Researchers Develop New Method for Spotting Critical Cancer Drivers

By Mark Schrope

Rapidly improving technologies for sequencing the human genome and the commercial availability of genetic tests for signs of predisposition to disease are just two indicators that an age of personalized medicine is rapidly approaching. In the relatively near future, doctors likely will be able to sequence targeted genes of a patient suffering from a disease and then order a highly effective drug tailored to treat the mutations that caused the person's illness.

Of course, such a future will depend on the ability to identify those specific gene mutations, known as drivers, that are responsible for a disease. To that end, Ali Torkamani, a newly arrived Scripps Research Institute scientist, along with Nicholas Schork, a professor in the Department of Molecular and Experimental Medicine at Scripps Research and the director of research for genomic medicine at Scripps Health, recently published as the cover paper in Cancer Research the details of a promising new method for identifying these driver mutations. The team's work may help move the concept of personalized medicine forward, and should also enable identification of new targets for more generalized cancer drugs.

Finding mutations associated with a disease such as a particular cancer is relatively straightforward using current technologies. The problem is that, in addition to driver mutations, cancer cells usually contain countless other mutations, known as passengers, that are not responsible for the disease.

"Only a small subset of those mutations actually cause the cancer by giving it some growth advantage," says Torkamani, who began the work while still a graduate student at the University of California, San Diego, "and there's no really good way to tell which ones they are."

Existing methods for pinpointing driver mutations, which have not proven as accurate as doctors and researchers would like, have relied on two basic strategies. The first is the sequence conservation approach. One of the ways cancer progresses is through mutations in genes that code for essential proteins, causing one or more changes in the amino acids used to construct the protein and leading to malfunctions. Sequence conservation methods involve analyzing the amino acids found in critical groups of proteins, because amino acids common to many key proteins are themselves assumed to be key to proper functioning. Mutations that lead to the substitution of one of these key amino acids are then assumed to be driver mutations, because they would likely have significant impacts.

The second approach involves statistical analyses of the actual structures of critical proteins. Here, mutations that have already been tied to disease are studied to determine the impacts they have on the proteins they code for, such as specific alterations to the way the protein folds. If troublesome mutations are found to commonly affect a particular aspect of folding, for instance, then other mutations that have similar impacts can be assumed important and likely drivers.

Narrowing the Focus

The work by Torkamani and Schork involved elements of both these strategies, but the researchers intentionally narrowed their focus to an important group of enzymes known as protein kinases, which modify other proteins and regulate many cellular functions. Mutated protein kinases are known to play important roles in causing certain cancers, and a number of successful drugs on the market for lung, breast, and other forms of cancer are tied to these kinases.

This narrowed focus allowed the team to look closely at amino acid substitutions and structural changes caused by mutations specific to the protein kinases. This is a key advantage, because in broader studies looking at a wider range of proteins, some of these changes would not be as discernible. For instance, a particular protein group can have subunits, known as functional domains, that perform specific functions. An analysis of all protein groups collectively would not be able to identify the importance of mutations in a particular functional domain unique to protein kinases.

Another benefit of focusing on the protein kinase family, given its proven importance in various cancers, is increased likelihood of identifying mutations that will be suitable targets for drug treatments—whether broad-based or personalized.

Fully measuring the success of the new method at this stage in the research is difficult, because most cancer driver mutations have not been definitively identified. If they had, the method wouldn't be needed in the first place. That means that determining if mutations the researchers have identified are in fact drivers will take further study.

Nonetheless, a number of preliminary tests are possible and have shown the method to be reliable. Researchers have been able to identify a number of driver mutations responsible for the subset of cancers that are completely hereditary, because the drivers are relatively easily identified through comparisons between those who have inherited the condition and non-carriers. Torkamani and Schork's method successfully picked out the mutations tied to these hereditary diseases almost 85 percent of the time, a more than 10 percent improvement compared to other methods available.

"If you want to take this on to personalized medicine, then it pays to be as accurate as possible when making predictions," says Torkamini, "and this is really important for identifying general drug targets as well."

Torkamani says the next step in their research will be analyzing the mutations identified as drivers to determine their specific impacts, enabling identification of how they might contribute to cancer progression.

Torkamani says that while their work could aid in the development of general cancer drugs, he says the work's payoff is likely to be greatest in personalized medicine. "I'm interested in the genetics of human disease in general," says Torkamani, "but I think the idea of finding out about someone's genetics and determining the best treatment for whatever disease they have is just really exciting."

Torkamani and Schork's paper, "Prediction of Cancer Driver Mutations in Protein Kinases," appeared in the March 15, 2008 issue of Cancer Research. See http://cancerres.aacrjournals.org/cgi/content/abstract/68/6/1675

Send comments to: mikaono[at]scripps.edu

A cover paper in Cancer Research by Ali Torkamani (right) and Nicholas Schork presents a promising new method for identifying genetic mutations that drive disease.