Analytica Chimica Acta 805 (2013) 80–86

Contents lists available at ScienceDirect

Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca

Ultra-fast cyclosporin A quantitation in whole blood by Laser Diode Thermal Desorption – Tandem Mass Spectrometry; comparison with High Performance Liquid Chromatography–Tandem Mass Spectrometry Jean-Franc¸ois Jourdil a,∗ , Pierre Picard b , Cécile Meunier a , Serge Auger b , Franc¸oise Stanke-Labesque a,c,d a

CHU, Hôpital A. Michalon, Laboratoire de Pharmacologie-Toxicologie, BP217, Grenoble 38043, France Phytronix Technologies, Québec, Canada c Univ. Grenoble Alpes, F38041, France d INSERM U1042, 38041 Grenoble, France b

h i g h l i g h t s

g r a p h i c a l

a b s t r a c t

• We analyzed cyclosporin A in whole blood in 9 s using ultra-fast LDTDAPCI-MS/MS method. • We managed a simple sample preparation with protein precipitation. • Intra- and inter-day accuracies (%) and RSD (%) were all acceptable and respectively ±15%. • Method comparison with HPLC–MS/MS showed a good agreement and allowed to use this innovative technique for clinical routine.

a r t i c l e

i n f o

Article history: Received 8 August 2013 Received in revised form 24 October 2013 Accepted 29 October 2013 Available online 6 November 2013 Keywords: Cyclosporin A Tandem mass spectrometry Laser diode thermal desorption Liquid chromatography

a b s t r a c t In the last decade the quantitation of immunosuppressive drugs has seen vast improvements in analytical methods, optimizing time, accuracy of analysis and cost. Laser Diode Thermal Desorption (LDTD) coupled to Atmospheric Pressure Chemical ionization–tandem mass spectrometry (APCI-MS/MS) represents a technological breakthrough that removes the chromatographic separation step and thereby significantly increases the analytical throughput for the quantitation of cyclosporin A (CsA) in whole blood for therapeutic drug monitoring (TDM). A simple protein precipitation step was used prior to depositing 5 ␮L of the extract on a 96-well LazWellTM plate and CsA was quantified by LDTD-APCI-MS/MS. The laser pattern was set to ramp from 0 to 45% laser power within 2 s. The APCI parameters were set to negative needle voltage (−2 ␮A), carrier gas temperature (30 ◦ C) and air flow rate (3 L min−1 ). The negative ion single reaction monitoring transitions for CsA and its internal standard cyclosporin D (CsD) were respectively m/z 1201.1/1088.9 and m/z 1214.8/1102.8; obtained with a collision energy of −40 V. The analysis was achieved within 9 s from sample to sample. The extraction procedure yielded high recovery (92%; RSD = 9.4%, n = 6). The lower limit of quantitation was fixed at the first level of calibration: 23.5 ng mL−1 (accuracy = 112.3%; RSD = 9.6%; n = 6) and a blank + 6 point linear regression up to 965 ng mL−1 was used. Using 4 levels of quality control (QC), intra-day assays (n = 6) ranged from 93.5 to 95.7% (bias) and from 3.4 to 13.1% (RSD) while inter-day assays (n = 6) ranged

∗ Corresponding author at: Laboratory of Pharmacology and Toxicology, Grenoble University Hospital BP217, Grenoble F-38043, France. Tel.: +33 4 76 76 54 92. E-mail address: [email protected] (J.-F. Jourdil). 0003-2670/$ – see front matter © 2013 Published by Elsevier B.V. http://dx.doi.org/10.1016/j.aca.2013.10.051

J.-F. Jourdil et al. / Analytica Chimica Acta 805 (2013) 80–86

81

from 92.9 to 105.3% (bias) and from 4.9 to 7.5% (RSD). An inter-sample contamination of CsA of 2.3% was calculated that was considered negligible with respect to the range of CsA concentrations. Whole blood samples (120) from patients under CsA treatment were analyzed by LDTD-APCI-MS/MS and HPLC–ESIMS/MS, the gold standard reference method for CsA quantification. Both methods agreed (P ≥ 0.99), with a coefficient of correlation of 0.99 (95% confidence interval 0.982–0.991). The Passing–Bablok regression revealed no significant deviation from linearity (Cusum test, P = 0.11). This method seems suitable for use in TDM of CsA. © 2013 Published by Elsevier B.V.

