sign in sign up
<<
Home , Mazipredone

Food Additives and Contaminants, September 2005; 22(9): 798–807

Androstadienetrione, a boldenone-like component, detected in cattle faeces with GC-MSn and LC-MSn

S. POELMANS1, K. DE WASCH1, H. NOPPE1, N. VAN HOOF1, M. VAN DE WIELE2, D. COURTHEYN2, W. GILLIS3, P. VANTHEMSCHE3, C. R. JANSSEN4,& H. F. DE BRABANDER1

1Laboratory of Chemical Analysis, Research Group of Veterinary Public Health and Zoonoses, Department of Veterinary Public Health and Food Safety, Faculty of Veterinary Medicine, Ghent University, Salisburylaan 133, B-9820 Merelbeke, Belgium, 2Federal Feed & Food Laboratory, Braemkasteelstraat 59, B-9050 Gentbrugge, Belgium, 3Federal Agency for the Safety of the Food Chain, WTC III, Simon Bolivarlaan 30, B-1000 Brussels, Belgium, and 4Laboratory of Environmental Toxicology and Aquatic Ecology, Department of Applied Ecology and Environmental Biology, Faculty of Bioscience Engineering, Ghent University, J. Plateaustraat 22, B-9000 Ghent, Belgium

(Received 29 March 2005; revised 10 June 2005; accepted 17 June 2005)

Abstract Boldenone (1,4-androstadiene-17-ol-3-one, Bol) has been the subject of a heated debate because of ongoing confusion about its endogenous or exogenous origin when detected in one of its forms in faecal or urine samples from cattle. An expert report was recently written on the presence and metabolism of Bol in various animal species. Androstadienedione (ADD) is a direct precursor of 17 -boldenone ( Bol). It is a 3,17-dione; ßBol is a 17-ol-3-one. Not much is published on 1,4-androstadiene- 3,17-diol, which is a 3,17-diol (ADL). If animals were exposed for a longer period to one of these analytes, a metabolic pathway would be initiated to eliminate these compounds. Similar to recent testosterone metabolism studies in the aquatic invertebrate Neomysis integer, ADD, ßBol and ADL could also be eliminated as hydroxymetabolites after exposure. The presence of 11-keto- or 11-hydroxy-metabolites in faecal samples can interfere with a confirmation method by gas chromatography-negative chemical ionization mass spectrometry (GC-NCI-MS), after oxidation of with a double bond in the A-ring (e.g. or its metabolite ). The presence of androstadienetrione (ADT) in faecal samples of cattle has never been reported. The origin of its presence can be explained through different pathways, which are presented in this paper.

Keywords: Prednisolone, , metabolite, androstadienedione, boldenone, androstadienetrione, GC-MSn, LC-MSn

Introduction lot of suspicion that these analytes could also be used as growth promoters in cattle. Residue laboratories Residue analysis of steroidal compounds in animals face difficulties confirming the origin of the analyte. used for food production is of great importance. Are they endogenous or have they been administered 17 -boldenone ( Bol, a 17-ol-3-one) is a illegally? There is still an unexplained excretion in with strong anabolic activity. An expert report on its untreated animals. presence and metabolism in various animal The metabolism of testosterone (17 -hydroxy- species was presented (De Brabander et al. 2004). androst-4-en-3-one), which differs only in a double Androstadienedione (ADD, a 3,17-dione) is the bound with Bol, has been studied intensively. We direct precursor of Bol. Androstadiene-3,17-diol have previously used testosterone as a substrate (ADL, a 3,17-diol) has been called ‘1,4-androdiol’ for metabolic studies with the aquatic invertebrate and is used by bodybuilders as a muscle enhancer. Neomysis integer (Crustacea; Mysidacea) (Verslycke Because of the presence of a double bond at the et al. 2002). These studies demonstrated the 1-position, it could have high oral activity. There is a presence of cytochrome P45011B1 (CYP11B1) in

Correspondence: H. F. De Brabander. E-mail: [email protected] ISSN 0265–203X print/ISSN 1464–5122 online ß 2005 Taylor & Francis DOI: 10.1080/02652030500239508 Androstadienetrione in cattle faeces 799

