PRI-002
An orally active, all-D-amino-acid peptide designed to pull apart toxic amyloid-beta oligomers rather than attack plaques. Its 304-patient phase 2 trial, reported at CTAD on 17 November 2026, met its safety endpoint and missed its efficacy endpoint.
PRI-002 is a twelve-amino-acid peptide built entirely from mirror-image amino acids, which is why it survives the gut and can be swallowed as a capsule — almost unheard of for a peptide drug. Instead of clearing amyloid plaques the way the approved antibodies do, it is meant to break up small clumps of amyloid-beta called oligomers, which many researchers think do the actual damage. Its phase 2 trial in 304 people with early Alzheimer's reported on 17 November 2026. The safety half passed: the drug was well tolerated and, unlike the antibodies, was not expected to cause brain swelling. The efficacy half failed: on the main measure, CDR-SB at 48 weeks, it did not beat placebo. The company highlighted exploratory analyses at 72 weeks, where a cognitive scale showed a 35.9% slowing of decline with p=0.022 — but that is one result below 0.05 out of twelve reported comparisons, in a per-protocol subgroup at a secondary timepoint, which is about what chance alone would produce. Full data are promised for 2027.
What is PRI-002?
PRI-002 — also called RD2, and contraloid in its clinical formulation — is an investigational oral peptide for Alzheimer's disease. It is a twelve-residue peptide with the sequence ptlhthnrrrrr, written in lower case by convention because every residue is a D-amino acid, the mirror image of the L-amino acids that make up natural proteins, with an amidated C-terminus. That mirror-image construction is the entire reason the drug can be taken by mouth. Proteases are stereospecific: the enzymes that digest peptides in the gut, plasma and liver recognise L-amino acids and leave D-peptides largely alone. Laboratory work found RD2 to be resistant to degradation in simulated gastrointestinal fluids, blood plasma and liver microsomes, and not to be a substrate for human D-amino acid oxidase, with reported oral bioavailability above 75%. For a peptide drug — a class that is almost universally injected — that is unusual enough to be the molecule's defining property. The mechanism is also deliberately different from the approved amyloid drugs. Lecanemab and donanemab are antibodies that bind amyloid and recruit immune clearance of plaques, an approach that produces measurable plaque removal, modest clinical slowing, and amyloid-related imaging abnormalities — the brain swelling and microbleeds known as ARIA. PRI-002 is not an antibody and does not recruit immune clearance. It is designed to bind amyloid-beta monomers and thereby destabilise and disassemble the small soluble oligomers that many researchers regard as the species that actually damages synapses, converting them back into harmless monomers. Its developers describe the approach as an “antiprionic” or “detangler” strategy, by analogy with the self-propagating behaviour of prion proteins. It was developed at Forschungszentrum Jülich and Heinrich Heine University Düsseldorf in the group of Dieter Willbold, and is being taken through the clinic by Priavoid GmbH. The phase 2 trial was sponsored by PRInnovation GmbH, an affiliate of SPRIND, the German Federal Agency for Disruptive Innovation.
What PRI-002 Is Investigated For
PRI-002 has one indication and one result, and the result is a miss that the surrounding coverage has not always described as one. PRImus-AD was a randomised, double-blind, placebo-controlled phase 2 trial that enrolled 304 people aged 55 to 80 with mild cognitive impairment or mild dementia due to Alzheimer's disease, confirmed by cerebrospinal fluid or amyloid PET, at 38 sites in six European countries. Participants were randomised 1:1:1 to oral PRI-002 at 300 mg or 600 mg daily or placebo, and treated for 48 to 96 weeks under a common-close design in which the study ended when the last participant reached week 48. There were two primary endpoints. The safety endpoint was met. The efficacy endpoint — change from baseline in the Clinical Dementia Rating Sum of Boxes at week 48 — was not. The company's topline, presented at CTAD in Boston on 17 November 2026, emphasised exploratory analyses, and those deserve to be read as a set rather than one at a time. With the two doses pooled, twelve comparisons were reported across three scales, two timepoints and two analysis populations. Eleven of them had p-values between 0.080 and 0.775. One — ADAS-Cog13 at week 72 in the per-protocol population, a 35.9% slowing of decline — came in at p=0.022. In twelve comparisons with no multiplicity correction, one result