Journal of Chromatography B, 967 (2014) 211–218
Contents lists available at ScienceDirect
Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb
Development of an LC–MS/MS method for high throughput quantification of metformin uptake in transporter inhibition assays Marianne Vath, Lizbeth Gallagher, Wilson Shou, Harold Weller, Lisa Elkin, Jun Zhang ∗ Bristol Myers Squibb, Research and Development, 5 Research Parkway, Wallingford, CT 06492, United States
a r t i c l e
i n f o
Article history: Received 10 June 2014 Accepted 12 July 2014 Available online 24 July 2014 Keywords: Metformin LC–MS/MS Transporter inhibition Porous graphite carbon Silica hydride
a b s t r a c t A high throughput LC–MS/MS method for quantification of metformin substrate uptake enables conversion of radiometric transporter inhibition assays for multidrug and toxin extrusion transporters (MATE 1 and 2) and organic cation transporter 2 (OCT2) to a nonradioactive format. Such conversion greatly simplifies assay complexity and reduces assay costs. The development of a quantitative LC–MS/MS method for metformin in support of the high throughput transporter inhibition assays faced specific challenges of achieving both adequate chromatographic retention and rapid analytical turnaround. Here we report a method that circumvents both challenges. The utilization of a porous graphitic carbon column (Hypercarb) ensured adequate retention of highly polar metformin in biological samples. The combined employment of a ballistic gradient on a 3 mm × 30 mm, 5 m Hypercarb column, and dual staggered chromatography coupled with multiple injection chromatography acquisition, yielded a fast injection-to-injection cycle time of 30 s. The method demonstrated good accuracy, precision and excellent robustness for high throughput applications, and has been successfully implemented in the development and validation of the nonradioactive transporter inhibition assays for MATEs and OCT2. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Due to the potential impact of transporter inhibition on the clinical development of drug candidates, there has been a growing interest in predicting transporter related drug–drug interactions in the drug discovery setting [1–3]. Multidrug and toxin extrusion transporters 1 and 2 (MATEs), and organic cation transporter 2 (OCT2) are uptake transporters playing a key role in the disposition and renal clearance of cationic drugs and endogenous compounds, and both transporters exhibit overlapping substrate specificities for organic cations of metformin (Fig. 1). Uptake transporter inhibition assays have been developed to assess potential inhibitors of MATE and OCT2 transport by measuring the extent to which metformin uptake is inhibited by the co-administered drug candidate. These assays have typically been conducted using radiolabeled metformin and rely upon liquid scintillation counting (LSC) for sample analysis, which is a high throughput but expensive method, due to the cost of reagents and the associated special handling requirements [4]. It would therefore be advantageous to convert assays to a nonradioactive format if an appropriate label-free detection method
∗ Corresponding author. Tel.: +1 203 677 5610; fax: +1 203 677 6984. E-mail address: [email protected] (J. Zhang). http://dx.doi.org/10.1016/j.jchromb.2014.07.024 1570-0232/© 2014 Elsevier B.V. All rights reserved.
for metformin could be developed. In order to enable the conversion, we set out to develop a high throughput LC–MS/MS method for metformin quantification with a runtime of less than 1 min, to support these high throughput 384-well transporter inhibition assays. To our knowledge, this is the first report of such an assay and LC–MS/MS bioanalysis in the literature. There are two primary challenges associated with developing an LC–MS/MS method for metformin in support of transporter inhibition assays. Firstly, metformin is a highly polar compound (log P = −2.6), which presents a significant challenge in achieving adequate chromatographic retention. Secondly, analysis capacity and speed must be sufficient to support the large number of compounds studied for early profiling of transporter inhibition, which are screened in 10-point dose response in 384-well plate format. The LC–MS/MS method for substrate detection must match the sensitivity, speed and robustness of the LSC method in context of the above challenges. Metformin is the most widely prescribed anti-diabetic drug, used either alone or co-administered with other antihyperglycemic drugs for treatment of type-II diabetes. A very extensive effort has been seen in the development of LC–MS/MS methods for metformin, along with concomitantly administered drugs in order to evaluate possible pharmacokinetic interactions [5–8]. The main chromatographic methods reported are
212
M. Vath et al. / J. Chromatogr. B 967 (2014) 211–218
acetate, ammonium bicarbonate and dimethyl sulfoxide (DMSO) were also purchased from EMD Chemicals. 2.2. Transporter assay conditions
Fig. 1. Structures of metformin (a) and internal standard d6-metformin (b).
reversed-phase [6,7,9] and hydrophilic interaction liquid (HILIC) [10–12]. Most of the reported methods have LC runtimes too long (>5 min) to be considered suitable for in vitro assay support. Two HILIC methods, however, showed some potential for high throughput applications. Discenza et al. achieved good retention for both metformin and the concomitant drug in 2 min, via a combined mechanism of ion-exchange and reversed-phase chromatography on a HILIC column, by conducting a gradient elution from low organic to medium organic composition [13]. Zhang et al. reported a conventional HILIC method with a runtime of 3 min, using a gradient from high organic to low organic. In order to maintain good peak shape, however, they had to dilute the samples extensively by 38-fold using ACN following a protein precipitation of high solvent ratio (1:25/sample: ACN) [14]. Our attempt to apply HILIC methods for analysis of metformin from transporter inhibition assays quenched in ACN were not successful, due to shifting retention time and distorted peak shape in batch analyses, even at a reduced injection volume of 2 L (data not shown). In this study we evaluated both a porous graphitic carbon (Hypercarb) column under reversed-phase conditions, and silica hydride columns under aqueous normal phase (ANP) mechanism for high throughput metformin bioanalysis. To our knowledge, this is the first report of an LC–MS/MS method for metformin using either of these two columns. The suitability of each developed method was assessed in terms of chromatographic retention, peak shape and analytical throughput. For the final adopted method using the Hypercarb column, analytical performance such as accuracy, precision and robustness were also characterized. Using the final adopted chromatographic method, sample analysis for transporter inhibition assays was performed on a multiplexed LC–MS/MS platform we previously reported [15], to further reduce analytical turnaround. The platform uses dual staggered LC systems, and switching the mass spectrometric detection between the two systems for peak eluting chromatographic window only, minimizes the mass spectrometer idle time while maintaining the chromatographic quality as on a single LC system. Data acquisition of the multiplexed system is done in the fashion of multiple injection chromatography (MIC), whereby data for all the injections pertaining to a 384-well plate from the multiplexed LCs is acquired to a single file, thus eliminating the mass spectrometer reset delay. The platform also utilizes an open-deck manifold for sample holder to eliminate the auto-sampler overhead time spent on opening and closing sample drawers, resulting in a significant time saving for bioanalysis of a large number of samples. 