G Model

ARTICLE IN PRESS

CHROMA-355834; No. of Pages 12

Journal of Chromatography A, xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Analysis of antithyroid drugs in surface water by using liquid chromatography–tandem mass spectrometry Virginia Pérez-Fernández b , Stefano Marchese a , Alessandra Gentili a,∗ , María Ángeles García b , Roberta Curini a , Fulvia Caretti a , Daniela Perret a a Dipartimento di Chimica, Facoltà di Scienze Matematiche Fisiche e Naturali, Università degli Studi di Roma “La Sapienza”, P.le Aldo Moro 5, 00185 Roma, Italy b Department of Analytical Chemistry, Physical Chemistry and Chemical Engineering, Faculty of Chemistry, University of Alcalá, Ctra. Madrid-Barcelona, Km. 33.600, 28871 Alcalá de Henares (Madrid), Spain

a r t i c l e

i n f o

Article history: Received 24 July 2014 Received in revised form 13 September 2014 Accepted 18 September 2014 Available online xxx Keywords: Liquid chromatography–tandem mass spectrometry Antithyroid drugs Thyreostatics Surface water analysis River water analysis Multi-residue analytical method

a b s t r a c t This paper describes development and validation of a new method for the simultaneous determination of six antithyroid drugs (ATDs) in surface waters by using liquid chromatography–triple quadrupole mass spectrometry (LC–MS/MS). Target compounds include two ATD classes: thiouracil derivatives (thiouracil (TU), methyl-thiouracil (MTU), propyl-thiouracil (PTU), phenyl-thiouracil (PhTU)) and imidazole derivatives (tapazole (TAP), and mercaptobenzimidazole (MBI)). Sensitivity and selectivity of the LC-multiple reaction monitoring (MRM) analysis allowed applying a simple pre-concentration procedure and “shooting” the concentrated sample into the LC–MS/MS system without any other treatment. Recoveries were higher than 75% for all analytes. Intra-day precision and inter-day precision, calculated as relative standard deviation (RSD), were below 19 and 22%, respectively. Limits of detection (LODs) ranged from 0.05 to 0.25 ␮g/L; limits of quantitation (LOQs) varied between 0.15 and 0.75 ␮g/L. The validated method was successfully applied to the analysis of ATD residues in surface water samples collected from the Tiber River basin and three lakes of Lazio (central Italy). The analytes were quantified based on matrixmatched calibration curves with mercaptobenzimidazole-d4 (MBI-d4 ) as the internal standard (IS). The most widespread compound was TAP, one of the most common ATDs used in human medicine, but also TU and MBI were often detected in the analysed samples. © 2014 Elsevier B.V. All rights reserved.

1. Introduction The ATDs are a group of relatively simple molecules chemically known as thionamides, which contain a sulfhydryl group and a thiourea moiety within a heterocyclic structure (Table 1). They are also named thyreostatic agents, thyreostatics, thyreostats or thyrotropic agents. Due to their chemical structures, these compounds interfere with thyroid peroxidase-mediated iodination of tyrosine residues in thyroglobulin, producing a decrease in the production of the thyroid hormones, triiodothyronine (T3) and thyroxine (T4) [1,2]. The ATDs are commonly used in human medicine for the management of hyperthyroidism, especially for patients with Graves’ disease. TAP, also well-known as methimazole or thiamazole, is

∗ Corresponding author. Tel.: +39 06 49693230; fax: +39 06 49693230. E-mail addresses: [email protected], [email protected] (A. Gentili).

used in the United States (US) and in most of Europe and Asia; it is also the biologically active metabolite of carbimazole (CMI), another ATD licensed mainly in the United Kingdom (UK). A recent study, providing an overview of the ATD use in the US from 1991 to 2008, has revealed that PTU held two-thirds of the market up to 1995 and that it was surpassed by TAP in 1996 [3]. The annual number of TAP prescriptions during these two decades increased from 158,000 to 1.36 million per year, whereas the increase for the PTU prescriptions was slower passing from 348,000 to 415,000 per year. In the UK, between 1981 and 2003, there were 5.23 million prescriptions for ATDs, 94% of which were for CMI [4]. Data obtained from the website of Italian Medicines Agency [5] indicates that the number of the defined daily doses (DDDs) of TAP distributed in Italy in 2000 was about 202 million and increased up to 445 million in 2011. Because it is unlikely that the incidence of thyrotoxicosis cases is increasing dramatically, it is plausible that such intensification reflects a growing preference for long-term pharmacotherapy rather than radioiodine treatment or thyroidectomy.

http://dx.doi.org/10.1016/j.chroma.2014.09.045 0021-9673/© 2014 Elsevier B.V. All rights reserved.

