Journal of Chromatography B, 991 (2015) 85–91
Contents lists available at ScienceDirect
Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb
A UHPLC–MS/MS bioanalytical assay for the determination of BMS-911543, a JAK2 inhibitor, in human plasma Jane Liu, Long Yuan ∗ , Guowen Liu, Jim X. Shen, Anne-Franc¸oise Aubry, Mark E. Arnold, Qin C. Ji Analytical and Bioanalytical Development, Bristol-Myers Squibb, Princeton NJ 08543, United States
a r t i c l e
i n f o
Article history: Received 30 January 2015 Accepted 8 April 2015 Available online 17 April 2015 Keywords: Liquid–liquid extraction LC–MS/MS Quantitation High-throughput Automation 96-Well format
a b s t r a c t Herein we report a rapid, accurate and robust UHPLC–MS/MS assay for the quantitation of BMS-911453, a Janus kinase 2 inhibitor under clinical development for the treatment of myeloproliferative disorders, in human plasma. A systematic method development approach was used to optimize the mass spectrometry, chromatography, and sample extraction conditions, and to minimize potential bioanalytical risks. The validated method utilizes stable-isotope labeled 13 C4 -BMS-911543 as the internal standard. Liquid-liquid extraction was used for sample preparation. Chromatographic separation was achieved within 2 min on a Zorbax Extend-C18 column with an isocratic elution. BMS-911543 and its internal standard were detected by positive ion electrospray tandem mass spectrometry. The assay range was from 1 to 500 ng/mL, and the standard curve was fitted with 1/x2 weighted linear regression. The intraassay precision was within 5.0% CV and the inter-assay precision was within 2.6% CV. The inter-assay mean accuracy, expressed as percents of theoretical, was between 99.8% and 102.3%. The assay has high recovery (∼80%) and minimal matrix effect (0.95–1.00). BMS-911543 was stable in human plasma for at least 24 h at room temperature, 90 days at −20 ◦ C, and following three freeze–thaw cycles. The validated method was successfully applied to sample analysis in clinical studies. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Janus kinase 2 (commonly called JAK2) is a non-receptor tyrosine kinase [1]. It is a member of the Janus kinase family and has been implicated in signaling by members of the type II cytokine receptor family (e.g., interferon receptors) [2]. JAK2 is a promising target for therapeutic intervention in myeloproliferative disorders (MPDs) [1] and several JAK inhibitors are in pre-clinical or clinical development for MPDs [3]. BMS-911543, as shown in Fig. 1A, is a potent and selective JAK2 inhibitor under clinical development for the treatment of MPDs [1]. BMS-911543 showed a high degree of functional selectivity toward JAK2 in a variety of pre-clinical models, which supported the therapeutic rationale for its clinical development [1]. In a multicenter phase 1/2a study in patients with myelofibrosis, BMS-911543 showed an acceptable safety and efficacy profile, and was generally well tolerated [4].
∗ Corresponding author. Tel.: +1 6092525336. E-mail address: [email protected] (L. Yuan). http://dx.doi.org/10.1016/j.jchromb.2015.04.013 1570-0232/© 2015 Elsevier B.V. All rights reserved.
Evaluations of BMS-911543 as a drug candidate requires an accurate, sensitive and rugged LC–MS/MS bioanalytical assay to support both preclinical and clinical studies. Ultra-high performance liquid chromatography coupled with tandem mass spectrometry (UHPLC–MS/MS) technology has been widely used for the bioanalytical assays to support dug development [5,6]. UHPLC uses sub-2 m particles stationary phases, therefore, has better resolving power than conventional LC and can usually achieve faster analysis and higher throughput with good chromatographic performance. In this manuscript, we report the method development and validation of a high-throughput UHPLC–MS/MS method for the quantification of BMS-911543 in human plasma. A systematic method development approach was applied to achieve optimized mass spectrometry, chromatography, and sample extraction conditions that used several previously published strategies to accelerate method development for high-throughput bioanalytical assays [5] and to ensure quality of the developed method [6]. Incurred samples were also used in method development to minimize bioanalytical risks from potential metabolite interference and ensure the quality of the method. Stable isotope labeled BMS-911543 (13 C4 -BMS-911543, Fig. 1B) was used as the internal standard. The assay was validated and successfully used to support clinical studies with excellent assay performance.
86
J. Liu et al. / J. Chromatogr. B 991 (2015) 85–91
samples were prepared using the same method at six different concentration levels: LLOQ QC (1.0 ng/mL), low QC (3.0 ng/mL), geometric mean QC (GM QC, 30 ng/mL), mid QC (250 ng/mL), high QC (400 ng/mL), and dilution QC (20,000 ng/mL). The QCs were aliquoted into polypropylene tubes and stored frozen at approximately −20 ◦ C. The internal standard stock standard solution was prepared by dissolving appropriate amounts of stable labeled standard in 1:1 (v/v) acetonitrile/DMSO to give a final concentration of 1.00 mg/mL of 13 C4 -BMS-911543. The internal standard working solution of 100 ng/mL of 13 C4 -BMS-911543 was prepared by adding 200 L of the internal standard stock solution to a 200-mL glass volumetric flask, then diluting to the volume with acetonitrile/water (70/30, v/v). 2.4. Sample extraction
Fig. 1. Chemical structures of BMS-911543 (A) and its internal standard, 13 C4 -BMS911543 (B).
2. Experimental 2.1. Chemicals and reagents HPLC-grade acetonitrile, methanol, concentrated ammonium hydroxide, ammonium acetate, methyl-tert butyl ether (MTBE), and ACS grade glacial acetic acid were purchased from Fisher (Pittsburgh, PA, USA). Formic acid (96%) was purchased from Sigma–Aldrich (St. Louis, MO, USA). The reference standards of BMS-911543 and its internal standard, 13 C4 -BMS-911543, were produced in Bristol-Myers Squibb Co. (BMS). Normal human plasma, with K2 EDTA as an anticoagulant, was purchased from Bioreclamation (Hicksville, NY, USA).
Samples were thawed at room temperature, and vortexed to ensure homogeneity. Each plasma sample tube was uncapped and 100 L of sample was transferred into the appropriate well of a 96well plate. Then, 50 L of the internal standard working solution (100 ng/mL of 13 C4 -BMS-911543) and 50 L of 1.0 M ammonium acetate in water solution were added to each sample, followed by a brief vortex to ensure mixing. Extraction solvent (MTBE, 600 L) was added to each sample and shaken at high speed for 20 min. After centrifugation to separate the phases, 500 L of the organic phase layer was transferred into a 96-well collection plate. The samples were dried with a Speed-Dry evaporator at 40 ◦ C under a nitrogen stream and reconstituted with 400 L of the reconstitution solution of acetonitrile/water (25/75, v/v). A volume of 5 L of the reconstituted samples was injected into the LC–MS/MS system for analysis. 2.5. Chromatographic conditions The analytical column used was a UHPLC column: Zorbax Extend-C18, 2.1 mm × 50 mm, 1.8 m particle size, Agilent (Santa Clara, CA, USA). An isocratic elution (60/40, mobile phase A/mobile phase B) was utilized and the analytical column was kept at 40 ◦ C during the analysis. A flow rate of 800 L/min was used. The mobile phase A consisted of 10 mM ammonium bicarbonate with 0.10% ammonium hydroxide in water and the mobile phase B consisted of acetonitrile.
2.2. Instrumentation
2.6. MS/MS detection
The UHPLC system consisted of a Flux pump and a SIL-HTC autosampler (LEAP Technologies, Carrboro, NC, USA). The API 4000 mass spectrometer controlled by AnalystTM software was from Applied Biosystems/MDS Sciex (Concord, ON, Canada). AnalystTM version 1.4.2 was used as the data acquisition software. The column heater HotDog 5090 was from Prolab Instruments (Reinach, AG, Switzerland). A JANUS Mini liquid handling system (Perkin Elmer, Downers Grove, IL, USA) was used for the automated liquid transfer in sample extraction (96-well plate format).