N. integer, which is the enzyme responsible for the In an attempt to explain the discrepancy between the hydroxylation in the 11-position. This enzyme is also results of the used detection methods, different present in mammals, and thus in cattle. We have possible pathways were investigated. The objective found interesting similarities between steroid meta- of this study was to investigate the identity of bolism in the invertebrate N. integer and vertebrate a compound that was detected as an derivatized models (De Wasch et al. 2002). 11-Ketosteroids oxidized product by GC-NCI-MS. have been described to play a role in the reproductive cycle of fish (Eckstein et al. 1982; Manire et al. Materials and methods 1999). An example of an 11-ketosteroid is adreno- sterone (4-androstene-3,11,17-trione). It is an Chemicals adrenal steroid derived mainly from the reversible All chemicals used for extraction were of analytical conversion of 11 -hydroxy-androstenedione. Adre- grade from Merck-Eurolab (Overijse, Belgium). The nosterone may also be derived directly from adrenal derivatization procedure (oxidation) was described metabolism of . Adrenosterone is a weak by Courtheyn et al. (1994). androgen but levels are elevated in hirsute female Standards were obtained from Steraloids and decreased in hypogonadic males. (Wilton, NY, USA): Prolon (prednisolone, Besides the above-mentioned anabolic com- P0650-000), Pron (prednisone, P0300-000), IFP pounds, another group of steroidal compounds, the (, P0553-000), ADT (androstadiene- glucocorticosteroids, are of importance in the residue trione, A0230-000), 11 OH-ADD (11 -hydroxy- analysis. Prednisolone (Prolon) is a synthetic gluco- 1,4-androstadiene-3,17-dione, A0170-000) and corticosteroid. In veterinary medicine, it (as the Bol ( -boldenone, A0200-000). Only Bol free alcohol) is included as an ingredient in a number ( -boldenone) was obtained from the Rijksinstituut of antibiotic preparations, which are indicated voor Volksgezondheid en Milieu (RIVM; National for intramammary administration for the treatment Institute for Public Health and the Environment, of bovine mastitis (EMEA/MRL/629/99 1999). Bilthoven, the Netherlands). Prolon can be metabolized to prednisone (Pron), 20 -dihydroxyprednisolone and 20 -dihydropredni- GC-MSn analysis sone. There is an interconversion of prednisone and prednisolone (Meikle et al. 1975; Garg and Jusko GC-NCI-MS. For GC-NCI-MS analysis, an ion- 1994). Maximum residue limits (MRLs) have been trap mass spectrometer PolarisQ (ThermoFinnigan, described in bovine species in muscle, fat, liver, Austin, TX, USA) coupled to a Trace GC 2000 kidney and milk at concentrations of 4, 4, 10, 10 and (ThermoFinnigan) was used. A Finnigan Mat 6 mgkg1, respectively, (EMEA/MRL/629/99 1999). A200S autosampler was used to inject the samples. Recently, a number of faecal samples from bulls as Analyses were performed using a non-polar 35% part of a monitoring programme were analysed for phenyl-polysilphenylene-siloxane SGE BPX-35 corticosteroids with gas chromatography-negative GC-column (25 m 0.22 mm i.d., 0.25 mm film chemical ionization-mass spectrometry (GC-NCI- thickness) (SGE, Inc., Austin, TX, USA). Glass MS). In this procedure, the target analyte is oxidized injector liners (10.5 cm 3 mm) were supplied by with potassium dichromate under acidic conditions SGE. The injector, ion source and transferline at 608C for 10 min (Courtheyn et al. 1998). temperature were, respectively, 250, 200 and After HPLC fractionation, the presence of the 2758C. A temperature gradient was used starting at derivatization product of prednisolone was identified 908C and increasing to 2708C in steps of 1 in a variety of concentrations in most of the faecal 908C min . In a second step the temperature was 1 samples. An important detail in this case is that increased to 2988C in steps of 38C min . A last 1 -boldenone ( Bol), androstenedione (AED) and ramp of 508C min to 3008C was performed. (P) were also identified in these faecal Helium was used as carrier gas at a flow of 1 samples with GC-MS electron impact (EI). The EC 1 ml min . A total of 1 ml of 25-ml sample was criteria 2002/657/EC (European Commission 2002) injected with a split-splitless injector (split flow 1 were checked and fulfilled for all identified analytes. 20 ml min , splitless time 1 min). Negative chemical After a second analysis with the same extraction and ionization mode (NCI) was used to ionize the detection method, the same results were obtained. analytes, with ammonia as the reaction gas. A third analysis was performed for the identification of Prolon and Pron with a different detection GC-EI-MS. A trace GC 2000 coupled to an ion-trap method: liquid chromatography atmospheric pres- mass spectrometer Polaris (both ThermoFinnigan) sure chemical ionization multiple mass spectrometry was used for GC-EI-MS analysis. A Carlo Erba (LC-APCI-MS2) in positive-ionization mode. No AS2000 (ThermoFinnigan) autosampler was used to Prolon and Pron were identified with this technique. inject. Analyses were performed using a non-polar 800 S. Poelmans et al.