2. Experimental 2.1. Chemicals and reagents Metformin hydrochloride and d6 metformin hydrochloride were purchase from Toronto Research Chemicals (Toronto, Ontario, Canada). HPLC grade water, acetonitrile, and methanol were purchased from EMD Chemicals (Gibbstown, NJ, USA). Ammonium
Transporter expressing cell lines were developed by subcloning cDNA for MATE1 (SLC47A1; NM 018242) into pIRESneo2 expression vector (Clontech, Mountain View, CA) and transfection into HEK293 cells (American Type Culture Collection, Manassas, VA, USA), which were maintained in Modified Eagle Medium (MEM) supplemented with 10% fetal calf serum (FCS) and selection antibiotics. Forty-eight hours prior to assay, MATE1-HEK293 cells were seeded at a density of 20,000 cells per well into columns 1–22 of a 384-well poly-d-lysine (PDL)-coated 384-well plate; the parental cell line (HEK293) was seeded at the same density into columns 23 and 24 of each plate for use as minimal signal controls (low control or 100% inhibition). Cells were grown for 48 h in MEM media containing 10% FBS at 37 ◦ C and 5% CO2 in a humidified incubator. Immediately prior to assay, media was removed and plates were washed in modified HBSS buffer (HBSS containing 10 mM HEPES, pH 8.4) using a BioTek 405 automated plate washer (BioTek, Winooski, VT). After complete aspiration of all wash buffer and addition of 20 L of modified HBSS, a LabCyte ECHO-550 noncontact acoustic dispenser (Sunnyvale, CA) was used to transfer 125 nL from each well of the compound source plate to a corresponding well of the cellular assay plate. After a 10 min pre-incubation in compound, 5 L of 25 M metformin in modified HBSS was added to each well of the assay plate using a MultidropTM Combi Reagent Dispenser (Thermo Scientific, Waltham, MA). Assay plates were incubated at room temperature for 40 min At the end of the incubation time, assay plates were aspirated and washed 3 times in modified HBSS using an automated plate washer. Plates were aspirated a final time and cells were lysed by addition of acetonitrile containing 50 nM d6 metformin as internal standard, and the supernatant was subjected to LC–MS/MS analysis. Assay procedures for transporter MATE2K and OCT2 were similar. 2.3. Instrument and LC–MS/MS conditions A modified Aria LX-2 multiplexed system (Thermo Scientific, CA, USA) was used as previously described [15]. The system consisted of two sets of independent 20AD binary pumps (Shimadzu Scientific Instruments, MD, USA) for dual staggered chromatography, a valve interface module (VIM, Thermo Scientific) for stream selection and two modified CTC HTS PAL auto-samplers (Leap Technologies, NC, USA). Sample plates in 384-well plate format were placed on a cooled, open-deck manifold for injection. Aria software (version 1.6.3, Thermo Scientific) controlled all peripheral components. A TSQ Quantum Ultra triple quadrupole mass spectrometer (Thermo Scientific) was used for mass spectrometric analysis. The MS was equipped with heated electrospray ionization (H-ESI) and used in positive ionization mode. XCalibur 2.2 was used for mass spectrometric data acquisition in the MIC mode. The final adopted LC–MS/MS method used a Hypercarb column (Thermo Scientific) with a dimension of 3 mm × 30 mm, 5 m. The chromatographic conditions are listed in Table 1. For MS/MS conditions, a universal set of source parameters was employed, including a spray voltage at 3.5 kV for positive ionization, a vaporizer temperature at 450 ◦ C and a capillary temperature at 275 ◦ C. The sheath gas and auxiliary gas settings were 70 and 60 (arbitrary unit), respectively. The selected reaction monitoring (SRM) transition was m/z 130 → m/z 60 for metformin, and m/z 136 → m/z 60 for d6-metformin, the tube lens offset was 46 and collision energy was 10 V. The scan widths for both scanning quadrupoles were 1 amu and the scan time was 0.1 s for each transition.
M. Vath et al. / J. Chromatogr. B 967 (2014) 211–218 Table 1 Chromatographic conditions for the finalized metformin method. Analytical column Mobile phase A Mobile phase B Injection volume Start acquisition End acquisition Gradient
Hypercarb, 3 mm × 30 mm, 5 m 10 mM ammonium acetate in 95:5/water:acetonitrile 10 mM ammonium acetate in 95:5/acetonitrile:water 15 L to a 5 L loop 0.15 min 0.55 min
Time (min)
A (%)
B (%)
Flow rate (mL/min)
0 0.1 0.35 0.6 0.62 0.92
70 70 0 0 70 70
30 30 100 100 30 30
0.9 0.9 0.9 1 1 1
2.4. Preparation of standard and QC samples The standard curve of metformin at 0.5, 1, 5, 50, 500 and 1000 nM, and quality control samples at 1, 4, 40 and 400 nM were made in the blank matrix of MATE1 assay for analytical performance evaluation of the finalized LC–MS/MS method. All
213
the samples were then diluted with acetonitrile containing 50 nM internal standard of d6-metformin, and centrifuged at 3500 rpm for 10 min before the supernatant was subjected to LC–MS/MS analysis.
2.5. Assay analysis Substrate uptake in each well was determined by calculating the peak area ratio of metformin relative to d6-metformin internal standard in each well. Signal to background ratio (S:B) was determined from the average maximal control (uptake in the cell line expressing the transporter of interest in the presence of 0.5% DMSO, defined as 0% inhibition) and average minimal control (uptake in the parental cell line HEK293 in the presence of 0.5% DMSO, defined as background or 100% inhibition). The Z factor was calculated from formula: 1 − Z = (3 maximal + 3 minimal )/|maximal − minimal | where symbols refer to standard deviations () and average () of the maximal and minimal control values [16]. Percent inhibition was determined by normalizing samples relative to the maximal and minimal controls, and IC50 were determined using 4-parameter non-linear regression.
100 90
Relative Abundance
80 70 60
Metformin
50 40 30 20 10 0 100 90 80 70 60
Internal
50
Standard 40 30 20 10 0 15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Time (min) Fig. 2. Representative chromatograms of metformin (top) and internal standard d6-metformin (bottom) using finalized LC–MS/MS method on Aria LX-2 system, transporter assay sample analysis starting from high inhibitor concentration (left) to low inhibitor concentration (right).
