COR388 · program

Atuzaginstat (COR388) for Alzheimer's disease

DiscontinuedDiscontinuedCortexyme, Inc. (now Quince Therapeutics, Inc.) (QNCX)

Atuzaginstat (COR388) for mild-to-moderate Alzheimer's disease dementia — Cortexyme's lead programme and the first large controlled test of the gingipain hypothesis. FAILED AND DISCONTINUED. The pivotal Phase 2/3 GAIN trial (NCT03823404, COR388-010; GingipAIN Inhibitor for Treatment of Alzheimer's Disease) randomised 643 patients with NIA-AA probable AD dementia (MMSE 12-24) to 40 mg BID, 80 mg BID or placebo for 48 weeks across 93 sites in the US and Europe, with co-primary endpoints of change from baseline in ADAS-Cog11 and ADCS-ADL. Topline on 2021-10-26: NEITHER co-primary was met in the overall cohort. In the pre-specified subgroup with P. gingivalis DNA detectable in saliva at baseline the 80 mg BID arm showed 57% slowing of decline on ADAS-Cog11 (p=0.02) and the 40 mg arm 42% (p=0.07), with no benefit on ADCS-ADL, and reductions in salivary P. gingivalis at week 24 correlated with better ADAS-Cog11, CDR-SB and MMSE outcomes — a mechanistically coherent but secondary finding. Safety drove the outcome: dose-related ALT/AST elevations >3x ULN occurred in 7% (40 mg) and 15% (80 mg) versus 2% on placebo, two 80 mg patients had concomitant bilirubin elevations without alternative explanation, and dropout was 40% on drug versus 25% on placebo. The FDA had already imposed a PARTIAL clinical hold on 2021-02-15 that stopped enrolment and dosing in the open-label extension (the double-blind phase was allowed to finish), and on 2022-01-25 imposed a FULL clinical hold on IND 134303. Cortexyme cut headcount by 67%, pivoted to the second-generation inhibitor COR588, agreed to acquire Novosteo, and renamed itself Quince Therapeutics on 2022-08-01 — announcing at the same time its intent to out-license its legacy neuroscience assets. On 2023-01-27 the whole protease-inhibitor portfolio, atuzaginstat included, was sold to Lighthouse Pharmaceuticals. Neither hold was ever publicly lifted and no further atuzaginstat trial has been registered. The programme's scientific legacy is the design of the second-generation Kgp inhibitor LHP588 (COR588), which carries the P. gingivalis-positive enrichment strategy and an improved liver-safety profile into a Phase 2 study in Pg-positive AD.