Please cite this article in press as: V. Pérez-Fernández, et al., Analysis of antithyroid drugs in surface water by using liquid chromatography–tandem mass spectrometry, J. Chromatogr. A (2014), http://dx.doi.org/10.1016/j.chroma.2014.09.045

G Model CHROMA-355834; No. of Pages 12

ARTICLE IN PRESS V. Pérez-Fernández et al. / J. Chromatogr. A xxx (2014) xxx–xxx

2 Table 1 Target analytes and their physicochemical properties.a Name, abbreviation and exact massb

Tautomers with corresponding distribution, pKa and log Pow c

Thiouracil (TU) 128.0044 u

Prevalent tautomer (68%) Log Pow = 0.40

Tautomer 2 (27%) Log Pow = 0.03

Tautomer 3 (4%) Log Pow = 0.38

Tautomer 2 (26%) Log Pow = 0.24

Tautomer 3 (5%) Log Pow = 0.64

6-Methyl-2-thiouracil (MTU) 142.0201 u

Prevalent tautomer (67%) Log Pow = 0.60

6-Propyl-2-thiouracil (PTU) 170.0514 u

Prevalent tautomer (70%) Log Pow = 1.57

Tautomer 2 (23%) Log Pow = 1.20

Tautomer 3 (6%) Log Pow = 1.78

6-Phenyl-2-thiouracil (PhTU) 204.0357 u

Prevalent tautomer (47%) Log Pow = 1.70

Tautomer 2 (41%) Log Pow = 1.33

Tautomer 3 (8%) Log Pow = 2.29



Tapazole (TAP) 114.0252 u

Prevalent tautomer (58%) Log Pow = 0.75

Tautomer 2 (42%)

Please cite this article in press as: V. Pérez-Fernández, et al., Analysis of antithyroid drugs in surface water by using liquid chromatography–tandem mass spectrometry, J. Chromatogr. A (2014), http://dx.doi.org/10.1016/j.chroma.2014.09.045

G Model CHROMA-355834; No. of Pages 12

ARTICLE IN PRESS V. Pérez-Fernández et al. / J. Chromatogr. A xxx (2014) xxx–xxx

3

Table 1 (Continued) Name, abbreviation and exact massb

Tautomers with corresponding distribution, pKa and log Pow c

2-Mercaptobenzimidazole (MBI) 150.0252 u





Prevalent tautomer (100%) Log Pow = 2.08 a b c

The shown tautomer distribution occurs at pH values lower than 7 for all ATDs with the only exception of TAP (pH values lower than 9). Mass exact values were determined by using Chem Calc Isotopic Distribution Calculator. Tautomer distribution, pKa and log Pow values were predicted with Marvin Calculator.

The ATDs are also used in veterinary medicine to treat companion and farm hyperthyroid animals, but, since the middle 1970s, there have been strong suspicions about their illicit administration to euthyroid cattle to obtain a quick fattening [2]. Between 1974 and 1979 in Belgium, the use of MTU was discouraged through selective controls [6]. Later, in the European Union (EU) countries, grounds for concern were the detection of MBI in thyroid of slaughtered animals [7] and an increased incidence of aplasia cutis in Spain, perhaps caused by the consumption of meat contaminated with ADTs [8,9]. Compared to other growth promoters, the ATD abuse may be favoured by the low cost and the simplicity of administration (addition to feed) [2]. In cases of illegal treatment, the urine of the cattle shows high concentrations of ATDs (100 ␮g/L) since 1–5 g/day doses have to be administered to observe the expected animal weight gain [10]. Unlike other anabolic agents such as the natural hormones [11], there is a world-wide agreement on the ban of these drugs for two major reasons: (i) their residues may be a serious risk to consumers [1]; (ii) the ATDs cause water retention in edible tissues and, consequently, meat results of inferior quality [2,12]. The EU has forbidden their administration to animals since 1981 [13,14] and has issued strict regulations on the use and the control of veterinary drugs in all member states. The measures to monitor certain substances and their residues in live animals and animal products are described in the Council Directive 96/23/EC [15]. Annex I of this directive also lists two groups of toxicants: Group A (banned substances) that comprises hormones, ATDs and substances included in Annex IV of the Commission Regulation 2377/90/EEC [16]; Group B (authorised substances) including other veterinary drugs and contaminants. To ensure a harmonised implementation of the Directive 96/23/EC, the Commission Decision 2002/657/EC [17] lays down the technical guidelines and performance criteria for the residue control. The criteria contained in this document have currently been adopted by many researchers to perform confirmation analyses also in other study areas (environmental pollution, clinical chemistry food analysis and so on). In the latest years, TU has been supposed to have a natural origin since low levels (1–10 ␮g/L) have been detected in the urine of untreated animals [10]. The Community Reference Laboratories’ guidance paper of 2007 acknowledged this by stating that TU concentrations below 10 ␮g/L might derive from the consumption of glucosinolate-rich Brassicaceae plants [18]. A more recent study has demonstrated that a Brassicaceae-rich diet enforces the presence of this analyte, but in a non-significant way, providing that it is not the only natural source of TU detected in human and animal urine [19]. Such kinds of studies are still going on. The ATDs are associated with a variety of minor side effects as well as potentially life-threatening complications [1]. Minor side

