Good morning, and welcome to the Applied Therapeutics Diabetic Cardiomyopathy Expert Forum. All participants will be in a listen-only mode, and should you need any assistance during the call, please signal a conference specialist by pressing the star key followed by zero. After today's presentation, there will be an opportunity to ask questions. To ask a question, you may press star, then one on your touch-tone phone, and to withdraw a question, please press star then two. Please also note that this event is being recorded. I would now like to turn the conference over to Dr. James Januzzi. Please go ahead. Yeah, thanks very much, and good morning to everyone joining us. My name is James Januzzi. I'm the Hutter Family Professor of Medicine at Harvard Medical School, and I direct heart failure and biomarker clinical trials at the Baim Institute for Clinical Research. I want to thank Applied Therapeutics for hosting this important forum because it gives us an opportunity from the academic side to talk to you all about the importance of heart failure and diabetes, and specifically about diabetic cardiomyopathy. It's important to recognize that this is supported by Applied Therapeutics, who are supporting the ARISE-HF clinical trial, and I'll be speaking about ARISE-HF. In addition, I'll be joined by my colleague, Dr. Greg Lewis from the Massachusetts General Hospital in just a few minutes. What I'd like to do to start things off first is to talk a bit about diabetic cardiomyopathy as a disease state. It's not a new diagnosis, actually, but only recently have therapies to treat diabetic cardiomyopathy been developed. Then we'll talk about the ARISE-HF registrational phase 3 clinical trial, for which I will be serving as the global chair. Heart failure. Anyone that knows the space knows that heart failure is the number 1 most increasing diagnosis among elderly individuals globally. It's reaching epidemic proportions, and so there's an increased emphasis on recognizing the risk factors for the development of heart failure at its earliest stages, with an idea to treat and prevent the disease, if possible. Diabetic cardiomyopathy is a diagnosis that was actually first recognized in the early 1970s as a heart muscle disease affecting individuals with diabetes who lacked other causes for heart failure, such as active coronary disease or hypertension. This is a form of Stage B heart failure at its earliest stages. Stage B, so what is Stage B heart failure? Well, in the universal definition of heart failure, we define heart failure in four stages. Stage A is the risk for heart failure. Stage B is asymptomatic structural heart disease, identifying individuals at high risk to progress to the symptomatic stages, which are stages C and D. So it's important to recognize that whether we treat at Stage A or Stage B, the goal is to avert onset of overt heart failure symptoms. So diabetic cardiomyopathy, as I said, is a heart muscle disease. It's caused by the underlying metabolic changes in heart tissue, which leads to scarring in the heart, and this scarring progresses on to congestion and other signs and symptoms leading to symptomatic heart failure. Diabetic cardiomyopathy may occur in all forms of dysglycemia, including both Type 1 and Type 2 diabetes, even despite presently adequate glucose control. I'll talk about the risk factors for what may lead to diabetic cardiomyopathy, but it's an important point to recognize that glycemic control alone is not sufficient once the disease is present. So how do we recognize diabetic cardiomyopathy? Well, I mean, it's a diagnosis in those with diabetes, right? So it makes sense that individuals with either Type I or Type II will be at risk. As I've already indicated, Stage B heart failure is the presence of either structural or functional cardiac abnormalities. So we may see abnormalities in heart function. We may see abnormalities in heart size. We will see impaired exercise tolerance, which we'll talk about in much more detail, as well as, and very importantly, in the universal definition of heart failure, we also acknowledged that elevations in prognostic biomarkers, such as natriuretic peptides and NT-proBNP or high-sensitivity troponin, may also mark the presence of heart muscle disease in these patients. Now, importantly, as I've already said, although coronary artery disease and hypertension can certainly coexist in people with diabetic cardiomyopathy, in order to really make the diagnosis of diabetic cardiomyopathy, it's necessary to exclude active coronary disease or uncontrolled hypertension as the cause of the heart muscle disease in these patients. So how common is it? This, to me, was one of the most eye-opening findings as we were designing the ARISE-HF trial and trying to come to some conclusion as to how many patients we would need to screen in order to identify the presence of diabetic cardiomyopathy. The take-home message from this slide, based on numerous population-based studies, it's estimated that somewhere in the range of, around 1 in 4 to 1 in 5 patients with diabetes have the presence of diabetic cardiomyopathy, even in the absence of other forms of heart disease. What this means is that nearly 100 million patients worldwide may have the diagnosis, nearly 10 million in North America alone. In those individuals with the diagnosis, about 1 in 4 will progress on to symptomatic heart failure in a very short period of time, only within 1 to 2 years, really indicating the high-risk nature of this diagnosis. So we, as clinicians, increasingly recognizing this diagnosis, need to try to figure out not only how to recognize and intervene, but, you know, the problem is the interventions at this point are nonspecific. We counsel our patients on heart failure risk reduction. We talk with them about the things that they can do to avoid progression to symptomatic heart failure. But unfortunately, there are no therapies that specifically target the metabolic derangements that cause diabetic cardiomyopathy. So how do we recognize it? As I've already said, there are tools that can be used... biomarkers like natriuretic peptides. I'll remind everyone that the American Diabetes Association now recommends a yearly screen with N-terminal pro-BNP or high-sensitivity troponin in order to identify the signal of risk for heart failure, so the identification of Stage B heart failure. But in whom do you look for it? Well, as I've already said, this is a diagnosis that may affect Type 1 or Type 2 diabetes. It is also related to advanced age and longer chronicity of hyperglycemia. In addition, as was originally described in the first manuscript that identified the presence of diabetic cardiomyopathy, concomitant renal impairment. I'll remind people that the presence of diabetic kidney disease is almost a one-for-one risk factor for onset of heart failure in individuals with diabetes, as well as the presence of other micro or macrovascular complications of diabetes, including neuropathy, nephropathy, and retinopathy. Now, how do we assess the implications of diabetic cardiomyopathy? In other words, what are the ramifications, not just in terms of the prognostic meaning and the risk for progression to symptomatic heart failure, but how does the patient feel? It's