From a quiet genome, a new cancer gene emerges

Chronic lymphocytic leukemia, the most common form of blood cancer, is a strikingly heterogeneous disease. In some cases, the disease is aggressive and fatal; in others, it causes few symptoms for years or decades. From a genome analysis perspective, CLL is also unusual: it is quiet. As the cost of...

Chronic lymphocytic leukemia, the most common form of blood cancer, is a strikingly heterogeneous disease. In some cases, the disease is aggressive and fatal; in others, it causes few symptoms for years or decades. From a genome analysis perspective, CLL is also unusual: it is quiet.

As the cost of sequencing has plummeted, research teams have begun reading the entire genomes of many cancer types – prostate cancer, melanoma, colon cancer, and now, CLL. One of the challenges in these projects is to separate out the genetic mutations responsible for a cancer’s deadly ability to grow unchecked – driver events – from those genetic changes that have little or no effect (passengers). Finding the events that drive cancer is a bit like finding needles in a haystack or singling out one person’s voice at a loud party. In cancer genomes that are riddled with aberrations, it’s difficult to hear the signal of a driving event over the din of background mutations. But in CLL, the background noise is eerily quiet.

“We’re looking at one far extreme of the spectrum of quiet to noisy tumors,” says Mike Lawrence, a computational biologist at the Broad Institute. Blood cancers in general are among the more quiet tumors in terms of the number of mutations that they have. “If we’re looking for a needle in a haystack, here the haystack is actually pretty meager.”

“This quiet quality highlights the mutations that are there and helps draw your attention to them,” says Cathy Wu, an assistant professor of medicine at Harvard Medical School and an associated researcher at the Broad. “You’re not necessarily distracted by all of the other noise.”

From this quiet landscape, a research team led by scientists from the Broad Institute and Dana-Farber Cancer Institute were able to pick out potentially important mutations in CLL, including a mutated gene not previously reported in cancer. Their results appeared in the New England Journal of Medicine in December.

Mike and other researchers in the Broad Cancer Genome Analysis Group, which is led by Gad Getz, are able to sort out the needles from the hay using a computational tool called MutSig. This tool helps researchers determine if detected mutations signify a driving event or are more likely due to chance differences between genomes from a patient’s cancer samples and normal tissue.

In order to detect key mutations in cancer genomes, researchers need access to a large number of samples. Cathy, a researcher in Dana-Farber's Cancer Vaccine Center who sees patients with CLL together with DFCI colleague Dr. Jennifer R. Brown, was able to obtain enough samples to give the project statistical power.

“It’s not enough to sequence just a couple of cancer genomes,” says Gaddy. “If you sequence only a handful of cases and then look for mutations in additional samples, that has a very severe power limitation and limits your ability to discover new cancer genes.”

By looking at genetic material from normal tissue and cancer samples from 91 CLL patients, researchers found nine commonly mutated genes. This includes a gene involved in splicing – or piecing together genetic material that cells use to create slightly different proteins from the same gene. Before this year, compromised RNA splicing had never been reported to be recurrently mutated in cancer. It’s not yet clear how these mutations lead to CLL, but mutations to this particular gene, known as SF3B1, were seen in 15% of the cancer samples.

“This is a gene and a pathway that, until recently, no one had ever thought to be mutated in cancer,” says Gad Getz. “There’s a notion that we know all of the cancer genes already or at least the pathways. But here is a new gene in a pathway that we didn’t even suspect would be deeply involved in cancer.”

Cathy adds that SF3B1 is a marker of prognosis – patients with the mutation tend to have a more aggressive form of CLL than those that do not. “This suggests that we can refine our ability to prognosticate and it does suggest that some sort of mutational analysis could be helpful in identifying those patients in whom conventional chemotherapy is just not going to cut it,” she explains.

For those patients, more aggressive treatment options such as allogeneic stem cell transplantation (in which a patient receives blood-forming stem cells from a donor) might be best.

“We’re moving in the direction of refining treatment,” Cathy says. “We’re cutting into this heterogeneity and moving away from the one-therapy-for-all approach to discern a little bit better differences between patien