In my talk, I show this banana—and it is kind of a joke—but if you think about the shape of a banana, this is what we are seeing in the PFS curves. At the beginning, everyone is the same, then there is a split while they are on the treatment, and then the split comes back together. The tails of the curves collapse, but the middle of the banana is still there; there is an effect that gets lost over time. We hypothesized in the past that maybe it was just because we pulled the drug away; some of the studies were limited exposure to the antiangiogenesis drug. However, even when we were allowed to use those drugs to natural progression, we still saw the same banana effect. The GOG-262 trial is a very good example of that kind of strategy in the frontline setting.
Why is this happening? We created this animal model in immuno-competent animals to basically recapitulate that. We treated all of these animals, and some of them developed resistance to the targeting while others were sensitive and stayed sensitive. We looked at the tumor at the beginning, middle, and at the time of resistance, and we found that one of the factors driving resistance has to do with the infiltration of an immune cell called macrophage. We identified that these patients, who have an emergence of adaptive resistance, are actually trafficking macrophages from the bone marrow into the microenvironment where there are stimulating factors of angiogenesis escape.
We have an immune response to an antiangiogenesis strategy. That is one of our focuses in a clinical trial that we have launched, which is actually targeting these macrophages at the point when we think they are trafficking into the tumor. The way that the trial is set up is that everyone receives standard chemotherapy plus bevacizumab, but the patients who are not responding after the first 2 cycles of therapy are actually then randomized to continuation of treatment versus continuation of treatment plus a CSF1R-inhibiting drug. Hopefully, we will identify if this particular mechanism can be exploited.
However, there are many others. We understand that there are regulatory factors on natural angiogenesis. For instance, vasohibin-1 (VASH1) is a microenvironment molecule that is controlled by the transcription factors and by EZH2. It just so happens that EZH2 inhibitors are starting to emerge on the scene. Therefore, EZH2 is an inhibitor of VASH, which is a natural angiogenesis inhibitor. By targeted EZH2, we release the restriction on VASH1, and potentially induce a resensitization to VEGF therapy. How do we make bevacizumab better? This is a way. Identifying the patients who are ultimately developing these resistance or escape mechanisms will help us learn how to target.
Is there potential to combine bevacizumab with immunotherapy agents or PARP inhibitors?
Right now, our excitement is around the immunotherapy and PARP inhibitor spaces. Both of those additional drugs are being used in combination with anti–VEGF-based strategies. Those are currently ongoing. In fact, even the National Cancer Institute is running 2 trials; one is in platinum-resistant patients and one in platinum-sensitive patients and is using the combination of an antiangiogenesis drug with a PARP inhibitor. The idea is that the antiangiogenesis inhibitor induces local hypoxia and hypoxia can then lead to homologous recombination (HR) deficiency. You can take an HR-competent tumor and make it incompetent via hypoxia, and then treat it with a PARP inhibitor.
Is there anything else you would like to add?
As we have learned more and more about what is going on in the microenvironment with angiogenesis, we have learned that there are many growth factors. That is what led to the development of things such as FGF inhibitors. We know that there are other elements within the microenvironment such as pericytes, which are supportive cells for endothelial cells, and those can be targeted. Therefore, we are looking at the endothelial cells as potential primary targets with vascular-disrupting agents and anti-VEGF therapies; we are looking for cytokines, we are looking at targeting the pericytes that may make the endothelial cells more vulnerable, and we are looking at the immune escape mechanisms.
One of the things that has always been problematic for us is that VEGF is a natural growth factor that is responsible for our own wound healing, so one of the side effects is the impact on normal processes. We have never been able to completely document why we see these catastrophic complications, such as bowel perforation, in our patients with ovarian cancer who are treated with this class of drug. One of the hypotheses is that microtrauma happens in us all the time. If you inhibit microtrauma in compromised patients, that microtrauma can lead to localized infection, and ultimately, to viscous disruption and catastrophic bowel perforation. We see them, so we know that they occur.
We were really interested in seeing if there was a way to distinguish tumor angiogenesis and normal angiogenesis. We did a series of experiments with my colleague Dr Anil Sood from The University of Texas MD Anderson Cancer Center, and we tried to distinguish factors that were driving tumor angiogenesis in patients, but were missing in the normal angiogenesis.