In vitro study of placental trophoblast calcium uptake using JEG-3 human choriocarcinoma cells

Mammalian fetal nutrition during development is wholly dependent on the transport of nutrients by the placenta (Boyd, 1987; Hill and Longo, 1980; Munro et al. 1983; Shennan and Boyd, 1987; Truman and Ford, 1984). Calcium, needed for skeletal formation, neuromuscular functions, and other physiological activities, is transported actively across the placenta from the maternal to the fetal circulation (Brunette, 1988; Pitkin, 1985; van Kreel and van Dijk, 1983). This process is carried out by the placental trophoblastic cells (Dearden and Ockleford, 1983; Loke and Whyte, 1983), which line the chorionic villi and represent the epithelial layer separating the maternal and fetal circulations. In the human placental chorionic villi, a layer of syncytiotrophoblaste lies over the cytotrophoblasts, and surrounds the internal mesoderm and fetal capillaries (Ramsey, 1975). Calcium transport by placental trophoblasts is therefore analogous to epithelial transport in general, i.e. calcium is moved in a transcellular manner. Furthermore, the calcium level is higher in the fetal circulation than in the maternal circulation. It is well established that calcium is actively transported against a concentration gradient by placental trophoblasts.

Our laboratory has been studying the cellular and molecular mechanism and regulation of placental calcium transport (Tuan, 1982, 1985; Tuan and Bigioni, 1990; Tuan and Cavanaugh, 1986; Tuan and Kirwin, 1988; Tuan and Kushner, 1987; Tuan et al. 1988). These studies, which are summarized below, have shown that mammalian placental trophoblasts express two marker molecules that are functional components of the calcium transport mechanism: a specific, high-M_r calcium-binding protein (CaBP) and an integral membrane Ca²⁺-activated ATPase. Both the CaBP and the Ca²⁺-ATPase are expressed as a function of embryonic development in a manner that parallels the onset of placental calcium transport. The trophoblastic localization of the CaBP and Ca²⁺-ATPase was revealed by immunohistochemistry and enzyme histochemistry, respectively. The functional involvement of the CaBP and Ca²⁺-ATPase in transmembrane calcium transport was demonstrated using cell-free placental membrane vesicles, whose active, ATP-dependent calcium uptake was inhibited by antibodies directed against the CaBP and inhibitors of the Ca²⁺-ATPase. Recently, a cDNA to the placental CaBP has been cloned, and has been used to probe its gene expression with respect to temporal profile (Northern blot hybridization) and cellular location (in situ hybridization). Thus, CaBP mRNA is found to increase in level during development, and is expressed only by placental trophoblasts. These studies have been carried out with both human and mouse placentae, with essentially similar results.

Although much information on the physiological, molecular and biochemical aspects of trophoblast calcium transport has been obtained using whole placental tissue, the cellular mechanism of the transport process remains unknown. The experimental means of analyzing transport are mostly limited to either whole organ perfusion (Sweiry et al. 1980; van Kreel and van Dijk, 1983) or uptake by cell-free vesicles (Bissonnette, 1982; Fisher et al. 1987; Shami et al. 1975; Tuan, 1985; Whitsett and Tsang, 1980), whereas whole cell studies dealing with isolated primary trophoblasts are made complicated by their varying homogeneity and availability (Contractor et al. 1984; Hunt et al. 1989). As in other epithelial transport systems (Rodriguez-Boulan and Nelson, 1989), these problems may be alleviated by identifying and studying a stable cell line in vitro, perhaps analogous to the popular Madin-Darby canine kidney (MDCK) cells. Towards this goal, we report here our findings on the characterization of a candidate in vitro cellular model of trophoblast calcium uptake, the human JEG-3 choriocarcinoma cells. Derived originally from the BeWo human choriocarcinoma cell line, JEG-3 cells in culture form large, multinucleated syncytia (Babalola et al. 1990), and express abundant human chorionic gonadotropin and chorionic somatomammotropin, hallmarks of trophoblastic cells (Kohler and Bridson, 1971; Patillo and Gey, 1968). The objective of this study is to evaluate the JEG-3 cells as a potential model system for placental trophoblasts with regard to calcium transport. For this purpose, the cellular and biochemical properties and regulation of calcium handling by the JEG-3 cells have been characterized. In addition, in view of evidence suggesting that placental calcium transport may be under vitamin D regulation (Brunette, 1988; Bruns et al. 1978; Danau et al. 1981; Halloran, 1989; Lester, 1986; Tanaka et al. 1979; Stumpf et al. 1983; van Bogaert, 1987), the response of JEG-3 cells to exogenous 1,25-dihydroxy vitamin D₃, the active vitamin D metabolite, was also characterized. The results reported here suggest that JEG-3 cells should serve as a valid and convenient in vitro model system for studying the cellular mechanism and regulation of transplacental calcium transport by trophoblasts.

