All right. Good afternoon, everyone. Final day of the Leerink Partners Global Healthcare Conference. Thanks for joining us. I'm Andy Berens, Senior Biotech Analyst. We're really excited to have Tango Therapeutics here with us. We have Barbara Weber, the CEO. Thank you, Barbara, for joining us.
Thanks for the invitation.
Great. Why don't we start with an overview of your company for those who may not be familiar with Tango?
Tango was established as a synthetic lethal company, and the first target that we selected out of that was what is now our lead molecules, PRMT5. I think the focus now remains solidly on moving those molecules forward in development. We have also one other clinical asset in development behind that. It's a novel program that's focusing on STK11 mutant lung cancers.
Okay. The class, the PRMT5s, are a number in development in the clinic, and there's a lot of enthusiasm. Can you just walk us through, I guess, the physiology of dosing, that pathway, and what you guys are doing relative to what, I guess, Bristol Myers Squibb and Amgen are doing?
Yeah. PRMT5 is a really interesting cancer target, and it's quite different from many of the cancer targets that we're used to in the past, which have been oncogenes that are activated in cancer. PRMT5 is an essential gene, meaning that if you inhibit it in any cell, it will die. How this comes to play with the synthetic lethal story is that it was discovered a number of years ago now that there's a deletion in a gene called MTAP that gives you a big therapeutic index for PRMT5. Those tumors, and this is where we're going with our development of PRMT5 inhibitors, in those tumors that have lost MTAP, you can dose the PRMT5 inhibitor up much higher than you can in the absence of a PRMT5 deletion, and that gives you the ability to kill the cancer cells without unduly harming the normal cells.
That is really the biology behind the PRMT5 inhibitors.
Okay. You are talking about an MTA-cooperative approach, the first-generation drugs that did not use that. What type of toxicities did they run into?
Right. In the absence of this MTA-cooperative mechanism, which is what takes advantage of the MTAP deletion, you are hitting normal cells to the same extent as cancer cells. What you see as dose-limiting toxicity is bone marrow toxicity, particularly anemia and thrombocytopenia. That is backed up by the biology in terms of knockout mice and what we see as well in our GLP-tox studies. The on-target toxicity is bone marrow toxicity.
Okay. Which tumors are MTAP deletion relevant?
MTAP deletion is a very common deletion across all human cancers, but the biggest patient populations with MTAP deletion are non-small cell lung cancer and pancreatic cancer. Particularly our primary focus, or our initial focus, I should say, in pancreatic cancer, probably about a third of all pancreatic cancer patients have a tumor that has an MTAP deletion. There are quite a number of other important tumors where you also will find MTAP deletions. Glioblastoma is a big one. Almost half of all glioblastomas have MTAP deletions. I mentioned lung cancer, probably around 15%-20% of lung cancers will have MTAP deletions, and quite a number of additional smaller subgroups, but important tumors that have MTAP deletions.
Okay. You guys are in the clinic, as well as Amgen, as well as Bristol Myers Squibb with their drug, and that was a big component of the deal that Bristol did. What do we know, or what has been shown for the class? There was quite a bit of enthusiasm, especially around the time that Bristol acquired Mirati. I felt like the updates from a number of the clinical programs, kind of the efficacy went down a little bit, and there seems to be less enthusiasm now. There are also other companies like IDEAYA that has a program. BeiGene has one. There are a number. In the industry, it seems to be people very enthusiastic, but I feel like the investor community seems to have lost the same level of enthusiasm.
Yeah, no disagreement there. I think that it's unjustified, though, and I think we and others will show why that lack of enthusiasm for investors isn't justified. I think the initial concerns have been around overall response rates and why aren't they better with PRMT5 inhibitors. I think the answer is that the real bar for measuring PRMT5 inhibitor success is around durability, less so overall response rate. The reason for that can be seen really in all of the data that have been released, our early data and the data from Amgen and Bristol Myers Squibb, which is that the benefit to the patients is almost the same in the patients with stable disease as in those with partial responses.
