Lexeo Therapeutics, Inc. (LXEO) Earnings Call Transcript & Summary

Good morning, and welcome to the Lexeo Therapeutics webcast presentation of interim Phase I/II clinical data of LX1001 for the treatment of APOE-associated Alzheimer's disease. [Operator Instructions] As a reminder, this call is being recorded today, Wednesday, October 30, 2024. I would now like to turn the conference call over to Rand Monaghan, Vice President of Finance of Lexeo Therapeutics. Rand, please go ahead.

This morning we issued a press release outlining the interim clinical data, which is available on our website at lexeotx.com and filed an 8-K with the SEC. Joining us on today's call will be Nolan Townsend, Chief Executive Officer; and Dr. Sandi See Tai, Chief Development Officer. Also available for questions is Dr. Sam Gandy, Mount Sinai Professor of Alzheimer's Disease Research and Chairman Emeritus of the National Medical and Scientific Advisory Council of the Alzheimer's Association. Before we begin, I would like to remind you that this call will contain forward-looking statements for the purposes of the safe harbor provision under the Private Securities Litigation Reform Act of 1995. Actual results may differ materially from those indicated by these as a result of various important factors, including those discussed in our filings with the SEC. With that, I will turn the call over to Nolan.

Thank you, Rand, and thank you, everyone, for joining us. Earlier today at the clinical trials on Alzheimer's Disease Conference, we are pleased to release interim data from the Phase I/II clinical trial of LX1001 for the treatment of APOE4-associated Alzheimer's disease. LX1001 is a unique genetic approach to treating this high-risk category of Alzheimer's patients. And the interim data we will discuss today show a highly encouraging and consistent pattern of reduction across CSF tau biomarkers and tau PET imaging alongside APOE2 protein expression in all participants. Furthermore, we observed a favorable safety profile with no reports of ARIA. Before we dive into the data today, I'd like to quickly remind everyone of our broader genetic medicines pipeline. LX1001 is the most advanced program in our pipeline of candidates targeting APOE4-associated Alzheimer's disease. And the Phase I/II study completed dosing late last year. We also have 2 preclinical programs in this therapeutic area, LX1021, which incorporates the potent Christchurch mutation into the APOE2 transgene and LX1020, which utilizes microRNA to suppress APOE4. We also have a pipeline in cardiovascular genetic medicine, including 2 clinical-stage programs, one in Friedreich's ataxia cardiomyopathy and the other in PKP2 Arrhythmogenic Cardiomyopathy. Importantly, we retain global rights to all programs in our pipeline. I'd like to now share more background on APOE4 homozygotes or those with 2 copies of the APOE4 allele. We estimate that there are approximately 900,000 APOE4 homozygotes with Alzheimer's in the U.S. and this population is the highest genetic risk factor for developing Alzheimer's disease, approximately 15x greater than the general population. Nearly all APOE4 homozygotes will develop the disease in their lifetime, with earlier symptom onset and faster rate of cognitive decline than other APOE genotypes. Importantly, APOE4 homozygotes also have limited treatment options as the currently approved amyloid-targeted therapies carry an increased risk of amyloid-related imaging abnormalities known as ARIA, which includes brain swelling and microhemorrhages and the risk of the adverse events are the highest in the E4/E4 population. For this reason, these therapies carry box warnings recommending genetic testing to