Download Determination of Clenbuterol Residues in Bovine Urine by Optical

HAUGHEY ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 84, NO. 4, 2001 1025
DRUGS, COSMETICS, FORENSIC SCIENCES
Determination of Clenbuterol Residues in Bovine Urine by
Optical Immunobiosensor Assay
SIMON A. HAUGHEY, G. ANDREW BAXTER, and CHRISTOPHER T. ELLIOTT
Department of Agriculture and Rural Development, Veterinary Sciences Division, Stoney Rd, Stormont, Belfast, BT4 3SD, UK
BJORN PERSSON, CARIN JONSON, and PETER BJURLING
Biacore AB, Rapsgatan 7, Uppsala, Sweden
Clenbuterol (CBL) is an orally active
b2-adrenoceptor agonist which has been used in
veterinary medicine as a broncodilator and an
agent of uterine relaxation. It has however become
better known as a drug used illegally to promote
growth in farm animals. A rapid and sensitive biosensor assay was developed to detect CBL residues in bovine urine. The method involved a simple extraction procedure using tert-butyl methyl
ether followed by analysis on the biosensor with
results obtained against a buffer calibration curve.
The assay allowed up to 88 samples to be analyzed
per working day, with each cycle on the biosensor
taking approximately 7 min to complete. The limit
of detection (LOD) was determined as 0.27 ng/mL
using 20 EU reference blank urine samples. The
intra-assay Sr ranged from 4.7–7.6% for 3 control
samples while the interassay Sr ranged from
9.2–12.7%. The recovery was found to be approximately 95%. A series of incurred urine samples
were assayed and the results compared by Enzyme immunoassay (EIA), radio-immunoassay
(RIA), and gas chromatography/mass spectrometry
(GC/MS) analysis. Urine samples taken from local
abattoirs were also analyzed by the biosensor
method and by EIA analysis. The antibody used in
the biosensor test exhibited high cross reactivity
with at least 7 other b-agonists allowing detection
of these compounds at less than 1 ng/mL in bovine
urine.
n veterinary medicine, the β2-adrenoceptor agonist
clenbuterol (CBL) has been used, under veterinary supervision, for the treatment of respiratory disease in cattle and
horses. It was also used as a tocolytic agent in cattle. However,
the properties of CBL, which have given the drug a high degree of notoriety, are not linked to these uses but rather the
side effect of marked growth promotion. When CBL is used
illegally at high doses (about 10 × therapeutic; 1) as a
repartitioning agent, i.e., promotes weight gain, increases pro-
I
Received August 22, 2000. Accepted by JM November 30, 2000.
tein deposition, and decreases the fat mass, the economic benefits accrued are substantial. The use of CBL as a
repartitioning or growth promoting agent has been prohibited
in Europe under European Council Directive 96/22/EC (2) but
it is clear that the problem of illicit use of this substance (and
analogues) is a global one.
CBL misuse will give rise to the presence of CBL residues
in foodstuffs intended for human consumption. Documented
cases of acute food poisoning resulting from the ingestion of
CBL contaminated meat exists. In Spain 367 cases were reported, with the majority in 2 major outbreaks (3, 4), and
thought to relate to the ingestion of veal liver. In France
22 cases were reported (5), and in Italy 16 people developed
food poisoning after eating fillet and rump steaks which were
found to contain high concentrations of CBL residues
(>0.5 µg/g; 6). In the United States, suspected illegal use of
the drug coupled with these instances of food poisoning due to
CBL residues caused concern with the U.S. Food and Drug
Administration. It advised on the illegal use of CBL and the
possible adverse effects on public health as well as alerting
U.S. Customs to illegal importation. Development of mass
spectrometric analysis to detect CBL residues has led to indictments, convictions, and other investigations (7).
In order to monitor and control the unwanted use of such
growth-promoting drugs, analytical strategies are required.
These strategies will vary from region to region based on the
size of the abuse problem which is likely to be encountered, the
size of the animal population present, and the degree of sophistication of laboratory facilities available to perform the testing.