As we know, that's generally not been the case, for example, with TKIs, where ORR is a good measure of what will ultimately be success as a commercial success, whereas with these PRMT5 inhibitors, we see patients not infrequently at all that have minimal tumor shrinkage but stay on for very long periods of time. I'll just cite one that we've talked about publicly, a pancreatic cancer patient with stable disease that's been on for well over a year. Really the story is durability, and that takes time to develop. Two things investors don't like: ORR, which ultimately won't be the right thing to be looking at, and time it takes to develop those data.
What did the animal models predict in terms of apoptosis versus stable disease?
Yeah, so I think what you're getting at is the sort of cytostatic, cytotoxic mechanism. I don't know that I think that maybe and maybe always has been a bit of an oversimplification. In the animal models, the tumors shrink and go away. In the patients, they really shrink, right? They don't just sit there. What actually happens in any tumor when it reaches best overall response, then just sits there and then eventually recurs is not really well described. That's what's seen with basically all oncology drugs, right? I would say there is apoptosis, they are cytotoxic, and it's just a slower mechanism to maximum tumor response with these.
Okay. Yeah, no, I mean, we were very excited about the durability, especially what Bristol Myers Squibb showed in lung cancers over 10 months, which is much better than some of the targeted agents we've shown.
Absolutely.
There was this hypothesis that I know, I mean, we've discussed it, but that you needed to also have an MTAP, a MAT2A agent to block the shift from MTA to SAM. Where are you guys with that hypothesis that you need to have a MAT2A as part of the equation to get the durability? Because we certainly have seen long durability with a monotherapy.
Absolutely. I think there's no question that biologically there is synergy between PRMT5 inhibition and MAT2A inhibition, but it's inhibition against the same target. What you're doing when you add a MAT2A inhibitor is you're potentially allowing further inhibition of PRMT5 as opposed to, say, our collaboration with RevMed, where you're inhibiting two orthogonal pathways. You're getting at the same enzyme target, which means that you can do it one of two ways. You can use that synergy and reduce the doses of both molecules to get the same effect, or you can have a really good PRMT5 inhibitor that gets that same level of PRMT5 inhibition as a single agent, which actually simplifies development quite a bit.
Okay. It was my understanding that the hypothesis was that there would be more SAM produced, that the tumor could shift from MTA production to SAM, and therefore the drugs that are MTA cooperative wouldn't be able to bind as well. Is that no longer the thought behind?
That was a biological hypothesis. We have not seen that effect either preclinically or, well, I should not say clinically because it is very difficult to measure clinically. We have not seen evidence that that is the case. Again, I do not want to completely discount the value of PRMT5 plus MAT2A inhibition. We are still waiting to see what that looks like, but we feel pretty good about our durability with single agent 462.
Right. Okay. Yeah, why don't we talk about the different programs? You guys have had several PRMT5s in the clinic. Why don't we go through what you've shown thus far and what the plans are going forward?
Sure. If I think about our PRMT5 inhibitors, I would say there are two characteristics that we've really focused on to some extent separately. One is the potency and selectivity that allows you to get maximum PRMT5 inhibition. The other is brain penetrance that allows you to get at this really important glioblastoma patient population. Our first molecule, TNG908, was somewhat potent and selective, but brain penetrant. Our second one is very potent and selective, but not brain penetrant. I think it's really, for us, it was a decision to say, is TNG908 going to work in glioblastoma? Turns out it doesn't because you just can't get enough drug in because of the potency and selectivity. TNG462 looks great. I would say that the potency and selectivity there is giving us both an efficacy and a tolerability profile that makes it potentially best in class.
We then are going forward, now we're going forward with TNG456. I know these numbers are unfortunately closely related to each other, but the molecules are really, the differentiation here is the brain penetrance. Our complete focus on 456 is whether it will in fact be active in glioblastoma.
Okay. What about other tumors that maybe they're not primary in the CNS, but they tend to metastasize to the CNS? Couldn't there be a reason to use a brain penetrant PRMT5 inhibitor there to be either empiric or treat patients that already have?
There absolutely could well be. I think the second thing you said, treating active brain metastases is probably the first way that we are thinking about this. Could you in fact prophylactically get better utility out of a molecule that has brain penetrance upfront for tumors like lung cancer, which are commonly producing brain metastases? Maybe, but it's a double-edged sword also because then you usually get CNS side effects that go along with that. Now you're giving a drug that may be less tolerable to a lot of patients who will never develop brain mets. The way we're thinking about it again right now, very straightforward, 462 for solid tumors that are not in the CNS and 456 for CNS tumors, whether they're primary glioblastoma or they're brain metastases. The rest of it can be sort of sorted out over time.
