Abstract
As a result of the growing availability of genetically engineered mouse lines, the pilocarpine post–status epilepticus (SE) model of temporal lobe epilepsy is increasingly used in mice. A discrepancy in pilocarpine sensitivity in FVB/N wild-type versus P-glycoprotein (PGP)–deficient mice precipitated the investigation of the interaction between pilocarpine and two major multidrug transporters at the blood-brain barrier. Doses of pilocarpine necessary for SE induction were determined in male and female wild-type and PGP-deficient mice. Brain and plasma concentrations were measured following low (30–50 mg⋅kg−1 i.p.) and/or high (200 mg⋅kg−1 i.p.) doses of pilocarpine in wild-type mice, and mice lacking PGP, breast cancer resistance protein (BCRP), or both transporters, as well as in rats with or without pretreatment with lithium chloride or tariquidar. Concentration equilibrium transport assays (CETA) were performed using cells overexpressing murine PGP or BCRP. Lower pilocarpine doses were necessary for SE induction in PGP-deficient mice. Brain-plasma ratios were higher in mice lacking PGP or PGP and BCRP, which was also observed after pretreatment with tariquidar in mice and in rats. Lithium chloride did not change brain penetration of pilocarpine. CETA confirmed transport of pilocarpine by PGP and BCRP. Pilocarpine is a substrate of PGP and BCRP at the rodent blood-brain barrier, which restricts its convulsive action. Future studies to reveal whether strain differences in pilocarpine sensitivity derive from differences in multidrug transporter expression levels are warranted.
Introduction
The pilocarpine-induced status epilepticus (SE) model is the most widely used rodent model of both SE itself and chronic temporal lobe epilepsy (TLE) developing as a long-term consequence of SE. In humans, SE is a life-threatening emergency, often refractory to anticonvulsive drug treatment (Treiman et al., 1998; Mayer et al., 2002; Young, 2006), resulting in a mortality rate of 40% or higher (Young, 2006), and is often followed by severe neuropathological consequences. Chronic TLE, the most frequent type of epilepsy, is associated with pharmacoresistance in more than two-thirds of cases (Semah et al., 1998). Post-SE animal models of TLE, such as the pilocarpine model, provide insight into pathologic mechanisms of epileptogenesis and pharmacoresistant epilepsy (Löscher and Brandt, 2010). The pilocarpine model is extensively applied in rats and, more recently with increasing frequency, in mice (Schauwecker, 2012). In this model, SE induction via systemic or intracerebral administration of the cholinergic muscarinic agonist pilocarpine serves as an initial insult for provoking epileptogenesis. Importantly, the model reflects many features of human TLE in terms of seizure characteristics and histopathological findings (Curia et al., 2008). Furthermore, the pilocarpine model is highly valuable for studying consequences of seizures and SE induced by cholinergic agents in chemical weapons and for developing counteractions (Tetz et al., 2006; de Araujo Furtado et al., 2012). Pilocarpine is also clinically relevant as a topical treatment of glaucoma (Lee and Higginbotham, 2005), and local or systemic treatment of dry mouth (xerostomia; Visvanathan and Nix, 2010).
Since the first description of the model by Turski et al. (1984), different protocols have been developed for systemic SE induction by pilocarpine. In many laboratories, SE is induced via injection of a single high pilocarpine dose. This approach is often accompanied by a high mortality rate and does not always lead to SE induction in a satisfactory proportion of animals (Turski et al., 1989; Glien et al., 2001; Buckmaster and Haney, 2012; Mazzuferi et al., 2012). Using a fractionated protocol with repetitive injection of smaller pilocarpine doses can improve the outcome in terms of SE induction rate and low mortality in rats and in mice (Glien et al., 2001; Gröticke et al., 2007; Schauwecker, 2012). Additionally, lithium chloride is often administered the evening before the pilocarpine experiment, which drastically reduces the required pilocarpine dose for SE induction, thereby increasing the proportion of animals surviving SE (Clifford et al., 1987; Bersudsky et al., 1994; Shaldubina et al., 2007).
Unfortunately, SE induction and survival rates can dramatically differ not only between rodent strains (Chen et al., 2005; Winawer et al., 2007a,b), but also between animals of the same strain bred by different vendors and even between inbred mice originating from different breeding rooms of the same vendor (Borges et al., 2003; Bankstahl et al., 2009, 2012; Müller et al., 2009). The mechanisms underlying this variable efficacy of pilocarpine are unknown. In our laboratory, we assessed the SE induction protocol in a batch of rodents from new strains or vendors before starting the main experiment. By this means, we found evidence that mice lacking the multidrug transporter P-glycoprotein [PGP; Mdr1a/b(−/−) mice] required lower doses of pilocarpine for SE induction than wild-type (WT; FVB/N) mice. Multidrug transporters are expressed at tissue barriers throughout the body, such as the blood-brain barrier, and serve as protective mechanisms, transporting a wide spectrum of xenobiotic compounds (Löscher and Potschka, 2005b). Since the substrate spectrum of the most widely investigated multidrug transporter PGP considerably overlaps with that of breast cancer resistance protein (BCRP) (Löscher and Potschka, 2005a), we hypothesized that pilocarpine is transported by each of these transporters at the blood-brain barrier in mice and rats, thereby influencing the uptake of pilocarpine by the brain, and, consequently, its convulsant action. In vivo experiments with genetic knockout or pharmacological inhibition of multidrug transporters were performed and validated by in vitro transport assays using cells overexpressing murine PGP or BCRP.
