Atovaquone and quinine anti-malarials inhibit ATP binding cassette transporter activity

[6,7-³H(N)]Estrone-sulphate ammonium salt ([³H]-E1S, specific activity 45.6 Ci/mmol), Tauro[carbonyl-3H]Cholic Acid sodium salt ([³H]TCA) (5 Ci/mmol) and [6,7-³H(N)]Estradiol 17-β-D-glucuronide ([³H]-E₂17βG) (34.3 Ci/mmol) were purchased from PerkinElmer Life and Analytical Sciences (Groningen, Netherlands). [³H(N)]-methyl quinidine ([³H]-NMQ) (80 Ci/mmol) and unlabelled NMQ [N-methyl-quinidine] were purchased from Solvo Biotechnology (Szeged, Hungary). Bac-to-Bac and Gateway systems, Dulbecco’s modified Eagle’s medium, GlutaMAX-I culture medium, and foetal calf serum were purchased from Life Technologies (Bleiswijk, Netherlands). Primers were purchased from Biolegio (Nijmegen, Netherlands), and a plasmid purification midiprep kit was from Genomed (Löhn, Germany). Triple flasks (500 cm²) were purchased from Sanbio BV Biological Products (Uden, Netherlands). Estradiol 17-β-D-glucuronide (E₂17βG), estrone-sulphate (E1S), taurocholic acid (TCA) adenosine 5’-triphosphate magnesium salt (bacterial source), goat-anti-mouse IgG antibody IRDye 800 and goat-anti-rabbit Alexa 680 secondary antibodies, chloroquine (CQ), quinine (Q), artemisinin (ART), mefloquine (MQ), lumefantrine (L), atovaquone (ATO), dihydroartemisinin (DHA), and proguanil (PG) were purchased from Sigma-Aldrich (Zwijndrecht, Netherlands). Protein concentrations were determined with a Bio-Rad protein assay kit from Bio-Rad Laboratories (Veenendaal, Netherlands), and 96-well filter plates were purchased from Millipore (Etten-leur, Netherlands).

Human P-gp, BCRP, BSEP and MRP1-4 had previously been cloned into the Gateway pDONR221 vector. Sequences matched accession numbers NM_000927, NM_004827, NM_003742, NM_004996, NM_000392, NM_00378, and NM_005845 respectively [15‐19]. Some sequences did hold silent mutations of described polymorphisms. Gateway cloning was used to transfer the genes into a VSV-G improved pFastBacDual vector for mammalian cell transduction. The production of baculovirus was executed according to the Invitrogen Bac-to-Bac manual.

HEK293 cells were grown to 40% confluency in Dulbecco’s modified Eagle’s medium-GlutaMAX-I containing 10% foetal calf serum at 5% CO₂ in 500 cm² triple flasks. Culture medium was removed and 25 mL of medium combined with 10 mL virus was added and incubated at RT for 20 min, followed by the addition of another 40 mL of complete medium including 5 mM sodium butyrate to enhance protein expression.

Cells were harvested three days post transduction by a 5-min centrifugation step at 3,000 g. Cells were resuspended in ice-cold hypotonic buffer (0.5 mM sodium phosphate, 0.1 mM EDTA, pH 7.0) containing protease inhibitors (100 mM phenylmethylsulfonyl fluoride, 5 mg/ml aprotinin, 5 mg/ml leupeptin, 1 mg/ml pepstatin and 1 mg/ml E-64) and shaken at 4°C for 30 min. This lysate was centrifuged 100,000 xg for 30 min at 4°C, after which the pellet was homogenized in ice-cold TS buffer (10 mM Tris-HEPES and 250 mM sucrose, pH 7.4) supplemented with protease inhibitors described before using a tight-fitting Dounce homogenizer for 25 strokes. Two subsequent centrifugation steps at 4°C of firstly 20 min at 4,000 g followed by supernatant centrifugation for 60 min at 100,000 g ensured harvesting of the membrane fraction. The pellet was resuspended in ice-cold protease free TS buffer and passed 25 times through a 27-gauge needle to enhance membrane vesicle formation. Protein concentration in these vesicles was determined using the Bio-Rad protein assay, vesicles were flash-frozen in N₂ and stored at -80°C.

A rapid filtration technique that has been described earlier was applied to evaluate uptake of transporter specific substrates into the vesicles; NMQ for P-gp, E1S for BCRP, E₂17βG for MRP1-4 and TCA for BSEP [20]. Briefly, 0.015-0.15 μCi of labelled substrate was combined with unlabelled substrates to a concentration of 0.1-1 μM in a 30 μL reaction mixture with 4 mM ATP, 10 mM MgCl₂ and 7.5 μg total protein membrane vesicles in TS buffer. Transport was allowed by transfer of the plates to 37°C during 1–5 min, a time-point within the linear phase of time-dependent transport, as previously determined [15‐19]. Hereafter, the reaction was rapidly stopped by placing the plates back on ice and the addition of 150 μL ice-cold TS buffer. Samples were subsequently transferred to a 96-well filter plate that had been pre-incubated with TS buffer, and filtered using a multiscreen HTS-vacuum manifold filtration device (Millipore). Filters were washed and extracted, after which 2 mL scintillation fluid was added to each filter. Radioactive signal on the filters was determined by liquid scintillation counting. Negative controls included eYFP-transduced vesicles and AMP instead of ATP in the reaction mixture.

