Morning. Thank you for joining us today for the third session of our Explorer Series featuring our lead myeloid reprogramming portfolio program, NGM-707. My name is Brian Schoelkopf, Head of Investor Relations at NGM Biopharmaceuticals. We will be making forward-looking statements during today's presentation, including statements about anticipated timing of events and potential benefit of our product candidates. We refer you to our most recent 10-K, which identifies factors that could cause actual results to differ materially from today's forward-looking statements. I encourage you to read the full language on this slide and all of the slides that will be presented today. It will be posted to the Investors and Media section of our website. With that, I'll hand it over to David Woodhouse, PhD, and Chief Executive Officer at NGM Biopharmaceuticals. David? Thank you, Brian. I'm pleased to welcome you today to the third installment in our multi-episode R&D Day we call our Explorer Series. As a reminder, in the first segment of this series, we shared with you details behind our discovery engine that has produced our entire pipeline from scratch and has a track record of producing approximately one IND every year. In the second episode, we gave you an introduction to two of our myeloid reprogramming portfolio programs, NGM-831 and NGM-438, that we think can demonstrate the power of overcoming stromal checkpoints that underpin immune resistance in certain solid tumors. If you didn't have a chance to catch those earlier segments, both parts are available on our website. Given the depth and breadth of our pipeline, we hope these R&D Day modules serve as a handy resource to get up to speed on the company. Now, as we look forward on our roadmap for the Explorer Series, what we have in store for you in the next two segments are sessions focusing on two programs in our pipeline that generate quite a bit of interest from industry and investors. Today, in session three, we'll be focusing on our lead product candidate in our myeloid checkpoint inhibitor portfolio, NGM-707. It targets inhibition of both ILT2 and ILT4 receptors on immune cells that we believe may help patients mount stronger immune responses to solid tumors. Our final session, segment four, will focus on NGM-621, our anti-complement C3 antibody that we think can be a category-leading treatment for geographic atrophy, a disease that threatens the vision of millions of people worldwide. Before I walk you through the specific items in the agenda for today's session on NGM-707, I'd like to share with you a bit of background on the program and a few themes I hope you take away from the presentations today. Seven years ago, we made a strategic decision to direct considerable focus of our discovery engine towards mechanisms of tumor resistance that drive poor prognosis for many patients. Despite important advancements over the last decade, cancer continues to be the second leading cause of death in the U.S. and touches the lives of nearly everyone, either directly or through close friends and family. That is certainly true at NGM, and for many of us, our motivation to advance programs like NGM-707 is driven by both scientific passion and a personal commitment to try to develop better treatments to help patients fight cancer. I lost both my parents to cancer while I was still in graduate school over 20 years ago and consider it a tremendous privilege to be able to work with our very capable scientists and clinicians to try and make a difference against this disease. As you'll hear today, NGM-707 is our lead myeloid checkpoint inhibitor program that stems from some important insights we developed by studying the biology that underlies tumor resistance to immune detection and attack. While T-cell checkpoint inhibitors like KEYTRUDA and Opdivo have been quite effective for some patients with cancer, most patients do not respond or they only respond transiently. We focus on what makes some patients' tumors resistant to T-cell checkpoint inhibitors and what targets we could identify to overcome that resistance. Basically trying to identify ways we could turn cold tumors into hot tumors that can be more readily detected and eliminated by the immune system. Our work honed in on fairly abundant immune cells in the tumor microenvironment from the myeloid lineage, which is macrophages and dendritic cells, to help direct and control T-cells and other effector cells from the lymphoid lineage. Unfortunately, in the instance of treatment-resistant tumors, these myeloid cells are promoting an immunosuppressing environment that protects tumors, even if anti-PD-1 antibodies have released the T-cell checkpoint that underlies another form of tumor resistance to immune detection. Essentially, the immune system has a T-cell checkpoint and a myeloid cell checkpoint that tumors can use to evade attack, and we believe by releasing both, we can help drive a more effective immune response to fight tumors. As you'll hear today, the story behind NGM-707 illustrates the power of our discovery engine by drawing from both our expertise in biology and protein engineering to create what we think will be an optimal myeloid checkpoint inhibitor. Our biology team, in hunting for promising myeloid checkpoint targets, intersected with important work done by Dr. Marco Colonna, who you'll hear from in a moment, as well as other myeloid biology pioneers, identifying a collection of related receptors that appear to play an important role in creating a myeloid cell checkpoint around certain solid tumors. Working in concert with our protein engineers, our team then created a broad set of tools to interrogate the biology behind these receptors. This toolbox included both monospecific antibodies to individual receptors in this family, as well as dual antagonist antibodies. NGM-707 represents our lead candidate from these efforts, with the important distinction that it is a dual antagonist antibody designed to block two types of receptors with a shared epitope to have the powerful effect of releasing the myeloid checkpoint and reprogramming tumor-associated myeloid cells from suppressing to stimulating immune attack. We have quite a bit of data to share with you today that illustrates the promise of this approach. Importantly, we advanced NGM-707 into the clinic last year and have recently progressed to the phase 1b portion of dose escalation in combination with pembrolizumab or Keytruda while we finish completing the final dose of the phase 1a monotherapy dose escalation portion of the trial. We plan to share an update on the phase 1a portion of the trial at a medical meeting in the second half of the year. With updates from the phase 1b to follow in 2023. Now with that brief overview, let's turn to how we have organized the time today. First, you'll hear from Marco Colonna on the groundbreaking research that his team has produced that has led to important insights about tumor-associated myeloid cell biology and opportunities to use those insights for drug development. Then our lead scientist for NGM-707, Dr. Jeff H. Sode, will walk you through some of the pre-clinical in vitro and in vivo research the team here at NGM has done over the years that has built our conviction in this program. Then finally, Dr. Dan Kaplan, our head of Translational Immune Oncology, will spend some time discussing our clinical development strategy for NGM-707 and share some promising kinetic data from our phase one trial seen here for the first time today. Now, it's my great pleasure to introduce Dr. Marco Colonna, the Robert Rock Belliveau, MD, Professor of Pathology and Immunology at Washington University, to speak today about the ILT family of receptors. Marco is a true pioneer of the ILT receptor field. 