Molecular Human Reproduction Vol.8, No.5 pp. 485–493, 2002

Human choriocarcinoma cell line JEG-3 produces andsecretes active retinoids from retinol

Loıc Blanchon1, Patrick Sauvant2, Claes Bavik5, Denis Gallot4, Francoise Charbonne3,Marie-Cecile Alexandre-Gouabau2, Didier Lemery4, Bernard Jacquetin4,Bernard Dastugue1, Simon Ward5 and Vincent Sapin1,6

1INSERM U.384, Faculte de Medecine, BP 38, 63000 Clermont-Ferrand, 2INRA, Unite des Maladies Metaboliques etMicronutriments, Equipe Vitamines, Clermont-Ferrand/Theix, 63122 Saint Genes Champanelle, 3Laboratoire de Virologie Medicale,CHU, 4Medecine Materno-Foetale, Maternite de l’Hotel-Dieu, 63000 Clermont-Ferrand, France and 5Biomedical Genetics, Universityof Sheffield Medical School, Royal Hallamshire Hospital Sheffield, South Yorks, S10 2RX, UK

6To whom correspondence should be addressed at: INSERM U.384, Laboratoire de Biochimie Medicale, Faculte de Medecine et dePharmacie, 28, Place Henri Dunant, BP.38, 63000 Clermont-Ferrand, France. E-mail: visapin@courrier.u-clermont1.fr

Vitamin A (retinol) and its active derivatives (the retinoids) are essential for growth and development of themammalian fetus. Maternally-derived retinol has to pass through the placenta to reach the developing fetus. Despiteits apparent importance, little is known about placental metabolism of retinol, and particularly placental productionand/or secretion of active retinoids. It has been previously considered that retinoids are recruited from the uterineenvironment to influence placental development and function during gestation. We have studied retinoid metabolismin the human choriocarcinoma cell line JEG-3 and demonstrate, for the first time, that active retinoids are producedendogenously by the JEG-3 cell line from retinol. These retinoids induce gene expression from a retinoic acid-responsive enhancer element reporter plasmid and modulate placental transglutaminase activity. Furthermore,retinoids are secreted from JEG-3, as shown by the activation of retinoic acid-responsive β lacZ reporter cellsgrown in conditioned media. These results suggest that there could be an active role for trophoblast-derivedretinoids during human development.

Key words: development/ethanol/placenta/retinoic acid

IntroductionIn addition to their essential roles in vision, growth andmaintenance of differentiated epithelia, vitamin A (retinol) andits active derivatives, the retinoic acids (retinoids) are requiredfor normal mammalian reproduction and fetal development(Wilson et al., 1953; Chambon, 1994). Maternal vitamin Adeficiency (VAD) can result in fetal death or in a spectrum ofmalformations called fetal VAD syndrome (Takahashi et al.,1975; Morriss-Kay and Ward, 1999). Excessive vitamin Aintake can also produce a spectrum of congenital defects in adose- and developmental stage-dependent manner (Miller et al.,1993; Morriss-Kay and Ward, 1999). Since there is node-novo fetal synthesis of retinol, the developing mammalianembryo is dependent on the maternal circulation for its vitaminA supply. The presence of measurable hepatic vitamin A storedat birth is indicative of the functionality of placental transportduring gestation (Satre et al., 1992; Ross and Gardner, 1994).Despite its apparent importance, little is known about placentalmetabolism of retinoids. This question has been approachedin animal models (sheep, mouse, monkey), but the placental

© European Society of Human Reproduction and Embryology 485

metabolism of retinoids is difficult to deduce from studies inthe intact animal and extrapolation of these results to humansremains uncertain (Donoghue et al., 1982; Ismadi and Olson,1982; Vahlquist and Nilsson, 1984). Measurements of humanplacental, maternal and umbilical cord retinoid levels havebeen reported, but they shed little light on the metabolicpathway (Dimenstein et al., 1996; Sapin et al., 2000c). Usingfull-term human placental explants and primary cell cultures,the ability of placental tissues to produce and store retinylesters from retinol has been demonstrated (Torma andVahlquist, 1986; Sapin et al., 2000a).

Mammalian placentas are also known to express nucleartranscription factors [retinoic acid receptors (RARs) and reti-noid X receptors (RXRs)] by which retinoic acids modulatethe expression of target genes (Sapin et al., 1997; Tarradeet al., 2000). Several genes modulated by retinoids have beendescribed in human and mouse placenta; for example, chorionicgonadotrophin hormone (CGH) (Guibourdenche et al., 1998a),placental lactogen hormone (Stephanou and Handwerger,1995), leptin (Guibourdenche et al., 2000), receptor of epi-

L.Blanchon et al.

dermal growth factor (Roulier et al., 1994), 17β hydroxysteroiddehydrogenase type 1 (Peltoketo et al., 1996) and STRA(stimulated by retinoic acid) genes (Sapin et al., 2000b). Thesedata question the origin of active retinoids detected in placenta.There is no evidence for a maternal and central endocrineproduction during development and it is clear that plasmaretinoic acid could not be the source of ligand for all placentalcells (Kurlandsky et al., 1995; Bavik et al., 1997). It hasrecently been suggested that the uterine environment producesretinoic acid to influence early embryonic and extra-embryonicdifferentiation, morphogenesis and development (Zhang et al.,2000). In order to propose the placenta as an additional siteof retinoid generation during development, we investigated theability of the human choriocarcinoma cell line JEG-3 toproduce and secrete active retinoids from retinol, detected bythe activation of retinoic acid-responsive promoter. Further-more, we investigated the presence of the binding protein(cellular retinol binding protein 1, CRBP1) and the twoenzymes, alcohol dehydrogenase (ADH) and aldehyde dehy-drogenase (ALDH), all of which are involved in the cytosolicgeneration of active retinoids from retinol.

