2024 Volume 49 Issue 2 Pages 69-77
Placental dysfunction can disrupt pregnancy. However, few studies have assessed the effects of chemical-induced toxicity on placental function. Here, we examined the effects of valproic acid (VPA) as a model chemical on production of hormones and on glucose uptake in human choriocarcinoma cell line BeWo. Cells were treated with forskolin to differentiate into syncytiotrophoblasts, which were then treated with VPA for 72 hr. Real-time RT-PCR analysis showed that VPA significantly increased the mRNA expression of chorionic gonadotropin β (CGB), a hormone that is produced by the placenta in the first trimester of pregnancy, relative to that in the forskolin-only group. It also suppressed the increase in intracellular glucose uptake and GLUT1 level observed in the forskolin-only group. RNA-seq analysis and pathway database analysis revealed that VPA consistently decreased the level of HIF-1α protein and expression of its downstream target genes HK2 and ADM in the hypoxia pathway. Cobalt chloride, a HIF-1α inducer, inhibited CGB upregulation in VPA-treated cells and rescued VPA-induced suppression of glucose uptake and GLUT1 level. Thus, HIF-1α-mediated elevation of CGB expression and suppression of glucose uptake by VPA is a novel mechanism of placental dysfunction.
Pregnant women and fetuses are vulnerable to many chemicals, including alcohol and nicotine (Brinchmann et al., 2023; Wikoff et al., 2017). Chemical exposure can be a risk factor for adverse outcomes such as spontaneous abortion and fetal congenital anomalies or developmental disorders (Gómez-Roig et al., 2021). Therefore, it is necessary to understand how chemicals can cause reproductive toxicity.
The reproductive toxicity of chemicals is determined mainly from the presence or absence of toxic signs, such as maternal general toxicity, maternal reproduction, and the development of the next generation. On the other hand, the evaluation of the placenta is limited (Cindrova-Davies and Sferruzzi-Perri, 2022). The developing fetus depends on the mother via the placenta until birth (Maltepe and Fisher, 2015). Therefore, it is easy to imagine that if the placenta is damaged, the fetus will be adversely affected. Thus, when evaluating the reproductive toxicity of chemicals, it is essential to evaluate their effect on placenta.
The placenta plays essential roles in the formation and maintenance of pregnancy, including protecting the fetus from foreign substances via the placental barrier (Costa, 2016). However, placental function can be harmed through exposure to various chemicals; for example, glucocorticoids, which are administered to pregnant women at risk of premature birth, reduce the expression of genes for amino acid transporters in placental cells (Audette et al., 2014), and bisphenol A, a compound used in plastic food containers, increases placental hormone production (Paulesu et al., 2018). In addition, children of pregnant women who received multiple doses of glucocorticoids had 3.8 × the childhood behavioral disorder scores of normal infants (Crowther et al., 2007; French et al., 2004), and treatment of experimental animals with bisphenol A restricts fetal growth via placental angiogenesis (Müller et al., 2018). Although a relationship between impairment of placental function due to chemical exposure and adverse pregnancy outcomes has been suggested, how chemical exposure causes reproductive toxicity by affecting placental function is still obscure.
Within the placenta, chorionic villi mediate material exchange between the maternal blood and the fetal blood. They consist of two types of cell layers, made of cytotrophoblasts and syncytiotrophoblasts (Knöfler et al., 2019). Syncytiotrophoblast cells are formed in early pregnancy by the syncytialization process, by which mononuclear cytotrophoblast cells fuse with each other (Knöfler et al., 2019) and play a central role in maintaining placental function, such as producing placental hormones to maintain pregnancy and supplying nutrients to the fetus (Turco and Moffett, 2019). Human chorionic gonadotropin β (hCGβ; gene name CGB), a hormone that is produced by syncytiotrophoblasts in the first trimester of pregnancy (Kovalevskaya et al., 2002), assists placental steroid hormone production (Nwabuobi et al., 2017), and human placental lactogen (hPL; gene name CSH1) maintains maternal blood glucose at a higher concentration than in non-pregnancy owing to its anti-insulin action and supplies maternal glucose to the fetus (Sibiak et al., 2020). Moreover, the fetus relies on the transplacental supply from maternal blood for most nutrients, and their uptake and transport to the fetus are essential for fetal development (Malhotra et al., 2019). GLUT1 is the major glucose transporter in syncytiotrophoblasts and is responsible for glucose uptake and its transport to the fetus (Baumann et al., 2007). Insufficient supply of glucose increases the risk of fetal cranial nerve damage (Antonow-Schlorke et al., 2011).
