A UPLC-MRM-MS method for comprehensive profiling of Amadori compound-modified phosphatidylethanolamines in human plasma

Abstract

Phosphatidylethanolamines (PEs) are targets of non-enzymatic glycation, a chemical process occurred between glucose and primary amine-containing biomolecules. As the early-stage non-enzymatic glycation products of PE, Amadori-PEs are implicated in the pathogenesis of various diseases. However, only a few Amadori-PE molecular species have been identified so far, a comprehensive profiling of these glycated PE species is needed to establish their roles in disease pathology. Herein, based on our previous work of using liquid chromatography coupled neutral loss scanning and product ion scanning tandem mass spectrometry (LC-NLS-MS and LC-PIS-MS) in tandem, we extend identification of Amadori-PE to the low abundant species, which is facilitated by using plasma lipids glycated in vitro. The confidence of identification is improved by high resolution tandem mass spectrometry and chromatographic retention time regression. A LC coupled multiple reaction monitoring mass spectrometry (LC-MRM-MS) assay is further developed for more sensitive quantitation of the Amadori compound modified lipids. Using synthesized stable isotope labeled Amadori lipids as internal standards, levels of 142 Amadori-PEs and 33 Amadori-LysoPEs are determined in the NIST human plasma standard reference material. These values may serve as important reference for future investigations of Amadori modified lipids in human diseases.

1. INTRODUCTION

Phosphatidylethanolamines (PEs), together with their partial hydrolysis products lysophosphatidylethanolamines (LPEs) are the second most abundant class of glycerophospholipids in human blood plasma [1]. Because of the primary amine in their ethanolamine head group, PEs and LPEs are targets of non-enzymatic glycation, also known as Maillard reaction, in which the carbonyl group of reducing sugars react with the primary amines to initially form an unstable Schiff base and then a more stable Amadori product after rearrangement [2]. Although most of the literature point to the importance of protein glycation in many physiological and pathological processes, such as normal aging and complications of diabetes mellitus, Amadori compound modified lipids are also found to play roles in vascular disease, diabetes, cancer, and in other diseases as well [3–8].

Despite a large number of PEs/LPEs listed in the LIPID MAPS database (http://www.lipidmaps.org), only a few Amadori-PEs/LPEs have been identified in biological samples [9–11]. To better understand the roles of Amadori-PE/LPE in physiological and pathological processes, it is essential to obtain an in-depth profiling of Amadori-PE/LPE species in human plasma.

We previously reported synthetic methods for isotopically labeled Amadori compound modified PE and LPE standards, optimized the extraction procedure for simultaneous enrichment of Amadori-PE and -LPE from human plasma, and confidently identified 20 Amadori-LPE and 62 Amadori-PE species using liquid chromatography tandem mass spectrometry in series, i.e., neutral loss scanning (NLS) followed by product ion scanning (PIS) [12, 13]. Considering the large number of ether linked PEs in human plasma, here in this work, we synthesized ¹³C-labeled, either linked Amadori-PE P-18:0/20:4 as internal standard for quantifying Amadori modified ether linked PEs. Using in vitro glycated human plasma lipid extract, we extended the strategy of performing NLS and PIS in tandem to confidently identify low abundant modified lipid species, which was accomplished together with high resolution mass spectrometry and chromatographic retention time regression. We further developed a LC coupled multiple reaction monitoring mass spectrometry (LC-MRM-MS) assay for more sensitive quantitation of the Amadori compound modified lipids in plasma. In view of the increasing popularity of SRM 1950 – the NIST standard reference human plasma in lipidomics community for quality control and platform standardization [14], we applied the LC-MRM-MS method and determined the levels of 142 Amadori-PEs and 33 Amadori-LPEs in SRM 1950 human plasma, by spiking in respective synthetic, isotope labeled Amadori-PEs as internal standard for the endogenous diacyl-, ether- and lyso-Amadori-PEs.

2. MATERIALS AND METHODS

2.1. Chemical, reagents and materials

1,2-dipentadecanoyl-sn-glycero-3-phosphoethanolamine (PE 15:0/15:0), 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (PE 16:0/18:1(9Z)), 1-(1Z-octadecenyl)-2-arachidonoyl-sn-glycero-3-phosphoethanolamine (PE P-18:0/20:4(5Z,8Z,11Z,14Z)), 1-tridecanoyl-sn-glycero-3-phosphoethanolamine (LPE 13:0/0:0) were purchased from Avanti Polar Lipids (Alabaster, AL, USA). All of them were >99% pure, except LPE 13:0/0:0 may contain up to 10% of sn-2 isomer. [U-¹³C₆]-D-glucose (99%) was from Cambridge Isotope Laboratories (Tewksbury, MA, USA). D-glucose and all solvents including methanol (MeOH), acetonitrile (ACN), isopropanol (IPA), chloroform (CHCl₃), and water (H₂O) of LC-MS grade and HPLC grade were purchased from Fisher Scientific (Pittsburgh, PA, USA). Pooled human plasma collected from healthy people with K2-EDTA as anticoagulant was purchased from BioIVT (Westbury, NY, Lot#: BRH1524451). The standard reference material—metabolites in frozen human plasma (NIST SRM-1950)—was purchased from the NIST (Gaithersburg, MD).

2.2. Standards preparation

Amadori-PE/LPE standards (Amadori-PE 15:0/15:0, -PE 16:0/18:1(9Z), -PE P-18:0/20:4(5Z,8Z,11Z,14Z), and -LPE 13:0/0:0) were synthesized by incubation of their respective substrates with D-glucose in anhydrous MeOH, as described previously with minor modifications [12, 13]. Optimized synthesis conditions are shown in Table 1. Lower temperature and shorter reaction time were used for the synthesis of Amdori-PE P-18:0/20:4 and Amadori-LPE 13:0/0:0 to decrease the yield of byproducts for more efficient downstream purification.

Table 1.

Experimental conditions for syntheses of Amadori-PE 15:0/15:0, -PE 16:0/18:1, -PE P-18:0/20:4, and -LPE 13:0/0:0 standards.

