Quince Therapeutics, Inc. (QNCX) Earnings Call Transcript & Summary

Greetings. Welcome to Cortexyme Conference Call. [Operator Instructions] Please note, this conference is being recorded. I will now turn the conference over to Stacy Roughan. Thank you. You may begin.

Welcome, and thank you for joining us to discuss top line results from our GAIN trial. With us from management are Casey Lynch, our Chief Executive Officer, Co-Founder and Chair; Mike Detke, Chief Medical Officer; and Chris Lowe, our Chief Operating Officer and Chief Financial Officer. You can find today's news release, along with the presentation that management will refer to in their prepared remarks, on our Investor Relations website at ir.cortexyme.com. During the course of our remarks today, we will be making forward-looking statements, including statements relating to the outcome and results of the GAIN trial, the timing of related announcements and updates, the potential of atuzaginstat to treat Alzheimer's disease and opportunities to treat other chronic health conditions. Forward-looking statements are based on Cortexyme's current expectations and involve risks and uncertainties. Our actual results and timing of events could differ materially from those anticipated. Factors that could cause results to differ include those described in the section titled Risk Factors in our annual report on Form 10-K filed with the Securities and Exchange Commission on March 1, 2021 and in our quarterly report on Form 10-Q filed with the SEC on August 6, 2021. Cortexyme cautions you not to place undue reliance on forward-looking statements, and the company undertakes no duty or obligation to update any forward-looking statements as a result of new information, future events or changes in its expectations. With that, I would now like to turn the call over to Chris Lowe.

Great. Thank you, Stacy, and good afternoon, everybody. Truly a great day for patients, a great day for Alzheimer’s research. Very pleased to be able to share positive outcomes across multiple fronts from the GAIN trial with you today. Today, our mission to alter the course of this disease has indeed taken a giant step forward. When our co-founders, Casey and Steve, set out on this journey over 7 years ago, the mission was to simply follow the evidence in search of a comprehensive cause of Alzheimer's. Their leadership to follow the evidence has led us to, indeed, this wonderful day. And so from me and on behalf of all the employees and the patients around the world, a heartfelt thank you to Casey and Steve. Now relative to GAIN, as we indicated in the press release, we did not achieve statistical significance on the co-primary endpoints in the broad study population. What we did see was confirmation of the role of P. gingivalis in Alzheimer's. We demonstrated a greater than 50% slowing of cognition loss in prespecified subpopulations representing about half of the overall population. Our ability to identify patients likely to benefit and show modification of their disease confirms our development strategy and that we are going to turn the tide of Alzheimer's disease and help patients maintain their cognition and, frankly, their precious memories a little bit longer. Now the agenda for the call will begin with Casey and her comments about our development strategy and what is indeed this milestone moment from the GAIN trial. Then Mike will walk through some of the specific data we released today, bearing in mind that we have a very detailed 30-minute presentation at CTAD on November 11. Today's presentation is simply meant to share and preview the top line result and, most importantly, the available data which is driving our development strategy today. After some closing remarks, we'll then take your questions as time allows. Before I turn the call over to Casey, I just want to add, on a personal note, I'm overwhelmed by the number of people flooding my e-mail, my phone and over the last 15, 20 minutes with messages of congratulations. It's truly humbling to be in this position to be a part of a team that is indeed making a difference in patients' lives. It literally gives me goosebumps thinking about the moment that I first heard about this data set. Casey, can you walk us through some of the highlights?

