Compugen Ltd. (CGEN) Earnings Call Transcript & Summary

All right. So it's 9:10. Let's go ahead and get started with day 2 of the Oppenheimer Fall Healthcare and MedTech Summit (sic) [ Fall Healthcare Life Sciences & MedTech Summit ]. My name is Mark Breidenbach. And I'm pleased to introduce our next presenting company, Compugen Ltd., based out of Holon, Israel and South San Francisco, California. Joining us this morning to walk us through the story, we have Compugen's CEO, Anat Cohen-Dayag. And for those new to the story, Compugen is a clinical stage therapeutics company focused on drugging novel immune checkpoints with applications in oncology and autoimmune disease. The company's lead program, COM701, has demonstrated monotherapy activity in a Phase I dose escalation trial and is currently being evaluated in expansion cohorts focusing on specific tumor types. COM701 also recently entered into an all-comers triple combination trial with Bristol-Myers Squibb's TIGIT antibody and PD-1 checkpoint antibody. As Anat will explain, COM701 acts on a novel target that belongs to the same signaling axis as TIGIT, which, of course, has received considerable attention lately from big biopharma companies, including Roche, Gilead, Merck, Bristol-Myers Squibb and maybe some others. So at the end of the presentation, I just want to say, we'll save some time for Q&A. [Operator Instructions] Feel free to type questions as the presentation is underway and they'll cumulate so we can go through them at the end of our session. So with that, I'll hand the mic over to Anat. Please take it away.