We selected the JEG-3 choriocarcinoma cell line as ourtrophoblastic model system. This cell line, derived from ahuman choriocarcinoma, appears to present many of thebiological and biochemical characteristics associated withsyncytiotrophoblasts, even though they are mononucleated andhighly proliferative (Matsuo and Strauss, 1994). The JEG-3cell line retains the capacity to produce progesterone, HCG(Chou, 1982), several steroids, other placental hormones andenzymes (Kato and Braunstein, 1991; Sun et al., 1998; Trem-blay et al., 1999); therefore, it has been proposed as a modelfor placental syncytiotrophoblasts. In addition, RARα andRXRα proteins have been detected by Western blotting andimmunocytochemistry as two majors receptors in JEG-3 cells(Guibourdenche et al., 1998b). In this way, this cell line hasbeen extensively used to investigate the regulation by retinoidsof proliferation, differentiation and secretion of placenta cells.For all these reasons, we chose JEG-3 cells in order to answerto our question focused on the placental generation of retinoicacid from retinol. As ethanol and 4-methylpyrazole (4MP) arewell known to alter the metabolism of retinoids (Xiang-Dong,1999) and particularly the conversion of retinol into retinoicacid by acting as a competitive inhibitor for cytosolic ADH(Chen et al., 1995a,b; Duester, 2000), the potential effects of4MP and ethanol on retinoid metabolism in the JEG-3 cellline were also investigated.

Materials and methods

Chemicals

All-trans retinol, all-trans retinoic acid (ATRA), trypsin, 4MP, phenyl-methylsulphonyl fluoride, aprotinin, leupeptin, pepstatin, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) andhuman purified retinol binding protein (RBP) were purchased fromSigma-Aldrich (St Quentin Fallavier, France). Horseradish peroxidase(HRP)-conjugated streptavidin, 5-(biotinamido) pentylamine and o-phenylene diamine dihydrochloride were obtained from Interchim(Montlucon, France). Culture medium [Dulbecco’s modified Eagle’s

486

medium (DMEM)], additives [glutamine, penicillin/streptomycin,dextran-coated charcoal-stripped fetal calf serum (FCS)] and lipofecta-mine were purchased from Life Technologie (Cergy Pontoise, France).For all experiments, ATRA and retinol were prepared as 1000� stocksolution in ethanol. The maximal ethanol concentration to which thecells and tissues were exposed was �0.1%. Retinol–RBP complex wasprepared and purified as previously described (Siegenthaler, 1990).

Cell culture

Human choriocarcinoma trophoblastic cells (JEG-3) and retinoidsreporter cells (F9-1.8-RARβ lacZ cells) (Sonneveld et al., 1999)were maintained in DMEM supplemented with 2 mmol/l glutamine,10 units/ml penicillin, 100 µg/ml streptomycin and 10% dextran-coated charcoal-stripped FCS (added with 0.2 mg/ml G418 forselective medium of F9-1.8-RARβ lacZ cells). No changes concerningthe cellular growth and HCG production of the JEG-3 cell line wereobserved during 15 days treatment with dextran-coated charcoal-stripped FCS. All experiments were performed in a humidifiedincubator with 5% CO2, 95% air at 37°C.

Immunocytochemistry

Antibodies against ADH and ALDH were generated using a specificantigenic amino peptide sequence conserved in all the members ofeach enzyme (Bavik et al., 1997). The antibody against CRBP1 wasgenerated as previously described (Bavik et al., 1997). JEG-3 cellswere grown on glass coverslips and fixed in paraformaldehyde at 4%[v/v in phosphate buffered saline (PBS)] and permeabilized usingacetone. Specific cellular protein expression was checked using theprimary polyclonal antibodies and a secondary antibody conjugatedwith HRP. HRP activity was demonstrated by using diaminobenzidineand H2O2 as substrate to reveal the immunoreactions.

Transfection of cultured cells

Plasmid DR5-tk-CAT contains one copy of the retinoic acid-responsiveelement DR5 (direct repeat 5) ligated to an herpes simplex thymidinekinase promoter upstream of a chimeric chloramphenicol acetyltransferase (CAT) reporter gene (Mader et al., 1993). JEG-3 cellswere trypsinized 16 h prior to transfection. A total of 5�105 cells in30 mm dishes were transfected using lipofectamine with 1 µg ofreporter DR5-tk-CAT plasmid and 1 µg of cytomegalovirus (CMV)–luciferase vector serving as internal control to normalize variationsin transfection efficiency. The CMV-luciferase plasmid (Pharmacia,St Quentin en Yvelines, France) contains a CMV promoter andenhancer sequences that drive a luciferase (LUC) gene. Paralleltransfections of the corresponding empty vectors CAT and LUC atequivalent concentrations were performed in all experiments. Afteran overnight incubation with DNA, cells were washed and incubatedwith fresh medium for an additional 12 h period. They were treatedfor another 24 h with retinoids in 1000� stock solutions in ethanol.In this case, the maximal ethanol concentration to which the cellswere exposed was �0.1%. In other experiments, 4MP and ethanolwere added at concentrations of 2 and 100 mmol/l respectively. Cellviability assays were performed for each treatment (4MP and ethanol)using MTT assays, as previously described (Godichaud et al., 2000).For this purpose, cells were seeded into 30 mm dishes at 5�104

cells/well and treated for 48 h with 4MP or ethanol at concentrationsranging from 1–10 and 10–200 mmol/l respectively. During this 48 hperiod, a second dose of 4MP or ethanol was added after 24 h ofincubation.

CAT and luciferase reporter gene assay

After incubation, cells were washed twice with PBS and treated with700 µl of cell lysis buffer for 1 h at 4°C. The lysed cells were scraped

Retinoid production and secretion by JEG-3

off and centrifuged at 950 g for 5 min at room temperature. CATwas measured by an immunoenzymatic assay (Roche Diagnostics,Meylan, France) on 100 µl of supernatant. In all studies, CAT assayswere normalized to luciferase activity, determined according to themanufacturer’s instructions. The supernatants were incubated withluciferase assay reagent based on an original protocol (De Wet et al.,1987) (Prodemat, Anduze, France). The number of relative light unitswas determined with a 3 s delay after a 10 s incubation.

