Research Article Received: 7 January 2014
Revised: 27 February 2014
Accepted: 27 February 2014
Published online in Wiley Online Library
Rapid Commun. Mass Spectrom. 2014, 28, 1101–1106 (wileyonlinelibrary.com) DOI: 10.1002/rcm.6883
Simultaneous quantitation and size characterization of apolipoprotein(a) by ultra-performance liquid chromatography/mass spectrometry Michael E. Lassman1*, Theresa M. McLaughlin1, Haihong Zhou1, Yi Pan1, Santica M. Marcovina2, Omar Laterza1 and Thomas P. Roddy1 1
Merck Research Laboratories, 126 E. Lincoln Ave., Rahway, NJ 07065, USA Northwest Lipid Metabolism and Diabetes Research Laboratories, University of Washington, 401 Queen Anne Ave, North Seattle, WA 98109, USA
2
RATIONALE: Apolipoprotein(a) is a polymorphic glycoprotein covalently bound to apoB100 in Lp(a) particles and has been described to be both atherogenic and prothrombotic, although its exact mechanism of action is not well defined. Apolipoprotein(a) is routinely measured by immunoassays. Unfortunately, the accuracy of the measurement can be affected by the apolipoprotein(a) size (number of kringles) polymorphism in Lp(a) particles. Here we describe an ultra-performance liquid chromatography/mass spectrometry (UPLC/MS) assay that is capable of measuring apolipoprotein(a) concentrations while simultaneously determining the number of kringles present per protein. METHODS: Plasma samples were diluted and proteins de-lipidated with deoxycholate prior to tryptic digestion. Distinct tryptic peptides from different regions of apolipoprotein(a) were measured to determine both concentration and the number of kringles present per protein. Separation and quantitation of tryptic peptides is carried out at 700 μL/min using a 1.7 μm C18 column (2.1 × 100 mm) coupled to a Thermo Vantage triple quadrupole (QQQ) mass spectrometer with a heated electrospray ionization (HESI) source. RESULTS: This method was compared to established methods for measuring concentration (monoclonal antibody based ELISA) and size (gel-electrophoresis) using 80 plasma samples proved by NWLRL. The slope and r2 value for the correlation of concentrations were determined to be 0.96 and 0.98, demonstrating excellent agreement of absolute values between the UPLC/MS and ELISA methods. As measured by UPLC/MS, the average kringle number or size is smaller than determined by the electrophoretic method. CONCLUSIONS: A single UPLC/MS method was developed capable of measuring apolipoprotein(a) concentration and size (by measuring the number of kringles per protein). This assay passes criteria required for ’fit for purpose’ assays including sensitivity, intra and interday reproducibility and freeze/thaw stability. While the agreement between UPLC/MS and ELISA is excellent for concentration and may provide researchers with additional tools for studying apolipoprotein(a), the dissimilarities between UPLC/MS and the electrophoretic method may also be exploited for understanding apolipoprotein(a) structure and function. Copyright © 2014 John Wiley & Sons, Ltd.
Lp(a) is a lipoprotein that has been reported as being both proatherogenic and prothrombotic.[1–3] Structurally, Lp(a) is similar to another proatherogenic particle, LDL, with the distinction that Lp(a) contains the polypeptide protein apolipoprotein(a) (Apo(a)), a polymorphic glycoprotein covalently bound to apoB100 by a disulfide bond.[4] The concentration of circulating Apo(a) is genetically determined[5,6] and has been reported to be inversely correlated with Apo(a) size. This inverse association between plasma Apo(a) concentration and size has been demonstrated to be primarily due to differences in Apo(a) production rates.[7] Multiple research studies have included measurement of Apo(a) kinetics as part of an effort to understand its role and function.[8–10] Still, despite many large studies, the exact function of Apo(a)
Rapid Commun. Mass Spectrom. 2014, 28, 1101–1106
Copyright © 2014 John Wiley & Sons, Ltd.
1101
* Correspondence to: M. E. Lassman, Molecular Biomarkers and Diagnostics Laboratory, Merck Sharp & Dohme Corp, Rahway, NJ 07065, USA. E-mail: [email protected]
particles is not completely understood, nor is the relationship between disease and concentration/size. Apo(a) size is defined primarily by the number of kringle 4 type 2 (KIV2) repeats in Apo(a)[4] which can be experimentally determined by high-resolution sodium dodecyl sulfateagarose gel electrophoresis followed by immunoblotting,[11] a technique that is performed by a limited number of laboratories. Apo(a) concentration can be accurately determined by a double monoclonal antibody-based enzyme-linked immunoassay that is by design, insensitive to Apo(a) size heterogeneity, NWLRL ELISA.[12] Alternatively, a limited number of commercially available assays exist including an immunoturbidimetric assay that uses an antibody that does not cross-react with plasminogen or apolipoproteinB as well as a lectin-based assay that captures Lp(a) and measures the cholesterol content in the particle.[13] Lamon-Fava et al. compared the commercially available assays to the Northwest Lipid Research Laboratories (NWLRL) ELISA and found that they were well correlated. However, the immunoturbidimetric assay measures concentration in mg/dL rather
M. E. Lassman et al. than nM, which is problematic due to the heterogeneity in Apo(a) molecular weight. As a result, the immunoturbidimetric assay was demonstrated to have a strong Apo(a) isoform size-dependent bias. The lectin-based method was determined to over-estimate the amount of Lp(a) cholesterol in samples with low Apo(a) values.[13] In order to advance the field and understanding of the role of Apo(a) in disease, readily available assays must be available that are capable of supporting large clinical studies, accurately measuring Apo(a) concentration in units of nM and characterizing Apo(a) size. Multiple laboratories have proposed the use of mass spectrometry (MS)-based approaches as an alternative to immunoassays,[14] including measurements of apolipoproteins.[15–18] An MS-based approach has advantages in that it does not require the generation or use of specific antibodies and instead relies on the specificity of liquid chromatography (LC) and multiple reaction monitoring (MRM) in quantitation of proteotypic peptides. Furthermore, these assays are readily multiplexed for simultaneous measurements of multiple analytes. For Apo(a), an MS-based approach can be used to characterize size by measuring the concentration of kringle 4 type 2 repeats. An MS-based approach also has the advantage that it can be readily modified and applied to Apo(a) research in non-human primates, which also have Apo(a), although the sequence is not completely conserved. The application of MS-based quantitation of proteins by MRM has been described repeatedly for plasma proteins, with single and multiplexed assays reported in the literature.[19–23] These assays have been used for the analysis of endogenous as well as therapeutic proteins,[24,25] but are essentially similar. In general, MRM assays use enzymatic digestion of the targeted protein within a complex protein matrix to generate proteotypic peptides that are measured by LC/MS. The use of stable isotope labeled standards corrects for variation in sample handling and peptide degradation as well as analytical variability and allows for quantitation of the peptide as a proxy for the intact protein. The throughput and instrument availability of LC/MS-based methods are limited compared to many immunoassays, but ultra-performance liquid chromatography (UPLC) methods have made throughput more competitive with immunoassays and the instrumentation required to perform these assays is becoming more commonplace in academic and clinical laboratories. Furthermore, compared to multiple immunoassays, a single multiplexed UPLC/MS-based method becomes even more competitive in terms of cost and throughput. Here, we demonstrate a single UPLC/MS assay based on an apolipoprotein platform previously reported by our group[17] capable of measuring apolipoproteins A1, B48 and B100 in multiple species which is now applied for the first time to measurements of Apo(a) concentration and size. The enzymatic digestion, sample preparation and UPLC/MS analysis of plasma samples are performed in 96-well microtiter plates, allowing for a daily throughput of over 200 samples per day.