JEG-3 cells were obtained from the American Tissue Type Culture and were grown in RPMI1640 (Cell-Gro) medium, 20 mM Hepes, pH7.4, supplemented with fetal calf serum (10%) and penicillin-streptomycin, and maintained at 37 °C in 5% CO₂. Optimal cellular attachment was obtained with Nunc or Costar tissue culture plastic ware. In some experiments, the cultures were supplemented with 1,25-dihydroxy vitamin D₃ (Biomol Research Lab., Inc.) at various concentrations for 24 h.

Immunodetection of CaBP in JEG-3 cells was carried out using rabbit antibodies produced against the human placental CaBP (HCaBP) (Tuan, 1982, 1985). Cells were washed in phosphate-buffered saline (PBS), and homogenized in a 20 mM Tris buffer (pH 7.4), centrifuged (31000 g, 30 min), and the soluble extract was fractionated by SDS-polyacrylamide gel electrophoresis (12% gel) (Laemmli and Favre, 1973). The gel was then electroblotted onto nitrocellulose and immunoreacted with anti-HCaBP followed by secondary antibodies conjugated with alkaline phosphatase, and then developed using a chromogenic substrate system consisting of bromochloroindolyl phosphate (BCIP) and nitroblue tétrazolium (NBT) (Ono and Tuan, 1990). The immunoreactive HCaBP band was scanned densitome-trically using a Hoeffer scanner. The relative level of CaBP in different JEG-3 samples was determined titrimetrically on the basis of signals generated from multiple serial dilutions of each sample.

Cells were fixed with 4% paraformaldehyde in PBS, rinsed with PBS, and incubated successively with anti-HCaBP antibodies, and fluorescein-conjugated goat antirabbit IgG antibodies (Tuan, 1985). After the final rinsing, the culture was mounted in PBS-glycerol and observed with epifluorescence and phase-contrast optics using an Olympus BH-2 microscope. Photography was done using Kodak Tri-X film.

Total cellular RNA was extracted from JEG-3 cells using the guanidine isothiocyanate procedure (Chomczynski and Saachi, 1987). The integrity of the RNA preparations was routinely examined by means of in vitro cell-free translation in a rabbit reticulocyte lysate system (Promega), followed by SDS–polyacrylamide gel electrophoresis, to ascertain the presence of high-M_r (⩾150×10³) translated products (Tuan and Kirwin, 1988). The RNA samples were fractionated by denaturing, formaldehyde-agarose (1%) gel electrophoresis (Lehrach et al. 1979), and blotted onto GeneScreen filter (New England Nuclear). The 0.7 kb CaBP cDNA insert was excised from pMCP (Tuan and Kirwin, 1988), ³²P-labelled by random priming (Feinberg and Vogelstein, 1983), and used to probe the RNA blot to detect CaBP mRNA (hybridization at 42 °C in 50% formamide, IM NaCl, and 1% SDS). The Northern blot was calibrated using RNA markers obtained from BRL. In addition, for CaBP mRNA quantitation, the RNA samples were slot-blotted in serial dilution onto GeneScreen filter (NEN-DuPont) and probed similarly with labelled pMCP or β-actin cDNA (provided by Dr D. Cleveland). Hybridization signals were quantified by densitométrie scanning of the autoradiograph.

