摘要:

Study of the metabolism of testosterone, nandrolone and estradiol in cattle† T. P. Samuels,* A. Nedderman,‡ M. A. Seymour and E. Houghton HFL, PO Box 150, Newmarket Road, Fordham, Ely, Cambridgeshire, UK CB7 5WP Received 16th July 1998, Accepted 23rd September 1998 The current metabolism study was undertaken to identify key analytes in urine, plasma and bile following testosterone, nandrolone and estradiol administrations to cull cows, heifers and steers. This information will be used to develop confirmatory analysis procedures.In the present study, mixtures (1 : 1) of testosterone, nandrolone or estradiol and their deuterium labelled analogues were administered to cull cows, heifers and steers. Two analogues of deuterium labelled testosterone were synthesised and administered, to facilitate identification of metabolites. Following administration, urine, plasma and bile samples were collected and subjected to solid phase extraction. The extracts were derivatised and analysed by GC-MS.The major analytes derived from the administered steroids were identified on the basis of the twin ion peaks produced for their non-labelled and deuterium labelled analogues and their stereochemistries determined by comparison of retention times with appropriate reference standards. Using suitable internal markers, excretion profiles for the major analytes in urine and plasma have been determined and levels in isolated bile samples estimated.This work is on-going, and this paper is a summary of some of the studies completed to date. Control of the abuse of endogenous hormones in meat producing animals presents the analyst with an interesting challenge. Screening and surveillance programmes are primarily based upon immunoassay methods, but definitive confirmatory analysis methods are not available. The development of such methods requires a detailed knowledge of both the metabolism of endogenous hormones and the normal steroid profiles in each species.Based on this information, rational strategies for confirmatory analysis procedures can then be developed. Little work has been published on the metabolism of endogenous steroids in cattle,1,2 so the aim is to acquire definitive knowledge of these profiles. Experimental Synthesis of deuterated steroids [16,16,17-2H3]-Testosterone and [16,16,17-2H3]-nandrolone. Deuteration was carried out using an adaptation of the five-step procedure of Schänzer and Donike.3 Testosterone (10 g; Sigma, Poole, Dorset, UK) was dissolved in benzene (40 ml).Ethylene glycol (12 ml) and p-toluenesulfonic acid (40 mg; Sigma) were added and the solution refluxed for 16 h. This yielded ~ 10 g of crude 17b-hydroxyandrost-5-en-3-one 3-ethylene ketal, which was oxidised to androst-5-en- 3,17-dione 3-ethylene ketal by addition of N,N-dimethylformamide (DMF, 100 ml; Sigma), dichloromethane (28 ml), and pyridinium dichromate (26 g; Sigma) and stirring for 1 h at room temperature.Ether (400 ml) and H2O (100 ml) were added, thoroughly mixed for 5 min and then the organic layer decanted off. This was dried, the ether removed and the product recrystallised from methanol (MeOH, 1 l), yielding ~ 4.1 g. Addition of deuteromethanol (MeOD) (20 ml; Sigma), dichloromethane (DCM, 20 ml), 2H2O (4 ml), and 40% sodium deuteroxide (NaOD) in 2H2O (80 ml), followed by reflux for 16 h produced 3.5 g of crude [16,16-2H2]-androst-5-en- 3,17-dione 3-ethylene ketal.This was redissolved in DCM (30 ml) and ether (90 ml), and the 17-keto group was then reduced using lithium aluminium deuteride (0.6 g; Sigma), stirring for 1 h at room temperature. Water (50 ml) was added, the product extracted eight times with diethyl ether (150 ml), and dried in vacuo. The incorporation of a deuterium atom at C17 yields crude [16,16,17-2H3]-hydroxyandrost-5-en-3-one 3-ethylene ketal ( ~ 3 g). Deprotection of the ketal was achieved by refluxing with 30% aqueous acetic acid (30 ml) for 30 min.Longer reaction times resulted in a small amount of acetylation of the 17-hydroxy group. Water (100 ml) was added, the product extracted three times with diethyl ether (150 ml), and dried in vacuo. The crude product was purified by flash column chromatography resulting in high purity [16,16,17-2H3]-testosterone (1.1 g), an overall yield of 11%. All solvents were purchased from Rathburn (Walkerburn, UK) and were HPLC grade unless otherwise stated. [16,16,17-2H3]-Nandrolone was prepared from nandrolone by the same method. [16,16,17-2H3]-Estradiol.[2H3]-Estradiol was synthesised by deuteration and subsequent reduction of its 17-oxo steroid analogue, estrone (500 mg). The solid was dissolved in deuteromethanol (25 ml) and 30% aq. NaOD (1 ml) and refluxed for 48 h. The product was purified to yield ~ 500 mg of [2H2]-estrone, which was dissolved in MeOD (25 ml) and reduced with ~ 100 mg of sodium borodeuteride (NaBD4; Sigma).The resulting mixture was then purified, yielding 380 mg [2H3]-estradiol. [2,2,4,6,6-2H5]-Testosterone. Deuteration of the A-ring of the testosterone, to yield the [2H5]-analogue, was carried out by a basic exchange reaction, similar to that described above for the deuteration of estrone. Characterisation of products. The identity and purity of the final products was assessed by GC-MS, TLC and, in the case of steroids containing a chromophore (i.e.testosterone and † Presented at the Third International Symposium on Hormone and Veterinary Drug Residue Analysis, Bruges, Belgium, June 2–5, 1998. ‡ Present address: Pfizer Central Research, Ramsgate Road, Sandwich, Kent, UK CT13 9NJ. Analyst, 1998, 123, 2401–2404 2401nandrolone), by HPLC-UV. GC-MS was also used to determine the deuterium incorporation of the four steroids. Animal administrations Animal administrations were carried out at the Centre for Dairy Research at Reading University.Steroids, either alone or in combination with their deuterated analogues, were given by intramuscular injection to a total of 16 animals (cows, heifers and steers). Urine was collected for 7 d after administration, except when bile samples were required. In such cases, animals were slaughtered ~ 48 h after administration and bile was harvested from the gall bladder. Extraction of urine samples Urine samples (5 ml) were adjusted to pH 6.8, using concentrated HCl.b-Glucuronidase (2000 units; from E. coli; Sigma) was added and the samples were incubated for 2 h at 50 °C. A C18 solid phase extraction (SPE) cartridge was pre-conditioned with MeOH (5 ml), then water (5 ml), and the sample was applied. The cartridge was washed with water (5 ml), then hexane (5 ml), drying the cartridge under vacuum after each wash. The analytes of interest were eluted with ether (5 ml). The extracts were washed with aqueous saturated Na2HCO3 solution (2 3 1 ml) and dried over anhydrous sodium sulfate before removal of the solvent in vacuo.Extraction of bile samples The same procedure was used as for urine except that 500 ml of bile was diluted with phosphate buffer (4.5 ml; 0.1 M; pH 6.8), then applied to a pre-conditioned SPE cartridge. Quantitation of metabolites For quantitation of metabolites in urine, calibration standards and quality control samples (QCs) were prepared in urine which had been passed through an XAD-2 column to remove endogenous steroids (steroid stripped urine).For bile, calibration standards and QCs were prepared in phosphate buffer (pH 6.8; 0.1 M). Calibration curves were established using [2H3]-androstenediol as an internal standard. Preparation of methoxime–trimethylsilyl/trimethylsilyl (MO–TMS/TMS) derivatives The extracts from urine or bile were heated in 8% methoxyamine hydrochloride (Sigma) in pyridine (50 ml) at 80 °C for 30 min. The pyridine was removed under oxygen-free nitrogen (OFN) at 80 °C and the residue heated with 2% Ntrimethylsilylimidazole (TMSI; Sigma) in N-methyl-N-(trimethylsilyl)- trifluoroacetamide (MSTFA; 50 ml; Sigma) at 80 °C for 1 h.The sample was dissolved in chloroform–hexane (1 : 1), passed through a mini sephadex column (Lipophilic sephadex LH-20; Sigma) and the solvent removed under OFN at 80 °C. The residue was dissolved in undecane (25 ml; Sigma) for analysis by GC-MS. Preparation of oxime-t-butyldimethylsilyl/tbutyldimethylsilyl (oxime–TBDMS/TBDMS) derivatives The extracts were heated in 8% hydroxylamine hydrochloride in pyridine (50 ml; Sigma) at 80 °C for 30 min.The pyridine was removed under OFN at 80 °C and the residue heated with 8% diethylamine hydrochloride (50 ml; Sigma) and N-(tert-butyldimethylsilyl)- N-methyltrifluoroacetamide (50 ml; Sigma) at 80 °C for 1 h. Water (1 ml) and ether (2 ml) were added and the mixture vortexed. The water was frozen at 220 °C, the ether transferred to a new vial and evaporated under OFN at 80 °C.The residue was dissolved in undecane (50 ml) for analysis by GC-MS. GC-MS GC-MS analysis of the extracts was performed on a Finnigan (Sunnyvale, CA, USA) GCQ instrument or a Fisons MD800 instrument (Finnigan Mat Instruments, Manchester, UK), using a BPX5 column (25 m 3 0.25 mm id; 0.25 micron; Scientific Glass Engineering). The initial column oven temperature was 150 °C held for 1 min and the injector temperature was 250 °C.The column oven was programmed to 250 °C at 20 °C min21, 310 °C at 5 °C min21 and held for 4 min, and then to 330 °C at 20 °C min21 and maintained at 330 °C for 2 min. For qualitative analyses, full scan data (m/z 100–800) were acquired on the GCQ or MD800 instruments, whilst for quantitative work, selected ion recording (SIR) on the MD800 was used. Results and discussion Co-administration of unlabelled and stable isotope labelled analogues of endogenous hormones facilitates the detection/ identification of metabolites using GC-MS analysis.The twin ion peaks (Fig. 1) produced in the mass spectra clearly differentiates compounds derived from the administered steroid from other components in the complex chromatograms of the derivatised urinary extracts. The main metabolites identified for testosterone were three isomers of androstane-3,17-diol, 5b-androstan-3a-ol-17-one (etiocholanolone) and epietiocholanolone. The mass spectrum of etiocholanolone shows two intense peaks at m/z 270 and 360 as an MO–TMS derivative (Fig. 2). Fig. 1 and 3 show the mass spectra of this metabolite isolated from two different animal experiments (A and B), in which [16,16,17-2H3]-testosterone and [2,2,4,6,6-2H5]-testosterone were administered, respectively, and demonstrate the advantage of the use of stable isotope labelled steroids. The twin peaks in the mass spectrum from experiment A show a mass difference of just 2 (Fig. 2; m/z 270, 272, 360, 362), due to oxidation of the hydroxy function at C17. However, following administration of [2,2,4,6,6-2H5]- Fig. 1 EI Mass spectrum of the MO–TMS derivative of etiocholanolone. 2402 Analyst, 1998, 123, 2401–2404testosterone, the twin peaks show a mass difference of 5 (Fig. 3; m/z 270, 275, 360, 365), indicating that all the deuterium atoms are retained in the metabolite. The major metabolites identified for nandrolone include 17anandrolone, four isomers of estrane-3,17-diol, estranetriol and noretiocholanolone.A similar profile was determined by Daeseleire and co-workers.2 The mass spectra of the MO–TMS derivatives of two of these metabolites (5b-estrane-3a,17a-diol and its 17b-epimer), extracted from post-administration urine samples, are shown in Fig. 4. In the case of 17a-estranediol, one deuterium has been lost giving ions at m/z 242 and 244, whilst for the 17b-epimer the 17a-deuterium is retained resulting in ions m/z 242 and 245.The loss/retention of the deuterium atom at C17 was thus particularly useful in determining the stereochemistry of the 3,17-diol metabolites resulting from epimerisation at this position. The use of two derivatives was advantageous in confirming the stereochemistry of metabolites, based on retention time data. The MO–TMS/TMS derivatives tend to show a number of diagnostic fragment ions, and are ideal for confirmatory/ identification purposes. The oxime–TBDMS/TBDMS derivatives also show fragmentation when run on the GCQ, an ion-trap instrument.However when the latter derivatives are run on the MD800, a quadrupole instrument, they show intense ions in the high mass region of the spectrum, and in many cases a single ion (M+-57), making them ideal derivatives for quantitative analysis. Quantification of the major metabolites of testosterone, nandrolone and estradiol in urine, bile and plasma has been carried out, using the oxime–TBDMS derivatives.Calibration curves have been set up, using [2H3]-androstenediol as an internal standard. As an example, a typical metabolic profile for nandrolone is shown in Fig. 5. These detailed metabolic studies have identified the major metabolites of the endogenous hormones testosterone, nandrolone and estradiol in the bovine. This information is necessary for the development of strategies to confirm the abuse of these substances. However, in order to further develop such strategies, additional work is necessary to determine the normal profiles of these steroids and their metabolites in the biological fluids of interest.Acknowledgements The authors would like to thank the UK Ministry of Agriculture, Fisheries and Food for the financial support for this work and the staff at the Centre for Dairy Research, Reading University, Reading, UK, where the animal administrations were carried out. References 1 A. G. Rico, P. Benard, J. P. Braun and V. Burgat-Sacaze, Ann. Rech. Vet., 1977, 8, 135. Fig. 2 EI Mass spectrum of the MO–TMS derivative of etiocholanolone, extracted from bovine urine following administration of [16,16,17-2H3]- testosterone. Fig. 3 EI Mass spectrum of the MO–TMS derivative of etiocholanolone, extracted from bovine urine following administration of [2,2,4,6,6-2H5]- testosterone. Fig. 4 Comparison of the EI Mass spectrum of the MO-TMS derivative of 5b-estrane-3a,17a-diol (top) and 5b-estrane-3a,17b-diol (bottom), extracted from bovine urine following the administration of [16,16,17-2H3]- nandrolone. Fig. 5 Excretion profiles of the major metabolites of nandrolone in bovine urine. A, aba estranediol; B, baa estranediol; C, bab estranediol and D, epinandrolone. Analyst, 1998, 123, 2401–2404 24032 E. Daeseleire, A. De Guesquiere and C. Van Peteghem, Anal. Chim. Acta, 1993, 275, 95. 3 W. Schanzer and M. Donike, Recent advances in doping analysis (2), Proceedings of the 12th Cologne Workshop on Dope Analysis, ed. M. Donike, H. Geyer, A. Gotzman and U. Mareck-Engelke, Sport und Buch Strauss, Germany, 10–15th April, 1994. Paper 8/05531I 2404 Analyst, 1998, 123, 2401–2404

ISSN:0003-2654    

DOI:10.1039/a805531i

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

The use of stable carbon isotope analysis to detect the abuse of testosterone in cattle¢Ó Paul M. Mason,¢Ôa Sarah E. Hall,b Iain Gilmour,a Edward Houghton,*b Colin Pillingera and Mark A. Seymourb a PSRI, The Open University, Milton Keynes, UK MK7 6AA bHFL, PO Box 150, Newmarket Road, Fordham, Ely, Cambridgeshire, UK CB7 5WP Received 16th July 1998, Accepted 30th October 1998 The use of stable carbon isotope analysis to detect the administration of anabolic steroids to cattle was investigated.Samples were extracted by solid-phase extraction on C18 cartridges. Stable isotope ratios (13C:12C) were measured by gas chromatography-isotope ratio mass spectrometry (GC-IRMS) of the underivatised extracts. A programmed temperature vaporiser (PTV) injector was installed in the GC-IRMS system, which conferred a number of advantages. First, it allowed large volumes of sample to be injected whilst the injector liner was cool. The solvent was subsequently vented to the atmosphere prior to transfer of the sample to the GC column.Thus a significantly greater amount of sample could be presented for analysis, thereby increasing the sensitivity. Second, by this means virtually all the solvent could be removed prior to analysis. This eliminates solvent peak tailing, which can be a major problem in GC-IRMS. Finally, the PTV allowed the use of higher initial GC oven temperatures, which in turn facilitated the analysis of underivatised steroids by reducing the GC run time and improving the chromatographic peak shape.The carbon isotope composition of 5b-androstane-3a,17a-diol, the major metabolite of testosterone found in bovine bile, was measured in bile samples from untreated cattle and from cattle injected intramuscularly with testosterone or a mixture of testosterone esters. There was considerable inter-animal variation in the values obtained and there was no significant difference between samples from treated and untreated animals. However, when the isotopic composition of the metabolite was normalised with respect to that of an endogenous reference compound (cholesterol) in the same sample, the difference between treated and untreated animals became statistically significant.Introduction Traditionally there have been two approaches to the detection of the abuse of endogenous steroids: measurement of the absolute concentration of the steroid or its metabolite(s) or determination of the ratio between the amount of the analyte and a second steroid which is not a metabolite of the first.In the former case, a threshold value must be established by statistical analysis of steroid concentrations in a large number of samples from untreated animals. For example, a threshold of 20 ng ml21 has been set by the international horseracing authorities for testosterone in the equine male castrate.1 An example of the ratio approach comes from human athletics, where the International Olympic Committee considers that a testosterone-toepitestosterone ratio of > 6 is indicative of testosterone administration.2,3 However, neither approach is capable of discriminating directly between endogenous steroids and administered material.Stable carbon isotope measurements, however, may provide a means of doing so because the 13C:12C ratio is dependent upon the source of the steroid. In this study, the use of gas chromatography-isotope ratio mass spectrometry (GC-IRMS) to detect the administration of testosterone to cattle was investigated.The use of an endogenous reference compound (ERC) to normalise the results was examined. By this means, each animal can act as its own control. Experimental Measurement of stable carbon isotope ratios Carbon occurs naturally in two stable isotopic forms, 13C and 12C. Atmospheric CO2 has an overall 13C:12C ratio of 0.01 : 1. However, processes including photosynthesis and metabolism cause isotopic fractionation because the rates of reaction of the two isotopes are different.Hence isotopic composition can vary, depending on the origin of the material. The 13C:12C ratio for individual compounds can be determined by IRMS and is normally expressed as a d13C value: d13 13 12 13 12 13 12 1000 C C C C C C C sample reference reference = ( ) -( ) ( ) E I II ¢� ¡Æ ¢«¢« ¡Í : : : ¢¶ The analyte is combusted in an on-line furnace to form CO2, then measured in a sensitive mass spectrometer which monitors the masses m/z 44 (corresponding to 12CO2) and m/z 45 (13CO2).A third mass, m/z 46, is also monitored and is used to correct for the relative abundance of the stable isotopes of oxygen in the sample. The isotope ratio measured for the analyte(s) of interest are normalised to the ratio measured for a CO2 standard analysed alongside the unknown sample. This standard is in turn calibrated against an international standard reference material, thus ensuring comparability of data between laboratories.All organic carbon is ultimately derived from atmospheric CO2 by photosynthesis. There are two main routes for the ¢Ó Presented at the Third International Symposium on Hormone and Veterinary Drug Residue Analysis, Bruges, Belgium, June 2¡©5, 1998. ¢Ô Present address: Exchem Organics, Great Oakley, Harwich, Essex, UK CO12 5JW. Analyst, 1998, 123, 2405¡©2408 2405fixation of CO2 into plant material, namely C3 and C4. This nomenclature refers to the carbon chain length of the photosynthetic intermediates in each case.The majority of plants fix CO2 via the C3 route. However, there are a few mass produced crops that use the C4 cycle, notably maize and sugar. During photosynthesis 12CO2 is preferentially incorporated into plant tissue compared with 13CO2, resulting in more negative d13C values. Owing to the different enzymes involved, C3 plants discriminate more strongly than C4 plants against 13CO2.Thus the range of d13C values for C3 plants is 225 to 235‰ whereas for C4 plants it is 211.9 to 215.2‰.4 For human athletes, the d13C composition of endogenous steroids reflects a diet containing both C3 and C4 foodstuffs, whereas synthetic testosterone is prepared exclusively from soy products, i.e. a C3 source.5 Consequently, a negative shift in the d13C value, from 225.1 to 229.1‰, is observed following the administration of testosterone to athletes and this shift is carried through to the major metabolites of testosterone.6 In contrast, the diet of cattle in the UK is reportedly predominantly of C3 origin. Any difference in the d13C values of steroids due to the administration of testosterone is therefore likely to be less pronounced in cattle (and other herbivores) than in the human.Materials Testosterone and b-glucuronidase (type IX-A, from E. coli) were obtained from Sigma (Poole, UK), solid-phase extraction (SPE) cartridges from Anachem (Luton, UK) and Durateston from Intervet (Cambridge, UK).All other chemicals were of analytical-reagent grade or better. Animal administrations Animal administrations were carried out at the Centre for Dairy Research at Reading University. Holstein–Friesian cattle were dosed with a single intramuscular dose of testosterone (heifers and cull cows 400 mg; steers 250 mg) or Durateston (heifers and cull cows 500 mg total esters; steers 400 mg). Durateston is a proprietary preparation containing testosterone propionate (6 mg ml21), testosterone phenylpropionate (12 mg ml21), testosterone isocaproate (12 mg ml21) and testosterone decanoate (20 mg ml21).The animals were slaughtered approximately 48 h after dosing and bile was collected from the gall bladder. Extraction of bile samples Bile samples from treated and untreated cows were extracted on C18 SPE cartridges. The procedure used is shown schematically in Fig. 1. The SPE extracts were further purified by HPLC, using a 5 mm Hypersil ODS column (100 3 4.6 mm id) with an acetonitrile–water gradient (0–100% acetonitrile over 18.5 min, then held at 100% acetonitrile for 7.5 min; flow rate 1 ml min21; column oven temperature 40 °C).The column was re-equilibrated for 15 min between runs. The analytes [5b-androstane-3a,17a-diol (the jor metabolite of testosterone found in bovine bile7) and cholesterol] were collected in timed fractions, using retention times relative to a reference compound (19-hydroxyandrostendione) which was added to each extract.The fractions were diluted with water (2 volumes), then concentrated on C18 SPE cartridges using the method shown in Fig. 1. The extracts were dried under a stream of oxygen-free nitrogen, the residues reconstituted in ethyl acetate and replicate aliquots analysed by GC–IRMS. GC-IRMS analysis GC-IRMS was performed on a Finnigan MAT Delta S instrument (San Jose, CA, USA), using a BPX-5 fused silica capillary column (25 m 3 0.32 mm id; 0.5 mm film thickness) (SGE, Milton Keynes, UK).The gas chromatograph was equipped with an Optic II programmed temperature vaporiser (PTV) injector (ATAS, Cambridge, UK), which allowed large volumes (up to 100 ml) of sample to be injected. The sample was introduced into a cool liner (40 °C) and the solvent was evaporated using a stream of helium and vented to the atmosphere. The injector temperature was then increased rapidly to 400 °C, transferring the sample on to the head of the GC column.Results The d13C values measured for 5b-androstane-3a,17a-diol and cholesterol are given in Table 1. The standard deviations for replicate measurements indicate that the method is reasonably precise. There was considerable variation between animals in the d13C values for cholesterol, with values ranging from 224.5 to 228.4‰. From Fig. 2, it can be seen that although the d13C values for androstanediol in each of the post-administration samples are more negative than the mean value for untreated animals, the difference is not statistically significant. However, when the ratios of the d13C value for androstanediol to that for cholesterol are plotted (see Fig. 3), the value for each of the post-administration samples is at least 2 s above the mean value for untreated animals. When tested using f-tests and Student’s ttests, 8,9 the post-administration values were found to be significantly different from the untreated values at the 95% confidence level.As expected, the ratio of the d13C values of the analyte to that of the ERC is close to unity for untreated animals. Discussion Compared with what was observed in human athletes, the change in the d13C values measured for the testosterone metabolite 5b-androstane-3a,17a-diol following administration of testosterone or its esters to cattle were small. It is proposed that this is due to the animals’ diets being exclusively C3 based, i.e.the carbon source for the endogenous compounds is similar to that of the synthetic testosterone. The d13C value for the synthetic material, measured by GC-IRMS, was 228.5‰. It is interesting that other workers have found larger differences between endogenous and administered steroids.10 This may be due to the fact that these studies were carried out in mainland Europe, where cattle tend to be fed on diets rich in maize. Fig. 1 Solid-phase extraction of bovine bile samples. 2406 Analyst, 1998, 123, 2405–2408The mean of the values measured for endogenous cholesterol was 227.5‰, compared with the value reported for humans of approximately 224‰.6 However, the variation observed in the isotopic composition of endogenous cholesterol reflects the fact that the d13C values of C3 plant tissue can vary by up to 10‰.4 The use of an ERC to normalise the results is therefore essential.The installation of the PTV injector conferred a number of important advantages.With conventional GC injection techniques, the amount of solvent introduced is many orders of magnitude greater than the amount of analyte so that, even though it represents only a tiny fraction of the total, the solvent peak tail can cause major problems in GC-IRMS. Using a PTV injector, the injection solvent can be completely removed by venting prior to transferring the analyte on to the GC column, thus dramatically reducing the size of the solvent peak.The PTV injector has also facilitated the analysis of underivatised steroids due to its ability to remove the solvent. As a result, the initial GC oven temperature can be increased, so dramatically reducing the analysis time. This would otherwise be prohibitively long and give very broad peaks. Using the PTV injector, the peak shape is improved markedly and the retention time is reduced by a factor of more than two. Another useful attribute of the Optic II is the facility for programming the column head pressure to increase in line with the oven temperature.This counteracts the increase in the viscosity of the carrier gas, and thus prevents the decrease in flow rate normally observed with conventional GC systems. The analysis of derivatised steroids (methoxime-trimethylsilylated or acetylated) by GC-IRMS was investigated. However, both derivatisation procedures introduce additional carbon atoms and thus change the isotopic composition of the analyte.Although shorter retention times and improved peak shapes could be obtained, the variability observed in the d13C values measured for these derivatives makes them unsuitable, in our hands, for this application. However, other workers have used them successfully.11 An alternative derivatisation method, oxidation using pyridinium chlorochromate, was also evaluated. Since this procedure does not involve addition or removal of carbon atoms, it should not affect the carbon isotope composition of the molecule.However, although oxidation gave good chromatographic properties combined with an acceptably small change in isotopic composition, the process is not stereospecific, i.e. a- and b-hydroxy isomers give an identical keto derivative. The oxidation procedure is also relatively complex. The determination of underivatised steroids using the PTV was therefore the preferred option. Under the conditions used, a single peak of acceptable shape was obtained for each steroid (see Fig. 4). GC-MS analysis of standards under similar GC conditions indicated that some steroids may occur, either in the PTV or on-column. However, this should not affect the isotope ratio as it does not involve loss or addition of carbon atoms. Horning et al.6,12 also successfully analysed steroids underivatised for their work in the human. The use of immunoaffinity chromatography (IAC) as an alternative to SPE for the extraction steroids from biological fluids was investigated.It was hoped that IAC would provide a means of achieving the selectivity required for GC-IRMS in a single step. The technique is also suitable for the trace enrichment of large volumes of dilute samples, such as bovine Table 1 GC-IRMS analysis of bovine bile samples d13C (‰) Androstanediol Cholesterol Sample No. of replicates Mean Standard deviation Mean Standard deviation Steer 5582 2 230.81 0.40 228.43 0.15 Steer 5620 3 225.74 0.20 225.56 0.07 Steer 5614 3 225.75 0.25 225.13 0.19 Steer 5621 3 226.14 0.85 225.52 0.62 Steer 5626 2 227.99 0.23 224.88 0.06 Steer 5628 4 225.43 0.27 224.86 0.22 Steer 5624 1 227.58 —a 227.75 — Steer mix 1 3 228.28 0.32 227.43 0.62 Steer mix 2 3 228.55 0.95 228.15 0.31 Heifer 9 2 228.30 0.25 227.43 0.35 Heifer 11 2 226.89 0.23 226.48 0.08 Heifer 12 2 — — 227.07 1.15 Heifer 13 2 231.89 0.01 227.59 0.28 Heifer 16 2 228.55 0.79 227.65 0.21 Heifer 17 1 229.36 — 227.07 — Heifer 18 2 228.07 0.22 227.11 0.40 Heifer 19 3 228.12 0.67 227.38 0.46 Heifer 21 3 226.53 0.25 226.32 0.12 Heifer 25 3 228.15 0.90 226.90 0.57 Heifer 26 2 226.48 0.47 226.37 0.17 Heifer 27 3 227.12 0.61 226.37 0.53 Heifer mix 1 3 229.37 0.70 226.07 2.02 Heifer mix 2 3 227.68 0.88 228.12 0.74 Exp MAFF 5/95 4 229.02 0.79 225.02 0.33 Exp MAFF 3/96 6 230.87 1.76 225.89 1.62 Exp MAFF 12/96 2 230.81 1.18 224.50 0.21 Exp MAFF 14/96 2 228.83 1.10 225.69 0.03 Exp MAFF 1/97 3 229.35 0.39 226.79 0.19 a —, Value not determined.Fig. 2 d13C values for 5b-androstane-3a,17a-diol in bile from untreated cattle and cattle treated with testosterone or testosterone esters. Mean value calculated for all ‘untreated’ samples. Analyst, 1998, 123, 2405–2408 2407urine. A range of antisera were produced by immunising sheep or rabbits with protein conjugates of the steroids of interest, i.e. testosterone, etiocholanolone (a metabolite of testosterone) and nandrolone. These were screened for cross-reactivity towards the parent steroids, their metabolites and other endogenous steroids.Using a protein G column, immunoglobulin G (IgG) fractions were prepared from antisera with appropriate crossreactivity profiles. The IgG fractions were covalently bound to a solid matrix (CNBr activated Sepharose, Pharmacia, St.Albans, UK; Biospher GM1000E, Presearch, Letchworth, UK; or POROS, PerSeptive Biosystems, Cambridge, MA, USA) and packed into low pressure chromatographic columns (Pharmacia). Samples were loaded on to the columns at pH 7, the columns were washed with phosphate buffered saline (pH 7), then the analytes were eluted with methanol or propionic acid– methanol.However, although good results were achieved with standards, extracts of biological fluids proved too complex for analysis by GC-IRMS. Work is continuing to try to overcome this problem, presumably due to low levels of non-specific binding to the support material. Future work will concentrate on continuing refinement of the extraction and analysis procedures and the size of the data set for untreated animals will be extended. The GC-IRMS analysis of metabolites of testosterone in urine and plasma, i.e. the principal on-farm sample types, will also be investigated and the work will be extended to cover two other endogenous steroids, 17b-[19-nortestosterone] (nandrolone) and 17b-estradiol.Acknowledgements The authors thank the UK Ministry of Agriculture, Fisheries and Food for financial support of this work and the staff at the Centre for Dairy Research, Reading University, Reading, UK, where the animal administrations were carried out.References 1 P. W. Tang, W. C. Law and D. L. Crone, in Proceedings of the 15th International Conference of Racing Analysts and Veterinarians, Queensland, Australia, ed. D. E. Auer and E. Houghton, R and W Publications, Newmarket, 1996, pp. 68–72. 2 M. Donike, K. R. Bärwald, K. Klostermann, W.Schänzer and J. Zimmerman, in Testosterone in Sport: Leistung und Gesundheit, ed. H. Heck, W. Hollmann, H. Liesen and R. Rost, Deutscher Ärtze- Verlag, Cologne, 1983, pp. 293–300. 3 D. H. Catlin, D. A. Cowan, R. de la Torre, M. Donike, D. Fraisse, H. Oftebro, C. K. Hatton, B. Starcevic, M. Becchi, X. de la Torre, H. Norli, H. Geyer and C. J. Walker, J. Mass Spectrom., 1996, 31, 397. 4 M. O’Leary, Phytochemistry, 1981, 20, 553. 5 C. H. L. Shackleton E. Roitman, A. Philips and T.Chang, Steroids, 1997, 62, 665. 6 S. Horning, H. Geyer, U. Flenker and W. Schänzer, in Recent Advances in Doping Analysis (5): Proceedings of the 15th Cologne Workshop on Dope Analysis, ed. W. Schänzer, H. Geyer, A. Gotzman and U. Mareck-Engelke, Sport und Buch Strauss, Cologne, 1998, pp. 135–148. 7 T. P. Samuels, A. Nedderman, M. A. Seymour and E. Houghton, Analyst, 1998, 123, 2401. 8 J. C. Millar and J. N. Millar, Statistics for Analytical Chemistry, Ellis Horwood, Chichester, 2nd edn., 1989. 9 T. J. Farrant, Practical Statistics for the Analytical Scientist; a Bench Guide, Royal Society of Chemistry, Cambridge, 1997. 10 V. Ferchaud, B. Le Bizec, F. Monteau, M.-P. Montrade and F. Andre, Analyst, 1998, 123, 2617. 11 M. Becchi, R. Aguilera, Y. Farizon, M. M. Flument, H. Casabianca and P. James, Rapid Commun. Mass Spectrom., 1994, 8, 304. 12 S. Horning, H. Geyer, M. Machnik, W. Schänzer, A. Hilkert and J. Oesselmann, in Recent Advances in Doping Analysis (4): Proceedings of the 14th Cologne Workshop on Dope Analysis, ed. W. Schänzer, H. Geyer, A. Gotzman and U. Mareck-Engelke, Sport und Buch Strauss, Cologne, 1997, pp. 275–283. Paper 8/05529G Fig. 3 Ratio of d13C values (androstanediol : cholesterol) in bile from untreated cattle and cattle treated with testosterone or testosterone esters. One ‘untreated’ sample, heifer 13, was shown to be a statistical outlier at the 95% confidence level using the Grubb test.9 The data for this sample were therefore omitted from the calculation of the mean. Fig. 4 GC-IRMS analysis of underivatised steroid standards. A mixture of 5b-androstane-3a,17a-diol and cholesterol (each 50 ng ml21 in ethyl acetate) was analysed by GC-IRMS, using a PTV injector, under conditions similar to those used to analyse bile extracts. 2408 Analyst, 1998, 123, 2405–2408

