spectrometry supercritical ﬂuid chromatography coupled tomass chromatography liquid andultrahigh-performance Analysis in human of oxylipins plasma: Comparison ofultrahigh-performance Journal of Chromatography A

(1)

ContentslistsavailableatScienceDirect

Journal of Chromatography A

jou rn al h om ep a g e : w w w . e l s e v i e r . c o m / l o c a t e / c h r o m a

Analysis of oxylipins in human plasma: Comparison of ultrahigh-performance liquid chromatography and

ultrahigh-performance supercritical ﬂuid chromatography coupled to mass spectrometry

Róbert Berkecz

^a^,^b

, Miroslav Lísa

^a

, Michal Holˇcapek

^a^,∗

aDepartmentofAnalyticalChemistry,FacultyofChemicalTechnology,UniversityofPardubice,Studentská573,53210Pardubice,CzechRepublic

bDepartmentofMedicalChemistry,FacultyofMedicine,UniversityofSzeged,Dómtér8,H-6720Szeged,Hungary

a r t i c l e i n f o

Articlehistory:

Received23January2017

Receivedinrevisedform23June2017 Accepted27June2017

Availableonline29June2017

Keywords:

Oxylipin Eicosanoid Prostaglandin Plasma UHPSFC/MS UHPLC/MS SPE ESI

a b s t r a c t

Thepotentialofultrahigh-performanceliquidchromatography–massspectrometry(UHPLC/MS)and ultrahigh-performancesupercriticalfluidchromatography(UHPSFC)coupledtonegative-ionelectro- sprayionizationmassspectrometry(ESI–MS)fortheanalysisof46oxylipinsand2fattyacidstandards iscomparedintermsoftheirchromatographicresolutionwiththeemphasisondistinguishingisobaric interferencesandthemethodsensitivity.UHPLCprovidesthebaselineseparationof24isobaricoxylipins within13min,whileUHPSFCenablestheseparationofonly20isobaricoxylipinswithin8min.More- over,theUHPLC/ESI–MSmethodprovidesanaverageimprovementofsensitivityby3.5-fold.Asimilar trendisobservedintheanalysisofhumanplasmasamples,butlowerionsuppressioneffectscaused bylysophospholipids(LPL)areobservedincaseofUHPSFC/ESI–MSduetobetterseparationofLPL.Both methodsarefullyapplicablefortheanalysisofoxylipins,butUHPLC/ESI–MSmethodispreferreddue tobetterseparationandhighersensitivity,whichresultsintheidentificationof31oxylipinsinhuman plasmabasedonavailablestandardsandadditionaltentative20identificationsbasedonaccuratem/z valuesandthefragmentationbehaviorknownfromtheliterature.

1. Introduction

Oxylipins, including eicosanoids, docosanoids, and octadecanoids,areanimportantfamily oflipids,whichareformedby theoxidationofpolyunsaturatedfattyacids(PUFA)[1].Theyare derivedmainlyfromarachidonicacid(AA),docosahexaenoicacid, eicosapentaenoic acid and dihomo-␥-linolenic acid. PUFAs are mostlygeneratedfromglycerophospholipids(PL)bytheenzyme phospholipase A2, and then they are further metabolized to eicosanoidsthroughenzymaticprocessesbyfreeseparateenzyme families,suchascyclooxygenase(COX),lipoxygenase(LOX),and cytochrome P450(CYP). In addition,PUFA canbe autooxidized formingbioactivelipidsvianonenzymaticpathwaysaswell[1,2].

Inhumans,eicosanoidshaveawiderangeofphysiologicaleffectsin inﬂammation,cardiovascularprotection,bloodclotting,andapo- ptosis.Theimportantbiologicalrolesandexcretionoftheselipid mediatorsintobodyﬂuids,suchasblood,urine,andcerebrospinal

∗Correspondingauthor.

E-mailaddress:Michal.Holcapek@upce.cz(M.Holˇcapek).

ﬂuid,makethempotentialbiomarkertargets[2–6].However,the concentrationofeicosanoids(pmol/mLrange)inhumanplasma is much lower compared toother endogenous lipids(nmol/mL range), such as glycerolipids, PL, sphingolipids, ceramides, and sterols.Thereforetheiranalysisrequiresmoresensitiveanalytical methods[7].

In past decades, the analysis of oxylipins was mainly per- formedbyimmunoassaytechniques,suchasradioimmunoassays (RIA) and enzyme immunoassays(EIA). The main disadvantage ofthesemethodsisthelimitedapplicabilityforplasmaandtis- suesamplesduetoimmunologicalcross-reactivityanddecreased sensitivity due to protein – eicosanoid interactions [2]. Gas chromatography–massspectrometry(GC/MS)provideshighersen- sitivityandbetterselectivitythanRIAandEIA,buttheessential derivatizationstepduringthesamplepreparationistimeconsum- ingandlaborious[8,9].Nowadays,liquidchromatography–mass spectrometry (LC/MS) is the most common technique used to simultaneouslyanalyzeeicosanoidsandotheroxylipins,especially in targetedtandemmassspectrometry(LC/MS/MS) modeusing triple quadrupole(QqQ)instrumentswiththeselectedreaction monitoring(SRM)scanningmode[8],becauseit providesaccu-

(2)

108 R.Berkeczetal./J.Chromatogr.A1511(2017)107–121

ratequantitativeinformationduetohighsensitivityandselectivity withoutderivatizationstepand theopportunityof onlinesam- pleextraction[9].Massspectrometry playsa majorrole inthe identificationandquantitationofeicosanoids.Negative-ionelec- trosprayionization(ESI)isthemostsensitiveionizationmodefor eicosanoidsduetothepresenceofcarboxylicacid[10].Thecom- prehensivenontargetedanalysisofeicosanoidsusingquadrupole time-of-flight(QTOF)massspectrometerprovidesmorequalita- tiveinformationonanalyzedsamples,suchastheidentification ofnewcompounds,informationonmatrixorotherendogenous compounds,whileSRMscansonQqQarebetterforsensitivequan- titation[8–10].

TheLCseparation is themostwidespreadtechnique forthe analysisofwiderangeofoxylipins.Reversed-phase(RP)modeper- formedmainlyonoctadecylsilica(C18)oroctylsilica(C8)column providesthehighestselectivityfortheresolutionofisobaricoxylip- ins[2,6,8,9,11,12,14–18].Oxylipinsaredeterminedinvariousbody fluids,suchas amnioticfluid, cerebrospinalfluid, urine, serum, and plasma [9,11–18]. LC/MS/MSmethod hasbeen applied for thesimultaneousdeterminationof32oxylipinsinhumanplasma samplesin29minusingC8columnandammoniumformate–ace- tonitrilegradient[12].Vreekenandcoworkershavedetected36 humanplasmaoxylipinsin26minbyUHPLC/MS/MSmethodusing C18columnandaqueousaceticacid–2-propanol–acetonitrile gradient[16].Recently,UHPLC/MS/MSmethodhasbeenusedto identifyandquantify60endogenousoxylipinsinhumanplasma usingC18columnin5minusingaqueousaceticacid–acetonitrile –2-propanolgradient[17].Faccioliandcoworkershaveprofiled 22oxylipinsinhumanplasmawithin30minwithUHPLC/MS/MS methodusingC18columnandaqueousaceticacid–acetonitrile– 2-propanolgradient[18].

Supercriticalﬂuidchromatography(SFC)successfullycombines advantagesofgasandliquidchromatography,andespeciallythe novelapproach basedontheuseofsub–2␮mparticlecolumns yieldsultrafastandefﬁcientseparations.Thistechniqueisreferred as UHPSFC by the analogy to UHPLC, and may be easily cou- pledtomassspectrometryaswell(UHPSFC/MS)[19].Applications forvarious compoundsclasses illustratethepowerofthis new approach,e.g.,exampleforsyntheticcannabinoids[20],pharma- ceuticalcompounds[21],phospholipids,andsphingolipids [22].

