Cms

NC

ARRAA

KBHHMG

1

ttsflsnmfaiba

Df

0h

Journal of Chromatography A, 1305 (2013) 234– 243

Contents lists available at SciVerse ScienceDirect

Journal of Chromatography A

j our nal homep age: www.elsev ier .com/ locate /chroma

omprehensive profiling analysis of bioamines and their acidicetabolites in human urine by gas chromatography/mass

pectrometry combined with selective derivatization

a-Hyun Park, Joo Yeon Hong, Hyun Ju Shin, Jongki Hong ∗

ollege of Pharmacy, Kyung Hee University, Seoul 130-701, Republic of Korea

a r t i c l e i n f o

rticle history:eceived 25 March 2013eceived in revised form 24 June 2013ccepted 1 July 2013vailable online 4 July 2013

eywords:iogenic aminesuman urineMDS/MBHFBA derivatizationCX SPEC/MS

a b s t r a c t

A comprehensive analytical method was developed for the profiling of biogenic amines in human urineusing GC/MS in SIM mode. Biogenic amines and their acidic metabolites were converted into their volatileO-trimethylsilyl/N-heptafluorobutyryl (OTMS/-NHFBA) derivatives for GC/MS analysis. Dual hexamethyl-disilazane (HMDS)/-N-methyl-bis-heptafluorobutyramide (MBHFBA) derivatizations have been shownto be quite effective, with high derivatization yields and the absence of side products for primary bio-genic amines. In this study, selective derivatization conditions by HMDS/MBHFBA were optimized interms of the reagent amount, reaction temperature and reaction time period. The highest derivatizationreaction yield was obtained at 40 ◦C for 10 min for OTMS derivatization and 80 ◦C for 5 min for N-HFBAderivatization. The use of MCX SPE cartridges with different SPE elution solvents was effective for thepre-concentration and selective cleanup of the biogenic amines and their acidic metabolites in humanurine. The selection of appropriate ions in SIM mode provided reliable quantification and identification

and a reduction in background effects. The established method was validated in terms of linearity, limitsof detection (LOD), limits of quantification (LOQ), precision, and accuracy. The present method was linear(r2 > 0.996), reproducible (relative standard deviation range 1.1–6.9%), and accurate (range 87.9–111.9%),with LOQs of 0.17–17.84 ng/mL. The biogenic amine profiling of human urine was successfully accom-plished by GC/MS in SIM mode combined with selective HMDS/MBHFBA derivatization and MCX SPEcleanup.

. Introduction

Biogenic amines, such as catecholamines and serotonin, andheir metabolites are important mammalian neurotransmittershat are released by the adrenal glands and the sympathetic nervousystem [1–3]. The profiling analysis of biogenic amines in biologicaluids has provided valuable information regarding physiologicaltatus and metabolic disorders that may be useful for diagnosingeuronal diseases [4,5]. In some cases, the concentrations of theajor acidic metabolites in urine can serve as indirect biomarkers

or the diagnosis of metabolic disorders. However, both biogenicmines and their metabolites should be quantified because phys-

ological levels are influenced by the concentrations of both theiosynthetic and metabolic enzymes [6]. Thus, the development ofnalytical methods for the determination of biogenic amines and

∗ Corresponding author at: College of Pharmacy, Kyung Hee University, Hoegi-ong, Dongdaemoon-Ku, Seoul 130-701, Republic of Korea. Tel.: +82 2 961 9255;

ax: +82 2 962 9255.E-mail address: jhong@khu.ac.kr (J. Hong).

021-9673/$ – see front matter © 2013 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.chroma.2013.07.003

© 2013 Elsevier B.V. All rights reserved.

their acidic metabolites has become an important subject in theclinical diagnosis of neuronal diseases.

Several analytical methods, such as HPLC and capillaryelectrophoresis combined with electrochemical detection [7], fluo-rescence detection [8,9] or mass spectrometric (MS) detection [10],GC/MS [11–13], and LC–MS/MS [14,7], have been developed for thebiogenic amine profiling analysis of biological samples. Recently,an LC–MS/MS multiple ion reaction monitoring (MRM) methodprovided high selectivity and sensitivity for the analysis of bio-genic amines and/or their metabolites in biological samples withcomplex matrixes [15,16]. The disadvantages of LC–MS/MS meth-ods applied toward the simultaneous determination of basic andacidic endogenous compounds, apart from sample preparation andmobile phase limitations, include several factors associated withsensitivity drift [17], the application of dual polarity in electro-spray ionization (ESI) according to the acidic and basic propertiesof endogenous compounds, and the susceptibility of certain ioniza-

tion techniques to matrix effects [18].

GC/MS has been widely used for the analysis of some biogenicamines and their acidic metabolites in biological samples due toits high peak capacity, high detection sensitivity, and positive peak

atogr

ctrltFtttmtta[ade(ahbebhiau

tftacMacuacwutbitr

fmdfioe(a

2

2

hdh

N.-H. Park et al. / J. Chrom

onfirmation. However, the GC/MS analysis of biogenic amines andheir acidic metabolites is relatively complicated due to the wideange of structures and functional groups of the analytes and theirow physiological concentrations in biological samples. In additiono these factors, other issues exist that complicate GC/MS analyses.irst, biogenic amines and their acidic metabolites require deriva-ization of their hydroxyl, amino and/or carboxyl groups to reduceheir polarity, to enhance the thermal stability and volatility, ando improve detection sensitivity [7,14,15]. The derivatization of

ultiple functional biogenic amines, especially compounds con-aining primary amines, often yields incomplete reactions, andhe formation of by-products, resulting in multiple derivativesnd leading to difficulties in the quantification of the analytes19]. Second, the simultaneous extraction and cleanup of biogenicmines and their acidic metabolites from human urine remains aifficult task. Current extraction methods, including liquid-liquidxtraction (LLE) [20,21], immersion solid-phase microextractionSPME) [17,22] combined with ethyl chloroformate derivatization,nd solid-phase extraction (SPE) using various adsorbents [23–25],ave been applied for the selective extraction of basic and acidicioamines from biological fluids due to their unique chemical prop-rties. Third, in many cases, the detection of trace amounts ofiogenic amines and their metabolites can be easily masked byigh-concentration biological constituents; thus, the analytes of

nterest remain undetected. Due to these reasons, the profilingnalysis of biogenic amines and their acidic metabolites in humanrine has not been fully investigated.

