Abstract
Rhodamine 123 (Rho123), a model substrate of P-glycoprotein (P-gp), was used to evaluate the functional activity of P-gp efflux transporter in the rat placental barrier. The dually perfused rat-term placenta method was used. In our experiments, the materno-fetal transplacental passage of Rho123 did not meet the criteria of the first-order pharmacokinetics, suggesting an involvement of transporter-mediated process. Inhibitors of P-gp, such as [3′-keto-Bmt1]-[Val2]-cyclosporine (PSC833), cyclosporine (CsA), quinidine, and chlorpromazine, increased significantly the materno-fetal transplacental passage of Rho123 in the experiments under steady-state conditions. On the other hand, PSC833, CsA, and quinidine decreased the feto-maternal passage of Rho123. Similarly, in the experiments carried out under nonsteady-state conditions, CsA accelerated the passage of Rho123 in the materno-fetal direction and decreased its passage in the opposite direction. Feto-maternal transplacental clearances of Rho123 were found to be considerably higher than those in the materno-fetal course. Potent P-gp inhibitors, such as PSC833 or CsA, partially canceled the asymmetry. Negligible metabolism of Rho123 into its major demethylated metabolite rhodamine 110 was observed in the rat placenta. Expression of P-gp genes was detected using immunohistochemical, Western blotting, and reverse transcription-polymerase chain reaction methods preferentially in the second rat syncytiotrophoblast layer. In conclusion, these data suggest that P-gp limits the entry of Rho123 into fetuses and at the same time it accelerates the feto-maternal elimination of the model compound. Therefore, it seems plausible that pharmacokinetics of xenobiotics in the rat placental barrier could be controlled by P-gp in both directions.
The function of P-glycoprotein (P-gp) has been proposed to be an ATP-dependent membrane efflux pump, whose primary mode of action is to remove numerous lipophilic compounds from the lipid bilayer and to pump them actively out of cells (Stein, 1997; Ambudkar et al., 1999). In this way, P-gp confers a multidrug resistance phenomenon (MDR) to tumor cells against a large spectrum of anticancer drugs. Substrates transported by P-gp include a variety of structurally and pharmacologically unrelated, hydrophobic compounds, such as various anticancer drugs, cardiac glycosides, β-blockers, calcium channel blockers, etc. (Ambudkar et al., 1999). P-gp has also been detected on the membranes of a wide range of normal tissues, such as the capillaries forming the blood-brain and blood-testis barriers, apical surface of the gastrointestinal tract, luminal membrane of the renal proximal tubular cells, and hepatocytes. At these sites, P-gp was found to serve as a barrier for the entry of lipophilic xenobiotics into the body and/or to the tissues that are sensitive to their toxic injury.
However, function of P-gp in the placental barrier has been less examined and is still not fully understood. High levels of P-gp have been detected in human and murine syncytiotrophoblast layers, which are the crucial parts of the hemochorial placental barrier (Cordon-Cardo et al., 1990; Sugawara, 1990; Bremer et al., 1992; Trezise et al., 1992; MacFarland et al., 1994; Nakamura et al., 1997; Lankas et al., 1998; Myloma et al., 1999; Smit et al., 1999; Ushigome et al., 2000; Tanabe et al., 2001). P-gp has been demonstrated to be integrated in the microvillous membrane of the human syncytiotrophoblast that faces directly maternal blood (MacFarland et al., 1994; Ushigome et al., 2000). Transport activity of P-gp in the placental barrier has been examined both in vivo and in vitro (Audus, 1999). Lankas and coworkers revealed that fetuses of CF-1 mice lacking the mdr1a gene isoform of P-gp were susceptible to cleft palate malformation induced by avermectin B1a. Conversely, the fetuses of wild-type mice were completely protected against the above-mentioned teratogen (Lankas et al., 1998). Similarly, administration of other P-gp substrates (digoxin, saquinavir, or paclitaxel) to mdr1a-/-/1b-/- knockout mice revealed 2.4-, 7-, and 16-fold higher transplacental transport of these drugs into the fetus compared with wild-type mice (Smit et al., 1999).
The in vitro action of the placental P-gp has been demonstrated in uptake studies using the BeWo cell line (a human choriocarcinoma trophoblastic cell line), primary cultures of the human cytotrophoblasts (Utoguchi et al., 2000), and human trophoblast membrane vesicles (Nakamura et al., 1997). Ushigome et al. (2000) studied the function of P-gp in the placenta using the BeWo cell line cultured in a confluent epithelial monolayer. The study suggests that due to the one-way functional activity of P-gp located in the apical membrane, transport of selected P-gp substrates is higher in the basolateral-to-apical and restricted in the apical-to-basolateral direction. On the basis of all these findings, it is believed that P-gp expressed in trophoblast cells of the placenta limits the entry of its substrates into the fetus by reverse pumping of the compounds from the trophoblast layers back into the maternal bloodstream. Moreover, as the in vitro study using BeWo cells suggests, P-gp could accelerate transplacental passage of P-gp substrates in the feto-maternal direction. Our previous in situ experiments demonstrated that the passage of CsA across the rat placenta is restricted in the materno-fetal direction due to the P-gp activity (Pavek et al., 2001).
The aim of the present article was to examine 1) the functional activity of P-gp in both the materno-fetal and fetomaternal directions using the dually perfused rat placenta method, and 2) to confirm at the level of the intact rat placental barrier the asymmetry of transplacental passage of P-gp substrates described in the in vitro epithelial model of BeWo cells. A fluorescent dye rhodamine 123 (Rho123), which is a well established model substrate for testing the functional transport activity of P-glycoprotein, was used (Masereeuw et al., 1997; van der Sandt et al., 2000). CsA, PSC 833 (a nonimmunosuppressive derivate of CsA), and chlorpromazine were used as selective inhibitors/modulators of P-gp, quinidine was used as a nonselective inhibitor for both P-gp and organic cation transporters (rOCT), and sodium azide was used as an inhibitor of ATP mitochondrial synthesis. Because there was only limited information on P-gp expression in the rat placenta in the current literature, we carried out immunochemical, immunohistochemical, and RT-PCR studies to confirm expression and localization of P-gp gene products in the rat-term placentas.
