The majority of rare diseases have an underlying genetic cause (Wright et al., 2018). Accurate and timely diagnoses are necessary to provide appropriate guidance and treatment, but are complicated by clinical heterogeneity and information scarcity that slow the progress of retrospective natural history research. Though highly beneficial, many rare genetic disorders lack a dedicated database of known variants and associated clinical data.
In this webinar, Genomenon co-founder and chief science officer Dr. Mark Kiel is joined by Dr. Gus Khursigara, VP of Medical Affairs and Clinical Strategy at Inozyme Pharma to discuss how a computational approach to curation of medical evidence has emerged as an efficient and economical method to systematically identify and analyze all reported cases of a rare disease and interpret all associated genetic variants. The resulting locus-specific variant databases and disease-specific patient databases act as retrospective natural history studies and clarify the clinical relevance of individual variants.
In particular, our speakers explored how a comprehensive, locus-specific patient database (or Patient Landscape) for ENPP1 deficiency is providing Inozyme with an important tool to better understand and target this condition.
You will learn how:
- Clinical heterogeneity and information scarcity complicate the study of rare disease
- AI and machine learning-assisted technology is streamlining medical evidence curation for rare disease
- Comprehensive locus-specific patient databases can guide rare disease diagnosis and treatment
Gus Khursigara, PhD
Medical Affairs and Clinical Strategy
Mark Kiel, MD, PhD
Co-founder and Chief Science Officer
Hello, everyone, and welcome to the webinar! I’m Kate Oesterle, and I’ll be your host today. Thank you so much for joining us for today’s live event. We’ll be discussing a computational approach for locus-specific databases for rare diseases. We have a lot of great information to share with you, so I’m going to get right to housekeeping and our introductions.
As you’re watching today’s event, please feel free to submit your questions in the go-to dropdown window. If time allows, we’ll get to those at the end of the presentations. This webinar is being recorded, and it will be emailed to you later today. I encourage you to keep an eye out for that email and share it with your contacts and colleagues who find this information valuable.
Now, I have the pleasure of introducing today’s speakers! I’d like to welcome Dr. Mark Kiel. He’s the Chief Scientific Officer at Genomenon, where he oversees the company’s scientific direction and product development. Mark received his MD/PhD in clinical pathology at the University of Michigan, and he founded Genomenon to address the challenge of connecting pharma researchers with evidence in the literature to help diagnose and treat patients with rare genetic diseases and cancer.
We’re pleased to also welcome Dr. Gus Khursigara, the Vice President of medical affairs and clinical strategy at Inozyme Pharma. Gus received his PhD in molecular neuroscience from the Weill Cornell Graduate School of Medical Sciences. He has almost 20 years of experience in the pharmaceutical industry, with his last 12 years being focused on bringing treatments to patients with rare and under-treated diseases. I’m looking forward to today’s discussion, so Gus, would you like to get us started?
GUS: Thank you, Kate, and thank you all for attending! I appreciate the opportunity to give a perspective from an early-stage company on the value of having a partnership with Genomenon, and how they’re going to help us with some of our objectives and our mission at Inozyme Pharma. For those that don’t know Inozyme, it’s sort of a classic startup. It’s five years old. We did an IPO not too long ago, and we are interested, as you see, in the unmet medical need for diseases with abnormal mineralization. Of course, we have an asset that we’ve been working on: an enzyme replacement therapy to address some abnormal mineralization. When you go to the next slide, just to give you a bit of background on the pathology, the clinical spectrum, and then what our objectives are — it just wants to set it up to show how we’re going to leverage our relationship with Genomenon.
To the left is a normal setup for mineralization. You see in the green in the bottom left there, ABCC6, it it takes ATP and accelerates it from intracellular to extracellular. Then the ENPP1 will hydrolyze ATP to pyrophosphate, or PPi, and to AMP. AMP then is broken down to adenosine. Now, the value here is, pyrophosphate, or PPi, is a very highly-used regulator for mineralization. Calcium and phosphate, the building blocks for mineralization, are necessary for bone. Too much of it, bones are brittle. Too little of it, bones are soft. Because it’s ubiquitous, calcium and phosphate, as you can imagine, are also circulating the body, so PPi becomes also another important player to try to minimize calcification in the vessels and tissue. Adenosine also plays a role in arterial stiffness or arterial thickness.
Now, to the right, if you have a deficiency in ENPP1, as with example 2 in the middle, you are not creating or producing PPi, pyrophosphate, and you’re not producing adenosine. The consequence from the physiological perspective is calcification of arteries and thickness of the arterial wall. Thickness can lead to stiffness, that can lead to stenosis. So how does this manifest in patients that do have a deficiency in ENPP1? What you’ll see, as we move to the next slide, is that this can present itself across a broad age range, and again, is not so different from other inherited enzyme diseases. In the case for ENPP1 deficiency, they can be identified as generalized arterial calcification of infancy, or GACI. When born, they have tremendous, extensive calcification of the arteries. You tend to see a lot of it in the aorta. It has a high mortality rate, as you see on the slide to the left, of 50% within the first six months of life. You can also see at the bottom, some x-rays of calcification around the heart and artery, you can see some stenosis from an echo. Even with survivors of GACI, the other 50%, they eventually move on to develop skeletal complications, cardiovascular complications, neurological complications. There is also a group of patients that are identified with hypophosphatemic rickets that did not present with GACI or were identified with arterial calcification. They too also present with rickets, due to the deficiency in ENPP1 enzyme. Again, as you can imagine, similarly to other diseases that are skeletal in nature, you can see that as they get older, this has also an impact on mobility, on pain, and just overall risk of heart valve defect.
