Transcript : ALX Oncology Holdings Inc. - Special Call

Good day, and thank you for standing by. Welcome to the ALX2004 R&D Webcast Conference Call. [Operator Instructions] Please be advised that today's conference is being recorded. I would now like to hand the conference over to your speaker today, Jason Lettmann, CEO. Please go ahead.

Welcome to our R&D event focusing on our new ADC-ALX2004. I'm Jason Lettmann, CEO of ALX Oncology. Excited to tell you more about this program today, and appreciate you all spending some time with us this morning. On Slide 2, before we start, just housekeeping. Here are our forward-looking statements. So for today, our goals are to walk you through the background of our differentiated ADC, ALX2004, the scientific rationale and significant data we've generated to support it, where we see the opportunity for patients and hopefully leave you as bullish and excited as we are on the program. So for today, in terms of the agenda, here's the plan. I'm going to start with giving you a program overview. Next, we will have Marija Vrljic , who leads the R&D effort at ALX, go through the rigorous work that her and her team have conducted to bring this potentially best and first-in-class EGFR ADC into clinical development.

Last, our CMO, Alan Sandler, will review in detail our clinical development program and how we envision taking this program forward and benefiting patients. On the next slide, and as a reminder to many of you, shares a snapshot of our clinical pipeline and where we're heading at ALX, both with evorpacept, which, as you know, is our differentiated CD47 blocker and ALX2004, which will be today's focus. As we communicated a few months ago back at our R&D Day in March, our priority is to advance evorpacept in combination with anticancer antibodies, given the consistent proof of concept that we have seen in various clinical studies with different monoclonal antibodies across many years. This is the primary MOA for evorpacept since the company's founding in 2015 as enhancing ADCP with Evo has been our focus since day 1.

We have 5 positive clinical studies with Evo, as you may know, and the efficacy and safety profile that we saw with Evo with 2 HER2-targeted agents in both breast and gastric cancer make us very excited about our new breast and colorectal trials. On that front, quickly, we remain on track and are executing with speed towards launching these 2 studies in breast and colorectal in combination with Herceptin and Erbitux, respectively, and remain on track to do so middle of this year. But the focus for today is on ALX2004. As you may recall, we have IND clearance, and we are going to walk through why we are excited about that program as we move to launching the Phase I trial also on track and planned for middle of this year. On Slide 5 highlights ALX2004 and provides a summary as to what are the key points that we would like you to take away from today's call.

First, ALX2004 was developed by a world-class team in our Palo Alto Labs with a vision starting back in 2021 to create the best and potentially first-in-class drug designed to maximize the therapeutic window and overcome the historical tox challenges that others have encountered in targeting EGFR with an ADC. Our team has a long track record of creating drugs that have led to multiple FDA approvals and deep protein engineering as well as deep chemistry expertise. They also have deep expertise in cracking the code and drugging known targets, and developing an ADC against EGFR is a good example of this. This team has created an ADC that has been optimized on all 3 components. The payload, linker, antibody have all been optimized to create a truly novel molecule against a validated target.

Last, the data that we will share today and our goal today is to show the data that supports this. The preclinical data, both in vitro and in animal models support dose-dependent activity, and the data also supports our conviction that this molecule could potentially demonstrate efficacy with, importantly, a manageable safety profile in patients. We look forward to investigating that in the Phase I study in patients with EGFR-expressing tumors, which Alan will walk through later, and we're slated, as I said, to begin that in the near term. We expect to have safety data here in the first half of 2026, which, again, given the challenges in the field will be an important catalyst for the program.

On the next slide, Slide 6, highlights the robustness of the work that has gotten us to this point. First step was to build a platform to take on this effort, where the team utilized substantial resources to take this on. On the left, you can see what was done in terms of designing and building the payload and linker. As I mentioned, this was internally designed and developed in the lab, starting with 600 payloads designed, then 60 payloads synthesized, then optimizing the ADC across 3 targets with almost 20 linker payloads, which then led to in vitro ADC potency work as well as demonstrating the activity in a CDX model. Last but not least, we took that into rats as well as primates and had 2 different linker payloads tested as part of the effort. On the right, you can see what we did in terms of antibody optimization.

