kynurenine amino acids, and of serotonin in humanplasma Development, application validation and of LC–MS/MS method forquantiﬁcation Journal of Pharmaceutical and Biomedical Analysis

(1)

Contents lists available atScienceDirect

Journal of Pharmaceutical and Biomedical Analysis

j o u r n a l h o m e p a g e :w w w . e l s e v i e r . c o m / l o c a t e / j p b a

Development, validation and application of LC–MS/MS method for quantiﬁcation of amino acids, kynurenine and serotonin in human plasma

Dávid Virág

^a

, Márton Király

^a

, László Drahos

^b

, Andrea Edit Édes

^c,d

, Kinga Gecse

^c,d

, György Bagdy

^d,e

, Gabriella Juhász

^c,d

, István Antal

^a

, Imre Klebovich

^a

,

Borbála Dalmadi Kiss

^a

, Krisztina Ludányi

^a,∗

aDepartmentofPharmaceutics,SemmelweisUniversity,H ˝ogyesEndreutca7,BudapestH-1092,Hungary

bMSProteomicsResearchGroup,InstituteofOrganicChemistry,ResearchCentreforNaturalSciencesoftheHungarianAcademyofSciences,Magyar tudósokkörútja2,BudapestH-1117,Hungary

cSE-NAP2GeneticBrainImagingMigraineResearchGroup,SemmelweisUniversity,Nagyváradtér4,BudapestH-1089,Hungary

dDepartmentofPharmacodynamics,FacultyofPharmacy,SemmelweisUniversity,Nagyváradtér4,BudapestH-1089,Hungary

eMTA-SENeuropsychopharmacologyandNeurochemistryResearchGroup,HungarianAcademyofSciences,SemmelweisUniversity,Nagyváradtér4, BudapestH-1089,Hungary

a r t i c l e i n f o

Articlehistory:

Received16August2019

Receivedinrevisedform6November2019 Accepted27November2019

Availableonline28November2019

Keywords:

LC–MS/MS Aminoacid Serotonin Kynurenine

Fit-for-purposevalidation Surrogatematrix

a b s t r a c t

Alteredserotonergicneurotransmissionisakeyfactorinseveralneurologicandpsychiatricdisorders suchasmigraine.Humanandanimalstudiessuggestthatchronicallylowinterictalserotoninlevels ofplasmaandbrainmayfacilitateincreasedactivityofthetrigeminovascularpathway,andmaycon- tributetodevelopmentofrepeatedmigraineattacks.However,brainserotoninsynthesisisaffectedby theconcentrationoftryptophan,itsmetabolitesandanumberofaminoacids.Inthisworkasimpleand robustLC–MS/MSmethodforthequantitativedeterminationofvaline,leucine,isoleucine,phenylalanine,tyrosine,tryptophan,serotoninandkynurenineinhumanplasmahasbeendevelopedandvalidated.

Samplepreparationwasachievedbyproteinprecipitation,usingtrifluoroaceticacid.Chromatographic separationwascarriedoutonaSupelcoAscentis^®ExpressC18column(3.0mmi.d.×150mm,2.7␮m) equippedwithanAgilentZorbaxEclipseXDBC8guard-columnunderisocraticconditionsataflowrate of0.4mL/min,overa6.5minruntime.Mobilephasewas0.2%trifluoroaceticacid–acetonitrile(85:15, v/v).Theeightanalytesandtwointernalstandardswereionizedbypositiveelectrosprayionizationand detectedinmultiplereactionmonitoringmode.

A“fit-for-purpose”validationapproachwasadoptedusingsurrogatematrixforthepreparationof calibrationsamples.Thecalibrationcurvesofallanalytesshowedexcellentlinearitieswithacorrelation coefficient(r²)of0.998orbetter.Spikedsurrogatematrixsamplesandpooledhumanplasmawere usedasqualitycontrolsamples.Intra-dayandinter-dayprecisionswerelessthan11.8%and14.3%, andaccuracieswerewithintherangesof87.4–114.3%and87.7–113.3%,respectively.Stabilityofthe componentsinstandardsolutions,surrogatematrix,pooledplasmaandprocessedsampleswerefound tobeacceptableunderallrelevantconditions.Nosignificantcarryovereffectwasobserved.Thesurrogate matrixbehavedparalleltohumanplasmawhenassessedbystandardadditionmethodanddilutingthe authenticmatrixwithsurrogatematrix.Themethodwassuccessfullyappliedforanalysisof800human plasmasamplestosupportaclinicalstudy.

∗Correspondingauthor.

E-mailaddress:ludanyi.krisztina@pharma.semmelweis-univ.hu(K.Ludányi).

1. Introduction

Alteredserotonergicneurotransmissionisakeyﬁndinginsev- eralneurologicandpsychiatricdisordersincludingmigraine[1].

Migraineisacommondisablingprimaryheadachedisorderwith a globalprevalenceof 15–18%[2]. It is characterizedby throb- https://doi.org/10.1016/j.jpba.2019.113018

(2)