In-situ tissue transglutaminase (tTG) assay

For in-situ tTG activity measurements, JEG-3 (5�105 cells in 30 mmdishes) was pre-incubated with 5-(biotinamido) pentylamine andincorporation of this reagent into synthesized proteins was determinedusing a streptavidin–peroxidase assay following a previously describedprocedure (Zhang et al., 1998). The results obtained by this methodare well correlated with those issued from the classical determinationof tTG by using an in-vitro putrescine incorporation assay (Perryet al., 1995; Zhang et al., 1998). The activity of tTG measured insitu was calculated as a percentage of basal activity and normalizedto cellular protein concentrations. Protein concentrations of thehomogenates were determined using the Biuret method (RocheDiagnostics) (Camara et al., 1991).

Retinoid activity assay in conditioned medium

JEG-3 cells (5�105 cells in 30 mm dishes) incubated overnight withretinol (with or without ethanol or 4MP) and the conditioned mediawere placed on the F9-1.8-RARβ lacZ cells to test the ability of themedia to activate the lacZ gene which is under the control of aretinoic acid-responsive promoter, as previously described (Sonneveldet al., 1999). β-galactosidase was determined by using immunoenzym-atic assay (Roche Diagnostics, Meylan, France). Normalization ofβ-galactosidase activity was performed with cellular protein concen-trations, measured as described above.

Statistical methods

Results expressed as mean � SD were an average value from ninevalues per each condition. Comparison of means was done by analysisof variance and Fisher’s t-test using the Statview II 1.03 software(Abacus Concepts Inc., Berkeley, CA, USA). For all the studies,values were considered significantly different when P � 0.05.

ResultsIn order to characterize the presence of proteins involved inthe metabolic generation of retinoic acid, we first showedthat JEG-3 cells presented cytoplasmic immunoreactivity forCRBP1 (Figure 1B), ADH (Figure 1C) and ALDH (Figure 1D).

The production of active retinoids by JEG-3 cells afterincubation with retinol was studied by using transient transfec-tions of a retinoid-sensitive reporter gene based on CATexpression (see Figure 2). First, we checked that the naıveJEG-3 cells (or JEG-3 cells transfected with an empty vector)did not produce CAT. CAT production by cells transfectedwith the DR5-tk-CAT plasmid and not treated was consideredas the basal level for comparison with and detection ofinduction recorded following exposure to retinoids.

ATRA treatment (1 µmol/l) for 24 h increased CAT response(6.5 � 0.5-fold induction). After retinol (1 µmol/l) treatment,CAT production was significantly enhanced (3.7 � 0.4-foldinduction) but not as much as with stimulation with 1 µmol/lof retinoic acid (Figure 2). CAT production (reflecting active

487

retinoid generation) was measured with different retinol con-centrations ranging from 10–11 to 2�10–6 mol/l. Retinol treat-ment increased CAT generation in a dose-dependent manner(Figure 3). The maximal CAT induction was obtained with1 µmol/l of retinol and the half maximal induction [EC50] ofCAT was experimentally deduced to be 40 nmol/l of retinol(versus 5 nmol/l for retinoic acid), which is a classical valuealready described in enzymatic retinoic acid generation fromretinol. CAT induction obtained with 1 µmol/l of retinol wassimilar to that which occurred with 0.01 µmol/l of retinoicacid (see Figure 2). In addition, we found that JEG-3 cellsgenerate only the all-trans isomer of the retinoic acid fromretinol (unpublished data obtained using high-performanceliquid chromatography experiments).

To confirm that active retinoids were generated by enzymaticconversion from retinol, we tested potential alterations of thisreaction by two well-known inhibitors of cytosolic ADH (Chenet al., 1995a,b), 4MP and ethanol, as presented in Figure 2.We first evaluated JEG-3 cells viability (using MTT assays)when they were exposed to 4MP and ethanol at usual and welldescribed experimental concentrations. The average values ofcells grown in regular medium was considered as 100%viability; no toxicity �10% was observed for 4MP concentra-tions between 0.5 and 10 mmol/l and for ethanol treatmentsbetween 10 and 200 mmol/l. 4MP (2 mmol/l) caused astatistically significant decrease in retinoic acid production byJEG-3 cells, causing only 54.1% of the induction observed incells incubated with 2 µmol/l retinol alone (Figure 2). Cellstreated with 100 mmol/l of ethanol also produced significantlylower amounts of retinoic acid, producing only 53.7% ofretinol-treated cells. The absolute values of internal standard(not dependant upon the enzymatic conversion caused byADH) did not change in response to 4MP and ethanoltreatments (data not shown), indicating the specific effect of4MP and ethanol on this inhibition of retinoic acid generation.

In-situ tTG is an enzymatic activity expressed by placentaand modulated by ATRA (Piacentini et al., 1992; Hageret al., 1997). In-situ tTG activity was assayed to confirm thegeneration of active retinoids from retinol in JEG-3 cells byusing an alternative, complementary and more physiologicalmethod than the transient transfection of a reporter plasmid.As presented in Figure 4, we first demonstrated that retinoicacid repressed tTG activity in the JEG-3 cell line. Cells treatedwith 4MP (2 mmol/l) or ethanol (100 mmol/l) expressed thesame tTG activity as the control cells, demonstrating thatneither compounds modulated tTG activity in JEG-3 cells(Figure 4). Retinoic acid- (1 µmol/l) and retinol- (1 µmol/l)treated cells had statistically lower tTG activity than controlcells (38 and 54% of basal activity respectively). Retinoicacid- (1 µmol/l) treated cells possessed lower tTG activitythan retinol- (1 µmol/l) treated cells (Figure 4). When retinol-treated JEG-3 cells were incubated with 4MP or ethanol, tTGactivity was not statistically different from that observed incontrol cells (Figure 4), suggesting an absence of retinoic acidgeneration. As with the CAT assays, the specificity of thisalteration by ethanol and 4MP was assessed by the absenceof alterations of absolute values on internal standards (concen-tration of total cellular proteins).