effects include cutaneous reactions, arthralgia, and gastrointestinal upset, whereas agranulocytosis, blood dyscrasias, hepatotoxicity and vasculitis are less common but very severe. The latter reactions are more frequent and dose-unrelated using PTU rather than TAP, but The Food and Drug Administration (FDA) has categorised both compounds as class D agents because of the potential for foetal hypothyroidism [20]. For this reason the presence of these drugs in several kinds of matrices is matter of concern. However, the highly polar and amphoteric character of the ATDs and their ability to adopt different tautomeric forms (see Table 1) affect negatively both their extraction and their chromatographic separation; moreover, the MS detection is generally poorly sensitive due to the reduced size of the molecules. Till the moment, all reports published previously delay with the ATD determination in food and biological samples (tissues, milk, urine and feed). Early methods, developed by De Brabander and Verbeke, were based on thin-layer chromatography (TLC) and high performance TLC with fluorimetric detection [6,21]. Later, LC has been the most used separation technique, followed by gas chromatography (GC) and capillary electrophoresis (CE). The few CE methods have been based on UV–vis [22], laser induced fluorescence [23] and electrochemical detection [24]. On the other hand, GC has been used by coupling it to MS [25–29] and nitrogen–phosphorus detection [30]. Most LC methods have relied on UV–vis [31,32] and MS detection [33–37,10,38–40]; some of the latters have included a preliminary derivatization step of the analytes in order to produce compounds with higher molecular weight and lower polarity, which are better retained on reversed phase (RP) columns and which generate a lower MS background resulting in higher S/N ratios [10,38–40]. Pharmacokinetic data suggest that the ATDs, similarly to other pharmaceutical compounds, are excreted both unmetabolized and conjugated with glucuronic acid in urine [41] or faeces [42]. Consequently, these drugs are discharged into domestic wastewater both in their free and conjugated forms along with unwanted or expired ATDs, which are disposed improperly. Besides domestic wastes, other sources could be wastes from hospitals, farms and manufacturing facilities. Therefore, it can be expected that the ATDs reach sewage treatment plants (STPs) in substantial amounts and that they can be released into surface water if they escape degradation or elimination through sludge (for example, ATD glucuronates can release their active moiety during treatment in STPs). In this sense, it has to be remarked that, surveying the literature, there is still no work delaying with the determination of these contaminants in natural waters in spite of their expected occurrence in the environment.

Please cite this article in press as: V. Pérez-Fernández, et al., Analysis of antithyroid drugs in surface water by using liquid chromatography–tandem mass spectrometry, J. Chromatogr. A (2014), http://dx.doi.org/10.1016/j.chroma.2014.09.045

