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

Good day, and thank you for standing by, and welcome to the Annexon Interim HD Data Conference Call. [Operator Instructions] Please be advised that this call is being recorded. [Operator Instructions] I would now like to hand the conference over to your host today, Jennifer Lew, Chief Financial Officer. Please go ahead.

Thank you, operator, and good afternoon, everyone. Earlier today, we issued a press release outlining interim data from our ongoing Phase II trial of ANX005 in patients with Huntington’s disease. That press release as well as the slides that we'll refer to on today's call can be found on the Investors section of our website at annexonbio.com. On the call with me are Doug Love, CEO; Dr. Ted Yednock, Executive Vice President and Chief Innovation Officer; and Dr. Sanjay Keswani, Executive Vice President and Chief Medical Officer. Before we begin, please note that 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 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. I'll now turn the call over to Doug.

Thank you, Jen, and thank all of you for joining us. Happy new year 2022, it's going to be a great year. We're excited to share with you interim data from our ongoing Phase II trial in Huntington’s disease. As many of you know, at Annexon, our mission is all about patients and rest on the belief that we can create game-changing medicines for patients living with serious complement-mediated diseases. Now looking back on 2021, we made significant progress and gained tremendous insights across all of our programs, which has set a strong foundation for an explosive next 2 years. Indeed, as we head into 2022, we are well positioned to more than deliver on our aspirations across 3 therapeutic franchises: autoimmune, neurodegeneration and ophthalmology. To do so, we're leveraging 3 clinical candidates in 5 ongoing Phase II trials and anticipate 7 clinical readouts by 2023. And importantly, we're well capitalized into 2024 to achieve our objectives. Turning to Slide 5. The first of our upcoming catalyst is the interim data that we're reporting today for our Phase II trial Huntington's disease with our lead clinical candidate, ANX005. The data we'll walk through today are from an open-label, multicenter Phase II trial that enrolled a total of 28 patients with or at risk of manifest Huntington's disease. And the trial was designed to assess safety and tolerability, PK/PD as well as clinical measures with patients treated over a 24-week period followed by a 12-week off-treatment follow-up period. The interim data being presented today is on those patients treated with ANX005 over the 24-week period and includes an assessment of 28 patients for safety, 13 patients for target engagement and PK/PD, all 23 patients who completed the 6-month treatment period for the clinical measures and 16 patients who have thus far completed the 6-month treatment period for NfL. Turning to Slide 6, and as reported in our press release, the initial findings from our Phase II trial build on the scientific hypothesis of our scientific founder, the late Ben Barres, who believed that blocking C1q protects synaptic loss and can lead to rapid functional impact in a host of neurodegenerative diseases. Barres' discovery has been extensively researched across the globe by both academia and industry. And this Phase II trial in Huntington's disease is the first clinical test of the hypothesis. Cutting to the chase, we're quite encouraged by the initial data generated with ANX005 in this study, which has shown robust target engagement in both the serum and CSF, rapid clinical improvement in patients and a favorable safety profile. Interestingly, findings have also shown that patients with excess complement activity at baseline had statistically better clinical improvement compared to those patients with lower baseline complement activity. The differentiated responses in the 2 patient populations suggest the responses are real, that the classical complement pathway is playing a key role in the neurodegenerative disease process and that ANX005 has the potential to provide meaningful benefit to Huntington’s disease patients. As it relates to NfL, no meaningful changes in levels was observed in this early readout, which as we learn more about NfL, may suggest that longer than 6 months of treatment is necessary to see a decline in NFL levels in classic CNS diseases. This is particularly relevant here where our anti-C1q approach targets the loss of synapses, which occurs prior to the loss of neurons. So overall, as I've said, we find the data set quite promising and are continuing to learn from it and we look forward to providing full trial results in Q2 of this year. With that, I'll now turn the call over to Ted Yednock to provide an overview of our anti-C1q platform approach in neurodegeneration, that is preserving synapses to improve clinical function, and Dr. Sanjay Keswani thereafter to walk through the interim data analysis. Thank you.