First, optimization of the epitope and affinity was optimized across a multitude of different constructs. This then led to in vitro cytotox work as well as bystander effect work, which ultimately then culminated soon in our IND-enabling studies. What we have as a result, is a proprietary molecule and IP that covers compounds across all of these attributes, and we're very excited about the next steps here. On Slide 7, again, highlights how this ADC has been meticulously designed to maximize the therapeutic window. Again, on the antibody front, the binding epitope was selected to minimize the off-tumor skin tox that have been seen and characterized with this target. The linker payload was also optimized to be similar to ENHERTU in many ways, yet minimize off-tumor payload release. And last but not least, we are using a Topo 1 payload, which again is the most clinically validated payload and has potency similar to deruxtecan as well as enhanced by bystander activity. Again, all 3 of these attributes have been optimized to bring us here today.

Again, on the next slide, this highlights how the differentiated design behind ALX2004 stacks up both with what has been tried in the past as well as what is currently being pursued by others. As we've mentioned, there have been failures in the past, which we have addressed. Later in the presentation here, Marija will discuss the failed programs from Amgen and AbbVie and why we believe we have addressed those. There are also other approaches to targeting ADCs in development, both monospecific approaches as well as bispecifics. And if you look at what we've done here, it's to optimize our ADC to address how we both stack up with competition as well as what's been done in the past. Again, in terms of payload, we're utilizing Topo 1, which is a well-known payload that has been validated across a number of different tumor types.

We are targeting EGFR, which again is a well-known and validated target. And unlike the bispecifics, we are targeting EGFR alone. With a bispecific, there comes increased complexity, which in the context of an ADC, in particular, with a payload present, may or may not ultimately prove to be beneficial in the clinic. And last but not least, we have optimized the antibody. We are using a differentiated epitope with differentiated affinity, and that compares favorably with many of the other ADCs in development that are using known and approved antibody constructs. Again, all of these 3 working together are why we believe we have a differentiated ADC that we're very excited about.

Next, on Slide 9, highlights specifically how this will, we hope, translate into the clinic. What you'll hear in more detail is the robust data to support all of these claims. And again, that was driven by years of engineering and chemistry efforts to get this right. Again, we've demonstrated dose-dependent activity, potent antitumor activity, both in xenograft models as well as CRC patient-derived model. And on the safety front, we're very encouraged by what we're seeing, particularly in our NHP tox work, where we feel our compound, our construct does not show the typical EGFR-related skin tox that you've seen. And again, when you think about how this compares to ENHERTU, we are also not seeing payload-related ILD, which has been a challenge in the past. Both of these attributes across safety and efficacy really has us, I think, well positioned for the clinic.

And before we go to that section, I'm going to now pass it over to Marija, who will provide a detailed overview on the scientific work and preclinical data we generated. Marija?

Thank you, Jason. In this section, I will walk you through the design of ALX2004 and the rationale behind its mechanism of action. As Jason alluded to, EGFR known and validated target as monoclonal antibody, has been hard to crack a successful ADC, primarily due to payload classes as illustrated in these 3 examples. EGFR remains one of the most validated and clinically actionable targets in oncology. Monoclonal antibodies like cetuximab and panitumumab have demonstrated meaningful clinical activity, but translating this success into effective ADCs has proven exceptionally challenging.

Historically, the primary barrier has been the choice of payload. Earlier EGFR-targeted ADCs use extremely potent or highly toxic payload classes, for example, tubulin inhibitors and DNA cross-linking agents such as MMAF or PBD, with toxicities prevented dose escalation to therapeutically meaningful levels. These experiences have shaped our thinking and informed the design of ALX2004. ALX2004 was intentionally engineered to overcome the historical liabilities of EGFR-targeted ADCs by integrating modern, best-in-class design features across 3 critical dimensions: linker payload, antibody and immunomodulatory effects.

