Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Accurate evaluation of hepatic disposition (including hepatic metabolism, protein binding, intracellular sequestration, and biliary excretion) is necessary in the development of clinically useful drugs, as well as for predicting the pharmacological and toxicological effects of drugs, pharmacokinetic properties in humans, and drug-drug interactions. Biliary excretion of substrates is a complex process involving translocation across the sinusoidal membrane, movement through the cytoplasm, and transport across the canalicular membrane. Numerous in vitro systems (e.g., isolated perfused livers, isolated hepatocytes, short-term cultured hepatocyte couplets, liver plasma membrane vesicles, and expressed transport proteins) have been used to investigate biliary excretion processes (Oude Elferink et al., 1995).

Cultured hepatocytes represent a potential model to study the biliary excretion of a large number of substrates. Short-term (3-8 h) cultured hepatocyte couplets have been used to directly examine the biliary excretion of fluorescent compounds utilizing fluorescence microscopy (Graf et al., 1984; Graf and Boyer, 1990). However, the application of cultured hepatocyte couplets to study biliary excretion of xenobiotics is limited because the substrate must contain a fluorescent chromophore. Long-term (more than 24 h) cultured hepatocytes have been reported to restore polarity with canalicular-like structures and to develop an asymmetrical distribution of the sinusoidal and canalicular membrane proteins (Barth and Schwarz, 1982; Maurice et al., 1988; Talamini et al., 1997). Although primary hepatocytes maintained under conventional culture conditions have been used to study drug metabolism and hepatotoxicity, long-term cultures of hepatocytes have not been a suitable model for studying hepatobiliary transport due to the rapid loss of liver-specific functions, including hepatic transport properties, and failure to re-establish normal bile canalicular networks and maintain normal hepatocyte morphology (Groothuis and Meijer, 1996; LeCluyse et al., 1996a).

Modifications to conventional culture conditions have resulted in dramatic improvements in the maintenance of hepatic function and longevity of hepatocyte cultures (Maher, 1988). One successful approach was to mimic the native extracellular matrix geometry by maintaining hepatocytes between two layers of a collagen gel in a collagen-sandwich configuration (Dunn et al., 1989; LeCluyse et al., 1994). Maintenance of hepatocytes in a collagen-sandwich configuration prolongs hepatocyte viability and preserves liver-specific protein synthesis (Dunn et al., 1989). Further studies demonstrated that long-term cultured hepatocytes in a collagen-sandwich configuration re-establish a bile canalicular network and show better maintenance of drug uptake and enzyme induction potential (Sidhu et al., 1993; LeCluyse et al., 1996b). Recently, we have demonstrated that Na⁺/taurocholate cotransporting polypeptide was partially maintained in hepatocytes cultured in a collagen-sandwich configuration for 4 to 5 days (Liu et al., 1998). Furthermore, in these sandwich-cultured hepatocytes, the functional activity of the canalicular bile acid transporter and the canalicular multispecific organic anion transporter was demonstrated. In addition, a technique was developed to quantitate the amount of substrate in the bile canaliculi by determination of substrate accumulation in the presence and absence of Ca²⁺ in the incubation medium (Liu et al., 1999). Ca²⁺ depletion was used to increase the permeability of tight junctions in this model. This approach allows the quantitative examination of biliary excretion of nonfluorescent compounds with higher efficiency and greater versatility than other existing approaches. However, the effects of Ca²⁺ depletion on the transport properties and tight junctions of sandwich-cultured hepatocytes have not been examined extensively. The primary objective of this study was to investigate further the effects of Ca²⁺ modulation on this in vitro model. A multi-experimental approach was used to examine the effect of Ca²⁺ depletion on taurocholate accumulation and the permeability of tight junctions in the sandwich-cultured hepatocytes. A kinetic model was developed to describe substrate accumulation and to examine the transport processes in this in vitro model. Results from the present study further demonstrate that hepatocytes cultured in a collagen-sandwich configuration represent a useful in vitro system that can be utilized to study hepatobiliary disposition of xenobiotics.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. Taurocholate, dexamethasone, ruthenium red, Hanks' balanced salt solution, and Ca²⁺, Mg²⁺-free Hanks' balanced salt solution were purchased from Sigma Chemical Co. (St. Louis, MO). [³H]Taurocholate (3.4 Ci/mmol, purity >97%) and [³H]inulin (1.3 Ci/mmol, purity >97%) were obtained from DuPont-NEN (Boston, MA). Collagenase (type I, class I) was obtained from Worthington Biochemical Corp. (Freehold, NJ). Dulbecco's modified Eagle's medium (DMEM), fetal bovine serum, and insulin were purchased from Gibco (Grand Island, NY). Rat tail collagen (type I) was obtained from Collaborative Biomedical Research (Bedford, MA). All other chemicals and reagents were of analytical grade and were readily available from commercial sources.

