Oxidized lipid detection

Abstract

The present invention is concerned with a method of extracting oxidized lipids from a lipid solution, the method comprising (a) a derivatisation step, comprising contacting a derivatisation agent with the lipid solution such that aldehydic oxidized lipids and/or ,-unsaturated oxidised lipids, if present in the lipid solution, are derivatised to include an anionic group, and (b) an oxidised lipid capture step, in which nanoparticles are contacted with the lipid solution, wherein the nanoparticles capture anionic-group containing oxidised lipids. The invention also includes a method of extracting aldehydic oxidized phospholipids from a lipid solution, the method comprising (a) a derivatisation step, comprising introduction of a anionic group to aldehydic oxidized lipids and/or ,-unsaturated oxidised lipids in the lipid solution, and (b) an oxidised lipid capture step, in which nanoparticles are contacted with the lipid solution, wherein the nanoparticles bind anionic-group containing oxidised lipids.

Claims

1. A method of extracting oxidised lipids from a lipid solution, the method comprising (a) a derivatisation step, comprising: (i) a reductive amination of aldehydic oxidised lipids present in the lipid solution, comprising contacting the lipid solution with a derivatisation agent selected from 2-aminoanthranilic acid (2AA), 4-aminoanthranilic acid (4AA), 3-(4-aminophenyl)-propionic acid (4APA) and 6-aminocaproic acid (6ACA), such that the aldehydic oxidised lipids present in the lipid solution are derivatised to include a carboxy group; or (ii) a hydrazone formation with aldehydic oxidised lipids and/or ,-unsaturated oxidised lipids present in the lipid solution, comprising contacting the lipid solution with a derivatisation agent that is a compound according to formula (I): ##STR00006## or a salt, solvate or hydrate thereof wherein: X represents a carboxy group each L independently represents a single bond or linker group n is an integer from 0 to 4, and A represents an optionally substituted heteroaryl or aryl group, such that the aldehydic oxidised lipids and/or ,-unsaturated oxidised lipids present in the lipid solution are derivatised to include a carboxy group; and (b) an oxidised lipid capture step, in which nanoparticles are contacted with the lipid solution, wherein the nanoparticles capture carboxy group containing oxidised lipids.

2. A method according to claim 1, wherein the nanoparticles comprise a metal oxide.

3. A method according to claim 1, wherein the nanoparticles comprise a magnetic core and a surface, and wherein the nanoparticles are functionalised such that the surface is electrophilic.

4. A method according to claim 3, wherein the magnetic core comprises iron oxide.

5. A method according to claim 3, wherein the magnetic core is surface functionalised such that the surface is hydrophilic; and/or the surface binds carboxy group containing oxidised lipids by an anion-exchange mechanism; and/or the surface comprises a polymer; and/or the surface comprises or consists of amines or polyetheramines.

6. A method according to claim 1, wherein for the reductive amination of (i), the aldehydic oxidised lipids present in the lipid solution are derivatised to include an aromatic group.

7. A method according to claim 1, wherein the derivatisation step comprises the hydrazone formation of (ii).

8. The method according to claim 7, wherein the derivatisation agent is ##STR00007## or a salt, solvate or hydrate thereof.

9. A method according to claim 1, further comprising an initial step of lipid extraction, comprising extraction of the lipid solution from a biological sample.

10. A method according to claim 1, wherein in step (b) the nanoparticles are added to the lipid solution, and further comprising after step (b), (ci) an extraction step, comprising extraction of the nanoparticles from the lipid solution.

11. A method according to claim 10, wherein the nanoparticles are extracted from the lipid solution by centrifugation or by using a magnet.

12. A method according to claim 10, wherein the nanoparticles are fixed to a surface, and further comprising after step (ci), (cii) a washing step, comprising washing the surface to remove excess lipid solution.

13. A method according to claim 12, further comprising after step (cii), (d) a data gathering step, in which data representative of the aldehydic oxidised lipids of (i) or the aldehydic oxidised lipids and/or ,-unsaturated oxidised lipids of (ii) is obtained.

14. A method according to claim 13, further comprising (e) a data analysis step, in which the aldehydic oxidised lipids of (i) or the aldehydic oxidised lipids and/or ,-unsaturated oxidised lipids of (ii) are detected, quantified and/or identified based on the data obtained.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) Embodiments of the invention will now be described in detail with reference to the accompanying figures, in which:

(2) FIG. 1 shows the influence of the 4AA concentration on the MALDI ionization efficiency of 0.2 M (100 g/L) ALDO-OxPCs. The ATT matrix ion at m/z 569 was used as internal standard (ISD) for relative quantitative evaluation.

(3) FIG. 2 shows the influence of different amounts (1 and 10 g) of MAG-PEA-NPs on the binding of CARBO- and 4AA-ALDO-OxPCs (5 pg each) spiked to mouse plasma (MP). The ATT matrix ion at m/z 569 was used as internal standard (ISD) for relative quantitative evaluation.

(4) FIG. 3 shows the influence of different NP-incubation temperatures on the auto-oxidation of PLs. 20 L human plasma (HP) lipid extracts dissolved in 200 L MeOH were incubated with 20 g MAG-PEA-NPs and incubated for 20 minutes at 4 C. and 25 C., respectively. The degree of oxidation was measured based on the signal intensity ratio of major ALDO-OxPLs (i.e. POVPC, SOVPC and PONPC) relative to their respective precursors (i.e. PAPC, SAPC and PLPC) detected by MALDI-MS.

(5) FIGS. 4A and 4B show the preferential binding of ZrO.sub.2 nanoparticles to carboxy-group containing OxPLs, over aldehydic OxPLs and long-chain OxPCs. In A) the MALDI mass spectra recorded from a mixture of OxPLs before and in B) after incubation with ZrO.sub.2NPs are shown.

(6) FIGS. 4C and 4D show the effect of the nanoparticle identity on MALDI ionisation properties and binding efficiencies. In C) the influence of ZrO2 and two surface-functionalize Fe.sub.3O.sub.4-core NPs (i.e. MAG-NH2 and MAG-PEA) on the ionization and in D) the trapping efficiency of individual short-chain OxPCs (5 pg each) spiked to mouse plasma (MP) is shown. The ATT matrix ion at m/z 569 was used as internal standard (ISD) for relative quantification of the peaks.