The non-radioactive, biotin-based procedure was described recently (Liebhaber et al. 1989; McDonald and Tuan, 1989; Tuan et al. 1988). The cells were plated on glass or plastic chamber slides (Miles or Nunc) that had been coated with the cell culture bioadhesive, Adhera-Cell (Genex Corp.), fixed in modified Carnoy’s fixative, dehydrated, digested with Proteinase K, denatured with formamide, and hybridized with pMCP, which had been biotin labelled by nick translation using biotin-dUTP (BRL). Hybridization was detected using streptavidin conjugated with alkaline phosphatase, followed by chromogenic histochemistry using BCIP and NBT. Controls included omission of probe or the use of irrelevant DNA as probe.

JEG-3 cells were extracted in a Tris buffer (pH 7.4) containing 1% Triton X-100 (Tuan and Bigioni, 1990; Tuan and Knowles, 1984; Tuan and Kushner, 1987). Total Ca²⁺-ATPase enzyme activity, of both plasma membrane and intracellular orgin, in the Triton X-100-solubilized JEG-3 extract was determined using the molybdate-Malachite Green assay as described previously (Tuan and Knowles, 1984; Tuan and Kushner, 1987), and expressed as nmol phosphate released min⁻¹.

Ca²⁺-ATPase activity was also detected histo-chemically using two previously described procedures (Tuan and Bigioni, 1990; Tuan and Knowles, 1984; Tuan and Kushner, 1987). The first procedure involved electrophoretic fractionation of the solubilized cell extract on a non-denaturing Triton X-100 polyacrylamide gel, which was then incubated with ATP in the presence of PbCl₂, followed by Na₂S to detect the precipitated Pb. Enzyme activity was visualized as a dark brown band on the gel. For the second, cytohistochemical procedure, cells were fixed with 4% paraformaldehyde in Hepes–Pipes buffer, pH 7.4, for 5 min at room temperature, permeabilized with 0.01% Triton X-100 for 15 s, rinsed with Tris-buffered saline, and incubated with ATP (3 mM) in the presence of PbCl₂ for 30 min at 37 °C. The enzyme reaction product was detected using Na₂S to produce a brownish PbS precipitate, and the reacted cells were examined with bright-field and phase-contrast optics. Photography was done using Kodak Pan-X and Ektachrome films.

For all of the Ca²⁺-ATPase procedures described above, enzyme activity was defined as the difference in ATPase activity measured with and without Ca²⁺. Thus, all controls included omission of Ca²⁺, or the addition of, or pre-treatment of the gel with, EGTA. An additional control for cytohistochemistry also involved the omission of ATP.

After thorough rinsing of the tissue culture dish with Hank’s Balanced Salt Solution (Ca²⁺, Mg²⁺-free HBSS), cells were incubated with gentle agitation in HBSS containing 0.01 mM CaCl₂ and trace amount of ⁴⁸Ca at 37 °C for various periods of time. At the end of the incubation, the cells were rinsed in cold PBS (3 times), solubilized with 2% SDS, and the radioactivity incorporated was determined by liquid scintillation counting in Ecolume (ICN). In other experiments, calcium concentration and incubation temperature were varied. The effect of various agents was tested by pre-incubating the cells for 30 min and then measuring uptake in the presence of the respective agents. Uptake activities were expressed as pmol calcium min⁻¹.

The BCA reagent of Smith et al. (1985) was used to determine protein concentrations according to the protocol provided by the manufacturer (Pierce Chemicals).

