Zymeworks Inc. (ZYME) Earnings Call Transcript & Summary

Okay. Well, welcome, everyone, to Zymeworks' 2024 R&D Day. My name is Paul Moore. I'm the Chief Scientific Officer at Zymeworks, and it's my privilege to get things kicked off. Just a legal disclaimer slide and then an agenda for today, what we'll be presenting. So let's just touch on this just now. I'll start off the day by giving an overview of our R&D vision and how we're building our portfolios of multifunctional therapeutics. I'll then hand it over to Dr. Susana Banerjee, who will cover the gynecological cancers; Dr. Hatim Husain will cover lung cancer; and Dr. Jaffer Ajani will cover the gastrointestinal cancers. And the reason we have particularly highlighted these cancer types is that this is the areas that we're focused on with our 5x5 solid tumor portfolio. After that, Jeff Smith, our Chief Medical Officer; and myself, along with the KOLs -- Jeff and I will talk through our programs that our 5x5 programs, the progress we're making on those. And then that will be followed by Q&A session that Jeff and I will be joined with the KOLs. And after the Q&A session, we'll then move to our advanced portfolio in the second half. Myself and Alexey will start that off talking about where we're going beyond solid tumors in the fields of hem/onc and autoimmune disease. That will then be finished off with a presentation from our research leadership team, both our ADC leads, Stuart and Jamie, and then also Nina and Thomas, who lead up our multi-specific antibody therapeutics. We'll talk a little bit about where we're going moving forward with our technologies. Ken will then close with some remarks, and then we'll have a final Q&A session. So on this slide, wanted to introduce Zymeworks here to you, and it's mapped on a time line slide showing how Zymeworks has evolved from its inception in 2010 through to today and where we envision going beyond today. So as you can see, Zymeworks launched as a -- based on technology, EFECT platforms and Azymetric platforms, which were based on modeling that designed those platform technologies, those heterodimeric Fcs. And that ability to form bispecifics was recognized by the broad industry as a leading platform in which to build bispecifics. And that hatched many collaborations. Those platform technologies are still being used by many of these companies. What Zymeworks then progressed was its own lead molecule, zanidatamab, which was devised and IND was in 2016. And if you've been following Zymeworks story, you'll know that, that just got approval last month for BTC. And so it's a tremendous excitement and pride that Zymeworks has and actually having its first developed molecule been approved for patients. During this time, during its evolution, Zymeworks recognized the power also of complementing our expertise in multispecifics, which is mapped in the bottom. In the top, we're showing you our evolution in ADCs. We brought in a company, Kairos, a unit that had expertise in ADCs. And what we've done is we've continued to invest both in that group and in the multispecifics and really developed a world-class group working on ADCs and working on multispecifics. Two years ago, we had our R&D Day, our last R&D Day. That's where we described our 5x5 platform or our 5x5 was launched, which was the initiative to have 5 new molecules in the clinic in 5 years. Today, we'll talk about the fifth of those programs, the ZW209, completing our nomination of that portfolio. And that molecule is tagged to reach the clinic in the first half of 2026, demonstrating that 5x5 has actually advanced quicker than we had anticipated, about 18 months ahead of schedule. What I'd like to -- so then beyond the 5x5, as I mentioned, what we'll talk about is our advanced portfolio and that will be discussed more in the second half of the presentation, where we're really trying to push the needle even further on the novelty, the diversity of our multifunctional proteins and develop true -- potentially truly first-in-class molecules. What I would like to say is that core to both the ADC and the multispecific group at Zymeworks is our holistic approach to design our care in really designing all features. We want to develop molecules that have stand-alone therapy, but also have combinability with other modalities, other standards of care and potentially combinable with themselves. On this slide, demonstrating sort of the 3 buckets of assets that have evolved from Zymeworks. I mentioned the strategic partnerships on the left. You can see that these are still very active, and they have multiple programs that are now in the clinic. Not all of these have been disclosed. There is one that's been disclosed, the J&J type 2 CDC bispecific, which is developing nicely in the clinic. We also have, of course, our internally developed approved drug, Ziihera, which is zanidatamab. That is moving forward by Jazz. They described their path for that yesterday. So you can follow that through their communication. Now in addition, what we made the commitment to was to develop our only wholly owned candidates. And for those, we have 6 wholly owned candidates now, 2 that are in clinical trials already, 171 and 191, which you maybe are familiar with and we'll talk more about, 2 INDs planned in '25 and now we have 2 INDs planned in '26. The 209 program, which is the fifth of the 5x5, our DLL3 CD3/CD28 TriTCE Co-Stim2 molecule. And then also today, we're introducing our first autoimmune program that were designated for development, which Alex will talk more about in the second half, which is our IL-4 receptor alpha IL-33 bispecific. What I do want to do now is just spend a few minutes on a few slides talking about our technologies, how we are differentiated on the multispecifics, what gives us our competitive edge and what we believe is important in our philosophy and design of multispecifics, and then I'll do the same for ADCs. So a cornerstone feature of our ADCs is what's called our Azymetric platform, which is at the engineering level, that's a set of mutations that are easily -- that enable heterodimer formation either for Fcs or for Fabs. And what that allows us to do is to generate very robust antibody-like IgG-like bispecifics. The flexibility of that approach is such that it allows us also to wire in different Fc isotypes, we can wire in Fc modifications. And importantly, it's also compatible with linker/payload strategies. One of the beauties of the Azymetric platform is also its compatibility with high throughput screening, which really allows Zymeworks to screen many molecules because of the efficiency of the Azymetric to form an Fc heterodimer, you can screen many molecules in high throughput because you get a quality product coming out -- straight out from small-scale expression. Beyond that, the molecules are also very highly manufacturable and that because of an antibody like properties, the robustness, and that's one of the other appeals to big pharma on this platform. Regarding the actual different structures you can form with the Azymetric, some of those are shown on the right and sort of highlight the formats that are possible and some of the applications that we can use those formats for. When we think about actually then designing a multispecific, it's a very iterative process between biology and engineering. So first of all, we'll identify a challenge. And then what we'll do is design panel of molecules to reach that challenge, continue to engineer screen and finally get the molecule that does what we want it to do, whether that's transtargeting in the case of a T cell engager, cis targeting or biparatopic epitope targeting in the case of Ziihera or in the case of our new molecule, a bispecific inhibitor of cytokine, something that can block both cytokines simultaneously. Here is 4 of the sort of lead molecules that have come from the multispecific approach. You're obviously familiar with zanidatamab, which is on the top left. That's a biparatopic where we combined 2 different epitopes for HER2. What was unique about -- what is particularly exciting about that molecule is that it gave biology that's different than just a pure combination of 2 antibodies. And this is a feature of the bispecifics that we develop at Zymeworks. On the bottom left is our first T cell engager, where there, the challenge was to develop a molecule that had supported a better therapeutic window between tumor and normal. And we did that by playing around -- by modifying the valency and the geometry of the molecule and if we come up with a 2+1 molecule, that supported a very -- a much bigger therapeutic window that had been presently achievable for mesothelin. We'll talk more about that program later. If you sort of then go to the right, you can see the increased complexity that we're designing in our molecules. If you first look at the T cell engager on the bottom left, you can see how that's then progressed into a trispecific T cell engager where instead of having 2 binding sites for mesothelin, we've now incorporated a third binding site, in this case, CD28. And we've designed in such a way that, that activation is conditional. I'll talk more about that later. And then on the top right, you can see where we can also then make bispecifics. Similar, more like the zanidatamab where we can block 2 pathways simultaneously. And again, Alex will talk more about that later. But there, we're also seeing some unique bispecific activity that could potentially suggest that it's superior to the combination. I want to switch now to our ADC philosophy. If you're familiar with the ADC literature and conferences, you'll notice knowing that Zymeworks has got a good presence there. We're considered pretty much thought leaders in the field. And some of the findings that we've discussed or described on this slide. So one of the features that we've understood is that when you look at the history of ADCs and you look at the toxicity profile, what ADCs do is not necessarily increase the tolerability of a payload. What it does is it increases the efficacy of that payload. And that then brings into mind the focus you have to pay on tolerability and why different payloads can drive more toxicity is because they're inherently more toxic. And that's shown in the middle -- on the second bullet. Here, we're comparing different TOPO payloads, the Daiichi payloads, which are more moderate in potency compared to other antibodies. You can see that their tolerability is lower than what you can get with the less potent DXd. You also have to bear in mind that when you build an ADC, only 1% of that gets to the tumor. And then you also have to bear in mind that bystander activity is very important in how you can factor that into your ADC design and that the right payload mechanism matters. So in the case of HER2, you can see that picking TOPO as a payload was improvement over a mitotic inhibitor that's in Kadcyla. So bearing that in mind, we also think about the antibody holistically, not just on the payload and the linker, but we also think very much about the antibody to target how that molecule internalizes and that's all features that are embedded in our thinking when we design. One thing that we published this year was our own topoisomerase payload 519 that is a payload that was derived from an active functional screen of hundreds of analogs of Camptothecin. We screen those. We conjugated them and then we characterized them. And what we looked for were payloads that gave us the profile that we really wanted. So rather than just taking what was available, which some other groups can do, which -- such as exatecan and attach, we wanted something that was actually within the sweet spot of potency, bystander activity and how well it played as an ADC. That we landed on 519, and that's now embedded in all our topoisomerase inhibitor ADCs, and we'll talk about those more later. Actually, we'll talk about them now. So here is 3 of the antibodies 191, 220, 251. These are all 3 molecules that are destined for clinical evaluation. 191 is already in the clinic, targets folate receptor; 220, targets NaPi2b, a target that others have attempted to target with ADC strategies. We believe our design overcomes the limitations and challenges that they encountered. And then we have our GPC3, our third program, our third ADC, which is designed for treatment of hepatocellular cancer. We'll talk more about these molecules later. On the right, I can show you that we're not done with the TOPO. We see a lot of applications for that payload. We're excited about it, and we feel like there's other targets and other tumor types that can benefit from that payload. And what we're doing there is working on additional formats and targets that can be supported through that payload. One thing I'd like to just finish this section off with is how we pick targets, the KOL that we think about when we select targets. This is an example of ovarian cancer, where you might wonder, well, why did you pick mesothelin? Why did you pick folate receptor? Why do you pick NaPi2b? I think this slide really elegantly shows the evidence of why those are such good targets in a disease such as ovarian cancer. Based on the expression profile alone, this is transcriptional profiling. You can see if you order and look at the expression profile of different -- of targets that are being pursued, you can see that these 3 targets are the highest express targets across the broadest number of patients. So here, we have all 3 of those. Why did you pick ADC versus T cell engager? I think a lot of that's been driven by the biology, the precedent clinical comparisons you have with that. We decided to pursue folate receptors in ADC and NaPi2b as an ADC. There was more biologies driving that as an ADC candidate, more precedent for that. And in the mesothelin, it's such an exciting target that we wanted to pursue with sort of a contrasting modality. So we picked the T cell engager strategy there. We'll talk more about those later. One of the things that we also feel that we achieved by having this complementary approach to targeting ovarian cancer is that we can use these either in separate patient populations that may respond more to one than the other or potentially in the future, they could even be combined or combined with other standards of care. So that's another thing that's very important when you think about our molecule design, we think about combination. So with that, I'd like to just finish this part with just sort of summarizing our pipeline today. As I mentioned, we have the 3 ADC programs that are listed on the top when they won 220 and the 251 all with TOPO payloads, so we have a balance, we have the ADCs. And then, we're balancing that then with 2 T cell engagers, the 171 CD3 2+1 and then the 209, which is our TriTCE targeting DLL3 that we'll go into more detail today. We've been running that off now with a third therapeutic area or a third -- sorry, third modality venturing into autoimmune inflammatory disease, the 1528, which is a dual cytokine blocker. And that we have scheduled for an IND in 2026. So you can see we've got a nice cadence of IND filings. We had 2 this year, 2 next year and 2 in '26. So with that, I'm going to switch over, hand the mic over to Dr. Banerjee, who will present on our gynecological cancers and provide an overview of the unmet medical need and the therapeutic opportunity there.

Thank you very much for the introduction. I am Professor Susana Banerjee. I specialize in the treatments of women with gynecological cancers in London, U.K. Here are my disclosures. And what I really want to achieve in the next few minutes, just highlight to you the significance of the burden of these collection of diseases for women around the world. So we know that far too many women are diagnosed and die from these conditions. 1.5 million women diagnosed with gynecological cancers worldwide and far too many deaths, almost 700,000. And these are mothers, daughters, wives, girlfriends, grandmothers contributing to our lives and society. The burden is overwhelming, and we really need to move things forward quickly to try and eradicate and diminish the fear of living with these cancers. So the good news is that there actually has been advances in gynecological cancers. Ovarian cancer, for example, there have been advances since I would argue beyond chemotherapy and surgery since 2011, good timing in the sense that I started as a consultant specializing in that area from that time. But the real backbone of treatment here for advanced disease is platinum-based chemotherapy and the anti-angiogenic vascular endothelial growth factor antibody, bevacizumab. Yet despite these advances with the majority of women responding to chemotherapy, 70% of women relapse within 3 years of first-line treatment. And the overall survival rate is 45% to 50%. So we need better frontline treatments to improve outcomes of ovarian cancer and one of the largest breakthroughs actually was the recognition and discovery of the significance of BRCA mutations. Looks like there's a party going on as I'm talking. So, thank you, everybody. So I hope you can hear me. So the main thing is the significance of BRCA mutations and the significance of targeting PARP inhibitors. That's transformed care. That was a real paradigm shift in treatment with maintenance therapies. And by that, I mean targeting minimal residual disease when patients have responded to surgery and chemotherapy and then actually attacking disease at that time when you've got the best chances of having long-term -- longer-term therapeutic benefit. So this slide is now over 10 years old, but still highly relevant. It really shows you and reminds me to tell you that ovarian cancer is not one disease. It's heterogeneous. There are multiple subtypes yet, we tend to treat even today with the same modalities of chemotherapy. But there are multiple types and these are just some of the mutations and genomic aberrations that are associated with some of the most common types of epithelial and non-epithelial cancers. So this is a summary of the ovarian cancer treatment landscape in the U.S. and it's just really for me to guide you and illustrate where we are. Surgery is critically important in terms of removing all disease if possible despite having advanced cancers. And it's platinum and taxane therapy with bevacizumab I mentioned and the maintenance approaches with PARP inhibitors or bevacizumab. We know how significant the benefit can be in women with BRCA mutations or indeed homologous recombination deficiency, or HRD, and that's with a multitude of PARP inhibitors, but also in those women who have HRD negative disease, there is some benefit in terms of progression-free survival with PARP inhibitors. That's in the first-line setting. For recurrent disease, second line or beyond, we still divide treatment into so-called platinum sensitive or platinum resistant. What do we mean by that? We mean when cancer has come back within 6 months of platinum, that so-called platinum-resistant disease and if it's more than 6 months platinum sensitive. But I'd really like to highlight the point that unfortunately, to date, relapsed disease is not curable. And that's where we really need to make improvements in outcomes. Where we've got a chance of cure and extended overall survival would be perhaps in that setting, but again, in the first line setting where we can actually have a chance of extending and increasing the number of women who are cured with overall long-term remission. So for platinum-sensitive disease, platinum doublet chemotherapy with or without bevacizumab and a maintenance PARP inhibitor if a patient hasn't had a PARP inhibitor previously. In platinum-resistant disease, with single agent chemotherapy, potentially with bevacizumab depending on the situation. And really, there hadn't been much change since 2014 in that regard until recently, in the last year, with antibody drug conjugates. So in the U.S. and -- so FDA and EMA approval, we have mirvetuximab and in the U.S. more recently, for HER2-high IHC 3+ trastuzumab deruxtecan. So from my point of view, I think this is really exciting. It shows us the proof of principle that ADCs can change practice. The other important point here, I would say, is that it also highlights the significance, mirvetuximab targets the folate receptor. It shows the significance of folate receptor as a biomarker and really the proof of principle of targeting the folate receptor here. So I'm going to move on to show you some of the benchmarks. I'm not going to go through every study. These are some selected, and I'm not going to talk through all this. But for you to see the responses that critically important beyond response is progression-free survival and overall survival. We're going to focus really on the recurrent disease setting. As I've told you, recurrent disease is not curable. And let's have a look a bit more closely in platinum-resistant disease. Response rates with chemotherapy overall of around 12%. We add bevacizumab, it goes up. What I really want to highlight is how dismal the survival rate is around a year for women with this disease. And once cancers relapse, the inevitable is platinum-resistant disease at some point. So it's great that we've seen the progress with mirvetuximab, as I mentioned, FR targeting ADC response rate is higher, 42% versus 16% of chemotherapy, PFS median 5.6 months and this is the first drug to have caused an overall survival improvement in platinum-resistant disease. So again, proof of principle of an ADC, proof of principle of targeting the folate receptor, but there are limitations in the sense of mirvetuximab, this is limited to folate receptor high and that accounts for around 33%, so 1 in 3 women with this condition in this setting. So certainly room for improvement, not only in the numbers of efficacy and survival, but also broadening the population. So now I'm going to move on to uterine cancers, endometrial cancer and I'm focusing on advanced disease, surgery, again, really important, radiotherapy, but really to highlight here the systemic therapy angle for decades. We've been left with chemotherapy, carboplatin, paclitaxel, doxorubicin and hormonal therapy. But there's been an explosion, thanks to the clinical trials in this area of newer therapies. Immunotherapy has changed the practice in the last year or slightly more PD-1, PD-L1 targeting. And also the recognition of the biological diversity and the subtypes, MMR deficient, MMR proficient, for example, and other biomarkers. And that's where we're really heading. And also, as you see here with HER2 trastuzumab deruxtecan in IHC 3+ showing promising activity in a relatively small pan-tumor type study that we were part of, but really showing how the field is moving and important to say, improvements in a disease that was considered a Cinderella, if you like, in terms of clinical trial development a few years ago. So I really want to focus here on the unmet needs, which remain an increase in gynecological cancers. So focusing on ovarian cancer and endometrial cancer today. We need more personalized therapeutic approaches. Hand-in-hand with efficacy. By that, I mean survival, progression-free survival, overall survival and, of course, response rate to manage the tumor burden. We need better quality of life, better toxicity profiles, which can limit efficacy and better survivorship as we have more women surviving these cancers. More maintenance options and molecular or genomic factors understanding this in terms of drug resistance, primary or subsequent drug resistance. If we could then focus a bit more on the actual treatments in ovarian cancer. We need to look at specific groups, first line, HRD negative, where outcomes are poorer than HRD positive or BRCA mutated, PARP resistance to those patients that have a prior PARP inhibitor, platinum resistance is a real challenge. And also in endometrial cancer with a field that's exploded with immunotherapy, we need to look beyond immunotherapy and look also at primary immunotherapy resistance. Looking at how we treat post-IO progression, looking at targets beyond immunotherapy or targeted together with immunotherapy. And there's certainly room for improvement in the mismatch repair proficient population. And as I've highlighted already, thinking about combinations. So a lot of unmet needs. We need to improve outcomes for women with gynecological cancers. It's really exciting times treating patients in the clinic where we've seen hope with these latest advances, but also hope for the future. So with new technologies in drug development, I think the future is bright in terms of reaching our goal of eradicating some of the difficulties we have in these cancers and I think through optimization of, for example, ADCs, different platforms, there really is a chance. And here, we have Zymeworks with a real opportunity to address the unmet medical needs for women with these diseases. So thank you for your attention.