On the linker payload front, we prioritized the topoisomerase 1 inhibitor payload for having the greatest potential to achieve a clinically meaningful therapeutic window with an EGFR-targeted ADC and paired it with stable tumor-selective cleavable linker. This ensures efficient tumor cell killing, both direct and bystander, while minimizing off-target toxicity in healthy tissues. With respect to immunomodulatory effects, the release payload triggers immunogenic cell death that can lead to modulation of adaptive immune response, while ALX2004 Fc engagement can lead to activation of innate immunity thus potentially amplifying long-term tumor control.

On the antibody front, the EGFR-targeting antibody retains intrinsic antitumor activity, including inhibition of EGFR signaling and, as just mentioned, engagement of innate immunity through functional Fc domain. Together, these features allow ALX2004 to deliver potent antitumor activity across EGFR-expressing cancers with a significantly broadened therapeutic window compared to legacy ADCs. In the following section, I'm going to dive a bit deeper into each of these properties.

Next slide. Our goal was to maintain potent killing of EGFR-positive tumor cells while maximizing the bystander effect to eliminate EGFR-negative cells within heterogeneous tumors. For the payload, we developed a proprietary topoisomerase 1 inhibitor with potent cytotoxic activity. It induces both direct tumor cell killing and a bystander effect that's effectively targeting neighboring EGFR-negative cells. Importantly, the payload also triggers immunogenic cell death, thereby priming an antitumor immune response.

Regarding the linker payload architecture, we focus on tumor-selective delivery. We stabilize linker payload deconjugation to increase stability in circulation and release payload specifically in the tumor microenvironment. This reduces systemic toxicity associated with off-target release of potent cytotoxin. On the linker payload front, we use native cysteine conjugation for simplicity and scalability but enhance it with proprietary chemistry to increase the plasma stability of the conjugate. The result was targeted payload release at the tumor site and reduced systemic exposure.

Next slide. This slide summarizes the robust cytotoxicity profiling we conducted. The top panel shows the in vitro screening results for 65 novel ALX payloads, each synthesized and tested across 8 tumor cell lines and benchmarked against payload of approved Topoisomerase 1 ADCs, Trodelvy's SN-38, and Enhertu's DXd. Each dashmark on the X-axis corresponds to a unique proprietary ALX payload. 14 payload candidates were selected for testing at full ADCs based on their payload properties, for example, cytotoxicity and permeability.

The lower panel depicts payload candidates converted into linker payloads then conjugated to 3 solid tumor targeting antibodies and tested as full ADCs on 6 tumor cell lines, and again, benchmarked against DXd linker payload. Each dashmark on the X-axis corresponds to unique proprietary ALX linker payload. ALX2004 linker payloads emerged as the most consistent and potent across antibody combinations, showing comparable activity to deruxtecan, the clinically validated DXd payload.

Next slide. We stabilized linker payload deconjugation to increase stability in circulation and increased release of payload specifically in the tumor microenvironment. Off-tumor deconjugation of linker payloads from their antibody remains a challenge even in the current generation of ADCs, which may lead to increased toxicity. To benchmark our linker performance, we conducted a head-to-head comparison with deruxtecan in nonhuman primates. We conjugated the ALX linker payload to trastuzumab in order to make a direct comparison to trastuzumab deruxtecan. We then compare drug to antibody ratios over time after administration in nonhuman primates. Our linker payload demonstrated superior stability, maintaining a higher conjugation level throughout circulation, as you can see on the graph. These data strongly suggests that ALX-2004 will deliver more payload to tumors while limiting exposure to healthy tissues. This is central to our strategy of improving the therapeutic index through intelligent design.

