Enhanced HIF2α expression during human trophoblast differentiation into syncytiotrophoblast suppresses transcription of placental growth factor

Placental growth factor (PlGF), abundantly produced from trophoblasts is involved in placental angiogenesis. The regulatory mechanism of its expression is poorly understood. Hypoxia inducible factors (HIFs) are centrally involved in the modulation of cellular function in response to low oxygen conditions. This study aimed to clarify HIF1α and HIF2α expression patterns during cytotrophoblast differentiation into syncytiotrophoblast and the impact of any changes on PlGF expression. HIF proteins were induced remarkably under low oxygen condition (2%). HIF1α expression decreased and HIF2α expression increased when syncytialization of cultured cytotrophoblasts is progressed. Those expression changes of HIF proteins in the process of in-vitro syncytialization was congruent with the immunohistochemical findings in preeclamptic placenta as well as uncomplicated placenta. Low oxygen condition was also associated with reduced PlGF production in syncytializing primary cells and BeWo choriocarcinoma cells. Small interfering RNA-mediated HIF2α knockdown in BeWo cells abrogated hypoxia-associated decreases in PlGF secretion; HIF1α silencing had no significant effect on PlGF secretion. In summary, HIF2α, rather than HIF1α, is most affected by reduced oxygen level during syncytialization and increases in HIF2α trigger a reduction of PlGF production. Our findings suggest new and important connections between HIF proteins and PlGF pathways in the regulation of placental angiogenesis.

conditions, HIF proteins are stabilized and enhance the expression of target genes 8 . Several biological characteristics of HIF1α and HIF2α differ. HIF1α expression is relatively ubiquitous and appears to be closely linked to the maintenance of homeostasis. HIF2α expression, however, is restricted to specific tissues and cell types 9 . As recently described in macrophages, the two HIF proteins can differ in their regulation of the vascular endothelial growth factor (VEGF) system. Here, HIF1α promoted VEGF expression, whereas HIF2α promoted soluble fms-like tyrosine kinase 1 (sFlt-1) expression under the presence of cytokine granulocyte-macrophage colony-stimulating factor 10 . Past studies have shown HIF proteins are essential in successful embryo implantation and placentation. Gene deletion of HIF1α or HIF2α resulted in impaired trophoblast differentiation, placental malformation and fetal death in mice 11 . The gene product of the von Hippel-Lindau (VHL) gene is essential for degradation of HIF proteins 12 . Uncontrolled enhancement of HIF proteins by VHL gene deletion causes deranged vasculogenesis in the murine placenta 13 . These observations suggest HIF proteins are critical for trophoblast differentiation and the regulation of placental vascular formation. In the human placenta, the significance of HIF proteins in trophoblast differentiation and placental angiogenesis has not been fully elucidated.
The VEGF system is central to the regulation of the angiogenic processes. In this system, vascular development and function are regulated by a balance between proangiogenic factors, represented by VEGF and placental growth factor (PlGF), and anti-angiogenic factors such as sFlt-1. Through its binding to VEGF and PlGF, sFlt-1 inhibits the biological actions of those proangiogenic factors 14 . Gestational stage specific regulation of proangiogenic and anti-angiogenic balance is critical especially in the early stage of pregnancy. Dominant action of anti-angiogenic factors is estimated to be vital for proper implantation and embryo development, while the shift to proangiogenic activities are required in the later stage of placental vasculature development 15 . Ample production of proangiogenic and anti-angiogenic members of the VEGF family by trophoblast cells aids in the establishment of appropriate fetomaternal circulation. The molecular mechanisms to regulate those angiogenic molecules in trophoblast cells are not fully elucidated.
In this study, we aimed to investigate the expression patterns of HIF1α and HIF2α in human trophoblast cells during differentiation from cytotrophobalsts (CTBs) to STB and the impact of these HIF proteins on PlGF expression.

