The blood–brain barrier (BBB) physically and metabolically functions as a neurovascular interface between the brain parenchyma and the systemic circulation. Non-fenestrated capillary endothelial cells are the major component of the BBB that participate in regulating the permeability of several endogenous substrates and xenobiotics into and out of the CNS (Eyal et al. 2009). In addition, several membrane-associated drug efflux transporters have been characterized at the luminal membrane of brain capillary endothelial cells and serve as efflux pumps to extrude substrates (i.e., pharmacological agents) from the CNS back into the systemic circulation. In particular, P-Glycoprotein (P-gp), belonging to the ATP Binding Cassette membrane-associated transporter superfamily, is one of the most extensively studied drug transporters in brain capillary endothelial cells. This transporter has been well recognized to actively efflux many structurally diverse molecules including, steroids, environmental toxins and clinically prescribed drugs (e.g. anticancer, antiretroviral, antihypertensive, antiarrhythmic, antimicrobial agents and others) (Schinkel et al. 1996; Eyal et al. 2009) and its high expression in brain capillary endothelial cells is considered to be an essential mechanism which restricts brain entry of many pharmacological agents (Bendayan et al. 2002; Eyal et al. 2009). In addition, its expression at the BBB is well-recognized to be up-regulated during several neurological disorders [e.g. seizure activity (Tishler et al. 1995; Dombrowski et al. 2001)], pathological stimuli [e.g. pro-inflammatory cytokines (Bauer et al. 2007; Yu et al. 2008) and HIV viral proteins (Hayashi et al. 2005)], as well as chronic exposure to xenobiotics including drugs [e.g. ritonavir (Perloff et al. 2004; Zastre et al. 2009) and rifampin (Bauer et al. 2006)]. Among many molecular pathways, one mechanism which explains P-gp induction at the BBB involves agonist-activation of nuclear receptors, such as the Pregnane X Receptor (PXR) (Bauer et al. 2006; Narang et al. 2008; Chan et al. 2011). Upon agonist activation, PXR can serve as a factor which promotes transcription of MDR1 and mdr1a/1b genes which encode P-gp in humans and rodents, respectively (Geick et al. 2001; Cui et al. 2010). A wide array of structurally diverse molecules, such as steroid hormones, bile acids, herbal and dietary constituents as well as therapeutic agents, are known ligands of human and/or rodent PXR (Chang and Waxman 2006). Potent agonists of rodent PXR have been used to study, in vivo, P-gp induction at the BBB. For example, dexamethasone (DEX), a potent agonist of rodent PXR, has been utilized to induce P-gp expression in rodent brain capillaries (Bauer et al. 2004). Determination of the in vivo disposition of a drug in rodent brain is commonly performed via the bioanalysis of brain homogenate and/or CSF obtained by the terminal collection of whole brain and CSF samples at various time intervals following dosing of the drug (De Lange and Danhof 2002; Shen et al. 2004). However, drug concentrations from these samples may not always accurately predict the unbound drug concentrations in the brain extracellular fluid (ECF) which are primarily related to drug concentrations at the site of action (Lin 2008; Westerhout et al. 2012, 2013). Intracerebral microdialysis is a powerful technique that offers a continuous in vivo monitoring of drug concentrations in the brain ECF (De Lange et al. 1998, 2000; Chaurasia et al. 2007; Hammarlund-Udenaes et al. 2009). With the use of selective drug transporter substrates such as the antiarrhythmic drug quinidine, this technique has been utilized in rats to examine the role of P-gp at the BBB in vivo (Krajcsi et al. 2012; Syvänen et al. 2012; Westerhout et al. 2013). Quinidine has been identified as a P-gp substrate and inhibitor by many groups (Kusuhara et al. 1997; Emi et al. 1998; Giacomini et al. 2010). For example, the use of P-gp inhibitors, such as PSC-833 and LY-335979, resulted in an increase in quinidine brain accumulation by seven to ten-fold in rodents (Wang et al. 1996; Kusuhara et al. 1997; Starling et al. 1997). Moreover, P-gp knockout mice showed a 27–50 times increase in quinidine total brain to plasma concentration ratios compared to wild-type controls (Kusuhara et al. 1997; Kodaira et al. 2011; Uchida et al. 2011). In addition, several in vitro and in vivo studies suggest that P-gp constitutes the primary efflux mechanism for quinidine, while other major efflux drug transporters found at the BBB, such as breast cancer resistance protein and multidrug resistance-associated proteins show no interactions with quinidine (Kalvass et al. 2007; Kodaira et al. 2011; Sziráki et al. 2011). Cytochrome P450 (CYP) enzyme 3A4 in humans and its mouse ortholog Cyp3a11 can metabolize quinidine, while 10–50% of the drug is excreted unchanged in urine (Martignoni et al. 2006). To date, to the best of our knowledge, no information is available regarding the effect of ligand-mediated P-gp induction at the BBB on the brain distribution of quinidine. In this study, we utilized quantitative intracerebral microdialysis to examine quinidine distribution in the brain ECF of mice following P-gp induction at the BBB mediated by DEX treatment.