L.Blanchon et al.

Figure 1. Immunolabelling of JEG-3 cells. JEG-3 cells were seeded at 1500 cells/cm2 and checked for expression of proteins involved inthe retinoic acid pathway: (B) CRBP 1, (C) ADH, (D) ALDH. (A) JEG-3 cells without primary antibody (negative control). Originalmagnifications �1000.

To establish the trophoblastic JEG-3 cells as a source ofactive retinoid secretion, we used conditioned medium assaysand the F9-1.8-RARβ2 lacZ cell line as a reporter system(Sonneveld et al., 1999). Thus, retinoids present in conditionedmedium will activate expression of the lacZ gene when placedon a confluent monolayer of our reporter cells. First, wedemonstrated that in our short time of incubation (applied forall experiments using this cell line), retinol did not induce β-galactosidase production in F9-1.8-RARβ2 as the value wasnot statistically different from reporter cells incubated withethanol of 4MP (Figure 5). Conditioned medium of JEG-3 afterincubation with 1 µmol/l of retinol activated β-galactosidasegeneration, indicating secretion of active retinoids by JEG-3(Figure 5). A decrease in β-galactosidase production wasobserved with conditioned medium from JEG-3 cells incubatedwith retinol and the addition of 4MP (2 µmol/l) or ethanol(100 mmol/l). β-galactosidase induction by retinoic acid wasnot blocked by the direct addition of 4MP (2 µmol/l) or ethanol(100 mmol/l) on F9-1.8-RARβ2 reporter cells (data not shown).These two combined results strongly suggest that the decreasein reporter cell activation by retinol-conditioned mediumsupplemented with 4MP and ethanol was due to a reductionof active production and/or secretion of retinoic acid byJEG-3 cells.

488

DiscussionVitamin A and its active derivatives, the retinoids, are funda-mental for development of the mammalian fetus (Chambon,1994; Ross and Gardner, 1994; Morriss-Kay and Ward, 1999).The placenta plays a crucial role in the regulation of transportand metabolism of maternal nutrients transferred to the fetus.Abnormalities in these placental functions could have deleteri-ous consequences for fetal development (Miller et al., 1993;Rossant and Cross, 2001). Little is known about humanplacental retinol transfer and metabolism. We previously estab-lished that human placental tissues—more precisely, the villousmesenchymal fibroblasts—are able to esterify retinol (Sapinet al., 2000a). Creech Kraft et al. have demonstrated that theearly human placenta is able to metabolise 13-cis retinoicacid (Creech Kraft et al., 1989). More recently, it has beenestablished that the all-trans retinyl acetate oxidating lipoxyg-enase and a specific cytochrome p450 implicated in retinoidmetabolism (CYP26) are present in the human placenta (Dattaand Kulkarni, 1996; Trofimova-Griffin and Juchau, 1998).Together, these results indicate that the human placenta hasthe ability to esterify, isomerize and oxidize retinoids. Never-theless, the generation and secretion of active retinoids fromretinol by human placental cells was never investigated. Toelucidate this mechanism, we chose JEG-3 choriocarcinoma

Retinoid production and secretion by JEG-3

Figure 2. Activation of a retinoic acid-responsive reporter gene by retinol in JEG-3 cells is inhibited by 4-methylpyrazole and ethanol.JEG-3 cells were transiently transfected with a retinoic acid-responsive CAT reporter gene (DR5-tk-CAT). The induction of CAT isexpressed as the ratio between treated cells [retinol (Rol), retinoic acid (RA), 4-methylpyrazole (4MP) or ethanol (EtOH)] and control cells(i.e. transfected with plasmid and not treated). CAT production was determined after 24 h incubation in different conditions and normalizedto luciferase activity. Each value represents the mean � SD of the nine experiments. The different superscripts (a–f) indicate statisticaldifferences between the different incubation conditions tested (P � 0.05).

Figure 3. Dose-dependent activation by all-trans retinol of aretinoic acid-responsive reporter in JEG-3 cells. JEG-3 cells weretransiently transfected with a retinoic acid-responsive CAT reportergene (DR5-tk-CAT). The induction of CAT is expressed as the ratiobetween (retinol) treated cells and control cells. CAT productionwas determined after 24 h incubation. Each value represents themean � SD of CAT activity normalized to luciferase activity fromfive experiments.

cells as a trophoblastic-like model for the investigation ofplacental retinoid metabolism.

Proteins implicated in the generation of retinoic acid fromretinol (such as CRBP1, ADH and ALDH) were shown to bepresent in JEG-3 cells. The presence of CRBP1 was strongly

489

expected because extracts from total term human placenta arealready known to express this intracellular protein which playsa pivotal role in intracellular retinol metabolism (Okuno et al.,1987; Johansson et al., 1999). For the first time, ADH andALDH were detected in JEG-3 cells. Our results corroboratea recent study suggesting the presence of a 9-cis retinolconversion to 9-cis retinoic acid in human tissues such placenta(Paik et al., 2000). Because of the predominant presence ofRXRα in mediating the biological effects of retinoids onJEG-3 cells, it has also been suggested that a possibleinduction of retinoic acid generation itself is a proper metabolicpathway (Guibourdenche et al., 1998b). Our present workshows, for the first time, that derived trophoblastic cells areable to produce and secrete retinoids at functional levels fromretinol. Two mammalian placentas have also been shown toproduce retinoic acid from retinol: the porcine (Parrow et al.,1998) and the mouse (yolk sac) placenta (Bavik et al., 1997).Due to the metabolic property demonstrated in this study, theJEG-3 cell line belongs to a little group of transformed andestablished cell lines experimentally shown to be able togenerate active retinoids from retinol; this group includes thepig kidney cell line LLC-PK1 (Napoli, 1986), the rat hepaticstellate cell line HSCT6 (Vogel et al., 2001) and the humanintestinal cell line Caco-2 (Lampen et al., 2000). For eachdetermination, induction obtained with retinol was alwaysless than with all-trans retinoic acid exposure. The enzymesinvolved in retinol metabolism, such as ADH and ALDH, aresaturable (Leo et al., 1987; Lindahl, 1992; Duester, 2000).Moreover, intracellular retinol can be metabolized followingdifferent pathways (oxidation but also decarboxylation,

L.Blanchon et al.

