- Email: [email protected]
MS method for determination of tacrolimus in oral fluids
G Model
ARTICLE IN PRESS
CHROMB-20290; No. of Pages 6
Journal of Chromatography B, xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb
Development and validation of a sensitive and selective LC–MS/MS method for determination of tacrolimus in oral fluids Mwlod Ghareeb (Ph.D.), Fatemeh Akhlaghi (Pharm.D., Ph.D.) ∗ Department of Biomedical and Pharmaceutical Sciences, University of Rhode Island, Kingston, RI, United States
a r t i c l e
i n f o
Article history: Received 31 July 2016 Received in revised form 28 September 2016 Accepted 8 October 2016 Available online xxx Keywords: Assay Immunosuppressant LC–MS/MS Oral fluid Saliva Tacrolimus Therapeutic drug monitoring Transplantation Validation
a b s t r a c t Tacrolimus is a commonly used immunosuppressive agent in organ transplant recipients. Therapeutic drug monitoring (TDM) of tacrolimus is essential to adjust the dose and achieve optimal immunosuppression level. Routine TDM is practiced using whole blood samples obtained through venipuncture. However, tacrolimus concentration that is present in oral fluid (OF) can theoretically represent the free or pharmacologically active form of tacrolimus. In this study, we report the development and validation of a rapid, sensitive and selective liquid chromatography-tandem mass spectrometry (LC–MS/MS) method for quantification of tacrolimus in OF. Chromatographic separation were achieved on an Acquity UPLC BEH C18 column by a gradient elution using 2 mM ammonium acetate/0.1% (v/v) formic acid in water (mobile phase A) and in methanol (mobile phase B) with a 2.2 min chromatographic run time. Tacrolimus was extracted from OF with acetonitrile as the precipitating solvent. Both extraction and chromatography was optimized to provide optimal sample cleanness, negligible matrix effect, and optimal specificity. The lower limit of quantification (LLOQ) for the assay was set at 10 pg/mL using a 50 L aliquot of OF obtained by passive drool. The method demonstrated adequate accuracy and precision with accuracy between 94.5–103.6%, and coefficient of variation ranging from 4 to 9.8%. Tacrolimus was stable in OF for up to one month at −80 ◦ C and the extracted matrix was stable up to 48 h in auto-sampler at 20 ◦ C. The method showed high reproducibility as confirmed by incurred sample reanalysis test. This assay was employed in several clinical pharmacokinetic studies and could successfully measure the concentration of tacrolimus in OF. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Tacrolimus (TAC) is a commonly prescribed immunosuppressive agent used to prevent allograft rejection after organ transplantation [1]. It has a narrow therapeutic index as well as significant intra- and inter-subject pharmacokinetic variability [2,3]. Therefore, therapeutic drug monitoring (TDM) of tacrolimus is an integral part of post-transplant pharmacotherapy. Currently,
Abbreviations: ACN, acetonitrile; ASC, ascomycin; Cmin , trough blood concentration; CV, coefficient of variation; DBS, dried blood spot; ESI, electrospray ionization; FDA, Food and Drug Administration; IS, internal standard; ISR, incurred sample reanalysis; LC–MS/MS, liquid chromatography–tandem mass spectrometry; LLOQ, lower limit of quantification; ME, matrix effect; MeOH, methanol; MRM, multiple reaction monitoring; MS, mass spectrometry; MW, molecular weight; OF, oral fluid; QCs, quality controls; RTR, renal transplant recipients; S/N, signal to noise ratio; SD, standard deviation; TAC, tacrolimus; TDM, therapeutic drug monitoring. ∗ Corresponding author at: Clinical Pharmacokinetics Research Laboratory, Department of Biomedical and Pharmaceutical Sciences, University of Rhode Island, 495A College of Pharmacy, 7 Greenhouse Road, Kingston, RI 02881, United States. E-mail address: [email protected] (F. Akhlaghi).
peripheral venous blood is used for TAC monitoring [4]. However, TAC blood level shows weak correlation with TAC level in lymphocytes [5,6], allograft tissue [6,7] or with TAC free fraction [8]. As a result, whole blood sampling fails to provide a reliable prediction of immunosuppressive activity or adverse drug reaction [5–9]. Intra-lymphocyte [5,6,10,11] and intra-tissue [6,7,12–14] concentrations of TAC has been investigated as alternative sampling methods for TAC monitoring. Both of the sampling methods proved to be good indicators of therapeutic efficacy and predictors for allograft rejection in liver [6,7] and kidney [10] transplant recipients [15]. Nonetheless, the clinical use of these techniques is hindered by relatively large amounts of blood needed for intracellular TAC determination, laborious lymphocytes isolation/TAC extraction or the need for obtaining a biopsy to measure intraallograft drug concentration [12]. Dried blood spot (DBS) sampling was proposed as a less invasive alternative to venipuncture [16–22]. Tacrolimus concentration obtained using DBS showed good association with TAC venous level [16–18,20,22]. However, technical difficulties associated with drug extraction from DBS precluded the possibility of widespread use of
http://dx.doi.org/10.1016/j.jchromb.2016.10.008 1570-0232/© 2016 Elsevier B.V. All rights reserved.
Please cite this article in press as: M. Ghareeb, F. Akhlaghi, Development and validation of a sensitive and selective LC–MS/MS method for determination of tacrolimus in oral fluids, J. Chromatogr. B (2016), http://dx.doi.org/10.1016/j.jchromb.2016.10.008
G Model CHROMB-20290; No. of Pages 6
ARTICLE IN PRESS M. Ghareeb, F. Akhlaghi / J. Chromatogr. B xxx (2016) xxx–xxx
2
this method. Moreover, since the sampling site is still venous blood, it correlates poorly with drug concentration at the site of action. Only 1% of the total concentration of TAC in blood is present as free or pharmacologically active form. The free drug can penetrate into allograft and shows a higher correlation with organ rejection [8,9]. For example, the free fraction of another immunosuppressive agent cyclosporine was associated to a greater extent with the incidence of heart transplant rejection than total concentration of cyclosporine [23,24]. Similarly, the free concentration of TAC in blood or plasma was significantly lower in patients experiencing allograft rejection than in other patients, whereas the concentration of total TAC was not different [8,9]. Therefore, monitoring free TAC concentration instead of total concentration, would provide a closer approximation to clinical outcomes. Tacrolimus is present in oral fluid (OF) at a concentration that associates with free or unbound drug concentration. It must be noted that theoretically, unbound concentration in plasma (fu,plasma) is in equilibrium with the unbound concentration in saliva (fu,saliva). However, measuring tacrolimus fu,saliva would be very difficult because of adsorption to the surface of devices needed to measure unbound drug and inadequate sensitivity of analytical instruments. Therefore, measuring TAC concentration in OF may provide a favorable alternative to monitoring tacrolimus non-invasively with the added advantage of measuring free or pharmacologically active drug level. We have previously reported LC–MS/MS assays for determination of cyclosporine [25] and mycophenolic acid [26] in OFs. However, the only previously published assay on TAC concentration in OF [27] lacks sufficient information about analytical method development and validation process. In this study, we have developed a simple, sensitive and selective liquid chromatography-tandem mass spectrometry (LC–MS/MS) method and have validated it according to guidelines set by Food and Drug Administration of the United States. 1.1. Chemicals and reagents Tacrolimus (C44 H69 NO12, MW = 804.02, 1.0 mg/mL solution in acetonitrile) and the internal standard (IS) ascomycin (ASC, C43 H69 NO12, MW = 792.01, 1.0 mg/mL solution in acetonitrile) were purchased from Cerilliant Corporation (Round Rock, Texas, USA). Acetonitrile (ACN) (Optima LC/MS), ammonium acetate (Crystalline/HPLC), formic acid (Optima LC/MS), and methanol (MeOH) (Optima LC/MS) were purchased from Fisher Scientific (Fair Lawn, NJ, USA). Deionized water was obtained from Milli-Q Synthesis system fitted with Q-Gard 2 Purification Pack (Millipore, Bedford, MA, USA). AquaSil Siliconizing Fluid was purchased from Thermo Fisher Scientific Inc. (Franklin, MA, USA). Drug-free human OF from six donors was obtained from Bioreclamation Inc. (Westbury, NY, USA). 1.2. Instruments Samples were sonicated using Branson® Sonicator (Danbury, CT, USA) to produce a homogeneous mixture. The supernatant was obtained using Eppendorf 5810 centrifuge from Micro and Nanotechnology (Urbana, IL, USA). Samples were analyzed using liquid chromatography-tandem mass spectrometry (LC–MS/MS). The LC–MS/MS system consisted of Acquity UPLC from Waters Corp. (Milford, MA, USA) connected to a Xevo TQ MS mass spectrometer from Waters Corp. (Milford, MA, USA). MassLynxTM (V 4.1) was used to control the system and data acquisition, and data was processed using TargetLynxTM . The UPLC system had a binary pump and was equipped with integrated column heater. A twenty microliter sample loop was used to deliver 10 L of the sample in partial loop mode. Blood contamination of OF samples was assessed using
