317© Springer Nature Switzerland AG 2019A. G. Woods, C. C. Darie (eds.), Advancements of Mass Spectrometry in Biomedical Research, Advances in… [616473]

317© Springer Nature Switzerland AG 2019A. G. Woods, C. C. Darie (eds.), Advancements of Mass Spectrometry in Biomedical Research, Advances in Experimental Medicine and Biology 1140, https://doi.org/10.1007/978-3-030-15950-4_18Application of Ion Mobility Mass Spectrometry in LipidomicsFereshteh Zandkarimi and Lewis M. Brown
AbstractLipids play significant roles in biological system, and the study of lipid metabolisms may provide a new insight into the diagnosis and pathophysiology of diseases. Recent developments in high-resolution mass spectrometry tech-niques combined with high-performance chromato-graphic methods provide deep insight into lipid analysis. Addition of ion mobility mass spectrometry orthogonal to LC-MS analysis workflow enhances separation of com-plex lipids, improve isomers resolution, and intensify confidence in lipid identification and characterization. In this chapter, we describe the principle of travelling wave ion mobility mass spectrometry (TWIMS) and its applica-tions in untargeted LC-MS analysis for characterizing the structural diversity and complexity of lipid species in bio-logical samples.KeywordsLipids · Lipidomics · LC-MS · Ion mobility mass spectrometry · MSE · TWIMS · HDMSE18.1 IntroductionLipids are a unique class of biomolecules mediating various structural and functional activities in cellular compartments to maintain homeostasis. The lipid species structurally are composed of a glycerol backbone with different polar head groups and various aliphatic chains differentially connected to the head groups (Fig. 18.1). These functional backbone structures are the foundations of the most widely used lipids classification system including eight categories; fatty acyls, glycerolipids, glycerophospholipids, sphingolipids, sterol lipids, prenol lipids, saccharolipids, and polyketides [1, 2]. The aliphatic chains are varied with different lengths (i.e., different numbers of carbon atoms), various degrees of unsaturation, different locations of double bonds, and branches.As the structural components of cell membranes, lipids affect the fluidity, curvature and membrane interactions. Moreover, lipids are involved in energy storage and transport as well as in cell communication and signaling [3, 4]. Hence, perturbations in lipid metabolism and their synthesis are associated with the pathogenesis of metabolic disorders, cardiovascular diseases, cancer, and inflammation [5–8]. Comprehensive study of lipids and their metabolism in the context of lipidomics is needed to better understand the mechanisms of pathogenesis of a disease. Lipidomics is the study of compositions and abundances of the complete set of lipids (the lipidome) produced in a given cell or organism as well as their interactions with other lipids, proteins and metabolites [4, 9].Due to the chemical complexity and diversity of lipids, advanced and high-throughput analytical platforms have been involved as the main step in the analyses and mea-surements of the lipid species. There are several analytical platforms that have been applied for comprehensive characterization of lipid species, mainly nuclear magnetic resonance (NMR) spectroscopy and mass spectrometry (MS).NMR is a widely used method in lipidomic profiling and metabolic flux analysis because of its robustness, repeatabil-ity as well as providing detailed structural information of the compounds [10]. NMR spectroscopy facilitates the analysis of a wide range of lipids in the micro-molar range. Moreover, it is a non-destructive method so samples can be recovered for further analysis. The main limitation of NMR is its rela-tively low sensitivity when compared to MS especially for the detection of low-abundant lipid species in the complex mixtures [10].18
F. Zandkarimi · L. M. Brown (
) Department of Biological Sciences, Quantitative Proteomics and Metabolomics Center, Columbia University, New York, NY , USAe-mail: LB2425@columbia.edu

318
MS has high selectivity and sensitivity for the identification, quantification, and structural elucidation of several hundreds of lipids in single measurements, the performance of MS can be enhanced by coupling with pre-separation techniques such as gas chromatography (GC), liquid chromatography (LC) and/or ion mobility, which can aid in reducing the complexity of lipids’ separation [11]. GC-MS analysis provides better metabolite separation than LC by minimizing ion suppression events, but the compounds should be thermally stable. GC is only applicable for volatile organic compounds and also requires chemical derivatization of the lipids prior to analysis [12]. LC-MS with a broad range of stationary phase options such as reversed-phase and normal-phase columns has become the most widely used analytical platform for detection and characterization of diverse classes of lipids in complex bio-logical samples [13–15]. Moreover, the integration of ion mobility separations with LC-MS provides both structural and conformational information orthogonal to chromatographic separation and MS-based detection [14, 16].There are two common approaches to analyze lipids including targeted analysis and untargeted analysis. The tar-geted