25 years ago, Marco first discovered and then cloned these receptors. He subsequently demonstrated that ILT-2 and ILT-4 are receptors for MHC class I family ligands. He and his team have, in the ensuing years, elucidated key immune regulatory biologies mediated by the ILT family. The ILT receptors are increasingly recognized as important therapeutic targets in oncology. We still have Marco to thank for bringing these receptors to light and enabling the development of potential new therapies for cancer patients. Marco. Hi, I'm Marco Colonna, Professor of Pathology and Immunology at Washington University School of Medicine in St. Louis. Today, I'd like to give you some insight into the biology of the immunoglobulin-like transcript or leukocyte Ig-like receptors and try to convince you that this important target for tumor immunotherapy. The impulse for the discovery of this receptor came from our interest in natural killer cells and their ability to kill many target cells, both tumor cells and virus-infected cells. We knew that NK cells do not express antigen-specific receptors, and yet they can recognize many different target cells, many different tumor cells, many different virus-infected cells. How is that possible? In the late eighties, a scientist from the Karolinska Institutet, Klas Kärre, came up with an interesting idea. At that time, at the height of the Cold War, the Swedish Coast Guard was constantly alerted by the local fishermen that submarines were present in the Swedish waters. Yet they could not distinguish between the Swedish submarines, the Russian submarines, or perhaps submarines from other countries. The Swedish Coast Guard distributed to the fishermen the silhouette of the only submarine that was produced by the Swedish army and told the fishermen to call them only if the submarine that they saw did not fit that particular silhouette, did not match with the Swedish submarine. Klas Kärre came up with this idea. He thought that perhaps NK cells work in a similar fashion. They recognize not different target cells, but only one target cell that expresses a particular silhouette molecule. For example, MHC class I, which is expressed by all normal cells in the body, and only when MHC class I is not expressed by tumor cells, by virus-infected cells. At this point, NK cell is alerted against these target cells. cells. We know that MHC class I is indeed down-regulated by tumor cells and also by virus-infected cells as an attempt to escape these T cell responses. We were able to use this idea, and we tried to identify the inhibitor receptor that would recognize MHC class I on NK cells and block NK cell alerting. In 1995, we finally identified a family of receptors with a killer cell Ig-like receptor or KIRs that recognize HLA class I. This family is a family of diverse receptors that recognize different MHC class I molecules, HLA-C, HLA-B, HLA-A proteins. We noticed that in the cytoplasmic domain of this receptor, we found the presence of two motifs which are now known as inhibitory receptor motifs or ITIMs, which carry protein tyrosine phosphatase SHP-1 and SHP-2, which are responsible for inhibition. We mapped these genes to a chromosomal region on human chromosome 19, which we now call the LRC, the leukocyte receptor complex. To our surprise, we found that this region contains not only the KIRs, but also contains additional receptors, highly homologous to KIRs that we cloned almost at the same time by homology. These receptors are now called immunoglobulin-like transcripts or ILT and also, leukocyte Ig-like receptors. They come in two different clusters, a tetrameric cluster and a trimeric cluster, separated by a pair of receptor chimeras. Homologs of these receptors are also found in mouse. However, one should note that in mouse, the equivalent of ILTs are separated in two different chromosomal locations, probably reflecting a separation of chromosomes that has occurred, a long time ago in the evolution. We identified a number of inhibitory isoforms. Five different inhibitory isoforms. Whereas in mouse we only identified two inhibitory isoforms. All these isoforms contain tyrosine-based inhibitory motifs that approach SHP-1 and SHP-2. We also identified activating isoforms. Precisely, we identified five different activating isoforms in humans and five different activating isoforms in mouse. Now, these isoforms do not contain cytoplasmic ITAM, but they do pair with, a transmembrane factor called the Fc receptor gamma chain, which contain cytoplasmic motifs known as immunoreceptor tyrosine-based activation motifs or ITAM. These motifs are responsible for the recruitment of protein tyrosine kinase C that is induced in cell activation. Today, in the interest of time, I will focus my presentation on inhibitory receptors, and particularly, I will focus on the first two receptors that we cloned and characterized, ILT2, which is also known as LILRB1, and ILT4, also known as LILRB2, both of which bind HLA class I. Let's start with ILT2 LILRB1. We found that ILT2 was expressed in T cell-like precursors but was also expressed in T cells, particularly on a subset of CD8 T cells. It was expressed in all peripheral B cells, in monocytes, and also most hematopoietic stem cells. The ILT2 has a much broader expression pattern compared to KIR. We define the specificity of ILT2 for MHC class I using different assays. Here I show you one assay in which it took the extracellular domain of ILT2, used the Fc portion of the immunoglobulin, and test the binding of this soluble ILT2 to cells that do not express MHC class I or cells that have been transfected with different MHC class I molecules. What you can see is that ILT2 Fc binds to a number of different class I molecules, HLA-A3, HLA-B27, and also HLA-B. Suggesting that ILT2 has a broad specificity for MHC class I molecule. Now, the crystal structure of ILT2 together with MHC class I was solved by Kannan Natarajan. What she found was the identification of two sites with interaction between ILT2 and the HLA-A2. One site was binding beta-2 microglobulin, and the other site was binding the alpha-3 domain. This is interesting because beta-2 microglobulin is identical in all MHC class I molecules. In alpha-3 domain is highly conserved in MHC class I molecules. That explains why ILT2 can bind to MHC class I molecules with a very broad specificity. We also knew that other receptors can bind to MHC class I, for example, homodimers of CD8α. Therefore, we're interested to see whether the binding of CD8α MHC class I and the binding of the ILT2 MHC class I has some partial or complete overlap. What you can see on the left here are the contact points of CD8α homodimers with MHC class I. What you can see on the right are the contact points between ILT2 and MHC class I. As you can see in the center, there's some partial overlap and some partial and steric clashes such that the CD8 alpha and the ILT2 will somehow impede reciprocally in binding to MHC class I. Now, we functionally characterized the ILT2 in NK cells. As I mentioned, inhibitory receptors block cytotoxicity of NK cells, and we demonstrated that indeed engagement of ILT2 reduces the NK cell cytotoxicity better than I'm now going to show you here, b ut we did also demonstrate that ILT2 can block CD8 T cell cytotoxicity. We isolated CD8 T cells that express ILT2, a subsurface unique subset of cells in peripheral blood. Then we tested their ability to kill a target cell 721.221 that express a MHC class II bound to superantigen, SST. We found that these T cells recognize superantigen kill the target cells very effectively. However, when we transfected the target cells with MHC class I, such that MHC class I could then engage ILT2, then we saw clear inhibition of CD8 T cell cytotoxicity. Finally, when we repeated this assay in the presence of an anti-ILT2 antibody that can block the interaction between ILT2 and MHC class I, we reconstituted cytotoxicity of CD8 T cells. We also did a lot of initial studies on ILT4, known as LILRB2. ILT4 is different from ILT2 because it has a more restricted tissue expression. It's mostly expressed on myeloid cells, including, for example, monocytes, tolerogenic dendritic cells, tissue dendritic cells, tissue macrophages. Whereas it's not expressed in B cells or T cells. We demonstrate ILT4, like ILT2 binds, MHC class I molecule with the close specificity. We use the same approach that I described before. We use an ILT4 Fc fusion protein to show the binding called ILT4 extracellular domain to cell expression in MHC class I. We also use an interesting complementary approach in which we generate HLA class I tetramers, and we tested their binding on cells that are like monocyte express ILT4. As you can see here, we saw a very nice binding of MHC class I tetramers to ILT4. The crystal structure of ILT4 with MHC class I, specifically HLA-G, was also solved. Once again, we identified two sites of interaction of ILT4, one with the β-sheet lobe and one with the extracellular domain, again, consistent with the close specificity of ILT4. Again, when we compare the CD8 alpha complex with MHC class I with the ILT4 complex with MHC class I, we found significant steric clashes between the two, suggesting that there's an antagonism between CD8 and ILT4 for binding MHC class I. This will be important in the rest of the talk. As I mentioned, we were not surprised to see that inhibitory receptors inhibit