5% phenyl-polysilphenylenesiloxane SGE BPX-5 301, respectively, were isolated with an isolation GC-column (25 m 0.22 mm i.d., 0.25 mm film width of 3 amu, with the exception of Prolon, which thickness) (SGE). Glass injector liners had an isolation width of 3.5 amu. All components (10.5 cm 3 mm) were supplied by SGE. The were fragmented with a relative energy of 30% injector, ion source and transferline temperature (absolute energy, dependent on mass component: were, respectively, 250, 200 and 2758C. The Pron: 0.86V, ADT and 11 OH-ADD: 0.79V), temperature gradient started at 1008C and increased except again Prolon, which was fragmented with by 178C min1 to 2508C. Further, the temperature 25% of relative energy (absolute energy: 0.72V). In increased to 2738C in steps of 28Cmin1 and was addition, a full-scan mass spectrum was acquired. finally ramped at 308C min1 to 3008C. Helium was used as carrier gas at a flow of 1 ml min1. A total of Extraction procedure 1 ml of 25-ml sample was injected with a split-splitless injector (split flow 20 ml min1, splitless time 1 min). The extraction procedure of corticosteroids from Electron ionization (EI) was used in the positive-ion faeces has previously been described by Courtheyn mode, MS full-scan acquisition. et al. (1994). In short, faecal samples were diluted with water and extracted with diethyl ether. Further

n clean-up comprised evaporation of the extracts and LC-MS analysis redissolving in methanol. This solution was purified LC-APCI-MS2 in positive-ion mode method 1 (Federal by a solid-phase extraction and HPLC fractionation. Feed and Food Laboratory). Chromatographic The derivatization before GC-MS analysis was separation was achieved using an Alltima C18 5 mm dependent on the ionization method. NCI (negative column (250 mm 3.0 mm i.d.) (Alltech Associates, chemical ionization) analysis was preceded by Deerfield, IL, USA). The HPLC apparatus consisted derivatization with potassium dichromate, while EI þþ of an Agilent 1100 system with quaternary gradient (electron impact) needed MSTFA derivatization. þþ pump, degasser, column oven and autosampler. MSTFA is a mixture of MSTFA (N-methyl- The mobile phase consisted of a gradient of 1% N-(trimethylsilyl)trifluoroacetamide), NH4I (ammo- acetic acid in water and methanol at a flow rate of nium iodide) and ethanediol. 0.5 ml min1. The mass spectrometrical analysis was The procedure used for the extraction of performed using an LCQdeca ion-trap mass spectro- (cortico)steroids from N. integer with ethylacetate meter (ThermoFinnigan). ADT and the internal was previously by Verslycke et al. (2002). standard (deuterated ) were measured in full- 2 scan and in MS . The pseudo molecular ion of ADT Results 299 was isolated with an isolation width of 2.5 amu and fragmented with a relative energy of 26% Mass spectral analysis, GC-NCI-MS (absolute energy 0.85V). Blank faecal samples and blank faecal material spiked at 2 mgkg1 Prolon and IFP (internal standard) LC-APCI-MS2 in positive-ion mode method 2 were analysed by GC-NCI-MS and used as quality (Veterinary Public Health and Food control samples. The diagnostic ions of Prolon Safety). Chromatographic separation was achieved (177, 297, 298, 299) eluted at a relative retention using an YMC column J’Sphere ODS M-80 time (relative to the internal standard IFP) of (Achrom, Zulte, Belgium). The HPLC apparatus 1.09 min 8.41E–4 min. The ion ratios and toler- comprised of a TSP P4000 pump and an AS3000 ance levels, as mentioned in Commission Decision autosampler (Thermo Separation Products, San 2002/657/EC (European Commission 2002), of the Jose, CA, USA). The mobile phase consisted of a derivatization product of Prolon are given in Table I. mixture of 45% acetonitrile and 55% water and was For ions with a relative intensity >50, >20–50, pumped at a flow rate of 0.5 ml min1. The gradient was changed from 5 to 20 min to 100% acetonitrile. Table I. Relative ion intensities of Prolon in the spike and the In between samples, there was an equilibration time tolerance levels, calculated as prescribed in Commission Decision of 5 min at initial conditions. The mass spectro- 2002/657/EC (European Commission 2002). metrical analysis was performed using an LCQdeca Relative ion-trap mass spectrometer (ThermoFinnigan). Diagnostic intensity Lower Upper The ions were acquired after atmospheric pressure ions (% of base peak) limit (%) limit (%) chemical ionization (APCI) in positive-ion mode 177 82 66 99 MS2. In a one-time segment five scan events were 297 16 11 21 performed. The pseudo-molecular ions of Prolon, 298 100 80 120 299 20 15 25 Pron, ADT and 11 OH-ADD; 361, 359, 299 and Androstadienetrione in cattle faeces 801

