Accepted Manuscript Title: A generic screening methodology for horse doping control by LC-TOF-MS, GC-HRMS and GC-MS Author: Maroula K. Kioussi Emmanouil M. Lyris Yiannis S. Angelis Maria Tsivou Michael A. Koupparis Costas G. Georgakopoulos PII: DOI: Reference:

S1570-0232(13)00546-1 http://dx.doi.org/doi:10.1016/j.jchromb.2013.10.008 CHROMB 18575

To appear in:

Journal of Chromatography B

Received date: Revised date: Accepted date:

18-6-2013 12-9-2013 8-10-2013

Please cite this article as: M.K. Kioussi, E.M. Lyris, Y.S. Angelis, M. Tsivou, M.A. Koupparis, C.G. Georgakopoulos, A generic screening methodology for horse doping control by LC-TOF-MS, GC-HRMS and GC-MS, Journal of Chromatography B (2013), http://dx.doi.org/10.1016/j.jchromb.2013.10.008 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

A generic screening methodology for horse doping control by LC-TOF-MS, GCHRMS and GC-MS a,b*

Maroula K. Kioussi

a

a

a

, Emmanouil M. Lyris , Yiannis S. Angelis , Maria Tsivou , Michael Α.

b

c*

a

ip t

Koupparis , Costas G. Georgakopoulos

Doping Control Laboratory of Athens, Olympic Athletic Centre of Athens “Spyros Louis” (OAKA),

b

cr

Kifissias 37, 15123 Maroussi, Greece.

Laboratory of Analytical Chemistry, Department of Chemistry, University of Athens, 15771

Anti Doping Laboratory, P.O Box 27775 Doha, State of Qatar.

an

c

us

Panepistimiopolis-Zographou, Athens, Greece.

* Corresponding author: M.K. Kioussi; E-mail address: [email protected]; Tel:+30-210-

M

6853074; Fax:+30-210-6834567.

ABSTRACT

d

In the present study a general screening protocol was developed to detect prohibited substances

te

and metabolites for doping control purposes in equine sports. It was based on the establishment of a unified sample preparation and on the combined implementation of liquid and gas

Ac ce p

chromatographic MS analysis. The sample pretreatment began with two parallel procedures: enzymatic hydrolysis of sulfate and glucuronide conjugates, and methanolysis of the 17β-sulfate steroid conjugates. The extracts were treated for LC-TOF-MS, GC-HRMS and GC-MS assays. The majority of the prohibited substances were identified through a high mass accuracy technique, such as LC-TOF-MS, without prior derivatization. The sample preparation procedure included the formation of methylated and trimethylsilylated derivatives common in toxicological GC-MS libraries. The screening method was enhanced by post-run library searching using Automated Mass spectral Deconvolution and Identification System (AMDIS) combined with Deconvolution Reporting Software (DRS). The current methodology is able to detect the presence of more than 350 target analytes in horse urine and may easily incorporate a lot of new substances without changes in chromatography. The full scan acquisition allows retrospective identification of prohibited substances in stored urine samples after reprocessing of the acquired data. Validation

1

Page 1 of 32

was performed for sixty representative compounds and included limit of detection, matrix interference - specificity, extraction recovery, precision, mass accuracy, matrix effect and carry over contamination. The suitability of the method was demonstrated with previously declared

ip t

positive horse urine samples.

cr

Keywords: screening; equine; urine; doping control; mass spectrometry; spectra deconvolution

1. INTRODUCTION

us

Accredited horse racing doping control laboratories have to examine hundreds of banned compounds in biological specimens. The prohibited list of Fédération Equestre Internationale (FEI)

an

[1] is not limited to 1156 banned and controlled medication substances, but also extends to other compounds with similar chemical structure or similar biological effect. Moreover, according to Article 6 of the International Agreement on Breeding, Racing and Wagering (IABRW) published by

M

the International Federation of Horseracing Authorities (IFHA), masking agents, as well as any substance that has the potential to affect the performance of a horse, acting on one or more of the

d

mammalian body systems are forbidden [2]. Furthermore, the presence of metabolites in urine

te

may provide additional information and evidence of drug abuse, even though the parent drug may not be detected. Therefore, doping control laboratories have to detect analytes with different

Ac ce p

physicochemical properties, with different ranges of polarity, acid-base properties, molecular size or solubility and pharmaceutical nature. Different types of mass spectrometers coupled with liquid and gas chromatography accomplish the analytical requirements in the particular field and they have become indispensable tools for doping control analysis [3-14]. The available libraries of standardized electron impact mass spectra are also useful for drug testing [15, 16]. The establishment of comprehensive screening approaches that allow the simultaneous

detection of as many prohibited compounds as possible in equine urine is essential. Nonetheless, procedures providing the simultaneous detection of several drug classes, which were traditionally handled separately, appear to be relatively limited in doping control in equine sports [17-21]. The development of a suitable sample preparation for broad screening is an uphill task allowing for the numerous substances with different properties to be examined, the interferences from the complex horse urine matrix and the low concentrations of the drugs to be screened.

2

Page 2 of 32

The existing screening methods for the identification of multi-class analytes in horse urine samples comprise a single deconjugation procedure and subsequently either a simultaneous analysis by LC-MS techniques with polarity switching [19, 20] or different separate analyses for basic and acidic/neutral fractions [17, 18]. Regarding LC-TOF-MS systems applied in drug

ip t

screening, the identification of the analytes is based on the accurate mass and the isotopic pattern of the detected compounds by conducting library searching in databases [13, 14]. Although LC-MS

cr

methods provide wide drug coverage, there are still compounds not detected due to low ionization efficiency. GC-MS methods have also been adopted for the identification of individual groups of

us

prohibited substances, such as anabolic steroids [6].

The aim of the hereby study was the development of a generic methodology for the

an

screening of a large number of prohibited compounds with different therapeutic activities and acidic/basic properties. The current approach has the advantage of the determination of over 350 drug targets, by using a unified sample preparation procedure that combines different

M

deconjugation and derivatization procedures. The analysis of doping agents was accomplished by LC-TOF-MS, GC-HRMS and GC-MS techniques that acted complementarily. The majority of the

d

prohibited substances were identified without prior derivatization and through a high mass

te

accuracy technique such as LC-TOF-MS in full scan mode which enables retrospective analysis. Furthermore, detection of prohibited drugs not targeted by the above screening method was

Ac ce p

achieved through library searching by using Automated Mass spectral Deconvolution and Identification System (AMDIS) incorporated into the Deconvolution Reporting Software (DRS).

2. MATERIALS AND METHODS 2.1. Materials

Reference materials were purchased from various suppliers which are presented in detail

in Supplement 1. Stock solutions of the reference substances were prepared in methanol and stored at -20 ºC. Multicompound solutions from the above stock solutions were also prepared. βGlucuronidase - arylsulfatase isolated from Helix pomatia was purchased from Sigma-Aldrich, -1

Steinheim, Germany (Type HP-2; 197114 units mL

-1

β-glucuronidase and 876 units mL

arylsulfatase). Isolute C18(EC) solid phase extraction columns (sorbent mass 500 mg) were obtained from Biotage AB (Uppsala, Sweden). Solvents used in the urine extraction procedures

3

Page 3 of 32

were of analytical grade and were acquired from Labscan (Dublin, Ireland). Sodium hydrogen carbonate (NaHCO3, 99.7%-100.3%) and sodium carbonate anhydrous (Na2CO3, 99.8%) were supplied by Panreac (Barcelona, Spain). Sodium sulfate anhydrous (Na2SO4, >99%) was purchased from Lach-Ner, S.r.o. (Neratovice, Czech Republic). Potassium carbonate (K2CO3, 99-

ip t

101%), methanolic hydrochloride (3N), iodomethane (CH3I, >99.0%), ammonium iodide (NH4I, 99.999%), ammonium formate (HCOONH4, 97%) and formic acid (HCOOH, 98-100%) were

cr

obtained from Sigma-Aldrich (Steinheim, Germany). N-methyl-N-trimethylsilyltrifluoroacetamide (MSTFA, 98-100%) was acquired from Pierce Biotechnology (Rockford, IL, USA). 1-Propanethiol

us

(PrSH) was supplied from Merck (Darmstadt, Germany). HPLC-grade water was obtained by purifying water in filtration system (Millipore, Billerica, MA, USA) and LC-MS grade acetonitrile was

an

from Carlo Erba Reagenti SpA (Rodano, Italy).