check the E4 allele before prescribing as well as MRI monitoring of E4/E4 patients should a physician choose to initiate therapy. Moreover, the anti-amyloid therapies also demonstrated lower efficacy on cognitive measures in the E4 homozygotes group. Given these safety risks and lower efficacy for E4/E4 patients, we believe that there continues to be high unmet need for these patients. To understand what is happening at the biological level, we first have to understand the role of APOE4, which is a driver of Alzheimer's disease and can independently impact amyloid pathology, tau pathology, neurotoxicity, neuroinflammation and eventual synaptic and neuronal dysfunction. These underlying genetics are not currently addressed by existing therapies or antibody mechanism of action and are the thesis for our approach that acts upstream of any particular pathway, potentially modifying pathology at multiple levels. In contrast to these toxic pathways, the APOE2 isoform has been shown to independently drive many neuroprotective pathways. This is most evident by the fact that E2/E4 heterozygotes have lower risk of Alzheimer's onset and slower disease progression and E2 homozygotes have significantly reduced risk of Alzheimer's disease compared to other genotypes. In terms of amyloid biology, the buildup of A-beta plaque can begin years or even decades before the onset of Alzheimer's symptoms. Though expression of APOE2 may not clear preexisting amyloid plaque, there's evidence suggesting that it can suppress A-beta toxicity and therefore, indirectly impact tau, the well-known driver of pathology in this disease. Independent of amyloid, there is also increasing evidence that APOE2 directly modulates tau toxicity by regulating tau phosphorylation, accumulation and inflammation. Tau pathology occurs later in the course of disease, and as we've come to understand in recent years, may be more closely correlated with cognitive decline in Alzheimer's disease. So if we can impact tau, we believe we are likely to address cognitive decline as well. Ultimately, the hypothesis we are exploring is that the delivery of APOE2 will increase the ratio of E2/E4 in an E4 brain, which will simultaneously dilute the effects of APOE4 toxicity while adding the protective benefits of APOE2. Similar to other recessive genetic diseases, we believe that a small amount of normal protein can provide meaningful clinical benefit. This is why we believe delivery of LX1001 could act upstream of any particular pathway or target, potentially mediating both efficacy and tolerability. Here, you can see that hypothesis in the LX1001 mechanism of action. We're utilizing the AAVrh10 vector administered via intrathecal or intrasternal injection into the central nervous system, to deliver the protective APOE2 allele for APOE4 homozygotes. To our knowledge, this is the first and only clinical-stage program -- clinical-stage gene therapy program, targeting the underlying genetic cause of APOE4-associated Alzheimer's disease, with a clinically validated capsid with a favorable safety profile and a differentiated single administration therapy. In this initial LX1001 Phase I/II study, we had several primary goals. We first wanted to demonstrate APOE2 expression in the CSF, observe signals of downstream biologic activity and demonstrate that LX1001 administration was well tolerated, which is especially important in this population that has been associated with higher ARIA rates. Overall, we are pleased with the results we have observed, and we believe that our objectives have been achieved and that this results clearly support further evaluation. I will now turn it over to Dr. Sandi See Tai to review the interim data.