So, this is why I bring it up for this discussion, it does highlight the pathology. It’s heterogeneous, these patients present at any age, that means they touch a lot of specialists. As it turns out, right now, this disease is considered very rare. You can see the challenges that we are facing when we talk about trying to bring therapies to patients. On the next slide, very quickly, we do have a product: an enzyme replacement therapy with ENPP1, the extracellular component. Just quickly, you can see in an animal model, in red to the left, is your ENPP1, and a knockout animal and that activity. You can see the next three points, as we increase the dose of the concentration, the ENPP1 activity is increased. What you see to the right, the output, the PPi as we highlighted earlier. In red there, the vehicle, there’s no PPi in the knockout, and you can see that there’s an increase of pyrophosphate when we introduce the enzyme. We’ve shown that we can recover failure to thrive, we can reduce vascular calcification, we see an increase in survival. With that IND, we had put in and had received approval to move into our first in-human study.
On the next slide is your typical multi-ascending dose safety study. We just announced on Monday, the first patient has been dosed, so we are excited to be moving from a pre-clinical stage to a clinical stage! This is typical, and not to get into the details, but we’ll have a cohort with a low dose. If the safety looks great, then we look at the PD, then we’ll move it to a second cohort with the second higher dose, the same principle. If it’s safe, we’ll move it to a third. Here, we’ll hopefully get safety as well as dosing parameters for a drug to move into further trials.
Even though we’re doing our first in-human study, as of this past week, there are still things that we are building within the community itself. The activities typical for rare disease launch — one of the big challenges is increasing awareness. Usually, there’s very few patients, and there’s very limited publications around them because of the few patients. So, how do you increase awareness for disease that hasn’t been well publicized? Number two, how do you accelerate diagnosis? As I highlighted in the slide earlier, the age and the spectrum and the heterogeneity of the disease is so broad, it can be misdiagnosed. Therefore, how do you get physicians or clinicians to think about a rare condition and accelerate the diagnosis? Three, we’re trying to understand the prognosis. Early in rare diseases, as some of you might know, there are few publications, few case studies, few clinician experiences. Trying to understand how the disease progresses longitudinally is usually very difficult, or at least, there’s not a lot of information available early launch. We want to determine the prevalence and we want to determine treatment response, to standard therapy as well as, obviously, to our potential drug of interest.
To the right, I’ll talk about why we started to consider working in partnership with Genomenon. One of the ways to overcome and appreciate the approach that we have is trying to get one source where the literature and the genetic information stands. We’ve talked to a lot of physicians that try to do genetic testing. They get the variant back, they’re not sure where to look it up, or it takes a lot of time and dedication to find literature that’s scant. We feel that this, having one source, will actually be very helpful for clinicians, and actually help increase awareness. Mark will talk a little bit more on the specifics around that. We want to determine prevalence, trying to identify and expand the variants of significance will help us determine the prevalence of the disease, based on penetration, incidence, and prevalence. We want to accelerate genotype/phenotype, is there something prognostic about certain variants that will tell us the patient will accelerate or not accelerate to a poor outcome? We want to expand genetic confirmation. Many of these rare diseases, as you know, don’t have disease-specific symptoms. Take rickets, for example, hypophosphatemic rickets can be due to a lot of reasons. Pathology just lands on rickets. So we are trying to expand genetic confirmation to try to accelerate diagnosis, as well as expand our potential patients that can benefit from the treatment that we’re trying to develop. Finally, we prepare: are there any patients that don’t respond to our treatment? Is there a genetic underlying reason for that? Then, can we try to identify that sooner rather than later?
So I think those make up the approach we’re taking, to the left, and how we think the partnership and what Genomenon is doing, to the right, is how they’re going to help work with us to increase the awareness, the diagnosis, and hopefully better outcomes for patients with ENPP1 deficiency. I’ll leave it over to Mark to sort of talk through a little bit more about what they do, specifically in terms of this challenge.
MARK: Thank you so much, Gus, for that introduction! Gus will be back when I’m done with my introduction about what Genomenon does, and a bit of more detail on what Inozyme and Genomenon have done together with respect to ENPP1 deficiency. As Kate alluded to at the beginning of the broadcast, if you have any questions, I invite you to enter them into the chat, and we’ll perhaps have time to go over those with Gus when he returns.