Thank you, Mark, for this terrific introduction, and thank you for Oppenheimer for inviting Compugen to present at the conference. Compugen, as Mark was saying, competent is a therapeutic discovery company and -- therapeutic discovery and development company. We discover new drug targets, new biological pathways. And our vision is to transform patient lives by developing first-in-class therapeutics that are addressing these drug targets. The way we discover new drug target is by target discovery platform, computational platform that we developed for many years in the company. So basically today, Compugen has 3 key building blocks: an innovative immuno-oncology portfolio, strategic collaborations and a computational discovery engine that feeds our own pipeline. On the front of the innovative immuno-oncology portfolio, as I indicated, the reason it is innovative is because we discover our own drug targets. Today, we have 3 drugs in clinical studies. COM701 is our leading program. It is an antibody targeting PVRIG, a new immune checkpoint that we identified. This program, we recently shared encouraging preliminary clinical data, and this program is under clinical collaboration with BMS, with Bristol-Myers Squibb. COM902 is the program behind COM701. This is our internal TIGIT antibody that we designed to work in combination with COM701 in blocking PVRIG and TIGIT in parallel. And this is due to what we call the DNAM axis hypothesis that I'll explain later in this presentation. I'll just remind, to those that are maybe not familiar with the story, that we discovered TIGIT. Back then in 2009. We sent it to publication back-to-back with Genentech. But the reason we develop our own TIGIT program is not in order to compete with the rest of the TIGIT programs, but really in order to extract the full value of COM701 in tumor types where both PVRIG and TIGIT should be blocked. The third program is a program under collaboration with Bayer. This is an antibody that is targeting ILDR2, a new immune checkpoint that we discovered, and this is now in Phase I studies that are handled by Bayer. Behind this clinical stage pipeline, we have an earlier-stage pipeline that consists of multiple programs addressing immunosuppressive tumor macro environment, mostly focused on myeloid programs with myeloid biology. The second key building block is the strategic collaborations, and these are collaborations with leading pharma companies that are very different from one another. This is with Bristol-Myers Squibb, Bayer and AstraZeneca, and I'll relate to it in a moment. The third key building block is the discovery engine that we developed for many years in the company, a computational platform. And I'll say 2 things about it. And mainly because there is a lot of recognition today of the contribution or the possible contribution of computational discovery in the field of life sciences all over. But Compugen is highly differentiated on 2 fronts. First, our engine is proving. From computer prediction to clinical studies, we pushed 3 new drug targets that we discovered: PVRIG, TIGIT and the ILDR2. And these were done through successful preclinical studies. And today, we're sitting here with COM701 targeting PVRIG showing some initial clinical validation. And for the TIGIT pathway, being validated already by others clinically. So this is one proof-of-concept of the engine. And the second differentiating factor is the fact that in one company, we have the computational discovery capabilities integrated with drug development capabilities. And when you start with the end in mind, it matters. You know better how to discover drug targets or proteins that could be translated to drug targets that would lead to products. So this is us. And the pipeline that we developed actually with -- based on our computation and discovery engine is very diversified. It consists of programs addressing new drug targets, at different stages, different drug targets, and some of them are internal and some of them are partnered. And you can see COM701 with the multiple clinical studies that are ongoing now as monotherapy and combination therapy in collaboration with BMS. With respect to the specific collaborations, I said that they are very different from one another. And what we're trying to do is really to tailor the collaboration to the specific development candidate and to our needs for growth as a company, both in terms of access to revenues, access to additional capabilities and capacity, but also in order to retain value for the company. So the Bristol-Myers Squibb collaboration, that's a clinical trial collaboration that we entered in October 2018. Under the collaboration, BMS supply Opdivo and its anti-TIGIT inhibitor, which is an investigational drug, and we conduct a combination study. We retain the ownership. We retain the ownership and the commercial rights for the program, and BMS has right to first negotiation during the exclusivity period. At the time that we entered the collaboration, BMS entered into strategic equity investment in 30% premium. And today, they own about 3% of the company. The Bayer collaboration is a collaboration that we entered at a very early stage, and it was in August 2013, very early stage in terms of the stage of the drug target program. This is a first-in-class candidate targeting ILDR2, a completely new immune checkpoint. It is now in Phase I studies, tested in monotherapy and in combination with Keytruda. Under the collaboration, we got up until today, over $30 million in upfront and milestone payments and we're eligible for over $250 million in milestone payments and mid- to high single-digit royalties. The third partnership is with AstraZeneca. That's actually a license agreement that we entered on March 2018. Under the partnership, we licensed to AstraZeneca the rights to develop bispecific antibodies to one of the programs that is in our pipeline. AstraZeneca is responsible for all the R&D and commercialization activity, and we retain the rights for -- the rest of the rights other than the specific bispecific rights that we licensed to AstraZeneca. For this agreement, we got $10 million upfront and we're eligible for packages of milestones and royalties for each and every program that is going to be developed based on this targeted antibody. Moving forward, I'll focus on COM701 and COM902, and I'll explain to you the signs behind the 2 pathways of PVRIG and TIGIT and how we translated, in science-driven manner, all these understandings and preclinical data that we have into a clinical strategy. And then I'll follow-up with some information about the initial clinical data that we presented and some guidance with respect