Judah Folkman, MD and Rakesh K. Jain, Ph.D., Harvard Medical School, Boston, MA
David J. Anderson, Ph.D., California Institute of Technology, Pasadena, CA
Peter Carmeliet, MD, Ph.D., Centre for Transgene Tech and Gene Therapy, Leuven, Belgium
Gullermo Garcia-Cardena, MD, Brigham and Women’s Hospital, Boston, MA
Robert S. Kerbel, Ph.D., Sunnybrook Health Science Center, Toronto, Canada
Daniel Linzer, Ph.D., Northwestern University, Evanston, IL
Donald M. McDonald, MD, Ph.D., University of California, San Francisco, CA
Renata Pasqualini, Ph.D., The Burnham Institute, La Jolla, CA
Brian Seed, Ph.D., Massachusetts General Hospital, Boston, MA
George D. Yancopoulos, MD, Ph.D., Regeneron Pharmaceuticals, Inc., Tarrytown, NY
Layout and Goals
- New Phenotypes of Tumor Angiogenesis
- Role of Host Microenvironment
- Characterization of Tumor Vessels
- Endothelial Control of Tissue Mass
Chemotherapy is the mainstay of treating tumors which have spread throughout the body. However, all cancer drugs, while potentially benefiting the patient, cause side effects which can limit their use. Doctors continue to strive to develop new approaches to destroy cancers without damaging normal tissues. This was the topic of our 1996 Forum, but it did not address the topic of drug delivery. More specific drugs and ways of delivering them selectively to tumors in the body could be the answer to treating many cancers. We know that the problems of delivery are formidable. Tumors need a blood supply to grow which should make the delivery of drugs simple. Unfortunately, cancers are very good at keeping compounds out which can readily accumulate in other organs. If one takes this argument to an extreme, one can have the most selective way of treating cancers imaginable, but if you cannot deliver the material to the tumor cells, it will not be effective. Ways to address this problem will be the focus of the 1998 Forum. The two chairmen are at the forefront of this research area; Dr. R. Jain who discovered that the fluid pressure inside tumors is higher than in normal organs and Dr. Judah Folkman, who described the ability of tumors to produce factors which encourage the growth of blood vessels, allowing cancers to maintain a good supply of nutrients.
Dr. Folkman opened the meeting by sketching a list of 15 central questions and interesting problems in the field of angiogenesis research. George Yancopoulos presented his discovery of a new family of angiogenesis stimulators, the angiopoietins. He showed that angiopoietin-1 may work together with another angiogenic factor, VEGF, to build normal new blood vessels in the embryo and possibly in wounds. Furthermore, a related member of this family, angiopoietin-2, appears to be made by the endothelial cells which line the blood vessels coming into the tumors. Angiopoietin-2 appears to instruct new tumor vessels not to coat themselves with an outer layer of smooth muscle cells, as is the case for new vessels in wound. This leaves tumor vessels with thin walls. Dr. Yancopoulos made transgenic mice which over-produced angiopoietin-1 in their skin. These mice produce extra vessels with large diameters, beneath the skin. The mice had red skin. This proved that angiopoietin-1 is a new member of the family of angiogenic stimulators, but more importantly, the vessels did not leak blood or reveal microbleeding, as was the case with another angiogenic factor, VEGF.
A new idea emerged in the discussion when Dr. Yancopoulos mentioned unpublished data that platelets, the smallest of the circulating cells in the bloodstream, carry very high levels of angiopoietib-1 (10,000 times more than VEGF). Judah Folkman suggested that angiopoietin-1 could be a long sought missing ‘hormone’ that has been thought to be carried by platelets and that acts as a continuous ‘sealant’ for the inside lining of blood vessels. For example, patients who have very low platelets from infection or from cancer therapy, have bleeding and oozing which is stopped by giving a platelet transfusion. While current dogma is that platelets plug leaks, this has never been proven. An alternative theory is that a platelet-derived protein may continuously repair microscopic holes in capillaries. However, such a putative protein has not been discovered. Thus, with several pieces of information from different scientists during the discussion, a completely new idea emerged that could be tested during the coming year.