Screening tests for CBL have generally used immunoassay
based procedures. Yamamoto and Iwata (8) reported the first
competitive-based enzyme test (enzyme immunoassay; EIA)
for CBL. Degand et al. (9) developed an EIA based on the
competition between CBL and CBL-horse radish peroxidase.
This test was capable of detecting CBL residues present in
urine samples in the region of 0.5 ng/mL. A similar EIA
method reported detection limits for urine analysis at
0.3 ng/mL (10). This test was later modified to allow the detection of CBL residues in a wide range of sample types (11).
Haasnoot et al. (12) have also described the development of a
CBL EIA. By using the cross reactivity of the antibody devel-
1026 HAUGHEY ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 84, NO. 4, 2001
Figure 1. Typical calibration curve (n = 4) for the assay with error bars showing the range of response for each
calibration point.
oped with structurally similar compounds to CBL, they were
able to detect at least 4 other β-agonist compounds which had
been linked to illegal use (cimbuterol, brombuterol,
mapenterol, and mabuterol).
Table 1. Assay validation data recorded for the CBL
urine test
Repeatability
Parameter
n
CBL concn.,
ng/mL
Mean
Sr, %
Intra-assay
5
0.3
0.285
7.0
5
0.4
0.416
4.7
0.6
0.637
7.6
5
Interassay
Recovery, %
3
0.3
0.267
9.2
3
0.4
0.395
12.7
3
0.6
0.601
10.7
12
95.7 ± 7.5
Radio-immunoassay procedures (RIA) have also been developed for the detection of CBL in extracts from a range of
samples (13, 14). A radio-receptor assay (RRA) has also been
developed which uses specific receptors isolated from plasma
membrane (15). Detection limits in the range of 2 ng/mL have
been claimed. The advantage of such a method has been described as the ability to detect a wide range of β-agonists residues present in a sample.
More recently, a biosensor immunoassay was developed
for the detection of salbutamol (a β-agonist compound similar
in structure to CBL) in bovine urine (16). Many now view biosensor technology as a method of producing highly robust,
sensitive, and fast means of performing analytical tests. The
aim of the present study was to develop an assay capable of
detecting CBL residues present in bovine urine using an optical biosensor assay. Following the validation of this test, the
results that it obtained in detecting CBL residues in samples
taken from treated animals were compared with those found
with the traditional immunoassay (EIA and RIA) and gas
chromatography/mass spectrometry (GC/MS) methods routinely used in testing laboratories.
HAUGHEY ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 84, NO. 4, 2001 1027
Table 2. Cross-reaction of F140 against b-agonists
Cross-reaction, %
Compound
Buffer
Extracted
Clenbuterol
100
100
Mabuterol
147.6
115
Salbutamol
140
5
Cimbuterol
91.9
77.1
Hydroxymethyl
clenbuterol
79.9
57.3
Brombuterol
72.8
60.1
Methylclenbuterol
62.5
51.5
Mapenterol
50
29.7
Cimaterol
37.3
19.7
Terbutaline
14.8
<0.01
Fenoterol
<0.01
<0.01
Salmeterol
<0.01
<0.01
Ritodrine
<0.01
<0.01
Ractopamine
<0.01
<0.01
Isoxsuprine
<0.01
<0.01
(c) Storage solution.—Sodium azide (0.975 g) and bovine
serum albumin (2.5 g) were dissolved in HBS-EP (1000 mL).
(d) Preparation of calibrants.—A stock solution was prepared by dissolving 10 mg clenbuterol hydrochloride in
10 mL deionized water, which is equivalent to 0.884 mg
clenbuterol/mL. The stock solution was diluted 1 to 100 with
deionized water to give an intermediate solution of
8.84 µg/mL. The calibrants were prepared by dilution of the
intermediate solution with the calibration buffer to give concentrations of 7.07, 3.53, 1.77, and 0.88 ng/mL. These concentrations were recalculated from clenbuterol equivalents in
buffer to clenbuterol equivalents in urine to give concentrations 0.88, 0.44, 0.22, and 0.11 ng/mL. The calibration curve
was constructed from the 4 calibration points run in duplicate
using a 4-parameter fit algorithm.
