INTRODUCTION

The adenosine triphosphate (ATP)-binding cassette (ABC) superfamily of efflux transport proteins contains subfamily B, often referred to as the multidrug-resistant subfamily (MDR), and subfamily C, also known as the multidrug-resistant related protein subfamily (MRP) (Germann, 1996; Carson et al, 2002). The most extensively studied ABC transporter is P-glycoprotein (P-gp). P-gp belongs to subfamily B and is a 170-kDa transmembrane protein encoded by the human gene ABCB1. Formerly known as MDR1 or PGY1, ABCB1 maps to chromosome 7q21.1 (Dean et al, 2001). It was the first ABC transporter cloned and characterized by its ability to confer a multidrug resistance phenotype to cancer cells that had become resistant to chemotherapy (Riordan et al, 1985; Roninson et al, 1986). In addition to expression in tumor cells, the ABCB1 protein (P-gp) is widely localized in normal tissues, including the apical membranes of the gastrointestinal tract, the biliary canalicular membranes of hepatocytes, the luminal membranes of proximal tubular epithelial cells in the kidney, and the luminal membranes of endothelial cells in cerebral capillaries forming the blood–brain barrier (Cordon-Cardo et al, 1989; Yu, 1999; Silverman, 2000).

The biological function of P-gp appears to be a protective role for major organs by limiting cellular uptake of xenobiotics by excreting these compounds into bile, urine, the intestinal lumen, and limiting accumulation in brain. This function as a drug efflux pump plays an important role in drug absorption, distribution, and elimination (Lin and Yamazaki, 2003). A large number of structurally diverse drugs from various pharmacologic classes are actively transported by P-gp (Schinkel et al, 1996; Schuetz et al, 1996; Mahar Doan et al, 2002; Weiss et al, 2003). P-gp is highly polymorphic and its expression in the endothelial cells of the blood–brain barrier limits substrate access to the central nervous system (CNS), thereby influencing the therapeutic consequences of psychoactive drugs (Thompson et al, 2000). A P-gp inhibitor was shown to increase the extracellular fluid concentration of morphine in the brain of rats, while chronic treatment with morphine leading to tolerance of the antinociceptive effect was shown to be mediated by a morphine-induced increase in the brain P-gp (Aquilante et al, 2000).

Several in vitro cell systems and a transgenic Abcb1a/b(−/−) mouse model have been used to study the functional activity of the protein in the blood–brain barrier (Lin and Yamazaki, 2003; Schwab et al, 2003). Drug-stimulated ATPase activity can be determined by monixtoring the release of inorganic phosphate (Sarkadi et al, 1992; Urbatsch et al, 1994). Alternatively, fluorescent indicators can be monitored where the inhibition of P-gp leads to cellular accumulation of rhodamine-123 or calcein AM (Eytan et al, 1997; Tiberghien and Loor, 1996). The Abcb1a(−/−) ‘knockout’ mouse model exhibits functional deficiency of P-gp in the blood–brain barrier (Cordon-Cardo et al, 1989; Schinkel et al, 1996, 1997; Silverman, 2000), and therefore has provided considerable insight into the role and importance of P-gp in the brain entry of drugs (Kim et al, 1998; Schinkel et al, 1996, 1997; Letrent et al, 1999; Aquilante et al, 2000). Mice lacking the Abcb1a gene display no anatomical, biochemical, or other abnormalities (Schinkel et al, 1994, 1997).

Data documenting transport of antipsychotic drugs with P-gp are sparse. No previous reports could be found using the Abcb1a (−/−) mouse to test antipsychotic drugs for transport by P-gp. Previous reports using in vitro methods have found that levomeprazine and risperidone may be substrates, while chlorpromazine, clozapine, haloperidol, chlorprothixene, and perphenazine may not (Mahar Doan et al, 2002; Schinkel et al, 1997). The data on chlorpromazine are conflicting with positive transport using the human CCRF-CEM leukemia cells (Syed et al, 1996) and low transport by multidrug resistance-transfected MDCK cells (Mahar Doan et al, 2002). Using ATPase activity as a marker for P-gp binding affinity, we previously found a rank order of activity from verapamil (positive control)>quetiapine>risperidone>olanzapine>chlorpromazaine>haloperidol=clozapine (Boulton et al, 2002). These results suggested that P-gp may influence the access to the brain of the antipsychotics to various degrees.