EXPERIMENTAL Chemicals
1102
Ammonium bicarbonate (ABC), dithiothreitol (DTT), iodoacetic acid (IAA), formic acid and sodium deoxycholate (DOC) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Trypsin
wileyonlinelibrary.com/journal/rcm
was Promega Gold (V5280) (Promega, Madison, WI, USA). The 96-well microtiter plates were purchased from Corning (NY, USA; part no. 3605) and Applied Biosystems (Grand Island, NY, USA; part no. N8010560). Plasma samples Human and rhesus EDTA plasma for method development was obtained from Bioreclamation (Hicksville, NY, USA). Apo(a) calibrators and assay comparison samples were provided by Dr. Marcovina, University of Washington Lipid Research Laboratory (Seattle, WA, USA). Apo(a) concentration and size were determined as reported previously.[8,9,26] In silico selection of proteotypic peptides Proteotypic peptides for each Apo(a) region were selected to maximize assay sensitivity and selectivity. Peptides containing methionine were not considered due to potential oxidation. Peptide sequences were Blast searched using the NCBI tool to ensure that they were unique to the target protein. Targeted peptides included LFLEPTQADIALLK for Apo(a) concentration and GTYSTTVTGR for the kringle type 4 repeating sequence. Preparation of isotope labeled peptide internal standards Isotope labeled peptides were purchased from New England Peptide (Gardner, MA, USA). Peptides were synthesized with universally labeled 13C and 15 N at the C-terminal lysines and arginines. Purity for each peptide was evaluated by LC/UV by the manufacturer and concentration determined by amino acid analysis. Labeled peptides were mixed to achieve a concentration of 13, 440 nM for Apo(a) (LFLEPTQADIALLK) and Apo(a) repeating unit (GTYSTTVTGR) in 50% acetonitrile, 0.1% formic acid and stored in 7 μL single use aliquots at –80 °C. Prior to analysis, single use aliquots were diluted with 193 μL of 50 mM ABC, pH 8.0, to achieve a final concentration of 470.4 nM. Plasma digestion Plasma (20 μL) was diluted with 180 μL 0.5% DOC in a 96-well plate and 20 μL of the resulting mixture transferred to a second plate with 180 μL reduction solution (55 mM ABC, 2.5 nM internal standard peptides, 0.5% DOC and 5.5 mM DTT), resulting in final concentrations 50 mM ABC, 2.3 nM internal standard peptides, 0.5% DOC and 5.0 mM DTT. Samples were reduced for 30 min at 60 °C in an Eppendorf Mastercyler thermocycler, allowed to cool to room temperature, and then alkylated with 2 μL of 1 M fresh IAA prepared in 1 M NaOH (to achieve a final concentration of 10 mM) for 60 min at room temperature in the dark. Samples were digested overnight with (10 μL of 0.075 μg/uL) trypsin gold. The next morning, 10 μL of 20% formic acid were added to stop digestion and precipitate the DOC. Samples were centrifuged for 15 min at 3000 rcf and 120 μL of the supernatant were transferred to a new plate for analysis. For time-course experiments to evaluate the extent of protein digestion, 16 μL of purchased plasma was digested in a total volume of 1.6 mL. After addition of trypsin at the same ratio as above, 200 μL aliquots were removed at various timepoints, immediately precipitated with formic acid, and
Copyright © 2014 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2014, 28, 1101–1106
Characterization of apolipoprotein(a) by UPLC/MS stored on ice. The samples were centrifuged to remove the precipitated DOC, and a 120 μL aliquot of each sample was stored at –20 °C until analysis. UPLC/MS/MS conditions Samples were analyzed on a Waters Acquity UPLC system with a Thermo Vantage triple quadrupole mass spectrometer: 40 μL of sample was injected on-column. The gradient conditions began with initial conditions 98%A (0.1% formic acid in water)/2%B (0.1% formic acid in acetonitrile) ramped to 92%A at 0.5 min, 85%A at 2.0 min, 71%A at 2.2 min and held at 71% until 3.0 min. The solvent conditions were then ramped to 5%A at 3.5 min and held for an additional 0.5 min and immediately returned to 98%A for column equilibration and regeneration for an additional 2 min. The column was an Acquity BEH C18 (1.7 μm, 2.1 × 100 mm, part no. 186002352, Waters, Milford, MA, USA), maintained at 60 °C. The flow rate was 0.7 mL/min. A modification of the column and gradient compared to the method previously reported[17] was required to separate the analyte peptides for Lp(a) from background signal. Values for MRM transitions are included in Fig. 1. Collision energies for GTYSTTVTGR and LFLEPTQADIALLK and their isotope labeled standards were 14 and 19, respectively. Calculation of Apo(a) concentration by UPLC/MS Six calibrator plasma samples were received from NWLRL with assigned Apo(a) concentration (11.2–202.6 nM) and size (14–27 kringle 4) and were aliquoted into 20 μL single-use vials and stored at –80 °C prior to analysis. Calibrator samples were processed as unknown plasma samples. During data analysis, calibrator samples were assigned the values determined by NWLRL and concentrations of unknown samples were calculated using a 1/x weighted linear fit to the ratio of analyte to internal standard area under the curve. Calculation of the average number of kringle 4 on Apo(a) molecules by UPLC/MS Recombinant Apo(a) with a defined number of kringles based on a single gene sequence (Abcam, ab80516, Cambridge, UK) was used to confirm the concentration of the GTYSTTVTG-
[13C15 6 N4]-R in the internal standard peptide solution which was subsequently used to calculate the concentration of GTYSTTVTGR in each plasma sample using single-point calibration. For each sample, the ratio of analyte to internal standard area under the curve was multiplied by [GTYSTTVTG-[13C15 6 N4]-R] to yield [GTYSTTVTGR]. The peptide sequence in K4 type 2 (GTYSTTVTGR) is shared in K4 types 1 and 5 but not by K4 types 3, 6, 7, 8, 9, or 10. Therefore, the total number of kringle 4 is equal to ([GTYSTTVTGR]/[LFLEPTQADIALLK] + 6). It is important to note that this size value represents the average Apo(a) size and most individuals have two expressed Apo(a) alleles of different sizes. Preparation of quality control (QC) plasma Purchased plasma from 20 volunteers was digested as described above. Plasma from four individuals were selected as quality control (QC) samples to achieve four QC levels for Apo(a). 20 μL of plasma from each donor was aliquoted into Sarstedt 0.5 mL screw-capped vials and stored for single use at –80 °C. Intra- and interday reproducibility Intraday reproducibility was determined by measuring the concentration of the four QC samples from six different singleuse vials prepared and analyzed on a single plate. Interday reproducibility was determined by measuring the concentration of the four QC samples from different single-use vials prepared on six different plates and analyzed on six different days. Freeze/thaw stability Freeze/thaw stability was determined by thawing three different QC single-use vials three, two and one additional times and compared to the interday value. QC plasma at each concentration level was allowed to thaw at room temperature for more than 2 h and then stored at –80 °C.
RESULTS Identification of peptides and MRM transitions In silico digestion of the peptide sequence for human Apo(a) using Protein Prospector from UCSF[27] generates multiple potential trypsin cleavage products. Candidate peptides GTYSTTVTGR and LFLEPTQADIALLK were Blast searched against the NCBI database and found to be unique to Apo(a).[28] Targeted product ion spectra for the doubly charged precursor ion for each candidate were acquired from digested plasma. 13C- and 15 N-labeled internal standards for the peptides described above were obtained and infused into the mass spectrometer to optimize instrument parameters and collision energy for each MRM transition. Product ions with m/z values greater than that of their precursor ion were selected for better specificity. Figure 1 shows the extracted ion chromatograms for Apo(a) peptides. Optimization of tryptic digestion conditions
Rapid Commun. Mass Spectrom. 2014, 28, 1101–1106
The utility of sodium deoxycholate (DOC) has been demonstrated for multiple plasma proteins.[29] Previously, our laboratory identified 0.5% DOC to be optimal for solubilizing
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm
1103
Figure 1. Extracted ion chromatograms for Apo(a) peptides GTYSTTVTGR and LFLEPTQADIALLK from a typical plasma digest.