2 D.Virág,M.Király,L.Drahosetal./JournalofPharmaceuticalandBiomedicalAnalysis180(2020)113018

bingandusuallyunilateralmoderateorseverepainaccompanied byphono-andphotophobia,nauseaorvomiting,andworsening byroutinephysicalactivity[3].Humanandanimalstudiessug- gestthatchronicallylowinterictalserotoninlevelofplasmaand brainpredisposetoincreasedsensitivityofthetrigeminovascu- larpathwaythatmaycontributetothedevelopmentofrepeated migraineattacks[1,2,4,5].Inlinewiththeserotonergicdysfunc- tionhypothesis,the5-HT_1B/1D(5-hydroxytryptaminereceptor1B and1Dsubtypes)agonisttriptansalleviatemigrainepain,while tryptophan depletion, which acutely decreases the brain serotonin concentration, increasesthe excitability of the brain and symptoms of the migraine attacks [1,5,6]. However, the brain serotoninconcentrationisaffectednotonlybytryptophanintake butalsobytheratiooftheplasmaconcentrationsoftryptophan andotherlargeneutralamino acids(LNAA)that arecompeting tooccupyatransporterattheblood brainbarrier,calledL-type amino acid transporter 1 (LAT1). Fernstrom deﬁned a ratio of plasmatryptophanand LNAAs, namelytyrosine, phenylalanine, leucine,isoleucineand valine,which inﬂuenced thebraintryp- tophanconcentrations and serotoninsynthesis in experimental studies. We selected these LNAAs for this measurement based onhisstudy[7].Investigationoftheseaminoacidsisimportant becausedietarymanipulationoftryptophan/LNAAratiopromptly abletoelicitbehavioralalterations,althoughtheexactmechanism isnotfullyunderstood.Forexample,tryptophandepletionwith ingestionofatryptophanfreeaminoacidmixtureisacutelyable toincrease anxiety,lowermoodand contributetomoresevere courseofmigraineattacksinvulnerablesubjects[6–8].Thusour aimwastodevelop a methodtosimultaneouslymeasure tryptophan, LNAAs and the peripheral concentrations of serotonin and kynurenine (another compound synthesized from trypto- phanand have been implicatedin migraine) in humanplasma [9].

During a typical bioanalytical validation procedure, calibra- tion and quality control (QC) samplesare prepared by spiking the same sort of biological matrix as the study samples with knownamountsofanalyte.Thisensuresthattheanalyteofinterest facessameconditionswithrespecttomatrixeffectsandextrac- tion recoveries in the validation samples as well as the study samples.In thecase of endogenouscompounds, preparation of validationsamplesarehamperedbytheabsenceofanalyte-free (blank)matrix[10].Asaccuratequantitativeanalysisofendoge- nousanalytesiscrucial foranumber ofclinicalandnonclinical applications,increasinginterestisarisingforalternative,so-called

“ﬁt-for-purpose” quantitation/validation approaches. There are fourcommonlyusedapproaches:standardaddition,background subtraction, surrogateanalyte in authentic matrix and authen- ticanalyteinsurrogatematrix[11–16].Applicability,advantages and drawbacksof these methods have recently been reviewed [17].

In the present article, we report a quick, accurate and reliable liquid chromatography-tandem mass spectrometry (LC–MS/MS) method for quantitative determination of valine, leucine, isoleucine, phenylalanine, tyrosine, tryptophan, sero- toninandkynureninein800humanplasmasamples.Twostable isotope-labeledinternalstandards(SIL-ISs)wereusedtoreduce assayvariation:leucine-5,5,5-d3(Leu-d3)foraliphaticaminoacids and tyrosine-(phenyl-3,5-d₂) (Tyr-d₂) for aromatic compounds, respectively.Duetothelargenumberofsamplesandcompounds tobeanalyzed,authentic analyte in surrogatematrix approach wasadoptedbyusingartiﬁcialplasmaasasurrogatematrixfor theconstructionofcalibrationstandardsand QCsamples.Some assayvalidationissuesofthemethodofchoicesuchasparallelism andlimitofquantitation(LOQ)arealsodiscussed.Fig.1showsthe eightanalytesandthetwoSIL-ISs.

2. Materialsandmethods

2.1. Chemicals

LC–MS grade acetonitrile waspurchased fromVWR Chemi- cals(Pool,England).LC–MSgradetriﬂuoroaceticacid(TFA)was purchased from Fisher Chemical (Loughborough, UK). Leucine (99.9%),isoleucine(99.5%),valine(99.3%),phenylalanine(99.0%), tyrosine (99.7%) were purchased from Dr. Ehrenstorfer GmbH (Augsburg,Germany).Tryptophan(≥99%)wasobtainedfromAlfa Aesar (Haverhill, MA, USA). Kynurenine (≥98%) was purchased fromCaymanChemicalCompany(AnnArbor,MI,USA).Serotonin (≥98%),l-leucine-5,5,5-d₃ (99%Datom),l-tyrosine-(phenyl-3,5- d₂) (98%D atom), phosphate bufferedsaline(PBS) and human serumalbumin(HSA)(≥96%)werepurchasedfromSigmaAldrich Inc.(St.Louis,MO,USA).K₃EDTAtubeswereobtainedfromGreiner Bio-OneInternationalGmbH(Kremsmünster,Austria).Ultrapure waterwaspreparedwithanin-houseSimplicity^®WaterPuriﬁca- tionSystem(MerckMillipore,Burlington,MA,USA).

2.2. Standardsolutions

Stock solutions of the analytesand internal standards were madeupin0.2%triﬂuoroaceticacid–acetonitrile(85:15,v/v)and storedat−20^◦C.Onthedayofanalysis,calibrationandQCstan- dardsolutions werepreparedbyserial dilutionoftwo separate primarystocksolutionswithwater.ISworkingsolutionwaspre- paredinwatertogiveaconcentrationof250␮g/mLforLeu-d3and 25␮g/mLforTyr-d₂.Surrogatematrixwaspreparedbyadding4g ofHSAto100mLPBSsolution.

2.3. Plasmasamples

Humanbloodsampleswerecollectedinto3mLK3EDTAtubes frommigrainepatientswithoutauraandhealthyvolunteers(total numberofparticipantswas98,totalnumberofbloodsampleswas 800)byintravenouscannulation.Thesampleswereimmediately centrifuged,thenplasmasampleswerefrozenandkeptat−80^◦C untiltheassay.Pooledplasmawaspreparedbymixingequalvol- umesofplasmasobtainedfrom6healthyindividuals.Thestudy protocolwasapprovedbytheScientiﬁcandResearchEthicsCom- mitteeoftheMedicalResearchCouncil,Budapest,Hungary.

