By Dean A. Haycock

Special to BioWorld Today

It seems as if everything about cancer cells is abnormal. Their appearance. Their uncontrolled growth. Their gene expression patterns. Their interactions with normal tissue. All set them apart. It is not surprising, therefore, to note that the blood vessels that nourish malignant tumors also are abnormal.

Researchers have been aware of tumor blood vessel abnormalities for more than half a century. They knew that the vessels feeding tumors were lined with normal endothelial cells, which line all blood vessels, and with cancer cells. This dual cellular constituency accounts for the name by which these vessels are known: "mosaic" tumor blood vessels.

Although scientists have known about the existence of mosaic vessels for decades, they didn't have a good idea of how many cancer cells were present in them. Now they do. Researchers from the Harvard Medical School and Massachusetts General Hospital in Boston and the University of California in San Francisco published a paper in the Dec. 19, 2000, issue of the Proceedings of the National Academy of Sciences that provides quantitative data describing mosaic blood vessels.

Author Lance Munn at the Harvard Medical School and Massachusetts General Hospital and his colleagues describe in their paper, "Mosaic blood vessels in tumors: Frequency of cancer cells in contact with flowing blood," how they managed to apply multiple staining procedures to the same animals to be able to count the number of cancer cells in contact with the lumen, or space, in tumor blood vessels.

Researchers Use GFP-Expressing Tumor Cells

The study was technically complex, according to Munn, because it involved developing a procedure to simultaneously visualize cancer cells, endothelial cells, blood vessel morphology and markers of blood flow through vessels.

"To my knowledge," he told BioWorld Today, "nobody has ever done a study that looked at this many different aspects [of tumor blood vessels] simultaneously. We did the best job we could to make sure we identified the cells correctly. That involved using the GFP-expressing tumor cells, which is probably the most significant advance."

This refers to a method the authors used to uniquely label cancer cells by transfecting a human colon adenocarcinoma cell line with a constitutively expressed green fluorescent protein (GFP) construct. "It allowed us unambiguously to say that a particular cell is a tumor cell right in the blood vessel wall," Munn explained.

To identify endothelial cells, the group labeled tissue sections with a mixture of antibodies directed against two endothelial cell markers, CD31 and CD105. "The labeling of the endothelium was pretty straightforward," Munn recalled, "but used a cocktail of markers to increase the sensitivity. The other advance was the lectin perfusion, which has been used in other systems to identify vasculature in perfused vessels. It was a challenge to get all the conditions optimized so all the various stains would be strong enough so we could see them all in the same tissue sample.

"The other major advance," he continued, "was the use of confocal microscopy. We could do 3-dimensional scans down through each region to verify that the mosaic regions carried through a significant depth. Once all this was set up, it was a matter of going through a thousand vessels to do the quantification, and that wasn't trivial in itself."

The reward for this effort was the finding that 15 percent of perfused vessels of tumors grafted at two different sites in mice were mosaic vessels that contained regions with no indication of CD31/CD105 immunoreactivity but where tumor cells appeared to contact the space in the vessel. These regions, the authors report, involved a quarter of the perimeter of the mosaic vessels. That accounts for 4 percent of the total vascular surface area in these colon cancer tumors.

Munn stressed, "It didn't appear as if the tumor cells were 'pretending' to be endothelial cells. It's more as though the endothelial cells are somehow damaged or defective in the tumor environment. They seem to develop breaks and may even peel off - so the tumor cells just happen to be there when the endothelial cells are lost. But this is still an open question that we are currently investigating using electron microscopy."

He added that the possibility still exists that the tumor cells are actively migrating into the vessels in some metastatic process. The team is developing in vivo assays to try to see this dynamically using intravital microscopy.

Importantly, the paper goes on to describe similar percentages for mosaic vessels in biopsy samples of human colon cancers.

Munn warned, "There are examples [of endothelial cells not producing certain markers], especially in tumors and in particular of CD31 not staining in some endothelial cells. That is why we tried to stack the odds in our favor by using a cocktail of two different endothelial cell markers. Although unlikely, it is possible that in some of these regions the endothelial cells are not expressing the markers that we were using. That is why we now are going to the EM [electron microscopic] level to see both if there is a cell there and whether or not it is expressing these markers."

Findings Suggest Targeting, Treating Strategies

The findings suggest a possible explanation for the effects of some cytotoxic agents on tumor vessels. "It is possible, Munn said, "that by treating and killing the cancer cells [in the mosaic regions] directly, you could cause a cascade that shuts down those vessels and cuts off blood flow to large regions of tumor."

The data also suggest potential new ways to target tumor vasculature. "We actually started this study to look at how these [mosaic blood vessels] might be important in drug delivery and in the permeability of the blood vessels," Munn said. "We now are injecting tracer particles to see how they extravasate [move from the bloodstream into the tissue] around these mosaic regions."

The group also is exploring new anticancer strategies suggested by the findings. "If we have tumor cells that are right in the vessel wall," Munn explained, "the obvious strategy would be to target those tumor cells and combine it with something that closes off the vessels or destroys the targeted vessels specifically. Since they are right there in the lumen, you don't have the traditional problems of transport to get to all the tumor cells, which is one of the big problems in treating a tumor."