For my part, I want to share with you my experience in working with rare disease pharmaceutical companies like Inozyme, and the work that we’ve been doing with Gus, lending a little bit of the insights that I’ve learned through several years of effort. I’d like to begin with some of the things that Gus talked about with specific relevance to ENPP1 deficiency by highlighting this as a problem across multiple rare diseases. In particular, rare disease diagnosis is challenged by a number of characteristics. The early identification and diagnosis and treatment of these patients can dramatically improve outcomes, as Gus alluded to.
For rare disease, most rare diseases, 70 to 80 percent, have a genetic component. Most of those individuals with rare disease remain undiagnosed, whether throughout their life, they remain undiagnosed, or there’s a great delay in receiving a confirmed diagnosis. That’s referred to clinically as a diagnostic odyssey. Statistics suggest that, for rare disease patients, on average, this diagnostic odyssey can last 4.8 years, and can result in seeing seven or eight specialists, as Gus suggested for ENPP1 deficiency, before they finally obtain the correct diagnosis.
The other trend that we’re seeing is that more and more patients are being sequenced with the advent of next generation sequencing technology and things like newborn screening expansion through sequencing, which is offering a great deal of promise for expediting these diagnoses for rare disease patients, improving outcome and mitigating this diagnostic odyssey. The challenge there is that many of the genetic variants that result from sequencing are poorly characterized, and that puts a limit on the value that sequencing can provide.
To improve rare disease diagnosis, we focus on capturing existing knowledge about the disease genetics, as I’ll speak to in subsequent slides here, and then ensuring optimal engagement with that information. I like to say, weaponizing that knowledge, not just keeping it precious and in a silo, but making sure that it gets out into the world on the front lines, where the clinicians are seeing these patients to diagnose and treat them. When you put these two things together, we like to refer to the comprehensive organization of this evidence as a genomic landscape, which is to say, every genetic variant and all the supporting evidence necessary to adjudicate its meaningfulness clinically is aggregated, and then integrated into clinical practice. I’ll speak to some of those specifics with the Inozyme/Genomenon partnership that Gus has alluded to.
Genomenon’s mission is to make genetic and genomic evidence actionable. We have a particular interest in rare disease, because there’s a great need in making this information actionable and getting it in the hands of clinicians in the service of better designing and better optimizing drug development and drug delivery. That’s our mission. Our goal subserving that mission is to curate the entire human genome, doing so one gene or one disease or one set of diseases at a time. We are a piece of the overall puzzle, a very significant piece, which is to say, the empirical evidence in the published literature that’s uniquely beneficial for clinicians to render their diagnoses, and uniquely challenging to organize and annotate and adjudicate that evidence. There’s a problem here in that extracting this meaningful evidence, especially for rare disease, from the published literature requires or hithertofore had required manual extraction and navigation of the complexities of genetics and genomics and idiosyncratic clinical phenotypes and presentations.
What Genomenon has done is automated that process. With sophisticated computational techniques, we take the published literature, identify the relevant aspects of that literature, the relevant articles, and extract the meaningful genomic concepts out of those articles. We then organize those into a network of genomic associations. What results from that process is the Mastermind database of genomic associations, which comprises all diseases, all genes, and all genetic variants in those genes, as well as the attendant clinical and functional significance of those variants found in any one of the patients that had been published. This is a clinical use tool that I’ll speak to here in a moment, but it is comprehensive across the entirety of human disease, the full breadth of the human genome, as well as every variant. No matter how poorly described a variant is by the authors, the Mastermind capability allows us to comprehensively ingest and annotate and organize that evidence.
We go further beyond just the computational organization of that evidence for maximal sensitivity: we also have internal capabilities that allow us to curate that information for maximal specificity. We have a curation engine capability that allows my team of curators to adjudicate that evidence, fit for purpose (disease-specific or drug-specific) and present that evidence back into Mastermind, which is what Inozyme and Genomenon have been working toward in our partnership. I’ll show you the fruits of that work here in a moment.
That computational approach to annotating and aggregating the evidence for maximal sensitivity, and the specific approach using our internal capability for maximal specificity by human curation, allows us to produce a comprehensive genomic variant landscape, in this case, for ENPP1, but we’re able to do this across the whole human genome. For every genetic variant and any rare disease or cancer or other disease, those results are interpreted to the clinical standards, annotated for clinical and functional significance, and all supported by the evidence. In addition to that, we are able to produce a comprehensive patient landscape. That is to say, every published patient for any rare disease, in this case, ENPP1 deficiency, all of whom are characterized by their demographics, the age and gender and ethnicity, the clinical phenotypes that attend their clinical presentation, the laboratory findings, the treatments that they received, and the outcomes of those treatments. Collectively, that information comprises a natural history study.