However,UHPSFC/MSmethodfortheanalysisofoxylipinshasnot beenreportedsofar.

ThemaingoalofthisstudyisacomparisonofUHPLC/ESI–MS andUHPSFC/ESI–MSmethodsintermsoftheirsuitabilityforthe analysisofwiderangeofoxylipinsinhumanplasmasamplesusing thesetof46standardstypicallyoccurringinbiologicalsystems.All parametersarecarefullyoptimizedtoachievethebestselectivity andsensitivityoftheﬁnalmethod,whichisthenappliedforthe analysisofendogenousoxylipinsinthehumanplasmaextract.

2. Experimental

2.1. Chemicalsandstandards

Methanol,acetonitrile,2-propanol,ethanol(allLC/MSgrade), chloroformstabilizedby0.5–1%ethanol(HPLCgrade),ammonium formate,ammoniumacetate,formicacid,andaceticacidwerepur- chasedfromSigma-Aldrich(St.Louis,MO,USA).Deionizedwater waspreparedwithaMilli-QReferenceWaterPuriﬁcationSystem (Molsheim,France).Carbondioxide4.5grade(99.995%)waspur- chasedfromMesserGroupGmbh(BadSoden,Germany).

8R-hydroxy-4Z,6E,10Z-hexadecatrienoic acid (tetranor- 12-HETE), 12S-hydroxy-5Z,8E,10E-heptadecatrienoic acid (12-HHTrE), 13-oxo-9Z,11E-octadecadienoic acid (13- OxoODE), 9-oxo-10E,12Z-octadecadienoic acid (9-OxoODE),

13S-hydroxy-9Z,11E,15Z-octadecatrienoic acid (13-HOTrE), (±)13-hydroxy-9Z,11E-octadecadienoic acid (13-HODE), (±)9-hydroxy-10E,12Z-octadecadienoic acid (9-HODE), (±)12,13-dihydroxy-9Z-octadecenoic acid (12,13-DiHOME), (±)15-hydroxy-5Z,8Z,11Z,13E,17Z-eicosapentaenoic acid (15- HEPE), (±)5-hydroxy-6E,8Z,11Z,14Z,17Z-eicosapentaenoic acid (5-HEPE),(±)11,(12)-epoxy-5Z,8Z,14Z-eicosatrienoicacid(11,12- EET), (±)5,6-epoxy-8Z,11Z,14Z-eicosatrienoic acid (5,6-EET), (±)12-hydroxy-5Z,8Z,10E,14Z-eicosatetraenoic acid (12-HETE), (±)15-hydroxy-5Z,8Z,11Z,13E-eicosatetraenoic acid (15-HETE), (±)11-hydroxy-5Z,8Z,12E,14Z-eicosatetraenoic acid (11-HETE), (±)5-hydroxy-6E,8Z,11Z,14Z-eicosatetraenoic acid (5-HETE), 15S-hydroxy-8Z,11Z,13E-eicosatrienoic acid (15-HETrE), 9S- hydroxy-11,15-dioxo-2,3,4,5-tetranor-prostan-1,20-dioic acid (tetranor-PGDM), 9-oxo-15S-hydroxy-5Z,10Z,13E-prostatrienoic acid (PGA2), 11-oxo-15S-hydroxy-5Z,9,13E-prostatrienoic acid (PGJ2), 9S-hydroxy-11-oxo-5Z,12E,14E-prostatrienoic acid (15-deoxy-␦-12,14-PGD2), 15S-hydroxy-9-oxo- 5Z,8(12),13E-prostatrienoic acid (PGB2), (±)5,6-dihydroxy- 8Z,11Z,14Z,17Z-eicosatetraenoic acid (5,6-DiHETE), 5S,15S-dihydroxy-6E,8Z,10Z,13E-eicosatetraenoic acid (5,15- DiHETE), 8S,15S-dihydroxy-5Z,9E,11Z,13E-eicosatetraenoic acid (8,15-DiHETE),5S,12R-dihydroxy-6Z,8E,10E,14Z-eicosatetraenoic acid (LTB4), 5S,12R-dihydroxy-6E,8E,10E,14Z-eicosatetraenoic acid (6-trans LTB4), (±)14,15-dihydroxy- 5Z,8Z,11Z-eicosatrienoic acid (14,15-DiHETrE), (±)5,6-dihydroxy-8Z,11Z,14Z-eicosatrienoic acid (5,6-DiHETrE), 6-oxo-9S,11R,15S-trihydroxy-2,3-dinor-13E-prostaenoic

acid, sodium salt (2,3-dinor-6-keto-PGF1␣), 14S-hydroxy- 4Z,7Z,10Z,12E,16Z,19Z-docosahexaenoic acid (14-HDoHE), (±)4-hydroxy-5E,7Z,10Z,13Z,16Z,19Z-docosahexaenoic acid (4- HDoHE), 9S,15S-dihydroxy-11-oxo-5Z,13E,17Z-prostatrienoic acid (PGD3), 9,15-dioxo-11R-hydroxy-5Z-prostenoic acid (13,14-dihydro-15-keto-PGE2), 9S,11R-epidioxy- 15S-hydroxy-5Z,13E-prostadienoic acid (PGH2), 9-oxo-11R,15S-dihydroxy-5Z,13E-prostadienoic acid (PGE2), 9S,15S-dihydroxy-11-oxo-5Z,13E-prostadienoic acid (PGD2), 9S,11R-dihydroxy-15-oxo-5Z,13E-prostadienoic acid (15-keto-PGF2␣), 9S,11S-dihydroxy-15-oxo- 5Z-prostenoic acid (13,14-dihydro-15-keto-PGF2␣), 9S,11R,15S-trihydroxy-5Z,13E-prostadienoic acid (PGF2␣), 9S,11R,15S-trihydroxy-5Z,13E-prostadienoic acid-cyclo[8S,12R] (8-iso-PGF2␣), (±)19,20-dihydroxy- 4Z,7Z,10Z,13Z,16Z-docosapentaenoic acid (19,20-DiHDPE), 9␣,11,15S-trihydroxythromba-5Z,13E-dien-1-oic acid(TXB2), 6- oxo-9S,11R,15S-trihydroxy-13E-prostenoicacid(6-keto-PGF1␣), 7S,8R,17S-trihydroxy-4Z,9E,11E,13Z,15E,19Z-docosahexaenoic acid (resolvin D1), and 5S-hydroxy,6R-(S- cysteinyl),7E,9E,11Z,14Z-eicosatetraenoic acid (LTE4) standards (see Table S1 for structures) were purchased from Cay- man Chemical (Ann Arbor, MI, USA). Fatty acid standards 6Z,9Z,12Z-octadecatrienoic acid (␥-linolenic acid), and 9Z,12Z,15Z-octadecatrienoic acid (␣-linolenic acid) were purchasedfromNu-Chek-Prep(Elysian,MN,USA).

TheoxylipinsnomenclatureandabbreviationsfollowtheLIPID MAPSstructuredatabasesystem[23].Humanplasmasampleswere obtainedfromhealthyvolunteersbasedontheethicalagreement.

2.2. Samplepreparation

Thestocksolutioncontaining46 eicosanoidand2 fattyacid standardsattheconcentrationof500pg/␮Lexceptfortetranor- PGDM(2500pg/␮L),2,3-dinor-6-keto-PGF1␣(2500pg/␮L),PGD3 (2500pg/␮L),5,6-EET(1000pg/␮L),PGH2(2500pg/␮L),andLTE4 (5000pg/␮L)waspreparedinethanol.ForUHPLCmeasurements, the stock solution was dried under nitrogen at ambient tem-

(3)

perature and redissolved in acetonitrile – water – acetic acid (45/55/0.02,v/v)mixturetopreparesuitableconcentration.Incase ofUHPSFCanalysis,workingsolutionswerepreparedinchloroform –ethanol(50/50,v/v)mixture.