analysis is based on the quantification and validation of one or a set of well-defined lipids in the known pathways [17, 18]. In contrast, the untargeted analysis, or discovery- based approach, aims to identify and quantify all possible lipids present in given samples concurrently without prior knowledge [17]. The main advantage of untargeted analysis is that it may lead to new discoveries in the cellular pathways linking to the biological mechanism.Here we focus mainly on the application of ion mobility in combination with liquid chromatography for analysis of lipid species using an untargeted approach.18.2 Ion Mobility Mass SpectrometryIon mobility spectrometry (IMS) is a gas-phase separation technique that lipid ions are resolved in the ion mobility region filled with the relatively high pressure buffer gas (e.g., nitro-gen) by size, shape, charge, and the interaction with the inert gas under the influence of electric filed [19–23]. There are three commercially available IMS-MS technologies that have been utilized recently for lipidomics analysis: (1) drift- time IMS-MS (DTIMS-MS) [24], (2) travelling wave IMS-MS (TWIMS-MS) [16, 25], and (3) field asymmetric IMS-MS (FAIMS-MS) or differential mobility spectrometry (DMS) [26, 27]. DTIMS separate ions by the time ions needed to pass through the mobility cell in the presence of linear electric field. FAIMS/DMS platforms separate ions by altering voltages and filtering ions in a space-dispersive manner. OOHOOON+P-OOOOOH
OOHOOOO
HHHHOHHHNHOHOOHH
OOPHOOOHPHOOFa/g425y acylsGlycerolipidsGlycerophospholipidsOOHHOHOOPHOOOPHOOOHOHONNHOOHNOOSphingolipids
Sterol lipidsPrenollipidsSaccharolipids
OPOOHOOHOHOOHOHPolyke/g415desFig. 18.1 The categories of major lipid according to their chemical structures along with examples of main classesF. Zandkarimi and L. M. Brown

319TWIMS-MS separates ions based on differences in mobility, after trapping and selectively releasing them. Comprehensive reviews about different ion mobility techniques have been dis-cussed in details previously [16, 19, 21, 28, 29].18.3 Travelling Wave Ion Mobility Mass SpectrometryThe first commercially available traveling wave ion mobility instrument was released in 2006 by Waters Corporation (Milford, MA) named Synapt HDMS [30]. Synapt G2 was released with improved ion mobility separation and resolving power through modifications in 2009. This instrument employs the traveling potential waveform, and it utilizes a six-fold increase in the drift section operational pressure by integrating a helium-filled ion entrance region before the TWIMS [31]. The instrument was modified in 2011 and 2013 (Synapt G2-S and G2-Si), while retaining the same TWIMS configuration, the company altered the source and ion transfer optics to enhance ion transmission through the IM-MS [21, 32]. In principle, the “tri-wave” section of the instrument con-sists of three main traveling-wave (T-wave) ion guide regions including Trap T-wave, ion mobility separation (IMS), and transfer T-wave region. In the trap region, ions are accumu-lated and released as packets into the ion- mobility separation device. After ion mobility separation, ions convey from the transfer region to the orthogonal acceleration time-of-flight analyzer. The “Tri-wave” region consists of a series of stacked ring ion guides (SRIG) filled with neutral background gas such as nitrogen. Ions move through the SRIG by a traveling wave potential that is created by voltage pulses dynamically applied across a stack of ring electrodes. Therefore, TWIMS separations are performed based on the interaction of differ-ent ions to the influence of the specific wave characteristics and may be described as the ability of ions to “surf” on these waves [21]. In the presence of a neutral gas in the TWIMS device, the transport of ions in the SRIG region is dependent on the traveling wave height (V), the velocity of the traveling waves (m/s) and gas pressure (mbar). These factors influence the dwell time of the ion in the SRIG and optimization of these parameters can be employed for separation of ions of the different collisional cross-sections. Therefore, the separa-tion of the ions in the “Tri-wave” region is based on the inter-action of the structural characteristics of the ions (i.e., size, shape, charge, and collision cross section) with the drift gas (normally N2) and the traveling wave. This process of interac-tion of the ions with the traveling wave provides mobility-based separation of the ions in the traveling wave ion mobility spectrometry [20, 23].Since the chromatographic separation occurs in seconds and ion mobility separations typically occur on the order of milliseconds, it is possible to couple IMS with LC and mass spectrometry for lipid analysis. This additional dimension of separation can provide increased peak capacity and substan-tial improvement in analytical results. Specifically, IMS-MS generates cleaner mass spectral data by reducing background noise, improve selectivity by removing interferences in com-plex biological samples, and also intensify confidence in lipid identification and confirmation [19, 22].An example of UPLC-TWIMS-MS chromatogram and corresponding driftogram and 3D plot containing orthogonal ion mobility separation of complex lipid mixtures from a dietary supplement is shown in Fig. 18.2.
