Proguanil and cycloguanil are organic cation transporter and multidrug and toxin extrusion substrates

Human embryonic kidney (HEK293) cells transiently expressing (a combination of) human transporters of interest [organic cation transporter 1 or 2 (OCT1; SLC22A1 or OCT2; SLC22A2), multidrug and toxin extrusion protein 1 or 2-K (MATE1; SLC47A1 or MATE2-K; SLC47A2)], or enhanced yellow fluorescent protein (eYFP) were obtained from PharmTox (Nijmegen, The Netherlands).

BioCoat poly-d-lysine coated 24- and 96-wells plates were purchased from Becton–Dickinson B.V. (Breda, The Netherlands). Dulbecco’s Modified Eagle’s Medium (DMEM) + GlutaMAX-I, Hank’s balanced salt solution (HBSS) and 4-(4-(dimethylamino)styryl)-N-methylpyridinium iodide (ASP; ≥ 95% purity) were purchased from Life Technologies Europe B.V. (Bleiswijk, The Netherlands). Fetal bovine serum was purchased from Greiner Bio-One B.V. (Alphen a/d Rijn, The Netherlands). Sodium butyrate, amodiaquine dihydrochloride dehydrate (AQ; analytical standard), artemisinin (ART; 98% purity), atovaquone (ATO; ≥ 98% purity), chloroquine diphosphate (CQ; ≥ 98% purity), dihydroartemisinin (DHA; ≥ 97% purity), lumefantrine (LUM; ≥ 98% purity), mefloquine hydrochloride (MQ; ≥ 98% purity), primaquine diphosphate (PQ; 98% purity), proguanil hydrochloride (PG; ≥ 95% purity), pyrimethamine (PYR; 98.8% purity), quinine (QN; ≥ 98% purity) and 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) were purchased from Sigma-Aldrich Chemie B.V. (Zwijndrecht, The Netherlands). Cycloguanil (CG; 95% purity) was purchased from Aurum Pharmatech (Franklin Park, NJ, United States). N-methyl-quinidine (NMQ) was purchased from Solvo Biotechnology (Szeged, Hungary). Bovine serum albumin fraction V (BSA) was purchased from Roche Diagnostics Nederland B.V. (Almere, The Netherlands). The Victor X3 multimode plate reader was purchased from PerkinElmer Nederland B.V. (Groningen, The Netherlands). Bio-Rad protein assay was purchased from Bio-Rad laboratories Inc. (Veenendaal, The Netherlands). The acquity ultra performance liquid chromatography (UPLC) and Xevo TQ-S micro mass spectrometer was purchased from Waters (Milford, MA, USA). Finally, the HSS T3 analytical column and VanGuard HSS T3 pre-column were purchased from Waters (Dublin, Ireland).

HEK293 cells were modified to transiently express (a combination of) human OCT1, OCT2, MATE1, MATE2-K, or eYFP. Briefly, cDNA of the respective genes was cloned downstream of a CMV promoter into a baculovirus, of which passage three was used to transduce HEK293 cells. These cells were seeded in poly-d-lysine coated 24- or 96-wells plates and grown in a 37 °C incubator at 5% CO₂. Growth was initiated at ~ 25% confluency using 375 or 125 μL cell and complete medium (DMEM + GlutaMAX™-I supplemented with 10% fetal bovine serum) suspension per well, respectively. After 24 h, cells were transduced with 30 µL virus (10 µL virus + 20 µL complete medium in case of 96-wells plates) or 2 × 15 µL virus for cells that required simultaneous expression of two transporters. Finally, 195 µL sodium butyrate (45 µL for 96-wells plates) was added to a final concentration of 2 mM to enhance protein expression [17].