Next slide. The goal of our rigorous linker payload selection process was to match or exceed the activity of the DXd linker payload, both in terms of direct cell killing and the bystander effect. The bystander effect is an important mechanism of cell killing in solid tumors. The ADC is internalized in a target-expressing tumor cell. Payload released within the cell directly kills the cell, and the payload is able to kill non-target expressing neighboring tumor cells through the bystander effect. First, we synthesize an ADC using the ALX-2004 antibody plus the DXd linker payload in order to make a direct comparison of the activity of the 2 linker payload platforms. Second, we compared ALX2004 to DXd ADC, consisting of our antibody conjugated to the DXd linker payload in several EGFR-expressing mouse models, including a model specifically for bystander effect.

Across multiple CDX models with varying levels of EGFR expression, ALX2004, shown in purple bars on the graph, demonstrated equivalent or superior tumor eradication as shown by the percent of mice in which tumors were completely eradicated. Importantly, in comparison of the antigen homogeneous model on the left side of this figure and the bystander effect model containing both EGFR high and EGFR ultra-low cells shown on the right side of the graph, ALX2004 outperformed the DXd comparator, again, achieving a higher rate of tumor eradication. This suggests enhanced bystander activity, consistent with the design goal of tumor selective and membrane permeable payload delivery. We attribute improved activity to combination of improved bystander effect and improved linker stability compared to deruxtecan.

In addition to direct cytotoxicity, ALX2004 is designed to harness the immune system. The payload induces immunogenic cell death, releasing signals that can activate an adaptive antitumor immune response. Further, ALX2004 Fc region mediates antibody-dependent cell-mediated cytotoxicity and antibody-dependent cellular phagocytosis, augmenting tumor clearance by engaging innate immune effector cells.

This slide illustrates ALX2004's ability to induce calreticulin expression, a key marker of immunogenic cell death in EGFR-positive cells. Cells were treated with ALX2004, ALX2004 payload as an unconjugated small molecule and compared to the control, a nontargeted ADC and the ALX2004 naked antibody. After treatment with ALX2004 and its payload as small molecule, surface levels of calreticulin are significantly elevated and much higher than controls, indicating that ALX2004 retains the immunogenic cell death activity that is characteristic of the topoisomerase 1 inhibitor payload class. Similar results were observed for other immunogenic cell death biomarkers, including HMGB1 and ATP. These findings support ALX2004's ability to trigger adaptive immune engagement.

Antibody drug conjugation can inhibit the Fc-mediated antibody-dependent activity of the ADC due to steric hindrance from the 8 payloads attached to the antibody. We verified that ALX2004 maintains Fc-mediated immune effector activity. In vitro ADCC and ADCP assays confirmed robust dose-dependent activity in EGFR high cells with no measurable effect in EGFR ultra-low cells. This confirms that ALX2004 Fc engages immune effector cells, but only when EGFR is present.

The antibody component of ALX2004 was designed to both block EGFR signaling and bind a unique epitope distinct from approved anti-EGFR antibodies. This design may overcome resistance developed to approved EGFR-targeted antibody therapy. We also tuned the antibody's affinity to maximize tumor uptake while minimizing binding to normal EGFR expressing tissues, thus increasing the therapeutic index.

To maximize likelihood of the ADC success in clinic, we designed an antibody backbone with a widest therapeutic window. ALX2004 was selected for binding epitope different and approved antibodies, thus allowing for first-in-human study in patients with resistances to approved therapies due to EGFR mutations. In addition, its affinity was selected for higher tumor uptake and lower tissue uptake based on published studies done by others relating antibody affinity and the distribution. In order to test whether we created a potential safety benefit for activity, we tested a series of antibodies with different epitopes and affinity gaining 500 fold. In mouse models, ADCs showed similar activity, suggesting there was no activity penalty for designing a potentially safer ADC through affinity tuning. The antibody selected for ALX2004 offered a balance between tumor activity and safety.

