CytomX Therapeutics, Inc. (CTMX) Earnings Call Transcript & Summary

Well, good morning, good afternoon, everybody. I hope everybody can hear me okay. I'm Sean McCarthy, President, CEO and Chairman of CytomX Therapeutics. And it's a pleasure to have you here today for our investor event focused on conditionally activated therapeutics for the treatment of cancer. We will, over the course of the next 2 hours, be making certain forward-looking statements, and I refer you to our safe harbor statements in our SEC filings. So really it's a pleasure to be spending time with you all today as we take a deep dive into our science and our programs at CytomX. CytomX is a clinical stage oncology-focused organization, advancing a broad pipeline of candidate therapeutics for the treatment of a wide range of cancers. Our company has built around a unique and differentiated approach to localizing the activity of anticancer biologics to tumor tissue. We believe that localization of the activity of biologics presents an enormous opportunity. The conditional activation technology that we call the Probody platform has led the industry, and we are coming towards key data readouts from 2 Phase II assets starting late this year. These are the conditional antibody-drug conjugates CX-2009 and CX-2029, first-in-class agents, which target the novel cancer antigen CD166 and CD71, respectively. In addition to these late-stage clinical assets, we have a rich early-stage pipeline, foundational alliances with key players in the oncology space, and our balance sheet remains strong following a successful financing in January of this year. Let me briefly review our agenda for today that includes 3 distinguished guest speakers and members of the CytomX management team. Our main focus will be our lead clinical assets, the conditional ADCs, CX-2009 and CX-2029. My colleague and CytomX Head of Research, Dr. Marcia Belvin will kick things off with a review of foundational aspects of the rational design of the CytomX platform. Marcia will be followed by Dr. John Lambert, CytomX Scientific Advisory Board Member and Former Chief Scientific Officer of ImmunoGen. Dr. Lambert will discuss core principles of antibody drug conjugate design, with a focus on payload considerations and the need of new targets for this increasingly validated therapeutic modality. CytomX's Chief Development Officer. Dr. Amy Peterson will then briefly review our clinical pipeline before handing over to Dr. Sara Tolaney from the Dana-Farber Cancer Institute, who will review and discuss our CX-2009 program, targeting CD166. Dr. Tolaney will be followed by Dr. Melissa Johnson, Director of the Sarah Cannon Research Institute, who will discuss our CX-2029 program targeting CD71. I'll come back to wrap up at the end, including a quick look at some exciting new work that we're doing in the conditional cytokine space. We'll then hold a Q&A session. Please hold your questions until this session at the end of the event. And I'm really looking forward to this. So let's get going. I'll now hand over to Marcia Belvin.

Thank you, Sean. You can advance to the next slide, please. Are the slides advancing? Okay. The slides are not advancing for me. So I will go ahead and speak to the slides looking at the printout. Cancer therapy has been improving over the past several decades with the development of increasingly more potent and selective molecules. Immuno-oncology has revolutionized the treatment of cancer therapy. However, patients who do not respond to immunotherapy leave a large unmet need, and new approaches are needed. Recently, conditionally active approaches have emerged as a potential new treatment paradigm. The power of this approach is to expand the target landscape and deliver therapy to new patients. The Probody therapeutic platform is a conditionally active platform that is based on the profound difference in protease biology between in tumor tissue and normal tissue. Upregulated protease activity is a hallmark of cancer and is required for such cancer phenotypes as migration, invasion and metastasis. In contrast, protease activity is tightly regulated in normal tissue to avoid unwanted tissue damage and aberrant clotting. Probodies are comprised of a therapeutic antibody to which is attached a peptide mask, here shown in blue, that blocks the antigen recognition site of the antibody. The peptide is attached to the antibody via a protease-cleavable linker, here shown in red, that is designed to be cleaved by proteases in the tumor microenvironment. When the Probody is in healthy tissue or in circulation, it's designed to remain in its intact form, which limit its ability to bind the target. However, when it enters the tumor microenvironment, where protease activity is upregulated, protease is now cleaved, the cleaveable linker, thereby releasing the mask and now allowing the Probody to engage its target. Next slide please. CytomX has been studying tumor protease biology for over 10 years and we're leading the field and applying our knowledge of protease biology to the development of conditionally active therapies. We are using our learnings from our ongoing research to continually optimize and improve out platform. We are investing heavily in systems biology approaches to understand protease expression profiles in tumor tissue and normal tissue. We have identified signatures of protease expression that are both common across indications as well as unique within individual indications. We have also identified those proteases that are the most differentially expressed between tumor tissue and normal tissue. And we are using these sequences and signatures to inform the design of our protease substrates. Next slide, please. I'll now describe how we're using what we've learned about tumor protease biology to develop multi-selective protease substrates. Well, as I mentioned, we're investing in systems biology approaches to look at protease expression profiles, solely focusing on the expression of proteases is not enough. And this is because proteases are highly regulated post translationally. The majority of proteases are translated as inactive zymogen, and they are also regulated by the presence of endogenous protease inhibitors. Therefore, the presence of a protease does not necessarily indicate that it is active. We have, therefore, developed novel methods to measure protease activity directly in tumors. And one such method is shown here. It's called immuno histozymography or IHZ, and we published on it last year. In this method, we use fluorescently labeled probodies as probes to detect protease activity directly in tissue splices. An example of that is shown on the right-hand side. Here, we're using IHZ and using 2 different anti-EGFR probodies as probes to investigate the xenograft model, H292. We're using 1 probe that is sensitive to cleavage by serine proteases, and 1 probe that is sensitive to cleavage by matrix metalloprotease or MMPs. In the control, you can see that both probes bind the tissue, indicating that H292 model has both serine protease activity as well as MMP activity. If we pretreat the tissue with an MMP inhibitor, we prevent the activation and binding of the MMP sensitive probe. Whereas if we pretreat with the serine protease inhibitor, we prevent the activation and binding of the serine protease cleavable probe. And if we pretreat with a broad spectrum protease inhibitor, we prevent the activation of both probes. It shows that we can use the IHZ method to assess protease activity directly in tissue. Next slide, please. We've used this method to assess protease activity in hundreds of tumor samples in multiple indications. And an example is shown on the left with 2 samples from triple-negative breast cancer. In this case, we're acquiring the tissue with 2 different Probody probes, 1 that is containing a substrate that's selective for urokinase plasminogen activator, or UPA; and 1 that is selective for MMPs. You can see that sample number one is cleaved by both probes or is lit up by both probes, demonstrating that it has both UPA and MMP activity, whereas sample number two only as MMP activity. We have identified signatures of protease activity across multiple tumor types and have merged this data with what we have learned from our systems biology approaches to develop a more complete understanding of the protease environment of tumors. We've also characterized the protease environment of preclinical models, and this is shown on the right. Here, we're looking at 2 xenograft models, H292 and the FaDu model. In this case, we're using antibodies that are selective for the activated forms of selected protease, such as matriptase and UPA. You can see that the H292 model has matriptase activity but not UPA activity, whereas the opposite is true of the FaDu model. This is important for 2 reasons: one, this shows that preclinical models have different profiles of protease activity similar to human tumors; and it also shows that we can use preclinical models to test and develop our protease substrates. I'll now describe how we're using what we've learned about tumor protease biology to develop and design what we call multiselective substrates. Multiselective substrates are substrates that are designed to be cleaved by multiple proteases in the tumor microenvironment. However, they only need to be cleaved by any one of those proteases in order to be fully activated. An example of that is shown below on the left again, using these same 2 xenograft models. Here, we're using 1 Probody probe that is specific for UPA, that's called a single substrate probe and 1 containing a multi-selective substrate that can cleaved by matriptase and UPA. You can see that while the FaDu model is activated by both probes, the H292 model only shows staining with the multi-selective probe because it lacks UPA activity, but the matriptase activity in H292 is sufficient to cleave the multi-selective probe. This carries forward into in vivo efficacy studies as shown on the right. We took the same 2 models into efficacy studies and treated them with anti-EGFR probodies containing either a single substrate or a multi-selective substrate. Similar to our ex vivo studies, you can see that the FaDu model was sensitive to both substrates. However, the H292 model required the multi-selective substrate in order to show efficacy. This demonstrates how multi selective substrates broaden the activity of protease activated agents. Next slide, please. So why is all this important? What we really care about, of course, is clinical activity in patients. We have been able to take all that we have learned about tumor protease biology and develop clinical therapeutic molecules that are designed to deliver broad activity. An example of this is pacmilimab, the first Probody that we put into the clinic. And this is a Probody targeting PD-L1. The waterfall plot on the left shows that pacmilimab has strong single-agent activity in multiple patients across multiple indications. We designed pacmilimab to contain one of those multi-selective substrates that's cleaved by multiple proteases in the tumor microenvironment. However, it only needs to be cleaved by any one of those proteases in order to be fully active, and this is why it shows such broad activity despite the likely heterogeneity of protease activity in these patients. We've also invested considerable effort in understanding the protease environment of metastatic lesions because we want to ensure that our agents are active in metastases as well as the primary tumor. This is most easily visualized in our imaging study. The patient on the right is an anaplastic thyroid cancer patient that achieved a confirmed partial response and was treated with Zirconium labeled pacmilimab. You can see that the tracer was taken up strongly by both the primary tumor as well as the metastatic lesions, showing that our probodies are activated in metastatic sites. You'll hear more about this asset later in the presentation because we are combining it with CX-2009 in 1 arm of our Phase II breast cancer study. Next slide, please. I'll now switch gears and spend just a few minutes talking about the versatility of the Probody platform. There are 2 components of our Probody platform that make the platform highly tunable. One is our affinity-based peptide masking strategy. We identify masks from large, highly diverse libraries of millions of peptides. And we select a range of masks that have a range of masking affinities against each particular therapeutic antibody. We also select our substrates from a panel of substrates that have a range of cleavability. And therefore, during the lead optimization process, we can select a particular mask and pair it with a particular substrate to generate the molecule with the widest possible therapeutic index. We know that conditional activation technology is not a one-size-fits-all approach. So our ability to tune our molecules based on the target and the therapeutic modality is a major advantage. Next slide, please. The Probody platform can be applied to multiple therapeutic modalities, and we've applied it to traditional antibodies, antibody drug conjugates, T cell bispecifics and cytokines. Antibody-drug conjugates are the topic of today's presentations with a particular focus on how our technology can unlock new targets, which addresses a major need in the field. And with that, I will -- thank you for your attention, and I would like to introduce Dr. John Lambert, the former CSO of ImmunoGen and a current member of the CytomX Scientific Advisory board. Thank you.