214
M. Vath et al. / J. Chromatogr. B 967 (2014) 211–218
Table 2 Chromatographic methods attempted. Method
Column
Mobile phase A
Mobile phase B
Gradient
Method A Method B
10 mM ammonium bicarbonate Isopropanol
Acetonitrile 1:1 methanol:water
5–100% B in 0.1 min 95–50% B in 0.1 min, 0.9 min hold at 50% B
Isopropanol
1:1 methanol:water
95–50% B in 0.3 min, 3 min hold at 50% B
Method D
XBridge C18, 2 mm × 30 mm, 3 m Cogent Diamond Hydride, 2 mm × 20 mm, 3 m Cogent Diamond Hydride, 2 mm × 20 mm, 3 m Hypercarb 2 mm × 30 mm, 3 m
Acetonitrile
30–100% B in 0.25 min
Method E
Hypercarb 2 mm × 30 mm, 3 m
10 mM ammonium bicarbonate in water 10 mM ammonium bicarbonate in water
10 mm ammonium bicarbonate in 80:20 acetonitrile:water
30–100% B in 0.25 min
Method C
3. Results and discussion
illustrated in the representative chromatograms in Fig. 2, the method achieved symmetrical peak shapes for both metformin and the internal standard of d6-metformin. In contrast to the typically early elution of polar compounds in highly aqueous mobile phase on a reversed-phase column, the elution of metformin on the Hypercarb column occurred at 100% organic of the gradient, making it ideal for desolvation and ionization of the molecule in the mass spectrometer and resulting in enhanced sensitivity for LC–MS/MS analysis. The lifetime of the column, and the long term sustainability of the LC method is heavily dependent on the robustness of the column stationary phase packing. Since there is no modification to the graphite surface of the Hypercarb column, column bleeding and deterioration should be minimal. As expected, the column easily maintained its performance and back pressure
3.1. Method development In the finalized LC–MS/MS method, a porous graphitic carbon based Hypercarb column (3 mm × 30 mm, 5 m) was used. It has been demonstrated that this stationary phase provides great retention for polar compounds based on a unique retention mechanism, namely, charge induced interaction between solutes and the polarizable surface of graphite [17]. The method we developed using the Hypercarb column yielded adequate retention for metformin at a retention time of 0.65 min, corresponding to 4× of column void volume (Vm ) at 0.16 min, where Vm = 0.7 × × (i.d./2)2 × L, i.d. is the inner diameter of the column and L is the column length. As
inj. #1
inj. #2 0.6
100
inj. #3
1.2
inj. #4
0.1
inj. #5
3.4
2.4
5.6
3.9
2.8
1.7
inj. #6
4.5
6.1
5.0
Metformin
50
0
2.3
100
50
1.8 1.3
Inhibitor in inj. #1 2.1
1.6
2.5
0
2.9
100
Relativ e Abundance
50
3.2
3.4
Inhibitor in inj. #2
2.4 2.5 3.6
0
5.0
100 4.6
5.6
50
Inhibitor in inj. #3
5.3
4.7
6.1
0
6.0
100
6.2 5.7
50
Inhibitor in inj. #4
5.8
5.4
0
6.0
100
5.7
6.2
50
Inhibitor in inj. #5
5.9
0 100
Inhibitor in inj. #6 (not eluted)
50
0 0.0
0.5
1.0
1.5
2.0
2. 5
3.0
3. 5
4.0
4.5
5.0
5.5
6.0
Time (min) Fig. 3. Injections of transporter inhibition assay samples, using silica hydride column (Method B), simultaneous MRM monitoring of metformin and inhibitor compounds. Top most panel: extracted ion chromatogram of metformin, other panels: carryover of inhibitor compounds into later injections.
M. Vath et al. / J. Chromatogr. B 967 (2014) 211–218
inj. #1 100
inj. #2
inj. #3
inj. #4
215
inj. #5
inj. #6
15.2
0.1
7.7
3.9
19.0
11.4
Metformin
50 0 100
0.7
Inhibitor in inj. #1
50 0 100
4. 5
Relative Abundance
50
Inhibitor in inj. #2
0 100
8.3
50
Inhibitor in inj. #3 0 100
12.6
50
Inhibitor in inj. #4
13.1
0 100
15.8
50
Inhibitor in inj. #5
0 100
19.9
Inhibitor in inj. #6
50 0 0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Time (min) Fig. 4. Injections of transporter inhibition assay samples, using silica hydride column (Method C), simultaneous MRM monitoring of metformin and inhibitor compounds. Top most panel: extracted ion chromatogram of metformin, bottom panels: elution of inhibitor compounds in the same injection as metformin.
after over 7000 injections of transporter assay samples. The column re-equilibration time was observed to be longer than the typical reversed-phase column of same dimension to resume to the initial pressure after a gradient run, employment of a higher flow rate at 1.0 mL/min for the organic hold and column re-equilibration alleviated the issue. Prior to finalizing the LC–MS/MS method described above, various combinations of column, mobile phase and gradient were explored, as summarized in Table 2. Method development using reverse phase XBridge column (Method A) and ammonium bicarbonate mobile phases yielded symmetrical peak shape (chromatogram not shown). However, the retention of metformin at 0.1 min, was in the calculated column void volume at 0.16 min Applications using silica hydride column have demonstrated its high capacity factor for polar compounds though the retention mechanism is not fully understood [18–20]. The silica hydride column is amenable to both aqueous and organic mobile phase as the traditional reversed-phase column, yet retention of polar compounds increases with increasing organic content of the mobile phase, a phenomenon similar to the traditional normal phase whereby aqueous solvent is usually prohibited. The aqueous normal phase (ANP) mechanism of the silica hydride column provides some unique retention and selectivity advantages for polar compounds that are not achievable with conventional reversed-phase or normal-phase chromatography. When we set out to develop
the high throughput LC–MS/MS method for metformin on the silica hydride based Cogent Diamond Hydride column, we were able to easily achieve the symmetrical peak shape, adequate retention and great reproducibility for metformin spiked in blank transporter assay matrix (Method B, data not shown). However, the method showed deteriorated precision when samples generated from the transporter assays were injected, even with reduced injection volume and various organic mobile phases (methanol, acetonitrile and isopropanol). Considering that the only difference between the two sets of samples was inhibitor compounds present in the real assay samples while absent in the spiked samples, the SRM transitions of inhibitors in the assay samples were monitored along with metformin. As shown in Fig. 3, under these chromatographic conditions, the inhibitor compounds were all trapped in the column and carried over to the later injections. This carryover probably disrupted the interaction between metformin and the silica hydride surface, and resulted in the poor reproducibility observed. To confirm the hypothesis, an extended LC method was employed to elute all the test compounds in the same injection as metformin (Method C, Fig. 4). Using the extended method, the reproducibility test via injections of assay samples showed a %CV of less than 5% for metformin, a significant improvement over the short LC conditions. However, in achieving the elution of the interfering compounds, we had to extend the injection-to-injection cycle time from 1.1 min to 3.8 min, making it far less attractive as a high throughput method for transporter inhibition assay support.
216
M. Vath et al. / J. Chromatogr. B 967 (2014) 211–218
inj. #1
inj. #2
inj. #3
inj. #4
100 90
Relative Abundance
80 70 60 50 40 30 20 10 0 100 90 80 70 60 50 40 30 20 10 0 0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6
2.8
3.0
3.2
3.4
3.6
3.8
4.0
Time (min) Fig. 5. Chromatogram of metformin (top) and d6-metformin (bottom) using Method D, mobile phase A: 10 mM ammonium bicarbonate and mobile phase B: acetonitrile.
The method development on the porous graphitic carbonbased Hypercarb column started with a column dimension of 2 mm × 30 mm, 3 m, and a mobile phase A of 10 mm ammonium bicarbonate and B of acetonitrile (Method D). As shown in the representative chromatograms in Fig. 5, metformin peak shape was broad and splitting using the method. Since the retention of polar molecules on Hypercarb column is due to specific interaction between solutes and delocalized electrons on the graphite surface, it was hypothesized that this interaction was disrupted while the concentration of ammonium bicarbonate as electronic modifier was decreased during the gradient elution from 10 mM ammonium bicarbonate (mobile phase A) to acetonitrile (mobile phase B), resulting in the poor peak shape of metformin. To confirm the hypothesis, we replaced the mobile phase B with 80:20 acetonitrile: water containing same concentration of ammonium bicarbonate (10 mM) as in mobile phase A, and were able to achieve symmetrical peak shape while maintaining the needed retention. The yielded injection-to-injection cycle time of the method was 0.9 min; any further refinement of the method to shorten it was deemed unlikely, especially with the high back pressure the column already reached (260 bar at 0.5 mL/min). Though the initial attempt using Hypercarb column (2 mm × 30 mm, 3 m) was not successful in achieving an optimal
LC method, it led us to the trial of the same column packing with a larger inner diameter and particle size, namely 3 mm × 30 mm, 5 m column. As a result of this column choice, the column back pressure was lowered drastically (80 bar at 0.9 mL/min); in addition, the higher flow rate enabled by the new column allowed faster gradient and enhanced the method throughput by nearly 40%. To further facilitate the analytical process, ammonium bicarbonate mobile phase which is custom made in our lab, was replaced with batch-made ammonium acetate mobile phase, without compromising the analytical performance of the method.