This slide demonstrates that ALX2004 inhibits EGFR tyrosine kinase activity in a dose-dependent manner in EGFR overexpressing cells. This further contributes to its antitumor mechanism beyond payload-mediated cytotoxicity.

I will now touch on ALX2004 in vivo efficacy in mouse models and its GLP toxicity evaluation in nonhuman primates. ALX2004 showed tumor suppression activity across a panel of xenograft models, representing a broad spectrum of cancer types and EGFR expression levels. Notably, ALX2004 was effective in models harboring KRAS, BRAF and p53 mutations.

ALX2004 showed excellent tumor suppression activity, especially for topoisomerase 1 payload class down to single 1 milligram per kilogram dose level, leading to complete tumor eradication in several models. Notably, complete tumor growth inhibition was seen at all dose levels, even with tumors having H-score of 100, for example, pancreatic models shown here. These results confirm the broad applicability of ALX2004 in targeting EGFR-positive malignancies.

In a patient-derived colorectal organoid xenograft model, ALX2004 showed dose-dependent tumor suppression. These data further support its translational potential in human tumors with low EGFR expression. PDX had an H-score of 60.

Our 6-week repeat dose with 6-week recovery period GLP toxicology study evaluated ALX2004 in nonhuman primates. All findings were minimal to moderate and fully recoverable. The GLP toxicology supports our design selection and safety margin for clinical use.

And next, our Chief Medical Officer, Dr.

Thanks very much, Marija. EGFR is overexpressed in many cancer types as well as normal tissues, but often to a lesser degree. This is illustrated in the bar graph on your right. The importance of EGFR is that it plays a prominent role in tumor initiation and growth through the various dysregulations of cell proliferation, differentiation, metabolism and cell death. Therefore, blocking this pathway is critically important. As you know, EGFR is a validated target with multiple FDA-approved EGFR-targeted antibody. ALX2004 is an EGFR ADC designed to further enhance activity with reduced toxicity.

Let's move to the next slide. So now we'll start to discuss our plans for our Phase I study. Given all we know about EGFR pathway and its importance in cancer, we've designed our Phase I study to maximize both efficiency and the probability of technical success. We will accomplish this by restricting eligible patients to 4 tumor types known to be dependent on the EGFR pathway. That is squamous cell carcinoma of the head and neck, colorectal cancer, non-small cell lung cancer and squamous cell carcinoma, the esophagus. Patients must have relapsed or refractory disease with no approved or standard therapy available for eligibility.

Looking at the table here, you'll see that these 4 disease entities meet the following criteria, including EGFR expression, sensitivity to Topo 1 inhibitors and have already shown sensitivity to EGFR antibodies. Importantly, this constellation of tumor types represent a significant unmet need, representing over 250,000 patients annually in the U.S. alone and of course, even larger on a global basis.

We'll now move to the next slide, where I'll give you a bit more detail about our Phase I clinical development study. Our Phase I trial consists of 3 cohorts. The initial one is dose escalation where we'll evaluate safety and establish the recommended dosing schema with an early opportunity to assess efficacy given the known sensitivity of these disease types selected. Upon successful completion of the dose escalation cohort, we'll move to the dose exploration phase where we'll have narrowed the dose and schedules for the Phase Ib expansion portion looking toward dose optimization.

In this last cohort dose expansion, we will, based on the results from the initial 2 cohorts, evaluate which of the tumor types we believe will have the highest opportunity for success. We look forward to providing initial safety data from the Phase I portion of this study in the first half of 2026. I'd now like to pass the baton over to Jason.

Great. Thanks, Alan. As we just walked through, we presented in detail the robust preclinical package that supports how we optimized our ADC over the last 4 years of internal development and is now ready for the next stage in the clinic. We believe EGFR remains one of the last validated targets for ADCs, as it remains without an approved drug and without a clear leader in the clinic. We believe ALX2004 has the potential to meet this need and to be the leader in the space. And so the team is now laser-focused on execution and, importantly, doing so at speed.

