FOR CANCER CELLS, IT’S A MAKE-OR-BREAK JOURNEY. To do their worst, the cells have to depart from a tumor in the breast, prostate or other site of origin and set up a colony in another part of the body. In nine out of 10 cases, it’s these metastases that kill patients. In short order, they pop up in multiple places, too many to yield to the surgeon’s scalpel and with genetic alterations that make them more resistant to chemotherapy and radiation than the original tumor ever was.

Yet very, very few cancer cells survive the trip. Just 0.01% of the millions of cells shed daily from a primary tumor will avoid being torn apart in the bloodstream, take hold in a new organ’s tissues and then recruit normal cells to aid in their lethal endeavor. “To metastasize, a tumor cell has to complete many steps, and if it misses even one, it’s out of the game,” says Isaiah J. Fidler, professor and chairman of cancer biology and director of the Metastasis Research Center at the University of Texas M.D. Anderson Cancer Center in Houston.

The long, uncertain nature of this process should, theoretically, provide ample opportunity for therapies that would disrupt a tumor cell, and much research has focused on comprehending exactly what occurs during each stage of the so-called metastasis-invasion cascade to pinpoint the cell’s vulnerabilities. Still, researchers are dealing with a vast ecosystem of genetically unstable cancer cells and normal cells, each with their own biological characteristics. Until recently, progress was slow. “The biology of metastatic cells has been bewildering, because there are so many genes and proteins involved,” says Robert Weinberg, director of the Massachusetts Institute of Technology Ludwig Center for Cancer Research.

During the past few years, however, work on metastasis has advanced quickly, with scientists first identifying, then deciphering many of the genes and cellular functions important to cancer’s spread. That has led to new ideas about how and when the process might be short-circuited. One theory sees cancer as an inflammatory disease that subverts immune-?system cells, such as macrophages, to assist in its spread, so disrupting the activities of these noncancerous cells could avert metastasis. According to another theory, metastatic cells may exploit a normal embryonic process to break free from the primary tumor and migrate, acquiring the properties of indestructible stem cells along the way. New research has taken aim at these cancer stem cells, searching for a compound that would arrest their ability to metastasize. A third theory is focused on the timing of cancer’s spread.

“We used to think metastasis was triggered very late in the evolution of the primary tumor, because that’s when we could physically detect the secondary tumors at distant sites,” says Sridhar Ramaswamy, an oncologist and researcher at the Massachusetts General Hospital Cancer Center. “Now it seems more likely that rare metastatic cells leave some tumors when they are still very small and that many genes that trigger a metastasis are actually active early during the formation of the primary tumors.” The evidence for early metastasis, Ramaswamy says, is the detection of individual cancer cells in the blood of many patients with early-stage cancers, made possible by new technologies such as a silicon chip etched with thousands of tiny, antibody-coated, cancer-cell-attracting pegs.

If confirmed, this controversial theory could open the door to a remarkable idea—that cancer-screening tests could detect disseminated cells even before symptoms of the primary tumor appear. That would permit therapies targeting the migrating cells and their mutations to be delivered much earlier than happens now. This clashes with some other ideas about metastasis, and much more experimentation lies ahead. But all these emerging theories are part of a sudden, hopeful acceleration in the battle against cancer’s lethal colonization.

SOME CANCERS ARE VERY LIMITED IN WHERE they can establish a home. Prostate cancer cells can live only in bone, for example. Others, like breast cancer, are more promiscuous, spreading to bone, lung and brain, and melanoma can make a home anywhere.

How fast each spreads is also cancer-specific. Some cancers may take decades to metastasize—if they ever do—whereas others will spread in months, even when the primary tumor is caught early. “Lung cancer cells go out with all guns blazing because they’ve acquired mutations that make them immediately capable of setting up shop in different organs,” says Joan Massagué, chair of the Cancer Biology and Genetics Program at Memorial Sloan-Kettering Cancer Center in New York City. “In contrast, it takes years for some breast cancer and prostate cancer cells to evolve the special abilities they need to establish metastatic colonies.”

Such wide variations have made it difficult to find the right targets—such as certain genes or processes—for cancer therapy. “Metastatic cells look different in every organ,” says Christoph Klein, head of the division of oncogenomics at the University of Regensburg in Germany. “And because chemotoxic agents target the cancer cell’s DNA, there is always the danger that cells that aren’t killed will evolve mutations that give them greater metastatic potential.”