To overcome these analytical challenges, we have attemptedo apply a selective derivatization method using HMDS/MBHFBAor the sensitive and selective detection of biogenic amines andheir acidic metabolites. It is well known that HMDS reagent isn effective TMS reagent for selective reactions on hydroxyl andarboxylic groups, and it does not react with amine groups [26].BHFBA was then applied for the selective derivatization of the

mine groups. Using these reagents, both acidic and basic biogenicompounds were successfully derivatized without any side prod-cts. For the effective extraction and selective cleanup of biogenicmines and their acidic metabolites from human urine, a mixedation exchange (MCX) SPE cartridge using different SPE eluentsas applied. Although MCX SPE cartridges have generally beensed for the extraction of basic compounds from biological samples,his SPE method was also applied for the simultaneous extraction ofoth biogenic amines and their acidic metabolites. GC/MS selected

on monitoring (SIM) was then applied for the sensitive and selec-ive detection of the targets of interest, selecting specific ions toeduce any potential matrix effects.

The aim of this study was to develop an analytical methodor the reliable quantification of biogenic amines and their acidic

etabolites in human urine by GC/MS. This study included (i) theevelopment of a novel derivatization method that prevented theormation of side products during the derivatization reaction tomprove detection sensitivity, (ii) the development of an extractionf the biogenic amines and their acidic metabolites from urine toffectively eliminate interferences using an MCX SPE cartridge, andiii) a demonstration of the effectiveness of the established methodpplied toward a diagnostic clinical test.

. Materials and methods

.1. Chemicals

Authentic catecholamines, dopamine hydrochloride (DA),omovanillic acid (HVA), 3,4-dihydroxy-l-phenylalanine (L-DOPA),,l-normetanephrine hydrochloride (NMN) and (±)-epinephrineydrochloride (EP) were purchased from Sigma Chemical

. A 1305 (2013) 234– 243 235

(St. Louis, MO, USA). d,l-noradrenaline hydrochloride (NE),3,4-dihydroxyphenylacetic acid (DOPAC), and 4-hydroxy-3-methoxymandelic acid (VMA) were obtained from Fluka (Buchs,Switzerland). Serotonin (ST) hydrochloride and its metabolite5-Hydroxyindole-3-acetic acid (5-HIAA) were also obtained fromSigma Chemical (St. Louis, MO, USA). Deuterium-labeled internalstandards (HVA-d5 and DA-d3) were obtained from CambridgeIsotope Laboratories (Woburn, MA, USA) and CDN Isotopes(Vaudreuil, Quebec, Canada), respectively. The purities of thesechemicals were greater than 97%. The biosynthesis pathways ofthese biogenic amines are depicted in Fig. 1.

The derivatization reagents N-methyl-N-(trimethylsilyl)-trifluoroacetamide (MSTFA) and hexamethyldisilazane (HMDS)were obtained from Sigma–Aldrich Chemical Co. (Milwaukee,WI, USA) and Supelco (Bellefonte, PA, USA), respectively. N-Methyl-bis(heptafluorobutyramide) (MBHFBA) was purchasedfrom Macherey-Nagel (Duren, Germany).

Ethyl acetate, methanol, and pyridine were obtained from J. T.Baker (Rockford, IL, USA). Hydrochloric acid was obtained fromMerck (Darmstadt, Germany), and phenanthrene-d10, used as arecovery internal standard, was obtained from Supelco.

2.2. Stock solutions

The biogenic amines and their metabolites were dissolved inmethanol at a concentration of 1 mg/mL. However, L-DOPA was dis-solved in methanol with the addition of 2 �L of 6 M HCl to improveits solubility. All standard solutions were stable for several weeks at−4 ◦C. The solutions were kept in amber vials to protect the materialfrom photooxidation.

2.3. Sample preparation and MCX SPE procedure

Urine samples from four healthy (two female and two male)subjects aged 24–31 years were collected without the additionof preservatives. Samples were stored in a deep freezer at −80 ◦Cuntil analysis. Prior to the experiments, the samples were thawedat room temperature. The labeled internal standards (HVA-d5 foracidic metabolites and DA-d3 for biogenic amines, each 0.1 �g)were added to 2 mL of urine sample, and then 150 �L of 0.1 MHCl was added. The resulting solution was heated at 90 ◦C for30 min. After the hydrolysis procedure, SPE was performed withmixed cation exchange (MCX, 1 mL, 30 mg) cartridges obtainedfrom Waters Corporation (Wexford, Ireland). The cartridges wereconditioned with 1 mL of methanol, followed by 1 mL of distilledwater. After adsorption of the sample, the cartridge was washedwith 1 mL of 0.1 M HCl and dried for 1 min under vacuum to removethe excess water. All analytes were retained on the solid surfaceof MCX cartridge by two mechanisms: п–п interactions with thephenyl rings and ionic interaction with the sulfonic acid moieties.Different pH values and volumes of the elution solutions weretested. The elution of the acidic metabolites from the MCX SPEcartridge was performed with 1 mL of methanol. The elution ofthe biogenic amines was then performed with 1 mL of 5% NH4OHin methanol (v/v). The eluate was evaporated to dryness under agentle stream of nitrogen gas.

