Article pubs.acs.org/ac
Plasma Lipidomic Profiling Method Based on Ultrasound Extraction and Liquid Chromatography Mass Spectrometry Consuelo Pizarro,*,‡ Irene Arenzana-Rámila,‡ Nuria Pérez-del-Notario,‡ Patricia Pérez-Matute,† and José-María González-Sáiz‡ ‡
Department of Chemistry, University of La Rioja, C/Madre de Dios 51, 26006 Logroño, La Rioja, Spain HIV and Associated Metabolic Alterations Unit, Infectious Diseases Area, Center for Biomedical Research of La Rioja(CIBIR), C/Piqueras 98, 26006 Logroño, La Rioja, Spain
†
S Supporting Information *
ABSTRACT: Lipidomics is an emerging field in biomedical research that includes the analysis of all the lipids present in complex biological samples. To evaluate the chemical and biological diversity of lipids, lipid extraction is usually the first step toward lipidomics analysis. Nevertheless, sample preparation is still a time-consuming and error prone analytical step. Therefore, the development of simple and robust methods suitable for high-throughput lipid analysis is of great interest. This study presents a new method for exhaustive lipid fingerprinting of human blood plasma samples based on the employment of methyl tert-butyl ether (MTBE) and ultrasound (US) energy combined with liquid chromatography-electrospray ionization quadrupole-time-offlight mass spectrometry (LC−ESIqToF-MS). First, the MTBE-US extraction step was optimized by means of experimental design methodology. After the optimization step, a comparative study was performed to assess the suitability of the proposed method. The new method allowed extraction time to be reduced to half, in comparison with previously reported methods. The proposed method also allowed increasing extraction repeatability (with RSDs below 5.55%) and efficiency (recoveries higher than 70% were obtained for all lipids evaluated). Moreover, the new proposed method enables more than 800 different features to be detected. Thus, the overall number of lipids identified with the databases for this novel extraction method (352) was the highest of the evaluated methods. The efficiency, precision, and feature detection capacity of the proposed method confirmed its suitability for the evaluation of the lipid profile of human blood plasma samples. Moreover, taking into account its simplicity, low time consumption, and compatibility with automation, the new proposed method could be a suitable alternative to previously reported methods for use in laboratories for comprehensive lipidomic profiling.
L
interest to obtain in-depth knowledge of lipids function in biological systems and their involvement in the development of diseases. It is in this context that lipidomics appears as an emerging field that aims to analyze lipid species in biological systems. Lipidomics provides insights into the specific roles of lipid molecular species and also assists in identifying potential biomarkers.5 To evaluate the chemical and biological diversity of lipids, lipid extraction is usually the first step toward lipidomics analysis. Sample preparation is of great importance
ipids are essential cellular components that comprise a large number of chemically distinct molecular species.1 The LIPID Metabolites and Pathways Strategy (LIPID MAPS) consortium (www.lipidmaps.org) has systematically defined and classified lipids into eight categories, according to their structural and biosynthetic complexity: fatty acyls, glycerolipids, glycerophospholipids, sphingolipids, sterol lipids, prenol lipids, saccharolipids, and polyketides, each with its own classification hierarchy.2 Lipids play an outstanding role in diverse biological functions, including structural, energetic, and regulatory functions,2 and have been shown to be directly involved in many human diseases, such as atherosclerosis, Alzheimer’s disease, and cancer.3−6 Therefore, the identification and quantification of lipids in biological samples could be of great © 2013 American Chemical Society
Received: October 3, 2013 Accepted: November 14, 2013 Published: November 14, 2013 12085
dx.doi.org/10.1021/ac403181c | Anal. Chem. 2013, 85, 12085−12092
Analytical Chemistry
■
for generating high quality lipidomic data sets. Without a deep understanding of the capabilities and limitations of the sample preparation method used in a given study, the accuracy of the biological interpretation of collected data may be compromised.7 Therefore, there is great interest in the development of simple and robust sample preparation methods that allow highthroughput analyses from complex biological matrices, such as blood fluids. Traditionally, several methods based on liquid− liquid extraction (LLE) have been applied to extract lipids from blood fluids. The most popular methods of LLE for lipids were those developed by Folch8 and by Bligh and Dyer.9 The former involves the use of a mixture of chloroform and methanol (2:1). The Bligh and Dyer method incorporated water as an extractant component and reduced solvent to sample ratios compared with the Folch method. However, these methods present certain disadvantages in terms of safety. Indeed, chloroform