Recent advances in sample preparation techniques for the lipid profiling based on chromatography-mass spectrometry

YANG Jina, LIU Danyang, ZHOU Ting

(School of Biology and Biological Engineering, South China University of Technology, Guangzhou 510006, China)

Abstract: As the main structural components of cellular and sub-cellular membranes, lipids are a major source of energy and play an important role in various biological processes such as cellular signaling. Lipid profiling has attracted increasing attention in recent years, and chromatography-mass spectrometry techniques for lipid profiling occupy a dominant position in this regard. Sample preparation, which aims at enriching trace substances and reducing matrix interferences, is a crucial step in the analysis of lipids for the intricacy of sample matrices. Here, we review recent developments in sample preparation techniques based on chromatography-mass spectrometry and their application to lipid profiling. Various sample preparation techniques are described and summarized. The extraction methods based on liquid phase included liquid-liquid extraction and single organic solvent extraction. The extraction methods based on solid phase involved solid-phase extraction and solid-phase microextraction. The field-assisted extraction methods cover supercritical fluid extraction, pressurized liquid extraction, microwave-assisted extraction, and ultrasound-assisted extraction. Moreover, online coupling sample preparation methods and sample preparation methods for in vivo lipid analysis are presented. Finally, the problems and trends in sample preparation techniques for lipid profiling based on chromatography-mass spectrometry are discussed. It is believed that efficient development of sample preparation techniques would help improve the sensitivity and selectivity of lipid profiling as well as the analysis speed.

Key words: sample preparation; lipid profiling; chromatography-mass spectrometry

As an important class of hydrophobic or amphipathic small molecules in a biological system, lipids play multiple vital roles in many biological processes [1]. Lipids form the matrices of cellular and sub-cellular membranes, acting as an energy source and ensuring cell membrane integrity. Moreover, lipids are involved in cellular signaling pathways and aid in membrane protein function and interactions [2]. Any defects or anomalies in lipid synthesis and metabolism would lead to obesity [3], diabetes [4], cardiovascular disease [5], Alzheimer’s disease [6] and cancers [7]. Therefore, much attention has been paid to the determination and quantification of lipids, such as endogenous lipid detection [5] and nutritious lipid analysis in plants [8]. Furthermore, lipidomics has emerged as a new technique for the determination of entire lipids in a given biological system, and it has been widely applied to the discovery of biomarkers as well as to unveiling pathogenesis mechanisms and exploring of new drugs [9,10].

Nowadays, techniques based on chromatography-mass spectrometry, including gas chromatography-mass spectrometry (GC-MS), liquid chromatography-mass spectrometry (LC-MS), and supercritical fluid chromatography-mass spectrometry (SFC-MS) [11], are being widely used for lipid analysis. With its longest development history, GC-MS covers the most abundant standard databases, facilitating the identification of lipids [12]. It should be noted that the high temperature required in GC would cause degradation of thermolabile lipids and a derivatization step is necessary for involatile lipids before analysis. Because of its high efficiency and sensitivity, LC-MS, which covers a diverse range of molecular species, has been the dominant method used for the detection of lipids [13-15]. Owing to its extraordinary separation ability for lipids and environment-friendly characteristics, SFC-MS has attracted increasing attention as an alternative to LC-MS and GC-MS for lipid analysis in recent years [16,17]. Currently, three chromatography-mass spectrometry techniques occupy a predominant position in the field of lipid analysis.

Although chromatography-mass spectrometry techniques have greatly impelled the development of lipid profiling, lipid analysis still faces great challenges, especially with regard to sample preparation, because of the wide range of polarities and molecular weights of lipid species [18,19]. Sample preparation, which aims at the concentration of trace analytes and suppression of matrix interference, plays a vital part in the lipid profiling [20]. Appropriate sample pretreatment can not only improve the analytical sensitivity but also protect the instruments from contamination. The importance of sample preparation is further emphasized by the challenges in lipidomics study, which requires comprehensive coverage of the lipidome [21]. Therefore, sample preparation techniques for lipid analysis have attracted increasing attention in recent years.