Target products	Concentration (mg/mL)	Concentration of D-glucose (mg/mL)	Reaction temperature (°C)	Reaction time (h)	Yield
Amadori-PE 15:0/15:0	1	40	60	72	56.1%
Amadori-PE 16:0/18:1	5	40	60	72	52.8%
Amadori-PE P-18:0/20:4	1.5	40	50	48	67.5%
Amadori-LPE 13:0/0:0	5	40	50	48	49.3%

Synthetic products were purified by duplicated preparative HPLC separation on a Shimadzu 20A HPLC system using a Kromasil C18 column (250 × 10 mm, 10 μm, AkzoNobel, Bohus, Sweden). Isocratic mobile phases of 100% and 80% MeOH both containing 5 mM ammonium formate were used for Amadori-PEs and Amadori-LPE, respectively, at a flow rate of 3 mL/min. Elution was monitored at 220 nm with a multiple wavelength UV detector.

The synthesis and purification of stable isotope-labeled Amadori-PE/LPE, i.e., [¹³C₆]Amadori-PE/LPE were similarly performed as described above with 20 mg/mL [U-¹³C₆]-D-glucose.

2.3. Lipids extraction from plasma samples

Ten times volume of MeOH containing 0.5 % acetic acid (v/v) (MeOH-AA) was added to human plasma samples. The samples were vortex mixed for ~20 s then centrifuged at 10,000 × g for 10 min to precipitate proteins. Supernatants were collected and nitrogen-dried.

2.4. Glycation of human plasma lipids in vitro

Pooled human plasma (7 mL) underwent MeOH-AA extraction described above. The resultant lipid extract was reconstituted in 0.5 mL of MeOH, mixed with 20 mg of D-glucose, and then reacted in a thermomixer at 50 °C for 48 h. The reaction product mixture was subjected to LC-MS/MS analysis for identification of Amadori-PE/LPE species.

2.5. Identification of Amadori-PE/LPE

Separation of various glycated lipid species was achieved on a core-shell Accucore C30 column (150×2.1 mm, 2.6 μm, ThermoFisher Scientific) using a Vanquish UPLC. Column oven temperature was 40 °C, injection volume was 5 μL. Elution conditions were the same as previously reported by our laboratory[12]. Briefly, the mobile phase was composed of solvent A (ACN:H₂O, 50:50, v/v) and B (IPA:ACN, 90:10, v/v), both containing 10 mM ammonium formate and 0.1% formic acid. The gradient was as follows: −3–0 min, 30% B for column equilibration; 0–5 min, 30–43% B; 5–5.1 min, 43–50% B; 5.1–14 min, 50–70% B; 14.1–23 min, 70–99% B; 23–26 min, 99% B; 26–26.1 min, 99–30% B; 26.1–30 min, 30% B for column re-equilibration. The total analysis time including column re-equilibration was 33 min. The flow rate of mobile phase was 350 μL/min.

The effluent from UPLC was detected by a TSQ Quantiva mass spectrometer (ThermoFisher Scientific) equipped with a heated electrospray ionization source and operated in neutral loss scan (NLS) and product ion scan (PIS) mode for species identification. Spray voltage was set at 3500 V in positive ion mode and 3000 V in negative ion mode. Sheath, auxiliary and sweep gases were set at 20, 7 and 1 (arbitrary units), respectively. Vaporizer and ion transfer tube temperatures were both 300 °C. Argon collision gas pressure was 1.5 mTorr. LC-NLS-MS was used to scan precursor ions having neutral loss of 303/309 Da in positive mode and 162/168 Da in negative mode after collision-induced dissociation for Amadori- and [¹³C₆]Amadori-PE/LPE species, respectively, with scan range m/z 550–1000 and collision energy 27 eV. For LC-PIS-MS, the optimized collision energy was 20 eV for Amadori-PE/LPE species in positive ion mode; while 35 eV for Amadori-PE and 27 eV for Amadori-LPE in negative ion mode.

A Q Exactive HF hybrid quadrupole-Orbitrap mass spectrometer (QEHF, Thermo Fisher Scientific) coupled to the aforementioned Vanquish UPLC, was also performed for high-resolution mass analysis. The QEHF ion source setting is the same as TSQ Quantiva. Full scan was used for mass accuracy analysis. Parallel reaction monitoring (PRM) was used to obtain fragment ion information of targeted precursor ions. Normalized collision energy was 15 for Amadori-PE and 18 for Amadori-LPE in positive ion mode; 18 for Amadori-PE and 19 for Amadori-LPE in negative ion mode.

2.6. Quantitation of Amadori-PE/LPE in human plasma

LC-MRM-MS analysis on Vanquish UPLC-TSQ Quantiva was performed for relative quantitation of Amadori-PE/LPE. Prior to establishment of the final MRM assay, the transitions were confirmed according to the following: in positive ion mode, transitions of [M+H]⁺ to [M+H-309]⁺ and [M+H-303]⁺ were used for ISTDs and endogenous analytes, and collision energy was 23 and 28 eV for Amadori-PE and -LPE, respectively; in negative ion mode, transition of [M-H]⁻ to [M-H-162]⁻ was used for confirmation purpose. Collision energy was 29 and 23 eV for Amadori-PE and -LPE, respectively. Scheduled MRM was applied with retention time window of 0.6 min.

Human plasma sample (200 μL) was spiked with 10 μL of ISTD mixture in MeOH containing 400 nM [¹³C₆]Amadori-PE 15:0/15:0, 400 nM [¹³C₆]Amadori-PE P-18:0/20:4, and 160 nM [¹³C₆]Amadori-LPE 13:0/0:0 to make the final concentrations of each standard in plasma at 20 nM, 20 nM and 8 nM, respectively. Then spiked plasma samples underwent MeOH-AA extraction described in Section 2.3. The resultant residues were reconstituted in 50 μL MeOH, followed by centrifugation at 10,000 × g for 10 min. Clear supernatants of 5 μL were subjected to LC-MRM-MS analysis in positive ion mode. Quantification was performed based on the ratio between peak areas of endogenously identified modified lipids and their corresponding ISTDs, the resulted concentration data were used directly for statistical analysis. All measurements were performed in quadruplicate.

