Immune System Impact of VEGF Targeting Under the Microscope

Robert S. Kerbel, PhD

For more than 25 years, Robert S. Kerbel, PhD, a senior scientist at Sunnybrook Research Institute and a professor of Medical Biophysics at the University of Toronto, has been studying tumor angiogenesis and the impact of inhibiting the VEGF pathway as an anticancer strategy.

OncLive: How has our understanding of the role of the VEGF pathway in cancer evolved in recent years?

Although the FDA has approved 10 VEGF-targeting agents in a variety of tumor types, the strategy has proved problematic. In an interview with OncLive, Kerbel, who also formerly served as Canada Research Chair in Tumor Biology, Angiogenesis, and Antiangiogenic Therapy, discussed the evolving landscape of VEGF targeting and some of the unique challenges posed by the strategy.Well, there are a number of things. One is that VEGF was first discovered in retrospect as a vascular permeability factor in the mid-1980s and since then it has been mainly evaluated and studied as an angiogenesis factor. Although VEGF is predominantly considered to be a factor that stimulates endothelial cells to grow, divide, and form blood vessels, it has other functions, and now it’s evolving to some extent into some other interesting therapeutic areas.

How have these changing views affected the use of VEGF-targeting drugs in the clinic or research relating to them?

One that’s beginning to attract much more attention is its possible role in regulating aspects of the immune system. It’s sort of evolving to some extent from an angiogenesis factor to an immune-modulating factor. Whether that will fully pan out, we still don’t know yet. I think a lot will depend on what happens with ongoing clinical studies evaluating, for example, immune checkpoint inhibitors like ipilimumab and nivolumab, and so forth, with drugs like ramucirumab and bevacizumab and some of the antiangiogenic tyrosine kinase inhibitors.I think another interesting thing about inhibiting this pathway is we still don’t really know all that well how it works. By not really having a clear answer to that question, it’s more difficult to make progress.

There are a couple of different mysteries. First, one would predict that if angiogenesis is really an important and significant driver of tumor growth, progression, and metastasis, that if you treat patients with a single-agent antiangiogenic drug only, like bevacizumab, the results would actually be much better than they have been in certain types of cancer such as breast, colorectal, lung, and prostate.

Patients, regardless of what type of cancer they have, rarely get these antibody drugs as monotherapy. They are almost always administered upfront as part of a small cocktail, usually with standard chemotherapy. The prevailing wisdom is that this makes the chemotherapy part of the cocktail work better.

Well, then, that raises the question of how does that actually happen? The first successful reports of this in the clinic came more than 12 years ago and we still don’t know why this type of combination works better than chemotherapy alone—when it does. There are a number of different theories, but really none of them have been validated as being definitively correct in the clinic, and the preclinical data are conflicting.

Another area in which there’s been a great deal of interest, because of the efficacy results, is trying to figure out why the drugs work and then stop working. This is obviously a big problem in oncology with virtually any type of cancer therapy, though immunotherapy in some cases may be an exception.

How has it informed the development of new VEGF pathway-targeting therapies?

With almost every type of targeted therapy or chemotherapy or radiation therapy, the story has been that initially sensitive tumors become resistant over time to these therapies and the same thing seems to be happening with drugs that target the VEGF pathway. So there has been a great deal of interest in trying to understand why that happens and then, by figuring out how it happens, trying to delay the onset of resistance or actually develop some strategies to treat resistant disease.I think most researchers and key opinion leaders, if they were told that Company X or Academic Lab Y is trying to develop another VEGF pathway— targeting drug, my suspicion would be that they would roll their eyes and say, “you’ve got to be kidding.” We have so many such drugs now that are approved, do we really need anything more?

Well, yes, there are some possibilities. The existing drugs are extremely expensive and one possible approach to get around that problem, which would then have a potential significant advantage, would be to induce patients to produce their own VEGF antibodies. There’s one approach being pursued, whereby the VEGF molecule is altered in various ways so it becomes immunogenic in patients—that is, by altering the basic structure of the molecule or conjugating it to a foreign protein so that it provokes an immune response. That could be an interesting strategy, very different from all of the other approaches which involve passive administration of either a synthetic drug or a genetically engineered protein drug. That could have some possible advantages if it works out.

Another possibility is to take something that targets the VEGF pathway, like an antibody for example, and then genetically engineer that protein so that it also disrupts or blocks an additional pathway. These are so-called bispecific or cross-specific antibodies; for example, Roche has been developing antibodies that simultaneously target the VEGF pathway but also block another major proangiogenic pathway and that’s the tie-2/angiopoietin-2 pathway. Until recently, it has been rather challenging to target this pathway specifically.

What are the most important unanswered questions or challenges?

