G Model

ARTICLE IN PRESS

CHROMA-356442; No. of Pages 10

Journal of Chromatography A, xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

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

Development and validation of a liquid chromatography–tandem mass spectrometry method for the quantitative determination of gamithromycin in animal plasma, lung tissue and pulmonary epithelial lining fluid Siegrid De Baere ∗ , Mathias Devreese, Anneleen Watteyn, Heidi Wyns, Elke Plessers, Patrick De Backer, Siska Croubels Department of Pharmacology, Toxicology and Biochemistry, Faculty of Veterinary Medicine, Ghent University, Salisburylaan 133, B-9820 Merelbeke, Belgium

a r t i c l e

i n f o

Article history: Received 30 January 2015 Received in revised form 5 April 2015 Accepted 13 April 2015 Available online xxx Keywords: Gamithromycin LC–MS/MS Plasma Pulmonary epithelial lining fluid Lung Pharmacokinetics

a b s t r a c t A sensitive and specific method for the quantitative determination of gamithromycin in animal plasma, lung tissue and pulmonary epithelial lining fluid (PELF) using liquid chromatography combined with heated electrospray ionization tandem mass spectrometry (LC–MS/MS) was developed. The sample preparation was rapid, straightforward and consisted of a deproteinization and phospholipid removal step using an Oasis® OstroTM 96-well plate (chicken, turkey and calf plasma) or HybridSPE® -Phospholipid SPE cartridges (pig plasma and turkey lung tissue), while a liquid–liquid extraction with diethyl ether in alkaline medium was used for PELF of turkey poults. Chromatography was performed on a C18 Hypersil GOLD column using 0.01 M ammonium acetate in water with a pH of 9, and acetonitrile as mobile phases. The MS/MS instrument was operated in the positive electrospray ionization mode and the following selected reaction monitoring transitions were monitored for gamithromycin (protonated molecule > product ion): m/z 777.45 > 619.35 and m/z 777.45 > 157.80 for quantification and identification, respectively. The method was validated in-house: matrix-matched calibration graphs were prepared and good linearity (r ≥ 0.99) was achieved over the concentration ranges tested (2.5–10,000 ng mL−1 for chicken, pig and calf plasma; 5.0–2500 ng mL−1 for turkey plasma; 50–10,000 ng g−1 for turkey lung tissue and 20–1000 ng mL−1 for turkey PELF). Limits of quantification (LOQ) were 2.5 ng mL−1 for chicken, pig and calf plasma and 5.0 ng mL−1 for turkey plasma, while the limits of detection (LOD) ranged between 0.007 and 0.07 ng mL−1 . For lung tissue and PELF, respective LOQ and LOD values of 50 ng g−1 and 0.76 ng g−1 (lung tissue) and 20 ng mL−1 and 0.1 ng mL−1 (PELF) were obtained. The results for the within-day and between-day precision, expressed as relative standard deviation (RSD), fell within the maximal RSD values. The accuracy fell within −30% to +10% (concentrations 1–10 ng mL−1 ) or −20% to +10% (concentrations > 10 ng mL−1 or ng g−1 ) of the theoretical concentration. The method was successfully applied for the quantitative determination of gamithromycin in plasma samples of chickens, turkeys, pigs and calves; and in lung tissues and PELF of turkeys, all derived from pharmacokinetic studies in these animal species. © 2015 Elsevier B.V. All rights reserved.

1. Introduction

∗ Corresponding author. Salisburylaan 133, 9820 Merelbeke, Belgium. Tel.: +32 09 264 73 48; fax: +32 09 264 74 97. E-mail addresses: [email protected] (S. De Baere), [email protected] (M. Devreese), [email protected] (A. Watteyn), [email protected] (H. Wyns), [email protected] (E. Plessers), [email protected] (P. De Backer), [email protected] (S. Croubels).

Gamithromycin (GAM, (2R,3S,4R,5S,8R,10R,11R,12S,13S,14R)13-[(2,6-Dideoxy-3-C-methyl-3-O-methyl-␣-l-ribo-hexopyranosyl)oxy]-2-ethyl-3,4,10-trihydroxy-3,5,8,10,12,14-hexamethyl-7propyl-11-{[3,4,6-trideoxy-3-(dimethylamino)-␤-d-xylohexopyranosyl]oxy}-1-oxa-7-azacyclopentadecan-15-one) is a 15membered semi-synthetic macrolide antibiotic of the azalide sub-class with an uniquely positioned alkylated nitrogen at the 7a-position of the lactone ring (Fig. 1A). It has been developed

http://dx.doi.org/10.1016/j.chroma.2015.04.022 0021-9673/© 2015 Elsevier B.V. All rights reserved.

Please cite this article in press as: S. De Baere, et al., Development and validation of a liquid chromatography–tandem mass spectrometry method for the quantitative determination of gamithromycin in animal plasma, lung tissue and pulmonary epithelial lining fluid, J. Chromatogr. A (2015), http://dx.doi.org/10.1016/j.chroma.2015.04.022

G Model CHROMA-356442; No. of Pages 10 2

ARTICLE IN PRESS S. De Baere et al. / J. Chromatogr. A xxx (2015) xxx–xxx

Fig. 1. (A) Chemical structure, chemical formula and molecular weight of GAM and d5-GAM; (B) MS/MS spectra of GAM and d5-GAM recorded at a collision energy of 35 V, with indication of the quantifier (Quan) and qualifier (Qual) product ions.

