Ladies and gentlemen, thank you for joining us today. Welcome to Compugen's Second Quarter 2021 Results Conference Call. At this time, all participants are in a listen-only mode. An audio webcast of this call is available on the Investors section of Compugen's website, www.cgen.com. As a reminder, today's call is being recorded. I would now like to introduce Yvonne Naughton, Head of Investor Relations and Corporate Communication.

Thank you, operator, and thank you for joining us on the call today. Joining me to present prepared remarks are Anat Cohen-Dayag, President and CEO; and Ari Krashin, CFO and COO. For the Q&A session, we will also be joined by Henry Adewoye, CMO; and Eran Ophir, Vice President, Research and Drug Discovery. Before we begin, I would like to remind you that during this call, the company may make projections or forward-looking statements regarding future events or business outlook, our development efforts and their outcome, our discovery platform, anticipated progress, results, and timelines for our programs, financial and accounting related matters as well as statements regarding our cash position. Such statements represent only the company's current beliefs, expectations and assumptions, while actual results, performance or achievements of the company may differ materially. These statements involve known and unknown risks and uncertainties and we refer you to our SEC filings for more details on these risks, including the company's most recent report on Form 20-F filed on February 25, 2021. The company undertakes no obligation to update projections or forward-looking statements in the future. With that, I will turn the call over to Anat.

Thank you, Yvonne. Good morning and good afternoon everyone, and welcome to our second quarter 2021 update. Our continued progress through 2021 has been strong with steady execution, solidifying our leading position in the DNAM axis and differentiating us in the TIGIT space as the only company targeting in a clinical setting, PVRIG, TIGIT and PD-1 as part of our three-pathway hypothesis. On today's call, I'm pleased to have the opportunity to remind you of our strategy, provide perspectives on our most recent data, our views on important developments in the field and what's to come for the remainder of 2021. We believe that the foundation that underlies our success at Compugen is our science and our people. We were the first to identify PVRIG and ILDR2 as novel checkpoints and we published on TIGIT the same year as Genentech. Both PVRIG and TIGIT are key parallel and complementary inhibitory pathways in the DNAM axis, which also intersect with the well-established PD-1 pathway. While TIGIT and PVRIG pathways share similarities, we observed key differences between the two with respect to their expression pattern and their ligand expression pattern on immune cells and tumor types. Furthermore, our recent data shows that PVRIG is expressed similarly to PD-1 and TIGIT in stem cell-like memory and exhausted T cells, an important cell population with the potential role in mediating antitumor effects. However, recently Eran Ophir, our VP of Research and Drug Discovery presented scientific data showing that PVRIG has a more dominant expression pattern on early differentiated stem-like memory T-cells than TIGIT and PD-1, further pointing to the possibility that PVRIG may act differently. We believe that the future of immuno-oncology will be driven by combination approaches. Research from Compugen suggests that the PVRIG, TIGIT and PD-1 pathways have different dominance in different…

Thank you, Anat. Good morning and good afternoon to everyone. Our financial results for the second quarter of 2021 released this morning continue to show our strong financial position as we execute across our growing clinical programs. Research and development expenses for the second quarter of 2021 were $6.8 million compared with $4.4 million for the same period in 2020. The increase in expenses reflects the continued execution and expansion of the various clinical trials, Phase 1 programs. Net loss for the second quarter of 2021 was $9.5 million or $0.11 per basic and diluted share, compared with a net loss of $6.2 million or $0.08 per basic and diluted share for the same period in 2020. As of June 30, 2021, we had approximately $111 million in cash and cash-related accounts compared with approximately $119 million as of March 31, 2021, the Company has no debt. As a reminder, we expect our gross cash expenditures for 2021 to be in the range between $40 million to $42 million without taking into consideration any potential cash inflows for the Company from existing and new collaborations. Thank you and with that, we will now open the call for questions.

[Operator Instructions]. Our first question is from Mark Breidenbach of Oppenheimer.

