Epithalon

Epithalon is the tetrapeptide Ala-Glu-Asp-Gly, four amino acids, the best-known product of the St Petersburg Institute of Bioregulation and Gerontology and the one the whole bioregulator series is built around. The institute's method is to take an extract of one animal tissue, analyse it, and synthesise a short peptide from what it finds; AEDG was designed from the composition of Epithalamin, a bovine pineal extract in Soviet clinical use since the 1970s, and its presence in the pineal complex itself was not confirmed until 2017. It is sold as an anti-ageing peptide, and that is what it is taken for here: the telomerase and lifespan reports, the pineal and sleep work, and its place beside the Cartalax and Sigumir courses in the same series.

This entry has two halves, like the NMN entry. The first covers the ageing literature at full depth, because that is the reason the peptide is taken. The second follows the chain from that biology to cartilage, one tested link at a time, so that the reader can see exactly where an anti-ageing compound could reach a joint and where the evidence stops.

The melatonin and circadian work

The pineal gland makes less melatonin with age, and the institute's claim for this peptide is that it restores the output. In old rhesus monkeys, mean age 22.8 years, intramuscular Epithalon raised evening melatonin and normalised the daily cortisol rhythm, with young monkeys largely unaffected; later work from the same group found the old animals' glucose response area falling from 479.6 to 388.9 mM/min on the peptide and drifting back a month after dosing stopped. The one human study at the dose people actually take is the institute's 75-women circadian study: 0.5 mg a day under the tongue for 20 days raised urinary 6-sulfatoxymelatonin, the metabolite that reports pineal output, 1.6 times relative to placebo, reduced Clock gene expression 1.8-fold, doubled Cry2 in leukocytes and lowered Csnk1e 2.1-fold.

Those are real endpoints in the right species. What has to be said beside them is that every one of these papers is the institute's, that the human study sits in a Russian gerontology journal no route delivers, and that the numbers here are the independent review's reading of it rather than the paper's own tables.

The telomerase and telomere work

The 2003 report that the peptide induces telomerase activity and lengthens telomeres in human fetal fibroblasts is the program's most-cited claim, and the 2004 follow-up had the same cells dividing past the 44th passage where controls stopped at the 34th. The one direct measurement in people, relative telomere length in stimulated lymphocytes from donors, changed significantly in 7 of 11, longer in 5 and shorter in 2.

In 2025 a Brunel University group with no institute author, no external funding and no competing interests put quantitative numbers on it for the first time. In two normal human lines, IBR.3 fibroblasts and HMEC mammary epithelial cells, three weeks of AEDG at 1 µg/ml raised hTERT expression, raised telomerase activity fourfold and 26-fold, and lengthened telomeres; four days did nothing. In two breast-cancer lines the telomeres also lengthened, and the route was different: telomerase activity did not rise, and the alternative-lengthening pathway, ALT, rose tenfold in 21NT and threefold in BT474, with the PML bodies that mark it increased. ALT did not activate in the normal cells. The authors conclude from that last fact that the peptide can be used safely in healthy people; the cancer-line finding is the first report of the peptide lengthening telomeres in cancer cells, and nothing has yet looked for it in an animal.

Lifespan and tumours in animals

In 108 female SHR mice injected for life, mean lifespan, food intake and body weight were unchanged; the last 10% of survivors lived 13.3% longer and maximum lifespan rose 12.3%; bone-marrow chromosome aberrations fell 17.1%; total tumour incidence was unchanged while leukaemia fell six-fold. In CBA mice four times as many treated animals reached 23 months, with the oldest control dead at 24 months and the oldest treated mouse at 34. Transgenic HER-2/neu mice grew fewer and smaller mammary tumours, Drosophila lived up to 16% longer, and in rats under different lighting regimens tumour incidence fell while mean lifespan in the males did not lengthen. A frailty-model reanalysis of the transgenic mice found one dosing scheme shortening maximum lifespan.

