OATP8/1B3-mediated Cotransport of Bile Acids and Glutathione

In cholestasis, the accumulation of organic anions in hepatocytes is reduced by transporters (multidrug resistance-associated proteins and OSTα-OSTβ) able to extrude them across the basolateral membrane. Here we investigated whether organic anion-transporting polypeptides (OATPs) may contribute to this function. Xenopus laevis oocytes expressing human carboxylesterase-1 efficiently loaded cholic acid (CA) methyl ester, which was cleaved to CA and exported. Expression of OATP8/1B3 enhanced CA efflux, which was trans-activated by taurocholate but trans-inhibited by reduced (GSH) and oxidized (GSSG) glutathione. Moreover, taurocholate and estradiol 17β-d-glucuronide, but not bicarbonate and glutamate, cis-inhibited OATP8/1B3-mediated bile acid transport, whereas glutathione cis-stimulated this process, which involved the transport of glutathione itself with a stoichiometry of 2:1 (GSH/bile acid). No cis-activation by glutathione of OATP-C/1B1 was found. Using real time quantitative reverse transcription-PCR, the absolute abundance of OATP-A/1A2, OATP-C/1B1, and OATP8/1B3 mRNA in human liver biopsies was measured. In healthy liver, expression levels of OATP-C/1B1 were ∼5-fold those of OATP8/1B3 and >100-fold those of OATP-A/1A2. This situation was not substantially modified in several cholestatic liver diseases studied here. In conclusion, although both OATP-C/1B1 and OATP8/1B3 are highly expressed, and able to transport bile acids, their mechanisms of action are different. OATP-C/1B1 may be involved in uptake processes, whereas OATP8/1B3 may mediate the extrusion of organic anions by symporting with glutathione as a normal route of exporting metabolites produced by hepatocytes or preventing their intracellular accumulation when their vectorial traffic toward the bile is impaired.

The removal from blood of endogenous and xenobiotic cholephilic organic anions, such as bile acids, bilirubin, and steroid hormones, is a major function carried out by the liver. The vectorial transport of these compounds by hepatocytes includes uptake across the basolateral plasma membrane and subsequent secretion into bile across the canalicular plasma membrane. Because for most of these substances simple diffusion plays a minor role in both processes, the overall system depends on the polarized expression of transport proteins (1).
Three OATP isoforms with the ability to transport bile acids are expressed in human liver, namely OATP-A/1A2 (gene symbol SLCO1A2), OATP-C/1B1 (SLCO1B1), and OATP8/1B3 (SLCO1B3). The role of OATP-A/1A2 in bile acid uptake by hepatocytes is probably minor as compared with other OATPs because of its low expression in these cells (7,8). In contrast, based on its broad specificity, high expression level, and localization at the hepatocyte basolateral plasma membrane, OATP-C/1B1 probably plays a major role in the hepatic uptake of organic anions (9). OATP8/1B3 is also expressed at the same domain of the hepatocyte plasma membrane (10) and exhibits similar broad substrate specificity to that of OATP-C/1B1 (11), which suggests that this transporter could contribute to transporting organic anions in the liver as well as in other epithelial cells expressing it (8).