2.7. Method validation

Method validation was performed on lower limit of detection (LLOD), lower limit of quantitation (LLOQ), precision, accuracy, processing recovery, and matrix effect. Briefly, serial dilutions of [¹³C₆]Amadori-PE 15:0/15:0, -PE P-18:0/20:4, and -LPE 13:0/0:0 standards were prepared and subjected to LC-MRM-MS analysis. LLOD and LLOQ were defined as the concentration of analyte with signal-to-noise ratio (S/N) of 3 and 10, respectively. The accuracies and intra-day (n = 4) and inter-day (n = 3) precisions were investigated using pooled human plasma spiked with Amadori-PE 15:0/15:0 and -LPE 13:0/0:0 at low (LQC, 1 nM), middle (MQC, 5 nM), and high (HQC, 25 nM) concentrations, and their corresponding ISTDs at 20 nM for [¹³C₆]Amadori-PE 15:0/15:0 and 8 nM for [¹³C₆]Amadori-LPE 13:0/0:0. Accuracy was expressed as recovery (%) of QC standards (the ratio of measured value over expected value). Precision was expressed as percent coefficient of variance (CV%, standard deviation divided by the mean).

To evaluate the processing recovery (PR%) and matrix effect (ME%) of the developed method, pooled human plasma and pooled human plasma spiked with ISTDs of 3 concentrations (final concentrations in spiked plasma: 1 nM, low; 5nM, medium; and 25 nM, high) were subjected to sample preparation in parallel (pre-spiked samples). Post-spiked samples were prepared by spiking ISTDs into pooled human plasma extract to make the final ISTD concentrations to be 1, 5, and 25 nM. ISTDs of 1, 5, and 25 nM were also prepared in MeOH as neat solution controls. Pre-spiked, post-spiked samples and neat solution controls were subjected to LC-MRM-MS analysis. Processing recoveries were calculated as peak areas of standards pre-spiked divided by peak areas of standards post-spiked; matrix effects were calculated as peak areas of standards post-spiked divided by peak areas of neat solutions.

3. RESULTS AND DISCUSSION

The complete workflow for this study, including standard synthesis, identification and verification of Amadori-PE/LPE species, quantitative method development and application, is shown in Figure 1.

Figure 1.

Workflow of the identification and quantification of Amadori compound modified PEs and LPEs in human plasma.

3.1. Synthesis of Amadori-PE P-18:0/20:4 and [¹³C₆]Amadori-PE P-18:0/20:4

The synthesis and characterization of Amadori-PE 15:0/15:0, -PE 16:0/18:1, -LPE 13:0/0:0 and their isotope labeled compounds have been reported previously [12, 13], only slight modifications were made in substrate concentration, reaction temperature and time (Table 1). In this study, Amadori-PE P-18:0/20:4 and [¹³C₆]Amadori-PE P-18:0/20:4 were newly synthesized and purified for using as internal standards.

As shown in Figure 2, the characteristic fragment ions matched perfectly with the fragmentation pattern reported previously. For example, cation at m/z 611.5 and anion at 750.5 reflected Amadori-PE characteristic neutral loss of head group in positive and loss of Amadori moiety in negative ion mode, respectively. In panel A and B, cations arising from neutral loss of 3H₂O + HCHO (m/z 830.5 and 835.5) and loss of glycerol backbone with sn-2 fatty acyl chain (m/z 554.4 and 560.4), and cations corresponding to glycerol backbone plus sn-2 fatty acyl group (m/z 361.3) and head group (m/z 304.1 and 310.1) could also be observed. In panel C and D, anions at m/z 303.3 and 464.4 were also observed, which reflects sn-2 fatty acid anion and loss of Amadori moiety plus sn-2 fatty acyl group as a ketene, respectively. Overall, the expected fragmentation patterns and the corresponding mass shifts observed unambiguously validated the identification of synthetic compounds.

Figure 2.

Representative product ion MS/MS spectra of synthetic (A) Amadori-PE P-18:0/20:4 and (B) [¹³C₆]Amadori-PE P-18:0/20:4 in positive mode, (C) Amadori-PE P-18:0/20:4 and (D) [¹³C₆]Amadori-PE P-18:0/20:4 in negative mode. Collision energy: 20 eV in positive mode and 35 eV in negative mode.

Amadori-PE P-18:0/20:4 exists naturally in human plasma, so it is essential to investigate the isotopic purity of synthesized [¹³C₆]Amadori-PE P-18:0/20:4 for using as ISTD in assay. Extracted ion chromatograms obtained on QEHF were integrated for each of the resolved isotopes related to [¹³C₆]Amadori-PE P-18:0/20:4. After subtracting the contribution from the non-fully labeled natural isotopes, peak area values of isotopes were used to calculate the isotopic enrichment and the overall isotopic purity (see Electronic Supplementary Material (ESM) Fig. S1). As a result, the synthesized compound was found to be 92.7% fully enriched (containing six labeled carbons), and the overall isotopic purities is 98.7%. This is in agreement with the 99% isotope purity of the starting material, [¹³C₆]D-glucose.

3.2. Identification of Amadori-PE/LPE species in human plasma

Identification of 20 Amadori-LPEs and 62 Amadori-PEs in human plasma has been reported previously [12], using a species-specific identification approach involving LC-NLS-MS and LC-PIS-MS performed in tandem. Briefly, in the first step, a list of candidate Amadori-PE/LPE species was achieved by LCNLS-MS, with precursor ions having neutral loss of 303 Da in positive mode and 162 Da in negative mode; in the second step, LC-PIS-MS was performed to obtain molecular fragment information of these candidate species, which helped to verify and determine specific structure features (subclass, carbon chain length, double bond number, fatty acyl sn-position). Detailed method of species identification using MS/MS spectra is described in ESM (Figs. S2 and S3), taking Am-LPE 0:0/18:2, Am-LPE 18:2/0:0, Am-PE 18:0/20:4, Am-PE P-18:0/20:4, Am-PE O-18:0/20:4 as examples. Additionally, chromatograms in MRM mode of these 4 representative Amadori-PE/LPE species are shown in ESM Fig. S4.

Since Amadori-modified products are resulted from early-stage non-enzymatic glycation, hypothetically, each PE/LPE should have its corresponding Amadori adduct in human plasma. It was estimated by Nakagawa et al. that about 0.05–0.10 mol% PEs exist in healthy human plasma in the form of Amadori adduct [9]. Accordingly, a large number of Amadori-PE/LPE species are expected to exist in too low a level to be identified. To identify these low abundant species more confidently, the volume of pooled human plasma used for lipid extraction and glycation in vitro was increased 7 times in this study, as described in Section 2.4, to boost the concentration of Amadori-PEs/LPEs for analysis. Meanwhile, high-resolution MS and MS/MS analysis was performed to help confirm the identification.

