Journal of Analytical Toxicology, Vol. 27, May/June 2003
Screening of Clostebol and its Metabolites in Bovine Urine with ELISAand Comparison with GC-MS Results in an Interlaboratory Study Patricia Crabbe 1,*, Ulrich J. Meyer 2, Zheng-tiang Zhi 2, Giuseppe PieraccinP, Michael O'Keeffe 4, and Carlos Van Peteghem 1
1Laboratory of Food Analysis (LFA), Faculty of Pharmaceutical Sciences, Ghent University, Harelbekestraat 72, 9000 Ghent, Belgium; 21nstitutf6r Chemo- und Biosensoflk (ICB), Mendelstr. 7, D-48149 M6nster, Germany; 3Centro Interdipartimentale di servizi di Spettrometria di Massa (CISM), University of Firenze, Viale G. Pieraccini 6, 50139 Firenze, Italy; and 4Teagasc, The National Food Centre, Dunsinea, Castleknock, Dublin 15, Ireland
I Abstract [ An extraction procedure for clostebol metabolites in urine is developed including enzymatic hydrolysis of conjugated metabolites with Helix pomatia juice (SHP) and solid.phase extraction (SPE) with further cleanup of sample extracts. For the enzymatic deconjugation step, variables such as buffer pH, amount of enzyme, incubation time, and temperature are examined. For the SPE step, different wash solutions and combinations with subsequent liquid-liquid extractions are examined. Incurred bovine urine samples,obtained through oral and intramuscular administration of clostebol acetate to animals, are used to test the performance of the developed method. In addition to the optimization of the sample pretreatment procedure, an interlaboratory study for the analysis of the incurred urine samples with ELISAand GC-MS is performed and good agreements are
of several authors (2,3,4) on the excretion and biotransformation of clostebol acetate in cattle, there is no doubt about the identity of the different metabolites, though their relative concentrations do not always coincide. In general, it is possible to distinguish two groups of major metabolites (Figure 1): after inOCOCH3
o
CI
CI
CIos~bol ece~e
,4-Chloro-4-andro~en-3,17-dlone = CLAD OH
OH
observed. c!
Introduction
o
Clostebol
Epl-clostebol
= 17~ -clostebol
= 17a-olostebol OH
Clostebol acetate is an anabolic steroid often used for illegal fattening purposes in livestock farming. To safeguard public health, the use of anabolic steroids as growth-promoting agents in cattle has been prohibited under Directive 86/469 of the European Commission (1). Clostebol acetate is strongly metabolised after oral or intramuscular administration. In general, no traces of the parent compound are found in urine of treated animals, therefore the choice of the right diagnostic marker metabolite for the detection of illegal use is of great importance. Comparing the work
o
CI
CI
4-Chloro.-4.-andro~ten-3a,1713-dlol
4-Cldoro.-4-androsten-3a-ol-17-one o
CI H 4-Chloroandrostane-3~ol-17-one * Author to whom correspondenceshould be addressed:PatriciaCrabbe, Departmentof Endocrinology(9K12),GhentUniversityHospital, De PinteIaan185, 9000 Ghent, Belgium. E-mail:
[email protected].
Figure1. Clostebol acetate and its main urinary metabolites.
Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission.
