Simultaneous determination of avermectins in bovine tissues by LC-MS/MS

Transcription

1 3596 Koichi Inoue 1 Yukiko Yoshimi 1 Tomoaki Hino 1 Hisao Oka 1,2 1 Department of Physical and Analytical Chemistry, School of Pharmacy, Kinjo Gakuin University, Moriyama-ku, Nagoya, Japan 2 Graduate School of Human Ecology, Human Ecology Major, Kinjo Gakuin University, Japan Received June 8, 2009 Revised September 13, 2009 Accepted September 15, 2009 Research Article Simultaneous determination of avermectins in bovine tissues by LC-MS/MS Analytical method for the simultaneous quantification of avermectins (AVMs), abamectin B1a, abamectin 8,9-Z isomer B1a, emamectine benzoate B1a, emamectine benzoate 8,9-Z isomer B1a, ivermectin, eprinomectin B1a, doramectin and moxidectin in bovine tissues (muscle, liver and fat) was developed by LC-MS/MS in electrospray positive ion mode. The separation was achieved on a short TSK-GEL ODS 100V column with the mobile phase consisting of acetonitrile and aquatic 0.1 mm ammonium formate containing 0.1 formic acid v/v at a flow rate of 0.2 ml/min with gradient elution. Liquid liquid extraction with isooctane was used for the sample extraction/preparation of analytes in bovine samples. The linearity of the calibration curves was excellent in matrix-matched standards, and yielded the coefficients (r , range from LOQ to 500, 1000 or 5000 ng/g) of determination of the target analytes. Recoveries were in the range of with associated precision values (within-day: , n 5 6, and between-day: for 3 days) for repeatability and reproducibility. LC-MS/MS method has been proven to be highly efficient and suitable for the simultaneous determinations of eight AVMs in bovine tissue samples. Keywords: Avermectins / Bovine tissues / LC-MS/MS / Matrix-matched calibration DOI /jssc Introduction Avermectins (AVMs) are veterinary drugs derived from the soil bacterium Streptomyces avermitilis that belong to the class of macrocyclic lactones. AVMs namely, abamectin (ABM), emamectine (EMA), ivermectin (IVM), eprinomectin (EPR), doramectin (DOR) and moxidectin (MOX), are widely used in food-producing animals to control parasitic diseases. The ABM consists of a mixture of AVM B1a (at least 80) and AVM B1b (not more than 20). In addition, EMA consists of 90 or more of the B1a component and not more than 10 of the B1b component. Each ABM and EMA has its delta-8,9-isomers (8,9-Z-ABM and 8,9-Z-EMA). Based toxicological information in the European Union, the maximum residue limits (MRLs) of AVMs have been fixed by Commission Regulations from 10 to 500 mg/kg for bovine tissues [1] ( Recently, the strict guideline of MRLs for veterinary Correspondence: Dr. Koichi Inoue, Department of Physical and Analytical Chemistry, School of Pharmacy, Kinjo Gakuin University, Omori, Moriyama-ku, Nagoya , Japan kinoue@kinjo-u.ac.jp Fax: Abbreviations: ABM, abamectin; AVMs, avermectins; DOR, doramectin; EMA, emamectine; EPR, eprinomectin; IVM, ivermectin; MEV, matrix effect value; MOX, moxidectin; MRLs, maximum residue limits medicines was executed by Japanese decision for agricultural chemical residues based on European Union, US FDA and global residue levels [2] ( FFCRHOME.nsf/pages/MRLs-p). For example, ABM (sum of AVM B1a and 8,9-Z-AVM B1a: 10 mg/kg for muscle, and 100 mg/kg for fat and liver), EMA (sum of EMA B1a and 8,9- Z-EMA B1a: 2 mg/kg for muscle and fat, and 10 mg/kg for liver), IVM (10 mg/kg for muscle, 40 mg/kg for fat, and 100 mg/kg for liver), EPR (100 mg/kg for muscle, 250 mg/kg for fat, and 2000 mg/kg for liver), DOR (10 mg/kg for muscle, 150 mg/kg for fat, and 100 mg/kg for liver) and MOX (20 mg/ kg for muscle, 500 mg/kg for fat, and 100 mg/kg for liver) in bovine tissues were established based on global MRLs [3] ( html). Therefore, a highly sensitive and selective method for the simultaneous determination of AVMs and 8,9-Z-isomers in bovine tissue samples