We have Barbara, who's the President and CEO of the company, who's come speak to you today. She's gonna run you through a presentation, and then there, if there's time, we might do a little bit of Q&A, for which I'll bring the mic around. Over to you, Barbara.
Thank you very much. I think I'll say that the first half of my presentation is a lot of preclinical data and preclinical background for the program. I'll try and get through that relatively quickly just to make sure everyone's on the same page, 'cause I suspect what most people are interested in is the clinical data. I do wanna say, in case any of you are here, 'cause you think I might be disclosing our clinical data, I'm not, so if that's why you're here, just don't be disappointed at the end. All right, first, a little bit about Tango. Tango is a company that we built and launched in 2017 to address synthetic lethality as a big unmet need in cancer targeting, and this is now what our clinical pipeline looks like.
I'll spend the time today really only talking about our PRMT5 programs, but our CoREST program, which I think is a really interesting potential advance for treating patients with STK11 mutations with lung cancer who universally don't respond to checkpoint inhibitors, is also already in the clinic, and we will be shortly starting dosing patients with our USP1 inhibitor, which is a synthetic lethal with BRCA1/2 mutations and other HRD-positive defects, so another fairly big market target. So we have two PRMT5 inhibitors currently in the clinic. And just to give you a little bit of background on this, some of you may be familiar with the first-generation PRMT5 inhibitors, which were not synthetic lethal with what the key genetic alteration here is, which is MTAP deletion. All four of those programs have been withdrawn now.
However, there was clearly evidence of activity from previous mechanism, and molecules with the previous mechanism, just not with sufficient patient selection ability to get to a development plan that was feasible. I think the finding that there was another way to approach this came from the discovery back in 2015 from Novartis and the Broad Collaboration, that MTAP deletion created a vulnerability in a large number of tumors, 10%-15% of all solid tumors, to make them vulnerable to PRMT5 inhibition. But to do that requires a very specific mechanism of action for the PRMT5 inhibitors, which is the reason that the previous four PRMT5 inhibitors failed in the clinic.
This briefly outlines what that is, and that is that when MTAP is deleted in human cancers, and it's deleted as a homozygous loss, meaning both copies of the gene is completely gone, the MTAP protein, which normally breaks down this cofactor called MTA, goes away, and MTA levels get very high in just those tumor cells, but not in the surrounding normal cells. That MTA partially inhibits PRMT5 itself, so if you can make an inhibitor that takes advantage of the partial inhibition that's already there with MTA and inhibit the enzyme the rest of the way, you get a big therapeutic index that allows you to inhibit PRMT5 in normal- in tumor cells with an MTAP deletion, but not in wild-type cells, and that's really the key point, and mechanism behind all of the PRMT5 inhibitors that are currently in the clinic.
This is a schematic that shows what that looks like from the standpoint of the PRMT5 enzymatic pocket. So this is a cartoon. PRMT5 is a methyltransferase, just meaning it moves a methyl group from the donor, which is SAM, into a protein and uses that to regulate some function of that protein. That's what it looks like in a normal cell. The protein comes into the pocket, the SAM with the methyl donor comes into the pocket, and the enzyme moves that methyl group onto the protein. When MTAP is gone and there's these high levels of MTA, MTA really takes the place of SAM. And you can see from looking at this cartoon, that structurally it looks just like SAM, except that it doesn't have that methyl donor, so that's why it inhibits PRMT5.
It gets stuck in there, and now it can't act as a Trans Methyltransferase, and it leaves a little extra space where you can put an inhibitor. That's how we made these inhibitors, and that's how our competitors, Amgen and Mirati, made similar, mechanistically similar inhibitors that are also in the clinic. What this shows on this slide is that by doing that, you can differentially inhibit the enzyme in the MTAP-deleted versus wild-type cells, and I won't spend a lot of time on this, being a little bit more esoteric, because we've now shown data from the clinic that we can actually do this in patients. I'll show you that slide in just a few minutes. But you can see here two important points on the far right graph.
One is, even with vehicle-treated cells, when the cells are MTAP wild type, you can see just from the fact that those dots there, in the middle, the red dots, that's telling you that PRMT5 activity is only about half normal, and that's the way you get it all the way inhibited in those MTAPnull cells. And what that does is translate into some quite remarkable preclinical activity. These are all xenograft models that are MTA, sorry, MTAP-deleted xenografts. You can see here with TNG908, you can get deep, sustained regressions, and in fact, in some of these cases, on the far right, you can see in those glioblastoma and squamous cell lung cancer models, that you actually have cured these xenografts, because you can withdraw the drug, and the, and the tumors in the mice don't come back.
That's not very common for small molecule inhibitors as cancer therapeutics. One important thing about our program is that one of our molecules, TNG908, is blood-brain barrier penetrant. The other one, TNG462, is not. And these are the data showing why that could be important. I think probably most of you know that glioblastoma is a common and very treatment-refractory adult brain cancer, and I think the possibility that we may be able to have an impact in glioblastoma is suggested by these data. About half, up 40%-50% of glioblastomas have an MTAP deletion, so this is a population that could be quite relevant for moving forward with this molecule, TNG908.
