Chapter 5: Membrane Transporters and Drug Response

Transporters in toxicological target organs or at barriers to such organs affect exposure of the target organs to drugs. Transporters expressed in tissues that may be targets for drug toxicity (e.g., brain) or in barriers to such tissues (e.g., the blood-brain barrier [BBB]) can tightly control local drug concentrations and thus control the exposure of these tissues to the drug (Figure 5–3, middle panel). For example, to restrict the penetration of compounds into the brain, endothelial cells in the BBB are closely linked by tight junctions, and some efflux transporters are expressed on the blood-facing (luminal) side. The importance of the ABC transporter multidrug resistance protein (ABCB1, MDR1; P-glycoprotein, P-gp) in the BBB has been demonstrated in mdr1a knockout mice (Schinkel et al., 1994). The brain concentrations of many P-glycoprotein substrates, such as digoxin, used in the treatment of heart failure (Chapter 28), and cyclosporin A (Chapter 35), an immunosuppressant, are increased dramatically in mdr1a(−/−) mice, whereas their plasma concentrations are not changed significantly.

Another example of transporter control of drug exposure can be seen in the interactions of loperamide and quinidine. Loperamide is a peripheral opioid used in the treatment of diarrhea and is a substrate of P-glycoprotein. Co-administration of loperamide and the potent P-glycoprotein inhibitor quinidine results in significant respiratory depression, an adverse response to the loperamide (Sadeque et al., 2000). Because plasma concentrations of loperamide are not changed in the presence of quinidine, it has been suggested that quinidine inhibits P-glycoprotein in the BBB, resulting in an increased exposure of the CNS to loperamide and bringing about the respiratory depression. Inhibition of P-glycoprotein-mediated efflux in the BBB thus would cause an increase in the concentration of substrates in the CNS and potentiate adverse effects.

The case of oseltamivir (the antiviral drug TAMIFLU) provides an example that dysfunction of an active barrier may cause a CNS effect. Abnormal behavior appears to be a rare adverse reaction of oseltamivir. Oseltamivir and its active form, Ro64-0802, undergo active efflux across the BBB by P-glycoprotein, organic anion transporter 3 (OAT3), and multidrug resistance-associated protein 4 (MRP4) (Ose et al., 2009). Decreased activities of these transporters at the BBB caused by concomitant drugs, ontogenetic and genetic factors, or disease may enhance the CNS exposure to oseltamivir and Ro64-0802, contributing to an adverse effect on the CNS.

Drug-induced toxicity sometimes is caused by the concentrative tissue distribution mediated by influx transporters. For example, biguanides (e.g., metformin and phenformin), widely used as oral hypoglycemic agents for the treatment of type II diabetes mellitus, can produce lactic acidosis, a lethal side effect. Phenformin was withdrawn from the market for this reason. Biguanides are substrates of the organic cation transporter OCT1, which is highly expressed in the liver. After oral administration of metformin, the distribution of the drug to the liver in oct1(−/−) mice is markedly reduced compared with the distribution in wild-type mice. Moreover, plasma lactic acid concentrations induced by metformin are reduced in oct1(−/−) mice compared with wild-type mice, although the plasma concentrations of metformin are similar in the wild-type and knockout mice. These results indicate that the OCT1-mediated hepatic uptake of biguanides plays an important role in lactic acidosis (Wang et al., 2003).

The organic anion transporter 1 (OAT1) and organic cation transporters (OCT1 and OCT2) provide other examples of transporter-related toxicity. OAT1 is expressed mainly in the kidney and is responsible for the renal tubular secretion of anionic compounds. Substrates of OAT1, such as cephaloridine (a β-lactam antibiotic), and adefovir and cidofovir (antiviral drugs), reportedly cause nephrotoxicity. In vitro experiments suggest that cephaloridine, adefovir, and cidofovir are substrates of OAT1 and that OAT1-expressing cells are more susceptible to the toxicity of these drugs than control cells (Ho et al., 2000; Takeda et al., 1999). Exogenous expression of OCT1 and OCT2 enhances the sensitivities of tumor cells to the cytotoxic effect of oxaliplatin for OCT1, and cisplatin and oxaliplatin for OCT2 (Zhang et al., 2006b).

Drugs may modulate transporters for endogenous ligands and thereby exert adverse effects (Figure 5–3, bottom panel). For example, bile acids are taken up mainly by Na⁺-taurocholate cotransporting polypeptide (NTCP) (Hagenbuch et al., 1991) and excreted into the bile by the bile salt export pump (BSEP, ABCB11) (Gerloff et al., 1998). Bilirubin is taken up by OATP1B1 and conjugated with glucuronic acid, and bilirubin glucuronide is excreted by the multidrug-resistance-associated protein (MRP2, ABCC2). Inhibition of these transporters by drugs may cause cholestasis or hyperbilirubinemia. Troglitazone, a thiazolidinedione insulin-sensitizing drug used for the treatment of type II diabetes mellitus, was withdrawn from the market because it caused hepatotoxicity. The mechanism for this troglitazone-induced hepatotoxicity remains unclear. One hypothesis is that troglitazone and its sulfate conjugate induced cholestasis. Troglitazone sulfate potently inhibits the efflux of taurocholate (K_i = 0.2 μM) mediated by the ABC transporter BSEP. These findings suggest that troglitazone sulfate induces cholestasis by inhibition of BSEP function. BSEP-mediated transport is also inhibited by other drugs, including cyclosporin A and the antibiotics rifamycin and rifampicin (Stieger et al., 2000).

Thus, uptake and efflux transporters determine the plasma and tissue concentrations of endogenous compounds and xenobiotics, and thereby can influence the systemic or site-specific toxicity of drugs.

Such diffusion may be described by the equation:

(Equation 5-1)

where Δμ is the potential gradient, z is the charge valence of the solute, E_m is the membrane voltage, F is the Faraday constant, R is the gas constant, T is the absolute temperature, C is the concentration of the solute inside (i) and outside (o) of the plasma membrane. The first term on the right side in Equation (5–1) represents the electrical potential, and the second represents the chemical potential.

