ARA-290 10mg Peptide

ARA-290 is an 11-residue synthetic peptide that reproduces the water-facing surface of helix B of human erythropoietin. Its international non-proprietary name is cibinetide, and the older literature calls it pHBSP, for pyroglutamate helix B surface peptide. It carries CAS number 1208243-50-8, the molecular formula C51H84N16O21 and an average mass of 1,257.3 g/mol.

The molecule exists to separate two things erythropoietin does at once. EPO raises red cell mass by binding an EPO receptor homodimer, and it protects injured tissue by binding a different receptor: a heterocomplex of one EPO receptor subunit and CD131, the beta common receptor shared by the GM-CSF, IL-3 and IL-5 receptors. Brines and Cerami's group established that split in 2004 and then asked which part of the EPO molecule was responsible for the second activity. Helix B, which faces the aqueous medium when EPO is docked on the erythropoietic homodimer, turned out to carry it. ARA-290 is what remains when you keep only that face.

The design is unusual in a research peptide catalogue because it is not a fragment. The 11 residues are not a contiguous stretch of the EPO chain. They are the amino acids that project into solvent from helix B, spanning EPO residues 58 to 82, plus three residues from the proximal portion of the BC loop, assembled into a linear peptide that mimics a surface rather than a sequence. The N-terminal glutamine is cyclised to pyroglutamate, which is where the pHBSP name comes from and why the peptide begins with a blocked amine.

Development reached phase 2 and stopped there. Four trials are registered on ClinicalTrials.gov, the largest a 64-subject phase 2b in sarcoidosis-associated small nerve fibre loss, and no phase 3 has ever been registered. That record is the honest frame for everything below: the receptor pharmacology is well evidenced, the preclinical file is large and international, and the human data are small, mostly from two academic centres, and mostly funded by the company that owns the molecule.

ARA-290 is a selective agonist at the EPOR/CD131 heteroreceptor, which the group that developed the peptide named the innate repair receptor. The evidence that this receptor is a real entity is loss-of-function rather than structural: EPO receptor co-immunoprecipitated with the beta common receptor from membranes prepared from rat brain, heart, liver and kidney, and neither EPO nor carbamylated EPO protected cardiomyocytes or spinal cord neurons in beta common receptor knockout mice (Brines et al., PNAS 2004). A 2024 mouse middle cerebral artery occlusion study made the same point for ARA-290 specifically: knocking down the beta common receptor with siRNA abolished the peptide's reduction in neuronal apoptosis (Wang et al., CNS Neuroscience and Therapeutics 2024).

The receptor's most useful property is where it is not. CD131 pairs with the EPO receptor only in tissue under metabolic stress, inflammation or injury, and the complex is not appreciably assembled in resting tissue. That gives the peptide a form of targeting that is not conferred by the peptide at all: it is conferred by the state of the tissue. ARA-290 has no measurable affinity for the EPO receptor homodimer, so it does not stimulate erythroid progenitors. It was inactive in the UT-7 EPO-dependent cell assay and produced no erythropoiesis in vivo (Brines et al., PNAS 2008), and a month of daily administration in streptozotocin-diabetic rats left haematocrit unchanged (McVicar et al., Diabetes 2011).

Downstream, engagement of the heteroreceptor runs through JAK2 into the survival kinases. Helix B surface peptide inhibited about 80% of TNF-alpha induced apoptosis in cultured neonatal rat cardiomyocytes and activated Akt, ERK1/2 and STAT3, with siRNA experiments showing Akt was the essential arm. In parallel the peptide suppresses the inflammatory transcription programme: in LPS-activated primary mouse macrophages its anti-inflammatory effect required both CD131 and functional JAK2 and was mediated by inhibition of NF-kB subunit p65 (Nairz et al., Scientific Reports 2017). In the same study cibinetide reduced myeloid infiltration and nitric oxide synthase-2 production in DSS colitis.

At least one action does not obviously belong to the receptor story. Calcium imaging and behavioural work in rodents found that ARA-290 inhibits TRPV1 channel activity directly and relieved capsaicin-induced mechanical hypersensitivity (Zhang et al., Peptides 2016). Whether that contributes materially to the analgesia seen in nerve injury models is unresolved, and it is one of the few places where the innate repair receptor account of this peptide is incomplete rather than merely unproven.