below 0.05 is roughly what chance produces. The point estimates were consistent in direction, all favouring the drug, which is worth something; it is not worth a conclusion. The safety picture is the more interesting half, and it is not quite the clean story the mechanism predicts. The drug was well tolerated, with headache, nausea and dizziness the most common adverse events, and an independent safety board was comfortable enough to recommend ending intensive central MRI monitoring for ARIA after the first 90 patients reached 24 weeks. But ARIA was not absent. ARIA-E, the oedema form, occurred in 2% on 300 mg, 0% on 600 mg and 1% on placebo — no signal. ARIA-H, the haemorrhagic form, occurred in 8% on 300 mg and 11.7% on 600 mg against 5% on placebo. Those are small numbers in a 304-patient trial and ARIA-H is common in amyloid-positive populations regardless of treatment, but a non-antibody drug whose selling point is the absence of ARIA reporting roughly double the placebo rate of ARIA-H is a finding that should be followed rather than waved past. The company also reported preliminary p-tau181 data showing an increasing change from baseline relative to placebo from week 24 onward. The release does not state the direction clearly enough to interpret, and p-tau181 moving away from placebo could mean either thing. It should be treated as unreadable until the full data appear. Behind the phase 2 sit three phase 1 studies — single and multiple ascending dose work in healthy volunteers, and a 20-patient phase 1b in MCI and mild Alzheimer's at 300 mg daily for 28 days — and a preclinical record in three transgenic mouse models and in aged beagle dogs. Priavoid says the results will inform next steps including a possible further study, with the full dataset expected at conferences in 2027.
History & Discovery
The molecule began as an answer to a chemistry problem rather than a biology one. Peptides make poor drugs because proteases destroy them, and the brain is behind a barrier that large molecules cross badly. Mirror-image phage display, developed in the 1990s, offered a route around the first problem: screen a conventional phage library of L-peptides against a synthetic mirror-image version of the target, and the mirror images of whatever binds will bind the real, natural target — while being invisible to stereospecific proteases. Dieter Willbold's group, working across Forschungszentrum Jülich and Heinrich Heine University Düsseldorf, applied this to amyloid-beta(1-42) and identified D3, the all-D twelve-mer rprtrlhthrnr. Work through the 2010s characterised D3's behaviour, its pharmacokinetics and its surprising stability in the media an oral drug has to survive. RD2 — ptlhthnrrrrr — came from rationally rearranging D3's sequence rather than from a new screen, and proved both more efficient at eliminating amyloid-beta oligomers and more bioavailable. A 2017 study comparing the four most promising D3 derivatives in a large-scale oral treatment experiment helped select it, and the Scientific Reports paper of the same year established the signature result: improved cognition in APP/PS1 mice with no change in plaque load. The mouse work extended through models of pyroglutamate-amyloid-beta accumulation and of established pathology, and in 2023 to aged beagle dogs, a spontaneously declining non-transgenic model that is a far harder test than a transgenic mouse dosed before symptoms appear. Clinical development began in 2020 with phase 1 single- and multiple-ascending-dose studies in healthy volunteers, followed by a 20-patient phase 1b published in Nature Communications in 2025. The phase 2 PRImus-AD opened in February 2024, sponsored by PRInnovation GmbH, an affiliate of the German Federal Agency for Disruptive Innovation, completed in April 2026, and reported topline at CTAD on 17 November 2026: safety endpoint met, efficacy endpoint not met. What happens next is undecided. The company says the full analysis, including biomarkers and brain volumetrics, will inform next steps and may include a further study, with presentations planned for 2027.
How It Works
Amyloid-beta is a sticky fragment that clumps together in Alzheimer's disease. The large clumps are the plaques everyone has heard of, but many researchers think the small clumps — oligomers of a handful of molecules — are what actually poison the connections between brain cells. PRI-002 is a small peptide designed to grab loose amyloid-beta molecules and, by doing so, pull the small clumps apart into harmless single molecules. It does not try to clear plaques. Because it is built from mirror-image amino acids that digestive enzymes cannot break down, it can be swallowed rather than infused.
