Eight years ago, Adrienne Boire had a conversation with a patient that would set the course of her research. Back then, she was dividing her time between her postdoctoral research in a lab studying metastasis at Memorial Sloan Kettering Cancer Center and treating patients there as part of a clinical fellowship. The patient, like Boire, was in her 30s, and she had recently been diagnosed with a condition called leptomeningeal metastasis, in which cancer cells invade the spinal fluid, causing death within a few months. “She asked me deceptively simple questions,” Boire remembers, “like, ‘Why did this happen to me? Why did my cancer go into the spinal fluid? How did it get there?’ . . . And, obviously, ‘What can we do to stop this?’” Boire didn’t have good answers. “And then she said, ‘I really hope somebody will study this someday.’”

Boire agreed...

Many unanswered questions remain, not just about leptomeningeal metastasis, but about metastasis in general—that is, the process by which cancer cells move out from the site where they initially arose and colonize new tissues. Once seen as a late stage of cancer, metastasis is now recognized as a complex process that can involve very early dissemination of cancer cells from primary tumors and is therefore unlikely to be averted simply by early screening and treatment. But as researchers work to understand cancer’s spread, they are uncovering clues about the factors that enable or block it and are working to develop treatments that specifically target disseminated cancer cells or the healthy tissues where they make their homes.

Their quest is an urgent one: there are few treatment options for metastatic disease, which is responsible for the vast majority of the nearly 10 million cancer deaths globally each year. Boire says she’s hopeful that researchers will ultimately find a way to head off metastasis completely and “put me and my clinic out of business. . . . It would be really lovely to not have any more patients to treat.”   

Sleeper cells

The scientific conceptualization of metastasis was long dominated by a model posited in a 1990 paper by Johns Hopkins Kimmel Cancer Center’s Bert Vogelstein and his colleague Eric Fearon. In that model, an accumulation of genetic mutations that activate oncogenes and tamp down tumor suppressor genes first makes normal cells form benign tumors, then turns benign tumors into malignancies, and finally enables metastatic cells to leave the primary tumor and establish themselves elsewhere in the body.

Yet hints emerged early on that there was more to the story, at least in some patients and cancer types. In the early 2000s, for example, Christoph Klein, a metastasis researcher now at University of Regensburg in Germany, his former thesis advisor Gert Riethmüller of the Ludwig Maximilians University in Munich, and their colleagues analyzed the genomes of individual cancer cells taken from the bone marrow of breast cancer patients who did not have metastases that could be detected via imaging. Disseminated cancer cells could be distinguished from normal cells in the bone marrow because they bore hallmarks of epithelial cells, breast cancer’s tissue type of origin. These markers pointed to the fact that the cells didn’t belong in the marrow and must have migrated there from elsewhere in the body. But rather than carrying more cancer-associated mutations than the primary tumors, the researchers found that these disseminated cells actually had fewer, says Klein. “We were quite frustrated to see that many of the expectations that we thought we would find did not fulfill.” 

Once seen as a late stage of cancer, metastasis is now recognized as a complex process that can involve very early dissemination of cancer cells from primary tumors.

Once a patient had detectable metastases somewhere in the body, however, the picture changed, with individual disseminated cells in the bone marrow harboring multiple genetic changes typical of primary and metastatic tumors. “That was very surprising, [that] all the information [about whether metastasis has occurred] is already in a single cell genome,” Klein says, whether or not those cells are near at the actual site of metastasis.

Klein and Riethmüller suspected that cancerous, latent cells can lodge not only in the bone marrow, but in organs distant from the primary tumor, where they later seed metastatic growth. The idea aligned with circumstantial evidence they’d uncovered earlier, including documented cases in which patients have metastatic disease with no detectable primary tumor, as well as cases in which metastatic cancer was inadvertently transferred from donors to transplant recipients via apparently healthy organs. This suggested, Klein and Riethmüller argued in a 2001 review article, that metastatic cells were able to disseminate to distant sites very early in a primary tumor’s formation, before the patient became symptomatic. But the conditions cited in that review are fairly rare, so it was a surprise when the 2003 study found early dissemination in most of the patients studied, Klein says. A few years later, he, Riethmüller, and their colleagues found further support for this idea when they detected early dissemination of metastatic cells in breast cancer–prone transgenic mice, and determined that in breast cancer patients, the size of the primary tumor wasn’t correlated with the number of disseminated tumor cells found in the bone marrow.