to data readouts. So we'll start with the basic pathway of TIGIT as it is described in the literature, by the way, also including the paper that we published in 2009. TIGIT is a negative costimulatory receptor. It is present on NK and T cells. And it is sending inhibitory signals to the immune system in the context of the cancer, and it binds PVR and PVRL2 as its 2 ligands. And the notion is that TIGIT blocking antibodies are actually releasing the inhibitory signal of TIGIT and releasing PVR and PVRL2 to stimulate DNAM, which is a positive costimulatory protein in this axis. So you actually block inhibition and you stimulate the stimulation by DNAM. And that's the notion out there. And as I said, also supported by the paper that we published in 2009. But then we came back to this pathway. And actually, we discovered that there is another protein here in this pathway. This is PVRIG. And this PVRIG is also a negative costimulatory protein. But this is the receptor, in our view, that binds preferentially PVRL2. And we think, based on our data, the PVRL2 does not bind TIGIT under physiological conditions. So basically, these are 2 parallel pathways, TIGIT PVR and PVRIG/PVRL2. And we are saying that in tumor types where the 2 pathways are operative, you need to block the 2, not only in order to release the 2 negative costimulatory signals of TIGIT and PVRIG, but also in order to be able to release PVR and PVRL2 to stimulate DNAM. And in this way, you'll maximize the signal in certain tumor types where the 2 pathways are operative. But there is one more piece to the story. And if we'll move to the next slide, you'll see on the right side, that's exactly what I described, PVRIG/PVRL2 and COM701 inhibit blocking PVRIG and TIGIT PVR and COM902 is blocking TIGIT. But on the right side, you'll also see the PD-1 pathway. Now in the literature, in the last 2 years, there are some evidences of molecular intersection between the PD-1 pathway and DNAM. And this is highly supported by our preclinical data. What we are saying is that it's a 3-pathway story. And that in different tumor types, you'll get different dominance and different expression of these 3 pathways. And therefore, you'll need to treat these patients with different drug combinations out of these 3 pathways, and I'll explain. In patients where the PD-1 pathway is mostly dominant, you treat these patients with a PD-1 blocker, and they may respond. But in some cases, it will not be enough. You'll still need to block 1 or 2 of the other pathways. And this is the basis of what we're saying. One of the key questions that we're asking us is actually in which tumor types the PVRIG/PVRL2 pathway is mostly dominant and where we could employ COM701? In which patient populations, we should employ it in order to maximize the potential of patients to benefit from this treatment. And when we were looking at the different expression profile of the different receptors and ligands of this axis, we found out that PVRIG/PVRL2 pathway, designated by high PVRL2, is actually present in tumor types like breast, ovarian, endometrial. These are tumor times that, as you know, have low PD-L1 and usually are not responsive to PD-1 blockers, which is supporting our hypothesis. We also identified that high PVRL2 can be present in tumor types like non-small cell lung cancer that are responsive to PD-1 blockers, but these are -- a PVRL2 is present in patient populations that have high or low PD-L1. So this gives a horizon for maybe being able to expand the response in this in these tumor types that are responsive to PD-1 blockers. One thing that we did identify, and you'll see that in the initial clinical data that I'll present to you, as we move forward to the -- in the clinical study, we identified that patients with MSS-CRC are also showing anti-tumor activity. And this was not in a cancer indication that we selected to begin with. And the reason for that is because the PVRIG/PVRL2 pathway is not highly expressed. So PVRL2 is not highly expressing in MSS colorectal. When we saw the data in MSS colorectal, we were highly encouraged, and that's because this tumor type is moderately expressing PVRL2. And from our perspective, showing some response in these tumor types was highly encouraging in terms of supporting our DNAM axis hypothesis. So with this understanding of the signaling, of how these pathways are parallel but also complementary, we moved ahead in order to design our clinical strategy and the clinical trial's design. And you can see that the Phase Ia Arm A (sic) [ Phase I Arm A ] is dealing with testing COM701 as a monotherapy agent, in dose escalation and in cohorts expansion in the indication that we tested preclinically as could be positive or dominant for PVRIG/PVRL2 expression. Phase I Arm B has to do with testing COM701 in combination with nivolumab in tumor types where we think that would make sense for the 2 pathways to be more dominant. And this is under clinical collaboration with BMS with Opdivo. And the Phase I/II study that we just initiated 2 weeks ago, it's a triple combination study where we test dose escalation and expansion cohorts in tumor types where we believe COM701 is most likely to affect. One arm of the triple combination expansion cohort has to do with a basket study that will select high PVRL2 expression patient population that will test this biomarker strategy that -- of the pathway. I'll just say that in terms of biomarker strategy, as a company, we take it step by step. So in the monotherapy expansion cohorts, and -- we selected the cancer indications based on the preclinical data that we have for the pathway. And this is a biomarker-informed selection of the indications. The second step is actually taking biopsies at the expansion cohort stage, and testing on these biopsies, this biomarker approach that we're taking, the different DNAM axis members, specifically PVRL2, PVRIG and others in order to be able to test the hypothesis and try to correlate to response. And you can see that in the triple combination expansion cohort, we took a more -- another step forward, and we're going actually to select patients to the expansion cohort, at least 1 arm, based on the biomarker approach with high PVRL2 expression. So this is the design of the study. And at the recent AACR, we shared data from the monotherapy dose escalation and the combination of COM701 plus Opdivo dose escalation. Basically, for the dose escalation of the combination, it was up to 10 mg per kg. And what we found out is the following. First, with respect to the safety profile. COM701 presented a well-tolerated profile in those doses that we tested it. So no issues on the safety profile, both