2.2. Instrumentation

M

2.2.1. LC-TOF-MS

The Liquid Chromatograph was an Agilent 1200 Series rapid resolution LC system

d

(Agilent Technologies, Waldbronn, Germany) comprising a vacuum degasser, a high-pressure

te

binary pump, an auto-sampler with a cooled sample tray at 12 ºC and a column compartment kept at 35 ºC. Chromatographic separation was carried out on a reversed phase Zorbax Eclipse Plus

Ac ce p

C18 column (100 x 2.1 mm i.d., 1.8 μm particle size; Agilent Technologies) protected with an inline filter (0.5 μm) (Agilent Technologies, Waldbronn, Germany). The mobile phase was composed of 5 mM ammonium formate / 0.1% formic acid in water (solvent A) and 5 mM ammonium formate / 0.1% formic acid in a mixture of acetonitrile:water, 90:10 v/v (solvent B). A linear gradient was run -1

at flow rate of 0.3 mL min , with 10% solvent B at initial condition (t=0 min), increasing to 80% in 9 min. Then, the proportion of solvent B increased to 100% from t=9 min to t=10 min, where it was held for 3 min. The gradient was then returned to 10% solvent B within 0.5 min and stabilized for 3.5 min (post-run equilibrium time) before the next injection. The injection volume was 5 μL. Liquid Chromatograph was coupled with an orthogonal acceleration quadrupole time-offlight mass spectrometer (6520 Accurate-Mass QTOF/MS; Agilent Technologies, Santa Clara, CA, USA) through an orthogonal electrospray ionization (ESI) interface. The instrument was operated in positive mode. Nitrogen was used both as a nebulizer gas (pressure 40 psi) and drying gas

4

Page 4 of 32

-1

(temperature 330 ºC, flow rate 10 L min ). Capillary voltage of the ion source was set at 3500 V and the fragmentor voltage at 140 V in order to reduce the fragmentation of the pseudo-molecular ions of the target analytes. The applied skimmer and octapole 1 RF voltages were 65 V and 750 V, respectively. All the other MS parameters were automatically optimised by the instrument

ip t

autotuning procedure, performed on a monthly basis. The mass spectral data were acquired within the range of m/z 100-1100.

cr

Calibration of TOFMS was performed prior to each sequence run covering the mass range of m/z 118.0863-1521.9715 by using a calibration solution provided by the manufacturer (ESI-L

us

Low Concentration Tuning Mix, G1969-85000, Agilent Technologies). The full width at half maximum (FWHM) mass resolution ranged from 4700 (at m/z 118.0863) to 12600 (at m/z

an

2721.8948). Reference mass correction was accomplished during the analysis, to achieve better mass accuracies, by continuously introducing two reference compounds, hexakis (1H,1H,3Htetrafluoropropoxy)

phosphazine

(Agilent

Technologies)

at

m/z

922.0098

and

M

benzyldimethylphenylammonium chloride (Sigma-Aldrich) at m/z 212.1434, simultaneously with

te

2.2.2. GC-MS

d

the samples into the ESI source from a second orthogonal nebulizer.

2.2.2.1. GC-MS for methylated derivatives

Ac ce p

The analysis of the methyl-derivatized target substances was performed on an Agilent

6890N GC system (Agilent Technologies, Palo Alto, CA, USA) coupled with an Agilent 5973 inert Mass Selective Detector. Chromatographic separation was implemented with a bonded and crosslinked 5% diphenyl, 95% dimethyl siloxan capillary column (12 m x 0.200 mm i.d., 0.33 μm film -1

thickness, HP ULTRA 2) with helium used as carrier gas at constant flow 0.6 mL min . A 2.0 μL sample volume was injected in split mode (8:1). The front inlet and the MSD transfer line heater temperatures were set at 250 and 310 ºC, respectively. The column oven temperature was -1

programmed to increase from 110 to 300 ºC at 15 ºC min and maintained at 300 ºC for 3 min. The run time was 15.67 min, inclusive of equilibration time (0.5 min). MS source and quadrupole temperatures were maintained at 230 and 150 ºC, respectively. The MS system was acquiring -1

data in full scan mode from 50 to 500 amu, at a rate of 3.25 scans s

5

and electron ionization

Page 5 of 32

energy of 70 eV. Perfluorotributylamine (PFTBA) mass-spectrometric grade (Agilent Technologies) was used as tuning standard.

2.2.2.2. GC-MS for trimethylsilylated derivatives

ip t

The trimethylsilylate (TMS) derivatized compounds were assayed by GC-MS in full scan mode (40-550 amu), for library searching purposes utilizing AMDIS and DRS programs. An Agilent

cr

6890N GC system (Agilent Technologies, Palo Alto, CA, USA) combined with an Agilent 5973

quadrupole Mass Selective Detector (MSD) and equipped with a bonded and cross-linked 5%

us

diphenyl, 95% dimethyl siloxan capillary column (19 m x 0.200 mm i.d., 0.33 μm film thickness, HP ULTRA 2) was used. GC was operating at constant flow of helium with a flow rate of 1.0 mL min

-1

an

and pressure of 10.43 psi. Two microlitres of sample were injected in split mode of 15:1 at 250 ºC. -1

Initial oven temperature was 100 ºC, then ramped at 20 ºC min to 290 ºC, held for 5.0 min. MSD transfer line heater, MSD source and quadrupole temperatures were maintained at 300, 230 and

M

150 ºC, respectively. The run time was 14.50 min. Perfluorotributylamine (PFTBA) mass-

d

spectrometric grade (Agilent Technologies) was used as tuning standard.

te

2.2.3. GC-HRMS

The assay of the TMS-derivatized target substances was carried out with a three-sector

Ac ce p

(electrostatic–magnetic–electrostatic) reverse geometry double focusing mass spectrometer (Micromass AutoSpec-Ultima) coupled to an Agilent 6890N network GC system (Agilent Technologies, Palo Alto, CA, USA). The GC was equipped with a bonded and cross-linked 100% dimethylpolysiloxan capillary column (12 m x 0.200 mm i.d, 0.33 μm film thickness, HP ULTRA 1). -1

Helium was used as carried gas at a flow rate of 1.1 mL min . The injector port was heated at 250 ºC and the transfer line at 350 ºC. One microlitre of sample was injected in split mode (40:1). Initial -1

oven temperature was 150 ºC for 0.5 min, then ramped at 12.5 ºC min to 310 ºC and held for 2.5 min. The MS source temperature was maintained at 220 ºC. The HRMS system was acquiring data in 10,000 resolution in selected ion recording (SIR) mode with eleven function groups. Perfluorokerosene (PFK) mass-spectrometric grade (Apollo, Manchester, UK) was used as calibration reference substance.

6

Page 6 of 32

2.3. Deconvolution software Spectral deconvolution of the MS data is a mathematical technique which separates overlapping mass spectra and contributes to the identification of analytes that are buried under coeluting matrix compounds. The fully automated deconvolution process includes treatment of noise,

ip t

correction for base line drift and extraction of pure component spectra and related information (eg peak shape, retention time), from complex chromatograms. This individual information can be

cr

library searched and matched against the spectra of the library.

In the present study an Agilent G1716 MSD-DRS reporting application was used,

us

combining results from the Agilent GC-MSD ChemStation (G1701DA), the AMDIS32 (version 2.1) and the NIST 2005 Mass Spectral Search Program (NIST05, version 2.0d). DRS worked with

an

three reference libraries: Association of Official Racing Chemists AORC_R6 Mass Spectral Library, NIST/EPA/NIH Mass Spectral library (NIST05) and an in-house library comprising the locked retention time and the mass spectrum of each entry. NIST05 Mass Spectral Library

M

contains 190,825 electron ionization (EI) spectra of 163,198 different chemical compounds [22]. AORC_R6 Mass Spectral Library and in-house library contain 6,840 and 377 entries, respectively.

te

derivatized samples.

d

DRS run in data reprocessing mode after GC-MS full scan analyses of the methyl- and TMS-

The automated identification through library searching and AMDIS-DRS programs was

Ac ce p

performed in duplicate. Primarily, the acquired data files of methylated and TMS derivatives were evaluated by DRS in a simple type of analysis based on matching the deconvoluted spectra with those of the reference libraries, with a minimum match factor of 70 and without participation of the retention times in the matching. At first, AMDIS deconvoluted the ChemStation data file and the resultant components were compared against the AORC_R6 Mass Spectral Library. This comparison used full scan spectra and was retention time independent. The results from the AORC library were also searched against a second library, NIST05 Mass Spectral Library. Finally, the generated DRS-report combined results from the processes described above, in one easy-toread format. The second automated detection by means of AMDIS-DRS programs pertained only TMS-derivatized substances and it was based on matching both the deconvoluted spectra and the retention times with those of the in-house library. The identification was accomplished by incorporating the retention index RI of the detected components in the search algorithm. The “use

7

Page 7 of 32

RI Calibration Data” analysis type was selected which included the deconvoluted spectra and the retention times in the matching. The AMDIS-based data analysis was performed with a minimum match factor of 30. A filter was also set in AMDIS, requiring the analyte’s retention times to fall within a time window of ±5s. When the difference between the compared retention times was

ip t

greater than ±5s, a penalty of 70 was abstracted from the match factor. Thus, the match factor

cr

was lower than 30 and the particular substance was not included in the detected compounds.

2.4. Sample preparation

Supplementary Figure 1. Procedure A

us

Urine samples were prepared according to

comprised preparation of 5.0 mL equine urine fortified with etiocholanolone glucuronide (5 μg),

an

mefruside (1 μg) and nalidixic acid (3.5 μg) as the internal standards (IS). For hydrolysis urine was adjusted to pH 4.8-5.5 with 1 M acetate buffer and incubated at (50±2) ºC for 3 h after adding 100 μL of glucuronidase/arylsulfatase from H. pomatia. After cooling, pH was adjusted to 9.0-10.0 with

M

a mixture of NaHCO3:Na2CO3 (10:1 w/w). Basic liquid-liquid extraction (LLE) was carried out with 7.0 mL ethylacetate, using anhydrous Na2SO4 for salting out and shaking for 10 min. The mixture

d

was centrifuged for 10 min and the organic phase was separated into three fractions (2.2 mL each)

te

and evaporated to dryness under a nitrogen stream at (50±5) ºC. The first fraction was reconstituted with 300 μL of a mixture of solvent A / solvent B (80/20 v/v) and was analyzed by LC-

Ac ce p

TOF-MS assay in positive ESI mode. The second fraction was methyl-derivatized with 20-30 mg K2CO3, 150 μL acetone and 40 μL CH3I under heating at hot block at (110±5) ºC for 30 min and then it was submitted to GC-MS analysis. The acquired data files were also reprocessed by AMDIS and DRS algorithms. The remaining third fraction was mixed with the extract from procedure B.