Several excellent reviews on this topic have been reported in the past decade. In 2011, Wei and Zeng [22] reviewed the sample extraction and instrumental analysis in the quantification of fatty acids. Kortz et al. [23] summarized the sample preparation, chromatographic separation, and mass spectrometric detection of eicosanoids and related lipids in human biological matrices by liquid chromatography-tandem mass spectrometry (LC-MS/MS). In 2015, Teo et al. [24] described sample preparation and analytical techniques for the lipidomics study of clinical samples and highlighted the importance of standardization in the lipidomics workflow in order to reduce variations among different research groups. Subsequently, Jurowski et al. [18] underlined the issues related to the storage and preparation of biological samples in modern medical lipidomics.

Up to now, the extraction methods based on liquid phase such as liquid-liquid extraction (LLE) [25] and single organic solvent extraction (SOSE) [24] have been most commonly used for lipid profiling because of their low cost and easy execution. The extraction methods based on solid phase, e. g., solid-phase extraction (SPE) [26] and solid-phase microextraction (SPME) [27], involve the use of special sorbents, which help minimize matrix interference and significantly improve detection sensitivity in lipid profiling. In contrast to conventional approaches, field-assisted extraction methods, i. e., supercritical fluid extraction (SFE) [28], pressurized liquid extraction (PLE) [29], microwave-assisted extraction (MAE) [30], and ultrasound-assisted extraction (UAE) [31], are also widely employed for lipid extraction since lipid recovery is enhanced with the assistance of pressure, heat, microwaves, and ultrasound. Furthermore, the development of sample preparation techniques, including online coupling sample preparation techniques [32] and sample preparation techniques forinvivolipid profiling, has recently drawn much attention [33].

This paper presents a review of the recent advances in sample preparation techniques for the lipid profiling based on chromatography-mass spectrometry. The sample preparation techniques are categorized into the extraction methods based on both liquid and solid phase, the field-assisted extraction methods, online coupling sample preparation methods, and sample preparation methods forinvivoanalysis. The development, applications, and challenges related to the abovementioned techniques are discussed and summarized. Finally, the problems and trends in sample preparation techniques for the lipid profiling based on chromatography-mass spectrometry are presented.

1 Extraction methods based on liquid phase

The extraction methods based on liquid phase, e. g., LLE and SOSE, which are the most commonly used in lipid profiling processing, are summarized in this section. The advantages and disadvantages of these methods are listed in Table 1.

Table 1 Advantages and disadvantages of extraction methods coupled with different analytical techniques used for lipid profiling of various samples

表 1 (续)

1.1 Liquid-liquid extraction

LLE is the most common sample preparation technique for lipid profiling, which relies on the difference in the solubilities of the target components in immiscible solvents. Almost all biological samples, including cell [34], tissue [40], serum [84], and plasma [85] samples, can be pretreated by this technique. LLE for lipid analysis originated from the chloroform/methanol extraction method proposed by Folch [86], wherein the solvent was chloroform/methanol/water in 8∶4∶3 (v/v/v) ratio. In order to prevent the degradation of the target compounds, Bligh and Dyer [87] modified Folch’s method by changing the solvent to chloroform/methanol/water with the ratio of 1∶2∶0.8 (v/v/v), so that the process could be executed in approximately 10 min. The Folch and Bligh-Dyer protocols were hitherto the most popular LLE approaches in lipids analysis based on chromatography-mass spectrometry [41,88].

However, the Folch and Bligh-Dyer methods suffer from drawbacks such as the need for highly toxic extraction solvents. Consequently, substitutes for chloroform, e. g., methylene chloride, were chosen due to their low toxicity and identical extraction yield as the chloroform/methanol/water system [89]. Chen et al. [36] utilized a methylene chloride/methanol/water mixture in 6∶3∶2 (v/v/v) ratio to extract lipids from mice plasma. A total of 284 lipids species in four categories, glycerolipids, glycerophospholipids, sphingolipids, and sterol lipids, were characterized. Lofgren et al. [90] also adopted a chloroform-free approach, utilizing butanol and methanol (BUME) in 3∶1 (v/v) ratio with LC-MS. With BUME, the abundant lipids and scarce lipids in 96 plasma samples were extracted automatically within 60 min.

