On the MAP

Excitement was no doubt in order. "Chris was really one of a handful of people who set the field on its way to understanding how cells signal. It's a field that will never finish, and he was right there at the start," says Alison Lloyd of the University College London (UCL), Marshall's first graduate student. His discovery of the human oncogene N-Ras in 1982 "was the catalyst," says Lloyd. "And everything he's done since then is a major breakthrough."

With an eye toward understanding how Ras proteins work—and how their misbehavior leads to cancer—Marshall and his colleagues determined how Ras proteins are linked to the plasma membrane, and elucidated the signaling pathway that allows Ras proteins to do their job. "He showed that Ras and Raf together activate ERK and that they were part of a linear pathway," says Tony Pawson of the Mt. Sinai Hospital in Toronto. "That was a seminal discovery because it really put that whole canonical signaling pathway on the map." That MAP kinase pathway is now a therapeutic target for drugs that are currently being tested for treating several types of cancers.

"He's made such huge contributions to understanding the role of Ras and Raf pathways in cell transformation, migration, and morphology—and what makes cancer cells behave the way they do," says Karen Vousden, director of the Beatson Institute for Cancer Research in Glasgow and Marshall's first postdoc. "If you do the thought experiment and ask if this person didn't exist, what would have happened, well, clearly there would have been a huge hole."

As a doctoral student at Oxford University, Marshall worked with Henry Harris, a pioneer of cell hybridization techniques, who was fusing tumor cells with nontumor cells to see what the union would beget. When these hybrid cells were injected into mice, they rarely formed tumors, suggesting that the tumor phenotype was recessive. "Those studies could give you vague ideas about recessiveness or dominance," says Marshall. "But they didn't give you the genes. So it was a little bit frustrating."

During a two-year turn as a Harvard Medical School postdoc, it was the work of Robert Weinberg, then at MIT, and Harvard's Geoffrey Cooper, that showed him the path to finding genes associated with cancer. The researchers had independently demonstrated that they could transform normal cells into tumor cells by adding viral- or cancer-cell DNA. "The penny dropped and I realized that this was the way to go," says Marshall, who had just launched his own lab at ICR. "It was quite a volte-face for me, because everything I'd done before suggested that the genes involved in cancer were recessive. And those would never show up in this sort of assay," he says, because only dominant genes would be able to overpower their wild-type counterparts to transform a cell. "But you can only go after the things for which you've got an assay. As Medawar says, science is the art of the soluble."

Marshall was hoping to fish additional cancer-causing genes directly from human tumor cell lines. Working with Alan Hall—who had arrived at ICR a few months after him—Marshall purified genomic DNA, collected it in a calcium-phosphate coprecipitate, and slathered it over NIH 3T3 cells which, he says, "by some miraculous means are able to take up genomic DNA and express genes." They then looked for cells with a "cancer phenotype," for example, those that proliferated into small piles on a culture plate.

Their success was by no means instant. "There's a fine art to making the coprecipitate," says Marshall. "The granularity is crucial and I was endlessly checking the pH of the buffers. People in lab thought I had OCD."

"Transfecting NIH 3T3 cells required a little bit of magic," recalls Vousden. "And Chris was always very, very strict in terms of the science. Everything had to be just so. But he led by example and didn't expect you to do anything he didn't do."

"We screened a lot of DNAs and hadn't gotten anything novel," says Marshall. "I remember one weekend, talking with Alan and saying, 'Well, we'll try 20-odd more DNAs and see what happens. If we don't get anything, we'll have to completely rethink what we'll do with our lives."

That's when they hit upon N-Ras, the third member of the family of genes involved in 10 to 15 percent of all human cancers. Cells that had taken up N-Ras showed the uncontrolled proliferation characteristic of cancer. That discovery, published in Nature, drew Marshall into the wonderful world of cell signaling. "Chris stuck with Ras and went on to study how it works in the cell," says UCL's Robin Weiss, who recruited Marshall to ICR in 1980. "I might have continued trying to crank out more oncogenes, cataloguing them. But Chris was more thoughtful, wanting to get at the mechanisms."