Regarding procedure B, a second aliquot of 2.5 mL horse urine was fortified with

androsterone sulfate (0.5 μg) and methyltestosterone (0.25 μg) as the IS. The urine was then diluted with 2.0 mL acetate buffer 1 M and the pH was adjusted to 4.8-5.5. Solid phase extraction (SPE) was performed on C-18 cartridges pre-conditioned with 5.0 mL methanol and 5.0 mL acetate buffer (1 M, pH 5.2). After loading the urine, the cartridge was washed with 5.0 mL acetate buffer (1 M, pH 5.2) followed by 5.0 mL n-hexane and the compounds of interest were eluted with 5.0 mL methanol. The eluate was evaporated under nitrogen at (60±5) ºC. Methanolysis was

8

Page 8 of 32

accomplished by dissolving the residue in 1.0 mL anhydrous methanolic HCl (3 M) and incubating at (60±5) ºC for 30 min. The basic extraction was implemented with 4.0 mL diethylether after pH adjustment at 9.5-10 with carbonate buffer (30 % w/v) and addition of 700 μL NaCl (4 M) for salting out. The organic extract was mixed with the third fraction from procedure A and evaporated

ip t

to dryness under a nitrogen stream at (50±5) ºC. The residue was dried in desiccator over P2O5 overnight. TMS-derivatization was performed by adding 150 μL MSTFA:NH4I:PrSH (1000:2:3,

cr

v/w/w) and incubating at (110±5) ºC for 60 min. Per-TMS derivatives were analyzed by GC-HRMS.

Additionally, the TMS-derivatized samples were subjected to GC-MS analysis and the acquired

us

full-scan data files were evaluated by AMDIS and DRS programs.

In both procedures, addition of CH3COOH (3 M in ethylacetate) solution preceded organic

an

layer evaporation in order to prevent loss of volatile compounds.

2.5. Method validation

M

The current screening method was validated according to Eurachem guidelines for validation of qualitative chromatographic methods [23]. Prior to validation, a tentative experimental

d

day was accomplished to two gelding horse urine samples fortified with the reference compounds -1

te

listed on Table 1, at seven concentration levels (5, 10, 25, 50, 100, 250, 500 ng mL ). Afterwards, a minimum concentration, Cmin, at which each compound was detected at both urine samples, was

Ac ce p

determined.

2.5.1. Limit of Detection – Method sensitivity In order to determine the limit of detection (LOD) ten different blank equine urine

substrates were fortified with the reference compounds at six concentration levels (0.25×Cmin, 0.5×Cmin, 1×Cmin, 2.5×Cmin, 5×Cmin, 10×Cmin) and prepared along with the respective blanks. The LOD was defined as the lowest concentration level at which a compound could be identified in all ten urine samples with a signal-to-noise (S/N) ratio above 3 in the selected ion chromatogram.

2.5.2. Matrix interference - Specificity

9

Page 9 of 32

Ten different drug-free urine specimens from three mares, three male and four gelding horses were prepared in order to check for possible naturally occurring interfering substances at the retention times of the examined analytes and at the selected extracted ion masses.

ip t

2.5.3. Extraction Recovery Extraction recovery was investigated at three concentration levels (5×LOD, 10×LOD,

cr

20×LOD) on three days by analyzing in duplicate six different urine samples which had been spiked with the target analytes either before or after sample extraction. Spiking to the organic

us

phase after LLE corresponded to 100% recovery. To evaluate extraction recovery, the peak area response of the analyte spiked initially before extraction into the urine sample was compared to

an

the peak area of the same analyte spiked after extraction into the blank urine extract.

2.5.4. Repeatability

M

The repeatability (intra-day and inter-day precision) of the method was assessed by analyzing in duplicate six aliquots of an equine specimen at three concentrations (5×LOD,

d

10×LOD, 20×LOD) on three different days over a two-month period. The intra-day and inter-day

te

precision were expressed as percentage relative standard deviations (RSD) of the relative

Ac ce p

retention times (RRTs) and the peak areas of the studied analytes.

2.5.5. Mass accuracy

Mass accuracy regarding only LC-TOF-MS analysis, was evaluated on three days in six

different urine substrates fortified at 10×LOD concentration.

2.5.6. Matrix Effect

Matrix effect associated with LC-MS analysis, was assessed at three concentration levels

(5×LOD, 10×LOD, 20×LOD) on three days by analyzing in duplicate six different urine samples which were first extracted and then fortified with the studied compounds. Matrix effect was expressed as the ratio of the peak response in the presence of matrix to the peak response in the absence of matrix in the neat reconstitution solvent. [24]

10

Page 10 of 32

2.5.7. Carry over contamination Carry over contamination was examined in drug free horse urine samples injected subsequently to fortified samples at high concentration level of 20×LOD in order to evaluate the presence of any peak at the expected retention time. Regarding the evaluation of carry over

ip t

contamination, the peak area of the analyte detected at the blank sample was expressed as a

cr

percentage of the peak area of the same analyte detected at the previous sample.

2.6. Screening of 302 drugs in equine urine

us

The developed and validated method was further examined for the identification of a large number of banned compounds belonging to different therapeutic classes. The studied substances

an

presented in Supplementary Table 1 were fortified to blank equine urine at concentrations of 100 -1

or 200 ng mL . The fortified and the blank horse urine samples were prepared and analyzed

M

according to Supplementary Figure 1.

2.7. Method applicability

d

In order to verify that the validated screening method was effective to real positive

te

samples, ten positive post-race horse urine samples and five urine specimens from the proficiency test of Association of Official Racing Chemists (AORC) 2012 were analyzed. The samples had

Ac ce p

been previously analyzed with established and validated methods and they had been reported to contain amitriptyline, bumetanide, clanobutin, clenbuterol, dexamethasone, flunixin, furosemide, ketoprofen, mepivacaine, modafinil, morphine, nandrolone and its metabolite 5α-estrane-3β,17αdiol, phenylbutazone and its metabolite oxyphenbutazone, procaine, phenobarbital, nordiazepam and oxazepam. The fifteen re-examined samples were prepared and analyzed along with blank urine and quality control samples.

In addition to the above, two horse urine samples tested positive with pentobarbital were

also prepared and analyzed. The particular substance was not included among the target drugs of the screening method and therefore the applicability of AMDIS-DRS for automated evaluation of data files of real samples was examined, as well.

3. RESULTS AND DISCUSSION

11

Page 11 of 32

3.1. Method Development An extensive method development was performed in order to establish a suitable sample pretreatment taking into account the laboratory instrumentation. Thirty six compounds included in the AORC list of proficiency tests were used for the method development, in concentrations similar

ip t

to the ones depicted in the list [25]. The selected compounds represent multi-class drugs with various molecular weights, polarities, acidic-basic properties, chemical structures and biological

cr

effects. In brief, the method development was performed by testing different deconjugation

processes, different solvent extractions and mixtures of the final extracts. The examined

us

deconjugation processes were i) enzymatic hydrolysis with β-glucuronidase/arylsulfatase from H.pomatia ii) methanolysis and iii) basic hydrolysis with NaOH. The extraction solvents tested

an

were i) diethylether:ethylacetate:dichloromethane (4:3:3.5) ii) ethylacetate:dichloromethane (7:3) iii) ethylacetate and iv) diethylether. Moreover, the extraction pH tested were 9-10 and 5-5.2. The screening method was selected due to the greater number of compounds detected

M

and the greater mean extraction recovery obtained. The best results came from the sample pretreatment which comprised enzymatic hydrolysis and LLE with ethylacetate at pH 9-10. The

d

17β-sulfate steroid conjugates were cleaved by methanolysis after solid phase extraction and they

Ac ce p

Supplementary Figure 1.

te

were extracted with diethylether at pH 9.5-10. A more detailed description is found in

3.2. Method validation

The developed method was validated for sixty prohibited compounds, listed in Table 1,

which covered a large variety of drugs. The studied substances were chosen on the basis of AORC list of proficiency test 2011 [25] and FEI prohibited substances list [1] taking also into consideration the classification guidelines for foreign substances of the Association of Racing Commissioners International [26]. In addition to the parent compounds, metabolites of caffeine (theophylline),

cocaine

(benzoylecgonine),

stanozolol

(16β-hydroxystanozolol),

amitriptyline

(nortriptyline) and diazepam (oxazepam, nordiazepam) were identified and utilized as target compounds as well [27,28]. Moreover, terbutaline and ethacrynic acid artefacts caused by sample preparation were detected. Terbutaline incorporated a methylene unit in the presence of formaldehyde

under

acidic

conditions

during

12

enzymatic

hydrolysis,

forming

its

Page 12 of 32

tetrahydroisoquinoline analogue [29]. In respect to ethacrynic acid, its artefact was generated during derivatization process. Furthermore, the selected compounds for validation comprised five threshold substances approved by IFHA [2]. The threshold substances presented on Table 1, concern

endogenous

compounds

(5α-estrane-3β,17α-diol

in

male

horses,

testosterone,

ip t

hydrocortisone, 3-methoxytyramine) or compounds of dietary origin (theobromine) and therefore, they occur naturally in horse urine. These substances were examined at the established threshold

cr

concentration in the validation procedure.