Figure 4. Decrease of tissue trans-glutaminase (tTG) activity by retinol in JEG-3 cells is inhibited by 4-methylpyrazole and ethanol. In-situtTG activity was assayed after 24 h incubation in the described conditions and was normalized by intracellular protein concentrations. Foreach parameter, the value represents the mean � SD from nine experiments. Different superscripts (a–d) indicate statistical differencesbetween tissue transglutaminase activity values obtained from control cells and from treated cells (P � 0.05).

Figure 5. JEG-3 cells are able to secrete functional retinoids from retinol. JEG-3 cells were incubated overnight with retinol (added or notwith ethanol or 4-methylpyrazole) and the conditioned media (JEG-3CM) were placed on the F9-1.8-RARβ lacZ cells to test its ability toactivate the lacZ gene. β-galactosidase was determined by using immunoenzymatic assay and normalized with cellular proteinconcentrations. For each parameter, the mean � SD from nine experiments are presented. Different superscripts (a–d) indicate statisticaldifferences between β-galactosidase activity values obtained from control and treated cells (P � 0.05).

conjugation or epoxydation). Together, these two reasons couldexplain why retinol is only partially used to form retinoic acid.

What could be the biological function of active retinoidsformed and secreted in this derived trophoblastic cell line,JEG-3? In-situ retinoid production is a single and efficient way

490

to regulate cellular proliferation, differentiation and adhesion. Ithas been well described that retinoids stimulate differentiationof JEG-3 cells (indicated by the rate of HCG secretion) butconsistently reduce, in a dose- and time-dependent manner, theproliferation and the adhesion of these cells to the endometrial

Retinoid production and secretion by JEG-3

uterine monolayer (Yamada et al., 1997; Hohn et al., 2000).In addition, regulation of tTG by retinoic acid in JEG-3 cellscould also participate in the regulation of apoptosis, a cellularevent indispensable for the harmonious development of thiscell line (Hager et al., 1997; Autuori et al., 1998). In-situgenerated retinoids could also be implicated in the implantationprocess. A previous study has demonstrated that RARα/RXRα(a functional heterodimer of nuclear receptors) is present inthe JEG-3 cell line, in the proliferative intermediate tropho-blasts and in the invasive extravillous trophoblasts duringinvasion process (Tarrade et al., 2000). Upon retinoic acidtreatment, a rat choriocarcinoma cell line was shown toirreversibly lose connexin31 gene expression, a moleculestrongly implicated in the implantation process (Grummeret al., 1996). Other placental events could also be regulatedby retinoids in the JEG-3 cell line via the modulation ofexpression of retinoic acid-sensitive genes (in terms of tran-scripts and proteins). These could include the following (i)hormone production: total HCG and its subunit α and β (Katoand Braunstein, 1991; Matsuo and Strauss, 1994; Yamadaet al., 1997); (ii) metabolism and transfer: an increase ingp330/megalin (a membrane-bound 550 kD calcium bindingglycoprotein belonging to the low density lipoprotein receptorfamily) mRNA expression has been seen in JEG-3 cells culturedwith retinoids (Liu et al., 1998); and (iii) steroidogenesis:progesterone production by the JEG-3 cell line is stimulatedby retinoic acid, without changes of intracellular levels ofcAMP (Kato and Braunstein, 1991). In addition, key enzymesof steroidogenesis are regulated by retinoic acid in the JEG-3cell line. These include 11β-hydroxysteroid dehydrogenasetype 2 mRNA expression and activity (Tremblay et al., 1999),17β-hydroxysteroid dehydrogenase type 1 mRNA expressionand activity (Piao et al., 1997) and CYP19 (p450 aromatase)regulated by a region upstream of exon 1 (Sun et al., 1998).

Extrapolating our findings on choriocarcinoma cells tonormal trophoblast function, our data point to a possiblelinkage between the nutrient supply of retinol to the placenta,and the generation of strong developmental morphogene andgene regulation and placental gene regulation and physiology.During pregnancy, placental cells may be exposed to deleteriousmaternal conditions including alcohol abuse. Links have beenestablished between alcohol abuse, fetal malformations [thefetal alcoholic syndrome (FAS)] and alterations of retinoidmetabolism (Zachman and Grummer, 1998; Leo and Lieber,1999). Previous studies and our present work have demon-strated the interference of alcohol on synthesis of functionalretinoids from retinol (Xiang-Dong, 1999). We suggest thatalterations of placental retinoid metabolism by maternal ethanolingestion could provide a novel and additional explanation forthe genesis of FAS pathology, by alteration of placentalphysiology. A molecular explanation due to the heterodimeriza-tion of RXR with peroxisome proliferator activated receptor(PPAR)γ could be also proposed in the same way. In fact, arecent study demonstrated that PPARγ expression and activityis indispensable for harmonious placentation (Barak et al.,1999). Alteration of retinoic acid generation from retinol inplacental cells could consequently alter the functionality ofthe PPARγ/RXR heterodimer, indispensable for trophoblastic

491

invasion (Tarrade et al., 2001) and JEG-3 physiology (Matsuoand Strauss, 1994). It is well established that peroxisomeproliferators and retinoids differentially regulate JEG-3 cellendocrine activities, suggesting that JEG-3 cells possess mech-anisms to respond to nutrient cues (active retinoids and/orperoxisome proliferators) using the activation of the PPARγ/RXR heterodimer, and this could be altered by toxins suchas ethanol.

