Introduction

The tight junctions of capillary endothelial cells of the brain serve to protect the central nervous system (CNS) from exposure to potential toxicants by limiting the paracellular movement of endogenous and exogenous compounds. Consequently, specific uptake transporters are required to carry essential molecules such as glucose and amino acids from the blood to the brain through capillary endothelial cells. Until recently, it was thought that only highly lipophilic drugs were capable of gaining access to the CNS. We now know that in addition to the physicochemical properties of a drug, such as charge, lipophilicity, and molecular weight, uptake and efflux drug transporter expression and their activity at the capillary endothelial cells of the blood−brain barrier are often the rate-limiting determinants of CNS drug entry. For this reason, drug transporters play an important role in determining exposure of the brain to drugs and therefore govern the CNS effects and toxicity of many drugs (Table 1).

Table 1 Expression, substrates, and effect of transporters on central nervous system (CNS) exposure to drugs
Full size table

Several uptake and efflux drug transporters are expressed at the blood−brain barrier, including organic anion-transporting polypeptide 1A2 (OATP1A2/SLCO1A2) organic anion transporter 3 (OAT3/SLC22A8), monocarboxylate transporter 1 (MCT1/SLC16A1), P-glycoprotein (P-gp; MDR1/ABCB1), breast-cancer-resistance protein (BCRP/ABCG2) and multidrug-resistance-associated protein 4 (MRP4/ABCC4). The objective of this brief report is to review the expression and activity of these drug transporters at the blood−brain barrier and, where appropriate, comment on the implications for response or toxicity observed during drug therapy.

Organic anion transporting polypeptide 1A2 (OATP1A2/SLCO1A2)

OATP drug transporters mediate the sodium-independent uptake of a broad range of substrates, including the drugs fexofenadine [1], levofloxacin [2], methotrexate [3], and ouabain [4], along with bile acids and the synthetic peptides deltorphin II and D-penicillamine (2,5)- enkephalin (DPDPE) [5]. Recent studies suggest that in addition to its importance to intestinal drug absorption [6], OATP1A2 plays a role in the transfer of drugs from the blood to the brain, as this transporter is highly expressed on the luminal membrane of capillary endothelial cells of the blood−brain barrier [5, 7]. Given the broad substrate spectrum associated with OATP1A2, this transporter is likely one of the reasons why certain drugs attain greater than expected CNS drug levels and, in some cases, CNS toxicity profile of a number of compounds (Fig. 1).

Fig. 1
figure 1

Drug transporter expression at the blood brain barrier (A) and choroid plexus (B). TJ represents tight junctions typical of these regions of the brain

Full size image

Methotrexate is a dihydrofolate reductase inhibitor used to treat several clinical conditions including rheumatoid arthritis, Crohn’s disease, and several types of cancer. Recently, methotrexate has been shown to be a substrate for OATP1A2 [3]. Interestingly, a common OATP1A2 polymorphism (Ile13Thr) present in 16% of European Americans was shown to confer a two-fold increase in OATP1A2-mediated methotrexate uptake. High-dose methotrexate is used to treat several cancers, and the resulting CNS side effects range from mild white-matter changes to severe CNS demyelination and encephalopathy [8]. Given the role of OATP1A2 in mediating methotrexate uptake and the high allele frequency of this gain of function polymorphism Ile13Thr, OATP1A2 should be considered as a factor in the observed CNS toxicity of methotrexate, although this remains to be confirmed in clinical studies.

Levofloxacin is a fluoroquinolone antibacterial agent used to treat respiratory tract infections and in the treatment and prevention of traveler’s diarrhea. Several CNS side effects have been associated with levofloxacin use, including seizures, toxic psychoses, increased intracranial pressure, and CNS stimulation. These side effects may even occur following a single dose of levofloxacin. Although the efflux of quinolone antibiotics by MRP, BCRP and P-gp drug transporters has been well established [9, 10], it was only recently that OATP1A2 was implicated in the uptake of levofloxacin [2]. Although P-gp effectively limits the CNS penetration of levofloxacin [10], it is possible that under conditions of reduced P-gp expression or activity (i.e., drug interactions), OATP1A2-mediated increases in brain accumulation of levofloxacin may be at least partly accountable for the adverse CNS side effects.