PRI-002 is the twelve-residue all-D peptide ptlhthnrrrrr with an amidated C-terminus. Its lineage starts with D3 (rprtrlhthrnr), identified by mirror-image phage display against amyloid-beta(1-42) — a technique in which a phage library of L-peptides is screened against the mirror image of the target, so that the mirror images of the hits bind the real target. RD2 was produced from D3 by rationally rearranging the amino acid sequence, and proved substantially more efficient at eliminating amyloid-beta oligomers in vitro than its parent, as well as having better bioavailability. The proposed mechanism is not aggregation inhibition in the conventional sense. The peptide is arginine-rich and binds amyloid-beta monomers; by sequestering monomers it shifts the equilibrium that holds oligomers together, so that existing oligomers disassemble into monomers rather than simply being prevented from forming. The developers call this an oligomer-eliminating or “detangler” mechanism and frame it as antiprionic, by analogy with the self-replicating behaviour by which small aggregates template further aggregation. Biophysical work has characterised the thermodynamics of D-peptide binding to amyloid-beta42 monomers and the effect on aggregation kinetics, and a cryo-EM and high-pressure literature has grown around the same question. The prediction this mechanism makes is distinctive and testable: cognitive benefit without plaque clearance. That is what the mouse work showed. In APP/PS1 mice, 28 days of treatment improved cognition and learning with no significant change in brain plaque load, though insoluble amyloid-beta42 fell significantly. The same pattern appeared in a model accumulating pyroglutamate-amyloid-beta, a particularly aggressive species, where oral treatment decelerated the neurodegenerative phenotype, and in a model with established full-blown pathology. The most interesting preclinical result is in aged beagle dogs — a non-transgenic model of genuinely sporadic Alzheimer's-like decline — where oral treatment reduced cognitive deficits and the benefit persisted beyond the end of dosing, alongside a treatment-dependent fall in cerebrospinal fluid tau oligomers. The second half of the mechanism is pharmacokinetic and is what makes the molecule unusual. D-amino acid residues are not substrates for stereospecific proteases, so RD2 resists degradation in simulated gastrointestinal fluids, plasma and liver microsomes across species, is not metabolised by human D-amino acid oxidase, and generates no human-specific metabolites. Reported oral bioavailability exceeds 75%, and human phase 1 work confirmed rapid absorption, dose-proportional exposure, about threefold accumulation on daily dosing and steady state within one to two weeks. What has never been demonstrated is target engagement in a human brain. There is no validated assay for amyloid-beta oligomer concentration in living patients, and the phase 1b found no significant change in cerebrospinal fluid oligomers, p-tau, t-tau, amyloid-beta40 or amyloid-beta42 after 28 days at 300 mg. A mechanism this specific that cannot be measured in the patients it is given to is a structural problem for the programme, not merely a missing dataset.
Evidence Snapshot
Human Clinical Evidence
Four trials, about 350 participants in total, and one missed primary endpoint. Phase 1 comprised a single-ascending-dose study (4 to 320 mg, 40 healthy volunteers) and a multiple-ascending-dose study (160 mg for 14 days or 320 mg for 28 days, 24 volunteers), both safe and well tolerated with dose-proportional exposure. The phase 1b randomised 20 patients with MCI or mild Alzheimer's to 300 mg daily or placebo for 28 days: all primary safety endpoints met, no ARIA, no significant change in any cerebrospinal fluid biomarker, and a better CERAD word-list performance on drug at day 56 (P ≤ 0.05) that was a secondary exploratory finding in nine treated patients. The phase 2 PRImus-AD randomised 304 patients to 300 mg, 600 mg or placebo for 48 to 96 weeks; the safety endpoint was met and the efficacy endpoint, CDR-SB change at week 48, was not.
Animal / Preclinical
The strongest part of the evidence base, and consistent across models and laboratories. RD2 improved cognition in APP/PS1 mice without reducing plaque load while significantly lowering insoluble amyloid-beta42; decelerated the neurodegenerative phenotype in TBA2.1 mice accumulating pyroglutamate-amyloid-beta; restored cognition in transgenic mice with full-blown pathology; and reduced cognitive deficits in aged beagle dogs, a non-transgenic model of sporadic disease, with benefit persisting beyond treatment end and a fall in cerebrospinal fluid tau oligomers. The developers note efficacy in three transgenic mouse models across three different laboratories, which is a stronger replication record than most preclinical Alzheimer's programmes can claim.
Mechanistic Rationale
Coherent and well characterised biophysically, but resting on a hypothesis that remains unsettled. The oligomer hypothesis — that soluble amyloid-beta oligomers rather than plaques drive synaptic damage — is widely held and not established, and PRI-002 is among its most direct tests. The peptide's stability and oral bioavailability are genuinely demonstrated. What is missing is the middle link: no assay exists to show that the drug eliminates oligomers in a living human brain, and the one trial that looked for cerebrospinal fluid biomarker change found none.
