Introduction

Lysophospholipids (LysoPLs) are deacylated products of phospholipids with single fatty acid chain and are produced by the action of phospholipase A (PLA1 or PLA2). Several types of LysoPLs were identified in biological samples such as plasma, with a change in their head groups, including lysophosphatidylcholine (LysoPC), lysophosphatidylethanolamine (LysoPE), lysophosphatidylinositol (LysoPI), lysophosphatidylserine (LysoPS), lysophosphatidylglycerol (LysoPG), and lysophosphatidic acid (LysoPA). Their biological activities are considered to be significantly different depending on their chemical nature as for example LysoPA is known to act as a lipid mediator [1, 2]. LysoPLs in plasma are generally bound to albumin or lipoproteins [3]. Numerous studies have suggested that changes in plasma levels of LysoPLs are linked to several diseases such as ovarian cancer, acute kidney injury, non-alcoholic fatty liver disease (NAFLD), and obesity [4,5,6,7]. LysoPC is the most abundant LysoPL in human plasma, and its levels are reported to be about several hundred micromoles [8] and it is known to involve in several biological functions such as vascular endothelial dilation, cellular immune response, and insulin secretion [9,10,11]. However, the in vivo roles and physiological functions of LysoPEs [12] were not well studied. Several in vitro studies demonstrate that the LysoPEs are a neurotrophic activator via activation of mitogen-activated protein kinase (MAPK) signaling [13], increase intracellular calcium levels, and stimulate chemotactic migration and cellular invasion in cancers [14]. Clinical studies also demonstrate the decreased LysoPE levels in serum samples of migraine patients [15]. Despite the significant biological importance of LysoPEs, most of the studies were focused on measurement of their relative concentrations [15, 16]. Although semi-quantitative analysis is useful as comprehensive one, it has fundamental drawbacks such as the inability to obtain an accurate concentration and less sensitivity. So far, an absolute quantitative method for LysoPE determination is limited, due to the lack of authentic standards and optimal internal standard (IS) for LC-MS/MS.

Herein, we chemically synthesized the authentic LysoPEs (LysoPE 18:2, 20:4, 20:5, and 22:6) and an odd chain IS (LysoPE 15:0), which are not commercially available and aimed to develop an absolute quantitative method for LysoPE species by a simple liquid-liquid extraction followed by highly sensitive targeted LC-MS/MS analysis. Seven LysoPE molecular species prepared in this study are major molecular species in previous studies [17, 18]. Furthermore, we employed our method as a clinical application to profile the concentration of LysoPEs in the serum from patients with non-alcoholic fatty liver disease (NAFLD) and healthy subjects. NAFLD is a progressive hepatic disease characterized by massive triglyceride (TG) accumulation in the liver and can range from simple steatosis (SS) to non-alcoholic steatohepatitis (NASH) [19].

Materials and methods

Materials

1,2-Dipalmitoyl-sn-glycero-3-phosphoethanolamine (DPPE), trityl bromide, cis-4,7,10,13,16,19-docosahexaenoic acid (DHA), linoleic acid, pentadecanoic acid, and arachidonic acid were purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). Dehydrated solvents (dichloromethane, pyridine, and N,N-dimethylformamide [DMF]), triethylamine, and sodium methoxide were purchased from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan). Chloroform, methanol (MeOH), and oxalyl dichloride were purchased from Kanto Chemical Co., Inc. (Tokyo, Japan). Eicosapentaenoic acid (EPA) was purchased from Cayman Chemicals (MI, USA). Acetic acid (AcOH) was purchased from Nacalai Tesque, Inc. (Kyoto, Japan).

