Skip to Content
MIT News magazine

Understanding Metastasis

Pioneering biologist Robert Weinberg, the first to discover a gene that causes cancer, is now studying how the disease spreads.
February 19, 2008

Cancer cells evade the immune system, travel down blood-vessel thoroughfares, colonize distant regions of the body, and recruit normal cells to support their cause. For all that, they aren’t very clever, says Robert Weinberg, a professor of biology and member of the Whitehead Institute for Biomedical Research. They pull off complex biological feats using a few surprisingly simple tricks, which Weinberg’s research is helping to reveal.

Weinberg has played a leading role in defining the battlegrounds in the war on cancer. He was the first to discover a cancer-­causing gene, and the first to isolate a human tumor-suppressing gene. And when it became clear that his successes had led to a baffling and increasingly complex picture of the disease, he came up with unifying, simplifying principles that helped move research in a promising new direction. Weinberg is now studying one of the most important and least understood processes in cancer: metastasis, the spread of malignant cells from an initial site to other parts of the body, where they cause secondary tumors.

“In the mid-1970s, we knew nothing about why cancer cells misbehaved,” recalls Weinberg. But in 1982, he identified a cancer­-­causing mutation in a human gene called RAS. Before this discovery, the notion that cancer originated in mutated genes was speculative; in the years since, hundreds of mutations have been associated with cancer.

The discovery of RAS was followed by what Weinberg calls an “avalanche” of research tying cancer cells and tumor traits to a multiplicity of damaged genes. These findings have led to new therapies. But the sheer number of cancer-causing genetic mistakes that biologists have uncovered has overwhelmed them even as it has empowered them. The variety of the genes implicated raises a basic question about cancer: Is it actually a few hundred different diseases, each with its own genetic causes and disease processes? Or is it indeed one disease, with characteristics that are common to tumors in different patients and tissue types?


  • Green fluorescent protein labels are used to track cancer cells.

In 2000, Weinberg and Douglas Hanahan of the University of California, San Francisco, addressed this question in an article that has already become a classic. “The Hallmarks of Cancer,” published in the journal Cell, argued that cancers should still be grouped together as one disease. Weinberg and Hanahan professed their faith in the possibility of simplifying cancer biology even in the face of a scientific literature “complex almost beyond measure.” They predicted that the disease’s complexity would be understood in terms of a few underlying principles, and they proposed six such “hallmarks”–acquired abilities that together make a successful tumor, be it in the breast, in the lung, or elsewhere. Cancer cells evade death, produce their own growth signals, resist antigrowth signals, continually encourage the construction of blood vessels, replicate on a potentially limitless scale, and can invade surrounding tissue to create distant tumors. The researchers suggested that this group of behaviors is characteristic of cancer–no matter what combination of genetic mistakes enables them, and no matter where in the body the cancer originates.

Weinberg and Hanahan are currently updating this work and may add one or two hallmarks, but Weinberg says the basic model has held up well over the past eight years. The global view it advocates–the concept of looking at groups of genes that cause common disease processes–is not just an interesting way of thinking about cancer. The work has been cited in hundreds of other papers. It has been put into practice in the emerging field of network or systems biology by researchers who have used it to make computer models that help them understand cancer cells and even predict individual patients’ response to therapies.

As head of the Ludwig Center for Molecular Oncology, which was established at MIT in 2006 to foster metastasis research and is now part of the Koch Institute, Weinberg concentrates primarily on breast cancer. Most deaths from this disease are caused not by the original tumors in the breast but by secondary tumors that arise elsewhere in the body.

“Until recently, we had no idea what caused cancer cells to metastasize,” Weinberg says. But it’s becoming clear that a complex cascade of events has to happen for any cancer cell to leave a tumor and become the basis of another. First, cells in the initial tumor produce enzymes that break down the surrounding tissue, clearing the way for the cells to invade it. Some, but not all, of these cells are then able to enter the bloodstream. Only some of those, in turn, can emerge from the bloodstream to lodge in a new place in the body. And only a portion of those cells remain malignant enough to persist at the new site and grow from a tiny colony into a large, life-threatening tumor.

In one sense, we’re lucky metastasis is so elaborate; this means that most cancer cells never make it all the way through the process to start their own tumors. But it presents a challenge to researchers. “The biology of metastatic cells seems to be so complex as to be bewildering beyond anyone’s ability to understand it, simply because there are so many genes and proteins involved,” Weinberg says.

His recent research suggests, however, that cancer cells orchestrate this complex achievement in a simple way: through ­master-regulator proteins called transcription factors, each of which activates or deactivates a large number of genes.

Transcription factors regulate genetic activity in all cells. A handful of them are particularly important during embryonic development, when cells travel to new areas and take on new functions to help turn a single-celled zygote into a complex animal made up of many kinds of carefully organized, highly differentiated tissues. “This movement in the embryo is at least superficially similar to metastasis,” says Weinberg. The embryonic transcription factors are generally inactive in adults; as a result, most normal adult cells (blood cells are an exception) are stationary and tightly bound to their neighbors. But Sendurai Mani and Jing Yang, former postdocs in Weinberg’s lab who studied the roles of two embryonic transcription factors important in breast cancer metastasis, found that certain tumor cells somehow reactivate these proteins, thus regaining the mobility and flexibility of embryonic cells.