For non-ionized compounds, the flux J owing to simple diffusion is given by Fick's first law (permeability multiplied by the concentration difference). For ionized compounds, the difference in electrical potential across the plasma membrane needs to be taken into consideration. Assuming that the electrical field is constant, the flux is given by the Goldman-Hodgkin-Katz equation:

(Equation 5-2)

where P represents the permeability. The lipid and water solubility and the molecular weight and shape of the solute are determinants of the flux in passive diffusion; they are incorporated in the permeability constant P. The permeability constant positively correlates with the lipophilicity, determined by the partition between water and organic solvents, such as octanol, and is also related to the inverse of the square root of the molecular weight of the solute. At steady state, the electrochemical potentials of all compounds become equal across the plasma membrane. In the case of non-ionized compounds, the steady-state concentrations are equal across the plasma membrane. For ionized compounds, however, the steady-state concentration ratio across the plasma membrane is affected by the membrane voltage and given by the Nernst equation:

(Equation 5-3)

The membrane voltage is maintained by the ion gradients across the membrane.

Depending on the transport direction of the solute, secondary active transporters are classified as either symporters or antiporters. Symporters, also termed co-transporters, transport S₂ and S₁ in the same direction, whereas antiporters, also termed exchangers, move their substrates in opposite directions (Figure 5–4). The free energy produced by one extracellular sodium ion (Na⁺) is given by the difference in the electrochemical potential across the plasma membrane:

(Equation 5-4)

The electrochemical potential of a non-ionized compound Δμ_sacquired from one extracellular Na⁺ is less than this value:

(Equation 5-5)

Therefore, the concentration ratio of the compound is given by the following equation:

(Equation 5-6)

Assuming that the concentration ratio of Na⁺ is 10 and that E_m is −60 mV, ideally, symport of one non-ionized organic compound with one Na⁺ ion can achieve a 100-fold difference in the intracellular substrate concentration compared with the extracellular concentration. When more than one Na⁺ ion is coupled to the movement of the solute, a synergistic driving force results. For the case in which two Na⁺ ions are involved,

(Equation 5-7)

In this case, the substrate ideally is concentrated intracellularly 1000-fold relative to the extracellular space under the same conditions. The Na⁺/Ca²⁺ antiporter shows the effect of this dependence in the square of the concentration ratio of Na⁺; Ca²⁺ is transported from the cytosol (0.1 μM < [Ca²⁺] < 1 μM) to the plasma [Ca²⁺]_free ∼1.25 mM.

In the liver, a number of transporters with different substrate specificities are localized on the sinusoidal membrane (facing blood). These transporters are involved in the uptake of bile acids, amphipathic organic anions, and hydrophilic organic cations into the hepatocytes. Similarly, ABC transporters on the canalicular membrane (facing bile) export such compounds into the bile. Multiple combinations of uptake (OATP1B1, OATP1B3, OATP2B1) and efflux transporters (MDR1, MRP2, and BCRP) are involved in the efficient transcellular transport of a wide variety of compounds in the liver by using a model cell system called "doubly transfected cells," which express both uptake and efflux transporter on each side (Ishiguro et al., 2008; Kopplow et al., 2005; Matsushima et al., 2005). In many cases, overlapping substrate specificities between the uptake transporters (OATP family) and efflux transporters (MRP family) make the vectorial transport of organic anions highly efficient. Similar transport systems also are present in the intestine, renal tubules, and endothelial cells of the brain capillaries (Figure 5–5).

DNA methylation is one of the mechanisms underlying the epigenetic control of gene expression. Reportedly, the tissue-selective expression of transporters is achieved by DNA methylation (silencing in the transporter-negative tissues) as well as by transactivation in the transporter-positive tissues. Transporters subjected to epigenetic control include OAT3, URAT1, OCT2, Oatp1b2, Ntcp, and PEPT2 in the SLC families; and MDR1, BCRP, BSEP, and ABCG5/ABCG8 (Aoki et al., 2008; Imai et al., 2009; Kikuchi et al., 2006; Turner et al., 2006; Uchiumi et al., 1993).

Transporters in the SLC superfamily transport diverse ionic and non-ionic endogenous compounds and xenobiotics. SLC superfamily transporters may be facilitated transporters or secondary active symporters or antiporters. The first SLC family transporter was cloned in 1987 by expression cloning in Xenopus laevis oocytes (Hediger et al., 1987). Since then, many transporters in the SLC superfamily have been cloned and characterized functionally. Predictive models defining important characteristics of substrate binding and knockout mouse models defining the in vivo role of specific transporters have been constructed for many SLC transporters (Chang et al., 2004; Ocheltree et al., 2004).

In 1976, Juliano and Ling reported that overexpression of a membrane protein in colchicine-resistant Chinese hamster ovary cells also resulted in acquired resistance to many structurally unrelated drugs (i.e., multidrug resistance) (Juliano and Ling, 1976). Since the cDNA cloning of this first mammalian ABC protein (P-glycoprotein/MDR1/ABCB1), knowledge of the ABC superfamily has grown to include 49 genes, each containing one or two conserved ABC regions (Borst and Elferink, 2002). The ABC region is a core catalytic domain of ATP hydrolysis and contains Walker A and B sequences and an ABC transporter-specific signature C sequence (Figure 5–6). The ABC regions of these proteins bind and hydrolyze ATP, and the proteins use the energy for uphill transport of their substrates across the membrane. Although some ABC superfamily transporters contain only a single ABC motif, they form homodimers (BCRP/ABCG2) or heterodimers (ABCG5 and ABCG8) that exhibit a transport function. ABC transporters (e.g., MsbA) (Figure 5–7) also are found in prokaryotes, where they are involved predominantly in the import of essential compounds that cannot be obtained by passive diffusion (sugars, vitamins, metals, etc.). By contrast, most ABC genes in eukaryotes transport compounds from the cytoplasm to the outside or into an intracellular compartment (endoplasmic reticulum, mitochondria, peroxisomes).

The kidney and liver are major organs for overall systemic drug elimination from the body. The liver also plays a role in presystemic drug elimination. Key to the vectorial excretion of drugs into urine or bile, ABC transporters are expressed in the polarized tissues of kidney and liver: MDR1, MRP2, and MRP4 (ABCC4) on the brush-border membrane of renal epithelia; MDR1, MRP2, and BCRP on the bile canalicular membrane of hepatocytes; and MRP3 and MRP4 on the sinusoidal membrane of hepatocytes. Some ABC transporters are expressed specifically on the blood side of the endothelial or epithelial cells that form barriers to the free entrance of toxic compounds into tissues: the BBB (MDR1 and MRP4 on the luminal side of brain capillary endothelial cells), the blood-cerebrospinal fluid (CSF) barrier (MRP1 and MRP4 on the basolateral blood side of choroid plexus epithelia), the blood-testis barrier (MRP1 on the basolateral membrane of mouse Sertoli cells and MDR1 in several types of human testicular cells), and the blood-placenta barrier (MDR1, MRP2, and BCRP on the luminal maternal side and MRP1 on the anti-luminal fetal side of placental trophoblasts).