There is no question that, preclinically, simultaneous or sequential targeting of those two pathways seems to bring about a consistent and superior benefit. If this holds up in clinical trials, it’s a potentially interesting strategy for extending and modifying the benefit of VEGF-targeting drugs.Probably the biggest hurdle for the development of antiangiogenic drugs as a class in the clinic is that there has been no validated predictive biomarker to prospectively select patients having a greater chance of benefiting from these drugs. With VEGF pathway—targeting drugs, you would think intuitively that patients who have high levels of VEGF in their tumors or in the circulation that could be detected would do better than patients that have low levels, but unfortunately that has not turned out to be the case, which has been very frustrating.

Investigators have been pursuing many other strategies, which might result in a predictive biomarker. All sorts of things have been attempted such as detecting changes in blood flow in tumors or looking at what are called gene signatures, but nothing yet has been validated.

What that means is that all of the trials that have been undertaken with antiangiogenic drugs have always enrolled “all-comers” within a particular type of cancer—there was no prior selection of patients. Consequently, in a negative trial, it may have been the case that, say, of 500 patients enrolled, only 50 or less were actually truly benefiting from that drug.

So when these drugs went into clinical trials, the chance of detecting an impressive effect, especially in overall survival, is reduced by the inability to somehow select patients beforehand to see which ones should be enrolled in the trial and which shouldn’t.

A second significant challenge is, as I just mentioned, to figure out what the main clinical mechanisms of resistance are. There are numerous theories describing how tumors might become resistant to VEGF—targeting drugs, with a number of really interesting, clever hypotheses. The problem is that almost all of them have come from preclinical studies and it has been difficult to determine whether they actually apply to the clinic, and whether we could do something about it.

The most cited theory of resistance goes like this: you administer the drug, it acts as an antiangiogenic agent, and thus you get a reduction in tumor vascularity, blood flow, and perfusion; that is, it does all the things you want it to do. The problem is that, although VEGF is considered the most important angiogenic factor in most types of cancer, it is not the only one. We now know that there are a very large number, many of which are normally “silent” in tumors.

VEGF is the major operational angiogenesis factor: you give a patient a drug that blocks it, and it works for a while, but then you simply create the evolutionary pressure for activating one of these bypass pathways (like angiopoietin- 2). If this hypothesis is correct, and there is certainly evidence for it from preclinical studies, you could start to give a drug that blocks the other pathway to extend the duration of response.

So far, there have been no meaningful successful attempts utilizing such a strategy in patients. In short, it has been a challenge to identify the major mechanisms of resistance in patients. A third challenge, one that I’ve been working on recently, is to reconsider aspects of the overall complexities of the hypothesis about targeting the tumor vasculature. If you go back and look at the literature that led to the development of antiangiogenic drugs, Judah Folkman’s idea has been widely accepted over the years, but there are still some skeptics around who proclaim that in mice there is little doubt that there is a lot of tumor angiogenesis going on and so if you give an antiangiogenic drug, you often get these really nice responses that are often difficult to recapitulate in patients.

In the mid-1990s, a few investigators started to report that when they examined spontaneously arising human tumors, the extent of tumor angiogenesis was very limited and in some cases wasn’t even detectable (eg, in primary or secondary lung tumors). The reason was that if a tumor mass is growing, say, in the lungs (or the liver, as another example), these organs already have a very rich, abundant supply of preexisting blood vessels. So when a tumor starts to develop in or invade these sites, it can simply “hijack” the normal vasculature in that organ. Hence, there’s a minimal or no need to make new blood vessels.

This phenomenon is called “vessel co-option” and there has only been a very small number of investigators who have worked in this area over the past 20 years, but that’s now changing.

There’s a growing feeling that one of the explanations for the modest clinical effects of most antiangiogenic drugs that have been developed so far could be related to the fact that the degree of angiogenesis that’s actually occurring in tumors is not always that robust. And these drugs targeting the VEGF pathway are programmed to target angiogenesis.

So a major challenge ahead in the field is to determine to what extent this phenomenon happens in patients and, if it is common, is there any way we can target co-opted vessels with a different type of “antivascular” therapy? Based on their leakiness, it is likely that such co-opted tumor blood vessels are not truly normal, and hence potentially susceptible to an alternative antivascular therapy. Perhaps combining an antivascular or antiangiogenic therapy with an immune checkpoint therapy would be one way of doing this, and that is something we are looking into.

And finally, one other important development to note: the endothelial cells of tumor blood vessels can also stimulate growth of tumor cells in a perfusion-independent fashion by secreting a number of paracrine growth factors called “angiocrines.”

Blocking these factors could also represent a new potential form of antivascular therapy.