Materials and Methods
Animals.
For the present study, 187 adult mice of both sexes [n = 98 FVB/N WT mice, n = 77 Mdr1a/b(−/−) mice, n = 6 Bcrp1(−/−) mice, n = 6 Mdr1a/b(−/−)Bcrp1(−/−) mice], and 18 adult male Sprague-Dawley rats (Harlan, Horst, The Netherlands) were used. Breeding pairs of the four mouse genotypes were kindly provided by the group of Dr. A. Schinkel (The Netherlands Cancer Institute, Amsterdam, The Netherlands), and offspring were used in this study. Twenty-one of the WT mice were obtained from Charles River (Sulzfeld, Germany) and used for the pharmacokinetic experiment. Mice and rats were housed in groups separated by sex under controlled environmental conditions (22 ± 1°C, 40–70% humidity) with a 12-hour light/dark cycle (lights on at 6:00 AM) and ad libitum access to food and water. All experiments were performed in accordance with the European Communities Council Directive of November 24, 1986 (86/609/EEC) and the German Law on Animal Protection (“Tierschutzgesetz”). Ethical approval for the study was granted by an ethical committee (according to §15 of the Tierschutzgesetz) and the government agency (Lower Saxony State Office for Consumer Protection and Food Safety; LAVES) responsible for approval of animal experiments in Lower Saxony. All efforts were made to minimize pain or discomfort in the animals used.
Chemicals and Cell Lines.
Pilocarpine hydrochloride, lithium chloride (LiCl), and methylscopolamine bromide were purchased from Sigma-Aldrich (Taufkirchen, Germany) and dissolved in injectible water. The PGP/BCRP inhibitor tariquidar dimesylate was kindly provided by Xenova Ltd. (Slough, UK) and freshly dissolved prior to each administration in 5% (w/v) aqueous dextrose solution. Administration volume was 10 mg⋅kg−1 for all compounds in mice, and 2 ml⋅kg−1 (LiCl, tariquidar) or 3 ml⋅kg−1 (pilocarpine) in rats. For the in vitro experiments, tariquidar, the BCRP inhibitor Ko143 hydrate [(3S,6S,12aS)-1,2,3,4,6,7,12,12a-octahydro-9-methoxy-6-( 2-methylpropyl)-1,4-dioxopyrazino[1′,2′:1,6]pyrido[3,4-b]indole-3-propanoic acid 1,1-dimethylethyl ester hydrate] (Sigma-Aldrich) and unlabeled digoxin (Sigma-Aldrich), as a reference substrate for PGP, were dissolved in dimethyl sulfoxide (<0.1% DMSO in final solution). The PGP inhibitor PSC833 [(3S,6S,9S,12R,15S,18S,21S,24S,30S,33S)-6,9,18,24-tetraisobutyl-3,21,30-triisopropyl-1,4,7,10,12,15,19,25,28-nonamethyl-33-[(2R,4E)-2-methyl-4-hexenoyl]-1,4,7,10,13,16,19,22,25,28,31-undecaazacyclotritriacontane-2,5,8,11,14,17,20,23,26,29,32-undecone] (kindly provided by Novartis Pharma GmbH, Nürnberg, Germany) was dissolved in ethanol/Tween 80 (9:1) and then diluted with double-distilled water (1:2). [14C]Mannitol and [3H]digoxin (specific activity: 1.85–2.22 kBq/μmol and 0.555–1.48 GBq/μmol, respectively) were obtained from Hartmann Analytic GmbH (Braunschweig, Germany).
For in vitro transport studies, Lilly Laboratories cell (LLC)-PK1 cells transduced with murine Mdr1a and respective WT LLC cells were used. Madin-Darby canine kidney type II (MDCK-II) cells transduced with murine Bcrp1 and respective WT MDCK-II cells were also used. Culturing of the cells was performed as described previously (Baltes et al., 2007). Both cell lines are widely used and were recommended by the US Food and Drug Administration for identifying PGP substrates (Food and Drug Administration, 2006). The sole use of these cell lines as a model system does not reflect all properties of the complex composition of a blood-brain barrier, and therefore, does not allow direct translation to the in vivo situation. In this study, cell assays were used to confirm the in vivo data.
SE Induction by Pilocarpine in Mice.