as monotherapy and also in combination as we presented up to 10 mg per kg. We also presented data showing anti-tumor activity, and that was very compelling. Even though this is initial data, it's really very small number of patient population. But also remember, these are patients that were heavily treated with multiple lines of therapy. And they really exhausted all these lines before they reached to this study. But we were able to present 2 confirmed partial responses. One response was for patients with platinum-resistant MSS primary peritoneal cancer, which is a type of ovarian cancer. First, it's the fact that getting partial response in ovarian cancer as monotherapy, that's a very nice achievement. But I'll say more than that. Remember, when we were looking to understand the pathway rationale in specific indications, ovarian indication was one of these indications that we said PVRIG/PVRL2 may play a role. And getting this response was really encouraging for us supporting our preclinical data and our biomarker strategy. More than that, as I already told you previously, we got confirmed response in one patient with MSS colorectal cancer in the combination dose escalation, together with nivo, COM701 plus nivo. And again, in this highly hard-to-treat cancer indication, we were able to obtain partial response. So we were highly encouraged in an indication where the pathway is moderately expressing. So this data was really highly supporting our understanding of the pathway and what we were expecting to see. Although with the current monotherapy landscape and the failures that have been in clinical studies, it was really nice for us to see some anti-tumor activity also in monotherapy. More than that, we were able to see high disease control rate, both in the monotherapy and in the combination arms; and durable response of over 6 months in some of the patient populations across treatment. The data also, as I said, supported the biomarker-informed approach and the predictive discovery capability that we employed in the company. So that was highly encouraging for us. One thing that I want to mention here is really relating to how the PVRIG and TIGIT pathways are distinct from one another. And I'll tell you why I'm relating to this. When we started to present this DNAM hypothesis and the 3-pathway story, actually only the PD-1 pathway was validated. But today, I'm sitting in front of you and actually, the PVRIG pathway already have some initial proof-of-concept, still initial. And TIGIT was validated as a clinical relevant pathway by others. Yes, there is a question whether it's relevant in PD-L1 high, low and some additional questions, but in the randomized study, it was showing some clinical activity. So the question came. Okay, so how do you know that PVRIG and TIGIT are not redundant and that you're just doubling down on the same pathway? So the answer is no, and these are the reasons for that. PVRIG and TIGIT are binding 2 different ligands. And the ligands are differentially expressed in different tumor types. PVRL2 is more dominant in its expression in tumor types like breast, endometrial, ovarian tumor types that are not really responsive to PD-1 blockers. And we'll need to see how responsive this will be for TIGIT blockade. But also the receptors themselves, TIGIT and PVRIG, they are differentially expressed in immune cells. They are both expressed on T cells and NK cells, but TIGIT is highly expressed on Tregs and PVRIG is not. And also, they are differentially expressed in the tumor microenvironment. PVRL2 has different -- has preferential expression in some -- to myeloid immune cells as opposed to PVR. So with the totality of our understanding of the preclinical data that we have, also supported by the data that we saw in indications that are not responsive to PD-1 blockade, we are focusing on cancer indications that are not responsive to PD-1 blockade. And this is a highly differentiated strategy. And we do think that this is -- that these 2 pathways are complementary but still distinct. And we -- we're looking forward to be able to test this in combinations. Double combinations of PVRIG, PD-1 blockade, PVRIG, TIGIT blockade and the triple combination with the BMS. So for that purpose, we also developed our TIGIT antibody and because we wanted to make sure that we can fully address the DNAM axis in combination with COM701. We developed an antibody that is high affinity. It has a femtomolar affinity TIGIT antibody. It has preclinical proof-of-concept that was demonstrating synergistic activity with COM701 and we recently initiated our first Phase I study for patients with advanced malignancies going -- it is now ongoing. The end goal for us for this specific anti-TIGIT inhibitor is that while we move ahead forward with the triplet inhibition study with BMS, we would like to make sure that we can act also on testing the doublet inhibition of PVRIG and TIGIT and its potential outside of blocking in combination with PD-1 -- in a PD-1, PD-L1 independent setting. So as I said, the study started with the monotherapy dose escalation stage. It's an all-comers population. The dosing regimen is Q3 weekly, IV. And initial data are expected in 2021. So if we go ahead to the clinical studies that are ongoing and the guidance on data readouts, we completed enrollment in the combination study of Opdivo plus COM701. And we also initiated enrollment of the monotherapy expansion cohort. And we're aiming to complete the monotherapy expansion cohort enrollment until the end of the year. And then in the first half of 2021 to share data from the monotherapy expansion cohorts and the rest of the data from the doublet dose escalation of COM701 and Opdivo. We also initiated 2 weeks ago, the Phase I/II triple combination with Opdivo and BMS TIGIT inhibitor. And since we just started it, we did not share guidance yet. We will progress with the study. And obviously, we share guidance with respect to data readouts from this study. On the COM902 front, as I said, we initiated the Phase I study on March, and initial data are expected in 2021. And as I said, the idea is for us to keep the potential to test the combination of COM701 plus COM902 in a PD-1, PD-L1 free regimen. And just one more slide, sharing the financial position of the company. We ended June 2020 with $136 million. With the current cash expenditures, gross cash expenditures, this gives us enough cash in order to reach key milestones, key drivers of the company, and this is obviously without taking into consideration any cash inflows from other sources. And so this is it. And I thank you for taking the time to listen to the corporate presentation of Compugen.