Thank you, Marcia, for the introduction. So I'm going to give a short presentation discussing the ADC landscape and with particular reference to targets and make some remarks about payload -- payloads. So the next slide, please. This slide depicts the basic components of an ADC. There is the antibody, which needs to be stable nonimmunogenic IgG. The target antigen ideally should be highly expressed on the surface of all target cells. It should be internalized efficiently and it should be largely absent from all other normal cells. The antibody is then linked to a cell killing agent via a conjugation process in our linker. Now the linker should be stable in the plasma. And ideally, the release of the payloads, the drug should be triggered only upon binding of the anybody to the target tumor cells. The payload itself should be a highly cytotoxic drug in the picomolar range so that it is toxic in the amounts that can be delivered by antibody delivery. It should kill cell tumor cells rapidly. It should be rapidly detoxified after release ideally. Another characteristic that the payload must be solidly enough for conjugation to proteins. After all an antibody is a protein and the chemistry must be done in an [indiscernible] environment. And importantly, the mechanism of cell killing of the payload, it should be active in the targeted cancer indication. Next slide, please. Now there are many challenges in designing and developing an effective ADC. First, there is the antibody from the blood plasma must penetrate from the blood stream into the tumor tissue. Once in the tumor tissue, the events of binding to the tumor target, and what happens after that. The target should be internalized, carrying the antibody and the ADC with it. First, in nonlysosomal compartments endosomes, ultimately to the lysosome where proteins go to be completely degraded. And then the linkers can be designed to release the payload in any of these places. But the other aspect of ADC design is what actually is released from the ADC? Is it the payload itself without any elements of the linker attached? Or is it a payload where a part of the linker chemistry still attached? And what the metabolites are then the metabolic profiles, are they charged? Are they neutral molecules? And how they distribute within the tumor is the next -- is another aspect of ADC design. Charge molecules will not be able to penetrate neighboring tumor cells, whereas neutral molecules that can diffuse across membranes can, in fact, diffuse into neighboring tumor cells and kill bystander killing activity. For example, tumor cells that have lower levels of target or a further away from the blood vessels, so they don't get enough payload delivered by antibody delivery alone. Next slide, please. An emerging aspect of ADCs -- actually, could you go back to the prior slide, is there the effect of all of these metabolites on the tumor microenvironment. This bystander effect can have effects on the growth of neovasculature and can also alter the immunosuppressive microenvironment of tumors. Also the mechanism of killing, the mechanism of cell death by the payload can also influence and suppress the immunosuppressive microenvironment. So all of these are aspects of have -- designing an effective molecule. Next slide, please. Now all of these challenges, they can be met. Each challenge is an opportunity after all the innovation. And up until now, I would say most of the innovation has been in the realm of medicinal chemistry. And in 2011, and in 2013, ADCETRIS and Kadcyla were approved and have become very effective agents, bringing benefit to cancer patients, and each of them now sell in excess of $1 billion. These approvals in the early 2000s sparked much innovation, the results of which are now being realized. When this slide was presented at the World ADC Meeting in October 2019, there were just 5 approvals at time. Now before the end of 2019, there were 2 further ADCs approved. And in 2020, 2 more were approved as denoted by the check marks in the bottom of the slide. And before -- right now, there are 2 ADCs that are -- for which BLAs have been filed and are under a regulatory view right now. These approvals, I would say that 1 thing to note here that it is clear that these ADCs are becoming an important class of drugs for treating cancer patients. But of the 9 approved ADCs, 5 were to targets for hematologic cancers. And there were only 3 different solid tumor targets amongst the 4 ADCs approved for solid tumors. This is a point I shall come back to later in the presentation. Before I come back to that, I would like to discuss what the ADC field has learned in regard to payloads and ADCs. Next slide, please. So I'll talk about the payload considerations. I use this slide to illustrate various aspects of the distribution of ADCs and their payloads in the body. Of course, the ADCs are administered intravenously. So they start off in the blood pool. And once an antibody diffuses into cancer tissue, it combined to the target and be taken up inside the cell releasing the payload. However, 1 must -- what has been learnt over the years is that the uptake of antibodies and ADCs by tumor tissue is about 0.01% of the injected dose per gram of tissue, about 24 hours after injection of the antibody or ADC. So for example, if 1 had a kilogram of tissue, that could take up about 10% of the ADC, you inject intravenously. It does mean that 90% of what you inject may never get into tumor tissue and doesn't have the opportunity to bind to a tumor cell before it gets metabolized elsewhere, and that's illustrated on the right side of the slide. So in the early days, if the linker between the payload and the antibody was unstable or was cleaved in plasma. Of course, the free payload would now distribute wherever the particular payload distributes in the body and is a source of potential toxicity. Also, of course, if there's less period attached per antibody molecule, less could end up on the ADC molecules that are in the cancer tissue. Now this aspect has been largely solved, thanks to the medicinal chemist designing better linkers and better ways of conjugation. So the upper panel, the upper route of metabolism of the payload is generally not so important anymore. However, that does leave the other aspects. So first, there is nontarget media uptake by fluid phase pinocytosis. This is entirely concentration dependent, independent of the target, independent of antibody type and such uptake will result in release of payload in those tissues. There is then also -- if I go to the next slide, as illustrated in the green circle is the target media uptake by normal cells, that is on target, but it's off tumor uptake. Now while the idea is to always have a tumor -- a target that is expressed only on cancer cells, actually, there are -- such targets are few and far between. Most targets do have -- for solid tumors do have some level of expression on some normal tissues. So that uptake can result in toxicity if the target expression is too high on particular cell types. Furthermore, such uptake by normal tissue, even if the antigen levels on the normal tissue a low enough such that it doesn't direct toxicity to those normal tissues, but a low level of expression on a lot of normal tissue will result in removal of ADC from the blood pool resulting in a shorter half-life and less ADC able to get to target tumor cells. If the concentration of ADC in the blood pool is rapidly decreased, it will compromise how much ADC can diffuse from the blood into tumor. And I would say before moving on, in an ideal case where no normal tissue expresses any target, all ADCs will have a maximum tolerated dose in the blood stream due to ultimately concentration dependent fluid phase pinocytosis, the middle panel on the right, and/or the effects of the payload on the elimination pathways since antibodies don't circulate forever, and there are processes that remove antibody from the circulation, albeit at a slow rate. So moving on to the next slide, I'll discuss some payload characteristics. This slide talks about the 2 payload classes, the auristatins as represented by monomethyl auristatin E and the maytansinoid as represented by DM4, which are well-validated payloads that interact with tubulin and kill cells by disrupting Microtubule dynamics. And this table shows that the ADCs made with these agents have predictable toxicity profiles. Looking at the upper panel with MMAE, there are 5 different targets represented there across several different indications. And yet the maximum tolerated dose on a 3-week schedule is -- for all of them, was about the same, about 2.5 milligrams per kilogram. And the dose-limiting toxicities doses exceeding the maximum tolerated dose were neutropenia, anemia and neuropathy for MMAE. So this is the ADC platform associated toxicity associated with MMAE ADCs, and it's independent