Thank you, Chris. Yes, this is an incredible milestone that we've achieved in our understanding of Alzheimer's disease and the profile of atuzaginstat. And I want to recognize the entire Cortexyme time team, the investigators and the participants in the study for doing such an incredible job and dedication to this study, especially during the global pandemic. As many of you know, this was the first study of this mechanism of action and this molecule. We took all comers into this study. We did not want to leave anyone behind that could benefit from treatment, and we also didn't want to pick a companion diagnostic at priority that might not be the right one to take forward. So we prespecified this. A huge goal of this study was to understand, do we need a companion diagnostic? Which one is the best? Is saliva the appropriate sample to look at? Do we need to look at systemic infection? Or do we need something more invasive from a cerebral spinal fluid sample, for example? And the great news is that we are fortunate that the best marker appears to be the least invasive, which is a simple saliva test for P. gingivalis DNA. So we're not surprised to see that this was a prespecified subgroup of responders. Again, we saw a 57% slowing with a p-value of 0.02. And we saw similar results in other infection-related subgroups, which are very consistent with this result and support the finding. Another very important and convincing analysis to us that was prespecified is to look at reduction of P. gingivalis and see if it correlated with improved outcomes. And indeed, as Mike will show you, we know when we are really reducing P. gingivalis, we're seeing benefits across multiple fixed scales very consistently, which is just really exciting confirmation that our hypothesis that driving down P. gingivalis load is going to slow or stop progression of disease. This is really exciting to see also in a mild-to-moderate population. This has been a traditionally very hard-to-treat population but the most needy population for a treatment, and we are seeing signals in both the mild and moderate cohorts. Also recall, this is an oral drug, so a twice-a-day pill with a differentiated safety profile that doesn't require MRI monitoring. So as far as the next steps, we are poised to execute very quickly on a confirmatory study. We know the right patients. We know the right dose. We know the best sites. And we can execute on that very quickly. That said, we're going to leave no stone unturned. We know atuzaginstat is benefiting patients, and we'll be engaging with global regulators very quickly to determine the most rapid path forward to advance this molecule. I'll turn it over to Mike to talk about some of the details.

