Lipid Metabolism and Lipidomics Applications in Cancer Research

Abstract

Lipids are the critical component of cellular and plasma membrane, which constitutes an impermeable barrier of cellular compartments, and play important roles on numerous cellular processes including cell growth, proliferation, differentiation, and signaling. Alterations in lipid metabolism have been implicated in development and progression of cancers. However, unlike other biomolecules, the diversity in the structures and characteristics of lipid species result in the limited understanding on lipid metabolic alterations in cancers. Lipidomics is an emerging discipline that studies lipids in a large-scale based on analytical chemistry principles and technological tools. Multi-dimensional mass spectrometry-based shotgun lipidomics (MDMS-SL) uses direct infusion to avoid difficulties from alterations in concentration, chromatographic anomalies, and ion-pairing alterations to improve resolution and achieve rapid and accurate qualitative and quantitative analysis. In this chapter, lipids and lipid metabolism relevant to cancer research are introduced, followed by a brief description of MDMS-SL and other shotgun lipidomics techniques, and some applications for cancer research.

1.1. Lipid Classes and Their Functions

Lipids are the crucial component of cellular membrane, which constitutes an impermeable barrier of cellular compartments and provides appropriate motifs for membrane protein function [1]. In addition to the role of energy storage, many lipids play distinct and critical roles in a variety of cellular functions such as signal transduction, in which lipids serve as active second messengers and hormones [2]. Lipids involved in metabolism are highly complex in terms of type and concentration, and vary constantly under physiological, pathological and environmental conditions [3].

According to the definition of the LIPID MAPS consortium, lipids are small hydrophobic or amphipathic molecules, which might originate in whole or in part by carbanion-based condensations of ketoacyl thioesters and/or by carbocation-based condensations of isoprene units [4]. The lipids have been classified into eight categories [4]. 1) Fatty acyls: The fatty acyl groups include various molecules synthesized by chain elongation of acetyl coenzyme A (CoA) primers with malonyl-CoA groups. The fatty acyls may contain cyclic functional groups and/or be substituted by heteroatoms. Fatty acyls, characterized in repeating a series of methylene groups, are the simplest lipids in structure. Therefore, fatty acyls are the basic components of other complex lipids [5]. 2) Glycerolipids: The glycerolipid groups include monoacylglycerol, diacylglycerol, triacylglycerol (TAG), and glycolipids. Glycerolipids are characterized in containing a glycerol backbone with fatty acyl chains connected to the hydroxyl groups of glycerol. Glycerolipids can be hydrolyzed into glycerol, non-esterified fatty acid (NEFA), and/or alkyl variants. Glycerolipids contribute to energy storage, energy metabolism and signal transduction [5]. 3) Glycerophospholipids: Glycerophospholipids (GPLs) are the lipids having at least one glycerol hydroxyl group esterified with one phosphate or phosphonate group, such as phosphatidylcholine (PC), phosphatidyl ethanolamine (PE), phosphatidylinositol (PI), phosphatidylserine (PS), phosphatidylglycerol (PG), phosphatidic acid (PA), cardiolipin, bis(monoacylglycerol)phosphate and lysophospholipids, etc. GPLs are the critical and predominant components of cellular and plasma membranes. GPLs contribute to second messenger generation and are involved in cellular metabolism and signal transduction [5]. 4) Sphingolipids: Sphingolipid groups are characteristic in containing a long chain sphingoid as the core structure, such as ceramide (Cer), sphingomyelin, cerebroside, sulfatide, and gangliosides. Sphingolipids are involved in plasma membranes and cellular signaling [5]. 5) Sterol lipids: Sterol lipids are compounds with core characteristic in four fused rings. Sterol lipids are divided into subcategories of cholesterol and its derivatives, steroids, bile acids and the derivatives, etc. [4]. Cholesterol and its derivatives are critical component of cellular membranes [6]. The steroids, act as hormones and signaling molecules, are involved in various biological and metabolic processes [7]. 6) Prenol lipids: The prenol lipids are synthesized from the five carbon precursors isopentenyl diphosphate and dimethylallyl diphosphate mainly produced through the mevalonate pathway [8]. 7) Saccharolipids: The saccharolipids are a group of lipids containing a sugar backbone straightly connected with fatty acids. Saccharolipids exist in the form of glycans or phosphorylated derivatives [5]. 8) Polyketides: The polyketides are a group of metabolites having the characteristics alike lipids. They come from plant and microbial [4].

The lipids act as cellular membrane components, energy storage depots and signaling molecules to exert their biological functions (Table 1) [5].

Table 1.

The major biological functions of lipid classes [5]

Cellular Functions	Lipid Classes
Membrane structure construction	Glycerophospholipids (e.g., PC, PE, PI, PS, PG, PA, etc.), sphingolipids (e.g., sphingomyelin, cardiolipin, cerebroside, sulfatide, gangliosides, etc.), glycolipids, sterol lipids (e.g., cholesterol, etc.)
Energy storage and metabolism	Glycerolipids (e.g., NEFA, TAG, diacylglycerol, monoacylglycerol, acyl CoA, acylcarnitine, etc.)
Signal transduction	Glycerolipids (e.g. diacylglycerol, monoacylglycerol, acyl CoA, acylcarnitine, NEFA, oxidized fatty acid), sphingolipids (e.g., Cer, sphingosine, sphingosine-1-phosphate), psychosine, N-acylethanolamine, Lyso-lipids, etc.
Other functions	Plasmalogen, acylcarnitine, cardiolipin, PS, etc.

Briefly, the polar lipids are largely involved in construction of cellular and plasma membranes, including the majority of glycerophospholipids and sphingolipids. Some of their metabolism intermediates (e.g., lyso-lipids, diacylglycerol, Cer, and NEFA) contribute to membrane structures with very low abundance, which could be crucial in regulation of membrane structure and cell functions. Any changes in membrane lipid compositions and lipid species could alter membrane permeability, fluidity, and stability, leading to alterations in cell functions. The effects of lipid composition alterations on cellular functions include: 1) alter the matrix that interacts with membrane proteins (e.g., ion channels), which in turn affect the protein configurations and functions; 2) affect the roles of microenvironments in cell communication and cytosolic ion distribution. The anionic lipids play important roles on the stability of microenvironments; 3) influence the processes of membrane fusion, vesicle transport, contact point formation, etc., particularly the changes in the lipids with different special shapes, such as plasmalogen, cardiolipin, and lyso-lipids. For example, cholesterol constitutes the functional domain of lipid rafts on membrane.

Some glycerolipids are involved in energy metabolism and act as energy storage depots, such as NEFA, acylcarnitine, acyl CoA, diacylglycerol, and TAG. The changes in these lipids suggest alterations in energy metabolism and energy homeostasis. Excessive energy deposited in the storage depots could result in increased diacylglycerol and TAG levels. Elevated acylcarnitine indicates either mitochondrial dysfunction or excess fatty acid oxidation. Elevated NEFA and acyl CoA species are usually associated with either increased lipid turnover including lipolysis or de novo biosynthesis. Therefore, accumulation of the lipids involved in energy metabolism could lead to lipotoxicity [9–12], which is related to obesity and insulin resistance [5]. The fatty acyl profiles in lipid species also provide a wealthy of information about fatty acid metabolism from lipid uptake, biosynthesis to lipolysis and lipids oxidation in mammalian systems. High levels of fatty acyls containing n-3 and/or n-6 fatty acid in lipid classes indicate either excess uptake of extracellular fatty acids or reduced utility of these fatty acids through oxidation. Accumulation of fatty acyls 16:0, 16:1, 18:0, and 18:1 strongly suggests increased de novo synthesis. Fatty acid de novo synthesis and their food sources are also related to the location and number of double bond(s) in fatty acids. For instance, the 18:1 (n-9) fatty acyl isomer is the most abundant monoenoic fatty acyl in plant and animal tissues. The 18:1 (n-7) fatty acyl isomer is a common monoenoic fatty acyl group in bacterial lipids and usually exist as minor component in most plants and animal tissue. The 18:1 (n-12) fatty acyl isomer accounts for 50% or more of lipids in seed oils of the umbelliferase family including carrot, parsley and coriander [13].

