Determination of Thyreostats in Urine Using Supported Liquid Extraction and Mixed-Mode Cation-Exchange Solid-Phase Extraction: Screening and Confirmatory Methods

Determination of Thyreostats in Urine Using Supported Liquid Extraction and Mixed-Mode... Abstract The application of thyreostats in livestock has been banned in the European Union since 1981, but these drugs are currently in the focus due to the natural occurrence of thiouracil (TU). Studies have been published on TU contamination in urine samples of animal and human origins without any drug administration of it. This paper presents new analytical methods to analyze thyreostats to support the legislation on the recommended concentration (RC) levels of these drugs. Both screening and confirmatory methods are developed for analyzing thyreostats in porcine and bovine urines using a liquid chromatography–tandem mass spectrometry technique. The new methods include a chemical derivatization with 3-iodobenzyl bromide, followed by novel purification approaches using supported liquid extraction and mixed-mode cation-exchange solid-phase extraction (SPE) for screening and confirmatory purposes, respectively. The optimized derivatization in combination with the cation-exchange SPE gives high sensitivity and reducing matrix effect of the analysis. The methods are validated in accordance with the guidelines for the validation of screening methods and European Commission Decision 2002/657/EC. The confirmatory method is used in the national monitoring plan. The detected levels of TU in urine samples are below the currently applicable RC level (10 μg L−1). Introduction Thyreostats decrease the production of thyroid hormones that causes the body weight increase due to enhanced water retention (1, 2). Thyreostats have xenobiotic, carcinogenic and teratogenic properties (3). The European Union (EU) banned the application of thyreostats, namely, tapazole (TAP), thiouracil (TU), methyltiouracil (MTU) and propylthiouracil (PTU) in livestock and set a minimum required performance limit (MRPL) as 100 μg L−1 in urine and 100 μg kg−1 in thyroid gland. This MRPL was changed to 10 μg kg−1 as a recommended concentration (RC) in 2007 (3, 4). Some studies have reported the natural occurrence of TU up to 145 μg L−1 in urine of animal and human origin without drug abuse (5–8). More than 80% of the detected concentrations were below the RC. Therefore, the RC is under further review and the levels may be increased up to 30 μg L−1 in urine (8). The natural contamination of TU in urine may be linked to some feed ingredients such as cabbage, cauliflower and rapeseed (species belonging to the family Brassicaceae) (5). Other studies also investigated the family Brassicaceae as the origin of TU and found correlation between the consumption of Brassicaceae-based feed and presence of TU in urine (8). The glucosinolate content of family Brassicaceae may be the indirect source of TU; the enzymatic (myrosinase) catalyzed metabolization may be responsible for the TU formulation (6, 7, 9). A detailed overview on the natural occurrence of TU summarizes the concentrations detected in six European Member states between 2010 and 2012 (8). The liquid chromatography–tandem mass spectrometry (LC–MS/MS) is the most suitable technique to analyze thyreostats in complex matrices at ≤RC (10–14). The four thyreostats mentioned above are polar and weak basic compounds (logP: −0.62–1.39, pKa: 4.41–8.24) and have small molecule weight (114–170 g mol−1), which give low retention on reversed-phase LC columns and less fragmentation in MS detectors (10). Moreover, the structure of target compounds possesses the instable thio group that has the ability to form different tautomers (Supplementary Table I). Consequently, chemical derivatization has been introduced to analyze thyreostats using 3-iodobenzyl bromide (3-IBBr), or 2,3,4,5,6-pentafluorobenzyl bromide, or 4-chloro-7-nitrobenzofurazan (8, 10, 15–17). After derivatization, the polarity of thyreostats decreases, but their weak basic character remains that can possibly result in higher retention on the LC columns and good sensitivity in the MS/MS with positive ionization (10). The derivatization is also important for the sample clean-up. The LC–MS/MS technique is widely applied to analyze contamination in food, but its application may be limited due to the occurrence of matrix effect in the ion source of the MS system. This effect is caused by the co-eluting matrix solutes that results in ion suppression or ion enhancement. Both sensitivity and precision are also influenced by the matrix effect. Influence of matrix effect can be reduced by optimizing sample clean-up such as solid-phase extraction (SPE) or liquid–liquid extraction (LLE). Matrix effect can also be compensated by using isotopically labeled standards that are used as internal standards (ISTDs). The isotope dilution approach gives accurate results, but it requires the application of all analog ISTDs. Isotopically labeled ISTDs are expensive and the application of multiple ISTDs would increase further the cost of the method. It should be mentioned that the isotopically labeled ISTDs can only compensate the matrix effect, the sensitivity is still influenced by the matrix effect (ion suppression), and consequently the optimized clean-up has a great role in the LC–MS/MS methods. In the lack of isotopically labeled ISTD, purification approaches are needed that utilize orthogonal separation theory to the subsequent LC separation. The application of SPE with ion-exchange cartridge for sample clean-up, and a subsequent reversed-phase high-performance liquid chromatography (HPLC) column for LC separation fulfills the orthogonality. This orthogonal approach can considerably minimize the matrix effect by eliminating several co-eluting matrices. Basically, in this strategy, the target molecules are separated during the clean-up from those background matrix solutes that have other characters (acidic, neutral or basic) toward the analytes. The aim of sample preparation prior to perform LC–MS/MS separation is to reduce the matrix effect and to pre-concentrate the sample. If the co-eluting matrices are not separated from the target compounds during clean-up, the concentration of matrices will also increase in the sample due to the pre-concentration, and hence higher matrix effect could occur in the ion source. In the present study, we report screening and confirmatory methods for thyreostats using an LC–MS/MS technique. The objectives of the current paper are: (i) to optimize a chemical derivatization of thyreostats that is appropriate to achieve separation using an LC–MS/MS method, (ii) to obtain high chromatographic resolution of thyreostats under LC–MS/MS separation conditions, (iii) to develop simple sample preparations for both screening and confirmatory purposes, (iv) to validate the methods, based on EU recommendations and (v) to demonstrate the validity of the developed methods by analyzing thyreostats in urine samples, obtained from the national monitoring system. Experimental Reagents and samples The analytical standards, tapazole (1-methyl-2-imidazolethiol), 2-thiouracil (4-hydroxy-2-mercaptopyrimidine), 6-methyl-2-thiouracil (4-hydroxy-2-mercapto 6-methylpyrimidine), 6-propyl-2-thiouracil (4-hydroxy-2-mercapto-6-propylpyrimidine), 5,6-dimethyl-2-thiouracil (4-hydroxy-2-mercapto 5,6-methylpyrimidine) and 3-iodobenzyl bromide (3-IBBr) were purchased from Sigma-Aldrich (Budapest, Hungary). Stock solutions of each standard (1 mg mL−1) were prepared separately in methanol and were kept at −20°C for up to 3 months. HPLC-grade methanol, acetonitrile, tert-butyl methyl ether (TBME), ethyl acetate (ETAC) were obtained from Promochem (Budapest, Hungary). LC–MS-grade acetonitrile was purchased from VWR (Budapest, Hungary). Suprapur formic acid (98%) and ammonia solution (25%) were acquired from Merck (Budapest, Hungary). Tris-(hydroxymethyl)-aminomethane (TRIS) was obtained from Sigma-Aldrich (Budapest, Hungary). TRIS buffer (0.1 M, pH 10) was prepared by dissolving 12.1 g of TRIS in water and the pH of it was adjusted to 10.0 by adding 1 M sodium hydroxide dropwise. Samples were originated from the Hungarian residue control monitoring program (from January 2016 to December 2016) and were stored at −20°C until subjected to analysis. Equipment and instruments Kinetex XB C-18 HPLC column (100 mm × 3 mm, 2.6 μm) and C-18 security guard column (4 mm × 2 mm); Strata-XL, Strata-XL-C cartridges (200 mg, 3 mL, 100 μm) and Novum SLE cartridges (1-mL tube) were obtained from Gen-lab Ltd (Budapest, Hungary). Filter-Bio syringe filters (polyvinylidene difluoride (PVDF)-L, 13 mm, 0.22 μm) were purchased from Gen-lab Ltd. Centrifugation was performed on a Sigma 3–18 K centrifuge (Osterode am Harz, Germany). A Caliper TurboVap LV (Hopkinton, MA, USA) was employed for sample evaporation. LC–MS/MS separation was performed using an Agilent 6410A LC–MS Triple Quad system (Agilent Technologies, Palo Alto, CA, USA), which includes G1379A degasser, G1312A binary pump, G1329A autosampler, G1316A column thermostat and Agilent 6410A MS/MS detector with ESI ion source. Data acquisition and evaluation were performed using the Agilent Mass Hunter B.01.04 software. Values of logP and pKa were calculated by using Pallas 3.1 software (CompuDrug International, Inc., FL, USA). Chemical derivatization Thyreostats in the urine samples were derivatized by adding 3-IBBr prior to clean-up. Urine samples (5 mL) were weighed in polypropylene centrifuge tubes (50 mL) and were fortified with 5,6-dimethyl-2-thiouracil (DMTU) as a surrogate standard (SSTD) to obtain a concentration of 5 μg L−1. TRIS buffer (0.1 M, 5 mL, pH 10) was added to the samples, followed by 3-IBBr (300 μL, 1 mg mL−1) standard solution. Samples were vortex-mixed for 5 s and derivatized at 37°C for 1 h. After 1 h, samples were let to cool down, vortex-mixed and cleaned up on either supported liquid extraction (SLE) or SPE cartridges depending on the purpose. Supported liquid extraction The derivatized samples were centrifuged at 10,000 rpm at 25°C for 1 min. SLE cartridges (Novum, 1 mL tube) were attached to vacuum manifold and an aliquot (250 μL) of samples was pipetted onto SLE columns. After the cartridges adsorbed the aqueous samples, a gentle vacuum was applied for 1 min, and then the samples completely filled out the sorbents. After waiting 1 min for equilibration, the adsorbed samples were extracted with ETAC–TBME (50/50, v/v) mixture. First, 400 μL of organic mixture was pipetted onto the cartridges. After the solvent passed through the cartridges, another 400 μL was pipetted onto the sorbents. The effluents were collected in glass tubes. The extraction was completed by pipetting 800 μL of organic solution mixture onto the cartridges. When all extraction solvents passed through the sorbents, the cartridges were dried under vacuum for 1 min. The collected effluents (1.6 mL) were evaporated to dryness at 45°C under a gentle stream of nitrogen. The evaporated samples were then re-dissolved in 500 μL of methanol–water mixture (50/50, v/v) by vortex-mixing for 15 s. As the last step, the samples were filtrated through PVDF-L (13 mm, 0.22 μm) syringe filters into HPLC vials. Solid-phase extraction An independent purification approach was developed using SPE for confirmatory analysis. After the chemical derivatization of the sample, the pH was adjusted to below 2.0 by adding HCl solution (0.1 M, 2 mL) to the derivatized samples that were vortex-mixed afterward. Samples were centrifuged at 10,000 rpm at 25°C for 1 min and an aliquot of samples (3 mL) was subsequently cleaned up on Strata-XL-C mixed-mode polymeric strong cation-exchange SPE cartridges (200 mg, 3 mL, 100 μm). SPE columns were first conditioned two times with 3 mL of methanol, followed by 3 mL of water and 3 mL of 0.1 M HCl solution. Then, the samples were slowly passed through dropwise. The cartridges were washed two times with 3 mL of water, followed by 3 mL of methanol. After the washing step, the cartridges were dried under vacuum for 1 min. Samples were eluted two times with 2.5 mL methanol–25% ammonia mixture (95/5, v/v) and collected in glass tubes. The cartridges were dried under vacuum for 1 min, and then the samples were evaporated to dryness at 45°C under a gentle stream of nitrogen. The evaporated samples were then re-dissolved in 500 μL of methanol–water mixture (50/50, v/v) by vortex-mixing for 15 s. In the last step, the samples were filtrated through PVDF-L (13 mm, 0.22 μm) syringe filters into HPLC vials. LC–MS/MS separation Derivatized thyreostats were separated onto Kinetex XB C-18 HPLC column using binary linear gradient elution. Solvent A consisted of 0.1% formic acid (v/v, pH 2.3) in water and solvent B was pure acetonitrile. The flow rate was 0.3 mL/min. The mobile phase contained 25% (v/v) of B at 0 min, 100% of B at 8 min, 100% of B at 12 min and 25% of B at 12.5 min. The stop time was 20 min. The column thermostat was maintained at 30°C. The injection volume was 10 μL. The compounds were iodized using electrospray (ESI) ion source with positive ionization and were detected in the MS/MS detector using the multiple-reaction monitoring (MRM) mode. The ESI settings were as follows: drying gas temperature 350°C, drying gas flow 8 L/min, nebulizer pressure 207 kPa (30 psi), capillary voltage 3000 V. The precursor ions were protonated molecule ions ((M + H)+). The detection parameters are summarized in Table I. Table I. MS/MS Settings for Derivatized Thyreostats Compounds  Time segment (min)  Precursor ion (m/z)  Product ions (m/z)  Fragmentor (V)  Collision energy (V)  ∆EMV (V)  Dwell time (ms)  Ion ratio (%)  Permitted tolerance range (%)  TAP  0–6.0  331.3  217.1  100  30  50  200  17.1  12.0–22.2  113.9  30  200  TU  6.0–20.0  345.2  217.1  100  20  0  100  26.6  20.0–33.3  90.1  30  100  MTU  359.0  217.1  100  20  0  100  21.1  15.8–26.4  90.1  30  100  PTU  387.0  217.1  100  25  0  100  12.5  8.8–16.3  90.1  30  100  DMTU  373.2  217.1  100  30  0  100  19.7  13.8–26.6  90.1  30  100  Compounds  Time segment (min)  Precursor ion (m/z)  Product ions (m/z)  Fragmentor (V)  Collision energy (V)  ∆EMV (V)  Dwell time (ms)  Ion ratio (%)  Permitted tolerance range (%)  TAP  0–6.0  331.3  217.1  100  30  50  200  17.1  12.0–22.2  113.9  30  200  TU  6.0–20.0  345.2  217.1  100  20  0  100  