In total, 53 (21 males, 32 females) WT mice and 65 (25 males, 40 females) Mdr1a/b(−/−) mice were used for SE induction experiments. To reduce peripheral side effects of pilocarpine, mice were pretreated with 1 mg⋅kg−1 i.p. methylscopolamine 30 minutes before the first injection of pilocarpine. Based on a prior fractionated protocol (Gröticke et al., 2007), the pilocarpine dose necessary to induce self-sustaining SE was adapted to the mouse genotypes by intraperitoneal administration of an initial bolus (100–200 mg⋅kg−1) and up to nine subsequent injections of 50–150 mg⋅kg−1 every 20 or 30 minutes. Animals were continuously monitored, and time of occurrence as well as number and severity [referred to Racine’s scale (Racine, 1972)] of each single seizure prior to SE was recorded. SE onset was defined as continuous ongoing seizure activity following occurrence of one or two generalized tonic-clonic seizures [stage 4 or 5 on the Racine scale (Racine, 1972)]. SE was characterized by head nodding in an upright (sitting) body position, Straub tail, and slight to moderate convulsions of forelimbs, sometimes interrupted by further generalized tonic-clonic seizures (see Bankstahl et al., 2012). Behavioral seizure activity and SE duration was confirmed by individual EEG recordings in a subset of animals (for methodological details, see Klein et al., 2015). For EEG recordings, animals were implanted with bipolar electrodes in the dorsal hippocampus [anteroposterior (AP) 1.8 mm, laterolateral (LL) 1.6 mm, and dorsoventral (DV) 2.0 mm relative to bregma] as described elsewhere (Gröticke et al., 2008).
Pharmacokinetic Profile of Pilocarpine in Mice.
Pilocarpine was administered intraperitoneally at a subconvulsive dose of 50 mg⋅kg−1 to WT mice, which were closely observed for potential occurrence of seizures. After 5, 10, 15, 20, 30, 60, and 120 minutes, mice were deeply anesthetized with carbon dioxide (n = 3 per time point) and blood was sampled by retro-orbital puncture into vials preloaded with 5 mM EDTA. After cervical dislocation, brains were immediately removed, weighed, and homogenized after addition of double distilled water (∼1 ml per 100 mg tissue), and frozen in aliquots of 200 μl. Blood samples were centrifuged (12,000 rpm, 3 minutes; room temperature) and the supernatant plasma was frozen. All samples were stored at –30°C until analysis. Plasma elimination half-life (t1/2) was estimated by PK Solutions 2.0 (Summit Research Services, Montrose, CO). Animals were not perfused before taking brain samples, as this can lead to a washout effect and falsely low brain concentrations (unpublished data). Thus, we cannot exclude that brain concentrations are overestimated owing to remaining pilocarpine in the brain vasculature, but as all animals in this study were treated identically, group comparisons are expected to be reliable.
Brain-Plasma Ratio of Pilocarpine in Mice and Rats with and without Tariquidar and LiCl Pretreatment.
Fifty milligrams per kilogram pilocarpine were administered intraperitoneally to all four mouse genotypes (n = 6 per group). A separate group of WT mice (n = 6) received an injection of 15 mg⋅kg−1 tariquidar intravenously 1 hour before pilocarpine for complete inhibition of PGP and partial inhibition of BCRP (Bankstahl et al., 2013). A further group (n = 6) were pretreated with 10 mEq⋅kg−1 (423.3 mg⋅kg−1) LiCl intraperitoneally 13 hours before pilocarpine (Bersudsky et al., 1994) to evaluate whether LiCl increases brain levels of pilocarpine as suggested previously (Marchi et al., 2009). Interestingly, the potency of LiCl to lower convulsive doses of pilocarpine is much higher in rats than in mice (Turski et al., 1989). Therefore, in the present study, we used LiCl doses known to successfully potentiate pilocarpine effects (10 mEq⋅kg−1 in mice and 3 mEq⋅kg−1 in rats). Further WT (n = 6) and Mdr1a/b(−/−) mice (n = 6) were treated with a convulsive dose of pilocarpine (200 mg⋅kg−1). All mice were closely observed for occurrence of seizures. Twenty minutes following pilocarpine injection, blood samples were taken and brains excised and stored as described above.
Three subsets of rats (n = 6 per group) were treated intraperitoneally with 30 mg⋅kg−1 pilocarpine and various pretreatments: untreated, pretreated with 15 mg⋅kg−1 tariquidar intravenously 1 hour before pilocarpine to achieve complete PGP inhibition (Kuntner et al., 2010), or pretreated with 3 mEq⋅kg−1 (127 mg⋅kg−1) LiCl by mouth 13 hours before pilocarpine (Glien et al., 2001). Rats were closely observed for occurrence of seizures. Thirty minutes following pilocarpine injection, blood and brains were sampled and stored as described above.
Concentration Equilibrium Transport Assays.