2.4. Samplepreparation

Samplepreparationwasachievedbysimpleproteinprecipita- tion.ForthepreparationofcalibrationsamplesandQCsamplesin surrogatematrix,100␮Lofspikingstandardsolutionsand20␮Lof ISworkingsolutionweremixedwith900␮Lofsurrogatematrix.

Incaseofstudysamples,100␮Lofwaterand20␮LofISworking solutionweremixedwith900␮Lofhumanplasma.Forproteinpre- cipitation,200␮Loftriﬂuoroaceticacidwasaddedtoeachsample andvortexmixed.Aftercentrifugationat4000gfor10minat4^◦C, aliquots(150␮L)ofthesupernatantweretransferredtoautosam- plervials.5␮Loftheresultingsolutionswereinjectedintothe LC–MS/MSsystem.

2.5. Liquidchromatography-tandemmassspectrometry

ThechromatographicseparationswereperformedonanAgilent 1260InﬁnityLCsystem(AgilentTechnologies,CA,USA).Theana- lyteswereseparatedonaSupelcoAscentis^®ExpressC18column (3.0mmi.d.×150mm,2.7␮m)equippedwithanAgilentZorbax EclipseXDBC8guard-column(4.6mmi.d.×12.5mm,5␮m).The columnandtheautosamplerweremaintainedat25^◦Cand4^◦C, respectively.Thealiquotsofsampleswereelutedunderisocratic

(3)

Fig.1. Chemicalstructureoftheeightanalytesandtwointernalstandards.

conditionsover6.5minataflowrateof0.4mL/min.Themobile phase was composed of 0.2% trifluoroacetic acid– acetonitrile (85:15,v/v).Theneedlewaswashedfor10sintheflushportbefore everyinjectioninordertominimizecarryovereffect.

SampleswereanalyzedbyanAgilent6460triple-quadrupole massspectrometer(AgilentTechnologies, SantaClara, CA,USA) usingpositive electrospray ionization(ESI)and scheduledmul- tiple reaction monitoring (MRM) mode. Agilent MassHunter DataAcquisitionsoftware(versionB.04.01)wasusedtocontrol theequipment,MassHunterQuantitative QQQAnalysis(version B.05.01)andQualitativeAnalysissoftware(versionB.05.00)were appliedforquantitationanddataprocessing.Settingswereasfol- lows:capillaryvoltage,+3.5kV;nozzlevoltage,+400V.Nitrogen wasappliedasanebulizergasof45psi,acarriergasof10L/min at350^◦C,andasheathgasof11L/minat350^◦C.MRMtransitions, collisionenergies(CE)andfragmentorvoltagesforallcompounds wereautooptimizedbyOptimizerofAgilentMassHunterworksta- tion.

2.6. Validation

2.6.1. Calibrationcurves,LOQ

Thespikingstandardsolutions ofcalibrationstandardswere dilutedfromthestocksolutiontoobtaineightcalibrationlevels, andraninduplicateatthebeginningandtheendofeachbatch.The lowestandhighestpointsofthecalibrationcurvecoincidedwith thelowerlimitofquantitation(LLOQ)aswellastheupperlimitof quantitation(ULOQ).Thecalibrationcurvewasderivedbyplotting theconcentrationofthestandardsversustheanalytetoISpeak arearatiosusing1/xweightedleast-squareslinearregressions.

2.6.2. Accuracyandprecision

FourconcentrationlevelsofQCsampleswereusedforamino acids:LLOQ,lowQC,mediumQC,highQC,andthreeconcentra- tionlevelsforserotoninandkynurenine(LLOQ,lowQC,highQC) inordertocoverthecalibrationrange.Concentrationswereset accordingtotheEMAGuidelineonbioanalyticalmethodvalida- tion[18].Furthermore,pooledplasmawasappliedasanadditional levelofQCtoevaluatepotentialerrorscausedbymatrixdiffer-

ences[10].Endogenouslevelsofthesesamplesweredetermined bythemethodofstandardaddition.Calibrationrangesandcon- centrationsoftheQCsamplesaresummarizedinTable1.Intra-day accuracyandprecisionwereassessedbyevaluatingfivereplicates ofeachQCsamplesdescribedabove.Theinter-dayaccuracyand precisionwereestablishedbytherepetitionoftheintra-dayvali- dationprocedureonfiveconsecutivedays.Accuracywasexpressed aspercentageofthenominalconcentrationandprecisionwascal- culatedasthecoefficientvariation(%CV).Theacceptancecriteria forbothparametersweresetat±15%exceptfortheLLOQforwhich itshouldbewithin±20%[18].

2.6.3. Extractionrecovery,parallelism,dilutionintegrity

Extractionrecoverieswereevaluatedbycomparingtheanalyte responsesobservedin pre-spikedandpost-spiked samples.The experimentwasperformedinsixreplicates.Pre-spikedsamples werepreparedbyadding100␮LofhighQCstandardsolutionand 20␮LofISstandardworkingsolutionto900␮Lofsurrogatematrix, thenprecipitatedwithTFAandcentrifuged.Inthecase ofpost- spikedsamples,1000␮Lofsurrogatematrix(blanksample)was processedinthesameway.90␮Loftheresultingsupernatantwere mixedwith10␮LofhighQCstandardsolutionand2␮LofISstan- dardworkingsolution.Extractionrecovery(%)wascalculatedas peakareaofpre-spikedsamples/peakareaofpost-spikedsamples x100.IS-normalizedextractionrecovery(%)wasalsoexpressed fortheanalytesasextractionrecoveryoftheanalyte/extraction recoveryoftheISx100.