Animals. Male Wistar rats (250-280 g) from Charles River Laboratories, Inc., (Raleigh, NC) were used as liver donors. Rats were housed individually in stainless steel cages in an alternating 12-h light/dark cycle at least 1 week before the study was performed, and were fed ad libitum until use. All procedures were approved by the Institutional Animal Care and Use Committee.

Hepatocyte Isolation. Hepatocytes were isolated under sterile conditions with a two-step perfusion method as reported previously (LeCluyse et al., 1996). Rats were anesthetized with ketamine and xylazine (60 and 12 mg/kg i.p., respectively) before portal vein cannulation. The liver was perfused in situ with oxygenated Ca²⁺-free Krebs-Henseleit bicarbonate buffer containing 5.5 mM glucose for 10 min at 37°C followed by perfusion with Krebs-Henseleit bicarbonate buffer containing collagenase type I (0.5 mg/ml) for 10 min. The hepatic capsule was removed with forceps. The hepatocytes were released by shaking the liver gently in 100 ml DMEM. The released cells were filtered through a sterile nylon mesh (70 µm) and centrifuged at 50g for 3 min. The cell pellet was resuspended in 25 ml DMEM and an equal volume of 90% isotonic Percoll (pH 7.4) and centrifuged at 150g for 5 min. The pellet was resuspended in 50 ml DMEM and the suspensions were combined into one tube followed by centrifugation at 50g for 3 min. Hepatocyte viability was determined by trypan blue exclusion. Only those hepatocyte preparations with viability greater than 90% were utilized for further studies.

Accumulation of Taurocholate in Isolated Hepatocytes. Taurocholate accumulation studies in freshly prepared hepatocyte suspensions were conducted with a modified method described by Studenberg and Brouwer (1993). Hepatocytes were suspended in ice-cold Hanks' balanced salt solution (standard buffer) to obtain a cellular protein concentration of approximately 2.0 mg/ml and stored in an ice bath. A 4-ml aliquot of the hepatocyte suspension was centrifuged at 50g for 2 min. The resulting pellet was suspended in 4 ml of standard buffer or Ca²⁺, Mg²⁺-free Hanks' balanced salt solution with 1 mM EGTA (Ca²⁺-free buffer) and incubated at 37°C with 95% O₂ and 5% CO₂ for 10 min. After addition of 0.1 ml [³H]taurocholate to the suspended hepatocytes, 0.1-ml aliquots were taken at designated times and added to 0.4 ml polyethylene microfuge tubes containing 0.05 ml silicone oil (diluted to a density of 1.03 with mineral oil) layered on top of 0.05 ml of 3 M KOH. The samples were centrifuged for 10 s in a table-top microfuge (Beckman Instruments, Inc., Fullerton, CA). The amount of [³H]taurocholate taken up into the hepatocytes was determined by cutting the tubes at the oil interface, placing the cell lysate in a scintillation vial with 5 ml of cocktail (Atomflow, Packard), and analyzing by liquid scintillation spectrometry. Adherent fluid volume on the surface of hepatocytes was determined with [³H]inulin (Bauer et al., 1975).

Preparation of Culture Dishes. Plastic culture dishes (60 mm) were precoated with rat tail collagen at least 1 day before preparing the hepatocyte cultures. To obtain a gelled collagen substratum, ice-cold neutralized collagen solution (0.1 ml, 1.5 mg/ml, pH 7.4) was spread onto each culture dish. Freshly coated dishes were placed at 37°C in a humidified incubator for approximately 1 h to allow the matrix material to gel, followed by addition of 3 ml DMEM to each dish and storage in a humidified incubator.