We employed concentration equilibrium transport assays (CETA), in which the influence of passive diffusion is minimized by adding the drug at equal concentrations to the apical and the basolateral compartments of the Transwell system (Luna-Tortós et al., 2008; Löscher et al., 2011). CETA were performed in triplicate as described previously (Luna-Tortós et al., 2008). LLC cells were seeded with a density of 0.3 × 106 cells cm−2 on transparent polyester membrane filters (Transwell, six-well, 24-mm diameter, 0.4-μm pore size; Corning Costar Corporation, Cambridge, MA). MDCK-II cells were seeded with a density of 0.4 × 106 cells cm−2 on translucent polyester membrane filters (ThinCert, six-well, 24.85-mm diameter, 0.4-μm pore size; Greiner Bio-One, Frickenhausen, Germany). Transport experiments were performed 5–7 days after the cells reached 100% confluence. At 1 hour preincubation, culture medium was replaced by serum-free Opti-MEM with or without inhibitor (10 μM PSC833 for PGP inhibition or 0.1 μM Ko143 for BCRP inhibition). For experiments with LLC cells, Ko143 (0.1 μM) was added to inhibit endogenous BCRP, and for experiments with MDCK-II cells, tariquidar (0.2 μM) was added to inhibit endogenous PGP (Kühnle et al., 2009). The subsequent incubation was performed with pilocarpine in fresh Opti-MEM on the apical and basolateral sides of the monolayer at a concentration of 25 μM. After an initial distribution phase, samples were taken from both compartments after 60, 120, 240, 360, 480, and 600 minutes and the amount of pilocarpine was quantified by high-performance liquid chromatography (HPLC) analysis. The results were calculated as percentage of initial concentration. The calculated values after the distribution phase may be associated with a loss of compound, which may relate to residual adherence to the walls or membrane of the inserts (Luna-Tortós et al., 2008). As control experiment, CETA were performed with [3H]digoxin (diluted with unlabeled digoxin to result in an activity of 1.85 kBq/ml and a final concentration of 10 nM) as a reference substrate. The membrane integrity of monolayers was determined by [14C]mannitol diffusion and transepithelial electrical resistance, following exclusion criteria described earlier (Luna-Tortós et al., 2008). [3H]Digoxin and [14C]mannitol were measured in disintegrations per minute (DPM) with a β-scintillation counter (MicroBeta Trilux; Wallac/PerkinElmer LAS GmbH, Rodgau, Germany).
Determination of Pilocarpine in Brain, Plasma, and Transport Assay Samples.
Concentrations of pilocarpine were determined via HPLC with UV detection. The chromatographic system was composed of a LC-6A pump (Shimadzu, Duisburg, Germany) and a UV detector (SPD-6A; Shimadzu) set at 213 nm. Separations were obtained on a C18 column (Refill Nucleosil 120–5 C18; Macherey and Nagel, Düren, Germany) preceded by a guard column (Nucleosil 120–5 C18; Knauer, Berlin, Germany). The mobile phase consisted of potassium dihydrogen phosphate buffer (0.05 M) and acetonitrile (70:40, v/v), and 0.1% triethylamine (final pH of 7.5). The flow rate was 1 ml⋅min−1, and the injections were carried out via a 50-μl loop. Data processing was handled by CBM-20A software (Shimadzu). For sample preparation, plasma (50 μl) or cell medium (40 μl) were mixed with 150 μl (plasma) or 60 μl (cell medium) of acetonitrile, followed by 1 minute of agitation and 5 minutes centrifugation (13,200 rpm, room temperature). The upper layer was removed and stored at –30°C until measurement. Homogenized brain samples (200-μl aliquots) were mixed with 200 μl of acetonitrile, followed by 5 minutes of agitation and 15 minutes centrifugation (15,000 rpm, 4°C). The supernatant was centrifuged in 0.2-μm filters (6000 rpm, 4°C) and the lower layer was used for injection into the HPLC apparatus. Mean recovery for pilocarpine was 99.8% in plasma, 100.7% in brain, and 102.8% in cell medium. The quantification limit was 200 ng/ml for plasma, 1.1 μg⋅g−1 for brain, and 625 ng⋅ml−1 for cell medium. Retention times under these conditions were 6.0 minutes in plasma samples and 6.5 minutes in brain and cell medium samples.
Statistical Analysis.
Statistical analysis was performed using Prism 5 software (GraphPad Software Inc., La Jolla, CA). For comparison of two groups, Student’s t test was used. In the case of more than two groups, a one-way analysis of variance followed by Bonferroni’s multiple comparison test was performed. Fisher’s exact test was used to assess sex differences in SE induction and mortality rates, respectively. CETA data were statistically analyzed by two-way analysis of variance followed by Bonferroni test to compare replicate means. All tests were used two-tailed, and a P value <0.05 was considered statistically significant. Unless stated otherwise, all values are given as mean ± S.E.M.
Results
SE Induction in Mdr1a/b(−/−) Mice Requires Lower Pilocarpine Doses Than in WT Mice.
SE induction in mice was performed by fractionated pilocarpine injection. As described previously for C57BL/6 and NMRI mice (Gröticke et al., 2007; Bankstahl et al., 2012), SE development was preceded by immobility, head nodding, Straub tail, tremor, chewing, and at least one generalized clonic-tonic seizure (type 4 or 5 on the Racine scale). SE itself was characterized by continuous limbic seizure activity (head nodding, forelimb convulsions, infrequent generalized tonic-clonic seizures). Convulsive activity characteristics and SE-associated EEG patterns were comparable in mouse genotypes. In 40% of EEG-monitored mice, SE was self-limiting before diazepam administration. In animals with self-limiting SE, EEG analysis revealed that SE duration was shorter in WT mice than in Mdr1a/b(−/−) mice (53.0 ± 11.2 minutes, n = 4, versus 83.5 ± 3.0 minutes, n = 4; P = 0.039).
