Researchers from newly formed diagnostic company Durin Technologies and the University of Medicine and Dentistry of New Jersey have discovered a set of serum autoantibody biomarkers that strongly distinguish Parkinson's disease from non-disease controls and that they plan to develop as a commercial diagnostic.
The group reported in PLoS One this month that the ten-marker signature showed very high sensitivity and specificity — 93 percent and 100 percent, respectively — in separating patients with Parkinson's from healthy controls and also showed high accuracy picking out PD patients from subjects with other neurological diseases in a small proof-of-principle study.
Robert Nagele, founder of Durin Technologies and the leader of the UMDNJ research group, told ProteoMonitor this week that the promising results from the group's initial discovery and proof-of-principle work has now led to a larger validation project with support from the Michael J. Fox foundation. The group is almost finished with that 75-subject test. The researchers also plan to follow this with a larger validation of their initial work with Alzheimer's, which followed the same strategy.
Nagele said the company is still working out its commercial plans, which will most likely involve partnering with another company to support the US Food and Drug Administration approval process for an eventual assay. The researchers have developed and tested their antibody signatures for both AD and PD using Invitrogen Human Protein Microarrays. But Nagele said that since the team narrowed the field in each disease to only a handful of markers, it is most likely that a commercial test would be better served by targeting just those.
Because there are only about ten autoantibodies in each marker set, one likely strategy would be to create a multi-disease protein array incorporating both AD and PD, and potentially other diseases still in the group's pipeline, Nagele said.
In the group's recent PLoS One report, the researchers describe their discovery work, identifying a predictive signature of ten autoantibodies, and testing its predictive power.
The team divided 69 total serum samples — 29 Parkinson's and 40 controls — into a 35-subject training set and a 34-subject testing set. Using the Invitrogen arrays, each containing about 9,000 antigens, the group identified which autoantibodies in the training samples had significantly higher prevalence in the PD group than the control group.
Initially, the group found about 780 antibodies to be associated with Parkinson's. The researchers then ranked these, taking the ten with the largest difference in group prevalence to serve as the diagnostic set for further testing.
The team then tested the ability of these ten to distinguish between PD and controls in the same training set, as well as the separate testing set. The researchers found that the ten autoantibody markers were able to classify the full set of 69 PD and control samples with 93 percent sensitivity and 100 percent specificity.
To test the markers for disease specificity, the group also tested the signature's ability to distinguish the 29 PD samples from Alzheimer's, breast cancer, and multiple sclerosis with accuracies of 86, 96, and 100 percent, respectively, the authors reported.
While the approach the group has taken is agnostic toward the functional role of the proteins targeted by the antibodies in the signature, Nagele said it was encouraging for the group to find that two of the ten markers selected for PD overlapped with the AD signature — matching to "pentatricopeptide repeat domain 2" and "FERM domain containing 8" on the arrays.
Nagele said that for most of the autoantibodies and their associated antigens in both the AD and PD signatures, little is known about their function in the brain or in disease. "There are a [few] that do make sense, but for the majority, we don't know how they are connected to the disease. Yet, it works great as a biomarker."
The researchers hope, however, "that by revealing [what they are], pharmaceutical companies or others maybe will look at them and find out what pathways they are involved in and maybe some [will] lead to the development of new therapies, or show us pathways we didn't know were important before," he said.
The team's autoantibody array approach for both Alzheimer's and Parkinson's has been based on a hypothesis that autoantibodies targeting the cellular debris that results from various diseases can be used as precise and sensitive diagnostic biomarkers.
"What we published in our previous Alzheimer's disease paper is that we discovered that essentially all people are walking around with thousands of autoantibodies in their blood that we really didn’t know were there before," Nagele said.
"What we think is going on in all people, regardless of whether they have disease or not, [is that] these autoantibodies are involved in the clearance of day-to-day debris," he said. "But when you have a disease, it usually targets a particular organ. So let's say if its kidney disease … the end result of that is your kidney will generate a disproportionate amount of debris."