213
Journal of Analytical Toxicology,Vol. 27, May/June 2003
tramuscular administration, the main metabolites are epiclostebol, 4-chloro-4-androsten-3oc-ol-17-one,and 4-chloro-4androsten-3,17-dione (CLAD),whereas after oral intake there is also a higher abundance of 4-chloro-4-androsten-3oc,171~-diol and 4-chloro-4-androstane-313-ol-17-one. Most steroid metabolites appear as hydrophilic glucuronide or sulfate conjugates, which need to be efficiently hydrolyzed prior to further analysis. In veterinary residue control, enzymatic hydrolysis using Helix pomatia juice (SHP) is a very popular mode of deconjugation. However, in literature, side reactions caused by the digestive juice of SHP have been reported by several authors (4--8). They describe the degradation of some androgens during hydrolysis and suggest that enzyme impurities present in SHP were responsible for a 3~OH-5-ene to 4-ene-3-oxo-steroid transformation. A previous study (9) also demonstrated the particular conversion of 3-OH-4-ene structure into a 3-oxo-4-ene structure, related to the use of SHP. There is no doubt that, next to the great variability of hydrolysis conditions used at several laboratories, this phenomenon also plays a role in the evident discrepancies when comparing the clostebol metabolites found in different studies. A study by Walshe et al. (2) reported an enzyme immunoassay for the screening of clostebol metabolites using a commercial kit for the analysis of urine samples. This paper describes the analysis of incurred urine samples in an enzyme-linked immunosorbent assay (ELISA) developed inhouse. For this purpose, different immunogens were synthesized, and antibodies were produced (10,11). Following the optimization of a sample pretreatment method including enzymatic hydrolysis and solid-phase extraction (SPE), the urine samples were then analyzed with this ELISA. Gas chromatography (GC)-mass spectrometry (MS) analysiswas performed to identi~ and quantitate the individual metabolites present in each urine sample, and results of both methods were compared in an interlaboratory study.
Experimental
Reagents The clostebol acetate (4-chlorotestosterone acetate), tetramethylbenzidine, and bovine serum albumin (BSA) were purchased from Sigma Chemical Co. (Bornem, Belgium). The SHP, which contained 100,000 Fishman units/mL of ~-glucuronidase and 800,000 Roy units/mL of arylsulfatase, was purchased from Boehringer (Mannheim, Germany). Clostebol (4-chloro-4-androsten-17~-ol-3-one) was obtained from Alltech (State College, PA), and CLAD was a gift from Dr. L. Leyssens (Dr. Willems-Instituut, Diepenbeek, Belgium). The epi-clostebol (4-chloro-4-androsten-17~c-ol-3-one)was generously donated by RIVM (Bilthoven, The Netherlands). The rabbit antimouse IgG was obtained from DAKO(Glostrup, Denmark). The ProClin 300 was purchased from Supelco (Bellfonte, PA).All organic solvents were of analytical-reagent grade. Metabolites 4-chioro-4-androsten-3oc,17[Miol and 4-chloro-4androsten-3(x-ol-17-one were synthesized by Prof. Antonio TM
214
Guarna and Dr. Vittoria Giacomelli, chemists at the University of Firenze (Firenze, Italy).
Buffers and solutions The assay buffer was phosphate-buffered saline (PBS),which contained 0.05% ~veen 20 and 0.1% BSA.The coating buffer was 0.2M NaH~PO4/H20 in water. The blocking solution was PBS, which contained 2% casein. The wash solution was PBS with 0.005% "I~een 20, 0.1% casein, and 0.004% antifoam A emulsion. The substrate solution A (pH 5) contained 10raM urea hydrogen peroxide, 100raM Na2HPO4/2H20, and 50raM citric acid/H20 in water. Solution B (pH 2.4) contained 2raM tetramethylbenzidine and 50raM citric acid/H20 in a mixture of water:dimethylsulfoxide (96:4, v/v). Both solutions were stored separately and protected from light at 4~ Before the assay, equal volumes of A and B were mixed. To all buffers and solutions, 0.05% ProClin 300 was added as a preservative.