(7) FIG. 5 illustrates the selective binding of OxPCs from human plasma (HP) using ZrO.sub.2NPs (100 nm). MALDI mass spectra of A) pure HP, B) pure OxPAPC (10 ng/L) and C) a mixture of HP+OxPAPC dissolved in MeOH are shown. In D) the MALDI mass spectrum of the mixture shown in C) but after incubation with 10 g ZrO.sub.2NPs and on-probe analysed by MALDI-MS is shown. Note the selective binding of short-chain OxPCs (e.g. POVPC, PGPC, PAZPC) but absence of any other abundant PLs present in HP (e.g. PLPC, SM16:0). The peak indicated by an asterisk (m/z 569) originates from ATT matrix used for sample analysis.

(8) FIG. 6 illustrates the selective binding of CARBO-OxPCs using magnetic surface-functionalized Fe.sub.3O.sub.4NPs (100 nm). In A) the MALDI mass spectrum of a mixture of four individual OxPCs (200 g/L) before NP-binding is shown. The mass spectrum of the same sample recorded after incubation with B) aminopropyl-(MAG-NH.sub.2) and C) polyetheramine-(MAG-PEA) functionalized NPs is shown. Note the selective NP-binding of the CARBO-OxPCs (i.e. PGPC, PAZPC) while the ALDO-OxPCs (i.e. POVPC, PONPC) were removed during the washing step.

(9) FIG. 7A shows MALDI spectra of PLs in the positive ionization mode using different types of 5000 nm MeO.sub.2-microparticles (MPs). The MALDI mass spectra of A) a pure PL-standard mixture (1 pmol/L dissolved in MeOH) and after incubation with B) TiO.sub.2-MPs, C) SnO.sub.2-MPs, and D) ZrO.sub.2-MPs are shown. Note that ATT was used as matrix substance. The signals represent [M+H].sup.+ ions of the individual PL species. The peak indicated by an asterisk (m/z 663) represents the [M+H].sup.+ ion of a plasticizer (Irganox168) that leached out of the plastic materials used.

(10) FIG. 7B shows MALDI spectra of PLs in the negative ionization mode using different types of 5 m MeO.sub.2-microparticles (MPs). The MALDI mass spectra of A) a pure PL-standard mixture (1 pmol/L dissolved in MeOH) and after incubation with B) TiO.sub.2-MPs, C) SnO.sub.2-MPs, and D) ZrO.sub.2-MPs are shown. Note that 9AA was used as matrix substance. The signals represent [MH].sup. ions of the individual PL species.

(11) FIG. 8 demonstrates the selective derivatisation of ALDO-OxPCs using 2-aminoantranilic acid (2AA). The MALDI mass spectra of A) a mixture of four individual OxPCs (50 g/L) and B) of the same sample after derivatization with 2AA reagent are NP-binding, are shown. Note the mass shift of 121 Da of the ALDO-OxPCs (i.e. POVPC, PONPC) while the m/z values of the CARBO-OxPCs (i.e. PGPC, PAZPC) remain unaffected.

(12) FIG. 9 demonstrates selective 2AA-derivatisation and NP-binding of ALDO-OxPCs from mixtures of lipid oxidation products. The MALDI mass spectra of 10 ng/L A) OxPAPC and C) OxPLPC before and B) OxPAPC and D) OxPLPC after 2AA-labeling and binding to MAG-NH2-NPs are shown. Note the selective NP-binding of CARBO-OxPCs and 2AA-labeled ALDO-OxPCs, while long-chain OxPLs (i.e. lipid hydroperoxides and epoxyisoprostanes) are removed during the washing step. The mass shift of 121 Da after 2AA-derivatization is indicated for major ALDO-OxPCs present in the samples. The peak indicated by an asterisk (m/z 569) originates from ATT matrix used for sample analysis.

(13) FIG. 10A shows the mechanism for derviatisation (reductive amination) of an example aldehydic OxPC (POVPC), using a 2AA derivatisation reagent. The w-terminal carbonyl-group (i.e. the aldehyde) of the POVPC initially forms an intermediate Schiff's base (imine), which is subsequently reduced to the amine using a reducing agent (e.g. cynanoborohydride). The resulting mass shift of +121 Da can be used as specific mass tag during MALDI-MS analysis

(14) FIGS. 10B and 100 show the results of the derivatisation procedure for four different derivatisation reagents. In B) the influence of different derivatization reagents on the ionization efficiency of different ALDO-OxPCs and in C) the binding efficiency to MAG-PEA-NPs compared to the unlabelled (pure) OxPCs are shown. The ATT matrix ion at m/z 569 was used as internal standard (ISD) for relative quantification of the peaks. 2AA: 2-aminoantranilic acid, 4AA: 4-aminoantranilic acid, 4APA: 3-(4-aminophenyl)-propionic acid, GACA: 6-aminocaproic acid

(15) FIG. 11 shows the influence of the 4AA concentration on the OxPC-binding efficiency and the capacity of MAG-PEA-NPs. In A) the effect of increasing concentration of 4AA (0.2-200 M) of CARBO- and 4AA-labeled ALDO-OxPCs on the binding to MAG-PEA-NPs and in B) the percentage of different amounts (0.4-20 nM) of CARBO- and 4AA-ALDO-OxPCs bound to 20 g MAG-PEA-NPs are shown. The ATT matrix ion at m/z 569 was used as internal standard (ISD) for relative quantitative evaluation.

(16) FIG. 12 shows the influence of 4AA-derivatization on the MS/MS properties and chromatographic behaviour of ALDO-OxPCs. In A) the MS/MS spectra of 4AA-POVPC (m/z 715) using MALDI-QIT-TOF and B) ESI-QQQ-MS/MS are shown. In C) the extracted ion chromatograms (EIC) of C) POVPC (m/z 594) and D) 4AA-POVPC (m/z 715) based on monitoring of the diagnostic ion at m/z 184 by LC-ESI-SRM-MS/MS are shown. A retention time shift of about 2.5 minutes and complete absence of the small isobaric peaks was observed after derivatisation. Note that the diagnostic fragment ion m/z 184 derived of the phosphocholine headgroup from 4AA-POVPC (m/z 715) is independent of the used ionization technique (i.e. ESI- or MALDI) and instrument types (i.e. QIT-TOF or QQQ).