LC–MS/MS detection was performed using a Sciex API 4000 triple quadrupole mass spectrometer with a Turbo Ionspray® ionization source operated in the positive ion mode. The ion spray voltage was 4500 V and the source temperature was 550 ◦ C. The CAD gas setting was 6 and curtain gas setting was 50. The ion source gas 1 (GS1) and ion source gas 2 (GS2) settings were 50 and 60, respectively. The de-cluster potential was 121 V. The collision energy (CE) was 49 eV and the collision cell exit potential (CXP) was 24 V. The selected reaction monitoring (SRM) monitored was m/z 433 → 281 for BMS-911543 and m/z 437 → 285 for 13 C4 BMS-911543. The dwell time for each transition monitored was 150 ms.
2.3. Preparation of standard and QC samples Standard and QC stock solutions of BMS-911543 (1.00 mg/mL) were prepared from separate weighings by dissolving the reference compound into 1:1 (v/v) acetonitrile/DMSO. The intermediate standard working solution containing 20,000 ng/mL of BMS911543 was prepared by dilution of the stock standard solution with human plasma. The calibration curves with eight levels (1.00, 2.50, 5.00, 10.0, 50.0, 100, 250 and 500 ng/mL BMS-911543) were prepared fresh daily by diluting appropriate volumes of intermediate standard working solutions with human plasma. The QC
2.7. Quantitation The peak areas of both analyte and its internal standard were determined using AnalystTM software version 1.4.2. For each analytical batch, a calibration curve was derived from the peak area ratios (analyte: internal standard) using weighted linear least squares regression of the area ratio vs. the concentration of the standards. A weighting of 1/x2 (where x is the concentration of a
J. Liu et al. / J. Chromatogr. B 991 (2015) 85–91
87
given standard) was used for curve fitting. The regression equation was used to back-calculate the measured concentration for each standard and QC. The results were compared to the theoretical concentration to obtain the accuracy, expressed as a percentage of the theoretical value, for each standard and QC measured. 3. Results and discussion 3.1. Method development A systematic method screening and optimization approach [6–9] was used to optimize the mass spectrometry, chromatography, and sample extraction conditions for the LC–MS/MS assay. Briefly, four aqueous mobile phases from acidic to basic and a total of 11 different UHPLC columns (C18, modified C18, phenyl and other types from several different manufacturers) were screened to achieve the best chromatographic separation and sensitivity. Liquid–liquid extraction (LLE) methods using different extraction buffers and solvents were evaluated for the optimized extraction recovery and cleanness of the extracts. We also used rat and dog incurred samples in method development to minimize the risk of potential interferences from metabolites, which ensured the quality and robustness of the developed method. 3.1.1. Mass spectrometry, chromatography, and sample extraction MS scan and product ion scan mass spectra were obtained by the infusion of BMS-911543 and its internal standard, 13 C4 -BMS911543, solutions via a tee connection between the LC column and mass spectrometer inlet. As shown in Fig. 2A and C, the protonated ions of m/z 433 and 437 are the predominant ion of BMS-911543 and its internal standard under positive electrospray ionization, respectively. The most abundant product ions of BMS-911543 and its internal standard are at m/z 281 and 285, respectively (Fig. 2B and D). Therefore, the SRM transitions m/z 433 → 281 and 437 → 285 were selected for the monitoring of BMS-911543 and its internal standard. A total of 11 different UHPLC columns (see Fig. 3 for detailed column information) were screened. As shown in Fig. 3, both
Fig. 3. LC–MS/MS chromatograms of BMS-911543 using different UHPLC columns (1: Agilent SB-C18, 2: Waters Aquity BEH shield RP18, 3: Agilent Zorbax ExtendC18, 4: Waters Aquity BEH phenyl, 5: Phenomenex Kinetex C18, 6: Agilent Eclipse PAH, 7: Agilent Eclipse plus C18, 8: Agilent XDB C18, 9: Thermo Gold aq, 10: Waters Aquity BEH C18, and 11: Waters Aquity HSS T3, all columns are 2.1 mm × 50 mm with sub-2 m particle size). The gradient method was as follows: 0–1.5 min 20–80% B, 1.5–2.0 min 80% B, 2.0–2.5 min 20% B. The flow rate was 0.80 mL/min.
Zorbax Extend-C18 column (column 3) and Phenomenex Kinetex C18 column (column 5) offered a good combination of peak shape and sensitivity. The Zorbax Extend-C18 column provided a better sensitivity than the Phenomenex Kinetex C18 column and, therefore, was selected for further optimization. Four different aqueous mobile phases from acidic to basic (0.1% formic acid in water, 10 mM ammonium acetate with 0.01% acetic acid in water, 10 mM ammonium formate with 0.05% formic acid in water, and 10 mM ammonium bicarbonate with 0.1% ammonium hydroxide in
Fig. 2. Electrospray positive ion MS scan ion spectra and product ion spectra of BMS-911543 (A and B) and 13 C4 -BMS-911543 (C and D).
88
J. Liu et al. / J. Chromatogr. B 991 (2015) 85–91
Fig. 4. LC–MS/MS chromatograms of BMS-911543 using different mobile phases (mobile phase A: I: 0.1% formic acid in water, II: 10 mM ammonium acetate with 0.01% acetic acid in water, III: 10 mM ammonium formate with 0.05% formic acid in water, and IV: 10 mM ammonium bicarbonate with 0.1% ammonium hydroxide in water; mobile phase B: acetonitrile). The column used was a Zorbax Extend-C18 column. The gradient method was as follows: 0–1.5 min 20–80% B, 1.5–2.0 min 80% B, 2.0–2.5 min 20% B. The flow rate was 0.80 mL/min.
water) were screened. Although similar peak shape was obtained under these different mobile phases, much better sensitivity was achieved using the basic mobile phase, 10 mM ammonium bicarbonate with 0.1% ammonium hydroxide in water (Fig. 4). After further optimization, quick and efficient chromatographic separation of BMS-911543 was obtained using an isocratic elution with 60% of 10 mM ammonium bicarbonate with 0.1% ammonium hydroxide in water and 40% of acetonitrile on a Zorbax Extend-C18 column. The retention time is approximately 0.7 min for BMS911543 and its internal standard. The total run time is only 1.5 min, which ensured the high-throughput of the assay. Liquid–liquid extraction has been routinely used in our laboratory for LC–MS/MS bioanalytical assays to support preclinical and clinical studies. MTBE as the extraction solvent gave a good and consistent extraction recovery for BMS-911543 in plasma. Matrix effect is the suppression or enhancement of the analyte ionization from co-eluting matrix constituents [10]. The abundant phospholipids in plasma samples are one major source of matrix effect, and may affect the accuracy, precision and sensitivity of LC–MS/MS bioanalytical assays [6]. We evaluated the phospholipids profiles of the extracted plasma sample using positive precursor ion scan of m/z 184 and negative precursor ion scan of m/z 153 and a longer 10-min gradient [11]. As shown in Fig. 5, the BMS-911543 peak was well separated from the major phospholipids peaks, indicating minimum risks of matrix effect from phospholipids. Consistent column performance has been maintained for injection of multiple sample batches throughout the sample analysis of a clinical study. In addition, a short column wash with high organic solvent could be added to clean up the column and further extend the column lifetime for continuous injection of a large number samples. 3.1.2. Evaluation of potential interference from metabolites To avoid potential interference from BMS-911543 metabolites, the SRM channels of all known and predicted metabolites (see Table 1) were monitored using a mixture of pooled rat and dog incurred samples. To minimize the risk of missing a metabolite peak eluting close to the analyte peak, the experiments were also done using a much longer (25 min) gradient. The SRM chromatograms of BMS-911543 and all detectable metabolites in the pooled incurred
Fig. 5. Chromatograms of BMS-911543 and phospholipids in a pooled rat and dog incurred sample. The column used was a Zorbax Extend-C18 column. The gradient method was as follows: 0–1.5 min 25–30% B, 1.5–2 min 30–40% B, 2–8 min 40–90% B, 8–9 min 90% B, 9–10 min 25% B. Table 1 SRM transitions used in the assessment of metabolite interference.