Disposablesand equipment Flat-bottomed microtiter plates (96 wells, Nunc Maxisorp) were obtained from Life Technologies (Merelbeke, Belgium). Optical activities were measured with a microplate reader model MPR-A4from Eurogenetics (Tessenderlo,Belgium). C18 columns were obtained from Mallinckrodt Baker (Griesheim, Germany).A commercial ELISAkit was supplied by Laboratoire d'Hormonologie (Marloie, Belgium). The GC-MS analyses were performed on an Agilent Technologies (Milan, Italy) GC-MS system, composed of a 6890 gas chromatograph and a 5973 mass spectrometric detector. The mass spectrometer operated in electron impact mode (70 eV). Acquisition was performed in select-ion monitoring (SIM) mode, which monitored at least two significant ions for each metabolite. The fused-silica capillary column was a Supelco MDN-5S (30-m x 0.25-mm i.d., 0.25-ram film thickness) (Sigma-Aldrich, Milan, Italy). Helium was the carrier gas at a constant flow of 0.8 mL/min. The oven temperature program was as follows: 100~ for 2 min, then increased 20~ to 220~ then increased 2~ to 265~ and finally 30~ to 310~ the final temperature was held for 5 min. Injector and GC-MS interface temperatures were 270~ and 290~ respectively. The MS source temperature was maintained at 230~
Urine samples Incurred bovine urine samples were obtained after oral and intramuscular administration of clostebol acetate, as described in detail previously (2). Urine samples were collected and filtered prior to storage at -20~ Samples included a control urine taken prior to administration of clostebol acetate (CLOS/L/E), a sample taken at 1 day after oral treatment with 500 mg clostebol acetate (CLOS/L/C),and a sample taken at 9 days following intramuscular treatment with 500 mg clostebol acetate (IM9). Urine sample IM9 was diluted with control urine ten-fold (CLOS/L/A) and fifty-fold (CLOS/L/B), urine sample CLOS/UC was used undiluted and diluted with control urine five-fold (CLOS/L/D). All urines were prepared as lyophilized samples (1 mL equivalent) and were reconstituted by adding 1 mL water before use.
Journal of Analytical Toxicology, Vol. 27, May/June2003 Methods Immunoreagents
Mice were immunized with the immunogen 4-chloro-4-androstenedione-3-carboxymethyloxime-BSA. The synthesis of this immunogen,the process of immunization, and the preparation of the enzyme-labeled analogue CLAD-horseradishperoxidase (HRP) have been described in previous publications (10,11).
Enzymeimmunoassay To one volume of rabbit antimouse IgG, six volumes of diluted HC1 (1:400, v/v) were added, and the mixture was left for 5 rain before further dilution in coating buffer at a concentration of 5 lag/mL.Amicrotiter plate was coated with this solution (200 IJL/well) and left for incubation overnight at 37~ in a humid container. After decanting the solution, the plate was blocked (300 IJL blocking solution/well) for 30 rain at morn temperature on a shaker. The plate was washed three times, and reagents were added in the following order: 50 IJL of standard solutions of CLADin assay buffer or sample solutions, 25 IlL of CLAD-HRP label (10 ng/well), and 25 IJL of monoclonal antibody (final dilution of 1:2500). After incubation for one hour at room tern1 mL Urine +
0.5 mL 2M Acetate buffer pH 5.0 + 50 IJL SliP 2 h, 52~
$
SPE step (0.5 g C18) - 5 mL Methanol - 2 • 5 mL Water - 5 mL 0.2M Acetate buffer pH 5.0 - Apply the hydrolized sample - Rinse with 2 • 1 mL 0.2M acetate buffer pH 5.0 - 2 • 5 mL Water - Dry (5 rain) - 2 mL Hexane - Dry (5 rain)
- 5 mL EtAC:MeOH (90:10)
$
2 • 2 mL 1M KOH
$
1 • 2 mL 1% NaCl
$
-Evaporate solvent - Dissolve in I mL assay buffer
$
ELISA
Figure 2. Optimized procedure for the sample pretreatment including enzymatic hydrolysiswith SliP, SPF,and liquid-liquid washing steps.