NK cell cytotoxicity, and we also expected inhibitory receptors to inhibit CD8 T cell cytotoxicity because of our work on KIRs. We didn't know what is the impact of ILTs on myeloid cells, and specifically on myeloid cells, like dendritic cells, that are involved in antigen presentation. Early on, we knew that dendritic cells in humans are heterogeneous, and we distinguished two very broad subset of dendritic cells. The Langerhans cell dendritic cells, which are important in priming CD8 T cells and CD4 T cells, and also the interstitial dendritic cells, which are important in priming follicular helper T cells and also initiating B cell responses. The Langerhans cells are important for cellular immunity, whereas interstitial DCs are important for humoral immunity. These cells can be isolated directly from tissue, for example, from the skin. Langerhans cells are present in the epidermis, whereas interstitial dendritic cells are found in dermis. The cells can also be obtained in vitro by culturing CD34 positive hematopoietic stem cells with IL alpha and GM-CSF. In both approaches, we'll end up with two subsets of dendritic cells. One expressing CD1A, which corresponds to Langerhans cells, and one expressing CD8, which corresponds to interstitial dendritic cells. To test the function of this subset of dendritic cells, we co-cultured these dendritic cells with allogeneic CD8 T cells to induce the expansion of NK cells or proliferation. We looked at their effector function, particularly cytotoxicity and interferon production after 7 days of reactivation. What we noticed is that Langerhans cells were very effective in inducing a population of CD8 T cells, which we call TCY, that had high-level expression of CDA and producing interferon gamma. On the other hand, dermal interstitial dendritic cells were effective in inducing a population of CD8 T cells that express low level of CDA and produce IL-13 and IL-5. We also performed a whole genome RNA profile of this subset of our dendritic cells, and we noticed that interstitial DCs, but not Langerhans cells, express some of the ILT receptors, specifically ILT2, ILT4, and ILT5. We asked the question whether the expression of ILT in this particular subset of cells could be important to modulate or to direct the type of T cells that were induced by these cells. To address this question, we took dermal dendritic cells express CD14 and cultured them with naive alloreactive allogeneic T cells. We used antibodies that would block ILTs. When we did it, instead of having the induction of CD8 T cells that express IL-13, IL-5, and IL-4, we obtained cells that expressed high levels of CD8 and producing interferon gamma. The blockade of ILT was effective in inducing highly cytotoxic interferon gamma-producing CD8 T cells. We also did the reciprocal experiment. We knew that Langerhans cells that do not express ILTs are very effective in inducing TC1, therefore using interferon gamma. We performed the same experiment in the presence of CD8 competitors. For example, an Fc fusion protein of ILT and CD8α. In this case, we reverse the ability of Langerhans cells to induce PC1 rather than induce PC2 cells that have lower expression of CD8, IL-13, and IL-5. Overall, this experiment told us that DC comprise different subsets with different immunological function. Langerhans cells are in PC1. Tumor interstitial DC are in PC2. This difference, this functional difference between the two subsets depends on the expression of ILT receptors on normal DC, which interfere with engagement of CD8. Of course, this experiment suggests that ILTs can be harnessed to modulate allogeneic transplantation DC. I'd like to spend a few words on ILT3. A little before, we identified ILT3 early on. We found that like ILT4 was mostly expressed in myeloid-related cells, monocytes in blood, DC in the blood, also monocyte IDC and CD52 positive monocytes. However, we were not able to find ligand for this molecule. We tried extensively to see whether we could find any ligand at all, but we could not find any positive answer. Recently, NGM Bio scientists have identified a ligand of ILT3 in fibronectin. Another molecule that can act as a surrogate for cell molecules because fibronectin, of course, is expressed in all normal tissue. Work by NGM Bio has shown that a blockade of ILT3 fibronectin interaction with specific monoclonal antibody can also modulate the function of DC cells. For example, DC cells cultured in the presence of fibronectin have a tolerogenic function, and we can see in the presence of fibronectin, reducing the proliferation of DC cells, the production of IL-3, the production of IL-4. However, in the presence of a blocking antibody that interrupts the interaction between ILT3 and fibronectin, we have a significant increase of the immunostimulatory function of DC, leading to a reprogramming of tolerogenic DC. Now, we think that this data really suggests that ILTs can be important target for therapeutic, particularly in tumor models. O f course, the question is can we test ILT therapeutics in mouse models? As I mentioned my introduction, the configurations of a gene transfers for ILT in mouse and humans are quite different. For example, when we look at the LILRB4 isoforms, we see five different LILRB4 isoforms in human. Whereas in mouse we see two different isoforms in mouse. In addition, the isoforms in human have four C-terminal immunoglobulin domains, whereas the mouse ILT has six C-terminal immunoglobulin domains, suggesting they may have also different specificities. To address this question, we decided to generate transgenic mice that express human ILT. We took a bacterial artificial chromosome covering the centromeric cluster of ILT, which you can see here, and we generated transgenic mice that we call ILT-centric, which express ILT-5, ILT4, LILRB1, among others. We generated a second mouse using a BAC, a bacterial artificial chromosome that encompass the telomeric cluster of ILT. We call these mice ILT telo. They express ILT2, ILT1, ILT3, among others. We've looked at the expression of this human molecule in mouse tissues. Here, for example, we look at expression in blood or in mouse tissue like the brain. We find that ILT1 and ILT3 are expressed in myeloid cells and also in macrophages like microglia. We found a low level of expression of ILT2 in T cells and B cells. We note a different level of expression. This is our observation, and this also is the case in human. In ILT-centric mice, we see a clear expression of ILT4, ILT5, ILT7, LILRB1 in myeloid cells. We see expression of ILT5 and LILRB1 in T cells consistent with the pattern of expression that we previously observed in human. We have a model in which we can recapitulate the expression of ILT receptor in mice. In addition, we have now crossed the ILT-centric with ILT telo mice to reconstitute in one single mouse the entire human ILT. Of course, we'll be able to use these mice to test various antibodies against ILT and to see their impact on tumor rejection or other functional assays. We think that ILTs have a bright future in immunotherapy. I've shown you that blockade of ILT2 can activate the cell and NK cell responses to tumor cells. I showed you that blockade of ILT2, ILT3, and ILT4 can remodel dendritic cells to prepare antigen-presenting cells to generate robust CD8 T cell responses. I showed you that blockade of ILT2, ILT3, and ILT4, they also have impact on tumor macrophages. Certainly, it will be important to test the effect of this blockade in tumor associated macrophages to see whether blockade of this receptor reduces immune suppressive function. We are showing you that strategies to block immune regulatory for enhanced immune rejection can now be tested directly using transgenic mice. I'd like to close with a few acknowledgments. I'd like to acknowledge people in my lab at Washington University, particularly Maria Scheller, Kimberly McPherson, Marina Cella, and Gaya Santacroce. I'd like to acknowledge scientists at NGM Bio, particularly Arthur Hsu, Daniel Kaplan, Julia Roga, Jeffrey Stone, and John C. Siegfried. We had a fantastic collaboration, and it will be very productive. I'm sure that this collaboration will advance our understanding of ILT biology, and most importantly, will generate new therapies for tumor immunotherapy. Thank you for your attention. Thank you, Marco. Hello, my name is Jeff Stone, and I'm the lead scientist for NGM-707 here at NGM Bio. During my presentation today, I'll further discuss the exciting biology of ILT2 and ILT4 and share preclinical work we have done demonstrating the powerful