Those in the audience may be aware that there are two basic flavors of natural history study: one, looking back, in aggregated data from the past and retrospective analyses, and the other, looking forward, where you start with nothing and you start to accrete patients, either in a cross-sectional study or in a longitudinal study. There’s no reason for these two versions of natural history studies to be at odds. In fact, they’re very complementary and very synergistic. This table summarizes that synergy. Suffice it to say that the retrospective approach allows you to gain keen functional insight into the consequences of these genetic variants, as well as expanding your cohort size, particularly by capturing more of an ethnic representation of the disease, whereas with a prospective study, you can a priori define your own collected parameters. By virtue of the longitudinal understanding of the patient’s course, you can get deeper and more specific progression information about your patients. Again, both of those benefits blend together in a unique synergy that makes for a very robust and unique natural history study. That’s what we’ve done at Genomenon with with Gus and Inozyme.
I just want to walk you through some of those data points and some of the details before I bring Gus back, and we’ll have a conversation about the challenges attending rare disease drug development and delivery. This is a summary of the source of the patients from the retrospective and the prospective natural history studies that I just alluded to for ENPP1 deficiency. In blue there is resulting from our comprehensive patient landscape from the survey of the medical literature. You can see a large number of patients, including 66 that were unique to the retrospective study. Then, we actually had, with our collaborators from the NIH and the University of Muenster, we had two separate natural history studies that were prospective. They contributed their unique numbers of patients to this consideration as well. I’m highlighting in this venn diagram not how these differ, but how they blend together, and how there’s a unique value in combining these two approaches.
You see the fruits of that unique value when you actually look at the variants that you uncover, when you have a wide breadth of understanding from multiple different sources to aggregate this patient data. Particularly, I’m showing here the unique ENPP1 variants, obviously that are the cause of ENPP1 deficiency, GACI, and ARHR2. Again, the venn diagram is meant to show that there’s complementarity here, as opposed to pitting these two resources, or multiple resources against each other. We feel strongly that combining these approaches provides unique value and insight.
I’m sorry that this is a bit detailed, I feel it’s a bit obligatory, but this is the ENPP1 protein. The diagram on the bottom is of the functional domains. Each of those dots or bars is a reflection of where those variants were found across all of the patients in our aggregated natural history study. I’ll emphasize, of course, that not only are these dots, they’re patients. They’re reflective of patients who stand to benefit from early diagnosis and intervention. All of these variants for each of these patients have been adjudicated according to clinical standards, based on a comprehensive reflection of all of the evidence. So it’s necessary and sufficient to ensure clinical appropriateness for diagnosis and treatment.
For the last couple slides here, before I bring Gus back on, I want to make this actionable. As I said earlier, that’s all well and good, we have a better understanding of the disease and how genetic variants cause disease. But Gus and I both are motivated to make this actionable. The way that Genomenon is able to do that is by leveraging our reach into clinical laboratories in the service of connecting the patients who are diagnosed by the clinicians, with the drug or the trial, through the data, through the evidence. For that purpose, we have the Mastermind clinical software tool that I alluded to earlier, used in hundreds of diagnostics labs across the globe.
This is what those clinicians see. Again, it’s the evidence that they see, in the case of our partnership with Inozyme, in addition to seeing the evidence, they are able to see right away that the patient and the variant in their clinical workflow is appropriate for enrollment in the trial or prescribing of that drug if it’s post-commercial. What I’m showcasing here is the clinician’s eye view. When they have a GACI patient who’s undergone sequencing, they can immediately see, in our clinical software, what that variant means, the fact that it causes disease, and the fact that there’s a trial that that patient is amenable for enrollment in.
So that’s, as I say, a whirlwind tour of Genomenon’s philosophy, the data composition that we have, and the way that we make value out of that data, and in particular, how the partnership with Inozyme and Genomenon is starting to bear fruit. With that, I will bring Gus back on. Hey, Gus. I’ll also invite any of our audience members to weigh in with their own questions. I have a series of my own. Gus, you and I have been working together for some time now, so we can forget what we both think we know about each other.
MARK: Let’s just start with an open-ended question: What are some unique challenges for pharma companies focused on rare disease? We can go in any direction that you want to take that.
GUS: Sure. I’ll say this, the experience in the industry is getting greater and greater. There are more smaller companies that are trying to address rare diseases, so what used to be considered unique maybe 10 years ago, I think is becoming standard. What is that? I think it all starts with just the definition. It’s a rare disease, so when you get into a disease early-stage, there are few patients, that means few clinicians, and that means there’s not a lot of information and awareness that’s available. The challenges of getting into a rare disease and launching a product in there is, how do you overcome those few patients, few clinicians? That means few publications, that means there’s uncertainty where the patients are. How do you overcome that? That’s the general challenge, I would say, Mark. A general response to what are some of the unique challenges. Now, we can break it down, if you wish, into different components. If there are too few patients, how do you find patients? If there’s too little experience, how do you broaden experience? We can go down that road as well, but I think that’s the underlying theme for any of these rare diseases. Few patients, little clinical experience, how do you start to build it up?
MARK: Let’s pull some of those threads. One thing that you alluded to on your last slide, I think on the left, one of the things that you at Inozyme are deeply engaged with now is the clinician education, which I feel like is a major challenge to mitigate the diagnostic odyssey and actually get the patient sequenced upfront. Can you speak to some of the efforts at Inozyme for augmenting and accelerating clinician education in the service of diagnosing these patients earlier?