The LC-NLS-MS and LC-PIS-MS strategy is capable of identifying lower abundance species in in vitro glycated plasma lipid extract. For example, a compound with molecular weight (MW) of 899.6 Da and retention time at 14.2 min was identified as Amadori-PE P-17:0/20:4, which had not been included in species reported previously. As shown in Fig. 3, two characteristic neutral losses of 303 Da in positive mode and 162 Da in negative mode were observed. The retention time for this species is 14.2 min, which excludes the possibility of an Amadori-LPE, because Amadori-LPE and -PE are always eluted at 2–5.5 and 11–18 min, respectively, under the LC condition in our study. Negative MS/MS spectrum (Fig. 3C) only had one fatty acid anion (m/z 303), indicating that it has a 20:4 carbon chain. This molecular ion ruled out the possibility that the two fatty acyls having the same chain length, therefore the 20:4 fatty acyl must be on the sn-2 position. In positive MS/MS spectrum, two ions, m/z 361 arising from loss of glycerol backbone plus sn-2 fatty acyl chain and m/z 304 corresponding to the Amadori headgroup, were observed very clearly (Figure 3A), which indicated that this compound should be P- but not O-Amadori-PE. Given the subclass, MW, and sn-2 substituent, sn-1 was assigned as P-17:0, which was also verified by the ionat m/z 540 arising from loss of glycerol backbone with sn-2 fatty acyl chain (characteristic of sn-1 position for P-Amadori-PE).

Figure 3.

Representative MS/MS spectra of Amadori-PE P-17:0/20:4 present in in vitro glycated human plasma lipid extract, by product ion scan on Quantiva in (A) positive ion mode and (C) negative ion mode; by parallel reaction monitoring on QEHF in (B) positive ion mode and (D) negative ion mode. Positions of the double bonds are unknown and not depicted in the molecular structure.

It is notable that we didn’t find PE P-17:0/20:4 in LIPID MAPS (https://www.lipidmaps.org/resources/databases/lmsd/index.php). To verify all identifications, mass accuracy of each Amadori-PE/LPE was evaluated. This was achieved using a high-resolution Orbitrap mass spectrometer — QEHF. As listed in ESM Table S1, mass errors range from −4.9 to 1.5 ppm for all identified Amadori lipids, showing the excellent mass accuracies and confidence in identification for these modified lipids. LC-PRM-MS analysis on QEHF also provided fragment ion information of targeted precursor ions in both positive and negative modes, as further verification. As shown in Figure 3B and 3D, fragment ions of Amadori-PE P-17:0/20:4 matched well with calculated values (< 1.6 ppm). Collectively, Amadori-PE P-17:0/20:4 was confidently identified in this study despite its corresponding PE P-17:0/20:4 wasn’t included in LIPID MAPS.

Retention of lipid species on a reversed phase column is influenced by the fatty acyl chain length and the degree of unsaturation. It has been observed by others and us that a linear correlation exists between the retention time and the total carbon number and the degree of unsaturation in fatty acyls, especially for the molecular species within the same lipid class [15, 16]. This rule was applied to verify the Amadori-compound modified lipid species identified in this work. As shown in Figure 4, we plotted the retention time of identified Amadori-PE/LPE versus either the total carbon number of fatty acyls or the number of double bonds in two fatty acyl chains, and observed a relationship of second degree polynomial regression for each Amadori-PE/LPE species. The correlation R² in most cases are greater than 0.96, indicating that all of the Amadori-PEs/LPEs were correctly identified. In these plots, isomers with different fatty acyl sn-position or C=C position have slightly different retention time which is in agreement with previous report [16].

Figure 4.

Representative relationship between retention times of Amadori-LPE (A), Amadori-PE (diacyl (B), O-Amadori-PE (C), and P-Amadori-PE (D) species and the total carbon number of fatty acyls (upper) or the total number of double bonds in two fatty acyl chains (bottom). X, number of carbon; Y, number of double bond.

In summary, using the approach described above, 80 Amadori-PE and 13 Amadori-LPE species were newly identified in this study. Including the species reported previously, a total of 142 Amadori-PE and 33 Amadori-LPE species were identified, their retention times and mass measurement accuracies are listed in ESM Table S1.

3.3. Quantification of Amadori-PE/LPE in human plasma

In the present study, [¹³C₆]Amadori-PE 15:0/15:0, -PE P-18:0/20:4, and -LPE 13:0/0:0 were used as the ISTDs corresponding to diacyl Amadori-PE, ether Amadori-PE (P- and O-), and Amadori-LPE, respectively. In contrast, all Amadori-PEs (diacyl, P-, and O-) used [¹³C₆]Amadori-PE(15:0/15:0) as ISTD in our previous work [12]. [¹³C₆]Amadori-PE(P-18:0/20:4) was introduced because we found neutral loss of 303 Da (head group) in P- and O-Amadori-PE was not as facile as that in diacyl Amadori-PE [12], which caused the MRM intensity much lower for P- and O-Amadori-PEs than diacyl Amadori-PEs. As such, spiking of [¹³C₆]Amadori-PE(P-18:0/20:4) into the sample can determine levels of P- and O-Amadori-PEs in human plasma more accurately.

With respect to method validation, LLOD, LLOQ, precision, accuracy, processing recovery, and matrix effect were investigated. The LLOD were 15, 11, and 5 pM, and LLOQ were 51, 35, and 18 pM, for diacyl Amadori-PE, ether Amadori-PE (P- and O-), and Amadori-LPE, respectively. In term of accuracy, the best accuracies were achieved at HQC (25 nM) for Amadori-PE, and at MQC (5 nM) for Amadori-LPE, because HQC and MQC had the closest concentration to ISTD of Amadori-PE (20 nM) and Amadori-LPE (8 nM), respectively. However, the accuracy decreased when the concentrations of analytes deviated from ISTDs’ concentration, as shown in Table 2. Overall, accuracy was high when the ISTD concentration was close to QC, regardless of the concentration of the QC sample. The intra-day (n = 4) and inter-day (n = 3) precisions were also summarized in Table 2, with the values of 1.8–8.3% and 1.5–11.1%, respectively, confirming that the developed method was sufficiently precise. Processing recovery and matrix effect of Amadori-PE/LPE at LQC, MQC, and HQC were summarized in Table 3. For diacyl Amadori-PE, ether Amadori-PE, and Amadori-LPE, processing recoveries were 83.9–90.1%, 86.6–89.1%, and 84.0–90.8%, respectively; and matrix effects were 36.3–40.5%, 38.6–42.6%, and 98.8–103.7%, respectively. The large matrix effect observed in diacyl and ether Amadori-PE are likely due to the retention time of these species on RPLC where other unmodified lipid species are coeluting, which is not the case for Amadori LPEs as they elute very early (RT <5 min) in the chromatogram, ahead of all other lipid species.