an interesting and concerning issue that our patients often will make adjustments in their physical activity to compensate for their limitations. But if you challenge patients with a cardiac functional capacity test, so for example, a cardiopulmonary exercise test, which my colleague, Dr. Greg Lewis, will be talking about, we can really begin to appreciate just how significantly impaired these individuals may be. So one way to assess this, as I said, is a cardiopulmonary exercise test, where we assess the peak oxygen consumption or peak VO2. This is a metric that's used to describe the oxygen level during maximal physical exertion, and this is a way to really, essentially, in an objective manner, identify the impairments that these patients may have. Various measures are used to ensure that maximal physical exertion is reached in these tests, which really helps us to use them in clinical trials, and this will be discussed in the next section by Dr. Lewis. But on your slide here, you see some examples of what the peak VO2 necessary is for that VO2 max. Indeed, when we look at individuals with diabetic cardiomyopathy, we can appreciate that cardiac functional capacity is impaired in these people. So on this slide, you see the various stages of heart failure, Stage A, Stage B, Stage C, and Stage D. A heart failure specialist, like my colleague, Dr. Greg Lewis from the Massachusetts General Hospital considers these stages when he and his colleagues are evaluating patients for the diagnosis of heart failure. The all-important stages where prevention is still possible, Stage A and B, are indicated on the slide. Very, very importantly, although overt symptoms may not be present among individuals with Stage B heart failure, it's not at all unusual on provocative testing to identify a decrease in functional capacity when compared to, say, a Stage A patient. So this ability to identify essentially heart failure in development that has not yet become completely overt is a prime example of how we might target a therapy to help improve outcomes in these patients. Mechanistically, how might we intervene? Diabetic cardiomyopathy is due to a number of different mechanisms, including impaired cardiac energetics. So increased fatty acid oxidation is one pathway that has been focused on using partial fatty acid oxidation inhibitors. Accumulation of advanced glycosylation endpoints is also a marker of the presence of diabetic cardiomyopathy, so treatments that reduce advanced glycosylation endpoints might be expected actually to improve progression or avert progression to the manifestation of the disease, and we're going to talk about that in a little bit. The presence of myocardial fibrosis, remodeling, which is the way that the heart responds to injury, and is a marker of risk for progression to heart failure, is also present in these individuals, as is the presence of inflammation and impaired calcium handling. So this really, you know, significant cascade of adverse events occurring in the heart muscle ultimately conspires to lead to a myocardial dropout, so-called apoptosis, replacement with scar, fibrosis, as well as thickening of the heart muscle and dysfunction, leading to an enhanced risk for heart failure. So that leads us to the concepts behind treatment for diabetic cardiomyopathy. And as I said in my opening slides, until recently, there was no treatment that was available. But now, thanks to a number of different research, you know, pathways being examined, a number of therapies, including, and importantly, aldose reductase inhibition, which is leading the charge for the treatment of diabetic cardiomyopathy, is poised to actually provide us a therapy for this important diagnosis. So I'm going to oversimplify things a little bit. I'm just a cardiologist. I'm not an endocrinologist. But we all know that glucose is an important fuel for the body, and the cardiomyocytes, the heart muscle cells, have the highest energy requirements of any cell type in the body, so they require ample amounts of fuel, whether it's from glucose or other sources, in order to generate energy and to provide squeezing strength for the heart. In the setting of diabetes, there are a number of deranged pathways that cause a shift in how glucose is handled in the myocardium.... So rather than following through the hexokinase pathway, the normal pathway, there are a number of reasons why, but there's a shift into the polyol pathway, leading to energy depletion due to impaired cardiac energetics, oxidative damage, formation of those advanced glycation end products that I mentioned earlier. This is the downstream implication of higher glucose levels causing tissue damage, ultimately leading to fibrosis and remodeling. To a large extent, much of this damage is related to the production of sorbitol in the myocardium. Sorbitol, a byproduct of the metabolism of glucose within the myocardium through the polyol pathway, is in part generated through the activity of aldose reductase. So aldose reductase inhibitors may be a very valid way to mitigate the tissue damage associated with sorbitol production in the setting of diabetic cardiomyopathy. The problem is that older generation aldose reductase inhibitors had actually been developed many years ago. The problem is they were limited by off-target hepatotoxicity, including competitive inhibition of aldehyde reductase leading to liver toxicity. So although aldose reductase inhibitors seemed like a valid option in early days, because of this off-target hepatotoxicity and relatively lower potency against aldose reductase, we were sort of stymied until the development of AT-001. So this drug is a highly potent aldose reductase inhibitor, 1,000 times more potent for its on-target efficacy. So target engagement for inhibition of aldose reductase is much, much greater with AT-001. Furthermore, it has no off-target inhibition of aldehyde reductase, avoiding the hepatotoxic side effects of older aldose reductase inhibitors. Furthermore, it has broad exposure, not only in heart tissue, but also nerve tissue. And for those that do not remember the back history of aldose reductase inhibitors, there was a lot of interest and enthusiasm to look at them as a treatment for peripheral neuropathy and diabetes as well. So there's upside there as well for AT-001. So when we look at comparative studies between AT-001, also known as caficrestat, versus zopolrestat, we see that there's a improved aldose reductase inhibition at lower doses with AT-001, allowing for target engagement at lower doses, which is very important from a drug development perspective. And as I've already indicated, we do not see any aldehyde reductase inhibition with AT-001. So that both of those findings, as I've already said a number of times, really put it forward as a robust candidate for treatment of individuals with possible diabetic cardiomyopathy. Now, I've already mentioned the importance of the downstream tissue damage associated with the presence of intracellular sorbitol. Through treatment with AT-001 inhibiting aldose reductase, one might expect, through reductions in sorbitol, that we would see attenuation of reactive oxygen species generation, reduction of advanced glycation end products, and other downstream