Cultured Rat Hepatocytes. Hepatocyte suspensions were prepared with DMEM containing 5% fetal calf serum, 1 µM dexamethasone, and 4 mg/l insulin. Hepatocyte suspensions were added to the precoated dishes at a density of 2 × 10⁶ cells/60-mm dish. Approximately 1 h after plating the cells, the medium was aspirated and 3 ml of fresh DMEM was added.

Electron Microscopy. Hepatocytes cultured on Permanox dishes (Nunc, Inc., Naperville, IL) in a sandwich configuration were incubated in standard buffer or Ca²⁺-free buffer for 10 min at 37°C and then fixed in a ruthenium red (0.25%)/glutaraldehyde (1.25%)/0.1 M sodium cacodylate (pH 7.3) buffer for 1 h at room temperature. After removal of the primary fixative solution, the cells were rinsed three times at room temperature in 0.1 M sodium cacodylate buffer. A solution of osmium tetroxide (1.3%)/ruthenium red (0.2%)/cacodylate (0.2 M) was applied and the cells were allowed to postfix for 1 h. Subsequently cells were rinsed three times with sodium cacodylate buffer, dehydrated, and embedded in Spurr resin. The embedded cultures were removed from the Permanox dishes and the area of interest was selected for re-embedding. Semi-thin sections were cut and stained with toluidine blue and examined before cutting ultra-thin sections. Ultra-thin sections were placed on copper grids and examined unstained with a Jeol 100C Transmission Electron Microscope (Jeol, Tokyo, Japan). Approximately 20 hepatocyte cultures from four individual preparations were examined.

Morphology and Accumulation Studies in Sandwich-Cultured Hepatocytes. Hepatocytes cultured in a collagen-sandwich configuration were incubated in 3 ml of standard buffer or Ca²⁺-free buffer at 37°C. Phase contrast micrographs of hepatocyte monolayers were obtained with an Olympus Light Microscope (Olympus, Tokyo, Japan). The cultured hepatocytes that were used for the morphology studies were from five individual preparations. Three or four separate sections were examined in each Petri dish. After removing the incubation buffer, uptake was initiated by addition of 3 ml of standard buffer containing [³H]taurocholate to each dish. After incubation for designated times, accumulation was terminated by aspirating the incubation solution and rinsing four times with 3 ml of ice-cold standard buffer to remove extracellular substrate (Liu et al., 1998). After washing, 2 ml of 1% Triton X-100 solution was added to culture dishes to lyse cells by shaking the dish on a shaker for 20 min at room temperature. An aliquot (1 ml) of lysate was analyzed by liquid scintillation spectrometry. Bio-Rad DC Protein Assay Kit (Bio-Rad Laboratories, Hercules, CA) was used to determine the protein concentration in the culture extracts using bovine serum albumin as standard. Triton X-100 (1%) did not interfere with the protein assay. All values for taurocholate accumulation into cell monolayers were corrected for nonspecific binding to the collagen by subtracting taurocholate accumulation determined in the appropriate control dishes in the absence of cells as described previously (Liu et al., 1998).

Model Development. The average accumulation versus time data for taurocholate (1-100 µM) were used in model development. Differential equations corresponding to a series of models shown schematically in Fig. 5 with combinations of first-order and Michaelis-Menten parameters provided in Table 1, were solved simultaneously with the nonlinear least-squares regression program WinNonlin (version 1.1, Scientific Consulting Inc., Apex, NC). Models incorporated two different compartment structures (Fig. 5). In models 1 to 5, cell and bile canaliculi were localized in the same compartment. In models 6 to 17, cell and bile canaliculi were localized in separate compartments. Each model was based on two assumptions: 1) preincubation in Ca²⁺-free buffer did not influence the membrane transport processes, and 2) the translocation processes were unidirectional. Model selection and assessment of goodness-of-fit were based on Akaike's Information Criterion (AIC; Akaike, 1976), the degree of colinearity of parameters, the S.E. of parameter estimates, the degree of bias in residual error, and visual inspection of the generated curves relative to the data. A weighting scheme of 1/Y was used for all fitting procedures.