Intracerebral microdialysis was utilized to investigate the effect of P-glycoprotein (a drug efflux transporter) induction at the mouse blood–brain barrier (BBB) on brain extracellular fluid concentrations of quinidine, an established substrate of P-glycoprotein. Induction was achieved by treating male CD-1 mice for 3 days with 5 mg/kg/day dexamethasone (DEX), a ligand of the nuclear receptor, pregnane X receptor, and a P-glycoprotein inducer. Tandem liquid chromatography mass spectrometric method was used to quantify analytes in dialysate, blood and plasma. P-glycoprotein, pregnane X receptor and Cyp3a11 (metabolizing enzyme for quinidine) protein expression in capillaries and brain homogenates was measured by immunoblot analysis. Following quinidine i.v. administration, the average ratio of unbound quinidine concentrations in brain extracellular fluid (determined from dialysate samples) to plasma at steady state (375–495 min) or Kp, uu, ECF/Plasma in the DEX-treated animals was 2.5-fold lower compared with vehicle-treated animals. In DEX-treated animals, P-glycoprotein expression in brain capillaries was 1.5-fold higher compared with vehicle-treated animals while Cyp3a11 expression in brain capillaries was not significantly different between the two groups. These data demonstrate that P-gp induction mediated by DEX at the BBB can significantly reduce quinidine brain extracellular fluid concentrations by decreasing its brain permeability and further suggest that drug–drug interactions as a result of P-gp induction at the BBB are possible.
Applying microdialysis, distribution of quinidine, a P-gp substrate, in mouse brain extracellular fluid (ECF) was investigated following ligand-mediated P-glycoprotein (P-gp) induction at the blood–brain barrier (BBB). We demonstrated that a PXR agonist (dexamethasone) significantly up-regulated P-gp in brain capillaries and reduced quinidine brain ECF concentrations. Our data suggest that drug–drug interactions as a result of P-gp induction at the BBB are possible.
artificial cerebrospinal fluid
central nervous system
pregnane X receptor
selected reaction monitoring
Materials and methods
Chemicals and reagents
Quinidine hydrochloride monohydrate, DEX, corn oil, phenylmethanesulfonylfluoride and protease inhibitor cocktail were purchased from Sigma-Aldrich (St. Louis, MO, USA). Deuterated quinidine (quinidine-D3) was obtained from Toronto Research Chemicals (Toronto, ON, Canada). Lidocaine was provided as a generous gift from AB Sciex (Concord, ON, Canada). Bovine Fraction V Heat Shock Serum Albumin and Ficoll 400 were acquired from Bioshop (Burlington, ON, Canada). Sterile Dulbecco's Phosphate Buffered Saline (PBS) was purchased from Invitrogen (Grand Island, NY, USA). Murine monoclonal C219 antibody against P-gp was purchased from ID Laboratories (London, ON, Canada). Murine monoclonal C-4 (sc-47778) antibody against β-actin, goat polyclonal A-20 (sc-7737) antibody against mouse PXR and goat polyclonal L-14 (sc-30621) antibody against Cyp3a were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Horseradish peroxidase-conjugated anti-mouse and anti-goat antibodies were obtained from Jackson Immunoresearch Laboratories, Inc. (West Grove, PA, USA) and Sigma-Aldrich (Oakville, ON, Canada), respectively. Immunoblot stripping solution and enhanced chemilumescent solution were purchased from Pierce Thermo Fisher Scientific Inc. (Waltham, MA, USA). MilliQ™ water and reagent grade salts from Sigma-Aldrich were used for the preparation of artificial CSF (aCSF). All organic solvents used for the sample bioanalysis were of analytical grade.