Data current as of

Development timeline

DiscontinuedAug 2022
  1. Programme discontinued by the sponsor. On 2022-08-01 the renamed company (Quince Therapeutics, Nasdaq: QNCX) announced a strategic shift to a bone-targeting platform and disclosed 'the company's intent to out-license its legacy neuroscience and antiviral assets' — the first unambiguous, company-stated exit from the atuzaginstat programme. This is the discontinuation anchor; the earlier 2022-01-25 full clinical hold is logged separately because that announcement still described continued Alzheimer's development (of COR588). Cause chain: dose-dependent hepatotoxicity plus a missed co-primary endpoint. The compound was sold to Lighthouse Pharmaceuticals on 2023-01-27.↗
Phase 2/3Mar 2019 – Jan 2022
  1. FDA FULL clinical hold. Cortexyme received an FDA letter on 2022-01-25 placing a full clinical hold on atuzaginstat's IND 134303 (announced 2022-01-26), following the dose-dependent hepatotoxicity signal. The company immediately began a cost-reduction programme, said it would prioritise the next-generation gingipain inhibitor COR588 in Alzheimer's disease instead, and would 'explore strategic alternatives for its coronavirus program and non-Alzheimer's indications for COR388'. Recorded as 'suspended' rather than 'discontinued' because at this date the company had not yet stated it was abandoning the asset. The hold was never publicly lifted.↗
  2. mixedAdditional GAIN top-line results presented at CTAD 2021 (Boston, 9-12 November 2021): expanded P. gingivalis-infection subgroup analyses reinforcing the ~50% slowing of cognitive decline in infected participants across multiple pre-specified infection-related subgroups and analysis methods, plus periodontal sub-study and safety detail.↗
  3. missedGAIN Phase 2/3 topline MISSED both co-primary endpoints: atuzaginstat 40 mg and 80 mg BID showed no significant benefit on ADAS-Cog11 or ADCS-ADL at Week 48 in the overall 643-patient mild-to-moderate Alzheimer's cohort. A pre-specified P. gingivalis-saliva-positive subgroup (n=242) showed 57% slowing of cognitive decline at 80 mg BID (p=0.02), with no ADCS-ADL benefit and dose-related liver enzyme elevations up to 15%.↗
  4. FDA PARTIAL clinical hold, imposed after its review of hepatic adverse events in the atuzaginstat trial. Scope was limited to the open-label extension: no new OLE enrolment and all enrolled OLE participants discontinued. The fully enrolled (N=643) double-blind, placebo-controlled phase was explicitly allowed to continue to its Q4 2021 topline, so the programme stays 'active' at this step rather than 'suspended'.↗
  5. Pivotal Phase 2/3 GAIN trial (NCT03823404, COR388-010) started per ClinicalTrials.gov: randomised, quadruple-masked, placebo-controlled, 48 weeks of 40 mg or 80 mg BID in mild-to-moderate AD dementia, co-primary endpoints ADAS-Cog11 and ADCS-ADL, with an open-label extension and a periodontal-disease sub-study.
Phase 1Dec 2017
  1. First-in-human. Phase 1 single ascending dose study COR388-001 (NCT03331900) started, 34 healthy adults, 5-250 mg oral capsules; completed 2018-04-02. Followed by the multiple ascending dose study COR388-002 (NCT03418688, started 2018-03-06, completed 2018-10-15) in 33 healthy older volunteers and patients with Alzheimer's disease. Status recorded as 'completed' because both Phase 1 studies did complete; no severe adverse events, drug-related TEAE rates 14% on COR388 vs 20% on placebo, and no dose-limiting toxicity.

Readouts

  • 2021-11-11ReportedFull resultsmixedNCT03823404

    Additional GAIN top-line results presented at CTAD 2021 (Boston, 9-12 November 2021): expanded P. gingivalis-infection subgroup analyses reinforcing the ~50% slowing of cognitive decline in infected participants across multiple pre-specified infection-related subgroups and analysis methods, plus periodontal sub-study and safety detail. ↗

  • 2021-10-26ReportedTopline datamissedNCT03823404

    GAIN Phase 2/3 topline MISSED both co-primary endpoints: atuzaginstat 40 mg and 80 mg BID showed no significant benefit on ADAS-Cog11 or ADCS-ADL at Week 48 in the overall 643-patient mild-to-moderate Alzheimer's cohort. A pre-specified P. gingivalis-saliva-positive subgroup (n=242) showed 57% slowing of cognitive decline at 80 mg BID (p=0.02), with no ADCS-ADL benefit and dose-related liver enzyme elevations up to 15%. ↗

Clinical trials in Alzheimer's disease

NCT03823404COR388-010Phase 2/3Completedn=643

GAIN Trial: A Randomized, Double-Blind, Placebo-Controlled Study of COR388 in Subjects With Alzheimer's Disease

Started Mar 2019· Primary completion Jan 2022· 📍 93 sites across 6 countries (United States, United Kingdom, Spain, Poland)

missedprimaryADAS-Cog11 total score, change from baseline to Week 48 (overall ITT cohort)

Co-primary cognitive endpoint NOT met. The 643-patient overall cohort showed no statistically significant difference from placebo on ADAS-Cog11 at the end of the 48-week treatment period. Effect size and p-value are deliberately null: the sponsor's topline release reported no numbers for the overall cohort, and the ClinicalTrials.gov results posting (2023-02-23) carries a malformed analysis row ('p= <0.0455') with no effect estimate or confidence interval, which is not reportable.

