Ultragenyx Pharmaceutical Inc. (RARE) Earnings Call Transcript & Summary

Ladies and gentlemen, thank you for standing by, and welcome to the Ultragenyx 2020 ASGCT Investor Conference Call. [Operator Instructions] Please be advised that today's conference is being recorded. [Operator Instructions] I would now like to hand the conference over to your speaker today, Danielle Keatley. Thank you. Please go ahead, ma'am.

Thank you, and good morning. Welcome to the Ultragenyx Conference Call to discuss the data presented at the American Society of Gene & Cell Therapy virtual meeting to speak. We issued press releases earlier detailing the data results from the DTX301 and DTX401 studies, and slides to accompany this call can be found on our website at ultragenyx.com. With me today is Emil Kakkis, Chief Executive Officer and President of Ultragenyx. I'd like to remind investors that this call will include forward-looking statements, including those related to management's outlook or predications of future results. These forward-looking statements represent our views only as of the date of this call and involve substantial risks and uncertainties, including many that are beyond our control. Please note that actual results could differ materially from those projected in any forward-looking statements. For a further description of the risks and uncertainties that could cause actual results to differ materially from those expressed in the forward-looking statements as well as risks relating to our business, see our periodic reports filed with the SEC. I'll now turn the call over to Emil.

Good morning, everyone, and thank you for joining us. Ultragenyx had 9 important data presentation at this year's virtual ASGCT meeting, highlighting our 5 clinical and preclinical gene therapy programs as well as our HeLa PCL and HEK293 manufacturing technology platform. During this call, I will focus on our 2 clinical stage program: DTX401 for the treatment of glycogen storage disease type 1a or GSDIa; and DTX301 for the treatment of ornithine transcarbamylase or OTC deficiency. Our slides are posted on our website. You can follow along with the slides as I discuss the results. I'll start with the GSDIa data beginning on Slide 4. This data is being presented later this morning by Dr. David Weinstein, Professor and Director of the Glycogen Storage Disease Program at Connecticut Children's Medical Center and UConn Health. Slide 5 describes the study design for DTX401, a Phase I/II dose escalation study for adult patients with GSDIa. We've previously announced results from the first 2 cohorts in the study where we showed that all patients responded to treatment, demonstrating meaningful and sustaining cornstarch reductions over time and significant increase in the time to hypoglycemia. Based on these data, we identified 6x10^12 genome copies per kilogram is the appropriate dose and moved to a third cohort to confirm these results. On Slide 6, you can see the 3 key modifications that we made to the protocol before starting the confirmatory Cohort 3. These changes were based on what we learned from the first 6 patients that were treated with DTX401. First, we reduced the cornstarch dose at the start of each controlled fasting challenge from 35 grams to 5 grams. By significantly reducing the amount of cornstarch given to these patients at baseline, we were able to see more clearly the effect of gene therapy alone without the results being confounded by a large intake of cornstarch at the beginning of the test. This change is also intended to mitigate an acute rise in glucose and the resulting insulin surge that was confounding the conduct of the controlled fasting challenge. Second, we implemented continuous glucose monitoring or CGM to allow physicians to assess blood sugars in real-time and more proactively manage cornstarch intake. These changes were important and they allowed us to better see the early effect of treatment DTX401 on glucose metabolism in the third cohort of patients. In addition, we began steroids earlier whenever alanine aminotransferase or ALT levels began to increase from baseline, even if they were still within the normal range. This was done in order to help reduce ALT increases as seen in Cohort 2. On Slide 7, we show the effect of reduced cornstarch dose on baseline time to hypoglycemia. So 1 hour after the evening meals provided, the patient receives 5 grams of cornstarch rather than the 35 grams as previously received. The decreased cornstarch dose led to a significantly shorter time to hypoglycemic baseline. In Cohort 3 when patients received only 5 grams of cornstarch, they became hypoglycemic in just a mean of 2.3 hours with 2 of the 3 patients under 2 hours. This rapid time to hypoglycemia shows how dangerous untreated GSDIa can be, crossing a low of 60 mg per deciliter of glucose in less than a couple of hours. This further highlights the unmet need in the fragile state of glucose control in these patients. Now Slide 8 shows CGM monitoring data prior to treatment and shortly after treatment for patient 9. Three days prior to treatment, this patient only had a couple of brief episodes of hyperglycemia, which is shown in the red circles on the top graph. Just 4 days after treatment with DTX401, the CGM data show substantial and prolonged periods of hyperglycemia you see on the bottom. We believe this is due to early transgene expression causing the liberty to release glucose and resulting in significant hyperglycemia. This then causes insulin level to surge and can lead to precipitous drop in glucose as shown on the bottom graph. These hyperglycemic periods and wide swings at glucose show that DTX401 is beginning to change the glucose controlled paradigm in these patients. So turning to Slide 9. The controlled fasting challenge test of Cohort 3 shows the time to hypoglycemia increase for all patients in Cohort 3 at week 12. As I pointed out earlier, baseline time to hypoglycemia was 40% shorter with Cohort 3 than the early cohorts due to the reduced cornstarch dose at the start of the fasting challenge. During the 12-week test, all 3 patients in Cohort 3 showed significant increase in time to hypoglycemia, ranging from a 23% increase to a 167% increase. These early results are promising, but keep in mind, these patients metabolisms have not yet fully adapted to the new glucose environment since it is early in the treatment course. Due to COVID-19, these patients were not able to attend their 24-week visits. As a result, we only have the data from the 12-week controlled fasting challenge. It will be important to monitor these patients over time, and we anticipate that we would see continued improvement in time to hypoglycemia as patients become metabolically stable. Next demand to bring patients back in for fasting challenges once sites begin to open back up. We plan to share data from these visits in the second half of the year. On Slide 10, we observed that all patients in the study experienced a significant reduction in cornstarch requirements. Cohort 3 was, in fact, on average, faster than the prior 2 cohorts. As a result, this cohort reduced cornstarch by 57% at week 12 compared with 38% in the first cohort and 14% in the second cohort. Across all 3 cohorts, 4 of the patients have reduced cornstarch by more than 85%, meaning they are receiving very small amount of cornstarch each day. Four of these 6 patients have discontinued daytime cornstarch, which is extremely encouraging and shows that patients in the long period of time were needing supplemental glucose. Two patients are essentially off cornstarch entirely and are able to sleep through the night without getting into a crisis with their glucose. We've been able to see this through their nightly glucose levels through the use of the continuous glucose monitoring. These are significant changes that weren't possible prior to treatment with this gene therapy. On Slide 11, you see an example of profound impact reducing starch can