>10–20% as a per cent of the base peak, the to keto-groups, the results had to be interpreted tolerances are, respectively, 20, 25 and 30%. cautiously. Different compounds can result in the In the blank faeces samples no signal was observed same derivatization product. An LC-MS2 procedure for the oxidation product of Prolon (Figures 1 and 2). was used to confirm the previous results. In almost 70% of the batch of 40 samples, the signal corresponding with the derivatization product of Prolon was found. The relative retention time was Mass spectral analysis, LC-MSn within the required limits and the relative abundance At first an LC-APCI-MS2 method in positive-ion of the diagnostic ions was situated within the mode, created at the Federal Feed and Food tolerance levels, according to Commission Decision Laboratory, was used to confirm the presence of 2002/657/EC (European Commission 2002). The Prolon and its major metabolite Pron in samples extraction and detection procedure was repeated and giving GC-NCI-MS detection for the derivatization the same results were obtained. Since the detection product of Prolon. The presence of Prolon or Pron method used for the faecal samples involved a could not be confirmed in these samples. Both the derivatization procedure in which the 17-side chain components were identified in the quality control was removed and all hydroxyl-groups were oxidized samples spiked at the level of 2 ppb. In a next stage, the Department of Veterinary O Public Health and Food Safety developed an OH O O 2 HO O HO LC-APCI-MS method in positive-ion mode to OH H H H detect Prolon, Pron, ADT and 11 OH-ADD. The 2 H H H H H H differences between both LC-MS methods were O O O because the samples were analysed in two different Figure 1. Structure of Prolon, ADT and 11 OH-ADD. laboratories.

c:\02. data\040212\040212s27 12/02/2004 23:29:45 tolueen

RT : 9.80 - 10.80 SM : 7G 297.9 NL: 2.17E4 100 100 10.49 m/z= 80 177.0 80 1 176.5-177.5+ 297.5-298.5 F: -c Full 60 60 ms [ 150.00-570.00] 40 MS 040212s02 40 20 20 297.1 299.0 9.86 9.91 9.95 10.05 10.1910.25 10.32 10.67 178.0 199.0 227.0 283.0 300.0 361.9403.3 423.3 453.1 492.7 512.1 547.1 Relative Abundance Relative Relative Abundance Relative 0 0 9.8 9.9 10.0 10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8 150 200 250 300 350 400 450 500 550 Time (min) m/z

RT : 9.80 - 10.80 SM : 7G NL: 3.00E4 100 298 100 10.45 m/z= 177 80 80 2 176.5-177.5+ 297.5-298.5 F: -c Full 60 60 ms [ 150.00-570.00] 40 MS 040212s06 40 20 297 20 178 299 9.87 9.98 10.0910.14 10.25 10.34 10.75 176 215 228 271 311341 355 375 397 407 432 485 488 493 546 557 Relative Abundance Relative 0 196

Relative Abundance Relative 0 9.89.910.010.110.210.310.410.510.610.710.8 150 200 250 300 350 400 450 500 550 Time (min) m/z RT : 9.80 - 10.80 SM : 7G 298 10.44 NL: 6.52E5 100 100 m/z= 80 177 80 3 176.5-177.5+ 60 60 297.5-298.5 F: -c Full ms [ 150.00-570.00] 40 40 MS 040212s29 20 20 297 299 9.84 9.96 10.1010.22 10.71 176 178 544 557 Relative Abundance Relative 0 191 246296 301 333 371 406433 458 487 522

Relative Abundance Relative 0 9.89.910.010.110.210.310.410.510.610.710.8 150 200 250 300 350 400 450 500 550 Time (min) m/z

RT : 9.80 - 10.80 SM : 7G 10.40 NL: 3.31E2 100 100 10.20 4 10.66 m/z= 80 80 9.87 10.51 176.5-177.5+ 10.02 10.78 60 60 297.5-298.5 F: -c Full 225 10.57 ms [ 150.00-570.00] 371 381 401 40 40 215 281 355 468 MS 040212s27 272 304 327 422 463 492 543 10.26 20 158 200 231 20 242 516 546

Relative Abundance Relative 0

Relative Abundance Relative 0 9.89.910.010.110.210.310.410.510.610.710.8 150 200 250 300 350 400 450 500 550 Time (min) m/z

Figure 2. (left) Trace 1 standard Prolon, trace 2 spiked faeces 2 mgkg1, trace 3 non-compliant faeces sample and trace 4 blank sample; (right) spectrum of Prolon oxidation product with m/z 177, 297, 298 and 299. 802 S. Poelmans et al.

