Okay, great. Well, thanks everybody for joining us here on the third and final day of Piper Sandler's annual healthcare conference. I'm Joe Catanzaro, one of Piper's biotech analysts. It's my pleasure to welcome to this session Tango Therapeutics; joining us is their President of R&D, Adam Crystal. Adam, thanks so much for joining us. Maybe before we jump into Q&A, I could sort of give you the opportunity over a minute or two to sort of introduce Tango to maybe those who aren't familiar and let them know what you've been up to and what we have to look forward to.
Absolutely, thank you. So Tango Therapeutics is a biotechnology focused on finding therapies by leveraging synthetic lethality to discover novel targets in oncology. It was founded about eight years ago, really timing-wise empowered by the advent of high-throughput CRISPR screening, through which Tango and others have walked the course of identifying novel synthetic lethal genes which can be targeted pharmacologically. Over that time, the company has gotten to the point we are now, which is with two molecules in the clinic, soon to be three, and a lead program which we have released some public data on and demonstrated what we believe to be a proof of concept and are moving towards pivotal studies. In addition to that, we have a second program we could also talk about, named TNG260, which is a molecule in the immuno-oncologic space. From there, I'll ask you to guide me.
Yeah, that was perfect. Maybe I'll start with just sort of a high-level sentiment question. So you mentioned synthetic lethality. It's a space kind of near and dear to my heart. Just sort of where you think sort of sentiment is there over the sort of various data sets, not maybe just with PRMT5 inhibitors, but sort of more generally and what CRISPR has allowed the space to do.
Yeah, so I start with synthetic lethality and thinking about BRCA and PARP inhibitors, which really provide very strong clinical proof of concept for the space. I think that there is space to discover additional clinically transformative targets based on the synthetic lethal principles. There have been other successes. Immediately I point to now the class of emerging molecules of MTA-cooperative PRMT5 inhibitors. Our lead asset is indeed based on a synthetic lethal interaction. We are extraordinarily excited about the proof of concept the field has really demonstrated. Synthetic lethality as a whole is no different than any other space. In my view, it really comes down to the target. There will be synthetic lethal targets that are incredibly powerful, like PARP inhibitors.
There will be others that may have less effect size and lack the ability to really be impactful to patients in the treatment of cancer. We have discovered novel targets, which we are advancing in our pipeline, led by PRMT5, which emerged from the pipeline, followed by CoREST and others that we believe have the space to be as impactful as PRMT5 does at the moment.
Great. So maybe with that, that's a good segue. We could talk about PRMT5 and your PRMT5 programs. But maybe first you could sort of give a little bit of an intro around PRMT5, MTAP biology, and what the next generation of PRMT5 inhibitors, MTA-cooperative, are looking to do and overcome what prior PRMT5 inhibitors sort of failed to achieve.
Yeah. You know, the PRMT5 inhibitor story is really simple once you wrap your head around it, but there's a little bit of sort of mental gymnastics that you have to do the first time you think about the pathway. PRMT5 is a cell essential gene, which is to say if you knock it out, if you fully inhibit it in any cell, that cell will die. So you can make a PRMT5 inhibitor, but it does not have therapeutic index. It kills cancer cells and it kills normal cells, which is to say it's no better than bleach. Our lead program is an MTA cooperative PRMT5 inhibitor, which leverages a genetic lesion of MTAP deletion to create therapeutic index. What our molecule does is the same mechanism which the Mirati molecule, MRTX1719, and the Amgen molecule have done, take advantage of the accumulation of MTA.
Our molecule effectively only binds to and inhibits PRMT5 when MTA has accumulated. That only happens in MTAP-deleted tumor cells, meaning when you give a patient with an MTAP-deleted tumor cell a molecule like TNG462, it specifically inhibits PRMT5 in the tumor while sparing normal tissue, creating that all-important therapeutic index. Now, what patients have MTAP deletion? The short answer is that it's a lot. There are probably 50,000 patients a year with metastatic cancers in the U.S. that are MTAP-deleted. That includes about 30% of pancreatic cancer and about 15% of non-small cell lung cancer, as well as 40%-50% of glioblastoma.
So you mentioned Amgen and Bristol Mirati programs. We saw updates from both of those programs just a couple of months ago back in the fall. Maybe before diving into 462 and your programs, we could sort of look at those data sets and sort of speak to what you see as the most important takeaways within those respective data sets for those programs and maybe, again, the field more broadly.