If you turn to the next slide, this summarizes again our pipeline, where we're headed, and the inflection points we have over the next year. As we mentioned at the beginning of the call, we're very excited about launching evorpacept trials in breast and in colorectal, 2 large unmet needs, and are on track to dose our first patient middle of this year. On the ALX2004 front, as we talked about here today, we now have the IND clearance in hand. Similarly, we are laser-focused on execution, driving towards the first patient in the middle of this year as well with the goal of delivering initial safety data first half of next year. Again, very excited about this program, pleased with the work to date. I appreciate all of your support and interest. And now I'll open it up to questions.

[Operator Instructions] And our first question is going to come from the line of Michael Yee with Jefferies.

This is Kyle Yang for Michael Yee. Just a few quick ones for us. The first one is, so on your Phase I data update, what would you like to see to support your decision to move this asset forward? And when do we expect to see initial efficacy data, please? And the second one real quick is, can you briefly talk about your capital allocation strategy and your various options to extend your runway?

Great. Thanks for the question. So on the first one, I think, as we mentioned, certainly encouraged by what we're seeing in the NHP work. Obviously, we want to see that translate. So the first step, and again, important for this class uniquely, is to demonstrate safety. So we think we'll have enough patients enrolled to have a good handle of that and a handle on the dose, and should have that first half of the year. I'll turn it over to Alan to weigh in on the efficacy. I think that's going to depend, right, on how things progress in terms of the dose escalation. But Alan, do you want to add a couple of comments on that front?

Sure. Thanks, Jason. Thanks for the question. One of the ways we've tried to design the study is by not having an all-comer study. So we've got 4 tumor types that are known to overexpress in EGFR and be sensitive to EGFR-naked antibodies. So we would anticipate some level as we go through. But of course, can't predict as to when that timing is going to be. The primary goal, of course, is going to be safety, but we do anticipate seeing efficacy along the way.

And then on your second question on capital allocation and sources of capital. We, as a reminder, have cash well into Q4 of next year. Our goal internally is to drive across these 3 studies, so the colorectal and breast study with Evo, and then the 2004 study, of course, here. That is our focus. Those are our priorities. And our goal is to deliver data on this capital with the money we have in hand. And so I think that's how we're executing. Of course, we have ongoing work with partners, for example, with Sanofi, where we're testing the combination with Evo plus SARCLISA in multiple myeloma, which is also in the clinic.

But those 3 studies in terms of our use of cash is where we're focused. I think from a fundraising perspective, I think that does give us runway here to put up additional catalysts. Of course, we're going to continue to assess options along the way, both in terms of equity as well as pharma, where we continue to have active conversations. So I think we're well-positioned with the cash on hand and are focused right now on execution.

Our next question is going to come from the line of Allison Bratzel with Piper Sandler.

This is Ashley on for Allison Bratzel. Can you hear me?

Yes, we can hear you.

Great. I was just wondering, I know this might be a horse before the cart kind of question, but I think our team was just interested in kind of understanding where even though at this really early stage, where this could potentially fit in the treatment paradigm or kind of how you guys are thinking about that? Or do we need to first kind of move through the early phases to have a better understanding of that? But any insights would be really helpful for us.

Yes, it's a great question, Ashley, and certainly not the horse before the cart or however you say that. But I think this team and the clinical effort started early here. So we are already gearing up on where we see the unmet need. And as Alan mentioned, and I'll let Alan expand on this, but we're starting out of the gates in the 4 tumor types that we think matter. And I think that's the right way to do it in this market. And I think we're going to be able to, out of the gates, hopefully treat patients where this drug should benefit. But Alan, do you want to talk a little more about any of those indications?

