Read the Full Video Transcript

Alicia Morgans: Hi, I'm so excited to be here today with Dr. David Braun, who is joining me after ESMO 2024, where he presented some really fascinating work in the CheckMate 9ER study in kidney cancer. Thank you so much for being here today, Dr. Braun, and please tell us a little bit about it.

David Braun: Absolutely. Very happy to do so, and thank you so much for having me on today. So I'm delighted to be able to chat with you today about the work that we really did collaboratively—I'm speaking really on behalf of my co-investigators—on trying to understand some novel serum glycoproteomic biomarkers, in this case, in relation to the CheckMate 9ER study.

So we know that the CheckMate 9ER study really helped to set one of the standards of care for advanced kidney cancer, showing that Nivolumab plus Cabozantinib, this anti-PD1-based combination, had benefit over the TKI Sunitinib alone with response, progression-free survival, and overall survival. But I would say generally for immunotherapy and kidney cancer, we don't have effective biomarkers to understand which patients are most likely to derive benefit, and unfortunately, which tumors are most likely to be resistant.

Now, our group and many others have explored a lot about the tumor microenvironment and about tumor mutations. So we've looked at genomics and transcriptomics, and I think this has made some progress in the field, but we also know that circulating factors, host factors, might play a really important role as well. And while DNA and RNA are obviously important, ultimately, the end product is protein, and we know that post-translational modification of proteins really plays a key role in biological function overall, but particularly in carcinogenesis—it really is a hallmark of malignant transformation.

And so understanding altered protein glycosylation, a form of post-translational modification, really is important to understanding questions of response and resistance. Just as an example, one form of glycosylation called hypersialylation really can impact immune responses. It's immunomodulatory and might be associated even with resistance to things like PD1 blockade.

So it's with this background—the idea of trying to understand this relatively nascent field of protein glycosylation—that we really performed this exploratory post-hoc analysis, using pre-treatment serum samples from the CheckMate 9ER study.

And this was using the InterVenn GlycoVision Glycoproteomics platform. And the question was really, "Is there something there? Is something potentially informative with respect to either prognosis or prediction that's really worthy of further investigation, validation, and biological investigation?"

So with that background, we asked a number of questions that I'll sort of summarize. One of the key questions was really surrounding prognosis: that if we ask the question of not an individual glycosylation event, but just the total amount of post-translational modification that takes place—the total amount of glycosylation—and there's many different forms of glycosylation, either sialylation shown on one side of the screen or fucosylation shown on the other, does the total amount make a difference? So if you have more glycosylation, is that better or worse?

And what we can see is that as we move from left to right in these individual boxes, we have an increase in the total amount of glycosylation: either sialylation or fucosylation. And what we can clearly see is that for both Nivolumab and Cabozantinib in blue and Sunitinib in orange, as you increase the number of sialic residues or fucose residues, there's a progressive increase in the hazard ratio—basically, a worsening in outcomes in progression-free survival and overall survival with both treatment arms.

And so overall, this data really suggests a potential prognostic role: that the more glycosylation there is—sialylation or fucosylation—the worse the outcome will be, independent of therapy. So that's something potentially prognostic.

The next question is: are there individual glycosylation events that might be predictive? Meaning, if there's a specific protein that's altered with a glycosylation, is it going to inform on benefit for Nivolumab plus Cabozantinib versus Sunitinib? And again, we examined pretty systematically a lot of proteins and a lot of glycosylation events. And I just want to provide an example of two of them.

So there's, I would say, two big patterns that emerge: that proteins involved in lipid metabolism seem to play a role; but I think largely proteins involved in complement. The complement cascade really seems to have some predictive nature, and I've shown just two clear examples of that here.

The first is C3, or complement protein 3, shown here on the left. And complement protein 3, or C3, is really a pro-inflammatory—and we think of it as potentially an anti-tumor complement factor. And so when glycosylation is low, shown on the left, we really don't see a lot of difference between outcomes—this is progression-free survival—between Sunitinib and Nivolumab plus Cabozantinib. But when glycosylation levels are high, here's where we really see a benefit for Nivolumab plus Cabozantinib over Sunitinib.

