Historical background

Pharmacokinetics comprises drug liberation, drug absorption, drug distribution, drug metabolism and drug excretion. The issue of xenobiotic elimination, also termed evasion, is defined in pharmacokinetics by two processes, drug metabolism and drug excretion, whereby metabolism is described to comprise sequential biotransformation steps termed phase 1 and phase 2 metabolism (Williams 1959; Gillette 1963; Gerok and Sickinger 1973). The “phase concept” has long been embedded in pharmacokinetics: R.T. Williams divided phenacetin detoxification into two parts, phase I which consists of an oxidative dealkylation reaction to yield 4-acetamidophenol, followed by phase II in which an organic acid is formed, yielding the highly polar 4-acetamidophenyl-β-glucuronide (Fig. 1). Emphasis was put on phase II as an essential detoxification phase because increase of polarity was considered the essential step for drug elimination via bile and urine.

Fig. 1
figure 1

Detoxification of phenacetin taken from Smith and Williams 1949. This description was a hallmark in drug metabolism and the first description of an O-dealkylation reaction of drugs

Full size image

Since the detoxification reactions occur within cells, it is surprising that the questions of how the drug reaches the endoplasmic reticulum of, for example, a liver parenchymal cell from the blood, and how the drug is released out of the hepatocyte into bile, have not been subject to equally intense studies as those concerning drug metabolism. In analogy to intestinal absorption, which in those days was thought to follow the physical diffusion concept of non-charged molecules (Brodie and Hogben 1957; Hogben et al. 1959; Brodie 1964), the permeation of organic compounds into and out of cells was considered to be governed by non-selective physical diffusion processes (Stein 1967; Diamond and Wright 1969; Elleroy and Lew 1977).

However, it was soon recognized that at least for highly polar organic compounds such as sugars, so-called facilitated passive diffusion, which involves a biological carrier with saturation and competition kinetics (Wilbrandt 1975) or even active, energy-dependent uphill transport by membrane carriers, is mandatory (LeFevre 1948; Crane and Krane 1959). An Na+-dependent intestinal glucose-carrier from a mammalian species was not cloned until 1987 (Hediger et al. 1987; Wright and Turk 2004). In general, biological non-electrolyte transport is a strong pillar of physiological cell functions. Its energetics comprises ion-coupled cotransport or ion-coupled countertransport, allowing secondary uphill transport, or pumping by ABC-carriers (ATP-binding cassette-carriers), allowing primary active transport, or countertransport with endogenous non-electrolytes or even simple facilitated diffusion along the compound´s concentration gradient (Fig. 2).

Fig. 2
figure 2

Classification of solute carriers (SLC), ATP-dependent carriers (ABC-carriers), and channels in the plasma membrane

Full size image

It has long been established that transporters exist for endogenous compounds such as glucose, amino acids, nucleosides, water soluble hormones and neurotransmitters. However, the perspective that xenobiotics are also substrates of membrane carriers has emerged only in the last two decades. Xenobiotics are by definition compounds which are not essential for the maintenance of a physiological function; they may, however, modulate, ameliorate or damage such functions, depending on whether they are drugs, diagnostics, or toxins. Living organisms, therefore, have also developed processes to eliminate “non-physiological” xenobiotics by carrier mediated transport (LeFevre 1975). This article focuses mainly on transporter-based elimination of drugs by the liver.

Functional indications of drug transporters

Early functional studies on the elimination of drugs by the liver (Klaassen and Watkins 1984), gut (Kramer and Lauterbach 1977; Jackson 1987), and kidney (Ullrich 1997) indicated the existence of carriers for xenobiotics, defined by transport saturation, cell specificity, and mutual transport interferences. Without any molecular knowledge of the carrier proteins themselves, descriptions were made of driving forces for uphill transport such as electrochemical gradients of ions (Heinz 1972; Semenza and Kinne 1985; Alvarado and Van Os 1986), counter- or antiport- transport (Rosenberg and Wilbrandt 1957; Heinz 1978), membrane potential (Ward 1970; Athayde and Ivory 1985) or direct energy transmission due to ATP-cleavage (Caldwell 1956, 1960; Keynes 1961; Carafoli and Scarpa 1982). Although it was suggested very early (Sperber 1959), it was later shown that drug excretion and absorption is strongly dominated by membrane carrier proteins, i.e. in liver (Petzinger et al. 1989; Petzinger 1994), kidney (Greger et al. 1981) and gut (Gilles-Baillieu and Gilles 1983). As with the detoxification enzymes of phase I and phase II, these carriers are subject to postnatal development in neonates (Gao et al. 2004), and regulation and induction by diseases may occur (Trauner and Boyer 2003; Kullak-Ublick et al. 2004). With the availability of cloned carrier genes, pharmacogenomics of drug transporters have recently been the subject of pharmacological and clinical research (Tirona et al. 2001; Tirona and Kim 2002; Ieiri et al. 2004).