Research Gaps & Open Questions
What the current literature has not yet settled about PRI-002:
- 01Target engagement has never been demonstrated in a human being. There is no validated assay for amyloid-beta oligomer concentration in a living brain, and the phase 1b found no significant change in cerebrospinal fluid oligomers after 28 days at 300 mg. A mechanism this specific that cannot be measured where it is supposed to act leaves every clinical result ambiguous.
- 02The full phase 2 dataset is unpublished. The adverse-event tables, the complete biomarker panel, the brain volumetric analysis and — most importantly — the dose-by-dose breakdown are all unavailable, because the topline pooled 300 mg and 600 mg for every reported comparison.
- 03The p-tau181 result, described as “an increasing change from baseline relative to placebo” from week 24, is reported too ambiguously to interpret in either direction.
- 04The ARIA-H imbalance — 8% and 11.7% against 5% on placebo — is unexplained and sits awkwardly against a mechanism that predicts no ARIA at all. Whether it reflects chance in small numbers, amyloid-related vascular pathology in the enrolled population, or something the drug does is unknown.
- 05Why the primary endpoint failed is undiagnosed. Underpowering, a 48-week window too short to show disease modification, the wrong dose, the wrong patient population, or a mechanism that does not operate in humans are all live and mutually compatible explanations.
- 06The exploratory week-72 results came from a common-close design in which only some participants were treated that long, so the week-72 population is a self-selected subset — the earliest-enrolled patients — and no analysis accounting for that has been published.
- 07Whether the food effect seen in mice applies to humans has not been tested, despite its potential to produce several-fold differences in exposure.
- 08No head-to-head or add-on comparison with an approved anti-amyloid antibody has ever been attempted, so where an oral oligomer-directed drug would sit relative to the approved standard is entirely unknown.
Forms & Administration
PRI-002 is an oral capsule, which is the single most unusual thing about it in a field where every approved amyloid-directed drug is infused. In the phase 2 trial participants took three capsules in the morning and three in the evening, delivering 300 mg or 600 mg per day, or matching placebo. The phase 1b used 300 mg once daily. No injectable, intranasal or other formulation has been developed, and none would make sense given that oral bioavailability above 75% is the molecule's principal pharmaceutical advantage. One administration detail is unresolved. Mouse pharmacokinetic work published in 2026 found a substantial food effect: at 200 mg/kg, plasma levels were about fifteen times higher in fasted than in fed animals, though the effect was minor at 600 mg/kg. The authors note this may apply to patients and is of interest for further clinical development. Whether and how food should be managed around dosing has not been settled in humans.
Dosing & Protocols
The ranges below reflect protocols commonly discussed in the literature and by clinicians — not a prescription. Actual dosing for any individual should be determined by a qualified healthcare provider who knows the patient.
Typical Range
300 mg or 600 mg per day orally in the phase 2 trial; 300 mg once daily in the phase 1b
Frequency
Once or twice daily. The phase 2 gave three capsules in the morning and three in the evening; the phase 1b gave 300 mg once daily
Timing Considerations
No specific timing requirements: can be administered at any time of day, with or without food, and is not tied to exercise timing. Consistency matters more than the specific clock — dose at roughly the same time each day (or same day each week, for weekly protocols) to keep exposure steady.
Protocol Notes
These are research doses from an investigational drug with no approved label, reported for completeness rather than as guidance. The dose range tested in humans is well characterised for safety. Single doses ran from 4 mg to 320 mg in healthy volunteers without issue, multiple dosing covered 160 mg for 14 days and 320 mg for 28 days, the phase 1b used 300 mg daily for 28 days, and the phase 2 ran 300 mg and 600 mg daily for 48 to 96 weeks. Exposure is dose-proportional, accumulation on daily dosing is about threefold, and steady state arrives within one to two weeks. What the phase 2 does not tell anyone is whether the two doses differed in effect, because the topline pooled them for every reported analysis. That is a meaningful gap: a dose-response relationship, or its absence, is one of the more informative things a failed trial can report, and it will have to wait for the full dataset. The open practical question is food. Mouse pharmacokinetics published in 2026 found plasma levels roughly fifteen times higher in fasted than in fed animals at 200 mg/kg, with the effect much smaller at 600 mg/kg. The authors explicitly flag this as potentially applying to patients. Whether PRImus-AD specified fasted or fed dosing, and whether that contributed to variability in exposure, is not public.
PRI-002 is not approved by any regulator, missed the primary efficacy endpoint of its only phase 2 trial, and is not lawfully available. Approved treatments for early Alzheimer's disease exist and are the appropriate subject of a conversation with a neurologist.