To investigate these aspects, we used valproic acid (VPA) as a model chemical. VPA has been used from the late 1900s clinically for the treatment of epilepsy and bipolar disorder (van Breemen et al., 2009). However, clinical reports show that the risk of the development of congenital abnormalities in fetuses rose from 2%–3% in healthy pregnant women to 24% in women who took high doses of VPA (Thomas et al., 2007), and that children born to women who took VPA during pregnancy had an average IQ value 8.9 points lower than those of women who did not take it (Tomson et al., 2011). Thus, VPA presents a risk of reproductive and developmental toxicity.
Here, we evaluated VPA effects on the hCGβ and hPL production and on glucose uptake in the human choriocarcinoma cell line BeWo, a widely used model of human trophoblast (Myllynen and Vähäkangas, 2013), and attempted to clarify the mechanisms of VPA action.
The human choriocarcinoma cell line BeWo was purchased from the Japanese Collection of Research Bioresources Cell Bank (JCRB9111; Osaka, Japan) and cultured in 10% inactivated fetal bovine serum (Biosera, Nuaille, France) and 1% (v/v) penicillin–streptomycin–amphotericin B suspension (FUJIFILM Wako Pure Chemical, Osaka, Japan) in Ham’s F-12 nutrient mixture (Wako) and maintained under standard cell-culture conditions at 37°C and >95% relative humidity (RH) under a 5% CO2 atmosphere (Sakahashi et al., 2022).
Treatments with forskolin, VPA, and cobalt chlorideCells were seeded at 1.5 × 105 cells/mL (for real-time RT-PCR) or at 1.5 × 105 cells/2 mL (for other experiments) per well in six-well flat plates, held overnight at 37°C and >95% RH. Forskolin (Frk; Cayman Chemical, Ann Arbor, MI, USA) was diluted in Ham’s F-12 (Wako) to a final concentration of 50 µM. Cells were treated with Frk for 24 hr at 37°C, >95% RH, in 5% CO2. The culture medium was decanted, and cells were treated for 72 hr with VPA (Wako) in Ham’s F-12 at 0.3, 0.6, 0.9, or 1.2 mM or a subset of these concentrations as indicated in the figures. In VPA+ CoCl2 treatment, cells were treated for 72 hr with VPA (0.6 mM) with CoCl2 (Nacalai Tesque, Kyoto, Japan) at 0, 5, or 10 µM.
Real-time RT-PCRTotal RNA was extracted by using a FastGene RNA Kit (Nippon Genetics, Tokyo, Japan) and reverse-transcribed into cDNA by using a High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, Waltham, MA, USA). cDNAs were used as templates. Primers (Table 1; Eurofins Genomics, Tokyo, Japan), a GeneAce SYBR qPCR Mix α Low ROX kit (Nippon Gene, Tokyo, Japan), and a CFX-384 Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA) were used. The level of each transcript was normalized to that of GAPDH.
| Target | Forward | Reverse |
|---|---|---|
| ADM | 5′-CCAGAGCATGAACAACTTCCA-3′ | 5′-TGTCCTTGTCCTTATCTGTGAACTG-3′ |
| CGB | 5′-GCTCACCCCAGCATCCTATC-3′ | 5′-CCTGGAACATCTCCATCCTTG-3′ |
| CSH1 | 5′-ACCTCGGACAGCGATGACTAT-3′ | 5′-CTGTAGGTCTGCTTGAGGATCTG-3′ |
| HK2 | 5′-CTTTGTGAGGTCCACTCCAGAT-3′ | 5′-GAGCCCATTGTCCGTTACTTT-3′ |
| GAPDH | 5′-GAAGGTGAAGGTCGGAGTC-3′ | 5′-GAAGATGGTGATGGGATTTC-3′ |
According to the manufacturer’s instructions provided with the Glucose Uptake Assay Kit-Green (Dojindo Laboratories, Kumamoto, Japan), treated cells were washed multiple times with serum- and glucose-free Ham’s F-12 medium (Research Institute for Functional Peptides, Yamagata, Japan), and then treated for 30 min with fluorophore-glucose diluted 1:500 with the same Ham’s F-12. Intracellular fluorescence was measured under a BZ-X800 all-in-one microscope (Keyence Engineering Corporation, Osaka, Japan).