Great. Thank you, Casey. I don't think it's an exaggeration at all to say that the GAIN trial has yielded the most impactful data of my career in clinical drug development. On the scientific side, there's evidence for a novel root cause of Alzheimer's and true disease modification; and on the clinical side, enormous impact for patients, caregivers, doctors and others. We developed atuzaginstat, formerly COR388, to effectively target and inhibit P. gingivalis, and these rigorous clinical data add to the extensive existing database that it is upstream of Alzheimer's disease pathology. Atuzaginstat is a first-in-class, orally administered, brain-penetrant, small molecule. It effectively blocks gingipains, the protease virulence factors, which are secreted by P. gingivalis and are required for survival but also responsible for the toxicity in Alzheimer's. Now for the study highlights. On the first slide, we see some of the demographics. There are 2 key takeaways from this slide. One is that we identified a definitive Alzheimer's population. You can see on the right that just about half of the patients were in the mild range and half in the moderate range on the MMSE rating scale for severity of Alzheimer's. And you can also see that about 3/4 or 2/3 of these patients were ApoE4 carriers, which is exactly what is commonly seen in mild-to-moderate Alzheimer's. The other key takeaway is that they were randomly distributed well across the placebo, low dose at 40 milligrams twice a day and high dose at 80 milligrams twice a day doses. Moving on to the next slide. This is a key slide for understanding the efficacy of atuzaginstat in the patient populations with higher levels of PG infections. As Casey mentioned, these are both prespecified being core parts of the hypothesis of P. gingivalis and atuzaginstat. On the left, you have the patients that had PG, which was detectable in their saliva. And on the right, you have the upper half of patients with evidence of infection, PG IgG in their serum or blood. And just above them, you can see there's -- the interaction effects show that the atuzaginstat was statistically significantly more effective in slowing the progression of cognitive decline in the high-PG group than in the low-PG group with a p-value of 0.02. And similarly, in the high -- anti-PG IgG group on the right, there was a statistically significant interaction again of 0.03, again indicating that atuzaginstat was statistically significantly better at slowing cognitive decline in the high-serum anti-PGG group. You can see on the upper left that in the primary analysis and population that there was 57% slowing in the high-dose group, which I'll focus on for here, and this, too, was statistically significant at a p of 0.02 in this subpopulation. On the right, you see a 37% slowing in the high-dose group in the anti-PGG IgG group. At the bottom are nonparametric analyses. This, too, was prespecified as a different kind of a statistical analysis, which we used because we saw that there were some evidence of nonnormality of distribution. And when that is seen, nonparametric analyses are used primarily to assess whether or not the effect size seen was approximately -- was impacted negatively by the nonnormality or outliers in the distribution. And you can see on the lower left, again, 56% slowing in the high-dose group; and on the lower right, 50% slowing. These data support the findings with the MMRM parametric approach above. So again, a lot of convergence of evidence here, 2 different methodologies for identifying the high-PG infection patients, both interactions and subgroup analyses and both parametric and nonparametric analysis showing a convergence of a lot of data supporting these findings. On the next slide, this, too, is really compelling data that Casey alluded to earlier. This is, again, prespecified, and this is the correlation between the levels of P. gingivalis in the saliva at 24 weeks. That's halfway through the study, about 6 months. And the reduction in PG at that point in time predicts the efficacy on multiple clinical endpoints, both at that time halfway -- half a year through the study and also later at the end of the study, 48 weeks or roughly a year at the end of the study. And you can see that all of these bars are on the right, which indicate that for all of these comparisons, the reduction in PG correlated with better clinical outcomes. Likewise, worsening of PG correlates with worse clinical outcomes. You can see that the p values for 6 out of the 8 comparisons are statistically significant at less than 0.05, several of them much lower than that and with the exception of the ADCS-ADL, which showed trends in the right direction, but they did not reach the traditional nominal level of statistical significance. So again, abundant converging evidence. On the next slide, turning to the summary of safety. We can see that the most common adverse events were gastrointestinal in nature, including diarrhea, nausea, transaminase elevations and some other GI findings. But the most common adverse events were present in about 12% to 16% of patients, as you can see for diarrhea, which is a very nice range to be in. So we consider it to be well tolerated overall. There were dose-dependent liver enzyme elevations, as referred to here. We have worked with hepatic experts who have said that based on these data and others, they are confident that simple implementation of a titration of the dose at the beginning of treatment should reduce these transaminase elevations. These lab abnormalities in things like ALT, AST, lipase and amylase were in virtually all cases, both clinically insignificant, and patients were completely asymptomatic. And in all cases, they were covered fully with no long-term sequelae. And finally, and importantly, we saw no signal on ARIA, the amyloid-related imaging abnormalities, which have been seen with amyloid-related molecules, and that can cause microhemorrhage or edema or related superficial siderosis. Next slide, please. So to summarize, the co-primary endpoints were not met in the broad intent-to-treat study population. However, the randomized -- these randomized, double-blind, placebo-controlled trial results support the extensive research already existing establishing P. gingivalis as a root cause of Alzheimer's disease. These PG, Alzheimer's patients are easily identified through saliva or simple blood tests, offering a convenient level of precision medicine, which is atypical in Alzheimer's disease. Prespecified interaction analyses showed significant interaction of high-PG subgroups with better outcomes. Likewise, prespecified subgroups with PG DNA detectable in saliva showed statistically and clinically significant superiority of atuzaginstat over placebo on cognitive decline. And multiple prespecified statistically and, again, clinically significant correlations were seen between the reduction of the bacterial load and improved clinical outcomes on several scales. And last, a clearly differentiated safety profile, the oral route of administration, the patient population, both the disease severity of mild to moderate and the evidence of PG infection, separates this drug significantly from other drugs that are available and in development that target amyloid. Thank you very much. And now back to Chris.

Great. Thank you, Mike. So look, our next steps are very clear. We're going to share this data set with global regulators and collaborate on the most efficient path to help patients. The data set, when you look at it in its totality, it's a very rich data set, which is indeed the capstone of over 7 years of research. It's another compelling layer of evidence about the role of P. gingivalis in Alzheimer's disease. We know that the need for disease modification grows every day. It's why we work diligently on behalf of our patients. Our mission has not wavered. It's never wavered and today represents a giant step towards stopping Alzheimer's disease. Now we're well capitalized, and we'll be able to promptly execute on the next steps of development for atuzaginstat. Our current cash stands at just over $140 million as of September 30, and it's sufficient to fund operations through 2024. And lastly, and most important for many, I've had the unique experience of working these past few years on the research side of the house with the company while also being a caregiver for my parents who are victims of Alzheimer's, both of them. I've seen firsthand the progression from early onset to mild to moderate to end stage. This is a disease of cognition. So our ability to clearly identify our population have a meaningful benefit to the cognition of these patients is nothing short of an incredible advancement in understanding this disease. Thank you for your attention. Thank you for your support. We would now be pleased to take any questions you have.