of target. It reflects the ADC payload class, and it is and these toxicities become familiar to clinicians, these are the toxicities that have -- that should be managed with this class of agent. Looking at the bottom side -- bottom part of the table, DM4 ADCs. There are 6 different ADCs to 6 different targets represented here to a whole variety of different tumors. And again, the maximum tolerated dose on a 3-week basis is all about the same. In this case, about 6 milligrams of antibody protein per kilogram body weight. And the toxicities for all of those agents irrespective of target is reversible blurred vision and associated corneal keratopathy. Again, it's an ADC platform associated toxicity that is independent of target. It reflects the payload type. And in this case, a stable disulfide linker that once you want all aspects of free payload toxicity have been removed, the payload now will distribute where antibodies distribute. And it's clear that there are corneal epithelial cells represent a site where antibodies and ADCs can be at high concentration in the corneal -- in the capillary bed around the eye, and it appears that corneal epithelial cells are sensitive to then the non-target mediated pinocytosis uptake. Again, the tolerated dose, these toxicities are reversible and well managed. The next slide, I want to make some remarks about target itself, having talked about the payload. So -- and introduce the concept of conditional activation. So this slide, the ideal target is depicted on the left-side panel. It should be tumor specific, have minimal normal tissue expression. It should have expression on the tumor cells, it should have uniform expression on all tumor cells in the tumor tissue ideally. It should be highly prevalent in the tumor cells so that nearly all the cancers of the type actually express the target on all the cells, and it should be internalizing. So these would describe an ideal target that fits the criteria for an effective ADC. Now on the right, 1 thing is for hematologic disease, 1 can use lineage specific markets in hematologic cancers. For example, 1 can use targets that occur on all normal B cells and the cells of B-cell lymphoma. And you can kill all the cell types because the normal tissue, normal B cells, in this case, will regenerate from bone marrow stem cells. So hematologic disease offer the opportunity to use linear specific targets. However, these are not available for solid tumors. HER2 is a target that is -- has a long history, it's a target on solid tumors. It's highly expressed to levels of 0.5 million to more than 1 million copies per cell on the surface of HER2-positive cancers, for example, HER2-positive breast cancer, which is some tenfold or more than the level of expression on normal tissues. But the fact that 2 of the 4 ADCs approved for solid tumors are directed to HER2. And as shown in the panel on this slide, more than 25 different HER2 targeting ADCs are in clinical evaluation as of today. It's a testament to how few targets that are -- that meet all of the criteria in the top left for an ADC for solid tumors. The -- and so if I go to the next slide, please. However, there are many targets that would met all of the criteria except that of having minimal normal tissue expression. Now such targets are considered undruggable by ordinary ADCs. The normal tissue with high levels of expression will remove the ADC from circulation. That rapid clearance compromising how much will ever get into the tumor tissue. And secondly, delivery of the payload to the normal issue may result in unwanted target mediated toxicity. So both of these aspects render such targets undruggable using conventional antibody technology, but provide the strong rationale for conditional activation of ADCs binding only within the tumor microenvironment. Next slide. So if 1 can bring to bear a technology where an antibody binding only occurs in the conditions of the tumor microenvironment, then the landscape of available targets suitable for exploitation by an ADC is greatly expanded. And it's -- and on the left of the slide, expanding the funnel of available targets, at the top of the slide, in the bright red, expands the number of molecules that eventually will meet all the other criteria needed to make an effective molecule. All the other criteria and preclinical hurdles, such that the blue bottom of the funnel, it will yield a larger number of ADC candidates have a strong rationale for clinical evaluation. Next slide, please. With conventional approaches to identifying targets and antibody selection, the hunt for targets where a subpopulation expresses the target uniformly and at a level of expression sufficiently amplified over normal tissue such that an ADC may have a therapeutic window, already then narrows the proportion of patients that one can address with a particular technology. I again refer to HER2 as an example. While it's overexpressed in HER2-positive breast cancer, HER2-positive breast cancer represents only 15%, perhaps up to 20% of breast cancer, which leaves 80% of breast cancers unaddressed by that target. However, with conditional activation, one can look for targets with high expression and uniform tumor expression without worrying about whether this high expression uniform expression on normal tissues. Thus many more target options that can serve a much larger patient population are available. On the next 2 slides, I'll just show a couple of examples. First example is CD166 or ALCAM. In -- on the right of the slide, in the top panel shows on a log scale, the expression of CD166 on breast cancer and ovarian cancer. And you can see from the tight bar on the left of the top panel, the level of expression is as high as HER2 and Trop2, for example, 2 targets that are validated as ADC targets. And that's true of both breast cancer in the top panel and ovarian cancer underneath it. However, the bottom panel also shows the immunohistochemistry staining with 3 different tumors, prostate, breast and non-small cell lung cancer with very, very strong membrane staining would make an ideal target. However, the very bottom set of 3 panels show that this target is highly expressed on a lot of normal tissue. And so this was a target that was undruggable by conventional antibody approaches. But it otherwise would be a highly desirable ADC target due to its high tumor expression. But it absolutely requires a conditional approach due to this expression on normal tissue. The next slide shows another example, that is CD71 or the transferrin receptor. Again, another previously undruggable target by conventional approaches. It's a transmembrane glycoprotein that's highly expressed on malignant cells because rapidly growing cells have a high need for iron for the metabolic enzymes that require iron. And on the right-side panel shows you an experiment about it's internalization. So at time 0, the total ADC shown in green and ADC on the surface, which is shown with the red fluorescence, they're one and the same at the starting point of this experiment all the ADCs bound to the surface. But you can see that within 4 minutes, all of the surface bound material is within the cell, which is transferrin receptor is a professional internalizer of transferrin carrying iron and in this case, carrying an antibody or an ADC. So it has these highly desirable ADC properties due to its highly high internalization rate, but it can't be targeted by traditional approaches because of the high expression on erythrocyte precursors. It's also, of course, need to import iron for the hemoglobin. And indeed, even unmasked the naked antibody itself has some toxicity. So an unmasked ADC while it has potent activity on -- in preclinical models, it has no therapeutic index because of the presence of transferrin receptor on erythrocyte precurses. Next slide. So in summary, I think the pace of ADC approvals has accelerated rapidly in the past couple of years. I think the era of medicinal chemistry have provided lots of medicinal chemistry tools of payloads and linkers and chemistries of linking the linker payload to antibodies. So now there are lots of medicinal chemistry tools in the toolbox. I think the next advances in ADC technology are going to be innovations at the target antibody part of the equation to take advantage of this plethora of additional chemistry linker payload tools. The ADC target landscape of approved agents has really been limited to heme targets, say, 5 of the 9 approved ADCs are to hematologic targets. And a few selected solid tumor targets, but conditional activation technology broadens the target space for ADCs enormously to be able to take advantage of this medicinal chemistry toolbox. And CytomX is advancing to such conditionally activating ADCs against novel targets that couldn't otherwise be exploited as ADCs into Phase II studies. And with that, thank you very much.