Data Analysis. Accumulation data were normalized to the protein content and expressed as mean ± S.D. from three to four separate preparations of hepatocytes. Differences in substrate accumulation between experimental conditions were analyzed by ANOVA with the appropriate post hoc tests. A P value of < .05 was considered significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Accumulation of Taurocholate in Freshly Isolated Hepatocytes. The effects of Ca²⁺ on taurocholate accumulation were examined in freshly isolated hepatocytes incubated in standard or Ca²⁺-free buffer for 10 min before the addition of [³H]taurocholate. Taurocholate accumulation at 4°C was 2 to 7% of that observed at 37°C, as expected for an active transport process (Fig. 1). The initial uptake rates of taurocholate in standard buffer (6.53 ± 0.13 nmol/min/mg protein) and in Ca²⁺-free buffer (5.91 ± 0.88 nmol/min/mg protein) were not significantly different (p > .05). The accumulation of taurocholate at 10 min in freshly isolated hepatocytes was not significantly different in the presence and absence of extracellular Ca²⁺ (p > .05).

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Accumulation of [3H]taurocholate (25 µM) in standard buffer (closed symbols) or Ca2+-free buffer (open symbols) in freshly isolated hepatocytes at 37°C (circles) or 4°C (triangles).

Electron Microscopy and Ruthenium Red Staining. Ruthenium red staining in cultured hepatocytes was investigated after a 10-min incubation in standard buffer and Ca²⁺-free buffer. For sandwich-cultured hepatocytes incubated in standard buffer, ruthenium red staining was visible on the plasma membranes that were in direct contact with the collagen gel layer and along intercellular membranes, but was not present on the membranes lining the canalicular space (Fig. 2A). However, in hepatocyte monolayers incubated in Ca²⁺-free buffer, ruthenium red staining was present not only on basolateral and intercellular membranes but also on membranes lining the canalicular space (Fig. 2B). These observations directly demonstrated that Ca²⁺ depletion disrupted the barrier function of the tight junctions between the canalicular and extracellular spaces.

View larger version (184K): [in this window] [in a new window]   Fig. 2.   Electron micrograph of hepatocytes cultured in a sandwich configuration for 4 days. In standard buffer (A), ruthenium red staining was present along intercellular membranes but not on the canalicular membranes (arrow). Tight junctions prevent ruthenium red from reaching the canalicular membrane surface. After incubation in Ca2+-free buffer for 10 min (B), ruthenium red staining was present along intercellular and canalicular membranes (arrow). Bar = 1 µm.

Effects of Ca²⁺ on Canalicular Morphology and Taurocholate Accumulation. Hepatocytes cultured in a sandwich configuration for 4 days formed dilated canaliculi between adjacent hepatocytes (Fig. 3A). After incubation of the monolayers for 10 min in Ca²⁺-free buffer, the size of the canaliculi was reduced drastically (Fig. 3B). Subsequently, the monolayers were incubated in standard buffer for designated recovery times to determine if the contracted canaliculi could be redilated. Canalicular size did not change considerably after incubation of the monolayers for 10 min in standard buffer (Fig. 3C), but after 60 min the majority of canaliculi had redilated (Fig. 3D). However, the apparent size of most of the canaliculi was not as great as before incubation in Ca²⁺-free buffer (Fig. 3A).