David Anderson of Caltech described his discovery of a new set of proteins, the ephrins, hat specify in the embryo which blood vessels will become arteries and which will become veins or capillaries. This is the beginning of unraveling a mystery that has been around for a long time. All of the blood vessels, their branches, and arteries, veins, and capillaries are in place in the embryo before the heart starts to beat. The ephrins seem to be the guiding molecules for this map of the vascular system. It will be of great importance to determine whether tumors also use these molecules when they recruit their own private blood supply.
Donald McDonald of the University of California San Francisco developed a new method that enabled him to examine the inside lumen or hollow part of the smallest tumor blood vessels under very high magnification. These beautiful photographs, for the first time, gave viewers the feeling that they were walking inside of a large tunnel. However, instead of being lined with smooth shiny thin endothelium as would normal vessels, the tumor vessels looked more like the insides of caves with stalactites on the ceiling and the walls. Dr. McDonald was able to see holes develop, that could be the cause of leaky tumor vessels. He was also able to see the endothelial cells behave in a wild and unorganized manner unlike any normal vessels. Furthermore, there seemed to be a second type of intruding cell which made up perhaps 10-20% of the lining cells. Dr. McDonald felt that some of these intruders may be invading tumor cells that take up a place in the lining of the blood vessels. This was the first time that any of us at the meeting had ever seen such a view directly inside of tumor vessels at high power. Once these are published other scientists in the world will have the same exciting experience we did. In the discussion of McDonald’s paper, another new idea emerges.
David Anderson asked whether the jumbled, tangled and disorganized endothelial lining of the tumor blood vessels could be the result of lack of nerves in tumors. It is known that small nerves normally follow vessels everywhere into normal tissues, but do not go into tumors. Dr. Anderson suggested the novel idea that perhaps certain molecules in nerves are able to put the brakes on growing endothelium. Thus, any absence of these nerves in the lining cells of tumor vessels would be unrestrained. Anderson suggested some experiments that could answer this question. He also pointed out that he was new to the field of nerve growth and that this conference stimulated him to think of an idea that would otherwise not have occurred to him.
Dr. Rakesh Jain of the Mass General Hospital grew prostate cancer in male mice and then castrated them to lower the level of testosterone. The tumor stopped growing and shrank after castration. He showed that new capillary blood vessels in the tumor regressed and disappeared before the tumor regression began. He concluded that tumor regression is dependent withdrawal. After several weeks, the residua dormant tumor cells escape their dependence on testosterone. They can now resume production of the angiogenic stimulator, VEGF. This induces new blood vessel growth which permits regrowth of the prostate cancer. This result may explain why in men with prostate cancer that after two years of suppression of testosterone (e.g. by Lupron therapy, or by castration), it is not uncommon for the tumor cells to become independent of testosterone and to resume their angiogenic activity leading to tumor recurrence.
Dr. Jain further showed that if a tumor employing mainly VEGF for its angiogenic activity was treated with antibody that blocked VEGF, the subsequent tumor regression would only be temporary (i.e. several weeks or a month). In some mice, tumors recurred and their cells were making a different angiogenic stimulator, bFGF. This finding has important implications for antiangiogenic therapy and suggests that if an angiogenesis inhibitor is targeted against only one angiogenic factor the tumor may eventually escape unless it is eradicated the first time. Therefore, it may be important to use angiogenic inhibitors against one or more different angiogenic factors. Furthermore, a second type of angiogenesis inhibitor that turns off endothelial cell response to a wide range of angiogenic stimulators may be more effective and this is the type of inhibitor represented by angiostatin or endostatin or vitataxin or TNP-470 and some others.