All right. So you've heard now from Doctors Belvin and Lambert on the versatility and functionality of our Probody platform, which is the underpinning of our deep and maturing pipeline, shown on Slide 38. What I'll ask for our speakers is that if they can occasionally state the slide number they're on so that the presenters can make sure that everything is staying on track. So Slide 38 is the pipeline. It's a high-level schematic indicating where we are today with regard to clinical development. We have 4 molecules in 11 Phase II studies spanning 9 different indications. Focusing on our clinical programs, praluzatamab ravtansine or CX-2009 is being evaluated in 2 subsets of breast cancer as monotherapy and in combination with pacmilimab, our Probody to PD-L1 in triple-negative breast cancer, and you'll hear more about this from Dr. Sara Tolaney. CX-2029, which is partnered with AbbVie, and for which we retain 35% of co-development rights in the U.S. is now being tested in 4 expansion cohorts as listed here, and you'll hear more about this from Dr. Melissa Johnson. For these 2 programs, we anticipate initial data emerging from certain cohorts by the end of this year. BMS just opened 3 single-arm cohorts to evaluate the Probody to ipilimumab, BMS-986249 in combination with nivolumab. And this, in addition to the previously initiated randomized study evaluating the combination versus nonmasked ipi plus nivolumab in frontline melanoma. So that's a great advancement from our perspective. BMS is also evaluating the safety tolerability of the Probody to a nonfucosylated potentially more potent, but also more toxic, if not masked version of ipilimumab, again, as monotherapy and in combination with nivolumab. We also have 2 new programs in the IND-enabling phase, which we hope to have in clinical testing in 2022. Slide 39, well, I'm going to introduce to you Sara Tolaney, who will talk about CX-2009 and CX-072. Dr. Tolaney is a highly regarded academic clinician focusing on developing novel treatments in women's cancer, through her appointment at the Dana-Farber. Dr. Tolaney is particularly interested in IO combinations, and we are pleased to be able to tap into her knowledge as a member of our steering committee for this program. Dr. Tolaney, please take the floor.