Table 2.

Accuracies and intra-day and inter-day precisions of the method for quantitation of Amadori-PE/LPE.

		Diacyl Amadori-PE	Amadori-LPE
Accuracy (%)	LQC	82.2	91.7
	MQC	88.7	95.6
	HQC	105.2	106.7
Intra-day precision (%)	LQC	3.4	8.3
	MQC	5.4	2.2
	HQC	1.8	4.0
Inter-day precision (%)	LQC	10.7	11.1
	MQC	1.5	2.2
	HQC	3.7	3.1

Table 3.

Processing recovery and matrix effect of the method for quantitation of Amadori-PE/LPE.

		Diacyl Amadori-PE	Ether Amadori-PE	Amadori-LPE
Processing recovery (PR% ± SD)	LQC	86.7 ± 1.4	89.1 ± 3.5	84.0 ± 3.5
	MQC	83.9 ± 1.6	86.9 ± 2.3	86.5 ± 1.8
	HQC	90.1 ± 4.3	86.6 ± 2.0	90.8 ± 3.6
Matrix effect (ME% ± SD)	LQC	39.2 ± 1.2	38.6 ± 3.8	103.7 ± 4.2
	MQC	40.5 ± 1.5	38.9 ± 2.4	102.7 ± 2.1
	HQC	36.3 ± 3.6	42.6 ± 3.3	98.8 ± 3.8

The developed quantitative method was successfully applied to determine Amadori-PEs/LPEs in standard reference human plasma (NIST SRM-1950). The levels of Amadori-PEs/LPEs in NIST SRM-1950 were summarized in Table 4, with concentration ranging from 0.01 nM to 22.77 nM. Overall, we quantified 142 Amadori-PEs and 33 Amadori-LPEs with 12 out of them below the LOD of our method. The retention time shifts (CV%, n=4) were 0.01–0.37% for Amadori-LPEs and 0.002–0.16% for Amadori-PEs, with median CV of 0.15% and 0.05%, respectively. The median CV of all quantified Amadori-PEs/LPEs levels was below 9.35% in quadruplicate measurements, revealing the excellent reproducibility achieved by this method. It is of note that the levels of P- and O-Amadori-PE are much higher than we reported previously [12]. Although different plasma was used in our previous work which may account for the difference, another more probable reason is that the levels of P- and O-Amadori-PE were underestimated previously. This is because ether-Amadori-PEs had lower fragmentation efficiencies than the diacyl-Amdaroi-PEs as evidenced by their fragment ion spectra under the same collision energies, as a result, the concentrations of ether-Amadori-PEs in our previous work were lower when using the diacyl [¹³C₆]Amadori-PE(15:0/15:0) as ISTD for ether-Amadori-PEs.

Table 4.

The levels of 142 Amadori-PEs and 33 Amadori-LPEs quantified in the NIST SRM-1950 reference human plasma. All measurements were performed in quadruplicate.