ISSN:0003-2654    

DOI:10.1039/a805529g

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

Determination of betamethasone and triamcinolone acetonide by GC-NCI-MS in excreta of treated animals and development of a fast oxidation procedure for derivatisation of corticosteroids† Dirk Courtheyn,*a Jan Vercammen,a Maureen Logghe,a Hilde Seghers,a Katia De Waschb and Hubert De Brabanderb a State Laboratory (ROLG), Braemkasteelstraat 59, B-9050 Ghent, Belgium b Faculty of Veterinary Medicine of the University of Ghent, Department of Veterinary Food Inspection, Laboratory of Chemical Analysis, Salisburylaan 133, B-9820 Merelbeke, Belgium Received 29th June 1998, Accepted 11th September 1998 The use of corticosteroids in combination with other hormonal substances has long been known to result in increased mass gain with bovines.Practice has demonstrated, however, that even the single use of a glucocorticoid may result in growth promoting effects. In addition to the popular dexamethasone, more recently other corticosteroids have also been misused for fattening purposes.The first part of this study deals with the detection of two of them, namely betamethasone and triamcinolone acetonide. Betamethasone was administered orally to a cow at a dose of 50 mg d21 for 5 d, then later the same cow was injected intramuscularly with a dose of 50 mg of betamethasone dipropionate. Excretion in urine and faeces was followed with both HPLC-enzyme immunoassay and a previously described method based on negative chemical ionization mass spectrometry (NCI-MS) after oxidation. For the triamcinolone acetonide study a cow was treated with 50 mg d21 of the drug during a 7 d period.Excretion in faeces was followed with GC-NCI-MS. As triamcinolone acetonide is resistant to the previously described oxidation procedure, however, a hydrolysis step had to be introduced prior to oxidation. In addition to this specific modification necessary for triamcinolone acetonide, in a subsequent part of this study the original oxidation procedure with pyridinium chlorochromate was re-investigated especially to shorten the procedure. With the introduction of potassium dichromate the reaction time could be decreased from 3 h to 10 min.Introduction In recent years, corticosteroids have become one of the most important groups of illegal growth promoters in livestock production.1,2 At first often combined with b-agonists and/or anabolic steroids,3,4 more recently corticosteroids seem to be applied alone. Their use results in improved feed intake, increased live mass gain and a reduced feed conversion ratio.5,6 Belgium has very strict regulations concerning illegal growth promoters and severe sanctions are taken against farmers who administer non-registered corticosteroids (e.g., betamethasone) or registered corticosteroids without a prescription from a veterinarian.Screening and confirmation procedures for these compounds are usually performed by immunoassays, followed by gas chromatography-mass spectrometry (GC-MS).GC-MS in the negative chemical ionization (NCI) mode after oxidation remains one of the most powerful and attractive confirmation techniques for the determination of corticosteroids. Indeed, by a simple oxidation reaction these relatively large molecules with numerous polar groups are transformed into derivatives which are very well suited for GC and which can be detected extremely sensitively in the NCI mode. This transformation also permits the detection of most corticosteroids in a very specific way, as the matrix interferences are not sensitized.The high sensitivity is thought to be due to the strong response generated by the 1,4-dien-3-one system in ring A, combined with the 11-keto function.7 In the first part of the present study, the elimination of one of the most popular corticosteroids, betamethasone, was followed in the urine and faeces of a cow, first after oral treatment with betamethasone and second after intramuscular injection with the dipropionate ester.Both the HPLC-enzyme immunoassay (EIA) method and the GC-NCI-MS method (after oxidation of the corticosteroid) reported earlier were applied.1 In a subsequent elimination study, a cow was treated with triamcinolone acetonide. This compound, in contrast to most of the important representatives of the group of corticosteroids, is resistant to oxidation. Therefore, the earlier described oxidation procedure, preceding GC-NCI-MS detection, had to be adapted.A streamlined procedure was elaborated in which the cleavage of the acetonide function under acid conditions was followed by oxidation. The growing number of samples to be analysed and especially those taken in slaughterhouses created a demand for a faster procedure. The original oxidation procedure with pyridinium chlorochromate was re-investigated and potassium dichromate was introduced. As outlined further in this study, by this means the oxidation reaction could be shortened from 3 h to only 10 min.Experimental Reference compounds Reference products were beclomethasone (9-chloro-11b,17,21- trihydroxy-16b-methylpregna-1,4-diene-3,20-dione), flume- † Presented at the Third International Symposium on Hormone and Veterinary Drug Residue Analysis, Bruges, Belgium, June 2–5, 1998. Analyst, 1998, 123, 2409–2414 2409thasone (6a,9-difluoro-11b,17,21-trihydroxy-16a-methylpregna- 1,4-diene-3,20-dione) and methylprednisolone (11b,17,21-trihydroxy-6a-methylpregna-1,4-diene- 3,20-dione) from Sigma (St.Louis, MO, USA), betamethasone (9-fluoro-11b,17,21-trihydroxy-16b-methylpregna-1,4-diene- 3,20-dione), dexamethasone (9-fluoro-11b,17,21-trihydroxy- 16a-methylpregna-1,4-diene-3,20-dione), prednisolone (11b,17,21-trihydroxypregna-1,4-diene-3,20-dione), prednisone (17,21-dihydroxypregna-1,4-diene-3,11,20-trione) and triamcinolone acetonide {9-fluoro-11,21-dihydroxy- 16,17-[1-methylethylidenebis(oxy)]pregna-1,4-diene- 3,20-dione} from Serva (Heidelberg, Germany) and betamethasone 17,21-dipropionate, fluorometholone (9-fluoro-11b,17-dihydroxy-6a-methylpregna-1,4-diene- 3,20-dione) and isoflupredone (9-fluoro-11,17,21-trihydroxypregna- 1,4-diene-3,20-dione) from Steraloids (Wilton, NY, USA). Reagents All solvents used were of analytical-reagent or HPLC grade and were obtained from Merck (Darmstadt, Germany).Sodium acetate, potassium dichromate and sodium hydroxide were of analytical-reagent grade from Merck and pyridinium chlorochromate was purchased from Sigma (St Louis, MO, USA).Trifluoroacetic acid was of Sequanal grade from Pierce (Rockford, IL, USA). EIA kits for corticosteroids were purchased from CER (Marloie, Belgium). Cross-reactivity for betamethasone was 50% (dexamethasone, 100%). Instrumentation The HPLC system for the follow-up of the oxidation reaction consisted of a SpectraSystem P1000XR pump equipped with a SpectraSystem UV6000LP diode array detector, both from Thermo Separation Products (San Jose, CA, USA), and a Merck-Hitachi AS-4000 automatic injector (Merck).The analytical column was an Alltima C18 (5 mm) column (15 cm 3 3.2 mm id) supplied by Alltech (Deerfield, IL, USA). Mass spectrometric determinations Analysis was carried out as earlier described.1 This included spiking with flumethasone, hydrolysis (for the urine samples only), extraction with diethyl ether, clean-up with HPLC fractionation and chemical oxidation as described below.For the HPLC fractionation the collection started at the retention time of flumethasone minus 0.5 min and ended at the retention time of betamethasone (or triamcinolone acetonide) plus 0.5 min. GC-MS determinations were performed with the equipment and conditions described earlier.1 Quantitative measurements were made in the selected ion monitoring (SIM) mode at the following m/z values: betamethasone 330, 310, triamcinolone acetonide, 316, 296 and flumethasone 348, 328.Enzyme immunoassay (EIA) procedure The samples were analysed as for the GC-MS measurements, but without spiking. From the HPLC fractions, collected 0.35 min before until 0.35 min after the retention time of betamethasone, a portion of 50 ml was taken. After four-fold dilution with the dilution buffer delivered with the kit, an aliquot of 50 ml was taken for the assay. The kit applied was the CER corticosteroids EIA kit.Subsequently the test procedure was performed following the instructions provided with the kit. Oxidation procedure Original procedure used for the betamethasone excretion study. The residue obtained after evaporation of the HPLC fraction in a vacuum concentrator (45 °C) was dissolved in 50 ml of acetonitrile and 200 ml of an aqueous solution containing 50 mg ml21 pyridinium chlorochromate and 25 mg ml21 sodium acetate. The mixture was vortex mixed and heated at 92 °C for 3 h.After cooling, the oxidized compounds were extracted with 3 ml of tert-butyl methyl ether–dichloromethane (2 + 1 v/v) with vortex mixing and ultrasonication. Complete separation of the layers was achieved by centrifugation at 3500g for 5 min. After freezing, the organic phase was decanted and evaporated in a vacuum concentrator (45 °C). The residue was reconstituted with 25 ml of toluene. Modified procedure for the oxidation of triamcinolone acetonide. The residue was dissolved in 50 ml of acetonitrile and 50 ml of 6 m HCl.The mixture was vortex mixed and heated at 85 °C for 30 min. After cooling, 50 ml of 5 m NaOH and 200 ml of an aqueous solution containing 50 mg ml21 pyridinium chlorochromate and 25 mg ml21 sodium acetate were added. The mixture was vortex mixed and heated at 92 °C for 3 h. Extraction was performed as above. Newly developed fast procedure for free corticosteroids. The residue was dissolved in 50 ml of acetonitrile and 50 ml of a solution of 1 g of K2Cr2O7 dissolved in 20 ml of H2O–H2SO4 (9 + 1 v/v).The mixture was vortex mixed and heated at 60 °C for 10 min. After cooling, 100 ml of 10% Na2CO3 and 800 ml of water were added, and the oxidized compounds were extracted with 3 ml of hexane–CH2Cl2 (2 + 1 v/v) with vortex mixing and ultrasonication. Separation and evaporation of the organic layer were performed as above. Kinetics of the new oxidation procedure For all of the corticosteroids listed in Table 4, six aliquots of 250 ml of their stock standard solutions containing 100 mg ml21 were pipetted into glass vials.After evaporation under nitrogen the residues (25 mg of the corticosteroids) were subjected to the fast oxidation procedure described above. Only the time– temperature combination was different: vial 1 of each series was worked up immediately after mixing with the reagent, whereas vials 2–6 were held at 60 °C for 2.5, 5, 7.5, 10 and 15 min, respectively.After work-up and extraction, half of the organic layer was evaporated under nitrogen. The residues were dissolved in 50 ml of acetonitrile with vortex mixing and 10 ml were injected into the HPLC system. Chromatography was carried out under isocratic conditions with a mixture of acetonitrile and an aqueous solution of 5% v/v acetonitrile and 0.01% m/v trifluoroacetic acid. The volume ratios were 20 : 80 for betamethasone, fluorometholone, prednisolone, prednisone and methylprednisolone; 25 : 75 for dexamethasone and 30 : 70 for beclomethasone and flumethasone.The flow rate was held at 1.0 ml min21. Betamethasone administration protocol Two excretion studies were performed on a 5 year old white dairy-cow, first an oral administration of betamethasone and second an intramuscular administration of betamethasone dipropionate. Between both trials a rest period of 40 d was introduced in order to avoid interaction between the two treatments. Oral administration and sample collection.After acclimatization, the cow (520 kg) received betamethasone for five 2410 Analyst, 1998, 123, 2409–2414consecutive days at a dose of 50 mg d21. The 50 mg of betamethasone were dissolved in 20 ml of ethanol and further diluted with 200 ml of water. This solution was directly introduced into the paunch by means of a probe, which was rinsed with 400 ml of water. Urine and faecal samples were collected from 8 h onwards after the first administration until 19 d after the last administration.A blank sample was taken 3 d before the start of the experiment. Samples were stored at 220 °C until analysis. Intramuscular administration and sample collection. A 15 ml volume of an oily suspension containing 50 mg of betamethasone dipropionate was injected intramuscularly into the right cervical muscles of the cow (550 kg) at 9.30 am. Samples of urine and faeces were collected from the first until the 19th day after injection.Triamcinolone acetonide administration protocol A 5 year old black pied cow of about 600 kg received triamcinolone acetonide orally for 7 d, at a daily dose of 50 mg. Samples of faeces were collected from start until 19 d after the end of treatment. Results and discussion Determination of betamethasone As described earlier, betamethasone, differing from dexamethasone only in the b-position instead of the a-position of the 16-methyl group, on oxidation gives rise to the same two reaction products as dexamethasone: 9a-fluoro-16a-methyl- 1,4-androstadiene-3,11,17-trione and its 16b-methyl epimer.1 Both products can easily be separated by GC.The retention time of the a-epimer is lower than that of the b-epimer, and the spectra are also different. Distinction between dexamethasone and betamethasone is possible, however, by the large difference in the ratio of the two oxidation products obtained: the ratio of a/b, in the described procedure,1 is about 5 for dexamethasone and about 0.2 for betamethasone.The ratios obtained are fairly constant, especially within one series. They are, however, influenced by several factors, of which temperature and reaction time seem to be important. This supports the thesis of Her and Watson8 that the two isomeric products are the result of the enolization of the C-17 ketone under oxidation conditions. During the optimization of the oxidation conditions, we observed that an increase in temperature of 10 °C (from 87 to 97 °C) resulted in a decrease in the b/a ratio for betamethasone from 6.6 to 5.9 and of the a/b ratio for dexamethasone from 5.6 to 5.4. Concerning the influence of the reaction time, with an increase from 2 to 4 h the ratio for betamethasone dropped from 6.7 to 5.8 and that for dexamethasone from 5.4 to 5.3.Therefore, it is advisable to include in a series of samples at least one spiked sample of each of the two corticosteroids. In case of doubt, one can always resort to standard additions.Another possibility to obtain complete evidence for the presence of either dexamethasone or betamethasone is the combined application of a generic EIA kit with reaction towards dexamethasone and betamethasone and a specific kit for dexamethasone. The ‘enhanced dexamethasone kit’ from Biognost (Wevelgem, Belgium) has a cross-reactivity for betamethasone of only 1.5%. For dexamethasone a reaction is obtained with both kits, whereas betamethasone only reacts in the first kit.Excretion profile of betamethasone. The results from the excretion study, for betamethasone, obtained by EIA and GCMS, for a cow after oral administration of the drug and intramuscular injection of the dipropionate ester are given in Tables 1 and 2, respectively. The excretion profiles in urine and faeces after oral administration, based on GC-MS measurements, are shown in Fig. 1. The elimination of betamethasone seems to proceed essentially via urine and only to a lesser extent via the faeces.This is different to dexamethasone, for which an earlier study1 indicated that elimination via urine and faeces was comparable. Steady state conditions for betamethasone in urine are reached very fast: the highest concentration was obtained already the third day of the 5 d treatment. The concentrations measured in the faeces gradually increased and reached a maximum at the first day after the end of the treatment, which may indicate an enterohepatic circulation. Resorption of the drug after oral administration seems to be complete, whereas elimination is fast.Also after intramuscular injection of betamethasone dipropionate, the elimination proceeds primarily via the urine (Fig. Table 1 Betamethasone concentrations observed in faeces and urine of a cow treated orally with 50 mg d21 of the drug for 5 d. Times are given according to the the 24 h clock Betamethasone in faeces/ng g21 Betamethasone in urine/ng ml21 Day Treatment with 50 mg betamethasone at Collection of faeces at EIA GC-MS Collection of urine at EIA GC-MS 1 1000 1930 0.7 1.1 2 0930 1600 29.5 35.2 1600 260 270 3 0830 1630 40.5 47.5 1630 434 346 4 0900 5 0900 1600 68.5 55.3 1600 347 346 6 1000 81 57.9 1000 196 192 7 0930 20 23.4 8 0900 6.5 7.3 0900 24 24 9 0900 1.7 3.4 0900 7.1 11 10 0900 < 2 < 2 0900 2.2 3.9 11 0900 < 2 < 2 1130 < 2 < 2 12 0815 < 2 < 2 0900 < 2 < 2 15 0900 < 2 0900 16 0900 0900 17 0900 0900 18 0900 0900 19 0900 0900 22 0900 0900 23 0900 0900 < 2 < 2 Analyst, 1998, 123, 2409–2414 24112).Here the elimination is slower than after oral administration, which also may originate from the application of the ester. Determination of triamcinolone acetonide The sensitive determination via GC-NCI-MS after oxidation is applicable to most of the important representatives of the group of corticosteroids, such as dexamethasone, betamethasone, flumethasone, isoflupredone, prednisolone, prednisone and methylprednisolone.Triamcinolone acetonide, however, which has an acetonide function between C-16 and C-17, is resistant to oxidation with pyridinium chlorochromate. LC-MS has been reported for the confirmation of triamcinolone acetonide.9 We tried to apply GC-NCI-MS after cleavage of the acetonide and subsequent oxidation. Although it has been reported that ethers may be cleaved by heating with concentrated hydrogen iodide and hydrogen bromide, and that hydrochloric acid is only seldom successful,10 we obtained the best results with the latter.After a number of preliminary experiments, the reaction mixture was chosen as 50 ml of acetonitrile, in order to dissolve the residue, and 50 ml of 6 m hydrochloric acid. Temperature–time studies gave the optimum conditions as 85 °C for a reaction time of 30 min. Fig. 3 shows the response found after 30 min at different temperatures.The use of hydrochloric acid allowed us to perform the oxidation step after a simple neutralisation without any further manipulations. In the chromatograms, two peaks with almost the same spectra are obtained, the ratio of which was found to be strongly dependent on the reaction conditions. Under the given condi- Fig. 1 Betamethasone excretion profiles in faeces and urine of a cow treated orally with 50 mg d21 of the drug for 5 d. Fig. 2 Betamethasone excretion profiles in faeces and urine of a cow injected intramuscularly with 50 mg of betamethasone dipropionate on day 1.Table 2 Betamethasone concentrations observed in faeces and urine of a cow treated intramuscularly with betamethasone dipropionate on day 1 at 9.30 am. Times are given according to the 24 h clock Betamethasone in faeces/ng g21 Betamethasone in urine/ng ml21 Day Collection of faeces at EIA GC-MS Collection of urine at EIA GC-MS 2 1100 3.9 5.5 0900 103 67 3 0900 4.3 6.6 0900 19.8 15.2 4 0900 3.1 5.7 0900 32.6 28.8 5 0900 1.7 < 2 0900 5.6 9.7 8 0900 < 2 < 2 0900 < 2 < 2 9 0900 < 2 0900 < 2 10 0900 < 2 0900 < 2 11 0900 < 2 0900 < 2 12 1100 < 2 1430 < 2 16 1100 < 2 1100 < 2 < 2 19 0830 < 2 < 2 0830 < 2 Fig. 3 Cleavage of the acetonide function of triamcinolone acetonide with 6 m HCl for 30 min at different temperatures. Response expressed as percentage relative to the optimum. Fig. 4 Selected ion chromatograms of ions of m/z 296 and 316 under NCI conditions for oxidized triamcinolone acetonide, with full scan spectra of the two observed peaks. 2412 Analyst, 1998, 123, 2409–2414tions the first peak is about 40% of the second. Fig. 4 shows the selected ion chromatogram of the ions of m/z 316 and 296 and the spectra of both observed peaks. The response of the oxidized triamcinolone acetonide obtained in this way is about 10 times lower than that for the same amount of dexamethasone.Excretion profile of triamcinolone acetonide. The administration protocol and GC-MS results for the excretion study of triamcinolone acetonide, given orally to a cow, are given in Table 3. In Fig. 5 the elimination profile of triamcinolone acetonide after oral administration for 7 d is compared with that of dexamethasone under identical conditions.1 Although the concentrations for both drugs during the first few days are more or less comparable, the concentrations of triamcinolone acetonide are generally lower.After treatment, the concentration of triamcinolone acetonide decreases very rapidly and can be hardly followed until 1 week after the end of treatment. Elimination in another form, such as free triamcinolone, may be a possible explanation. Fast oxidation procedure The original oxidation procedure with pyridinium chlorochromate described earlier has been widely used for various matrices.11 One of the more important applications nowadays is the determination of corticosteroids in liver.One of the drawbacks to the application of the procedure to urgent samples from slaughterhouses is the long oxidation time of 3 h. As the oxidation temperature already was high, namely 92°C, we investigated the use of stronger oxidizing agents. One of these, potassium dichromate in acidic conditions, seemed very suitable, and we started to optimize the reaction conditions. Of important help here was that we were able to follow the reaction kinetics by HPLC (Table 4).Therefore, we first optimized the HPLC conditions so that we could follow both the corticosteroid and the oxidation products, which was impossible with GCMS. For the 11-hydroxycorticosteroids the oxidation reaction includes two steps: oxidation of the 17-hydroxy function with cleavage of the C-17–C-20 bond and oxidation of the 11-hydroxy function. These two steps could be followed during reaction: the HPLC traces from the reaction mixtures showed in addition to the corticosteroid and the 3,11,17-trione, also an intermediate compound.This compound was readily obtained by mixing the standard with the reagent at room temperature, while the second step with formation of the 3,11,17-trione proceeded much more slowly. As Her and Watson8 reported that the 17-OH function can be selectively oxidized, without affecting the 11-OH function, with sodium bismuthate under acidic conditions, we first thought that this intermediate was the 17-keto derivative.However, the oxidation of prednisone, the 11-keto analogue of prednisolone, was completely comparable to that of the intermediate of the reaction of prednisolone, proving that not the 17-OH function, but the 11-OH function, was oxidized first. Another indication for this was that the retention times for the glucocorticosteroid and the intermediate 11-keto analogue were very close, which was not expected for the 17-keto analogue, which should be much more apolar owing to the loss of the side chain. The final proof was obtained in an experiment in which the intermediate compound of the oxidation of prednisolone, obtained at room temperature within 1 min, was identified as prednisone by LC-MS.After some preliminary oxidation experiments, the reaction temperature was chosen as 60 °C and sampling was performed immediately after addition of the oxidation reagent and subsequently at 2.5, 5, 7.5, 10 and 15 min.The oxidation was followed for the nine corticosteroids listed in Table 4. The areas, given as a percentage of the maximum area measured of the 3,11,17-trione, of both the intermediate 3,11-dione and the 3,11,17-trione, are given in Table 4. In general, the intermediate was completely oxidized after 10 min, and the maximum Table 3 Triamcinolone acetonide concentrations measured by GC-NCI–MS in faeces of a cow treated orally with 50 mg d21 of the drug for 7 d. Times are given according to the 24 h clock Day Treatment with 50 mg triamcinolone acetonide at Collection of faeces at Triamcinolone acetonide in faeces/ng g21 Corresponding dexamethasone levels/ng g21 1 0800 2000 15.1 2 0800 0800 108 146 3 0800 2000 142 269 4 0800 0800 242 636 5 0800 2000 191 744 6 0800 0800 156 468 7 0800 2000 120 422 8 0800 596 9 0800 80 473 10 0800 231 11 0800 3.93 132 12 0800 18.3 13 0800 1.48 12.5 14 0800 4.89 15 0800 0.43 3.41 16 0800 1.05 17 0800 0.81 18 0800 0.65 19 0800 0.48 Fig. 5 Triamcinolone acetonide (TrmAt) excretion profile in the faeces of a cow treated orally with 50 mg d21 of the drug for 7 d, in comparison with the excretion profile of dexamethasone (Dxm). Analyst, 1998, 123, 2409–2414 2413response of the trione was also reached. The reaction conditions adopted with potassium dichromate under acidic conditions were therefore 60 °C for 10 min. Fig. 6 shows the intermediate 3,11-dione and the 3,11,17-trione from oxidation of prednisolone after an oxidation time of 2.5 min at 60 °C.In this oxidation procedure also, the earlier used tert-butyl methyl ether–dichloromethane (2 + 1 v/v) for extraction of the triones prior to GC-MS analysis was replaced with hexane– dichloromethane (2 + 1 v/v), as the latter was found to be more selective in not dissolving the oxidation reagents. Changing the original oxidation procedure to the fast procedure did not influence the visual aspect of the chromatograms. References 1 D.Courtheyn, J. Vercammen, H. De Brabander, I. Vandenreyt, P. Batjoens, K. Vanoosthuyse and C. Van Peteghem, Analyst, 1994, 119, 2557, and references cited therein. 2 K. De Wasch, H. F. De Brabander, D. Courtheyn and C. Van Peteghem, Analyst, 1998, 123, 2415. 3 M. L. J. Rijckaert, and H. P. J. Vlemmix, The Growth Promoting Effect of Glucocorticosteroids, Department of Chemical Engineering, Eindhoven University of Technology, Eindhoven, 1992. 4 M. J. Groot, P. L. M.Berende, R. Schilt, W. Haasnoot, H. Hooijerink and J. S. Ossenkoppele De Effecten van Lage Doseringen Betaagonisten al of Niet Gecombineerd met Oestradiol, Methylthiouracil en Dexamethason bij Vleeskalveren, Rikilt-DLO Rapport 94.22, Rikilt-DLO, Wageningen, 1994. 5 L. Istasse, V. De Haan, C. Van Eenaeme, B. Buts, P. Baldwin, M. Gielen, D. Demeyer and J. M. Bienfait, J. Anim. Physiol. Anim. Nutr., 1989, 62, 150. 6 N. R. Adams and M. R. Sanders, Aust. Vet. J., 1992, 69, 209. 7 J. Negriolli, PhD Thesis, Faculty of Sciences and Techniques, University of Nantes, 1997. 8 G. R. Her and J. T. Watson, Biomed. Environ. Mass Spectrom., 1986, 13, 57. 9 M. R. Koupai-Abyazani, N. Yu, B. Esaw and B. Laviolette, J. Anal. Toxicol., 1995, 19, 182. 10 J. March, in Advanced Organic Chemistry: Reactions, Mechanisms, and Structure, ed. D. N. Hume, G. Stork, E. L. King, D. R. Herschbach and J. A. Pople, McGraw-Hill, New York, 1968, p. 344. 11 Ph. Delahaut, P. Jacquemin, Y. Colemonts, M. Dubois, J. De Graeve and H. Deluyker, J. Chromatogr. B, 1997, 696, 203. Paper 8/04921A Table 4 Kinetics of the oxidation reactions of some corticosteroids with K2Cr2O7 at 60 °C, expressed as a percentage relative to the maximum response of the 3,11,17-trione Reaction time/min Oxidation of Reaction product 0 2.5 5 7.5 10 15 Beclomethasone 3,11-Dione 236 115 19 4a 8a 9a 3,11,17-Trione 0 85 93 90 100 88 Betamethasone 3,11-Dione 42 60 9 5 0 0 3,11,17-Trione 0 76 99 100 99 98 Dexamethasone 3,11-Dione 26 38 69 37 16 9 3,11,17-Trione 3 25 42 70 93 100 Flumethasone 3,11-Dione 13 79 58 43 32 15 3,11,17-Trione 0 34 45 72 89 100 Fluorometholone 3,11-Dione 35 199 187 170 160 133 3,11,17-Trione 0 23 37 49 67 100 Isoflupredone 3,11-Dione 9 36 22 19 11 4 3,11,17-Trione 0 36 64 76 86 100 Methylprednisolone 3,11-Dione 76 37 33 16 15 4 3,11,17-Trione 13 51 79 84 99 100 Prednisolone 3,11-Dione 29 24 18 10 8 3 3,11,17-Trione 2 40 61 80 100 96 Prednisone 3,11-Dione 40 21 23 10 8 3 3,11,17-Trione 2 41 77 85 100 90 a Disturbed by side products. Fig. 6 HPLC trace of the reaction mixture, with the intermediate 3,11-dione and the 3,11,17-trione of prednisolone, after oxidation with K2Cr2O7 at 60 °C for 2.5 min. 2414 Analyst, 1998, 123, 2409–2414

