Comprehensive Lipoprotein Characterization Using Lipidomics

Jun 26, 2017 - Comprehensive Lipoprotein Characterization Using Lipidomics. Analysis of Human Plasma. Nicolas Christinat and Mojgan Masoodi*...
0 downloads 0 Views 2MB Size
Subscriber access provided by Olson Library | Northern Michigan University

Article

Comprehensive Lipoprotein Characterization using Lipidomics Analysis of Human Plasma Nicolas Christinat, and Mojgan Masoodi J. Proteome Res., Just Accepted Manuscript • DOI: 10.1021/acs.jproteome.7b00236 • Publication Date (Web): 26 Jun 2017 Downloaded from http://pubs.acs.org on July 1, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Proteome Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Comprehensive Lipoprotein Characterization using Lipidomics Analysis of Human Plasma Nicolas Christinat1 and Mojgan Masoodi1* 1

Lipid Biology, Nestlé Institute of Health Sciences, EPFL Innovation Park, Bâtiment H, 1015

Lausanne

ABSTRACT

Lipoproteins are responsible for the transport of lipids and other nutrients in the circulation and therefore play an important role in lipid metabolism and dyslipidemia. They have also been linked to multiple metabolic disorders including cardiovascular disease, thus understanding their lipid composition is of crucial importance. Characterization of lipoproteins is a challenging task due to their heterogeneity. In particular, their fractionation is often laborious and time-consuming, making large sets of clinical samples difficult to analyze. We developed and validated lipidomics analysis of lipoproteins including chylomicrons, very low-density, low-density, and high-density lipoproteins. Lipoproteins were first fractionated by polyacrylamide tube gel electrophoresis and, after liquid-liquid extraction, lipids were analyzed by direct-infusion mass spectrometry. About 100 unique lipid species were detected with good reproducibility and reliability. In addition to their lipid composition, valuable information on the fatty acid composition of lipoproteins and lipids was obtained. The presented method offers in-depth analysis of the lipid as well as fatty acid composition of 1 ACS Paragon Plus Environment

Journal of Proteome Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 26

lipoproteins while allowing a good sample throughput. It is thus especially suited for studying lipid associated diseases in clinical cohorts.

Keywords: Plasma lipoproteins, tube gel electrophoresis, Lipids, Lipidomics, Fatty acids. INTRODUCTION Lipoproteins are complex spherical assemblies found in circulation and composed mainly of proteins and lipids. They are responsible for the transport of apolar biomolecules in waterrich (polar) environments such as biofluids. As transporters of lipids between different tissues and organs, plasma lipoproteins play a crucial biological role. They are commonly categorized according to their size and density in four major classes: chylomicrons (CM), very low-density lipoprotein (VLDL), low-density lipoprotein (LDL) and high-density lipoprotein (HDL). Postprandial dyslipidemia is associated with the development of atherosclerosis1-4 and increased cardiovascular disease (CVD), especially in individuals with metabolic syndrome linked to lipoprotein abnormalities. The postprandial state is defined by circulating lipoprotein particles absorbed and processed through the intestine and liver including chylomicrons, VLDL and remnant particles. Chylomicrons are intestinally-derived lipoproteins formed in the intestine after a meal. These triglyceride-rich lipoproteins (TGRL) enter circulation and contribute to postprandial TG concentrations. VLDL particles are also TGRL but their biogenesis occurs in hepatocytes and are present in the plasma during both fasted and fed states. It is worth mentioning that these TGRL are heterogeneous in size and their composition is a key determinant of their health impact such as association with CVD1,2. HDL particles are also heterogeneous in size and composition. Several studies reported a variety of anti-atherogenic and cardioprotective properties of HDL5-8. Understanding the composition of fatty acids (FAs) within these lipoprotein particles is very important to investigate the mechanism of actions especially associate with insulin secretion during 2 ACS Paragon Plus Environment