Substrate Specificity of ABC Transporters. MDR1/ABCB1 substrates tend to share a hydrophobic planar structure with positively charged or neutral moieties (see Table 5–4 and Ambudkar et al., 1998). These include structurally and pharmacologically unrelated compounds, many of which are also substrates of CYP3A4, a major drug-metabolizing enzyme in the human liver and GI tract. Such overlapping substrate specificity implies a synergistic role for MDR1 and CYP3A4 in protecting the body by reducing the intestinal absorption of xenobiotics (Zhang and Benet, 2001). After being taken up by enterocytes, some drug molecules are metabolized by CYP3A4. Drug molecules that escape metabolic conversion are eliminated from the cells by MDR1 and then reenter the enterocytes. The intestinal residence time of the drug is prolonged with the aid of MDR1, thereby increasing the chance of local metabolic conversion by the CYP3A4 (see Chapter 6).

MRP/ABCC Family. The substrates of transporters in the MRP/ABCC family are mostly organic anions. The substrate specificities of MRP1 and MRP2 are similar: both accept glutathione and glucuronide conjugates, sulfated conjugates of bile salts, and non-conjugated organic anions of an amphipathic nature (at least one negative charge and some degree of hydrophobicity). They also transport neutral or cationic anticancer drugs, such as vinca alkaloids and anthracyclines, possibly by means of a cotransport or symport mechanism with reduced glutathione (GSH).

MRP3 also has a substrate specificity that is similar to that of MRP2 but with a lower transport affinity for glutathione conjugates compared with MRP1 and MRP2. Most characteristic MRP3 substrates are monovalent bile salts, which are never transported by MRP1 and MRP2. Because MRP3 is expressed on the sinusoidal side of hepatocytes and is induced under cholestatic conditions, backflux of toxic bile salts and bilirubin glucuronides into the blood circulation is considered to be its physiological function.

MRP4 accepts negatively charged molecules, including cytotoxic compounds (e.g., 6-mercaptopurine and methotrexate), cyclic nucleotides, antiviral drugs (e.g., adefovir and tenofovir), diuretics (e.g., furosemide and trichlorothiazide), and cephalosporins (e.g., ceftizoxime and cefazolin). Glutathione enables MRP4 to accept taurocholate and leukotriene B₄.

MRP5 has a narrower substrate specificity and accepts nucleotide analog and clinically important anti–human immunodeficiency virus (HIV) drugs. Although some transport substrates have been identified for MRP6, no physiologically important endogenous substrates have been identified that explain the mechanism of the MRP6-associated disease, pseudoxanthoma.

BCRP/ABCG2. BCRP accepts both neutral and negatively charged molecules, including cytotoxic compounds (e.g., topotecan, flavopiridol, and methotrexate), sulfated conjugates of therapeutic drugs and hormones (e.g., estrogen sulfate), antibiotics (e.g., nitrofurantoin and fluoroquinolones), statins (e.g., pitavastatin and rosuvastatin), and toxic compounds found in normal food [phytoestrogens, 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) and pheophorbide A, a chlorophyll catabolite].

Mice deficient in MDR1 function are viable and fertile and do not display obvious phenotypic abnormalities other than hypersensitivity to toxic drugs, including the neurotoxic pesticide ivermectin (100-fold) and the cytotoxic drug vinblastine (3-fold) (Schinkel et al., 1994). Mrp1 (−/−) mice are also viable and fertile without any obvious difference in litter size. However, these mice are hypersensitive to the anticancer drug etoposide. Damage is especially severe in the testis, kidney, and oropharyngeal mucosa, where MRP1 is expressed on the basolateral membrane. Moreover, these mice have an impaired response to an arachidonic acid–induced inflammatory stimulus, which is likely due to a reduced secretion of leukotriene C₄ from mast cells, macrophages, and granulocytes. MRP2-deficient rats (TR– and EHBR) and Dubin-Johnson syndrome patients are normal in appearance except for mild jaundice owing to impaired biliary excretion of bilirubin glucuronide (B¨chler M et al., 1996; Ito et al., 1997; Paulusma et al., 1996). Mrp4 (−/−) mice are hypersensitive to adefovir and thiopurine. Damage induced by adefovir is especially severe in the bone marrow, thymus, spleen, and intestine, and that by thiopurine is in the myeloid progenitor cells (Belinsky et al., 2007; Krishnamurthy et al., 2008; Takenaka et al., 2007). MRP4 modulates the signal transduction by active efflux of cAMP and cGMP across the plasma membrane. Mrp4 (−/−) mice are sensitive to CFTR-mediated secretory diarrhea (Li et al., 2007), and exhibit an increase in inflammatory pain threshold. In addition, MRP4 is associated with proliferation of smooth muscle through modulation of cAMP/cGMP signaling (Sassi et al., 2008).

BCRP (−/−) mice are viable but highly sensitive to the dietary chlorophyll catabolite phenophorbide, which induces phototoxicity (Jonker et al., 2002). These mice also exhibit protoporphyria, with a 10-fold increase in protoporphyrin IX accumulation in erythrocytes, resulting in photosensitivity. This protoporphyria is caused by the impaired function of BCRP in bone marrow: knockout mice transplanted with bone marrow from wild-type mice become normal with respect to protoporphyrin IX level in the erythrocytes and photosensitivity.

BSEP has a narrower substrate specificity. It accepts bile acids (Gerloff et al., 1998) and its physiological role is to provide the osmotic driving force for bile flow by biliary excretion of bile acids in an ATP-dependent manner. Hereditary functional defect of this transporter results in the acquisition of progressive familial intrahepatic cholestasis (PFIC2), a fatal liver disease (Strautnieks et al., 1998). In addition, several drugs such as cyclosporin A inhibit the transport function of BSEP at clinically relevant concentrations, thereby causing the disruption of bile formation usually called drug-induced cholestasis (Stieger et al., 2000).

ABC Transporters in Drug Absorption and Elimination. With respect to clinical medicine, MDR1 is the most important ABC transporter yet identified, and digoxin is one of the most widely studied of its substrates. The systemic exposure to orally administered digoxin (as assessed by the area under the plasma digoxin concentration–time curve) is decreased by coadministration of rifampin (an MDR1 inducer) and is negatively correlated with the MDR1 protein expression in the human intestine. MDR1 is also expressed on the brush-border membrane of renal epithelia, and its function can be monitored using digoxin as a probe. Digoxin undergoes very little degradation in the liver, and renal excretion is the major elimination pathway (>70%) in humans. Several studies in healthy subjects have been performed with MDR1 inhibitors (e.g., quinidine, verapamil, vaspodar, spironolactone,clarythromycin, and ritonavir) with digoxin as a probe drug, and all resulted in a marked reduction in the renal excretion of digoxin.