Materials and Methods
Material and Reagents. Rhodamine 123 and rhodamine 110 (demethylated Rho123) were purchased from Sigma-Aldrich (St. Louis, MO; catalog no. R-8004, batch 127H3707) and Fluka (Buchs, Switzerland; catalog no. 83695, EC.2369447), respectively. Substances of CsA and SDZ PSC833 (PSC 833, [3′-keto-Bmt1]-[Val2]cyclosporine) were obtained thanks to the kindness of Dr. Andrýsek (IVAX Ltd., Opava, Czech Republic). Quinidine sulfate (QND) and diamond fuchsin were purchased from Lachema Ltd. (Brno, Czech Republic). Chlorpromazine chloride (CPZ), bovine serum albumin, RPMI 1640 medium, and other substances were purchased from Sigma-Aldrich. Tris, glycine, leupeptin, pepstatin, and phenylmethylsulfonyl fluoride were obtained from Serva GmbH (Heidelberg, Germany). 4-Nitrophenyl phosphate disodium salt hexahydrate was purchased from Fluka. Acetonitrile and methanol (both HPLC grade) were purchase from Merck (Darmstadt, Germany).
Stock solutions of CsA and PSC833 were used in a concentration of 1 mg/ml containing 19% of ethanol and 1% of cremorphor (v/v). The final concentration of cremorphor in perfusion media was less that 0.01%. Aqueous stock solutions of QND, CPZ, and sodium azide were used in a concentration of 10, 25, and 100 mg/ml, respectively.
Animals. All experiments were approved by the Ethical Committee of the Faculty of Pharmacy (Hradec Králové, Charles University in Prague) and were carried our in accordance with the Guide for the Care and Use of Laboratory Animals, 1996; and the European Convention for the protection of vertebrate animals used for experimental and other scientific purposes, Strasbourg, 1986.
Pregnant Wistar rats were purchased from Biotest Ltd. (Konárovice, Czech Republic) and were bred in 12/12-h day/night standard conditions with water and pellets ad libitum. Experiments were performed on day 22 of gestation. Fasted rats were anesthetized with pentobarbital (Nembutal; Abbott Laboratories, North Chicago, IL) in a dose of 40 mg/kg administered intravenously into the tail vein.
Perfusion Method. The method of dually perfused rat placenta was used as described previously (Mohammed et al., 1993; Pavek et al., 2001). Briefly, the placenta was excised and allowed to dive in the heated Ringer saline. A catheter was inserted into the uterine artery proximal to the blood vessel supplying the selected placenta and connected with the peristaltic pump bringing Krebs' perfusion liquid containing 1% bovine serum albumin from the maternal reservoir. The uterine vein, including the anastomoses to other fetuses, was ligated behind the perfused placenta and carefully cut so that maternal solution could leave the perfused placenta. The chosen fetus was separated from the neighboring ones by ligatures. The umbilical artery was catheterized using the 24-gauge catheter and connected with the tubing by which the fetal perfusion liquid from the fetal reservoir was supplied. The umbilical vein was catheterized in a similar manner and the selected fetus was removed. After a successful umbilical catheterization, the fetal vein effluent was collected into preweighted glass vials to check a possible leakage of perfusion solutions from the placenta. In the case of leakage, the experiment was discarded. Perfusion experiments did not last longer than 36 min, because the integrity of the placental barrier could be affected in later intervals. This notion is based on our previous experiments with l-[3H]glucose, a marker of paracellular diffusion, where the transplacental passage of l-[3H]glucose started to significantly increase in later intervals of the perfusion (Pavek et al., 2001). Wet weights of the placentas used in experiments were 0.54 ± 0.13 g.
To study the influence of inhibitors on both the materno-fetal and feto-maternal transport of Rho123, experiments were carried out under both steady-state and nonsteady-state conditions as described below.
Evaluation of Rho123 Metabolism in the Rat Placenta. In our pilot experiments, we studied possible placental metabolism of Rho123 during its transplacental passage. In the experiments carried out under both steady-state and nonsteady-state conditions (for experimental designs of steady-state and nonsteady-state experiments see below), Rho123 was present in the maternal perfusion solution in concentrations of 0.65 or 1.3 μM. Fetal effluent samples were analyzed using HPLC.
Effect of CsA on Transplacental Passage of Rho123 under Nonsteady-State Conditions. After successful establishment of the dual perfusion with the Rho123-free and inhibitor-free perfusion solution, the maternal and/or fetal perfusion inflows were switched to another prewarmed perfusion reservoirs containing examined compounds. The experiments started after 15 s of delay (time 0) to make it possible to fill tubing with the new liquids. The fetal umbilical vein outflow was collected in 5-min intervals into preweighted glass vials for 25 min of the experiment.
To examine possible influence of CsA on the transplacental passage of Rho123 in the materno-fetal direction, CsA was added in a concentration of 40 μM to the maternal reservoir together with Rho123 (1.3 μM). Samples were collected in 5-min intervals from the fetal umbilical vein outflow.
The feto-maternal transplacental passage of Rho123 was followed in experiments where Rho123 was added in a concentration of 1.3 μM into the fetal solution (control experiments). The effect of CsA on the Rho123 feto-maternal passage was examined by simultaneous addition of CsA (40 μM) into the maternal solution and Rho123 (1.3 μM) into the fetal solution at time 0 of the experiment. Samples were collected in 5-min intervals from the fetal umbilical vein outflow.
