FOXO4-DRI: The Mouse That Grew Its Fur Back Has a DNA Repair Disease
No human has received it in any study I could find, every animal dose I could read is the one number picked in 2017, and the certificates I opened cannot separate the active molecule from its inert mirror image.
What it is
FOXO4-DRI is a synthetic peptide of 46 amino acids, molecular weight 5358.2 in the paper that introduced it. (PMID 28340339) D-retro-inverso is two changes at once: every amino acid swapped for its mirror image form, the way a left hand mirrors a right, then the sequence written backwards. (PMID 20382735) Living things use the other hand, so enzymes that chop peptides have nothing familiar to grip.
Write the sequence backwards and it falls into three pieces: ten letters of HIV-1 Tat, the standard hitchhiker sequence for dragging cargo through a cell membrane, two prolines, and 34 letters matching human FOXO4 residues 91 to 124. Check it against two public sequence records.
FOXO4 is a transcription factor, a protein that switches genes on and off, and the peptide is meant to prise it off p53, the tumour suppressor that orders a badly damaged cell to destroy itself. Senescent cells, the argument goes, survive by parking p53 where it cannot act.
What the trials found
There are none I could find. The United States government registry returns zero studies with FOXO4-DRI as an intervention, and the European register answered that the query did not match any clinical trials.
The founding paper is Baar and colleagues in Cell 2017, from Peter de Keizer's group at Erasmus MC in Rotterdam. (PMID 28340339) In a dish, they incubated senescent and control human IMR90 lung fibroblasts across a rising concentration series and reported a selectivity index of 11.73, meaning it took roughly twelve times as much peptide to halve the viability of ordinary cells as senescent ones. That is a ratio of two concentrations, not a count of cells killed at any one dose.
Then three mouse experiments. Two of them have their route and dose written down: the chemotherapy one, three doses of 5 mg per kilogram into a vein, and the naturally aged one, the same amount into the belly cavity. The fast-ageing mutant experiment, the one the pictures come from, has neither.
XpdTTD/TTD mice, treated between 26 and 60 weeks of age and followed 30 days, 7 to 8 per treatment for the kidney measurement, showed robust improvement in fur density, more running wheel activity in most of them, and raised plasma urea, a waste the kidney clears, coming back down. Intravenous is only an inference, drawn from the paper flagging one belly cavity experiment in that strain as a departure. The strain is not old either: it carries a mutation in Xpd, a gene used both to repair DNA and to read it, and models trichothiodystrophy, a human genetic disease named for its brittle hair. De Boer and colleagues described the mice in Science in 2002 as showing early greying, wasting and a short life. (PMID 11950998) The photograph is a DNA repair disease responding.
The third is naturally aged mice, roughly two years old, given the same dose into the belly cavity three times every other day, assessed at 30 days, and here the wording goes soft: the peptide could improve fur density and responsiveness, against robust a few figures earlier. The running wheel was abandoned because the animals varied too much, and no survival or lifespan endpoint appears anywhere in the paper.
Eight further studies followed, from at least six groups, none of them de Keizer's, and every one whose dose I could read gave mice that same 5 mg per kilogram into the belly cavity: in 20 aged males, six per group, testosterone was higher 30 days after three doses. (PMID 31959736)
The mirror image in the vial
The founding paper's own control makes this a question: the identical sequence built from ordinary L-amino acids did nothing to senescent cells. (PMID 28340339) So the mirroring is not a durability upgrade on an active molecule. It is the activity.
An all-D peptide and the all-L peptide of one written sequence have the same formula, the same mass and the same behaviour on an ordinary column, so a mass spectrum cannot separate them. It takes a method built for handedness, which is why Adams and Zubarev had to construct one. (PMID 16013875)
The certificates on the listings I opened are HPLC and mass spectrometry: 98.03 percent purity from a research supplier, 99 percent or better from a grey market vendor, with no chiral purity figure, no Marfey's analysis and no circular dichroism spectrum anywhere on them. I have not tested a vial, but the narrow point stands: a standard certificate cannot tell this compound from its inert mirror image. How to read a certificate of analysis covers the rest.
How it compares
Epithalon sells on a bovine extract's clinical record, TB-500 on thymosin beta-4's. This one is different. Its animal record is its own, and what it inherited is a photograph. Independent groups took up the cell killing and the mechanism, in Circulation, JCI Insight and Communications Biology. The fur, the activity and the kidney function, everything a seller means by anti-ageing, come from Baar 2017 alone.