This was measured by the Fura-2 method as reported previously (Akins et al. 1988; Akins and Tuan, 1989; Grynkiewicz et al. 1985). Cells were rinsed with Ca²⁺, Mg²⁺-free Hank’s Balanced Salt Solution (CMF-HBSS) to remove excess serum and calcium from the culture medium, and then treated with a solution of 5 μM Fura-2 acetoxy methylester (Fura-2/AM) for 15 min at room temperature followed by 15 min at 37°C. Cells were washed with CMF-HBSS, with 1 mM EGTA to remove unincorporated Fura-2/AM, and then mounted in a microscope cell chamber for observation. Changes in cytosolic [Ca²⁺], as a function of perfusion of the cells with 1–5 mM CaCl₂, were measured on the basis of the Ca²⁺-dependent Fura-2 differential fluorescence ratios using the Deltascan calcium-imaging system of Photon Technology International (South Brunswick, NJ) connected to a Leitz Fluovert inverted microscope. In this procedure, cells loaded with Fura-2 were initially viewed by phase-contrast microscopy to allow for optimal initial focusing and then alternately excited with either 350 or 380 nm (8 nm band pass) light. A single cell (or subsection of a cell) was selected visually and extraneous images were blocked with a field aperture. Emitted light passed through a 495 nm long pass filter, and was detected by a photomultiplier tube. The collected signals were corrected for background fluorescence, and analyzed using the Deltascan System software. The ratio, #350/330, corresponded to cytosolic [Ca²⁺], and could be determined on the fly during the experiment. All determinations were made using cells that yielded stable baseline fluorescence signals at each of the excitation wavelengths, indicating that no dye leak had occurred.

The Fura-2 ratiometric signal was calibrated in vitro using 100 mM KC1, ImM MgCl₂, 10 mM EGTA, 5mM Fura-2 and 10 mM Pipes, pH 7.0, with CaCl₂ added to set the free [Ca²⁺], An empirical calibration curve was generated, and all experimental R values were converted into cytosolic [Ca²⁺] values accordingly.

Cultured JEG-3 cells grew as large, connecting clusters, with a well-defined border surrounding the cellular mass (Fig-1). During the early phase of culture, the generally circular clusters (Fig. 1A and B) showed occasional fibroblast-like cellular extensions when expansion or spreading was apparent. The cells appeared mostly syncytial in the cluster, with a prominent nucleolar apparatus as well as various cytoplasmic particles (Fig. 1C); in addition, clear vacuole-like structures were often observed along the cellular cytoplasmic cortex. Because of their syncytial nature, JEG-3 cells were routinely passaged either by simply dividing scraped cells into new cultures or lifting them off the dish with a non-enzymatic cell dissociation solution (Specialty Media Inc.). Generally, cell scraping produced a culture of initial clusters containing about 30 cells each, whereas more uniformly sized aggregates were usually obtained using the cell dissociation solution. With vigorous scraping, single cells could sometimes be released. Even during long-term culture, JEG-3 cells tended to remain as largely syncytial aggregates; however, they did not exhibit any contact inhibition, since at very high densities during long-term culture, the aggregates eventually grew on top of one another, giving rise to three-dimensional masses that were clearly visible to the naked eye.

Since the placental CaBP was found to be specific for placental membranes, it was of great interest to ascertain whether it was also expressed by the JEG-3 cells, particularly if they were to be considered as potential candidate cells for the study of trophoblast calcium transport. The expression of CaBP was analyzed at both the protein and the mRNA levels.

The presence of CaBP in JEG-3 cells was demonstrated using antibodies specific for the human placental CaBP (Tuan, 1982, 1985). Thus, both immunoblotting (Fig. 2) and immunohistochemistry (Fig. 3) revealed the presence of the CaBP. As shown in Fig. 2, a single high-M_r (∼75×10³) immunoreactive protein band was observed on the immunoblot, and corresponded to that found in the human term placenta. Immunohistochemistry (Fig. 3) revealed that the CaBP was associated with the cytoplasmic areas of the JEG-3 cell (Fig. 3A and B). Immunospecificity was demonstrated by the lack of staining in controls where antibodies to CaBP were omitted (Fig. 3C and D). Interestingly, not all cells found within an island stained positively for CaBP; distinct cells showing a total absence of immunostaining were observed. This finding strongly suggested that the cell population was heterogeneous with respect to CaBP expression, and further supported the notion that the cell islands were not complete syncytia.

