🎧 Listen to this article

When a pharmaceutical company spends millions developing a drug that targets bacteria from your mouth inside the brain, it tells you something about where the science has landed. When that drug fails — and another company buys the wreckage and starts over — it tells you something even more interesting.

A 2019 study in Science Advances found Porphyromonas gingivalis — the keystone pathogen behind chronic gum disease — in the majority of postmortem Alzheimer’s brains examined. The toxic enzymes it produces, called gingipains, correlated directly with tau pathology and amyloid buildup. More gingipains meant worse disease markers (Dominy et al., 2019).

We’ve covered the gingipain inhibitor approach to Alzheimer’s and how oral bacteria reach your brain through two distinct pathways. But the business story behind this drug — the trial results, the FDA hold, the company pivot, and the new trial now underway — reveals what happens when an idea is too promising to abandon but too complicated to get right on the first try.

What did the original research actually find?

Dominy’s team didn’t just find bacteria in brain tissue. They found P. gingivalis in the majority of Alzheimer’s samples, along with its toxic proteases (gingipains). The levels of those gingipains correlated with tau and amyloid pathology — the two hallmark proteins of Alzheimer’s disease. More gingipains, worse markers.

Then they infected mice with P. gingivalis through the mouth, mimicking human gum disease. The oral infection led to brain colonization, increased amyloid-beta production, neuroinflammation, and hippocampal neuron damage. Treating those mice with small-molecule inhibitors designed to block gingipain activity reduced bacterial load in the brain, decreased amyloid production, and rescued hippocampal neurons. Block the toxin, protect the brain.

This wasn’t a vague statistical association. The bacteria were in the brain. Their toxic enzymes were in the brain. And the amount of enzyme tracked with disease severity.

How did Cortexyme’s Alzheimer’s drug actually perform?

Based on this preclinical evidence, a company called Cortexyme was founded to develop a gingipain inhibitor for Alzheimer’s. Their lead compound, atuzaginstat (COR388), was designed to cross the blood-brain barrier and block the toxic proteases that P. gingivalis produces once it reaches brain tissue.

The drug entered Phase II/III clinical trials — the GAIN trial — for mild to moderate Alzheimer’s disease. That’s hundreds of millions in development costs. When a pharmaceutical company bets that much on a mechanism, it means the preclinical evidence cleared a very high bar.

A 2020 study in Pharmacology Research & Perspectives tested the compound in aged dogs naturally infected with P. gulae, the canine equivalent of P. gingivalis. The treatment reduced bacterial infection, decreased neuroinflammation markers, and improved cognitive function. The animal data kept building.

The GAIN trial results, announced in October 2021, told a complicated story. The drug didn’t meet its co-primary endpoints in the overall population. On the surface, that looks like failure.

But buried in the data was something striking. A pre-specified subgroup analysis — not a post-hoc fishing expedition, but something the researchers planned before unblinding — found that participants with detectable P. gingivalis DNA in their saliva who received the higher dose experienced a statistically significant 57% slowing of cognitive decline as measured by ADAS-Cog11.

A 57% slowing in a neurodegenerative disease is meaningful. The problem was that the drug didn’t work in Alzheimer’s patients who didn’t have detectable P. gingivalis. The overall trial population included everyone — oral-bacteria-positive and oral-bacteria-negative — and the negative patients diluted the effect.

Then came the liver issue. The FDA placed a partial clinical hold on atuzaginstat in February 2021 due to liver abnormalities in some participants. By January 2022, that escalated to a full clinical hold. In August 2022, Cortexyme announced it was discontinuing its gingipain inhibitor program and offering the portfolio for external licensing.

Why did another company buy a failed Alzheimer’s drug?

Cortexyme changed its name to Quince Therapeutics and shifted its focus to bone disease. In January 2023, Quince sold the former Cortexyme small-molecule protease inhibitor portfolio — including the rights to the gingipain inhibition approach — to Lighthouse Pharmaceuticals.

Lighthouse didn’t buy a failure. They bought a 57% subgroup finding and a mechanism that made biological sense. Their approach differs from Cortexyme’s in one critical way: patient selection. Instead of enrolling all Alzheimer’s patients and hoping the drug works in enough of them, Lighthouse is running the SPRING trial (Stopping PRogression of P. gINGivalis-Positive AD with Gingipain Inhibition) with a second-generation compound called LHP588. The trial aims to enroll 300 patients with mild-to-moderate Alzheimer’s disease who test positive for P. gingivalis in saliva — the exact subgroup where Cortexyme’s original drug showed the strongest effect (Lighthouse Pharmaceuticals, CTAD 2025).

