Krystal Biotech, Inc. (KRYS) Earnings Call Transcript & Summary

All right, everyone. I'd like to thank everyone for joining us today at the Goldman Sachs Global Healthcare Conference, and we're so pleased to be joined by Krish Krishnan from Krystal Biotech to give us a good update on where things are, where things headed into, there's really kind of, no pun intended, pivotal year for the company.

So Krish, let's start at a big picture level. Why don't we walk through Krystal's herpes simplex virus or HSV gene therapy platform, including a rundown of both pipeline candidates and ongoing clinical studies.

Sure. First, thanks for having me. Thanks for having the company. So Krystal's HSV-1 is a DNA virus with many favorable properties as a delivery vector for therapeutic genes, including other exogenous generic materials to cells. Those favorable properties are like large genome size, ability to transfect cells with a high degree of efficiency, nonreplicating nature. The fact that vector is episomal with a bell characterized immunization properties. All that make it a great backbone to redose, and I want to make that point, I want to stress redose genetic materials to cells. The virus has good tropism to the skin. So our founding hypothesis was that -- and there have been prior clinical studies with HSV, where they found the vector stays in the skin. And so this localization in the skin can be harnessed for derm indications, and that's how kind of how we got going. And we've evolved from the hypotheses to now realizing, hey, you could target epithelial cells in other organs, too. As you can see, we have a depth in skin, but we're also slowly trying to transition to pulmonary indications. So that's a quick rundown on the broad properties of our backbone.

Thanks a lot. It's a great introduction of the company. So starting out with your lead drug called B-VEC and the lead indication of dystrophic epidermolysis bullosa or DEB. Why don't we start by describing the condition of DEB, its prevalence and how it's currently treated and managed?

Yes. So it's in normal healthy people, the 2 layers of the skin, the epidermis and dermis are held together by what are known as anchoring fibrils. This is essentially a proxy for collagen VII. Now EV is a genetic condition that causes the skin to be extremely fragile and blister easily, blisters to rubbing, scratching like pretty much any kind of friction. And it's caused by mutations in collagen VII. That gene -- the gene is actually called 7A1, provides instructions to assemble this type VII collagen or anchoring fibrils, which holds the skin together. So when you have a mutation, it kind of disrupts the production of collagen VII and the 2 skin layers tend to separate. DEB is a segment within -- it's one of the segments within the broad EV indication, and it's got both recessive and dominant. And the symptoms blistering of the hands, the feet, the knees, the elbows, sometimes if you have a severe condition, it could lead to vision loss and disfigurement. But one of the biggest risks from this disease is people with DEB have a high propensity to develop a form of skin cancer called squamous cell carcinoma of the skin, which is unusually aggressive and life-threatening. And so even without a treatment, there are no treatments for this disease. People spend quite a bit of money, like between $200,000 and $400,000 on palliative approaches just to delay the onset of squamous carcinoma. And roughly speaking, and although our commercial team is honing in on prevalence and incidents, we have always said it's about 10,000 patients worldwide faced with this disease.

Okay. Great. So following from that, given the kind of description of the pathology of the disease, what would you consider the most clinically meaningful endpoints for individuals afflicted with DEB?

Yes. So yes -- so they are open wounds, right, as I mentioned. And so given the nature of the disease, I would consider complete wound healing or complete wound closure would be a very meaningful endpoint. The durability, like once it's closed, how long it stays closed to be a clinically meaningful endpoint. And indirectly, because if you durably hold the skin closed, you are slowing down the epithelialization and reepithelialization of the skin, which is what is regarded as one of the prime drivers of the squamous carcinoma of the skin. Now that said, besides closing and keeping the wound closed, there is a third attribute. If you make a meaningful difference, time to wound closure can be a meaningful attribute. And I'm not saying like because the skin sometimes randomly open and close, if your drug is able to close the wound faster, obviously, it's better for the patients. But there, the meaningful difference I mentioned is not by a couple of days, but if you're faster by a week to 10 days, then the patient sees that, like as opposed to 2 weeks versus 1. So I would say complete wound healing, durability of wound closure are the 2 primary clinical drivers and time to wound closure can be beneficial to patients if you have -- if you end up with a meaningful difference.

Okay. So with that in mind, can you start by walking us through the mechanistic rationale for how B-VEC works to treat DEB?