as the commercial formulation Zactran® (Merial, Duluth, GA, USA), to treat bovine respiratory disease caused by Gram-negative bacteria such as Mannheimia haemolytica, Pasteurella multocida and Histophilus somni [1,2]. GAM, like other macrolides, acts by inhibiting the bacterial protein synthesis by binding to the 50S subunit of the prokaryotic ribosomes [1]. In addition to their anti-infectious properties, macrolides have been reported to have immunomodulatory properties, such as inhibiting the release of pro-inflammatory cytokines and mediators, the migration of neutrophils and the oxidative burst in phagocytes [3–5]. In order to study the efficacy of GAM against respiratory infections in other animal species than cattle, pharmacokinetic (PK) data are mandatory. Moreover, macrolides accumulate in the target tissue, namely lung and pulmonary epithelial lining fluid (PELF), implying the need for PK data in those matrices as well and not only in plasma. To date, the PK of GAM in plasma as well as lung tissue and PELF has only been described in cattle [2,6] and foals [1]. Previous studies from our group reported the PK of GAM in plasma from broiler chickens [7] and pigs [8]. In order to investigate the PK of GAM in plasma as well as lung tissue and PELF from different animal species, sensitive analytical methods are mandatory. To date, only one method has been described and applied by several authors to determine GAM in biological samples, such as plasma, PELF and lung tissue [1,2,6], but this method had some disadvantages. The sample preparation for plasma, PELF and lung tissue used by Berghaus et al. [1], Huang et al. [2] and Giguère et al. [6] was rather time consuming, since it consisted of a dilution/homogenization in phosphate buffer, followed by centrifugation and further cleanup of the supernatant using an Oasis® MCX 96-well solid-phase extraction (SPE) plate (plasma, PELF) or cartridges (lung tissue). Chromatography was performed by pre-loading the sample onto an on-line Oasis® HLB SPE cartridge column, followed by elution and analysis onto a PLRP-S polymeric analytical column. This twostage sample preparation procedure is complex, labour intensive

and requires specific apparatus. Furthermore, two separate highperformance liquid chromatography (HPLC) pumps were necessary to deliver mobile phase to both the on-line SPE cartridge and the analytical column, which is not a standard configuration in an LC–MS/MS laboratory. Another drawback of the cited reports is the lack of validation results. The authors state that the method is validated, although no details are presented. For other macrolide antibiotics, off-line SPE using different types of sorbents has been adopted as a routine technique for sample preparation in various matrices (food, biological and environmental samples) [9]. However, these sample preparation procedures are rather time consuming and expensive as well, which is a disadvantage if a large amount of samples has to be analyzed such as for PK studies. Therefore, the aim and also the novelty of the present study was to develop and in-house validate a generic straightforward sample preparation procedure consisting of deproteinization and phospholipid removal (plasma and lung tissue) or liquid–liquid extraction (LLE, PELF), in combination with a sensitive and specific liquid chromatographic tandem mass spectrometric (LC–MS/MS) method that allowed the analysis of ≥100 samples per day. Incurred samples were analyzed to demonstrate the suitability of the developed method to determine the PK of GAM in plasma from different animal species (broiler chickens, turkey poults, pigs and calves) as well as in target tissues (PELF and lung) of turkey poults. 2. Material and methods 2.1. Chemicals, products and reagents The analytical standards of GAM (purity: 99.1%) and d5gamithromycin (d5-GAM, purity: 94.4%) were kindly donated by Merial and stored at 2–8 ◦ C. Water and acetonitrile (ACN) used for the mobile phases were of LC–MS grade and obtained from Biosolve (Valkenswaard, The Netherlands). All other solvents and reagents were of HPLC grade (water, ACN, methanol (MeOH)) or analytical

Please cite this article in press as: S. De Baere, et al., Development and validation of a liquid chromatography–tandem mass spectrometry method for the quantitative determination of gamithromycin in animal plasma, lung tissue and pulmonary epithelial lining fluid, J. Chromatogr. A (2015), http://dx.doi.org/10.1016/j.chroma.2015.04.022