Antibody
A mouse derived anticlenbuterol monoclonal antibody
(F140) was developed according to the method of
Teh et al. (17). The antibody was precipitated from ascites
fluid using saturated ammonium sulfate solution and then dialyzed extensively against phosphate buffered saline (PBS).
The clenbuterol antibody was diluted to give an intermediate
dilution of 1/200 prepared in storage solution. The working
antibody dilution (1/1000) was prepared by taking 1 part intermediate solution and diluting it with 4 parts HBS-EP.
Sensor Surface
Experimental
CBL was immobilized onto the surface of a sensor
chip (CM5, certified grade, Biacore AB) using the fol-
Instrumentation
An optical biosensor, BIACORE Q, was obtained from
Biacore AB, Uppsala, Sweden. BIACORE Q control software was used for instrument operation and for data handling.
Chemicals
All chemicals are AnalaR grade unless otherwise stated.
Sodium hydroxide (1M), sodium dihydrogen phosphate 1-hydrate, disodium hydrogen phosphate 2-hydrate, sodium chloride, potassium chloride, urea, ammonium sulfate, and
tert-butyl methyl ether (TBME; LC grade) were obtained
from BDH Laboratory Supplies (Poole, UK).
Creatinine and clenbuterol hydrochloride were obtained
from Sigma (Poole, UK). The clenbuterol derivative, prepared
by modification of the anilinic amino group for immobilization to the surface, HBS-EP buffer, and Amine Coupling Kit
were obtained from Biacore AB.
Reagents
(a) Phosphate buffer.—Sodium dihydrogen phosphate 1-hydrate (0.12 g), disodium hydrogen phosphate
2-hydrate (3.4 g), NaCl (117 g), and KCl (149 g) were
dissolved in deionized water up to a final volume of
1000 mL. The pH should be 7.2.
(b) Calibrant buffer.—Urea (0.4 g) and creatinine
(0.027 g) were dissolved in 10 mL phosphate buffer.
Table 3. Suggested limits of detection of the biosensor
assay for a range of b-agonist compounds based on
extraction efficiences and CBL validation data
Compound
Suggested LOD for each
compound, ng/mL urine
Clenbuterol
0.27
Mabuterol
0.23
Salbutamol
5.4
Cimbuterol
0.35
Hydroxymethyl clenbuterol
0.47
Brombuterol
0.45
Methylclenbuterol
0.52
Mapenterol
0.91
Cimaterol
1.37
Terbutaline
>10
Fenoterol
>10
Salmetero
>10
Ritodrine
>10
Ractopamine
>10
Isoxsuprine
>10
1028 HAUGHEY ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 84, NO. 4, 2001
Table 4. Results (in ng/mL) of incurred calf urines taken from 3 days after withdrawal of medication and analyzed by
various assay formats
Calf 1
Calf 2
Calf 3
Day
RIA
Biosensor
EIA
GC/MS
RIA
Biosensor
EIA
GC/MS
RIA
Biosensor
EIA
GC/MS
7
1.16
>0.88
1.29
1.13
2.03
>0.88
0.95
1.53
1.67
>0.88
1.39
1.05
8
0.89
>0.88
1.24
0.75
1.36
>0.88
0.73
1.22
1.32
>0.88
1.22
0.77
9
0.65
>0.88
0.54
0.47
0.83
>0.88
0.75
0.92
0.69
0.83
0.74
0.57
10
0.51
0.41
1.09
0.47
0.50
0.55
0.70
0.61
0.52
0.56
0.80
0.48
11
0.42
0.33
0.72
0.28
0.45
0.42
0.45
0.41
0.32
0.28
0.77
0.38
12
0.30
0.36
0.27
<0.20
0.43
0.40
0.63
0.41
0.41
0.32
0.58
0.38
13
0.34
0.31
0.25
<0.20
0.25
0.27
0.50
0.31
0.37
0.28
0.7
0.29
14
0.24
0.26
0.15
<0.20
0.19
0.26
0.61
0.31
0.28
0.22
0.57
<0.20
lowing procedure: The carboxymethyl dextran surface
was activated by the injection of 35 µL of a 1:1 mixture of
0.1M
N-hydroxy
succinimide
(NHS)
and
0.4M 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide
hydrochloride (EDC; both contained in the Amine Coupling Kit) at a flow rate of 5 µL/min. CBL derivative solution (35 µL 0.1 mM in 50 mM boric acid, pH 8.5) was injected over the activated surface. Unreacted sites were then
deactivated by injection of 15 µL 1M ethanolamine, pH 8.5.