Verapamil, quetiapine, and risperidone each are partially metabolized by CYP3A4 (Kroemer et al, 1993; DeVane and Nemeroff, 2001a, 2001b). Olanzapine is predominantly biotransformed by CYP1A2 and glucuronyl transferases, yet a small contribution by CYP3A4 has been implicated (Ring et al, 1996; Markowitz et al, 2001). An inter-relationship in substrate specificity between cytochrome P450 (CYP) 3A4 and P-gp has been proposed (Kim et al, 1999). The ABCB1 and CYP3A genes are located on the same chromosome with proximal gene loci in humans (Tsui et al, 1995); however, many drugs that are CYP3A4 substrates (midazolam, triazolam, alprazolam) are not transported by P-gp and vice versa (digoxin, fexofenadine; Perloff et al, 1999; Schuetz et al, 1996; von Moltke et al, 1999). Nevertheless, the overlap in CYP3A4 and P-gp substrates is consistent with our previous in vitro results based on ATPase activity, suggesting that quetiapine among the antipsychotics had the greatest degree of affinity for P-gp (Boulton et al, 2002). Olanzapine has the least evidence as a P-gp substrate. To further assess the role of P-gp in limiting brain entry of the antipsychotics, we compared the tissue distribution of olanzapine in Abcb1a(−/−) transgenic mice with that in the FVB control Abcb1a(+/+) mice.

MATERIALS AND METHODS

Chemicals

Olanzapine and RS0244 (internal standard) were obtained from Eli Lilly Inc. (Indianapolis, ID, USA). Hank's balanced salt solution (HBSS) was purchased from Gibco BRL (Carlsbad, CA, USA). Other chemicals and reagents were the purest grade available and were obtained from Fisher Scientific Co. (Fairlawn, NJ, USA).

Experimental Animals

Male Abcb1a(−/−) mice (FVB/TacfBR-[KO]MDR1aN7) and genetically matched male Abcb1a(+/+) mice (FVB/MTtacfBR) 9–10 weeks of age weighing 20–30 g were obtained from Taconic (Germantown, NY, USA). The founding of this colony has been described previously (Schinkel et al, 1994). Animals were housed individually and maintained at a 12:12 h light/dark cycle. The animals were cared for in accordance with the US Public Health Service policy for the Care and Use of Laboratory Animals, and the experimental studies were approved by the Medical University of South Carolina Animal Care Committee.

Experimental Procedures

Olanzapine was dissolved in 0.9.% saline with minimal use of 1 N HCL (1‰, v/v). A dose of 2.5 μg olanzapine/g mouse was administered intraperitoneally (i.p.). The entire volume injected was 10 μl/g mouse. At 1 h after the injection, the mice were anesthetized with halothane and decapitated. Trunk blood was collected in EDTA-coated tubes and centrifuged at 3000g for 20 min to determine plasma concentrations of olanzapine.

The brain (brain tissue plus meninges), liver, kidney, spleen, and penis were dissected and weighed, and then homogenized in five-fold volume of an HBSS and 0.02 M HEPES buffer, pH 7.2. The homogenates were frozen at −70°C until high-performance liquid chromatography (HPLC) analysis.