ISSN:0003-2654    

DOI:10.1039/a804921a

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

Detection of corticosteroids in injection sites and cocktails by MSn† Katia De Wasch,*a Hubert De Brabander,a Dirk Courtheynb and Carlos Van Peteghemc a Faculty of Veterinary Medicine of the University of Ghent, Department of Veterinary Food Inspection, Laboratory of Chemical Analysis, Salisburylaan 133, B-9820 Merelbeke, Belgium. E-mail: katia.dewasch@rug.ac.be b State Laboratory (ROL), Braemkasteelstraat 59, B-9050 Gentbrugge, Belgium c Laboratory of Bromatology, Harelbekestraat 72, B-9000 Ghent, Belgium Received 29th June 1998, Accepted 8th September 1998 In the European Union, the use of growth promoting substances such as thyreostats, anabolics (products with estrogenic, androgenic or gestagenic action) and beta-agonists in animal fattening is forbidden.Corticosteroids, such as dexamethasone, although considered catabolic substances, have been administered to food producing animals in order to achieve mass gains. For the analysis of injection sites and of suspect cocktails (found at the farm), a number of HPTLC and HPLC methods are used.However, in injection sites and also in cocktails found at the farm, sometimes many unknown substances are found. In this investigation, a multiple mass spectrometric (MSn) method was developed. The method is based on rapid extraction of the matrix with methanol and direct infusion of the extract into the interface of the mass spectrometer. Tables that summarise the masses of corticosteroids and their possible esters are presented. 1. Introduction In Europe, public opinion rejects the use of growth promoters in animal fattening. Increasing surveillance by the inspection services (Belgian Veterinary Food Inspection and Ministry of Agriculture) has resulted in a decrease in the use of the drugs for which adequate analytical methods were available. The use of thyreostats, stilbenes, most anabolic steroids and beta-agonists has decreased considerably or even vanished.However, the socalled “hormone Mafia” discovered that the use of corticosteroids (CoST) in animal fattening could lead to a substantial profit. At first sight and/or from a pharmacological point of view, this was a surprise because CoST are catabolic agents and their use in animal fattening is contra-indicated. However, in practice and also in the literature, indications of the growth promoting effect of corticosteroids was found.1–4 Also in sports (e.g. pigeon racing) CoST are abused.5 At first, the most important CoST abused was dexamethasone (Dxm). Later, other substances such as betamethasone (Btm), triamcinolone (Trm) and/or their esters were detected in injection sites, preparations (cocktails) and animal feed.The consumption of these highly contaminated injection sites (mainly in minced meat) can be a considerable threat to human health or interfere with doping control.6 Moreover, the administration of these drugs to animals may result in a decrease in animal welfare.In Belgium, high performance thin layer chromatographic (HPTLC) methods are used for the screening and detection of CoST in injection sites and other matrices containing substantial amounts of CoST.7–9 These methods are adequate for the identification of substantial amounts of target CoST. “Unknown” TLC spots or “unknown” HPLC peaks are sometimes observed when analysing these matrices. These unidentified responses to the standard operating procedure (SOP) may be due to interferences from the matrix but also to “unknown” growth promoting substances.In some cases, these “unknowns” may interfere with “target” components. Therefore, it is obligatory that suspect samples should be confirmed by spectrometric techniques. A gas chromatographic-mass spectrometric method with negative chemical ionisation (GC-NCI-MS) for the detection of Dxm in urine or faeces of treated animals has been described,10 and other GC-MS methods have been published.11–13 They provide good sensitivity in analysing samples with low concentrations of analytes but all require time consuming derivatisations which change the structure and also the molecular mass of the molecule.When trying to identify an unknown molecule, it is easier to have a pseudo-molecular species (e.g. MH+) to start with. LC-MS methods have also been reported.14–16 The relationship between an observed signal in GC-MS or LC-MS and an “unknown“ spot in HPTLC is not always unequivocal (derivatisation in GC-MS may change the molecular mass; a different phase in LC-MS to that in HPTLC may result in a different elution order, irreversible adsorption, etc.).Therefore, we tried to use the power of the recently introduced multiple mass spectrometric (MSn) instruments based on ion trap technology (Finnigan MAT LCQ). Throughout this paper the abbreviation MSn (e.g., MS3) is preferred to MS-MS-MS, etc., because of its simplicity when n is > 2.MSn gives structural information on an underivatised compound by sequential fragmentation. No information was found in the literature on the identification of corticosteroids using MSn. It was found that an extract of an injection site or of an unknown cocktail could be directly infused into the interface of the mass spectrometer. In this way, all the components of the extract are transferred into the mass spectrometer at the same time. In this paper, a rapid detection method following a TLC screening is described.A method was developed for the identification of a number of target and “unknown” CoST by direct MSn analysis. Tables that summarise the mass data for † Presented at the Third International Symposium on Hormone and Veterinary Drug Residue Analysis, Bruges, Belgium, June 2–5, 1998. Analyst, 1998, 123, 2415–2422 2415target and “unknown” CoST are presented. An example of the identification of a “new” corticosteroid in animal fattening, clobetasol propionate, is presented. 2. Experimental 2.1. Apparatus The following apparatus was used: homogeniser (e.g., Waring blender with 250 ml reservoir, Ultra-Turrax), Stomacher 400 Lab Blender (Seward Medical, London, UK), microwave oven, centrifuge, rotary vacuum evaporator, water-bath, extraction flasks (250 and 500 ml), vacuum manifold (e.g., Sample Preparation Unit, Analytichem International, Harbor City, CA, USA), nitrogen evaporator (e.g., Techni Dry Block) or other types of evaporators (e.g., Speedvac SVC 200, SC 210A, Savant Instruments, Farmingdale, NY, USA; Howe Gyrovap, VA Howe & Co.Ltd., Banbury, Oxon, UK), chromatographic columns and tanks. The mass spectrometer used was an LCQ Ion Trap Mass Analyzer (Finnigan MAT, San Jose, CA, USA) with an electrospray interface. 2.2. Reagents and reference components Most reference CoST were obtained from Steraloids (St) (Wilton, NY, USA) or Sigma (Si) (St-Louis, MO, USA). Other CoST were gifts from various sources. All recent standards were obtained through the Belgian NRL [National Reference Laboratory, WIV-LP (formerly IHE), Brussels, Belgium] to ensure that all the field laboratories used the same standards.17 The most important CoST and their common abbreviations in Belgium (cited in order of increasing molecular mass) used in this investigation are as follows: prednisone (Pron, St P300) (17,21-dihydroxypregna-1,4-diene-3,11,20-trione); predniso- Fig. 1 Full MS1 spectrum of a mixture of eight CoST. Fig. 2 Full MS2, MS3 and MS4 spectra of dexamethasone. Table 1 Diagnostic ions (m/z) for the identification of target CoST Analyte MH+ Diagnostic ions in MS2 Btm (betamethasone) 393 373, 355, 337 Dxm (dexamethasone) 393 373, 355, 337 FlM (flumethasone) 411 391, 371, 335 FML (fluorometholone) 377 357, 339, 321 Bcm-DP (beclomethasone dipropionate) 521 503, 411, 319 Clol-P (clobetasolpropionaat) 467 447, 373, 355 Dom (desoximetasone) (internal standard) 377 357, 339, 321 2416 Analyst, 1998, 123, 2415–2422lone (Prolon, Si P6004) (11b,17,21-trihydroxypregna- 1,4-diene-3,20-dione); cortisone (Cron, St Q2500) (17,21-dihydroxypregn- 4-ene-3,11,20-trione); cortisol (Crol, St Q3880) (11b,17,21-trihydroxypregn-4-ene-3,20-dione); methylprednisolone (MProlon, Si M0639) (11b,17,21-Trihydroxy-6amethylpregna- 1,4-diene-3,20-dione); fluorometholone (FML, F9381) (9-fluoro-11b,17-dihydroxy-6a-methylpregna- 1,4-diene-3,20-dione) ; dexamethasone (Dxm; Si D1756) (9-fluoro-11b,17,21-trihydroxy-16a-methylpregna-1,4-diene- 3,20-dione); betamethasone (Btm, Si B7005) (9-fluoro- 11b,17,21-trihydroxy-16b-methylpregna-1,4-diene- 3,20-dione); triamcinolone (Trm, Si T6376) (9-fluoro-11b,16a,17,21-tetrahydroxypregna-1,4-diene- 3,20-dione); beclomethasone (Bcm, Si B0385) (9-chloro- 11b,17,21-trihydroxy-16b-methylpregna-1,4-diene- 3,20-dione); flumethasone (Flm, Si F9507) (6a,9-difluoro-11b,17,21-trihydroxy-16a-methylpregna- 1,4-diene-3,20-dione); clobetasol (Clol) (21-chloro-9-fluoro- 11b,17-dihydroxy-16b-methylpregna-1,4-diene-3,20-dione); and the propionate esters of Btm, Bcm and Clol.Clol itself is not commercially available. The internal standard was desoximetasone (Dom, Si D6038) (9-fluoro-11b,21-dihydroxy-16a-methylpregna-1,4-diene- 3,20-dione). All solvents were of analytical-reagent grade from Merck (Darmstadt, Germany). 2.3. Solutions Stock standard solutions of CoST were prepared at 200 ng ml21 in ethanol. Tenfold dilutions of these solutions gave working standard solutions of CoST at a concentration of 20 ng ml21.Table 2 List 01 Molecular mass CoST 358.2 360.2 360.2 362.2 374.2 376.2 392.2 394.2 408.2 Ester MM MM + Pron Prolon Cron Crol MProlon FML Dxm Trm Bcm Acetonide 40.0 398.2 400.2 400.2 402.3 414.3 416.2 432.2 434.2 448.2 Acetate 60.1 42.0 400.2 402.2 402.2 404.2 416.2 418.2 434.2 436.2 450.2 Propionate 74.1 56.1 414.2 416.3 416.3 418.3 430.3 432.3 448.3 450.2 464.2 Butyrate 88.1 70.1 428.3 430.3 430.3 432.3 444.3 446.3 462.3 464.3 478.3 Diacetate 102.1 84.0 442.2 444.2 444.2 446.3 458.3 460.2 476.2 478.2 492.2 Valerate 102.1 84.1 442.3 444.3 444.3 446.3 458.3 460.3 476.3 478.3 492.3 Pivalate 102.1 84.1 442.3 444.3 444.3 446.3 458.3 460.3 476.3 478.3 492.3 Caproate 116.2 98.1 456.3 458.3 458.3 460.4 472.4 474.3 490.3 492.3 506.3 Benzoate 122.1 104.1 462.3 464.3 464.3 466.3 478.3 480.3 496.3 498.3 512.3 Dipropionate 130.1 112.1 470.3 472.3 472.3 474.3 486.3 488.3 504.3 506.3 520.3 Enanthate 130.2 112.2 470.3 472.4 472.4 474.4 486.4 488.4 504.4 506.3 520.3 Phosphate (di-Na) 124.0 482.1 484.2 484.2 486.2 498.2 500.2 516.2 518.1 532.1 Cypionate 142.2 124.1 482.3 484.3 484.3 486.3 498.3 500.3 516.3 518.3 532.3 Caprylate 144.2 126.2 484.4 486.4 486.4 488.4 500.4 502.4 518.4 520.4 534.4 Phenylpropionate 150.2 132.2 490.3 492.4 492.4 494.4 506.4 508.4 524.4 526.3 540.3 Nonanoate 158.2 140.2 498.4 500.4 500.4 502.4 514.4 516.4 532.4 534.4 548.4 Tosylate 172.2 154.2 512.4 514.4 514.4 516.4 528.4 530.4 546.4 548.4 562.4 Decanoate 172.3 154.2 512.4 514.4 514.4 516.5 528.5 530.4 546.4 548.4 562.4 Divalerate 186.2 168.2 526.4 528.4 528.4 530.4 542.4 544.4 560.4 562.4 576.4 Undecylate 186.3 168.3 526.4 528.4 528.4 530.5 542.5 544.5 560.5 562.4 576.4 Laurate 200.3 182.3 540.5 542.5 542.5 544.5 556.5 558.5 574.5 576.5 590.5 Fig. 3 Overview of the fragmentation of list 01 components.Fig. 4 MS1 spectrum during infusion of injection site with unknown component. Analyst, 1998, 123, 2415–2422 24172.4. HPTLC screening of corticosteroids The injection site is sampled by cutting at least 2 g suspect material with a bistoury and transferring it into a double bag of a stomacher. The number of the sample is marked on the outer bag. Methanol is added at a ratio of 2.5 ml g21 material (with a minimum of 5 ml). Internal standard desoximetasone (Dom) (15 ml = 3000 ng per 2.5 ml of methanol or 3 mg kg21) is added.For the destruction of the matrix, a stomacher is used for at least 1 min. Overnight extraction allows the matrix to react longer with the extraction solvent and therefore a higher extraction yield is obtained. After extraction overnight, the mixture is filtered into a disposable plastic recipient. This primary extract is prepared in a laboratory room separated from the laboratory for residue analysis to avoid contamination problems.A primary extract of a suspect cocktail is prepared by mixing or extraction with methanol in an analogous way. If necessary, an aliquot (90%) of this extract may be concentrated: 4.5 ml of extract are evaporated to dryness and the residue is dissolved in 0.8 ml of methanol. Subsequently 1.2 ml of water is added and the mixture is placed in an ultrasonic bath until the solution is clear. A C18 cartridge (3 ml, 500 mg) is conditioned with 2 3 2 ml of methanol followed by 2 3 2 ml of water.The extract is transferred into the cartridge and the CoST are eluted with methanol–water (70 + 30 v/v) (2 ml) followed by methanol (2 ml). This eluate is evaporated to dryness and the residue is dissolved in 50 ml of ethanol. The concentrated extracts (10 ml) are spotted on an NH2-F254 plate together with standard mixtures of the target CoST (600 ng). The plate is developed with ethanol–chloroform–ethyl acetate (10 + 20 + 20 v/v). If this plate shows a response for esters (at the front), another plate is developed with hexane– acetone (65 + 35 v/v).After drying the plates, compounds are revealed by heating the plates at 200 ± 10 °C for 3 min. The plates are inspected under UV and visible light. At 254 nm the CoST are visible as blue–purple spots on a white background. At 366 nm the spots are beige or blue on a blue–purple background. If suspect spots are present, the extract is transferred for MSn analysis. Other procedures for the TLC analysis of CoST have been described earlier8,9,17 but their compatibility with MSn was not tested.Fig. 5 A, MS2, B, MS3 and C, MS4 spectra of the unknown component in the injection site. 2418 Analyst, 1998, 123, 2415–24222.5. MSn apparatus and conditions For MSn experiments, an LCQ ion trap mass spectrometer was used with a built-in syringe pump. The analytes were ionised through the electrospray interface producing MH+ or M2H2 ions. Infusion. The remainder of the concentrated extract used for HPTLC was directly infused into the mass spectrometer.Infusion into the mass spectrometer was performed as follows: the incoming flow of sample (5 ml min21) was mixed with an eluent flow of, methanol–1% acetic acid (60 + 40 v/v) at 0.3 ml min21 through a T-piece. In order to exclude contamination, the spray shield, heated capillary and infusion line were thoroughly cleaned with methanol before starting the infusion and acquisition, and between samples.Eluent was infused into the mass spectrometer through the infusion line and checked for the presence of known parent (MH+) and daughter ions, to ensure that no contamination or residues of previous standards or positive samples were present. If this check proved to be negative, a new sample was infused. Tuning. In theory, a tune file can be made for each compound individually. Since during one acquisition many different compounds are investigated, and since in practice there is only a slight difference in parameter settings to obtain optimum conditions for compounds with a small molecular mass difference and a similar structure, Dxm was used for tuning purposes.An amount (40 ng ml21) which produces a fairly readily distinguished pseudo-molecular ion (MH+) was directly infused into the mass spectrometer and the different MS parameters (capillary voltage, tube lens offset, ESI voltage, etc.) were optimised and saved in a tune file.This tune file was used during the subsequent investigation. Interpretation. During infusion, the spectrum was searched for MH+ ions in the range 350–510 mu which rise above the normal background originating from the infusion liquid and the electronics. For first line quality control, the MH+ ion of Dom (377u) should be present, otherwise sampling and the extraction procedure should be repeated. A “suspect” ion can be defined as an ion that rises above the background noise and that needs further fragmentation to match the spectrum with a standard mass spectrum.If a suspect ion does not match any known standard compound, it becomes an unknown. If such an ion was observed the MS2 and MS3 spectra of the suspect peak were acquired. For the target components the diagnostic ions are given in Table 1. The sample is considered to be positive when the MS2 spectrum matches that of the previously infused standard. If, next to or instead of target component ions, other ions are observed, an attempt at the identification of these components is made using the so-called list 01 and 02 (see Results and discussion). 3. Results and discussion 3.1. MSn of corticosteroid standards Mixtures of standards of CoST are infused into the mass spectrometer. In Fig. 1 a full MS1 spectrum of a mixture of eight CoST is shown. Very abundant MH+ ions for all CoST infused are found. This is normal because electrospray is a soft ionisation technique. The relative energy of collision applied to pseudo-molecular ions is chosen so the intensity of the most intense daughter ion is maximum.Further fragmentation is performed on the MH+ for MS2 and for MSn, as a general rule, on the most abundant daughter ion. During one acquisition, MS1 up to MS3 or MS4 spectra are acquired. In Fig. 2, as an example, the MS2, MS3 and MS4 spectra for dexamethasone are given. In Fig. 3 an overview of the fragmentation of the most important CoST (so-called list 01 components; see below) is given.The relative values of the collision energy are comparable for related compounds. The collision energy is reported on a relative scale (percentage) and no correlation is given with an absolute voltage. As the percentage is increased, more fragment ions will be formed. For this application we preferred to apply energy to the MH+ ion until it disappeared and only daughter ions were present. For further fragmentation of the most intense daughter ions, the same rule was applied.A collision energy of about 20% will be sufficient to generate a good response of fragment ions. 3.2. MSn of corticosteroids in injection sites It was found that the extract of an injection site (or an unknown cocktail) could be directly infused into the interface of the mass spectrometer. When a fairly high concentration of analyte(s) is present (which is mostly the case in a “positive” injection site), the MH+ ions will rise above the background ions.If the concentration is lower, the pseudo-molecular ions will disappear into the background. For the identification of CoST, the following three stage strategy is used. In addition to the target components (stage 1) a so called “list 01” is programmed in Microsoft Excel (stage 2). In this list 12 important “known” CoST are listed in columns from left to right in order of increasing molecular mass. In this list 01, Dxm, Btm and Pam (paramethasone) are in the same column and thus indistinguishable.In the rows, 21 acids (possibly used for esterification of CoST) are listed also in order of increasing molecular mass. The combination of columns and rows yields 212 possible esters of the CoST. For the mass of the different esters the loss of water (218) upon formation of the ester is taken into account. In Table 2 the list 01 is given. The Table 3 Part of list 02: some corticosteroid esters matching the MH+ of the unknown (2361, clobetasol; 2368, clocortolone; 3126, diflorasone) Molecular mass 410.17 410.17 410.19 Ester MM MM+ 2361 2368 3126 Acetonide 40.04 450.21 450.21 450.23 Acetate 60.05 42.03 452.20 452.20 452.22 Propionate 74.08 56.06 466.23 466.23 466.25 Isobutyrate 88.10 70.08 480.25 480.25 480.27 Butyrate 88.10 70.08 480.25 480.25 480.27 Diacetate 102.06 84.04 494.21 494.21 494.23 Valerate 102.13 84.11 494.28 494.28 494.30 Pivalate 102.13 84.11 494.28 494.28 494.30 Caproate 116.16 98.14 508.31 508.31 508.33 Benzoate 122.12 104.10 514.27 514.27 514.29 Dipropionate 130.12 112.10 522.27 522.27 522.29 Enanthate 130.18 112.16 522.33 522.33 522.35 Phosphate (di-Na) 123.96 534.13 534.13 534.15 Cypionate 142.15 124.13 534.30 534.30 534.32 Caprylate 144.21 126.19 536.36 536.36 536.38 Phenylpropionate 150.18 132.16 542.33 542.33 542.35 Nonanoate 158.23 140.21 550.38 550.38 550.40 Tosylate 172.20 154.18 564.35 564.35 564.37 Decanoate 172.26 154.24 564.41 564.41 564.43 Divalerate 186.22 168.20 578.37 578.37 578.39 Undecylate 186.27 168.25 578.42 578.42 578.44 Laurate 200.31 182.29 592.46 592.46 592.48 Analyst, 1998, 123, 2415–2422 2419masses of the esters increase from left to right and from top to bottom, making the search for a certain mass easy.If, in an MS1 spectrum of a sample an abundant “non-target” MH+ ion is observed, list 01 is searched for a match with the corresponding molecular mass (stage 2). If a match is found, a search is made to establish if the component is available in one of our laboratories.If so, MSn spectra of both sample and standard are acquired and compared. 3.3. Validation of the procedure In the list 01, only Dxm, Btm, Flm, Fml, Bcm-DP and Clol-P (see section 3.5) are validated because they are the target components in Belgium. Validation was carried out by fortifying blank tissue samples resembling the structure of an injection site with known amounts of CoST at the level of 2 ppm and carrying out the procedure described several times ( > 20).Fig. 6 A, MS1, B, MS2, C, MS3 and D, MS4 spectra of clobetasol propionate. 2420 Analyst, 1998, 123, 2415–2422The number of times that the CoST added are detected is statistically evaluated. To our clients (in this case the inspection services) it can be guaranteed that target CoST present in injection sites at the level of 2 ppm will be detected by the laboratory with a probability (frequency) of > 95%. 3.4. MSn identification of a “new” corticosteroid In an injection site, a suspect HPTLC spot at the correct RF value of beclomethasone dipropionate (Bcm-DP) and/or betamethasone dipropionate (Btm-DP) was observed (one-dimensional HPTLC). Using co-chromatography a new, not completely separated, TLC spot occurs. The suspect spot also has a slightly different colour than the Btm and Bcm esters. According to the quality criteria described in EC 93/256,18,19 the sample was considered to be negative ( = analyte absent or lower than the action limit).Further investigations with MSn were carried out. The extract of the injection site was infused in the LCQ, next to Bcm-DP and Btm-DP. The spectrum of the unknown showed an intense ion at m/z 466.9 (most probably MH+ ) (Fig. 4). This ion is clearly different from the MH+ ion observed for the standards of Bcm-DP and Btm-DP (m/z 521.8 and 505.7 respectively; see list 01) and is also different from the target components.In list 01 only one match with cortisol benzoate was found. However, this molecule does not correspond with the other characteristics of the MS1 spectrum. For the identification of this “unknown”, the following points were taken into account: the spot migrates a long distance and has a similar RF value to esters of CoST. A CoST ester is a possibility. The spot is also present in the 2D-HPTLC of anabolics. The presence of two isotope peaks with an m/z difference of 2 and a ratio of 3:1 indicate that the analyte contains one chlorine atom.The loss of 20 u in MS2 indicates a loss of HF. This corresponds with the findings for other CoST standards containing fluorine (Fig 3). The loss of 74 u indicated the presence of propionic acid. Both Bcm-DP and Btm-DP lose the 74 u fragment twice. The fragmention of the molecule is shown in Fig. 5. One of our laboratories has observed this spot also more than once in illegal cocktails in addition to other target components. In Microsoft Excel, a combination of all CoST in the Merck Index (except those in list 01) with all the acids already used in list 01 was made (stage 3).The CoST are indicated by their Merck Index number and listed in order of increasing molecular mass. This so-called list 02 contains 20 CoST combined with 22 esters, yielding 440 possible CoST esters. In this list 02, three matches with the molecular mass of the unknown are found (Table 3): the monopropionate esters of 2361 (clobetasol), 2368 (clocortolone) and 3126 (diflorasone). Diflorasone is eliminated because the molecule does not contain chlorine.In the Sigma catalogue, one of these products was found and ordered: clobetasol propionate. A solution of the standard was infused and it was found that the MSn spectra matched the spectra of the unknown (Fig. 6). Only in the MS1 spectrum were substantial differences in the low mass region observed (Fig. 4). These were most probably formed by co-extracted components from the matrix.Based on the so-called “intellectual owner’s right” (a Belgian tradition which states that the discoverer of a “new” component may propose an abbreviation), the abbreviation ClolP was proposed for clobetasol propionate. The samples and the analytical data were transferred to the Belgian NRL and the Community Reference Laboratory (CRL), the RIVM at Bilthoven. The formula of clobetasol propionate and possible fragmentations is shown in Fig. 7. The base peak in the MS2 spectrum (m/z 447) is formed by the loss of HF (220 u) (like all other fluorine-containing CoST). In the MS3 spectrum the loss of propionic acid (274 u) is predominant (base peak at m/z 373). The MS4 spectrum shows several losses of water (218 u) with the formation of ions at m/z 355, 337, 319 and 301. By increasing the isolation width of the parent ion and the most intense daughter ions, ions containing the chlorine isotopes are also isolated.During further fragmentation the isotope peaks remain present in the spectrum. This means that chlorine is not split off. 4. Conclusion A possible strategy for the control of the abuse of CoST in cattle fattening through the analysis of injection sites and illegal cocktails has been described. The presence or absence of CoST is screened by HPTLC, which is a very fast and robust analytical technique. The results for negative samples may be produced very quickly (e.g., within 2 d).If, on the TLC trace, suspect spots are present, the extract is subjected to the powerful three stage MSn identification procedure. This confirmation step by direct infusion is also very fast because no chromatographic run is needed. Moreover, with this technique not only the presence of target (stage 1) or “known” (stage 2) CoST may be confirmed, but also “unknown” components (stage 3) may be identified. An example of such an identification is that of clobetasol propionate. However, this technique also has its limitations: the differentiation power between isomers such as betamethasone and dexamethasone is low.For such differentiation, other techniques have to be used. Also, true unknowns, not belonging to any group of analytes with which our laboratory is used to dealing, can be present in the sample. If this unknown is seen fairly often when performing routine screening tests and our confirmation technique is unable to make a valid identification, other techniques such as NMR spectroscopy are advisable options.Acknowledgement The authors are indebted to W. De Rycke for skilful operation of the LCQ. References 1 N. R. Adams and M. R. Sanders, Aust. Vet. J., 1992, 69, 209. 2 W. R. Dayton and M. R. Hathaway, in Growth Regulation in Farm Animals—Advances in Meat Research, vol. 7, ed. A. M. Pearson and T. R. Dutson Elsevier Applied Science, London, 1991, pp. 17–45. 3 L. Istasse, V. de Haan, C. Van Eenaeme, B.Buts, P. Baldwin, M. Gielen, D. Demeyer and J. M. Bienfait, J. Anim. Physiol. Anim. Nutr., 1989, 62 150. 4 P. M. Keen, Vet. Ann., 1987, 27, 45. 5 G. Dhondt, R. Ectors, L. Mathys and R. Ducatelle, Vlaams Diergeneesk. Tijdschr., 1993, 62, 35. 6 G. De Bruykere and C. Van Peteghem, Anal. Chim. Acta, 1992, 275, 49. 7 D. Courtheyn, N. Verheye V. Bakeroot, V. Dal, R. Schilt, H. Hooijerink, E. O. Van Bennekom, W. Haasnoot, P. Stouten and F. A. Huf, in Proceedings of Euroresidue II, Veldhoven, 1993, ed. N. Fig. 7 Clobetasol propionate and possible fragmentation. Analyst, 1998, 123, 2415–2422 2421Haagsma, A. Ruiter and P. B. Czedik-Eysenberg, 1993, pp. 251- 256. 8 K. Vanoosthuyze, L. S. G. Van Poucke, A. C. A. Deloof and C. H. Van Peteghem, Anal. Chim. Acta, 1993, 275, 177. 9 J. Hoebus, E. Daneels, E. Roets and J. Hoogmartens, J. Planar Chromatogr., 1993, 6, 269. 10 D. Courtheyn, J. Vercammen, H. De Brabander, I. Vandereyt, P. Batjoens, K. Vanoosthuyze and C. Van Peteghem, Analyst, 1994, 119, 2557. 11 D. G. Watson, J. M. Midgley and C. N. J. McGhee, Rapid Commun. Mass Spectrom., 1989, 1, 8. 12 J. M. Midgley, D. G. Watson, T. Healy and M. Noble, J. Pharm. Pharmacol., 1987, 39 (Suppl.), 51. 13 E. Houghton, P. Teale and M. C. Dumasia, Analyst, 1984, 109, 273. 14 Y. Kim, T. Kim and W. Lee, Rapid Commun. Mass Spectrom., 1997, 11, 863. 15 S. Rizea-Savu, L. Silvestro, A. Haag and F. Soergel, J. Mass Spectrom., 1996, 31, 1351. 16 S. J. Park, Y. J. Kim, H. S. Pyo and J. Park, J. Anal. Toxicol., 1990, 14, 102. 17 F. Smets, G. Pottie, H. F. De Brabander, P. Batjoens, L. Hendriks, D. Courtheyn, B. Lancival and Ph. Delahaut, in Analyst, 1994, 119, 2571. 18 EEC (1993) document 93/256/EEC, Official J. Eur. Commun., 1993, No. L 118/64. 19 R. W. Stephany and L. G. Van Ginkel, Fresenius’ J. Anal. Chem., 1990, 338, 370. Paper 8/04932G 2422 Analyst, 1998, 123, 2415–2422