Common Questions
Did PRI-002 work?
No, on the measure the trial was built to answer. PRImus-AD had two primary endpoints. The safety endpoint was met. The efficacy endpoint — change in the Clinical Dementia Rating Sum of Boxes at week 48 — was not met, and the company's own release says so. Everything positive that has been reported comes from exploratory analyses at week 72, in subgroups, with the doses pooled. The strongest of those, a 35.9% slowing on ADAS-Cog13 in the per-protocol population at week 72, had p=0.022, and it is one of twelve reported comparisons of which the other eleven ranged from p=0.080 to p=0.775. That is the pattern chance produces. The right summary is a failed trial with a direction of effect that was consistently, non-significantly favourable — which is a reason to look at the full data in 2027, not a reason to conclude the drug works.
How can a peptide be taken as a pill?
By being built from the wrong amino acids on purpose. Natural proteins and peptides are made of L-amino acids, and the proteases that digest them in the stomach, gut, plasma and liver are stereospecific — they recognise that geometry and nothing else. PRI-002 is made entirely of D-amino acids, the mirror images, which those enzymes largely cannot process. Laboratory studies found it stable in simulated gastric and intestinal fluids, in plasma and in liver microsomes from several species, and not a substrate for human D-amino acid oxidase, with reported oral bioavailability above 75%. The phase 1 work in healthy volunteers confirmed rapid absorption, dose-proportional exposure, roughly threefold accumulation on daily dosing, and steady state in one to two weeks. One practical wrinkle: mouse pharmacokinetics show a large food effect at lower doses, with plasma levels about fifteen times higher fasted than fed, which could matter for how it would eventually be taken.
How is it different from lecanemab or donanemab?
Three ways that all follow from it not being an antibody. First, route: PRI-002 is an oral capsule, while lecanemab and donanemab are intravenous infusions. Second, target: the antibodies bind amyloid aggregates and recruit immune clearance of plaques, while PRI-002 binds amyloid-beta monomers to destabilise soluble oligomers into monomers, and explicitly does not aim to clear plaques — in mice it improved cognition with no significant change in plaque load. Third, the expected safety trade: ARIA arises from antibody-mediated clearance of amyloid from vessel walls, and a non-antibody drug has no mechanism to produce it, which is the main safety argument for this approach. The phase 2 complicates that argument slightly, with ARIA-H reported at 8% and 11.7% against 5% on placebo. The decisive difference, though, is efficacy: lecanemab and donanemab are approved on the basis of positive phase 3 trials, and PRI-002 missed its phase 2 primary endpoint.
What is the oligomer hypothesis and does this trial say anything about it?
The hypothesis is that amyloid plaques are relatively inert and the real damage is done by small soluble aggregates of amyloid-beta — oligomers — that disrupt synapses. PRI-002 is one of the more direct tests of it, because it was designed to eliminate oligomers specifically and the mouse work showed cognitive improvement with no reduction in plaque load, which is the signature the hypothesis predicts. A clear positive phase 2 would have been meaningful support for it. A missed primary endpoint is weaker evidence against: the trial may have been underpowered for the effect size, 48 weeks may be too short, the dose may have been wrong, or the drug may not be reaching and eliminating oligomers in human brain at all. Measuring brain oligomer concentrations in living patients remains the hard unsolved problem, and without it the mechanism cannot be confirmed or excluded from a clinical result.
Is PRI-002 available anywhere?
No. It is not approved by the FDA, the EMA or any other regulator, and it has never been. The phase 2 trial completed in April 2026 and there is no open study to enrol in. Approved treatments for early Alzheimer's exist — lecanemab and donanemab, both infused, both with positive phase 3 data — and are the relevant conversation with a neurologist. Anything sold as RD2 or contraloid outside a trial should be assumed not to be the drug: an all-D twelve-mer is a specialist synthesis and there is no legitimate supply chain for it.
Who is behind it?
The science came from Forschungszentrum Jülich and Heinrich Heine University Düsseldorf in Dieter Willbold's group, where the parent compound D3 was found by mirror-image phage display against amyloid-beta(1-42) and RD2 was derived from it by rationally rearranging the sequence. Clinical development is run by Priavoid GmbH. The phase 2 trial was sponsored by PRInnovation GmbH, an affiliate of SPRIND, the German Federal Agency for Disruptive Innovation — public funding rather than venture capital, which is unusual for a phase 2 Alzheimer's trial and worth knowing when reading the company's framing of a missed endpoint.