2.4. Derivatization

An authentic standard mixture solution (each 1 �g) containinginternal standard was dried under nitrogen in a silylated reaction

vial prior to derivatization. The vial was dried under vacuum for5 min at room temperature. Trimethylsilylation (TMS) of the bio-genic amines was achieved by the addition of 20 �L of ethyl acetateand 40 �L of HMDS.

236 N.-H. Park et al. / J. Chromatogr. A 1305 (2013) 234– 243

biogen

HaprTapaw2

Fig. 1. Biosynthesis pathways of

To optimize the TMS derivatization of the biogenic amines byMDS, the influence of the reagent amounts, the reaction temper-ture and reaction time was investigated. The reaction mixture wasurged with argon for 30 s, tightly capped, and vortexed, and theeaction temperature was tested at 0, 20, 30, 40, 60, 80, 120 ◦C.he reaction time course for the TMS derivatization of the biogenicmines by HMDS at 40 ◦C was performed with varying reaction

eriods of 5, 10, 20, and 40 min. A total of 20 �L of MBHFBA wasdded to the solution to acylate the amine group, and the solutionas heated at 80 ◦C for 5 min. After cooling the reaction mixture,

0 �L of the internal standard phenanthrene-d10 (1 �g/mL) was

Fig. 2. Overall analytical procedure for the analysis of biogenic amin

ic amines and their metabolites.

added to the solution. The derivatized sample was injected into theGC/MS system. The overall analytical procedure for the determina-tion of biogenic amines and their acidic metabolites is depicted inFig. 2.

2.5. GC/MS conditions

An Agilent 5975 N (Palo Alto, CA, USA) mass spectrometer (EImode, 70 eV) connected to an Agilent 6890 gas chromatographequipped with a DB-5MS capillary column (15 m × 0.25 mm i.d.,0.25 �m film thickness, J & W Scientific, Folsom, CA, USA) was

es and their acidic metabolites in urine by GC/MS SIM mode.

N.-H. Park et al. / J. Chromatogr. A 1305 (2013) 234– 243 237

Table 1Retention times (min) and characteristic ions (m/z) of biogenic amines and their metabolite-OTMS,-NHFBA derivatives by GC/MS scan mode.

Compound RT (min) M.W. EI-characteristic ion value (%)

HVA 5.64 326 209(100), 326(77), 179(55), 311(46), 267(42), 296(17), 149(15), 237(11)DOPAC 6.29 384 179(100), 384(75), 267(73), 237(22), 207(12), 369(12), 193(4)DA 6.44 493 267(100), 193(52), 179(45), 493(41), 280(35), 478(5)VMA 6.94 414 297((100), 147(6), 267(4), 371(3), 399(2), 414(0.7)NMN 7.02 523 297((100), 298(16), 267(3), 508(2), 523(0.7)NEP 7.45 581 355((100), 297(8), 265(3), 281(2), 147(2), 476(2) 581(0.8)EP 7.89 595 355((100), 284(5), 240(4), 580(2), 267(2), 147(2), 595(0.4)L-DOPA 8.65 609 267(100), 179(33), 396(5), 594(4), 609(3), 219(2)5-HIAA 10.30 407 218((100), 335(38), 292(10), 320(8), 407(2), 392(0.8)ST 10.35 444 218((100), 444(17), 231(11), 146(4), 429(0.8)HVA-D5 5.61 331 214(100), 331(83), 184(56), 316(45), 272(40), 301(15), 242(10)DA-D3 6.41 496 270(100), 496(47), 182(46), 193(42), 283(33), 481(8)

B

u1aaaostocuatTwt

2

epaaSaibl1Lttvmbttiaw

3

3

sa

phenanthrene-d10 5.26 188

old: selected ion in SIM mode.

sed for analysis. The samples were introduced via split (ratio0:1) injection with the port heated at 280 ◦C. Helium was useds the carrier gas at a flow rate of 1.0 mL/min. The oven temper-ture, initially 120 ◦C and held for 2 min, was ramped to 175 ◦C at

rate of 10 ◦C/min and held for 1.5 min, raised to 250 ◦C at a ratef 15 ◦C/min, then ramped to 280 ◦C and held for 3 min. The masspectrometer interface temperature was set at 280 ◦C. The manifoldemperature was maintained at 230 ◦C. The mass spectrometer wasperated in scan mode from 70 to 650 amu. For the monitoring andonfirmation analysis, the selected ion monitoring (SIM) mode wassed. In GC/MS SIM mode, two characteristic ions for each biogenicmine and acidic metabolite were used for peak identification, andhe most abundant ion was selected for quantification (Table 1).he dwell time of each ion was set at 50 ms. The relative peak areaas obtained by dividing the integrated area of the base peak by

hat of the internal standard.