The LysoPE species analyzed in the present study included 1-pentadecanoyl-2-hydroxy-sn-glycero-3-phosphoethanolamine (LysoPE 15:0), 1-palmitoyl-2-hydroxy-sn-glycero-3-phosphoethanolamine (LysoPE 16:0), 1-stearoyl-2-hydroxy-sn-glycero-3-phosphoethanolamine (LysoPE 18:0), 1-oleoyl-2-hydroxy-sn-glycero-3-phosphoethanolamine (LysoPE 18:1), 1-linoleoyl-2-hydroxy-sn-glycero-3-phosphoethanolamine (LysoPE 18:2), 1-arachidonyl-2-hydroxy-sn-glycero-3-phosphoethanolamine (LysoPE 20:4), 1-eicosapentaenoyl-2-hydroxy-sn-glycero-3-phosphoethanolamine (LysoPE 20:5), and 1-docosahexaenoyl-2-hydroxy-sn-glycero-3-phosphoethanolamine (LysoPE 22:6). Among them, LysoPE 16:0, 18:0, and 18:1 were purchased from Avanti Polar Lipids, Inc. (AL, USA), while the other LysoPE species were synthesized in our laboratory following the method below.

General procedure for synthesis of LysoPEs

The synthetic route for the preparation of LysoPEs is shown in Fig. 1. The LysoPE standards were synthesized by a method earlier established in our laboratory with minor modifications [17, 20]. Briefly, the protection of an amine moiety of DPPE with trityl bromide gives compound 2 trityl 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine (Tr-DPPE) and subsequent deacylation using sodium methoxide gives compound 3 trityl sn-glycerol-phosphoethanolamine (Tr-GPEA) with 69% yield [17]. The compounds 4a–e trityl lysophosphatidylethanolamine (Tr-LysoPE 15:0, 18:2, 20:4, 20:5, and 22:6) were obtained by direct acylation of Tr-GPEA using oxalyl dichloride and pyridine instead of dibutyltin oxide (Bu2SnO) and isopropanol (iPrOH), which often gives poor yield or no reaction [21]. In brief, the respective fatty acids (15:0, 18:2, 20:4, 20:5, and 22:6) were activated by converting them into acid chlorides by adding oxalyl chloride (2 eq) in DCM:DMF (1000:1) and stirring at room temperature for 3 h. The activated acid chlorides were re-dissolved in DCM having 0.1 M pyridine and reacted with Tr-GPEA for an overnight at 0 °C-room temperature. The reaction was monitored by thin-layer chromatography, and the reaction mixture is extracted with chloroform and washed with 1 N hydrochloric acid and a brine solution. The chloroform layer was dried in a vacuum, and respective Tr-LysoPEs were purified by silica gel column chromatography with a yield about 21–55%. Finally, the compounds 5a–e (LysoPE 15:0, 18:2, 20:4, 20:5, and 22:6) were obtained by the deprotection of trityl group using 90% aqueous AcOH [17]. The IS, compound 5e, LysoPE 15:0 was synthesized for the first time in this study. The detailed experimental protocols, high-resolution masses, and their spectroscopic characterizations (1H and 13C-nuclear magnetic resonance [NMR]) were provided in the Electronic Supplementary Material (ESM, Figs. S1–S7).

Fig. 1
figure 1

Synthetic scheme for each LysoPE standard and internal standard (IS) [17]. Reagents for the process are shown above the arrow, and solvents are shown below. The yield is shown in parentheses. Compound 5e (R = pentadecanoic acid) is IS. AcOH, acetic acid; (COCl)2, oxalyl chloride; DCM, dichloromethane; DHA, docosahexaenoic acid; DMF, N,N-dimethylformamide; DPPE, 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine; EPA, eicosapentaenoic acid; LysoPE, lysophosphatidylethanolamine; Tr-DPPE, trityl 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine; Tr-GPEA, trityl sn-glycerol-phosphoethanolamine; Tr-LysoPE, trityl lysophosphatidylethanolamine

Full size image

Sample collections

In order to obtain serum samples from healthy subjects, blood samples were collected from 8 young healthy volunteers and are stored at − 80 °C. The serum from patients with NAFLD including SS (n = 9) and NASH (n = 27) are obtained from Okayama University Hospital and are stored at − 80 °C prior to analysis. The NAFLD patients were diagnosed histologically by liver biopsy using the NAS score and Brunt classification. The details about the classification of the serum from NAFLD patients are provided in our previous paper [22]. Clinical parameters of both healthy subjects and NAFLD patients are provided in Table 1.