GUS: Yeah, it’s the biggest challenge, and it’s a challenge that I think will stay with us. If we can develop the products, if we get approval after we get payers to purchase and we get into the commercial space, we will constantly be trying to educate physicians. Early on, the approach is this: there are a few physicians, usually fewer than a handful, that have published or that have a very keen interest in some of these diseases. It’s to start there and learn from them, and it’s a very good collaboration to start with those physicians, because you get to work together. As drug developers, you start to understand, or try to understand, what are some of the end points that are important for a clinical development program? Clinicians will tell you what’s practical and what’s obvious. Now, from that perspective, you start to grow. You start to look for avenues of, how do you target certain specialists? How are these patients coming to the clinic in the first place? Who’s looking at them, who’s diagnosing, how are they managed? In our case, and I think in many other rare diseases, there’s multiple specialists that see this patient or can touch this patient. We have to start to calculate, are we looking for people that are diagnosing? Are we looking for physicians that are managing? What are the key conferences that we go to? Where do we put our announcements? It’s not one specific path, Mark. I think in the end, we start there, and we try to build the experience.
Now, something like having a database like yours is interesting to us, because the first question that will come up is, how do I know it’s really a GACI, or a rickets type, or ENPP1? As you know, we are now also looking to accelerate diagnosis by saying, you should go and genetically test this patient profile. When they go and test, and you probably know more than I do, clinicians will get that variant back and just not understand what that means, right? Some physicians are okay, they see the phenotype, they see the genotype and say, okay, I think this is ENPP1 deficiency. Others might say, I’m not sure what that means. Having a data set like this early, even though we’re in our first in-human studies, is getting physicians to understand, to genetically test. Then, when we drive them here, as you highlighted, the literature is highlighted. Patient advocacy could be highlighted. Clinical trials could be highlighted. It gives a single source for that clinician to get more positive. I think I rattled on a few little points there, Mark, but hopefully that sort of sets up the stage for some of the other questions we might get, as well.
MARK: Yeah. Just a little bit more detail about the gap between recognizing there’s a challenge and pulling the trigger on a sequencing effort — I remember, during my training a decade or more ago now, there was uncertainty about the value of sequencing, because if you can’t take action on it, or if there’s an interpretation challenge, why bother? was the sentiment. I think it’s now two decades, actually, since training. Are you seeing a continued trepidation about casting a wide net, and doing, say, exome sequencing, or even large gene panel sequencing? Or are you seeing that barrier break down with the recognition that there’s real value to come out of the sequencing, especially when you fully understand what the results may mean?
GUS: That’s a great question. I think I’m still a little bit more toward your experience. My impression, what I’ve seen is that clinicians will diagnose to a treatment. There are very few physicians that will continue to try to keep pursuing, what is the rationale for the pathology? But, in general, and that’s not a judgment, physicians treat to a treatment. Now, why would they look for GACI type of patients if the treatment is just going to be the same in terms of phosphate and calcitriol, etc.? When you have a company, a pharma company, that’s now looking for a therapeutic, it does create some motivation to further diagnose that patient, because you can tend code down to sequence. Can they be recruited into a clinical trial? Would they be available for the therapy down the line? Again, I think that moves from diagnosis to treatment. In a rare disease, there are many diseases that don’t have disease-specific symptoms, they all look alike. Again, I’ll focus on rickets, since we’re on this topic. That’s where the genetics comes in. We will now go out and talk to physicians and talk about the prognosis, and how this phenotype can look similar to other phenotypes. Genetics become important because, when we’re recruiting too, we also now understand that some patients will not benefit from other treatments if they have our mutation. That’s what I think inspires physicians/clinicians to move toward genetics and sequencing. Then, as we hit upon, are these physicians comfortable once they get the genetic sequence back? What did they do with it? They have to be inspired.
MARK: One other thing that you touched on in that last slide that you presented was that this is a rare disease, and that makes it challenging to estimate the number of patients that you can successfully treat. I wonder if you can speak to some of your early challenges in estimating the prevalence, understanding, perhaps, geographic distribution of these patients, and how that informed your strategy thinking about developing the trials or thinking about post-commercial phases, even, looking forward.
GUS: Mark, I have to tell you, that’s part of the reason why we did a partnership with you. That is exactly the information we need to try to figure out, to be honest. So what have we done? We’ve gone to different institutions and we’ve done some simple marketing. We’ve tried to understand, through ICD-9 codes, similar phenotypes, and maybe they’re captured in there. I think some experts have tried to look at some of the genetics, and are using what we know from the number of mutations, looking at penetrance, and therefore trying to size. What spurred us on is to try to understand from you. What I think we’re getting in our results is, we’ve been able to identify more and more variants. As we continue to identify more variants, either through the literature or unpublished, but yet, physicians continuing to add to the database. We are starting to see a broader number of patients. Our approach has been a little bit manual. I would say it’s probably not so successful, but it’s given us a perspective or a landmark to start with.