Yes, so in the DEB or open skin setting, well, there are a lot of cells in the vicinity, primarily the 2 predominant cells are fibroblast and keratinocytes, and HSV is one of those vectors that transfects both fibroblast and keratinocytes, which is not the case for AAV or lentivirus or other types of vectors. And these 2 cells line the basement membrane zone of the skin. And essentially, when you transfect what B-VEC does, as I mentioned, it's an episomal virus, it's in the nucleus of keratinocytes and fibroblast where because of the genetic information, the collagen VII gets secreted, which then enters the basement membrane zone and then aligns itself to form anchoring fibrils. And if you look at some of the mechanistic data we showed in Phase I/II, we believe we generate robust anchoring fibrils because we show AF is being anchored to both the NC1 domain and the NC2 domain, something we believe that no other company has ever shown a functional robust anchoring fibril with both domains. And that's kind of critical. So that's essentially the mechanism of action. It's localized, transfects the cells, collagen VII secreted in the BMZ. And then the vector itself, the skin cells turn over pretty fast. So they stay in the nucleus until the cells divide a couple of times, and then they get eliminated. And that's how the vector is eliminated from the system.

Okay. And just for clarity's sake because this is important based on what you just said. Since collagen VII is a secretive protein, it matters less which cell type gets transduced and expressed as collagen VII as much as the field of cells gets transduced and expressed and secrete collagen VII.

Right. I totally -- I agree with that point. The only comment I made was if you are transfecting cells that line both sides of the basement membrane zone, then you are kind of having collagen VII come from one end and the other end, and they can align themselves nicely to form anchoring fibrils.

Excellent. Great. So following from that, why don't we start with the Phase I/II, the GEM-1/2 trial of B-VEC in DEB. So could you please walk through the existing clinical data for your HSV program and based on the Phase I/II?

Yes. So I'll summarize it and pretty much safety, efficacy in detail by wound is out there on our website or in publications. But essentially, it was a 90-day intra-patient, randomized, placebo-controlled, repeat dose study that was both dose escalating and increasing with respect to frequency of application. So we had to figure out what that would be for the pivotal trial. The study predominantly enrolled RDEB patients following discussions with the FDA. [Technical Difficulty] It was studied at Stanford. There were no safety concerns in the study. We've presented a lot more safety data at SID recently. An important point to note about the study is the patients who were enrolled and treated, a majority of them like we would expect the normal populations had baseline antibodies to HSV-1. So it kind of shows there was no impact to tolerability, efficacy, redosing. But from a pure efficacy perspective or wound healing or complete wound closure, we saw clear separation from placebo at weeks 8 and 10 and 10 and 12. We did not have any placebo wound that closed completely at these endpoints. At least, if not higher, 80% of the B-VEC wounds hit that endpoint. And when you look at -- when you retrospectively, statistically analyze either using Cochran-Mantel or McNemar, the separation was striking, and that's a slide in our website. The time -- so that's with respect to both keep closing the wound and keeping it closed. Once the wound stay closed, even though we didn't follow the patient following the Phase I/II study, anecdotally, we've heard wounds have -- some wounds have stayed closed as long as 6 months. And from a time to wound closure perspective, we were meaningfully different, like at least, I think the average time to wound closure was half that of placebo, that got a very high level. It varied by wound, of course. So all in all, it was a very positive study for us and set the stage for our design and the ongoing pivotal study.

Excellent. So with that Phase I/II data in hand, could you kind of -- in some level of DEB describe the Phase III, GEM-3 trial, including aspects of trial design, dosing schedule, dosing decisions and then key clinical endpoints?

Yes. So Phase III is a double blinded, randomized, intra-patient, placebo-controlled study. We enrolled 31 DEB patients. When I say DEB, I'm using them as a proxy for recessive and dominant, both adult and pediatric as young as 6 months. Enrollment completed, I believe, at the end of 1Q 2021. Data anticipated in 4Q 2021. So patients had a primary wound pair chosen for randomization. The wound pairs were anatomically similar. One was randomized to B-VEC. One was randomized to placebo. Wounds were dosed once a week until the wounds closed. The dose that we administer is a function of the size of the wound. But on a per centimeter squared basis, like PFU per centimeter squared, it was consistent between Phase I/II and the pivotal trial. So the doses were administered until they closed. And if during the study duration of wound were to open, then we would redose until it closed again. And that goes on for about 6 months. The primary endpoint is looking for complete wound closure at weeks 22 and 24 or 24 and 26. And the reason for the 2-week is that's by FDA guidelines. If you look at guidelines for wounds, chronic wounds, they look at durability as being closed at least for a 2-week time point. So that's how that end point came about. And because these wounds have a randomness to them, it's good that we have 22 and 24 or 24 and 26, it gives us a bigger opportunity when compared to how the placebo behaves. So that's -- and in terms of dosing, we have this concept of max dose per week. So every patient gets a certain amount of that dose gets consumed by the primary wound pair. And whatever is left is an open-label applied to other wounds on the patient. There in an open-label setting, we're able to track things like time to wound closure, duration of durability that we could potentially publish someday. So that -- and that concept is very similar to the way in logic works, where they have a max week per dose. They don't know how many tumors you have or how many injections or what the size of the tumors are. And it's kind of a very similar concept, so it's a well-established way of bringing a drug to market.