The expression of CaBP by JEG-3 cells was also detected by RNA analysis involving blot hybridization and in situ hybridization. The Northern RNA blot (Fig. 4) clearly indicated the presence of a 2.95 kb RNA hybridizing to the CaBP cDNA probe in both placental and JEG-3 total RNA isolates. When /J-actin mRNA hybridization signal was used as an internal standard, JEG-3 cells were found to have a 1.7-fold higher relative CaBP mRNA level than the placenta. In situ hybridization of JEG-3 cells, as shown in Fig. 5, also revealed distinct cells with positive signals for the CaBP. Interestingly, as described above for immunohistochemistry of the CaBP (Fig, 3), the cDNA-mRNA hybridization signals were also heterogenously distributed within the cell islands, since positive cells were found mingled with negative cells.

The analysis of the placental Ca²⁺-ATPase was based on enzyme activity in solubilized JEG-3 cell extract as well as in whole cells in situ.

A Triton X-100-solubilized extract of the JEG-3 cells exhibited significant Ca²⁺-activated ATPase activity, at a specific activity level of 13 nmol P₁ min⁻¹ mg⁻¹, which was slightly lower than that reported for the human term placenta (Tuan and Kushner, 1987). The membranous nature of the enzyme activity was indicated as the activity was undetectable in nondetergent extracted cells. The enzyme activity of the JEG-3 cells was sensitive to a number of pharmacochemical agents (Fig. 6A), similar to that of the human placenta (Tuan and Kushner, 1987). Further evidence supporting the identity between the two enzymes was observed upon electrophoretic fractionation of the activities (Fig. 6B). Thus, histochemical detection of enzyme activity on a nondenaturing electrophoretic gel showed that JEG-3 cells and placental microsomes contained an activity band with identical electrophoretic mobility, which corresponded to that previously identified as associated with trophoblasts (Tuan and Kushner, 1987).

Ca²⁺-ATPase activity was also detected histochemically in paraformaldehyde-fixed JEG-3 cells (Fig. 7A,B). The staining was most intense in the cellular periphery and in general appeared to be associated with areas of high cell density. Again, areas lacking any staining were also observed within a given island of cells, consistent with the cellular heterogeneity with respect to CaBP expression. The enzyme specificity for Ca²⁺ and ATP was also demonstrated in the histochemical reaction (Fig. 7C–K).

JEG-3 cells exhibited active uptake of extracellular calcium under the experimental conditions described in Materials and methods. The characteristics of JEG-3 calcium uptake are shown in Fig. 8A and B. Thus, calcium uptake was temperature-dependent and exhibited near-linear kinetics at 37°C for at least 7–8min (Fig. 8A). The uptake activity also increased proportionally with medium calcium concentration (Fig. 8B). Calcium uptake was dependent on the viability and membrane integrity of the JEG-3 cells, as paraformaldehyde fixation and digitonin treatment both significantly reduced calcium uptake activity (data not shown). In addition, JEG-3 calcium uptake was sensitive to a number of pharmacochemical agents. As shown in Fig. 8C, agents that were previously found to inhibit the Ca²⁺-ATPase activity (Fig. 6A), in particular quercetin and phenothiazin also effectively reduced calcium uptake by the JEG-3 cells. However, erythrosin B, which strongly inhibited Ca²⁺-ATPase activity (Fig. 6A), was ineffective on cellular calcium uptake (Fig. 8C), possibly due to membrane impermeability to the drug. Overall, the similarity between the pharmacochemical sensitivity of the calcium uptake and Ca²⁺-ATPase enzyme activities of the JEG-3 cells strongly suggested that these two activities might be related.

To gain insight into the process of calcium handling by the JEG-3 cells during active calcium uptake, cytosolic [Ca²⁺] was analyzed by Fura-2 microfluorometry. As shown in Fig. 9, washed JEG-3 cells exposed to external calcium (5 mM) showed an immediate, rapid, and transient rise in cytosolic [Ca²⁺], preceding by a considerable time interval the actual cellular uptake of ‘^t5Ca (Fig. 8A). In fact, cytosolic [Ca²⁺] rapidly decayed to a plateau only slightly higher than the starting value. In addition, 0.5 MM 1,25-dihydroxy vitamin D3 treatment, which greatly enhanced calcium uptake (see below), did not significantly alter the cytosolic [Ca²⁺] profile (Fig. 9B). From this kinetic analysis, it appeared unlikely that the calcium uptake activity of JEG-3 cells could be accounted for simply by a rise in cytosolic free [Ca²⁺]; an alternative calcium sequestration mechanism was thus necessary.