Fig. 18.2 Chromatogram and corresponding 2D (drift time vs. retention time, bottom left panel), and 3D (m/z vs. drift time vs. counts; right panel) driftograms from UPLC-TWIMS-MS separation of different lipid classes extracted from dietary supplements [52]
2.04.06.07.08.09.010.03.05.0Reten/g415on Time (min)Dri/g332 Time (ms)Rela/g415ve Intensity (%)IMS separa/g415on of chromatographically co-elu/g415ng lipidsLysophospholipidsPC, PE, PG, & SMChoE, DAG, & TAG
Chromatographic separa/g415on18 Application of Ion Mobility Mass Spectrometry in Lipidomics

32018.4 Coupling Ion Mobility with LC-MS for Enhancing Structural CharacterizationThe coupling of IMS with MS/MS fragmentation tech-niques is also of particular importance for structural eluci-dation of complex lipid species with the same elemental composition. High-quality MS/MS data provide clean frag-ment ion information representing different parts of the molecule of interest and hence improve structural charac-terization of the lipid. For MS/MS investigation of a mole-cule in LC-MS analysis, two common acquisition methods are popular including data dependent acquisition (DDA) and data-independent acquisition (DIA). The DDA method produces MS/MS data from the most abundant ions in a targeted approach [33].Recent advances in high resolution MS instruments facili-tate data-independent analysis (DIA) strategies such as MSE (E stands for and elevated energy) to obtain unbiased all product ions for all precursor ions simultaneously, thus increasing the coverage of lipids and reduce the identifica-tion of false negatives [34–36]. The main advantage of MSE is the ability to provide the information of the precursor and product ions in parallel alternating scans simultaneously in one analytical run. The outputs of LC-MSE are lists of time- aligned precursors and product lipid ions annotated with their respective accurate mass, retention time, intensity, charge state and other physicochemical properties [37–39] (Fig. 18.3).Both the trap T-wave and transfer T-wave regions of Tri- wave regions of the ion mobility chamber are capable of fragmentation process through collision induced dissociation (CID) as shown in Fig. 18.4. If the high collision energy applies only in the trap region, fragmentation occurs prior entering IMS T-wave region (Fig. 18.4a). Then, both products and precursor ions undergo mobility separation, and are guided to the ToF detector with no collision at transfer T wave cell. Therefore, structural information for both pre cursor and product ions is obtained simultaneously. Alternatively, if high collision energy applies only in the transfer T-wave region, fragmentation only occurs after ion mobility separation (Fig. 18.4b). Therefore, product ions are aligned in drift time correlation with their respective precursor ions. This is par-ticularly beneficial in case of co-eluting and isobaric lipids, which results in cleaner MS/MS (product-ion) spectra and reduce false-positive assignments in complex matrices [19]. Another approach to obtain more structural information on the lipid ions is to apply high collision energy in both the trap T-wave and transfer T-wave regions, which is called time-aligned parallel (TAP) fragmentation [40, 41] (Fig. 18.4c). Association of secondary fragment ions produced in the transfer collision cell with specific drift times of first gener-ated fragment ions in the trap collision cell favors the produc-tion of pseudo-MS [3] spectra [40]. All the generated MS/MS and ion mobility data can be processed by software algo-rithms designed by the Waters Corporation such as MSE data viewer and DriftScope. The main function of DriftScope algorithm is to deconvolute multidimensional IMS data
Fig. 18.3 An overview of UPLC-MSE analysis [38]. Reprinted under the terms of the Creative Commons Attribution Non Commercial License from the publisher
Low Energy
m SecUPLC(Chromatographic Separa/g415on)MSEElevated EnergyF. Zandkarimi and L. M. Brown

321
(m/z, RT, peak intensity, and drift time) to a two- dimensional format (for example m/z and drift time).Moreover, calculating collision cross section (CCS) values, which represents a physiochemical property of an ion facilitate the confidence of lipid structural characterization, leading to enhanced specificity of lipid identification [42]. In traveling wave ion mobility, the relationship between CCS and measured drift time cannot be calculated directly, since the motion of the ions in “Tri-wave” region in the pres-ence of electrodynamic waveform potential is complicated, and to date, is not fully addressed [20]. Therefore, the TWIMS device is usually calibrated using mixture of molecule ions with known CCS values that have been previously measured in DT-IMS instrument [42]. Finally, the experimentally deter-mined CCS values of the lipid ions will be compared with theoretical CCS values from databases to confirm the assign-ment of structural information of the given lipid spices [25, 43, 44]. Therefore, the incorporation of CCS values to lipidomics workflow considerably enhanced the accuracy of lipid assignment [25].