Thank you so much, Amy. I appreciate it. So I'm going to be giving an overview of the clinical data surrounding CX-2009. So if you go to the next slide, [Technical Difficulty] cancer represents about 30% of all cancers in women, and it's estimated that there are about 275,000 new cases and about 42,000 new deaths that occur each year in the United States. About 80% of breast cancer is HER2-negative and despite recent advances, new therapies are urgently needed for patients with metastatic HER2-negative breast cancer. We know that CD166 is highly expressed in HER2-negative breast cancer, and therefore, really represents a very attractive target. If you go to the next slide. We know that antibody-drug conjugates have been an effective therapy for breast cancer. And as of today, we have 3 ADCs that are available for treating patients with breast cancer with approved agents currently targeting either HER2, such as trastuzumab emtansine or trastuzumab deruxtecan or targeting Trop2 as in the case of sacituzumab govitecan. We also have promising data that has led to several registration designs in patients with hormone receptor positive HER2-negative disease, such as the TROP-02 study, looking at sacituzumab. Also registration trials ongoing, looking at HER2 low positive breast cancer, specifically with trastuzumab deruxtecan and also ADCs in registration focused on HER2-positive disease, such as trastuzumab duocarmazine. And finally, there are multiple earlier stage trials looking at ADCs that are targeting either Trop2, HER2, HER3, PD-L1 and LIV-1. And so ADC targeting CD166 represents a really novel strategy for patients that have either hormone receptor positive or triple-negative breast cancer. Next slide. So CX-2009 is a first-in-class Probody drug conjugate targeting CD166 with the maytansinoid DM4 payload, which is a well-characterized and potent microtubule inhibitor. CD166 is an adhesion molecule that has a role in cell growth, survival and motility, contributing to tumor invasion and progression. In breast cancer, there's certainly some literature to suggest that CD166 has a specific role in the survival of breast cancer cells and the references are shown here on the slide. CD166 expression in normal cells really precludes the development of a conventional ADC, given the likelihood of substantial toxicity to normal tissue. We do know that CD166 is expressed on many cancer types, and there is Phase I activity signals that have been detected in other cancers such as ovarian, lung and head and neck based on previously published work that has been presented at ASCO last year. Today, though, we're going to focus on patients with breast cancer that have been treated in the Phase I trial. If you go to the next slide. The Phase I study enrolled a total of 99 patients, of which 39 had been diagnosed with metastatic HER2-negative breast cancer. This slide presents their baseline demographics. And you can see that the trial enrolled 28 patients who had hormone receptor positive HER2-negative metastatic breast cancer and 11 patients with metastatic triple-negative breast cancer. And as you can see, these patients were heavily pretreated, with a median of about 7 prior regimens for their metastatic disease, and this includes either a cytotoxic or endocrine therapy. And I think it's also important to note that this really represents a modern Europe population because as you can see, most triple-negative patients had received prior platinum and about 40% had prior checkpoint inhibition. And you can see that the majority of the hormone receptor positive patients had a prior CDK4/6 inhibitor. When looking at CD166 expression in breast cancer patients, as you can see in the figure to the right, nearly all hormone receptor positive patients expressed CD166, as you can see in about 26 of the 28 patients, whereas triple-negative breast cancer has a different immunohistochemical profile with about half or 6 of the 11 patients expressing CD166. And this data certainly led to the development of the Phase II trial that really incorporated this information using a selection strategy for CD166 positivity only in the triple-negative subgroup of patients. And we can go to the next slide. In this toxicity table, you can see that doses between 0.25 and 6 mg per kg IV of CX-2009 administered every 3 weeks are shown and combined on this table. The selection of the 7 mg per kg 3-week dose as the recommended Phase II dose was supported by tolerability of this dose for chronic administration signs of antitumor activity and the PK/PD modeling using a trough concentration that was established in vivo mouse efficacy experiments. We see a toxicity profile really consistent with the DM4 payload, including ocular, neurotoxic and hepatic toxicity. Looking at higher dose levels, there is a dose-dependent increase in grade 3 treatment-related toxicity, specifically with ocular toxicity in both all grade and grade 3 and higher toxicity. But I think it's really important to note when looking at this Phase I trial that use of mandatory ocular prophylaxis did not occur in this study that prophylaxis was initially optional at the 8 mg per kg dose and then only became mandatory at 9 and 10 mg per kg. But as we move into Phase II testing, use of prophylactic ocular medications will be mandatory. Now we can go to the next slide. When looking across toxicities of payloads that have been used in ADCs, we see that DM4 as an ADC payload has been in clinical trials for solid tumors as well as for hematologic malignancies for more than 10 years. And toxicities associated with this payload have been very well characterized. In general, we know that patients who have solid tumors treated with a DM4 containing ADC do not experience hematologic toxicity. However, both peripheral neuropathy and ocular toxicities can be seen. Other DM4 containing ADCs do not typically show severe toxicities in patients with solid tumor malignancies, but mild-to-moderate neuropathy and ocular toxicities can be seen. Next slide. Overall, ocular toxicity is the most common adverse event associated with any DM4 conjugate ADC, including mirvetuximab, which is also an ADC with the DM4 payload that is moving forward in ovarian cancer. The shared toxicity across ADCs that target different antigens suggest ocular toxicity is off-target and an antigen independent toxicity. The mechanism is presumed to be due to pinocytosis of DM4 by the epithelial layer of the cornea and common symptoms can include blurry vision, keratitis, dry eyes, microcytic epithelial damage. And if the toxicity is grade 1 to 2, patients are usually asymptomatic or demonstrate very mild symptoms. And usually, the intervention could be limited to something like artificial tears. However, if the toxicity is grade 3, there are usually changes that can be found in visual acuity. And then treatment is needed with either steroid eyedrops, vasoconstricting drops on the day of infusion and cold compresses as tolerated during infusion. And so the Phase II trial of CX-2009 in breast cancer will incorporate these mitigation strategies. At the recommended CX-2009 dose of 7 mg per kg and less, most ADC-related ocular AEs are not severe and are usually reversible. Next slide. Here, you can see there are 2 waterfall plots. The one on the left is for patients with hormone receptor positive HER2-negative breast cancer, and on the right is the waterfall for patients with triple-negative disease. These waterfalls only show those patients who received a dose of 4 mg per kg or higher. And you can see an H score for each patient is represented either above or below the individual bar. Patients who have hormone receptor positive breast cancer, as you can see, have very high CD166 expression by IHC. And as noted previously, patients who have triple-negative disease have a more variable pattern of CD166 expression. Of note, if you look at the triple-negative breast cancer patient with the best response, that patient was not evaluable for CD166 expression due to inadequate number of tumor cells within the specimen. Looking at the efficacy evaluable subgroup of patients, which in this trial was defined to be those patients who had at least 1 measurable lesion at baseline and then went on to have at least 1 on study scan, there were 2 confirmed responses by resist in the hormone receptor positive group and 3 unconfirmed responses in the triple-negative group. The clinical benefit rate at 16 weeks was 41% and at 24 weeks was 28%. For a population that has had a median of 7 to 8 prior regimens in the metastatic setting, I think many of us do feel that clinical benefit rate sometimes is more meaningful than objective response in this particular population. And these numbers suggest moving forward into a Phase II clinical setting with CX-2009 in patients who have advanced breast cancer. If you go to the next slide, you can see the spider plot for both subgroups of breast cancer patients combined into one. And as you can see here, there are several patients that had stable disease that went out beyond 24 weeks with 3 patients reaching nearly 1 year of therapy. One patient with hormone receptor positive breast cancer did achieve a complete response in the 2 measurable target lesions, but was felt to be a partial responder due to the evaluable disease remaining on her bone scan. Next slide. Here are the CTs and photographs of cutaneous lesions in a patient who had metastatic triple-negative disease, this patient had previously received upfront therapy with a taxane and checkpoint inhibitor, then went on to receive sacituzumab govitecan and then went on to this trial with CX-2009. You can see on the CT scan that there was involvement of the chest wall and the axilla, but you can see, I think, much more clearly this involvement on the photographs. You can see that the skin lesions show both reduction in size and in epithelial healing following treatment with single-agent CX-2009. Next slide. This shows the schema for the Phase II trial design. This trial is for patients who have either hormone receptor positive HER2-negative metastatic disease or for patients who have triple-negative breast cancer. For those patients who have metastatic hormone receptor positive disease, they're allowed to have 0 to 2 prior lines of chemotherapy for their advanced disease and must have progressed on a prior CDK4/6 inhibitor, either in the metastatic setting or having received one in the adjuvant setting. For those patients with triple-negative disease, these patients do need to be prescreened for CD166 and need to have either 2 plus or 3 plus staining by IHC on either archival tumor tissue or a prescreening biopsy. Patients with triple-negative disease must have received one to 3 prior lines of cytotoxic therapy in the advanced disease setting. And if someone has an early relapse after adjuvant therapy that would be counted as a line of therapy. The study is a 3-arm trial. As you can see, there are 2 arms that are investigating CX-2009 as monotherapy. One that is looking at the metastatic hormone receptor positive patients and another that is looking at those patients with triple-negative disease. And then the third arm will look at the metastatic triple-negative patients using a combination of CX-2009 with the CytomX checkpoint inhibitor. Since these patients are going to be treated with an immune checkpoint inhibitor. These patients do have some additional eligibility requirements. So for this particular cohort, patient cannot have known PD-L1 negative disease or unknown status or if the patient has received a prior checkpoint inhibitor, they cannot have what is defined to be checkpoint refractory breast cancer, meaning that they can't have disease progression within 120 days of their first dose of checkpoint inhibitor. And certainly, those patients going on to the immunotherapy arm cannot have an active autoimmune condition. So just as a way of background for the combination arm, arm C. We wanted to review some of the data that supports combining this particular ADC with a checkpoint inhibitor. So here, you can see that CX-072 is a conditionally activated or masked checkpoint inhibitor targeting PD-L1. It was studied in a Phase II clinical trial, enrolling patients with multiple solid tumor malignancies that did include patients who had triple-negative disease. The waterfall plot from this study is shown on the left of this slide. There were 12 triple-negative breast cancer patients who were evaluable for activity, with 3 patients showing reduction in their target lesions. Two of the patients with cutaneous lesions are shown in the slide, and you can see there is clearing of their skin lesions in excess of 72 weeks and 20 weeks, respectively. Of note, patient A actually had PD-L1 low positive disease and patient B was PD-L1 negative. So I think this suggests that both CX-2009 and CX-072 have shown signs of single-agent activity in triple-negative disease. If you go to the next slide. We do know that antibody-drug conjugates are able to both kill cancer cells and modulate immune response. And we know that these cytotoxic payloads can induce immunogenic cell death, both in vitro and in vivo. And cytotoxic payloads also provoke phenotypic and functional dendritic cell maturation and activation, and so there is potential for synergistic antitumor activity when combining an ADC and a checkpoint inhibitor. And this has been explored in mouse models. As we'll see on the subsequent slides, neither CX-2009 nor the anti PD-L1 Probody monotherapy result in regressions in mice, but the combination did produce tumor regressions in about half the mice. Next slide. So the upper left-hand figure shows vehicle control. The upper right-hand figure here shows the CytomX Probody checkpoint inhibitor alone. The lower left figure shows CX-2009 alone. And you can see that the combination in the lower right-hand figure with approximately half the mice, as highlighted in red, do demonstrate tumor regression in this syngeneic model. Next slide. So in addition to the CytomX nonclinical data, there is clinical data to support the combination of a maytansinoid ADC with a checkpoint inhibitor in patients with breast cancer. We do have data from the randomized Phase II K2 trial, which looked at patients with metastatic HER2-positive breast cancer and have randomized them to receive T-DM1 with or without atezolizumab. This study did demonstrate the combination of T-DM1 with atezolizumab did improve objective response from 33% to 54%, and importantly, extended progression-free survival with a hazard ratio of 0.6, and this was specific in those patients who had PD-L1 positive disease. And I think these data indicate that there is potential clinical benefit for combining a PD-L1 antibody and maytansinoid ADC in a breast cancer population. Next slide. So in conclusion, the Phase I study of CX-2009 demonstrated that CD166 is a viable first-in-class therapeutic target for patients with advanced malignancies, including breast cancer, Probody technology that utilizes masking the drug until it is activated by proteases in the tumor microenvironment, enables the administration of CD166-directed antibody drug conjugate, at tolerable doses with signs of clinical benefit in patients with either HER2 nonamplified breast cancer that's hormone receptor positive or that is triple negative. The trial really established a recommended Phase II dose of every 3-week treatment and supported the activity of this agent as well as the tolerability. There were ocular toxicities that we're seeing that are consistent with having a DM4 payload, and it's hoped that utilizing a prophylaxis strategy that was started in the Phase I trial can translate into less severe and manageable toxicity in the Phase II study. Optimization of CD166 IHC assay is ongoing to support a potential selection strategy, particularly for those patients with triple-negative breast cancer and enrollment onto the Phase II study in breast cancer is currently underway in that 3-arm design that we had shown. Thank you so much.