Reference Blank Urines
Lyophilized reference blank bovine urine samples (Project
389002, code BOV01-BOV20) were obtained from
RIVM-ARO/CRL, Bilthoven, The Netherlands. These samples were taken from a broad range of animals including veal
calves, fattening bulls, heifers, pregnant cows, and mature
bulls. Each sample was reconstituted in 5 mL deionized water.
Production of Incurred Urine Samples
Incurred urine samples were obtained from TNO Voeding
(Zeist, The Netherlands). Three calves (weight: 116.2 kg,
118.4 kg, and 111.4 kg) were treated with 4 µg clenbuterol hydrochloride per kg body weight, twice daily, administered
orally in milk replacer and continued for 4 days. Urine samples were collected prior to the commencement of medication
as blank reference samples, then daily for 4 days of medication and the withdrawal period of 10 days. All samples were
stored at –20°C until analyzed.
Routine Urine Sample Collection
Urine samples were randomly selected from animals submitted for slaughter at local abattoirs in Northern Ireland. The
samples were placed in sealed containers and transported to
the laboratory.
Urine Sample Extraction Procedure
Sample urine (2 mL) was pipetted into a labeled test tube
which had been conditioned with TBME (2 mL) and air-dried.
NaOH (0.5 mL, 0.1M) was added followed by a volume of 4
mL TBME. Each tube was mixed gently on a Vortex mixer for
3 × 5 s. If an emulsion formed at this stage, the test tube was
centrifuged at 800 × g for 5 min. The aqueous layer of each
tube was frozen using an aluminum block, precooled in liquid
nitrogen. The TBME layer was transferred into a clean labeled
test tube. The TBME was evaporated to dryness under a gentle
stream of nitrogen on a dri-block sample concentrator (Techne
DB-3A) at 50°C. The resultant sample residue was reconstituted in 250 µL phosphate buffer, which had been diluted 1:1
with deionized water, and a minimum of 120 µL of each sample extract was transferred to the well of a microtitre plate
(Greiner 650101).
Biosensor Assay
Assay principle.—The biosensor assay developed was
based on inhibition principles and detected the CBL antibody
when it bound to the drug immobilized on the sensor chip surface. A fixed concentration of the CBL antibody was mixed
with the sample prior to injection. Any CBL present in the
sample bound to the antibody and inhibited it from binding to
the surface of the sensor chip. The higher the concentration of
CBL in the sample, the higher the level of inhibition and hence
the lower the response of the biosensor. The analysis cycle is
recorded in the form of a sensorgram with report points recorded before and after each analysis cycle. The surface was
then regenerated, ready for the next sample.
Urine analysis.—Each calibrant or reconstituted extracted
urine (120 µL) was placed in the appropriate well of the
microtitre plate, which was subsequently transferred to the
biosensor instrument. Within the biosensor, an autosampler
proceeded to take a 1:4 mixture of antibody solution and
calibrant or reconstituted urine sample from a predetermined
well on the microtitre plate. The biosensor then injected 80 µL
of this solution over the CBL coated sensor chip at a flow rate
of 40 µL/min. Report points were recorded before and after
the sample injection, after which the surface was prepared for
the next sample with a 35 µL injection of the regeneration solution (0.1M NaOH). A complete analysis for each sample
HAUGHEY ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 84, NO. 4, 2001 1029
(i.e., mixing with antibody, sample injection, measurements,
and regeneration) lasted for ca 7 min.