Effect of CsA, PSC833, QND, CPZ, and Sodium Azide on Transplacental Passage of Rho123 under Steady-State Conditions. In experiments examining the materno-fetal passage of Rho123 under the steady-state conditions, Rho123 (0.65 μM) was brought to the perfused placenta via the catheterized uterine artery immediately after the catheterization. Sample collection started (time 0) after 15-min delay. Within this delay, steady-state conditions for transplacental passage of Rho123 were achieved as suggested by a nearly constant time course of transplacental clearances during the experiment (Fig. 3). Samples were collected in 3-min intervals from the fetal umbilical vein for 36 min of the experiment. In the 12th min of the experiment, examined P-gp inhibitors were added to the maternal reservoir to reach the concentrations of 10 μM (CsA and PSC833), 40 μM (QND and CPZ), and 5 mM (sodium azide), respectively. This experimental design enables observation of the direct effect of the selected inhibitor on the steady-state transplacental passage of Rho123 within one experiment. In addition to this, the steady-state experimental approach eliminates the variability between two groups of experiments carried out under nonsteadystate experiments. On the other hand, the effect of inhibitors on the passage of Rho123 is less evident under steady-state in comparison with the data obtained in the nonsteady-state experiments.
For the examination of the feto-maternal passage of Rho123 under steady-state conditions, fetal solution containing Rho123 (0.65 μM) was used to perfuse the selected placenta via the catheter in the fetal umbilical artery immediately after catheterization. Sample collection started 10 min after the installation of the catheter (time 0). In the 12th min of the experiment, one of the P-glycoprotein inhibitors was added to the maternal reservoir to reach a concentration of 10 μM (CsA and PSC833), 40 μM (QND), and 5 mM (sodium azide), respectively. Samples were collected in 3-min intervals from the fetal umbilical vein for 36 min of experiments.
Our previous results suggested that inhibition of P-gp was more significant when an inhibitor of P-gp was present in the maternal compartment rather than in the fetal one (Pavek et al., 2001). That is why the inhibitors/modulators of P-gp were given into the maternal solution even in the feto-maternal experiments.
In our previous study (Pavek et al., 2001), we also demonstrated that sodium azide (5 mM) increased the transplacental paracellular passage of l-[3H]glucose in later intervals, which could suggest impairment of the placental barrier. Therefore, we conducted experiments with sodium azide only for 27 min after catheterization.
Effect of Maternal Inflow Concentration on the Transplacental Clearance of Rho123. The dependence of the materno-fetal transplacental passage of Rho123 on the concentration of Rho123 in the maternal reservoir was examined in the steady state. Rho123 was brought to the perfused placenta via the uterine artery in various concentrations of 0.42, 0.65, 1.3, 2.0, and 4.0 μM, respectively. Samples were collected in 3-min intervals from the fetal umbilical vein for 30 min of experiments. The mean materno-fetal clearance of Rho123 was calculated for every concentration from all measured intervals.
Calculations. To describe the transfer of Rho123 across the dually perfused rat placenta in the materno-fetal direction in both nonsteady-state and steady-state experiments, its materno-fetal transplacental clearance (CLmf) was calculated according to eq. 1 (Mohammed et al., 1993). where Cma (in micromolar) is the concentration of Rho123 in the maternal reservoir, Cfv (in micromolar) is the concentration of Rho123 in the umbilical vein effluent, Qf (milliliters per minute) is the umbilical flow rate and wp (in grams) is the wet weight of the placenta. The amount of Rho123 that passed the placenta within any interval in the materno-fetal direction was calculated by multiplication of the volume of perfusion liquid collected from the umbilical vein within this interval by the concentration of Rho123 in the sample. This amount of Rho123 was recalculated per wet weight of the perfused placenta and expressed in nanomoles per gram units. Similarly, the feto-maternal clearance (CLfm) was calculated according to eq. 2. where Cfa (in micromolar) is the concentration of Rho123 in the fetal reservoir entering the perfused placenta via the umbilical artery, Cfv (in micromolar) is the concentration of Rho123 in the umbilical vein effluent, Qf (in milliliters per minute) is the umbilical flow and wp (in grams) is the wet weight of the placenta. The amount of Rho123 that passed across the perfused placenta in the feto-maternal direction during any interval was calculated as the difference between the concentrations of the fetal artery inflow and the fetal vein outflow (the concentration in the sample) multiplied by the volume of perfusion liquid collected from the umbilical vein within the interval indicated.
In the nonsteady-state experiment, the total cumulative amount of Rho123 that passed across the placenta within 25 min of the experiment was calculated to assess the influence of CsA on the transplacental passage of Rho123 in both directions. To eliminate variations of the cumulative amount caused by the weight of perfused placentas, the amount of Rho123 was expressed as the amount per wet weight of the perfused placenta (nanomoles per gram).
In the steady-state experiments, the effects of P-gp inhibitors given to the maternal solution were evaluated from the following inhibitory ratio (eq. 3). where X12–30 is the amount of Rho123 that passed the placenta from the 12th to 30th minute of experiments (i.e., the period in which an inhibitor was present) and X0–12 is the amount of Rho123 that passed the placenta from zero to the 12th minute (i.e., a period with inhibitor-free solution).
If sodium azide was examined as an inhibitor of P-gp, both the materno-fetal and the feto-maternal experiments were performed up to 27 min, and the ratio was calculated as X12–27/X0–12.
In the studies where the effect of maternal inflow concentration on the transplacental clearance of Rho123 was examined, the data were fitted by eq. 4. where CLpassive diffusion is a clearance of passive transport of Rho123 across the placenta in the materno-fetal direction, CLmax is the maximum reverse feto-maternal clearance of Rho123 mediated by P-gp function and Km is the Michaelis-Menten constant of the P-gpmediated reverse transport.
