The Uromigos Episode 101: Paper of the Month—Belzutifan in Renal Cancer

Dr. Eric Jonasch gives an overview of his paper published in Nature Medicine.

Episode Transcript

Tom:
Hey everybody, welcome to the 101st podcast. We’re persisting with this despite mixed feedback. We’re honored to be joined by my great friend and colleague, Eric Jonasch from the MD Anderson. Eric and I go back a long way and I remember a particularly fun walk around Berlin, Eric, but we’ll talk about that on a different day.

Today, we’re here to talk about your excellent Nature Medicine publication, which you last author. Do you want to just talk maybe for a couple of minutes about targeting and a little bit about what you did?

Eric Jonasch:
So, VHL mutations are functional deficiencies of the VHL protein exist in the vast majority of clear cell original cell carcinomas. And obviously also in the hereditary VHL disease syndrome. This is an E three ubiquitin ligase, which basically means that it’s capable of in together with other proteins to ubiquitinate and essentially target for execution, a variety of many different proteins.

And so VHL’s job as an E3 ubiquitin ligase is to help with the ubiquitination of hypoxia-inducible factors, 1-alpha and 2-alpha. And these two transcription factors actually post-translational regulated. So, the transcriptional regulation is moderate, but what really matters is how quickly the protein is degraded postproduction.

Tom:
It definitely sounds like you know what you’re talking about there.

Brian:
So, Eric

Eric Jonasch:
I can fool people. So yes.

Brian:
Hey Eric, this is Brian. Thanks for joining. So, first of all, on our last podcast, Dave McDermot dropped an F bomb. So, we’re hoping that this podcast goes a little smoother than that one did tell us about in your paper.

Obviously, it was the dose escalation sort of one of the initial studies, but maybe just, you could even just talk broadly about sort of the clinical activity that we’ve seen with this compound. And then maybe I’d like to go back and talk about erythropoietin and some mechanistic stuff.

Eric Jonasch:
Yeah, sure. So, this was a phase one study with, with an expansion cohort and clear cell renal cell carcinoma. And the primer endpoints really were toxicity and recommended phase two dose and the dose escalation cohort, and then efficacy in the RCC cohort.

43 patients were in the dose escalation cohort, 55 in the RCC cohort. And actually, the majority of patients, even in the dose escalation cohort were renal carcinoma patients or a smattering of other diseases. Interestingly, what was seen, number one is MTD was not reached, went all the way up to 160 milligrams, but in terms of pharmacodynamics, the dose of 120 milligrams was selected and then moved on to the RCC cohort where there were 55 patients. And the primary endpoint there really was efficacy.

And what was seen was a 25% objective response rate, a median progression free survival of 14.5 months. And this is in individuals who had received a median of three prior therapies, and they had all, the vast majority of them had either received a VEGF receptor inhibitor, checkpoint inhibitor or both. So, this is a little interesting from a positioning perspective.

Tom:
Eric, can I stop you there? So, 71% I think received both, which is a heavily treated population of renal cancer. There are two or three things around the efficacy endpoint that you preset for success or failure.

And then the second question I wanted to ask was about the assessment of resist response, assessed it by resist. Was that confirmed resist response, or was that just investigator assess response?

Eric Jonasch:
It was confirmed, and it was confirmed response. It was by independent review. The null hypothesis was an objective response rate of 15%, I believe it was. So, it certainly in the alternate was in the twenties. So, it did meet its primary endpoint in terms of objective response rate.

But I think the most important element of this from an efficacy perspective really is the fact that we are seeing this prolonged progression free survival. And also just from a kinetics perspective, we’re seeing that in individual who receive this agent, that the time to response ends up being somewhat longer than what we’ve seen with some of the other agents.

Tom:
Can I ask one more question than I’m going to have over Brian? So, Eric, one of the issues is that maybe the third and fourth and fifth- and sixth-line renal cancer population actually have better outcomes than second- and third-line patients, because there’s a group of professional patients that do exceptionally well.

It’s terrific that they exist, and they’ve got lymph nodes and maybe one or two lung lesions. And maybe these patients are not represented of the classic clear cell renal cancer population. How do you feel about that?