ISSN:0003-2654    

DOI:10.1039/a804932g

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

New database on hormone and veterinary drug residue determination in animal products† Nico J. Van Eeckhout,*a Carlos H. Van Peteghem,a Vincent C. Helbo,b Guy C. Maghuin-Rogisterb and Marc R. Cornelisc a Laboratory of Food Analysis, Faculty of Pharmaceutical Sciences, University of Ghent, Harelbekestraat 72, 9000 Ghent, Belgium. E-mail: nico.vaneeckhout@rug.ac.be b Laboratory of Analysis of Foodstuffs of Animal Origin, Faculty of Veterinary Science, Department of Food Sciences, University of Liège, Boulevard de Colonster 20, Bât.B43 Bis, Sart-Tilman, 4000 Liège, Belgium c Institute for Veterinary Inspection, Wetstraat 56, 1040 Brussels, Belgium, Received 29th June 1998, Accepted 24th August 1998 A new database was created which provides a carefully judged inventory of analytical methods available for the determination of residues of growth promoters (steroidal anabolic hormones, b-agonists and glucocorticoids) and veterinary drugs (antibiotics and growth inhibitors), which are or will be regulated by EU legal acts.This database is available on the Internet at http://cemu10.fmv.ulg.ac.be/OSTC. Introduction The European Union requires its Member States to monitor certain substances and residues thereof in live animals and animal products (Directive 96/23/EC).1 Two groups of substances are included in this Directive. Group A involves substances having a hormonal or thyrostatic action and bagonists applied in stockfarming (Directive 96/22/EC)2 and veterinary drugs that have now been banned [included in Annex IV of Council Regulation (EEC) No. 2377/90].3 Group B involves other veterinary drugs and a list of contaminants. Details of the substances belonging to groups A and B are presented in Table 1. One of the most important principles laid down by these Directives is that foodstuffs obtained from treated animals must not contain residues which might constitute a health hazard for the consumer, in order to protect the health of the consumer.To facilitate the uniform application of this principle throughout the Community, and in order to ensure that differences in the assessment of the effects of residues by Member States do not create barriers to the free movement of veterinary medicinal products or to the movement of foodstuffs of animal origin, the Council adopted on 26 June 1990 Regulation (EEC) No. 2377/903 laying down a Community procedure for the establishment of maximum residue limits (MRLs) for veterinary medicinal products in foodstuffs of animal origin.The aim of this project was to establish, in the framework of the Directives and Regulation cited above, an inventory of analytical methods for residues of specific classes of growth promoters (sex hormones, b-agonists and glucocorticoids) and veterinary drugs (antibiotics) which could be standardized after validation. This inventory is organized in the form of a database which is accessible on the Internet at http://cemu10.fmv.ulg.ac.be/OSTC.The database is subdivided into five databases which contain information about the following topics: 1. The Methods database contains a compilation of analytical methods for the screening and confirmation of residues of growth promoters and veterinary drugs. Furthermore, the methods were critically evaluated according to the criteria described in Commission Decisions 93/256/EEC4 and 93/257/EEC.5 2. In the Molecules database, general information about residues found in foodstuffs of animal origin is presented.Aspects such as physical and chemical data, available reference materials and specific reagents and toxicological data concerning consumers health are discussed. 3. An inventory of European and Belgian legislation with regard to residues in animal products is included in the Legislation database. † Presented at the Third International Symposium on Hormone and Veterinary Drug Residue Analysis, Bruges, Belgium, June 2–5, 1998.Table 1 Survey of the compounds included in Groups A and B of Directive 96/23/EC1 Group A: Substances having an anabolic effect and unauthorized substances (1) Stilbenes, stilbene derivatives and their esters and salts (2) Antithyroid agents (3) Steroids (4) Resorcyclic acid lactones including zeranol (5) b-Agonists (6) Compounds included in Annex IV of Council Regulation (EEC) No. 2377/19903 Group B: Veterinary drugsa and contaminants (1) Antibacterial substances, including sulfonamides and quinolones (2) Other veterinary drugs— (a) Anthelmintics (b) Anticoccidials including nitroimidazoles (c) Carbamates and pyrethroids (d) Sedatives (e) Non-steroidal anti-inflammatory drugs (NSAIDs) (f) Other pharmacologically active substances (3) Other substances and environmental contaminants— (a) Organochlorine compounds including PCBs (b) Organophosphorus compounds (c) Chemical elements (d) Mycotoxins (e) Dyes (f) Others a Including unlicensed substances which could be used for veterinary purposes.Analyst, 1998, 123, 2423–2427 24234. Qualitative and quantitative data about the monitoring programs by the Belgian Institute for Veterinary Inspection and the Belgian Ministry of Agriculture are included in the Monitoring Programs database. 5. An estimation of the consumption of animal products in Belgium for the last 40 years is presented in the Food Consumption database. Methods database When putting a monitoring program into practice, two analytical steps are generally performed, namely a screening phase and a confirmation phase.In the first phase, a large scale screening is put into place, allowing the analysis of a large number of samples using a cheap, simple and rapid method. Immunoassays (radio- and enzyme immunoassays) are well suited to this purpose. Screening methods yield negative results, which are generally accepted, or positive results, which must be confirmed in a second phase to distinguish between true positive results and false positive results.Analytical methods based on chromatography, most often coupled to spectrometry (UV, FT-IR, MS) are methods of choice for the second phase (confirmatory analysis). Despite the fact that many methods for both screening and confirmation have been published, very few, if any, have been tested on an interlaboratory scale. Until a few years ago, an analytical method was considered suitable only after having been tested successfully in a collaborative study involving a circular analysis of the same sample by several laboratories.This view has now changed owing to the high cost of such studies, the long time needed to perform the tests, the rapid progress made in the development of analytical methods and the large number of residues for which analytical methods are needed. Within the European Union, criteria have been fixed that must be fulfilled (Commission Decisions 93/256/EEC4 and 93/257/EEC).5 These criteria are very helpful for verifying the sound basis of a method and provide guidelines for the analyst who develops methods other than reference methods.The approach by these criteria does not mean that interlaboratory testing is no longer needed. Circular analysis is the ultimate test to demonstrate the quality of a method and it is a very efficient tool in the training of laboratories and the harmonization of the quality of their work.Analytical methods for the screening and confirmation of residues were obtained through a systematic search of the literature and completed with data obtained from the proceedings of scientific meetings dedicated to the subject: Fig. 1 Example of the information provided for an analytical method which is included in the database. 2424 Analyst, 1998, 123, 2423–24274 International Symposium on the Analysis of Anabolizing and Doping Agents in Biosamples (Ghent, 1988 and 1990); 4 International Symposium on Hormone and Veterinary Drug Residue Analysis (Ghent, 1992; Bruges, 1994); 4 Euroresidue Conference on Residues of Veterinary Drugs in Food (The Netherlands, 1990, 1993 and 1996); 4 EC Flair Workshops, Concerted Action No. 8 (Swansea, 1991; Ghent; 1992; Thessaloniki, 1992; Liège, 1993; Nantes, 1993; Porto, 1994); 4 Scientific Conference on Growth Promotion in Meat Production (Brussels, 1995); 4 Reference manual edited by Heitzman,6 1994.The analytical methods which are included in the database were critically evaluated according to the criteria described in Commission Decisions 93/256/EEC4 and 93/257/EEC.5 These Commission Decisions prescribe definitions to be met for screening tests and confirmation analyses, analytical methods which may be used and detailed criteria which should be met by the techniques. Aspects such as specificity, limits of detection and determination, accuracy and other quality criteria are discussed.After evaluation the methods were classified in two categories: 4 A: high reliability: method is complete in accordance with the prescribed criteria; 4 B: limited reliability: method is partially in accordance with prescribed criteria. For the analytical methods included in the database, the following information is provided: 4 bibliographic references (authors, journal, year of publication, address of the first author, etc.) 4 sample type (e.g., plasma, urine, kidney, muscle) used to evaluate the method; 4 residues which can be detected and/or confirmed by the method; 4 sample preparation (extraction type, clean-up, fractionation, etc.); 4 method characteristics; 4 limit of detection and/or limit of quantification; 4 grade A or B. The way in which an analytical method is described in the database is presented in Fig. 1. Fig. 2 Example of the information provided by the discussion of compounds included in the database. Analyst, 1998, 123, 2423–2427 2425Molecules database General information about the residues found in food products of animal origin is included in the Molecules database.For all the compounds discussed, the following information is provided: therapeutic category, Chemical Abstracts Service (CAS) name and synonyms, CAS registry number, number of references for analytical methods included in the methods database for the determination of the compound, reference to the legislation which applies to the compound, molecular formula, molecular mass, structure, derivatives and toxicological data.A compilation of physical data (e.g., solubility, optical rotation, molar absorptivity, melting point) is included for the compound, where available. In Fig. 2 this is demonstrated for sulfamethazine. Legislation database In the Legislation database, an inventory of European and Belgian legislation with regard to residues found in foodstuffs is presented. The following legislations are included: 4 Council Directive 96/23/EC1 of 29 April 1996 on measures to monitor certain substances and residues thereof in live animals and animal products; 4 Council Directive 96/22/EC2 of 29 April 1996 concerning the prohibition of the use in stockfarming of certain substances having a hormonal or thyrostatic action and of bagonists; 4 Council Regulation (EEC) No. 2377/903 of 26 June 1990 laying down a Community procedure for the establishment of maximum residue limits of veterinary medicinal products in foodstuffs of animal origin; 4 Commission Decision 93/256/EEC4 of 14 April 1993 laying down the methods to be used for detecting residues of substances having a hormonal or a thyrostatic action; 4 Commission Decision 93/257/EEC5 of 15 April 1993 laying down the reference methods and the list of national reference laboratories for detecting residues; 4 Commission Decision 98/179/EC7 of 23 February 1998 laying down detailed rules on official sampling for the monitoring of certain substances and residues thereof in live animals and animal products; 4 Council Directive 70/524/EEC8 of 23 November 1970 concerning additives in feeding-stuffs. For controlling the application of these Regulations and Directives and of the corresponding Belgian legislation, a progam for the control of residues, concerning mainly substances illegally used as growth promoters and veterinary drugs, including unauthorized substances which could be used as veterinary drugs, must be put into place every year and executed by Member States of the EU after approval by the Commission.Monitoring programs database Control of residues, particularly growth promoters, in animal products is performed by two governmental authorities in Belgium: 4 Department of Public Health: the Institute for Veterinary Inspection is responsible for collecting samples at slaughter- Fig. 2 continued 2426 Analyst, 1998, 123, 2423–2427houses and occasionally on farms, and the Inspection Service for Foodstuffs is responsible for sampling at the retail level; 4 Department of Agriculture in Farms, Veterinary Inspection and Inspection of Raw Materials.Analyses are performed by approved laboratories. The hierarchy of laboratories in charge of residue control in the European Union is well defined (Directive 96/23/EC).1 There are four Community Reference Laboratories (CRLs) that are each responsible for certain categories of substances.These CRLs are in charge of controlling and training the National Reference Laboratories (NRLs). In Belgium, the Scientific Institute of Public Health Louis Pasteur is responsible for the analysis of all types of residues in live animals and animal products. Other Belgian laboratories (about six) are involved in the routine analysis of residues, especially of growth promoters. They are supervised by the NRLs. The results of the monitoring programs conducted by the Institute for Veterinary Inspection in Belgium for 1994, 1995 and 1996 and by the Belgian Ministry of Agriculture for 1996 are presented in the Monitoring programs database.Food consumption database A survey of the Belgian consumption of meat (kg per inhabitant per year, expressed as carcass mass) for the last 40 years is presented for different animal species. The data were obtained from the Belgian Ministry of Agriculture. Conclusion A database was created that contains a carefully judged inventory of analytical methods which may be considered as candidates for standardization after validation.The methods were classified as ‘high reliability’ or ‘limited reliability’ according to their conformity with the criteria described in Commission Decisions 93/256/EEC4 and 93/257/EEC.5 The database also contains toxicological data on consumer health, an estimation of the Belgian consumption of meat, an inventory of the European and Belgian legislation with regard to residues in animal products, commercially available equipment (antisera, immunoassay kits, radioactive and other tracers etc.), qualitative and quantitative data about the monitoring programs performed by the Belgian Institute of Veterinary Inspection and Ministry of Agriculture and chemical and physical data about the residues found in food products of animal origin.Acknowledgements This project was financially supported by the Belgian State, Services of the Prime Minister—Federal Office for Scientific, Technical and Cultural Affairs (OSTC) as part of the ‘Scientific Support Plan for a Sustainable Development Policy—Standards for Foodproducts’. The authors also thank the Centre d’Etude Multimedia Universitaire (CEMU) of the University of Liège for formatting the collected data for introduction into an electronic network. References 1 EC Council Directive 96/23, Off. J. Eur. Commun., 1996, No. L 125/10. 2 EC Council Directive 96/22, Off. J. Eur. Commun., 1996, No. L 125/3. 3 EEC Council Regulation No. 2377/90, Off. J. Eur. Commun., 1990, No. L 224/1. 4 EEC Commission Decision 93/256, Off. J. Eur. Commun., 1993, No. L 118/64. 5 EEC Commission Decision 93/257, Off. J. Eur. Commun., 1993, No. L 118/75. 6 Veterinary Drug Residues. Residues in Food Producing Animals and Their Products: Reference Materials and Methods, ed. R. J. Heitzman, Published on behalf of the Commission of the European Communities, Blackwell, Oxford, 2nd edn., 1994. 7 EC Commission Decision 98/179, Off. J. Eur. Commun., 1998, No. L 65/31. 8 EEC Council Directive 70/254, Off. J. Eur. Commun., 1970, No. L 270/1. Paper 8/04941F Analyst, 1998, 123, 2423–2427 2427