Recently, some lipid species are found to act as second messengers involved in cellular signal transduction, such as sphingolipids and glycerophospholipids. These lipid species could straightly bind to and subsequently activate a target protein such as a receptor, kinase, and phosphatase, leading to a specific cellular functional regulation. Lipids contributed to the signaling regulation have been summarized by Han et al (Table 2) [5].

Table 2.

Summary of lipid species involved in cellular signaling

Lipid Species	Signaling Pathways
Sphingolipids	1. Cell stress response, programmed cell death [14, 15] and cell aging [16] (Cer) 2. Activate phospholipase A2 to release arachidonic acid through formation of ionophore [17] (Ceramide-1-phospate) 3. Vesicular trafficking, cell division, survival and phagocytosis [18–20] (Ceramide-1-phospate) 4. Endocytosis, cell cycle, and apoptosis through interaction with protein kinases [21] (Sphingosine) 5. Cell survival and migration, inflammation [22] (Sphingosine-1-phsphate)
Glycerophospholipids	1. Agonize inward rectifying potassium channels [23] (PIP2) 2. Generate second messengers IP3 and diacylglycerol (PIP2) 3. Release intracellular calcium ions (IP3) 4. Active the members of protein kinase C family [24, 25]. 5. A protein kinase B to increase binding to extracellular proteins and ultimately enhance cell survival [26] (PIP3)
G-protein-coupled receptors activators	Activate G-protein-coupled receptors such as lysoglycerophospholipids (e.g., Lyso-PA), sphingosine-1-phsphate, platelet-activating factor, endocannabinoids, eicosanoids, fatty acid-hydroxy fatty acid, and retinol derivatives, etc.
Nuclear receptors and transcription factors activators	Bind to nuclear receptors and activate transcription factors, e.g., steroid hormones, retinoic acid, eicosanoids, NEFA, etc.

Additionally, some of lipid classes exert specific functions in various cellular processes. For instance, acylcarnitine is involved in fatty acyl transportation in and out mitochondria. Cardiolipin is involved in electron transport chain to generate ATP in mitochondria. LysoPC species contribute to inflammatory disorders [27–30]. Plasmalogen species, the component of biological membrane, play a plenty of roles in cellular functions such as antioxidize. PS is precursor of mitochondrial PE in mammalian cells [5].

1.2. Lipid Metabolism and Cancer

Alterations in lipid metabolism affect numerous cellular processes including cell growth, proliferation, differentiation, and motility [2]. The crosstalk amongst lipids and cell functions usually occurs during metabolic processes including non-oncogenic and oncogenic metabolic states [5]. There are at least three metabolic networks that are common and associated with lipids. 1) Sphingolipid metabolic pathway and network. Ceramide is the center of this pathway, which could interpret the alterations in sphingolipid classes and the subclasses containing different sphingoid backbones, e.g., dihydrosphingolipids. 2) Glycerophospholipid biosynthesis pathway and network. This network is different from species of animals/plants, yeast and/or bacteria. The fatty acyl chain remodeling is observed in most newly synthesized glycerophospholipid species, which is associated with the activities of transacylase/acyltransferase and fatty acyl hydrolysis. Alterations in individual lipid species indicate hydrolysis and remodeling activities. The alterations in lipid classes suggest a possible combination outcomes of hydrolysis activity, which can be supported by the changes of the corresponding lyso glycerophspholipid levels. 3) Glycerolipid metabolism pathway and network. Monoacylglycerol, diacylglycerol, and TAG are included in this network, in which diacylglycerol could be composed in various cellular compartments and therefore can be composed in TAG species. There is homeostasis amongst monoacylglycerol, diacylglycerol, and TAG synthesis. The activity of sphingomyelin synthase is closely linked to the above network. Acyl CoA is a critical metabolite that participates in the synthesis, degradation and remodeling of many lipids, and thus changes the lipid profiles [5].

1.2.1. Lipid Metabolic Alterations in Cancer

The ‘metabolic switch’ in cancer cells has been observed [31]. The carcinogenic phenotypes are produced by a series of mutation events that combine to alter a variety of signal transduction pathways, which converge to alter core cellular metabolism to meet the needs for rapid dividing cells such as rapid ATP generation to maintain high rate of metabolism, increased molecular biosynthesis, strengthened maintenance of proper cellular redox status, and dissemination of cancer cells to form distant metastases [5]. Since lipids are one of the main energy sources and basic component of living cells, it is not doubt that development of cancer (i.e., uncontrolled cell proliferation and growth) is closely associated with lipid metabolism. Cancer cells exhibit specific alterations in different aspects of lipid metabolism (Fig 1) such as a high rate of de novo lipid synthesis, cholesterol synthesis through mevalonate pathway, increased dependence on lipid oxidation, etc. [2].

Fig 1.

Regulation of lipid metabolism by oncogenic signaling transduction pathways. Briefly, glucose utilization for ATP production decreases, although PI₃K/Akt pathway activation promotes glucose uptake. High rates of de novo lipid synthesis exhibits. Extra cellular fatty acids and fatty acids mobilized from lipid stores can be degraded in the mitochondria through β-oxidation to provide energy. AMP kinase is activated to prevent lipid synthesis and stimulate β-oxidation. Mutant p53 increases cholesterol synthesis via the mevalonate pathway. Sterol regulatory element-binding protein contributes to the regulation of fatty acid and cholesterol biosynthesis pathways.

1.2.1.1. Alterations in Lipid Synthesis

It was found that fatty acid synthase expression was enhanced in the earliest stages of cancer development in lung cancer, prostate cancer and breast cancer [32–34] and was more obvious as cancer progresses [35–37], indicating that the cancer cells exhibit high rates of de novo lipid synthesis starting at a relatively early stage tumor development [33, 38]. It has been observed that cancer cells endogenously synthesize 95% of fatty acids, despite having adequate extracellular fatty acids [39]. Cholesterol is one of the key components of biological membranes contributing to modulate the fluidity of the lipid bilayer, and forms lipid rafts that coordinate the activation of some signal transduction pathway [40]. Accumulation of cholesterol has been observed in prostate cancer cells. The abnormal cholesterol metabolism could influence signal transduction events at the membrane by promoting tumor cell growth, inhibiting apoptotic signals and potentially stimulating other malignant cellular behaviors [41]. The enzyme activities of fatty acid and cholesterol biosynthesis in cancer cells are regulated by sterol regulatory element-binding proteins [42, 43] and phosphoinositide 3-kinase (PI₃K)/Akt/PKB (protein kinase B) [44]. Mutant p53 increases the expression of genes involved in the cholesterol biosynthesis pathway [45].

1.2.1.2. Lipid Remodeling

The rapid growth and proliferation in cancer cells require a large amount of lipids for lipogenesis, biological membrane construction and functioning maintenance. Cellular and plasma membrane are largely composed of cholesterol, glycerophospholipids and sphingolipids [40, 46]. Activation of de novo lipogenesis results in a dramatical difference on membrane lipid composition from that of normal cells, i.e., a lipid remodeling in plasma membrane. For example, phospholipid composition towards more saturated and mono-unsaturated acyl species, leading to modulate membrane biophysics and functions [47]. One example of relationship between genetic changes in cancer and membrane lipid remodeling that drives tumor growth has recently been reported. The lysophosphatidylcholine acyltransferase, a key membrane lipid remodeling enzyme [48], could shape the lipid composition of plasma membrane in growth factor receptor-driven cancers by increasing saturated phosphatidylcholine content which is involved in oncogenic signals transduction [49]. The structures of hormones and growth factors, such as prostaglandins, leukotrienes, lysophosphatidic acid, or steroid hormones, are also composed of lipids [50]. Therefore, the levels of fatty acyls are associated with lipid hormone synthesis and affect tumor-promoting signaling processes [51].