Statistical analysisStatistical analyses using Tukey’s method were conducted in GraphPad Prism v. 9.3.1 software. Data are expressed as means ± SD. P-values < 0.05 were considered statistically significant.
First, we evaluated the effects of VPA on placental hormone production. Cells were treated with VPA, and the levels of CGB and CSH1 transcripts were analyzed by real-time RT-PCR. Forskolin (Frk) increases intracellular cAMP concentration, and its addition induces cytotrophoblasts fuse and differentiate into multinucleated syncytiotrophoblasts (Nampoothiri et al., 2007). Multinucleated cells were observed in Frk-treated BeWo cells (Supplementary Fig. 1) and the expression of ERVFRD-1 which plays important roles in cell fusion during the trophoblast syncytialization process and SDC-1 which mediates cell signaling during the placental formation process (Sakahashi et al., 2022) was elevated in Frk-treated BeWo cells compared to the untreated cell (Supplementary Fig. 2). These results suggest that Frk induced syncytialization in BeWo cells. Frk treatment significantly increased CGB expression relative to the untreated cells, and subsequent VPA treatment further increased it concentration-dependently (Fig. 1a). Frk did not significantly change CSH1 expression relative to the untreated group, but VPA decreased it concentration-dependently (Fig. 1b). VPA increased hCGβ production and decreased hPL production concentration-dependently (Supplementary Fig. 3a). These results show that VPA can disrupt placental hormone production in differentiated BeWo cells.
Valproic acid changes placental hormone expression and suppresses glucose uptake in BeWo cells. BeWo cells were treated with forskolin (Frk, 50 µM) for 24 hr and then with valproic acid (VPA, 0.3, 0.6, 0.9, or 1.2 mM) for 72 hr. Expression of (a) CGB and (b) CSH1 normalized to that of GAPDH was evaluated by real-time RT-PCR. Data are means ± SD (n = 3). **P < 0.01, ****P < 0.0001. (c) After treatment with VPA (0.3, 0.6, or 0.9 mM) for 72 hr, BeWo cells were treated with fluorophore–glucose for 30 min. Uptake was visualized by fluorescence microscopy. Three to five fields of view were randomly taken for each treatment group of cells. Scale bars = 500 µm. All experiments were repeated with similar results.
Since VPA can cause cranial neuropathy in offspring (Safdar and Ismail, 2023), we thought that it might affect the supply of nutrients to the placenta, and thus evaluated its effect on glucose uptake in Frk-treated BeWo cells. We evaluated the effect as the amount of intracellular fluorescence as an index when tagged glucose was added. Frk increased the intracellular uptake of fluorescent glucose relative to the untreated group, but VPA decreased it concentration-dependently relative to Frk alone (Fig. 1c). Since VPA decreased glucose uptake, we evaluated its effect on the level of GLUT1 by western blotting analysis. Frk increased the production of GLUT1 and VPA decreased it (Supplementary Fig. 3b). These results suggest that the suppression of glucose uptake by VPA in differentiated BeWo cells involves VPA-induced downregulation of GLUT1.
Valproic acid inhibits HIF-1 transcriptional activity pathwayTo elucidate how VPA harms placental function, we analyzed genes whose expression it altered in BeWo cells. Frk decreased the expression of 456 genes and VPA increased it (Cluster B, Fig. Supplementary Fig. 4a). Frk increased the expression of 450 genes and VPA decreased it (Cluster D, Supplementary Fig. 4a). Pathway analysis of these genes (Cluster D) identified the “HIF-1 transcriptional activity in hypoxia” pathway, as having the most significant q-value (Supplementary Fig. 4b). In the early stage of pregnancy, blood flow in the placenta is low, so the oxygen partial pressure and concentration in the placenta are low, resulting in a hypoxic state (Burton et al., 2021). Transcription factor hypoxia-inducible factor (HIF) mediates the transcriptional responses to hypoxia. It has a regulatory α-subunit (HIF-1α) and constitutive β-subunit (HIF-1β). Under normoxia, the former is posttranslationally degraded, but under hypoxia, it is not degraded (Mimura et al., 2012). In vitro, HIF pathway is activated in hypoxic conditions and regulates placental villus growth, invasion, and differentiation (Burton et al., 2021; Highet et al., 2015).
