Ascorbic Acid Transported by Sodium-Dependent Vitamin C Transporter 2 Stimulates Steroidogenesis in Human Choriocarcinoma Cells

Reduced vitamin C [ascorbic acid (AA)], which is taken up into cells by sodium-dependent vitamin C transporter (SVCT) 1 and 2, is believed to be important for hormone synthesis, but its role in generating placental steroids needed to maintain pregnancy and fetal development is not clear. To determine the steroidogenic effect of AA and the role of SVCT2 in AA-induced steroidogenesis, we tested the effects of AA treatment and SVCT2 knockdown on steroidogenesis in human choriocarcinoma cell lines. AA treatment of JEG-3, BeWo, and JAR cells for 48-h dose dependently increased progesterone and estradiol levels. In JEG-3 cells, AA increased the mRNA expression of P450 cholesterol side-chain cleavage enzyme, 3β-hydroxysteroid dehydrogenase type 1, and aromatase, key enzymes for steroidogenesis. Stable knockdown of SVCT2 in JEG-3 cells by retrovirally mediated RNA interference decreased the maximal velocity of AA uptake by approximately 50%, but apparent affinity values were not affected. SVCT2 knockdown in JEG-3 cells significantly suppressed the AA-induced mRNA expression of placental P450 cholesterol side-chain cleavage enzyme, 3β-hydroxysteroid dehydrogenase type 1, and aromatase. This suppression of the AA-induced mRNA expression of steroidogenic enzymes subsequently decreased progesterone and estradiol production. In addition, inhibition of MAPK kinase-ERK signaling, which is a major pathway for AA-regulated gene expression, failed to affect AA-induced steroidogenesis. Our observations indicate that SVCT2-mediated AA uptake into cells is necessary for AA-induced steroidogenesis in human choriocarcinoma cell, but MAPK kinase-ERK signaling is not involved in AA-induced steroidogenesis.

THE PLACENTA, a unique organ of the fetomaternal unit, is vital to the development and maintenance of pregnancy. The production of steroid hormones, such as progesterone (P) and estrogens, is a crucial function of the primate placenta. In humans, by 7-wk gestation, nearly all P and estrogens in circulation are synthesized by the placenta (1, 2). In human placenta, steroid biosynthesis is regulated by various steroidogenic enzymes (Fig. 1). P450 cholesterol side-chain cleavage enzyme (P450scc) catalyzes the conversion of cholesterol into pregnenolone, which 3β-hydroxysteroid dehydrogenase type 1 (3β-HSD1) converts into P. Unlike the ovary, which synthesizes pregnancy hormones in rodents and other animals, the human placenta is not capable of performing de novo estrogen synthesis from cholesterol because it lacks P450c17, which converts pregnenolone and P to their 17α-hydroxylated products and then to dehydroepiandrosterone (DHEA) and androstenedione (A-dione), respectively. Instead, the placenta uses C19 androgens derived from the fetus and maternal adrenal glands, such as DHEA and DHEA-sulfate, to synthesize estrogens. In this pathway, 3β-HSD1 catalyzes the conversion of DHEA into A-dione, which is subsequently converted to estradiol (E2) by aromatase, which catalyzes the conversion of C19 androgens to C18 estrogens, and 17β-HSD1 type, which catalyzes the conversion of low-activity estrone to the more active E2.

Fig. 1

The pathway of steroid hormone biosynthesis in human placenta. STS, Steroid sulfatase.

Vitamin C, an essential cellular nutrient, is involved in a variety of metabolic reactions and is an important antioxidant in biological systems. Although tissues vary widely in their reduced vitamin C [ascorbic acid (AA)] content, the highest concentrations occur in the pituitary, adrenal gland, and gonads (3, 4). The high concentrations of AA in such endocrine tissues attest to its importance in hormone synthesis. The neurohypophysial synthesis of oxytocin and vasopressin involves an AA-dependent posttranslational amidation, and neurotransmitter synthesis is similarly promoted by AA in other areas of the brain (5, 6). Catecholamines and corticosterone require AA for their synthesis in the adrenals and to protect them from oxidation (7, 8). In the reproductive system, the ovary has long been recognized as a site of AA accumulation and turnover, and the ability of LH to block the uptake of AA by gonadotropin-primed ovaries provided the basis for an LH bioassay that was used frequently before the advent of immunoassays (9). Previous studies with luteinizing granulosa cells showed that AA stimulates P and oxytocin secretion (10, 11). The ovaries of AA-requiring osteogenic disorder Shionogi/Shi-od/od rats, which are genetically unable to synthesize AA, have significantly increased ovarian aromatase activity, but no effect on P metabolism has been observed (12). All of these studies suggested that AA functions as an important endocrine modulator.

Specific nonoverlapping transport proteins mediate the transport of AA and the oxidized forms of vitamin C, dehydroascorbic acid (DHA), across biological membranes. DHA is transported by the isoforms of facilitated-diffusion glucose transporter (GLUT), GLUT1, GLUT3, and GLUT4 (13). In human placenta, DHA is supposed to be taken up via these isoforms of GLUTs because human placenta and choriocarcinoma cell lines, including JEG-3, BeWo, and JAR cells, have abundantly expressed GLUT1 and GLUT3 (14–16). On the other hand, AA uptake is via sodium-dependent vitamin C transporters (SVCTs) (17–20). Two isoforms of SVCTs (SVCT1 and SVCT2) have been well characterized in rodents and humans (17, 18). Although they exhibit very similar functional properties, SVCT1 and SVCT2 differ in their tissue distribution. The SVCT1 isoform is largely confined to epithelial surfaces involved in bulk transport, such as those of the intestine and kidney, whereas the SVCT2 isoform appears to account for tissue-specific uptake of vitamin C (18–21). SVCT2 expression is widespread, occurring in neurons, the endocrine systems, bone, and other tissues. SVCT2 is abundantly expressed in human placenta and choriocarcinoma JAR cells (19, 20, 22), and several in vitro experiments using human choriocarcinoma lines have indicated that the Na⁺/AA cotransporter is most likely to be the major pathway responsible for the entry of AA from the maternal circulation into the placenta (22, 23). In addition, AA levels are undetectable or markedly reduced in the blood and tissues of SVCT2-null mouse pups, as well as the associated placenta. Providing the pregnant females with prenatal AA supplements does not elevate blood AA in SVCT2-null fetuses, and SVCT2-null mice die within a few minutes of birth with respiratory failure and intraparenchymal brain hemorrhage (24). These observations suggest that SVCT2 is important for placental AA transport and fetal development.