G Model CHROMA-355834; No. of Pages 12

ARTICLE IN PRESS V. Pérez-Fernández et al. / J. Chromatogr. A xxx (2014) xxx–xxx

4

On the basis of the foregoing considerations, it is evident that there is a critical need to develop new, simple and fast methods for the determination of these drugs in natural waters. This paper just describes development, optimisation and validation of a very simple method for the preconcentration and the subsequent analysis of six ATDs (TU, MTU, PTU, PhTU, TAP, MBI) in surface waters by using LC–MS/MS. Advantages and drawbacks are presented and discussed in detail. The method was applied to assess the distribution of the selected ATDs in surface waters (Lazio, Italy) affected by different anthropogenic activities (for example, river water receiving the effluents of Roman STPs (the Tiber River-Rome Centre) or runoff from animal agricultural operations (the Tiber River-Farfa Oasis)). 2. Materials and methods 2.1. Reagents and samples All drug standards and the IS were of high purity grade (>90%). TU, TAP, MTU, PTU, PhTU, MBI and MBI-d4 were purchased at Sigma Aldrich Chemical (St. Louis, MO, USA). Individual stock solutions of the standards and the IS were prepared by dissolving the correct amount of each analyte in methanol at 0.5 mg/mL for TU and MTU and at 1 mg/mL for the other ATDs and the IS. The regularity of preparation was established on the basis of the results achieved by a stability study carried out over a period of 70 days: the individual stock solutions were prepared every two months. Working standard solutions were prepared from the individual solutions after appropriate dilution with methanol. All the solutions were stored at −20 ◦ C. For extraction studies, the solid phase extraction (SPE) cartridges were tested: Supelclean ENVI-Carb (6 cc, 500 mg) and Discovery DPA-6S (1 cc, 50 mg) from Supelco (Bellefonte, PA, USA); Oasis-HLB (6 cc, 500 mg) from Waters Corporation (Milford, MA, USA); SPE-ED-SCX (3 cc, 500 mg) from Applied Separation (Allentown, PA, USA); Sampli-Q Evidex (3 cc, 200 mg) from Agilent (Palo Alto, CA, USA). Acetonitrile, methanol, ethyl acetate, dichloromethane (all of them of RS Plus grade) and hydrochloric acid 37% (HCl) were supplied by Carlo Erba (Milano, Italy). Ammonia, formic acid puriss. p.a., trifluoroacetic acid (TFA) and ethylenediaminetetraacetic acid disodium salt dehydrate (EDTA) were acquired from Sigma Aldrich Chemical. Tetrabutylammonium chloride (TBACl) was from Fluka (Buchs, Switzerland). For HPLC, distilled water was further purified by passing it through the Milli-Q Plus apparatus from Millipore (Bedford, MA, USA). 2.2. Collection of natural water samples Natural water samples were collected from Vico, Bracciano and Martignano lakes and from different points of the Tiber River basin (natural area of Farfa, Rome centre, Rome suburb, Fiumicino). Samplings were carried out in two periods: from March to May 2012 and from May to June 2014. The samples were collected in glass bottles and forthwith spiked with EDTA to obtain the concentration of 0.4 g/L (1.4 mM). Once in the laboratory, the samples were filtered through 1.2 ␮m Whatman glass microfiber filters (Whatman International Ltd, Maidstone, UK) and stored at 4 ◦ C until the extraction, which was performed within 48 h in order to avoid any degradation (see Section 3.3 on the stability study of ATDs in surface water samples). 2.3. Preconcentration For the extraction of the ATDs from surface water two different approaches were developed based on evaporation under reduced pressure and SPE as described below.

2.3.1. Evaporation under reduced pressure A 250-mL volume of surface water was filled into 1 L roundbottomed flask and spiked with MBI-d4 (2 ␮g/L). After a 30 min period for equilibration at room temperature, the sample preconcentration was carried out under reduced pressure (40 mbar) at 70 ◦ C in a rotary evaporator (Rotavapor R II from Büchi, Milan, Italy); it was operated at a moderately low speed, which avoided excessive distribution of the sample, for approximately 1/2 h. The dry residue was reconstituted in 5 mL of methanol, centrifuged for 10 min at 6000 rpm and, subsequently, evaporated in a water bath at 50 ◦ C under a stream of nitrogen to a final volume of 500 ␮L. Prior to injection, the extract was centrifuged for 10 min at 6000 rpm and then filtered with 0.45 ␮m PTFE syringe filters to remove precipitated EDTA. Twenty microlitres of the final extract was injected into the LC–MS/MS system. 2.3.2. Solid-phase extraction As an alternative to the above-described method, we tried to develop SPE protocols based on methods that are currently in use for polar substances [43,44] in order to clean-up the aqueous samples. The extraction of the ADTs was evaluated by testing different commercial cartridges (Envicarb, Discovery DPA-6S, SCX, Oasis-HLB and Sampli-Q Evidex). Each SPE cartridge was conditioned according to the specifications of the commercial brand at a flow rate of 1 mL/min using a Baker vacuum system (J.T. Baker, The Netherlands). After the conditioning step, the water samples (250 mL of river and lake water spiked with the IS at 2 ␮g/L) were percolated through the cartridges at a flow rate of 10 mL/min. Finally, each cartridge was rinsed with 5 mL of MilliQ water to eliminate the possible interfering compounds and then was dried under vacuum for the time necessary to remove the water excess. The analytes were eluted with 12 mL of methanol from the SCX, Oasis-HLB and Sampli-Q Evidex cartridges, with 6 mL of acetone and 6 mL of ethyl acetate from Discovery DPA-6S cartridge and with 12 mL of dicloromethane/methanol (9/1, v/v) containing 5 × 10−3 mol/L TBACl from Envicarb cartridge. The obtained extract was evaporated in a water bath at 50 ◦ C under a stream of nitrogen till 500 ␮L and 20 ␮L was injected in the LC–MS/MS system. 2.4. Liquid chromatography–tandem mass spectrometry Liquid chromatography was performed by means of a micro HPLC/autosampler/vacuum degasser system PE Series 200 (Perkin Elmer, Norwalk, CT). The analytes were separated on a X-Terra MS C18-column (4.6 × 150 mm, 5 ␮m, Waters Corporation, USA) protected by a guard column of the same type (4.0 × 10 mm, 5 ␮m, Waters Corporation, USA). In separating the ATDs, phase A was water/acetonitrile (50/50, v/v) containing 6 × 10−3 mol/L formic acid and phase B was water. The mobile phase gradient profile was as follows: t0 , A = 0%; t5 , A = 100%; t10 , A = 100%. A mixture water/acetonitrile (50/50, v/v) was used as washing solution of the autosampler injection system. The flow rate of the LC eluent was 1 mL/min, but 80% of it was split by means of a T-junction; the advantages in diverting only the 20% of the column effluent to the Turbo V source of the mass spectrometer were consistent with the best S/N chromatographic ratios and the decreased instrumental contamination. The analytes were identified and quantified by a 4000 Qtrap® (AB SCIEX, Foster City, CA, USA). The detection was performed in dual polarity ionisation, placing the electrospray (ESI) probe in the Turbo V source and setting a capillary ion voltage of 5000 V for the positive ionisation mode and −4500 V for the negative ionisation mode; the temperature to heat the drying gas was set at 450 ◦ C. High-purity nitrogen was used as curtain gas (5 L/min) and collision