Yes, sure. Thanks for the question. And it's an important one and one that we start to think about even right in the early design phase, even in Phase I. We'll be taking the approach initially looking at this agent within the diseases, second line, third line-ish depending on the disease itself, seeing what type of activity we have and of course, attempt to move that earlier. We'll have lots of opportunities to evaluate the 4 different disease entities and look at how they are doing from an activity perspective and adjust accordingly. So again, not too early to start to think about it, but this will be data generated, and we're really anxious to get this started and see what we have because we're pretty excited about this agent.

Our next question comes from the line of Li Watsek with Cantor.

Maybe just first one, can you comment on the efficiency of internalization of this molecule? And then any resistance that you may have observed preclinically associated with efflux pump? And the second is, it looks like you're going after head and neck cancer. How are you thinking about the landscape here as some bispecifics are moving to the front line? Where do you think you can fit into the landscape? And any biomarker strategy here to select for EGFR expression levels?

Yes. Great question, Li. So on the first one on internalization and resistance, I think I'll have Marija weigh in on that. I think it's an important question, I think, particularly when you think about TOPO 1, right, in terms of where else it's utilized and how we would fit. But Marija, do you want to add anything there in terms of how we test it for that or how we think about resistance modalities?

Thank you for the question, and thank you, Jason. On the internalization front, we have chosen an antibody that delivers very efficient internalization of full ADC. So we are confident that we can target both the high-expressing cells and mid- to low-expressing cells, as you have seen across different tumor models, and that we have activity across varying levels of EGFR expression from high to pretty low. Now on the efflux pumps and resistances, as we have pointed out, we have tried to match our payload with certain properties very similar to deruxtecan. And on that front, we do have kind of a similar range to what deruxtecan is experiencing. And now kind of for the second half, maybe Alan or Jason, you can go, would you like to comment on that?

Yes, I'll start. It's also a good question. I think in head and neck, certainly excited about what we see. I think we know that cetuximab historically has been a predominant drug in the frontline, of course, until KEYNOTE-048, which disrupted that. But again, I think that provides nice validation as to what an EGFR can do. Certainly, exciting early data coming from both Merus and Vicara with their bispecifics. But I think, again, to have a targeted ADC going just at EGFR, which, again, in head and neck is very validated, is exciting. And I think there's opportunities, both first line and second line. But I don't know, Alan, I may have just answered the whole question. But if you have anything to add, go forward.

Jason, I think you did a great job. I answered the question. I think the again, you've got naked antibodies with success. We believe the ADC will build upon that. We'll be evaluating activity, and there's always room for active agents in metastatic disease. And so we'll keep an eye on that and adjust accordingly as things move forward and as this study begins to get started and approve patients, and we'll be able to assess as we go. But yes, similar thoughts, and thank you for the question.

Our next question comes from the line of Sam Slutsky with LifeSci Capital.

Two for me. I guess, first, could you just talk about the kinetics of when toxicities have been observed with EGFR ADCs in the clinic? And at what point would you feel confident that the safety profile of 2004 is clearly better than prior attempts at this target? And then second one, just how translatable are the preclinical safety models for ADCs historically? And then anything that you feel is quite derisked in humans already based on the preclinical studies?

Thanks, Sam. I think both are good questions, and I'll pull on Marija here. I mean, I think on the dose question and what has been seen elsewhere, the team went to school on the efforts, both by AbbVie and Amgen before. And I think what you see there that's encouraging for us, not encouraging for those programs at the time is that they ended up with a really tight therapeutic window of down to 1 mg per kg in one of the programs. So I think that's where we feel those programs stumbled in terms of safety and what we've optimized to. So, Marija, do you want to add to that any further?

No, that was quite a good summary, Jason. And I just kind of want to remind everyone that all of our findings were minimal to moderate, fully recoverable. And kind of as Jason pointed out, legacy ADCs, their primary liability was a payload tox of different classes and topoisomerase 1. And we feel that we have made a good choice by choosing topoisomerase going forward on that front.