Discussion were retained for further investigation: , Three pathways were investigated for the explanation , , , of the presence of the derivatization product of etabonate, mazipredone, , prednimus- Prolon, described in a screening procedure with tine, prednisolone, prednival, prednisolone GC-NCI-MS. First, the possibility of the presence of 21-diethylaminoacetate and prednisolone sodium another corticosteroid, with the same derivatization phosphate. Figure 3 shows the chemical structure product as Prolon, in the samples is described. of the analytes. The clean-up procedure of the faecal Second, metabolic pathways of oxidized boldenone samples contains an HPLC fractionation step. are considered. Third, the possible formation Prednisolone and prednisone both eluted in the of ADT, derived from oxidized phytosterols, is same fraction that gave the positive response discussed. with GC-NCI-MS. Since the presence of both The results obtained after using a derivatization compounds was not confirmed with an LC-MS procedure, which oxidizes the 17-side chain of procedure, prednisolone and prednisolone esters compounds, need to be interpreted with caution. (prednisolone 21 diethyl aminoacetate, prednisolone Different analytes leading to the same derivatization sodium phosphate, prednicarbate, product could be the cause of the interference and prednival) could also be eliminated from the detected with screening method GC-NCI-MS. list. The esters could be eliminated because illegal Therefore, possible analytes leading to the same administration of prednisolone-esters would result in derivatization product as Prolon were searched for in the detection of prednisolone or prednisone in the The Merck index (ChemOffice, CambridgeSoft faeces samples. From the remaining six analytes, we Corp., Cambridge, MA, USA). The search was checked which could react to the same derivatization performed using the basic structure of Prolon, product as prednisolone and elute in the same HPLC without any substitution on the 17-position. Forty- fraction. Budesonide, ciclesonide and desonide are six analytes were selected from the Merck index acetonides, deflazacort is an oxazoline. In theory, (Table II). these products can react to the Prolon derivatization The analytes containing a fluor or chlorine were product. However, acetonides and oxazolines would first eliminated since the Prolon derivatization elute later than prednisolone on a C18 column. This product does not contain fluor or chlorine. means that even though the acetonides and the Compounds that cannot be oxidized to the Prolon oxazoline would react to form the same derivatiza- derivatization product because of the presence of a tion product as Prolon, they would not be captured substitution are formebolone (2-COH), methylpred- in the correct HPLC fraction. nisolone (6-CH3), (16-CH2) and Mazipredone and loteprednol (etabonate) (16-CH3, 17-CH3). Twelve analytes remained as last possible candidates responsible

Table II. Forty-six analytes containing the basic structure of Prolon without any substitution in the 17 position.

HO

H

H H

O

Aclometasone acetate Prednicarbate Prednimustine Beclomethasone propionate Prednisolone Formebolone Prednisolone Budesonide Flumethasone 21-diethylaminoacetate Ciclesonide Halobetasol propionate Prednisolone sodium phosphate acetonide acetate Prednival Loteprednol etabonate Prednylidene Deflazacort butyl Mazipredone Rimexolone Desonide fuorate acetate Androstadienetrione in cattle faeces 803

O O CH O 3 CH3 O CH3 O OH O H CH CH 3 O N O 3 CH O CH 3 H3C HO 3 HO HO O O O H CH CH3 H CH3 3 H

H H H H H H O O O Budesonide Ciclesonide Deflazacort

CH3 O O OCl O HO N CH CH O O CH 3 O CH 3 3 CH3 OH HO O 3 HO HO N CH CH CH H O 3 3 H 3 CH3 H

H H H H H H O O O Desonide Loteprednol etabonate Mazipredone

Cl O N O Cl H3C O O O O HO CH3 O O CH3 O CH3 O HO CH3 OH HO HO CH3 CH3 H O CH3 H O CH3 H H H H H H H O O O Prednicarbate Prednimustine Prednival

Figure 3. Chemical structures of analytes that can interfere with the detection of the oxidation product of prednisolone.

O O OCl OCl

HO HO OH O O H H derivatisation: O oxidation H H H H O O O loteprednol etabonate loteprednol in faeces after administration HO of loteprednol etabonate H