View larger version (198K): [in this window] [in a new window]   Fig. 3.   Phase contrast micrographs of rat hepatocyte monolayers cultured in a sandwich configuration for 4 days (A). The monolayers were incubated in Ca2+-free buffer for 10 min (B), followed by incubation in standard buffer for 10 min (C) and 60 min (D). After incubation of the monolayers for 10 min in Ca2+-free buffer, the size of the canaliculi was reduced drastically (Fig. 3B, arrow). Canalicular size did not change considerably after incubation of the monolayers in standard buffer for 10 min (Fig. 3C, arrow), but after incubation in standard buffer for 60 min, the majority of canaliculi had redilated (Fig. 3D, arrow). Bar = 50 µm.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   [3H]Taurocholate accumulation in hepatocytes cultured in a sandwich configuration for 4 days. Effects of Ca2+ depletion (A). After preincubation in Ca2+-free buffer for designated times, taurocholate accumulation was determined by incubation of the hepatocytes in standard buffer containing 1 µM [3H]taurocholate for 10 min. *Accumulation at designated times after incubation in Ca2+-free buffer was statistically different (p < .05) from accumulation after incubation in Ca2+-free buffer for 10 min. Effects of Ca2+-repletion following Ca2+-depletion (B). Hepatocyte monolayers were preincubated in Ca2+-free buffer for 10 min before incubation in standard buffer for the designated recovery times. Accumulation was determined by incubation of the hepatocytes in standard buffer containing 1 µM [3H]taurocholate for 10 min. Control represents 10-min accumulation in hepatocytes not preincubated in Ca2+-free buffer. *Accumulation at designated recovery times after incubation in standard buffer was statistically different (p < .05) from time 0 (no incubation in standard buffer before initiation of 10-min accumulation studies).

Kinetic Analysis of Taurocholate Accumulation in Sandwich-Cultured Hepatocyte Monolayers. Accumulation of taurocholate (1-100 µM in standard buffer) was examined in hepatocyte monolayers cultured for 4 days in a sandwich configuration that had been preincubated in standard buffer or Ca²⁺-free buffer at 37°C for 10 min to characterize the kinetic processes involved in basolateral uptake and canalicular excretion. Seventeen different models were used to fit the data (Fig. 5, Table 1) to define an appropriate model to describe the kinetics of taurocholate accumulation. All models were of full rank, indicating that there were sufficient data to precisely estimate all the parameters. The condition number of the matrix of partial derivatives was less than 10⁶, suggesting a low degree of colinearity between parameters in the models. According to AIC and visual examination, model 13 provided the best description of taurocholate accumulation in the sandwich-cultured hepatocytes (Table 2, Fig. 6). The differential equations corresponding to model 13 are:

(1)

(2)

where X_standard is the cumulative amount of taurocholate taken up in standard buffer, X_Ca²⁺free is the cumulative amount of taurocholate taken up in Ca²⁺-free buffer, C is the taurocholate concentration in the incubation buffer, V_ma is the maximal velocity for uptake, K_ma is the apparent Michaelis-Menten constant for uptake, K_e4 is the first order rate constant for elimination from the bile compartment in standard buffer, V_mb is the maximal velocity for canalicular (biliary) excretion, and K_mb is the apparent Michaelis-Menten constant for canalicular (biliary) excretion. Kinetic parameter estimates for model 13 are presented in Table 2.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   Model scheme describing the accumulation of taurocholate in hepatocytes cultured in a sandwich configuration for 4 days. Parameter designations are provided in Table 1.

View larger version (25K): [in this window] [in a new window]   Fig. 6.   Accumulation of taurocholate in hepatocytes cultured in a sandwich configuration for 4 days. The hepatocyte monolayers were preincubated in standard buffer (A) or Ca2+-free buffer (B) for 10 min before the transport studies were initiated in standard buffer containing [3H]taurocholate (1-100 µM). Lines represent the computer-generated best fit of the compartmental model equations (model 13) to the data.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study, a variety of techniques were used to investigate the effects of Ca²⁺ depletion on the transport properties and tight junctions of hepatocytes cultured in a sandwich configuration. The results indicate that: 1) Ca²⁺ depletion does not alter taurocholate transport; 2) Ca²⁺ depletion increases the permeability of tight junctions, thus disrupting the barrier between the canalicular lumen and the extracellular space; 3) integrity of the disrupted tight junctions cannot be re-established completely by incubation in the presence of Ca²⁺ for 1 h; and 4) taurocholate accumulation involves Michaelis-Menten nonlinear processes for uptake and biliary excretion.