The driving forces accounting for OATP-mediated transport are not yet well known (12). Rat Oatp1/1a1-mediated transport * This work was supported in part by the Junta de Castilla y Leon Grants SA013/04 and SA059A05, Ministerio de Ciencia y Tecnologia, Plan Nacional de Investigacion Cientifica, Desarrollo e Innovacion Tecnologica Grant BFI2003-03208, Fundacion Investigacion Medica Mutua Madrileñ a Conv-III, 2006, Instituto de Salud Carlos III, FIS Grants CP03/00093 and PI051547. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1  has been reported to occur through pH-dependent and electroneutral anion exchange, in which the uptake of organic anions is coupled to the efflux of bicarbonate (13). However, the ability of reduced glutathione (GSH) to participate in exchange with organic anions across this antiporter has been also suggested (14). The efflux of intracellular GSH or glutathione S-conjugates also seems to be able to account for the driving force involved in rat Oatp2/1a4-mediated substrate uptake (15). It has been demonstrated that the organic substrate transport carried out by Oatp1/1a1 and Oatp2/1a4 can be potentially bidirectional (15,16), which implies that the overall directionality of the net transport is dependent on substrate gradients across the basolateral membrane of hepatocytes. Studies carried out with human hepatoblastoma HepG2 cells have indicated that the uptake and extrusion of cholephilic organic anions are modulated by intracellular GSH (17) but not by proton gradients (18). In contrast, OATP-B/2B1 is a pHsensitive transporter that does not transport GSH and is not activated by the outwardly directed GSH gradient (19). Whether other human members of the OATP family may work as anion exchangers or GSH-mediated transporters remains to be elucidated. In preliminary experiments, we unexpectedly found evidence that OATP8/1B3 could behave as a GSH-dependent symporter. Accordingly, we wondered whether this could play a role in cis-activating bile acid transport and, if so, what the functional meaning of this mechanism of action would be.
Under physiological circumstances, several ATP-binding cassette proteins export cholephilic organic anions into bile across the canalicular plasma membrane of hepatocytes, whereas other ATP-binding cassette proteins, such as the isoforms 1, 3, and 4 of the multidrug resistance-associated protein (MRP), are up-regulated in response to cholestasis and become a route for the extrusion of cholephilic organic anions across the basolateral membrane (for review see Ref. 20). This may limit the accumulation in hepatocytes of potentially toxic compounds. Therefore, another goal of this study was to investigate whether OATPs, in particular OATP8/1B3, may contribute to this function.  (21). [ 14 C]Cholic acid (CA) methyl ester (CA-ME) was synthesized by reaction of CA dissolved in methanol/ chloroform (1:2, v/v) with an excess of freshly distilled diazomethane in diethyl ether at room temperature for 12 h (22). Unlabeled sodium salts of bile acids, CA-ME, estradiol 17␤-D-glucuronide (E 2 17␤G), GSH and oxidized (GSSG) glutathione, L-glutamic acid, and sodium bicarbonate were purchased from Sigma. As indicated by the supplier, the purity of the bile acids, GSH, and GSSG was more than 98% as determined by thin layer chromatography. All other reagents were of analytical grade.

Uptake Studies in Xenopus laevis Oocytes
Harvesting and preparation of oocytes were carried out as described elsewhere (23) from mature female frogs (X. laevis), purchased from Regine Olig (Hamburg, Germany) and treated in accordance with the indications of current Spanish and European laws (RD 223/88 and European Union Directive 86/609/CEE) and supervised by the Ethical Committee for Laboratory Animals of the University of Salamanca.
Synthesis of cRNAs was performed as described previously (23) using recombinant plasmids containing the open reading frame cDNA of the following proteins: human OATP8/1B3 and OATP-C/1B1 cloned in pCMV6-XL4 and rat Bsep and human carboxylesterase-1 (CES1) cloned in pSPORT1. Incubation time after injection of the cRNA in oocytes was selected based on preliminary experiments on the time course of the functional expression for these carriers (results not shown).
Uptake studies were carried out using groups of at least eight oocytes per data point, and experiments were repeated using three different frogs. Oocytes were washed with substrate-free uptake (U) medium (100 mM choline chloride, 2 mM KCl, 1 mM CaCl 2 , 1 mM MgCl 2 , 10 mM Hepes, and pH adjusted to 7.5 with Tris) and incubated with 100 l of U medium containing the desired amount of the substrate to be tested at 25°C for the indicated time.
To investigate the cis-effects (activation or inhibition), the compounds were added to the U medium together with the substrate. Uptake was stopped by the addition of 4 ml of ice-cold U medium, and oocytes were washed three more times before being collected. They were then placed individually in vials to measure radioactive substrate by liquid scintillation.

Efflux Experiments from X. laevis Oocytes
Three different sets of experiments were carried out to investigate the glutathione-induced activation of OATP8/1B3-mediated efflux of bile acids from oocytes.