Assay Validation
The limit of detection (LOD) of the biosensor test was determined by assaying a panel of European Union (EU) reference blank urine samples (n = 20). The mean and the standard
deviation in response units (RU) were calculated from the assay data. The LOD was calculated as the mean minus 3 times
the standard deviation (s) (mean – 3s) which gives a concentration value by entering the result in the calibration curve.
Assay reproducibility was determined by performing repeated analysis (within and between assays) of samples fortified with CBL and concentrations ranging from
0.3 to 0.6 ng/mL. The recovery efficiencies were determined
by spiking negative urine with CBL at the aforementioned
concentrations and comparing the calculated result from the
assay with the spiked level. The antibody cross reactivity profile against a range of β-agonists was determined as according
to the method of Elliott et al. (10). In addition, the range of
β-agonist standards were also added to blank urine samples
and extracted as described previously.
Results and Discussion
The type of calibration curves (n = 4) obtained with the biosensor assay are shown in Figure 1, where the error bars represent the spread of response units for each calibration point. To
determine the LOD of the assay, the mean response and the
standard deviation (s) for the reference blank urine samples
were calculated as 347.1 and 44.8, respectively. From these
values, the LOD was calculated as 212.7 RU, which gave a
value of 0.27 ng/mL for the LOD when entered in the calibration curve. This value is similar to the values obtained by other
immunoassay based screening procedures for β-agonists
(9–13). The validation data obtained for the assay has been
outlined in Table 1. The assay exhibited good repeatability
data for intra- and interassay variations. The intra-assay variation was 7.0, 4.7, and 7.6% from samples spiked with CBL
concentrations of 0.3, 0.4, and 0.6 ng/mL, respectively, while
the interassay variation was 9.2, 12.7, and 10.7%, respectively. The mean value for recovery was calculated at
95.7 ± 7.5% (n = 12).
The cross-reactivity profile for the monoclonal antibody
(F140) has been outlined in Table 2. It can be seen that the antibody selected for the assay had pronounced cross reactivity
with a wide range of β-agonists such as mabuterol, salbutamol, cimbuterol, and brombuterol. There was an apparent decrease in the degree of cross reactivity observed in urine samples spiked with β-agonists and extracted prior to analysis.
These decreased values are likely to be related to the parameter of extraction efficiencies rather than changes in antibody
behavior. Most notably, the phenolic drugs such as salbutamol
and terbutaline were affected. This information tends to suggest that for the present assay to be classified as a truly
‘multi-β-agonist’ screening test, some modifications to the extraction procedure will be required. Based on the calculated
LOD for the assay and ‘extracted cross-reactivity data,’ Table 3 has been incorporated to suggest the LODs likely to be
experienced with other β-agonists. From this data it can be
seen that the biosensor assay is capable of detecting at least 7
of the most widely abused β-agonists at concentrations less
than 1 ng/mL in urine.
The results of the analyses of the urine samples, taken from
the calves treated with CBL, from 3 days after withdrawal of
medication (i.e., Day 7 of the experiment) have been detailed in
Table 4. RIA results were determined by the procedure outlined
in Elliott et al. (16), the EIA results by the method of Elliott et
al. (10), and the GC/MS results by the technique of
Blanchflower et al. (18). The biosensor assay showed good correlation with GC/MS confirmatory results, with a correlation
coefficient calculated as 0.84. The correlation coefficients for
the RIA–GC/MS and EIA–GC/MS were 0.93 and 0.56, respectively. In one of the samples tested (Calf 2, Day 14), the GC/MS
result was 0.31 ng/mL and the corresponding biosensor assay
value was 0.26 ng/mL. Though these results are quantitatively
very close, the biosensor result is slightly lower than the defined
LOD for the assay. In contrast, 3 samples taken during the study
remained above the LOD for the biosensor assay while the
GC/MS procedure could not detect CBL residues at its predetermined LOD. This data demonstrates that when residue concentrations are at or close to the LODs of analytical methods,
minor differences in actual values can result in samples being
deemed ‘positive’ or ‘negative.’
Ninety-eight sample urine samples, obtained from local
abattoirs, were also tested by both biosensor and EIA assay
conditions. Positive results (i.e., above LODs) were not found
during this study.