damage in the tissue. This is an example of the inhibition of reactive oxygen species generation, both in tissue as well as in mitochondria associated with treatment with AT-001. You can see on the left panel a significant upregulation of reactive oxygen species, which is attenuated in the right panels on both sides of the screen through treatment with AT-001. So by preventing production and accumulation of reactive oxygen species, we control the amount of myocardial injury in the heart and thereby reduce senescence, which is a sciencey word for tissue aging, which ultimately leads to fibrosis, cellular dropout, and dysfunction. So this is encouraging in vitro data. Furthermore, in animal models, we can actually see that treatment with AT-001 improves cardiac energetics, it improves cardiac function, and prevents left ventricular hypertrophy, and reduces fibrosis and adverse remodeling in an animal model of diabetic cardiomyopathy. So we have in vitro data, we have animal model in vivo data showing that through treatment with this highly potent aldose reductase inhibitor, we can reduce myocardial dysfunction in a well-validated model of diabetic cardiomyopathy. So for all these reasons, this leads us to the design and execution of the ARISE-HF, clinical trial. Let's talk about ARISE-HF. Very exciting clinical trial. I'm very grateful for the opportunity to lead this study, which is the first, essentially first human trial for the prevention of progression of diabetic cardiomyopathy. This is a randomized, placebo-controlled trial of patients with diabetic cardiomyopathy at high risk of progression to overt heart failure. And I'll describe how we identify these people and how we stratified our patients to be at high risk for overt heart failure, really essentially providing us the ideal opportunity to evaluate the effects of AT-001. The study population of individuals, around 675, were randomized 1:1:1 in an ascending dose fashion of placebo versus 1,000 or 1,500 milligrams twice daily of AT-001. They were treated for just over a year, with the primary endpoint of functional capacity, peak VO2, and we're gonna be looking at the change from baseline. Dr. Lewis will explain the importance of this endpoint and how we'll go about doing it. With key secondary endpoints that look at, among other things, a physical activity scale known as the PACE, NT-proBNP. I've already mentioned the importance of that biomarker to identify risk for heart failure onset, health status as assessed by the Kansas City Cardiomyopathy Questionnaire, as well as downstream, following these patients to attempt to see if AT-001 has an impact on progression to overt heart failure. It's important to emphasize, however, the primary endpoint at 15 months is gonna focus on functional capacity, but of course, we have the opportunity at a longer time point to evaluate for the clinical endpoints of heart failure progression, change in echocardiography, as well as potential change in exploratory cardiac biomarkers. We're looking forward to our readout at year's end this year, in the next month or so. This is a study that was global. We enrolled in numerous studies in Asia Pacific, Europe, North America. And as I said, we're just under 700 patients, enrolled very well, actually, in this clinical trial, in order to identify individuals with diabetic cardiomyopathy at, at high risk to progression to overt heart failure. We included individuals with type 2 diabetes, either at an advanced age or with prolonged diabetes. So again, if you recall earlier, the risk factors for diabetic cardiomyopathy include either of those. In addition, we required the presence of Stage B heart failure. So these patients either had to have echocardiographic abnormalities, elevation in relevant cardiac biomarkers, or both. And in addition, we also required the presence of an impaired functional capacity on a maximal cardiopulmonary exercise test. In order to get into the trial, study participants or potential study participants needed to provide us an RER of 1.05, and Dr. Lewis will explain why that is critically important for the validity of the data from CPET, as well as to show impairment in VO2 max, less than 75% of their age, gender-predicted normal. Now, we excluded individuals who actually had heart failure because at this point, they have Stage C or D. That includes people using a water pill, a loop diuretic. We don't want people who have symptoms but just hadn't yet been identified as having heart failure. We excluded people with active ischemic heart disease, revascularization, et cetera, because, again, we wanna try to create as pure a phenotype as possible. And that also included excluding people who had valve disease or other structural heart disease that might explain their impairments in exercise capacity. And other exclusionary criteria, very notably, included out-of-control blood pressure or out-of-control diabetes, because once again, we wanted to make sure that while that may have been present in the past, at least right now, these patients are well managed to really take off the table the question of what, what would one expect from tighter glycemic control or tighter blood pressure control? That is not applicable in this trial because these patients actually have well-controlled diabetes and well-controlled blood pressure. So to ensure balance between our active and randomized groups... I'm sorry, active and placebo-treated groups, we stratify patients at baseline based on their age, their sex, the presence of concomitant treatment with a GLP-1 receptor agonist or an SGLT2 inhibitor, as well as baseline cardiac functional capacity. So these patients were well stratified for relevant contemporary risk factors in 2023, including the pharmacotherapy. In addition, there are numerous committees that oversee that have been overseeing the execution of the ARISE-HF trial. This includes a steering committee, executive committee that I'm glad to lead, as well as a blinded cardiac endpoint adjudication committee for events, as well as safety, firewalled by firewall from the data. The data itself are unblinded for an unblinded data monitoring committee, which the steering committee is firewalled from. Sorry about that. So then, our baseline characteristics. As I said, we've enrolled 675 individuals with diabetic cardiomyopathy. The take-home message from the baseline data, it's really quite fascinating. This is an older population, as one might expect, notably equally balanced between men and women. So 50/50 in this clinical trial, men and women. They have long-standing diabetes. They have prevalent risk factors, as well as abnormal cardiovascular biomarkers and abnormal imaging, as one might expect for a Stage B heart failure population. When we look at the concentrations of natriuretic peptides in these individuals, again, that was one of the inclusion criteria to get into the study. We can appreciate that the mean value, it's a non-normal distribution, so we use medians here, of 71 nanograms per liter is elevated for an age-matched population. But people who know this biomarker might say, "Geez, you know, that's a little left-shifted, a little bit lower than I might have expected." But I'll remind you that this population generally has