Although much is known about AA function as a regulator of endocrine function in the adrenal glands and ovaries, the knowledge of its physiological role in placenta/trophoblasts has been restricted to its antioxidative function (25, 26). Moreover, although fertility is associated with the involvement of AA in P and E2 production (27, 28), its precise physiological role in human placental/trophoblastic steroidogenesis remains unclear. P and estrogen production occurs in the human choriocarcinoma cell lines JAR, JEG-3, and BeWo, as well as in primary human trophoblast cells (29), and these cell lines have proven to be a very useful and convenient model for the study of various nutrient transport systems in the human placenta (15, 16, 20, 30, 31). In addition, the great majority of research on human placental endocrine function has also been performed on these choriocarcinoma cells (29, 32–35). Here, we investigated the steroidogenic effect of AA and the role of SVCT2 in steroidogenesis using these human choriocarcinoma lines as a human trophoblast model.

Materials and Methods

Cell cultures

Cells of the human choriocarcinoma cell lines JAR, JEG-3, and BeWo were obtained from American Type Culture Collection (Manassas, VA) and cultured as described previously (33). RetroPack PT67 packaging cells (Clontech, Palo Alto, CA) were cultured in DMEM containing 10% fetal calf serum (FCS). All lines were maintained in a humidified atmosphere containing 5% CO₂ at 37 C. In all experiments, cells were seeded, precultured for 24 h in regular culture medium supplemented with 10% normal FCS, and then treated with AA (Nacalai Tesque, Kyoto, Japan) or α-tocopherol (Nacalai Tesque) in regular culture medium supplemented with 5% charcoal-stripped FCS instead of 10% normal FCS. The AA- and α-tocopherol-containing media were replaced with fresh medium every 24 h. PD98059 (Sigma-Aldrich, St. Louis, MO) was applied with or without 100 μm AA for 48 h.

P and E2 measurement in AA-treated human choriocarcinoma cells

The P and E2 assays were described previously (33). After the cells were plated, they were treated with AA or α-tocopherol for 48 h. The cells were then cultured for an additional 4 h in serum-free medium containing 20 μg/ml 25-hydroxycholesterol (Sigma-Aldrich) to determine P production, and 10 μm 4-androstene-3,17-dione (Nacalai Tesque) or testosterone (Sigma-Aldrich) to determine E2 production. After incubation, the media were harvested, and their P and E2 contents were determined with Correlate-EIA P and E2 Enzyme Immunoassay kits (Assay Designs, Ann Arbor, MI).

Isolation of RNA, synthesis of cDNA, and RT-PCR assay

Cells were plated at a density of 5 × 10⁵ cells in 10-cm dishes and were treated with various concentrations of AA for the indicated periods. Total RNA was extracted from the cells using trizol reagent (Invitrogen, Carlsbad, CA). mRNA expression of steroidogenic enzymes in JEG-3 cells was determined by quantitative RT-PCR. We placed 5 μg total RNA extracted from cells in a total volume of 20 μl and reverse-transcribed it using SuperScript III reagent (Invitrogen) with oligo-(dT) as a primer and a 1-h incubation at 42 C. After the termination of cDNA synthesis, each reaction mixture was diluted with 80 μl Tris-EDTA buffer. For the quantitative RT-PCR assay, aliquots (2 μl) of diluted RT products were amplified using LightCycler (Roche Diagnostics, Mannheim, Germany) in a reaction mixture containing QuantiTect SYBR Green PCR reagent (QIAGEN, Valencia, CA) and 0.5 μm of each primer. After preincubation of the reaction mixtures at 95 C for 15 min, real-time PCR amplification was performed using the PCR conditions and primer pairs described in Table 1. The relative amounts of the mRNAs of the steroidogenic enzymes, MLN64, SVCT2, and MAPK1 were determined by calculating the ratio between these mRNAs and the mRNA of the housekeeping gene β-actin. In our experimental systems, AA treatment did not change the expression of β-actin. We achieved similar results when using another housekeeping gene, glyceraldehyde-3-phosphate-dehydrogenase, to normalize the target gene expression (data not shown).

Table 1

Oligonucleotide primer sequences and PCR conditions for real-time PCR

Gene (accession no.)		Primer sequences (5′–3′)	PCR condition				Product length (bp)
			Denaturation	Annealing	Elongation	Cycle no.
SVCT2	Forward	GCA GAG CTG TTG CAC ACA GAA	95 C	62 C	72 C	40	103
(EF032501)	Reverse	CGA GGA GGC CGA TGA CTA CTT	15 sec	30 sec	10 sec
ERK2 (MAPK1)	Forward	GGG CTA CAC CAA GTC CAT TGA TA	95 C	62 C	72 C	35	107
(NM_002745)	Reverse	GGT TCA GCT GGT CAA GAT AAT GCT	15 sec	30 sec	10 sec
P450scc	Forward	CAA TGG CTG GCT AAA CCT GTA CCA	95 C	67 C	72 C	50	104
(M14565)	Reverse	CCC TGT AAA TCG GGC CAT ACT TCT	15 sec	30 sec	10 sec
3β-HSD1	Forward	CAT TGA TGT CTT CGG TGT CAC TCA	95 C	65 C	72 C	40	102
(X53321)	Reverse	ACT GGC ACA CTA GCT TGG ACA CAG	15 sec	30 sec	10 sec
Aromatase	Forward	GAG CTC TGG AAA ACA ACT CGA CCC	95 C	65 C	72 C	40	103
(Y07508)	Reverse	CCA GAT GTG TTT TGA GGG ATT CAG C	15 sec	30 sec	10 sec
17β-HSD1	Forward	GGG CTG CCT TTC AAT GAC GTT T	95 C	65 C	72 C	40	107
(M36263)	Reverse	ATC AGG CTC AAG TGG ACC CCA A	15 sec	30 sec	10 sec
MLN64	Forward	GCC CAG GAA GAG AAC TGG AAG TTT GAG	95 C	62 C	72 C	40	118
(X80198)	Reverse	CAG GAC AGG GCA GGA AGG TC	15 sec	30 sec	10 sec
β-actin	Forward	CCT CGC CTT TGC CGA TC	95 C	62 C	72 C	40	110
(BC013380)	Reverse	AAG CCG GCC TTG CAC AT	15 sec	30 sec	10 sec