HPLC Analysis of Rho123 and Its Metabolites in the Perfusion Media. Solid phase extraction of analytes from perfusate samples was performed according to the method of Sweatman et al. (1990) with slight modifications. Visiprep solid phase extraction vacuum manifold (12-port; Supelco, Bellefonte, PA) with SPE columns (Supelclean LC-18, 1-ml tubes; Supelco) were used for the solid phase extraction. The dry extract in the glass tube was reconstituted in 600 μl of the mobile phase, centrifuged, and transferred into a vial of the autosampler. The sample (100 μl) was injected into the chromatographic column.
Chromatographic analyses were performed using chromatograph (Thermo Separation Products, Minneapolis, MN; formerly Spectra Physics). The chromatographic system consisted of a SCM400 solvent degasser, P4000 quaternary gradient pump, AS 3500 autosampler with 100-μl sample loop, SpectraFOCUS high-speed scanning UV-visible detector, SN4000 system controller, and data station (Intel-Pentium 166 MMX, RAM 64 MB, HDD 2GB) with the analytical software ChromQuest 2.1 (ThermoQuest, Inc., San Jose, CA). A LiChroCART 125-4 mm analytical column packed with Purospher RP-18e, 5 μm and precolumn LiChroCART 4-4 mm with the same stationary phase (Merck) were used for analyses. The mobile phase was composed of acetonitrile/0.01 M phosphate buffer, pH 3 (3:7, v/v). Flow rate was 1 ml · min-1.
UV-visible detection was performed either in dual wavelength mode at 500 nm (for rhodamines) and 550 nm (for diamond fuchsin used as an internal standard) or in high-speed scanning mode (range 195–750 nm with 1-nm distance, used for UV-visible spectra collection).
The retention times of Rho110, Rho123, and Diamond fuchsin under the above-mentioned chromatographic conditions were 2.40, 4.09, and 8.44 min, respectively. The whole HPLC analysis lasted 14 min.
Fluorimetric Determination of Rho123 in Perfusion Media. A commercial SIA system with an eight-port selection valve and a fluorometric detector equipped with a flow cell was used for fluorometric determination of Rho123 in samples (Sklenarova et al., 2002).
Cell Line Cultivation and Isolation of the Membrane Fraction from the Cells. The lymphoid macrophage cell line P388 and its resistant P-gp expressing subline P388-MDR were gifts from Dr. St'astny (Institute of Microbiology, Academy of Sciences of the Czech Republic). The membrane fractions of the cell lines were used as positive and negative controls for immunochemical determination of P-gp in the rat placenta. Cells were cultured as was reported previously (St'astny et al., 1999).
Isolation of the Total Membrane Fraction from the Rat Placentas. Rat placentas were collected on the 22nd day of gestation and were homogenized in ice-cold buffer (1:1, v/w) containing 250 mM sucrose, 10 mM Tris, 5 mM EDTA, 10 μg/ml leupeptin, 10 μg/ml pepstatin, and 1 mM phenylmethylsulfonyl fluoride, pH 7.4, in a Potter-Elvehjem tissue homogenizer. Homogenized tissues were centrifuged at 15,000g for 15 min at 4°C. The supernatants were further spun at 100,000g for 1 h at 4°C and the pellets containing the membrane fraction were sonicated and resuspended in ice-cold phosphate-buffered saline, pH 7.4.
Isolation of Apical Membrane Fraction of the Rat Syncytiotrophoblast. Apical membrane fraction of the rat placentas was isolated using the method described by Malandro et al. (1996) with slight modifications. Briefly, the total membrane fraction was resuspended in Tris-mannitol buffer (300 mM mannitol, 2 mM Tris base, pH 7.0) and homogenized in a glass homogenizer with Teflon pestle. MgCl2 was added to a final concentration of 10 mM. Then membranes were again homogenized, incubated for 10 min at room temperature, and centrifuged at 2200g for 12 min. The pellet was discarded and supernatant was centrifuged at 100,000g for 60 min to pellet the apical-enriched membrane vesicles. The membrane fractions were resuspended in HEPES-sucrose buffer (300 mM sucrose, 10 mM HEPES-Tris base, pH 7.4) and frozen at -74°C.
Isolation of Basal (basolateral) Membrane Fraction of the Rat Syncytiotrophoblast. The basal membrane fraction of the rat syncytiotrophoblast was isolated according to the method of Malandro et al. (1996). Bicinchoninic acid protein assay reagent kit (Pierce Chemical, Rockford, IL) was used to determine the protein content in the samples. The purity of membrane fractions was analyzed by determination of alkaline phosphatase activity (the marker enzyme of the apical membrane) and Ca2+-ATPase (the marker enzymes of the basolateral membrane, according to the method of Malandro et al., 1996).
Alkaline Phosphate Activity Assay. Apical and total membrane fractions were dissolved in glycine buffer (0.1 M glycine, 1 mM MgCl2, 1 mM ZnCl2, pH 10) to final protein concentration of 500 μg/ml. 4-Nitrophenyl phosphate disodium salt was added to a final concentration of 0.75 mg/ml. The mixture was incubated for 20 min at 37°C and then absorbance was measured at 410 nm (Helios gamma spectrophotometer; Thermo Spectronic, Rochester, NY).
