1. Introduction

Obesity remains one of the pressing problems of modern society, therefore, studies of the mechanisms underlying its occurrence and the therapies used to treat it continue to be relevant. Depending on the hypothesis, a wide range of research methods can be used, ranging from purely assessing biochemical parameters to deep psychological research. However, research usually involves standard procedures such as measuring body mass index (BMI) or fat content in human subjects. Among the biochemical parameters, those that undergo the greatest changes with obesity (lipid profile, fasting glucose, insulin, etc.) are first examined.

A large amount of accumulated data on obesity allows for their meta-analysis and underlies a large number of systematic and retrospective reviews [1]. In particular, studies related to bariatric surgery are a powerful source of data because these types of surgery involve altering the stomach, intestines, or both to induce weight loss ^{[2][3][4][5][6][7]}[2,3,4,5,6,7]. In addition, a large number of studies of obesity are associated with cardiovascular diseases ^{[8][9][10][11]}[8,9,10,11], diabetes and metabolic syndrome ^[12][13][14][12,13,14]; many studies are in the field of diets, psychology and neurology ^{[15][16][17][18][19][20][21][22][23]}[15,16,17,18,19,20,21,22,23].

While the number of parameters in the study of obesity itself is limited only by the imagination of scientists and the equipment available in the laboratory ^{[24][25][26][27][28][29]}[24,25,26,27,28,29], the study of obesity in connection with other diseases is strictly subordinate to the study area and is often limited to several parameters, such as BMI and total fat content. ^{[30][31][32][33]}[30,31,32,33]. In addition, determining the diagnosis of obesity is always primary in that work; for this purpose, the most commonly used method is the calculation of BMI. The World Health Organization used BMI to categorize humans into underweight (< 18.5), normal weight (18.5–24.9), overweight (25–29.9) and obese (BMI ≥ 30) categories [34]. Since BMI may not be a good indicator of obesity for bodybuilders and other groups of athletes ^[35][36][35,36], body fat ^[37][38][39][37,38,39] and total body water ^[40][41][40,41] can be determined in these groups, as can concomitant states of lipid alterations in blood (dyslipidaemia) [42], hyperinsulinemia [43], etc. Since obesity is directly related to lipid metabolism, it is interesting to study not only standard plasma parameters but also alterations in specific lipid groups in serum [44]. However, due to the much higher lipid accumulation in obesity (Figure 1) and many alterations in the lipid composition, the methods used for sample preparations for lipidomic studies in samples from obese subjects sometimes have to be modified.

Figure 1. Lipid alterations in obesity. Lipids in red are elevated in obesity, and lipids in green are reduced. BCAA—branched chain amino acids; BCFA—branched chain fatty acids; DAG—diacylglycerols; FFA—free fatty acids; HDL—high density lipoproteins; LDL—low density lipoproteins; LPP—lipid peroxidation products; MUFA—monounsaturated fatty acids; OCFA—odd chain fatty acids; SFA—saturated fatty acids; PUFA—polyunsaturated fatty acids; TAG—triacyclglycerols.

In the study of obesity, determining triglyceride (TG) and cholesterol levels is of great clinical importance. Basic blood test results for total cholesterol (TC), TG and cholesterol in lipoprotein fractions (low density lipoproteins (LDL) and high density lipoproteins (HDL)) should be considered together. To study these indicators and related and often required for standard clinical practice indicators (C-reactive protein, glucose, insulin levels, etc.), there are many standard methods and their modifications that make these analyses routine in clinical practice.

2. Methods of Sample Preparation for Lipidomic Studies

2.1. Sample Collection and Storage

Fat metabolism disorders are detected by determining the lipid spectrum of the blood. Blood for a study is taken from a vein, always on an empty stomach (12–14 h after eating); otherwise, the results of the study are distorted, since 1-4 h after eating, alimentary hyperlipaemia occurs [45]. During blood sampling, adverse events such as haemolysis, coagulation, and platelet activation should be avoided, but the class of anticoagulants used should also be taken into account since calcium-chelating coagulants (ethylenediaminetetraacetic acid (EDTA) and citrate) can cause the calcium-dependent formation or degradation of certain classes of lipids ex vivo [46].