Thesolidphaseextraction(SPE)wasperformedaccordingto themethoddescribedbySchebbetal.[24]withminormodiﬁca- tions.Aliquotsof500␮Lcontrolplasmaweredilutedwith1.5mL 90% aqueousmethanol and vortexedfor 10s. Thentheplasma solutionwascentrifuged(Hettich^® EBA20centrifugeEBA20)at 6000rpmfor10minatambienttemperature.Theoxylipinswere extractedusingStrata^TM-X33␮m(200mg/3mL)polymericRPcar- tridges (8B-S100-FBJ-T, Phenomenex,Aschaffenburg, Germany).

Cartridgeswereconditionedwith3mLofmethanolandthenwith 3mLofwater.Next,2mLofsampleswereloadedoncartridges, withasubsequentwashusing3mLofwater.Thenoxylipinswere elutedwith3mLofmethanol.Theeluentwasdriedundernitrogen atambienttemperatureandredissolvedin50␮LoftheeluentA (acetonitrile–water–aceticacid(45/55/0.02,v/v/v)mixture)for UHPLCor50␮Lchloroform–ethanol(50/50,v/v)mixtureincase ofUHPSFCmeasurements.Finally,thesolutionwascentrifugedat 6000rpmfor1minpriortotransferringinto300␮Lconicalinsert.

2.3. UHPLC/MSandUHPSFC/MSconditions

UHPLC/MSanalyseswereperformedonaliquidchromatograph Agilent 1290 Infinity series (Agilent Technologies, Waldbronn, Germany)consistingofanAgilent1290InfinityLCbinarypump,an Agilent1290InfinitywellplateautosamplerandanAgilent1290 Infinitycolumnthermostat.TheUHPLCsystemwascoupledtothe WatersSynaptG2-Siquadrupole–ionmobility–timeofflightmass spectrometerwithESI(Waters,Milford,MA,USA).

TheﬁnalUHPLC/MSmethodfortheanalysisofoxylipinswas asfollows:AcquityUPLCBEHC18column(150×2.1mm,1.7␮m, 130Å, Waters),injectionvolume1␮L,and columntemperature 40^◦C. The mobile phase A consisted of acetonitrile – water – aceticacid(45/55/0.02,v/v/v)mixture,and theeluent Bwas2- propanol–acetonitrile(50/50,v/v)mixture.Thegradientprogram wasthefollowing:0min–2%B,12min–62%B,12.1min–99%B, 14.5min–99%B,15min–2%B,and20min–2%B.Theﬂowrate was0.4mL/minduringtheanalysis(from0to13.5min)andthe equilibrationbeforethenextinjection(from18.5to20min),and 0.65mL/minduringthecolumnwashingaftertheanalysis(from 13.5to18.5min).Theinjectorneedlewaswashedwithstrong(hex- ane–2-propanol–water(2:2:1,v/v/v)mixture)andweaksolvents (Aeluent)aftereachinjection.

AllUHPSFC/MS experimentsweredoneon anAcquity UPC² system(Waters,Milford,MA,USA) equippedwitha binarysol- vent manager deliverypump,a samplemanager maintainedat 8^◦C,10␮Linjectionloop,acolumnoven,abackpressureregula- tor,andaWatersModel515pumpfordeliveringmake-upsolvent.

Thechromatographicsystemwascoupledtothemassspectrome- terviadedicatedinterfacekit(Waters)composedoftwoT-pieces enablingthebackpressurecontrol,andmixingofcolumnefﬂuent withamake-upsolvent.

ThefollowingUHPSFCcolumnswiththesamecolumndimen- sion(100×3mm,1.7␮m) weretested,i.e.,AcquityUPC2Torus 1-Aminoanthracene(1-AA),AcquityUPC2Torus2-Piycolylamine (2-PIC),Acquity UPC2Torus Diol (1-DIOL), Acquity UPC2Torus Diethylamine(DEA),andAcquityUPC2HSSC18SB(1.8␮m)(HSS C18).

The final UHPSFCmethod for the analysisof oxylipins was asfollows: Acquity UPC² Torus 1-Aminoanthracene(1-AA)column(100×3mm,1.7␮m,Waters),activebackpressureregulator (ABPR)pressure1800psi(124bar),flowrate1.7mL/min,injection volume1␮L,and column temperature50^◦C.The modifier0.1%

aceticacidinmethanolwasusedastheeluentB.Thegradientpro-

gramwasthefollowing:0min–4%B,10min–30%B,11min– 30%B,11.5min–4%B,and15min–4%B.Theinjectorneedlewas washedwithhexane–2-propanol–water(2:2:1,v/v/v)mixture aftereachinjection.Puremethanolwasdeliveredasamake-up eluentattheﬂowrateof0.3mL/min.

Themassspectrometerwasoperatedinthenegative-ionESI sensitivitymodeunderthefollowingconditions:massrangeofm/z 50–950,capillaryvoltage2.2kV,sourcetemperature150^◦C,sam- plingcone20V,sourceoffset90V,dryingtemperature500^◦C,cone gasﬂowrate50L/h,desolvationgasﬂowrate1000L/h,andneb- ulizergaspressure4bar.Leucineenkephalinewasusedasalock massforallexperiments.Thedatadependentacquisition(DDA) experimentswereperformedonatransfercellwiththecollision energyrampfrom17to40eVwiththeselectionofupto3ions.

ThesettingofmassspectrometerwasidenticalforbothUHPLC/MS andUHPSFC/MSmeasurementsexceptfordecreaseddesolvation gasﬂowrate(800L/h)usedforUHPSFC/MSanalysisinorderto achievethestablespray.

2.4. Dataprocessing

TheUHPLCsystemwascontrolledwithAgilentOpenLABsoft- ware. The control of UHPSFC system,MS data acquisition and processingwereconductedbyMassLynxV4.1SCN901andinte- gratedTargetLynxsoftwares.Calculationswereperformedusing MicrosoftExcelsoftware.

3. Resultsanddiscussion

Asetof46oxylipinstandardsrepresenting36eicosanoids,4 docosanoids,6octadecanoids,and2fattyacids(seeTableS1for thecompletelistandstructures)wasusedfortheoptimizationof UHPLC/MSandUHPSFC/MSconditions.Thepreviousoccurrenceof thesestandardsinhumanplasmawasanimportantaspectduring theirselectionbasedonpublishedworks[7,12,16–18].Chromato- graphic and mass spectrometricparameters were optimized to improvetheseparationandsensitivity.TheﬁnalUHPLC/MSand UHPSFC/MS methods were compared, and the former one was appliedforproﬁlingofhumanplasmaoxylipins.TheMSsensitivity andchromatographicresolutionwerethemainaspectsduringthe optimizationduetolownaturalabundancesofoxylipins.

3.1. OptimizationofUHPLC/MSmethod 3.1.1. Columnselectionandinitialconditions

TheAcquityUPLCBEHC18columnwasselectedfortheRPanal- ysisofoxylipinsbasedontheliterature[14,17,25].Thecoelution ofoxylipinisomersmakestheiridentiﬁcationandquantiﬁcation complicated,thereforelongercolumn(150mm)wasselectedto improve the isomeric separation. The starting conditions were based onthemobilephase consisted ofeluent A acetonitrile– water–aceticacid(60/40/0.02,v/v/v),eluentB2-propanol–acetonitrile(50/50,v/v),andthegradientprogramfrom0min–0.1%

B to6min– 55.5% B, but toolow retention was observed for tetranor-PGDM, 2,3-dinor-6-keto-PGF1␣, PGD3, 8-iso-PGF2␣, PGF2␣,6-keto-PGF1␣,PGE2,PGH2,andTXB2(retentionfactors lowerthan1.2).Therefore,theamountofwaterintheeluentAwas increasedto55%watertoimprovetheirretention.Inparallel,the gradientprogramwasalsochanged(0min–0.1%B,6min–70.0%B, 6.75min–99.0%B,7.50min–99%B,7.65min–0.1%B,and15min –0.1%B).