The intestinal absorption of cyclosporine is also related mainly to the MDR1 level rather than to the CYP3A4 level, although cyclosporine is a substrate of both CYP3A4 and MDR1. Alteration of MDR1 activity by inhibitors (drug-drug interactions) affects oral absorption and renal clearance. Drugs with narrow therapeutic windows (such as the cardiac glycoside digoxin and the immunosuppressants cyclosporine and tacrolimus) should be used with great care if MDR1-based drug-drug interactions are likely.

Despite the broad substrate specificity and distinct localization of MRP2 and BCRP in drug-handling tissues (both are expressed on the canalicular membrane of hepatocytes and the brush-border membrane of enterocytes), there has been little integration of clinically relevant information. Part of the problem lies in distinguishing the biliary transport activities of MRP2 and BCRP from the contribution of the hepatic uptake transporters of the OATP family. Most MRP2 or BCRP substrates also can be transported by the OATP family transporters on the sinusoidal membrane. The rate-limiting process for systemic elimination is uptake in most cases. Under such conditions, the effect of drug-drug interactions (or genetic variants) in these biliary transporters may be difficult to identify. Despite such practical difficulties, there is a steady increase in the information about genetic variants and their effects on transporter expression and activity in vitro. Variants of BCRP with high allele frequencies (0.184 for V12M and 0.239 for Q141K) have been found to alter the protein expression in cellular assays. One variant is associated with greater oral availability of rosuvastatin and sulfasalazine following oral administration although the impact of the variant on the intestinal absorption and biliary excretion has not been separately evaluated (Yamasaki et al., 2008; Zhang et al., 2006c).

MRP3 is characterized by its localization in the plasma membrane of epithelial cells facing blood (basolateral membrane in the intestine epithelial cells, and sinusoidal membrane in the liver). In the intestine, MRP3 can mediate the intestinal absorption in conjunction with uptake transporters. MRP3 mediates sinusoidal efflux in the liver, decreasing the efficacy of the biliary excretion from the blood, and excretion of intracellularly formed metabolites, particularly, glucuronide conjugates. Thus, dysfunction of MRP3 results in a shortening of the elimination half-life (Kitamura et al., 2008; Zelcer et al., 2005).

MRP4 substrates also can be transported by the OAT family transporters (OAT1 and OAT3) on the basolateral membrane of the epithelial cells in the kidney. MRP4 is involved in the directional transport in conjunction with OAT1 and/or OAT3 in the kidney (Hasegawa et al., 2007). The rate-limiting process in renal tubular secretion is likely the uptake process at the basolateral surface. Dysfunction of MRP4 enhances the renal concentration, but has limited impact on the blood concentration.

HMG-CoA Reductase Inhibitors. Statins are cholesterol-lowering agents that reversibly inhibit HMG-CoA reductase, which catalyzes a rate-limiting step in cholesterol biosynthesis (see Chapter 31). Statins affect serum cholesterol by inhibiting cholesterol biosynthesis in the liver, and this organ is their main target. On the other hand, exposure of extrahepatic cells in smooth muscle to these drugs may cause adverse effects. Among the statins, pravastatin, fluvastatin, cerivastatin, atorvastatin, rosuvastatin, and pitavastatin are given in a biologically active open-acid form, whereas simvastatin and lovastatin are administered as inactive prodrugs with lactone rings. The open-acid statins are relatively hydrophilic and have low membrane permeabilities. However, most of the statins in the acid form are substrates of uptake transporters, so they are taken up efficiently by the liver and undergo enterohepatic circulation (Figures 5–5 and 5–11). In this process, hepatic uptake transporters such as OATP1B1 and efflux transporters such as MRP2 act cooperatively to produce vectorial trans-cellular transport of bisubstrates in the liver. The efficient first-pass hepatic uptake of these statins by OATP1B1 after their oral administration helps to exert the pharmacological effect and also minimizes the escape of drug molecules into the circulating blood, thereby minimizing the exposure in a target of adverse response, smooth muscle. Recent studies indicate that the genetic polymorphism of OATP1B1 also affects the function of this transporter (Tirona et al., 2001).

Temocapril. Temocapril is an ACE inhibitor (see Chapter 26). Its active metabolite, temocaprilat, is excreted both in the bile and in the urine by the liver and kidney, respectively, whereas other ACE inhibitors are excreted mainly by the kidney. The special feature of temocapril among ACE inhibitors is that the plasma concentration of temocaprilat remains relatively unchanged even in patients with renal failure. However, the plasma area under the curve (AUC) of enalaprilat and other ACE inhibitors is markedly increased in patients with renal disorders. Temocaprilat is a bisubstrate of the OATP family and MRP2, whereas other ACE inhibitors are not good substrates of MRP2 (although they are taken up into the liver by the OATP family). Taking these findings into consideration, the affinity for MRP2 may dominate in determining the biliary excretion of any series of ACE inhibitors. Drugs that are excreted into both the bile and urine to the same degree thus are expected to exhibit minimum inter-individual differences in their pharmacokinetics.

Irinotecan (CPT-11). Irinotecan hydrochloride (CPT-11) is a potent anticancer drug, but late-onset gastrointestinal toxic effects, such as severe diarrhea, make it difficult to use CPT-11 safely. After intravenous administration, CPT-11 is converted to SN-38, an active metabolite, by carboxylesterase. SN-38 is subsequently conjugated with glucuronic acid in the liver. SN-38 and SN-38 glucuronide are then excreted into the bile by MRP2. The inhibition of MRP2-mediated biliary excretion of SN-38 and its glucuronide by co-administration of probenecid reduces the drug-induced diarrhea, at least in rats. For additional details, see Figures 6–5 and 6–7.