I would think in the next phase, and why we’re doing this so early with you, is we want to encourage physicians to use and add their variants and their interest into the database so we can get a better prevalence. As we move forward with regulatory agencies, and when we really look forward toward payers, we need to understand what this potential patient population looks like. It sets up how we are going to strategically roll out, whether it’s due to pricing, whether it’s due to, as you say, geographical areas. There are certain countries we may or may not go to, so we’re not there, Mark. That’s exactly what we’re trying to spend the next two, three years building up this part of the database.
MARK: A particularly unique challenge in rare disease, and in conversations with all other groups, is that it’s hard to know, what is the right answer? I don’t feel like there is a right answer. It’s such a big question. You have to rely on assumptions and estimates, and the quality and care with which you make those assumptions and apply them, I think that’s really where you triangulate around the actual truth. It starts with really understanding the disease, as you said. You talked about the clinical heterogeneity and the diagnostic challenges absent genetics, but even within genetics, what we’ve been working on is clarifying and minimizing that uncertainty. Particularly for GACI, you had that nice progression of GACI and rickets and osteomalacia across the different age groups. With specific relevance to GACI, can you speak to the phenotypic presentation and the timing and the acuity of the diagnosis for earlier intervention, leading to improved outcomes?
GUS: Yeah, I’m happy to! Really, this ENPP1 deficiency was initiated, or the gene was identified, because of the GACI phenotype, generalized arterial calcification in infants. You can see, even prenatally on ultrasounds, some calcification that seems to be occurring in key arteries, and around the heart, quite frankly. That’s what started identifying this phenotype. Because of the calcification of the heart and artery, you get extensive calcification. There are many of these patients that actually die prematurely before birth, and others, we see even after birth, within days and weeks. I think I said up to six months, but really, even earlier than that, you see that these patients will die from either of the cardiopulmonary issues. Even looking at the genotypes that have been attempted, are there particular mutations that are prognostic to patients, or that that will die in patients that don’t? We haven’t seen that so far. There’s no predictive risk factor, at this point, that can be identified. That is one of the things that has always been of interest to many of the clinicians. Can we predict that? We can’t so far.
Now, we don’t know why some survive and some don’t. There’s some speculation that it’s because of the change in the metabolism, and there’s a difference in phosphate to pyrophosphate ratio, but the body then starts to elevate FGF23. What FGF23 does is force phosphate wasting. Patients start to pee out phosphate. Because of that, since phosphate is part of that calcium building block, this might be part of the reason why we see some calcification resolve in some of these patients. As they grow, what hadn’t been well documented until we got involved was that they start to develop rickets, because, guess what, they’re wasting phosphate, and now they’re under-mineralizing. It’s a bit of a contrast between extensive calcification and under-mineralization in these patients. That phenotype early on for GACI is quite obvious and avert. That hasn’t been an issue. Understanding why some patients survive and some don’t has been the challenge. Once they survive, it looks like they all move into that hypophosphatemic rickets phenotype. I will say, whether it’s FGF23, whether it’s a PHEX mutation, whether it’s an ENPP1 mutation, the phenotype looks very similar. That’s where the genetic confirmation becomes more critical to us, both from a diagnostic perspective, but also as we try to identify patients and recruit patients for any future trials. Hopefully, that gave a little bit of an answer there.
MARK: To build on that, you mentioned PHEX, and we actually have an audience question that asks more generally how many diseases have more than one genetic anomaly. You talk about PHEX with respect to the clinical overlap in phenotypes. On, I think, your second or third slide, you mentioned ABCC6 being part of this ENPP1 pathway, and so where I know that ABCC6 mutations do not often cause GACI, they can. Can you speak to the genetic complexity of the this constellation of genetic defects, and how the value of the data that we produce can start to penetrate that complex?
GUS: Yeah, I’m going to go in reverse. That’s what we need, right, Mark, we need this type of data to address that type of question. It’s a good question, especially since I hit on FGF23. What we’ve seen, or what some clinicians have told us, is that they will see patients that come in, and they diagnose them as XLH. XLH is hypophosphatemic rickets due to a PHEX mutation. When they go through genetic testing, in some cases, payers are asking for genetic confirmation that it is truly a PHEX mutation, to be given treatment. It turns out that they don’t actually have a PHEX mutation, they have an ENPP1 mutation. Why is that? Because, as I highlighted, ENPP1 deficiency elevates FGF23, PHEX mutation elevates FGF23, so it’s not a surprise you get to see the same phenotype.