3.1.2. Effectofmobilephasecomposition

EffectsofadditivesintheeluentAontheretentionbehaviorand MSsensitivityof oxylipinswerestudiedforthefollowingaddi- tives:aceticacid(0.02and0.05%),formicacid(0.02and0.05%),

(4)

Fig.1.Effectofadditivesinacetonitrile–water(45/55,v/v)mobilephasesonnegative-ionUHPLC/ESI–MSextractedionchromatogramsofthestandardmixtureofisomeric HETE/EETforthedifferentcompositionofeluentA:(A)5mMammoniumacetate,(B)5mMammoniumformate,(C)0.05%formicacid,(D)0.02%formicacid,(E)0.05%acetic acid,(F)0.02%aceticacid,and(G)withoutanyadditive.Conditions:eluentB–2-propanol–acetonitrile(50/50,v/v),AcquityUPLCBEHC18column(2.1×150mm,1.7␮m, 130Å,Waters),ﬂowrate0.5mL/min,columntemperature40^◦C,andgradientprogram0min–0.1%B,6min–70.0%B,and6.75min–99.0%B.

5mMammoniumacetate,and5mMammoniumformate(Fig.1).

Thelowestretentionwasobservedwithammoniumacetate,while aceticandformicacidsresultedinthehighestretentiontimesat 0.05%.Theuseofaceticandformicacidsatthesameconcentration (0.02%or0.05%)resultedindifferentpHvaluesofthemobilephase, buttheretentionandresolutionofisomersdidnotchangesigniﬁ- cantly.Forallcompounds,higherretentiontimeswereobserved for higher acid content (0.05%). The comparison of chromato- graphicresolutionforisomersobtainedwithandwithoutadditives showthathigherresolutionwasachievedwithaceticacid,formic acidand without any additive,except for ␥-linolenic acid, ␣- linolenicacid,15-HETE,11-HETE,12-HETE,5-HETE,11,12-EET, 5,15-DiHETE,6-trans LTB4, 14-HDHA, and 4-HDHA,where the mobilephaseswithammoniumacetateprovided betterresults.

Peakshapeswerenot inﬂuencedconsiderablybyadditives.The seriouspeaktailingwasobservedforLTE4intheabsenceofany additiveand inammoniumacetateorammoniumformatecon- tainingmobilephases.Thepeakshapewasimprovedathigheracid concentrations,whichcanbeexplainedbysuppressedzwitterion formationofcysteinylsidechain(isoelectricpointatpH5.07)in moreacidicenvironment[26].Theoxalicacidtreatmentofsta- tionaryphaseusingtheinjectionof20␮Lof10mMoxalicacidtwo times[27]hadalsoasubstantialeffectontheimprovementofpeak shapeofLTE4.

Fig. 2 demonstrates the dependence of negative-ion ESI responseonthetypeandconcentrationofadditives.Aceticacid

Fig. 2.Dependencies of peak areas of oxylipins standards in negative-ion UHPLC/ESI–MSonthecompositionofeluentAcontainingvariousadditivesincom- parisontoconditionswithoutanyadditives:(A)5mMammoniumacetate,(B)5mM ammoniumformate,(C)0.05%formicacid,(D)0.02%formicacid,(E)0.05%acetic acid,and(F)0.02%aceticacid.OtherconditionsareidenticalasforFig.1.

(5)

Table1

AveragerelativechangesinUHPLCforretentiontimes(tR),peakwidths,peak areas,andresolution(Rs)forincreasedﬂowratesof0.4and0.5mL/minincompar- isontotheinitialvalueof0.3mL/min.

Flowrate[mL/min] tR peakwidth peakarea Rs

0.4 −33% −21% −31% −6%

0.5 −54% −33% −51% −14%

enhancedtheionizationcomparedtothemobilephasewithout anyadditive(ca.90%ontheaveragefor0.02%and40%for0.05%of aceticacid).Theuseofformicacidcausedthesignalsuppression (ca.15%for0.02%and45%for0.05%offormicacid).Theadditionof ammoniumacetateorformateresultedinthedrasticlossofsen- sitivity.0.02%ofaceticacidwasselectedasoptimalforthefurther optimizationofotherparameters.

Theretentionandselectivityoftheseparationcanbetunedby thecompositionoftheorganicsolventB.FourtypesofeluentsB wereinvestigated,suchas2-propanol–acetonitrile(50/50,v/v), 2-propanol–acetonitrile(10/90,v/v),acetonitrile,andmethanol.

Theconcentrationof2-propanolexhibitedonlyaslighteffecton retention,resolution, and sensitivity.Thehighest retentionwas observedformethanol,butitnegativelyinﬂuencedtheseparation, especiallyincaseof␥-linolenicacid,␣-linolenicacid,13-HODE, 9-HODE,andallHETE.Finally,2-propanol–acetonitrile(50/50,v/v) wasselected,becausethiswasthemostefﬁcientintheremoval of plasma phospholipid contaminants (in spite of SPE sample

treatment)fromthecolumnduringthewashingstepofgradient program.

3.1.3. Effectofﬂowrate

Theincreasedflowratereducestheanalysistime,butonthe otherhanditalsoinfluencestheresolutionandsensitivity.Fig.3 demonstratestheeffectofflowrateintherangeof0.3–0.5mL/min forisomers15-HETE,11-HETE,12-HETE,5-HETE,11,12-EET,and 5,6-EET.Gradientsfordifferentflowratesarerecalculatedbasedon theratioofsquaredcolumninternaldiameters.Averageretention timesandpeakwidthsweredecreasedforallcompounds,when theflowrateincreased(Table1),butitalsoreducedpeakareas.

Themoderatereductionofresolutionwasobservedwithincreased flowratesforisomers.0.4mL/minischoseninthefinalmethod asacompromiseamongtheshortanalysistime,highselectivity, andESIresponse.Theoxylipinandfattyacidstandardselutedin therangeof1–12min(Fig.4andTable2).Theflowrategradient programwasusedduringthewashingandequilibrationcycleto reducethetotalruntime.

3.2. DevelopmentofUHPSFC/MSmethod 3.2.1. Columnselection

Five UHPSFC columns with the same column dimension (100×3mm, 1.7␮m) and different chemistries were used in our column screening, i.e., 1-AA, 2-PIC, 1-DIOL, DEA, and HSS C18,inorder toﬁndtheoptimalstationaryphase fortheanal-

Fig.3.Effectoftheﬂowrateonnegative-ionUHPLC/ESI–MSextractedionchromatogramsofthestandardmixtureofisomericHETE/EETusing0.02%aceticacidasthe additiveintheeluentA:(A)0.5mL/min,(B)0.4mL/min,and(C)0.3mL/min.Gradientprogramsareadjustedforindividualﬂowrates:(A)0min–0.1%B,6min–70.0%B, and6.75min–99.0%B,(B)0min–0.1%B,7.52min–70.0%B,and8.44min–99.0%B,and(C)0min–0.1%B,10.02min–70.0%B,and11.25min–99.0%B.Otherconditions areidenticalasforFig.1.

(6)

112R.Berkeczetal./J.Chromatogr.A1511(2017)107–121

Table2

ChromatographicandmassspectrometricdataforoxylipinandfattyacidstandardsintheﬁnalUHPLC/MSmethod.