perature, the plate was washed four times. Substrate solution was added (100 IJLMelI)and the plate was incubated for 30 rain at room temperature in the dark. The reaction xi~s stopped by adding 100 IlL 1M H2SO4to each well, and the absorbance was read at 450 nm. Development and optimization of a urine sample pretreatment method
Optimization of the SPE step In a first experiment, blank urine samples were analyzed in ELISAafter a common hydrolysis step--using incubation with 50 I~LSHP/1 mL urine for 2 h at 52~ and pH 5.2--combined with variations of the SPE and/or subsequent liquid-liquid washings with alkali and salt solutions. One batch was analyzed according to the following procedure (method a): the hydrolyzed sample was applied to a C18 column, preconditioned with 5 mL methanol, 2 x 5 mL water, and 5 mL 0.2M acetate buffer (pH 5.2). The column was rinsed with 2 x 1 mL 0.2M acetate buffer and 2 x 5 mL water, dried for 5 min, and rinsed further with 2 mL hexane. After drying for 5 rain, the sample was eluted with 5 mL ethylacetate:methanol (90:10). The organic phase was subsequently washed with 2 x I mL 1M KOH and I x 2 mL 1% NaCl. Other batches were analyzed after changing or omitting the hexane washing step of the SPE procedure or the additional liquid-liquid washing steps with KOH or NaC1 (methods c-g). Finally, one batch was analyzed directly in ELISAafter a simple pH adjustment to 7.5 (method b). Some of these methods were also selected to test the influence on the recovery of urine spiked with 10 IJg/LCLAD.
Optimization of the enzymatic hydrolysis step In a second experiment, the optimal conditions for complete hydrolysis of conjugated forms were studied. Incurred samples CLOS/UA-Ewere incubated with SI-IPusing a variation of time (2 or 16 h), temperature (37~ or 52~ pH (5.2 or 7.5), and amount of SHP (0, 25, 50, or 200 laUreL urine). After hydrolysis, samples were applied to SPE including liquid-liquid washing, according to method (a) of the first experiment, before measurement in ELISA.
Evaluation of different sample pretreatment methods In a third experiment, the incurred urine samples were analyzed in a commercial ELISA kit following sample pretreatment according to: (a) the instructions of the manufacturer with a few modifications, (b) a procedure described by Walshe et al. (2), and (c) the inhouse optimized procedure (Figure 2). In method (a), 1 mL urine was mixed with 1.5 mL hydrolysis buffer (1 mL PBS containing 33 laL of SHP) and incubated for 1 h at 37~ The sample was then applied to a C18 column previously equilibrated with 5 mL of methanol, 5 mL water, and 5 mL PBS. The column was rinsed with 2 x I mL PBS and 5 mL methanol.~ater (50:50), dried for 5 min, and rinsed with 2 mL hexane. After drying, again, for 5 rain, the sample was eluted with 3 mL ethylacetate. The eluate was evaporated, and the residue dissolvedin 1 mL dilution buffer.A ten-times diluted solution was then measured in the ELISAkit. Method (b) included a hydrolysis step for 16 h at 37~ and pH 7.5 followedby an SPE step. The column was conditioned with 215
Journal of Analytical Toxicology,Vol. 27, May/June2003
5 mL methanol, 5 mL water, and 5 mL PBS. After applying the sample, the column was rinsed with 10 mL water and 3 mL hexane, and finally the sample was eluted with 2 x 3 mL ethylacetate methanol (70:30) and washed with 2 x 5 mL 1M KOH before evaporation. All the samples were then measured in the commercial ELISAkit. Validation of the overall assay procedure The complete inhouse-developedassay procedure, including enzymatic hydrolysis, SPE, and liquid-liquid washing steps (according to the procedure described in Figure 2) with subsequent screening in ELISA (according to the procedure described previously), was validated. For the determination of the limit of detection (LOD) and limit of quantitation (LOQ), 20 blank urine samples were analyzed for the determination of accuracy and precision; control urine samples were spiked with 2, 5, and 10 IJg/LCLAD;and were then assayed (n = 10).