missedprimaryADCS-ADL total score, change from baseline to Week 48 (overall ITT cohort)

Co-primary functional endpoint NOT met in the overall cohort, and — importantly — also not met in the P. gingivalis-positive subgroup where the cognitive signal appeared: 'Significant benefits in this subgroup were not seen on the other co-primary, ADCS-ADL.' The functional miss is the reason the subgroup cognitive finding could not carry the programme. Numbers null for the same registry/disclosure reasons as ADAS-Cog11.

missedsafetyHepatic safety — ALT/AST elevations >3x upper limit of normal

Dose-related liver enzyme elevations >3x ULN: 2% on placebo, 7% on 40 mg BID, 15% on 80 mg BID. Two 80 mg BID participants had concomitant bilirubin elevations without alternative explanation. The sponsor characterised the elevations as asymptomatic and reversible (resolving on drug or after withdrawal). Alzforum's independent write-up of the same disclosure adds figures the press release omitted: overall dropout of 40% in the atuzaginstat arms versus 25% on placebo, transaminase-related discontinuations of 1 (placebo), 5 (40 mg) and 17 (80 mg), and six deaths (5 high dose, 1 low dose), all adjudicated unrelated — see https://www.alzforum.org/therapeutics/atuzaginstat, declared in sources. This signal produced the FDA partial hold (2021-02-15) and then the full hold (2022-01-25) and is the proximate cause of the programme's termination. Most common non-hepatic AEs were gastrointestinal: diarrhoea up to 16% and nausea 6% on drug vs 3% and 2% on placebo. No increase in ARIA, microhaemorrhage, oedema or superficial siderosis.

metpharmacodynamicSalivary P. gingivalis DNA reduction at Week 24 vs clinical outcomes at end of treatment (target-engagement correlation) (0.0007)

Reductions in salivary P. gingivalis DNA at Week 24 were significantly correlated with improved end-of-treatment outcomes on ADAS-Cog11 (p=0.0007), CDR-Sum of Boxes (p=0.004) and MMSE (p=0.007), with a beneficial trend on ADCS-ADL (p=0.08). p_value column carries the ADAS-Cog11 correlation. This is the pharmacodynamic evidence that the drug engaged its bacterial target in humans; it is a correlation within treated patients, not a randomised comparison.

mixedefficacy_subgroupADAS-Cog11, pre-specified subgroup with P. gingivalis DNA detectable in saliva at baseline (PG-DS) (0.02)

In the pre-specified PG-DS subgroup (n=242 per the sponsor's release; n=244 per the 2026 ASCP abstract) atuzaginstat slowed cognitive decline on ADAS-Cog11 by 57% in the 80 mg BID arm (p=0.02; p=0.020 in the ASCP abstract) and by 42% in the 40 mg BID arm (p=0.07; p=0.066 in the ASCP abstract, i.e. not significant), a dose-response. Recorded as 'mixed', not 'met': this is a subgroup on one of two co-primaries, with no accompanying ADCS-ADL benefit. effect_size is null because the 57% figure is a percentage slowing of decline, not a scale-unit treatment difference, and no LS-mean difference was disclosed.

mixedexploratoryCSF p-tau181, mean percent change from baseline to Week 48 (pre-specified exploratory) (0.048)

Post hoc-reported CSF biomarker analysis presented at the 2026 ASCP Annual Meeting, ~4.5 years after topline. CSF p-tau181 rose from baseline to Week 48 in both the ITT and PG-DS placebo groups. That rise was significantly inhibited by atuzaginstat 40 mg BID in the ITT population (p=0.048, n=79), with a trend in the PG-DS subgroup (p=0.083, n=34). The 80 mg BID arms had fewer evaluable subjects and reached significance in neither ITT (p=0.174, n=65) nor PG-DS (p=0.778, n=20). No significant longitudinal change was seen for CSF Abeta42/40 or total tau. Recorded as 'mixed': significant only at the lower dose, in a pre-specified exploratory analysis with small evaluable Ns and no dose-response. This result informed the enrichment design of the LHP588 SPRING trial.