H H

O O N oxidation product HO OH N H derivatisation: H H oxidation O mazipredone

Figure 4. Possible formation of the derivatized oxidation product of Prolon from mazipredone and loteprednol etabonate. for the positive GC-NCI-MS screening results Besides the analytes that produced the same (Figure 4), where the LC-MS confirmation method derivatization product of Prolon by oxidation, would be negative for Prolon. The presence of these metabolism also needs to be considered. An impor- analytes in the faeces samples could be confirmed tant consideration is the link to the presence of with a specific LC-MS method and the use of boldenone in faeces samples. Boldenone, ADD standard solutions. and ADL (androstadienediol), all three having two 804 S. Poelmans et al. double bounds in the A-ring, can metabolize to with MSTFAþþ. Therefore, an LC-MS2 confirma- an 11-hydroxy metabolite. Cattle do have steroid tion gave additional information. Using this LC- 11 -hydroxylases, by which the 11-position of APCI-MS2 method, ADT was clearly confirmed in steroids can be hydroxylated (Brown 1998; Go¨ro¨g the three examined Prolon suspected samples 2003; Popinigis et al. 2004). The oxidized derivati- (Figure 5). zation product of the 11-hydroxy metabolite This is the first study to report the presence should be the same as the derivatization product of of ADT in faeces of cattle. In fish, however, the prednisolone. In the review of De Brabander et al. presence of 11-keto-steroids (especially 11-keto- (2004), the boldenone-containing samples originate testosterone) has been demonstrated. These from male animals (bulls). Similarly, the faeces 11-keto-steroids play a crucial role in the reproduc- samples described in the present paper were from tive cycle (Eckstein et al. 1982; Manire et al. 1999). male animals. Two standards were purchased, the In vitro experiments have been performed with oxidation product androstadienetrione (ADT) and testicular tissue of the European eel in which tritiated 11 OH-ADD. The two compounds eluted in the androstenedione and progesterone were used as same HPLC fraction as Prolon and prednisone, substrates. 11-Keto and 11-hydroxy steroids were hence were highly suspected of causing the false produced. In addition, adrenosterone was identified screening signal for Prolon with GC-NCI-MS. This (Eckstein et al. 1982). An additional 1(2) dehydro- hypothesis of the production of 11-hydroxy metabo- genation of adrenosterone would form ADT. lites of boldenone, ADD or ADL was indeed valid Microbial transformation of adrenosterone by as Bol was detected in most of the suspect faecal Cephalosporium aphidicola produced three major samples. In addition, when Bol was detected in metabolites; ADT, 17 -hydroxyandrost-4-ene-3, routinely screened faeces, the GC-NCI-MS proce- 11-dione and 17 -hydroxyandrosta-1,4-diene-3, dure was used to detect the Prolon derivatization 11-dione (Musharraf et al. 2002). product. In most cases, the presence of the deriva- The invertebrate N. integer can be used as an tization product was confirmed. alternative model to vertebrates to study metabolic An attempt was made to identify the analyte pathways (De Wasch et al. 2002). An exposure of present in the faeces samples. Spiked samples N. integer to Prolon was performed and the with ADT and 11 OH-ADD were investigated produced metabolites were detected with LC- after dervatization with GC-NCI-MS and compared APCI-MS2.11 OH-ADD was detected in a high with the samples giving a signal for Prolon. The concentration, prednisone was only detected in very standards ADT and 11 OH-ADD were also injected low amounts and a trace of ADT was present. This on column without preceding derivatization. The indicates that the 17-side chain (-COCH2OH) was response of ADT with and without derivatization was removed to form a major metabolite. The removal nearly the same. 11 OH-ADD on the other hand, of the side chain is shown in Figure 6. Figure 7 showed a low response without oxidation and a 1000 shows the presence of the metabolites formed by N. times better response with derivatization. The integer after exposure to Prolon. With this knowl- response of the samples suspected for Prolon and edge, the misuse of Prolon can still be the cause the faeces spiked with ADT and 11 OH-ADD were of the suspected faeces. Knowing that the fraction also compared with and without derivatization. of excreted dose of unchanged Prolon after intra- The response of the Prolon suspected faeces and venous administration to humans is very small and the ADT spike were better without oxidation, while that elimination occurs mainly by metabolism 11 OH-ADD in faeces showed only a low response (Habet and Rogers 1989), more research need to without oxidation. These experiments indicated that be performed to study the metabolite formation of the most probable interfering analyte present in the Prolon in cattle. Prolon suspected faeces was ADT. In order to An explanation for the presence of ADT in faeces confirm unequivocally the identity of the analyte, is the oxidation of phytosterols. Phytosterols are other methods were used for confirmation. plant sterols and are therefore present in the feed of The faecal samples were also analysed by GC-MS herbivores, like cattle. Nowadays, phytosterols are electron impact (EI) method in which full-scan mass present in higher concentrations since the use of spectra were acquired to look for ‘unknowns’ and the animal fat is banned in animal feed. If phytosterols formation of metabolites. The expected m/z values were oxidized at the 11-position by the steroid were calculated from the MSTFAþþ derivative. In 11 -hydroxylase, ADT could be formed. It has none of the faeces could ADT be detected in high been proven by experiments with N. integer that the concentrations comparable with the high concentra- sidechain of the phytosterols can be removed by tions of the Prolon derivatization product screened metabolization (Poelmans et al. 2003). The admin- with GC-NCI-MS. This anomaly could be explained istration of -sitosterol was responsible for the by the fact that ADT does not derivatize very well formation of Bol and after exposure to stigmastanol Androstadienetrione in cattle faeces 805

RT : 0.00 - 11.98 SM : 7G 7.07 NL: 5.64E6 100 m/z= 270.5-271.5 F: 95 + c APCI Full ms2 90 [email protected] [ 85 160.00-290.00] MS 80 cortst151 75 70 65 60 55 50 45 40

Relative Abundance 35 30 25 20 15 10 7.48 8.08 5 8.70 0.28 1.011.90 2.34 2.88 3.46 4.21 4.976.13 6.56 9.21 9.67 10.79 0 0 1 2 3 4 5 6 7 8 9 10 11 Time (min)

cortst151 #465-492 RT:6.99-7.19 AV:14 NL:9.57E6 F:+ c APCI Full ms2 [email protected] [160.00-290.00] 281 100 95 90 85 80 75 70 65 60 55 263 50 271 45 40

Relative Abundance 35 253 30 25 20 15 223 10 221 5 211 239 241 280 169 171 175 183 185 191197 201 213 226 237 245 255 262 269 282 285 0 160 170 180 190 200 210 220 230 240 250 260 270 280 290 m/z

Figure 5. Chromatogram and spectrum of ADT detected by LC-APCI-MS2 in a faeces sample.