obesity, and obesity has a tendency to reduce natriuretic peptide concentrations, which is why we used an inclusion value of 50 nanograms per liter in order to get into the trial. So this is an at-risk population with higher concentrations of natriuretic peptides, as well as, of course, abnormalities in cardiac echo, structure and function. Notably, the N-terminal pro-BNP data, these are from baseline, results in ARISE-HF, really tells an important story, and across the quartiles of NT-proBNP.... We can see that concentrations of NT-proBNP are associated with worse health status, as assessed by the Kansas City Cardiomyopathy Questionnaire, as well as reduced activity levels. So higher NT-proBNP is associated with lower PACE scores, one of the endpoints that we're looking at in this trial. Furthermore, as we might expect, but it's reassuring to see, NT-proBNP at higher concentration in this trial is also associated with decreased exercise capacity, with a meaningful reduction, for example, in the duration of the cardiopulmonary exercise test, a meaningful reduction in the VO2 max, as well as a meaningful increase in something called the VE/VCO2 slope, something that Dr. Lewis will be talking about. So these findings, I think, are really striking because it indicates how NT-proBNP, after we have a successful trial, may be useful to identify the presence of diabetic cardiomyopathy. And indeed, when we look at cardiac functional capacity, VO2 max or peak VO2, it not only correlates with NT-proBNP, but it also correlates with the PACE score, something that will be an endpoint. So each of these, these numbers, the PACE score, peak VO2, NT-proBNP, these are all things that we're going to be looking at as important endpoints in the ARISE-HF trial. So in conclusion, I'm really grateful for the opportunity to speak today about diabetic cardiomyopathy. This is a serious complication of an incredibly prevalent diagnosis, that being diabetes. So diabetes is exceedingly prevalent, and diabetic cardiomyopathy affects probably about one in five individuals with the diagnosis of diabetes. There are currently no treatments specifically approved for diabetic cardiomyopathy, and ARISE-HF is the first registrational study among individuals with diabetic cardiomyopathy to evaluate an important therapy, for this diagnosis. The baseline data from ARISE-HF suggests that we have the right target population. These are patients with abnormalities in biomarkers, abnormalities in echocardiography, and importantly, have impaired exercise capacity. Our primary readout is expected at the end of this calendar year. Only about a month from now, we should be starting to get towards data. AT-001 may be the first treatment approved for this highly prevalent diagnosis. I look forward to discussing the results once we have them available. With that, I'm going to turn it over to my colleague, Dr. Greg Lewis. Dr. Lewis is the head of our heart failure section at the Massachusetts General Hospital and directs the Cardiopulmonary Exercise Testing Laboratory at the Massachusetts General Hospital. Very importantly, Dr. Lewis is a world expert on the importance of cardiopulmonary exercise testing in clinical trials and its clinical applications. There's nobody better to discuss cardiac functional capacity than Greg. Thanks for joining me today. Thank you very much, Jim, and I'm gonna pick up on the talk that Dr. Januzzi just provided to you to give you a little bit more detail on this primary endpoint measurement, the cardiac functional capacity. So by way of background, the peak oxygen uptake, the gold standard measurement of cardiac functional capacity, and the derivation of peak oxygen uptake is essentially the product of the heart rate, the cardiac stroke volume, and the ability to utilize the oxygen that gets delivered to our peripheral tissues. We can measure peak VO2 directly using a metabolic cart or a cardiopulmonary exercise testing machine that I'll tell you about in a little bit more detail. And you can see on the right there, the scale of different activity levels and what the metabolic costs of those activity levels are, as measured in oxygen uptake. So all of us in a state of rest are using about 3.5 milliliters of oxygen per kilogram per minute. Then when we do simple activities like walking at 2 miles per hour, that increases metabolic rate by about 2- to 3-fold. If we increase our pace of walking to 4 miles per hour, you can see that increases metabolic rate to about 4-fold. Now we start to get into the height of what some of these patients that you've been hearing about with diabetic cardiomyopathy can afford in terms of their peak metabolic capacity or their peak oxygen uptake. So we'll talk about the relevance of this endpoint for the ARISE clinical trial. First, a little bit more background about cardiopulmonary exercise testing. Here is visualized the cardiopulmonary exercise testing machine. What we do is we monitor patients closely during either cycle ergometry or treadmill exercise. We measure oxygen levels with a pulse oximeter. We measure electrocardiogram response to exercise, so this also serves as a type of a stress test that many of you may be familiar with. Then we have a gas analyzer that importantly quantifies breath-by-breath expired gas exchange, so that we can actually measure the amount of oxygen being taken up and utilized by the body. Again, this is a gold standard measurement of cardiorespiratory fitness. We can use these measurements in order to measure the peak oxygen uptake, and that is the primary endpoint of this trial. Some more details about cardiopulmonary exercise testing. First of all, it's a reliable form of testing. These measurements of peak oxygen uptake are reproducible, in normal individuals, in individuals with overt heart failure, and by way of translation, in individuals with diabetic cardiomyopathy. It's important to maintain and service the equipment. This is the role of the core laboratory for this trial, where all of the sites that participated had to go through a qualification protocol, had to perform testing on standardized individuals in order to ensure reproducibility of their results. And then they all follow the same protocol such that these individuals that are participating in the trial, perform the same cardiopulmonary exercise testing protocol throughout ARISE, and of course, the same protocol at baseline and at that 15-month endpoint that we heard about from Dr. Januzzi. The central core lab receives all of this data, importantly, receives the data in breath-by-breath fashion, so that we can look at the same exact way at all of this data in order to determine what the peak oxygen uptake was, both at baseline and following the 15-month or 27-month enrollment period in this trial. Dr. Januzzi mentioned the respiratory exchange ratio. This is critically important. The respiratory exchange ratio remains below one when we are performing aerobic metabolism. So low level of exercise results in a normal Respiratory Exchange Ratio of less than one. When you begin to exceed one, that's because your body is producing lactic acid at a rate that exceeds the ability to metabolize lactic acid. And as a result, that respiratory exchange ratio begins to increase as we