How did you first get started working on angiogenesis and integrin a _vb ₃, the subject of your highly cited 1994 Science paper (PC Brooks, RAF Clark, DA Cheresh, "Requirement of vascular integrin a _vb ₃ for angiogenesis" Science 264[5158]: 561-71, 22 April 1994)?

"The goal is to make a smart bomb for the tumor endothelium, and to increase the efficacy of agents by directing them straight to a vb 3."

It all started back in 1984, which is when I first went to work at Scripps. I made a monoclonal antibody to a ganglioside that blocked melanoma cell adhesion in vitro to the extracellular matrix. We then started collaborating with Erkki Ruoslahti, who identified the first fibronectin receptor and vitronectin receptor, which were later incorporated into the family of proteins now called integrins. In collaboration with his laboratory we determined that melanoma cells had a huge amount of the vitronectin receptor and very little fibronection receptor. So we generated a monoclonal antibody to the melanoma cell vitronectin receptor which later turned out to be known as integrin a _vb ₃. At that point we began a study looking at melanoma and cancer biopsies as well as normal tissues and we were very pleased to see that all melanomas had high levels of this. But we also noticed that the blood vessels of all of the tumors did as well. Initially, that was a little disappointing and naturally we became much less interested in the fact that this was expressed on melanoma as a tumor-specific antigen. Upon further analysis we determined that this antibody, termed LM609, recognized angiogenic blood vessels but not normal ones. This prompted me to ask whether this antibody might somehow block angiogenesis. Thus, I made plans to visit the laboratory of Judah Folkman. He was doing pioneering work with angiogenesis. He showed me how to do a certain kind of experiment where you take a three-day-old chick embryo, crack open the egg, put in a test subject—in this case, our LM609 antibody, which bound to the integrin a _vb ₃—and evaluate whether or not it causes a zone of clearance in the blood vessels. At least one of the interpretations of a positive result is that you have an anti-angiogenic agent.

Well, it’s called a chick-CAM test. "CAM" stands for "chorioallantoic membrane," which is what surrounds the embryo. It is a highly vascularized tissue where the oxygen exchange takes place during development. It is a very popular membrane on which to evaluate the growth of blood vessels because it is so highly vascularized. We did this test on a three-day-old chick embryo with LM609 and we had good news and bad news. The good news was that there were no blood vessels on the CAM. The bad news was all the embryos were dead 24 hours after administering the antibody. So I assumed I had a very toxic antibody that reacted with a number of embryonic tissues. At that point I let the model drop.

In 1992, Peter Brooks came to the lab as a post-doc. He had been working with chick embryos, and so we decided to reinvestigate that whole idea. But he had been doing studies on 10-day-old chick embryos, looking at tumor cell invasion. And so we got a developmental biology book about chick embryos, and learned that blood vessels stop growing by day 10. So we decided to repeat the original studies with 10-day-old embryos. We reasoned that maybe the chicks needed integrin a _vb ₃ for development and maybe that’s why the antibody killed them in the previous CAM studies. We reasoned that by day 10, blood vessel development was finished and thus did not depend on this integrin. So we induced new blood vessels with a growth factor, and determined whether LM609 could block this process. In essence this was one of the primary results we reported in Science in 1994.

The antibody blocked only the new vessel growth and had no effect on existing blood vessels. Also, when we placed a tumor fragment on the CAM and injected the embryo with the antibody, it effectively starved the tumor, stopping its growth while not influencing the growth of the embryo. This was the first indication that any kind of adhesion receptor might be involved somehow in angiogenesis.