Thus, to evaluate the production of molecules involved in that pathway, we assessed the effect of VPA on HIF-1α protein production. Frk increased the production of HIF-1α and VPA decreased it concentration-dependently (Supplementary Fig. 5). Real-time RT-PCR analysis of two genes downstream of HIF pathway, HK2 (Fig. 2a) and ADM (Fig. 2b), showed that Frk increased their expression and VPA decreased it. These results indicate that VPA suppresses the expression of two genes involved in the HIF-1 transcriptional activity in the hypoxia pathway, suggesting that VPA inhibits HIF pathway activity.
Valproic acid downregulates downstream genes of HIF-1α in BeWo cells. Cells were treated with forskolin (Frk, 50 µM) for 24 hr and then with valproic acid (VPA, 0.3, 0.6, 0.9, or 1.2 mM) for 72 hr. The experiment was repeated with similar results. Expression of (a) HK2 and (b) ADM normalized to that of GAPDH was evaluated by real-time RT-PCR. Data are means ± SD (n = 3). ****P < 0.0001. The experiment was repeated with similar results.
Since we showed that VPA inhibits HIF-1α production in Frk-treated BeWo cells, we investigated whether its effect on the hormone production and on glucose uptake was mediated by this inhibition. Under normoxic conditions, HIF-1α is hydroxylated by hydroxylases (Ivan et al., 2001) and then degraded by the ubiquitin-proteasome system; cobalt chloride (CoCl2) inhibits hydroxylase activity and thus increases HIF-1α production (Befani et al., 2013). Therefore, we evaluated placental dysfunction induced by VPA after HIF-1α activity was increased by co-treatment with VPA and CoCl2 in Frk-treated BeWo cells. The morphology of Frk-treated BeWo cells was not changed between VPA- and VPA + CoCl2-treated cells (Supplementary Fig. 6), and there was no significant change in the cell viability between VPA- and VPA + CoCl2-treated cells (Supplementary Fig. 7). VPA decreased HIF-1α production relative to Frk, but co-treatment with VPA + CoCl2 increased it relative to VPA alone (Supplementary Fig. 8a). VPA + CoCl2 increased the expression of both HK2 (Fig. 3a) and ADM (Fig. 3b) relative to VPA alone. This confirms that VPA + CoCl2 increased HIF pathway activity.
CoCl2, a HIF-1α inducer, upregulates downstream genes of HIF-1α. BeWo cells were treated with forskolin (Frk, 50 µM) for 24 hr and then with valproic acid (VPA, 0.6 mM) or VPA + CoCl2 (0, 5, or 10 µM) for 72 hr. Expression of (a) HK2 and (b) ADM normalized to that of GAPDH was evaluated by real-time RT-PCR. Data are means ± SD (n = 3). ****P < 0.0001. All experiments were repeated with similar results.
We evaluated placental hormone production. Co-treatment with VPA + CoCl2 inhibited the increase in CGB expression induced by VPA, and the expression of CGB decreased as CoCl2 concentration increased (Fig. 4a). On the other hand, VPA + CoCl2 did not increase the expression of CSH1, but rather decreased it more than VPA alone (Fig. 4b). These results suggest that VPA modulates CGB expression through inhibition of HIF pathway activity.
HIF-1α inducer inhibits CGB upregulation in valproic acid–treated BeWo cells and rescues valproic acid-induced suppression of glucose uptake. BeWo cells were treated with forskolin (Frk, 50 µM) for 24 hr and then with valproic acid (VPA, 0.6 mM) or VPA + CoCl2 (0, 5, or 10 µM) for 72 hr. Expression of (a) CGB and (b) CSH1 normalized to that of GAPDH was evaluated by real-time RT-PCR. Data are means ± SD (n = 3). *P < 0.05, ***P < 0.001, ****P < 0.0001. (c) Cells were treated as above, and then with fluorophore–glucose for 30 min. Uptake was visualized by fluorescence microscopy. Three to five fields of view were randomly taken for each treatment group of cells. Scale bars = 500 µm. The experiment was repeated with similar results.
We examined whether the inhibition of glucose uptake by VPA was mediated by the inhibition of HIF pathway activity. VPA suppressed glucose uptake relative to Frk alone, while VPA + CoCl2 increased it to the same extent as Frk alone (Fig. 4c). VPA decreased the production of GLUT1 relative to Frk alone, while VPA + CoCl2 increased it (Supplementary Fig. 8b). These results suggest that VPA inhibits glucose uptake and GLUT1 production through inhibition of HIF pathway activity in differentiated BeWo cells.