Intracerebral microdialysis study
Adult male CD-1® mice (8–10 weeks old, 25–35 g) were purchased from Charles River Laboratories (St. Constant, QC, Canada). Mice were housed in the animal facility at NoAb BioDiscoveries Inc. (Mississauga, Ontario, Canada) and maintained on a 12 h light–dark cycle. Mice had access to water and to Lab Diet® 5015 Mouse Diet (Ren's Feed, Milton, ON, Canada) ad libitum. All procedures were reviewed by the NoAb BioDiscoveries Inc. Animal Care Committee and were performed in accordance with the principles of the Canadian Council on Animal Care. Newly arrived animals were acclimatized to their environment for at least 5 days prior to the insertion of jugular vein and carotid artery catheters under isoflurane anesthesia. The animals were allowed to recover for 1–2 days prior to the implantation of a microdialysis guide cannula and dummy probe (CMA Microdialysis AB, Stockholm, Sweden) in the striatum (AP 0.38, ML −1.80, DV −2.25). Before the microdialysis experiment, animals were administered subcutaneously with either DEX (5 mg/kg/day) or corn oil (vehicle control) for 3 days (n = 5 per group). One day after initiation of DEX treatment, a microdialysis guide cannula with a dummy probe was inserted into the striatum. The next day, the animals were allowed to acclimatize to individual RATURN cages (Bioanalytical Systems Inc. (BASi), West Lafayette, IN, USA). On that evening, the dummy probe was replaced with a 2 mm CMA/7 microdialysis probe (CMA Microdialysis AB) and the probe was equilibrated with aCSF (147 mM NaCl, 2.7 mM KCl, 1.2 mM CaCl2, 0.85 mM MgCl2) overnight. On the following morning, animals were administered quinidine prepared in saline (4 mg/kg i.v. loading dose and constant rate infusion of 0.8 mg/mL at 6 μL/min) via the jugular vein catheter and the microdialysis probes were perfused with aCSF containing 100 ng/mL quinidine-D3 at 0.4 μL/min using a 30–40 cm long Fluorinated Ethylene Propylene tubing (CMA Microdialysis AB) connected to a syringe pump (Harvard Apparatus, Holliston, MA, USA). Dialysate samples were collected at 30 min intervals over 510 min using a refrigerated fraction collector (BASi) at 4°C. Blood (15 μL) was serially collected using an automated sampling system (BASi) from the carotid artery catheter at 15, 45, 75, 165, 255, 375, 435 and 495 min following the start of quinidine infusion. Blood was diluted with heparinized saline (50 μL) after each sample collection and then stored at −80°C until tandem liquid chromatography mass spectrometric (LC-MS/MS) analysis. At the end of the experiment, animals were sacrificed to collect whole blood and plasma (obtained by centrifugation at 4°C) samples to determine quinidine concentrations. Whole brain tissues were harvested to evaluate protein expression in isolated brain capillaries and brain homogenates using immunoblot analysis. Probe perfusates were collected at the beginning and end of the experiment.
In vitro and in vivo recovery/loss from the microdialysis probes
Prior to the day of the in vitro recovery experiment, three CMA/7 microdialysis probes were first immersed in and perfused with aCSF solution overnight. On the following day, probes were immersed in a 1.5 mL reservoir of aCSF containing 100 ng/mL of quinidine at 37°C and perfused with aCSF containing 100 ng/mL of quinidine-D3 at 0.4 μL/min using Fluorinated Ethylene Propylene tubing (CMA Microdialysis AB). The dialysate was collected at 30 min intervals for 4 h and then stored at −20°C prior to analysis. The relative probe recovery and loss of quinidine and quinidine-D3, respectively, were calculated as described in Eqns (1) and (2).
The responsiveness of the dialysis system to a rapid change in quinidine and quinidine-D3 concentrations was also subsequently ascertained during a 2 h washout period by perfusing aCSF solution through the probes that were immersed in aCSF solution.
In vivo loss of quinidine and quinidine-D3 was compared in some of the animals 1 day after the microdialysis experiment. Following perfusion of the probes with aCSF solution overnight, a dialysate sample was collected to confirm the washout of the analytes and then the probes were perfused with a combination of 100 ng/mL quinidine and quinidine-D3 in aCSF solution. Dialysate was collected over 30 min intervals for 4 h and stored at −20°C prior to analysis.
Determination of quinidine unbound fractions in mouse plasma
The unbound quinidine fractions in mouse plasma were determined using a 96-well Teflon equilibrium dialysis apparatus utilizing dialysis membranes with a molecular cut-off at 12-14 kDa (HTDialysis LLC, Gales Ferry, CT, USA). Plasma was obtained from drug-naïve adult male CD-1 mice (8–10 week-old). The donor side of the dialysis chamber contained 300 μL of plasma containing 2100 ng/mL quinidine while the receiver side contained the same volume of blank PBS buffer (pH. 7.4). The dialysis apparatus was shaken gently (100 rpm) at 37°C for 16 h and equilibrium was observed after 12 h. A volume of 100 μL was collected from the receiver side and mixed with 50 μL of blank plasma. Meanwhile, 50 μL of plasma was collected from donor side and was mixed with 100 μL of blank PBS buffer. Samples were stored at −80°C prior to LC-MS/MS analysis.