(17) FIG. 13 shows the detection of endogenous OxPLs from mouse plasma using different NP-types. In A) the MALDI-MS spectra measured before and in B) after incubation with MAG-NH.sub.2NPs, in C) with ZrO.sub.2NPs, and in D) with MAG-PEA-NPs are shown. The peaks marked by asterisks represent OxPLs bound to the different NPs. In E) the identification of the different peaks based on m/z values and the retention times using LC-ESI-MS/MS and in F) a comparison between the relative signal intensities of the OxPLs detected by LC-MS and the approach described for the first time herein (termed Nano-MALDI) are shown. The peak at m/z 663 represents a background signal related to a plasticizer (Irganox168) leached out of the plastic materials used..sup.[19]

(18) FIG. 14 shows the detection and quantification of OxPLs in mouse plasma and human oxidized lipoproteins. In A) the calibration curves of different OxPC standards detected by MALDI-MS after binding to MAG-PEA-NPs using 1-O-octadecyl-2-butyryl-sn-glycero-3-phosphocholine (C4-PAF) (m/z 580) as a structurally closely related internal standard (ISD) and in B) the calculated concentrations of OxPLs (i.e. OxPCs and OxPAFs) are shown. In panel C) the levels of LPCs and OxPLs displayed relative to the signal intensity of their precursor species and in D) the relative levels of the individual OxPL species detected by the Nano-MALDI approach are displayed. Plasma samples of ApoE/ mice after a 3-months high-fat diet (ApoE-HF) were used.

(19) FIG. 15 shows MALDI-MS/MS analysis of endogenous OxPLs present in mouse plasma. The MALDI-post source decay (PSD) mass spectra of the peaks at A) m/z 610, B) m/z 624, C) m/z 638, D) m/z 652, and E) m/z 666 related to OxPLs detected on-probe by MALDI-MS after NP-binding from crude lipid extracts (see FIG. 13). Note the detection of the prominent diagnostic ion m/z 184 from the phosphocholine headgroup in all mass spectra (see left panel). A detection window of 5 Da was used for selection of the precursor ions for MALDI-PSD analysis (see right panel).

(20) FIG. 16 shows MALDI spectra of endogenous OxPLs present in plasma of ApoE-deficient mice in vivo. The MALDI mass spectra of plasma lipid extracts of A) C57BL/6 wild type mice, B) ApoE/ mice under normal diet (ApoE-ND), and C) ApoE/ mice under high-fat diet (ApoE-HF) detected on-probe after binding to MAG-PEA-NPs are shown. Note the increased intensity of OxPC-related signals in the plasma samples of the ApoE-HF mice. The peak indicated by an asterisk (m/z 663) represents the [M+H].sup.+ ion of a plasticizer (Irganox168) leached out of the plastic materials used.

(21) FIG. 17 shows MALDI spectra of endogenous OxPLs present in HOCl-modified LDL in vitro. The MALDI mass spectra of lipid extracts of A) HOCL-modified bovine serum albumin (HOCl-BSA) and B) HOCL-modified LDL (HOCl-LDL) detected on-probe after binding to MAG-PEA-NPs are shown. Note the absence of any OxPC-related signals in the mass spectrum of HOCl-BSA containing no oxidisable PLs. The peak indicated by an asterisk (m/z 663) represents the [M+H].sup.+ ion of a plasticizer (Irganox168) leached out of the plastic materials used for sample preparation.

(22) FIG. 18 shows the binding efficiency of ALDO-OxPCs after derivatisation with 4-AA and 4-CBH.

DETAILED DESCRIPTION

(23) The embodiments described below are illustrative and by way of example only. They should not be taken to limit the scope of the invention in any way.

EXAMPLES

Methodology

(24) Lipids

(25) The lipids used in the experiments detailed below are listed in table 1.

(26) Oxidation of Native PLs Containing Esterified PUFAs

(27) Mixtures of short- and long-chain OxPLs (e.g. OxPAPC) derived from oxidation of the sn-2 PUFA residues of native PLs (e.g. PAPC) were obtained by auto-oxidation of completely dried samples at room temperature (25 C.) over night.

(28) The samples were then re-suspended in chloroform and stored under argon at 80 C. until ESI-MS or MALDI-MS analysis.

(29) Isolation and Modification of Human Low-Density Lipoprotein

(30) Immediately after venipuncture blood was supplemented with EDTA (1 mM) and BHT (20 M). Native LDL (n-LDL) was isolated from 3 mL human plasma by preparative ultracentrifugation (Sorvall ETD, SW-41 swing out rotor, 41000 rpm, 10 C., 24 h) using a discontinuous KBr density gradient.

(31) For preparation of oxidized LDL (OxLDL) the EDTA-stabilized samples (0.25 mg/ml protein) were extensively dialyzed in the cold against PBS (pH 7.4). The samples were subsequently incubated with 50 M CuSO.sub.4 for 24 hours at 37 C. for preparation of Cu(II)-oxidized LDL (Cu-LDL) or with 8 mM NaOCl in 0.8 mM NaOH for the preparation of HOCl-oxidized LDL (HOCl-LDL).

(32) Samples were stored under nitrogen in the dark at 4 C. to perform the subsequent experiments.

(33) Lipid Extraction and Micro Solid-Phase Extraction of Mouse Plasma and Lipoproteins

(34) 500 L MeOH was added to 20 L of the samples, mixed vigorously for 1 minute and kept on ice for 30 minutes for lipid extraction. Afterwards, vials were centrifuged for 5 min (10000 rpm at +4 C.) to separate the lipids from the precipitated proteins. Finally, supernatants were carefully transferred to 2 mL glass vials and stored at 70 C. before further analysis..sup.[20]

(35) Before LC-ESI-MS/MS analysis, OxPLs were separated from bulk plasma lipids using our recently established micro-preparative high-performance solid-phase extraction (HP-SPE) method..sup.[7]

(36) Briefly, C18-SPE columns (PepClean, Pierce) were first equilibrated by 500 L pure MeOH followed by an equal volume of MeOH:0.2% formic acid=70:30 (v/v) (=loading buffer). Afterwards the SPE columns were loaded 3-times by the lipid extracts (dissolved in loading buffer) followed by a 500 L washing step using loading buffer to remove most of the LPCs, followed by a single elution using 700 L MeOH: 0.2% formic acid=82:18 containing the majority of OxPCs (>60-70%) and 800 L MeOH:0.2% formic acid=98:2 containing >95% of the unoxidized PCs. The OxPC containing fraction was directly used for the subsequent LC-ESI measurements.

(37) Derivatisation of Aldehydic OxPLs

(38) Stock solutions containing 4 mM of the derivatisation reagents (2AA, 4AA, 4APA, GACA) and 8 mM sodium cyanoborohydride (CB) as reducing agent dissolved in MeOH were prepared. Additionally a stock solution of 4 mM 4-CBH in MeOH without the addition of CB was prepared.

(39) 2 L of each of these solutions was mixed with 20 L OxPLs dissolved in MeOH, or a 1:5-1:20 diluted mouse or human plasma lipid extract in MeOH, resulting in a final concentration of 400 M of the labelling reagents in the sample.