Molecular classification of SLC- drug transporters

Drug transporter gene identification started in the 1990s by cloning Oatps (organic anion transporting polypeptides) from the liver (Jacquemin et al. 1994; Hagenbuch and Meier 2003) and Oats/Octs (organic anion transporters/organic cation transporters) from the kidney (Gründemann et al. 1994; Sekine et al. 1997; Sweet et al. 1997; Burckhardt and Burckhardt 2003; Koepsell and Endou 2004). These carriers have now been grouped in the large family of solute carriers SLC (Hediger et al. 2004). Table 1 shows a list of selected families of SLCs, which currently comprise 43 families with an expanding number of subfamilies and individual carriers. Over 300 human genes are assumed to encode SLC carriers. Many of these carriers can be considered to transport physiological endogenous substrates alone; several transport a few foreign compounds in addition, and some even transport xenobiotics predominantly. The SLC-families receive sequential numbers from the HUGO (human genome organisation) nomenclature committee HGNC. Overviews of several families of drug transporters have been published. Since this field is rapidly expanding, this article can only present a momentary snapshot, and several other reviews are recommended (Pritchard and Miller 1993; Burckhardt and Wolff 2000; Kullak-Ublick et al. 2000; Ayrton and Morgan 2001; Dean et al. 2001; Dresser et al. 2001; Borst and Oude Elferink 2002; Russel et al. 2002; Burckhardt and Burckhardt 2003; Daniel and Rubio-Aliaga 2003; Hagenbuch and Meier 2003; Mizuno et al. 2003; Daniel 2004; Koepsell and Endou 2004; Hediger et al. 2004; Pauli-Magnus and Meier 2004, and also reviews in this journal issue, i.e. König et al. 2006 and Geyer et al. 2006).

Table 1 List of solute carrier (SLC) families, according to the HUGO Gene Nomenclature Committee (HGNC), depicting the SLC-transporters performing drug transport
Full size table

An extended phase concept

Drug transporter / drug metabolism interplay represents a new challenge in cellular pharmaco- and toxicokinetics (Liu and Pang 2005). Reflecting the meaning of drug transport in cellular pharmacokinetics, the historical two-phase concept, which has only considered the relevance of biotransformation of drugs for drug evasion, needs extension.

Figure 3 shows an extended phase concept encompassing interactions between transporter phases for uptake and excretion, and the metabolic steps. The metabolic phases 1 and 2 are flanked by drug transporter phases 0 and 4, while intracellular cytoplasmic drug traffic is regarded as phase 3. In the following, some implications of this model are presented.

Fig. 3
figure 3

Schematic principle of vectorial drug evasion in liver and kidney. Phase 0 = drug uptake out of blood, Phases 1 and 2 = biotransformation exemplified by hydroxylation and glucuronidation, Phase 3 = transport of xenobiotics/metabolites towards excretion, Phase 4 = efflux into excreted fluids and/or backward into blood. ☼ xenobiotic