Okay. Thanks very much, Anat, and that was a terrific update. And let's go ahead and kick off the Q&A.

I see there is -- there are some questions coming in from the audience. But let me start with some kind of big picture questions. Obviously, we've seen some pretty serious investment recently in the TIGIT space from big biopharma, especially Genentech, which is pushing tiragolumab into pivotal studies. And also at this year's ESMO conference, we have new data from Merck's TIGIT antibody showing modest monotherapy activity in patients who have failed higher PD-1 treatment. So I guess the 2 important questions I really want to start with is, why we should be confident that the TIGIT DNAM signaling axis is not going to be idle all over again? And second, I'm wondering if you've seen anything in the data that's coming out of Merck's and Genentech's TIGIT trials that reinforces your hypothesis about PVRIG and PVRL2 and its role in that signaling axis. Maybe start with those 2 questions.

Okay. Thank you. Yes. So with respect to the data that is out there, obviously, its initial data. There is a huge investment that is ongoing now in multiple tumor types, and that just highlights the potential of this axis, at least in the view of Genentech and Merck. I think it's one thing that we need to remember when we look at the data, and obviously, everyone -- many were burned on the IDO example. I think that here, we're dealing with a study, it's a Phase II study, but it's a randomized study. And as much as everyone will wait for the Phase III study and wait to see the data, but it's a randomized study that was randomized to atezo, and it was seen that TIGIT was adding. The question is, how exactly? How much? What will be the magnitude? What will be the patient population? Is it only PD-L1 high? Will it be in PD-L1 low? In which patient populations? But remember, we saw only data in non-small cell lung cancer. So there's still additional indications, there's lines of therapies, data from different pharma companies. The story still need to be told and still need to be uncovered. But when you're talking about Compugen, I'll say 2 things. First, we have our data that we're following for 3 years now, more than 3 years. Yes, it's preclinical data. That's correct. It's not clinical data. But for 3 years, we're building the science behind this axis and the interaction between these 2 pathways. Now it could be that this whole axis will not be translated to clinical as we think about it. But hey, we have evidences now much more than we had even a year ago. So COM701 was showing some initial clinical data, and that's encouraging for us. And TIGIT was showing some initial data, and that's encouraging for us, and it fits what we're saying for quite some time. And when you look at the biology, and this is why I cover the difference between PVR and TIGIT, but also how they complement one another, there is some basis to think that this would work together, either with PD-1 and without PD-1 or in combos with PD-1. But the story still need to be uncovered. So we're more encouraged with the new data, but there are still a lot of questions to answer. That's for sure.

Okay. Fair enough. I'm going to try and distill some of the audience questions down into maybe shorter format questions than they arrived. I'm seeing some level of frustration over the sort of the pacing of data readouts from the COM701 monotherapy trial. Given that it's an open-label study, why wait so long before giving the Street updates on data from this trial? What's the rationale behind waiting all the way until next year before the next update in terms of clinical results?

So look, we shared data on 6/8/2019 and then at AACR 2020. So we're just between April and August, there are not that many. But I understand the question going forward. And I think that this is clear that we not only need to enroll in the monotherapy, but we also need to follow-up on the patients. And this takes time. And mainly, you saw the data and we got some durable responses with COM701. All of this takes time, and we need to make sure that we complete the enrollment and then share the data. So this is -- we're not trying to delay time lines. We're not trying to really hide data. We really take the time to be responsible enough to share data that we think are -- data that are representative of what we will be able to see. That's it.

Right. And I guess if there's no financing overhang, there's no need to salami slice out the data from this trial.

That's correct, but I wouldn't say that we could share the monotherapy data earlier than that. We really need to enroll the patients. And we said that we'll enroll them until the end of the year. And then we also need to monitor these patients. So it takes time.