Please cite this article in press as: V. Pérez-Fernández, et al., Analysis of antithyroid drugs in surface water by using liquid chromatography–tandem mass spectrometry, J. Chromatogr. A (2014), http://dx.doi.org/10.1016/j.chroma.2014.09.045

G Model

ARTICLE IN PRESS

CHROMA-355834; No. of Pages 12

V. Pérez-Fernández et al. / J. Chromatogr. A xxx (2014) xxx–xxx

5

Table 2 LC–MS/MS parameters for the analysis of the target analytes by MRM negative and positive ionisation mode. Analyte

Retention time Meana (±SDn=10 ) (min)

MRM transitionsb (m/z)

Declustering potential (V)

Collision energy (eV)

Ion ratios Meanc (RSDn=10 ) (%)

Permitted tolerances for ion ratiosd (%)

TU

3.07 (±0.05)

41 (±18)

±25

MTU

3.57 (±0.05)

39 (±22)

±25

PTU

4.78 (±0.04)

58 (±16)

±20

MBI

5.12 (±0.04)

85 (±15)

±20

PhTU

5.55 (±0.05)

41 (±18)

±25

MBI-d4 (IS)

5.10 (±0.04)

−26 23 25 42 −30 23 −35 27 35 −31 −46 39 −31 38

±25

3.57 (±0.05)

−52 50 38 38 −45 42 −56 55 44 −82 −31 89 −71 44

22 (±20)

TAP

127.0/58.0 (−) 129.0/112.0 (+) 115.1/88.0 (+) 115.1/81.1 (+) 141.0/58.0 (−) 143.0/125.8 (+) 169.1/58.0 (−) 171.0/112.0 (+) 151.0/118.0 (+) 149.0/117.0 (−) 203.0/58.0 (−) 205.0/85.9 (+) 152.9/120.9 (−) 155.0/122.0 (+)





a

The retention times are reported as arithmetic average of ten replicates plus the corresponding standard deviation (SD). The first line reports the most intense MRM transition (quantifier) and the second line the least intense one (qualifier). c The ion ratio (relative abundance) between the two MRM transitions is calculated as percentage ratio of qualifier intensity/quantifier intensity; the results are reported as arithmetic average of ten replicates plus the corresponding relative standard deviation (RSD). d Tolerances recommended by the Commission Decision 2002/657/EC. b

gas (4 mTorr), while air as nebuliser gas (2 L/min) and drying gas (30 psi). A preliminary mass axis calibration of each quadrupole massanalyser, Q1 and Q3, was carried out by the infusion of a polypropylene glycol solution at 10 ␮L/min. Unit mass resolution was established and kept in each mass-resolving quadrupole by maintaining a full width at half maximum (FWHM) of approximately 0.7 ± 0.1 u. The instrumental parameters of the ESI source and the analysers were optimised for every compound by infusing the corresponding standard solutions with a syringe pump (concentration between 0.1 and 1 ng/␮L and flow rate of 10 ␮L/min). The fragmentation study was carried out by acquiring in product ion scan mode. The quantitative analysis of the target analytes was performed in MRM scan mode. For each ATD, the two most intense MRM transitions were selected after studying their corresponding fragmentation spectra. Table 2 summarises the quantitation channels of the analytes (IS included) and the parameters used for their identification in surface water samples (retention time, two MRM transitions for each analyte and their relative abundance). Fig. 1 shows the LCMRM profiles of a lake water sample spiked with the ATDs at 0.50 ␮g/L.