Full size image

Phase 0

SLC transporters mediate the earliest phase in drug kinetics on cells, which is here termed phase zero (phase 0). Phase 0 is the first step of drug elimination from blood via uptake across the basolateral membrane into the cells, i.e. of the proximal tubule or into hepatocytes respectively, or the first step of absorption from the gut, i.e. the uptake across the luminal membrane into enterocytes. Thus, SLC transporters are located in both the basolateral and the luminal cell membrane. This type of carrier imposes a selection, although not a really stringent “filter” gate, for certain classes of xenobiotics. SLC-drug carriers are multispecific, which means they allow permeation of a spectrum of compounds with variable chemical structures. An example of this type of multispecificity are members of the organic anion transporting polypeptide carriers Oatps (SLC21 / SLCO) which transport weak organic acids, neutral compounds and even a few cationic compounds (Hagenbuch and Meier 2003). Quantitative structure–activity relationships (QSARs) (Yarim et al. 2005) were elaborated to predict structural criteria of compounds essentially needed to fulfill the requirements for transport by individual carrier proteins. Systems in the liver and kidney have been reviewed in detail (Van Montfoort et al. 2003; Hagenbuch and Meier 2003; Chandra and Brouwer 2004). Similarly, another family of organic anion transporters, the Oat-family, transports negatively and positively charged compounds, and is very closely related to the Oct-family, specialised for organic cation transport (Burckhardt and Wolff 2000; Burckhardt and Burckhardt 2003; Koepsell and Endou 2004). Oats and octs, therefore, are members of the same SLC22 family (Table 1). Phase 0 drug transporters jointly influence compound allocation for drug metabolism.

Phase 4

Phase 4 pathways are well-defined. They comprise final steps of excretion, e.g. in the bile canaliculus of the liver, or secretory steps in the luminal membrane, e.g. in the gut, counteracting absorption. Phase 4 is predominantly maintained by directly driven “primary” uphill transport of xenobiotics across cell membranes, which is achieved by ATP-consuming transport ATPases belonging to the ABC-carrier proteins (Müller and Jansen 1998; Chan et al. 2004) (Table 2). A prominent member of these transport proteins, first detected in drug-resistant tumour cells (Juliano and Ling 1976), is P-glycoprotein. This drug resistance was conferred by a gene named multidrug resistance (MDR) gene, which was characterised in 1986 (Roninson et al. 1986). P-glycoprotein substrates are lipophilic, and are generally non-conjugated compounds, whereas water-soluble drug conjugates (sulfated, glucuronidated and glutathione-conjugated drugs) are transported in the liver by the multidrug resistance-associated protein MRP2 (Homolya et al. 2003; Hoffmann and Kroemer 2004; Fardel et al. 2005). Other members of the ABCC-transporter family (MRPs 1, 3, 4, 5, and 6), in conjunction with the BCRP (breast cancer resistance protein)-carrier belonging to the ABCG-family (Table 2) provide a panel of export pumps in the liver and elsewhere in the body (Haimeur et al. 2004) (Fig. 4). Carrier genes comprise at least 5% of all human genes with 49 genes belonging to human ABC-carrier genes (Dean et al. 2001).

Table 2 List of selected human ATP-binding cassette (ABC) transporters enabling drug transport
Full size table
Fig. 4
figure 4

Individual carriers in the hepatocyte which are involved in drug uptake and secretion. OAT/Oat = organic anion transporter belonging to the SLC22 family; OCT/Oct = organic cation transporter belonging to the SLC22 family; NTCP/Ntcp = Na + /taurocholate cotransporting polypeptide belonging to the SLC10 family; OATP/Oatp = organic anion transporting polypeptide belonging to the SLCO family (previously called SLC21). MDR1/Mdr1 = multidrug resistance protein (ABCB1), BCRP/Bcrp = breast cancer resistance protein (ABCG2), MRP2/Mrp2 = multidrug resistance-associated protein (ABCC2), BSEP/Bsep = bile salt export pump (ABCB11) all belonging to ABC-carrier proteins; OA− = organic anion; BS− = bile salt; OC+ = organic cation; DC− = dicarboxylate; GSH = glutathion

Full size image

ABC-carriers for xenobiotics, i.e. P-glycoprotein, MRP2 and BCRP, are mainly located at the luminal membrane of a cell facing, for example, the bile canaliculus, the lumen of the gut or the tubule lumen of a nephron. The general function which results from this location is drug excretion, providing evidence that these types of ABC-carriers convey protection against xenobiotics (Leslie et al. 2005). In the gut, these ABC-carriers are gatekeepers of drug absorption, limiting drug bioavailability (Dietrich et al. 2003).

