Science Editor

Disease-specific antibodies that could be isolated from blood serum would make great diagnostic markers. But to date, few such antibodies have been identified. Now, a research team suggests that this may be because researchers have been looking for binding in all the wrong places – and showed that by using synthetic peptoids, they were able to isolate diagnostic antibodies for both an animal model of multiple sclerosis, and for Alzheimer's disease.

Currently, the way to isolate an antibody is to first get the antigen that binds to it. Thomas Kodadek, who is a research professor at the Florida campus of the Scripps Research Institute and senior author of the new paper, believes that may be because in many autoimmune diseases, the immune response is initially not to a normal protein, but to one that is somehow chemically altered.

The idea that autoimmune responses may be triggered by malformed proteins is "very much my theory, and I want to be careful to not be too vociferous about it," Kodadek cautioned.

"But if you think about how the immune system works, it's just not clear to me why a plain vanilla protein that is always in your body would trigger an autoimmune response."

In their paper, Kodadek and his team described making synthetic versions of such abnormal proteins in the form of so-called peptoids, which are based on peptides, or short amino acid chains. But "they have different connections between the atoms, and so they have different shapes and chemical properties" than regular peptides, Kodadek said.

One big advantage of peptoids, Kodadek said, is that they are "vehicles of convenience" in that it is easy to synthesize large numbers of them. "I think the important point was to present the antibodies with a huge collection of molecular shapes."

For their paper, which appeared in the Jan. 7, 2011, issue of Cell, Kodadek and his team synthesized a library of such peptoids and tested them both in an animal model of multiple sclerosis and serum samples from human Alzheimer's disease patients. In both cases, they were able to isolate peptoids that bound specifically to serum samples from disease tissue, indicating that there are disease-specific antibodies in the samples.

That such antibodies exist in multiple sclerosis is less of a surprise than that they exist in Alzheimer's disease, which is not an autoimmune disorder. Kodadek said that his team's results might be a sign of an unknown immune component to the disease, but also that "your immune system, even if it's not involved in initiating a disease or driving a disease, probably knows about a lot of disease states."

Overall, he said, "I'd like to think that the scope of this is extremely broad." In addition to the markers described in the current paper, his team has unvalidated markers for Parkinson's disease, lung cancer, pancreatic cancer and neuromyelitis optica, an autoimmune disorder that – like multiple sclerosis – attacks the myelin sheath, but mainly affects the optic nerves and spinal cord.

Kodadek and his team stressed in their paper that while they can identify individual antibodies that are diagnostic for the presence of disease, "this strategy is not an attempt to obtain a 'fingerprint' or 'signature' of a disease state." The reason, Kodadek said, is that for such signatures, binding strength needs to always be roughly the same, to enable a quantitative analysis.

"The data were not sufficiently reproducible to make a signature feasible." Such reproducibility might be possible with further tinkering, he said.

But even as the technology currently stands, the data were "certainly reproducible enough to tell the difference between a really bright spot and a really dark spot – and that's all we're trying to do."

The technology is licensed to specialty health care firm OPKO Health Inc. Kodadek said that OPKO plans to develop "a more standard assay format such as ELISA" that incorporates the approach. Once that is done, the company wants to make kits for specific diseases based on the technology. "We're planning to focus on diseases where early stage detection would have a major impact. . . . Cancers certainly come to mind immediately."

The Alzheimer's marker is furthest along, however, and Kodadek said that it may soon be used to stratify patients for clinical trials because "we have strong indications that we can detect early stage Alzheimer's disease, maybe even presymptomatically."

Such diagnostic abilities, he said, would be "a huge boon" to companies running clinical trials, and "probably" the best use of the test in the near-term future. "The current situation is nearly impossible for them. . . . They are trying to test these drugs on already demented patients, and no drug is going to regrow neurons."