Trans-activation Experiments-Oocytes were injected with cRNA of either human CES1 and OATP8/1B3 or CES1 and rat Bsep, which were used as a positive control of bile acid efflux (22). After 2 days, the oocytes were incubated with 50 M [ 14 C]CA-ME at 25°C for 1 h, which, as has been reported previously (22), permitted the cells to be loaded with [ 14 C]CA-ME, which was subsequently hydrolyzed to free [ 14 C]CA by CES1. Based on previous studies (22), because the magnitude of [ 14 C]CA-ME hydrolysis by endogenous esterases is variable, cotransfection with CES1 was carried out to obtain higher and standardized cleavage activity. Bile acid efflux was determined after transferring the oocytes to fresh radioactivity-free medium with or without the activators or inhibitors to be tested and incubating at 25°C for 2 h.
Cis-activation of the Efflux from Oocytes Loaded by Incubation-The oocytes were incubated with 10 M [ 14 C]taurocholic acid in the absence of GSH at 25°C for 120 min or in the presence of 20 mM [ 3 H]GSH at 25°C for 60 min. The efflux of both compounds was determined at different time points after transferring the oocytes to radioactive substrate-free medium at 25°C.

Cis-activation of the Efflux from Oocytes Loaded by
Microinjection-The oocytes were directly loaded with 50 nl of U medium containing 300 M [ 14 C] taurocholic acid alone or with 20 mM [ 3 H]GSH. The efflux of both compounds was determined at different time points after immediately transferring the oocytes to radioactive substrate-free U medium at 25°C. Some oocytes were used to carry out the analysis of the intracellular balance between reduced and oxidized radioactive glutathione. Groups of three oocytes were treated with 100 l of lysis/extraction solution (1-butanol, 2-propanol, glacial acetic acid, and methanol, 2.5:2:0.5:1), sonicated in an ice-cold bath for 1 min, and centrifuged at 20,000 ϫ g at 4°C for 5 min. The supernatant was analyzed by TLC using a mixture of 1-butanol, 2-propanol, glacial acetic acid, methanol, and water (2.5:2:0.5: 1:4) as eluent. In this system, R f values were 0.45 for GSH and 0.28 for GSSG. In a different set of oocytes, the magnitude of the leak of injected compounds through the hole made in the plasma membrane by the microcapillary was evaluated by measuring the efflux of [ 3 H]inulin that was similarly loaded by microinjection.
In these sets of experiments the efflux process was stopped as in the uptake experiments, and the oocytes were processed as described elsewhere (22) for the determination of the amount of radiolabeled compound retained.

Measurement of Absolute mRNA Levels in Human Liver
Samples from human liver biopsies were obtained at the University Hospital of Salamanca (Spain) in accordance with established protocols, and consent forms were reviewed and approved by the Human Subjects Committee of this hospital. Forty two patients (age 49.6 Ϯ 14.7 S.D.; 50% female) were included in the study after being diagnosed at the Gastroenterology Division of the Salamanca University Hospital with one of the liver diseases indicated in Table 1 and showing biochemical signs of cholestasis. These were elevations above normal values of the serum concentrations of one or more of the following: total bilirubin, alkaline phosphatase, and ␥-glutamyl transpeptidase.
Samples were collected for diagnostic purposes, and only the remaining tissue was used for this study. This was immediately immersed in the RNA stabilization reagent RNAlater (Qiagen, Izasa) and stored at Ϫ80°C until total RNA was isolated and absolute abundance measured by real time quantitative RT-PCR using AmpliTaq Gold polymerase (Applied Biosystems, Madrid) in an ABI Prism 5700 Sequence Detection System (Applied Biosystems) as described previously (8). Detection of amplification products was carried out using SYBR Green I (Applied Biosystems). No nonspecific product of PCR, as detected by 2.5% agarose gel electrophoresis and DNA melting temperature curves, was found in any case. The results of mRNA abundance for the target genes in each sample were normalized on the basis of its 18 S rRNA content, which was measured using the TaqMan ribosomal RNA control reagents kit (Applied Biosystems). The primer oligonucleotides sequences and conditions for measuring the absolute abundances of OATP-A/1A2, OATP-C/1B1, and OATP8/ 1B3 have been reported elsewhere (8).