BMS-911543 (P) P+O P-CH2 + O P-CH2 P + 2O P + O + glucuronide P-CH2 + glucuronide P-CH2 + O + glucuronide P-CH2 + GSH
[MH]+
Product ions
433.2 449.2 435.2 419.2 465.2 625.2 595.3 611.2 724.2
281.0 281.0, 297.0 283.0, 267.0 267.0 313.0 449.2, 297.2 443.1, 281.0 435.2, 283.2 433.2
sample are shown in Fig. 6. No metabolite peak interfering with BMS-911543 was observed using either the short or long gradient. In addition, the same experiments were also done using a column with a different stationary phase (Aquity BEH phenyl) and a mobile phase at a different pH (acidic: 10 mM ammonium formate with 0.05% formic acid in water), and no interfering metabolite peak was observed in either case (data not shown). The evaluations using incurred samples ensured the quality of the method and minimized the risk of re-doing the method development due to unexpected metabolite interference. Unlike animal incurred samples that can be obtained from discovery studies, it is impractical to get human incurred sample for method development before the start of human clinical studies. In most cases, humans have similar metabolic pathways to animal species. For BMS-911543, there was no unique human metabolite identified in in vitro studies. Thus, the pooled animal incurred sample can well represent the human incurred samples and help to ensure the quality of the human assay. 3.2. Assay validation Validation experiments were designed with reference to the Guidance for Industry-Bioanalytical Method Validation by the Food and Drug Administration (FDA) of the United States [12] and the Guideline on Bioanalytical Method Validation by the European Medicines Agency (EMA) [13]. Standard curve linearity, lower limit
J. Liu et al. / J. Chromatogr. B 991 (2015) 85–91
89
Fig. 6. SRM chromatograms of BMS-911543 and its metabolites in a pooled rat and dog incurred sample. The chromatographic separation was with a Zorbax Extend-C18 column using a 2.5 min gradient (left), or a 25 min gradient (right). The 2.5 min gradient method was as follows: 0–1.5 min 25–30% B, 1.5–2.0 min 30–40% B, 2.0–2.5 min 25% B. The 25 min gradient method was as follows: 0–18.5 min 5–30% B, 18.5–22 min 30–90% B, 22–24 min 90% B, 24–25 min 5% B. The flow rate was 0.80 mL/min.
of quantitation (LLOQ), accuracy and precision, selectivity, recovery, matrix effect, and stability were evaluated. The experimental design and results of method validation are presented in the following sections. 3.2.1. Standard curve linearity Assay linearity was evaluated using eight calibration standards analyzed in duplicate over the nominal concentration range of 1.00 ng/mL and 500.00 ng/mL for BMS-911543. For a total of eight validation runs, a linear 1/x2 weighted regression model provided the best fit for BMS-911543 in human plasma with the coefficient of determination (r2 ) between 0.9973 and 0.9997, and therefore was selected for curve fitting. The mean back-calculated concentrations of the standards were between 90.0% and 107.0% of the theoretical concentrations. 3.2.2. Accuracy, precision, and dilution The accuracy and precision data of BMS-911543 in human plasma are summarized in Table 2. Based on four levels of analytical QCs (low, GM, mid and high), the intra-assay precision, expressed as coefficient of variation (CV), was within 5.0%, and the interassay precision was within 2.6%. The accuracy, expressed as %Dev, was within ±2.3% Dev of the nominal concentration. The results
demonstrated that the method was accurate and precise for the analysis of BMS-911543 in human plasma. The suitability of study samples being diluted with drug-free plasma on the day of assay (without undergoing an additional freeze–thaw cycle) was also evaluated as part of the validation. A QC level (dilution QC), used specifically for dilutions, was prepared with the concentrations of analyte at 20,000 ng/mL. To achieve a 100-fold dilution, 10 L of dilution QC and 990 L of blank matrix were combined, and 100 L of resulting sample was extracted. As shown in Table 2, the accuracy and precision for dilution QCs were all well below the 15% acceptance criteria, demonstrating the reliability of the dilution. 3.2.3. LLOQ and selectivity The LLOQ for the analyte was assessed using human K2EDTA plasma samples at 1.0 ng/mL of BMS-911543, the lowest concentrations for the analyte in the standard curve. Six different lots of control human K2EDTA plasma were spiked to obtain the six LLOQ samples. The deviations of the measured concentrations from the nominal LLOQ values were within ±17.0% for all the six lots (acceptance criteria is to achieve within ±20.0% for at least five of the six lots). Six different lots of human plasma were analyzed with and without internal standard to determine whether any endogenous constituents interfered with the analyte or the internal standard.
Table 2 QC accuracy and precision for BMS-911543 in human plasma in validation. QC type (nominal conc. in ng/mL)
LLOQ (1.00)
Low (3.00)
GM (30.0)
Mid (250)
High (400)
Dilution (20,000)
Mean observed conc. %Dev Between run precision (%CV) Within run precision (%CV) Total variation (%CV) n Number of runs
0.99 −1.0 4.2 6.2 7.5 30 5
3.07 2.3 0.0 5.0 4.9 38 8
29.93 −0.2 2.6 1.3 2.9 38 8
248.03 −0.8 2.4 1.4 2.8 38 8
394.44 −1.4 1.8 1.4 2.3 38 8
19030.74 −4.8 4.5 2.0 4.9 30 5
GM, geometric mean.
90
J. Liu et al. / J. Chromatogr. B 991 (2015) 85–91 Table 4 QC accuracy and precision for BMS-911543 in human plasma for a clinical study. QC type (nominal conc. in ng/mL) Mean observed conc. %Dev Between run precision (%CV) Within run precision (%CV) Total variation (%CV) n Number of runs
Low (3.00)
GM (30.0)
Mid (250)
High (400)
30.36
248.29
391.22
−0.2 5.9
1.2 5.0
−0.7 4.1
−2.2 4.1
4.1
3.6
3.5
4.6
7.2
6.2
5.4
6.2
2.99
125 37
125 37
125 37
125 37
GM, geometric mean.
Fig. 7. Representative SRM chromatograms of BMS-911543 in blank plasma (A), LLOQ of 1.0 ng/mL (B), and a patient sample (C).
No significant interfering peaks from the plasma were found at the retention time of either the analyte or its internal standard, which demonstrated the assay has good specificity. Representative LC–MS/MS chromatograms of a blank plasma and a LLOQ sample are shown in Fig. 7. 3.2.4. Matrix effect The matrix effect was determined at concentrations of 3.0 and 400 ng/mL for BMS-911543 by dividing the analyte response in human plasma spiked post-extraction with BMS-911543 by the analyte response (peak area) of those spiked in reconstitution solution. The matrix effect of the internal standard was determined similarly. Matrix effect of 1 indicates ion enhancement. The matrix effects were 0.95–1.00 for BMS-911543, and 1.00–1.01 for the internal standard. This demonstrates that there was no or minimal matrix effect on the measurement of the analyte. 3.2.5. Extraction recovery The recovery of the analyte from human plasma during extraction was determined at 3.0 and 400 ng/mL by comparing the response ratios in human plasma samples spiked with the analyte prior to extraction with those spiked post-extraction. The recovery of the internal standard was determined similarly at 100 ng/mL. The recovery of BMS-911543 and its internal standard was 79.3–79.7% and 78.9–83.5%, respectively. 3.2.6. Stability The room temperature, freeze–thaw, and frozen storage stability of BMS-911543 in human plasma were evaluated in triplicate using QCs. The reinjection integrity was assessed by re-injecting an entire run. The processed sample stability was assessed by testing processed and stored QCs against freshly prepared standards. Stability of processed sample in reconstitution solution was investigated in order to demonstrate the stability of the extracted analyte
before and during the injection of an analytical run. To establish the stability of the analyte, the deviations of the mean measured concentrations of the test samples have to be within ±15% of the nominal concentrations. As shown in Table 3, at least 24 h room temperature, 3 cycles freeze–thaw, 90 day frozen storage stability at −20 ◦ C, 5 day processed sample stability at 5 ◦ C, and 48 h reinjection integrity at 5 ◦ C were established. To ensure the stability of the analyte during human blood sample collection and processing, whole blood stability of BMS-911543 was also evaluated. Whole blood stability was evaluated by adding BMS-911543 to fresh human blood at two concentration levels. Plasma was derived immediately after the addition of the compounds (zero time), and after being stored for 2 h on ice and at room temperature. Blood stability was based on the deviation of the test time point value from the zero (0) time value. To establish the stability of the analyte, the deviations of the mean measured concentrations of the test samples have to be within ±15% of the zero time value. BMS-911543 was determined to be stable for at least 2 h at RT or 4 ◦ C (on ice) in human blood. Stock solution stability of BMS-911543 and its internal standard in acetonitrile/DMSO was also evaluated at both ambient temperatures and at 2–8 ◦ C as part of the validation. Results for the determination of stock solution stability were calculated by comparing mean response ratios (area of response per unit of concentration) of stability solutions to mean response ratios of freshly prepared control solutions. To establish the stock solution stability of the analyte, the means of the responses of the two solutions have to be within 5.0% of each other. BMS-911543 stock solutions were determined to be stable for at least 24 h at room temperature and 113 days at 2–8 ◦ C. 3.2.7. Assay application for clinical studies The validated assay was successfully used for the bioanalytical support of clinical studies in patients. As shown in Table 4, for one supported clinical study, excellent assay performance was achieved throughout the study. For a total of 37 runs, the accuracy was within ±2.2% Dev of the nominal concentration, the intraassay precision was within 4.6%, and the inter-assay precision was
Table 3 Stability of BMS-911543 in human plasma. QC type (nominal conc.)
