Researchers from the University of California, San Francisco, and biomarker-discovery firm PPD have used a liquid chromatography-mass spectrometry based method to identify a battery of proteins differentially expressed in cerebrospinal spinal fluid, in what they say is the first study in which a proteomic analysis of CSF yielded biomarkers with greater sensitivity than cytology.
The researchers combined high-resolution LC/MS data for profiling differential quantification without isotope labeling with LC/MS/MS for protein identification and found 76 proteins in CSF associated with central nervous system lymphoma, including a majority that had not been previously identified.
The study, which appears in the Jan. 1 edition of the Journal of Clinical Oncology
, demonstrates that “the CSF proteome contains a wealth of potential diagnostic and prognostic information,” the authors report.
According to the researchers, trying to pinpoint the cause of focal brain lesions in patients with unexplained neurological symptoms is a clinical challenge. Diagnostic methods for CNS lymphoma include flow cytometry, and in some cases, measuring beta-human chorionic gonadotropin or alpha-fetoprotein levels.
In an interview with ProteoMonitor this week, James Rubenstein, senior author of the study and an assistant professor of hematology and oncology at UCSF, said that among approaches using CSF, cytology is the “gold standard” in diagnosing brain cancer. However, the sensitivity of the method to diagnose lymphoma and other cancers is less than 50 percent.
The diagnostic method of last resort is brain biopsy, but in addition to a risk for brain hemorrhaging, the method carriers a failure rate as high as 35 percent.
Earlier research has suggested that the elevation of total CSF proteins may be associated with adverse prognosis in CNS lymphoma, although the identity of specific proteins and peptides had not been known. In addition, mass spectrometry has been shown to be able to detect such proteins to help diagnose certain cancers early in their development.
But confounding the ability to evaluate these proteins include a lack of high-resolution detection and peptide identification in blood samples, and a lack of proper validation of the studies evaluating mass spec-based methods, Rubenstein and his colleagues report.
The most common method for studying proteins in CSF has been high-resolution 2D gel electrophoresis with mass spectrometry, but the approach is plagued by low sensitivity and often yield results that are no better than those obtained by cytology, the authors claim.
For their work, Rubenstein and his colleagues identified the proteins using Thermo Fisher Scientific’s LTQ mass spectrometer. To profile their work they chose a shotgun, label-free approach, using online reverse-phase capillary HPLC coupled with high-resolution time-of-flight mass spectrometry.
The researchers say the findings demonstrate that “the CSF proteome contains a wealth of potential diagnostic and prognostic information.”
According to Chris Becker, executive director of chemistry at PPD, whose company funded the study in part and performed the proteomics work, one notable fact of the research is that it identified proteins on such a large scale: Rather than looking at a few proteins, which was the method used in many other proteomics-based cancer studies, the UCSF/PPD team looked at more than 1,000 proteins.
Additional funding for the study came from the National Institutes of Health, National Cancer Institute, American Cancer Society, American Society of Clinical Oncology, and other sources.
In addition, the label-free approach provided greater accuracy and the largest dynamic range, and provided for differential quantification, Becker told ProteoMonitor this week.
The team chose CSF as its medium because it has several properties “that facilitate differential proteomic profiling” compared to blood, including a lower volume — 150 milliliters on average, compared to 5 liters of blood, the authors write in the study. In addition, because it bathes exclusively in the CNS, CSF is not exposed to systemic circulation or to multiple organs.
“Both features favor the relative over-representation of brain and brain-tumor related proteins in the CSF,” the researchers write.
Alternative to CSF Cytology?
In the first phase of their work, the researchers compared CSF from eight patients with CNS lymphoma to CSF from nine controls and found 196 proteins expressed at different levels between the two groups. To validate the results, they then looked at the CSF of seven new patients with CNS lymphoma and the CSF of seven new controls.
Eventually, the team validated 76 proteins that were differentially expressed in CNS lymphoma. Most had never been previously described in the CSF in CNS lymphoma, according to the authors, and a number of proteins detected at the highest relative concentration are implicated “in tumor invasion and in the generation of tumor-associated extracellular matrix.”
In addition, the researchers found that several proteins known to play a role in regulating normal brain function were expressed at lower concentrations than their counterparts in control CSF samples.
The researchers then further validated their results by measuring with a sandwich ELISA the levels of a candidate marker, antithrombin III, in the CSF samples of 101 patients — 24 with primary and secondary CNS involvement by high-grade B-cell non-Hodgkin’s lymphoma — and 77 controls. They chose ATIII because preliminary data suggested it was a marker for CNS lymphoma, according to Rubenstein.
“We had independent microarray data suggesting it was expressed, and it had prognostic significance according to our microarray data,” Rubenstein said. His team confirmed those findings by finding that ATIII concentration in CSF was significantly higher in patients with CNS lymphoma than in controls. The results were unchanged after the team normalized for total CSF protein concentrations.
At a cutoff CSF ATIII concentration of 1.2 micrograms per milliliter the researchers observed 75 percent sensitivity and 98.7 percent specificity. By comparison, cytologic evaluation of CSF in the same 24 CNS lymphoma patients yielded 8 percent sensitivity, they report.
According to Rubenstein, follow-up work has uncovered “preliminary evidence to validate several other putative biomarker proteins among the 76 proteins described. Also, the majority of the proteins up-regulated in CSF in patients with brain tumors may be involved in angiogenesis, inflammation, and/or the immune response.”
The authors have filed for a US patent on their method with PPD as the assignee. The company is not currently developing a diagnostic based on the method, Becker said, though it may at some point seek a partner to do so.
In the study, Rubenstein and colleagues say that the proteins represent the “first novel” CSF biomarker linked to CNS lymphoma to be identified by LC-MS technology.
“Prospective measurement of individual CSF proteins such as ATIII may facilitate early, noninvasive diagnosis, risk stratification, and determination of the need and/or timing of therapeutic interventions in patients with indeterminate radiographic or cytologic findings,” they write.