The present study has led to the development of a rapid,
sensitive, optical biosensor assay for the detection of CBL residues at levels comparable to other immunoassay based
screening tests. Up to 88 samples can be assayed in 1 day with
a relatively short analysis time (about 7 min) for each cycle. It
can already be seen that the biosensor assay can detect a wide
range β-agonists and there is potential for the assay, following
some modifications to the extraction procedure, to become an
even broader spectrum multi β-agonist screening test.
Acknowledgments
We would like to thank Cor Arts and Natascha Overmars,
TNO Voeding, The Netherlands, for the incurred bovine urine
samples and the RIA results. We would also like to thank the
staff of the Drug Residues Unit Statutory Testing Laboratory
and the Chemical Confirmation Unit at the Veterinary Science
Division, Belfast, for the EIA and GC/MS analysis of the
urine samples.
References
(1) Witkamp, R.F. (1995) in Proceedings of the Scientific Conference on growth promotion in meat production, Brussels,
Belgium, pp 297–323
1030 HAUGHEY ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 84, NO. 4, 2001
(2) Official Journal of the European Communities (1996) Council Directive 96/22/EC, No. L 125/3-9
(3) Martinez-Navarro, J.F. (1990) Lancet 336, 1311
(4) Salleras, L., Dominguez, A., Meta, E., Taberner, J.L., Moro,
I., & Salvà, P. (1995) Public Health Reports 110, 338–342
(5) Pulce, C., Lamaison, D., Keck, G., Bostvironnois, C.,
Nicolas, J., & Descotes, J. (1991) Veterinary and Human
Toxicology 33, 480–482
(6) Maistro, S., Chisea, E., Angeletti, R., & Brambilla, G. (1995)
Lancet 346, 180
(7) Mitchell, G.A., & Dunnavan, G. (1998) J. Anim. Sci. 76,
208–211
(8) Yamamoto, I., & Iwata, K. (1982) J. Immunoassay 3,
155–171
(9) Degand, G., Bernes-Duyckaerts, A., & Maghuin-Rogister, G.
(1992) J. Agric. Food Chem. 40, 70–75
(10) Elliott, C.T., McCaughey, W.J., & Shortt, H.B. (1993) Food
Add. Contam. 10, 231–244
(11) Elliott, C.T., Crooks, S.R.H., McEvoy, J.G.D., McCaughey,
W.J., Hewitt, S.A., Patterson, D., & Kilpatrick D. (1993) Vet.
Res. Commun. 17, 459–468
(12) Haasnoot, W., Cazemier, G., Stouten, P., &
Kemmers-Voncken, A. (1996) in Immunoassays for Residue
Analysis—Food Safety, R.C. Beier & L.H. Stanker (Eds),
American Chemical Society, Washington, DC, pp 60–73
(13) Delahaut, P., Dubois, M., Pri-bar, I., Buchman, O., Degand,
G., & Ectors, F. (1991) Food Add. Contam. 8, 43–53
(14) Boyd, D., O’Keefe, M., & Smyth, M.R. (1994) Analyst 119,
1467–1470
(15) Helbo, V., Vandenbroeck, M., & Maghuin-Rogister, G.
(1993) in Proceedings of the Euroresidue 2 Conference, N.
Haagsma, A. Ruiter, & P.B. Czedik-Eijsenberg (Eds),
CIP-Gegevens Koninklijke Bibliothiek, Den Haag, The Netherlands, pp 362–366
(16) Elliott, C.T., Baxter, G.A., Hewitt, S.A., Arts, C.J.H.,
VanBaaK, M., Hellenas, K.E., & Johansson, A. (1998) Analyst 123, 2469–2473
(17) Teh, C.Z., Wong, E., & Lee, C.Y.G. (1984) J. Appl.
Biochem. 6, 48–55
(18) Blanchflower, W.J., Hewitt, S.A., Cannavan A., Elliott, C.T.,
& Kennedy D.G. (1993) Biol. Mass Spect. 22, 326–333