Low QC 3.00 ng/mL
Sample condition
Mean conc.
%Dev
Mean conc.
%Dev
Mean conc.
%Dev
24 h at RT After 3rd freeze–thaw cycle 90 Day at −20 ◦ C 5 Day at 5 ◦ C Processed samples 48 h at 5 ◦ C re-injection integrity
2.99 3.01 3.08 3.08 3.00
−0.3 0.3 2.7 2.7 0.0
390.01 384.46 404.79 392.95 388.01
−2.5 3.9 1.2 −1.8 −3.0
18884.03 19363.98 20439.25 18725.02 18398.52
−5.6 −3.2 2.2 −6.4 −8.0
High QC 400 ng/mL
Dilution QC 20,000 ng/mL
J. Liu et al. / J. Chromatogr. B 991 (2015) 85–91
91
References
Fig. 8. Plasma concentration vs. time curve of a patient after a daily oral dosing of 80 mg of BMS-911543.
within 5.9%. Incurred sample reanalysis (ISR) was performed with 179 out of 1630 study samples, and 96% and 81% of the ISR samples were within ±10% and ±5% from the mean of initial and repeat test results, respectively, demonstrating the excellent reproducibility of the assay. A representative concentration vs. time profile of BMS-911543 in a patient from the clinical study is shown in Fig. 8. 4. Conclusions Here, we reported the method development and validation of a UHPLC–MS/MS bioanalytical assay for BMS-911543 in human plasma over the concentration range of 1.0–500 ng/mL. Rapid, accurate and rugged analysis of BMS-911543 was achieved using liquid–liquid extraction and an isocratic elution. The use of incurred samples during method development minimized potential bioanalytical risks due to the interferences from metabolites and ensured the quality of the method. The assay was successfully used for supporting clinical studies. Acknowledgement Authors would like to take this opportunity to thank Kai Cao from Bristol-Myers Squibb Radiosynthesis team for providing the stable isotope labeled internal standard.
[1] A.V. Purandare, T.M. McDevitt, H. Wan, D. You, B. Penhallow, X. Han, R. Vuppugalla, Y. Zhang, S.U. Ruepp, G.L. Trainor, L. Lombardo, D. Pedicord, M.M. Gottardis, P. Ross-Macdonald, H. de Silva, J. Hosbach, S.L. Emanuel, Y. Blat, E. Fitzpatrick, T.L. Taylor, K.W. McIntyre, E. Michaud, C. Mulligan, F.Y. Lee, A. Woolfson, T.L. Lasho, A. Pardanani, A. Tefferi, M.V. Lorenzi, Characterization of BMS-911543, a functionally selective small-molecule inhibitor of JAK2, Leukemia 26 (2012) 280–288. [2] P.J. Murray, The JAK-STAT signaling pathway: input and output integration, J. Immunol. 178 (2007) 2623–2629. [3] C. Kumar, A.V. Purandare, F.Y. Lee, M.V. Lorenzi, Kinase drug discovery approaches in chronic myeloproliferative disorders, Oncogene 28 (2009) 2305–2313. [4] A. Pardanani, A.W. Roberts, J.F. Seymour, K. Burbury, S. Verstovsek, H.M. Kantarjian, K. Begna, H. Yoshitsugu, T.A. Gestone, P. Phillips, G. Xing, G. Peltz, M.V. Lorenzi, L. Alland, A. Woolfson, A. Tefferi, BMS-911543, A Selective JAK2 Inhibitor: A Multicenter Phase 1/2a Study In Myelofibrosis, 55th American Society of Hematology Annual Meeting and Exposition. New Orleans, LA., 2013. [5] G. Liu, H.M. Snapp, Q.C. Ji, M.E. Arnold, Strategy of accelerated method development for high-throughput bioanalytical assays using ultra high-performance liquid chromatography coupled with mass spectrometry, Anal. Chem. 81 (2009) 9225–9232. [6] M. Jemal, Z. Ouyang, Y.Q. Xia, Systematic LC–MS/MS bioanalytical method development that incorporates plasma phospholipids risk avoidance, usage of incurred sample and well thought-out chromatography, Biomed. Chromatogr. 24 (2010) 2–19. [7] L. Yuan, H. Jiang, Z. Ouyang, Y.-Q. Xia, J. Zeng, Q. Peng, R.W. Lange, Y. Deng, M.E. Arnold, A.-F. Aubry, A rugged and accurate liquid chromatography–tandem mass spectrometry method for the determination of asunaprevir, an NS3 protease inhibitor, in plasma, J. Chromatogr. B 921–922 (2013) 81–86. [8] L. Yuan, Y. Fu, D. Zhang, Y.-Q. Xia, Q. Peng, A.-F. Aubry, M.E. Arnold, Use of a carboxylesterase inhibitor of phenylmethanesulfonyl fluoride to stabilize epothilone D in rat plasma for a validated UHPLC–MS/MS assay, J. Chromatogr. B 969 (2014) 60–68. [9] L. Yuan, H. Jiang, N. Zheng, Y.-Q. Xia, Z. Ouyang, J. Zeng, B. Akinsanya, J.L. Valentine, J.D. Moehlenkamp, Y. Deng, A.-F. Aubry, M.E. Arnold, A validated LC–MS/MS method for the simultaneous determination of BMS-791325, a hepatitis C virus NS5B RNA polymerase inhibitor, and its metabolite in plasma, J. Chromatogr. B 973 (2014) 1–8. [10] B.K. Matuszewski, M.L. Constanzer, C.M. Chavez-Eng, Strategies for the assessment of matrix effect in quantitative bioanalytical methods based on HPLC–MS/MS, Anal. Chem. 75 (2003) 3019–3030. [11] Y.Q. Xia, M. Jemal, Phospholipids in liquid chromatography/mass spectrometry bioanalysis: comparison of three tandem mass spectrometric techniques for monitoring plasma phospholipids, the effect of mobile phase composition on phospholipids elution and the association of phospholipids with matrix effects, Rapid Commun. Mass Spectrom. 23 (2009) 2125–2138. [12] FDA, Guidance for Industry: Bioanalytical Method Validation, 2001 http:// www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/ Guidances/ucm070107.pdf [13] European Medicines Agency, Guideline on Bioanalytical Method Validation, http://www.ema.europa.eu/docs/en GB/document library/Scientific 2011 guideline/2011/08/WC500109686.pdf
Contents lists available at ScienceDirect
Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb
A UHPLC–MS/MS bioanalytical assay for the determination of BMS-911543, a JAK2 inhibitor, in human plasma Jane Liu, Long Yuan ∗ , Guowen Liu, Jim X. Shen, Anne-Franc¸oise Aubry, Mark E. Arnold, Qin C. Ji Analytical and Bioanalytical Development, Bristol-Myers Squibb, Princeton NJ 08543, United States
a r t i c l e
i n f o
Article history: Received 30 January 2015 Accepted 8 April 2015 Available online 17 April 2015 Keywords: Liquid–liquid extraction LC–MS/MS Quantitation High-throughput Automation 96-Well format