M. E. Lassman et al. and enhancing digestion efficiency for apolipoproteins.[17] Here, we have maintained the same concentration of DOC, but increased the dilution of plasma to 1:100 and reduced the concentration of trypsin used four-fold, from 3 μg to 0.75 μg per sample. The rationale for this modification is to allow digestion of an entire 96-well microtiter plate using a single 100 μg aliquot of trypsin. We confirmed that production of tryptic peptides used for quantitation of Apo(a) is complete after approximately 8 h digestion with no reduction in product ion observed up to 21 h under these conditions. Overnight digestion kinetics (data not shown) for production of peptides LFLEPTQADIALLK and GTYSTTVTGR demonstrate that equilibrium has been achieved for these peptides in less than 8 h. Assay performance This limit of quantitation (LOQ) of the assay for Apo(a) was determined by preparing a serial dilution of the highest concentration calibrator sample into purchased human plasma for which LFLEPTQADIALLK could not be detected. Using this methodology, the LOQ for Apo(a) was established to be 6 nM with the coefficient of variation (CV) and relative error less than 20%. The LOQ for the repeating sequence was not determined as the GTYSTTVTGR concentration in the lowest concentration purchased plasma was 39 nM and the CV and relative error is less than 20% at this concentration. All other plasma samples were measured to have a GTYSTTVTGR concentration above 39 nM. Results for intra- and interassay reproducibility are displayed for each QC level in Table 1. Values are considered
Table 1. Assay performance Intra (n = 6)
Inter (n = 6)
Mean, nMa CV, % GTYSTTVTGR QC1 QC2 QC3 QC4 LFLEPTQADIALLK QC1 QC2 QC3 QC4
Mean, nMa
CV, %
1331 661 39 298
5 7 4 5
1219 667 38 312
4 4 14 5
149 76 LOQ 24
3 9
142 72 LOQ 27
6 4
10
10
acceptable for all protein QC levels except the lowest concentrations for Apo(a), where the concentrations are considered to be below the limit of quantitation (LOQ). Spike recovery experiments were not performed because purchased protein standards do not mimic the endogenous lipid-rich protein particles in plasma. Alternatively, mixing experiments were implemented with QC samples diluted with each other 1:1 to demonstrate accuracy and investigate bias. Values obtained from these mixing experiments are included in Table 2. No bias was observed during these experiments. Sample stability was assessed through multiple freeze/thaw cycles. QC sample sets were thawed and frozen for 1, 2 and 3 cycles. Comparison of the measured values with the interday values are calculated as percent differences and demonstrated in Table 3. For GTYSTTVTGR and LFLEPTQADIALLK, the measured concentration difference is less than 20% for all concentration levels above LOQ. We therefore conclude that this assay is insensitive up to 3 freeze/thaw cycles. Comparison of UPLC/MS and immunochemical/electrophoresis assays Eighty plasma samples with pre-assigned values for Apo(a) concentration and predominant size were received from NWLRL and analyzed for Apo(a) by MS. The concentrations for Apo(a) ranged from 8.7 nM to 276.1 nM as measured by the monoclonal antibody-based ELISA method. The correlation of the two methods (UPLC/MS and ELISA) is demonstrated in Fig. 2. Slope and r2 were determined to be 0.96 and 0.98 using Graphpad Prism 6 software, demonstrating excellent correlation and agreement of absolute values between the two assays for Apo(a) concentration measurement. The LC/MS and electrophoresis methods measure Apo(a) size differently. While the electrophoresis method is capable of separating and qualitatively identifying Apo(a) polymorphs based on electrophoretic mobility, the UPLC/MS method quantitates the concentration of kringles that contain the peptide sequence GTYSTTVTGR and uses this concentration measurement to calculate the average number of kringles per Apo(a) protein quantitated using the LFLEPTQADIALLK peptide. Because most individuals have two Apo(a) alleles that result in two different sized Apo(a) polymorphs circulating in plasma, it is important to understand the fundamental differences between the two assays. Figure 3 describes the size measurements for both the UPLC/MS method and the electrophoresis measurement. Not surprisingly, the two size measurements do not match as well as they do for Apo(a) concentration. For many samples, the average size (UPLC/MS) is smaller than the predominant size (electrophoresis). The differences are greater at lower Apo(a)
Table 2. Demonstration of recovery by QC mixing QC1 + QC3 Apolipoprotein(a)
1104
GTYSTTVTGR LFLEPTQADIALLK
QC1 + QC2
QC2 + QC4
Cal6 + QC3
Calc, nm
Meas, nm
Calc, nm
Meas, nm
Calc, nm
Meas, nm
Calc, nm
Meas, nm
639 72
627 75
930 104
947 111
439 49
495 49
715 98
686 102
wileyonlinelibrary.com/journal/rcm
Copyright © 2014 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2014, 28, 1101–1106
Characterization of apolipoprotein(a) by UPLC/MS
DISCUSSION
Table 3. Freeze-thaw stability for Apo(a) concentration # Cycles Concentration
1
1219 677 38 312
–8 1 –5 –8
142 72 LOQ 27
–10 –1
10 –9
1 –4
11
0
7
GTYSTTVTGR
LFLEPTQADIALLK
2
3
–16 –15 –1 –9 –8 –16 –4 1
Figure 2. Correlation of Apo(a) concentration values assigned using the NWLRL ELISA method and the LC/MS assay.
Figure 3. Correlation of Apo(a) concentration values assigned using the NWLRL ELISA method and size determined using the electrophoretic method and the LC/MS assay.
Here we have described a modification of our previously published UPLC/MS-based method[17] for the inclusion of Apo(a) as well as characterization of its size through quantitation of the number of kringles. In this single assay, multiple apolipoproteins can be measured in a multiplexed fashion with a minimal volume of plasma (2 μL). As described previously, the limit of quantitation for apolipoproteins A1 and B are much lower than typical plasma concentrations for clinical samples; however, concentrations for Apo(a) span a lower and much broader dynamic range. For this UPLC/MS assay, the limit of quantitation for Apo(a) was determined to be 6 nM, which enables the quantitation of Apo(a) for most subjects. This UPLC/MS assay has passed all typical requirements for use as a ’fit for purpose’ assay to be used for the analysis of clinical research samples, including assay precision, sample-mixing and freeze/thaw stability. Like most UPLC/MS assays, this assay has lower throughput and higher cost relative to immunological methods due to the cost of UPLC/MS instrumentation and 6 min cycle time per sample. However, as a multiplexed assay, the ability to measure multiple proteins from a single sample and the reduced requirement for multiple analyses may reduce the overall cost of sample analysis. Also, the UPLC/MS assay does not require specific antibodies and can be performed using most commercially available triplequadrupole mass spectrometers, thus allowing wide access to researchers interested in the function and biology of Apo(a) which is especially important given the unknown function and correlation of Apo(a) size and disease. Apo(a) concentrations determined using this assay showed excellent correlation and agreement of absolute values with a highly validated and accurate ELISA method, giving confidence to both assays. However, the determinations of size are different. While the established high-resolution SDS agarose gel electrophoretic method is capable of independent measurements of determining the Apo(a) size for the two expressed alleles, the UPLC/MS method measures the average Apo(a) size using a quantitative measurement. This difference in approach may explain the observed differences in size determined for the two methods. It is unclear what impact this measurement of average Apo(a) size can have for clinical research and for elucidating the function of Apo(a) in disease, but this assay may provide an additional tool for solving the relationship between Apo(a) size and function. A potential future application may be to combine measurements of Apo(a) concentration and isotope enrichment as part of tracer-based kinetics studies.[7–9] For this version of the assay, there is not sufficient sensitivity to measure small percentages of isotope enrichment for many plasma samples, but we have demonstrated the ability to measure isotope enrichment using the LFLEPTQADIALLK peptide in fractioned plasma samples (HDL and LDL) using a similar digestion strategy and microflow LC/MS.[30]
Acknowledgements
Rapid Commun. Mass Spectrom. 2014, 28, 1101–1106
The authors would like to thank Dr. Gissette Reyes-Soffer and Dr. Henry Ginsberg of Columbia University for helpful discussion regarding Lp(a) biology and function.
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm
1105
concentrations. Furthermore, the inverse relationship between average Apo(a) size and concentration is less pronounced, as indicated by a more shallow slope.
