An expert panel at this year's Hupo meeting discussed proteomics' lack of success at clinical biomarker implementation.
Very few biomarkers developed by mass spectrometry have been successfully introduced into preclinical experimentation or clinical practice.
Patrick Brown MD, PhD, professor in the department of biochemistry at the Stanford University School of Medicine, stated that the 'biggest shortcoming' of proteomics has been biomarker discovery, in particular the way in which hype from a few years ago and the consequent failure to live up to it 'unfairly damaged the proteomics community at large'.
He added that the field, unlike genomics, has had an 'extreme neglect' of variations: whereas genomics has focused on genetic differences and their possible meanings, the proteomics field has not focused on those areas, which has contributed to the field's comparatively smaller role in the clinic.
Lack of standardisation, reliability, and reproducibility has hampered the ability of the proteomics community to move biomarkers into the clinical arena.
The investment in complex mass spectrometers has been made.
Analysis of complex proteomes remains a daunting task, however.
Modern proteomic mass spectrometry (MS) of complex samples involves a workflow that culminates in proteolytic digestion of proteins followed by liquid chromatography and MS.
Spectra are matched with peptide hits using a searching algorithm such as Sequest, Mascot, or X-tandem.
Two overarching hurdles are the inability to adequately sample all proteins in a complex matrix due to the high dynamic range of protein concentrations (1011 for plasma), and to identify the same proteins reproducibly in different laboratories using different MS platforms.
The former problem has led to multidimensional prefractionation techniques prior to analysis on the mass spectrometer.
Invariably, investigators separate their proteins with SDS-Page, 2-D electrophoresis, Mudpit, or isoelectric focusing methods prior to liquid chromatography and MS.
However, each of these prefractionation methods suffers from lack of reproducibility.
Some of these methods, such as Mudpit, cause information loss as aspects of protein heterogeneity, based on post-translational modifications, are thrown away.
The latter problem can only be addressed when MS instrument manufacturers and software developers can unify MS/MS data searching in a way that is comprehensive and platform-independent.
Protein Forest has attacked the MS sample-prep reproducibility issue by developing a digitally formatted chip (digital proteomechip - dPC) that separates and concentrates proteins by their net charge.
The value of the digital format is that the chips can be manufactured with a high degree of precision such that the pH of each component within the chips has little variability.
This leads to a method where many of the uncontrolled variables apparent in isoelectric focusing strips and SDS-Page are eliminated.
Accordingly, total cell lysate data from dPC showed that the overlap in proteins identified by LC/MS/MS after multiple chips was 85 per cent with a CV of 8.2 per cent (n=6 pairwise comparisons.
This high degree of reproducibility has also been observed with other sample types, including plasma, CSF, urine, and semen samples.
This underscores the utility of the dPC for biomarker discovery and validation, and in diagnostics and forensics, where reproducibility is critical.
SDS-Page can be insensitive to post-translational modifications due to the small shifts in molecular weight engendered by the modification.
Complex lysates from the PC3 human prostate tumour cell line were separated by SDS-Page or dPC.
The mock and drug-treated samples from the SDS-Page separation showed no differences for the protein nucleolin.
DPC separation revealed three charge isoforms for the nucleolin protein; one form was found primarily in the mock sample and two more basic forms were found in the treated sample.
The data represents the sum of two replicates for each type of separation.
Clearly, isoelectric focusing mediated separation of nucleolin and other proteins with charge-based post-translational modifications allow for analysis that SDS-Page does not permit.
Charge-based separation of proteins is an important method to assess post-translational modifications such as phosphorylation and glycosylation.
Protein glycosylation represents a major post-translational modification and can have significant effect on protein function.
Changes in carbohydrate structure have been recognised as an important modification associated with cancer.
As an example, serum samples from diseased and healthy individuals were processed by an optimised glycoprotein sample-preparation protocol.
The goal was to identify changes in the concentration level and/or the carbohydrate structure of the glycoprotein(s) found.
The alpha one antitrypsin showed a change in peptide counts after dPC separation.
Parallel isoelectric focusing was first explained in 2003 by scientists at Protein Forest.
The dPC uses a discontinuous pH gradient, where each gel feature has a distinct pH (for example, a dPC 4.20 - 6.20 has 41 gel features separated by 0.05 pH units) with the electrical field perpendicular to the orientation of the dPC.
High electrical field strength is created, allowing rapid (30-45min) separation wherein the proteins are separated and concentrated at or near the isoelectric point of the protein.
Achieving a durable preclinical role for biomarker discovery relies on methods that are reproducible and reliable.
A new tool for biomarker discovery was created in response to the strong need for a simple, reliable, and reproducible method of protein prefractionation.
Digitising the separation of proteins by isoelectric point and changing the format to a parallel method has led to three measureable benefits: high reproducibility, improved coverage, and rapid separation.
Using dPC, complex protein samples such as plasma and cell lysates can be separated with a high degree of reproducibility in 30min.
The digital format dictates low lot-to-lot variability, leading to results that are repeatable across multiple laboratories using multiple MS platforms.
Coupled with the post-MSMS bioinformatics software platform, MSRAT, which is platform- and search engine-independent, it is now possible to reliably identify new and repeatable biomarkers.
Together dPC and MSRAT address sample reproducibility and the confounding effects of multiple-platform analysis that have limited proteomic biomarker identification and utility.