Hepatocytes cultured in a collagen-sandwich configuration for 6 days form complete junctional complexes composed of a tight junction, intermediate junction, and desmosomes (LeCluyse et al., 1994). Recently, Talamini et al. (1997) demonstrated the existence of junctional protein, uvomorulin (E-cadherin), in hepatocytes cultured in a sandwich configuration. Hepatocytes cultured in this configuration for 4 to 5 days consist of two compartments: the intracellular space and the canalicular lumen. The present studies provide direct evidence that Ca²⁺ depletion leads to loss of integrity of the tight junctions, resulting in enhanced permeability and thereby loss of a canalicular space distinct from the extracellular space.

Localization of ruthenium red staining was utilized to directly examine the barrier function of tight junctions in the monolayers. Ruthenium red does not penetrate intact plasma membranes, but binds to intercellular membranes, and will penetrate to the level of the tight junction in nonleaky epithelia (van Deurs et al., 1996; Mullin et al., 1997). Disruption of the tight junctions due to Ca²⁺ depletion allowed ruthenium red to access the interior of the canaliculi. Morphologic data obtained by light microscopy indicated that the canaliculi "collapsed" after the monolayers were exposed to Ca²⁺-free buffer (Fig. 3B). Previous confocal fluorescence microscopy studies in hepatocyte monolayers demonstrated that the fluorescence intensity of the canalicular networks in the monolayers after incubation with carboxydichlorofluorescein diacetate attenuated considerably after exposure to Ca²⁺-free buffer for 1 to 2 min, and the fluorescent canalicular networks disappeared completely in approximately 5 min (Liu et al., 1999). Consistent with the fluorescence studies, taurocholate accumulation decreased significantly after exposure of the hepatocyte monolayers to Ca²⁺-free buffer for 1 to 2 min. These observations further indicate that tight junction integrity is disrupted by Ca²⁺ depletion relatively quickly. Rapid disruption of the barrier function of tight junctions by exposure to Ca²⁺-free buffer with 1 mM EGTA has been described by Citi (1992) in cultured Madin-Darby canine kidney cells; the transepithelial electrical resistance was reduced within 5 min. To disrupt the integrity of the tight junctions, but to avoid the potential interfering effects of prolonged Ca²⁺ depletion on cellular function, 10-min incubations in Ca²⁺-free buffer were used in the present studies.

Although Ca²⁺ depletion did not interfere with accumulation of the model substrate taurocholate, it is possible that Ca²⁺ depletion may interfere with the accumulation of other substrates. Therefore, all accumulation studies were conducted in standard buffer to prevent potential interfering effects of Ca²⁺ depletion on substrate transport in the hepatocyte monolayers. This approach was based on the assumption that the functional integrity of tight junctions could not be re-established during the short duration of transport studies. To test this hypothesis, canalicular morphology and taurocholate accumulation were examined at designated recovery times during incubation in standard buffer after monolayers were incubated for 10 min in Ca²⁺-free buffer. The morphology and transport studies suggested that the integrity of the disrupted tight junctions recovered slowly during incubation in standard buffer. Based on these results, quantitation of biliary excretion in the sandwich-cultured hepatocytes should not be influenced by the re-establishment of tight junction integrity if substrate accumulation studies are completed within 10 min.

To further examine the effects of Ca²⁺ on the transport properties of sandwich-cultured hepatocytes, and to examine the utility of this in vitro model to study hepatobiliary disposition, kinetic modeling was utilized to analyze taurocholate accumulation in the monolayers preincubated in standard or Ca²⁺-free buffer. Taurocholate disposition in sandwich-cultured hepatocytes involves multiple kinetic processes, including uptake across the basolateral membrane and excretion across the canalicular membrane. More than one kinetic process may be responsible for taurocholate translocation across each membrane domain. Parameter estimates obtained from fitting kinetic models to taurocholate accumulation-time data in sandwich-cultured hepatocyte monolayers may reveal information obscured by conventional mass-balance analysis (Studenberg and Brouwer, 1993; Booth et al., 1996). All models were based on the assumption that activity of the membrane transporters was the same in hepatocyte monolayers preincubated in standard or Ca²⁺-free buffer, and each kinetic process was unidirectional. Freshly isolated hepatocytes that lose hepatic architecture and intact canalicular tight junctions (Graf and Boyer, 1990) represent an ideal model to investigate this assumption. The present study demonstrated that the initial uptake rate as well as the 10-min accumulation of taurocholate were independent of extracellular Ca²⁺ concentrations in freshly isolated hepatocytes, suggesting that Ca²⁺ modulation did not alter taurocholate transport processes. These findings were consistent with previous observations that hepatic uptake and secretion of taurocholate in isolated hepatocytes are not dependent on extracellular Ca²⁺ concentrations (Anwer and Clayton, 1985).