Okay. Excellent. So digging in on the statistical analyses that you described for GEM-3. So if you -- where are you recently announced a shift in the statistical test being employed as you mentioned earlier, from CHM (sic) [ CMH ] to McNemar. And admittedly, I'm not a statistics expert, but could you briefly describe the McNemar test and how it will be employed in GEM-3?

Yes. So McNemar is a stat test used on paired nominal data. Paired is easy to understand. We have a pair of primary wounds in each patient. So it's for paired data. Nominal implies that the result is not numerical, it's a yes or no. And what I mean by that is, is the wound completely closed at the endpoint, yes or no. So it's a perfect test. But so is Cochran-Haenszel, CMH. The only advantage, like if you believe there's any correlation between the 2 wounds in a patient, people tend to believe anything associated with the patient, something could be potentially correlated to the other, even though we're local administration, there's nothing systemic about our drug. If you believe -- if you think there could be correlation, then McNemar is a preferred way of analyzing it and could potentially lead to a better outcome if there was a correlation. We evaluate wound closure at week 22 and 24 or 24 and 26, which are 2 weeks on either side of a 6-month time point. And in McNemar, one tends to look at discordant pairs for the most part, which is -- which essentially means how many patients hit the active endpoint versus placebo. And while in placebo, how many opposite happens. So you're not looking for commonality, you're looking for how different is the active from the placebo. As I mentioned earlier, we had originally proposed CMH. But the agency, like it is difficult in DEB to figure out if there's a correlation between 2 wounds in a patient. There are no publications on it. And ours is localized. And when we applied both these tests retrospectively to Phase I/II, the results were very striking. And so we're aligned with the agency on McNemar.

Okay. Excellent. So in our conversations with investors around this use of McNemar, people have kind of considered the scenario in which there are a limited number of discordant pairs, where even if the study of GEM-3 were to succeed statistically on that test, will there be a risk that one might not view that difference is clinically meaningful. How do you think about this overhang related to the kind of limited numbers of discordant pairs? And what practically do you think is the minimum wound closure rate that would be considered clinically meaningful with or without McNemar as your statistical test?

So let's think sequentially about your question, which is stat multiple. First, it's a rare disease. We have 2 other companies that are working on corrective therapies in P3. And if you look at -- they've all aligned with the agency on enrolling less than 20 patients. So from a pure sample size perspective, we're at 31, right? So we're at least 50% more. So the agency is definitely aligned with the sample size of our pivotal trial, which is the question you're asking, what if the discordant pairs are not enough, will the agency say, this would not be clinically meaningful. So I think you can answer that by saying, not just us, but everybody else is aligned on the number of patients given in a pivotal trial, given that we're at 31. So that's one point. But if you -- then the relevance of the 31. If you assume 30 patients, a 20% dropout rate, which, thankfully, to date, has been a very conservative assumption. We announced that -- we actually said this in one of our slides, we have a 90% power to detect the 50% difference between the active and the placebo. So if you look at our Phase I/II, we saw a significant value on 11 or 12 patients because the active was so far different than placebo, especially with this high bar of complete wound closure which is difficult for the placebo to hit because it's a high bar. So we are aligned with the agency on the design. The McNemar on the sample size and all that. The McNemar on Phase I/II was positive. So if you think about risk in a pivotal study, the risk is only whether we would separate or not with respect to the McNemar. If we do, then it will be clinically meaningful as it shows that even in 31 patients, the active is robust, separates nicely from the placebo in a double blinded, randomized controlled study, which is the gold standard for pivotal trials, whether it's rare disease or not. And so the -- if you ask me, the other -- if the question was asked the other way that hey, what if you don't meet the endpoint, but the data is so significant and the disease is so rare, would the agency consider allowing the drug? That would have been a more pragmatic way of looking at it as opposed to, hey, you're going to separate on a well-defined study, will the FDA come in and say, it's not clinically meaningful. I think that's less of a risk environment.