Daniel Linzer from Northwestern University showed that the placenta makes its own angiogenic factor early in pregnancy. This is called proliferin. Towards the middle of pregnancy, the placenta gradually turns off proliferin and turns on an inhibitor of blood vessel growth called proliferin-related protein (because it has a structure which is only slightly different from this stimulator). One of the functions of the inhibitor protein is to prevent the mother’s vessels and the fetal vessels from invading or overlapping each other so the two circulations remain separate. This prevents the mother from becoming immunized by the baby’s cells under normal conditions. Both of these proteins are being found in certain tumors. This suggests that these tumors may be reactivating older genes that were used in the placenta and then were turned off.
Renata Pasqualini from the Burnham Institute in La Jolla, CA and her colleague (Erik Ruoslahti) have made the important discovery that small proteins (peptides) when injected into the circulation will bind specifically to the lining endothelial cells at different ‘addresses’. They have found that essentially a ‘zipcode’ allows a protein to be targeted to any organ in the body. Also a protein can be sent only to growing vessels in tumors and not to other vessels. She showed that a standard chemotherapeutic drug could be directed only to the blood vessels of a tumor and not to the rest of the body. This lowered the toxicity of the drug and increased its efficacy by optimizing the drug for its antiangiogenic activity. Tumor regression resulted. She also showed that the drug could be directed just to the prostate and would completely shrink it in mice without causing any other damage. This discovery has enormous implications for the pharmaceutical industry because it may in the future be possible to send a drug to any part of the body without flooding the rest of the body and incurring ‘side effects’.
Guillermo Garcia-Cardena from the Harvard Medical School is the co-discoverer of the finding that leptin is an angiogenic factor. Leptin was previously discovered to be a protein made by fat cells. It enters the circulation and tells the brain to reduce appetite. When fat accumulates, fat cells start to grow and may use leptin to stimulate blood vessels along with other angiogenic proteins. It has been estimated that a pound of fat contains a mile of capillary blood vessels. In unpublished data, Garcia-Cardena has recently found that certain tumors over-produce leptin abnormally and that this may explain the severe weight loss and cachexia in many cancer patients. Thus, it could be possible to design a drug that would interfere with leptin and stop the severe weight loss and return the appetite of cancer patients. The discussion of this paper led to a completely new idea suggested by Brian Seed. He noted that in patients with the ‘late onset form of diabetes’ there are very high circulating levels of leptin which comes from the excess fat that these patients have. However, apparently the receptors in the brain that can detect leptin are missing or are abnormal, so that the brain does not know to shut off the appetite. The patients also have severe angiogenesis in their eyes in the retina and this causes blindness. He suggested that perhaps the excess levels of leptin are driving the angiogenesis in the eye. Therefore, the development of a novel drug which could interfere with these leptin levels could perhaps inhibit the eye angiogenesis and thus prevent blindness. This most interesting idea would probably not have occurred to any individual scientist if they had not been in the same conference room together.
Dr. Chenggang Li, a Forbeck Scholar from the University of Manchester Medical School in the United Kingdom showed a new addressing marker on endothelial cells and this particular protein is increased on endothelial cells that are growing. He was able to send a specific antibody to this address and the antibody is currently just starting to be used as a method of antiangiogenic therapy in cancer patients. Judah Folkman presented two new concepts and experimental data to support them. The first is that proteins of the clotting system may share in the regulation of angiogenesis. There are at least three types of this interaction: (i) Platelets carry or produce angiogenic stimulators such as VEGF, angiopoietin-1 and TGF as well as angiogenic inhibitors thrombospondin and platelet factor-4; (ii) Certain internal fragments of clotting proteins inhibit angiogenesis. Example are, angiostatin an internal fragment of plasminogen, and domain 5, an internal fragment of high molecular weight kininogen (unpublished data from Temple University); (iii) A change in molecular confirmation of antithrombin III inactivates its anticoagulant activity, but activates its antiendothelial and antiangiogenic activity. The latter was discovered recently by Michael O’Reilly in the Folkman lab when it was found that a human lung cancer made an enzyme that initiated this change in molecular confirmation. This finding has been confirmed by Oliver Kisker and Steven Pirie-Shephard in another human tumor (pancreatic cancer) in the Folkman laboratory and also a third finding by Jie Lin in the Folkman laboratory. The antiangiogenic form of antithrombin III without its anticoagulant activity is currently being manufactured for clinical trial as an angiogenesis inhibitor.