The comparison that paper drew for itself was ABT-263 and ABT-737, senolytics, meaning drugs built to kill senescent cells, which cause severe thrombocytopenia, a dangerous fall in the cells that clot blood. FOXO4-DRI did not noticeably move platelet levels in treated mice, the advantage claimed. Hold it next to the Circulation result below, where both were tested in one model and both looked bad.
Where the dose came from
The 25 mg figure has a derivation, which is rare. On 5 March 2018 the writer Reason published a self-experiment outline at Fight Aging!, method stated: take the mouse figure, apply a published animal to human conversion factor, and out comes 25 mg for a 60 kg adult by vein, three times on alternate days. The author disclaimed encouraging anyone to follow it, and named commenters report having run it on themselves anyway.
A second and incompatible convention is a ladder under the skin, 250 micrograms daily rising to 500 across 16 weeks, labelled on the page carrying it as reported community dosing that no clinical trial has validated for schedule, route or dose. That is roughly a fiftieth of the intravenous figure given about thirty seven times as many times, and I found no published study giving this peptide under the skin to any animal.
Under both sits one decision: every animal dose I could read is 5 mg per kilogram, and in nine years I found no living animal given more or less. It is a recipe copied from one card, and I found nobody who went back and cooked it at half the quantity to see whether the amount mattered.
What could go wrong
There is no human safety information, because I found no human, and the founding paper's own tolerability sentence is hedged.
Then the result pointing the other way. Born and colleagues in Circulation 2023 tested senolytics in pulmonary hypertension: in mice overexpressing the serotonin transporter, given ABT263 or FOXO4-DRI, pulmonary blood flow deteriorated and lung endothelial cells were lost against controls. Clearing those cells, they conclude, may worsen pulmonary haemodynamics. (PMID 36515093) I could not reach the full text, so that arm's dose is unread.
The testosterone group flagged it themselves: FOXO4 protein sits in testis, placenta and muscle, so special attention must be paid to muscle damage and especially to cardiotoxicity, harm to the heart muscle itself. (PMID 31959736)
Bourgeois and colleagues in 2025 measured the peptide binding a p53 fragment covering its first 94 residues at 400 plus or minus 280 nanomolar, a soft measurement, which they call five times stronger than that fragment's grip on the forkhead domain the peptide copies. (PMID 40593617) p53 is the most important tumour suppressor in the human genome, and I found no carcinogenicity study of a molecule built to grip it, nor repeat dose toxicity or immunogenicity work in any species.
What no one seems to have checked
Whether the aged mouse result replicates. Nine years, no independent repeat that I found, and later work in aged animals measured narrower things: testosterone, sperm, aortic stiffness.
What happens to it in a body. I found no pharmacokinetic study in any species, so no half-life and no basis for a dosing interval. The nearest thing I found in the founding paper is an antibody detecting it inside cultured cells at 2 to 4 hours and still at 72 hours, which is a cell in a dish rather than a concentration in a living animal.
Whether the design premise holds. Li and colleagues found that reversing and mirroring a peptide cost 3.2 to 3.3 kilocalories per mole of binding energy on the p53 helix that grips MDM2 and MDMX, two partner proteins of p53, and lost 3.0 and 3.4 in two unrelated proteins, which is what makes the finding general rather than one off. They concluded the strategy works poorly for helical peptides. (PMID 20382735) Different peptide and target, so not a test of this one.
Whose evidence this is
What changed my reading was checking who wrote the follow-up. The 2025 structural paper looks like independent confirmation and is not: de Keizer is an author on it, with two other co-authors of the 2017 paper. That is the test this site runs when a seller waves a citation, and it runs in this direction too.
Then what the inventors did next. Cleara Biotech, spun out of Erasmus MC on this molecule, calls FOXO4-DRI the third generation of its anti-senescence drugs on its history page and says why it built a fourth: ageing was not an easy clinical trial target, and the third generation's selectivity needed improving. Its leads now are CL04177 and CL04183, aimed at solid tumours and blood cancers. It has designed Phase 1a and 1b trials of CL04183, whose primary readout is safety, and says it is looking for Series A financing or a pharmaceutical partner to run them. The vials hold what its inventors moved on from.
If you have seen FOXO4-DRI for sale, I want to know whether the listing showed you the mouse photograph, and whether its certificate named any method that tells one hand from the other.
Frequently asked
Has FOXO4-DRI been tested in humans?