Who PRI-002 Is NOT For
- •No clinical contraindication list exists, because the drug is unapproved and has never had a label — what follows derives from trial exclusion criteria and from the limits of what has been studied
- •Dementia of moderate or greater severity — every trial enrolled mild cognitive impairment or mild dementia only, with amyloid confirmed by cerebrospinal fluid or PET
- •Age outside 55 to 80, the phase 2 enrolment window, and 50 to 80 in the phase 1b
- •Cognitive impairment not due to Alzheimer's disease — the trials required biomarker confirmation, and nothing is known about other dementias
- •Pregnancy and breastfeeding — no data
- •Children and adolescents — never studied in any age group below 50
- •Concurrent anti-amyloid antibody therapy, which has never been combined with this drug and whose ARIA risk profile would complicate interpretation of the haemorrhagic imaging findings already seen
Drug & Supplement Interactions
No drug-interaction studies have been published for PRI-002, and its pharmacology suggests few would be expected. The molecule is metabolically inert in the ways that generate conventional interactions. Laboratory work found it resistant to degradation in plasma and in liver microsomes from human, rodent and non-rodent species, not a substrate for human D-amino acid oxidase, and — importantly for interaction risk — it did not influence the activity of any of the enzymes tested. A peptide that is neither metabolised by nor inhibitory toward the hepatic enzyme systems is unlikely to participate in cytochrome P450 interactions in either direction. The theoretical interactions that remain are pharmacodynamic and untested. Combination with an anti-amyloid antibody is the obvious one, and nobody has tried it; given that both act on amyloid-beta by different routes, and that the phase 2 reported ARIA-H at 8% and 11.7% against 5% on placebo, a combination would need careful imaging surveillance before anything else. The practical interaction with the best evidence behind it is not a drug at all but food. Mouse pharmacokinetics show a large fasted-versus-fed difference in plasma exposure at lower doses, which has not been characterised in humans and which would be the first thing a dosing instruction would need to settle. Standard Alzheimer's co-medications — cholinesterase inhibitors, memantine — were presumably permitted in the trials as background therapy, but no interaction analysis has been reported.
Safety Profile
Common Side Effects
Cautions
- • Investigational everywhere — not approved by any regulator, with no open trial to enrol in as of November 2026
- • The phase 2 trial missed its primary efficacy endpoint; nothing here supports clinical use
- • The ARIA-H imbalance deserves attention rather than dismissal, precisely because the mechanism predicts there should be none
- • Mouse pharmacokinetics show a substantial food effect, with plasma levels roughly fifteen times higher fasted than fed at lower doses — relevant to any future dosing instruction
- • Total human exposure is about 350 people across all four trials, with the longest treatment at 48 to 96 weeks in one study
What We Don't Know
The full phase 2 dataset has not been released; Priavoid says it will present it at scientific conferences in 2027. Until then the adverse-event tables, the biomarker results in full, the brain volumetric data and the dose-by-dose efficacy breakdown are unavailable — the topline pooled the two doses, so whether 300 mg and 600 mg behaved differently is not public. The p-tau181 observation reported as “an increasing change from baseline relative to placebo” from week 24 is not stated clearly enough to interpret in either direction. Most fundamentally, there is no validated way to measure amyloid-beta oligomer concentrations in a living human brain, so whether the drug engages its target in patients has never been demonstrated — the phase 1b found no significant change in cerebrospinal fluid amyloid-beta oligomers, among other biomarkers, after 28 days.
Legal Status
United States
Investigational and never studied in the United States. PRI-002 has no FDA approval, no FDA-reviewed application, and does not appear in any FDA approved-drug database. All four of its clinical trials ran in Europe. It is not a controlled substance and not eligible for compounding. There is no lawful route of access in the United States, and no open trial to enrol in as of November 2026.
International
Not approved in any jurisdiction. Development has been European throughout: the phase 1 trials and the 20-patient phase 1b ran in Germany, and PRImus-AD ran at 38 sites across six European countries under EU clinical trial regulation, registered as EudraCT 2020-003416-27 for the phase 1b and NCT06182085 for the phase 2. No marketing application has been submitted anywhere.