Page 3 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

postprandial state as well as investigating the sources of FAs. The previous studies demonstrated that, while the majority of FAs used for VLDL synthesis in healthy subjects is from serum non-esterified fatty acids (NEFAs) pool, during postprandial state other sources such as dietary FA as well as hepatic de novo lipogenesis (DNL) contribute to VLDL biosynthesis9,10. Given the heterogeneity of lipoprotein particles and the importance of their composition on their biological function, we aim to develop a comprehensive method to characterize these lipoprotein particles in human plasma. In order to gain knowledge on lipoproteins’ composition, they must be fractionated. Lipoprotein fractionation is typically performed using density gradient ultracentrifugation11 or non-denaturating gradient gel electrophoresis12. Combining one of these fractionation techniques with analytical techniques such as liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) or matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) allows for characterization of lipoprotein composition. Apolipoproteins were extensively studied using classical proteomics techniques13. For instance, Heller et al. published an in-depth study on molecular protein characterization of human plasma lipoprotein by ultracentrifugation, gel electrophoresis, LC-MS/MS and MALDI-TOF MS14. Similarly, ultracentrifugation coupled to MALDI-TOF MS15 or LC-MS16 allowed the profiling of lipid classes such as cholesteryl ester, triglycerides, phospholipids, lysophospholipids, and sphingolipids. Although well established, density-gradient centrifugation and non-denaturating gradient gel electrophoresis are technically demanding and time-consuming, and are thus not well suited to clinical investigations. Alternative lipoprotein fractionation techniques such as sizeexclusion chromatography17,18 or polyacrylamide tube gel electrophoresis19,20 were also developed. Compared to traditional methods, they represent reliable, fast and easy-to-use alternatives. 3 ACS Paragon Plus Environment

Journal of Proteome Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 26

Recently, Castro-Perez et al. reported the lipidomics analysis of Syrian golden hamster plasma lipoprotein fractions, separated with the commercially available Lipoprint™ LDL system21. The same approach was also used to study incorporation of isotopically labelled oleic acid in rodent plasma VLDL, LDL, and HDL lipids22. Liebisch and coworkers used offline fast-performance liquid chromatography to separate human plasma lipoproteins in VLDL, LDL, and HDL fractions. Lipids were extracted with a Bligh and Dyer procedure before being analyzed by electrospray-ionization tandem mass spectrometry23. Here we report the development and validation of a new method for lipidomics analysis of human plasma lipoproteins. In this method, lipoproteins are fractionated in CM, VLDL, LDL, and HDL fractions using the Lipoprint™ LDL system, a commercial available polyacrylamide tube gel electrophoresis system. After extraction with methyl tert-butyl ether/methanol24, lipids are analyzed by a mass spectrometry-based shotgun lipidomics approach25. This three steps approach is technically less demanding and faster than existing methods, resulting in a higher throughput. The latter characteristic makes our methodology particularly well suited for the analysis of large set of samples such as those found in clinical studies. In addition, full characterization of lipid composition including profiling of lipid classes as well as identifying fatty acid composition of different lipid classes within lipoproteins can be investigated with a single measurement. The possibility to simultaneously measure both lipid classes and fatty acid compositions of plasma lipoproteins has so far not been reported and represents a clear advantage over existing methods. EXPERIMENTAL SECTION Chemicals and Reagents Water, methanol and isopropanol (LC-MS grade) were purchased from Fischer Scientific (Waltham, MA USA). Methyl tert-butyl ether, chloroform, ammonium bicarbonate and 4 ACS Paragon Plus Environment

Page 5 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

ammonium acetate were purchased from Sigma-Aldrich (St-Louis, MO, USA). Internal standards were purchased from Avanti Polar Lipids (Alabaster, AL, USA), Larodan (Solna, Sweden) and Sigma-Aldrich (St-Louis, MO, USA). Plasma Sample Collection and Handling Pooled human plasma (2K EDTA) was purchased from Biopredic International (Rennes, France) and was collected from blood donors in blood centers. All blood donors were asked to sign a consent form, and personnel of the Nestlé Institute of Health Sciences did not have access to any direct identifiers or key codes that could link biological material to its donor. Use of human plasma for this project was approved by an independent ethical committee, and the material was handled according to the Swiss Human Research Act. Plasma lipoprotein separation by gel electrophoresis Gel electrophoresis was performed using Lipoprint® LDL kit (Quantimetrix Corporation, Redondo Beach, CA). 25 µL of EDTA plasma follow by 200 µL of Sudan black B dye were loaded onto the gel and gels were run according to manufacturer’s instructions. After electrophoresis was completed, gels were allow to rest for 60 minutes at room temperature before being measured for their cholesterol content. Lipids extraction from electrophoresis gel CM, VLDL, LDL, and HDL bands were cut from gels and placed in a snap-lock tube with 500 µL (CM, VLDL and HDL bands) or 1 mL (LDL band) 150mM aqueous ammonium bicarbonate buffer. The mixture was homogenized by shaking the tubes at 30 Hz for 15 seconds in a Tissue Lyser (Qiagen) prior to centrifugation at 17500 rpm for 10 minutes at 10°C. A defined volume of supernatant (150 µL for CM and VLDL, 120 µL for LDL, and 100 µL for HDL) was transferred in a deep well-plate and further diluted to 180 µL with 5 ACS Paragon Plus Environment