You can see here, both with a flank xenograft and on the right, an orthotopic xenograft, meaning that xenograft has been actually put in the brain of these mice, that TNG908 has a very significant impact on these tumors, that we hope will translate in the clinic. In this case, showing that at least based on historical controls of the standard of care of Avastin and temozolomide, you get a very marked increase of the benefit of TNG908 versus those clinical molecules. So this is the design of our clinical study. We are currently still dose escalating with both molecules. TNG908 was started in the clinic about, well, in June-July of last summer, TNG462, July of this summer.
But the dose escalation of TNG462 has gone quite a bit faster for a number of reasons, so these two molecules are starting to get closer and closer to development timelines. Once we identify the phase 2 dose for TNG908, we'll move into a dose expansion that you see here. And I'll maybe at this point say that Mirati, in August, announced the first proof of concept for these MTA-cooperative PRMT5 inhibitors by showing that a range of different tumor types, you could... They, they identified partial responses in a range of different histologies, and that those responses have been quite sustained over time. They've presented really only data as vignettes, but we continue to hear that all of those patients that were described as partial responses back in August remain on study.
That's consistent with the data from the previous PRMT5 inhibitors, suggesting that in those patients who respond, they can actually stay on drug for quite a long time. J&J's data from the previous generation suggested patients even out 2-3 years with stable disease on some of these molecules. So where we are with 908, as I said, far along in dose escalation, we have not yet encountered a DLT and have really no tolerability issues, so we continue to dose escalate, and our guidance for this molecule is clinical data next year in 2024. What we did show back in May was this clinical proof of mechanism.
I mentioned that you can inhibit, differentially inhibit PRMT5 in normal cells versus MTAP-deleted cancer cells, and show that thus the mechanism, and therefore the therapeutic index that we expect with these molecules, was demonstrated clinically and really for the first time, that this was possible. At the time we showed these data, we were not yet in the therapeutic range, so we didn't expect that these would be active cohorts, but we can still show the proof of mechanism of the drug. So that's shown here on the next slide. SDMA is the PD marker for PRMT5 inhibition. It is an immunohistochemical assay that's measuring the immediate downstream methyl mark of PRMT5.
And again, you can see here across the top from cohorts 1 and 2, the normal tissue in dark blue, the tumor tissue in red, and you can see that at baseline, the normal tissue has higher SDMA staining, more SDMA staining than a tumor, but particularly marked in the cohort 2 ovarian cancer patient on the far right. You can see that you get very little change in the SDMA staining in the normal cells, with strong suppression of the staining in the tumor cells. So that's what's giving the therapeutic index there and showing that we can actually, with this mechanism, differentially inhibit the enzyme in normal versus tumor cells. What you can see actually on the next slide, is that you can do this in cells that are even immediately adjacent to each other.
So this is the same ovarian cancer patient, that in this case, her tumor has metastasized in her liver, and what you're looking at on the left-hand side, in brown stain, is the normal liver or the normal liver cells, and then adjacent to that, the cells with absence of staining, which makes them blue, the tumor cells. And you can actually see, if you look carefully, that there's a blood vessel on the far right-hand side, sort of lower quadrant on the right-hand side, of brown-staining cells that's running through the tumor. So it's clear that you can not only differentially inhibit PRMT5 in MTAP null versus wild-type cells in general, but you can do it even when those cells are in a tissue immediately adjacent to each other.
Which was important finding, because there were concerns people had before this, that those cells being adjacent to each other would either damage the normal cells or would dampen the anti-tumor response, either because MTA was leaching out, which it actually is not actively transported. It's a little small molecule that just passively diffuses in and out of cells. So imagine if there's a tumor cell with a lot of MTA in it, the MTA is following a gradient, does it actually go into the normal cells and make those normal cells more susceptible to PRMT5 inhibition, raising the concern that you damage a normal organ with these kinds of molecules, and this clearly shows that that's not the case.
Or does, in fact, that leaching out of the MTA lower the MTA levels in the tumor cell and prevent them from having the complete inhibition of the enzyme, and you can see also that that doesn't appear to be the case either. So, we, as I said, haven't released our clinical data yet, but Mirati and Amgen have both now shown data that their PRMT5 inhibitors have in the range of a 30% response rate across multiple tumor types, with no evidence of any histology, bias in the data that they've released, and that to date, and that data, the numbers are small, and the data are early, but those tend to be durable responses, and that those patients have stayed on study. Amgen actually reported a 100% partial response in a pancreatic cancer patient.
Presumably, that wasn't a complete response because there were other non-target lesions, but that pancreatic cancer patient stayed on for almost a year. Just by way of comparison here, 908, where again, I said we haven't released any clinical data, in preclinical data, is comparable to superior, or superior to the Mirati molecule in every xenograft that we've tried, and we've done quite a few. And I would also make the point that, in our opinion, TNG908, TNG462, the Mirati molecule, and the Amgen molecule are actually probably in honesty, more similar than they are different. They're structurally very different, but the mechanism is the same, and I think, the Amgen data and the Mirati data look very similar. So I think, you know, you can sort of, think about what that might mean going forward.