is highly toxic and carcinogenic. To replace the old, chloroform-based methods, dichloromethane was proposed as an alternative to lipid extraction.10,11 However, this method requires the use of a chlorinated solvent which is also environmentally harmful. More recently, methyl tert-butyl ether (MTBE) has been introduced as a very efficient solvent for high-throughput lipid extraction in biological samples.12,13 It has been proven that MTBE provides comparable results to the Bligh and Dyer method. Moreover, MTBE has lower density than water, which facilitates the collection of the extract. To be useful for large-scale functional lipidomics approaches, the procedure had to meet certain requirements,14 one being the elimination/reduction of time-consuming extraction steps. Despite the advantages of the MTBE lipid extraction protocols proposed previously in blood plasma, these involve long extraction times.12,13 The application of ultrasonic radiation is a powerful aid in the acceleration of various steps in the analytical process.15−17 Consequently, in recent years ultrasound-assisted extraction (USAE) has been widely used to facilitate mass transfer between immiscible phases.18 USAE is an efficient extraction technique that provides high reproducibility at shorter times, simplifies manipulation, and increases yields and often the quality of the extract.19,20 Therefore, following the new trends focused on the development of increasingly simple, sensitive and efficient methods for lipidomic profiling, the main objective of this research was to present a novel sensitive method based on extraction with MTBE and ultrasound (US) energy combined with LC−ESIqToF-MS for exhaustive lipid fingerprinting of human blood plasma samples. Experimental design methodology was applied to achieve adequate conditions for the extraction of lipids. This methodology allowed the authors to investigate and optimize the different parameters that affect the MTBE-USAE procedure. Then, the proposed lipid extraction method was directly compared with some of the methods previously proposed in the bibliography for lipidomic profiling.11,13 The main advantage of the new MTBE-USAE approach is that it can be used to determine the multiple classes of lipids inherent to plasma samples with minimal bias and high efficiency, increasing the number of features/lipids identified, reducing the time required and economic costs. Therefore, the new MTBE-USAE lipid extraction procedure could become the method of reference for comprehensive lipidomic profiling and biomarker identification.
Article
MATERIALS AND METHODS
Chemicals. Ultrapure water, used to prepare all the aqueous solutions, was obtained from a Milli-Q system (Milipore, Bedford, MA, USA). LC−MS grade acetonitrile (ACN), isopropyl alcohol (IPA), and ammonium formate, as well as high-performance liquid chromatography (HPLC) grade methanol and MTBE, were supplied by Aldrich Chemie (Steinheim, Germany). HPLC grade dichloromethane (DCM) was supplied by Scharlau (Barcelona, Spain). Tridecanoic acid (FA (13:0)), glyceryl triheptadecanoate (TG (17:0/17:0/17:0)), 1,2-didodecanoyl-sn-glycero-3-phosphocholine (PC (12:0/12:0)), 1,2-dipalmitoyl-sn-glycero-3phosphoethanolamine (PE (16:0/16:0)), and n-hexanoyl-dsphingomyelin (SM (d18:1/c6:0)) were purchased from Aldrich Chemie (Steinheim, Germany). The purity of all standards was above 95%. Preparation of Lipid Standards. Stock solutions were prepared by dissolving lipid standards in DCM or MTBE at a concentration of 2 mg/mL and stored at −20 °C. Working solutions were diluted to 10 μg/mL in DCM or MTBE prior to spiking studies depending on the experiment. Collection and Handling of Plasma Samples. Plasma samples were provided by La Rioja Blood Bank and came from anonymous donors. Venous blood samples were drawn via antecubital venipuncture from each subject in a sitting position. Fasting conditions were not required. Blood was processed to obtain plasma by centrifugation at 2200g, 15 min at 4 °C. As soon as the plasma samples were delivered to the laboratory, they were frozen and then stored at −80 °C until further use. Lipid Extraction. Lipids were extracted from 30 μL of human blood plasma aliquots. First, 15 μL of Milli-Q water was added to the plasma. Then, 60 μL of methanol was added to precipitate proteins by vortex-mixing for 2 min. Then, to perform the USAE process, different amounts of MTBE (depending on the experiment) containing 10 μg/mL of internal standard were added and dispersed with the aid of ultrasonic radiation. The lipid extraction process was carried out by immersing the mixture into an ultrasonic water bath supplied by ATU Utrasonidos (Valencia, Spain). The level of both liquids (water and sample) was the same. The ultrasound frequency and power were 40 kHz and 100 W, respectively, and temperature and time conditions were controlled for each analysis. Once US extraction was performed, 75 μL of miliQ water was added to the mixture and the organic phase was separated by centrifugation at 3000 rpm for 10 min at 10 °C in a Rotina 38 (Hettich, Tuttlingen, Germany). The lipid extracts contained in the upper phase were collected and poured into an autosampler vial. DCM Method (Bligh−Dyer). The total lipid fraction was extracted from 30 μL of human blood plasma according to the method described by Bligh and Dyer,9 substituting the chloroform with DCM.11 Briefly, 200 μL of MeOH was added to 30 μL of human blood plasma. Once the mixture had been vortexed for 20 s, 400 μL of DCM, containing 10 μg/mL of internal standard, was added and the mixture was vortexed again for 20 s. Then, 120 μL of water was added to induce phase separation and vortexed during 10 s. Then, after equilibration for 10 min at room temperature, the extract was centrifuged at 8000 rpm for 10 min at 10 °C. The lipid extract from the lower organic phase was then collected and poured into an autosampler vial. 12086
dx.doi.org/10.1021/ac403181c | Anal. Chem. 2013, 85, 12085−12092
Analytical Chemistry
Article
MTBE Method (Vortex). MTBE lipid extraction was performed according to previously described protocols.12,13 A total of 20 μL of plasma was mixed with 10 μL of Milli-Q water. Then, proteins were precipitated by adding 40 μL of methanol with a vortex and mixed for 2 min. Then, 200 μL of MTBE, containing 10 μg/mL of internal standard was added, followed by mixing of the solution at room temperature with a vortex during 1 h. After addition of 50 μL of miliQ water, the sample was mixed and centrifuged at 3000 rpm for 10 min at 10 °C. The MTBE upper phase was then collected and poured into an autosampler vial. LC Conditions and Experiments. A Warters Acquity UPLC chromatograph (Milford, MA, USA), equipped with a Waters Acquity HSS T3 100 × 2.1 (i.d.) mm 1.8 μm particle size column and a Waters VanGuard precolumn of the same material, coupled to a Microtof-Q (Q-TOF) mass spectrometer from Bruker Daltonik (GMBH, Germany) with an electrospray interface (ESI), was employed to obtain the blood plasma lipid profile. Chromatographic and mass spectrometry data were acquired with the software Data Analysis Version 4.0 from Bruker Daltonik (GMBH, Germany). The lipid extracts were diluted 5 times with injection solvent prior to analysis. The sample tray was kept at 5 °C, and the column at 55 °C. UPLC separation conditions were as previously described.3 Mobile phase consisted of an acetonitrile−water mixture (60:40, v/v) with 10 mM ammonium acetate (solvent A) and an acetonitrile−isopropanol mixture (10:90, v/v) with 10 mM ammonium acetate (solvent B). The mobile phase composition was varied according to a linear gradient that increased from 40% to 100% B within 10 min, and was held at 100% B for an additional 2 min. Then, it was increased from 0% to 60% A within 3 min, then returned to the initial conditions and maintained at 60% A for 0.5 min. The total run time was 15.50 min. The flow rate was set at 0.4 mL min−1 and the injection volume was 10 μL. The electrospray source was operated both in positive and negative modes. A capillary potential of 4.5 and 3.5 kV, respectively, was used for both polarities. A coaxial nebulizer N2 gas flow (9.0 L min−1) at 200 °C and 3.0 bar of pressure around the ESI emitter was used to assist the generation of ions. Spectra were acquired in full scan mode between m/z 50 and 1500 in positive mode and from 120 to 3000 in negative mode. The mass spectrometer was calibrated across the mass range using internal references. Data Analysis and Lipid Identification. The construction and analyses of the experimental design and the response surfaces were carried out using the Nemrod-W statistical package.21 In terms of the lipid profile studies, raw LC−MS data files were converted from the original Bruker format into netCDF format (*.cdf) to ensure format compatibility with the latter using bioinformatics packages. Then, the XCMS opensource R package incorporating nonlinear retention time alignment, matched filtration, peak detection, and peak patching was used.22 The data were first deconvolved using the xcmsSet command, setting the different parameters as follows: matched filter method, fwhm at 10 and snthresh at 10. Alignment was performed with the group command, fixing bw at 2, minfrac at 0.5 and mzwid at 0.7. Then, retention times were corrected using the retcor command and the Gaussian family argument. Data were aligned again with the group command and the same arguments, and missing data were filled with the fillpeaks
command. As a result, a matrix of metabolite features (with accurate m/z and retention time) and related peak area for each sample was obtained. Then, the R CAMERA package was applied to cluster the different mass features coming from the same metabolite (including the molecular ion, 13C isotopes, adducts, and in-source fragments) into spectra.23 The required correlation both across peaks and across samples was set to 0.75. Background noise was removed by subtracting masses found in blank runs from filtered masses with the enviMass workflow.24 Metabolite features were then matched between the different extraction protocols evaluated by identifying those that had a m/z difference of