ISSN:0003-2654    

DOI:10.1039/a804941f

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

Increased milk levels of insulin-like growth factor 1 (IGF-1) for the identification of bovine somatotropin (bST) treated cows† Andreas Daxenberger,a Bernhard H. Breierb and Helga Sauerweina a Institute for Physiology, Research Centre For Milk and Food, 85354 Freising-Weihenstephan, Germany b Research Centre for Developmental Medicine and Biology, Auckland, New Zealand Received 29th June 1998, Accepted 4th September 1998 The present EU moratorium banning the use of bST to increase milk yield implies the need for official controls.Our study aimed to identify milk from bST treated cows via the induced increase of insulin-like growth factor 1 (IGF-1) concentrations. A non-extraction radioimmunoassay for IGF-1 was improved and thoroughly validated for milk. Accuracy was 99% recovery in a fortified sample material, the precision was 5.1% intra-assay variation and 13.4% inter-assay variation. Parallelism was proved by a dilution experiment which yielded a regression line with a slope (20.7%) not significantly different from zero (P = 0.534).Naturally occurring milk IGF-1 levels were recorded in 5777 random milk samples from the Bavarian dairy cow population. In samples from lactation week 7 to 33, the effect of somatic cell count (SCC), protein content and parity could be quantified and corrected; thus a normal distribution (20.068 mean ± 0.440 s) of the corrected logarithmic IGF-1 levels (corr ln IGF-1) was obtained. IGF-1 concentrations occurring in milk from bST treated cows were recorded in 33 Brown Swiss cows treated once with rbST (POSILAC®).Mean corr ln IGF-1 levels increased by 0.828 and 0.477 in first parity and older cows, respectively. Thus 60% and 29%, respectively, of the positives could be detected at a 95% probability. If our results are confirmed in experiments with more bST treated cows and with prolonged treatment intervals, IGF-1 measurements might be useful to monitor for bST application in milk samples. Introduction In 1994 the FDA approval and the introduction of a bST preparation (POSILAC®, Monsanto Company, Inc., St.Louis, MO) in USA and other countries to increase milk yield in dairy cows again raised public concern in the EU where the moratorium prohibiting the use of bST was prolonged until 2000.1 Regardless of a further extension of the EU moratorium, labelling of milk from cows which have not been treated with bST initiated a new marketplace for milk and milk products.2 Both the ban of bST use and the labelling of milk produced without bST, imply the need for official control.As generally accepted, the stimulative effect of exogenous bST on milk yield is mediated via the somatotropic axis, in particular through insulin-like growth factor 1 (IGF-1).3,4 With bST-application, the concentrations of both bST and IGF-1 are increased several times in plasma.5–8 For milk, no increase of bST concentrations is prevailingly documented when recommended dosages are used.9,10 In contrast to bST, IGF-1 concentrations in milk are consistently reported to rise during bST application.8,11,12 Because milk samples are easily available in food quality control, our study aimed to identify milk from bST treated cows via the induced increase of IGF-1 concentrations in milk.The objective of our investigation was to identify and to quantify the parameters affecting the IGF-1 content of milk and to describe the frequency curve of naturally occurring concentrations.In the present paper we accurately quantify and evaluate the extent of the increase in IGF-1 concentrations in milk after bST treatment and describe a screening method based on a refined and thoroughly validated radioimmunoassay (RIA) for the detection of IGF-1 in milk. The IGF-1 is tightly bound to specific IGF binding proteins (IGFBP) in biological fluids. From at least six related forms of IGFBP known at present, IGFBP-2, -3 and -4 are predominant in milk.11,13–15 Their affinities (Kd 10210 to 10211 nM) reach the affinities of antibodies commonly used in RIAs,16 causing major problems in IGF-1 measurements-based immunological techniques.17–19 Our approach to eliminate IGFBP interferences applies the functional separation of IGF-1 from IGFBPs by the IGFBP-blocked assay described by Blum and Breier.16 IGFBPbinding sites are blocked by an excess of IGF-2, which shows very little cross-reactivity with the highly specific IGF- 1-antiserum.Experimental Standards and reagents The IGF-1 antiserum was raised in rabbits (#878/4, see Breier et al.17 for details). Standard preparations were purchased from GroPep, Adelaide, Australia (IGF-1 and IGF-2, recombinant, human, “receptor grade”, cat. #CM001 and FU100). Tracer was prepared according to a chloramin-T method described in the standard manufacturer’s Technical Bulletin. The specific activity was between 1.5 and 3 3 1016 Bq mol21, the amount of added tracer was 160–200 Bq per vial.Unspecific rabbit IgG was purchased from Sigma (Deisenhofen, Germany). Anti rabbit IgG antiserum was prepared in sheep by immunisation against unspecific rabbit IgG (reagent grade). All other reagents were of analytical-reagent grade. Water (ultra pure) was prepared with a Milli-Ro10+ system (Millipore, † Presented at the Third International Symposium on Hormone and Veterinary Drug Residue Analysis, Bruges, Belgium, June 2–5, 1998.Analyst, 1998, 123, 2429–2435 2429Eschborn, Germany) and was used throughout all experiments. Milk samples Naturally occurring IGF-1-levels were measured in 5777 random milk samples from the Bavarian dairy cow population taken during the regular milk recording in Bavaria over one year covering all regions. All routine data for milk recording were made available by the routine laboratory. The animal experiment with bST was carried out at our institute’s experimental farm during March and April 1997. 34 Brown Swiss dairy cows were treated once with POSILAC® (Lot #96F03/10) according to the manufacturer’s directions for use. Milk samples were collected in analogy to the procedure of the milk recording during each routine milking process twice per day from 2 weeks before until 4 weeks after bST application. The milk yield of the entire farm was discarded during treatment, i.e. for three weeks after bST application. The milk constituents were measured in the same laboratory where the milk recording was carried out.Sample preparation Milk samples were conserved with sodium azide according to the regular procedure during milk recording (8 mg per 40 ml milk). Stability during storage was proved until acidification or putrefaction (minimum 3 days at 25 °C or 10 days at 10 °C). Conservation had no effect on measurable IGF-1 content. All samples were defatted by a 20 min upside-down centrifugation (4000g) at 4 °C.In this article ‘milk sample’ always refers to ‘defatted milk samples’. Radioimmunoassay The principle of our non-extraction method is as follows: milk samples (100 ml) were diluted by addition of 750 ml acidic buffer (20 mM sodium phosphate of pH 2.8, 0.1 M NaCl, 0.2% BSA and 0.5% Triton X-100), 0.5 M phosphoric acid (25 ml) and 60 mM EDTA in acidic buffer (75 ml). Thus a complete pH decrease to pH 2.8 and a complete complexation of the milk’s calcium ions were achieved.At last 50 ng IGF-2 in 50 ml acidic buffer were added in order to displace IGF-1 from the IGFBPs. The resulting sample dilution was 1 + 9, and the addition of IGF-2 was 500 ng per ml milk. With the help of relatively large amounts of a nonionic detergent (Triton X-100) in the buffer and the calcium complexation, the casein precipitation during the pH 2.8 incubation (minimum 15 min) was reversible when the pH was brought to neutral conditions later for the RIA. Before pipetting, the diluted sample had only to be vortexed shortly in order to solubilize the precipitate and thus eliminate casein interference.To 100 ml diluted sample specific antibody (878/4, final dilution of 1/105 000) and tracer (100 ml each) were added in a slightly alkaline assay–buffer (100 mM sodium phosphate of pH 7.8, 0.05 M NaCl, 0.2% BSA, 0.5% Triton X-100 and 0.02% NaN3) to raise the pH after the acidic incubation to a physiological range again for optimal performance of the specific antibody.The addition of unspecific rabbit IgG (50 mg per tube) together with the specific antiserum was necessary for complete formation of the immunocomplexes. After two days of incubation at 4 °C, bound and free ligands were separated by addition of 100 ml anti rabbit IgG antiserum (diluted 1 + 11.5 in assay–buffer) and 1 h of incubation at 4 °C. Immunoprecipitation was supported by addition of 1.5 ml of polyethylene glycol 6000 (4% m/v) to each vial before 45 min centrifugation at 4000g at 4 °C.Statistical analysis All statistics were performed with Sigma Stat Scientific Software (Jandel GmbH, Erkrath, Germany). Data management In this study the detection of bST treated animals is based on the description of the frequency of naturally occurring IGF-1 concentrations versus increased levels of IGF-1 under the influence of bST. Only mathematically defined frequency distributions such as the normal distribution allow for the calculation of threshold values and their probabilities, and thus for the calculation of their corresponding rates of correct and false estimations of bST treated cows.Therefore we used normalized data derived from the IGF-1 concentrations in milk (in ng ml21) by natural logarithmic transformation and the correction of the influence factors affecting IGF-1 levels. Influence factors were quantified with the help of the random milk samples by the following multiple linear regression formula model, ln IGF-1 = corr ln IGF-1 + a 3 ln SCC + b 3 prot + i2 + i36 + ihf where the natural logarithm (ln IGF-1) of the measured concentration of IGF-1 (ng ml21) is the sum of the corrected IGF-1 level (corr ln IGF-1) and the values for the influence factors.As shown in the results, the influence factors are the somatic cell count (ln SCC, as natural logarithm of thousands per ml, multiplied by the coefficient a), the protein content (prot, as % m/m, multiplied by the coefficient b), the lactation number (incremental values for second - i2 - or third to sixth - i36 - lactation) and breed (incremental value - ihf - for Holstein– Friesian animals).After the transformation of the linear regression equation, all corr ln IGF-1 values were obtained from the measured logarithmic IGF-1 concentration (ln IGF-1) and the parameters (SCC, protein content, lactation number and breed) by the equation corr ln IGF-1 = ln IGF-1 - a 3 ln SCC - b 3 prot - i2 - i36 - ihf The corr ln IGF-1 value shall be seen as a number without dimension, which serves only for the calculation of threshold values and corresponding detection efficiencies. Results RIA validation Presently there are no national or international regulations concerning the quantification of naturally occurring proteohormones or growth factors.Our validation follows the Commission Decision (93/256/EEC) laying down the methods to be used for detecting residues of substances having a hormonal or a thyreostatic action.20 Specificity The specificity of the antiserum used can be sufficiently described by the cross-reactivity to IGF-2, the substance most closely related to the analyte.Cross-reactivity is commonly given as the quotient of the half maximum displacement concentrations of the analyte and the cross-reacting substance. Determined by this manner, the antiserum showed a crossreactivity for IGF-2 of 0.0016%. However, this does not provide valid data for the interfering influx of small concentrations of 2430 Analyst, 1998, 123, 2429–2435IGF-2, which are needed to find the optimum dosage of IGF-2 added to block IGFBPs.If increasing IGF-2 concentrations are included in the assay, the measured IGF-1 concentrations follow the exponential increase shown in Fig. 1. The pretended IGF-1 concentrations can be calculated by the formula: c(IGF- 1) = 0.00327 3 c (IGF-2)0.4861 (R2 = 99.1%; concentrations c in ng ml21).For all amounts of added IGF-2 and all dilutions the faked IGF-1 can thus be predicted. Accuracy Because reference material is not available, accuracy can only be evaluated by analysing fortified (‘spiked’) sample material. One milk sample with a low (2.9 ng ml21 ± 0.23 s) and one with a high (10.2 ng ml21 ± 0,82 s) IGF-1 content each were spiked with 1, 2, 5, 10 and 20 ng ml21 IGF-1. Recovery was 99.2% (± 2.8 SEM; n = 88) with no significant differences between the 5 different spikes (Kruskal–Wallis one way ANOVA on ranks, P = 0.180) and between the two samples (t-test, P = 0.245).Precision The repeatability of the method is 5.1% evaluated as intra-assay variation. Reproducibility can not be described, as the method is not yet carried out in other laboratories. Under conditions of reproducibility within our laboratory, the inter-assay variation is 13.4%, determined on the basis of 5 control milk samples measured throughout all 64 assays carried out for this study.Limit of determination In competitive immunoassays the limit of determination is set by the location of the sigmoid calibration curve, mathematically best described by a four parameter logistic model. Conventionally the limit of determination is the 20% displacement concentration of the analyte and is about 0.1 ng IGF-1 per ml test solution. With a sample dilution of 1 + 9 the determination limit is 1 ng IGF-1 per ml milk in our test system, so that all naturally occurring IGF-levels in milk can be measured with the specified accuracy and precision. Therefore a statement of the limit of detection is not necessary.Susceptibility to interference Parallelism is the central element of any validation of measurements. Independent of the sample’s dilution, the concentration of the analyte found by the test system has to be constant, as long as the concentration in the diluted test solution is within the working range of the calibration curve.Fig. 2 shows the dilution graph of an exemplary milk sample with a relatively high content of IGF-1 ( > 8 ng ml21) and thus with a wide dilution range in the working range. The measured content of IGF-1 is dependent on the dilution, as demonstrated by the regression line [c(IGF-1) = 0.0259 3 dilution factor + 9.06; R2 = 0.43]. The following considerations indicate, that crossreactivity to IGF-2 causes non-parallelism (see Table 1). The addition of 500 ng IGF-2 per ml milk results in effective additional IGF-2 concentrations between 125 and 7.8 ng ml21 within the sample dilutions of 1 + 3 to 1 + 63.The crossreactivity can be calculated according to the regression formula of Fig. 1; faked IGF-1 is revealed by multiplication with the dilution factor. Measured IGF-1 concentrations minus the IGF- 1 faked by cross-reactivity turn out to be independent of the sample dilution, which is demonstrated by the regression line of the corrected values in Fig. 2 [c(IGF-1) = 20.0069 3 dilution factor + 8.76; R2 = 0.43; P = 0.534]. Optimum test conditions The favourable dilution of 1 + 9 ensures that all milk samples are detectable in the working range of the calibration curve. IGF-2 shall be added as much as necessary, but as little as possible. Fig. 3 shows the effect of increasing amounts of added IGF-2: faked IGF-1 is calculated by the regression formula (Fig.1), not displaced (not detectable) IGF-1 is estimated under the simplified assumptions that all IGF-1 is bound to IGFBP but that there is no excessive IGFBP, that the sample itself does not contain IGF-2 and that IGF-1 and -2 bind to IGFBP with the same affinity. Then the amount of non-displaced IGF-1 is derived from a simple displacement formula resulting in the decreasing trend depicted in Fig. 3. The corrected IGF-1 is the sum of IGF-1 faked by IGF-2, the non-displaced IGF-1 and the measured IGF-1. Optimal IGF-2 addition proves to be 500 ng per ml milk sample, an amount at which faked IGF-1, as well as non-displaced IGF-1, are negligible. Fig. 1 Increasing concentrations of IGF-2 test solutions (x-axis) measured as IGF-1 (y-axis). Table 1 Calculation of the RIA parallelism considering the cross-reactivity to IGF-2 after the addition of 500 ng IGF-2 per ml milk sample Sample dilution 1 + 3 1 + 5 1 + 7 1 + 11 1 + 15 1 + 23 1 + 31 1 + 47 1 + 63 Measured IGF-1 8.5 9.1 9.1 9.6 9.7 9.4 11.0 10.9 9.8 Effective addition IGF-2 125 83.3 62.5 41.7 31.3 20.8 15.6 10.4 7.8 Calculated cross-reactivity 0.083 0.074 0.067 0.060 0.055 0.048 0.04 0.039 0.036 Cross-reactivity 3 dilution 0.33 0.44 0.54 0.72 0.88 1.16 1.42 1.89 2.31 Corrected IGF-1 8.2 8.6 8.6 8.9 8.8 8.2 9.6 9.0 7.5 Fig. 2 Parallelism of the RIA: A, of the measured IGF-1 concentrations; B, after correction of the cross-reactivity influx. At a sample dilution of 1 + 9 cross-reactivity was negligible. Analyst, 1998, 123, 2429–2435 2431With the help of a Scatchard Plot analysis the dissociation constant Kd of the used antibody and IGF-1 was calculated to be 2.5 3 10211 M.The affinity of the antiserum to IGF-1 reaches the affinity of IGFBPs.16 Any valid IGF-1 quantification should be carried out with antibodies which have a similar affinity and cross-reaction as ‘878/4’. For each antiserum, the test has to be validated in similar manner as presented herein. Naturally occurring milk IGF-1 levels IGF-1 concentrations in the 5777 recorded samples were spread over a wide range from 1 to over 83 ng ml21, occasionally.The distribution was strongly skewed to the right with a median of 4.4 ng ml21 and percentiles for 90 and 95% of 9.5 and 12.5 ng ml21, respectively. There was no detectable effect of region, season, the cow’s milk performance and the milk sample’s fat content on the IGF-1 concentration. Only for the three major dairy breeds in Bavaria (Simmental, Brown Swiss and Holstein- Friesian) the numbers of observations allowed for statistical considerations.Besides, for regressions and tests for significance, normalized IGF-1 data gained by naturally logarithmic transformation (ln IGF-1) had to be used. Factors affecting IGF-1 levels As shown in many previous studies, IGF-1 levels are strongly dependent on the lactational state (see Fig. 4). Particularly at the beginning (colostrum phase), but also in late lactation, IGF-1 concentrations were relatively high.During weeks 7 until 33 of lactation the IGF-1 levels were constantly low and the variances were smaller than in the other periods. Elevated IGF-1 levels caused by bST will be best detectable before week 34. The IGF-1 milk concentrations prior to the ninth week of lactation were not considered because bST use is not recommended prior to this time. The somatic cell count (SCC) also contributes to the IGFconcentrations in milk. The linear regression in Fig. 5 (A; all 5777 milk samples) illustrates the correlation between natural logarithmic SCC (ln SCC) and ln IGF-1 values.Although the extent of the ln SCCs influence is rather small (R2 = 0.093), it was statistically significant (P < 0.001). Although not seen in Fig. 5 but apparent after mathematical calculation, above a SCC of 2 millions per ml (ln SCC > 7.6; calculated as ln 2000) the variance increases and the respective data have to be omitted for the multiple linear regression model used later.The correlation of the protein content and IGF-1 (see Fig. 5, B) is similar to SCC (R2 = 0.081), but also highly significant (P < 0.001). Because there is, with the exception of the colostrum phase, no causal relationship between SCC and the protein, both parameters have to be taken into account mathematically. In milk samples obtained from cows at their first, second or third to sixth lactation, different IGF-1 concentrations were observed (Kruskal–Wallis One Way Analysis of Variance on Ranks; all paired multiple comparison, Dunn’s Method, P < 0.05); higher lactation numbers than 6 stood out for their high variance in IGF-1 values and could not be included into the following model. Above that, samples of Holstein–Friesian cows show slightly elevated IGF-1 levels compared to the other breeds.Frequency and limits of naturally occurring IGF-1 concentrations To describe their frequency and their limits, naturally occurring IGF-1 concentrations (in logarithmic transformation) had to be corrected by the influence of SCC, protein content, number of lactation and breed according to the procedure described in the data management section.The following multiple linear regression includes all 3113 samples with a SCC < 2 millions per ml, originating from lactation week 7–33, lactation number Fig 3 Appropriate addition of IGF-2: A, detectable IGF-1; B, not displaced IGF-1; C, pretended IGF-1 by cross-reaction and D, mathematically corrected IGF-1 levels. An addition of 500 ng IGF-2 per ml milk sample implied a security factor also for large amounts of IGFBPs.Fig. 4 Mean IGF-1 concentrations (+ SEM) in milk during the entire lactation. Values and variances during lactation week 7–33 allow for mathematical considerations. Fig 5 Relation between (A) SCC or (B) protein content, and IGF-1. The regressions were significant (P < 0.001; n = 5777) and their influx could be computed by the regressions. 2432 Analyst, 1998, 123, 2429–24351-6 and from the three major breeds in Bavaria (Simmental, Brown Swiss and Holstein–Friesian). The coefficients including the P-values and incremental sum of squares (SSincr) are shown in Table 2. The P-values smaller than 0.001 (for the variables ln SCC, protein and the incremental value for the third to sixth lactation), smaller than 0.002 (incremental value for the breed Holstein– Friesian) and smaller than 0.025 (incremental value for the second lactation) confirm that the coefficients are significantly different from 0.However, the SCC and the protein content are the most important influence factors on IGF-1, as revealed by the SSincr values. The efficiency of the multiple linear regression is confirmed by its statistical power (P = 1.000 with a = 0.05), the constant variance (homoscedasticity) test (P = 0.653; computed as Spearman rank correlation between the absolute values of the residuals and the observed values of the dependent variable) and the normality test (P = 0.760; Kolmogorov–Smirnov test with Lilliefors’ correction).The Pvalues confirm the theory, that the differences between the ln IGF-1 of the observations and the constant of the regression are independent of the values of the variables and are normally distributed. After correction of their influence parameters, the natural logarithmic IGF-1 concentrations form a typical bellshaped Gaussian curve (Fig. 6) with a mean of 20.0677 and a standard deviation of 0.440. Normal distributions suggest that the variation is caused by random variables. Logarithmic normal distributions usually are the result of a multiplicative combination of factors being normally distributed themselves. For all corr ln IGF-1 values, according to the mathematical implications of the normal distribution, frequencies and the respective probabilities can be computed for milk samples derived from cows not treated with bST.Table 3 shows selected limits for the one-sided posing of question in the normal distribution (P, probability; z, factor of the standard deviation) and the values of the corr ln IGF-1 they are affiliated to (threshold values). Increase of IGF-1 concentrations after treatment with bST Fig. 7 demonstrates the increase of the mean IGF-1 concentrations of all 34 cows in the course of time. 10 days after bST treatment IGF-1 levels reached a maximum and declined in the following two weeks back to initial values.The statistical analysis focuses on the main effect period (14 milkings from day 7 to 13 after treatment) compared to 14 milkings in the 7 days before treatment. IGF-1 concentrations of all animals increased significantly (t-test, P < 0.001) with the exception of one cow (P = 0.150). To compare elevated IGF-1 levels with naturally occurring values, logarithmic transformation and correction of the influence factors had to be carried out.One cow suffered from acetonaemia during the treatment period and was not further considered. The results of a two-way ANOVA (general linear model) for corr ln IGF-1 on the factors number of lactation (1 or 2–6) and bST-treatment (treated or untreated) revealed a significant influence of the interaction lactation–treatment on corr ln IGF-1; i.e., primiparous (n = 16) and older cows (n = 17) react differently to the bST treatment (P < 0.001). Therefore a differentiated evaluation concerning calving numbers is necessary.Presently it is not possible to decide whether this difference actually reflects natural facts or is just the result of the limited number of animals in the dairy herd investigated herein. The detection of bST treated animals The standard deviations within the 462 recorded milk samples (14 milkings of 33 cows in each period) before (s = 0.447) and after treatment (s = 0.400) are similar to the standard deviation gained from the 3114 samples of different non-treated cows (s = 0.440).Therefore calculation of detection rates for treated cows is based on the shifting of the Gaussian curve for corr ln IGF-1 to the right, as demonstrated in Fig. 8 for animals in the first lactation (threshold value for 1% false estimations). For all levels of corr ln IGF-1 the fraction of milk samples gained from non-treated cows (threshold values) can be computed as well as the number of samples from treated cows, that can be detected at the corresponding threshold value. Table 4 summarises the detection rates of treated cows at threshold values for 0.1, 1 and 5% under differentiation of primiparous and multiparous cows. Table 2 Multiple linear regression for ln IGF-1 values in the influence factors SCC, protein content, number of lactation and breed.Corrected ln IGF-1 values were calculated by subtraction of the influence increments from the observed value. Parameter Coefficient P SSincr Constant 20.0677 0.0888 0.446 1n SCC (1n 1000 ml21) 0.0974 0.00736 < 0.001 54.294 Protein concentration (% m/m) 0.275 0.0252 < 0.001 20.358 Lactation 2 0.0479 0.0214 0.025 0.239 Lactation 3–6 0.123 0.0188 < 0.001 8.249 Holstein–Friesian 0.134 0.0423 0.002 1.934 Fig. 6 Normal distribution of logarithmic IGF-1 concentrations, corrected for the influence of SCC, protein content, lactation and breed (n = 3113, s = 0.440, mean = 20.067). Normality allows for stochastic considerations.Table 3 Threshold values for corr ln IGF-1. The z values are the multipliers for the standard deviation, P is the probability of the random variable to exceed the threshold value. P z Threshold value corr 1n IGF-1 0.001 3.090 1.192 0.01 2.326 0.956 0.5 1.645 9.656 Fig. 7 Mean IGF-1 concentrations in milk after bST treatment. Statistical analysis was based on (A) the control period and (B) main effect period. Analyst, 1998, 123, 2429–2435 2433Discussion Mean IGF-1 concentrations in milk reportedly range between 4 to 35 ng ml21; after bST application the levels rise within physiological limits.9,11,15 The wide range of IGF-1 concentrations and the different statements about the increase8,9,11,12 after bST application might be due to different analytical methods concerning the elimination of interference caused by IGFBP.Confirmation methods for IGF-1 assays, e.g. by MStechniques, are not available so far. Separation of IGF-1 from IGFBP by size exclusion chromatography (SEC) prior to IGF-1 quantification is widely recommended and accepted,16–19 but not possible in milk.Because of the protein precipitation occurring under the acid conditions which are necessary to dissociate IGFs from IGFBPs, milk samples can not be put completely on SEC columns, and casein separation from the milk samples causes the general problem of IGF-1 losses. In the assay described here there are no extraction steps which basically bear the risk of reduced recoveries.Because of the lack of confirmation techniques, complete validation has to prove the reliability of the assay. Absolute parallelism suggests that IGFBP binding sites are efficiently blocked by excessive IGF-2 and that there is no other interference. The recovery of 99% confirms the good accuracy of the method. Therefore, our thoroughly validated assay provides reliable data for IGF-1 concentrations in milk and their elevation under the influence of bST.The measurement of increased concentrations of IGF-1 in milk from bST treated cows allows for stochastic considerations, but in principle not for qualitative statements. According to the choice of threshold values, different numbers of treated individual animals can be detected. After recording two or more samples taken from each animal the efficacy of the method can be improved. Moreover, the wide use of bST in complete herds is intended by the manufacturer and will be expected in the case of authorization in the EU.In accordance with the stochastic laws, for bST treated herds reliable evaluations should be possible by combining the frequencies for correct and false estimations for individual animals. Admittedly, our results have to be confirmed with the help of larger studies considering other breeds and longer periods of application time. Especially the different reaction of primiparous and older animals has to be critically examined with larger numbers of experimental animals.However, presently the method proposed herein is to our knowledge the only possibility to monitor bST treatment using milk samples. In addition, non-governmental agricultural associations (see acknowledgements) which carry out the milk recording have the logistic background to draw milk samples comprehensively and have the ability for the necessary analytical work. A valuable extension for the examination of milk might be a screening method based on increased levels of both bST and IGF-1 in plasma described by Schams et al.(96% correct evaluations with about 20% false estimations).21 However, in food quality control, milk and dairy products themselves, and not blood, are the primary matter of concern. Collection of blood samples is an invasive procedure and might bring about legal difficulties for the food controlling authorities. Generally, a confirmation method to prove bST treatment does not yet exist.One interesting attempt might be the MSbased differentiation of recombinant and natural bST. Naturally derived pituitary bST is a proteohormone consisting of 4 isoforms with 190 or 191 amino acids and a heterogeneity at position 127 (126 respectively) with valine or leucine.3 Recombinant bST is derived from the 191 amino acid–127 valine isoform, but the preparations produced by different pharmaceutical companies differ in kind and number of additional N-terminal amino acids.22 In POSILAC®, the only bST product approved by the FDA up to now, the N-terminal alanine is replaced by methionine for technical reasons, resulting in a higher molecular weight (21 851 g mol21) compared to the natural isoforms (21 791 g mol21 or less).This mass difference could be detected by electrospray MS of the pure substance compared to pituitary extracts, but it has not been successfully used to identify recombinant bST in biological matrices.22 Detection of differences in immunoreactivity between pituitary and a recombinant form of bST with an additional octapeptide at its N-terminal end (Eli Lilly Italia S.p.A., Florence, Italy) was possible with the help of monoclonal antibodies,23 but immunological differences between POSILAC® and pituitary bST have not yet been observed.Due to their differences at the N-terminus compared to natural bST, the recombinant preparations might have an immunogenic effect in the animals they are applied to.Antibody formation against recombinant bST has been reported,22,24,25 but detection of bST-treated cows by measurement of antigenic activities against recombinant bST seems to be difficult because of the cows great individual variability. Altogether, at the moment it is not possible to distinguish between bST-treated and untreated cows by means of qualitative differences; quantitative changes show the best prospects for the control of bST-application. Besides IGF-1 and bST, levels of other bST-sensitive factors of Fig. 8 Shifting of the bell-shaped curve to the right after bST application (first lactation): A, before treatment; B, after treatment; C, threshold value 0.956; D, 1% false estimations; E, 32.6% detection rate for treated animals. Table 4 Efficiency of the detection of bST treated cows by means of increased concentrations of IGF-1 in milk. Lactation number Increase (D) corr ln IGF-1 Corr ln IGF-1 after treatment (20.068 + D) Tolerable fraction false estimations (%) Threshold value Difference threshold value 2 corr ln IGF-1 after treatment Differences as factor of s (difference/0.440) Detectable fraction of bST treated cows (%) 0.1 1.292 0.532 1.209 11.3 1 0.828 0.760 1.0 0.956 0.196 0.445 32.6 5.0 0.656 20.104 20.236 59.5 0.1 1.292 0.883 2.007 1.1 2–6 0.477 0.409 1.0 0.956 0.547 1.234 10.8 5.0 0.656 0.247 0.561 28.8 0.1 1.292 0.708 1.609 5.4 No discrimination 0.652 0.584 1.0 0.956 0.372 0.845 19.8 5.0 0.656 0.072 0.164 43.6 2434 Analyst, 1998, 123, 2429–2435the somatotropic axis, e.g.growth hormone binding protein (GHBP) and IGF binding proteins 2 and 3 (IGFBP-2, -3), seem to be altered after bST-application: bST decreased the serum and mammary lymph levels of IGFBP-2, while IGFBP-3 concentrations were increased only in plasma.26 The presence of GHBP in cattle plasma and milk is reported,27 but treatment of different species with GH caused variable response in serum GHBP.28 However, because of the lack of precise and accurate immunological test methods, these factors are not suitable parameters for bST-control at the moment, but they might be worthwhile being focused at in future. Another attempt to detect bST treatment presented by the Cornell Research Foundation aimed at the decrease in phosphorylation of the fatty acid binding protein in the milk fat globule membrane caused by bST application.29 However, the detection of bST treated animals is also based on quantitative differences; the patent description is not detailed enough to evaluate the limits and the reliability of the method.IGF-1 measurements in milk samples might be useful to monitor bST treatment, although they can not be used for forensic purposes. The technique presented in this paper can be applied as a screening procedure prior to further confirmatory testing, for which commonly accepted methods do not yet exist. Acknowledgements This work was supported by the Bavarian Ministry for Nutrition, Agriculture and Forestry (Bayerisches Staatsminiserium für Ernährung, Landwirtschaft und Forsten).We thank Dr. Duda from the ‘Landeskuratorium der Erzeugerringe für tierische Veredelung in Bayern e.V.’ for enabling the collection of 5777 random milk samples and the “Milchprüfring Bayern e.V.” for the measurement of the milk components during the bST animal experiment. References 1 Council Decision, 94/936, 1994, No L 366/19. 2 T.J. Centner and K. W. Lathrop, J. Dairy Sci., 1997, 80, 215. 3 D. E. Bauman, J. Dairy Sci., 1992, 75, 3432. 4 D. E. Bauman and R. G. Vernon, Ann. Rev. Nutr., 1993, 13, 437. 5 D. Schams, F. Graf, J. Meyer, B. Graule, M. Mauthner and C. Wollny, J. Anim. Sci., 1991, 69, 1583. 6 O. L. Hadsell, C. R. Baumrucker and R. S. Kensinger, J. Endocrinol., 1993, 137, 223. 7 B. K. Sharma, M. J. Vandehaar and N. K. Ames, J. Dairy Sci., 1994, 77, 2232. 8 X. Zhao, B. W. McBridge, L. M. Trouten-Radford, L.Golfman and J. H. Burton, Dom. Anim. Endocr., 1994, 11(2), 209. 9 D. Schams, in Use of Somatotropin in Livestock Production, ed. K. Sejersen, M. Vesterhaard and A. Neiman-Sorensen, Elsevier, London, 1989, p. 192. 10 A. R. Torkelson, K. A. Dwyer, G. J. Rogan and R. L. Ryan, J. Dairy Sci., 1987, 70 (Suppl 1), 146. 11 C. G. Prosser, I. R., Fleet and A. N. Corps, J. Dairy Res., 1989, 56, 17. 12 B. G. Hammond, R. J. Collier, M. A. Miller, M. McGrath, D. L. Hartzell, C. Kotts and W. Vandaele, Ann. Rech. Vét., 1990, 21, (Suppl. 1), 107. 13 P. G. Campbell, and C. R. Baumrucker, J. Endocrinol., 1989, 120, 21. 14 C. E. Grosvenor, M. Frances Picciano and C. R. Baumrucker, Endocr. Rev., 1992, 14, 710, and references cited therein. 15 C. R. Baumrucker, W. M. Campana, C. A. Gibson and D. E. Kerr, Endocrine Regul., 1993, 27, 157. 16 W. F. Blum and B. H. Breier, Growth Regulation, 1994, 4, (Suppl.), 11. 17 B. H. Breier, B. W. Gallaher and P. D. Gluckman, J. Endocrinol., 1991, 128, 347. 18 P. Bang, R. C. Baxter, W. F. Blum. B. H. Breier, D. R. Clemmons, K. Hall, R. L. Hintz, J. M. P. Holly, R. G. Rosenfeld and J. Zapf, Endocrinology, 1995, 136, 816. 19 J. M. P. Holly and S. C. Cwyfan Hughes, J. Endocrinol., 1994, 140, 165. 20 Commission Decision, 93/256, 1993, No L 118/64. 21 D. Schams, P. Matzke, P. Hollwich and H. Karg, Z. Ernährungswiss, 1990, 29, 154. 22 M.-L. Scippo, G. Degand, A. Duyckaerts, and G. Maghuin-Rogister, Ann. Méd. Vét., 1997, 141, 381. 23 A. Berrini, V., Borromeo and C. Secchi, Hybridoma, 1994, 13, 485. 24 C. M. Zwickl, H. W. Smith, R. N. Tamura and P. H. Bick, J. Dairy Sci., 1990, 73, 2888. 25 P. J. Eppard, G. J. Rogan, B. G. Boysen, M. A. Miller, R. L. Hintz, B. G. Hammond, A. R. Torkelson, R. J. Collier and G. M. Lanza, J. Dairy Sci., 1992, 75, 2959. 26 W. S. Cohick, M. A. McGuire, D. R. Clemmons and D. E. Bauman, Endocrinology, 1992, 130, 1508. 27 A. Devolder, R. Renaville, M. Sneyers, I. Callebaut, S. Massart, A. Goffinet, A. Burny and D. Portetelle, J. Endocrinol., 1993, 136, 91. 28 S. L. Davis, N. B. Wehr, D. M. Laird and A. C. Hammond, J. Anim. Sci., 1994, 72, 1719. 29 Cornell Research Foundation, Int. Pat., G01N 33/06, 1997. Paper 8/04923H Analyst, 1998, 123, 2429–2435 2435