Now, 462 is our second molecule. The reason that we made this molecule and brought it into the clinic was because of this question of how big of a therapeutic index do you need? How much selectivity for MTAP-deleted cells do you need? So for 908, it's about 15-fold selective for MTAPnull cells. For 462, it's 45x, and I would say Mirati and Amgen are also in that sort of 40-45x selectivity for MTAP deletion. You can see what happens for the selective killing of cells as you increase the selectivity. So on the far right, the GSK compound is one of the first-generation molecules, not MTA-cooperative at all, no selectivity for MTAP deletion, and no ability to distinguish between the normal and tumor cells, which are equally distributed.
TNG908 in the middle, and then by the time you get to the 45x selectivity of TNG462, you get almost a complete perfect separation of the killing of the MTAPnull cells versus the wild-type cells. What that translates to when you move into xenografts is on this slide, which is that with TNG462, we get regressions in about 55% of the xenografts that we've tested, and some of those are complete and durable regressions, complete regressions. For TNG908, that number is about 30%. Now, again, it's all extrapolating and correlating preclinical data, but TNG908 and Mirati are pretty similar preclinically. As I said, TNG908 is as good or better than Mirati in any xenograft model.
Our estimate is that translates into about a 30% regression rate in xenografts, and as it turns out, that's the approximate response rate in the clinic. So we'll see what happens as we move forward with our molecules, but what you get from the higher selectivity of 462 is a higher percentage of regressions in the animal models. We'll see if that translates. So just to summarize, as I mentioned earlier, these two molecules, although they went into the clinic about a year apart, are starting to catch up with each other. The first one, 908, I think, as we started to do these trials, investigators really were used to thinking about MTAP deletion as a way to select patients, so the initial patient enrollment was a bit slower.
We also were not allowed by the FDA to dose double in our initial expansion cohort. That's changed now with 462, so we're dose doubling. So between investigator enthusiasm and the ability now to start at a higher dose with 462 and dose double, those programs are getting closer together. And as I mentioned, our guidance for 908 is 2024. We haven't given specific guidance for 462, but they're starting to converge. So maybe I'll stop there and see if people have specific questions about the program.
If you have a question, just raise your hand and I'll bring it around.
Hi, I saw your initial slide. You also had a MAT2 or a MAT2A inhibitor. Would those work together? Would you be able to combine them with PRMT5? I think somebody is doing that, and what would you think would happen to toxicity in that case?
So we don't have a MAT2A inhibitor, but you're correct, IDEAYA has a MAT2A inhibitor that they've started recently in a trial, combining with the Amgen PRMT5 inhibitor. What MAT2A inhibitors do is reduce SAM levels in cells, but they reduce it equally in normal and MTAP-deleted cells. So the idea that they have is that if you widen the gap between M, MTA level and SAM level by lowering SAM, you increase further the potential benefit. I think the way we think about this. It's not in this deck, but it's on our website, on our corporate deck. If you have a really good PRMT5 inhibitor, and you can fully inhibit the PRMT5 enzyme with the single PRMT5 inhibitor, you don't get added benefit of adding a MAT2A inhibitor.
However, MAT2A inhibitors by themselves don't have sufficient activity to do that, so the reverse isn't true. If you have a MAT2A inhibitor, to get full activity, you have to add an MTA cooperative PRMT5 inhibitor. We're very interested to see what the results of those data look like. We believe the synergy is real, but to see it with our molecules, you have to dramatically reduce the dose of our molecule to see that synergy. So we'll see what happens.
Yeah, hello. So, perhaps I missed this, so the selectivity of PRMT5 inhibitor, how are you getting the selectivity to more specificity?
The selectivity.
Is it coming from the MTAP, deletion?
The selectivity actually is an inherent molecular characteristic of these inhibitors, that the medicinal chemists can actually build into the molecule, and it has to do with how tightly they bind in cooperating with MTA. So it's an experimental finding of once, you know, how much you get out of any given structure. Does that, is that the question you're asking?
I think perhaps we can follow up, but.
Yeah, sure.
Congratulations. Yeah.
Thank you.
I think we have time for, like, one or two more questions, maybe.
Thank you. Maybe if you could just talk more about the amount of 908 data you could show in 2024 and the types of patients that you're enrolling?
I mean, I think that with the proof of concept in humans out there now, from Amgen and Mirati, I think what we'd like to be able to show when we show our data are, one, what 908 looks like in glioblastoma. We are enrolling glioblastoma patients in the dose escalation phase of our study, not in the escalation cohorts, but in backfill cohorts, so we will, when we release data, have GBM data. And I think we would like to be able to see and say, you know, what are we thinking about with 908 versus 462? So we'd like to...
You know, we haven't actually committed to doing them at the same time, but we'd like to be able to be far enough along with both molecules to have an idea what our development plans will be when we talk about it.
I think that's all we have time for, so thanks a lot, Barbara, for the presentation.