This is precision medicine applied to a pathogen hypothesis. Don’t treat everyone with Alzheimer’s. Treat the people whose Alzheimer’s has a specific infectious driver. If it works, it changes the diagnostic conversation entirely — suddenly your dentist’s periodontal chart becomes relevant to your neurologist’s assessment.

What does the subgroup finding actually tell us?

Whether LHP588 succeeds or fails in clinical trials, the implications of this story extend beyond one compound.

The subgroup finding validates the mechanism. If P. gingivalis in the brain were irrelevant to Alzheimer’s progression, you wouldn’t see a 57% slowing of cognitive decline in people who carry the bacterium. The fact that the drug only worked in P. gingivalis-positive patients is itself evidence that the bacteria are contributing to the disease in those individuals.

The two-pathway model explains the inconsistency. Not everyone with gum disease develops Alzheimer’s, and not everyone with Alzheimer’s has detectable P. gingivalis. But research shows the bacteria can reach the brain through at least two routes. The first is direct bloodstream invasion — when chronically inflamed gum tissue becomes permeable, P. gingivalis outer membrane vesicles cross the blood-brain barrier and trigger neuroinflammation (Li et al., 2023). The second route runs through the gut: swallowing oral bacteria disrupts your intestinal microbiome, alters kynurenine metabolism, and produces neurotoxic metabolites that contribute to neurodegeneration (Lee et al., 2025). Different patients may have different dominant pathways.

The business continuity signals confidence. Pharmaceutical companies don’t buy failed drug programs out of nostalgia. Lighthouse evaluated the same data everyone else saw — the liver toxicity, the missed primary endpoints, the FDA hold — and decided the 57% subgroup finding plus a refined patient selection strategy justified a second attempt. That’s a calculated bet on the underlying biology.

What can you do about your gums right now?

You don’t need to wait for the SPRING trial results to act on this research. The preclinical and clinical evidence already supports taking periodontal health seriously as a brain health intervention.

Get a periodontal evaluation. Not a routine cleaning — ask specifically about pocket depth measurements and signs of bone loss. A standard dental cleaning doesn’t assess these. If you’re over 40, request periodontal charting.

Treat bleeding gums as an active inflammatory process. Bleeding when you brush or floss isn’t normal. That inflammation is the entry point for bacteria to reach your bloodstream. See a periodontist if your dentist flags concerns.

Understand what’s at stake. Periodontal disease affects roughly half of adults over 35. By your 60s and 70s, that climbs to 70–80% with significant disease. And it’s largely painless in early stages. The damage accumulates silently for years — possibly decades — before anyone connects it to cognitive changes.

Reduce your total inflammatory burden. Everything that reduces systemic inflammation also reduces your susceptibility to the downstream effects of oral bacterial translocation. Sleep, stress management, diet quality, and regular movement all contribute. The goal isn’t perfection; it’s lowering the baseline inflammation that makes bacterial invasion more damaging.

This article is for informational purposes only and does not constitute medical advice. Always consult with a qualified healthcare provider before making changes to your health regimen.

What are gingipain inhibitors? Gingipain inhibitors are small-molecule drugs designed to block the toxic proteases (gingipains) that Porphyromonas gingivalis produces. These enzymes damage brain tissue and correlate with Alzheimer’s pathology. The approach targets the bacterial toxin rather than the bacteria itself.

Did the gingipain inhibitor Alzheimer’s drug fail? The GAIN trial missed its primary endpoints in the overall Alzheimer’s population. However, a pre-specified subgroup analysis found a 57% slowing of cognitive decline in patients with detectable P. gingivalis DNA. The drug failed broadly but showed meaningful efficacy in a specific subgroup.

What is LHP588? LHP588 is a second-generation gingipain inhibitor developed by Lighthouse Pharmaceuticals. It’s being tested in the SPRING trial, which enrolls only Alzheimer’s patients who test positive for P. gingivalis in saliva — the subgroup where the original drug showed the strongest effect.

How do oral bacteria reach the brain? Research identifies two pathways. First, P. gingivalis outer membrane vesicles can cross the blood-brain barrier directly from inflamed gum tissue. Second, swallowed oral bacteria disrupt gut microbiome balance, alter kynurenine metabolism, and produce neurotoxic metabolites that contribute to neurodegeneration.

Should I get my gums checked if I’m worried about Alzheimer’s? Periodontal disease affects roughly half of adults over 35 and is largely painless in early stages. A periodontal evaluation — including pocket depth measurements and bone loss assessment — is a reasonable step if you’re concerned about the oral-brain connection. Standard dental cleanings don’t assess these factors.