Concentration-dependent inhibition was assessed of anti-malarial compounds that inhibited ASP uptake by more than 67% in the initial screens. For each compound, a range of seven concentrations was used to determine the concentration at which ASP uptake was inhibited half maximally (IC₅₀). At least three independent experiments (4 for QN) were performed for each drug in triplicate as described above. Inhibition curves were plotted by nonlinear regression analyses of the data using GraphPad Prism version 5.03. Minimum uptake was set to be greater than 0% and the curve-fitted top within each experiment was adjusted to 100% as follows: each uptake percentage value was multiplied by 100/top in order to let the curve fit start at 100% at the lowest drug concentration. To generate the final dose–response inhibition curve per anti-malarial, percentage of uptake was averaged for each drug concentration per experiment and the resulting mean ± SEM corresponding to three or four independent experiments were plotted.

In order to predict whether future pharmacokinetic studies are recommended to study transporter inhibition in vivo, DDI indices were calculated by using the formula DDI index = fraction unbound * C_max/IC₅₀ with a cut-off value of 0.1 [18]. As C_max the (geometric) mean or median maximum plasma concentrations obtained from previous in vivo pharmacokinetic studies after therapeutic dosing was used.

HEK293 cells were seeded in 24-wells poly-d-lysine coated plates and transduced the subsequent day with OCT1 or OCT2 as described above. Three days following transduction, OCT1- or OCT2-mediated uptake of ten anti-malarials (AMO, ART, ATO, CQ, DHA, LUM, MQ, PQ, PG, and QN) was performed. NMQ served as a positive substrate for both OCT1 and OCT2 import [19]. In short, cells were washed with 400 µL 37 °C HBSS-HEPES buffer followed by adding 150 µL HBSS buffer with 10 µM anti-malarial or NMQ to each well. After incubation for 15 min at 37 °C, import was stopped by washing with 400 µL ice-cold HBSS-HEPES buffer supplemented with 0.5% BSA. Following a second wash with 400 µL ice-cold HBSS-HEPES buffer, cells were lysed with 200 µL 50% (v/v) MeOH and 0.1% (v/v) formic acid (HCOOH) in water and samples were sent for quantification by mass spectrometry analysis (see “LC–MS/MS quantification of anti-malarials” section). Anti-malarial import mediated by OCT1 or OCT2 was performed in two independent experiments in triplicate per drug and normalized to background (mock-transduced cells). Percentage of anti-malarial uptake (compared to background control) was plotted as mean (6 values) ± SEM in GraphPad Prism version 5.03 and drugs with at least a twofold increase versus background were selected for further analysis.

The affinity (K_M) and maximum transport rate (V_max) for OCT1 and OCT2 of PG were determined by analysing concentration-dependent uptake data from three independent experiments. The experimental setup was similar to the cellular import assay, except that a range of four concentrations (2.0, 3.9, 7.8 and 16 µM) and an incubation time of 1 min was used to determine the K_M of proguanil. The Michaelis–Menten curve was plotted as mean ± SEM in GraphPad version 5.03 after fixing the V_max of each individual experiment to 100% and subtracting mock-transduced background for each concentration used.

Import and export of PG and cycloguanil (CG) was assessed by single and double transduced HEK293 cells. In short, HEK293 cells seeded in 24-wells plates as described, were transduced separately with eYFP, OCT1, OCT2, MATE1 or MATE2-K, or with combinations of these viruses. Following the same procedure as for the cellular import assay described above, the import of 100 µM PG and CG was measured after 5 min incubation and data was plotted as mean pmol/mg/min ± SEM. For export (following OCT1 uptake) of 1 µM PG after 1 min of transport at pH 7.4, percentage of remaining cellular PG (compared to background) was plotted as mean ± SEM in GraphPad version 5.03. Significant reduction in uptake (i.e. export by MATEs) was determined using ln transformation and a one-way repeated measures ANOVA with Dunnet’s post test.