ISSN:0003-2654    

DOI:10.1039/a804923h

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

Development of a one step strip test for the detection of sulfadimidine residues† Ron Verheijen,* Piet Stouten, Geert Cazemier and Willem Haasnoot State Institute for Quality Control of Agricultural Products (RIKILT-DLO), P.O. Box 230, 6700 AE Wageningen, The Netherlands Received 22nd June 1998, Accepted 27th August 1998 The one step strip test described is a competitive immunoassay in which the detector reagent consists of colloidal gold particles coated with affinity purified polyclonal anti-sulfadimidine (SDD) antibodies.The capture reagent in the assay is an SDD–ovalbumin conjugate which is immobilised on the lateral flow membrane of the test device. In the test procedure, 150 ml (four drops) of a liquid sample (buffer, urine or milk) are brought into the sample well of the test device and allowed to migrate over the membrane. The more analyte present in the sample, the more effectively it will compete with the SDD immobilised on the membrane for binding to the limited amount of antibodies of the detector reagent.A sufficient amount of SDD in the sample will therefore prevent the binding of the detector reagent to the SDD immobilised on the membrane. Therefore, a positive sample will not show a test line in the read-out zone. With spiked buffer or calf urine this was obtained at a level of > 10 ng ml21 of SDD and with spiked (diluted) fresh cow milk at a level > 20 ng ml21 of SDD. At these levels, the test is applicable only as a qualitative assay. The presence or absence of a test line indicates lower or higher levels of SDD, respectively.The major advantages of the one step strip test are that results can be obtained within 10 min and that all reagents are included in the test device. Introduction In 1980, Leuvering et al.1 reported the use of colloidal particles, not necessarily gold particles, as labels for immunoassays and proposed the term sol particle immunoassay (SPIA). Macromolecules such as proteins (e.g., immunoglobulins) can be adsorbed on the dyed colloidal particles2 [e.g., gold (red), carbon (black), latex3 (several colours)].Such immunoglobulin loaded particles can be applied to make antigens directly visible. One step strip test devices, using antigen immobilised on a membrane and antibody–gold conjugates as mobile phase, are commercially available for an increasing number of antigens (high and low molecular mass). The first major target analyte for this test format was human chorionic gonadotropin (HCG) for the detection of pregnancy.A great variety of test strip assays are now available,2,3 e.g., for the detection of hormones (pregnancy, fertility, ovulation, menopause, sexual disorder, thyroid functions), tumour markers (prostate, colorectal, etc.), viruses (HIV, hepatitis B and C), bacteria (Streptococcus A and B, Chlamydia trachomatis, Treponema pallidum, Helicobacter pylori, etc.), allergy (IgE) and cardiac monitoring (troponin T4).All these analytes are measured on the basis of their presence or absence. For such assays, a binding agent specific for the analyte (usually, but not necessarily, an antibody) is immobilised on the membrane. The detector reagent, typically an antibody coupled to latex or colloidal metal, is deposited (but remains unbound) in the conjugate pad. When liquid sample (urine, serum, plasma, whole blood, milk, etc.) is added to the conjugate pad, the detector reagent is solubilised and begins to move with the sample flow front up the membrane strip.The analyte present in the sample is bound by the antibody coupled to the detector reagent. As the sample passes over the zone to which the capture reagent has been immobilised, the analyte–detector reagent complex is trapped. Here, colour development is proportional to the amount of analyte present in the sample, which is only applicable for analytes with more than one epitope (high molecular mass analytes).There are also commercially available assays for low molecular mass analytes, e.g., for drugs of abuse, for steroidbased ovulation prediction and for progesterone in milk,5 which are based on competitive immunoassay protocols. In this type of assay, the detector reagent can be the analyte bound to a protein or the antibody bound to latex or a colloidal metal. As the analyte containing sample and the detector reagent pass over the zone to which the capture reagent (antibody or analyte–protein conjugate) has been immobilised, some of the analyte and some of the detector reagent are bound and trapped.The more analyte present in the sample, the more effectively it will be able to compete with the immobilised analyte on the membrane for binding to the limited amount of antibodies of the detector reagent. Hence an increase in the amount of analyte in such competitive immunoassays will result in a decrease in signal in the read-out zone.So far, these tests have been mainly intended for human diagnostics where, especially for the low molecular mass analytes, relatively high concentrations of analytes are measured. In food diagnostics, however, lower detection levels are needed. For the detection of veterinary drug residues, for example, detection limits at ppb level (nanograms of analyte per gram of sample) are necessary. In cooperation with a diagnostic company [Princeton BioMeditech (PBM), Princeton, NJ, USA], an SPIA based one step strip test for SDD, using polyclonal anti-SDD, was developed.6 By adjusting the amount of essential reagents in the strip test device, two types of competitive tests (low- and high-level tests) were prepared.In the low-level test, almost complete disappearance of the signal in the read-out zone was obtained at a level of > 100 ng ml21 and in the highlevel test at a level of > 1000 ng ml21. The low- and high-level tests were intended for the monitoring of plasma samples and for urine samples, respectively. Owing to the higher levels of SDD and the presence of cross-reacting metabolites in plasma † Presented at the Third International Symposium on Hormone and Veterinary Drug Residue Analysis, Bruges, Belgium, June 2–5, 1998.Analyst, 1998, 123, 2437–2441 2437and especially in urine compared with tissue, levels of 200 ng ml21 in plasma and 1000 ng ml21 in urine would compare with about 20–40 ng g21 of the parent compound in tissue.7 These action levels included safety margins for the variation of SDD concentrations and were recommended for the control on a maximum residue level (MRL) of 100 ng g21.In this study, with SDD as the target analyte, all ingredients of the test device were reinvestigated and/or selected with the aim of acquiring expertise about all aspects of the test principle to be used and to develop as sensitive a test as possible. With a higher sensitivity, such tests could also be suitable for analytes with lower MRLs.Experimental If not stated otherwise, procedures were performed at room temperature. Chemicals and materials Sheep serum was obtained from the DLO Institute for Animal Science and Health (ID-DLO, Lelystad, The Netherlands). Bovine g-globulins (BGG), bovine serum albumin (BSA), 1-ethyl-3-[3-(dimethylaminopropyl]carbodiimide hydrochloride (EDC), sodium azide, ovalbumin grade V, sulfadimidine (SDD), tetrachloroauric(iii) acid trihydrate (HAuCl4·3H2O), Igepal CA-630 (octylphenoxypolyethoxyethanol) and purified rabbit anti-sheep IgG (whole molecule) (2.5 mg ml21) were obtained from Sigma (St.Louis, MO, USA). Acetic anhydride, d-(+)-glucose monohydrate, glutaraldehyde (25% m/m solution in water), glycerol (extra pure), potassium carbonate (extra pure), lactose monohydrate, sodium acetate, sodium carbonate, sodium dihydrogenphosphate monohydrate, sodium ethylenediaminetetraacetate, sodium hydrogencarbonate, sodium hydroxide, trisodium citrate dihydrate and sucrose were supplied by Merck (Darmstadt, Germany).The bicinchoninic acid (BCA) protein assay and PharmaLink immobilization kit were obtained from Pierce (Rockford, IL, USA) and sodium dodecyl sulfate and Visking dialysis tubing (44110, 20/32, diameter 16 mm) from Serva (Heidelberg, Germany). Both a surfactant starter kit (manufactured by Pragmatics) and diagnostic materials kit were obtained from BioDot (Irvine, CA, USA).AE 100 (12 mm) nitrocellulose membrane strips (300 3 25 mm) on a Mylar backing were supplied by Schleicher and Schuell (Dassel, Germany), 5 mm Acrodisc filter by Gelman Sciences (Ann Arbor, MI, USA) and Centriprep-30 concentrators by Amicon (Beverly, MA, USA). Varian (Harbor City, CA, USA) supplied a Bond Elut reservoir (8 ml). Centriprep concentrators (Type 30) were supplied by Amicon. Equipment The BioDot system consisted of two BioJets Quanti3000 attached to a BioDot XYZ-3000 dispensing platform.An AZCON Sur-Size automatic guillotine cutter (Model SS-4) was supplied by AZCON (Elmwood Park, NJ, USA). Sealing equipment (Magneta, Model 421) was purchased from Kenosha (Amstelveen, The Netherlands). Antibodies Preparation of the immunogen, by coupling of succinylated SDD to BSA, and the raising of polyclonal antibodies in a rabbit against the SDD-BSA obtained conjugate, have been described previously.7 Specific anti-SDD immunoglobulins were isolated from the obtained complete rabbit serum (pool 464B) by affinity chromatography with immobilised SDD.Preparation of an affinity column with immobilised SDD SDD was coupled to PharmaLink Gel, obtained from a PharmaLink immobilization kit, as follows. A 2 ml volume of gel slurry of 50% m/v DADPA (diaminodipropylamine)- activated 6% cross-linked agarose in 0.1 m MES [2-(Nmorpholino) ethanesulfonic acid] (pH 4.7) (PharmaLink coupling buffer) was poured into a Varian column.After draining the storage solution, the gel was washed with 20 ml of demineralized water (water) and transferred into a PharmaLink reaction bottle. After addition of 2 ml of 0.2 m sodium acetate and 2 ml of acetic anhydride, the mixture was gently swirled for 30 min at 4 °C on a rock and roll shaking platform. Subsequently, another 1 ml of acetic anhydride was added, followed by another 30 min of incubation. After transferring the gel slurry into the column, the reaction mixture was drained completely and washed successively with 10 ml of water, 10 ml of 0.1 m sodium hydroxide and approximately 25 ml of water to pH 7.Coupling was performed in a PharmaLink reaction bottle by suspending the gel in 2 ml of 0.1 m MES (pH 4.7) (PharmaLink coupling buffer) and adding 20 mg of SDD in 500 ml of methanol and 60 mg of EDC. The reaction mixture was incubated for 24 h while being swirled on a rock and roll shaking platform. As the pH of the reaction mixture increased slightly during the first 8 h of the incubation, 1 m hydrochloric acid was added at regular intervals to maintain a pH of approximately 5.After the incubation, the gel slurry was once again transferred into the column. The reaction mixture was drained completely, then the gel was washed successively with 20 ml of water, 20 ml of 0.1 m sodium acetate, 20 ml of 0.5 m sodium hydrogencarbonate, 20 ml of water, 6 ml of 0.1 m acetic acid and 10 ml of phosphate buffered saline (PBS) (pH 7.4), containing 0.02% m/v sodium azide.Isolation of rabbit anti-SDD specific antibodies The SDD affinity column was washed with 20 ml of PBS to remove the sodium azide. For absorption, 2 ml of 464B serum were diluted with 3 ml of PBS and poured on to the column. The diluted serum was allowed to drain completely, then it was poured once again on to the column. Washing was performed with 10 ml of PBS, followed by desorption with 15 ml of 0.1 m acetic acid.During desorption, the pH of the eluate was kept between 5 and 7 by the addition of small amounts of 1 m sodium hydroxide. The entire eluate of approximately 15 ml was adjusted to pH 7.5 with 1 m sodium hydroxide solution and concentrated to about 2 ml in a Centriprep-30 concentrator according the manufacturer’s operating instructions. Centrifugation was performed in a Roty III centrifuge with a swingingbucket rotor. The purified and concentrated anti-SDD antibody preparation was stored in 1 ml aliquots at 220 °C.The protein concentration of the antibody preparation obtained was determined using the BCA protein assay. Regeneration of the column was performed directly after elution by washing with 10 ml of PBS until the pH of the eluate was neutral. Isolation of total IgG from sheep serum After mixing 10 ml of raw sheep serum with 20 ml of PBS, 30 ml of saturated ammonium sulfate solution (80 g in 100 ml of water) were slowly added with constant stirring.After standing for 30 min, immunoglobulins were collected by centrifugation for 10 min at 11 000 g in a Sigma-320K centrifuge with a swinging-bucket rotor. After dissolving the pellet in as small a volume of PBS as possible (a few millitres), the solution was dialysed for 48 h at 4 °C against PBS. The protein content was determined using the BCA protein assay. The IgG obtained was stored in small aliquots at -20 °C until used. 2438 Analyst, 1998, 123, 2437–2441Protein quantification The concentration of proteins in solution was determined by the BCA protein assay (Pierce) in microtitre plates using either ovalbumin or bovine g-globulins (BGG) as standard proteins.Preparation of colloidal gold particles (G40) Colloidal gold with an average particle diameter of 40 nm (G40) was prepared by controlled reduction of gold chloride with sodium citrate using the procedure described by Frens.8 The solutions of the reagents were prepared in water that was purified using a Milli-Q 185 Plus water purification system (Millipore, Bedford, MA, USA).The water obtained in this way had a resistivity of > 18.2 M½ cm. The procedure for preparing a G40 suspension was as follows: 100 ml of 0.01% m/v tetrachloroauric(iii) acid trihydrate were heated to boiling under reflux conditions, then 1 ml of 1% m/v trisodium citrate dihydrate solution was added under constant stirring. After about 25 s the slightly yellow solution turned faintly blue (nucleation).After approximately 70 s the blue colour then suddenly changed to brilliant red, indicating the formation of monodispersed spherical particles. The solution was allowed to boil for another 5 min to complete the reduction of the gold chloride. The absorbance, measured at 540 nm (A540) of such G40 suspensions was about 0.9. The size of gold particles obtained in one such procedure was checked using transmission electron microscopy (IPO-DLO). The particle diameter was found to vary between 38 and 43 nm.When supplemented with 0.05% m/v sodium azide, the G40 suspensions obtained could be stored at 4 °C for several months. Preparation of antibody-coated colloidal gold (detector and control reagent) The pH of 50 ml of G40 sol with an A540 of approximately 0.9 was adjusted to 8.5 with 0.2 m potassium carbonate solution. Coating of the G40 particles was performed by incubating 7.5 mg of antibody per millilitre G40 sol for 45 min while being gently swirled on a rock and roll shaking platform.The antibodies used were either affinity purified rabbit anti-SDD antibodies obtained from serum 464B (test detector reagent) or rabbit anti-sheep IgG (control detector reagent). For both types of antibody, the minimum protecting amount (MPA), i.e., the minimum amount of protein needed to protect the sol from saltinduced precipitation, was found to be approximately 5 mg per millilitre of G40 sol as determined by the procedure of Horisberger and Rosset.9 After coating, the sol was further stabilised by adding 5 ml of 1% m/v bovine milk casein mixture (Na, K and Ca salts, obtained from the Department of Quality Control, RIKILT-DLO, Wageningen, The Netherlands) that had been adjusted to pH 8.5 with 0.2 m potassium carbonate.After incubation for 60 min on a rock and roll shaking platform, 27.5 ml of the detector or control reagent were centrifuged in a 50 ml Greiner tube over a discontinuous glycerol gradient consisting of 5 ml of 80% v/v glycerol in water (pH 8.5) and 7.5 ml of 50% v/v glycerol in water (pH 8.5).Centrifugation was performed for 45 min at 2000 g in a Roty III centrifuge with a swinging-bucket rotor. Under these conditions, the detector or control reagent was found entirely in the 50% v/v glycerol layer. The particles were harvested by piercing the centrifuge tube with a syringe needle and removing the dark red layer sideways out of the tube.The 4–5 ml of detector reagent obtained were then diluted to 15 ml with water (pH 8.5), transferred into a Centriprep-30 concentrator and centrifuged in a Roty III centrifuge for 15 min at 1500 g. This was repeated four times, then in an additional run the reagent was further concentrated to 2 ml. After addition of 400 ml of 6% m/v bovine milk casein (pH 8.5) in water (1% m/v end concentration) and 12 ml of 10% m/v sodium azide (0.05% m/v end concentration), both the detector and control reagents were stored at 4 °C until used.The following composition of the reagent mixture was applied on the conjugate pads (Whatman Glass Fiber Paper with F 075-17 binder, 6 34 mm): 30 ml of anti-SDD antibody coated G40 sol (test detector reagent), 50 ml of rabbit anti-sheep coated G40 sol (control detector reagent), 20 ml of 12% m/v sucrose in water, 4 ml of 1% v/v surfactant S24 in water and 16 ml of water. Each conjugate pad was loaded with 12 ml of the detector reagent mixture and dried for 2 h at 40 °C.Thereafter, the pads were immediately assembled into the test devices. Preparation of capture reagents Test capture reagent. The SDD–ovalbumin conjugate that was used as the test capture reagent was prepared as follows: 10 mg of SDD were dissolved in 1 ml of methanol and added to 99 ml of PBS (solution I). Solution II was prepared by dissolving 300 mg of ovalbumin in 30 ml of PBS, after which the solution was filtered over a 5 mm Acrodisc filter (Gelman Sciences, Model 4489). The reaction mixture consisted of 30 ml of solution I, 30 ml of solution II and 30 ml of PBS to which 900 ml of 25% m/m glutaraldehyde solution were added slowly with constant stirring.Approximately 10 min after addition of the glutaraldehyde, the clear solution took on a yellowish appearance. The mixture was stirred for 4–5 h and exhaustively dialysed against PBS for 4–5 days. The dialysate was then concentrated to 50 mg ml21 protein using a Centriprep-30 concentrator and stored in 1 ml aliquots at 220 °C.Control capture reagent. Ammonium precipitated total sheep immunoglobulin fraction was used as the control capture reagent. Immobilization of capture reagents Both the test and control capture reagent were dispensed on the nitrocellulose membrane strips (300 3 25 mm) at a concentration of 0.5 mg cm21 protein using a BioJet Quanti3000, attached to a BioDot XYZ-3000 dispensing platform.After drying the membrane strips for 2 h at 40 °C, blocking was performed by incubating the strips for 30 min in 0.9 g l21 sodium dihydrogenphosphate monohydrate, 2% m/v non-fat milk powder and 0.02% m/v SDS. After blocking, the membrane strips were washed three times for 5 min with 0.9 g l21 sodium dihydrogenphosphate monohydrate containing 0.01% (v/v) surfactant (S24). After drying for 2 h at 40 °C, the membrane was divided into 6 mm strips using the cutter. The strips were stored under dry conditions (desiccator) at room temperature until used.Test production As shown in Fig. 1, the test strip consists of a backing plate on which the nitrocellulose membrane, detector conjugate pad, sample pad and absorbent pad are pasted. In all assays Gelman Type 133 cellulose was used for both the absorbent and sample pads. The test strip as shown in Fig. 1 was mounted in a plastic housing provided with a sample well and read-out zone and sealed in a pouch in the presence of a drying agent.Test procedure The principle of the assay is depicted in Fig. 2. Four drops (about 150 ml) of the liquid sample were pipetted into the sample well and allowed to migrate upwards. After 10 min, the test result was judged. When SDD was present in the sample, it Analyst, 1998, 123, 2437–2441 2439competed with the immobilised SDD on the capture line for the limited amount of antibodies of the detector reagent. When a sufficient amount of SDD was present, it therefore prevented the binding of detector reagent on the membrane.Hence a positive sample gave no visible test line in the sample read-out zone. When the test is performed properly, the control test line is always visible. Results and discussion A polyclonal rabbit antiserum raised against an SDD–BSA conjugate was used as the antibody source. This serum (pool 464B) was previously successfully applied for the development of a sensitive enzyme linked immunosorbent assay (ELISA) for SDD and its major metabolites.7 In general, a polyclonal rabbit serum contains about 60 mg ml21 of protein, of which approximately 10 mg is IgG and about 1 mg of this IgG is specific towards the hapten.The specific anti-SDD IgGs were isolated by affinity chromatography using a column with immobilised SDD as described in the Experimental section. In this way, two batches of affinity purified antibodies were prepared. For each batch the procedure was performed five times (10 ml of raw serum per batch). The amount of protein in these batches, as determined by the BCA protein assay, was found to be 0.76 mg ml21 in an end volume of 8 ml ( = 6.1 mg of specific IgG) and 0.5 mg ml21 in an end volume of 14 ml ( = 7.0 mg of IgG), respectively. To prepare the detector reagent, 7.5 mg of the affinity purified antibodies were used per millilitre of G40 sol.After coating and stabilization with a casein mixture, the protein sol preparation obtained was centrifuged over a discontinuous glycerol gradient in a 50 ml Greiner tube.The conditions of centrifugation were chosen such that the coated G40 particles appeared entirely in the 50% v/v glycerol layer. In this way, the particles were separated from large and/or aggregated particles that were centrifuged through the 80% v/v glycerol layer on to the bottom of the tube and also from loosely or non-bound proteins (free antibody and casein proteins) that remained in the aqueous upper layer of the gradient.The particles in the 50% v/v glycerol layer were recovered by piercing the centrifuge tube with a syringe needle and removing the dark red layer sideways out of the tube. In this way, 4–5 ml of concentrated G40 sol in 50% v/v glycerol were obtained from each 25 ml of the original sol. For loading on the conjugate pad, the detector reagent had to be washed and further concentrated. This was achieved by four successive washes of the G40 sol with water (pH 8.5) to a 15 ml end volume, each wash followed by a concentration step in a Centriprep-30 concentrator.Removal in this way of (most of) the glycerol from the G40 sol appeared to be essential as, at too high a concentration, the glycerol caused problems in drying the conjugate pads and in running the test, i.e., slow migration of the sol over the membrane and insufficient binding of the test detector reagent at the capture line in the case of a negative sample.Finally, in an additional run the G40 sol was concentrated to a volume of 2 ml. After addition of a bovine milk casein mixture, the sol was stored at 4 °C until further use. By this procedure, the initial sol was concentrated 10-fold. A comparable procedure was applied to produce the control sol. Both G40 sols were mixed in the appropriate ratio, followed by addition of sucrose and surfactant (see Experimental). Sucrose as additive was tested at concentrations of 0.5, 1.0, 2, 3 and 4% m/v respectively.The best results were obtained with 2% m/v sucrose: concentrations lower than 2% m/v resulted in poor solubilization of the detector reagent from the conjugate pad and consequently in less intensively stained capture lines, and at concentrations higher than 2% m/v the viscosity of the detector reagent apparently became too high, resulting in an irregular and slow moving flow front and capture lines that were not stained uniformly.The use of glucose or lactose as additives at concentrations ranging from 0.5 to 5.0% m/v resulted in poor solubilization of the detector reagent from the conjugate pad and consequently in weakly stained capture lines. The influence of the surfactant (S24) concentration in the detector reagent mixture was also tested. It was found that solubilization of the detector reagent from the conjugate pad was optimum at S24 concentrations between 0.02 and 0.05% v/v. Pads made from different materials (6 3 4 mm) were loaded with 12 ml of the detector reagent mixture and dried for 3–4 h at 40 °C.The following materials, present in the BioDot diagnostic materials kit, were tested for use as conjugate pads: rayon, non-woven 5S (Schleicher and Schuell), Cellulostic Fig. 1 Schematic diagram of a test strip showing its several components. A, Top view; B, cross-section. A complete one step strip test device consists of a test strip as shown here, packed in a plastic housing.Fig. 2 Principle of a one step strip test for the detection of low molecular mass analytes. A, Sample is brought on to the sample pad (not shown) and SDD present in the sample is bound by the test detector reagent in the conjugate pad. This pad contains two distinct detector reagents: gold particles coated with specific anti-SDD antibodies (test detector reagent) and gold particles coated with rabbit anti-sheep IgG (control detector reagent). B, Control detector reagent, SDD-bound test detector reagent, free SDD or free test detector reagent migrates up the strip with the sample.C, As the sample passes over the capture zones, test detector reagent (anti-SDD IgG) that is free of analyte binds to the test capture reagent (SDD– ovalbumin conjugate) at test line T, whereas the control detector reagent (rabbit anti-sheep IgG) binds to the control capture reagent (sheep IgG) at control line C. D, Reading the results. Colour develops at the capture line T if the sample contains no SDD (negative sample), whereas no colour is seen if the sample contains SDD (positive sample).Colour develops at the negative control line (C) if the test has been used properly. 2440 Analyst, 1998, 123, 2437–2441Paper 593 (Schleicher and Schuell), Glass Fiber Matrix T5 NM- 10 (Millipore), Glass Fiber Paper with F 075-17 binder (Whatman) and Glass Fiber 8980 (Gelman). Considering the ease with which the conjugate pad was wetted, the capacity of the pad to absorb fluid, the distribution of the detector reagent on the pad and the amount of detector reagent that remained on the pad after running the assay, the Whatman Glass Fiber Paper with F 075-17 binder was chosen as the most appropriate pad for further use.After immobilization of the capture reagents on the nitrocellulose membrane, the reactive groups on the membrane had to be blocked. The effect of several types of blocking buffers on the lateral flow rate of the nitrocellulose strip and on the intensity of the capture line was determined.The best results were obtained with either Blocker BLOTTO in PBS (Pierce, 37526) or with 2% m/v non-fat milk powder supplemented with 0.02% m/v sodium dodecyl sulfate (SDS). In the present study, milk powder was chosen as the blocking agent, mainly because of its low cost. As an excess of blocking agent may dramatically interfere with the membranes capillary flow properties, unadsorbed blocking agent had to be removed by washing the membrane in a weak buffer solution.To promote uniform rewetting of the blocked membrane, the presence of a low concentration of surfactant is essential. Washing of the blocked membrane was performed by swirling the membrane three times for 5 min on a rock and roll platform in a solution of 0.9 g l21 sodium dihydrogenphosphate monohydrate (pH 7.5), supplemented with different surfactants. The following surfactants (all at a concentration of 0.025% m/v) were tested: SDS, Standapol ES-1 (S7), Triton X-100 (S14), Triton X-305 (S15), Tween 20 (S19) and Surfactant 10G (S24).The S-numbers refer to the surfactants present in the BioDot surfactant starter kit. After washing, the excess of buffer was removed from the membranes with a piece of filter-paper. After drying for 2 h at 40 °C, each membrane (25 3 300 mm) was divided into strips (25 3 6 mm). When monitoring the appearance of the test capture line on the strips, the highest staining intensity and best sharpness of the line were obtained with washing buffer containing S24.In an additional experiment, washing buffer with different concentrations of S24 (0.01 , 0.025 and 0.05% v/v) was tested and no significant difference was found. Therefore, the lowest concentration tested, i.e., 0.01% v/v, was applied in the washing buffer in all further experiments. All test ingredients (nitrocellulose, gold conjugate pad and sample and solvent absorbent pads) were pasted on to the backing plate and mounted in the plastic housing.This ready to use strip test device was sealed in a pouch in the presence of a bag of drying agent. The lower detection limit (LDL) is defined here as the amount of SDD in the sample solution that just causes total invisibility of the capture line. The LDL of the test was first determined using different amounts of SDD in 200 ml of PBS. Although the presence of 2.5 ng of SDD per millilitre of PBS already gave a considerable decrease in the assay signal, the LDL was estimated as 10 ng of SDD per millilitre of PBS (10 ppb).Further, five calf urine samples (pH Å 7) and ten bovine urine samples (pH 7.8–8.4) were tested. Also with urine the LDL of the assay was found to be 10 ppb. When running the assay with either PBS or urine samples, test results were visible in approximately 5 min. When undiluted milk samples were tested, it took a long time for a 200 ml sample to be completely absorbed by the sample pad.Consequently, the assay running time was extended to 20–25 min. A sample volume of 100 ml, on the other hand, appeared to be insufficient for running the assay, as concluded from the incomplete solubilization of the detector reagent from the conjugate pad and the abrogated lateral flow on the membrane during the assay performance. The most appropriate sample volume for milk samples in these assays was found to be 150 ml.The assay could then be performed in approximately 10 min. Only when using diluted milk samples (e.g., 1:1 in PBS), could the assay running time be brought back to 5–10 min. For spiked milk samples, the LDL of the assay appeared to be approximately 20 ng of SDD per millilitre of milk. These observations, however, only apply to fresh milk. When milk samples were used that had been cooled, frozen or kept at room temperature for a long time (8 h or longer), the sample pads became silted up with milk-fat, resulting in a considerable redardation to even complete abrogation of the lateral flow. In an attempt to make old milk samples suitable again for use on a strip test, such samples were submitted to several treatments.In doing so, some improvement concerning the assay time was observed when the samples were first heated to 40 °C for about 10 min and then cooled to 20 °C. However, the use of such heated samples also resulted in unacceptably long assay times of 30 min or more. In another attempt to make old milk samples more suitable for being used in the strip test, several surfactants (SDS, S5, S22, S24 and Igepal CA-630) were tested at concentrations of 0.05 and 0.1% m/v or v/v.With none of these surfactants could the assay running time be improved. Dilution of the samples with PBS (1:4 v/v and higher) appeared to be the only method by which the assay time could be shortened to acceptable values (5–10 min). The results of this study demonstrate that the one step strip test works for a veterinary drug at the low ppb level in buffer, urine and diluted milk. However, when using a polyclonal antiserum, the amount of tests to be produced is limited (approximately 2500–3000 tests per millilitre of raw serum). In the future, the work will be continued using monoclonal antibodies and combinations of assays in panel tests will be studied. Acknowledgement The authors thank the Department of Electron Microscopy of IPO-DLO (Wageningen, The Netherlands) for performing the transmission electron microscopy. References 1 J. H. W. Leuvering, P. J. H. M. Thal, M. Van der Waart and A. H. W. M. Schuurs, J. Immunoassay, 1980, 1 (1), 77. 2 Syllabus of a Two-day Seminar on Solid Phase Membrane-Based Immunoassays, Paris, September 25–26th, 1997, Millipore Corporation, Bedford, MA, 1997. 3 Syllabus of The Latex Course, London, October 1–3rd, 1997, organised by Bangs Laboratories, Inc., Fishers, IN, USA, 1997. 4 M. Mullerbardorff, H. Freitag, T. Scheffold, A. Remppis, W. Kubler and H. A. Katus, Circulation, 1995, 92 (10), 2869. 5 M. P. A. Laitinen and M. Vuento, Acta Chem. Scand., 1996, 50, 141. 6 W. Haasnoot, K.-A. Kim, G. Cazemier, W. Kang and J. Kang, in Proceedings of the EuroResidueIII Conference, Veldhoven, May 6–8, 1996, ed. N. Haagsma and A. Ruiter, University of Utrecht, Faculty of Veterinary Medicine, Utrecht, 1996, p. 461. 7 W. Haasnoot, G. O. Korsrud, G. Cazemier, F. Maneval, H. J. Keukens and J. F. M. Nouws, Food Addit. Contam., 1996, 13 (7), 811. 8 G. Frens, Nature (London) Phys. Sci., 1973, 241, 20. 9 M. Horisberger and J. Rosset, J. Histochem. Cytochem., 1977, 25, 295. Paper 8/04695F Analyst, 1998, 123, 2437–2441 2441

