In an effort led by Thermo Fisher Scientific's Biomarkers Research Initiatives in Mass Spectrometry Center, several leading proteomics teams have completed a study examining the inter-lab reproducibility of selected-reaction monitoring assays applied to clinical samples treated with upfront protein enrichment techniques.
According to one of the participants, George Mason University researcher Emanuel Petricoin, the study, which was published last week in the online edition of the Journal of Proteome Research, suggests growing clinical viability for mass-spec based proteomics. Additionally, it offers insight into certain instrument vendors' ongoing evolution into diagnostic players.
The project – which brought together scientists from facilities including Thermo Fisher's BRIMS Center, GMU's Center for Applied Proteomics and Molecular Medicine, the University of Toronto's Advanced Center for Detection of Cancer, Johns Hopkins University's Center for Biomarker Discovery, and several others – set out to measure cross-lab coefficients of variation of SRM-MS assays for a number of clinically relevant protein biomarkers isolated using upfront enrichment techniques.
To do this they developed standard operating procedures on a Thermo Scientific TSQ Quantum Ultra triple quadrupole instrument for four sets of analytes: prostate specific antigen in human serum enriched by an N-linked glycosylated enrichment procedure developed by Johns Hopkins researcher Daniel Chan; pituitary human growth hormone in human urine enriched by nanoparticle technology developed by GMU researchers Alessandra Luchini, Lance Liotta, and Petricoin; PSA in human serum enriched by Thermo Fisher's MSIA technology; and unenriched seminal plasma proteins.
Running SRM assays for these analytes across the four labs, the researchers achieved inter-lab CVs as low as 5 percent – for one of the peptides from the seminal plasma proteins – and as high as 30 percent, for the nanoparticle PHGH. Of the nine SRM assays tested, seven demonstrated cross-lab CVs of 21 percent or lower.
The results, Petricoin told ProteoMonitor, suggest that, while considerable testing and optimization remains to be done, "you can perhaps start to see the potential for how clinical mass spectrometry in the proteomics arena can match what has been happening in the small-molecule field for decades now."
He described the challenges facing mass spec-based clinical proteomics as a "kind of three-dimensional cube of barriers," observing that researchers are simultaneously beset by interlocking issues of throughput, technical performance, and cost.
As in other similar projects – perhaps most notably SISCAPA Assay Technologies CEO Leigh Anderson's ongoing collaborations with Agilent and Bruker (PM 12/2/2011) – the JPR team tackled these problems via the use of upfront enrichment procedures designed to concentrate the analytes of interest prior to mass spec analysis.
Enriching the target analytes during sample prep increases their concentration, making them easier to detect via mass spec, which in turn improves the SRM assays' CVs. It also reduces the complexity of the sample, allowing the use of high-flow chromatography as opposed to the nanoflow systems often used in proteomics work. The use of this more robust chromatography further improves CVs while also increasing throughput, which is key for clinical implementation.
"You might be able to get good CVs using very sensitive [nanoflow-LC] approaches, but then your throughput would be untenably low," Petricoin said. "So we were trying to hit a sweet spot with throughput while not sacrificing analytical sensitivity or precision and accuracy."
"What we were able to show is that even using sample enrichment steps in real clinical material, we were able to get CVs in a kind of clinical diagnostic range," he added. "So that in theory suggests that if you were to develop a protocol that uses a sample enrichment step and a high flow rate SRM-based workflow, you could develop a workflow that allowed lots of samples per day to be analyzed and with CVs that are in the strike zone of what would be doable in a clinical diagnostic setting."
According to Mary Lopez, director of the BRIMS Center, using its MSIA technology company researchers can run several hundred samples per day. In the JPR study the MSIA SRM-MS assay for PSA isoforms demonstrated CVs of 13 and 18 percent.
The CVs measured in the study accounted only for the SRM-MS portion of the workflow, not the sample prep or enrichment steps. Measuring CVs for this part of the process is a goal of the researchers moving forward, Petricoin said.
In this round "we wanted to just isolate the SRM assay itself to see if the downstream mass spec side of things could generate acceptable cross-laboratory precision and accuracy," he said. "Now that we have that, the next step would be to take the sample prep methods and have those samples run in the four different laboratories independently and then we could assess the impact of that variable on the CV."