"It was a very exciting time because oncogenes were just being discovered," says Lloyd. "How the outside of the cell speaks to the inside of the cell was a complete black box. If you look at signaling pathways now, they're networks. But back then they were like two proteins with a question mark in between."

Some of the most fruitful experiments were essentially fishing expeditions. "My PhD was very open-ended," says Sally Leevers of Cancer Research UK at Lincoln's Inn Fields: "Ras must activate other protein kinases. So let's find some." Again, it was easier said than done. "The first year was tough. Basically nothing worked," she says.

"I had this naïve idea that Sally could take a scrape-loading technique we'd been using for a few years to get activated Ras into cells, and combine it with a renaturation assay to see what kinases get activated," says Marshall. Cells filled with active Ras were broken open and their proteins were denatured, run out on a gel, and transferred to a membrane. "Then, instead of developing the membrane with an antibody, as you would for a Western blot, you renature the protein and incubate with hot ATP to allow any active kinases to phosphorylate themselves," he says. The treatment should uncover any kinases that had been activated by Ras—which is what Leevers initially found. But then she couldn't reproduce the results. "It was just horrible," says Marshall. "For six or seven months, Sally could not get it to work again."

"It was a long and torturous troubleshooting process," says Leevers. "But Chris was really supportive and never made me feel like it was my fault." He even helped out with some of the experiments, prepping the Ras protein and handling the radiolabeled phosphate. "Sometimes when he was in the lab it was a little chaotic," laughs Leevers. "He'd steal your reagents and you'd have to tidy up the day after. But he is quite obsessive about the science." ("I don't have OCD for tidiness," Marshall admits.)

Eventually Leevers got the assay to work again by renaturing the proteins in the gel, rather than on the membrane—a kinder, gentler, and less tricky treatment. "And she showed that when you put activated Ras into cells, MAP kinase becomes activated," says Marshall. "That was the first link between Ras and MAP kinase." They then went on to show that cells transformed by an oncogenic form of Raf also showed elevated MAP kinase activity. "That really suggested that there might be a Ras-Raf-MAP kinase cascade." And working with Philip Cohen and his colleagues at the University of Dundee, they reconstituted the cascade in vitro. "That was the start of the whole Ras-MAP kinase-ology," says Marshall, "which is still the most clearly validated pathway downstream of Ras."

In addition to running his own lab, Marshall has also been a great supporter of others' efforts, including the large-scale sequencing of human cancers. "Mike Stratton and Andy Futreal came up with the idea of the cancer genome project," says Marais. "It was an audacious idea because they proposed it before the human genome had been fully sequenced. It was Chris who suggested that they start by sequencing components of MAP kinase pathway," in part because Ras was a known oncogene. "So if you don't find Ras mutations, you know the system isn't working." As luck would have it, the first mutation they came up with was in B-Raf, a protein that Marshall and Marais had been studying. What's more, they found that B-Raf is mutated in about 7 percent of human cancers, and in more than half of all melanomas. "That discovery changed everything," says Marais. "It changed the Raf field completely, because we then knew that Raf was a major human oncogene."

These days Marshall has turned his attention to cell movement, and the signaling pathways that govern how cancer cells navigate through three-dimensional space. He's found that melanoma cells can switch back and forth between two forms of movement: one in which they extend elongated protrusions, another in which they remain more rounded. And he's showed that the two forms are controlled by the small GTPases Rac and Rho, findings published last year in Cell.

"Cell movement and cell shape are taken to be such complex properties," says Pawson. "It's really quite remarkable that you can, in a biochemical sense, sort out what are the signals regulating those events. But it's especially important because metastasis is really the killer in cancer, and we actually understand very little about how cancer cells move."

So there's no resting on one's laurels in the Marshall lab. "He's still doing exciting new experiments," says Vousden. And he still has occasion to work in the lab. "I called him on his 60th birthday [in January] and he was doing a phosphate-labeling experiment," she says. "That sums Chris up. He's a huge figure, an internationally renowned scientist. And he's in the cold room on his birthday doing an experiment. How could you not love that? How could that not inspire you?"