Angiotensin II Receptor Antagonists. Angiotensin II receptor antagonists are used for the treatment of hypertension, acting on AT₁ receptors expressed in vascular smooth muscle, proximal tubule, and adrenal medullary cells, and elsewhere. For most of these drugs, hepatic uptake and biliary excretion are important factors for their pharmacokinetics as well as pharmacological effects. Telmisartan is taken up into human hepatocytes in a saturable manner, predominantly via OATP1B3 (Ishiguro et al., 2006). On the other hand, both OATPs 1B1 and 1B3 are responsible for the hepatic uptake of valsartan and olmesartan, although the relative contributions of these transporters are unclear. Studies using doubly transfected cells with hepatic uptake transporters and biliary excretion transporters have clarified that MRP2 plays the most important role in the biliary excretion of valsartan and olmesartan.

Repaglinide and Nateglinide. Repaglinide is a meglitinide analog anti-diabetic drug. Although it is eliminated almost completely by the metabolism mediated by CYPs 2C8 and 3A4, transporter-mediated hepatic uptake is one of the determinants of its elimination rate. In subjects with SLCO1B1 (gene code for OATP1B1) 521CC genotype, a significant change in the pharmacokinetics of repaglinide was observed (Niemi et al., 2005). Genetic polymorphism in SLCO1B1 521T>C results in altered pharmacokinetics of nateglinide, suggesting OATP1B1 is a determinant of its elimination, although it is subsequently metabolized by CYPs 2C9, 3A4, and 2D6 (Zhang et al., 2006a).

Fexofenadine. Fexofenadine, a histamine H₁ receptor antagonist, is taken up in the liver by OATP1B1 and OATP1B3 and excreted in the bile via transporters including MRP2 and BSEP (Matsushima et al., 2008). Patients with genetic polymorphism in SLCO1B1 521T>C, show altered pharmacokinetics.

Bosentan. Bosentan is an endothelin antagonist used to treat pulmonary arterial hypertension. It is taken up in the liver by OATP1B1 and OATP1B3, and subsequently metabolized by CYP2C9 and CYP3A4 (Treiber et al., 2007). Transporter-mediated hepatic uptake can be a determinant of elimination of bosentan, and the inhibition of hepatic uptake by cyclosporin A, rifampicin, and sildenafil can affect its pharmacokinetics.

Cerivastatin (currently withdrawn), an HMG-CoA reductase inhibitor, is taken up into the liver via transporters (especially OATP1B1) and subsequently metabolized by CYPs 2C8 and 3A4. Its plasma concentration is increased 1-5 fold when co-administered with cyclosporin A. Transport studies using cryopreserved human hepatocytes and OATP1B1-expressing cells suggest that this drug-drug interaction is caused by inhibition of OATP1B1-mediated hepatic uptake (Shitara et al., 2003). However, cyclosporin A inhibits the metabolism of cerivastatin only to a limited extent, suggesting a low possibility of serious drug-drug interactions involving the inhibition of metabolism. Cyclosporin A also increases the plasma concentrations of other HMG-CoA reductase inhibitors. It markedly increases the plasma AUC of pravastatin, pitavastatin, and rosuvastatin, which are minimally metabolized and eliminated from the body by transporter-mediated mechanisms. Therefore, these pharmacokinetic interactions also may be due to transporter-mediated hepatic uptake. However, the interactions of cyclosporin A with pro-drug statins (lactone form) such as simvastatin and lovastatin are mediated by CYP3A4.

Gemfibrozil is another cholesterol-lowering agent that acts by a different mechanism and also causes a severe pharmacokinetic interaction with cerivastatin. Gemfibrozil glucuronide inhibits the CYP2C8-mediated metabolism and OATP1B1-mediated uptake of cerivastatin more potently than does gemfibrozil. Laboratory data show that the glucuronide is highly concentrated in the liver versus plasma probably owing to transporter-mediated active uptake and intracellular formation of the conjugate. Therefore, it may be that gemfibrozil glucuronide, concentrated in the hepatocytes, inhibits the CYP2C8-mediated metabolism of cerivastatin. In addition, gemfibrozil glucuronide is an inhibitor of CYP2C8. Gemfibrozil markedly (4-5 fold) increases the plasma concentration of cerivastatin but does not greatly increase (1.3-2 times) that of unmetabolized statins pravastatin, pitavastatin, and rosuvastatin, a result that also suggests that this interaction is caused by inhibition of metabolism. Thus, when an inhibitor of drug-metabolizing enzymes is highly concentrated in hepatocytes by active transport, extensive inhibition of the drug-metabolizing enzymes may be observed because of the high concentration of the inhibitor in the vicinity of the drug-metabolizing enzymes.

The contribution of transporters to hepatic uptake can be estimated from in vitro studies. Injection of cRNA results in transporter expression on the plasma membrane of Xenopus laevis oocytes (Hagenbuch et al., 1996). Subsequent hybridization of the cRNA with its antisense oligonucleotide specifically reduces its expression. Comparison of the drug uptake into cRNA-injected oocytes in the presence and absence of antisense oligonucleotides clarifies the contribution of a specific transporter. Second, a method using reference compounds for specific transporters has been proposed. The reference compounds should be specific substrates for a particular transporter. The contribution of a specific transporter can be calculated from the uptake of test compounds and reference compounds into hepatocytes and transporter-expressing systems (Hirano et al., 2004):

(Equation 5-13)

where CL_hep,ref and CL_exp,ref represent the uptake of reference compounds into hepatocytes and transporter-expressing cells, respectively, and CL_hep,test and CL_exp,test represent the uptake of test compounds into the corresponding systems. For example, the contributions of OATP1B1 and OATP1B3 to the hepatic uptake of pitavastatin have been estimated using estrone-3-sulfate and cholecystokinine octapeptide (CCK8) as reference compounds for OATP1B1 and OATP1B3, respectively. However, for many transporters, reference compounds specific to the transporter are not available. Another approach to estimate the relative contribution of OATP1B1 and OATP1B3 is to use estrone-3-sulfate as a selective inhibitor for OATP1B1 (Ishiguro et al., 2006). The difference in uptake clearance of test compound in human hepatocytes in the absence and presence of estrone-3-sulfate represents OATP1B1-mediated hepatic uptake.