[Operator Instructions] Our first question is from Jason Butler with JMP Securities.

Just first one, just thinking about the saliva -- measuring PG in the saliva, so all patients essentially had anti-PG IgG, but only 40% had PG in the saliva. Any thoughts on what explains the difference and why saliva is a better measure? And then you're obviously showing concordance of effect between saliva and serum. Did you also look at concordance with PG and CSF?

Yes. Thanks for the question, Jason. So the PG detection in saliva is based on a rinse. It's about a 5-milliliter saliva saline rinse. And so there are certain people who are above a detection limit. We know that rinse dilutes the sample a little bit, so a drool sample captures more like 50% of patients using this rinse. It's about 37% of patients in the study. So this is directly measuring live PG, PG DNA. The IgG test is different. It's measuring a host response to P. gingivalis. And that could be -- in the past, it could be acute. So it's a different measure, and we are very interested in it because it's more of a systemic marker. And it turns out, as you noted, both are good. The PG DNA is very nice because you can look at reductions, and we do see a nice dose-dependent reduction with the drug. We are still looking at the CSF data. That assay is a little bit more involved. But we do believe that, that one also is consistent, and we're still just dotting Is and crossing the Ts.

Okay. Great. And then...

But the nice thing -- sorry, just the nice thing is we don't need to go to an invasive test like that.

Got it. Great. Just in terms of the correlations you're seeing with the PG reductions and the clinical markers, you're showing reductions at 48 weeks to predict -- sorry, 24 weeks to predict 48-week clinical efficacy. Do you see similar results if you look at reductions in PG at 48 weeks versus baseline? Or I guess why did you look at the 24-week reduction to be predictive of 48-week benefit?

We plan on looking at the change to 24 weeks because we knew that should be enough time for the drug to have the benefit on P. gingivalis reduction. So that was our plan all along. Also, the end-of-study biomarker samples are still under analysis. So we can look at those, but actually, our plan had to look at the 6-month samples for PG reduction, which should have a downstream effect on cognitive endpoints, both at 6 months and 12 months, which is what we saw.

And I would just add that, obviously, seeing the temporal pattern of the PG change at week 24 and the clinical results, both at that time and later, it doesn't absolutely prove causality, but it's certainly consistent with and very supportive of the causality relationship. So we'll look at a number of different temporal relationships, and we look forward to sharing some more data with you at CTAD and even more after that.

Great. Last one for me. Could you talk through the liver enzyme signal a bit more? When is this occurring after initiation of treatment? Is it all early? Or do you see any patients that don't have enzyme elevations until later in treatment? And then any more details on the patients that also have the increased bilirubin?

Sure. The timing is a really important point. I'm glad you brought that up. These tend to peak at about 6 weeks after initiating drug with a lot of drugs that cause transaminase elevations, and that's great because once you get past 8, 10, 12 weeks, you're really -- the risk at that point is close to 0% for -- so both doctors and patients can relax after that. And yes, the 2 patients that had bilirubin elevations, they both recovered fully with no long-term sequelae. And the, really, big-picture, important thing to take away is the overall risk-benefit relationship. We think with 50% of the population slowing cognitive decline by 50-plus percent in a mild-to-moderate population, it's never seen that kind of efficacy before. We're very pleased, and we think the overall risk-benefit profile is very, very attractive.

Our next question is from Tazeen Ahmad with Bank of America.

I have a few. Can we start with the ADAS-Cog versus ADCS-ADL? How important is it that you were stat sig for the PG subpopulation for ADAS but not for ADL? If it's the case that there were co-primary end points, how does that become a topic of discussion with FDA?

Well, first of all, the [indiscernible] of Alzheimer's is cognition. It's a disease of cognition. So the finding on ADAS-Cog is very impactful. The findings on ADCS-ADL, as you saw, the correlations were in the right direction, so supportive of the finding. But it's pretty well known in the literature that the activities of daily living, ADLs, usually take longer periods of time to see a signal on, and they're challenging to see our KOL, so you never see improvement on those. You're just -- the best you can hope for is some slowing. The -- so...