In conclusion, our study establishes that a human-derivedtrophoblastic cell line produces and secretes active retinoidsfrom retinol. It is well established that retinoids regulateproliferation, differentiation and target gene expression introphoblastic and uterine cells. These results suggest thatendogenous and trophoblast-derived retinoids could have activeroles in human early development. By blocking this metabolicpathway, maternal ethanol consumption may have deleteriouseffects during placental implantation and development.

AcknowledgementsThe authors want to thank Jean-Marc Lobaccaro for critically readingthe manuscript, and all the midwives for their professional competenceand kindness, which have made this study possible. All thanks totechnicians of the Laboratories of Biochemistry and Immunology ofCentre Hospitalier Universitaire (Clermont-Ferrand) for their kindhelp, and to Dr Paul Van der Saag for the kind gift of the F9-1.8-RARβ lacZ cells. L.B. was supported by a grant from Ministerede l’Education Nationale, de la Recherche et de la Technologie(MENRT, France).

ReferencesAutuori, F., Farrace, M.G., Oliverio, S., Piredda, L. and Piacentini, M. (1998)

‘Tissue’ trans-glutaminase and apoptosis. Adv. Biochem. Eng. Biotechnol.,62, 129–136.

Barak, Y., Nelson, M.C., Ong, E.S., Jones, Y.Z., Ruiz-Lozano, P., Chien, K.R.,Koder, A. and Evans, R.M. (1999) PPAR gamma is required for placental,cardiac, and adipose tissue development. Mol. Cell., 4, 585–589.

Bavik, C., Ward, S.J. and Ong, D.E. (1997) Identification of a mechanism tolocalize generation of retinoic acid in rat embryos. Mech. Dev., 69, 155–167.

Camara, P.D., Wright, C., Dextraze, P. and Griffiths, W.C. (1991) Comparisonof a commercial method for total protein with a candidate reference method.Ann. Clin. Lab Sci., 21, 335–339.

Chambon, P. (1994) The retinoid signalling pathway: molecular and geneticanalyses. Sem. Cell. Biol., 5, 115–125.

Chen, H., Namkung, M.J. and Juchau, M.R. (1995a) Biotransformation of all-trans retinol and all-trans retinal to all-trans retinoic acid in rat conceptalhomogenated. Biochem. Pharmacol., 50, 1257–1264.

Chen, W.J., McAlhany, R.E. and West, J.R. (1995b) 4-methylpyrazole, analcohol dehydrogenase inhibitor, exacerbates alcohol-inducedmicroencephaly during the brain growth spurt. Alcohol, 12, 351–355.

Chou, J.Y. (1982) Effects of retinoic acid on differentiation of choriocarcinomacells in vitro. J. Clin. Endocrinol. Metab., 54, 1174–1180.

Creech Kraft, J., Lofberg, B., Chahoud, I., Bochert, G. and Nau, H. (1989)Teratogenicity and placental transfer of all-trans-, 13-cis-, 4-oxo-all-trans-,and 4-oxo-13-cis-retinoic acid after administration of a low oral dose duringorganogenesis in mice. Toxicol. Appl. Pharmacol., 100, 162–176.

Datta, K. and Kulkarni, A.P. (1996) Co-oxidation of all-trans retinol acetateby human term placental lipoxygenase and soybean lipoxygenase. Reprod.Toxicol., 10, 105–112.

De Wet, J.R., Wood, K.V., DeLuca, M., Helinski, D.R. and Subramani, S.(1987) Firefly luciferase gene: structure and expression in mammalian cells.Mol. Cell. Biol., 7, 725–737.

Dimenstein, R., Trugo, N.M., Donangelo, C.M., Trugo, L.C. and Anastacio,A.S. (1996) Effect of subadequate maternal vitamin-A status on placentaltransfer of retinol and beta-carotene to the human fetus. Biol. Neonate, 69,230–234.

L.Blanchon et al.

Donoghue, S., Richardson, D.W., Sklan, D. and Kronfeld, D.S. (1982) Placentaltransport of retinol in sheep. J. Nutr., 112, 2197–2203.

Duester, G. (2000) Families of retinoid dehydrogenases regulating vitamin Afunction: production of visual pigment and retinoic acid. Eur. J. Biochem.,267, 4315–4324.

Godichaud, S., Krisa, S., Couronne, B. Dubuisson, L., Merillon, J.M.,Desmouliere, A. and Rosenbaum, J. (2000) Deactivation of cultured humanliver myofibroblasts by trans-resveratrol, a grapevine-derived polyphenol.Hepatology, 31, 922–931.

Grummer, R., Hellmann, P., Traub, O., Soares, M.J., El-Sabban, M.E. andWinterhager, E. (1996) Regulation of connexin31 gene expression uponretinoic acid treatment in rat choriocarcinoma cells. Exp. Cell Res., 227,23–32.

Guibourdenche, J., Alsat, E., Soncin, F., Rochette-Egly, C. and Evain-Brion, D. (1998a) Retinoid receptors expression in human term placenta:involvement of RXR alpha in retinoid induced-hCG secretion. J. Clin.Endocrinol. Metab., 83, 1384–1387.

Guibourdenche, J., Roulier, S., Rochette-Egly, C. and Evain-Brion, D. (1998b)High retinoid X receptor expression in JEG-3 choriocarcinoma cells:involvement in cell function modulation by retinoids. J. Cell. Physiol.,176, 595–601.

Guibourdenche, J., Tarrade, A., Laurendeau, I., Rochette-Egly, C., Chambon,P., Vidaud, M. and Evain-Brion, D. (2000) Retinoids stimulate leptinsynthesis and secretion in human syncytiotrophoblast. J. Clin. Endocrinol.Metab., 85, 2550–2555.