(40) Finally, the sample was incubated for 15 minutes at room temperature resulting in an almost complete reaction using 0.2 M ALDO-OxPCs (see FIG. 1), which is below the normal physiological range reported in plasma..sup.[21]

(41) Nanoparticle Binding of OxPLs

(42) For extraction of the OxPLs from solution, 20 L of the lipid samples were mixed with 180 L MeOH and 20 g NPs were added. The solution was briefly mixed in order to suspend the NPs and incubated 20 minutes at 4 C. These were found the optimal conditions in regard to the binding efficiency and recovery of OxPLs without affecting their endogenous levels due to auto-oxidation of the samples during incubation (see FIGS. 2 and 3).

(43) Subsequently, the NPs were sedimented by centrifugation (e.g. ZrO.sub.2NPs) or by using a magnet (e.g. Fe.sub.3O.sub.4NPs), and the supernatant containing unbound material was removed.

(44) Finally, the remaining NPs were washed by re-suspension in 500 L MeOH, sedimented again and the supernatant discarded.

(45) MALDI Sample Preparation

(46) For the detection of OxPCs in positive ionisation mode (MALDI), the NPs were mixed with 2 L ATT matrix (10 mg/mL) dissolved in EtOH:H.sub.2O=90:10 (v/v) containing 2 mM GUA and 5 mM DAHC..sup.[7]

(47) For detection of OxPCs in the negative ionisation mode (MALDI), a 9AA matrix (10 mg/mL) dissolved in ISO:ACN=60:40 containing 5 mM GUA was used..sup.[22] In either case, 1 L of the NP-matrix suspension was deposited on FlexiMass-DS (FLDS) sample plates (Shimadzu Biotech, Manchester, UK) and inserted into the MALDI instrument using a specific adaptor carrying up to four sample plates (AXIMA-Precision, Shimadzu, Manchester, UK).

(48) MALDI Mass Spectrometry

(49) Mass spectra were obtained using an AXIMA-CFRplus (Shimadzu Biotech, Manchester, UK) curved-field reflectron time-of-flight (RTOF) mass spectrometer equipped with a 337 nm pulsed nitrogen laser (3-ns pulse width).

(50) Measurements were performed either in positive or negative mode using delayed ion extraction for optimized mass resolution. The ion acceleration voltage was set to 20 kV and the reflectron analyzer was operated at 25 kV. The laser energy was adjusted to 5-10% above threshold irradiation. An integrated video imaging system (25 magnification) allows direct observation of the sample spots under investigation.

(51) A hybrid MALDI-quadrupole ion trap (QIT)-TOF mass spectrometer (AXIMA-Resonance, Shimadzu, Manchester, UK) was employed for recording of MS/MS spectra (mass range m/z 300-1000) based on low-energy collision-induced dissociation (CID) of the selected precursor ions using argon as collision gas.sup.[23]. The collision energy during the MS/MS experiments was adjusted to 100% precursor ion suppression.

(52) In MS mode 300-500 single laser shots were accumulated. In MS/MS mode 500-1000 single laser shots were accumulated. An external calibration was performed, based on the exact mass values of the [M+H].sup.+ and [MH].sup. ions using mixtures of the defined PL standards.

(53) ESI Mass Spectrometry

(54) ESI mass spectra were acquired using a 4000 QTrap triple quadrupole linear ion trap hybrid mass spectrometer equipped with a Turbo V electrospray ion source (Applied Biosystems, Foster City, Calif., USA).

(55) The lipid samples were dissolved in 50 l of methanol/UHQ 85:15 (v/v) containing 5 mM ammonium formate and 0.1% formic acid.

(56) For LC-MS analysis 5-10 L were injected onto a core-shell type C18 column (Kinetex 2.6 m, 503.0 mm ID; Phenomenex, Torrance, Calif., USA), which was kept at 20 C. A linear binary gradient consisting of 5 mM ammonium formate and 0.1% (v/v) formic acid in UHQ (eluent A) and MeOH (eluent B) was used as mobile phases at a flow rate of 400 L/min over 20 min total run time.

(57) The ion source was operated in positive ion mode using an electrospray ionization voltage of 4500 V and an ion source temperature of 550 C. Nitrogen was used as nebulizer, heater, curtain, and collision gas for MS/MS experiments. Detection was carried out by selected reaction monitoring (SRM) of the m/z 184 product ion corresponding to the PC headgroup.

(58) MS Data Analysis

(59) MS data were processed by the manufacturer supplied instrument software versions Launchpad 2.9.1 (Shimadzu, Manchester, UK) and Analyst 1.5 (Applied Biosystems), respectively. MALDI mass spectra were routinely calibrated using the exact mass values of the lipid standards and smoothed using Savitzky-Golay algorithm.

(60) Results

(61) Evaluation of Different NP-Types for Selective Detection of OxPLs

(62) Metal-Oxide Nanoparticles

(63) Metal-oxide (MeO.sub.2) particles (e.g. SnO.sub.2, TiO.sub.2, ZrO.sub.2) have been described previously for the selective analysis of phosphopetides by MALDI-MS for proteomics studies or for HPLC-ESI-MS/MS analysis of PLs.sup.[24][25][26].

(64) The inventors have previously shown that such particles can be used for the selective analysis of oxidised phospholipids containing carboxy-groups.

(65) Metal-oxide nanoparticles contacted with mixtures of abundant unoxidized PLs and short- and long-chain OxPLs (e.g. OxPAPC spiked mouse plasma), showed a preferential binding and enrichment of carboxy-group containing OxPCs (CARBO-OxPCs) (e.g. PGPC, PAzPC) but not aldehydic OxPCs (ALDO-OxPCs) (e.g. POVPC) and long-chain OxPCs (e.g. PEIPC). See FIGS. 4A, 4B and 5.

(66) The same enrichment effect of short-chain CARBO-OxPCs (e.g. PGPC) in contrast to neutral long-chain OxPLs (e.g. PEIPC) and the abundant unoxidized PLs (e.g. PLPC) was observed when OxPAPC-spiked mouse plasma samples were used (see FIG. 6).

(67) Without wishing to be bound by theory, it is believed that this selectivity relies on an ion-pairing mechanism based on the electrophilic interaction of the anionic OxPLs (e.g. CARBO-OxPLs) and the nanoparticle surfaces.

(68) Further, the inventors have found that, of the metal oxide nanoparticles tested, that only ZrO.sub.2 particles allowed the binding and detection of PLs from mixtures in positive (+) and negative () MALDI ionization mode (see FIGS. 7A and 7B). As can be clearly seen ZrO.sub.2 particles show a preferred binding of acidic (anionic) PLs (e.g. PA, PE, PS) from a lipid mixture (FIG. 7B). Nevertheless some contribution of the phosphate headgroup to the binding in case of LPC, PC and SM cannot be ruled out using this type of particles (Figure A).