Absolutely. You know, even before doing that, I'll go back to what was really the first generation of PRMT5 inhibitors. There was a GSK molecule, there was a BMS molecule, and others. And they worked the same way by inhibiting PRMT5, but not through leveraging MTA, which is to say they didn't have therapeutic index. Those molecules were active, some tumors shrunk, but there was too much toxicity. And patients did not tolerate the drug for long periods of time because of on-target toxicities of thrombocytopenia, anemia, and fatigue. In that setting, it was difficult to understand how efficacious the molecules could be because they couldn't be dosed high enough or for long enough. In contrast, with BMS and Amgen, what we're seeing is very exciting. They have demonstrated clear efficacy across different histologies, which are MTAP-deleted.
That includes critically non-small cell lung cancer, pancreatic cancer, biliary tract cancers, and others. And as a class, they are very well tolerated. The on-target dose-limiting toxicities are hematologic, thrombocytopenia, and anemia. And the safety profiles of those molecules overall are good, but leave room for improvement. What we've seen across them is good activity. And more importantly than a fair number of patients achieving partial responses is a durability of treatment. So with both of those molecules from the data sets released, we can see patients staying on for a long period of time, sometimes up to or over a year, with continuous shrinkage over that time. And we think that's an important feature of this class of molecules.
So you mentioned sort of dose-limiting toxicities. I think you guys have spoken to how you think it's notable that both the Amgen and Bristol programs were not dose-limited by myelosuppression, rather other things. Why is that important to you? What does it sort of say about what's going on?
Yeah, it's a critically important point, so our molecules are selected for MTAP-deleted tumors, but not perfectly. They all have a range of about 45%-65% fold preference for deleted over wild-type tissue, which is to say if you dose high enough, you'll eventually suppress the marrow and cause anemia and thrombocytopenia. We saw a little bit of anemia and thrombocytopenia with our first molecule, TNG908, but before we could really hit the marrow hard, we hit some other toxicities that forced us to stop dose escalating, which were a CK elevation in one patient and a change in mental status in another, which means that we could not escalate all the way up to take advantage of the full selectivity of the molecule because of this non-specific toxicity. That's also what Amgen is seeing very clearly, with nausea, vomiting, and diarrhea being dose-limiting.
They escalate to a point and can escalate no higher because of those non-specific toxicities, so they can't fully take advantage of their 45X selectivity index. With TNG462, we saw exactly the opposite. We escalated from a starting dose of 20 all the way up to 300 without any dose-limiting toxicities. At 300, we were very deeply suppressing the target, both by PD assessments and by efficacy assessments, frankly, and when we got to 600, the single DLT which we saw was thrombocytopenia with a nose bleed. We looked at the data, we saw the rates of cytopenia, and we expanded two doses of 200 and 300, knowing that at those dose levels, we were achieving the maximum inhibition of PRMT5 in tumor as possible. That's very different than the other molecules. They can't go all the way up because they're limited by other things.
So maybe you answered this a little bit in your response there, and it's sort of a two-parter. But when you look at those clinical data sets and compare the two for Amgen and Bristol, what does it tell you about how preclinical potency and selectivity translates into the clinic? And the second part of that is how does 462 preclinical profile sort of stack up and compare to those two programs?
You know, I would actually add something to that, which is PK. And listen, it shouldn't be surprising that potency, selectivity, and pharmacokinetic properties are critical for a small molecule oncology. It's no different for this class. I would say that TNG462, Amgen, and Mirati all have very similar selectivities. TNG462 is the most potent, but not by a heck of a lot. What we have learned from those molecules is that selectivity and potency matter. Selectivity here can be thought of as a proxy for toxicity. And potency is really the extent to which you are covering target and resulting in efficacy. With that potency and selectivity come tolerability and efficacy. We've seen that very clearly. We even saw with TNG908, that molecule in many ways looks very much like Amgen from a PK perspective, as well as the overall activity, as you could assess by a waterfall.
That's because their PK is very similar. At the achieved exposures, they are inhibiting the target to the same degree. 462 is much different. TNG908, for example, let's say it's about twofold above what we think would be the exposure required for maximal activity based on preclinical assessments. 462 is much higher, achieving sixfold, sevenfold, eightfold above that required efficacy. And I think that will be critical in time, as we've seen in really the history of oncology, small molecule development. When you hit the target harder, patients tend to have more durable responses.
So I want to maybe dive into the initial data disclosures that you provided a short while ago. I would say heading into those updates, the number one question I think I got was sort of how many patients in lung and pancreatic are we going to see and what's the bar? And so we got the update and we didn't really get that for 462. We didn't get any pancreatic and lung data. So maybe you could speak to why that was the case and then maybe have some follow-ups.
Absolutely. So what we disclosed recently were a few things. First is that we had halted development on TNG908. TNG908 looks like a drug, frankly. It's a good molecule. We always believed preclinically that 462 had some advantages to it. So we're not surprised that that translated the way it did to 462 looking better. But TNG908 has PK in many ways similar to Amgen and has activity, we believe, similar to Amgen. We saw two of nine responses in pancreatic cancer. And of the four evaluable non-small cell lung cancer patients, one of them responded. Despite this, we decided to stop development of the drug because the data tells us that TNG462 is a better molecule and we believe that will translate to the clinical data.