A second new concept is that endothelial cells not only control tumor growth, but they may control the growth of normal tissue mass. Folkman developed this concept from experiments in which the liver is partially removed in animals and it grows back to normal, and stops at normal size. However if an angiogenesis stimulator is injected in the animals the liver overgrows, while if an angiogenesis inhibitor is injected, the liver is significantly reduced in size. Furthermore, in a model of fat growth in leptin-null mice who continuously eat because they do not have the leptin gene, the large increase in fat which grows 22 times faster than any mouse tumor, can be regressed rapidly by giving either of 3 types of angiogenesis inhibitors (TNP-470, angiostatin or endostatin). In these experiments the rapid loss of fat is suddenly reversed when the endothelial inhibitors are discontinued, but this is unrelated to any major changes in food intake. Both new concepts may enlarge our understanding of how to administer antiangiogenic therapy in patients.
Dr. Robert Kerbel from the Sunnybrook Health Science Center in Toronto, Canada showed that in a growing tumor that tumor cells hugging capillary blood vessels could be separated from tumor cells growing at a distance from blood vessels. This was accompanied by using a nuclear stain that could distinguish cells that came from a low oxygen environment versus a high oxygen environment. The provocative result was that the tumor cells most remote from the capillary vessel were more growth proficient when cultured in vitro or as spheroids. Thus, tumor cells in vivo undergo a selection process within a tumor according to their capacity to live at different levels of oxygen. Dr. Kerbel also reported that tumor cells that are never tumorigenic in vivo and are not angiogenic per se can be mixed with similar tumor cells that have been transfected with VEGF121 or VEGF165, but not VEGF-C. Tumors grow and they contain both types of tumor cells. This means that a subpopulation of tumor cells that is angiogenic can recruit blood vessels which will support other tumor cells that are not themselves angiogenic.
Brian Seed of the Massachusetts General Hospital showed in elegant experiments carried out in collaboration with Rakish Jain that certain tumors can induce neighboring normal host cells to upregulate expression of the angiogenic factor VEGF. This result was obtained by using transgenic mice in which the VEGF promoter was hybridized to green fluorescent protein. This important finding suggests that the angiogenic switch may not only involve the elaboration of angiogenic proteins from tumor cells per se, but that tumors may secrete another unknown mediator which induces the neighboring host cells to release their own angiogenic activity. This new angiogenic mechanism would augment the total level of angiogenesis in a tumor bed. Peter Carmeliet from the Center for Transgene Technology and Gene Therapy in Leuven, Belgium produced mice in which the gene for hypoxia inducible factor-1 (HIF-1) was knocked out. They died during embryonic development. He then obtained embryonic stem cells from these mice. These cells were also deficient in HIF-1. When these cells were injected into animals, tumors results (teratocarcinomas) that had normal or low levels of VEGF production, as would be expected without the hypoxia inducible system that increases expression of VEGF. While the tumors were angiogenic, they appeared to be less angiogenic than tumors from embryonic stem cells expressing HIF-1. Tumor cell apoptosis was lower in these large tumors. These results were unexpected, but suggested the novel possibility that HIF-1 may be involved in the control of tumor cell apoptosis.
The two days of the Forbeck Forum on angiogenesis were very exciting and productive. At least 4 new major ideas emerged that had never been considered before. Several new collaborations were formed. The results of many experiments improved our understanding of how to develop and discover new angiogenesis inhibitors. Furthermore, from the meeting, it gradually became clear that when angiogenesis inhibitors enter into the clinic that they could be added to chemotherapy or added to radiotherapy or added to vaccine therapy, immune therapy or gene therapy. Finally, the many results discussed at the Forbeck Conference provided many new insights into the angiogenic process.