Congrats on the progress. So let me maybe start with the kind of science you've won. We certainly appreciated Eran's presentation showing the fraction of TCF7 positive cells is important for response to checkpoint therapy, and that there's expression of PVRIG in stem-like T cells. I guess I'm wondering if you see TCF7 as a potential biomarker for patients who walks in, would you expect there to be many patients in your Phase 1 dose escalation trials who still have stem-like T cells? And is that the kind of thing that can be simply monitored from peripheral blood without needing biopsies?

Mark, so I think it's an interesting notion, which is not relevant only to PVRIG but to the field at all, the expansion of TCF1. I think it's might be challenging biomarker in terms of its prevalence, an expression in TME but we are looking in addition to the obvious aspects in DNAM pathway, et cetera. We are looking quite broad in terms of exploratory biomarker identification, obviously, TCF7 is one of them and peripheral blood, I'm not sure it's correlated enough the presence of the cells in blood compared to tumor.

Okay, fair enough. And just in terms of what we should be expecting as far as patient numbers for the upcoming readouts in -- from the triple combination dose escalation study and COM902 monotherapy dose escalation study and also might we see additional kind of translational biomarker data compared to archival biopsies for any of these patients or will the focus really be more on safety and efficacy in the fourth quarter?

So just, I'll let Henry relate to it but just for the translational data, we're accumulating data, we're working on translational data generation from our studies. When we will have the data, we'll present, not sure that this will be ready for the Q4 or maybe we'll have only very initial data. And as for the expectations, I'll let Henry relate to it.

Yes. So it'll be along the -- Mark, thank you very much for your question. It'll be along the same lines of the subjects that we were previously enrolled in dose escalation, so an approximate range would be between the strength of 15 to about 20 patients for the dose escalation.

And that's for both the triple combination and for COM902 as well?

Yes. That is correct, Mark.

The next question is from Chris Howerton of Jefferies.

Congratulations on the progress and look forward to the end of the year here. So maybe just a couple of questions from me. One is maybe just more of a theoretical question in terms of some of the checkpoint axes, obviously, the DNAM axis is the focus for your team, but there's also been some activity with VISTA in the associated macrophage lineage. So I guess I'd be curious to hear your thoughts on how those may or may not synergize or work together. I mean, then the second question I have is I know Anat, you were mentioning around the good strategy and the hypothesis is that you have around the Fc domain of your antibodies. There was a recent update from Arcus. Obviously, a little scant on details but curious to hear your thoughts as to how that reads through to your programs. Thank you.

For first -- for your first question, not sure what business specifically, but if you're talking about MIALU checkpoints in general, this is something we entered few years ago. Yes, we do think that MIALU plays an important role in immune suppression in tumor microenvironment and we have early programs on this -- in this front. And we are evaluating, of course, efficacy in combination, its synergies with current pipelines, assets, et cetera.

And for the --

Go ahead. Sorry.

Yes, no, I just wanted to add for the Fc domain question and the Arcus data. So obviously, we were not surprised to see this descriptive -- data positive descriptive data. We're waiting to see the data itself. But remember that this is an approach that we were taking ourselves and we believe in this approach, we selected an IgG4 purposely. So it was encouraging for us to say that we were expecting it.

Yes, okay. And then maybe if I may just one other quick question with respect to kind of the tumor microenvironment that you're describing like head and neck where it was inflamed but not necessarily responding to immunotherapy. I guess what is the mechanistic hypothesis there in terms of specifically what the checkpoints are and how is it inflamed, yet no appreciable immune response to the actual tumor?

So we're moving in few fronts, evaluating few different hypotheses, the -- as discussed in the script, the expression of PVRIG in stem-like T cells, the expression of PVRL2 in dendritic cells, all of this make us think that PVRIG may be a dominant checkpoint in terms of enhancing T cell priming, expansion and thereby enabling targeting also patients have less inflamed tumor microenvironment and we presented in ASCO preliminary data to show that we see some early signs of antitumor activity in patients, which have a PD-L1 low less inflamed tumor types. But obviously, every check when should work better in a more inflamed tumor microenvironment because the T cells are already there. So you might need to push it a bit less to get the efficacy and two most like head and neck and even other more inflamed might also have a potential for PVRIG, in terms of enhancing and expanding the T cells which are already these, releasing another break that might be the one which is relevant if PD-1 alone is not itself and head and neck, and non-small cell lung and other hotter indications have a quite dominant expression of the pathway of PVRL2 and others. So I think that we are testing in a clinical strategy, multiple indications, which fits different while hypotheses all stemming from the same biology of PVRIG and PVRL2 and the role in T cell expansion.