Other labs have also found evidence for the phenomenon of early dissemination followed by dormancy. Icahn School of Medicine at Mount Sinai metastasis researcher Julio Aguirre-Ghiso says that a series of mouse studies in his and other labs suggests “that in breast cancer patients and in melanoma, cells can leave very, very early before cancers acquire a certain size or are detectable. And the cells can colonize organs and eventually give origin to metastasis.”

This idea is now well-accepted among metastasis researchers, and it comes with a sobering implication: namely, that catching and treating a primary tumor early in its growth won’t necessarily head off metastasis. “I don’t think we can prevent [early] dissemination,” says Maria Soledad Sosa, a cancer researcher at Icahn School of Medicine. “That won’t be a viable therapy.” 

Indeed, the fact that today’s most effective cancer therapies, such as surgical removal and radiation, are largely aimed at localized, primary tumors rather than metastases may explain why most cancer fatalities are ascribed to metastatic disease, says Matt Vander Heiden, a cancer researcher at MIT’s Koch Institute for Integrative Cancer Research and a medical oncologist at the Dana-Farber Cancer Institute. “We’re not as good at treating metastatic cancer.”

Many researchers have now turned their attention to dormant metastatic cells in hopes of better understanding these cells’ characteristics and what factors determine whether they remain quiescent or wake up and begin dividing. Ideally, this will lay the groundwork for new therapies that could kill the cells before they begin to proliferate or ensure they stay in a dormant state. 

The Road to Metastasis

In some cancers, such as breast cancer and melanoma, tumor cells can leave the primary tumor site early in the tumor’s formation and colonize new tissues, where they may receive molecular signals from surrounding cells, known as the niche, that keep them dormant for long periods. Mutations in the cancer cells themselves or changes to the niche may later awake these dormant cells, enabling them to proliferate and form metastatic tumors.


Dissemination: In some cancers, including breast cancer, cancer cells can move away from the site of the primary tumor very early in the progression of the disease, before doctors can even detect a primary tumor.

Dormancy: It’s thought that most of these cells die, but a few disseminated cancer cells survive the bloodstream. These cells may already have mutations needed to colonize a new niche, such as the lungs, or they may adapt once they arrive. The cells tend to stay close to blood vessels, where they receive signals from epithelial cells directing them to stay dormant.

Proliferation: If something changes—either in the surrounding healthy tissue, where stress or other factors can alter the dormancy signals that cancer cells receive, or in the cells themselves, which sometimes stop responding to the signals, or both—the cancer cells can begin to proliferate, forming metastatic tumors.

See full infographic: WEB | PDF

Bone marrow: A sentinel 

The presence of disseminated tumor cells in the bone marrow—which can be sampled from patients relatively easily—can indicate that such cells are present elsewhere in the body as well, predicting future metastasis. The bone marrow can also play a direct role in metastases at other sites by producing dormancy cues, or, conversely, by awakening resident, dormant cancer cells, which then enter the bloodstream and travel to other tissues, where they proliferate.

Preventing metastasis

The key to preventing dormant, disseminated cancer cells from growing into metastatic tumors, experts say, lies in signals the cells receive from their environments. Although a comprehensive picture of these factors has yet to emerge, and it’s likely that not all metastatic cancers involve a dormant phase, researchers have identified numerous signals involved in inducing dormancy or waking cells up. 

For example, while doing a postdoc in Aguirre-Ghiso’s lab, Sosa and her colleagues found that a transcription factor called NR2F1 helped induce dormancy in mouse models of head and neck, breast, and prostate cancers by, among other things, bringing about cell cycle arrest, causing epigenetic changes, and activating the retinoic acid pathway, which is involved in development. Further work showed that a combination of azacitidine and retinoic acid, two cancer drugs approved by the US Food and Drug Administration (FDA), induced dormancy in cultured cancer cells. The combo is now being tested in a clinical trial among patients with prostate cancer at Mount Sinai to see if it can delay or prevent metastasis, Sosa says.

