Technology helps trace cells' genetic fingerprints
It's often said that cancer isn't one disease, but hundreds. If new research at Dana-Farber bears fruit, that may prove to be something of an underestimate.
Using equipment with the ungainly name "oligonucleotide microarrays," investigators are exploring whether cancer comes in more varieties — potentially many more — than is now recognized. They're taking the genetic "fingerprints" of tumor cells to determine whether differences in those fingerprints reflect differences in the way the cells behave and respond to treatment.
The benefit to patients could be immense. By making finer distinctions among different types and subtypes of cancers, doctors hope to be able to tailor treatments to each patient's condition.
"The more precisely we can diagnose a condition and predict its course of progression, the better we can devise therapies against it," says Todd Golub, M.D., (left) who is leading the research at Dana-Farber. "That is the longrange hope this technology represents."
Applying labels
Tumors are traditionally named for the organ in which they originate. Tumors that begin in the breast, for example, are called breast cancer even if they break free and spread to other parts of the body. While doctors have long recognized that not all breast cancers are alike — some are fast-growing, some are slow-growing; some are likely to spread, others are not — distinguishing one from another is currently a very inexact science.
In some cases, like leukemia, where doctors have identified several different types, determining the type contracted by an individual patient is the job of pathologists and other specialists who scrutinize tumor cells for certain distinctive patterns and markings. The trouble is, cancer cells that look alike under the microscope don't always act alike or respond to treatment in the same way.
For years, physicians and researchers have sought to improve their ability to diagnose different subtypes of cancer. One idea was to examine the genetic make-up of tumor cells — to look not just at the handful of genes that are defective in a cell, but to get a snapshot of the cell's entire genetic inventory, a read-out of the genes that are activated (and their level of activity) and those that are inactive.
The only thing standing between doctors and that ambitious goal was a lack of technology. In the last few years, however, that barrier has been overcome.
DNA on a chip
Golub and his colleagues at Dana-Farber and the Whitehead Institute for Biomedical Research are pioneering the use of DNA microarrays, not just as laboratory tools, but also for the clinical diagnosis of cancer in human patients.
"This opens the door to testing breast cancer, lung cancer, and a variety of other malignancies for distinctions that have not been recognized in the past."
— Todd Golub, M.D.
The technology, which is shared by Dana-Farber and Whitehead researchers, consists of a series of plastic cartridges and a computerized analyzer. The cartridges, which are the size of a small pocket calculator, are embedded with thousands of tiny compartments, each containing the DNA for a different gene. When a tumor's RNA — a chemical "mirror image" of its activated genes — is placed on the microchip, it binds to the corresponding DNA, and the reaction is scanned and recorded by the computer. The result is a molecular profile of the genes that are switched on within the cell.
But how to make sense of this windfall of data? There are roughly 100,000 genes in the human genome; the microarray technology used by Golub and his colleagues scans about 7,000 of them. Which of those genes could help researchers distinguish one type of cancer from another?
There are two ways of finding out. Researchers could make an educated guess, but, as Golub admits, they'd be likely to discover what they already know. Or they could simply record the data and see what it tells them. "We're not smart enough to know in advance which genes will prove to be important and which won't," Golub says. "We decided to collect as much data as we could and have it show us what's important."
Used with the new DNA cartridges, a computerized analyzer can produce this instantaneous recognition pattern distinguishing between different types of leukemia by identifying which genes are "switched on." Red indicates a high level of activity (the gene is producing products such as enzymes), blue a low level of activity.
Their first effort in that direction involved leukemia. Doctors normally distinguish between the two main acute types of the disease — acute myeloid leukemia (AML) and acute lymphoblastic leukemia (ALL) — by running tissue samples through a battery of tests, each performed in a specialized laboratory and requiring the services of experienced pathologists. To determine whether microarray technology could do the job just as well, Golub and his colleagues scanned bone marrow samples from 72 patients with either type of the disease. Using a mathematical model developed at the Whitehead Institute, they identified the genetic signatures of the two types. When they crosschecked diagnoses made by the chip with those made with a microscope, they found the chip could distinguish between the two varieties with nearly 100 percent accuracy.
"This opens the door to testing breast cancer, lung cancer, and a variety of other malignancies for distinctions that have not been recognized in the past, and that suggest new forms of treatment," Golub says.
"With prostate cancer, for example, we know that after surgery to remove the prostate gland, some patients remain disease-free for decades, while in others the disease roars back and is fatal," he adds. "Right now, we have no way of telling in advance which type a patient has. But by looking at the pattern of gene activation within a tumor, we may be able to."
The technology may also enable researchers to find tumor-causing genes that are good targets for new therapies. "If we really understand the genetic abnormalities that define prostate cancer, we'll be better able to target therapies at critical genes," Golub remarks.
The research even has potential to radically alter the way cancers are classified and categorized. If the technology proves as powerful as some think it is, doctors may one day label tumors not by the organ where they originated, but by their genetic profile — and treat them accordingly.