immune activity that we have seen unleashed in preclinical models when ILT2 and ILT4 are blocked by NGM-707. ILT2 and ILT4 are closely related immunosuppressive receptors. They're both expressed on myeloid cells in the tumor microenvironment. ILT2 is also expressed on NK cells, B cells, and T cells. As you just heard from Marco, these receptors are part of an important pathway driving suppression of myeloid cells, NK cells, and T cells. ILT2 and ILT4 together can suppress all of these immune cells. There's also redundancy between these two receptors, suggesting that inhibiting just one of the receptors may be insufficient to inhibit the pathway, and therefore, dual inhibition may be important. In the coming slides, we'll look at preclinical evidence of the described overlapping biology of ILT2 and ILT4 in human tumors. We'll then look at the complementary activity of ILT2 and ILT4 blockade in vitro, and we'll finish by reviewing the effect of the combination of ILT2 and ILT4 blockade in vivo. Human tumor data provide us with initial evidence that ILT2 and ILT4 could be key contributors to poor outcomes in cancer. Shown in the violin plots on the left, ILT2 and ILT4 are upregulated in melanoma patients who do not respond to T cell checkpoint therapy, providing a compelling hypothesis that ILT2 and ILT4 act as T cell checkpoint inhibitor resistance mechanisms. Patients who responded to T cell checkpoint inhibitors had lower expression of ILT2 and ILT4, shown in the blue plots, while patients who failed to respond to checkpoint inhibitors, shown in the red plots, had higher expression of ILT2 and ILT4. In addition, high ILT2 and ILT4 expression in the tumor is associated with poor survival and faster recurrence. This is shown for ILT2 in gastric cancer in the middle panel and for ILT4 in colorectal cancer in the right panel. Patients with high levels of expression of these receptors, the red lines in the middle charts and the blue lines in the charts on the right-hand side, show worse survival and higher recurrence in patients with lower ILT2 or ILT4 expression. These are just a few examples of the numerous tumor types where ILT2 and ILT4 are associated with poor cancer patient outcomes. On the next slide, we'll review gene expression data that can also tell us a lot about the biology of ILT2 and ILT4. Here we are looking at single-cell RNA-seq data showing the expression of ILT4 in breast cancer. ILT4 expressing cells are shown as green dots. ILT4 is expressed specifically on tumor-infiltrating myeloid cells, with some myeloid cells showing high ILT4 expression and other myeloid cells showing moderate expression. In this next chart, we are looking at ILT2 expression in the same patients. As you can see, ILT2 is expressed not only in the myeloid T cluster, but also in the tumor-infiltrating B cell cluster and in a cluster including NK cells and CD8-positive T cells. This insight suggests that ILT2 may be involved in a broad range of immune suppression. Finally, here we see the expression of both ILT2 and ILT4. Notice how many of the myeloid cell dots are now yellow, denoting dual expression of ILT2 and ILT4. There are also cells expressing only ILT2 or only ILT4. There are two important takeaways from this. First, there is a high level of redundancy of ILT2 and ILT4 on tumor-infiltrating myeloid cells. This is very important because it suggests that inhibition of both pathways should be required to have an effect on these cells. Second, broad expression of ILT2 and ILT4 on many different cell types suggests there are multiple components of the innate and adaptive immune response that may be activated with dual blockade of ILT2 and ILT4. Through the tireless effort of our biologics team, NGM was able to find a unique antibody that can block both ILT2 and ILT4 through a shared epitope. NGM-707 is a first-in-class antibody targeting both ILT2 and ILT4 interaction with MHC class I, including both classical HLA-A and HLA-B, as well as the non-classical molecule HLA-G. On the charts on the right side of the slide, NGM-707 in green shows robust inhibition of the interaction of ILT2 with MHC class I on the left, as well as robust inhibition of ILT4 interaction with MHC class I on the right. This suggests NGM-707 can capture the biology of both ILT2 and ILT4 blockade. As I said earlier, NGM-707 has the potential to impact a broad array of immune cells. What you see here is a list of immune cells that can be impacted by NGM-707 and the range of immune mechanisms that are affected. Activation of CD8-positive T cells will have a potent tumor-killing effect, both directly and through increased expression of cytokines by monocytes and macrophages. Myeloid activation and antigen presentation may also amplify the effect of PD-1 therapies by switching myeloid cells from an immune-suppressive state to an immune-stimulatory state. Non-T cell activation of NK cells, stimulation of phagocytosis, and increased antigen presentation will provide other mechanisms of immune response against the tumor. The power of this biology is that unlike PD-1 therapies, we are activating multiple components of the immune response. Based on this broad activity, we really believe that NGM-707 can dramatically heat up the immune response against cancer. Initially, we looked at the effect of ILT2 blockade on tumor cell killing by NGM-707. We saw increased killing by ILT2 blockade shown in red, but not ILT4 blockade shown in blue, which makes sense since ILT4 is not expressed in these cells. NGM-707, shown in green, was highly active, which was really encouraging because it confirmed that NGM-707 is effectively inhibiting ILT2 and stimulating T cells in line with what we saw with anti-ILT2. The case of killing of tumor cells was also enhanced by ILT2 blockade, shown in the red line and red bar, suggesting ILT2 blockade can directly impact the case of killing in tumor cells. ILT4 blockade in blue has no activity, as we would anticipate based on the lack of ILT4 expression on NK cells. Again, the important takeaway is how NGM-707, shown in green, activated NK cell-mediated killing to the same degree as ILT2 inhibition. Many myeloid cells express both ILT2 and ILT4, so we were surprised to see a distinct biology of ILT2 and ILT4 in some assays. For example, here is a myeloid reprogramming assay. When we mixed myeloid cells in culture with T cells, there was very little cytokine secretion. When we add NGM-707, the cells respond with increased TNF-alpha. ILT4 blockade enhanced this secretion while ILT2 had no effect, despite high expression of both ILT2 and ILT4 on these cells. Importantly, NGM-707 was able to exceed the response by anti-ILT4 in this model. This highlights the importance of hitting both receptors. Because of their redundant nature, which receptor is driving immune activation is not always apparent, and so inhibiting both of these redundant receptors is critical. While ILT4 blockade drove response in the MDSC assay, we were surprised to see this to be opposite in the macrophage phagocytosis assay. Here, ILT2 blockade, shown in red, induced a potent increase in phagocytosis, while ILT4 blockade, shown in the blue triangle, gave no response. NGM-707 in green matched the activity of anti-ILT2, nicely capturing this biology as well. Again, this analysis underscores the difficulty of predicting which of these two receptors drives immune suppression and therefore the value of inhibiting both. Taking the results of these analyses of NGM-707's impact on T cells, NK cells, and myeloid cells together, the conclusion is that while anti-ILT2 was only active on lymphocytes and phagocytosis and anti-ILT4 was only active on MDSC, NGM-707 has the power to increase the activation of all of these immune cells. A number of other assays, such as this assay that measures myeloid cell activation, showed IL-4-specific activity in a standard culture. What really got us excited was the additive effect of IL-2 and IL-4 in the presence of MHC class I. In this experiment, we measured inhibition of IL-10, a cytokine that drives immune suppression. Unlike the prior experiments we've looked at, a lower level of this cytokine is what you would hope to see because it indicates reduction of immune suppression. When we block IL-2 or IL-4 alone, we see a modest response. When we use NGM-707, shown in green, we see a much stronger response that matches the combination of anti-IL-2 and anti-IL-4, shown in brown. NGM-707 is capturing the combined biology of IL-4 and IL-2 blockade. This experiment and similar experiments are teaching us that both ILT2 blockade and