Annexon, Inc. (ANNX) Earnings Call Transcript & Summary

Hello, and welcome to Annexon's Conference Call. [Operator Instructions] Please be advised that this call is being recorded at Annexon's request. I would now like to turn the call over to Jen Lew, Chief Financial Officer. You may begin.

Thank you, and good afternoon. Yesterday, we issued a press release outlining results from our ARCHER Phase 2 clinical trial evaluating ANX007 in patients with geographic atrophy that were presented during an oral presentation at the ASRS 2023 Annual Meeting. The press release, podium presentations and slides used for this conference call are available under the Events and Presentations page within the Investors section of our website. Today's discussion will include forward-looking statements related to Annexon's current plans and expectations, which are subject to certain risks and uncertainties. Actual results may differ materially due to various important factors, including those described in the Risk Factors section of our most recent Form 10-Q and other SEC filings. These forward-looking statements represent our views as of this call and should not be relied upon as representing our views as of any subsequent date. We undertake no obligation to publicly update these statements. Now let me turn the call over to Annexon'S CEO, Doug Love.

Thank you, Jen, and thank you all for joining us today. On the heels of the top line ARCHER results that we shared in May, we presented the Phase 2 data for the first time at a medical meeting, ASRS, this past weekend in Seattle. The data presented included additional analyses supporting ANX007's protection against vision loss in patients with geographic atrophy. And we're excited to share the data more broadly today. Joining me on the call is our Chief Innovation Officer, Dr. Ted Yednock, who will walk us through anti-C1q's neuroprotective mechanism of action in GA. Importantly, and distinct from downstream complement approaches, our upstream approach is designed to preserve the function of photoreceptor synapses in cells to protect against vision loss. Our Vice President and Head of Ophthalmology, Dr. Donald Fong, is also joining us to provide an overview of ARCHER, including the study's endpoints, patient demographics and ANX007's well-designed drug candidate properties. And finally, we are delighted to be joined by Dr. Jeff Heier, the Director of Retinal Service and Retinal Research at Ophthalmic Consultants of Boston and an investigator in the ARCHER study. Dr. Heier will review ARCHER's efficacy and safety data, including the trial's protection against vision loss as demonstrated by several analyses. At the end, we will open the call for Q&A with Dr. Heier and our team. As many of you know, at Annexon, our bold mission is to free the body, the brain and the eye from complement-mediated diseases. Our aim is to enable every person impacted by complement-mediated diseases to live out their dreams and talent without being routed at their physical and cognitive health due to the disease. To do so, we're sharply focused on stopping classical complements' aggressive inflammatory cascade where it starts on disease tissue to provide the most complete protection against harmful classical pathway inflammation and resulting in functional loss. Our steadfast pursuit of our mission has led us here today to discuss the promising results from our GA program. As described on Slide 5, our broad portfolio was built on the understanding that not all complement inhibitors are the same, and that upstream C1q inhibition can drive powerful functional benefit. This differentiated approach has now been demonstrated in the clinic with robust functional improvement shown in Guillain-Barre Syndrome, Huntington's disease and now geographic atrophy and puts us on the path to potentially helping millions of patients through our programs. While the focus today is on our GA program, we have multiple additional value-creating catalysts expected across our broad complement portfolio throughout the remainder of 2023 and into 2024, which are shown on Slide 6. More to come on these readouts, but needless to say, we're quite encouraged by the data we've generated to date clinically and preclinically in support of each of these programs. Before I turn the call over to the team to review the ARCHER trial, I want to quickly highlight the key learnings from our in-depth analysis of the study. As outlined on Slide 7, there are 4 key takeaways from ARCHER. First, the findings support our founding neuroprotective hypothesis that blocking C1q to preserve photoreceptor synapses in cells to protect vision loss is quite distinct from inhibiting downstream complement such as C3 and C5. Second, ARCHER is the first study to demonstrate consistent visual function benefit in GA. With ANX007 treatment, there was statistically significant dose and time definitive protection against vision loss on 2 prespecified clinical measures, BCVA and LLVD. BCVA is a well-established endpoint to assess visual acuity and will be important for future development. These findings are strengthened by a host of related analyses, adding breadth and depth to the veracity of the vision protection data demonstrated in the general GA population. Third, while the primary endpoint of rate of lesion growth did not reach statistical significance, evaluating the first and second 6 months of treatment showed greater effect on lesion size with time. This suggests that lesion growth may slow with longer than 12 months treatment with ANX007. And finally, ANX007 was generally well tolerated as monthly and every other month treatment through 12 months with no increase in CMV rates between the treated arms and sham and no reported cases of retinal vasculitis. As you know, GA is a devastating disease of aging that greatly limits a person's ability to read, perform activities of daily living and even to see the faces of their loved ones. To be clear, protecting against vision loss for patients with GA is our seminal goal, and it is of the utmost importance for healthcare professionals and patients. Based on the strength of ARCHER's data, we believe that ANX007 has the potential to be a new and distinct treatment option for patients with GA, marking an important step toward achieving our mission of delivering game-changing therapies to patients living with classical complement-mediated diseases. We look forward to discussing next steps for ANX007 with regulators later this year, and we'll update after we've done so. With that, I'll turn the call over to Ted to walk us through our anti-C1q neuroprotective mechanism of action. Ted?