eliminate more CO2 than we take in VO2. We've set a threshold of 1.05 for this trial, meaning that when patients in the trial exceed that threshold, we can be confident that they've given us a maximum volitional effort. This is important for the sake of validity of comparing baseline data and follow-up data. And as Dr. Januzzi just showed in one of his slides, the actual levels of respiratory exchange ratio achieved in ARISE were up in the 1.17-1.18 range, which is excellent in terms of ensuring that there was maximum volitional effort for the measurement of this important primary endpoint. So in terms of the core lab, as mentioned, we are responsible for the oversight of all the sites performing these cardiopulmonary exercise tests. We've taken them through training, through qualification steps, central oversight, and importantly, we only permit individuals to enter the trial if they meet the criteria set by the core lab in terms of a Respiratory Exchange Ratio being at least 1.05, and that Peak VO2 being less than 75% of predicted as measured and adjudicated by the central core lab. If there's incorrect performance of a test, then there's prompt resolution of that by repeating the test, if there's a technical issue that comes up during the acquisition of this data, and then all of these reports are read centrally by the core lab. Importantly, a change in peak VO2, even a modest 6% change in peak VO2, is clinically meaningful for patients. In one of the largest heart failure trials looking at peak oxygen uptake, a 6% change in peak VO2 predicted hospitalization for heart failure, predicted all-cause mortality, and as you can see here in the table, was related to all these important endpoints for patients that are descending into a state of heart failure with diabetic cardiomyopathy. So, peak VO2 offers several advantages. You know, there are different ways to measure functional capacity or to measure exercise tolerance, things like a six-minute walk test, for example. But we know that the measurement of peak VO2 is more precise, more reproducible than other measurements of physical function that are sometimes used in clinical trials. Importantly, we gain insight into why patients are limited when we perform a cardiopulmonary exercise test. If we just ask somebody to walk down the hall or if we just ask them about their physical activity levels, then we don't know what limited them when they are performing a given activity, such as a 6-minute hall walk. Whereas the cardiopulmonary exercise test, as I showed you, has this added advantage of we're measuring blood pressure during the test, we're measuring the ECG response, we're measuring the gas exchange patterns like the peak VO2. And so we collect a great deal of physiologic data that's gonna provide insight for us into diabetic cardiomyopathy over time and the hopeful prevention of its progression to overt heart failure. Importantly, if you repeat a cardiopulmonary exercise test on an individual, then you do not see a training effect. You're just as likely to have a higher than a lower value of peak VO2 with repeated measures, and this also differs from other measures of physical activity, such as a six-minute hall walk. So we aimed for, and achieved precision measurement in this trial, and we are relying on the high reproducibility, of this measurement, for the ARISE primary endpoint. This is the peak VO2 that is, been achieved at baseline, 15.7 milliliters per kilogram per minute, similar mean and median values, and you can see the interquartile range is there. And again, to remind you in terms of what this means for our patients, you can see that a peak VO2 of that 14-17.5 range is what it takes in order to achieve a walking speed of 4 miles per hour or to accomplish household tasks. And so if one wants to go beyond this peak metabolic rate, then they begin to exceed what they're able to afford in terms of more vigorous physical activity. And as Dr. Januzzi mentioned previously, this leads patients to curtail their physical activity and to limit what they're doing in order to essentially guard against exceeding what they can afford metabolically. And so this is the range in which the enrollees in ARISE fell. And if we think about this in context of what other activities cost, again, you can see that this is where our ARISE heart failure population falls. They're really right on the cusp of heart failure. If we think about patients with overt heart failure, preserved or reduced ejection fraction, they typically come in at a peak VO2 right around 14-15 milliliters per kilogram per minute. So these patients, we've been successful in identifying and recruiting a patient cohort that are right on the edge of this heart failure phenotype. And then, again, you can see what these different activities cost. A resting metabolic rate of 3.5 mL per kilo per minute. Walking 3 miles per hour comes at a large percentage of what these individuals can afford. Then you can see things like carrying groceries or trying to jog at five miles per hour, comes at a metabolic cost that's beyond what the ARISE-HF population can afford. So the peak VO2 is a primary endpoint in ARISE-HF. We designed this to measure the difference in peak VO2 between the active high-dose AT-001 and the placebo group. We're looking at the natural history of the diabetic cardiomyopathy over this period of 15 months. And what we're expecting to see, particularly with this population that's right on the cusp of heart failure, is this decline over time in peak VO2 in the placebo group of this study. However, with the introduction of the AT-001 and the prevention of further damage, from the aldose reductase activity and the opportunity to improve cardiac energetics, our goal is to preserve peak VO2 or potentially improve upon it, unlike the placebo arm of this population, where there tends to be this progressive decline in functional capacity as these patients descend into the state of overt heart failure. As far as the powering, this is a well-powered study. I can certainly attest to that, having run core labs for many trials with peak VO2 primary endpoints. There's 90% power to demonstrate a 6% difference in the peak VO2, from baseline to 15 months between the active and the placebo-treated groups. We chose the 6% threshold because it's been shown to represent a meaningful clinical difference. Meaningful in terms of future outcomes of these patients, and meaningful in terms of what these patients can afford. Remember, I showed you that they're on the steep of this curve in terms of the metabolic costs of activities of daily living and what they can afford. So even a 6% increment in Peak VO2 for these patients, or 6% difference, is quite meaningful in terms of what they can and can't do on a day-to-day basis. The ARISE-HF study, at baseline, there's a significant correlation, as Dr. Januzzi showed, between Peak VO2 measurements and the physical activity scores. There's significant correlation between peak VO2 and the very important biomarker of NT-proBNP, and this suggests that if the study meets the primary endpoint, that several secondary endpoints will be met as well. Thank you for your attention, and we'll be happy to take questions. ... Okay, I think we're ready to open the