Journal of Proteome Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 26

150mM aqueous ammonium bicarbonate buffer. The resulting solution was extracted as previously described26, using a Star robotic unit (Hamilton Bonaduz AG, Bonaduz, Switzerland) for all pipetting steps. Briefly, the diluted lipoproteins solutions were extracted with 810 µL of methyl tert-butyl ether-methanol (7/2) containing the internal standards. Internal standards and their respective amounts were: 500 pmol 1,2-diheptadecanoyl-snglycero-3-phosphocholine (PC(17:0/17:0)), 50 pmol 1,2-dipalmitoyl-sn-glycero-3-phospho(1'-myo-inositol) (PI(16:0/16:0)), 50 pmol 1,2-diheptadecanoyl-sn-glycero-3-phosphate (PA(17:0/17:0)), 50 pmol 1-dodecanoyl-2-hydroxy-sn-glycero-3-phosphocholine (LPC 12:0), 50 pmol 1-(10Z-heptadecenoyl)-sn-glycero-3-phosphoethanolamine (LPE 17:1), 50 pmol 1tridecanoyl-sn-glycero-3-phospho-(1'-myo-inositol) (LPI 13:0), 50 pmol 1-tridecanoyl-snglycero-3-phospho-L-serine (LPS 13:0), 50 pmol 1-(10Z-heptadecenoyl)-sn-glycero-3phospho-(1'-rac-glycerol) (LPG 17:1), 50 pmol 1-heptadecanoyl-2-hydroxy-sn-glycero-3phosphate (LPA 17:0), 1 nmol cholesterol-d6 (Chol), 200 pmol cholesteryl eicosanoate (SE 20:0), 100 pmol diheptadecanoin (DG(17:0/17:0)), and 100 pmol triheptadecanoin (TG(17:0/17:0/17:0)). The plate was then shaken for 15 min at 4°C in a Thermomixer Comfort C (Eppendorf AG, Hamburg, Germany) and centrifuged at 3000 g for 5 minutes. 100 µL of the organic phase was transferred to a skirted twintec PCR Plate 96 (Eppendorf AG, Hamburg, Germany) and dried in a speed vacuum concentrator. After solvent evaporation in a Concentrator Plus speedvac (Eppendorf AG, Hamburg, Germany), the extract was reconstituted in 40 µL of 7.5 mM ammonium acetate in chloroform/methanol/propanol (1:2:4). The plate was sealed with an aluminum foil and immediately analyzed. DI-MS data acquisition