have on patients' daily lives. Here you see a photo that patient 3 shared with us from a trip he was able to take about a year after receiving DTX401. On the left, you can see the amount of cornstarch he would have needed prior to treatment, 11 boxes of cornstarch or 1 box for each day of the trip. After 1 year of treatment, at that time, his cornstarch requirements decreased by 75%, he was able to dramatically reduce the amount of cornstarch needed for his trip. During that time, his cornstarch requirement -- since that time, these cornstarch requirements have decreased even more by 92% after 85 weeks. In summary, on Slide 12, patients in Cohort 3 had more rapid reduction in cornstarch compared with first 2 cohorts since physicians could reduce cornstarch more rapidly based on continuous real-time glucose reading. We saw early implementation of steroids with the optimized regimen, which was more effectively mitigated in ALT levels. All patients in the study have responded to therapy and have demonstrated meaningful and sustained cornstarch reductions over time and significant increase in time to hypoglycemia. The safety profile has been acceptable and consistent. Turning next to the next steps for the DTX401 program on Slide 13. We continue to collect longer-term data from the Phase I/II study, which we will share in the second half of this year, barring further delays related to COVID-19. We are moving ahead with end of Phase II meeting with the FDA, which we also expect to hold in the second half of 2020. For the Phase III, we're looking at endpoints that are all based fundamentally on glucose metabolism with advantage of having glucose as an accepted biomarker. In the Phase I/II study, we looked at glucose metabolism in 3 ways. When we started the study, the best available way to monitor glucose was through the in-house controlled fasting challenge. This data from the controlled fasting challenge has been meaningful though we have learned that it takes time for patients to move out of the original treatment period to become stabilized on their lower cornstarch regimen. Accurate continuous glucose monitoring with the Dexcom 6 device became possible during the study. Importantly, it allows us to actively observe lower glucose levels on a regular basis in a real-world setting, not just during fasting challenges. With CGM, we can look at the reduction of hypoglycemic events that patients experience as well as the quality in their glucose control with large amounts of data being collected. Third, we could look at cornstarch dose reduction, which is a measure of how much amount of glucose is required to maintain euglycemia. We have observed the need for profound and meaningful changes in cornstarch requirements early on after treatment with DTX401. We anticipate the tally of data, including cornstarch, time to hypoglycemia through controlled fasting challenges and overall reduction of hypoglycemic events or quality of glucose control as measured by CGM could all address the efficacy of DTX401 and GSDIa, and could be incorporated as endpoints in the Phase III study. We expect that we will meet with the FDA in the second half of 2020 and we could begin a Phase III towards the end of the year barring any further COVID impact. I'll now turn to Slide 15 to briefly discuss the data from the Phase I/II study of DTX301 in ornithine transcarbamylase or OTC deficiency. This data was presented to ASGCT on Wednesday. We're encouraged by these longer-term data from the first 3 cohorts in the study. Six of 9 patients in the study have responded, including all 3 patients in the third cohort. The patients in the early cohorts continue to do well and they are showing durable responses that are being maintained over longer periods of time. There are 3 patients classified as complete responders in the study. These patients have all discontinued ammonia scavenger medications and liberalized diet, and continue to remain metabolically stable without recurrence of hyperammonemia. On Slide 16, I will highlight some of the new updates, which you can see in green. Starting with patient 9 in Cohort 3, this patient was previously deemed a potential responder at an earlier update in January. He is now a confirmed responder with 2 increased ureagenesis measures and a 188% increase in the rate of ureagenesis from baseline. His ammonia levels have remained stable in the normal range since treatment. This patient is still on tapering steroids and it's not yet discontinued alternate pathway medications or liberalized his diet, but we anticipate he'll begin to make these changes once he's able to return to the clinic after the COVID-19 restrictions have been lifted. The other new update in this third cohort is Patient 8. This patient demonstrated a 90% decrease in ammonia levels. She has now discontinued 1 of her 2 ammonia scavenger medications and has increased her protein intake. She continues to do well clinically and the plan is to begin to taper her second alternate pathway medication. Patient 7 remains a complete responder. This patient had a meaningful change in the rate of ureagenesis, has maintained normal ammonia levels since treatment. She continues to be clinically and metabolically stable after liberalizing diet and discontinuing ammonia scavenger medications. Going to the earlier cohorts, we had 2 complete responders, Patient 1 and 4. These 2 patients are showing ongoing durable responses and they've been followed for 2 years and 1.5 years at this point. Both patients have discontinued alternate pathway medications, liberalized their diet for more than a year and remain stable with ammonia levels maintained in the normal range. They are in excellent clinical condition with no significant adverse events, hospitalizations, or other events related to OTC deficiency. The other responder is Patient 6 in Cohort 2. This patient has had a significant improvement in ureagenesis of 218%. She has maintained her normalized ammonia levels at week 78. She continues to do well, and is tapering her medication and liberalizing her diet. To summarize on Slide 17, the durable responses we are seeing in DTX301 study are encouraging. The recent data reinforce the response rates with the Cohort 3 dose of 1x10^13 genome copies per kilogram. These results give us confidence that this is the right dose for the Phase III study. We plan to enroll a fourth cohort at the same dose by using prophylactic steroids. We believe the prophylactic steroids could further enhance the level and consistency of expression that has been demonstrated in the study so far based on our own work and published work of others. Because of COVID-19, dosing this first -- fourth cohort is currently on hold though patients were already screened but just could not be enrolled. We still expect data in the second half of 2020 for those patients barring further delays related to COVID 19. In parallel, we are planning the Phase III study. We intend to hold an end of Phase II meeting with the FDA in the second half of this year. We'll base that discussion on the existing data from the first 3 cohorts. We anticipate using prophylactic steroids in the Phase III study, and we will confirm this decision with the data from the fourth cohort once available. For the Phase III study, we expect ammonia will be a primary endpoint based on direct FDA feedback to date. Barring any major COVID related delays, we should be able to begin enrollment in the first half of 2021. To conclude, these data give us additional confidence in advancing our 2 clinical stage gene therapy programs into Phase III studies. In both programs, we have selected a dose for the pivotal programs and are moving forward with planning those studies. We are actively conducting conversations with FDA around the Phase III studies and look forward to provide more updates on both these programs in the coming months. With that, I'll open up the call to questions. Operator, please provide the instructions for the Q&A.