3.2. Method performance evaluation The analytical performance of the finalized method is summarized in Table 3. The method demonstrated a good accuracy of 94–106% at all the concentration levels of the standard curve, representing a deviation ≤±6% from the nominal concentrations. The relative standard deviation (%CV) was ≤12% for all the concentrations except the lower limit of quantification (LLOQ) at 0.5 nM, whereby %CV was 16%. All the QC samples showed a %CV within 7% for intra-assay reproducibility, and a relative error ≤±11%.
M. Vath et al. / J. Chromatogr. B 967 (2014) 211–218
217
Fig. 6. IC50 curves of assay control compounds (S:B = 8.5, Z = 0.5) (a) dasatinib, (b) sunitinib, (c) imatinib and (d) pyrimethamine.
Table 3 Calibration curve and QC performance of metformin.
4. Conclusions
Standard Curve Performance Nominal concentration (nM)
Determined concentration (nM), Avg. n = 4
%CV
%Accuracy
0.5 1 5 50 500 1000
0.48 0.90 4.9 49 506 1063
16% 12% 6% 3% 7% 2%
96% 94% 97% 98% 101% 106%
QC Performance QC level
Determined concentration (nM), Avg. n = 4
%CV
% Accuracy
QC-1.0 QC-4 QC-40 QC-400
0.89 3.6 36 360
2% 7% 1% 5%
89% 90% 91% 90%
A high throughput LC–MS/MS method was successfully developed for the quantification of metformin uptake in transporter inhibition assays. The method included utilization of a porous graphitic carbon based Hypercarb column with a dimension of 3 mm × 30 mm, 5 m. Running the chromatography on a multiplexed LX-2 system coupled with multiple injection chromatography acquisition, the method yielded a fast injection-to-injection cycle time of 30 s. The method showed good precision, accuracy and robustness in analyzing transporter inhibition samples. The successful development and implementation of the LC–MS/MS method for metformin enabled the development and validation of transporter inhibition assays for MATE1, MATE2K and OCT2 using unlabeled metformin as a substrate. This is the first published high throughout LC–MS/MS method for metformin in support of transporter inhibition profiling assays; the application of this method can be extended to other bioanalytical areas requiring quantification of metformin. Appendix A. Supplementary data
Table 4 Comparison of IC50 values. Assay controls
IC50 determined (M) (n = 3)
Dasatinib Sunitinib Imatinib pyrimethamine
0.51 0.31 0.043 0.02
± ± ± ±
0.2 0.1 0.01 0.005
IC50, literature (M) 0.8 ± 0.2 [21] 0.28 ± 0.05 [21] 0.05 ± 0.005 [21] 0.04 ± 0.01 [22]
3.3. Transporter assay validation Substrate uptake in transporter inhibition assays was assessed by quantification of metformin uptake in HEK293 cells expressing the transporter of interest relative to uptake in the parental cell line, HEK293 cells. After establishing linearity of substrate uptake at the Km for metformin, a reaction endpoint was selected to maximize signal: noise (S:N) and assay robustness (Z > 0.5) (data not shown). Further assay validation was performed by assessing the ability of reference inhibitors to inhibit metformin uptake in this cell system. The typical IC50 curves for the assay control compounds are presented in Fig. 6. The IC50 values determined for the controls agreed well with the literature values (Table 4), validating the conversion to unlabeled transporter inhibition assays.
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jchromb. 2014.07.024. References [1] A.D. Wagner, J.M. Kolb, C.C. Özbal, J.J. Herbst, T.V. Olah, H.N. Weller, T.A. Zvyaga, W.Z. Shou, Rapid Commun. Mass Spectrom. 25 (2011) 1231. [2] S. Summerfield, P. Jeffrey, Expert Opin. Drug Discov. 4 (2009) 207. [3] P. Ballard, P. Brassil, K.H. Bui, H. Dolgos, C. Petersson, A. Tunek, P.J.H. Webborn, Drug Metab. Rev. 44 (2012) 224. [4] Y. Cheng, L.J. Martinez-Guerrero, S.H. Wright, R.K. Kuester, M.J. Hooth, I.G. Sipes, Drug Metab. Dispos. 39 (2011) 1755. [5] J.S. Yang, H.J. Oh, J.A. Jung, J.R. Kim, T.E. Kim, J.W. Ko, S.Y. Lee, W. Huh, Bull. Korean Chem. Soc. 34 (2013) 3284. [6] K. Sharma, G. Pawar, S. Yadam, S. Giri, S. Rajagopal, R. Mullangi, Biomed. Chromatogr. 27 (2013) 356. [7] S.L. Bonde, R.P. Bhadane, A. Gaikwad, A.S. Narendiran, B. Srinivas, Int. J. Pharm. Pharm. Sci. 5 (2013) 463. [8] Y. Chen, H.Q. Li, J.J. Xu, X.F. Jiu, C.H. Deng, X.G. Li, L. Li, X.Q. Xu, T.Y. Zhou, W. Lu, J. Chin. Pharm. Sci. 21 (2012) 211. [9] N. Sultana, M.S. Arayne, N. Shafi, F.A. Siddiqui, A. Hussain, J. Chromatogr. Sci. 49 (2011) 774. [10] W. Zeng, Y. Xu, M. Constanzer, E.J. Woolf, J. Chromatogr. B 878 (2010) 1817. [11] A. Liu, S.P. Coleman, J. Chromatogr. B 877 (2009) 3695. [12] Y. Hsieh, G. Galviz, J.J. Hwa, Bioanalysis 1 (2009) 1073. [13] L. Discenza, C. D’Arienzo, T. Olah, M. Jemal, J. Chromatogr. B 878 (2010) 1583. [14] W. Zhang, F. Han, H. Zhao, Z.J. Lin, Q.M. Huang, N. Weng, Biomed. Chromatogr. 26 (2012) 1163.
218
M. Vath et al. / J. Chromatogr. B 967 (2014) 211–218
[15] J. Zhang, M. Vath, C. Ferraro, Y. Li, K. Murphy, T. Zvyaga, H. Weller, W. Shou, Rapid Commun. Mass Spectrom. 27 (2013) 731. [16] J.H. Zhang, T.D.Y. Chung, K.R. Oldenburg, J. Biomol. Screen. 4 (1999) 67. [17] Thermo Scientific, Method Development Guide for Hypercarb Columns, http://www.thermoscientific.com [18] J.J. Pesek, M.T. Matyska, R.I. Boysen, Y. Yang, M.T.W. Hearn, TrAC – Trends Anal. Chem. 42 (2013) 64.
[19] J.J. Pesek, M.T. Matyska, Bioanalysis 4 (2012) 845. [20] J.J. Pesek, M.T. Matyska, J.A. Loo, S.M. Fischer, T.R. Sana, J. Sep. Sci. 32 (2009) 2200. [21] T. Minematsu, K.M. Giacomini, Mol. Cancer Ther. 10 (2011) 531. [22] B. Astorga, S. Ekins, M. Morales, S.H. Wright, J. Pharmacol. Exp. Ther. 341 (2012) 743.