lines for Q&A. Operator, are there any questions on the line? Yes. Currently, we have one question here from Brian Skorney with Baird. Please go ahead. Hey, good morning, everyone. Thanks for going through this, really helpful. I mean, I guess to start, can you comment at all about your assumptions under the powering in terms of standard deviation? And from what you're seeing in the blinded data right now, is standard deviation coming in line or seeing better or worse standard deviation? And then, just are you looking at any event rates in the studies, hospitalizations, MACE? And is there anything potentially important here that you could see enough of accrued to have any sort of potential information on trends? Yes. So maybe I'll start off a little bit and hand it over to the rest of the panel. So it is still a blinded study, so none of the four of us have seen any unblinded data. But we will soon because the study is coming to a conclusion and primary endpoint readout. We have done as much as we can based on the blinded data. And we have analyzed standard deviations. We've looked at, in a blinded fashion, the percent of patients that decline on peak VO2, that stabilize, that improve. And it almost perfectly falls out into a third, a third, and a third. So of course, it's still blinded. We don't know who those patients are that are declining in the one-third of the study population. But it should be the placebo patients because it was randomized at a 1:1:1 ratio. And, you know, going by that, we would hope that the one-third that are pretty stable might be the low dose group, and perhaps the third that are actually improving on Peak VO2 might be the high dose group. So what we're seeing based on the blinded data set is a very good reproducibility, and we're seeing what we hope to see in terms of the spread of the trial. We are looking at progression to overt heart failure and hospitalizations and some of the hard heart failure outcomes. Because although that's not the primary endpoint, it is a secondary. We believe that that could really help in the long-term implementation of use of this drug if we achieve the primary endpoint, and the drug is approved. Jim, did you want to say anything? Yeah. No, thanks very much. So for the hard cardiovascular endpoints, Brian, the endpoints are being adjudicated by a blinded endpoint committee. And so we'll have those outcomes to analyze after completion of the trial. Of course, we're powered for the VO2 max. But nonetheless, it's important to gather those endpoints, if at all possible. A question to Greg Lewis. So, Greg, the mean and median actual values of VO2 max in the study are actually nearly identical, so it's a pretty normal distribution. What's your interpretation of the standard deviations on this? Yes, thank you. And thank you for the question. It's an important one. You know, this study was designed to afford the standard deviation and repeated measures of VO2 that was larger than our anticipated standard deviation. And, you know, we know that based on what was shared in terms of the 90% power in order to detect a 6% difference between groups. That 6% happens to fall right at that, you know, 1 mL per kg per minute, which we also think is a clinically meaningful difference when you're starting off in the 15s for your peak VO2. And so we think you have excellent power in this study. If you look at other studies, using peak VO2 as a primary endpoint, in the heart failure populations, this sample size exceeds the sample size of other trials. And so we've afforded the ability to have that standard deviation in repeated measures of peak VO2, be in significant excess of what we are anticipated to see in ARISE. So it's a, it's a well-powered study for a, a Peak VO2, primary endpoint. Thanks, Brian. Operator, do we have another question? That was the only question in queue at this time. Okay, we do have a question online. So the question online asks: "If the glucose levels are controlled, how do we think about sorbitol levels and continued impact on cardiomyocytes? And how are the shape and distribution of these 3-thirds that we just spoke about in the blinded data sets?" Okay, there's a lot of questions. "And, how much progression of VO2 max do we expect, in a non-treated population or the placebo group?" Okay, so let's take those one at a time. Okay, so, we didn't talk a lot today about our phase 1/2 study with AT-001, but now is probably the time to speak about it a little bit. So we did a phase 1/2 trial, of course, before this phase 3 ARISE-HF study with AT-001 in diabetic patients. The phase one and phase two portions were in the same well-controlled diabetic patient population that we're treating in ARISE-HF. So the same criteria of well-controlled glucose, well-controlled hypertension. Patients in our phase one, two study also had to have less than or equal to an HbA1c of 7.5, which is the criteria for ARISE-HF as well. And even when patients with diabetes are quite well controlled, I think we would all agree that an HbA1c of less than 7.5 is very tight glucose control. Those patients have twice as high levels of sorbitol in their blood as compared to healthy volunteers. So we did a lot of work on that at the beginning of our clinical development program, showing what a normal or a healthy person's level of sorbitol should be, showing that that doubles in a well-controlled diabetic patient. So the aldose reductase enzyme is quite sensitive. It does become activated with elevated glucose levels, even when that elevation is relatively moderate in a well-controlled diabetic patient population. And what we know from our phase 1/2 study is that when we treated those patients with AT-001, at the same dose that we're treating the high dose group with in ARISE-HF, we normalized their sorbitol levels down to that of healthy volunteers. So we have a very good understanding of the aldose reductase, higher enzyme activity that we see in a diabetic patient population, and how we shut off that enzyme activity and normalize sorbitol levels with AT-001 treatment. And of course, Dr. Januzzi spoke about how that correlates in animal models and in cells with a positive impact on cardiomyocytes. Okay, that was the first question. The next two- Actually, another quick thing to jump in, though. Yeah. With response to sorbitol, there was an impact on NT-proBNP, wasn't there? Absolutely. Mm-hmm. Yes. Yeah. So, I mean, I think that adds, I think, further evidence that through attenuation of sorbitol concentrations, and as a biomarker researcher, to me, that immediately makes me ask: Well, you know, can we measure sorbitol and use it as a, an indicator of disease activity? You know, that really adds a lot of encouragement. Sorry. No, I was actually going to hand it over to you for the second part of that question, which was about the distributions that we see and correlations at baseline. And that was the slide you spoke to. There's a quite tight correlation- Yeah ... all three of these endpoints at baseline. Yeah, with respect, at least to the NT-proBNP, it's performing the way we would expect to see in a heart failure population, which is it. If you look in individuals with overt heart failure, NT-proBNP is probably the best predictor of health