1. Introduction Cyclosporin A (CsA) is one of the most widely used immunosuppressive drugs for the prevention of allograft rejection in solid organ and bone marrow transplantation. However, CsA has a narrow therapeutic window and large intra- and inter-patient variability in its pharmacokinetics. Thus, CsA therapeutic drug monitoring (TDM) is essential to optimize immunosuppressive therapy. At present, routine TDM of CsA is performed by measuring whole blood concentrations before CsA oral intake (C0 = through concentration ranging from 50 to 300 ng mL−1 ) or 2-h post dose (C2 concentration ranging from 400 to 1200 ng mL−1 after induction therapy) [1]. The quantitation of CsA and other immunosuppressants in whole blood has been the target of numerous technological developments. High Performance Liquid Chromatography in combination with Electrospray Tandem Mass Spectrometry (HPLC–ESI-MS/MS) [2–9] is now largely used because of its high specificity compared to immunoassay methods [10–13], and its capability to simultaneously analyze a range of immunosuppressive drugs. The HPLC–MS/MS methods were first optimized with offline sample preparation, either with liquid–liquid extraction (LLE) or solid phase extraction (SPE). While these two techniques were very effective in purifying samples, they were time consuming because several steps of sample purification and evaporation were needed. Sample preparation using protein precipitation procedures have been described for immunosuppressant analyses that allow twodimensional chromatography including an online sample clean up step. This new generation of analysis techniques has provided considerable progress for pharmacology laboratories because it offers good reproducibility, repeatability and accuracy with analysis times compatible with the requirements for urgency of routine hospital use. While HPLC–ESI-MS/MS remains the gold standard, recent technological advances have focused on faster methods providing the same sensitivity, precision and accuracy. Currently, the fastest method allows the analysis of the four main immunosuppressive drugs within 1.2 min [2]. Despite the emergence of Ultra High Performance Liquid Chromatography (UHPLC), the separation step remains the major limiting factor to shortening the analytical runs. A second time consuming step is the automatic injector that requires time to take the sample, inject it into the HPLC system and rinse the needle, ∼0.5–1 min/sample depending of the injector. Simple and rapid methods for introducing samples directly into mass spectrometers for chemical identification have been recently described for pharmaceutical and other compounds, including ambient ionization methods, such as desorption electrospray ionization [14–16], direct analysis in real time [17] or laser diode thermal desorption (LDTD) [18–28]. In the present study, we explored the use of LDTD as a rapid sample introduction technique in conjunction with atmospheric pressure chemical ionization (APCI) and tandem mass spectrometry (MS/MS) for the quantitation of CsA in whole blood. Then, using 120 blood samples from patients under CsA treatment, we

compared the concentrations of CsA obtained using this new method with our routine HPLC–ESI-MS/MS method. 2. Materials and methods 2.1. Chemical and reagents CsA and cyclosporin D (CsD, as internal standard (IS)) were kindly provided by Novartis Pharma (Basel, Switzerland). LC-MS-grade methanol (MeOH) and formic acid were purchased from Carlo Erba Reagents (Val de Reuil, France), zinc sulfate heptahydrate (ZnSO4 ·7H2 O) and Normapur® -grade ethyl acetate were respectively purchased from Alfa Aesar (Ward Hill, USA) and VWR (Radnor, USA). Ultrapure water (resistivity ≥ 18.0 M cm−1 ) was obtained using a Milli-Q Plus (Millipore, Molsheim, France). Polypropylene 2-milliliter (mL) centrifuge tubes and pipette tips were purchased from Eppendorf (Le Pecq, France) and Gilson (Middletown, WI, USA) respectively. LazWellTM 96-well plates were obtained from Phytronix Technologies (Quebec, Canada). Drug-free whole blood from volunteers was provided by the “Etablissement Franc¸ais du Sang” (Grenoble, France). Surplus whole blood samples from patients undergoing routine CsA TDM were also analyzed by LDTD-APCI-MS/MS. Whole blood samples were collected in EDTA and stored at +4 ◦ C until analysis. 2.2. Preparation of working solutions, calibration standards, and quality control samples Stock solutions of CsA and CsD were prepared in MeOH at 0.5 g L−1 , ultrasonicated for 10 min to enhance dissolution and stored at −20 ◦ C. Stock solutions of CsA and CsD were diluted to 200 ng mL−1 in methanol in order to optimize laser desorption and APCI ionization. The validation of the method was performed with the commercial 6PLUS1® multilevel whole blood calibrator (lot no. 4909) set and the 4 level MassCheck® internal quality control kit (lot no. 3510) supplied by Chromsystems (Munich, Germany). 2.3. Sample preparation for HPLC–ESI-MS/MS and for LDTD-APCI-MS/MS For both techniques, 100 ␮L of EDTA-treated whole blood was treated in Eppendorf polypropylene tubes with 200 ␮l precipitation reagent (methanol/0.2 M ZnSO4 (80/20, v/v)) including CsD at a concentration of 200 ng mL−1 . Due to the high distribution of CsA into erythrocytes, zinc sulfate solution was added to perform hemolysis and methanol to precipitate proteins. Samples were immediately vortexed during 20 s and centrifuged for 15 min at 20,800 × g at room temperature. For HPLC–ESI-MS/MS analysis, supernatants of 200 ␮L were transferred into integrated microinsert polypropylene HPLC vials ready to be injected (30 ␮L). For LDTD-APCI-MS/MS analysis, supernatants of 200 ␮L were evaporated to dryness under nitrogen prior to reconstitution with water (25 ␮L) and ethyl acetate (50 ␮L). This mixture of water and ethyl