In the present study, we found that in Frk-treated BeWo cells, VPA increased CGB expression and suppressed glucose uptake under mediation by HIF pathway activity. However, how VPA inhibits HIF pathway activity, the relationship between HIF pathway activity and CGB expression, and how VPA reduces CSH1 expression have not been elucidated. Understanding these will lead to an understanding of the reproductive toxicity of VPA.
The morphology of Frk-treated BeWo cells was not change after VPA treatment, however, changes in the localization of ZO-1 were observed. Since not all cytotrophoblasts differentiate into syncytiotrophoblasts due to Frk treatment, it is possible that VPA could effect on the formation process of syncytiotrophoblasts. Besides, RNA-Seq analysis showed that the increased expression ratio of ERVFRD-1 and SDC-1 in only Frk-treated cells tends to reduce in Frk + VPA treated cells. Moreover, a significant decrease in cell viability was observed in VPA-treated cells compared to only Frk-treated cells, indicating that VPA could decrease mitochondrial activity. In the future, it is expected that investigating the effect of VPA on the cell fusion process will lead to a more detailed mechanism elucidation.
Histone deacetylases (HDACs) remove acetyl from the ε-N-acetyl lysine of histones, allowing histones to wrap DNA more tightly (Peng et al., 2020). The transcription factor HIF-1 is associated with epigenetic regulation of gene expression; the class I HDACs (HDAC1 and HDAC3) promote the stabilization of HIF-1α and HIF transactivation function independent of oxygen concentration (Kim et al., 2007). As VPA is suggested to be an HDAC inhibitor (Göttlicher et al., 2001), it could inhibit HIF-1α stabilization, thereby suppressing HIF-1α production in BeWo cells. To clarify this hypothesis, we need to examine the relationship between HIF pathway activity and HDAC1 activity in differentiated BeWo cells by using valpromide, an amide derivative of VPA that is not an HDAC inhibitor (Gorres et al., 2016).
There are few reports of a direct relationship between HIF pathway activity and CGB expression. The gene for Glial Cells Missing Transcription Factor 1 (GCM1) is activated by the cAMP pathway and promotes CGB transcription (Cheong et al., 2015). The GCM1 degradation is promoted under conditions that induce HIF-1α production, such as in the presence of CoCl2 or under hypoxic conditions (Chiang et al., 2009). Therefore, VPA could suppress the GCM1 degradation by inhibiting HIF-1α, thus elevating CGB level.
VPA decreased CSH1 expression, but its action this effect was not mediated by inhibition of HIF pathway activity. In BeWo cells, activation of the MAPK/ERK pathway downregulates CSH1 expression (Dawid et al., 2019). In RAW264.7 cells, the MAPK/ERK pathway is activated by enhancement of the GABA signaling, a pharmacological activity of VPA (Sugiura et al., 2011), suggesting that VPA may act via the MAPK/ERK pathway in BeWo cells. In H9c2 cells, the MAPK/ERK pathway is activated by CoCl2 (Cheng et al., 2017). Thus, both VPA and CoCl2 might activate the MAPK/ERK pathway, and such treatment could further decrease CSH1 expression.
Although we showed that VPA elevates HIF-1α-mediated CGB expression and suppresses glucose uptake in Frk-treated BeWo cells, it is also essential to assess these effects in vivo. Moreover, for the evaluation of toxicity to placental function, it is necessary to verify whether impaired function triggers adverse pregnancy outcomes. Therefore, we need to identify other chemicals that cause placental dysfunction and evaluate their effects on pregnancy outcomes, allowing us to verify whether specific placental dysfunction causes adverse pregnancy outcomes.
We thank Dr. Daisuke Okuzaki (Osaka University) for his technical assistance in RNA-seq analysis. This research was partially supported by the Platform Project for Supporting Drug Discovery and Life Science Research (Basis for Supporting Innovative Drug Discovery and Life Science Research (BINDS)) from the Japan Agency for Medical Research and Development under Grant Number JP22ama121054.
FundingThis study was supported by a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (no. 22H03749 to K.H.) and by a grant from the Ministry of Health, Labor and Welfare of Japan (no. 21KD1002 to Y.T.).
Conflict of interestThe authors declare that there is no conflict of interest.