Okay. Thank you so much, Dr. Banerjee. And I'm going to extend on some of these points to speak about the current landscape of lung cancer, and we're going to go through a deep dive into how we typically think about this disease, both from our first line, second line, from a non-small cell lung cancer perspective and a small cell lung cancer perspective. So my name is Hatim Husain. I'm a medical oncologist at the University of California in San Diego, and I've been there for about 12 years. So here are my disclosures. So to start, actually, I really want to emphasize the significance of where this disease fits in the current landscape of the incidence and prevalence of the disease. We can see that across both male patients and female patients, lung cancer ranks among the top in terms of those patients who have the disease. And we have seen that there has even been increases in the rates of nonsmoking-related lung cancer globally. Furthermore, while the death rates for lung cancer have declined, largely due to patterns of smoking cessation as well as the introductions of targeted therapy and immune therapy, we can still see that the death rate in lung cancer is the highest among all the cancers that are listed. And this really has a significant global burden. If we really want to make an impact in how the global rate of cancer-related death impacts the world, lung cancer is a big target still. So along those lines, we have made tremendous advances in lung cancer. We can see over here that this is first-line advanced non-small cell lung cancer, particularly patients without activating mutations here. We can see that the introduction and early introduction of immune therapy really has set the stage for the treatment of the disease in the stratification of the treatment paradigm based around PD-L1 expression, where the higher rates of PD-L1 expressed tumors are more reliant on that pathway in treatment paradigms that can include a monotherapy approach or a PD-1 focus-centric approach really can define that treatment paradigm. As we have patients who have lower PD-L1 expression, the integration of chemotherapy really becomes a prime component. And we have actually even seen better responses in patients who have PD-L1 negative disease with the introduction of CTLA-4 in addition to PD-1 in addition to chemotherapy. What we have also seen is for those patients who have activating genomic oncogenes, genomic drivers, we have seen a plethora of investments and therapeutic benefit seen across a number of different targets. And here, we can just -- we can appreciate the magnitude of the drug development in the targeted therapy space. And I want to highlight the fact that while many of these agents are approved in the first line, what we do after that first or second line of therapy really has been an unmet need here. So while we've seen new drugs, many of these have been within the same pathway against the same target and new novel approaches are really needed here. We have seen some encouraging data looking at antibody drug conjugates even in the oncogene-addicted space. And I will actually think and reference the fact that I think that is kind of an active area of further exploration and development here. So this is a slide which depicts the landscape across first line, second line and third line. This slide does include actionable genomic alteration patients as well as those without actionable genomic alterations. I think the key to take away from here, though, is that while we have somewhat of a saturated first-line space with biomarker selection and active consideration across multiple different targets, as we go into the second and third lines, we can really see a clear unmet need. And this actually -- I was just chatting actually that this really is an important space partly because this is a place where an impact can be made. Patients, even in the second and third lines, are living long periods of time relative to the historical standards of where this disease was at, and this has been a space where antibody drug conjugates as well as T cell engagers have made a difference, and we'll speak about where some of those have integrated. So here shows some of the benchmarks for where the approved regimens are at. We can see response rates actually that can be upwards of 50%. You can see PFSs that span anywhere from 6 to 8 months, even in later line approaches. And overall survival actually that exceeds 1 year to 1.5 years across whatever line is being looked at, obviously, longer in the first-line space. Not only have we seen advances in non-small cell lung cancer, but we have also seen advances in small cell lung cancer. So historically, small cell lung cancer was more common than non-small cell lung cancer. But because of declines in smoking patterns, now non-small cell lung cancer is more common. Our current standard hasn't shifted dramatically in the first line. We have had the introduction of immune therapy and approvals for PD-L1 antibody in that space -- antibodies in that space. In the second line, the paradigm really can include and is focused in chemotherapy with the recent approval of tarlatamab in the second line. This I thought actually is an important advance because it is the first bispecific T cell engager to enter into lung cancer and has the potential to show a tipping point on how both the biomarker selected strategy with DLL3 as well as a novel mechanism with T cell engagement and the constructs created in that paradigm can really enter into the space. We can see that if patients have had a progression that is longer than 6 months, we can have the reintroduction of platinum-based chemotherapy. And then we can see other chemotherapeutic agents that are listed with different mechanisms that can span both alkylating agents, topoisomerase-related agents and such. I think it's staggering and kind of somewhat disconcerting point as if we go to third line and beyond, there's really not a clear preferred regimen as we advance into that space, really, truly kind of an opportunity for patients. So here just shows some of the benchmarks for current approved regimens in small cell lung cancer. We can see that while some of the regimens do have high response rates, the durations of response and median progression-free survival do leave more to be desired. And I think this is an unmet need where new mechanisms to integrate into this space, particularly around where the field has seen advances, both in ADCs and T cell engagement is an active area of interest. We have now seen also kind of the importance of some biomarker selection around active programs in the DLL3 space. So with that actually, I really want to highlight just kind of where we are at with the unmet needs in lung cancer treatment. We can see that the field has sent real close signal that personalization of the treatment strategy is an inherently important element of this disease. We have really been appreciative of the fact actually that patients are getting more lines of treatment now. So really, while the first line and the second line are where active areas of interest have been, it is important to think about and consider how the sequence of therapy and the applicability of second-line and third-line agents are considered. Survivorship and quality of life is an important element of the space. As patients are living longer, treatment decisions are based around anticipation of tolerance of the therapy and the quality of life that patients may have on the therapy, and that is more and more becoming an important part of the decision-making in our practices. How that reflects the maintenance approach for patients is also a key consideration here. So as we think about this, I think some other unmet needs are how are we looking at non-small cell lung cancer as a diverse histological subgroup. We know that squamous patients have underperformance on certain therapies compared to non-squamous patients. How are those patients being taken care of? How do we address brain metastases? These are active areas of interest and concern. And then really kind of small cell lung cancer is a true unmet need where we are, I believe, on a tipping point of where kind of we may see better outcomes for patients. So with that, I actually want to thank you.

Good morning, everyone. I'm very pleased to be here speaking to you about GI malignancies. I'm going to highlight certain ones and then focus on pancreatic cancer and hepatocellular carcinoma. So I work at MD Anderson Cancer Center for a long time. And over all these years, what I've seen is many patients with different GI malignancies are doing better. So examples will be colon cancer, MSI-high patients in different GI types. But what is really disturbing is the young people that are coming down with cancer. And I see -- I think you see in the news about colon cancer all the time. But it's all GI cancers and many other cancers where -- and we don't understand why that is happening. So what will be really important is to develop a blood test to find these cancers early and study the young individuals much more than we do now. These are my disclosures. So one thing you will notice is GI malignancies are very common. If we look at the totals, and this does not include all GI system cancers, that is over 3 million. And on the right side, a little comparison with other tumor types. And you can see the number of deaths that are occurring are very high. And when we look at the ranking of GI malignancies in terms of number of deaths compared to other cancers, it's all in single digit. So the -- even though I mentioned we have made some progress, a lot more progress needs to be made. The other thing is that pancreas cancer is listed as #6, but it is projected to be #2 cause of death in about 15 or 20 years. And the incidence is increasing rapidly around the globe. We don't understand what the reasons are. There may be similar reasons why young people are getting cancer. So now focusing on pancreatic cancer. I think if you just look at the left upper box, you will just see chemotherapy, although hundreds and hundreds of randomized trials have been done, but they were not -- not all of them are focused on any kind of biomarker or a driver of cancer. These are just throwing the dice. And that's how a lot of trials have been done, very little success. So we do have success in using chemotherapy, but as we all know, chemotherapy has limited level of activity you can get a response and then it doesn't last and doesn't really provide significant prolongation and survival. So what is -- the only so-called biomarker in pancreatic cancer is the germline mutations that you inherit from your mom and dad and that can be BRCA1, BRCA2, PALB2. These are very rare. But once we recognize those, those patients do well anyway. They're going to live long time, whether you treat them or not because the immune system is on top of those cancers. But if you treat them with olaparib, you can delay progression of those malignancies. So start deficiency of a biomarker. So in other words, next patient walks in with pancreatic cancer, you can't take the tumor and start looking for this and that. It is recommended, but we don't have any treatment to say to the patient or family. So here is a target we have identified, and we can use that to improve the outcome. So in the first line, most of these patients are doing poorly anyway because we don't have early detection strategies. So clinically, they have a lot of symptoms and it's been dragging on for some time from PCP to subspecialists and the diagnosis. However, there are some patients with reasonable performance status, minimum symptoms and they can get 3 drug chemotherapy. If they're not doing so well in the clinic, you give 2 drugs. And if they are doing very poorly, which does occur in the community setting in the VA hospitals, then sometimes they don't even get any treatment. When you come to the second-line space, it's basically the treatment that you did not give in the first line. You can use it in second line. And sometimes, you can drag it out in the third line. But the attrition rate is very high. So you start with 10 patients in the first line and you end up with 1 or 2 patients in the third line. So tremendous problem in the clinic with the cancer where the incidence is going up. So this is sort of a landscape based on what I mentioned to you. So pancreatic cancer often has KRAS mutation, and there's a lot of excitement in targeting KRAS, different types, either specific types of mutations or even pan KRAS approaches. And we hope that, that will be productive. But otherwise, what you see is let's just look at the overall survival is less than a year. So many patients are going to die within a year. The medium PFS is also very, very low, sometimes almost nonexistent. And the number of treatments are quite limited, as I mentioned before. So the next one I want to mention is the hepatocellular carcinoma. The incidence of hepatocellular carcinoma also is going up particularly in the southern states because of the migration from the countries in the -- below the southern border of the United States. They are highly susceptible to hepatocellular carcinoma plus also in Asia because of the viral infections. So hepatocellular carcinoma is sort of unique when you compare it to pancreatic cancer because in hepatocellular carcinoma -- in pancreas, also you can make that argument, but this is stark, where you have diseased liver. So you can have a liver disease because of multiple reasons, for a long, long time. So I'm talking about 10 years, 15 years. So there can be hepatitis. It can be the alcohol or obesity related. So you -- we have a chance to identify high-risk population and perhaps do early detection with various techniques, including blood test. So what you will notice here also is that we have made some headways in adding the immunotherapy agents, checkpoint inhibitors, PD-L1 or PD-1. And that is the first line, but this is not based on any kind of selection. So the next patient comes in, if they are eligible to receive those 2 drugs, bevacizumab, which the contraindications would be uncontrolled hypertension or active bleeding, et cetera. But most patients should be able to receive that. So you end up giving that. And then after that, you have 2 or 3 drugs, which are multi-kinase inhibitors, fairly toxic, lowly efficient. And as I mentioned, there is no selection based on -- you can't take the blood and say to the patient or family that we identified a target, and you cannot take the tumor and do the same. In the middle is the second line is -- basically depends on what you use in the first line. You can use some of the drugs again, and you can drag it out into the third line. But again, the attrition is high because of 2 reasons. One is they already got sick liver and there are limitations to what you can do with altered liver functions. And the second thing is, of course, hepatocellular carcinoma. So a lot of room for improvement. This again is the landscape. We'll look at the overall survival. This is a little bit better than pancreatic cancer. But again, majority of the patients are -- we're going to lose them within 2 years of diagnosis. And fortunately, some are surviving long time. So we don't get very high response rate. We don't get very long progression-free survival. And the limitation of treatments as well as limitation of smart treatment. So if you talk about colon cancer, gastric cancer, we can be smarter than when we are treating hepatocellular carcinoma or pancreatic cancer. So as I mentioned before, I think the most important thing will be to find these tumors early. So it's probably more possible for hepatocellular carcinoma. And some individuals argue that it's not going to be very useful in pancreatic cancer because many of these blood tests are detecting tumors when they are much more advanced. In other words, you can't just surgically remove and patient will be fine. That's not the case. Nevertheless, I think that line of research is progressing fairly rapidly. So we will see that there will be a blood test for all of us. There are blood tests if you want to pay from your pocket even today. So what is missing, particularly for pancreatic and hepatocellular carcinoma is a smarter approach. These are heterogeneous tumors being driven by different oncogenes and the tumor microenvironment is very complex, which is orchestrated by cancer cells and oncogenes. And we're not able to take advantage of any of that. So I think very empiric approach, very dissatisfying to the patient and family and we're regularly kind of disappointing when they find out what we can do and what they can expect. And if you tell a patient or family you can live 3 to 6 months to 12 months longer, that doesn't make anybody happy because patients don't want to know that I can live another 12 extra months because that's really not meaningful. So we hope that as we go forward, that we will become much smarter in treating these tumors. These are being driven by specific pathways, and we haven't been able to take advantage. So what you heard earlier today, hopefully, some of those molecules will make a difference and we certainly need a lot more research than we can imagine today. Thank you.