Quinidine quantification in blood, plasma, perfusate, dialysate and dosing solution
Sample analysis was conducted using an AB Sciex API4000 LC-MS/MS system equipped with an Agilent 1100 series binary pump (Agilent Technologies, Santa Clara, CA, USA), solvent degasser, CTC autosampler and a Valco VICI divert valve. The Agilent 1100 mixer was replaced with an Upchurch® U466S mixer (Upchurch Scientific, Oak Harbor, WA, USA) to decrease the LC system volume. Dialysate samples and calibration standards containing quinidine and quinidine-D3 were prepared by diluting an aliquot (5 μL) with mobile phase containing lidocaine as an internal standard. Samples of dosing solution were diluted with mouse plasma and were analyzed to confirm the infusion solution concentration and determine the infusion rate. Protein precipitation in plasma and diluted mouse blood samples (10 μL) was performed using 50/50 methanol/acetonitrile containing quinidine-D3 (an internal standard) to extract the analyte from these samples. For the diluted blood samples, the supernatant of the protein precipitated mixture was dried down and reconstituted in the mobile phase prior to the LC-MS/MS analysis, whereas for the plasma samples, the supernatant was diluted in mobile phase prior to analysis. A Zorbax XDB-C18 column (2.1 × 30 mm) (Agilent Technologies) was utilized for the chromatographic separation and the analytes were eluted using a combination of 10 mM ammonium formate pH 3.0 in water (mobile phase A) and 95/5 (v/v) methanol/10 mM ammonium formate pH 3.0 in water (mobile phase B, MPB) at 0.7 mL/min flow rate. The analytes were trapped on the column with an initial isocratic condition of 10% MPB delivered over 0.4 min, and then eluted from the column using an increasing step gradient (0.1 min) to 40% MPB and then the latter condition was held for 0.6 min. The column was subsequently washed with 95% MPB prior to equilibration with 10% MPB. The cycle time (from injection to injection) was 2.7 min. Analytes were monitored using selected reaction monitoring in positive ion electrospray mode and quantified using peak area ratio of analyte to internal standard. The selected reaction monitorings for quinidine, quinidine-D3 and lidocaine were m/z 325 to 172 and 79, m/z 328 to 175 and 79, and m/z 235 to 86, respectively. Calibration curves were generated using Analyst™ software (AB SCIEX, Framingham, MA, USA) from at least six concentrations of standards prepared in aCSF (0.2–200 ng/mL), mouse plasma (10–10 000 ng/mL) and diluted mouse blood (10 to 10 000 ng/mL undiluted concentration).
Mouse brain capillary isolation and brain homogenate preparation
Procedures to isolate mouse brain capillaries were adapted from previously published methods with slight modifications (Bauer et al. 2004). In brief, fresh brains collected at the end of the experiment were rinsed with ice-cold PBS and stored on dry ice. The cortical gray matter was later isolated and homogenized at 400 rpm in ice-cold PBS. The mixture was centrifuged at 5800 g for 20 min at 4°C after the addition of ice-cold Ficoll 400 (final concentration 15%). The resulting pellet was resuspended in ice-cold PBS containing 1% Bovine Fraction V Heat Shock Serum Albumin and filtered through a 200 μm nylon mesh. The filtrate containing capillaries was passed over a glass bead column and washed with ice-cold PBS. The column effluent was collected to serve as capillary-depleted brain homogenate samples. Capillaries retained in the column were collected by agitation of the glass beads, then centrifuged (5 min of 500 g) and the pellet was washed three times in ice-cold PBS to remove impurities before snap-freezing in liquid nitrogen and storing at −80°C.
Protein expression of P-gp, PXR, Cyp3a11 and β-actin in isolated tissues was determined by sodium dodecyl sulfate–polyacrylamide gel electrophoresis according to previously published protocols from our group (Chan et al. 2011). Tissue lysates were prepared in lysis buffer (1% (v/v) NP-40 in 20 mM Tris, 150 mM NaCl, 5 mM EDTA at pH 7.5 containing 1 mM phenylmethanesulfonylfluoride and 0.1% (v/v) protease inhibitor cocktail), sonicated for 10 s and centrifuged at 20 000 g for 10 min at 4°C. Lysates containing 50 μg of protein from brain capillary and brain homogenate samples were mixed in Laemmli buffer and resolved on 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Mouse liver lysate (isolated from control non-surgical animals) containing 25–30 μg of protein was used as a positive control for P-gp, PXR and Cyp3a11 expression. After electrophoresis, gels were washed three times (5 min each) in transfer buffer (25 mM Tris-HCl, pH 8.0, 200 mM glycine) containing 20% (v/v) methanol and then electrotransferred onto polyvinylidene difluoride membranes. The membranes were blocked for 2 h in Tris-buffered saline containing 0.1% Tween 20 and with 5% (w/v) skim milk. The membranes were incubated with the appropriate primary antibody overnight at 4°C. On the next day, the membranes were washed three times (10 min each) in Tris-buffered saline containing 0.1% Tween 20 and were incubated with anti-mouse (Jackson immunoresearch Laboratories) or anti-goat (Sigma-Aldrich) horseradish peroxidase-conjugated secondary antibody at 1 : 3000 and 1 : 4000 dilution, respectively. P-gp expression was detected using a 1:500 dilution of mouse monoclonal C219 antibody (ID Laboratories), which recognizes an internal, highly conserved amino acid sequence VQEALD and VQAALD of P-gp. PXR expression was detected using a 1 : 100 dilution of goat polyclonal A-20 antibody (Santa Cruz Biotechnology, Inc.), which recognizes a unique domain of the mouse PXR.1 (protein accession # O75469). Cyp3a11 expression was detected using a 1 : 100 dilution of goat polyclonal L-14 antibody (Santa Cruz Biotechnology, Inc.), which recognizes the N-terminus of CYP3a11 (protein accession # Q64459). β-actin expression (protein accession # P60710) was detected using a 1 : 3000 dilution of mouse monoclonal C-4 antibody (Santa Cruz Biotechnology, Inc.). Protein bands were detected using an enhanced chemiluminescence kit. Densitometric analysis was performed using AlphaDigiDoc RT2 software (San Leandro, CA, USA) to quantify relative protein expression.