Thetermparallelismisgenerallyunderstoodtomeanhowwell asetofcalibrationstandardstracktheresponseoftheanalyteof interestinthebiologicalmatrix[19].Evaluationofparallelismis ofcriticalrelevancewhenusingsurrogatematrix,ashighdegree ofsuchparameterindicatesthatthereisnosigniﬁcantdifference betweentheauthenticandthesurrogatematrixwithrespectto extractionrecoveriesandmatrixeffects.Determinationofparal- lelismwascarriedoutbasedupon theexperiment proposedby Houghton etal.[20]. Endogenousconcentrationsof sixindivid- ualplasma samplesweredeterminedin triplicates bystandard addition.ThesamesampleswerespikedwithmediumQCstan- dardsolutionandwereseriallydilutedwithsurrogatematrixtwo,

(4)

Table1

CalibrationrangesandQClevelsoftheanalytes.

QCtype Surrogatematrix Pooledplasma

Analyte Calibrationrange(␮g/ml) QClevel(␮g/ml)

LLOQ Low Medium High

Valine 1–100 1 3 20 80 24.05

Leucine 1–100 1 3 20 80 13.09

Isoleucine 0.5–50 0.5 1.5 10 40 7.82

Phenylalanine 1–100 1 3 20 80 6.98

Tyrosine 0.5–50 0.5 1.5 10 40 11.88

Tryptophan 0.5–50 0.5 1.5 10 40 7.81

Serotonin 0.1–2.5 0.1 0.3 – 2 0.20

Kynurenine 0.1–2.5 0.1 0.3 – 2 0.44

ﬁveandtentimes.Thenthedilutedandundilutedsampleswere measuredagainstsurrogatecalibrators.The%REand%CVofthe back-calculatedconcentrationsweremaximizedin±15%[20].By trackingtheeffectofdilutingtheplasmawithsurrogatematrix,the methodcanalsobeconsideredasadilutionintegritytest.

2.6.4. Stability

Stability of thestock solutions and working solutions were assessed atroom temperature (20^◦C) for 24h and at 10^◦C for tendays.Working solutionsof theeight analyteswerestudied usingbracketingapproach:onlythelowestandthehighestcal- ibrationstandardsweremeasured.Thelatterwasalsothestock solution.Stabilityoftheanalytesinmatrixwereevaluatedusing lowandhighQCsamplespreparedinsurrogateplasmaandpooled humanplasmasamples.Eachlevelswereanalyzedintriplicates after beingexposed to different conditions:room temperature for 24h, −20^◦C for two months and three freeze-thaw cycles.

Stability of processed samples were determined at room temperature for 12hand autosampler temperature (4^◦C) for 24h.

Sampleswereconsideredstablewhen the%RE waswithin15%

[18].

2.6.5. Carryover

Carryoverwasanalyzedbyassayingblanksamples(processed surrogatematrixsamples)injectedaftertheanalysisofapooled plasmasamplespikedwiththehighestcalibrationstandard.The acceptancecriteriaweresetat20%analyteresponseoftheLLOQ fortheanalytesofinterestand5%fortheISs[18].

2.7. Applicationofthemethodtoclinicalhumanplasmasamples Thevalidated methodwasappliedto800samplescollected duringtheclinicalstudydescribedearlier.Thesamplesweremea- suredintenseparateanalyticalruns.Surrogatecalibratorswere runatthebeginningandtheendofeachbatch.Low,medium,high QCsandpooledplasma sampleswereanalyzed induplicatesat eachbatchtokeeptrackofaccuracyandprecision.Reproducibil- ityofthedescribedmethodforincurredstudysampleswastested byreanalysisof10%ofthesamples(n=80).Reanalyzedsamples wereconsideredacceptablewhenthedifferencebetweenthepairs oftheresultswerewithin20%foratleast66.7%ofthesamples [18].

3. Resultsanddiscussion

3.1. Liquidchromatography-tandemmassspectrometry

Thereisavastamountofliteratureonthequantitativedeter- minationof amino acids in plants, animal and humansamples [21–23].However,tothebestofourknowledge,nobioanalytical LC–MS/MSmethodhasbeenreportedforthesimultaneousanalysis

Table2

MRMparametersoftheanalytesandinternalstandards.

Compound Transition(m/z) Fragmentorvoltage(V) Collisionenergy(eV)

Valine 118→72 10 8

Leucine 132→86 2 8

Isoleucine 132→86 2 8

Phenylalanine 166→120 2 8

Tyrosine 182→136 30 12

Tryptophan 205→188 40 4

Serotonin 177→160 45 4

Kynurenine 209→192 45 4

Leu-d3 135→89 45 4

Tyr-d2 184→138 45 10

ofaminoacids,serotoninandkynurenineinhumanplasma.Amino acidsareveryoftenderivatizedinorder toenhancefluorescent detection selectivity, sensitivity or chromatographic resolution [23,24].However,thederivatizationagentcanbringanadditional sourceofinterferencesand errorsintothesystemcomplicating methodvalidation[24].Itshouldbealsotakenintoaccountthat derivatizationcansignificantlyincreasethelaborintensityofthe method,which can becrucial when thenumber of samplesto beanalyzedislarge. Inthis particularcase,where theaimwas themeasurementof800(withvalidationsamplesfarover1000) plasmasamples,weprimarilyfocusedonspeed,simplicityandhigh throughputduringmethoddevelopment.MSconditionsincluding MRMtransitions,collisionenergies(CE)andfragmentorvoltages wereindividuallyoptimizedforallanalytesandSIL-ISsinpositive ionmodebyOptimizerofAgilentMassHunterworkstationinorder toenhanceselectivityandsensitivityofthemethod.Asshownin Table2,theoptimizationprocedureresultedinsameMRMtransi- tionsforisomericaminoacidsleucineandisoleucine.Becauseof theinsufficientmassspectrometricselectivity,properchromato- graphic resolution wasnecessary to separatesuch compounds withoutcompromisingpeakshape orareasonablyshortanalyt- icalrun time. As representative MRMchromatograms in Fig.2 demonstrates,aSupelcoAscentis^® ExpressC18column(3.0mm i.d.×150mm,2.7␮m)providedsatisfactoryseparationoftheana- lyteswithina6.5minrun.Anisocraticconditionconsistingof0.2%

TFA-acetonitrile(85:15,v/v)atﬂowrateof0.4mL/minwascho- seninordertoavoidtheneedforcolumnreequilibrationbetween consecutiveinjections.AZorbaxEclipseXDBC8guard-columnwas usedtoprotecttheanalyticalcolumn,therefore,performanceof theseparationintermsofpeakshapesandretentiontimeswas consistentthroughoutthestudy.