What is even more interesting is, in that pathway I showed you, there’s an ABCC6 that translocates intracellular/extracellular ATP. ENPP1 breaks down PPi, AMP. You have a CD73, which I didn’t highlight, that breaks the AMP and adenosine, but there’s also then ALP that breaks down PPi. So PPi and adenosine can be managed by other enzymes. Are the mutations there? Is some deficiency there? There are some patients we’ve seen, a small amount in this case, that have both an ENPP1 and an ABCC6 mutation, just found from a personal conversation and no other evidence. We’ve seen there are some PHEX mutations that look like they also carry an ENPP1 mutation, so this is where this type of database will be helpful. It’s not from a competition, and this is what’s so interesting, it’s not a competition to try to understand what’s causing the phenotype. I think what is valued here is dissecting out what is the dominant gene or the dominant loss of function that’s leading the phenotype. Any company that’s working to replace some of these enzymes will want value and response from your drug. You don’t want to introduce a therapeutic if it’s not addressing your gene. So I don’t have a good answer for that. What I’d say is, I think we recognize there’s some overlap. Does it make a difference? Is ENPP1 more or less important than PHEX mutation? We don’t know.
MARK: I like the way you frame that. In addition to not wanting to needlessly treat a patient with a treatment that will be inefficacious, if you’re in the context of a trial, you want to be sure that you optimize the chance for successful outcomes for that trial. Clearly, understanding what the genetics are, underpinning that individual’s pathogenesis, and obviously, understanding your drug mechanism, making sure that there’s a match. I can actually address, from my experience as a molecular diagnostician, this is true of cancer. Cancer in the 60s and 70s, the war on cancer was like this single enemy. What we learned in that war was that there’s dozens or hundreds of different enemies with each of their own idiosyncrasies. Even during my training in hematopathology, there was a parsing of the diseases that looked the same, but the underlying genetics were different. This is also true in neurodegenerative disease, with ALS. It’s particularly important, as precision medicine comes to the fore, whether it’s enzyme replacement therapy or antisense oligonucleotide technology, it drives home the point that really clarifying the aboriginal cause of these diseases at the genetic level is going to be really important in most of these circumstances. Another question that I had while you were speaking is, as I alluded to earlier, the clinical reach of Mastermind is across thousands of diagnostics labs across the globe. Can you tell me what it means for you to be able to find one of these rare disease patients? I imagine that there’s a bit of euphoria when you’re able actually to identify one of these patients, because you feel that you can have a positive impact on their lives.
GUS: It goes both ways. One, it’s a positive impact. Sometimes, we find clinicians that are treating patients or families of patients that just desperately want to understand why they’re not responding to current therapy. Physicians also want to get the right therapy as well. Sometimes, just identifying why a therapy is not working, because they have a mutation in the ENPP1, it’s very joyful, I’ll say that first. We get excited. As anybody in a rare disease knows, every patient does count. It’s not a saying, it’s just the way you have to live, especially when you’re in a stage where we are, for example, at Inozyme, because we are still trying to learn about the progression of the disease. We are still trying to learn how they were identified. So every time we see a patient, it is euphoric, to your point, both from just a family perspective as well as from a scientific perspective, and then maybe at some point, from a research perspective, from a clinical development.
I didn’t hit upon one of your questions earlier. As we’re going to physicians, the other thing we are trying to do is go to diagnostic labs. Do they include a genetic panel that includes ENPP1? Some did and some didn’t. As we continue to build evidence and the value of why they should offer it, we’re seeing more and more labs pick it up. I don’t think that’s unique to us, but I think as labs get understanding on value, they’re going to add more to their genetic panels. Again, this is where I think genes will play a role. Our hope is to see that anybody with hypophosphatemic rickets, when they’re getting a genetic screen, it will include all the genes that we’ve been highlighting, and then some. You’re not going in to talk about one disease, but you’re going in to talk about a panel, and what is the cause, and what’s the direct cause of that. That’s what our hope is, working with diagnostic labs to include ENPP1. Then, Mark, that’s one of the things that we think through. As these diagnostic labs are identifying genetic variants, it’s how we’re going to make sure they get captured and pulled into Mastermind, so we continue to grow our own understanding of genetics within our own diseases.
MARK: Pursuant to a question, again from an audience member, but also harkening back to some of my slides, I wonder if — I’ll remind you what we did here, we merged the retrospective with two different prospective studies, and produced in aggregate a larger and more insightful data set. The audience member asks about how these clinical complexities and genetic complexities could potentially confound that merger, and I’ll just emphasize, that’s exactly the painstaking approach that we took, was to be sure that we did not confound the analysis. We understood very clearly what the genetics was for each patient, and very clearly what the clinical features were, not just what the diagnosis was, but what the attendant phenotypes were, the age of onset, the different diagnostic modalities, clinical lab values, etc. Data is messy, especially clinical data, and you have to start from the cleanest, most comprehensive dataset that you can. Then, when you marry the retrospective with the prospective, which were treated similarly, can you speak to the value of synergizing those two data sets, particularly for rare disease?
