$CRTX | Cortexyme
A post-mortem and a look at what’s next
Cortexyme $CRTX is a clinical stage biotech that has been developing gingipain inhibitors primarily for the treatment of Alzheimer’s Disease (AD). However, recent trial failures for their lead asset have forced the company to pivot to a secondary asset that is being explored not only for AD but also to treat periodontis as well. Below, we recap what went wrong with Cortexyme’s initial attempt at an AD drug and speculate as to the company’s future.
FDA Clinical Hold on COR388:
On January 25th 2022, Cortexyme received a letter from the FDA announcing a full clinical hold on their lead asset, COR388 (also referred to Atuzaginstat)1. On February 1st 2022, the CEO and CSO departed.2 According to Cortexyme, the reason for the prior clinical hold is related to concerns about potential liver toxicity. It is important to recall that in February 2021 the FDA informed Cortexyme of a partial clinical hold for a similar reason.3 This most recent hold comes on the heels of the Phase 2/3 GAIN trial of COR388 failing to meet statistical significance in October 2021.4
Cortexyme proposed that Porphyromonas gingivalis played a significant role the progression of AD and that by treating patients with a protease inhibitor that was lethal to P. gingivalis they could slow disease progression.5 However, correlation does not imply causation and while the data from the murine model was convincing, the predictive power of such models is limited.
What’s Next for Cortexyme
Cortexyme announced that they are pivoting away from COR388 to focus on COR588 as a treatment for AD and periodontitis1. Currently, COR588 is in the process of entering the clinic with recruiting being conducted for their Phase I clinical trial.6
Excerpt Press Release on results Phase 2/3 GAIN Trial Announced in October 2021:
The 643-participant study in mild to moderate patients with Alzheimer’s disease did not meet statistical significance in its co-primary cognitive and functional endpoints as measured by ADAS-Cog11 and ADCS-ADL at end of the treatment period in the overall cohort.
The pre-specified subgroup of participants with P. gingivalis DNA detectable in saliva at baseline (PG-DS; n=242) showed a dose response, with a 57% slowing of cognitive decline as measured by ADAS-Cog11 in the 80 mg BID arm (p=0.02) and a 42% slowing in the 40 mg BID arm (p=0.07) vs. placebo. Significant benefits in this subgroup were not seen on the other co-primary, ADCS-ADL. The cognitive benefit of atuzaginstat in patients with high P. gingivalis infection was reinforced by similar results in multiple pre-specified infection related subgroups and with multiple methods of analysis. Additionally, 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), Mini-Mental State Exam (MMSE) (p=0.007), and a beneficial trend on ADCS-ADL (p=0.08).
The sub-study in periodontal disease demonstrated a trend to benefit on the primary clinical endpoint of pocket depth in the same pre-specified sub-group with P. gingivalis DNA detectable in saliva. Further results will inform the next stage of development in periodontitis and will be presented at a future scientific conference.
Source: Cortexyme Press Release from October 26th, 20214
We like that they used both a functional and cognitive endpoint and controlled for with P. gingivalis DNA detectable in saliva.
Cortexyme claims they have identified a subgroup, patients with P. gingivalis DNA detectable in saliva, that may respond to treatment. This claim is contentious in our opinion since only one of the two groups has a p-value < 0.05 for the cognitive endpoint and the functional endpoint was not statistically significant.
It is unclear how large this subgroup was in the trial. according to their investor presentation only 435 (with 173 in placebo group, 133 in the 40 mg dose group and 129 in the 80 mg dose group) of the 643 patients completed the trial. Based on our understanding ~38% of patients had detectable P. gingivalis DNA in the initial 643 but the number completed has not been reported. To better assess the power of this observation it could be useful to know the final end count.
The sub study in periodontal disease found a reduction in gingival pocket depth in patients with detectable levels of P. gingivalis. Data corresponding to this conclusion was not shown. However, these findings are consistent with the preclinical data presented in their investor presentation.
The Parallax View:
While we disagreed with Cortexyme’s AD hypothesis from the start, we applaud Cortexyme for the thorough scientific research they conducted – providing a reminder that most ships do not make it back to harbor in clinical biotech research. Cortexyme proposed a novel pathway to describe the progression of AD, then developed a lead asset with a proposed mechanism of action that would test their proposed pathway. Cortexyme proceeded with a logical progression of experiments: first developing murine models to test the role of P. gingivalis in AD progression and the efficacy of their lead asset in slowing disease progression before initiating clinical trials. Unfortunately for them, it seems their proposed therapy did not pan out in the clinic.