Contents lists available at ScienceDirect
Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb
Development of an LC–MS/MS method for high throughput quantification of metformin uptake in transporter inhibition assays Marianne Vath, Lizbeth Gallagher, Wilson Shou, Harold Weller, Lisa Elkin, Jun Zhang ∗ Bristol Myers Squibb, Research and Development, 5 Research Parkway, Wallingford, CT 06492, United States
a r t i c l e
i n f o
Article history: Received 10 June 2014 Accepted 12 July 2014 Available online 24 July 2014 Keywords: Metformin LC–MS/MS Transporter inhibition Porous graphite carbon Silica hydride
a b s t r a c t A high throughput LC–MS/MS method for quantification of metformin substrate uptake enables conversion of radiometric transporter inhibition assays for multidrug and toxin extrusion transporters (MATE 1 and 2) and organic cation transporter 2 (OCT2) to a nonradioactive format. Such conversion greatly simplifies assay complexity and reduces assay costs. The development of a quantitative LC–MS/MS method for metformin in support of the high throughput transporter inhibition assays faced specific challenges of achieving both adequate chromatographic retention and rapid analytical turnaround. Here we report a method that circumvents both challenges. The utilization of a porous graphitic carbon column (Hypercarb) ensured adequate retention of highly polar metformin in biological samples. The combined employment of a ballistic gradient on a 3 mm × 30 mm, 5 m Hypercarb column, and dual staggered chromatography coupled with multiple injection chromatography acquisition, yielded a fast injection-to-injection cycle time of 30 s. The method demonstrated good accuracy, precision and excellent robustness for high throughput applications, and has been successfully implemented in the development and validation of the nonradioactive transporter inhibition assays for MATEs and OCT2. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Due to the potential impact of transporter inhibition on the clinical development of drug candidates, there has been a growing interest in predicting transporter related drug–drug interactions in the drug discovery setting [1–3]. Multidrug and toxin extrusion transporters 1 and 2 (MATEs), and organic cation transporter 2 (OCT2) are uptake transporters playing a key role in the disposition and renal clearance of cationic drugs and endogenous compounds, and both transporters exhibit overlapping substrate specificities for organic cations of metformin (Fig. 1). Uptake transporter inhibition assays have been developed to assess potential inhibitors of MATE and OCT2 transport by measuring the extent to which metformin uptake is inhibited by the co-administered drug candidate. These assays have typically been conducted using radiolabeled metformin and rely upon liquid scintillation counting (LSC) for sample analysis, which is a high throughput but expensive method, due to the cost of reagents and the associated special handling requirements [4]. It would therefore be advantageous to convert assays to a nonradioactive format if an appropriate label-free detection method
∗ Corresponding author. Tel.: +1 203 677 5610; fax: +1 203 677 6984. E-mail address: [email protected] (J. Zhang). http://dx.doi.org/10.1016/j.jchromb.2014.07.024 1570-0232/© 2014 Elsevier B.V. All rights reserved.
for metformin could be developed. In order to enable the conversion, we set out to develop a high throughput LC–MS/MS method for metformin quantification with a runtime of less than 1 min, to support these high throughput 384-well transporter inhibition assays. To our knowledge, this is the first report of such an assay and LC–MS/MS bioanalysis in the literature. There are two primary challenges associated with developing an LC–MS/MS method for metformin in support of transporter inhibition assays. Firstly, metformin is a highly polar compound (log P = −2.6), which presents a significant challenge in achieving adequate chromatographic retention. Secondly, analysis capacity and speed must be sufficient to support the large number of compounds studied for early profiling of transporter inhibition, which are screened in 10-point dose response in 384-well plate format. The LC–MS/MS method for substrate detection must match the sensitivity, speed and robustness of the LSC method in context of the above challenges. Metformin is the most widely prescribed anti-diabetic drug, used either alone or co-administered with other antihyperglycemic drugs for treatment of type-II diabetes. A very extensive effort has been seen in the development of LC–MS/MS methods for metformin, along with concomitantly administered drugs in order to evaluate possible pharmacokinetic interactions [5–8]. The main chromatographic methods reported are
212
M. Vath et al. / J. Chromatogr. B 967 (2014) 211–218
acetate, ammonium bicarbonate and dimethyl sulfoxide (DMSO) were also purchased from EMD Chemicals. 2.2. Transporter assay conditions
Fig. 1. Structures of metformin (a) and internal standard d6-metformin (b).
reversed-phase [6,7,9] and hydrophilic interaction liquid (HILIC) [10–12]. Most of the reported methods have LC runtimes too long (>5 min) to be considered suitable for in vitro assay support. Two HILIC methods, however, showed some potential for high throughput applications. Discenza et al. achieved good retention for both metformin and the concomitant drug in 2 min, via a combined mechanism of ion-exchange and reversed-phase chromatography on a HILIC column, by conducting a gradient elution from low organic to medium organic composition [13]. Zhang et al. reported a conventional HILIC method with a runtime of 3 min, using a gradient from high organic to low organic. In order to maintain good peak shape, however, they had to dilute the samples extensively by 38-fold using ACN following a protein precipitation of high solvent ratio (1:25/sample: ACN) [14]. Our attempt to apply HILIC methods for analysis of metformin from transporter inhibition assays quenched in ACN were not successful, due to shifting retention time and distorted peak shape in batch analyses, even at a reduced injection volume of 2 L (data not shown). In this study we evaluated both a porous graphitic carbon (Hypercarb) column under reversed-phase conditions, and silica hydride columns under aqueous normal phase (ANP) mechanism for high throughput metformin bioanalysis. To our knowledge, this is the first report of an LC–MS/MS method for metformin using either of these two columns. The suitability of each developed method was assessed in terms of chromatographic retention, peak shape and analytical throughput. For the final adopted method using the Hypercarb column, analytical performance such as accuracy, precision and robustness were also characterized. Using the final adopted chromatographic method, sample analysis for transporter inhibition assays was performed on a multiplexed LC–MS/MS platform we previously reported [15], to further reduce analytical turnaround. The platform uses dual staggered LC systems, and switching the mass spectrometric detection between the two systems for peak eluting chromatographic window only, minimizes the mass spectrometer idle time while maintaining the chromatographic quality as on a single LC system. Data acquisition of the multiplexed system is done in the fashion of multiple injection chromatography (MIC), whereby data for all the injections pertaining to a 384-well plate from the multiplexed LCs is acquired to a single file, thus eliminating the mass spectrometer reset delay. The platform also utilizes an open-deck manifold for sample holder to eliminate the auto-sampler overhead time spent on opening and closing sample drawers, resulting in a significant time saving for bioanalysis of a large number of samples. 