Eric Jonasch:
Yeah, so when we enrolled the largest number, or the plurality of patients were enrolled at MD Anderson. And so, we do have a lot of anecdotal personal experience with this. I can tell you that the individuals that were enrolled in this were not cherry-picked individuals with indolent disease there were a fair number of individuals who did really have quite aggressively progressive disease.

And this was certainly our sort of, we did not select for individuals who had favorable features. And so, I think this is really representative of the population of individuals who need treatment and who are at a critical juncture.

Brian:
Hey Eric, I have the same question. That’s why I hate when Tom goes first because he steals my question.

Tom:
I apologize, Brian, I apologize. [crosstalk 00:06:33]

Brian:
I think overall it’s challenging to interpret time to event endpoints in single arm studies. And I think Tom’s point is a good one that even if you are not selecting patients, they’re selecting themselves, they don’t make it to MD Anderson on a third, fourth, fifth line trial without having good biology, but be that as it may, it’s an impressive number. And we’ll see in the registration trial, certainly what it turns out to be.

My question is around recent checkpoint inhibition, how many patients do you know percent who are coming off a checkpoint inhibitor who then went on study where maybe there was some overlap, some lingering antibody in their system. And this might even reflect some combo activity.

Eric Jonasch:
It’s an interesting question. And in terms of what percentage of patients had checkpoint is their most as their most proximate prior therapy? I don’t know the answer to that, but it was undoubtedly a certain percentage, but that would imply that there would be some synergy between this agent and checkpoint inhibition. And that in of itself actually at this point in time has not been proven.

Brian:
Right. Just curious how many other responders were coming off checkpoint. We’ve seen that before with some of the TKI in this post IO setting, patients coming off a checkpoint who have both toxicity, but also potentially enhanced efficacy.

And I guess the follow up question then is how much tumor shrinkage? Response rates VEGF TKIs in a refractory setting are about 20%. Most of the studies are sort of in that from an objective response rate, but I’m wondering just terms of a tumor shrinkage sort of waterfall look at the data.

Eric Jonasch:
Yeah. So, what’s the specific question with regards to the?

Brian:
How many patients had tumor shrinkage? What percent of patients had any degree of tumor shrinkage?

Eric Jonasch:
Yeah, it was about two thirds. So there definitely was about one third or so of the patients. I’m looking at the waterfall plots on the paper here. That there definitely was a subset of individuals who had PD as best response. And then there was a majority that were kind of in between the plus 20 and the minus 30. And then there was this group that were greater than 30.

Brian:
And last question on response. What do you think accounts for these delayed responses? Because that’s distinctly unusual with VEGF TKIs, generally first or second scan you know what you’re in for, but it’s clearly a feature of this drug is mechanistically or what do you think what’s the hypothesis for why that is?

Eric Jonasch:
You know, I think it’s because it’s acting at the level of the tumor cell. And so, what we’re seeing is we’re seeing a relatively slow perhaps modulation of the inherent biology of the tumor cell itself, which then has a lag in terms of how it starts influencing the microenvironment as opposed to a therapy that is immediate working on elements of the microenvironment, like the endothelia.

And so, it’s that reprogramming of the tumor cell and then the delayed downstream consequences of that.

Brian:
And as you know, this drug’s being combined with checkpoint inhibitors, TKIs, et cetera, in Merck’s portfolio. Do you think this drug would have synergy with the VEGF TKI? I have a hypothesis that I’d love to see this drug replace VEGF TKIs, right? I mean, that’s one of the toxicity downsides of TKIs.

And I’m not saying it can, but I guess I just wonder what your thought about if you were king of the world, how you would most optimally combine this drug presuming, you think it’s future is in combination.

Eric Jonasch:
Yeah. I mean, obviously if you had all the money and all the time, you would both take a somewhat empirical approach to combine it with what you can, which would be obviously checkpoint antibodies and VEGF receptor inhibitors. But mechanistically, I think there are a couple of questions.

So, I’m not answer your question directly because there’s the obvious trials that are happening, but I think there are some questions we need to answer so that we can really design the ideal trials. Number one, when we’re blocking hif-2-alpha, we’re thinking of this as a tumor cell autonomous event that we’re doing stuff to the tumor cell and that’s going to do stuff to the microenvironment.

