Profiling over 1500 Lipids in Induced Lung Sputum and the

University Medical Center Groningen, University of Groningen, Groningen, 9700 RB Groningen, The Netherlands. Anal. Chem. , 2015, 87 (9), pp 4957â€...
3 downloads 14 Views 1MB Size
Download PDF
Article pubs.acs.org/ac

Profiling over 1500 Lipids in Induced Lung Sputum and the Implications in Studying Lung Diseases Ruben t’Kindt,† Eef D. Telenga,‡,§ Lucie Jorge,† Antoon J. M. Van Oosterhout,§,∥ Pat Sandra,† Nick H. T. Ten Hacken,‡,§ and Koen Sandra*,† †

Metablys, Research Institute for Chromatography, President Kennedypark 26, Kortrijk, 8500 Belgium Department of Pulmonary Diseases, §Groningen Research Institute for Asthma and COPD, and ∥Department of Pathology and Medical Biology, Laboratory of Allergology and Pulmonary Diseases, University Medical Center Groningen, University of Groningen, Groningen, 9700 RB Groningen, The Netherlands

‡

S Supporting Information *

ABSTRACT: Induced lung sputum is a valuable matrix in the study of respiratory diseases. Although the methodology of sputum collection has evolved to a point where it is repeatable and responsive to inflammation, its use in molecular profiling studies is still limited. Here, an in-depth lipid profiling of induced lung sputum using high-resolution liquid chromatography coupled to quadrupole time-of-flight mass spectrometry (LC-Q-TOF MS) is described. An enormous complexity in lipid composition could be revealed. Over 1500 intact lipids, originating from 6 major lipid classes, have been accurately identified in 120 μL of induced sputum. By number and measured intensity, glycerophospholipids represent the largest lipid class, followed by sphingolipids, glycerolipids, fatty acyls, sterol lipids, and prenol lipids. Several prenol lipids, originating from tobacco, could be detected in the lung sputum of smokers. To illustrate the utility of the methodology in studying respiratory diseases, a comparative lipid screening was performed on lung sputum extracts in order to study the effect of Chronic Obstructive Pulmonary Disease (COPD) on the lung barrier lipidome. Results show that sphingolipid expression in induced sputum significantly differs between smokers with and without COPD.

N

of sampling procedures, such as bronchoscopy for obtaining bronchoalveolar lavage (BAL), bronchial biopsies, or bronchial epithelial brushes, precludes sampling from a wide range of patients and on repeated occasions.10 Lung sputum induction has been acknowledged as a convenient sampling method to assess pathological airway processes noninvasively. In 2002 a European Respiratory Task Force described its safety, methods of processing, analysis of its constituents, clinical applications, and future directions.11 Although lung sputum is composed of a rich mixture of pulmonary surfactant from the alveoli, inflammatory cells and material secreted from mucous and other glands in the airway space,12,13 most research in COPD focused on the cellular composition of induced sputum. Neutrophils in induced sputum have been found to be significantly higher in COPD than in healthy smokers and nonsmokers in numerous studies; on the other hand there is often considerable overlap for individual values.14 Additionally, sputum neutrophils associate poorly with the severity and prognosis of COPD.14 Therefore, it is not surprising that, along with the cellular fraction of

owadays, lipids are not only referred to as building materials of biological membranes and energy providers. Various lipid species have been described as key components in biological processes such as cell signaling and cell−cell interactions.1−4 Moreover, lipid metabolic enzymes and pathways have been proposed to play a critical role in cancer, diabetes, neurodegenerative diseases, infections, and inflammation.5,6 This knowledge has paved the way for the emergence of the discipline of lipidomics which aims at the comprehensive measurement of lipids present in a biological matrix and the concomitant detection of the individual lipid responses to various stimuli, for example disease and pharmaceutical treatment.4,7 Although the relevance of lipids in respiratory diseases has been recognized,8 the complex composition and function of the lung lipidome is still poorly understood. For Chronic Obstructive Pulmonary Disease (COPD), a disease that affects millions worldwide and is associated with a high disease burden and mortality rate, little is known about the role and the impact of the lipidome. To study the underlying mechanisms of complex heterogeneous diseases like COPD using lipidomics, it would be ideal to have an abundance of high quality lung specimen available. However, in obstructive airway diseases this kind of research is mainly impeded by the relative inaccessibility of samples from the lung.9 The invasive character © XXXX American Chemical Society