Interlaboratorystudy The samples CLOS/L/A-E were used for an interlaboratory study, which compared ELISAand GC--MSresults. Theywere hydrolyzed and pretreated by SPE according to the optimized procedure (Figure 2). At the Laboratory of Food Analysis(LFA, Ghent, Belgium) and the Institut ~r Chemo- und Biosensorik (ICB, Mi~nster,Germany), each of these samples were prepared four times. Each partner analyzed one batch at their own laboratory and sent one pretreated batch to the other partner for further analysis in ELISA.At LFA,the inhouse-developedELISA was used, whereas at the ICB another ELISAformat was used that combined direct coating with the same monoclonal antibody and competitivereaction with a CLAD-alkalinephophatase conjugate (12). The remaining pretreated batches from LFA and ICB were sent to the project partners of Centro Interdipartimentale di servizi di Spettrometria di Massa (CISM, Firenze,
,oo 1 80
70 6O
I
o
50
$
.o t 2O I0. 0
~
I00
O, I
Figure 3. Mean CLA
216
ISA
Italy) for further quantitative analysis of all metabolites by GC-MS. For this purpose, methyltestosterone was added as the internal standard (IS) (50 ng IS/mL urine) prior to sample cleanup. Afterwards,the samples were derivatized (trimethylsilylester and -ether) using 100 IJL of a mixture MSTFA:NH4I:ethanethiol(1000:4:6, v/w/v) for 30 rain at 70~ and injected (1 I~L,splitless mode, SIM acquisition) in the GC-MS system.
Results and Discussion Enzymeimmunoassay The mean CLAD calibration curve (Figure 3) was characterized by a working range between 0 and 100 IJglL. The results for the cross-reactivity study are presented in Table I. The monoclonal antibodies recognized epi-clostebol relatively well (35%), but 4-chloro-4-androstene-3oc-17-one only slightly (6%). Little cross-reactivity with clostebol and clostebol acetate was observed. There was no significant cross-reactionwith endogenous hormones such as testosterone, estradiol, and progesterone nor with other structural analogues or anabolics (e.g., bolde-
Table i. Cross-Reactivity of the MonoclonaiAntibody %Cross-Reactivity CLAD C]ostebol acetate Epi-clostebol Clostebol 4-Choloroandrostene-30c-ol-17-one 4-Choloroandrostene-3~-I7~-diol Testosterone Epi-testosterone Testosterone-17-sulfate Epi-testosterone-sulfate Testosterone-17-glucuronide 19-Nortestosterone Epi-19-nortestosterone 17~-Me-testosterone Progesterone a-Estradiol 13-Estradiol Androsterone Androstene-3,17-dione 19-Norandrostene-3,17-dione Dehydroepiandrosterone Dehydroepiandrosterone-sulfate 5cr 5~-Androstane-3,17-dione 50~-Androstane-313-ol-I7-one 5~.-Androstane-3oc-ol-17-one Boldenone Me-boldenone Fluoxymesterone Stanozolol Urobiline Creatinine
100 < 0.01 35 1 6 0.05 16 h at 37~ (E) > 16 h at 52~ (F). The very low concentrations determined for the procedure without hydrolysis (treatment H) confirmed that only a minor
portion (5%) of the metabolites occur unconjugated in urine (13). On the other hand, these results also indicated that the antibodies used show no cross-reactivity towards the conjugated forms of the metabolites. For further experiments, the combination of 501JL SHP/mL urine, 2h at 52~ pH 5.2 (treatment D) was chosen. The complete procedure for sample pretreatment is summarized in Figure 2.