The study, he noted, was part of a longer-term effort initiated by Thermo Fisher in 2007 to investigate the potential of SRM-MS as a clinical proteomics platform. In 2010 the researchers published an initial paper – also in JPR – investigating cross-lab reproducibility of SRM-MS for measuring peptides and standards spiked into human plasma.
The first two stages of the project complete, Thermo Fisher is now looking to implement some of the workflows "soup-to-nuts," Lopez told ProteoMonitor, and has partnered with Chan's lab at Johns Hopkins to incorporate the MSIA technology into SRM assays for PSA that Chan is investigating as part of the National Cancer Institute's Clinical Proteomic Tumor Analysis Consortium. The company, she added, is also working with Washington University in St. Louis researcher Reid Townsend to incorporate MSIA into his CPTAC work.
Petricoin suggested, as well, that the participants might in the future set up a validation trial of sorts "where we could maybe take [one of the researchers'] specific biomarkers that they are trying to validate anyway and build an interlaboratory analysis into that process."
Several of the researchers have discussed this idea among themselves, Petricoin added. However, he said, they have struggled with the question of where to get funding for such a study.
Who, exactly, will foot the bill for protein biomarker validation has proven a vexing question in the effort to drive mass spec-based proteomics into the clinic. While there's been no shortage of studies identifying potential biomarkers for all manner of diseases, tools and funds for rigorously validating these markers have proven more difficult to come by.
At a workshop held at the American Society for Mass Spectrometry annual meeting last month, representatives from several leading clinical reference laboratories offered a pessimistic take on the current potential of clinical proteomics (PM 5/25/2012).
Yet, despite the gloom, there are some positive signs for mass spec-based proteomics. This week, for instance, Quest Diagnostics introduced its new LC-MS/MS Thyroglobulin Post-treatment Monitoring Test – a mass spec test for measuring thyroglobulin to aid in monitoring recurrence of thyroid cancer following surgery.
On a larger scale, there is the continued investment by large mass spec vendors in clinical mass spec studies like the BRIMS-led JPR work. While many researchers and clinicians remain frustrated about the lack of tools and money to bridge the gap between discovery and clinical implementation, a number of large mass spec firms, including Thermo Fisher, Agilent, and Bruker, are engaged in efforts to develop clinical validation workflows and, suggested Petricoin, potentially transition into full-fledged diagnostics outfits.
"There's a reason why Thermo buys something like Intrinsic Bioprobes, and they're not doing it just to sell more [mass spec] boxes," he said. "I think the writing is on the wall, especially given the economy, that there are only going to be so many people who buy a [Thermo Scientific Orbitrap] Velos for $800,000 a pop. So the market saturates quickly, and these vendors are either going to go into the kind of Gillette mode where they try to sell massive amounts of consumables for clinical assays or they're going to develop the assays themselves and start offering diagnostic tests and developing [intellectual property] around the test method."
"I would suspect in the next few years we're going to see announcements where [companies like] AB Sciex and Thermo are going to be announcing clinical diagnostics, 510(k) approvals for specific diseases," Petricoin said. "There's precedent for this," he added, citing imaging companies like GE, Siemens, and Philips that "developed things like MRI and PET and are turning into diagnostic companies."
Indeed, efforts like the BRIMs-led collaboration or Agilent and Bruker's work with Anderson aside, mass spec companies have made a number of moves in recent months suggestive of clinical ambitions – from Agilent's acquisition of cancer diagnostics firm Dako (GWDN 5/17/2012) and its registration of its Infinity Series 1200 liquid chromatography systems and 6000 Series mass spectrometry systems as Class I medical devices with the US Food and Drug Administration (PM 1/20/2012); to AB Sciex's announcement of plans to likewise register certain of its mass spec instruments with the FDA (PM 3/16/2012).
Lopez agreed that vendors are increasingly exhibiting a clinical bent. "Thermo is a great example," she said. "You have this biomarker translational center at BRIMS where we have an internal customer – [Thermo Fisher's] clinical diagnostic division – and we do discovery for them and then they look at the commercial possibilities and whether these things can be developed into diagnostics. And you see this happening with other vendors."