Okay. Then beyond statistics, do you think it's important to consider some of the other aspects of GEM-3 that differ from GEM-1/2. So I think one that certainly comes to mind in investor conversations is the difference in timing from 12 weeks in GEM-1 to around 24 weeks in GEM-3, as you mentioned earlier. So what gives you confidence of the efficacy you observed in GEM-1/2 at around 12 weeks of treatment will translate into efficacy up to 24 weeks of treatment?

So two things give me confidence. An optimized dosing regimen and the ability to redose should the wound open. So that straight answers up and let me explain that. With regard to optimized regimen, in Phase III, we've taken the guessing out. In Phase I/II, because it was the first time and it's a topical gene therapy, we didn't know whether you had to front-load the dose or spread it over time. And so we had -- we dealt with varying frequencies to kind of figure out. But what we found is, no matter what the frequency, applying like if you look at Phase III, we keep dosing until the wound closes. And that, if you look at Phase I/II, happened in like a 2- to 3-week time point for an average wound. Average wound being between 10 and 20 centimeter square. Second, if you look at -- with respect to redosing, let's say, behaved exactly like the Phase I/II, and it closed at 90 days. If it opened after 90, we have the ability to redose following 90 days and then evaluate at the 120, at the 180 time point. So a simple way to think about it, the confidence is I view pivotal as 2 cycles of Phase I/II essentially. And the good part of the second cycle is when we say the wound opens, you can redose, it's not like they're fully opened to the original state before we start redosing. So even increment, if it's not 100% closed, we're going to start redosing the wound. So we think it will end up closing faster and that kind of what gives me the confidence that, well, the pivotal is not that different than the Phase I/II trial.

Okay. So another consideration, as you mentioned just now, is the dosing schedule in the GEM-1/2 trial was kind of patient-specific and kind of, as you say, it was an evolving dose escalation study. While in GEM-3, it is standardized weekly dosing up to wound closure. So what gives you confidence that, that weekly dosing schedule will achieve the same level of kind of effective exposure to mechanistically provide that anchoring fibral formation to ensure wound closure to the same degree that you saw in GEM-1/2?

Yes. So in Phase I/II, if you think about it, Patient 1 got essentially treatment 30 days apart. Patient 2 got treatment 15 days apart. Starting with Phase II, which is Patient 3 to 8 or 9 or whatever or 10. It was like, hey, let's try upfront 5x stop as opposed to spreading it out and see what happens. And what we found, so if you look at the Patient 1 that was 30 days apart, the wound closed nicely and stayed closed. 15-day apart, didn't make a difference. And so when you think about a commercial setting, you can't have like multiple -- so here we say, look, here is a max weekly dose. Some of it may be applicable to the primary wound on a given week. And whatever is left in the vial, feel free applying to secondary wounds in the patient. So given -- so what I'm trying to say is given like data in the Phase I/II that no matter what the frequency, whether it was once every 15 or once every 30, we saw the wound closing and showed durability of closing. And so weekly application, which is a slight variant of that shouldn't, in our opinion, have any meaningful differences.

Okay. So then another difference between the Phase I/II and the Phase III is, as you mentioned earlier, in the GEM-1/2, the patients were -- had the recessive form of DEB. While in GEM-3, 3 patients who are understanding have dominant DEB. So to what extent could the inclusion of dominant DEB patients affect B-VEC's activity in the Phase III trial?

That's good. So it is rationale to expect B-VEC to provide benefits in dominant patients because we have a high expressing system. We encode 2 copies of collagen VII and we make a lot of collagen VII that we think will potentially displace the nonfunctional collagen VII that you see in dominant patients. So that's the hypothesis for why we think the drug would work in dominant patients. That said, you're right, we don't have clinical experience with dominant patients in the Phase I/II study. So what we did, from an agency perspective, they did not force us to any specific number of dominant what we're going for the indication of DEB. We believe that if the data pans out, the label would run across recessive and dominant. But what we did -- so we did 2 things, given the lack of experience in dominant. We powered the study adequately, meaning to account for the dominant patients. So we had 31 patients. And the [Technical Difficulty] is that we don't anticipate the DEB patients to be more than 10% of total enrollment. Although we haven't, we are yet to disclose what exactly that number was, given the lack of experience and given it's not a big number and the study is being powered, that's how we're going about approaching that problem that you raised.

