Skip to main content
Premium Trial:

Request an Annual Quote

Hutch, Broad Teams Publish Comprehensive Methods for MS-Based Biomarker Discovery and Verification


By Adam Bonislawski

Two recently published workflows for protein biomarker verification offer a potential way forward for mass spec-based biomarker research, said scientists from San Diego's Proteogenomics Research Institute for Systems Medicine.

In a commentary published in this month's edition of Nature Biotechnology, PRISM director Jan Schnitzer and professor Martin Latterich highlighted a pair of studies - also published this month in the journal - that they said demonstrate "that mass spectrometry is becoming increasingly attractive for clinical diagnostics and has the potential to displace [enzyme-linked immunosorbent assays], the current gold standard for biomarker validation."

Mass spectrometry's specificity and multiplexing capability make it an appealing platform for protein biomarker research, but while it has been widely used for candidate discovery, issues with throughput, variability, and sensitivity have typically hampered verification and validation efforts.

According to Schnitzer and Latterick, the two Nature Biotech studies - one, on breast cancer markers, led by scientists at Seattle's Fred Hutchinson Cancer Research Center and the other, on cardiovascular disease markers, led by researchers at the Broad Institute - suggest strategies for overcoming these problems.

In contrast to the bulk of published protein biomarker studies, the two studies "cover [biomarker] discovery from beginning to end," Latterich told ProteoMonitor. "Historically, a lot of biomarker discovery efforts have focused on just the initial screening but have spent very little time on ultimately validating those original findings."

In particular, both studies used a series of mass spec approaches to winnow candidate biomarkers down to a number amenable to analysis by high-throughput, targeted quantitation.

The Hutchinson researchers began with 1,908 putative circulating biomarkers identified by integrating a series of independent genomic and proteomic datasets, including proteins that had been discovered in tumor tissue and plasma.

They then performed accurate inclusion mass screening of these 1,908 proteins on a Thermo Scientific LTQ-Orbitrap to determine which were detectable in plasma, identifying 572 candidates. These they then analyzed via semiquantitative selected-reaction monitoring mass spec on an AB Sciex 4000 QTRAP, identifying 49 candidates - along with seven housekeeping proteins - for which they built a multiplex SRM-MS assay that they ran on an AB Sciex QTRAP 5500. They also selected for SISCAPA analysis - which combines antibody-based peptide enrichment with SRM mass spec - an additional 35 proteins whose signals in the SQ-SRM stage had proven too low to measure.

The researchers ran these 88 assays in triplicate on 80 plasma samples derived from three cohorts of normal and tumor-bearing mice, finding that of the 57 proteins measured via SRM-MS, 30 showed significant elevation in tumor-bearing animals, and that six of the 17 proteins measured using SISCAPA were elevated.

The Broad team similarly began with a large candidate list, using LC-MS/MS on a Thermo Scientific LTQ-Orbitrap FT instrument to identify 1,105 unique proteins in coronary sinus plasma samples from patients undergoing a therapeutic, planned myocardial infarction for treatment of hypertrophic cardiomyopathy. They then performed label-free, relative quantitation of these proteins, identifying 121 that showed increases of more than five-fold compared to baseline.

These proteins they analyzed via AIMS to determine which could also be detected in peripheral blood samples, generating a prioritized list of 52 candidate proteins for SRM-MS assay development.

SRM-MS has drawn interest as a platform for protein biomarker validation and clinical use because of it multiplexing ability and the relative ease of developing such assays compared to ELISAs. For the breast cancer study, the Hutchinson researchers developed and characterized 88 SRM assays in less than one year. By comparison, they noted, "it would be extraordinary for a single academic laboratory to successfully configure 1-10 ELISA assays de novo in 1-2 years, especially at comparable cost."

"Once you have a very good antibody that is highly specific, that's a tremendous reagent to have," Schnitzer said. But, he noted, with mass spec "you don't have to wait for the development of a reagent - that's the big point."

The methods presented in the Nature Biotech papers, he said, offer workflows for combining several mass spec techniques to efficiently move from biomarker discovery through to verification and validation by targeted assays like SRM.

"You can think of it as using mass spec in the discovery phase and now dialing in to specific candidates and using the mass spec in a more [quantitative] fashion [for measuring] a subset of proteins you're interested in," he told ProteoMonitor.

One concern surrounding the use of SRM-MS for biomarker validation or as a clinical platform has been its variability - in particular the variability associated with the sample prep steps that typically precede triple-quadrupole analysis.

As George Mason University professor Emanuel Petricoin noted last year in an interview with ProteoMonitor, "triple quadrupoles can have fantastic [coefficients of variation], but then what's the CV of the trypsin digestion, what's the CV of the immunocapture [in the case of SISCAPA-based assays], what's the CV of the entire process? The 600-pound gorilla in the room is that the CVs of that process are actually very high" (PM 10/22/2010).

The Hutchinson researchers, however, observed CVs of less than 15 percent for the majority of their SRM assays, which is in line with CVs of some existing clinical assays.

Automation of the workflows could also improve CVs, Latterich said.

"The biggest source of experimental variability at this point is at the sample preparation and separation steps," he said. "Those are relatively manually intensive, so various people have started to look into how to possibly automate those early steps."

In particular, a number of efforts are underway to automate the SISCAPA workflow, including a platform developed by researchers at the Translational Genomics Research Institute that is capable of preparing roughly 1,000 SISCAPA samples for mass spec analysis every 24 hours (PM 02/11/2011).

Leigh Anderson, one of the inventors of the SISCAPA method, has collaborated with Agilent to also develop an automated SISCAPA workflow using an Agilent Bravo liquid handling system for sample prep attached to an Agilent 1200 series LC system and an Agilent 6490 triple quadrupole mass spec.

In June, Anderson launched a new firm, SISCAPA Assay Technologies, to commercialize the assay (PM 06/17/2011). Thus far the company has licensed the technology to Pfizer for internal use and development and this week announced that contract research organization Advion BioServices will be offering SISCAPA assay services.

Have topics you'd like to see covered in ProteoMonitor? Contact the editor at abonislawski [at] genomeweb [.] com.