Both in female and male WT mice, the pilocarpine dose necessary to induce SE was higher compared with Mdr1a/b(−/−) mice (female, 465 ± 32 mg⋅kg−1 versus 295 ± 12 mg⋅kg−1, P < 0.0001; male, 340 ± 22 mg⋅kg−1 versus 271 ± 15 mg⋅kg−1, P = 0.013; Fig. 1). Interestingly, female WT mice required higher pilocarpine doses compared with male WT mice (P = 0.030), whereas this sex difference was not found in PGP-deficient mice. SE induction rates were comparable between female and male WT (female, 71.9% versus male, 71.4%, P = 1.00) as well as female and male Mdr1a/b(−/−) mice (female, 77.5% versus male, 76.0%, P = 1.00). No differences were found between mouse genotypes in the number of stage 3–to–stage 5 seizures prior to reaching SE [stage 3, 0.39 ± 0.12 versus 0.16 ± 0.066; stage 4, 1.7 ± 0.24 versus 1.5 ± 0.18; stage 5, 1.1 ± 0.18 versus 1.2 ± 0.19, WT versus Mdr1a/b(−/−), respectively]. Mortality rate in WT mice was not statistically different between sexes (female, 21.4% versus male, 6.7%; P = 0.11). However, in PGP-deficient mice, mortality was higher in male compared with female mice (female, 26.1% versus male, 47.2%; P = 0.025). This mortality in PGP-deficient male mice was also higher than in male WT mice [Mdr1a/b(−/−), 47.2% versus WT, 6.7%; P = 0.020], whereas mortality in female mice did not differ between mouse genotypes (P = 0.321).
Pharmacokinetic Properties of Pilocarpine in WT Mice.
To gain information on blood-brain barrier penetration of pilocarpine in WT mice, we determined pilocarpine concentrations in brain and plasma over a sampling period of 2 hours following intraperitoneal administration of a single subconvulsive dose of pilocarpine (50 mg⋅kg−1). Maximum plasma concentration (29.20 ± 0.86 μg⋅ml−1) was measured 10 minutes following administration (Fig. 2). Estimated plasma elimination half-life (t1/2) was 35 minutes. Brain concentrations peaked between 10–30 minutes after injection (about 22 μg⋅g−1; Fig. 2). On the basis of these results, the time point of 20 minutes postinjection was used to compare pilocarpine brain uptake in different mouse genotypes and following pharmacological treatment (see below). Brain-plasma ratio was approximately 1 at 30 minutes after injection and remained constant thereafter.
Brain Accumulation of Pilocarpine Is Limited by PGP and BCRP.
To evaluate the influence of PGP on pilocarpine brain uptake, we compared brain-plasma ratios in WT and Mdr1a/b(−/−) mice. As PGP and BCRP are characterized by an overlapping substrate spectrum, we included Bcrp1(−/−) mice as well as triple knockout mice [Mdr1a/b(−/−)Bcrp1(−/−) mice] in this part of the study. Unexpectedly, after administration of a subconvulsive dose of 50 mg⋅kg−1 pilocarpine, brain concentrations in WT and Mdr1a/b(−/−) were not statistically different. Only a tendency toward higher concentration in Mdr1a/b(−/−) compared with WT mice was identified (Fig. 3A; P = 0.0843, Student’s t test). However, following administration of 200 mg⋅kg−1 pilocarpine, an increase in brain-plasma ratio of Mdr1a/b(−/−) mice was statistically significant compared with WT mice (Fig. 3B). An unexpected decrease in brain-plasma ratio was found in mice lacking only BCRP compared with WT and triple-knockout mice (Fig. 3A), suggesting compensatory overexpression of other transporters in this mouse genotype. In Mdr1a/b(−/−)Bcrp1(−/−) mice, a statistically significant increase in brain-plasma ratio of pilocarpine was observed compared with WT mice that was higher than the increase in the PGP-deficient mice. For comparison, we added a group of WT mice pretreated with the dual PGP and BCRP inhibitor tariquidar and found a similar ratio to the triple knockout mice (Fig. 3C). By contrast, LiCl pretreatment did not alter pilocarpine brain uptake (Fig. 3C).
For species comparison, we administered 30 mg⋅kg−1 pilocarpine in rats. Consistent with our mouse data, increased brain-plasma ratio of tariquidar pretreated rats was statistically significant compared with control animals, whereas LiCl pretreatment had no effect on pilocarpine brain uptake (Fig. 3D).
Pilocarpine Is Transported by Murine PGP and BCRP In Vitro.
