The discovery of angiogenesis inhibitors: A new class of drugs: Transcript: Part 3
Judah Folkman:
31. Heterogeneity of Angiogenesis |
Now a surprising finding from recent work in our lab is that after a tumor has become angiogenic, tumor cells that are not angiogenic remain in the tumor. In other words, many human and animal tumors appear to contain two populations of tumor cells, angiogenic and non-angiogenic. What is the evidence for this? It is well known that many human tumors taken directly from the operating room fail to "take" when implanted into immunodeficient mice (such as SCID mice). There has not been an adequate explanation for the high incidence of "no take" of human tumors in mice, except for the finding by Josh Fidler of Houston, that certain human tumors such as colon cancer will not grow well under the skin of mice but will grow well when implanted into the colon, i.e., in their normal "orthotopic" position. Fidler further found that the normal skin contains high levels of interferon-beta, an angiogenesis inhibitor for those tumors which rely on production of bFGF to be angiogenic.
Achilles, however, employed tumors which also produced VEGF, and implanted tumors at a variety of different sites. He implanted tumor cells or explants into a large number of SCID mice and observed them for more than 160 days. Because each mouse day is equivalent to approximately 35 human days, this would be equivalent to about 15 human years.
This slide shows the results of one of his Achilles' experiments (in the Jul 18, 2001, Journal of the National Cancer Institute). Only a few tumors grew by about 30--40 days. In the majority of animals, however, there appeared to be "no take." But, when the skin was opened virtually all of the "no take" animals revealed the presence of a tiny white tumor of less than 1millimeter diameter that was not angiogenic and was not growing. Histology showed proliferating tumor cells balanced by dying tumor cells and the absence of capillary blood vessels. Thus, an additional explanation for "no take" is that a relatively large number of tumor cells in an angiogenic tumor are non-angiogenic. Unless numerous samples of a human tumor are implanted there is a high probability that mainly non-angiogenic tumor cells will be implanted. Currently in our lab we are more interested in
tumors that don't "take," i.e., that are not angiogenic, than tumors which grow easily.
In this slide, Taturo Udagawa a post-doctoral fellow in Robert D'Amato's lab in our department, isolated a human osteogenic sarcoma that is non-angiogenic despite repeated passage. Fortuitously, it secretes a hair growth factor, which allows its location under the skin to be identified by a tuft of hair. This is new, unpublished data.
Histology of these barely visible tumors shows them to be avascular and thin (247 microns thick). This is consistent with size restriction by the oxygen diffusion limit of an in situ tumor sandwiched between two layers of quiescent microvessels in the subcutaneous space.
The tumor cells are undergoing proliferation and apoptosis. In very recent unpublished experiments, Taturo has been able to switch these (and other non-angiogenic tumor cells) to the angiogenic phenotype by transfecting them with the ras oncogene or with the gene for VEGF. Rapidly growing tumors result. Furthermore, he has transfected the non-angiogenic tumor cells with green fluorescent protein. When they are mixed with angiogenic tumor cells, Taturo can prove that the two tumor cell populations, angiogenic and non-angiogenic, exist side by side in a growing angiogenic tumor, and, further, that only a minority of angiogenic tumors are needed to generate a large angiogenic tumor.
You may have noticed in Eike Achilles' experiment, that a few tumors, which had lain dormant and non-angiogenic for 50-60 days, suddenly switched to the angiogenic phenotype and started growing. Why the sudden switch after a long period of dormancy?
Genetic Regulation of Cancer:
|
39. Horizontal transfer of oncogenes by uptake of apoptotic bodies |
Let me conclude this part of the talk about a new level of genetic regulation of cancer that may provide some insight into the angiogenic switch problem, especially after a long period of tumor dormancy. It is a recently reported experiment in the May 22, 2001 issue of the Proceedings of the National Academy of Sciences.