Okay. What type of liability do you see by penetrating the CNS? Is it hitting like track or what's the side effect that's?
You know, I don't know that it's ever been sorted out at that molecular level, but almost all drugs that have CNS penetrance have this kind of side effects that patients will describe as brain fog, dizziness, memory lapses, word-finding difference. You're just like putting drug into normal brain cells that doesn't do good things for how your brain cells function.
Okay. Of the tumors that you highlighted earlier, which ones do you think would not benefit necessarily from CNS penetration?
Pancreatic cancer rarely metastasizes to brain. Lung does maybe, what, 30 or so % of the time. I think pancreatic cancer for sure. I do not see a scenario where we would be wanting to use a brain penetrant molecule. A number of the other GI tumors, for example, gastric, esophageal, cholangiocarcinoma, those are all tumors where we have good activity and may explore in the future, but those rarely go to brain.
Okay. So that.
It's really about lung is the big one.
Okay. Lung might benefit from one that's potent, but also CNS penetrant, like some of the small molecules that have been successful now.
Right.
Especially if you're going into like the adjuvant setting where.
Absolutely. I think that's where it really might make the most difference. That may be very much where it would be useful to have that as an upfront treatment where you're truly preventing the development of brain metastases or treating micrometastases.
Right. Okay. Why don't we go through the last update? I think we thought it was encouraging, especially relative to your competitors. I think investors were looking for maybe some lung data. I guess, why was there no lung data? When can we see the larger opportunity that might get people reinvigorated?
Yeah, I agree. I mean, we were hoping that what we could do was to show that sort of some early signals that were encouraging. That's certainly how we viewed them. Obviously, it wasn't how the market viewed it. The short answer is we didn't show lung data or pancreatic cancer data because we had very few patients that had been enrolled so far in the trial. What we did show was the one tumor type where we had enough patients where we could compare directly to our competitors, and that was cholangiocarcinoma. Not because we see this, unfortunately, in the post-IRA world as a path forward, at least initially for development, but because we had enough patients to say, "Look at our response rate in cholangiocarcinoma versus the competitors and see that it's considerably better," right?
Saying that at the least our molecules are as good, if not better, than the competition. We hear loud and clearly that what we need to do for our next data update is have a fulsome data update that includes a significant number of lung and pancreatic cancer patients. That is what we're driving to this year, an update with pancreatic cancer patients with at least 20 pancreatic cancer patients that have been followed for six months and a similar data set with lung cancer.
Okay. In terms of the line of therapy that these patients would be, what are you looking for in terms of the response rates and maybe some durability that would get people excited? I guess also, what have the competitors shown? Because that'll give us some idea of the bar too.
It's a really good question. Maybe I'll start with what's the bar for us to launch a pivotal study, because that's the key point here, right? That is that the standard of care in second-line pancreatic cancer and second-line lung cancer is somewhere in the range of 12-14 weeks of PFS on the current standard of care chemotherapy. We are setting the bar for ourselves to be 18-24 weeks. That gives us a good range where we're above that, we think we can beat the second-line standard of care chemotherapy.
That would be docetaxel in lung?
Docetaxel in lung and a mixture of gemcitabine or 5-FU-containing regimens for pancreatic cancer. That is the bar for us. Now, in terms of what our competitors have shown and what we have shown, our data update was really focused on our dose escalation cohort. I think it is really important to go back and think just how strong those data were. That was a subset of patients where the patients who were treated at active doses were a mixture of late-line, hard-to-treat cancers, cholangiocarcinoma, esophageal cancer, like really third-line-plus patients. In that patient population, we saw a median PFS of 24 weeks. That is pretty spectacular, right? If you think about phase one dose escalation, 24 weeks for these late-line, hard-to-treat cancers, that is telling you that this is going to be really something important going forward.
Right. Okay. To get more of the lung and pancreatic patients, I guess it was competitive with some of the other programs, some of your bigger competitors, Amgen and Bristol. Have you expanded the trial sites to get some of these patients?