Nonetheless, in the Folch and Bligh-Dyer methods, lipids are likely to be concentrated in the lower layer rather than the upper layer, which renders the collection inconvenient and leads to lipid loss. Matyash et al. [91] developed a new LLE approach using methyl-t-butyl ether (MTBE), methanol, and water in 10∶3∶2.5 (v/v/v) ratio. In this method, MTBE preferentially concentrates in the upper layer during phase separation because of its low density, thus making the collection faster and cleaner. Therefore, the MTBE method has gained popularity for the lipid profiling based on the chromatography-mass spectrometry. Zhang et al. [92] developed an MTBE method coupled with LC-MS to analyze the lipid metabolomics of mitochondria-damaged Saccharomyces cerevisiae after heavy ion beam radiation, and a total of 54 lipids were identified as metabolic markers.

However, some studies have reported that conventional LLE approaches are still the most suitable for lipid profiling under certain conditions. Ulmer et al. [93] investigated the Folch, Bligh-Dyer, and MTBE methods for human plasma-based lipidomics studies, and found that the former two methods give the highest extraction yields, with a sample to total solvent ratio of 1∶20 (v/v).

A three-phase liquid extraction (3PLE) method, which differs from the abovementioned extraction methods, was proposed by Vale G et al. [94]. The 3PLE involves the use of two organic phases and one aqueous phase to extract lipids from different biological samples such as bovine liver, human plasma, and mouse liver. The 3PLE method was successfully used to obtain 78 more lipids as compared with the Bligh-Dyer method, with reduced interfering substances as well as ion suppression in the direct-infusion workflow. Future development of LLE would focus on the use of green solvents and the enlargement of lipid coverage.

1.2 Single organic solvent extraction

SOSE is a sample preparation method that uses only one organic solvent, and consequently, is simpler and cheaper than both two-phase and three-phase liquid-extraction methods. Methanol is most frequently used in SOSE to extract the polar metabolites from different matrices. Zhao et al. [47] established an SOSE method coupled with LC-MS using methanol to extract lysophospholipids and phospholipids from human serum and plasma, which required a minimal amount of blood (10 μL) and gave high yields (>80.0%) over 100 lipids. SOSE with methanol coupled with LC-MS was also applied to the extraction of lipids from feces [52], bone tissue [44], and hepatocellular carcinoma and para-carcinoma tissue [95]. Furthermore, Lee et al. [48] used SOSE with methanol to extract polar lipids in sheep plasma, which was profiled by SFC-MS, and a total of seven polar lipids were successfully analyzed in only 7 min.

Isopropanol is the second most commonly used solvent in SOSE for lipid profiling. Christinat et al. [49] utilized SOSE with isopropanol for the extraction of nonesterified fatty acids in 96 human plasma samples. Merrill et al. [96] used an optimized nanoflow LC-MS technique coupled with SOSE to detect the lipids in neurons identified from live mammalian brains, and obtained more than 40 lipid species.

In addition, ethanol could be used as a low-toxicity SOSE solvent for lipid study. Ludovici et al. [97] found that ethanol gave an extraction yield comparable to that obtained with methanol during the extraction of lipids from stratum corneum samples using SOSE coupled with LC-MS.

Although SOSE is a very simple method, its poor ability to extract non-polar components has restricted its application in lipid profiling. For example, diacylglycerols and triacylglycerols must be extracted with other organic solvents such as chloroform in two-phase LLE procedures [24].

2 Extraction methods based on solid phase

In the past few years, the extraction methods based on solid phase such as SPE and SPME with special sorbents have been reported for use in lipid profiling. The advantages and limitations of these methods are outlined in Table 1.