Additionally, TNG908, because of its PK properties and its ability to penetrate the blood-brain barrier, did not achieve adequate exposures to be active in GBM. And that was where we had focused the molecule. So we stopped developing it. So that brings us to TNG462 and the data that we did provide. At the time of the data update, we were fortunate in some ways that in escalation, we had a couple of very enthusiastic investigators who put a good number of cholangiocarcinoma patients on. This is a disease with a terrible prognosis. And it's a disease that both BMS and Amgen have published response rates on of 18% and 15% respectively. We saw responses in three of seven patients. That's an overall response rate of 43%. And while that's a relatively small number of patients, we did believe it provided an opportunity to make a reasonable apples-to-apples comparison.
In contrast with non-small cell lung cancer and pancreatic cancer, we had neither adequate patient numbers nor duration of follow-up to provide data that we felt would be fully informative. We believe we're a better place to let that data mature and that that release data would be clearer to the community, and we look forward to presenting that data in 25. I think that answers your question.
Yeah, no, that was perfect. I guess maybe my follow-up would be there were some pancreatic and lung cancer patients enrolled in 462, albeit sort of small sample sizes to make definitive conclusions. But anything you can say at a high level within those few patients, what you may or may not be seeing with 462?
Absolutely. We have said that we've seen activity, by that I mean PRs, in a number of indications beyond cholangiocarcinoma. That includes both non-small cell and pancreatic cancer. We have also said that, you know, in the patients that we looked at, which were the patients treated at active doses and escalation, and it's about 30 patients, the median PFS at the time of the data release was 24. And many patients were ongoing, such that that median had an opportunity to increase, which it may do. We haven't updated it. That is a meaningful number. It's a meaningful number in a lot of ways. First is that from Amgen's data set, their median PFS seems to be lower at about 15.5 weeks. BMS did not release a PFS, but looking at their entire data set, I believe they may be in a comparable range to us.
But just taken in a vacuum, six months in a phase one-ish population of very heavily pretreated oncology patients with hard-to-treat tumors is a number that I think speaks well of the molecule. A lot of patients do well on this drug for a fair amount of time. And that data set points to us to real opportunities for further development.
So you mentioned the comparison to the molecules within cholangio and how 462 is performing early on relative to what Bristol and Amgen have shown in those settings. Do you think it's fair? Where are sort of the pitfalls if we try to think about what Amgen and Bristol showed in lung and pancreatic and what that could mean for 462, again, using the cholangio sort of experience as a starting point?
Sure. So I mean, the fairest caveat of a 43% overall response rate that we released is that it's seven patients. So the overall likelihood is that that will change. And will a 43% overall response rate be sustained? I have no idea. But these seven patients did very well and compared favorably to Amgen and BMS. The point that we're trying to make is that our molecule does have properties which are better than Amgen and BMS. We believe the PK looks best in class. We believe the safety profile looks best in class. At our 200 milligram dose, of the 20 patients that we disclosed at the time of data release, we saw zero dose reductions and a very clean safety profile with nothing really more than 5% of grade 1, 2 effects and no grade 3 or 4 adverse events. That is an improvement over Amgen.
That is an improvement over BMS. Now, our 300 milligram dose level did indeed have some on-target cytopenias. We already talked about that. They were in the range of 25% for both thrombocytopenia and anemia, and they resulted in a fair number of dose reductions, so I want to be clear about two things. Three things, I would say. It wasn't a slam dunk that 300 wasn't tolerable. A lot of patients stayed on. A lot of patients stayed on for a long time. In fact, I would say some of our investigators continue to believe that 300 is the best dose. We don't think so. We think 40%-50% dose reduction is too much in this patient population, so we're exploring 200-250.
The other thing I want to be entirely clear about is that when I look at the data sets for 200 and 300 milligrams and ask the question, is there a difference in efficacy? The unequivocal answer to my eye is that they look the same. So listen, I do, as a trained oncologist, believe that the harder you hit target may be better. For that reason, we are exploring 250 milligrams because it might be the right dose as intermediate between 200 and 300. And we have one opportunity to get the single-agent dose for further exploration, right? But 200 milligrams is a dose that I'm very comfortable with in terms of safety for sure and efficacy as well.
Great. So I think you've guided towards the next monotherapy update for 462 for some time in 2025. For that update, what sort of questions are you hoping to be able to get answers for and what we could expect within that data set?