Evaluation of different sample pretreatment methods The developed method for sample pretreatment was compared with two other methods, one described in the literature and one described for use with a commercial ELISAkit. The commercial ELISAkit is the only one on the market for analysis of clostebol and its metabolites. In contrast with the ELISAdescribed in this work, the commercial ELISAkit uses polyclonal antibodies generated against clostebol-17-hemisuccinate,which showed the following cross-reactivity: CLAD,100%; clostebol, 28%; and clostebol acetate, 120%. Results (Figure 5) for the intercomparison showed that the procedure for sample pretreatment described for the commercial ELISAkit (method 1) yields incomplete hydrolysis of conjugates. As was found for the monoclonal antibody used in this work, the polyclonal antibodies in the commercial ELISA kit do not recognize unhydrolyzed conjugates of the metabolites. With the exception of sample CLOS/L/C (analyzed by method 2), the highest results, suggesting a more complete hydrolysis, were found with the inhouse optimized procedure (method 3). For further comparison, the results for analysis of samples prepared by method 3 using the developed monodonal antibody based ELISA are included. The higher concentrations found with the latter method could be because of higher cross-reactivity with epiclostebol, which is present in high amounts in samples CLOS/L/Aand CLOS/L/C as determined by GC-MS (Table Ill). Validation of the overall assay procedure The LODs and LOQs--defined as the concentration of CLAD that corresponds with the mean response of 20 blank urine
45 40
~ 35
l Method 1
0
~= 30
[] Method 2 I Method 3 noptim, proc. Mab
"E 25
Table III. Determination of Metabolites in Incurred Urine Samples by GC-MS Following Sample Pretreatment as in Figure 4
g 20 Concenlrations measured by GC-MS (pg/t)
tO
Samples
10 G)
0 A
C D E Samples Figure 5. Analysis of incurred urine samples (CLOS/L/A-E) with commercial kit using polyclonal antibodies.The following methodsof sample pretreatmenthavebeen followed: method I according to the instructions of the manufacturer, method 2 according to Walshe et al. (2), and method 3 according to inhouse optimized procedure (seeFigure 4). For comparison, the resultsfollowing method 3 in combination with analysis in ELISAusing monoclonal antibodies have been incorporated.
218
B
Clostebol
CLOS/L/A PretreatmentLFA PretreatmentICB CLOS/L/B PretreatmentLFA Pretreatmen~IC8
CLOS/L/C PretreatmentLFA PretreatmentICB C[OSILID PretreatmentLFA Pretrealment ICB
3.3
Epi-clostebol
CLAD
46.7 48.3
12.8 12.8
9.7 9.5
5.2 5.2
55.2 75.3
7.5 9.7
13.8 10.1
4.8 3.6
Journal of Analytical Toxicology, Vol. 27, May/June 2003
samples minus three and six times, respectively, the standard deviation--were 0.4 and 0.7 IJg/L CLAD,respectively. At concentrations of 2, 5, and 10 IJg/LCLAD,the accuracy (expressed as %recovery) was 99%, 83%, and 85%, respectively. The precision (expressed as %CV) was 5.1%, 2.4%, and 6.1%, respectively. These values represent good accuracy and precision. Another approach for the evaluation of accuracy is to work with incurred samples, which contained conjugated forms of the analytes. Because the concentrations in such samples are unknown, the test results from the ELISAmethod have to be compared with results from another well-characterized and validated method. This approach corresponds with the experiments described in the interlaboratory study, where the ELISA results are compared with those of a GC-MS method.
Interlaboratorystudy An overviewof the main metabolites found by GC-MS in the incurred urines is given in Table III. The amount of each molecule was calculated using a calibration curve prepared by adding scalar amounts of standards (epi-clostebol, clostebol, CLAD,and 4-chloro-4-androsten-3(x-ol-17-one) and a constant amount of IS to a blank urine, which was processed like the samples. In general, all the samples pretreated in the two laboratories (LFA and ICB) compared well with each other. The overall results of the interlaboratory study are presented in Figure 6. In order to compare the concentration of metabolites detected by GC-MS with the results obtained by immunoassay, only those metabolites for which significant cross-reactivity data are available are included. Therefore, 35% of the determined epiclostebol concentration plus the determined CLAD concentration by GC-MS are presented (Figure 6A). It can be concluded that the immunoassay results from both laboratories compare well with GC-MS results. It may be observed that the concentrations for samp]es pretreated by ICB are higher than those of LFA.These differences can be confirmed by comparing the mean concentrations for the urine samples CLOS/L/A-E obtained by the involved partners for both sample preparations separately (Figure 6B). Taking the errors of both pretreatments into account, only for sample CLOS/L/C is there a significant difference in concentration for the results from both laboratories. The mean results for all six methods in the interlaboratory study are presented in Figure 6C. Based on a method described by Meyer (12), the precision of the methods in the comparative study was evaluated by comparing the relative errors (%CV) for the immunoassay used in LFA (Table II, method D) with the relative errors for the interlaboratory analyses using the same pretreated samples of LFA (Figure 6B, pretreatment LFA)and the relative errors for the interlaboratory analyses after interlaboratory sample pretreatment (Figure 6C). For samples A, B, C, and D, the following mean variations have been calculated for the developed immunoassays of LFAand the comparative study: ELISA,20.4%; interlaboratory analyses after sample pretreatment by LFA, 29.3%; and interlaboratory analyses after interlaboratory sample pretreatments, 31.8%. By comparing the coefficients of variation of the different labs, it can be concluded that the results correlate well.