Until three or four years ago, Weinberg says, biologists thought cancer cells within a tumor were pretty much equivalent: any given cell was just as likely to initiate metastasis as its neighbors. His lab’s work suggests that this is not the case. Mani and Yang studied four breast cancer cell lines derived from a single mouse tumor. Despite their common ancestry, these cells have divergent properties. All four types can form tumors, but only three can make it into the circulatory system; only two can reach the lung, a common site of secondary breast cancer tumors; and only one can form successful new tumors there. This last cell line over­produces two transcription factors, which are at least partly responsible for its metastatic success.

The discovery that cancer cells rely on a few transcription factors is heartening, says Weinberg. It gives researchers a small number of very important targets to go after, instead of a large number of less important ones. The highly malignant activities of some cancer cells can be traced to a few central regulators, “each of which acts by modulating the expression of a whole cohort of responder genes,” Weinberg says. “Rather than having to cobble together complex behavioral programs, cancer cells simply resort opportunistically to resurrecting behaviors that are normally suppressed in adult tissue.”

The Complicated Role of Healthy Cells
And that’s not the only secret to metastasis. “Highly malignant cancer-cell traits [are] not dictated exclusively by the genes residing inside the cancer cells,” says Weinberg. Another key to the cells’ success is their generous neighbors.

Tumors cannot survive without enlisting the support of many normal cells. Most famously, they recruit cells to build new blood vessels that carry oxygen and nutrients in and carry carbon dioxide and other waste out. But these blood-vessel-­forming cells are only one of a multitude of cell types lured into tumors. Research by Weinberg and his postdoc Antoine Karnoub shows that normal cells called mesenchymal stem cells also play an important role in metastasis.

Mesenchymal stem cells are adult stem cells that originate in the bone marrow. Although poorly understood, they are known to travel to infected and inflamed sites in the body, where they play an important role in healing. Once in a tumor, ­Weinberg says, mesenchymal stem cells “educate” cancer cells on how to become metastatic. In mouse models, breast cancer cells that have been exposed to mesenchymal stem cells can form new tumors in the lung five to six times as efficiently as cells that have not been exposed.

Karnoub performed a series of mouse experiments with mesenchymal stem cells and breast cancer cells. First, he found that when injected into the bloodstream of mice with breast cancer, stem cells labeled with imaging dyes ended up concentrated in the tumors. This suggested that the tumor cells were “talking” to the stem cells, sending signals to draw them in. When such stem cells and tumor cells were mixed together and then implanted into the mice, the tumors were more metastatic than cancer cells implanted alone. This suggested that the stem cells were “talking back” to the tumors, sending them some kind of message that made them behave more aggressively.

Further work revealed that the cancer cells induce the stem cells to produce a signaling protein called CCL5, though the precise mechanism is uncertain. Cancer cells exposed to CCL5 move more readily from the bloodstream into the lungs, Karnoub found. When they leave the tumor, they lose contact with the stem cells; after reaching the lungs, they forget their CCL5 education and revert to a less metastatic form.

Karnoub and Weinberg’s work upends the conventional wisdom about metastasis. For 30 years, says Weinberg, the “prevailing paradigm in cancer research” has been that cancer cells’ behavior is dictated primarily by their genes. But Karnoub’s work shows that at least in some cases, highly aggressive behavior is transient, sustained not by the cancer cells themselves but by signals from normal cells nearby. “We now realize that an equally important determinant of the behavior of the cancer cell is its context, its environment within the tissue,” says Weinberg.

This revelation adds a new dimension to the emerging understanding of metastasis as a process that only some cancer cells are capable of, involving reactivation of embryonic behaviors by master genetic regulators–an understanding that has taken shape only over the past five years. It’s unclear what kind of therapies all this work might lead to as the new framework of knowledge is filled in over the next decade or so. Weinberg says that the full implications for cancer in general aren’t yet apparent but that his findings probably apply broadly to carcinomas, the class to which breast cancer belongs. Carcinomas arise in the skin and in the tissues that line organs including the lungs, liver, prostate, and colon.

“It may ultimately become a small part of the picture,” Weinberg says, “but it’s a hint of things to come.”

Associate editor Katherine Bourzac, SM ‘04, welcomes your feedback at

Keep Reading

Most Popular

This new data poisoning tool lets artists fight back against generative AI

The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models. 

The Biggest Questions: What is death?

New neuroscience is challenging our understanding of the dying process—bringing opportunities for the living.

Rogue superintelligence and merging with machines: Inside the mind of OpenAI’s chief scientist

An exclusive conversation with Ilya Sutskever on his fears for the future of AI and why they’ve made him change the focus of his life’s work.

How to fix the internet

If we want online discourse to improve, we need to move beyond the big platforms.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at with a list of newsletters you’d like to receive.