an ELISA method utilizing a SpectraMax M5e Microplate Reader (Sunnyvale, CA, USA). 1.3. Chromatographic condition An Acquity UPLC BEH C18 (50 × 2.1 mm) column with 1.7 mparticle size and 130 Å pores size (Waters Corp) was used for chromatographic separation. An Acquity UPLC BEH C18, (2.1 × 5 mm) VanGuard pre-column with 1.7 m particle size and 130 Å porosity (Waters Corp) was connected immediately to the inlet of the analytical column. The temperature of the column was kept at 60 ◦ C, and the auto-sampler temperature was maintained at 20 ◦ C. The chromatographic conditions were optimized, and the gradient elution described here provided the best separation of TAC and IS from suppressing phospholipid ions. Chromatographic separation was achieved using a mobile phase which consisted of water containing 2 mM ammonium acetate/0.1% (v/v) formic acid (solvent A); and MeOH containing 2 mM ammonium acetate/0.1% (v/v) formic acid (solvent B). The mobile phase was delivered at 0.4 mL/min flow rate. The run cycle started at 50% solvent (B) and increased gradually to 98% over 0.5 min and was maintained at this level until 1.8 min. To re-equilibrate the column for the next run, solvent (B) decreased within 0.1 min to 50% and remained at 50% until the end of the run at 2.2 min. The diversion valve was set to deliver the first 0.70 min and from 1.20 min until the end of each run to waste. The elution time of TAC and IS was 1.0 min. 1.4. Mass spectrometry condition Mass spectrometry detection and quantification of TAC and ASC was performed in positive electrospray ionization (ESI) and multiple reaction monitoring (MRM) modes. Intellistart tool was used to obtain initial mass spectrometry parameters in low mass resolution analysis mode followed by manual tuning to achieve highest possible sensitivity. Final mass spectrometry parameters were as following: collision energy (CE) = 22 and 20 for ASC and TAC respectively, cone voltage (CV) = 28, capillary voltage (kV) = 1.50, source temperature (◦ C) = 150, cone gas flow (L/h) = 25, desolvation gas flow (L/h) = 1000, and collision gas flow (mL/min) = 0.15. Ammonium adducts [M + NH4 ]+ were selected as precursors for MRM with transitions (m/z, Q1 → Q3) of (m/z, 809.30 → 756.30) and (m/z, 821.30 → 768.35) for ASC and TAC, respectively. 1.5. Preparation of standard, quality control, and internal standard solutions Sub-stock and working stock solutions of ASC and TAC were prepared from the original stock (1 mg/mL) using ACN and MeOH, respectively, and stored at −20 ◦ C. Standards and quality controls (QCs) were prepared by spiking the OF with serially diluted working stock solutions (<5% of total OF volume) to achieve desired concentrations. A final concentration of 400 pg/mL ASC solution in ACN was used as the precipitating solvent. 1.6. Clinical samples Study protocols were approved by Institutional Review Board at Rhode Island Hospital (Providence, RI). After giving informed consent, kidney transplant recipients attending kidney transplant clinics were recruited. All patients were on a triple immunosuppressive regimen including tacrolimus, prednisone, and mycophenolic or azathioprine. In two studies, patients were asked to give venous blood samples (about 4 mL collected with ethylenediaminetetraacetic acid as anti-coagulant) and matching OF samples. The OF samples were collected by passive drool into
Please cite this article in press as: M. Ghareeb, F. Akhlaghi, Development and validation of a sensitive and selective LC–MS/MS method for determination of tacrolimus in oral fluids, J. Chromatogr. B (2016), http://dx.doi.org/10.1016/j.jchromb.2016.10.008
G Model CHROMB-20290; No. of Pages 6
ARTICLE IN PRESS M. Ghareeb, F. Akhlaghi / J. Chromatogr. B xxx (2016) xxx–xxx
siliconized plastic cups. All blood and OF samples were kept on ice until transferred to the Biomedical and Pharmaceutical Sciences (BPS) department at University of Rhode Island and stored at −80 ◦ C until analyzed. 1.7. Sample extraction Calibration standards, quality controls (QCs), blank, and patients’ OF samples were allowed to thaw at room temperature. After vortex mixing for 5 s, the samples were sonicated for 5–10 s (depending on samples volume) to disintegrate salivary components and to produce a homogenous mixture. This is a crucial step that will ensure adequate recovery of TAC from OF. A 50 L aliquot of homogenized OF sample was transferred into a 1.5 mL polypropylene tube, and 100 L of precipitating solvent was added (IS final concentration was 400 pg/mL). After vortexing for 10 s, samples were centrifuged at 10,000g for 5 min at 20 ◦ C. The supernatant was then transferred into an auto-sampler vial, and 10 L was injected onto the column. 1.8. Statistical data analysis Statistical analysis was performed using the SPSS software (version 23, SPSS Inc., Chicago, IL, USA) and GraphPad Prism (version 5.0, GraphPad Software, Inc., La Jolla, CA, USA). Normal distribution of the data was checked graphically and confirmed with the Shapiro-Wilk test. 2. Assay validation 2.1. Standards and QCs The method was validated according to the current version of FDA guidance for industry bioanalytical method validation [28]. Tacrolimus to IS peak ratio against nominal tacrolimus concentration was used to construct the calibration curve and fitted using (1/x) weighting method. This weighing option gave the best fit as compared to not weighted or other weighing options. Calibration curve points were, 10, 20, 50, 100, 250, 750, 1440, and 1600 pg/mL. Quality control concentrations were set at 30, 200, and 1200 pg/mL. To determine accuracy and precision of the assay, three different batches of OF were spiked with the working stock solution to achieve standards and QCs (six replicates) and extracted as described in sample extraction section. 2.2. Sensitivity and selectivity The lower limit of quantification (LLOQ) was determined as the concentration that had signal to noise ratio (S/N) of at least 10, accuracy between 80 and 120%, and coefficient of variation (CV) less than 20%. Acceptance criteria for QCs were accuracy between 85 and 115% and CV less than 15%. Selectivity assessed by inspecting the presence of noise or peaks in chromatograms represent blank OF samples injections (from 6 donors) compared with LLOQ sample chromatogram. 2.3. Stability Stability studies were performed by measuring TAC concentrations at QC1 and QC3, in three replicates. Freeze and thaw (after three freeze and thaw cycles), bench-top, auto-sampler (by reinjecting one of validation batch after it was left in the auto-sampler for 24 h and 48 h), and short-term stability up to one month were investigated.
3
2.4. Matrix effect and recovery The presence and potential influence of matrix effect (ME) in OF were studied using two different methods. First, chromatograms were obtained from the post-column infusion test. This test involved a continuous infusion of 98% methanol (which represents the composition of mobile phase at the elution time of ASC and TAC) containing 1 ng/mL of ASC and TAC at 20 L/min flow rate after the column through a three-way (Tee) connection. After establishing the baseline, a 10 L of blank extracted OF sample was injected by HPLC. The obtained chromatogram was then checked for signs of ion suppression or enhancement in comparison to blank injection of a neat solution (1:2, water: ACN). Second, the possible interference of OF components, namely the phospholipids, was studied. To do so, MRM transitions of abundant phospholipids were added to MS method to enable us to visually locate the elution region of these phospholipids. The influence of increasing ratio of precipitating solvents on the ME was also studied to select the ratio that offers the cleanest chromatography trace. Two different sets of QC1 and QC3 samples were prepared, in triplicate, either by (1) QCs samples (set 1) prepared by adding ACN to OF samples spiked with TAC as prescribed in sample extraction section (pre-extraction spiked samples); (2) QCs samples (set 2) prepared using a mixture of de-ionized water: ACN (neat solution). In each set, different ratios of ACN were added (1:1, 1:2 and 1:3). In total, 18 samples were analyzed, 9 samples in each set. The absolute ME was measured by calculating the percentage of the ratio of mean peaks area of pre-extracted samples to samples prepared in de-ionized water/ACN mixture. Recovery was assessed by analyzing the third set of QCs samples (set 3) prepared by extracting blank OF first with 1:1, 1:2 and 1:3 ACN, followed by adding TAC working standard solutions to achieve required concentrations (post-extraction spiked samples). The recovery was determined by calculating the ratio of mean peaks area of pre-extraction samples (set 1) to post-extracted spiked samples (set 3) and was expressed as a percentage.
