Sixth Annual Meeting on Proteomic Sample Preparation, part of the second annual Getting Optimized Targets Summit

The Sixth Annual Conference on Proteomics Sample Preparation was held at the World Trade Center in Boston (MA, USA) on April 24–25, 2006. This meeting, one of eight concurrent or sequential Getting Optimized Targets Summits (GOTS) conferences within the week, highlighted issues and the most recent approaches to prepare samples for production of reliable and high-quality proteomics data. The critical importance of proteomics sample preparation is often overlooked, but it is very important to set high standards and develop rigorous protocols at the very beginning of research programs, since the final results are greatly dependent on the quality and processing of the original biological samples.

An excellent overview was provided by Scott Patterson’s keynote address (see [101] for a detailed summary of the talk), which focused on approaches used at Amgen (CA, USA) to discover pharmocodynamic biomarkers using flow cytometry and multiplexed planar arrays on stimulated whole blood. He highlighted the importance of paying careful attention to collection, storage and transportation of samples used for gene expression (RNA transcription) and proteomics analysis. Importantly, Scott provided strong data suggesting that the discovery of low-molecular-mass peptides in both serum and plasma as putative biomarkers (‘peptidomics’) is enormously complicated by the extensive role of ex vivo proteolysis – an essentially uncontrollable process. Peptides often better correlate with collection and storage parameters than with a disease state.

Richard Hegele (James Hogg iCAPTURE Center, Vancouver, Canada) gave the second keynote address, summarizing the value of his institution’s very large bank of cardiovascular and pulmonary tissues and associated clinical data. In addition to the technical aspects of tissue processing, preservation and storage, the center has managed issues of ethics, computational resources, and data security and quality.

The first breakout session, Sample Partitioning and Fractionation, emphasized the critical need to simplify and enrich samples prior to subsequent analytical steps. Typical tissue sources, such as serum/plasma or other biofluids, or tissue or cell extracts, are far too complex and have too great a dynamic range of protein concentrations for effective analysis by direct mass spectrometry (MS). Three speakers described their approaches to simplify and enrich samples for biomarker discovery.

Moira Lynch (BG Medicine, MA, USA) explained her company’s implementation of a high-throughput, high-capacity, automated and robust proteomics workflow. The system employs plasma delipidation, Beckman Coulter’s (CA, USA) avian antibody-based immunoglobulin (IgY-12) liquid chromatography immunoaffinity columns for selective partitioning of 12 highly abundant plasma proteins (5–10% of the starting plasma protein mass), followed by reverse phase capture of the highly enriched flow-through material, iTRAQ™ labeling, strong cation exchange chromatography, and MS analysis of differentially expressed proteins by both matrix-assisted laser desorption/ionization (MALDI) and liquid chromatography (LC)–mass spectrometry-MS/MS. An in-house-developed post-MS bioinformatics system was deployed for confident protein identifications, as well as careful sample randomization. The complicated multistep workflow was successfully deployed to run large-scale studies involving many hundreds of clinical samples.

Supporting the need to partition and fractionate plasma, Sunny Tam (U. Mass. Medical School, MA, USA) described a variety of methods to enrich low abundance proteins prior to 2D gels, MS or reverse protein microarrays. These included various immunoaffinity devices to selectively partition human serum albumin (HSA) and IgG (ABI and camel (VHH) regions), mouse albumin, IgG and transferrin (Agilent, CA, UAS), and 12 highly abundant proteins (IgY-12, Beckman Coulter). Fractionation methods included pI-based protein separation of a partitioned sample: PSL’s Multi-Chamber Electrolyser (dividing into three fractions) and Beckman Coulter’s protein fractionation-2D (PF-2D) (dividing into 700–1200 fractions with profiling software and imaging), or ion-exchange fractionation through a Mono-Q™ column.

Luke Schneider (Target Discovery, CA, USA) took a different approach by using a variety of immunoaffinity capture reagents to directly enrich specific proteins up to 400 times, combined with mass defect tags, for extremely sensitive isoform analysis by capillary electrophoresis or MS. This ‘needle magnet’ approach for deeper biological understanding and absolute quantitation of known targets nicely complements the ‘haystack remover’ approach to enrich middle- and low-abundance proteins for broad biomarker discovery. Labeling lysine or cysteine residues has been shown to work well with a variety of proteins, including bovine serum albumin (BSA), transferrin, prostate specific antigen (PSA) and α-fetoprotein and other biomarker discovery projects, as well as to measure postcollection protein degradation effects. Despite nearly quantitative labeling, capture with affinity-purified polyclonal IgY antibodies was only slightly affected.

