Most compounds on a research-peptide shelf have an empty human file. Ara-290 does not. It has three completed placebo-controlled trials in one patient population, a published phase 2b report in a peer-reviewed journal, and something rarer than any of those: a full results table posted on ClinicalTrials.gov, numbers and all.
That makes it an unusually good test case. When the primary data are actually available, does the sentence everyone repeats survive contact with them?
A peptide carved off the surface of a hormone
Erythropoietin does two jobs. It drives red cell production, and it protects tissue locally after injury. Brines and colleagues proposed that these are separate functions running through separate receptors — hematopoiesis through the EPO receptor homodimer, tissue protection through a heterocomplex of the EPO receptor and CD131, the beta common receptor.
Their 2008 paper in the Proceedings of the National Academy of Sciences (105:10925-30) took that idea to its conclusion. They mapped tissue-protective activity to helix B of erythropoietin, then to an 11-amino-acid stretch forming the aqueous-facing surface of that helix — the side pointing away from the receptor when erythropoietin is bound to its homodimer. That fragment showed activity in injury models and, as designed, none of the erythropoietic activity.
That 11-mer is Ara-290. PubChem lists it under cibinetide (CID 91810664) as C51H84N16O21, roughly 1257 Da, sequence pGlu-Glu-Gln-Leu-Glu-Arg-Ala-Leu-Asn-Ser-Ser with a free acid C-terminus. The literature also calls it the pyroglutamate helix B surface peptide.
The design logic is unusual. This is not a fragment chosen because it is a ligand's active site. It is the part of the surface that was left over — free to do something else while the rest of the molecule was docked.
The receptor is the claim, not the background
On most compound pages the receptor is settled and the biology is argued about. Here it is the reverse.
"Innate repair receptor" is a name coined by the group that proposed it, and it embeds the conclusion. It says what the receptor is for before anyone has shown that is all it does. This site made the same objection about the phrase "immune booster" in the immune class primer: a receptor has a direction of effect in a context, not a purpose.
There is direct evidence for the objection. A 2018 study in Physiological Reports (6:e13751, University of Florida) used an EPO-stimulating peptide that binds the EPO receptor/beta common receptor heterodimer and found that activating it selectively abolished acetylcholine-mediated vasodilation in mouse mesenteric arterioles, an effect a beta common receptor inhibitory peptide prevented. Endothelium-independent responses were untouched. In that preparation, activating the complex was not benign.
There is also a competing account of the phenomenon. A 2026 review in Frontiers in Neurology (16:1665405) covering engineered erythropoietin derivatives describes the non-erythropoietic compounds — carbamylated EPO, asialo-EPO and cibinetide among them — as biasing signalling at the EPO receptor toward PI3K-AKT and away from JAK2-STAT5. That is a biased-agonism explanation, not a separate-receptor explanation. The two are not mutually exclusive, but they are not the same claim, and a page stating the heterodimer as established fact reports one lineage's hypothesis as settled. The vocabulary problem is the one covered in the receptor pharmacology primer.
Three trials, one population
The human work is narrow and consistent in target: sarcoidosis-associated small fiber neuropathy.
The 2012 pilot (Molecular Medicine 18:1430-6, Leiden) randomized 22 patients, double-blind and placebo-controlled, and reported improvement on a small fiber neuropathy screening score against placebo. It also reported something easy to skim past: pain inventory and fatigue scores improved significantly but equivalently in both arms. The placebo response in this population is large.
A larger blinded trial followed in 2013 (Molecular Medicine 19:334-45), reporting improved neuropathic symptoms, increased corneal small nerve fiber density and increased six-minute walk distance. That paper carries a published erratum (Molecular Medicine 2016;22:674) — routine, conclusions intact, but the corrected version is the one to cite, per yesterday's piece on verifying citations.
Then the phase 2b: NCT02039687, sponsored by Araim Pharmaceuticals, 64 participants, randomized, parallel, quadruple-masked, three active dose groups against placebo, 28 days. Published as Culver et al., Investigative Ophthalmology & Visual Science 2017;58:BIO52-BIO60. The primary endpoint was change in corneal nerve fiber area measured by corneal confocal microscopy.
Reading the posted table beside the paper
The short version that circulates is that Ara-290 met its phase 2 endpoint. The longer version is more interesting.
The paper reports the placebo-corrected mean change in corneal nerve fiber area as 109 (95% CI −429 to 647) in the lowest dose group, 697 (159 to 1236, P = 0.012) in the middle group, and 431 (−130 to 992) in the highest. Only the middle group cleared significance. That is a non-monotonic result on the primary endpoint — informative, but not the clean dose-response a mechanism story wants.
The registry's own posted results add three things the abstract does not.
- No statistical analyses are posted at all. The results module carries group means and standard deviations and stops there. The dispersion is wide relative to the means, with standard deviations of several hundred to over a thousand on the primary measure.