Organic anion transporter 3 (OAT3/SLC22A8)

Oat3 was originally isolated from rat brain and shown to mediate the transport of cimetidine, estrone sulfate, para-aminohippurate (PAH), and ochratoxin A [11]. Oat3 has been detected on the apical membrane of the choroid plexus [12, 13] and the luminal and abluminal membranes of rat brain endothelial cells [14]. The function of Oat3 in the choroid plexus appears to mediate the transport of organic anions from the cerebrospinal fluid (CSF) to the blood, whereas at the blood−brain barrier, it transfers substrate drugs from the brain to the blood. In terms of drug action, rOat3 has been implicated in the brain to blood transport of the thiopurines 6-mercaptopurine and 6-thioguanine [15]. Thiopurines are used to treat acute lymphoblastic leukemia (ALL). During treatment of ALL, patients often have relapse due to the entry of leukemic cells into the brain. It has been proposed that Oat3-mediated transport of thiopurines out of the brain allows the proliferation of these cells in the CNS, and thereby result in the observed leukemia relapse during chemotherapy [15].

Oseltamivir (Tamiflu) is a prodrug of Ro 64-0802, which is used for the symptomatic treatment of illnesses caused by the influenza virus. It is a potent inhibitor of the enzyme neuraminidase and such functional effects results in decreased release of viral particles from infected cells. The use of this drug has been implicated with CNS side effects in animals and more recently in humans [16, 17]. A recent study investigated the impact of Oat3 on the CNS penetration of the active metabolite Ro 64-0802 using Oat3 knockout mice [17]. Ro 64-0802 accumulation in the brain was significantly greater in knockout animals at 1 and 2 h following administration compared with wild-type control. Interestingly, there was no difference in the brain to plasma ratio in the Oat3 knockout animals, and therefore other transporters may also contribute to the CNS exposure of oseltamivir. Accordingly, Oat3 likely plays a significant role in the clearance of Ro 64-0802 from the brain, and modulation of Oat3 activity may result in CNS toxicity associated with oseltamivir. Although the results implicating Oat3-mediated drug transport in rodent brain are intriguing, at this time, experimental evidence demonstrating OAT3 expression and/or activity in the human brain is lacking.

Monocarboxylate transporter 1 (MCT1/SLC16A1)

MCT1 is expressed on the luminal membrane of endothelial cells at the blood−brain barrier [18]. Its primary role is in the transport of monocarboxylate solutes, such as lactate and pyruvate, into the brain, although more recently, drug substrates of MCT1 and other members of the MCT family have been studied. The most widely studied drug associated with MCT1 is γ-hydroxy butyrate (GHB), a controlled substance that has been used clinically to treat insomnia, cataplexy, and narcolepsy. CNS-related side effects of GHB include respiratory depression, seizure, and loss of consciousness and may even result in coma or death. Recent studies using an in situ rat brain perfusion model show saturable GHB transport at the blood−brain barrier [19]. Further, transport is inhibited by the MCT1 substrates lactate and pyruvate, suggesting that blood−brain barrier GHB transport is predominantly mediated by MCT1. The direct role of MCT1 in mediating CNS activity and side effects of GHB remain to be elucidated.

P-glycoprotein (MDR1/ABCB1)

P-gp is a member of the adenosine triphosphate (ATP)-binding cassette family of efflux drug transporters and is the most widely studied of all efflux drug transporters. P-gp is known for is critical role in mediating cellular resistance to many chemotherapeutic agents and in limiting the tissue penetration of a broad range of chemically diverse substrate drugs at many blood−tissue barriers. In the brain, P-gp is expressed on the luminal membrane of capillary endothelial cells as well as the epithelium of the choroid plexus, therefore functioning to limit the brain and CSF distribution of drugs.

