Educational content on VJDementia is intended for healthcare professionals only. By visiting this website and accessing this information you confirm that you are a healthcare professional.

Share this video  

AAIC 2021 | Targeting P. gingivalis infection to treat Alzheimer’s: the evidence

The gingipain hypothesis of Alzheimer’s disease proposes that P. gingivalis infection of brain tissue causes disease through the secretion of gingipains (proteases) that promote neuronal damage. COR388 (atuzaginstat) is a novel investigational virulence factor inhibitor that targets gingipains for the treatment of Alzheimer’s disease. Michael Detke, MD, PhD, Cortexyme, South San Francisco, CA, shares the key preclinical and early phase data from investigations of COR388 gathered to date. Post-mortem tissue analysis has demonstrated the presence of gingipains in a statistically significantly higher proportion of Alzheimer’s brains compared to age-matched controls. Additionally, the level of gingipains was found to correlate with other biomarkers of Alzheimer’s severity, including tau and ubiquitin levels. Infection of mice with P. gingivalis was shown to induce brain infiltration and downstream Alzheimer’s pathology, including amyloid beta production, microglial activation, and hippocampal neurodegeneration. The introduction of COR388 blocked these changes. Data from early phase clinical trials showed COR388 to be well tolerated with few adverse events, and trends towards benefit were seen in exploratory cognitive measures, including the mini-mental state examination and the Cambridge Neuropsychological Test Automated Battery (CANTAB). This interview took place during the Alzheimer’s Association International Conference 2021.

Transcript (edited for clarity)

Our preclinical colleagues have shown that the drug binds to the gingipains. One of the key preclinical findings is in an animal model. And so this is a really a key piece of evidence for causation too, because prior to starting the company, there was some scientific literature showing that one of the most common factors you’ll see as a risk factor for development of Alzheimer’s disease is periodontal disease...

Our preclinical colleagues have shown that the drug binds to the gingipains. One of the key preclinical findings is in an animal model. And so this is a really a key piece of evidence for causation too, because prior to starting the company, there was some scientific literature showing that one of the most common factors you’ll see as a risk factor for development of Alzheimer’s disease is periodontal disease. Not very intuitively, but they overlap substantially. We knew that that was an overlap. We knew that P. Gingivalis was a cause of periodontal disease. So we went looking for P. Gingivalis in Alzheimer’s brains.

One of the key pieces of data was work we did with the University of Auckland, which has an excellent brain bank of patients who passed away and donated their brains. And so we looked at Alzheimer’s brains and found the gingipains in 90 to 100% of the Alzheimer’s brains. We also looked in age-matched controls, and it was much lower. It was statically significant at a p-value less than 0.0001. But very interestingly, the age-matched controls, some of them had gingipains, not 95%, but about 30%. And that’s exactly consistent with our hypothesis that this infection is upstream from all the other effects. We know that there’s a lot of pathology in Alzheimer’s years, even decades before the cognitive impairment sets in. Plaques, tangles and other things.

So if this infection were to be upstream of it, you’d expect it to be in some healthy elderly in the 50 to 80 year old range that went on to die for other reasons, a stroke or heart attack or something like that. So that’s exactly what we saw. It’s very consistent with that hypothesis. These gingipains, the level of the gingipains correlate with other biomarkers of Alzheimer’s severity, such as tau, and ubiquitin, but this doesn’t prove causality. The gingipains are there in the brains of almost all Alzheimer’s patients, but the next key preclinical study was a mouse model.

This is a deceptively simple model in which you take regular wild type mice, meaning they have no genetic modifications, and you take P. Gingivalis from humans, and you rub it on the teeth and gums of the mice. That’s the whole model. And then you wait. These on the left, the Alzheimer’s brains and the control brains. And above one on that scale is positive, below that is basically background noise. So you can see almost all the Alzheimer’s brains have P. Gingivalis in them. And as I said, a minority, but a significant 30% or so of the controls did too. And this is correlating with tau and ubiquitin. And then in the animal model, these majors findings in this animal model have now been… We’ve generated them but they’ve now been replicated at five different independent scientific laboratories.