The doping control method described was based on the combination of liquid and gas

us

chromatography that covers probably all small molecules for anti-doping interest. The method used two technologies, LC-TOF-MS and GC-HRMS, which enabled accurate mass determination

an

at high resolving power allowing the target analytes to be distinguished from the complex matrix. Besides, LC-TOF-MS and GC-MS techniques offered acquisition in full scan mode and thus the carrying out of retrospective analysis of the samples by reprocessing their initially acquired data

M

files. Regarding LC-TOF-MS analysis, the detection criteria included retention time of the monitoring ions (RT) and monoisotopic exact mass by applying mass tolerance ±20 ppm. All the +

d

studied compounds in method validation were ionized to their protonated molecules [M+H] in

te

positive ESI mode, except hydrochlorothiazide which was detected as ammonium adduct [M+NH4]+. Moreover, retention time and two diagnostic ions, the molecular and a fragment ion,

Ac ce p

were used for the identification of the methyl-derivatized substances. Furthermore, the detection of TMS-derivatized compounds through GC-HRMS analysis was based both on retention time and selected ion recording of one or two diagnostic ions at 10,000 resolution. The present method incorporated quality assurance measures covering the deconjugation

processes and the TMS-derivatization step. Etiocholanolone-glucuronide was added in urine at procedure A, prior to enzymatic hydrolysis, in order to evaluate the hydrolysis efficiency. Moreover, the addition of androsterone-sulphate in urine at procedure B allowed for the evaluation of methanolysis efficiency. Besides, by monitoring the mono-TMS derivatized etiocholanolone and androsterone ions at GC-HRMS analysis, the TMS-derivatization efficiency could be estimated.

3.2.1. Limit of Detection – Method sensitivity

13

Page 13 of 32

Each compound of interest was identified by means of the technique in which the lower detection limit was achieved. When the same detection limit was obtained by two techniques, LCTOF-MS was the technique of choice (eg salbutamol and clenbuterol). Regarding furosemide and phenylbutazone, methyl-derivatization was preferred instead of their unstable per-TMS derivatives,

ip t

while for ethacrynic acid, the artefact formed during TMS derivatization benefited GC-HRMS analysis against methyl-derivatization and GC-MS analysis.

cr

The estimated LOD values are given in Table 1. All fifty five target substances could be -1

-1

detected at 62.5 ng mL or lower, except furosemide that presented an LOD at 125 ng mL . For -1

us

thirty five analytes (64% of the examined compounds) the LOD ranged from 1.25 ng mL to 5.0 ng -1

mL , whereas fourteen analytes (25% of the studied compounds) showed LODs in the range 10-1

an

25 ng mL . The minimum detecting concentrations required by IFHA in Article 6 of the IABRW -1

-1

were met with the exception of benzoylecgonine (20 ng mL ), bumetanide (10 ng mL ) and -1

3.2.2. Matrix interference - Selectivity

M

dexamethasone (2 ng mL ) [2].

d

Ten negative post-race horse urine specimens were analyzed and demonstrated that

te

interferences from the different matrices at the expected retention times of the target ions were negligible, except flufenamic acid TMS derivative during GC-HRMS analysis. Thus, the particular

Ac ce p

analyte was detected by GC-MS assay after methyl-derivatization.

3.2.3. Extraction recovery

The extraction recovery data shown in Table 1 are the mean extraction recoveries

assessed on three experimental days (n=3) in six horse urine substrates per day. The mean calculated recovery values ranged from 13% (benzoylecgonine) to 92% (ractopamine), with overall average values (n=55 analytes) of 62%, 59% and 71% at the concentrations of 5×LOD, 10×LOD and 20×LOD, respectively. Regarding threshold substances, the obtained mean extraction recoveries varied between 26% (3-methoxytyramine) and 118% (testosterone). At least forty seven of the sixty compounds studied (78%) demonstrated mean recoveries above 50% at all three concentration levels. The hydrophilic and zwitterion benzoylecgonine (pKa1=2.3, pKa2=11.2) as well as the acidic compounds phenylbutazone (pKa=4.5) and ethacrynic acid (pKa=3.5) were

14

Page 14 of 32

extracted with recoveries around or below 20%. The poor extraction recovery of benzoylecgonine could be attributed to the degradation of its ester structure under basic conditions during sample preparation [30]. The calculated recovery values suggested recovery’s dependence upon analyte

ip t

concentration. Most of the examined compounds showed an increase of 10% in recovery at the highest concentration, 20×LOD, comparatively to the lower concentrations 5×LOD and 10×LOD.

cr

Besides, the polar compounds trifluoperazine, atenolol, salbutamol, terbutaline, ketoprofen and

benzoylecgonine demonstrated variation in extraction between days because their extraction is

us

highly affected by slight alterations in the pH during extraction. The particular validation parameter was examined on six different urine substrates and the results (not shown here) proved recovery’s

an

independence of the urine substrate, except for individual analytes such as fluphenazine, trifluoperazine, bumetanide, theophylline, which exhibited RSD>20% for at least two experimental days. Moreover, recovery’s RSD for substances analyzed by GC-HRMS and GC-MS was higher

M

than the RSD of those analyzed by LC-TOF-MS because of the additional step of derivatization

d

required for GC analysis.

te

3.2.4. Repeatability

Regarding the GC-MS analysis of the methylated compounds, the intra-day and inter-day

Ac ce p

precision of the RRTs were less than 0.08%. With respect to GC-HRMS analysis of the TMSderivatized substances, the intra-day precision of the RRTs ranged from 0.03 to 0.16% whereas the respective inter-day precision ranged between 0.03 and 0.36%. The RRTs for the selected compounds analyzed by LC-TOF-MS were also stable within day, as their RSD values did not exceed 1%. Moreover, inter-day precision lower than 1% was observed for forty one of the analytes detected by LC-TOF-MS, exclusive of salbutamol, terbutaline, hydrochlorothiazide, amphetamine,

atenolol,

ephedrine,

morphine,

clonidine,

ergonovine,

3-methoxytyramine,

theobromine at overall concentration levels. The variation of RRTs for the latter analytes can be attributed to their early elution in the beginning of the gradient or to their peak shape eg asymmetrical and wide peaks. The results of the precision (of peak areas stated in terms of RSD) are summarized in Table 1. The intra-day precision for the compounds analyzed by LC-TOF-MS was in the range of

15

Page 15 of 32

0.9% (timolol) to 23% (trifluoperazine), with mean RSD values of

7.8, 4.6 and 2.7% at the

concentrations of 5×LOD, 10×LOD, 20×LOD, respectively. The inter-day precision varied from 0.9% (timolol) to 41% (phentermine), with mean RSD values of 12, 11 and 14% at the concentrations of 5×LOD, 10×LOD, 20×LOD, respectively. The poorest within-days repeatability

ip t

was obtained for phentermine because of the double chromatographic peak observed in the first and the second experimental day. For at least 81% of the compounds analyzed by LC-TOF-MS

cr

the inter-day precision was lower than 20% at the examined concentrations.

For GC-MS analysis, the intra-day precision of the methylated substances ranged

us

between 5.0% (ketoprofen) and 30% (diflunisal) and the inter-day precision was around 24-61%. The highest variation was observed in diflunisal due to the formation of mono- and bis-methylated

an

diflunisal during the derivatization step.

In respect of GC-HRMS analysis, the intra-day precision of the trimethylsilylated target compounds was less than 25% at the examined concentration levels except for the ethacrynic acid

M

artefact due to its non-constant formation during the sample preparation procedure. The inter-day precision ranged from 13% (theophylline) to 58% (ethacrynic acid artefact).

d

Overall, the precision results derived from LC-TOF-MS analysis were better than those

Ac ce p

3.2.5. Mass accuracy

te

obtained from GC-MS and GC-HRMS analyses because of the absence of the derivatization step.

Mass accuracy was evaluated at a medium concentration level of 10×LOD, at which most 4

5

of the ion abundances were in the range 1×10 - 5×10 . Bristow et al. [31] observed a clear 3

6

deterioration of mass accuracy at both very low (1×10 ) ion abundances. The mass error values presented in Table 1 are the mean mass errors calculated on three days in a period of four months, using lock mass correction. The mean mass measurement errors ranged from -2.8 ppm (diphenhydramine) to 5.0 ppm (3-methoxytyramine) with an overall average value (n=60 substances) of 1.0 ppm. 3-Methoxytyramine is a small molecule (MW=167) eluting in the beginning of the gradient (RT=1.43 min) along with many matrix compounds, resulting often in broad and less symmetrical peak shapes which shifted the centroid mass to a higher m/z value.

3.2.6. Matrix Effect

16

Page 16 of 32

The matrix effect results reported on Table 1 are the mean values from the three experimental days with the corresponding relative standard deviations. A matrix effect value of 100% indicates that the response in the mobile phase and the urine extract are the same and thus no matrix effect is observed. A matrix effect value greater than 100% indicates ionization

ip t

enhancement and a value smaller than 100% indicates ionization suppression. Signal suppression, which is more common than signal enhancement in LC-ESI-MS, was the effect

cr

observed for almost all the analytes. The highest ion suppression, about 20%, occurred in phentermine (MW=149), whereas the highest ion enhancement, about 160%, was caused for

us

oxazepam at all three concentration levels. Despite the high ion suppression observed, -1

phentermine could still be detected at 2.5 ng mL . It is known that smaller and more polar analytes

an

are more susceptible to undergo ion suppression [32-33]. Regarding the antipsychotic compounds fluphenazine and trifluoperazine, the kind of matrix effect was not constant since both signal suppression and signal enhancement were observed at different concentrations.