3.2. Validation

3.2.1. Calibrationcurves,LLOQ

Themeanregressioncoefﬁcients(r²)foralltheanalyteswere over0.998indicatingexcellentlinearities.Thecalibrationranges weresufﬁcienttocoveranalyteconcentrationsinalmostallhuman

(5)

Fig.2.ExtractedMRMchromatogramsoftheanalytesandinternalstandards.

plasmasamples.LLOQ is describedas thelowestconcentration withacceptableaccuracyandprecisionwhichlargelydependson thesensitivityoftheinstrument[18].However,inthisstudy,only theLLOQofserotoninandkynurenineweredeterminedwiththis approach.In thecaseof aminoacids,measurement ofaccuracy andprecisionattheLLOQmaynotbeadequate,consideringthat suchconcentrationsareoutsidethephysiologicallyrelevantrange [25].Thusthelowestcalibrationlevels,followingamorepragmatic approach,weresettobeapproximatelyonetenthoftheconcen- trationsmeasuredinpooledplasma.Thismethodsimpliﬁesthe preparationand dilutionofcalibrationstandardswhenworking withalargenumberofanalytes,whilecoveringtherelevantcon-

centrationrange,evenifdownregulationoftheanalytesisexpected [20,25].

3.2.2. Accuracyandprecision

Intra-day and inter-day accuracies and precisions for surrogate matrix andpooled plasma QC samplesaresummarized in Table3.Inthisstudy,theintra-dayandinter-dayprecisionswere lessthan11.8%and14.3%,andaccuracieswerewithintheranges of87.4–114.3%and87.7–113.3%,respectively.Theresultsindicate thatthemethodisaccurateandpreciseenoughforthemeasure- mentofplasmasamples.

(6)

Table3

Intra-dayandinter-dayaccuracyandprecisionoftheQCsamples.

QCtype Surrogatematrix Pooledplasma

QClevel LLOQ Low Medium High

Accuracy(%) CV(%) Accuracy(%) CV(%) Accuracy(%) CV(%) Accuracy(%) CV(%) Accuracy(%) CV(%)

Intra-day

Valine 111.7 7.4 93.0 11.8 100.4 11.3 110.8 8.8 92.4 3.7

Leucine 102.7 5.7 95.3 2.5 93.8 4.9 89.7 3.3 90.9 4.0

Isoleucine 114.3 7.2 104.8 3.3 108 3.2 113.3 7.0 101.3 3.9

Phenylalanine 100.2 5.2 102.1 5.6 99.8 6.4 102.9 2.8 102.3 6.4

Tyrosine 98.6 4.3 87.4 2.5 88.9 10.5 103.7 4.2 107.0 3.3

Tryptophan 106.8 6.8 106.0 6.5 95.8 10.2 108.0 4.6 97.2 6.7

Serotonin 101.0 8.1 92.7 2.9 – – 87.8 4.7 95.1 7.2

Kynurenine 110.7 8.2 114.3 10.1 – – 113.4 9.4 110.9 9.7

Inter-day

Valine 107.6 10.2 96.5 11.9 97.0 9.7 111.7 9.0 91.5 5.3

Leucine 102.5 6.4 95.6 7.8 91.5 5.6 92.4 7.1 95.8 8.2

Isoleucine 114.4 8.2 106.7 9.6 103.0 6.8 112.6 8.3 106.4 8.7

Phenylalanine 102.1 7,4 100.1 7.2 96.5 6.9 104.6 8.8 108.8 5.9

Tyrosine 111.2 14.3 102.3 13.3 102.9 10.7 109.5 7.6 104.5 3.5

Tryptophan 102.9 8.1 96.7 7.6 95.4 6.3 113.3 4.7 109.5 7.0

Serotonin 99.5 14.2 87.7 7.7 – – 91.1 7.4 89.2 9.1

Kynurenine 92.4 13.8 90.7 10.6 – – 95.0 12.8 91.2 11.4

3.2.3. Extractionrecovery,parallelism,dilutionintegrity

Extraction recovery of the method was evaluated by com- paringthemeanpeak areaof pre-extractedand post-extracted surrogatematrix samples. The sample preparation process did not inﬂuence considerably the concentrations of the analytes anISs(extraction recoveriesvariedbetween85.2–101.4%).Fur- thermore,theIS-normalizedextractionrecoveryvalueshighlight that the selected SIL-ISs are appropriate references of the analytes (IS-normalized extraction recoveries varied between 93.6–111.8%). Results of extraction efﬁciency evaluation are detailedinTable4.

Fortheparallelismexperiment,backgroundconcentrationsof theindividualplasmasamplesweredeterminedbystandardaddi- tion.Evaluationofthesamesampleswithsurrogatecalibrators resulted very similar concentrations for all theanalytes (RE is lessthan10%).Fig.3illustrates theuseof standardadditionto assessparallelismwithsurrogatematrix methodin thecase of isoleucine.Moreover,themeasuredanalyteconcentrationsinthe spikedsamples at each dilution levels were within the accep- tancecriteria,whichindicatesthatthesurrogatematrixbehaves similarlytotheauthenticbiologicalmatrix.Fig.4illustratesthe relatively balanced accuracy proﬁle of a representative plasma samplewhendilutedzero-two-,ﬁve-,ten-timeswithsurrogate matrix.