Yes. The other thing that we're digging into a little more is the impact of the pandemic and quarantine. There are many questions in ADL that relate directly to this. For example, did you go to the grocery store? Did you have a haircut -- did you keep a haircut appointment? Did you travel? So that scale, in particular, has a lot of questions that may have been impacted in unpredictable ways by quarantine, and we're starting to dig into the data to understand more about that as well.

And finally, the CDR is also an overall functioning or global efficacy scale, and seeing good signals on that were also reassuring to us.

Okay. But did FDA ask you to look at both of these measurements? Or was it a choice that you guys had made?

The FDA's position, I think this is published in their guidance documents that a well-validated cognitive scale like the ADAS-Cog is needed in this population of mild to moderate and a co-primary that is also a measure of function or global improvement like either the ADCS-ADL or the CDR. So that's -- I think that's the FDA's position on it, which is in their guidance.

Okay. Now is it correct that you were stat sig at the PG subset level but not on the IgG subset? And if that's the case, any idea why that would be?

The important thing to look at in both of those first is the interaction term, which was statistically significant in both groups. And thus, the interaction -- the statistical interaction shows you that the drug was statistically significantly better in the high-PG group than in the low. Then what -- after that, you can look at the subgroups, but you have to realize that each time you're looking at a smaller subgroup, you have a smaller and a less statistical power. So not seeing statistical significance on every one of those subgroups is not at all damning to the overall hypothesis. The interaction term is really the primary thing to focus on.

Okay. Then when you refer to the data today being addressable and, let's say, 50% of the Alzheimer's population potentially, is that inclusive of this IgG subset?

Yes. The IgG subset supports the fact that -- and what we looked at was a median split here. So it was just about 50% in the high-IgG and low-IgG subset. And as Casey was saying earlier, we saw 37% of patients with the PG DNA in their saliva, but we do have some plans to even improve upon that assay going forward. And we already have some data with this better form of -- non-rinse form of collecting raw saliva that identifies at least 50% of well-characterized Alzheimer's patients.

Okay. And then my last question, I promise. As it relates to the data presented today, how should we be thinking about the step that would be needed to remove the clinical hold that's currently in place for the OLE portion of the GAIN study?

Yes. I think we're going to sit down with regulators. I mean FDA was -- certainly wanted to see the efficacy from this study. And we are already in the process of scheduling that meeting to be able to show them the data that we have. And we'll update you promptly as soon as we have a little bit tighter window on timing, for sure.

Our next question is from Sumant Kulkarni with Canaccord.

Yes. I have a few. So first one, how did the stage of disease split in the patients with P. gingivalis detectable in saliva? And what fraction of the P. gingivalis saliva positive patients also tested positive for IgG in the CSF scans?

We haven't looked at the exact overlap between the 2 subgroups yet, but we'll again be sharing more at CTAD as we dig through the data, look at the data in more detail. The first question was about the mild to moderate, I think. Can you repeat the first part?

Yes. Well, the first part was how did the mild-to-moderate split look in the patients with P. gingivalis detectable in saliva?

Yes. It was very similar. The PDGS subgroup, as we're calling it, P. gingivalis detectable subgroup in saliva subgroup was very well balanced for ApoE4 and on the MMSE scale as well. So that was nice to see.

Was there any correlation between the severity of gum disease versus the effect of atuzaginstat?

I'm sorry. There was some static. I couldn't...

There was some static. Can you repeat, Sumant?

Sorry. Was there any correlation between the severity of gum disease versus the effect of atuzaginstat?

Yes. We need to look at that in additional detail.

Okay. And my last question for now is in the fraction of patients that did not have P. gingivalis present in saliva, how did the atuzaginstat treatment groups perform?

Again, we'll be addressing that more at CTAD just to show everyone all the different graphs and subgroups.

Our next question is from [ Miriam Delgati ] with [ Lakeside Capital ].

My first question is along with the ADAS-Cog11, what do you see as the most important confirmatory data points among all the other secondary endpoints you looked at?

Yes. I think the -- well, I think there's a couple of ways of looking at this. One is looking at different scales, as you implied. And as we saw in the correlation findings, the ADAS-Cog; the CDR Sum of Boxes, which is in global function scale; and the MMSE, which is primarily another cognition scale all showed statistically significant and clinically significant correlation with reduction in P. gingivalis. But it's also important with looking at subgroups to look at multiple different definitions of the high-PG infection, which we've done some different statistical analysis methods like parametric and nonparametric. And all of this makes for a very compelling case with the full weight of evidence for this hypothesis showing that PG is a root cause of Alzheimer's and that knocking it down is a way to bring about really significant clinical slowing of cognitive decline.