26.6  20.0–33.3  90.1  30  100  MTU  359.0  217.1  100  20  0  100  21.1  15.8–26.4  90.1  30  100  PTU  387.0  217.1  100  25  0  100  12.5  8.8–16.3  90.1  30  100  DMTU  373.2  217.1  100  30  0  100  19.7  13.8–26.6  90.1  30  100  ∆EMV, delta electron multiplier voltage. Quantifier ion transition is highlighted with bold. Table I. MS/MS Settings for Derivatized Thyreostats Compounds  Time segment (min)  Precursor ion (m/z)  Product ions (m/z)  Fragmentor (V)  Collision energy (V)  ∆EMV (V)  Dwell time (ms)  Ion ratio (%)  Permitted tolerance range (%)  TAP  0–6.0  331.3  217.1  100  30  50  200  17.1  12.0–22.2  113.9  30  200  TU  6.0–20.0  345.2  217.1  100  20  0  100  26.6  20.0–33.3  90.1  30  100  MTU  359.0  217.1  100  20  0  100  21.1  15.8–26.4  90.1  30  100  PTU  387.0  217.1  100  25  0  100  12.5  8.8–16.3  90.1  30  100  DMTU  373.2  217.1  100  30  0  100  19.7  13.8–26.6  90.1  30  100  Compounds  Time segment (min)  Precursor ion (m/z)  Product ions (m/z)  Fragmentor (V)  Collision energy (V)  ∆EMV (V)  Dwell time (ms)  Ion ratio (%)  Permitted tolerance range (%)  TAP  0–6.0  331.3  217.1  100  30  50  200  17.1  12.0–22.2  113.9  30  200  TU  6.0–20.0  345.2  217.1  100  20  0  100  26.6  20.0–33.3  90.1  30  100  MTU  359.0  217.1  100  20  0  100  21.1  15.8–26.4  90.1  30  100  PTU  387.0  217.1  100  25  0  100  12.5  8.8–16.3  90.1  30  100  DMTU  373.2  217.1  100  30  0  100  19.7  13.8–26.6  90.1  30  100  ∆EMV, delta electron multiplier voltage. Quantifier ion transition is highlighted with bold. Results Chemical derivatization for thyreostats The optimization of derivatization was carried out by using experimental design. The design of experiments (DoE) was performed with the statistical software R, version 3.0.2 for Windows (http://www.r-project.org). Three parameters were optimized as factors in the DoE, namely, the time, pH and the amount of reagent (3-IBBr). The range of factor level for time (min), pH (unit) and reagent (μL) were 30–60 min, 6–10 and 100–300 μL, respectively. The concentration of 3-IBBr solution (reagent) was 1 mg mL−1. The urine sample used for the experiment was naturally contaminated with TU (1.5 μg L−1) and fortified with the other three native thyreostat standards to obtain 10 μg L−1 concentration (RC level) for these compounds. The response surface methodology was used together with a central composite design to evaluate the effects and interactions of the three variables. Very similar response surfaces were obtained for TU, MTU and PTU due to their same basic structure (Supplementary Table I). Figure 1a,b shows the response surface of TU (slice at 45 min or slice at reagent of 200 μL). The pH and the amount of reagent are highly proportional, the highest responses could be achieved at pH 10.0 with 300 μL reagent; the 45-min reaction time gave the optimum at pH 10 for these three compounds (Figure 1b). In the case of TAP, which has different structure to TU-type compounds, pH has less effect and correlation with reagent. However, the reaction time significantly increases the response of TAP (Figure 1c). Therefore, the optimum of derivatization for all compounds was found to be at pH 10.0, for 1 h and with 300 μL reagent. Figure 1. View largeDownload slide Response surface of TU: slice at 45 min (a) and slice at reagent of 200 μL (b). Response surface of TAP: slice at reagent of 200 μL (c). Figure 1. View largeDownload slide Response surface of TU: slice at 45 min (a) and slice at reagent of 200 μL (b). Response surface of TAP: slice at reagent of 200 μL (c). General conditions for LC–MS/MS separation The ion transitions in the MS/MS detector were tuned using a flow injection analysis and the two essential parameters of ion transitions (fragmentor voltage and collision energy (CE)) were individually optimized with the derivatized standards (18–20). The positive ionization mode was optimized. Thyreostats have basic character that allows enhanced ionization using acidic mobile-phase composition and positive ionization. The acidified eluent was also needed due to the basic character of target compounds to obtain narrower chromatographic peak shapes, thus improving the HPLC separation. The acidic eluent increased the positive ionization which resulted in sensitive precursor ions. The mass spectra of derivatized standards (MS2 spectra) was recorded by scanning with only one analyzer (MS2 scan mode) from 200 to 500 m/z. Protonated molecule ions ([M + H]+) were only seen above the noise range. The precursor ion was then fragmented using a product ion scan mode. The fragmentation of TU, MTU, DMTU and PTU gave the same product ions, the two most intense ions were the 217.1 and 90.0 m/z. The 217.1 m/z also appeared in the mass spectra of TAP. This product ion derives from the derivatization reagent. Afterward, the selected ion transitions were optimized in MRM mode by recoding the mass spectra with different CEs. The CE was optimized (0–30 V) individually for all ion transitions to find the highest responses and signal-to-noise ratio (SNR). Then, the fragmentor voltage was optimized (70–130 V) also in MRM mode. The fragmentor links to the precursor ion, so it was enough to optimize it for one ion transition and could also be applied to the second ion transition of the target compound. Generally, 100 V of fragmentor gave the highest responses and SNR. The application of a high-resolution HPLC column was required because the derivatization decreased the selectivity of MS/MS detection. The most intensive product ion of the derivatized thyreostats (217.1 m/z) originates from the derivatization reagent and thus all derivatized matrices and target compounds also possess this daughter ion. Urine matrix contains several isobaric matrix solutes, if they also reacted with 3-IBBr, also have the X > 217 m/z ion transition (Figure 2). Therefore, baseline separation is needed between target compounds and matrix solutes appeared in the chromatogram. Moreover, enhanced selectivity of HPLC separation was also needed because the SLE clean-up could only lower the number and concentration of the hydrophilic matrix compounds of the samples. Figure 2. View largeDownload slide Total ion chromatogram and the quantifier ion transitions of derivatized thyreostats in urine at 5 μg L−1 concentration. Figure 2. View largeDownload slide Total ion chromatogram and the quantifier ion transitions of derivatized thyreostats in urine at 5 μg L−1 concentration. Supported liquid extraction In this study, the SLE purification of urine was tested with polymeric based SLE cartridge (21). The recovery (n = 3), precision and absolute matrix effect were evaluated at 10 μg L−1 level for all thyreostats including the SSTD under various conditions. The matrix effect was calculated using the Matuszewski approach (22) at one concentration level (10 μg L−1). For the extraction conditions, different solvents were tried as ETAC, ETAC–TBME (50/50, v/v) and TBME. Various sample loading and extraction solvent volumes were investigated. The results are summarized in Table II. Table II. Recovery (n = 3) and Precision Data at 10 μg L−1 Level After SLE Clean-Up with Different Elution Conditions Compound  Extraction with ETAC  Extraction with ETAC–TBME (50/50, v/v)  Extraction with TBME  Matrix effect %  Recovery % (RSD%)  TAP  107 (3.7)  104 (0.2)  107 (2.2)  +23  TU  104 (4.4)  106 (1.2)  103 (2.2)  +3  MTU  110 (2.8)  108 (1.6)  106 (0.6)  +9  PTU  105 (2.2)  102 (1.1)  101 (2.2)  +38  Compound  Extraction with ETAC  Extraction with ETAC–TBME (50/50, v/v)  Extraction with TBME  Matrix effect %  Recovery % (RSD%)  TAP  107 (3.7)  104 (0.2)  107 (2.2)  +23  TU  104 (4.4)  106 (1.2)  103 (2.2)  +3  MTU  110 (2.8)  108 (1.6)  106 (0.6)  +9  PTU  105 (2.2)  102 (1.1)  101 (2.2)  +38  Table II. Recovery (n = 3) and Precision Data at 10 μg L−1 Level After SLE Clean-Up with Different Elution Conditions Compound  Extraction with ETAC  Extraction with ETAC–TBME (50/50, v/v)  Extraction with TBME  Matrix effect %  Recovery % (RSD%)  TAP  107 (3.7)  104 (0.2)  107 (2.2)  +23  TU  104 (4.4)  106 (1.2)  103 (2.2)  +3  MTU  110 (2.8)  108 (1.6)  106 (0.6)  +9  PTU  105 (2.2)  102 (1.1)  101 (2.2)  +38  Compound  Extraction with ETAC  Extraction with ETAC–TBME (50/50, v/v)  Extraction with TBME  Matrix effect %  Recovery % (RSD%)  TAP  107 (3.7)  104 (0.2)  107 (2.2)  +23  TU  104 (4.4)  106 (1.2)  103 (2.2)  +3  MTU  110 (2.8)  108 (1.6)  106 (0.6)  +9  PTU  105 (2.2)  102 (1.1)  101 (2.2)  +38  The recommended loading volume is 200 μL (sample plus diluent) (21); however, the derivatized urine could be adsorbed onto the SLE sorbent up to 250 μL. The recoveries were between 101 and 110% (RSD = 0.2–4.4%) for the target compounds. No considerable change could be seen under different conditions. It should be pointed out that the precision improved when the extraction solution contained TBME. The extraction with ETAC–TBME (50/50, v/v) mixture resulted in the best precision (0.2–1.6%), and hence we used this composition during validation. The matrix effect showed 3–38% ion enhancement depending on the compound of interest. The recovery for DMTU (91–95%) was different from those obtained for the target compounds, but the precision (RSD = 0.2–4.4%) was also high for the SSTD. The point of the application of ISTD in LC–MS/MS analysis is to compensate the matrix effect. As Figure 2 shows, DMTU does not co-elute with the other compounds and hence a different matrix effect (+15%) influences it. Thus, DMTU cannot compensate the matrix effect of other thyreostats and cannot be used as an ISTD. However, DMTU is a good choice for SSTD to check the goodness of derivatization and the entire analysis. The recommended volume for extraction is two times of 600 μL, according to the user guide of Novum (21). However, it was found that the suggested volume is not enough, because target compounds could be extracted from the cartridge with the addition of 600 μL of extraction solvent. When we tried the elution using volumes of 400 μL, 400 μL and 800 μL, no additional amount of target compounds could be extracted from the SLE cartridge. These volumes of extraction solvent were necessary to complete the extraction in our developed method. Solid-phase extraction Thyreostats are weak basic compounds, therefore, we applied mixed-mode polymeric strong cation-exchange SPE cartridge (MCX) for clean-up. At pH below 2.0, the thyreostats link to the sulfonic acid groups of MCX via ionic interaction. Acidic and neutral matrices adsorb on the reversed phase of the cartridge and can be eluted with water and pure methanol. Consequently, only the basic matrix solutes, which concentrate together with the thyreostats, remain in the sample. These remaining matrices are separated from the target compounds on the reversed-phase HPLC column. In this case, the purification was based on ion-exchange theory that is orthogonal to the reversed-phase HPLC separation. The SPE clean-up was also tested using a reversed-phase hydrophilic modified copolymer cartridge (Strata-XL) and pure methanol as an elution solvent. This cartridge also showed good retention for thyreostats. The absolute matrix effect was therefore compared between the two SPE purifications (mixed-mode ion exchange and reversed phase) at levels from 2.5 μg L−1 to 12.5 μg L−1. Matrix-matched calibrations were prepared using post-spiked blank urines. Two calibrations were done using either MCX or the reversed-phase cartridges. The slopes of calibrations were compared to the slope of a matrix-free calibration curve. After MCX clean-up, the matrix effects were −49%, −15%, +10%, −5% and −2% for TU, MTU, PTU, TAP and DMTU, respectively. Negative and positive results mean ion suppression and ion enhancement, respectively. The reversed-phase copolymer SPE resulted in higher matrix effect: −48%, −27%, +22%, −25% and +3% for TU, MTU, PTU, TAP and DMTU, respectively. This can be due to the same separation theory (reversed phase) of purification and LC determination. The matrix effect for TU was the same with both MCX and reversed-phase SPE, suggesting that the co-eluting basic matrix solutes causes the ~50% ion suppression. Again, the matrix effect for DMTU was different to other compounds due to the different retention time, so it could not be used as an ISTD. The best way to compensate the matrix effect is to use isotopically labeled analogs, but it will not be cost effective in this case. In lack of labeled ISTDs, the purification of urine with MCX is suitable for sample clean-up, which gives low matrix effect and selective analysis. DMTU is only applicable as an SSTD. Validation Screening validation Screening method was validated for porcine and bovine samples in line with the CRL guideline. The guideline allows the combination of species (e.g., 10–10 pig and bovine samples) if the method is validated for the same matrix (e.g., urine) (23). Twenty blank samples (10 pig and 10 bovine urines), originated from different sources, were analyzed, followed by fortification of the 20 samples at half of RC level (screening target concentration, 5 μg L−1) before derivatization. The selection of 0.5 × RC as the screening target concentration was based on the recommendation of the guideline (23). The responses obtained in the blank and spiked samples were compared according to CRLs (23). The detection capability (CCβ) and the cut-off level were evaluated for all compounds, except for SSTD that was used only to control the analysis. The responses in blanks did not overlap the response range obtained from the chromatograms of the spiked samples; consequently, the CCβ is equal to the screening target concentration (5 μg L−1) for all thyreostats. This means that 5 μg L−1 can be detected with the screening method with an error of 5%. The cut-off level was determined as the lowest response obtained from the chromatogram of spiked samples. The validation results are summarized in Table III. During the screening analysis, a matrix-matched sample is done at screening target concentration and the response obtained from this sample can be used to set the cut-off level for the analyzed batch (23). If the responses from the test samples of the batch are below the response obtained from the matrix-matched sample, it means that the test samples are complained. In addition, the screening analysis is useful because we can get a semi-quantitative