To confirm the in vivo findings, in vitro CETA using Mdr1a- or Bcrp1-transduced cells were performed. CETA allow evaluation of active transport almost independently of the passive permeability component (Luna-Tortós et al., 2008; Löscher et al., 2011). Since endogenous PGP or BCRP of LLC or MDCK-II cells may influence pilocarpine transport, we inhibited function of these transporters by adding tariquidar [at a concentration (0.2 μM) that is reported to selectively inhibit PGP (Kühnle et al., 2009)], to Bcrp1-overexpressing cells; and a BCRP inhibitor (Ko143) to Mdr1a-overexpressing cells. As we did not test whether this concentration inhibits only PGP in our assays, we cannot exclude that murine BCRP was partially inhibited by 0.2 μM tariquidar. However, as BCRP is the predominantly expressed transporter in the transduced MDCK-II cells and transport is clearly present, any crossinhibition appears to be negligible. In Mdr1a-overexpressing cells, transport of pilocarpine from the basolateral to the apical compartment was detected (Fig. 4A). Transport was eliminated in the presence of PGP inhibitor PSC833 (Fig. 4B). Unexpectedly, in LLC WT cells, pilocarpine transport was observed as well, regardless of PSC833 presence (Fig. 4, C and D), suggesting pilocarpine transport by another noninhibited transporter expressed to higher extent in WT cells than in PGP-overexpressing cells. In Bcrp1-overexpressing cells, distinct transport of pilocarpine into the apical compartment was observed, which was not detected in WT cells (Fig. 4, E and G). Following inhibition of murine BCRP by Ko143, pilocarpine transport was ablated (Fig. 4F). As a routine positive control experiment, CETA with reference substrate digoxin was performed in LLC cells overexpressing Mdr1a. Distinct transport of digoxin was observed (Fig. 4H), which could be completely inhibited by PSC833 (not shown).
Discussion
The major findings of our study are that higher doses of pilocarpine are necessary to induce a self-sustaining SE in mice expressing PGP as compared with mice in which the PGP-encoding gene is knocked out. This phenomenon is reflected by increased pilocarpine brain uptake in PGP- and PGP/BCRP-deficient mice, as well as in WT mice and rats following pharmacological inhibition of PGP and BCRP. The transport of pilocarpine by PGP and BCRP is further substantiated by in vitro transport assays.
SE Induction by Pilocarpine in FVB/N WT and Mdr1a/b(−/−) Mice.
The mouse pilocarpine model of TLE has recently gained increasing relevance in preclinical epilepsy research (Buckmaster and Haney, 2012). In this study, we used a fractionated pilocarpine injection protocol for SE induction in the inbred mouse strain FVB/N, which is a common background strain for generation of genetically modified mice (Taketo et al., 1991). Here, this protocol showed superior outcome compared with a recently published work in FVB mice by Buckmaster and Haney (2012). Although the reported bolus protocol led to an almost identical SE induction rate of 69.2% versus 71.7% in our study, mortality was much higher (51.5% versus 21.1%). This supports the advantage of individual dosing in the fractionated pilocarpine model.
The finding that Mdr1a/b(−/−) mice were more susceptible to pilocarpine’s convulsive effects than FVB/N WT mice was somewhat surprising, as there is no evidence from the literature for pilocarpine transport by PGP. According to the “rule of fours,” compounds with a molecular mass <400 g/mol, sum of N and O atoms ≤4, and base pKa <8 are probably not PGP substrates (Didziapetris et al., 2003). Thus, it was not self-evident to expect pilocarpine (molecular mass, 208 g/mol, sum of N and O atoms = 4, pKa = 6.78) to be transported by PGP. However, considering the presence of phytotoxins during evolution, it is plausible that pilocarpine, which is contained in the leaves of a tropical plant, would be among substrates of multidrug transporters that have protective function at tissue barriers throughout the body. Interestingly, we also found that higher pilocarpine doses are needed for SE induction in female than in male WT mice. This sex difference was not observed in PGP-deficient mice and therefore could be explained by a PGP-dependent mechanism. Similar sex differences in SE induction rate have previously been described in mice (Buckmaster and Haney, 2012) and rats (Mejı́as-Aponte et al., 2002), and are probably not based on sex differences in pharmacokinetics of pilocarpine (Omori et al., 2004). Rather, they may be attributed to sex-dependent sexual hormone levels occurring in the estrus cycle, which have been shown to influence expression levels of PGP (Schiengold et al., 2006). Indeed, evidence suggests that the degree of multidrug transporter–mediated efflux activity at the blood-brain barrier might be sex-related (Morris et al., 2003). For example, brain uptake of the PGP substrate verapamil was lower in PGP-competent female compared with male mice, which may indicate an increased functional PGP activity in female mice (Dagenais et al., 2001).
Transport of Pilocarpine by PGP and BCRP.