But then the next question this is a systemic treatment, how is this affecting myeloid derived suppressor cells? How is this affecting tumor associated macrophages? How is this affecting effector T-cells?

So, what’s it doing to the microenvironment? And what do we need to know from a patient-to-patient perspective about the microenvironment, to know which patients this is good for or bad for, from a microenvironmental perspective.

And the second thing is if it is modulating certain things in the microenvironment, what other agents are needed to then either enhance or counterbalance those effects. So, these are the sorts of questions that we don’t have answers for. And I’m hoping we’ll be able to obtain over the next few years.

Tom:
We do know a little bit about this. So [Tony Ary, geo ASCO 00:11:17] 52 individuals Belzutifan plus Cabozantinib after nine months progression follow up, the response rate was 22% and I guess higher disease control rate, but the combination with Cabo that response rate looks modest. It doesn’t look like the sort of 70% or 50% that one might see with axi pembro to other drugs, I don’t know, what do we, what do we feel about that response?

Brian:
I mean, I think it was that data was pretty early, I guess, would be the only caveat, as there, I don’t remember the median follow up, but there are a whole lot of patients I.

Tom:
Median follow up was nine months.

Eric Jonasch:
It’s I think a highly, especially with this agent and the kinetics of this agent, I think those data are not mature enough to reach any conclusions. I mean, the PFS and again is a heavily censored, PFS was 16 and a half months, I believe.

So, there’s potential for if you’re thinking about vertical integration, starting with a hif-2-alpha inhibitor, and then also hitting some of the downstream events. But the thing about hif-2-alpha inhibition is its much more than just a hypoxia or hypoxia regulating gene or regulator of hypoxia type genes. It also regulates metabolism and things like that.

So, to think of it as purely vertical, as opposed to perhaps complementary again, I think the proof of the pudding here is going to be in the eating. We’re going to have to see what happens with that trial. And obviously there’s now other TKI Belzutifan trials that are underway. We’re going to need to take the empirical information and backtrack into mechanisms.

Tom:
Eric, let’s move to safety. There is an event profile was a phase one trial. Do you want to just talk a little bit about that?

Eric Jonasch:
Yeah, so it overall, it was really quite well tolerated. So, what was seen is that there were no dose limiting toxicities observed in the dose escalation cohort all the way up to 160 milligrams once daily, there were a couple of just in the dose escalation phase, some toxicities, but then in the RCC cohort, what was seen was that anemia, which is an on-target effect.

EPO is a hif-2-alpha client gene was seen in 76%, fatigue in 71%, dyspnea in 49%, nausea in about 36%. In terms of grade three treatment emergent events, anemia was seen in 27% and hypoxia in 16%. And these were the hypoxia sort of an interesting thing was definitely more in this study than in the VHL study.

Often was associated with some intercurrent issues that these patients had mainly from disease driven in their lungs and did result in either dose reduction or dose discontinuation in a small number of individuals but did not look like it was a deal breaker from a safety perspective.

Brian:
Hey, Eric. I have a question about out the anemia and the erythropoietin level. So, I know EPO changes correlated with plasma concentrations. So, I’m wondering if this is kind of like the hypertension with VEGF TKIs story where hypertension generally correlates with higher plasma concentrations generally correlates with better outcome, but in an individual patient, we don’t really use hypertension.

We’re not titrating to hypertension, are EPO levels more titratable if you understand my question. Can we use them to titrate dose or titrate to a response? Or is it just more of a classified, like the hypertension TKI analogy?

Eric Jonasch:
It’s actually kind at an individual patient level. There’s, just hold on a second here. [silence 00:15:26] At an individual patient level, there clearly were individuals who did not have substantial anemia, did not have substantial diminution of their EPO levels who still got benefits.

So, I don’t think we can use it as an individual patient biomarker for pharmacodynamics and for response. But it’s interesting, I think just from at a population level.

Brian:
So then pretty similar to hypertension with TKIs, right? It’s population wide. I think it tells us a little something about the biology and plasma concentrations, but of course there’s patients, you could give them all the drug in the world and they’re not going to respond. Right. So not everybody is who’s dose responsive.

Tom:
Eric is giving, because you gave, I think a lot of patients who got anemia, they received erythropoietin rather than blood transfusions. Erythropoietin has a plethora of different actions.