M. E. Lassman et al.
REFERENCES
1106
[1] S. Tsimikas, L. D. Tsironis, A. D. Tselepis. New insights into the role of lipoprotein(a)-associated lipoproteinassociated phospholipase A2 in atherosclerosis and cardiovascular disease. Arterioscler. Thromb. Vasc. Biol. 2007, 27, 2094. [2] M. Ignatescu, K. Kostner, G. Zorn, M. Kneussl, G. Maurer, I. M. Lang, K. Huber. Plasma Lp(a) levels are increased in patients with chronic thromboembolic pulmonary hypertension. Thromb. Haemost. 1998, 80, 231. [3] B. G. Nordestgaard, M. J. Chapman, K. Ray, J. Boren, F. Andreotti, G. F. Watts, H. Ginsberg, P. Amarenco, A. Catapano, O. S. Descamps, E. Fisher, P. T. Kovanen, J. A. Kuivenhoven, P. Lesnik, L. Masana, Z. Reiner, M. R. Taskinen, L. Tokgozoglu, A. Tybjaerg-Hansen. Lipoprotein (a) as a cardiovascular risk factor: current status. Eur. Heart J. 2010, 31, 2844. [4] G. Utermann, W. Weber. Protein composition of Lp(a) lipoprotein from human plasma. FEBS Lett. 1983, 154, 357. [5] G. Utermann, H. J. Menzel, H. G. Kraft, H. C. Duba, H. G. Kemmler, C. Seltz. Lp(a) glycoprotein phenotypes. Inheritance and relation to Lp(a)-lipoprotein concentrations in plasma. J. Clin. Invest. 1987, 80, 458. [6] E. Boerwinkle, C. C. Leffert, J. Lin, C. Lackner, G. Chiesa, H. H. Hobbs. Apolipoprotein(a) gene accounts for greater than 90% of the variation in plasma lipoprotein(a) concentrations. J. Clin. Invest. 1992, 90, 52. [7] D. J. Rader, W. Cain, K. Ikewaki, G. Talley, L. Zech, D. Usher, H. B. Brewer. The inverse association of plasma Lp(a) concentrations with apolipoprotein(a) isoform size is not due to differences in Lp(a) catabolism but to differences in production rate. J. Clin. Invest. 1994, 93, 2758. [8] M. E. Frischmann, K. Ikewaki, E. Trenkwalder, C. Lamina, B. Dieplinger, S. Muhidien, H. Schweer, J. R. Schaefer, P. Konig, F. Kronenberg, H. Dieplinger. In vivo stableisotope kinetic study suggests intracellular assembly of lipoprotein(a). Atherosclerosis 2012, 225, 322. [9] J. L. Jenner, L. J. Seman, J. S. Millar, S. Lamon-Fava, F. K. Welty, G. G. Dolnikowski, S. M. Marcovina, A. H. Lichtenstein, P. H. R. Barrett, C. DeLuca, E. J. Schaefer. The metabolism of apolipoproteins (a) and B-100 within plasma lipoprotein (a) in human beings. Metabolism 2005, 54, 361. [10] T. Demant, D. Seeberg, A. Bedynek, D. Seidel. The metabolism of lipoprotein(a) and other apolipoprotein B-containing lipoproteins: a kinetic study in humans. Atherosclerosis 2001, 157, 325. [11] S. M. Marcovina, H. H. Hobbs, J. J. Albers. Relation between number of apolipoprotein(a) kringle 4 repeats and mobility of isoforms in agarose gel: basis for a standardized isoform nomenclature. Clin. Chem. 1996, 42, 436. [12] S. M. Marcovina, J. J. Albers, B. Gabel, M. L. Koschinsky, V. P. Gaur. Effect of the number of apolipoprotein(a) kringle 4 domains on immunochemical measurements of lipoprotein(a). Clin. Chem. 1995, 41, 246. [13] S. Lamon-Fava, S. M. Marcovina, J. J. Albers, H. Kennedy, C. DeLuca, C. C. White, L. A. Cupples, J. R. McNamara, L. J. Seman, V. Bongard, E. J. Schaefer. Lipoprotein(a) levels, apo(a) isoform size, and coronary heart disease risk in the Framingham Offspring Study. J. Lipid Res. 2011, 52, 1181. [14] A. N. Hoofnagle, M. H. Wener. The fundamental flaws of immunoassays and potential solutions using tandem mass spectrometry. J. Immunol. Methods 2009, 347, 3. [15] R. G. Kay, B. Gregory, P. G. Grace, S. Pleasance. The application of ultra-performance liquid chromatography/ tandem mass spectrometry to the detection and quantitation of apolipoproteins in human serum. Rapid Commun. Mass Spectrom. 2007, 21, 2585.
wileyonlinelibrary.com/journal/rcm
[16] S. A. Agger, L. C. Marney, N. A. Hoofnagle. Simultaneous quantification of apolipoprotein A-I and apolipoprotein B by liquid-chromatography–multiple-reaction–monitoring mass spectrometry. Clin. Chem. 2010, 56, 1804. [17] M. E. Lassman, T. M. McLaughlin, E. P. Somers, A. C. Stefanni, Z. Chen, B. A. Murphy, K. K. Bierilo, A. M. Flattery, K. K. Wong, J. M. Castro-Perez, B. K. Hubbard, T. P. Roddy. A rapid method for cross species quantitation of apolipoproteins A1, B48 and B100 in plasma by ultra-performance liquid chromatography/tandem mass spectrometry. Rapid Commun. Mass Spectrom. 2012, 26, 101. [18] R. Simon, M. Girod, C. Fonbonne, A. Salvador, Y. Clément, P. Lantéri, P. Amouyel, J. C. Lambert, J. Lemoine. Total ApoE and ApoE4 isoform assays in an Alzheimer’s Disease case-control study by targeted mass spectrometry (n = 669): A pilot assay for methionine-containing proteotypic peptides. Mol. Cell. Proteomics 2012, 11, 1389. [19] L. Anderson, C. Hunter. Quantitative mass spectrometric multiple reaction monitoring assays for major plasma proteins. Mol. Cell. Proteomics 2005, 5, 573. [20] T. A. Addona, S. E. Abbatiello, B. Schilling, S. J. Skates, D. R. Mani, D. M. Bunk, C. H. Spiegelman, L. J. Zimmerman, A. J. L. Ham, H. Keshishian, S. C. Hall, S. Allen, R. K. Blackman, C. H. Borchers, C. Buch, H. L. Cardasis, M. P. Cusack, N. G. Dodder, B. W. Gibson, J. M. Held, T. Hiltke, A. Jackson, E. B. Johnson, C. R. Kinsinger, J. Li, M. Mesri, T. A. Neubert, R. K. Niles, T. C. Pulsipher, D. Ransohoff, H. Rodriguez, P. A. Rudnick, D. Smith, D. L. Tabb, T. J. Tegeler, A. M. Variyath, L. J. Vega-Montoto, A. Wahlander, S. Waldemarson, J. R. Whiteaker, L. Zhao, N. L. Anderson, S. J. Fisher, D. C. Liebler, A. G. Paulovich, F. E. Regnier, P. Tempst, S. A. Carr. Multi-site assessment of the precision and reproducibility of multiple reaction monitoring-based measurements of proteins in plasma. Nat. Biotechnol. 2009, 27, 633. [21] M. A. Kuzyk, D. Smith, J. Yang, T. J. Cross, A. M. Jackson, D. B. Hardie, N. L. Anderson, C. H. Borchers. MRM-based, multiplexed, absolute quantitation of 45 proteins in human plasma. Mol. Cell. Proteomics 2009, 8, 1860. [22] S. E. Ong, M. Mann. Mass spectrometry-based proteomics turns quantitative. Nat. Chem. Biol. 2005, 1, 252. [23] V. Brun, C. Masselon, J. Garin, A. Dupuis. Isotope dilution strategies for absolute quantitative proteomics. J. Proteomics 2009, 72, 740. [24] E. Ezan, M. Dubois, F. Becher. Bioanalysis of recombinant proteins and antibodies by mass spectrometry. Analyst 2009, 134, 825. [25] E. E. Chambers, C. Legido-Quigley, N. Smith, K. J. Fountain. Development of a fast method for direct analysis of intact synthetic insulins in human plasma: the large peptide challenge. Bioanalysis 2013, 5, 65. [26] S. M. Marcovina, Z. H. Zhang, V. P. Gaur, J. J. Albers. Identification of 34 apolipoprotein(a) isoforms: Differential expression of apolipoprotein(a) alleles between American blacks and whites. Biochem. Biophys. Res. Commun. 1993, 191, 1192. [27] http://prospector.ucsf.edu/prospector/mshome.htm. [28] http://blast.ncbi.nlm.nih.gov/Blast.cgi. [29] J. L. Proc, M. A. Kuzyk, D. B. Hardie, J. Yang, D. S. Smith, A. M. Jackson, C. E. Parker, C. H. Borchers. A quantitative study of the effects of chaotropic agents, surfactants, and solvents on the digestion efficiency of human plasma proteins by trypsin. J. Proteome Res. 2010, 9, 5422. [30] H. Zhou, J. Castro-Perez, M. E. Lassman, T. Thomas, W. Li, T. M. McLaughlin, D. Xie, P. Jumes, J. A. Wagner, D. E. Gutstein, B. K. Hubbard, D. J. Rader, J. S. Millar, H. N. Ginsberg, G. Reyes-Soffer, M. Cleary, S. F. Previs, T. P. Roddy. Measurement of apo(a) kinetics in human subjects using a microfluidic device with tandem mass spectrometry. Rapid Commun. Mass Spectrom. 2013, 27, 1294.