3.2.4. Stability

Thestocksolutionsandworkingsolutionsoftheanalytesand internalstandardswerestable foratleast24hat20^◦C and for 10daysat10^◦C,respectively.Humanplasmasamplesfrozento

−20^◦Candthawedtoroomtemperatureforthreecyclesdidnot changeconsiderablytheconcentrationsof theanalytes.Human plasmaandsurrogatematrixspikedwithQCstandardswerestable atroomtemperaturefor24hand−20^◦Cfor2months.Noapparent changewasobservedintheconcentrationsoftheprocessedsam- plesafterbeing24hatautosamplertemperature(4^◦C)and12hat roomtemperature.

3.2.5. Carryover

Carryoverwasanalyzedbyinjectingblanksamplesafterpooled plasmaspikedwiththehighest calibrationstandard.AsTable5 demonstrates, responses of the analytes were 0.7–17.2% and 0.8–1.0%fortheinternalstandards,respectively.

3.3. Applicationofthemethodtoclinicalhumanplasmasamples Thevalidatedmethodwassuccessfullyappliedinsupportof aclinicalstudytomeasuretheplasmaconcentrationoftheeight analytesin 800humanplasmasamplesobtainedfrommigraine patientsand healthyindividuals.Valine,isoleucineandtyrosine concentrationswereabovetheULOQinlessthan1%ofthesamples.Insuchcases,anotheraliquotofthesampleweredilutedtwo timeswithsurrogatematrix,thenprocessedandmeasuredagain.

AccuracyandprecisionoftheQCsamplesdidnotchangeconsid- erablycomparedtotheresultsofthemethodvalidation,incurred samplereanalysiswasalsowithintheacceptancecriteria.Table6 summarizestheconcentrationrangesin800studysamples.

3.4. Strategiesforquantitativeanalysisofendogenous compounds

Regulatoryguidancesare designedprimarilytoaddressvali- dationofdrugmoleculesaswellastheirmetabolites.However, reﬂecting on the challenges of endogenous analytes, the draft version ofthe ICHguideline of bioanalytical methodvalidation dedicatesasectionforthequantitativedeterminationofthesecom- pounds,inwhichtheabove-mentionedquantitationstrategiesare detailed[26].Consideringthatbothmethodshaveadvantagesand limitations,theirapplicabilityisdeterminedbythespeciﬁcanalyt- icalchallenge.Inthepresentcase,whenthenumberofsamplesto beanalyzedislargeandtheamountofeachsampleisrelatively low(insomecases1mLorless),theuseofstandardadditionis narrowedbytherequirementformultiplemeasurementsofevery studysamples.Comparedtohealthyindividuals,downregulation ofcertainanalytesisexpectedinmigrainepatients,whichmeans that sampleswithlowerconcentration thanthecontrol matrix shouldbetakenintoaccount.Therefore, theuseofbackground subtractionis not recommendeddue tothepotentialmeasure- menterrors introducedbyextrapolationbeyondthecalibration range[20]. Thesituationisfurthercomplicatedwhen thereare multipleanalyteswithvariableendogenouslevels[17].Surrogate analyteinauthenticmatrixtypicallyrequirestwoversionsofSIL- ISsoftheanalyteofinterest[10].Despitebeingawell-accepted approach,itsapplicationislimitedbytheavailabilityandhighprice ofstableisotope-labeledanalogs,especiallywhentheanalysisof numerousanalytesisneeded.Undersuchcircumstances,surro- gatematrixapproachmaybethemethodofchoice,whichenables reliablequantitationofmultipleanalyteswithrelativelylowlabor

(7)

Table4

Meanextractionrecoveriesoftheanalytesandinternalstandardsinsurrogatematrix.

Valine Leucine Isoleucine Phenylalanine Tyrosine Tryptophan Serotonin Kynurenine Leu-d3 Tyr-d2

Extractionrecovery(%) 87.5 101.3 88.6 101.4 95.0 95.1 85.2 97.2 90.6 91.0

IS-normalizedextractionrecovery(%) 96.6 111.8 97.8 111.3 104.4 104.5 93.6 106.7 – –

Fig.3.Illustrationoftheuseofstandardadditiontoassessparallelism.Endogenousconcentrationsweredeterminedbysurrogatecalibrationandbyextrapolatingtothe negativex-interceptfromthestandardadditioncalibrationcurve.Similarityoftheresultsindicatingparallelismbetweenhumanplasmaandtheartiﬁcialmatrix.

Fig.4.Assessmentofparallelismusingdilutionlinearitytest.Spikedhumanplasmasamplesweremeasuredagainstsurrogatecalibratorswhendilutedtwo-,ﬁve-,ten-times andnondiluted.Thedatawasthemeanoftriplicates.Theanalyteconcentrationsweremultipliedbythedilutionfactors.

Table5

Carryoveroftheanalytesandinternalstandards.

Valine Leucine Isoleucine Phenylalanine Tyrosine Tryptophan Serotonin Kynurenine Leu-d3 Tyr-d2

Carryover(%) 17.2 4.8 8.3 2.1 8.7 0.7 2.6 2.7 1.0 0.8

intensity,evenwhendownregulationoftheanalytesisexpected [1,7,9].Apotentialdrawbackofthemethodisthesystematicerror thatmay beintroducedbythelack ofparallelismbetweenthe bioﬂuidandthesurrogatematrix[19,20].However,parallelismcan beensuredbythepropercompositionofthesurrogatematrixand monitoredbydilutionlinearitytestingorstandardaddition[19].