[Operator Instructions] Our first question comes from the line of Paul (sic) [ Maurice ] Raycroft from Jefferies.

And congrats on the update today. I was wondering for the glucose -- continuous glucose monitoring, if you can provide more specifics about how you're doing that? Are the patients typically do what they're supposed to do? And how reliable or variable can results speed? And then for the expression revealed by the continuous glucose monitoring data, can you talk more about trends that you're seeing across the patients? And in a commercial setting, I guess, form an algorithm for how a doctor can wean the patient off his supplements?

Very good. So with a continuous glucose monitoring data, we're certainly developing our insights from it now, and it's just the early days for us in looking at it and analyzing it. The device we're using is a Dexcom 6 device. The reason this is important is that device has a more accurate measure of glucose, it is not interfered by lactate. Lactate is -- can be elevated in these patients. So it's real important to have a device that's very -- has a very accurate glucose measuring, particularly in the lower range and in the presence of any lactate. So there is -- the devices are put on the patient at the clinic. They're relatively easy to manage, but there are some techniques in terms of how often to change your probes and how to update. And there's ways to mark when you take starch on the tracing so it can be known when starch was taken. So those are things that patients can do and they'll be instructed in follow-up on how to manage that. We see that these patients genetically have significant glucose variation in the pretreatment, the limited -- we have limited pretreatment data at this point, but they are having variation when they take cornstarch and sugars are going up and down. So it's pretty clear that they're a relatively brittle glucose control in the state where your liver doesn't make any glucose. What we're seeing and certainly we've shown you on Slide 8 was that like more extreme hyperglycemia and hypoglycemic activity because the cornstarch dose is just too high and we need -- this is giving us that ability to see the what's going on and to understand that sometimes when they're having hypoglycemia, it's because they had hyperglycemic just 2 hours before. So that is going to help doctors interpret what's really happening in a real-world setting and be more accurate about understanding that an episode of hypoglycemia may actually be due to a period of hyperglycemia. So what we're seeing over time is we are seeing -- in the patients who are later in the study, we have been looking, we are seeing a relatively stabilization, a less smaller standard deviation in the glucose control and we think it -- we certainly -- because to the patient essentially off glucose, with monitoring, we can see that their management of glucose during the night without starch. And by seeing that on the same 24-hour cycle repeatedly, you can kind of basically get equivalent of a time to hypoglycemia test done at home and done multiple times. So what we're seeing in those patients are -- further along in the course is a stable glucose levels, and we feel good about the way this looks over time, but it's pretty clear these patients' bodies are distorted by the extreme levels of starch and control for 2 decades, 3 decades. And so weaning them off of that physiologic setting to a new setting where their liver is doing the work and starch is not is -- it takes some physiological adaptation. And that's what we're learning from the data so far with CGM.

Got it. That's really helpful. And maybe just a quick one on the 5 -- the decision to take 5 mg for the go-forward cornstarch dose. Just wanted to check the box to see if there are situations where that could be too low. And just anything else you could provide on how you got to the 5 mg number?

Well, the discussion of how much -- how low to go, there was an ad board conducted with investigators. And the balance of issues here have to do with, when you take a meal, you will get some insulin response and the doctors are concerned that without any starch that the patient could bottom out very quickly and put them in danger. So the balancing is it's -- well, going to 0 would allow you to see only what the liver is doing. There was a little bit of concern that they might go down too quickly and that, that might put them in danger. And so the agreement was to go at less than 10 on the protocol with some judgment left to the investigator based on how much starch that patient is still on. So at the beginning, they felt since the patients were on starch, they went with 5 grams, but they stuck with the 5 grams as they went through, at least in the first few months of the protocol. But later they could take the 5 grams away if they felt it wasn't really necessary. So the 5-gram dose is just essentially a token amount of starch, but it was enough to give the investigators confidence that the patients weren't going to crash so quickly that it will put them in danger after a meal.