Copyright © 2014 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2014, 28, 1101–1106
Revised: 27 February 2014
Accepted: 27 February 2014
Published online in Wiley Online Library
Rapid Commun. Mass Spectrom. 2014, 28, 1101–1106 (wileyonlinelibrary.com) DOI: 10.1002/rcm.6883
Simultaneous quantitation and size characterization of apolipoprotein(a) by ultra-performance liquid chromatography/mass spectrometry Michael E. Lassman1*, Theresa M. McLaughlin1, Haihong Zhou1, Yi Pan1, Santica M. Marcovina2, Omar Laterza1 and Thomas P. Roddy1 1
Merck Research Laboratories, 126 E. Lincoln Ave., Rahway, NJ 07065, USA Northwest Lipid Metabolism and Diabetes Research Laboratories, University of Washington, 401 Queen Anne Ave, North Seattle, WA 98109, USA
2
RATIONALE: Apolipoprotein(a) is a polymorphic glycoprotein covalently bound to apoB100 in Lp(a) particles and has been described to be both atherogenic and prothrombotic, although its exact mechanism of action is not well defined. Apolipoprotein(a) is routinely measured by immunoassays. Unfortunately, the accuracy of the measurement can be affected by the apolipoprotein(a) size (number of kringles) polymorphism in Lp(a) particles. Here we describe an ultra-performance liquid chromatography/mass spectrometry (UPLC/MS) assay that is capable of measuring apolipoprotein(a) concentrations while simultaneously determining the number of kringles present per protein. METHODS: Plasma samples were diluted and proteins de-lipidated with deoxycholate prior to tryptic digestion. Distinct tryptic peptides from different regions of apolipoprotein(a) were measured to determine both concentration and the number of kringles present per protein. Separation and quantitation of tryptic peptides is carried out at 700 μL/min using a 1.7 μm C18 column (2.1 × 100 mm) coupled to a Thermo Vantage triple quadrupole (QQQ) mass spectrometer with a heated electrospray ionization (HESI) source. RESULTS: This method was compared to established methods for measuring concentration (monoclonal antibody based ELISA) and size (gel-electrophoresis) using 80 plasma samples proved by NWLRL. The slope and r2 value for the correlation of concentrations were determined to be 0.96 and 0.98, demonstrating excellent agreement of absolute values between the UPLC/MS and ELISA methods. As measured by UPLC/MS, the average kringle number or size is smaller than determined by the electrophoretic method. CONCLUSIONS: A single UPLC/MS method was developed capable of measuring apolipoprotein(a) concentration and size (by measuring the number of kringles per protein). This assay passes criteria required for ’fit for purpose’ assays including sensitivity, intra and interday reproducibility and freeze/thaw stability. While the agreement between UPLC/MS and ELISA is excellent for concentration and may provide researchers with additional tools for studying apolipoprotein(a), the dissimilarities between UPLC/MS and the electrophoretic method may also be exploited for understanding apolipoprotein(a) structure and function. Copyright © 2014 John Wiley & Sons, Ltd.
Lp(a) is a lipoprotein that has been reported as being both proatherogenic and prothrombotic.[1–3] Structurally, Lp(a) is similar to another proatherogenic particle, LDL, with the distinction that Lp(a) contains the polypeptide protein apolipoprotein(a) (Apo(a)), a polymorphic glycoprotein covalently bound to apoB100 by a disulfide bond.[4] The concentration of circulating Apo(a) is genetically determined[5,6] and has been reported to be inversely correlated with Apo(a) size. This inverse association between plasma Apo(a) concentration and size has been demonstrated to be primarily due to differences in Apo(a) production rates.[7] Multiple research studies have included measurement of Apo(a) kinetics as part of an effort to understand its role and function.[8–10] Still, despite many large studies, the exact function of Apo(a)
Rapid Commun. Mass Spectrom. 2014, 28, 1101–1106
Copyright © 2014 John Wiley & Sons, Ltd.
1101
* Correspondence to: M. E. Lassman, Molecular Biomarkers and Diagnostics Laboratory, Merck Sharp & Dohme Corp, Rahway, NJ 07065, USA. E-mail: [email protected]
particles is not completely understood, nor is the relationship between disease and concentration/size. Apo(a) size is defined primarily by the number of kringle 4 type 2 (KIV2) repeats in Apo(a)[4] which can be experimentally determined by high-resolution sodium dodecyl sulfateagarose gel electrophoresis followed by immunoblotting,[11] a technique that is performed by a limited number of laboratories. Apo(a) concentration can be accurately determined by a double monoclonal antibody-based enzyme-linked immunoassay that is by design, insensitive to Apo(a) size heterogeneity, NWLRL ELISA.[12] Alternatively, a limited number of commercially available assays exist including an immunoturbidimetric assay that uses an antibody that does not cross-react with plasminogen or apolipoproteinB as well as a lectin-based assay that captures Lp(a) and measures the cholesterol content in the particle.[13] Lamon-Fava et al. compared the commercially available assays to the Northwest Lipid Research Laboratories (NWLRL) ELISA and found that they were well correlated. However, the immunoturbidimetric assay measures concentration in mg/dL rather
M. E. Lassman et al. than nM, which is problematic due to the heterogeneity in Apo(a) molecular weight. As a result, the immunoturbidimetric assay was demonstrated to have a strong Apo(a) isoform size-dependent bias. The lectin-based method was determined to over-estimate the amount of Lp(a) cholesterol in samples with low Apo(a) values.[13] In order to advance the field and understanding of the role of Apo(a) in disease, readily available assays must be available that are capable of supporting large clinical studies, accurately measuring Apo(a) concentration in units of nM and characterizing Apo(a) size. Multiple laboratories have proposed the use of mass spectrometry (MS)-based approaches as an alternative to immunoassays,[14] including measurements of apolipoproteins.[15–18] An MS-based approach has advantages in that it does not require the generation or use of specific antibodies and instead relies on the specificity of liquid chromatography (LC) and multiple reaction monitoring (MRM) in quantitation of proteotypic peptides. Furthermore, these assays are readily multiplexed for simultaneous measurements of multiple analytes. For Apo(a), an MS-based approach can be used to characterize size by measuring the concentration of kringle 4 type 2 repeats. An MS-based approach also has the advantage that it can be readily modified and applied to Apo(a) research in non-human primates, which also have Apo(a), although the sequence is not completely conserved. The application of MS-based quantitation of proteins by MRM has been described repeatedly for plasma proteins, with single and multiplexed assays reported in the literature.[19–23] These assays have been used for the analysis of endogenous as well as therapeutic proteins,[24,25] but are essentially similar. In general, MRM assays use enzymatic digestion of the targeted protein within a complex protein matrix to generate proteotypic peptides that are measured by LC/MS. The use of stable isotope labeled standards corrects for variation in sample handling and peptide degradation as well as analytical variability and allows for quantitation of the peptide as a proxy for the intact protein. The throughput and instrument availability of LC/MS-based methods are limited compared to many immunoassays, but ultra-performance liquid chromatography (UPLC) methods have made throughput more competitive with immunoassays and the instrumentation required to perform these assays is becoming more commonplace in academic and clinical laboratories. Furthermore, compared to multiple immunoassays, a single multiplexed UPLC/MS-based method becomes even more competitive in terms of cost and throughput. Here, we demonstrate a single UPLC/MS assay based on an apolipoprotein platform previously reported by our group[17] capable of measuring apolipoproteins A1, B48 and B100 in multiple species which is now applied for the first time to measurements of Apo(a) concentration and size. The enzymatic digestion, sample preparation and UPLC/MS analysis of plasma samples are performed in 96-well microtiter plates, allowing for a daily throughput of over 200 samples per day.