result before confirmation if the sample is positive. Table III. Screening and Confirmatory Validation Results for Thyreostats in Urine Compound  Screening with SLE clean-up  Confirmation with SPE clean-up  CCβ (μg L−1)  The highest response in blank samples  The lowest response in spiked samples  Cut-off (cps)  LOQ (μg L−1)  Recovery (%)  Repeatability (RSD%)  Within-laboratory precision (RSD%)  CCα (μg L−1)  LOQ (μg L−1)  TAP  5  67  1431  1431  2.5  74.5–91.1  2.61–20.5  11.8–19.9  0.8  2.5  TU  5  1902  12,022  12,022  0.5  70–102  3.62–23.5  13.8–23  0.05  0.15  MTU  5  415  35,162  35,162  1  78.2–95.6  4.84–15.6  11.9–15.7  0.1  0.3  PTU  5  3423  43,917  43,917  1  78.6–91.7  4.37–14.3  13.2–16.1  0.15  0.5  Compound  Screening with SLE clean-up  Confirmation with SPE clean-up  CCβ (μg L−1)  The highest response in blank samples  The lowest response in spiked samples  Cut-off (cps)  LOQ (μg L−1)  Recovery (%)  Repeatability (RSD%)  Within-laboratory precision (RSD%)  CCα (μg L−1)  LOQ (μg L−1)  TAP  5  67  1431  1431  2.5  74.5–91.1  2.61–20.5  11.8–19.9  0.8  2.5  TU  5  1902  12,022  12,022  0.5  70–102  3.62–23.5  13.8–23  0.05  0.15  MTU  5  415  35,162  35,162  1  78.2–95.6  4.84–15.6  11.9–15.7  0.1  0.3  PTU  5  3423  43,917  43,917  1  78.6–91.7  4.37–14.3  13.2–16.1  0.15  0.5  Table III. Screening and Confirmatory Validation Results for Thyreostats in Urine Compound  Screening with SLE clean-up  Confirmation with SPE clean-up  CCβ (μg L−1)  The highest response in blank samples  The lowest response in spiked samples  Cut-off (cps)  LOQ (μg L−1)  Recovery (%)  Repeatability (RSD%)  Within-laboratory precision (RSD%)  CCα (μg L−1)  LOQ (μg L−1)  TAP  5  67  1431  1431  2.5  74.5–91.1  2.61–20.5  11.8–19.9  0.8  2.5  TU  5  1902  12,022  12,022  0.5  70–102  3.62–23.5  13.8–23  0.05  0.15  MTU  5  415  35,162  35,162  1  78.2–95.6  4.84–15.6  11.9–15.7  0.1  0.3  PTU  5  3423  43,917  43,917  1  78.6–91.7  4.37–14.3  13.2–16.1  0.15  0.5  Compound  Screening with SLE clean-up  Confirmation with SPE clean-up  CCβ (μg L−1)  The highest response in blank samples  The lowest response in spiked samples  Cut-off (cps)  LOQ (μg L−1)  Recovery (%)  Repeatability (RSD%)  Within-laboratory precision (RSD%)  CCα (μg L−1)  LOQ (μg L−1)  TAP  5  67  1431  1431  2.5  74.5–91.1  2.61–20.5  11.8–19.9  0.8  2.5  TU  5  1902  12,022  12,022  0.5  70–102  3.62–23.5  13.8–23  0.05  0.15  MTU  5  415  35,162  35,162  1  78.2–95.6  4.84–15.6  11.9–15.7  0.1  0.3  PTU  5  3423  43,917  43,917  1  78.6–91.7  4.37–14.3  13.2–16.1  0.15  0.5  Confirmatory validation Confirmatory method was validated according to EU 2002/657/EC guideline (24). The selectivity was proven by comparing chromatograms obtained from spiked and blank samples (Figures 2 and 3). The identification was based on the ion ratio that is the intensity ratio of qualifier and quantifier ion transitions (Table I). The permitted tolerance ranges for ion ratios were calculated in accordance with Commission Decision 2002/657/EC (24). The matrix-matched calibration levels were 2.5, 5, 7.5, 10 and 12.5 μg L−1. The determination coefficient (r2) was higher than 0.9422. Six different blank samples, which were a mixture of pig and bovine urines (1/1, v/v), were fortified at three levels before derivatization and analyses. This was repeated on additional 2 days under different conditions (e.g., different operators, lots of solvents and cartridges). The spiking levels were 5 μg L−1 (0.5 × RC), 7.5 μg L−1 (0.75 × RC) and 10 μg L−1 (1 × RC). The recovery and within-laboratory precision were evaluated at all levels from 18 results per level. The repeatability was calculated from the data of the first day using six results per level. The repeatability ranged from 2.61% to 23.5%. The recovery was between 70 and 102% for the four target compounds and the within-laboratory precision ranged from 11.8% to 23%. These data meet the requirements: recovery should be between 70% and 110% and the RSD should be below 30%. The recovery of SSTD was 81.8–96.4%. The calculation of decision limit (CCα) was based on the SNR and evaluated as three times of SNR. The limit of quantification (LOQ) was calculated as ten times of SNR (Table III). The analytical limits were verified by spiking blank samples at the estimated concentrations with six replicates and analyzed. The CCα and LOQ were confirmed by obtaining SNR higher than 3 and 10 in all samples, respectively. Figure 3. View largeDownload slide Ion transitions of TU in a naturally contaminated and blank urine samples. Figure 3. View largeDownload slide Ion transitions of TU in a naturally contaminated and blank urine samples. The robustness was evaluated using Plackett–Burman design using two-level screening method using statistical software R. Four factors, expected to have an impact on the reliability of results, were tested. The chosen factors were: derivatization reagent volume (285–315 μL), derivatization time (57–63 min), pH adjustment before SPE clean-up (1.9–2.1 ml HCl) and elution solvent volume (4.75–5.25 mL). Blank samples were fortified at 10 μg L−1 level for each analyte. The tested ranges for the factors were chosen considering the highest level of the uncertainty under the current experimental conditions. Half-normal plots and interaction plots drawn for the different responses revealed no critical effects, which were statistically significant confirming the robustness of the method. Real-sample analysis The Food Toxicological National Reference Laboratory has been accredited for the measurement of thyreostat in urine by LC–MS/MS since 2009; the used method was based on a published paper (17). This method utilizes derivatization with 3-IBBr, followed by LLE and SPE clean-up on silica-based normal-phase cartridges. After the adoption of this earlier method, CCα between 5.2 μg L−1 (TU) and 10 μg L−1 (TAP) could be achieved and the low level of natural contamination of samples with TU could not be detected. The improvement of the method with MCX clean-up allowed to reduce the CCα in conjunction with LOQ. Some urine samples were randomly selected and analyzed with the presented confirmatory method at this location in 2016. A quantity of pig urines contained low levels of TU in the range from 0.148 μg L−1 to 3.79 μg L−1. Most of them were below 2 μg L−1. TU was also measured in bovine urines in lesser samples between 0.139 μg L−1 and 4.39 μg L−1. These levels are similar to those found by other researchers in the naturally contaminated urine samples (5, 8). Discussion Sample stabilization prior to preparation Based on a recent collaborative study on TU, one laboratory suggests the stabilization of samples before sample preparation that can be performed by adding HCl (37%) and ethylenediaminetetraacetic acid (EDTA) (0.25 M) to the sample to decrease the pH below 3 and to eliminate the copper (8). This stabilization is recommended by Bussche et al. who extensively studied the effects of preservation, number of freeze–thaw cycles, and matrix-related variables on the stability of thyreostats in urine (16). Most of the methods in the collaborative study (six approaches out of eight), however, utilized only sample freezing at -20°C and immediate derivatization after melting the sample (8). These laboratories passed the EU-RL proficiency test (PT) on thyreostats in urine in 2013. Only one out of eight laboratories applied stabilization for TU, but all other participants also detected satisfactory concentrations in the PT (8). We did not add HCl and EDTA to the samples before freezing and did not observe problem so far, however, the stabilization of urine shall be studied in the future to obtain more information about those relevant factors that can influence the thyreostat contamination of samples. The reduction of thyreostat concentration in urine can also be caused to the farming style (8) or the endogenous composition of urine that is based on the geographical origin of sample (e.g., feeding). This was not studied up to now. Method development In this study, the thyreostats were determined with precolumn derivatization. The aim of derivatization was (i) to reduce the polarity of thyreostats, (ii) to increase the sensitivity of MS/MS detection and (iii) to stabilize the unstable thiol group on the structure of target compounds. According to the literature, different derivatization agents could be used for thyreostats (introduction section) from which 3-IBBr was selected because the reaction was sufficiently fast. Additionally, the thyreostat standards, derivatized with 3-IBBr, showed high sensitivity in neat solution during the MS/MS detection with positive ionization due to the basic character of target compounds. The general conditions for derivatization of thyreostats in urine with 3-IBBr were as follows: 100 μL of 3-IBBr (2 mg mL−1 or 5 mg mL−1), temperature 40°C, time 1 h, pH 8 (5, 8, 10, 15–17). Based on the experimental designed, which was carried out for thyreostat derivatization in urine, however, we found that pH of 10 can considerably increase the efficiency of derivatization. An overnight derivatization (16 h) at pH 10.0 with 300 μL reagent was compared to the optimum time (1 h). There was no considerable difference in the responses under these two conditions. This suggests that the reaction is almost finished at pH 10 after 1 h. This reaction time (1 h) is the same that was found by other researchers. HPLC columns containing core–shell packing material are known for their high resolution, good batch-to-batch reproducibility and robustness (18, 20). In this work, a HPLC column, especially developed for basic molecules, was also tested for derivatized thyreostats. Kinetex XB C-18 has been used successfully in our previous studies to analyze antibiotics, steroids and mycotoxins (18–20). This column has shown good selectivity and sensitivity in complex matrices such as body fluids, foods of animal and plant origin by separating the co-eluting matrix solutes, and thus reducing the matrix effect of LC–MS/MS analysis. Consequently, during the method development in the present study, we applied the Kinetex XB HPLC column packed with C-18 core–shell particles for thyreostats for the first time. The separation on core–shell column together with gradient elution enabled narrow peak shapes that further improved the analytical limits. Furthermore, the HPLC conditions allowed the separation of derivatized thyreostats on baseline (Figure 2) and hence the crosstalk effect could be eliminated. The crosstalk effect may appear when the co-eluting target compounds possess same fragment ions and in this case the fragment ion of one compound may be detected on the mass channel of another target compound. Novum SLE cartridges filled with polymeric packing material were introduced recently (21). The synthetic sorbent has superior batch-to-batch reproducibility. SLE can also be carried out with diatomaceous earth-based adsorbents, but it was not tested in this work. This clean-up is suitable only for screening purpose, because background matrices cannot be fully eliminated with SLE, so it does not provide sufficient purification. Therefore, a more robust clean-up is needed for confirmation. The SLE can only be used for aqueous samples; the cartridge completely adsorbs the sample up to a certain amount. The target compounds can be extracted from the adsorbed sample using immiscible organic solvents such as TBME, ETAC or chlorinated hydrocarbons. We found that both TBME and ETAC are suitable solvents for extraction and also the TBME–ETAC mixture. Even though this clean-up approach eliminates only the polar matrices as proteins, salts and other hydrophilic compounds from the sample, it provides a simple, fast, precise and cost-effective clean-up for screening purpose. For confirmatory analysis of thyreostats, a more effective clean-up approach was used to pre-concentrate the sample and to eliminate the disturbing matrices. SPE is generally applied for sample purification (8, 10). If the separation theory of SPE is orthogonal to the subsequent HPLC separation, the best selectivity and sensitivity can be achieved during the LC–MS/MS analysis. In this approach, those co-eluting matrices are removed from the samples that have different characters to the target compounds (e.g., acidic and neutral matrices). During the LC–MS/MS analysis, we used the reversed-phase separation on C-18 column, therefore the purification was orthogonal. The combination of cation-exchange SPE with reversed-phase LC–MS/MS separation on column packed with core–shell particles is described for thyreostats for the first time. In the case of MS/MS detection, it is important to obtain relative clean samples before detection to eliminate the matrix effect, thus the analytical limits can be lowered. Comparison of the present LC–MS method to existing ones Method developments for thyreostats in urine have been done by research groups to investigate the natural occurrence of TU and to set up methods for the analysis of banned thyreostats by LC–MS (Table IV). There are two ways to treat the urine samples containing thyreostats before analysis: one is with derivatization and the other is without. The methods with derivatization can stabilize the tautomer form of compounds analyzed and enhance the retention in LC separation as well as improve the fragmentation during MS/MS detection. Most of the papers suggest derivatization with 3-IBBr at pH 8 (5, 6, 8, 10, 15, 17, 25). In the present study, we found that the higher pH can increase the responses of derivatized thyreostats and pH 10 is the optimum obtained from the experimental design. The disadvantage of derivatization is the higher noise and matrix effect; and also the less selectivity caused by those isobaric matrix solutes that also react with the derivatization agent. Sample purifications therefore have been introduced in the methods to eliminate the number and concentration of matrices. LLE was enough only in one study that applied derivatization (25), mainly LLE in combination with normal-phase SPE was needed for clean-up. The consecutive extraction (LLE plus SPE) can lead to losses during the purification that increase the analytical limits. Moreover, this approach needs relatively high amount of organic solvents and is also time consuming. Methods without derivatization applied only LLE after a reduction