Our next question comes from the line of Yaron Werber from Cowen.

Great. And also my congrats, it's really nice to see this coming together. So Emil, I have a couple of questions on OTC that are related. So when you're talking about a primary of ammonia levels, is it -- do you think it's going to end up being a percent of patients who are normalized? Or is it going to be a certain percent reduction from baseline too, I assume, and again, maybe I'm making an assumption that you'll have a responder endpoint? So either achieving normal level or a certain degree decrease from baseline. Is that the way you're thinking about it?

Well, we -- as a general rule, we don't do percent of patients that respond because it's not a very powerful way to do the analysis. So we will look at absolute ammonia levels and their change from baseline and use repeated measures, which will give you a lot more power. Our expectation at baseline, though, is that not all the patients will have elevated ammonia, there'll be a wide range of ammonias in the study and we'll allow some elevated ammonias to be in there, so it will be this big spread. Our expectation is that by 24 weeks in the study that all the higher ones will contract down, and we'll be able to show the difference in the population as a group. Over time though we will be -- if you're removing the drugs because they've got no need for the drugs or diet, perhaps at week 52, we might want to look then at how are they doing ammonia control without drugs or diet compared to how they did when they were on the drugs, right? That would be kind of comparing them with 1 set of treatment versus another set of treatment. So that will be another way to use the ammonia as a comparison to 2 different states for a patient off drugs and diet versus on their drugs and diet. And so those are the 2 ways I envision currently. And we still need to have our discussion with FDA on how to do it and really nail down exactly what the patient population will enroll. But we do expect to enroll some patients with more problems with ammonia control than in the first study which were intended, first of all, safety to assure that we were getting people that weren't going to be harmed. And the data we have today, we can tell that that the AAV gene therapy safety profile is excellent and we feel that there is reason now to look at more severe looking patients.

Okay. So it's possible that the primary technically will be in patients with elevated ammonia versus baseline, even though not everybody is going to have that, and you might have a need to enroll a certain number of patients you can power it accordingly?

Right. But you still have to include all the patients in the analysis even if they -- some of them were at the normal range. They still need to stay normal and the other ones would have to come back. But you actually -- because we're doing the actual ammonia levels, it's pretty powerful. So you should be able to see that very easily.

Okay. And so it sounds like you'll speak with FDA, perhaps, would you wait to have the prophy steroid data before you meet with FDA for the end of Phase II? Or would you meet with FDA sort of as you're generating that data and as long as that data continues to hold up, that's going to be the dosing in the Phase III?

Yes. We're decided not to wait for the prophy steroid data, which, as you know, is now delayed. Because we've screened the patients, we were ready to dose, but then the sites got shut down, right, as this was getting ready. Our view is that prophy steroids is something FDA sees repeatedly across many programs. We are using steroids in all the patients who've had no concerns with steroids. So we will proceed with the idea of using prophylactic steroids. We just want to test a few patients before we actually start the Phase III. Our expectation is will we talk to FDA and hopefully have that data at the time or by the time we're talking to them, but we're trying to do -- we're not going to wait for that at this point because it would further delay us.

Our next question comes from the line of Adam Walsh from Stifel.

Emil, my question is on 401. You've talked about having more data from the third cohort in the back half of this year and potentially starting Phase III by year-end. And the data in the back half is pending kind of COVID issues. But in these data, we have week 12 data only because the fasting challenge at 24 weeks in the hospital couldn't be collected. Will you be collecting 36-week data or just 52-week data? And does it matter that, say, you don't have the commensurate 24-week data? And how will you get enough confidence based on the totality of the data set to actually start the Phase III by year-end '20, if that's possible? That's my first and I have a follow-up.

Sure. Well, we will get -- so as soon as the sites open, we're going to have them come in, we will test them for a glucose test that is, we will have them come in as soon as we can. So whatever week it is, it will be that week. It'll -- it's obviously past 24 weeks, but if it's 30, we'll have them come in then. So we're not going to wait for a defined milestone. We have the protocol flexibility to actually pull them in at any point in time. So we'll do that as soon as we can. I don't think it matters to understand how these patients are doing and the control fasting challenge is one part of the story. But with the CGM data, so we actually know what's going on with their glucoses all the time and we can follow them during the night. So we can see how they're evolving. Are they doing better? Is this cornstarch weaning? Is their hyperglycemic in control? Hypoglycemic in control? So in some ways, this is a far better situation than having these milestone, than we only get a little bit of data every so often. So we should be able to actually monitor them during this period and get a feel for how they're doing. So we're actually more encouraged in our ability to do this in the COVID environment than before. But we would like the formal in-house testing because there's also the controlled fasting challenge. Remember, there's also other things we draw during that test and there's also other metabolites and things we're measuring. In addition, the MRIs, looking at their livers and other things like lactate that we want to look at. So those things would be helped by having the hospital test to give us a full picture. But I think from the standpoint of glucose metabolism, we will have enough ongoing data from CGM and cornstarch to tell us how the patients are doing.