Transport of organic cations from cell to tubular lumen across the apical membrane occurs via an electroneutral proton–organic cation exchange mechanism in a variety of species, including human, dog, rabbit, and cat. The recent discovery of a new transporter family, SLC47A, multidrug and toxin extrusion family (MATEs), has provided the molecular identities of the previously characterized electroneutral proton–organic cation antiport mechanism (Otsuka et al., 2005; Tanihara et al., 2007). Transporters in the MATE family, assigned to the apical membrane of the proximal tubule, appear to play a key role in moving hydrophilic organic cations from tubule cell to lumen. In addition, novel organic cation transporters (OCTNs), located on the apical membrane, appear to contribute to organic cation flux across the proximal tubule. In humans, these include OCTN1 (SLC22A4) and OCTN2 (SLC22A5). These bifunctional transporters are involved not only in organic cation secretion but also in carnitine reabsorption. In the reuptake mode, the transporters function as Na⁺ co-transporters, relying on the inwardly driven Na⁺ gradient created by Na⁺,K⁺-ATPase to move carnitine from tubular lumen to cell. In the secretory mode, the transporters appear to function as proton–organic cation exchangers. That is, protons move from tubular lumen to cell interior in exchange for organic cations, which move from cytosol to tubular lumen. The inwardly directed proton gradient (from tubular lumen to cytosol) is maintained by transporters in the SLC9 family, which are Na⁺/H⁺ exchangers (NHEs, antiporters). Of the two steps involved in secretory transport, transport across the luminal membrane appears to be rate-limiting.

OCT2 (SLC22A2). OCT2 (SLC22A2) was first cloned from a rat kidney cDNA library in 1996 (Okuda et al., 1996). Human, rabbit, mouse, and pig orthologs all have been cloned. Mammalian orthologs range in length from 553 through 555 amino acids. OCT2 is predicted to have 12 transmembrane domains, including one N-linked glycosylation site. OCT2 is located adjacent to OCT1 on chromosome 6 (6q26). A single splice variant of human OCT2, termed OCT2-A, has been identified in human kidney. OCT2-A, which is a truncated form of OCT2, appears to have a lower K_m (or greater affinity) for substrates than OCT2, although a lower affinity has been observed for some inhibitors (Urakami et al., 2002). Human, mouse, and rat orthologs of OCT2 are expressed in abundance in human kidney and to some extent in neuronal tissue such as choroid plexus. In the kidney, OCT2 is localized to the proximal tubule and to distal tubules and collecting ducts. In the proximal tubule, OCT2 is restricted to the basolateral membrane. OCT2 mammalian species orthologs are > 80% identical, whereas the OCT2 paralog found primarily in the liver, OCT1, is ∼70% identical to OCT2. OCT2-mediated transport of model organic cations MPP⁺ and TEA is electrogenic, and both OCT2 and OCT1 can support organic cation–organic cation exchange (Koepsell et al., 2007). OCT2 generally accepts a wide array of monovalent organic cations with molecular weights < 400 daltons (Ciarimboli, 2008; Koepsell et al., 2007). The apparent affinities of the human paralogs, OCT1 and OCT2, for some organic cation substrates and inhibitors are different in side-by-side comparison studies. Isoform-specific inhibitors of the OCTs are needed to determine the relative importance of OCT2 and OCT1 in the renal clearance of compounds in rodents, in which both isoforms are present in kidney. OCT2 is also present in neuronal tissues; however, monoamine neurotransmitters have low affinities for OCT2. OCT2 may play a housekeeping role in neurons, taking up only excess concentrations of neurotransmitters. OCT2 also may be involved in recycling of neurotransmitters by taking up breakdown products, which in turn enter monoamine synthetic pathways.

OCT3 (SLC22A3). OCT3 (SLC22A3) was cloned initially from rat placenta (Kekuda et al., 1998). Human and mouse orthologs have also been cloned. OCT3 consists of 551 amino acids and is predicted to have 12 transmembrane domains, including three N-linked glycosylation sites. hOCT3 is located in tandem with OCT1 and OCT2 on chromosome 6. Tissue distribution studies suggest that human OCT3 is expressed in liver, kidney, intestine, and placenta, although it appears to be expressed in considerably less abundance than OCT2 in the kidney. Like OCT1 and OCT2, OCT3 appears to support electrogenic potential-sensitive organic cation transport. Although the specificity of OCT3 is similar to that of OCT1 and OCT2, it appears to have quantitative differences in its affinities for many organic cations. Some studies have suggested that OCT3 is the extraneuronal monoamine transporter based on its substrate specificity and potency of interaction with monoamine neurotransmitters. Because of its relatively low abundance in the kidney (in the basolateral membrane of the proximal tubule (Koepsell et al., 2007)), OCT3 may play only a limited role in renal drug elimination.

OCTN1 (SLC22A4). OCTN1, cloned originally from human fetal liver, is expressed in the adult kidney, trachea, and bone marrow (Tamai et al., 1997). The functional characteristics of OCTN1 suggest that it operates as an organic cation–proton exchanger. OCTN1-mediated influx of model organic cations is enhanced at alkaline pH, whereas efflux is increased by an inwardly directed proton gradient. OCTN1 contains a nucleotide-binding sequence motif, and transport of its substrates appears to be stimulated by cellular ATP. OCTN1 also can function as an organic cation–organic cation exchanger. Although the subcellular localization of OCTN1 has not been demonstrated clearly, available data collectively suggest that OCTN1 functions as a bidirectional pH- and ATP-dependent transporter at the apical membrane in renal tubular epithelial cells. OCTN1 appears to transport the anti-epileptic agent, gabapentin, in the kidney (Urban et al., 2007).

OCTN2 (SLC22A5). OCTN2 was first cloned from human kidney and determined to be the transporter responsible for systemic carnitine deficiency (Tamai et al., 1998). Rat OCTN2 mRNA is expressed predominantly in the cortex, with very little expression in the medulla, and is localized to the apical membrane of the proximal tubule.

OCTN2 is a bifunctional transporter, i.e., it transports L-carnitine with high affinity in an Na⁺-dependent manner, whereas, Na⁺ does not influence OCTN2-mediated transport of organic cations such as TEA. Thus, OCTN2 is thought to function as both an Na⁺-dependent carnitine transporter and an Na⁺-independent organic cation transporter. Similar to OCTN1, OCTN2 transport of organic cations is sensitive to pH, suggesting that OCTN2 may function as an organic cation exchanger. The transport of L-carnitine by OCTN2 is a Na⁺-dependent electrogenic process, and mutations in OCTN2 appear to be the cause of primary systemic carnitine deficiency (Nezu et al., 1999).

