INmune Bio Inc. (INMB) Earnings Call Transcript & Summary

Okay. Welcome to the INmune Bio key opinion leader webinar on our Phase Ib data on targeting neuroinflammation with XPro1595. I'm David Moss. CFO of INmune Bio, and we're excited to have a distinguished group of panelists today to share with you expanded data on our Phase Ib trial, targeting neuroinflammation with XPro1595 in patients with Alzheimer's disease. This event is being recorded and will be made available on the company's YouTube channel, which can be found on the company's website a few hours after the close of this discussion. We expect today's webinar will last about an hour, with 45 minutes of time allocated the panel discussion and the rest open for Q&A. And then finally, concluding remarks from RJ Tesi. At this time, all attendees have been placed on mute in a listen-only mode. If you would like to ask a question, please use the Q&A button in the Zoom window and type your question in the dialogue box, and we will read it. Alternatively, you can e-mail your question to me at [email protected], and I'll sort them and read them to the panel during the time allotted for Q&A. Before we begin, I remind everyone that except for historical -- except for the statements of historical facts, the statements made by management and responses to questions on this conference call are forward-looking statements under the safe harbor provisions of the Private Securities Litigation Reform Act of 1995. These statements involve risks and uncertainties that can cause actual results to differ materially from those such as forward-looking statements. Please see the forward-looking statements disclaimer on the company's earnings press release as well as the risk factors in the company's SEC filings included in our most recent quarterly filings with the SEC. There's no assurance of any specific outcome. Undue reliance should not be placed on forward-looking statements, which speak only as of the date they are made as the facts and circumstances underlying these forward-looking statements may change. Except as required by law, INmune Bio disclaims any obligation to update these forward-looking statements to reflect future information, events or circumstances. So with the forward-looking statements behind us, we're very excited to share with you data from our ongoing Phase Ib trial. Without further ado, I'd like to pass the webinar over to RJ Tesi, CEO. RJ?

Thank you, David. 6 months ago, we presented early data on INmune Bio's Phase I study using XPro1595 to treat neuroinflammation in patients with Alzheimer's disease. Using white matter-free water, a validated biomarker of neuroinflammation, we showed that XPro1595 decreased neuroinflammation in the CNS of patients with Alzheimer's disease. Those data gave INmune Bio confidence that XPro1595 is a drug that targets neuroinflammation and which is an important contributor to the cognitive decline in patients with dementia. At this time, we remain committed to initiating the Phase I -- Phase II trial this year. If anything, we are more committed to that plan. We understand white matter-free water as a biomarker was new to many. Today, we will connect the dots between white matter-free water, a noninvasive biomarker of neuroinflammation with more traditional biomarkers of neuroinflammation seen in the CSF or found in the CSF. We hope to convince you that decreasing neuroinflammation has positive downstream effects on neuronal and synaptic health and function. In our quest to be the leading-edge of biomarker development and CNS drugs development, we will introduce cortical disarray measurement and apparent fiber density, gray and white matter analytics, respectively, that bring greater understanding to the therapeutic potential of XPro in Alzheimer's disease. Many of you know, CJ Barnum, INmune Bio's Head of Neurosciences. CJ, floor is yours.