Abbreviation Subclass Isomericgroup Theoreticalm/zof[M−H]⁻ tRinUHPLC/MS[min] RS Indentiﬁedinplasma

tetranor-PGDM Prostaglandins – 327.1449 0.96 No

6-keto-PGF1␣ Prostaglandins 1 369.2283 1.24 1.25 No

TXB2 Thromboxanes 1.53 Yes

8-iso-PGF2␣ Prostaglandins 2 353.2334 1.52 1.53 Yes

PGF2␣ 1.73 4.96 Yes

13,14-dihydro-15-keto-PGF2␣ 2.43 No

2,3-dinor-6-keto-PGF1␣ Prostaglandins – 341.1970 1.71 No

PGD3 Prostaglandins – 349.2021 1.71 No

PGH2 Prostaglandins 3 351.2177 1.88 No

PGE2 1.88 1.34 Yes

15-keto-PGF2␣ 2.03 0.68 No

PGD2 2.08 3.35 Yes

13,14-dihydro-15-keto-PGE2 2.48 No

ResolvinD1 Docosanoids – 375.2177 2.37 No

PGA2 Prostaglandins 4 333.2071 3.15 0.96 No

PGJ2 3.33 No

PGB2 3.33 7.51 Yes

15-deoxy-␦-12,14-PGD2 4.51 No

8,15-DiHETE Hydroxy/hydroperoxyeicosatetraenoic acids

5 335.2228 3.85 2.15 Yes

5,15-DiHETE 4.14 0.71 Yes

6-transLTB4 Leukotrienes 4.22 1.55 Yes

LTB4 4.48 6.33 No

5,6-DiHETE Hydroxy/hydroperoxyeicosatetraenoicacids 5.54 Yes

12,13-DiHOME Otheroctadecanoids – 313.2384 4.75 Yes

19,20-DiHDPE Docosanoids – 361.2384 5.33 Yes

tetranor-12-HETE Hydroxyfattyacids – 265.1809 5.34 No

14,15-DiHETrE Hydroxy/hydroperoxyeicosatrienoic acids

6 337.2384 5.40 8.46 Yes

5,6-DiHETrE 7.18 Yes

12-HHTrE Hydroxy/hydroperoxyeicosatrienoicacids – 279.1966 5.50 Yes

LTE4 Leukotrienes – 438.2320 5.60 No

13-HOTrE Otheroctadecanoids 7 293.2122 6.34 7.77 Yes

13-OxoODE 7.83 1.72 Yes

9-OxoODE 8.22 Yes

15-HEPE Hydroxy/hydroperoxyeicosapentaenoicacids 8 317.2122 6.70 4.94 Yes

5-HEPE acids 7.55 Yes

13-HODE Otheroctadecanoids 9 295.2279 7.56 1.10 Yes

9-HODE 7.73 Yes

15-HETE Hydroxy/hydroperoxyeicosatetraenoic

acids

10 319.2279 7.85 2.23 Yes

11-HETE 8.26 1.20 Yes

12-HETE 8.48 3.39 Yes

5-HETE 9.13 3.05 Yes

11,12-EET Epoxyeicosatrienoic

acids

9.72 2.32 No

5,6-EET 10.16 No

14-HDoHE Docosanoids 11 343.2279 8.17 6.26 Yes

4-HDoHE 9.34 Yes

15-HETrE Hydroxy/hydroperoxyeicosatrienoicacids – 321.2435 8.68 Yes

␣-Linolenicacids Unsaturatedfattyacids 12 277.2173 11.70 0.91 Yes

␥-Linolenicacids 11.90 Yes

(7)

Fig.4.Negative-ionUHPLC/ESI–MSextractedionchromatogramsofstandardmixtureobtainedusingﬁnalconditions:AcquityUPLCBEHC18column(2.1×150mm,1.7␮m, 130Å,Waters),ﬂowrate0.4mL/min,columntemperature40^◦C,eluentAacetonitrile–water–aceticacid(45/55/0.02,v/v/v),eluentB2-propanol–acetonitrile(50/50,v/v), andgradientprogram0min–2%B,12min–62%B,and12.1min–99%B.DetailsareintheExperimental.

ysis of oxylipins. The following parameters were chosen for theinitialoptimization: eluent B10mM ammoniumacetate in methanol,methanolasthemake-upeluent,ABPRpressure1800psi (124.1bar),ﬂowrate1.0mL/min,injectionvolume1␮L,column temperature50^◦C,andthegradientprogram0min–1%B,5min – 50% B, 12min – 50% B, and 13min– 1% B. First, the reten- tionbehaviorof selectedstandardsrepresentingdifferenttypes ofisomerismwasstudied(PGA2,15-deoxy-␦-12,14-PGD2,PGB2, PGD2,PGE2,PGF2␣,8-iso-PGF2␣,TXB2,and6-keto-PGF1␣).For allcompounds,thelowestretentionwasobservedonHSSC18col- umncompared withotherstationaryphases,moreoverPGF2␣, 8-iso-PGF2␣,PGD2,andPGE2arenotseparated.Theobtainedlow retentionand poorselectivityindicatedtheimportanceofelec- trostaticattractionbetweenthecarboxylategroupfromanalytes andthebasicfunctionalityinthestationaryphase,whichisnot occurringatalloronlywithmild effectsonHSS C18stationary phase due tothelack of additionalionizable functional groups besideresidualsilanols(pKa=7–8)[28,29].DEAhadthemostbasic functionality(calculatedbasicpKa=9.5),providingstrongelectro- staticinteractionswithdeprotonatedcarboxylgroupsofoxylipins, thusresultinginthehighestretentiontimesforallanalytes,and generallyaccompaniedbyhighestpeakwidth,whichresultedin thelowerresolution.HSSC18andDEA columnswereexcluded fromfurtheroptimizationbasedonabovementionedobservations.

Comparableretentionandselectivitywereobservedontheremain- ingthreestationaryphases,thereforethosewerefurthertested.

1-DIOLprovidedlowerretentionofoxylipinscomparedtothebasic

stationaryphases.Lowerretentiontimeswereobservedforallana- lyteson1-AArelativeto2-PICcolumn[29].Fig.5demonstrated that1-AAstationaryphaseprovidedasuperiorchromatographic resolution for 11,12-EET, 12-HETE, 15-HETE, 11-HETE, and 5- HETEisomersover1-DIOLand2-PICcolumns.For␥-linolenicacid and␣-linolenicacid,thepartialseparationwasobtainedonlyon 1-AAstationaryphase.The bulkyanthraceneringmayimprove therecognitionprocessofisomersinatleasttwoways,suchas theformation ofadditional hydrophobicinteractions and steric hindrance. Finally, 1-AA column was selected for the develop- mentofUHPSFC/MSmethodconsideringthechromatographicdata obtainedforHETE,whichexhibitnumerousbiologicallyimportant isomersinplasma.

3.2.2. Effectofﬂowrate

OneofthemainbenefitsofUHPSFCisthefeasibilityofusing highflowrate[30,19],therefore1.0,1.5,1.7,2.0,and2.3mL/min flowratesweretested.Theincreasedtheflowratefrom1.0mL/min resultedinthefollowingaveragereductionsofretentiontimes:36%

at1.5mL/min,44%at1.7mL/min,53%at2.0mL/min,and60%at 2.3mL/min.Theaverageresolutionofisomerschangedonlymod- erately,andthemaximumwasfoundat1.5mL/min.Theresolution wasonlyslightly lowerat 1.7mL/min.Theincrease oftheﬂow ratefrom1.5to1.7mL/mindecreasedpeakwidths,especiallyfor tetranor-PGDMelutingatthebeginning.Thedecreaseofobserved peakareaswasverysmall(only3%).Therefore, 1.7mL/minwas usedintheﬁnalmethod.

(8)

Fig.5.Effectofthecolumntypeonnegative-ionUHPSFC/ESI–MSextractedionchromatogramsofthestandardmixtureofisomericHETE/EET:(A)AcquityUPC²Torus 1-Aminoanthracene(1-AA),(B)AcquityUPC²Torus2-Pycolylamine(2-PIC),and(C)AcquityUPC²TorusDiol(1-DIOL)columns.Conditions:ABPRpressure124bar,flowrate 1mL/min,columntemperature50^◦C,modifier10mMammoniumacetate,gradientprogram0min–1%B,10min–50%B,andmake-uppumpflowrate0.2mL/min.

Theeffect offlow rate ofmake-up eluent onthe sensitivity wasalsostudied,andthetrendofaveragesignalreductionwith increasingflowratefrom0.15mL/minwasthefollowing:10%at 0.2mL/min,13%at0.25mL/minand20%at0.30mL/min.Despite thelowersensitivityathigherflowrate,0.30mL/minisessential toachievestablesprayinESI–MS,especiallyatthebeginningof analysiswithlowmodifiercontent.