That's helpful. And then just one quick one on 301. You've mentioned, I think in the last earnings call, your partnership with Arcturus, and they're developing a self-replicating mRNA approach to OTC. I know you have an investment in that company as well, I guess and over the next 5 years or so, how do you kind of compare and contrast in your mind, a gene therapy approach versus an mRNA approach?

Well, our general view would be that if you can do gene therapy, one -- one time and correct the patient permanently, it's going to be the preferred option for patients. However, that there may be patients, particularly at a young age before their livers are fully grown or more severe types that it might be beneficial to use an mRNA, for example, in neonatal or in young children before their livers have grown because we look at the gene therapy is requiring liver to be kind of fully grown up to avoid the effective dilution of liver growth. So I can see mRNA being a companion treatment that's used in the earlier stages of the disease for some severe patients and would not necessarily detract from a gene therapy in older patients that we're doing. So that's one way to look at it. I think there -- there's always room in a disease with 10,000 patients in the world for treatments and we feel very comfortable that both could coexist and be used in different ways. We're obviously working closely with them. We're also working on our own mRNA. So we're highly familiar with it. We're also collaborative with them. And we believe the mRNA strategies have a role. But I would think the average patient to take one-shot therapy and not have to do it again. Anything again will be always a superior choice if it can be done.

Our next question comes from the line of Gena Wang from Barclays.

First also want to congratulate the data, very good data. So Emil, how do you measure cornstarch requirement? The reason I'm asking is just trying to understand how consistent this data point would be. And given that we -- seems like with the other primary endpoint you were measuring time to hypoglycemia, if you have a higher protein expression, means like if you have a better clinical benefit, you actually got punished and you will continue to go down with the cornstarch and now using 5 gram. So just wondering if any possibility to use cornstarch requirement as a major measurement rather than the time to hypoglycemia?

Well, cornstarch, the way it's prescribed is the doctor tells the patient. They usually had a set number of grams and they have a set schedule, how many times a day to take it, which can be anywhere from 5x to 7x, perhaps around that range and a certain number of grams every setting. And so the patients are pretty work by clockwork. I mean they have regimens. So they are pretty regular in how they do the cornstarch. So that part can be managed very accurately. The thing that's important is when we talk about cornstarch requirements, we also have to make sure they're not getting hypoglycemic. So if you pair the CGM data and say they're not getting hypoglycemic, so the starch reduction is not putting them more at risk, then you can interpret the cornstarch requirement is being reduced. Does that make sense?

Yes.

So it's a reduction in cornstarch without inducing significant hypoglycemia. And so the 2 things together are required. You have to have CGM data to help verify that the reduction you're giving them is not changing their hypoglycemia risk. And if you put them together, then I think you can accurately then assess to what degree the liver gene therapy is actually in controlling glucose versus cornstarch.

Okay. Will you measure across certain days or just like 1 data point for the cornstarch requirement?

Well, in the graph you're seeing, the -- it's measured at different points in time. It was being measured at all the milestones that is recorded in our EDC system. On the graph, we sent on Slide 10, you can see the little X and dotted lines. Those Xs are basically call the doctor to determine what they were telling the patient what they were taking at that point in time. So that is the -- you can ask for that cornstarch prescription at any time. But by protocols coming in at certain points in the EDC -- in the data capture system. But we could get that information any time because it's changing as that is looking at their glucose data over time.

Okay. And my second question is regarding the steroid treatment. So you do modify to be more reactive use to begin early as soon as the ALT level increased from baseline. Just wondering why not just apply prophylactic steroid?

Well, it's a good question, Gena. I think the question in this particular GSDIa disease is that steroids can cause severe hypoglycemia because the steroids increase the uptake of glucose, but normal people also stimulate the release of glucose from the liver. But these patients don't have an ability to release glucose. So there is fear among the doctors that treat these patients, that giving them steroids without any gene therapy done will cause them to go severely hypoglycemic can be dangerous. So that is the reason there's a trepidation in going with steroids, let's say, the week or 2 before you start the dose. So what we did is moved the steroids up earlier in the course by triggering on any movement off baseline. And we're looking to doing what we call a post dose prophylactic steroids where we -- now that we know that looks like expressions beginning within a few days and the doctors feel more comfortable that we can give them steroids as soon as -- at a set time after the treatment and feel comfortable that's going to be safe. So we're trying to move it up to be as close to prophylactic as we can, but because of their disease state, prophylactic steroids in these patients is thought to be more concerned. So at this point, we've not pushed for prophylactic, but we do think the earlier and regular application of steroids will be far easier to apply in the commercial setting. And we think the steroid management of these patients has been very simple and there hasn't been a problem when they're treated with the gene therapy.

Our next question comes from the line of Laura Chico from Wedbush Securities.

This is Kenneth on for Laura. This is on 401, could you talk a little bit about the planned 24-week controlled fasting challenge. I believe you said that you would anticipate subjects will show increases at this point? Or should we be looking for stabilization versus the 12-week levels? And related to that, could you help us understand how this translates into clinical meaningfulness versus their baselines?

