Tango Therapeutics, Inc. (TNGX) Earnings Call Transcript & Summary

Hi, everyone. My name is Kevin Strang. I'm one of the biotech associates here at Jefferies. It's my pleasure today to introduce Adam Crystal, the President and Head of R&D at Tango Therapeutics.

Thank you much for the introduction, and thank you for your attention today and the opportunity to introduce you to Tango. My name is Adam Crystal. I'm the Head of R&D at Tango. I'm very excited today to really give you a summary into the foundings of Tango, but then go into some more depth on recent proof-of-mechanism data, which we have put out there for our lead program, the MTA-cooperative PRMT5 inhibitor. So Tango Therapeutics is now a clinical-stage precision oncology company, leveraging the concept of synthetic lethality with really a state-of-the-art CRISPR platform to develop a sustainable pipeline of novel oncology targets. You can see that at the moment, we're quite busy with our lead program, TNG908, in escalation to other molecules, one of which is also a PRMT5 inhibitor, status post IND approval and in study startup mode; and our fourth program, TNG348, planning an IND filing in the second half of this year. Things -- we are well funded with the runway which extends into 2026. The idea of Tango is to leverage synthetic lethality to go after cancers which were previously untargetable. Specifically, I'm talking about tumor suppressor genes. When there is no protein in the cell, it cannot be inhibited. Most of the oncogenes we're familiar with where success has been had, whether it be RAF or EGFR or KRAS, are inhibitable. They are oncogenes. Our platform leverages synthetic lethality to target genetic context of tumor-suppressor loss to treat those cancers effectively in a way which previously could not be done. This is about half of cancers and really is a substantial opportunity to discover novel targets. This concept has been clinically validated. PARP inhibitors now in the clinic and approved for over 10 years are the example where this has been tremendously successful. As you know, in the setting of BRCA mutation, PARP inhibitors are quite effective in killing cancer cells while sparing normal tissue. This is the idea on which therapeutic index emerges in synthetic lethal relationships. And we believe that hundreds of such examples exist within cancer biology. So the company has now extended to the fifth phase of the idea that drove it forward, which is to identify genetic context in which targeting genes can be specifically selected, and then leveraging that genetic context to select those patients who are most likely to respond. We have identified novel oncogenes or novel targetable genes, enabled discovery, identified drug candidates and now put them into the clinic. This is the status of our pipeline. We have 4 clinical programs. And -- well, we intend to have 4 clinical programs within the next year. As I mentioned, our lead program, PRMT5, an MTA-cooperative PRMT5 inhibitor, began dosing last summer. We'll discuss some of the data which has emerged from that trial today. The second and third studies are now in study startup mode, and we intend to file the IND for that fourth study in the second half of this year. The pipeline does extend into earlier discovery where we have multiple targets undergoing that discovery-enabling stages, and we look forward to sharing those in the not-too-distant future. I'll now spend the rest of my time talking about our lead programs, the MTA-cooperative PRMT5 inhibitors. PRMT5 inhibition has been pursued clinically before. They are our first generation of PRMT5 inhibitors which demonstrated activity. These molecules, however, did not have therapeutic index because they inhibited normal tissue to the same degree which they inhibited tumor. This program is leveraged on the idea of MTAP deletion to create that therapeutic index. Said another way, in the setting of MTAP deletion, a synthetic lethal relationship is created where PRMT5 can be specifically inhibited in tumor cells, resulting in activity and a positive therapeutic index. This is a large opportunity. It's seen in 10% to 15% of tumors, period. We see MTAP deletion across multiple histologies and believe that there is a great opportunity for patients to benefit as well as commercial opportunities in these large indications. And as I mentioned, TNG908 has established proof of mechanism in Phase I. We'll go over that data, and we look forward to presenting additional data in 2024. This data speaks to the opportunity for an MTA-cooperative PRMT5 inhibitor. Across solid tumors, its prevalence is between 10% and 15%. And here, you can see how that breaks down by indication. There are multiple indications of commercial significance where the incidence is greater than 10%. Of particular interest to us is GBM, where somewhere in the range of 40% of GBMs are MTAP-deleted. Other indications which are worth highlighting are non-small cell lung cancer, where both squamous as well as adenocarcinomas have a prevalence of greater than 10% for MTAP deletion as well as bladder and pancreatic cancers. We have 2 molecules in this program. The molecules are differentiated as we'll discuss going forward. We think both of them have a clear path to clinical proof of concept. TNG908 is uniquely brain penetrant. Our second molecule, TNG462, is not. TNG462 is, however, more potent as well as more selective for mutant rather than wild-type -- mutant tumors rather than wild-type tissue. And we're excited to see where both of these may be developed as we observe the clinical data emerging. This slide speaks to the mechanism by which this molecule in [ the ] class of molecules works. It's actually quite straightforward once you get your head around the basic presets. MTAP deletion, as we mentioned, is present in 10% to 15% of solid tumors. MTA