a b s t r a c t Herein we report a rapid, accurate and robust UHPLC–MS/MS assay for the quantitation of BMS-911453, a Janus kinase 2 inhibitor under clinical development for the treatment of myeloproliferative disorders, in human plasma. A systematic method development approach was used to optimize the mass spectrometry, chromatography, and sample extraction conditions, and to minimize potential bioanalytical risks. The validated method utilizes stable-isotope labeled 13 C4 -BMS-911543 as the internal standard. Liquid-liquid extraction was used for sample preparation. Chromatographic separation was achieved within 2 min on a Zorbax Extend-C18 column with an isocratic elution. BMS-911543 and its internal standard were detected by positive ion electrospray tandem mass spectrometry. The assay range was from 1 to 500 ng/mL, and the standard curve was fitted with 1/x2 weighted linear regression. The intraassay precision was within 5.0% CV and the inter-assay precision was within 2.6% CV. The inter-assay mean accuracy, expressed as percents of theoretical, was between 99.8% and 102.3%. The assay has high recovery (∼80%) and minimal matrix effect (0.95–1.00). BMS-911543 was stable in human plasma for at least 24 h at room temperature, 90 days at −20 ◦ C, and following three freeze–thaw cycles. The validated method was successfully applied to sample analysis in clinical studies. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Janus kinase 2 (commonly called JAK2) is a non-receptor tyrosine kinase [1]. It is a member of the Janus kinase family and has been implicated in signaling by members of the type II cytokine receptor family (e.g., interferon receptors) [2]. JAK2 is a promising target for therapeutic intervention in myeloproliferative disorders (MPDs) [1] and several JAK inhibitors are in pre-clinical or clinical development for MPDs [3]. BMS-911543, as shown in Fig. 1A, is a potent and selective JAK2 inhibitor under clinical development for the treatment of MPDs [1]. BMS-911543 showed a high degree of functional selectivity toward JAK2 in a variety of pre-clinical models, which supported the therapeutic rationale for its clinical development [1]. In a multicenter phase 1/2a study in patients with myelofibrosis, BMS-911543 showed an acceptable safety and efficacy profile, and was generally well tolerated [4].
∗ Corresponding author. Tel.: +1 6092525336. E-mail address: [email protected] (L. Yuan). http://dx.doi.org/10.1016/j.jchromb.2015.04.013 1570-0232/© 2015 Elsevier B.V. All rights reserved.
Evaluations of BMS-911543 as a drug candidate requires an accurate, sensitive and rugged LC–MS/MS bioanalytical assay to support both preclinical and clinical studies. Ultra-high performance liquid chromatography coupled with tandem mass spectrometry (UHPLC–MS/MS) technology has been widely used for the bioanalytical assays to support dug development [5,6]. UHPLC uses sub-2 m particles stationary phases, therefore, has better resolving power than conventional LC and can usually achieve faster analysis and higher throughput with good chromatographic performance. In this manuscript, we report the method development and validation of a high-throughput UHPLC–MS/MS method for the quantification of BMS-911543 in human plasma. A systematic method development approach was applied to achieve optimized mass spectrometry, chromatography, and sample extraction conditions that used several previously published strategies to accelerate method development for high-throughput bioanalytical assays [5] and to ensure quality of the developed method [6]. Incurred samples were also used in method development to minimize bioanalytical risks from potential metabolite interference and ensure the quality of the method. Stable isotope labeled BMS-911543 (13 C4 -BMS-911543, Fig. 1B) was used as the internal standard. The assay was validated and successfully used to support clinical studies with excellent assay performance.
86
J. Liu et al. / J. Chromatogr. B 991 (2015) 85–91
samples were prepared using the same method at six different concentration levels: LLOQ QC (1.0 ng/mL), low QC (3.0 ng/mL), geometric mean QC (GM QC, 30 ng/mL), mid QC (250 ng/mL), high QC (400 ng/mL), and dilution QC (20,000 ng/mL). The QCs were aliquoted into polypropylene tubes and stored frozen at approximately −20 ◦ C. The internal standard stock standard solution was prepared by dissolving appropriate amounts of stable labeled standard in 1:1 (v/v) acetonitrile/DMSO to give a final concentration of 1.00 mg/mL of 13 C4 -BMS-911543. The internal standard working solution of 100 ng/mL of 13 C4 -BMS-911543 was prepared by adding 200 L of the internal standard stock solution to a 200-mL glass volumetric flask, then diluting to the volume with acetonitrile/water (70/30, v/v). 2.4. Sample extraction
Fig. 1. Chemical structures of BMS-911543 (A) and its internal standard, 13 C4 -BMS911543 (B).
2. Experimental 2.1. Chemicals and reagents HPLC-grade acetonitrile, methanol, concentrated ammonium hydroxide, ammonium acetate, methyl-tert butyl ether (MTBE), and ACS grade glacial acetic acid were purchased from Fisher (Pittsburgh, PA, USA). Formic acid (96%) was purchased from Sigma–Aldrich (St. Louis, MO, USA). The reference standards of BMS-911543 and its internal standard, 13 C4 -BMS-911543, were produced in Bristol-Myers Squibb Co. (BMS). Normal human plasma, with K2 EDTA as an anticoagulant, was purchased from Bioreclamation (Hicksville, NY, USA).
Samples were thawed at room temperature, and vortexed to ensure homogeneity. Each plasma sample tube was uncapped and 100 L of sample was transferred into the appropriate well of a 96well plate. Then, 50 L of the internal standard working solution (100 ng/mL of 13 C4 -BMS-911543) and 50 L of 1.0 M ammonium acetate in water solution were added to each sample, followed by a brief vortex to ensure mixing. Extraction solvent (MTBE, 600 L) was added to each sample and shaken at high speed for 20 min. After centrifugation to separate the phases, 500 L of the organic phase layer was transferred into a 96-well collection plate. The samples were dried with a Speed-Dry evaporator at 40 ◦ C under a nitrogen stream and reconstituted with 400 L of the reconstitution solution of acetonitrile/water (25/75, v/v). A volume of 5 L of the reconstituted samples was injected into the LC–MS/MS system for analysis. 2.5. Chromatographic conditions The analytical column used was a UHPLC column: Zorbax Extend-C18, 2.1 mm × 50 mm, 1.8 m particle size, Agilent (Santa Clara, CA, USA). An isocratic elution (60/40, mobile phase A/mobile phase B) was utilized and the analytical column was kept at 40 ◦ C during the analysis. A flow rate of 800 L/min was used. The mobile phase A consisted of 10 mM ammonium bicarbonate with 0.10% ammonium hydroxide in water and the mobile phase B consisted of acetonitrile.
2.2. Instrumentation
2.6. MS/MS detection
The UHPLC system consisted of a Flux pump and a SIL-HTC autosampler (LEAP Technologies, Carrboro, NC, USA). The API 4000 mass spectrometer controlled by AnalystTM software was from Applied Biosystems/MDS Sciex (Concord, ON, Canada). AnalystTM version 1.4.2 was used as the data acquisition software. The column heater HotDog 5090 was from Prolab Instruments (Reinach, AG, Switzerland). A JANUS Mini liquid handling system (Perkin Elmer, Downers Grove, IL, USA) was used for the automated liquid transfer in sample extraction (96-well plate format).