3.2.3. Effectoftemperatureandbackpressure

TheselectivityandmainlytheretentioninUHPSFCcanbeinflu- encedbythedensityofmobilephasethroughtheregulationof temperatureandbackpressure[31].Inorder tostudytheeffect oftemperatureonthechromatographicbehaviorofoxylipins,40, 50,and60^◦Cwereselectedat124barbackpressure.Forallcom- pounds,higherretentiontimesareobservedathighertemperature, whiletheseparationofisomerswasnotinfluencedsignificantly.

Theeffectoftemperatureonthechangeofretentionwasreduced byincreasingthecontentoforganicmodifierinthemobilephase duringthe gradient program (Fig. S1). Thedistribution of data correlateswithgraphsobtainedfordiffusioncoefficientsofantho- cyaninsdependenceonthemethanolcontentinthemobilephase [32]. This retention behavior might be explained by decreased diffusioncoefficientsofoxylipinswithincreasedorganicsolvent contentandprobablybythetransitionfromsupercriticaltosub- criticalstateintherangeof12–16%methanol[32,33].Thesimilar chromatographicbehavior wasfoundwithdecreased backpres- surefrom138to103bar.Thecolumntemperatureof50^◦C and thebackpressureof124barwereselectedinthefinalmethod.

3.2.4. Effectofgradientsteepness

TheretentionandselectivityinUHPSFCcanbeinfluencedby thegradientsteepnesssimilarlyasforotherchromatographictech- niques [31].30% of organicmodifier providessufficient elution strengtheven for theelutionofmore retainedpolarstandards, forexampletetranor-PGDMand2,3-dinor-6-keto-PGF1␣.Thus, therangefrom4%to30%oftheeluent Bwasusedfortheini- tialoptimizationstep.Inordertoinvestigatetheeffectofgradient steepnessonthechromatographicretention,astudywascarried outwiththegradientslopeof2.6,5.2,and10.4%ofB/min.Retention timesincreasedwithdecreasedgradientsteepness.Thisincrease was36%at5.2%ofB/minand86%at2.6%ofB/minontheaver- age.For theseparationofisomers,asimilartrendwasfoundin theincreaseofresolution(57%and152%)withthedecreaseofthe gradientsteepness.

3.2.5. Selectionoforganicco-solventandmobilephaseadditives Thepolarityofsupercriticalcarbondioxideissimilartohex- ane,thustheuseofpolarorganicmodifierinthemobilephaseis essentialfortheelutionofmorepolarcompounds[19].Methanol isthemostcommonorganicmodifierinUHPSFCowingtoitsphys- icalandchemicalproperties,suchasitshighelutionstrength,and thecomplete miscibility withsupercritical fluid carbondioxide atgiventemperatureandpressureranges.Moreover,themobile phaseadditives(e.g.,ammoniumacetateorformate)arewellsol- ubleinmethanolattypicallyusedconcentrations[31].Methanol with10mMofammoniumacetatewasusedastheorganicmodi- fierintheinitialstageofoptimization.Theeffectofotherorganic

(9)

Fig.6. Effectofdifferentadditivesinmethanolasamodiﬁeronnegative-ionUHPSFC/ESI–MSextractedionchromatogramsofthestandardmixtureofisomericHETE/EET:

(A)10mMammoniumacetate,(B)10mMammoniumformate,(C)0.02%formicacid,(D)0.02%aceticacid,and(E)withoutanyadditives.Conditions:1-AAcolumn,ABPR pressure124bar,ﬂowrate1.7mL/min,columntemperature50^◦C,gradientprogram0min–4%B,10min–30%B,andmake-uppumpﬂowrate0.25mL/minofmethanol.

modiﬁersin methanolwastestedfor10%or20%of acetonitrile or 2-propanol in methanol.For allcompounds, theaddition of eitheracetonitrileor2-propanolincreasedtheretentionmoder- ately,whilenoinﬂuenceontheseparationofisomerswasdetected, thereforeonlymethanolisusedinfurtheroptimizationsteps.

Theapplicationofpolaradditives,suchasacid,base,neutralsalt orbufferinthemobilephasecaninfluencetheretention,selectiv- ityandsensitivitythroughdifferentwaysdependingonproperties (mainlyacid–base)ofanalyte,typesofpackedstationaryphaseand thedetectionmode[29,19,31].Consideringbasicpropertiesof1-AA columnandthedetectionofacidicoxylipinsbynegative-ionmode ESI–MS,theeffectsofaceticacid,formicacid,ammoniumacetate, andammoniumformateonthechromatographicbehaviorandthe sensitivitywereinvestigatedandcomparedwiththemobilephase withoutanyadditive.Thenatureofadditivesshowedonlymod- erateeffectsontheretention,whilethesensitivitywasinfluenced considerably.Fig.6revealsthattheuseofdifferentmobilephase additivesfor1,12-EET,12-HETE,15-HETE,11-HETE,and5-HETE isomers,slightly lowerretention was observed for formic acid, whilehigherretentiontimeswereobtainedwhenusingsalts,espe- ciallywithammoniumacetate.Generally,increasedretentionand slightlyimprovedresolutionwereobtainedforincreasedconcen- trationofammoniumacetate(1,5,10,20,and30mM).However, theincreaseofaceticacidcontent(0.01,0.02,0.05,0.1,and0.2%)did notinfluencetheretentionandselectivity.Theeffectofadditives onpeakshapewasnegligible.

Theconcentrationofaceticacidintherange of0–0.2% does notshowanyvisibleeffectontheretention.Fig.6showstherole

ofadditives intheMS response.Theuseofammoniumacetate provided the lowest responses, as general trend for all compounds(47%),whilethehighestpeakareaswereobtainedwithout any additives (i.e.,100%) in methanol. Slightly higher response wasfoundforammoniumformate(61%).Thereductionofsignal wasmoreconsiderableforincreasedconcentrationofammonium acetate,forexamplethemeanintensitywasreducedtoapprox- imately50%withthechangeofconcentrationfrom1to30mM.

Aceticacid(96%) provided a bettersensitivitythan formic acid (85%).Themeanpeakareashowedamaximumat0.1%ofacetic acidwithinthetestedrangefrom0to0.2%.Theuseofbuffersin themobilephaseslightlyimprovedtheseparationofisomers,but ontheotherhanditresultedinaseriouslossofsensitivity,therefore 0.1%aceticacidwasselectedasthemobilephaseadditive.

Theeffectofwatercontent(1,2,and5%)inthemobilephase containing0.1%aceticacidontheretention andselectivitywas studied.Aslightimprovementofpeakshapeswasachievedwith 5%ofwater,buttheESIresponsewasdramaticallyreducedby65%.

Thesimilarsignalreductionwasobservedfor1%waterinmethanol usedasamake-upsolvent,thereforewaterwasnotusedeitherin themobilephaseormake-upsolventinfurtherexperiments.

3.2.6. Effectofsamplesolventandinjectedvolume

Thetypeofsolventandinjectedvolumecaninﬂuencethepeak shape and theretention time aswell. Hexaneand heptaneare thebestchoicefor dissolutionsolventduetothesimilarpolar- itywithsupercriticalcarbondioxide[19].Hexaneandchloroform wereselectedascommonlyusednonpolarsolventsinlipidomics.

(10)

Fig.7. Negative-ionUHPSFC/ESI–MSextractedionchromatogramsofthestandardmixtureobtainedusingfinalconditions:1-AAcolumn,modifier0.1%aceticacidin methanol,ABPRpressure124bar,flowrate1.7mL/min,columntemperature50^◦C,gradientprogram0min–4%B,10min–30%B,andmake-uppumpflowrate0.30mL/min ofmethanol.DetailsareintheExperimental.

