Klotho & Memory
Klotho, memory, and the questions that matter
The Arising · Klotho & memory · November 2026 · John Griffin, MD
Memory is personal long before it becomes a research measurement. It is the conversation we want to remember, the familiar route, the name that matters to someone we love.
That is why a finding about memory can travel quickly. A protein helps older animals perform better on a task, and the next question almost asks itself: could it help us?
Klotho is one such protein. A 2023 study tested it in older rhesus monkeys and reported improved performance after a low dose. Higher doses did not show the same clear benefit. The finding is interesting, but its meaning depends on the details: what was given, what the animals did, how the experiment was organized and what it did not measure. [1]
This is a closer look at that animal study, not a report of a proven human memory treatment. The distinction leaves room for both hope and a useful next question.
What is klotho?
Klotho is a protein involved in several aspects of biology and studied in connection with aging and brain function. The word appears in research about naturally occurring levels, genetic differences and experimental interventions. Those are related subjects, but they do not answer the same question.
A study that measures klotho in blood asks something different from an experiment that administers a manufactured form of the protein. A supplement advertised to influence a person's level is another intervention again. The shared name does not make the evidence interchangeable.
In this study, the researchers first tested the rhesus form of klotho in mice, then administered it to older nonhuman primates. The intervention was a manufactured version of the secreted hormonal form of rhesus alpha-klotho, delivered by injection under the skin. They wanted to know whether an effect seen in earlier animal work could extend to an animal closer to humans. [1]
That is a meaningful step in research. It is not the final step between a promising molecule and clinical care.
How Klotho Was Tested in Older Monkeys
The cognitive cohort included 18 rhesus monkeys, 13 female and five male, aged 15 to 28 years. These were the animals involved across the study's memory experiments, not 18 animals in each treatment group. [1]
The task tested spatial memory: remembering where something had been placed after a delay. The researchers used versions with normal and higher memory demands. These are useful experimental tasks because they give investigators a defined activity to measure repeatedly.
But a task is still a task. Better performance does not tell us whether the animals had less brain disease, whether they would retain the benefit for years, or whether a person would remember appointments or manage daily life more easily.
Human memory is not one ability. Holding information briefly, learning a new route and recalling a conversation draw on overlapping but different processes. A result from one kind of task should not become a claim that every form of memory improved.
Low-Dose Klotho and Monkey Memory Performance
The lower-dose condition was tested in nine animals, with two animals tested twice. The study reported improvement beginning four hours after administration and persisting across later assessments for roughly two weeks. Some later, higher-demand assessments occurred 14 to 23 days after treatment. [1]
That duration makes the result worth investigating. It suggests more than an immediate change measured only once. It does not establish a lasting treatment effect in people or tell us what would happen with repeated administration over months or years.
The right-sized conclusion is that the researchers observed a favorable signal in an older-monkey memory task after a low-dose intervention. It supports further work on the idea. It does not supply a human prescription.
Higher Klotho Doses Did Not Clearly Improve Monkey Memory
Two higher-dose conditions involved seven and 13 animals, respectively. The groups overlapped: some monkeys took part in more than one condition after returning to their prior performance level. The counts should therefore not be added together as though they describe separate groups of new animals. [1]
The higher doses did not produce statistically significant improvements in the reported tests. In plain language, the results of those comparisons were too uncertain to establish a clear benefit under the analysis used. This is not proof that a higher dose was harmful, and it does not precisely identify a best dose for humans.
It does challenge an easy assumption: if a molecule is associated with healthy aging, more of it must be better.
Biology rarely owes us a straight line. The timing, formulation, amount and tissue response can all matter. A low-versus-higher contrast can guide the next experiment without becoming a dosing rule outside the experiment.
Why the study design matters
This was not a randomized clinical trial, nor were the monkeys randomly allocated to treatment conditions. Baseline and vehicle comparisons were part of the design, and the vehicle controls were in animals that had not previously received klotho. A vehicle is the preparation used to deliver a substance without the active protein being tested. [1]
More than half of the monkey testing sessions were blinded, meaning the person assessing the task did not know the treatment condition. Not every session was blinded. Blinding helps reduce the chance that expectations influence how a test is conducted or scored; partial blinding provides less protection than full blinding. [1]
Repeated testing can also affect performance. Reusing animals can be helpful because researchers learn how the same individual performs across conditions, but it makes the analysis more complicated. Measurements from the same animal are connected to one another. They are not equivalent to measurements from many independent animals.
These details do not make the result disappear. They explain why an encouraging small experiment needs careful analysis and follow-up rather than a sweeping conclusion.
Reanalysis of the Klotho Memory Study
A 2026 reanalysis broadly reproduced the original findings while identifying statistical shortcomings and additional comparisons. It used the same data, not a new animal experiment, and does not establish benefit in people. [3]
That distinction matters. Reanalysis asks whether the conclusions hold when the existing measurements are examined again. Independent replication asks whether a new experiment produces a similar result. Both can be useful, but they add different kinds of evidence.
A memory signal is not dementia prevention
The study did not show that klotho prevents dementia in people. It did not establish that it treats Alzheimer's disease, reverses human brain aging or extends human lifespan.