We have never been optimistic about the long term prospects of any of Cortexyme’s drug candidates as treatments for AD and this most recent setback reaffirms our opinion. It seemed to us that Cortexyme was advancing a microbiome influenced version of the amyloid hypothesis. We refrained from previously commenting about this company because we do not think debating the validity of the amyloid hypothesis and stating why we think an AD drug is going to fail in a late-stage clinical trial is a worthwhile intellectual exercise. However, given the reported improvement in gum pocketing observed in in the GAIN trial announced in October of 2021, we think gingipain inhibitors may represent a nontrivial opportunity for treating gingivitis and preventing periodontitis (roughly a $1B market size in 2027 by one estimate, where something like half of all American adults suffer from gum disease).7 For this reason, we are interested to see how the rest of the Cortexyme story plays out. Unfortunately, ~15% of patients who received the highest doses of COR388 had elevated liver enzymes. Viewed in the context of the most recent clinical hold, it makes sense that Cortexyme is advancing COR588, a second generation gingipain inhibitor that is structurally distinct from COR388, as a treatment for periodontitis. However, we have seen little data relevant to COR588 that suggests improved safety and will instead wait until the Phase I data for more insight.
Given the available data and established FDA precedent rejecting elevated liver enzymes, even in a small fraction of patients, as an appropriate price for healthier gums, we are cautiously optimistic at best. If their second-generation inhibitor has an improved safety profile, then we believe the treatment and prevention of gum disease is a compelling use case for this class of protease inhibitors. However, it is challenging for us to understand the relevant unit economics associated with this use case. From an intellectual viewpoint, we hope that COR588 is successful as a treatment for gum disease since we are very interested to see how pricing this type of drug would shake out.
Periodontis, especially severe periodontis, is currently treated by primarily surgical methods as well as conventional antibiotics when an indentifialbe infection is implicated, making a protease inhibitor a relatively unique approach that may be viewed by patients and dentists as less invasive. However, validation of P. gingivalis’ direct link to gum disease must be conducted. Furthermore, pricing would likely hinge on the longer-term "costs” of gum disease which can be extrapolated to overall health declines when comparing patient cohorts given dental health’s longstanding correlation with overall health. Whether this type of pricing strategy would hold up under insurer scrutiny (as it does for things like routine dental cleanings or dental coverage within Medicare, for example) is another question as broad claims of periodontis leading to costly outcomes elsewhere may be met with reasonable skepticism.
Relevant Scientific Publications implicating the role of P. gingivalis in the progression of AD:
Singhrao, S. K. J. Alzheimers. Dis. 2013, 36, 665: Identified the presence of P. gingivalis in 4 out of 10 brain samples harvested from AD patients shortly after the time of death.
de Leon, M. J. Neurobiol Aging, 2015, 36, 627: Found that Aβ brain accumulation in cognitively normal and healthy elderly correlated periodontal disease. In the study the researchers controlled for relevant confounders resulting from medical history and oral heath behavior. Findings were consistent with previous findings in animal models.
Watanabe, K. Plos One, 2018, 13, e0204941: Continually orally exposed mice to P. gingivalis and observed neurodegeneration and increased production of Aβ42. Findings suggested that chronic P. gingivalis infection can lead to the development of a neuropathology associated with AD.
Potempa, J. Sci. Adv. 2019, 5, eaau333: Identified COR388, a small molecule gingipain inhibitor as a potential AD drug. AD mouse model experiments suggested that treatment with COR388 blocked Aβ production and decreased the presence of P. gingivalis in the brain.
Cortexyme Announces Clinical Hold on Atuzaginstat’s Investigational New Drug Application
Two Cortexyme Founders Leave in Executive Shake-Up
Cortexyme Provided Regulatory Update on Development Program for Atuzaginstat in Alzheimer’s Disease https://www.cortexyme.com/cortexyme-provides-regulatory-update-on-development-program-for-atuzaginstat-in-alzheimers-disease/
Cortexyme Reports GAIN Trial Data Demonstrated Relationship Between Reduction of P. gingivalis Infection and Slowing of Alzheimer’s Disease Progression
Comprehensive Alzheimer’s pathology is induced by P. gingivalis infection: COR388 and other gingipain inhibitors protect against synaptic loss
A Single and Multiple Ascending Dose Study of COR588 https://clinicaltrials.gov/ct2/show/NCT04920903
Global Periodontal Disease Therapeutics Market $1.1 Billion by 2027