A team led by researchers from Stanford University has developed a multi-omics approach for profiling individuals' immune systems.
Using a combination of proteomic and genomic techniques, the scientists identified nine markers with which they were able to predict patient response to an influenza vaccine with 84 percent accuracy.
Detailed in a paper published last week in Molecular Systems Biology, the study offers insights into mechanisms underlying immune function and could ultimately help identify metrics for quantifying immune health, David Furman, a post-doc in the lab of Stanford professor Mark Davis and first author on the paper, told ProteoMonitor.
"If you were to go to the doctor today and ask whether you have a healthy immune system, nobody would really be able to tell you," he said, noting that while doctors could provide information like blood platelet counts, such measures would provide little indication of whether a person was, for instance, likely to respond to a vaccine or develop an infectious disease.
In an effort to discover measures that could provide this information, Furman and his colleagues adopted a broad systems biology approach with the aim of identifying characteristics that would allow them to distinguish between patient populations who responded well or poorly to an influenza vaccine.
"One of the beauties of such an approach is that you can identify immunological states and identify clusters of measurements that will separate individuals into, in this case, [vaccine] responders and non-responders," he said. "So our first idea was to do this broad analysis of different levels of the immune system ... kind of a fishing expedition to see what comes out."
To that end the researchers applied a variety of omics methods including peptide arrays to identify pre-existing anti-influenza antibodies in patient serum and correlate them with vaccine response; immunoassays to measure serum levels of 50 cytokines; pFlow assays to assess phosphorylation levels of intracellular proteins in T cells, B cells, and monocytes; immune cell subset phenotyping of peripheral blood mononuclear cells; and microarray-based gene expression profiling.
From this data emerged a potential link between immune health and apoptosis, with higher levels of cytokines and genes linked to apoptosis being associated with better immune response. This finding, Furman noted, fits with previous research identifying apoptosis as critical to eliminating older immune cells in order to make way for fresh ones.
"There is previous knowledge about apoptosis being critical for clearing out the [immune cells] that are basically responding to historical viral exposure," he said. "In the absence of functional apoptosis machinery, these [old] T cells that you want to get rid of just keep accumulating and fill up the immunological space so that there's no room for new incomers."
The team also identified from the peptide array data a set of anti-influenza antibodies that when detected in patient serum appeared to indicate a patient would be a poor vaccine responder.
In the case of this finding, Furman said, the researchers hypothesized that the presence of these particular antibodies was a form of "negative feedback [telling the body] not to spend energy in producing more of this type of antibody," which could result in poor vaccine response when large numbers of these antibodies are present.
"There are a number of [antibodies] that we found ... [where] if you have them you won't be able to mount effective immune responses, in this case to vaccination," he said.
The researchers are still working to piece these two findings together, Furman said, but he suggested a model in which the apoptosis-deficient T cells identified via the gene expression and cytokine assays are the same cells responsible for promoting production of the antibodies associated with poor vaccine response.
"We're not quite sure how the [apoptosis and antibody findings] are related to each other, but I'm assuming there are a bunch of T cells that accumulate – are apoptosis deficient – and those are probably the ones that are [promoting production] of these pathogenic [antibodies]," he said. "This is a hunch we need to test now."
Furman and his colleagues collected data from 91 young and older subjects, ultimately identifying nine variables that they said allowed them to distinguish between good and poor influenza vaccine responders with 84 percent accuracy.
Furman cautioned, however, that it is too early to comment on any clinical potential the research might have. Before considering this issue, the researchers would need to collect data on larger and more diverse sets of subjects, he said, noting that the MSB paper looked only at two major age groups – individuals ranging from 20 to 30 years old and the elderly.
"We need independent data sets. We need to increase the sample size, [and] to test additional cohorts, ethnicities, ages," Furman said. He added that in additional data collected since publication of the initial study, the researchers had observed significant differences along gender lines.
"That tells you that there are other things that we have to consider," he said.
The researchers have now collected data on around 500 patients, Furman said, including roughly 80 for which they have five years of longitudinal data tracking their response to vaccination.
In addition to collecting data on more patients, Furman and his colleagues are adding new data sources. For instance, the original antibody profiling was done using conventional peptide arrays, which allowed the researchers to look at antibodies to roughly 40 percent of the influenza hemagglutinin protein.
Going forward, Furman said, they hope to expand their arrays to represent the entire length of the HA protein using technology being developed by Stanford researcher Paul Utz, a co-author on the MSB paper. Utz is currently collaborating with Intel on silicon-based peptide arrays that could prove a more reproducible, higher-throughput alternative to conventional glass slide arrays (PM 8/24/2012).
The team is also adding data from DVS Science's CyTOF mass cytometry instrument to their studies. That device, of which Stanford researcher Garry Nolan has been an early developer and proponent (PM 8/24/2012), uses antibodies linked to stable isotopes of elements, which can then be read with high resolution via time-of-flight mass spectrometry. It allows researchers to quantify as many as 100 protein biomarkers in individual cells at a rate of roughly 1,000 cells per second.
Furman estimated the researchers have CyTOF data for roughly 80 percent of the 500 patients from whom they've collected measurements.
That data, he noted, would allow them to look at associations between immune response and levels of cell subtypes "with a much finer resolution" than in the MSB study.
"It's much richer data," he said. "We can do a lot of [additional] things."