Thank you, RJ, and thank you, everybody, for joining us. Before I start, I just want to make a brief statement. None of this will be possible without funding from The Alzheimer's Association's Part the Cloud. So we would be remised in not reminding ourselves that this is really funded from a fantastic institution, and we thank them once again. So I want to put this -- let's start from the beginning and make sure we understand what this trial is really all about. This is a trial about showing that we can reduce biomarkers of neuroinflammation in patients with Alzheimer's disease that have the pathology we're looking at. So I think what's unique about this study is not only that these patients have a diagnosis of AD, but that they have a biomarker of inflammation. And what I put down here are the biomarkers that we've used to enroll patients into this study. And now these biomarkers were chosen based on their translational utility that they could be done at virtually any lab around the world and they're linked to neuroinflammation. I will make one brief comment about this. Most patients had at least 2 of these biomarkers to get into the study. And I think only 1 patient only had 1 biomarker. So the fact that we see correspondence among multiple biomarkers, I think, is really telling as to how these patients really fit as it relates to neuroinflammation. Again, the goal of the study is to demonstrate that peripheral administration of XPro can reduce biomarkers of neuroinflammation. As RJ mentioned, we presented some imaging data in July. And today, we're going to follow that up with some of the cerebral spinal fluid data that we have. These patients were treated for 3 months. And the biomarkers we're going to be looking at today, at least as it relates to CSF, we're looking at a baseline and a 12-week change. We have some additional time points we get to MRI, and I will talk about that momentarily. So this is a basic demographic slide of the patients that have completed 12 weeks. And this is the patients that we'll be talking about today. We do have additional patients that are in the study, but they are not a part of the analysis at this time. So one of the things I want to point out here is this baseline MMSE. And the MMSE reflects the status, the current status of the patient. And what's unique about this trial is most Alzheimer's studies will have a very homogenous group. And what I mean by that, there will be all mild patients or moderate patients. And you can see here that we have patients that span all different severities. And that's because, again, the goal is to show that we reduce inflammation, and we are recruiting patients that had inflammation. It will become apparent why this is important a little bit later on in the talk. Okay. So I want to give a status update on the patients that we will be talking about today. So we have 9 patients that have completed the 12-week study. All 6 patients that were treated with a high dose, which we will focus on today, have elected to enroll in the extension study. Three of those patients have completed the 9-month extension study and have been on drug for 12 months. Importantly, I'd like to point out that all 3 patients that have completed 12 months, have requested to stay on the drug due to an expanded access program, and we are in the process of making sure we make that happen. So again, the goal of the study was to show that we could reduce biomarkers of neuroinflammation. And our approach is pretty simple. If we want to have confidence that the drug is doing what we think it does, we want to make sure that we can show it in multiple modalities and multiple assays. And so what I'm going to show you today is some data from -- updated data from Imeka as it relates to free water. And some data as it relates to cerebral spinal fluid. And these will be in assays from Olink in proteome sciences. And the extent to which we can show neuroinflammation is reduced in all these different approaches. It will give us the greatest confidence that we have a drug that's doing what it's intended to do. The key element here is showing the totality of it. We want to see the pattern because the ends are very small, the pattern of change is extremely important. So what I want to point out here is we're talking about patients in the high dose. The patients that received the low dose of XPro had much more variability as it relates to neuroinflammation. So we think the most appropriate dose is the 1 mg per kg. And I'm going to talk about that in the next 2 slides, but we'll go back and forth. And I'll point out when we talk -- when we include both patient groups as well. So as we mentioned before, with white matter-free water, we get a reduction in MRI. In this white matter-free water, safe mask is about a 4%, 5 -- almost a 5% reduction. When we start to look at markers of inflammation within the CNS, what we can see is that these are changing in the right direction. So let me just give you a little bit of background on what this assay does. So this is an assay from Olink, which has a target at 48 cytokine platform. And the data that you're seeing here is every single data point that is within the limit of detection. And there are some of them that are not present, but the reason they're not present is because the assay couldn't detect them. And what you can see here is that in every single analyte, we get a reduction with the exception of interfering gamma in which there's no change. Now again, the numbers are very small. There's a little bit of variability, and we're less interested on the significance of it and more interested on the pattern. The other thing I will point out as it relates to TNF, CCL7. This change in CCL7 is important, at least from my perspective, because CCL7, I like to think of as TNF's shadow. It is highly reactive to changes in TNF. It goes up very quickly and robustly when TNF has increased, and it reduces very quickly and robustly as it relate to TNF when it is attenuated. So what we're seeing here in each of these values is a percent reduction from baseline. And this mean score here is an average of all of these different inflammatory factors. So again, the pattern that all these factors are changing in the right direction is extremely important from our perspective. So what does this mean? I want to take this a little bit further and just give sort of -- anchor these CSF findings to what we found with MRI inflammation. And what we're looking here is a correlation between the 2. And this CSF inflammation is a composite of all those different factors that I showed you on the previous slide. And what we can see here is that when all the patients, regardless of dose, are correlated, we see a very large and nice correlation between the 2, suggesting that MRI may actually be a very good sensitive, noninvasive metric of inflammatory burden in the brain. So the next step for us is really to look at what are the consequences of reducing neuroinflammation. And again, our approach is similar. We want to make sure that we can assess this through multiple different assays or methods. Then what I'm going to talk about first is some of the proteomics, and then I will anchor that with some of the newer metrics that RJ mentioned previously as it relates to looking at gray matter and white matter. So we worked with a company called Proteome Sciences that has this technique called TMTcalibrator. And basically, what they do is they take AD brain tissue, and they analyze it in parallel with the CSF from the XPro-treated patients. And using that high level of AD protein relative to the CSF actually triggers the mass spec, the machine that detects the proteins, on those AD-related peptides to determine whether it's present. The unique aspect of this is it increases your sensitivity, so your ability to detect the protein and your specificity so that it's really looking at proteins that are present in the brain. And using an unbiased approach, the conclusion of this study was that markers of microglial activation, synaptic anecdotal dysfunction were significantly regulated in CSF from the AD patients that were treated with XPro. So they change over time. The functional analysis, a more specific analysis showed that the most significant changes were found in CNS neuronal function, immune-inflammatory response, cytoskeletal, metabolic processes and dendritic morphogenesis and synaptic plasticity. And they did this by looking at 35,000 distinct peptides and about 5,000 different protein groups. To give you an example of what this -- of what we're talking about, I have asked them to look at a prespecified analysis of a few targets that are relevant to AD neuroinflammatory, neurodegeneration and synaptic proteins. Of the targets that I asked them to look at, 26 of them were detectable, and 20 of the 26 were statistically significant. And what I'm showing here are 3 different examples in different categories of where this changed. So as it relates to immune function, what they found with YKL-40, which is a marker of neuroinflammation in Alzheimer's disease, is an almost fourfold change following treatment with XPro. If we look at C-reactive protein, which is one of the biomarkers that we like to use in selecting patients and treatment response, we see about a twofold decrease over the course of treatment in 12 weeks. If we look at synaptic proteins, I have 2 examples here. One is contactin-2, which is about a 2.5 fold change. And contactin-2 