- The conventional skin measure did not move in the expected direction. Posted mean change in intraepidermal nerve fiber density, the standard biopsy metric, was slightly higher in the placebo group than in any active group. The paper's skin finding rests on a different measure — GAP-43-positive regenerating fibers, reported as increased in the middle group (P = 0.035). Both statements are true. They are not the same statement, and "increased nerve fiber abundance in the cornea and skin" reads as though they were.
- The symptom endpoint was not significant on placebo correction. The paper states pain improved significantly in all groups, and reports the placebo-corrected decrease in the middle group at P = 0.157. Posted questionnaire means show placebo changes broadly comparable to active ones — the same placebo response the 2012 pilot flagged.
The honest summary is that a 64-participant trial produced a significant result on an imaging surrogate in one of three dose groups, a supporting result on a regenerating-fiber subset, and no placebo-corrected separation on how patients said they felt. That is a signal worth a phase 3. It is not a demonstrated effect, and the distinction between a prespecified primary and everything else is the one set out in the evidence hierarchy piece.
The trial that ran out of drug
There is a fourth registry record, and it is new to this site.
NCT06626971 is an investigator-led phase 2 trial of Ara-290 in diabetic macular oedema, run by Belfast Health and Social Care Trust. It started in April 2016, completed in August 2017, and is marked TERMINATED with 9 participants enrolled. The reason field reads: "Expiry of study drug - no replacement available."
The trial did not stop for futility or for safety. It stopped because the vials passed their date and nobody could send more.
Two further details matter. The record was first posted in October 2024, roughly seven years after the study finished — so while the work was running, and for years afterward, no public trace of it existed. And the developer, Araim Pharmaceuticals, wound down without a phase 3, leaving phase 2 data with no sponsor to carry it. Vendor pages assert orphan and fast-track designations; those claims trace only to vendor-adjacent sources and are not repeated here.
An empty pipeline is not a negative result. This is a compound stranded by economics rather than by evidence, and the habits in the trial registry literacy post are what surface it.
What 2026 has added
Not much, and what there is cuts against inflation.
The most substantive new primary work is a June 2026 study in Neurobiology of Disease (223:107376) from Kuwait University and the Dasman Diabetes Institute — a group with no connection to the originating lineage. Using a touchscreen operant platform in diabetic db/db mice, Ara-290 improved insulin sensitivity and altered circulating monocyte proportions but did not rescue deficits in executive cognitive flexibility. The authors' conclusion is that peripheral metabolic and immune improvement was insufficient to restore that cognitive domain. A partial, independent, honestly reported animal result.
The rest of 2026 is review traffic. An August 2026 narrative review in Current Pain and Headache Reports (30:109) from Brigham and Women's Hospital places cibinetide alongside BPC-157, TB-500 and GHK-Cu, noting most remain unapproved with limited human evidence. A Current Neuropharmacology review of diabetic neuropathy names it among emerging targets. Neither adds data.
What the paperwork can and cannot settle
Ara-290 is short, chemically ordinary, and forces a chemical synthesis route. A few specifics follow from the sequence itself, in the manner of the sequence reading primer.
The N-terminal pyroglutamate is part of the molecule. It costs about 17 Da relative to a free glutamine and, like the pGlu on gonadorelin, blocks aminopeptidase entry. It also removes the N-terminal positive charge. With two glutamate side chains, one arginine and a free acid C-terminus, this is a net anionic peptide — unusual on a shelf dominated by cationic sequences, and relevant to the charge and solubility primer. A theoretical mass calculated from bare letters will disagree with the product by 17 Da: a paperwork error, not a product error, but identical in appearance on a certificate.
Asparagine followed by serine is a deamidation-prone motif, and it sits at positions 9 and 10. Deamidation adds about 0.98 Da — near-mass, and the kind of change that a rounded average mass on a unit-resolution spectrum cannot exclude. See degradation pathways and reading a chromatogram and spectrum.
There is no tryptophan and no tyrosine, so ultraviolet quantitation at 280 nm is unavailable and low-ultraviolet detection is comparatively weak — the same limitation carried by Epithalon and KPV. Quantitation belongs to net peptide content, not to a purity percentage. More at /quality/.
Questions researchers ask
Did Ara-290 fail? No. Nothing published is a failed trial. A phase 2b produced a significant primary-endpoint result in one of three dose groups and no placebo-corrected symptom separation, then the sponsor ceased operations before a confirmatory trial. Those are different outcomes and collapse into neither "it worked" nor "it failed."
Why is it filed under healing-recovery rather than immune? Shelving, not pharmacology. The proposed mechanism is anti-inflammatory and repair-directed, which is why it appears in the immune class primer alongside VIP, LL-37 and KPV despite its catalog category. Category labels here are navigation.
Does the posted results table make the published paper wrong? No. Both are accurate about what they report. The paper ran the analyses; the registry posted the raw group summaries without them. Read together they give a fuller picture than either alone, which is the argument for reading primary records rather than summaries of them. More context in /library/.
This article is educational and for the laboratory research community. Trulogic Labs products are sold for laboratory and research use only and are not for human consumption.