(69) Magnetic Core Nanoparticles

(70) Although ZrO.sub.2NPs were found to bind OxPLs well, the inventors found that they are sometimes unstable in solution, and needing repeated vortexing during sample incubation.

(71) In contrast, so-called magnetic-core NPs (e.g. silanized iron oxide, carbon-coated cobalt) have been shown to be very suitable for sample handling and NP-harvesting using magnetic separation..sup.[27][28]

(72) The inventors tested the selectivity of nanoparticles comprising a magnetic iron oxide (Fe.sub.3O.sub.4)-core, surface functionalised with hydrophilic polymers. The hydrodynamic diameter of the nanoparticles is approximately 100 nm. The surface layer was believed to be approximately 10-60 nm thick.

(73) Two nanoparticles types were tested. Each had a superparamagnetic magnetite (Fe.sub.3O.sub.4) core surrounded by an aminosilane matrix. One was surface-functionalized with propylamine (MAG-NH2) and the other with poly-(dimethylamin-co-epichlorhydrin-co-ethylendiamin) (MAG-PEA) respectively. These groups were selected for their anion exchange properties, and showed the same selectivity for CARBO-OxPCs (see FIG. 6). In contrast, to MeO.sub.2 particles (e.g. ZrO.sub.2) the contribution of the phosphate group for binding can be ruled out, as is demonstrated by the complete absence of signals related to POVPC and PONPC (FIG. 6). This further supports the electrophilic interaction binding hypothesis described above.

(74) In contrast, to gold nanoparticles (GNPs) and also the MeO.sub.2NPs, the MAG-NH.sub.2 and particularly the MAG-PEA nanoparticles showed full compatibility with ethanol, methanol or even acetone. This allows the analysis of OxPLs by MALDI-MS directly from lipid solutions extracted using these solvents (see Table 2).

(75) TABLE-US-00001 TABLE 2 H.sub.2O MeOH EtOH AcOH GNPs + MeO.sub.2-NPs + + +/ NH.sub.2-NPs + + + PEA-NPs + + + + Stability of different NP-types in solvents used for lipid extraction of biological samples (e.g. mouse and human plasma). GNPs: gold-NPs, MeO2: metaloxide NPs, NH2-NPs: aminpropyl-NPs, PEA-NPs, polyetheramine-NPs. MeOH: methanol, EtOH: ethanol, AcOH: acetone.
Comparison of Different Nanoparticle Binding Efficiency

(76) Following these initial tests, methanolic lipid extracts of mouse plasma (MP) were spiked with OxPC-mixtures (e.g. OxPAPC) or defined OxPC standards (i.e. POVPC, PGPC, PONPC, PAzPC) at very low concentrations (1-5 g/A) and directly (i.e. on-probe) analyzed by MALDI-MS.

(77) The inventors have found that different NPs influence the MALDI ionization properties to different extents, and that different NPs have quite different binding efficiencies (see FIGS. 4C and 4D). Using defined amounts of OxPC standards and of nanoparticles for incubation, the figures in Table 3 were determined.

(78) TABLE-US-00002 TABLE 3 ZrO2 MAG-NH2 MAG-PEA 3.3 2.9 3.8 pg OxPCs/10 g NPs on target 65.9 58.6 77.0 Recovery (%) 0.53 0.47 0.62 fmol OxPCs/g NPs not specified ~2 10.sup.9 ~2 10.sup.9 NPs/g (according to the manufacturer) 1.41 10.sup.1 1.85 10.sup.1 Molecules OxPC/NP ~0.2 nmol/L ~0.2 nmol/L ~0.2 nmol/L OxPC conc. of the samples Experimental figures for different NPs for the detection of OxPLs by MALDI-MS. A mixture of 0.125 pg OxPCs/L (~0.2 nmol/L) dissolved in MeOH was used for evaluation. The total amount of OxPCs (in pg/10 g NPs) was determined based on comparison of the signal intensity of individual OxPCs measured on-probe before and after NP-binding.

(79) The OxPC recovery rate based on comparison of the individual signal intensity of the peaks from an equimolar OxPC mixtures measured on-probe together with the different NPs before and after incubation was found to be 59-77%.

(80) The number of OxPC molecules bound per NP was found to be 1.610.sup.1 for MAG-NH2 and MAG-PEA. This was sufficient to detect and quantify OxPLs at a sample concentration of 0.2 nmol/L (=200 pM) representing a 3-4 orders of magnitude higher sensitivity than the physiological concentrations of OxPLs in human or animal plasma (2 nmol/L-6 mol/L).sup.[29][30][31].

(81) In summary, based on this careful evaluation the suitability of NPs for the selective detection of OxPLs by MALDI-MS was found in the following order: MAG-PEA>ZrO.sub.2>MAG-NH2.

(82) Testing of Derivatisation Reagents for MS-Tagging and NP-Binding of OxPLs

(83) The inventors have found that, while CARBO-OxPLs (e.g. PGPC, PAzPC) can be readily extracted using nanoparticles, ALDO-OxPLs (e.g. POVPC, PONPC) cannot (see for example, FIG. 6).

(84) The inventors use chemical derivatisation in order to introduce a negatively charged chemical group (e.g. a carboxyl group) into ALDO-OxPLs via chemical derivatisation of the -terminal aldehyde groups of the sn-2 FA residues. This allows capture of ALDO-OxPLs by nanoparticles with electrophilic surface properties, using the ion-pairing mechanism described above.

(85) Chemical derivatisation to aid in MALDI-MS analysis is known, as discussed above. The prior art strategies are intended to introduce specific mass shifts (i.e. MS-tagging) and to use characteristic reporter ions allowing a better differentiation of oxidized lipid molecules containing carbonyl-groups (i.e. aldehydes and/or ketones) from more abundant isobaric unoxidized lipid species. Another method.sup.[13] uses 2,4-Dinitrophenylhydrazine (DNPH) as a reactive matrix for the analysis of OxPLs by MALDI-MS to allow the detection of small volatile aldehydes as hydrazones without the need for an additional MALDI matrix.

(86) In contrast, the inventors' method improves in detection/analysis by a) mass-tagging the derivatised aldehydes, b) improving the ionization efficiency of the OxPLs, and c) allowing nanoparticle enrichment of the OxPLs.

(87) The method allows direct (on-probe) analysis of the nanoparticles by MALDI-MS.