LC–MS/MS detection was performed using a Sciex API 4000 triple quadrupole mass spectrometer with a Turbo Ionspray® ionization source operated in the positive ion mode. The ion spray voltage was 4500 V and the source temperature was 550 ◦ C. The CAD gas setting was 6 and curtain gas setting was 50. The ion source gas 1 (GS1) and ion source gas 2 (GS2) settings were 50 and 60, respectively. The de-cluster potential was 121 V. The collision energy (CE) was 49 eV and the collision cell exit potential (CXP) was 24 V. The selected reaction monitoring (SRM) monitored was m/z 433 → 281 for BMS-911543 and m/z 437 → 285 for 13 C4 BMS-911543. The dwell time for each transition monitored was 150 ms.
2.3. Preparation of standard and QC samples Standard and QC stock solutions of BMS-911543 (1.00 mg/mL) were prepared from separate weighings by dissolving the reference compound into 1:1 (v/v) acetonitrile/DMSO. The intermediate standard working solution containing 20,000 ng/mL of BMS911543 was prepared by dilution of the stock standard solution with human plasma. The calibration curves with eight levels (1.00, 2.50, 5.00, 10.0, 50.0, 100, 250 and 500 ng/mL BMS-911543) were prepared fresh daily by diluting appropriate volumes of intermediate standard working solutions with human plasma. The QC
2.7. Quantitation The peak areas of both analyte and its internal standard were determined using AnalystTM software version 1.4.2. For each analytical batch, a calibration curve was derived from the peak area ratios (analyte: internal standard) using weighted linear least squares regression of the area ratio vs. the concentration of the standards. A weighting of 1/x2 (where x is the concentration of a
J. Liu et al. / J. Chromatogr. B 991 (2015) 85–91
87
given standard) was used for curve fitting. The regression equation was used to back-calculate the measured concentration for each standard and QC. The results were compared to the theoretical concentration to obtain the accuracy, expressed as a percentage of the theoretical value, for each standard and QC measured. 3. Results and discussion 3.1. Method development A systematic method screening and optimization approach [6–9] was used to optimize the mass spectrometry, chromatography, and sample extraction conditions for the LC–MS/MS assay. Briefly, four aqueous mobile phases from acidic to basic and a total of 11 different UHPLC columns (C18, modified C18, phenyl and other types from several different manufacturers) were screened to achieve the best chromatographic separation and sensitivity. Liquid–liquid extraction (LLE) methods using different extraction buffers and solvents were evaluated for the optimized extraction recovery and cleanness of the extracts. We also used rat and dog incurred samples in method development to minimize the risk of potential interferences from metabolites, which ensured the quality and robustness of the developed method. 3.1.1. Mass spectrometry, chromatography, and sample extraction MS scan and product ion scan mass spectra were obtained by the infusion of BMS-911543 and its internal standard, 13 C4 -BMS911543, solutions via a tee connection between the LC column and mass spectrometer inlet. As shown in Fig. 2A and C, the protonated ions of m/z 433 and 437 are the predominant ion of BMS-911543 and its internal standard under positive electrospray ionization, respectively. The most abundant product ions of BMS-911543 and its internal standard are at m/z 281 and 285, respectively (Fig. 2B and D). Therefore, the SRM transitions m/z 433 → 281 and 437 → 285 were selected for the monitoring of BMS-911543 and its internal standard. A total of 11 different UHPLC columns (see Fig. 3 for detailed column information) were screened. As shown in Fig. 3, both
Fig. 3. LC–MS/MS chromatograms of BMS-911543 using different UHPLC columns (1: Agilent SB-C18, 2: Waters Aquity BEH shield RP18, 3: Agilent Zorbax ExtendC18, 4: Waters Aquity BEH phenyl, 5: Phenomenex Kinetex C18, 6: Agilent Eclipse PAH, 7: Agilent Eclipse plus C18, 8: Agilent XDB C18, 9: Thermo Gold aq, 10: Waters Aquity BEH C18, and 11: Waters Aquity HSS T3, all columns are 2.1 mm × 50 mm with sub-2 m particle size). The gradient method was as follows: 0–1.5 min 20–80% B, 1.5–2.0 min 80% B, 2.0–2.5 min 20% B. The flow rate was 0.80 mL/min.
Zorbax Extend-C18 column (column 3) and Phenomenex Kinetex C18 column (column 5) offered a good combination of peak shape and sensitivity. The Zorbax Extend-C18 column provided a better sensitivity than the Phenomenex Kinetex C18 column and, therefore, was selected for further optimization. Four different aqueous mobile phases from acidic to basic (0.1% formic acid in water, 10 mM ammonium acetate with 0.01% acetic acid in water, 10 mM ammonium formate with 0.05% formic acid in water, and 10 mM ammonium bicarbonate with 0.1% ammonium hydroxide in
Fig. 2. Electrospray positive ion MS scan ion spectra and product ion spectra of BMS-911543 (A and B) and 13 C4 -BMS-911543 (C and D).
88
J. Liu et al. / J. Chromatogr. B 991 (2015) 85–91
Fig. 4. LC–MS/MS chromatograms of BMS-911543 using different mobile phases (mobile phase A: I: 0.1% formic acid in water, II: 10 mM ammonium acetate with 0.01% acetic acid in water, III: 10 mM ammonium formate with 0.05% formic acid in water, and IV: 10 mM ammonium bicarbonate with 0.1% ammonium hydroxide in water; mobile phase B: acetonitrile). The column used was a Zorbax Extend-C18 column. The gradient method was as follows: 0–1.5 min 20–80% B, 1.5–2.0 min 80% B, 2.0–2.5 min 20% B. The flow rate was 0.80 mL/min.
water) were screened. Although similar peak shape was obtained under these different mobile phases, much better sensitivity was achieved using the basic mobile phase, 10 mM ammonium bicarbonate with 0.1% ammonium hydroxide in water (Fig. 4). After further optimization, quick and efficient chromatographic separation of BMS-911543 was obtained using an isocratic elution with 60% of 10 mM ammonium bicarbonate with 0.1% ammonium hydroxide in water and 40% of acetonitrile on a Zorbax Extend-C18 column. The retention time is approximately 0.7 min for BMS911543 and its internal standard. The total run time is only 1.5 min, which ensured the high-throughput of the assay. Liquid–liquid extraction has been routinely used in our laboratory for LC–MS/MS bioanalytical assays to support preclinical and clinical studies. MTBE as the extraction solvent gave a good and consistent extraction recovery for BMS-911543 in plasma. Matrix effect is the suppression or enhancement of the analyte ionization from co-eluting matrix constituents [10]. The abundant phospholipids in plasma samples are one major source of matrix effect, and may affect the accuracy, precision and sensitivity of LC–MS/MS bioanalytical assays [6]. We evaluated the phospholipids profiles of the extracted plasma sample using positive precursor ion scan of m/z 184 and negative precursor ion scan of m/z 153 and a longer 10-min gradient [11]. As shown in Fig. 5, the BMS-911543 peak was well separated from the major phospholipids peaks, indicating minimum risks of matrix effect from phospholipids. Consistent column performance has been maintained for injection of multiple sample batches throughout the sample analysis of a clinical study. In addition, a short column wash with high organic solvent could be added to clean up the column and further extend the column lifetime for continuous injection of a large number samples. 3.1.2. Evaluation of potential interference from metabolites To avoid potential interference from BMS-911543 metabolites, the SRM channels of all known and predicted metabolites (see Table 1) were monitored using a mixture of pooled rat and dog incurred samples. To minimize the risk of missing a metabolite peak eluting close to the analyte peak, the experiments were also done using a much longer (25 min) gradient. The SRM chromatograms of BMS-911543 and all detectable metabolites in the pooled incurred
Fig. 5. Chromatograms of BMS-911543 and phospholipids in a pooled rat and dog incurred sample. The column used was a Zorbax Extend-C18 column. The gradient method was as follows: 0–1.5 min 25–30% B, 1.5–2 min 30–40% B, 2–8 min 40–90% B, 8–9 min 90% B, 9–10 min 25% B. Table 1 SRM transitions used in the assessment of metabolite interference.