And so the hypothesis here, the biological hypothesis, was that having more glycosylation really amplifies its function, amplifies the anti-tumor role of the complement cascade, and may allow for benefit from an anti-PD1-based therapy.

If we see on the right, we see the opposite pattern with complement factor H. Now, complement factor H is a negative regulator of the alternative complement cascade, and so in this case, when the levels are high—so when you have this inhibitory factor that's blocking alternative complement activation—we really don't see benefit from Nivolumab plus Cabozantinib. But when this factor is low, allowing alternative complement pathway activation, then all of a sudden we see a benefit for Nivolumab plus Cabozantinib.

And so the biological hypothesis here might be that essentially, the complement factor H is acting as an innate immunologic checkpoint, and so when the levels are high, it's blocking effectiveness of an immunotherapy; when it's low, it's allowing that immunotherapy to have a benefit.

And so overall, I want to really emphasize that this is an exploratory analysis, that there's not a lot of glycoproteomics work done in the biomarker space, and that the overall question was, "Is this something worthy of further investigation?" I think the answer to that is clearly yes. There's really some interesting signals, but they're preliminary, and we have to really note that because this is an exploratory post-hoc analysis.

There are factors that really seem to be associated with prognosis, and that's essentially the total amount of glycosylation—fucosylation and sialylation. But also, there seem to be individual factors related to the complement cascade and, not discussed today, but lipid metabolism, that have the potential to maybe even be predictive for benefit of Nivolumab plus Cabozantinib versus Sunitinib.

And so it really is next steps. We have to explore, are these real? We need to really perform vigorous validation, external validation; and if it is, then really try to understand the biology of these factors: what's driving it to impact response that way? Thank you so much.

Alicia Morgans: Well, David, that was really, really interesting, and I commend you and the team for actually being able to make that collaboration happen within a registration trial. I think that the dedication of the team and then of course the sponsor is very clear, because it's not always possible to do such innovative science in that kind of a situation or setting, so congratulations, and of course, congratulations to the patients involved.

Now, I want to dig in a little bit. We'll get to glycosylation in a second, but I think that the findings around complement are particularly intriguing, and that's not necessarily something I would expect in most tumor types. There's sort of a distinct phenotype that we often see with renal cell carcinoma causing clotting to happen, causing that to be a problem through the renal vein and other places potentially.

Do you think that some of the effects on complement may be leading to some of the thromboses that we see? Is this something that's specific to renal cell carcinoma or is it something you expect to see across other solid tumors?

David Braun: It's a really good question, and I think underlying that is this idea, as you mentioned, that renal cell carcinoma really has a pretty distinct biology—immunobiology—compared to a lot of other solid tumors. It's always been sort of the outlier when it comes to immunotherapy responses. It doesn't have a lot of mutations like melanoma or lots of non-small cell lung cancer, but here it is, being historically and contemporarily an immunotherapy-responsive tumor.

And so I think we don't know yet whether this is going to be more generalizable. I think there's at least some initial hints to suggest it might be worthwhile to look at it in other tumors. There's been a similar study led by Dr. Genevieve Boland's group looking in melanoma and using that same InterVenn GlycoVision platform, looking at serum glycoproteomics in relation to immune checkpoint inhibition in melanoma.

And actually some of these findings were very, very similar: higher rates of sialylation and fucosylation really being associated, for instance, with worse outcomes with immune checkpoint inhibitors. So I think it at least provides some hope that there might be something more generalizable, but given the sort of unique immunobiology of kidney cancer, I think we really need to validate that.

With respect to complement, I think it's an interesting question. I think the true answer is me, and I think a lot of people in the field haven't thought nearly as deeply as probably we should have about complement, because I think a lot about CD8 T cell-mediated immunity, but we know there's entirely other branches of both adaptive and innate immunity that's really important for tumor control.