However,theadditionofpolarorganicsolventintochloroformwas necessaryformorepolarstandards.Thus,theeffectofmethanol andethanolcontentandtypeofnonpolarsolventswasinvestigated onchromatographicresults.Nosigniﬁcantchangeswereobserved forretentiontimes.Theuseofhexaneandchloroformwith50%

ethanolcontentshowedthatchloroformresultedinnarrowerpeak widths.Concerningcontributionofpolarorganicsolvent(50%)in chloroform,theadditionofethanolresultedinbetterpeakshape thanmethanol,additionallybetterseparationoccurredincaseof lowretainedlinolenicacidisomers.Interestingly,higherchloro- formcontentdidnotprovideanyimprovement inpeakshapes.

Theselectionofchloroform –ethanol(50/50,v/v) wasthebest compromiseconsideringpeak proﬁles,properpolarity,and low volatility.Theinjectedvolumemayhavethemaineffectonthe peakdistortion[19],thereforemeasurementswerecarriedoutfor theinjectionvolumesof1,2,3,and5␮L.Theincreaseofinjection volumeresultedinnochangeintheretention,butitdecreasedthe chromatographicresolution,asexpected.Thistrendwasmoreevi- dentforlessretainedcompounds.Thus,theinjectedvolumewas limitedto1␮L.

TheﬁnalUHPSFC/MSmethodprovidedbaselineseparationfor 20oxylipinstandardswithin7min(Fig.7andTable3).

3.3. ComparisonofUHPLC/MSandUHPSFC/MS

Fig. 8 demonstrated different retention mechanisms with a low degree of orthogonality (r²=0.8) between UHPLC/MS and

UHPSFC/MS methods. Generally, theincrease of the compound hydrophobicityresultedinhigherretentioninRP-UHPLC,whilethe oppositetrendwasobservedinUHPSFC.Adifferentelutionorder betweenUHPLCandUHPSFCwasobservedforisomericgroups1,2, 3,4,5,7,10,and12.For␥-linolenicacid,␣-linolenicacid,11,12- EET,and5,6-EET,thehighestretentionwasobservedinUHPLC, whilelowestretentiontimeswereobtainedinUHPSFC.Resolvin D1, 2,3-dinor-6-keto-PGF1␣, TXB2, and 6-keto-PGF1␣ with 3 hydroxylgroupsprovidedhighestretentiontimesinUHPSFC.For TXB2,broad(inUHPSFC)andtailing(inUHPLC)elutionproﬁlesmay beassociatedwiththeinterconversionbetweenhemiacetaliso- mers[34].Thepresenceofadditionalcarboxylgroupinmorepolar analyte tetranor-PGDM caused thetailingpeak withincreased retentioninUHPSFC.Thecontributionofpolarcysteinylsidechain ofLTE4maycontributetothepeak shapedistortion.Thisunfa- vorableeffectwasmorepronouncedinUHPSFC,whereitdisabled thedetectionofLTE4duetoextremelytailingpeaks.Thedegree ofunsaturationalsoinﬂuenced theretention,mainlyinUHPLC.

Table2 reveals thatfor 15-HETrE, 15-HETE,and 15-HEPE,the reductionin retentiontimescorrelateswiththeincreaseinthe numberofdoublebonds.

Tables 2 and 3 show that UHPLC generally provides better separationofisomersthanUHPSFCmethodexceptforpositional isomeric groups of2,4, 5, and 9.Higher retention and resolu- tionwereobservedforpositionalisomersPGJ2,PGA2,andPGB2 inUHPSFC.ItindicatesthatthepositionofthecarbonylgroupatC9 wasmorefavorablefordipole–dipoleinteractionthanatC11,thus

(11)

R.Berkeczetal./J.Chromatogr.A1511(2017)107–121117 ChromatographicandmassspectrometricdataforoxylipinandfattyacidstandardsintheﬁnalUHPSFC/MSmethod.

Abbreviation Subclass Isomericgroup Theoreticalm/zof[M−H]⁻ tRinUHPSFC/MS[min] RS Indentiﬁedinplasma

tetranor-PGDM Prostaglandins – 327.1449 7.03 No

TXB2 Thromboxanes 1 369.2283 5.36 1.02 Yes

6-keto-PGF1␣ Prostaglandins 5.86 No

13,14-dihydro-15-keto-PGF2␣ Prostaglandins 2 353.2333 4.37 7.23 No

PGF2␣ 5.41 1.75 Yes

8-iso-PGF2␣ 5.66 Yes

2,3-dinor-6-keto-PGF1␣ Prostaglandins – 341.1970 5.90 No

PGD3 Prostaglandins – 349.2021 4.82 No

13,14-dihydro-15-keto-PGE2 Prostaglandins 3 351.2177 4.03 4.81 No

PGH2 4.71 No

PGE2 4.71 0.68 Yes

PGD2 4.77 0.78 Yes

15-keto-PGF2a 4.83 No

ResolvinD1 Docosanoids – 375.2177 6.30 No

PGJ2 Prostaglandins 4 333.2071 3.61 1.02 No

PGA2 3.73 3.07 No

15-deoxy-␦-12,14-PGD2 4.08 2.10 No

PGB2 4.28 Yes

5,6-DiHETE Hydroxy/hydroperoxyeicosatetraenoic acids

5 335.2228 3.91 9.12 Yes

5,15-DiHETE 4.60 0.73 Yes

8,15-DiHETE 4.67 2.40 Yes

LTB4 Leukotrienes 5.02 1.32 No

6-transLTB4 5.22 Yes

12,13-DiHOME Otheroctadecanoids – 313.2384 3.47 Yes

19,20-DiHDPE Docosanoids – 361.2384 3.86 Yes

tetranor-12-HETE Hydroxyfattyacids – 265.1809 2.78 No

14,15-DiHETrE Hydroxy/hydroperoxyeicosatrienoic acids

6 337.2384 3.37 2.60 Yes

5,6-DiHETrE 3.82 Yes

12-HHTrE Hydroxy/hydroperoxyeicosatrienoicacids – 279.1966 2.87 Yes

LTE4 Leukotrienes – 438.2320 n.d. No

13-OxoODE Otheroctadecanoids 7 293.2122 2.10 0.44 Yes

9-OxoODE 2.13 6.29 Yes

13-HOTrE 2.72 Yes

15-HEPE Hydroxy/hydroperoxyeicosapentaenoic

acids

8 317.2122 2.69 4.09 Yes

5-HEPE 3.23 Yes

13-HODE Otheroctadecanoids 9 295.2279 2.68 1.48 Yes

9-HODE 2.85 Yes

11,12-EET Epoxyeicosatrienoic

acids

10 319.2279 1.70 1.09 No

5,6-EET 1.84 6.18 No

12-HETE Hydroxy/hydroperoxyeicosatetraenoic

acids

2.54 0.96 Yes

15-HETE 2.64 0.90 Yes

11-HETE 2.75 2.76 Yes

5-HETE 3.14 Yes

14-HDoHE Docosanoids 11 343.2279 2.90 3.78 Yes

4-HDoHE 3.43 Yes

15-HETrE Hydroxy/hydroperoxyeicosatrienoicacids – 321.2435 2.78 Yes

␥-Linolenicacids Unsaturatedfattyacids 12 277.2173 1.27 0.42 Yes

␣-Linolenicacids 1.30 Yes

(12)

Fig.8.CorrelationofretentiontimesinUHPLC/ESI–MSandUHPSFC/ESI–MSmethodsusingﬁnalconditions.

positionofdoublebondinthecyclopentaneringhadhighercon- tributiontotheseparationofisomers.FordiastereoisomersPGF2␣ and8-iso-PGF2␣andfunctionalisomersof15-keto-PGF2␣and PGF2␣,higherretention wasobservedinUHPSFC, andthiswas accompaniedbyhigherresolution.

Thesensitivityofgivenanalyticalmethodwasalsoanimportant aspectduetolowconcentrationsofoxylipinsinbiologicalsamples.

