NEW YORK (GenomeWeb News) – In a study appearing online today in BMC Medical Genomics, researchers from the University of Virginia reported that they tracked down a gene expression signature that may help predict atherosclerosis risk.
Using published microarray data, the group initially found a few hundred genes with distinct expression patterns in a subset of white blood cells from individuals with a family history of hyperlipidemia, a condition that increases atherosclerosis, compared with unaffected controls. They then trimmed this potential blood-based biomarker set down to 56 genes, concentrating on those with functions that are possibly related to atherosclerosis.
Expression patterns for the 56 genes appeared to accurately classify individuals into high- or low-atherosclerosis risk groups, researchers reported, based on their analysis of available datasets representing expression in several white blood cell types.
Researchers believe a subset of these biomarkers may eventually prove useful for clinically detecting early atherosclerotic events. The findings also provide clues about pathways that may contribute to early stages of the disease.
"This data provides some information, not only for the prediction of atherosclerosis, but also about the development of atherosclerosis," the study's first author, Feng Cheng, told GenomeWeb Daily News.
Cheng was a researcher in senior author Jae Lee's biophysics lab at the University of Virginia when the research was done. He is now a post-doctoral researcher at Yale University.
Atherosclerosis, artery-clogging plaques that can eventually lead to debilitating conditions such as heart disease or stroke, appears to arise as a consequence of multiple biological processes. But detecting early signs of impending artherosclerosis can be difficult and often requires arterial imaging, the study's authors noted.
"[U]se of these imaging methods for plaque detection in human coronary arteries is costly and associated with the risk of adverse events, so are often restricted to a small proportion of patients who already show high-risk features and clinical symptoms," they wrote. "Consequently, despite these biotechnical and molecular efforts, diagnosis of sub-clinical atherosclerosis remains difficult."
In their effort to find blood-based markers for simplifying this process, the researchers started by looking at gene expression in individuals with a condition known as familial hyperlipidemia who came from families known to be at risk of the condition owing to a genetic mutation that elevates levels of certain lipids in the blood.
"These people have a very high risk of atherosclerosis and atherosclerosis can develop at an early age," Cheng explained. By studying these individuals, he added, "we can see which genes might be important for development of the atherosclerosis."
Using Affymetrix microarray data, the team compared genome-wide gene expression patterns in a type of white blood cell known as monocytes for 10 individuals with a family history of hyperlipidemia with those in 13 unaffected control individuals.
The team's comparison uncovered 363 genes with distinct expression profiles in the familial hyperlipidemia group relative to the control group.
"A lot of those genes may or may not be relevant to the atherosclerosis or cardiovascular functions," Lee told GWDN.
To find those most likely to play a role in these processes, he explained, the researchers looked through the list of differentially expressed genes to find those with biological functions suspected of being most relevant to atherosclerosis risk, including genes involved in inflammation, lipid metabolism, and so on — an approach that helped them pare the potential biomarker set down to 56 genes.
Because the data used for their initial analysis was specifically generated from monocytes, the investigators were also interested in seeing how the potential biomarkers performed in expression datasets representing other white blood cell types.
For instance, they looked at whether the 56-gene signature could distinguish between the same 10 familial hyperlipidemia patients and 13 controls in the original analysis using gene expression data from another type of white blood cells, the T-lymphocytes.
The team did follow-up analyses using published expression data on two other cohorts as well: a group comprised of five more individuals with familial hyperlipidemia and five matched controls and a group of 15 individuals with asymptomatic atherosclerosis and 15 controls.
Individuals in the first group had had gene expression profiling done on their white blood cell samples as a whole, while gene expression data was available on a subset of monocyte cells for the latter group.
Along with statistical analyses showing that the complete 56-gene set could distinguish between high- and low-risk individuals using expression data on each of the white blood cell types, the researchers used an algorithm known as COXEN to look at which potential biomarkers shared similar expression patterns in all of the white blood cell types.
Though they did find subsets of genes that were more informative depending on the white blood cell type tested, Lee noted that the overall expression patterns were quite concordant across the white blood cell sub-types for which data was available.
"Once we combined those multi-gene signatures, we realized that that information is quite consistently preserved across the multiple [white blood] cell types," he said.
The group is now working on a prospective study of the proposed biomarkers using blood samples collected from individuals from a more general patient population for whom follow-up clinical information is available.
"Our goal is to generalize this kind of predictive signature to the general atherosclerosis population," Lee said.
If they are able to validate their current findings, the team ultimately hopes to see the atherosclerosis biomarkers used in a clinical diagnostics setting.
The type of platform used to test these markers may depend on the minimum number of informative genes that are needed to distinguish between the high- and low-risk individuals, Lee noted. "Our hope is to combine a specific number of genes into a simpler assay, such as RT-PCR," he said, "and use that as a diagnostic tool."
The group is working with the University of Virginia's patent foundation to lay the groundwork for possible commercialization of such a diagnostic test. They are also in preliminary talks with some undisclosed companies who may be interested in partnering on the development of such a test, Lee said.