EXPERIMENTAL Chemicals
1102
Ammonium bicarbonate (ABC), dithiothreitol (DTT), iodoacetic acid (IAA), formic acid and sodium deoxycholate (DOC) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Trypsin
wileyonlinelibrary.com/journal/rcm
was Promega Gold (V5280) (Promega, Madison, WI, USA). The 96-well microtiter plates were purchased from Corning (NY, USA; part no. 3605) and Applied Biosystems (Grand Island, NY, USA; part no. N8010560). Plasma samples Human and rhesus EDTA plasma for method development was obtained from Bioreclamation (Hicksville, NY, USA). Apo(a) calibrators and assay comparison samples were provided by Dr. Marcovina, University of Washington Lipid Research Laboratory (Seattle, WA, USA). Apo(a) concentration and size were determined as reported previously.[8,9,26] In silico selection of proteotypic peptides Proteotypic peptides for each Apo(a) region were selected to maximize assay sensitivity and selectivity. Peptides containing methionine were not considered due to potential oxidation. Peptide sequences were Blast searched using the NCBI tool to ensure that they were unique to the target protein. Targeted peptides included LFLEPTQADIALLK for Apo(a) concentration and GTYSTTVTGR for the kringle type 4 repeating sequence. Preparation of isotope labeled peptide internal standards Isotope labeled peptides were purchased from New England Peptide (Gardner, MA, USA). Peptides were synthesized with universally labeled 13C and 15 N at the C-terminal lysines and arginines. Purity for each peptide was evaluated by LC/UV by the manufacturer and concentration determined by amino acid analysis. Labeled peptides were mixed to achieve a concentration of 13, 440 nM for Apo(a) (LFLEPTQADIALLK) and Apo(a) repeating unit (GTYSTTVTGR) in 50% acetonitrile, 0.1% formic acid and stored in 7 μL single use aliquots at –80 °C. Prior to analysis, single use aliquots were diluted with 193 μL of 50 mM ABC, pH 8.0, to achieve a final concentration of 470.4 nM. Plasma digestion Plasma (20 μL) was diluted with 180 μL 0.5% DOC in a 96-well plate and 20 μL of the resulting mixture transferred to a second plate with 180 μL reduction solution (55 mM ABC, 2.5 nM internal standard peptides, 0.5% DOC and 5.5 mM DTT), resulting in final concentrations 50 mM ABC, 2.3 nM internal standard peptides, 0.5% DOC and 5.0 mM DTT. Samples were reduced for 30 min at 60 °C in an Eppendorf Mastercyler thermocycler, allowed to cool to room temperature, and then alkylated with 2 μL of 1 M fresh IAA prepared in 1 M NaOH (to achieve a final concentration of 10 mM) for 60 min at room temperature in the dark. Samples were digested overnight with (10 μL of 0.075 μg/uL) trypsin gold. The next morning, 10 μL of 20% formic acid were added to stop digestion and precipitate the DOC. Samples were centrifuged for 15 min at 3000 rcf and 120 μL of the supernatant were transferred to a new plate for analysis. For time-course experiments to evaluate the extent of protein digestion, 16 μL of purchased plasma was digested in a total volume of 1.6 mL. After addition of trypsin at the same ratio as above, 200 μL aliquots were removed at various timepoints, immediately precipitated with formic acid, and
Copyright © 2014 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2014, 28, 1101–1106
Characterization of apolipoprotein(a) by UPLC/MS stored on ice. The samples were centrifuged to remove the precipitated DOC, and a 120 μL aliquot of each sample was stored at –20 °C until analysis. UPLC/MS/MS conditions Samples were analyzed on a Waters Acquity UPLC system with a Thermo Vantage triple quadrupole mass spectrometer: 40 μL of sample was injected on-column. The gradient conditions began with initial conditions 98%A (0.1% formic acid in water)/2%B (0.1% formic acid in acetonitrile) ramped to 92%A at 0.5 min, 85%A at 2.0 min, 71%A at 2.2 min and held at 71% until 3.0 min. The solvent conditions were then ramped to 5%A at 3.5 min and held for an additional 0.5 min and immediately returned to 98%A for column equilibration and regeneration for an additional 2 min. The column was an Acquity BEH C18 (1.7 μm, 2.1 × 100 mm, part no. 186002352, Waters, Milford, MA, USA), maintained at 60 °C. The flow rate was 0.7 mL/min. A modification of the column and gradient compared to the method previously reported[17] was required to separate the analyte peptides for Lp(a) from background signal. Values for MRM transitions are included in Fig. 1. Collision energies for GTYSTTVTGR and LFLEPTQADIALLK and their isotope labeled standards were 14 and 19, respectively. Calculation of Apo(a) concentration by UPLC/MS Six calibrator plasma samples were received from NWLRL with assigned Apo(a) concentration (11.2–202.6 nM) and size (14–27 kringle 4) and were aliquoted into 20 μL single-use vials and stored at –80 °C prior to analysis. Calibrator samples were processed as unknown plasma samples. During data analysis, calibrator samples were assigned the values determined by NWLRL and concentrations of unknown samples were calculated using a 1/x weighted linear fit to the ratio of analyte to internal standard area under the curve. Calculation of the average number of kringle 4 on Apo(a) molecules by UPLC/MS Recombinant Apo(a) with a defined number of kringles based on a single gene sequence (Abcam, ab80516, Cambridge, UK) was used to confirm the concentration of the GTYSTTVTG-
[13C15 6 N4]-R in the internal standard peptide solution which was subsequently used to calculate the concentration of GTYSTTVTGR in each plasma sample using single-point calibration. For each sample, the ratio of analyte to internal standard area under the curve was multiplied by [GTYSTTVTG-[13C15 6 N4]-R] to yield [GTYSTTVTGR]. The peptide sequence in K4 type 2 (GTYSTTVTGR) is shared in K4 types 1 and 5 but not by K4 types 3, 6, 7, 8, 9, or 10. Therefore, the total number of kringle 4 is equal to ([GTYSTTVTGR]/[LFLEPTQADIALLK] + 6). It is important to note that this size value represents the average Apo(a) size and most individuals have two expressed Apo(a) alleles of different sizes. Preparation of quality control (QC) plasma Purchased plasma from 20 volunteers was digested as described above. Plasma from four individuals were selected as quality control (QC) samples to achieve four QC levels for Apo(a). 20 μL of plasma from each donor was aliquoted into Sarstedt 0.5 mL screw-capped vials and stored for single use at –80 °C. Intra- and interday reproducibility Intraday reproducibility was determined by measuring the concentration of the four QC samples from six different singleuse vials prepared and analyzed on a single plate. Interday reproducibility was determined by measuring the concentration of the four QC samples from different single-use vials prepared on six different plates and analyzed on six different days. Freeze/thaw stability Freeze/thaw stability was determined by thawing three different QC single-use vials three, two and one additional times and compared to the interday value. QC plasma at each concentration level was allowed to thaw at room temperature for more than 2 h and then stored at –80 °C.
RESULTS Identification of peptides and MRM transitions In silico digestion of the peptide sequence for human Apo(a) using Protein Prospector from UCSF[27] generates multiple potential trypsin cleavage products. Candidate peptides GTYSTTVTGR and LFLEPTQADIALLK were Blast searched against the NCBI database and found to be unique to Apo(a).[28] Targeted product ion spectra for the doubly charged precursor ion for each candidate were acquired from digested plasma. 13C- and 15 N-labeled internal standards for the peptides described above were obtained and infused into the mass spectrometer to optimize instrument parameters and collision energy for each MRM transition. Product ions with m/z values greater than that of their precursor ion were selected for better specificity. Figure 1 shows the extracted ion chromatograms for Apo(a) peptides. Optimization of tryptic digestion conditions
Rapid Commun. Mass Spectrom. 2014, 28, 1101–1106
The utility of sodium deoxycholate (DOC) has been demonstrated for multiple plasma proteins.[29] Previously, our laboratory identified 0.5% DOC to be optimal for solubilizing
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm
1103
Figure 1. Extracted ion chromatograms for Apo(a) peptides GTYSTTVTGR and LFLEPTQADIALLK from a typical plasma digest.