Later that year, we published another paper in Cell, which is also very highly cited although it doesn’t show up on your list (PC Brooks et al., "Integrin a _vb ₃ antagonists promote tumor regression by inducing apoptosis of angiogenic blood vessels," Cell 79[7]:1157-64, 30 December 1994). That one turned up on the front page of the New York Times and in Time and many other papers. We showed that the inhibitors of integrin a _vb ₃ also could actually cause the regression of preexisting large tumors, and could do so by causing apoptosis of the endothelial cells. And we reasoned that the receptor, a _vb ₃, is involved in endothelial cell survival during angiogenesis. It provides the survival signal for neo-vessels as they are growing in and toward the tumor. The combination of those two papers got us very entrenched early on in the angiogenesis field.

That antibody has now been humanized and is in fairly middle to late-stage clinical trials, not only for cancer but also for arthritis and psoriasis. At the same time, a small molecule we helped develop that binds that receptor is also doing well in cancer, including late-stage brain cancer, where it’s showing some efficacy. The safety issues are essentially nonexistent—that is to say, these drugs are nontoxic.

Well, we were aware this was the first time anybody had shown or even suggested that you could regress tumors. That’s why once the news media got hold of it and it appeared on the front page of the New York Times we had an inkling of what would happen. Then Judah Folkman invited me to give a talk at Harvard. In many ways, we had supported what Folkman had predicted would happen. With the right angiogenic agent, you would expect to see such a dramatic effect. Under the appropriate conditions and under certain circumstances with the right agent at the right time you can see this kind of major anti-tumor activity. Since then we’ve seen some evidence for this kind of effect in people. Not in all people, and not in all tumors, but certainly there are some that seem to respond very significantly to this agent.

Yes it is, but in addition we’ve taken a bit of a tangent. In the course of our studies on the signal-transduction events associated with how growth factors and cell adhesion receptors stimulate blood vessel growth, we made the discovery that certain key enzymes, kinases, play a role in the actual biology of endothelial cell function. One pathway had to do with vascular permeability, and so we’ve taken a big jump into this area. Vascular permeability is a unique outcome associated with the growth factor VEGF, which is the only angiogenic growth factor that induces permeability—leakiness—in blood vessels. We discovered that vascular permeability could be turned off selectively by using a Src-kinase inhibitor and that mice deficient in Src were resistant to the ill effects of stroke, because they didn’t have this leakage response. They had no edema following the stroke. So if you either genetically knock out Src or give a normal mouse a Src inhibitor, and then give it a stroke, you will effectively preserve the brain of that animal. We just published a paper showing that the same thing happens after a myocardial infarction—you preserve most of the heart muscle following this ischemic injury by simply blocking vascular permeability. So you can imagine this has dramatic effects on the amount of heart damage and the ultimate survival of those animals. We’re only a month or two away from initiating a clinical trial on this Src inhibitor. This drug would be administered when the heart attack patient comes into the emergency room, and you have up to six hours following to get the drug on board to produce a protective effect. So far we have seen this effect in pigs, rats, and mice, and the toxicity is essentially non-existent. By blocking the edema response within the first several hours of an acute event, we can prevent much of the injury that usually ensues.

Well, we’re looking to understand how these inhibitors work on a molecular level. We are attempting to investigate how genetic or pharmacological blockade of the permeability pathway functions within the intact endothelium. What are the specific mechanisms of action? We’re also focused somewhat on vascular targeting. Going back to the original 1994 Science paper and LM609, we’ve developed a strategy to target genes and target drugs to these new blood vessels. So we’ve come up with a method by which we can eradicate tumors very effectively by targeting the suicide gene to a _vb _3, so it takes it directly into the blood vessels of the tumor and causes those blood vessels to die. We published that in Science in 2002 (JD Hood et al. "Tumor regression by targeted gene delivery to the neovasculature," Science 296[5577]: 2404-7, 28 June 2002). We want to further develop that area and target not only suicide genes but drugs, radionucleotides, etc. to neovascular tissue. The goal is to make a smart bomb for the tumor endothelium, and to increase the efficacy of agents by directing them straight to a _vb _3.

David A. Cheresh, Ph.D.
The Scripps Research Institute
La Jolla, CA, USA