Anti-malarial concentration in the cell lysates was quantified using an LC–MS/MS system consisting of a UPLC, a binary solvent manager, a vacuum degasser and an autosampler, coupled to a Xevo TQ-S micro triple quadrupole mass spectrometer. Liquid chromatographic separation of the samples (stored at 10 °C; injection volume of 10 µL) was performed at 40 °C using a HSS T3 analytical column (1.8 μm; 100 × 2.1 mm) coupled to a VanGuard HSS T3 pre-column (1.8 µm; 5 × 2.1 mm). The mobile phase (run time: 10 min) consisted of solvent A [20 mM ammonium formate and 0.5% (v/v) formic acid (HCOOH) in water] and solvent B [0.5% (v/v) formic acid (HCOOH) in acetonitrile] using the following gradient with a flow of 200 µL/min: 0–0.5 min, 50% A; 0.5–4.5 min, 5% A; 4.5–10 min, 50% A. Electrospray ionization (ESI) was operated at a capillary voltage of + 1.0 kV and desolvation and source temperatures of 600 and 150 °C, respectively. Nitrogen was used as desolvation gas with a gas flow of 1000 L/h. Argon was used as collision gas at a pressure of 1.5 mTorr. Positive ion mode was used with selected reaction monitoring (SRM) for the quantitative analysis of AMO, ART, ATO, CQ, CG, DHA, LUM, MQ, PQ, PG, and QN. The peak area of the most abundant product ion was used for quantification. Ions with their corresponding mass/charge (m/z) ratio are shown in Table 1.

Using the baculovirus transduction system as described previously [11], OCT1 and OCT2 were transiently expressed in HEK293 cells. These cells were used to study the inhibitory effect of eleven anti-malarials (AMO, ART, ATO, CQ, DHA, LUM, MQ, PQ, PG, PYR, and QN) on OCT1- and OCT2-mediated ASP transport. Uptake of ASP by mock-transduced cells served as a negative control, which was used for background correction. Simultaneous incubation of 10 µM ASP, a known substrate of OCT1 [20] and OCT2 [21], and 50 µM anti-malarial drug (25 µM ATO and LUM, due to precipitation beyond these concentrations; 100 µM QN was used as positive control) for 10 min at 37 °C resulted in different inhibition profiles for OCT1 and OCT2. OCT1-mediated ASP transport in HEK293-transduced cells was only partially inhibited to 49 ± 3% by 100 µM QN (p < 0.05) and even stimulated by MQ to 150 ± 21% (p < 0.05) (Fig. 1a). Four other anti-malarials, AQ, PQ, PG, and PYR inhibited ASP uptake at a concentration of 50 µM for more than 30%, to a relative uptake of 67 ± 4, 59 ± 3, 60 ± 7, and 59 ± 6%, respectively, although these were non-significant inhibitions. In OCT2-transduced HEK293 cells, ASP uptake was reduced to 61 ± 13 and 60 ± 7% by ART and PQ, respectively. AQ, CQ, PG, PYR, and QN significantly reduced ASP uptake to 12.1 ± 0.6% (p < 0.001), 48 ± 11% (p < 0.05), 19 ± 7% (p < 0.001), 6 ± 2% (p < 0.001), and 3 ± 6% (p < 0.001), respectively (Fig. 1b).

The anti-malarials that most potently inhibited OCT1- and/or OCT2-mediated ASP import (more than 67%) were selected for further analysis by determining their IC₅₀-values. Using a larger concentration range, inhibition of OCT2-mediated transport of 10 µM ASP by AQ, PG, PYR, and QN was assessed using a similar set-up as above. Sigmoidal inhibition curves were generated by plotting ASP uptake rates as percentage of control versus increasing drug concentrations (Fig. 1c–f). This resulted in the strongest inhibition by PYR and QN with low micromolar IC₅₀ concentrations of 1.6 µM (95% CI 1.0–2.5 µM) and 3.4 µM (95% CI 1.1–11 µM), respectively. These were closely followed by IC₅₀-concentrations of 11 µM (95% CI 6–22 µM) and 13 µM (95% CI 10–17 µM) for AQ and PG, respectively. In order to assess the clinical relevance of these values, DDI indices were calculated, and a cut-off value of 0.1 was applied to indicate the requirement for future in vivo pharmacokinetic studies [18]. Based on reported C_max and unbound fraction values, only QN, with a DDI index of 1.1, fulfilled the criterion (Table 2).