Yes, so we do look forward to presenting that data in 2025, and in short, I think it's the data set that many folks, as you articulated, were hoping to get with this data update. I'd like to remind everyone that TNG462 opened only in June of 2023. That was a year or more after BMS and Amgen. So the program is a year less mature. We opened expansion only in June of this year, meaning it's only been open for about five months. There has not been a long opportunity to put patients on expansion and get interpretable data. That is the data that we want to share with the 2025 update, the totality of the escalation and expansion data with the data that we've accrued from pancreatic cancer, non-small cell lung cancer, as well as the other histologies on the study.
So I think among the investor community, and maybe this is fair or unfair, that I think there's still a debate of whether there's monotherapy opportunities for this drug class, just given the totality of all the data sets. How do you think about that? And where do you see potential monotherapy opportunities still in play?
For TNG462 and BMS, both of which are not blood-brain barrier penetrant, I think the clear opportunities to meaningfully benefit patients as single-agent are in non-small cell lung cancer after standard of care and in pancreatic cancer after first-line chemotherapy. I think we need more data to be confident that those are the paths. But that is where our goals are focused. And we believe that we can have a first pivotal study in second-line pancreatic cancer as single-agent. You asked the question for the class of molecules, and I would point to GBM as another opportunity for what could be a clear single-agent registration path. And that could be in either the adjuvant or relapsed refractory setting. But I think the adjuvant setting is attractive. And that's why we're really developing our next-Gen molecule, TNG456.
So I think the other thing you sort of spoke to with the most recent update was intention to initiate a number of different combination strategies with 462. Maybe before getting into those specific combinations, what gives you the confidence that with the data you have to date, that that's an appropriate step to be taking now?
Yeah, so that's a great question. I would say the first thing is we're convinced of the activity of the molecule, and we're convinced of the safety profile of the molecule. This molecule is active in the indications in which we are doing combinations, and we think there's an opportunity to change standard of care with these combinations. The combinations we're pursuing really fall into what I call two bins. One are really combinations with the standard of care therapies, which would be required to get to the front line in both pancreatic cancer and non-small cell lung cancer, and those which are more novel combinations with basically going after genetically defined targets. To that end, we're combining TNG462 with gemcitabine and modified FOLFIRINOX in this study.
And what we're very excited about is in the same study, we're combining with the KRAS inhibitors from RevMed, both the pan-RAS-6236 molecule and the G12D-specific inhibitor 9805. Given the safety profile of the RevMed molecules, given the activity they've demonstrated, and the data that we've seen in pancreatic cancer, we think the combination really has a high likelihood of success to transform standard of care in pancreatic cancer, whether it be in second or first line.
And then within lung cancer, sort of any efforts there in combination strategies that you're?
So there are really two other efforts that we're making in the short term. One is adding pembrolizumab to TNG462. That will be done within the existing first-in-human. And it's really to demonstrate the tolerability of pembrolizumab with 462, critical in non-small cell lung cancer, as well as a lot of other indications. And the intent is, once we understand the profile, to layer standard of care non-small cell chemotherapy on top of that.
Sorry, go ahead. Finish your thought.
One other combination we're moving forward is with the CDK4/6 inhibitor, abemaciclib. Now, there is a biologic rationale for that, which is relevant in 100% of MTAP-deleted patients: CDKN2A is deleted. I didn't speak to it, but the reason that so many cancers have MTAP deletion is because it sits next to the tumor suppressor, CDKN2A, which is a real driver of so many malignancies, and when CDKN2A is deleted, MTAP comes along for the ride. That's why every MTAP-deleted cancer is CDKN2A deletion. To make a long biologic story short, that CDKN2A deletion creates a susceptibility to CDK4/6 inhibitors. We've shown preclinically how well the molecules work together, and what we're planning to do is combine TNG456 with CDK4/6 inhibitor, abemaciclib, focusing on GBM as well as non-small cell, knowing that abemaciclib is both CNS active and has demonstrated activity in GBM.
So I'm going to try to squeeze another question or two here in the last minute. So you talked about the opportunity for PRMT5 inhibitors potentially in GBM. You just disclosed a new program, TNG456. So maybe speak to that, what it potentially might offer over 462, and sort of how the strategy there may be different?
Yeah. So I think it's fair to think of 456 simply as 462 with blood-brain barrier penetrability. The molecule has similar potency and similar selectivity, both of them superior to TNG908. In short, we believe that TNG908 was not successful in GBM because we didn't get enough drug into the brain relative to activity. But what we understand about 456 is that we should get many-fold above the exposures we need to be active. That program will move forward in GBM where really almost 50% of patients are MTAP-deleted. But the mechanism is as relevant in every other tumor as it is in glioblastoma.
Perfect. I don't think I have an opportunity for that second question. So we'll leave it there. We're out of time. Adam, thanks so much for your time and thoughts. And thanks everybody for joining and tuning in. Take care. Enjoy the rest of your day. Thanks.
Thank you.