M. E. Lassman et al. and enhancing digestion efficiency for apolipoproteins.[17] Here, we have maintained the same concentration of DOC, but increased the dilution of plasma to 1:100 and reduced the concentration of trypsin used four-fold, from 3 μg to 0.75 μg per sample. The rationale for this modification is to allow digestion of an entire 96-well microtiter plate using a single 100 μg aliquot of trypsin. We confirmed that production of tryptic peptides used for quantitation of Apo(a) is complete after approximately 8 h digestion with no reduction in product ion observed up to 21 h under these conditions. Overnight digestion kinetics (data not shown) for production of peptides LFLEPTQADIALLK and GTYSTTVTGR demonstrate that equilibrium has been achieved for these peptides in less than 8 h. Assay performance This limit of quantitation (LOQ) of the assay for Apo(a) was determined by preparing a serial dilution of the highest concentration calibrator sample into purchased human plasma for which LFLEPTQADIALLK could not be detected. Using this methodology, the LOQ for Apo(a) was established to be 6 nM with the coefficient of variation (CV) and relative error less than 20%. The LOQ for the repeating sequence was not determined as the GTYSTTVTGR concentration in the lowest concentration purchased plasma was 39 nM and the CV and relative error is less than 20% at this concentration. All other plasma samples were measured to have a GTYSTTVTGR concentration above 39 nM. Results for intra- and interassay reproducibility are displayed for each QC level in Table 1. Values are considered
Table 1. Assay performance Intra (n = 6)
Inter (n = 6)
Mean, nMa CV, % GTYSTTVTGR QC1 QC2 QC3 QC4 LFLEPTQADIALLK QC1 QC2 QC3 QC4
Mean, nMa
CV, %
1331 661 39 298
5 7 4 5
1219 667 38 312
4 4 14 5
149 76 LOQ 24
3 9
142 72 LOQ 27
6 4
10
10
acceptable for all protein QC levels except the lowest concentrations for Apo(a), where the concentrations are considered to be below the limit of quantitation (LOQ). Spike recovery experiments were not performed because purchased protein standards do not mimic the endogenous lipid-rich protein particles in plasma. Alternatively, mixing experiments were implemented with QC samples diluted with each other 1:1 to demonstrate accuracy and investigate bias. Values obtained from these mixing experiments are included in Table 2. No bias was observed during these experiments. Sample stability was assessed through multiple freeze/thaw cycles. QC sample sets were thawed and frozen for 1, 2 and 3 cycles. Comparison of the measured values with the interday values are calculated as percent differences and demonstrated in Table 3. For GTYSTTVTGR and LFLEPTQADIALLK, the measured concentration difference is less than 20% for all concentration levels above LOQ. We therefore conclude that this assay is insensitive up to 3 freeze/thaw cycles. Comparison of UPLC/MS and immunochemical/electrophoresis assays Eighty plasma samples with pre-assigned values for Apo(a) concentration and predominant size were received from NWLRL and analyzed for Apo(a) by MS. The concentrations for Apo(a) ranged from 8.7 nM to 276.1 nM as measured by the monoclonal antibody-based ELISA method. The correlation of the two methods (UPLC/MS and ELISA) is demonstrated in Fig. 2. Slope and r2 were determined to be 0.96 and 0.98 using Graphpad Prism 6 software, demonstrating excellent correlation and agreement of absolute values between the two assays for Apo(a) concentration measurement. The LC/MS and electrophoresis methods measure Apo(a) size differently. While the electrophoresis method is capable of separating and qualitatively identifying Apo(a) polymorphs based on electrophoretic mobility, the UPLC/MS method quantitates the concentration of kringles that contain the peptide sequence GTYSTTVTGR and uses this concentration measurement to calculate the average number of kringles per Apo(a) protein quantitated using the LFLEPTQADIALLK peptide. Because most individuals have two Apo(a) alleles that result in two different sized Apo(a) polymorphs circulating in plasma, it is important to understand the fundamental differences between the two assays. Figure 3 describes the size measurements for both the UPLC/MS method and the electrophoresis measurement. Not surprisingly, the two size measurements do not match as well as they do for Apo(a) concentration. For many samples, the average size (UPLC/MS) is smaller than the predominant size (electrophoresis). The differences are greater at lower Apo(a)
Table 2. Demonstration of recovery by QC mixing QC1 + QC3 Apolipoprotein(a)
1104
GTYSTTVTGR LFLEPTQADIALLK
QC1 + QC2
QC2 + QC4
Cal6 + QC3
Calc, nm
Meas, nm
Calc, nm
Meas, nm
Calc, nm
Meas, nm
Calc, nm
Meas, nm
639 72
627 75
930 104
947 111
439 49
495 49
715 98
686 102
wileyonlinelibrary.com/journal/rcm
Copyright © 2014 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2014, 28, 1101–1106
Characterization of apolipoprotein(a) by UPLC/MS
DISCUSSION
Table 3. Freeze-thaw stability for Apo(a) concentration # Cycles Concentration
1
1219 677 38 312
–8 1 –5 –8
142 72 LOQ 27
–10 –1
10 –9
1 –4
11
0
7
GTYSTTVTGR
LFLEPTQADIALLK
2
3
–16 –15 –1 –9 –8 –16 –4 1
Figure 2. Correlation of Apo(a) concentration values assigned using the NWLRL ELISA method and the LC/MS assay.
Figure 3. Correlation of Apo(a) concentration values assigned using the NWLRL ELISA method and size determined using the electrophoretic method and the LC/MS assay.
Here we have described a modification of our previously published UPLC/MS-based method[17] for the inclusion of Apo(a) as well as characterization of its size through quantitation of the number of kringles. In this single assay, multiple apolipoproteins can be measured in a multiplexed fashion with a minimal volume of plasma (2 μL). As described previously, the limit of quantitation for apolipoproteins A1 and B are much lower than typical plasma concentrations for clinical samples; however, concentrations for Apo(a) span a lower and much broader dynamic range. For this UPLC/MS assay, the limit of quantitation for Apo(a) was determined to be 6 nM, which enables the quantitation of Apo(a) for most subjects. This UPLC/MS assay has passed all typical requirements for use as a ’fit for purpose’ assay to be used for the analysis of clinical research samples, including assay precision, sample-mixing and freeze/thaw stability. Like most UPLC/MS assays, this assay has lower throughput and higher cost relative to immunological methods due to the cost of UPLC/MS instrumentation and 6 min cycle time per sample. However, as a multiplexed assay, the ability to measure multiple proteins from a single sample and the reduced requirement for multiple analyses may reduce the overall cost of sample analysis. Also, the UPLC/MS assay does not require specific antibodies and can be performed using most commercially available triplequadrupole mass spectrometers, thus allowing wide access to researchers interested in the function and biology of Apo(a) which is especially important given the unknown function and correlation of Apo(a) size and disease. Apo(a) concentrations determined using this assay showed excellent correlation and agreement of absolute values with a highly validated and accurate ELISA method, giving confidence to both assays. However, the determinations of size are different. While the established high-resolution SDS agarose gel electrophoretic method is capable of independent measurements of determining the Apo(a) size for the two expressed alleles, the UPLC/MS method measures the average Apo(a) size using a quantitative measurement. This difference in approach may explain the observed differences in size determined for the two methods. It is unclear what impact this measurement of average Apo(a) size can have for clinical research and for elucidating the function of Apo(a) in disease, but this assay may provide an additional tool for solving the relationship between Apo(a) size and function. A potential future application may be to combine measurements of Apo(a) concentration and isotope enrichment as part of tracer-based kinetics studies.[7–9] For this version of the assay, there is not sufficient sensitivity to measure small percentages of isotope enrichment for many plasma samples, but we have demonstrated the ability to measure isotope enrichment using the LFLEPTQADIALLK peptide in fractioned plasma samples (HDL and LDL) using a similar digestion strategy and microflow LC/MS.[30]
Acknowledgements
Rapid Commun. Mass Spectrom. 2014, 28, 1101–1106
The authors would like to thank Dr. Gissette Reyes-Soffer and Dr. Henry Ginsberg of Columbia University for helpful discussion regarding Lp(a) biology and function.
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm
1105
concentrations. Furthermore, the inverse relationship between average Apo(a) size and concentration is less pronounced, as indicated by a more shallow slope.