Thank you, Doug, and hello, everyone. Huntington's disease is a devastating neurodegenerative disorder that impacts both motor and cognitive function. It affects a large part of the global population, and there's emerging evidence that indicates that the classical complement cascade drives synaptic loss and disability. On Slide 9 shows our mechanism of action for a C1q inhibition. Our scientific founder, Dr. Ben Barres, in his lab, discovered the C1q recognize synapses in development and then again in neurodegenerative disease. By binding to the synapses and neurodegeneration, C1q causes activation of the classical complement cascade, resulting in neuroinflammation and synapse removal causing further neural damage. In this way, C1q becomes an important driver of the neurodegenerative process independent of the initiating trigger. So by targeting C1q, we're blocking the complement cascade upstream of other complement components to fully shut down this neurodegenerative process and potentially allowing a rapid restoration of synaptic function in preventing neurodegeneration. On Slide 10, it has previously been shown the complement activity is elevated in Huntington's disease patients and the complement activity correlates with disease severity. The graph on the left shows that as disease progresses through premanifest and manifest stages, complement activity increases as measured by the activated complement component, C4a, in the CSF. Well, C4 is the first component of the classical complement cascade acted on by C1q. So its measurement, C4a measurement, in the CSF is a good proximal marker for C1q activity. Of note, patients in the current study largely had early manifest disease with a broad range of C4a levels. This is highlighted by the circled area. Showing on the table on the right, CSF C4a levels correlate with several measures of clinical severity, including both motor and cognitive function. These measures are part of the standard clinical outcome measured in HD. Slide 11 shows that synapse loss can also be measured in Huntington's patients. On the graph on the left, synapse number is measured with a PET ligand at living Huntington's patients. It shows that even in premanifest disease, which is the tan bar in the middle, there's been a significant loss of synapses. At this early stage of disease prior to the onset of overt symptoms, there's little loss of neurons, supporting a finding from other neurodegenerative disorders that synapse loss occurs prior to a loss of neurons. Now on the graph on the right, as disease progresses from premanifest to manifest Huntington's, synapse loss correlates with functional decline as measured by a standard Unified Huntington's Rating Scale or the UHDRS. So synapse loss is an important feature of disease progression. Shown on Slide 12, preclinical studies in animal models of Huntington's disease have linked complement activation synapse loss. On the left, systemic treatment of animals with ANX005 against C1q blocks complement deposition on synapses. And on the right, by blocking complement activation in these animals, we've protected against the loss of synapses over a 1-month treating period. So treated animals end up with more synapses in control. So in summary, on Slide 13, complement activity is increased in Huntington's patients along the synapse loss and is correlated with functional decline prior to the loss of neurons. Inhibition of C1q in preclinical models prevents synapse loss and protects neurons. So now I'm passing it on to Dr. Sanjay Keswani, our Chief Medical Officer.

