Molecular fingerprinting
may lead to new breast cancer treatments
BY KRISTIN WEIDENBACH
Researchers studying the
DNA patterns obtained from different breast tumors have
found that each cancer has a unique "molecular
fingerprint." In the future, physicians may be able
to use this information to design treatment programs
specific for each kind of tumor.
Researchers at Stanford
and in Norway analyzed 65 breast tumor samples from 42
women. They surveyed 8,102 genes in each of these samples
and monitored the activity of the genes to build a
biological picture of the gene pattern of each tissue
sample. The gene expression patterns obtained from the
breast cancer samples were compared with patterns from
samples of normal breast tissue.
Using a subset of 496
genes, Chuck Perou, PhD, and his colleagues detected four
distinct gene patterns that corresponded to four classes
of samples. The researchers specifically selected this
subset to include genes that were expressed similarly
within the same tumor yet varied among different tumors.
The first pattern that
could be reliably identified belonged to samples of
normal breast tissue. Normal breast tissue contains two
kinds of epithelial cells, basal and luminal. The gene
expression pattern associated with normal breast tissue
is typified by high expression of genes characteristic of
basal epithelial cells and low expression of genes
characteristic of luminal epithelial cells.
The second pattern was
distinguished mostly by expression of a cancer gene,
Erb-B2/HER2. The remaining two groups had features
corresponding to the two kinds of epithelial cells. One
pattern identified breast tumors derived from luminal
epithelial cells and the other identified tumors derived
from basal epithelial cells.
According to Perou,
basal-cell-derived breast tumors have not traditionally
been diagnosed and treated as a distinct group. "Our
paper shows that each epithelial cell type gives rise to
breast tumors," he said. "It wasn't clinically
appreciated before that one cell type, breast basal
epithelial cells, gives rise to a significant number of
breast tumors."
Most of the study samples
of tumors derived from luminal epithelial cells expressed
the estrogen receptor gene. The estrogen receptor protein
is an important clinical marker that physicians look for
when diagnosing breast cancer. The tumors are typically
defined as estrogen receptor positive or negative, and
patients are assigned a treatment plan accordingly.
Unlike the luminal-cell-derived tumors, neither the
tumors expressing the cancer gene nor the
basal-cell-derived tumors expressed the estrogen receptor
gene.
"One of the most
practical findings of this study may be that clinically
what are described as estrogen receptor negative tumors
may be at least two distinct subtypes that are likely
derived from two different cell types and therefore
should be treated differently," said Perou.
"Now that we recognize at least two different
diseases, we need to figure out what chemotherapy works
best on these different subtypes, which is an important
focus of my future research," he added.
The results of the study
were published in the August 17 issue of Nature. The
authors note in the report of their findings that one of
the original goals of the study was to develop a system
for classifying tumors based on their gene expression
patterns. Two particular aspects of their work indicate
to them that monitoring gene expression patterns is a
robust technique for classification of breast tumors.
The team was able to study
breast tumors from 20 patients, both before and after
they underwent a 16-week course of chemotherapy. Some of
the tumors shrank to half their original size following
the treatment, but their molecular fingerprint hardly
varied. "Even though we know these tumors were
shrinking by 50 percent or greater, the patterns looked
very similar before and after therapy," said Perou.
He says that this was somewhat surprising to researchers
who expected the gene expression pattern to change as a
result of bombarding the cells with anticancer drugs.
But the stability of the
pattern is good news for the technique because it shows
that the scientists can reliably identify a tumor from
among a potentially confusing background of other cell
types and follow its progress over time without losing
its signal.
For two patients, samples
were taken from the primary breast tumor and from lymph
nodes to which the tumor had migrated. The molecular
fingerprint from each of these samples was very similar,
showing that a tumor did not seem to lose its distinctive
characteristics upon metastasis and further confirming
that each tumor could be reliably recognized. The
research team will reveal more about the way the gene
expression patterns correlate with clinical data in a
separate report to be published later this year.
Perou is a postdoctoral
fellow in the laboratory of David Botstein, PhD,
professor of genetics. He and other members of Bostein's
lab are working closely with researchers in the
laboratory of Patrick Brown, MD, PhD, a Howard Hughes
investigator and associate professor in biochemistry,
developing methods to detect the molecular fingerprints
of a variety of human cancers. In addition to breast
cancer, the scientists and their clinical collaborators
are looking at glioblastoma -- a type of brain tumor --
and cancers of the lung, liver, prostate and ovary.
Additional Stanford
co-authors on the Nature report include Matt van de Rijn,
MD, PhD, assistant professor of pathology; Stefanie
Jeffrey, MD, assistant professor of surgery; postdoctoral
fellows Christian Rees, PhD, Jonathan Pollack, MD, PhD,
and Douglas Ross, MD, PhD; and research assistants
Alexander Pergamenschikov, Cheryl Williams and Shirley
Zhu. Michael Eisen, PhD, now at the University of
California, Berkeley, contributed to the study when he
was a research fellow in Botstein's lab.
The National Cancer
Institute and the Howard Hughes Medical Institute
provided funding for the study. Perou is a SmithKline
Beecham Pharmaceuticals Fellow of the Life Sciences
Research Foundation and Ross is a Walter and Idun Berry
Fellow. SR
|