step with DL-dithiothreitol at pH 7.0 (7, 26). The presented method utilizes a novel SPE purification with mixed-mode polymeric strong cation-exchange cartridge (MCX). This clean-up eliminates the matrices having acidic and neutral characters; therefore, an LLE step was not necessary to minimize the matrix effect of analysis; the method is thus less time consuming and is cost effective. The pre-concentrated basic solutes, however, caused ~50% ion suppression for TU. The MCX purification resulted in high extraction efficiency and clean extracts that enabled to lower the CCα below 1 μg L−1 for the first time (Table IV). The lowest TU concentration that could be confirmed with acceptable ion ratio was 0.148 μg L−1 (Figure 3). A novel screening method using SLE is also described. This clean-up is fast and inexpensive, but eliminates only the polar matrices, consequently it is recommended only for screening purpose. Table IV. Existing and the Presented LC–MS Methods for Thyreostats in Urine Analyzed compounds  Derivatization reagent  Purification  Technique  Analytical limit (μg L−1)  Reference  TU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with negative ionization  LOD = 0.8  5  TU  3-IBBr (pH 8)  LLE and normal phase SPE  LC–MS2 (iontrap) with negative ionization  LOQ = 1  6  TU    LLE  LC–MS/MS with positive ionization  CCα = 2.2  7  TU  3-IBBr (pH 8), or 2,3,4,5,6-pentafluorobenzyl bromide, or 4-chloro-7-nitrobenzofurazan  LLE and SPE normal phase SPE  LC–MS/MS with positive ionization  The highest CCα = 3.9  8  Seven thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with positive/negative ionization  CCα = 0.6–4.1  15  Eight thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with negative ionization  CCα = 0.1–5.1  17  Five thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE  LC–MS2 (iontrap) with positive ionization and LC–MS/MS with positive ionization  CCα = 0.84–1.24  25  Eight thyreostats including TAP, TU, MTU, PTU    LLE  LC–MS/MS with positive ionization  CCα = 1.1–5.5  26  TAP, TU, MTU, PTU  3-IBBr (pH 10)  Cation-exchange SPE for confirmation SLE for screening  LC–MS/MS with positive ionization  CCα = 0.05–0.8, CCβ = 5  This study  Analyzed compounds  Derivatization reagent  Purification  Technique  Analytical limit (μg L−1)  Reference  TU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with negative ionization  LOD = 0.8  5  TU  3-IBBr (pH 8)  LLE and normal phase SPE  LC–MS2 (iontrap) with negative ionization  LOQ = 1  6  TU    LLE  LC–MS/MS with positive ionization  CCα = 2.2  7  TU  3-IBBr (pH 8), or 2,3,4,5,6-pentafluorobenzyl bromide, or 4-chloro-7-nitrobenzofurazan  LLE and SPE normal phase SPE  LC–MS/MS with positive ionization  The highest CCα = 3.9  8  Seven thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with positive/negative ionization  CCα = 0.6–4.1  15  Eight thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with negative ionization  CCα = 0.1–5.1  17  Five thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE  LC–MS2 (iontrap) with positive ionization and LC–MS/MS with positive ionization  CCα = 0.84–1.24  25  Eight thyreostats including TAP, TU, MTU, PTU    LLE  LC–MS/MS with positive ionization  CCα = 1.1–5.5  26  TAP, TU, MTU, PTU  3-IBBr (pH 10)  Cation-exchange SPE for confirmation SLE for screening  LC–MS/MS with positive ionization  CCα = 0.05–0.8, CCβ = 5  This study  Table IV. Existing and the Presented LC–MS Methods for Thyreostats in Urine Analyzed compounds  Derivatization reagent  Purification  Technique  Analytical limit (μg L−1)  Reference  TU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with negative ionization  LOD = 0.8  5  TU  3-IBBr (pH 8)  LLE and normal phase SPE  LC–MS2 (iontrap) with negative ionization  LOQ = 1  6  TU    LLE  LC–MS/MS with positive ionization  CCα = 2.2  7  TU  3-IBBr (pH 8), or 2,3,4,5,6-pentafluorobenzyl bromide, or 4-chloro-7-nitrobenzofurazan  LLE and SPE normal phase SPE  LC–MS/MS with positive ionization  The highest CCα = 3.9  8  Seven thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with positive/negative ionization  CCα = 0.6–4.1  15  Eight thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with negative ionization  CCα = 0.1–5.1  17  Five thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE  LC–MS2 (iontrap) with positive ionization and LC–MS/MS with positive ionization  CCα = 0.84–1.24  25  Eight thyreostats including TAP, TU, MTU, PTU    LLE  LC–MS/MS with positive ionization  CCα = 1.1–5.5  26  TAP, TU, MTU, PTU  3-IBBr (pH 10)  Cation-exchange SPE for confirmation SLE for screening  LC–MS/MS with positive ionization  CCα = 0.05–0.8, CCβ = 5  This study  Analyzed compounds  Derivatization reagent  Purification  Technique  Analytical limit (μg L−1)  Reference  TU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with negative ionization  LOD = 0.8  5  TU  3-IBBr (pH 8)  LLE and normal phase SPE  LC–MS2 (iontrap) with negative ionization  LOQ = 1  6  TU    LLE  LC–MS/MS with positive ionization  CCα = 2.2  7  TU  3-IBBr (pH 8), or 2,3,4,5,6-pentafluorobenzyl bromide, or 4-chloro-7-nitrobenzofurazan  LLE and SPE normal phase SPE  LC–MS/MS with positive ionization  The highest CCα = 3.9  8  Seven thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with positive/negative ionization  CCα = 0.6–4.1  15  Eight thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with negative ionization  CCα = 0.1–5.1  17  Five thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE  LC–MS2 (iontrap) with positive ionization and LC–MS/MS with positive ionization  CCα = 0.84–1.24  25  Eight thyreostats including TAP, TU, MTU, PTU    LLE  LC–MS/MS with positive ionization  CCα = 1.1–5.5  26  TAP, TU, MTU, PTU  3-IBBr (pH 10)  Cation-exchange SPE for confirmation SLE for screening  LC–MS/MS with positive ionization  CCα = 0.05–0.8, CCβ = 5  This study  Conclusions A new LC–MS/MS method for both screening and confirmation of thyreostats in urine is developed. The method uses an optimized chemical derivatization with 3-IBBr prior to clean-up. The sample purification is performed on SLE and MCX SPE cartridges for screening and confirmatory approaches, respectively. The derivatized thyreostats were separated on HPLC column packed with core–shell particles. The confirmatory method allowed reducing the analytical limits (CCα below 1 μg L−1) and naturally contamination of urine with TU could be detected. The methods are validated according to the EU recommendations and the confirmatory method is currently used in the national monitoring. Supplementary Data Supplementary material is available at Journal of Chromatographic Science online. Acknowledgments Authors wish to thank Éva Pálffy, Tímea Horváth and Viktória Lipcsei for their help in the sample preparations. Conflict of interest statement The authors have no conflict of interest on the submission of the manuscript. References 1 Gentili, A., De Pratti, V., Caretti, F., Pérez-Fernández, V., Tomai, P., Curini, R., et al.  .; Residue analysis of thyreostats in baby foods via matrix solid phase dispersion and liquid chromatography–dual-polarity electrospray–tandem mass spectrometry; Food Additive and Contaminants , ( 2016); 33: 1793– 1802. Google Scholar CrossRef Search ADS   2 Bertuzzi, T., Rastelli, S., Mulazzi, A., Pietri, A.; Trace monitoring of goitrin and thiouracil in milk and cheese; Food Analytical Methods , ( 2016); 9: 2952– 2959. Google Scholar CrossRef Search ADS   3 European Union. Council Directive 81/602/EC; Official Journal of the European Union, ( 1981); L/222. 4 CRL Guidance Paper.—CRLs view on state of the art analytical methods for the national residue control plans; ( 2007); http://www.rivm.nl/bibliotheek/digitaaldepot/crlguidance2007.pdf 5 Pinel, G., Mathieu, S., Cesbron, N., Maume, D., De Brabander, H.F., Andre, F., et al.  .; Evidence that urinary excretion of thiouracil in adult bovine submitted to a cruciferous diet can give erroneous indications of the possible illegal use of thyrostats in meat production; Food Additive and Contaminants , ( 2006); 23: 974– 980. Google Scholar CrossRef Search ADS   6 Vanden Bussche, J., Kiebooms, J.A.L., De Clercq, N., Deceuninck, Y., Le Bizec, B., De Brabander, H.F., et al.  .; Feed or food responsible for the presence of low-level thiouracil in urine of livestock and humans?; Journal of Agricultural and Food Chemistry , ( 2011); 59: 5786– 5792. Google Scholar CrossRef Search ADS PubMed  7 Vanden Bussche, J., Vanhaecke, L., Deceuninck, Y., Wille, K., Bekaert, K., Le Bizec, B., et al.  .; Ultra-high performance liquid chromatography coupled to triple quadrupole mass spectrometry detection of naturally occurring thiouracil in urine of untreated livestock, domesticated animals and humans; Food Additive and Contaminants , ( 2011); 28: 166– 172. Google Scholar CrossRef Search ADS   8 Wauters, J., Vanden Bussche, J., Le Bizec, B., Kiebooms, J.A.L., Dervilly-Pinel, G., Prevost, S., et al.  .; Toward a new European threshold to discriminate illegally administered from naturally occurring thiouracil in livestock; Journal of Agricultural and Food Chemistry , ( 2015); 63: 1339– 1346. Google Scholar CrossRef Search ADS PubMed  9 Kiebooms, J.A., Vanden Bussche, J., Hemeryck, L.Y., Fievez, V., Vanhaecke, L.; Intestinal microbiota contribute to the endogenous formation of thiouracil in livestock; Journal of Agricultural and Food Chemistry , ( 2012); 60: 7769– 7776. Google Scholar CrossRef Search ADS PubMed  10 Vanden Bussche, J., Noppe, H., Verheyden, K., Wille, K., Pinel, G., Le Bizec, B., et al.  .; Analysis of thyreostats: A history of 35 years; Analytica Chimica Acta , ( 2009); 637: 2– 12. Google Scholar CrossRef Search ADS PubMed  11 Lega, F., Contiero, L., Biancotto, G., Angeletti, R.; Determination of thyreostats in muscle and thyroid tissues by QuEChERS extraction and ultra-performance liquid chromatography tandem mass spectrometry; Food Additive and Contaminants , ( 2013); 30: 949– 957. Google Scholar CrossRef Search ADS   12 Chiesa, L.M., Labella, G.F., Pasquale, E., Panseri, S., Pavlovic, R., Arioli, F.; Determination of thyreostats in bovine urine and thyroid glands by HPLC–MS/MS; Chromatographia , ( 2016); 79: 591– 599. Google Scholar CrossRef Search ADS   13 Schmidt, K.S.; In-house validation and factorial effect analysis of a liquid chromatography-tandem mass spectrometry method for the determination of thyreostats in bovine blood plasma; Analytical and Bioanalytical Chemistry , ( 2014); 406: 735– 743. Google Scholar CrossRef Search ADS PubMed  14 Kiebooms, J., Wauters, J., Vanden Bussche, J., Vanhaecke, L.; Validation of a quantitative method using liquid chromatography coupled to multiple mass spectrometry for thiouracil in feedstuffs used in animal husbandry; Analytical and Bioanalytical Chemistry , ( 2015); 407: 4376– 4384. Google Scholar CrossRef Search ADS   15 Lõhmus, M., Kallaste, K., Le Bizec, B.; Determination of thyreostats in urine and thyroid gland by ultra high performance liquid chromatography tandem mass spectrometry; Journal of Chromatography. A , ( 2009); 1216: 8080– 8089. Google Scholar CrossRef Search ADS PubMed  16 Vanden Bussche, J., Sterk, S.S., De Brabander, H.F., Blokland, M.H., Deceuninck, Y., Le Bizec, B., et al.  .; Thyreostatic drugs, stability in bovine and porcine urine; Analytical and Bioanalytical Chemistry , ( 2012); 403( 10): 2973– 2982. Google Scholar CrossRef Search ADS PubMed  17 Pinel, G., Bichon, E., Pouponneau, K., Maume, D., André, F., Le Bizec, B.; Multi-residue method for the determination of thyreostats in urine samples using liquid chromatography coupled to tandem mass spectrometry after derivatisation with 3-iodobenzylbromide; Journal of Chromatography A , ( 2005); 1085: 247– 252. Google Scholar CrossRef Search ADS PubMed  18 Tölgyesi, Á., Sharma, V.K., Fekete, J.; Confirmatory analysis of stanozolol metabolites in bovine, pig and sheep urines using an optimized clean-up and liquid chromatography-tandem mass spectrometry; Journal of Pharmaceutical and Biomedical Analysis , ( 2014); 88: 45– 52. Google Scholar CrossRef Search ADS PubMed  19 Tölgyesi, Á., Kunsági, Z.; Quantification of T-2 and HT-2 mycotoxins in cereals by liquid chromatography-multimode ionization-tandem mass spectrometry; Microchemical Journal , ( 2013); 106: 300– 306. Google Scholar CrossRef Search ADS   20 Tölgyesi, Á., Sharma, V.K., Fekete, S., Fekete, J., Simon, A., Farkas, S.; Development of a rapid method for the determination and confirmation of nitroimidazoles in six matrices by fast liquid chromatography-tandem mass spectrometry; Journal of Pharmaceutical and Biomedical Analysis , ( 2012); 64–65: 40– 48. Google Scholar CrossRef Search ADS PubMed  21 Novum simplified liquid extraction User Guide; ( 2015); https://az621941.vo.msecnd.net/documents/7a8aa89b-c764-4f0f-ab75-f85459f260e3.pdf 22 Matuszewski, B.K., Constanzer, M.L., Chavez-Eng, C.M.; Strategies for the assessment of matrix effect in quantitative bioanalytical methods based on HPLC−MS/MS; Analytical Chemistry , ( 2003); 75: 3019– 3030. Google Scholar CrossRef Search ADS PubMed  23 CRLs. ( 2010): Guidelines for the validation of screening methods for residues of veterinary medicines (Initial validation and transfer). Community Reference Laboratories 20/1/2010. http://ec.europa.eu/food/food/chemicalsafety/residues/Guideline_Validation_Screening_en.pdf 24 Commission Decision 2002/657/EC; Official Journal of the European Union, ( 2002); L221/8. 25 Wozniak, B., Zuchowska, I.M., Zmudzki, J., Jedziniak, P., Korycinska, B., Sielska, K., et al.  .; Liquid chromatography tandem mass spectrometry with ion trap and triple quadrupole analyzers for determination of thyreostatic drugs in urine and muscle tissue; Analytica Chimica Acta , ( 2011); 700: 155– 166. Google Scholar CrossRef Search ADS PubMed  26 Vanden Bussche, J., Vanhaecke, L., Deceuninck, Y., Verheyden, K., Wille, K., Bekaert, K., et al.  .; Development and validation of an ultra-high performance liquid chromatography tandem mass spectrometry method for quantifying thyreostats in urine without derivatisation; Journal of Chromatography A , ( 2010); 1217: 4285– 4293. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Chromatographic Science Oxford University Press

Determination of Thyreostats in Urine Using Supported Liquid Extraction and Mixed-Mode Cation-Exchange Solid-Phase Extraction: Screening and Confirmatory Methods

Loading next page...