M

The relative standard deviations of the matrix effect values were also calculated in order to evaluate the variation between six different horse urine matrices (data not shown). The particular

d

RSD values demonstrated the dependence of matrix effect upon substrate and especially for

te

triamcinolone acetonide, amphetamine, chlorpromazine, flunixin, morphine, celecoxib, trenbolone, fluphenazine, trifluoperazine, nortriptyline, phentermine, ergonovine and 16β-hydroxystanozolol

Ac ce p

the matrix effect varied considerably between different urine samples. Furthermore, the particular validation parameter was examined in three concentration levels, by keeping constant the matrix amount. For some substances, such as nortriptyline, hydrochlorothiazide, nordiazepam, the extent of signal suppression was inversely proportional to their concentration; the higher the concentration of the target analyte, the lower the level of suppression. However, there were substances, such as verapamil, dexamethasone, fentanyl, in which the matrix effect was constant and independent of the concentration of the analyte.

3.2.7. Carry over contamination Carry over contamination was found for furosemide (2%) and bumetanide (1%) which were identified by GC-MS and GC-HRMS respectively. Both analytes detected at the blank sample

17

Page 17 of 32

originated from the former fortified sample at concentration of 20×LOD. No carry over was detected for the rest of the compounds.

3.3. Identification of 302 drugs in equine urine

ip t

The hereby proposed method aimed at the determination of a wide range of prohibited compounds in equine urine. Supplementary Table 1 shows the 302 analytes related to equine

cr

doping control that were covered by the current screening method. The majority of the examined compounds could be detected through the LC-TOF-MS analysis in full scan acquisition mode with

us

high mass accuracy. Some NSAIDs and carbonic anhydrase inhibitors were detected through methyl-derivatization and GC-MS analysis, while anabolic steroids were identified by LC-TOF-MS

an

[34-40] and GC-HRMS assays. The studied analytes comprised parent compounds, phase 1 metabolites, degradation products and artefacts [41].

M

3.4. Method applicability

Seventeen previously reported positive horse urine samples were re-examined by

d

applying the current screening method. Figures

1, 2 and 3 present the extracted-ion-

te

chromatograms (EIC) of the target analytes in the blank urine, the quality control sample and the positive samples after LC-TOF-MS, GC-MS and GC-HRMS analysis, respectively. The EIC of the

Ac ce p

blank urine did not show any peak at the expected retention time. All previously reported prohibited substances were detected by the present method with good S/N ratios and with retention times which did not vary from that in the reference sample by more than 1.6% (morphine, Figure 1). There were compounds that could be identified by more than one analysis (chromatograms not shown), such as bumetanide which was detected by all the three analyses (LC-TOF-MS, GC-MS and GC-HRMS). Furosemide and clenbuterol were also detected after TMSderivatization and GC-HRMS analysis, while phenylbutazone and oxyphenbutazone could be identified by LC-TOF-MS besides GC-MS analysis of their methylated derivatives. Regarding the positive samples in nandrolone and phenylbutazone, the major urinary metabolites 5α-estrane3β,17α-diol and oxyphenbutazone were clearly detected in addition to the corresponding parent analytes. Nandrolone and its metabolite 5α-estrane-3β,17α-diol were found in urine from female

18

Page 18 of 32

horse and therefore, doping control was not based on threshold concentration. Moreover, the detection of oxazepam and nordiazepam indicated administration of diazepam. Furthermore, the applicability of the library searching using AMDIS-DRS software was demonstrated by the detection of pentobarbital in two urine samples after GC-MS analysis of the

ip t

methyl-derivatized samples. Pentobarbital was not detected by the targeted screening method, since it was not included in the examined analytes. The application of library searching using

cr

AMDIS-DRS software opens the window of the detectable compounds.

us

4. CONCLUSIONS

A unified, sensitive and suitable method was developed for the qualitative identification of

an

a wide range of prohibited substances in equine urine for doping control purposes. The phase 2 metabolites were cleaved through enzymatic hydrolysis and methanolysis, while the solid phase and liquid-liquid extractions were implemented for the extraction of the compounds of interest from

M

the complex matrix. The analysis of the prepared urine samples was performed in a complementary way by LC-TOF-MS, GC-HRMS and GC-MS techniques. Thus the substances

d

with difficulty in positive ESI ionization could be detected by GC, whereas the thermo labile, the

te

nonvolatile molecules and those with marginal GC properties even after derivatization, could be identified by LC. The method was validated for sixty selected compounds and in addition it

Ac ce p

facilitated a broad screening of 302 additional banned substances. Moreover, the validated method was supported by automated identification of analytes through library searching by using the common AMDIS-DRS program, which enabled the detection of banned drugs not targeted by the screening method. Both approaches, targeted and non-targeted doping control, were successfully applied to previously reported positive authentic horse urine samples. Overall, this is a generic and robust racing animal protocol that provides MS library

searching, retrospective evaluation of analyzed samples and easy incorporation of new released drugs.

ACKNOWLEDGEMENTS

19

Page 19 of 32

The authors wish to thank Agilent Technologies for providing AMDIS-DRS programs. S.Loui, F.Hlapana and I.Koutsouli are gratefully acknowledged for their technical assistance as well as

Ac ce p

te

d

M

an

us

cr

ip t

P.Kiousi and R.Fragkaki for their support in LC-TOF-MS analysis.

REFERENCES

20

Page 20 of 32

[1] Federation Equestre Internationale (FEI), 2013 Equine Prohibited Substances List. Available at:

http://www.feicleansport.org/2013%20Equine%20Prohibited%20List.pdf

(accessed

on

June 2013).

Racing

and

Wagering,

March

ip t

[2] International Federation of Horseracing Authorities. International Agreement on Breeding, 2012.

Available

at:

cr

http://www.horseracingintfed.com/resources/2012_choose_eng.pdf (accessed on June 2013).

[3] M. Thevis, W. Schänzer, Mass Spectrometry in sports drug testing: structure characterization

us

and analytical assays, Mass Spectrom. Rev. 26 (2007) 79.

[4] O.J. Pozo, J. Marcos, J. Segura, R. Ventura, Recent developments in MS for small molecules:

an

application to human doping control analysis, Bioanalysis 4(2) (2012) 197.

[5] I. Ojanperä, M. Kolmonen, A. Pelander, Current use of high-resolution mass spectrometry in drug screening relevant to clinical and forensic toxicology and doping control, Anal. Bioanal.

M

Chem. 403 (2012) 1203.

[6] M. Yamada, S. Aramaki, M. Kurosawa, I. Kijima-Suda, K. Saito, H. Nakazawa, Simultaneous

d

doping analysis of main urinary metabolites of anabolic steroids in horse by ion-trap gas

te

chromatography – tandem mass spectrometry, Anal. Sci. 24 (2008) 1199. [7] M.E. Touber, M.C. van Engelen, C. Georgakopoulos, J.A. van Rhijn, M.W.F. Nielen, Multi-

Ac ce p

detection of corticosteroids in sports doping and veterinary control using high-resolution liquid chromatography/time of flight mass spectrometry, Anal. Chim. Acta 586 (2007) 137.

[8] C.H.F. Wong, D.K.K. Leung, F.P.W. Tang, J.K.Y. Wong, N.H. Yu, T.S.M. Wan, Rapid screening

of

anabolic

steroids

in

horse

urine

with

ultra-high-performance

liquid

chromatography tandem mass spectrometry after chemical derivatisation, J. Chromatogr. A 1232 (2012) 257.

[9] Y. Liu, C.E. Uboh, L.R. Soma, X. Li, F. Guan, Y. You, J.-W. Chen, Efficient use of retention time for the analysis of 302 drugs in equine plasma by liquid chromatography-MS/MS with scheduled multiple reaction monitoring and instant library searching for doping control, Anal. Chem. 83 (2011) 6834. [10] P. Garcia, A.-C. Paris, J. Gil, M.-A. Popot, Y. Bonnaire, Analysis of β-agonists by HPLC/ESIMSn in horse doping control, Biomed. Chromatogr. 25 (2011) 147.

21

Page 21 of 32

[11] R.J.B. Peters, A.A.M. Stolker, J.G.J. Mol, A. Lommen, E. Lyris, Y. Angelis, A. Vonaparti, M. Stamou, C.G. Georgakopoulos, M.W.F. Nielen, Screening in veterinary drug analysis and sports doping control based on full-scan, accurate-mass specrtrometry, Trends Anal. Chem. 29(11) (2010) 1250.

ip t

[12] A. Kaufmann, P. Butcher. K. Maden, M. Widmer, Ultra-performance liquid chromatography coupled to time of flight mass spectrometry (UPLC-TOF): A novel tool for multiresidue

cr

screening of veterinary drugs in urine, Anal. Chim. Acta 586 (2007) 13.

[13] R.G. Howitt, M. Pelzing and G.D. Beresford, Routine drug screening by accurate mass using

th

Wada and J.F. Wade (Eds.), Proceedings of the 16

us

Liquid Chromatography/Time of Flight Mass Spectrometry, in: E. Houghton, I. Kijima-Suda, R. International Conference of Racing

an

Analysts and Veterinarians, R & W Publications, Newmarket, 2006, pp. 222.