is important for neurites. So these neurite outgrowth and extension. So when the neurons are starting to branch out and connect, contactin-2 is really important for that. And I also have an example from the dendrites of neurogranin, which is about a half fold log change -- a half fold change from baseline. And this speaks to the dendrite. So we're seeing changes in not only synaptic proteins from these axons here but also with the level of dendrite suggesting that's having an overall -- potentially an overall effect. And then finally, markers of neurodegeneration, we see about a log, a twofold change in both VILIP-1 and neurofilament, 2 markers of neurodegeneration which are extremely important and relevant for the Alzheimer's disease. The important piece to -- the proteomics from our perspective is that this reflects exactly what we see in the animal models. We have 75 -- nearly 70 publications with XPro. And this is exactly what you would expect. So the fact that we are seeing these same changes in a very small number of patients using an unbiased approach is quite remarkable from our perspective in this short-term study. So as RJ mentioned, CSF is wonderful, and the proteomics is fantastic, and we have a lot to learn from that. But it turns out that we might be able to look at how some of these consequences of neuroinflammation are changing with novel MRI technology. And so what I'm going to talk about here is 2 ways to assess brain quality through what we like to call the virtual biopsy using a measurement called impaired fiber density in white matter and cortical disarray measurement in gray matter. So let me start here on the left with white matter. What you can see with white matter in this healthy column here is that the white matter has all these wonderful axons with these little coverings called myland, which helps insulate and propagate signals. They also have immune cells that are in other cells that provide support. And in a healthy system, you've got wonderfully connected axons, great myland, good quality and supportive immune cells. In a disease patient, these things can change. It could be fewer axons. It could be fewer myland. You can have different types of cells. And of course, there's pathways going back and forth that are taken into account. As it relates to gray matter, not only looking at volume, which is a traditional metric of assessing brain, the amount of brain tissue that you have, you can also look at the quality of gray matter. And you can see here on the far right in this healthy column, what you would expect is to have all these neurons that are nicely organized together that are connected together. And in a diseased patient, you have these disconnections, and sometimes, it's loss of cells. Sometimes, it's just loss of neurons. And so this cortical disarray measurement that Oxford Brain Diagnostics in the apparent fiber density that Imeka has come up with gives us the ability to do this. So why is this important? Well, traditional volumetric measures of brain tissue are quite insensitive to change over time. In fact, it can take years to see meaningful changes. And so the extent to which we can see changes in the quality of the brain, despite not seeing changes in volume, I think, is potentially very important and relates back to some of the proteomic data. So as an attempt to explore the potential of this technology, what we have done is we have taken the patient that we thought had the best clinical response to see if these specific metrics held any promise. And so what I'm going to start here is with the gray matter that -- from Oxford Diagnostic. Again, this is a -- the metric is called cortical disarray measurement. And what you're seeing is the bit from the baseline brain, the colors are the changes over the course of time. The green color reflects improvement, and the yellow represents no change. And what you can see is that within 3 months, this patient shows improvement in multiple areas of the brain that continues over the course of 9 months. I will make -- what I think is really important about this is that if you look at this 9-month scan, the regions that are implicated or that are changing include the temporal lobe, which is this circle part here. And the temporal lobe is the lobe critical for dementia in Alzheimer's disease. And this change over 9 months was about 6%. And the suggested clinical meaningful change is believed to be about 5%. So this suggests that this metric may have utility in understanding quality of gray matter and the biomarker may be useful in a very short period of time. So if we go back to white matter, looking at how those neuron, those gray matters connect to each other, we see a very similar pattern. So again, if we start in the top left, we look at the baseline, you can see this as the fiber tracks, the white matter tracks all in purple. And by the time you get to 3 months, you have improvement. And this is all color that's not purple, is an improvement in the quality of white matter tracts that peaks at 6 months. Now this is a little bit different but what they've done is they've looked at the number of fiber tracts. And there's 33 fiber tracts that Imeka has identified as important to Alzheimer's disease. And this increase is about a 13% on average for 31 of the 33 fiber pathways, which is extremely exciting to us. So we now think we have 2 noninvasive visual biomarkers that can be used in the same scan. By the way, this can be done with free water. The free water, the white matter, and the cortical disarray measurement gray matter, can all be done in a single scan with patients that will potentially allow us to see changes that we believe were reflected in the CSF in a very short period of time. So I want to take this back to trying to putting this in context of how inflammation may be affecting clinical change. And this slide, I just want to put the caveat out there, disclaimer, that this is not a slide about cognition, and I'll get to that in just a moment. It's -- there's -- it's very difficult to assess -- virtually impossible to assess cognition in a small study. Why? This is a small end. There is a stream disease heterogeneity. Some of these patients can't even complete some of these scales. The short time period is very difficult. And again, this is uncontrolled. But to try to understand the relationships, the inflammation and cognition, I have asked Dr. Judith Jaeger, who has eons of experience and is the top of her field at understanding these psychometric tests that we administered and how they play in different disease states to come up with a metric to compare across these patients. And so after some conversation, she decided on a qualitative score that ranges from a minus 2 to a 2 based on her assessment of the overall change in these patients over 3 months and given all those caveats of where they are in their disease status, the clinician report and those -- and all those factors. And so as you can see here in the 9 patients, and this, by the way, includes patients in the low-dose group as well. We have 1 patient that she classified as minor progression. The majority of patients, 5 would -- she would classify as stable disease, which is probably what you would expect at this course of time. One, she would put in a minor improvement category; and two, she would put in the meaningful improvement category. And again, this is not a slide about cognition. It's about how cognition relates to inflammation. And the reason we did this is because it turns out that these 2 patients with the greatest clinical improvement over the course of 3 months had the largest reduction in neuroinflammation. And to give you an example of that, I've pulled out 4 of -- 4 inflammatory factors that were measured in the CSF, as an example, and what you can see on the bottom here is the clinical change. So 2, over here means the patients that had the best clinical response. And on this y-axis here, this is a percent change over time. So what we're looking for are patients that had -- are in this bottom quadrant that had the highest clinical response with the greatest percent of neuroinflammation. You can see that irrespective of the measure we looked at, we see that. So this is giving us great confidence as an early marker that we're looking at, and we're predicting that inflammation may have -- may be associated with clinical response may actually be the case. Now these are really small numbers of patients, but at least the direction is trending in the right way. So just to summarize what we've discussed. We've shown data that XPro reduces neuroinflammation across multiple measures and assays. The pathway proteomic analysis, unbiased, showed a significant effect of XPro in AD-relevant pathways, and we may actually be able to support this by some of these novel MRI quality metrics. And if you drill down in the pathway analysis and the proteomics, you can see that XPro significantly reduced these biomarkers of neuroinflammation, neuronal injury and synaptic proteins. I think a key point here is that these changes in biomarkers all appear within 3 months and were sustained over the duration of the trial. The drug is generally -- it's well tolerated. So with the exception of a few injection site reactions, which are common with this type of administration. And overall, we feel that these data support advancing to a Phase II. So with that, I will turn it over to RJ, and I thank you for your attention.