Brain ECF concentrations of quinidine were calculated by dividing each dialysate sample concentration of quinidine by the relative loss of quinidine-D3 determined for that sample. Blood sample concentrations were converted to plasma concentrations using the blood to plasma concentration ratio determined at the end of the in vivo microdialysis experiment for each animal. Quinidine unbound plasma concentration was estimated using experimentally determined unbound fraction in mouse plasma (0.08) as described above. In Fig. 2, averages of quinidine brain ECF concentration to unbound plasma concentration ratios ± SD at each time point were calculated from ratios obtained in five individual animals per treatment group. Each ratio per animal per time point was determined from the corresponding individual quinidine brain ECF concentration and unbound plasma concentration at the midpoint of the dialysate collection interval. The total plasma clearance of quinidine was estimated as the infusion rate divided by the steady state (375–495 min) plasma concentration. The Student's two-tailed t-test for unpaired data was used to determine statistical significance for the comparison of unbound quinidine concentrations in plasma and brain ECF, ratios of quinidine brain ECF concentration to unbound plasma concentration and protein expression in isolated brain capillaries and capillary-depleted brain homogenates between the vehicle-treated and DEX-treated groups. Data were analyzed by SPSS software (Chicago, IL, USA) and a p-value of < 0.05 was considered to be statistically significant.
A reliable recovery of quinidine is absolutely essential to correctly estimate quinidine brain ECF concentrations from the dialysate samples. The relative recovery for quinidine in these experiments was estimated by the relative loss of deuterium labeled quinidine (quinidine-D3). To estimate relative recovery using an internal calibrator, one must assume that relative drug recovery from the brain ECF surrounding the microdialysis probe is equal to the relative drug loss from the probe perfusate into the ECF. We first validated this assumption in vitro by confirming that the in vitro relative loss of quinidine-D3 (35.1 ± 3.1%) and relative recovery of quinidine (38.7 ± 5.3%) were essentially the same, with the use of three probes. During a 2 h wash-out period, dialysate concentrations of quinidine and quinidine-D3 decreased significantly within the first 30 min collection interval and then were essentially non-detectable over the remainder of the 2 h (data not shown), indicating that rapid changes in quinidine dialysate concentrations would be measurable. As well, in vivo relative loss of quinidine-D3 was 45.0 ± 7.2% and was found to be essentially identical to that of quinidine (44.8 ± 7.4%) during equilibrium, which was attained 150 min after the start of the perfusion. This was estimated in three of the animals after an overnight washout period following the cessation of the constant rate i.v. infusion.