Great. Thank you, Ted. So I'm happy to present, firstly, the demographics for our Phase II trial as well as safety and target engagement data. Moving to Slide 15. This slide illustrates those baseline characteristics. Of note, our patients were primarily early manifest HD patients and the demographics were consistent with prior natural history cohorts, including the TRACK-HD cohort, which for reference, is in the column on the right-hand portion of the slide. In particular, the cohorts were fairly well matched for key prognostic factors of progression, such as age, CAG repeats and baseline clinical scores. Moving to Slide 16. This shows our safety data of 005, which is our intravenous anti-C1q antibody. Chronic dosing of 005 was generally well tolerated. We essentially did an interim analysis on all patients who received at least 1 dose of 005, so that comprised 28 patients. The majority of the adverse events were related to the infusion reaction associated with the first dose on day 1. And these reactions were essentially limited to the skin, specifically transient skin rashes. Five patients discontinued 005 treatment, 2 of them for nondrug-related reasons and 3 of them due to 2 SAEs and 1 drug-related adverse event. With respect to the 2 SAEs, this included one individual had symptoms and signs of mucocutaneous lupus, which resolved after stopping drug, as well as the case of idiopathic pneumonitis that stabled post-dosing cessation with follow-up ongoing. There were no deaths and no serious infections in the study. Moving to Slide 17. This shows our target engagement data. Of note, we measured C1q target engagement in both blood and CSF. The CSF data is reflected on this slide. Of note, we saw full target engagement in both blood and CSF for the entire treatment period. And this was with the dosing regimen that was utilized in the Phase II study of 100 milligrams per kilo dosed every 2 weeks. On the left graph, one can see that we achieved robust drug levels in CSF in HD patients. And on the right, we see full target engagement, the blue line representing 3 C1q levels. Now of note, the first time point for the lumber puncture was at 6 weeks, but our expectation is that full target engagement was achieved at earlier than 6 weeks, closely resembling our blood target engagement data where we saw full target engagement within a few hours. Okay. So now going to move to our Phase II trial, clinical NfL data. Before I do that, and this is outlined on Slide 19, so I want to give a quick background on the clinical scores that are utilized to measure HD progression, specifically the composite UHDRS or cUHDRS which was a scale proposed by an industry academic consortium in 2017 on the basis of analysis of natural history data for more than 1,600 patients with early Huntington's disease. cUHDRS is a composite score that's composed of 4 constituent domains. The Total Motor Score, which is the top right-hand portion of the slide, which measures motor function; Total Functional Capacity, which measures activities of daily living; and 2 scales that measure cognition and specifically cognitive processing, the Symbol Digit Modality Test and Stroop Word Reading. Moving to Slide 20. Natural history data demonstrate consistent decline in the composite UHDRS. This is in data comprising of 1,600 patients with early HD as well as decline in constituent domains over time. Specifically, the composite declines by about 1 point per year in early HD patients. And all 4 constituent domains consistently decline over time with spontaneous improvement being unusual. Here on this slide we have data from the TRACK-HD natural history cohort, which demonstrates declines in each of these 4 constituent domains. For reference, the red dots represent HD patients and the blue green dots represent age-matched healthy controls. And as one can see in each of these 4 domains, there's progressive decline over time. I should note that for the Total Motor Score, an increase in score is worsening of function as opposed to a decrease in score representing worsening in the other 3 scales. Moving to Slide 21. This shows our data for the overall cohort for composite UHDRS. On the left, we show change from baseline for cUHDRS for all time points through 24 weeks. And of note, 24 weeks was the length of treatment duration in this study. As one can see, clinical function in this overall cohort was fairly well maintained as opposed to the blue shadow, which represents natural history data from the 2017 Schobel paper. In other words, clinical function seems to be maintained in the overall cohort as opposed to the expected decline predicted from natural history data. On the right, we're looking at the data in another analysis, which just relates to change from the UHDRS baseline at the last time point, specifically at week 24. Here, we see just over half of patients had improved from their baseline, specifically 56% of all patients. And again, spontaneous improvement will be considered unlikely on analyzing natural history data. So moving to Slide 22. We then perform sub-analysis of patients according to their baseline complement activity. And as Ted had referenced earlier in his presentation, our method of action C1q is the initiator of the classical complement pathway. And hence, we would expect that patients with high evidence of classical complement pathway activity to be most responsive to our anti-C1q drug. So I'll now hand it over to Ted.

Thank you, Sanjay. So in showing on Slide 23, as mentioned before, there is prior evidence of elevated complement activity in Huntington's patients. In the graph on the left, approximately half of these patients showed elevated levels of C4a, higher than control. And these results were true in 2 separate natural history cohorts of patients. In the graph on the right, in the current study, Huntington's patients showed a similar range of C4a in their CSF at baseline. So we reasoned, as Sanjay mentioned, that patients with higher levels of C4a may respond differently to complement [ inhibition ]. On Slide 24, we use baseline CSF levels of C4a normalized to baseline C4 as a way to divide the Huntington's cohort in 2 populations: those with high baseline levels of complement activity being above the median value and those below complement activity being below the median value. Importantly, C4a/C4 is an objective biomarker of complement activity. It was not examined until the end of the study, so patients, investigators and the sponsor were blinded to C4a/C4 levels throughout the study. Also importantly, the other baseline characteristics were well balanced between the groups, including age, CAG repeats and baseline disease score. The subsequent slides will compare the 2 subsets in functional measures following treatment. So back to you, Sanjay.