ISSN:0003-2654    

DOI:10.1039/a804695f

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

Metabolism of chlortetracycline: drug accumulation and excretion in the hen’s egg† D. Glenn Kennedy,* Robert J. McCracken, S. Armstrong Hewitt and John D. G. McEvoy Veterinary Sciences Division, Department of Agriculture for Northern Ireland, Stoney Road, Stormont, Belfast, UK BT4 3SD. E-mail: glenn.kennedy@dani.gov.uk Received 26th June 1998, Accepted 17th August 1998 Chlortetracycline (CTC) is one of the few antibiotics that can be used without any withdrawal period in chickens laying eggs intended for human consumption. 6-Iso-CTC and 4-epi-6-iso-CTC have recently been identified as the principal metabolites of CTC in eggs. Although not covered by the European Union (EU) maximum residue limit (MRL) for CTC, these compounds, taken together, accumulate in the eggs of birds treated therapeutically with CTC to a mean concentration equivalent to more than twice the EU MRL (200 mg kg21) in eggs. Plateau concentrations in eggs were achieved after approximately 3 d of medication.Following withdrawal of medication, mean egg concentrations of these compounds were maintained for 48 h, before falling below a level equivalent to the MRL after 5 d. Feeds containing typical sub-therapeutic contamination concentrations of CTC did not produce mean concentrations of 6-iso-CTC plus 4-epi-6-iso-CTC, combined, greater than 200 mg kg21. It is not known whether these compounds are formed as a result of metabolism or of chemical degradation. However, analysis of ovules pre-lay showed that all of the CTC present in this matrix was in the form of 6-iso-CTC and 4-epi-6-iso-CTC, and not as the parent drug.Although microbiologically inactive, the toxicological properties of 6-iso-CTC and 4-epi-6-iso-CTC are not known. Introduction Chlortetracycline [CTC, Fig. 1(a)] is a broad spectrum antibacterial, active against a wide range of Gram-positive and Gram-negative bacteria. It was originally isolated from cultures of Streptomyces aureofaciens. In the UK, CTC is licensed for use in feedingstuffs for egg-laying chickens, at a concentration of 300 mg kg21.The withdrawal period for eggs is set at zero days. The chemistry of CTC is complex, the molecule being capable of a wide range of chemical reactions. Under mildly acidic conditions, CTC can reversibly epimerise to form 4-epi- CTC [Fig. 1(b)], a compound which has little antimicrobial activity. The European Union (EU) has established a maximum residue limit (MRL) in a number of foods of animal origin for CTC, which is listed in Annexe I of Council Decision 2377/90.1 For eggs, this is 200 mg kg21, the marker residue being the sum of the concentrations of CTC and 4-epi-CTC.2 Relatively few studies on the kinetics of CTC accumulation and depletion in eggs have been reported.Katz et al.,3 using a microbial growth inhibition assay, reported that daily feeding of CTC to egg-laying birds at 200 mg kg21 for 4 months produced mean egg CTC concentrations of approximately 60 mg kg21.More recently, Roudaut et al.,4 again using a microbial growth inhibition assay, demonstrated that feeding CTC to egg-laying birds at 600 mg kg21 for 7 d produced mean egg CTC concentrations of 190 mg kg21. A recent report from this laboratory5 has studied the accumulation of CTC residues in eggs using liquid chromatography- atmospheric pressure chemical ionisation mass spectrometry (LC-APCI-MS). This study confirmed that neither CTC nor 4-epi-CTC occurred in eggs laid by CTC-treated birds to any significant extent.However, we showed that two other compounds, 6-iso-CTC [Fig. 1(c)] and 4-epi-6-iso-CTC, occurred in eggs at combined concentrations equivalent to approximately twice the MRL. These compounds, which also have minimal antibacterial activity (unpublished observations), have not previously been described as significant metabolites of CTC. CTC is the commonest antibacterial agent detected in animal feedingstuffs in Northern Ireland.6 Of 408 animal feeds tested, 101 were found to contain at least one undeclared antibacterial agent.CTC was present in 62 of these 101 contaminated feeds. Most of the CTC-containing feeds contained less than 20% of the full therapeutic dose. On occasions, however, ‘unmedicated’ feeds have been found to contain CTC at concentrations of up to 450 mg kg21. The present study was initiated to examine the rate of accumulation and depletion of 6-iso-CTC and 4-epi-6-iso-CTC in eggs collected from egg-laying chickens treated with a therapeutic dose of CTC.We also examined the accumulation of total CTC residues in eggs collected from birds fed levels of CTC that were typical of those found in cases of feed contamination. In addition, the chemical identity of the CTC species present in the developing ovule is reported. Experimental Determination of CTC in feedingstuffs CTC in animal feedingstuffs was determined using a modification of an assay originally used for the determination of CTC in animal tissues.7 Briefly, milled feed (50 g) was extracted for 1 h by shaking with methanol–HCl (98 + 2, v/v).After settling, an aliquot (10 ml) was diluted 10-fold with methanol. An aliquot (2.5 ml) was then diluted to 50 ml in HPLC mobile phase7 and allowed to stand for 3 h prior to analysis as described previously.7 Using this method, the mean recovery of CTC (25–300 mg kg21) from control meals prepared and stored at 220 °C ranges from 64 to 70%.† Presented at the Third International Symposium on Hormone and Veterinary Drug Residue Analysis, Bruges, Belgium, June 2–5, 1998. Analyst, 1998, 123, 2443–2447 2443LC-APCI-MS analysis of CTC in eggs This was carried out using a VG Platform LC-APCI-MS system, essentially as described in an earlier report from this laboratory by Blanchflower et al.8 Fig. 1 shows the positive ion LC-APCI mass spectra obtained for CTC, 4-epi-CTC and 6-iso- CTC.Both CTC and 4-epi-CTC show the molecular ion [M+H]+ at m/z 479, as well as ions corresponding to the successive loss of ammonia and water at m/z 462 and 444, respectively. In contrast, 6-iso-CTC forms only two main ions at m/z 479 and 462. Any CTC and 4-epi-CTC present in eggs were quantified by comparison with a freshly prepared standard of CTC. Any 6-iso-CTC and 4-epi-6-iso-CTC present in eggs were quantified by comparison with a freshly prepared 6-iso- CTC standard.This was done to minimise difficulties caused by the in vitro epimerisation of 4-epi-CTC solutions (very rapid) and 6-iso-CTC (slow). Study 1. Accumulation and depletion of CTC residues in eggs from layers fed a therapeutic dose of CTC Nine layer chickens in mid-lay (approximately 35 weeks old) were purchased from a local supplier. The birds were individually housed on wire floors and were allowed ad libitum access to fresh water from nipple drinkers. The birds were maintained on a repeating cycle of 4 h dark and 4 h light.Once daily they were fed 120 g each of a finely milled, commercially prepared layer ration, that had been shown to be CTC-free. This acclimatisation period lasted for 1 week. At this point, eggs were tested and found to be CTC-free. The birds were then fed 120 g per head per day of the same ration supplemented with CTC (Sigma, Poole, Dorset, UK) to give a final concentration of 300 mg of CTC per kilogram of diet.The CTC supplied by Sigma contained no iso-CTC. Once daily for 9 d, eggs were collected, homogenised and stored at 220 °C until analysis. At the end of this period, six birds were transferred to clean cages and fed the CTC-free ration for a further 9 d. Eggs were collected and treated as described above. The remaining three birds were killed by asphyxia using carbon dioxide, without any withdrawal period, to examine the chemical form of the CTC residues present in ovules.Their abdomens were opened; eggs and ovules were removed, individually weighed and placed in descending weight order for each bird before homogenisation and storage as described above. For each bird, they were assigned an ‘ovule number’ in which increasing ovule number corresponds to decreasing ovule mass. Study 2. Accumulation of CTC residues in eggs from layers fed CTC at concentrations typical of feed contamination CTC was added to a commercially available layer ration to give a final concentration of 450 mg kg21.This was then diluted appropriately with CTC-free feed to give a series of feeds containing CTC at 450, 300, 50, 30 and 10 mg kg21. Thirty layers (five groups of six birds each) were purchased, housed and acclimatised to their conditions as described above. Each group was then fed one of the experimental diets at a rate of 120 g per head per day for 9 d. Eggs were collected and treated as described above. Results LC-APCI-MS analysis of CTC in eggs Under the conditions used, the CTC standard [Fig. 2, Panel (A)] is relatively stable. Only one peak (Å10.1 min) is normally Fig. 1 Chemical structures and positive ion LC-APCI-MS scans of chlortetracycline (a); 4-epi-chlortetracycline (b); and 6-iso-chlortetracycline (c). 2444 Analyst, 1998, 123, 2443–2447present. 4-Epi-CTC is, however, relatively unstable. Within 1 h of preparation, peaks corresponding to keto-4-epi-CTC8 (Å7.7 min) and CTC (Å10.1 min) are present in the standard in addition to 4-epi-CTC (Å9.1 min) [Panel (B)].Freshly prepared 6-iso-CTC standards contained only trace amounts of 4-epi- 6-iso-CTC, or none at all [Panel (C)]. After several hours, the 6-iso-CTC standard (Å8.2 min) [Panel (C)] shows some epimerisation to 4-epi-6-iso-CTC (Å7.4 min) [Panel (D)]. One trace [Panel (E)] of an incurred egg is shown. Only 6-iso-CTC and 4-epi-6-iso-CTC were quantifiable in this egg. A trace amount of CTC was present (10.1 min), but at a concentration that was too low to permit quantification.Accumulation of CTC metabolites in eggs during medication Neither CTC nor 4-epi-CTC was detected in the eggs of birds fed a therapeutic (300 mg kg21 diet) dose of CTC (Fig. 3). However, substantial concentrations of 6-iso-CTC and 4-epi- 6-iso-CTC were detected. The concentrations of these metabolites rose rapidly, and reached a plateau level after approximately 4 d of medication with CTC. The proportion of the drug that was present as 6-iso-CTC was slightly higher than that in the form of 4-epi-6-iso-CTC for the first 5 d of treatment.However, thereafter, the ratio of 6-iso-CTC to 4-epi-6-iso-CTC was approximately 1 : 1. The mean concentration of both metabolites, combined, in eggs exceeded 200 mg kg21 from day 3 onwards. Depletion of CTC metabolites from eggs following drug withdrawal The mean concentration of (6-iso-CTC plus 4-epi-6-iso-CTC) in eggs did not change during the first 48 h following withdrawal of the drug (Fig. 4). Thereafter, mean (6-iso-CTC plus 4-epi-6-iso-CTC) concentrations fell, falling below 200 Fig. 2 Selected ion monitoring traces at m/z 479, 462 and 444 for a CTC standard (A); a 4-epi-CTC standard, aged for approximately 2 h (B); a fresh 6-iso-CTC standard (C); an iso-CTC standard, aged for approximately 6 h (D); and an incurred positive egg (E). All standards were prepared to contain the equivalent of 500 mg kg21. Fig. 3 Accumulation of CTC residues in eggs of layer chickens (n = 6) fed 300 mg CTC per kg diet.Whole egg concentrations of 6-iso-CTC (8) and 4-epi-6-iso-CTC (-) and their arithmetic sum (5) were measured using LC-APCI-MS. Data represent means ± standard error of means. Fig. 4 Depletion of CTC residues from eggs of layer chickens (n = 6), following withdrawal of medication (300 mg kg21). Whole egg concentrations of 6-iso-CTC (8) and 4-epi-6-iso-CTC (-) and their arithmetic sum (5) were measured using LC-APCI-MS.Data represent means ± standard error of means. Analyst, 1998, 123, 2443–2447 2445mg kg21 6 d after withdrawal. Both metabolites were still detectable in eggs 9 d after withdrawal of the drug. The ratio of 6-iso-CTC to 4-epi-6-iso-CTC remained approximately 1 : 1 during the depletion period. CTC metabolites in eggs from birds fed contamination levels of CTC Exposure of egg-laying chickens to sub-therapeutic concentrations of CTC, typically found in antibiotic-contaminated feeds, resulted in mean plateau (6-iso-CTC plus 4-epi-6-iso-CTC) concentrations of less than 200 mg kg21 (Fig. 5). However, feeding birds a supra-therapeutic dose of CTC (450 mg kg21) resulted in a mean (6-iso-CTC plus 4-epi-6-iso-CTC) concentration in eggs of approximately 900 mg kg21. There was a strong correlation between the dietary CTC concentration and the mean (6-iso-CTC plus 4-epi-6-iso-CTC) concentration in eggs. The equation of the line-of-best fit was: Egg concentration (mg kg21) = 1.85 3 meal CTC concentration (mg kg21) + 7.7 (r = 0.999) Chemical form of CTC in ovules from CTC-treated layers LC-APCI-MS was used to measure the distribution of CTC amongst its various chemical forms in ovules (pre-lay eggs) from chickens fed a medicated feed that contained CTC at a concentration of 300 mg kg21.In all of the ovules studied, neither CTC nor 4-epi-CTC was detectable (Fig. 6). Only the iso metabolites were present. Furthermore, the mean (6-iso- CTC plus 4-epi-6-iso-CTC) concentration was similar in all of the ovules. Discussion LC-APCI-MS analysis of CTC in eggs The present study has confirmed our earlier report5 on the occurrence of substantial amounts of 6-iso-CTC and 4-epi- 6-iso-CTC in the eggs of chickens treated with a therapeutic dose of CTC. Although the parent drug and its 4-epimer were present in many egg extracts in the present study, they were present at concentrations that precluded quantification. This is in contrast to our previous study,5 in which all four compounds could be quantified.However, in that study, the measured concentrations of CTC and 4-epi-CTC were very low (approximately 10% of their corresponding iso derivatives). It is possible that a slight variation in the detection limit of the LCMS system between the two occasions could account for the difference between the two studies. Fig. 3 shows that both metabolites accumulated very rapidly in eggs.They were detectable 1 d after commencing feeding of a medicated diet to layers. Plateau concentrations (Å500 mg kg21) were achieved after approximately 3 d. Given that 6-iso-CTC concentrations were higher than those of its 4-epimer in the early stages of accumulation, it is possible that it is the primary metabolite, with epimerisation to 4-epi-6-iso- CTC occurring more slowly. Following withdrawal of medication, mean (6-iso-CTC plus 4-epi-6-iso-CTC) concentrations remained at approximately 500 mg kg21 for 48 h before declining (Fig. 4). They were still present at combined concentrations equivalent to the EU MRL after 5 d, and were still detectable 9 d after cessation of treatment. We have previously shown that inadvertent carry-over of medication from batches of medicated feed to subsequent, ostensibly unmedicated feeds can cause residues of lasalocid9 and monensin and salinomycin in eggs.10 The same has now been shown to be true for CTC in eggs.CTC, the commonest antibacterial contaminant found in unmedicated feedingstuffs in Northern Ireland, can cause accumulation of (6-iso-CTC plus 4-epi-6-iso-CTC) residues in eggs. However, the mean plateau (6-iso-CTC plus 4-epi-6-iso-CTC) concentrations in eggs from birds fed sub-therapeutic CTC concentrations were well below 200 mg kg21. The current EU marker residue for CTC is the sum of the concentrations of CTC and 4-epi-CTC.2 Therefore, neither 6-iso-CTC nor its 4-epimer counts towards the MRL. Since the existence of these compounds as major metabolites in eggs has only been reported recently by this laboratory,5 there has, as yet, been no opportunity to assess the need for a re-examination of either the MRL in eggs or the withdrawal period for this drug in laying hens.Until the publication of that report, 6-iso-CTC had been described only as a very minor metabolite in canine urine and faeces.11 Most previous studies of CTC accumulation in eggs have relied on microbial growth inhibition tests.3,4 These tests would not have detected 6-iso-CTC and its 4-epimer, because they have no antibacterial activity. Other workers who may have analysed eggs using HPLC assays that relied on UV Fig. 5 Relationship between the mean plateau concentration of (6-iso- CTC plus 4-epi-6-iso-CTC) in whole egg homogenates (defined as the mean concentration measured on days 7, 8 and 9 after the start of the experiment) and the dietary concentration of CTC.Egg residue concentrations were measured using LC-APCI-MS, and dietary CTC levels by HPLC with fluorescence detection. Data represent means ± standard error of means. Fig. 6 Chemical form of CTC residues in ovules dissected from three layer chickens following the administration of a therapeutic dose of CTC (300 mg kg21) for 9 d. Concentrations of 6-iso-CTC (8) and 4-epi-6-iso- CTC (-) and their arithmetic sum (5) were measured using LC-APCI-MS. Increasing ovule number is equivalent to decreasing ovule mass.The data below the ovule number are the mean ovule masses in grams. All data represent means ± standard error of means. 2446 Analyst, 1998, 123, 2443–2447absorption of CTC in an acidic mobile phase12 would not have detected these metabolites because, under those conditions, they exhibit no UV absorbance.5 Other assays, relying on the formation of metal complexes either for the purposes of sample clean-up13 or detection,14 will also not detect these metabolites.This is because the metal binding site in the iso-tetracyclines (at C11 and C12) is altered, and the iso-tetracyclines do not bind metal ions (see Discussion in ref. 5). However, the present study, and our previous report,5 clearly demonstrate that hitherto undiscovered metabolites of CTC may be present in eggs at concentrations equivalent to more than twice the current EU MRL. To the best of our knowledge, no studies concerning the toxicology of 6-iso-CTC have been conducted. Such studies are now clearly required to assess the potential significance of these residues to the consumer.Preliminary studies carried out at this laboratory have shown that the iso derivatives of CTC are relatively minor metabolites in pig and poultry tissues. These findings will be published elsewhere. It is still unclear if these compounds arise as the result of enzymatic metabolism or chemical degradation. One possibility, that formation of the iso metabolites of CTC occurs during the process of shell formation, has been eliminated. Fig. 6 shows that, in all of the ovules studied, the only detectable CTC metabolites were 6-iso-CTC and its 4-epimer. 6-Iso-CTC forms readily from CTC under alkaline conditions.15 The developing ovule (yolk) has a pH of approximately 6.5. These conditions will not cause the chemical conversion of CTC to iso-CTC. Egg albumen, on the other hand, can be alkaline (pH 7.6 in the fresh egg, rising to 9.2 over several days).However, since albumen is laid down shortly before shelling and laying, it cannot contribute to the formation of iso-CTC in pre-existing ovules which have no albumen. The concentration of iso-CTC and its 4-epimer was similar in all of the ovules examined. This may result from a rough equality between the rate of drug accumulation in tissue and the rate of growth of the developing yolk. A further unresolved question is whether or not oxytetracycline forms similar metabolites in eggs.CTC is especially prone to form the phthalide iso-CTC following exposure to alkali.15 However, all tetracyclines that possess a C6 hydroxyl group can readily form these derivatives.15 Oxytetracycline is therefore no exception. It has, however, long been accepted that oxytetracycline is relatively resistant to such changes. Despite this, a number of analytical methods for oxytetracycline have been described16,17 that rely on alkali-induced fluorescence as the detection step.Further studies to examine the possibility that iso-oxytetracycline accumulates in the eggs of treated layers are planned and will also be reported elsewhere. In conclusion, we have shown that (6-iso-CTC plus 4-epi- 6-iso-CTC) accumulate rapidly, to concentrations equivalent to more than twice the current EU MRL, in the eggs of layer chickens treated with a therapeutic dose of CTC. We have also shown that these metabolites persist at mean concentrations above 200 mg kg21 for 5 d following withdrawal of medication.Low-level contamination of feedingstuffs with CTC is unlikely to produce mean (6-iso-CTC plus 4-epi-6-iso-CTC) concentrations in excess of 200 mg kg21. Although possessing no antibacterial action, the toxicological properties of these compounds are not known. References 1 Off. J. Eur. Commun., 1990, L224, 1. 2 Off. J. Eur. Commun., 1996, L37, 9. 3 S. E. Katz, C. A. Fassbender and J. J. Dowling, Assoc. Off. Anal. Chem., 1972, 55, 128. 4 B. Roudaut, J. P. Moretain and J. Boisseau, Food Addit. Contam., 1989, 6, 71. 5 D. G. Kennedy, R. J. McCracken, M. P. Carey, W. J. Blanchflower and S. A. Hewitt, J. Chromatogr. A, 1998, 812, 327. 6 L. Lynas, D. Currie, W. J. McCaughey, J. D. G. McEvoy and D. G. Kennedy, Food Addit. Contam., 1998, 15, 162. 7 W. J. Blanchflower, R. J. McCracken and D. A. Rice, Analyst, 1989, 114, 421. 8 W. J. Blanchflower, R. J. McCracken, S. A. Haggan and D. G. Kennedy, J. Chromatogr. B, 1997, 692, 351. 9 D. G. Kennedy, W. J. Blanchflower, P. J. Hughes and W. J. McCaughey, Food Addit. Contam., 1996, 13, 787. 10 D. G. Kennedy, P. J. Hughes and W. J. Blanchflower, Food Addit. Contam., 1998, 15, 535. 11 H. J. Eisner and R. J. Wulf, J. Pharmacol. Exp. Ther., 1963, 142, 122. 12 H. Oka, H. Matsumoto, K. Uno, K.-I. Harada, S. Kadowaki and M. Suzuki, J. Chromatogr., 1985, 325, 265. 13 W. H. H. Farrington, J. Tarbin, J. Bygrave and G. Shearer, Food Addit. Contam., 1991, 8, 55. 14 R. J. McCracken, W. J. Blanchflower, S. A. Haggan and D. G. Kennedy, Analyst, 1995, 120, 1763. 15 L. A. Mitscher, The Chemistry of the Tetracycline Antibiotics, Marcel Dekker, New York, 1978, p. 123. 16 T. Agasøster and K. E. Rasmussen, J. Pharm. Biomed. Anal., 1992, 10, 349. 17 S. Horii, J. Liq. Chromatogr., 1994, 17, 213. Paper 8/04913K Analyst, 1998, 123, 2443–2447 2447