is a substrate of MTAP. So when you delete MTAP, MTA accumulates uniquely in the tumor, not in the normal tissue. Our molecules, TNG908 and TNG462, are built to take advantage of that MTA accumulation, which is to say they bind to and inhibit PRMT5 only in the setting of the MTA accumulation, thus inhibiting tumor cells and sparing wild-type tissue, creating that therapeutic index. This schema illustrates that concept in exactly how it works. What we're showing here is the binding pocket of -- the enzymatic pocket and the binding pocket of PRMT5. The outline is indeed that binding pocket. Within that pocket normally sits SAM. And when SAM sits there, it's a methyl donor. What it does is it methylates the client protein, resulting in these SDMA methyl marks. However, when MTA accumulates, our molecule is built to sit in that pocket with MTA specifically and block that otherwise normal methylation. The molecule is designed such that it cannot access the pocket when SAM is there, meaning it only works in the setting of MTA, not SAM, which means it only works in the setting of tumor, which is MTAP deleted, not normal tissue. This results in a distinctly differentiated profile from first-generation inhibitors, critical to understanding why this molecule should be both active as were the first-generation inhibitors while also having a positive therapeutic index, enabling more activity as well as enhance tolerability. What you see on the left are isogenic cell lines in blue, normal, and in red, MTAP deleted, and various concentrations of our drug -- I'm sorry, of GSK and J&J's drug. Those are 2 of the first-generation molecules. What you can see is that these molecules are dose dependent but do not discriminate normal from mutant tissue. That is they inhibit PRMT5 similarly in normal tissue as cancer. And so once you are killing the cancer cells, you're indeed causing toxicity, and what this read out clinically as was thrombocytopenia and anemia. In contrast, in the middle panel here, you see TNG908, our lead molecule, is indeed selective for the MTAP-deleted setting or the tumor setting. It is indeed dose dependent. And what you can see in the area circled by that box is that there is a dose range at which SDMA, the direct downstream methyl mark of PRMT5, is ablated or nearly ablated, whereas that signal is relatively less inhibited in normal cells. An important point to make here is that we do not anticipate toxicity from wild-type inhibition based on the first-gen data as well as our own preclinical data until that blue bar is decreased by about 90%, meaning we believe we'll be able to achieve activity in the absence of toxicity. An important point to make here on the right here -- on the right is the difference between inhibition in xenografts, in wild-type versus no mice. And here, you can see very clearly in mice, we achieved complete or near-complete inhibition of SDMA at 30 milligrams per kg a dose, which is -- results in tumor growth inhibition in these xenograft models; or 90, same SDMA suppression, which results in a complete regression in many of these models. Here, you can see the difference. Six different models in which our molecule results in complete regression at 120 mg per kg, a dose which we anticipate will be tolerable clinically. You can see that, that 30 mg per kg dose, which resulted in near-complete SDMA ablation in the prior slide, results in a robust tumor growth inhibition. In contrast, when we increase the dose, the PD readout is the same, complete ablation of SDMA, but now we're achieving a near-complete regression. As mentioned, TNG908 is blood-brain barrier penetrant. We're very excited. It's very important for us to see how this molecule works clinically in GBM, an attractive path for development. Here, you see a preclinical model, demonstrating efficacy of this molecule in an orthotopic GBM model, a subcutaneous GBM model on the left. And similarly, in the middle or on the right here is an orthotopic GBM model where we see an approval of mouse survival in comparison to a vehicle. And historically, this is much better than has observed with approved standards of care, including Avastin and temozolomide. It's worth noting that this activity was observed in the setting of 15% penetration into the rodent brain. We expect much higher in humans because the penetration or Kpuu [ in sino ] is about 92%. So we saw this despite what we expect is inferior penetration in comparison to what we will see in humans. This brings us to the design of our first-in-human study. This is a study which dose finds in the setting of MTAP-deleted tumors or patients with MTAP-deleted tumors exclusively. We are prospectively selecting for that. And all solid tumors are permitted. Once we have identified a dose at which we want to move forward, we intend to expand into either of 6 arms. We are particularly interested in GBM as well as non-small cell lung cancer. And there are other opportunities for development here, including malignant peripheral nerve sheath tumor, MPNST, where approximately half of such tumors are MTAP-deleted as well as mesothelioma and cholangiocarcinoma. We will also have an arm which is histology-agnostic, meaning if you have a solid tumor with MTAP deletion, you are eligible for those criteria. It's worth noting that there are multiple paths for registration depending on how the molecule looks clinically. One could envision taking this molecule forward into GBM or non-small cell lung cancer. There are also paths we could consider for histology-agnostic approvals. So on the next slide, I will speak to the proof-of-mechanism data which we recently released in May. This data includes PK/PD data from 2 cohorts, I will show you that PD data shortly, as well as safety data from 4 cohorts. In these cohorts, we enrolled 16 patients across 12 different histologies. And