One of the best examples describing the prominent role P-gp plays in limiting CNS penetration of drugs is the antidiarrheal agent loperamide [20]. Loperamide is a potent opiate. However, due to its high affinity for P-gp as a substrate, it is unable to cross the blood−brain barrier. For this reason, loperamide is available over the counter to reduce gut motility and does not result in the central effects of euphoria and respiratory depression typical of opiates, even when ingested at high doses. A similar example exists with the dopaminergic antagonist domperidone. Unlike other dopaminergic receptor blockers that are used as neuroleptic agents, domperidone does not penetrate the brain and is therefore ineffective as an antipsychotic [21]. Elegant studies have shown that domperidone is actively extruded by P-gp and that its lack of CNS activity is a result of P-gp-mediated extrusion from the brain [20]. In contrast, the neuroleptic agent haloperidol, a non-P-gp substrate, displays significant CNS penetration and is an effective antipsychotic.

Aside from being a substrate for OATP1A2-mediated cellular uptake, fexofenadine is also efficiently extruded from cells by P-gp [1]. The colocalization of these transporters in many tissues, including the brain, suggests that disposition of dual substrates such as fexofenadine will be greatly influenced by transporter expression. Indeed, it has been shown that mice with disruption of the mdr1a gene have a nine-fold increase in the brain level of fexofenadine compared with wild-type controls [1]. As fexofenadine is relatively impermeable by passive diffusion, P-gp is essential in limiting the brain distribution of this nonsedating antihistamine despite its high affinity for OATP1A2-mediated uptake. Importantly, even at supraclinical doses (360 mg), fexofenadine is devoid of CNS side effects [22], likely owing to its low brain penetration.

The late stage of HIV infection is often characterized by neurological complications such as confusion, forgetfulness, and behavioral changes. HIV protease inhibitors are one of the mainstay treatments of HIV infection. Several HIV protease inhibitor drugs have been shown to be substrates of P-gp. Studies using mdr1a(-/-) mice show that the brain concentration and brain to plasma ratio of indinavir, nelfinavir, and saquinavir are significantly increased compared with wild-type control [23]. It is clear from these studies that P-gp plays a fundamental role in limiting the CNS penetration of HIV protease inhibitors and therefore may limit the action of these drugs in treating some of the neurological complications of disease.

Breast-cancer-resistance protein (BCRP/ABCG2)

BCRP was originally identified in a breast cancer cell line that exhibited resistance to anthracyclines [24]. Subsequent studies have clearly shown that BCRP is expressed on the luminal membrane of endothelial cells lining the brain [25], and BCRP is now a widely appreciated component of the blood−brain barrier. Not surprisingly, much effort has gone into defining the role of this transporter as a mediator of drug resistance to chemotherapeutic agents.

Topotecan is a camptothecin derivative that is used primarily to treat recurrent small-cell lung and ovarian cancer, although some studies suggest it may be moderately effective in treating brain metastases [26, 27] and malignant gliomas [28, 29]. Topotecan is an excellent substrate of BCRP and a moderate substrate of P-gp. As BCRP and P-gp often have overlapping substrates and are both highly expressed at tissues, such as the brain, that represent barriers to drug entry, it is important to evaluate the effect that both of these transporters have to the CNS penetration of drugs such as topotecan. Through the use of Bcrp1(-/-), Mdr1a/b(-/-), and Bcrp1(-/-)Mdr1a/b(-/-) mouse models, de Vries et al. clearly demonstrate the important role that Bcrp plays in limiting the brain penetration of topotecan [30]. In Bcrp1 (-/-) Mdr1 (a/b) (-/-) mice, the brain to plasma area under the curve (AUC) ratio of topotecan was 3.2-fold higher than wild-type control. Importantly, the authors point out that when investigating drugs such as topotecan, it is important to perform experiments in a P-gp-deficient background, as P-gp clearly plays a role in limiting the entry of shared substrates.

Imatinib mesylate is indicated for the treatment of chronic myelogenous leukemia and gastrointestinal stromal tumors and functions as a potent inhibitor of Bcr-Abl, platelet-derived growth-factor receptor (PDGFR)-α, PDGFR-β c-Fms, and C-Kit tyrosine kinases. In vitro evidence suggests that imatinib mesylate may also be effective in the treatment of malignant gliomas owing to its potent (IC50 = 0.1 μmol/L) inhibition of PDGFR [31]. However a recent clinical study suggests that imatinib mesylate has minimal activity in the treatment of malignant gliomas in humans [32]. Studies by Breedveld et al. not only show that imatinib mesylate is a substrate of BCRP but also that Bcrp1 knockout mice have a 2.5-fold increased imatinib brain penetration compared with wild-type controls [33]. Similarly, pretreatment with the BCRP inhibitor pantoprazole increases the brain exposure of imatinib by approximately five fold. Taken together, these results suggest that BCRP limits the brain penetration of imatinib and at least in part explains its minimal efficacy in the treatment of malignant gliomas.