Species	Subclass	Rt (min)	Concentration (nM)	CV% (n=4)	Non-glycated Concentration (μM)#
Am-LPE 0:0/14:0	Amadori-LPE	2.06	<LOD
Am-LPE 0:0/16:0	Amadori-LPE	3.19	0.25	3.8	0.91
Am-LPE 0:0/16:1	Amadori-LPE	2.25	<LOD
Am-LPE 0:0/17:0	Amadori-LPE	4.04	0.02	15.2
Am-LPE 0:0/18:0	Amadori-LPE	4.99	0.30	10.4	1.6
Am-LPE 0:0/18:1	Amadori-LPE	3.48	0.15	11.4	1.4
Am-LPE 0:0/18:2	Amadori-LPE	2.56	0.40	6.5	1.9
Am-LPE 0:0/18:3	Amadori-LPE	2.15	<LOD
Am-LPE 0:0/20:2	Amadori-LPE	3.87	<LOD
Am-LPE 0:0/20:3	Amadori-LPE	2.98	0.04	7.4	0.52
Am-LPE 0:0/20:4	Amadori-LPE	2.52	0.11	3.9	1.1
Am-LPE 0:0/20:5	Amadori-LPE	1.99	<LOD
Am-LPE 0:0/22:4	Amadori-LPE	3.49	0.01	15.7
Am-LPE 0:0/22:5 (1*)	Amadori-LPE	2.69	0.02	32.3
Am-LPE 0:0/22:5 (2)	Amadori-LPE	3.09	0.01	8.5
Am-LPE 0:0/22:6	Amadori-LPE	2.39	0.08	14.3	0.52
Am-LPE 14:0/0:0	Amadori-LPE	2.21	0.01	19.3
Am-LPE 16:0/0:0	Amadori-LPE	3.48	1.20	11.2	0.91
Am-LPE 16:1/0:0	Amadori-LPE	2.45	0.10	6.9
Am-LPE 17:0/0:0	Amadori-LPE	4.37	0.03	8.6
Am-LPE 18:0/0:0	Amadori-LPE	5.41	2.51	9.9	1.6
Am-LPE 18:1/0:0 (1)	Amadori-LPE	3.78	1.59	10.4	1.4
Am-LPE 18:1/0:0 (2)	Amadori-LPE	4.05	0.09	6.4	1.4
Am-LPE 18:2/0:0	Amadori-LPE	2.76	2.37	11.6	1.9
Am-LPE 18:3/0:0	Amadori-LPE	2.37	<LOD
Am-LPE 20:2/0:0	Amadori-LPE	4.17	0.01	13.9
Am-LPE 20:3/0:0	Amadori-LPE	3.22	0.11	10.3	0.52
Am-LPE 20:4/0:0	Amadori-LPE	2.70	0.72	11.9	1.1
Am-LPE 20:5/0:0	Amadori-LPE	2.09	0.04	16.8
Am-LPE 22:4/0:0	Amadori-LPE	3.76	0.05	10.2
Am-LPE 22:5/0:0 (1)	Amadori-LPE	2.90	0.07	8.5
Am-LPE 22:5/0:0 (2)	Amadori-LPE	3.30	0.06	10.9
Am-LPE 22:6/0:0	Amadori-LPE	2.55	0.34	12.1	0.52
Am-PE 15:0/18:2	diacyl Amadori-PE	12.16	0.05	15.0
Am-PE 15:0/22:6	diacyl Amadori-PE	11.66	0.05	12.4
Am-PE 16:0/18:0	diacyl Amadori-PE	15.43	0.26	4.0	1.6
Am-PE 16:0/18:1	diacyl Amadori-PE	14.14	1.46	11.1	1.2
Am-PE 16:0/18:2 (1)	diacyl Amadori-PE	13.01	2.21	8.2	2.2
Am-PE 16:0/18:2 (2)	diacyl Amadori-PE	13.22	0.39	8.9	2.2
Am-PE 16:0/18:2 (3)	diacyl Amadori-PE	13.60	0.02	9.5	2.2
Am-PE 16:0/18:3 (1)	diacyl Amadori-PE	11.97	0.24	6.6
Am-PE 16:0/18:3 (2)	diacyl Amadori-PE	12.19	0.04	1.8
Am-PE 16:0/20:3	diacyl Amadori-PE	13.36	0.98	10.6	2.4
Am-PE 16:0/20:4	diacyl Amadori-PE	12.86	2.19	9.1	3.1
Am-PE 16:0/20:5	diacyl Amadori-PE	11.83	0.33	4.4	0.26
Am-PE 16:0/22:4	diacyl Amadori-PE	13.89	0.66	10.5	8.1
Am-PE 16:0/22:6	diacyl Amadori-PE	12.51	5.25	12.7	3.2
Am-PE 16:1/16:0	diacyl Amadori-PE	12.64	0.51	13.5	0.34
Am-PE 16:1/18:1	diacyl Amadori-PE	12.86	0.03	12.5	2.2
Am-PE 16:1/18:2	diacyl Amadori-PE	11.55	0.16	9.6
Am-PE 16:1/20:4	diacyl Amadori-PE	11.39	0.09	9.7	0.26
Am-PE 17:0/18:1	diacyl Amadori-PE	14.86	0.15	12.8
Am-PE 17:0/18:2 (1)	diacyl Amadori-PE	13.52	0.07	3.6	0.93
Am-PE 17:0/18:2 (2)	diacyl Amadori-PE	13.76	0.31	8.8	0.93
Am-PE 17:0/20:4	diacyl Amadori-PE	13.61	0.34	6.7	0.94
Am-PE 18:0/18:1	diacyl Amadori-PE	15.58	2.96	8.5	1.3
Am-PE 18:0/18:2	diacyl Amadori-PE	14.50	5.82	6.3	6.7
Am-PE 18:0/18:3 (1)	diacyl Amadori-PE	13.58	0.20	9.1	2.4
Am-PE 18:0/18:3 (2)	diacyl Amadori-PE	13.83	0.06	4.7	2.4
Am-PE 18:0/20:1	diacyl Amadori-PE	16.72	0.07	4.7	2.6
Am-PE 18:0/20:2	diacyl Amadori-PE	15.77	0.07	8.0	1.9
Am-PE 18:0/20:3 (1)	diacyl Amadori-PE	14.92	1.65	8.7	0.95
Am-PE 18:0/20:3 (2)	diacyl Amadori-PE	15.38	0.11	8.0	0.95
Am-PE 18:0/20:4	diacyl Amadori-PE	14.42	11.47	8.1	8.1
Am-PE 18:0/20:5	diacyl Amadori-PE	13.46	1.15	11.1	2.7
Am-PE 18:0/22:4	diacyl Amadori-PE	15.36	1.22	10.3	0.26
Am-PE 18:0/22:5 (1)	diacyl Amadori-PE	14.47	0.33	13.0	0.73
Am-PE 18:0/22:5 (2)	diacyl Amadori-PE	14.97	0.68	10.2	0.73
Am-PE 18:0/22:6	diacyl Amadori-PE	14.09	1.03	11.6	1.8
Am-PE 18:1/18:1	diacyl Amadori-PE	14.27	1.78	9.0	6.7
Am-PE 18:1/18:2	diacyl Amadori-PE	13.13	3.63	10.0	2.4
Am-PE 18:1/20:4 (1)	diacyl Amadori-PE	13.00	0.55	9.8	2.7
Am-PE 18:1/20:4 (2)	diacyl Amadori-PE	13.21	2.26	7.1	2.7
Am-PE 18:1/20:5	diacyl Amadori-PE	12.00	0.12	7.3	3.2
Am-PE 18:1/22:5 (1)	diacyl Amadori-PE	13.14	0.32	4.9	1.8
Am-PE 18:1/22:5 (2)	diacyl Amadori-PE	13.57	0.11	10.9	1.8
Am-PE 18:1/22:6 (1)	diacyl Amadori-PE	12.65	1.11	13.5	0.77
Am-PE 18:1/22:6 (2)	diacyl Amadori-PE	12.88	0.03	14.0	0.77
Am-PE 18:2/18:2	diacyl Amadori-PE	11.89	2.51	3.6	3.1
Am-PE 18:2/20:4	diacyl Amadori-PE	11.77	0.69	5.5	3.2
Am-PE 18:3/18:2	diacyl Amadori-PE	10.90	0.06	3.5	0.26
Am-PE 19:0/18:2	diacyl Amadori-PE	15.30	0.09	8.9
Am-PE 19:0/20:4	diacyl Amadori-PE	15.14	0.07	3.0
Am-PE 20:0/18:2	diacyl Amadori-PE	15.98	0.08	3.6	1.9
Am-PE 20:0/20:4	diacyl Amadori-PE	15.84	0.07	8.3	0.26
Am-PE 20:1/18:1	diacyl Amadori-PE	15.59	0.02	6.0	1.9
