April 12, 2007 | What is cancer? That’s a truly big question. Researchers and clinicians have had varying success — from spectacular to dismal — describing and treating cancer’s myriad forms. Vast resources have been spent trying to tease out pieces of the cancer jigsaw puzzle. Sadly, producing a coherent and comprehensive picture of cancer remains elusive or at least problematic.
Today, researchers at the three-year-old Center for Cancer Systems Biology (CCSB), part of the Dana Farber Cancer Institute (DFCI) in Boston, are working toward identifying the high level “wiring diagram” for cancer. In this instance, that means an accurate map of all cancer-associated protein-protein interactions.
It’s a daunting task. Even as CCSB succeeds in producing these cancer circuit descriptions, translating them into research programs and clinical treatment is a long-term effort requiring an army of researchers.
Yet the challenge is part of the attraction for Marc Vidal, director of CCSB. Trained first in engineering agronomy in Belgium, Vidal was swept up in the molecular biology revolution. He received his Ph.D. in 1991 from Belgium’s Gembloux University for work performed at Northwestern University. He identified a pair of global transcriptional regulators in yeast, one of which, RPD3, was later found to encode histone deacetylase. During his postdoc at Massachusetts General Hospital, he developed the reverse two-hybrid system to genetically characterize protein-protein interactions.
“In 1993, after a year and a half of my post-doc, it just hit me one morning,” says Vidal. “I remember talking to my previous advisor [Richard Gaber, Northwestern U], and saying, ‘You know, all those people who want to get the next protein-protein interaction, I want to know what’s next. With all the genomes coming up I wonder what the [entire] network of interactions is going to look like.’ He said, ‘I think you’re crazy,’ and I said, ‘Well, I think it’s feasible.’”
Turning that idea into reality has occupied much of Vidal’s time since then. He joined DFCI in 2000, focusing mostly on global and local properties of interactome networks. The flagship project of the CCSB is “to generate a high quality human cancer interactome map that together with other functional genomic and proteomic information, will serve as a backbone for the drawing of a global functional wiring diagram between both already known and yet to be discovered human cancer gene products,” according to its web site.
No Lack of Challenges
The path to such a cancer interactome leads first through an assembly of the normal human interactome. Against such a gold standard, cancer interactomes could then be compared and interpreted. Vidal called for a Human Interactome Project in a commentary in The Scientist last year, suggesting a $100 million price tag. His group published a first pass at a proteome-scale map of the human protein-protein interaction network in Nature in October 2005.
Another necessary step is compiling a library of protein-coding gene constructs, or open-reading frames — the ORFeome — which would allow protein interactions to be mapped. Vidal has worked extensively on the ORFeome for the nematode Caenorhabditis elegans. Vidal says, “Actually, despite what people said, C. elegans do get cancer. It’s sort of a simplified form of cancer. If you give them a Ras mutation, one set of cells will divide a few more times than they should.” Together with CCSB’s David Hill, Vidal has a paper coming out in Genomics on the human ORFeome, version 3.1 — a compilation of more than 10,000 open reading frames.
The choice to emphasize systems biology — even to the extent of incorporating it into the center’s name — reflects Vidal’s intention to look at cancer in a broader context, as well as his commitment to the multidisciplinary approach espoused by systems biology. For example, he has drawn heavily on Notre Dame mathematical physicist Albert-László Barabási’s research into scale-free networks and biological networks to interpret protein interaction networks.
“For me, the word ‘systems’ is really well defined,” says Vidal. “You can Google it and you might see various definitions, but there’s a trend. It’s essentially an ensemble of components that interact with each other and give rise to emerging properties, or something like that. That applies to economy, all the way to sociology, engineering, and biology. I go down inside cells where tens of thousands, if not hundreds of thousands of gene products form this complex structure we call the interactome. That’s when you do systems biology.”
Systems biology has been used to describe many approaches. Some researchers, such as Bernhard Palsson (UCSD), have tackled whole organism metabolism in bacteria and yeast and built complicated, quantitative models that permit “what if” kinds of experiments. Others have conducted extensive modeling of individual protein-ligand interactions and their associated pathways.
“My view is we all do systems biology, we just take the problem from different angles, and hopefully at some point it will be integratable,” says Vidal. “[There’s a] difference between local and global emerging properties. You can do beautiful systems biology with two proteins. Say you put them together in a circuit where you have two transcription factors, and they activate each other’s transcription, that’s a circuit. That’s a wonderful system right there with properties that can be studied forever.”
Sometimes, Vidal says, “people look at us and say, “Oh, you guys are not real systems biologists because you don’t really do mathematical modeling, you want to stay global.” Well, we want to stay global because we think that that’s also an interesting unit of study of the cell, because at the end of the day, the cell is a unit and there is a network of interacting proteins in there and the point is that they are properties that are global that radiate back to the biology.”
Vidal’s believes that some of those global properties might be perturbed in cancer cells. Suppose a cell has ten genes, and seven are active. “How they interact will be my interactome,” he says. In a cancer cell, a different mix of genes is switched on and off. “Interactions are rewired a little bit. That’s my new interactome, and would correlate with the cancer phenotype.”
One concrete example is CCSB’s work with tumorigenic viruses that focus on identifying interaction partners for viral proteins. Vidal ticks off the key questions: “Where do the proteins go? What do they contact on the human interactome? Can we see anything in the global properties of the interaction network that would relate back to what the viral proteins touch physically?”
CCSB is also working with mutated genes associated with tumorogenesis. Vidal says they are just now learning how to effectively clone, express, and characterize these genes, “in the context of the network and of all the [variants] associated with a phenotype such as tumorogenesis.”
There are lots of technical obstacles in pursing the Cancer Interactome. It will be necessary, for example, to account for important proteins resulting from alternative splicing or post-translation modification. You’ll never get them all, concedes Vidal, but he believes it’s possible to achieve high coverage and to draw important conclusions. Likewise, the yeast two-hybrid method of identifying protein-protein interactions can be mishandled and misused, producing inaccurate information. Vidal cautions, “You have to use it with the appropriate controls.”
So far, Vidal seems content with CCSB’s progress. Its publishing record to date is admirable, and funding is strong. Asked when systems biology could become relevant to drug discovery, Vidal says, “I would take it the other way around. Imagine designing drugs and therapeutic strategies in five years time without considering the system. Is that even conceivable? I mean seriously! So, if you asked, “Do you think systems biology will merge with the pharmaceutical industry?” I don’t know how it wouldn’t.”
Subscribe to Bio-IT World magazine.