Yes, if you could. Sorry, password issues, but we should be okay in a second. I'll introduce myself. I'm Jeff Smith, I'm the Chief Medical Officer of Zymeworks. I just wanted to take this opportunity to thank Dr. Banerjee, Dr. Husain and Dr. Ajani for bringing us up to speed on gynecological, thoracic and GI cancers. It was very illuminating. And I'm going to talk primarily about ZW171 and ZW191, then hand over to Paul, who's going to tell you about the other 3 in the 5x5 portfolio that we have. Thanks, Paul. Okay. Well, having had those really good talks, I thought it would be important that we just outlined in a slide where we felt the 5x5, the different ADCs and T cell engagers in the 5x5 that we have would play in these different cancers. So here at Zymeworks, we aspire to address the unmet need in gynecological, lung and GI malignancies with our 5x5 pipeline. With these molecules, we aim to develop therapies that are tolerable, practical, combinable, improve survival and, more importantly, also improve quality of life of patients. Just a quick slide showing you 171 and 191. So we've, as you know, initiated 2 global Phase I first into human clinical trials this year. These trials are evaluating 171 and 191. 171 is a mesothelin CD3 bispecific T cell engager designed to enhance potential safety and efficacy. 171 has a 2 plus 1 antibody format that drives avidity-dependent mesothelin binding of 2 mesothelin paratopes to enable selective cytotoxicity to tumor cells versus normal tissues and minimize off-tumor on-target toxicities. Additionally, the 2 plus 1 format reduces the impact of soluble mesothelin on potency. Reduced T cell binding was incorporated to mitigate the risk of cytokine release syndrome. ZW191 is a folate receptor alpha targeting ADC. It consists of a proprietary topoisomerase 1 inhibitor payload with a drug antibody ratio of 8. So this slide just shows that mesothelin has high expression in gynecological, GI and thoracic tumors. And this just lists some of them. So ZW171 exhibits a wider therapeutic index compared to other mesothelin-targeted T cell engagers, and that's illustrated in the top right hand of this slide. The reduced T cell binding of ZW171 is evident in the bottom left graph, which show that ZW171 has lower T cell binding compared to the all other mesothelin targeted T cell engagers and demonstrating the tumor selective binding and cytotoxicity. ZW171 mediates enhanced cytotoxicity on high mesothelin expressing tumor cells and remains inactive on low mesothelin expressing healthy cells. So ZW171 mediates strong antitumor activity in patient-derived ex vivo and in vivo models. We see this activity in an ovarian cancer organoid model shown in the left panel. We also see activity in a patient-derived non-small cell lung cancer humanized mouse model, which is in the middle panel. And finally, in a pancreatic cancer humanized mouse model shown in the right-hand panel. So ZW171, a global Phase I study is currently enrolling participants. Part 1 of the study is dose escalation and allows the enrollment of various histological subtypes of advanced ovarian and non-small cell lung cancers. Mesothelin status is tested retrospectively during the dose escalation. Once a recommended dose is reached, we'll start part 2 of the study, looking at dose optimization and expansion. Part 2 will enroll ovarian cancer, non-small cell lung cancer and a basket cohort consisting of endometrial, pancreatic and others. This is a global trial. We have regulatory approval to proceed with dose escalation in the United States, United Kingdom, South Korea and we have pending approval in Germany. The global footprint we feel will enhance patient recruitments and also patient diversity in the clinical trial. Moving on to folate receptor alpha. ZW191 is Zymeworks' folate receptor alpha ADC, utilizing our proprietary 519 payload. And folate receptor alpha, as a lot of you are aware, is expressed in ovarian, endometrial and non-small cell lung cancer. This presents an opportunity to not only deliver a potential best-in-class therapy in folate receptor alpha-positive ovarian cancer, but also opportunities in the endometrial and non-small cell lung cancer. Our preclinical studies demonstrate that ZW191's novel monoclonal antibody drives superior internalization, payload delivery and tissue penetration compared to other folate receptor alpha targeting ADCs. We also see antitumor activity in a range of PDX models across a range of cancer types of interest. The ZW191 global Phase I study is currently enrolling participants. Part 1 of this study, a dose escalation -- is dose escalation allows the enrollment of advanced ovarian, endometrial and non-small cell lung cancer. Folate receptor alpha expression status will be tested retrospectively. Once the recommended dose is reached, Part 2a will consist of dose optimization in ovarian cancer and Part 2b will be expansion cohorts, including endometrial cancer and non-small cell lung cancer. Again, this is a global study. We have approvals to proceed in the U.S., Japan, Australia, Singapore and South Korea. And the global footprint, as we said before, will facilitate diversity and also help a lot with recruitment. So with ZW171 and ZW191 currently in the clinic and others planned for next year, there will be several strategic pivot points. For each molecule, we will determine the clinical development pathway based on safety, tolerability, early signs of antitumor activity as well as the competitive landscape. Based on these factors, we will consider accelerated development combination approaches, partnership opportunities or moving the next best asset forward. We anticipate submitting trial and progress posters at a peer-reviewed medical conference in 2025 for both ZW171 and ZW191. And we look forward to being able to share initial clinical data from our solid tumor product portfolio, starting potentially during 2025. And now I'd like to invite Paul up to talk about the other, in the 5x5.

Okay. Well, thank you, Jeff. Yes, as Jeff mentioned, I'll take over here and talk about the 3 other programs that are part of our 5x5. And first of all, I'll talk about 220 and 251, which are 2 additional TOPO1 inhibitor ADCs. ZW220 targets NaPi2b. What we've done there is we've selected an antibody that we believe has good internalization. We'll show you some data to support that. We've toggled in a DAR4 for this molecule. We realized there's been challenges with NaPi2b as a target with other ADCs. What we believe here is with the different payload, the TOPO payload, we think that gives us opportunity that was not -- that gives an opportunity for this molecule that was not -- that overcomes limitations of payloads previously used against NaPi2b. And then we've also toggled the DAR here to DAR4 to enable more of a therapeutic window that we can play with as we move the molecule into the clinic. Importantly, when we do develop our ADCs, we develop a DAR4 and a DAR8 molecule. So this was based on an extensive preclinical assessment, which, again, I'll touch base on in a future slide. One thing we also did with this antibody was we decided to silence the Fc. That's also -- has been hypothesized and demonstrate -- or demonstrated from preclinical data to also potentially lead to toxicity issues associated with Fc gamma receptor interaction. And we thought for this program and for this target, that was a choice that we decided to toggle in as well. For 251, that uses the same payload. That's a target that we are developing for hepatocellular carcinoma, as you heard from Dr. Ajani, there's a big need for new treatments there. We've got some -- a lot of exciting data, preclinical data for that molecule. And we really feel that this could bring benefit to patients with HCC. A couple of slides. Well, first of all, on NaPi2b, as I showed earlier in ovarian cancer. It's one of -- it's among the highest expressed cell surface targets. It's also very highly expressed in endometrial and non-small cell lung cancer. Other areas that are within our focus range. Regarding activity for 220. We've just shown you 3 pieces of key data on this slide. On the left, you can see the efficient internalization of the antibody that would be selected for NaPi2b to go into 220 compared to the antibody from Mersana and from Roche, previous antibody targeting drug conjugates against NaPi2b. And you can see that 220 has a very competitive profile for internalization. On activity in models, you can see -- we've just shown you 2 examples. We get very nice activity with 220 both in ovarian PDXs and lung PDXs, and we've expanded this analysis to multiple PDX models. What's particularly impressive about this molecule, however, is its tolerability and in nonhuman primates. Here, we have shown you the non-GLP tox data. There, you can see the doses that we treated. We treated up to 90 mg per kg for ADC, TOPO ADC, and that was well tolerated for that molecule, suggesting that we have indeed got the opportunity to have a bigger tolerability profile for this molecule than prior ADCs against NaPi2b. On the next slide, I have a similar slide for 251. This is the HCC target. GPC3, I should mention, is a target that if you profile hepatocellular and look for cells over targets is a very attractive for that indications. Others are exploring that target in the context of other modalities, but we are really the first or one of the first to really push this target forward as an ADC target. That is really based on our impressive preclinical data. Again, on the left, you can see the ADC -- internalization of the antibody is very efficient. And then that then translates into efficient in vitro cytotoxicity and what we're showing here is very efficient activity in in vivo models, again, using PDX models. And this is just 2 representative model showing activity in a span of HCC expressing PDX models. Again, here, what is also important, of course, is the tolerability profile. And, in this case, we actually went up to 120 mg per kg with this ADC and again, you can see that the antibody -- the ADC was well tolerated and gave us nice dose proportional PK as expected for an ADC. So for those molecules, they're both slated for 8 INDs in 2025. We're on track for that. We're going through the final stages of IND-enabling studies, and we look forward to reporting more on them as they progress. What I'd like to do now is actually switch a little bit to the fifth program that we're nominating today. This is our DLL3 CD3/CD28 TriTCE, so it's a trispecific T cell engager. As you heard from Dr. Hatim that this is a target -- DLL3 is a target in small cell lung cancer, where there is indeed a therapeutic need for better modalities. Here, what I'd like to do first is just talk about what is it that we've done with this T cell engager that is the advance. So if you follow the T cell engager space, certainly, there's been challenges for T cell engagers in solid tumors where due to low T cell infiltration or due to T cells that are exhausted, you don't have the substrate or the T cell to really drive the killing of the tumor cells that you're trying to do with your T cell engager. So this has really gained attention to adding in co-stimulation, the so-called second signal, so you'll get CD3 activation. And then what you also add is -- can you add in co-stimulation to energize the T cell or make the T cell more active. CD28 is a particularly attractive co-stimulatory signal for this for many of the biology reasons and for engineering reasons that are shown on the right-hand side of the slide. So this, of course, is not -- we are not the first to think of this. I mean it's been an active area of pursuit CD28 in the field. There has been companies spend effort on bispecific CD28 T cell engagers. Those have been evaluated and tested alone or in the combination of a molecule that can also provide a second signal, either regenerate the T cell with PD-1 or can actually combine it, in fact, with another T cell engager. So you have a CD28 T cell engager targeting a tumor antigen and you combine that with the CD3 separate molecule of CD3 T cell engager and that way you get both signals. Unfortunately, I mean, those programs have shown promise, but they do have limitations with toxicity that's seen with those. This is led to the elegant suggestion of generating a trispecific CD3, CD28 T cell engager. Others have done -- there has been some first-generation molecules there, but they likewise have suffered from challenges associated with too much T cell activation or cross-transbinding between the CD3 and CD28. What our solution was, was really to try and focus on generating a new format for CD3, CD28 dual targeting within a T cell engager that really balanced the CD3 and the CD28 engagement so that you've got the right balance when you engage both. And we did this in a way that the CD28 activation is purely conditional on the CD3 engagement. This profile, we believed, would lead to a more favorable molecule that could actually then be moved forward towards clinical treatment -- or clinical testing. So to get there, though, we had to do a lot of work. This was an extensive screen profiling. Again, I mentioned the ability to use Azymetric platform to screen many formats. This is what we did. What you maybe can appreciate here is a number of different geometries that we tested with the CD3, the CD28, the number of different affinities, the variations that we tested there, and that was an extensive screen and then that led to a format that we've owned that we could reproduce on multiple target antigens that gave us the profile that we wanted and that's the platform that we're moving forward. And that -- those key features of that platform are shown on this slide. What I will tell you is that cartoon that I'm showing on the left of the CD3, CD28 engager with the tumor antigen that's a representative format. That's not our secret sauce, but it does give you the concept of what we're trying to achieve here, okay? If you look on the right-hand side of the slide, this really lists the design criteria that we were looking for and what we've achieved. So we wanted a balanced activation of CD3 and CD28. So only -- you only get the right amount, the right balance of signaling there. You don't overstimulate on CD3 or CD28. They're both low affinity, but you have an avidity gain when you add the CD28. The binding for CD3 and CD28 is in cis, so there's no T cell to T cell bridging. And that engagement of CD28, as I mentioned, is conditional. This then leads to a molecule by having that CD3/CD28, has the safety profile you want, but when you actually then use it to treat or test for cytotoxicity, it has an enhanced activity. So with that molecule platform, what we've decided to do is pursue DLL3 as the first target that we're going to evaluate with this platform clinically. We think that, that target has a nice exciting profile. It's in a therapeutic area, there's unmet need. And also, there's, of course, now the precedent of T cell engagers against DLL3 work having meaningful benefit. What we believe, however, is that with this molecule that we can extend that benefit, have more sustainable -- potentially more sustainable responses, and potentially treat a broader breadth of patients with small cell lung cancer. What I'll do -- a slide missing. Okay. There was a slide -- Okay. Well, essentially, the slide that we had was a slide that demonstrates the activity of DLL3 in vitro. It shows the binding -- preferential binding independence of addition of CD28 to get additional activity with that molecule. What that has really though translated into is an increased benefit in antitumor activity. What you can see here is a comparison against tarlatamab, AMG 757, and you can see in these models that we have increased benefit of efficacy compared to that molecule. In the safety profile, we have looked at that as well. That's, of course, a big key thing here for ZW209. We look for cytokine activation. That's a big question mark. Will you get cytokine relaxation with PBMCs alone? And what you can see here is in testing in the sort of typical in vitro cell-based assays, you don't see any activation there. This also -- then we also tried and tested the molecule in humanized mouse CRS model. And you can see that no systemic release of cytokines. So this is a model where you've humanized the mouse with T cells, human T cells. And so you can see the positive control there works, the anti-CD28 control, but you see no activation with our molecule. And then furthermore, we've taken this program into nonhuman primates. We've gone up to 10 mg per kg with this molecule. And while we see some mild spike, we see some mild spikes in select cytokines. This is not a significant level, and we feel very manageable from a preclinical perspective. With this data, we decided to take this forward and move towards preclinical IND-enabling studies. So just to wrap up, the DLL3 molecule is summarized here. The key design features that I mentioned are a trispecific T cell engager that now has incorporated into it, the CD28 co-stimulation. We think of this advance something similar to what the advance was in CAR-Ts, where by adding in co-stimulation, you've got the potential to move the needle and get better responses because when you activate the T cells, you have that second signal, which then gives T cells more prolongation of survival and activity and sustained activity. That binding, though, was very carefully dialed in by incorporating the CD28 in such a way that it only engages the T cells upon first binding with CD3. We believe that mechanism of targeting tumor cells with CD3 and CD28 will translate into additional benefit than you would see with just C3 engagement alone. And right now, where we are, we see this as a first-in-class molecule. We're using the co-stimulation incorporation. And we think that can lead to increased durability of responses in DLL3 expressing cancers such as small cell lung cancer. And we're on track for an IND submission for this in the first half of 2026. So thank you.

That's great. Good morning, everyone. Ken Galbraith, CEO of Zymeworks. I think we're going to get started with our first Q&A session. If our presenters would come up, that would be great. While they do so, I really appreciate all of you taking the time to attend our session in person and get an update on our R&D efforts that we've been working on since our last R&D day in 2022. We sincerely appreciate these 3 amazing physicians taking time out of their practice to come join us today to share their thoughts on their practices and managing patients in the areas they spoke about today. So we'll take a pause in our program to give you a chance to ask some questions of these folks here. They will be available at the end of the presentation after our formal remarks for more detailed discussions. But if you have some questions that you'd like to put to the folks here, they're available now. If you would like to ask a question in the audience, please put up your hand. We'll bring you a microphone. If you could please identify yourself for us, that would be appreciated and for the people online. And if you would please tell us who you're directing your question to, that would be fantastic. We have provided the availability for those who are following online and not in person here to ask some questions. So we will try to take some of those as well. But this is your opportunity now to ask some questions of 3 amazing physicians here up on our panel as well as we've just disclosed for the first time today our fifth of our 5x5 strategy, ZW209 and, obviously, Paul is happy to address questions related to that as well. So are there any questions? Great. Well, thank you.

It's Jon Miller from Evercore ISI. I guess a question for the Zymeworks team. Just on what we just heard about the CD28 containing TriTCE. I'm curious if you would expect that the most obvious delta versus Amgen's DLL3 to be on durability given the mechanism of action CD28? Or would you expect to be able to see benefits from extra T cell stimulation on first scan ORR, for instance?

Yes. No, that's a great question. I think we anticipate the potential for both, to be frank. Direct antitumor activity, we've observed in models I showed you there. Obviously, again, you're right that when you think about the biology, it is also the sustainability, is really the beauty of having that co-stimulation there. So I think that should be reflected also in durability of response as well as initial response. And I think we do sort of think of parallels with CAR-Ts that when no initial CAR-Ts didn't have secondary signal, you observed improvement there, and we would anticipate the potential for a similar profile and increased delta that was observed with those modalities.

We'd love to see that missing slide.

Yes, yes. Yes, maybe we can find that. Okay.

Yes, sorry.

No, I think it's a really essential question. And a really important question is that if we look at kind of where bispecifics are moving, we think about, is there a way through a more selective interaction to both increase efficacy and also mitigate toxicity. So I will say that perhaps kind of with some of the considerations on how, now we think about different iterations of bispecifics, how can the construct also affect in limit CRS to the capacity that it can. If we have patients who are not being able to get through the step-up dosing, that's a rate-limiting step for them to be able to get the product and kind of move forward. So I think with changes and kind of key reflection on this point, both from an efficacy standpoint, whether it'd be ORR or duration, but also how we think through limiting the toxicity, I think those will be the magic combination to move it forward and to create differentiation.

Other questions, please raise your hand. We'll get you a microphone, yes.

This is Brian Cheng from JPMorgan. For the doctors on the panel, this is a 2-part question. So one is -- so it's a fair comprehensive overview of what the current treatment response and the duration of response that you showed today. So I'm just curious, for new products who are entering the space of your -- of the respective indications that you talked about today, what did they need to show in terms of particularly the response rate and the durability of response? And then secondly is how translatable is the mouse data that Zymework has shown today that you think will be translatable to clinical output? And -- or -- and also, do you have kind of a go-to mouse model or NHP model that we can kind of glean on to get a sense of potential clinical efficacy?