Certain ABC-carriers, i.e. MRPs 1, 3, 4, and 6, are mainly found in the basolateral membrane, where they confer secretion of organic compounds into blood. In the liver MRP1 and MRP6 are strongly expressed, whereas MRP3 and MRP4 expression is low in normal liver. MRP1 transports non-metabolized cytotoxic drugs including certain cytostatics (daunomycin, doxorubicin, vincristine, but not cisplatin) known to be also P-glycoprotein transportates. In addition, ATP-dependent efflux of conjugated endogenous compounds, i.e. bilirubin glucuronides, cysteinyl-leukotriene C4, estradiol 17β-glucuronide and dianionic bile salts, is promoted, and efflux of xenobiotic conjugates via MRP1 may also occur (Leslie et al. 2001).

MRP3, normally present in the liver at low expression level, is highly expressed if hepatobiliary drug excretion is impaired due to extrahepatic cholestasis (Ogawa et al. 2000; Soroka et al. 2001; Scheffer et al. 2002). Under cholestasis (Donner and Keppler 2001) and in patients suffering from Dubin-Johnson syndrome (König et al. 1999), this carrier allows reflux of drug metabolites out of the hepatocyte into blood (small arrow of Fig. 3). At certain blood/tissue barriers such as the blood/brain and blood/testis barrier, reflux out of the cell into the blood occurs via P-glycoprotein (MDR1) and other carriers of the ABCC and ABCG family (Fromm 2004). P-glycoprotein could act as a “vacuum cleaner” (Bolhuis et al. 1997), sucking lipophilic drugs out of membrane phospholipids. Such drugs will not have penetrated through the membrane into the cell cytoplasm, and are thus not subjected to phase 1/phase 2 metabolism. This mechanism would impose a strict barrier at the membrane level between blood and tissues.

Perspective: intracellular transport of xenobiotics within a polarised cell

A mandatory step following phase 0 and also the biotransformation phases 1 or 2 is a cytoplasmic shuttle service of xenobiotics between differing cell poles. Pharmacokinetic phases 1 and 2 of drug metabolism preferably, but not exclusively, occur in the endoplasmic reticulum (ER), and are related to drug metabolism by membrane-bound cytochrome P450s (CYPs) and comparable enzymes (phase 1), followed by reactions of phase 2. Here, conjugating enzymes within or outside of the ER, such as glucuronosyltransferases (UGT-GTs), glutathione-S-transferases (GSTs), sulfotransferases (SULTs), cytoplasmic N-acetylases (NATs) and others, generate drug metabolites which are suited for rapid excretion (Fig. 3, main arrow). The metabolism phases afford delivery of drugs from the basolateral cell membrane to the ER and, later, of drug metabolites from the ER to the secretory sites of the cell membrane. This step is referred to as transport phase 3 in this review (Fig. 3). Cytoplasmic phase 3 transport represents a vectorial transport phase which has not yet been described in detail in any of the various pharmacokinetic cell transport models that have to date been published. It appears unlikely that cytoplasmic drug trafficking is a matter of simple physical diffusion, because diffusion would account for neither the velocity nor the efficiency of drug excretion by polarized cells. Thus, phase 3 drug transport needs floating carrier proteins or carrier particles. Previously, a protein named ligandin was considered as an important protein for cytoplasmic drug delivery in the liver (Levi et al. 1969). Later it was recognized that this protein was a glutathione-S-transferase (Habig et al. 1974). Other travelling proteins considered were cytosolic fatty acid-binding proteins (cFABPs), ileal lipid-binding proteins (ILBP) and bile acid-binding proteins (for a review see Petzinger 1994; Kramer et al. 2001). They comprise intracellular steps of the enterohepatic circulation of bile acids and drugs. Such binding proteins are assumed to travel together with transporter ligands along cytoskeletal tracks to their distinctive membrane sites (Bilej and Vetvicka 1989). Modern reflexions consider also endocytic trafficking pathways as important routes of drugs in the cell (Watson et al. 2005). For example, some weakly basic drugs such as doxorubicin may accumulate within acidic lysosomal organelles (Weaver et al. 1991), and thus may follow the traffic via vesicular pathways either in the lysosomal compartment (Lloyd 2000) or after fusion with endosomes in the multivesicular body (MVB) compartment. Vesicular cytoplasmic drug traffic may also require the participation of the cytoskeleton (Murray and Wolkoff 2003), and could explain cholestatic effects of toxins destroying the cytoskeleton (Ohashi et al. 2002). During cholestasis, the redistribution of some membrane carrier proteins would change the hepatobiliary secretion route for drugs which, instead of circulating into bile, reflux back to the blood (small arrow Fig. 3) (Rost et al. 1999). New approaches use fluorescent drugs or drugs coupled to fluorescent dyes for live cell imaging of intracellular transport pathways (Watson et al. 2005). This will better help to elucidate phase 3 drug transport in the future.