ILT4 blockade are required to fully unleash myeloid cell activation, highlighting the need for a dual antagonist such as NGM-707. We also see an additive effect with NGM-707 when we activate a key stimulator of dendritic cells, the Fc receptor. NGM-707 in green shows significantly higher TNF-alpha secretion levels compared to anti-ILT2 in red and anti-ILT4 in blue. This stimulation of dendritic cells is another important aspect of the immune response that is stimulated with NGM-707 with a combined blockade of ILT2 and ILT4. Overall, we've shown you that NGM-707 activates a broad range of cells, including dendritic cells, macrophages, MDSC, as well as NK cells and T cells in preclinical models. NGM-707 can capture powerful and complementary biology on all these cell types. Next, we want to ask how powerful the biology of NGM-707 is in the tumor microenvironment, where fibroblast stromal cells are known to suppress the immune response. We use cancer-associated fibroblast cells as a physiologically relevant model to evaluate suppression in the tumor microenvironment and as a bridge to in vivo tumor studies. We cultured fibroblasts with dendritic cells and looked at immune suppression. As shown in the graph on the right, dendritic cells by themselves, the first bar, show an immune activating phenotype with high levels of CD86, a marker of immune activation, and low levels of CD163, a marker of immune suppression. In the middle bar, fibroblast cells dramatically reduce this immune activating phenotype, consistent with the fibroblasts making dendritic cells immune suppressive. What was exciting was when we added NGM-707, shown on the right bar. Here, the dendritic cells returned back to their immune activating phenotype. I really like the plot at the bottom right, which shows the actual data. Notice how fibroblasts make the dendritic cells reduce CD86 and increase CD163. Adding NGM-707 made the dendritic cells return to the bottom right box. They're just as activated as they were in the absence of fibroblasts. This is powerful biology. NGM-707 made dendritic cells act as if the suppressive tumor microenvironment was not even there. Next, I'll show you how NGM-707's ability to reprogram myeloid cells from an immune suppressive state to an immune stimulatory state while also directly activating lymphocyte cells can impact the growth rate of a living tumor. ILT2 and ILT4 are primate-specific receptors not present in the mouse, which makes them very challenging to model in vivo. This is the reason there's not a lot of in vivo analysis of this mechanism currently available. To create a viable in vivo model, we needed to use a humanized mouse system to mimic the interaction of human tumors with human immune cells. Even more than that, we needed a model that includes more human myeloid cells than what is currently available in most humanized mouse models. We are very excited to share that we were able to capture NGM-707's powerful myeloid biology in these two mouse models. First, we tested NGM-707 in a CD34 humanized NOG-SCID model using the human melanoma cancer cell line A375. Tumor volume is on the y-axis while days post implantation are on the x-axis. Isotype is in black and NGM-707 is in green. Remarkably, NGM-707 drove a 70% mean tumor growth inhibition by day 47. Very promising data in this model system. Next, we tested human PBMC humanized mouse model, where we compared NGM-707 to the powerful checkpoint inhibitor anti-PD-1. Again, we used A375 tumors, which are known to be PD-1 responsive. We saw very promising results in this model as well, with stronger activity of NGM-707 as a monotherapy compared to anti-PD-1. In the graph, isotype is shown in black, NGM-707 is shown in green, and anti-PD-1 is shown in purple. We observed 64% tumor growth inhibition by NGM-707 compared to 38% tumor growth inhibition by anti-PD-1. How much did ILT2 or ILT4 contribute to this response? ILT2 blockade in red is tumor growth inhibition similar to anti-PD-1. ILT4 blockade, shown in blue, gave a strong response in this model. In contrast, in other mouse tumor models, we see ILT2 blockade driving the response with minimal contribution from anti-ILT4. What is very encouraging is that in all these models, NGM-707 consistently showed the strongest activity. This also infers that having an antitumor effect in more patients may require blocking both ILT2 and ILT4, as the receptor driving the response is likely to be variable from patient to patient. We also evaluated the combination of NGM-707 and anti-PD-1. Isotype is shown in black, anti-PD-1 is shown in purple, NGM-707 is shown in green, and the combination is shown in brown. Again, I want to remind you this is a tough model. PD-1, which typically has a potent impact in mouse models, here shows only a modest response. What is encouraging to us is that NGM-707 by itself shows a stronger response than PD-1 alone. Most importantly, the maximal effect is seen when combining PD-1 and NGM-707, validating our rationale to study the combination of these two therapies in the clinic. Importantly, NGM-707 has an impact even in a difficult tumor model such as this, suggesting that NGM-707 may have an impact across a broad set of tumors in patients, particularly in combination with anti-PD-1. I want to leave you with some takeaways about the NGM-707 program. ILT2 and ILT4 are expressed by many myeloid and lymphocyte cells in human tumors. ILT2 and ILT4 may be key checkpoint inhibitor resistance mechanisms in cancer. Our dual antagonist antibody, NGM-707, captures a powerful biology for both ILT2 and ILT4 blockade, synergistically enhancing myeloid and lymphocyte responses. NGM-707 overcomes the immune suppression environment generated by cancer-associated fibroblast cells. NGM-707, alone or in combination with anti-PD-1, can suppress the growth of human tumors in humanized xenograft models. With that, I'll hand it over to our NGM-707 product development team leader, Daniel Kaplan, to talk about our clinical development strategy for NGM-707. Dan? Thank you, Jeff. We've learned a lot today from Marco and Jeff about the ILT family and the distinct as well as overlapping immune suppressive activities of ILT2 and ILT4. It should now be clear why we developed NGM-707 as a first-in-class antibody targeting ILT2 and ILT4. We've seen NGM-707 has potential to reprogram ILT4-expressing myeloid cells and stimulate the activity of ILT2-expressing myeloid and lymphoid cells. As we design the first-in-human clinical study to test NGM-707, we gave a lot of thought to how to best bring the potential benefits of dual ILT2 and ILT4 blockade to patients. We hope that this clinical study will provide an opportunity to demonstrate the benefit of dual ILT2 and ILT4 blockade over targeting either individual receptor, as well as the potential superiority of this strategy as compared to targeting next-generation T-cell checkpoints such as LAG-3 or TIGIT. The objectives of the first-in-human study of NGM-707 are to determine the safety, PK, PD, and efficacy of NGM-707 alone and in combination with the anti-PD-1 antibody Keytruda. We ultimately hope to test three therapeutic hypotheses. These three therapeutic hypotheses are represented by different indications and lines of therapy. The first hypothesis is that NGM-707, in combination with anti-PD-1, may deepen and broaden responses in inflamed tumor microenvironments. Here, we would be treating IO-naive patients whose tumors are traditionally at least somewhat responsive to PD-1 blockade. We hope that NGM-707 in combination with Keytruda may enable a larger fraction of these patients to benefit than would benefit from Keytruda alone. Furthermore, we hope that these patients could experience deeper and more durable responses. The second hypothesis is that in patients who have relapsed on or been refractory to prior anti-PD-1 or PD-L1 therapies, NGM-707 in combination with Keytruda may be able to rescue these patients and enable them to respond where they would not respond to Keytruda alone. Our third therapeutic hypothesis is that NGM-707 may reprogram tumor-associated macrophages in immunologically cold tumors to an inflamed phenotype, enabling responses in these difficult tumors that do not traditionally respond to T-cell checkpoint therapies. In some of the coming slides, we'll discuss the specific indications that will enable testing of these three therapeutic hypotheses. Our clinical development plan for NGM-707 is shown here. Part 1-A, shown at the left, is a monotherapy dose escalation in solid tumor types that have high expression of ILT2 and ILT4. This monotherapy dose escalation is ongoing and progressing well. We're currently at the 1,200 mg dose level. Part 1-B, shown at the top middle, is a combination dose finding with anti-PD-1, which in our case is Keytruda. We're currently at the 600 milligram dose level and plan to progress up to 1,200 and 1,800 milligrams. 