Hager, H., Gliemann, J., Hamilton-Dutoit, S., Ebbesen, P., Koppelhus, U. andJensen, P.H. (1997) Developmental regulation of tissue transglutaminaseduring human placentation and expression in neoplastic trophoblast.J. Pathol., 181, 106–110.

Hohn, H.P., Linke, M. and Denker, H.W. (2000) Adhesion of trophoblast touterine epithelium as related to the state of trophoblast differentiation:in vitro studies using cell lines. Mol. Reprod. Dev., 57, 135–145.

Ismadi, S.D. and Olson, J.A. (1982) Dynamics of the fetal distribution andtransfer of vitamin A between rat fetuses and their mother. Int. J. Vit. Nutr.Res., 52, 112–119.

Johansson, S., Gustafson, A.L., Donovan, M., Eriksson, U. and Denker, L.(1999) Retinoid binding proteins—expression patterns in the humanplacenta. Placenta, 20, 459–465.

Kato, Y. and Braunstein, G.D. (1991) Retinoic acid stimulates placentalhormone secretion by choriocarcinoma cell lines in vitro. Endocrinology,128, 401–407.

Kurlandsky, S.B., Gamble, M.V., Ramakrishnan, R. and Blaner W.S. (1995)Plasma delivery of retinoic acid to tissues in the rat. J. Biol. Chem., 270,17850–17857.

Lampen, A., Meyer, S., Arnhold, T. and Nau, H. (2000) Metabolism ofvitamin A and its active metabolite all-trans-retinoic acid in small intestinalenterocytes. J. Pharmacol. Exp. Ther., 295, 979–985.

Leo, M.A. and Lieber, C.S. (1999) Alcohol, vitamin A, and beta-carotene:adverse interactions, including hepatotoxicity and carcinogenicity. Am. J.Clin. Nutr., 69, 1071–1085.

Leo, M.A., Kim, C.I. and Lieber, C.S. (1987) NAD�-dependant retinoldehydrogenase in liver microsomes. Arch. Biochem. Biophys., 259, 241–249.

Lindahl, R. (1992) Aldehyde dehydrogenases and their role in carcinogenesis.Crit. Rev. Biochem. Mol. Biol., 27, 283–335.

Liu, W., Yu, W.R., Carling, T., Juhlin, C., Rastad, J., Ridefelt, P., Akerstrom,G. and Hellman P. (1998) Regulation of gp330/megalin expression byvitamins A and D. Eur. J. Clin. Invest., 28, 100–107.

Mader, S., Chen, J.Y., Chen, Z, White, J., Chambon, P. and Gronemeyer, H.(1993) The patterns of binding of RAR, RXR and TR homo- andheterodimers to direct repeats are dictated by the binding specificites ofthe DNA binding domains. EMBO J., 12, 5029–5041.

Matsuo, H. and Strauss, J.F. 3rd (1994) Peroxisome proliferators and retinoidsaffect JEG-3 choriocarcinoma cell function. Endocrinology, 135, 1135–1145.

Miller, R.K., Faber, W., Asai, M., D’Gregorio, R.P., Ng, W.W., Shah, Y. andNeth-Jessee, L. (1993) The role of the human placenta in embryonicnutrition. Impact of environmental and social factors. Ann. NY Acad. Sci.,678, 92–107.

Morriss-Kay, G.M. and Ward, S.J. (1999) Retinoids and embryonicdevelopment. Int. Rev. Cytol., 188, 73–133.

Napoli, J.L. (1986) Retinol metabolism in LLC-PK1 cells. Characterizationof retinoic acid synthesis by an established mammalian cell line. J. Biol.Chem., 261, 13592–13597.

492

Okuno, M., Kato, M., Moriwaki, H., Kanai, M. and Muto, Y. (1987) Purificationand partial characterization of cellular retinoic acid-binding protein fromhuman placenta. Biochem. Biophys. Acta, 923, 116–124.

Paik, J., Vogel, S., Piantedosi, R., Sykes, A., Blaner, W.S. and Swisshelm, K.(2000) 9-cis-retinoids: biosynthesis of 9-cis-retinoic acid. Biochemistry, 39,8073–8084.

Parrow, V., Horton, C., Maden, M., Laurie, S. and Notarianni, E. (1998)Retinoids are endogenous to the porcine blastocyst and secreted bytrophectoderm cells at functionally-active levels. Int. J. Dev. Biol., 42,629–632.

Peltoketo, H., Isomaa, V., Poutanen, M. and Vihko, R. (1996) Expression andregulation of 17 beta-hydroxysteroid dehydrogenase type 1. J. Endocrinol.,150, S21–S30.

Perry, M.J., Mahoney, S.A. and Haynes, L.W. (1995) Transglutaminase C incerebellar granule neurons: regulation and localization of substrate cross-linking. Neuroscience, 65, 1063–1067.

Piacentini, M., Ceru, M.P., Dini, L., DiRao, M., Piredda, L., Thomazy, V.,Davies, P.J. and Fesus, L. (1992) In vivo and in vitro induction of ‘tissue’transglutaminase in rat hepatocytes by retinoic acid. Biochem. Biophys.Acta, 1135, 171–179.

Piao, Y.S., Peltoketo, H., Jouppila, A. and Vihko, R. (1997) Retinoic acidsincrease 17 beta-hydroxysteroid dehydrogenase type 1 expression in JEG-3 and T47D cells, but the stimulation is potentiated by epidermal growthfactor, 12-O-tetradecanoylphorbol-13-acetate, and cyclic adenosine 3�,5�-monophosphate only in JEG-3 cells. Endocrinology, 138, 898–904.

Ross, A.C. and Gardner, E.M. (1994) The function of vitamin A in cellulargrowth and differentiation, and its roles during pregnancy and lactation.Adv. Exp. Med. Biol., 352, 187–200.

Rossant, J. and Cross, J.C. (2001) Placental development: lessons from mousemutants. Nat. Rev. Genet., 2, 538–548.