G Model CHROMA-356442; No. of Pages 10

ARTICLE IN PRESS S. De Baere et al. / J. Chromatogr. A xxx (2015) xxx–xxx

grade (ethyl acetate (EtAc), diethyl ether (ether), formic acid, acetic acid, ammonium acetate, ammonium hydroxide, trichloroacetic acid (TCA), trifluoroacetic acid (TFA), sodium hydroxide (NaOH)) and were purchased from VWR (Leuven, Belgium). HybridSPE® -Phospholipid cartridges (30 mg mL−1 ) were obtained from Sigma–Aldrich (Diegem, Belgium). Oasis® OstroTM protein precipitation & phospholipid removal 96-well plates (25 mg) were obtained from Waters (Zellik, Belgium). Millex® -GN Nylon (0.20 ␮m) and Millex® -GV PVDF (0.22 ␮m) syringe filters were obtained from Merck-Millipore (Overijse, Belgium). 2.2. Preparation of standard solutions Stock solutions of GAM and the internal standard (IS, d5-GAM), both 1 mg mL−1 , were prepared in MeOH. Working solutions of 0.010, 0.025, 0.050, 0.10, 0.25, 0.50, 1.0, 2.5, 5.0, 10.0, 25.0, 50.0 and 100.0 ␮g mL−1 of GAM were prepared by appropriate dilutions of the stock and working solutions with water. For the IS, working solutions of 1.0 and 10.0 ␮g mL−1 were prepared in water. The stock solutions were stored at ≤−15 ◦ C, whereas working solutions were kept at 2–8 ◦ C. 2.3. Biological samples 2.3.1. Blank samples For the preparation of matrix-matched calibrator, quality control (QC) and validation samples, blank plasma was obtained from chickens, pigs, calves and turkeys that were not treated with macrolide antibiotics. In addition, blank PELF and lung tissue samples were obtained from the same turkeys of which plasma was derived. Blank samples were stored at ≤−15 ◦ C until analysis. 2.3.2. Incurred plasma samples Plasma samples were obtained from broiler chickens (April 2012), turkey poults (September 2013), pigs (December 2011) and calves (March 2013) which were subcutaneously (SC) administered 6 mg GAM kg body weight−1 (BW) (Zactran® , Merial). The blood samples were centrifuged (1500 × g, 10 min, 4 ◦ C) within 2 h after sampling and plasma was stored at ≤−15 ◦ C until analysis. The animal experiments were approved by the Ethical Committee of the Faculty of Veterinary Medicine and Bioscience Engineering of Ghent University (approval number EC 2011/159, EC 2012/031, EC 2013/107, EC 2013/12 for pigs, chickens, turkey poults and calves, respectively). 2.3.3. Incurred PELF and lung tissue samples Incurred PELF and lung tissue samples from turkey poults were obtained from the same birds of which plasma was sampled. The birds were euthanized by exsanguination, preceded by sedation with a combination of xylazine (XylM 2%, VMD, Arendonk, Belgium), zolazepam and tiletamine (Zoletil 100, Virbac, Wavre, Belgium). The whole right lung was weighed and homogenized with an equal weight of water using an Ultra Turrax mixer (Ika, Staufen, Germany). A 0.5 g aliquot was weighed into a 15 mL falcon tube. The complete left lung was preelevated to collect PELF as described by Bottje et al. [10]. Both the lung tissue and the PELF were stored at ≤−15 ◦ C until analysis. 2.4. Sample extraction and clean-up 2.4.1. Pig plasma To 250 ␮L of pig plasma were added 25 ␮L of the IS working solution (1.0 ␮g mL−1 ) and 750 ␮L of 1% formic acid in ACN, followed by a vortex mixing (15 s) and centrifugation step (10 min, 7825 × g). The supernatant was transferred to a HybridSPE® -Phospholipid

3

cartridge. The sample was passed drop wise through the cartridge and collected into an Eppendorf cup. Vacuum (15 mm Hg) was applied in order to remove all liquid drops. A 50-␮L aliquot of the filtrate was transferred to an autosampler vial and diluted with LC–MS grade water by a factor 1/2 for concentrations between 2.5 and 1000 ng mL−1 or by a factor 1/10 for concentrations >1000 ng mL−1 . After vortex mixing, a 2.5-␮L aliquot was injected onto the LC–MS/MS instrument. 2.4.2. Chicken, turkey and calf plasma To 100 ␮L of plasma were added 12.5 ␮L of the IS working solution (1.0 ␮g mL−1 ), followed by vortex mixing and loading onto the OstroTM 96-well plate. Thereafter, 300 ␮L of 1% formic acid in ACN were added and the sample was aspirated three times to enhance protein precipitation. The sample was passed through the 96-well plate by the application of a vacuum (15 mm Hg) for 10 min. The samples were diluted by the addition of 500 ␮L of LC–MS grade water, followed by gently vortex mixing of the 96-well plate (15 s). A 2.5-␮L aliquot was injected onto the LC–MS/MS instrument. 2.4.3. Lung tissue of turkey poults To 0.5 g of lung tissue homogenate (corresponding with 0.25 g of lung tissue), 50 ␮L of the IS working solution (10.0 ␮g mL−1 ) and 500 ␮L of water were added. After vortex mixing, the samples were equilibrated for 5 min at room temperature. Thereafter, 3 mL of a 1% solution of formic acid in ACN were added, followed by a vortex mixing (30 s) and centrifugation (10 min, 2851 × g, 4 ◦ C) step. The supernatant was transferred to a HybridSPE® -Phospholipid cartridge and the eluate was collected in a pyrex tube. Next, the samples were evaporated under a gentle nitrogen (N2 ) stream (40 ± 5 ◦ C) and the dry residue was reconstituted in 1 mL of water. After vortex mixing (30 s), a 200 ␮L aliquot was passed through a Millex® -GN Nylon syringe filter. The filtrate was collected in an autosampler vial and 800 ␮L of LC–MS grade water were added, followed by a vortex mixing step. A 2.0-␮L aliquot was injected onto the LC–MS/MS instrument. 2.4.4. Pulmonary epithelial lining fluid of turkey poults To 1 mL of PELF, 50 ␮L of the IS working solution (10.0 ␮g mL−1 ) and 100 ␮L of water were added. After vortex mixing, the samples were equilibrated for 5 min at room temperature. Next, 50 ␮L of a 10 M NaOH solution were added followed by vortex mixing (30 sec). Three millilitres of ether were added and the samples were extracted for 20 min on a roller mixer (Stuart Scientific, Surrey, UK) and centrifuged (2851 × g, 3 min, 4 ◦ C). The supernatant was transferred to another tube and evaporated under a gentle N2 stream (40 ± 5 ◦ C). The dry residue was reconstituted in 250 ␮L of LC–MS grade water. After vortex mixing (30 s), the sample was passed through a Millex® -GN Nylon syringe filter and transferred to an autosampler vial. A 2.0-␮L aliquot was injected onto the LC–MS/MS instrument. 2.5. Evaluation of adsorption of GAM to syringe filter membranes Prior to LC–MS/MS analysis, sample extracts were passed through a 0.2 ␮m syringe filter. In order to evaluate the adsorption of GAM and the IS, two types of filter membranes were tested, i.e. Millex® -GN Nylon and Millex® -GV PVDF. Two hundred microlitres of a 100 ng mL−1 solution of GAM and IS in water (A), water/ACN (90/10, v/v) (B) and 1% acetic acid in water (C) were passed through a syringe filter device and the filtrate was transferred to an autosampler vial. A 2.5-␮L aliquot was injected onto the LC–MS/MS system. The peak areas of GAM and the IS were compared with the peak areas in a corresponding standard

