MSACL 2016 EU Abstract

Multiplexed MRM-based Protein Quantitation in Human Plasma Using Two Different Stable Isotope Labeled Peptides for Calibration

André LeBlanc (Presenter)
McGill University - Lady Davis Institute

Authorship: André LeBlanc (1,2); Sarah Michaud (3); Andrew Percy (1); Darryl Hardie (1); Juncong Yang (1); Nicholas Sinclair (1); Jillaine Proudfoot (1); Adam Pistawka (1); Derek Smith (1); Christoph Borchers (1,2,4)
(1) University of Victoria - Genome BC Proteomics Centre, Victoria, BC, Canada; (2) McGill University - Lady Davis Institute, Montreal, QC, Canada; (3) MRM Proteomics, Inc., Victoria, BC; (4) Department of Biochemistry and Microbiology, University of Victoria, Victoria, BC, Canada

Short Abstract

Precise and accurate quantitation of the endogenous plasma proteome has tremendous potential for clinical applications. Targeted detection of peptides in a bottom-up strategy is the most common and precise MS-based quantitation approach when combined with the use of stable isotope labeled peptides. However, when measuring protein in plasma, these unknown endogenous levels prevent the implementation of ideal calibration strategies since no blank matrix is available. Consequently, several alternate calibration strategies are employed by different laboratories. In this study, these methods were compared to a new approach using two stable isotope standard (SIS) peptides, differentially labeled, enabling an external calibration curve and quality control samples to both be prepared in pooled human plasma without interference from endogenous peptides.

Long Abstract

Introduction

Precise and accurate quantitation of the endogenous plasma proteome has tremendous potential for clinical applications. Targeted detection of peptides in a bottom-up strategy is the most common and precise MS-based quantitation approach when combined with the use of stable isotope labeled peptides. However, when measuring large panels of protein in plasma, these unknown endogenous levels prevent the implementation of ideal calibration strategies involving standard peptides prepared in human plasma since no blank matrix is available. Consequently, several alternate calibration strategies are employed by different laboratories. These strategies include generating “reverse” standard curves (where endogenous and/or light peptides are employed to normalize the standards while the heavy peptides are used to normalize unknown samples) or using surrogate matrices for preparing standards. In this study, these methods were compared to a new approach using two stable isotope standard (SIS) peptides, differentially labeled, enabling an external calibration curve and quality control samples to both be prepared in pooled human plasma without interference from endogenous peptides. The ability to prepare samples of known concentration in human plasma with one SIS peptide while using the second SIS peptide as the internal standard in all samples is useful as a tool to evaluate the accuracy of peptide measurements in human plasma. It also allows us to compare different calibration strategies and evaluate surrogate matrices. Furthermore, this approach has the potential to simplify method development and validation of peptide-based assays in more regulated environments since the standard curve matrix is matched with the sample matrix and quality control samples can be prepared in human plasma without using reverse curves.

Methods

Standard peptides

Thirty-one peptides chosen from human plasma protein with varying hydrophobicities and endogenous concentrations were synthesized in-house (via Fmoc chemistry). These 13C/15N-labeled tryptic peptides were purified by RP-HPLC and characterized by MALDI-TOF-MS, amino acid analysis (AAA), and capillary zone electrophoresis (CZE) with UV detection. Two different labeled versions of the chosen peptides were synthesized: one with a labeled C-terminus lysine or arginine (Stable Isotope Standard; SIS 1) and another with an internal 13C/15N-labeled phenylalanine or leucine in addition to the C-terminal labeled amino acid (SIS 2).

Sample processing

The following steps were all performed by the Tecan Evo™ liquid handling robot: Raw pooled normal human plasma (10 µl) or surrogate matrix (pre-spiked human plasma, chicken plasma, dimethylated human plasma digest, phosphate buffered saline (PBS) and BSA solution (10 mg/ml in PBS)) were prepared by denaturation and reduction with 9M urea/20mM dithiothreitol. Denatured proteins were alkylated with 40mM iodoacetamide and diluted prior to overnight tryptic digestion. Digests were acidified and SIS peptide mixtures were added appropriately. Samples were finally concentrated by solid phase extraction using a mixed-mode reverse phase cartridge (Waters Oasis HLB).

For pre-spiked human plasma samples, SIS peptides mixtures were added prior the digestion step. For the dimethylated human plasma, reductive methylation of amines was performed according to litterature [1] immediately after digestion. Briefly, all primary amines in the human plasma digest were reductively dimethylated, shifting the masses of all endogenous peptides in order to create a human plasma-based blank matrix. SIS peptide mixtures were added after the reductive methylation reaction, prior to solid phase extraction.