And as we've been sort of picking this apart over the years, we're starting to see roles for B cells, for plasma cells, for all these different branches of the immune response. And so I think this is really opening the door to saying, "Actually, this pathway—the complement cascade—could actually be extremely important, as well."

We need to validate first, make sure these results are robust, but if they really do hold up, I think that's something that really does require further investigation.

Alicia Morgans: It really is fascinating. Now, one other thing I think I saw on the slides, but you'll have to correct me if I'm wrong, is that it looked like although the progression-free survival and the overall survival were associated with the levels of glycosylation in both settings, the response rates did not necessarily seem to be as tightly tied to that, and I wonder if you have thoughts on that or if that was a misunderstanding on my part related to the figures.

David Braun: No, I think that's absolutely right. There wasn't a very, very clear ordering of response, and I think part of that is these are decent-sized cohorts to start, but the amount of biological material we have is only a subset of the total cohort. And then when you begin to sort of slice and dice it in various ways where you're now only taking the subset of patients that have one sialic acid residue or two or three or zero, we get to actually very, very small numbers.

And so I think it's hard to know whether there really is a lack of association or just a lack of power, but I think that is a very astute point. I think it's something that, if it does hold up in subsequent validation, is a little bit perplexing. Maybe it's something that doesn't affect things like shrinkage, which is response, but might affect durability—so someone who has a stable disease, but the length of time the disease is stable really impacts progression-free response, but won't show up necessarily in objective response criteria. So I think really interesting point. It's something that we need to look into further.

Alicia Morgans: And to your point, the immune response is different than our traditional RECIST measurements, and response rates certainly is based on historical shrinkage of tumor, and we don't necessarily always see that. And so just an interesting thing that, again, supports that judging response in immunotherapy is something that I think, as a field, from a clinical perspective, we still need to work on.

But not your charge at this point: you're enlightening us with your laboratory savvy. So as you think about all of this and you imagine where it might go, give us some forward-looking ideas about where these kinds of analysis may take us in next steps.

David Braun: Absolutely. So I think there's some really tangible and practical next steps. I think this is some exciting early signal that there might be something there. I think there's a lot that's appealing about this. It's something that is relatively straightforward to measure, meaning it requires serum, so it's a relatively simple blood measurement, doesn't require invasive action.

Because it's the serum as well, it's something that essentially captures the heterogeneity of cancer, meaning we know that when we look at a single biopsy from a single site of metastasis or a single primary tumor, that might not be reflective of other sites of disease, and having a circulating factor, like a proteomic measurement, might actually overcome some of those limitations.

But I think really a key next thing is external validation. I think this is an initial signal that's really interesting and exciting, but we have to make sure this is real. We have to make sure we test systematically and confirm that the signal is true. I think we have some early indications which are promising.

I didn't show this data, but the analysis was actually run in multiple independent batches, so this is an internal validation. But at least internally between independent batches, it was really a signal that held up, but now we have to make sure to externally validate. That's the first thing: really making sure it's true and validates.

If it does, I think that the steps after that are really wide open and exciting, which is: what is the biology of this? So there's the aspect that's a practical component, which is: could this be used as a biomarker? Could this help select the right drug for the right patients—the ones most likely to benefit, the ones least likely to benefit from a particular regimen? But then, how can we understand what's actually happening at a molecular level that's actually impacting the response? It's actually impacting whether a tumor is more or less likely to respond. And how might that actually be modifiable in a way?

I think that's really exciting because it opens up the idea, at least, to new therapeutic avenues as well. So I think that lots of promise in the future, but I think the hard work of validating is sort of the first step.

Alicia Morgans: Wonderful. Well, I sincerely appreciate your time today and congratulations again to you, the team, and of course the patients for participating in this work that is quite exciting and hopefully will continue to move the field of renal cancer and perhaps others forward as we consider immunologic approaches. Thank you so much.

David Braun: Thank you again. I really appreciate it.