.6. Method validation

The established method was validated by determining the lin-arity, limits of detection (LOD), limits of quantification (LOQ),recision and accuracy. The linearity of the calibration was evalu-ted by combining 2 mL of urine with the analyte standard solutionst six different concentrations, subjecting the mixtures to thePE procedure, derivatization and analysis by GC/MS SIM. Thenalyte/IS peak area ratios were plotted against the correspond-ng concentrations, and the calibration curves were constructedy means of the least-squares method. The considered analyte

inearity ranges were 1–2000 ng/mL for the acidic metabolites,0–1000 ng/mL for DA, NE, ST and NMN and 20–2500 ng/mL for-DOPA and EP. The LODs and LOQs were determined at signal-o-noise (S/N) ratios of 3 and 10, respectively. The precision ofhe developed method was evaluated using intra- and inter-dayariations. The relative standard deviation was determined as aeasurement of the precision. The intra- and inter-day repeata-

ilities were determined on triplicates within one day and overhree consecutive days, respectively. To evaluate the accuracy ofhe analytical method, the recovery study was measured by spik-ng known amounts of standards in known real samples in triplicatet three concentration levels, 50, 250 and 1000 ng/mL. The mixtureas analyzed using the method described in Section 2.4.

. Results and discussion

.1. Optimization of derivatization conditions by HMDS/MBHFBA

Six derivatization techniques were investigated in our previoustudy to find an effective derivatization method for biogenicmine profiling analysis by GC/MS [27]. Among them, the

188(100), 160(8), 133(3)

two-step derivatization involving O-silylation and N-perfluoroacylation with MSTFA and MBHFBA, respectively,possessed several advantages compared to single derivatizationtechniques, including the production of thermally stable deriva-tives, improvements in the sensitivities of higher-mass fragmentions with decreasing interference effects, and improvements inthe chromatographic properties.

However, the MSTFA/MBHFBA method still produced side prod-ucts from the primary amine compounds, even though it was ahighly sensitive and selective derivatization. As seen in Fig. 3A andFig. 4A, biogenic amines containing primary amines, such as DA,NE, NMN, L-DOPA and ST, produced significant amounts of sideproducts when MSTFA/MBHFBA was used. Their correspondingside products were identified as the biogenic amine-OTMS,-NTMS,-NHFBA derivatives using GC/MS scan mode. The formation of theresulting side products was most likely due to the favorable reactiv-ity of the primary amines with MSTFA. The amounts of these sideproducts tended to increase as the reaction period and reactiontemperature increased. Thus, the irreproducible derivatization ofthe primary amines could lead to unreliable quantification results.

To prevent the formation of side products, an HMDS reagent wasapplied for the TMS derivatization of the biogenic amines. HMDS isknown to selectively derivatize hydroxyl or carboxylic acid groups,but not amine groups [16] due to the strong basicity of the aminegroup in HMDS. As seen in Fig. 4B, HMDS prevents the formation ofside product -NTMS derivatives (e) during the TMS derivatizationof the primary amines. In other words, the excess amount of HMDScannot further react with the biogenic primary amine compoundsto form any -OTMS,-NTMS derivatives because the basicity of theHMDS and the reaction intermediate of HMDS (f) is greater thanthe reaction products (b), as shown in Fig. 4B.

To optimize the TMS derivatization of the biogenic amine-OTMSconditions, the amount of HMDS reagent, reaction temperature,and reaction time were examined (Fig. 5). The overall reaction yieldfor the biogenic amine-OTMS derivatives increased proportionallyas the amount of HMDS was increased to 40 �L (Fig. 5A). However,the reaction yield did not increase at volumes of HMDS greaterthan 40 �L. The reaction temperature (Fig. 5B) and reaction time(Fig. 5C) of the HMDS derivatization of biogenic amines stronglyaffected the kinetics of the TMS derivatization reaction. As shownin Fig. 5B, the TMS derivatization yields were shown to decreasedat temperatures greater than 40 ◦C and reaction times longer than10 min. Temperatures greater than 40 ◦C can lead to a retardationof the TMS derivatization because HMDS is an endothermic reagent(�H = 42.40 kJ/mol) [28]. The longer reaction period produced a

white precipitate in the reaction solution that decreased the TMSderivatization yield.

The N-HFBA derivatization by MBHFBA was performed at 80 ◦Cfor 5 min by using previously optimized conditions [27]. The alkyl

238 N.-H. Park et al. / J. Chromatogr. A 1305 (2013) 234– 243

F BHFBp , (3) N4 -NTMS

arHdgdzbOd

ig. 3. TICs of biogenic primary amines with derivatization using (A) MSTFA/Mhenanthrene-d10, (2) dopamine-(OTMS)2,-NHFBA, 2′ . DA-(OTMS)2,-NTMS,-NHFBA′ . NE-(OTMS)3,-NTMS,-NHFBA, (5) L-DOPA-(OTMS)3,-NHFBA, 5′ . L-DOPA-(OTMS)3,

mine and primary amine groups of the biogenic amines wereeadily reacted to form the N-HFBA derivative. For serotonin and 5-IAA, the amine group of the indole ring did not react with MBHFBAue to low basicity of the indole amine. However, the indole amineroup of 5-HIAA did react with MSTFA to form the–OTMS,-NTMSerivative as a side product. Thus, in this study, a dual derivati-

ation was successfully applied for the selective derivatization ofioamine-OTMS,-NHFBA derivatives and their acidic metabolite-TMS derivatives. As seen in Fig. 3B, the single peaks of theerivatives were observed for the primary amine compounds of the

Fig. 4. Chemical derivatization of biogenic compounds containing primary

A and (B) HMDS/MBHFBA by GC/MS scan mode. Peak identities as follow: (1)MN-(OTMS)2,-NHFBA, 3′ . NMN-(OTMS)2,-NTMS,-NHFBA, (4) NE-(OTMS)3,-NHFBA,,-NHFBA, (6) ST-OTMS,-NHFBA and 6′ . ST-OTMS,-NTMS,-NHFBA.

biogenic amines and their acidic metabolites, in contrast to the mul-tiple peaks when derivatized with MSTFA/MBHFBA. The retentiontimes and characteristic ions of the biogenic amine-OTMS,-NHFBAderivatives and the acidic metabolite-OTMS derivatives are sum-marized in Table 1.