NCT03418688COR388-002Phase 1Completedn=33

A Randomized, Double-Blind, Placebo-Controlled, Multiple Ascending Dose Study of COR388 in Older Healthy Volunteers and Patients With Alzheimer's Disease

Started Mar 2018· Primary completion Oct 2018· 📍 3 sites across 1 country (United States)

NCT03331900COR388-001Phase 1Completedn=34

A Phase 1 Single Ascending Dose Study of COR388 HCl

Started Dec 2017· Primary completion Feb 2018

Formulations

FormulationRouteRegimenPharmacokinetics
Atuzaginstat oral capsule (40 mg / 80 mg twice daily)
Phase 2/3 GAIN (NCT03823404): COR388 HCl 40 mg or 80 mg oral capsule twice daily, approximately 12 hours apart and no less than 6 hours apart, for 48 weeks. The (halted) open-label extension used the same 40 mg or 80 mg BID for a further 48 weeks. Phase 1 single ascending dose (NCT03331900): 5-250 mg capsules in 34 healthy adults. Phase 1 multiple ascending dose (NCT03418688): 25, 50 or 100 mg every 12 hours for 10 days in healthy older volunteers, and 50 mg every 12 hours for 28 days in patients with Alzheimer's disease.
OralTwice daily—

Mechanism of action (compound-wide)

Irreversible covalent inhibitor of lysine-gingipain (Kgp), a bacterial cysteine protease secreted by Porphyromonas gingivalis — an anti-virulence-factor mechanism, not a host-directed one and not an antibiotic in the conventional sense. In a fluorogenic assay on purified Kgp (Z-His-Glu-Lys-MCA substrate, 100 mM Tris-HCl / 75 mM NaCl / 2.5 mM CaCl2 / 10 mM cysteine / 1% DMSO, pH 7.5, 90 min at 37 C) atuzaginstat gave an IC50 of <=50 pM, and a Morrison inhibition constant Ki of <0.01 nM; dilution-recovery kinetics showed irreversible binding (no Koff calculable), consistent with covalent alkylation of the catalytic cysteine by the aryloxymethyl ketone warhead. Selectivity is broad: no significant activity on the arginine gingipains RgpA/RgpB, IC50 >=10 uM on trypsin, and IC50 >10 uM on cathepsin S, calpain, tryptase, thrombin, plasmin, factor Xa, factor VIIa, BACE1, DPP4, the proteasome, deubiquitinating peptidases and the caspase family. In vitro it blocked P. gingivalis growth in defined medium (Kgp generates the peptide nutrients this asaccharolytic organism depends on) without selecting resistance over 16 serial passages, whereas moxifloxacin's MIC rose >1000-fold in 12 passages. In mice with established P. gingivalis brain infection, oral atuzaginstat at 10 and 30 mg/kg twice daily reduced brain bacterial load, Abeta1-42 and TNF-alpha; 3 mg/kg reduced bacterial load only. It blocked P. gingivalis-induced ApoE proteolysis in iPSC-derived astrocytes. In humans it was detected in CSF and reduced pathological ApoE fragments in CSF in the 28-day Phase 1 study in AD patients, and in GAIN it reduced salivary P. gingivalis DNA — the pharmacodynamic anchor for the trial's subgroup findings. The clinical hypothesis was that inhibiting Kgp halts gingipain-driven neurodegeneration upstream of amyloid and tau.

TargetActionAffinity
KgpprimarykgpInhibitor—ⓘ

← Full COR388 compound page (identity, identifiers, all indications)