Thank you, Doug. Turning to Slide 9. Annexon was founded based on the work of Dr. Ben Barres, the former Chair of Neurobiology at Stanford. Dr. Barres made the discovery of the C1q, the initiating molecule of the classical complement cascade recognizes synapses in neurodegenerative disease triggering aberrant synapse removal. This pathway causes neuroinflammation and damage to become a key driver in the neurodegenerative process and was the founding technology of Annexon over a decade ago. Dr. Barres' work spawned an entire shield that has been validated in multiple laboratories in neurodegenerative models, including neurodegenerative diseases of the eye, brain as well as acute brain injury. A consistent finding in these studies is that synapse loss precedes loss of neurons and correlates with functional decline. So protecting synapses with anti-C1q is a novel neuroprotective approach that starts by protecting function. Slide 10 describes C1q in the role of the classical cascade in neurodegeneration. The left diagram shows how C1q triggers activation of downstream C4 and C2, which then activates C3 and C5 within the classical pathway. Inhibition of C1q at the top blocks all downstream components of the classical cascade, including C3 and C5 within the classical pathway. However, it leaves the lectin and alternative complement pathways in place, retaining normal homeostatic and immune functions of both C3 and C5 within those pathways. The diagram on the right shows how the classical cascade contributes to pathology in GA. The diagram represents a photoreceptor cell with a synapse on top that connects the photoreceptor cell to the rest of the retina. C1q localizes on the photoreceptor cell synapses in GA, where it anchors classical cascade activation on the synapse surface. Each of the activated component components plays a specific role in driving aberrant synapse removal by microglial cells, causing loss of visible function, neuroinflammation and photoreceptor cell loss. Inhibiting C1Q blocks all of these downstream components to protect synapses will provide a unique neuroprotective mechanism. This process has been validated in animal models, as shown on Slide 11, where anti-C1q is protected against photoreceptor cell damage. In the left panel, photoreceptor synapses are labeled in red and C1q with green. After exposure to damaging bright light, C1q binds to the synapses of photoreceptor cells within 3 days, triggering classical cascade activation and synapse removal. On the right of the slide, in a published study with light-induced damage, intravitreal administration and anti-C1q shown in the maroon bars, protected against photoreceptor cell loss and protected photoreceptor cell function. Slide 12 is a key slide because it shows how this process relates to geographic atrophy and importantly, illustrates how measurement of visual function is distinct from lesion growth. The image shows a cross-section of retinal tissue taken from a patient with geographic atrophy at the edge of a GA lesion. The green labeled cells on the bottom are retinal pigmented epithelial cells or RPE. The GA lesion is on the right where the RPE cells have already died and are missing in the orange box. Above the RPE are the photoreceptor cells in blue, and directly above those are photoreceptor synapses, which appear as red dots, highlighted by the green and white arrows. These dots are physical structures that connect the photoreceptors to the rest of the retina and so a critical to visual function. The photoreceptor synapses are very dense on the left side of the image away from the lesion and outside of the white box. But moving towards the lesion on the right within the white box, the synapses become sparse and are entirely absent in many places. Every loss of the synapse represents a loss of visual function since they are the only connection for the photoreceptors to the rest of the retina. Anti-C1q is designed to protect synapses and their function as demonstrated in multiple animal models of neurodegeneration. So anti-C1q would have the potential to protect visible function nearly across this entire image. In contrast, lesion growth is shown by the orange arrow on the right, illustrating the average growth of lesion in a year. This growth is measured by FAF, only tracks loss of the green RPE cells, such as those above the orange arrow. And this occurs in a region of the retina where there are very few functional synapses remaining. Therefore, it's not surprising that lesion growth, as measured by FAF, correlates poorly with visual function. We anticipate that by protecting photoreceptor cell integrity and visual function across a broad region of the retina, anti-C1q will slow lesion growth over time rather than just saving RPE cells at the very edge of the lesion. So to summarize, anti-C1q provides a unique approach for protecting vision by preserving functional synapses. This was the founding technology of Annexon and is distinct from downstream complement approaches. This mechanism has been validated in multiple animal models of neurodegeneration, including models of photoreceptor cell loss. In geographic atrophy, anti-C1q offers the potential to protect photoreceptor synapses, cell and visual function across a broad region of the retina for lasting benefit. So thank you for your attention. And with that, I will turn the mic over to Don to talk about the ARCHER trial.