To compare the reaction yields of the biogenic amine deriva-

tives formed by HMDS/MBHFBA and MSTFA/MBHFBA, the relativeresponse factors were calculated by the peak area ratios of theanalytes versus the internal standard (phenanthrene-d10). Therelative response factors of the biogenic amine-OTMS,-NHFBA

amines using (A) MSTFA/MBHFBA and (B) HMDS/MBHFBA reagents.

N.-H. Park et al. / J. Chromatogr. A 1305 (2013) 234– 243 239

Fig. 5. Optimization of derivatization conditions of bioamines containing p

Table 2Area ratio of biogenic amine-OTMS,-NHFBA derivatives formed with differentderivatization reagents by GC/MS scan mode. (n = 3).

MSTFA/MBHFBA HMDS/MBHFBA

HVA 1.97 ± 7.7 1.81 ± 1.82DOPAC 1.62 ± 7.35 1.66 ± 2.42DA 3.41 ± 10.24 3.64 ± 1.65VMA 3.38 ± 6.07 3.32 ± 3.03NMN 4.03 ± 4.18 3.87 ± 3.89NE 4.86 ± 8.01 5.70 ± 3.00EP 3.09 ± 4.61 2.86 ± 2.21L-DOPA 1.09 ± 4.09 2.62 ± 2.27

dtEHwdcdp

3

iidchwia

5-HIAA 2.90 ± 8.70 4.95 ± 1.03ST 4.09 ± 3.31 4.80 ± 1.37

erivatives produced by HMDS/MBHFBA were equivalent or betterhan those produced by MSTFA/MBHFBA, as indicated in Table 2.specially, the relative response factors of the bioamine O-TMS/N-FBA derivatives of the primary amines formed by HMDS/MBHFBAere higher than those formed by MSTFA/MBHFBA. This result wasue to the HMDS/MBHFBA derivatization of the primary amineompounds producing single products, but the MSTFA/MBHFBAerivatization produced both bioamine-OTMS/-NHFBA as a majorroduct and -OTMS,-NTMS,-NHFBA derivatives as side products.

.2. Sample cleanup by MCX SPE

In this study, the applicability of MCX SPE cartridges was exam-ned for the cleanup of bioamines and their acidic metabolitesn urine. The MCX adsorbent, containing N-divinylpyrrolidone,ivinylbenzene, and sulfonic acid groups, can be useful for theleanup of acidic and basic components in biological samples viaydrogen bonds, van der Waals interactions, and ionic interactions

hen appropriate elution solvents are used. The mechanisms of

nteractions between the biogenic amines and acidic metabolitesnd the MCX sorbent are depicted in Fig. 6.

rimary amines using HMDS reagent followed by HFBA derivatization.

The elution patterns of the acidic metabolites were studied usingthe MCX SPE with methanol solvent. After loading the sample ontothe MCX SPE cartridge, the carboxylic acid and benzene groups inthe acidic metabolites interact with the N-divinylpyrrolidone anddivinylbenzene of the MCX sorbent, respectively, through hydro-gen bonding and van der Waals interaction, respectively. The acidicmetabolites retained on the MCX SPE cartridges were easily elutedby methanol. For all the acidic metabolites, the elution recoverywith MCX SPE using 0.5 mL of methanol was found to be greaterthan 93%.

The elution patterns of the biogenic amines were also investi-gated using the MCX SPE cartridge with 5% ammonium hydroxidein methanol. During sample loading onto the MCX SPE at pH val-ues less than at least 3, most of the biogenic amines are convertedinto the protonated amine forms. Although L-DOPA has both car-boxylic acid and amine groups, its amine group can be protonated inacidic media. The positively charged amine groups of the biogenicamines strongly interact with the sulfonic anion of the MCX sor-bent. To elute the biogenic amines from the MCX SPE cartridges, abasic 5% NH4OH in MeOH eluent was used to cleave the electrostaticinteractions between the ammonium ion and sulfonic anion. Mostof the basic compounds, except for serotonin, were readily elutedwithin 0.5 mL of 5% NH4OH in methanol. Serotonin contains twoamine groups and interacted more strongly with the MCX sorbentcompared to the other basic compounds. Thus, to effectively elutethe basic compounds, the amount of elution solvent was increasedto 1 mL. For all the biogenic amines, the elution recovery with theMCX SPE using 1.0 mL of 5% NH4OH in methanol was found to begreater than 92%.

The biogenic amines and their acidic metabolites in urine sam-ple were successfully purified on MCX SPE cartridges using differenteluting solvents. MCX SPE cleanup method could effectively applyfor the profiling analysis of biogenic amines and acidic metabolites,respectively, in urine even though sample procedure was relatively

complicated by the addition of SPE procedure, compared with HPLCbased analytical methods [7–10].

240 N.-H. Park et al. / J. Chromatogr. A 1305 (2013) 234– 243

Fig. 6. Three main interactions between MCX sorbent and (A) acidic metabolites and (B) biogenic amines.

Table 3Calibration curves, correlation coefficients, LODs and LOQs of biogenic amine and acidic metabolites-OTMS, -NHFBA derivatives obtained by GC/MS SIM.