would be needed. Many researchers have developed LC-MS for purely confirmatory purposes, with LC-FD being used for quantitative determination of AVMs in bovine tissue samples [4]. Danaher et al. suggested that the detection systems for determination of AVMs residues are discussed with a particular emphasis placed on new developments in screening technologies [4]. Based on the recent analytical studies, the papers have been increased using LC-MS/MS for the determination of AVMs in environmental and food samples [5 8]. LC-MS/MS technique has both the confirmatory and the quantitative performance for selective and sensitive analysis of AVMs. A few LC-MS/MS methods were aimed to quantify AVMs in

2 Liquid Chromatography 3597 bovine tissue samples [9 11]. However, no analytical method to date, however, targets the specific AVMs and 8,9-Z-isomers in various bovine samples. The quantitative procedure of LC-MS/MS has a problem that the co-extracting compounds from samples cause the ionization suppression and/or enhance of analytes. Recently, Ferrer et al. reported that the LC-MS/MS quantitation was carried out using matrix-matched calibration for the determination of 100 pesticides in food samples [12]. In addition, Hernando et al. reported that the matrix-matched calibration was proven as reliable alternative to compensate matrix effects for the quantitative analysis of ABM, EMA, IVM and DOR in salmon muscle and pepper samples [7]. Therefore, the matrix-matched calibration would be useful for the quantitative analysis of AVMs in bovine tissues. In this study, we focused on the LC-MS/MS quantitative analysis of AVMs and 8,9-Z-isomers in bovine tissue samples using matrix-matched calibration. Milford, MA). LC column was an TSK-GEL ODS 100 V ( mm, 3 mm: Tosoh, Tokyo, Japan) maintained at 401C. The mobile phase consisted of 0.1 FA and 0.1 mm ammonium formate in water (Solvent A) and 0.1 FA in acetonitrile (Solvent B). The LC gradient was as follows: 50 Solvent B at 0 min, 95 Solvent B at 15 min, 98 Solvent B at 15.1 min, 98 Solvent B at 20 min, and 50 Solvent B at 20.1 min with a flow rate of 0.2 ml/min. The injection volume was 10 ml. The ESI source conditions in the positive ionization mode were as follows: capillary voltage 3.0 kv, extractor voltage 3 V, RF lens voltage 0 V, source temperature of 1201C, and desolvation temperature of 4001C. The cone and desolvation gas flows were 53 and 900 L/h, respectively, and were obtained using a nitrogen source. We used argon as the collision gas and regulated it at 0.35 ml/h, setting the multipliers to 650 V. Cone voltages and collision energies of analytes were summarized in Table 1. 2 Materials and methods 2.1 Chemicals and reagents ABM (purity: 92.3), 8,9-Z-ABM (96.8), EMA (90.7), 8,9- Z-EMA (81.4), IVM (99.2), and EPR (99.3) were obtained from Hayashi Pure Chemical (Osaka, Japan). DOR (81.1) and MOX (92.3) were obtained from Sigma-Aldrich (St. Louis, MO). Pure standard solutions were prepared by diluting an aliquot of the stock solution in methanol. HPLCgrade water, methanol, acetonitrile, formic acid (FA, LC/MSgrade), isooctane, analytical grade ammonium formate, n-hexane, ammonia (25), acetone, and sodium chloride (NaCl) were obtained from Wako Chemical (Osaka, Japan). Purified water was obtained from a Milli-Q purifying system (Millipore, Bedford, MA). Bovine liver, muscle and fat were obtained from a local store in Nagoya, Japan. 