The next question is from Reni Benjamin of JMP Securities.

Congratulations on all the clinical progress. I guess just sticking with the TIGIT theme, we've seen some significant BD activity most recently, I guess with BMS and Agenus. And I just wanted to get your thoughts as to, I guess, how we should be thinking about your own BD discussions and your thoughts regarding the potential to partner or is there something that you really want to keep in-house? And sticking with the TIGIT theme, when I think about, I think, I know you had made some comments about how the strategy with having a inactive or slightly less functioning Fc effector prevent CD8 T cell depletion. I always thought about the Fc regions is mainly impacting NK cells. Is there something that I'm missing there or something you can help educate me on in regards to T cells and Fc receptors?

I'll relate to the business front and then Eran can take the Fc, NK relationship. So first, COM902 was developed at Compugen in order to ensure that we extract the full potential of COM701. So we're still on this path and we want to make sure that we remain competitive on this front and that we remain flexible owning the two arms of this DNAM axis. And as you know, we just started recently the COM701, COM902 study. Having said that, we're obviously open to discuss collaborative arrangements. This should allow us to continue extract the potential of COM701 and to generate some differentiation for our pipeline. So we're open to discuss collaborative arrangements spending [ph], it will allow us to keep what we need for our own pipeline.

About the -- And then just -- yes. So, but the role of Fc, so there are few assumptions to why preclinically people think the TIGIT is superior when it have the Fc binding. I would say maybe the leading one indeed relates in a way to enhancing NK activity, so when you use the antibody which have capability to bind Fc receptors, you enhance few mechanism that enhance, at the end of the day, depletion of the cells which carry the target TIGIT in this case. NK activation is a major one, especially in mice. By the way, in human tumor microenvironment, you can see less NK than in mice, but also other mechanisms and people again some claim that you can get the depletion of TIGIT positive cells and since TIGIT has high expression on regulatory T cells, this is desirable, however, TIGIT has also high and overlapping expression on the CD8 T cells, the same cells you want to unleash from the TIGIT inhibitory effects and if you're going to deplete anything, you have carried the risk also of depleting CD8 T-cells in addition to Tregs and that's why we and others chose to focus on our non-Fc binders to avoid the risk of depletion of effector CD8 T cells due to enhancement of NK activity or other mechanisms of T cell depletion inside tumor microenvironment. Does this answer your question?

Yes, that does. I guess just as a follow-up with 902, we're expecting the data from the monotherapy in the fourth quarter and you've already started the combination study, which -- what is the recommended dose for 902 or have you started at kind of the low end of the range when combining with 701 and/or escalating?

Yes, Reni, thank you very much for your question. Yes, we did start at the low end of the range for COM902 during dose escalation. The disclosure will identify what the recommended dose for expansion for COM902 will be.

Sorry, Henry, I just want to make sure. So in the fourth quarter, we'll know what the recommended...

That's correct.

Got you.

Yes, so the disclosure in the fourth quarter that Anat mentioned in her prepared remarks.

The next question is from Asthika Goonewardene of Truist Securities.

So Anat, you mentioned about the cut-offs that [indiscernible] PVRL2 and PVRIG expression as selection criteria to some of the expansion cohorts. Can you tell us what was the level of expression you are looking for inclusion? And can you maybe just clarify for me please which studies you're going to be applying this inclusion criteria?