Thank you, CJ. I think before we go to Q&A, I'd like to engage the panel in some discussion here. And I'll start out with Dr. Tansey, Malu. You have decades of experience with XPro and with neuroinflammation. And I presume the cytokine results don't really surprise you. What do you think about the proteomic results and the structural changes in the white and gray matter?

Yes. I think I'm not surprised that given the preclinical data we've seen, the inflammatory markers are moving in the right direction. What I am surprised about is that so many of them are moving in the right direction. Certainly, in the high-dose, the response is pretty remarkable with targeting soluble TNF. One thing to bring up is that there's not a lot to compare it to because this is one of the few examples where you've really used this biomarker-directed trial, right? And you're enrolling the subjects based on inflammation and you're testing a hypothesis based on inflammation. And I think that's unique. You're not taking all comers. And so that's -- your population is homogeneous within the context of inflammation. And so that, I think, is one thing that is important to realize. And so that is I wish the mice had responded that way in a way we could have measured them in real-time. So I'm not surprised from that respect. What is really exciting to me is that you're able to connect the dots between the inflammation and the proteomics that points to synaptic protein and integrity with the MRI measures that appears to suggest in early results that there's some protection of integrity and that there's less degradation with the protein, the hits that you're getting. So that's exciting. And again, what we don't -- what we didn't have in the preclinical models is the ability to sample while the animals were still live and seeing how that was getting protected. We just had endpoint measures. And to be honest, our models were not that -- are never as good as the patient. So you're able to really connect the dots between the inflammation and what the ultimate -- where the rubber meets the road in terms of being able to protect the synaptic function that we want to protect.