2.1 Solid-phase extraction

SPE, which is based on selective solid-phase adsorption, has recently become popular in lipid extraction as it allows for significant reduction of matrix effects [98]. Depending on the type of packing material, commercial SPE columns are classified as reversed-phase, normal-phase, and ion-exchange columns [24]. The use of the appropriate SPE column in lipid analysis helps suppress matrix interferences, thereby improving the detection sensitivity and recovery of low-abundance constituents as well as expanding the detection coverage of lipid species.

reversed-phase columns, typically C8and C18columns, are the most popular for the separation of lipids, including eicosanoids [99], polyunsaturated fatty acids [54], and oxylipins [55]. Wang et al. [99] comprehensively profiled 184 eicosanoids in human plasma using a Strata-X reversed-phase SPE column coupled with LC-MS/MS. With preconcentration of the SPE column, the limits of quantification (LOQs) were less than 10.0 pg for 86.0% of the analytes, and the matrix effects were in the range of 70.4%-113% for all analytes. In our previous study, SPE with two kinds of reversed-phase columns was compared with the modified Folch method to extract the endogenous lipid mediators in ischemic stroke rats. All of the extraction recoveries obtained by SPE were higher than that of modified Folch method [58]. Based on these results, we adopted a Strata-X SPE approach followed by SFC-MS for the untargeted lipidomics study of the protective effects of isosteviol sodium on stroke rats [57]. Fifteen differential lipids that presented significant differences between the sham group and the model group were screened and identified within the 9-min profiling.

Normal-phase columns for lipid profiling are usually packed with silica [100] and applied to the separation of neutral and polar lipids, e. g., phospholipids [56]. Lopez-Bascon et al. [59] used a normal-phase column packed with silica to extract the polar lipids in human visceral adipose tissue by LC-MS/MS; the detected glycerophospholipid species increased by 50% compared with that in the traditional LLE. Ferreiro-vera et al. [56] developed an SPE method using zirconia-bonded silica particles to extract the phospholipids in human serum by LC-MS/MS and obtained 2 667 molecular features, which was 1.4-fold as the LLE strategy.

Ion-exchange columns used for lipid analysis are typically packed with an aminopropyl stationary phase and are suitable for the separation of neutral lipids and polar lipids such as free fatty acids [101], phospholipids [102], and sphingolipids [103]. Schlotterbeck et al. [101] profiled the free fatty acids in marine algae utilizing a 3-aminopropyl silica ion-exchange column prior to LC-MS/MS. The limits of detection (LODs) and LOQs were 0.230 ng/mL and 1.00 ng/mL, respectively. Zhang et al. [103] developed a lipidomics platform based on LC-MS for the profiling of phospholipids and sphingolipids in brain tissues using an aminopropyl ion-exchange column. More than 300 lipid species were quantified with recoveries in the range of 80.2%-108%.

2.2 Solid-phase microextraction

SPME, introduced by Arthur and Pawliszyn in 1990, is a sample preparation approach that makes use of fused silica fibers coated with an appropriate stationery phase, which can capture and enrich lipids in complex biological samples [104]. SPME efficiently reduces matrix effects [105] and shows great flexibility for the analysis of various samples such as gas, liquid, and solid samples. However, the application scope of SPME is impeded by its small extraction capacity. This method is usually coupled with LC or LC-MS as well as GC or GC-MS.

Much effort has been undertaken to explore new coating materials and thus extend the application of SPME to lipid profiling. Cha et al. [61] employed laboratory-made sol-gel derived fibers with butylmethacrylate/hydroxy-terminated silicone oil coating coupled with GC-MS for the detection of fatty acids in lung tissue. The recoveries ranged from 76.4% to 107%, with relative standard deviations (RSDs) of precision less than 13.3%. Birjandi et al. [60] used a biocompatible C18-polyacrylonitrile 96-blade SPME prior to LC-MS for the untargeted lipidomics study of cell line cultures. The SPME technique yielded 77 differential lipid species, comparable to the peer list obtained by the Bligh-Dyer method. The matrix effects for all the detected lipids varied from 80.0% to 120%, with RSDs of precision between 5.00% and 18.0%.

3 Field-assisted extraction methods

Field effects, such as pressure, heat, ultrasound, and microwave energy, may increase the mass transfer rate, thereby improving the extraction yield [106]. Recently, SFE, MAE, PLE, and UAE based on different field effects have been found to be favorable for the profiling of lipids due to the enhanced extraction efficiency.