O HO

H

H H O O OH 11 βOH-ADD HO OH H O H H O O H

prednisolone H H O ADT

Figure 6. Removal of the C17-side chain of Prolon to form 11 OH-ADD and ADT by Neomysis integer. 806 S. Poelmans et al.

D:\01.XCalibur\...\040906s33 07/09/2004 19:57:17S 362E

RT : 2.72 - 12.94 SM : 7B 040906s33 #1546-1613 RT: 6.46-6.71 AV: 14 SB: 32 1.23-1.87 NL: 5.94 E7 6.57 NL: F:+ c APCI Full ms2 [email protected] [ 100.00-365.00] 100 2.20E8 343.1 100 TIC F: + c APCI 80 1 Full ms2 80 [email protected] [ 1 100.00-365.00] 60 325.1 60 MS 040906s33 40 307.1

40 20 279.1 289.2 135.0 147.1 173.1211.2 223.1 265.1 344.1 Relative Abundance 0

Relative Abundance 20 100 150 200 250 300 350 m/z 0 3.65 4.36 5.487.12 8.29 8.60 10.27 10.89 12.07 7.13 NL: 040906s33 #1680-1751 RT: 7.00-7.28 AV: 14 SB: 321.23-1.87 NL: 8.75E4 100 5.16E5 F:+ c APCI Full ms2 [email protected] [ 100.00-365.00] TIC F: + c APCI 341.0 100 80 2 Full ms2 [email protected] [ 80 100.00-365.00] 2 60 MS 040906s33 60 323.1 40 313.1 40 265.2 5.31 20 295.1 6.71 183.2 213.1 225.1 261.1 277.2 342.0 6.50 10.92 135.0 171.3 Relative Abundance

Relative Abundance 20 11.69 0 3.29 4.53 9.37 8.12 9.30 10.34 100 150 200 250 300 350 m/z 0 10.91 NL: 100 1.56E8 040906s33 #2577-2635 RT: 10.79-10.98 AV: 11 SB: 311.23-1.87 NL: 7.72E7 TIC F: + c APCI F:+ c APCI Full ms2 [email protected] [ 100.00-310.00] 80 3 Full ms2 283.0 [email protected] [ 100 100.00-310.00] 80 60 6.58 MS 040906s33 3 60 40 40

20 187.1 Relative Abundance 20 173.1 265.2 121.1 135.1 147.0 201.2225.2 238.6 283.9 Relative Abundance 0 3.00 3.45 4.63 6.26 7.11 7.92 8.59 9.16 9.92 11.66 0 6.59 NL: 100 120 140 160 180 200 220 240 260 280 300 100 3.40E6 m/z TIC F: + c APCI 040906s33 #2707-2760 RT: 11.30-11.51 AV: 11 SB: 280.97-1.54 NL: 1.66E5 80 Full ms2 [email protected] [ F:+ c APCI Full ms2 [email protected] [ 100.00-310.00] 281.1 100.00-310.00] 100 60 MS 040906s33 80 263.2 10.99 4 40 60 4 11.39 40 253.0 Relative Abundance 20 223.0 271.0 7.08 299.0 7.31 11.64 20 138.8 147.4 3.454.46 6.04 8.83 9.40 9.95 220.9 226.0 300.3 5.18 120.6 161.3 187.3 298.3 0 Relative Abundance 0 3 4 5 6 7 8 9 10 11 12 100 120 140 160 180 200 220 240 260 280 300 Time (min) m/z

Figure 7. Prolon (trace 1) metabolization by Neomysis integer. Formation of metabolites Pron (trace 2), 11OH-ADD (trace 3) and a trace of ADT (trace 4) confirmed by LC-APCI-MS2.

O HO H H H O HO 11OH-ADD H H O H H H H O HO HO H β-sitosterol H H O ADT

Figure 8. Transformation of a phytosterol to the oxidation products (11OH-ADD and ADT).