2.5. Validation of the analytical procedure The developed method was validated in accordance with the main guidelines of the Commission Decision 2002/657/EC [17] for a quantitative confirmation method. The method performance was assessed valuating qualitative parameters (identification and selectivity) and quantitative parameters (recovery, precision, linearity, sensitivity, limit of detection (LOD) and limit of quantitation (LOQ)). The criteria taken from the Commission Decision 2002/657/EC establish that at least four identification points (IPs) are needed to confirm the identity of substances belonging to Group A such as the ATDs. To this end, two MRM transitions (1IP for the precursor ion + (2 × 1.5) IPs for the two product ions) were selected for each target analyte. The most intense MRM transition (quantifier transition) was used to perform the quantitative analysis, while the second most intense MRM transition (qualifier) was used for identification purpose. Preliminary analyses showed that the water samples collected from Vico lake did not present any of the studied analytes above

the established LOD. Therefore, they were suitable for being used as blank matrices for the method development and validation. The software used for acquiring and elaborating LC–MS data was Analyst 1.5.1 (AB Sciex, Foster City, CA, USA). Linear regression, mean and standard deviation (SD) were calculated using Microsoft Excel 2010. Statistical data treatment was also performed employing Statgraphics Centurion XVI software. 2.5.1. Identification and selectivity The selectivity study was carried on reference standards, IS and real samples. Since the method is intended to determine more than one analyte, the IS and each authentic standard was injected into the LC–MS system at increasing concentrations to ensure the absence of interference and/or contamination. By using highly sensitive instruments, procedural blanks were periodically prepared to check potential contamination of glassware, syringes and solvents used during the analyses. Solvent blanks were prepared by simulating extraction of real samples, i.e. substituting 250 mL of Milli-Q water for the 250 mL of surface water. The presence of potentially interfering substances from matrix was investigated by analysing 20 blank samples and checking for any intrusive peaks in the retention time window where the target analyte is expected to elute. In analysing real samples, selectivity was assessed by applying the identification criteria drawn from the Commission Decision 2002/657/EC. The occurrence of each analyte in the real samples was confirmed by matching the retention time (tr ± 2.5%) and the ion ratio of the two MRM transitions (within the tolerances recommended by the Commission Decision 2002/657/EC; see Table 2) with the average values obtained for the reference standards in matrix (blank water samples from Vico Lake spiked with the analytes at 0.50 ␮g/L; see Fig. 1). Using a low resolution MS as chromatographic detector, these criteria allow one to recognise the coelution of an isobaric interference with a specific ATD through an altered ion ratio of the analyte. 2.5.2. Recovery and precision Recovery and precision were estimated by spiking blank water samples (six 250-mL aliquots) with MBI-d4 at 2.0 ␮g/L and with the selected ATDs at the following levels: 0.5, 2.0 and 5.0 ␮g/L. In order to calculate extraction efficiencies, a matrix blank was spiked post-extraction with the same nominal concentrations of the analytes and IS. All the samples were preconcentrated by evaporating

Please cite this article in press as: V. Pérez-Fernández, et al., Analysis of antithyroid drugs in surface water by using liquid chromatography–tandem mass spectrometry, J. Chromatogr. A (2014), http://dx.doi.org/10.1016/j.chroma.2014.09.045

G Model CHROMA-355834; No. of Pages 12 6

ARTICLE IN PRESS V. Pérez-Fernández et al. / J. Chromatogr. A xxx (2014) xxx–xxx

Fig. 1. LC-MRM-chromatogram relative to the analysis of a water sample collected from Vico Lake (250 mL) and spiked post-extraction with the ATDs at 0.50 ␮g/L.

under reduced pressure (see Section 2.3.1) and the obtained extract was twice injected into the LC–MS/MS system. Six replicates were analysed in one day to evaluate recovery and intra-day precision. The same set of samples was prepared and analysed on two additional days by different operators and under different environmental conditions to estimate the inter-day precision. The method precision was expressed as the relative standard deviation (RSD) of replicate measurements. 2.5.3. Linearity The analytes were quantified by using matrix matched calibration in conjunction with a surrogate IS (MBI-d4 ) [45]. For each calibration curve, seven 250-mL aliquots of water collected from Vico lake were spiked with increasing concentrations of the ATDs (0.5, 1, 2, 5, 7, 10 and 18 ␮g/L) and with the same concentration of MBI-d4 (2 ␮g/L); a zero sample (i.e. a blank sample spiked only with the IS at 2 ␮g/L). After fortification, a 30-min period was allowed for equilibration at room temperature and in the dark under continuous stirring. Extractions and analyses were performed according to