Clinical importance of implementing drug transporter phases

Drug transporters of the SLC-families preferably transport hydrophilic and amphiphilic xenobiotics. These carriers could confer some organ selectivity in xenobiotic elimination and toxicity. E.g. the Green Death Cap mushroom toxin phalloidin has long been recognised to be a liver-selective toxin. It was suggested by the author in the early 1980s that the bicyclic heptapeptide phalloidin should be transported by a liver-specific bile acid transporter (Petzinger et al. 1979; Petzinger 1981). In 2003 the liver-specific human OATP1B1, syn. OATP-C was shown to transport phalloidin (Fehrenbach et al. 2003). Only in the hepatocyte can the peptide accumulate to such an extent that the cytoskeletal protein actin is blocked. OATP carriers transport bile acids (Trauner and Boyer 2003) and a plethora of drugs (Hagenbuch and Meier 2003) including HMG-CoA reductase inhibitors such as pravastatin (Ziegler and Stünkel 1992; Ziegler and Hummelsiep 1993; Hsiang et al. 1999). Oatps were addressed by drugs conjugated with bile acids in a bile acid-based drug targeting approach for the liver (Kramer et al. 1992; Meijer 1993; Petzinger et al. 1995; Kramer and Wess 1996).

Drug–drug interferences already occur at the level of phase 0. A known example is inhibition of tubular secretion of β-lactam antibiotics penicillin G and cephalosporine by probenecid (Burckhardt and Burckhardt 2003). In the past this interaction was used clinically to diminish renal excretion of penicillins and to prolong their half-life times in serum. The timely development of β-lactam antibiotics with longer half-lives now makes the co-application of probenecid unnecessary. Dibromosulfophthalein is a substrate of the basolateral Oatp1a1 and of canalicular Mrp2 in rat liver (compare Fig. 4). It inhibited the hepatobiliary excretion of fexofenadine into rat bile by blockage of fexofenadine entry into the hepatocyte and, thereby, markedly reduced fexofenadine liver-tissue concentration (Milne et al. 2000). Decreased absorption of fexofenadine from gut after concomitant intake of grapefruit, orange, or apple juice has been reported (Dresser et al. 2002) and was partly related to transport competition. Further examples of clinically relevant drug interactions versus phase-0 carriers have been noted elsewhere (Ayrton and Morgan 2001).

Better known are drug–drug interactions resulting from transport competition on phase 4 ABC-carriers. For example, hepatobiliary liver excretion of the antihistaminic drug fexofenadine is reduced by 46% if co-administered with erythromycin, due to competition on P-glycoprotein (Milne et al. 2000). In contrast to dibromosulfophthalein (see above), erythromycin did not reduce the intrahepatic concentration of fexofenadine. The β-blocker talinolol is transported by P-glycoprotein (Grammatté et al. 1996). Therefore, the P-glycoprotein inhibitor verapamil inhibited the excretion of talinolol via the gut (Grammatté and Oertel 1999) and enhanced the oral exposure to the drug (Spahn-Langguth et al. 1998). Co-substrates of P-glycoprotein are the commonly used drugs ketoconazole/itraconazole, erythromycin, and quinidine. Previous clinical observations of a significant increase of plasma concentrations of digoxin, a cardiac glycoside, by simultaneous application of the antiarrhythmic drug quinidine (Doering 1979; Dahlquist et al. 1980; Pedersen et al. 1983) are now explained by transport competition on P-glycoprotein (Fromm et al. 1999; Drescher et al. 2003). This clinically important interaction was previously interpreted to be caused predominantly by phase 1 interactions (Dresser et al. 2001). The implementation of drug–carrier interactions (phase 4) into the picture of drug–CYP interactions (phase 1) (Ito et al. 1998) may also resolve certain drug–drug interactions among the modern anti-HIV-protease inhibitors ritonavir and saquinavir. Combinations of ritonavir with saquinavir boosted saquinavir blood levels several-fold (Van Heeswijk et al. 2001). Both compounds are P-glycoprotein substrates (Kim et al. 1998; Washington et al. 2000) but are also metabolized by CYP3A4 enzyme. A recent clinical study indicated that ritonavir exerted both types of interactions, namely inhibition of first-pass metabolism and increased bioavailability due to inhibition of intestinal P-glycoprotein secretion (Kilby et al. 2002).