ISSN:0003-2654    

DOI:10.1039/a804913k

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

Metabolites in feces can be important markers for the abuse of anabolic steroids in cattle† Mark Van Puymbroeck,*ab Luc Leyssens,a Dirk Vanderzande,b Jan Gelanb and Jef Rausab a Dr. L. Willems-Instituut, Department of Drug and Residue Analysis, B-3590 Diepenbeek, Belgium b Limburgs Universitair Centrum, Department SBG, B-3590 Diepenbeek, Belgium Received 30th June 1998, Accepted 7th September 1998 In Belgium, to control the abuse of anabolic steroids in cattle, urine samples have been gradually replaced by feces samples, because the latter can be obtained more easily from living animals.Urine and feces samples were collected from heifers after administration of boldenone, norethandrolone or ethylestrenol. Metabolites present in feces or urine were determined by GC-MS. Large qualitative and quantitative differences in the metabolic profiles were observed. In feces, in contrast to urine, the parent compounds or their major metabolites were detectable only shortly after administration.On the other hand, metabolites resulting from the reduction of the 3-oxo group and the unsaturated carbon–carbon bonds, present on the A-ring, allow for long-term detection in feces. A-ring reduced metabolites have been identified in samples found positive for norgestrel, boldenone, methylboldenone and methyltestosterone, respectively. These results are in agreement with concomitant in vivo experiments. Introduction Several matrices, such as meat, injection sites, kidney fat, urine and feces, have been used to screen for the illegal use of anabolic steroids in cattle.The importance of fecal matter has increased in recent years in Belgium, because the sampling is easy to perform.1 However, since fecal matter is heterogeneous and hydrophobic, the isolation of steroids from feces is more cumbersome than isolation from an aqueous urine sample. In addition, extensive metabolic activity in the gastrointestinal tract is caused by bacteria.As a consequence, the excretion of metabolites in feces can be different from that in urine.2,3 Steroids that are excreted in the gastrointestinal tract may be reabsorbed and subsequently metabolized. Hydrolysis, esterification, dehydroxylation, hydrogenation, dehydrogenation and side chain cleavage have been reported as possible anaerobic transformations of steroids and sterols by intestinal bacteria of humans and other mammals.2,3 In feces, steroid hormones conjugated with sulfuric acid or glucuronic acid, mainly at the 3a-, 3b-, 17b- and 21-positions, can be hydrolyzed by microorganisms.Hydrolysis is achieved mainly through the metabolic activities of Bacteroides, Escherichia coli and Clostridium spp.2 It is therefore generally accepted that an enzymatic hydrolysis step in steroid analysis of fecal material is of no use.1 The possibility of additional metabolic conversions by microorganisms and the scarce data on the excretion of anabolic steroids in feces gave rise to the present study.The metabolite excretion profiles in urine and feces samples obtained from animals treated with ethylestrenol (EES), norethandrolone (NE) and boldenone (BOL) were evaluated. The presence of metabolic conversion products with a reduced A-ring seemed to be of great importance for the long-term detection of treatment with EES, NE and BOL. The significance of the results obtained was evaluated using field samples found positive for NE, norgestrel (NG), methyltestosterone (MT) or methylboldenone (MeBOL).Experimental Materials and standards All reagents and solvents were of analytical-reagent grade. Milli-Q water was obtained from a Milli RO water purification system (Waters, Bedford, MA, USA). HPLC-grade methanol was obtained from Labscan (Dublin, Ireland) and diethyl ether from Merck (Darmstadt, Germany). Bond-Elut silica 3 ml, Bond-Elut amino 1 ml and Chem-ELut 20 ml columns were supplied by Varian (Harbor City, CA, USA).All anabolic steroids were obtained from the Belgian National Reference Laboratory (Wetenschappelijk Instituut voor Volksgezondheid –Louis Pasteur, Brussels, Belgium). All reagents used for the chemical synthesis of metabolites were purchased from Aldrich (Milwaukee, WI, USA). Norgestrel positive routine feces samples were obtained from Dr. Delahaut (Centre d Economie Rurale, Division Hormonologie Animale, Marloie, Belgium), methyltestosterone and methylboldenone positive feces samples from Dr.D. Courtheyn (Rijksontledingslaboratorium Gent, ROLG, Ghent, Belgium) and the NE positive routine feces samples from our laboratory. Experimental animal studies A 36 mg amount of EES was administered intramuscularly to a 12 month old heifer (about 100 kg) and 250 mg of NE to a 15 month old heifer (about 300 kg). A 50 mg amount of EES was dissolved in 10 ml and 250 mg of NE in 12 ml of a 1+1 v/v mixture of pharmaceutical grade Miglyol and N-methyl- 2-pyrrolidone (Sigma, St.Louis, MO, USA). Urine and feces samples from a heifer (625 kg) treated intramuscularly with 700 mg of 17b-boldenone were provided by Professor H. De † Presented at the Third International Symposium on Hormone and Veterinary Drug Residue Analysis, Bruges, Belgium, June 2–5, 1998. ‡ Predoctoral Fellow of the LUC Universiteitsfonds. Analyst, 1998, 123, 2449–2452 2449Brabander (State University of Ghent, Faculty of Veterinary Medicine, Ghent, Belgium).Small scale syntheses of metabolites The selective reduction of the double bond at position 4 to form the 5b,3a-ol isomers and 5b-androst-1-ene-3,17-dione on a milligram scale was performed following the method described by Schänzer and Donike.4 To establish basic reaction conditions, methanol–6 mol l21 sodium hydroxide (20+1 v/v) was used as the solvent for selective reduction with hydrogen. Hydrogen was generated in situ, by slowly adding 12 mol l21 concentrated hydrochloric acid to an aqueous solution of 12% m/v sodium tetrahydroborate.To stabilize the sodium tetrahydroborate solution, 0.2 mol l21 sodium hydroxide was added. Palladium (10%) on charcoal was used as a catalyst. After 30 min of vigorous stirring at room temperature, the reaction mixture was diluted with water and extracted with tert-butyl methyl ether. The organic layers were evaporated to dryness. Further reduction of the 3-oxo function was performed with aluminium lithium hydride.Tetrahydronorgestrel was synthesized according to Hübner et al.5 The selective hydroxylation of boldenone and androsta-1,4-diene-3,17-dione was performed according to Schänzer and Donike.6 Sample preparation The extraction method of urine samples has been described elsewhere.7 A 20 g amount of feces was extracted for 2 h with 50 ml of a mixture of diethyl ether and Milli-Q water (4+1 v/v), followed by 1 h with 50 ml of diethyl ether. Equilinin (500 ng) was added as internal standard before extraction.To remove solids, the extracts were centrifuged for 10 min at 1500 g and poured over a cotton-wool plug in a round-bottomed flask. Diethyl ether was removed under reduced pressure at 60 °C and the remainder was dissolved in 8 ml of methanol. A 4 ml volume of 0.33 mol l21 phosphoric acid solution was added to eliminate fat. This solution was filtered before pouring it on a Chem-Elut column. The column was washed with hexane and eluted with dichloromethane. Subsequently, the dichloromethane was removed under reduced pressure and the residue dissolved in 600 ml of chloroform.A 6 ml volume of hexane was added to the solution for a further clean-up of the extracts on a Bond-Elut silica column on top of an amino column, as described for urine samples.7 These extracts were fractionated by HPLC on a modular system equipped with a Model 231 high-pressure pump (Spectra-Physics, San Jose, CA, USA) with a Rheodyne (Cotati, CA, USA) Model 7010 injection valve, a Model 2151 variable wavelength monitor (LKB, Bromma, Sweden) operating at 244 nm, a Model 202 fraction collector (Gilson, Villiersle- Bel, France) and a Chromatopack CR-1B recorder (Shimadzu, Kyoto, Japan).The HPLC column was a 250 3 10 mm i.d. stainless-steel tube packed with 5 mm silica (ODS Ultrasphere) (Beckman, Fullerton, CA, USA.). Methanol–water (73+23 v/v) was used as the mobile phase for isocratic elution at a flow rate of 2.5 ml min21.The collector was operated in the time mode. Eight fractions were collected with time windows set relative to naphthalene as internal standard, ranging from 0.46 to 0.61 (fraction I), 0.61 to 0.73 (fraction II), 0.73 to 0.86 (fraction III), 0.86 to 1.04 (fraction IV), 1.04 to 1.24 (fraction V), 1.24 to 1.44 (fraction VI), 1.44 to 1.65 (fraction VII) and 1.65 to 1.85 (fraction VIII). All fractions were dried under a stream of nitrogen at 60 °C, dissolved in 500 ml of methanol and injected separately.Aliquots of 250 ml from the 500 ml solutions were dried in GC-MS vials and derivatized as ethoxime trimethylsilyl (EO-TMS) derivatives. Gas chromatography-mass spectrometry The GC-MS analyses were performed on a Finnigan (San Jose, CA, USA) GCQ system or a Varian (Walnut Creek, CA, USA) Saturn I system. The GC column used was a DB-5 MS 30 m3 0.32 mm i.d. fused silica column with a 0.25 mm film thickness (J & W Scientific, Folsom, CA, USA).The temperature settings were as follows: injector 260 °C, transfer line 275 °C, ion source 200 °C and oven program, initial temperature 50 °C for 20 s, increased from 50 to 190 °C at 50 °C min21 and from 190 to 320 °C at 4.6 °C min21, held at 320 °C for 4 min. The carrier gas was helium at a linear velocity of 60 cm s21 for the GCQ and a flow rate of 1 ml min21 for the Saturn I. Aliquots of 1 ml were injected using a split–splitless injector in the split mode with a splitting ratio of 1 : 10.High-performance thin layer chromatography (HPTLC) For the HPTLC screening of routine samples, the 4 3 4 mode according to De Brabander et al.8 was used. Briefly, 8 ml of the HPLC fractions were applied to a 10 3 10 cm silica gel 60 plate (Merck). Chromatography was carried out in two dimensions using chloroform–acetone (27+3 v/v) as solvent I and cyclohexane –ethyl acetate–ethanol (18+12+0.75, v/v) as solvent II. For the staining and detection of NE, a 5% sulfuric acid solution in acetic anhydride was used.9 Results and discussion Comparison of in vivo excretion profiles of NE, EES and BOL in urine and feces NE, a major intermediate in the metabolism of EES,10 was not found during the first 3 d after treatment with EES.During a period of 3 d after the lag period, NE was detected in urine at levels below 2 ng ml21, whereas feces samples remained negative. However, 17a-ethyl-5b-estrane-3a,17b-diol (EED) was found in both urine and feces.Semi-quantitative results for NE and EED (Fig. 1) showed that EED is the only conversion product present in feces that points to the administration of EES. Unaltered EES was not detectable in either urine or feces. Fig. 1 Excretion profile of EED and NE in (a) urine and (b) feces after intramuscular injection of 36 mg of EES into a 12 month old heifer. 2450 Analyst, 1998, 123, 2449–2452After administration of NE, large amounts of the parent compound were excreted in urine during the first week.In feces, the presence of NE was present up to 36 h after NE administration (Fig. 2). Again, EED levels exceded those of NE in both urine and feces (Fig. 2). The large amounts of EED excreted via the intestinal route prolonged the detection period for NE treatment for almost a week. No major oxidation products were found in urine and feces. After administering 17b-BOL to an animal, 17a-boldenone (17a-BOL) was among the most prominent metabolites present in urine (Fig. 3). Additional metabolites, previously also identified in human urine,6 such as androsta-1,4-diene- 3,17-dione (ADD), 6-hydroxyandrosta-1,4-diene-3,17-dione (6-HO-ADD), 6-hydroxy-17b-boldenone (6-HO-BOL) and 5bandrostene- 3,17-dione (5b-AED) were found. The identity of ADD was confirmed by comparison with a commercially available reference standard. Six other metabolites, tentatively identified in cattle urine as isomers of either 5x-androst-1-en- 3x-ol-17-one or 5x-androst-1-en-3-on-17x-ol (5x-AE) were also observed.Relatively low concentrations of 17a-BOL were found in feces, as compared with urine, obtained from the animal treated with 17b-BOL (Fig. 3). None of the above mentioned hydroxylation or oxidation products of 17b-BOL were detected in fecal matter. Only the reduced metabolites of either BOL or ADD were detected. After application of 17b-BOL, small amounts of 5b-AED were found for a limited period only.Two isomers of the six 5x-AE found in urine were identified. The isomer of 5x-AE that elutes in HPLC fraction VII was detected beyond 17a-BOL. Complete reduction of the A-ring by reducing the two double bonds and 3-oxo function of ADD and BOL can result in the formation of stereoisomers of androstanolones and androstanediols. The amounts of some endogenous androstanolones and androstanediols increased far above the endogenous levels present in blank samples collected before administering 17b- BOL (Fig. 4). The amounts of stereoisomers of androstanolones and androstanediols were much higher than in blank urine samples. These in vivo experiments demonstrate the potential of a screening for the A-ring reduced metabolites of anabolic steroids in feces. A-ring reduced metabolites as biological markers in routine residue analysis In feces samples from one batch of cattle all from the same owner, NE itself was identified. As a result of this positive screening, 175 additional feces samples were collected at the suspected farm, and 7.5% of these samples were found positive for NE by means of HPTLC and GC-MS.The NE-positive samples, and 80 randomly chosen NE-negative samples, were analyzed by GC-MS and revealed the presence of EED. All NE positive feces samples were found to be positive for EED also and the EED levels always exceeded the NE levels. Therefore, by accepting EED as a biological marker for NE abuse, 77% of the NE-negative feces samples were found positive for EED.Routine samples found positive for synthetic anabolic steroids such as norgestrel (NG), methyltestosterone (MT) and methylboldenone (MeBOL) were analyzed for the presence of metabolic conversion products, analogous to those found after treatment with EES and NE. All samples, analyzed in the full scan mode, were screened for a series of predicted fragment ions, originating from theoretically possible, either reduced or oxidized, derivatives of the parent compounds.The results indicated the presence of tetrahydro metabolites at significantly higher levels than the parent compounds, which were present in only trace amounts. Structure elucidation of the tetrahydro metabolites was based on mass spectral data, shown in Fig. 5. The absence of an oxo group was demonstrated by the fact that the spectra were not affected by the use of an oximation reagent. Fig. 2 Excretion profile of EED and NE in (a) urine and (b) feces after intramuscular injection of 250 mg of NE into a 15 month old heifer.Fig. 3 Excretion profile of 17b-BOL and 17a-BOL in (a) urine and (b) feces after intramuscular injection of 700 mg of 17b-boldenone. Fig. 4 Increased presence of androstanediols and androstanolones in feces samples collected after treatment with 17b-BOL. For each compound one isomer was chosen as an example. BL = blank feces sample before application of 17b-diol. Ratio = ratio relative to internal standard (5 ng ml21). Analyst, 1998, 123, 2449–2452 2451Two fragments, i.e.[M+290] and [M+2180], correspond to the consecutive loss of two trimethylsilylhydroxyl moieties (TMSOH) from the molecular ions. This loss generates ions at m/z 370 and 280 in the mass spectrum of the tetrahydrometabolite of NG (Mr 460) and at m/z 360 and 270 in the spectrum of the tetrahydro-metabolite of methyltestosterone (Mr 450). Methylandrostanediol-TMS-2 shows a base peak at m/z 435 [M+215] and prominent peaks at m/z 345 [M+2105] and 255 [M+2195].The last two originate from the loss of a methyl and one and two TMSOH groups, respectively. The ion at m/z 143 is characteristic of the intact D-ring carrying a 17-methyl group. For tetrahydronorgestrel, the characteristic loss of the ethyl group at position 18 generates a base peak at m/z 431 [M+229]. Ions at m/z 341 [M+2119] and 251 [M+2209] are generated by further cleavage of one and two TMSOH groups, respectively. The stereoisomerism of the tetrahydro metabolites was clarified further. The compounds were synthesized by chemical reduction of NG and MT with NaBH4 in aqueous pyridine.This reaction is selective towards the 3a,5b- and 3b,5a-isomers, while 3a,5a- and 3b,5b-isomers are only minor reaction products. No reduction of the 17a-ethynyl group of NG occurs by this method. 3b,5a-Methylandrostanediol was commercially available but was not identical with the product found in MTand MeBOL-positive samples.Therefore, it was concluded that the latter was identical with 3a,5b-methylandrostanediol (MeAD), which was one of major isomers formed by reduction of MT. The exact stereoisomeric configuration of tetrahydronorgestel was postulated on the basis of existing chromatographic data, describing the relationship between configuration and relative retention times, for isomers of estranediol, 17a-ethylestranediol and tetrahydro-3-ketodesorgestrel.2,6,11,12 Based on these data, the tetrahydro metabolite of NG was identified as 3a,5b-tetrahydronorgestrel (THNG).From in vivo experiments and from the analysis of positive routine samples, involving four structurally different synthetic anabolic steroids, the potential of metabolite screening and, in particular, the potential of screening for the A-ring reduced metabolites of anabolic steroids in feces was evaluated. The present study revealed large differences between the excretion of several metabolites in urine and feces. In feces, the reduced forms, originating from a reduction of the 3-oxo function and the unsaturated bonds, at either position 4 or 1 of the corresponding steroids, are predominantly present.In contrast, in urine either the parent compound or their corresponding hydroxylated or reduced metabolites were detected. The possibility of screening for the abusive application of synthetic anabolic steroids by means of feces analysis is significantly improved by the monitoring of matrix specific metabolites.References 1 T. Hamoir, D. Courtheyn, H. De Brabander, P. Delahaut, F. Smets, L. Hendriks and G. Pottie, in Proceedings of Euroresidue III, Veldhoven, 6–8 May 1996, ed. N. Haagsma and A. Ruiter, distributed by: Department of Food of Animal Origin, Section of Food Chemistry, Faculty of Veterinary Medicine, University of Utrecht, Utrecht, The Netherlands, 1996, p. 471. 2 H. Groh, K. Schade and C. Hörhold-Schübert, J. Basic Microbiol., 1993, 33, 59. 3 J. Winter, C. H. L. Shackleton, S. O’Rourke and V. D. Bokkenheuser, J. Steroid Biochem., 1984, 21, 563. 4 W. Schänzer and M. Donike, Anal. Chim. Acta, 1993, 275, 23. 5 M. Hübner, I. Füssel and K. Ponsveld, Pharmazie, 1984, 39, 462. 6 W. Schänzer and M. Donike, Biol. Mass Spectrom., 1992, 21, 3. 7 L. Leyssens, E. Royackers, B. Gielen, M. Missotten, J. Schoofs, J. Czech, J. P. Noben, L. Hendriks and J. Raus, J. Chromatogr. B, 1994, 654, 43. 8 H. F. De Brabander, F. Smets and G. Pottie, J. Planar. Chromatogr., 1988, 1, 369. 9 R. Verbeke, J. Chromatogr., 1979, 177, 69. 10 M. Van Puymbroeck, E. Royackers, R. F. Witkamp, L. Leyssens, A. S. Van Miert, J. Gelan, D. Vanderzande and J. Raus, in Proceedings of Euroresidue III, Veldhoven, 6–8 May 1996, ed. N. Haagsma and A. Ruiter, distributed by: Department of Food of Animal Origin, Section of Food Chemistry, Faculty of Veterinary Medicine, University of Utrecht, Utrecht, The Netherlands, 1996, p. 808. 11 H. Groh, R. Schon, M. Ritzau, H. Kasch, K. Undizs and G. Hobe, Steroids, 1997, 62, 437. 12 E. Houghton, A. Ginn, P. Teale, M. Dumasia and J. Copsey, J. Chromatogr., 1989, 479, 73. Paper 8/05009K Fig. 5 Full scan electron impact mass spectrum of the trimethylsilyl derivatives of (a) THNG and (b) MeAD. 2452 Analyst, 1998, 123, 2449–2452

ISSN:0003-2654    

DOI:10.1039/a805009k

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

In vitro liver models are important tools to monitor the abuse of anabolic steroids in cattle† M. Van Puymbroeck,‡ab M. E. M. Kuilman,c R. F. M. Maas,c R. F. Witkamp,cd L. Leyssens,a D. Vanderzande,b J. Gelanb and J. Rausab a Dr. L. Willems-Instituut, Department of Drug- and Residue Analysis, B-3590 Diepenbeek, Belgium b Limburgs Universitair Centrum, Department SBM, B-3590 Diepenbeek, Belgium c University of Utrecht, Department of Veterinary Pharmacology, Pharmacy and Toxicology, NL 3508 TD Utrecht, The Netherlands d TNO Pharma, 3700 AJ Zeist, The Netherlands Received 30th June 1998, Accepted 13th August 1998 Current veterinary residue analysis mainly focuses on the monitoring of residues of the administered parent compound.However, it is possible that larger amounts of metabolites are excreted and that they can have a prolonged excretion period. In order to unravel specific metabolic steps and to identify possible biological markers, two in vitro liver models were used, i.e.monolayer cultures of isolated hepatocytes and liver microsomes, both prepared from liver tissue of cattle. Clostebol, boldenone, norethandrolone (NE) and ethylestrenol (EES) were used as model substrates. Results show that the metabolic profiles derived from in vitro experiments are predictive for the in vivo metabolic pathways of the steroids evaluated in this study. By means of this strategy, it is possible to identify 17a-ethyl-5b-estrane-3a,17b-diol (EED) as a common biological marker for NE and EES.By in vivo experiments it was shown that EED is particularly important for the detection of the abuse of NE or EES because of its high excretion levels and its prolonged presence as compared with the parent compounds or any other metabolite. Aim of investigation A knowledge of the metabolic pathways of a particular synthetic anabolic steroid can be important to improve the screening for its abuse in cattle. Therefore, the investigation of the metabolism of some of these steroids in vitro was initiated.The liver is quantitatively the main organ responsible for a wide variety of biotransformations of xenobiotics. There are several possibilities for using liver preparations to assess the hepatic metabolism, e.g., perfusion of liver as such, incubation of subcellular fractions, liver slices, liver tissue preparations and cells.1,2 The phase I oxidative biotransformations are the most important for the metabolism of steroids.The enzymes involved are almost exclusively localised in the endoplasmic reticulum. The microsomal fraction represents the ‘pinched off’ and vesiculated fragments of the original endoplasmic reticulum that retain most of their enzymatic activity.1,3 The microsomal fraction was therefore used to study the oxidations of the steroid nucleus in a low matrix environment. Primary cultures of hepatocytes isolated from cattle livers allow a more accurate quantitative study of the different metabolic routes.In addition, the formation of cytosolic metabolites can be monitored directly within the system.3 The phase I and II metabolic conversions, the metabolic routes of the endoplasmic reticulum and the cytosol, can be studied in combination. Our results in both cattle liver microsomal preparations and primary cultures of isolated hepatocytes made it possible to clarify and predict the structure of in vivo metabolites of clostebol (CLT), norethandrolone (NE) and ethylestrenol (EES).As the metabolites of CLT in cattle urine were identified in a previous study,4 CLT was used as a reference compound to evaluate the in vitro assays to approximate the in vivo situation. The metabolism of NE and EES in cattle was not yet known. The predictive value of the in vitro methods was further confirmed by the successful unravelling of the metabolism of NE and EES. In vivo experiments and case studies in control procedures to detect forbidden growth promoters verified these conclusions.Experimental Experimental animals and samples Microsomes were prepared from livers of four different animals: two adult females (Friesian–Holstein and Meuse– Rhine–Yssel) of approximately 600 kg and a cow and a bull, both of mixed breed and 36 weeks old. Hepatocytes were isolated from livers from two Friesian– Holstein bulls of about 1 year old. For the in vivo experiments, 36 mg of EES were administered intramuscularly to a 12 month old heifer (±100 kg) and 250 mg of NE to a 15 month old heifer (±300 kg).A 50 mg amount of EES was dissolved in 10 ml and 250 mg of NE in 12 ml of a 1 + 1 (v/v) mixture of pharmaceutical grade Miglyol and Nmethyl- 2-pyrrolidone (Sigma, St. Louis, MO, USA). Urine samples from animals treated with 17b-boldenone undecanoate were obtained from RIKILT-DLO (Wageningen, The Netherlands). The urine metabolites of 4-chlorotestosterone were identified in an earlier study in our laboratory.4 For this study, two 18 month old heifers and two 10 month old bulls were † Presented at the Third International Symposium on Hormone and Veterinary Drug Residue Analysis, Bruges, Belgium, June 2–5, 1998.‡ Pre-doctoral fellow of the LUC Universiteitsfonds. Analyst, 1998, 123, 2453–2456 2453injected intramuscularly with different amounts of 4-chlorotestosterone acetate. Chemicals and standards All reagents and solvents were of analytical-reagent grade.All anabolic steroids were obtained from the Belgian Reference Laboratory (Wetenschappelijk Instituut voor Volksgezondheid- Louis Pasteur, Brussels, Belgium). Tri-Sil-TBT (Pierce, Rockford, IL, USA), which consists of trimethylsilylimidazole, N,Obis( trimethylsilyl)acetamide and trimethylsilylchlorosilane (3 + 3 + 2, v/v/v) was used for trimethylsilylation. Ethoxyaminehydrochloride (Fluka, Buchs, Switzerland) was used as a 2% solution in pyridine (Sigma). 17a-Ethyl-5b-estrane-3a,17bdiol (EED) was provided by the Institut für Biochemie, Deutsche Sporthochschule Köln, Germany.All reagents used for the chemical synthesis of the metabolites were from Aldrich (Milwaukee, WI, USA). In vitro experiments Microsomes were prepared and incubated with steroids as described previously.5 Protein concentrations were assessed by the method of Lowry et al.6 using bovine serum albumin as a standard. Hepatocytes were isolated according to Van’t Klooster et al.,3 based on the method of Seglen.7 Cells were cultured at a density of 4.106 cells per 60 mm culture dish (Greiner, Alphen a/d Rijn, The Netherlands) in 4 ml of Williams’ E supplemented with 4% newborn calf serum, 1.67 mmol l21 glutamine, 50 mg ml21 gentamicin sulfate, 1 mmol l21 hydrocortisone, 1 mmol l21 insulin, 0.5 mmol l21 CaCl2 and 0.5 mmol l21 MgCl2. Cells were incubated for 4 h in a humidified atmosphere of air (95%) and CO2 (5%) at 37 °C.The medium was then replaced by a medium without serum, CaCl2 and MgCl2.After incubation for 20 h, the hepatocytes were incubated with either 100 mmol l21 of the steroid for 6 and 24 h or with 10 mmol l21 for 24 h. The steroids were dissolved in methanol (final concentration of methanol 0.1%). Sample preparation To isolate steroids from microsomal incubation mixtures a liquid–liquid extraction of 1 ml of the incubation mixture with 6 ml of diethyl ether and subsequently with 6 ml of tert-butyl methyl ether was performed.The tubes were capped and shaken by hand. After freezing the aqueous layer, the organic solvent was removed, and evaporated under a gentle stream of nitrogen. The extracts were analyzed by HPLC and GC-MS as described below. A 2 ml volume of the liver cell incubation mixtures of hepatocytes was adjusted to pH 5.2 with acetate buffer (0.2 mol l21, pH 4.8) after which 50 ml of Succus Helix Pomatia were added and the mixture was incubated overnight at 37 °C.The mixture was then loaded on to a Chem-ElutTM column and the steroids were eluted by additions of 5 ml of tert-butyl methyl ether and 5 ml of chloroform. All extracts were analyzed by GC-MS, as described below. The clean-up of urine and faeces samples by solid-phase extraction (SPE) and HPLC fractionation was carried out as described previously.4 For the detection of the 6-hydroxy metabolites of 17b-boldenone (17b-BOL) and 1,4-androstadiene- 3,17-dione (ADD), the HPLC fractionation was expanded with three additional fractions.These were collected from the drain prior to the actual collection of the standard steroids. The collector was operated in the time mode. The three fractions were collected with time windows, relative to naphthalene as internal standard, ranging from 0.23 to 0.29, from 0.29 to 0.35 and from 0.35 to 0.46. Gas chromatography-mass spectrometry In order to obtain ethoxime-trimethylsilyl (EO-TMS) derivatives, dry residues were derivatised with ethoxyamine and a silylating mixture (Tri-Sil-TBT).5 The GC-MS analyses were performed on a Finnigan GCQ system (San Jose, CA, USA).The GC column used was a DB-5 MS 30 m 3 0.32 mm id fused-silica column with a 0.25 mm film thickness (J&W Scientific, Folsom, CA, USA). Temperature settings were as follows: injector, 260 °C; transfer line, 275 °C; ion source, 200 °C; oven program: from 50 °C (held for 20 s) to 190 °C at 50 °C min21, then to 320 °C (held for 4 min) at 4.6 °C min21.The carrier gas was helium at a velocity of 40 cm s21 for the GCQ. Aliquots (1 ml) were injected on a split–splitless injector in the split mode. Results and discussion CLT CLT was used as a reference compound. Its urine metabolites were identified in an earlier study on three animals.4 For several years, these metabolites have been successfully used in routine analysis to detect the misuse of CLT in cattle breeding. The isotope ratios of the chlorine group present in the CLT molecule facilitated the location of formerly unknown chlorinated metabolites and are an excellent marker for their exogenous character.The three major in vivo biodegradation products of CLT, i.e., 4-chloroandrost-4-ene-3,17-dione (CLAD), 4-chloroandrost- 4-ene-3a,17b-diol, and 4-chloroandrost-4-ene-3-ol- 17-one, were present in microsomal liver preparations and incubation mixtures of hepatocytes. Although epimerisation is a common in vivo biotransformation route, no 17a-CLT was found.Only a trace amount of an important hydroxylated urine metabolite, 4-chloroandrost-4-ene-3x,17x,x-triol, was found in vitro.4 In addition, new compounds, not yet described in the urine of CLT treated animals, were present in the microsomal preparations yielding mass spectra with similar electron impact (EI) fragmentation patterns to those of some of the hydroxylated metabolites found in cattle urine, but eluting at different retention times.EES and NE Incubation of microsomal liver preparations from four different animals with EES, revealed NE as the major biotransformation product [Fig. 1(A)]. Because of its strong hydrophobic nature, the oxidation of EES is of major importance to eliminate it from the body. NE itself was used as substrate in an incubation experiment with liver microsomes. Only minor hydroxylated products were tentatively identified. Additional experiments on the metabolism of NE were performed with isolated liver cells.To identify possible metabolites, NE was incubated for 24 h at 100 and 10 mmol l21, and for 6 h at 100 mmol l21. The major conversion product was the reduced form 17a-ethyl-5b-estrane-3a,17b-diol (EED). Hydroxylation products were of minor importance. A typical chromatogram and the EI mass spectrum of EED are shown in Fig. 1(B) and 2, respectively. The two fragment ions of the Dring at m/z 144 and 157 relate the product to NE and EES (Fig. 2) and prove its exogenous character.This makes it very useful as a biological marker for NE and EES. The loss of two [M 290] is typical of the presence of two trimethylsilylhydroxy groups. The molecular ion of the trimethylsilyl derivative (Mr 450) is not present owing to the loss of the ethyl group [M 2 29] at position 17. This generates the ion at m/z 421 in the spectrum. The ions at m/z 331 and 241 originate from the subsequent loss of the two trimethylsilylhydroxy groups. 2454 Analyst, 1998, 123, 2453–2456Despite differences in race, sex and feeding habits between the studied animals, all in vitro experiments pointed towards the same metabolic pathway, as described in Fig. 3. As a consequence, these experiments indicated that EED might be a reliable marker for the detection of both NE and EES abuse in cattle. This hypothesis was confirmed by in vivo experiments. Urine and faeces samples of the heifers treated with EES and NE were collected at regular intervals and analyzed by GCMS.EES itself was not detectable in urine or faeces. The metabolite NE was only detectable in urine samples for 3 d after a 3 d lag period at concentrations lower than 2 ng ml21, whereas faeces samples remained negative for NE. However, EED could be detected in both urine and faeces. EED was the only conversion product in faeces that points to the administration of EES. After administration of NE, large amounts of the parent compound were excreted in urine during the first week.NE was present in faeces up to 36 h after NE administration. EED prolonged the detection of NE treatment in faeces for almost 1 week. The longer detectability of EED in faeces as compared with NE can be explained by the fast conversion of NE to EED in liver. The latter metabolite was seen in incubations of primary cultures of hepatocytes. Therefore, it was concluded that EED is an excellent biological marker to reveal the abuse of NE and EES in cattle, particularly when faeces samples are screened. 17b-BOL The oxidation of 17b-boldenone (17b-BOL) to ADD is the major metabolic pathway in microsomes. Additionally, small amounts of 6-hydroxy-17b-boldenone and 6-hydroxy-1,4-androstadiene- 3,17-dione are formed. The 6-hydroxy metabolites were synthesised according to the method of Schänzer.8 After 24 h of incubation with isolated liver cells, a significant portion of 17b-BOL and ADD is hydroxylated at position 6. Only a small portion of either 17b-BOL or ADD is subjected to further reductions of the unsaturated bonds and/or keto functions.a- Epimerisation to 17a-boldenone (17a-BOL) is also detected. Incubations of 17b-BOL with liver microsomes and isolated hepatocytes indicate that hydroxylations of the androstadiene skeleton, especially at the allylic position 6, are important metabolic pathways. To study the in vivo situation, urine samples obtained after application of 17b-boldenone undecanoate were screened for the presence of 17b-BOL, the earlier detected in vitro metabolites of 17b-BOL and 5b-androst-1-ene-3,17-dione (5b- AED). 5b-AED was reported as a metabolite of 17b-BOL in human urine,8 but was not found after in vitro incubations. 5b- AED was synthesised according to the method of Schänzer and Donike.8,9 17b-BOL, 17a-BOL, ADD and 5b-AED were detected in cattle urine. In addition to 5b-AED, several reduction products were tentatively identified. These included all the spectra found in the incubations with hepatocytes, but also some additional isomers.Also, a small amount of 6-hydroxy-17b-boldenone was found. All urine metabolites found were predicted by the in vitro screening except for 5b- AED and some of the isomers of the reduction products. Conclusions The search for metabolites that could act as biological markers for the illegal use of anabolic steroids as growth promoters in Fig. 1 (A) Total ion current and single ion chromatograms are shown for the three most important diagnostic ions of EES and NE, pointing towards NE as the most important metabolite of EES after incubation with liver microsomal preparations. (B) EED as the major in vitro metabolite after a 6 h incubation of NE with isolated hepatocytes.Total ion current and ion chromatograms for the three most important diagnostic ions of EED and NE are shown. Fig. 2 (A) Full scan EI mass spectrum of EED. (B) D-ring fragmentation in the trimethylsilyl derivative of EED gives rise to the fragment ions at m/z 157 and 144 present in the spectrum of EED.Immediate loss of the ethyl group at position 17 causes a loss of the molecular ion at m/z 450. Analyst, 1998, 123, 2453–2456 2455cattle led to in vitro techniques, such as microsomal preparations and monolayer cultures of intact liver cells, prepared from liver tissue of slaughtered cattle. The combined use of the two techniques allowed a study of the prominent changes of the steroid skeleton by the liver.For all synthetic steroids included in this study, the liver microsomes were able to predict the major hydroxylations and oxidations of the steroid skeleton by membrane-bound mixed function oxidative enzymes, the cyt P450 systems. Hepatocytes more closely reflect the in vivo situation, since in addition to the oxidations and hydroxylations, reductions of the unsaturated bonds are also performed. In conclusion, in vitro studies reduce the need for in vivo experiments and provide a low matrix environment for fast localisation of diagnostic metabolites, as illustrated by the example of EED.References 1 Introduction to Drug Metabolism, ed. G. G. Gibson and P. Skett, Blackie Academic & Professional, Glasgow, 2nd edn., 1994, pp. 191–198. 2 M. N. Berry, A. M. Edwards and G. J . Barrit, in Isolated Hepatocytes, Preparation, Properties and Applications, Laboratory Techniques in Biochemistry and Molecular Biology, ed R. H. Burdon and P. H. van Knippenberg, Elsevier, Amsterdam, 1991, pp. 59–80. 3 G. A. E. Van’t Klooster, F. M. A. Woutersen-van Nijnanten, W. R. Klein, B. J. Blaauboer, J. Noordhoek and A. S. J. P. A. M. van Miert, Xenobiotica, 1992, 22, 523. 4 L. Leyssens, E. Royackers, B. Gielen, M. Missotten, J. Schoofs, J. Czech, J. P. Noben, L. Hendriks and J. Raus, J. Chromatogr., 1994, 654, 43. 5 M. Van Puymbroeck, E. Royackers, R. F. Witkamp, L. Leyssens, A. S. Van Miert, J. Gelan, D. Vanderzande and J. Raus, in Proceedings of the Euroresidue III Conference, Veldhoven, May 6–8, 1996, ed. N. Haagsma and A. Ruiter, Department of Science of Food of Animal Origin, Section of Food Chemistry, Faculty of Veterinary Medicine, University of Utrecht, Utrecht, The Netherlands, 1996, p. 808. 6 O. H. Lowry, N. J. Rosenbrough, A. L. Farr and R. J. Randall, J. Biol. Chem., 1951, 193, 265. 7 P. O. Seglen, Cell Biol., 1976, 13, 29. 8 W. Schänzer, Clin. Chem., 1996, 42, 1001. 9 W. Schänzer and M. Donike, Anal. Chim. Acta, 1993, 275, 23. Paper 8/05013I Fig. 3 In vitro metabolic conversion of EES and NE to EED. NE is an intermediate in the conversion of EES to EED. 2456 Analyst, 1998, 123, 2453–2456

ISSN:0003-2654    

DOI:10.1039/a805013i

出版商:RSC    

年代:1998

数据来源: RSC