Please cite this article in press as: S. De Baere, et al., Development and validation of a liquid chromatography–tandem mass spectrometry method for the quantitative determination of gamithromycin in animal plasma, lung tissue and pulmonary epithelial lining fluid, J. Chromatogr. A (2015), http://dx.doi.org/10.1016/j.chroma.2015.04.022

G Model CHROMA-356442; No. of Pages 10

ARTICLE IN PRESS S. De Baere et al. / J. Chromatogr. A xxx (2015) xxx–xxx

4

solution that was analyzed without filtration. The experiment was performed in duplicate for each solvent and filter type.

2.6. LC–MS/MS analysis 2.6.1. Chromatography The LC system consisted of a quaternary, low-pressure mixing pump with vacuum degassing, type Surveyor MSpump Plus and an autosampler with temperature controlled tray and column oven, type Autosampler Plus, from ThermoFisher Scientific (Breda, The Netherlands). Chromatographic separation was achieved on a reversed-phase Hypersil GOLD (50 mm × 2.1 mm i.d., dp: 1.9 ␮m, part number: 25002-052130, C18) column in combination with a pre-column of the same type (10 mm × 2.1 mm i.d., dp: 3 ␮m), both from ThermoFisher Scientific. The mobile phase A consisted of 0.01 M ammonium acetate in water, adjusted with ammonium hydroxide to a pH of 9, whereas mobile phase B was ACN. The following gradient was applied: 0–0.5 min (70% A, 30% B), 0.5–1.5 min (linear gradient to 95% B), 1.5–4.0 min (5% A, 95% B), 4.0–4.3 min (linear gradient to 30% B), 4.3–7.0 min (70% A, 30% B). The flow-rate was 300 ␮L min−1 . The temperatures of the column oven and autosampler tray were set at 50 ◦ C and 5 ◦ C, respectively.

2.7. Evaluation of matrix effects, extraction recovery and process efficiency Extraction recovery (RE ), matrix effects (ME ), and process efficiency (PE ) were quantitatively assessed by preparing three sets of samples: set A consisted of standard solutions containing GAM at a concentration of 100 ng mL−1 ; the other sets consisted of matrixmatched samples that were prepared by spiking blank matrix after (set B) and before (set C) extraction at a concentration of 100 ng mL−1 or ng g−1 of GAM. The RE , ME and PE were determined by dividing the peak areas of GAM in the respective samples, i.e. RE = C/B × 100, ME = B/A × 100 and PE = C/A × 100. All experiments were performed in triplicate. Matrix effects were also qualitatively assessed by applying the post-column infusion method as described by Bonfiglio et al. [12]. Briefly, mobile phase was delivered to the HESI interface at a flow rate of 300 ␮L min−1 . A standard solution of GAM (10 ␮g mL−1 ) was infused post-column at a flow-rate of 20 ␮L min−1 through a T-piece, using a Harvard Pump 11 syringe pump (Uno B.V, Zevenaar, The Netherlands). Blank plasma sample extracts (10 ␮L) were injected onto the Hypersil GOLD column and the HPLC column effluent, combined with the infused analyte entered the HESI interface. MS/MS acquisition was performed as described above in Section 2.6.2. 2.8. Method validation

2.6.2. Mass spectrometry The LC column effluent was interfaced to a TSQ® Quantum Ultra triple quadrupole mass spectrometer (MS), equipped with a heated electrospray ionization (HESI) probe operating in the positive mode (all from ThermoFisher Scientific). Due to the combination of ESI with heated auxiliary gas, the HESI probe delivers better desolvation and improved nozzle performance for enhanced sensitivity. A divert valve was used and the LC effluent was directed to the MS instrument from 2.6 to 4.2 min. Instrument parameters were optimized by syringe infusion of working solutions of 1 ␮g mL−1 of each compound (flow-rate: 10 ␮L min−1 ) in combination with the mobile phase (50% A, 50% B, flow-rate: 200 ␮L min−1 ). The resolution of the first and third quadrupole (Q1 and Q3) was set at 0.7 peak width at half maximum. The following parameters were used: spray voltage: 3000 V, vaporizer temperature: 300 ◦ C, sheath gas pressure: 40 au (arbitrary units), ion sweep gas pressure: 2.0 au, auxiliary gas pressure: 20 au, capillary temperature: 300 ◦ C, tube lens offset: 100 V, source CID collision energy: −5 eV, collision pressure: 1.5 mTorr and quad MS/MS bias: −2.0. MS/MS acquisition was performed in the selected reaction monitoring (SRM) mode. For GAM, the two most intense precursor ion > product ion transitions were monitored, i.e. at m/z 777.45 > 619.35 (collision energy (CE): 35 V) and m/z 777.45 > 157.80 (CE: 45 V) for quantification and identification, respectively. For the IS, only one SRM transition was monitored, namely m/z 782.45 > 624.35 (CE: 35 V). The scan width and scan time were set at 0.01 and 0.3 s, respectively. The presence of residual phospholipids in chicken and pig sample extracts was evaluated according to Chambers et al. [11] by monitoring the SRM transitions of the following phospholipids: 1-palmitoyl-2-hydroxy-sn-glycero-3-phosphocholine (m/z 496.35 > 184.30), 1-stearoyl-2-hydroxy-sn-glycero-3-phosphocholine (m/z 524.37 > 184.30), 1-hexadecanoyl-2-(9Z,12Z-octadecadienoyl)-sn-glycero-3-phoshocholine (m/z 758.57 > 184.30), 1-(9Z,12Z-octadecadienoyl)-2-(5Z,8Z,11Z,14Z-eicosatetranoyl)sn-glycero-3-phosphocholine (m/z 806.57 > 184.30) and a fifth glycerophosphocholine lipid corresponding with the chemical formula C38 H74 NO8 P (m/z 703.57 > 184.30).