Thanks to you, Nolan. On this slide, you can find the key details of the Phase I/II study design. It's a 52-week open-label trial that enrolled 15 APOE4 homozygotes with mild cognitive impairment, mild dementia or moderate dementia due to Alzheimer's across 4 ascending dose cohorts. Enrollment was completed in the fourth quarter of last year. And today, we will be sharing 12-month safety and biomarker data from cohorts 1, 2 and 3, and 6-month data from cohort 4. I will note that although 15 patients were dosed with LX1001, 2 patients did not complete the protocol-required visits or testing at 6 and 12 months. So we have noted the end for each analysis on the following slides. Here you can see the baseline characteristics for all 15 patients dosed in this study. Importantly, you can see that the patients with moderate dementia were clustered in dose Cohorts 1 and 3, whereas Cohort 2 primarily included patients with mild cognitive impairment. So keep that mind as we review the data, as you will see a general pattern of greater treatment effect in these moderate dementia patients. These patients overall also demonstrated high amyloid and tau burden at baseline. Moving on to safety. LX1001 was well tolerated across all 4 dose cohorts. And importantly, there were no reports of ARIA, which was an elevated risk for APOE4 homozygotes in the amyloid-targeting therapy studies, as Nolan discussed earlier. Transient CSF pleocytosis was observed, but it was not correlated with dose levels and most resolved by 12 months with no significant adverse events associated with the pleocytosis. 4 serious adverse events were reported in the study, 3 were assessed as unrelated to treatment and 1 event of mild to moderate sensorineural hearing loss was assessed as possibly related to treatment. I can share that this patient's hearing is reportedly improving and repeat audiometry is pending. Finally, cognitive testing was performed primarily for safety in this study. No safety signal was observed across the ADAS-Cog13, CDR-SB or MMSE scales. No clear pattern of change was observed across these scales, which was not surprising, given that large sample sizes are needed to assess Alzheimer's compression given the wide inter-individual variability. You can see the mean change from baseline at 12 months in the LX1001 study across the measures below. And on average, these rates of progression are generally in line with what you would expect to see based on reported data in patients with early Alzheimer's and high tau burden. We hope in the next study to further evaluate these measures with a larger trial size. Now looking at APOE2 protein expression, it's important to note that our goal in this study was to look for expression with increases in a dose-dependent manner and then observe the impact on downstream markers associated with Alzheimer's pathology. We did not have a predefined target level of expression in mind. And ultimately, we did achieve our objective of seeing expression alongside corresponding improvement on biomarkers, which we will show you later in the presentation. On this slide, you can see APOE2 protein expression in the CSF, which was increased in all participants with a dose-dependent increase and durability out to 12 months. This demonstrates that we are achieving both transduction and expression of APOE2 in the brain as expected. In the middle chart, you can see that APOE4 expression is largely stable between baseline and month flow. Looking at the ratio of APOE2 expression relative to the existing APOE4, we then see a consistent pattern of dose and time-dependent increase in expression with continuing durability at month 12. We did have 1 patient in Cohort 3 with the highest expression, which we have called out on the slide and not included in the mean as we do not have 12-month data from this patent. Of note, you will see on the following slides that these protein levels appear to be associated with improvement in key clinical biomarkers, which we will review. Looking first at amyloid, let me orient everyone to this slide. Each bar represents an individual patient in the study and their change versus baseline. All bars are at last follow-up with most at 12 months, except Cohort 4, where we are still awaiting the 12-month data and 1 patient in Cohort 3, that outlier with the highest expression who unfortunately did not consent to the lumbar puncture at 12 months. So looking first at A-beta 42/40 ratio, LX1001 appears to stabilize A-beta 42/40 in the CSF as most patients are staying relatively close to baseline. This result aligns with our understanding of APOE2 and its role in the normalization of soluble A-beta. In amyloid PET on the right, we see it again, the majority of patients, 8 of the 11 have minimal change from baseline. So again, amyloid progression appears to be relatively stabilized. This next slide further cements our excitement on the data. In these 2 charts, we see a consistent reduction in CSF T-tau and P-tau181 in the majority of patients, 9 of 13, and across all dose