2.2 LC-MS/MS analysis LC-MS/MS analyses were performed using a Waters Alliance 2695/Micromass Quattro Premier system (Waters, 2.3 Extraction procedure For simple and easy sample preparation of AVMs in bovine tissues, the liquid liquid extraction (LLE) was adopted based on preliminary report [13]. In this report, the samples are extracted with isooctane, and followed by delipidation using n-hexane [13]. First step, the necessary standard (MRLs levels) was spiked and then stored for 30 min. Then, 5 g of bovine sample were added to 30 ml of acetone/0.5 aquatic ammonia (50:50, v/v) and 5 g of NaCl in 100 ml of glass tube. This mixture was homogenized for 2 min again (homogenizer: YELLOW LINE, D125 basic, IKA Japan K.K., Nara, Japan), and then this sample was added to 60 ml of isooctane. This sample solution was vortex-mixed for 5 min, and then centrifuged at 2500 rpm (2851 g) for 5 min by using Kubota 5420 type (Tokyo, Japan). The organic layer was selected to other glass tube. This procedure was performed two times. This isooctane solution was evaporated to dryness at 301C. Next step, this residual sample was added to 20 ml of hexane, and 20 ml acetonitrile saturated with hexane. This solution was mixed using separatory funnel for 5 min. After storing the solution for separation, the lower solution (acetonitrile) was trans- Table 1. LC-MS/MS conditions of AVMs Compound Identity Precursor ion (m/z) Product ion (m/z) Collision energy (ev) Cone voltage (V) Dwell time (ms) Retention time (min) Avermectine B1a [M1NH 4 ] a) / ,9-Z-Avermectine B1a [M1NH 4 ] a) / Emamectine B1a [M1H] a) / ,9-Z-Emamectine B1a [M1H] a) / Ivermectin [M1NH 4 ] a) / Eprinomectine B1a [M1H] a) / Doramectin [M1NH 4 ] a) / Moxidectin [M1H] a) / a) The most abundant ion (also used for analyte quantification).

3 3598 K. Inoue et al. ferred to another tube and evaporated to dryness at 301C. The sample was redissolved in 5 ml of methanol. 2.4 Analytical validation Recovery Eight AVMs were quantified by a 6-point matrix-matched calibration. Analytical ranges were summarized in Table 2. The calibration curves were prepared daily using stock solutions. The area of each analytes was plotted against analyte concentration. Matrix-matched calibrations were prepared by adding 100 ml of the working standards to individual extracts of control bovine tissue before the last solvent evaporation in the procedure. Spiked levels for the recovery test in bovine muscle, fat and liver samples were selected of the Japanese MRLs in positive list decision for agricultural chemical residues [3]. As for the reason, these MRLs were the strict guideline based on European Union, US FDA and global residue levels [2]. Thus, these MRLs for recovery test would be useful to evaluate the validation of analytical method. However, the total amount of B1a and 8,9-Z-isomers forms was set for the MRLs of ABM and EMA [3]. These evaluations of ABM and EMA were similar to the proposal MRLs in New Zealand [14] ( govt.nz/). The guideline for analytical validation of veterinary medicine in foods in Japan indicated that the 1/2 MRLs could be used if it is difficult to evaluate the main compound in animal tissues [15] (Japanese version: Thus, this study was decided to evaluate the 1/2 MRLs values of each B1a and 8,9-Z-isomer for recovery test Precision Precision was assessed by analyzing spiked bovine tissues at MRLs concentration. For within-day repeatability, MRLs concentration levels (Table 3) were analyzed six times within a day, while for between-day precision, samples of spiked bovine tissues at the same concentrations were analyzed two times per day for three days (n 5 6) Sensitivity The sensitivity of the developed analytical method has been checked in terms of LOD and LOQ. The calculations for LOD were based on signal per noise (control bovine Table 2. LOD, LOQ and matrix-matched calibration of AVMs for bovine samples Analytes Bovine samples LOD (ng/g) a) LOQ (ng/g) b) Matrix-matched calibration c) Linearity (r 2 ) Analytical range (ng/g) Avermectine B1a Muscle Fat Liver ,9-Z-Avermectine B1a Muscle Fat Liver Emamectine B1a Muscle Fat Liver ,9-Z-Emamectine B1a Muscle Fat Liver Ivermectin Muscle Fat Liver Eprinomectine B1a Muscle Fat Liver Doramectin Muscle Fat Liver Moxidectin Muscle Fat Liver a) LOD is S/N 5 3. b) LOQ is S/N410. c) Curve type: linear, Origin: exclude, Weighting: 1/x.