Table 1 Clinical parameters of serum samples. aValues are expressed as mean ± SD; N.D. means not determined in this study; n.s. means not significant; SS simple steatosis, NASH non-alcoholic steatohepatitis, TC total cholesterol, TG triglyceride, PL phospholipid, LDL-C low-density lipoprotein cholesterol, HDL-C high-density cholesterol, ApoAI apoprotein AI, ApoB apoprotein B, AST aspartate aminotransferase, ALT alanine aminotransferase, γ-GT gamma-glutamyl transferase
Full size table

Lipid extraction

Lipid extraction was performed according to the Bligh and Dyer method with minor modifications [23]. Briefly, the serum sample (10 μL) was added into 400 μL of MeOH (with 62.5 nmol/mL of IS [LysoPE 15:0]). Then, 400 μL of chloroform and 190 μL of Milli-Q were added into the mixture and vortexed at 3500 rpm for 5 min (MSV-3500, Funakoshi Co., Ltd., Tokyo, Japan). After centrifugation at 9000 rpm, 4 °C for 5 min (CF-15R, Hitachi Co., Ltd., Tokyo, Japan), the chloroform layer was collected and dried during re-extraction, and the upper aqueous layer was re-extracted with 400 μL of chloroform. Next, combined organic layer was dried under vacuum for 30 min with a centrifugal concentrator (CC-105, Tomy Seiko Inc., Tokyo, Japan). The obtained lipid extract was re-dissolved in 200 μL of MeOH, centrifuged at 15,000 rpm, 4 °C for 15 min, to remove any residues, and transferred to LC-MS vials.

LC-MS/MS conditions

The high-performance liquid chromatography (HPLC) system (Thermo Fisher Scientific Inc., MA, USA) equipped with a Hypersil GOLD column (50 × 2.1 mm, 1.9 μm, Thermo Fisher Scientific Inc., MA, USA) was used for separation of LysoPE. The oven temperature was maintained at 35 °C, the sample tray was kept at 4 °C, and the injection volume was set at 5 μL. The mobile phase consisted of 5 mmol/L aqueous ammonium acetate (A), iPrOH (B), and MeOH (C) at a flow rate of 200 μL/min. The following gradient elution was applied: the first 0.5 min, 20% A, 40% B, and 40% C; 0.5–5.0 min, gradually changed to 10% A, 40% B, and 50% C; then 5.0–5.5 min, 5% A, 50% B, and 45% C, and kept this ratio until 13.0 min; the final 13.0–13.5 min, returned to the initial ratio, and kept until 15.0 min for re-equilibration.

The TSQ Quantum Access mass spectrometer (Thermo Fisher Scientific, Inc., MA, USA) was used in electrospray ionization (ESI) negative ion mode. The capillary temperature was 200 °C. The spray voltage was 4000 V. The sheath gas (nitrogen) pressure was 30 psi, the ion sweep gas (nitrogen) pressure was 4 psi, and the aux gas (nitrogen) pressure was 35 psi. The selected reaction monitoring (SRM) by collision-induced dissociation (CID) was optimized for each analyte, and the detailed parameters are listed in Table 2.

Table 2 MS/MS conditions. Electrospray ionization (ESI) negative ion mode
Full size table

HPLC-Orbitrap MS conditions

The semi-quantitative profiling of phosphatidylethanolamine (PE) for confirming its changes was carried out according to our previous method [24] by using an HPLC-Orbitrap MS (Thermo Fisher Scientific, Inc.). Chromatographic separation was conducted by a T3 C18 column (2 × 150, 3 μm, Waters, Milford, MA) with the mobile phase consisted of 10 mM aqueous ammonium acetate (A), iPrOH (B), and MeOH (C) following the optimized gradient elution. High-resolution MS was utilized for detection, with ranges of m/z 150–1100 and 220–1650 for ESI positive mode, while MS/MS fragmentation was acquired by collision-induced dissociation in ion trap mode. The raw data were processed by Xcalibur 2.3 (Thermo Fisher Scientific, Inc.) within the mass tolerance of 5.0 ppm. The identification of fatty acyl composition was executed with the LIPIDMAPS database [25].