Different classes of lipids are subject to different changes during storage. Long-term storage of plasma at room temperature (RT) leads to an increase in lysophosphatidylethanolamines (LPE), lysophosphatidylcholines (LPC) and fatty acids (FAs), while phosphatidylethanolamines (PE) and phosphatidylcholines (PC) decrease, which suggests the breakdown of ester bonds in these phospholipids [47]. Avoiding freeze-thaw cycles is no less important because with their increase, the number of lipid metabolites decreases significantly [48].

2.2. Pre-Extraction Additives

Additives used during or before extraction serve a variety of purposes. Internal standards are a measure of extraction efficiency. In many cases, lipidomic studies of obesity are accompanied by the determination of obesity-associated hormone levels (such as ghrelin, obestatin, glucagon, leptin, and adiponectin); therefore, protease inhibitor cocktails are added to serum/plasma samples to increase hormone stability [49]. In addition, various detergents serve to facilitate cell destruction during homogenization, and buffers are used to maintain a stable pH. The most commonly added substances to prevent oxidative processes during extraction are antioxidants and radical scavengers such as butylated hydroxytoluene (BHT) ^[50][51][50,51]. This is especially important when studying unstable compounds such as oxylipins ^[52][53][54][52,53,54], which are the metabolites of polyunsaturated fatty acids.

2.3. Sample Stability

It is generally assumed that lipids are highly stable at RT, while it is advisable to not allow them to overheat during homogenization and to prevent oxidation by the addition of antioxidants [55]. Despite this, many studies consider the effects of storage conditions, the number of freeze/thaw cycles and the behaviour of organic compounds in experimental conditions. Jiang et al. validated a liquid chromatography-tandem mass spectrometry (LC-MS/MS) method for the determination of ceramides (Cer) in human plasma and determined the stability of each analyte at low- and high-quality control concentrations under long-term storage (39 days at -80 °C), freeze/thawing (five times), tabletop mode (14 h at RT before sample extraction) and autosampler conditions (3 days). The results showed that Cer (22:0) and Cer (24:0) were stable in human plasma under all conditions [56]. Ferreiro-Vera et al. assessed the stability of eicosanoids in serum under experimental conditions; every hour for 8 h, they analysed samples spiked with eicosanoids, and no significant differences in analyte concentrations were found [57]. Zeng and Cao also showed sufficient stability of short-chain FAs (SCFAs) and ketone body derivatives during autosampler storage (5 °C; 48 h), after 2 h at RT and after three freeze/thaw cycles [58]. Klawitter et al. showed that freeze/thaw cycles and long-term storage of plasma (6 h; RT) should be avoided to prevent changes in the composition of lipid classes of very low-density lipoprotein (VLDL) (loss of cholesterol esters and phospholipids), while free fatty acid (FFA) concentrations did not change under the same conditions [59]. Oxylipins are especially unstable in this regard, and improper collection and storage of samples can lead both to a significant decrease in their level and to an increase in their content due to enzymatic and non-enzymatic oxidation [60]. Some oxylipins (resolvins and prostanoids) are unstable even at −20 °C [61], so the manufacturers of their standards recommend storing them at −80 °C, while the concentration of prostaglandins can significantly decrease with prolonged storage at even −80 °C [62].

2.4. Extraction Methods

2.4.1. Protein Precipitation

Protein precipitation (PPT) is used to remove protein from samples, therefore, when carrying out PPT during sample preparation, it is important that the chosen solvent causes protein denaturation and, at the same time, is a good solvent for lipids [55]. In addition, precipitation of proteins that make up a large volume of the analysed matrix is necessary since some groups of lipids are present in the matrix in trace amounts. This helps to minimize the risk of lack of detection or misidentification and to release protein-bound compounds prior to target lipid extraction [63]. Most often, PPT is preceded by subsequent solid-phase extraction (SPE) and liquid-liquid extraction (LLE).