Yes. If I could follow-up, aducanumab and the new Lilly results have -- with anti-amyloid drugs have really produced some excitement with their positive clinical data. And by definition, all of the patients in this trial had amyloid, and most likely had tau. Do these data, our neuroinflammation data, say anything about the role of amyloid or tau as it relates to neuroinflammation and dementia?

Yes. So the current view of inflammation is that it appears very late. And what the field usually thinks about when they talk about amyloid is that you have an amyloid in tau-driven cell death and dysfunction event, and inflammation and microglial activation appear as a result of that, right? And that you should, therefore, target the amyloid, and that's what all our immunotherapy has done. And then the inflammation is a by standard or a result of that. And what we think really is the opposite that you really -- because of the genetics and [indiscernible] hits that they're really saying that you have innate immune dysfunction and inflammation as a result of that, that starts much earlier, and inflammation is -- chronic inflammation and dysfunction is a cause of that, and it starts early. And so if you target that instead, you are able to then potentially slow down the cascade of amyloid and tau, and it's really what triggers and contributes to that. And so I think your data is consistent with that idea in that you have MCI, mild, moderate and other patients that, potentially, some of them may or may not be robustly tau and amyloid-positive, right? They may be. But maybe if you had some memory forgetters and others that could potentially qualify for a trial, they may not be super robustly amyloid and tau-positive, but they might have inflammation. And so the idea would be target those subjects and prevent all of that to be able to really save off the cell death and dysfunction that could come as a result of that, right? That would be the exciting thing and never wait until you have a lot of amyloid plaques and tangles because I really think at that point, the ship sales.

Thank you. Dr. Jaeger. Moving to you. Drug approval in AD requires cognitive endpoints, and we're perfectly comfortable with that. It seems there are a lot of options to choose from. I've been impressed as I read these protocols that the number of cognitive tests that one can do. So I have 2 questions. First, does the biology of the drug affect the choice of cognitive endpoints or metrics that you use? And two, how do the findings of a study like this help you guide us because you and CJ are leading the charge on the study design. How do you -- how does it guide you to which cognitive scales we should put into the study?

So actually, it doesn't vary much. We choose cognitive scales -- we choose endpoints in clinical trials to demonstrate clinically meaningful improvement in the target clinical manifestation of the disorder. So the way -- this study will help dictate which sample we're going to target perhaps. Will we seek to treat the mildest or a more severe population? And by all means, the measures that we use will differ as a function of the level of severity of the syndrome, and that's for psychometric reasons. That's related to the fact that in order to detect change, you want to avoid being near the ceiling or the floor on any measure. So the molecule and the disease biology doesn't typically affect the outcomes we choose. We're seeking to change the life of a person via the biology, but it's -- we measure how we affect that goal, not the biology when we look for clinical endpoints.

Thank you. CJ?

And Dr. Jaeger, just to follow-up on that. One of the challenges is heterogeneity as it relates to length of time a trial needs to be done, right? I mean, you and I have had this discussion about -- if you make them a more homogenous group of patients, is there an opportunity to see a difference? What are your thoughts on length of time as it relates to clinically meaningful endpoints if you can reduce the heterogeneity? You are on mute.

That's an interesting question because I think it depends on what you think your drug does. So in this field, we've typically, over the years, sort of classified drugs as either symptomatic or disease-modifying. And symptomatic trials, we are typically happy in Phase II even, registration trials to do a 6-month trial. We seek to see the benefit within a very short period, and then we seek to sustain it for a minimum of 6 months to assure ourselves that we're not developing a tolerance to whatever the mechanism is. On the other hand, the disease-modifying drug development strategy assumes that without the drug, there would be a decline as a result of the natural disease progression. And that we are seeking to demonstrate the degree to which we foresaw that decline. For those trials, a fairly long duration is required because of the heterogeneity, not only in the -- between patients in their state at a given moment, but the heterogeneity in the rate and slope of decline, the rate and the shape of the curve. So not everyone is linear. Here, we're dealing with a molecule that is probably a little bit of both. A symptomatic and a disease modifier. And I think that's unique. I'm not sure there's another molecule in development that has that property. And that will uniquely affect the way we decide to develop it. It may make it possible to conduct a pivotal trial following a symptomatic model, meaning it could be shorter. With then longer investigations to determine a degree to which, in addition to being producing symptomatic treatment in a short time window, we also foresaw further decline.

Fantastic. Thank you. And we certainly will have an excellent fun time working that out moving forward. so thanks, Dr. Jaeger.

It's a great company.