200 milligrams is a backfill dose level to provide a fuller understanding of the dose-response relationship of pharmacometric modeling. Based on the Part 1-A dose escalation and the Part 1-B dose finding, we'll identify a dose or dose levels to bring forward into dose expansion cohorts in specific indications. The phase 2, Part 2-A monotherapy expansion cohorts, shown in the bottom middle, are considered optional and will be guided based on our observations in Part 1-A. The recommended phase 2 dose of NGM-707 monotherapy would be tested in specific tumor types and lines of therapy in cohorts of 10 patients each and could be expanded up to 40 patients each. The phase 2, part 2-B, which is shown at the right, will test a combination of the recommended NGM-707, phase 2 dose of NGM-707 in combination with the standard of care dose of Keytruda in specific tumor types and lines of therapy in cohorts of 20 patients each that could be expanded up to 40 patients each. In the coming slides, we'll discuss some of our thinking about how the specific indications for these expansion cohorts will be selected. As we think about indication selection, it's useful to consider the characteristics of a tumor that would make it a good candidate for NGM-707 treatment, as well as looking back to the three therapeutic hypotheses that we hope to test. In general, tumors with substantial myeloid content and expression of ILT2 and/or ILT4 would be good candidates for NGM-707 treatment. The vast majority of tumors contain myeloid cells and express ILT2 and ILT4, but we do not yet know how many myeloid cells or how much ILT2 or ILT4 expression is enough to enable responses. Importantly, in cancer types that express ILT2 and ILT4, we believe that by inhibiting these two different receptors, NGM-707 will have a stronger effect than if you were to inhibit just one of these receptors. Through our ongoing clinical study, we just have to better understand which tumor types will be most sensitive to dual ILT2 and ILT4 blockade. We view most of the tumor types shown here, with the exception of those at the very bottom, as good candidates for testing the three therapeutic hypotheses that we laid out. To test our first hypothesis that NGM-707, in combination with anti-PD-1, may deepen and broaden responses in inflamed tumor microenvironments, we'd be looking to treat IO-naive patients whose tumors are traditionally at least somewhat responsive to PD-1 blockade. These tumor types can be identified as those in the final column on the right, shown in green and yellow as those high or medium in responsiveness to PD-1 or PD-L1 therapies. These same tumor types, but in an IO-experienced rather than IO-naive setting, could be used to test our second hypothesis that in patients who have relapsed on and been refractory to prior anti-PD-1 or PD-L1 therapies, NGM-707 in combination with KEYTRUDA may be able to rescue these patients and enable them to respond where they would not respond to KEYTRUDA alone. To test our third therapeutic hypothesis that NGM-707 may reprogram tumor-associated macrophages in immunologically cold tumors to an inflamed phenotype, the tumor types shown in red in the final column on the right would be excellent candidates. These are generally very difficult tumors to treat. If NGM-707 could enable responses in these tumors that do not traditionally respond to T-cell checkpoint inhibitors, this could have a profound impact on addressing the tremendous unmet need in patients suffering from these tumors. It's also worth noting that Merck has seen responses with their anti-ILT4 antibody in colorectal, gastric, head and neck, non-small cell lung, Merkel cell, ovarian, and thyroid tumors. These are certainly tumor types that deserve consideration. The categorization of ILT2 and ILT4 tumor expression on the previous slide was based on RNA expression data. I wanted to confirm these data at the protein level, which we accomplished using ILT2 and ILT4 immunohistochemistry assays that were developed right here at NGM. We analyzed ILT2 expression in 211 tumors and ILT4 expression in 117 tumors across seven cancer types. You can see examples of ILT2 and ILT4 staining in non-small cell lung, head and neck, and kidney cancers on the left side of the slide, and a quantification of the expression across tumor types on the right side of the slide. These data confirm that ILT2 and ILT4 are broadly expressed across these tumor types. If we dig into these immunohistochemistry data evaluating the expression of ILT2 and ILT4, we can gain some key insights that will help to guide our indication selection. Here, we're looking at non-small cell lung adenocarcinomas in orange, non-small cell lung squamous cell carcinomas in yellow, and renal cell carcinomas in black. The first key takeaway here is that every tumor expresses both ILT2 and ILT4. You don't see any dots along the axes indicating expression of one receptor but not the other. This is an important finding because it means that all of these tumors may have immune suppression driven by both ILT2 and ILT4, and the dual blockade provided by NGM-707 may be advantageous. Furthermore, you can dig deeper and find tumors like those highlighted by the blue oval. These are tumors that express both ILT2 and ILT4, but express particularly high levels of ILT2. These are tumors in which dual blockade of ILT2 and ILT4 by NGM-707 may be superior to blockade of ILT4 alone. These data also point to the potential utility of developing co-diagnostic strategies in which measurement of ILT2 and ILT4 prior to treatment could be used to select those patients who might be most likely to respond to NGM-707 treatment. We would now like to share with you for the first time some preliminary PK and receptor occupancy data from the ongoing NGM-707 clinical trial. The PK data are shown at the left, and you can see at the lower doses from 6-60 milligrams, target-mediated drug disposition is observed. You can see that the PK curves are nonlinear because NGM-707 is being cleared quickly due to its interactions with a relatively abundant target, ILT2 and ILT4. However, at doses of 200 milligrams and above, the PK curves become linear, indicating that all of the ILT2 and ILT4 have been saturated by NGM-707. The receptor occupancy data, which is shown on the right, backs up this observation. Full receptor occupancy on both ILT2 and ILT4 is observed even from the lowest dose of 6 milligrams, but the duration for which the receptor occupancy is maintained increases with the dose level. At the dose level of 200 milligrams, complete receptor occupancy is maintained for the entire 3-week dosing interval in most patients, and at the 600 milligrams dose level, complete receptor occupancy is maintained for the 3-week dosing interval in all patients. In addition to measuring PK and receptor occupancy, we have a comprehensive biomarker strategy for NGM-707 that is being used to assess target engagement and guide dose selection, to demonstrate proof of mechanism, and to potentially enable patient selection strategies. We're collecting blood samples at multiple time points to perform immune cell phenotyping, which is coupled with evidence of myeloid cell reprogramming, measuring circulating biomarkers, which may provide evidence of immune activation status, and we're evaluating genomic and circulating tumor DNA to assess tumor mutation burden and MSI status, genomic alterations, and driver mutations. Additionally, we're collecting tumor biopsies at baseline from all patients, as well as at baseline and then matched on-treatment biopsies from some patients, which are being assessed using multiplex immunohistochemistry panels we have developed in-house here at NGM, as well as using RNA-seq for global gene expression profiling. These analyses can help us to understand questions such as whether ILT2 and ILT4 expression correlate with clinical responses, as well as to see evidence of myeloid reprogramming or changes in immune infiltration in the tumor in response to NGM-707 treatment. I'd like to wrap up with a few key takeaways. Our first human clinical trial of NGM-707 is ongoing. This trial has the objectives of determining the safety, PK, PD, and efficacy of NGM-707 alone and in combination with Keytruda, and ultimately testing three therapeutic hypotheses in various cancer types. First, if NGM-707 in combination with anti-PD-1 may deepen and broaden responses in inflamed tumor microenvironments. Second, if NGM-707 in combination with anti-PD-1 may rescue responses in PD-1 or PD-L1 relapse or refractory tumors. Third, if NGM-707 may reprogram tumor-associated macrophages to an inflamed phenotype, enabling responses in immunologically cold tumors. The NGM-707 clinical trial is progressing well, with the monotherapy dose escalation nearing completion and the dose finding in combination with KEYTRUDA underway. Give me a good understanding of NGM-707's PK and the relationship between exposure and receptor occupancy on ILT2 and ILT4. 