Another tack for treating metastasis is to target not the disseminated cancer cells themselves, but the niche where they reside. “There’s a lot of thought and effort in the field into how to basically get the microenvironment to fight against the tumor,” says Ekrem Emrah Er, who studies metastasis at the University of Illinois at Chicago. For instance, healthy endothelium gives off cues that can drive disseminated breast cancer cells into quiescence and keep them there, explains Cyrus Ghajar, a metastasis researcher at the Fred Hutchinson Cancer Research Center in Seattle. “And when that homeostasis is lost,” through wounding, local inflammation, or other insult, Ghajar says, “that’s one of the triggers that can now catalyze the outgrowth of tumor cells in these distant organs.” 

In a 2019 study, Ghajar and colleagues found that by disrupting two molecular signals that enable communication between disseminated cancer cells and surrounding endothelial cells alongside blood vessels in mice, “you sensitize [cancer] cells to chemotherapy, and importantly, you do so without waking them up,” he explains. “By disrupting these interactions, you can deplete the reservoir of disseminated disease, and . . . doing so prevents metastasis down the road.”

Much of the research on this phenomenon has taken place in animals, and Lewis Chodosh, a researcher and physician at the University of Pennsylvania who studies cancer dormancy, cautions that it hasn’t been conclusively demonstrated that dormant disseminated cancer cells exist and cause recurrent metastatic cancers in humans. “It’s very rare that you ever have the opportunity to look in a patient and find dormant cells,” he notes. “It’s very, very hard in a clinical situation to know to what extent does dormancy really exist.”

Moving on out

In addition to dormancy, another property of metastatic cells that has captured researchers’ attention is their ability to adapt to and thrive in new environments within the body. This represents a shift in emphasis over the past 5 to 10 years from a narrower focus on cancer cells’ ability to escape the primary tumor site, says Purdue University’s Mike Wendt. 

“You can imagine if you’re a tumor cell, and you’re growing in a primary tumor—primary tumors are very hard and collagen-rich and acidic and hypoxic—and then you’re thrust into the bloodstream or the lymphatics, then you’re deposited shortly thereafter into a normal tissue. . . . These are very oxygen-rich . . . very soft tissue,” he says. It’s not so much dissemination, then, but the abilities of tumor cells to survive and later initiate a new tumor that “really seem to be the rate-limiting steps of metastasis.”

One factor that likely enables this adaptation in at least some tumor cells, Wendt says, is a mechanism called fibroblast growth factor receptor (FGFR) signaling. He notes that several organs in the body produce ligands that interact with FGFR. His group’s hypothesis is that “if a tumor cell can upregulate the receptor, that might put it in a better position to respond to some of these FGF ligands that we know are important in normal organ physiology . . . allowing those cells to adapt and eventually grow in the context of a new and different organ.” One piece of evidence for this comes from a 2016 study in which researchers led by Razelle Kurzrock of the Moores Cancer Center at UC San Diego Health found that aberrations—such as amplification, in which the copy number of the gene increases—in the gene for this protein were relatively common, occurring in 7 percent of the thousands of tumors of various types that they analyzed. FGFR amplification has also been found in the metastatic tumors of patients whose primary tumors lacked such abnormalities, and metastatic cells can ramp up FGFR production even without such genetic changes, Wendt notes. “All of these pieces of evidence point to importance of FGFR signaling as a metastatic driver.” Multiple groups have worked to identify inhibitors of FGFR or other components of its pathway, and in 2019, the FDA approved one such drug, erdafitinib, for patients with metastatic urothelial cancer with genetic alterations in FGFR.

Another property of metastatic cells that has captured researchers’ attention is their ability to adapt to and thrive in new environments within the body.

Yibin Kang, a metastasis researcher at Princeton University, says that in recent years there’s been a change in emphasis in metastasis research from mainly trying to target what are known as oncogenic driver genes—genetic elements that are needed for cells to become cancerous and maintain malignant growth—to going after cancer fitness genes that enable cells to cope with stress. Such cancer fitness genes, he says, “are critically important for highly aggressive metastatic cancer.” In addition to their need to adapt to new environments and their unchecked growth if and when they escape dormancy, metastatic cells face stressors such as chemotherapy and attack by the immune system. In studies of one such cancer fitness gene, metadherin (MTDH), which codes for a signaling protein that’s involved in multiple pathways for functions such as apoptosis, Kang’s group has linked its amplified expression to poor prognosis in several human cancers. Furthermore, the team showed that knocking out MTDH in mice inhibited metastasis without any apparent ill effects on the animals. He’s now working on identifying inhibitors of the metadherin protein that could be used therapeutically.