Statistical Methods
Results are expressed as means Ϯ S.D. For kinetic analyses, values were fitted to Michaelis-Menten or Hill equations. Linear regression analyses were carried out by the method of least squares. To calculate the statistical significance of the differences between groups, the paired t test or the Bonferroni method for multiple range testing was used as appropriate.

OATP8/1B3-mediated Uptake of Bile Acids in X. laevis
Oocytes-OATP8/1B3 expression markedly enhanced the ability of the oocytes to take up E 2 17␤G, a prototypic substrate of OATP8/1B3 (10), as compared with noninjected oocytes (Fig.  1A). Because in preliminary experiments no significant difference between oocytes injected with TE buffer and noninjected oocytes regarding their ability to take up bile acids or E 2 17␤G was found (data not shown), in this study, noninjected oocytes (wild) were used to determine nonspecific uptake. Similarly, the uptake of several bile acid species differing in their conjugation states and degree of hydroxylation was enhanced in oocytes injected with the cRNA of OATP8/1B3 (Fig. 1A). However, this effect was lower than for E 2 17␤G (Fig. 1B).
Trans-activation of OATP8/1B3-mediated Bile Acid Efflux-Because bile acid transport by several OATP isoforms, such as rat Oatp1/1a1 and Oatp2/1a4, has been reported to be potentially bidirectional and trans-stimulated by glutathione (14 -16), we investigated these characteristics for OATP8/1B3 using an experimental model described previously (22). To measure bile acid efflux, oocytes expressing CES1 were incubated with radiolabeled CA-ME. This permitted efficient loading of these cells with CA-ME. This was subsequently cleaved to free CA, which then slowly effluxed from the oocytes. This process was enhanced when rat Bsep was expressed in these cells (Fig. 2), which was used here as positive control of the experiment. Although to a lower extent, the expression of OATP8/1B3 also

levels of mRNA for OATPs in control liver tissue and that of patients with cholestatic liver diseases
Samples were collected from control healthy liver tissue (n ϭ 6) or 42 biopsies obtained from patients with hepatocellular cholestasis (n ϭ 12), primary biliary cirrhosis I-II (n ϭ 7), nonalcoholic steatohepatitis (n ϭ 7), nonviral hepatitis (n ϭ 8), and hemochromatosis (n ϭ 8). Expression levels were normalized on the basis of the content of 18 S rRNA measured in the same sample. Values are means Ϯ S.D. Determinations were carried out in triplicate on each sample by real time quantitative RT-PCR using total RNA obtained from human liver. No significant difference ( p Ͻ 0.05) as compared with control by the Bonferroni method of multiple range testing was found.

OATP-A/1A2
OATP-C/1B1 OATP8/1B3 1B1/1B3 ratio copies/10 6  OATP8/1B3-mediated Bile Acid/GSH Cotransport enhanced CA efflux (Fig. 2). On placing taurocholic acid in the extracellular medium, a trans-activation of the OATP8/1B3mediated efflux of CA was observed (Fig. 2); this is usually interpreted as an indication of the bidirectional transport ability of the carrier. To test whether OATP8/1B3 could work as an exchanger trans-activated by glutathione, inwardly directed glutathione gradients were imposed. Surprisingly, extracellular GSH and GSSG, which had no effect on the CA efflux from noninjected oocytes, induced a significant inhibition of OATP8/1B3-mediated CA efflux. Indeed, the reduction in CA content was even lower than that observed in oocytes expressing only human CES1 (Fig. 2). We speculated then on the possibility that CA molecules that had effluxed from the oocytes could in part be taken up again via OATP8/1B3 because of activation by the inwardly directed glutathione gradient, resulting in a net reduction of bile acid efflux. Cis-activation of OATP8/1B3-mediated Bile Acid Uptake-The next experiments were carried out to elucidate whether re-uptake of the bile acid was indeed involved in the diminished net CA efflux observed when glutathione was placed in the extracellular medium. Both GSH and GSSG, but not other anions such as bicarbonate and glutamate, were able to induce cis-stimulation of OATP8/1B3-mediated uptake of E 2 17␤G, taurocholic acid, glycocholic acid (GCA), and CA (Fig. 3). As expected, unlabeled E 2 17␤G and taurocholic acid inhibited the uptake of the radiolabeled substrates.