Yet, as understanding of metastases has increased, researchers have begun studying the primary tumor’s microenvironment—the nonmalignant cells in and around the tumor—for ways to prevent metastasis. “We now think the tumor co-opts these cells to promote its own survival, growth and invasion,” says Jeffrey W. Pollard, deputy director of the Cancer Center at the Albert Einstein College of Medicine in New York City.

Pollard is focusing on the white blood cells known as macrophages, which normally clean up debris in the wake of disease or injury and help identify invading viruses and bacteria. Macrophages should also identify cancer cells and mount an immune response against them. But cancer cells emit proteins called growth factors, which send out chemical distress signals to attract the macrophages to the tumor site. After they’ve been subverted by the cancer cells’ distress signals, which send the macrophages scurrying to the tumor as if it were a wound or an infection, the macrophages release enzymes that break down the matrix that binds the cancer cells to one another—just as they remodel the cell matrix during wound healing or inflammation. Once free of their cellular bonds, the cancer cells can disperse.

The macrophages are also tricked by the cancer cells into producing vascular endothelial growth factor (VEGF), which stimulates the production of small, abnormal blood vessels around the tumor that enable the cancer cells to enter the bloodstream through leaks in the vessel walls. And in response to the cancer cells’ production of colony-stimulating factor-1 (Csf-1), which promotes the formation of macrophage colonies and controls their function, the macrophages synthesize another growth factor—epidermal growth factor (EGF)—which attaches to a receptor on the cancer cells and lures them out of the tumor mass and toward the blood vessels. In all these ways, the cancer cells hijack the normal function of macrophages to further their own agenda.

In his experiments, Pollard used mice in which a breast-cancer-causing gene had been turned on. He then introduced a gene mutation that halts production of Csf-1. Though the mice developed tumors, the tumors were deficient in macrophages, and few of the cancers metastasized. “We found that if you stop the activity of the macrophages by inhibiting their signaling pathways, you can stop the migration of tumor cells and metastasis,” says Pollard, who thinks that drugs targeting macrophages or other noncancerous cells known to aid metastasis could slow or stop their proliferation.


ANOTHER THEORY ABOUT HOW METASTASIZING CELLS ESCAPE their bonds in the original tumor could explain many other steps of the metastatic process. In most normal, noncancerous tissue, cells are packed tightly together and can’t move. But when an embryo is developing, cells have to migrate from one location to another to take on new functions and build the many types of tissue that will form a complete organism. To aid in that process, something known as the epithelial-mesenchymal transition, or EMT, occurs. It enables epithelial cells, which will form the skin and line many body cavities, to lose their characteristically tight junctions and become mesenchymal cells that act like stem cells, traveling through the embryo and differentiating into the cells that form connective tissue, blood vessels and lymphatic tissue. In adults the EMT program is dormant in healthy cells, though it is sometimes briefly reactivated during wound healing as the body builds new tissue.

MIT’s Weinberg and other researchers now think the same process may explain how a cancer cell is able to move from the primary tumor and become a micrometastasis—the small clump of cells in a distant tissue that will eventually grow into a detectable macroscopic mass. “The EMT program, which lies latent in all cells’ genes, is exploited by the cancer cell, enabling it in one fell swoop to invade the tissue around the primary tumor, survive in the circulation, escape the bloodstream and seed a micrometastasis,” Weinberg says. “So cancer cells don’t have to cobble together all these distinct capabilities.” And if there’s just one master program orchestrating this complex series of steps, there could be a single target for disabling it.

EMT was discovered 20 years ago by scientists studying how the embryos of frogs and flies develop. It wasn’t until recently, though, that accumulating evidence began to suggest that EMT could also explain part of the metastasis process. “At first it seemed implausible that a normal embryonic process could be appropriated and exploited by cancer cells,” says Weinberg. But biologist Jean-Paul Thiery of France’s National Center for Scientific Research found that rat cancer cells in culture could transform themselves into mesenchymal cells, and in 2004, in human tumors, Weinberg found a transcription factor that’s crucial to EMT. “Now, with the benefit of that additional research, it makes perfect sense and has proved to be absolutely critical to advancing our knowledge of malignant progression,” he says.