82

J.-F. Jourdil et al. / Analytica Chimica Acta 805 (2013) 80–86

Table 1 MRM transitions, voltage settings declustering potential (DP), entrance potential (EP), collision energy (CE), and collision cell exit potential (CXP) for CsA and its IS CsD. Dwell times were optimized to obtain at least 15 points per peak. Analyte

Q3 m/z

Dwell time (ms)

DP (V)

EP (V)

CE (V)

CXP (V)

(A) LDTD-APCI-MS/MS parameters 1201.1 CsA CsD 1214.8

Q1 m/z

1088.9 1102.8

25 25

−140 −140

−10 −10

−40 −40

−9 −9

(B) HPLC–ESI-MS/MS parameters CsA 1219.9 CsD 1233.8

1202.8 1216.9

80 50

51 41

41 45

14 14

acetate was optimized in order to obtain a high signal sensitivity and reproducibility (optimization data not shown). 5 ␮L of the upper layer were spotted on a 96-well LazWellTM plate. Extracts were allowed to evaporate to dryness at room temperature (drying time: 5 min) before loading the plate into the LDTD apparatus for analysis.

9.5 9.5

60 psi, ion source temperature at 600 ◦ C, nebulizer gas setting at 50 psi, and curtain gas setting at 40 psi, collision gas flow rate fixed at 8 arbitrary units. Ammonium adducts [m + NH4 ]+ were chosen as precursor ions for both CsA and CsD. The optimized parameters for each ion transition are shown in Table 1B. 2.5. LDTD-APCI-MSMS method validation

2.4. Instrumentation and analytical conditions 2.4.1. LDTD-APCI-MS/MS analysis Analysis was performed using a model S-960 LDTD ionization source from Phytronix Technologies coupled to an API 4000 triple quadrupole mass spectrometer (ABSciex, Toronto, Canada) and operated using LazSoft (version 4.1, Phytronix Technologies) and Analyst (version 1.5.2) software. CsA and CsD ionization, and laser desorption parameters were optimized by desorbing both analytes (200 ng mL−1 ) from the LazWellTM plate. The laser pattern used to control the power of the laser was optimized to give the best CsA and CsD desorption while minimizing the release of potentially interfering matrix residues. It was set as follows: 0.5 s at 0%; ramp to 45% over 2 s (0.5–2.5 s); hold for 2 s (2.5–4.5 s); drop to 0% in 0.1 s, then hold at 0% for 2.19 s (4.51–7.2 s), using a 20 W diode laser. Including the displacement of the well of the plate in front of the laser the analysis was achieved within 9 s from sample to sample. The APCI parameter settings were the following: negative corona discharge needle voltage (−2 ␮A), vaporization at ambient temperature, and carrier gas air flow rate set at 3 L min−1 . The curtain gas was set at 10 psi while the flow rate of the collision gas (CAD) was set at 4 arbitrary units. Although the LDTD ion source heats up (145 ◦ C within 3 s, i.e. 50 ◦ C s−1 ) no thermal fragmentation of the parent compounds was observed, leading us to choose some appropriate and reproducible [m − H]− precursor ions for multiple reaction monitoring (MRM). The optimized parameters for each ion transition are shown in Table 1A. 2.4.2. HPLC–ESI-MS/MS analysis Analyses were performed on the same mass spectrometer used for LDTD. The HPLC system includes two Ultimate 3000 RS quaternary pumps (Pump A and B), an autosampler and a column compartment (Ultimate 3000 RS, Dionex, Thermo Scientific, Germering, Germany). Online sample clean-up was performed isocratically at 4 mL min−1 with a first mobile phase (methanol/water 50/50, v/v) on a Perfusion Chromatography® column (Poros R1/20, 20 ␮m, 2.1 mm × 30 mm, Applied Biosystems, Darmstadt, Germany) at room temperature. Chromatographic separation was performed isocratically at 0.5 mL min−1 with a second mobile phase (methanol/ammonium acetate 20 mM 97/3 (v/v) + 0.1% acetic acid) on a Luna Phenyl-Hexyl analytical column (Phenomenex, 2 mm × 50 mm, 5 ␮m, Aschaffenburg, Germany) maintained at 60 ◦ C. The source was operated in positive ion mode with an ESI potential of +5500 V and the following parameters: turbo heater gas at

The selectivity was tested by the analysis of 10 blank plasma samples, treated with the precipitation reagent without IS, and 10 blank plasma samples treated with the precipitation reagent containing the IS to measure putative interference. For each molecule (CsA and CsD), the MRM signal was monitored to confirm the absence of interfering matrix compounds. Intra- and inter-day accuracy (bias) and precision were examined by replicate analyses (n = 6) of the 4 levels of the IQC. Inter-day accuracy and precision were assessed by desorption (n = 6) of the same IQCs on 6 different days (including 3 consecutive days among these 6) by several analysts, and were calculated as the percentage deviation of the average calculated concentration from the nominal concentration. Precision was expressed as the relative standard deviation (%RSD). The acceptance limits were