Following a bolus dose (4 mg/kg) and continuous constant i.v. infusion (5 μg/min) of quinidine, unbound quinidine concentrations in plasma reached a steady state by 6 h in vehicle-treated animals (159.3 ± 5.2 ng/mL, n = 5) and DEX-treated animals (181.9 ± 3.0 ng/mL, n = 5, Fig. 1a). Blood concentrations were similar to plasma concentrations and the plasma to blood concentration ratio averaged 0.92 ± 1.4 and 0.86 ± 0.06 in vehicle- and DEX-treated animals, respectively. The estimated total plasma clearance of quinidine did not differ between the DEX- and vehicle-treated animals (DEX-treated group: 9.3 ± 1.8 L/kg/h and vehicle-treated group: 9.5 ± 1.5 L/kg/h). Brain ECF drug concentrations also reached a steady state by 6 h. However, quinidine brain ECF concentrations at steady state in DEX-treated animals (10.6 ± 1.6 ng/mL) were significantly lower than those of the vehicle-treated animals (24.7 ± 2.9 ng/mL, Fig. 1b). The average ratio of the unbound quinidine concentrations in brain ECF to that in plasma at steady state (375–495 min) or Kp, uu, ECF/Plasma in DEX-treated animals (0.063 ± 0.025) was approximately 2.5-fold lower than the ratio observed in vehicle-treated animals (0.149 ± 0.031, Fig. 2). Immunoblot analysis of the mouse brain capillary samples isolated from the animals treated with 5 mg/kg/day DEX for 3 days showed a significant induction (~ 1.5 fold) of P-gp protein expression when compared with vehicle-treated animals (Fig. 3). P-gp expression in mouse brain homogenate samples depleted of brain capillaries was also examined to investigate whether P-gp was induced in brain parenchyma, and was not found to be significantly different between DEX- and vehicle-treated groups (Fig. 4). In addition, expression of PXR protein in mouse brain capillaries and brain homogenates depleted of brain capillaries was not significantly different between the groups (Figs 5 and 6). As quinidine can be metabolized by mouse Cyp3a11 (Martignoni et al. 2006), the induction of Cyp3a11 by the mouse PXR agonist, DEX, could potentially also decrease brain ECF concentrations of quinidine. Therefore, Cyp3a11 expression in brain capillaries and brain homogenates was investigated in the DEX- and vehicle-treated animals. In this study, although Cyp3a11 expression was not detected in mouse brain capillary samples isolated from the two groups of animals (Fig. 7), Cyp3a11 was expressed in mouse brain homogenates, but no significant differences were observed between the two groups (Fig. 8).
Pharmacological mechanisms of CNS drugs, that is, those that bind to or modulate cellular membrane receptors as their therapeutic target, are primarily related to unbound drug concentrations in the brain ECF (Hammarlund-Udenaes 2010). Intracerebral microdialysis, the gold-standard approach to measure in vivo chemical concentrations in brain ECF, has been a useful tool to examine drug disposition in this compartment (De Lange et al. 1998, 2000; Chaurasia et al. 2007; Hammarlund-Udenaes et al. 2009). As drug concentrations in the brain ECF are affected by transport processes at the BBB, intracerebral microdialysis following the administration of substrates of drug transporters, can be used to study the in vivo role of drug transporters at the BBB (De Lange et al. 1998; Sawchuk and Elmquist 2000; Hammarlund-Udenaes et al. 2009; Krajcsi et al. 2012). For example, intracerebral microdialysis using quinidine as a selective P-gp substrate has been utilized to investigate P-gp function at the rat BBB (Liu et al. 2009; Sziráki et al. 2011; Syvänen et al. 2012). However, to our knowledge, quinidine distribution in the brain ECF of rodents following ligand-mediated P-gp induction at the BBB has not been previously examined. In this study, we utilized quantitative intracerebral microdialysis to investigate quinidine distribution in the brain ECF of mice following P-gp induction at the BBB mediated by a 3-day treatment with DEX, a potent rodent PXR agonist and P-gp inducer (Bauer et al. 2004; Chang and Waxman 2006). The average ratio of the unbound quinidine concentrations in brain ECF to that in plasma at steady state (375–495 min) or Kp, uu, ECF/Plasma in DEX-treated mice was significantly lower when compared with vehicle-treated animals (Fig. 2), indicating that quinidine distribution in the brain ECF was reduced in DEX-treated animals. In addition, P-gp expression in brain capillaries isolated from DEX-treated animals was significantly induced by 1.5-fold compared with vehicle-treated animals, while P-gp expression in brain homogenate samples depleted of brain capillaries remained unchanged between the two groups. As well, we did not observe a significant induction of Cyp3a11, which is believed to metabolize quinidine in mouse (Martignoni et al. 2006), neither in brain capillary nor in brain homogenate samples in DEX-treated animals when compared with vehicle-treated animals. These findings suggest that the reduction in brain ECF concentrations of quinidine as a result of DEX treatment is primarily because of the induction of P-gp at the mouse BBB, which ultimately results in enhanced efflux of quinidine from brain endothelial cells.
The use of a well-established P-gp substrate which is not known to interact with other major efflux drug transporters at the BBB is important to detect changes in P-gp function following its induction. Among many other lipophilic P-gp substrates, quinidine was chosen in this study because it does not significantly bind to the microdialysis probe and tubing, which would otherwise make the use of relative loss as an estimate of its relative recovery difficult (Chaurasia et al. 2007). Our in vitro and in vivo relative recovery/relative loss for quinidine and quinidine-D3 was essentially the same (approximately 45%), confirming that the deuterium labeled analog of quinidine can be utilized for the quantification of quinidine in the brain ECF. Furthermore, the relative recovery/relative loss of quinidine from the short (2 mm) microdialysis probe used in mice is relatively high and brain ECF concentrations of quinidine were high enough in the DEX-treated animals (Css = 10.6 ± 1.6 ng/mL) to be accurately quantified (the lower level of quantification of quinidine in ECF was 0.2 ng/mL).