Thank you, Dr. Tolaney. Very much appreciate your participation and look forward to the Q&A. Our next guest speaker, Dr. Melissa Johnson has been working with CytomX for some time actually. Participating both on the Phase I dose escalation of CX-2029, our conditionally activated ADC to CD71, the transparent receptor, and continues to participate in the Phase II expansion cohorts. Dr. Johnson works not only in early drug development. She is also the program director for lung cancer at, and we're pleased to have her expertise and experience with this asset in Phase I and in lung cancer as we embark on the expansions in the Phase II studies in targeted indications including squamous non-small cell lung. Dr. Johnson, I'll pass the mic to you. Thank you.

Thanks, Amy, for the introduction. And thanks for the opportunity to participate today. Let's jump right in, talk about CX-2029. Let's -- we're going to go to Slide 58. So as we heard earlier, CD71, also known as the transferrin receptor aids in the internalization of iron into many rapidly dividing cells. And so while it is a highly expressed tumor antigen in tumor types, as you see on this slide, including lung, squamous, esophageal and lymphoma, it is also highly expressed in red blood cell precursors, also known as reticulocytes, which has made it difficult to target as a traditional antibody drug conjugate. So it also lends itself very nicely to the Probody strategy. We've heard about earlier, where only when the drug is in the tumor microenvironment and the mask can be cleaved by the proteases, does the antibody bind to CD71. Next slide. So we're going to discuss or review some of the data that was reported at ASCO in 2020. It was a phase -- we reported a Phase I dose escalation study, evaluating 6 202one administered IV once every 3 weeks in 48 patients with advanced solid tumors. You can see that doses from 0.one milligrams per kilogram all the way up to 5 milligrams per kilogram were explored. The dose is 2 milligrams per kilogram and above or thought to be biologically active doses. And patients were enrolled with advanced unrespectable solid tumors, they had to have archival tissue or be willing to undergo a biopsy at the time that they enrolled and stable brain metastasis were allowed. Only patients with transfusion-dependent anemias were excluded as well as grade 2 or higher neuropathy. On the right side of the slide, you see the key patient demographics the 45 patients enrolled in the Phase I study. The median age was 60. Patients had a good performance status. You see that the CD71 expressed in the tumor sample varied about 1/3 had high expression, 1/3 low expression. And 1/3 was unknown, either due to the tumor tissue, not expressing it reliably according to assay or not having tissue that was able to be analyzable. The tumor types that were included are also shown in this slide. Importantly, about 20% were lung cancer, 20 -- around 20% were head and neck, squamous cancer and another 20% colorectal cancer and almost 50% other solid tumors. This was a heavily pretreated group of patients. We'll move on to Slide 60. This slide is a summary of the tolerability that supports the recommended Phase II dose of 3 milligrams per kilogram. You can see some of the adverse events of special interest across this trial given the MMAE payload are shown in the table, anemia and neutropenia. We also saw infusion-related reactions and then the DLTs are summarized at the bottom. You can see that both 4 milligrams per kilogram and 5 milligrams per kilogram exceeded the tolerable doses with 2 DLTs reported in each of those cohorts. And you also see from a macroscopic view, we'll dive into the anemia a little bit -- in a little bit more detail on the next slide, but you can see that anemia existed in -- across all cohorts. It was not -- but it was dose dependent, as you see, increasing over the cohorts. Likewise, neutropenia seems to be dose dependent. By contrast, infusion-related reactions happened seemed irrespective of dose. We'll see some PK as we go through these slides, but we know that greater than 90% of the Probody drug conjugate was masked in the circulation, and we'll review the Grade 3 adverse events further on subsequent slides. We can go on to Slide 61. Thank you. So let's talk a little bit more about the anemia. The -- as I told you, the CD71 expression on red blood cell progenitors is quite high. And so it makes sense that it is perhaps -- in targeting those cells that we can perhaps explain the anemia. We'll show you in some more slides the fact that when you have a patient that is experiencing anemia, the normal response in the bone marrow would be to increase the reticulocytes, the progenitor activities. Basically, the bone marrow pumps out more reticulocytes, and you see an increase in the reticulocyte count. By contrast, we actually saw a significant decrease in the peripheral reticulocyte counts of patients in the trial who were treated at doses greater than 2 milligrams per kilogram, suggesting this on-target, off-tumor toxicity. Other mechanisms of anemia were ruled out -- or we believe have been ruled out by the fact that the anemia that we see is typically normocytic and normochromic, and so made it less likely to be a mechanism of blood loss or decreased production due to decreased iron metabolism, for example, or a mechanism of increased destruction like hemolysis. Next slide. So here, you see a nice representation of the data that I just summarized. At higher doses of CX-2029, we see that is associated with a steeper decline in hemoglobin. So on the left-hand side, that graph summarizes all patients in the lower nonbiologically active doses, 0.1 to 1 milligram per kilogram, without a decline in hemoglobin over the first 21 days of cycle 1. But on the right, you see the hemoglobin of patients treated at 2 to 5 milligrams per kilogram with a pretty predictable slope downward in the first 2 weeks of therapy with some then rebound prior to cycle 2. Next. On this slide, you see summarized the dose-dependent effect of CX-2029 on reticulocytes. So with increasing doses of 2029, we see different effects on the reticulocytes. First, on the very left hand panel, you see almost no effect in those lower doses. In the middle panel, you see patients reticulocytes treated at doses of 0.5 to 2 milligrams per kilogram, where you see the hemoglobin drops by day 8 or by day 10. And then the reticulocyte count predictably goes back up, which is what we would expect as the bone marrow is trying to respond to a state of anemia. By contrast, on the right-hand side, you see a panel that shows patients treated at doses 3, 4 and 5 milligrams per kilogram. The reticulocytes drop in the first 10 days of treatment and then they are unable to rebound because reticulocytes have been targeted. Next slide. Here we see a summary of the pharmacokinetic data. We see that the drug remains intact in the circulation, greater than 90%. You see that the PK parameters were well-behaved and dose proportional with a terminal half-life, somewhere between 2 and 9 days, and the free payload MMAE that was circulating only about 4% of the total. Next slide. On this slide, you see summarized the clinical activity of patients treated in our Phase I study at doses of 2 milligrams per kilogram and higher. So first, you see a waterfall plot on the left, showing the dose that patients received, color coded: yellow for 2 milligrams per kilogram, light green for 3, dark green for 4, and orange for 5 milligrams per kilogram. And convincingly, we see confirmed partial responses in a couple head and neck cancer patients and a squamous non-small cell lung cancer patient. And notably, all but one had doses of 3 milligrams per kilogram. You can see this data also summarized nicely across all of the patients in the trial in the spider plot on the right, showing that not only did we achieve response, but in many cases, it was durable. Next slide. So this is drilling down on the same efficacy endpoints, the waterfall plot and the spider plot illustrating just those patients with squamous lung cancer and squamous head and neck cancers as those are 2 of the target populations in the expansion efforts treated at 2 milligrams per kilogram every 3 weeks and above. And so you can see them labeled nice responses in the squamous patients. And these are also among the patients with more durable responses seen in the spider plot. Next. So what follows now are 2 case studies 2 particular patients from our experience, first, a 66-year-old patient with squamous head and neck cancer, nasopharyngeal histology diagnosed initially in February of '18. He was previously treated with chemoradiation with high-dose cisplatin as well as with an investigational agent in combination with pembro as a second line therapy, pretty refractory to these therapies about 3 months on both of these lines before joining our trial, CX-2029 in January of 2020. The patient was enrolled in the 3-milligram per kilogram dose cohort was dose reduced for anemia in March. And as you can see from the CT scans at the bottom, patient had a partial response, 43% shrinkage. In particular, in the liver lesion noted on his week scan -- his week 8 scan, sorry, confirmed again 8 weeks later. Continues to dose with continued shrinkage at 2 milligrams per kilogram. Next slide. Second case study then a 75-year-old patient with squamous non-small cell lung cancer presented with a stage III locally advanced cancer in August of '17 received chemoradiation, followed by maintenance durvalumab for about a year before progressing on 2 other lines of cytotoxic chemotherapy, gemcitabine and then docetaxel ramucirumab, before joining the trial also in January 2020. This patient, likewise, was enrolled at 3 milligrams per kilogram. It's notable that the patient did receive both red blood cells as well as erythropoietin stimulating agent, ESA. Darbepoetin was subsequently dose reduced 2 milligrams per kilogram in April a year ago. Irrespective of the dose reduction, you see a nice partial response at week 8 in both the right perifigural target lesion as well as the left lower lobe large target lesion and the sort of ground glass on the right lower lobe, that was then confirmed at week 16. Next slide. So based on this Phase I data, the Phase II expansions were launched to evaluate, in particular, squamous lung cancer -- squamous head and neck cancer as well as esophageal cancers. The eligibility are spelled out on the left-hand side of this slide. Patients should have received a prior platinum as well as a checkpoint inhibitor alone or in combination per local health authority for squamous -- or sorry, for esophageal cancer, squamous or adenocarcinoma or GE junction is allowed prior HER2 therapy required if the tumor has been found to express HER2. And then at least -- progression on at least one prior line of therapy in the advanced setting. A fourth cohort -- and sorry, you just go back. Not quite ready. A fourth cohort in expansion was the diffuse large B-cell lymphoma cohort. Patients are eligible after progression on at least 2 prior lines of therapy, one of which should be a CD20-containing regimen, and these are patients that would not be eligible for stem cell transplant. As you can see, 25 patients in each of these cohorts will be enrolled with primary endpoint objective response and secondary endpoints, including PFS, DCR, clinical benefit rate at 24 months as well as duration of response. Overall survival, safety, PK, 88 and time to treatment recurrence. Next slide. So just to summarize the data that I've described today, we found CX-2029 to be safe. It produces dose-dependent hematologic toxicities consistent with the MMAE payload and likely also due to the expression of CD71 on reticulocytes and on target off-tumor effect. Anemia was the most common toxicity seen also seen in preclinical species and the best order and mechanisms of responding that anemia are part of the investigators focus currently. CX-2029 at 3 milligrams per kilogram will be studied in the dose expansion phases of the study. There was no dose-limiting toxicities with this dose during the Phase I portion and no patients discontinued for toxicity at this level. Patients who are treated at 3 mg per kg had a manageable grade 3 anemia, frequently assessed with routine labs treated with transfusions as well as ESAs, dose reduction and dose delay. Finally, clinical activity was observed at 2 milligrams per kilogram and so dose reduction and the PK ramifications are also under investigation. Thanks. To conclude, the results of this first-in-human trial validates CD71 as a viable therapeutic target in cancer. The Probody technology enables administration of a CD71-directed antibody drug conjugate at tolerable doses with clinical antitumor activity, which would not have been possible with a conventional ADC. And the safety profile and clinical activity support dose expansion in head and neck cancers, squamous lung cancer, esophageal cancers and diffuse large B cell. Thanks.

