A Hard Boiled Look at Metastatic Remodeling Molecules

By Jason Socrates Bardi

The very worst kind of tumors are those drifters that pick up and move to another part of the body, an event called metastasis.

When a tumor metastasizes, it does so by expressing proteins that allow its cells to break free of the colony, enter the bloodstream, survive in the circulation, and arrest in the vessels of another organ. Then these founder cells express more proteins that keep them alive, allow them to divide and divide, bring them blood and nutrients, and help them grow into new metastatic tumor colonies.

Professor James Quigley of the Department of Vascular Biology calls the expression changes within a tumor that begins metastasis remodeling.

"If a cell is going to spread into a neighboring tissue," he says, "there is going to be a lot of molecular remodeling... We're after the early events. What are the molecules that determine why a human tumor cell migrates to and survives in a different organ—a different environment with different growth factors, adhesion molecules, hormones, and glandular products nearby?"

If one could identify these molecules, they would yield information about both the basic biology of metastatic cancer and possible therapeutic targets, and that is exactly what Quigley is trying to do.

Under Glass, In Profile, Under Reason

This is easier said then done, though, because trying to observe metastasis in the laboratory is not trivial. First, scientists must study cells in vivo, because metastasis is a complex, three-dimensional cascade event with too many processes involved to be imitated in vitro. "Metastasis cannot be mimicked at all in culture," says Quigley.

Second, some changes may not be morphological or otherwise observable via some easily detected change or signal. In order to detect the tumors with a microscope, a standard method, one would have to wait for the tumors to grow large enough to be seen through exhaustive searches of tissue sections, a time-consuming processes. Using pigmented melanomas to increase the visibility of the tumors helps, but these melanomas do not necessarily have the desired metastatic properties.

Finally, there is the problem of false positives, which arises from the fact that some of the most important molecules that contribute to the phenotype of metastasis may not be the ones that are most widely expressed. The observation that a protein is up or down-regulated in cell metastasis does not mean that protein is necessary for metastasis.

Plus, the metastatic phenotype may be brought about by extensive combinations of subtle up and down regulations of proteins that activate or suppress other proteins that are normally there, in which case the more interesting observation would be how the expression of the activator and suppressor proteins change.

"We want to identify the molecules that are not just associated or correlated with a given process, but functionally involved," Quigley says.

One way to get at these functionally involved molecules is to reason out which proteins might be necessary for metastasis, given everything we know about the basic biology, and study those. Serine and metalloproteases, for instance, are necessary to free a potentially metastatic cell from the collagens and proteoglycans that make up the stromal tissues to which they are bound. Quigley has worked on such proteases for years, and he continues to devote a significant portion of his time to them.

Quigley and his laboratory have also developed another method, which Quigley calls an unbiased approach to identifying the proteins involved in metastasis. This method has the advantage of working without any preconceived notion as to what these proteins are.

The Unbiased Way

This approach involves first generating many monoclonal antibodies raised against "crude" tumor cell antigen populations—whole cells and cell membranes—and then screening for those that block the metastatic ability of the cell. Quigley reasons that any antibody that arrests the metastasis must have as its target some antigen involved in the process.

"[We do this] without having any idea as to what the nature of the target of that antibody is," says Quigley. "Once we screen for a blocker, then we try to identify its target."

The difficulty in using monoclonal antibodies against tumor cells is that most of the antibodies raised will be against immunodominant antigens—all the high-visibility proteins that the immune system recognizes. But these antigens may not be critical for a biological process like metastasis. One would really like to make antibody against only those minor or low-abundance antigens that are critical.

To accomplish this, Quigley uses a trick called "subtractive immunization" that increases the proportion of non-immunodominant antibody raised.

The trick with subtractive immunization is to first tolerize an organism -with a non-metastatic tumor cell, using an immune suppressant chemical like cyclophosphamide to kill off all the immune cells that recognize the immunodominant antigens.

Once this tolerance is achieved, the next step is to challenge the same organism with a second tumor cell. The second cell should be similar in every way to the first except that it is aggressively metastatic. "It will still have those same common immunodominant antigens on the surface, but the tolerized immune system won't mount a defense against them," says Quigley.

Instead, the immune system will produce antibodies against those proteins that are unique to or enriched in the metastatic cell. Some of the proteins may directly contribute to metastatic spread. Then the B cells that make these antibodies can be used to make antibody-producing cell lines, called hybridomas, and the antibodies produced by these hybridomas can be screened in an assay Quigley has developed to look for the phenotype of blocking metastasis. Then the antigens of those antibodies that do block metastasis can be isolated and sequenced.