Am-PE 22:6/17:0	diacyl Amadori-PE	13.28	0.12	9.4	1.3
Am-PE O-16:0/18:1	O-Amadori-PE	15.06	0.58	11.7	0.46
Am-PE O-16:0/18:2	O-Amadori-PE	13.97	0.33	9.2	0.78
Am-PE O-16:0/18:3	O-Amadori-PE	13.02	<LOD		1.5
Am-PE O-16:0/22:4	O-Amadori-PE	14.83	0.87	8.5	0.94
Am-PE O-16:0/22:5	O-Amadori-PE	13.90	2.44	9.0	5.8
Am-PE O-16:0/22:6	O-Amadori-PE	13.46	1.25	10.2	4.9
Am-PE O-17:0/20:4	O-Amadori-PE	14.61	0.07	8.7
Am-PE O-18:0/16:0	O-Amadori-PE	16.33	0.11	6.7
Am-PE O-18:0/18:1	O-Amadori-PE	16.45	0.36	9.0
Am-PE O-18:0/18:2	O-Amadori-PE	15.50	0.46	12.8	0.93
Am-PE O-18:0/18:3	O-Amadori-PE	14.43	<LOD		3.2
Am-PE O-18:0/20:4	O-Amadori-PE	15.33	3.69	10.5	0.94
Am-PE O-18:0/22:5 (1)	O-Amadori-PE	15.40	0.90	8.7	0.73
Am-PE O-18:0/22:5 (2)	O-Amadori-PE	15.87	0.14	10.0	0.73
Am-PE O-18:0/22:6	O-Amadori-PE	15.02	1.66	9.0	1.8
Am-PE O-18:1/18:1	O-Amadori-PE	15.15	0.11	3.0	0.93
Am-PE O-18:1/18:2	O-Amadori-PE	14.10	<LOD		3.2
Am-PE O-18:1/20:1	O-Amadori-PE	16.37	0.06	5.0
Am-PE O-20:0/18:2	O-Amadori-PE	16.81	0.06	5.0
Am-PE O-20:0/20:4 (1)	O-Amadori-PE	16.24	0.49	10.8
Am-PE O-20:0/20:4 (2)	O-Amadori-PE	16.67	0.16	9.0
Am-PE O-20:0/22:5	O-Amadori-PE	16.73	0.04	6.6
Am-PE P-16:0/16:0	P-Amadori-PE	14.71	0.06	17.8
Am-PE P-16:0/18:1	P-Amadori-PE	14.84	1.44	8.9	0.78
Am-PE P-16:0/18:2	P-Amadori-PE	13.74	2.87	12.2	1.5
Am-PE P-16:0/18:3	P-Amadori-PE	12.67	0.02	15.4
Am-PE P-16:0/20:3 (1)	P-Amadori-PE	14.16	0.62	5.8	1.6
Am-PE P-16:0/20:3 (2)	P-Amadori-PE	14.31	<LOD		1.6
Am-PE P-16:0/20:4	P-Amadori-PE	13.56	9.61	10.4	4.9
Am-PE P-16:0/20:5	P-Amadori-PE	12.59	0.54	13.9	0.7
Am-PE P-16:0/22:4	P-Amadori-PE	14.61	0.90	6.9	5.8
Am-PE P-16:0/22:5	P-Amadori-PE	14.18	1.23	7.1	4.9
Am-PE P-16:0/22:6	P-Amadori-PE	13.19	3.92	10.8	3.5
Am-PE P-17:0/18:1 (1)	P-Amadori-PE	15.31	0.04	22.0
Am-PE P-17:0/18:1 (2)	P-Amadori-PE	15.59	0.04	7.3
Am-PE P-17:0/20:4 (1)	P-Amadori-PE	14.01	0.09	12.7
Am-PE P-17:0/20:4 (2)	P-Amadori-PE	14.33	0.27	11.3
Am-PE P-17:1/18:1 (1)	P-Amadori-PE	14.24	0.08	12.2
Am-PE P-17:1/18:1 (2)	P-Amadori-PE	14.58	0.04	21.6
Am-PE P-18:0/16:0	P-Amadori-PE	16.14	0.11	11.3	0.46
Am-PE P-18:0/18:1 (1)	P-Amadori-PE	16.26	1.64	8.8	0.93
Am-PE P-18:0/18:1 (2)	P-Amadori-PE	16.50	0.03	8.1	0.93
Am-PE P-18:0/18:2	P-Amadori-PE	15.26	7.24	9.9	3.2
Am-PE P-18:0/20:1	P-Amadori-PE	17.28	0.03	4.6
Am-PE P-18:0/20:3 (1)	P-Amadori-PE	15.63	0.48	4.4	0.94
Am-PE P-18:0/20:3 (2)	P-Amadori-PE	15.78	0.22	7.2	0.94
Am-PE P-18:0/20:3 (3)	P-Amadori-PE	16.05	0.06	7.0	0.94
Am-PE P-18:0/20:4	P-Amadori-PE	15.08	22.77	9.5	5.8
Am-PE P-18:0/22:4	P-Amadori-PE	16.02	1.03	4.7	0.73
Am-PE P-18:0/22:5 (1)	P-Amadori-PE	13.99	0.20	8.3	1.3
Am-PE P-18:0/22:5 (2)	P-Amadori-PE	15.17	1.66	7.3	1.3
Am-PE P-18:0/22:5 (3)	P-Amadori-PE	15.64	0.22	7.0	1.3
Am-PE P-18:0/22:6 (1)	P-Amadori-PE	13.55	0.65	11.9	2.5
Am-PE P-18:0/22:6 (2)	P-Amadori-PE	14.73	4.41	8.2	2.5
Am-PE P-18:1/18:1	P-Amadori-PE	14.99	0.67	5.2	3.2
Am-PE P-18:1/18:2	P-Amadori-PE	13.81	4.74	9.4	1.6
Am-PE P-18:1/20:3	P-Amadori-PE	14.43	0.15	8.7	5.8
Am-PE P-18:1/20:4 (1)	P-Amadori-PE	13.66	11.45	10.8	4.9
Am-PE P-18:1/20:4 (2)	P-Amadori-PE	13.95	1.42	8.8	4.9
Am-PE P-18:1/20:5	P-Amadori-PE	12.70	0.23	4.7	3.5
Am-PE P-18:1/22:4	P-Amadori-PE	14.47	0.01	10.3	1.3
Am-PE P-18:1/22:5 (1)	P-Amadori-PE	13.80	0.73	8.3	2.5
Am-PE P-18:1/22:5 (2)	P-Amadori-PE	14.25	0.13	4.4	2.5
Am-PE P-18:1/22:6 (1)	P-Amadori-PE	12.35	0.01	24.5
Am-PE P-18:1/22:6 (2)	P-Amadori-PE	12.63	0.04	9.3
Am-PE P-18:1/22:6 (3)	P-Amadori-PE	13.30	1.28	8.7
Am-PE P-18:1/22:6 (4)	P-Amadori-PE	13.58	0.64	7.3
Am-PE P-18:2/18:2	P-Amadori-PE	12.86	<LOD		4.9
Am-PE P-18:2/20:4	P-Amadori-PE	12.53	0.25	10.0	3.5
Am-PE P-18:2/22:6	P-Amadori-PE	12.15	0.11	15.1
Am-PE P-19:0/18:2	P-Amadori-PE	15.98	0.06	17.0
Am-PE P-19:0/20:4	P-Amadori-PE	15.80	0.10	10.4
Am-PE P-20:0/18:1	P-Amadori-PE	17.42	0.15	10.4
Am-PE P-20:0/18:2	P-Amadori-PE	16.63	0.56	10.5
Am-PE P-20:0/20:3	P-Amadori-PE	16.92	0.06	11.5
Am-PE P-20:0/20:4	P-Amadori-PE	16.47	1.09	9.2	0.73
Am-PE P-20:0/22:4	P-Amadori-PE	17.22	0.02	10.5
Am-PE P-20:0/22:5 (1)	P-Amadori-PE	16.52	0.03	8.1
Am-PE P-20:0/22:5 (2)	P-Amadori-PE	16.94	0.03	16.5
Am-PE P-20:0/22:6	P-Amadori-PE	16.17	0.29	11.0
Am-PE P-20:1/18:1	P-Amadori-PE	16.25	0.08	11.5
Am-PE P-20:1/22:6	P-Amadori-PE	14.72	0.15	9.1
Am-PE P-21:0/20:4	P-Amadori-PE	17.07	0.02	15.6
Am-PE P-22:0/18:2	P-Amadori-PE	17.73	0.14	9.9
Am-PE P-22:0/20:4	P-Amadori-PE	17.61	0.20	9.8
Am-PE P-22:1/20:4	P-Amadori-PE	16.37	0.25	9.1
Am-PE P-22:1/22:6	P-Amadori-PE	16.08	0.07	15.6
Am-PE P-22:2/20:4	P-Amadori-PE	15.31	0.07	11.1