BMS-911543 (P) P+O P-CH2 + O P-CH2 P + 2O P + O + glucuronide P-CH2 + glucuronide P-CH2 + O + glucuronide P-CH2 + GSH
[MH]+
Product ions
433.2 449.2 435.2 419.2 465.2 625.2 595.3 611.2 724.2
281.0 281.0, 297.0 283.0, 267.0 267.0 313.0 449.2, 297.2 443.1, 281.0 435.2, 283.2 433.2
sample are shown in Fig. 6. No metabolite peak interfering with BMS-911543 was observed using either the short or long gradient. In addition, the same experiments were also done using a column with a different stationary phase (Aquity BEH phenyl) and a mobile phase at a different pH (acidic: 10 mM ammonium formate with 0.05% formic acid in water), and no interfering metabolite peak was observed in either case (data not shown). The evaluations using incurred samples ensured the quality of the method and minimized the risk of re-doing the method development due to unexpected metabolite interference. Unlike animal incurred samples that can be obtained from discovery studies, it is impractical to get human incurred sample for method development before the start of human clinical studies. In most cases, humans have similar metabolic pathways to animal species. For BMS-911543, there was no unique human metabolite identified in in vitro studies. Thus, the pooled animal incurred sample can well represent the human incurred samples and help to ensure the quality of the human assay. 3.2. Assay validation Validation experiments were designed with reference to the Guidance for Industry-Bioanalytical Method Validation by the Food and Drug Administration (FDA) of the United States [12] and the Guideline on Bioanalytical Method Validation by the European Medicines Agency (EMA) [13]. Standard curve linearity, lower limit
J. Liu et al. / J. Chromatogr. B 991 (2015) 85–91
89
Fig. 6. SRM chromatograms of BMS-911543 and its metabolites in a pooled rat and dog incurred sample. The chromatographic separation was with a Zorbax Extend-C18 column using a 2.5 min gradient (left), or a 25 min gradient (right). The 2.5 min gradient method was as follows: 0–1.5 min 25–30% B, 1.5–2.0 min 30–40% B, 2.0–2.5 min 25% B. The 25 min gradient method was as follows: 0–18.5 min 5–30% B, 18.5–22 min 30–90% B, 22–24 min 90% B, 24–25 min 5% B. The flow rate was 0.80 mL/min.
of quantitation (LLOQ), accuracy and precision, selectivity, recovery, matrix effect, and stability were evaluated. The experimental design and results of method validation are presented in the following sections. 3.2.1. Standard curve linearity Assay linearity was evaluated using eight calibration standards analyzed in duplicate over the nominal concentration range of 1.00 ng/mL and 500.00 ng/mL for BMS-911543. For a total of eight validation runs, a linear 1/x2 weighted regression model provided the best fit for BMS-911543 in human plasma with the coefficient of determination (r2 ) between 0.9973 and 0.9997, and therefore was selected for curve fitting. The mean back-calculated concentrations of the standards were between 90.0% and 107.0% of the theoretical concentrations. 3.2.2. Accuracy, precision, and dilution The accuracy and precision data of BMS-911543 in human plasma are summarized in Table 2. Based on four levels of analytical QCs (low, GM, mid and high), the intra-assay precision, expressed as coefficient of variation (CV), was within 5.0%, and the interassay precision was within 2.6%. The accuracy, expressed as %Dev, was within ±2.3% Dev of the nominal concentration. The results
demonstrated that the method was accurate and precise for the analysis of BMS-911543 in human plasma. The suitability of study samples being diluted with drug-free plasma on the day of assay (without undergoing an additional freeze–thaw cycle) was also evaluated as part of the validation. A QC level (dilution QC), used specifically for dilutions, was prepared with the concentrations of analyte at 20,000 ng/mL. To achieve a 100-fold dilution, 10 L of dilution QC and 990 L of blank matrix were combined, and 100 L of resulting sample was extracted. As shown in Table 2, the accuracy and precision for dilution QCs were all well below the 15% acceptance criteria, demonstrating the reliability of the dilution. 3.2.3. LLOQ and selectivity The LLOQ for the analyte was assessed using human K2EDTA plasma samples at 1.0 ng/mL of BMS-911543, the lowest concentrations for the analyte in the standard curve. Six different lots of control human K2EDTA plasma were spiked to obtain the six LLOQ samples. The deviations of the measured concentrations from the nominal LLOQ values were within ±17.0% for all the six lots (acceptance criteria is to achieve within ±20.0% for at least five of the six lots). Six different lots of human plasma were analyzed with and without internal standard to determine whether any endogenous constituents interfered with the analyte or the internal standard.
Table 2 QC accuracy and precision for BMS-911543 in human plasma in validation. QC type (nominal conc. in ng/mL)
LLOQ (1.00)
Low (3.00)
GM (30.0)
Mid (250)
High (400)
Dilution (20,000)
Mean observed conc. %Dev Between run precision (%CV) Within run precision (%CV) Total variation (%CV) n Number of runs
0.99 −1.0 4.2 6.2 7.5 30 5
3.07 2.3 0.0 5.0 4.9 38 8
29.93 −0.2 2.6 1.3 2.9 38 8
248.03 −0.8 2.4 1.4 2.8 38 8
394.44 −1.4 1.8 1.4 2.3 38 8
19030.74 −4.8 4.5 2.0 4.9 30 5
GM, geometric mean.
90
J. Liu et al. / J. Chromatogr. B 991 (2015) 85–91 Table 4 QC accuracy and precision for BMS-911543 in human plasma for a clinical study. QC type (nominal conc. in ng/mL) Mean observed conc. %Dev Between run precision (%CV) Within run precision (%CV) Total variation (%CV) n Number of runs
Low (3.00)
GM (30.0)
Mid (250)
High (400)
30.36
248.29
391.22
−0.2 5.9
1.2 5.0
−0.7 4.1
−2.2 4.1
4.1
3.6
3.5
4.6
7.2
6.2
5.4
6.2
2.99
125 37
125 37
125 37
125 37
GM, geometric mean.
Fig. 7. Representative SRM chromatograms of BMS-911543 in blank plasma (A), LLOQ of 1.0 ng/mL (B), and a patient sample (C).
No significant interfering peaks from the plasma were found at the retention time of either the analyte or its internal standard, which demonstrated the assay has good specificity. Representative LC–MS/MS chromatograms of a blank plasma and a LLOQ sample are shown in Fig. 7. 3.2.4. Matrix effect The matrix effect was determined at concentrations of 3.0 and 400 ng/mL for BMS-911543 by dividing the analyte response in human plasma spiked post-extraction with BMS-911543 by the analyte response (peak area) of those spiked in reconstitution solution. The matrix effect of the internal standard was determined similarly. Matrix effect of 1 indicates ion enhancement. The matrix effects were 0.95–1.00 for BMS-911543, and 1.00–1.01 for the internal standard. This demonstrates that there was no or minimal matrix effect on the measurement of the analyte. 3.2.5. Extraction recovery The recovery of the analyte from human plasma during extraction was determined at 3.0 and 400 ng/mL by comparing the response ratios in human plasma samples spiked with the analyte prior to extraction with those spiked post-extraction. The recovery of the internal standard was determined similarly at 100 ng/mL. The recovery of BMS-911543 and its internal standard was 79.3–79.7% and 78.9–83.5%, respectively. 3.2.6. Stability The room temperature, freeze–thaw, and frozen storage stability of BMS-911543 in human plasma were evaluated in triplicate using QCs. The reinjection integrity was assessed by re-injecting an entire run. The processed sample stability was assessed by testing processed and stored QCs against freshly prepared standards. Stability of processed sample in reconstitution solution was investigated in order to demonstrate the stability of the extracted analyte
before and during the injection of an analytical run. To establish the stability of the analyte, the deviations of the mean measured concentrations of the test samples have to be within ±15% of the nominal concentrations. As shown in Table 3, at least 24 h room temperature, 3 cycles freeze–thaw, 90 day frozen storage stability at −20 ◦ C, 5 day processed sample stability at 5 ◦ C, and 48 h reinjection integrity at 5 ◦ C were established. To ensure the stability of the analyte during human blood sample collection and processing, whole blood stability of BMS-911543 was also evaluated. Whole blood stability was evaluated by adding BMS-911543 to fresh human blood at two concentration levels. Plasma was derived immediately after the addition of the compounds (zero time), and after being stored for 2 h on ice and at room temperature. Blood stability was based on the deviation of the test time point value from the zero (0) time value. To establish the stability of the analyte, the deviations of the mean measured concentrations of the test samples have to be within ±15% of the zero time value. BMS-911543 was determined to be stable for at least 2 h at RT or 4 ◦ C (on ice) in human blood. Stock solution stability of BMS-911543 and its internal standard in acetonitrile/DMSO was also evaluated at both ambient temperatures and at 2–8 ◦ C as part of the validation. Results for the determination of stock solution stability were calculated by comparing mean response ratios (area of response per unit of concentration) of stability solutions to mean response ratios of freshly prepared control solutions. To establish the stock solution stability of the analyte, the means of the responses of the two solutions have to be within 5.0% of each other. BMS-911543 stock solutions were determined to be stable for at least 24 h at room temperature and 113 days at 2–8 ◦ C. 3.2.7. Assay application for clinical studies The validated assay was successfully used for the bioanalytical support of clinical studies in patients. As shown in Table 4, for one supported clinical study, excellent assay performance was achieved throughout the study. For a total of 37 runs, the accuracy was within ±2.2% Dev of the nominal concentration, the intraassay precision was within 4.6%, and the inter-assay precision was
Table 3 Stability of BMS-911543 in human plasma. QC type (nominal conc.)