Tri-wave Region
21Quadrupole isolateion of interestm/zDrift timem/zDrift timeProduct ions Product ionsseparated by IMSm/zDrift timem/zDrift timem/zvPrecursor ionDrift timeIMS separation of a precursor ionDrift timePrecursor & product ions are drift time aligned1
2
m/zm/zDrift timem/zDrift time1stproduct ions 21
m/zDrift timeIMS separation of 1stproduct ions1stproduct ions and 2ndproductions are drift time alignedTrap FragmentationTransfer Fragmentation
TAP Fragmentationa)b)c)
Fig. 18.4 The principle of different fragmentations possibilities in Tri-wave IMS region including elevated collision induced dissociation in (a) only trap region, (b) only transfer region, and (c) both regions (TAP) [52]18 Application of Ion Mobility Mass Spectrometry in Lipidomics

32218.5 ApplicationsWith the support of the high-resolution, high-throughput, and structural elucidation capabilities, an ion mobility sys-tem is remarkably well-suited for untargeted lipidomics applications. Several studies have shown the advantages of ion mobility combined with LC-MS for detection and sepa-ration of lipids. For example, Damen et al. applied reversed phase LC using charged surface hybrid technology com-bined with travelling wave ion mobility mass spectrometry to resolve co-eluting lipid isomers in human plasma extracts and its application for quantitative and qualitative analysis of lipid species [45].We also applied ion mobility in combination with UPLC and high resolution mass spectrometry to separate co-eluted lipid species such as plasmalogen phosphatidylethanolamine (PE p-) lipids on mice brain tissue lipidome samples. For example, co-eluted PE (p-36:1) and PE (p-38:2) at retention time around 9 min were separated clearly in the ion mobility region with different drift time bin numbers, 136.39 and 141.14, respectively as shown in Fig. 18.5. Additional selec-tivity provided by IMS drift time following high collisional energy in transfer region resulted in much cleaner fragment ion spectra making structural elucidation feasible. The time- aligned high energy spectrum for PE (p-36:1) yielded two signature fragments at m/z 281.2459 representing the C18:1 carboxylate anion and m/z 464.3125 indicating the presence of LPE 18:0p (Fig. 18.5c). And, the time-aligned high energy spectrum for PE (p-38:2) resulted in two signature fragments at m/z 309.2761 representing the C20:1 carboxylate anion and m/z 462.2955 indicating the presence of LPE 18:1p (Fig. 18.5d).Moreover, using ion mobility in combination with MSE allows us to separate facial isobaric isomers such as PE p-18:0/20:4 and PE p-16:0/22:4 as shown Fig. 18.6 in a lipid extract from homogenized brain tissue samples. For
9.059.04PE(p38:2)PE(p36:1)728.55911035.6743590.5339702.5421729.5610756.5800728.5590754.5744826.6614953.6595331.2620255.2309281.2475464.3131303.2321a)
b)
Fig. 18.5 (a) Extracted ion chromatograms for co-eluting plasmalo-gen phosphatidylethanolamine. (b) Low and high energy spectra at 9.05–9.07 min shows many co-eluting species. (c) Low energy spec-trum of [PE p18:0/18:1-H]− at RT = 9.05 min; DT = 136.39 bins and the corresponding time-aligned high energy spectrum. (d) Low energy spectrum of [PE p18:1/20:1-H]− a t R T = 9.06 min; DT = 141.14 bins and the corresponding time-aligned high energy spectrum [52]F. Zandkarimi and L. M. Brown

323
c)d)
Fig. 18.5 (continued)
Fig. 18.6 IMS-MSE resolved isobar of ether-linked plasmalogen phos-phatidylethanolamine, PE (p-38:4). (a) Extracted ion chromatograms at m/z 750.5429. (b) Low energy spectrum of [PE p-16:0/22:4-H]− and the corresponding time-aligned high energy spectrum in IMS-MSE (c) Low energy spectrum of [PE p18:0/20:4-H]− and the corresponding time-aligned high energy spectrum in IMS-MSE [52]18 Application of Ion Mobility Mass Spectrometry in Lipidomics