Pilocarpine has an oil-water coefficient (log P) of 0.95–1.15 and a low molecular mass (208.26 g⋅mol−1). Additionally, low plasma protein binding (<5%) has been reported (Omori et al., 2004). Surprisingly, although the pilocarpine model has been used in epilepsy research for decades, only limited information is available concerning brain penetration in rodents (Table 1). It was recently hypothesized that pilocarpine-mediated secondary effects like peripheral inflammation or elevated potassium levels in the brain are required for SE induction (Marchi et al., 2007, 2009). Furthermore, it was suggested that pilocarpine bears only modest blood-brain barrier permeability following systemic administration in rats (Table 1) (Marchi et al., 2007). By contrast, our data suggest good penetration of pilocarpine across the rat blood-brain barrier, demonstrated by a brain-plasma ratio of about 1 at 30 minutes postadministration. This is consistent with data from Zgrajka et al. (2010), who report equal or even higher brain-plasma ratios (Table 1). In Naval Medical Research Insitute (NMRI) mice, Mazzuferi and colleagues (2012) report brain-plasma ratios comparable to the present study (Table 1).
For comparison of pilocarpine brain uptake in the different mouse genotypes, we initially administered a subconvulsive dose to exclude influence of a seizure-induced blood-brain barrier disruption (Sahin et al., 2003; van Vliet et al., 2007; Li et al., 2013). Unexpectedly, no increase in brain levels was observed in PGP-deficient mice. However, increasing the dose to levels used for SE induction revealed a statistically significant increase in brain uptake among PGP-deficient mice (Fig. 3B). It has been suggested that in the absence of a multidrug transporter, compensatory changes in expression of other multidrug transporters may occur (Cisternino et al., 2004; Hoffmann and Löscher, 2007). For example, Cisternino and colleagues (2004) showed that in mutant Mdr1a-deficient mice, mRNA expression of Bcrp1 is increased ∼3-fold. We consider it possible that such a compensatory increase in transporter expression in Mdr1a/b(−/−) mice may have prevented increased uptake of the subconvulsive pilocarpine dose, but that this potential compensatory transport was at least partially saturated by the convulsive dose. However, a direct translation of mRNA data to the protein level is not possible, and conclusions should be drawn with caution. In combined PGP/BCRP-deficient mice, the increase of pilocarpine brain uptake was higher than in either PGP or BCRP knockout mice, suggesting an additive function of PGP and BCRP. Similar observations have been made for other substrates of PGP and BCRP (de Vries et al., 2007; Chen et al., 2009; Polli et al., 2009). Furthermore, pretreatment of WT mice with 15 mg⋅kg−1 of the dual PGP and BCRP inhibitor tariquidar (Bankstahl et al., 2013) resulted in increased pilocarpine brain uptake to a degree comparable with PGP/BCRP-deficient mice, underscoring that pilocarpine is indeed transported by both multidrug transporters. Interestingly, in Bcrp1(−/−) mice we did not observe increased pilocarpine brain uptake, although in vitro results clearly show BCRP-mediated transport. Again, this finding may be explained by a compensatory overexpression of other transporters in Bcrp1(−/−) mice. In addition, many compounds interacting with multidrug transporters are described to behave dose dependently as both substrate and inhibitor (Bankstahl et al., 2013). Further studies are needed to determine whether this concept may be applicable for pilocarpine.
Published studies suggest differences between mouse strains with respect to their susceptibility to pilocarpine-induced seizures/SE and resulting neuropathology (Bankstahl et al., 2012; Schauwecker, 2012). FVB/N mice appear to exhibit high vulnerability compared with other inbred (C57BL/6) or outbred strains (CD1, NMRI) (Chen et al., 2005). Conversely, C57BL/6 mice are highly resistant to pilocarpine in terms of SE induction (Borges et al., 2003; Müller et al., 2009; Bankstahl et al., 2012). In view of the present results, these strain differences might be explained in part by different transporter expression levels at the blood-brain barrier, leading to mouse strain–dependent brain uptake of pilocarpine. Therefore, it would be of interest to compare brain-plasma ratios of pilocarpine in commonly used outbred and inbred mouse strains for epilepsy research in future studies.
As the pilocarpine model is commonly used in rats, we investigated whether similar efflux transport of pilocarpine was present in this species. Consistent with the mouse data, pharmacological transporter inhibition with tariquidar led to comparably increased pilocarpine brain uptake in rats, supporting its classification as a multidrug transporter substrate across species (Fig. 3D).
In vitro CETA clearly confirmed transport of pilocarpine by overexpressed murine PGP and BCRP, which could be blocked by respective specific inhibitors of both transporters (Fig. 4). Nevertheless, transport of pilocarpine was less distinct than transport of digoxin, which is frequently used as a model PGP substrate (Schwab et al., 2003). Unexpectedly, pilocarpine transport was also present in LLC WT cells without PGP overexpression (Fig. 4, C and D). This finding cannot be interpreted as pilocarpine transport mediated by endogenous PGP or BCRP, as both were blocked in WT cells by respective inhibitors. Rather, persistent transport despite PGP/BCRP inhibition suggests transport mediated by other endogenous transporters expressed to a higher degree in LLC WT cells than in PGP-overexpressing cells.