Compound Linear range (ng/mL) Slope Intercept Correlation coefficient (r2) LOD (ng/mL) LOQ (ng/mL)

HVA 1–2000 0.104 0.1338 0.998 0.27 0.75DOPAC 1–2000 0.0062 −0.1321 0.996 0.12 0.68DA 10–1000 0.0272 1.4787 0.997 1.42 6.28VMA 1–2000 0.0067 0.1187 0.999 0.05 0.17NMN 10–1000 0.09 −1.466 0.999 0.92 2.85NEP 10–1000 0.0038 0.0098 0.998 1.15 5.48EP 20–2500 0.0045 0.2449 0.998 6.24 17.84L-DOPA 20–2500 0.0005 0.0801 0.997 4.34 13.885-HIAA 1–2000 0.0067 0.0564 0.997 0.07 0.27ST 10–1000 0.1412 2.4097 0.998 0.75 2.31

Table 4Inter- and intra-day data for acidic metabolite–OTMS,-NHFBA derivatives obtained by GC/MS SIM.

Compound Concentration (ng/mL) Intra-day (n = 3) Inter-day (n = 3)Mean Acuuracy (%) RSD (%) Mean Acuuracy (%) RSD (%)

VMA50 53.6 107.3 3.5 55.9 111.9 4.1

250 259.9 103.9 3.9 262.4 104.9 1.91000 1049.3 104.9 2.3 1052.6 105.3 2.8

DOPAC50 51.6 102.1 1.4 52.1 104.2 3.4

250 249.4 100.8 3.7 251.7 100.7 3.01000 964.5 96.5 1.5 951.9 95.2 2.3

HVA50 50.1 100.1 2.7 53.9 107.9 4.3

250 264.5 105.8 3.4 266.7 106.7 4.31000 1039.8 103.9 3.3 1053.4 105.3 3.6

5-HIAA50 47.3 94.6 1.3 47.5 95.1 2.3

250 239.2 95.7 4.9 241.3 96.5 3.21000 980.4 98.0 3.0 985.7 98.6 2.1

Table 5Inter- and intra-day data for biogenic amine–OTMS, -NHFBA derivatives obtained by GC/MS SIM.

Compound Concentration (ng/mL) Intra-day (n = 3) Inter-day (n = 3)Mean Acuuracy (%) RSD (%) Mean Acuuracy (%) RSD (%)

DA50 48.6 97.1 4.8 49.3 98.6 1.7

250 238.3 95.3 5.1 240.2 96.1 5.41000 963.7 96.4 2.1 970.6 97.1 3.6

L-DOPA50 50.2 100.4 3.0 46.1 92.3 3.5

250 242.5 97.0 3.5 238.6 95.4 4.91000 978.3 97.8 2.7 996.2 99.6 2.5

NE50 53.5 107.0 4.0 52.9 105.9 3.0

250 218.3 87.9 3.9 241.6 96.7 5.51000 879.1 87.9 2.9 899.8 89.9 3.5

NMN50 46.6 93.2 5.5 49.0 98.0 5.9

250 256.2 102.5 2.3 260.1 104.0 4.81000 964.0 96.4 1.8 967.2 96.7 2.3

EP50 47.9 104.4 6.9 48.5 96.9 5.1

250 256.6 102.6 3.4 254.2 101.7 3.91000 1004.8 100.5 1.9 978.3 97.8 1.8

ST50 49.5 99.0 5.8 51.4 99.6 3.0

250 247.6 95.6 2.2 239.4 98.2 2.51000 990.5 99.0 2.2 994.5 98.2 1.1

N.-H. Park et al. / J. Chromatogr. A 1305 (2013) 234– 243 241

F rivatim

3

t1f22fdayiiTc

F(

ig. 7. Extracted ion chromatograms of (A) m/z 179 and 384 for DOPAC-(OTMS)3 deode.

.3. Method validation

To evaluate the linearity of the developed method, calibra-ion spiked solutions (n = 3 each) with a concentration range of–2000 ng/mL urine (1, 50, 250, 500, 1000 and 2000 ng/mL urine)or the acidic metabolites; of 10–1000 ng/mL urine (10, 50, 100,50, 500 and 1000 ng/mL urine) for DA, NE, ST and NMN; and of0–2500 ng/mL urine (20, 50, 250, 500, 1000 and 2500 ng/mL urine)or EP and L-DOPA were prepared by spiking the stock solutionsiluted with human urine. The calibration curves were constructednd calculated according to the method of least squares, relating

(the peak area ratio of the biogenic amine-derivatives to the

nternal standard) to x (the concentration of the biogenic aminesn �g/mL). The calibration solutions were analyzed in triplicate.able 3 lists the data for the equations of the calibration curves,orrelation coefficients, and the limits of detection (LOD). The

ig. 8. TICs of (A) standard mixture of acidic metabolites (each 10 ng/mL) and (B) healthy h1) HVA-(OTMS)2, 1′ HVA-d5-(OTMS)2, (2) Dopac-(OTMS)3, (3) VMA-(OTMS)3, and (4) 5-H

ve and (B) m/z 193 and 493 for DA-(OTMS)2,-NHFBA in human urine by GC/MS SIM

calibration curves of the analytes showed good linearity withinthe given concentration ranges, with correlation coefficients (R2)greater than 0.996.

In many cases during GC/MS analysis, the LOD can be affectedby several factors, such as GC injection volume, injection splitratio, and final solvent volume. Therefore, the sensitivity ofthis developed method could be enhanced by adjusting thesefactors. The LODs that were obtained for this method rangedfrom 0.12 to 6.24 ng/mL and were lower or equivalent to thoseobtained by other HPLC-based methods [12,22]. Although recentlypublished GC/MS method using O-ethoxycarbonyl (EOC)/tert-butyldimethylsilyl (TBDMS) derivatization [11] provided higher

sensitivity for acidic metabolites than this established method.However, the previously developed method could not be appliedfor the analysis of biogenic amines due to less reaction yieldof EOC for amine compounds. The established method provided

uman urine obtained by GC/MS SIM. Peak identities as follow: IS. phenanthrene-d10

IAA-(OTMS)2.