There was no difference in the in vivo relative recovery of quinidine between the vehicle- and DEX-treated groups (40.9 ± 4.5 and 46.9 ± 9.5, respectively). This finding is consistent with a previous report demonstrating that the relative recovery of the analyte was not affected by inhibition of P-gp in rats (Sun et al. 2001). In this study, Kp,uu,ECF/Plasma of quinidine in vehicle-treated mice was 0.149 ± 0.031, which is approximately six times higher than the Kp,uu,Brain/Plasma obtained from rats in a study using the whole brain homogenate method (Kodaira et al. 2011). Interestingly, Liu et al., reported that the unbound quinidine brain concentrations at steady state determined by the brain homogenate method can underestimate the unbound quinidine ECF concentrations by approximately three-fold (Liu et al. 2009).
PXR, a xenobiotic-activating nuclear receptor, has been demonstrated to regulate P-gp expression in intestinal and hepatic tissues (Geick et al. 2001). Recently, several publications including ours have reported that P-gp is regulated by PXR in in vitro and ex vivo models of the BBB (e.g., human and rat brain microvessel endothelial cell culture systems and isolated rodent and porcine brain microvessels) (Bauer et al. 2004, 2006; Narang et al. 2008; Ott et al. 2009; Zastre et al. 2009; Chan et al. 2011). Several xenobiotics such as DEX, identified to serve as ligands of PXR, have been utilized to examine the inductive role of this nuclear receptor in rodents (Jones et al. 2000). Although DEX is a known anti-inflammatory agent, it is utilized as a P-gp inducer and PXR ligand in our current study. The microdialysis probe-related brain injury is anticipated to be minimal in our current protocol and we do not anticipate that a significant inflammatory reaction is produced (Benveniste and Diemer 1987; De Lange et al. 1995). The DEX dosage of 5 mg/kg/day used to treat our animals was selected based on previous in vivo studies performed in mice which showed induction of PXR targets (e.g., P-gp and Cyp3a11) (Bauer et al. 2004; Scheer et al. 2010). In this study, we confirmed the protein expression of P-gp and PXR in brain capillaries and demonstrated that treatment of mice with DEX (5 mg/kg/day for 3 days) resulted in a 1.5-fold P-gp induction in mouse brain capillaries. This finding is in agreement with previous published study where the same dosing regimen of DEX induced P-gp expression by two-fold in rat brain capillaries (Bauer et al. 2004). It is important to note that although PXR is functionally expressed in in vitro and in vivo BBB models in rodents, both the transcript and protein expression of human PXR have not been previously detected in human brain capillaries (Dauchy et al. 2008; Shawahna et al. 2011). However, human PXR transcript expression was reported by another group in specific regions of the human brain such as thalamus, pons and medulla (Nishimura et al. 2004; Miki et al. 2005). As well, our laboratory has demonstrated PXR protein expression in human fetal brain tissue (Chan et al. 2010). Our current findings provide in vivo evidence that a PXR pathway may regulate P-gp functional expression at the mouse BBB. However, since DEX is also a known glucocorticoid receptor ligand, this pathway cannot be excluded. Further studies are needed to fully elucidate, in vivo, the potential role of these nuclear receptors in the regulation of P-gp at the BBB.
To date, no information is available regarding the effect of ligand-mediated P-gp induction at the BBB on the brain distribution of a P-gp substrate. However, a few studies have examined the effect of P-gp induction at the BBB on P-gp substrate distribution in the brain ECF in different rodent models of disease states. For instance, Bauer et al. reported very interesting data showing that the anti-nociceptive effect of methadone (a P-gp substrate) was reduced in mice exhibiting P-gp induction in brain capillaries, although methadone concentrations in the brain were not determined (Bauer et al. 2006). As well, Wu et al. showed that P-gp induction in the brain capillaries of an inbred type II diabetic mouse model could lead to significantly lower concentrations of Rhodamine 123, an established P-gp substrate, in brain ECF compared with wild-type mice (Wu et al. 2009). These observations along with our current findings support the concept that concentrations of P-gp substrates in brain ECF are expected to be reduced following P-gp induction at the BBB and this could be explained by the increased clearance of substrates from the brain ECF back into the circulation. On the other hand, Bankstahl and Löscher demonstrated that while pilocarpine-induced epileptic rats showed a higher P-gp expression in brain capillaries, these animals exhibited higher phenytoin (a weak P-gp substrate) concentrations in brain ECF from the hippocampus when compared to non-treated controls as determined by non-quantitative microdialysis (Bankstahl and Löscher 2008). As well, a kainate-induced epileptic rat model, which was previously demonstrated to have a higher P-gp expression in brain capillaries when compared with non-treated animals, exhibited no significant change in quinidine permeability across the BBB but higher brain ECF concentrations of quinidine in hippocampus regions (Syvänen et al. 2012). Interestingly, these authors also reported a decrease in total brain concentrations of quinidine following drug-induced epileptic condition, however, this effect did not reach significance. Although, this finding may appear to contradict our results, Syvänen et al. had also suggested that neuro-inflammation in brain parenchyma and disease-mediated BBB dysfunction could occur during epileptic states and that these factors, in addition to P-gp induction at the BBB, glial cells and neurons, could affect quinidine distribution in total brain and the ECF compartment (Syvänen et al. 2012). Therefore, these discrepancies may be explained by differences in the mechanism and cellular locations by which P-gp induction was mediated between the epileptic rat models and our wild-type mouse model that was triggered with a P-gp inducer systemically.