(88) Comparison of Reagents

(89) One example of the derivatisation method reacts 2AA with ALDO-OxPLs in the presence of a reducing agent. This reaction introduces a carboxy-group to the OxPL allowing nanoparticle enrichment. It also introduces an aromatic group, which is readily ionisable in mass spectrometry analysis. Further, there is also a characteristic mass shift of 121 Da, allowing ready differentiation from CARBO-OxPLs (see FIGS. 8 and 9).

(90) Other 2AA-derivatives were tested to determine their effect on the MALDI ionization efficiency and NP-binding of ALDO-OxPLs. When compared to 2AA; the use of 4-AA was found to improve the signal intensity of ALDO-OxPLs (e.g. POVPC and PONPC) and increase the binding efficiency to the NPs (e.g. PEA-NPs); the use of 4-APA was found to reduce the signal intensity and reduce the binding efficiency; the use of 6-ACA, which has no phenyl group, was found to have no significant impact on the ionization efficiency but to dramatically reduce the NP-binding properties of ALDO-OxPLs.

(91) See FIGS. 10B and 10C.

(92) These results clearly demonstrate the influence of the physico-chemical properties of the different labelling reagents (e.g. gas-phase basicity, and steric properties) on the MALDI process, as well as the importance of optimized stereochemistry for interaction with functional groups on the NP-surface.

(93) Accordingly, the suitability of the derivatisation agents tested has been determined to be 4-AA>2AA>6ACA>4APA.

(94) 4-carboxybenzohydrazide (4-CBH) was also tested. This molecule was found to introduce a carboxy group to ALDO-OxPLs and additionally to ,-unsaturated OxPLs containing no w-terminal aldehydic groups (e.g. KDdiA-PC).

(95) 4-AA>2AA>6ACA>4APA all react by reductive amination. In contrast, 4-CBH reacts to form a hydrazine. 4-CBH provides comparative signal intensity and binding efficacy results as 4-AA, but allows analysis of additional oxidised lipids.

(96) Concentration of Derivatisation Agent

(97) The optimum 4AA-concentration has been found to be 20 M 4AA in the incubation solution, which maximised binding of CARBO- and ALDO-OxPCs to MAG-PEA-NPs.

(98) Signal quenching was observed at higher concentration (FIG. 11A), indicating saturation of the binding sites on the NP-surface by the excess, unreacted 4AA in the solution.

(99) Thus, the surface capacity of 20 g MAG-PEA-NPs is sufficient to capture 80-100% of total OxPCs present at a concentration of 0.4-4 nM in the incubation solution (FIG. 11B).

(100) OxPL Profiling and Quantification from Mouse Samples

(101) The suitability of MS/MS for the targeted detection of the NP-captured OxPLs was tested using MALDI-QIT-TOF- and ESI-MS/MS as reference methods.

(102) The MS/MS spectra of both the CARBO-OxPCs and 4AA-labeled ALDO-OxPCs showed only one predominant m/z 184 fragment ion derived from the PC headgroup (see FIGS. 12A and 12B). This was observed independent of the ionization process (i.e. MALDI or ESI) and the instrument type.

(103) Moreover, the incorporation of 4AA into the sn-2 acyl-group of the molecules leads to a retention time shift (2.5 minutes) of the peak at m/z 715 (4AA-POVPC) compared to the unlabeled POVPC at m/z 594 (see FIGS. 12C and 12D) which allows the chromatographic separation of the derivatised OxPC molecules from their unreacted precursors.

(104) Using crude mouse plasma samples seven major peaks could be detected in the range of m/z 590-680 dependent on the used NP-type (see FIGS. 13 and 14), whereby the best enrichment of OxPLs was achieved using the MAG-PEA-NPs.

(105) The detection of the m/z 184 fragment ion within the MALDI-MS/MS spectra indicated that they correspond to OxPCs and/or OxPAFs (see FIG. 15). The identity of the peaks was confirmed by LC-ESI-SRM-MS/MS using authentic OxPC and OxPAF standards as reference compounds (FIG. 13E). These types of OxPLs are known to be present in atherosclerotic lesions and OxLDL.sup.[32], whereby OxPAFs were found to be biologically more active at same or even lower concentrations than the structurally corresponding OxPCs.sup.[33][34].

(106) A comparison between the results obtained by Nano-MALDI and LC-ESI-MS shows distinct differences in the enrichment of the individual endogenous OxPL species from the mouse plasma samples. OxPLs containing very short-chain (C4-6) oxidized sn-2 residues (e.g. SPPC) were more effectively captured by the MAG-PEA-NPs compared to OxPL enrichment using the C18 micro-columns (HP-SPE) method..sup.[21] (FIG. 13F).

(107) By comparing plasma samples from C57BL/6 mice with those of ApoE/ mice (i.e. an animal model for atherosclerosis), elevated levels of OxPLs in those fed a high-fat diet (ApoE-HF) were found (FIG. 16). Quantification of selected OxPLs in the ApoE/ mouse plasma revealed a concentration range of 0.2 up to 0.4 M for the detected lipid species (FIGS. 14A and 14B).

(108) This data is in agreement with previous studies demonstrating that these OxPL concentrations were sufficient to exhibit biological and pro-atherogenic activities (e.g. stimulation of monocytes or platelet activation).sup.[35][36].

(109) OxPL Profiling and Quantification from Human Samples

(110) The above method was used for OxPL-profiling from in vitro oxidized human lipoproteins.

(111) A number of peaks attributed to individual OxPC species (which were confirmed by LC-ESI-SRM-MS/MS) were detected in the MALDI mass spectra after NP-binding from the OxLDL samples (see FIG. 17).

(112) The highest levels of OxPCs were found in HOCl-LDL and those of lyso-phosphatidylcholines (LPCs) (which are major degradation products of OxPCs) in Cu-LDL (FIG. 14C). A higher level of OxPLs were found in n-LDL compared to Cu-LDL suggesting more advanced breakdown of OxPLs to LPCs in the latter.

(113) In a similar way to the mouse plasma testing (described above), HHdiA-SPC was found to be the major oxidation product within HOCl-LDL (FIG. 14D).

(114) Consequently, these experiments demonstrate the suitability of this (Nano-MALDI) approach to deliver information about the composition and identity of oxidised lipid molecules (i.e. OxPLs) which are known to represent oxidative stress biomarkers of biological and clinical samples (e.g. human plasma and lipoproteins)

(115) Thus this approach can be used in the development of novel (clinical) screening assays of these important oxidative stress biomarkers.