So I would like to pose a question to both Max Descoteaux and Steven Chance at Oxford regarding the metrics, the new metrics that we talked about today. So let me start with you Max. So the data we reported today shows that the AFD, the apparent fiber density, in 31 of the 33 fiber bundles increased in average 13% over 6 months. How would you consider this in terms of a meaningful change? And do you have any data as to how this compares with other AD patients over the same time period?

Yes. Thanks, CJ. That's a great question. This is an important change. With disease, we typically see the opposite, so a decrease in fiber density and in overall white matter quality. So integrity tends to go down with disease, and this is true for Alzheimer's disease, but on other neurodegenerative diseases. And we do have data on this from public databases such as ADNI, the Alzheimer's Disease Neuroimaging consortium with hundreds of patients followed over time. And what I can tell you is over 6 months and then a year and then 2 years in time, apparent fiber density time tends to steadily decrease up to 10% decrease. And what you're seeing here is the opposite. The 13% increase, almost, overall, white matter connections and bundles. So this is, to me, extremely positive and probably means some sort of axonal repair of some sort or increased integrity of the white matter so while inflammation-reducing as well. So this is amazing to me.

Thanks. We appreciate that. Dr. Chance. So similarly, with the gray matter structural changes that you see using your cortical disarray measurement. And again, I want to just remind everybody, this is a single patient, just so we're clear about that. To us, the visuals of both of yours and Imeka's data are quite dramatic and really localized to the areas that are of interest in patients with AD. So based on your research, can you speculate on those structural changes and what they are and what they might mean?

Sure. Thanks, yes. So I mean, we created our cortical delay measurement as a unique set of our algorithms to investigate the structural integrity of the computation of new units in the cortex. So that's structure and function in the cerebral cortex. And really those measurements in our CDM assessment are sensitive to combination of cells and their connections in terms of axons and dendrites and the synapses as was illustrated in the diagram that you showed, CJ. And so these are the fundamental elements of what enables the brain to function and the connections to work in the brain. And they are lost in Alzheimer's disease is the typical picture. So normally, we would see cortical disarray increasing, more disruption and breakdown of that structural integrity. What's clearly striking, quite compelling in this case, in this individual case, is that we have the opposite directional effect. And as you said, CJ, the area in which we see that, as illustrated by the gradually accumulating green illustration there in the inferior temporal cortex, we could have found changes anywhere in the brain, and we might have found no changes, or indeed, we might have found the typical picture, which is that it got worse. But in fact, the measurement, as far as we understand it, gets better and it gets better in that inferior temporal cortex, which is critical in Alzheimer's disease. So that's an intriguing and compelling results, I think.

Great. Thank you, Dr. Chance. So this is really for both Dr. Descoteaux and Dr. Chance. So I think you both mentioned this, the question that I had is does -- do these changes surprise you? And I think the answer is yes. I guess -- so let me add on that a little bit. Does the time frame in which the changes are occurring surprise you? And can you tie it to any other diseases or drugs that you have studied that show a similar result? In other words, what are we -- how do we put this into context across different diseases and potentially treatment? So Max, would you like to go first and then...

Yes. My answer is a bit short, not much experience with other drugs because not that many companies go after different drugs with this kind of MRI as a marker longitudinally. So -- but I'm quite impressed at the speed of change here. As you mentioned before in your talk, volumetric changes tend to appear very slowly and probably too late in the disease progression. We seem to be capturing things here very early and much faster than what I see in my typical public databases in Parkinson's disease or Alzheimer's disease. So yes, both ends are impressed and surprised but also impressed by the speed at which these changes occur.

Dr. Chance?

Yes. I mean, I don't think we should be surprised that neuroinflammation could have a profound effect. We know -- and there has been increasing work over recent years sort of realizing that it should be a focus of interest because it's a key part of the neuropathology of Alzheimer's disease. But having said that, I would exactly echo the previous comment that the time span of which we see changes is usually relatively gradual. I mean certainly, the volumetric data has some real limits in terms of how fast one would typically see changes. So that's why we developed a measurement of quality rather than quantity in the brain. You pointed out, CJ, that there was the sort of the hint of changes starting at 3 months. I mean those are small changes, and it could be within the bounds of typical variation. But in that subject, what was particularly compelling to me is that you see that steady progression over those 3-month intervals. So it's quite clear that it forms part of a pattern of ongoing change it seems in that individual.

Great. Thank you. RJ?

This is great stuff. Dr. Pike from Proteome, your system, which is incredibly elegant, is really a big data approach that produces really an embarrassment of riches. I mean the -- CJ just presented what is literally the tip of the iceberg of this fascinating dataset. I'm just a country transplant surgeon. So this is really quite complicated. But from a drug development point of view, there almost seems to be too many choices, right? And because at the end of the day, we need to -- or I guess the question is, do we need to narrow the window to fewer variables so we can make progress? And to narrow -- if we do need to narrow those -- that window, do we do it by increasing the number of samples to improve precision? Or there are other techniques to do this?