Well, we're seeing that once you get out of this, the cornstarch reduced. You're not having hypoglycemias, going to have less of the wild swings that are -- were affecting these patients early on. And so we believe at 24 and at later time points, particularly as we've talked about with the other patients that have gone way down in their starch, they're not having as wild the swings and their variation in glucose during the day is a lot more stable. So that's why we feel we'll get to a point where you'll get a little more stable, consistent response in the time to hypoglycemia. In terms of clinical meaningfulness, the time to hypoglycemia, we talked about sleeping through the night. I think what we're seeing over time these patients are able to get less and less starch and able to sleep through the night and to the patients essentially off cornstarch, are doing that fine or sleeping through the night, maintaining their glucose, do not bottom out or have any glucose crises. So I'd look at -- as we progress toward the 52-week point as the patient's physiology adapts to this new environment that we would look for their ability to wean off cornstarch and to then be maintaining their glucose during the night, either by CGM or a time of hypoglycemia test for more and more time. So, so far, we think is encouraging what's happening, I think the lack -- less cornstarch of the test is not causing the acute drops that we were seeing before 24 or 12 weeks, all of the gene therapy programs are really looking at 52 weeks of data in our Phase III studies in general. So that's -- what we've heard from FDA is that they want to see a year of data in Phase III study. So we look at this period as sort of the evolution of the disease in response to the effective treatment. But ultimately, 52 weeks' worth of data will be what the FDA will want to look at in terms of a Phase III data package.

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

Emil, a couple on 301. Would you expect superior efficacy in terms of ammonia change if you do use steroid prophylaxis? And can you give us a sense of what degree of efficacy improvements you expect there? And then based on what you're seeing for Cohort 3, when do you think is the right time to measure the primary endpoint on ammonia? And then I have a follow-up.

Sure. So I think with steroid prophylaxis, what we would do is measure their ammonia, the 24-hour AUC of ammonia before they'd start the steroid regimen because if you put them on steroids, that will clearly alter their ammonia and that would happen for placebo. So if you put them on steroids, their ammonia went up and then you took the steroids away, now the placebos probably won't get steroids. So you would create a little bit of noise there. If we were to do the AUC, then you would treat the steroids and follow the patients and the placebos will just get placebo. The time point for following ammonia, it's pretty clear from the data we have so far that when they go down, they go down within 6 weeks, like, very quickly. We think the regenesis takes a little more time to build up, more, like, 24 weeks for some of the patients, some of them even continue to grow between 24 and 52 weeks. The challenge is we don't want to remove their drugs and diet before we take the first look at ammonia superiority. And so I'd expect us looking at 24 weeks. But if you want to look at the patients taken off their drugs and diet, that's probably going to have to happen at 52 weeks because it just would take time to wean them off and see how they're doing to verify that they're stable. So it's somewhere between 24 and 52 weeks, and maybe we're looking at it 2 different ways at those 2 time points.

Our next question comes from the line of Cory Kasimov from JPMorgan.

This is Turner on for Cory. I'm just hoping you could help us think about the time to hypoglycemia in Cohort 3 versus cohorts 1 and 2, just given the change in the cornstarch dose? Relatively, do you think the data is more or less robust? Or is it just too difficult to compare?

Well, it's pretty clear the cornstarch was accounting for 2, 3 hours of control. So I think it does make it hard to compare head-to-head the change significantly. And the degree of severity, these are small cohorts. So it's hard to verify exactly how comparable the patients are and there was some variation in cohort -- Cohort 1 and 2 in terms of timing of time to hypoglycemia on 35 grams as well. So I think it's a little bit hard to go straight to it that way. And I think what we can see is in cohort -- Cohort 1 and 2, whatever start tests they were on, as their starch gets weaned down and as we look at them with a CGM without the test, we can see them getting in control and stabilizing. So we feel that when we do the 5-gram starch approach with them or the less, it will tell us how well they're doing through the night, and we can tell already from the CGM data that they're doing better. So the comparison of the 1 test and the other, I think, is going to be hard to do because of the variable and confounding effect of it. But I do think we have enough data to interpret whether the early cohorts are doing well or not. I think they are doing well. And I think that the time to hypoglycemia will improve as time goes on and as their insulin -- hyper insulin state kind of settles back as they get deeper into the study.

Our next question comes from the line of Yigal Nochomovitz from Citi.

This is Samantha on for Yigal. Congrats on all the updated data. I kind of want to follow-up on a prior question. You were talking about DTX301 and potential endpoints and time points. How would you standardize the removal of drugs and diets in these patients? I mean I think the data so far, I believe, has been investigator-driven for when you would remove or liberalize the diet or remove the scavengers. What happens to the patient if he's ready to go off drug and diet at a predefined time?

Well, yes, I think the -- what's been happening is we've been usually looking at ureagenesis in case of ammonia. If they're achieved what we think is a significant improvement and the removal of drugs usually first. And what's interesting is we were trying to taper patients in a standard way, but patients want to come off immediately. And so what we're seeing with the drug, most of the drug patients are just stopping immediately because they don't like taking them. So their desire is to go off completely in some, and in Patient 6 case, she took it off. She took herself off herself and admitted to it afterwards. But -- so that -- the way the drugs come off is kind of immediate. People feel well, ammonia is under control, they want to come off. And so far, that looks like that works. If they have a significant response to ureagenesis, then they could come off their drugs essentially immediately and watch them. The diet part, I think, is more of a tapering effect and doing it, I think what would happen generally is if patients achieved a certain level of ureagenesis at week 24, we would enable their doctors to begin loosening their controls. But doing it in a blinded study is tricky because we'd have to have someone independent looking at that information and allowing patients to move forward. So working through the logistics of that. I think what you're getting is how do you control the actual amount? But what I'm saying to you is that the endpoint would be getting off the drug completely or getting -- and getting off the diet completely. No approaching restriction. Since we're able to achieve that in some of the patients, we really want to look at getting all the way down rather than, let's say, certain degrees of change, which I think would be very hard to standardize because patients are eating whatever they're eating, and you have to -- it's very hard to accurately measure various degrees of compliance. So we'll be looking mostly for patients coming off drug or off diet completely as a sign of being a complete responder.