LC-MS/MS

Digested samples (15 µg protein) were separated on-line by RP-UHPLC (Zorbax Eclipse column, 2.1 x 150 mm; Agilent) at 0.4 mL/min over a 30 min multi-step LC gradient. The LC system (Agilent 1290 Infinity) was interfaced to an Agilent 6490 QqQ via a standard-flow ESI source, operated in the positive MRM mode. The equivalent 5 optimized transitions were used to monitor all three isotopes of each peptide, light (endogenous) peptide, SIS 1 and SIS 2.

Data processing

The raw data was processed and the integration performed by Skyline software version 3.5. Quantitation was performed via regression analysis of peptide standard curves (1/x weighting). All standard and QC samples contained a constant amount of internal standard (SIS 1) and a variable amount of SIS 2 peptide.

Calibration strategies and surrogate matrices were evaluated and compared based on precision and accuracy via quality control samples (n=6, 3 QC levels in human plasma), using three isotopes (light, SIS 1, and SIS 2). Different calibration strategies used different isotope ratios to construct the standard curve and to calculate the concentrations of the quality control samples. Single point measurement and reverse curve strategies (where the endogenous peptide is used to normalize) were included in the evaluation. Intra-day and inter-day (2) precision and accuracy was assessed.

Results

Assay development

MRM parameters for each of the 31 peptides were optimized for the 5 most sensitive transitions. Then, assay concentration ranges were established by running a dilution curves of SIS peptides in human plasma digest. The limits of detection (LOD) and potential lower limits of quantitation (LLOQ) were estimated based on signal to noise ratios in human plasma and the endogenous concentration levels were also estimated using the dilution curve. The final assay concentration ranges were established based on the estimated LLOQ (spanning a 2000 fold range) and, if needed, were adjusted upwards to include the endogenous concentrations near the middle of the range. The final LLOQs for all peptides lie between 0.5 and 25 fmol on column, or 2.3 and 116.7 pmol per ml of plasma.

Calibration strategies

First, the different calibration strategies were compared. The two SIS peptide system in human plasma was compared to the reverse curve methodology as well as single point measurements. In order for the comparison to be fair, the SIS 2 concentration of the same QC samples were calculated using each strategy. The two SIS peptide method clearly provides consistently more accurate and precise results when compared to the reverse curve methods and single point measurements.

Surrogate matrix evaluation

Secondly, five different surrogate matrices were compared using the two SIS peptide system and all displayed comparably excellent accuracies at all QC levels as well as similar slopes. Calibration curves were prepared in chicken plasma, dimethylated human plasma digest (where all peptides are dimethylated in order to shift the masses of all endogenous human peptides), phosphate buffered saline (PBS) and BSA solution (10 mg/ml in PBS), in addition to human plasma. The only examples of matrix effects, for these 31 peptides, were displayed with chicken plasma and dimethylated human plasma. These matrices are more complex and the inaccuracies were a result of interferences in specific transitions. Once those transitions were discarded, the results for that peptide closely reflected those obtained from other matrices. It is important to note that, while all the peptides tested during this study performed well under the chosen conditions, the simpler matrices such as PBS buffer are more prone to variability due to peptide adsorption to the labware during sample preparation.

Pre-digestion spiking of plasma

Using two different labeled peptides also allowed for spiking standards and IS into human plasma prior to digestion. The calibration curve prepared in this fashion performed equally well as the calibration curve prepared when adding both SIS peptides after digestion, with negligible loss in sensitivity. Such an approach could increase sample throughput and facilitate automation since unknown samples, calibration standards and QC samples would undergo identical sample processing steps.

Conclusions

The calibration approach utilizing two stable isotope labeled peptides, one as the calibrator and the other as the internal standard added uniformly to all samples, has not previously been performed. This method allows standard and quality control samples to be prepared in human plasma without complications due to interference from endogenous proteins. With this method, assays can be developed that more closely reflect the standards set by regulated bioanalysis -- for example, peptide accuracy can be determined directly in human plasma, which is not the case when only one labeled peptide is available. The development and validation of peptide assays can be simplified since matrix effect evaluations, recovery tests and linear range determinations, to name a few, can be performed directly with appropriate standard concentrations in desired matrix.

Using this system we have shown the ability to assess the assay accuracy of endogenous peptide human plasma at any desired concentration, while using a stable labeled internal standard and without resorting to reverse curves. We show that two peptide calibration outperforms other strategies such as single point measurement and reverse curves. We also show that two distinct isotope standards can be used to directly evaluate surrogate matrices for standard curves.


References & Acknowledgements:

[1] Boersema et. al. (2009) Nature Protocols (4), 484 – 494.


Financial Disclosure

DescriptionY/NSource
Grantsno
Salaryno
Board Memberno
Stockno
Expensesno

IP Royalty: no

Planning to mention or discuss specific products or technology of the company(ies) listed above:

no