In the first screen, all anti-malarial compounds were added to the reaction mixture to evaluate transport inhibition at a concentration of 100 μM. Solvents were used as negative controls, as CQ was dissolved in milliQ, Q and ART in methanol, MQ, L, ATO and DHA in DMSO and PG in 50% ethanol. When ATP-dependent uptake was reduced more than 66.7%, the compound was considered a potential inhibitor, and multiple concentrations were tested in the reaction mixture to determine the IC₅₀ value. All concentrations were tested in duplicates or triplicates in two individual biological replicates containing vesicles of independent transductions. Results were depicted and statistically analysed using Graphpad Prism, version 5.03. IC₅₀ values were determined by nonlinear regression analysis of (log) inhibitor-response curves with variable slope. Maximal transport was restricted to 100%, and the minimum was set to be equal or greater than 0%. Statistical analysis was performed using IBM SPSS Statistics 20, applying one-way ANOVA (Analysis of variance).

The inhibitory characteristics against the ABC transporters of eight well-known anti-malarials; CQ, Q, ART, MQ, L, ATO, DHA and PG, was investigated at a 100 μM concentration. For each transporter protein, specific radio-labelled substrates were applied to measure ATP-dependent transport into the vesicular overexpression system; N-methyl quinidine (7nM radio-labelled diluted with 90 nM non-radio-labelled) for P-gp, estrone sulphate (74 nM) for BCRP, estradiol 17-β-D glucuronide (150 nM) for MRP1-4 and taurocholic acid (1 μM) for BSEP [16‐19].A significant inhibitory effect of 100 μM CQ, Q, MQ and PG on P-gp-mediated NMQ transport was observed. CQ reduced NMQ transport to 50% (p < 0.001) and PG to 76% (p < 0.001), whereas Q and MQ gave more pronounced inhibitory effects to 15% (p < 0.001) and 30% (p < 0.001) P-gp-mediated NMQ transport, respectively. ART and DHA slightly induced transport activity to 131% (p < 0.001) and 112% (p = 0.033), respectively (Figure 1A). All anti-malarials inhibited BCRP-mediated estrone sulphate transport activity at 100 μM concentrations. Most potent inhibitors were MQ, ATO and PG, which reduced estrone sulphate transport to 8.5%, 22% and 36% with p < 0.001, respectively (Figure 1B). CQ reduced transport to 69%, Q to 45%, ART to 62%, L to 44%, and DHA to 70% of solvent-exposed BCRP-mediated transport capacity (p < 0.001). Significant inhibition of taurocholic acid transport by BSEP was observed for ATO, which reduced uptake to 54% (p < 0.001) and MQ, which reduced uptake to 72% (p = 0.037). Furthermore, induction of BSEP transport activity was found for CQ (117%, p < 0.001), ART (117%, p < 0.001) and DHA (114%, p < 0.001) (Figure 1C). MQ was found to have a modest but significant inhibitory effect on estradiol 17-β-D glucuronide transport by MRP1 as this was reduced to 50% (p < 0.001), whereas ATO was observed to induce this process to 141% (p < 0.001) (Figure 1D). Induction was also observed for ART and ATO on MRP2-mediated estradiol 17-β-D glucuronide transport to 151% (p = 0.015) and 162% (p = 0.020), respectively. However, no significant inhibition was measured for any of the anti-malarials tested (Figure 1E). MRP3-mediated translocation of estradiol 17-β-D glucuronide was significantly inhibited by MQ at a 100 μM concentration to 70% (p = 0.001), whereas ART and DHA induced this process to 122% (p = 0.016) and 121% (p = 0.020), respectively (Figure 1F). No significant estradiol 17-β-D glucuronide transport inhibition of MRP4 could be detected (Figure 1F). As the 100 μM concentration is not within the physiological range of compound exposure, the most potent inhibitors were selected for further investigation. Inhibition of Q and MQ on P-gp-mediated transport, as well as BCRP inhibition by MQ, ATO and PG, were studied in more detail to determine their potencies.

Subsequently, transport inhibition assays were performed for a larger concentration range of Q, MQ, ATO and PG to evaluate P-gp or BCRP activity. Inhibition of transport was measured in a similar fashion applying the same specific radio-labelled substrates. Drug concentrations were logarithmically depicted, and a sigmoidal, inhibitor-response, variable slope equation was fitted to the data to determine the inhibition curve. Maximal inhibition to 0% transport was not always reached, which might be due to endogenous transport present in the vesicular membranes.

The strongest inhibitory effect for ATO on BCRP-mediated transport was found at median nanomolar range. Transport of estrone sulphate was inhibited with 50% by this compound at 0.23 μM (95% CI 0.17-0.29 μM) (Figure 2A), whereas MQ and PG required the addition of 18 μM (95% CI 17–20 μM) (Figure 2B) and 118 μM (95% CI 93–148 μM) (Figure 2C) to achieve a similar effect on BCRP activity, respectively. Also for the other compound-transporter combinations, IC₅₀ values were found in the low to median micromolar range. The effect of Q on P-gp-mediated NMQ transport inhibition was the strongest, and the IC₅₀ was defined at 6.8 μM (95% CI 5.9-7.8 μM) (Figure 2D). MQ was a less potent inhibitor with an IC₅₀ of 72 μM (95% CI 49–104 μM) (Figure 2E). The inhibitory concentration of ATO and Q transport were within the therapeutic range of blood plasma concentrations after both prophylactic and curative anti-malarial dosing.