3. Results and discussion Recommended TAC trough blood concentration (Cmin ) in kidney transplant recipients ranges between 15 and 20 ng/mL immediately after transplantation [3]. TAC dose tapered gradually and maintenance Cmin can be set as low as 5–7 ng/mL after 12-months of transplant operation [3]. Since only 1% of TAC amount found in unbound form was capable of reaching OF, the expected OF concentration would range from 50 to 200 pg/mL. Therefore, optimum mass spectrometry and chromatographic conditions were sought to develop a method with adequate selectivity and sensitivity. To achieve the highest feasible selectivity, different analytical columns were tested in conjunction with mobile phases of varying composition and gradients. An Acquity UPLC BEH C18 was the best choice as it gave sharp and symmetrical peaks for TAC and IS. Given the above-mentioned UPLC and mass spectrometry conditions, it was possible to set LLOQ at 10 pg/mL with signal/noise ratio more than 10 (Fig. 1A). No carryover was detected when a double blank OF sample was injected following the highest calibration concentration (Fig. 1B). The calibration curve was constructed by plotting nominal standard concentrations against peak area ratios of the analyte to IS and fitted with 1/x weighted least squares linear regression. The method demonstrated adequate accuracy and precision with QCs accuracy between 94.5-103.6%, and CV within 4–9.8% (Table 1). The coefficient of determination (R-squared) calculated for validation batches (n = 3) was between 0.998–0.999. Stability studies, namely, freeze and thaw, bench top, autosampler, and short-term storage at − 80 ◦ C for up to four weeks
Please cite this article in press as: M. Ghareeb, F. Akhlaghi, Development and validation of a sensitive and selective LC–MS/MS method for determination of tacrolimus in oral fluids, J. Chromatogr. B (2016), http://dx.doi.org/10.1016/j.jchromb.2016.10.008
G Model
ARTICLE IN PRESS
CHROMB-20290; No. of Pages 6
M. Ghareeb, F. Akhlaghi / J. Chromatogr. B xxx (2016) xxx–xxx
4
Fig. 1. Shows Chromatograms of TAC at LLOQ (10 pg/mL) (1A) and the internal standard ASC (400 pg/mL) (1D). Chromatograms (1B) and (1E) show traces of TAC; and (1C) and (1F) show traces of ASC in injections of pooled blank OF and blank solvent, respectively, injected immediately after highest concentration of the calibration curve (1600 pg/mL).
Table 1 Summary of accuracy and precision data obtained from quality control (QC) samples obtained from three individual runs (mean ±% CV, each QC had 6 replicates in each validation run, Total = 18).
Quality control samples (pg/mL) Accuracy (%) Precision (%)
LLOQ
QC1
QC2
QC3
10 103.6 9.8
30 103.0 6.1
200 94.5 5.9
1200 99.4 4.0
Accuracy = (mean concentration/nominal concentration) × 100. Precision expressed as % CV = (standard deviation/mean) × 100.
were conducted (Table 2). TAC was stable in extracted matrix for up to 48 h, and no loss of stability was noticed during different stability studies. Possible interference from endogenous substances in OF was investigated and is represented in chromatograms obtained from a pooled blank OF from six donors (Fig. 1B) and a blank or neat solution (66% ACN) (Fig. 1C) were compared. Using the assay conditions described above, no signs of interference was noticed. Using methanol instead of ACN as an organic solvent improved the sensitivity. Utilization of LC/MS grade methanol showed to increase the sensitivity by approximately 20%. Positive mode
ionization and monitoring ammonium adduct [M+NH4 ]+ at (m/z 821.30 → 768.35) provided better signal compared to [M]+ and [M+Na]+ . 3.1. Matrix effect and recovery Co-eluting of the analytes with endogenous substances in OF may lead to either ion suppression or enhancement, which is collectively called ME [29–31]. The presence of ME could compromise the reproducibility and may result in bias in the analysis [32]. Different cleaning procedures were used in other methods that aimed to measure the concentration of immunosuppressive agents in OF. These techniques included solid phase extraction [25,26], analyte pre-concentration by drying and reconstituting [33] and simple protein precipitation using organic solvents [27,34]. Type and percentage of the precipitating solution could influence sensitivity and selectivity through its effect on the yield of the analytes and cleanness of the extracted sample. Acetonitrile has been reported to provide satisfactory protein precipitation in oral fluid samples [31]. Recovery of a number of drugs and ME achieved using MeOH and ACN as the precipitating solvents in plasma are comparable. However, MeOH tends to retain about 40% more phospholipids [35].
Table 2 The result of stability studies (mean ±% CV, N = 3) indicates adequate bench top, freeze-thaw and short-term stability of tacrolimus in oral fluids. QC concentrations
Bench top
Freeze & thaw
Auto-sampler
Short-term
24 h
48 h
1 weeks
4 weeks
QC1 (30 pg/mL) %Accuracy (%CV) QC3 (1200 pg/mL) %Accuracy (%CV)
111.1 (9.4)
108.2 (5.8)
105.3 (9.6)
102.5 (4.3)
98.4 (11.9)
100.8 (6.7)
98.8 (2.4)
111.9 (2.0)
102.2 (2.7)
105.6 (3.6)
98.9 (3.4)
102.9 (1.9)
Please cite this article in press as: M. Ghareeb, F. Akhlaghi, Development and validation of a sensitive and selective LC–MS/MS method for determination of tacrolimus in oral fluids, J. Chromatogr. B (2016), http://dx.doi.org/10.1016/j.jchromb.2016.10.008
G Model
ARTICLE IN PRESS
CHROMB-20290; No. of Pages 6
M. Ghareeb, F. Akhlaghi / J. Chromatogr. B xxx (2016) xxx–xxx
5
Table 3 Effect of different ratios of oral fluid sample: extraction solvent (ACN) on recovery and absolute matrix effect, expressed as mean peak area ± SD (n = 3). Matrix
De-ionized water Extracted OF Post-extraction Recovery of Spiked OF (%)
QC1 (30 pg/mL)
QC2 (200 pg/mL)
QC3 (1200 pg/mL)
1:1
1:2
1:3
1:1
1:2
1:3
1:1
1:2
1:3
104 ± 8 83 ± 7 82 ±8 101.6
92 ± 4 106 ± 8 106 ± 6 100.0
85 ± 15 112 ± 14 96 ± 12 116.3
1592 ± 45 1485 ± 139 1317 ± 36 112.7
1739 ± 50 1689 ± 54 1484 ± 67 113.8
1719 ± 142 1775 ± 47 1559 ± 67 113.8
4260 ± 577 3916 ± 139 3760 ± 245 104.1
4653 ± 109 4452 ± 139 4090 ± 20 108.8
4409 ± 138 4452 ± 63 3579 ± 59 124.3
Fig. 3. Composite chromatogram shows traces of MRM transitions of 6 major phospholipids, obtained from a chromatogram of extracted blank pooled OF injection overlaid on TAC injection at concentration.
Fig. 2. Effect of blank OF and blank solvent injections on chromatograms obtained from continues post-column infused mixture of tacrolimus (TAC) and ascomycin (ASC) overlaid on TAC at QC2 concentration (400 pg/mL) (2A) and ASC (2B).
Therefore, ACN was chosen as the extracting solvent. Belostotsky, et al. [27] used 3:1, ACN:OF in a study to measure TAC salivary concentration. To the best of our knowledge, no other study has investigated an optimal proportion of extracting solvent (ACN) that gives maximum recovery and sample cleaning up. To examine the effect of using different proportions of ACN on recovery and absolute ME, OF samples were extracted with an equal, double and triple amount of ACN (Table 3). As it can be seen from Table 3, there was slightly less variability in areas count in samples extracted with the double volume of ACN compared to other two categories. Based on these values, it is obvious that samples extracted with ACN three times of their volume gave exaggerated recoveries, while the other two groups showed comparable recovery ranges. For ME, the first group showed to have significant ion suppression of about 20% in QC1 samples. Considering variability in the acquired data, adequate sample cleaning, adequate recovery, and minimum sample dilution, two-folds of ACN was chosen as the sample to solvent ratio to ensure optimal protein precipitation. Matrix effect was explored visually using post-column infusion technique [30,31]. The composite chromatograms in Fig. 2A and B were obtained by overlying chromatograms acquired from inject-
Fig. 4. Bland-Altman plot of % difference between the repeated measurements plotted against mean differences.
ing a neat solution (66% ACN), a blank OF with a continuous infusion of a mixture of ASC and TAC (1 ng/mL) and a chromatogram of QC2 injection. The only areas of chromatograms that showed ion suppression are between 0.2–0.5 min, which is far enough from the retention times of ASC and TAC. Finally, potential co-elution of phospholipids was examined by adding MRM of transitions of the most common phospholipids to the mass spectrometry method [30]. Phospholipid transitions included were (m/z 496 → 184, 520 → 184, 522 → 184, 524 → 184, 758 → 184, 782 → 184). In early stages of method development, we have observed that ASC and TAC peaks were co-eluting with low molecular weight phospholipids (m/z 496 to 524).The other two phospholipids that have m/z more than 700 were less problematic and eluted long after analytes of interest. By manipulating the mobile phase gradient or composition, the full separation between the analytes of interest and the phospholipids was achieved (Fig. 3). To our knowledge, the interference of phospholipids with LC–MS/MS analysis of drugs in OF and resolution of this issue has not been reported previously. This assay presents an excellent level of sensitivity to enable determination of TAC concentration in OF obtained from stable
Please cite this article in press as: M. Ghareeb, F. Akhlaghi, Development and validation of a sensitive and selective LC–MS/MS method for determination of tacrolimus in oral fluids, J. Chromatogr. B (2016), http://dx.doi.org/10.1016/j.jchromb.2016.10.008
G Model CHROMB-20290; No. of Pages 6
ARTICLE IN PRESS M. Ghareeb, F. Akhlaghi / J. Chromatogr. B xxx (2016) xxx–xxx
6
Fig. 5. Chromatograms represent TAC at concentration near LLOQ (11.7 pg/mL), (A) and corresponding internal standard (B) in an oral fluid sample obtained from a kidney transplant recipient at 10 h after a 2 mg tacrolimus dose. The estimated concentration after applying the dilution factor of 3 was 35 pg/mL.