Method validation

A series of diluted LysoPE standard solutions was prepared, with the concentrations 0.005, 0.01, 0.05, 0.25, and 2.0 pmol/μL for each standard. The limits of quantification (LOQ) and the limits of detection (LOD) were evaluated as the signal-to-noise ratio (S/N) as 10 and 3, respectively (except for LysoPE 20:5, of which the S/N for LOQ was 4). Reproducibility was tested using serum from a healthy subject in intraday assay for 10 replicates and interday assay for 3 days. The recovery test was conducted by adding three levels of LysoPE species into the serum from a healthy subject with n = 6 replicates.

Statistical analysis

GraphPad Prism 7.0e (GraphPad Software Inc., La Jolla, CA, USA) was used as the software for all statistical analyses. The one-way ANOVA test followed by the Tukey-Kramer test for multiple comparison was applied in all clinical studies except the analysis of the female/male ratio in clinical parameters. The chi-square test was applied in the analysis of the female/male ratio in clinical parameters. P < 0.05 was considered to be statistically significant. All results were expressed as means ± standard deviation (SD).

Results

Chemical synthesis of authentic LysoPEs and their characterizations

The LysoPE authentic standards which are commercially unavailable were synthesized in our laboratory by the method earlier established in our laboratory with minor modifications. In particular, LysoPE standards such as LysoPE 18:2 (5a), LysoPE 20:4 (5b), LysoPE 20:5 (5c), LysoPE 22:6 (5d), and LysoPE 15:0 (5e: IS) are obtained in a four-step reaction starting from commercially available DPPE (1). The protection of an amine group of DPPE was achieved with trityl bromide to give Tr-DPPE (2) with 90% yield, and further the fatty acyls of 2 were hydrolyzed by sodium methoxide to give Tr-GPEA with 69% yield. To obtain the required LysoPEs, desired fatty acids of each molecular species were activated by converting them to acyl chlorides using carbonyl dichloride in pyridine and re-acylated to Tr-GPEA to yield respective Tr-LysoPEs 4a–e with a yield (21–53%). Earlier, Bu2SnO and iPrOH were used to prepare Tr-LysoPE, but in this study, we found that use of carbonyl dichloride improved the reaction yield. Finally, deprotection of trityl moiety of Tr-LysoPEs using aqueous acetic acid gives the LysoPEs (5a–e) with a 30–97% yield. All the prepared standards were characterized by their high-resolution masses and structural confirmation was carried out by 1H and 13C-NMR. The detailed NMR assignments are provided in the ESM (Figs. S1–S7).

Method validation

Linearity and sensitivity

Under the optimized conditions, the linearity of each LysoPE was evaluated (shown in Table 3). For all the LysoPE analytes, the R2 were more than 0.9972, showing good linearity. The LOQ for all the LysoPE ranged from 0.001 to 0.005 pmol/μL and the LOD ranged from 0.0005 to 0.0033 pmol/μL, showing sufficient sensitivity.

Table 3 Analysis of LOD and LOQ. aLOQ limits of quantification, bLOD limits of detection
Full size table

Reproducibility and recovery

The results of reproducibility and recovery tests are listed in Table 4. The coefficients of variation (CVs) of intraday and interday assays for all LysoPE species were within 12.6% and 9.6%, respectively, while the recoveries for all the LysoPE ranged from 87.3 to 111.1%, of CVs within 6.4%.

Table 4 Result of reproducibility and recovery test. aValues are expressed as mean ± SD; bCV coefficient of variation
Full size table

Freezing and thawing test and stability test

In order to check the stability of serum samples, the serum from n = 3 healthy subjects were frozen and thawed three times within 3 days for the freezing and thawing test, and measured 1 week, 4 weeks, and 8 weeks after serum collection for the stability test. During the freezing and thawing test, there was no significant influence on the LysoPE concentration (ESM Table S1). And in the stability test, between week 0 and week 8, there was also no significant influence on it (ESM Table S2).