The second session, Downstream Sample Preparation, featured three presentations on applications of sample enrichment to a variety of systems. The first was by Brett Chromy (Lawrence Livermore National Laboratory, CA, USA), explaining how they study the host response of certain animals to pathogens of biowarfare potential, such as Yersinia pestis, Bacillus anthracis, Salmonella enterica var. typhimurium and Staphylococcus aureus. The major approach was to deplete high-abundance proteins using the Agilent multiple affinity removal system (MARS) and Beckman BSA columns, and use 2D DIGE and surface-enhanced laser desorption ionization (SELDI)-MS to detect differentially expressed proteins in multiple tissue samples.

Jerry Feitelson (Beckman Coulter, CA, USA) then illustrated the benefits and use of IgY antibodies for sample partitioning of primate and rodent biofluids, with applications to a variety of biomarker discovery systems. His narrated talk and the one following addressing proteome fractionation by 2D liquid chromatography in the third session are available online [102].

Vyomesh Patel (National Institute of Dental and Craniofacial Research, MD, USA) applied laser capture microdissection techniques of formalin-fixed/paraffin-embedded tissues to study the proteomics of head and neck cancers. A variety of up- and down-regulated proteins were identified in tumor cells that could be biomarker candidates after extensive validation work is carried out.

At the conclusion of the second session, meeting participants could choose to participate in six well-attended give-and-take roundtable panels for lively discussions on a variety of interesting topics. The roundtable topics included:

•	The Importance of Protein Partitioning and Fractionation: Enrichment before Downstream Analysis – Jerry Feitelson (Beckman Coulter)
•	Proteomic Sample Preparation – Alex Lazarev (Pressure BioSciences, MA, USA)
•	Genomic Sample Preparation – Marcia Armstrong (QIAGEN, CA, USA)
•	Tissue Technologies – Sandra Gaston (Beth Israel Deaconess Medical Center, MA, USA)
•	Laser Capture Microscopy – Jim Wittliff (University of Louisville, KY, USA)
•	Labels and Stains for Tissues and Cells.

The topic of the third session was changed to ‘Advanced Sample Preparation’ to better reflect the speakers’ focus on novel methods to lyse, fractionate and analyze samples for proteomics analysis.

Jerry Feitelson (Beckman Coulter) delivered the second part of two talks, entitled ‘Multidimensional Liquid Chromatography of Proteomes,’ illustrating the principles and many applications of 2D liquid chromatography to proteome fractionation. The advantage of this top-down proteomics fractionation method is that post-translational protein modifications will be detected, which can then be exploited for biomarker discovery. The ProteomeLab™ PF-2D system separates proteins in the first dimension using chromatofocusing followed by reversed phase chromatography in the second dimension, thereby separating intact proteins based on pI and hydrophobicity. Examples were presented from a variety of laboratories and different biological systems, including the unambiguous identification of tissue-specific isoforms of troponin T from the heart and troponin C from fast-twitch skeletal muscle. This result illustrates that the combination of partitioning and fractionation allows identification of serum biomarkers in the unit ng/ml range, spanning a remarkable 10⁸ orders of magnitude below albumin.

The second speaker in this session, Alexander Lazarev (Pressure Biosciences, MA, USA) discussed his company’s very flexible, nonchemical approach to complete tissue and cellular lysis using pressure cycling technology, including disposable lysis chambers, appropriate buffers and programmable instrumentation. Data were presented on the successful use of this technology on a variety of sample types to recover DNA, RNA or protein, including very difficult-to-lyse Bacillus spores and bone fragments.

Finally, Chris Russell (Amgen) illustrated the perils and pleasures of developing proteomics expression assays on cells in whole blood ex vivo. The critical take-home lesson, similar to the Scott Patterson’s keynote address, is that studies must be designed so that the resulting biomarker candidates reflect differences between patients, patient populations or drug treatments, and not artifactual differences based on collection sites, study personnel, macromolecule isolation, transport and storage, or laboratory runs. The sample collection/storage process validated on one gene or protein should be generalizable to all genes or proteins. Several very careful studies on transcriptional analysis of signal transduction pathways in response to an ex vivo stimulus highlighted the key issues, such as measurements of sample stability and effects of citrate on coagulation that induces a strong inflammatory response, a ‘cytokine storm’. Enzyme-linked immunosorbent assays (ELISA) or TaqMan® or bDNA assays were used to confirm expression of candidate biomarker genes.

This meeting met its objectives of highlighting important and often overlooked issues in proteomic sample preparation crucial to the discovery of valid biomarker candidates and provided participants with several technologies and methods to successfully overcome these obstacles. There was general agreement in many of the lectures, hallway conversations and roundtable discussions that effective partitioning and fractionation to enrich low-abundance proteins is necessary to ‘dig deeper’ into the proteome, particularly when combined with high-sensitivity downstream analytical methods. While there are several competing technologies all along the sample prep continuum and there are not yet generally accepted standards, the future looks bright for discovery of disease-specific biomarkers when all factors are appropriately addressed.