Cortexyme Alzheimer’s Study Fails; Drug Slows Cognitive Decline In Some Patients
Subpopulation Efficacy Limited To One Of Two Primary Endpoints
Executive Summary
Atuzaginstat targets a bacteria associated with Alzheimer’s disease. The higher of two doses showed a 57% slowing in cognitive decline in the Phase II/III GAIN trial in mild to moderate patients with P. gingivalis infection confirmed by saliva test.
Cortexyme, Inc.’s drug atuzaginstat did not meet either of the two primary endpoints in the Phase II/III GAIN clinical trial that enrolled 643 patients with mild to moderate Alzheimer’s disease. However, the company said when it revealed top-line results on 26 October that the drug did slow cognitive decline by up to 57% in a patient subgroup that aligns with the biologically relevant population for atuzaginstat.
South San Francisco-based Cortexyme’s value surged in the aftermath of US Food and Drug Administration accelerated approval in June for Biogen, Inc./Eisai Co., Ltd.’s Aduhelm (aducanumab) based on biomarker results – the antibody’s ability to clear amyloid-beta plaques from the brains of patients with mild Alzheimer’s disease. It remains to be seen if Cortexyme can pursue a similar biomarker-driven approval for atuzaginstat (COR388), but the company plans to speak with regulators and other stakeholders about the path forward for its drug based on the GAIN results.
Cortexyme’s stock price was up 107.6% year-to-date when it closed at $57.68 on 26 October, including a residual boost from Aduhelm’s approval, anticipation that GAIN would generate positive data and biomarker data that the company reported from the study in August. Cortexyme fell 73.2% to $15.45 in after-hours trading on 26 October following the top-line data announcement.
Previously, the stock surged to $120.21 on 9 August when Cortexyme said it would announce results from GAIN in mid-November. In addition to the top-line results now reported, the company will present detailed data on 11 November during the Clinical Trials in Alzheimer’s Disease (CTAD) meeting.
Atuzaginstat targets the bacteria P. gingivalis in the brains of Alzheimer’s patients, which has been associated with inflammation and degeneration. P. Gingivalis is a bacterium associated with periodontal infections and inflammation but gains access to the brain through the bloodstream. It generates gingipains that destroy neurons in the brain. Cortexyme’s drug blocks gingipains with the goal of preventing neuronal death that leads to neuroinflammation and the accumulation of amyloid.
The company recently initiated a Phase I study for its second-generation lysine-gingipain inhibitor known as COR588 – an oral drug dosed once daily versus the twice-daily dosing for atuzaginstat.
“We've been building this company for eight years based on an understanding of a bacteria called p. gingivalis and its ability to drive Alzheimer's pathology localized in the brains of patients,” CEO Casey Lynch told Scrip. “This Phase II/III trial was the culmination of work to develop a drug that could target this bacteria very specifically and reduce bacterial load, not only systemically, but also in the brain.”
Primary Endpoints Missed, Cognition Endpoint Met In Subgroup
In GAIN, atuzaginstat-treated patients across the entire Phase II/III population did not experience statistically significant slowing in the rates of cognitive or functional declines as measured by the Alzheimer's Disease Assessment Scale-Cognitive Subscale 11 (ADAS-Cog11) and Alzheimer's Disease Cooperative Study Group-Activities of Daily Living (ADCS-ADL) at 48 weeks, respectively – the co-primary endpoints of the study.
However, among 242 patients with P. gingivalis DNA detectable in saliva at baseline (37%), there was a 57% slowing of cognitive decline as measured by ADAS-Cog11 among trial participants treated with an 80mg twice-daily dose of atuzaginstat (p=0.02) and a 42% slowing in cognitive decline in the 40mg twice-daily arm of the study (p=0.07) compared with those who received a placebo. Changes in functional decline as measured by the ADCS-ADL were not statistically significant for either atuzaginstat dose in the subgroup.
Looking at patients with P. gingivalis detected in the blood (50%), cognitive decline assessed via ADAS-Cog11 slowed by 37% compared to placebo in the 80mg atuzaginstat group (p=0.12) and by 27% in the 40mg dose group (p=0.25).
“We think the subgroup response actually really reinforces the hypothesis,” Lynch said. “We're not talking about a subgroup that was on cholinesterase inhibitors or a subgroup that was within a very specific cognitive range. We're talking about the biologically relevant subgroup that is sensible to respond, so we think it actually even reinforces the mechanism.”
Cortexyme said reductions in P. gingivalis in saliva at week 24 were significantly correlated with improved outcomes at the end of the treatment period as measured by ADAS-Cog11 (p=0.0007), Clinical Dementia Rating–Sum of Boxes (CDR; p=0.004) and Mini-Mental State Exam (MMSE; p=0.007) with a positive trend on ADCS-ADL (p=0.08).
“We took all-comers into the study, but we said, ‘These are the most likely groups to respond – we can use a saliva diagnostic, we can use a blood-based diagnostic,’ and both of those subgroups showed this very nice slowing of disease,” Lynch said. “The people who had the most reduction in their bacterial load also had the best outcomes which really reinforces the mechanism of action.”
However, Cortexyme did raise the possibility that atuzaginstat could impact Alzheimer’s patients more broadly in August when it reported that P. gingivalis antibodies were detected in the cerebrospinal fluid (CSF) of all 472 patients with CSF samples available at the time they enrolled in GAIN. The antibodies in the CSF were “only very weakly correlated to the albumin index (r=0.22), indicating production in the central nervous system in addition to that shown previously in serum,” the company said on 25 August.
H.C. Wainwright analyst Andrew Fein said in a 26 August note that these and other biomarker data previously released by the company supported Cortexyme’s P. gingivalis hypothesis in a majority of Alzheimer’s patients.