Okay. That's actually -- that's very helpful. And can you remind us in that because I think you had 3 complete responders, what weeks did they become complete responders?

Well, it was varying. I think we had some of the patients coming off their drugs at around 24 weeks and then around 36 or 32, 36 weeks they're coming off the diet, I think, is the first 2 complete responders. Patient 6 actually came off her drugs earlier, I think. So there was some variation between them. But I would say, in general, that they were able to get off tracking off drugs that are around 24 weeks and diet was between there and 52 weeks.

Okay. That's helpful. And then just one more on 301. Was there anything unique about the 3 nonresponders particularly the one that had the elevated ammonia at baseline?

Well, not that we can tell right now. Originally, we had this issue, whether they were female or not, a male seemed to be the responders, the few most for us. Obviously, at higher doses, we're getting both males and females response. So that was one consideration. There's no tighter like antibody tighter, like a slight tighter or differential on antibodies that would have explained the difference. We think there's just differences in any immunity perhaps or other aspects that simply providing more virus is able to overcome.

Our next question is from the line of Jeff Hung from Morgan Stanley.

They're both on 401. I guess the first one is, can you provide a little bit more color on the asymptomatic and transient rise in ALT for patient 8? Do you know what the cause might be? And how often does it happen in patients?

Well, the Patient 8 rise we saw in Cohort 2, we had all 3 patients had a rise. And so it's just I would call it, traditional AAV transaminase response. It's response to the steroids, how we're -- the concern when you have elevations like that is that you might get loss of expression of the transgene, is the same, for example, in the hemophilia program. So that's why we increased -- we increased the speed of application of steroids in this cohort, which in 2 out of 3 patients, help keep them in control without any rise. We'll probably go to a more, what we call post treatment, prophylactic response where we put them on steroids a week after dosing, for example, and we hope that early application will keep the inflammation and control, which to assure it doesn't have any impact on expression, which is the main thing we want to watch out for.

Okay. And then I guess for Cohort 3, how much of the time to hypoglycemia at baseline do you attribute to the modified protocol? And then can you talk about the disease severity in these patients versus those in the prior cohorts?

Well, the modified protocol, you can see in patient 7 and patient 9 are at like 1.5 hours and 1.9 hours. So that's what the 5 grams of starch is doing. Without any starch at all, it probably would have been nothing now. At the time point 12, they are getting the 5 grams of starch too. So they're both done the same way. But I think you can see that they have a substantial increase in the amount of time. And the truth is that these patients have a history of making a lot of insulin to manage the starch they've been given. And we think that they're probably on the hyperinsulinemic side based on how high their insulins are even when they're fasting. So I think there is a little bit of the challenge of having a patient that is producing more insulin than they need. And this is the kind of thing we think we might take some time to adapt as a change from a starch insulin based regimen to a liver based control regimen during fasting.

Our next question comes from the line of Salveen Richter from Goldman Sachs.

Regarding patient 8 from Cohort 3 of DTX401, can you just discuss why that patient demonstrated a lower magnitude increase in the controlled fasting challenge? And a follow-up.

Yes. So patient 8 start is the 1 patient started a little bit higher, just a less severe hypoglycemia in that particular patient, that gene type. So they started higher. If you look, they actually ended up at close to the same place as the other 2. I don't think -- and you can see on -- if you look at Patient 8 on the cornstarch reduction on Slide 10, you can see they're already at 71% reduction. So you have to ask yourself, how do they get all the way down to 71% reduction, majority reduction in cornstarch without having hypoglycemia, but they had a 23% increase? So this is somewhat discrepancy that's going on is that the truth is that there -- their metabolism is so cornstarch dependent that during this period, we're weaning off their cornstarch, which is affecting their starting glucose level. So when you do a fasting challenge, should be started at 110 or you start wherever you start, has some impact on when you cross 60. And so I think you're just seeing what we've been understanding now as we've learned and done the study is that the patients need to have their starches weaned promptly, which will affect how their cornstarch reduction goes and you can see this patient being at honestly hit 70% at week 12. It's probably a factor in how it's time to hyperglycemia was going because they need to wean his cornstarch. As we get further away and further out, then we think the stabilization will allow the time of hyperglycemias to improve. So I would tell you, even though there was a 23% increase there, starch reduction tells us he's having an excellent response to treatment.

Our next question comes from the line of Vincent Chen from Bernstein.

This is Brian on for Vincent. I guess just a quick one on DTX301. Is there a reason to believe that the prophy steroids should alter the time course of response? Or just trying to dimensionalize what kind of data we might be able to see into H '20 given the kinetics we've seen so far?

Well, we would expect that the time course would probably be similar. I would point out 2 things. One, the steroids are -- we expect to help the prevalence and consistency of expression level of expression. But you also know that the steroids also impair the regenesis for a period while they're on them. And when they come off steroids is when we see the really big improvements in regenesis. So because the steroids caused catabolism to some degree. So we'd expect that in the early periods of the study, we will be able to put on the steroid sooner and wean them off a little sooner. And so that might help us see their ureagenesis a little bit earlier because they're going to be off the steroids at an earlier point in time, we'd expect.

Okay. Makes sense. I guess just as a quick follow up, I'm just thinking about that one and I guess we still don't have a great sense of why exactly, the response facility, but one of the hypothesis you guys have floated was kind of sum in immune response. I was wondering if steroids might mitigate that or you might see quicker response in any other patients.