Several interesting issues are apparent after examination of the model structure and parameter estimates. The fact that a two-compartment model (cell compartment and bile compartment) described the accumulation data better than a one-compartment model is consistent with observations from confocal fluorescence microscopy studies (Liu et al., 1999) and electron microscopy studies discussed above. Taurocholate uptake was described best by a Michaelis-Menten kinetic process (K_m = 28.0 ± 3.6 µM). This value was close to the range of K_m values (30-50 µM; Boyer and Meier, 1990) for taurocholate uptake in rat hepatic sinusoidal membrane vesicles. V_max values for taurocholate uptake in the present study were greater than the values determined in a previous study (Liu et al., 1998). Taurocholate uptake in hepatocytes is mediated predominantly by Na⁺/taurocholate cotransporting polypeptide and to a lesser extent by a Na⁺/independent organic anion transporter (Zimmerli et al., 1989; Oude Elferink et al., 1995). In the present study, addition of a parallel first order uptake process to the Michaelis-Menten equation slightly improved the fit based on the sum of square residuals, however, the improved fit was not statistically significant. In previous studies (Liu et al., 1998), kinetic analysis of the initial rate of taurocholate uptake in hepatocytes cultured in a sandwich configuration was described best by a Michaelis-Menten process in parallel with a first order process. These apparent differences may be because the concentration range of taurocholate used in the present study (1-100 µM) was lower than in previous work (1-200 µM).

The elimination of taurocholate from the sandwich-cultured hepatocyte cell compartment in the presence of Ca²⁺-free buffer represents the biliary excretion process in the monolayers. A Michaelis-Menten kinetic process best described the biliary excretion data, suggesting that a carrier-mediated elimination process was involved in canalicular excretion, as demonstrated previously (Muller et al., 1991; Stieger et al., 1992). The estimated maximal velocity for taurocholate biliary excretion was 1.82 ± 0.36 nmol/min per mg of cellular protein. Considering protein content for liver tissue is 0.20 mg/mg liver (Seglen, 1976), the estimated maximal taurocholate secretion by normal rat liver would be approximately 364 nmol/min/g liver. This value is consistent with the maximal excretion rates reported for bile salts (170-350 nmol/min/g liver; Stieger et al., 1992; Klos et al., 1979; Yousef et al., 1987).

The modeling analysis in this study suggested that first order elimination of taurocholate occurred from the bile compartment but not the cell compartment when the sandwich-cultured hepatocytes were incubated in standard buffer. As expected, first order elimination directly from the cell compartment should be negligible because simple diffusion of taurocholate across the canalicular membrane is negligible (Liu et al., 1999). First order elimination from the bile compartment in standard buffer presumably represents "leakage" from the canaliculi. The canalicular lumen undergoes cycles of contraction and dilation, which cause the expulsion of bile contents in vivo and in cultured hepatocytes (Phillips et al., 1982; Watanabe et al., 1991). Alternatively, bile motility may be due to the noncontractile collapse of canaliculi in response to secretory pressure, resulting in rupture of the canaliculi (Boyer, 1987; Graf and Boyer, 1990). The translocation of bile from bile canaliculi into the medium via an apparent first order elimination process from the bile compartment is consistent with proposed mechanisms of bile flow.

In summary, results from this study directly demonstrate that tight junctions are the diffusional barrier between the bile canalicular lumen and the extracellular space in sandwich-cultured hepatocytes. This barrier can be disrupted rapidly by depletion of extracellular Ca²⁺ without altering taurocholate transport. Kinetic modeling analysis indicates that taurocholate uptake and biliary excretion occur via carrier-mediated transport processes. Hepatocytes cultured in a sandwich configuration represent a useful in vitro model system that may be utilized to study hepatobiliary disposition of compounds.