Results
In-vitro syncytialization of isolated cytotrophoblasts. After seeding the isolated cytotrophoblasts, spontaneous cell fusion occurred similarly to syncytialization in vivo. Morphological change into multinuclear cell clusters was evident at 72 hours. By 96 hours in culture, the culture wells were covered by a monolayer of syncytialized trophoblast. It is known that syncytiotrophoblast is characterized by abundant secretion of human chorionic gonadotropin (HCG) 16 . After seeding, culture supernatants were collected and new medium was replenished every 24 hours up to 96 hours in culture. HCG concentrations in the collected supernatants were measured using ELISA to confirm that the observed morphological changes were accompanied by secretory properties corresponding to syncytiotrophoblast. HCG secretion increased rapidly over time, particularly after 72 hours (n = 5-6, respectively, Fig. 1A). The acquisition of HCG secretion capacity suggested that the stages of cell differentiation in our cultured trophoblast model corresponded to cytotrophoblasts during the period of 0 to 24 hours in culture and syncytiotrophoblast during the period of 72 to 96 hours in culture.
We next examined the impact of the oxygen environment on our cultured trophoblasts. Cultures were exposed to 2% O 2 for 24 hours at distinct periods after seeding, from 0 to 24 hours (to model cytotrophoblast cells prior to fusion) and from 72 to 96 hours (to model syncytiotrophoblast after cell fusion). No remarkable alteration in the cell morphology and the syncytialization status was observed after 2% O 2 exposure in compared with the cells cultured in ambient air. The culture medium was replenished at the beginning of low oxygen exposure and was collected at the end of the exposure. HCG secretion during 2% O 2 exposure periods was compared with that from the cultured trophoblasts under 20% O 2 condition (n = 4, respectively). There were no significant differences in HCG secretion levels between 2% O 2 and 20% O 2 conditions in either of the exposure period groups (Fig. 1B). These findings suggested that O 2 concentration did not affect HCG production capacity in cytotrophoblasts or syncytiotrophoblast.
Alternations in HIF1α and HIF2α expression during the process of syncytialization. HIF1α and HIF2α expression patterns during trophoblast differentiation were analyzed. Isolated trophoblasts were incubated for 0, 24, 48 and 72 hours in ambient oxygen conditions (20% O 2 ) and were further incubated under 2% O 2 or 20% O 2 conditions for the final 24 hours. HIF1α and HIF2α proteins expression levels were consistently low under 20% O 2 both in cytotrophoblast cells at 24 hours and syncytiotrophoblast at 96 hours of culture. HIF1α and HIF2α protein levels were increased in all cultures grown under hypoxic conditions (2% O 2) when compared to those grown under ambient oxygen conditions ( Fig. 2A). Additional analysis revealed that syncytiotrophoblast HIF1α levels at 96 hours after 24 hours exposure to 2% O 2 conditions were lower than those from cytotrophoblast cells exposed to hypoxia for 24 hours. In contrast, HIF2α levels were increased in syncytiotrophoblast compared to cytotrophoblast after 24 hours of exposure to hypoxic conditions. To further delineate the relationship between trophoblast differentiation stage and HIFα isoform expression, we examined alterations in HIF1α and HIF2α protein levels across the entirety of the 96 hours of culture. The final 24 hours in culture were in the presence of 2% O 2 . HIF1α protein induction under 2% O 2 condition progressively diminished over time as the amount of syncytialization increased, with the largest decrease noted between 24 and 48 hours. In contrast, HIF2α protein induction in response to hypoxia increased over time as syncytialization progressed (Fig. 2B). We also assessed changes in transcription of HIF1α and HIF2α in this culture model. As predicted by protein assessments, the hypoxia-induced increase in HIF1α mRNA expression in syncytiotrophoblast was significantly lower than that in cytotrophoblast stage cells after 24 hours exposure to 2% O 2 . HIF2α mRNA expression after 24 hours of 2% O 2 exposure was significantly higher in syncytiotrophoblast (Fig. 2C). These observations indicated that hypoxic HIF1α and HIF2α expression in villous human trophoblast. Immunohistochemical analysis was conducted to confirm the localization of HIF proteins in human placental tissues. HIF1α and HIF2α protein expression patterns in cytotrophoblast cells and syncytiotrophoblast were examined in three uncomplicated term placentas and three placentas complicated by early-onset PE. Representative images are shown in Fig. 3. Clear expression of HIF1α was confirmed in the cytotrophoblast cells residing at the inner surface of syncytiotrophoblast layer, but HIF1α-specific staining intensity was relatively dim in syncytiotrophoblast. Nuclear HIF2α was observed equally both in villous cytotrophoblasts and syncytiotrophoblast. These results support data from primary in vitro syncytialization experiments showing relatively enhanced expression of HIF2α compared to HIF1α in syncytiotrophoblast. These characteristics of HIF1α and HIF2α expression patterns were shared between uncomplicated and preeclamptic placentas. In comparison of staining intensities for HIF proteins, the difference between preeclampsia and uncomplicated pregnancy was not consistent among the placental samples examined.