Quigley calls this isolation "a hard chore," but once done, these antigens should be cell molecules that are essential components for metastasis and, perhaps, eventual targets for therapeutics.

The Assay

Quigley uses an assay that has the advantages of being inexpensive, fast, simple, not requiring complicated surgeries, animal protocols, and large spaces.

The assay itself uses chicken eggs with their shells removed that are placed in an incubator to develop for a few days. Because the eggs are only a few days old, they are immunologically naïve and will tolerate human tumor cells.

After ten days, a tumor is placed on the soft chorioallantoic membrane on the inside of the shell and an antibody against the tumor cells is injected into the egg. Quigley uses an aggressive tumor cell that will grow and form metastases in under a week. An antibody of interest can be injected into the egg and tested for the ability to block metastasis. Another advantage of this model is that the volume is small and dilution of antibody over the course of the assay is minimal. "It stays in and does its job," says Quigley.

After a week with the tumor, metastatic cells can be detected by looking for evidence of human (tumor) DNA in a part of the egg that is distinct from where the tumor was implanted a week before. Human DNA has thousands of copies of 30 to 50 base "alu" repeats spread throughout its genome, and finding copies of these alu repeats is analogous to finding a visible tumor. The DNA can be extracted from the sample using a simple prep and amplified through the polymerase chain reaction with alu-specific primers.

Then the amount of DNA extracted from the sample can be quantified by comparing the signal to a standard curve. The number of copies of alu repeats progresses linearly with the number of tumor cells. "We can detect anywhere from 100 to 10,000 cells per sample," says Quigley. "And 100 cells is really seeing micrometastasis—cells that you might never see [looking at sections under a microscope]."

By testing different antibodies against a control, those which alter the metastatic phenotype of the tumor cells can be easily identified. Once an antibody is found that modulates metastasis, its antigen must then be identified by using the same antibody to isolate the correct protein, which then gets sequenced.

Occasionally a protein is found that, when blocked, stops metastasis. Quigley's group has found several of these thus far. Some were to be expected, such as a membrane-spanning integrin-associated protein they identified that is necessary for helping the integrins loose their grip on other cells and on the extracellular matrix, an important first step in metastasis.

Others, however, have turned out to be a surprise. One antigen that was identified was that of an novel protein whose cDNA has been sequenced in an expression database but not yet annotated.

"We have no idea what it is," Quigley says. "But we have cloned it, and we are trying to find out how it works."

Metastasis and Angiogenesis—The Complete Picture

In addition to metastasis, Quigley's laboratory studies angiogenesis, the process where blood vessels are formed and differentiated. The goal of both the metastasis and the angiogenesis work is to identify molecules that could become targets for intervention, and he employs the same basic techniques in both, using subtractive immunization and chicken egg in vivo models to study them.

In the angiogenic models, blood vessels can be easily counted and observed under a microscope, and this forms a basic assay that can then be used to screen for compounds that inhibit, stimulate, or otherwise modulate the angiogenic process.

In fact, the laboratory's first major paper in the field, recently published in Blood, was about growth factor induced angiogenesis. After implanting native collagen onto a chick embryo, they injected angiogenesis growth factors into the collagen, and then studied the effect of new vessel growth on the surrounding area to uncover the active molecules that contributed to the remodeling of the new tissue. Some of the active molecules turned out to be enzymes that break down proteins, the same proteolytic enzymes that Quigley had been studying for years as part of his research into the tumor invasion process. Now his research has turned up the fact that the same enzymes are also involved in the formation of new blood vessels, a gratifying breakthrough.

In other instances, as well, these two problems merge. For instance, tumor angiogenesis—the process whereby tumors will cause a proliferation of blood vessel growth—is an important first step in metastasis. A tumor placed on a membrane will cause new vessels to grow into a drop of collagen that is laid on top of the tumor, and this process can be studied.

"In some cases," he notes, "all the areas in the laboratory merge."

Go back to News & Views Index

James Quigley is a professor in the Department of Vascular Biology.

Basis for quantifying new blood vessel growth (angiogenesis) involves implanting a gridded nylon mesh surrounded by collagen onto the chorioallantoic membrane of a chick embryo (A and B). When the collagen contain specific growth factors, bFGF and VEGF, enhanced appearance of new blood vessels occur in the upper grid of the nylon mesh (C vs. D), which can be easily quantified.