Previous studies (2-4) reported on the wide variety of metabolites found in urine after oral or intramuscular administration. It was noticed that there was a lot of variation in the results. Dramatic differenceswere present for molecules having a 3-ol-4-ene or 3-oxo-4-ene structure. Until now, no clear explanation was given for these discrepancies. During one of our previous studies (9), however,all the possible causes of these differences were studied, and the enzymatic hydrolysis step was found to be the principal source of this problem. In combination with the variation in parameters (time, temperature, pH, and source of SHP) used for the hydrolysis step, the origin of these differenceswas mainly found in a side-reaction because of an oxidase present in the enzymatic preparations. This enzyme was able to convert the 3-ol group to a 3-oxo group in presence of a double bond between positions 4 and 5 of the A-ring of the steroid molecule. This results in a conversion of metabolite 4chloro-4-androsten-3cc-ol-17-one into CLAD and 4-chloro-4androsten-3(~,17[3-diol into clostebol. Because 4-chloro-4-
so 45
A
rlH/SPE LFA ", ELISALFA m
9
40
__
30111111
IIH/SPELFA* ELISAIC8 mH/SPELFA'~GC-MS
Jill
25 15 10 5 0
B
A
B
C
D
36,9
l'lPreb'eatmeritLFA
35,3
~35i
E
2 ~
I~PretreatmentICB
25
820 15
5,9
~1 10
7,2
6,0 6,4
0,2 o.j z
o
A
C
0
E
O
E
C 32,7
29,6
30
0 B
C
Incurred urine samples
Figure 6. Results of the comparative study of steroid determination in bovine urine by LFA and ICB. Mean concentrations obtained for each combination of separate pretreatment (LFA or ICB) and analysis (ELISA LFA, ELISA ICB, or GC-MS) (A). Mean concentration obtained from the interlaboratory analyses for the two separate pretreatments (B). Mean concentration obtained from all the interlaboratory combinations of pretreatments and analyses (C).
219
Journal of Analytical Toxicology, Vol. 27, May/June2003
androsten-3oc,1713-diol was not converted into 4-chloro-4androsten-3mol-17-one or epi-clostebol into CLAD, the oxidase activity appears to be restricted to position 3 only, which left position 17 unchanged. The percentage conversion from 3ol to 3-oxo differedaccording to the brand of SHP used and was also related to the hydrolysisconditions (9). Results in Table III show that in both samples following intramuscular (A-B) or oral (C-D) treatment, epi-closteboland CLADare present. Furthermore, no 4-chloro-4-androsten-3mol-17-one was found in any sample after hydrolysis and SPE, which could be explained by its oxidative conversion into CLAD.
Conclusions This paper describes the development, optimization, and validation of sample pretreatment and ELISA methods for the analysis of clostebol and its metabolites in bovine urine. It is clear that steroid analysis in urine with immunoassaysis highly dependent on the cross-reactivity of the antibodies used towards the different metabolites in the target matrix. The sample pretreatment discussed in this paper offers the opportunity-through a combination of sulfatase and oxidoreductase during the enzymatic hydrolysis step--to influence the conversion of metabolites in favor of the antibodies' target analytes with higher cross-reactivities (i.e., CLADand epi-clostebol). Comparison of the overall assay procedure in an interlaboratory study showed a good agreement between the results of ELISA and GC-MS analyses. It can be concluded that our ELISA method shows a good feasibility,and clostebol and its metabolites can be determined semiquantitatively.