Great. Thank you, Ted. So now we're on Slide 25. And as Ted mentioned, we analyzed patients with a high C4a/C4 ratio, which is illustrated by the green line; as well as patients with a low C4a/C4 ratio, which is illustrated by the blue line on this graph. And of note, we saw a difference in response between these 2 lines or between these 2 patient cohorts, specifically those with a high ratio showed improvement from baseline. Whereas those with a low ratio had a decline that was roughly similar to what would be predicted by natural history, that's in that blue shadow. The difference between these 2 lines was statistically significant with a p-value of 0.045. As Ted mentioned, the 2 groups, the high and the low-ratio groups, were fairly well matched with respect to key prognostic factors for progression such as age, baseline NfL and CAG repeats. As Ted also pointed out, the C4a/C4 ratio is an objective biomarker that was blinded to patients, investigators and sponsors, so-called triple blinding. And hence, it would be unlikely that the effect we're seeing is due to a placebo effect as one would expect this effect to not disproportionately affect the high-ratio group. In addition, a prior hypothesis was that patients with excess complement activity would respond best to our drug. Moving to Slide 26. Confirming the veracity of this efficacy signal, we also observed improvements in 3 of the 4 constituent domains with the high-ratio group performing superiorly to the low-ratio group. This specifically relates to SDMT, a measure of cognition, which is the top left of this slide; the Total Functional Capacity on the right, which measures active daily living; as well as the Total Motor Score on the bottom right. And of note, for the Total Motor Score, we saw a p-value of 0.001 with respect to the difference between the low and the high-ratio groups. And again, with respect to TMS, a better score is a smaller score. So the arrow is downwards for TMS. We do not see separation of these 2 cohorts for SWR or Stroop Word Reading. The SWR is known to be a relatively insensitive scale that does not change much over a short period of time. Moving on to Slide 27. This is looking at the data with the second analysis that I talked about, which is the change of the UHDRS from baseline just at that last time point, specifically at week 24 at the end of treatment duration. And here, we see that, again, there is a superior response in the high-ratio group with 75% of patients improving from their baseline versus 36% improving in those individuals with a low C4a ratio. So essentially double the effect size in the high-ratio group. And the difference between these 2 groups was again statistically significant with a p value of 0.03. So I'm now going to move on to the analysis of our NfL data and this related to a cohort size of 16 patients, and we measured NfL in both blood as well as CSF. On Slide 29, we have our plasma NfL data. Here, we saw essentially no change of plasma NfL from baseline at the 6-month time point. This compared to a 3.6% increase in NfL as predicted by natural history. On Slide 30, we have the CSF NfL data. 005 was associated with a numeric increase of 9.8% at week 24 and this compared with a range of 7.3% to 19% with respect to NfL increase over this time period. So roughly in the same ballpark of NfL increases with respect to natural history over this time frame. Our assumption is that we would likely have needed to treat these patients for longer than 6 months to see an impact on NfL. And as Ted mentioned, synapse loss happens relatively quickly and precedes neuronal axonal loss, the latter being the main producer of NfL. So with that, I'm going to hand it to Doug to close the session.

Great. Thanks, Sanjay, Ted and Jen, really nice job. And thank all of you for your attention today. As you've heard, we're quite encouraged by the data from this interim readout from our ongoing HD Phase II study, which we'll summarize with a few observations. First, our scientific platform of preserving synapses to drive functional improvement has been extensively researched with consistent outcomes in animal models in an array of acute and chronic neurodegenerative indications. And this research underpins our approach in the clinic. Second, ANX005 appears to be a strong anti-C1q drug candidate in that it has been generally well tolerated and robustly inhibits C1q in both serum and CSF. Third, the interim HD data from this trial appears to support a growing body of evidence that blocking C1q can provide meaningful clinical benefit to patients suffering from devastating diseases. In this initial HD data set, blocking C1q protected clinical function in the overall patient cohort at 6 months and was consistent across both the composite measure as well as its subdomains. Blocking C1q also provided a statistically heightened improvement in patients with excess complement activity at baseline versus those with lower complement activity. As Ted and Sanjay pointed out, it's important to note that these 2 groups were well balanced at baseline and that the cohorts were triple blinded and that neither the patients, investigators nor Annexon were aware of baseline complement levels. It's also worth noting that we have extensively researched the role of C4a and excess complement activities for purposes of potentially ferreting out differentiated patient responses to our approach. Indeed, we most recently published on the role of excess complement activity in Huntington's disease last fall, and we're currently using this approach in 2 autoimmune indications, warm autoimmune hemolytic anemia as well as lupus nephritis. Fourth, the interim HD data set is the second demonstration of anti-C1q approach driving important clinical responses. Indeed, we previously saw a meaningful clinical benefit in our placebo-controlled GBS proof-of-concept trial. And thus, we're encouraged by the growing body of support for an anti-C1q approach. Fifth, as it relates to NfL, the levels remained generally unchanged over the 6-month treatment period. And our thinking is that a longer treatment period is likely required to see a decline given that synapse loss and improvement in synaptic function occurred prior to neuronal loss and NfL changes. Finally, with all of that, we certainly appreciate this is an open-label trial for 6 months and that a well-controlled trial will provide a definitive answer here. In other words, we appreciate that more work is to be done to fully answer the role of anti-C1q in Huntington’s disease. Nevertheless, we're excited about what we've seen thus far, and we look forward to continuing the journey of assessing ANX005 as a potential disease-modifying approach to treat patients with Huntington's disease. Turning to Slide 33. These HD data kick off what we believe will be a meaningful year ahead at Annexon. This HD trial is just one of multiple clinical studies underway for ANX005 based on its differentiated mechanism of action. And we are also advancing both ANX007 and ANX009 in clinical trials. We further have an active research organization to further enrich and expand our pipeline in autoimmune, neurodegenerative and ophthalmic diseases, again, tackling all 3 therapeutic areas in parallel, leveraging synergies and balancing risk and value creation. With the compelling early data in ANX005 in both Huntington's disease and GBS and 7 clinical readouts anticipated over the next 2 years as well as a strong balance sheet to fund our operations through key clinical milestones, we're well positioned to deliver on the promise of Ben Barres' early discoveries and in bringing important therapies to scores of patients in need. In close, I believe 2022 will be a transformational year. And we are very much looking forward to reporting more progress in the months to come. With that, we'll now open up the lines to take your questions. Thank you.