The performance characteristics of the method were evaluated by a set of parameters that were in compliance with the recommendations as defined by the European Community [13,14], by international guidelines [15] and in the literature [16]. Following parameters were evaluated: linearity, within-day and between-day accuracy and precision, limit of quantification (LOQ), limit of detection (LOD), specificity and carry-over. The validation protocol and the acceptance criteria used were previously described by De Baere et al. [17]. In addition, the storage stability of GAM was evaluated in stock (1 mg mL−1 ) and working (1 ␮g mL−1 ) solutions during storage at ≤−15 ◦ C and 2–8 ◦ C, respectively. A solution was considered stable if the peak area of GAM fell within ±10% of the corresponding peak area in a freshly prepared solution. Furthermore, the stability of GAM was evaluated in extracted samples during storage at 2–8 ◦ C and during 3 freeze–thaw cycles (freezing at ≤−15 ◦ C and thawing to room temperature). For each stability experiment, three chicken, pig, calf and turkey plasma samples were spiked with GAM at concentrations of 100 and 1000 ng mL−1 . GAM concentrations in the stability samples were determined using freshly prepared calibration curves and had to fall within −20% to +10% of the theoretical concentrations. The stability of GAM in matrix during storage at ≤−15 ◦ C was evaluated by re-analysing plasma samples of one turkey 16 months after the first analysis. 3. Results and discussion Several points of interest were taken into account during the optimization of a method for the analysis of GAM in animal plasma, lung tissue and PELF. First, the sample preparation procedure had to be straightforward and cost effective, since it was the final aim to analyze a high number of samples per day, especially for plasma analysis (n ≥ 100). To be able to analyze all these samples within 24 h, the LC–MS/MS analysis had to be accomplished within a short run-time (≤10 min). Secondly, simplicity and speed of analysis should not be reached at the expense of specificity, reliability and sensitivity of analysis. To evaluate this, a thorough method validation was performed.

Please cite this article in press as: S. De Baere, et al., Development and validation of a liquid chromatography–tandem mass spectrometry method for the quantitative determination of gamithromycin in animal plasma, lung tissue and pulmonary epithelial lining fluid, J. Chromatogr. A (2015), http://dx.doi.org/10.1016/j.chroma.2015.04.022

G Model CHROMA-356442; No. of Pages 10

ARTICLE IN PRESS S. De Baere et al. / J. Chromatogr. A xxx (2015) xxx–xxx

5

3.1. Sample extraction and clean-up 3.1.1. Plasma The preferred extraction method for macrolides in biological matrices (plasma, urine, broncho-alveolar cells) is liquid-liquid extraction (LLE) [9]. Labour-intensive SPE procedures with different types of sorbents (reversed-phase C18 or C2, cation-exchange MCX or CBA, hydrophylic–lipophylic balanced (HLB) co-polymer) were used in a limited number of studies, but were not an option for the present study, since it was the aim to analyze a high number of samples per day (n ≥ 100). Other sample preparation procedures consisted of the use of molecular mass cut-off filters, the application of the “dilute-and-shoot” principle or protein precipitation (PPT) using organic solvents such as ACN or MeOH [9]. Although “dilute-and-shoot” is the simplest sample preparation technique, it was not considered for the analysis of GAM in animal plasma since it can lead to clotting and pollution of the MS/MS instrument which impairs the sustainability of the apparatus. Ultrafiltration with molecular mass cut-off filters was not performed since only free analytes are determined, not including the protein-bound fractions. Therefore, clean-up of the plasma samples by protein precipitation, followed by high-speed centrifugation and analysis of an aliquot of the supernatant, was the first choice. During preliminary experiments on pig plasma, deproteinization was performed using organic solvents, since this sample preparation procedure had already successfully been applied at our laboratory for the analysis of tylosin – another macrolide antibiotic – in swine and chicken plasma [18]. Initially, the method seemed to be quite promising for the analysis of GAM with an extraction recovery of ∼60–75% and an LOQ of 2.5 ng mL−1 . However, when large batches (n > 50 per day) of chicken plasma samples were analyzed, problems occurred with the injector of the HPLC instrument (i.e. visible white precipitate in the sample transfer tube resulting in a blockage of the tubing and rotor seal). These problems did only occur with chicken plasma samples and could possibly be attributed to the presence of high amounts of phospholipids (PL) in the deproteinized sample. Ferlazzo et al. [19] reported indeed that phosphatidylcholine and sphingomyelin levels were significantly higher in chicken plasma compared to other species tested (pigs, cows, horses and ostriches). Moreover, Chambers et al. [11] mentioned that matrix components and endogenous phospholipids in particular, can be a serious source of imprecision in quantitative LC–MS/MS analysis. Devreese et al. [18] did not observe these issues probably due to the rather limited number of samples analyzed (n = 176) over a two-week period. To overcome the problem with PL, additional sample preparation methods were evaluated with the aim to remove interfering PL as much as possible (see supplementary material, Table S1): protein precipitation using strong acids (TCA 20%, TFA), LLE in alkaline medium using organic solvents (EtAc, ether) and protein precipitation in combination with PL removal using HybridSPE® Phospholipid cartridges or OstroTM 96-well plates. Supplementary table related to this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chroma.2015.04.022 Although clear extracts were obtained with TFA, the MS/MS sensitivity for GAM and the IS was low, which excluded the use of TFA for precipitation of proteins. Using TCA 20%, the extraction recovery (RE ) was ∼20% and could be slightly improved by the performance of a LLE step using ether (RE ∼ 30%). As shown in Fig. 2A, much better extraction recoveries (RE ∼ 90%) could be obtained using HybridSPE® -Phospholipid cartridges and the OstroTM 96well plate. In addition, PL were removed much more efficiently from the plasma matrix using these two latter techniques, compared with deproteinization using TCA 20% and LLE extraction using ether (Fig. 2B). This can be explained by the fact that none of the