For the regular semiquantitative RT-PCR assay, diluted RT products or a human liver cDNA library (Clontech) were amplified with GoTaq DNA Polymerase (Promega, Madison, WI) using the PCR conditions and primer pairs as described in Table 2. The PCR products were electrophoresed in agarose gels and visualized with ethidium bromide staining. Expression of β-actin was used as a positive control. The intensity of the DNA bands was calculated using National Institutes of Health Image software (available at http://rsb.info.nih.gov/nih-image/). The amounts of PCR products were measured within the exponential phase of the amplification curve when amplification had not been saturated.

Table 2

Oligonucleotide primer sequences and PCR conditions for regular RT-PCR

Gene (accession no.)		Primer sequences (5′–3′)	PCR condition				PCR product length (bp)
			Denaturation	Annealing	Elongation	Cycle no.
SVCT1	Forward	CAC CGA AGG CAT TTG CTG CAT CAT C	95 C	58 C	72 C	40	140
(BC050261)	Reverse	GCA TGA TAG CCG CAC CAT ACT G	30 sec	30 sec	15 sec
SVCT2	Forward	GAT GCC ATG TGT GTG GGG TA	95 C	52 C	72 C	37	456
(EF032501)	Reverse	TAT TGT CAG CAT GGC AAT GC	30 sec	30 sec	30 sec
β-Actin	Forward	CGA GCA CAG AGC CTC GCC TTT G	95 C	62 C	72 C	23	598
(BC013380)	Reverse	GGT CCA GAC GCA GGA TGG CAT G	30 sec	30 sec	30 sec

Creation of a small interfering RNA (siRNA)-expressing vector and establishment of stable SVCT2- and ERK2-knockdown JEG-3 cells

HiPerformance 2-For-Silencing siRNA Duplexes for SVCT2 (GenBank accession no. NM_005116) were designed and synthesized by QIAGEN. Duplex 2, with a sense sequence of r(ACUGUUCCGUACAAAUAAA)dTdT and antisense sequence of r(UUUAUUUGUACGGAACAGU)dTdT, suppressed the SVCT2 expression by 61% at the mRNA level when transiently transfected into JEG-3 cells. For construction of a retroviral siRNA-expressing vector, the hairpin siRNA template of complementary oligonucleotides (SVCT2 target sequence: nt 6908–nt 6926) containing BamHI and ClaI overhangs was designed; the sense sequence for SVCT2 was 5′-GATCCACTGTTCCGTACAAATAAATTCAAGAGATTTATTTGTACGGAACAGTCTTTTTAT - 3′, and the antisense sequence was 5′-CGATAAAAAGACTGTTCCGTACAAATAAATCTCTTGAATTTATTTGTACGGAACAGTG-3′. The sense sequence for the negative control (SVCT2 target sequence: nt 2182–nt 2201) was 5′ - GATCCAAAGAGTAGGTAAGAGGCCTTCAAGAGAGGCCTCTTAC - CTACTCTTTTTTTTTAT-3′, and the antisense was 5′-CGATAAAAAAAAAGAGTAGGTAA GAGGCCTCTCTTGAAGGCCTCTTACCTACTCTTTG-3′. The sense sequence for ERK2 (MAPK1: GenBank accession no. NM_002745, target sequence: nt 1083–nt 1103) was 5′-GATCCATGCTGACTCCAAAGCT- CTGTTCAAGAGACAGAGCTTTGGAGTCAGCATTTTTTTAT-3′, and the antisense sequence was 5′-CGATAAAAAAATGCTGACTCCAAAGCTCTGTCTCTTGAACAGAGCTTTGGAGTCAGCATG-3′. The synthesized complementary oligonucleotides were annealed and inserted into the BamHI/ClaI sites of a retroviral siRNA expression vector, pSINsi-hU6 (Takara Bio, Otsu, Japan). The integrity of the siRNA-expressing construct was verified with a DNA sequencer (ABI Prism 310; Applied Biosystems, Foster City, CA). RetroPack PT67 packaging cells were transfected with 5 μg of each construct using Lipofectamine regent (Invitrogen); 48 h after transfection, retrovirus-containing supernatants were harvested, and retroviral supernatants with a titer more than 10⁶ colony forming U/ml were used for infection. JEG-3 cells were exposed to retrovirus-containing medium for 24 h in the presence of 8 μg/ml Polybrene (Nacalai Tesque). Stable selection of infected cells was achieved with the use of 800 μg/ml G418 (Nacalai Tesque) for 2 wk, and the negative control, SVCT2-knockdown and ERK2-knockdown cells were expanded and maintained in G418-containing MEM medium.

Western blotting

Immunoblotting was performed as described previously (36, 37). Cells were plated at a density of 5 × 10⁵ cells in 10-cm dishes; 48 h later, cell lysates from the negative control and SVCT2-knockdown JEG-3 cells were prepared using lysis buffer [10 mm HEPES (pH 7.6), 250 mm NaCl, 5 mm EDTA, and 1% Triton X-100]. Protein samples (40 μg) were subjected to SDS-PAGE electrophoresis in a 10% gel and transferred to a membrane. Immunoblotting was performed with an SVCT2 polyclonal antibody developed by immunizing a rabbit with CYARTDARKGVLLVAP (mSVCT2 288–302) conjugated to keyhole limpet hemocyanin. A peroxidase-conjugated antirabbit IgG was used as a secondary antibody, and the immunoreactive bands were visualized with an ECL system (GE Healthcare Bio-Sciences, Buckinghamshire, UK). National Institutes of Health Image software was used to quantify the immunoreactive bands, and the normalized antigen signals were calculated from SVCT2-derived and α-tubulin-derived signals. The mean density of bands from negative control cells was set to 1.

