Seeing cancer with new eyes
The molecular biology revolution of the last ten years
has resulted in a dizzying stream of new discoveries about cancer. Of late,
scientists have been using a technique called transcription profiling with
DNA chips or microarrays to describe cancers in ways not possible before.
This technique measures the expressed levels of tens of thousands of different
genes produced by cancer cells. Through comparisons with other cancers and
with normal cells, scientists may then draw conclusions about what genes
may be malfunctioning, which groups of genes can predict metastasis, or
even what drugs may be most effective in treating a particular cancer.
 |
| Both Pomeroy, left, and Golub believe that the
full potential of transcription filing has yet to be tapped. (Photo
by Jeff Thiebauth, courtesy of Children’s Hospital, Boston) |
HMI World spoke with two Harvard scientists who have used
transcription profiling and asked them to give their impressions of the
effects this
technology has had on cancer research and where it will take us in
the future.
Todd Golub, associate professor of pediatrics at the Dana-Farber Cancer
Institute, has performed some of the groundbreaking research in this
field, working first with the blood cancer acute lymphocytic leukemia
before moving
on to solid tumors. He used DNA microchips to study the genetic profile
of these cancers and identify distinct subgroups within cancers that
were morphologically indistinguishable.
“Ten years ago, we couldn’t really see much about what we were doing.
It was hard to see into the molecular workings of a tumor,” Golub
said. “The reason we used to think of cancer classification systems
relating to anatomical location, was simply because that was the best
system we had. When you only have blunt instruments, you don’t
see very much. Now the tools we have are giving us a richness that is
far beyond what we had before, and from this is coming the notion that
the strategy for classifying cancer probably needs to be amended.”
Through transcription profiling, scientists have discovered that cancers
found in distant parts of the body may be more alike than two tumors
originating in the same organ, depending on the type of cancer-causing
mutations they harbor. “In fact, now we realize that a cancer in
the breast and one in the brain may be more alike than different,” Golub
said.
Scott Pomeroy, HMS associate professor of neurology, has worked on the
cancer medulloblastoma, a rare childhood brain tumor. He found that a
certain subgroup of medulloblastoma cases was characterized by expression
of a developmental gene named (by a playful developmental biologist in
the mid-90s) “sonic hedgehog” that is supposed to be shut
off after birth.
According to Pomeroy, microarrays and transcription profiling have dramatically
changed both the pace at which cancer research is done and the way hypotheses
are tested. “Transcription profiling enables you to generate a
large number of focused hypotheses from these data that you can test.
In the past you either made an educated guess or found a mutated gene.
You would identify A or B by looking for them specifically. In this technique,
A and B, as well as many other genes, can be looked at all at once. We
are able to look quickly at a large number of molecules and find the
ones directly related to cancer, so the rate of discovery is much higher
than it used to be.”
Not only is the research itself accelerated, but the information that
can be gleaned from it is increased too when scientists share their data
online. “When you put the data on the computer, other people can
do research by logging on and checking for their gene of interest in
your data,” said Pomeroy. “There have already been papers
published that reanalyzed our data.”
So when will transcription profiling be used to diagnose and plan treatment
for real cancer patients? Pomeroy is optimistic and believes that if
all goes well, it could be soon—but not right away. “I think
that if one can use medulloblastomas as a metric, it will be a few years
because the findings that have come out so far have been kind of backwards-looking
findings. As the treatments have evolved over the years, the patients
haven’t all been treated uniformly. So the findings have to be
validated in a group of patients treated under modern conditions. The
basic concept—using multiple expressions of genes to measure a
clinical outcome—will probably happen in the next five years,” he
said.
The advent of Gleevec for chronic myelogenous leukemia has also shown
that drugs targeted to the specific molecular changes of cancer cells
can produce dramatic results—but with much less toxicity to the
patient. “We’re reaching the end of what we can do with drugs
that work in a nonspecific fashion to kill tumors,” said Golub. “What
needs to happen now is to develop more specifically targeted treatments
in order to maximize cancer cell killing but at the same time minimize
toxicity.”
This realization has also begun to shed light on a long-time cancer treatment
mystery: why patients with what are thought to be the same cancers respond
to treatment in radically different ways. “It’s now been
realized there’s a great diversity in the way patients respond
to treatments, and we’re just now starting to find with genomic
tools the molecular reasons for why patient responses can be so different.
We’re beginning to see how it might be feasible in the future to
translate that into the selection of appropriate treatment regimens for
patients,” Golub added.
Clinical testing is currently underway to make sonic hedgehog blockers
for medulloblastoma, and Pomeroy, like Golub, sees this as just the beginning. “I
think the trend that really sticks out and has been a long time coming
is that you understand the basic mechanism of a given cancer—what’s
causing it to grow—and specifically target that process,” Pomeroy
said. “Angiogenesis inhibitors and Gleevec, for example, target
the specific biology of the tumor and shut it down by specifically targeting
the thing that allows the tumor to grow. I think that’s definitely
the way we’re going.”
|