Our group has previously demonstrated that, in addition to brain capillaries, P-gp is expressed in different cellular compartments of brain parenchyma such as astrocytes and microglia (Lee et al. 2001; Ronaldson et al. 2004; Bendayan et al. 2006). The function of P-gp in brain parenchyma could potentially result in the efflux of quinidine from these cellular compartments into brain ECF for further distribution in the brain or removal across the BBB. Therefore, alterations of P-gp expression in these cellular compartments could potentially affect quinidine concentrations in the brain ECF. In this study, samples of mouse brain homogenate depleted from brain capillaries were used for the measurement of P-gp protein expression in brain parenchyma. Expression of P-gp in brain homogenates did not significantly differ between DEX- and vehicle-treated mice suggesting that the lower quinidine concentrations observed in brain ECF of DEX-treated animals was likely not a result of P-gp induction in brain parenchyma.
It is well recognized that the cellular expression and activity of CYP enzymes in brain parenchyma and microvessel endothelial cells is significantly lower relative to their functional expression in hepatocytes (Woodland et al. 2008). Low expression implies that CYP enzymes may not participate substantially in the overall brain drug clearance. However, the potential that drug disposition in brain ECF could be affected by these enzymes still exists. Furthermore, the expression of human CYP3A4 and mouse Cyp3a11 enzymes, which can metabolize quinidine in the liver (Martignoni et al. 2006), are known to be regulated by PXR and can be induced by DEX in rodents (Perloff et al. 2004). Although little is known on Cyp3a11 induction mediated by PXR in the brain, this potential effect at the BBB and in brain parenchyma could reduce quinidine concentrations in brain ECF. In this study, Cyp3a11 expression was not detected in mouse brain capillaries; these data are consistent with previous findings demonstrating that CYP3A4 is not present in human brain capillaries (Dauchy et al. 2008). Furthermore, Cyp3a11 induction in brain homogenates of DEX-treated animals was not observed. Together, these data indicate that the lower quinidine ECF concentrations observed in the DEX-treated animals mice were not a result of Cyp3a11 induction at the BBB or in brain parenchyma.
In summary, this study shows that intracerebral microdialysis utilizing quinidine and an internal calibrator (quinidine-D3) can be used to assess P-gp induction mediated by a P-gp inducer/PXR agonist (DEX) at the mouse BBB. Our findings demonstrate that P-gp induction mediated by DEX at the BBB can reduce quinidine concentrations in the brain ECF. These data support the concept that P-gp induction can further restrict the brain permeability of its substrates and illustrate the major role of P-gp at the BBB in protecting the brain against xenobiotics including therapeutic agents. These data further suggest that drug–drug interactions as a result of P-gp induction at the BBB are possible.
The authors thank Mrs. Tennile Tavares (NoAb BioDiscoveries Inc., Mississauga, Canada) for the preparation of the samples for LC/MS/MS analysis and Jianghong Fan for performing quinidine plasma binding experiments (NoAb BioDiscoveries Inc., Mississauga, Ontario, Canada, present address: InterVivo Solutions, Toronto, Ontario, Canada). This research was funded in part by a grant from the Canadian Institutes of Health Research (CIHR Grant # MOP56976) awarded to Dr. Reina Bendayan and in part by NoAb BioDiscoveries Inc. Dr. Bendayan is a recipient of a Career Scientist Award from the Ontario HIV Treatment Network, Ministry of Health of Ontario, Canada. Mr. Gary Chan was a recipient of a Natural Sciences and Engineering Research Council doctoral scholarship award.
Chan, de Lannoy and Bendayan participated in research design. Chan, Saldivia, Yang and Pang conducted experiments. Chan, Pang, de Lannoy and Bendayan performed data analysis and wrote or contributed to the writing of the manuscript.
Conflicts of interest