transplant recipients. In total, out of 181 OF samples analyzed, only one sample had a concentration lower than LLOQ with a calculated concentration around 8.5 pg/mL. The concentration of TAC ranged from 11.7–2864.4 pg/mL and 1.7–46.1 ng/mL for OF and whole blood respectively. The clinical finding of this study will be presented in a separate manuscript. 3.2. Incurred sample reanalysis The incurred sample reanalysis test was performed by reanalyzing about 10% of the samples (19 samples) [36,37]. Whenever many samples were available per patient, two samples were selected to represent absorption and elimination phases. The differences between the paired measurements were normally distributed; therefore, the use of Bland-Altman method was justified [36]. Repeatability was tested visually (Fig. 4) and a statistically good agreement between the two repeated measurements was observed (Fig. 4) since all points lie between or near the 95% confidence interval limit. The 95% limit of agreement was from −19.16 to +31.98% and the bias (the mean difference between two occasions) was 6.40%. Fig. 5A depict a representative TAC chromatogram near the LLOQ concentration and Fig. 5B shows the chromatogram of IS for the same sample. 4. Conclusion We report a highly sensitive, selective, simple and robust method for quantitative determination of TAC in oral fluids. A simple sample preparation and extraction protocol was developed and optimized to provide adequate sensitivity, appropriate samples cleanliness, excellent recovery and minimal interference with endogenous components. In addition to lowest reported LLOQ of TAC, this work is the first to study the effect of different proportions of precipitating solvent (ACN) on the ME and recovery. In addition, this is the first report that investigated and described interference of phospholipids with LC–MS/MS determination of a small molecule compound in OF. Disclosure No conflict of interest is declared. This study was funded by the Rhode Island Science and Technology Advisory Council (STAC) and the University of Rhode Island, College of Pharmacy. References [1] L.J. Scott, K. McKeage, S.J. Keam, G.L. Plosker, Drugs 63 (2003) 1247. [2] L.C. Borra, J.I. Roodnat, J.A. Kal, R.A. Mathot, W. Weimar, T. van Gelder, Nephrol. Dial. Transplant. 25 (2010) 2757. [3] C.E. Staatz, S.E. Tett, Clin. Pharmacokinet. 43 (2004) 623.
[4] P.J. Taylor, C.H. Tai, M.E. Franklin, P.I. Pillans, Clin. Biochem. 44 (2011) 14. [5] F. Lemaitre, M. Antignac, C. Fernandez, Clin. Biochem. 46 (2013) 1538. [6] A. Capron, J. Lerut, D. Latinne, J. Rahier, V. Haufroid, P. Wallemacq, Transpl. Int. 25 (2012) 41. [7] W.J. Sandborn, G.M. Lawson, T.J. Cody, M.K. Porayko, J.E. Hay, G.J. Gores, J.L. Steers, R.A. Krom, R.H. Wiesner, Hepatology 21 (1995) 70. [8] H. Zahir, G. McCaughan, M. Gleeson, R.A. Nand, A.J. McLachlan, Ther. Drug Monit. 26 (2004) 506. [9] H. Zahir, G. McCaughan, M. Gleeson, R.A. Nand, A.J. McLachlan, Br. J. Clin. Pharmacol. 57 (2004) 298. [10] A. Capron, M. Mourad, M. De Meyer, L. De Pauw, D.C. Eddour, D. Latinne, L. Elens, V. Haufroid, P. Wallemacq, Pharmacogenomics 11 (2010) 703. [11] A. Capron, F. Musuamba, D. Latinne, M. Mourad, J. Lerut, V. Haufroid, P.E. Wallemacq, Ther. Drug Monit. 31 (2009) 178. [12] B.D. Noll, J.K. Coller, A.A. Somogyi, R.G. Morris, G.R. Russ, D.A. Hesselink, T. Van Gelder, B.C. Sallustio, Ther. Drug Monit. 35 (2013) 617. [13] L. Elens, A. Capron, V.V. Kerckhove, J. Lerut, M. Mourad, D. Lison, P. Wallemacq, V. Haufroid, Pharmacogenet. Genomics 17 (2007) 873. [14] A. Capron, J. Lerut, C. Verbaandert, J. Mathys, O. Ciccarelli, R. Vanbinst, F. Roggen, C. De Reyck, J. Lemaire, P.E. Wallemacq, Ther. Drug Monit. 29 (2007) 340. [15] M. Ghareeb, F. Akhlaghi, Bioanalysis 7 (2015) 1037. [16] K. Hoogtanders, J. van der Heijden, M. Christiaans, A. van de Plas, J. van Hooff, L. Stolk, Transplantation 83 (2007) 237. [17] K. Hoogtanders, J. van der Heijden, M. Christiaans, P. Edelbroek, J.P. van Hooff, L.M. Stolk, J. Pharm. Biomed. Anal. 44 (2007) 658. [18] C.Y. Cheung, J. van der Heijden, K. Hoogtanders, M. Christiaans, Y.L. Liu, Y.H. Chan, K.S. Choi, A. van de Plas, C.C. Shek, K.F. Chau, C.S. Li, J. van Hooff, L. Stolk, Transpl. Int. 21 (2008) 140. [19] D.R. Koop, L.A. Bleyle, M. Munar, G. Cherala, A. Al-Uzri, J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 926 (2013) 54. [20] E. Hinchliffe, J.E. Adaway, B.G. Keevil, J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 883–884 (2012) 102. [21] J.C. den Burger, A.J. Wilhelm, A. Chahbouni, R.M. Vos, A. Sinjewel, E.L. Swart, Anal. Bioanal. Chem. 404 (2012) 1803. [22] N.J. Webb, D. Roberts, R. Preziosi, B.G. Keevil, Pediatr. Transplant. 9 (2005) 729. [23] F. Akhlaghi, J. Ashley, A. Keogh, K. Brown, Ther. Drug Monit. 21 (1999) 8. [24] F. Akhlaghi, A.M. Keogh, K.F. Brown, Transplantation 67 (1999) 54. [25] A. Mendonza, R. Gohh, F. Akhlaghi, Ther. Drug Monit. 26 (2004) 569. [26] A.E. Mendonza, R.Y. Gohh, F. Akhlaghi, Ther. Drug Monit. 28 (2006) 402. [27] V. Belostotsky, J. Adaway, B.G. Keevil, D.R. Cohen, N.J. Webb, Pediatr. Nephrol. 26 (2011) 133. [28] US Department of Health and Human Services. Food and Drug Administration (FDA). Center for Drug Evaluation and Research (CDER). 2001. http://www. fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/ Guidances/ucm070107.pdf (last accessed: Oct 2016). [29] H. Mei, Y. Hsieh, C. Nardo, X. Xu, S. Wang, K. Ng, W.A. Korfmacher, Rapid Commun. Mass Spectrom. 17 (2003) 97. [30] G. Zhang, C.E. Wujcik, J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 877 (2009) 2003. [31] R. Dams, M.A. Huestis, W.E. Lambert, C.M. Murphy, J. Am. Soc. Mass Spectrom. 14 (2003) 1290. [32] B.K. Matuszewski, M.L. Constanzer, C.M. Chavez-Eng, Anal. Chem. 75 (2003) 3019. [33] M.H. Wiesen, F. Farowski, M. Feldkotter, B. Hoppe, C. Muller, J. Chromatogr. A 1241 (2012) 52. [34] B. Shen, S. Li, Y. Zhang, X. Yuan, Y. Fan, Z. Liu, Q. Hu, C. Yu, J. Pharm. Biomed. Anal. 50 (2009) 515. [35] E. Chambers, D.M. Wagrowski-Diehl, Z. Lu, J.R. Mazzeo, J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 852 (2007) 22. [36] J.M. Bland, D.G. Altman, Lancet 1 (1986) 307. [37] P. van Amsterdam, A. Companjen, M. Brudny-Kloeppel, M. Golob, S. Luedtke, P. Timmerman, Bioanalysis 5 (2013) 645.
