It may be a strange concept to think about, but 30 years ago it was only a pipe dream that researchers would be able to simply look up a gene’s coding sequence, find orthologs in other organisms, and then order designer mutations in cells and animals. But today’s reality is that an abundance of data has amassed on genetic variation——information that has proven indispensible to our understanding of basic biological processes and population genetics. Perhaps the most salient point for human health has been the identification of DNA sequence variation associated with various disease phenotypes.
While hundreds of genome-wide association studies have been conducted, identifying myriad genetic risk factors for diseases ranging from Alzheimer’s to type 1 diabetes, finding associated protein biomarkers of disease has proven more challenging for scientists. But with expanding proteome catalogues and advancing technologies, the identification of specific proteins associated with disease phenotypes could become more commonplace.
“Both of these illnesses are characterized by rather debilitating fatigue, but when you tell a doctor that you’re suffering from fatigue, the doctor yawns and says ‘me too,’” says Benjamin Natelson, Emeritus professor of neurology and neuroscience at the University of Medicine and Dentistry of New Jersey (UMDNJ). “It’s such a common complaint that most doctors just dismiss it, but for some people it’s so disabling that they can’t function normally.”
Because CFS and nPTLS are collections of symptoms with unknown etiologies rather than diseases, they are often regarded as “trashcan diagnoses”—that is, diagnoses that include a wide variety of symptoms and complaints. For instance, many causes have been shown to account for the symptoms seen in CFS: mercury poisoning; acute or chronic chemical exposure; and chronic viral, parasitic, or bacterial infections.
In the case of nPTLS, it remains unclear whether individuals who have been treated for Lyme disease exhibit symptoms of fatigue and cognitive impairment as a result nPTLS, or because they have a persisting bacterial infection. To this effect, it has been proposed that the spirochete bacteria responsible for Lyme disease often persist despite aggressive treatment with antibiotics, causing a chronic infection that manifests as symptoms of nPTLS.
“There are a number of symptom-based illnesses out there: firbromyalgia, migraines, depression, schizophrenia, et cetera,” notes Natelson. “Clinicians who develop case definitions for these conditions try to be very specific in defining them so that they’re relatively homogeneous, but because they have no specific diagnostic test, by definition they’re heterogeneous.”
Adding to this diagnostic dilemma is the fact that no biological markers exist that can distinguish CFS from nPTLS, resulting in frequent misdiagnoses and impeding research efforts into understanding each individual syndrome.
Mining the proteome
Natelson and colleagues previously tested the hypothesis that CFS had neurological origins by comparing protein profiles of cerebral spinal fluid (CSF) in patients who fulfilled the case definition for CFS with those of age-matched healthy volunteers (1). Their results showed that significantly more CFS patients had elevated cell counts and protein levels compared to healthy controls. In another study, Natelson also showed that CFS patients had reduced cerebral blood flow (CBF) compared with healthy individuals (2). To go beyond these studies, Natelson needed to survey how the proteins actually differed in CSF samples.
Given that CFS and nPTLS both fell under the rubric of neurologic disorders, Natelson, in collaboration with post-doctoral fellow Thomas Angel and the Richard Smith group at Pacific Northwest National Laboratory, as well as Steven Schutzer at UMDNJ, speculated that discrete proteins specific to each condition might be more abundant in the CSF of patients exhibiting characteristic symptoms.
“Not only is CSF an accessible liquid extension of the brain, it is in intimate contact with the tissues of the central nervous system, making it an ideal biological fluid to examine signature protein profiles of neurologic symptoms,” says Angel.
The group did have a starting point for their comparisons, in 2010 they completed the laborious task of surveying the proteome of normal CSF, a feat that enabled the group to then compare protein markers in patients with CFS and nPTLS.
CSF from three different patient groups was analyzed. The first group consisted of 43 patients who fulfilled the case definition for CFS; the second group consisted of 25 patients who had been diagnosed with, and treated for, Lyme disease but did not completely recover; the third group consisted of 11 healthy control subjects.
Pooled samples from all three groups were analyzed by liquid chromatography-mass spectrometry (LC-MS), a procedure that involves the separation of molecules within a mixed sample according to relative proportions and mass. During the LC phase, a sample is passed through a long, silica particle filled column. Because of the polar nature of silica, polar molecules within a sample remain attracted to the silica for longer, while less- or non-polar molecules pass quickly through the column. In this way, a mixed sample is separated into many sub-samples containing molecules of a specific charge. These purified samples are then vaporized by an integrated mass spectrometer and ionized via an electron beam. This results in the formation of charged particles which can be separated according to mass by an electromagnetic field, generating signatures that can be used to determine specific protein peptides.
By coupling high-throughput LC-MS with immunoaffinity depletion—a molecular technique by which large, abundant proteins are removed from a sample—the investigators were able to distinguish the less-abundant proteins specific to each pathology that may have otherwise been masked by more abundant ones.
“This was a wonderful opportunity,” says Natelson. “What the method does is provide multiple inputs for samples…so instead of just one LC platform feeding into one MS machine, it had 17 or 18.”
Applying these techniques to their CSF samples, Natelson and his colleagues identified more than 2500 proteins in each group. Of these proteins, 738 were specific to the CFS group, 692 were identified only in the nPTLS group, and 305 were common to both conditions. There were significantly fewer common proteins between each condition and the control.
What’s more, the group found that proteins specific to the complement cascade—part of the immune response that facilitates the clearance of infectious organisms—were increased in both CFS and nPTLS groups compared with the control. They also identified proteins involved in the CDK5 signaling pathway. Alterations in the CDK5 pathway have been previously linked to Parkinson’s and Alzheimer’s disease.
“Because we had two illnesses that were phenoypically similar, I would have expected that the shared proteins between them would be more common than the shared proteins between each individual illness and the controls… and that’s exactly what we found,” notes Natelson.
The promise of the proteome
For all the proteins identified, questions still remain as to why CFS and nPTLS occur at all, and what biochemical pathways are involved in the pathogenesis of these conditions. Furthermore when it comes to clinical applications, the use of CSF for diagnostic procedures presents several logistical problems: CSF collection is significantly more invasive than blood or urine and sample size is generally limited.
A crucial next step will involve the comparison of plasma and CSF proteomes in healthy and diseased states to determine whether specific markers present in the CSF of individuals with a particular condition reflect those present in the plasma. This would allow for more accurate diagnoses using blood.
At the same time, researchers will have to validate implicated proteins using targeted approaches—such as MS-based, immune-based, or a combination of these techniques—that will enable much higher sensitivity and specificity, more accurate quantification, and much higher throughput for measuring biomarker candidates in large clinical cohorts.
“We have and are continuing to develop unique technological capabilities that will allow for population scale proteomic studies—that is, quantification of the proteome for many thousands of human subjects in a single study,” says Angel. “Additionally, we are broadening the types of measurements we make on a sample from simply protein identification and abundance to other ’omics such as phosphoproteomics, glycomics, metabolomics, and activity based proteomics—collectively termed our “panomics” initiative.”
The paper “Distinct cerebrospinal fluid proteomes differentiate post-treatment Lyme disease from chronic fatigue syndrome,” was published 23 Feb 2011 in PLoS ONE.