The pattern across the strains is a longer tail rather than a longer mean, and fewer tumours in the institute's hands. Two results come from elsewhere. A Shandong group with provincial funding found the peptide, at 0.1 mM, protecting mouse oocytes ageing in culture: reactive oxygen species down, fragmentation from 13% to 5.8%, spindle and mitochondrial defects partly rescued, DNA damage and early apoptosis lower, and no effect on reactive oxygen species at the two highest concentrations. The Chieti collaborations, with the institute supplying the peptide, found it restoring wound closure in glucose-injured retinal cells at 20 to 60 ng/ml and raising neurogenic markers 1.6 to 1.8-fold in gingival stem cells at 0.01 µg/ml.

The human record

Two clinical studies of the peptide exist, both from the institute, and the independent review is the only route to either. The retinitis pigmentosa series gave 162 patients aged 18 to 72 the peptide by injection beside each eye for 10 days, against the conventional treatment of 2002, and reports visual acuity up 0.15 to 0.20 on average, the visual field widened by 90 to 120 degrees in 64.8% of patients, and no side effects. The 75-women circadian study is above. No registered trial of the peptide exists anywhere as of September 2026.

For a joint, there is nothing. PubMed holds 172 records for the peptide under any of its names and none of them combines it with arthritis, osteoarthritis, cartilage, chondrocytes or synovium. The one paper in the orthopaedic literature that names it, a 2026 narrative review of therapeutic peptides in the American Academy of Orthopaedic Surgeons' open-access journal, lists it among recovery-enhancing agents and says in its own words that it is not directly targeting musculoskeletal tissue, proposing only an indirect benefit through sleep and hormonal regulation; it cites two institute papers for that and no joint study, and describes the whole peptide field as dominated by animal models, small cohorts and case series with very limited randomised trials.

The closest thing to a joint result is the report the tier rests on: 266 elderly people in St Petersburg and Kiev given the thymic extract Thymalin and the pineal extract Epithalamin in the first 2 to 3 years of a 6 to 8 year observation, with mortality 2.0 to 2.5-fold lower than an untreated control group, 4.1-fold lower in a subgroup dosed annually for six years, and lower reported incidence of ischaemic heart disease, hypertension, deforming osteoarthrosis and osteoporosis. It is indexed as a randomised trial and describes no randomisation procedure, no allocation concealment, no blinding and no confidence intervals, and it tested the extract rather than the peptide.

Where the evidence comes from

Of the records on the peptide itself, three carry no author from the institute: the 2025 review by pharmacy and geriatrics groups in Warsaw and Bydgoszcz, the Brunel telomere study, and the Shandong oocyte study. The Chieti work is a collaboration with the institute's founder as a co-author and the peptide supplied from St Petersburg. The institute's own papers from 2000 to 2021, the monkey work and both clinical studies among them, sit behind Springer, Elsevier and Russian-journal paywalls, and none has been read here.

The independent review is worth the reader's attention. It read 25 years of this literature, accepts that the geroprotective and neuroendocrine effects are there in the record, and concludes that the mechanism remains unclear, that physico-chemical and structural work on the peptide is very limited, and that information on its safety is missing.

The chain from ageing biology to cartilage

Every link below has been tested in human tissue or an animal. None has been tested with this peptide.

Telomeres shorten in ageing and osteoarthritic cartilage. Across 27 donors aged 1 to 87, mean chondrocyte telomere length fell from 11.8 kbp in a 13-year-old to 8.7 kbp in an 87-year-old. Chondrocytes from affected sites have shorter telomeres than unaffected ones in 15 hip and 30 knee patients against 11 controls; leukocyte telomeres were 178 bp shorter in 160 people with hand osteoarthritis than in 926 without, and no different in 34 patients against 130 controls in another cohort. In three end-stage knees biopsied at paired positions, the load of ultra-short telomeres rose towards the lesion in 19 of 20 directions and tracked grade and senescence more closely than mean length did. Against that, one autopsy comparison of osteoarthritic and normal cartilage found more DNA damage but no evidence of critical telomere shortening. And osteoarthritis cartilage is epigenetically 3.7 years older than control cartilage while the same patients' bone and blood are not, which puts the premature ageing in the tissue rather than in the body.