Effect of Pretreatment with LiCl on Pilocarpine Brain Uptake.
It is well known that pretreatment with LiCl potentiates the SE-inducing effect of pilocarpine in rats as well as in mice (Jope et al., 1986; Clifford et al., 1987; Bersudsky et al., 1994; Chaudhary et al., 1999; Shaldubina et al., 2007). The underlying mechanism is not completely understood. One hypothesis suggests that LiCl-mediated proinflammatory processes affect blood-brain barrier integrity (Marchi et al., 2009), subsequently resulting in increased brain uptake of pilocarpine. This concept is supported by the findings of Marchi et al. (2007) that pretreatment with LiCl leads to activation of white blood cells, increased blood interleukin-1β levels, and, importantly, increased blood-brain barrier permeability of fluorescein isothiocyanate–labeled albumin. Therefore, we evaluated whether LiCl pretreatment would lead to higher pilocarpine brain levels compared with vehicle-treated animals as a consequence of blood-brain barrier disturbance. However, our results do not support a conclusion that potentiation of pilocarpine convulsant activity can be attributed to LiCl-mediated blood-brain barrier impairment, as a subsequent increase in brain concentration of pilocarpine does not occur.
Potential Implication for Clinical Use of Pilocarpine.
Awareness of pilocarpine as a transporter substrate is of relevance for clinical use of this compound, especially relating to drug interactions. Although we have not directly investigated transport of pilocarpine by multidrug transporters in the periphery, it is probable that transporters also play a role in peripheral pharmacokinetics of pilocarpine. For example, the combination of clinically-used, transporter-inhibiting compounds, such as verapamil or cyclosporine, with pilocarpine in patients with xerostomia may lead to increased side effects, or even toxic effects, of pilocarpine owing to increased tissue concentrations.
Conclusion.
In conclusion, our findings demonstrate that pilocarpine has good brain penetration in mice and rats that is not further increased by pretreatment with LiCl. However, cerebral uptake is limited by at least two major multidrug transporters at the blood-brain barrier, leading to increased convulsant potency of pilocarpine in transporter knockout mice. Differences in multidrug transporter expression levels at the blood-brain barrier may therefore explain differences in pilocarpine susceptibility. Besides genetic modification or mutation, alterations in transporter expression can be disease-associated, e.g., in mouse models of Alzheimer’s disease (Cirrito et al., 2005; Hartz et al., 2010). Therefore, the efflux-transporter-substrate properties of pilocarpine should be considered when applied for SE induction in rodents with altered multidrug transporter expression or function.
Acknowledgments
The authors thank Piet Borst and Alfred Schinkel (The Netherlands Cancer Institute) for providing the cell lines and breeding pairs of transporter knockout mice used in this study. The skillful technical assistance of Martina Gramer and Maria Hausknecht with HPLC analysis is gratefully acknowledged. The authors thank Manfred Kietzmann for help with pharmacokinetic analysis. The authors are grateful to James Thakeray for carefully revising the manuscript.
Authorship Contributions
Participated in research design: M. Bankstahl, J. P. Bankstahl, Löscher.
Conducted experiments: M. Bankstahl, J. P. Bankstahl, Römermann.
Performed data analysis: M. Bankstahl, J. P. Bankstahl, Römermann.
Wrote or contributed to the writing of the manuscript: M. Bankstahl, J. P. Bankstahl, Römermann, Löscher.
Footnotes
- Received December 11, 2014.
- Accepted March 6, 2015.
↵1 Current affiliation: Department of Nuclear Medicine, Preclinical Molecular Imaging, Hannover Medical School, Hannover, Germany.
K.R. and J.P.B. contributed equally to this work.
This work was supported by a grant from the German Research Foundation (DFG) [Grant Lo 274/10-2]. The authors have no conflicts of interest.
Abbreviations
- BCRP
- breast cancer resistance protein
- CETA
- concentration equilibrium transport assays
- HPLC
- high-performance liquid chromatography
- Ko143 hydrate
- (3S,6S,12aS)-1,2,3,4,6,7,12,12a-octahydro-9-methoxy-6-( 2-methylpropyl)-1,4-dioxopyrazino[1′,2′:1,6]pyrido[3,4-b]indole-3-propanoic acid 1,1-dimethylethyl ester hydrate
- LLC cells
- Lilly Laboratories cells
- MDCK-II cells
- Madin-Darby canine kidney type II cells
- PGP
- P-glycoprotein
- PSC833
- (3S,6S,9S,12R,15S,18S,21S,24S,30S,33S)-6,9,18,24-tetraisobutyl-3,21,30-triisopropyl-1,4,7,10,12,15,19,25,28-nonamethyl-33-[(2R,4E)-2-methyl-4-hexenoyl]-1,4,7,10,13,16,19,22,25,28,31-undecaazacyclotritriacontane-2,5,8,11,14,17,20,23,26,29,32-undecone
- SE
- status epilepticus
- TLE
- temporal lobe epilepsy
- WT
- wild-type
- Copyright © 2015 by The American Society for Pharmacology and Experimental Therapeutics