We then investigated whether OATP-C/1B1-mediated transport was also sensitive to cis-activation by glutathione. In contrast to what was found for OATP8/1B3, the addition of 20 mM GSH or GSSG to the incubation medium significantly reduced OATP-C/1B1-mediated taurocholic acid uptake (Fig.  4). Similar results were obtained when bicarbonate, taurocholic acid, or E 2 17␤G was added to the incubation medium (Fig. 4). These results are consistent with the expected behavior of OATP-C/1B1, similar to other members of the OATP family, as an organic anion exchanger (5).
The net intracellular accumulation of taurocholic acid reflects the balance between both uptake and efflux mechanisms. Accordingly, in the presence of glutathione or bicarbonate in the medium, the efflux of substrate taken up previously may decrease net taurocholic acid accumulation as compared with that occurring in the absence of inwardly directed glutathione or bicarbonate gradients. The reduction in radiolabeled taurocholic acid uptake by E 2 17␤G and taurocholic acid was consistent with competition between these substrates for OATP-C/1B1-mediated uptake (Fig. 4).
To perform kinetic analyses under appropriate initial velocity conditions, the time course of taurocholic acid uptake by oocytes expressing OATP8/1B3 was determined (Fig. 5A). The results indicated that 15 min was an appropriate time to obtain a marked signal of uptake still within the linear range of the uptake process. The addition of 20 mM GSH to the extracellular medium had little effect on nonspecific taurocholic acid uptake but markedly increased OATP8/1B3-mediated taurocholic acid uptake (Fig. 5B). Saturation curves were obtained by fitting the values of net taurocholic acid uptake against taurocholic acid concentrations. The best fits for OATP8/1B3-mediated taurocholic acid uptake were to Michaelis-Menten equations. Values are means Ϯ S.D. from at least 24 determinations per data point obtained using oocytes from at least three different frogs. OATP8/1B3-mediated uptake was calculated by subtracting the amount of radioactivity taken up by noninjected oocytes (wild) from that determined in oocytes injected with OATP8/1B3 cRNA 2 days before carrying out transport studies. *, p Ͻ 0.05, on comparing OATP8/1B3-expressing oocytes with noninjected oocytes by paired t test. †, p Ͻ 0.05, as compared with the E 2 17␤G uptake by the Bonferroni method of multiple range testing.

OATP8/1B3-mediated Bile Acid/GSH Cotransport
Kinetic parameters were calculated from double-reciprocal Lineweaver-Burk plots for taurocholic acid (Fig. 5C) and similarly for other bile acids (Table 2). These results revealed that the GSH-induced enhancement in the efficiency of OATP8/ 1B3-mediated bile acid transport (E T ), defined as the ratio between the maximal velocity of transport (V max ) and the apparent affinity constant (K m ), was mainly because of the increased V max value, whereas the K m value was not significantly modified by the presence of GSH.
OATP8/1B3-mediated Cotransport of Glutathione and Bile Acids-As kinetic analysis suggested that GSH activates the translocation of the substrate, whether this was because of an interaction of GSH with the protein with or without cotransport of the activator was investigated. Using radiolabeled GSH, it was found that nonspecific GSH uptake was much lower than that observed in oocytes expressing OATP8/1B3 (Figs. 6 and 7A). Moreover, when the values of net OATP8/1B3-mediated GSH uptake versus GSH concentrations were plotted, a saturation curve was obtained. The best fit was a Michaelis-Menten equation, whose kinetic parameters ( Fig. 7B and Table 2), revealed values for E T of the same range as those for bile acids, although both V max and K m values were markedly higher. When the ability of bile acids to affect OATP8/1B3-mediated GSH transport was investigated, we found that in the presence of extracellular concentrations of GCA much higher than those of GSH, the uptake of the latter was stimulated. Nevertheless, even when GCA concentrations in U medium were much lower than those of GSH, GSH uptake was also stimulated (Fig. 6). Moreover, the kinetic study of GSH uptake in the presence of GCA (Fig. 7B) revealed that this induced a significant increase in E T mainly because of enhanced V max , whereas the K m value was not significantly changed (Table 2).