3.1 Supercritical fluid extraction

As a green extraction approach, SFE has been widely adopted for the separation of lipids from diverse matrices, generally using supercritical carbon dioxide (scCO2) as the extraction solvent [107]. In comparison with the conventional LLE method, SFE has noteworthy advantages such as high extraction efficiency, less organic solvent consumption, and the realization of solvent-free extracts by decreasing the pressure [28,108]. Furthermore, the inertness and low critical temperature (31.06 ℃) of scCO2prevent latent degradation of the unstable compounds due to high temperature and atmospheric oxygen. Nevertheless, the low polarity of scCO2could result in inefficient extraction of polar molecules. Luckily, this issue could be resolved by adding a polar solvent, i. e., a modifier [109]. Subsequently, the SFE extracts are determined by GC-MS, LC-MS, and SFC-MS.

GC-MS has the most abundant standard databases befitting the identification of unknown SFE extracts. Esquivel-Hernandez et al. [73] utilized SFE prior to GC-MS for separating functional lipophilic compounds fromArthrospiraplatensisbiomass. The yields of tocopherols and fatty acids by SFE were twice those by MAE extraction. Zanqui [66] et al. profiledβ-tocopherol and phytosterols from flaxseed oil by SFE coupled with GC-MS. The oil obtained by SFE was up to 5 times more resistant to lipid oxidation as compared with that extracted by the conventional solvent extraction methods. In addition, SFE coupled with GC-MS was widely used to analyze other bioactive compounds in the manufacture of foodstuffs and pharmaceuticals, e. g., the extraction and identification of non-polar lipids and pigments fromNannochloropsisgaditanabiomass [110], unsaturated fatty acids fromMoringaoleifera[111], and turmeric oil from Curcuma Longa herb [63].

SFE coupled with LC-MS is suitable for the profiling of the involatile lipids, thus making up for the limited utility of SFE-GC-MS. Matsubara et al. [62] employed SFE coupled with LC-MS/MS for the analysis of hydrophobic metabolites in dried blood spots. A total of 144 lipid species extracted by SFE were detected, while only 127 lipids extracted by the traditional organic solvent extraction were analyzed. The glucosinolates, phenolic compounds, and unsaturated fatty acids in Eruca sativa were extracted by SFE with water prior to LC-MS, and the extracts showed significant activity in antioxidant tests [64].

Since the polarity of scCO2is similar to that of lipids, SFC-MS shows excellent separation ability for the profiling of lipids extracted by SFE. Jumaah et al. [65] applied SFE to the profiling of the lipids from bilberry followed by SFC-MS. The amounts of fatty acid methyl esters and total lipids obtained by SFE were equivalent to those in the case of the conventional Bligh-Dyer extraction, and a much lesser amount of organic solvent was consumed by SFE.

3.2 Pressurized liquid extraction

PLE, which was first reported by Richter et al. [112], combines elevated temperature and pressure with liquid solvents, and is thus both time-saving and solvent-saving. When using low-polarity extraction solvents such as dichloromethane/methanol solution (2∶1, v/v), PLE showed comparable efficiency as SFE for the extraction of fatty acid and triacylglycerols in chia seeds [71].

Isaac et al. [70] used PLE coupled with LC-MS to analyze phosphatidylcholine and sphingomyelin molecular species in human brain, and the obtained yields twice those in the case of the Folch method with the same extraction solvent. PLE was also applied to the extraction of phosphatidylethanolamine molecular species from four different food matrices (soy, egg yolk, ox liver, and krill oil), and 20 different phosphatidylethanolamines were detected in total [68].

In 2014, Golmakani et al. [69] developed a green PLE method using biodegradable solvents (limonene and ethanol) to replace the traditional toxic solvents for the extraction of fatty acids from diverse microorganisms in the ocean. Under the optimized conditions, the extraction was completed in 15 min, with extraction yields and recoveries comparable to those of the traditional methods.