242 N.-H. Park et al. / J. Chromatogr. A 1305 (2013) 234– 243

F y humd , (3) NS

rlcLtca

daa6da

3

tiwisstSibcrt

uFecwecSp

ig. 9. TICs of (A) standard mixture of biogenic amines (each 10 ng/mL) and (B) health10, (1) DA-(OTMS)2,-NHFBA, 1′ . DA-d3-(OTMS)2,-NHFBA, (2) NMN-(OTMS)2,-NHFBAT-OTMS,-NHFBA.

easonable sensitivity on both biogenic amines and acidic metabo-ites for clinical test. It is important to note that clinically significantoncentrations [9,29,30] of most target analytes (exept for EP and-DOPA) in urine are usually 10- to 100-fold higher concentrationshan the LODs achieved in this study. Thus, the established methodould be effectively applied for the profiling analysis of biogenicmines and their metabolites in human urine.

The relative standard deviations (RSDs) for the intra-and inter-ay variations in precision of the acidic metabolites and biogenicmines, are indicated in Tables 4 and 5. The recoveries of all thenalytes ranged from 87.9 and 111.9%, and the RSDs were less than.9%, which demonstrated that the present method was repro-ucible and accurate for determining biogenic amines and theircidic metabolites in urine samples.

.4. Method application

The analytical conditions were optimized for the determina-ion of trace amounts of biogenic amines and their metabolitesn human urine. In this study, the GC oven temperature program

as optimized to sufficiently separate the target compounds fromnterfering peaks. The GC column length was also tested for theeparation of target analyte derivatives. The 15 m capillary columnhowed better separation efficiency for the biogenic amines andheir metabolite derivatives than the 30 m capillary column (seeupplementary Fig. 1). Furthermore, the overall detection sensitiv-ty of the biogenic amines and their metabolite derivatives obtainedy the 15 m column was much higher than that of the 30 m capillaryolumn. Thus, the use of a 15 m column for GC/MS provided moreapid analyses, better sensitivity, and increased separation powerhan using a 30 m column.

Following the method developed and described above, humanrine samples were extracted, purified, derivatized, and analyzed.or the sensitive detection by GC/MS SIM, two abundant ions forach biogenic amine and metabolite were selected for quantifi-ation and peak confirmation. The base peak for each compoundas typically used for quantification in SIM mode (Table 1). How-

ver, in certain cases, the base peak ion or second abundant ionhromatograms could be interfered with other matrix peaks in theIM chromatogram. For example, the ion chromatogram of the baseeak at m/z 179 for dopac-(OTMS)3 exhibited higher baseline and

an urine obtained by GC/MS-SIM mode. Peak identities as follow: IS. phenanthrene-E-(OTMS)3,-NHFBA, (4) EP-(OTMS)3,-NHFBA, (5) L-DOPA-(OTMS)3,-NHFBA, and (6)

overlapped with other peaks, as shown in Fig. 7A. To eliminateany possible interferences, an appropriate abundant ion (m/z 384),instead of the base peak at m/z 179, was selected for quantitationfor the SIM analysis. As another example, the ion chromatogramof m/z 193 as the second abundant ion for DA-(OTMS)2,-NHFBAderivative exhibited a high baseline (Fig. 7B). By the selection of ahigher-mass ion at m/z 493 instead of m/z 193 or 179, the peak couldbe clearly detected, providing positive the peak identification andthe precise quantification of DA. Fig. 7 shows two components thatwere baseline separated from other co-extractive interferences bythe selection of appropriate higher mass ions. Although the peakintensities of the two components were reduced by selecting theappropriate abundant ions, the selectivity and specificity were sig-nificantly improved.

The SIM chromatograms and extracted ion chromatograms ofthe acidic metabolite-OTMS derivatives (Fig. 8) and biogenic amine-OTMS,-NHFBA derivatives (Fig. 9) from human urine samples wereobtained using the developed method. The peak abundance of thephenanthrene-d10 was used to estimate the overall recovery ofthe analytes from the urine sample using the analytical procedure.The deuterium-labeled HVA-d5 and DA-d3 compounds were usedas internal standards of the acidic metabolites and the biogenicamines, respectively, which improved the precision and accuracyof the quantification. In practice, the LOQs of most target analyteswere at least 10-fold lower than the clinically significant concen-trations [9,29,30]. In view of the recoveries and sensitivities ofthe target analytes and the removal of interferences, the MCX SPEcleanup and selective derivatization approach was effective for thereliable profiling analysis of bioamines and their acidic metabolitesin human urine.

4. Conclusion

A comprehensive analytical method for the determination ofbioamines and their acidic metabolites in urine was developedusing selective derivatization, MCX SPE cleanup and GC/MS in SIMmode. A selective derivatization using HMDS/MBHFBA reagents

provided improvements in the chromatographic properties andMS detection sensitivity without any side product formation, thusenabling the reliable quantification of bioamines and their acidicmetabolites in human urine by GC/MS. The sample cleanup of

atogr

humet

itnatmtmo

A

Fis

A

fj

R

[[

[

[[

[

[

[[

[

[[

[

[

[

[[

N.-H. Park et al. / J. Chrom

uman urine was effectively performed with an MCX SPE proceduresing methanol eluent for the acidic metabolites and 5% NH4OH inethanol eluent for the biogenic amines. The established method

nabled the effective extraction, purification and selective deriva-ization of both biogenic amines and acidic metabolites in urine.