Western Blotting of P-glycoprotein in the Rat Placenta. The Western blotting analysis was performed to detect P-glycoprotein in the rat placenta membrane fractions. Membrane fractions were suspended in equal volumes of Laemmli sample buffer, and 20 μg of protein per lane was resolved by 7% polyacrylamide SDS-PAGE gel (50 mA) and eletrotransferred (25 V, 300 mA) to nitrocellulose Hybond membranes (Amersham Biosciences UK, Ltd., Little Chalfont, Buckinghamshire, UK). The nitrocellulose membranes were blocked in 5% blocking solution (Amersham Biosciences UK, Ltd.) and incubated in Tris-buffered saline-Tween 0.01% buffer with murine monoclonal antibody (mAb) C219 (Signet Laboratories, Dedham, MA) diluted 1:500, or with hamster mAb Ab-2/F4 (p170/P-glycoprotein/MDR Ab-2; LabVision Neomarkers, Fremont, CA). The anti-mouse secondary horseradish peroxidase-conjugated antibody (1:1000 dilution) and ECL Western blotting detection kit (both Amersham Biosciences UK, Ltd.) were used for autoradiographic detection of P-gp on FOMA blue medical X-ray films (Foma Bohemia A.S, Hradec Králové, Czech Republic). Densitometric analysis was performed using a high-resolution scanner HP 5400c (Hewlett Packard, Palo Alto, CA) and LabImage gel densitometric software version. 2.62 (Kaplan GmbH, Halle, Germany).
Immunohistochemical Localization of P-gp in the Rat Placenta. Specimens of the placentas were fixed in 4% paraformaldehyde (pH 7.35) or in Bouin's fixative fluid and then were paraffinembedded. Sections of placentas (thickness, 5 μm) were mounted on an object slide, dewaxed, and rehydrated through a series of ethanol solutions. Endogenous peroxidase activity was blocked with 3% H2O2 in 50% methanol solution for 20 min. For heat-induced antigen retrieval, the slides were boiled in 0.1 M Tris-HCl buffer, pH 1.5, for 15 min in a microwave oven at 750 W. Blocking of nonspecific background staining was performed with 10% normal goat serum (Sigma Chemie, Steinheim, Germany) in phosphate-buffered saline (PBS) solution (pH 7.4) for 30 min. Slides were incubated with primary antibody for P-glycoprotein (C219; Signet Laboratories) diluted 1:50 in PBS solution for 15 to 18 h at 4°C. After a PBS rinse, the slides were developed with the secondary antibody goat anti-mouse Ig conjugated to peroxidase-labeled polymer (DAKO EnVision+ ready-to-use; DAKO, Carpinteria, CA). Secondary antibodies were visualized with diaminobenzidine (DAKO DAB substrate-chromogen solution; DAKO) and hematoxylin counterstained. As a control for background staining, control slides were treated in the same manner, except PBS solution was substituted for the primary antibody to P-glycoprotein. Slides were examined using computer image analysis (Hund h500 light microscope, Helmut Hund, Wetzlar, Germany; JVC TK C1380E color video camera, JVC, Tokyo, Japan; LUCIA, version 4.61software, Laboratory Imaging Prague, spol. s r. o., Prague, Czech Republic).
Examination of P-gp Genes Expression by RT-PCR Method. Rat placentas were collected into liquid nitrogen and homogenized in TRIzol reagent (Invitrogen, Carlsbad, CA) using 1 ml of TRIzol/50 mg of tissue. Total RNA was extracted from the tissue homogenate according to the manufacturer's instructions. RNA integrity was assessed by electrophoresis on a 1% agarose gel. First-strand cDNA was prepared from 1 μg of total RNA with AMV transcriptase (Finnzymes Oy, Espoo, Finland) using oligo(dT) primer under conditions recommended by the manufacturer. cDNA prepared from 3 ng of total RNA was amplified by PCR with HotStar Taq polymerase (QIAGEN, Valencia, CA); 3 mM MgCl2, 0.2 mM dNTP, 0.3 μM each primer, 0.03 U/μl polymerase; 95°C for 15 min followed by 35 cycles of 95°C for 30 s, 54°C for 1 min, 72°C for 1 min. The sequences of antisense primers were identical for both mdr1a and mdr1b isoforms: 5′-AGCATTTCTGTATGGTATCTGCAAGC-3′. Sequence of mdr1a sense primer was 5′-CTGCTCAAGTGAAAGGGGCTACA-3′ (product length 329 bp) and mdr1b sense primer 5′-CGCTTCTAATGTTAAAGGGGCTATG-3′ (product length 331 bp). β2-Microglobulin was used as a housekeeping gene (B2M antisense: 5′-TACATGTCTCGGTCCCAGGTGA-3′; B2M sense: 5′-TGCCATTCAGAAAACTCCCCA-3′, product length 303 bp). Amplified segments were analyzed by electrophoresis on 1.5% agarose gel and visualized using ethidium bromide.
Statistical Analysis. Differences between group mean values were assessed by unpaired Student's t test using STATISTICA software version. 6 (StatSoft, Tulsa, OK). Differences of p < 0.05 were considered statistically significant.
Results
Evaluation of Rho123 Metabolism in the Rat Placenta. We detected Rho123 and its major metabolite Rho110 in fetal effluent samples using HPLC (data are not presented). The concentrations of Rho110 in the fetal compartment ranged from 1.9 to 2.5% of Rho123 concentrations entering the placenta from the maternal compartment. None of the used inhibitors had any effect on Rho110 levels in the fetal umbilical effluent samples during experiments. We suppose that the presence of Rho110 in the fetal umbilical effluent is predominantly a consequence of contamination of Rho123 with Rho110 (5.27%) used in our experiments (Rhodamine 123, catalog no. R-8004, batch 127H3707; Sigma-Aldrich). Due to no or negligible biotransformation of Rho123 in the rat placenta, we could employ faster and easier fluorometric method for analysis of Rho123 in our samples.
Effect of Maternal Inflow Concentration on the Transplacental Clearance of Rho123. The materno-fetal transplacental passage of Rho123 was found to be dependent on the maternal inflow concentration in a range of 0.42 to 4.0 μM (Fig. 1). This saturable kinetics of Rho123 indicates that transplacental passage of Rho123 is a transporter-mediated process. After fitting the data to eq. 4, Clmax describing the efficiency of P-gp to pump Rho123 back into the maternal compartment was calculated to be 0.07 ± 0.01 ml · min-1 · g-1. Clpassive diffusion was calculated to be 0.11 ± 0.01 ml · min-1 · g-1. Therefore, our data indicate that in low concentrations of Rho123 (threshold concentration <0.10 μM), P-gp could completely reverse passive diffusion of Rho123 across the barrier (Fig. 1). In higher concentrations of Rho123 (>4.0 μM), however, P-gp-mediated transport of Rho123 become saturated and is not able to return all passively diffused Rho123 back to the maternal circulation. Unfortunately, fluorometric analysis of Rho123 used in this study did not allow examining lower concentrations of Rho123 than 0.42 μM in the maternal perfusion liquid.