MATE1 and MATE2-K (SLC47A1 and SLC47A2) Database searches for human orthologs of bacterial multidrug resistance transporters have identified two genes in the human genome that code for membrane transporters (Otsuka et al., 2005). Multidrug and toxin extrusion family members MATE1 (SLC47A1) and MATE2-K (SLC47A2) interact with structurally diverse hydrophilic organic cations including the anti-diabetic drug metformin, the H₂ antagonist cimetidine, and the anticancer drug, topotecan (Tanihara et al., 2007). In addition to cationic compounds, the transporters also recognize some anions, including the antiviral agents acyclovir and ganciclovir. The zwitterions cephalexin and cephradine are specific substrates of MATE1, but not MATE2-K. The herbicide paraquat, a bisquaternary ammonium compound, which is nephrotoxic in humans, is a potent substrate of MATE1 (Chen et al., 2007). Both MATE1 and MATE2-K have been localized to the apical membrane of the proximal tubule (Tanihara et al., 2007). MATE1, but not MATE2-K, is also expressed on the canalicular membrane of the hepatocyte. These transporters appear to be the long-searched-for organic cation proton antiporters on the apical membrane of the proximal tubule, i.e., an oppositely directed proton gradient can drive the movement of organic cations via MATE1 or MATE2-K. The antibiotics, levofloxacin and ciprofloxacin, though potent inhibitors, are not translocated by either MATE1 or MATE2-K.

Polymorphisms of OCTs. Polymorphisms of OCTs have been identified in large post–human genome SNP discovery projects (Kerb et al., 2002; Leabman et al., 2003; Shu et al., 2003). OCT1 exhibits the greatest number of amino acid polymorphisms, followed by OCT2 and then OCT3. Furthermore, allele frequencies of OCT1 amino acid variants in human populations generally are greater than those of OCT2 and OCT3 amino acid variants. Functional studies of OCT1 and OCT2 polymorphisms have been performed. OCT1 exhibits five variants with reduced function. These variants may have important implications clinically in terms of hepatic drug disposition and targeting of OCT1 substrates. In particular, individuals with OCT1 variants may have reduced liver uptake of OCT1 substrates and therefore reduced metabolism. Recent studies suggest that genetic variants of OCT1 and OCT2 are associated with alterations in the renal elimination and response to the anti-diabetic drug, metformin (Shu et al., 2007; Song et al., 2008; Wang et al., 2008).

A current model for the transepithelial flux of organic anions in the proximal tubule is shown in Figure 5–13. Two primary transporters on the basolateral membrane mediate the flux of organic anions from interstitial fluid to tubule cell: OAT1 (SLC22A6) and OAT3 (SLC22A8). Energetically, hydrophilic organic anions are transported across the basolateral membrane against an electrochemical gradient in exchange with intracellular α-ketoglutarate, which moves down its concentration gradient from cytosol to blood. The outwardly directed gradient of α-ketoglutarate is maintained at least in part by a basolateral Na⁺-dicarboxylate transporter (NaDC3). The Na⁺ gradient that drives NaDC3 is maintained by Na⁺,K⁺-ATPase. Transport of small-molecular-weight organic anions by the cloned transporters OAT1 and OAT3 can be driven by α-ketoglutarate; coupled transport of α-ketoglutarate and small-molecular-weight organic anions (e.g., p-aminohippurate) occurs in isolated basolateral membrane vesicles. The molecular pharmacology and molecular biology of OATs have recently been reviewed (El-Sheikh et al., 2008; Srimaroeng et al., 2008).

The mechanism responsible for the apical membrane transport of organic anions from tubule cell cytosol to tubular lumen remains controversial. Some studies suggest that OAT4 may serve as the luminal membrane transporter for organic anions. However, recent studies show that the movement of substrates via this transporter can be driven by exchange with α-ketoglutarate, suggesting that OAT4 may function in the reabsorptive, rather than secretory, flux of organic anions. Other studies have suggested that in the pig kidney, OATV1 serves as an electrogenic facilitated transporter on the apical membrane (Jutabha et al., 2003). The human ortholog of OATV1 is NPT1, or NaPi-1, originally cloned as a phosphate transporter. NPT1 can support the low-affinity transport of hydrophilic organic anions such as PAH. Other transporters that may play a role in transport across the apical membrane include MRP2 and MRP4, multidrug-resistance transporters in the ATP binding cassette family C (ABCC). Both transporters interact with some organic anions and may actively pump their substrates from tubule cell cytosol to tubular lumen.

OAT1 (SLC22A6). OAT1 was cloned from rat kidney (Sekine et al., 1997; Sweet et al., 1997). This transporter is > 30% identical to OCTs in the SLC22 family. Mouse, human, pig, and rabbit orthologs have been cloned and are ∼80% identical to human OAT1. Mammalian isoforms of OAT1 vary in length from 545-551 amino acids. The gene for the human OAT1 is mapped to chromosome 11 and is found in an SLC22 cluster that includes OAT3 and OAT4. There are four splice variants in human tissues, termed OAT1-1, OAT1-2, OAT1-3, and OAT1-4. OAT1-2, which includes a 13-amino-acid deletion, transports PAH at a rate comparable with OAT1-1. These two splice variants use the alternative 5′-splice sites in exon 9. OAT1-3 and OAT1-4, which result from a 132-bp (44-amino-acid) deletion near the carboxyl terminus of OAT1, do not transport PAH. In humans, rat, and mouse, OAT1 is expressed primarily in the kidney, with some expression in brain and skeletal muscle.

Immunohistochemical studies suggest that OAT1 is expressed on the basolateral membrane of the proximal tubule in human and rat, with highest expression in the middle segment, S2. Based on quantitative PCR, OAT1 is expressed at a third of the level of OAT3. OAT1 exhibits saturable transport of organic anions such as PAH. This transport is trans-stimulated by other organic anions, including α-ketoglutarate. Thus, the inside negative-potential difference drives the efflux of the dicarboxylate α-ketoglutarate, which, in turn, supports the influx of monocarboxylates such as PAH. Regulation of expression levels of OAT1 in the kidney appears to be controlled by sex steroids.

OAT1 generally transports small-molecular-weight organic anions that may be endogenous (e.g., PGE₂ and urate) or ingested drugs and toxins. Some neutral compounds are also transported by OAT1 at a lower affinity (e.g., cimetidine). Key residues that contribute to transport by OAT1 include the conserved K394 and R478, which are involved in the PAH–glutarate exchange mechanism.

OAT2 (SLC22A7) OAT2 was cloned first from rat liver (and named NLT at the time) (Sekine et al., 1998; Simonson et al., 1994). OAT2 is present in both kidney and liver. In the kidney, the transporter is localized to the basolateral membrane of the proximal tubule, and appears to function as a transporter for nucleotides, particularly guanine nucleotides such as cyclic GMP (Cropp et al., 2008). Cellular studies suggest that OAT2 functions in both the influx and efflux of guanine nucleotides. Organic anions such as PAH and methotrexate are also transported with low affinity by OAT2, which transports PGE₂ with a high affinity.

