By Ingfei Chen
Cancer's Complexities in the Brain and Pancreas
Genomic analyses uncover dozens of mutations in each patient’s tumor
By Ingfei Chen
A trio of studies last fall has produced the first drafts of maps that comprehensively detail what goes genetically wrong in two of the deadliest cancers: brain and pancreatic tumors.
In addition to cataloging previously known major DNA glitches, the genome analyses turned up hundreds of other rare mutations. On average, a tumor carried just dozens of genetic abnormalities, but the pattern of defects was unique from one patient to another.
The data will help direct drug developers to the best targets for attacking tumors, the studies’ authors say. However, some of the scientists point out that the findings also highlight the challenges of creating personalized therapeutic cocktails if every person’s cancer is distinctive.
While the new information will not yield any immediate changes in treatment, “we’re finally getting the tools in hand to deal with the complexity of what is really happening in the cancer genome,” says Stanley F. Nelson, a cancer genetics investigator at the University of California, Los Angeles, who was not involved in the three studies.
One report, published in Nature on Oct. 23, 2008, came from a $100 million pilot project of the Cancer Genome Atlas (TCGA), a federally funded, multicenter collaboration. A team led by cancer geneticist Lynda Chin and oncologist Matthew Meyerson of the Dana-Farber Cancer Institute and of Harvard Medical School in Boston checked for genetic alterations in tumor samples from 206 patients with glioblastoma multiforme, the most common brain cancer.
The researchers detected 453 mutations in 601 selected genes. Although most of those genes were known or suspected cancer culprits, looking broadly at cancer’s genomic makeup “allowed us to find things that we didn’t expect,” Chin says. For instance, three of the eight most significant mutations had not been considered consequential in the illness. The data also hinted that some tumors treated with the chemotherapy drug temozolomide (Temodar) outsmart it by acquiring mutations in particular DNA damage-repair genes. That finding, if confirmed, could inspire new ways to counter temozolomide resistance, says Chin.
The other two genomics studies appeared in Science on Sept. 26, 2008. Johns Hopkins University researchers and their colleagues scrutinized 20,661 genes in about two dozen samples each of glioblastoma and pancreatic cancers. They identified 748 mutations among the brain cancers and 1,163 mutations among the pancreatic cancers; each individual tumor averaged about 60 genetic alterations.
Many of the problem genes were not appreciated as cancer contributors before, says oncology researcher Kenneth Kinzler of the Johns Hopkins Kimmel Cancer Center in Baltimore. One gene, called IDH1, was defective in 12 percent of glioblastoma patients—suggesting it is a promising drug target or a marker for early detection.
Cancer appears to arise from a few major genetic players that are mutated in most tumors, plus many minor players that are less commonly mutated, Kinzler says. Although that complex picture seems daunting, it is still preliminary. The genomics studies do not tell if all the observed genetic alterations actually foster cancer.
As scientists analyze many more tumor samples, the picture becomes easier to understand, notes Meyerson, because “you see a constant pattern of recurrence of the same [genetic] events. That makes it very clear that you’re not going to need a different approach to treating every cancer.”
What’s more, mutations in most pancreatic tumors potentially disrupt genes within the same 12 core biochemical pathways in the cell, the Hopkins team found. And many TCGA brain cancers’ mutations clustered in three well-studied pathways. In such cases, rather than developing a medicine against each single culprit gene, drug designers should target these biochemical pathways, the Hopkins scientists say.
Yet Chin notes that core cancer pathways are interconnected; tweaking one process could trigger responses in the others. Thus it is crucial to first understand how they all interact, she says.
TCGA and the Hopkins efforts are producing valuable data, according to Nelson. But they only examine genes that code for proteins—less than 2 percent of the human genome. Scientists should also investigate so-called junk DNA, he says, because some of it might regulate cancer genes. “Without looking, we won’t know,” says Nelson, whose lab has begun such studies.