Low QC 3.00 ng/mL
Sample condition
Mean conc.
%Dev
Mean conc.
%Dev
Mean conc.
%Dev
24 h at RT After 3rd freeze–thaw cycle 90 Day at −20 ◦ C 5 Day at 5 ◦ C Processed samples 48 h at 5 ◦ C re-injection integrity
2.99 3.01 3.08 3.08 3.00
−0.3 0.3 2.7 2.7 0.0
390.01 384.46 404.79 392.95 388.01
−2.5 3.9 1.2 −1.8 −3.0
18884.03 19363.98 20439.25 18725.02 18398.52
−5.6 −3.2 2.2 −6.4 −8.0
High QC 400 ng/mL
Dilution QC 20,000 ng/mL
J. Liu et al. / J. Chromatogr. B 991 (2015) 85–91
91
References
Fig. 8. Plasma concentration vs. time curve of a patient after a daily oral dosing of 80 mg of BMS-911543.
within 5.9%. Incurred sample reanalysis (ISR) was performed with 179 out of 1630 study samples, and 96% and 81% of the ISR samples were within ±10% and ±5% from the mean of initial and repeat test results, respectively, demonstrating the excellent reproducibility of the assay. A representative concentration vs. time profile of BMS-911543 in a patient from the clinical study is shown in Fig. 8. 4. Conclusions Here, we reported the method development and validation of a UHPLC–MS/MS bioanalytical assay for BMS-911543 in human plasma over the concentration range of 1.0–500 ng/mL. Rapid, accurate and rugged analysis of BMS-911543 was achieved using liquid–liquid extraction and an isocratic elution. The use of incurred samples during method development minimized potential bioanalytical risks due to the interferences from metabolites and ensured the quality of the method. The assay was successfully used for supporting clinical studies. Acknowledgement Authors would like to take this opportunity to thank Kai Cao from Bristol-Myers Squibb Radiosynthesis team for providing the stable isotope labeled internal standard.
[1] A.V. Purandare, T.M. McDevitt, H. Wan, D. You, B. Penhallow, X. Han, R. Vuppugalla, Y. Zhang, S.U. Ruepp, G.L. Trainor, L. Lombardo, D. Pedicord, M.M. Gottardis, P. Ross-Macdonald, H. de Silva, J. Hosbach, S.L. Emanuel, Y. Blat, E. Fitzpatrick, T.L. Taylor, K.W. McIntyre, E. Michaud, C. Mulligan, F.Y. Lee, A. Woolfson, T.L. Lasho, A. Pardanani, A. Tefferi, M.V. Lorenzi, Characterization of BMS-911543, a functionally selective small-molecule inhibitor of JAK2, Leukemia 26 (2012) 280–288. [2] P.J. Murray, The JAK-STAT signaling pathway: input and output integration, J. Immunol. 178 (2007) 2623–2629. [3] C. Kumar, A.V. Purandare, F.Y. Lee, M.V. Lorenzi, Kinase drug discovery approaches in chronic myeloproliferative disorders, Oncogene 28 (2009) 2305–2313. [4] A. Pardanani, A.W. Roberts, J.F. Seymour, K. Burbury, S. Verstovsek, H.M. Kantarjian, K. Begna, H. Yoshitsugu, T.A. Gestone, P. Phillips, G. Xing, G. Peltz, M.V. Lorenzi, L. Alland, A. Woolfson, A. Tefferi, BMS-911543, A Selective JAK2 Inhibitor: A Multicenter Phase 1/2a Study In Myelofibrosis, 55th American Society of Hematology Annual Meeting and Exposition. New Orleans, LA., 2013. [5] G. Liu, H.M. Snapp, Q.C. Ji, M.E. Arnold, Strategy of accelerated method development for high-throughput bioanalytical assays using ultra high-performance liquid chromatography coupled with mass spectrometry, Anal. Chem. 81 (2009) 9225–9232. [6] M. Jemal, Z. Ouyang, Y.Q. Xia, Systematic LC–MS/MS bioanalytical method development that incorporates plasma phospholipids risk avoidance, usage of incurred sample and well thought-out chromatography, Biomed. Chromatogr. 24 (2010) 2–19. [7] L. Yuan, H. Jiang, Z. Ouyang, Y.-Q. Xia, J. Zeng, Q. Peng, R.W. Lange, Y. Deng, M.E. Arnold, A.-F. Aubry, A rugged and accurate liquid chromatography–tandem mass spectrometry method for the determination of asunaprevir, an NS3 protease inhibitor, in plasma, J. Chromatogr. B 921–922 (2013) 81–86. [8] L. Yuan, Y. Fu, D. Zhang, Y.-Q. Xia, Q. Peng, A.-F. Aubry, M.E. Arnold, Use of a carboxylesterase inhibitor of phenylmethanesulfonyl fluoride to stabilize epothilone D in rat plasma for a validated UHPLC–MS/MS assay, J. Chromatogr. B 969 (2014) 60–68. [9] L. Yuan, H. Jiang, N. Zheng, Y.-Q. Xia, Z. Ouyang, J. Zeng, B. Akinsanya, J.L. Valentine, J.D. Moehlenkamp, Y. Deng, A.-F. Aubry, M.E. Arnold, A validated LC–MS/MS method for the simultaneous determination of BMS-791325, a hepatitis C virus NS5B RNA polymerase inhibitor, and its metabolite in plasma, J. Chromatogr. B 973 (2014) 1–8. [10] B.K. Matuszewski, M.L. Constanzer, C.M. Chavez-Eng, Strategies for the assessment of matrix effect in quantitative bioanalytical methods based on HPLC–MS/MS, Anal. Chem. 75 (2003) 3019–3030. [11] Y.Q. Xia, M. Jemal, Phospholipids in liquid chromatography/mass spectrometry bioanalysis: comparison of three tandem mass spectrometric techniques for monitoring plasma phospholipids, the effect of mobile phase composition on phospholipids elution and the association of phospholipids with matrix effects, Rapid Commun. Mass Spectrom. 23 (2009) 2125–2138. [12] FDA, Guidance for Industry: Bioanalytical Method Validation, 2001 http:// www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/ Guidances/ucm070107.pdf [13] European Medicines Agency, Guideline on Bioanalytical Method Validation, http://www.ema.europa.eu/docs/en GB/document library/Scientific 2011 guideline/2011/08/WC500109686.pdf