In addition to adapting to new niches they colonize, metastatic cells can in some cases mold the niches to suit their needs. Using a mouse model of leptomeningeal metastasis, Memorial Sloan Kettering’s Boire and her colleagues reported in 2017 that cancerous cells in spinal fluid overproduce a protein called complement component 3 (C3), that cerebrospinal fluid (CSF) from patients with the disease also has a relatively high level of the protein, and that C3 disrupts the barrier between the blood and the CSF, allowing substances that the cancer cells need to leak across. Boire explains that the oxygen- and nutrient-poor space that houses CSF is a tough place for cancer cells to make their home, and that C3 helps them change the environment to make it more hospitable. The brain is a site where metastases seem to require particularly drastic adaptations in order to thrive. (See “Metastasis in the Brain” below.)

Researchers have yet to identify a commonality among metastatic cells that can be targeted therapeutically without deal-breaking side effects on other tissues. Although no master key that stops metastases from forming has yet been found, Boire says she thinks it’s possible that one exists, and she’s hopeful that the field is making its way toward finding it. She still has the notebook where she took notes during her conversation with the leptomeningeal metastasis patient, who died a few weeks later. While she hasn’t yet arrived at any definitive answers, Boire says that now “I have much more complete ways of asking the questions.” 

Metastasis in the Brain

Tumors visible in a brain scan

While some systemic cancer treatments can prolong patient survival by delaying metastasis, this isn’t true of brain metastasis. In fact, by extending patients’ lives, such treatments may even increase the chance that cancer will have time to spread to the brain.

Prior to the advent of Herceptin and other targeted therapies, brain metastases generally only occurred in very late-stage cancer patients who were already dying of systemic disease, explains National Cancer Institute (NCI) researcher Patricia Steeg. Despite years of experience in metastasis research, she hadn’t heard of cancer spreading to the brain until about a decade ago, when she attended a conference where some colleagues mentioned seeing the phenomenon in their patients. “What started happening was this explosion of patients that had brain [metastases] who were not at that advanced stage—they had a life to live. And they were often responding from the neck down to Herceptin and the eventual other” therapies for breast cancers that overproduce the HER2 receptor. 

Steeg says that while patients were living longer with Herceptin treatment—long enough for their cancers to metastasize to the brain—she suspects that another factor in the uptick in brain metastases was that Herceptin can’t penetrate brain tumors. In addition, notes Matt Vander Heiden of the Koch Institute for Integrative Cancer Research at MIT, brain metastases have more metabolic differences from the primary tumor than do metastases elsewhere in the body, which may explain their resistance to therapies that work on the primary tumor.

See “Bisrat Debeb: Studying Cancer’s Spread

Steeg has been studying brain metastasis in animal models ever since she first learned about it and says she now sees some hope for more-effective treatment and prevention of the condition. She points, for instance, to an ongoing NCI-funded trial led by Priscilla Brastianos of Massachusetts General Cancer Center in which researchers are analyzing brain metastasis tissue and, on that basis, trying to match patients to an existing approved treatment. There’s reason to think this approach could be promising; last year, Brastianos and her colleagues reported that in more than one-third of breast cancer patients with brain metastases, the receptors displayed on the tumor cells in the breast differ from those displayed on the cells in the brain tumor, such that the cancers at the two sites would call for different targeted treatments. For example, a cancer might overexpress the estrogen receptor in the primary breast tumor but not in the brain metastasis.

The findings offer a glimmer of hope for patients suffering from brain metastases, which are exceptionally deadly and come with debilitating symptoms such as cognitive changes, seizures, and loss of limb function, from both the cancer itself and the treatment, Steeg notes. “This is not how these patients care to live their lives. This is why we have to do something.”

Interested in reading more?

Magaizne Cover

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!