4 Liquid Chromatography 3599 Table 3. Recovery tests of AVMs for bovine samples Analytes Bovine samples Spiked levels (mg/g, ppm) Within-day assay (n 5 6) Between-day assay (n 5 6, 3 days) Mean of recovery () RSD () Mean of recovery () RSD () Avermectine B1a Muscle Fat Liver ,9-Z-Avermectine B1a Muscle Fat Liver Emamectine B1a Muscle Fat Liver ,9-Z-Emamectine B1a Muscle Fat Liver Ivermectin Muscle Fat Liver Eprinomectine B1a Muscle Fat Liver Doramectin Muscle Fat Liver Moxidectin Muscle Fat Liver samples) three times. The calculations for LOQ were based on signal per noise (control bovine sample) 410. Lower concentrations on matrix-matched calibration were used of LOQ for the quantification of AVMs in bovine samples Matrix effects Based on the approaches of Matuszewski et al. [16], the matrix effects were evaluated by comparing the MS/MS responses of standard and test solution. Diluted samples (0.1 mg/g; 0.1 ppm) were used to validate the findings obtained with the diluted standard solution and to verify the absence of a matrix effect. Three bovine samples (n 5 3) were used for the evaluation of matrix effects. The first set of experiments (set 1) was conducted to evaluate the MS/MS response to the injected standard in methanol. The second set of experiments (set 2) was conducted using samples originating from each tissue and spiked after extraction. The validation data obtained in the above manner enabled us to determine the matrix effect value (MEV) for the extraction procedure by comparing the absolute peak areas for the target compound obtained in sets 1 and 2. By calculating the peak areas obtained using a standard solution in set 1 (S1), and the corresponding peak areas for the standards spiked after (set 2: S2) extraction, MEV was calculated as follows: MEV () 5 (S2/S1) 100 [16]. A value greater than 100 indicates ionization enhancement and a value less than 100 indicates ionization suppression. 3 Results and discussion 3.1 LC-MS/MS analysis To measure AVMs using the multiple reaction-monitoring (MRM) mode, full scan and product ion spectra of the analytes were investigated under the LC conditions described in Section 2. AVMs could be detected under ESI-MS conditions following 0.1 FA in mobile phase of water/acetonitrile. From preliminary experiments, it was clear that a positive ionization might give greater sensitivity than the negative mode. However, it seemed possible that the [M1Na] 1 adduct ions of ABM, 8,9-Z-ABM, IVM, and DOR might be unstable and lower responses. Durden reported that the [M1NH 4 ] 1 adduct ions in ESI-positive mode were useful for the detection of these analytes [3]. Therefore, the precursor ions of [M1NH 4 ] 1 were investigated using ammonium formate in mobile phase (solvent A). In this results, the good responses of [M1NH 4 ] 1 of ABM, 8,9-Z-ABM, IVM, and DOR was observed in 0.1 mm ammonium formate and selected to the corresponding these precursor ions of Table 2. When precursor ions of [M

5 3600 K. Inoue et al. 1H] 1 and [M1NH 4 ] 1 were used, the major product ions were observed at m/z (ABM and 8,9-Z-ABM), m/z (EMA, 8,9-Z-EMA), m/z (IVM), m/z (EPR), m/z (DOR), and m/z (MOX), respectively, were observed in daughter scan mode. This result of MS/MS condition was almost same data regarding to Durden s report [8]. Recent reports indicated fast separation with LC-MS/MS because the analytical running was short time [17, 18]. The shortrunning method, however, could not resolve ABM or EMA and these 8,9-Z-isomers. Moreover, a 0.1 mm ammonium formate mobile phase was fixed for this ESI-MS/MS method of AVMs detection, which is not changed for LC separation. Thus, a few of C 18 columns were investigated for the separation of AVMs and 8,9-Z-isomers. In this result, an ESI ionization source with an ammonium formate and FAbased mobile phase and reversed phase C 18 column (TSK- GEL ODS 100V, mm, 3 mm) was chosen for the ionization source and separation of AVMs and 8,9-Zisomers in this experiment. LC-MS/MS analytical factors, such as cone voltage (V) and collision energy (ev), were then investigated to achieve highly sensitive and selective detection of AVMs (Table 1). The MRM chromatograms of standard solution (100 ng/ml) were shown in Fig. 1. These optimal LC-MS/MS condition and validation can be applied to monitoring AVMs in bovine tissue samples. 