Conversely, examples of co-stimulation of phase 2 enzyme and MRP-transporter expression under drug therapy are also known (Catania et al. 2004). For example, the herbal medicine St. John’s Wort is a strong inducer of several human drug-metabolizing enzymes (Delgoda and Westlake 2004) but also induces MRP2 transcription in rats and in human HepG2 cells respectively (Shibayama et al. 2004; Krusekopf and Roots 2005). Via the same PXR-mediated transactivation, human CYP3A4 and P-glycoprotein expression is enhanced, too (Synold et al. 2001). It is due to this mode of action that St. John’s Wort had a detrimental effect on therapy with immunosuppressants cyclosporin A and tacrolimus, with antineoplastic agents irinotecan and imatinib mesylate, and with hormonal contraceptives (Mannel 2004). Several drugs, i.e. oltipraz, garlic compound allyl sulphide, rifampicin, tamoxifen, and phenobarbital—already known inducers of phase 1 CYP enzymes—also induce or enhance MRP2- and P-glycoprotein-related drug transport (Fardel et al. 2005). For example, the well-known CYP inducers rifampicin/rifampin and phenobarbital are co-inducers of P-glycoprotein. Such changes complicate the prediction of drug–drug interactions enormously (Ito et al. 1998). Under rifampin, the plasma concentration of digoxin and talinolol decreased due to induction of intestinal digoxin secretion via P-glycoprotein (Greiner et al. 1999; Westphal et al. 2000). It is the extent of either inhibition or stimulation of metabolism, together with effects on transport, which alters the pharmacokinetic balance between the phases and which finally modulates drug–drug interactions in a very subtle manner (Zhang et al. 1998; Wandel et al. 1999).

Drug carrier polymorphisms and pathologies

Genetic carrier polymorphisms cause functional alterations of drug kinetics and their effects, and also of hereditary diseases. There are reports addressing pharmacokinetic consequences related to MDR1 polymorphism (Hoffmeyer et al. 2000; Drescher et al. 2002, 2003; Ieiri et al. 2004) and OATP-polymorphisms (Tirona et al. 2001; Nozawa et al. 2002; König et al. 2006). A frequent polymorphism of phase 4 MDR1 carrier occurring in about 12% of patients is the homozygous C3435T exchange. In these patients, (i) less than half of the normal MDR1 gene expression in the intestine, (ii) much lower gene induction by rifampin, and (iii) decreased excretion of digoxin into the gut was reported. Therefore, the homozygous T/T patients had digoxin plasma levels under rifampin stimulation four times higher than those of rifampin-stimulated homozygous C/C MDR1 patients (Hoffmeyer et al. 2000). On the other hand, plasma kinetics of P-glycoprotein substrate fexofenadine was not altered in patients bearing the C3435T or a G2677T mutation, although the activity of P-glycoprotein was markedly reduced (Drescher et al. 2002). Therefore, the known variability of the plasmakinetics of this drug in normal populations may have other reasons, and alterations in the phase 0 uptake transporters were considered (Drescher et al. 2002).

Other authors reported that certain SNPs and haplotypes of MDR1 affecting P-glycoprotein expression are associated with treatment outcome and/or host susceptibility to renal epithelial tumors (Siegsmund et al. 2002), Balkan endemic nephropathie (Atanasova et al. 2004), Parkinson’s disease (Drozdzik et al. 2003), breast cancer (Kafka et al. 2003), inflammatory large-bowel disease (IBD) (Brant et al. 2003) and ulcerative colitis (Schwab et al. 2003).

