Searching for less invasive screening tests for cancer,
Johns Hopkins scientists
have discovered proteins present in blood that accurately
identify colon cancer
and precancerous polyps.
Initial studies of the proteins CCSA-3 and CCSA-4
suggest they could be
used to develop a blood test to identify at-risk
individuals.
"The reality is that many people are not getting
regular screening
colonoscopies, so ideally we'd like to identify those with
some molecular
fingerprint for the disease and really need them," said
cancer researcher
Robert Getzenberg, a professor of urology and director of
research at Johns
Hopkins' Brady Urological Institute."
Current screening guidelines for healthy people call
for a baseline
colonoscopy — the insertion of a flexible
optical-scanning scope through the
rectum into the colon, preceded by colonic cleansing,
fasting and heavy
sedation — at age 50, followed by rescreening at
least every five to 10 years.
Colonoscopy is not foolproof; cancers can develop between
screenings.
First discovered by Getzenberg and colleagues at the
University of
Pittsburgh through a protein scan, the two blood-dwelling
proteins are thought
to be remnants of cellular debris cast off from dead cancer
cells. Although the
proteins' roles are not entirely clear, the Johns Hopkins
scientists say they
are part of the scaffolding that supports structures within
a cell's control
center, the nucleus.
Alteration of such nuclear scaffolding is a hallmark
of cancer cells that
is easily detectable under the microscope as a misshapen
and discolored nucleus.
That fact led Getzenberg to the notion that "there must be
something at the
molecular level that would form a molecular flag for cancer
via a blood test."
To find the flag, Getzenberg's team drew blood samples
from 107 apparently
healthy individuals the day before their scheduled
colonoscopies, and from 28
colorectal cancer patients.
Using a particular concentration of scaffold-proteins
as a marker for
disease, the Johns Hopkins team — which did not know
the colonoscopy results in
advance — was 100 percent accurate in identifying the
28 existing cancers. Using
the same protein markers, investigators also correctly
identified 51 of 53
individuals (96.2 percent) with normal colons and 14 of 18
people (77.8 percent)
with advanced precancerous polyps, which Getzenberg says
are the most important
to detect through routine screening.
When researchers combined samples, they correctly
identified 42 of 46 (91.3
percent) containing both cancers and advanced precancerous
polyps. Protein
levels were accurate in correctly assessing additional
blood samples from 125
people with benign conditions and other cancers.
"These proteins seem very good at separating normal
samples from cancerous
ones and identifying other groups with precancers at high
risk for disease as
well," Getzenberg said. Results are published in the June
15 issue of Cancer
Research.
The researchers are planning larger studies at several
hospitals over the
next several months. It may take several years to complete
the full range of
testing.
Getzenberg says that storing and processing the
samples are among the major
hurdles in biomarker development, a field that spans
ongoing research on many
cancers and various body fluids. "It is difficult to get
many facilities to
adhere to precise storage and processing conditions
important for keeping
proteins stable," he said. "Different conditions could
create incorrect
results." Researchers also differ in the type of biomarkers
they seek, with
some, like Getzenberg, looking for proteins, and others
searching for DNA
components.
Getzenberg and the University of Pittsburgh hold a
patent for the
technology described above, which is licensed to Onconome.
Funding for the study
described in the article was provided by Onconome and the
National Cancer
Institute. Under a licensing agreement between Onconome and
the University of
Pittsburgh, Getzenberg is entitled to a share of royalties
received by the
university on sales of products described in this article.
Getzenberg also is a
paid consultant to Onconome, which has a licensing
agreement with The Johns
Hopkins University covering CCSA-3 and -4 related
technologies. The terms of
this arrangement are being managed by Johns Hopkins in
accordance with its
conflict-of-interest policies.
Additional authors are Eddy S. Leman, Grant W. Cannon,
Lori J. Sokoll and
Daniel W. Chan, all of Johns Hopkins; and Robert E. Schoen
and Joel L.
Weissfeld, of the University of Pittsburgh Cancer
Institute.