To determine the stoichiometry of the process, the uptake of bile acids (taurocholic acid or GCA) at a fixed concentration (10 M) was measured in the presence of varying concentrations (1-20 mM) of GSH. The results were then fitted to Hill equations for different Hill numbers (n H ), i.e. the number of molecules of activator per molecule of substrate cotransported (Fig.  8). Both for taurocholic acid and GCA the best fit was for n H ϭ 2, suggesting that the most probable stoichiometry of OATP8/ 1B3-mediated cotransport of GSH and bile acids was 2:1, respectively.
Because these studies were carried out in uptake experiments, assuming bidirectional properties of the transporter, whereas in the in vivo situation the glutathione gradient in liver cells is expected to be directed outward, we investigated whether GSH was able to activate OATP8/1B3-mediated efflux. To carry out these experiments, oocytes were first loaded with radiolabeled taurocholic acid with or without GSH (Fig. 9). Because taurocholic acid uptake was activated by GSH, to obtain a similar initial intracellular load these cells were loaded for a shorter time (60 min), in the presence of GSH or for a longer period (120 min) in its absence. The efflux of taurocholic Values of net uptake are means Ϯ S.D. from at least 24 determinations per data point obtained using oocytes from at least three different frogs and were calculated by subtracting the amount of radioactivity measured in noninjected oocytes from that found in oocytes injected with OATP8/1B3 cRNA 2 days before carrying out transport studies. *, p Ͻ 0.05, as compared with control by the Bonferroni method of multiple-range testing. Ϫ ) or 50 M E 2 17␤G and taurocholic acid. Values of net uptake are means Ϯ S.D. from at least 24 determinations per data point obtained using oocytes from at least three different frogs and were calculated by subtracting the amount of radioactivity measured in noninjected oocytes from that found in oocytes injected with OATP-C/1B1 cRNA 2 days before carrying out transport studies. *, p Ͻ 0.05, as compared with control by the Bonferroni method of multiple range testing.
Because during the uptake/efflux period GSH could be converted to GSSG, an additional set of experiments was carried out by direct microinjection in the oocytes. [ 3 H]Inulin loaded by microinjection effluxed slowly from these cells (Fig. 10A), probably in part through the hole made by the microcapillary. GSH was more rapidly effluxed from oocytes expressing OATP8/1B3 (Fig. 10A). To determine whether part of the microinjected radioactivity remained as GSH throughout the experiment and hence could be used as a substrate by OATP8/ 1B3, TLC analysis of cell lysates was carried out at different time points. Even though a certain degree of oxidation of GSH to GSSG during the analytical procedure cannot be ruled out, this revealed that an important proportion of the radioactivity recovered from these cells was found to be GSH (min 0, 93 Ϯ 3%; min 10, 48 Ϯ 28%; min 30, 51 Ϯ 29%; and min 60, 54 Ϯ 23%). These experiments also confirmed that taurocholic acid efflux was faster from oocytes expressing OATP8/1B3 and was further stimulated by coinjection with GSH (Fig. 10B).