22 patients have been enrolled to date, and additional clinical sites are opening on a regular basis. With that, I'd like to thank you for your interest and attention, and at this point, I'd like to hand it back to David for some concluding remarks. Thank you, Dan. Before we transition to the Q&A portion of today's event, I want to remind you of the list of exciting milestones that we plan to achieve in 2022. Already this year, we've completed enrollment in our ALPINE 4 study of aldafermin for the treatment of patients with cirrhotic NASH and initiated phase 1 clinical trials for NGM831 and NGM438. Our most meaningful milestones for the year are still ahead of us. In the second half of this year, we plan to share updated clinical data from our phase 1A/1B trial of NGM120 in patients with cancer, provide an initial clinical update on the phase 1A portion of our NGM-707 trial in patients with advanced solid tumors, and provide a top-line readout from our phase 2 proof-of-concept CATALINA trial for NGM621 in patients with geographic atrophy. We look forward to sharing these updates with you as we continue to drive our pipeline of therapeutic candidates forward for the benefit of patients. With that, I'd like to open up our Q&A session. Good morning, and thank you again for joining us today for the third session of our Explorer series. This is Siobhan Nolan Mangini, Chief Medical Officer at NGM Bio. In the room with me are David Woodhouse, CEO, and Daniel Kaplan, Head of Translational Immuno-Oncology. I invite you to submit your questions through the Q&A box in the webcast platform at this time. The first question comes from Steve Seedhouse at Raymond James. Can you discuss any potentially unique safety and tolerability considerations for dual inhibition and how your clinical development plan concentrates any of those unique considerations for a dual ILT inhibitor? Hi, Steve. It's David. Thanks for the question. I wouldn't say it's anything unique. We are pushing on a couple of mechanisms as has been discussed today around immunosuppression. The type of safety signals we'll be looking for will be related to immune activation or inflammation. That's certainly what we're looking for as we dose escalate through phase 1a. There's nothing unique from a safety or tolerability point of view beyond that. Great. The next question from Steve is, monotherapy responses to ILT4 or ILT3 monoclonal antibodies has been few. Do you expect that to hold true for this dual targeting antibody, or are you hopeful for multiple monotherapy responses to demonstrate bona fide activity? Yeah. We're not expecting more from the monotherapy aspect of just the phase 1a, looking for PD activity necessarily. We're trying to get through the dose escalation as quickly as possible. We're not looking at it in too many patients. We also are aware, of course, of the clinical data that's seen so far, and we're really just trying to get as quickly as possible to where we think the most meaningful responses will be, which is in combination with anti-PD-1. Mechanistically, as you've heard, there's a reason to believe there could be some monotherapy responses, but we're not sort of holding up the molecule in that portion of development, to see that. Great. The final question comes Steve Seedhouse. Can you comment on early evidence of antitumor activity and if you are observing any in either mono or combo dose escalation portions? Yeah. Because we've guided towards the phase 1a, we'll be sharing at a conference in the second half of the year. That's when we'll be sharing what we're seeing both from a safety, further PK, and then any efficacy endpoints we have at that point. Great. The next questions are from Mayank Mamtani at B. Riley. It would be helpful to understand the rationale for monotherapy dose-expansion cohort being prioritized for RCC, CRC, and ovarian cohorts while pembro combination cohort is being explored in lung and neck cancer. Also, could you clarify if there is any other GI cancers being explored in this space, such as pancreatic, gastric, and biliary tract cancer? Great. That's a good one for Daniel Kaplan. I'll hand it off to Daniel. Great. Yeah. Thank you for the question, Mayank. Your question really ties into the three therapeutic hypotheses that we've described for NGM-707. In terms of evaluating the NGM-707 monotherapy in CRC and ovarian cancers, these are cold tumor types that do not typically respond to T-cell checkpoint inhibitors. Thus, we believe that these are good tumor types to evaluate the hypothesis that blocking myeloid checkpoints such as through ILT4 can heat up the tumor microenvironment and enable responses in these difficult tumors. HCC is a different case. This is actually the only tumor type in which we've observed really robust expression of HLA-G in our study. Thus, we believe that this is an excellent tumor type to address the question of whether reversal of HLA-G-induced immunosuppression can enhance antitumor immune responses. To lung and head and neck cancers, these are tumor types that provide great opportunities for evaluating NGM-707 in combination with Keytruda. These tumors allow us to test the other two therapeutic hypotheses that we've described. In the IO-naive setting, testing NGM-707 plus Keytruda will allow us to address whether NGM-707 can broaden and deepen responses in this IO-responsive setting, and then testing NGM-707 and Keytruda in lung and head and neck cancer patients who've relapsed on or whose tumors are refractory to anti-PD-1 or PD-L1 therapy, allow us to address the hypothesis that NGM-707 can help rescue these patients and enable responses to combination therapy where we would not expect to see responses to IO alone. The final part of your question around GI cancers being explored. These are certainly on our radar and as we finalize the expansion cohorts that we're going to be evaluating there, we're considering a number of different factors and some of these tumor types may indeed make the cut. Great. The next question from Mayank Mamtani is, since ILT proteins are not expressed in rodents and HLA-G expression heterogeneity issues, what patient enrichment strategies have emerged from the work that you and your peers have done in optimally interrogating the PD responses to ILT2 antagonism? Right. The issue we're dealing with here, of course, is this is relatively specific biology to human beings. As you've heard and seen in our work, we do a lot of work in vitro setting with human cells. We also shared some rodent models today, as you've seen. That really makes what we think is the exciting aspect of biology is we're working with something that seems really relevant with humans, not necessarily rodents. As Dan has gone through in his presentation and was just referencing, really our objective here with our biomarker strategy, as well as just interrogating this molecule in the clinic, is to understand the very question you're asking, which is which patient enrichment strategies can we use? We're obviously starting with tissue type the tumor originates in, but there are all sorts of other biomarkers that we're measuring. Dan, would you add anything to that? I'd just say, certainly if we focus on human expression data, the expression of ILT2 as well as ILT4, of course, is associated with poor outcomes, lack of response to checkpoint inhibitors in a number of cancers. We can use that information to guide us. Specifically on the predictive biomarker strategies. We are measuring ILT2, ILT4, and HLA-G in our study and assessing whether they will be useful as predictive biomarkers. If they do indeed prove to be associated with patient responses, we'll have assays in place that can be used to measure these proteins for patient selection. Great. The next question is from Ritu Baral at Cowen, and it builds on what we were just discussing. Are there obvious PD