Thank you, Ted. By way of background on Slide 14, GA is an advanced form of age-related macular degeneration, or AMD. It's chronic, progressive and neurodegenerative disease that can lead to blindness for about 1 million people in the U.S. and another 4 million worldwide. In the early stages of GA, patients have trouble reading or seeing in low light conditions. This is because the light-sensitive cells or the photoreceptors start to break down. As the photoreceptors start to die, patients will lose vision and the associated independence that comes with that. Slide 15 describes our drug ANX007. I want to spend some time highlighting the key attributes. ANX007 is a 50kD Fab antibody modeled after standard fabs like LUCENTIS. It is not pegulated. The formulation is low viscosity. ANX007 fully inhibits the classical pathway. Importantly, it doesn't affect the alternate and lectin pathway and leaves in place their normal immune and homeostatic functions. In summary, ANX007 is designed for intravitreal injection. Moving to the trial design on Slide 16. The ARCHER trial is pivotal design and is a randomized, multicenter, double-masked, sham-controlled study comparing the safety and efficacy of ANX007 to sham in patients with GA. The study enrolled 270 patients, stratified by GA lesion size, location and fellow eye CNV to minimize biases or imbalances in the data. Patients were randomized to receive 5 milligrams intravitreally either monthly or every other month or sham for a total of 12 months, followed by a 6-month off-treatment period. The primary end point of the study was the rate of GA area change measured by fundus autofluorescence from baseline through 12 months. Importantly, the ARCHER trial included multiple prespecified endpoints to assess visual function. These are best corrected visual acuity referred to as BCVA, low luminance visual acuity or LLVA and low luminance visual deficit or LLvD.These endpoints are important because prevention of vision loss is what we as ophthalmologists want and what our patients demand. Turning to Slide 17. Patient demographics and study eye characteristics were generally well balanced across groups. In total, 270 patients were enrolled across 3 arms, 89 in ANX007-monthly, 92 in the every other month group, and 89 in the sham group. The trial was predominantly conducted in the United States with approximately 95% of study patients in the U.S. The overall demographics of the population is consistent with other GA studies, including age, gender and ethnicity. The average age in the study was 80 years old. I mentioned this because at this age, patients do not have a lot of time to wait for vision preservation. Best corrected visual acuity was the same across all 3 arms. Patients were split almost equally between foveal and non-foveal. As background, foveal patients are those whose disease has involved the central macula, which is responsible for central vision. Non-foveal patients are those whose lesions are not yet affecting the central macula. Taking a look at patient discontinuations on Slide 18, between 11% and 15% of patients discontinued from ARCHER, which is consistent with previous GA studies. Slide 19. BCVA is a measure of visual function and has been a widely accepted functional endpoint for FDA-approved ophthalmology products. The assessment of BCVA 15 letter loss is considered clinically meaningful because it represents a 50% loss of a patient's visual acuity and is the gold standard for assessing efficacy in many retinal diseases. For veracity in our study, we defined vision loss as losing 15 or more letters on 2 consecutive visits or if the loss occurred at the end of study. This approach is designed to reduce variability in vision testing between visits, adding confidence to our data set. We are honored today to be joined by Dr. Jeff Heier, an investigator in our ARCHER trial. I'd like to turn the call over to Dr. Heier now to walk us through the ARCHER Phase 2 results.