3.2 Investigation of extraction solutions We think that the procedure is easy and useful to adopt LLE for routine, simple and inexpensive extraction and preparation of analyte in animal tissues. Therefore, the samples are extracted with isooctane [13], followed by delipidation using n-hexane. Preliminary extraction recovery of AVMs from liver samples using water/acetone (50:50 v/v) was over 80 except for EMA (50.2) and 8,9-Z-EMA (59.2). This data indicated that EMA has the different behavior of others AVMs. Based on the constitutional feature of AVMs, only EMA has methylamine group regarding to basic substances. Thus, the influence of ph on the extraction solvents was investigated in liver samples spiked with AVMs. Available volatile basic solution of aquatic ammonia/acetone (50:50) was used for the extraction, and indicated to the greater increase in EMA and 8,9-EMA recoveries than water/ acetone (50:50). In this result, this is fortunate to find the good effects for EMA. Moreover, the preparation with aquatic ammonia/acetone indicated that satisfactory values of other AVMs extraction recoveries were detected using various biological samples. The recovery values of AVMs in bovine tissue samples were listed in Table Matrix effects We evaluated matrix effects of AVMs subjected to the LLE procedure. The AVMs value obtained from spiked extract of 1 EMA 41. ABM ,9-Z-EMA ,9-Z-ABM 13.4 IVM EPR DOR 13.4 MOX > e > e > e > e > e > e Retention Time (min) Figure 1. LC-MS/MS chromatograms of AVMs standard solution in MRM mode. LC-MS/MS: Waters Alliance 2695/Micromass Quattro Premier system (Waters, Milford, MA). LC column: TSK- GEL ODS 100V ( mm, 3 mm: Tosoh, Tokyo, Japan). Mobile phase: 0.1 FA and 0.1 mm ammonium formate in water (solvent A) and 0.1 FA in acetonitrile (solvent B). ESIpositive MRM mode. Muscle Fat Liver Ion suppression Ion enhancement ABM 8,9-Z-ABM EMA 8,9-Z-EMA IVM EPR DOR MOX ME value () Figure 2. Matrix effects of AVMs from bovine tissue samples. By calculating the peak areas obtained using a standard solution (100 ng/ml) in set 1 (S1), and the corresponding peak areas for the standards spiked after (set 2: S2) extraction (100 ppb), MEV was calculated as follows: MEV () 5 (S2/S1) 100. A value greater than 100 indicates ionization enhancement and a value less than 100 indicates ionization suppression. bovine samples (100 ppb) was compared with that of the same analyte prepared in a methanol solution. These values (ME, ) of matrix effect were shown in Fig. 2. These bovine

6 Liquid Chromatography 3601 samples prepared by LLE-induced ion suppression and enhance in ESI-MS/MS. Villagrasa et al. suggested that the use of internal standard, the application of standard dilution method, and the dilution of the extracts before instrumental determination were useful for avoiding matrix signal suppression and/or enhancement [19]. Thus, we tried to study other two methods for avoiding ion suppression because of impossible obtaining the internal standards such as substituted d- and/or 13 C-AVMs from commonly routes. The dilution of extracts indicated that the MEV was better than non-dilution. However, it is suggested to decrease the quantitative and accuracy levels of analytes in samples by using the dilution method. Thus, we tried to study the application of standard dilution (matrix-matched standard) for getting the good-linear quantitative values (Table 2). Based on these results, the matrix-matched standard completely inhibited ion suppression and/or enhancement in ESI-MS/MS detection. 