Unfortunately, the so-called cancer stem cells that EMT produces have proved particularly difficult to study. When found in solid tumors, the cells were too few in number to analyze, and if they were propagated in the laboratory, they lost their stem-cell-like properties. Recently, however, Weinberg’s research team was able to generate large numbers using a reagent that coaxed cancer cells to undergo EMT, and a group of Harvard and MIT researchers has performed a large-scale analysis to gauge the effects of thousands of chemical compounds on these cells.

Among several compounds that showed promise in eradicating the cells, one called salinomycin had surprising potency. Salinomycin was more than 100 times as effective as Taxol (paclitaxel)—a chemotherapy drug commonly prescribed for breast cancer—in reducing the number of cancer cells, and it also diminished breast tumor growth in mice. If a drug like salinomycin proves to be as effective against cancer stem cells in humans as it is in mice, adding it to a chemotherapy regimen could diminish or destroy cancer cells’ metastatic powers.

IN ABOUT 5% OF ALL CANCER PATIENTS, metastases are discovered before there’s even evidence of a primary tumor. In about half of those cases the original cancer is eventually found, but in the others, according to Sloan-Kettering’s Massagué, the primary tumor remains too small to detect. “There are also many patients who have a very small tumor removed but then develop a metastasis years later,” he says. “This can only mean that the tumor was shedding cells very early on.” That, of course, contradicts the traditional view, which holds that cancer cells must mutate several times in the primary tumor to acquire the ability to disseminate and seed a distant organ, by which time the primary tumor has grown large.

“Some cancer cells may disseminate when the primary tumor is only 1 millimeter to 4 millimeters in size, too small to be detected,” Klein of the University of Regensburg says. Breast cancer, for example, is typically diagnosed when tumors are almost 10 times that large, in the range of 1 centimeter to 3 centimeters. And if metastasis starts very early, it’s possible that some mutations in migrating cells occur after they’ve left the main tumor, says Klein, who thinks that could have implications for how cancer is treated. “We’re finding that you can have a mutation in the metastasis that doesn’t appear in the primary tumor, and that means you can’t rely on the primary tumor to predict the best therapy,” he says. The right treatment for the tumor may be the wrong one for the metastases because different gene mutations are in play—and it’s usually the metastases that are deadly.

Klein studies esophageal cancer. He’s discovered disseminated cells that have lodged in patients’ bone marrow but haven’t yet formed a metastasis. A genetic profile of those cells revealed that some contained a mutation of the HER2 gene. That turned out to be very bad news for the patients. Everyone who had that mutation in a disseminated cell went on to develop metastases in the bone and was dead within 23 months. But merely having the HER2 gene in the primary tumor and not in a disseminated cell had no value in predicting a patient’s outcome.

“The drug Herceptin targets the HER2 gene, but it’s currently given only to women with breast cancer whose primary tumor expresses the mutation,” says Klein. He would like Roche, the maker of Herceptin, to study the drug’s effect on patients with esophageal cancer whose disseminated cancer cells also show the presence of the mutation—because the drug might prevent the metastases that invariably kill patients with HER2 mutations in those cells. “If that works, it could support the idea of basing therapy on the genetic profile of the disseminated cells rather than on that of the primary tumor.”

Klein’s hypothesis is controversial. While other scientists don’t dispute that cancer cells can be found in distant sites very early in the primary tumor’s progression, they don’t think those initial cells have the power to proliferate and do damage. “It takes many years and many cycles of cell proliferation in the primary tumor to sustain the randomly occurring mutations that evolve into malignant cells,” Weinberg says. “Cancer cells that disseminate early don’t have the chance to undergo that Darwinian evolution and acquire the series of mutations that enable them to create a macroscopic metastasis. Those early disseminated cancer cells just sit there like bumps on a log.”

Klein concedes he can’t yet prove that disseminated cells form macrometastases. Indeed, it’s a difficult hypothesis to test because those cells are in transition and their behavior in the laboratory won’t be telling. “The disseminated cells I isolate on my slide will definitely never develop into metastases,” he says. Rather, “we have to identify the changes in disseminated tumor cells that are essential for disease progression and then link those molecular profiles with the outcome of patients’ cancers. And that will take a lot of patients.”

And a lot of time. But there’s little question now that progress on metastasis is possible. “The cancer research community has been criticized for not making progress on the thing that really worries doctors and patients, which is metastasis,” says Massagué. “Well, we’re making important advances now. A corner has been turned, and while there’s still much to be done, there is a path forward for those who want to join the effort.”