So who would like to answer Brian's question? Please go ahead.

I'm happy to start. Just we can take it in turn. So I hope you can hear me. So I think we all are familiar with looking at response rates, and we love seeing tumor shrinkage and patients love it, too. But the reality is that some of these responses are short lived and that's not reflected in the response rate. And that's what really is more meaningful. So duration of response. But equally, I would say -- and it may differ from different tumor types, but I would say, delaying progression, delaying worsening of the disease, I think stabilization of cancers or some degree of shrinkage, but not necessarily the RECIST 30% is highly important for patients living with cancer and also delaying the time before a need for a subsequent line of therapy. And along with that goes the tolerability. I cannot emphasize that more because we, as a community, have always focused on responses and efficacy. But with newer drugs, also as, hopefully, people are living longer, we need to equally think about toxicity profiles. It's not a secondary aspect. It's -- I would say, it's together because that's what's limiting efficacy, getting enough drug in for enough duration.

And I would agree with a lot of those points as well. I think in lung cancer, we typically think of about a 5- to 6-month PFS and about a 10- to 12-month overall survival is being key benchmarks that we think about. I think in an ORR standpoint, about somewhere between 25% to 30% ORR has led to FDA approval and the PFSs and -- I mean, PFSs are kind of the ones that have led to FDA approval as well in second-line spaces. So those are some of the benchmarks that I think about. When it comes to preclinical models, I think that more and more having patient-derived preclinical models, I think, are important. And there isn't necessarily one cell line or one kind of mouse model, in particular, that one -- that at least I think about, it's more about are there adequate numbers of preclinical models that are demonstrating efficacy? Are there patient-derived models? And more and more, we are integrating from kind of real-world scenarios, plasma-based models or blood-based detection strategies to give us better clues. And I would say it's more of a composite picture that determines on-target efficacy. And I think we've seen this now across in lung cancer, both the IO space as well as the targeted therapy space. So I think it's great, actually a new drug development as from what we've heard, the integration of some of these strategies early and upfront to inform the strategy. IO in lung cancer has taught us that clinical responses may not exactly correlate with the model that was used preemptively. So I think that in the integration of these trials, seeing how both on-target efficacy is assessed, blood-based detection and blood-based strategies for surveillance and monitoring are integrated. I think those are going to be important elements as well.

Yes, I can just add a few more things. So I think it's good to see responses with a novel compound. But as mentioned earlier, it's not all that important sometimes, especially if you are engaging the host immune system, sometimes the progression-free survival is more important than -- so progression-free survival, I think is a better measure of an activity, but you have to have enough patients because as you heard over and over, in each tumor type, there is considerable heterogeneity. So if you study 10 patients, you will have a different conclusion. But if you studied 100 patients of the same tumor type, you will get a more complete picture. So if the immune system is being engaged, then you can have prolonged stability and that will be reflected in PFS. So if you had 10 patients and 1 responded, then you will say, oh, there is only 1 response. But if those 10 patients, 8 of them had prolonged progression-free survival, that could translate into benefit in overall survival. So I think that is one way to think about it. So the days of chemotherapy are almost gone. Very few cytotoxic agents are being developed. So now we are focusing on targeted therapy and immune therapy. And I think going forward, that is probably the best way. Another way to think about it is the -- how the cancer is surviving. First of all, it's got this immense intensity to survive in the body of the patient and it's using human genome to do that. So it knows what fibroblast can do, macrophages can do, T regulatory cells can do, CD8 killer T cell can do. It knows all that. It can read all the signals. So if you think about cancer on one side, driven by oncogenes and immune system on this side, there's a lot of traffic of these molecules. And they are all target of therapy, in fact. So you need to target the cancer cell, target the intermediary and then do something to enhance CD8 cells. A lot of CD8 cells inside the tumor are not tumor specific. And they are either exhausted or they're reacting to infection or inflammatory molecules rather than to cancer. So we have a lot to learn in that respect and have to kind of have more broad global approach to cancer. And one drug can make a difference. You saw the agnostic approvals. You get 50% response rate sometimes, but that's not the common phenomenon in the clinic. Most of the time, it's a very complex system you're dealing with and it's not going to be one drug that's going to make a big difference. So...

Thank you for the question, Brian. Other questions?

I think -- go ahead, Steve.

Yes. Steve Willey from Stifel. Maybe a couple of physician questions and a quick one for the company. So Dr. Banerjee, you have a lot of different ADCs that are being developed right now in the gyn space. Just curious how you think about whether or not these agents become available at some point in the future, things like CDH6, B7-H4 folate receptor, all with TOPO payloads. How do you think about this sequencing of these agents? And what are the implications for the competitive landscape?

Right. So I think it's great that there's lots of different potential targets. We've seen the example with folate receptor showing the sign that we can get a meaningful activity. I think we need to develop all these different targets, but then there's different concoctions within terms of the antibody with the different payloads, for example, in linker. And not all of them, I predict, are going to be as effective as each other so -- but we need to test these. And as you probably know, I'm involved in a number of the clinical trials across ADCs for that very reason because it's that balance of efficacy plus toxicity. So some may well end up scoring highly efficacy-wise, but are limited effectively from toxicities. And although we may have several drugs or compounds that may be looking at the same antibody targets, receptor target, that doesn't mean that they're all going to have the same efficacy because of the payload, for example. And to your point about when that payload is -- or that drug is being used because I do think the resistance mechanisms are going to be very much dependent on what the tumor has seen previously. So we may be able to, and I think we will, sequence a number of -- let's just take folate receptor, for example, depending on what the payload is and the resistance mechanism may well be different. It might be due to down-regulation of the receptor. That may be one aspect, which is very important to consider. That's where biopsies are going to be key to have a look as in pre-starting the next line of therapy, but it may well be actually payload resistance. So the -- when we design trials, I would say, beyond the initial dose escalation, dose-finding stage, I think it's really important to think about the sequencing when we design those parts. And it would differ between different tumor types and different settings. But I can't -- I don't think we can say just the prior number of lines of therapy, for example, what needs to be -- what does that cytotoxic look like, and therefore, the resistance mechanisms to that. And extending from that about potential combinations to try and overcome a resistant mechanism. I'm hoping that we see the day where when we have a newly diagnosed patient, we do a panel of looking for targets. For example, as we do for HRD and BRCA at the moment, folate receptor going forward, but other markers as well. For example, you mentioned CD86, Claudin-6 as well, TROP2, for example, HER2. And so that we have more information for an individual patient upfront, taking into note that this may well change over time, potentially depending on down-regulation of certain receptors. But then at least we have that information upfront. For many of these biomarkers, there isn't direct overlap. So you end up having different niches of patients within a tumor histology group, for example. And therefore, the sequencing is likely to be also influenced by that lack of overlap of certain biomarkers. So we're not there yet, and that's what we need to do alongside the clinical trials in the Phase I space at the moment looking at different targets. I hope that gives you some indication.

Okay. Next question, Steve?

For Dr. Husain, so it looks like small cell is kind of becoming the proven ground for the combination of both ADCs and bispecifics in terms of CD3 engagement. So just kind of wondering what your enthusiasm for that approach is.

Yes. No, I think it's -- for me, actually, I'm quite enthusiastic about this. I think that small cell is a unique beast. We spoke about the unmet need. We spoke about DLL3 as a biomarker. We spoke about the fact that there is active drug development around this target, both for ADCs and T cell engagers. And we've seen responses in both strategies. I think one of the promises that I see with the bispecific approach versus an ADC in this strategy is as we look around some of the landmark analysis, we do see that perhaps -- I mean, I think kind of the vetting ground in the space will be around durability of response. And across an ADC strategy or bispecific strategy, how will the durability play out is a key differentiator. And I think if there are strategies actually that can be more combinatorial or such, I think perhaps we can rely on the benefits and advantages of each in that regard.

Do you have a third question, Steve?

Yes, sorry, just quickly for the company. I know the prior versions of NaPi2b, ADCs probably aren't the best surrogates to benchmark yourself against, but I know that there was some suggestion, I believe, of on-target off-tumor toxicity in the lung that was seen, I think, both with the Mersana compound and with the Genentech compound. I'm just curious what level of confidence you have from the preclinical data that you generated to date that you're avoiding this?

Yes. Good question, Steve. That's definitely a point. I think there was more of that issue with the Mersana drug. I think with the Roche drug, I think, the -- that didn't report such toxicity. But certainly, with the Mersana drug, there was a lot of toxicity associated with their programs. And I think there, they did dissect that to some degree. And certainly a contribution to that was the high DAR and the first-generation payload that they had, that type of payload in the way that, that payload was attached to the linker and just how that high DAR molecule behaved that led to bleeding, which was actually off-target toxicity, they believe. And then they had a second-generation molecule that they also saw toxicity with. For us, when we look at that data, we really believe there's a -- that there's a limitation in that payload linker strategy that they deployed and really with the payload linker strategy that we deployed with different payload, different stability of linker, we see -- we don't anticipate that limitation with our strategy. So we are confident in the design. And then I think that's also then reflected in the toxicity profile that we see in nonhuman primates. We're actually also doing RAT studies. And so far, we've been able to go up to a very high doses without any evidence of tox or tissue damage there. I think what we also did was we did learn from that, when we were making the decision on NaPi2b molecule, thinking about the DAR. We could have done DAR8 or DAR4. Actually, our DAR8 molecule also was very well tolerated, but we did make that decision in this case just to sort of obviate potential challenges you have. There's no doubt that NaPi2b is expressed in normal lung. And so -- but we feel that we have -- we've navigated that both by the payload choice, the DAR and then I also mentioned the Fc modulation. So we've pulled all the toggles that we can and that's led to a very nice tolerable profile in preclinical models.

Next question?

Yigal Nochomovitz from Citi. So it's kind of following on Steve's question, but more specifically on the DLL3, obviously, we've got a lot going on in the approved drug, tarlatamab. We have your new one, the triple. We have the ADCs where there are several. So I know you probably can't really answer this because we don't know all the answers yet. But in terms of how you're thinking about this, will it be that you look at the particular patient like you mentioned and that patient should be an ADC patient or a TriTCE patient? Or will you take the view more that there's a one approach, which is sort of the preferred approach and that's kind of where you go first line? And then if you get a response, great, and then you try ADC, for example, next. Can you just kind of be more specific about -- because in small cell lung center, how do you see this evolving?

Yes. I think if all agents were approved and available and if I saw a small cell lung cancer patient, I think it's really a great question to figure out kind of how would one take that next step of what agent to choose. I think the reality is that with the current bispecific, tarlatamab, how we think through the step-up dosing and also how we think through the access-related issues can really define how a patient can get one medicine versus another. ADCs have been deployed in outpatient clinics fairly easily. Bispecifics require a little bit of investment in terms of logistics, in terms of how to build out that type of program in center. I appreciate the point around in current bispecific design how one thinks through taking toxicity into consideration as a distinguishing point even in the construct of a bispecific because as we deploy these, obviously, having agents that are more easily tolerable, more easily able to get up through the step-up dosing strategies, how some agents may have less CRS versus another. I think those are key points that will facilitate active differentiation and kind of further beyond efficacy across the portfolios of new agents coming into the space. I think in terms of ADCs, one of the -- obviously, solid tumor oncology has embraced ADCs to some degree around -- as we heard HER2, other ADCs kind of commercially available across multiple targets and multiple solid tumors. I think the challenge that I see in this space in small cell lung cancer is that the PFS and durations of response haven't been as long, in my opinion, with the ADCs for DLL3 compared to what we have seen with the duration of response with tarlatamab and I think in that context, I think we do take that into consideration. And if we can get a patient through and into a center that can do the step-up dosing and the bispecific, I would favor that approach over an ADC. If a patient cannot go through that or is not within range of the center that can do the bispecific administration, then I think ADCs would be the way to go. And ultimately, in my slides, I mentioned how patients aren't really getting beyond the second line. I hope and I think actually that maybe patients will get multiple lines and be able to integrate the ADC as well as the bispecific. Hopefully, that answers your question. I think if I could -- I would prioritize the bispecific over an ADC, but this really relies on access, I think, as well.

I think we have time for one more question in this session if...

I'm happy to. I already went. So if I can jump in...

Ren Benjamin, Citizens JMP. I'd love to understand what might be the biomarker strategy ultimately when you're developing these drugs. So you have these studies that are ongoing right now. What sort of -- so what's the expression level? Like, how are you going to determine the expression levels that you'll ultimately treat, right? And is there an entry criteria right now? Or is this something that you wind up determining, right, from these studies? Because I imagine that would be regardless of the target would be very, very important.

Paul, do you want to talk about that?

Yes. I'll answer that question first. Okay. So yes, we certainly believe the target is important. I mean, I think I showed you the data that drove our selection of the targets that we're pursuing is that we believe that they're highly expressed in the indications that we're good after with that. And so our approach is going to be that we're going to -- for targets like mesothelin, folate receptor, NaPi2b, DLL3 and for Glypican-3 in the disease indications that are in, they are already highly expressed. So we still will though follow those targets, the expression profile of those targets during clinical trial, Phase I. So our plan will be to analyze all the patients that we test for expression of the target. That's the obvious biomarker to use. We will deploy immunohistochemistry methods to do that. That's typically what's used in clinical trial studies. There are other approaches that you can look for biomarkers that we're aware of and can deploy. So we'll really let the data drive us there. For the targets such as folate receptor, I think, from our preclinical data, one of the attractions there is to go deeper than MIRV. So MIRV has a subset of patients that it treats. We're hoping that we can go to a deeper pool of patients with our design and our antibody. But at some point, we will also appreciate that we do have to understand which patients have the best benefit. And by having multiple targets within a certain therapeutic area, that does afford us a possibility in the future of maybe having the right molecule for the right patient, right? So becoming more personalized. That is potentially possible. Obviously, we're hoping each of the molecules will benefit most patients, but we appreciate that, that's not the reality. That's not how biology is driven. That it should be based on the target. We're not believers in just target not being important, if that makes sense.

Go ahead, Jon.

If I could expand?

Yes, you may. Yes, you can do that.

For the physicians on the panel, given that all these targets were chosen for widespread expression given that, obviously, in early-stage studies, you don't expect to see a ton of patient selection. What do you want to see in sort of that early to mid-stage development transition? Would you prefer to see programs be more heavily gated on selection? Or would you prefer to see programs dominated by broadest possible patient inclusion criteria? What would excite you in terms of patient enrollment?

So I can mention there are ways to deal with that. But for ADC, you don't need a lot of enrichment of the target. HER2 is a good example. But for antibody, you probably do. But even for ADC, if you look at the Claudin ADCs, there seems to be a threshold. So you certainly need that. In terms of approaching without biomarkers that is happening in the vaccine space where they're just going after 10 to 12 TAAs and sort of nonspecifically stimulating the patient's immune system and hope that the cancer has the target, some of them are selecting the TAA and others are not because what they chose was commonly expressed TAAs. So that's a nonspecific strategy, but we would like to see more specific strategies. And then the gradient, et cetera, one has to figure out.

Other quick comments?

Yes, I would just say, actually, I think in lung cancer, it's kind of been a really remarkable process to view this because some ADCs actually, I think, really do require some level of selection and some may not, right? And I think I've been surprised actually that ADCs that have a response rate of greater than 60% or kind of 50%, you don't need selection. An example of that, even across some cancers is trastuzumab deruxtecan, kind of where tumor agnostic indication and gyn malignancies and in lung cancer and other cancers kind of high response rates. The TROP2 ADC kind of with a lower response, it's been really a reflective experience to see how it brought to an all-comer patient population. First, there was histological kind of divides across non-small cell. And we're speaking about datopotamab deruxtecan in that context where non-squamous versus squam. And then now kind of BLA filing more for just the EGFR population. So really kind of getting into a very specific niche with that ADC. So I think actually that one has to reflect on the target. If the target has a response rate kind of in that range of about 30% to 40%, I think a layer of biomarker selection in that range is probably important. And whether that be through digital pathology or kind of how that may be done, I mean, there's many techniques that are being explored, none that have been prospectively defined. And I think that's going to be an area for the growth here in the ADC component around the biomarkers. It's how we prospectively define novel approaches, perhaps utilizing digital pathology and other kind of strategies like that.

Just to add to your point about the clinical trial design. I think that at an early stage, it needs to be broad because we have seen responses in patients who have tumors that have low level of expression as well. We've even seen that with HER2. The other complexity here is that we just -- we're looking at one biopsy at one point in time. And we know by taking biopsies in multiple areas -- let's just take ovarian cancer, for example, that the level of expression may vary. So that's really complex. It's the whole notion of the heterogeneity. So therefore, to completely base on an expression level may be too simplistic, and that's what we've learned. So therefore, I think in the development of a trial strategy, we need to look at more broad, but also look at cut-offs and then take that forward.