Usingstandardadditiontoinvestigateparallelismclearlyindicates thatﬁt-for-purposestrategiesshouldbeconsideredascomponents ofananalyticaltoolbox,thosecanbeappliedindividuallyandin combinationtomakesurethatnoneofthevalidationcriteriaare compromised.

Table6

Concentrationrangesoftheanalytesin800studysamples.

Analyte Concentrationrange(␮g/ml)

Valine 4.63–115.58

Leucine 6.31–71.42

Isoleucine 1.68–97.66

Phenylalanine 1.65–36.76

Tyrosine 1.16–82.20

Tryptophan 1.17–39.35

Serotonin 0.27–1.70

Kynurenine 0.10–2.04

(8)

4. Conclusion

A simple, fast and reproducible LC–MS/MSbased bioanalyt- icalmethod hasbeen developed for simultaneous quantitation ofvaline,leucine,isoleucine,phenylalanine,tyrosine,tryptophan, serotoninandkynurenineinhumanplasma.Thechromatographic method wascapable to separate isomeric amino acids leucine andisoleucine.Aﬁt-for-purposevalidationapproachwasadopted byemployingsurrogatematrixforcalibration,andpartlyforthe constructionofQC samples.Thesurrogatematrix wasfoundto beparallelwithhumanplasma,when investigatedbystandard additionanddilutionintegritytesting.Somequantitationandval- idationissues of endogenous compounds have been discussed.

Themethodwasabletomeetthecriteriaimposedbytheregu- latoryguidances ofEMAand FDA[18,27],and demonstratedits valueintheanalysisofatotalnumber800humanplasmasam- ples.

Funding

This work wassupported by the Hungarian Brain Research Program–GrantNo.KTIANAP13-2-2015-0001(MTA-SE-NAPB GeneticBrain ImagingMigraine Research Group), by the Hun- garianAcademyofSciences(MTA-SENeuropsychopharmacology and Neurochemistry Research Group), and by the Hungarian Brain Research Program – Grant No. 2017-1.2.1- NKP-2017- 00002. GK was supported by the New National Excellence ProgramoftheMinistryofHuman Capacities(ÚNKP-18-2-I-SE- 86).

ThestudywasalsoﬁnancedbytheHigherEducationInstitu- tionalExcellenceProgrammeoftheMinistryofHumanCapacities inHungary,withintheframeworkofthemolecularbiologythe- maticprogrammeofSemmelweisUniversity.

DeclarationofCompetingInterest Authorsdeclarenoconﬂictofinterest.

References

[1]M.Deen,C.E.Christensen,A.Hougaard,H.D.Hansen,G.M.Knudsen,M.

Ashina,Serotonergicmechanismsinthemigrainebrain—asystematicreview, Cephalalgia37(2017)251–264,http://dx.doi.org/10.1177/

0333102416640501.

[2]P.J.Goadsby,P.R.Holland,M.Martins-Oliveira,J.Hoffmann,C.Schankin,S.

Akerman,Pathophysiologyofmigraine:adisorderofsensoryprocessing, Physiol.Rev.97(2017)553–622,http://dx.doi.org/10.1152/physrev.00034.

2015.

[3]HeadacheClassiﬁcationCommitteeoftheInternationalHeadacheSociety (IHS),Theinternationalclassiﬁcationofheadachedisorders,3rdedition, Cephalalgia38(2018)1–211,http://dx.doi.org/10.1177/0333102417738202.

[4]G.Juhasz,T.Zsombok,E.A.Modos,S.Olajos,B.Jakab,J.Nemeth,J.Szolcsanyi,J.

Vitrai,G.Bagdy,NO-inducedmigraineattack:strongincreaseinplasma calcitoningene-relatedpeptide(CGRP)concentrationandnegative correlationwithplateletserotoninrelease,Pain106(2003)461–470.

[5]R.C.A.Guedes,M.dasG.R.deAraújo,T.C.Verc¸osa,F.M.Bion,A.L.deSá,A.

Pereira,R.Abadie-Guedes,Evidenceofaninversecorrelationbetween serotonergicactivityandspreadingdepressionpropagationintheratcortex, BrainRes.1672(2017)29–34,http://dx.doi.org/10.1016/j.brainres.2017.07.

011.

[6]P.D.Drummond,Tryptophandepletionincreasesnausea,headacheand photophobiainmigrainesufferers,Cephalalgia26(2006)1225–1233,http://

dx.doi.org/10.1111/j.1468-2982.2006.01212.x.

[7]J.D.Fernstrom,Largeneutralaminoacids:dietaryeffectsonbrain

neurochemistryandfunction,AminoAcids45(2013)419–430,http://dx.doi.

org/10.1007/s00726-012-1330-y.

[8]S.N.Young,Acutetryptophandepletioninhumans:areviewoftheoretical, practicalandethicalaspects,J.PsychiatryNeurosci.38(2013)294–305, http://dx.doi.org/10.1503/jpn.120209.

[9]L.Vécsei,L.Szalárdy,F.Fülöp,J.Toldi,KynureninesintheCNS:recent advancesandnewquestions,Nat.Rev.DrugDiscov.12(2013)64–82,http://

dx.doi.org/10.1038/nrd3793.

[10]W.Jian,R.W.Edom,N.Weng,Importantconsiderationsforquantitationof small-moleculebiomarkersusingLC–MS,Bioanalysis4(2012)2431–2434, http://dx.doi.org/10.4155/bio.12.247.

[11]Y.Huang,A.Louie,Q.Yang,N.Massenkoff,C.Xu,P.W.Hunt,W.Gee,Asimple LC–MS/MSmethodfordeterminationofkynurenineandtryptophan concentrationsinhumanplasmafromHIV-infectedpatients,Bioanalysis5 (2013)1397–1407,http://dx.doi.org/10.4155/bio.13.74.