what described in Section 2.3.1. In this method, the surrogate plays the dual role of correcting for extraction efficiency and instrumental variability. The curves were constructed by linear regression, using the quantifier transition and plotting the relative peak area (area of analyte/area of IS) against the fortification level. Squared linear regression coefficients (R2 ) greater than 0.90 were considered acceptable. Linearity was also valued applying a “Lack-of-fit F-test”. This test states that, if the p-value obtained is larger than the significance level ˛ (0.05), the null hypothesis is accepted, i.e. there is not enough evidence at ˛ level to conclude that there is lack of linear fit. 2.5.4. Limit of detection (LOD) and limit of quantitation (LOQ) LOD and LOQ were estimated from spiked water samples as mean of six replicates using the qualifier MRM transition so that both ion currents could be detected with their characteristic ion ratio also at the LOD level. LOD and LOQ were intended as the minimum detectable and quantifiable amount of analyte with a S/N ratio of 3 and 9, respectively.

Please cite this article in press as: V. Pérez-Fernández, et al., Analysis of antithyroid drugs in surface water by using liquid chromatography–tandem mass spectrometry, J. Chromatogr. A (2014), http://dx.doi.org/10.1016/j.chroma.2014.09.045

G Model

ARTICLE IN PRESS

CHROMA-355834; No. of Pages 12

V. Pérez-Fernández et al. / J. Chromatogr. A xxx (2014) xxx–xxx

7

Table 3 Recovery (%) of the ATDs (2 ␮g/L) from distilled water samples by using different commercial SPE cartridges.

OASIS HLBa ENVICARB SCX DISCOVERY DPA-6S SAMPLI-Q EVIDEX a

TU

MTU

PTU

PhTU

TAP

MBI

0–3.16% 12.8% – – –

3.3–15.1% 12.2% – – –

21.5–45.3% 10.0% – – –

46.8–55.8% – – 9.1% –

2.2–5.1% – 5.2% – –

10.5–25.8% – – – –

Values dependent on the modifier added to the sample (HCl, formic acid, TBACl or ammonia) and on the processed volume of water (100 mL and 250 mL).

3. Results and discussion 3.1. Optimisation of the LC conditions and MS detection The chromatographic separation of the ATDs is complicated by more than one factor: (i) The low molecular weight and the high polarity are characteristics responsible for the poor retention of these analytes on typical RP columns. (ii) The amphoteric nature causes the peak splitting of the ATDs with lower molecular weights. (iii) The basic centres on the heterocyclic system produce peak tailing by interacting with free silanols on silica surface of the stationary phase. In this work, the LC conditions were optimised in order to improve analyte retention and peak form; the tailing was considered unacceptable when the peak asymmetry factor (As ) was >1.5 [46]. The definitive LC conditions for the ATD separation were established after a series of preliminary trials addressed to the choice of the chromatographic column (Alltima C18, 4.6 mm × 250 mm, 5 ␮m; XTerra MS C18, 4.6 mm × 250 mm, 5 ␮m; Gemini C18, 4.6 mm × 100 mm, 5 ␮m; Gemini C6 Phenyl, 4.6 mm × 150 mm, 3 ␮m), mobile phase composition (solvents and additives), gradient elution and flow rate (0.7, 1.0 and 1.2 mL/min). Regardless the kind of the RP column chosen, the use of a mobile phase with a high percentage of water was indispensable to assure a good retention of all the analytes and, in particular, to avoid the splitting peak phenomenon of TU, MTU and TAP (ATDs with the smaller molecular weight). By chromatographing the ATDs, the Alltima C18 and XTerra MS C18 columns have given the best performances in reducing tailing problems. The Alltima C18 column, based on high-purity monomeric silica, exhibited slight tendency to generate asymmetrical peaks because of the low surface acidity [47]. The high degree of deactivation of the XTerra MS C18 column was due to both the methyl groups incorporated into the matrix of the organic/inorganic hybrid silica and the low activity of the residual free silanols. The analytes were eluted from both columns by adopting a linear decrease of pH (from pH ∼7 to ∼3; see Section 2.4) to suppress the ionisation of the residual silanols [48] and that of the ATDs in relation to their pKa (see Table 1); at the same time,