Effect of CsA on Transplacental Passage of Rho123 under Nonsteady-State Conditions. Addition of 30× molar excess CsA to Rho123 solution in the maternal reservoir resulted in an increase of the materno-fetal transplacental clearance (Fig. 2A). Correspondingly, the amount of Rho123 that entered the fetal compartment rose significantly (p < 0.05) compared with controls (1.28 ± 0.43 nmol · g-1 and 2.17 ± 0.42 nmol · g-1, respectively). Doubling the amount of CsA in the maternal solution resulted in a comparable increase in the materno-fetal passage of Rho123 (data not shown). On the other hand, CsA decreased the feto-maternal clearance of Rho123 (Fig. 2B) and significantly lowered the amount of Rho123 passing from the fetal compartment into the maternal one (22.60 ± 2.18 and 19.47 ± 2.54 nmol · g-1, respectively).
In control experiments with Rho123 (0.65 μM), its materno-fetal clearances were rising to reach a plateau (steady state) from minute 15 of the experiment (data not shown). In the case of feto-maternal passage of Rho123, steady state was reached within 10 min of the experiments (Fig. 2B). The plateau phase of the transplacental passage of Rho123 enabled us to perform experiments under steady-state conditions (see the following paragraph).
Effect of CsA, PSC833, QND, CPZ, and Sodium Azide on Transplacental Passage of Rho123 in the Steady-State Experiments. Materno-fetal transplacental clearances of Rho123 were nearly constant during the control experiments (Fig. 3). An addition of PSC833, CsA, QND, or CPZ into the maternal reservoir in the 12th min of the experiment to reach a concentration of 10 μM (PSC833 and CsA) or 40 μM (QND and CPZ), resulted in an increase of materno-fetal transplacental passage of Rho123 (Table 1 and Fig. 3; data for CPZ are not shown). The most significant effect on the materno-fetal transplacental passage of Rho123 was observed in the case of PSC833 followed by QND and CsA (Table 1). On the contrary, in experiments where fetomaternal transplacental passage of Rho123 was examined, PSC833, CsA, and QND decreased feto-maternal passage of Rho123 (Table 1 and Fig. 3). QND had the most potent effect on the Rho123 feto-maternal transplacental passage followed by CsA and PSC833 (Table 1). These data indicate that inhibitors of P-gp accelerate materno-fetal passage of Rho123; on the other hand, inhibition of P-gp decreases the feto-maternal passage of Rho123. This corresponds well with the same phenomenon observed in epithelial cell lines cultured in monolayers (placental BeWo, Caco-2, etc.), where the inhibition of a one-way activity of P-gp increases the apical-to-basal and decreases the basal-to-apical transport of P-gp substrates (Yumoto et al., 1999; van der Sandt et al., 2000).
Apart from P-gp specific inhibitors, an ATP-synthesis inhibitor sodium azide was used to study the influence of ATP depletion on the transplacental passage of Rho123. Sodium azide affected the passage of Rho123 across the rat placenta supporting the hypothesis that P-glycoprotein, an ATP-dependent transporter, is involved in the regulation of the passage of Rho123 across the rat placenta (Fig. 4). The impact of sodium azide was comparable with that of PSC833, which is supposed to be one of the most potent specific inhibitors of P-gp (Table 1). On the other hand, the effect of sodium azide on the feto-maternal transplacental passage of Rho123 was not significant (Table 1; Fig. 4).
Comparison of Materno-Fetal and Feto-Maternal Passages of Rho123. Feto-maternal clearances of Rho123 were found to be significantly higher (p < 0.01) than clearances in the opposite direction both in steady-state and nonsteady-state experiments (Fig. 5). PSC833 and CsA, however, were able to partly annul this asymmetry between the materno-fetal and feto-maternal passages of Rho123 (Table 2 and Fig. 3). Thus, our results show that the passage of Rho123 across the intact rat placental barrier is asymmetric due to the activity of P-gp similarly as in the case of in vitro cultures of epithelial cell lines (intestinal Caco-2, Yumoto et al., 1999; kidney LLC-PK1:MDR1, van der Sandt et al., 2000). Unlike the placental BeWo epithelial cell line, where inhibition of P-gp lead to the same materno-fetal and fetomaternal passages of P-gp substrates (Ushigome et al., 2000), we did not observe a complete annulment of P-gp function by inhibitors in the intact rat placenta. This discrepancy suggests that transplacental passage across the intact placental barrier is a more complex process in comparison with cellular epithelial models.
Western Blotting of P-glycoprotein in Placental Membrane Fractions. P-glycoprotein (molecular mass 150 kDa) was detected in the placental membrane fractions using C219 and Ab-2/F4 monoclonal antibodies (Fig. 6). Monoclonal antibody Ab-2/F4 immunoreacts with P-glycoproteins encoded by human (MDR1) and rat (mdr1a, mdr1b) genes (Huang et al., 2001); mAb C219 shows additional immunoreactivity with mdr2 and sister of P-glycoprotein gene products. The highest levels of P-glycoprotein were found in the apical membrane fraction of the rat syncytiotrophoblast (enrichment for alkaline phosphatase activity, 10.40 ± 0.49). On the other hand, the basal membrane fractions showed weak signal for P-glycoprotein.