Kinetics of AA transport

For the kinetic assay of AA transport, negative control and SVCT2-knockdown JEG-3 cells were plated at an initial density of 2 × 10⁴ cells per well of a 24-well plate. After 24 h, the cells were subjected to a [1-¹⁴C]AA uptake assay performed as described previously (36, 37). Briefly, after the cells were rinsed with incubation buffer [in mm: 140 NaCl, 4.2 KHCO₃, 5.8 KCl, 1.3 CaCl₂, 0.5 MgCl₂, 10 HEPES, and 1 dithiothreitol (pH 7.4)], they were incubated with various concentrations of [1-¹⁴C]AA (PerkinElmer Life Science, Boston, MA) at 37 C for 30 min. The amount of incorporated radioactivity was measured with a liquid scintillation counter. Nonspecific incorporation was determined using an excess amount of cold AA at a concentration of 10 mm. Specific incorporation was calculated as total incorporation minus nonspecific incorporation. The kinetic constants of transport, the apparent affinity (Km) and maximal velocity (Vmax), were calculated by fitting the data to a double reciprocal plot. AA uptake appeared to increase linearly within 1 h under these experimental conditions.

Statistics

All data from the control and treatment groups were obtained from the same numbers of replicated experiments. In addition, all the experiments were performed independently two or three times. First, we converted the data to relative values to eliminate sampling bias apart from the case of α-tocopherol. We then analyzed the data sets by linear regression analysis using an algorithm that considered the effect of the errors caused by repeated measurements for the samples of the same concentrations, and we tested the regression coefficients for statistical significance. When we logarithm transformed the concentration data, we added a small positive value, 0.01, to all of the concentration values to avoid undefined log0 values (38). The value of 0.01 was adopted after considering the magnitude of the concentration values. Then we converted the concentration values of AA or α-tocopherol to logarithms and applied a linear regression analysis to the transformed data. The null hypothesis was that the regression coefficient would be equal to zero; therefore, the results showing that the coefficient is significant indicate that the dose- or time-dependent relationships for the effects of AA or α-tocopherol are significant. Differences between the coefficients calculated from the data were evaluated with a test of the differences between regression coefficients. The null hypothesis of the test is that both coefficients are equal. If the null hypothesis is rejected, the dose- or time-dependent relationships for the effects are significantly different.

In the other case of the comparison of the effects, Tukey’s multiple comparisons test or Student’s t test were applied to the raw data. SPSS 13.0J software (SPSS, Inc., Chicago, IL) was used for all the analyses. Results from representative experiments are shown as figures. P < 0.05 was taken as the significant level.

Results

AA stimulates P and E2 production in human choriocarcinoma cells

The steroidogenic effect of AA was examined in JEG-3, JAR, and BeWo cells. In all cell lines, treatment with AA for 48 h significantly stimulated P production in a dose-dependent manner (Fig. 2, A–C). AA also significantly induced E2 production in JEG-3 and JAR cells (Fig. 2, D and E). On the other hand, E2 production in BeWo cells increased numerically but not significantly with increasing AA dose (0.05 < P < 0.10; Fig. 2F). The responsiveness to AA was most robust in JAR cells and weakest in BeWo cells. In vehicle-treated JAR cells, the E2 level was too low to be detected, but treatment with 100 μm AA strikingly induced E2 production.

Fig. 2

AA-induced P and E2 production in human choriocarcinoma cell lines. JEG-3 (A and D), JAR (B and E), and BeWo (C and F) cells were plated at a density of 4 × 10⁴ cells per well in 24-well plates. Cells were treated with the indicated concentrations of AA for 48 h, after which the medium was changed to serum-free medium containing 25-hydroxycholesterol (A–C) or A-dione (D–F), and the cells were cultured for another 4 h. The culture media were harvested for the determination of P (A–C) and E2 (D–F) production. Results are expressed as the mean ± 1 sd of quadruplicate cultures. In the nontreated JAR cells (AA 0 μm), no E2 was found above the limit of detection (2 pg/well) of the method used. Statistical significance was determined with a linear regression analysis after log transformation of the concentration values as described in Materials and Methods. The broken line is the regression line, and b indicates the regression coefficient ± 1 sd. *, P < 0.05; and **, P < 0.01 for the results of the tests of regression coefficient significance.

AA stimulates mRNA expression of steroidogenic enzymes in JEG-3 cells

We next used JEG-3 cells as a representative to examine the mRNA expression levels of steroidogenic enzymes after exposure to AA because JEG-3 cells showed a good response to AA stimulation, and their basal levels of P and E2 production were detectable. Treatment of JEG-3 cells with AA for 48 h increased the mRNA expression of P450scc, 3β-HSD1, and aromatase, but not 17β-HSD1, in a dose-dependent manner (Fig. 3, A–D). We did not detect the mRNA expression of the other isoforms of 17β-HSD, types 2 and 5, in human choriocarcinoma cells by RT-PCR (data not shown). We also examined whether AA-induced hormone production was resulted from the up-regulation of expression of MLN64, a protein that is thought to be responsible for transferring cholesterol into mitochondria in placenta (39). However, treatment with concentrations of AA up to 100 μm for 48 h failed to induce MLN64 mRNA expression in JEG-3 cells (Fig. 3E).