Thank you, Don. It's my privilege to present the ARCHER results now. If we go to Slide 21, we see the ARCHER primary endpoint. And as Doug mentioned before, ANX007 did not meet its primary endpoint, a reduction in GA lesion area change from baseline to month 12 versus sham injection. It slowed lesion area change by roughly 6% from baseline to month 12 versus sham injection. However, if we look at Slide 22, showing the effect on lesion size over time, we see that looking at the first 6 months relative to the second 6 months, the separation between the sham and in treated groups increases after month 6. And this suggests that lesion growth may slow with longer treatment with ANX007. However, if we go to Slide 23, we see that ANX007 treatment did result in a dose-dependent protection against vision loss in this prespecified endpoint. Persistent vision loss was defined as 15 letter or greater loss at 2 consecutive monthly visits or at the 12-month visit. And again, we see a dose-dependent effect here. Slide 24 is a Kaplan-Meier analysis that shows that 15 letter loss occurred quicker and in much greater numbers in sham-treated patients versus monthly ANX007-treated patients. And ANX-007 treatment resulted in a 72% risk reduction with monthly therapy and a 48% risk reduction with every other month therapy. If we look at Slide 25, this is an even more stringent definition of persistent vision loss, 15 letter loss at 2 consecutive months and eliminating the patients with a 15 letter loss at only their last visit. And we see a similar significant dose-dependent reduction in risk of visual loss was maintained with the risk of 15 letter loss reduced by 66% relative to sham with monthly dosing. The completeness of the data, the consistency of the data is really highlighted on Slide 26. And we see that a reduction in risk of visual loss was not influenced by baseline demographics. When adjusted for a range of baseline characteristics, the risk reduction was maintained in both the monthly and every other month treatment groups with all point estimates to the left of 1 thus favoring treatment. Slide 27 shows a consistent protection for vision loss was further supported by elements in the fellow untreated eye. 15 letter best-corrected visual acuity loss at month 12 was evaluated in patients with GA in both their study eye and fellow eye. Among patients in the sham group, the rate of 15 letter loss was consistent in the sham and fellow eye. Among the treated patients with bilateral GA, again, we see a dose-dependent protective effect of ANX007 treatment in the study eye relative to the fellow eye that did not receive treatment. The risk of 15 letter loss was reduced by 74% and 47% with monthly and every other month treatment, respectively, again, highlighting the theme that we're seeing in all of these outcomes. If we look at Slide 28, a similar dose-dependent result with other cutoffs of best corrected visual losses show a dose-dependent trend was seen in 10 letter loss which is considered noisier 10 letter loss with 15.7% of patients treated monthly with ANX007, experienced a persistent 10 letter loss compared to 25.8% of patients in sham. A significant dose-dependent change in persistent 20 letter loss, again, less noisy with higher criteria here were demonstrated with 4.5% of every other month treated patients compared to 16.9% in the sham group. If we look at Slide 29, we see a trend in mean change in best corrected visual acuity at 12 months were also consistent with the dose-dependent effect of ANX007 treatment on vision protection in GA patients. At month 12, the sham group had lost a mean of 4.1 letters, which was consistent with previous GA trials. While the every month and every other month routes lost a mean of 1.8 and 2.7 letters, respectively. Looking at Slide 30, we see ANX007 treatment protected against vision loss even after adjusting for baseline lesion location. Among patients with foveal and non-foveal disease, both monthly and every other month treated groups showed a significant treatment effect compared to sham. Slide 31 looks at a classification of retinal health, which is classified by low luminance visual deficit. We chose low luminance visual deficit of less than 30 letters as healthier and greater than or equal to 30 as less healthy. ANX007 treatment compared to sham stratified by healthy and less healthy retina, reduced 15 letter best-corrected visual acuity letter loss in both the monthly and every other month group. If you look just at the healthier eyes, 0% of monthly treated eyes suffered 15 letter loss. In Slide 32, the protection from vision loss seen with ANX007 is further supported by looking at early data from the off-treatment period. This is from 12 months to 18 months where patients are off treatment. During treatment, there is a steady separation between the sham and treatment arms showing protection against vision loss. However, during the off-treatment period, the drug arms progress at a rate similar to sham. These results showing a change on and off treatment strongly support the beneficial impact of ANX007. We plan to present the full data set from the off-treatment period later this year. But given that these data represent approximately 50% to 85% of subjects who will ultimately complete the study, we do not expect to see that these results will change. Looking at Slide 33. Visual function protection was evident on other measures of visual function as well. There was a trend in reduction of 15 letter loss as measured by low luminance visual acuity, and there was significant protection from 15 letter loss in low luminance visual deficit with monthly ANX007 treatment. And I hate to keep harping on it, but what we're looking at is a remarkable consistency of the visual function data. Slide 34 addresses safety. ANX007 was generally well tolerated with monthly or every other month dosing. There is no apparent difference in the incidence of CNV across the 3 arms. 3 cases of endophthalmitis were observed. There was only 1 case of retinal vascular occlusion, which is listed as a branch retinal artery occlusion under serious adverse reactions. There were no cases of retinal vasculitis or any cases of ischemic optic neuropathy. In terms of serious adverse reactions, there was 1 case of iritis. There was 1 case of vitritis and 1 case of vitreous debris as well as the same case of retinal artery occlusion. So based on the totality of this data, the consistency of the visual function benefits. I continue to be excited by the opportunity of ANX007 to treat this devastating disease. And I look forward to seeing its continued development. I'll now turn the call back to Doug to close this out.