[12]J.Jiang,C.A.James,P.Wong,Bioanalyticalmethoddevelopmentand validationforthedeterminationofglycineinhumancerebrospinalﬂuidby ion-pairreversed-phaseliquidchromatography-tandemmassspectrometry, J.Pharm.Biomed.Anal.128(2016)132–140,http://dx.doi.org/10.1016/j.jpba.

2016.05.019.

[13]X.Lai,J.A.Kline,M.Wang,Development,validation,andcomparisonoffour methodstosimultaneouslyquantifyl-arginine,citrulline,andornithinein humanplasmausinghydrophilicinteractionliquidchromatographyand electrospraytandemmassspectrometry,J.Chromatogr.BAnal.Technol.

Biomed.LifeSci.1005(2015)47–55,http://dx.doi.org/10.1016/j.jchromb.

2015.10.001.

[14]N.C.vandeMerbel,K.J.Bronsema,M.W.J.vanHout,R.Nilsson,H.Sillén,A validatedliquidchromatography-tandemmassspectrometrymethodforthe quantitativedeterminationof4␤-hydroxycholesterolinhumanplasma,J.

Pharm.Biomed.Anal.55(2011)1089–1095,http://dx.doi.org/10.1016/j.jpba.

2011.03.017.

[15]Y.Shi,X.Xu,M.Fang,M.Zhang,Y.Li,B.Gillespie,S.Yorke,N.Yang,J.C.

McKew,W.A.Gahl,M.Huizing,N.Carrillo-Carrasco,A.Q.Wang,Quantitative hydrophilicinteractionchromatography-massspectrometryanalysisof N-acetylneuraminicacidandN-acetylmannosamineinhumanplasma,J.

Chromatogr.BAnal.Technol.Biomed.LifeSci.1000(2015)105–111,http://

dx.doi.org/10.1016/j.jchromb.2015.07.018.

[16]A.Q.Wang,B.P.Wei,Y.Zhang,Y.J.Wang,L.Xu,K.Lan,Anultra-highsensitive bioanalyticalmethodforplasmamelatoninbyliquid

chromatography-tandemmassspectrometryusingwaterascalibration matrix,J.Chromatogr.BAnal.Technol.Biomed.LifeSci.879(2011) 2259–2264,http://dx.doi.org/10.1016/j.jchromb.2011.06.010.

[17]R.Thakare,Y.S.Chhonker,N.Gautam,J.A.Alamoudi,Y.Alnouti,Quantitative analysisofendogenouscompounds,J.Pharm.Biomed.Anal.128(2016) 426–437,http://dx.doi.org/10.1016/j.jpba.2016.06.017.

[18]EuropeanMedicinesAgency,CommitteeforMedicinalProductsforHuman Use,GuidelineonBioanalyticalMethodValidation,2011,July2011https://

www.ema.europa.eu/en/documents/scientiﬁc-guideline/guideline- bioanalytical-method-validationen.pdf(Accessed4November2019).

[19]B.R.Jones,G.A.Schultz,J.A.Eckstein,B.L.Ackermann,Surrogatematrixand surrogateanalyteapproachesfordeﬁnitivequantitationofendogenous biomolecules,Bioanalysis4(2012)2343–2356,http://dx.doi.org/10.4155/bio.

12.200.

[20]R.Houghton,C.HorroPita,I.Ward,R.Macarthur,Genericapproachto validationofsmall-moleculeLC–MS/MSbiomarkerassays,Bioanalysis1 (2009)1365–1374,http://dx.doi.org/10.4155/bio.09.139.

[21]N.Zheng,H.Xiao,Z.Zhang,X.Gao,J.Zhao,Rapidandsensitivemethodfor determiningfreeaminoacidsinplanttissuebyhigh-performanceliquid chromatographywithﬂuorescencedetection,ActaGeochim.36(2017) 680–696,http://dx.doi.org/10.1007/s11631-017-0244-5.

[22]M.K.Dhillon,S.Kumar,G.T.Gujar,AcommonHPLC-PDAmethodforamino acidanalysisininsectsandplants,IndianJ.Exp.Biol.52(2014)73–79.

[23]G.Sharma,S.V.Attri,B.Behra,S.Bhisikar,P.Kumar,M.Tageja,S.Sharda,P.

Singhi,S.Singhi,Analysisof26aminoacidsinhumanplasmabyHPLCusing AQCasderivatizingagentanditsapplicationinmetaboliclaboratory,Amino Acids46(2014)1253–1263,http://dx.doi.org/10.1007/s00726-014-1682-6.

[24]J.Pawliszyn,HandbookofSolidPhaseMicroextraction,Elsevier,Amsterdam, 2012,http://dx.doi.org/10.1016/C2011-0-04297-7.

[25]N.C.vandeMerbel,Quantitativedeterminationofendogenouscompoundsin biologicalsamplesusingchromatographictechniques,TrACTrendsinAnal.

Chem.27(2008)924–933,http://dx.doi.org/10.1016/j.trac.2008.09.002.

[26]EuropeanMedicinesAgency,CommitteeforMedicinalProductforHuman Use,ICHGuidelineM10onBioanalyticalMethodValidation,DraftGuidance, 2019,March2019https://www.ema.europa.eu/en/documents/scientiﬁc- guideline/draft-ich-guideline-m10-bioanalytical-method-validation-step-2b en.pdf(Accessed4November2019).

[27]U.S.DepartmentofHealthandHumanServices,FoodandDrug

Administration,GuidanceforIndustry,BioanalyticalMethodValidation,2018, May2018https://www.fda.gov/ﬁles/drugs/published/Bioanalytical-Method- Validation-Guidance-for-Industry.pdf(Accessed4November2019).