M. E. Lassman et al.
REFERENCES
1106
[1] S. Tsimikas, L. D. Tsironis, A. D. Tselepis. New insights into the role of lipoprotein(a)-associated lipoproteinassociated phospholipase A2 in atherosclerosis and cardiovascular disease. Arterioscler. Thromb. Vasc. Biol. 2007, 27, 2094. [2] M. Ignatescu, K. Kostner, G. Zorn, M. Kneussl, G. Maurer, I. M. Lang, K. Huber. Plasma Lp(a) levels are increased in patients with chronic thromboembolic pulmonary hypertension. Thromb. Haemost. 1998, 80, 231. [3] B. G. Nordestgaard, M. J. Chapman, K. Ray, J. Boren, F. Andreotti, G. F. Watts, H. Ginsberg, P. Amarenco, A. Catapano, O. S. Descamps, E. Fisher, P. T. Kovanen, J. A. Kuivenhoven, P. Lesnik, L. Masana, Z. Reiner, M. R. Taskinen, L. Tokgozoglu, A. Tybjaerg-Hansen. Lipoprotein (a) as a cardiovascular risk factor: current status. Eur. Heart J. 2010, 31, 2844. [4] G. Utermann, W. Weber. Protein composition of Lp(a) lipoprotein from human plasma. FEBS Lett. 1983, 154, 357. [5] G. Utermann, H. J. Menzel, H. G. Kraft, H. C. Duba, H. G. Kemmler, C. Seltz. Lp(a) glycoprotein phenotypes. Inheritance and relation to Lp(a)-lipoprotein concentrations in plasma. J. Clin. Invest. 1987, 80, 458. [6] E. Boerwinkle, C. C. Leffert, J. Lin, C. Lackner, G. Chiesa, H. H. Hobbs. Apolipoprotein(a) gene accounts for greater than 90% of the variation in plasma lipoprotein(a) concentrations. J. Clin. Invest. 1992, 90, 52. [7] D. J. Rader, W. Cain, K. Ikewaki, G. Talley, L. Zech, D. Usher, H. B. Brewer. The inverse association of plasma Lp(a) concentrations with apolipoprotein(a) isoform size is not due to differences in Lp(a) catabolism but to differences in production rate. J. Clin. Invest. 1994, 93, 2758. [8] M. E. Frischmann, K. Ikewaki, E. Trenkwalder, C. Lamina, B. Dieplinger, S. Muhidien, H. Schweer, J. R. Schaefer, P. Konig, F. Kronenberg, H. Dieplinger. In vivo stableisotope kinetic study suggests intracellular assembly of lipoprotein(a). Atherosclerosis 2012, 225, 322. [9] J. L. Jenner, L. J. Seman, J. S. Millar, S. Lamon-Fava, F. K. Welty, G. G. Dolnikowski, S. M. Marcovina, A. H. Lichtenstein, P. H. R. Barrett, C. DeLuca, E. J. Schaefer. The metabolism of apolipoproteins (a) and B-100 within plasma lipoprotein (a) in human beings. Metabolism 2005, 54, 361. [10] T. Demant, D. Seeberg, A. Bedynek, D. Seidel. The metabolism of lipoprotein(a) and other apolipoprotein B-containing lipoproteins: a kinetic study in humans. Atherosclerosis 2001, 157, 325. [11] S. M. Marcovina, H. H. Hobbs, J. J. Albers. Relation between number of apolipoprotein(a) kringle 4 repeats and mobility of isoforms in agarose gel: basis for a standardized isoform nomenclature. Clin. Chem. 1996, 42, 436. [12] S. M. Marcovina, J. J. Albers, B. Gabel, M. L. Koschinsky, V. P. Gaur. Effect of the number of apolipoprotein(a) kringle 4 domains on immunochemical measurements of lipoprotein(a). Clin. Chem. 1995, 41, 246. [13] S. Lamon-Fava, S. M. Marcovina, J. J. Albers, H. Kennedy, C. DeLuca, C. C. White, L. A. Cupples, J. R. McNamara, L. J. Seman, V. Bongard, E. J. Schaefer. Lipoprotein(a) levels, apo(a) isoform size, and coronary heart disease risk in the Framingham Offspring Study. J. Lipid Res. 2011, 52, 1181. [14] A. N. Hoofnagle, M. H. Wener. The fundamental flaws of immunoassays and potential solutions using tandem mass spectrometry. J. Immunol. Methods 2009, 347, 3. [15] R. G. Kay, B. Gregory, P. G. Grace, S. Pleasance. The application of ultra-performance liquid chromatography/ tandem mass spectrometry to the detection and quantitation of apolipoproteins in human serum. Rapid Commun. Mass Spectrom. 2007, 21, 2585.
wileyonlinelibrary.com/journal/rcm
[16] S. A. Agger, L. C. Marney, N. A. Hoofnagle. Simultaneous quantification of apolipoprotein A-I and apolipoprotein B by liquid-chromatography–multiple-reaction–monitoring mass spectrometry. Clin. Chem. 2010, 56, 1804. [17] M. E. Lassman, T. M. McLaughlin, E. P. Somers, A. C. Stefanni, Z. Chen, B. A. Murphy, K. K. Bierilo, A. M. Flattery, K. K. Wong, J. M. Castro-Perez, B. K. Hubbard, T. P. Roddy. A rapid method for cross species quantitation of apolipoproteins A1, B48 and B100 in plasma by ultra-performance liquid chromatography/tandem mass spectrometry. Rapid Commun. Mass Spectrom. 2012, 26, 101. [18] R. Simon, M. Girod, C. Fonbonne, A. Salvador, Y. Clément, P. Lantéri, P. Amouyel, J. C. Lambert, J. Lemoine. Total ApoE and ApoE4 isoform assays in an Alzheimer’s Disease case-control study by targeted mass spectrometry (n = 669): A pilot assay for methionine-containing proteotypic peptides. Mol. Cell. Proteomics 2012, 11, 1389. [19] L. Anderson, C. Hunter. Quantitative mass spectrometric multiple reaction monitoring assays for major plasma proteins. Mol. Cell. Proteomics 2005, 5, 573. [20] T. A. Addona, S. E. Abbatiello, B. Schilling, S. J. Skates, D. R. Mani, D. M. Bunk, C. H. Spiegelman, L. J. Zimmerman, A. J. L. Ham, H. Keshishian, S. C. Hall, S. Allen, R. K. Blackman, C. H. Borchers, C. Buch, H. L. Cardasis, M. P. Cusack, N. G. Dodder, B. W. Gibson, J. M. Held, T. Hiltke, A. Jackson, E. B. Johnson, C. R. Kinsinger, J. Li, M. Mesri, T. A. Neubert, R. K. Niles, T. C. Pulsipher, D. Ransohoff, H. Rodriguez, P. A. Rudnick, D. Smith, D. L. Tabb, T. J. Tegeler, A. M. Variyath, L. J. Vega-Montoto, A. Wahlander, S. Waldemarson, J. R. Whiteaker, L. Zhao, N. L. Anderson, S. J. Fisher, D. C. Liebler, A. G. Paulovich, F. E. Regnier, P. Tempst, S. A. Carr. Multi-site assessment of the precision and reproducibility of multiple reaction monitoring-based measurements of proteins in plasma. Nat. Biotechnol. 2009, 27, 633. [21] M. A. Kuzyk, D. Smith, J. Yang, T. J. Cross, A. M. Jackson, D. B. Hardie, N. L. Anderson, C. H. Borchers. MRM-based, multiplexed, absolute quantitation of 45 proteins in human plasma. Mol. Cell. Proteomics 2009, 8, 1860. [22] S. E. Ong, M. Mann. Mass spectrometry-based proteomics turns quantitative. Nat. Chem. Biol. 2005, 1, 252. [23] V. Brun, C. Masselon, J. Garin, A. Dupuis. Isotope dilution strategies for absolute quantitative proteomics. J. Proteomics 2009, 72, 740. [24] E. Ezan, M. Dubois, F. Becher. Bioanalysis of recombinant proteins and antibodies by mass spectrometry. Analyst 2009, 134, 825. [25] E. E. Chambers, C. Legido-Quigley, N. Smith, K. J. Fountain. Development of a fast method for direct analysis of intact synthetic insulins in human plasma: the large peptide challenge. Bioanalysis 2013, 5, 65. [26] S. M. Marcovina, Z. H. Zhang, V. P. Gaur, J. J. Albers. Identification of 34 apolipoprotein(a) isoforms: Differential expression of apolipoprotein(a) alleles between American blacks and whites. Biochem. Biophys. Res. Commun. 1993, 191, 1192. [27] http://prospector.ucsf.edu/prospector/mshome.htm. [28] http://blast.ncbi.nlm.nih.gov/Blast.cgi. [29] J. L. Proc, M. A. Kuzyk, D. B. Hardie, J. Yang, D. S. Smith, A. M. Jackson, C. E. Parker, C. H. Borchers. A quantitative study of the effects of chaotropic agents, surfactants, and solvents on the digestion efficiency of human plasma proteins by trypsin. J. Proteome Res. 2010, 9, 5422. [30] H. Zhou, J. Castro-Perez, M. E. Lassman, T. Thomas, W. Li, T. M. McLaughlin, D. Xie, P. Jumes, J. A. Wagner, D. E. Gutstein, B. K. Hubbard, D. J. Rader, J. S. Millar, H. N. Ginsberg, G. Reyes-Soffer, M. Cleary, S. F. Previs, T. P. Roddy. Measurement of apo(a) kinetics in human subjects using a microfluidic device with tandem mass spectrometry. Rapid Commun. Mass Spectrom. 2013, 27, 1294.
Copyright © 2014 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2014, 28, 1101–1106