The specificity of SIM mode through the selection of appropriateons provided an improvement in the analytical performance inerms of linearity and sensitivity by the increase in the signal-to-oise ratios. The development of a reproducible analytical methodnd the incorporation of isotopically labeled internal standards ledo improvements in accuracy and precision. Thus, this proposed

ethod can be used for the reliable diagnosis of neuronal diseaseshrough the measurement of both biogenic amines and their acidic

etabolites in biological samples, allowing for further applicationsf clinical tests of neuronal diseases due to metabolic alterations.

cknowledgments

This study was financially supported by the National Researchoundation of Korea (Development of analytical method for profil-ng analysis of biogenic amines and their metabolites in biologicalamples, No. 2011-0012671).

ppendix A. Supplementary data

Supplementary data associated with this article can beound, in the online version, at http://dx.doi.org/10.1016/.chroma.2013.07.003.

eferences

[1] K. Hyland, J. Inher. Metab. Dis. 22 (1999) 353.[2] T. Nakazato, A. Akiyama, Brain Res. 930 (2002) 134.[3] W.M. Manger, R.W. Gifford, J. Clin. Hypertens 4 (2002) 62.

[

[[[

. A 1305 (2013) 234– 243 243

[4] Y. Jinsmaa, V.R. Florang, J.N. Rees, L.M. Mexas, L.L. Eckert, E.M.G. Allen, D.G.Anderson, J.A. Doorn, Chem-Biol. Interact. 192 (2011) 118.

[5] B. Thomas, M. Flint Beal, Hum. Mol. Genet. 16 (2007) 183.[6] M. Yoshitake, H. Nohta, H. Yoshida, T. Yoshitake, K. Todoroki, M. Yamaguchi,

Anal. Chem. 78 (2006) 920.[7] E. Tareke, J.F. Bowyer, D.R. Doerge, Rapid Commun. Mass Spectrom. 21 (2007)

3898.[8] L. Liu, Q. Li, N. Li, J. Ling, R. Liu, Y. Wang, L. Sun, X.H. Chen, K. Bi, J. Sep. Sci. 34

(2011) 1198.[9] Y. Sakaguchi, H. Yoshida, T. Hayama, M. Itoyama, K. Todoroki, M. Yamaguchi,

H. Nohta, J. Chromatogr. A 1218 (2011) 5581.10] V. Carrera, E. Sabater, E. Vilanova, M.A. Sogorb, J. Chromatogr. B 847 (2007) 88.11] D.T. Nguyen, I.S. Cho, J.W. Kim, K.R. Kim, G. Lee, M.J. Paik, Biomed. Chromatogr.

27 (2013) 216.12] E. Combet, M.E.J. Lean, J.G. Boyle, A. Crozier, D.F. Davidson, Clin. Chim. Acta 412

(2011) 165.13] K.C. Joanna, Crit. Rev. Anal. Chem. 41 (2011) 114.14] M.E.P. Hows, L. Lacroix, C. Heidbreder, A.J. Organ, A.J. Shah, J. Neurosci. Meth.

138 (2004) 123.15] R.N. Xu, L. Fan, M.J. Rieser, T.A. El-Shourbagy, J. Pharmaceut. Biomed. 44 (2007)

342.16] E. Chambers, D.M. Wagrowski-Diehl, Z. Lu, J.R. Mazzeo, J. Chromatogr. B 852

(2007) 22.17] A. El-Beqqali, A. Kussak, M. Abdel-Rehim, J. Sep. Sci. 30 (2007) 421.18] J. Bicker, A. Fortuna, G. Alves, A. Falão, Anal. Chim. Acta (2013),

http://dx.doi.org/10.1016/j.aca.2012.12.030.19] A.A. Zoerner, K. Heusser, F.M. Gutzki, A. Mitschke, J. Tank, D.O. Stichtenoth, J.

Jordan, D. Tsikas, J. Chromatogr. B 879 (2011) 1444.20] E.C.Y Chan, P.Y. Wee, P.Y. Ho, P.C. Ho, J. Chromatogr. B 749 (2000) 179.21] M.M. Kushnir, F.M. Urray, E.L. Frank, W.L. Roberts, B. Shushan, Clin. Chem. 48

(2002) 323.22] M. Monteleone, A. Naccarato, G. Sindona, A. Tagarelli, Anal. Chim. Acta 759

(2013) 66.23] L. Mercolini, G. Gerra, M. Consorti, L. Somaini, M.A. Raggi, Talanta 78 (2009)

150.24] M.A. Saracino, R. Mandrioli, L. Mercolini, A. Ferranti, A. Zaimovic, C. Leonardi,

M.A. Raggi, J. Pharm. Biomed. Anal. 42 (2006) 107.25] Z.D. Clark, E.L. Frank, J. Chromatogr. B 879 (2011) 3673.26] S.T. Kadam, S.S. Kim, J. Organomet. Chem. 694 (2009) 2562.

27] S.Y. Park, B.X. Kang, Q. Li, H.S. Kim, J. Lee, J. Hong, Bull. Korean Chem. Soc. 30

(2009) 1497.28] R.S. Myers, H.L. Clever, J. Chem. Eng. Data 14 (1969) 161.29] M.J. Whiting, Ann. Clin. Biochem. 46 (2009) 129.30] J. Gabler, A. Miller, S. Wang, Clin. Chem. Lab. Med. 49 (2011) 1213.