Please cite this article in press as: M. Ghareeb, F. Akhlaghi, Development and validation of a sensitive and selective LC–MS/MS method for determination of tacrolimus in oral fluids, J. Chromatogr. B (2016), http://dx.doi.org/10.1016/j.jchromb.2016.10.008
ARTICLE IN PRESS
CHROMB-20290; No. of Pages 6
Journal of Chromatography B, xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb
Development and validation of a sensitive and selective LC–MS/MS method for determination of tacrolimus in oral fluids Mwlod Ghareeb (Ph.D.), Fatemeh Akhlaghi (Pharm.D., Ph.D.) ∗ Department of Biomedical and Pharmaceutical Sciences, University of Rhode Island, Kingston, RI, United States
a r t i c l e
i n f o
Article history: Received 31 July 2016 Received in revised form 28 September 2016 Accepted 8 October 2016 Available online xxx Keywords: Assay Immunosuppressant LC–MS/MS Oral fluid Saliva Tacrolimus Therapeutic drug monitoring Transplantation Validation
a b s t r a c t Tacrolimus is a commonly used immunosuppressive agent in organ transplant recipients. Therapeutic drug monitoring (TDM) of tacrolimus is essential to adjust the dose and achieve optimal immunosuppression level. Routine TDM is practiced using whole blood samples obtained through venipuncture. However, tacrolimus concentration that is present in oral fluid (OF) can theoretically represent the free or pharmacologically active form of tacrolimus. In this study, we report the development and validation of a rapid, sensitive and selective liquid chromatography-tandem mass spectrometry (LC–MS/MS) method for quantification of tacrolimus in OF. Chromatographic separation were achieved on an Acquity UPLC BEH C18 column by a gradient elution using 2 mM ammonium acetate/0.1% (v/v) formic acid in water (mobile phase A) and in methanol (mobile phase B) with a 2.2 min chromatographic run time. Tacrolimus was extracted from OF with acetonitrile as the precipitating solvent. Both extraction and chromatography was optimized to provide optimal sample cleanness, negligible matrix effect, and optimal specificity. The lower limit of quantification (LLOQ) for the assay was set at 10 pg/mL using a 50 L aliquot of OF obtained by passive drool. The method demonstrated adequate accuracy and precision with accuracy between 94.5–103.6%, and coefficient of variation ranging from 4 to 9.8%. Tacrolimus was stable in OF for up to one month at −80 ◦ C and the extracted matrix was stable up to 48 h in auto-sampler at 20 ◦ C. The method showed high reproducibility as confirmed by incurred sample reanalysis test. This assay was employed in several clinical pharmacokinetic studies and could successfully measure the concentration of tacrolimus in OF. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Tacrolimus (TAC) is a commonly prescribed immunosuppressive agent used to prevent allograft rejection after organ transplantation [1]. It has a narrow therapeutic index as well as significant intra- and inter-subject pharmacokinetic variability [2,3]. Therefore, therapeutic drug monitoring (TDM) of tacrolimus is an integral part of post-transplant pharmacotherapy. Currently,
Abbreviations: ACN, acetonitrile; ASC, ascomycin; Cmin , trough blood concentration; CV, coefficient of variation; DBS, dried blood spot; ESI, electrospray ionization; FDA, Food and Drug Administration; IS, internal standard; ISR, incurred sample reanalysis; LC–MS/MS, liquid chromatography–tandem mass spectrometry; LLOQ, lower limit of quantification; ME, matrix effect; MeOH, methanol; MRM, multiple reaction monitoring; MS, mass spectrometry; MW, molecular weight; OF, oral fluid; QCs, quality controls; RTR, renal transplant recipients; S/N, signal to noise ratio; SD, standard deviation; TAC, tacrolimus; TDM, therapeutic drug monitoring. ∗ Corresponding author at: Clinical Pharmacokinetics Research Laboratory, Department of Biomedical and Pharmaceutical Sciences, University of Rhode Island, 495A College of Pharmacy, 7 Greenhouse Road, Kingston, RI 02881, United States. E-mail address: [email protected] (F. Akhlaghi).
peripheral venous blood is used for TAC monitoring [4]. However, TAC blood level shows weak correlation with TAC level in lymphocytes [5,6], allograft tissue [6,7] or with TAC free fraction [8]. As a result, whole blood sampling fails to provide a reliable prediction of immunosuppressive activity or adverse drug reaction [5–9]. Intra-lymphocyte [5,6,10,11] and intra-tissue [6,7,12–14] concentrations of TAC has been investigated as alternative sampling methods for TAC monitoring. Both of the sampling methods proved to be good indicators of therapeutic efficacy and predictors for allograft rejection in liver [6,7] and kidney [10] transplant recipients [15]. Nonetheless, the clinical use of these techniques is hindered by relatively large amounts of blood needed for intracellular TAC determination, laborious lymphocytes isolation/TAC extraction or the need for obtaining a biopsy to measure intraallograft drug concentration [12]. Dried blood spot (DBS) sampling was proposed as a less invasive alternative to venipuncture [16–22]. Tacrolimus concentration obtained using DBS showed good association with TAC venous level [16–18,20,22]. However, technical difficulties associated with drug extraction from DBS precluded the possibility of widespread use of
http://dx.doi.org/10.1016/j.jchromb.2016.10.008 1570-0232/© 2016 Elsevier B.V. All rights reserved.
Please cite this article in press as: M. Ghareeb, F. Akhlaghi, Development and validation of a sensitive and selective LC–MS/MS method for determination of tacrolimus in oral fluids, J. Chromatogr. B (2016), http://dx.doi.org/10.1016/j.jchromb.2016.10.008
G Model CHROMB-20290; No. of Pages 6
ARTICLE IN PRESS M. Ghareeb, F. Akhlaghi / J. Chromatogr. B xxx (2016) xxx–xxx
2
this method. Moreover, since the sampling site is still venous blood, it correlates poorly with drug concentration at the site of action. Only 1% of the total concentration of TAC in blood is present as free or pharmacologically active form. The free drug can penetrate into allograft and shows a higher correlation with organ rejection [8,9]. For example, the free fraction of another immunosuppressive agent cyclosporine was associated to a greater extent with the incidence of heart transplant rejection than total concentration of cyclosporine [23,24]. Similarly, the free concentration of TAC in blood or plasma was significantly lower in patients experiencing allograft rejection than in other patients, whereas the concentration of total TAC was not different [8,9]. Therefore, monitoring free TAC concentration instead of total concentration, would provide a closer approximation to clinical outcomes. Tacrolimus is present in oral fluid (OF) at a concentration that associates with free or unbound drug concentration. It must be noted that theoretically, unbound concentration in plasma (fu,plasma) is in equilibrium with the unbound concentration in saliva (fu,saliva). However, measuring tacrolimus fu,saliva would be very difficult because of adsorption to the surface of devices needed to measure unbound drug and inadequate sensitivity of analytical instruments. Therefore, measuring TAC concentration in OF may provide a favorable alternative to monitoring tacrolimus non-invasively with the added advantage of measuring free or pharmacologically active drug level. We have previously reported LC–MS/MS assays for determination of cyclosporine [25] and mycophenolic acid [26] in OFs. However, the only previously published assay on TAC concentration in OF [27] lacks sufficient information about analytical method development and validation process. In this study, we have developed a simple, sensitive and selective liquid chromatography-tandem mass spectrometry (LC–MS/MS) method and have validated it according to guidelines set by Food and Drug Administration of the United States. 1.1. Chemicals and reagents Tacrolimus (C44 H69 NO12, MW = 804.02, 1.0 mg/mL solution in acetonitrile) and the internal standard (IS) ascomycin (ASC, C43 H69 NO12, MW = 792.01, 1.0 mg/mL solution in acetonitrile) were purchased from Cerilliant Corporation (Round Rock, Texas, USA). Acetonitrile (ACN) (Optima LC/MS), ammonium acetate (Crystalline/HPLC), formic acid (Optima LC/MS), and methanol (MeOH) (Optima LC/MS) were purchased from Fisher Scientific (Fair Lawn, NJ, USA). Deionized water was obtained from Milli-Q Synthesis system fitted with Q-Gard 2 Purification Pack (Millipore, Bedford, MA, USA). AquaSil Siliconizing Fluid was purchased from Thermo Fisher Scientific Inc. (Franklin, MA, USA). Drug-free human OF from six donors was obtained from Bioreclamation Inc. (Westbury, NY, USA). 1.2. Instruments Samples were sonicated using Branson® Sonicator (Danbury, CT, USA) to produce a homogeneous mixture. The supernatant was obtained using Eppendorf 5810 centrifuge from Micro and Nanotechnology (Urbana, IL, USA). Samples were analyzed using liquid chromatography-tandem mass spectrometry (LC–MS/MS). The LC–MS/MS system consisted of Acquity UPLC from Waters Corp. (Milford, MA, USA) connected to a Xevo TQ MS mass spectrometer from Waters Corp. (Milford, MA, USA). MassLynxTM (V 4.1) was used to control the system and data acquisition, and data was processed using TargetLynxTM . The UPLC system had a binary pump and was equipped with integrated column heater. A twenty microliter sample loop was used to deliver 10 L of the sample in partial loop mode. Blood contamination of OF samples was assessed using