Yes. Great observations, RJ. And yes, I won't apologize for the fact that we produce huge amounts of data with each of these studies because I think it's fundamentally important at this stage that we can really look at the whole system and what's happening at a system-wide level. Biology is complicated, you move to dabble on one thing, something might compensate for that. So it's important to look at that strictly. The joy of how we produce a data is you can follow different approaches through it. So you can take a very hypothesis-led approach and say, I know I'm hitting inflammation through a TNF mechanism. Do I see signals across the protium that are consistent with that? And I think the early analysis we've done of this dataset suggests that there are hallmarks around that would appear consistent. We can then look at hypothesis-free. This is an unbiased approach. So we shouldn't necessarily impose a worldview on it. And what's really starting is when you do that, you pull out things like the contactin, which is documented in the literature as being downregulated substantially in AD CSF. And with your treatment, you're seeing it elevate after 12 weeks. I mean that's remarkable. As we have many examples of how just filtering through the data because of the way we provide it Excel, you can do lots of simple filtering tools based on your particular interest in any aspect of biology, and because we have so much data there, of course, you can probably see things that relate to each aspect that you're interested in. Of course, more patient samples at more time points, different doses will all help increase the confidence you have. Do we need to narrow that down to a few things at the end? Yes, we probably do. It's not really practical to do this kind of workflow on a patient in a clinical setting. But if we can narrow that down to 5 hallmark features, we can develop a targeted method to measure those with immunoassays or with mass spec methods that will give you a recruitment tool, a patient stratification tool and indeed a way of monitoring the optimal response to treatment. So I think it is important to start very big and narrow it down to a few things at the end.

Yes. One follow-up question before we move on is, this was CSF. And one of the things we found is that CSF, getting CSF is a challenge. Patients don't like it. Clinicians often bark at it. That's one of the reasons we're very enthusiastic about the MRI assay or systems that we're using. I mean can you comment just on how these CSF observations translate? Well, can you measure them in blood or plasma? And have you had experience on how well they translate?

Yes. So great question. And the straightforward answer is some do, some don't. As you progress through AD pathology. You do get a weakening of the blood-brain barrier. You do get more leakage of CNS proteins into the periphery. But I'm not sure what we measure in plasma is actually CNS-derived most of the time. I think most of it's produced peripherally, but it's altering because of AD pathology. Now we've done lots of TMTcalibrator projects with the brain trigger in plasma with different people in different contexts, and we see lots of the same things moving consistently with what we see in CSF. So it is translatable. It's much more challenging in an abundance level because you get the dilution effect going into plasma. But with the brain trigger there to help you focus in the early discovery experiment, you can then narrow down your targets and get into some of the hallmark pathology markers in plasma samples. And there are methods of doing that a very high sensitivity with immunoassays. So there are assays [ phosphate ] out, for example, that are being evaluated in plasma at the moment that seem to be delivering a credible signal.

Very good. CJ?

Well, thank you, everybody. I'd like to move now to the question and answer.

And Malu, I found a perfect one for you. The question from [ Richard O'Mahoney ] says, what is the advantage to targeting soluble TNF over TNF? Are the side effects the same? And what is the advantage to using soluble TNF over other anti-inflammatories like NSAIDs?

Right. So yes, one maybe less well-known fact about the TNF pathway that it has 2 functional arms. So TNF is made as a transmembrane protein. It's inserted on the membrane. So it first acts as a membrane-bound protein. So membrane-bound TNF is critically important for immune function and fighting infection. And as a membrane-bound protein, it's also important for myelination. And so when you give a drug like an antibody, or decoy receptor and all the currently marketed anti-TNF drugs block both the membrane-bound and then the species that is cleaved by matrix metalloproteinases like taste, and that creates the soluble form of TNF, which is the target of XPro1595, then the blockage of both the membrane-bound and the soluble creates a problem in that you're blocking the protective myelinating and immune function of the membrane-bound as well as the inflammatory side of the soluble TNF. And so people who are chronically dosed with those drugs are prone to opportunistic infections. And after some time, they also appear to develop sort of MS-like demyelinating symptoms. Both of which are reversible if you come off the drugs, and then you, of course, have the inflammation problem. So that, when this drug was being developed, was one of the concerns that how could you potentially compete with those drugs if you couldn't block both arms. But it actually appears that being selective to the soluble arm is what you want to do to be able to just block the inflammatory pathologies when there's excess soluble TNF that you can neutralize with the XPro. So that's the advantage of the soluble targeting that you don't have the liability of blocking the protective immune function of the membrane-bound arm of TNF. And just to add and answer the question also by [ Alex Parent ], which is the fact that this targeting of soluble TNF, are we looking at a drug that could be used for AD prevention and delaying of disease or mostly to reduce inflammation for patients with Alzheimer's already? I think it could be both, right, because inflammation starts early, but it persists through the disease is what we think is what we're looking at. It's not -- inflammation is not a 1 size fits all, but it has different roles throughout these neurodegenerative diseases.