Received: February 23, 2015 Accepted: April 17, 2015

A

DOI: 10.1021/acs.analchem.5b00732 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

for RP-LC measurements. Lipid extracts were analyzed on an Acquity UPLC BEH Shield RP18 column (2.1 × 100 mm; 1.7 μm; Waters, Milford, MA, USA) placed in a Polaratherm 9000 series oven (Selerity Technologies, Salt Lake City, UT, USA) at 80 °C. Elution was carried out with a multistep gradient of (A) 20 mM ammonium formate pH 5 and (B) methanol, starting from 50% B to 70% B in 5 min, followed by a slow gradient of 70−90% B in 30 min. Mobile phase B subsequently reached 100% in 0.1 min where it was maintained for an additional 5 min. The flow rate was 0.5 mL/min. The whole system was allowed to re-equilibrate under starting conditions for 15 min. High-resolution accurate mass measurements were obtained on an Agilent 6530 Q-TOF mass spectrometer (Agilent Technologies) equipped with a Jetstream ESI source. The instrument was operated in both positive and negative electrospray ionization mode. Study samples were analyzed in randomized order in both ionization modes, with QC samples analyzed every 5 study samples. Samples were kept at 4 °C in the autosampler tray while waiting for injection. MS/MS experiments were performed in the data dependent acquisition mode (DDA). More details on the MS and MS/MS settings can be found in the Supporting Information. Data Analysis. For comparative lipidomics, raw LC-MS data files were processed in an untargeted fashion using the Molecular Feature Extraction (MFE) algorithm incorporated in the MassHunter Qualitative Analysis 5.0 software package (Agilent Technologies). The resulting feature files were subsequently imported in MassProfiler Professional 12.0 (Agilent Technologies) which aligned, visualized, and filtered the features. The filtered feature list was exported for recursive peak integration (i.e., targeted feature extraction) using the Find by Ion extraction algorithm in MassHunter Qualitative Analysis 5.0 using predefined mass (25 ppm) and retention time extraction windows (15 s). The Agile integrator was used, with an absolute peak height cutoff of 1000 counts. After targeted extraction of features, samples were normalized using total lipid abundance, correcting for sputum dilution, and were imported again in MassProfiler Professional for further statistical analysis. Differences in lipid expression were calculated with the Mann−Whitney-U test. In these analyses over 1500 different lipids were investigated; therefore, we corrected for multiple testing with the Benjamini Hochberg false discovery rate.20 Lipid identification results from an identification strategy that fully exploits the features of the Q-TOF MS system.7,19,21 One representative COPD lung sputum lipid extract was used to identify detected features. Generation of molecular formulas, based on accurate mass, isotope abundance, and isotope spacing both in positive and negative ionization mode, was complemented with accurate mass database searching in an inhouse build database (populated with LIPID MAPS and HMDB entries and theoretical lipid structures) and MS/MS measurements in both ionization modes. Lipid fragmentation mechanisms/spectra in both positive and negative ionization modes have extensively been reported in the literature.19,21−30 An overview of the qualitative fragment ions for the identification of common lipid classes is displayed in Supporting Information Table S-1. Other parameters such as lipid elution behavior and adduct formation interpretation further assisted in the identification.7,19,21 An in-house accurate mass retention time (AMRT) library was subsequently built with formula, exact mass, and retention time of all identified lipids in comma-separated values (.csv) format (compatible

sputum, there is increasing interest in the analysis of its soluble factors in COPD.15 For the most part, the fluid phase of induced lung sputum originates from epithelial lining fluid (ELF), a thin layer covering the alveoli and the mucosa of small and large airways. ELF consists of about 80−90% of lipid species and has biophysical as well as nonbiophysical functions. It prevents alveolar collapse, preserves bronchiolar patency, and protects the lungs from injuries and infections caused by inhaled particles and micro-organisms.12 Because cigarette smoke is the main risk factor for the induction of COPD, studying the lung barrier lipidome might unravel changes in the composition caused by smoking which may be implicated in the development of COPD.8 Contributing to a better understanding of the role of lipids in respiratory disease, comprehensive measurement of lipids in induced lung sputum has been performed using a hyphenated setup composed of high-resolution reversed-phase liquid chromatography (RP-LC) on sub-2 μm particles and elevated temperature coupled to Jetstream electrospray ionization (ESI) quadrupole-time-of-flight (Q-TOF) mass spectrometry (MS). As a proof of concept of the methodology, a comparative screening of induced sputum lipid extracts of healthy neversmokers, chronic smokers without COPD, and chronic smokers with COPD was performed.

