Science Editor

Cancer cells produce specific antigens, and killer T cells can and do recognize those antigens, making T cells a much-studied weapon in the fight against cancer.

But cancer cells "are often genetically unstable," Bin Zhang told BioWorld Today. As a result, they can become invisible to killer T cells by several mechanisms.

The antigens themselves can disappear; but tumor cells also can lose antigen-presenting MHC molecules, or antigen processing can go awry. The net result is that for all their promise, T cells have not, to date, delivered in the clinic.

However, tumor cells do not operate in isolation. They are embedded in the so-called stroma, a mix of extracellular matrix, endothelial cells and microvasculature. Indeed, stromal cells can make up 90 percent of a tumor in certain types of cancers, and blockbuster antiangiogenesis drugs such as Avastin (bevacizumab), from Genentech Inc. and ImClone Systems Inc.'s Erbitux (cetuximab,) can be thought of as stroma-targeting agents.

Zhang, an assistant professor of medicine at the University of Texas Health Sciences Center in San Antonio, has focused his own attention on whether T-cell therapy could be used to target them.

From the genetic standpoint, stroma cells are "very, very stable," Zhang explained - much more so than the tumor cells. But they also express tumor-specific antigens on their surface, because such antigens are released into the tumor microenvironment by the tumor cells themselves and taken up by stromal cells.

In two recent papers, Zhang and his colleagues from the University of Chicago, where he was a postdoctoral fellow when the research was conducted, reported that T-cell therapy targeting the stroma can stop tumor growth under some circumstances, and even eliminate established tumors when the conditions are right.

In the first of his recent papers, published in the March 1, 2008, issue of Cancer Research, Zhang and his colleagues - including senior author Hans Schreiber, in whose lab Zhang worked at the time of the study - reported on targeting stroma cells by injecting mice with killer T cells that recognized tumor antigens being presented by so-called myeloid cells, a type of cell that normally is found in the bone marrow, but makes up part of the stroma.

Even though the tumor cells themselves were not recognized by the T cells, the treatment stopped tumor growth for nearly three months, which, as the authors noted, "would be an admirable achievement for many cancers." Through their effects on the stroma, the T-cell injection also reduced tumor size.

In their second paper, published online March 3, 2008, in the Journal of Clinical Investigation, Zhang and his team reported on the molecular mechanisms that enable an even greater success: the eradication of established tumors by killer T cells that target the stroma.

Zhang and his colleagues had shown that such tumor eradication is possible in a paper published last year in the Journal of Experimental Medicine; in their current JCI paper, they investigate its molecular mechanism.

They found that for killer T cells to go after tumor cells that no longer present the antigens - a process known as "bystander elimination," though the bystanders in question are anything but innocent - several conditions must be met. First, the majority of tumor cells must express the antigen that the T cells recognize, and tumor cells must produce sufficiently high levels of antigen for stromal cells to take it up and present it to the killer T cells. Second, Zhang and his colleagues found that the secretion of cytokines - interferon-gamma and tumor necrosis factor - is critical for T cell's ability to eradicate tumors. T cells from mice lacking either of the cytokines stopped tumor growth at first, but the animals quickly relapsed. Lastly, if stromal cells did not express receptors for the cytokines, T-cell therapy was also unsuccessful.

Just how the tumor cells are killed when those three conditions come together is not yet clear, but one possibility is that vessel damage and relevant stromal destruction could kill cancer cells, including those that don't have the antigen, by anoxia.

Alternatively, Zhang said, stromal destruction "may also be associated with the local induction of various types of nonantigen-specific tumoricidal effector cells" such as macrophages that kill the bystander cells in an antigen-independent manner.

Whatever the ultimate mechanism, Zhang and his colleagues concluded that "cancer cells and stroma are both essential, nonredundant targets for eradicating cancer."