HPLC Analysis of Olanzapine Plasma and Tissue Concentrations

The plasma and tissue samples obtained at 1 h following administration of olanzapine were extracted and determined using HPLC method previously developed and validated in our laboratory and modified for tissue samples (Boulton et al, 2001; Llorca et al, 2001). Briefly, olanzapine was extracted from 100 or 200 μl of plasma or organ samples after the addition of 50 μl of internal standard working solution (RS0244 10 μg/ml), 0.2 ml of 1 M bicarbonate buffer (pH 11), and 5 ml of a mixture of hexane–isoamyl alcohol (98:2, v/v). The mixture was shaken for 15 min and centrifuged at 3000g for 5 min. The organic layer was transferred to another tube and back-extracted with 0.2 ml 0.2 M sulfuric acid. After shaking and centrifugation, the aqueous layer was re-extracted with 0.2 ml of 1 M bicarbonate buffer. The organic phase was evaporated under nitrogen at 45°C. The residue was dissolved in 100 μl of mobile phase (65% methanol:35% water containing 0.2% triethylamine and was adjust to pH 7.0 with 85% phosphoric acid), and 50 μl was injected into the chromatographic system. Olanzapine and internal standard were separated by a Luna 5u C18(2) column (250 × 4.6 mm; Phenomenex), and the effluent was monitored at a UV wavelength of 273 nm and at a flow rate of 1 ml/min at ambient temperature. The detection limit of olanzapine was 5 ng/ml. The intraday and interday coefficients of variation were lower than 7% at relevant concentrations (n=7). The plasma and tissue extraction recoveries ranged from 83 to 98%.

Data Analysis

Drug quantitation was performed by comparing peak height ratios of olanzapine to the internal standard with ratios derived from calibration curve of standards containing known amounts of drug extracted from plasma or tissues as described above. Final concentrations of olanzapine were expressed as either μg/ml for plasma or μg/g for tissues.

An unpaired t-test was used to compare the differences between the concentrations of olanzapine in the Abcb1a(−/−) mice and control mice and their tissue:plasma and tissue:spleen ratios. A two-tailed P-value was used and the level of statistical significance was P<0.01.

RESULTS

The plasma and organ concentrations of olanzapine 1 h after the i.p. injection of 2.5 μg olanzapine/g mouse are shown in Table 1 and Figure 1. Substantially and statistically higher olanzapine concentrations were found in the brain (three-fold), liver (2.6-fold), and kidney (1.8-fold) of Abcb1a(−/−) mice than those of the control mice (Table 1, Figure 1). Consistently higher ratios were also found for the organ:plasma and the organ:spleen concentrations of olanzapine in the Abcb1a(−/−) mice than in the control mice (Table 1, Figure 2). However, there was no statistically significant difference for concentrations of olanzapine in the plasma, spleen, and penis between the two groups, nor for the tissue:plasma ratios of spleen and penis (Table 1, Figures 1 and 2).

Table 1 Tissue Concentrations and Tissue:Plasma Ratios of Olanzapine in Abcb1a(−/−) (n=6) and Wild-Type Mice (n=6) after 1 h i.p. Injection of Olanzapine 2.5 μg/g Mouse
Full size table
Figure 1
figure 1

Plasma (μg/ml) and tissue (μg/g) concentrations of olanzapine in Abcb1a(−/−) (solid bars) and Abcb1a(+/+) mice (open bars) 1 h after the i.p. injection of 2.5 μg olanzapine/g mouse. Values are shown as mean±SD (n=6). The P-values shown were calculated by two-tailed unpaired t-test.

Full size image
Figure 2
figure 2

Organ:plasma ratios of olanzapine in Abcb1a(−/−) (solid bars) and Abcb1a(+/+) mice (open bars) 1 h after the i.p. injection of 2.5 μg olanzapine/g mouse. Values are shown as mean±SD (n=6). The P-values shown were calculated by two-tailed unpaired t-test.

Full size image

DISCUSSION

The current findings demonstrate that the brain penetration of olanzapine is limited by P-gp. The three-fold difference in the brain concentration of olanzapine between the P-gp-deficient Abcb1a(−/−) mice and the FVB control Abcb1a(+/+) mice were most likely due to the functional activity of P-gp residing in the blood–brain barrier. This conclusion is also in good agreement with the 2.6-fold difference observed in brain:plasma and brain:spleen ratios of olanzapine between the two groups, indicating that the difference in the brain olanzapine concentration between the two groups was not caused by differences of plasma olanzapine concentration that were identical. Similarly, differences in concentration in the spleen were absent, since the expression of P-gp in the spleen is very low (Cordon-Cardo et al, 1989; Schinkel et al, 1997). These findings of no drug concentration differences in organs with low P-gp expression (plasma and spleen) are consistent with the previous results of Uhr et al (2000) using amitriptyline.