■

EXPERIMENTAL SECTION Samples. Induced sputum supernatant samples from three studies were analyzed (study-I: clinicaltrials.gov NCT00848406, study-II: Lo Tam et al.,16 study-III: Willemse et al.17). Seventeen smokers were included with COPD, 20 smokers without COPD, and 14 never-smokers. All studies were approved by the University Medical Center Groningen medical ethical committee, and all subjects provided written informed consent. 60 μL aliquots of the study samples were collected in a Quality Control (QC) pool. This QC pool was then divided into 120 μL aliquots to obtain representative QC samples. More details on samples and the lung sputum induction procedure can be found in the Supporting Information. Lipid Extraction. The lipid extraction method was based on the procedure described by Matyash et al.18 120 μL of lung sputum was transferred into a 2 mL eppendorf tube together with 300 μL of methanol (ULC-MS grade; Biosolve BV, Valkenswaard, The Netherlands). After mixing for 10 s, 1 mL of tert-butyl methyl ether (MTBE; HPLC-grade; Sigma-Aldrich, Bornem, Belgium) was added, and the sample was incubated for 1 h at room temperature in a shaker. Phase separation was subsequently induced by adding 260 μL of water (ULC-MS grade; Biosolve BV). After 10 min incubation at room temperature, the extract was centrifuged for 10 min at 1000 × g thereby generating a lower hydrophilic and upper lipophilic phase with a protein layer on the bottom. One mL of the upper phase was removed and dried in a centrifugal vacuum evaporator (miVac duoconcentrator, Genevac Lim., Ipswich, UK). The dried sample has been reconstituted in 40 μL of MTBE/isopropyl alcohol (HPLC grade; Sigma-Aldrich) 50/50 (v/v). Two and 6 μL of this extract have been used for LC-MS analysis in positive ESI and negative ESI mode, respectively. Randomization of all study and QC samples was performed before the sample preparation. Liquid Chromatography−Mass Spectrometry. The LCMS method was adapted from Sandra et al.19 A 1200 RRLC system (Agilent Technologies, Waldbronn, Germany) was used B

DOI: 10.1021/acs.analchem.5b00732 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry with the MassHunter software). This provided an automated and targeted data-processing of LC-MS lipid profiles using Find Compounds by Formula from MassHunter Qualitative Analysis 5.0.7,19,31 The AMRT library was used to calculate the distribution of lipids in the sample groups with a peak area cutoff of 5000 counts. Lipid Nomenclature. Throughout the manuscript, the nomenclature of the International Lipid Classification and Nomenclature Committee (ILCNC) has been used, i.e. ‘‘Comprehensive Classification System for Lipids.’’32−34 Lipids for which the exact structure of the two acyl chains could not be elucidated are listed with the total number of carbon atoms and double bonds of the fatty acyl moiety, e.g. PI(34:2). The aliphatic chain composition and regiospecificity of these compounds can be identified with further MS/MS analyses. Regarding glycosphingolipids, LC-MS cannot discriminate between galactose and glucose or N-acetylglucosamine and N-acetylgalactosamine. Therefore, lipid nomenclature of these species contains hexose (Hex) instead of glucose or galactose.