Sports & Competition
Not a sports-doping concern. PRI-002 has no anabolic, stimulant, hormonal or oxygen-carrying activity and no conceivable performance application; it is a cognition-directed peptide that failed its efficacy endpoint. It does not appear on the WADA Prohibited List and there is no category it would plausibly fall under. The generic caution applies: a substance sold under this name outside a trial is not this substance, and unknown material is a strict-liability risk for any tested athlete.
Regulatory status changes over time. Verify current local rules with a qualified professional.
Myths & Misconceptions
Myth
PRI-002 showed a statistically significant 35.9% slowing of cognitive decline.
Reality
That number is real, and it is one of twelve comparisons the company reported. The other eleven had p-values from 0.080 to 0.775. The 35.9% figure is ADAS-Cog13 at week 72 in the per-protocol population with both doses pooled — an exploratory analysis, at a secondary timepoint, in a selected subgroup, on a secondary scale, with no correction for multiplicity. One result below 0.05 out of twelve is about what chance delivers. The trial's registered primary efficacy endpoint, CDR-SB at week 48, was not met.
Myth
Because it is not an antibody, PRI-002 cannot cause ARIA.
Reality
That is the mechanistic prediction, and the phase 2 data complicate it. ARIA-E, the oedema form that antibody clearance of vascular amyloid produces, showed no signal: 2%, 0% and 1% for 300 mg, 600 mg and placebo. ARIA-H, the haemorrhagic form, was reported at 8% and 11.7% against 5% on placebo. Those are small numbers, ARIA-H occurs in amyloid-positive people without any treatment, and an imbalance this size in a 304-patient trial could be chance. But it is a finding to follow, not one to discount because the mechanism says it shouldn't happen.
Myth
An oral peptide is impossible, so this must not really be a peptide.
Reality
It is really a peptide — twelve amino acid residues, ptlhthnrrrrr — and it really is oral, with reported bioavailability above 75%. The trick is that every residue is a D-amino acid, the mirror image of the natural form. Digestive and plasma proteases are stereospecific and largely cannot process them, and human D-amino acid oxidase does not act on this sequence either. The result is a peptide that behaves pharmacokinetically more like a small molecule: rapid absorption, dose-proportional exposure, steady state in one to two weeks.
Myth
It is an amyloid drug, and amyloid drugs have all failed.
Reality
Both halves of that need qualifying. Amyloid-directed drugs have not all failed — lecanemab and donanemab are approved on positive phase 3 trials, with modest effects and real risks. And PRI-002 targets a different species: not plaques, which it makes no attempt to clear, but soluble oligomers. In mice it improved cognition with no change in plaque load. The failure of its phase 2 primary endpoint is a fact about this drug and this trial, not a verdict on the oligomer hypothesis, which it was not adequately powered to settle either way.
Myth
RD2, contraloid and PRI-002 are different compounds.
Reality
They are three names for the same molecule. RD2 is the laboratory designation used throughout the preclinical literature from Willbold's group; contraloid (as contraloid acetate) is the name used in clinical and regulatory contexts; PRI-002 is Priavoid's development code. A 2026 pharmacokinetics paper states the identity directly. Searching under only one of the three names will miss most of the published evidence.
Published Research
14 studiesPharmacokinetic properties of the clinical drug candidate PRI-002 with regard to genotype, sex, age, dose-dependence and food effects in mice.
Oral PRI-002 treatment in patients with MCI or mild AD: a randomized, double-blind phase 1b trial.
Kutzsche, Cosma, Kauselmann and colleagues, Nature Communications, May 2025 — the study that justified the phase 2, and a good illustration of how a small positive signal can look. 20 patients aged 50 to 80 with MCI or mild dementia due to Alzheimer's, 19 randomised to 300 mg daily (n=9) or placebo (n=10) for 28 days. All primary safety endpoints met; 16 adverse events on drug against 27 on placebo; no serious adverse events; ECG, EEG and MRI unchanged, with no ARIA. Crucially for the mechanism, no significant change was detected in cerebrospinal fluid p-tau, t-tau, amyloid-beta40, amyloid-beta42 or amyloid-beta oligomers. Patients on PRI-002 performed significantly better on the CERAD word list at day 56 (P ≤ 0.05) — an exploratory secondary finding in nine treated patients, and the kind of result that is worth pursuing and incapable of settling anything.
Oral treatment with the all-d-peptide RD2 enhances cognition in aged beagle dogs - A model of sporadic Alzheimer's disease.