this elution mode also alleviated the ESI signal suppression of the ATDs detected in negative ionisation mode. The Gemini C18 column, like XTerra MS C18, is based on organo-silane particles which can tolerate extreme pH values (i.e. 2 < pH < 12). In this case, the best LC conditions were accomplished by eluting with a water/acetonitrile (50/50, v/v) solution containing 1 × 10−3 mol/L formic acid as phase A and water containing 7 × 10−4 mol/L NH3 (pH ∼8) as phase B (from 0 to 100% of A in 10 min). Using this stationary phase, NH3 was a better tail-suppressing compound for the poorly retained ATDs. A disadvantage was a longer conditioning time compared to that of the Alltima C18 and XTerra MS C18 columns. Gemini C6 Phenyl is another hybrid column tested for the ATD separation. It is characterised by an organo-silica material endcapped with tetramethylsilane and functionalised with phenyl hexyl groups. The high steric selectivity, offered for the separation of aromatic, polar or basic compounds, is due to both the ␲–␲ interaction between its phenyl groups and the analyte’s planar structure and the hydrophobic action of the C6 spacer. The use of different organic solvents allows one to modulate the retention of a specific compound, activating (with methanol) or suppressing (with acetonitrile) the ␲ interactivity. In this work, the Gemini C6 Phenyl column was used at the optimal flow of 0.7 mL/min testing both acetonitrile and methanol as organic modifiers. Notwithstanding the good resolution, some performances (broader peaks, higher backpressures, and less reproducibility of the retention times) made it less attractive than the other RP columns. In conclusion and considering the preliminary results obtained with the various columns tested, the XTerra MS C18 column was selected as the most adequate one for the separation of the six selected drugs. Since the mass to charge ratios of the ATDs appear into the typical range of the MS background noise, their ESI detection is not particularly sensitive. Some researchers have derivatised the ATDs with 3-iodobenzylbromide in order to stabilise them under single tautomeric form, to give them a non-polar behaviour and to increase their masses [10,38,40]. Other researchers have preferred the direct LC–MS detection in positive ion mode to avoid false identification [37] and to make the analysis of these substances simpler and faster when a high number of food or biological samples has to be controlled [34–36]. In this work, taking advantage from the amphoteric nature of the ATDs, the performances of the

Table 4 Validation parameters for the developed method. Analyte

TU TAP MTU PTU MBI PhTU

Recovery and intra-day precision (%)

Inter-day precision (%)

0.5 ␮g/L

2.0 ␮g/L

0.5 ␮g/L

2.0 ␮g/L

5.0 ␮g/L

Slope

R

p-value (lack of fit test)

102 (19) 78 (13) 99 (16) 105 (11) 90 (10) 81 (14)

84 (15) 88 (13) 102 (17) 104 (13) 85 (13) 99 (9)

22 18 20 15 14 7

12 13 17 15 18 7

4 8 5 6 7 5

4.185 0.869 2.327 1.738 1.239 0.945

0.900 0.969 0.954 0.972 0.982 0.953

0.255 0.117 0.108 0.825 0.256 0.459

5.0 ␮g/L 107 (3) 100 (5) 105 (8) 106 (4) 75 (5) 95 (3)

Linear equation parameters 2

LOD (␮g/L)

LOQ (␮g/L)

0.22 0.25 0.13 0.08 0.05 0.17

0.66 0.75 0.39 0.24 0.15 0.51

Please cite this article in press as: V. Pérez-Fernández, et al., Analysis of antithyroid drugs in surface water by using liquid chromatography–tandem mass spectrometry, J. Chromatogr. A (2014), http://dx.doi.org/10.1016/j.chroma.2014.09.045

G Model CHROMA-355834; No. of Pages 12

ARTICLE IN PRESS V. Pérez-Fernández et al. / J. Chromatogr. A xxx (2014) xxx–xxx

8

Table 5 Average levels (n = 3) measured for the target analytes in 10 points from the Tiber River basin and some lakes of Lazio (central Italy). Year

2012

2014

a b

Sampling points

Lakes Vico Lake Vico Lake groundwater Martignano Lake Bracciano Lake (Bracciano) Bracciano Lake (Anguillara) Bracciano Lake (Trevignano) The Tiber River The Tiber River (Farfa Natural Area) The Tiber River (IsolaTiberina) The Tiber River (Magliana) The Tiber River (Fiumicino) Lakes Vico Lake Vico Lake groundwater Martignano Lake Bracciano Lake (Bracciano) Bracciano Lake (Anguillara) Bracciano Lake (Trevignano) The Tiber River The Tiber River (Farfa Natural Area) The Tiber River (IsolaTiberina) The Tiber River (Magliana) The Tiber River (Fiumicino)

Analyte concentration (␮g/L) TUa

TAP

MTU

PTU

MBI

PhTU

– – – – – –

n.d.b n.d.

Analysis of antithyroid drugs in surface water by using liquid chromatography-tandem mass spectrometry.

This paper describes development and validation of a new method for the simultaneous determination of six antithyroid drugs (ATDs) in surface waters b...
2MB Sizes 8 Downloads 10 Views