So we are still studying it. We're following samples, we are following patients' correlation to response and this is exactly the goal of what we are trying to do to define in this stage retrospectively, which is the expression level cut-off, which could be percentage of PVRIG in the tumor microenvironment, it could be PVLR2 on the tumor cells on MIALU cells, I mean as other -- shown for other [indiscernible]. So we're looking at all of it correlating to response and of course, looking for the right cut-off. So this is yet to be defined and it is will be defined based on data and correlations.

Got it. And then also, I think at our last call, you mentioned -- you might have mentioned that you might be looking to present some more of the biomarker data perhaps at a conference later in 2021. Are you still on track for that or is this something we should expect more in 2022?

We didn't disclose when exactly we'll present the data. I can only comment that we are actively and aggressively pursuing sample acquisition and correlation and then we'll present the data in the future. Ideally, we'd like to present it in scientific conferences, we still didn't disclose when.

And then maybe just to tag on to Reni's question about Bristol [ph], but maybe approaching the other angle, so given that with -- given that Bristol may have acquired an asset that could have maybe some activity on PCRIG through the Agenus bispecific antibody deal ph , just want to get an idea of how involved and committed they are to doing the kind of data exploration with you or is it something that you are controlling?

So few things, indeed this licensing involve the TIGIT bispecific, the identity of the second arm is not known and it could be various other target. And so just to make sure that it is recognized as a product opportunity, a separate product opportunity than TIGIT. Having said that, I think it is pointing to the interest -- increased interest of BMS in the TIGIT space and that's the only thing that has shown on the behalf of BMS because obviously that's their decision-making and priority. I will tell you that and you know it's affecting the public domain, BMS is doing studies with their own TIGIT that is part of our studies as well, but they are having studies conducted on their own with their own TIGIT. And they are committed to the collaboration with say [ph], we're sponsoring obviously the study, the triplet study and pushing forward and that's critical for us because it's -- COM701 is the leading program in our pipeline and we want to push it forward and to have the leadership on the execution. But at the end of the day, there is investments on this program, on the TIGIT and no competition with the bispecific.

The next question is from Daina Graybosch of SVB Leerink.

Couple of questions from me. First, on the triplet, I wonder if you can clarify what cohort you're starting to enroll in the expansion. I think initially you had plan to enroll a tumor agnostic based on PVRL2 expression with Eran's comments just now, I wonder if you wouldn't to be ready to start to enroll that patient set yet. And so could you just clarify exactly which one do you start to enroll and you'll bring on other cohorts over time? And then with that how much is BMS involved in making these clinical development strategic decisions?

So for the second part of the question, I'll defer to Anat to answer. For the first part of the question, as you recall for the expansion cohort, we have cohorts of patients with ovarian cancer. So that includes patients with fallopian tube cancer, that includes patients with primary peritoneal cancer and patients with epithelial ovarian cancer. The other cohort of patients in the expansion cohorts are patients with relapsed endometrial cancer. So those cohorts open through enrollment and those are the ones that we're currently exploring right now.

And Daina for your question about decision making, we're making joint decisions with Bristol-Myers Squibb. Obviously, we are leading this program, we're looking at every bit and piece in the -- in this program, but we are making decisions in good collaboration with the Bristol-Myers Squibb. We have very good relationship, we keep an open line of communication and we're making decisions that are good for the program together. So this is how this collaboration is being conducted.

And then so I understand that you have ovarian and endometrial open. If I recall correctly, you had planned other cohorts in the expansion, are those follow or is this reflecting some strategic change as you've seen the dose escalation data?

No, no, no, there is no strategic change -- sorry. Henry, do you want to take it?

No, go ahead and I can clarify if any.

Okay. No, there is no strategic change. As we stated on -- as Henry said, in the ovarian, we decided that due to the low response rate and the days total data that we have, we can continue to relate to historical data, and we'll not do the randomization. We will add the head and neck as an inflamed tumor type to this and we'll continue with the endometrial that was planned in advance. So these are three changes with respect to the basket arm that we planned to do. This is still in planning what we were saying is that we are continuing to collect data that will correlate -- continue to correlate the PVRIG pathway with the treatment response and when we will have the data that can allow us to define the cut-off for enrollment, we will start this bucket arm. So there is no change in plans on this front.