Is that safe in view? Is it an on-target part of the drug or is it just a passenger and it’s fine to give erythropoietin, remember back in the day there was a randomized data or the retrospective data suggesting giving erythropoietin may not be a great idea to patients on treatment with cancer. What’s your take on that issue?

Eric Jonasch:
We definitely so long-term safety related impact. I don’t know whether we can state anything about that right now, because these effects might be fairly late and delayed, but we definitely did treat a significant number of individuals with erythropoietin and about two thirds of the patients that required some sort of intervention got erythropoietin, and one third got blood transfusion.

We didn’t see any of change in disease kinetics that we could really identify with the erythropoietin. The erythropoietin was extremely effective in really remediating this. And we pushed fairly on to try to actually use the, the EPO agents as opposed to blood transfusions, because we felt that that was probably the best thing for the patients. And it was similar sort of experience in the VHL study. The EPO use was certainly a highly effective way of managing this.

Brian:
So, Eric. Maybe my last question is, I mean, I know there’s big, big combo studies. There’s a registration study versus [Everolimus 00:18:04]. There’s big combo studies frontline study, but where else do you think we need to go? It’s clearly an active drug, it’s well tolerated.

Right? So, integrating it seems to be a no-brainer, but where would you take it? What are we not doing that we need to do might be the best way to ask that?

Eric Jonasch:
Well, beyond the obvious, which is to, to look at this agent in combination to see how we can synergize with other existing agencies. I think we really need to understand the biology a little bit better so that we can refine.

Right now, we’re looking under the lamp post in terms of how we’re designing clinical trials. I think we need to be able to actually inject a little bit of scientific knowledge into that so that we can perhaps go laterally as opposed to linearly with our drug development.

Now from a positioning perspective, should this be a frontline agent, should this be a rescue agent? I think, we don’t have a lot of information in the frontline setting and that’s a sort of I think a potential whether or not this is going to have some unique benefit in individuals who are previously untreated. We’ll have to see.

Tom:
Eric, last question about that. So, you’ve done a study previously looking at individuals with VHL who develop renal cancer and you showed a year ago with the rest of your team, a great response rate in those groups, potentially practice changing in my opinion.

Early in the disease process. And Brian, you’ve done some of this work up we’ve shown that actually good risk patients and perhaps even surgical type patients, operatable patients have more VEGF driven tumors. Whereas the poorest patients appear to have more complex tumors that perhaps are less VEGF driven with the strong relationship between hif-2-alpha and VEGF.

Does that suggest that this drug is going to work better earlier in the disease process rather than waiting until the tumor becomes very heterogeneous and, and, and is largely VEGF independent. What’s your take on that theory?

Eric Jonasch:
That’s a likely outcome that this is going to work even better in genomically less complex tumors, ones that are more purely VHL loss, hif driven as opposed to ones that have what I would sort of call secondary mutations and a more complex microenvironment that arises from that. So, yes, I believe that will be the case.

Tom:
And Brian

Brian:
And do you think, I don’t remember from the paper or other presentations, whether there was evidence that it worked better in favorable risk patients versus intermediate poor? I don’t remember that, is that true or?

Eric Jonasch:
There was actually a breakdown looking at intermediate and poor risk patients as, and the objective response rates were pretty similar. By breakdown, but there’re small numbers and you’re taking 55 patients and slicing them into smaller cohorts.

Brian:
Yeah. Fair enough.

Tom:
I’d like to stop by asking both of you one question, there’s some data from the public domain and then probably as some and trials being designed, which are not, but could, you just give me what you think if you had, as you said, the one trial that you are going to do with this and Brian, I know you’ve thought about this, a great deal. How would you be positioning this?

Brian:
You want me to go first? So.

Tom:
yeah. Cool. And Brian,

Brian:
I think most intriguingly is whether it could be a VEGF sparing agent, right? We, IO TKIs are potent, but you sort of saddled with that TKI toxicity and we, don’t no longer to continue, et cetera, et cetera.

And so, I wonder just most intriguingly if this could replace the TKI and IO TKI combinations, and other triplet, et cetera, combinations so that you have the activity without toxicity or with limited toxicity, that’s sort of where I would go. I think what’s being done is appropriate. I just think if there’s something that’s not being done, that’s what I would pursue.