Okay. Well, thank you for those questions. That will conclude the first of our 2 question-and-answer sessions. While our panel exits the stage, if you could show your appreciation for the physicians on our panel who've traveled to be with us today, including one on the red eye flights, we really appreciate that. And so we'll let them exit the stage. We'll continue with the program and talk about Zymeworks' research over the past 2 years beyond solid tumors. I do remind you that our 3 physicians on the panel will be available at the end of our prepared remarks. If you didn't get a chance to ask your question, you have additional multipart questions, I'm sure they'd be happy to engage in a longer discussion with you at the end of our prepared remarks. But I think we'll pass the mic back over to Paul and talk about the second part of our prepared program. So thank you.

Okay. So I'll kick off the second half of the presentation. As Ken alluded to, what we're going to talk to now is about what else we've been doing behind the 5x5 portfolio, what we call our advanced portfolio, where we're focusing on really trying to push the envelope on the multifunctional nature of our molecules really to try and focus -- develop novel first-in-class therapeutics. This is because we really -- as you've heard, there's a lot of factors that can contribute to disease state in cancer or in other diseases that I'll allude to. And so having molecules that can tackle or overcome axes that are in play may bring benefit beyond what we've seen. And as you know, we've shown with zanidatamab bispecifics, by [ en grouping ] that multifunctionality into a single molecule can also uncover biology that can be leveraged to have antitherapeutic activity. So our goal here will be to continue to build on our ADC cell engager cytokine engineering, our expertise there. And then what we want to do then is to extend this novelty and diversity that's in those molecules. And later, you'll hear from our ADC leadership team. They'll talk about the novel payloads that we're looking at and evaluating beyond TOPO. And then they'll also talk about the targeting moieties and how we design those targeting moieties to really leverage selective binding and powerful internalization for our ADCs. On the cell engagers in cytokine engineering, then Nina and Thomas will cover those and talk about how we're using that Costim platform and other platforms to other targets beyond the DLL3 target and further advancing other engineering capability to really drive the best-in-class molecules. Okay. Within the therapeutic focus, as Ken -- as we alluded to, we're going to talk about what we're doing beyond solid tumors. What I will say though is that we're still very interested and focused on select solid tumor indications that I'll talk about shortly. But we also want to expand our technology beyond solid tumors to hem/onc, to liquid tumors and also into autoimmune and inflammatory disease. And I'll give you a flavor of why we think those are opportunities. So this slide really maps out the sort of sphere of tumor types that we've been focusing on. We tend to think about focusing on those tumor types with the poor survival -- with the poor 5-year survival. And then red, these are the target indications within that sphere that we're going to be pursuing with our advanced portfolio. In blue, you can see also the indications or the molecules that we've actually already developed and where they sit. So you can see they really primarily sit within that sphere. Within -- so then within that sphere, though, what you can see is that beyond solid tumors, beyond the GI tract and the thoracic cancers that we tend to focus on, there are other opportunities in the hem malignancy space. And here, there are sort of really 3 buckets of malignancies that we feel could benefit. In particular, AML has really got maybe the -- has the poorest survival amongst the hem malignancies. And there, in AML, there has been some precedent for T cell engagers and for ADCs to have functionality there. But we feel like with our technology and with the improvements that we can design, that there's opportunity to actually take that to lift that bar and really make more meaningful difference there. Within the B-cell malignancies, multiple myeloma and non-Hodgkin's lymphoma, there's obviously been great progress there, but there's still patients who don't respond to treatment or relapse. And can we then actually overcome that by the design of our molecules. So that's a sphere that we will be focusing on. Beyond oncology, the obvious other sort of flip of oncology is autoimmune disease. So often in oncology, you have an immune system that's being sort of underactivated and so you're trying to boost the immune system. In immunity, you've got a hyperimmune system and you're trying to dampen the immune system. So what that opens up is overlapping pathways and the applicability of molecules to sort of flip the switch. So instead of downplaying or upplaying the immune system in oncology, you want to downplay the immune system in autoimmunity. And so a lot of the thinking, a lot of design of molecules can be extrapolated to autoimmunity, many of the biological models that we use. The modalities themselves can be reprogrammed for that purpose. And importantly, also at Zymeworks, we have knowledge base and expertise. Jeff and myself have worked on the development of autoimmune molecules as well. So we have that knowledge base to take the molecules through preclinical development to IND and then into the clinic and through clinical development. Within the autoimmune space, the sort of a slide here that's sort of kind of buckets where we see some opportunities. We're going to be selective in where we go with our modalities and our technology. It's a little bit simplistic side, but you can think of autoimmune disease and inflammatory disease, hence the AIID. Within the autoimmune disease, there, what you tend to have are immune responses driven by lymphocytes or self-antigens. And that population, if you could modulate that population -- potentially deplete that population in some cases or reprogram that population, you may have a chance to sort of overcome and really drive meaningful benefit to patients with that type of disease. And we think of lupus, RA, type 1 diabetes as disease indications that fit that opportunity. In there, that space, that's where we see the potential for B-cell depletion or other immune cell reprogramming being compatible with Zymeworks' next-generation molecules, including our T cell engagers. So for example, in something like B-cell malignancy, could you get global B cell, pan-B-cell depletion with the TriTCE Costim molecule that targets, say, all B cells or do you want to actually engage a molecule or engineer a molecule that can be more selective and take out a single population, the autoreactive B cell? And we're thinking about that. The other approach that you can toggle with a bispecific is actually to embrace the idea of blocking multiple pathways can actually lead to benefit beyond what you can achieve by just blocking one pathway. So in this case, we think about inflammatory diseases where cytokines are driving disease physiology and molecules that could block those cytokines, but block multiple cytokines because blocking only one is not enough, could actually then see benefit from a new generation molecule. And that's where we've actually been working quite extensively for several years at Zymeworks. You may be aware, we had a collaboration with LEO Pharma. And there, we had developed, at Zymeworks, a lot of antibody specificities against various targets. And what we've done is we've continued that work ourselves quite intensively and that's led to the development of a few bispecifics, one of which Alex is about -- will talk about, which is an IL-4 receptor, IL-33 bispecific that we see benefit for in COPD, and Alex will talk more about that. So Alex will be next on the podium. And then following Alex, Jamie and Stuart from our ADC team will talk about the ADC advances that we're pursuing. And then as I mentioned Nina and Thomas then will take over the podium and talk about progress in the multispecifics and our protein engineering.

Great. Thanks, Paul. Good morning, everybody. As Paul mentioned, my name is Alex Berezhnoy. I'm Director of Immunology at Zymeworks. And on behalf of the team, I would be presenting you key highlights for the ZW1528. That's fine. Anyway, we believe that bispecific antibodies could be the answer to the complex biology of autoimmune and inflammatory diseases. And we see it as a combination on 3 reasons related to the patients, to their advances in clinical science and to our technology. Autoimmunity and inflammation could be serious difficult-to-treat diseases. And they affect in large population of patients. Many of those patients can benefit from advanced therapeutics, but actually, very few of them receiving them. The biology of autoimmune disease is relatively well understood and there is a number of clinically validated targets, including the targets with our demonstrated benefits of co-blockade. Unfortunately, the unconveniences and high costs of that co-treatment prevent access to that potentially the most effective approach for multiple patients. At Zymeworks, we excel at making bispecific antibodies. We have that validated as a metric platform that could be utilized to create bispecific with a native IgG-like geometry and, therefore, can have potential advantages such as good manufacturability and extended half-life. We can create therapeutics that both are highly efficient and convenient and also cost effective. And additionally, as Paul mentioned, our team has very extensive experience in this area, both from internal expertise as well as our prior collaboration in the space. There is a strong rationale for selecting IL-4 receptor alpha as the anchor arm. This receptor is an important target that transmit signals from 2 cytokines, IL-4 and IL-13, both driving Th2 inflammation. And importance is really illustrated by the clinical impact of its blockers such as Dupixent, for example, on patients with atopic diseases. It's important to know that pathology of autoimmune diseases is complex and the multiple cytokines contributing to the progression. For example, IL-33 is driving non-type 2 inflammation in respiratory diseases. And that represents a problem in the clinic, but it's also an opportunity to combine to specificities into a single molecule and this way create a medicine that could be potentially more efficacious than they are currently available one. The challenge here is to overcome a loss of that blocking potential than we move from bivalent antibody format to a single-arm blockade of the bispecific. And as you can see in a few slides, that challenge was successfully solved for Zymeworks' advanced protein engineering approaches. IL-33 is another cytokine involved in airway inflammation. It's a tissue alarming. It's released in response to epithelial damage, and it acts on a range of cells, including immune and non-immune population. It's known to contribute to their initiation and amplification of inflammatory cascade in COPD and other respiratory diseases. Essentially, it's maintaining that chronic inflamed state, and it also may be involved in tissue remodeling. So all in all, we believe that our -- especially because there is evidence of clinical activity of IL-33 blockade in our -- in patients with COPD, we believe that adding the IL-33 blocking capability to ZW1528 can significantly increase therapeutic potential for this molecule. This slide illustrates current clinical trajectory in treating patients with COPD with advanced therapeutics. Anti-IL-4 receptor alpha, Dupixent and anti-IL-33 molecule, itepekimab, are being developed to treat different COPD populations. Itepekimab may prove to be effective in non-type 2 diseases, but only for a subset of patients. Dupixent can act on other populations, but only for type 2 disease. If we can make a bispecific that can combine effect of 2 antibodies in a single molecule, it provides the opportunity to treat a larger set of COPD patients, and it also might increase efficacy in those who are responsive to monotherapy. This slide highlights key features of ZW1528. It is a bispecific antibody with our native IgG-like geometry and it's constructed with -- as a metric platform based on IgG4 Fc backbone with the incorporation of half-life extended mutations. The protein is highly manufacturable and it's compatible with our high concentrations for small volume administration. One arm binds to IL-4 receptor alpha, another arm binds to IL-33, the cytokine. And binding to IL-4 receptor alpha blocks to cytokines, IL-4 and IL-13. And because the receptor is expressed on cells that are enriched in the area of Th2 inflammation, that also may retain some of the bispecific within that inflamed area rather than binding the cytokine and move into the circulation. The second arm neutralizes IL-33 and we designed the molecule so it can utilize that advantages of that local retention in the inflamed tissues in terms of advantages, in terms of high local concentration and potentially shortening the time between IL-33 release and its neutralization. Next few slides would illustrate core properties of the molecule. Here, we tested in vitro blockade of the cytokine-induced signaling. We used our reporter cell lines. ZW1528 effectively blocks IL-4 signal with a potency similar to that of term dupilumab, you can see red and blue. And the small insert here illustrates similar IL-13 blockade. On the right side, we show IL-33 signal inhibition, also in the engineered cells with the potency similar to that of the clinical benchmark antibody, itepekimab. Well, conclusion here is that 1528 demonstrate potency similar to that of benchmark antibodies. But unlike those benchmark antibodies, it can effectively block both targets. We extended that characterization to natural primary human cells. And at the left graph here, we are looking at our activation of monocytes with IL-4, and we measure that activation by their IL-23 of regulation. Potency of 1528 was similar to that of dupilumab and that's a clinical benchmark built on IgG4 backbone isotype. However, the potency of ZW1528 was much superior to that of IL-4 receptor blocking antibody with IgG1 factor negative backbone. And this is in agreement with previously demonstrated role of Fc interactions. The graph on the right illustrates the experiment where primary human cells from healthy donors were activated with -- for IL-33 resulting in the interferon gamma release. Itepekimab here blocks that activation with a potency that is similar to that previously shown potency in the engineered cell lines, while dupilumab in blue doesn't have any effect. ZW1528 demonstrated significantly increased inhibitor effect and that potentially reflects benefits of their bispecific blockade. Now we moved in vivo and we investigated molecule for pharmacokinetics and IL-4 receptor blockade in the live system. And in Tg32 transgenic mice on the left, ZW1528 demonstrated expected biodistribution and extended half-life. The experiment on the right was performed with mice, humanized by the IL-4 and IL-4 alpha receptor blockers. And these animals received systemic administrations of test articles and they're challenged with 12 inhalations of the house dust mite, the allergen for about a month. And as expected, the inhaled allergen exposure resulted in a significant increase of serum IgE in that untreated animals. Administration of 1528 suppressed that allergen-mediated IgE response, and there was a dose-dependent manner and the potency of the molecule was similar to that of dupilumab. So this slide illustrates pharmacokinetics and biomarkers of target suppression in nonhuman primates. On the left side of the slide, you can see pharmacokinetics profile of 1528 after a single intravenous administration. And then in the middle, we show our in vivo target engagement, which we analyzed by the receptor occupancy assay. And on the right graph, we demonstrate prolonged reduction of serum IgE compared to the [ molecular ] control. And that reduction was pretty long in time after a single administration. So conclusion here is that the molecule demonstrated IgG-like pharmacokinetics in nonhuman primates, biomarkers of IL-4, IL-33 blockade up to 6 weeks after a single administration. We extended that characterization to the models where both receptors are expressed. And one such example such cells are keratinocytes and those cells could be activated to express CCL2, chemokine, which is also known as a macrophage chemoattractant protein 1 that is involved in attracting pro-inflammatory cells to the site of inflammation. An addition of IL-13 here or IL-33 resulted in enhanced expansion of CCL2. And moreover, there are a combination of 2 cytokines, enhance that activation even more. In this model, the monospecific blockade of IL-4, for example, can partially reduce CCL2 expression, which is driven by these 2 cytokines together. 1528 potently blocks keratinocytes activation, almost completely suppressing CCL2. And of note here is that the potency of the molecule was superior to that of the combination of the antibodies, potentially reflecting the benefits of co-localized dual target blockade. At this slide, I would like to mention other programs where pursuing -- using that validated alpha receptor alpha blocker. The first one I just mentioned, ZW1528 is currently in the IND-enabling studies and additionally, we are working on ZW1572, a bispecific inhibitor of IL-4 receptor alpha and IL-31. Both targets are relevant for atopic dermatitis with expected benefits of dual blockade. We're also developing a next-generation therapeutic molecule, targeting pathways, driving asthma and IBD, and we believe that these 3 molecules can represent significant opportunity for Zymeworks to create a difference -- make a difference for patients with COPD, with asthma and with atopic dermatitis. And just as an example, this is the one side summary for ZW1572, the bispecific inhibitor of IL-4 receptor alpha and IL-31. The molecule is utilized in the same IL-4 receptor alpha arm and also the Fc region, just like ZW1528. However, the second arm was replaced with a high potency IL-31 neutralizing arm. In the middle graph, you can see the example of in vitro blockade of IL-31 signal and 1572 demonstrated superior blocking potency as compared to the bivalent clinical benchmarks, an antibody and bispecific. On the right side of the slide, we show that experimental data gained using keratinocytes, the human skin cells that naturally express both receptors. The cells can be activated with additional IL-13 and IL-31. And while the combination of 2 cytokines drives even higher expression of CCL2, here antibody can reduce CCL2 expression to a certain degree, but they cannot fully block it. On contrary, the equivalent amount -- same amount of the bispecific demonstrate superior suppression of CCL2 leading to the complete blockade of that immune activation. And in conclusion, the ZW1528 potently blocks 2 complementary pathways of respiratory inflammation. It represses type 2-driven inflammation by blocking IL-4 receptor alpha. It adds inhibition of non-type 2 inflammation driven by IL-33. And it shows favorable pharmacokinetics profile and expected PD biomarkers in nonhuman primates up to 6 weeks after a single administration. We also noted preliminary evidence of advantage of the bispecific blockade compared to the combination of 2 antibodies and that's potentially reflecting benefits of co-localized dual target blockade. ZW1528 is a stable IgG-like bispecific molecule. It's potentially easier to manufacture, and it's designed to support subcutaneous administration and less frequent dosing to increase patient compliance and convenience. We believe that 1528 and other bispecifics utilizing the same arm -- building arms are -- can make a meaningful difference for patients with COPD and other autoimmune diseases paving our way to this new therapeutic area. And with that, I would like to thank you for your attention and hand it over to Jamie.