PLE could also be coupled with other sample preparation methods. Del Pilar Sanchez-Camargo et al. [110] utilized PLE coupled with SFC to extract fatty acids fromNannochloropsisgaditanaand then determined these by GC-MS. Compared with SFE, this combined method gave improved yields, with reduced waste generation and solvent consumption. Compared with traditional methods like LLE and SOSE, PLE was not widely applied to the extraction of lipids because of the sophisticated instrument and the high temperature during the extraction which may cause the degradation of some unstable lipids.

3.3 Microwave-assisted extraction

MAE is a fast and efficient sample preparation approach, which has been adopted diffusely in lipid profiling [109]. The thermal effects of the ionic conduction and dipole rotation caused by microwave energy lead to accelerated extraction [30], making MAE better suited for the extraction of thermostable molecules. In MAE, the target molecules are selectively heated depending on the difference in the microwave absorption capacities of the lipids and extraction solvents; hence, MAE is a more time-saving, more effective, and less solvent-consuming strategy compared with the traditional LLE [113].

Safder et al. [72] used MAE coupled with GC-MS to analyze lipids from spent hen tissue, and obtained over 95.0% of the lipids within 10 min. Functional lipophilic compounds inArthrospiraplatensiswere extracted by MAE prior to GC-MS, and the carotenoid yields obtained by MAE were twice those by the SFE approach [73]. Pomegranate seeds were pretreated by MAE and then extracted by SFE to obtain seed oil, which was determined by GC-MS [114]. Even a 2-min microwave pretreatment increased the yield of the oil from 21.6% to 25.5%.

3.4 Ultrasound-assisted extraction

Compared with conventional extraction methods such as LLE, UAE is more efficient for lipid analysis [18]. Liu et al. [74] utilized UAE coupled with GC-MS to determine 16 free fatty acids from the liver samples of mice, with recoveries ranging from 87.0% to 120%. Pizarro et al. [75] developed a lipid profiling method based on UAE and LC-MS using MTBE as the extraction solvent for human plasma samples. With this method, the extraction time was reduced to one-half of that in the previously reported Bligh-Dyer method. Recoveries higher than 70.0% were obtained for all evaluated lipids, with RSDs below 5.50%. Moreover, the proposed method enabled the detection of more than 800 different features. Zhang et al. [76] used UAE prior to GC-MS for the extraction of lipids in two kinds of oleaginous microorganisms. This ultrasonication chloroform/methanol extraction allowed for the recovery of total lipids in 15 min at 25 ℃, whereas the conventional chloroform/methanol extraction required 12 h at 60 ℃. Ultrasonication, which could dramatically reduce the extraction time, was promising for the high-throughput analysis of lipids.

4 Online coupling methods

Online coupling of sample preparation techniques with chromatography and MS has always been a hotspot and has recently attracted increasing attention in the field of lipid analysis [115]. Online hyphenation of sample preparation techniques with chromatography and MS offers many obvious advantages such as automation, high throughput, and minimum artificial errors or sample loss; further, this strategy is time-saving [80]. Automation of online hyphenation also improves the accuracy, precision, and sensitivity of the analysis. In addition, closed online coupling can minimize the oxidative decomposition of unstable lipids and possible contamination [32]. Currently, SPE-LC-MS and SFE-SFC-MS are the two major online coupling configurations for lipid profiling.

Owning to the dominance of LC-MS in separation and detection science, SPE hyphenated with LC-MS is most commonly used for lipid analysis. An online SPE-LC-MS system constructed by Luque De Castro et al. [116] was efficaciously used to profile epoxyeicosatrienosic acids in human serum, with LODs and LOQs of less than 0.150 ng/mL and 0.500 ng/mL, respectively. Besides, the RSDs of repeatability ranged from 2.50% to 9.90% and the recoveries were in the range of 87.0%-96.0%. The same research group also applied the online SPE-LC-MS system to the identification of eicosanoid inflammation biomarkers [77], prostanoids [117], omega-6-derived eicosanoids [118], and hydroxyeicosatetraenoic acids [78] in human serum with satisfactory detection limits. Besides the excellent detection sensitivity, the online SPE-LC-MS system was greatly time-saving. This system required only 7.1 min for the pretreatment and quantification of prostaglandin E-2, D-2 and thromboxane B-2 in cell samples [80]. Similarly, only 6.5 min was required to simultaneously quantify 26 hydroxylated polyunsaturated fatty acids in cell and plasma [79].