cohorts. These data demonstrate that the APOE2 protein being expressed by LX1001 is having the desired impact on downstream biology linked to disease progression. In an effort to put these reductions into context and with limited longitudinal data on the natural history of progression in APOE4 homozygotes, we have concluded some references to certain mean data reported from the lecanemab Phase III study on the CSF biomarker measures, which can be seen as the dotted blue line on the charts. Of course, these are different studies with different designs, inclusion criteria and patient baselines. And in particular, we would point out that the mean data shown from the lecanemab trial includes all APOE genotypes and does not include any patients with moderate dementia at baseline. The top dotted line reflects the reported change from baseline in the placebo arm of the lecanemab Phase III study, which we believe highlights the natural trend of increasing tau pathology in Alzheimer's disease. Now when looking at LX1001, we are seeing a reduction or an improvement in over 2/3 of our patients from baseline. We are also seeing the most significant treatment effect in patients with moderate dementia, the light blue bars, which is notable as it demonstrates our potential to treat a population with more severe disease that is typically excluded from clinical research. On this slide, we want to share additional measures of CSF P-tau that we believe confirm the pattern of improvement shown on the previous slide. These measures were added to the study, and today, we only have data to share for Cohorts 2 and 3 as results from Cohort 4 are not yet available. But again, you can see improvement in the majority of patients relative to baseline. For CSF P-tau217, you can also see improvement in 5 of 6 patients relative to natural history data in Alzheimer's disease, although this is not specific to APOE4 homozygotes which may demonstrate even faster progression without treatment. Overall, the consistent pattern of improvement we are seeing here across all CSF tau biomarkers and across the majority of participants gives us confidence in the potential of LX1001 to have a reliable and consistent biological impact on tau pathology. This is critical because accumulation of tau is correlated with cognitive decline in Alzheimer's disease, and we believe that if we can impact tau, we may be able to impact cognitive and functional decline over time as well. In tau PET, we are also seeing consistent reduction, which is encouraging because, of course, these results measure tau burden in the brain and not just CSF. Tau PET analysis was added to the study after Cohort 2, so we are only able to share results from Cohorts 3 and 4 today with 12-month data for Cohort 4 still pending. On the left, you can see that, at 6 months, tau burden is reduced globally from baseline in 5 of 6 patients, and we have also highlighted the outsized reduction of the Cohort 3 participant with the highest APOE2 expression. On the right, you can see the 12-month imaging from another patient in Cohort 3. Tau is represented in red on the images, and you can see this participant showed noticeable improvement in tau burden 12 months when compared to baseline. On this slide, we are showing reductions in tau PET across all regions of interest. The majority of patients, 4 out of 6 in this sample, experienced reductions in tau burden in all 6 brain regions, and we look forward to understanding the pattern observed at 12 months once we have additional data from Cohort 4. So to summarize our finding, LX1001 has been well tolerated across all dose cohorts with no reported events of ARIA. We have achieved APOE2 expression in the CSF of all participants with a dose and time-dependent increase in E2/E4 expression with stabilization of amyloid pathology and a consistent reduction in key tau biomarkers. Together, these data give us confidence on the ability to utilize gene therapy, delivering APOE2 to treat APOE4-associated Alzheimer's disease and support continued development. Finally, to close this section, we'd point out have some of the themes we've observed over the last several years as it relates to our regulatory strategy for this program. The first is the heightened collaborative environment in Alzheimer's and in gene therapy from CEBR, which highlights the potential for accelerated approval based on surrogate endpoints and to which there is regulatory precedent using amyloid PET. The second tenet of the strategy is confidence in exploring tau PET as a surrogate endpoint for Alzheimer's disease. Studies have shown that tau pathology is closely correlated with cognitive outcomes in Alzheimer's disease, essentially more closely correlated than amyloid, and we believe this may be an appropriate surrogate endpoints for future studies. We have initiated FDA engagement to discuss the next phase of development. so we expect to provide an update on regulatory progress in 2025. With that, I will now turn it back over to Nolan to close out today's call.