[Operator Instructions] And our first question comes from Phil Nadeau from Cowen and Company.

A few from us. I guess first on the safety profile, can you talk a bit more about the SLE and idiopathic pneumonitis that was observed. Were those thought to be a consequence of complement inhibition? I guess how were they assessed to be drug related?

Yes, Phil. Good question. Yes, as you know, C1q deficiency has been associated with lupus and so we anticipated that we could see this. However, we were encouraged by the fact that it was resolved once we knew it's through drug. I'm going to turn it over to Sanjay and Ted to maybe elaborate on your question.

Great. Thanks, Phil. So with respect to the case of lupus. This was the case of mucocutaneous lupus. Specifically, this patient had onset of mouth ulceration as well as skin lesions compatible with lupus. And of note, post-drug cessation, these symptoms actually resolved. And that's actually consistent with what we know about congenital C1q deficiency that I've alluded to with respect to those patients also having their symptoms and signs are reversed when C1q is actually replete in that particular case with fresh frozen plasma. So specifically, we were monitoring for lupus, and in this particular context, detected that case and that resolved with drug cessation. In terms of the second case of idiopathic pneumonitis, it was really called idiopathic because it was difficult to ascertain a causal relationship with respect to anti-C1q being mechanistically related to the pneumonitis case. And this was something that we had not expected. We're not sure, quite frankly, if we're going to see this in future studies. Temporarily, it was actually related to drug onset, to drug treatment. And hence, it was designated as related. That patient has stabilized with respect to their lung function. And we continue to follow that patient.

Perfect. That's very helpful. And then a couple of efficacy questions. In terms of the natural history data for the primary end point, can you talk about how -- what the standard deviation of the changes is on each domain over 6 months? So how much variability would you anticipate in a natural history population kind of over the duration of the study?

Yes. Good question, Phil. Sanjay, you want to take that?

Sure. So actually, this is denoted on this slide, which relates to the TRACK-HD cohort. I would like...

Slide 20.

Slide 20, yes. And so the 4 graphs actually relate to the progression of the 4 constituent domains of the composite UHDRS and associated with that are standard deviations. And so one can actually see it in each of them. So I hope that helps, Phil.

Yes. So I guess what I'm asking is it generally linear over the first 6 months? It seems like there's kind of a gap between, I guess, month 0 and 12. Should we just interpolate kind of what happens before that?

That's right. Yes. So we have data sets from 4 different natural history cohorts. In fact, it was meta analysis by Schobel, et. al. in 2017. And so we did interpolate to some extent with respect to 6-month data versus some of these longer periods of duration, which are typically 1 year-plus. There was also some natural history data, which was 6 months in duration. All of them relate to this roughly 1 point decline over a year or 0.5 point [indiscernible] over a 6-month time period and the declines that we see in all these constituent domains as well.

Perfect. Then last question from us is in terms of the low-complement versus high-complement patients, were there any other notable differences at baseline between those 2 groups? Was one more severe, one further along in their disease or anything notable in addition to the difference in complement levels?

Yes. I'll start, and then, Ted, you should follow up on this. But they were actually pretty well matched. And so we looked specifically at known prognostic factors for progression. They include age, a number of CAG repeats and then baseline clinical scores for composites as well as the 4 constituent domains as well as CSF and plasma NfL levels as well. There were some minor differences, but overall are pretty well matched. So I don't know, Ted, if you want to extrapolate on that.