To investigate the effect of 1,25-dihydroxy vitamin D₃, JEG-3 cells were treated with the hormone at various doses for 24 h and then analyzed for their calcium uptake activity and level of CaBP expression. As shown in Fig. 10A, 1,25-dihydroxy vitamin D₃ treatment significantly stimulated JEG-3 calcium uptake in a dosedependent manner, the stimulation being maximal, 200%, at 0.5 HM. Interestingly, the enhancement of cellular calcium uptake was accompanied by a concomitant increase in the steady-state level of the CaBP, as determined by quantitative immunoblotting. The results in Fig. 10B, based on densitometry of the immunoreactive CaBP band on a Western blot, clearly indicated that 1,25-dihydroxy vitamin D₃ treatment stimulated CaBP by approximately twofold, compared to control. The stimulation of CaBP expression was also evident when JEG-3 cells were immunostained with antibodies to CaBP; thus, cells treated with 1,25-dihydroxy vitamin D₃ (Fig. 3F) exhibited significantly more intense staining than the untreated control (Fig. 3E). On the other hand, quantitative analysis of CaBP mRNA by serial RNA slot-blot hybridization with radiolabelled pMCP, using β-actin mRNA as an internal standard, did not reveal a concomitant increase in the steady-state level of CaBP mRNA in the JEG-3 cells treated with 1,25-dihydroxy vitamin D₃ (data not shown). Finally, the level of Ca²⁺-activated ATPase also remained unchanged in JEG-3 cells after 1,25-dihydroxy vitamin D₃ treatment (data not shown).

We have characterized the human choriocarcinoma cell line, JEG-3, as a candidate in vitro system for the study of placental calcium transport. The results reported here strongly indicate that the JEG-3 cells mimic placental trophoblasts with respect to many of the characteristics and biochemical activities associated with calcium transport. These include the expression of two previously identified marker molecules, a high-Af_r CaBP and a Ca²⁺-activated ATPase, temperature- and substrate-dependent and kinetically linear calcium uptake, and responsiveness to 1,25-dihydroxy vitamin D₃. Taken together with the syncytial morphology of the JEG-3 cells, which resembles that of the syncytiotrophoblasts, these findings support the validity of these cells as an in vitro model for studying placental calcium transport.

The important function of the placenta as the sole tissue responsible for nutrient translocation from the maternal to fetal circulation depends on the developmental differentiation of the trophoblasts into a specialized transporting epithelium (Dearden and Ockleford, 1983; Loke and Whyte, 1983). Thus, the syncytiotrophoblasts constitute a thin, tight epithelial interface between the maternal and fetal circulations. Detailed understanding of the mechanism of transport therefore requires studying the trophoblasts as an isolated cellular epithelium. Ideally, this should be a defined cell line that may be propagated reproducibly and also continuously express properties characteristic of the trophoblasts. The findings reported here are thus significant, since they establish the JEG-3 cells, a stable cell line, as an in vitro experimental model of placental calcium transport.

The biochemical activities of the JEG-3 cells mimic those of the term placental trophoblasts, with respect to the CaBP and Ca -ATPase, both of which appear to be identical to their placental counterparts. Thus, the JEG-3 CaBP is immunoreactive with anti-human CaBP antibodies and has identical Af_r and subcellular distribution with human placental CaBP; furthermore, the electrophoretic mobility and pharmacochemical sensitivity of the two Ca²⁺-ATPase activities are also similar. (Note: total cellular Ca²⁺-ATPase activities are analyzed, and thus both plasma membrane and intracellular activities are compared.) In addition, previous studies have also firmly established that the endocrinological (Kohler and Bridson, 1971; Patillo and Gey, 1968) and biochemical (Bahn et al. 1981; Hamilton et al. 1979) properties of the JEG-3 cells are similar to those of the placental syncytiotrophoblasts.