Previously, the absence of P-gp in Abcb1a(−/−) mice has been shown as a pervasive influence on the tissue distribution and pharmacokinetics of drugs such as indinavir, nelfinavir, saqunavir, loperamide, ondansetron, ivermectin, vinblastin, cyclosporin A, amitriptyline, morphine, some endogenous steroid hormones, citalopram, trimipramine, carbamazepine, phenobarbital, lamotrigine, and felbamate (Kim et al, 1998; Schinkel et al, 1996, 1997; Letrent et al, 1999; Aquilante et al, 2000; Uhr et al, 2000, 2002; Fromm, 2000; Potschka et al, 2001, 2002; Uhr and Grauer, 2003). The brain concentrations of these drugs were remarkably higher in Abcb1a(−/−) mice than that of Abcb1a (+/+) mice. However, differences in the liver and kidney were smaller compared to those in the brain or not statistically significant. In the present study, similar ratios of olanzapine to plasma were found for the brain, liver, and kidney (3.0, 2.6, 1.8, respectively, Table 1) in Abcb1a(−/−) compared to Abcb1a(+/+) mice. When studies have found large differences in the brain concentration of drugs transported by P-gp in Abcb1a(−/−) animals compared to controls, the liver and kidney differences in drug concentration were frequently small, but for some drugs (experimental anticonvulsants and doxorubicin; Van Asperen et al, 1999; Cox et al, 2002) such differences have been substantial. The exact mechanism(s) for this inconsistency between study results is unclear, but it suggests that the tissue distribution of drugs transported by P-gp in the two animal groups is highly dependent upon the specific substrate involved. Nevertheless, there is consistency in our results that statistically significant differences in drug concentrations were found in tissues normally containing P-gp (brain, liver, kidney), while no differences in concentrations were found in those tissues devoid of P-glycoprotein (plasma, spleen, and penis).

Olanzapine is one of the most widely used antipsychotic agents that has been marketed post clozapine. It is a serotonin/dopamine antagonist that has proven efficacy against both the positive and negative symptoms of schizophrenia, and is widely used in other psychotic disorders (Bhana et al, 2001). It has a markedly better side-effect profile than older conventional agents such as haloperidol and chlorpromazine (Beasley et al, 1996). However, even with these improvements, not all patients respond adequately to olanzapine and dosage requirements vary widely. Approximately 20–30% of patients with schizophrenia show an inadequate efficacy and 5–10% are intolerant to olanzapine therapy (Bhana et al, 2001).

The exact reason for refractory symptoms in olanzapine-treated patients is unknown. However, recent findings of functional polymorphisms of ABCB1 in humans may provide an insight into one mechanism (Hoffmeyer et al, 2000). Single-nucleotide polymorphisms (SNP) in exon 26 (C3435T) and exon 21 (G2677T) have been associated with the functional activity of P-gp. The homozygous T/T genotype mutant alleles were associated with more than two-fold lower intestinal P-glycoprotein expression levels and higher bioavailability of digoxin, a prototype P-gp probe (Kurata et al, 2002; Nakamura et al, 2002). High frequencies of the mutant alleles are present in the population (19% for the T/T 3435 allele and 28% for the T/T2677 allele; Lin and Yamazaki, 2003). These functional SNPs of P-gp in humans may represent a major source of interindividual variability in the therapeutic efficacy and toxicity of P-gp substrates. This effect may be even more remarkable for psychoactive drugs, which need to penetrate through the P-gp residing in the blood–brain barrier to exert their desirable pharmacological effects. Quite recently, Siddiqui et al (2003) reported that the ABCB1 C3435T polymorphism was associated with drug responses of epilepsy. Patients with drug-resistant epilepsy were more likely to have the CC genotype at ABCB1 3435 than the TT genotype compared with patients with drug-responsive epilepsy. In addition, the homozygosity for 3435T alleles of ABCB1 has been suggested as a risk factor for the occurrence of nortriptyline-induced postural hypotension, a CNS side effect related to the protective role of P-gp in the brain penetration of nortriptyline (Roberts et al, 2002). Accordingly, it is conceivable that patients who have a high P-gp expression in their blood–brain barrier may have low brain concentrations of olanzapine contributing to therapeutic resistance or failure, and patients who have a low P-gp expression in their blood–brain barrier may have a relatively high brain olanzapine concentration, resulting in adverse events, toxicity, or beneficial therapeutic effects with lower than usual daily drug doses.