No. I think that you've covered it well, Sanjay.

And our next question comes from Anupam Rama from JPMorgan.

What preclinically or otherwise gives you confidence in NfL kind of declines with longer-term treatment? Or are we really thinking about stability on NfL now as a win in the context of how natural history progresses?

Yes. Maybe I'll start and then turn it over to Ted. So Phil (sic) [ Anupam ], we were able to show reductions in NfL, both in our Huntington's disease models, which is a very aggressive fast-acting model as well as in a SOD1 ALS model. So preclinically, we have some reason to believe that over time, you can indeed reduce NfL but it does appear to play out longer than 6 months period of time. And Ted, I'll turn it over to you to elaborating on all of that.

Yes. I was to add that with the Huntington's mouse model, the, R6/2 model, very, very aggressive. And so we began treating early in the disease process and then treated throughout over basically 2 months, which takes you to a rather advanced stage of disease, again, in this very rapid mouse model. So in covering that length of disease progression in the mouse, we were able to see a significant reduction in NfL. So you could assume them with patients who are well into their disease process in a much more slowly progressing disease that it would take longer to see that kind of impact.

And our next question comes from Joseph Stringer from Needham & Company.

Two from us. Just on the 16 patients with the NfL data comparing to the 23 patients on the UHDRS clinical function score, I suppose is there any reason to believe that the 7 patients there could -- we saw the NfL reductions for those. Is there any reason to believe that it would be different than sort of the 16 that you're showing? Or I guess of the 16 you're showing, is there any -- are those stratified in any way in terms of high versus low complement activity? And then we have a follow-up with that.

It's a good question. Maybe I'll just kind of summarize and make sure and kick it over to Sanjay and Ted. So your question being with the 16 patients, can we see a difference between high and low complement levels? And is there any reason to believe that the additional patients coming into the study that you would see an overall change in the scores here? So gentlemen, I'll turn it over to you two to respond. Thanks, Joey.

Yes. I don't think that there would be -- just looking at the variance in the population, there's quite a bit of overlap between all the different patients. And with respect to low versus high complement, we did not see anything significant in that regard either. So I don't see these 7 patients changing the ultimate outcome.

Got it. And then just a follow-up. I know this is the initial data, but maybe just your early thoughts on the next steps and how to proceed here with additional studies. Would you be taking a look at or focusing more on this high versus low complement -- patients with high versus low complement activity going forward.

Thanks, Joey. Well, it's certainly intriguing what we've seen with the high and low patient population. That being said, we saw improvement in both populations and so we'll continue to assess that. We're not going to make any decisions on the Phase III program in any form until we see the final data set in Q2, which will include the 3-month off-treatment period. I will note, however, the difference between what we see clinically and on the NfL measures is that we do have the full data set of all 23 patients that completed the clinical study, the 6 months on treatment side for the clinical piece. Yes. I don't know if, Sanjay, you or Ted would elaborate on that at all?

Yes. At this moment, we're kind of digesting this with the benefit of our key opinion leader relationships. But I think we have some interesting options going forward. I think you're alluding to a precision medicine approach. And clearly, that's something we're utilizing across the border in Annexon. By focusing on the right patient subset, potentially we should see higher probability of success. And so as Doug mentioned, we do have the ability to enrich potentially for responsiveness here, utilizing this high C4a versus low ratio. I think the other thing to note is that the clinical function improved quite quickly. It also gives options down the road in terms of future clinical studies, which respect to adaptive designs. So one could treat for a period of time and then make decisions based on that initial response. So hopefully that's helpful. Clearly, food for thought for us at the moment.

And our next question comes from Tazeen Ahmad from Bank of America.

Follow-up on NfL. So Doug, should we expect to see any impact on NfL when you release the full data set in 2Q? Just based on what you said in your prepared remarks, it seems like it does take some time, but how much time do you think would be needed? And if you don't see an impact on NfL, how would that potentially impact your decision to move forward to a pivotal study?