(116) TABLE-US-00003 TABLE 1 CN:DB exact ion chemical name.sup.(a) abbreviation number.sup.(b) mass species supplier 1-palmitoyl-2-hydroxy-sn-glycero-3- LPC16:0 sn-1 16:0.sup.(c) 496.34 [M + H].sup.+ Avanti phosphocholine 1-heptadecanoyl-2-hydroxy-sn- LPC17:0 sn-1 17:0.sup.(c) 510.36 [M + H].sup.+ Avanti glycero-3-phosphocholine 1-linolenoyl-2-hydroxy-sn-glycero-3- LPC18:0 sn-1 18:0.sup.(c) 524.37 [M + H].sup.+ Avanti phosphocholine 1-O-octadecyl-2-butyryl-sn-glycero- C4-PAF C18-4:0.sup.(e) 580.43 [M + H].sup.+ Avanti 3-phosphocholine (ISD) 1,2-dimyristoyl-sn-glycero-3- DMPA 28:0 591.40 [M H].sup. Avanti phosphate 1-palmitoyl-2-(5-oxo-valeroyl)-sn- POVPC 21:0.sup.(f) 594.38 [M + H].sup.+ Avanti glycero-3-phosphocholine (ALDO) 1-palmitoyl-2-succinoyl-sn-glycero- S-PPC 20:0.sup.(f) 596.36 [M + H].sup.+ n.a. 3-phosphocholine (CARBO) 1-palmitoyl-2-glutaroyl-sn-glycero-3- PGPC 21:0.sup.(f) 610.37 [M + H].sup.+ Avanti phosphocholine (CARBO) 1-stearoyl-2-(5-oxo-valeroyl)-sn- SOVPC 22:0 622.41 [M + H].sup.+ n.a. glycero-3-phosphocholine (ALDO) 1-stearoyl-2-succinoyl-sn-glycero-3- S-SPC 22:0 624.39 [M + H].sup.+ n.a. phosphocholine (CARBO) 1-O-octadecyl-2-glutaroyl-sn- SG-PAF C18-5:0.sup.(e) 624.42 [M + H].sup.+ n.a. glycero-3-phosphocholine (CARBO) 1,2-dimyristoyl-sn-glycero-3- DMPE 28:0 634.45 [M H].sup. Avanti phosphoethanolamine 1-stearoyl-2-glutaroyl-sn-glycero-3- SGPC 22:0 638.40 [M + H].sup.+ n.a. phosphocholine (CARBO) 1-palmitoyl-2-(9-oxo-nonanoyl)-sn- PONPC 25:0.sup.(f) 650.44 [M + H].sup.+ Avanti glycero-3-phosphocholine (ALDO) 1-palmitoyl-2-(5-hydroxy-8-oxoocta- HOOA-PPC 24:1.sup.(f) 650.40 [M + H].sup.+ n.a. 6-enoyl)-sn-glycero-3- (ALDO) 1-palmitoyl-2-(4-hydroxy-7-carboxy- HHdiA-PPC 23:1.sup.(f) 652.38 [M + H].sup.+ n.a. hex-5-enoyl)-sn-glycero-3- (CARBO) phosphocholine 1-O-hexadecyl-2-azelaoyl-sn- PAz-PAF C16-9:0.sup.(e) 652.46 [M + H].sup.+ Sigma glycero-3-phosphocholine (CARBO) 1-palmitoyl-2-azelaoyl-sn-glycero-3- PAzPC 25:0.sup.(f) 666.44 [M + H].sup.+ Avanti phosphocholine (CARBO) 1-palmitoyl-2-(7-carboxy-5- HOdiA-PPC 24:1.sup.(f) 666.40 [M + H].sup.+ n.a. hydroxyhept-6-enoyl)-sn-glycero-3- (CARBO) phosphocholine 1,2-dimyristoyl-sn-glycero-3- DMPS 28:0 678.44 [M H].sup. Avanti phosphoserine 1,2-dimyristoyl-sn-glycero-3- DMPC 28:0 678.51 [M + H].sup.+ Avanti phosphocholine 1-stearoyl-2-(4-hydroxy-7- HHdiA-SPC 25:1.sup.(f) 680.41 [M + H].sup.+ n.a. carboxyhex-5-enoyl)-sn-glycero-3- (CARBO) phosphocholine 1-O-octadecyl-2-azelaoyl-sn- SAz-PAF 27:0.sup.(e) 680.45 [M + H].sup.+ n.a. glycero-3-phosphocholine (CARBO) 1-stearoyl-2-azelaoyl-sn-glycero-3- SAzPC 27:0 694.47 [M + H].sup.+ n.a. phosphocholine (CARBO) N-palmitoyl-D-erythro- SM16:0 C18-16:0.sup.(d) 703.58 [M + H].sup.+ Avanti sphingosylphosphorylcholine 1-palmitoyl-2-(9,12-dioxododec-10- KODA-PPC 28:1.sup.(f) 704.45 [M + H].sup.+ n.a. enoyl)-sn-glycero-3-phosphocholine (ALDO) 1-palmitoyl-2-(11-carboxy-9- KDdiA-PPC 28:1.sup.(f) 720.45 [M + H].sup.+ Cayman oxoundec-6-enoyl)-sn-glycero-3- (CARBO) phosphocholine 1-palmitoyl-2-(9-hydroxy-11- HDdiA-PPC 28:1.sup.(f) 722.46 [M + H].sup.+ n.a. carboxyundec-6-enoyl)-sn-glycero- (CARBO) 3-phosphocholine 1-stearoyl-2-(9-hydroxy-12- HODA-SPC 30:1.sup.(f) 734.50 [M + H].sup.+ n.a. oxododec-10-enoyl)-sn-glycero-3- (ALDO) phosphocholine 1-palmitoyl-2-linoleoyl-sn-glycero-3- PLPC 34:2 758.57 [M + H].sup.+ Avanti phosphocholine 1-palmitoyl-2-arachidonoyl-sn- PAPC 36:4 782.57 [M + H].sup.+ Avanti glycero-3-phosphocholine 1-stearoyl-2-linoleoyl-sn-glycero-3- SLPC 36:2 786.60 [M + H].sup.+ Avanti phosphocholine 1-palmitoyl-2-(9-hydroxy-linoleoyl)- PLPC-OH 34:2.sup.(g) 774.57 [M + H].sup.+ n.a. sn-glycero-3-phosphocholine 1-palmitoyl-2-(9-hydroperoxy- PLPC-OOH 34:2.sup.(g) 790.56 [M + H].sup.+ n.a. linoleoyl)-sn-glycero-3- phosphocholine 1-heptadecanoyl-2-(9Z- HMPI 31:1 793.49 [M H].sup. Avanti tetradecenoyl)-sn-glycero-3- phospho-(1-myo-inositol) 1-palmitoyl-2-(9-hydroxy-14- PLPC- 34:2.sup.(g) 806.56 [M + H].sup.+ n.a. hydroperoxy-linoleoyl)-sn-glycero-3- (OH)OOH phosphocholine 1-palmitoyl-2-(5,6- PECPC 36:3.sup.(h) 810.53 [M + H].sup.+ n.a. epoxycyclopentenone)-sn-glycero- 3-phosphocholine 1-palmitoyl-2-(5,6- PEIPC 36:2.sup.(h) 828.54 [M + H].sup.+ n.a. epoxyisoprostane)-sn-glycero-3- phosphocholine 1-stearoyl-2-docosahexaenoyl-sn- SDPC 40:6 834.60 [M + H].sup.+ Avanti glycero-3-phosphocholine 1-palmitoyl-2-(11,15-dihydroperoxy- PAPC- 36:4.sup.(g) 846.55 [M + H].sup.+ n.a. arachidonoyl)-sn-glycero-3- (OOH).sub.2 phosphocholine 1,3-bis[1,2-dimyristoyl-sn-glycero- TMCL 56:0 1239.84 [M H].sup. Avanti 3-phospho]-sn-glycerol .sup.(a)Nomenclature according to the LIPID MAPS classification standard .sup.(b)Total number of carbon atoms and double bonds of the sn-1 and sn-2 fatty acid residues esterified to the glycerol backbone of the PL molecules. .sup.(c)LPC contain only one fatty acid together with a free hydroxyl group either in sn-1 or sn-2 position. In human plasma saturated fatty acids (e.g. 16:0 or 18:0) are preferentially linked to the sn-1 position.sup.[37] .sup.(d)SM contains one fatty acid linked via amide bond to a sphingosine backbone (long-chain base consisting of C18 or C20 carbon atoms).sup.[38] .sup.(e)PAFs contain one sn-2 fatty acid residue (usually C2-C4 carbon atoms) and one sn-1 ether-linked carbon chain (with C16 or C18 carbon atoms).sup.[39] .sup.(f)For the chemical structures see.sup.[40] .sup.(g)For the isobaric structures.sup.[41] .sup.(h)For the chemical structures.sup.[42] ALDO - OxPC containing sn-2 short-chain -terminal aldehydic fatty acid residue CARBO - OxPC containing sn-2 short-chain -terminal carboxylic fatty acid residue ISD, internal standard (synthetic compound)