■

RESULTS AND DISCUSSION The Lung Sputum Lipidome Revealed by LC-MS. Lung sputum was centrifuged, and only the supernatant or fluid fraction, representing the humoral fraction of ELF, was subjected to lipid extraction using MTBE. This centrifugation step removes cells and large cellular debris of, most likely, various cell types. The lipids extracted after this step theoretically can be excreted and/or released upon destruction of these cells. Lipid extraction using MTBE generates an upper lipophilic layer, which allows a faster and cleaner lipid recovery, thus more suitable for large profiling studies, compared to the chloroform-based Folch or Bligh and Dyer lipid extraction.18 The lipid analysis platform combines high-resolution RP-LC with accurate mass and high-resolution Q-TOF MS and covers a broad range of lipid species with high precision.7,19 All major lipid classes are covered due to the combination of positive and negative electrospray ionization (ESI) analysis, i.e. glycerophospholipids (GPs), glycerolipids (GLs), sphingolipids (SPs), fatty acyls (FAs), sterol lipids (STs), and prenol lipids (PRs). An abundant lipid extract of one COPD patient was used to identify lipid species based upon accurate mass, MS/MS fragmentation spectra, retention time behavior, and adduct formation.7,19,21 A total of 1584 unique lipids could be accurately identified in the lung sputum extract. 1394 lipid species were detected in positive ESI mode next to 1119 in negative ESI mode. Figure 1 displays the acquired profile plots of all identified lipid species in the fluid fraction of lung sputum, both in positive and negative ESI mode. Table 1 gives a summary of the number of species identified in each lipid class/ subclass for each ionization mode. Supporting Information Table S-2 reports all individual lipid species, representing the lipid class, compound name, formula, the identified molecular ion, and m/z in positive and negative ESI mode. By number, GPs represent the largest lipid class (n = 743), followed by SPs (n = 443), GLs (n = 264), FAs (n = 82), PRs (n = 31), and STs (n = 21). All identified lipids have been listed in a lung sputum accurate mass and retention time (AMRT) library, as described before.7,21 Lipids included in the AMRT library were extracted from all lung sputum extracts in a targeted data processing workflow, which searches for the specific lipid ions in the raw data files in an automated manner. The global lipid distribution acquired in lung sputum of healthy never-smokers, smokers without COPD, and COPD patients, based on the measured

Figure 1. Lipid profiles plotting mass versus retention time of identified lipid species in lung sputum of a COPD patient, representing (a) 1394 individual lipids in positive ESI mode and (b) 1119 individual lipids in negative ESI mode (FA: fatty acyls; PC: phosphatidylcholines; PE: phosphatidylethanolamines; PG: phosphatidylglycerols; PI: phosphatidylinositols; PS: phosphatidylserines; PA: phosphatidic acids; DG: diacylglycerols; TG: triacylglycerols; CE: cholesterol esters; SOLE: solanesol esters; PR: prenol lipids; SLBPA: semilyso bis-phosphatidic acids; CL: cardiolipins; SM: sphingomyelins; Cer: ceramides; Hex: hexose; HexNAc: N-acetylhexosamine; NeuAc: N-acetylneuraminic acid).

lipid abundance in each ionization mode, is shown in Figure 2. The lung sputum lipid composition is discussed below for each main lipid class, typically considered from the perspective of healthy never-smokers (n = 14). Representative fragmentation spectra for uncommon lipids that were detected are shown in Supporting Information Figures S-1−S-10, next to several within-class abundance distribution plots, i.e. subclass distribution, carbon atom chain length distribution, distribution of saturation degree, etc. (Supporting Information Figures S-11− S-16). Glycerophospholipids. Glycerophospholipids (GPs) constitute the bulk components of all mammalian membranes.35 These compounds are essential in provision of membranes required for protein synthesis and export, cholesterol homeostasis, and triacylglycerol storage and secretion.36 GPs, especially phosphatidylcholines (PC), encompass the largest part of pulmonary surfactant.37 Accordingly, the lung sputum fluid phase mainly consisted of GPs. About 66.7% of the total measured lipid abundance in positive ESI (56.6% in negative ESI) could be attributed to this lipid class. All common GP subclasses could be detected in the lipid extract, but the overwhelming majority of GPs were phosphatidylcholines (PC, 67.5%), followed by phosphatidylglycerols (PG, 15.5%), C

DOI: 10.1021/acs.analchem.5b00732 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Table 1. Lipid Categories (in Analogy with the LipidMaps Database)32−34 and Number of Species Identified in Lung Sputum of a COPD Patient (Positive/Negative ESI) main class FA (n = 82) GL (n = 264) GP (n = 743)

SP (n = 443)