an ELISA method utilizing a SpectraMax M5e Microplate Reader (Sunnyvale, CA, USA). 1.3. Chromatographic condition An Acquity UPLC BEH C18 (50 × 2.1 mm) column with 1.7 mparticle size and 130 Å pores size (Waters Corp) was used for chromatographic separation. An Acquity UPLC BEH C18, (2.1 × 5 mm) VanGuard pre-column with 1.7 m particle size and 130 Å porosity (Waters Corp) was connected immediately to the inlet of the analytical column. The temperature of the column was kept at 60 ◦ C, and the auto-sampler temperature was maintained at 20 ◦ C. The chromatographic conditions were optimized, and the gradient elution described here provided the best separation of TAC and IS from suppressing phospholipid ions. Chromatographic separation was achieved using a mobile phase which consisted of water containing 2 mM ammonium acetate/0.1% (v/v) formic acid (solvent A); and MeOH containing 2 mM ammonium acetate/0.1% (v/v) formic acid (solvent B). The mobile phase was delivered at 0.4 mL/min flow rate. The run cycle started at 50% solvent (B) and increased gradually to 98% over 0.5 min and was maintained at this level until 1.8 min. To re-equilibrate the column for the next run, solvent (B) decreased within 0.1 min to 50% and remained at 50% until the end of the run at 2.2 min. The diversion valve was set to deliver the first 0.70 min and from 1.20 min until the end of each run to waste. The elution time of TAC and IS was 1.0 min. 1.4. Mass spectrometry condition Mass spectrometry detection and quantification of TAC and ASC was performed in positive electrospray ionization (ESI) and multiple reaction monitoring (MRM) modes. Intellistart tool was used to obtain initial mass spectrometry parameters in low mass resolution analysis mode followed by manual tuning to achieve highest possible sensitivity. Final mass spectrometry parameters were as following: collision energy (CE) = 22 and 20 for ASC and TAC respectively, cone voltage (CV) = 28, capillary voltage (kV) = 1.50, source temperature (◦ C) = 150, cone gas flow (L/h) = 25, desolvation gas flow (L/h) = 1000, and collision gas flow (mL/min) = 0.15. Ammonium adducts [M + NH4 ]+ were selected as precursors for MRM with transitions (m/z, Q1 → Q3) of (m/z, 809.30 → 756.30) and (m/z, 821.30 → 768.35) for ASC and TAC, respectively. 1.5. Preparation of standard, quality control, and internal standard solutions Sub-stock and working stock solutions of ASC and TAC were prepared from the original stock (1 mg/mL) using ACN and MeOH, respectively, and stored at −20 ◦ C. Standards and quality controls (QCs) were prepared by spiking the OF with serially diluted working stock solutions (<5% of total OF volume) to achieve desired concentrations. A final concentration of 400 pg/mL ASC solution in ACN was used as the precipitating solvent. 1.6. Clinical samples Study protocols were approved by Institutional Review Board at Rhode Island Hospital (Providence, RI). After giving informed consent, kidney transplant recipients attending kidney transplant clinics were recruited. All patients were on a triple immunosuppressive regimen including tacrolimus, prednisone, and mycophenolic or azathioprine. In two studies, patients were asked to give venous blood samples (about 4 mL collected with ethylenediaminetetraacetic acid as anti-coagulant) and matching OF samples. The OF samples were collected by passive drool into
Please cite this article in press as: M. Ghareeb, F. Akhlaghi, Development and validation of a sensitive and selective LC–MS/MS method for determination of tacrolimus in oral fluids, J. Chromatogr. B (2016), http://dx.doi.org/10.1016/j.jchromb.2016.10.008
G Model CHROMB-20290; No. of Pages 6
ARTICLE IN PRESS M. Ghareeb, F. Akhlaghi / J. Chromatogr. B xxx (2016) xxx–xxx
siliconized plastic cups. All blood and OF samples were kept on ice until transferred to the Biomedical and Pharmaceutical Sciences (BPS) department at University of Rhode Island and stored at −80 ◦ C until analyzed. 1.7. Sample extraction Calibration standards, quality controls (QCs), blank, and patients’ OF samples were allowed to thaw at room temperature. After vortex mixing for 5 s, the samples were sonicated for 5–10 s (depending on samples volume) to disintegrate salivary components and to produce a homogenous mixture. This is a crucial step that will ensure adequate recovery of TAC from OF. A 50 L aliquot of homogenized OF sample was transferred into a 1.5 mL polypropylene tube, and 100 L of precipitating solvent was added (IS final concentration was 400 pg/mL). After vortexing for 10 s, samples were centrifuged at 10,000g for 5 min at 20 ◦ C. The supernatant was then transferred into an auto-sampler vial, and 10 L was injected onto the column. 1.8. Statistical data analysis Statistical analysis was performed using the SPSS software (version 23, SPSS Inc., Chicago, IL, USA) and GraphPad Prism (version 5.0, GraphPad Software, Inc., La Jolla, CA, USA). Normal distribution of the data was checked graphically and confirmed with the Shapiro-Wilk test. 2. Assay validation 2.1. Standards and QCs The method was validated according to the current version of FDA guidance for industry bioanalytical method validation [28]. Tacrolimus to IS peak ratio against nominal tacrolimus concentration was used to construct the calibration curve and fitted using (1/x) weighting method. This weighing option gave the best fit as compared to not weighted or other weighing options. Calibration curve points were, 10, 20, 50, 100, 250, 750, 1440, and 1600 pg/mL. Quality control concentrations were set at 30, 200, and 1200 pg/mL. To determine accuracy and precision of the assay, three different batches of OF were spiked with the working stock solution to achieve standards and QCs (six replicates) and extracted as described in sample extraction section. 2.2. Sensitivity and selectivity The lower limit of quantification (LLOQ) was determined as the concentration that had signal to noise ratio (S/N) of at least 10, accuracy between 80 and 120%, and coefficient of variation (CV) less than 20%. Acceptance criteria for QCs were accuracy between 85 and 115% and CV less than 15%. Selectivity assessed by inspecting the presence of noise or peaks in chromatograms represent blank OF samples injections (from 6 donors) compared with LLOQ sample chromatogram. 2.3. Stability Stability studies were performed by measuring TAC concentrations at QC1 and QC3, in three replicates. Freeze and thaw (after three freeze and thaw cycles), bench-top, auto-sampler (by reinjecting one of validation batch after it was left in the auto-sampler for 24 h and 48 h), and short-term stability up to one month were investigated.
3
2.4. Matrix effect and recovery The presence and potential influence of matrix effect (ME) in OF were studied using two different methods. First, chromatograms were obtained from the post-column infusion test. This test involved a continuous infusion of 98% methanol (which represents the composition of mobile phase at the elution time of ASC and TAC) containing 1 ng/mL of ASC and TAC at 20 L/min flow rate after the column through a three-way (Tee) connection. After establishing the baseline, a 10 L of blank extracted OF sample was injected by HPLC. The obtained chromatogram was then checked for signs of ion suppression or enhancement in comparison to blank injection of a neat solution (1:2, water: ACN). Second, the possible interference of OF components, namely the phospholipids, was studied. To do so, MRM transitions of abundant phospholipids were added to MS method to enable us to visually locate the elution region of these phospholipids. The influence of increasing ratio of precipitating solvents on the ME was also studied to select the ratio that offers the cleanest chromatography trace. Two different sets of QC1 and QC3 samples were prepared, in triplicate, either by (1) QCs samples (set 1) prepared by adding ACN to OF samples spiked with TAC as prescribed in sample extraction section (pre-extraction spiked samples); (2) QCs samples (set 2) prepared using a mixture of de-ionized water: ACN (neat solution). In each set, different ratios of ACN were added (1:1, 1:2 and 1:3). In total, 18 samples were analyzed, 9 samples in each set. The absolute ME was measured by calculating the percentage of the ratio of mean peaks area of pre-extracted samples to samples prepared in de-ionized water/ACN mixture. Recovery was assessed by analyzing the third set of QCs samples (set 3) prepared by extracting blank OF first with 1:1, 1:2 and 1:3 ACN, followed by adding TAC working standard solutions to achieve required concentrations (post-extraction spiked samples). The recovery was determined by calculating the ratio of mean peaks area of pre-extraction samples (set 1) to post-extracted spiked samples (set 3) and was expressed as a percentage.