However, Citi analyst Neena Bitritto-Garg predicted in a 13 October note that the GAIN study was unlikely to succeed in Alzheimer’s disease broadly “because: 1) generally low [probability of success] in Alzheimer's, 2) novel mechanism with limited preclinical/mechanistic evidence, 3) limited previous clinical data (only 6 Alzheimer's patients dosed with atuzaginstat for 28 days prior to GAIN), 4) lack of adequate testing for futility in interim analysis, 5) high bar for success (co-primary endpoints, mild-moderate Alzheimer's population, etc.), 6) potentially underpowered study due to multiple endpoint changes, lack of rigorous screening criteria, and more.”
Could Safety Flag The Need For Another Trial?
Most adverse events in the GAIN study were mild to moderate in severity with gastrointestinal side effects as the most common adverse events. Diarrhea was experienced by 16% of atuzaginstat-treated patients with 6% who received the drug experiencing nausea versus diarrhea and nausea rates of 3% and 2%, respectively, for individuals in the placebo group.
Cortexyme also reported that dose-related liver enzyme elevations of greater than three times the upper limit of normal were observed in 2% of patients who received a placebo, 7% of people in the 40mg atuzaginstat dose group and 15% in the 80mg group. Nearly all patients were asymptomatic, but two trial participants in the 80mg arm of the study has concomitant bilirubin elevations. The company said liver enzyme elevations resolved while patients remained on drug or after discontinuing treatment without any known long-term adverse effects.
“Overall, we were really pleased with the tolerability and safety profile,” chief medical officer Michael Detke told Scrip. “We've got some hepatic expert advisors who have told us based on the data in our study and lots of other data that simply implementing something as easy as titrating up the dose may significantly mitigate those [elevated liver enzyme levels].”
Liver enzyme elevations previously were disclosed in February when the FDA put the open-label extension (OLE) phase of the GAIN trial on a partial clinical hold due to hepatic adverse events, which meant that no new participants could be enrolled and patients already in the OLE would discontinue treatment.
Cortexyme noted in a statement to Scrip that while the OLE was discontinued, “the FDA supported the continuation of the randomized, double-blind, placebo-controlled phase of the GAIN trial, which is the more rigorous way to assess both efficacy and safety, and it led to the results we're sharing today – game-changing efficacy in [P. gingivalis]-infected mild-moderate AD patients, with a highly differentiated oral drug and safety profile.”
Next Steps: Determining Whether Another Trial Is Needed
The next step for Cortexyme’s development of atuzaginstat will be sitting down with global regulators to figure out the most efficient path forward, chief operating officer Chris Lowe said.
“Is that another study? What might that study look like in terms of the realm of expanding safety, replicating efficacy, or is there another path?” Lowe said. “We don't presume to know what it is at this moment, we just know we've had a very collaborative relationship with both European and US regulators, and so we're looking forward to sitting down to share this data set with them and figure out how we can most efficiently help patients.”
Lynch noted that “regulatory flexibility is very important in this area of very high unmet medical need,” but she said Cortexyme is not looking at the FDA’s accelerated approval decision for Aduhelm based on the drug’s effects on amyloid as indicative of what might happen with atuzaginstat based on its ability to block P. gingivalis.
“We're really focused on our own program, which is differentiated,” she said. “We're in a symptomatic patient population, we've got an oral pill, we don't need PET imaging at the beginning, we can use a saliva test, we don't need MRI monitoring at the end, and it's a different mechanism. So, from our perspective, it's really important for patients to have multiple different options and different mechanisms, moving forward.”
Aduhelm is approved to treat people with mild cognitive impairment or mild dementia and its label says that Alzheimer’s patients should have had an MRI within the prior year before initiating treatment and they need MRIs prior to the seventh and 12th doses to monitor patients for amyloid-related imaging abnormalities (ARIA). The label stops short of requiring PET scans to confirm the presence of amyloid in the brain, but payers are likely to require such testing before covering Aduhelm’s cost. (Also see "Aduhelm’s Revised Labeling Could Help Payers Navigate Reimbursement" - Scrip, 8 Jul, 2021.)
Will Cortexyme Proceed Alone Or With A Partner?
In addition to the development and regulatory path forward for atuzaginstat, another unanswered question is whether Cortexyme will continue on its own or with a partner to advance its drug.
“First and foremost, we're going to pursue a development path of whatever is in the best interests of patients. And if a partner can add that value, we're open to have that conversation,” Lowe said. “I feel really lucky to be a part of a team that has the capabilities to take this the distance, if that's what's in the best interests of everybody.”
Cortexyme had $153.5m in cash on its balance sheet at the end of the second quarter, which the company said should fund its operations through the end of 2023.
Cortexyme has tried to nimbly develop its drug on its own to date, Lowe said, noting that “while it took three decades for other mechanisms to get to wherever they are presently, it's taken us about seven or eight years to amass a mountain of evidence, and so we want a partner that is going to recognize that and be able to match our nimbleness to help patients.”
Based on the saliva and blood test results in the GAIN study, it appears that about 50% of Alzheimer’s patients test positive for P. gingivalis, Lynch said, which means that about 1.5m of the 3m US patients with mild to moderate disease in the US may be eligible for treatment with atuzaginstat.
“While I've been in this industry a long time, I've also been a caregiver to both of my parents for a good number of years, and I saw their disease go from early onset to mild to moderate to late-stage firsthand,” Lowe said. “And so, to have this data set in hand, and to think about what it means to not just the millions of patients, but to the millions of caregivers, their families and the impact, and to be able to sit here and say that we have evidence that over half the patients should experience a 50% or greater slowing of disease on cognition, I mean, that's the most powerful thing that I, frankly, have ever heard in my entire career, let alone just here at Cortexyme.”