1.2.1.3. Lipid Oxidation

It has been recognized that reprogramming of energy metabolism is one of the emerging hallmarks in cancer progress [52]. Although increased glucose uptake are reprogrammed in most cancer cells, the glucose in cancer cells tends to be used for anabolic processes, such as ribose production, protein glycosylation and serine synthesis, rather than oxidized for ATP production [31]. Fatty acid is an extremely substitute energy source. Fatty acid oxidation is highly relevant to ATP production and NADPH homeostasis, which might be crucial for cancer cell growth and survival [31]. AMP kinase activity is essential for the cancer cells to respond the stressful environments and activate a catabolic switch to increase ATP and NADPH reserves [31]. Increased dependence on lipid oxidation as main energy source in cancer cells has been observed. One such example is prostate cancer, in which generally display a low rate of glucose utilization [53], and overexpression of some β-oxidation enzymes [54], indicating that fatty acid oxidation might be a dominant bioenergetic pathway [38].

1.2.1.4. Signaling Lipids

Bioactive lipids, acting as second messengers and hormones, play important role in signaling regulation. For example, fatty acids, fatty acid and sterol derivatives, and eicosanoids modulate the gene expression through binding and activation of the nuclear hormone receptors, e.g., peroxisome proliferator-activated receptors. These transcription factors control genes that regulate lipid homeostasis that promotes the progression of many diseases including cancer [55]. LysoGPL species are involved in cell proliferation, survival, and migration through the regulation of G-protein-coupled receptors [56]. Increased LysoPA level exhibits in malignant effusions, and its receptors are aberrantly expressed in several human cancers [57]. The intimate and causal relationship between metabolic alterations and cancers makes the metabolic alterations to be a kind of potential target for cancer treatment. For example, the use of inhibitory drugs directed against LysoPA receptors could be effective in suppressing tumor metastasis [57]. Hydrolysis products of PI and its phosphorylated derivatives such as PIP, PIP₂ and PIP₃, or themselves are second messengers and cellular regulators to activate the PI₃K/AKT signaling pathway, which is significant in chemotherapy and radiotherapy for human cancer [58].

In addition to the alternations aforementioned, cancer cells exhibit another dimension of complexity that contains a repertoire of recruited, ostensibly normal cells to create the “tumor microenvironment” of developing tumor [52]. Adipocytes are important producers of a variety of cytokines that contribute to inflammation and angiogenesis which associated with cancer development [59]. Adipocytes have been recognized recently as important components of tumor microenvironment [59], indicating that the lipid metabolism is involved in cancer development.

1.2.2. Lipid Metabolism and Tumor Immunity

Cancer is characteristic in uncontrolled cell growth and proliferation which required fatty acids for biosynthesis of membrane structure and signaling molecules [60]. The metabolic pathways or networks disrupted in various cancer cells might be different from their genetic and tissue background [61], tissue origin and functional phenotype [62, 63], and tumor microenvironment [64]. Metabolite-mediated communication and metabolic changes between tumor cells and tumor-infiltrating immune cells have been observed [65]. Tumor-derived cytokines and the subsequent signal transduction induce the expression of lipid transport receptors, leading to increased lipids uptake. In the tumor microenvironment, the availability and use of fatty acids in T cells are affected by competition with tumor cells. The lipid accumulation in the tumor microenvironment increase the oxidative metabolism and activate the immunosuppressive mechanisms [66]. The lipid accumulation in dendritic cells could reduce their ability to process and present antigens and thereby stimulating T cells. Reduction in lipid content due to the inhibition of fatty acid synthesis could restore the functions of dendritic cells [67], indicating that lipid accumulation in tumor microenvironment can affect the functions of immune cells.

Accumulation of cholesterol, sphingomyelin and saturated phosphatidylcholine species has been found in immunoisolated T cell receptor activation domains, suggesting that the lipid composition in immune cell membrane is involved in its function and signaling [68]. It has been reported that various fatty acids drive the differentiation and proliferation of T cells in the gut. The medium- and long-chain fatty acids support the differentiation of pro-inflammatory Th1 and Th17 cells, whereas the short-chain fatty acid, propionate, promotes the development of T_reg cells [69]. Moreover, a high rate of cholesterol esterification in the tumor could impair T cell responses. Elevated cholesterol content in plasma membranes of CD8⁺ T cells might increase their proliferation and improve their effector function [70]. The studies indicate that lipid metabolism in tumor and immune cells in the tumor microenvironment plays a crucial role on immunosuppression regulation [65].

Macrophages might be affected by abnormal lipid metabolism in cancer. Macrophages are important components of innate immunity that assistance the host defense against infections, but also maintain the tissue homeostasis [71]. One of the important characteristics of macrophages is their plasticity and ability to adopt various activation states in response to their microenvironment and fit their functional requirements. They are crucial partners for tumor cell migration, invasion, and metastasis [71]. Macrophages are elite producers of eicosanoids, which is hydrolyzed from phospholipid membranes by cytosolic phospholipase A2, and other related lipid mediators during inflammation [72]. The metabolic phenotype amongst macrophages is different in the energy metabolism. Fatty acid synthesis predominates in M1 macrophages (normal macrophage), which is characterized by aerobic glycolysis, fatty acid synthesis, and a truncated tricarboxylic acid cycle, leading to accumulation of succinate and citrate, contributing in a specific manner of their pro-inflammatory phenotype [71]. M2 macrophages is a phenotype of tumor-associated microphages. They are dependent on fatty acid oxidation for fuel their bioenergetic demands. The metabolic signature of M2 macrophages is characterized by fatty acid oxidation and an oxidative tricarboxylic acid cycle, functioning as anti-inflammatory component and mediator to tissue homeostasis [71, 73]. Thus, studies on macrophages lipidomes might reveal their metabolic reprogramming, which in turn shed light on the therapeutic approaches to the diseases with a high macrophage involvement, such as cancer.

1.3. Introduction to Lipidomics in Cancer Research

1.3.1. Introduction to Lipidomics

There are hundreds of thousands of individual lipid molecular species in cells. Regarding the lipid distribution in normal mammalian cell, phospholipids (a phosphodiester linked to the sn - 3 position of glycerides) are one of the predominant lipids in mammalian cells accounting for about 60 mol% of total lipids. Glycolipids and sphingolipids account for about 10 mol% of total lipids. The distribution of non-polar lipids, such as TAG and cholesterol, are different from cell types and compartments, ranging from 0.1 to 40 mol%. Metabolites, such as NEFAs, lysolipids, diacylglycerol, ceramides, acyl carnitines, acyl CoA, are usually less than 5 mol% of total lipids, but they can be accumulated and contribute to deleterious pathophysiologic sequelae during different pathologic conditions [74]. Each lipid aforementioned interacts in different compartments and within each bilayer compartment. Many lateral separational domains regulate specific interactions and cell functions (e.g., caveolae [75]). Moreover, the neighbor interactions (annular lipids) significantly regulate membrane dynamics, which in turn affect transmembrane protein kinetics and functions, leading to cellular physiologic function regulation and adaptation [74]. Accordingly, the lipid profiling of cells might be different under different pathologic conditions, closely related to cellular functions. Given the modern lipidomics approaches, a snapshot of entire spectrum of lipids in a cellular/organismal lipidome could be obtained to explore the alterations of lipid redistribution, remodeling and degradation under pathological conditions.

The definition of lipidome refers to “the entire collection of chemical distinct lipid species in a cell, an organ, or a biological system” [76]. According to the analogy to other “omics” disciplines, lipidomics is an emerging research field based on analytical chemistry, studying lipidomes in a large scale and at the levels of intact molecular species [5]. Lipidomic research focus on the structures and functions of the global lipids (profiles) in a certain cell or organism, and their interactions among cellular components, which includes 1) accurately identify the structures of cellular lipid classes and their species including the number of atoms, the number and location of double bonds, the core structures and head groups, individual fatty acyl chains, and the regional specificity of each isomer, etc.; 2) accurately quantify individual lipid species to explore the underlying mechanisms that are involved in cellular signaling/pathway, and discover associated biomarker signatures; 3) reveal the interactions of individual molecular species among lipid classes, proteins, and metabolites; 4) disclose the nutritional or therapeutic status related to disease prevention or intervention [5].