biomarkers for this mechanism you'll be monitoring? Which ones did Merck look at? Yes. Dan, do you wanna cover that one? Yeah. I think the PD biomarkers associated with myeloid reprogramming would be top of our list, and this can be assessed on immune cells in peripheral blood as well as in tumor biopsies. It can also be assessed by measuring cytokines in the peripheral blood. In terms of what Merck has reported so far, all they've reported are those peripheral cytokine measurements where they actually did not see any significant changes even in those patients who had responses. I think that points to some of the difficulty in seeing changes in the peripheral blood where the real action happening actually in the tumor. In our study, we are measuring, everything we can in the peripheral blood. Still hoping to see evidence of myeloid reprogramming there. Crucially, we're also looking at biopsies, on-treatment biopsies from patients, whenever we can get these. There we'll actually be able to assess evidence of myeloid reprogramming in the tumor using immunohistochemistry assays we've developed as well as RNA-seq assays. Great. Congratulations on the phase 1b dosing. Next question is, in your indication selection chart, you showed different parameters, myeloid cell content, ILT2 expression, ILT4 expression, PD-1 response across a broad set of tumors. Are there any parameters that you think will be more or most important? Yeah, thanks, Ritu. We're certainly thrilled to be moving on to the phase 1b portion. Exciting milestone for us. I'll let Dan continue to expand on this idea of the indication selection chart and relative importance you may find on either. Dan? Yeah. Thank you. I think the important takeaway here is that there's a wide opportunity to see effects based both on ILT2 expression and ILT4 expression. Of course, we group cancers by tissue types, as we showed in the presentation today. Really, each tumor is different, each patient is different. What we've seen in our preclinical work is that it's quite hard to determine which receptor, ILT2 or ILT4, is gonna drive a response in a particular model. This is one of the reasons why we think that hitting both receptors is so valuable. There we cover both bases, so we're better able to address the heterogeneity of tumors. We really think that a heterogeneity in immune suppression in the tumor microenvironment is kind of one of the key issues that you have to overcome to enable responses. We think that this dual-hit strategy can, best help us address that. The next question comes from Paul Choi at Goldman Sachs. How do you think about injection volumes for PD-1/NGM-707 combos as well as scheduling? Thanks for the question, Paul. It's David here. We've actually designed NGM-707 to really be in sync with an anti-PD-1, so every 3-week infusions, they're sort of done sequentially. There's nothing unusual about the volumes being used. We thought ahead for that. In other words, in terms of type of cycle the patients would be coming in on. As you've seen from the PK data, that is supported by the length of receptor occupancy that we shared with you today. The next question's from Paul Choi. Does it make sense to combine NGM-707 with a SHP2 inhibitor? Any potential tox issues? Sure. That's not something we spent much time thinking through. Dan, do you wanna expand on any thoughts on that? Yeah. I mean, I think it's quite a different approach from what we're trying to achieve with NGM-707. An SHP2 inhibitor would be, you know, much less selective, targeting many pathways all at once with, I'd say in my opinion, somewhat unpredictable effects there. Whereas we're really trying to go after these two pathways that we believe are strongly connected to immune suppression in the tumor microenvironment, and where their specific reversal may enable responses. Great. Our last questions will come from Swapnil Malekar at Piper Sandler. What was the rationale for specifically targeting ILT2/4 versus the other ILTs? Yeah. It's a good question, Swapnil. As you can probably gather, we have antibodies against this entire family, and so what you're really seeing in the instance of our three clinical stage programs are what we believe are the optimal strategies. I think you've heard today there's a pretty strong redundancy argument for ILT2 and ILT4 together, just in terms of structural homology of the receptors and expression patterns, namely in myeloid cells for both. There is intriguing biology of ILT2 in effector cells, and we think that for lymphoid lineage cells could actually be an interesting pairing of those two approaches. Importantly, they also share the same ligand receptor relationships. Again, kinda coming back and overlapping on this redundancy argument, that seems like biology that is shared between the two receptors that would have a rationale for blocking both. I'd remind you, too, that in non-human primates, there's actually one version of this receptor, and it seems somewhere in evolution between monkeys and humans, this is split into two, and thus again the rationale using one antibody to block both. That was the versus, let's say, ILT3, which we've spoken a little bit about today. That has a different ligand receptor relationship. We've talked about in our previous Explorer Series the fibronectin binding. In fact, Dr. Colonna actually touched on it today, too. That's a different biology of more of a stromal checkpoint that we didn't wanna overlap too broadly in the system. We think the dual antagonism mixture really makes a lot of sense. Great. This is the final question for today, which you were addressing just already, but how closely are ILT-2 and ILT-4 related to ILT-3, and what is the read-through from Merck's ILT-3 update at ASCO as it pertains to NGM-707? Sure. Maybe since I began to touch on it, Dan, do you wanna expand on it, and then I'll touch on the Merck data? Yeah. No, I think David already did touch on some of this, but you know, there are similarities as well as important distinctions between ILT2, ILT4, and ILT3. The expression of all three of these receptors is quite highly correlated in myeloid cells, but ILT4 does show somewhat broader expression in some myeloid subsets with expression in granulocytes, granulocytic MDSCs, neutrophils, where ILT3 is not expressed. Of course ILT2 is also expressed in T cells, NK cells, and B cells. Then as David also spoke to, these receptors have distinct ligands with immunoglobulin G1 ligands for ILT2, ILT4, and then fibronectin A2E function as ligands for ILT3. Yeah. Maybe just to touch on the second part of the question as it relates to Merck's MK-0482, which is some phase 1 data for their anti-ILT3 antibody that they released an abstract for ASCO that, presumably there'll be a little more information later this week. Those of you who didn't see it, I saw some modest antitumor activity with an overall response rate of about 15%. Obviously, there's more to dig into here generally just around dose levels and what type of tumor types those partial responses were seen in. It seems as typical in these early phase studies that these are patients that are fairly advanced, getting through most of them a couple of rounds of previous therapies to get here. While this is a modest response, I think it is encouraging, mainly because we have an important source of differentiation that, we think, is important to the effect of inhibiting ILT3, which is this interaction we discovered at NGM with fibronectin being a ligand for ILT3, is something that our antibody blocks, but Merck's does not. In sort of a baseline level of activity, relatively modest antitumor activity in combination with the PD-1 receptor, PD-1 inhibitor, this abstract at ASCO we view as being encouraging for our program, and sort of think of it as a floor of activity perhaps related to other ligand interactions, and we think fibronectin will be one to block as well. We'll look forward to learning more about it later this week, and also, obviously seeing data from our program. Just to remind you, our NGM-831 program, our anti-ILT3, is in the clinic and in phase 1a dose escalation right now. Great. Thank you, David, and thank you all for your questions today. With that, we will conclude the Q&A session in the third module of our Explorer Series. I look forward to speaking with you again soon for our next Explorer Series module, which will detail NGM-621, our monoclonal antibody product candidate engineered to potently inhibit complement C3 in patients with geographic atrophy. We look forward to speaking with you again soon during June. Thank you.