[14] R.G. Howitt, G.D. Beresford, M. Pelzing and I. Krebs, Routine drug screening by accurate mass using Liquid Chromatography/Time of Flight Mass Spectrometry-Part 2, in: E. Houghton,

M

F. Ipek Keskin, J.F.Wade and N. Yazicioglu (Eds.),

Proceedings of the 17

th

International

Conference of Racing Analysts and Veterinarians, R & W Publications, Newmarket, 2008, pp.

d

191.

te

[15] M.R. Meyer, F.T. Peters, H.H. Maurer, Automated Mass spectral Deconvolution and Identification System for GC-MS screening for drugs, poisons and metabolites in urine, Clin.

Ac ce p

Chem. 56(4) (2010) 575.

[16] M.R. Meyer, P. Du, F. Schuster, H.H. Maurer, Studies on the metabolism of the αpyrrolidinophenone designer drug methylenedioxy-pyrovalerone (MDPV) in rat and human urine and human liver microsomes using GC-MS and LC-high-resolution-MS and its detect ability in urine by GC-MS, J. Mass Spectrom. 45(12) (2010) 1426.

[17] G.N.W. Leung, E.W. Chung, E.N.M. Ho, W.H. Kwok, D.K.K. Leung, F.P.W. Tang, T.S.M. Wan, N.H Yu, High-throughput screening of corticosteroids and basic drugs in horse urine by liquid chromatography-tandem mass spectrometry, J. Chromatogr. B 825 (2005) 47. [18] E.N.M. Ho, D.K.K. Leung, T.S.M. Wan, N.H. Yu, Comprehensive screening of anabolic steroids, corticosteroids and acidic drugs in horse urine by solid-phase extraction and liquid chromatography-mass spectrometry, J. Chromatogr. A 1120 (2006) 38.

22

Page 22 of 32

[19] C.H.F. Wong, F.P.W. Tang, T.S.M. Wan, A broad-spectrum equine urine screening method for free and enzyme-hydrolyzed conjugated drugs with ultra performance liquid chromatography tandem mass spectrometry, Anal. Chim. Acta 697 (2011) 48. [20] Y. Moulard, L. Bailly-Chouriberry, S. Boyer, P. Garcia, M.-A. Popot, Y. Bonnaire, Use of

ip t

benchtop exactive high resolution and high mass accuracy orbitrap mass spectrometer for screening in horse doping control, Anal. Chim. Acta, 700 (2011) 126.

cr

[21] E.N.M. Ho, W.H. Kwok, A.S.Y. Wong, T.S.M. Wan, High resolution accurate mass screening

of prohibited substances in equine plasma using liquid chromatography-orbitrap mass

us

spectrometry, Drug Test. Analysis (2012) 5 (2013) 509.

[22] NIST, available at: http://www.sisweb.com/software/ms/nist.htm (accessed on June 2013).

an

[23] Eurachem Guide, The Fitness for Purpose of Analytical Methods: A Laboratory Guide to Method Validation and Related Topics, Eurachem Teddington, 1998. Available at: http://www.eurachem.org/images/stories/Guides/pdf/valid.pdf (accessed on June 2013).

M

[24] B.K. Matuszewski, M.L. Constanzer, C.M. Chavez, Strategies for the assessment of matrix effect in quantitative bioanalytical methods based on HPLC–MS/MS. Anal. Chem. 75 (2003)

Proficiency

Testing

Drug

List

2011.

Available

at:

http://www.aorc-

te

[25] AORC

d

3019.

online.org/documents/aorc-proficiency-testing-drug-list-2011 (accessed on June 2013)

Ac ce p

[26] Association of Racing Commissioners International, Inc. (ARCI). Uniform Classification Guidelines for Foreign Substances and Recommended Penalties and Model Rule, Available at: http://www.arci.com/druglisting.pdf (accessed on June 2013)

[27] J.P. Scarth, P. Teale, T. Kuuranne, Drug metabolism in the horse: a review, Drug Test. Analysis, 3(1) (2011) 19.

[28] A.R. McKinney, C.J. Suann, A.J. Dunstan, S.L. Mulley, D.D. Ridley, A.M Stenhouse, Detection of stanozolol and its metabolites in equine urine by liquid chromatography-electrospray ionization ion trap mass spectrometry, J. Chromatogr. B 811 (2004) 75. [29] M. Thevis, G. Opfermann, W. Schanzer, Liquid chromatography/electrospray ionization tandem mass spectrometric screening and confirmation methods for β2-agonists in human or equine urine, J. Mass Spectrom. 38 (2003) 1197.

23

Page 23 of 32

[30] M. Kolmonen, A. Leinonen, T. Kuuranne, A. Pelander, I. Ojanperä, Generic sample preparation and dual polarity liquid chromatography-time-of-flight mass spectrometry for highthrouput screening in doping analysis, Drug Test. Analysis 1 (2009) 250. [31] T. Bristow, J. Constantine, M. Harrison, F. Cavoit, Performance optimisation of a neworthogonal-acceleration quadrupole-time-of-flight mass spectrometer, Rapid

ip t

generation

Commun. Mass Spectrom. 22 (2008) 1213.

cr

[32] R. Bonfiglio, R.C. King, T.V. Olah, K. Merkle. The effect sample preparation methods on the

variability of the electrospray ionization response for model drug compounds. Rapid Commun.

us

Mass Spectrom. 13 (1999) 1175.

[33] A. Kloepfer, J. Quintana, T. Reemtsma, Operational options to reduce matrix effects in liquid

an

chromatography - electrospray ionization - mass spectrometry analysis of aqueous environmental samples. J. Chromatogr. A 1067 (2005) 153.

[34] H.-W. Hagedorn, R. Schulz, A. Friedrich, Detection of methandienone (methandrostenolone)

Biomed. Applic. 577 (1992) 195.

M

and metabolites in horse urine by gas chromatography-mass spectrometry, J. Chromatogr.

d

[35] M. Yamada, S. Aramaki, M. Kurosawa, K. Saito, H. Nakazawa, Detection of urinary

te

metabolites common to structurally related 17α-alkyl anabolic steroids in horses and application to doping tests in racehorses: methandienone, methandriol and oxymetholone, J.

Ac ce p

Anal. Toxicol. 32 (2008) 387.

[36] M. Yamada, S. Aramaki, T. Hosoe, M. Kurosawa, I. Kijima-Suda, K. Saito, H. Nakazawa, Characterization and quantification of fluoxymesterone metabolite in horse urine by Gas Chromatography / Mass Spectrometry, Anal. Sci. 24 (2008) 911.

[37] J. Fox, C. Dumasia, S. Stanley, E. Houghton, The in vivo biotransformation of normethandrolone in the horse, in: R.B. Williams, E. Houghton, J.F. Wade (Eds.), Proceedings th

of the 13 International Conference of Racing Analysts and Veterinarians, R & W Publications, Newmarket, 2000, pp. 479.

[38] E. Houghton, S. Maynard, Some aspects of doping and medication control in equine Sports, in: D. Thieme, P. Hemmersbach (Eds.), Doping in Sports: Biochemical principles, affects and analysis, Handbook of experimental Pharmacology, Springer, Heidelberg, 2010, vol. 195, pp. 369.

24

Page 24 of 32

[39] G.N.M. Leung, F.P.W. Tang, T.S.M. Wan, C.H.F..Wong, K.K.H. Lam, B.D. Stewart, In vitro and in vivo studies of androst-4-ene-3,6,17-trione in horses by gas chromatography-mass spectrometry, Biomed. Chromatogr. 24 (2010) 744. [40] M. Machnik, M. Thevis, C. Von Kuk, S. Guddat, W. Schänzer, Gestrinone analysis in equine

ip t

urine by LC-MS/MS, in W. Schänzer, H. Geyer, A. Gotzmann, U. Mareck (Eds.), Recent Advances in Doping Analysis (12) Sport and Buch Strauβ, Cologne, 2004b, pp 149.

cr

[41] K. Deventer, O.J. Pozo, P. Van Eenoo, F.T. Delbeke, Detection of urinary markers for

thiazides diuretics after oral administration of hydrochlorothiazide and altizide-relevance to

Ac ce p

te

d

M

an

us

doping control analysis, J. Chromatogr. A 1216 (2009) 2466.