And then on -- and then one more from me on TIGIT. I think that you guys starting back in 2009 have had your hands on a lot of different TIGIT antibody. I think you had a whole panel in your initial paper and of course, you work with the BMS, TIGIT as well as COM902. And you note the potential differentiating element of COM902 is a very high affinity. As you work up all these different TIGIT antibody, have you noted any other potential differentiating element? And I guess again asking the question, I wonder as we see your dose escalation data, you'll be the first Fc-silent TIGIT, and if you expect anything pre-clinically to show up clinically and what might we look for as we see this data in the second half or in fourth quarter?

So yes, we did the comparison of COM902 compared to most of the leading TIGIT antibodies. What we saw that indeed COM902 has a higher affinity, higher binding to effector T cells and has good or better activity in actual functional assay. So I don't believe we saw any major differences between the other antibodies. But just COM902 is a bit superior in that -- in this regard. I would say that is difficult to say at this stage, if any of this will translate to clinical observations. Yes, potentially in terms of dosing and PK, in terms of efficacy, it'll be difficult to say if there is any -- going to be any different from the other TIGIT assets that we've seen until now.

The next question is from Stephen Willey of Stifel.

Maybe just to follow up on the Bristol collaborative question, so is your ability to I guess pursuing a triplet regimen, whether it's I guess either with or without nivo using your TIGIT COM902, is that somehow constrained by the existing Bristol collaboration? I'm just kind of curious if you see an interesting signal emerging out of the triplet, how quickly can you start to think about swapping in 902 for the Bristol TIGIT?

So COM902 is totally unrelated to the Bristol collaboration. This is an independent asset. Discussing the triplet, it's a little bit more complicated because the -- under the collaboration that we have with Bristol-Myers Squibb, there is an exclusivity for COM701 plus PD-1 inhibitors for a fixed period of time. So we will need to take this thing to consideration, but COM902 is totally unrelated to the collaboration with Bristol-Myers Squibb.

Okay. So I guess your ability to pursue a triplet, that would include COM701, COM902 and a PD-1 inhibitor with -- overlap with the fact that you have exclusivity to nivo tethered to COM701, is that the correct way of --

There are some ways for us to do it with marketed PD-1 inhibitors. So that could be done.

And then I guess you've talked on the biomarker side about doing this I guess retrospective look, whether it's a function of evaluating PVRIG expression in the tumor microenvironment and/or PVRL2, I guess on various immune cell subsets. I know that Roche had kind of previously commented that I think in their -- in one of their TIGIT presentations that they didn't believe that PVR, which is the -- I guess one of the TIGIT ligands, wasn't a great surrogate biomarker in the sense that it's kind of broadly expressed on a variety of different cell types. Is PVRL2 akin to PVR in the sense that it has that same broad pattern of expression?

So PVR and PVRL2 in a way have similar pattern of expression in terms they do have expression also normal tissues. But both of them are upregulated in tumor microenvironment and are not -- in some tumor, you can find higher PVRL2 and some higher PVR. I think that for PVR, [indiscernible] also some publication may be due to its role also in enhancing other mechanisms for tumor invasive [ph], et cetera. It was published also in addition to Genentech data in relation to other trials but specifically TIGIT that it correlates with worse prognosis. So yes, what Genentech have shown for PVR that in non-small cell lung cancer, it didn't seem to have correlation since it is a biomarker, it doesn't suggest a lot about PVRL2. And we'll need to -- and we are evaluating it ourselves. So we don't know. But we're to think why PVRL2 will be different than PVR.

The next question is from Tony Butler of Roth Capital.

Three brief questions, and let me just state them because they all relate to one another. In ClinTrials.GOV on the triple, you've made some comments today regarding the basket cohort or that of cohort three, so am I to understand that, yet the notion of what defines high expression PVRL2 is not yet known? That's question one. And if it is any color would be great? Number two is, does PVRL2 high correlate with PD-L1 being positive or negative? And I guess a similar question may be. Are there tumors that are actually negative with respect to both markers? And then thirdly and this goes back to the SITC presentation which was quite helpful. But is there -- and this is a question that I think I hear a lot and it's never clear to me if there is a true correlation, but do peripheral effector memory CD8s and even NK-Ts, are they actually able to be found in the periphery equal to that of which is found intratumorally? That is to say, is there a direct correlation such that you don't really need the biopsy or may not need it in the future? Thank you very much.