Got it. I was also wondering if you think the data set you provided today is enough for filing for approval.

Yes. So [ Miriam ], thanks for that. It's Chris. I think, again, we're going to -- I think we obviously have conviction around how this hypothesis has been proven with this evidence, and we look forward to sharing the evidence with regulators to see what makes sense next.

Our next question is from Charles Duncan with Cantor.

Super. I don't currently cover the stock, but I'm intrigued with the results, definitely. Congratulations on the observation. First question, though, is -- and I know you're going to be presenting this at CTAD coming up here, but the effect over time, I'm wondering if you can provide us any kind of insights or color on the temporal pattern for the activity over time as you evaluated these patients.

Thank you, Charles. Yes, we'll be sharing more at CTAD, but it does look like a disease-modification effect, as you'd expect, with a change in slope.

Okay. Very good. That's helpful. Second question is if you were to do another study, would you do an all-comer study? Or would you do one based on the diagnostic test that basically you have -- to have that be a confirmatory step? And if -- just given the effect size here, which is pretty interesting, what size of another study would you -- what would be the sample size of another study if you were to move forward with that?

Yes. That's a great question. I mean we're still digesting the data, and there's a lot to look at. But clearly, we found our patient population that can benefit from this drug. And we would certainly be using one or both of the PG detectable in saliva or the IgG test to identify and a focused precision medicine patient population probably using the CDR Sum of Boxes as the global co-primary, improving the saliva assay as we've talked about. And with all those things in mind, we'll have to sharpen our pencils and do some thinking about it in the exact numbers. But you would expect with all those things improving signal detection that you could still -- you could see a really strong result and a statistically and clinically significant results with a smaller population. So we'll be thinking about that more as we go forward.

Yes. It would seem to me, given the elderly population, you wouldn't want to rely on a host response, but the actual saliva test may be more useful. Question on GI tolerability. Any pattern on that? I guess, similar to Jason's earlier question regarding transaminase change, was there any pattern in terms of GI tolerability? Did it subside?

Yes.

Or improved over time?

As I mentioned earlier, most side effects tend to be early in treatment, and certainly, the transaminases tended to peak about 6 weeks after starting drug. We know that P. gingivalis colonizes both the mouth and the rest of the GI tract. So we're not surprised to see that knocking out P. gingivalis changes the microbiome a little bit, and you get GI side effects. However, overall, the most common side effect being in the mid-teens is a side effect profile that a lot of drugs would be very envious of.

For sure and especially given the emerging standard of care in Alzheimer's. Last question for Chris. I guess you mentioned, Chris, cash through, I think, '24. Does that contemplate additional clinical work with this candidate, either in Alzheimer's or any other neurodegenerative disease?

It does, Charles. That was always kind of in our base plan, the idea that there would need to be a second study-based financial plan. So we -- with our runway, we've assumed an equivalent study in terms of expense to what GAIN was. So absolutely, we feel very comfortable with our runway, and I think the efficiency of the company that it's had historically will continue to serve us well as we move forward with this set of data.

We have reached the end of our question-and-answer session. I would like to turn the conference back over to management for closing remarks.

Great. Thank you. So thanks, everybody, for joining today. Look, I think to put this in perspective that in the last 7 years, this company has -- with its efficiency clinically, has come further than many other targets come in decades. And so to sit here today and to be able to clearly identify the patient that is likely to respond with a greater than 50% slowing of cognition in this huge unmet medical need and, in fact, one of the leading killers in the world, this, to me, is a very great place to be today as we move forward, and I feel like we've done the industry a great service today. So thank you for joining. As Mike and Casey both alluded to, we will have a much more detailed update at CTAD. And we will also have more data coming out from this study in the weeks and months after that as well. Thank you again for your support, and we look forward to speaking with you soon.

Thank you. This does conclude today's conference. You may disconnect your lines at this time, and thank you for your participation.