Fig. 3

Dose-dependent effects of AA on the mRNA expression of steroidogenic enzymes and MLN64 in JEG-3 cells as detected by real-time RT-PCR. Cells were plated at a density of 5 × 10⁵ cells per 10-cm dish and treated with AA at the indicated concentrations for 48 h. The expression of P450scc (A), 3β-HSD1 (B), aromatase (C), 17β-HSD1 (D), and MLN64 (E) were normalized to the expression level of β-actin. The data are expressed relative to the mRNA levels of cells treated with 0 μm AA, which were set to one. Results are expressed as means ± 1 sd of three independent cultures. Statistical significance was determined with linear regression after log transformation of the concentration values as described in Materials and Methods. The broken line is the regression line, and b indicates the regression coefficient ± 1 sd. *, P < 0.05; and **, P < 0.01 for the results of the tests of regression coefficient significance.

To determine further the time-dependent effects of AA on the mRNA expression of P450scc, 3β-HSD1, and aromatase, JEG-3 cells were treated with 100 μm AA for up to 48 h. The AA treatment increased the mRNA expression of P450scc, 3β-HSD1, and aromatase in a time-dependent manner (Fig. 4). The induction of each mRNA was maximal 48 h after AA treatment, the last time point examined (Fig. 4), consistent with the results of the dose-response curve (Fig. 3).

Fig. 4

Time-dependent effects of AA on the mRNA expression of steroidogenic enzymes in JEG-3 cells as detected by real-time RT-PCR. Cells were treated with 0 (open circles) or 100 μm (closed circles) AA for the indicated times. The expression of P450scc (A), 3β-HSD1 (B), and aromatase (C) were normalized to the expression level of β-actin. The data are expressed relative to the mRNA levels of cells treated with 0 μm AA at 0 h, which were set to one. Results are expressed as means ± 1 sd of three independent cultures. Statistical significance was determined with a linear regression analysis as described in Materials and Methods. The broken line is the regression line, and b indicates the regression coefficient ± 1 sd. *, P < 0.05; and **, P < 0.01 for the results of the tests of regression coefficients significance, indicating that the time-dependent relationship for the effect of AA are significant. ††, P < 0.01 for the results of the significance tests of the differences between coefficients calculated from the data at 0 and 100 μm, indicating that the responses to the different doses of AA differ significantly.

SVCT2 mRNA expression and protein levels in SVCT2-knockdown JEG-3 cells

We analyzed the mRNA expression of the AA transporters SVCT1 and SVCT2 in human choriocarcinoma cells and human liver tissue by RT-PCR. Both SVCT1 and SVCT2 were expressed in human liver, whereas SVCT2, but not SVCT1, was abundantly expressed in human choriocarcinoma cells (Fig. 5A), suggesting that AA may be transported into the cells mainly by SVCT2 and not by SVCT1. To assess the biological significance of the association of AA-induced steroidogenesis with SVCT2, we used a retrovirally delivered siRNA against SVCT2 to suppress endogenous SVCT2 expression in JEG-3 cells. The SVCT2 mRNA level in JEG-3 cells stably expressing the SVCT2 siRNA was significantly lower than that in negative control cells (Fig. 5B). In addition, Western blotting showed that knocking down SVCT2 expression in JEG-3 cells significantly decreased the intensity of the approximate 70-kDa anti-SVCT2 immunoreactive bands, which correspond to the expected molecular mass of SVCT2 of 71 kDa (Fig. 5C). In three independent cultures, the SVCT2 protein levels in SVCT2-knockdown JEG-3 cells decreased 64 ± 11%, relative to the negative control cells (P < 0.01).

Fig. 5

Analysis of SVCT2 expression and AA uptake in SVCT2-knockdown JEG-3 cells. A, RT-PCR analysis of SVCT isotypes in human choriocarcinoma cells. Total RNA from JEG-3, JAR, and BeWo cells was reverse-transcribed, and 25 ng of the resultant cDNA and a human liver cDNA library were amplified by PCR. In the case of JAR cells, we also amplified 250 ng of the resultant cDNA by PCR because SVCT2 expression in JAR cells was too low to detect by PCR using 25 ng of the resultant cDNA. The human liver cDNA library was used as a positive control for both SVCT1 and SVCT2 expression because both types of SVCT are expressed in human liver (19, 20 ). Reaction products were electrophoresed in 1% agarose gels for the detection of SVCT2 and β-actin or in 1.5% agarose gels for the detection of SVCT1, and visualized by ethidium bromide staining. Data are representative of at least three independent experiments. B, Quantitative RT-PCR analysis of SVCT2 in normal JEG-3 cells (Normal), negative control JEG-3 cells (NEGAsi), and SVCT2-knockdown JEG-3 cells (SVCT2si). The mRNA expression of SVCT2 was normalized to the expression of β-actin. The data are expressed relative to the mRNA levels of normal JEG-3 cells, which were set to one, and are expressed as means ± 1 sd of three independent cultures. **, P < 0.01 vs. normal JEG-3 cells; and ††, P < 0.01 vs. negative control JEG-3 cells for the results of Tukey’s multiple comparisons test. C, Western blot analysis of SVCT2 in SVCT2-knockdown (SVCT2si) JEG-3 cells. The relative SVCT2 protein levels were normalized to α-tubulin-derived immunoreactive signals. D, Lineweaver-Burk plots of AA kinetic transport in negative control (open circle) and SVCT2-knockdown (closed circle) JEG-3 cells. For the uptake assay, cells were plated at a density of 2 × 10⁴ cells per well in 24-well plates. After 24 h, [1-¹⁴C]AA uptake over 30 min was measured at the indicated concentrations of AA. The incorporated [1-¹⁴C]AA was determined with a scintillation counter. Results are expressed as means ± 1 sd of quadruplicate cultures. E, Kinetics of AA transport in negative control (open circle) and SVCT2-knockdown (closed circle) JEG-3 cells. The Km and Vmax were calculated by fitting the data to a double reciprocal plot. Results are expressed as means ± 1 sd of quadruplicate cultures.