Thank you very much, Alex. Good morning, everyone. I'm Jamie Rich, Senior Director of Technology in the ADC Therapeutics Development Group at Zymeworks. As part of our enthusiasm for antibody-drug conjugates, we're continuing to innovate beyond our topoisomerase 1 inhibitor platform. The goal of the Zymeworks ADVANCE initiative is to provide a portfolio of disruptive first-in-class multifunctional therapeutics, and a key element of this effort will involve discovering and applying novel ADC payload mechanisms that we pair with optimized antibody formats in a way that's considered both the disease and the target biology. ADCs that stem from ADVANCE will include both solid and hematological tumor targets and indications such as pancreatic ductal adenocarcinoma, colorectal cancer, esophageal cancer, head and neck squamous cell carcinoma and AML. So serious interest in the therapeutic potential of ADCs has now spanned on the order of over 40 years. And during this time, a lot of different payload mechanisms have been explored. Our figure in the upper left-hand corner of this slide is an effort to comprehensively outline the clinical successes, failures and ongoing ADC efforts as of August earlier this year. While emtansine, [indiscernible] and PBD ADCs have all scored approvals, it's the oral statins and to an even greater extent, the camptothecins that are the mechanistic focus of current clinical ADCs. So given this concentration of interest, which includes ZW191 in our own clinical pipeline as well as ZW220 and ZW251, we feel that there's both an opportunity and an imperative in the discovery of next-generation ADC payloads. So we're focusing specifically on mechanisms that should, at a minimum, fit well with the priority oncology indications that I've described previously. So we are interested in broadly cytotoxic payload mechanisms. This is the only class that has been applied successfully to ADCs to date. And it offers the likelihood of broad applicability across indications. But we're also interested in the more novel concept of conditionally cytotoxic payloads, which would feature -- could feature less normal tissue toxicity and provide potentially a built-in biomarker. Finally, we're also actively exploring protein degrader payloads, which offer an alternative to inhibitors and promise to expand the range of intracellular protein targets available. So we put a lot of thinking into the ADC mechanism, and we published some of that in recent years. Ultimately, our hypotheses around how ADCs function guide us in our approaches to novel payload discovery. Some of this thinking is outlined here. First of all, we have observed that conjugation doesn't improve the maximum tolerated dose of the payload. However, ADCs can improve efficacy over small molecules by altering the disposition of that payload. Furthermore, this efficacy can come through multiple different mechanisms, including tumor targeted delivery as well as contributions from free payload. Secondly, we're keenly aware that in comparing payloads from the same class, the more potent the payload is the lower the tolerated ADC dose will be. And we feel that by having a higher protein dose, we can exert some advantage. So for instance, in the center figure, the blue bars represent the dosing of various deruxtecan or DXd ADCs, while the orange bars represent the doses achieved by exatecan ADCs, which use the more potent camptothecin payload. Similarly, the other bars to the right of the orange exatecan bars represent ADCs with other highly potent camptothecin payloads that are emerging and currently a clinical focus. What's clear from this analysis is that the moderately potent ADCs achieve higher normalized payload doses and higher protein doses compared to the ADCs with more potent camptothecin payloads. Finally, we need to acknowledge that the vast majority of ADC metabolism occurs in normal tissue. This means that we not only have to design and optimize molecules for the small fraction of the ADC dose that actually gets into the tumor, in the tumor microenvironment, but also for the large fraction of the ADC dose, which is catabolized in normal tissue. So as a result, our payload criteria include moderate potency and bystander activity with evidence of single-agent activity of the small molecule or close relative, providing us with added confidence in our selection process. So we're sharing today an overview of some of the payload classes that we're working on currently. As a broadly cytotoxic payload, we're investigating an inhibitor of protein synthesis and after a significant structure activity relationship campaign, we've optimized the potency and the PK of this payload and have observed good efficacy and tolerability thus far. We anticipate the potential application of this mechanism in various solid tumor settings. We have 2 different payload discovery efforts that we term conditionally cytotoxic. The first is an inhibitor of a key oncoprotein that offers promise in a number of different solid tumors. The activity of this payload is conditional upon the tumor mutational status, but it's a frequently enough mutated gene that there's a large opportunity here. The second program exploits a synthetic lethal relationship. Beyond the promising activity of these small molecule drugs, what's interesting to us about both of these approaches is that we have a well understand biomarker for patient selection. So the last program, the fourth program that's underway explores the antibody targeted delivery of protein degraders and these degrader antibody conjugates provide us with optionality with respect to both disease and in oncology tumor indication. So we want to provide a little bit of a glimpse into a couple of these efforts. First off, here is some data from our protein synthesis inhibitor effort. This is a completely novel payload mechanism that targets a very clearly essential process in cancer cell growth. We started off with a synthetic small molecule library, from which we were able to identify inhibitors with potency in the desired range and with other key features, including bystander activity and desirable biophysical characteristics. Selected molecules were conjugated antibodies to assess potency and to ensure that the observed cytotoxicity was specific to a tumor targeting concept in ADC. A couple of examples of this potency and the tumor target-specific ADC activity are shown in the graph, that's the second from the left. We next evaluated the in vivo antitumor effects in several different ADCs, a couple of which are shown here to cause complete tumor regression in an SKBR3 cell line-derived xenograft model after a single dose of 7 milligrams per kilogram. And then these same ADCs were tested in tolerability models and found to have comparable or superior tolerability to vedotin ADCs in rats. So this is a promising discovery effort, and we're currently assessing indication fit before we apply it to a pipeline program. The second example that we discussed here is an earlier stage effort to develop an inhibitor of a key oncoprotein as an ADC payload. The thinking here is that an ADC approach could offer improved efficacy through sustained exposure and tumor targeting, while potentially improving the tolerability profile compared to the chemotherapeutic approach. Mechanistically, this payload complements many of our preferred tumor indications and offers upside in other indications with unmet need and significant patient populations. So our initial goal was to establish in vitro proof that an ADC could at least maintain the potency of a clinically promising benchmark inhibitor so mechanistic POC, if you will. At left here, the cytotoxicity of our novel ADC is evident shown comparable to a vedotin ADC in a lung cell line. Next, we modified those clinical benchmark inhibitors and demonstrated that our novel small molecules could retain or improve on the potency of that benchmark in the same setting. And we also improved on the hydrophobic character of these molecules with an eye to the overall character of ADC. So next, these novel inhibitors were conjugated on to antibodies and the graph at the right illustrates how the ADCs were at least as potent in killing tumor cells as was the benchmark ADC. So at this stage, we've developed novel and potent ADCs with these novel payloads. And we're currently evaluating the antitumor activity of these ADCs in vivo in cell line drug xenograft models, and we're continuing to investigate the structure activity relationship in this mechanistic class because there are a number of different scaffolds that we can interrogate and optimize. So I'd like to now turn things over to Stuart Barnscher, who's going to talk about how we approach targeting these payloads.

Thank you, Jamie, and good morning, everyone. My name is Stuart Barnscher, and I'm the Senior Director of Preclinical ADC programs. And I'll be talking to you about 3 of our advanced ADC programs and give you a snapshot of how our antibody discovery and engineering capabilities gives us a unique edge in the optimization of these assets. As we strive to discover multifunctional ADC therapeutics with novel payload mechanisms, we can't overlook the importance of the delivery vehicle, the antibody. Thankfully, Zymeworks has an antibody discovery engine that when paired with our antibody engineering capabilities allows for careful selection and combination of valency, geometry, affinity and target epitope to match the specific requirements of the target and disease biology. We utilize both wild-type and transgenic animal platforms paired with either internal B cell culture methods or more traditional methods to find antibodies that are not only specific to the target of interest, but also possess properties well suited to the ADC modality, including internalization, payload delivery and tumor penetration. Our engineering technologies allowed for modular evaluation of multiple antibody formats, including biparatopic and bispecific formats with varied valences and geometries. Taken together with Fc receptor binding modulation and drug antibody ratio optimization, the final engineering step allows for us to specifically tune each of our ADC candidates to their respective targets with efficacy, tolerability, PK and the patient in mind. Our Azymetric platform, paired with our flexible antibody discovery workflow, unlocks both diverse antibody formats and optimal paratope selection. Ultimately, target expression and biology drive the appropriate format selection for each program. In some cases, a monospecific antibody may be the preferred format for a given target. For instance, antigens with significant normal tissue expression or limited epitope space may be better suited to a monospecific antibody approach. Conversely, a biparatopic format may be favored to enhance internalization for antigens with limited normal tissue expression and a large enough epitope space to generate a panel of diverse paratopes. A bispecific format may be best utilized for targets with clear biological association or targets that, when combined, provide extensive tumor coverage either within an individual patient's tumor or across a patient population. For a bispecific ADC, there are more levers to pull as valency, format, affinity and geometry can be tuned for each target arm to maximize on-target tumor binding and minimize normal tissue expression. The first program I'll share with you is a program where we've selected a monospecific antibody as our lead format due to the small extracellular domain of the target antigen leading to restricted epitope space, which limited the ability to generate a diverse pool of paratopes and in turn an effective biparatopic ADC. Moreover, we were able to find a monospecific antibody with superior properties when compared to the clinical benchmark. As you can see on the left, the target is highly prevalent in a number of solid tumor indications, either with high unmet need or in our digestive tract indications of interest. In the middle panel, you'll see that the monospecific antibody we discovered and selected as our lead has 7x greater internalization compared to the only known ADC benchmark to this target. This enhanced internalization profile can lead to greater payload delivery at the tumor site and better antitumor activity. And that's exactly what we see on the right-hand side. You'll see that our monospecific ADC demonstrate superior antitumor activity compared to the benchmark ADC at a common dose of 6 milligrams per kilogram in a tumor xenograft model representing expression levels we'd expect to see in patients. The program provides a nice example where a more complexed antibody format was not achievable, but also not required to differentiate from the benchmark antibody and highlights our capabilities at discovering best-in-class antibodies for drug conjugate development. The next program we have in the works is a solid tumor program where we're exploring biparatopic antibody format. This target is highly and prevalently expressed in non-small cell lung cancer, both adenocarcinoma and squamous cell carcinoma, triple-negative breast cancer, and we are actively evaluating the expression of this target in a number of digestive tract cancers. In the middle panel, you can see that we compared the internalization of our lead biparatopic to our lead monospecific antibody with the lead biparatopic demonstrating 3x greater internalization compared to the monospecific lead. In the right panel, you can see that we compared the lead biparatopic ADC to lead monospecific in a non-small cell lung cancer, tumor xenograft model. And at a single dose of 6 milligrams per kilogram, we see that the lead biparatopic ADC demonstrate superior antitumor activity compared to the lead monospecific ADC. Taken together, these data suggest that the biparatopic format may offer superior antitumor activity over a monospecific antibody for this target. This program highlights both our antibody discovery capabilities in finding a panel of paratopes with diverse epitope binding profiles and our engineering capabilities in combining those diverse paratopes into dozens of biparatopic combinations to find a lead with enhanced ADC function. Finally, we have an acute myeloid leukemia ADC program in our advanced portfolio where we are exploring bispecific targeting. The rationale for exploring bispecific targeting for this program is to overcome individual target heterogeneity both -- on both tumor cells and leukemic stem cells and to enhance the ADC functionality by co-targeting cells that express both antigens. We've generated a proof-of-concept bispecific antibody where we see enhanced tumor cell binding and decoration compared to either of the monospecific antibodies in a tumor cell line that expresses both targets. And on the right, we see that, that enhanced tumor cell binding profile of the bispecific led to productive internalization events, as the bispecific antibody demonstrates greater levels of internalization compared to either of the monospecific antibodies. This is just a brief example of the power and capability of bispecific targeting, and we expect to see even greater enhancements in functionality and activity for this program once the target paratope, format and valency have been optimized for this program. As you've seen, we have a lot of exciting technology and development on both the payload and antibody side of our advanced ADC portfolio, which should lead to differentiated first-in-class molecules contributing to our 2027-and-beyond INDs. We're focused on developing assets for solid tumor indications with an emphasis on GI tract cancers and for hematologic cancers, including AML, by bringing together our novel payload efforts, guided by our hypothesis on ADC mechanism and our antibody discovery and engineering capabilities to create fit-for-purpose ADC assets suited to our specific targets of interest, disease indications and patient populations. With that, I'll now pass things over to my colleague, Nina Weisser.

Thanks, Stuart. Good day, everyone. My name is Nina Weisser. I'm Zymeworks' Senior Director of Multispecific Antibody Therapeutics. Today, I will provide an overview of our ADVANCE cell engager strategy and focus on building next-generation T cell engagers that we believe may have the potential to be disruptive and bring high benefit to patients. I would like to add the mystery missing slide that vanished from Paul's section has been inserted to my section. So I will be able to address some of the data that Paul wasn't able to speak to in his section on the 209 program. At Zymeworks, we are well positioned to develop and deliver disruptive, next-generation T cell engagers. With our core competency in protein engineering, we have several platforms and engineering solutions that enable the rational design of T cell engagers with enhanced functionality and specificity compared to traditional bispecific T cell engagers. Azymetric is our foundational multispecific platform that enables bi and multispecific antibody assembly. Our EFFECT platform enables modulation of Fc gamma R effector function. Integration of the Azymetric and EFFECT platforms allows for interrogation of tailored designs and screening of multiple antibody formats and geometries to identify candidates with a desired biology. As we look forward and drive the development of next-generation T cell engagers, we are focused on building molecules with enhanced functionality to improve antitumor responses and with enhanced specificity to further widen the therapeutic window and bring greater benefit to patients. Built off of the Azymetric and EFFECT platforms, we have trispecific T cell engagers with enhanced mechanistic properties. This includes, for example, our co-stimulatory trispecific, or TriTCE Costim, platform introduced earlier by Paul that has integrated CD28 co-stimulation to enhance T cell responses and antitumor activity. Furthermore, we have engineering solutions to enhance tumor-targeting selectivity to further drive activity within the tumor microenvironment and away from healthy tissues. As visualized by the puzzle pieces, we can easily pair different platforms and engineering solutions in a plug-and-play fashion to address complex disease biologies. As these plug-and-play solutions can be applied to diverse therapeutic areas, including oncology, our focus to date, as well as hematology and autoimmunity, which are new therapeutic areas of focus in our advanced portfolio. Today, I will focus on examples that demonstrate how we are further developing our TriTCE Costim platform and how we are applying engineering solutions to improve targeting selectivity in oncology and hematology. Examples of these approaches include 2 plus 1 formats and multi-antigen targeting to drive activity to the tumor as well as novel or intracellular targeting with exquisite tumor specificity. Additionally, for targets with more broad normal tissue expression, we have proprietary masking and cleavage sequence solutions for conditional activation, which my colleague, Thomas, will discuss further in his section. As indicated with the bottom flow arrows, our development process starts with a focus on our prioritized disease indications with high unmet need in which we identify the key biological challenges, limiting treatment responses, and from their design and test solutions to address. The ZW239 is our second, most advanced TriTCE Costim program under early preclinical evaluation that targets Claudin18.2, which is highly expressed in gastric and pancreatic cancers. As these cancers are characterized with low T cell infiltration and poor T cell function, ZW239 has a potential to improve responses in these tumors by enhancing T cell function and antitumor activity via CD28 mediated co-stimulation. As reviewed earlier by Paul, one of the key parameters in development of the TriTCE Costim platform was to optimize T cell engagement for optimal safety and antitumor activity. Specifically, ZW239 has CD28 binding that is conditional on CD3 engagement shown in the upper graph top left. You can see in the blue is a trispecific that's binding to T cells. In green is a CD3 bispecific where the CD28 is knocked out. So you can see that bump of avidity. But in the orange, where the CD3 is knocked out, you see no binding, thus, illustrating the conditional binding requiring CD3 engagement first. Sorry, I'm just finding my stuff -- spot on the slide here. In the upper graph and -- so additionally, T cell binding occurs in an obligate cis manner shown in the left bottom graph to ensure that no T cell bridging or fracture site occurs, which would further reduce safety and activity. Of note, T cell bridging is observed with the first-generation trispecific T cell engager, which is shown in cyan. The benefit of integrated CD28 co-stimulation is illustrated in the middle panel where ZW239 shows superior antitumor activity at low effector-to-target cell ratios and sustained cytotoxicity over repeated tumor cell challenges compared to clinical benchmark, bispecifics AMG 910 and ASP 2138. The loss of cytotoxicity observed with the bispecific benchmarks after the third round of tumor cell challenge is a result of insufficient T cell stimulation and function, which is not unexpected as this is characteristic of T cell engagers that activate via Signal 1 or CD3 only. This enhanced trispecific functionality translates in vivo, where ZW239 mediates superior antitumor activity in an established tumor model compared to AMG 910 shown on the panel at right. Now I'll switch to the missing slide here, ZW209. Essentially, these slides before and this slide here are -- the data is essentially the same, which is the beauty of the TriTCE Costim platform, illustrating not only with the Claudin-targeting molecule, but also with the DLL3-targeting molecule. We have the conditional CD28 binding. It's binding in obligate cis. We see that in the middle panel, enhanced T cell proliferation in blue with our trispecific relative to the bispecific in this case, tarlatamab, or AMG 757. We see that sustained cytotoxicity over repeated tumor cell challenges as well as that improved cytotoxicity in low effector to tumor cell settings. Overall, our data supports the potential for ZW239 and ZW209 to enhance the depth and durability of responses in Claudin18.2 or DLL3 expressing tumors. Data from both the Claudin18.2 targeted and DLL3-targeted trispecific programs illustrate the potential of the TriTCE Costim platform. Having identified a format that is transferable across targets, we are focused on advancing diverse applications of the TriTCE platform to maximize benefit to patients with difficult-to-treat tumors. I will now provide 2 strategies under evaluation to enhance tumor selectivity that we are applying to the advanced Costim portfolio. The first strategy we have prioritized to enhance tumor selectivity is to identify novel tumor target with exquisite tumor specificity. To do so, we have collaborated with a partner to identify novel membrane expressed proteins with broad tumor expression profiles and minimal to no normal tissue expression. In contrast to the majority of targets that have been identified using a combination of mRNA and immunohistochemistry methods with our partner, we have used a novel high-sensitivity proteomic method that enables discovery of tumor targets not identifiable via standard mRNA approaches. As shown in the graphs at bottom left, the target shown has diverse and broad expression across multiple tumors, including thoracic and gastrointestinal tumors that have low T cell numbers and function. Importantly, this target has minimal to no normal tissue expression and expression levels on a biology that supports application of the T cell engager modality. We are excited about the potential of this novel target and application of the TriTCE Costim platform as there are no active programs in this target space. This approach provides an opportunity with high commercial potential. Now I will discuss our TriTCE developments in hematology with a focus in AML. As we develop our next-generation T cell engager strategy in AML, we first identified the biological challenges that have limited activity of mono antigen target therapies, including bispecific T cell engagers and CAR-Ts. These include heterogeneous intratumoral antigen expression, which has led to antigen escape and treatment failure. Two, the lack of a clean single target between AML blast leukemic stem cells versus healthy tissues, which results in a narrow therapeutic window as there is limited differential expression between tumor and healthy tissue. And third, T cell dysfunction, which has been shown to contribute to treatment resistance and a lack of long-term responses with bispecific T cell engagers. To overcome both heterogeneous antigen expression and the lack of a single clean tumor target, we have identified 3 targets that have high overlapping expression on AML cells, which do not co-express on normal tissues. To address T cell dysfunction, we apply the Costim platform. Collectively, we are designing a co-stimulatory multispecific molecule that targets 3 tumor target antigens. As illustrated in the graphic at bottom, this molecule is designed to mean potency when 3 targets in blue or 2 targets in purple are expressed, as would be the case in a heterogeneous tumor, while mediating minimal to no cytotoxicity in a single target expressing cell shown in green, as would be the case for healthy normal tissue. Overall, this design is expected to widen the therapeutic window to enhance and broaden cytotoxicity in a heterogeneous tumor population and avoid targeting normal tissue while increasing T cell responses to drive enhanced antitumor activity. To identify candidates with a desired profile, we have employed multiparameter screening using the Azymetric platform to screen multiple antibody formats, geometries and paratopes in parallel. Shown in the illustration at center, we have identified candidates that show high potency in a triple target expressing cells in blue and minimal to no cytotoxicity in a single target expressing cell in green. Moreover, several candidates show -- also show high potency in the dual target expressing cell population as shown at the bottom left. Added to the enhanced tumor targeting profile, incorporation of the Costim platform will further enhance T cell responses in this dysfunctional T cell environment. In summary, we are excited about the potential of the ADVANCE cell engager portfolio and ability to design disruptive next-generation therapies. We have protein engineering solutions that are FDA approved, clinically validated as well as novel engineering solutions that enable plug-and-play building blocks to address complex disease biologies. We are addressing biological challenges and indications with high unmet need. We are designing next-generation T cell engagers to overcome biological challenges that are not addressable with traditional bispecific T cell engagers. And finally, we are driving the forefront of next-generation T cell engagers, where we are enhancing functionality and specificity to drive deep and durable responses in difficult-to-treat tumors. And with that, I'd like to thank you for your time. Next, my colleague, Thomas, will present on our advanced technology and engineering developments.