6 ACS Paragon Plus Environment

Page 7 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Samples were analyzed by direct infusion in a QExactive mass spectrometer (Thermo Fisher Scientific) equipped with a TriVersa NanoMate ion source (Advion Biosciences) as previously described26. Briefly, 5 µL of extract was infused during 5 minutes and analyzed in both polarities in a single acquisition. SE, DAG, and TAG were monitored in FTMS positive mode over the mass range m/z 550-1000 with a resolving power of 140 000 (at m/z = 200). The ammonium adduct of cholesterol was monitor in FTMS positive mode over the mass range m/z 402-412 with a resolving power of 140 000 (at m/z = 200). Common background ions m/z = 680.48022 was used as lock masses for FTMS+ modes. LPG, LPA, LPI, LPS, LPE, and LPC were monitored in FTMS negative mode over the mass range m/z 400-650 with a resolving power of 140 000 (at m/z = 200). Phospholipids were monitored in negative mode over the mass range m/z 520-940 with a resolving power of 140 000 (at m/z = 200). Each MS full scan was followed by a MS/MS data independent analysis over a similar mass range with a resolving power of 17 500 (at m/z = 200) and an isolation width of 1.0 Da. Common background ions m/z = 529.46262 was used as lock masses for FTMS- modes. Data analysis and post-processing All data was analyzed with an in-house developed lipid identification software based on LipidXplorer27,28. Tolerance for MS and MSMS identification was set to 2 ppm in scans where we have lock mass activated and 8 ppm when lock mass was not available. Data visualization was performed with Prism 5.0 software (GraphPad Software Inc.) and for the interrelationships between Lipoprotein fractions, lipid classes and selected fatty acids Chord plots was used and plots were generated using the R circlize package29. RESULTS AND DISCUSSION

7 ACS Paragon Plus Environment

Journal of Proteome Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 26

Optimization of sample preparation for DI-MS analysis Plasma lipoproteins were separated using the Lipoprint® LDL kit. When operated according to manufacturer’s instructions, this tube gel electrophoresis system separates plasma lipoproteins in four fractions: CM, VLDL, LDL, and HDL (Figure 1). In addition, chylomicrons that do not migrate on the gel and remained trapped in the loading gel can also be isolated. Using a lipophilic dye not only allows for cholesterol quantification in lipoproteins sub-fractions but also for precise gel-band cutting. Accurate cutting of the polyacrylamide gel is crucial to ensure analytical reproducibility and to minimize the amount of polyacrylamide gel in the downstream processes. Once cut, gel pieces were homogenized in 150 mM ammonium bicarbonate. After centrifugation, part of the supernatant was collected and further diluted to 180 µL using the same buffer. For each lipoprotein fraction, different sample loads (20-180 µL) were tested in triplicates and linearity range of different sample load was verified. Within the linearity range, the sample load having the highest as well as the most stable total lipid amount was chosen. Optimal values were found to be 100 µL for HDL, 120 µL for LDL, and 150 µL for both VLDL and chylomicrons. After centrifugation, the organic layer was pipetted and dried under vacuum. The residue was reconstituted in a MS buffer and immediately analyzed. For all four fractions, spray stability was good (CV< 15%). Interference of background ions was tested on all fractions as well as on blank samples (gel pieces that did not contain lipoproteins). Visual inspection of spectra did not show any polymer characteristic patterns or intense background ions common to all fractions. The absence of such signals indicates that the polyacrylamide gel did not release significant amounts of impurities during lipid extraction. On the other hand, multiple fraction specific background ions could be detected, in particular in the VLDL and

8 ACS Paragon Plus Environment

Page 9 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

chylomicrons fractions. For instance, two ions at 683 and 743 m/z were detected in the negative mode spectra of these two fractions. None of them represent isobaric interferences with common lipid signals but as their intensities are comparable to highly abundant lipids, it may impact detection of lower abundant species. Lipid detection and identification First, internal standards suitable for analysis were identified. Potential interferences with endogenous species were evaluated in LDL and HDL, the two lipoprotein fractions containing the highest lipid amounts. Although a few isobaric impurities were found in non-spiked HDL and LDL, their signal-to-noise ratio was lower than 10. We thus concluded that these S/N values were too low to have a significant impact on internal standards. Specific amount of internal standards covering 13 different lipid classes were spiked. All of them could be detected after sample extraction and were considered suitable for sample normalization as their coefficient of variation throughout the two batches of 12 identical samples was generally within the 10-25% range which is acceptable based on our experience. Standard titration with these internal standards allowed for the determination of the method’s linear range of each lipid class. Although all lipids in a class do not have exactly the same ionization characteristics and the same limit of detection, this approach remains valid because ionization of a lipid has been shown to be predominantly affected by its charge head group, especially at low concentration30,31. The eight different concentrations were used to test linearity and proportionality for each lipid class and lipoprotein fraction. Among the different fractions, LDL showed the broadest range, followed by HDL, and finally VLDL and CM. In LDL, all lipid classes could be measured at low concentrations (