Acknowledgement We thank Ms. A. Desmet (Laboratoryof Food Analysis,Ghent, Belgium) for her technical assistance and Dr. F. Kohen (The Weizmann Institute of Science, Rehovot, Israel) for the production of the monoclonal CLADantibodies and the synthesis of the enzyme-labeled conjugate. This work was supported by the Commission of the European Community in the framework of the European Community Research Programme (EU project SMT4-CT96-2092) with coordinator Dr. M. Salden (Euro-Diagnostica, Arnhem, The Netherlands) and by the BOF (Bijzonder Onderzoeksfonds, Ghent University, Ghent, The Netherlands). The authors acknowledge financial support from these programs.
220
References I. Council of the European Communities. EEC Directive 86/469. Official Journal of the European Communities; 26/09/86; Official Publications of the European Communities, Luxembourg, Belgium, 1986, pp 36-45. 2. M. Walshe, M. O'Keeffe, and B. Le Bizec. Studies on the determination of chlorotestosterone and its metabolites in bovine urine. The Analyst123: 2687-2691 (1998). 3. B. Le Bizec, M.P. Montrade, F. Monteau, I. Gaudin, and E Andre. 4-Chlorotestosterone acetate metabolites in cattle after intramuscular and oral administration. Clin. Chem. 44" 973-984 (1998). 4. L. Leyssens, M. Van Puymbroeck, and J. Raus. Interference of Helix pomatia extracts on the determination of methandriol in veterinary residue control analysis. The Analyst 123:2643-2644 (1998). 5. E. Vanluchene, W. Eechaute, and D. Vandekerckhove. Conversion of free 3b-hydroxy-5-ene-steroids by incubation with Helix pomatia. J. Steroid Biochem. 16:701-703 (1982). 6. G. Messeri, G. Cugnetto, G. Moneti, and M. Serio. Helixpomatia induced conversion of some 3[3-hydroxysteroids. J. Steroid Biochem. 20:793-796 (1984). 7. R. Mass~, C. Ayotte, and R. Dugal. Studies on anabolic steroids. I. Integrated methodological approach to the gas chromatographicmass spectrometric analysis of anabolic steroid metabolites in urine. J. Chromatogr. B 489:23-5,0 (1989). 8. M.-L. Scippo and G. Maghuin-Rogister. Primary extraction technologies. In Residue analysis in food--Principles and applications; M. O'Keeffe, Ed. Harwood Academic Publishers, Amsterdam, The Netherlands, 2000, p 32. 9. P. Crabbe, G. Pieraccini, G. Bartolucci, G. Moneti, and C. Van Peteghem. Influence of Helix pomatia enzyme preparations on the oxidative conversion of some clostebol acetate metabolites in urine. J. Anal Tox. 26:73-80 (2002). 10. P. Crabbe, C. Van Peteghem, M. Salden, and E Kohen. Influence of the hapten conjugation site on the characteristics of antibodies generated against metabolites of clostebol acetate. J. Agric. Food Chem. 48:3633-3638 (2000). 11. P. Crabbe, M. Salden, and C. Van Peteghem. Characteristics of antibodies to derivatives of clostebol acetate and their use in sample clean-up and rapid immunological screening of bovine urine. Proceedings Euro-Residue IV, May 8-10, 2000, Veldhoven, The Netherlands, pp 315-321. 12. U. Meyer. Fliebinjektionsimmunanalyse (FIIA) zur Bestimmung von Pestizid- und Medikament-R0ckst~nden in Realmatrizes. Inaugural-Dissertation zur Erlangung des Doktorgrades der Naturwissenschaften im Fachbereich Chemie und Pharmazie der Mathematisch-Naturwissenschaftlischen Fakult~it der Westf~lischen Wilhelms-Universit~it, MOnster, Germany, (2000). 13. B. Le 8izec, M.-P. Montrade, F. Monteau, and F. Andre. Detection and identification of anabolic steroids in bovine urine by gas chromatography-mass spectrometry. Anal. Chim. Acta. 275:123-133 (1993).
Manuscript received July 12, 2001 ; revision received September 6, 2002.