2. Experimental 2.1. Chemicals and reagents Metformin hydrochloride and d6 metformin hydrochloride were purchase from Toronto Research Chemicals (Toronto, Ontario, Canada). HPLC grade water, acetonitrile, and methanol were purchased from EMD Chemicals (Gibbstown, NJ, USA). Ammonium
Transporter expressing cell lines were developed by subcloning cDNA for MATE1 (SLC47A1; NM 018242) into pIRESneo2 expression vector (Clontech, Mountain View, CA) and transfection into HEK293 cells (American Type Culture Collection, Manassas, VA, USA), which were maintained in Modified Eagle Medium (MEM) supplemented with 10% fetal calf serum (FCS) and selection antibiotics. Forty-eight hours prior to assay, MATE1-HEK293 cells were seeded at a density of 20,000 cells per well into columns 1–22 of a 384-well poly-d-lysine (PDL)-coated 384-well plate; the parental cell line (HEK293) was seeded at the same density into columns 23 and 24 of each plate for use as minimal signal controls (low control or 100% inhibition). Cells were grown for 48 h in MEM media containing 10% FBS at 37 ◦ C and 5% CO2 in a humidified incubator. Immediately prior to assay, media was removed and plates were washed in modified HBSS buffer (HBSS containing 10 mM HEPES, pH 8.4) using a BioTek 405 automated plate washer (BioTek, Winooski, VT). After complete aspiration of all wash buffer and addition of 20 L of modified HBSS, a LabCyte ECHO-550 noncontact acoustic dispenser (Sunnyvale, CA) was used to transfer 125 nL from each well of the compound source plate to a corresponding well of the cellular assay plate. After a 10 min pre-incubation in compound, 5 L of 25 M metformin in modified HBSS was added to each well of the assay plate using a MultidropTM Combi Reagent Dispenser (Thermo Scientific, Waltham, MA). Assay plates were incubated at room temperature for 40 min At the end of the incubation time, assay plates were aspirated and washed 3 times in modified HBSS using an automated plate washer. Plates were aspirated a final time and cells were lysed by addition of acetonitrile containing 50 nM d6 metformin as internal standard, and the supernatant was subjected to LC–MS/MS analysis. Assay procedures for transporter MATE2K and OCT2 were similar. 2.3. Instrument and LC–MS/MS conditions A modified Aria LX-2 multiplexed system (Thermo Scientific, CA, USA) was used as previously described [15]. The system consisted of two sets of independent 20AD binary pumps (Shimadzu Scientific Instruments, MD, USA) for dual staggered chromatography, a valve interface module (VIM, Thermo Scientific) for stream selection and two modified CTC HTS PAL auto-samplers (Leap Technologies, NC, USA). Sample plates in 384-well plate format were placed on a cooled, open-deck manifold for injection. Aria software (version 1.6.3, Thermo Scientific) controlled all peripheral components. A TSQ Quantum Ultra triple quadrupole mass spectrometer (Thermo Scientific) was used for mass spectrometric analysis. The MS was equipped with heated electrospray ionization (H-ESI) and used in positive ionization mode. XCalibur 2.2 was used for mass spectrometric data acquisition in the MIC mode. The final adopted LC–MS/MS method used a Hypercarb column (Thermo Scientific) with a dimension of 3 mm × 30 mm, 5 m. The chromatographic conditions are listed in Table 1. For MS/MS conditions, a universal set of source parameters was employed, including a spray voltage at 3.5 kV for positive ionization, a vaporizer temperature at 450 ◦ C and a capillary temperature at 275 ◦ C. The sheath gas and auxiliary gas settings were 70 and 60 (arbitrary unit), respectively. The selected reaction monitoring (SRM) transition was m/z 130 → m/z 60 for metformin, and m/z 136 → m/z 60 for d6-metformin, the tube lens offset was 46 and collision energy was 10 V. The scan widths for both scanning quadrupoles were 1 amu and the scan time was 0.1 s for each transition.
M. Vath et al. / J. Chromatogr. B 967 (2014) 211–218 Table 1 Chromatographic conditions for the finalized metformin method. Analytical column Mobile phase A Mobile phase B Injection volume Start acquisition End acquisition Gradient
Hypercarb, 3 mm × 30 mm, 5 m 10 mM ammonium acetate in 95:5/water:acetonitrile 10 mM ammonium acetate in 95:5/acetonitrile:water 15 L to a 5 L loop 0.15 min 0.55 min
Time (min)
A (%)
B (%)
Flow rate (mL/min)
0 0.1 0.35 0.6 0.62 0.92
70 70 0 0 70 70
30 30 100 100 30 30
0.9 0.9 0.9 1 1 1
2.4. Preparation of standard and QC samples The standard curve of metformin at 0.5, 1, 5, 50, 500 and 1000 nM, and quality control samples at 1, 4, 40 and 400 nM were made in the blank matrix of MATE1 assay for analytical performance evaluation of the finalized LC–MS/MS method. All
213
the samples were then diluted with acetonitrile containing 50 nM internal standard of d6-metformin, and centrifuged at 3500 rpm for 10 min before the supernatant was subjected to LC–MS/MS analysis.
2.5. Assay analysis Substrate uptake in each well was determined by calculating the peak area ratio of metformin relative to d6-metformin internal standard in each well. Signal to background ratio (S:B) was determined from the average maximal control (uptake in the cell line expressing the transporter of interest in the presence of 0.5% DMSO, defined as 0% inhibition) and average minimal control (uptake in the parental cell line HEK293 in the presence of 0.5% DMSO, defined as background or 100% inhibition). The Z factor was calculated from formula: 1 − Z = (3 maximal + 3 minimal )/|maximal − minimal | where symbols refer to standard deviations () and average () of the maximal and minimal control values [16]. Percent inhibition was determined by normalizing samples relative to the maximal and minimal controls, and IC50 were determined using 4-parameter non-linear regression.
100 90
Relative Abundance
80 70 60
Metformin
50 40 30 20 10 0 100 90 80 70 60
Internal
50
Standard 40 30 20 10 0 15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Time (min) Fig. 2. Representative chromatograms of metformin (top) and internal standard d6-metformin (bottom) using finalized LC–MS/MS method on Aria LX-2 system, transporter assay sample analysis starting from high inhibitor concentration (left) to low inhibitor concentration (right).