OAT3 (SLC22A8). OAT3 (SLC22A8) was cloned originally from rat kidney (Kusuhara et al., 1999). Human OAT3 consists of two variants, one of which transports a wide variety of organic anions, including model compounds, PAH and estrone sulfate, as well as many drug products (e.g., pravastatin, cimetidine, 6-mercaptopurine, and methotrexate) (Srimaroeng et al., 2008). The longer OAT3 in humans, a 568-amino-acid protein, does not support transport. It is likely that the two OAT3 variants are splice variants. Human OAT3 is confined to the basolateral membrane of the proximal tubule.

OAT3 clearly has overlapping specificities with OAT1, although kinetic parameters differ. For example, estrone sulfate is transported by both OAT1 and OAT3, but OAT3 has a much higher affinity in comparison with OAT1. The weak base cimetidine (an H₂ receptor antagonist) is transported with high affinity by OAT1, whereas the cation TEA is not transported.

OAT4 (SLC22A9). OAT4 (SLC22A9) was cloned from a human kidney cDNA library (Cha et al., 2000). OAT4 is expressed in human kidney and placenta; in human kidney, OAT4 is present on the luminal membrane of the proximal tubule. At first, OAT4 was thought to be involved in the second step of secretion of organic anions, i.e., transport across the apical membrane from cell to tubular lumen. However, studies demonstrate that organic anion transport by OAT4 can be stimulated by transgradients of α-ketoglutarate (Ekaratanawong et al., 2004), suggesting that OAT4 may be involved in the reabsorption of organic anions from tubular lumen into cell. The specificity of OAT4 includes the model compounds estrone sulfate and PAH, as well as zidovudine, tetracycline, and methotrexate (El-Sheikh et al., 2008; Srimaroeng et al., 2008). Interestingly, the affinity for PAH is low (>1 mM). Collectively, emerging studies suggest that OAT4 may be involved not in secretory flux of organic anions but in reabsorption instead.

Other Anion Transporters. URAT1 (SLC22A12), first cloned from human kidney, is a kidney-specific transporter confined to the apical membrane of the proximal tubule (Enomoto et al., 2002). Data suggest that URAT1 is primarily responsible for urate reabsorption, mediating electroneutral urate transport that can be trans-stimulated by Cl⁻ gradients. The mouse ortholog of URAT1 is involved in the renal secretory flux of organic anions including benzylpenicillin and urate.

NPT1 (SLC17A1), cloned originally as a phosphate transporter in humans, is expressed in abundance on the luminal membrane of the proximal tubule as well as in the brain (Werner et al., 1991). NPT1 transports PAH, probenecid, and penicillin G. It appears to be part of the system involved in organic anion efflux from tubule cell to lumen.

MRP2 (ABCC2), an ABC transporter, initially called the GS-X pump (El-Sheikh et al., 2008; Ishikawa et al., 1990; Toyoda et al., 2008), has been considered to be the primary transporter involved in efflux of many drug conjugates such as glutathione conjugates across the canalicular membrane of the hepatocyte. However, MRP2 is also found on the apical membrane of the proximal tubule, where it is thought to play a role in the efflux of organic anions into the tubular lumen. Its role in the kidney may be to secrete glutathione conjugates of drugs, but it also may support the translocation (with glutathione) of various non-conjugated substrates. In general, MRP2 transports larger, bulkier compounds than do most of the organic anion transporters in the SLC22 family.

MRP4 (ABCC4) is found on the apical membrane of the proximal tubule and transports a wide array of conjugated anions, including glucuronide and glutathione conjugates (El-Sheikh et al., 2008; Toyoda et al., 2008). However, unlike MRP2, MRP4 appears to interact with various drugs, including methotrexate, cyclic nucleotide analogs, and antiviral nucleoside analogs. Recent studies in Mrp4 knockout mice suggest that the transporter is involved in the renal elimination of the antiviral drugs adefovir and tenofovir (Imaoka et al., 2006). Other MRP efflux transporters also have been identified in human kidney, including MRP3 and MRP6, both on the basolateral membrane. Their roles in the kidney are not yet known.

Polymorphisms of OATs Polymorphisms in OAT1 and OAT3 have been identified in ethnically diverse human populations (Leabman et al., 2003; Srimaroeng et al., 2008). Two amino acid polymorphisms (allele frequencies >1%) in OAT1 have been identified in African-American populations. Three amino acid polymorphisms and seven rare amino acid variants in OAT3 have been identified in ethnically diverse U.S. populations (see www.pharmgkb.org).

In addition to P-glycoprotein, there is accumulating evidence for the roles of BCRP and MRP4 in limiting the brain penetration of drugs at the BBB (Belinsky et al., 2007; Breedveld, et al., 2005; Enokizono et al., 2007; Leggas et al., 2004; Ose et al., 2009). Furthermore, because of overlapped substrate specificities of P-glycoprotein and BCRP, they act as active barrier cooperatively in the BBB (Enokizono et al., 2008). Thus, dysfunction of both P-glycoprotein and BCRP exhibits synergy on the increase in the brain-to-plasma concentration ratio of the common substrates compared with that observed when either P-glycoprotein or BCRP is dysfunctional (Oostendorp et al., 2009; Polli et al., 2009).

The transporters involved in the efflux of organic anions from the CNS are being identified in the BBB and the blood-CSF barrier and include the members of organic anion transporting polypeptide (OATP1A4 and OATP1A5) and organic anion transporter (OAT3) families (Kikuchi et al., 2004; Mori et al., 2003). They mediate the uptake of organic compounds such as β-lactam antibiotics, statins, p-aminohippurate, H₂ receptor antagonists, and bile acids on the plasma membrane facing the brain-CSF in the net efflux across the endothelial and epithelial cells. The transporters mediating the efflux on the membranes that face the blood, have not been fully elucidated both in the BBB and blood-CSF barrier. MRP4 has been shown to mediate luminal efflux in the directional transport on an anionic drug, Ro64-0802 (an active form of oseltamivir), across the BBB in which the abluminal uptake is mediated by OAT3 (Ose et al., 2009). Members of the organic anion transporting polypeptide family also mediate uptake from the blood on the plasma membrane facing blood. Further clarification of influx and efflux transporters in the barriers will enable delivery of CNS drugs efficiently into the brain while avoiding undesirable CNS side effects and help to define the mechanisms of drug-drug interactions and interindividual differences in the therapeutic CNS effects.