Great. Thank you, Malu. Appreciate that. So we have time for one more question here. And I want to -- so this is from Jason McCarthy, and Max and Steven, I'm going to ask you to weigh in on this as well. And the question is, where are the regulatory agencies in terms of comfort level and the reliance of imaging approaches in AD trials? Meaning the prior Lilly P3 is heavily relied on imaging at least for enrollment to mitigate heterogeneity and ultimately failed. Do you need much more extensive imaging work in a larger Phase II to validate this approach as a go-forward component for XPro for the use as a surrogate potentially for cognition changes? So let me start by saying the answer to that is yes. This is not an approvable endpoint at this stage. But again, what we're looking for are biomarkers that may reflect a clinically meaningful change. And the extent that we can use that in a Phase II study to help identify patients or to help us or to give us an idea of whether or not there will be an ultimate clinical approval, I think, is lots of utility. But Steven and Max, what are your thoughts on how the FDA or other regulatory agencies have -- think about this as surrogate endpoints? I'm guessing it's probably -- or way too early for that conversation, but interested in your thoughts.

If I can comment, which is -- cheers, which is, yes, it's early. CJ, you're absolutely right in all of those statements. Clearly, the data presented here, this was a single case, as you pointed out, much -- it would be great to do a lot more work on this. I would add, it's certainly public knowledge that Oxford Brain Diagnostics, our method to CDM analysis has a breakthrough device designation from the FDA. So we know that we're engaged with the FDA, and we know that they are very interested in looking at these sorts of imaging markers and also markers that take things beyond as you said, the more traditional methods of volume to try and pick up changes at a more sensitive sort of level and sort of detect changes earlier as well. So I think there's an appetite for it. But obviously, it needs more work.

Max, any comment?

Same pitch, very early for us as well. But on the way there, and I'm really looking towards more trials where primary endpoints are imaging endpoints, not just secondary endpoints because we need this to be more specific and sensitive.

Well, excellent. Thank you, both. RJ?

Yes. So I think we'll close here, and I want to thank the panel. Super, super discussion. And I would finish by saying, our goal today was to connect the dots between white matter-free water and a novel, which is a novel biomarker of neuroinflammation. And the traditional biomarkers that people are used to in the CSF. We introduced new ways to measure the effects of XPro therapy on gray and white matter, and we previewed the impact of decreased neuroinflammation on neurodegeneration and synaptic function using a powerful big data approach to study the CSF proteome. Targeting neuroinflammation was expected to produce biologic changes. But I have to say I am very surprised and intrigued by the structural changes suggested by the neuroimaging modalities. These data reinforce our belief that XPro decreases neuroinflammation in patients with Alzheimer's disease and supports our commitment to initiating the Phase II trial with XPro this year. Before we ended the session, I'd like to make one more point. We always like to say that XPro targets innate immune disregulation. CJ, do you want to advance the slide there? And so we have begun to think about using AI by combining all these interesting biomarkers into a -- to help us sort through this biomarker, shall we say, forest. And we're using -- working with the company called NetraMark, which is artificial intelligence company that helps biopharma develop new ways to understand complex diseases. So on the following figure, what I found was that we may have actually an image here that shows that what we like to say all the time is that XPro forces the wayward dysfunctional, chaotic innate immune system that's associated with chronic inflammation to actually behave normally. So we gave NetraMark really a small portion of the Proteome data. They fed it into their AI algorithm. And the result is shown here, and my interpretation maybe it's certainly preliminary and may be wrong. But let me tell you how I read this. As you can see in the top blue dots, this is all the patients from the 1 milligram group. We thought we were enrolling a very homogeneous group of patients. And as you can see, they were all a bit different as you become more sophisticated in your biomarker analytics. But remarkably, after 3 months, they really converged on a single point. And to me, this looks what looks -- this is what normalizing looks like. In other words, we took the dysfunctional chaotic innate immune system and brought it back to normal. So time will tell if that interpretation is right. But we believe the future is bright. So stay safe, wear your masks and please get vaccinated. Thank you very much.

So thank you, everybody. If you have additional questions, please do email them. We didn't have a chance to get to all of them. We thank our panelists. We thank all the people that came to join us today on the webinar and please be safe, wear what are your masks and get vaccinated.