Many of the known polymorphisms in the human MDR1 gene are either functionally silent, or only marginally alter P-glycoprotein-mediated transport. None of them resulted in a complete loss of transport. This, however, was observed when a deletion mutation was found to occur frequently in the Collie dog MDR1 gene (Mealey et al. 2001; Roulet et al. 2003; Geyer et al. 2005a,b). These dogs suffer from neurotoxic symptoms of the antiparasitic drugs ivermectin and moxidectin that normally do not reach the central nervous system (Geyer et al. 2005a). In these dogs, lethal outcomes following drug application have even occurred (Pulliam et al. 1985; Paul et al. 1987).

Single nucleotide polymorphisms are also frequent in human phase zero OATP carriers. In the OATP-C gene (OATP1B1), which is specifically expressed only in the liver, the cholesterol-lowering effect of pravastatin decreased in patients exhibiting multiple SNPs, due to the decreased drug uptake into hepatocytes where the drug inhibits HMG-Co-A reductase (Kim 2004; Niemi et al. 2004, 2005a). An analogous kinetic change was reported for fexofenadine, which uses this carrier for hepatobiliary clearance. Patients with the T512C-SNP had higher plasma levels of this H1-receptor antihistamine when compared with patients lacking or exhibiting non-functional SNPs (Niemi et al. 2005b). SNPs have been also reported for other human OATP members, namely OATP1A2, 1B3, and 2B1 (Iida et al. 2001).

Pathologies of the liver related to single carrier defects and associated with severe inherited diseases are rare in humans. Of those that have been identified, all are caused by loss of function of phase 4 carriers (Jansen 2001; Kubitz et al. 2005). Examples are Dubin–Johnson syndrome caused by MRP2 (ABCC2) loss (Paulusma and Oude Elferink 1997) and subtypes of progressive familial intrahepatic cholestasis (PFIC), of which PFIC type 2 and PFIC type 3 are the most progressive cholestases caused by functional loss of BSEP (ABCB11) and of MDR3 (ABCB4; syn. Mdr2 in rodents), respectively (DeVree et al. 1998; Maisonnette et al. 2005; Wagner and Trauner 2005). Dubin–Johnson syndrome results in black pigmentation of the liver and high serum levels of glucuronidated bilirubins. Since conjugated bilirubin is not toxic, the MRP2 defect does not elevate serum cytotoxicity markers, and prognosis is good (Jansen 2001). In Dubin–Johnson syndrome, the defect of MRP2 is counteracted by an overexpression of MRP3, enabling the release of glucuronidated bilirubin into blood. PFIC2, also named Byler’s syndrome, is characterised by marginal biliary bile acid secretion (1% of normal), jaundice and progressive cholestasis requiring liver transplantation within the first decade (Wagner and Trauner 2005). PFIC3 is characterised by defective phospholipid, particularly phosphatidylcholine secretion into bile, and high serum γ-glutamyltransferases levels (DeVree et al. 1998). The prognosis of this disease is infaust.

On the other hand, and already mentioned, remarkable reconstruction of carrier expression resulting in synchronous up- and down-regulation of multiple transporters in the basolateral and canalicular membrane is observed in patients with intra- or extrahepatic cholestasis (Trauner et al. 1998, 2005; Shoda et al. 2001). These phenomena reflect adaptive transport modulation aiming to prevent excessive load of liver cells with toxic bile acids and cholephilic xenobiotics. Whilst the understanding of the molecular mechanisms of cholestasis is becoming more and more detailed on this level, this knowledge also offers an intriguing approach for a rational treatment of liver diseases (Wagner and Trauner 2005).

In conclusion

We would like to emphasize an integrated drug metabolism/drug transporter concept consisting of the “old” metabolism phases 1 and 2 but extended for new transport phases: the sequential processes of carrier-mediated drug uptake, intracellular drug transport, and carrier-mediated drug excretion. A carrier-mediated uptake phase 0 precedes the metabolism phases 1 and 2. A transcellular transport of a xenobiotic/metabolite through the cytosol of a polar excretory cell is termed phase 3, and the final carrier-mediated excretory process across the cell membrane is named phase 4. Each phase is prone to drug–drug interactions, both by blockade or induction. Prediction models for drug pharmacokinetics based hitherto on metabolism need to be expanded to include transport phases. Clinical and toxicological relevancies have been reported in the literature.