Great. Thank you, Melissa, and thank you to all of our speakers. Hello, everybody, again. Before wrapping up and moving to Q&A, which we'll do in just a few minutes, I know we have a number of questions queued. I wanted to spend a few minutes actually on an exciting new research frontier at CytomX as we continue to innovate with our technology. We focus most of our time today on the antibody drug conjugate space, which, of course, is one of the key applications of our technology. And Marcia commented earlier on, in the session that one crucial attribute of the Probody platform is, of course, its versatility and its tunability. So having applied our technology to date and over the last few years to checkpoint inhibitor antibodies, antibody drug conjugates, as you've heard, bispecifics, which were not the focus of today's discussion. We're now also directing the power of our platform to modification of cytokines, and I want to close by briefly reviewing some exciting preclinical data on a masked conditional form of the potent immunemodulated interferon alpha. So as we all know, cytokines are highly potent immune modulators and major regulators of the innate and adaptive immune systems. Through their immune and nonimmune functions, many cytokines have been shown to elicit broad-based anticancer activity in preclinical models. And in fact, these results have previously translated into the early approvals of interferon alpha and IL-2, of course, for cancer treatment. But that was well before the era of checkpoint immunotherapy. However, the clinical success of cytokines has been limited by systemic toxicity and poor exposure due to the very high potency of these agents. So you might imagine that localizing cytokine therapeutics into cancer tissue with conditional technology like the CytomX Probody platform could have significant potential to reduce systemic toxicity, increase therapeutic index and enable single-agent and combination strategies. So a new work at CytomX, which was recently reported at the 2021 virtual Cytokine-Based Cancer Immunotherapies Summit, we have explored our hallmark affinity peptide masking approach, in addition to novel steric masking approaches, to creating protease conditional cytokines. We have an increasingly broad program in this area at the company across multiple cytokines and cytokine families, but I wanted to give you a snapshot -- a quick snapshot of our work today with the engineering of a conditional interferon alpha 2B. Interferon Alpha 2B is a member of the type 1 interferon family that has pleiotropic activities, including antiviral activity and immunomodulatory functions. Clinically, interferon alpha 2b has been approved for both antiviral and anticancer therapy, but systemic administration is accompanied by dose-dependent toxicities. Recent clinical data from others suggest that local delivery of interferon alpha 2b can be a safe and effective treatment in BCG unresponsive bladder cancer. So taken together, these factors suggest that a localized conditional version of interferon alpha 2B could be a more effective, systemically administered anticancer treatment. So we have successfully engineered in our labs and optimized protease-activatable versions of interferon alpha 2B as shown on this slide that have more than 5,000 fold masking in cell-based assays. This is something that we've done now across many different modalities, as you've seen, and it's exciting now to be moving into the cytokine space. The masked interferon alpha 2B induces complete tumor regressions in preclinical models with comparable efficacy to unmasked recombinant interferon alpha, as shown here in the Daudi tumor model. And importantly, masking appears to confer the anticipated safety benefit that would accrue from masking, as evidenced by analysis here in this hamster model of weight loss and liver function, where you can see that the masked agents have significantly improved safety compared to the unmasked interferon. So taken together, these encouraging data demonstrate the potential of using CytomX technology to engineer cytokines towards improved therapeutic index. And we continue our work in this exciting area of R&D, and we very much look forward to providing updates as this work progresses in the future. The full presentation from the recent cytokine summit is available on our website. I just highlighted a few key slides from that presentation today, given how excited we are about this new avenue of R&D at our company. So we certainly covered a lot of ground today, and thanks to all of you for staying with us, and particularly for those of you who may have had a few technical difficulties here and there. But to summarize, CytomX is the leading innovator in protease-activated conditional biologics. We believe we are uniquely positioned to leverage our platform across multiple modalities and cancer types. We have ongoing clinical studies of 5 Probody therapeutics, 4 Phase II programs across 9 cancer types. Initial readouts from CX-2009 and 2029 that we took a deep dive on today are anticipated beginning in Q4 of this year and into 2022. And we also have multiple emerging preclinical programs, including now conditional cytokines. So thank you for your attention. The -- I think with the time remaining, let's move to Q&A. We do have our guest speakers still online here. So with that, let's move to Q&A. And I think operators are ready to open the line.

[Operator Instructions] Your first question comes from the line of Mara Goldstein with Mizuho.

First, I wanted to ask Dr. Johnson about the anemia-associated effect with CX-2009 -- excuse me, 2029 and my question relates to the comparability of, let's say, chemotherapy treatments that utilize supportive therapy due to anemia-related effects, and how comparable that is to the baseline effect observed with CX-2029, which appears to be higher than other anemia-inducing therapeutics. So based on your experience, do you think that similar supportive therapies can modulate toxicity to levels observed with other agents? And then secondarily, for Dr. Tolaney, I also have another adverse side effect-related question. And that has to do with the use of paclitaxel in breast cancer and the potential for patients who have already experienced neuropathy to experience synergistic neuropathy with CX-2009.

Maybe I'll go first. There were a lot of questions, but I'll try to hit them all, and let me know how I do. This is, to me, a little -- I have to say, patients, despite a 2-gram hemoglobin drop that we find on their labs, feel well. They don't necessarily have the dyspnea with exertion, the shortest of breadth that we associate with anemias with red blood cells in the 7, 8, 9 range, which is where the hemoglobins tend to fall. Over time, patients do look a little bit more pale. Usually all still feeling well, but they look anemic. So I think that I'm excited and optimistic about the ongoing trial to help us out how to use ESAs to help us. I think in lung cancer, in particular, we haven't used ESAs much in the last 10 years to mitigate the anemia associated with chemotherapy, but I think the mechanism is different here, so I think it's worth a shot. And many of the other investigators share that view. So I think we'll see. I will -- the last comment I'll make is, it doesn't seem that transfusions alone will mitigate it. And so it's going to be -- I think, our focus has turned increasingly to the reticulocytes.

I'm happy to take the question regarding paclitaxel utilization in breast cancer and potential concern for any neurotoxicity related to CX-2009. Certainly, paclitaxel is a very commonly utilized chemotherapy for metastatic disease, both in hormone receptor positive as well as in triple-negative breast cancer patients. And so there are some patients who can certainly be left with some residual neuropathy. Although that's certainly being a small proportion of our overall population. When thinking about sequencing to a subsequent line of therapy such as CX-2009, certainly, one does keep underlying neurotoxicity as a thought in one's mind when making treatment decisions. But I think of note, it's important to realize that the rate of grade 3, 4 neurotoxicity, at least from the Phase I study, did appear to be quite low and was only 8%. And so it's not sort of very common to have high-grade neurotoxicity. And certainly, we have no data to suggest that there would be synergistic toxicity. Obviously, these agents are not being given together. And I don't think that, that would be of concern here. It would be more if someone had residual neuropathies from previously administered therapies coming into treatment that you'd be concerned about.

Your next question comes from Roger Song with Jefferies.

Thank you for all the speakers for all those great presentations. I have 2 questions. One is for Marcia. So it's very intriguing, you have this kind of protease activity screening library. My question is just how broad is the tumor screening library? Specifically, do you see -- like do you have those tumor sample pre, post-treatment and different type of treatments and other kind of factors will potentially impact the protease activities?

Thank you for the question. Yes, our library of tumor samples is based on hundreds of tumor samples in multiple indications. We do have metastatic samples. It is a mixture of patients that have received prior therapies. We have not, at this point, categorized the protease environment by previous lines of therapy or previous modalities of therapy to see how that influences the tumor protease environment, is something we could look into.

Okay. Sure. That's great. And another question for Dr. Tolaney. So you mentioned, given the heavily pretreated patients in Phase I, the durable clinical benefit rate may be more important than the resist response. Before Phase II, what would you consider clinically meaningful efficacy for each cohort, given those patients are less heavily pretreated?