status, KCCQ, activity levels, much better than ejection fraction, for example. So we see a very similar pattern in ARISE-HF, but it's encouraging to also see that NT-proBNP associates and correlates significantly with the PACE score, as well as with the results from the cardiopulmonary exercise test. So that, you know, again, lends further encouragement to the fact that we have the right patients in this study. Yeah. And that is actually the question that's now showing on the web. And I think that the slide that Jim is referring to, and all of these slides will be available to download after the webcast, actually show a statistically significant correlation between many of these endpoints. Yeah. So the question is really about change in VO2max. And Greg, it'll be interesting to hear your thoughts on to answer this question, but we haven't unblinded yet, so we don't know, you know, how the change in VO2max will stack up relative to change in NT-proBNP, because it's a really great question to ask is: How does one monitor the effects of AT-001 longitudinally? Greg, in a placebo-treated patient, one of the questions that was up there, that's been shifted off the screen, how much decline in VO2max do you expect over about a year, year and a half's time? Yes. Well, thank you for the question. I think really we're looking at that 6% difference, right? So whether that gets achieved by some improvement in the active arm or the mitigation of the decline, so this is a patient population that's sitting right on the edge of overt heart failure. And so those that develop heart failure, we can see different rates of progression over time in terms of their further decline in Peak VO2. We know in everybody, you know, across, you know, at a societal level, that the components of Peak VO2 do decline on a year-by-year basis. The peak heart rate achieved, for example, typically will fall by about 1 beat per minute per year in the general population and in individuals that are living with underlying cardiovascular disease. And so we expect these patients because we've essentially captured them right on the edge of heart failure, you know, this is stage B heart failure, that they are going to be having descent into more symptoms and more exercise intolerance. You know, honestly, we don't have enough kind of population-based data on diabetic cardiomyopathy to map the progression of decline in VO2 in this patient population, but we have identified individuals that are sitting at that 15.5 mL/kg/min, which is a very vulnerable point in terms of their functional reserve capacity at the baseline level and the risk of kind of descent into overt heart failure. ... Well, I can say from the study powering perspective that we assumed that the placebo group would decline, and we do have some very nice natural history in diabetic cardiomyopathy patients. That's thanks in part to a lot of work done by Tom Marwick, for example. So we know how the diabetic cardiomyopathy population will progress in an untreated setting and how they'll decline over time. Our assumptions for powering the study were that the placebo group would decline by a minimum of 6%, probably approximately 6% over the course of the study. Which, as Greg mentioned, with our baseline peak VO2, which we didn't know going into it, of about 15, is about 1 mL per kg per kg over time. And so our assumptions for powering were that treatment with AT-001 would halt or slow decline, and the placebo group would decline. What we do know based on the animal model and some of our phase one, two data is that we do have a possibility of upside of AT-001 actually improving peak VO2 based on improvements in cardiac energetics, and that would be additional upside to the study, but we had not anticipated any improvement in our powering assumptions. I see that there's a question here about commercial potential, leading into that a little bit, and talking about how our design of the study and identification of patients would transition into a potential commercial marketplace. So, I'll start off, and then I think Greg is probably best poised to answer the rest of that. We do know, based on the, the natural history of the disease and prevalence and incidence, and Greg flashed a few of... Sorry, Jim flashed a few of those studies on the slide. The studies show us that 17%-24% of the diabetic patient population has diabetic cardiomyopathy, so maybe we split that down the middle. We usually talk about it as approximately 20% of the diabetic patient population, which unfortunately, based on how many diabetic patients we have, in this country and globally, that's a huge number. So the commercial opportunity here is about 20% of the number of diabetic patients. And what we've tried to do with design of the ARISE-HF study is to ensure that we mirrored the patient population in the study, based on how they're diagnosed, based on how that would work in an actual treatment setting. So, I'm not sure if people know on the panel here, Dr. Januzzi is a cardiologist. Riccardo Perfetti, our Chief Medical Officer, is an endocrinologist by training and formerly ran a large diabetes center. We do anticipate the diabetic cardiomyopathy patients will be treated by two different prescribing physician groups. One is a cardiologist, like Dr. Januzzi, and another is the endocrinologist or diabetologist, or maybe sometimes even general practitioners, you know, as would be the case for Dr. Perfetti. And how they identify their patients and diagnose them might be a little bit different. So for example, the endocrinologist and diabetologist would probably prefer to run a simple blood test for NT-proBNP to proactively diagnose diabetic cardiomyopathy. Jim, are you more often running echoes or NT-proBNP? I do both. It's a good question. You know, I mean, this is a time of great change in the diabetes care space, and that's been fueled by the development of therapies that reduce cardiovascular risk, but also an increasing recognition in the diabetes community of the importance of heart failure. And, the American Diabetes Association has been particularly aggressive about recommending early diagnosis of heart failure, utilizing NT-proBNP as a first step, followed by imaging for people with abnormalities in imaging. But that doesn't mean we only use biomarkers. And indeed, you know, we will image people with echo if there's a clinical indication of something going on, changes on EKG, changes in functional capacity, other aspects that might indicate need for imaging. One of the challenges with NT-proBNP, although it's very useful for the identification of stage B heart failure in many populations, including those with diabetes, is the fact that it can be downwardly suppressed with the presence of obesity. And so one needs to be attuned to that fact, because in certain circumstances, imaging might be the better approach. Yeah. Yeah. Perhaps if I can add, in the practice of the endocrine clinic or the diabetes clinic, in addition to monitoring glucose, we monitor risk factors. So the fact that NT-proBNP measurement could become part of the routine of how we identify patient is very foreseeable. Yeah. Yeah. So that actually is the next three questions, all wrapped up into one: How do we identify these patients, and