Fig. 2. (A) Evaluation of the extraction recovery (RE , %), matrix effect (ME , %) and process efficiency (PE , %) of GAM in pig plasma after several sample preparation procedures: protein precipitation using 20% trichloroacetic acid (TCA20%), acetonitrile (ACN) and methanol (MeOH), protein precipitation in combination with liquid–liquid extraction using diethyl ether (LLE-ether), protein precipitation in combination with phospholipid removal (Hybrid-SPE and Ostro). Experiments were performed in triplicate; mean values + standard deviations (SD) have been shown; (B) evaluation of the efficiency of phospholipid (PL) removal from blank chicken (Ch) and pig (P) plasma that was subjected to protein precipitation using 20% TCA (TCA20%), LLE using diethyl ether (LLEether) and protein precipitation in combination with phospholipid removal (Hybrid-SPE and Ostro); the peak areas of the following PL were determined: (m/z = 496.35 > 184.30), 1-palmitoyl-2-hydroxy-sn-glycero-3-phosphocholine (m/z = 524.37 > 184.30), 1-stearoyl-2-hydroxy-sn-glycero-3-phosphocholine 1-hexadecanoyl-2-(9Z,12Z-octadecadienoyl)-sn-glycero-3-phoshocholine (m/z = 758.57 > 184.30), 1-(9Z,12Z-octadecadienoyl)-2-(5Z,8Z,11Z,14Zeicosatetranoyl)-sn-glycero-3-phosphocholine (m/z = 806.57 > 184.30) and a fifth glycerophosphocholine lipid corresponding with the chemical formula C38 H74 NO8 P (m/z = 703.57 > 184.30) [11]; *the peak area of this PL fell out of scale and was 1.16 × 109 ; (C) selected reaction monitoring chromatogram (m/z = 777.45 > 619.35) recorded during the post-column infusion of a GAM standard solution (concentration: 10 ␮g mL−1 , infusion flow-rate: 20 ␮L min−1 ) in combination with a 10-␮L on-column injection of a blank pig plasma extract, showing a signal suppression zone between 0.5 and 1.0 min and a signal enhancement zone >3.0 min; the elution zone of GAM in combination with a mobile phase at pH = 3.5 and pH = 9 is indicated with an arrow.

Please cite this article in press as: S. De Baere, et al., Development and validation of a liquid chromatography–tandem mass spectrometry method for the quantitative determination of gamithromycin in animal plasma, lung tissue and pulmonary epithelial lining fluid, J. Chromatogr. A (2015), http://dx.doi.org/10.1016/j.chroma.2015.04.022

G Model CHROMA-356442; No. of Pages 10

0.10 20 33.1 67.7 48.9 0.9987 ± 0.00088 −0.0284 ± 0.00098 0.0022 ± 0.00014

8.7 ± 2.09

0.76 50 41.0 50.4 81.4 0.9996 ± 0.00025 −0.0017 ± 0.00538 0.0005 ± 2.60 × 10−5

5.0 ± 0.52

0.07 5 83.4 91.9 90.7 0.9989 ± 0.00047 −0.0642 ± 0.00853

50–10,000

20–1000 Pulmonary epithelial lining fluid

5–2500

5, 10, 25, 50, 100, 250, 500, 1000, 2500 50, 100, 200, 500, 800, 1000, 2000, 5000, 8000, 10,000 20, 50, 80, 100, 200, 500, 800, 1000

0.0378 ± 3.61 × 10−5

8.0 ± 2.19

LOQ (ng mL−1 or ng g−1 ) RA (%) SSE (%) RE (%) r g (%) b a

Lung tissue

The results of the evaluation of the adsorption of GAM and the IS to a 0.22 ␮m Millex® -GV PVDF or a 0.20 ␮m Millex® -GN Nylon filter membrane are shown in Figure S1 of the supplementary material.

Plasma

3.2. Adsorption of GAM to syringe filter membranes

Spike levels (ng mL−1 or ng g−1 )

3.1.3. Pulmonary epithelial lining fluid PELF consists of lung exudate extracted from lung alveoli by flushing the lungs with saline. It does not contain PL and therefore, it was opted not to use a PL removal step in contrast to the sample preparation for plasma and lung tissue. To extract GAM and to remove salts from the PELF matrix, a LLE with ether was performed. As PELF mainly consists of saline and has a neutral pH, GAM was present in its ionized form (pKa of both N-groups: 9.78 and 8.88). To facilitate the extraction of GAM from the polar saline using an organic solvent, the compound had to be in its unionized form, which could be accomplished by the alkalinization of the sample using NaOH. As for lung tissue, the chromatographic properties of GAM were improved by evaporation of the supernatant and resuspension in water, followed by filtration to remove possible remaining solid particles. Although extraction recovery was intermediate using this simple extraction method (RE = 48.9%, Table 1), the method was sensitive enough for the aforementioned purpose. In addition, a good accuracy and precision could be obtained due to the use of a deuterated IS (GAM-d5), which fully compensated for analyte loss during the sample preparation procedure.