So for the first question, so what is high PVRL2? So this is exactly what we're looking at, to define what is the cut-off, [indiscernible] score of 100 or 200 or percentage of immune cells expressing PVRL2 or other DNAM axis members, what is the cut-off that determines and correlates best with response? So this will be determined retrospectively and then tested clinically prospectively. This is for the question of the what is high PVLR2. About PVLR2 versus PD-L1, so this is a very important question and PVRL2 in contrast to PD-L1 is not induced by inflammation. And therefore, you can find high PVRL2 in PD-L1 positive but also in PD-L1 negative tumor types. And that's why we pre-identified, for example, ovarian and endometrial and other tumor types which sits in some of them the expression of PD-L1 is not very high. The response to PD-1 is not very high, but PVLR2 and in general, the PVRIG pathway is very dominant there and that's why we're trying to attack these kinds of tumor types. About the effector memory and NK-T, so the correlation between effector memory cells proliferating in periphery and specific effector clones proliferating inside tumor macroenvironment exits. I wouldn't say that we and the entire scientific community is in a place that we can rely solely on the peripheral measurements and that's why we have third biopsies in the expansion course of our trials to look inside tumor microenvironment but there is a correlation and in many cases, you see specific clones that are proliferating peripherally and they're also expanding inside tumor microenvironment. NK-T, you can find a bit less of those in tumor microenvironment but there are studies to show that during they are present, they are correlated to mostly to good prognosis, the correlation between peripheral and intratumoral expansion of NK-T, I wouldn't believe it was investigated much or it is, I'm not aware of any studies to correlate that, but just to mention, in addition to what it reflects about the tumor microenvironment, this kind of intense proliferation also supports that the drug COM701 is active and it does -- again this is preliminary data. But what it suggests that it does, what you think it should do, enhance immune cells activity, enhance interferon gamma. So this is another surrogate marker for drug activity in addition to the direct correlation to what is happening in tumor microenvironment. And maybe just to add that this is may -- why may be in many cases also for other drugs when you, especially in early trials has heavily pre-treated patients, you can find this kind of pharmacodynamic changes in the tumor, in the periphery. And they're not always correlated to response.

One follow-up if I may though, this is back to the second part of PVRL2 and PD-L1, but if you were able, post-therapy, if in fact you actually drive PD-L1 expression, even in PD-L1 minus tumors, when you look at a tumor initially having been treated with something else that is in fact relapse or a relapsed patient or a resistant patient, a PD-L1 minus patient and in fact then treated with COM701, for example, would drive PD-L1 expression. Is that true?

If indeed and as mentioned before, the biology of PVRIG and PVRL2, if we will be able to drive and expand this in tumor microenvironment then potentially, we'll increase the inflammatory status of tumor macroenvironment and then potentially PD-L1 could be upregulated as well as a marker to indicate and again this is one of the reasons PD-L1 is a biomarker for PD-L1 activity while it reflects [indiscernible] is more inflamed and has more T cells inside.

This concludes our Q&A session. I will now turn the call back to Compugen's President and CEO, Dr. Cohen-Dayag. Would you like to make your concluding statement?

Thank you. To conclude, we're excited about what's to come and are proud of our remarkable progress. As always, we will seek to continue to drive the science to uncover important insights into the biology and mechanism of our candidate and we remain uniquely positioned to unlock the promise of these potentially foundational immunotherapy candidates in the clinic across patient population typically considered unresponsive to current treatment. Thank you for joining us today and your continued support. Stay safe and healthy.

Thank you. This concludes Compugen Limited second quarter 2021 financial results conference call. Thank you for your participation. You may go ahead and disconnect.