3.4 Recovery test and validations EMA ABM ,9-Z-EMA IVM EPR > e4 89-Z-ABM 8,9 Z > e > e > e6 For food analysis of veterinary residues, the special attention must be paid to the possible determination levels of analytes from various samples according to the MRLs. For recovery test, the results of the sample extraction of AVMs using MRLs concentration levels for bovine samples are summarized in Table 3. The extraction recoveries were indicated to the range of with associated precision values (within-day: , n 5 6, and between-day: for 3 days) for repeatability and reproducibility (Table 3). The recoveries and quantification levels are much higher for AVMs from bovine tissues than other chromatographic methods. The LC-MS/MS chromatograms of AVMs in recovery test of bovine liver sample were shown in Fig. 3. DOR 9.7 MOX > e > e4 3.5 Application for monitoring AVMs in the commercially available samples This developed method was applied to monitoring AVMs in the commercially available bovine samples (muscle n 5 3; liver n 5 3; fat n 5 1) in Japan. The AVMs in all samples was not detected by this analytical method with LC-MS/MS and LLE procedure. 4 Concluding remarks A simple, sensitive, and specific LC-MS/MS method was developed for the quantitative determination of ABM, 8,9-Z- ABM, EMA, 8,9-Z-EMA, IVM, EPR, DOR, and MOX in bovine tissues. This reliability of the LC-MS/MS with matrix-matched standard and LLE procedure was more useful than other chromatographic methods, and fulfilled the validation criteria for the performance of analytical methods and interpretation of results Retention Time (min) Figure 3. LC-MS/MS chromatograms of AVMs for the recovery test from liver sample. Spiked level was shown in Table 3. This study was supported in part by the Development of official analytical methods for the introduction of the positive list system for agricultural chemical residues in foods Research from the Ministry of Health, Labour and Welfare. The authors have declared no conflict of interest. 5 References [1] The European Agency of the Evaluation of Medicinal Products: veterinary medicines evaluation unit. [2] The Japanese Positive List System for Agricultural Chemical Residues in Foods.

7 3602 K. Inoue et al. [3] The Japan Food Chemical Research Foundation: Maximum Residue Limits (MRLs) List of Agricultural Chemicals in Foods. [4] Danaher, M., Howells, L. C., Crooks, S. R., Cerkvenik- Flajs, V., O Keeffe, M., J. Chromatogr. B 2006, 844, [5] Thompson, T. S., Noot, D. K., Forrest, F., van den Heever, J. P. et al., Anal. Chim. Acta 2009, 633, [6] Krogh, K. A., Björklund, E., Loeffler, D., Fink, G. et al., J. Chromatogr. A 2008, 1211, [7] Hernando, M. D., Suárez-Barcena, J. M., Bueno, M. J., Garcia-Reyes, J. F., Fernández-Alba, A. R., J. Chromatogr. A 2007, 1155, [8] Durden, D. A., J. Chromatogr. B 2007, 850, [9] Hou, X., Jiang, H., Ding, S., Zhang, S., Li, X., Shen, J., J. AOAC Int. 2006, 89, [10] Howells, L., Sauer, M. J., Analyst 2001, 126, [11] Kinsella, B., Lehotay, S. J., Mastovska, K., Lightfield, A. R. et al., Anal. Chim. Acta 2009, 637, [12] Ferrer, I., Thurman, E. M., Zweigenbaum, J. A., Rapid Commun. Mass Spectrom. 2007, 21, [13] Danaher, M., O Keeffe, M., Glennon, J. D., Analyst 2000, 125, [14] New Zealand Food Safety Authority. [15] National Institute of Health Sciences in Japan, Division of Foods. [16] Matuszewski, B. K., Constanzer, M. L., Chavez-Eng, C. M., Anal. Chem. 2003, 75, [17] Hernando, M. D., Suárez-Barcena, J. M., Bueno, M. J., Garcia-Reyes, J. F., Fernández-Alba, A. R., J. Chromatogr. A 2007, 1155, [18] Thompson, T. S., Noot, D. K., Forrest, F., van den Heever, J. P. et al., Anal. Chim. Acta 2009, 633, [19] Villagrasa, M., Guillamón, M., Eljarrat, E., Barceló, D., J. Chromatogr. A 2007, 1157,