 
/lp/ou_press/determination-of-thyreostats-in-urine-using-supported-liquid-4VjyGhnwQ0
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com
ISSN
0021-9665
eISSN
1945-239X
D.O.I.
10.1093/chromsci/bmy054
Publisher site
See Article on Publisher Site

Abstract

Abstract The application of thyreostats in livestock has been banned in the European Union since 1981, but these drugs are currently in the focus due to the natural occurrence of thiouracil (TU). Studies have been published on TU contamination in urine samples of animal and human origins without any drug administration of it. This paper presents new analytical methods to analyze thyreostats to support the legislation on the recommended concentration (RC) levels of these drugs. Both screening and confirmatory methods are developed for analyzing thyreostats in porcine and bovine urines using a liquid chromatography–tandem mass spectrometry technique. The new methods include a chemical derivatization with 3-iodobenzyl bromide, followed by novel purification approaches using supported liquid extraction and mixed-mode cation-exchange solid-phase extraction (SPE) for screening and confirmatory purposes, respectively. The optimized derivatization in combination with the cation-exchange SPE gives high sensitivity and reducing matrix effect of the analysis. The methods are validated in accordance with the guidelines for the validation of screening methods and European Commission Decision 2002/657/EC. The confirmatory method is used in the national monitoring plan. The detected levels of TU in urine samples are below the currently applicable RC level (10 μg L−1). Introduction Thyreostats decrease the production of thyroid hormones that causes the body weight increase due to enhanced water retention (1, 2). Thyreostats have xenobiotic, carcinogenic and teratogenic properties (3). The European Union (EU) banned the application of thyreostats, namely, tapazole (TAP), thiouracil (TU), methyltiouracil (MTU) and propylthiouracil (PTU) in livestock and set a minimum required performance limit (MRPL) as 100 μg L−1 in urine and 100 μg kg−1 in thyroid gland. This MRPL was changed to 10 μg kg−1 as a recommended concentration (RC) in 2007 (3, 4). Some studies have reported the natural occurrence of TU up to 145 μg L−1 in urine of animal and human origin without drug abuse (5–8). More than 80% of the detected concentrations were below the RC. Therefore, the RC is under further review and the levels may be increased up to 30 μg L−1 in urine (8). The natural contamination of TU in urine may be linked to some feed ingredients such as cabbage, cauliflower and rapeseed (species belonging to the family Brassicaceae) (5). Other studies also investigated the family Brassicaceae as the origin of TU and found correlation between the consumption of Brassicaceae-based feed and presence of TU in urine (8). The glucosinolate content of family Brassicaceae may be the indirect source of TU; the enzymatic (myrosinase) catalyzed metabolization may be responsible for the TU formulation (6, 7, 9). A detailed overview on the natural occurrence of TU summarizes the concentrations detected in six European Member states between 2010 and 2012 (8). The liquid chromatography–tandem mass spectrometry (LC–MS/MS) is the most suitable technique to analyze thyreostats in complex matrices at ≤RC (10–14). The four thyreostats mentioned above are polar and weak basic compounds (logP: −0.62–1.39, pKa: 4.41–8.24) and have small molecule weight (114–170 g mol−1), which give low retention on reversed-phase LC columns and less fragmentation in MS detectors (10). Moreover, the structure of target compounds possesses the instable thio group that has the ability to form different tautomers (Supplementary Table I). Consequently, chemical derivatization has been introduced to analyze thyreostats using 3-iodobenzyl bromide (3-IBBr), or 2,3,4,5,6-pentafluorobenzyl bromide, or 4-chloro-7-nitrobenzofurazan (8, 10, 15–17). After derivatization, the polarity of thyreostats decreases, but their weak basic character remains that can possibly result in higher retention on the LC columns and good sensitivity in the MS/MS with positive ionization (10). The derivatization is also important for the sample clean-up. The LC–MS/MS technique is widely applied to analyze contamination in food, but its application may be limited due to the occurrence of matrix effect in the ion source of the MS system. This effect is caused by the co-eluting matrix solutes that results in ion suppression or ion enhancement. Both sensitivity and precision are also influenced by the matrix effect. Influence of matrix effect can be reduced by optimizing sample clean-up such as solid-phase extraction (SPE) or liquid–liquid extraction (LLE). Matrix effect can also be compensated by using isotopically labeled standards that are used as internal standards (ISTDs). The isotope dilution approach gives accurate results, but it requires the application of all analog ISTDs. Isotopically labeled ISTDs are expensive and the application of multiple ISTDs would increase further the cost of the method. It should be mentioned that the isotopically labeled ISTDs can only compensate the matrix effect, the sensitivity is still influenced by the matrix effect (ion suppression), and consequently the optimized clean-up has a great role in the LC–MS/MS methods. In the lack of isotopically labeled ISTD, purification approaches are needed that utilize orthogonal separation theory to the subsequent LC separation. The application of SPE with ion-exchange cartridge for sample clean-up, and a subsequent reversed-phase high-performance liquid chromatography (HPLC) column for LC separation fulfills the orthogonality. This orthogonal approach can considerably minimize the matrix effect by eliminating several co-eluting matrices. Basically, in this strategy, the target molecules are separated during the clean-up from those background matrix solutes that have other characters (acidic, neutral or basic) toward the analytes. The aim of sample preparation prior to perform LC–MS/MS separation is to reduce the matrix effect and to pre-concentrate the sample. If the co-eluting matrices are not separated from the target compounds during clean-up, the concentration of matrices will also increase in the sample due to the pre-concentration, and hence higher matrix effect could occur in the ion source. In the present study, we report screening and confirmatory methods for thyreostats using an LC–MS/MS technique. The objectives of the current paper are: (i) to optimize a chemical derivatization of thyreostats that is appropriate to achieve separation using an LC–MS/MS method, (ii) to obtain high chromatographic resolution of thyreostats under LC–MS/MS separation conditions, (iii) to develop simple sample preparations for both screening and confirmatory purposes, (iv) to validate the methods, based on EU recommendations and (v) to demonstrate the validity of the developed methods by analyzing thyreostats in urine samples, obtained from the national monitoring system. Experimental Reagents and samples The analytical standards, tapazole (1-methyl-2-imidazolethiol), 2-thiouracil (4-hydroxy-2-mercaptopyrimidine), 6-methyl-2-thiouracil (4-hydroxy-2-mercapto 6-methylpyrimidine), 6-propyl-2-thiouracil (4-hydroxy-2-mercapto-6-propylpyrimidine), 5,6-dimethyl-2-thiouracil (4-hydroxy-2-mercapto 5,6-methylpyrimidine) and 3-iodobenzyl bromide (3-IBBr) were purchased from Sigma-Aldrich (Budapest, Hungary). Stock solutions of each standard (1 mg mL−1) were prepared separately in methanol and were kept at −20°C for up to 3 months. HPLC-grade methanol, acetonitrile, tert-butyl methyl ether (TBME), ethyl acetate (ETAC) were obtained from Promochem (Budapest, Hungary). LC–MS-grade acetonitrile was purchased from VWR (Budapest, Hungary). Suprapur formic acid (98%) and ammonia solution (25%) were acquired from Merck (Budapest, Hungary). Tris-(hydroxymethyl)-aminomethane (TRIS) was obtained from Sigma-Aldrich (Budapest, Hungary). TRIS buffer (0.1 M, pH 10) was prepared by dissolving 12.1 g of TRIS in water and the pH of it was adjusted to 10.0 by adding 1 M sodium hydroxide dropwise. Samples were originated from the Hungarian residue control monitoring program (from January 2016 to December 2016) and were stored at −20°C until subjected to analysis. Equipment and instruments Kinetex XB C-18 HPLC column (100 mm × 3 mm, 2.6 μm) and C-18 security guard column (4 mm × 2 mm); Strata-XL, Strata-XL-C cartridges (200 mg, 3 mL, 100 μm) and Novum SLE cartridges (1-mL tube) were obtained from Gen-lab Ltd (Budapest, Hungary). Filter-Bio syringe filters (polyvinylidene difluoride (PVDF)-L, 13 mm, 0.22 μm) were purchased from Gen-lab Ltd. Centrifugation was performed on a Sigma 3–18 K centrifuge (Osterode am Harz, Germany). A Caliper TurboVap LV (Hopkinton, MA, USA) was employed for sample evaporation. LC–MS/MS separation was performed using an Agilent 6410A LC–MS Triple Quad system (Agilent Technologies, Palo Alto, CA, USA), which includes G1379A degasser, G1312A binary pump, G1329A autosampler, G1316A column thermostat and Agilent 6410A MS/MS detector with ESI ion source. Data acquisition and evaluation were performed using the Agilent Mass Hunter B.01.04 software. Values of logP and pKa were calculated by using Pallas 3.1 software (CompuDrug International, Inc., FL, USA). Chemical derivatization Thyreostats in the urine samples were derivatized by adding 3-IBBr prior to clean-up. Urine samples (5 mL) were weighed in polypropylene centrifuge tubes (50 mL) and were fortified with 5,6-dimethyl-2-thiouracil (DMTU) as a surrogate standard (SSTD) to obtain a concentration of 5 μg L−1. TRIS buffer (0.1 M, 5 mL, pH 10) was added to the samples, followed by 3-IBBr (300 μL, 1 mg mL−1) standard solution. Samples were vortex-mixed for 5 s and derivatized at 37°C for 1 h. After 1 h, samples were let to cool down, vortex-mixed and cleaned up on either supported liquid extraction (SLE) or SPE cartridges depending on the purpose. Supported liquid extraction The derivatized samples were centrifuged at 10,000 rpm at 25°C for 1 min. SLE cartridges (Novum, 1 mL tube) were attached to vacuum manifold and an aliquot (250 μL) of samples was pipetted onto SLE columns. After the cartridges adsorbed the aqueous samples, a gentle vacuum was applied for 1 min, and then the samples completely filled out the sorbents. After waiting 1 min for equilibration, the adsorbed samples were extracted with ETAC–TBME (50/50, v/v) mixture. First, 400 μL of organic mixture was pipetted onto the cartridges. After the solvent passed through the cartridges, another 400 μL was pipetted onto the sorbents. The effluents were collected in glass tubes. The extraction was completed by pipetting 800 μL of organic solution mixture onto the cartridges. When all extraction solvents passed through the sorbents, the cartridges were dried under vacuum for 1 min. The collected effluents (1.6 mL) were evaporated to dryness at 45°C under a gentle stream of nitrogen. The evaporated samples were then re-dissolved in 500 μL of methanol–water mixture (50/50, v/v) by vortex-mixing for 15 s. As the last step, the samples were filtrated through PVDF-L (13 mm, 0.22 μm) syringe filters into HPLC vials. Solid-phase extraction An independent purification approach was developed using SPE for confirmatory analysis. After the chemical derivatization of the sample, the pH was adjusted to below 2.0 by adding HCl solution (0.1 M, 2 mL) to the derivatized samples that were vortex-mixed afterward. Samples were centrifuged at 10,000 rpm at 25°C for 1 min and an aliquot of samples (3 mL) was subsequently cleaned up on Strata-XL-C mixed-mode polymeric strong cation-exchange SPE cartridges (200 mg, 3 mL, 100 μm). SPE columns were first conditioned two times with 3 mL of methanol, followed by 3 mL of water and 3 mL of 0.1 M HCl solution. Then, the samples were slowly passed through dropwise. The cartridges were washed two times with 3 mL of water, followed by 3 mL of methanol. After the washing step, the cartridges were dried under vacuum for 1 min. Samples were eluted two times with 2.5 mL methanol–25% ammonia mixture (95/5, v/v) and collected in glass tubes. The cartridges were dried under vacuum for 1 min, and then the samples were evaporated to dryness at 45°C under a gentle stream of nitrogen. The evaporated samples were then re-dissolved in 500 μL of methanol–water mixture (50/50, v/v) by vortex-mixing for 15 s. In the last step, the samples were filtrated through PVDF-L (13 mm, 0.22 μm) syringe filters into HPLC vials. LC–MS/MS separation Derivatized thyreostats were separated onto Kinetex XB C-18 HPLC column using binary linear gradient elution. Solvent A consisted of 0.1% formic acid (v/v, pH 2.3) in water and solvent B was pure acetonitrile. The flow rate was 0.3 mL/min. The mobile phase contained 25% (v/v) of B at 0 min, 100% of B at 8 min, 100% of B at 12 min and 25% of B at 12.5 min. The stop time was 20 min. The column thermostat was maintained at 30°C. The injection volume was 10 μL. The compounds were iodized using electrospray (ESI) ion source with positive ionization and were detected in the MS/MS detector using the multiple-reaction monitoring (MRM) mode. The ESI settings were as follows: drying gas temperature 350°C, drying gas flow 8 L/min, nebulizer pressure 207 kPa (30 psi), capillary voltage 3000 V. The precursor ions were protonated molecule ions ((M + H)+). The detection parameters are summarized in Table I. Table I. MS/MS Settings for Derivatized Thyreostats Compounds  Time segment (min)  Precursor ion (m/z)  Product ions (m/z)  Fragmentor (V)  Collision energy (V)  ∆EMV (V)  Dwell time (ms)  Ion ratio (%)  Permitted tolerance range (%)  TAP  0–6.0  331.3  217.1  100  30  50  200  17.1  12.0–22.2  113.9  30  200  TU  6.0–20.0  345.2  217.1  100  20  0  100  26.6  20.0–33.3  90.1  30  100  MTU  359.0  217.1  100  20  0  100  21.1  15.8–26.4  90.1  30  100  PTU  387.0  217.1  100  25  0  100  12.5  8.8–16.3  90.1  30  100  DMTU  373.2  217.1  100  30  0  100  19.7  13.8–26.6  90.1  30  100  Compounds  Time segment (min)  Precursor ion (m/z)  Product ions (m/z)  Fragmentor (V)  Collision energy (V)  ∆EMV (V)  Dwell time (ms)  Ion ratio (%)  Permitted tolerance range (%)  TAP  0–6.0  331.3  217.1  100  30  50  200  17.1  12.0–22.2  113.9  30  200  TU  6.0–20.0  345.2  