Consistent with these results, the functional AA uptake into cells was inhibited in the SVCT2 siRNA-expressing JEG-3 cells. When we performed a kinetic assay in the presence of up to 50 μm [1-¹⁴C]AA, the Vmax of AA uptake was 1386 pmol/10⁵ cells·h for the negative control cells, whereas SVCT2-knockdown JEG-3 cells exhibited a Vmax of 702 pmol/10⁵ cells·h. In contrast, the Km for negative control cells, 30.5 μm, was almost the same as that for SVCT2-knockdown cells, 28.4 μm (Fig. 5, D and E), suggesting that the observed effects on AA uptake kinetics resulted from the knockdown of SVCT2 expression.

Effects of AA on steroidogenesis in SVCT2-knockdown JEG-3 cells

To test whether the knockdown of endogenous SVCT2 in JEG-3 cells decreases AA-induced steroidogenesis, we examined the effect of AA on the mRNA expression of steroidogenic enzymes in negative controls and stably SVCT2-knockdown JEG-3 cells. In SVCT2-knockdown JEG-3 cells, SVCT2 levels were approximately 33% of those in the negative control cells. AA did not affect the mRNA expression of SVCT2 in either cell type (data not shown) but dose dependently induced P450scc, 3β-HSD1, and aromatase mRNA expression in both the negative control and SVCT2-knockdown JEG-3 cells. However, the AA-induced mRNA transcription of P450scc, 3β-HSD1, and aromatase was significantly lower in SVCT2-knockdown JEG-3 cells than in negative control cells (Fig. 6, A–C).

Fig. 6

Effects of AA on the mRNA expression of steroidogenic enzymes and steroid hormone production in SVCT2-knockdown JEG-3 cells. A–C, Negative control (open circles) and SVCT2-knockdown (closed circles) JEG-3 cells were treated with the indicated concentrations of AA for 48 h. After that, total RNA was isolated for the preparation of cDNA. The expression level of P450scc (A), 3β-HSD1 (B), and aromatase (C) were normalized to the expression level of β-actin. The data are expressed relative to the mRNA levels of cells treated with 0 μm AA, which were set to one. Results are expressed as means ± 1 sd of three independent cultures. Statistical significance was determined with a linear regression analysis after log transformation of the concentration values as described in Materials and Methods. The broken line is the regression line, and b indicates the regression coefficient ± 1 sd. **, P < 0.01 for the results of the significance tests of regression coefficients. †, P < 0.05; and ††, P < 0.01 for the results of the significance tests of the differences between coefficients calculated from data in the treatments of negative control JEG-3 cells and SVCT2-knockdown JEG-3 cells. D and E, Negative control and SVCT2-knockdown JEG-3 cells were treated without (open columns) or with (closed columns) 100 μm AA for 48 h. The culture medium was changed to serum-free medium containing 25-hydroxycholesterol (D) or testosterone (E), and the cells were cultured for another 4 h. The culture media were then harvested for the determination of P (D) and E2 (E) levels. All results are expressed as means ± 1 sd of at least three independent cultures. **, P < 0.01 vs. AA 0 m; and ††, P < 0.01 vs. negative control cells treated with AA at same concentration for the results of Tukey’s multiple comparisons test.

We also examined the effect of AA on the steroidogenic enzyme activity in negative control and stably SVCT2-knockdown JEG-3 cells. We assessed P450scc and 3β-HSD1 activity to determine the conversion of cholesterol to P and aromatase activity to determine the conversion of testosterone to E2 (Fig. 1). Basal levels of P and E2 production were almost the same in SVCT2-knockdown JEG-3 cells and negative control cells (Fig. 6, D and E). Consistent with the results of the AA-induced mRNA expression, treatment of SVCT2-knockdown JEG-3 cells with 100 μm AA for 48 h stimulated both P and E2 production, but the amount of stimulation was significantly lower than in negative control cells. These results suggest that AA-induced steroidogenesis in human placenta may require AA transport by SVCT2.

Potential mechanisms for AA-induced hormone production in JEG-3 cells

We showed that AA induces P and E2 production via up-regulation of P450scc, 3β-HSD1, and aromatase mRNA transcription, however, further details of the mechanisms by which AA stimulates steroidogenic enzyme expression and hormone production remained unknown. AA serves as an effective antioxidant, and is responsible for maintaining intracellular redox status and protecting cells against oxidative stresses (3, 40). To investigate whether the antioxidative effect of AA regulates hormone production, we treated JEG-3 cells with another antioxidant, α-tocopherol, and observed the subsequent hormone production. When α-tocopherol was used in place of AA, P and E2 production increased numerically but not significantly with increasing α-tocopherol dose (0.05 < P < 0.10; Fig. 7). This contrasts with the statistical significance (P < 0.05) of same analysis using AA (Fig. 2). In addition, the effects of 100 μm AA and 100 μm α-tocopherol significantly differed from each other (Student’s t test; P < 0.05). Thus, α-tocopherol was less effective than AA at stimulating P and the E2 production in JEG-3 cells.

Fig. 7

Effects of α-tocopherol on P and E2 production in JEG-3 cells. Cells were treated without (open columns) or with the indicated concentrations of α-tocopherol (closed columns), or 100 μm AA (hatched columns) for 48 h, after which the medium was changed to serum-free medium containing 25-hydroxycholesterol (A) or A-dione (B), and the cells were cultured for another 4 h. The culture media were harvested for the determination of P (A) and E2 (B) production. Results are expressed as the mean ± 1 sd of quadruplicate cultures. Statistical significance was determined with a linear regression analysis after log transformation of the concentration values as described in Materials and Methods. The broken line is the regression line, and b indicates the regression coefficient ± 1 sd. No null hypotheses were rejected at the significance level of P < 0.05, although both tests were close to significance (P < 0.1). Differences between the averages calculated from the data in the effects of AA and α-tocopherol at 100 μm were evaluated by Student’s t test. *, P < 0.05; and **, P < 0.01 for the results of Student’s t test.