Roulier, S., Rochette-Egly, C., Rebut-Bonneton, C., Porquet, D. and Evain-Brion, D. (1994) Nuclear retinoic acid receptor characterization in culturedhuman trophoblast cells: effect of retinoic acid on epidermal growth factorreceptor expression. Mol. Cell. Endocrinol., 105, 165–173.

Sapin, V., Ward, S.J., Bronner, S., Chambon, P. and Dolle, P. (1997) Differentialexpression of transcripts encoding retinoid binding proteins and retinoicacid receptors during placentation of the mouse. Dev. Dyn., 208, 199–210.

Sapin, V., Chaib, S., Blanchon, L., Alexandre-Gouabau, M.C., Lemery, D.,Charbonne, F., Gallot, D., Jacquetin, B., Dastugue, B. and Azaıs-Braesco,V. (2000a) Esterification of vitamin A by the human placenta involvesvillous mesenchymal fibroblasts. Pediatr. Res., 48, 565–572.

Sapin, V., Bouillet, P., Oulad-Abdelghani, M., Dastugue, B., Chambon, P. andDolle, P. (2000b) Differential expression of retinoic acid-inducible (Stra)genes during mouse placentation. Mech. Dev., 92, 295–299.

Sapin, V., Alexandre-Gouabau, M.C., Chaib, S., Bournazeau, J.A., Sauvant, P.,Borel, P., Jacquetin, B., Grolier, P., Lemery, D., Dastugue, B. et al. (2000c)Effect of vitamin A status at the end of term pregnancy on the saturation ofretinol-binding protein saturation with retinol. Am J. Clin. Nutr., 71, 537–543.

Satre, M.A., Ugen, K.E. and Kochhar, D.M. (1992) Developmental changesin endogenous retinoids during pregnancy and embryogenesis in the mouse.Biol. Reprod., 46, 802–810.

Siegenthaler, G. (1990) Gel electrophoresis of Cellular Retinoic Acid BindingProteins, Cellular Retinol Binding Proteins and serum Retinol BindingProtein. Methods Enzymol., 189, 299–307.

Sonneveld, E., Van den Brink, C.E., Van der Leede, B.J., Maden, M. and Vander Saag, P.T. (1999) Embryonal carcinoma cell lines stably transfectedwith mRARbeta2-lacZ: sensitive system for measuring levels of activeretinoids. Exp. Cell Res., 250, 284–297.

Stephanou, A. and Handwerger, S. (1995) Retinoic acid and thyroid hormoneregulate placental lactogen expression in human trophoblast cells.Endocrinology, 136, 933–938.

Sun, T., Zhao, Y., Mangelsdorf, D.J. and Simpson, E.R. (1998) Characterizationof a region upstream of exon I.1 of the human CYP19 (aromatase) genethat mediates regulation by retinoids in human choriocarcinoma cells.Endocrinology, 139, 1684–1691.

Takahashi, Y.I., Smith, J.E., Winick, M. and Goodman, D.S. (1975) VitaminA deficiency and fetal growth and development in the rat. Am. J. Physiol.,233, 263–271.

Tarrade, A., Rochette-Egly, C., Guibourdenche, J. and Evain-Brion, D. (2000)The expression of nuclear retinoid receptors in human implantation.Placenta, 21, 703–710.

Tarrade, A., Schoonjans, K., Pavan, L., Auwerx, J., Rochette-Egly, C., Evain-Brion, D. and Fournier, T. (2001) Ppar gamma/rxr alpha heterodimerscontrol human trophoblast invasion. J. Clin. Endocrinol. Metab., 86,5017–5024.

Retinoid production and secretion by JEG-3

Torma, H. and Vahlquist, A. (1986) Uptake of vitamin A and retinol-bindingprotein by human placenta in vitro. Placenta, 7, 295–305.

Tremblay, J., Hardy D.B., Pereira, L.E. and Yang, K. (1999) Retinoicacid stimulates the expression of 11beta-hydroxysteroid dehydrogenasetype 2 in human choriocarcinoma JEG-3 cells. Biol. Reprod., 60,541–545.

Trofimova-Griffin, M.E. and Juchau, M.R. (1998) Expression of cytochromeP450RAI (CYP26) in human fetal hepatic and cephalic tissues. Biochem.Biophys. Res. Commun., 252, 487–491.

Vahlquist, A. and Nilsson, S. (1984) Vitamin A transfer to the fetus and tothe amniotic fluid in rhesus monkey (Macaca mulatta). Ann. Nutr. Metab.,28, 321–333.

Vogel, S., Piantedosi, R., Frank, J., Lalazar, A., Rockey, D.C., Friedman, S.L.and Blaner, W.S. (2001) An immortalized rat liver stellate cell line (HSC-T6): a new cell model for the study of retinoid metabolism in vitro. J. LipidRes., 41, 882–893.

493

Wilson, J.G., Roth, C.B. and Warkany, J. (1953) An analysis of the syndromeof malformations induced by maternal A deficiency. Effects of restorationof vitamin A at various times in gestation. Am. J. Anat., 92, 189–217.

Xiang-Dong, W. (1999) Chronic alcohol intake interferes with retinoidmetabolism and signaling. Nutr. Rev., 57, 51–59.

Yamada, S., Okamoto, T., Nakanishi, T., Nomura, S. and Tomoda, Y. (1997)Effects of all-trans retinoic acid on choriocarcinoma cells in vitro. J. Obstet.Gynaecol. Res., 23, 125–132.

Zachman, R.D. and Grummer, M.A. (1998) The interaction of ethanol andvitamin A as a potential mechanism for the pathogenesis of Fetal AlcoholSyndrome. Alcohol Clin. Exp. Res., 22, 1544–1556.

Zhang, J., Lesort, M., Guttmann, R.P. and Johnson, G.V. (1998) Modulationof the in situ activity of tissue transglutaminase by calcium and GTP.J. Biol. Chem., 273, 2288–2295.

Submitted on October 26, 2001; accepted on February 8, 2002