REFERENCES

(117) A number of publications are cited above in order to more fully describe and disclose the invention and the state of the art to which the invention pertains. Full citations for these references are provided below. The entirety of each of these references is incorporated herein. .sup.1 P. Libby Nature 2002, 420:868-874 .sup.2 Podrez et al. Nat. Med. 2007, 13:1086-1095 .sup.3 J. A. Berliner and A. D. Watson N. Engl. J. Med. 2005, 353:9-11 .sup.4 F. H. Greig et al. Free Rad. Biol. Med. 2012, 52, 266-280 .sup.5 N. Leitinger Mol. Aspects Med. 2003, 24, 239-250 .sup.6 L. J. Sparvero et al. J. Neurochem. 2010, 115, 1322-1336 .sup.7 Stbiger et al. Anal. Chem. 2010, 82, 5502-5510 .sup.8 M. Schlame et al. J. Lipid Res. 1996, 37, 2608-2615 .sup.9 Frey et al. J. Lipid Res. 2000, 41, 1145-1153 .sup.10 A. D. Watson et al. J. Biol. Chem. 1999, 274, 24787-24798 .sup.11 K. A. Harrison et al. J. Mass Spectrom 2000, 35, 224-236 .sup.12 X. Chen et al. J. Lipid Res. 2008, 49, 832-846 .sup.13 I. Milic et al. Anal. Chem. 2013, 85, 156-162 .sup.14 K. Teuber et al. Anal. Lett. 2012, 45, 968-976 .sup.15 Stbiger et al. Atherosclerosis 2012, 224, 177-186 .sup.16 Reis A. et al., Biochim. Biophys. Acta 2012, 1818, 2374-2387 .sup.17 Moumtzi A. et al., 2007, J. Lipid Res. 48:565-582 .sup.18 N. Auge et al. Arterioscler. Thromb. Vasc. Biol. 1990, 22, 1990-1995 .sup.19 J. Wu et al. Chem. Phys. Lipids 2011, 164, 1-8. .sup.20 Z. Zhao and Y. Xu J. Lipid Res. 2010; 51:652-659 .sup.21 Podrez et al. Nat. Med. 2007, 13:1086-1095 .sup.22 G. Sun et al. Anal. Chem. 2008, 80, 7576-7585 .sup.23 R. L. Martin and F. L. Brancia 2003, Rapid Commun. Mass Spectrom. 17, 1358-1365 .sup.24 Yan. Li et al. J. Proteome Res. 2008, 7, 2526-2538. .sup.25 Ana Gonzlvez et al. Anal. Bioanal. Chem. 2010, 396, 2965-2975 .sup.26 Grady R. Blacken et al. Anal. Chem. 2007, 79, 5449-5456 .sup.27 Robert N. Grass et al. Angew. Chem. Int. Ed. 2007, 46, 4909-4912 .sup.28 Yan. Li et al. J. Proteome Res. 2008, 7, 2526-2538 .sup.29 Podrez et al. Nat. Med. 2007, 13:1086-1095. .sup.30 B. Frey et al. J. Lipid Res. 2000; 41:1145-1153 .sup.31 Stbiger et al. Atherosclerosis 2012, 224, 177-186 .sup.32 K. Gopal et al. Trends Cardiovasc. Med. 2001, 11, 139-142 .sup.33 H. Kern et al. Biochim. Biophys. Acta 1998, 1394, 33-42 .sup.34 X. Chen et al. J. Lipid Res. 2008, 49, 832-846 .sup.36 H. Kern et al. Biochim. Biophys. Acta 1998, 1394, 33-42 .sup.36 E. A. Podrez et al. Nat. Med. 2007, 13:1086-1095 .sup.37 P V. Subbaiah et al. Biochim. Biophys. Acta 1992; 1128:83-92 .sup.38 B. Ramstedt et al. Eur. J. Biochem. 1999; 266:997-1002 .sup.39 S M. Prescott Annu. Rev. Biochem. 2000; 69:419-445 .sup.40 S L. Hazen J. Biol. Chem. 2008, 283, 15527-15531. .sup.41 A. Reis J. Mass Spectrom. 2004; 39: 1513-1522 .sup.42 G. Subbanagounder J. Biol. Chem. 2002, 277, 7271-7281.