217.1  100  20  0  100  26.6  20.0–33.3  90.1  30  100  MTU  359.0  217.1  100  20  0  100  21.1  15.8–26.4  90.1  30  100  PTU  387.0  217.1  100  25  0  100  12.5  8.8–16.3  90.1  30  100  DMTU  373.2  217.1  100  30  0  100  19.7  13.8–26.6  90.1  30  100  ∆EMV, delta electron multiplier voltage. Quantifier ion transition is highlighted with bold. Table I. MS/MS Settings for Derivatized Thyreostats Compounds  Time segment (min)  Precursor ion (m/z)  Product ions (m/z)  Fragmentor (V)  Collision energy (V)  ∆EMV (V)  Dwell time (ms)  Ion ratio (%)  Permitted tolerance range (%)  TAP  0–6.0  331.3  217.1  100  30  50  200  17.1  12.0–22.2  113.9  30  200  TU  6.0–20.0  345.2  217.1  100  20  0  100  26.6  20.0–33.3  90.1  30  100  MTU  359.0  217.1  100  20  0  100  21.1  15.8–26.4  90.1  30  100  PTU  387.0  217.1  100  25  0  100  12.5  8.8–16.3  90.1  30  100  DMTU  373.2  217.1  100  30  0  100  19.7  13.8–26.6  90.1  30  100  Compounds  Time segment (min)  Precursor ion (m/z)  Product ions (m/z)  Fragmentor (V)  Collision energy (V)  ∆EMV (V)  Dwell time (ms)  Ion ratio (%)  Permitted tolerance range (%)  TAP  0–6.0  331.3  217.1  100  30  50  200  17.1  12.0–22.2  113.9  30  200  TU  6.0–20.0  345.2  217.1  100  20  0  100  26.6  20.0–33.3  90.1  30  100  MTU  359.0  217.1  100  20  0  100  21.1  15.8–26.4  90.1  30  100  PTU  387.0  217.1  100  25  0  100  12.5  8.8–16.3  90.1  30  100  DMTU  373.2  217.1  100  30  0  100  19.7  13.8–26.6  90.1  30  100  ∆EMV, delta electron multiplier voltage. Quantifier ion transition is highlighted with bold. Results Chemical derivatization for thyreostats The optimization of derivatization was carried out by using experimental design. The design of experiments (DoE) was performed with the statistical software R, version 3.0.2 for Windows (http://www.r-project.org). Three parameters were optimized as factors in the DoE, namely, the time, pH and the amount of reagent (3-IBBr). The range of factor level for time (min), pH (unit) and reagent (μL) were 30–60 min, 6–10 and 100–300 μL, respectively. The concentration of 3-IBBr solution (reagent) was 1 mg mL−1. The urine sample used for the experiment was naturally contaminated with TU (1.5 μg L−1) and fortified with the other three native thyreostat standards to obtain 10 μg L−1 concentration (RC level) for these compounds. The response surface methodology was used together with a central composite design to evaluate the effects and interactions of the three variables. Very similar response surfaces were obtained for TU, MTU and PTU due to their same basic structure (Supplementary Table I). Figure 1a,b shows the response surface of TU (slice at 45 min or slice at reagent of 200 μL). The pH and the amount of reagent are highly proportional, the highest responses could be achieved at pH 10.0 with 300 μL reagent; the 45-min reaction time gave the optimum at pH 10 for these three compounds (Figure 1b). In the case of TAP, which has different structure to TU-type compounds, pH has less effect and correlation with reagent. However, the reaction time significantly increases the response of TAP (Figure 1c). Therefore, the optimum of derivatization for all compounds was found to be at pH 10.0, for 1 h and with 300 μL reagent. Figure 1. View largeDownload slide Response surface of TU: slice at 45 min (a) and slice at reagent of 200 μL (b). Response surface of TAP: slice at reagent of 200 μL (c). Figure 1. View largeDownload slide Response surface of TU: slice at 45 min (a) and slice at reagent of 200 μL (b). Response surface of TAP: slice at reagent of 200 μL (c). General conditions for LC–MS/MS separation The ion transitions in the MS/MS detector were tuned using a flow injection analysis and the two essential parameters of ion transitions (fragmentor voltage and collision energy (CE)) were individually optimized with the derivatized standards (18–20). The positive ionization mode was optimized. Thyreostats have basic character that allows enhanced ionization using acidic mobile-phase composition and positive ionization. The acidified eluent was also needed due to the basic character of target compounds to obtain narrower chromatographic peak shapes, thus improving the HPLC separation. The acidic eluent increased the positive ionization which resulted in sensitive precursor ions. The mass spectra of derivatized standards (MS2 spectra) was recorded by scanning with only one analyzer (MS2 scan mode) from 200 to 500 m/z. Protonated molecule ions ([M + H]+) were only seen above the noise range. The precursor ion was then fragmented using a product ion scan mode. The fragmentation of TU, MTU, DMTU and PTU gave the same product ions, the two most intense ions were the 217.1 and 90.0 m/z. The 217.1 m/z also appeared in the mass spectra of TAP. This product ion derives from the derivatization reagent. Afterward, the selected ion transitions were optimized in MRM mode by recoding the mass spectra with different CEs. The CE was optimized (0–30 V) individually for all ion transitions to find the highest responses and signal-to-noise ratio (SNR). Then, the fragmentor voltage was optimized (70–130 V) also in MRM mode. The fragmentor links to the precursor ion, so it was enough to optimize it for one ion transition and could also be applied to the second ion transition of the target compound. Generally, 100 V of fragmentor gave the highest responses and SNR. The application of a high-resolution HPLC column was required because the derivatization decreased the selectivity of MS/MS detection. The most intensive product ion of the derivatized thyreostats (217.1 m/z) originates from the derivatization reagent and thus all derivatized matrices and target compounds also possess this daughter ion. Urine matrix contains several isobaric matrix solutes, if they also reacted with 3-IBBr, also have the X > 217 m/z ion transition (Figure 2). Therefore, baseline separation is needed between target compounds and matrix solutes appeared in the chromatogram. Moreover, enhanced selectivity of HPLC separation was also needed because the SLE clean-up could only lower the number and concentration of the hydrophilic matrix compounds of the samples. Figure 2. View largeDownload slide Total ion chromatogram and the quantifier ion transitions of derivatized thyreostats in urine at 5 μg L−1 concentration. Figure 2. View largeDownload slide Total ion chromatogram and the quantifier ion transitions of derivatized thyreostats in urine at 5 μg L−1 concentration. Supported liquid extraction In this study, the SLE purification of urine was tested with polymeric based SLE cartridge (21). The recovery (n = 3), precision and absolute matrix effect were evaluated at 10 μg L−1 level for all thyreostats including the SSTD under various conditions. The matrix effect was calculated using the Matuszewski approach (22) at one concentration level (10 μg L−1). For the extraction conditions, different solvents were tried as ETAC, ETAC–TBME (50/50, v/v) and TBME. Various sample loading and extraction solvent volumes were investigated. The results are summarized in Table II. Table II. Recovery (n = 3) and Precision Data at 10 μg L−1 Level After SLE Clean-Up with Different Elution Conditions Compound  Extraction with ETAC  Extraction with ETAC–TBME (50/50, v/v)  Extraction with TBME  Matrix effect %  Recovery % (RSD%)  TAP  107 (3.7)  104 (0.2)  107 (2.2)  +23  TU  104 (4.4)  106 (1.2)  103 (2.2)  +3  MTU  110 (2.8)  108 (1.6)  106 (0.6)  +9  PTU  105 (2.2)  102 (1.1)  101 (2.2)  +38  Compound  Extraction with ETAC  Extraction with ETAC–TBME (50/50, v/v)  Extraction with TBME  Matrix effect %  Recovery % (RSD%)  TAP  107 (3.7)  104 (0.2)  107 (2.2)  +23  TU  104 (4.4)  106 (1.2)  103 (2.2)  +3  MTU  110 (2.8)  108 (1.6)  106 (0.6)  +9  PTU  105 (2.2)  102 (1.1)  101 (2.2)  +38  Table II. Recovery (n = 3) and Precision Data at 10 μg L−1 Level After SLE Clean-Up with Different Elution Conditions Compound  Extraction with ETAC  Extraction with ETAC–TBME (50/50, v/v)  Extraction with TBME  Matrix effect %  Recovery % (RSD%)  TAP  107 (3.7)  104 (0.2)  107 (2.2)  +23  TU  104 (4.4)  106 (1.2)  103 (2.2)  +3  MTU  110 (2.8)  108 (1.6)  106 (0.6)  +9  PTU  105 (2.2)  102 (1.1)  101 (2.2)  +38  Compound  Extraction with ETAC  Extraction with ETAC–TBME (50/50, v/v)  Extraction with TBME  Matrix effect %  Recovery % (RSD%)  TAP  107 (3.7)  104 (0.2)  107 (2.2)  +23  TU  104 (4.4)  106 (1.2)  103 (2.2)  +3  MTU  110 (2.8)  108 (1.6)  106 (0.6)  +9  PTU  105 (2.2)  102 (1.1)  101 (2.2)  +38  The recommended loading volume is 200 μL (sample plus diluent) (21); however, the derivatized urine could be adsorbed onto the SLE sorbent up to 250 μL. The recoveries were between 101 and 110% (RSD = 0.2–4.4%) for the target compounds. No considerable change could be seen under different conditions. It should be pointed out that the precision improved when the extraction solution contained TBME. The extraction with ETAC–TBME (50/50, v/v) mixture resulted in the best precision (0.2–1.6%), and hence we used this composition during validation. The matrix effect showed 3–38% ion enhancement depending on the compound of interest. The recovery for DMTU (91–95%) was different from those obtained for the target compounds, but the precision (RSD = 0.2–4.4%) was also high for the SSTD. The point of the application of ISTD in LC–MS/MS analysis is to compensate the matrix effect. As Figure 2 shows, DMTU does not co-elute with the other compounds and hence a different matrix effect (+15%) influences it. Thus, DMTU cannot compensate the matrix effect of other thyreostats and cannot be used as an ISTD. However, DMTU is a good choice for SSTD to check the goodness of derivatization and the entire analysis. The recommended volume for extraction is two times of 600 μL, according to the user guide of Novum (21). However, it was found that the suggested volume is not enough, because target compounds could be extracted from the cartridge with the addition of 600 μL of extraction solvent. When we tried the elution using volumes of 400 μL, 400 μL and 800 μL, no additional amount of target compounds could be extracted from the SLE cartridge. These volumes of extraction solvent were necessary to complete the extraction in our developed method. Solid-phase extraction Thyreostats are weak basic compounds, therefore, we applied mixed-mode polymeric strong cation-exchange SPE cartridge (MCX) for clean-up. At pH below 2.0, the thyreostats link to the sulfonic acid groups of MCX via ionic interaction. Acidic and neutral matrices adsorb on the reversed phase of the cartridge and can be eluted with water and pure methanol. Consequently, only the basic matrix solutes, which concentrate together with the thyreostats, remain in the sample. These remaining matrices are separated from the target compounds on the reversed-phase HPLC column. In this case, the purification was based on ion-exchange theory that is orthogonal to the reversed-phase HPLC separation. The SPE clean-up was also tested using a reversed-phase hydrophilic modified copolymer cartridge (Strata-XL) and pure methanol as an elution solvent. This cartridge also showed good retention for thyreostats. The absolute matrix effect was therefore compared between the two SPE purifications (mixed-mode ion exchange and reversed phase) at levels from 2.5 μg L−1 to 12.5 μg L−1. Matrix-matched calibrations were prepared using post-spiked blank urines. Two calibrations were done using either MCX or the reversed-phase cartridges. The slopes of calibrations were compared to the slope of a matrix-free calibration curve. After MCX clean-up, the matrix effects were −49%, −15%, +10%, −5% and −2% for TU, MTU, PTU, TAP and DMTU, respectively. Negative and positive results mean ion suppression and ion enhancement, respectively. The reversed-phase copolymer SPE resulted in higher matrix effect: −48%, −27%, +22%, −25% and +3% for TU, MTU, PTU, TAP and DMTU, respectively. This can be due to the same separation theory (reversed phase) of purification and LC determination. The matrix effect for TU was the same with both MCX and reversed-phase SPE, suggesting that the co-eluting basic matrix solutes causes the ~50% ion suppression. Again, the matrix effect for DMTU was different to other compounds due to the different retention time, so it could not be used as an ISTD. The best way to compensate the matrix effect is to use isotopically labeled analogs, but it will not be cost effective in this case. In lack of labeled ISTDs, the purification of urine with MCX is suitable for sample clean-up, which gives low matrix effect and selective analysis. DMTU is only applicable as an SSTD. Validation Screening validation Screening method was validated for porcine and bovine samples in line with the CRL guideline. The guideline allows the combination of species (e.g., 10–10 pig and bovine samples) if the method is validated for the same matrix (e.g., urine) (23). Twenty blank samples (10 pig and 10 bovine urines), originated from different sources, were analyzed, followed by fortification of the 20 samples at half of RC level (screening target concentration, 5 μg L−1) before derivatization. The selection of 0.5 × RC as the screening target concentration was based on the recommendation of the guideline (23). The responses obtained in the blank and spiked samples were compared according to CRLs (23). The detection capability (CCβ) and the cut-off level were evaluated for all compounds, except for SSTD that was used only to control the analysis. The responses in blanks did not overlap the response range obtained from the chromatograms of the spiked samples; consequently, the CCβ is equal to the screening target concentration (5 μg L−1) for all thyreostats. This means that 5 μg L−1 can be detected with the screening method with an error of 5%. The cut-off level was determined as the lowest response obtained from the chromatogram of spiked samples. The validation results are summarized in Table III. During the screening analysis, a matrix-matched sample is done at screening target concentration and the response obtained from this sample can be used to set the cut-off level for the analyzed batch (23). If the responses from the test samples of the batch are below the response obtained from the matrix-matched sample, it means that the test samples are complained. In addition, the screening analysis is