Lipidomics has been widely applied in many research fields and developed sub-discipline categories, e.g., molecular/structural lipidomics, functional lipidomics, nutritional lipidomics, dynamic lipidomics, oxidized lipidomics, mediator lipidomics, neurolipidomics, sphingolipidomics, fatty acidomics, etc., focusing on particular fields ([5] and therein references). The dynamic alterations in lipids during cellular physiological or pathological processes are also revealed by the analysis of lipid structures, cell functions, and their interactions in a spatial and temporal manner. The cellular lipidomes research has provided many new insights into disease conditions through the detailed quantitation of cell’s lipidome (e.g., lipid classes, subclasses, and their molecular species) of diseases, study on the lipid metabolic kinetics and the interactions between lipids and cellular proteomes (e.g., Hazen et al. [77]; Han et al. [78–81]). In short, lipidomics exerts a critical role in revealing the underlying mechanisms of lipid associated diseases by identifying alterations in cellular lipid signaling, metabolism, trafficking, and homeostasis.

1.3.2. Mass Spectrometry for Lipidomics

Unlike other biological molecules, lipids are not characterized by a certain individual structure. The unique chemical structure of most lipid molecular species consists of linear combinations of a small number of building blocks including backbones, head groups, and aliphatic chains [82, 83]. Lipidomics analysis presents a challenge due to the diversities in the structure and characteristics of lipids. Over the past years, a variety of conventional separation technologies for comprehensive analysis of lipids in complex samples has been applied such as thin-layer chromatography, gas chromatography, liquid chromatography (LC), supercritical fluid chromatography, capillary electrophoresis. Recently, mass spectrometry (MS), nuclear magnetic resonance, and other spectroscopic approaches have been introduced and become powerful in lipid analysis due to their advantages. The two major platforms used for lipidomic analysis include the direct infusion approach (i.e., shotgun lipidomics) and chromatographic separation coupled mass spectrometry.

Mass spectrometry is an analytical discipline that use mass spectrometers to study the mass-to-charge (m/z) ratio of individual analytes for structural elucidation and quantitation including steps of molecular ions and related fragments generation, ions separation according to their m/z, signal detection and the intensity of individual ion measurement [5]. A mass spectrometer usually consists of an ion source, a mass analyzer system, a detector, and a data processing system (Fig 2).

Fig 2.

Mass spectrometer composition. A mass spectrometer typically consists of an ion source, a mass analyzer system, a detector, and a data processing system.

1.3.2.1. Ionization Source

An ion source is the part of a mass spectrometer where analytes are ionized. The resulting ions are then transmitted to the mass analyzer. Yang and Han [3] have summarized the ionization techniques that are usually used in modern mass spectrometry for lipidomics.

• Electrospray Ionization (ESI) A soft ionization technique used in mass spectrometry. A fine aerosol spray was generated through the desolation effect by applying a strong electric field on a capillary tube which inside a stream of liquid flowing, and then the electrospray was futher used in the mass analyzer.

• Matrix-assisted Laser Desorption/Ionization (MALDI) A soft ionization technique used in mass spectrometry. Specialized in analyzing large and/or labile molecules such as peptides, proteins, lipids, and polymers etc. also can perform MS imaging analysis of tissue or cell samples. By shining a pulsed laser onto the matrix embedded with analytes, the analyte molecules is ionized with the help of the matrix material by absorbing energy from the laser and form a hot plume of ablated gases and then utilized by the analyzer.

• Atmospheric Pressure Chemical Ionization A soft ionization technique that utilizes gas-phase ion-molecule reactions at atmospheric pressure. A corona discharge electrode was used to ionize the analyte, where either proton transfer/abstraction or adduct formation occurred between the relative proton affinities of the reactant gas ions (e.g., evaporated mobile phase or solvent in most cases) and the gaseous analyte molecules to generate the molecular ions.

• Atmospheric Pressure Photoionization An alternative ionization technique when the analyte was not effectively ionized using ESI or atmosphere pressure chemical ionization method. In this method, A source of 10-eV photons is generated by a vacuum-ultraviolet lamp designed for photoionization detection in gas chromatography. The advantage of this technique is the high collision rate between the analyte and solvent molecules where the vapor of both are introduced into the photoionization region and the dopant photoions react to completion.

• Secondary Ion Mass Spectrometry The most sensitive surface analytical technique of analyzing the composition of thin films or solid surfaces. A primary ion beam (e.g., silver, or gold ion beam) is used to bombard the surface of the analyte to generate secondary ions ejected from the surface and later introduced to the mass analyzer.

• Desorption ESI A combination of ESI and desorption ionization technique, an ambient ionization method. A charged electrospray mist is pneumatically directed to the sample surface where the analytes is desorbed, ionized and subsequently carried by splashed droplets that travel in to the atmospheric pressure interface of the mass spectrometer.

The ESI and MALDI method are two of the most prominently used techniques in lipidomics among previously described ionization techniques. In lipid analysis, several remarkable advantages of ESI-MS are as follows, 1) due to the charge propensity of certain lipid classes, the ion source could serve as a separation device to selectively ionize certain lipid or class of lipid of interest without prior LC separation; 2) high sensitivity and for low concentration of lipid detection fmol/μL (i.e., nM); 3) high quantitation quality and instrumentation response factors for individual molecular species in a polar lipid, when using low lipid concentration to avoid lipid aggregation and after 13C de-isotoping correction; 4) broad linear dynamic range between the ion peaks intensity of a polar lipid species; and 5) an over 95% reproducibility for shotgun lipidomics which utilize direct infusion for any lipid extraction with the presence of internal standards [5].

1.3.2.2. Mass Analyzer

From the ion source the analyte is ionized and then transmitted to the mass analyzer which can separate ions by their m/z. Conventional mass analyzer like magnetic sector field analyzer are too large in size and has limited applications for lipidomics. Much more effective and easier to use analyzers such as quadrupole (Q), time of flight (TOF), and ion trap analyzers are often used for lipidomics study.

• Quadrupole Quadrupole mass analyzers are relatively small and inexpensive. They utilize at least four parallel rods that could generate a radio frequency quadurpole field to select ions by either stabilize or destabilize the ions that go through the path using oscillation electrical fields. They can only allow the ions within a certain range of m/z to pass through at a time. They can also allow a wide range of m/z to be swept quickly by swapping potentials on their rods. One of the most popular instruments in lipidomics is the triple quadrupole (QqQ) mass spectrometer. By setting three consecutive quadrupoles together, a QqQ-type is a powerful platform for both LC-MS and shotgun lipidomics. It offers high sensitivity, very good quantitation quality and multiple scanning modes (e.g. selected/multiple reaction monitoring, neutral-loss scan, precursor-ion scan, and product-ion analysis) with broad linear dynamic range.

• Time of Flight (TOF) In a TOF analyzer, the ions was accelerated in the drift tube and fly towards the detector, and the time of their flight was recorded for the identification of different ions. TOF mass analyzers have been broadly used for lipid analysis in lipidomics ([5] and therein references).

• Ion Trap In a TOF analyzer, the ions was accelerated in the drift tube and fly towards the detector, and the time of their flight was recorded for the identification of different ions. TOF mass analyzers have been broadly used for lipid analysis in lipidomics ([5] and therein references). The three-dimensional ion trap also generates an electric field to select ions with its ring electrode and two end-cap electrode, and has the same principles as a quadrupole mass analyzer, only in ion traps the ions are trapped and sequentially ejected mainly by using an RF field. A linear quadrupole ion trap is similar to 3D ion trap, but it traps ions in a 2D quadrupole field. Ion trap is widely used in lipidomics ([5] and therein references).

1.3.2.3. Detector

The detector is the final component of a mass spectrometer. Detector records either the induced charge or the produced current when an ion passes by or hits a surface. To obtain more meaningful signal, some types of electron multipliers are usually used to amplify the signals.

1.3.2.4. Tandem Mass Spectrometry Techniques

Too much in-source fragmentation can trouble in mass analysis by introducing fragment ions of lipids, but soft ionization techniques (e.g., ESI and MALDI) yield minimal in-source fragmentation under appropriate experiment conditions. Even though in-source fragmentation can lead to complication of lipid analysis, if used properly, in-source fragmentation can also provide structural information. However, in-source fragmentation is rarely used for identification of lipid species in lipidomics compared to tandem mass spectrometry after collision-induced dissociation. There are four main MS/MS modes, including product ion analysis, neutral-loss scan, precursor ion scan, and selected/multiple reaction monitoring, that are particularly useful in lipidomics ([5] and therein references). Listed are the tandem mass spectrometric techniques.