324
Fig. 18.7 Ultra-high resolution IMS separations of sodiated GalSphingosine (d18:1) and GlcSphingosine(d18:1) (m/z 484.622). Separation of the individual isomers with (a) the 1.25 m path; and their mixture with (b) two (30.6 m) and (c) four passes (59.9 m). The hydroxyl group differences between the two isomers are high-lighted in blue and orange [49]. Reprinted under Creative Commons CC BY 4.0 license from the publisher. Adapted from Wojcik et al. [49]instance, the extracted ion chromatogram of 750.5429 at 1 ppm mass accuracy yielded two peaks at 8.39 and 8.56 min. The time aligned low and high collisional energy in HDMSE of m/z 750.5429 and m/z 750.5442 correspond to the deprotonated PE (p-38:4) in negative at respective time points. The fragmentation of the first peak in HDMSE leads to a prominent ion at m/z 436.2792 (PE 16:0p) and m/z 331.2621 (carboxylate anion of C22:4 fatty acid), whereas the second peak fragments to m/z 464.3122 (PE 18:0p) and 303.2326 (carboxylate anion of C20:4 fatty acid). The clarity of the product ion spectra obtained using IMS-MSE to isolate plasmalogen phosphatidylethanolamine molecular species make identification of these molecules explicit and apparent.Untargeted UPLC-HDMSE was also applied for the char-acterization of lipid species such as LPC, SM, and TAG in plasma and lipoprotein samples of hypertriglyceridemic patients treated with extended-release nicotinic acid [46].The benefit of TWIMS separation combined with reversed phase-UPLC is also demonstrated for detection of low abun-dant N-acyl PE lipids in the highly complex lipidome of a mouse brain model of neuroinflammation [47].Hines et al. [48] have applied untargeted lipidomic HILIC separation coupled with HDMSE to assess the lipid homeo-stasis changes in neuroblastoma cell line treated with benzal-konium chlorides with different alkyl chain length. They demonstrated that ion mobility dimension orthogonal to HILIC enhances separation of individual lipid species within glycerophospholipids, glycerolipids, and sphingolipids lipid classes before MS detection, in addition to characterization of isobaric lipids.The separation of galactosylceramide from its glucosyl-ceramide isomer which differs only in the stereochemistry of their glycan C1-hydroxyl moiety is not feasible using con-ventional travelling wave ion mobility due to its low mobility resolution. Wojcik et al. [49] could achieve partial separation of the dehydrated form of these species using high-resolution TWIMS as shown in Fig. 18.7. The high-resolution TWIMS applies travelling waves in a serpentine-shaped structure for lossless ion manipulations (SLIM) platform which provides 30-fold increase in ion mobility resolution compared to tra-ditional drift tubes and TWIMS [50]. They also demonstrated significant improvement in resolution for isomeric PC lipids differing in either their double bond positions or cis/trans locations.18.6 Conclusion Remarks and PerspectivesLipids play significant roles in biological system, and the study of lipid metabolisms may provide a new insight into the diagnosis and pathophysiology of diseases.Current development of high-resolution mass spectrome-try with chromatographic methods provides deep insight into lipid analysis. The development of new ionization techniques and also imaging mass spectrometry make the in situ and F. Zandkarimi and L. M. Brown