It was less about expanding the trial sites rather than just making a very concerted effort to get those two histologies. Pancreatic is easier and has been very fast. I think we feel, you probably imagine that I wouldn't be saying patient numbers if I didn't already know that we had that to deliver. I think that's something that's been very fast. The lung cancer patients just take more effort because there's more competition. I don't know that it's whether it's Bristol Myers Squibb or Amgen and PRMT5 or just the range of trial options for lung cancer patients. It takes longer, but we're making good progress. I feel confident we'll have that as well.
Okay. That's by the end of this year?
Mm-hmm.
Okay. That's going to be an important update for sure. I guess, and then in terms of the Glio program, that sounds like where you're going to focus on 456, at least for proof of concept?
Correct. And that's really the gating data points for 456. If it doesn't work in GBM, then we focus those resources fully back on 462.
Okay. How much more potent is that agent than 908?
It's quite a bit more potent, but probably even the more important thing is that it's much more selective. TNG908 is 15x selective and 456 is 55x selective. It's really that selectivity that's going to, even if the brain penetrance is the same for those two molecules, that's the molecular feature that will allow us to get enough drug into the brain to be over what we know to be the activity exposure threshold that we've observed from 462 and 908.
Okay. Now you guys have already announced some combination strategies. Can you walk us through the drugs you've chosen and the rationale behind those decisions?
There are two really important categories of combinations that we're focusing on. The really important one to me strategically and for patients is the combination with the RevMed RAS inhibitors. I think that combination of 462 plus one or the other of the RAS inhibitors, the pan-RAS or the G12D, really have the potential to change the face of pancreatic cancer. The other combinations for us are tolerability studies for first-line registration trials. We need to make sure that you can combine 462 with pembrolizumab because that's a mainstay of first-line lung cancer treatment and with the other chemotherapy regimens that are first-line for either lung or pancreatic cancer. Those are more enabling combination studies to ensure that we can give those things together. The real exciting data will come from the RevMed combinations.
Okay. Is this going to be a synergistic? Is there some physiological rationale for there to be synergy or will it be mostly additive? What's the interplay between the two pathways?
It is the inhibition of two very important orthogonal pathways added together rather than the synergy, for example, that you see with MAT2A and PRMT5. I think what we see preclinically and we certainly hope to see clinically is that you can get good responses and very good durability with single-agent 462 or single-agent 6236 or 9805, put them together and you get even better responses and durability. That's the goal.
Okay. The mutations are they're overlapping, obviously.
Yes. I think you can, let's focus on pancreatic cancer to start with. You can make the valid assumption that almost all, if not all, pancreatic cancers have a RAS mutation. There's a few that don't, but by and large, they all do. About 35% of pancreatic cancers have an MTAP deletion. We're looking at that 35% of pancreatic cancer that has an MTAP deletion, and within that, they all have RAS mutations. The way our trial will go is that all G12D mutant patients will go on the 462-9805 combination, and the other RAS mutants will go on to the 6236 combination.
I see. Okay. To demonstrate contribution of parts, incremental benefit, what are you looking for in the different tumor types?
I think in the initial signal-seeking study, the phase one, the dose escalation and expansion will have to be looking at historical data, both ours and RevMed's, to see whether or not there's a significant difference from what the single agents do. What the FDA asks us to do later, we'll have to sort of sort out.
Right. Okay. And then the combination with checkpoint, are these tumors that tend to be immunosensitive?
They generally are. Getting actually to our next program, in lung cancer, the lung cancers that tend to have primary resistance to pembrolizumab or other checkpoint inhibitors are the STK11 mutants. I would say only 10-15% of MTAP-deleted cancers have an STK11 mutation. For the most part, we're talking about pembrolizumab-sensitive tumors.
Okay. That would enable you, I mean, that would enable you eventually to get into the front line by combining with KEYTRUDA versus trying to compete with it.
Correct.
Okay. That would be in PD-L1, low and intermediate or high, or what would you try to enroll for that, you think?
I think at this point, it's largely all comers, although we would probably exclude the STK11 mutant patients.
Okay. Do you think that the combination could resensitize patients to a checkpoint inhibitor, maybe?
The 462 combination?
Yeah.