This is some particularly elegant work that was done at the University of Illinois at Chicago. So you just put P. Gingivalis in the mouth, on the teeth and gums of regular mice, and you see P. Gingivalis infiltrating the brain, you see neuroinflammation with multiple markers, here’s TNF alpha. You see amyloid plaques build up and you see tau tangle-like neurons. You also see activation of microglia, which are the immune cells of the brain. And then finally, and perhaps most importantly, you see that neurons die off in the hippocampus, one of the key memory areas of the brain.

And then we went on to test COR388 in this model, and here are the data. The prior slide, I showed you was an academic group that treated for 22 months. So they showed really big differences. We were working more quickly here in a drug development timeline and so we treated for only six weeks with P. Gingivalis. Exposed them for the infection for six weeks, and then treated with COR388 or atuzaginstat for five more weeks. And you can see that it knocked down the level of the bacteria significantly, and it essentially fully reverses the increase in amyloid, the increase in infectious markers like TNF alpha, and it’s actually effectively, fully protected against the death of the interneurons in the hippocampus.

We did a Phase I program with a standard single ascending dose or SAD study and a multiple ascending dose or MAD study and the key takeaways in Phase I are usually with regard to safety and pharmacokinetics. And our safety signal across all these studies was very encouraging. Across all the safety measures there were really no clinically significant findings. On the pharmacokinetics, it’s a drug with a really nice, what we call drug-like properties. It’s well absorbed. It distributes well throughout the body. We know it gets into the brain, and this is roughly linear exposure in the blood as correlated with dose, what we call linear PK, within the dose range that we’re testing. The half-life is about five hours consistent with twice daily dosing, but getting to the really more interesting efficacy and biomarker data, these are a couple of the biomarkers that we looked at in our Phase I study. On the left is something called RANTES, it’s also called CCL5, which is an inflammatory biomarker. So again, we expected to see this, we know there’s inflammation in Alzheimer’s and all sorts of other things that suggest an infection, but we expected that our treatment with atuzaginstat would reduce this. And in fact, we saw about a 30% reduction in only 28 days, and this was statistically significant.

Another thing that we know the gingipains from P. Gingivalis do is they cleave certain important proteins like tau and also APOE, which is known to be involved in Alzheimer’s. And here we show that… We’ve hypothesized since the P. Gingivalis gingipains proteases are cleaving APOE and making these fragments, we predicted that treatment with COR388 or atuzaginstat would reduce the number of the small fragments. And again, we saw about a 30% reduction and again, statistically significant. And then the clinical data, this was a small study. So it wasn’t really designed or powered to see statistical significance on these. Nonetheless, we saw a nice change in the direction of improvement on a standard Mini Mental State Exam assessment. This is computerized cognitive assessment. We also saw a trend in the right direction here. And then this is a really interesting company, Winterlight, that’s developed an AI based analysis of speech. And on measurements of speech that are correlated with decline in Alzheimer’s, such as word finding difficulty and loss of sentence complexity, we saw statistically significant improvements.

So all in all these data in our phase one study were very encouraging and led us to go on to the larger Phase II study, Phase II/III potentially pivotal GAIN trial. We’re really excited about this molecule, the efficacy we’ve seen so far, the safety we’ve seen so far, and not only that, but the convergence of data. I’ve walked you through some of the data, but there are all sorts of other things that this PG hypothesis explains. I’ll give you just a couple of examples that I think anyone can, can resonate with anyone.

One of the findings out there that everybody who treats Alzheimer’s knows about is that a lot of Alzheimer’s patients lose their sense of smell. They develop anosmia. Sometimes even before they have Alzheimer’s. And what we’ve shown is that in the mouse model, when we have P. Gingivalis in their mouth, it goes into the nose and it goes directly into the olfactory nerve and into that part of the brain that controls olfaction. So that’s one of those things that’s been out there in the literature and not understood well, and our theory explains beautifully.

There’s a publication that came out before our company was founded that showed that spouses of Alzheimer’s patients are at six times as high risk of getting Alzheimer’s as the general population. They speculated in the paper that this might be due to caregiver stress because usually the spouse of the Alzheimer’s patient is the caregiver, but I’m a psychiatrist and a psychologist and stress is very important and it does impact things and it can make things 50% worse or something like that. It doesn’t make things six times worse. However, a very easily shared oral infection could easily make something six times harder. So again, our theory explains that. So as a scientist, I’m really excited about all the converging evidence that is really convincing us to be very enthusiastic about the data coming out next quarter.



Dr Detke is a full-time employee of Cortexyme and has CRTX stock options.