2.4.2. Liquid–Liquid Extraction

The high solubility of the hydrocarbon chains of lipids in organic solvents allows the use of LLE for the separation of lipids in various immiscible liquids. Widely used methods such as those of Folch ^[64][67] and Bligh and Dyer ^[65][68] have the drawback of using toxic solvents ^[66][69]; in addition, some classes of lipids (for example, lysophospholipids (LPL)) can remain in the aqueous phase ^[67][70]; however, many proposed modifications of these methods can overcome the above disadvantages, and these methods are still widely used in lipidomics of obesity samples ^{[67][68][69][70][71][72][73]}[70,71,72,73,74,75,76]. Methyl tert-butyl ether (MTBE) extraction, which has been popular recently, is undergoing various modifications and shows very good efficiency over classical methods ^[66][69]. In the study of obesity, MTBE extraction is used to isolate lipids from liver tissue ^[74][77], skeletal muscle ^[75][78], adipose tissue ^[76][79] and plasma ^[50][69][50,72].

2.4.3. Solid-Phase Extraction

SPE is more suitable than LLE for target lipidomics because it allows fractionation of specific lipid classes after LLE ^[77][78][80,81]. Therefore, in a lipidomics study, SPE is resorted to when it is necessary to isolate specific lipid groups or species that are present in the sample in a small amount, such as eicosanoids [57], LPL ^[67][70], oxidized phospholipids ^[72][75], serum sterols ^[79][82] oxysterols, endocannabinoids, and Cer ^[80][83], non-esterified FA and oxylipins ^{[53][81][82][83]}[53,84,85,86]. Due to the wide variety of SPE protocols and commercially available SPE columns, there are studies in which these parameters are compared, for example, in studies of fatty acid esters of hydroxy fatty acids (FAHFAs) in serum ^[84][87] or oxylipins in human plasma ^[85][86][88,89]. Additionally, SPE helps to separate lipids in complex matrices with a large lipid abundance, such as adipose tissue ^[80][87][83,90] or brain tissue ^[77][80].

2.4.4. Other Extraction Methods

In addition to the well-established routine extraction methods described above, such as LLE, SPE, and PPT, also more modern but at the same time rarer extraction methods are used in the studies of obesity, such as solid-phase microextraction (SPME), stir bar sorptive extraction (SBSE), dispersive liquid–liquid microextraction (DLLME) and their variants. The main disadvantage of the above-listed solvent extraction is the use of organic solvents that have such disadvantages as toxicity and harmfulness to the environment. In addition, they must be of high purity, which increases the cost of analysis ^[88][105]. However, SPME, SBSE, DLLME are a solvent-free sample preparation method that is easy to use, does not require preliminary sample preparation, and is easily automated ^[89][94].

The most widely used technique of above mentioned is SPME. In combination with gas chromatography-mass spectrometry (GC-MS) it can be used not only for analysis of volatile organic compounds, but also for the extraction of fatty acids and fatty acid esters from solid tissues and biofluids, which requires a very small sample volume and reduces the matrix effect ^[90][106]. Although SPME can be used for lipidomics studies, in obesity studies these methods are also used to study non-lipid compounds. SPME followed by GC/MS was used to analyze aroma compound headspace release from extra virgin olive oil after the interaction of saliva in obese and overweight individuals ^[91][107], to evaluate volatile organic compounds of gut microbiota of obese patients ^[92][93][108,109], and for urinary volatile organic compounds profiling in overweight children ^[94][110].

The SBSE method, like SPME, is a method of sample preparation without the use of solvents and with the use of a solid sorbent for preliminary concentration of the analyte before analysis. The surface area of the sorbing polymer is greater in SBSE than in SPME ^[95][111]. Eslami et al. used SBSE followed by HPLC for quantification of ghrelin in human plasma ^[96][112].

The DLLME method is based on the rapid mixing of dispersing and extraction solvents with an aqueous sample, resulting in the formation of an emulsion consisting of fine particles of the extraction solvent dispersed in the aqueous phase, then the solvent is separated from the sample by centrifugation ^[97][113]. Amin et al. used DLLME following GC/MS method for the evaluation of urinary Bisphenol A in obese subjects ^[98][114]. Krawczyńska et al. applied DLLME technique for the determination of vitamin D in obese patients plasma ^[99][115].

Thus, the relatively small number of studies in lipidomics using above methods is explained by their recent appearance, while such advantages as relative easiness of implementation, accuracy, small sample volume and lack of organic solvents make these extraction methods promising.