Thank you, Sandi. Like you, I'm very encouraged by the interim data shared today and the potential for LX1001 to deliver a targeted approach to treating APOE4-associated Alzheimer's disease. As we've shown today, we believe this unique mechanism has the potential to act upstream of Alzheimer's biology and impact multiple pathways associated with the disease. And we believe the biomarker data presented today validates this unique exciting approach for APOE4 homozygous patients for whom the limited treatment options available today may not be suitable. In terms of next steps, we have initiated engagement with the FDA, which will determine the next development steps for the program, including the possibility for accelerated approval. In addition, ongoing business development discussions are underway. We expect to provide an update on regulatory interaction and further LX1001 plans in 2025. Thank you for your time this morning, and we will now take questions. And as mentioned at the top of the call, we are joined by Dr. Sam Gandy as well for the Q&A.

Your question comes from the line of Tess Romero with JPMorgan.

A couple from us. So first on the reported safety, what makes you believe the event of mild-moderate sensorineural hearing loss is possibly related to treatment? And can you provide a little bit more additional color on the timing of this event? And as you think about what the etiology could be here, how do you and your investigators think about this? And we just wanted to clarify, were there any other treatment-related AEs? And then we have one follow-up.

Okay. Thanks, Tess, for the question. I think I'll ask Sandi to comment on the hearing loss adverse event and the other SAEs.

Yes. So thanks for the question. So just to -- and to that question, this event occurred approximately 3 months after dosing. I think what led to us saying that it may be possibly related is that it was a relatively acute onset, though I think that we have ruled out other causes. So it's not clear that this particular patient has a family history of hearing loss. What we do know in follow-up is that this patient appears to have some improvement in their hearing loss. And so a repeat hearing testing is pending. In terms of the etiology, we've tried to explore -- we've explored the literature as well as spoken with several experts in the area. We've also reviewed the event with our data safety monitoring board. And at this stage, do not feel that it required any changes to our ongoing protocol. And we certainly continue to monitor the safety in the study, and in particular, with the follow-up for this patient. And in terms of other treatment-related AEs, we have not seen any pattern or reports in that regard.

Okay. Great. Very helpful. And as you think about the path forward here bridging off this data, aside from moderate dementia patients observing it looks like a more pronounced response to treatment, are there any other characteristics of patients that could be a predictor of response?

Thanks, Tess. Maybe I'll take that one. I think what we were excited about in this data set was that we were seeing an impact on tau biomarkers that is roughly equivalent to some of the commercially approved therapies, but we are seeing this without events -- safety events of ARIA, which is obviously very pronounced in some of the commercially approved therapies. We're seeing this with the APOE4 population, which today, the benefit-risk profile and being treated with some of the commercially approved therapies may not make sense for that group. So I think today, in terms of exploration, obviously, tau becomes a very important theme for the path forward and as a surrogate endpoint and obviously, the more pronounced treatment effect in moderate is something that is notable. I would also note that moderates have not been well studied in some of the other therapies. And so that is a possibility for a population of very high unmet need as well. So I would just say with that summary, that's the direction of travel and more to come on earlier populations. But I think we can see here that probably the treatment specs across the moderates is one that needs to be further explored.

Your next question comes from the line of Paul Matteis with Stifel.

This is James on for Paul. Just 2 quick ones from us. One, I guess, based on these data, how are you thinking about kind of taking a go-forward dose? And what is the right dose to kind of carry forward? And two, I guess, just what level or have you had any sort of indication from regulators yet around the potential of tau as a surrogate?

Yes. So thanks for the question. I think one of the elements of this study is that we enroll patients across the range of the disease. So as you've seen, Cohorts 1 and 3 primarily consist of moderate patients, Cohorts 2 and 4 primarily consists of mild MCI patients. So I think, in terms of observing the treatment effect in biomarkers like tau, you cannot necessarily look at them across cohorts on a like-to-like basis. So I think tau is -- we're seeing great trends in tau, we're certainly seeing those more substantially in the moderate patients. But I think to compare them across dose cohorts because of the different starting baseline is a challenge. But where we do have, let's say, an equal starting point is with respect to E2 expression. So all the patients obviously start with 0 of APOE2 at baseline, and then we see these dose-dependent increases in APOE2 across the cohorts. And as we are observing, to some degree, a dose response, and the best example of that is the Cohort 3 patient who had the high E2 response and that corresponded to greater reductions in tau PET. We do believe there is a dose response associated with the therapy. So given that, the endpoints we would likely consider for a pivotal study would be something in the direction of tau PET, potentially combined with E2 expression, I would indicate that E2 expression levels and achieving the highest E2 expression levels would be the focus of a going-forward dose. So in that vein, I would suggest that probably the cohort 4 dose would be the right one to take forward. But I think this will be a function of conversations with the FDA about that strategy and those endpoints. It's early to comment on the FDA's position on the surrogate endpoints. Those discussions are ongoing. We will be able to give an update on that in 2025 in respect to those endpoints, but I would note, I think the precedent exists to support such a strategy. As you're aware, lecanemab was accelerated approved on amyloid PET. And there's increasing evidence that tau is more closely correlated to cognitive decline. So it would follow that tau PET would be a potentially compelling surrogate endpoint to indicate this link to cognitive decline. So more to come on that. We're in ongoing discussions with the FDA, and we will be able to provide an update on the regulatory path forward in addition to the path forward from a business development perspective as well in 2025.