Well, I think it's possible that suppressing an immunity might allow the expression to come on a little bit sooner if it's being repressed by -- through some of the mechanisms that can repress DNA expression. So it's possible that is something we can look for. We think, though, even with the delay or later response, by week 52, we still are seeing it. So we think it's within the range of what we're already planning for the trial. The question is that in the second half of the year? Is it the first half of the year? I think the prophy steroids cohort will tell us what we're seeing in the first half of that year time frame. And that might give us a hint that we can accelerate. But -- and certainly, that would be beneficial if that were true.

Our next question comes from the line of Arlinda Lee from Canaccord.

Congrats on the progress. I maybe wanted to follow-up a little bit more on the cadence of response in 301 and the -- I guess, the effect of the steroid on that? And I'm kind of curious if you can maybe talk further about the steroid prophylactic and the level that you're going to, are they -- is it like a short course and then a tapering? Or -- and how do you decide if you need to give more?

Well, the idea is to put them on around a 4-week regimen, which is kind of standard and then a taper that takes a month to do. So it's a month and then a taper month. That's pretty what's happening now, but it's happening now as a trigger by inflammation. The idea is to do it prophylactically. And the hope would be then that after the end of that, you would not have rises in ALT. However, the protocol will certainly allow us to if -- to monitor their ALTs, and if the ALT goes up after the end of that protocol, after that end of that standard steroid protocol, we certainly can apply steroids again. We're just trying to -- we're hoping that the regimen will prevent steroid -- prevent ALT rises in general so that we can apply that in the commercial setting which would be way better for people. But it still may require a little bit of monitoring, but we hope not a monitoring that is required every few days repeatedly for several months, which is a burden. But if you can get 80% to 90% of the patients covered with a regimen and not have any rises, then we could do monitoring on the rest to assure whether they'll need any secondary support. And we'll learn a little more about that from the Cohort 4, but also within the Phase III as we use prophylaxis steroids in the regimen there.

Great. And then maybe on the cadence of response in 301. Can you maybe talk about -- you said that it was within the range of what you were expecting. But I was wondering if you can maybe speculate a little bit further on if there's a way to tell ahead of time and if it makes this difference in the longer run if you're getting into the -- the range that you're expecting?

Can you ask that question again? I just didn't quite get the beginning of it.

On the 301, the cadence of response. Wondering, there was a patient that looked like they had a more delayed response? And I'm wondering if you could speculate on why that might be and if they're within the range of your estimated time frame in any way if that makes a difference? And how you're thinking about this?

Yes. Well, we had a woman with a more delayed response in Patient 6 and that patient was toward the -- closer to the end of the year, second half of the year that she had a response. But within the 1 year time frame, that's why I don't think it will matter. The trial will include that, that delay shouldn't matter. I think that, as we just mentioned, maybe the steroids might help increase the speed of response, but we don't know that for sure. But I don't think it will matter. I think at a higher dose, I think we would build to capture an effect if it happens later in the first half versus the first half. But our expectation is the trial will have a 52-week time course in it. So that should be a sufficient time to capture any effect.

Our next question comes from the line of Gena Wang from Barclays.

My follow-up. Just one quick question regarding the Cohort 1 and 2. I'm sorry, this is regarding the GSDIa data. So Emil, just wondering how -- what dose do you use for these patients? Would that be 5 gram cornstarch? And how do you compare the time to hypoglycemia to the early time since like the baseline be quite different, between 35 versus 5?

Yes. Well, we are using the shorter -- I mean the 5 gram dose later and it will definitely reduce the time of hypoglycemia because there's no starch supporting it as well. But -- so it's going to be harder to compare, the 2 tests were done in different ways. I don't think it matters, though, Gena, because I think we can look at the patient as they are. We know these patients can't tolerate life without cornstarch, right, that is how they will achieve their survival. So if we get to see those patients later, the time of hypoglycemia comparison is less important than understanding where they are at that moment in time. If they get 5 grams starch, how long are they going? If we follow them on CGM at night, just with their current regimen, including patients that are nearly off starch completely or are off, it gives us the ability to see how glucose control is working and if we can get them to a place where they're not needing starch and they're not getting hypoglycemic, the comparison baseline doesn't matter because that means their lives have changed. So the comparison to baseline issue, I'm not really concerned about the Cohort 1, Cohort 2, but it will change what happens in the test because the starting glucose level will be higher, if it's a little bit higher at start, then the time changes. We could potentially look at flow, how fast they're declining as a measure of how good the glucose support is from the liver, and that might help normalize between different months of starch. But again, it's complicated because the starch was causing insulin surges as well. So -- and because of physiology is changing, then the test in the early time versus the post time are also changing. So we're learning about how this works. But I think the good part is when we look at those Cohort 1, 2 over the long haul, we're seeing them come down very substantially in their total starch requirements and some of those patients are -- lives are pretty changed as we showed you on Slide 11. Even if you're not down all the way, it was a big enough change that patient felt, he was going to take a picture and send it to his doctor. So that tells you how important it was to them.

At this time, I'm showing no further questions. I would like to turn the call back over to Danielle Keatley for closing remarks.

Thank you. This concludes our call today, and a replay will be available soon on our website. If you have additional questions, please contact us by phone or at [email protected]. Thank you for joining today.

Ladies and gentlemen, this concludes today's conference call. Thank you for participating. You may now disconnect.