3. Results and discussion Recommended TAC trough blood concentration (Cmin ) in kidney transplant recipients ranges between 15 and 20 ng/mL immediately after transplantation [3]. TAC dose tapered gradually and maintenance Cmin can be set as low as 5–7 ng/mL after 12-months of transplant operation [3]. Since only 1% of TAC amount found in unbound form was capable of reaching OF, the expected OF concentration would range from 50 to 200 pg/mL. Therefore, optimum mass spectrometry and chromatographic conditions were sought to develop a method with adequate selectivity and sensitivity. To achieve the highest feasible selectivity, different analytical columns were tested in conjunction with mobile phases of varying composition and gradients. An Acquity UPLC BEH C18 was the best choice as it gave sharp and symmetrical peaks for TAC and IS. Given the above-mentioned UPLC and mass spectrometry conditions, it was possible to set LLOQ at 10 pg/mL with signal/noise ratio more than 10 (Fig. 1A). No carryover was detected when a double blank OF sample was injected following the highest calibration concentration (Fig. 1B). The calibration curve was constructed by plotting nominal standard concentrations against peak area ratios of the analyte to IS and fitted with 1/x weighted least squares linear regression. The method demonstrated adequate accuracy and precision with QCs accuracy between 94.5-103.6%, and CV within 4–9.8% (Table 1). The coefficient of determination (R-squared) calculated for validation batches (n = 3) was between 0.998–0.999. Stability studies, namely, freeze and thaw, bench top, autosampler, and short-term storage at − 80 ◦ C for up to four weeks
Please cite this article in press as: M. Ghareeb, F. Akhlaghi, Development and validation of a sensitive and selective LC–MS/MS method for determination of tacrolimus in oral fluids, J. Chromatogr. B (2016), http://dx.doi.org/10.1016/j.jchromb.2016.10.008
G Model
ARTICLE IN PRESS
CHROMB-20290; No. of Pages 6
M. Ghareeb, F. Akhlaghi / J. Chromatogr. B xxx (2016) xxx–xxx
4
Fig. 1. Shows Chromatograms of TAC at LLOQ (10 pg/mL) (1A) and the internal standard ASC (400 pg/mL) (1D). Chromatograms (1B) and (1E) show traces of TAC; and (1C) and (1F) show traces of ASC in injections of pooled blank OF and blank solvent, respectively, injected immediately after highest concentration of the calibration curve (1600 pg/mL).
Table 1 Summary of accuracy and precision data obtained from quality control (QC) samples obtained from three individual runs (mean ±% CV, each QC had 6 replicates in each validation run, Total = 18).
Quality control samples (pg/mL) Accuracy (%) Precision (%)
LLOQ
QC1
QC2
QC3
10 103.6 9.8
30 103.0 6.1
200 94.5 5.9
1200 99.4 4.0
Accuracy = (mean concentration/nominal concentration) × 100. Precision expressed as % CV = (standard deviation/mean) × 100.
were conducted (Table 2). TAC was stable in extracted matrix for up to 48 h, and no loss of stability was noticed during different stability studies. Possible interference from endogenous substances in OF was investigated and is represented in chromatograms obtained from a pooled blank OF from six donors (Fig. 1B) and a blank or neat solution (66% ACN) (Fig. 1C) were compared. Using the assay conditions described above, no signs of interference was noticed. Using methanol instead of ACN as an organic solvent improved the sensitivity. Utilization of LC/MS grade methanol showed to increase the sensitivity by approximately 20%. Positive mode
ionization and monitoring ammonium adduct [M+NH4 ]+ at (m/z 821.30 → 768.35) provided better signal compared to [M]+ and [M+Na]+ . 3.1. Matrix effect and recovery Co-eluting of the analytes with endogenous substances in OF may lead to either ion suppression or enhancement, which is collectively called ME [29–31]. The presence of ME could compromise the reproducibility and may result in bias in the analysis [32]. Different cleaning procedures were used in other methods that aimed to measure the concentration of immunosuppressive agents in OF. These techniques included solid phase extraction [25,26], analyte pre-concentration by drying and reconstituting [33] and simple protein precipitation using organic solvents [27,34]. Type and percentage of the precipitating solution could influence sensitivity and selectivity through its effect on the yield of the analytes and cleanness of the extracted sample. Acetonitrile has been reported to provide satisfactory protein precipitation in oral fluid samples [31]. Recovery of a number of drugs and ME achieved using MeOH and ACN as the precipitating solvents in plasma are comparable. However, MeOH tends to retain about 40% more phospholipids [35].
Table 2 The result of stability studies (mean ±% CV, N = 3) indicates adequate bench top, freeze-thaw and short-term stability of tacrolimus in oral fluids. QC concentrations
Bench top
Freeze & thaw
Auto-sampler
Short-term
24 h
48 h
1 weeks
4 weeks
QC1 (30 pg/mL) %Accuracy (%CV) QC3 (1200 pg/mL) %Accuracy (%CV)
111.1 (9.4)
108.2 (5.8)
105.3 (9.6)
102.5 (4.3)
98.4 (11.9)
100.8 (6.7)
98.8 (2.4)
111.9 (2.0)
102.2 (2.7)
105.6 (3.6)
98.9 (3.4)
102.9 (1.9)
Please cite this article in press as: M. Ghareeb, F. Akhlaghi, Development and validation of a sensitive and selective LC–MS/MS method for determination of tacrolimus in oral fluids, J. Chromatogr. B (2016), http://dx.doi.org/10.1016/j.jchromb.2016.10.008
G Model
ARTICLE IN PRESS
CHROMB-20290; No. of Pages 6
M. Ghareeb, F. Akhlaghi / J. Chromatogr. B xxx (2016) xxx–xxx
5
Table 3 Effect of different ratios of oral fluid sample: extraction solvent (ACN) on recovery and absolute matrix effect, expressed as mean peak area ± SD (n = 3). Matrix
De-ionized water Extracted OF Post-extraction Recovery of Spiked OF (%)
QC1 (30 pg/mL)
QC2 (200 pg/mL)
QC3 (1200 pg/mL)
1:1
1:2
1:3
1:1
1:2
1:3
1:1
1:2
1:3
104 ± 8 83 ± 7 82 ±8 101.6
92 ± 4 106 ± 8 106 ± 6 100.0
85 ± 15 112 ± 14 96 ± 12 116.3
1592 ± 45 1485 ± 139 1317 ± 36 112.7
1739 ± 50 1689 ± 54 1484 ± 67 113.8
1719 ± 142 1775 ± 47 1559 ± 67 113.8
4260 ± 577 3916 ± 139 3760 ± 245 104.1
4653 ± 109 4452 ± 139 4090 ± 20 108.8
4409 ± 138 4452 ± 63 3579 ± 59 124.3
Fig. 3. Composite chromatogram shows traces of MRM transitions of 6 major phospholipids, obtained from a chromatogram of extracted blank pooled OF injection overlaid on TAC injection at concentration.
Fig. 2. Effect of blank OF and blank solvent injections on chromatograms obtained from continues post-column infused mixture of tacrolimus (TAC) and ascomycin (ASC) overlaid on TAC at QC2 concentration (400 pg/mL) (2A) and ASC (2B).
Therefore, ACN was chosen as the extracting solvent. Belostotsky, et al. [27] used 3:1, ACN:OF in a study to measure TAC salivary concentration. To the best of our knowledge, no other study has investigated an optimal proportion of extracting solvent (ACN) that gives maximum recovery and sample cleaning up. To examine the effect of using different proportions of ACN on recovery and absolute ME, OF samples were extracted with an equal, double and triple amount of ACN (Table 3). As it can be seen from Table 3, there was slightly less variability in areas count in samples extracted with the double volume of ACN compared to other two categories. Based on these values, it is obvious that samples extracted with ACN three times of their volume gave exaggerated recoveries, while the other two groups showed comparable recovery ranges. For ME, the first group showed to have significant ion suppression of about 20% in QC1 samples. Considering variability in the acquired data, adequate sample cleaning, adequate recovery, and minimum sample dilution, two-folds of ACN was chosen as the sample to solvent ratio to ensure optimal protein precipitation. Matrix effect was explored visually using post-column infusion technique [30,31]. The composite chromatograms in Fig. 2A and B were obtained by overlying chromatograms acquired from inject-
Fig. 4. Bland-Altman plot of % difference between the repeated measurements plotted against mean differences.
ing a neat solution (66% ACN), a blank OF with a continuous infusion of a mixture of ASC and TAC (1 ng/mL) and a chromatogram of QC2 injection. The only areas of chromatograms that showed ion suppression are between 0.2–0.5 min, which is far enough from the retention times of ASC and TAC. Finally, potential co-elution of phospholipids was examined by adding MRM of transitions of the most common phospholipids to the mass spectrometry method [30]. Phospholipid transitions included were (m/z 496 → 184, 520 → 184, 522 → 184, 524 → 184, 758 → 184, 782 → 184). In early stages of method development, we have observed that ASC and TAC peaks were co-eluting with low molecular weight phospholipids (m/z 496 to 524).The other two phospholipids that have m/z more than 700 were less problematic and eluted long after analytes of interest. By manipulating the mobile phase gradient or composition, the full separation between the analytes of interest and the phospholipids was achieved (Fig. 3). To our knowledge, the interference of phospholipids with LC–MS/MS analysis of drugs in OF and resolution of this issue has not been reported previously. This assay presents an excellent level of sensitivity to enable determination of TAC concentration in OF obtained from stable
Please cite this article in press as: M. Ghareeb, F. Akhlaghi, Development and validation of a sensitive and selective LC–MS/MS method for determination of tacrolimus in oral fluids, J. Chromatogr. B (2016), http://dx.doi.org/10.1016/j.jchromb.2016.10.008
G Model CHROMB-20290; No. of Pages 6
ARTICLE IN PRESS M. Ghareeb, F. Akhlaghi / J. Chromatogr. B xxx (2016) xxx–xxx
6
Fig. 5. Chromatograms represent TAC at concentration near LLOQ (11.7 pg/mL), (A) and corresponding internal standard (B) in an oral fluid sample obtained from a kidney transplant recipient at 10 h after a 2 mg tacrolimus dose. The estimated concentration after applying the dilution factor of 3 was 35 pg/mL.