• Product Ion Scan In the first mass analyzer a specific precursor is selected and then passed to be fragmented. The second mass analyzer is applied to records all the result ions of the selected precursor.

• Neutral Loss Scan The first mass analyzer scans all the precursor ions while the second mass analyzer scans the fragment ions set at an offset mass from the first mass analyzer. This neutral loss is related to the precursor structure; therefore, all the precursors that undergo the loss of the specified neutral fragment are monitored. The neutral loss scan mode is very useful in shotgun lipidomics to detect a class or a group of lipids that structurally unique for example certain head group [66–69].

• Precursor Ion Scan A tandem mass spectrometric technique, in which the first mass analyzer scans all the precursor ions while the second mass analyzer only monitors selected fragment ions. The selected fragment ions correspond to the common fragment ions of the precursor. Therefore, all precursors that produce specified fragment ion during fragmentation process are monitored. The precursor ion scan model has also been utilized to effectively detect a given class or a group of lipids that yield a given fragment ion after collision-induced dissociation [66–69].

• Selected/Multiple Reaction Monitoring This is a non-scanning technique for targeted analysis on QqQ-type instruments and it uses two of the mass analyzers as static mass windows to perform the detection of certain fragmentation ions from selected precursor ion. From the precursor ion to the fragment ion generate a related m/z pair called ‘transition’. Multiple reaction monitoring is used to indicate the parallel acquisition of multiple SRM transitions [3]. The selected/multiple reaction monitoring techniques have been widely used for quantitative analysis of individual lipid species in lipidomics when a mass spectrometer is coupled with LC [84–87].

1.3.3. Strategies of Mass Spectrometry-based Shotgun Lipidomics

Shotgun lipidomics, originally described in 1994 [88], is a widely used and promising technology in the field of lipidomics research. Direct infusion is used to avoid the difficulties caused by concentration and ion-pairing changes and chromatographic anomalies to improve the resolution [89], and to achieve fast and accurate qualitative and quantitative analysis of lipid species. The mass spectrometry analysis of molecular ions of individual molecular species in targeted lipid, and the alterations in fragmentation energies and reagent gases can be performed under the same infusion and constant concentration conditions. The characteristic allows shotgun lipidomics to achieve detailed tandem mass spectrometry with multiple fragmentation strategies including precursor ion scan, neutral loss scan, and a variety of other fragmentation techniques [89]. Currently, at least three platforms of shotgun lipidomics have been developed:

1.3.3.1. Tandem Mass Spectrometry-Based Shotgun Lipidomics

Neutral loss scan or precursor ion scan of a characteristic fragment of lipid class specifically detects individual species of this class. A unique building block of a certain kind of lipid class is usually involved. This method is very simple and efficient, yet the process and result of the technique is very accurate, easy to manage and less expensive. Only in one MS/MS acquisition, all individual species in a particular class can be confirmed from a total lipid extract with any commercially available QqQ type mass spectrometer. Neutral loss scan or precursor ion scan can also easily identify lipids that have a stable isotope which could give more insights to the kinetics of lipid turnover, biosynthesis, lipid trafficking and homeostasis, etc. [89].

1.3.3.2. High Mass Accuracy-Based Shotgun Lipidomics

This platform rapidly and efficiently acquires a product-ion spectrum of each protonated/deprotonated molecule ion with a Q-TOF or Q-Orbitrap mass spectrometer. The TOF or Orbitrap analyzer records numerous virtual precursor ion scan in parallel, and the high mass resolution and accuracy inherent in the instrument records the accurate mass of fragment ions (0.1 amu) to minimize any false positive identifications [90]. Identification can be performed from bioinformatic reconstruction of the fragments from precursor ion scan or neutral loss scan. Quantitation can be achieved with a comparison of the sum of the intensities of extracted fragments of an ion to that of a pre-selected internal standard [89].

1.3.3.3. Multidimensional Mass Spectrometry-Based Shotgun Lipidomics (MDMS-SL)

MDMS-Based shotgun lipidomics [83, 91, 92] is a well-known approach to analyze individual lipid species straightly in lipid extracts of biological samples, with which the collision-induced dissociation process depends on the chemical structure of each individual molecular species. The technology maximizes the utility of unique chemical methods inherent in discrete lipid classes to analyze lipids including the molecular species with low abundance. Different types of lipids have different structures and charge properties, which are used to selectively ionize under various experimental conditions to separate specific lipid classes from source. After collision-induced dissociation, the targeted lipid class has its unique fragment pattern, which can usually be predicted according to the covalent structures of these lipid classes. The informational fragment ions derived from the head group or the neutral loss of the head group are used to identify the target lipid class, while precursor ion scan or neutral loss scan of fatty acyl chains is used to identify the individual molecular species present in the lipid class. Additionally, diagnostic ions can be identified and quantified by comparing the peak intensity of a certain ion to that of the selected standard in the same mass spectrum after correction for isotopologue [93, 94]. For example, the presence of a primary amine in phosphoethanolamine-containing species is unique in the cellular lipidome and has been exploited to tag the phosphoethanolamine-containing lipid species with fluorenylmethoxylcarbonyl chloride [95]. The facile loss of fluorenylmethoxylcarbonyl from the tagged lipid species allows one to readily identify and quantify these species with unprecedented sensitivity at an amol/μL level [89, 95].

1.3.4. Techniques of LC-MS-Based Lipidomics

The typical lipidomic analyses for biological samples include sample preparation and lipid extraction, lipid separation, data acquisition and processing.

1.3.4.1. The sample preparation and lipid extraction

Proper sample preparation, storage and processing are required prior to analysis. The ideal methods for sample preparation should be fast, reproducible, and able to extract a wide range of lipids with different polarities, and compatible with the instrumental technique [96]. In the field of lipidomics research, the commonly used sample-preparation methods with high lipid recovery for biological samples include liquid-liquid extraction, organic solvent precipitation, and solid-phase extraction [97]. Organic solvent precipitation methods are till commonly used for most lipidomics research. Their advantages and disadvantages of are summarized in Table 3 [3]. solid-phase extraction, a method of lipids separation, can be used for lipidomic analysis. It is highly recommended if specific lipid fractions interfere with LC-MS measurements, or wherein-depth characterizations of lipid classes are required in lieu of more comprehensive lipidomic profiling [96].

Table 3.

Advantages and disadvantages of organic solvent precipitation methods for lipid extraction

Methods	Solvents	Advantages	Disadvantages
Modified Bligh and Dyer	Chloroform:methanol:H2O (1:1:0.9, v:v:v)	Well-established; Widely used	Hazardous solvents used; Difficult to avoid aqueous component contaminants
Modified Folch	Chloroform:methanol (2:1, v:v), 0.9% NaCl (0.2 volume)	Well-established; Widely used	Hazardous solvents used; Difficult to avoid aqueous component contaminants
Methyl tert-butyl ether	Methyl tert-butyl ether (MTBE):methanol:water (5:1.5:1.45, v:v:v).	Feasible for high throughput and automation	Contains aqueous component contaminants.
Butanol:methanol	Butanol:methanol (3:1, v:v), heptane:ethyl acetate (3:1, v:v), 1% acetic acid	Less water-soluble contaminants	The butanol component in the organic phase is difficult to evaporate.

1.3.4.2. Lipid separation by LC

LC could separate or concentrate different types of compounds according to their physical and chemical properties [89, 96], eliminate the interaction of many lipid species [89], and enables comprehensive analysis of complex samples with trace level species [3].

• Reversed-phase LC This type of LC, which separates lipid based on their hydrophobicity, is the most widely used for the analysis of complex lipids, although the retention time various with lipids. The short (50-150 mm; typically 100 mm) microbore columns (1-2.1 mm I.D.) with sub-2-μm or 2.6-2.8-μm (fused-core) particle size and C18 or C8-modified sorbents are usually used in lipidomics research [96].