we're pleased to say that we saw what we believe to be dose-dependent inhibition of target in a tumor-specific manner. PK looks good thus far. We're seeing dose dependency across the cohorts. And it's important to emphasize that at the first 2 dose levels, dose level 1, 25 milligrams BID; dose level 2, 50 milligrams BID; we did not anticipate being nor worry in the efficacious range in terms of exposure. We did expect to see deep exposure in a tumor-selected manner of SDMA, and I'll show you that data shortly. And critically, in terms of safety, we were quite pleased to see 0 DLTs; no evidence of bone marrow suppression, which is the known on-target toxicity of wild-type PRMT5 suppression in MTAP wild-type cells; and nothing greater than a grade 2-related adverse event. Escalation is ongoing. We're very excited to see brisk enrollments. Our investigators are quite engaged, and we have waitlist in multiple sites. When we open up cohorts, we are getting very brisk enrollment as we move forward. We're looking forward to presenting additional clinical data in 2024. So here is the proof of mechanism which I spoke to. To orient you to this slide, what I'm showing you are paired tumor biopsies from the patients in cohort 1 and cohort 2. It's worth mentioning, we're presenting all of the data which we have in hand, which is pair 4 for paired tumor biopsies. These were scored blindly by pathologists. And it's worth mentioning that there were 2 other patients in cohort 2 who we attempted to biopsy on treatment. One was a technical fail, and we could not present that data here today. And the other one, they looked for the tumor and couldn't find it in the ultrasound suite, so they did not biopsy. We were excited. The tumor had not disappeared. It was simply technical reasons that it couldn't be assessed. So what you see here, to say it very succinctly, is, at baseline, a lot of SDMA staining and on treatment that, that signal has been dampened or near completely ablated in that cohort 2 ovarian patients. You can see the red bars above in which this inhibition of SDMA is quantitated. And you can see near complete ablation of that signal in that ovarian tumor patient. I'll show you in detail that patient on the subsequent slide. But here, the most important scrap of data is comparing the cohort 2, cycle 2, day 1 in blue to the cohort 2, cycle 2, day 1 in red. What you see there is the absence of inhibition of SDMA in the normal tissue in contrast to deep suppression of SDMA in the tumor sample. Critically, we have room to run in terms of escalation. Why do I believe that? If you can see that dotted line in each of the graphs, that's where we expect to see on-target toxicity when PRMT5 is that inhibited in the wild-type cells. We're not there, and we don't believe we will be there at the dose levels which follow up for some time, meaning we have room to escalate from cohort 2 and beyond to achieve deeper suppression of PRMT5 more activity in a manner which continues to be tolerable. This slide is from a single patient. This is that ovarian tumor patient who you saw on cohort 2. And this is actually a single biopsy taken on treatment. It is a biopsy of a ovarian metastasis in the liver core biopsy. And what you can see, and this is actually from a single slide. You can see normal tissue, which was also acquired adjacent to the ovarian metastasis on the left and the metastasis on the right. In the normal tissue, you can see persistent deep staining of SDMA because this MTA-cooperative PRMT5 inhibitor does not inhibit it in the absence of MTAP-deletion, whereas in the ovarian metastasis, which is MTAP-deleted, there is ablation of SDMA signal. One piece of data which is interesting to point out can be seen on the slide on the right here in this ovarian metastasis. You can see in the bottom right at about [ 4 30 ], a few darkly brown-staining nuclei. One could look at this and wonder if this is tumor heterogeneity. That is not the case. What you're seeing there is normal vasculature. So this is normal vasculature supplying the blood supply to this tumor. It is, in fact, MTAP intact. And therefore, SDMA is not inhibited. So even within this box on the right, I believe one can conclude that the molecule has achieved proof of mechanism by demonstrating tumor-specific deep inhibition of SDMA. Our second program, TNG462, is 45-fold selective for MTAP deletion. This is about threefold more selective than TNG908. You can see on this slide how this plays out across multiple 180 cancer cell lines. In the middle, you can see that the molecule highly enriches for MTAP deletion in red, whereas TNG462 demonstrates almost perfect selectivity for tumors which are MTAP-deleted. This molecule will be daily in the clinic in contrast to TNG908, which is BID. It is not blood-brain barrier penetrant. Here, you can see the activity of this molecule across 22 xenografts. Conclusions to be reached here is that it results in regression in about half of these models, and that it shows no preference for histology active across all of these histologies. On the left, you can see a cholangiocarcinoma, contrasting 908 in gray to 462 in red. And you can see that while 908 results in a deep regression, we eventually do see that tumor growing out. Excellent positive data in contrast, 462 achieves a complete response which stays down for longer. As I mentioned, these molecules are both moving forward clinically. 908 in escalation, having dosed the first patient last year; and 462 in study startup mode after clearing -- after recently clearing the IND. With that, I'll close by saying we have multiple milestones over 2023, most recently with the presentation of proof-of-mechanism data, and we are well financed with the cash runway taking us into 2026. I'll stop there and invite questions from the audience. And thank you for your time.