Not in any study I could find. The United States government registry returns no study with FOXO4-DRI as an intervention, the European register returns nothing for FOXO4, and a PubMed search for the name returns 19 records, none of them a clinical trial and none involving dosing a person. The closest thing to human tissue I found is a 2025 paper that applied the peptide to keloid scar tissue kept alive outside the body.
Is FOXO4-DRI safe?
There is no human safety information, because I found no study in which a person received it. In mice, the 2017 paper reported that platelet levels were not noticeably affected, which is the advantage it claims over ABT-263 and ABT-737, senolytic drugs built to kill senescent cells, which cause a dangerous fall in clotting cells, and it hedged its own tolerability statement. An independent group publishing in Circulation in 2023 reported the opposite of a benefit: in mice overexpressing the serotonin transporter, given either ABT263 or FOXO4-DRI, pulmonary blood flow deteriorated and lung endothelial cells were lost. I could find no repeat dose toxicity, carcinogenicity, genotoxicity or immunogenicity study in any species.
What dose has actually been studied?
In animals only, and it is one figure: 5 mg per kilogram in mice, injected into a vein or the belly cavity, usually three doses every other day, in groups of three to eight animals where a count is reported, and the founding paper's chemotherapy experiment reports none. I found no study in any living animal testing a lower or a higher dose. The two human figures that circulate are conventions rather than tested doses, and neither is a recommendation: a 25 mg intravenous figure that a 2018 blog post derived by scaling the mouse dose to a 60 kg adult by body surface area, and a ladder of 250 to 500 micrograms daily under the skin for 16 weeks that the page carrying it labels as reported community dosing that no clinical trial has validated for schedule, route or dose.
Did FOXO4-DRI reverse ageing in mice?
The famous fur and activity results came from XpdTTD/TTD mice, a strain carrying a mutation in a DNA repair gene that models the human genetic disease trichothiodystrophy, in which sparse brittle fur is present from early life. In naturally aged mice in the same paper, the authors wrote only that the peptide could improve fur density and responsiveness, and the running wheel test could not be used because the animals varied too much. No survival or lifespan endpoint appears anywhere in that paper, and the longest follow-up was 30 days.
Can a certificate of analysis show that a vial holds real FOXO4-DRI?
Not the usual ones. The activity depends on the peptide being built from mirror image D-amino acids, and the founding paper found that the identical sequence made from ordinary L-amino acids did nothing to senescent cells. Those two versions are mirror images with the same molecular formula and the same mass, so mass spectrometry and ordinary HPLC cannot tell them apart. Distinguishing them needs a method built for handedness, such as Marfey's analysis, circular dichroism or a purpose-built tandem mass spectrometry approach, and I found none of those on any product listing I opened.
Why does a PubMed search for FOXO4-DRI miss the main paper?
Because the 2017 Cell paper that everything rests on never uses the string FOXO4-DRI in its title or abstract, saying instead that the authors designed a FOXO4 peptide. A search for the name on the vial returns 19 records and that paper is not among them, so someone checking the literature in the obvious way can read the entire indexed corpus for the name and still miss the study underneath it.
This content is for educational and informational purposes only and is not medical advice. Peptides discussed are research compounds and may not be approved for human use. Nothing here should be used to diagnose, treat, cure, or prevent any disease. Always consult a qualified healthcare professional before starting any peptide, supplement, or protocol. Individual responses vary. Do not self-administer compounds without proper medical supervision.
Get the next one in your inbox
Free. One email a week at most. Unsubscribe in one click.
Double opt-in · No spam · Unsubscribe anytime
More in Longevity and bioregulators
- Glutathione: One Infusion Trial in FDA's Search and in Mine, and Nine of Twenty Five Stopped EarlyThe case for putting it in a vein rests on seven people watched for four and a half hours. A six month trial of capsules found the opposite, and I found nobody selling infusions who quotes it.
- The Two Human Claims in Pinealon's Main Review Lead to a Review and a PatentIts inventors say the peptide was lifted out of a cattle brain cortex extract, not a pineal gland, and the only human study I found with a control arm and a dose sits inside a patent that lapsed in 2025.
- Epithalon: The Human Evidence Belongs to a Different SubstanceThe 266 person life extension trial, the 4.1 fold mortality figure and the Kiev follow up all used Epithalamin, a bovine pineal extract. The words Epithalon and Ala-Glu-Asp-Gly appear nowhere in that paper.