Apical and basal membrane isolation methods used in this article (see Materials and Methods) have been found to yield the apical membrane fraction of the rat syncytiotrophoblast layer II and the basal membrane fraction of the rat syncytiotrophoblast layer III (Novak et al., 1997). Thus, we suggest that P-glycoprotein is localized predominantly in the apical membrane of the layer II, which forms crucial part of the rat materno-fetal placental barrier.
Immunohistochemical Localization of P-gp in the Rat Placenta. Antigen retrieval immunohistochemistry was performed for localization of P-gp in the rat placenta using mAb C219. The rat chorioallantoic placenta is composed of two distinct zones, junctional and labyrinthine. Strong immunoreactivity of P-glycoprotein was observed only in the inner layers (second or third layer) of the syncytiotrophoblast of the labyrinth zone (Fig. 7). This well corresponds with the fact that the labyrinth zone is thought to be the exchange area of nutriments and drugs between mother and fetuses (the placental “barrier”). Surprisingly, the brush-border membrane of the first syncytiotrophoblast layer, which directly faces maternal blood, was not immunostained as observed in the human placenta (Nakamura et al., 1997; Lankas et al., 1998). Fetal capillaries of the labyrinth zone were without any immunoreactivity. Spongiotrophoblast cells of the rat placental junctional zone, which are important in placental hormone production, were found negative (data not shown). These data further support the hypothesis that Pglycoprotein is present especially in the apical membrane of the second syncytiotrophoblast layer of the rat term placenta.
Expression of Genes Encoding P-gp in the Rat Placenta. Both mdr1a and mdr1b gene expressions were determined in the rat term placenta by RT-PCR analysis (Fig. 8). We detect mRNAs of both isoforms of mdr1 genes encoding rat P-glycoprotein in the rat placentas on the 22nd day of gestation. Our preliminary results show that expression of mbr1b gene could be dominant in the rat term placenta. Nevertheless, confirmation of these data using real-time RT-PCR method is in progress.
Discussion
P-gp has been demonstrated to be an important determinant of the pharmacokinetics of some lipophilic compounds in various body tissues. However, little is known about the functional expression of P-gp in the placental barrier. Moreover, clinical studies are not feasible in this case because of safety of the developing fetus. Substrates recognized by P-gp include tens of drugs, some of them applied in pregnancy.
The present study examines the functional activity of P-gp in the intact chorioallantoic rat placenta. Both the rat and human placentas are of the hemochorial type, thus the placental barriers of both species are very similar from the morphological and histological points of view. A recent study employing the human trophoblast BeWo cell line brought a new view on the function of P-gp in the placental trophoblast, suggesting that 1) the passage of P-gp substrates is different in the apical-to-basolateral and the basolateral-to-apical directions and that 2) P-gp accelerates the passage of some substrates in the basolateral-to-apical direction (Ushigome et al., 2000). On the basis of the in vitro data achieved using BeWo cells, we speculated that, in addition to regulating the materno-fetal passage, P-gp could also stimulate the elimination of its substrates from the fetal circulation in the feto-maternal direction. To confirm this speculation, the pharmacokinetics of Rho123 across the dually perfused rat placenta in situ was investigated. Rhodamine 123, a fluorescent dye, is a well established model compound for the evaluation of the transport activity of P-gp in different sites of the body and for testing of tumor cells for MDR mediated by P-gp (Ludescher et al., 1992; Masereeuw et al., 1997; de Lange et al., 1998; Yumoto et al., 1999; van der Sandt et al., 2000). Nevertheless, some authors suggest that a rOCT could also participate in Rho123 transport (Masereeuw et al., 1997; van der Sandt et al., 2000). To assess the permeability of the intact placental barrier for Rho123, transplacental passage of Rho123 across the placenta was expressed as the transplacental clearance per the wet weight of the placenta (milliliters per minute per gram). Similarly to the permeability coefficient calculated in in vitro epithelial studies, clearance is a constant that characterizes the ability of a drug to pass through the barrier in the in vivo experiments.
The materno-fetal transplacental passage of Rho123 did not show the characteristics of linear pharmacokinetics, which suggests involvement of a transport process different from passive transport mechanisms (Fig. 1). The same saturation process was found in the case of CsA in the perfused rat placenta, but not in the case of l-[3H]glucose, a marker of passive transport (Pavek et al., 2001). One can speculate that in very low concentrations, P-gp completely returns its substrates back into the maternal compartment. In higher concentrations, however, P-gp becomes saturated and drugs may pass the barrier by passive transport. Under the steady-state conditions, the transplacental clearance of Rho123 was found to be 8.48 times higher in the feto-maternal than in the materno-fetal direction (Table 2; Fig. 3). In BeWo cell monolayers, the transepithelial passages of other substrates of P-gp, such as vinblastine, vincristine, and digoxin, were found to be 6.2-, 3.7-, and 5.0-fold higher, respectively, in the basolateral-to-apical direction than those in the opposite direction (Ushigome et al., 2000). Thus, the asymmetry is higher in the rat placenta compared with the BeWo model. We found that inhibitors of P-gp, such as PSC833, CsA, and CZP, were able to increase significantly the materno-fetal transplacental passage of Rho123 (Fig. 3 and Table 1). In addition, PSC833 and QND significantly decreased the fetomaternal passage of Rho123 across the placental barrier (Fig. 2 and Table 1). Surprisingly, the inhibitors had less influence on the feto-maternal transport than on the opposite one (Fig. 3 and Table 1). In BeWo cultures, CsA was able to completely cancel the asymmetry of passage, resulting in the same apical-to-basolateral and basolateral-to-apical transport of three examined model compounds (Ushigome et al., 2000). In the intact rat placental barrier, however, P-gp inhibitors did not lead to a complete loss of P-gp function. The feto-maternal clearance of Rho123 was found to be higher than the materno-fetal one even in the last intervals