MAPK signaling is a major pathway for AA-regulated gene expression (41, 42). To determine whether the MAPK signaling pathway is involved in AA-induced hormone production, we knocked down ERK2 (MAPK1) in JEG-3 cells using a retrovirally delivered siRNA against ERK2. The level of ERK2 mRNA in JEG-3 cells stably expressing the ERK2 siRNA was significantly lower than that in negative control cells (Fig. 8A). However, stable expression of the ERK2 siRNA did not affect the expression of SVCT2 mRNA. We then examined the effect of AA on the steroidogenic enzyme activity in negative control and stable ERK2-knockdown JEG-3 cells. Neither P nor E2 production in response to AA was significantly different between negative control and ERK2-knockdown cells (Fig. 8, B and C). In addition, we applied a specific inhibitor of MAPK kinase (MEK)1/MEK2 (an upstream regulator of ERK1/ERK2), PD98059, to inhibit the ERK1/ERK2 activity. Although treatment with PD98059 alone significantly induced P and E2 production in JEG-3 cells, inhibition of MEK1/MEK2 activity by PD98059 did not inhibit the P and E2 production induced by AA (Fig. 8, D and E).

Fig. 8

Effects of ERK2 knockdown and PD98059 treatment on AA-induced P and E2 production in JEG-3 cells. A, Quantitative RT-PCR analysis of SVCT2 (open columns) and ERK2 (hatched columns) in normal JEG-3 cells (Normal), negative control JEG-3 cells (NEGAsi), and ERK2-knockdown JEG-3 cells (ERK2si). The mRNA expression of SVCT2 was normalized to the expression of β-actin. The data are expressed relative to the mRNA level of normal JEG-3 cells, which was set to one, and are expressed as means ± 1 sd of three independent cultures. **, P < 0.01 vs. normal JEG-3 cells and ††, P < 0.01 vs. negative control JEG-3 cells for the results of Tukey’s multiple comparisons test. B and C, Negative control and ERK2-knockdown JEG-3 cells were treated without (open columns) or with (closed columns) 100 μm AA for 48 h, after which the medium was changed to serum-free medium containing 25-hydroxycholesterol (B) or A-dione (C), and the cells were cultured for another 4 h. The culture media were harvested for the determination of P (B) and E2 (C) production. Results are expressed as the mean ± 1 sd of quadruplicate cultures. *, P < 0.05, **, P < 0.01 vs. AA 0 m and †, P < 0.05, ††, P < 0.01 vs. negative control cells treated with AA at same concentration for the results of Tukey’s multiple comparisons test. D and E, Normal JEG-3 cells were treated without (control) or with 50 μm PD98059 in the absence (open columns) or presence (closed columns) of 100 μm AA for 48 h, after which the medium was changed to serum-free medium containing 25-hydroxycholesterol (D) or A-dione (E), and the cells were cultured for another 4 h. The culture media were harvested for the determination of P (D) and E2 (E) production. Results are expressed as the mean ± 1 sd of quadruplicate cultures. *, P < 0.05; **, P < 0.01 vs. AA 0 m; and †, P < 0.05; ††, P < 0.01 vs. control cells treated with AA at same concentration for the results of Tukey’s multiple comparisons test.

Discussion

AA, an essential component in the diet of humans and some animals, such as guinea pig and monkeys, is associated with hormone synthesis. Lower concentrations of AA stimulate the catalytic activity of P450scc in bovine adrenal glands, but higher concentrations inhibit the activity (43). AA stimulates the catalytic activity of 3β-HSD in rat adrenal glands (44). In luteinizing granulosa cells, AA stimulates P and oxytocin secretion, and acts synergistically with neurotransmitters to stimulate the secretion of other hormones (10, 11). In addition, human fertility is believed to be associated with the effect of AA on P and E2 production because AA supplementation effectively prevents abortion (27, 28), significantly stimulates serum P and E2 levels, and substantially increases pregnancy rates in patients with luteal phase defects (28). However, its precise physiological role in steroidogenesis and reproduction has been uncertain. Here, we first demonstrated that physiological concentrations of AA (∼100 μm) time and dose dependently induced P450scc, 3β-HSD1, and aromatase mRNA expression (Figs. 3 and 4), which subsequently increased P and E2 production in human placental choriocarcinoma cell lines (Fig. 2). In addition, the stable knockdown of SVCT2 expression in JEG-3 cells markedly decreased AA uptake into the cells, significantly inhibited AA-induced P450scc, 3β-HSD1, and aromatase mRNA expression, and decreased P and E2 production. Our data suggest that SVCT2-mediated AA transport plays a key role in AA-induced steroidogenesis in human choriocarcinoma cell lines. Thus, AA may be an essential nutrient to establish and maintain human pregnancy with the result of up-regulation of placental/trophoblastic steroidogenesis.

AA transport is sodium dependent and is mediated by SVCT2, whereas transport of DHA may be mediated by GLUT isoforms 1 and 3 because these GLUT isoforms are abundantly expressed in human placenta and choriocarcinoma cell lines (14–16). Although which forms of transporter are responsible for vitamin C uptake into cells is still controversial, there are several reports favoring the hypothesis that transport of DHA via GLUTs is the principle pathway of vitamin C uptake into placenta (45, 46), and the transported DHA is reduced back to AA inside the cells. However, other studies have indicated that although both GLUTs and SVCT are expressed in human placenta and choriocarcinoma cell lines, SVCT serves as the predominant pathway responsible for vitamin C uptake in JAR cells and the entry of vitamin C from the maternal circulation into the placenta (23). Moreover, it is likely that extracellular AA was partially converted to DHA during the 48-h culture in our experiments. Therefore, it is possible that this newly generated DHA may have stimulated human placental steroidogenesis. However, the DHA concentration in maternal blood (∼2 μm) is much lower than the AA concentration (∼100 μm). In addition, the affinity of GLUTs for DHA is in the millimolar range in the absence of any competing monosaccharide (45, 47), whereas the results of our present study show that AA uptake by SVCT2 is of a higher affinity, with a Km of 30 μm in JEG-3 cells. In our experiments the culture media for choriocarcinoma cells contained approximately 5 mm (for JEG-3 and BeWo cells) or 22.5 mm (for JAR cells) of glucose, which can powerfully compete with DHA for entry into cells (48). Furthermore, the physiological concentration of DHA (2 μm) failed to induce hormone production in our culture conditions (data not shown), indicating that DHA is not the right form of AA for the induction of steroidogenesis in human choriocarcinoma cell lines. Thus, we concluded that the increases in P450scc, 3β-HSD1, and aromatase mRNA expression, and the subsequent induction of P and E2 production in JEG-3 cells predominantly result from an increase in SVCT2-mediated AA uptake rather than GLUT-mediated DHA uptake. This conclusion is also strongly supported by our observations that the suppression of SVCT2 expression using a retrovirally mediated RNAi knockdown technique decreased AA-induced steroidogenesis in JEG-3 cells and robustly decreased P450scc, 3β-HSD1, and aromatase mRNA expression (Fig. 6).