Thank you, Dr. Heier. Really nice job, also Ted and Don. Well, in summary, the novel neuroprotective mechanism of ANX007 translated into consistent and clinically meaningful protection of visual function in ARCHER trial, as you just heard. This effect was significant dose and time-dependent and robustly demonstrated across multiple measures and patient subtypes and was not dependent upon a significant change in lesion growth. Moreover, ANX007 was well tolerated throughout the study, which is important for this aging patient population. As a result, we believe ANX007 has the potential to fundamentally change the treatment paradigm in geographic atrophy, offering a chance to preserve vision in a broad population of patients. We're planning to engage quickly with regulators to discuss next steps and determine the optimal path to advance ANX007 toward patients. Our work in Annexon is only made possible by the passion, commitment and know-how of many. We sincerely thank our employees, collaborators and advisers who have greatly contributed to this milestone. We also thank the scores of physicians, medical teams and most importantly, the patients, families and caregivers who participated in our clinical trial. We are grateful and humbled by your support and contributions, and we are intent on bringing a new therapy to help solve this devastating disease. With that, we're happy to take your questions. Operator?

[Operator Instructions] Our first question comes from the line of Andrew Tsai with Jefferies.

I appreciate all these analyses. So clearly, based on this presentation, there is some consistent functional signals here to warrant further evaluation. So as we think about next steps, when can we expect you specifically to meet with the FDA? And when can we expect to hear the outcome around your Phase 3 plans? And secondly to that is what would be the ideal outcome in terms of the Phase 3 program strategy as well as the Phase 3 design itself?

We appreciate you joining today. Yes, we are actively working on a plan for discussions with the regulators, and we anticipate speaking with them by the end of the year, and we'll certainly update the market shortly after we have done so. With regard to the discussions there, we will be discussing the Phase 3 study design with an emphasis on BCVA 15 letter loss as the primary endpoint in that study. And so again, we will update after we've had concrete discussions with regulators on that this year.

Great. And then maybe just a quick follow-up for both you guys and the doctor. Another industry player has unfortunately disclosed some retinitis vasculitis. So maybe talk about your own views around that safety, AE, what's your hypothesis why that other drug is seeing vasculitis and why 007 is different and should not show any cases, especially since both of these drugs are both complement inhibitors.