what's the treatment potential in the real-life setting? So I think that's probably all the questions that we have time for. Operator, I'll do one check on the line. Are there any more questions on the line? We do currently have a follow-up from Brian Skorney with Baird. Okay, Brian, one follow-up, and then we'll close it up. Go ahead. All right. Thanks. I know you have pre-specification on the concomitant use of GLP-1s, but I think there's a lot of concern out there in the market broadly about the paradigm of treating complications of type 2 diabetes and obesity with the emergence of Ozempic and Mounjaro. So maybe you could just help kind of characterize how DbCM is likely to evolve and more widely used? And do you think it's necessary to show an effect in that pre-specified subgroup of patients on GLP-1s? How does the commercial paradigm change if the delta is not there in that subgroup? And more specifically, do you even need to look at data on patients using Ozempic or Mounjaro specifically to kind of build out the commercial opportunity? Yes, that's a great question. So that will probably involve a combination of all of us jumping in. I can say, at least from a study design and from our perspective when we went into this trial ensuring that we thought forward, like, we knew that this was going to be a large and relatively long clinical trial. We didn't anticipate COVID, so it's about a year longer than we thought it would be. But we, we did anticipate that use of SGLT2s and GLP-1s would increase over time. And when we designed the study, we needed to be thinking not about use of these agents at the time we started the study, which was 2019, but how did we anticipate that this would look, you know, three or four years out, which is kind of where we are now. And so, you know, I think that there's a few important factors to think about here. SGLT2s and GLP-1s do play an important role in glucose modification and obesity. Although, you know, as we've discussed today, we do have a pretty good understanding of the mechanisms of disease leading to diabetic cardiomyopathy, and it primarily has to do with activation of aldose reductase, which happens whether or not a patient is obese. So I'm sure we'll all be very glad if use of, you know, these agents decreases obesity in our population, but we don't need to be worried about that effect altering the impact of our drug or, or our study design. Now, there have been these sort of nondescript, we don't really know the mechanism of action, kind of effects of, of GLP-1s and SGLT2s on generalized, heart failure, which is important. And that's why we ensured that we stratified across the groups for use of these agents. We didn't want there to be an imbalance in one group where, you know, everyone was having a generalized beneficial effect because of an SGLT2, but we don't really know how it works, but it might throw the study off. So, you know, we ensured that it was stratified that way. And then the last thing I'll say before I hand it over to everyone else is that we didn't show the results today, but we have presented them at medical conferences before. We did actually look at the diabetic cardiomyopathy mouse model, plus and minus and SGLT2 with our drug. And AT-001 impacts the diabetic cardiomyopathy mouse model in all the ways that Jim showed you, cardiac energetics, structural abnormalities, as well as fibrosis. This happened regardless of concomitant use of an SGLT2. So we really have shown that these are very distinct mechanisms of action. SGLT2s and GLP-1s are beneficial in a lot of other ways for diabetic patients, but not impacting diabetic cardiomyopathy. But Jim, you see this every day. Yeah. Yeah, sure do. So great question, Brian. And also, there's a question about SGLT2s here online. What I would say is that, first, from a trial design perspective, you know, I congratulate Applied Therapeutics for thinking about this and stratifying. I'm doing a number of trials in the diabetes space where we're not stratifying, and the argument for stratification, you've already heard. The argument against stratification by some trialists is that although, yes, GLP-1s are on the rise, and although, yes, SGLT2 inhibitors are important and on the rise, they are still used in a very small percentage of the overall populations of the patients that we're studying. And so you can see the argument in both directions, but from a pharmacoeconomic perspective, from a perspective of Applied Therapeutics, thinking ahead, I think it was a brilliant move to stratify. Whether we'll be sufficiently powered will be a different question, but we can at least inform, you know, patterns of response in these different populations. Although the basic science data would suggest that they're going to be, if anything, additive, they're not gonna cancel each other out. To that extent, I did have a slide, Brian, where I went through the various pathways that are thought to be involved in diabetic cardiomyopathy. It's kind of a wonky biology slide, but the take-home message is that these therapies, the therapeutic targets here are likely orthogonal. In other words, one therapy is gonna be additive to the other because they do different things. And so if GLP-1s, for example, reduce lipotoxicity through reduction in myocardial steatosis, that's probably a different pathway of benefit than the way aldose reductase inhibition is going to exert its benefits. It'll be an interesting question to address, but, you know, the preclinical data really argue that this shouldn't interfere with the effects of the drug. Yeah. And maybe just one last point, and then we'll close the call. Brian, I think you were asking about, do we need to demonstrate an effect of the drug within that subgroup analysis, which is pre-specified as one of the secondaries of the ARISE-HF trial? I don't think that we do. Again, our focus was really just ensuring that we stratified the trial appropriately. To Jim's point, it, although we're hearing about this sort of blowout use of GLP-1s as than SGLT2s, it's a much smaller percent of the population than, you know, I think people really realize. We powered the study to ensure that we had less than one third of the patients on SGLT2s or GLP-1s, which we've well achieved. And but you do have to remember that it... When we do a subgroup analysis, that will be less than one third of the patient population. So to Jim's point, we've not powered for that. It's not the primary endpoint of the study. We will have some interesting data on that subgroup analysis, and we think that there shouldn't be any lessening of effect of AT-001 there. However, it was something that we thought about very carefully, and I think we've achieved what we wanted to in design of the trial. And that's probably a good way to close this. We're all very excited to see the results. They're coming soon. Hopefully, our takeaway here is that we have a good understanding of the disease, we've designed the trial well, we have the right patients. The baseline criteria are basically as good as we could have expected here, and the blinded analyses are pointing in the right direction. We really look forward to sharing the unblinded data with everyone soon. We thank you for joining us today. Thank you.