The lead-in to these experiments began in my laboratory approximately 6 years ago. At a small symposium on the subject of apoptosis convened by George Klein in Israel (at Ein Gedi), data was shown that tumor cells can eat or phagocytize apoptotic or dead cells that are nearby. The freshly dead tumor cell consists of disintegrating DNA wrapped in lipids from the original cell membrane. I asked, "What is the difference between phagocytosis of a lipid coated apoptotic body versus transfection by lipofectin, a technique by which DNA can be inserted into cells for the passage of genes?" The consensus of the group was that DNA from phagocytosis of apoptotic bodies was destroyed, but that DNA from lipofectin transfection went to the nucleus and was incorporated into the genome.
When I returned to my lab from the meeting, I discussed this problem with Lars Holmgren, then a post-doctoral fellow. He immediately saw a clean way to test the conventional wisdom that DNA from an apoptotic body was destroyed in the cell that ate it. He transfected cells with neomycin, which served as a genetic marker. He made them die and form apoptotic bodies by incubating them in nutrient-free saline overnight. He then layered these apoptotic bodies over naïve cells that did not express neomycin. Within a few days, 80% of the naïve cells were expressing neomycin. The neomycin gene had been eaten and had become transfected. Then we repeated the experiment with hygramycin and then with combinations of neomycin and hygramycin, always obtaining the same positive results.
After Holmgren completed his fellowship in my lab, he returned to the Karolinska Institute as a junior faculty member and continued these experiments. He published several papers on the horizontal transmission of genetic information by phagocytosis of apoptotic bodies, including a mechanism of Epstein Barr virus (EBV) transfection into endothelial cells. Holmgren's most recent paper in this series is shown in this slide (Proceedings of the National Academy of Sciences, May 22, 2001). Oncogenes can be horizontally transferred by the uptake of apoptotic bodies. In this experiment, the donor cells are rat embryonic fibroblasts and they are transfected with the oncogene H-ras or with c-myc. The cells become apoptotic after they are incubated in the absence of nutrients for 24 hours or if they are irradiated briefly (150 Gray). The apoptotic bodies are then co-cultured with mouse embryonic fibroblasts that are made p53 deficient. Both p53 positive and p53 negative mouse fibroblasts phagocytize the apoptotic bodies, and in both cases there is horizontal transfer of the rat oncogenes into the mouse genome. However, the p53 positive recipient cells died when they begin to proliferate, most likely due to the foreign DNA detected by the p53. In contrast, the p53 negative mouse fibroblasts become transformed and acquired the characteristics of cancer cells, including aneuploidy (excess of chromosomal material). This novel mechanism of transfer of genetic information between cells within a tumor may provide a mechanism by which genetic instability and genetic diversity are generated within a tumor. It is also possible that drug resistance can be acquired rapidly by many cells in a tumor by this mechanism.
One can speculate that if angiogenesis can be turned on by transfection of ras or VEGF (as Taturo Udagawa showed), then it may be possible that a microscopic, dormant, non-angiogenic tumor may become transfected with an oncogene that confers angiogenic activity, by the process of horizontal transfer of DNA.
Now, how can angiogenesis be inhibited? This slide shows some of the angiogenesis inhibitors developed in our laboratory. The first one was interferon-alpha. Bruce Zetter, a post-doctoral fellow in my lab, demonstrated in 1980 that this normal protein, whose known function was to interfere with viral infection, could also inhibit migration of endothelial cells in vitro. Subsequently, Robert Auerbach's lab in Wisconsin and Harold Dvorak's lab at the Beth Israel Hospital in Boston showed that interferon alpha could also inhibit angiogenesis in vivo. When compared to today's angiogenesis inhibitors, however, it is relatively weak.
By the 1990s there was steadily accumulating experimental evidence in animals that angiogenesis inhibitors were emerging as a new class of drugs. However, while basic scientists were becoming convinced that angiogenesis inhibitors should begin to be translated to clinical application, there was considerable resistance by clinicians in the early 1990s to the concept of translating anti-angiogenic therapy to clinical trials for cancer. Pharmaceutical leaders were also underwhelmed. The CEO of a large pharmaceutical company told me in 1990, "What would we do with an angiogenesis inhibitor?"
|
< previous section | next section >
More: The discovery of angiogenesis inhibitors: A new class of drugs
|