214
M. Vath et al. / J. Chromatogr. B 967 (2014) 211–218
Table 2 Chromatographic methods attempted. Method
Column
Mobile phase A
Mobile phase B
Gradient
Method A Method B
10 mM ammonium bicarbonate Isopropanol
Acetonitrile 1:1 methanol:water
5–100% B in 0.1 min 95–50% B in 0.1 min, 0.9 min hold at 50% B
Isopropanol
1:1 methanol:water
95–50% B in 0.3 min, 3 min hold at 50% B
Method D
XBridge C18, 2 mm × 30 mm, 3 m Cogent Diamond Hydride, 2 mm × 20 mm, 3 m Cogent Diamond Hydride, 2 mm × 20 mm, 3 m Hypercarb 2 mm × 30 mm, 3 m
Acetonitrile
30–100% B in 0.25 min
Method E
Hypercarb 2 mm × 30 mm, 3 m
10 mM ammonium bicarbonate in water 10 mM ammonium bicarbonate in water
10 mm ammonium bicarbonate in 80:20 acetonitrile:water
30–100% B in 0.25 min
Method C
3. Results and discussion
illustrated in the representative chromatograms in Fig. 2, the method achieved symmetrical peak shapes for both metformin and the internal standard of d6-metformin. In contrast to the typically early elution of polar compounds in highly aqueous mobile phase on a reversed-phase column, the elution of metformin on the Hypercarb column occurred at 100% organic of the gradient, making it ideal for desolvation and ionization of the molecule in the mass spectrometer and resulting in enhanced sensitivity for LC–MS/MS analysis. The lifetime of the column, and the long term sustainability of the LC method is heavily dependent on the robustness of the column stationary phase packing. Since there is no modification to the graphite surface of the Hypercarb column, column bleeding and deterioration should be minimal. As expected, the column easily maintained its performance and back pressure
3.1. Method development In the finalized LC–MS/MS method, a porous graphitic carbon based Hypercarb column (3 mm × 30 mm, 5 m) was used. It has been demonstrated that this stationary phase provides great retention for polar compounds based on a unique retention mechanism, namely, charge induced interaction between solutes and the polarizable surface of graphite [17]. The method we developed using the Hypercarb column yielded adequate retention for metformin at a retention time of 0.65 min, corresponding to 4× of column void volume (Vm ) at 0.16 min, where Vm = 0.7 × × (i.d./2)2 × L, i.d. is the inner diameter of the column and L is the column length. As
inj. #1
inj. #2 0.6
100
inj. #3
1.2
inj. #4
0.1
inj. #5
3.4
2.4
5.6
3.9
2.8
1.7
inj. #6
4.5
6.1
5.0
Metformin
50
0
2.3
100
50
1.8 1.3
Inhibitor in inj. #1 2.1
1.6
2.5
0
2.9
100
Relativ e Abundance
50
3.2
3.4
Inhibitor in inj. #2
2.4 2.5 3.6
0
5.0
100 4.6
5.6
50
Inhibitor in inj. #3
5.3
4.7
6.1
0
6.0
100
6.2 5.7
50
Inhibitor in inj. #4
5.8
5.4
0
6.0
100
5.7
6.2
50
Inhibitor in inj. #5
5.9
0 100
Inhibitor in inj. #6 (not eluted)
50
0 0.0
0.5
1.0
1.5
2.0
2. 5
3.0
3. 5
4.0
4.5
5.0
5.5
6.0
Time (min) Fig. 3. Injections of transporter inhibition assay samples, using silica hydride column (Method B), simultaneous MRM monitoring of metformin and inhibitor compounds. Top most panel: extracted ion chromatogram of metformin, other panels: carryover of inhibitor compounds into later injections.
M. Vath et al. / J. Chromatogr. B 967 (2014) 211–218
inj. #1 100
inj. #2
inj. #3
inj. #4
215
inj. #5
inj. #6
15.2
0.1
7.7
3.9
19.0
11.4
Metformin
50 0 100
0.7
Inhibitor in inj. #1
50 0 100
4. 5
Relative Abundance
50
Inhibitor in inj. #2
0 100
8.3
50
Inhibitor in inj. #3 0 100
12.6
50
Inhibitor in inj. #4
13.1
0 100
15.8
50
Inhibitor in inj. #5
0 100
19.9
Inhibitor in inj. #6
50 0 0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Time (min) Fig. 4. Injections of transporter inhibition assay samples, using silica hydride column (Method C), simultaneous MRM monitoring of metformin and inhibitor compounds. Top most panel: extracted ion chromatogram of metformin, bottom panels: elution of inhibitor compounds in the same injection as metformin.
after over 7000 injections of transporter assay samples. The column re-equilibration time was observed to be longer than the typical reversed-phase column of same dimension to resume to the initial pressure after a gradient run, employment of a higher flow rate at 1.0 mL/min for the organic hold and column re-equilibration alleviated the issue. Prior to finalizing the LC–MS/MS method described above, various combinations of column, mobile phase and gradient were explored, as summarized in Table 2. Method development using reverse phase XBridge column (Method A) and ammonium bicarbonate mobile phases yielded symmetrical peak shape (chromatogram not shown). However, the retention of metformin at 0.1 min, was in the calculated column void volume at 0.16 min Applications using silica hydride column have demonstrated its high capacity factor for polar compounds though the retention mechanism is not fully understood [18–20]. The silica hydride column is amenable to both aqueous and organic mobile phase as the traditional reversed-phase column, yet retention of polar compounds increases with increasing organic content of the mobile phase, a phenomenon similar to the traditional normal phase whereby aqueous solvent is usually prohibited. The aqueous normal phase (ANP) mechanism of the silica hydride column provides some unique retention and selectivity advantages for polar compounds that are not achievable with conventional reversed-phase or normal-phase chromatography. When we set out to develop
the high throughput LC–MS/MS method for metformin on the silica hydride based Cogent Diamond Hydride column, we were able to easily achieve the symmetrical peak shape, adequate retention and great reproducibility for metformin spiked in blank transporter assay matrix (Method B, data not shown). However, the method showed deteriorated precision when samples generated from the transporter assays were injected, even with reduced injection volume and various organic mobile phases (methanol, acetonitrile and isopropanol). Considering that the only difference between the two sets of samples was inhibitor compounds present in the real assay samples while absent in the spiked samples, the SRM transitions of inhibitors in the assay samples were monitored along with metformin. As shown in Fig. 3, under these chromatographic conditions, the inhibitor compounds were all trapped in the column and carried over to the later injections. This carryover probably disrupted the interaction between metformin and the silica hydride surface, and resulted in the poor reproducibility observed. To confirm the hypothesis, an extended LC method was employed to elute all the test compounds in the same injection as metformin (Method C, Fig. 4). Using the extended method, the reproducibility test via injections of assay samples showed a %CV of less than 5% for metformin, a significant improvement over the short LC conditions. However, in achieving the elution of the interfering compounds, we had to extend the injection-to-injection cycle time from 1.1 min to 3.8 min, making it far less attractive as a high throughput method for transporter inhibition assay support.
216
M. Vath et al. / J. Chromatogr. B 967 (2014) 211–218
inj. #1
inj. #2
inj. #3
inj. #4
100 90
Relative Abundance
80 70 60 50 40 30 20 10 0 100 90 80 70 60 50 40 30 20 10 0 0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6
2.8
3.0
3.2
3.4
3.6
3.8
4.0
Time (min) Fig. 5. Chromatogram of metformin (top) and d6-metformin (bottom) using Method D, mobile phase A: 10 mM ammonium bicarbonate and mobile phase B: acetonitrile.
The method development on the porous graphitic carbonbased Hypercarb column started with a column dimension of 2 mm × 30 mm, 3 m, and a mobile phase A of 10 mm ammonium bicarbonate and B of acetonitrile (Method D). As shown in the representative chromatograms in Fig. 5, metformin peak shape was broad and splitting using the method. Since the retention of polar molecules on Hypercarb column is due to specific interaction between solutes and delocalized electrons on the graphite surface, it was hypothesized that this interaction was disrupted while the concentration of ammonium bicarbonate as electronic modifier was decreased during the gradient elution from 10 mM ammonium bicarbonate (mobile phase A) to acetonitrile (mobile phase B), resulting in the poor peak shape of metformin. To confirm the hypothesis, we replaced the mobile phase B with 80:20 acetonitrile: water containing same concentration of ammonium bicarbonate (10 mM) as in mobile phase A, and were able to achieve symmetrical peak shape while maintaining the needed retention. The yielded injection-to-injection cycle time of the method was 0.9 min; any further refinement of the method to shorten it was deemed unlikely, especially with the high back pressure the column already reached (260 bar at 0.5 mL/min). Though the initial attempt using Hypercarb column (2 mm × 30 mm, 3 m) was not successful in achieving an optimal
LC method, it led us to the trial of the same column packing with a larger inner diameter and particle size, namely 3 mm × 30 mm, 5 m column. As a result of this column choice, the column back pressure was lowered drastically (80 bar at 0.9 mL/min); in addition, the higher flow rate enabled by the new column allowed faster gradient and enhanced the method throughput by nearly 40%. To further facilitate the analytical process, ammonium bicarbonate mobile phase which is custom made in our lab, was replaced with batch-made ammonium acetate mobile phase, without compromising the analytical performance of the method.
