published data on combining the two. A small biotech in Boston has filed a provisional patent on a P21 analog optimized for oral delivery, but the patent application does not include in vivo data. This is an editorial discussion of published research. It is not a treatment plan.
Why neurogenesis alone may not be enough
New neurons are fragile. In an Alzheimer's brain, they must survive in an environment filled with amyloid oligomers, hyperphosphorylated tau, and inflamed glia. A 2021 study in Cell Stem Cell showed that newborn neurons in Alzheimer's mice die within weeks if the local inflammatory environment is not controlled. That suggests any neurogenesis-boosting strategy will need a partner that quiets neuroinflammation. Selank has shown anti-inflammatory effects in rodent models of brain injury, reducing IL-6 and TNF-alpha. Whether that translates to better survival of new neurons is unknown.
The P21 preprint did not measure inflammation markers. The SAMP8 mouse does have elevated brain inflammation, so the fact that P21 increased neuron survival at all is notable. But the study lasted only 30 days. Longer studies are needed to see if the new neurons persist and integrate into circuits. The Morris water maze improvement suggests some functional integration, but the test is a blunt measure. Electrophysiology would be more convincing.
Another open question is whether P21 affects amyloid or tau pathology. The preprint did not measure these. Most Alzheimer's researchers now believe that amyloid removal alone is insufficient; the recent approvals of lecanemab and donanemab have shown modest clinical benefits at best. A combination approach that clears pathology and restores lost neurons might be more effective. That is speculative, but it is the logic driving interest in neurogenesis as a target.
The peptide landscape and what comes next
Peptide therapeutics for brain disorders have a checkered history. The blood-brain barrier stops most peptides. P21 is one of the few designed to cross it, using a sequence that engages the CNTF receptor's transport mechanism. Semax and Selank also cross, likely via passive diffusion due to their small size and proline-rich structure. Dihexa crosses easily because it is lipophilic. But crossing the barrier is only the first hurdle. Peptides are rapidly degraded in plasma, requiring frequent injections or continuous infusion. An oral P21 analog would be a significant advance, but no such compound has been reported in the peer-reviewed literature.
The preprint on P21 is a single study from one lab. It has not been replicated. The sample size is small. The mouse model, while widely used, is not Alzheimer's; it is accelerated aging. The cognitive benefits were modest. Yet the study is important because it provides a clear, quantitative demonstration that a peptide can boost hippocampal neurogenesis in a model where that process is failing. It also arrives at a time when GLP-1 drugs are forcing the field to reconsider metabolic and trophic factors in dementia.
The next step for P21 would be a study in a transgenic Alzheimer's mouse, ideally with a longer treatment period and measures of both neurogenesis and pathology. A combination study with a GLP-1 agonist would be logical but expensive. Funding for peptide research is scarce; most grants go to monoclonal antibodies or small molecules. The P21 preprint was supported by a university grant of less than $10,000, according to the acknowledgments. That is a fraction of what a single amyloid antibody trial costs.
Discussion of any compound's effects refers to outcomes observed