Erythropoietin protects tissue in a long list of preclinical injury models, and for a while that looked like a clinical opportunity. It was not, because the dose needed for tissue protection also raises haematocrit, activates platelets and increases selectin expression, and large trials of erythropoiesis-stimulating agents ran into thrombotic and cardiovascular harm. The field's response was to look for a version of the molecule that kept the second activity and lost the first.

Carbamylated EPO was the first answer. It does not bind the EPO receptor and does not stimulate erythropoiesis, yet it remained tissue-protective across in vitro and in vivo models. That result forced the question of which receptor was doing the work. Brines and colleagues reported in PNAS in 2004 that EPO receptor could be pulled down in a covalent complex with the beta common receptor from rat brain, heart, liver and kidney membranes, that the two colocalised on spinal cord neurons and cardiomyocytes, and that in beta common receptor knockout mice neither EPO nor carbamylated EPO protected those cells. Erythropoiesis was normal in the same animals, which is the clean dissociation the whole programme rests on.

The 2008 PNAS paper cut the molecule down. Helix B, EPO residues 58 to 82, faces the aqueous medium when EPO is bound to the erythropoietic homodimer, and it proved neuroprotective in vitro and protective in vivo in ischaemic stroke, diabetes-induced retinal oedema and peripheral nerve trauma. An 11-residue peptide built from the solvent-facing amino acids of that helix plus three residues of the adjacent BC loop reproduced the effect in stroke and renal ischaemia-reperfusion models, accelerated wound healing and improved rodent cognitive performance. Neither construct was erythropoietic. The doses in that paper are worth quoting for scale, because they are nanomolar per kilogram rather than milligram: 1.5 nmol/kg intravenously in stroke, 0.08 to 8.0 nmol/kg intraperitoneally in renal ischaemia-reperfusion, 0.3 nmol/kg intravenously after sciatic nerve injury, and 24 nmol/kg for wound healing and cognition.

Cyclising the N-terminal glutamine to pyroglutamate produced pHBSP, which is the form now called cibinetide and sold as ARA-290. The modification blocks the free alpha-amine, which removes the obvious aminopeptidase entry point at that end of the chain.

Small nerve fibre loss complicates sarcoidosis often, produces neuropathic pain and autonomic dysfunction, and responds poorly to immune suppression. It became the lead indication because it offered a measurable structural endpoint: corneal confocal microscopy quantifies small nerve fibres non-invasively and can detect regeneration, which questionnaires cannot.

The 2012 pilot randomised 22 patients, 12 to the peptide and 10 to placebo, using intravenous administration three times weekly for four weeks. The small fibre neuropathy screening list score fell by 11.5 points against 2.9 on placebo, and the pain and physical functioning dimensions of the SF-36 moved from baseline. The honest reading of that trial includes what did not separate: Brief Pain Inventory and Fatigue Assessment Scale scores improved significantly in both arms, and the Inventory of Depressive Symptomatology did not change at all.

The 2013 follow-up switched to 28 days of daily subcutaneous administration and added the structural measure. Corneal small nerve fibre density increased significantly, cutaneous temperature sensitivity changed, and 6-minute walk distance rose, alongside the symptom scores.

The phase 2b trial, NCT02039687, is the largest human study of this molecule and the one to read closely. Sixty-four subjects were randomised across 1 mg, 4 mg, 8 mg and placebo arms for 28 days, with change in corneal nerve fibre area at day 28 as the primary endpoint. The placebo-corrected change was 109 square micrometres at 1 mg (95% CI -429 to 647), 697 at 4 mg (159 to 1236, p = 0.012), and 431 at 8 mg (-130 to 992). Only the middle arm separated from placebo. Regenerating GAP-43 positive intraepidermal fibres rose in the same arm (p = 0.035), and the corneal change tracked both the skin marker (rho = 0.575) and 6-minute walk distance (rho = 0.645, p = 0.009). Pain improved in every arm including placebo, and in the moderate-to-severe subgroup the placebo-corrected fall in pain intensity at 4 mg reached p = 0.157, which is not significance. The structural signal is stronger than the symptomatic one, and the inverted dose response is a design fact anyone modelling this compound should account for rather than average away.

The type 2 diabetes trial published in Molecular Medicine in 2015 used the 4 mg arm for 28 days and then followed subjects for a further month with no treatment. Haemoglobin A1c and lipid profiles improved across the whole 56-day window rather than only during administration, PainDetect scores improved significantly against placebo, and corneal nerve fibre density rose in the subgroup whose baseline density sat more than one standard deviation below normal. Metabolic endpoints were secondary, and the trial was not powered to settle them.