Those would be different claims, requiring different participants, outcomes and follow-up. A trial in healthy older adults would answer a different question from a trial in people with mild cognitive impairment or a defined neurodegenerative disease.
Even in a human trial, a better test score would need interpretation. How much did performance change? Compared with what? Did the difference matter in everyday life? Did it last? What unwanted effects accompanied it?
A promising signal becomes more useful when it helps sharpen those questions, not when it makes them seem unnecessary.
The safety question needs its own evidence
It is tempting to assume that a protein already present in the body must be safe to administer. That does not follow. The amount, formulation, route, timing and duration can create conditions different from normal biology.
This small animal study cannot establish the safety of repeated klotho treatment in people. It does not provide a human dose, a clinical delivery strategy or a complete account of long-term risks.
A future human program would need to measure both exposure and effects. Exposure means how much of the intervention reaches the body and how long it remains available. Researchers would need to watch for unwanted changes as carefully as they watch for memory improvements.
A treatment is more than a molecule with an appealing name. It is a defined product, delivered in a defined way, to a defined group of people, with benefits and harms that can be weighed together.
A blood level and a treatment answer different questions
A blood test can tell researchers something about a molecule at the time it is measured. It does not automatically tell someone whether they should try to raise that number.
To justify a treatment, evidence must connect a practical intervention to a useful outcome, with an acceptable burden of harm. Simply moving a biomarker, a measurable sign of biology, is not enough.
The same applies to a biological-age score. A separate paper on interpreting epigenetic clocks helps explain why a younger-looking molecular reading is not, by itself, proof of better health. That paper does not test klotho or independently confirm this monkey experiment. It offers a useful caution about matching a claim to what was actually measured. [2]
For a memory treatment, the important endpoint is not merely a more favorable number somewhere in the body. It is a meaningful change in memory or function, understood alongside safety.
Who supported the research?
The original study reported support from the US National Institutes of Health, philanthropic sources and Unity Biotechnology. Unity participated in the design of the monkey study. Disclosures included patents related to klotho and cognition, consulting and equity relationships, and biotechnology founder interests. [1]
These relationships do not decide whether a finding is correct. They are relevant context, and they reinforce the value of transparent data and testing by other teams.
What would make the next study more useful?
Independent replication. Another team should be able to test the important findings in new animals. That would add evidence beyond another description or reanalysis of the same experiment.
A clearer dose-response picture. The low-versus-higher pattern deserves careful examination. Researchers need to understand the range where effects appear, how consistent they are and what happens outside that range.
Strong controls and blinding. Random allocation, appropriate comparison groups and blinded assessments can help separate a treatment effect from expectations, practice and other differences.
Transparent harms and longer follow-up. A short favorable result is not a substitute for watching what happens over time, especially if repeated treatment is contemplated.
Human outcomes that matter. Controlled human studies would need to define who might benefit and whether test changes translate into a difference people notice in their lives.
These are directions for future evidence, not achievements already demonstrated by the 2023 paper.
Three ways to keep the evidence attached to the claim
You read that a longevity protein improved memory. Ask which species, how many animals or people, and which memory test. Here the answer leads to a small older-monkey study, not a clinical trial showing that people avoided dementia.
A product says it “supports klotho.” Ask whether the exact product, formulation and use were studied. A blood test, a supplement, an injected protein and a gene-delivery proposal are not substitutes for one another simply because they share a word.
Someone you love is noticing memory changes. An experimental finding is not a reason to put off an assessment. New or persistent memory concerns deserve discussion with a qualified clinician; a research headline cannot identify their cause or determine the right care.
These are illustrative situations, not reports of actual patients or individual treatment advice.
Room for wonder, room for the next question
The interesting thing about klotho is not that its name gives us an answer. It is that an intervention produced a signal worth examining in an animal model relevant to cognitive aging.
The lower dose looked more promising than the higher doses. The benefit appeared on a defined task. The study was small, and its design leaves questions that another experiment can address.
A discovery does not have to be a prescription to matter. Sometimes its value is that it makes a better experiment possible.
The hope is not only for a better score. It is for more of the moments that make memory part of a life. That hope deserves evidence careful enough to tell us when a signal is becoming something people can use.
Sources and further reading
[1] Stacy A. Castner, Shweta Gupta, Dan Wang and colleagues. Longevity factor klotho enhances cognition in aged nonhuman primates. Nature Aging. 2023;3(8):931–937. Primary publication record. Full text.
[2] Daniel S. Borrus, Raghav Sehgal, Jenel Fraij Armstrong and colleagues. When to Trust Epigenetic Clocks: Avoiding False Positives in Aging Interventions. Aging Cell. 2026;25(10):e70756. Primary publication record. Full text.
[3] Olivia C. Robertson, Erik Parker, Beate Henschel and colleagues. Reanalyzing Castner et al. confirms most conclusions, reveals additional findings, and detects some statistical errors. Journal of Robustness Reports. 2026;2-rr1. Publication record. Author-posted article.
Continue exploring
Read Brain health, Biological-age clocks and Regenerative frontiers, or explore the magazine collection.