No, I think it's a very good question. For the hormone receptor-positive population, this is a population that allows up to 2 prior lines of cytotoxic therapy. So in fact, they will be pretty pretreated in a sense that most people will have had mandatory prior CDK46. So many of these patients may have had several lines of endocrine therapy and then could have gone on to 1 or 2 prior chemos even before going on to study. And so in that population, if you look at prior registration trials that went into that setting, in essence, what we're seeing from chemotherapy in that line of treatment is usually a response rate of around 10% or 11%. Certainly, you could look at data from eribulin which, I think, is a fairly reasonable comparator in a sense that it would be in a more pretreated population that led to registration where, again, that was the response rate. PFS generally has been in the 3-4 month range in that line of treatment, and so I think seeing something beyond that would be very encouraging. In the triple-negative population, unfortunately, outcomes for this population are quite poor outside of the first-line setting. I think certainly, the data from this ASCENT trial and the control arm does demonstrate that where you see PFS is really, in essence, around 2 months to most standard chemotherapy. And so I think, again, seeing something beyond that is going to be encouraging to -- I think that this has a potential path forward. Obviously, we have sacituzumab response rate of around 30% in heavily pretreated patients, so I think seeing something in the 20% to 30% response range is going to be a good signal.

Your next question comes from Anupam Rama of JPMorgan.

Thank for hosting this. I've got a quick one on [indiscernible]. Can you remind us -- for the company, can you remind us what cohorts we should be focused on for [indiscernible]?

Anupam, this is Sean. We had a lot of feedback on that question, but I noticed that you actually -- I think you asked at least half of it in the chat as well, so I'm going to answer what -- I think one of the questions was, what can we expect data for by the end of this year for the 2029 program. And so as you heard, 4 cohorts being enrolled, our guidance is initial data from the first 2 cohorts, so from the head and neck and the squamous non-small cell lung cohorts by the end of this year. The esophageal and DLBCL likely in 2022. I didn't get the rest of the question, unfortunately.

Yes. I have a follow-up on that one for the KOL for the cohorts that we're going to get [indiscernible] the win scenario for them.

Oh, I see. The question, I think, what the 6 months looks like, yes.

What would the win scenario look like for -- maybe we'll start with the squamous lung cancer cohort. I think just as Dr. Tolaney said, I think in the second and third line chemotherapy comparator would be docetaxel with a 3-month PFS and maybe a little bit better in the REVEL trial with ramucirumab. So anything beyond 3 months, 4, 5 months, 6 months would be a win. I think the analogous is true for any squamous head and neck cancer. The esophageal cohort is a little bit more very squamous adeno. And then the lymphoma, all bets are off there because those cancers respond a lot easier. But I think its durability, that is challenged in those patients. So I think -- I guess, to sum it all up, because these are different tumor types, I would say that we're looking at refractory patients who progressed on at least one hefty standard of care, and so it's on the order of 3 to 6 months.

Our next question comes from Terence Flynn with Goldman Sachs.

Great. Maybe one follow-up for Dr. Johnson. You touched on this a little bit in your remarks, but just can you expand in terms of the management of the anemia, just how easy is that? I know you have to collect some additional lab data from patients. And again, do you expect these strategies to be effective at mitigating that anemia? And then on the -- Sean, on the cytokine platform that you guys have rolled out, can you give us any more precise timing of when you might be able to move an asset into the clinic? Is this 12 months out or is this beyond that? And when might we see some preliminary data there?

Yes. Sean, I'll go first, then you can take the other question if you want.

Yes, please. Please, go ahead, Melissa.

Anytime that we have a cancer patient in the clinic with anemia, we start with the same work up. We will check their iron stores. We'll check their erythropoietin level. We'll check their B12, folate. We'll check TSH. And so that doesn't change here. If we find that the patient has some degree of iron deficiency, and many do because they're not eating, because they don't feel well, that's easy to correct with both oral and even IV iron in the clinic. If the EPO level and the hemoglobin allow, then erythropoietin-stimulating agents are propria. And in many cases, for these patients, they are, especially given the reticulocyte count being low. So it's an easy slide to do it, just oncologists -- at least oncologists that are specialists in a particular field are not always skilled hematologists. So it's teaching the oncologist in the clinical trial how to work up that anemia. We all know how to write for blood transfusion. It's certainly what I did early and often for my patients, but that alone, I don't think, is going to solve this problem. So I think it's solvable. I do. I think we need to pay more attention to facing all of the mechanisms of anemia. And with more knowledge, perhaps we'll be able to come up with the right technology -- or the right sequence to which to do these different tricks.

Great. Thank you, Melissa. And Terence, thanks for the question on the cytokines. Really too early to guide on timing at this point, but as you can tell, we're excited to be moving into this really new area of research for the company. I would track back to the presentation that Marcia gave at the beginning of this session, talking about the depth of understanding that we continue to develop in fundamental protease biology, in the design of protease cleavable linkers, obviously, new marketing strategies. And taking this all together, we plan to build a broad program in the cytokine space. So right now, we're really focused on breadth. Timing will be communicated at a later date.

Your next question comes from the line of Joseph Catanzaro with Piper Sandler.

Thanks for hosting this. Maybe one first for Dr. Tolaney. I was wondering with regards to CX-2009's ocular toxicity, if you've spoken to maybe some of your colleagues at the women's cancer center there about their experience with mirvetuximab in ovarian cancer, if they have any, and what they've told you about the manageability of ocular toxicity and being able to keep patients on drug once prophylaxis is implemented. And I have a follow-up.

And it's an excellent question, and certainly, I think we can learn a lot from the mirvetuximab experience here. And so this group did have a lot of consultation with Ursula Matulonis, who has been very involved in mirvetuximab development as well as other consultants. [indiscernible] has also been very involved in providing guidance here. And so I think again, utilizing that experience is key, and that is actually what led to these guidelines for using the combination of with [indiscernible] prophylaxis with steroids, vasoconstriction, cold compresses, rewetting drops, that really has been due to the mirvetuximab experience, and with the ocular consultant, who has been involved in mirvetuximab development as well. So again, yes, I think, very important to take those lessons learned moving forward with this.

Okay. And if I could just ask a quick follow-up, maybe one for the company. So as I look at the CX-2029 sort of baseline data, I'm reminded that the CD71 expression was maybe not uniformly high across the full Phase I cohort. So just wondering how CD71 expression levels break out specifically for squamous lung and the other Phase II tumor types and whether it makes sense at all to screen for CD71 expression in any of those cohorts.

Yes. Joe, maybe I can pass that question on to Amy to talk just briefly about some of the ongoing strategies or thinking. It's still relatively early days, but it's a good question. Amy, I think you're on mute. Let's give it one second. Let me continue to chip in and certainly ask Dr. Johnson to comment as well. I would just say, Joe, that we're still in the early stages of characterizing relationship between target and response. And as we've said before, it's such an interesting target because CD71 transports so rapidly off the cell surface. It may or may not have the similar types of relationships, but it's something that we're obviously going to try and pin down. And we'll learn a lot more from the ongoing expansions being run over the course of this year and into next year.

Yes. There's Amy. Maybe she's -- I don't know if she is audible now. No. I would just add that it was a question we asked earlier on as we looked at the data from Phase I. And unfortunately, it isn't as easy as the responders had high levels of CD71 expression across the board and the progressors have little. But it does make sense to me that it wouldn't be that easy because there is such rapid turnover of the transmembrane receptor. So I think it is a little bit more difficult to measure than EGFR, for example.

I note that we are at the top of the hour, so -- and I know that we have -- our guest speakers have hard stops, so I think what we'll do is we'll move to wrap up here. Thank you, everybody, for -- well, first of all, thank you to our distinguished speakers for their time today and for giving such terrific presentations on our science, on our clinical work, on the -- what we're trying to do for patients with our platform, keeping the patient front and center in all of our work. I do hope that for all of you who have participated today in this, that this detailed review has been useful, both the technology and the clinical programs. I hope it's also clear that CytomX continues to -- really, to innovate the core of our company, and we continue to blaze this trail of conditional activation using our Probody platform, so like I said, with an unwavering focus on the patient. Thank you all for your time, and we look forward to following up with you all in the coming weeks and months. And thank you again to Doctors Johnson, Tolaney and Lambert. Very much appreciate it.