Impact of oxygen concentration on PlGF expression levels in primary trophoblast cells and trophoblast cell lines.
In this series of experiments, we examined whether trophoblasts PlGF mRNA expression levels were affected by oxygen levels in this culture. PlGF mRNA expression was evaluated using real-time PCR in the same syncytialization model described above. Freshly isolated cytotrophoblasts or syncytiotrophoblast formed 72 hours after seeding were exposed to 2% O 2 or 20% O 2 conditions for 24 hours (n = 8-9). Under 20% O 2 conditions, PlGF mRNA expression levels were substantially higher in syncytiotrophoblast when compared with cytotrophoblast cells, linking PlGF expression to the process of in vitro syncytialization. PlGF mRNA expression was significantly suppressed in both cytotrophoblast cells and syncytiotrophoblast in response to 2% O 2 /hypoxic conditions. In summary, like HIF2α, PlGF transcription increased with syncytialization. Unlike the HIF proteins, PlGF transcription decreased under hypoxic conditions (Fig. 4A). We also examined PlGF responses to oxygen conditions in two common trophoblast cell lines, BeWo and HTR-8/SVneo. PlGF mRNA expression in BeWo and HTR-8/SVneo cells were evaluated following exposure to 2% O2 and 20% O2 for 24 hours. PlGF mRNA expression was significantly lower under 2% O 2 than that under 20% O 2 in BeWo cells but no change was noted in HTR-8/SVneo cells (Fig. 4B). The PlGF responsivity to hypoxia noted in primary trophoblast cultures was recognized in BeWo but not in HTR-8/SVneo cells.
HIF1α and HIF2α exert distinct regulatory effects on PlGF expression. We hypothesized that HIF proteins might be involved in the regulation of PlGF expression in trophoblast cells. To clarify this point, we separately suppressed HIF1α and HIF2α gene expression in BeWo cells using specific small interfering RNAs (siR-NAs). Two pairs of siRNAs were prepared for HIF1α and for HIF2α. Since expression of HIF proteins in BeWo cells at baseline is low in 20% O 2 conditions, no obvious effect of the siRNAs on HIF1α and HIF2α expression levels could be detected. (Fig. 5A). However, comparison with negative controls since the baseline expressions of target gene-specific suppression could be detected for both HIF1α and HIF2α under 2% O 2 conditions. We next used these siRNAs to assess the impact of HIF1α or HIF2α gene suppression on PlGF mRNA expression using real-time PCR. As expected, PlGF mRNA expression under 2% O 2 was lower than under 20% O 2 in BeWo cells transfected with negative control siRNAs. Both of the prepared HIF2α siRNAs (Si#3 and Si#4 in Fig. 5B) reversed this reduction in PlGF under hypoxia. PlGF expression levels following HIF2α siRNA   Finally, PlGF concentrations in culture supernatants of siRNA-transfected BeWo cells were also evaluated using ELISA. In agreement with mRNA expression results, HIF2α siRNAs reversed the suppression of PlGF secretion noted in response to hypoxic exposure. This alteration was not observed after transfection of HIF1α siRNAs (Fig. 5C).