The retinal work is older and preclinical. Rats given streptozotocin and left diabetic for six months, then treated with 10 micrograms per kilogram daily for one month, showed reduced Muller glial GFAP expression (p < 0.001), fewer CD11b positive microglia, lower proinflammatory cytokine expression, less 8-hydroxydeoxyguanosine and TUNEL positivity, and no acellular capillary formation. Two negative controls in that study matter as much as the positive results: haematocrit did not change, and in the oxygen-induced retinopathy model the peptide did not increase preretinal neovascularisation at any dose from 1 to 30 micrograms per kilogram. Angiogenic tissue-protective agents in the eye have to clear that second bar and EPO itself does not.

The clinical translation into eye disease did not work. A phase 2 study in diabetic macular oedema at Belfast, registered as NCT06626971, recruited 9 patients between May 2016 and April 2017 and was terminated. Over 12 weeks there was no improvement in best corrected visual acuity, central retinal thickness, central retinal sensitivity or tear production. The visual function questionnaire composite score rose slightly, individual participants showed changes in central retinal thickness, tear production, glycaemic control and albuminuria, and no anti-cibinetide antibodies were detected. The investigators reported it as safe and inconclusive, which is the accurate description.

Wound healing in diabetes is the strongest of the rodent metabolic results. In db/db mice given 30 micrograms per kilogram subcutaneously daily, wound VEGF, phospho-Akt, phospho-eNOS and nitrite/nitrate rose while malondialdehyde fell, and re-epithelialisation, angiogenesis, wound breaking strength and time to closure all improved against vehicle across 3, 7 and 14 day timepoints.

The single most interesting pharmacological fact about ARA-290 is the mismatch between its exposure and its effect. Plasma half-life is roughly 2 minutes in rat and rabbit, against 4 to 6 hours for EPO and its carbamylated derivative. Yet in the rat spared nerve injury model, five injections given over ten days suppressed mechanical and cold allodynia for 20 weeks, and spinal microglial Iba-1 reactivity was still lower in treated animals at that point. In the human trials the metabolic and symptom improvements continued through a month of follow-up after administration stopped.

Collino, Thiemermann, Cerami and Brines reviewed the pharmacokinetic and pharmacodynamic data in 2015 and framed it as a switch rather than an occupancy: a brief receptor engagement during a window when the heteroreceptor is expressed changes the state of the tissue, and the tissue holds the new state. Mechanistically the candidate explanations are changes in myeloid cell phenotype and in transcriptional programmes with long persistence, not sustained receptor binding. For anyone designing an experiment this reverses the usual intuition. Continuous exposure is not obviously the goal, and endpoints measured while the peptide is still in circulation will miss most of what it does.

The clearance route is exopeptidase-driven. Incubating ARA-290 with EDTA, heparin and citrate plasma and with serum produced extensive metabolism with one main product, PyrEQLERALN, the peptide minus its two C-terminal serines. The blocked pyroglutamate N-terminus means degradation runs from the other end. That metabolite is also the analyte doping control laboratories look for, since ARA-290 is named on the prohibited list in sport.

Against erythropoietin, the comparison is the reason the molecule exists. EPO binds an EPO receptor homodimer with high affinity and drives erythroid maturation; ARA-290 has no meaningful activity there and drives nothing in a resting cell. EPO circulates for hours; ARA-290 for minutes. EPO's tissue protection is inseparable from raised haematocrit and platelet activation at protective doses; the peptide leaves haematocrit alone in every model where it has been measured. Carbamylated EPO sits between the two: also non-erythropoietic, also protective, but a full 165-residue glycoprotein rather than an 11-mer, with the immunogenicity and manufacturing profile that implies.

Against the other repair peptides in this catalogue, the difference is what is known about the receptor. BPC-157 and TB-500 are both supported by rodent injury models, and both have mechanisms that are described in terms of pathways rather than a defined binding partner. TB-500 is a fragment of Thymosin Beta-4 whose best-defined molecular action is sequestering monomeric G-actin. BPC-157 has no identified receptor at all. ARA-290 is the opposite case: the receptor is named, its two subunits are named, knocking out or knocking down either subunit removes the effect, and the tissue-selectivity argument follows from receptor expression rather than from assertion. What ARA-290 does not have is the volume of published rodent tendon and gastrointestinal work those two accumulated.