Adult chondrocytes have no telomerase. The same series that measured the shortening demonstrated that chondrocytes are devoid of telomerase activity, and telomerase overexpression avoids chondrocyte senescence only if cell-cycle inhibitors are also inactivated. Because mitosis in normal cartilage is very low, chondrocyte telomere loss is attributed to oxidative stress and abnormal loading rather than to replication. A telomerase-inducing peptide is therefore aimed at cells that do not run the enzyme, in a tissue whose telomere loss is not mainly replicative. That is the break in the chain, and it is where any future study of this peptide in cartilage would have to start.

Senescent chondrocytes drive the disease, and removing them has been tried. Osteoarthritis chondrocytes carry raised p16, an activated DNA-damage response and a secretory phenotype of cytokines and matrix enzymes. The intra-articular senolytic UBX0101 reduced osteoarthritis features in 10-week-old mice, worked less well at 72 weeks, and in a phase 2 trial in adults with symptomatic knee osteoarthritis did not reduce pain or improve function at 12 weeks; oral fisetin is in a knee trial with results unreported; raising NAD+ by blocking CD38 reduced osteoarthritis severity in a preclinical model. The senescence route to a joint is real and its first human trial was null.

Melatonin protects cartilage in cells and animals. Melatonin protected a human chondrocyte line from hydrogen-peroxide injury and reduced cartilage destruction after intra-articular injection in a rabbit model, with both effects abolished when SIRT1 was inhibited or knocked down; in human osteoarthritic knee cartilage the clock gene BMAL1 and NAD+ are both reduced, and restoring either reverses the inflammatory gene changes. Two later reviews assemble the chondrocyte and animal work on mitochondria, reactive oxygen species and matrix enzymes, and name no human trial of melatonin with a joint endpoint. This is the link that connects to the peptide: the 1.6-fold rise in urinary melatonin metabolite in the 75-women study. Whether raised endogenous melatonin does to a human joint what injected melatonin does to a rabbit's has not been tested by anyone.

Dosing, safety and regulatory status

The 75-women study used 0.5 mg a day sublingually for 20 days, which is the course the disclosure page carries. The animal work is injected at 0.1 to 1 µg per mouse or 2 to 10 µg/kg in rats, and the cell work runs across four orders of magnitude, from 0.01 µg/ml to 0.1 mM. The peptide has no FDA status of any kind and is sold as a research chemical; the Brunel group bought theirs from an online vendor, which is how it reaches most people who take it, and products labelled as Epithalon have had to be checked by mass spectrometry against a counterfeit-prone isomer. Short peptides of this kind degrade rapidly in vivo, no stabilised form has been evaluated, and of its eight stereoisomers only the natural one has been studied.

No human safety study exists. The retinitis series reports no side effects, the rat work reports kidney protection rather than harm, and the institute's mouse work reports fewer tumours in several strains. The Brunel cancer-line result, telomere extension through ALT in two breast-cancer lines, is the one finding that points the other way, and no study has looked for it in an animal. This entry is an appraisal of published research, not dosing guidance.

What would move this

One study of the tetrapeptide in a chondrocyte, an explant, a joint-injury model or an osteoarthritis model would move the tier to preclinical, whatever it found, because it would be the first cartilage endpoint the peptide has ever had. The interesting version is the one the chain points at: senescent human chondrocytes, which have no telomerase, given the peptide and measured for the secretory phenotype the Cartalax work already measures for the tripeptide. Short of that, a pharmacokinetic study in any species, a replication of the 75-women melatonin result outside the institute, and an animal look for the ALT activation the Brunel paper found in cancer cells are the three results the literature most needs, and the last of them is the one a person taking it should want.