• Normal-phase LC This type of LC, which is separates lipids based on their polarities, represents a separation mechanism complementary to reversed-phase LC. Highly non-polar solvents with low ionization capacity are used during normal-phase LC. Normal-phase LC is good for the separation of phospholipids, particular PA. The analysis time in Normal-phase LC is typically longer (30–60 min), which can separate on long columns with large particle-size sorbents. Normal-phase LC can completely separate lipid classes [96].

• Hydrophilic interaction chromatography This type of LC, which compromises the physical properties of both reversed-phase and normal-phase columns to a certain degree, provides better reproducibility and robustness, and is more compatible with MS [98].

• Supercritical fluid chromatography Using columns packed with sub-2-μm particles, this type of setting is a newly emerging technique that can be used for fast lipid profiling (<20 min) [96].

• Two-dimensional liquid chromatography This type of setting, combining two different types of columns described above, separates lipids according to two independent molecular properties. It can optimize the separation conditions in both dimensions, which enables the characterization of further complex lipidomes [99]. However, it is more laborious and more time consuming [96].

1.3.4.3. Electrospray ionization using mass spectrometry (MS) and Data Processing

The LC-MS techniques include selected ion extraction, spectra acquisition using selected or multiple reaction monitoring or full mass spectra acquisition, and data-depend analysis [89]. The ion extraction approach is usually used for global lipid analysis, in which mass spectra are continuously acquired during the elution of the chromatographic column, and then targeted ions are extracted from the acquired data array after chromatographic separation [89]. The ionization modes used in LC-MS based lipidomics differ among lipid classes. ESI positive mode is the most common mode in LC-MS, because it can effectively ionize a variety of lipids [96], while negative mode is usually used for specific lipid classes, including PI, PS, PA [100].

The monitoring of selected or multiple reaction would be specific, if the monitored fragment ions are precursor-specific to the combination of LC separation and on interfering transition [89]. The data analysis elucidates the structures of detected lipid ions [89]. The data processing usually includes a couple of steps. 1) Filtering. Filtering processes the raw signal to remove the noise and/or baselines. 2) Feature detection. Feature detection is used to identify the true ion signals and avoid detecting false positive signals. 3) Normalization. Data normalization eliminates the cofounding factors that cause systematic bias in ion intensities during measurements and/or sample-preparation procedures. Finally, the three-dimensional data will be merged into a two-dimensional data matrix [96].

1.3.5. Lipidomics in Cancer Research

Lipidomics is a rapidly developing field of study that focuses on the identification and quantitation of various lipid species in the lipidome. Additionally, as the boundary of cells and organelles, lipids directly expose to the biochemical changes of intra- and extra-cellular environments [101]. Therefore, alterations in the lipidome might reflect the biochemical changes of the cellular system in diseases such as cancer. Analysis lipids for biological samples will help understanding of lipid-driven mechanisms and identifying lipid-based biomarkers in cancer [101].

Cancer initiation and progression is associated with specific alterations of cellular metabolism that are not only by-products of the disease, but also drive the disease [102]. Warburg first reported the “Warburg effect”, which is that cancer cells prefer to produce ATP by glycolysis, a less efficient pathway than oxidative phosphorylation [103]. Adaptive alterations, such as increased rates of glycolysis, amino acid and lipid turnover are known to be important features of the neoplastic process that promotes unlimited cell proliferation and tumor expansion [104]. Accordingly, altered metabolism of lipid is related to cellular metabolism in cancer, and could be considered a hallmark characteristic of many malignancies [105]. Disturbances in lipid metabolism, such as changes in metabolite patterns, have been reported in ovarian cancer verses benign or healthy controls [106, 107] and in different cancer status, e.g. phospholipids levels reduce in plasma with stage III/IV disease relative to benign controls, but increased in early-stage cancers [108].

Due to its rapid, accurate and high throughput in data acquisition, lipidomics is expected to be a potential approach to identify the biochemical homeostasis and abnormalities caused by interference in early biochemical processes, such as early stage of cancer or cancer initiation [101]. For example, palmitoyl sphingomyelin in urine provides the strongest predictive power for differentiate bladder cancer from non-cancer samples. Urine arachidonate level is higher in bladder cancer versus non-cancer controls [109]. The metabolites involved in fatty acid β-oxidation, such as carnitine and acylcarnitines (carnitine, isovalerylcarnitine, glutarylcarnitine, octenoylcarnitine, and decanoylcarnitine) in urine can clearly distinguished bladder cancer patients form control groups [110]. A prospective study on pre-diagnostic metabolic alteration and cancer risk observed that higher levels of LysoPCs, particular LysoPC (C18:0), are related to lower risks of breast, prostate, and colorectal cancer, but higher levels of PC (C30:0) are associated with increased cancer risk [111], indicating that lipidomics might precede the diagnosis of common malignancies by a couple of years [111].

Lipidomics could reflect the cancer progression. Eicosanoids and sphingolipids have been linked to inflammation associated with cancer development and progression, tumor growth and maintenance for the cancers of colon, breast, and lung [112–114]. Many aggressive cancer cells exhibit decreased PC content and increased LysoPC content, which links with phospholipase A₂ activation, an initial rate-limiting enzyme in eicosanoid biosynthesis [115]. Overexpression of phospholipase A₂ is also associated with the malignant potential in human breast cancer [116]. It has been observed that carnitine and acetylcarnitine in serum decreased in remission and increased in relapse in multiple myeloma patients, indicating that mitochondrial β-oxidation is altered at different cancer stages [117]. A longitudinal lipidomics approach could monitor the cancer progression, although the longitudinal lipidomics could be affected by various confounding factors such as treatment. On the other hand, the longitudinal lipidomics might provide a window into designing a targeted therapies and monitoring.

1.4. Application of MS-based Lipidomics in Cancer Research

Lipids are involved in all of the basic processes essential for tumor development [2]. One of the roles of lipidomics is to identify the molecular mechanism(s) responsible for the alter lipid levels induced by a stimulus [5]. For example, using lipidomics analysis, it has been found that epithelial cell cyclooxygenase-2-dependent eicosanoids might mediate tumor development [118]. Application of lipidomics, particularly the MS-based shotgun lipidomics strategies, in cancer research has been spread reported in past decades from cancer cellular remodeling, metabolism and functions alterations to attempts in biomarkers for cancer diagnosis and progression.

1.4.1. Lipid Profiling and Cellular Remodeling in Cancer

Lipids, particular phospholipids, constitute vital components of cells. Therefore, changed cell number in tumor proliferation might lead to alterations in lipid profiles. Dynamic remodeling in tumor was also reported in a study using global LC-MS based lipidomics [119]. This study on prostate cancer included 76 prostatic cancer patients and 19 benign prostatic hyperplasia patients. Global lipidomics profiling was applied for qualitative and quantitative characterization of lipidome in prostatic tissue samples [119]. The investigators analyzed 350 lipid species including fatty acyls, sterol lipids, sphingolipids, glycerolipids, glycerophospholipids, etc. About 140 distinct alterations of lipid species in prostatic cancer patients were found. PC, PE, PG, PI, Cer, diglyceride, cholesteryl ester and NEFA significantly increased in prostatic cancer tumor by 1.65-15.87 folds. Relative composition of lipidome in prostatic cancer was also found in the study. Diacyl-PC and diacyl-PE percentages increased in prostatic cancer whereas ether-linked PCs (alkyl/acyl-PCs, PC with alkyl substituent) and PEs (alkenyl/acyl-PEs, ethanolamine plasmalogens (pPE)) decreased; percentages of free mono- and poly-unsaturated fatty acids elevated, while the percentage of free saturated fatty acids reduced. In NEFA species composition, SFA% was significantly attenuated by 20%, whereas the percentages of mono- and poly-unsaturated fatty acids were enhanced by 40-50%. In the categories of phospholipids (PC, PE, PI, PS, and PG), mono-unsaturated fatty acids -acyl residues increased by 10-40%, while poly-unsaturated fatty acids -acyl and ether-linked chains were reduced by 10-20% and 20-40%, respectively. These results of lipidomic analysis indicated that dynamic remodeling exists in prostatic cancer tumor.