Next, the potential substrates of OCT1 or OCT2 amongst the tested anti-malarials (AMO, ART, ATO, CQ, DHA, LUM, MQ, PQ, PG, and QN) were identified. Uptake of anti-malarials (10 µM) in HEK293 cells expressing these transporters was determined following 15 min incubation at 37 °C by LC–MS/MS after cell lysis. Uptake was represented as percentage of mock-transduced HEK293 control cells. Anti-malarials with more than twofold increase in uptake of control versus transporter cells were regarded as potential substrates. NMQ was used as a positive OCT1 and OCT2 control substrate. NMQ was taken up by OCT1 and OCT2 as expected. Only one anti-malarial, PG, showed an uptake of more than 200% of the control value. PG was taken up by OCT1 (301 ± 25% of control) as well as OCT2 (207 ± 17%) into HEK293 cells (Fig. 2a, b).

Next, the transport kinetics of PG uptake by OCT1 and OCT2 (eYFP-transduced cells served as background control) were studied. Concentration-dependent uptake of PG was measured during 1 min of incubation at 37 °C and its affinity (K_M) and maximum transport rate (V_max) were determined. Uptake of PG by OCT1 was characterized by a K_M of 8.1 ± 1.6 µM and V_max of 1840 ± 510 pmol/mg protein/min. PG affinity for OCT2 was similar with a K_M of 9.0 ± 1.1 µM and V_max of 4440 ± 1500 pmol/mg protein/min (Fig. 2c, d).

As PG is metabolized in the liver to its active metabolite CG by CYP2C19 [22], CG transport by OCT1 or OCT2 was investigated. Uptake studies of CG were performed similarly as with the previous anti-malarials, with a 5 min incubation time at 37 °C. This resulted in significant uptake by OCT1 and OCT2 into HEK293 cells (Fig. 2e).

As PG is likely imported into hepatocytes and proximal tubule cells via OCT1 and OCT2, it was tested whether this compound could be exported by the extrusion transporters MATE1 and MATE2-K. These transporters are able to export and import compounds depending on the physiological conditions. For this purpose, HEK293 cells were transduced with eYFP (mock), MATE1, or MATE2-K. Uptake was represented as percentage of eYFP control cells. Incubation of 100 µM PG for 5 min resulted in small significant uptake by MATE1 (Fig. 3a).

To study PG export capacity of MATE1 and MATE2-K, HEK293 cells were co-transduced with eYFP (mock) and OCT1, OCT1 and MATE1, or OCT1 and MATE2-K. Uptake was represented as percentage of eYFP control cells. Incubation of 1 µM PG for 1 min resulted in eightfold uptake by OCT1 (850 ± 60% of control) (Fig. 3b). This uptake was similar in cells expressing OCT1 and MATE1 (780 ± 70% of control) simultaneously. However, uptake was significantly reduced in cells expressing OCT1 and MATE2-K (470 ± 80% of control, p < 0.01), indicating that approximately 50% of PG was exported by MATE2-K as compared to eYFP/OCT1 transduced cells.

Finally, CG was analysed as a potential MATE1 and MATE2-K substrate by transfecting HEK293 cells with these transporters. Incubation of 100 µM CG for 5 min resulted in significant uptake by MATE1 (40 ± 9 pmol/mg/min) and MATE2-K (56 ± 15 pmol/mg/min) compared to control cells (11.8 ± 0.9 pmol/mg/min) (Fig. 3c). This highlights the involvement of MATE1 and MATE2-K in excretion of PG and its metabolite CG.