^*Isobaric species that couldn’t be further differentiated in this study were numbered. These species were proposed as isomers with different double bond positions.

^#The concentration of non-glycated substrate as reported by Bowden et al (Ref 16).

We calculated the glycation rate based on the concentration ratio between the identified Amadori-PEs/LPEs and their unmodified analogs in the reference human plasma NIST SRM-1950, with the latter values obtained from a lipidomics interlaboratory harmonization study [14]. Although the glycation rate varies to some degree depending on the subtype of PE (Figure 5A), the average glycation rate was estimated to be 0.07 mol%, which is in agreement with previously reported 0.05–0.10 mol% [9, 17–19]. For example, the plasma concentration of Amadori-PE 16:0/18:1 measured in this study is 1.46 nM with 0.12 mol% glycation rate, and it was reported in healthy plasma as 3 nM with 0.08 mol% glycation rate [9]. Similarly, the average glycation rate of PE 16:0/20:4 and PE 16:0/22:6 in healthy plasma has been reported as ~0.05–0.10 mol% [9], which is consistent with results in this study (0.07–0.16 mol%). To see if a higher substrate concentration results in a higher glycation rate, we also did a correlation analysis between the concentrations of PE/LPE reported in NIST SRM-1950 human reference plasma [14] and the glycation rate of the corresponding Amadori-PE/LPE in this study. We obtained a Pearson correlation coefficient of 0.634 (p < 0.0001) (Figure 5B), indicating a positive, but not very strong correlation between the Amadori-PE/LPE and the PE/LPE levels in this plasma sample.

Figure 5.

(A) The glycation rate (mol%) of all Amadori-PEs/LPEs to PEs/LPEs in the NIST SRM-1950 human reference plasma, categorized in subtypes of PE. (B) The Pearson correlation of the glycation rate to the concentrations of PE/LPE reported in the NIST SRM-1950 human reference plasma by a lipidomics interlaboratory harmonization study [14].

4. CONCLUSION

In this study, we developed a more sensitive LC-MRM-MS method for comprehensive profiling of Amadori-PEs/LPEs in human plasma, as compared to the previously published procedures. To improve identification of the low abundant Amadori-PEs/LPEs, we used in vitro glycated lipid extracts and performed LC-NLS-MS and LC-PIS-MS in tandem, and verified the tentatively identified Amadori-PEs/LPEs using high mass accuracy analysis as well as the correlation between retention time, fatty acyl chain length and degree of unsaturation. As a result, 80 Amadori-PEs and 13 Amadori-LPEs were newly identified in this study. Using LC-MRM-MS and respective stable isotope labeled standards for each subclass of Amadori-PEs, we developed a rapid and robust quantitative method for analysis of Amadori-PE/LPE in human plasm. The developed method was successfully applied to determine the levels of 142 Amadori-PEs and 33 Amadori-LPEs in the reference human plasma NIST SRM-1950. To our knowledge, this study presents the most comprehensive profile of Amadori-PEs/LPEs in human plasma. The measurement of Amadori-PE/LPE in the reference human plasma also provides reference values for future studies of Amadori-PEs in human health and disease. We also note the limitation of this study in using Amadori-PE P-18:0/20:4 as internal standards for quantifying both P- and O- ether linked Amadori-PEs. A more accurate quantitation of O-Amadori-PEs would be achieved if using authentic isotope labeled O-Amadori-PE species as internal standards. In addition, although the common practice in quantitative lipidomics in the research community is to use one isotope labeled standard per lipid class, using more than one Amadori-PE internal standards for each subclass of Am-PE (diacyl, P- and O-ether linked) would provide more accurate values of these modified lipids in clinical specimen.