Sources

  1. Cortexyme Provides Regulatory Update on Development Program for Atuzaginstat in Alzheimer's Disease (press release dated 2021-02-15, filed as Exhibit 99.1 to Form 8-K) — FDA partial clinical hold on the GAIN open-label extension following review of hepatic adverse events; double-blind phase (N=643, fully enrolled) continues to a Q4 2021 topline — Cortexyme, Inc. via U.S. Securities and Exchange Commission (EDGAR)
  2. Cortexyme Reports GAIN Trial Data Demonstrated Relationship Between Reduction of P. gingivalis Infection and Slowing of Alzheimer's Disease Progression (press release dated 2021-10-26) — co-primary endpoints ADAS-Cog11 and ADCS-ADL not met in the overall cohort; PG-DS subgroup (n=242) 57% slowing at 80 mg BID (p=0.02); liver enzyme elevations >3x ULN in 7%/15% — Cortexyme, Inc. (archived by Lighthouse Pharmaceuticals, the current owner of the asset)
  3. Cortexyme, Inc. Form 8-K, Item 8.01, filed 2022-01-26 (earliest event 2022-01-25) — FDA letter placing a FULL clinical hold on atuzaginstat's (COR388) IND 134303; cost-reduction programme; intent to prioritise COR588 in Alzheimer's disease — Cortexyme, Inc. via U.S. Securities and Exchange Commission (EDGAR)
  4. Cortexyme's Atuzaginstat Slowed Cognitive Decline in Participants With Alzheimer's Disease and P. gingivalis Infection in Phase 2/3 GAIN Trial: Additional Top-line Data Presented at CTAD 2021 — BioSpace
  5. Detke M, et al. The phase 2/3 GAIN trial of the first-generation P. gingivalis gingipain inhibitor atuzaginstat in mild-to-moderate probable Alzheimer's disease: new cerebrospinal fluid biomarker data. ASCP 2026 Annual Meeting, poster W1 — states that the GAIN CSF p-tau181 finding informed the design of the Phase 2 SPRING study of LHP588 and its p-tau217 inclusion criterion — American Society of Clinical Psychopharmacology (via CNS Pulse)
  6. Dominy SS, Lynch C, Ermini F, et al. Porphyromonas gingivalis in Alzheimer's disease brains: Evidence for disease causation and treatment with small-molecule inhibitors. Sci Adv. 2019;5(1):eaau3333 — COR388 potency (IC50 <=50 pM on Kgp, Morrison Ki <0.01 nM), irreversible binding kinetics, protease selectivity panel, CNS penetration, murine efficacy — Science Advances (AAAS) via PubMed Central
  7. NCT03331900 (COR388-001) — A Phase 1 Single Ascending Dose Study of COR388 HCl; 34 healthy subjects, 2017-12-11 to 2018-04-02 — ClinicalTrials.gov (U.S. National Library of Medicine)
  8. NCT03418688 (COR388-002) — A Randomized, Double-Blind, Placebo-Controlled, Multiple Ascending Dose Study of COR388 in Older Healthy Volunteers and Patients With Alzheimer's Disease; 33 enrolled, 2018-03-06 to 2018-10-15 — ClinicalTrials.gov (U.S. National Library of Medicine)
  9. NCT03823404 (COR388-010) — GAIN Trial: A Randomized, Double-Blind, Placebo-Controlled Study of COR388 in Subjects With Alzheimer's Disease; Phase 2/3, 643 enrolled, started 2019-03-28, completed 2022-01-01, results posted 2023-02-23 — ClinicalTrials.gov (U.S. National Library of Medicine)
  10. Quince Therapeutics Details Strategic Growth Plan with Launch of New Corporate Name (press release dated 2022-08-01, Exhibit 99.1 to Form 8-K) — Cortexyme renamed Quince Therapeutics (Nasdaq: QNCX), strategic shift to a bone-targeting platform, and disclosure of the 'intent to out-license its legacy neuroscience and antiviral assets' — Quince Therapeutics, Inc. via U.S. Securities and Exchange Commission (EDGAR)
  11. Sabbagh MN, Decourt B. COR388 (atuzaginstat): an investigational gingipain inhibitor for the treatment of Alzheimer disease. Expert Opin Investig Drugs. 2022;31(10):1049-1058 — Phase 1 PK (Tmax 0.5-1.5 h, steady-state t1/2 4.5-4.9 h, Cmax 25 ng/mL and AUC0-24 178 h*ng/mL at 25 mg BID), CSF detection, GAIN design, hepatotoxicity and the 2022-01-25 halt — Expert Opinion on Investigational Drugs (Taylor & Francis) via PubMed Central