So actually that leads -- your answer leads to kind of interesting follow-up question, somewhat simple-minded question. What level relative to normal collagen VII do you think is necessary to provide kind of anchoring fibril formation in recessive patients? And then to your point about dominant patients, what level is necessary for displacement of a dominant negative form of collagen VII?

Yes. So what I've heard and read with respect to collagen levels is 50% of collagen VII is sufficient to not have the fragility of the skin or the RDEB. There are a lot of -- where one parent has both genes and the other one is mixed, or whatever. And so there's a lot of literature that it's 50%. Some KOLs talk about 30% of normal collagen, although I go, I have read, so 50% is a good number to hang your hat on. Now the only thing we know is we have 2 copies of the collagen. In our animal studies, especially early on, we actually did the animal studies, we had levels that were comparable, if not higher than normal collagen type levels based on animal models. So in general, we've always believed, but I don't particularly know the answer to your question in dominant, which is what percentage of collagen do you need to produce [Technical Difficulty] is anything north of 50% should help in both platforms or both types of the disease.

Excellent. So one last question about GEM-3 versus GEM-1/2. As you mentioned earlier, GEM-3 has this feature where if a wound that's closed reopens, you can be begin redosing, which, as we understand it was not necessarily the case in every patient in GEM-1/2. So to what extent do you think that could influence the trial? And how much could that contribute outperformance for B-VEC in the Phase III versus the Phase I/II?

No. I think, look, given -- it's a fair point. My view on the matter is that when the wound opens for redosing, it incrementally opens, right? It's not like -- it's not like 1 day the whole 10 centimeter squared is open. And so being able to apply a sufficient amount of dose on a skin that it just fortifies and [Technical Difficulty] it provides a high level of confidence to me that you would close within the time frame and stay closed nicely. That's -- and that's the way -- but in a commercial setting, I think, the dosing frequency is a function of the half-life of the collagen VII. And so it's about 30 days. So wounds that are expected to stay closed or, I mean, collagen VII is supposed to stay for about 90 days before you have to replenish it. So at a median level, we think dosing frequency will be a lot lower than once a week once the wound gets closed like anything durability would be 90 days, but we'll capture that in the pivotal study.

Excellent. So kind of shifting gears to more granularity. So can you remind us when you might expect top line data from GEM-3 given kind of the current pace of timing for the trial?

Well, we’re expecting that Q4 of 2021.

Okay. Great. So let's shift gears over to some of your other orphan skin programs, and we'll start with KB105 in ichthyosis. So could you please briefly walk through this program? And remind us where you are with the study? And kind of when and what we might expect from updates from this trial?

Yes. So KB105 is for the development of ichthyosis. Unlike DEB, which is fragile skin, here, it's a disease of flaky skin. It's essentially, the skin is open. It's a simple way of describing the disease because the area -- because of -- it's a mutation caused in TGM1 and when that happens, there is no TGM1 protein expression, which doesn't then let loricrin and loricrin and TGM1 kind of essentially close the gaps between the skin and keep the skin closed. So we developed KB105 as a topical application, transfuse keratinocytes, produce TGM1, repair this open skin or barrier and some of the secondary complications. So in Phase I/II, we had -- we showed great molecular data upon repeat dosing. We showed TGM1 being produced, a lot of them being co-located. But we were treating adults whose skins are extremely scaly. And while they moved nicely on a scale, meaning there's a scale used to determine before and after how it worked. When you look at a picture, it's not as prominent as DEB. DEB, it's very easy to see an open skin and a closed skin. Here, it's gradients of a super scaly skin. So we had good data on Phase I/II. We started Phase II, treating a larger area, about 10x the area we treated on Phase I. And the reason for the large dose is we then want to move -- we are working to move towards a pediatric indication where if the drug was given -- hypothesis being if the drug was given to an infant whose skin is not that much scaly, maybe the effect would be predominant, at least visually. Mechanistically, I don't think there's a difference between the way it works, but to be able to see on a scale. So that's where we are. We're close to announcing data on the fourth patient in the Phase II study. And along with that, we'll kind of describe what the next steps are in the program.

Okay. Great. So then shifting gears beyond the skin indications, let's spend some time discussing KB 407 in cystic fibrosis. And so could you start by walking through the mechanism for your HSV program in CF?