Approximately3.5-foldhighersensitivityonaveragewasobserved inUHPLC/MSincomparisontoUHPSFC/MS.Figs.4and7withthe samescaleofintensityillustratethedifferentsensitivityinboth methods.InUHPSFC,thesignalreductioncouldbeattributedtothe followingreasons:1/appliedhigherflowrate,2/lowerionization efficiencyduetodecreasedsamplesolubilityduringcarbondioxide expansion,and3/dilutionandsplittingofelutedsamplepriorto MSdetection[35].Thelowerconcentrationofcompoundsinthe effluentmightbecausedbydifferentdiametersbetweenUHPSFC (3mm)andUHPLC(2.1mm)columns[36].

3.4. Identiﬁcationofoxylipinsinhumanplasma

Theenrichmentofoxylipinsthroughpolarandnonpolarmatrix removalisacriticalsteppriortotheiranalysisduetotheirtrace concentrationinbiologicalsamples.Nowadays,SPEisoneofthe mostefﬁcientandwidespreadtechniqueinthesamplepreparation procedure.ThepolymericStrataXRPcolumnshowedagoodrecov- eryfor oxylipins [7,17,24,37].The applicationof relativelyhigh volumeofplasmasamplewasneededduetothelimitedinjection volumeinUHPSFC.Numerousglycerophospholipids(PL),mainly lysospecies,enrichedwithoxylipinsandfattyacidsintheextracts, causedmatrixeffectsduetothecoelutionwithoxylipins,which elutedintherangeof2–7mininUHPSFCand1–10mininUHPLC (Tables2and3).Thisunfavorableeffectwasmorepronouncedin UHPLCowingtocomparableretentionofoxylipinsandLPLsinthe RPseparation.Forexample,highabundantLPI20:4andLPI18:2,

coelutedwithPGE2andPGD2at2.1min.Retentiontimesofother LPEandLPCwerefoundintherangeof8–10min,e.g.,LPE18:2, LPE22:6,LPC16:0,and LPC22:6,whileinUHPSFCtheyeluted after9min,whichreducestheriskofionsuppressioneffects.Fatty acidshadhigherretentiontimesinUHPLC(after9.5min)andlower retentioninUHPSFC(before2.3min),thereforetheriskofionsup- pressioneffectsisrelativelylowinbothcases.

Tables2and3show31oxylipinsidentifiedinhumanplasma basedon48standardsbybothmethods.Consideringbettersepa- rationofisomersandhighersensitivity,theoptimizedUHPLC/MS methodwasselectedfor theidentificationofadditionalplasma oxylipins.Thefurtheridentificationofadditionaloxylipinswith- out availableauthentic standardswas basedon theirretention times, accurate masses of precursor ions (7ppm mass toler- anceatmaximum),andrelatedfragmentionsusinghomemade database created based on own measurements and literature sources [1–7,12,16–18,23,38,39]. Relevant chromatographic and MSdataofadditional20possiblecompoundsinplasmainclud- ingretentiontimes,m/zofprecursorions,andobservedfragment ionscorrespondingtostructuresaregiveninTable4.

4. Conclusions

OptimizedUHPLC/ESI–MSandUHPSFC/ESI–MSmethodshave beenappliedfortheanalysisof46oxylipinsand2fattyacidsstan- dards.InUHPLC,mobilephaseadditivesandtheirconcentrations significantlyinfluencethechromatographicretentionandMSsen- sitivity, especiallytheuseofaceticacidbringsadvantages over ammoniumacetateorformateintermsoftheresolutionandsensi- tivity.TheUHPSFC/ESI–MSmethodforoxylipinsisreportedforthe firsttimehere.1-AAcolumnyieldsabetterseparationofoxylipins thanotherdedicatedsub–2␮mUHPSFCstationaryphases(2-PIC, 1-DIOL, DEA,and HSS C18). The chromatographic resolution of

(13)

R.Berkeczetal./J.Chromatogr.A1511(2017)107–121119 UHPLC/MS/MSdatafortentativelyidentiﬁedoxylipinswithoutgivenstandardsinthehumanplasmasample.

Name Abbreviation Subclass Experimental

m/zof [M−H]⁻

Mass accuracy [ppm]

tRin UHPLC/MS [min]

m/zofobservedfragmentions

5,9S,11R-trihydroxy-6E,14Z-prostadienoicacid-cyclo[8S,12R] 5-iso-PGF2␣VI Isoprostanes 353.2317 4.8 1.89 335,317,309,115

9-oxo-11S,15S-dihydroxy-5Z,13E-prostadienoicacid 11␤-PGE2 Prostaglandins 351.2164 3.7 1.95 333,315,271,189

9,10-dihydroxy-12Z-octadecenoicacid 9,10-DiHOME Otheroctadecanoids 313.2371 4.2 5.14 295,277,201,177

9-hydroxy-10E,12Z,15Z-octadecatrienoicacid 9-HOTrE 293.2106 5.5 6.20 275,231,185,171,121

(+/−)-11-hydroxy-5Z,8Z,12E,14Z,17Z-eicosapentaenoicacid 11-HEPE Hydroxy/hydroperoxyeicosapentaenoicacids 317.2115 2.2 6.28 299,259,255,215,195,167, 149,121

(+/−)-14(15)-epoxy-5Z,8Z,11Z,17Z-eicosatetraenoicacid 14(15)-EpETE Othereicosanoids 317.2106 5.0 6.58 299,255,219,207 (+/−)-20-hydroxy-4Z,7Z,10Z,13Z,16Z,18E-docosahexaenoicacid 20-HDoHE Docosanoids 343.2280 −0.3 7.51 325,299,285,281,241,187,

159,133,107

(+/−)-17-hydroxy-4Z,7Z,10Z,13Z,15E,19Z-docosahexaenoicacid 17-HDoHE 343.2276 0.9 8.02 325,281,273,245,229,227,

201,173,147,121,111 15-oxo-5Z,8Z,11Z,13E-eicosatetraenoicacid 15-oxo-ETE Hydroxy/hydroperoxyeicosatetraenoicacids 317.2107 4.7 8.05 299,273,139,113

(+/−)-9,10-epoxy-12Z-octadecenoicacid 9(10)-EpOME Otheroctadecanoids 295.2271 2.7 8.20 277,183,201,171

(+/−)-10-hydroxy-4Z,7Z,11E,13Z,16Z,19Z-docosahexaenoicacid 10-HDoHE Docosanoids 343.2278 0.3 8.22 325,299,281,227,181,161, 153,121

(+/−)-11-hydroxy-4Z,7Z,9E,13Z,16Z,19Z-docosahexaenoicacid 11-HDoHE 343.2277 0.6 8.41 325,281,227,194,165,149,

133,121,95 8-hydroxy-5Z,9E,11Z,14Z-eicosatetraenoicacid 8-HETE Hydroxy/hydroperoxyeicosatetraenoicacids 319.2276 0.9 8.60 301,257,163,155

(+/−)-7-hydroxy-4Z,8E,10Z,13Z,16Z,19Z-docosahexaenoicacid 7-HDoHE Docosanoids 343.2278 0.3 8.61 325,281,245,227,201,147,

141,121,113,97

(+/−)-8-hydroxy-4Z,6E,10Z,13Z,16Z,19Z-docosahexaenoicacid 8-HDoHE 343.2276 0.9 8.69 299,281,243,189,135,109

9-hydroxy-5Z,7E,11Z,14Z-eicosatetraenoicacid 9-HETE Hydroxy/hydroperoxyeicosatetraenoic acids

319.2272 2.2 8.77 301,275,257,229,203,179, 167,139,123,69

12-oxo-5Z,8Z,10E,14Z-eicosatetraenoicacid 12-oxo-ETE 317.2122 0.0 9.01 299,273,235,153

(6E,8Z,11Z)-5-hydroxyicosa-6,8,11-trienoicacid 5-HETrE 321.2415 6.2 9.10 303,259,205,115

(+/−)-12(13)-epoxy-9Z-octadecenoicacid 12(13)-EpOME Otheroctadecanoids 295.2275 1.4 9.20 277,195,183

5-oxo-6E,8Z,11Z,14Z-eicosatetraenoicacid 5-oxo-ETE Hydroxy/hydroperoxyeicosatetraenoicacids 317.2102 6.3 9.63 299,273,245,203,129