ST (n = 21) PR (n = 31) total lipid number

subclass

neg ESI ID

pos ESI ID

unique ID

fatty acids hydroxy fatty acids diacylglycerols triacylglycerols glycerophosphocholines glycerophosphethanolamines glycerophosphoglycerols glycerophosphoinositols glycerophosphoserines phosphatidic acids ceramides phosphosphingolipids glycosphingolipids sterols polyprenols quinones/hydroxyquinones

69 13 N/D N/D 269 187 139 36 29 5 105 44 215 1 4 3 1119

N/D N/D 63 201 326 172 122 30 24 N/D 77 53 276 20 28 2 1394

69 13 63 201 331 191 144 37 35 5 106 54 283 21 28 3 1584

a

SLBPA are semilyso bis-phosphatidic acids or diacylglycerophosphomonoradylglycerols. phospho)-1′-3′-sn-glycerols.

b

comment

acyl, plasmenyl, and plasmanyl species 23 SLBPAa identified, 12 CLb identified

only sphingomyelins (SM) mainly cholesterol esters (CE) mainly tobacco compounds

CL are cardiolipins or bis(1,2-diacyl-sn-glycero-3-

16.9% of the GP fraction, representing 185 unique species. Based on the elution order of LPL regioisomers, the location of the FA chain could be unravelled.7,21 LPL isomers containing an acyl substituent at the sn-2 position elute before their sn-1 isomeric counterparts on RP-LC.38 Furthermore, regioisomers can be discriminated by their fragmentation behavior (Supporting Information Figure S-1).39 Semilyso bis-phosphatidic acids (SLBPA) were identified in small amounts, together about 1.5% of the total GP fraction. These compounds contain three acyl substituents and show a total chain length distribution from C48 to C58. Twelve cardiolipins (CL) were discovered at very low concentrations, i.e. 0.1% of the total phospholipid abundance. Cardiolipins stabilize the electron transfer complex in the inner mitochondrial membrane, which is crucial for physiological ATP production. Their general structure includes a unique dimeric phosphatidyl lipid moiety whereby two phosphatidylglycerols are connected via a glycerol backbone. These so-called polyglycerophospholipids thus contain four fatty acyl moieties, adding up to C74. The total carbon atom number of the conjugated FAs from GPs ranged from 14 to 74 carbon atoms. LPLs showed one aliphatic chain varying from C14 to C24, primarily with 16 or 18 carbon atoms. GPs linked to two FA chains showed a total chain length distribution from C24 to C46, with C32 and C34 as the most abundant ones. The majority of GPs were linked to fully saturated (43.1%) or mono- or diunsaturated FA side chains (23.4% and 23.5%), but the degree of saturation ranged from 0 up to 12. Substantial amounts of ether phospholipids were detected, with either a 1-O-alkyl (plasmanyl) or a 1-O-alk-1′enyl (plasmenyl, also called plasmalogen) linkage at the sn-1 position of the glycerol backbone. These species belonged exclusively to the PC and PE subclass. Plasmanyl species typically elute after their plasmenyl equivalents, which on their turn elute after the acyl substituted equivalents. The majority of PEs were identified as plasmalogens (50.7% of the PE fraction), in contrast to 4.3% of plasmanyl PEs. Next to that, the PC subclass only contained 3.3% of plasmanyl and 0.5% of plasmenyl species. These percentages are in accordance with the percentages of plasmanyl and plasmenyl PE/PC in human

Figure 2. Lipid distribution pie charts measured in positive (left) and negative (right) ESI in 14 healthy nonsmokers (HNS), 20 smokers without COPD (HS), and 17 COPD patients. Peak areas of all identified lipid species were summed up per lipid class for each individual sample, for both the analysis in positive and negative ESI mode. After normalization for sputum dilution, relative quantities for each lipid class were calculated within the sample group (table denotes average percentage ± standard deviation) (GP: glycerophospholipids; GL: glycerolipids; SP: sphingolipids; ST: sterol lipids; PR: prenol lipids; FA: fatty acyls).

phosphatidylethanolamines (PE, 9.9%), phosphatidylinositols (PI, 5.4%), phosphatidylserines (PS, 1.7%), and phosphatidic acids (PA,