N. integer produced AED. The substitution of the Conclusions side chain from the phytosterols by a keto-group was responsible for the formation of these steroids. In faeces of cattle, a new boldenone-alike substance, 11OH-phytosterols, in which the removal of the androstadienetrione (ADT), was detected. It can C17-side chain and the oxidation at the C17-position interfere with corticosteroid detection methods using occurs, could be an origin of the presence of the a derivatization technique, where an important side oxidation products (11 OH-ADD or ADT) in the chain of the corticosteroid structure is eliminated on faeces samples (Figure 8). the C17 position. Although the interpreted data Androstadienetrione in cattle faeces 807 matched all criteria as described in the 2002/657/EC De Wasch K, Poelmans S, Verslycke T, Janssen C, Van Hoof N, decision, a false ‘non-compliant’ result could be De Brabander H. 2002. An alternative for vertebrate animal experiments in the study of metabolisation of illegal growth obtained. Laboratories should be aware of this pitfall promoters and veterinary drugs. Analytica et Chimica Acta and use this GC-MS method only as a screening 473:59–69. method. The results should be confirmed by a Eckstein B, Cohen S, Hilge V. 1982. Steroid production confirmation method. in testicular tissue of the European eel. Endocrinology Possible causes for the presence of the derivatiza- 110:916–919. tion product screened with GC-NCI-MS were European Commission. 2002. Decision 2002/657/EC. Official Journal of the European Communities no. L221: 8–36. investigated and described here. First, the oxidation EMEA/MRL/629/99-FINAL July 1999. Committee for of other corticosteroids was discussed. Mazipredone veterinary medicinal products prednisolone (as free alcohol). and loteprednol etabonate are two corticosteroids Summary Report. pp 1–7. that are potentially responsible for the presence of Garg V, Jusko WJ. 1994. Bioavailability and reversible metabolism the derivatization product. Second, the presence of a of prednisone and prednisolone in man. Biopharmaceutics and Drug Disposition 15:163–172. triketon steroid (ADT) was investigated. This ADT Go¨ro¨g S. 2003. Chemical and analytical characterization of related could be present in the faeces samples through the organic impurities in drugs. Analytical and Bioanalytical metabolization of boldenone or through the conver- Chemistry 377:852–862. sion of phytosterols. The present authors previously Habet SM, Rogers HJ. 1989. Urinary excretion of prednisolone demonstrated that the side chain of the phytosterols following intravenous administration in humans. Journal of could be removed to form steroids by using an Clinical Pharmacology 29:922–927. Manire CA, Rasmussen LE, Gross TS. 1999. Serum steroid invertebrate model. Further research is needed to hormones including 11-ketotestosterone, 11-ketoandrostene- clarify the exact origin of ADT. The endogenous dione and dihydroprogesterone in juvenile and adult origin of ADT is not yet proven. bonnethead sharks, Sphyrna tiburo. Journal of Experimental Zoology 284:595–603. Meikle AW, Weed JA, Tyler FH. 1975. Kinetics and intercon- version of prednisolone and prednisone studied with new References radioimmunoassays. Journal of Clinical Endocrinology and Metabolism 41:717–721. Brown GD. 1998. The biosynthesis of steroids and triterpenoids. Musharraf SG, Atta-Ur-Rahman, Choudhary MI, Sultan S. 2002. National Products Reports 653–696. Microbial transformation of (þ)-adrenosterone. Natural Courtheyn D, Vercammen J, De Brabander HF, Vandenreyt I, Product Letters 16:345–349. Batjoens P, Vanoosthuyze K, Van Peteghem C. 1994. Poelmans S, De Wasch K, Martele´ Y, Schilt R, Van Hoof N, Determination of dexamethasone in urine and faeces of treated Noppe H, Verslycke T, Janssen CR, Courtheyn D, cattle with negative chemical-ionization mass spectrometry. De Brabander HF. 2003. The possible transformation of Analyst 119:2557–2564. phytosterols to boldenone. Proceedings of Euro Food Courtheyn D, Vercammen J, Logghe M, Seghers H, De Wasch K, Chemistry XII. pp 74–78. De Brabander H. 1998. Determination of betamethasone and Popinigis J, Antosiewicz J, Kaczor J, Rauchova´ H, Lenaz G. triamcinolone acetonide by GC-NCI-MS in excreta of treated 2004. Oxidation of glycerol-3-phosphate in porcine and bovine animals and development of a fast oxidation procedure for adrenal cortex mitochondria. Acta Biochimica Polonica derivatisation of corticosteroids. Analyst 123:2409–2414. 51:1075–1080. De Brabander HF, Poelmans S, Schilt R, Stephany RW, Verslycke T, De Wasch K, De Brabander HF, Janssen CR. Le Bizec B, Draisci R, Sterk SS, Van Ginkel LA, 2002. Testosterone metabolism in the estuarine mysid Courtheyn D, Van Hoof N, et al. 2004. The presence Neomysis integer (Crustacea; Mysidacea): Identification of and metabolism of the anabolic steroid Boldenone in various testosterone metabolites and endogenous vertebrate-type animal species: A review. Food Additives and Contaminants steroids. General and Comparative Endocrinology 21:515–525. 126:190–199.