25

Page 25 of 32

ip t cr

Molecular Formula

LC-TOF-MS analysis 1-(3-Chlorophenyl) Psychoactive C10H13N2Cl piperazine Anabolic steroid C21H32N2O2 2 16β-ΟΗ-stanozolol C9H13NO2 3 3-methoxytyramine# Neurotransmitter Tricyclic 4 Amitriptyline* C20H23N antidepressant 5 Amphetamine* Stimulant C9H13N 6 Anastrozole Aromatase inhibitor C17H19N5 7 Atenolol* Beta blocker C14H22N2O3 8 Benzoylecgonine* Stimulant C16H19NO4 9 Butorphanol* Opiod C21H29NO2 Cardiac-respiratory 10 Caffeine* C8H10N4O2 stimulant; diuretic 11 Capsaicin Topical analgesic C18H27NO3 12 Celecoxib NSAID C17H14F3N3O2S 13 Chlorpromazine* Sedative C17H19ClN2S 14 Clenbuterol* Bronchodilator C12H18Cl2N2O 15 Clonidine Antihypertensive C9H9Cl2N3 16 Dexamethasone* Corticosteroid C22H29FO5 17 Diphenhydramine Antihistamine C17H21NO 18 Ephedrine* Stimulant C10H15NO 19 Ergonovine Vasoconstrictor C19H23N3O2 20 Fentanyl Opiod analgesic C22H28N2O 21 Flunixin* NSAID C14H11F3N2O2 22 Fluphenazine Antipsychotic C22H26F3N3OS 23 Hydrochlorothiazide* Diuretic C7H8ClN3O4S2 # Corticosteroid C21H30O5 24 Hydrocortisone 25 Imipramine Antidepressant C19H24N2 26 Isoxsuprine Vasodilator C18H23NO3 27 Ketoprofen* NSAID C16H14O3 28 Lidocaine* Local anesthetic C14H22N2O 29 Meloxicam* NSAID C14H13N3O4S2 30 Mepivacaine* Local anesthetic C15H22N2O 31 Methocarbamol* Muscle relaxant C11H15NO5 32 Methylphenidate* Stimulant C14H19NO2 33 Morphine* Opiod analgesic C17H19NO3

[M+H]+ isotope [M+H]+ [M+H]+

a

Target ion

RT LOD -1 (min) (ng mL )

199.0812

4.90

7.41 1.43

345.2537 168.1019

Ac

% ER20 b (RSD)

Precision, % RSD intraday 5c intraday 10c intraday 20c (interday5)b (interday 10)b (interday 20)b

% ER 5 b (RSD)

% ER10 b (RSD)

62.5

72 (13)

74 (10.6)

82 (9.6)

8.9 (13)

2.6 (12)

12.5 4000

64 (8.8) 26 (19)

57 (15)

70 (8.6)

14 (11) 4.6 (1.9)

6.3 (6.8)

7.48

1.25

57 (6.9)

54 (8.8)

64 (9.9)

8.0 (5.2)

4.6 (3.4)

3.55 7.37 2.05 4.08 5.58

1.25 2.5 2.5 25 1.25

55 (6.1) 77 (0.9) 38 (38) 13 (53) 68 (6.7)

58 (3.0) 71 (7.7) 35 (41) 14 (59) 64 (5.7)

74 (2.9) 86 (6.2) 44 (38) 18 (57) 79 (4.4)

5.6 (20) 6.2 (9.4) 8.7 (8.6) 5.2 (24) 6.2 (8.3)

8.0 (19) 3.9 (17) 1.6 (10) 2.5 (16) 3.8 (3.4)

Mass Error10, ppm

% ME5 b (RSD)

% ME10 b (RSD)

% ME20 b (RSD)

5.3 (13)

0.22

51 (4.9)

58 (7.6)

60 (9.0)

3.5 (16)

1.1 5.0

21 (30) 76 (9.1)

25 (32)

28 (19)

2.6 (8.9)

2.7

55 (4.9)

57 (8.5)

58 (3.0)

6.7 (23) 1.9 (21) 2.8 (15) 4.2 (12) 2.2 (9.2)

1.6 0.94 0.25 -0.5 1.2 ()

47 (30) 39 (26) 84 (4.3) 35 (11) 69 (1.6)

69 (36) 42 (27) 75 (11) 39 (19) 68 (9.7)

52 (35) 46 (27) 79 (5.9) 46 (19) 71 (8.0)

[M+H]+

278.1903

[M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+

136.1121 294.1713 267.1703 290.1387 328.2271

[M+H]+

195.0877

3.39

62.5

81 (13)

68 (2.6)

84 (1.3)

5.5 (9.0)

2.3 (6.1)

1.5 (11)

2.2

44 (24)

46 (25)

51 (17)

[M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+NH4]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+

306.2064 382.0832 319.1030 277.0869 230.0246 393.2072 256.1696 166.1226 326.1863 337.2274 297.0845 438.1821 314.9978 363.2166 281.2012 302.1751 255.1016 235.1805 352.0420 247.1805 242.1023 234.1489 286.1438

9.46 10.37 7.81 4.82 3.26 7.19 6.48 3.00 3.65 6.39 8.27 8.27 3.51 6.43 7.27 5.40 8.20 4.31 7.51 4.28 5.06 4.87 1.54

1.25 12.5 5.0 2.5 1.25 12.5 1.25 25 12.5 1.25 5.0 2.5 12.5 1000 1.25 1.25 25 1.25 1.25 1.25 5.0 1.25 2.5

61 (4.6) 51 (6.7) 48 (9.2) 73 (4.4) 79 (9.6) 74 (5.9) 69 (5.8) 57 (8.7) 73 (7.4) 65 (0.8) 74 (5.1) 53 (9.9) 68 (6.9) 75 (5.7) 62 (9.6) 77 (2.1) 50 (25) 72 (9.4) 63 (13) 76 (6.8) 78 (7.4) 63 (5.8) 88 (0.6)

58 (6.4) 48 (17) 47 (9.4) 71 (4.0) 71 (3.1) 70 (8.0) 66 (11) 60 (15) 67 (2.7) 57 (4.4) 71 (4.3) 45 (21) 68 (9.3)

72 (3.4) 65 (15) 59 (8.5) 89 (4.3) 83 (0.7) 87 (6.5) 73 (9.0) 69 (3.7) 76 (7.6) 68 (0.6) 90 (2.7) 51 (4.7) 83 (7.0)

6.8 (7.2) 7.5 (10) 6.9 (8.7) 2.4 (6.5) 2.9 (10) 2.8 (15) 5.6 (5.8) 4.1 (8.9) 3.0 (17) 5.7 (4.2) 2.3 (2.3) 13 (8.6) 3.7 (5.8)

2.9 (12) 3.1 (15) 2.6 (18) 1.6 (11) 1.9 (16) 1.1 (15) 2.6 (12) 4.0 (12) 1.2 (14) 2.3 (11) 1.9 (7.7) 2.7 (14) 3.9 (10)

5.1 (13) 3.0 (9.3) 2.8 (8.1) 3.9 (3.8) 5.2 (1.5) 1.8 (16) 2.5 (4.4) 2.6 (7.8) 5.7 (6.5)

1.4 (16) 1.1 (10) 1.7 (18) 2.7 (4.1) 2.2 (8.7) 0.9 (7.1) 1.3 (7.5) 2.5 (13) 3.9 (11)

53 (9.8) 45 (18) 55 (6.5) 61 (4.5) 69 (7.7) 47 (13) 52 (20) 78 (6.4) 61 (17) 65 (6.7) 35 (6.2) 55 (14) 61 (4.6) 73 (16) 48 (9.7) 63 (9.4) 60 (8.8) 62 (3.1) 42 (12) 64 (3.2) 55 (5.1) 63 (5.4) 50 (17)

51 (9.3) 38 (21) 136 (39) 68 (11) 72 (1.7) 47 (13) 53 (19) 87 (8.4) 68 (18) 63 (6.1) 40 (15) 108 (44) 76 (14)

71 (11) 83 (4.3) 63 (30) 81 (0.1) 80 (13) 89 (4.2) 92 (1.8) 73 (9.1) 79 (11)

1.2 0.07 2.4 -0.38 1.5 3.0 -2.8 2.5 0.53 1.2 2.9 1.6 1.4 0.4 1.0 0.94 2.0 -2.4 0.63 0.0 0.50 -1.2 1.5

50 (3.0) 42 (13) 82 (9.2) 61 (14) 71 (7.8) 45 (12) 58 (24) 86 (6.0) 60 (17) 63 (5.8) 37 (16) 100 (38) 68 (7.2)

59 (12) 70 (4.7) 52 (31) 68 (4.6) 61 (18) 71 (3.9) 73 (3.9) 62 (9.5) 64 (9.4)

9.7 (3.0) 12 (19) 10 (12) 5.3 (6.0) 4.4 (5.3) 8.4 (18) 5.3 (8.7) 8.6 (4.4) 3.5 (18) 9.1 (7.1) 5.6 (3.4) 14 (21) 9.6 (5.0) 8.4 (3.7) 8.8 (6.8) 6.0 (4.9) 5.6 (8.0) 6.3 (11) 8.3 (14) 5.6 (23) 8.6 (16) 4.7 (5.7) 8.0 (3.0)

52 (10) 58 (14) 64 (12) 72 (12) 45 (7.6) 68 (19) 58 (13) 76 (12) 35 (34)

53 (3.6) 61 (12) 69 (15) 66 (6.2) 49 (15) 68 (12) 63 (11) 67 (8.9) 35 (26)

ce pt

1

Adduct / derivative

M an

Therapeutic Activity

ed

No. Substance

us

Table 1. Results of method validation for retention time (RT), limit of detection (LOD), extraction recovery (ER), repeatability on peak areas (intra-day and inter-day precision), mass error and matrix effect (ME) with the corresponding mean values and the percentage relative standard deviations (RSD) in parenthesis. The subscripts 5, 10 and 20 correspond to concentrations of 5×LOD, 10×LOD and 20×LOD, respectively.