Do you know how many patients you're going to be able to enroll? And then obviously, you can't really biopsy those patients very well. Is there anything you can look at biomarker-wise for those patients?

So for GBM, how many patients do we intend to enroll and how are we going to assess them. You're right, it's unlikely that many of those patients will get on-treatment biopsies, it being so invasive. I think that interpreting GBM data is more complex than many in terms of responsiveness. We're fortunate to have expert KOLs who have engaged us in saying, it's not just the imaging, talk to us in totality about how the patient is doing. This is important. We have and we will do this. We're currently admitting patients in backfill cohorts with GBM. And so I think we anticipate having substantial numbers, which would enable us to say how it does across a reasonable number of GBM patients. And so that's helpful in framing out the answer to the first part of your question, which is what kind of a data package will we have. What we intend to present in 2024 is a convincing clinical proof of concept, where we believe that this data package demonstrates not only that this is effective, safe and developable in MTAP-deleted cancers, where we can assess that in larger numbers of patients, not just the singlets of histologies that we have now. So our goal would be able to provide data in which one could assess not just, is it active in MTAP deletion, but what does it look like in GBM, or what does it look like potentially in non-small cell lung cancer.

Great. And just one more follow-up. For the data you showed, it looks very clear with the SDMA repression there. Just wondering because you didn't -- you don't expect cohorts 1 and 2 to be efficacious, but you do expect going upwards. What is the difference there in terms of the amount of repression that you need or anything else that would be predictive of that?

It's a critically important question. And I think the most concise way to answer the question is that we understand that SDMA signal is completely ablated preclinically, and not one could argue clinically before PRMT5 inhibition is completely achieved, meaning the assay is potentially not as sensitive as ideally it might be. It is the best assay out there, which is to say, at cohort 2, we expected to see SDMA suppression and no efficacy. That is indeed what we saw. We expect that as we escalate, we'll in fact see deeper PRMT5 inhibition, which will be accompanied by activity.

Great. Thank you very much.

Thank you so much.