useful because we can get a semi-quantitative result before confirmation if the sample is positive. Table III. Screening and Confirmatory Validation Results for Thyreostats in Urine Compound  Screening with SLE clean-up  Confirmation with SPE clean-up  CCβ (μg L−1)  The highest response in blank samples  The lowest response in spiked samples  Cut-off (cps)  LOQ (μg L−1)  Recovery (%)  Repeatability (RSD%)  Within-laboratory precision (RSD%)  CCα (μg L−1)  LOQ (μg L−1)  TAP  5  67  1431  1431  2.5  74.5–91.1  2.61–20.5  11.8–19.9  0.8  2.5  TU  5  1902  12,022  12,022  0.5  70–102  3.62–23.5  13.8–23  0.05  0.15  MTU  5  415  35,162  35,162  1  78.2–95.6  4.84–15.6  11.9–15.7  0.1  0.3  PTU  5  3423  43,917  43,917  1  78.6–91.7  4.37–14.3  13.2–16.1  0.15  0.5  Compound  Screening with SLE clean-up  Confirmation with SPE clean-up  CCβ (μg L−1)  The highest response in blank samples  The lowest response in spiked samples  Cut-off (cps)  LOQ (μg L−1)  Recovery (%)  Repeatability (RSD%)  Within-laboratory precision (RSD%)  CCα (μg L−1)  LOQ (μg L−1)  TAP  5  67  1431  1431  2.5  74.5–91.1  2.61–20.5  11.8–19.9  0.8  2.5  TU  5  1902  12,022  12,022  0.5  70–102  3.62–23.5  13.8–23  0.05  0.15  MTU  5  415  35,162  35,162  1  78.2–95.6  4.84–15.6  11.9–15.7  0.1  0.3  PTU  5  3423  43,917  43,917  1  78.6–91.7  4.37–14.3  13.2–16.1  0.15  0.5  Table III. Screening and Confirmatory Validation Results for Thyreostats in Urine Compound  Screening with SLE clean-up  Confirmation with SPE clean-up  CCβ (μg L−1)  The highest response in blank samples  The lowest response in spiked samples  Cut-off (cps)  LOQ (μg L−1)  Recovery (%)  Repeatability (RSD%)  Within-laboratory precision (RSD%)  CCα (μg L−1)  LOQ (μg L−1)  TAP  5  67  1431  1431  2.5  74.5–91.1  2.61–20.5  11.8–19.9  0.8  2.5  TU  5  1902  12,022  12,022  0.5  70–102  3.62–23.5  13.8–23  0.05  0.15  MTU  5  415  35,162  35,162  1  78.2–95.6  4.84–15.6  11.9–15.7  0.1  0.3  PTU  5  3423  43,917  43,917  1  78.6–91.7  4.37–14.3  13.2–16.1  0.15  0.5  Compound  Screening with SLE clean-up  Confirmation with SPE clean-up  CCβ (μg L−1)  The highest response in blank samples  The lowest response in spiked samples  Cut-off (cps)  LOQ (μg L−1)  Recovery (%)  Repeatability (RSD%)  Within-laboratory precision (RSD%)  CCα (μg L−1)  LOQ (μg L−1)  TAP  5  67  1431  1431  2.5  74.5–91.1  2.61–20.5  11.8–19.9  0.8  2.5  TU  5  1902  12,022  12,022  0.5  70–102  3.62–23.5  13.8–23  0.05  0.15  MTU  5  415  35,162  35,162  1  78.2–95.6  4.84–15.6  11.9–15.7  0.1  0.3  PTU  5  3423  43,917  43,917  1  78.6–91.7  4.37–14.3  13.2–16.1  0.15  0.5  Confirmatory validation Confirmatory method was validated according to EU 2002/657/EC guideline (24). The selectivity was proven by comparing chromatograms obtained from spiked and blank samples (Figures 2 and 3). The identification was based on the ion ratio that is the intensity ratio of qualifier and quantifier ion transitions (Table I). The permitted tolerance ranges for ion ratios were calculated in accordance with Commission Decision 2002/657/EC (24). The matrix-matched calibration levels were 2.5, 5, 7.5, 10 and 12.5 μg L−1. The determination coefficient (r2) was higher than 0.9422. Six different blank samples, which were a mixture of pig and bovine urines (1/1, v/v), were fortified at three levels before derivatization and analyses. This was repeated on additional 2 days under different conditions (e.g., different operators, lots of solvents and cartridges). The spiking levels were 5 μg L−1 (0.5 × RC), 7.5 μg L−1 (0.75 × RC) and 10 μg L−1 (1 × RC). The recovery and within-laboratory precision were evaluated at all levels from 18 results per level. The repeatability was calculated from the data of the first day using six results per level. The repeatability ranged from 2.61% to 23.5%. The recovery was between 70 and 102% for the four target compounds and the within-laboratory precision ranged from 11.8% to 23%. These data meet the requirements: recovery should be between 70% and 110% and the RSD should be below 30%. The recovery of SSTD was 81.8–96.4%. The calculation of decision limit (CCα) was based on the SNR and evaluated as three times of SNR. The limit of quantification (LOQ) was calculated as ten times of SNR (Table III). The analytical limits were verified by spiking blank samples at the estimated concentrations with six replicates and analyzed. The CCα and LOQ were confirmed by obtaining SNR higher than 3 and 10 in all samples, respectively. Figure 3. View largeDownload slide Ion transitions of TU in a naturally contaminated and blank urine samples. Figure 3. View largeDownload slide Ion transitions of TU in a naturally contaminated and blank urine samples. The robustness was evaluated using Plackett–Burman design using two-level screening method using statistical software R. Four factors, expected to have an impact on the reliability of results, were tested. The chosen factors were: derivatization reagent volume (285–315 μL), derivatization time (57–63 min), pH adjustment before SPE clean-up (1.9–2.1 ml HCl) and elution solvent volume (4.75–5.25 mL). Blank samples were fortified at 10 μg L−1 level for each analyte. The tested ranges for the factors were chosen considering the highest level of the uncertainty under the current experimental conditions. Half-normal plots and interaction plots drawn for the different responses revealed no critical effects, which were statistically significant confirming the robustness of the method. Real-sample analysis The Food Toxicological National Reference Laboratory has been accredited for the measurement of thyreostat in urine by LC–MS/MS since 2009; the used method was based on a published paper (17). This method utilizes derivatization with 3-IBBr, followed by LLE and SPE clean-up on silica-based normal-phase cartridges. After the adoption of this earlier method, CCα between 5.2 μg L−1 (TU) and 10 μg L−1 (TAP) could be achieved and the low level of natural contamination of samples with TU could not be detected. The improvement of the method with MCX clean-up allowed to reduce the CCα in conjunction with LOQ. Some urine samples were randomly selected and analyzed with the presented confirmatory method at this location in 2016. A quantity of pig urines contained low levels of TU in the range from 0.148 μg L−1 to 3.79 μg L−1. Most of them were below 2 μg L−1. TU was also measured in bovine urines in lesser samples between 0.139 μg L−1 and 4.39 μg L−1. These levels are similar to those found by other researchers in the naturally contaminated urine samples (5, 8). Discussion Sample stabilization prior to preparation Based on a recent collaborative study on TU, one laboratory suggests the stabilization of samples before sample preparation that can be performed by adding HCl (37%) and ethylenediaminetetraacetic acid (EDTA) (0.25 M) to the sample to decrease the pH below 3 and to eliminate the copper (8). This stabilization is recommended by Bussche et al. who extensively studied the effects of preservation, number of freeze–thaw cycles, and matrix-related variables on the stability of thyreostats in urine (16). Most of the methods in the collaborative study (six approaches out of eight), however, utilized only sample freezing at -20°C and immediate derivatization after melting the sample (8). These laboratories passed the EU-RL proficiency test (PT) on thyreostats in urine in 2013. Only one out of eight laboratories applied stabilization for TU, but all other participants also detected satisfactory concentrations in the PT (8). We did not add HCl and EDTA to the samples before freezing and did not observe problem so far, however, the stabilization of urine shall be studied in the future to obtain more information about those relevant factors that can influence the thyreostat contamination of samples. The reduction of thyreostat concentration in urine can also be caused to the farming style (8) or the endogenous composition of urine that is based on the geographical origin of sample (e.g., feeding). This was not studied up to now. Method development In this study, the thyreostats were determined with precolumn derivatization. The aim of derivatization was (i) to reduce the polarity of thyreostats, (ii) to increase the sensitivity of MS/MS detection and (iii) to stabilize the unstable thiol group on the structure of target compounds. According to the literature, different derivatization agents could be used for thyreostats (introduction section) from which 3-IBBr was selected because the reaction was sufficiently fast. Additionally, the thyreostat standards, derivatized with 3-IBBr, showed high sensitivity in neat solution during the MS/MS detection with positive ionization due to the basic character of target compounds. The general conditions for derivatization of thyreostats in urine with 3-IBBr were as follows: 100 μL of 3-IBBr (2 mg mL−1 or 5 mg mL−1), temperature 40°C, time 1 h, pH 8 (5, 8, 10, 15–17). Based on the experimental designed, which was carried out for thyreostat derivatization in urine, however, we found that pH of 10 can considerably increase the efficiency of derivatization. An overnight derivatization (16 h) at pH 10.0 with 300 μL reagent was compared to the optimum time (1 h). There was no considerable difference in the responses under these two conditions. This suggests that the reaction is almost finished at pH 10 after 1 h. This reaction time (1 h) is the same that was found by other researchers. HPLC columns containing core–shell packing material are known for their high resolution, good batch-to-batch reproducibility and robustness (18, 20). In this work, a HPLC column, especially developed for basic molecules, was also tested for derivatized thyreostats. Kinetex XB C-18 has been used successfully in our previous studies to analyze antibiotics, steroids and mycotoxins (18–20). This column has shown good selectivity and sensitivity in complex matrices such as body fluids, foods of animal and plant origin by separating the co-eluting matrix solutes, and thus reducing the matrix effect of LC–MS/MS analysis. Consequently, during the method development in the present study, we applied the Kinetex XB HPLC column packed with C-18 core–shell particles for thyreostats for the first time. The separation on core–shell column together with gradient elution enabled narrow peak shapes that further improved the analytical limits. Furthermore, the HPLC conditions allowed the separation of derivatized thyreostats on baseline (Figure 2) and hence the crosstalk effect could be eliminated. The crosstalk effect may appear when the co-eluting target compounds possess same fragment ions and in this case the fragment ion of one compound may be detected on the mass channel of another target compound. Novum SLE cartridges filled with polymeric packing material were introduced recently (21). The synthetic sorbent has superior batch-to-batch reproducibility. SLE can also be carried out with diatomaceous earth-based adsorbents, but it was not tested in this work. This clean-up is suitable only for screening purpose, because background matrices cannot be fully eliminated with SLE, so it does not provide sufficient purification. Therefore, a more robust clean-up is needed for confirmation. The SLE can only be used for aqueous samples; the cartridge completely adsorbs the sample up to a certain amount. The target compounds can be extracted from the adsorbed sample using immiscible organic solvents such as TBME, ETAC or chlorinated hydrocarbons. We found that both TBME and ETAC are suitable solvents for extraction and also the TBME–ETAC mixture. Even though this clean-up approach eliminates only the polar matrices as proteins, salts and other hydrophilic compounds from the sample, it provides a simple, fast, precise and cost-effective clean-up for screening purpose. For confirmatory analysis of thyreostats, a more effective clean-up approach was used to pre-concentrate the sample and to eliminate the disturbing matrices. SPE is generally applied for sample purification (8, 10). If the separation theory of SPE is orthogonal to the subsequent HPLC separation, the best selectivity and sensitivity can be achieved during the LC–MS/MS analysis. In this approach, those co-eluting matrices are removed from the samples that have different characters to the target compounds (e.g., acidic and neutral matrices). During the LC–MS/MS analysis, we used the reversed-phase separation on C-18 column, therefore the purification was orthogonal. The combination of cation-exchange SPE with reversed-phase LC–MS/MS separation on column packed with core–shell particles is described for thyreostats for the first time. In the case of MS/MS detection, it is important to obtain relative clean samples before detection to eliminate the matrix effect, thus the analytical limits can be lowered. Comparison of the present LC–MS method to existing ones Method developments for thyreostats in urine have been done by research groups to investigate the natural occurrence of TU and to set up methods for the analysis of banned thyreostats by LC–MS (Table IV). There are two ways to treat the urine samples containing thyreostats before analysis: one is with derivatization and the other is without. The methods with derivatization can stabilize the tautomer form of compounds analyzed and enhance the retention in LC separation as well as improve the fragmentation during MS/MS detection. Most of the papers suggest derivatization with 3-IBBr at pH 8 (5, 6, 8, 10, 15, 17, 25). In the present study, we found that the higher pH can increase the responses of derivatized thyreostats and pH 10 is the optimum obtained from the experimental design. The disadvantage of derivatization is the higher noise and matrix effect; and also the less selectivity caused by those isobaric matrix solutes that also react with the derivatization agent. Sample purifications therefore have been introduced in the methods to eliminate the number and concentration of matrices. LLE was enough only in one study that applied derivatization (25), mainly LLE in combination with normal-phase SPE was needed for clean-up. The consecutive extraction (LLE plus SPE) can lead to losses during the purification that increase the analytical limits. Moreover, this approach needs relatively high amount of organic solvents and is also time consuming. Methods without