Calibration range (ng mL−1 or ng g−1 )

3.1.2. Lung tissue The extraction procedure for lung tissue was based on that for plasma analysis. For tissue samples, it was not possible to perform an on-line protein precipitation and phospholipid removal, using the OstroTM 96-well plate. Therefore, it was decided to perform an off-line protein precipitation using 1% formic acid in ACN, followed by the removal of PL using the HybridSPE® -Phospholipid cartridges. Compared to SPE procedures, which are generally applied for tissue analysis, the presented sample preparation was much less time consuming since no conditioning and washing steps of the SPE cartridge were needed. To improve the chromatographic properties of GAM onto the Hypersil Gold column, the eluate – consisting of 1% formic acid in ACN – was evaporated and the dry residue was reconstituted in water. Eventual solid particles were removed by filtrating the sample through a Millex® -GN Nylon filter, which was necessary prior to chromatography on a column packed with stationary phase particles of 1.9 ␮m diameter.

Matrix

mainstream sample preparation techniques (PPT, LLE and SPE) offer the possibility to remove PL from the plasma matrix due to the specific nature of these compounds: by PPT only gross levels of proteins are removed, whereas using LLE, PL can co-extract with the analytes of interest due to their physicochemical characteristics. Depending on the type of sorbent of SPE cartridges, PL co-extract due to their hydrophobic tail (reversed-phase SPE) or their zwitterionic polar head group (mixed-mode/ion exchange SPE) [20]. Since the best results for RE and matrix effects were obtained using the HybridSPE® -Phospholipid removal columns, this procedure was selected for the sample clean-up of pig plasma. However, for the subsequent extraction of chicken, turkey and calf plasma, it was decided to switch to the sample preparation procedure with the high-throughput OstroTM 96-well plate, because protein precipitation and PL removal could be performed in one action, which was an advantage compared to the HybridSPE® Phospholipid removal columns. By using these techniques, the simplicity of PPT could be combined with the selectivity of SPE by targeting the specific removal of both plasma proteins and PL, which are important contributors to ion-suppression in LC–MS/MS analysis, from the plasma sample.

LOD (ng mL−1 or ng g−1 )

S. De Baere et al. / J. Chromatogr. A xxx (2015) xxx–xxx Table 1 Results of the evaluation of linearity (slope (a), intercept (b), goodness-of-fit coefficient (g), correlation coefficient (r)), extraction recovery (RE ), signal suppression/enhancement (SSE), apparent recovery (RA ), limit of quantification (LOQ), limit of detection (LOD) for gamithromycin in turkey plasma, lung tissue and pulmonary epithelial lining fluid.

6

ARTICLE IN PRESS

Please cite this article in press as: S. De Baere, et al., Development and validation of a liquid chromatography–tandem mass spectrometry method for the quantitative determination of gamithromycin in animal plasma, lung tissue and pulmonary epithelial lining fluid, J. Chromatogr. A (2015), http://dx.doi.org/10.1016/j.chroma.2015.04.022

G Model CHROMA-356442; No. of Pages 10

ARTICLE IN PRESS S. De Baere et al. / J. Chromatogr. A xxx (2015) xxx–xxx

As can be seen, GAM and IS are strongly retained by the Millex® GV PVDF filter for all solvents (A: water, B: water/ACN (90/10, v/v) and C: 1% acetic acid in water). Using the Millex® -GN Nylon filters, a recovery of 88.5–103.5% was obtained if GAM and the IS were dissolved in pure aqueous solvents (A and C). The addition of 10% ACN to the dissolution solvent (B), resulted in higher adsorption of the analytes to the filter membrane (recovery ∼48%). Hence, it was decided to dissolve the sample extracts in 200 ␮L of water prior to filtration through Millex® -GN Nylon filters. Supplementary figure related to this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chroma.2015.04.022 3.3. LC–MS/MS analysis 3.3.1. Chromatography The aim of the present method was to develop an LC–MS/MS method that could be accomplished in less than 10 min, since a large amount of samples (n ≥ 100) had to be analyzed in a 24 h period. Therefore, a Hypersil GOLD column of 50 mm × 2.1 mm i.d. with 1.9 ␮m stationary phase particles was chosen. This column had ultra high-performance liquid chromatography (UHPLC) properties, but could be operated at a pressure of ∼150–200 bar, which is compatible with the Surveyor MS Plus HPLC pump. By the addition of 0.01 M ammonium acetate to the aqueous part of the mobile phase, a good peak-shape of GAM and the IS could be obtained.

