Targeted proteomic assays for the verification of global proteomics insights

Introduction

In addition to traditional antibody-based targeted assays, recent rapid advances in mass spectrometry (MS)-based proteomic technologies have positioned selected reaction monitoring (SRM), parallel reaction monitoring (PRM), and Sequential Windowed Acquisition of All Theoretical Fragment Ion Mass Spectra (SWATH-MS) as powerful targeted MS-based methods that can be used to verify biological insights gained from discovery-phase proteomics studies. However, the robustness of these targeted approaches is largely dependent on factors that are instrument platform-independent, such as careful attention to experimental design, strict adherence to best practices, and the proper statistical control of the data processing.

Chasm between proteomics discovery and the development of targeted assays for verification

The field of proteomics has relatively few fully validated assays for clinical use. Despite the large number of protein candidates discovered (>1000 cancer-associated protein candidates), only a handful have been clinically qualified and cleared for use by the US FDA [1]. It can be argued that the lack of FDA-approved protein tests is largely caused by a lack of standardized technologies and methods for verifying the candidates, the majority of which have limited clinical utility.

There is an expectation for recent advances in MS-based proteomics technology and the associated analytical approaches to drive improvements in health care that are reliant upon the development of sensitive and specific clinical protein tests. Indeed, proteomics technologies have revolutionized cell biology and biochemistry by providing powerful new tools to characterize complex proteomes, multiprotein complexes, and posttranslational modifications. However, the widespread unrealized potential for proteomics technologies to address critical problems in clinical and translational bio-medical research needs to be fulfilled. The chasm between target discovery and the development of clinically applicable assays needs to be closed to provide an appreciable return on the large investment that has been made in MS-based protein assay development efforts.

SRM: gold standard targeted MS-based quantification requiring gold standard best practices

SRM is a targeted MS-based technology that has considerable potential to narrow the gap between protein target discovery and the translation to clinical utility. SRM was first applied to the quantification of peptides in biological tissues more than 30 years ago [2]. In contrast to MS-based discovery proteomics experiments, SRM entails the targeted, simultaneous measurements of peptides that serve as surrogates for the protein targets of interest that are often culled from discovery proteomics experiments. SRM-based assays are considered to be the ‘gold standard’ for MS-based targeted protein quantification because they are highly specific, precise, and accurate, and they can be multiplexed (hundreds of peptides can be quantified in a single assay), standardized, and readily reproduced. Kennedy et al. recently demonstrated the feasibility of the large-scale development of 645 standardized SRM assays [3].

SRM experiments typically consist of the following steps: (1) generation of a list of target proteins and a fit-for-purpose quantification strategy; (2) study design and experimental planning; (3) synthetic peptide preparation; (4) method development and refinement; (5) data acquisition; and (6) analysis and modeling (Figure 1). A common fit-for-purpose quantification strategy entails the use of stable isotope-labeled internal peptide standards for relative quantification based on the establishment of calibration curves of dilutions of the peptide mixtures. To facilitate the synthetic peptide preparation stage, practical recommendations are available for the generation, quantification, storage, and handling of peptides used for SRM assays [4]. Additionally, the method development and refinement stage has been aided by the availability of several public databases containing lists of peptide analytes and transitions including SRMAtlas (www.srmatlas.org) and the PeptideAtlas SRM Experimental Library (PASSEL; www.peptideatlas.org).

Figure 1.

Common workflow for SRM assay development.

Efforts to develop SRM assays for the verification of global proteomics insights are resource-intensive; hence, the establishment of commonly accepted best practices is essential. A primary goal of establishing best practices for these Tier 2 SRM assays is to achieve precise, relative quantification that can be harmonized across laboratories, increasing the replicability of research and enabling the robust quantification of peptides and proteins. The major challenges to achieving this analytical bar for SRM best practices include (1) selecting peptides that can be measured with high precision and repeatability in the matrix of interest; (2) generating well-characterized, synthetic peptide internal standards, and calibrators; (3) assuring the quality (e.g. concentration and stability) of the peptide internal standards and calibrators in lyophilized form and in solution over time, during storage and handling; and (4) properly interpreting peptide-based measurements. It should be noted that crude synthetic peptides are acceptable for the development of Tier 2 SRM assays that are deployed to assess the relative quantification of proteins across several samples. However, Tier 1 SRM assays that are used to determine absolute protein quantification for clinical applications must be developed using purified synthetic peptides with accurately determined concentrations. Toward the rapid and standardized performance of SRM-based verification studies, assay assessment standards, methods (for acquisition and analysis), and data analysis tools such as Qualis-SIS [5] and MRMPlus [6] have been developed.