LysoPE variation among healthy subject, SS, and NASH sera

By using the developed method, we acquired the LysoPE molecular species in serum from healthy subject, SS, and NASH groups. Although, we have synthesized LysoPEs of sn-1 type and performed analysis of serum, we have not distinguished sn-1 or sn-2 in the samples. Hence, the analysis results are shown by omitting the appropriate stereochemistry because LysoPEs of sn-1 and sn-2 types are quickly converted to each other and LysoPEs of sn-1 type account for almost all LysoPEs in the plasma [18]. The LysoPE profile in healthy subjects showed the highest concentration of LysoPE 18:2 of 4.485 nmol/mL, followed by LysoPE 22:6 of 3.249 nmol/mL and LysoPE 20:4 of 3.020 nmol/mL (Fig. 2). In addition, both SS and NASH significantly decreased in LysoPE molecular species except for LysoPE 18:0 when compared with healthy subjects. However, LysoPE 18:0 concentrations have been shown to have no significant difference between the three groups.

Fig. 2
figure 2

Quantitation of LysoPE molecular species in human serum samples. Concentrations of seven molecular species of LysoPEs and combined LysoPEs in the serum from healthy subject, SS, and NASH patients. LysoPE 20:4 in the serum from both patients were undetected. *P < 0.05; **P < 0.01; ***P < 0.001. n.s. means not significant. LysoPE, lysophosphatidylethanolamine; SS, simple steatosis; NASH, non-alcoholic steatohepatitis

Full size image

Combined LysoPE concentration in serum was found to be 18.030 nmol/mL in healthy subjects, 4.867 nmol/mL in SS, and 5.497 nmol/mL in NASH (Fig. 2). A significant difference between healthy subjects and SS or NASH was observed, though there was no significant difference on combined LysoPE concentration between SS and NASH groups of any molecular species.

PE variation among healthy subject, SS, and NASH sera by using HPLC-Orbitrap MS

We measured PE species, as they are the precursors of LysoPEs in serum from healthy subjects, SS, and NASH using HPLC-Orbitrap MS, and their semi-quantitative profiles are shown in Fig. 3. PEs having polyunsaturated fatty acyl chain in serum from SS and NASH were significantly lower than those of the healthy subjects.

Fig. 3
figure 3

Profiling of PEs in serum from healthy subject, SS, and NASH patients by using HPLC-Orbitrap MS. *P < 0.05; **P < 0.01; ***P < 0.001. SS, simple steatosis; NASH, non-alcoholic steatohepatitis; PEs, phosphatidylethanolamines

Full size image

Discussion

The phospholipids are the most abundant class of membrane lipids, which undergo hydrolysis by the action of PLA1 or PLA2 to produce LysoPEs. Previous studies have demonstrated the involvement of LysoPEs in cell signaling process, which act as a neurotrophic activator [13], increase the intracellular calcium levels [14], and are biomarkers for migraine [15]. Most of the reported LysoPE measurement methods are semi-quantitative and less sensitive and often represent only the variation trends. To know the physiological concentration and significance of LysoPEs, absolute quantitation methods are necessary; LC-MS/MS with SRM methods are most promising as they are more sensitive and accurate. So far, absolute quantitative methods are limited in the literature due to the lack of authentic standards and hence, roles of LysoPEs in clinical science were a few. To overcome this problem in this study, we synthesized the authentic standards by a four-step chemical reaction along with a novel, odd uncommon LysoPE 15:0 molecular species as an internal standard. A simple oxalyl chloride–mediated acylation step was employed in this study to prepare the good yields of LysoPEs in the third step (Fig. 1), which is more efficient and simpler than the previous report involving Bu2SnO [17].

The data of LOQ ranges from 0.001 to 0.005 nmol/mL and LOD from 0.0005 to 0.0033 nmol/mL (Table 3) showed that our method is highly sensitive enough to detect the LysoPEs in serum samples and it is about 10 times more sensitive and requires less sample volume than in a previous report [18]. This sensitivity enables us to detect small amount of each LysoPE species in the serum from NAFLD patients. Further, the comprehensive method validation which was performed as demonstrated in Table 4 showed that the recovery rate is > 90% for most of the LysoPEs, with a low coefficient of variation, and suggests the high precision of our methodology. Previously, the values of LysoPE were evaluated by a semi-quantitative method or a method using inappropriate internal standard, but not an absolute quantitative one as shown in the present study. To overcome this problem, we have completed a novel unnatural IS (LysoPE 15:0) and seven standards and established a more accurate and absolute measurement method for LysoPEs using sensitive targeted LC-MS/MS.