transplant recipients. In total, out of 181 OF samples analyzed, only one sample had a concentration lower than LLOQ with a calculated concentration around 8.5 pg/mL. The concentration of TAC ranged from 11.7–2864.4 pg/mL and 1.7–46.1 ng/mL for OF and whole blood respectively. The clinical finding of this study will be presented in a separate manuscript. 3.2. Incurred sample reanalysis The incurred sample reanalysis test was performed by reanalyzing about 10% of the samples (19 samples) [36,37]. Whenever many samples were available per patient, two samples were selected to represent absorption and elimination phases. The differences between the paired measurements were normally distributed; therefore, the use of Bland-Altman method was justified [36]. Repeatability was tested visually (Fig. 4) and a statistically good agreement between the two repeated measurements was observed (Fig. 4) since all points lie between or near the 95% confidence interval limit. The 95% limit of agreement was from −19.16 to +31.98% and the bias (the mean difference between two occasions) was 6.40%. Fig. 5A depict a representative TAC chromatogram near the LLOQ concentration and Fig. 5B shows the chromatogram of IS for the same sample. 4. Conclusion We report a highly sensitive, selective, simple and robust method for quantitative determination of TAC in oral fluids. A simple sample preparation and extraction protocol was developed and optimized to provide adequate sensitivity, appropriate samples cleanliness, excellent recovery and minimal interference with endogenous components. In addition to lowest reported LLOQ of TAC, this work is the first to study the effect of different proportions of precipitating solvent (ACN) on the ME and recovery. In addition, this is the first report that investigated and described interference of phospholipids with LC–MS/MS determination of a small molecule compound in OF. Disclosure No conflict of interest is declared. This study was funded by the Rhode Island Science and Technology Advisory Council (STAC) and the University of Rhode Island, College of Pharmacy. References [1] L.J. Scott, K. McKeage, S.J. Keam, G.L. Plosker, Drugs 63 (2003) 1247. [2] L.C. Borra, J.I. Roodnat, J.A. Kal, R.A. Mathot, W. Weimar, T. van Gelder, Nephrol. Dial. Transplant. 25 (2010) 2757. [3] C.E. Staatz, S.E. Tett, Clin. Pharmacokinet. 43 (2004) 623.
[4] P.J. Taylor, C.H. Tai, M.E. Franklin, P.I. Pillans, Clin. Biochem. 44 (2011) 14. [5] F. Lemaitre, M. Antignac, C. Fernandez, Clin. Biochem. 46 (2013) 1538. [6] A. Capron, J. Lerut, D. Latinne, J. Rahier, V. Haufroid, P. Wallemacq, Transpl. Int. 25 (2012) 41. [7] W.J. Sandborn, G.M. Lawson, T.J. Cody, M.K. Porayko, J.E. Hay, G.J. Gores, J.L. Steers, R.A. Krom, R.H. Wiesner, Hepatology 21 (1995) 70. [8] H. Zahir, G. McCaughan, M. Gleeson, R.A. Nand, A.J. McLachlan, Ther. Drug Monit. 26 (2004) 506. [9] H. Zahir, G. McCaughan, M. Gleeson, R.A. Nand, A.J. McLachlan, Br. J. Clin. Pharmacol. 57 (2004) 298. [10] A. Capron, M. Mourad, M. De Meyer, L. De Pauw, D.C. Eddour, D. Latinne, L. Elens, V. Haufroid, P. Wallemacq, Pharmacogenomics 11 (2010) 703. [11] A. Capron, F. Musuamba, D. Latinne, M. Mourad, J. Lerut, V. Haufroid, P.E. Wallemacq, Ther. Drug Monit. 31 (2009) 178. [12] B.D. Noll, J.K. Coller, A.A. Somogyi, R.G. Morris, G.R. Russ, D.A. Hesselink, T. Van Gelder, B.C. Sallustio, Ther. Drug Monit. 35 (2013) 617. [13] L. Elens, A. Capron, V.V. Kerckhove, J. Lerut, M. Mourad, D. Lison, P. Wallemacq, V. Haufroid, Pharmacogenet. Genomics 17 (2007) 873. [14] A. Capron, J. Lerut, C. Verbaandert, J. Mathys, O. Ciccarelli, R. Vanbinst, F. Roggen, C. De Reyck, J. Lemaire, P.E. Wallemacq, Ther. Drug Monit. 29 (2007) 340. [15] M. Ghareeb, F. Akhlaghi, Bioanalysis 7 (2015) 1037. [16] K. Hoogtanders, J. van der Heijden, M. Christiaans, A. van de Plas, J. van Hooff, L. Stolk, Transplantation 83 (2007) 237. [17] K. Hoogtanders, J. van der Heijden, M. Christiaans, P. Edelbroek, J.P. van Hooff, L.M. Stolk, J. Pharm. Biomed. Anal. 44 (2007) 658. [18] C.Y. Cheung, J. van der Heijden, K. Hoogtanders, M. Christiaans, Y.L. Liu, Y.H. Chan, K.S. Choi, A. van de Plas, C.C. Shek, K.F. Chau, C.S. Li, J. van Hooff, L. Stolk, Transpl. Int. 21 (2008) 140. [19] D.R. Koop, L.A. Bleyle, M. Munar, G. Cherala, A. Al-Uzri, J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 926 (2013) 54. [20] E. Hinchliffe, J.E. Adaway, B.G. Keevil, J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 883–884 (2012) 102. [21] J.C. den Burger, A.J. Wilhelm, A. Chahbouni, R.M. Vos, A. Sinjewel, E.L. Swart, Anal. Bioanal. Chem. 404 (2012) 1803. [22] N.J. Webb, D. Roberts, R. Preziosi, B.G. Keevil, Pediatr. Transplant. 9 (2005) 729. [23] F. Akhlaghi, J. Ashley, A. Keogh, K. Brown, Ther. Drug Monit. 21 (1999) 8. [24] F. Akhlaghi, A.M. Keogh, K.F. Brown, Transplantation 67 (1999) 54. [25] A. Mendonza, R. Gohh, F. Akhlaghi, Ther. Drug Monit. 26 (2004) 569. [26] A.E. Mendonza, R.Y. Gohh, F. Akhlaghi, Ther. Drug Monit. 28 (2006) 402. [27] V. Belostotsky, J. Adaway, B.G. Keevil, D.R. Cohen, N.J. Webb, Pediatr. Nephrol. 26 (2011) 133. [28] US Department of Health and Human Services. Food and Drug Administration (FDA). Center for Drug Evaluation and Research (CDER). 2001. http://www. fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/ Guidances/ucm070107.pdf (last accessed: Oct 2016). [29] H. Mei, Y. Hsieh, C. Nardo, X. Xu, S. Wang, K. Ng, W.A. Korfmacher, Rapid Commun. Mass Spectrom. 17 (2003) 97. [30] G. Zhang, C.E. Wujcik, J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 877 (2009) 2003. [31] R. Dams, M.A. Huestis, W.E. Lambert, C.M. Murphy, J. Am. Soc. Mass Spectrom. 14 (2003) 1290. [32] B.K. Matuszewski, M.L. Constanzer, C.M. Chavez-Eng, Anal. Chem. 75 (2003) 3019. [33] M.H. Wiesen, F. Farowski, M. Feldkotter, B. Hoppe, C. Muller, J. Chromatogr. A 1241 (2012) 52. [34] B. Shen, S. Li, Y. Zhang, X. Yuan, Y. Fan, Z. Liu, Q. Hu, C. Yu, J. Pharm. Biomed. Anal. 50 (2009) 515. [35] E. Chambers, D.M. Wagrowski-Diehl, Z. Lu, J.R. Mazzeo, J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 852 (2007) 22. [36] J.M. Bland, D.G. Altman, Lancet 1 (1986) 307. [37] P. van Amsterdam, A. Companjen, M. Brudny-Kloeppel, M. Golob, S. Luedtke, P. Timmerman, Bioanalysis 5 (2013) 645.
Please cite this article in press as: M. Ghareeb, F. Akhlaghi, Development and validation of a sensitive and selective LC–MS/MS method for determination of tacrolimus in oral fluids, J. Chromatogr. B (2016), http://dx.doi.org/10.1016/j.jchromb.2016.10.008