Against Thymosin Alpha-1, also in the Healing and Recovery group, the contrast is between two anti-inflammatory peptides working from opposite ends of the immune system. Thymosin Alpha-1 is an agonist at Toll-like receptors on antigen-presenting cells that pushes T cell differentiation. ARA-290 works on the myeloid side, damping macrophage NF-kB signalling and shifting monocyte phenotype, with no direct action on adaptive priming that has been demonstrated.

The peptide is short, strongly hydrophilic and free of cysteine, so the failure modes that dominate larger peptides do not apply. There is no disulfide bond, therefore no scrambling and no need to control redox conditions. Two glutamate side chains and a free C-terminal carboxyl against a single arginine make it net anionic at neutral pH and freely water soluble. A 10 mg vial holds about 7.95 micromoles, which is the number to work from when preparing a molar stock.

Two chemistry-specific points are worth knowing. The pyroglutamate N-terminus is a blocked amine: it resists aminopeptidases, and it also blocks Edman degradation, so N-terminal sequencing of this peptide requires pyroglutamate aminopeptidase treatment first. A laboratory that reports a failed Edman run on ARA-290 has confirmed the pyroglutamate, not a bad batch. Second, the Asn9-Ser10 motif near the C-terminus is a classic deamidation hotspot. Deamidation converts asparagine to a mixture of aspartate and isoaspartate and adds 1 Da, which is resolvable by high-resolution mass spectrometry and shows as a shoulder or a new early-eluting peak by reversed-phase HPLC. Warm, wet and neutral-to-alkaline conditions accelerate it, which is the real reason to keep lyophilised material cold, dry and dark rather than any concern about the peptide bond backbone.

For identity, the mass to confirm is 1,257.3 g/mol for the free peptide, with the formula C51H84N16O21. A certificate quoting about 1,274 has probably characterised the linear glutamine precursor rather than the cyclised pyroglutamate form, since cyclisation releases a molecule of ammonia and costs 17 Da. A mass near 1,083 is the exopeptidase metabolite PyrEQLERALN, the peptide stripped of both C-terminal serines. Purity by reversed-phase HPLC and identity by LC-MS are the two assays that matter; a purity figure without an accompanying mass tells you the material is homogeneous, not that it is the right molecule.

One practical warning about published listings. The correct sequence is pyroGlu-Glu-Gln-Leu-Glu-Arg-Ala-Leu-Asn-Ser-Ser. Sequences circulating in product listings for this compound are frequently wrong: some drop the glutamine at position 3 and print a 10-residue chain, and at least one prints an 11-character string containing residue codes that do not exist while quoting the correct 1,257.3 mass beside it. The mass and the formula are the reliable identifiers to check a document against.

ARA-290 holds no marketing authorisation in any jurisdiction. Four trials are registered on ClinicalTrials.gov under cibinetide or ARA 290: the phase 2b in sarcoidosis neuropathy (NCT02039687, 64 subjects, completed 2015), a phase 2 in prediabetes and type 2 diabetes (NCT01933529, 24 subjects, status unknown since), a phase 1/2 cognitive and neural study in depression (NCT02070783, 36 subjects, completed), and the terminated diabetic macular oedema study (NCT06626971, 9 subjects). No phase 3 has been registered, and no new registered trial of this compound has started since 2016.

The compound is prohibited in sport. It appears by name in doping control method development literature as a prohibited peptide under 2 kDa, and the metabolite PyrEQLERALN is the target analyte for urine screening by liquid chromatography and high-resolution mass spectrometry.

One paper in the preclinical file has been withdrawn. The 2012 Molecular Medicine report that delayed pHBSP attenuated acute kidney injury in rats was retracted on 26 February 2026 after the editor identified duplicated bands in the western blot figures and, on review of the original autoradiograms, concluded that a number of the published bands were not representative of the experimental results. Three of the eleven authors disagreed with the retraction and the remainder did not respond. The renal protection finding is replicated in other laboratories and other models, so the conclusion is not orphaned, but a citation to that specific paper should be dropped from any new work.

The other structural caveat is concentration of authorship. Most of the human data come from Leiden University Medical Center and the Cleveland Clinic, several papers share the same Araim Pharmaceuticals authors, and the corresponding competing-interest statements are declared in each. The independent replications are largely on the preclinical side, where groups in Innsbruck, Messina, Turin, Tel Aviv, Stockholm, Belfast, Beijing and Chongqing have reproduced pieces of the effect in colitis, wound healing, bone, islet transplantation, stroke and nerve crush models.