Yes. And I think the second question you had was just on translation, which is also a good one, Sam. Again, I think and the NHP work to see, in our case, 10 mg per kg, no AEL level, when we've reviewed this data, both internally and externally with experts, that's quite encouraging in terms of ADC development in general. And then again, when you think about the on-target related issues here, I think we're particularly encouraged. Of course, we have to temper that. It's animal models, and we need to see it translate. But again, in ADCs, there's some translation here that's happened. For example, ILD within HER2 was observed in primates, and we didn't see that. So again, nothing translates perfectly, of course. But I think when you stack all this up together across the disease models and the great work that Marija and team have done across all the modeling, I think it's a really strong package. Does that answer your question, Sam? I don't know if we missed one in there.

[Operator Instructions] Our next question comes from the line of Ting Liu with UBS.

I have a question, maybe for Marija. So you covered details around the linker and the payload in depth. Thanks for walking us through that. Could you also provide additional colors on the antibody part of 2004? Noted 2004 has EGFR binding epitope not overlapping with cetuximab and panitumumab. You discussed AbbVie's DART-2M and the payload in that molecule. So something also unique about that asset is it had a distinct binding site versus cetuximab and panitumumab and has very weak affinity to EGFR in normal tissue. And then eventually, that asset did not turn to clinical success.

On the other hand, there are some other recently developed EGFR ADCs, which they had some preliminary clinical success. For example, there's a Phase III asset from China. I think it's called MRG-003. That one was designed with overlapping binding site to cetuximab, but with much higher affinity. So 2 questions from my side. One, what about 2004's affinity to the non-overexpressing EGFRs in the normal tissues? I think you've shown some data in the EGFR ultra-low target cell line. Should I consider those equal to what we would have in the normal tissues? And 2, how important do you think it is for the ADCs to have a strong affinity to the EGFRs?

Thank you for your question. So kind of to go through our -- as you've noted and kind of caught up on yourself, the epitope is different than the approved EGFR antibodies. However, for example, AbbVie program they targeted a particular variant of EGFR variant 3. This molecule is not specific for that epitope. It targets a different epitope on the wild-type EGFR. However, we may have the ability to bind to the variant 3 epitope as well. Now on the affinity range, as we have shown you, we have screened different affinity ranges and based on and concluded that we are not sacrificing antitumor activity within the mouse models. And I think, did I miss any of your other questions?

Yes. I just wonder, in general, do you think it's necessary that the antibodies you choose for the EGFR ADC to have a very strong affinity, essentially it binds to all the EGFRs no matter the expression state or it's actually preferred to have some like kind of weak to moderate affinity, so it only binds to the overexpressed EGFRs.

So, as I kind of just kind of like try to state, we have tested that hypothesis and we have concluded that, regardless of essentially affinity that we have chosen does not sacrifice antitumor activity. So, for example, in our hands, very high-affinity antibodies and the affinity range that we have chosen do not compromise the activity. They have a similar activity. So on that front, like maybe it depends on the program. However, we believe that the lower range affinity antibodies are better suited with respect to their ability to distribute more homogeneously within the tumor tissues.

And then I think the other part of your question, was just on competition. And there is competition, of course, there always is. But I think for us, when we think about where we are and where we have taken this program, we have definitely the potential to be first in the program that you mentioned, MRG-003, we certainly tracked. Those trials to date have all been based in China as we understand it. And so we see a lot of open field here. And again, I think spent the time to really optimize this ADC internally and feel with good execution, we can be the leader here. So also a good question.

And I'm showing no further questions at this time. And I would like to hand the conference back to Jason Lettmann for any further remarks.

Great. Well, again, I appreciate the interest here and grateful for the time and all the good questions and look forward to taking this forward and future updates. So thanks again, everybody, and we'll talk soon.

This concludes today's conference call. Thank you for participating. You may now disconnect.