Good question. I think it's important to bear in mind that NfL is a biomarker to assess neurodegeneration. Ultimately, our objective is to slow the rate of progression clinically in the neurodegeneration, or in this case, show some improvement for patients. So that is ultimately the final objective for us with regard -- and all companies with regard to the neurodegenetive programs. So that's going to weigh quite importantly in our decision as we continue to move this program forward. Secondly, whether or not we will see a change on NfL when we have the final additional patients, just really difficult to say. You could see that there is some variance at each of the time points. And so we'll wait to get the final data before we opine on anything more with regard to that. I think it's important also to note with regard to NfL is that it appears to be relatively stable or consistent. And so patients are not getting worse. And I think that's a really important concept for people to understand as we assess NfL as a decision-enabling tool. And then coupling that with the clinical improvement, we think, overall, we're encouraged by how all this has come together.

So as it relates to change in UHDRS, what's the FDA's position on what you would need to be able to show [ import ] has to be viewed as a marker of efficacy in HD. Is there a particular percent improvement that the agency has indicated that you'd need to show?

I'm going to turn that over to Sanjay to tackle.

Yes. So the dogma with respect to neurodegenerative disease, including HD is prevention of further decline. And so obviously, we're quite interested in our results, which show an improvement above baseline. With respect to the first paradigm, i.e., in [ relation ] to decline, so clearly, other companies specifically Roche had discussions in terms of clinical endpoints. We've guided to some extent by the choice of the endpoints and how they power their studies or the Phase III program, specifically for tominersen. And they're essentially looking at just over 1 point in terms of complete UHDRS change over time being clinically meaningful, I believe it was 1.2 points. Clearly, we'll have our own discussions with health authorities in terms of what would be regarded as meaningful changes in these endpoints.

Okay. And then on safety, with the instances of SLE and IP that were observed, is there any reason to think that the other studies that you have underway for the other indications for 005 might see the same side effects? Or is this something that you think would be limited to HD patients?

Yes, back to you on that, Sanjay.

That's a really good question. So specifically, we saw this one case of mucocutaneous lupus in -- or with our drug candidate 005, which is our intravenous drug candidate. And we have a fairly wide range of drug candidates, as you know, that we're utilizing with or other indications, specifically our subcu candidate, which is 009. And also, we have an oral molecule coming up as well. So our sense is that the amount of target engagement we see just in the blood compartment may have a differential safety profile specifically with respect to the reduction of lupus versus 005, which achieves rather complete C1q blockade throughout the body. Again, that's the hypothesis that we have at the moment, but that's our sense going forward with our other drug candidates in these other indications.

But with respect to different indications, we don't see anything particularly special about Huntington's disease for the development of lupus. And so that could be mechanistic related and we'll have to see going forward. Importantly, it was something that can be monitored and appears to be reversible as it is in these [indiscernible] patients. So yes, we don't see anything particularly special about Huntington's disease in that regard.

And our next question comes from Pete Stavropoulos from Cantor Fitzgerald.

Congratulations on the data.

Good to hear from you.

So on Slide 21, 56% of the patients improved at week 24 on clinical function. Was there anything about these patients' baseline characteristics for those who improved versus those that did not, such as baseline UHDRS independent score or CAP score?

Actually they're pretty well -- sorry.

Go ahead. Sorry. No, go ahead.

They're pretty well matched between the 2 groups, specifically high and low. The only real difference was actually the ratio, the C4a/C4 ratio. So our hypothesis is that, that was more akin to our mechanism of actions anti-C1q.

Okay. And in terms of the AEs, I know you discussed them, but can you give a sense of timing for when the subjects discontinued dosing, after how many doses? And was there anything about the patient who experienced the SLE, was there anything about the patient's medical issue or background that made them susceptible?

Yes. A really good question. So in terms of discontinuation, it was roughly 3 months of dosing is that discontinuation time for that specific patient. We've been thinking carefully in evaluating any specific factors in that specific patient that increased risk with respect to development of lupus, specifically scrutinizing the titer of antinuclear antibody at baseline. And that's something that we're looking at with respect to further screening or monitoring in our other studies. So specifically utilizing a slightly lower threshold with respect to ANA titers that would be acceptable for entry into the study. But apart from that, there was nothing remarkable about this patient.

I am showing no further questions. I would now like to turn the call back over to Doug Love for closing remarks.

Thank you, operator, and thank all of you. We appreciate your questions and tuning in with us today. As you hear, we are encouraged by the initial data presented today and we look forward to sharing the full data from the study in the second quarter of this year. I want to wish you all a Happy New Year, and have a great afternoon. Thanks.

This concludes today's conference call. Thank you for participating. You may now disconnect.