Thank you, Nina, and good morning, ladies and gentlemen. My name is Thomas Spreter, Senior Director of Protein Engineering at Zymeworks. And I would like to shift focus to our next-generation multispecific technologies and give you a brief overview of how we're using technology development to advance our multispecific pipeline and what's coming next. The core strategy that we implemented over 10 years ago with Zymeworks has been the full integration of protein engineering with biology and drug development. And this cross functional approach has been critical to enable focused technology innovation and differentiation. And in our hands, this has been particularly important for bispecific and multispecific developments and for really understanding the interplay of antibody geometry with paratope affinity, valency and target epitope. And we've seen that through this integrated approach, we're able to find unique and differentiated biologies as we have demonstrated for zanidatamab and also 171 and our TriTCE platform. What might not be as obvious to everybody in the room is that Zymeworks actually started out as a -- in silico protein engineering and design company, and we have over 15 years of experience in in-house software and tool development. And the recent innovations of AI and machine learning provide significant opportunities to accelerate and advance drug development in a number of different areas. And over the past years, we have carefully looked at internal and external opportunities, how to combine them with our in-house expertise and what would provide the highest impact for us. And I'd like to just show you a few examples of this. We first started implementing machine learning and large language models in our protein engineering workflows a few years back, initially focusing on enhancing multispecific engineering of stability, affinity and PK, as shown here on the top right. But in parallel, building on our strength of our Azymetric platform, we have developed a proprietary high-throughput screening system that yields high purity of final format bispecific and multispecific and can directly be applied to primary screening. And combining this high-throughput workflow with AI/ML predicted smaller paratope libraries really allows us for parallel screening of different paratopes, affinities and geometries in the final multispecific format. And this lead discovery cycle of AI/ML-guided focused lab design and rapid screening in the final format, as shown in the figure on the top left, allows significant acceleration of the development cycle and really extends our ability to rapidly find novel multispecifics with unique activity. Two other previously difficult and labor-intensive or even intractable areas that we believe AI/ML innovations open up new opportunities, our 2-in-1 multispecifics and targeting of therapeutic development of cytokines and natural receptors or ligands. 2-in-1 antibodies are paratopes that are able to bind 2 distinct antigens, as shown here on the bottom left. And these 2-in-1 paratopes can be incorporated into our multispecifics in different ways to alter and optimize targeting and we're currently pursuing this approach as part of a multispecific. And in addition, just briefly, as shown here on the bottom right, a number of therapeutically promising targets have previously been difficult or intractable due to very low stability and expression. And the combination of AI innovations with our physics-based methods to enhance expression, stability and immunogenicity makes those cytokines and targets accessible for therapeutic development. An additional technology we have been working on for a number of years is our conditional masking approach. We have previously developed highly efficient proprietary cleavable sequences and validated them against clinical benchmarks. And as you can see on the bottom left graph, the lead sequences we have developed are more efficiently cleaved in vitro than the comparative benchmark sequence, but are also very stable in circulation as validated here in a nonhuman primate studies showing less than 1% peripheral cleavage. And then for fast and efficient mass development, we have further -- we're further combining this with AI/ML approach for mass discovery and engineering. So this work we have done allows us to couple our cleavable sequence and masking approach with our 2 plus 1 TCEs and the TriTCE platform. And based on more recent clinical data and learnings over the past years, we believe this technology is well suited for development of masked 2 plus 1 TCEs and TriTCEs, and we're currently applying this to a multispecific portfolio, as shown on the top right. We have further been using our conditional masking approach for the development of a novel tumor targeted IL-12, as illustrated on the bottom right. IL-12 is one of the most potent cytokines. And what makes this cytokine still very attractive is its unique ability to remodel the tumor microenvironment of cold tumors while also reducing direct tumor cytotoxicity. But so far, most therapeutic approaches have failed due to severe toxicities and lack of a therapeutic window. We have developed a novel conditional masked IL-12 fusion that combines our masking approach with selective affinity attenuation. And shown here on the bottom right panel is the efficacy and tolerability of our IL-12 lead in comparison to the clinical competitor approaches in a syngeneic mouse model. And as you can see, our masked IL-12 lead is highly active, but has a higher therapeutic window shown by no body weight loss in the mice. As Paul introduced earlier, there is a significant need and opportunity to develop better biologic for deeper and more durable depletion of B cells in autoimmune indications and T cell engagers present in a promising modality. But in difference to cancer and AID tolerability and the toxicity profiles of T cell engagers are really significant concern. And for successful T cell engagers in AID, it will be critical to develop an effective, but also safe molecule. And one hypothesis in the field to achieve this has been to fine-tune the CD3 affinity and synapse formation to separate strong cytotoxic activity from cytokine release. An early clinical data suggests that the CD3 affinity or more specifically, a fast Kon/Koff is critical for this. We have developed previously a panel of proprietary CD3 paratopes that have this desired fast Kon/Koff profile as shown here on the bottom left in the table and also the associated SPR affinity graphs. And we believe having this ability to screen and fine-tune both CD3 epitope and affinity for the development of T cell engagers and potentially TriTCE for AID will be critical, and we're currently pursuing this in a preclinical setting. To further diversify our AID portfolio, we have been working on other AID bispecifics that can complement or extend our IL-4 alpha access. And the focus has been to identify targets and bispecifics that have the potential to be superior to the combination of single agents and potentially provide differentiated biology. The bottom right shows an in vitro assay for a new bispecific program. And as you can see, the lead bispecific is more potent than dupilumab and also the combination of the single agents. And we're currently pursuing this bispecific to extend our AID portfolio. I hope I was able to give you a brief snapshot of what's coming next and how we're integrating technology development with our multispecific pipeline to further drive differentiation of our TCE portfolio and expansion to AID. We have over 10 years of structure-based protein engineering and design and in-house tool development expertise and the integration of new AI/ML methods with our Azymetric workflows enable acceleration of multispecific discovery and novel targeting approaches. And we have a continued commitment to novelty and new approaches and are currently applying them to upcoming multispecific and AID programs. And with this, I would like to thank you and pass it over to Ken.

Great. Thank you, Thomas, and all our speakers from the research leadership team. I just want to make a few concluding remarks before we bring them up here and open up our second question-and-answer period. So we hope that the presentation today has provided you with a clear understanding of our near-term clinical programs, which will come together more than 18 months ahead of our original target that we laid out in our 2022 R&D Day. At the same time, hopefully, you see our R&D team has been working on a longer-term R&D strategy for our product portfolio. I think over the past 2 years, we have advanced certain of our preclinical programs and platforms in order to broaden our therapeutic area of interest beyond solid tumors and into hematological cancers and AID. Today, we announced our first IND from the ADVANCE portfolio, ZW1528, which is expected in the second half of 2026, with the potential for more beyond this in 2027. Members of our research team do bring a wealth of experience in the discovery and development of FDA-approved clinical stage therapeutic products in both autoimmune and inflammatory disease and in hematological cancers. I do believe these broader therapeutic areas will provide more diversity to our R&D portfolio and also enhanced optionality for our partnering and collaborative strategies. We do clearly see opportunities for us to develop unique and differentiated multifunctional therapeutics in hematological cancers and AID as we have, so far, in solid tumors. Over the past 18 months, we have built the internal organization globally to be able to actively manage about 5 or 6 early-stage wholly-owned clinical programs, which is what we're looking at now. Our CMC and TMO group has the ability now to actively manage about 10 or 12 products at various stages of preclinical and clinical development. Paul's research group, which we built in Vancouver, does provide the necessary capabilities and capacity to support a long-term R&D strategy, and you're seeing some of those elements today. We will need to manage our future R&D portfolio carefully with additions, but also with attrition and with partnering strategies in mind. So we can manage our resources and capital effectively and with the ability to maintain focus and speed of development. Obviously, our strong balance sheet and by maintaining financial diligence, which I think we've done over the past 2 years while we've accelerated our R&D strategy, we have worked within our capital constraints without partnering to date. And we'll continue to maintain a strong financial position will give us the optionality to support our best products as development advances. So we do believe strong financial management, combined with world-leading science, will continue to be very important for us to be successful. We have a very active publication plan for 2025 in order to regularly report on our ongoing work in scientific and medical meetings in a peer-reviewed format. We do hope to report further on advancement of the interesting research areas discussed today, including some of the programs with undisclosed targets, undisclosed payloads or undisclosed structures that you've seen in this presentation today. We are also very hopeful to be able to share initial clinical data from our initial clinical studies during the course of 2025 at relevant scientific medical meetings in a peer-reviewed format based on the continued progress in our global clinical development programs. So far to date, we have been able to meet or sometimes exceed our publicly stated targets for development activities, and we'll strive to continue to exceed expectations for progress in our R&D programs. With ZW209 nominated today, we have completed the construction of the 5x5 portfolio ahead of the original timetable, and we are hopeful that all 5 programs will be able to generate clinical data to validate our preclinical hypothesis and meet or exceed our very strict TPPs for each product. The advanced R&D strategy provides a broad range of new opportunities with advanced novelty and diversity as well as broader therapeutic areas of interest, starting with our planned IND for ZW1528 in 2026. So just to leave you some key takeaways from the R&D Day today. I do believe we have the R&D organization and you're seeing elements of it today and also the financial position to pursue the ambition that you see in the R&D strategy today. We didn't speak a lot about Ziihera as Jazz Pharmaceuticals had a chance to provide an update yesterday. But very clearly, the recent FDA approval of our very first medicine at Zymeworks has provided us with the confidence to continue to stive further for further unique and differentiated opportunities, utilizing our technology, expertise and experience that's very specific to Zymeworks. So with those concluding remarks and these key takeaways, I would like to invite our research leadership team back up here, and we'll formally open our second Q&A session. So you're welcome to come up. Hopefully, we're not one stool short, but could be.

So I would remind you if you do have questions for the panel, if you can just raise your hand, we'll bring a mic to you. Please identify yourself for online so it could be helpful. If you'd like to direct your question to a specific person, please do that. I do give preference to folks who took the time to be present here today for questions, so we'll answer them. But I will try to get to some of the online questions as well. As a reminder, at the end of our session today, all of the internal research leaders and our invited KOLs will be available for more detailed discussions or questions if I don't get to them. So first question.

Charles Zhu from LifeSci Capital. First of all, congrats to everyone here on all the progress. And I also want to say congrats especially to Stuart, Jamie, Nina and Thomas. Two years ago at the R&D Day, you guys were directors, and now you all are senior directors. So belated congrats on that. A question for Alex. I may have missed this, but could you help me understand, in your 2x2 matrix, when you're kind of describing the presence of IL-4 and IL-33 across these different COPD patient populations, how relevant is IL-4 biology in areas where IL-33 is being evaluated? And vice versa, how do you think about going bispecific, but monovalent for each target as opposed to, let's say, the more traditional bivalence, monospecific? And I guess one kind of related follow-up to that is, how relevant is, I believe, there was that was -- the challenge where the -- the serum IgE were dupi performing at least as well as your bispecific, how are you thinking about that?

Maybe Alex and Paul, do you want to take that?

Yes. Maybe I can start that off just in general terms, and then Alex can get into some more of the specifics on that. So I think you covered a lot of ground there on your questions. I think regarding the coverage of where IL-4 and IL-33 play, certainly, there is evidence that both are important players in COPD clinically. That has been demonstrated from clinical trials there. So -- and I think Alex mapped out where we think the combination of those specificities could actually expand potentially the utility of those axes. I think then you had some questions about the differences in IL-4 versus dupilimab versus 1528 in a specific experiment, that Alex ran the in vivo experiment, which really is an experiment to look at IL-4 blocking. And maybe there, I can let Alex expand on that.

Very great. So you're probably referring to the mouse challenge experiment. So these guys have been made with for human IL-4 and IL-4 receptor. So there is no really IL-33 because it's not cross-reactive. So the only thing you can assess there is the IL-4 blockade, which our molecule does really well. There's no difference between dupi, which is a bispecific -- bivalent and that one arm of 1528. And what I can add to the Paul's comment on general approach, I think what the molecule clearly can do, it can do the job of both antibodies. It's just a single molecule, which can do exactly the same as IL-4 blocker as an antibody and exactly the same as IL-33 blocker as an antibody. But bispecific can do them both. There's no loss of anything as far as we can tell from the clinical piece.

Yes, Jon?

Jon Miller from Evercore ISI. Ken, I'd love to hear you talk about attrition in pipeline, and I'd like to hear responsible talk from management. But I'd love to hear from the entire development team. We hear from a lot of strong, early-stage drug developers a commitment to be data-driven in stepping away from programs that aren't driving value, but that doesn't always happen as you're very well aware. So can you tell me a little bit more about what you view as the key milestones, key opportunities to make those choices and how you intend to make them more responsibly than people who have let good money go after bad?