Another example is the study on the relationship among oxidative stress, aberrant lipid metabolism, and proinflammatory cytokines in Systemic Lupus Erythematosus (SLE) using MDMS-SL [120]. SLE is a chronic inflammatory autoimmune disease characterized by dysfunction of immunocytes and genetic and/or environmental factors. Oxidative stress associated with cardiovascular disease [121] is a major causal factor for the morbidity and mortality in SLE [122]. Increasing in very-low-density lipoprotein cholesterol and TAG, and decreasing in high-density lipoprotein cholesterol are dyslipoproteinemia characteristics in SLE [123]. Using MDMS-SL, Hu et al. [120] analyzed the serum lipid species and their metabolites in SLE female patients aged 20-55 years. They found that the levels of pPE species in SLE patients were significantly lower than those of controls by 27%. The most reduced serum pPE species in SLE patients were 16:0–20:4 (by ~40 mol% reduction) and 18:0–20:4 (by ~38 mol% reduction). The LysoPE (~46 mol%) level in SLE patients increased significantly. Plasmenylcholine (pPC) (18:0-18:0, 16:0-18:2), usually low abundance in human plasma, was detected in SLE patients using MDMS-SL. It was found in the study that serum pPC levels decreased in SLE by ~21 mol%. The results suggested that lipid peroxidation might exist in SLE patients. Since the majority of PE species possess poly-unsaturated fatty acid chains at sn-2 position, increasing in LysoPE species level strongly suggests a molecular mechanism that leads to breakdown of sn-1 aliphatic chains in PE species, which leads to reduction of pPE [120].

The lipid profile alterations contributed to cellular signal transduction were reported by the same research team [124]. Although no alternations were found in serum phosphatidylethanolamine (dPE), PI, SM and Cer levels, the profiles of their species in serum were significantly different between SLE patients and healthy controls. The serum dPE species of 16:0-18:2, 18:0-18:2, 16:0-22:6, and 18:0-22:6 increased in SLE by ~34%, 18%, 54%, and 39%, respectively. PI (18:0-18:2) increased by ~16%. However, the dPE and PI species containing 20:4 fatty acyl chain at the sn-2 position appeared reduction to a certain degree, e.g. dPE (18:0/20:4) reduced by ~26%, dPE (20:0/20:4) reduced by ~66%, and PI (18:1/20:4) reduced significantly. Cer (N22:0 and N23:0) reduced by ~10 %, whereas Cer (N24:1) increased by 26% [124]. In general, PUFA (e.g., arachidonic acid) incorporations are one the most common alterations in glycerophospholipids remodeling. The PUFA species in PE, such as arachidonic acid, linoleic acid, eicosapentaenoic acid, and docosahexaenic acid, are major sources of lipid mediators and endocannabinoids [114]. The acyl moieties of PE species are remodeled by the action of phospholipases and lysophospholipid acyltransferases, which to PE diversity and lipid mediator production. The dPE species containing arachidonic acid remodeling was observed in SLE. The arachidonic acid-derived eicosanoids, such as prostaglandins and leukotrienes, are generated to involve in inflammation, allergy and immune responses in SLE. For example, a series of these reactions resulted in dPE (18:0-20:4) reduction. Elevation of dPE (16:0-18:2, 18:0-18:2, 16:0-22:6, and 18:0-22:6) may compensate for the reduction dPE (18:0-20:4) [124]. Meanwhile, the glycerophospholipids containing docosahexaenoic acid were synthesized in the remodeling pathway. They increase membrane fluidity and regulate the biophysical properties of the membrane to maintain the cellular functions, and also act as the reservoir for producing anti-inflammatory mediators such as resolvins and protectins [125–127]. The cellular lipid remodeling is consistent with and might contribute to the inflammatory and oxidative stress characters observed in SLE patients.

It is noteworthy that the dynamic remodeling might associate with the regulation of metabolism and bio-molecular functions to a certain degree, and in turn be involved in pathophysiological process in cancer. The MDMS-SL technology reveals the changes of lipids profiles, e.g., oxidative stress-related lipids, such as plasmalogens, lysophospholipids, and 4-Hydroxynonenal species, and then discloses the underlying mechanism(s) of the alterations through the analysis on lipid species.

1.4.2. Alterations of Lipids Involved in the Metabolic Pathway in Cancer

The rapid cell proliferation and growth, one of the characteristics of cancer cells, can affect the demand and metabolism in all lipid classes. The different lipid classes play different roles. First, lipids act as components of cellular construction. For instance, lipid biosynthesis and remodeling request NEFAs as basic building blocks. Cellular and plasma membrane are composed largely by cholesterol, glycerophospholipids and sphingolipids. TAG, along with acyl-CoA and acylcarnitine, contributes to energy storage, energy metabolism and ATP generation [5]. Second, lipids act as second messengers and/or hormone involving cell signaling. Lysophospholipids and oxidized lipids are involved in cancer cell proliferation, survival and migration [56]. The hydrolysis products of phosphatidylinositol and its phosphorylated derivatives activate PI₃K/AKT signaling pathway [58], contributing to chemotherapy and radiotherapy for human cancers [58].

Hu et al. [120] investigated the relationship among oxidative stress, abnormal lipid metabolism, and proinflammatory cytokines in SLE to explore the underlying mechanisms pathogenesis and development in SLE. In addition to the presence of lipid peroxidation suggested by the alterations in pPE, LysoPE, they found that the levels of 4-hydroxyalkenals, a sensitive indicator of lipid peroxidation to evaluate oxidative stress, and 4-hydroxynonenal, an indicator of oxidative stress, increased significantly in SLE patients. Given that 4-hydroxy-2(E)-nonenal and 4-hydroxy-2(E)-hexenal are the products of lipid peroxidation of n-6 and n-3 polyunsaturated fatty acids [128, 129], Hu’s study indicates a high oxidative stress existing in SLE patients. The proinflammatory cytokines including IL-6, IL-10, and TNF-α in serum were found increased in SLE patients. A positive correlation was found among IL-10 and some clinical indicators such as the systemic inflammation marker, autoantibody titres, and complement component C3 [130]. It is interesting that serum IL-10 was significantly correlated among pPE species in SLE patients. In this study, the pPE species could predict 95.9% of the variability of IL-10 level. Lipid species including pPE (16:0-20:4, 16:1-20:4, 18:1-20:4, 18:0-20:4, 18:0-22:5, and 18:0-22:4), LysoPE (20:4), and 4-Hydroxynonenal also significantly predicted IL-10 level by 60% in SLE patients. Hu’s study suggested that peroxisomal dysfunction, phospholipase A₂ activation, and peroxidation-mediated degradation might result in pPE species reduction in SLE, indicating that accumulation of sn-2 acyl-containing LysoPE and 4-Hydroxynonenal supported an increased oxidative stress at SLE. The subsequent lipid peroxidation was the underlying mechanism that leads to the pPE reduction in SLE, because plasmalogens function as endogenous antioxidant reagents. The oxidation products, reactive aldehydes, can spread oxidative stress to other intracellular organelles through the bloodstream. With shotgun lipidomics approach, the findings in lipids redistribution and remodeling might reveal the SLE pathogenesis including 1) the elevated 4-hydroxyalkenals suggests the presence of oxidative stress and severe oxidized injury in SLE; 2) as one of the endogenous antioxidants, the pPE reduction in SLE serum suggests that blood cells and/or apolipoprotein are severely damage in SLE; 3) hydroxyalkenals are stable and can escape from the cells, thereby attacking nucleophilic targets far away from the original event site through the bloodstream [129]; 4) lysolipids generated from plasmalogen peroxidation are very toxic, and could lead to cell death and cellular dysfunction [131–133]. The alterations in pPE species and the peroxidation products may serve as novel biomarkers for diagnosis, progression, effectiveness of therapy in SLE [120].

As aforementioned, the upregulation of fatty acid biosynthetic pathway and reprogrammed composition in membrane phospholipids has been observed in prostatic tumors [119]. For example, the key genes in de novo lipogenesis, and genes encoding fatty-acid-transporter protein, plasma membrane fatty-acid-binding protein, scavenger receptor class B type I and phospholipase A₂ were found elevated in tumor tissues, consistent with the aforementioned result of overall lipid abundance up-regulated and poly-unsaturated fatty acid accumulation [119].