In this study we have characterized calcium uptake by JEG-3 cells cultured on tissue culture plastic. The substrate and temperature dependence, linear kinetics and membrane integrity requirement of the uptake function are all consistent with an energy-requiring, cellular uptake process. Interestingly, uptake activity is stimulated by 1,25-dihydroxy vitamin D₃ treatment of the JEG-3 cells, which also signficantly increases the level of the cytosolic CaBP, but not the Ca -ATPase. It should be noted that calcium uptake as measured here is a net result of influx and subsequent efflux of extracellular ⁴⁵Ca. Consequently, it is reasonable to speculate that the effect of 1,25-dihydroxy vitamin D₃ may be, first, the enhanced expression of the cytosolic CaBP, which then acts to increase retention of the influxed ^4BCa. Thus, the high-M_r CaBP of the placental trophoblasts may function in a manner similar to that of the low-Af_r calcium-binding protein, calbindin-9K. Calbindin, first discovered as the vitamin D-dependent calcium-binding protein of the intestinal mucosa (Kallfelz et al. 1967; Marche et al. 1977), is present in the placenta (Brims et al. 1978; Marche et al. 1978) and appears to increase in level as a function of gestation (Delorme et al. 1979). Various studies have suggested that calbindin may act as a cytosolic calcium sink to facilitate transepithelial calcium transport (Cara-foli, 1987; Wasserman and Fullmer, 1983), particularly in the intestinal mucosa, although its exact functional role remains to be established. It is thus noteworthy that in the JEG-3 cells, total cytosolic [Ca²⁺] did not rise concomitantly with calcium uptake, implicating a possible role for the CaBP as an intracellular calcium sequestrator. A number of recent studies have also implicated calbindin as functionally involved in placental transport; in addition, another calcium-binding protein, oncomodulin (Brewer and MacManus, 1985), first identified in transformed cells, has also been identified in the placenta. How all these proteins may participate in calcium transport, or possibly other metabolic functions of the placenta, remains to be elucidated. Nevertheless, the regulatory role of 1,25-dihydroxy vitamin D3 in placental calcium transport is strongly suggested, since both the high-M_r CaBP studied here and calbindin-9K are vitamin D-dependent. It should be noted that, unlike calbindin-9K (Mathieu et al. 1989), the vitamin D-stimulated increase in the level of CaBP appears to be unaccompanied by a concomitant increase in mRNA. The exact mechanism of vitamin D stimulation of CaBP expression therefore remains to be established.

Finally, the study of transepithelial transport in vitro ideally requires the use of a tight, polarized cellular epithelial sheet that will mimic the vectorial translocation of solutes (Aimers and Stirling, 1984; Rodriguez-Boulan and Nelson, 1989; Sabatini et al. 1983; Simons and Fuller, 1985), such as the widely studied MOCK cell line. As described earlier (Fig. 1), the JEG-3 cells will develop in long-term culture into a cellular sheet consisting of overlapping cell clusters. Although these cell islands are heterogeneous with respect to the expression of the CaBP and Ca²⁺-ATPase, the fact that they form a cell sheet is a distinct advantage in the study of transepithelial transport. Experiments are underway to examine the tightness and the polarity of the JEG-3 cell sheet. Initial observations on JEG-3 cells cultured on permeable membrane substrata suggest that a tight epithelium is indeed formed in vitro.

In summary, the studies presented here have provided a firm basis for the application of the human choriocarcinoma cell line, JEG-3, as an in vitro cellular model for the study of calcium transport by placental trophoblasts. In particular, the phenotypic stability and the hormone responsiveness of these cells should greatly facilitate the understanding of the mechanism and regulation of trophoblast calcium transport.

This work was supported in part by grants from the NIH (HD 15822, HD 21355), March of Dimes Birth Defects Foundation (1-1146), and the U.S. Department of Agriculture (88-37200-3746). The technical assistance of Ken Shepley in the in situ hybridization experiment is also acknowledged.