of the experiments. The discrepancies between the data obtained from BeWo cell line and those from the perfused rat placenta demonstrate that the transplacental passage across the intact placental barrier is a more complex process. Other factors, such as presence of additional transport mechanisms, plasma and/or tissue binding, pressures and flows of the perfusion solutions, etc., can influence the transplacental passage of Rho123 across the intact placenta. Despite the fact that syncytiotrophoblast is thought to be the principal barrier component, it seems plausible that others layers, such as endothelia of fetal capillaries, cytotrophoblast cells, and basal laminas, could partly influence transplacental passage of Rho123 in the intact placenta. Because P-gp showed lower effect on the feto-maternal passage of Rho123 in comparison with the materno-fetal passage of Rho123 (Table 1), passive diffusion seems to be important transport mechanism of Rho123 across the membranes in the feto-maternal direction as originally suggested by Stein (1997) in MDR-resistant cell lines. QND influenced the passage of Rho123 to the same extent as specific and highly potent inhibitors of P-gp, such as PSC833 and CsA. Because QND also inhibits rOCT (Pritchard and Miller, 1993), our results suggest that there could be a coinvolvement of an rOCT transporter in the regulation of the transplacental passage of Rho123. P-gp and rOCT1 may provide parallel transport mechanisms in the rat placenta similarly as it was reported in the kidney (Miller, 1995). Sodium azide also increased the materno-fetal passage of Rho123. Thus, ATP-depletion could result in abolition of P-gp protective function and in an increased passage of Rho123 across the placental barrier in the materno-fetal direction. Similar observations have been found in Caco-2 epithelial cell model (Augustijns et al., 1993). Based on these findings, we can speculate that placental hypoxia and placental ATP depletion could result in a decreased function of P-gp, and consequently in a higher exposure of human fetuses to some lipophilic xenobiotics from the maternal circulation. This speculation further emphasizes the importance of studying the functional activity of P-gp in the placental barrier.
To exclude possible interference of Rho123 metabolism with the examined transplacental passage of Rho123, HPLC method was used to determine possible biotransformation of Rho123 in the rat placental barrier. As reported by Sweatman et al. (1990), metabolites of Rho123, such as Rho110 (deacylated metabolite of Rho123) and/or its glucuronide conjugate, are formed in the rat. We found that only negligible amount of Rho123 passing the rat placenta was metabolized into Rho110, and therefore metabolism of Rho123 did not interfere with the study of P-gp-mediated transport processes.
To detect and localize P-gp in the rat placentas, immunochemical and immunohistochemical methods using mAb C219 and Ab-2/F4 have been used. As opposed to the human placental barrier that comprises one syncytiotrophoblast layer, there are three trophoblastic layers in the rat placental labyrinth; two of them are thought to be syncytium (layer II and III). Layer I facing maternal blood is of cellular nature with numerous fenestrations without diaphragmata and thus do not represent a barrier to small compounds (Metz et al., 1976, 1978). Syncytial layers II and III are connected each other by numerous gap junction channels composed preferentially of connexin26 (Metz et al., 1976; Shin et al., 1996). The gap junctions enable exchange of small watersoluble substances, e.g., glucose, between syncytiotrophoblast layers as suggested by Shin et al. (1997) and as was clearly demonstrated by Gabriel et al. (1998) in connexin26-defective mice. We found the highest levels of P-gp predominantly in the apical membranes of the second syncytiotrophoblast layer, which is considered the “barrier” layer of the rat chorioallantoic placenta to maternal blood (Metz et al., 1978; Enders and Blankenship, 1999). Review on both human and rat placental histology and morphology has been published recently by Enders and Blankenship (1999). This localization of P-glycoprotein corresponds well with the function of the membrane. Continuous location of the P-gp in the apical membrane of the syncytiotrophoblast layer II suggests that P-gp is strategically located to face the incoming xenobiotics from the maternal blood before they enter fetal vessels and other fetal tissues. Our results also suggest that both mdr1a and mdr1b genes encoding rat P-glycoprotein are expressed in the rat-term placenta (Fig. 8). Our preliminary results show that mbr1b gene product could prevail over mdr1a, however, we are currently using real-time-RT-PCR method to confirm the observation.
We conclude that our results support the hypothesis that P-gp contributes to the barrier function of the placenta. In addition, P-gp was found to accelerate the feto-maternal elimination of its substrates. Both these P-gp functions contribute simultaneously to the protection of the fetus against toxic injury. Consequently, P-gp inhibitors could increase the amount of substrates entering the fetus. Together, the present study emphasizes the importance of assessing the functional activity of P-gp in the human placental barrier since this knowledge could have important toxicological and therapeutic implications in pregnancy.
Acknowledgments
We thank Dana Soucková and Anezka Kunová for excellent assistance with the dually perfused rat placenta method.
Footnotes
-
This work was supported by the Grant Agency of the Czech Republic (Grants 305/01/0D89 and 305/01/0441), the Grant Agency of the Charles University in Prague (Grant 94/200/C), the Research Center LN00B125, and Research Project No. J13/98:11600002 obtained from the Czech Ministry of Education. The results of the article were presented in part at the XIVth World Congress of Pharmacology, San Francisco, 2002 and the 62nd Congress of FIP, Pharmacy and Pharmaceutical Sciences World Congress, Nice, France 2002.
-
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
-
DOI: 10.1124/jpet.102.048470.
-
ABBREVIATIONS: P-gp, P-glycoprotein; MDR, multidrug resistance; Rho123, rhodamine 123; rOCT, rat organic cation transporter; RT-PCR, reverse transcription-polymerase chain reaction; CsA, cyclosporine; QND, quinidine; bp, base pair(s); CPZ, chlorpromazine; HPLC, high-performance liquid chromatography; mAb, monoclonal antibody.
- Received December 21, 2002.
- Accepted February 25, 2003.
- The American Society for Pharmacology and Experimental Therapeutics