Yes, I mean, conceptually, we encode for the CFTR gene. But unlike topical, we're now delivering this through a nebulizer and the hypothesis being that we know that our vector transfects epithelial cells. And so that's kind of what started the program thinking, hey, maybe we could get the program into the lung. But that's essentially the mechanism of action. What our preclinical studies -- we went very systematically through the preclinical studies. We wanted to know that the vector made it through the nebulizer. We wanted to know that it actually -- it had a good dispersion in the lung tissue. We did some organized studies that were very positive for us. And then from a safety perspective, we conducted this broad NHP, nonhuman primate study about 39, 40 monkeys, primarily for safety but give us the opportunity to look at signs of efficacy, like is there enough of the vector in the lung, where in the lung are you able to repeat dose. So the preclinical data is kind of, we believe, is robust, and that's what's guiding us to move forward with the clinical study that is expected to start, like, I believe, in Q3, or so.

Okay. So also, I'll follow that up with a kind of naive question. So you mentioned before, there's good historical data about HSV infection of the skin. Is there like a kind of clinical data set that indicates people getting HSV infections into the lung epithelium, given the kind of experimental tropism that exists across the epithelial population?

Like HSV in its -- wild-type HSV has tropism to the skin, right? If you look at Krystal's approach, it's more of a gene delivery story. If we can get the vector close to a cell that it is very oriented towards transfecting, that could be keratinocytes, it could be a fibroblast, in the case of skin RSV cells in the case of pulmonary indications. What we're doing is essentially, we can get that vector safely close to the epithelial cells, then we believe the vector will transfect. It's an episomal virus. It behaves more like a gene delivery-type story. And so for us, the hurdle in the study -- in the preclinical was, a, does it -- does the vector survive the nebulizer? Will the virus survive the mucus in the lung? Will the -- once we believe that we can safely deliver this vector to the epithelial cell in the lung overcoming the nebulizer, overcoming the mucus, then we believe we have a really good shot of transfecting it and expressing the protein, right? That's kind of -- the thing about the negative history that people always remind me about gene therapy hasn't worked for CF. Well, there are reasons for that, packaging capacity limits, which we don't have; tropism issue to cells, which we do not have; immunogenicity or lack of redosability, which is not something the HSV has. And so you can actually make a case that -- you can actually answer that question very easily that if you figure out why gene therapy did not work for other gene delivery stories, you actually -- they serve as a great form for actually arguing why HSV would be a great vector for CF.

Okay. Cool. So shifting gears to the balance sheet. Can you remind me of Krystal's cash, cash runway and kind of what development programs that runway assumes?

Yes. So we ended 3Q with about -- 3Q, no. We ended...

In this pandemic, we're all losing track.

Exactly. Which is 3Q or -- $400 million in cash. Look at our pipeline, we've got a Phase III. We got a P2 in 105, starting a Phase I in CF. We have a couple of preclinical, 1 alpha, 1 antitrypsin, which is in early stages and other one methadone syndrome in skin preclinical. But the big optionality -- one big optionality on the story, which does not get -- I never -- barely get any questions on it, is the ability to use this vector in an aesthetic setting. The safety study was good. We're walking into an efficacy study from a pure science perspective, whether we deliver collagen VII or CFTR or whether it's collagen III, the science is the same. Now the markets are different, the application is different, the KOLs are different, the analysts sometimes are different, but we think it's a big optionality on the story that is underappreciated in my mind.

Okay. Excellent. So I'll end our conversation with a question we ask every company. And this one might be more trivial for Krystal given the setup, but what is the reason to own Krystal shares for the next 12 months?

Pivotal readout, KB103, I'll start with that. I'm just talking -- look, when I say to own Krystal's share more, it's based on data. So pivotal readout KB103. Phase I data, KB 407, hopefully. We know when we're starting the study. We haven't guided that we would be done, but that's kind of like a good thing to talk about. Same with 301, 105, or I forgot, which is not too far away. So we definitely have guided to 2 clinical readouts this year, possibility of 2 more if things work out and we don't run into some operational issues, which we don't anticipate. That's great. And the 301 is not just another pipeline. To me, it's a big optionality on the story that you have to kind of spend some time thinking about those indications. And that -- we're well financed. We have a great team. We're starting to think about commercial. So we're really on the cusp of positioning ourselves as a fully integrated rare disease company. So as the largest shareholder, I have to say, look, I think there are enough reasons to hold the stock.

Excellent. Well, thank you so much for your time this morning, Krish. We really appreciate it, and thank everyone for joining us today here at the Goldman Sachs Global Healthcare Conference.

Thanks for having me. That's super useful.

Thanks, Krish.