Discussion
In this study, we evaluated the protein and mRNA expression patterns of HIF1α and HIF 2α in an in-vitro syncytialization model in which highly purified cytotrophoblasts are differentiated into syncytiotrophoblast over the course of 96 hours. HIF2α expression increased with syncytialization while HIF1α expression decreased. Immunohistochemical analysis supported these in-vitro observations, showing intense HIF2α expression in syncytiotrophoblast and preferential HIF1α expression in villous cytotrophoblast cells. Finally, we demonstrated that HIF2α negatively regulates PlGF production in BeWo cells.
To our knowledge, this is the first report to evaluate the expression characteristics of HIF1α and HIF2α during the process of syncytialization. Although past immunohistochemical studies of human placenta have described the presence of HIF proteins in trophoblast cells [17][18][19] , none have specifically addressed expression patterns during trophoblast differentiation. Variations in HIF expression during cell differentiation have been reported in macrophage biology. In response to Th1 and Th2 cytokines, HIF1α and HIF2α mRNA is differentially expressed in M1-and M2-polarized macrophages 20 . HIF2α produced by macrophages in response to Th2 cytokines promotes arginase1 gene expression which is central to the development of the M2 phenotype 20 . Reflecting their restricted antigen presenting ability and suppressive effect on T-cell activities, M2 macrophages play a critical role in immune tolerance and immune evasion of cancer cells 21,22 . These properties are similarly important in the immunobiology of placental syncytiotrophoblast, the outer layer of fetal derived tissue that must avoid allogenic attack while bathed in maternal blood. Like M2 macrophages, syncytiotrophoblast exhibits HIF2α-biased expression responses. The molecular mechanism linking HIF2α activities to immunological regulation in syncytiotrophoblast could form the basis of future investigations on the role of the maternal immune system in normal and disordered human placentation.
PlGF is a VEGF family protein which is abundantly produced in the placenta, especially in syncytiotrophoblast 23 . PlGF can stimulate placental angiogenesis through binding to Flt-1 24 , although its angiogenic impact is relatively weak compared to VEGF. Regulation of PlGF production remains poorly understood. Ahmed et al., have shown that higher oxygen tensions are associated with enhanced PlGF protein production from term placental villous explants, whereas hypoxia decreases PlGF production in BeWo cells 25 . Consistent with their work, we confirmed in this study that PlGF expression was lower under hypoxic conditions when compared to ambient oxygen levels in both cultured primary syncytiotrophoblast and BeWo cells. In contrast, reduction of PlGF expression under hypoxia was not observed in HTR-8/SVneo cells. The ability for syncytialization is preserved in BeWo but not in HTR-8/SVneo. Considering that PlGF expression was highly upregulated following the syncytialization in primary trophoblasts, poor PlGF expression response to hypoxia in HTR-8/SVneo cells might be associated with their loss of ability for syncytialization. Additionally, we demonstrated that increases in trophoblast HIF2α expression lead to a down-regulation of PlGF gene expression under 2% O 2 conditions. This is a novel link between HIF signaling and the control of PlGF production in the placenta. Since the knowledge on transcriptional regulation of PlGF gene is quite limited, the interaction of HIF2α with PlGF gene promoter is a subject for the future works. HIF1α-induced upregulation of VEGF is important to activation of angiogenesis by hypoxia 26 . PlGF is generally categorized as a proangiogenic factor, albeit a relatively weak one when compared to VEGF. Still, HIF2α-induced reductions in PlGF in response to low oxygen levels seem somewhat contradictory to the common biological phenomenon of angiogenesis in hypoxic conditions. It is possible that HIF2α-induced PlGF suppression and HIF1α-enhanced VEGF under lower oxygen condition might be particularly favorable for the rapid vascular development required for early placental development prior to unplugging of the uterine spiral arteries. In contrast, elevations in local oxygen concentrations following the establishment of inter villous flow should attenuate HIF1α protein activities, leading to diminished angiogenesis with a shift to a high PlGF/ low VEGF balance.
