Yuri van der Burgt(1,2), Irene van den Broek(2), Nico Smit(2), Fred Romijn(2), Jan Nouta(2), Marco Bladergroen(1), Arnoud van der Laarse(2), Wilma Mesker(3), Rob Tollenaar(3), Christa Cobbaert(2)
(1) Center for Proteomics and Metabolomics (2) Department of Clinical Chemistry (3) Department of Surgery
Numerous serum protein profiling efforts have aimed for biomarker discovery since the early days of mass spectrometry(MS)-based proteomics. These pipelines have yielded many promising protein candidates, but disappointingly none of these has been translated into a truly validated medical test. Alternatively, strategies have been followed to convert existing routine uniplex protein assays into an MS-based in-vitro diagnostic test, however also in this case the success rate has been low so far. In this presentation both approaches will be discussed with regard to requirements needed for translation or implementation these into a clinical chemistry laboratory.
Numerous serum protein profiling efforts have aimed for biomarker discovery since the early days of mass spectrometry(MS)-based proteomics. These pipelines have yielded many promising protein candidates, but disappointingly none of these has been translated into a truly validated medical test. Alternatively, strategies have been followed to convert existing routine uniplex protein assays into an MS-based in-vitro diagnostic test, however also in this case the success rate has been low so far. In this presentation both approaches will be discussed with regard to their potential to translate or implement these into a clinical chemistry laboratory. Crucial is awareness of the conditions, or requirements, needed for such routine applications (Smit et al. Translational Proteomics 2014).
State-of-the-art MS-based proteomics pipelines will be overviewed in the context of providing diagnostic or prognostic markers, often not as a single species but rather as a panel or signature of proteins. As an example, we have developed fully automated and standardized sample preparation strategies in combination with ultrahigh resolution MS to obtain cancer-specific peptide and protein signatures. Also, we have used the inherent multiplexing capacity of an MS-based assay to quantify different apolipoproteins via targeted peptide measurements (multiple reaction monitoring, MRM). Finally, the approach of combining protein immunocapture with MS-based quantification will be discussed and exemplified.
Three different MS-based proteomics strategies have been followed to either profile or quantify serum proteins in patient cohorts. In the first case, serum samples were obtained from cancer patients prior to surgery, and from healthy volunteers (“controls”) with informed consent. This study was approved by the Medical Ethical Committee of the LUMC. Samples were collected and processed following an in-house developed fully automated high-throughput magnetic bead-based solid-phase extraction (SPE) procedure to enable peptide and protein profiling studies. Briefly, the isolation of peptides and proteins was performed using 10 microliter of Invitrogen Dynabeads for each analysis of 5 microliter human serum on a fully automated Hamilton liquid-handling robot platform. After washing and desorption two microliter of eluate material was used for matrix-assisted laser desorption ionization (MALDI) spotting and profiles were acquired on a Fourier transform ion cyclotron resonance (FTICR) MS system equipped with a 15T magnet. In a second strategy, an SPE-protocol was used on the same robotic platform to first immunocapture specific tryptic peptides with commercially available anti-peptide antibodies (SISCAPA) and then quantify the peptide using a MALDI time-of-flight system (van den Broek et al. Methods 2015). In the third case, serum apolipoproteins were quantified via a bottom-up workflow combined with MRM-measurement on an Agilent tripleQ MS-system. A careful design of experiments was followed to optimize digestion conditions and native, value-assigned sera were used for external calibration thus allowing accurate protein quantification.
Profiles of complex mixtures of peptides and proteins were analyzed at a high-throughput using fully automated sample processing strategies in combination with ultrahigh resolution 15T MALDI-FTICR MS. This platform provided “next-generation profiles” of general serum peptides and proteins that were used for “case-control” classification based on a set of previously characterized peptides using relative intensities. Full automation of the preparation and analysis process ensured standardization together with high discriminative power. With this approach, endogenous peptide and protein signatures were found for colorectal cancer and pancreatic cancer. In the latter case, the MS-results were compared with the mostly used clinical serum biomarker, namely carbohydrate antigen 19-9. Although it was concluded that these peptide signatures are a potential candidate for early disease detection, the road towards clinical implementation is not straight-forward. This is partly due to the difficulties encountered with regard to quantifying peptides and proteins in such profiles and the lack of external calibrators. Moreover, robustness is an issue since these profiles are sensitive for any change in sample collection as well as -processing procedure. In this context, quantification of proteolytic peptides may have superior advantage, provided the digestion step is carefully controlled. Requirements for implementation (or translation) of such strategies into routine clinical practice will be discussed.