The results demonstrated in Fig. 2 showed that most of the LysoPE molecular species serum concentrations were decreased in NAFLD (including SS and NASH) compared to those in healthy controls and these results are consistent with those in a previous report which showed the decreased total LysoPE levels in patients with NAFLD [6]. To our knowledge, this is the first report to profile individual molecular species variations in the serum from NAFLD patients. However, the reason behind the decreased serum concentration of LysoPEs and molecular mechanisms involving it remains unknown and needs to be elucidated. There are two possible reasons for lowering serum LysoPE in NAFLD patients. The first reason for lowering serum LysoPE might be lysophospholipid acyltransferase. It has been reported that the hepatic levels of lysophosphatidylcholine acyltransferase (LPCAT 1, 2, 3, and 4), which is a specific enzyme to metabolize LysoPC to phosphatidylcholine, were increased with an increase of inflammation cytokines, tumor necrosis factor-α, and transforming growth factor-β1, thereby leading to reductions in the amount of LysoPC in the serum [26]. Similar to these enzymes, LPCAT3, 4 and lysophosphatidylethanolamine acyltransferase 1 (LPEAT 1) were identified as specific lysophospholipid acyltransferase to metabolize LysoPE to PE in the liver [27]. Because both alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels were over normal ranges in SS and NASH patients recruited in this study, the liver inflammation might induce the increase of the expression of LPCAT3, 4 and LPEAT 1 followed by the decreases of combined LysoPE levels in serum. However, acyltransferases were not measured in this study, so it may be necessary to demonstrate them in mice in the future since the human liver is difficult to obtain.

The second reason for lowering serum LysoPE might be the decrease of PEs in the serum. Semi-quantitative analysis results of PEs demonstrated the decreased levels of PE enriched in linoleic acyl, arachidonic acyl, and DHA (e.g., PE 34:2 [16:0/18:2], PE 36:2 [18:0/18:2], PE 38:4 [18:0/20:4], PE 38:5 [18:1/20:4], and PE 40:6 [18:0/22:6]) in the NAFLD group, whereas PE 34:0, PE 34:3, PE 36:1, and PE 36:2 were different between SS and NASH, which might be potential biomarkers for NASH (Fig. 3). And the decrease of serum PEs in NAFLD patients could be linked to the reduction of serum LysoPEs. It was considered that there are two possible reasons on the decreases of these PE molecular species. First, since polyunsaturated fatty acids are generally susceptible to oxidation [28], the PE may be degraded due to the oxidative stress toward its containing linoleic acid, arachidonic acid, and DHA. Second, since PE is mainly biosynthesized in the Kennedy pathway of the endoplasmic reticulum (ER) [29], total PE production might be possibly reduced by ER stress [30]. It has been reported that ER stress is induced in NAFLD along with cellular fat accumulation, followed by the decreased levels of PE [31]. And finally, LysoPEs in the serum from NAFLD might be decreased.

Thus, our proposed highly sensitive method could be useful for future research on the analysis of LysoPE species. Since the receptor for LysoPE is unknown and the LysoPE metabolism in the cell and the blood is unclear [32], it was difficult to explain the decrease of LysoPE levels both in SS and in NASH clearly.

Conclusion

We successfully established the method for absolute quantitation of seven major LysoPE species. The proposed method will be a powerful tool for obtaining the profile of the LysoPE species as well as the comparisons in the concentrations of LysoPE species. Our proposed method has some advantages such as small sample volumes, multiple species measurement, and high sensitivity and could contribute to elucidating metabolism of LysoPEs. The detailed molecular mechanisms behind the change in PE and LysoPE concentration in SS and NASH patients will be of our future interest.