When a person is sick, there is a tell-tale sign in their
blood: a different mix of the various types of immune cells called leukocytes.
A group of scientists at several institutions including Brown University
has discovered a way to determine that mix from the DNA in archival or fresh
blood samples, potentially providing a practical new technology not only for
medical research, but also for clinical diagnosis and treatment monitoring of
ailments including some cancers.
The key to the new technique, described in two recent
papers, is that scientists have identified in each kind of leukocyte a unique
chemical alteration to its DNA, called methylation. By detecting these
methylation signatures in a patient’s blood sample and applying a mathematical
analysis, the researchers are able to determine the relative levels of
different leukocytes and correlate those with specific diseases.
“You can simply look at the DNA and discern from the
methylation marks the relative abundance of different type of leukocytes,” said
Karl Kelsey, professor of pathology and laboratory medicine in the Warren
Alpert Medical School of Brown University and a senior author on both papers. “It’s a way to more easily interrogate the immune system of a lot of people.”
Other tests, using flow cytometry, can already sort through
the abundance of different leukocytes in a blood sample, but they require the
blood to be fresh and leukocyte cell membranes to be intact. Because the DNA in
a blood sample remains even after cells have died and degraded, tests based on
detecting methylation could help doctors or researchers analyze a patient’s
blood sample that has either aged or has simply not been kept fresh.
In a paper published online by Cancer Epidemiology, Biomarkers, and Prevention, the researchers
describe using their technique to distinguish accurately which blood samples
came from patients with head and neck squamous cell carcinoma, ovarian cancer,
or bladder cancer. By using methylation to determine the leukocyte populations
in each sample, they could predict that the same samples were as much as 10
times more likely to have come from a patient with ovarian cancer than a
healthy control patient, six times more likely to be from a head and neck
cancer patient than a healthy control, or twice as likely to be from a bladder
cancer patient than a control.
“Our approach represents a simple, yet powerful and
important new tool for medical research and may serve as a catalyst for future
blood-based disease diagnostics,” wrote the authors, who hail from Dartmouth, Oregon State
University, the University of Minnesota,
and the University of California–San Francisco, as well as Brown. Several
authors worked with Kelsey at Brown during the research.
They describe the technique and its analytical methods in
deep mathematical detail in another paper published in BMC Bioinformatics. They also report experiments that included
analyses of the leukocyte mix of noncancer conditions such as Down syndrome and
The paper found many examples of differences between the
immune cell mix of healthy controls and people with specific illnesses. For
example, obese African Americans had an estimated increase in granulocyte
leukocytes of about 12 percentage points. People with Down syndrome, had 4.8
percentage points fewer B cells. For head and neck cancer, they noted a 10.4
percentage point drop in CD4+ T-lymphocytes.
“Any disease that has an immune-cell mediated component to
it would have applicability,” Kelsey said.
In both papers, the authors said they expect that the
technique will be applied in clinical and research efforts.
“Our approach provides a completely novel tool for the study
of the immune profiles of diseases where only DNA can be accessed,” the authors
wrote in Cancer Epidemiology Biomarkers
and Prevention. “That is, we believe this approach has utility not only in
cancer diagnostics and risk-prediction, but can also be applied to future
research (including stored specimens) for any disease where the immune profile
holds medical information.”
Source: Brown University