26

Page 26 of 32

Corticosteroid

C24H31FO6

[M+H]+

435.2177

Antipsychotic Anti-arrhythmic

C21H24F3N3S C27H38N2O4

[M+H]+ [M+H]+

408.1716 455.2904

Analgesic NSAID Diuretic NSAID Sedative NSAID

56 5α-estrane-3β,17α-

Anabolic steroid

57 Bumetanide* 58 Ethacrynic acid artefact* 59 Testosterone# 60 Theophylline

Diuretic Diuretic

diol*,#

a

C13H8F2O3 bis-Me C14H10F3NO2 mono-Me tris-Me C12H11ClN2O5S mono-Me C16H14O3 bis-Me C12H12N2O3 mono-Me C19H20N2O2 C18H30O2

C17H20N2O5S C13H12Cl2O4

Anabolic steroid Bronchodilator

4.43 8.22 7.32 5.85 7.33 5.30 4.01 6.22 4.25 1.86 1.87 1.75 4.56 7.63

2.5 1.25 12.5 12.5 2.5 1.25 2.5 1.25 5 25 10 2000 1.25 12.5

68 (9.0) 64 (5.8) 62 (16) 81 (0.6) 71 (7.2) 72 (5.6) 74 (4.3) 69 (9.0) 86 (3.0) 37 (28) 41 (27) 46 (13) 77 (6.1) 71 (8.2)

61 (2.1) 63 (6.7) 60 (19) 77 (2.9) 72 (7.2) 69 (7.9) 75 (6.4) 66 (9.1) 79 (3.1) 43 (31) 54 (31)

82 (2.1) 80 (6.5) 68 (12) 88 (2.7) 89 (5.7) 80 (8.7) 84 (4.0) 78 (11) 92 (3.2) 54 (29) 59 (29)

247; 278 263; 295 81; 372 209; 268 232; 117 322; 266

C19H28O2 C7H8N4O2

bis-TMS

6.8 (8.6) 4.1 (6.6) 3.9 (23) 2.7 (9.9) 2.8 (27) 2.8 (9.6) 3.2 (36) 4.5 (11) 2.4 (19) 2.5 (7.2) 4.2 (25)

7.2 (2.7) 1.4 (9.8) 2.3 (24) 1.0 (11) 2.2 (34) 2.5 (15) 8.4 (25) 1.2 (14) 1.5 (9.7) 2.1 (7.6) 5.7 (22)

89 (2.9) 82 (3.3)

13 (15) 5.6 (10) 7.3 (23) 5.5 (5.6) 6.8 (20) 5.3 (10) 9.6 (41) 6.8 (6.2) 2.6 (28) 7.5 (6.4) 7.1 (21) 5.7 (8.7) 4.5 (2.7) 13 (5.6)

72 (2.9) 69 (1.1)

2.1 (0.9) 8.1 (16)

7.57

5.0

70 (8.0)

8.72 7.46

5.0 1.25

46 (30) 54 (3.9)

66 (5.3)

83 (3.9)

8.9 (9.9)

37 (26) 51 (3.9)

40 (12) 61 (5.4)

23 (35) 6.1 (8.9)

5.0 5.0 125 25 5.0 62.5

68 (11) 64 (5.7) 39 (15) 48 (28) 64 (14) 22 (7.0)

61 (11) 64 (14) 36 (34) 51 (37) 68 (15) 20 (12)

75 (19) 79 (12) 44 (16) 59 (34) 82 (21) 27 (14)

12 (61) 10 (26) 14 (39) 8.9 (27) 10 (32) 9.0 (27)

45

103 (6.3)

25 50

37 (6.4) 21 (20)

34 (18) 20 (45)

35 (15) 19 (9.7)

20 50

118 (11) 40 (17)

40 (28)

50 (29)

us

179.1179 271.0633 264.1747 300.1958 287.0582 248.1645 150.1277 286.1914 302.1751 240.1594 226.1438 181.0720 317.1642 271.1693

cr

ip t 51 52 53 27 54 55

[M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+ [M+H]+

M an

49 50

C10H14N2O C15H11ClN2O C19H21N C19H25NO2 C15H11ClN2O2 C15H21NO2 C10H15N C17H23N3O C18H24NO3 C13H21NO3 C12H19NO3 C7H8N4O2 C13H24N4O3S C18H22O2

Stimulant Tranquiliser Antidepressant Sympathomimetic Anxiolytic Opiod analgesic Anticonvulsant Antihistamine Beta agonist Bronchodilator Bronchodilator Vasodilator Beta blocker Anabolic steroid

ed

48

Nikethamide* Nordiazepam* Nortriptyline Nylidrin Oxazepam* Pethidine* Phentermine Pyrilamine Ractopamine Salbutamol* Terbutaline* Theobromine# Timolol* Trenbolone Triamcinolone acetonide* Trifluoperazine Verapamil GC-MS analysis Diflunisal Flufenamic acid* Furosemide* Ketoprofen Phenobarbital* Phenylbutazone* GC-HRMS analysis

ce pt

34 35 36 37 38 39 40 41 42 43 44 45 46 47

6.89 6.21 11.66 7.74 6.02 8.82

332.2536; 7.69 407.2802 bis-TMS 508.1884 10.52 mono-TMS 315.0375; 8.46 317.0375 bis-TMS 432.2880 8.76 mono-TMS 237.0808 3.55

0.9 (5.2) 4.2 (19)

1.2 0.64 0.23 0.39 0.79 0.16 -0.52 1.1 0.17 2.2 1.2 2.5 0.56 -0.10

49 (17) 53 (15) 43 (7.4) 72 (8.9) 159 (13) 63 (0.2) 18 (12) 89 (12) 38 (17) 85 (5.7) 77 (6.5) 58 (7.0) 62 (4.7) 42 (7.8)

77 (50) 62 (12) 48 (12) 76 (10) 166(12) 72 (14) 24 (16) 90 (12) 43 (18) 87 (10) 94 (14)

57 (13) 69 (16) 52 (4.5) 78 (7.3) 165(14) 60 (10) 19 (7.7) 83 (2.0) 46 (20) 97 (7.0) 111 (12)

61 (14) 42 (5.3)

66 (12) 48 (5.5)

3.1 (15)

1.5 (19)

1.7

43 (13)

40 (16)

42 (12)

22 (22) 5.2 (5.8)

5.3 (27) 2.6 (7.8)

1.3 1.5

56 (31) 59 (13)

115 (43) 58 (9.3)

107 (57) 59 (6.9)

17 (51) 8.5 (26) 8.8 (32) 6.9 (25) 12 (30) 6.5 (27)

30 (35) 6.6 (27) 19 (31) 5.0 (24) 16 (22) 6.4 (27)

19 (31) 54(58)

9.9 (37) 43(39)

5.0 (49) 27(39)

11 (18) 14 (14)

12 (13)

9.3 (18)

6.0 (46)

Ac

n=6 different horse urine matrices. For etiocholanolone(LC-TOF-MS) RTIS=9.98 min, for mefruside(GC-MS) RTIS=11.87 min and for methyltestosterone(GCHRMS) RTIS=9.28 min. b n=3 different days. c n=12; six aliquots of a horse urine analyzed in duplicate. * Substances used for the method development. # Threshold substances with threshold concentration presented in LOD column

27

Page 27 of 32

FIGURE CAPTIONS

Supplementary Figure 1. Sample preparation procedure of the screening method.

ip t

Figure 1. Accurate mass extracted ion chromatograms obtained from LC-TOF-MS analysis of i) blank horse urine sample; first column namely ‘BLANK’, ii) quality control horse urine -1

-1

-1

-1

-1

cr

sample fortified with amitriptyline at 500 ng mL , clanobutin at 200 ng mL , clenbuterol at 4 -1

ng mL , dexamethasone at 50 ng mL , flunixin at 200 ng mL , ketoprofen at 200 ng mL , -1

-1

-1

-1

us

mepivacaine at 50 ng mL , modafinil at 500 ng mL , morphine at 1 μg mL , nordiazepam at -1

-1

100 ng mL , oxazepam at 100 ng mL and procaine at 500 ng mL ; second column namely

an

‘QUAL. CTRL’, iii) authentic horse urine samples; third column namely ‘SAMPLE’.

Figure 2. Extracted ion chromatograms of methylated derivatives (Me) obtained from GC-MS

M

analysis of i) blank horse urine sample; first column namely ‘BLANK’, ii) quality control horse -1

-1

urine sample fortified with furosemide at 500 ng mL , phenobarbital at 100 ng mL , -1

-1

-1

d

phenylbutazone at 500 ng mL , oxyphenbutazone at 2 μg mL , ketoprofen at 200 ng mL ;

Ac ce p

‘SAMPLE’.

te

second column namely ‘QUAL. CTRL’, iii) authentic horse urine samples; third column namely

Figure 3. Accurate mass extracted ion chromatograms of trimethylsilylated derivatives (TMS)

obtained from GC-HRMS analysis of i) blank horse urine sample; first column namely -1

‘BLANK’, ii) quality control horse urine sample fortified with bumetanide at 20 ng mL , -1

-1

nandrolone at 25 ng mL and 5α-estrane-3β,17α-diol at 50 ng mL ; second column namely ‘QUAL. CTRL’, iii) authentic horse urine samples; third column namely ‘SAMPLE’.

28

Page 28 of 32

*Highlights (for review)

HIGHLIGHTS

 Detection of 362 targeted prohibited compounds in equine urine for doping control.  A unified and generic sample preparation is followed by LC-MS and GC-MS analysis.  The method enables accurate mass determinations and acquisition in full scan mode.

ip t

 The method allows retrospective analysis and easy incorporation of new substances.

Ac

ce pt

ed

M

an

us

cr

 Supplementary automated detection through library searching by using AMDIS-DRS.

Page 29 of 32

Ac ce p

te

d

M

an

us

cr

ip t

Figure

Page 30 of 32

Ac ce p

te

d

M

an

us

cr

ip t

Figure

Page 31 of 32

Ac ce p

te

d

M

an

us

cr

ip t

Figure

Page 32 of 32