Expression of OATP-A/1A2, OATP-C/1B1, and OATP8/1B3 in Human Liver-To find some clue for understanding the physiological relevance of these findings, the steady-state expression levels in human liver of the three isoforms of OATPs with the ability to transport bile acids were measured by real time quantitative RT-PCR. As compared with OATP-C/1B1 and OATP8/1B3, the abundance of OATP-A/1A2 mRNA was very low both in normal liver and in several cholestatic liver diseases studied here (Table 1). In healthy livers, the expression of OATP-C/1B1 was higher than that of OATP8/1B3. No significant reduction in the expression of these isoforms was observed in any of the groups of cholestatic livers studied here. Moreover, a significant correlation between the expression of both OATP-C/1B1 and OATP8/1B3 was observed when all samples were plotted together (Fig. 11), indicating that the quantitative relationship between mRNA of both isoforms was ϳ5:1 (OATP-C/1B1 versus OATP8/1B3). No significant correlation between the abundance of mRNA of OATP-C/1B1 or OATP8/ 1B3 and any of the biochemical markers of cholestasis used here, namely serum levels of total bilirubin, alkaline phosphatase, and ␥-glutamyl transpeptidase, was found (data not shown).

DISCUSSION
By performing simultaneous determinations of OATP-A/1A2, OATP-C/1B1, and OATP8/1B3 expression, this study complements previous studies in which this has been measured in percutaneous liver biopsies from patients with cholestatic liver diseases (24 -26). Our results confirm previous reports (7,8) indicating that in healthy livers the expression of OATP-A/1A2 is too low to play a major role in the overall hepatic clearance of organic anions from blood. Moreover, this study suggests that no marked up-regula-

Substrate
Activator OATP8/1B3-mediated Bile Acid/GSH Cotransport OCTOBER 13, 2006 • VOLUME 281 • NUMBER 41 tion of this carrier occurs in cholestatic liver diseases that could change the relevance of OATP-A/1A2 in this function. However, an important role in the re-absorption of bile acids from bile toward the periductular capillary plexus cannot be ruled out (27). In contrast, the high expression of OATP-C/1B1 and OATP8/1B3 suggests that both may play a major role in the transport of organic compounds across the basolateral membrane of hepatocytes. In the cholestatic conditions studied here, the expression of OATP-C/1B1 and OATP8/1B3 was not significantly impaired. However this is probably not a general rule. Thus, in a previous study of four patients with primary scleros-ing cholangitis, the amount of OATP-C/1B1 mRNA was found to be reduced by half (24). The expression of both OATP-C/1B1 and OATP8/1B3 also has been found decreased in types 2 and 3 of progressive familial intrahepatic cholestasis (28). Moreover, in patients with primary biliary cirrhosis III, a significant decrease in the expression of OATP-C/1B1 has been reported (26). In this study, a tendency toward a decreased expression of OATP-C/1B1, although still not significant, in patients with primary biliary cirrhosis I and II was found.
The functional results of this study support, although they do not prove, the concept that OATP-C/1B1, like other members of the OATP family, behaves as an anion exchanger (5). Thus, several compounds induced a marked inhibition in overall OATP-C/1B1-mediated taurocholic acid uptake, which was consistent with competition with the substrate, as was probably the case of unlabeled taurocholic acid and E 2 17␤G, but also with activation of OATP-C/1B1-mediated efflux of taurocholic acid that had been taken up previously by the cells, as was probably the case of extracellular glutathione and bicarbonate.
Regarding the ability of OATP8/1B3 to transport bile acids, our results are apparently controversial with the initial studies by König et al. (10) who had found that OATP8/1B3-transfected HEK293 cells were not able to take up bile acids. However, an artificial mutation was inadvertently created during the construction of the recombinant plasmid used to transfect these cells, which decreased the ability of the expressed protein to transport some substrates, such as sulfobromophthalein, and abolished that of others, including bile acids (29). Moreover, the ability of OATP8/1B3 to transport taurocholic acid and GCA, when this protein was expressed in X. laevis oocytes, has been already described by others (11).