Tom:
Eric, what about you?

Eric Jonasch:
Yeah. I think this is one of the first drugs that’s really able to act at the tumor cell level. What we really need to learn now and because it’s clear, we need to modulate microenvironment.

We clearly need to learn what this is doing to the microenvironment so that we can continue to successfully target the tumor cell with this agent but understand what it is that we need to do at a microenvironmental level to synergize with this agent.

I don’t know what that is, but it requires discovery so that we can actually do it. It could be that the strategies that are being rolled out are the correct ones, but I think we need data to be able to actually really back that up.

Tom:
I’d like to see in triplets now, but I’d also like to see it in single agent, as you said, Eric, in those metastatic disease, but super good risk population. Cause it’s super well tolerated. And there’s an opportunity there maybe to get long term of disease control, because there’s a debate about whether you need, I mean, let’s see how the, the adjuvant space pans out, but there is this kind of space around early disease where I think we might be, make a big difference and I’m not quite sure what that trial looks like, but that’s often the case in my life.

Eric Jonasch:
All right. So let me ask you a question. So favorable risk pays, let’s say you take favorable risk previous the untreated patients where we know there’s been no overall survival advantage yet with any IO or IO TKI combination compared to sunitinib, how would you design a trial for that population?

Speaker 2:
Well, Eric, I’m not used to getting asked questions at all on the podcast. So, I found terrible trouble.

Quite good at the asking bit, but when you put it, listen, how would I think about this? I think there are two groups. There is a surveillance group and Brian, you’ve done the randomized trial, sorry, you did a prospective, the prospective [axibnib 00:23:40] trial. And we all have patients with good risk disease with one or two lung metastasis that will be put on surveillance.

Now my kind of issue around that is if the adjuvant immune therapy trials are really positive might actually think, well, these patients probably are going to be benefiting from an immune type therapy, if we could hit a survival signal, but as it currently stands, there is a big cohort of patients kicking around where we say, well, wait into a scan a few months and see if it’s progressing with early disease.

Now that population does currently exist, and you could do a randomized trial in that space. The other area that I’d be really interested in is post adjuvant therapy. We need to do more trial. We do some trials in this space and we have very slow progression of disease.

I suspect for the patients, post adjuvant therapy and in those who relapse post therapy with good or intermediate risk disease, you might say, well, if they’ve, if those patients have progressed after immune checkpoint inhibitors is unlikely to be a curative setting.

And can we intervene at that point with a really well tolerated agent like this that probably has good activity in any disease. So that’s the trial I’d like to combine those two populations and do a study in that group. That’s what I’d do if I had.

Eric Jonasch:
And I guess the big question there is what’s the comparator agent or a combination that you use to, to actually do that because you obviously, it would be hard to convince any regulatory authority to use [Sunitinib 00:25:12] again.

Brian:
[crosstalk 00:25:13] I guess what I would say is, I think you could do an IO hif versus hif study in those indol patients, right? To your point of, can we get away with a well-tolerated single agent in those patients? I would argue people need IO to get cured.

They don’t necessarily need it up front. Right? You could argue, well, if they fail hif, I can give them an IO agent and they’re potentially still curable, but I think that’s the kind of trial you’d have to do. I just, I don’t think it’ll ever happen.

Tom:
We do need to look at this drug early in the disease process. And the problem is that’s a hard thing because the temptation is to pile in currently with triple therapy.

Brian:
Yeah. But it may, as hopefully investigator-initiated studies get rolled out, you can do some neoadjuvant work for mechanistic reasons like Eric suggested or some of these more interesting sorts of clinical questions that we’re talking about that aren’t necessarily interesting to companies but are clearly clinically relevant. So, we’ll see.

Tom:
Brian, do you want to finish?

Brian:
I think I just did. No. Hey, Hey Eric. Thanks. This is great. Congratulations on all your hif work. I think again, as Tom said, I think the VHL work is practice changing. And this sort of just continues in this story. I think just keeps getting more and more interesting over time and I look forward to seeing where it goes next.

Eric Jonasch:
Yeah. It’s exciting. Thank you for the invitation and really nice chatting with you, guys.