The stable suppression of SVCT2 was not lethal in cell culture and, in fact, hardly affected the growth of the knockdown cells compared with the cells transfected with the negative control vector. In addition, knocking down SVCT2 had little effect on the steady-state steroidogenesis (Fig. 6) or the mRNA levels of P450scc, 3β-HSD1, 17β-HSD1, and aromatase (data not shown). These results suggest that AA may not be essential for cell growth or steady-state steroidogenesis in human placenta. In contrast, knockdown of SVCT2 expression strongly affected the AA-induced steroidogenesis in JEG-3 cells and robustly decreased P450scc, 3β-HSD1, and aromatase mRNA expression (Fig. 6). Thus, our stable SVCT2 knockdown system using JEG-3 cells may be a valuable tool for investigating the physiological roles of AA and SVCT2 in human placenta in pathways other than AA-induced steroidogenesis.

AA serves as a powerful antioxidant and is responsible for maintaining intracellular redox status and protecting cells against oxidative stresses (3, 40). In our experimental conditions, α-tocopherol, another antioxidant, showed a tendency to enhance P and E2 production in JEG-3 cells in a dose-dependent manner (Fig. 7), suggesting that the antioxidative properties of AA might account for its effect on human placental steroidogenesis. However, α-tocopherol failed to significantly induce P and E2 production in JEG-3 cells unlike AA (Figs. 2 and 7). In addition, 100 μm α-tocopherol induced significantly less hormone production than 100 μm AA. If the antioxidative effect is crucial for AA-induced steroidogenesis, these observations suggest that the lower ability of α-tocopherol to induce steroidogenesis might result from its lower antioxidative capacity. Further studies using other antioxidants, such as glutathione, should be performed to clarify this issue.

MAPK signaling is the major pathway for AA-regulated gene expression (41, 42, 49). In this study we knocked down ERK2 in JEG-3 cells to investigate the role of MAPK in AA-induced steroidogenesis, but knocking down ERK2 in JEG-3 cells did not result in an obvious decrease of AA-induced steroidogenesis compared with that in negative control cells (Fig. 8, B and C). We also treated JEG-3 cells with PD98059, a MEK1/MEK2 inhibitor, followed by AA treatment. Although the treatment with PD98059 alone significantly induced P and E2 production in JEG-3 cells, inhibition of MEK1/MEK2 activity did not inhibit AA-induced steroidogenesis (Fig. 8, D and E). These data suggest that the MEK-ERK signaling pathway is not likely to be involved in AA-induced steroidogenesis in JEG-3 cells. Therefore, the exact mechanisms by which AA induces the expression of these genes and hormone production remain to be established. Although steroidogenesis induced by epidermal growth factor, a simulator of MEK-ERK signaling pathway, occurs in several hours (50), AA-induced mRNA expression of steroidogenic enzymes in JEG-3 cells was not found by 24 h or less of AA treatment, suggesting that AA might indirectly affect human placental steroidogenesis, like AA-induced osteocalcin expression in MC3T3-E1 preosteoblast cells (51). Future studies are needed to clarify the precise mechanism of action of AA in human placental endocrine function in vitro and in vivo. However, it is interesting that knocking down ERK2 had little effect on the steady-state P and E2 production, whereas inhibition of MEK1/MEK2 activity significantly enhanced the production of these hormones. Therefore, human placental steroidogenesis may involve PD98059-sensitive MEK-ERK cascades other than ERK2.

In summary, we provide the first evidence that AA potentially promotes steroidogenic action to enhance the mRNA expression of P450scc, 3β-HSD1, and aromatase in human placenta. In addition, we succeeded in establishing SVCT2-knockdown JEG-3 cells by retrovirally mediated RNAi, and demonstrated that the knockdown of SVCT2 expression in human choriocarcinoma cells significantly inhibits both AA uptake into the cells and AA-induced steroidogenesis. Although it has been established that SVCT2 is important in adrenal endocrine function, placental AA transport, fetal development, and bone formation (20, 52), our data suggest that SVCT2 potentially plays a dual role in placenta: stimulation of steroidogenesis for maintaining pregnancy and facilitation of AA accumulation in the fetus.

Acknowledgments

This research was supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports, and Culture of Japan; and by Health and Labor Sciences Research Grants (Research on Advanced Medical Technology) from the Ministry of Health, Labor and Welfare of Japan. X.W. was the recipient of a postdoctoral fellowship from the Society of Japanese Pharmacopoeia.

Disclosure Statement: The authors have nothing to disclose.

Abbreviations

 
• AA

  Ascorbic acid

 
• A-dione

  androstenedione

 
• DHA

  dehydroascorbic acid

 
• DHEA

  dehydroepiandrosterone

 
• E2

  estradiol

 
• FCS

  fetal calf serum

 
• GLUT

  glucose transporter

 
• HSD

  hydroxysteroid dehydrogenase

 
• Km

  apparent affinity

 
• P450scc

  P450 cholesterol side-chain cleavage enzyme

 
• MEK

  MAPK kinase

 
• P

  progesterone

 
• siRNA

  small interfering RNA

 
• SVCT

  sodium-dependent vitamin C transporter

 
• Vmax

  maximal velocity.

Author notes

^*

X.W. and T.I. contributed equally to this work.