The Durin researchers posit that the immune system responds to the production of this increased debris by generating a larger number of autoantibodies that target such organ-specific damage. "Our diagnostic test detects [these] autoantibodies," Nagele explained.
Repeating the team's Alzheimer's success in Parkinson's disease has supported the researchers' initial hypothesis on the association of serum auto-antibodies with disease, Nagele said, and suggests the method could be a strong diagnostic approach for a variety of diseases.
"The [Parkinson's study] is absolutely a clone of what we did before. The methodology did not deviate one iota," he said. "It may be pie in the sky to a certain extent, but almost all diseases work by degenerating tissue and generating debris so theoretically this diagnostic approach should be usable for the detection of essentially all diseases, and that would be wonderful if that's true because it’s a very sensitive method — so we could get some of these diseases much earlier than we could previously."
Durin is not alone in using autoantibody signatures to detect neurological disease. In January 2011, researchers from the Scripps Research Institute and Opko Health published a paper in Cell on using synthetic molecules called peptoids to screen blood for autoantibodies associated with Alzheimer's disease (PM 1/7/2011).
According to Nagele, the researchers are almost finished with a Michael J. Fox foundation-funded follow-up to the PLoS One PD study in which they are testing the 10-marker signature on samples from 75 patients.
He said that of the 75 subjects, 25 have two samples available — one from early in their disease, and one from later. So the Durin team is also evaluating the test's ability to distinguish different stages of the disease.
"Not only will we use the 75 to validate the biomarkers, we will look at the 25 with two samples to see if we can distinguish stages of the disease as another proof of principle," he said. "And if we are successful … we'll go back to [the foundation], and say let's do a larger study with 50 at each stage of the five stages and see if we could diagnose them all."
This would be exciting for two reasons, he said. First, physicians would benefit from a way to track progression of the disease molecularly. Secondly, a hope of drug development for Parkinson's is to find treatments that can halt progression of the disease at early stages. Biomarkers that can reliably indicate stage could help with that research, Nagele said.
Ultimately, Nagele said, the hope is that his team will be able to establish the signatures for both PD and AD as not only diagnostic in the midst of disease, but also long before physical symptoms begin.
"The reason we think we might be successful is because … people [currently] diagnose Parkinson's by whether or not your hands are shaking," he said.
But researchers suspect that in both PD and AD, physical damage may begin several years before symptoms begin. "For Parkinson's disease … by the time your hands shake, some believe that as many as 50 percent of the neurons in the substantia nigra are already dead," Nagele said. "What you've been doing along the way is compensating for their loss, but you reach a point where you can't compensate any more."
Durin Technologies' hypothesis that the autoantibodies its researchers are detecting are involved in immune response to disease damage and debris bodes well for early detection.
"If cells are dying five years before you show symptoms, and we detect the fact that cells are dying …
we would hope our diagnostic should be able to detect disease years before symptoms appear … which would be great for pharmaceutical companies because everybody understands that any treatments [that are] going to be applied to these diseases are going to work best if given early," Nagele said.
The group is now planning a first test to try to demonstrate earlier detection in Alzheimer's, he said, working with patients suffering mild cognitive impairment, of whom typically only about 50 percent go on to be diagnosed with Alzheimer's.
Nagele said Durin is still unsure about its strategy for eventually commercializing a PD or AD test. Most likely, he said, the company would seek a partner to support the expense of the FDA approval process.
The company is also moving rapidly to expand its approach to other diseases. According to Nagele, the researchers are close to completing a study of multiple sclerosis mirroring those published for Alzheimer's and Parkinson's.
"The process we've developed is so fast now that we can actually develop a disease diagnostic [signature] in about a week and a half," Nagele said, as long as the group has access to patients and samples. "I envision if things take off — and I do think we're poised to take off — we'll be able to roll out disease diagnostics in rapid succession."
Have topics you'd like to see covered in ProteoMonitor? Contact the editor at mashford [at] genomeweb [.] com.