PRM, first published in 2012 [7], is a targeted proteomics strategy where all product ions of the target peptides are simultaneously monitored at high-resolution and high-mass accuracy. PRM analyses exhibit performance characteristics (dynamic range and lower limits of detection and quantification) that are similar to those of SRM. In PRM, the third quad-rupole of a QQQ (triple quadrupole) mass spectrometer is substituted with a high-resolution and accurate mass analyzer to permit the parallel detection of all target product ions in one high-resolution mass analysis. Although PRM reduces the burden of method development and refinement, it is still essential to adhere to best practices for MS-based targeted proteomics methods to ensure robust peptide and protein quantification.

Among the persistent challenges in MS-based targeted proteomics strategies are the high throughput, reproducible, and ideally automated sample preparation methods to generate peptides from the targeted proteins from a large number of specimens, need for methods for enrichment of low abundance targets, assays for peptides with specific modifications such as phosphorylation or glycosylation, and the need to increase the degree of multiplexing in a given analytical run [8–11]. To address the detection sensitivity challenge, SRM combined with immunoaffinity enrichment strategies such as stable isotope standards and capture by anti-peptide antibodies or the specific enrichment of sub-proteomes have been shown to improve detection sensitivity by several orders of magnitude [9,12,13].

SWATH-MS: untargeted data acquisition, but targeted data analysis

An untargeted MS-based proteomic method with the potential to overcome some of the throughput limitations of SRM is SWATH-MS. This technique was first published in 2012 as an implementation of the data-independent acquisition strategy that aims to maintain the robust quantitative characteristics of targeted proteomics but on the scale of thousands of proteins [14]. SWATH-MS analyses are typically carried out using a Triple-TOF mass spectrometer that includes features such as broad dynamic range, rapid acquisition rate, and dynamically adjusted MS/MS acquisition time based on precursor ion intensity. An advantage of SWATH compared to quantitative data-dependent proteomic acquisition strategies used in discovery proteomics is its potential to minimize under-sampling and stochastic and irreproducible precursor ion selection.

SWATH has not yet been widely accepted as a targeted quantification tool due to its status as a recently developed technique that was first published in 2012 [14]. The results from a systematic comparison of the performance of SWATH with SRM indicate similar performance of the two methods with a twofold to threefold less favorable limit of detection for SWATH compared to SRM, but with high throughput and number of peptides and proteins analyzed [15]. However, SWATH certainly has the potential to generate highly reproducible and quantitatively accurate results, and the method is well suited for the orthogonal verification of targets that are identified from discovery-phase proteomic studies [16]. An average coefficient of variation (CV) of 20% was achieved from the SWATH analysis of >340 plasma proteomes while maintaining high reproducibility [15].

Several groups have demonstrated that SWATH achieves robust protein and peptide quantification and excellent correspondence of protein identifications [16–19]. For example, a recent study coupled SWATH and SRM in a biomarker discovery and verification effort, respectively [18]. The results from these and other studies indicate that SWATH enables precise label-free quantification on a proteome scale; thus, the method has great potential for clinically oriented studies.

One of the main challenges associated with SWATH analyses is the accurate processing of the data resulting from the multiplexed MS/MS spectra containing fragment ions that originate from multiple precursor ions. Although there are commercially available software programs for the processing of SWATH data, improvements in the quality of data analysis can be made through the widespread availability of open source programs such as OpenSWATH [20] and DIA-Umpire [21]. Additionally, similar to the analysis of data from data-dependent experiments, precise control of the false discovery rate (FDR) at the peptide and protein levels is critical for the accurate processing of SWATH data. For projects entailing large-scale SWATH-based analysis, it is important to not only control the FDR at the single SWATH analysis level, but also within a global context including all the files that are part of the analysis.

Summary

SWATH and SRM are powerful targeted MS-based methods that can be used to verify biological insights gained from discovery-phase proteomics studies. However, the successful implementation of these approaches to facilitate their future clinical utility is largely dependent on instrument platform-independent factors including experimental design, strict adherence to best practices, and the proper statistical control of the data processing.