Thus although our results indicate that OATP8/1B3 was able to transport all major bile acids, either unconjugated or amidated with taurine or glycine, in the absence of glutathione, the magnitude of this process was similar to that carried out by OATP-C/1B1 only when OATP8/1B3 was activated by GSH or GSSG. In contrast, bicarbonate failed to activate OATP8/1B3mediated bile acid transport. Because glutathione enhanced V max without affecting K m , and GSH itself was transported by OATP8/1B3 in the same direction as bile acids, it could be suggested that this carrier may work as a symporter of organic anions and glutathione. The most probable ratio of the symport process is two molecules of GSH per each cholephilic organic anion cotransported. Moreover, our results suggest that this symporter is potentially bidirectional. This is a very interesting characteristic that has permitted kinetic studies by imposing precise extracellular concentrations of substrate and activator in uptake experiments, which is  OATP8/1B3-mediated Bile Acid/GSH Cotransport not possible in efflux studies using X. laevis oocytes because these cells have endogenous GSH export mechanisms, which are major confounding variables in GSH transport measurements (30). Thus, the results of the kinetic analyses are in agreement with those obtained for other GSH exporters recently reviewed by Ballatori et al. (12). In general, these carriers exhibit low catalytic efficiency, i.e. the K m values for GSH are relatively high and the transport velocities (V max values) are only moderate, leading to a low V max /K m ratio. The high K m values are not unexpected, given that GSH is present in high concentrations within cells (1-10 mM), whereas 100-fold lower K m values for bile acids are consistent with much lower concentrations of these compounds in hepatocytes. The values of K m for conjugated and unconjugated CA were similar to those reported previously for the transport of these bile acids by OATP-C/1B1 and rat Oatp1/ 1a1 and Oatp2/1a4, which lie in the range of 10 -35 and 35-54 M, respectively (11,31,32).
These findings have important functional implications. Because the glutathione gradient across the hepatocyte basolateral membrane is directed outwardly, it could be postulated that glutathione extrusion across the OATP8/1B3 may behave as a pathway for exporting organic anions. The balance between uptake and efflux across the basolateral membrane of the hepatocyte would be determined by the magnitude and direction of the gradient of a given substrate and the contribution to the overall process of uptake transporters, i.e. sodium-taurocholate cotransporting polypeptide and OATP-C/1B1 and export transporter, such as OATP8/1B3 whose expression levels are ϳ5-fold lower than those of OATP-C/1B1 both in healthy liver and in at least the cholestasis liver diseases studied here. Recently, two additional mechanisms have been suggested to contribute with OATP8/1B3 to carrier-mediated bile acid export across the basolateral membrane of the hepatocyte. One of them is MRP4, which, as we have described for OATP8/1B3, is also able to cotransport bile acids with reduced glutathione (33). The other is OST␣-OST␤, which is up-regulated in cholestasis (34).
Under physiological circumstances, the function of OATP8/ 1B3 may involve a reduction in the efficiency of the overall uptake process, but in compensation, it may also represent a rapid mechanism to export to the blood the derivatives produced by the biotransformation of cholephilic organic anions, which are subsequently taken up by other tissues or eliminated by the kidney. Moreover, because of the fact that the export into the extracellular space of GSH and its adducts is an important step in their turnover, a role of OATP8/1B3 in glutathione homeostasis, as has been proposed for rat Oatp1/1a1 (35), cannot be ruled out.
In addition, when the biliary excretory pathway is acutely impaired, the extrusion of organic anions across the basolateral membrane through OATP8/1B3 may constitute an alternative excretory pathway that is useful for protecting hepatocytes when the expression of OST␣-OST␤ and basolateral MRP isoforms is still very low and adaptive changes in transporter expression, aimed at reducing the toxic insult because of retained biliary constituents, have not yet taken place. More-over, because the expression of OATP8/1B3 is not markedly impaired in several cholestatic liver diseases, it could be suggested that, at least in these conditions, OATP8/1B3 would contribute to exporting cholephilic organic anions toward the blood.
In conclusion, although both OATP-C/1B1 and OATP8/1B3 are highly expressed, even in several cholestatic liver diseases, and are able to transport bile acids, their mechanisms of action are different. OATP-C/1B1 may be involved in uptake processes, whereas OATP8/1B3 may mediate the extrusion of organic anions by symporting with glutathione as a normal route of exporting metabolites produced by hepatocytes or preventing their intracellular accumulation when their vectorial traffic toward the bile is impaired.