Yes. Let me try and -- thank you, Jon, for that question. And for noticing the word attrition, I think, in an R&D strategy at the outset. I think we've clearly been at this, at least since I've been here in 2022, is we took the opportunity in 2022 to state some pretty ambitious objectives of moving 5 molecules into clinical studies within 5 years. This group was able to do that about 18 months faster to make our life a little bit more interesting. But we've been able to sustain that ability through the financial constraints that we had. At the same time, and I don't think it's been recognized probably until today, where we had a chance to talk about it, we've been able to advance a number of our preclinical programs and technologies and spend time adding more novelty and more diversity into them. So I think what that lets us be in a position to be is that those 5 programs, which we picked into -- moving the clinical studies in the 5x5, there was a choice of some really excellent molecules. So I think, we hope we're adding more valuable and interesting compounds in the clinical studies to start with, with some really good internal competition amongst a whole range of potential compounds. The fact we do both ADCs and T cell engagers and other multispecific formats gives us a pretty broad internal competition of product modalities beyond targets and payload choices. So I think the quality going in is better. I think we selected 5 compounds to move into the clinic because we felt we had to rely upon then clinical data to provide us validation of which ones were truly best-in-class, especially as additional products move along in the process. We weren't expecting that all 5 of these would be able to at the end of an early-stage clinical proof-of-concept stage, be able to be still be best-in-class against all the competition. It gives us the optionality to decide then if we want to rally around some which are clearly best-in-class. It gives us the ability to then replenish the portfolio on a regular basis, which is why we were working on advanced or in the last 2 years. So as there is attrition in the portfolio, we'll have interesting compounds, which are also an internal competition to replace and get back into that circle. We went a little bit beyond the 5x5, and Alex joining us was helpful on that to drive our first autoimmune program into the clinic. So hopefully, we'll be in a position where we have 6 and 3.5 or something like that. So that's really great to exceed the ambitions that we put down before. But I think this attrition is embedded in the company from the very beginning in the quality of the molecules and the quality of the work that we take to decide to take something in the clinic. The strong TPPs and rigor, I think, we have inside the company, a lot of it's been driven by extensive KOL input from the outside. I mean you saw 3 of these people today, but in each of these therapeutic areas, we have really strong KOL input as to what's really going to make a difference for the patients that they see every day. And we're going to hold true to that. So you should expect these are all excellent molecules. We think the preclinical hypothesis models behind them are fantastic. We think we're well positioned to be best-in-class. Clinical data will clearly tell us that and when we read it objectively and decide then what attrition might happen in the portfolio, which ones we'd like to partner with someone because we'd like to share development expertise and cost to move forward, which ones we may want to keep ourselves a little bit longer. We've built a group inside of the company now, as I said in our early clinical group, that could handle about 5 to 6 active early-stage clinical programs. We're not looking to go beyond that. I think that constraint somewhat allows us to then really push the quality of the medicines in clinical stage coming out the other way. And at the same time, these advanced molecules, many of which were undisclosed by targets or payloads are today, you'll hear more about in 2025, those are going to push the quality of what we see in our clinical stage portfolio. And I think we've tried to set ourselves up that way by 2022. It's not something that many biotechs at our stage get the chance to do. We have the opportunity because of the optionality we have with being able to move 5 programs in the clinic in a very short time period, to build the portfolio and then manage actively that portfolio looking at not just capital constraints, obviously, opportunities and where our science is really best-in-class. We want to find ways to move those forward. But I think it does give us optionality within the scale and the financial constraints that we have, and I'm hoping that quality of that portfolio leads to a higher probability of success, higher chance of getting best-in-class and, most importantly, a much higher probability of really moving the needle for patients and the indications where we're at and we have the experience of Ziihera to know how to do that. And we didn't talk a lot about Ziihera today, but if you look at the original U.S. patent for that drug, ZW25, you'll see the first name and that patent is sitting on the stage. If you want to look at the architect of Azymetric, you'll see that person sitting on this stage. If you want to know how we developed our current portfolio of 519 payload ADCs, those people are sitting on this stage. We've added Alex to our team. Paul has brought him on board to give us the expertise to then broaden the therapeutic area of focus, but now that we have the experience capabilities to do that, so I think it's very positive for a small biotech like us to have immense continuity in a research leadership team like we have on this stage, and Paul joined us in 2022 to really push that forward. And we have the experience of already having our first FDA approval for the first medicine that we came up with and it just gave us confidence and encouragement and ambition to do that again and maybe a couple of times if we're able to do that. And so that's what we're doing. So this attrition word you will hear from us. It's not a pretty word for a lot of biotechs, but it's the only way to really generate quality products at the end of the day. And so that's a mantra that's inside the company, and you'll hear us talk about that frequently during the course of 2025 and 2026 as clinical data starts to come out of the 5x5 portfolio and as these new agents from ADVANCE start to push up to find their place in the clinical portfolio that is inside Zymeworks. So thank you for that. Next question?

I can take this one. I have had a question from online from Yaron Werber at Cowen. Can you discuss how you manage to use the Azymetric screening to optimize the geometry of how -- of the low CD3/CD28 affinity while still optimizing DLL3 antitumor activity with ZW209?

Paul or Nina, do you want to take that?

Yes. Yes, I think you have Thomas and Nina here. I mean, there, I think, again, it comes down to the screening capability you can do with Azymetric, which allows you to really toggle a lot of variables. And then you can actually, as Thomas explained and Nina explained, you can then screen a lot of molecules in the final format. I mean if you've got a strict target product profile or molecular profile that you want, you can find that molecule. I mean that's sort of the general sort of strategy that we take when we're tackling an axis that we want to develop a therapeutic agent. But maybe I'll let Thomas maybe expand a little bit.

Sure. I mean, I think we covered that. I think as we showed, right, one key idea that we always had is that you really need to fine-tune that geometry, particularly when we're talking about you want to have CD28, CD3 and TAA. And you know for a T cell engager, it's critical to fine-tune the synapse, right? That's why we always use very different geometries as well. And the strength of the Azymetric is really is that you can make all these different formats with very little contamination. So that, I think, allowed us to set up systems we can express the final format in pretty much one system. So we can start with now different paratopes or focused paratope libraries, put it into multiple different formats and really do that very quickly. And that's I think I showed you in this lab on a loop cycle at the very end, right, so we can try to optimize that process using AI as well. So that's, I think, where we can do this really, really quickly, do a couple of iterations of it and then find the best molecule in the end.

Thanks, Steve. You have a microphone. Less than 3 questions.

Less than 3 at this time. Sorry. So you guys talked a lot about novel payloads, I guess, on the small molecule and degrader fronts. And I think you talked a lot about the acumen of the antibody engineering at the company. Just kind of curious around the small molecule acumen of the company as it pertains to these novel payloads. And I guess, whether or not there may be a higher ROI to go outside and perhaps partner with some of those companies that specialize in those modalities specifically.

Yes, thank you for the question. Certainly, we have an additional chemistry with a lot of experience in developing antibody drug conjugate payloads. The topoisomerase 1 inhibitor platform was not the first platform that we've worked on. But certainly, where opportunities exist to partner with other people around payload technologies, we're keen to do that. But I think we have developed our own sort of philosophy around ADCs, how they work, some of the key properties of payloads. And I think we're well positioned to execute on that.

I think we've illustrated this before. I think if you look back to 2014, when I wasn't CEO but I was on the Board, we were a very strong bispecific antibody company with great protein engineering skills. We decided we want to try to apply those in ADC formats. And so we made the choice of making a small acquisition of a company called Kairos, which is how we got Stu and Jamie and brought them in and integrated that within the company in a way that allowed all of those drugs to stay together and then give us the optionality of looking at ADC formats or bispecific or T cell engagers formats at that time for different targets of interest that we're looking to. So we're certainly open to and do look for collaborations in areas where we think we can benefit from some expertise or experience that we may not have all of and bring that in and try to find a way to integrate that with the way that we work. It doesn't have to be done through an acquisition like we did with Kairos. We're happy to do collaborative efforts and have been looking at those as we look at all of these novel areas, and we've done that with respect to targets, with respect to potential payloads and we've done it with respect to other areas that we're looking at going into. So we see that as a way to kind of maybe advance a little bit faster, maybe with a little bit more expertise than we have, which gives the higher probability to success. We'll look to do those things. Do you have another -- second question? Other questions in the room before I go back online? Okay. I'll go back to you, Jon. Yes, sorry.

I got an easy one for you, Nina. Just am I correct when I look at the cartoons in assuming -- these cartons are the best evidence for what's going on really in multispecific antibodies. But the obligate cis binding interaction on CD28/CD3 bispecific, is that driven primarily by steric constraints in the architecture of the antibody, not by the epitope -- not by the binding epitopes in the antigens?

It's definitely -- it's a combination of multiparameters in the sense of the affinities of the CD3 and CD28. They're both very low, in particular, the CD28. Yes, absolutely, the format matters and the steric positioning of the CD3 relative to the CD28 is really important to obtain that obligate cis binding and conditional CD28 binding as well. And we haven't interrogated fully, but there is definitely absolutely contribution of the individual epitope being targeted on the CD3 and CD28, which will contribute further. But certainly, the affinity and the geometry are big drivers for that.

Diana, do you want to go to another online question?

Yes. I have one from Amy at Needham. Across your ADC assets, how have you ensured you can achieve better safety profile to allow for longer duration of treatment or response? For ZW220, can you talk about how you believe it can be safer than some of the previous treatments?

Yes. Yes, Amy, we've thought a lot about that. That's been a really key thing that Jamie and Stuart and the ADC team have wired in, thinking about how do you control tolerability. I think Stuart and maybe I'll hand it to Stuart, you can get into some more specifics on that.

Sure. Thanks, Paul. So for ZW220, when we selected the payload for our topoisomerase 1 inhibitor platform, we had hundreds of compounds to choose from and we intentionally selected a more moderate potency payload because our hypothesis at the time was that if you increase the potency of the free molecule, you're going to increase the toxicity of the ADC and probably see more severe toxicity and potentially a lower protein doses or lower ADC doses. So moderate potency payload. Additionally, specifically with ZW220, we optimized the drug antibody ratio from -- you could go as high as 8. We chose a DAR4. And then we also silenced the Fc binding potential of that molecule with some very specific mutations that are probably fairly unique to Zymeworks. And we think that adds into the potential to enhance the safety profile. So safety with ZW220 and with all of our ADC assets is always top of mind. We determine that both empirically with our preclinical data, how things look in terms of tolerability and PK, but also in terms of the target expression profile and the patients we're trying to treat. So those are some of the levers we pulled with ZW220.

Other questions?

I have one from David Martin at Bloom Burton.

Okay.

What drives your decision of whether an ADC or TCE for each target?

Paul, do you want to talk about that?

Yes, sure. Yes. I mean, I think a lot of that comes down to the biology of the target. I think maybe a good example I can use here is why did we pick a mesothelin T cell engager and a full receptor ADC or NaPi2b ADC. That's -- they all overlap in the therapeutic area. There, it really was -- it partly is driven by the biology and driven by the properties of that molecule. We -- often, as you -- we will take ADCs and T cell engagers within the same space. But what we find -- what was a major driver, obviously, is the preclinical profile. So we're not afraid to double up on targets within the same space. That's kind of maybe a philosophy that's maybe a little bit different than other companies. We do see ultimately the ability to combine either a TCE with an ADC or with standard of care an option, and we often also see the possibility of combining the TCE with our own ADC in the future because we engineered in or considered having good tolerability. But when we actually -- if you want to take that question to a different level, then why did we pick mesothelin such the CD3 folate receptors in ADC? I think in that case, it's relatively -- it's quite obvious. There, we've had folate receptor has been validated target for an ADC. Previous efforts with TCEs were challenging for that target. So you do balance the level of expression of the target on normal tissue and the susceptibility, say, to direct on-target activity with the TCE. So you are tending to be a little bit more careful with your target when you're going with a TCE. And what we described here for a target like mesothelin, we can engineer ability around that by incorporating in more specificity towards the tumor target -- tumor expressed mesothelin than the normal expressed mesothelin -- or the normal level mesothelin in normal tissue. And I think one piece of data we included today we haven't shown before is where we've compared in that case are bispecific 2 plus 1 with other molecules are being developed. And I think what you could see there was that we did open that window quite considerably. We had more potency of activity in the tumor cell lines or the tumor populations while we had no activity or really block -- no evidence of any cytotoxicity on the model cell line, it was a normal tissue. That was not the profile that we saw for other molecules. So you can really see there we can navigate that. But the really -- it does come down to target expression, compatibility with the modality. Is it an internalizing target? Or is it a better target for a T cell-mediated response? Some of that we can predict from the literature from prior programs, some of it we would just establish empirically when we're testing our molecules.

Another question?

Ken, more of a question on just your thought about operation. I mean as you think about your R&D operation expanding into autoimmune, just how do you see your R&D operation evolve over time? I guess in the sense as, are each team kind of pitching the case to your BoD and the execs to make the case to move the program forward? How do you see allocation of resources and operations ramping up as you expand autoimmune? And then the second part to it is that autoimmune indications that you laid out here, they are much bigger than the oncology indications that you've trafficked in the past. So how much of a ramp up do you think that you need from a clinical operations side? And just any color on your outreach to any appropriate KOLs to start building the network so that when you are at the point of IND, you'll be ready to go on day 1?

Yes. No, good question. I mean, I think we like the diversity of having additional patient populations. And we've clearly seen opportunities where our technology platform can play a role. I think to the greatest extent, I think the elements of our R&D organization can remain integrated regardless of therapeutic area of choice. And I think we do have people on the stage who have experience in multiple therapeutic categories. So we'll stay integrated. There's no predefined portfolio goals in each of those therapeutic areas. So we still will let the highest quality ideas that we -- that flow from our technology platforms move forward. And I think as we see those rise up beyond 2026 and into 2027, we'll get a better idea of the opportunities for us in hem/onc and autoimmune that we see whether we're able to translate those into clinical candidates and we're also still participating very heavily in solid tumors a little bit more towards digestive system cancers. But ideally, we're just going to leave everything pretty integrated for now. We'll let the portfolio evolve by the quality of the opportunities and our ability to translate a good preclinical opportunity into a clinical agent. And then I think as we see the need to build some additional elements in autoimmune or hem/onc, then I think we can do that. But we'll let the portfolio drive that. So I think, at least for the next period of time, we'll stay as an integrated unit. We've obviously added some expertise to allow us to think about from a preclinical development standpoint moving into some areas outside of solid tumors. But there is no predefined number of candidates in any of those therapeutic categories. We'll just let the quality agents flow in the portfolio and then be prepared to manage that as we move forward over the next couple of years. There may be a couple of areas where we need something very specific in autoimmune or hem/onc, but we'll find a way to get that in the right way. I think looking at the profile of the integrated portfolio though, we do like the differential in risk associated with a more diversified portfolio, whether that's preclinical risk, translation of preclinical data into clinical models, how quickly you can get proof of concept in the clinic with some of these areas outside of solid tumors is pretty interesting to us. The commercial risk associated with different of those areas is quite different. So we like the diversity of how to manage our portfolio across the risk and the upside pieces of it. Obviously, partnering strategies might need to be different in terms of the way we think about them for some of these larger indications as you mentioned. So we'll think about how those partnering strategies might be different and where we engage someone to come on board maybe sooner or at what stage. So I think we can do it in a very integrated manner. We do like the outlook of how our diversified portfolio might come together, even though we haven't defined what it should look like by the number of candidates in which therapeutic area. We just want a nice quality portfolio that, taken together, gives us a range of opportunities to really be best-in-class and help patients well beyond where the standard of care is. That's what we do with Ziihera. I think we want to do that again. And we've got a nice opportunity to start with now of 6 agents that are in our 5x5 portfolio and a whole range of other opportunities coming forward to make sure that we push the quality, give ourselves chances to diversify. And then we'll just see how the portfolio evolves in a quality, data-driven manner, and then be able to differentiate our strategies later on to be able to actively manage that portfolio the same way we do today where it's a primarily solid tumor portfolio with one autoimmune agent added to it now. I hope that answers your question, Brian. Okay. Well, with that, and 13 seconds left, I'd like to really thank you for your patience and perseverance and attention to what we're talking about. I think we've come a long way as an R&D organization since 2022. I think if you look back at that R&D day from 2022, we are ahead of where we wanted to be. I think the quality of where we are right now with our initial 5x5 portfolio has not only accelerated, but some amazing opportunities for us to be best-in-class. The execution to get those agents into the clinic was very, very high quality, very fast. And so far, our execution in the clinic with our first 2 programs is trending the same way. And we hope to, in 2025, have many opportunities to update and report on the work you saw today around the clinical agents and the advanced portfolio, and we look forward to providing those updates next year to you. Thank you for your time and attention. Please remember all of our physician KOLs, at this panel up here, happy to address any questions or further discussion with you since then. Other than that, thank you, and have a very nice day.