Aside from its role in mediating drug efflux, BCRP is also noted for its protective role in limiting the exposure of plant-derived compounds [34]. Phytoestrogens such as genistein and daidzein are obtained in the diet predominantly from soy beans. Recently, they were shown to inhibit the proliferation of neuroblastoma cells in vitro [35] and therefore may be effective agents in the treatment of certain brain tumors. These compounds are known to have low brain penetration [36]. Studies using mouse and human BCRP overexpressing Madin-Darby canine kidney (MDCK) cell lines clearly demonstrate that genistein and daidzein are substrates for BCRP [37]. In a Bcrp knockout mouse model, the brain penetration of genistein and daidzein was significantly increased in knockout animals compared with wild-type controls, suggesting that the low brain penetration of these compounds is mediated by efflux drug transport by BCRP. Should these compounds be further evaluated for treatment of neuroblastoma, strategies to overcome their low brain penetration need to be addressed.

Nevertheless, it should be noted that the role BCRP plays in mediating drug exposure in the brain remains controversial. Brain penetration of known BCRP substrates mitoxantrone and dehydroepiandrosterone sulfate were shown to be minimally affected in a Bcrp knockout mouse model [38]. Similarly, a recent study showed a minor role of BCRP in the brain distribution of cimetidine, alfuzosin, dipyridamole, and LY22288220 despite strong in vitro evidence in Bcrp transfected cell lines [39]. This may be in part related to the overlap in the substrate specificity between BCRP and P-gp. Accordingly, caution is warranted when interpreting results from in vitro experiments and extrapolating them to in vivo human situation.

Multidrug-resistance-associated protein 4 (MRP4/ABCC4)

MRP4 is an efflux transporter known to play a role in the extrusion of organic anions and antiretroviral drugs and has also been implicated in multidrug resistance during chemotherapy. In the brain, MRP4 is expressed on the luminal membrane of capillary endothelial cells and the basolateral membrane of the choroid plexus [40, 41]. In an Mrp4 knockout mouse model, the chemotherapeutic drug topotecan was not only elevated in the brain but also in the CSF, emphasizing the important role that Mrp4 plays in determining the CNS distribution of this drug [40]. It is likely that multiple efflux drug transporters including MRP4 govern the brain penetration and activity of this anticancer agent.

Similar to OAT3, MRP4 has been implicated in CNS exposure to the active metabolite of oseltamivir, Ro 64-0802. Both brain exposure and brain to plasma ratio of Ro 64-0802 are significantly increased in Mrp4 knockout mice compared with wild-type control [17]. As there are several known polymorphisms in MRP4 that are known to affect transport activity, it would be interesting to evaluate their impact on the CNS side effects of Ro 64-0802 in humans. It is likely that the interplay of OAT3 and MRP4 activity in the brain determines the brain exposure of oseltamivir and Ro 64-0802, and their expression or function may predict the risk for CNS toxicity associated with this drug.

Summary

The endothelial cells that line the brain limit the CNS entry of many drugs in clinical use today. Although this barrier is beneficial in terms of its role in preventing xenobiotic-associated CNS toxicity, it can often block the delivery of drugs needed for optimal treatment of CNS diseases. It is clear that for many drugs, their CNS entry is determined by the interplay of certain uptake and efflux transport proteins expressed at the blood−brain barrier. This will allow us to optimize therapy and prevent the toxicity commonly associated with some pharmacological interventions. Indeed, although the efflux transporter P-gp appears to be the master gatekeeper with respect to limiting drug penetration in the brain, recent evidence also suggests important and emerging roles for BCRP and MRP4, as well as uptake transporters such as OATP1A2. Clearly, additional studies are needed to further our understanding of the clinical relevance of drug transporters expressed at the level of the blood−brain barrier to CNS exposure and efficacy of drug substrates.