NEW YORK (GenomeWeb) – Karolinska Institute researchers have developed a mass spec method for identifying antibody variable regions that could prove useful as disease biomarkers.
In a study published this month in Nature Scientific Reports, they used the approach to identify markers for distinguishing between patients with Alzheimer's disease and dementia with Lewy bodies (DLB), finding that the addition of these molecules to their analysis significantly improved their ability to discriminate between the two conditions.
More generally, the method offers the possibility of using mass spec to detect immune responses characteristic of particular diseases, said Roman Zubarev, professor of medical proteomics at Karolinska and senior author on the study.
Such an approach, he noted, offers a much broader space for discovery than conventional protein biomarker work.
"The whole proteome is around 20,000 proteins, but there's experimental evidence for only around 14,500 proteins so far," he said. "The number of IgG peptides is much greater. It's a potentially much bigger field for discovery."
The notion of using patient immune responses as biomarkers rather than levels of disease-associated antigens is not a new one. This, for instance, is the idea behind diagnostics firm HealthTell, which uses a random-sequence peptide microarray-based platform developed at Arizona State University's Biodesign Institute to generate patient immunosignatures — antibody expression profiles that can be correlated with various disease states.
Opko Health has explored a similar approach to profiling immune responses, using arrays of synthetic molecules to bind patient antibodies with the goal of identifying patterns characteristic of diseases including Alzheimer's.
Zubarev suggested, though, that these approaches are inherently limited in that they rely on having molecules in their arrays that bind to the patient antibodies.
"You bind antibodies to this array of peptides and you see whether or not an antibody binds to that particular peptide," he said. "Then you know then the patient has an antibody against that sequence. You know the antigen — your array is an antigen array."
With his lab's mass spec-based approach, "we are looking at the sequence of antibodies. We don't know what [antigens] the antibodies are against," he said.
Stephen Johnston, a professor at the Biodesign Institute and inventor of the immunosignature approach used by HealthTell noted, though, that because the peptides used in this method are mimotopes, not epitopes, they can likewise "recognize any antibody, no matter what the source antigen."
In both cases this means the researchers can also capture information on immune responses to unknown and non-protein antigens that could be relevant to a given disease state.
"It might be a [protein] modification, or a protein aggregate," Zubarev said. "Sometimes antibodies are raised not against proteins but against different molecules — lipids, or modified lipids. So you don't know what to look for; you don't know how to purify [and] isolate your sample. You don't know where to look for the antigen."
Another potential benefit of looking at antibodies is that the immune cell replication involved in immune response provides a natural amplification of the signal, which could make it possible to detect disease earlier than with conventional protein markers, whose low abundance have often stymied biomarker researchers.
Essential to the method, which Zubarev and his colleagues have named SpotLight, is the ability to obtain complete sequence coverage for the antibody variable regions they are analyzing. To do this they enriched for IgG molecules via Melon Gel enrichment, then used a Thermo Fisher Orbitrap Fusion to fragment peptides using both higher-energy collisional dissocation (HCD) and electron transfer dissociation (ETD), which provided the coverage necessary to perform de novo sequencing.
The two methods "give different and complementary patters of dissociation, preferences, so what is not sequenced by an HCD is sequenced by ETD and the other way around," Zubarev said. He noted, as well, that the antibody variable regions the method focuses on are only around ten amino acids long on average, making full sequence coverage feasible.
The researchers focused their efforts on antibody variable regions that appeared to be shared across populations. As they wrote in the Nature Scientific Reports paper, in theory, "antibody recombination and point mutations can result in over 1015 different antigen-binding sites in humans," making for what would seem an impossibly vast search space.
However, they noted, studies have demonstrated that "antigen-specific antibody homology is more frequent than would be expected by pure chance," and that when different individuals are exposed to the same antigen, "the antibodies raised against this challenge should bind to it efficiently, which puts restraint on sequence variability of these antigen-specific [antibodies]."
This, Zubarev said, suggests that some antigen-specific antibodies may be present in large portions of a population and abundant enough to be measured using mass spec.
"What we're interested in is this gray area, the area of limited variability in the IgG sequences," he said. "We're looking at peptides that are found in at least half of the cohort of patients. The peptides that show up only once or twice, we discard. And even with that limited scope we still discovered a lot of new sequences that correlate with specific diseases."
In the recent study, Zubarev and his colleagues looked at whether addition of the IgG data improved their ability to distinguish between patients with Alzheimer's and DLB. Using only peptides identified via analysis of the non-enriched proteome, they were able to separate the two groups with an area under the receiver operating cure of .88. Adding the IgG data, the AUC rose to .94.
Zubarev said he is now working with other labs to apply the technique to a variety of diseases, though he declined to name any specific conditions they were investigating.
"There's a lot of interest from our colleagues studying different diseases because it's a new dimension of analysis that was previously unexplored," he said "So they give us samples, the cohorts, and we are looking at and finding new [antibody] sequences specific for that disease. Our long-term goal is to create a database of specific sequences that are characteristic of this or that disease."