Since both SFE and SFC use a supercritical fluid as the extraction solvent and the mobile phase, solvent incompatibility can be avoided in online SFE-SFC-MS during the extraction and analysis of lipids. Uchikata et al. [119] proposed an online SFE-SFC-MS/MS system for phospholipid profiling in dried plasma spots. This high-throughput system was able to simultaneously extract and separate 134 phospholipids with the LODs in the range of 0.402-11.8 pmol; among these, 74 phospholipids were analyzed with good repeatability. Recently, a commercial online SFE-SFC-MS system, the Nexera-UC system from Shimadzu Corporation, appeared on the market. Zoccali et al. [82] employed the Nexera-UC system for the direct online extraction and determination of targeted carotenoids from red Habanero peppers. Compared with the traditional solid/liquid extraction, which required a few hours, the online system extracted and identified 21 targeted analytes within 17 min, with the extraction yields ranging between 37.4% and 65.4%. Subsequently, the Nexera-UC system was applied to the detection of carotenoids and apocarotenoids in intact human blood without any preliminary treatment [81]. The recoveries varied from 98.0% to 109%, and the LODs and LOQs were in the range of 2.00-10.0 ng/mL and 8.00-40.0 ng/mL, respectively.

Despite the higher instrument cost and more sophisticated equipment, the outstanding merits of online coupling systems cannot be neglected; hence, more variations of hyphenated techniques are expected to be applied to lipid profiling in the future.

5 Sample preparation methods for in vivo lipid profiling

Invivoanalysis is the ideal methodology to study the transformation and composition of lipids [120], and it allows for real-time monitoring of the physiological reactions of lipids. Appropriate sample preparation techniques forinvivolipid analysis are of great significance in physiopathology and clinical applications [121].

Bessonneau et al. [83] developed a fastinvivoSPME-LC-MS/MS method to monitor rapid changes in blood concentrations of arachidonic acid and 12-hydroxyeicosatetraenoic acid after lipopolysaccharide-induced inflammation in Sprague-Dawley rats. During the experiment, the SPME fiber was inserted into the catheters of the rats for sampling. The blood concentrations of the target eicosanoids obtained by the proposedinvivoSPME method were significantly correlated with those obtained by the conventional protein precipitation method. Thus,invivoSPME is a suitable alternative to current methods based on sample depletion for accurate monitoring of the rapid concentration changes of lipids in blood.

6 Conclusions and trends

Sample preparation is a crucial procedure in the analysis of lipids. Many improvements have been made to the extraction methods based on liquid phase, which are the most commonly used sample preparation techniques for lipid analysis, in order to achieve low toxicity and convenience. The extraction methods based on solid phase have also seen rapid development because of the use of new selective sorbents. Field-assisted extraction methods aided by pressure, heat, microwaves, and ultrasound are more time-saving and efficient as compared with the traditional extraction methods.

The wide variety of lipid species and the complicated matrices of biological samples pose several challenges in sample preparation. There has always been a trade-off between lipid coverage and matrix effects in lipidomics study. It must be emphasized that appropriate pretreatment approaches should be chosen based on the lipid profiling issues (targeted lipid profiling or untargeted lipidomics study), specific sample under investigation, and subsequent analytical method.

In the future, online coupling sample preparation techniques would be promising for lipid analysis due to their significant advantages such as automation, time-saving features, low solvent consumption, and minimal sample loss. Meanwhile, more environment-friendly techniques are also expected to be developed. Besides, sample preparation methods based oninvivoanalysis for the real-time monitoring of the physiological reactions of lipids, which are of great significant in physiopathology and clinical applications, are still scarce. In conclusion, much effort remains to be undertaken toward further developments in sample preparation technology for lipid profiling, in order to understand biological processes and disease mechanisms as well as to explore new drugs.