Yes. Well, we only have the public information that you all have. And so we will not speculate on what the potential drivers are for the safety events you were referring to. What we will say is that we are encouraged by the profile of ANX007. This study in geographic atrophy is the second study that's being used with this therapy in an ophthalmic patient population. The first population was glaucoma on top of IOP-lowering drops, where it demonstrated a robust safety profile, and we're seeing very well tolerability here in the geographic atrophy study. We do think, as you've heard over the course of this discussion that targeting upstream C1q and the classical pathway specifically does have the potential to be differentiated, both on the efficacy side as well as on the safety side. On the safety side, by blocking only C1q and the classical pathway down through C3 and C5 and allowing the lectin and alternative pathways to remain intact for their immune function that has the potential over time chronically to perhaps be a well-tolerated therapy. And so that's our approach, and that's what we know today. I'll open it up to the team members to see if they'd like to comment further whether Dr. Heier or Don.

I think just to focus on ANX007, as Doug pointed out, there, we're very happy with the safety profile that we saw in the study. With respect to any other programs, I defer to the work that's going on with those programs and prefer to just focus on ANX007.

Our next question comes from the line of Pete Stavropoulos with Cantor Fitzgerald.

Congratulations on the ASRS presentation. First question I have is, when you look at the data for patients with foveal involvement, do you believe the difference observed for every month versus every other month was due to not having enough drug on board for the every other month group? Or is there some other explanation? And does the data for the foveal, non-foveal involvement suggest to you that there's an opportunity for a different treatment paradigm or dosing regimen for the 2 groups of patients?

Yes, a really good question. I'm going to turn that over to the experts. Don and Dr. Heier, please weigh in on that.

Jeff, do you want to start? Or do you want me to respond to...

I'll let you start, Don...

That's fine. So we did demonstrate an effect with foveal patients. This has not been seen previously prospectively. And what's really interesting is that we know now this year and an earlier analysis from lampalizumab that foveal patients often have eccentric fixation. And this eccentric fixation is very susceptible to synaptic loss and photoreceptor loss. So our mechanism is shown to be effective in the foveal patients as well as the non-foveal patients. Now in terms of whether lack of effect is due to a mechanistic difference. I think that you can sort of see that with our non-foveal patients, there are just not enough events. And I think that with longer studies and with our Phase 3 study, we will see an effect with our every other month treatment as well.

Yes. Those are excellent points. Don. One other thing to add is that we think of foveal patients as potentially not having viable tissue or good vision. And it's important to remember that there are many patients who have foveal disease who actually do have good vision and do have preserved foveal photoreceptors and synapses. And so the effects can be similar in some of these patients between foveal and non-foveal patients.

Just one follow-up question, if you don't mind. So the data for LLVA and LLVD, Dr. Heier, can you help us understand what those 2 measures mean in terms of predictive value or prognosis for clinical course and outcomes in general? And can you provide us with your interpretation of the data with respect to a reduction observed in the 007 arm versus sham?

Yes. So low luminance VA and low luminance visual deficits are other means of measuring different types of visual function. And the patients, both in terms of aging and with age-related macular degeneration have decreases in each of these. So this is another means of measuring visual function. And in many advanced GA patients, it's hard to pick up a difference here. But what we're seeing is the consistency here of being able to see a difference, both in terms of low luminance visual acuity and the low luminance visual deficit just highlights more that we're having a consistent effect on visual function. We also spoke to the low luminance visual deficit, the differences between regular visual acuity and low-luminance visual acuity often speaks to the health of the eye. And the closer those numbers are together, overall, the healthier the eye is, the healthier the tissue. And so when you're seeing these very different, you're looking at eyes that are considered healthy versus eyes that are considered not as healthy. Don, you may want to have some comment about this as well.

Yes, I'm happy to add to that. One of the ways to assess sort of the functioning of the retina is to look at how the photoreceptors and the synapse is working different luminous conditions. And in a low luminance condition where the light is lower, the photoreceptors have to work harder. And that difference tells you whether the photoreceptors are able to work hard or not. If there is a big difference between what an eye or a patient can read in normal light versus low light that tells you that the photoreceptors aren't working as well and that it's less healthy, okay? So that deficit gives us a way to assess photoreceptor and synaptic function by looking at low luminance VA and VD.

And congratulations once again.

Our next question comes from the line of Tazeen Ahmad with Bank of America.