3.2. Method performance evaluation The analytical performance of the finalized method is summarized in Table 3. The method demonstrated a good accuracy of 94–106% at all the concentration levels of the standard curve, representing a deviation ≤±6% from the nominal concentrations. The relative standard deviation (%CV) was ≤12% for all the concentrations except the lower limit of quantification (LLOQ) at 0.5 nM, whereby %CV was 16%. All the QC samples showed a %CV within 7% for intra-assay reproducibility, and a relative error ≤±11%.
M. Vath et al. / J. Chromatogr. B 967 (2014) 211–218
217
Fig. 6. IC50 curves of assay control compounds (S:B = 8.5, Z = 0.5) (a) dasatinib, (b) sunitinib, (c) imatinib and (d) pyrimethamine.
Table 3 Calibration curve and QC performance of metformin.
4. Conclusions
Standard Curve Performance Nominal concentration (nM)
Determined concentration (nM), Avg. n = 4
%CV
%Accuracy
0.5 1 5 50 500 1000
0.48 0.90 4.9 49 506 1063
16% 12% 6% 3% 7% 2%
96% 94% 97% 98% 101% 106%
QC Performance QC level
Determined concentration (nM), Avg. n = 4
%CV
% Accuracy
QC-1.0 QC-4 QC-40 QC-400
0.89 3.6 36 360
2% 7% 1% 5%
89% 90% 91% 90%
A high throughput LC–MS/MS method was successfully developed for the quantification of metformin uptake in transporter inhibition assays. The method included utilization of a porous graphitic carbon based Hypercarb column with a dimension of 3 mm × 30 mm, 5 m. Running the chromatography on a multiplexed LX-2 system coupled with multiple injection chromatography acquisition, the method yielded a fast injection-to-injection cycle time of 30 s. The method showed good precision, accuracy and robustness in analyzing transporter inhibition samples. The successful development and implementation of the LC–MS/MS method for metformin enabled the development and validation of transporter inhibition assays for MATE1, MATE2K and OCT2 using unlabeled metformin as a substrate. This is the first published high throughout LC–MS/MS method for metformin in support of transporter inhibition profiling assays; the application of this method can be extended to other bioanalytical areas requiring quantification of metformin. Appendix A. Supplementary data
Table 4 Comparison of IC50 values. Assay controls
IC50 determined (M) (n = 3)
Dasatinib Sunitinib Imatinib pyrimethamine
0.51 0.31 0.043 0.02
± ± ± ±
0.2 0.1 0.01 0.005
IC50, literature (M) 0.8 ± 0.2 [21] 0.28 ± 0.05 [21] 0.05 ± 0.005 [21] 0.04 ± 0.01 [22]
3.3. Transporter assay validation Substrate uptake in transporter inhibition assays was assessed by quantification of metformin uptake in HEK293 cells expressing the transporter of interest relative to uptake in the parental cell line, HEK293 cells. After establishing linearity of substrate uptake at the Km for metformin, a reaction endpoint was selected to maximize signal: noise (S:N) and assay robustness (Z > 0.5) (data not shown). Further assay validation was performed by assessing the ability of reference inhibitors to inhibit metformin uptake in this cell system. The typical IC50 curves for the assay control compounds are presented in Fig. 6. The IC50 values determined for the controls agreed well with the literature values (Table 4), validating the conversion to unlabeled transporter inhibition assays.
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jchromb. 2014.07.024. References [1] A.D. Wagner, J.M. Kolb, C.C. Özbal, J.J. Herbst, T.V. Olah, H.N. Weller, T.A. Zvyaga, W.Z. Shou, Rapid Commun. Mass Spectrom. 25 (2011) 1231. [2] S. Summerfield, P. Jeffrey, Expert Opin. Drug Discov. 4 (2009) 207. [3] P. Ballard, P. Brassil, K.H. Bui, H. Dolgos, C. Petersson, A. Tunek, P.J.H. Webborn, Drug Metab. Rev. 44 (2012) 224. [4] Y. Cheng, L.J. Martinez-Guerrero, S.H. Wright, R.K. Kuester, M.J. Hooth, I.G. Sipes, Drug Metab. Dispos. 39 (2011) 1755. [5] J.S. Yang, H.J. Oh, J.A. Jung, J.R. Kim, T.E. Kim, J.W. Ko, S.Y. Lee, W. Huh, Bull. Korean Chem. Soc. 34 (2013) 3284. [6] K. Sharma, G. Pawar, S. Yadam, S. Giri, S. Rajagopal, R. Mullangi, Biomed. Chromatogr. 27 (2013) 356. [7] S.L. Bonde, R.P. Bhadane, A. Gaikwad, A.S. Narendiran, B. Srinivas, Int. J. Pharm. Pharm. Sci. 5 (2013) 463. [8] Y. Chen, H.Q. Li, J.J. Xu, X.F. Jiu, C.H. Deng, X.G. Li, L. Li, X.Q. Xu, T.Y. Zhou, W. Lu, J. Chin. Pharm. Sci. 21 (2012) 211. [9] N. Sultana, M.S. Arayne, N. Shafi, F.A. Siddiqui, A. Hussain, J. Chromatogr. Sci. 49 (2011) 774. [10] W. Zeng, Y. Xu, M. Constanzer, E.J. Woolf, J. Chromatogr. B 878 (2010) 1817. [11] A. Liu, S.P. Coleman, J. Chromatogr. B 877 (2009) 3695. [12] Y. Hsieh, G. Galviz, J.J. Hwa, Bioanalysis 1 (2009) 1073. [13] L. Discenza, C. D’Arienzo, T. Olah, M. Jemal, J. Chromatogr. B 878 (2010) 1583. [14] W. Zhang, F. Han, H. Zhao, Z.J. Lin, Q.M. Huang, N. Weng, Biomed. Chromatogr. 26 (2012) 1163.
218
M. Vath et al. / J. Chromatogr. B 967 (2014) 211–218
[15] J. Zhang, M. Vath, C. Ferraro, Y. Li, K. Murphy, T. Zvyaga, H. Weller, W. Shou, Rapid Commun. Mass Spectrom. 27 (2013) 731. [16] J.H. Zhang, T.D.Y. Chung, K.R. Oldenburg, J. Biomol. Screen. 4 (1999) 67. [17] Thermo Scientific, Method Development Guide for Hypercarb Columns, http://www.thermoscientific.com [18] J.J. Pesek, M.T. Matyska, R.I. Boysen, Y. Yang, M.T.W. Hearn, TrAC – Trends Anal. Chem. 42 (2013) 64.
[19] J.J. Pesek, M.T. Matyska, Bioanalysis 4 (2012) 845. [20] J.J. Pesek, M.T. Matyska, J.A. Loo, S.M. Fischer, T.R. Sana, J. Sep. Sci. 32 (2009) 2200. [21] T. Minematsu, K.M. Giacomini, Mol. Cancer Ther. 10 (2011) 531. [22] B. Astorga, S. Ekins, M. Morales, S.H. Wright, J. Pharmacol. Exp. Ther. 341 (2012) 743.