1.4.3. Biomarker Identification for Cancer

MS-based lipidomics approach enables reliable and accurate characterization of lipid structure and quantitation in given biological samples. The alterations of lipid metabolism in cancer has been identified including down- or up-regulation of lipid abundance, lipid redistribution, and phospholipids composition and remodeling, etc. In order to explore the potential biological characteristics of pathogenesis of cancer, the correlation between lipids alterations revealed by MS-based lipidomics and the progression of cancer has been investigated [119, 120, 124, 134], e.g., focused on identification of novel molecular biomarker(s) which is extremely important for early diagnosis, prognosis and treatment of cancer.

With the information of global lipidomics profiling and quantitative characterization of lipidome in prostatic tissue samples [119], analysis of the relationship between tumor lipidome and cancer diagnosis and progression revealed that cholesteryl ester species (20:1, 24:5, 24:4, 18:1, and 22:6), Cer species (d18:2-20:0 and d18:1-20:1), NEFA species (22:3) and TAG species (58:1) showed a high discriminant power in distinguishing prostatic cancer tumor from non-tumor [119]. Cholesteryl ester exhibited greatest expression increase by 10.5-45.5 folds in prostatic cancer tumor [119], indicating that the accumulation of cholesteryl oleates could be a potential biomarker for prostatic cancer diagnosis and progression.

The multivariate and multiple regression analysis on the difference of the serum lipid profile between SLE patients and healthy volunteers indicated that the lipid species pattern including pPE (18:0-20:4, 16:0-20:4, 18:0-22:4, 18:0-22:5, and 18:1-22:4), 4-Hydroxynonenal, and LysoPE(18:0) could be a potential biomarkers for SLE [120].

Lipid species profiling could be used as plasma biomarkers for diagnosis of breast cancer [135, 136]. Using a normal-phase/reversed-phase two-dimensional LC-mass spectrometry method, Yang et al. conducted comprehensive lipid profiling included 512 lipid species in human plasma from 6 benign breast tumor patients, 5 breast cancer patients and 9 healthy controls. It was found that lipid species of PI (16:0-16:1) and PI (18:0-20:4) in plasma could be utilized as potential breast cancer biomarkers. While PI (16:0-18:1), PG (36:3), and glucosylceramide (d18:1-15:1) in plasma were demonstrated to be potential biomarkers to evaluate the level of malignancy of breast tumor [136].

Using triple quadrupole liquid chromatography electrospray ionization tandem mass spectrometry, Chen et al. [135] analyzed plasma lipid profiling included 367 lipid species from 13 classes of phospholipids as well as cholesterol esters in 84 patients with early-stage breast cancer (stage 0–II) and 110 patients with benign breast disease. A lipid species pattern of combination of 15 lipid species was reported having diagnostic value according to the predictive model including LysoPC (18:3, 20:2, 20:1, and 20:0), cholesterol esters (C19:1, C19:1, and C20:0), PC (32:1, 34:4, 38:3, 40:5, 40:3, and 44:11), ether-containing PC (32:2, 38:3). The sensitivity, specificity, positive predictive value and negative predictive value of the combination of these 15 lipid species were 83.3%, 92.7%, 89.7%, and 87.9%, respectively, in training set, and were 81.0%, 94.5%, 91.9%, and 86.7%, respectively, in validation set. The area under receiver operating characteristic curve was 0.926 in training set, and 0.938 in the validation set [135]. The results suggested that the pattern of lipid profile is potential biomarkers for diagnosis of breast cancer.

Using ultra-performance LC-electrospray ionization-QTOF-MS, combined with multivariate data analysis, Zhang et al [137]. studied plasma lipid profile in 27 ovarian cancer and 27 benign gynecological tumor patients, and 11 healthy women. Higher LysoPC, lower PC and TAG with specific fatty acid chains in plasma were found in ovarian cancer patients. The lipid species pattern of combination of LysoPC (14:0, 16:0, 18:1, 20:3, 20:4, and 22:6), PC (16:0-18:1, 16:0-18:2, 16:0-20:3, 16:0-20:4, 18:0-18:1, 18:0-18:2, 18:0-20:5, and 18:2-18:2) and TAG (18:2-18:2-16:0) were identified to be potential biomarkers for distinguishing ovarian cancer from healthy population [137].

Using LC-MS, Buas et al [108]. Performed global (untargeted) and targeted metabolite profiling of plasmas isolated from 50 serous ovarian carcinoma and 50 benign controls. Global lipidomics analysis identified 34 metabolites (of 372 assessed) differing significantly between cases and controls in both training and testing sets of analysis. Compared to the controls, 17 glycerolipids and glycerophospholipids decreased in abundance in ovarian carcinoma plasmas including PS (O-18:0-0:0, O-18:0-16:1, O-20:0-0:0), Cholesteryl ester (18:3), TAG (16:0-16:1-16:1, 16:1-16:1-16:1, 17:2-17:2-20:5, 16:1-17:2-17:2, 16:1-17:1-17:2, 16:0-16:0-16:1, 16:1-17:0-17:2), PG (P-20:0-12:0), PE(18:1-20:3), PE(18:1-20:3), etc. The top 4 lipids of them were selected and conducted multivariate modeling analysis, and then compared to CA125, one of the tumor markers. The estimated specificity at 95% sensitivity was ~35% in 4 lipid metabolites, 87% in CA125 alone, and 91% in combination of 4 lipid metabolites with CA125. The authors also assessed the classification accuracy of a multi-marker model between CA125 alone and the combination of 4 lipid metabolites with CA125 (hybrid model). A significant higher level of estimated specificity was found in hybrid model (~43%) versus CA125 alone (~10%). This study provides insight into lipid metabolic alterations that is potentially associated with ovarian tumor. Lipid metabolites in plasma might be used to differentiate cancer from benign and/or healthy controls.

Low-invasive or non-invasive samples, such as plasma, serum and urine, are usually applied in the above biomarker related studies. Body fluids such as whole blood or blood cells, plasma, and serum, are worthy exploring to be a source of biomarker(s) for early diagnosis and progression of cancers, as their lipid profiles reflect the general condition of the whole organism [5]. These samples are also considered as one of choices for cancer monitoring as they offer minimal patient discomfort and can justifiably be taken from healthy population as a control.

1.5. Conclusion

Lipids are the crucial component of cellular membrane, which constitutes an impermeable barrier of cellular compartments, and play important roles in numerous cellular processes including cell growth, proliferation, differentiation, and signaling. Cancer cells undergo profound alterations in lipid homeostasis from remodeling, reprogramming, metabolism to signaling. This offers new diagnostic and therapeutic strategies that could be explored by high throughout lipidomics approach. MDMS-SL improves resolution and achieves rapid and accurate qualitative and quantitative analysis. Direct infusion is used to avoid difficulties from alterations in concentration, chromatographic anomalies, and ion-pairing alterations in MDMS-SL. Due to the diversities in the structures and characteristics of lipid classes and their species, it is promising that MDMS-SL has been widely exploited in cancer research from metabolism(s) exploration to biomarker(s) or biomarker signature(s) recognition and identification.

Abbreviations

Cer

ceramide

dPE

phosphatidylethanolamine

ESI

electrospray ionization.

GPL

glycerophospholipids

IP3

inositol triphosphate

LC

liquid chromatography

LysoPA

lysophosphatidic acid

LysoPC

choline lysoglycerophospholipid

LysoPE

ethanolamine lysoglycerophospholipid

LysoPI

lyso-phosphatidylinositol

Lyso-pPE

lysoplasmenylethanolamine

LysoPS

lyso-phosphatidylserine

MALDI

matrix-assisted laser desorption/ionization

MDMS-SL

multidimensional mass spectrometry-based shotgun lipidomics

MS

mass spectrometry

NEFA

nonesterified fatty acid

PA

phosphatidic acid

PC

choline glycerophospholipid

PE

ethanolamine glycerophospholipid

PG

phosphatidylglycerol

PI

phosphatidylinositol

PIP2

phosphatidylinositol diphosphate

pPC

plasmenylcholine

pPE

plasmenylethanolamine

PS

serine glycerophospholipid

Q

quadrupole

QqQ

triple quadrupole

SLE

Systemic Lupus Erythematosus

TAG

triacylglycerol

TOF

time of flight