GUS: Yes, and just to be clear, the prospective that we’re highlighting in our terminology here isn’t from a classic prospective design study, where we went ahead and did a study. “Prospectively” here is, at the NIH, they prospectively collect data over time, because again, they have a physician there that is very interested in this disease and how it progresses. So from that perspective, you can call it prospective, but really, this was a chart review. Now it was really good from both Muenster from Europe as well as the NIH in the U.S. Because it’s a rare disease, these two work together. The data that was collected was agreed upon in terms of a similar protocol, if you will. It’s a chart review, just to be clear. They both agreed to capture the same information because they wanted to aggregate and increase the patient numbers. That became very helpful for obvious reasons, and I think that might address some of the questions of where there’s some confounders in there, as Mark highlighted. We attracted the same dataset, the same clinical outcomes, and so we understood much better how the phenotype looked in these patients, and at the age of that phenotype for these quotations.
MARK: I’ll speak to my own research experience. In my PhD phase, you had to fight the tendency to change your experimental modality when you’ve learned something from the first iteration, because then, you can’t have compatible data. There’s the necessity to be sure that you lock down the right parameters and the right collection techniques so that there is compatibility across data sets, which is hard when you have two sites. It’s hard when you have clinical heterogeneity or genetic complexity. It’s especially hard when you’re trying to make inferences across multiple related diseases, because you’ve got to coordinate those different groups. One of the ways that you and our other clients have benefited is understanding what those outcomes and those parameters may look like, a priori. Not only is there synergy between the retrospective and the prospective, but there’s benefit to doing them in series. We worked on these in parallel and came together, but doing that in series, looking at the retrospective to inform the way that you set up your prospective, particularly if there’s complexities and nuances that you need to disentangle to get a real good handle on, say, the genotype/phenotype correlations that you’re talking about.
GUS: Yeah, and we didn’t touch upon that. While we’re looking at the genetics and are focused here, what collaborating with those two investigators and their sites and their information is trying to help us determine is what clinical endpoints should we be looking at and determining in our clinical trial. You see this in many other instances, at least in my experience, usually when you start off in a rare disease, it’s the overt symptom that catches everyone’s attention.
GUS: But as you start to break down whether it’s the genetics or the protein, as you get to inspect, you start to see other patterns develop when you get more and more patients involved. What that retrospective and that chart review is doing for us is answering, are there other patterns that weren’t so overt? but, guess what, multiple patients are showing the same thing, and can you explain it with our physiology? Maybe that should be an end point. That’s part of it as well, why these these prospective retrospective sort of progression natural history studies are critical, even, again, as I say, at our stage at this point.
MARK: Yeah, in this context and in others, I’ve heard that you’re very right. Sometimes on paper, the clinical presentation or phenotype that seems like it’s the most important or the most appropriate for an endpoint is actually not the thing that would improve quality of life if corrected, or that should actually be tracked for final outcomes, which is to say, survivability. Sometimes, it’s diarrhea, it’s not seizures, as an example. Really appreciating that beforehand, before you design your prospective endpoints and data parameters, is important.
I’m looking at the clock here, so just one one last question, and it’s forward-looking and somewhat optimistic. With the benefit of this aggregated, comprehensive dataset, if you do see a heterogeneity in response, if you do happen to have differential pathogenesis, even slight, because of this mutation or that mutation, and that may have bearing on the success of your intervention, I wonder if you could speak to how the data would be used after the trial when the data has emerged, in looking at and reflecting on what that data may mean?
GUS: That’s a really good and scary question in that sense. We like to think that, when we offer a therapy to patients in general, all patients will respond, and they all will respond with equal magnitude, and that’s not usually the case. There are many studies that show that either some patients don’t respond, it’s very marginal and you’d call them a non-responder, or they don’t hit the target. What the genetics will do is try to figure out if it’s the way our drug is designed and targeting, or if there’s another biology behind why they’re not seeing what they’re seeing. For example, we’ve seen this in the old days with general drugs being metabolized much quicker through the live. We come back with some of the isozymes of the liver with through metabolism. Back in the day, our patients would have no response sometimes. They were just metabolizing the drug too quick, and you’d have to adjust. Then, we started to look for some of these isoforms to understand if they’re going to respond or not. I look at it the same way. Just to be clear, I don’t look at it like we’re only going to take the top echelon, and we’re not going to provide some sort of general support, but it helps us, one, manage expectations. You never want to tell a patient, this is going to work at 100%. You’ve got to set expectations, and this will help us do that. Number two, are there particular confounders and particular reasons why we’re not seeing as great a response in some patients versus the other? Is there something we can do about that? That’s how I look at it. Sometimes, I’ve worked with others, where they get worried that we’re gonna cut down the patient population from here to here, but that’s not what this is about. This is about how you optimize what you have, and that’s the way we’ve been looking at it from a genetics perspective and response to treatment perspective.
MARK: Well, Kate, I’ll call you back to the stage. Gus, that was awesome! As always, I always enjoy our conversations.
GUS: Yeah, that was good, thank you! I appreciate the time.
MARK: Thanks also to the audience! I will be signing off and pass it over to Kate.
KATE: Thank you both so much! Yes, thank you, Gus, and thank you, everyone for watching today! If you have any questions for our teams, feel free to contact us at the websites you see on your screen. Again, I’ll be sending the recording to this webinar to you today. Thank you again for joining us! Have a great day.