7

The chromatographic retention of GAM on the HPLC column was determined by the mobile phase pH and could be increased from 1.0 min to 3.0 min by changing the pH from 3.5 to 9.0 (Fig. 3C). The latter mobile phase pH was chosen in the final procedure, after the qualitative investigation of matrix effects in pig plasma using the post-column infusion technique [12]. Using the acidic mobile phase at a pH of 3.5, GAM eluted at the end of the signal suppression zone caused by the solvent front (between 0.5–1.0 min), whereas GAM eluted just prior to a signal enhancement zone (>3.0 min) using the alkaline mobile phase at a pH of 9 (Fig. 2C). As can be seen from Figs. 3 and 4, some changes in retention time of GAM and the IS could be observed between the analysis of different matrices. This could be attributed to the difference in redissolution solvents for the different type of sample extracts (cf. Section 2.4) and to the large time spread between the analysis of GAM in the different matrices (i.e. up to 2 years), resulting in a change in column performance over time and in the use of several analytical columns with different batch numbers. 3.3.2. Mass spectrometry GAM and d5-GAM were tuned in the positive HESI mode. After optimization of MS instrument parameters, the protonated molecular ions of GAM ([M-H]+ , m/z = 777.45) and the IS ([M−H]+ , m/z = 782.45) were fragmented and MS/MS spectra were recorded (Fig. 1B). As can be seen, the most intense product ions were

Fig. 3. LC–MS/MS chromatogram from the analysis of (A) a blank turkey plasma sample, (B) a blank turkey plasma sample spiked with GAM at a concentration of 100 ng mL−1 , (C) a standard sample with a GAM concentration of 1000 ng mL−1 and eluted with a mobile phase at pH = 9 and (D) a turkey plasma sample that was taken at 0.5 h after subcutaneous administration of 6 mg GAM kg BW−1 ; GAM concentration: 115.0 ng mL−1 .

Please cite this article in press as: S. De Baere, et al., Development and validation of a liquid chromatography–tandem mass spectrometry method for the quantitative determination of gamithromycin in animal plasma, lung tissue and pulmonary epithelial lining fluid, J. Chromatogr. A (2015), http://dx.doi.org/10.1016/j.chroma.2015.04.022

G Model

ARTICLE IN PRESS

CHROMA-356442; No. of Pages 10

S. De Baere et al. / J. Chromatogr. A xxx (2015) xxx–xxx

8

Fig. 4. LC–MS/MS chromatogram from the analysis of (A) a turkey lung sample and (B) a turkey PELF sample that were taken at 2 days after the subcutaneous administration of 6 mg GAM kg BW−1 ; GAM concentration: 2677.1 ng g−1 and 35.5 ng mL−1 in lung and PELF, respectively.

observed at m/z = 619.35 (GAM) and 624.04 (d5-GAM) and selected for quantitative purposes. These quantifier ions were probably formed by the loss of the {[3,4,6-trideoxy-3-(dimethylamino)␤-d-xylohexopyranosyl]oxy} moiety at the C-11 position of the macrocyclic lactone ring. The second most intense product ion was observed at m/z = 158.08 for both GAM and d5-GAM and selected as qualifier ion. This can be attributed to the [(2,6-Dideoxy-3C-methyl-3-O-methyl-␣-l-ribo-hexopyranosyl)oxy] moiety at the C-13 position. However, the structure of the suggested product ions should be further confirmed by high-resolution mass spectrometric experiments. 3.4. Evaluation of matrix effects LC–MS/MS is known for its specificity and selectivity, but it has been shown that co-eluting matrix components may affect the ionization efficiency [11,21]. This phenomenon can be reduced by

performing an extensive sample clean-up and by optimizing the chromatographic separation. In the present method the influences of matrix effects have been further minimized by preparing matrix-matched calibrator samples and by the use of an isotope-labelled internal standard. The method of internal standardization was applied in order to compensate for variability during sample preparation and for matrix effects (ME ) during LC–MS/MS analysis. In the present method, an isotope-labelled IS (d5-GAM) was chosen, because of the identical structural and physico-chemical properties as the analyte of interest. This IS did indeed perfectly compensate for analyte loss during the sample preparation procedure, as RE and ME values for d5-GAM generally fell within 14% of those observed for GAM (results not shown). In addition, the IS co-eluted with GAM and therefore was subjected to the same matrix effects as GAM on the MS instrument [22].

Table 2 Results of the within-run and between-run precision and accuracy evaluation for the analysis of gamithromycin in turkey plasma, lung tissue and lung epithelial lining fluid. Matrix

Theoretical concentration (ng mL−1 or ng g−1 )

Mean concentration ± SD (ng mL−1 or ng g−1 )

Precision, RSD (%)

Accuracy (%)

Plasma

a

5 10a 10b 100a 100b 1000a 1000b

5.1 9.3 9.9 105.4 99.7 1036.8 943.5

± ± ± ± ± ± ±

0.99 1.27 1.08 4.38 11.58 40.22 65.34

19.5 13.6 10.9 4.2 11.6 3.9 6.9

1.6 −6.7 −1.0 5.4 −0.3 3.7 −5.7

Lung tissue

50a 100a 100b 1000a 1000b

46.6 93.0 109.7 988.1 962.7

± ± ± ± ±

5.62 8.31 14.36 42.12 81.12

12.1 8.9 13.1 4.3 8.4

−6.9 −7.0 9.7 −1.2 −3.7

20a 50a 50b 500a 500b

20.6 49.6 43.8 484.0 501.2

± ± ± ± ±

0.82 3.53 7.09 27.16 6.51

4.0 7.1 16.2 5.6 1.3

3.0 −0.7 −12.4 −3.2 0.2

Epithelial lining fluid

SD: standard deviation; RSD: relative standard deviation. Acceptance criteria: accuracy: 1–10 ng mL−1 : −30% to +10%, >10 ng mL−1 : −20% to +10%, within-run precision (RSDmax ):

Development and validation of a liquid chromatography-tandem mass spectrometry method for the quantitative determination of gamithromycin in animal plasma, lung tissue and pulmonary epithelial lining fluid.

A sensitive and specific method for the quantitative determination of gamithromycin in animal plasma, lung tissue and pulmonary epithelial lining flui...
2MB Sizes 0 Downloads 12 Views