Kutzsche, Schemmert, Bujnicki and colleagues, Heliyon, August 2023 — the strongest preclinical study in the programme, because of the model rather than the effect size. Transgenic mice develop amyloid pathology on a predictable schedule, which makes preventive dosing easy and curative dosing nearly untestable; cognitively impaired aged beagles decline spontaneously, as humans do. Oral RD2 significantly reduced cognitive deficits in these dogs, the benefit persisted beyond the end of treatment, and cerebrospinal fluid tau oligomers fell in a treatment-dependent way. The authors read the combination as evidence of disease modification. It remains an animal study, and the phase 2 that followed missed its endpoint.
Interaction of Therapeutic d-Peptides with Aβ42 Monomers, Thermodynamics, and Binding Analysis.
Safety and pharmacokinetics of the orally available antiprionic compound PRI-002: A single and multiple ascending dose phase I study.
Toward the Mode of Action of the Clinical Stage All-d-Enantiomeric Peptide RD2 on Aβ42 Aggregation.
Do We Need Anti-Prion Compounds to Treat Alzheimer's Disease?
Metabolic resistance of the D-peptide RD2 developed for direct elimination of amyloid-β oligomers.
Deceleration of the neurodegenerative phenotype in pyroglutamate-Aβ accumulating transgenic mice by oral treatment with the Aβ oligomer eliminating compound RD2.
Aβ Oligomer Elimination Restores Cognition in Transgenic Alzheimer's Mice with Full-blown Pathology.
The Aβ oligomer eliminating D-enantiomeric peptide RD2 improves cognition without changing plaque pathology.
van Groen, Schemmert, Brener and colleagues, Scientific Reports, November 2017 — the paper that gives the sequence, the design rationale and the mechanism's signature finding. RD2 (ptlhthnrrrrr, amidated) was produced by rearranging the sequence of D3 (rprtrlhthrnr), itself found by mirror-image phage display against amyloid-beta(1-42), and eliminated amyloid-beta oligomers in vitro significantly more efficiently than its parent. After 28 days of treatment, APP/PS1 mice showed enhanced cognition and learning with no significant change in brain plaque load but a significant reduction in insoluble amyloid-beta42. Cognitive benefit without plaque change is precisely the pattern the oligomer hypothesis predicts, which is why this result carries more weight for the mechanism than the efficacy numbers do.
Priavoid Announces Topline Phase 2 Data for Oral All-D-Peptide PRI-002 in Alzheimer's Disease at CTAD 2026
Priavoid, 17 November 2026, presented at CTAD in Boston — the source for every phase 2 figure here, and a document worth reading against itself. The headline frames “promising signals”; the body states that the study did not meet its primary efficacy endpoint. 304 patients aged 55 to 80 with MCI or mild dementia due to Alzheimer's at 38 sites in six European countries, randomised 1:1:1 to 300 mg, 600 mg or placebo for 48 to 96 weeks under a common-close design. Twelve exploratory comparisons are reported across CDR-SB, ADAS-Cog13 and ADCS-ADL at weeks 48 and 72 in mITT and per-protocol populations, with the doses pooled; eleven had p-values from 0.080 to 0.775, and one — ADAS-Cog13 at week 72, per protocol, 35.9% slowing — reached p=0.022. ARIA-E 2%/0%/1% and ARIA-H 8%/11.7%/5% for 300 mg, 600 mg and placebo. Full dataset promised for 2027.
Study to Assess Safety and Efficacy of PRI-002 in Patients With MCI to Mild Dementia Due to Alzheimer's Disease (PRImus-AD)
The registry record, which confirms what the trial was designed to show before anyone knew what it did show. Two primary endpoints are registered: safety and tolerability by incidence of drug-related adverse events from baseline to week 48, and efficacy by change in the Clinical Dementia Rating Sum of Boxes from baseline to week 48. Sponsored by PRInnovation GmbH, 304 participants actual, phase 2, started 15 February 2024, completed 15 April 2026, status COMPLETED. Everything reported at week 72 is outside the registered primary analysis.
PRImus-AD study: Design of the phase 2 study treating patients with MCI or mild dementia due to Alzheimer's disease with the orally available PRI-002
Quick Facts
- Class
- All-D-Enantiomeric Amyloid-Beta Oligomer Disassembler
- Tier
- C
- Evidence
- Preliminary
- Safety
- Moderate Data
- Updated
- Nov 2026
- Citations
- 14PubMed
Also known as
Tags
Related Goals
Evidence Score
Clinical Trials
View Clinical TrialsLinks to ClinicalTrials.gov for reference. Listing does not imply endorsement.