derivatization applied only LLE after a reduction step with DL-dithiothreitol at pH 7.0 (7, 26). The presented method utilizes a novel SPE purification with mixed-mode polymeric strong cation-exchange cartridge (MCX). This clean-up eliminates the matrices having acidic and neutral characters; therefore, an LLE step was not necessary to minimize the matrix effect of analysis; the method is thus less time consuming and is cost effective. The pre-concentrated basic solutes, however, caused ~50% ion suppression for TU. The MCX purification resulted in high extraction efficiency and clean extracts that enabled to lower the CCα below 1 μg L−1 for the first time (Table IV). The lowest TU concentration that could be confirmed with acceptable ion ratio was 0.148 μg L−1 (Figure 3). A novel screening method using SLE is also described. This clean-up is fast and inexpensive, but eliminates only the polar matrices, consequently it is recommended only for screening purpose. Table IV. Existing and the Presented LC–MS Methods for Thyreostats in Urine Analyzed compounds  Derivatization reagent  Purification  Technique  Analytical limit (μg L−1)  Reference  TU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with negative ionization  LOD = 0.8  5  TU  3-IBBr (pH 8)  LLE and normal phase SPE  LC–MS2 (iontrap) with negative ionization  LOQ = 1  6  TU    LLE  LC–MS/MS with positive ionization  CCα = 2.2  7  TU  3-IBBr (pH 8), or 2,3,4,5,6-pentafluorobenzyl bromide, or 4-chloro-7-nitrobenzofurazan  LLE and SPE normal phase SPE  LC–MS/MS with positive ionization  The highest CCα = 3.9  8  Seven thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with positive/negative ionization  CCα = 0.6–4.1  15  Eight thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with negative ionization  CCα = 0.1–5.1  17  Five thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE  LC–MS2 (iontrap) with positive ionization and LC–MS/MS with positive ionization  CCα = 0.84–1.24  25  Eight thyreostats including TAP, TU, MTU, PTU    LLE  LC–MS/MS with positive ionization  CCα = 1.1–5.5  26  TAP, TU, MTU, PTU  3-IBBr (pH 10)  Cation-exchange SPE for confirmation SLE for screening  LC–MS/MS with positive ionization  CCα = 0.05–0.8, CCβ = 5  This study  Analyzed compounds  Derivatization reagent  Purification  Technique  Analytical limit (μg L−1)  Reference  TU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with negative ionization  LOD = 0.8  5  TU  3-IBBr (pH 8)  LLE and normal phase SPE  LC–MS2 (iontrap) with negative ionization  LOQ = 1  6  TU    LLE  LC–MS/MS with positive ionization  CCα = 2.2  7  TU  3-IBBr (pH 8), or 2,3,4,5,6-pentafluorobenzyl bromide, or 4-chloro-7-nitrobenzofurazan  LLE and SPE normal phase SPE  LC–MS/MS with positive ionization  The highest CCα = 3.9  8  Seven thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with positive/negative ionization  CCα = 0.6–4.1  15  Eight thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with negative ionization  CCα = 0.1–5.1  17  Five thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE  LC–MS2 (iontrap) with positive ionization and LC–MS/MS with positive ionization  CCα = 0.84–1.24  25  Eight thyreostats including TAP, TU, MTU, PTU    LLE  LC–MS/MS with positive ionization  CCα = 1.1–5.5  26  TAP, TU, MTU, PTU  3-IBBr (pH 10)  Cation-exchange SPE for confirmation SLE for screening  LC–MS/MS with positive ionization  CCα = 0.05–0.8, CCβ = 5  This study  Table IV. Existing and the Presented LC–MS Methods for Thyreostats in Urine Analyzed compounds  Derivatization reagent  Purification  Technique  Analytical limit (μg L−1)  Reference  TU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with negative ionization  LOD = 0.8  5  TU  3-IBBr (pH 8)  LLE and normal phase SPE  LC–MS2 (iontrap) with negative ionization  LOQ = 1  6  TU    LLE  LC–MS/MS with positive ionization  CCα = 2.2  7  TU  3-IBBr (pH 8), or 2,3,4,5,6-pentafluorobenzyl bromide, or 4-chloro-7-nitrobenzofurazan  LLE and SPE normal phase SPE  LC–MS/MS with positive ionization  The highest CCα = 3.9  8  Seven thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with positive/negative ionization  CCα = 0.6–4.1  15  Eight thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with negative ionization  CCα = 0.1–5.1  17  Five thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE  LC–MS2 (iontrap) with positive ionization and LC–MS/MS with positive ionization  CCα = 0.84–1.24  25  Eight thyreostats including TAP, TU, MTU, PTU    LLE  LC–MS/MS with positive ionization  CCα = 1.1–5.5  26  TAP, TU, MTU, PTU  3-IBBr (pH 10)  Cation-exchange SPE for confirmation SLE for screening  LC–MS/MS with positive ionization  CCα = 0.05–0.8, CCβ = 5  This study  Analyzed compounds  Derivatization reagent  Purification  Technique  Analytical limit (μg L−1)  Reference  TU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with negative ionization  LOD = 0.8  5  TU  3-IBBr (pH 8)  LLE and normal phase SPE  LC–MS2 (iontrap) with negative ionization  LOQ = 1  6  TU    LLE  LC–MS/MS with positive ionization  CCα = 2.2  7  TU  3-IBBr (pH 8), or 2,3,4,5,6-pentafluorobenzyl bromide, or 4-chloro-7-nitrobenzofurazan  LLE and SPE normal phase SPE  LC–MS/MS with positive ionization  The highest CCα = 3.9  8  Seven thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with positive/negative ionization  CCα = 0.6–4.1  15  Eight thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE and SPE normal phase SPE  LC–MS/MS with negative ionization  CCα = 0.1–5.1  17  Five thyreostats including TAP, TU, MTU, PTU  3-IBBr (pH 8)  LLE  LC–MS2 (iontrap) with positive ionization and LC–MS/MS with positive ionization  CCα = 0.84–1.24  25  Eight thyreostats including TAP, TU, MTU, PTU    LLE  LC–MS/MS with positive ionization  CCα = 1.1–5.5  26  TAP, TU, MTU, PTU  3-IBBr (pH 10)  Cation-exchange SPE for confirmation SLE for screening  LC–MS/MS with positive ionization  CCα = 0.05–0.8, CCβ = 5  This study  Conclusions A new LC–MS/MS method for both screening and confirmation of thyreostats in urine is developed. The method uses an optimized chemical derivatization with 3-IBBr prior to clean-up. The sample purification is performed on SLE and MCX SPE cartridges for screening and confirmatory approaches, respectively. The derivatized thyreostats were separated on HPLC column packed with core–shell particles. The confirmatory method allowed reducing the analytical limits (CCα below 1 μg L−1) and naturally contamination of urine with TU could be detected. The methods are validated according to the EU recommendations and the confirmatory method is currently used in the national monitoring. Supplementary Data Supplementary material is available at Journal of Chromatographic Science online. Acknowledgments Authors wish to thank Éva Pálffy, Tímea Horváth and Viktória Lipcsei for their help in the sample preparations. Conflict of interest statement The authors have no conflict of interest on the submission of the manuscript. References 1 Gentili, A., De Pratti, V., Caretti, F., Pérez-Fernández, V., Tomai, P., Curini, R., et al.  .; Residue analysis of thyreostats in baby foods via matrix solid phase dispersion and liquid chromatography–dual-polarity electrospray–tandem mass spectrometry; Food Additive and Contaminants , ( 2016); 33: 1793– 1802. Google Scholar CrossRef Search ADS   2 Bertuzzi, T., Rastelli, S., Mulazzi, A., Pietri, A.; Trace monitoring of goitrin and thiouracil in milk and cheese; Food Analytical Methods , ( 2016); 9: 2952– 2959. Google Scholar CrossRef Search ADS   3 European Union. Council Directive 81/602/EC; Official Journal of the European Union, ( 1981); L/222. 4 CRL Guidance Paper.—CRLs view on state of the art analytical methods for the national residue control plans; ( 2007); http://www.rivm.nl/bibliotheek/digitaaldepot/crlguidance2007.pdf 5 Pinel, G., Mathieu, S., Cesbron, N., Maume, D., De Brabander, H.F., Andre, F., et al.  .; Evidence that urinary excretion of thiouracil in adult bovine submitted to a cruciferous diet can give erroneous indications of the possible illegal use of thyrostats in meat production; Food Additive and Contaminants , ( 2006); 23: 974– 980. Google Scholar CrossRef Search ADS   6 Vanden Bussche, J., Kiebooms, J.A.L., De Clercq, N., Deceuninck, Y., Le Bizec, B., De Brabander, H.F., et al.  .; Feed or food responsible for the presence of low-level thiouracil in urine of livestock and humans?; Journal of Agricultural and Food Chemistry , ( 2011); 59: 5786– 5792. Google Scholar CrossRef Search ADS PubMed  7 Vanden Bussche, J., Vanhaecke, L., Deceuninck, Y., Wille, K., Bekaert, K., Le Bizec, B., et al.  .; Ultra-high performance liquid chromatography coupled to triple quadrupole mass spectrometry detection of naturally occurring thiouracil in urine of untreated livestock, domesticated animals and humans; Food Additive and Contaminants , ( 2011); 28: 166– 172. Google Scholar CrossRef Search ADS   8 Wauters, J., Vanden Bussche, J., Le Bizec, B., Kiebooms, J.A.L., Dervilly-Pinel, G., Prevost, S., et al.  .; Toward a new European threshold to discriminate illegally administered from naturally occurring thiouracil in livestock; Journal of Agricultural and Food Chemistry , ( 2015); 63: 1339– 1346. Google Scholar CrossRef Search ADS PubMed  9 Kiebooms, J.A., Vanden Bussche, J., Hemeryck, L.Y., Fievez, V., Vanhaecke, L.; Intestinal microbiota contribute to the endogenous formation of thiouracil in livestock; Journal of Agricultural and Food Chemistry , ( 2012); 60: 7769– 7776. Google Scholar CrossRef Search ADS PubMed  10 Vanden Bussche, J., Noppe, H., Verheyden, K., Wille, K., Pinel, G., Le Bizec, B., et al.  .; Analysis of thyreostats: A history of 35 years; Analytica Chimica Acta , ( 2009); 637: 2– 12. Google Scholar CrossRef Search ADS PubMed  11 Lega, F., Contiero, L., Biancotto, G., Angeletti, R.; Determination of thyreostats in muscle and thyroid tissues by QuEChERS extraction and ultra-performance liquid chromatography tandem mass spectrometry; Food Additive and Contaminants , ( 2013); 30: 949– 957. Google Scholar CrossRef Search ADS   12 Chiesa, L.M., Labella, G.F., Pasquale, E., Panseri, S., Pavlovic, R., Arioli, F.; Determination of thyreostats in bovine urine and thyroid glands by HPLC–MS/MS; Chromatographia , ( 2016); 79: 591– 599. Google Scholar CrossRef Search ADS   13 Schmidt, K.S.; In-house validation and factorial effect analysis of a liquid chromatography-tandem mass spectrometry method for the determination of thyreostats in bovine blood plasma; Analytical and Bioanalytical Chemistry , ( 2014); 406: 735– 743. Google Scholar CrossRef Search ADS PubMed  14 Kiebooms, J., Wauters, J., Vanden Bussche, J., Vanhaecke, L.; Validation of a quantitative method using liquid chromatography coupled to multiple mass spectrometry for thiouracil in feedstuffs used in animal husbandry; Analytical and Bioanalytical Chemistry , ( 2015); 407: 4376– 4384. Google Scholar CrossRef Search ADS   15 Lõhmus, M., Kallaste, K., Le Bizec, B.; Determination of thyreostats in urine and thyroid gland by ultra high performance liquid chromatography tandem mass spectrometry; Journal of Chromatography. A , ( 2009); 1216: 8080– 8089. Google Scholar CrossRef Search ADS PubMed  16 Vanden Bussche, J., Sterk, S.S., De Brabander, H.F., Blokland, M.H., Deceuninck, Y., Le Bizec, B., et al.  .; Thyreostatic drugs, stability in bovine and porcine urine; Analytical and Bioanalytical Chemistry , ( 2012); 403( 10): 2973– 2982. Google Scholar CrossRef Search ADS PubMed  17 Pinel, G., Bichon, E., Pouponneau, K., Maume, D., André, F., Le Bizec, B.; Multi-residue method for the determination of thyreostats in urine samples using liquid chromatography coupled to tandem mass spectrometry after derivatisation with 3-iodobenzylbromide; Journal of Chromatography A , ( 2005); 1085: 247– 252. Google Scholar CrossRef Search ADS PubMed  18 Tölgyesi, Á., Sharma, V.K., Fekete, J.; Confirmatory analysis of stanozolol metabolites in bovine, pig and sheep urines using an optimized clean-up and liquid chromatography-tandem mass spectrometry; Journal of Pharmaceutical and Biomedical Analysis , ( 2014); 88: 45– 52. Google Scholar CrossRef Search ADS PubMed  19 Tölgyesi, Á., Kunsági, Z.; Quantification of T-2 and HT-2 mycotoxins in cereals by liquid chromatography-multimode ionization-tandem mass spectrometry; Microchemical Journal , ( 2013); 106: 300– 306. Google Scholar CrossRef Search ADS   20 Tölgyesi, Á., Sharma, V.K., Fekete, S., Fekete, J., Simon, A., Farkas, S.; Development of a rapid method for the determination and confirmation of nitroimidazoles in six matrices by fast liquid chromatography-tandem mass spectrometry; Journal of Pharmaceutical and Biomedical Analysis , ( 2012); 64–65: 40– 48. Google Scholar CrossRef Search ADS PubMed  21 Novum simplified liquid extraction User Guide; ( 2015); https://az621941.vo.msecnd.net/documents/7a8aa89b-c764-4f0f-ab75-f85459f260e3.pdf 22 Matuszewski, B.K., Constanzer, M.L., Chavez-Eng, C.M.; Strategies for the assessment of matrix effect in quantitative bioanalytical methods based on HPLC−MS/MS; Analytical Chemistry , ( 2003); 75: 3019– 3030. Google Scholar CrossRef Search ADS PubMed  23 CRLs. ( 2010): Guidelines for the validation of screening methods for residues of veterinary medicines (Initial validation and transfer). Community Reference Laboratories 20/1/2010. http://ec.europa.eu/food/food/chemicalsafety/residues/Guideline_Validation_Screening_en.pdf 24 Commission Decision 2002/657/EC; Official Journal of the European Union, ( 2002); L221/8. 25 Wozniak, B., Zuchowska, I.M., Zmudzki, J., Jedziniak, P., Korycinska, B., Sielska, K., et al.  .; Liquid chromatography tandem mass spectrometry with ion trap and triple quadrupole analyzers for determination of thyreostatic drugs in urine and muscle tissue; Analytica Chimica Acta , ( 2011); 700: 155– 166. Google Scholar CrossRef Search ADS PubMed  26 Vanden Bussche, J., Vanhaecke, L., Deceuninck, Y., Verheyden, K., Wille, K., Bekaert, K., et al.  .; Development and validation of an ultra-high performance liquid chromatography tandem mass spectrometry method for quantifying thyreostats in urine without derivatisation; Journal of Chromatography A , ( 2010); 1217: 4285– 4293. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Journal of Chromatographic ScienceOxford University Press

Published: Jun 6, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off