Your next question comes from the line of Brian Skorney with Baird.

Congrats on the data. I guess my first question, was there a difference in terms of baseline tau levels across different groups, like did the moderate dementia patients have a substantially higher baseline tau measurement than the MCI or mild patients? I'm just wondering if you calculated [ NR2 ] or a correlation between the APOE2 expression and any of tau measurements?

Thanks, Brian. I mean, the moderate patients did have higher starting tau burden. So that is the attribute of the disease that as the disease progresses, you typically see greater accumulation of tau tangles. And that was certainly represented within the moderate population. In terms of a direct correlation, we did not conduct an analysis of that profile. Obviously, with the small sample set in a 15-patient, phase I study, drawing such statistical correlations are a challenge. But I think we do see very clear trends in tau within the population, and we see even more substantial trends in that moderate group. We did not conduct the statistical analysis against that from E2 relative to tau.

And your next question comes from the line of Mani Foroohar with Leerink.

I guess a quick one for Dr. Gandy. How would you interpret the numeracy of APOE2/4 ratio that we're seeing on Slide 11 of the slide deck that Nolan and team were kind enough to share? And separately, how should we think about the distribution across areas of the brain, in terms of thinking about frontal lobe versus other areas, [ corporeal ], et cetera, as seen on Slide 16. How should we interpret that? And at what point is there an adequate pool of data in terms of number of patients, duration of follow-up for that to be something meaningful?

Yes. Thanks for the question. And I think with the E2/E4 ratio, you would ideally like to get as close to 1:1 as possible. So since the safety signal shows that this is very safe, and you have a long way to go before you would get to that, I think that you have plenty of room to go in terms of increasing the dose to try and get towards that. And the idea with humans is that 2:4s -- 2 is so protective that 2:4s are roughly the same as 3:3 in terms of risk. That's the sort of theoretical goal. The importance of the tau PET is to we look at that [ 4 ] regions of the brain that you know are involved in cognition, and speech and to try and correlate symptoms with those changes. And you can do that much better with tau PET than with a biofluid biomarker, which averages all the brain regions into the -- into the CSF or the blood.

Great. I guess as a broader follow-up. Looking over to Slide 11. Obviously, we have a -- that was a different follow-up in the different cohorts. It is for the Lexeo team as well as for you, Dr. Gandy. How should we think about Cohort 4 going forward, given what we've seen from Cohorts 1 to 3 so far, do you expect improvement in 2:4 ratio between month 6 to month 12? Or is that too hard to call for at this point?

Yes. So maybe I'll provide some additional comments on the 2:4 ratio. So while I agree, genetically heterozygotes potentially have this 1:1 or even potentially lower than that ratio, we did not go into this study with the predefined level of APOE2 that we were seeking to achieve. What we wanted to look at was the correlation of levels of E2 that we did achieve to some of the downstream Alzheimer's biomarkers. And I think that, that's the most important aspect in regards to treating this disease. I mean, what matters here is what will impact cognition. And we think tau is the best indicator of a biomarker that will actually predict the impact on cognition. So irrespective of the level of E2 expression, the fact that we are seeing these improvements in tau biomarkers, from our point of view, is a positive signal. In respect to the difference between month 6 and month 12, I wouldn't expect to see any dramatic shifts in the E2 expression, I would expect it to look a lot like month 6 as we saw in the other cohorts. But I would note that we did see, for example, in the patients that had higher expressing E2, we saw greater impact on tau on average, including this Cohort 3 outlier. So I think that, that's -- the focus for us going forward is how does E2 impact tau, and then tau, obviously, as a potential surrogate endpoint for the future study.

Your next question comes from the line of Geulah Livshits with Chardan.