Previous studies have indicated that the modulation of P-gp-mediated drug transport can dramatically increase or decrease drugs' penetration into the CNS and affect their therapeutic efficacy (Letrent et al, 1999; Aquilante et al, 2000). For example, the analgesic effect of morphine in rats was enhanced through the use of a specific P-gp inhibitor (Letrent et al, 1999). Accordingly, the current experimental findings provide a rational for the potential use of P-gp modulators with olanzapine or other P-gp substrates, since a reversal of the functional activity of P-gp may result in improved therapeutic efficacy by increasing CNS drug concentration. However, the inhibition of P-gp activity in humans would likely increase both plasma and brain concentrations of olanzapine and other orally administered substrates, since P-gp is also present in the gastrointestinal tract (Fricker and Miller, 2002; Verstuyft et al, 2003; Becquemont et al, 2001; Hennessy et al, 2002). This is not necessarily a deterrent for the use of safe P-gp inhibitors. By artificially increasing the brain:plasma drug concentration ratio through the inhibition of P-gp, smaller oral substrate doses could be used resulting in lower plasma drug concentration, while maintaining an equivalent or higher concentration in the CNS. Adverse events related to systemic drug exposure should be decreased as a result of using lower oral doses while promoting increased efficacy at the site of action as a result of increased brain drug concentration. Such an approach is currently being investigated to improve the therapeutic efficacy of the protease inhibitors (Choo et al, 2000). A requisite step to applying this approach in neuropsychopharmacology is the identification of relevant P-gp substrates.

By measuring P-gp ATPase activity, we previously reported that the antipsychotics have various degrees of binding affinity with P-gp in vitro, with rank order of affinity (by measuring Vmax/Km) as: quetiapine (1.7)>risperidone (1.4)>olanzapine (0.8)>chlorpromazine (0.7)>haloperidol (0.3)=clozapine (0.3) (Boulton et al, 2002). Accordingly, olanzapine was identified as an intermediate affinity substrate among the antipsychotics (Boulton et al, 2002). Consistent with the present findings, a previous report using monolayer efflux in multidrug resistence-transfected cells identified that cellular uptake of risperidone is mediated by P-gp (Mahar Doan et al, 2002). In another study, the conventional antipsychotic chlorpromazine has been shown to be actively transported by P-gp in cancer-derived cells (Syed et al, 1996). Unfortunately, binding studies and in vitro transport measurements in monolayer cell systems under-represent the complexity of the in vivo situation. Inconsistencies have been noted (Polli et al, 2001). This is the first report using an in vivo model to identify that the brain entry of olanzapine is limited by P-gp. Unfortunately, no chemical, clinical, or pharmacological properties appear to allow accurate predictions within or between drug classifications of those compounds that are transported by P-gp. Even the widely used in vitro methods to evaluate transport have produced conflicting findings with approximately half of the studied compounds displaying significant variation in P-gp activity dependent on the specific assay used (Adachi et al, 2001; Yamazaki et al, 2001). The single and double knockout mouse model appears to produce the most unequivocal results (Schinkel et al, 1996).

In conclusion, the present results showed that the expression of P-gp in the blood–brain barrier significantly limits the penetration of olanzapine into the CNS. Functional polymorphisms of P-gp in humans may represent an unrecognized source of interindividual variability in the therapeutic efficacy of olanzapine. The recognition of brain entry of olanzapine is limited by P-gp offering intriguing possibilities for manipulating its therapeutic utility through the use of adjunctive P-gp modulators.