I don't think so. I think that those that have primary resistance to a checkpoint inhibitor would stay that way. It's really the 260 program that has the potential to sensitize patients who have a genetic alteration that's making them primarily resistant to checkpoint inhibitors.
Okay. You are also testing it with a CDK4/6 inhibitor, right?
Yep. That will be with the 456 in the glioblastoma study. The rationale for that is both that all of the patients who have an MTAP deletion also have a CDKN2A deletion, and there's preclinical data to suggest that combination can definitely have combination benefit. There's also a clinical study showing with abemaciclib, which is the only commercially available CDK4/6 inhibitor that's brain penetrant, that there's a PFS advantage in glioblastoma with single-agent abemaciclib. Our thinking is to say, we know that all of our eligible GBM patients will have a CDKN2A deletion, and there's potential benefit from single-agent abema, and put those together.
Okay. Interesting. Now, that'll be, I mean, we're obviously very interested in the CDK class of drugs, and it'll be interesting to see it expand more broadly. Okay. Why don't we move to 260? And you touched on it a bit. Can you just give us a summary of that program and what you're planning to do in relation to the checkpoint inhibitors?
That is such an interesting program, really, from the beginning. The idea goes all the way back to our sort of a twist on our synthetic lethal approach, which is that we said, can we identify genetic features of cancers that make them sensitive, sorry, make them resistant to immune cell killing? Can we then find a target that we can hit that would reverse that? In an independent stream from what was happening clinically that we were not aware of, we discovered STK11 mutation in the lab as a genetic alteration that made tumors insensitive to checkpoint inhibition. We discovered the target of 260, which is the CoREST complex, as the target that can reverse that.
At the same time, a consortium of academic investigators, primarily from MD Anderson, Dana-Farber, and Memorial, were showing that STK11 mutations looked at in retrospective analyses of patients treated with checkpoint inhibitors was a predictor for primary resistance. Those two things came together kind of at the same time, validating each other, one from the biology side and one from the clinical side. That paper was just recently published. Now the next step was to do the clinical trial we're in the midst of, which is to show that in those patients which are STK11 mutant, that you can treat them with 260 plus pembro or a checkpoint inhibitor. Our study is pembro. You can show that those patients who historically have not been sensitive to checkpoint inhibition are now sensitive.
Okay. When you say they're not sensitive, they get no responses or no durability at all?
Very little. Very little. The overall response rate is 10% or less, and the PFS is two to three months.
Okay. Where is that program now? Your dose expansion?
Right. We've completed the dose escalation. We've selected the dose going forward, which is 80 milligrams a day. We're now expanding in the lung cancer cohort to look at the efficacy signal there. Those data will be available, we think, by the end of this year.
Okay. The guidance is broadly for 2025? Is it more likely the second half, or?
I think that's the most likely scenario, but we haven't really given full guidance there.
Okay. That'll be interesting. I don't think that's on a lot of people's radar. I mean, PRMT5 clearly is, but this one's kind of been under the radar for.
Absolutely.
Okay.
I think it's a really interesting program, and it's 20% of lung cancer, so it's something to watch.
Okay. You will be looking for when we see the dose expansion patients, how many patients do you think we will have and what is going to be a sign that you are successful, like more than a 10% response rate?
I would say a couple of things. First of all, these are going to be smaller numbers of patients than the PRMT5 dataset. That's a substantial dataset. This will be early signals coming from smaller, probably numbers of patients with less follow-up. We'll just have to see what that looks like.
Okay. TNG260 is an HDAC inhibitor, right?
It's targeting the CoREST complex, which contains HDAC, but it doesn't inhibit most of the other HDACs.
Okay. All right. What safety is worth looking for with that agent when we look at the therapeutic window?
Yeah. To be fair, then what does end up being the dose-limiting toxicity is what you would expect from an HDAC1 inhibition, which is bone marrow toxicity and some GI symptoms. That was our dose-limiting toxicity, bone marrow.
Okay. Great. Let me see if there are any questions from the audience. Any questions for Barbara? All right.
I think everybody wants to go to the beach.
I think so. I think some people are already there. Thank you.
Thanks very much.
Congrats on all the progress, and you're going to have big data here, so look forward to seeing the updates.
Thanks very much.
Thank you, Barbara. Thank you, everyone.