In previous studies on HIFα isoform expression in preeclamptic placentas, HIF2α protein levels were clearly increased when compared to unaffected placentas 27,28 . Minimal changes were noted in HIF1α expression. Our study demonstrates that HIF2α rather than HIF1α is dominantly expressed under low oxygen conditions in syncytiotrophoblast. Additionally, our immunohistochemical study demonstrated that HIF2α was dominant in syncytiotrophoblast, although no obvious enhancement in HIF protein expression levels in preeclamptic placentas was confirmed. Taken together, HIF2α upregulation previously reported in preeclamptic placentas may reflect the response of syncytiotrophoblast to local ischemia and/or oxidative stress accumulated by ischemia-reperfusion damage as a consequence of spontaneous vascular constriction in incomplete uterine artery remodeling. Further, we provide new insight into the well-known phenomenon of reduced serum PlGF levels in preeclamptic women 29,30 . It is possible that enhanced HIF2α expression in preeclamptic placenta may lead to the decreases in placental PlGF secretion in PE patients. However, this speculation might be too simplistic, since there is a controversy over whether PlGF production is reduced in preeclamptic placenta. In some past studies, reduced PlGF mRNA expression was described in the placenta complicated with PE 31,32 . In contrast, recent systematic review and meta-analysis of the placenta denies diminished placental PlGF expression in PE patients 33 . Thus lower serum PlGF concentrations in the preeclamptic women is no more than the secondary phenomenon caused by increased placental sFlt-1 secretion which reduces unbound PlGF in the serum. On the other hand, it has been reported that in the first trimester of pregnancy, lower serum PlGF without remarkable increase in sFlt-1 in women developing PE in the later stage of gestation 34,35 . Taken together, it is possible that deranged PlGF expression resulting from increased HIF2α might be involved in the pathological event of preeclampsia in the early pregnancy.
The present study demonstrates that trophoblast cell differentiation is closely associated with the expression patterns of HIFα isoforms. In addition, HIF2α expression has a negative impact on PlGF production in syncytiotrophoblast. These findings suggest a novel relationship between HIFα protein activities and the VEGF system in human placenta.

Isolation of human villous cytotrophoblasts. This study is approved by the institutional review board
of Faculty of Medicine, the University of Tokyo, and all experiments were performed in accordance with relevant guidelines and regulations (IRB number in our facility: 10979). After written informed consent was obtained, normal term placentas were collected from healthy pregnant women at 37 to 38 weeks of gestation at the time of selective cesarean section. We isolated villous cytotrophoblasts by modifying previously described protocols 36,37 . Briefly chorionic villous tissues were dissected away from the chorionic plate, basal plate and main vessels. Minced villous tissues were digested in 100 ml HBSS supplemented with 0.125% trypsin (Thermo Fisher Scientific, Waltham), 0.5 mg/ml DNase type1, 250 mg Dispase 2 (Sigma-Aldrich, St. Louis), 0.1 mM CaCl2 and 0.8 mM MgSO4 for 20 min at 37 °C. The cell suspension was collected and 5 ml FBS was added to inactivate trypsin. After passing through a 100 μm nylon filter, the collected cell suspension was layered over Percoll (GE Healthcare, Chicago) density gradients with four layers (50%, 45%, 25% and 20% Percoll layers). After the centrifugation, the floating cells between the 25 and 45% Percoll layerd were collected. Collected cells were incubated with an anti HLA-ABC antibody (Affymetrix, Santa Clara) followed by anti-mouse-IgG antibody-microbeads (Miltenyi Biotec, Bergisch Gladbach). Only HLA-ABC negative cells were collected using a Mini MACS TM separator (Miltenyi Biotec). Purified cytotrophoblasts were diluted in IMDM (GE Healthcare) supplemented with 10% fetal bovine serum (FBS), 1% Antibiotic-Antimycotic (Thermo Fisher Scientific), 200mM L-glutamine and 10ng/ml EGF and plated at a density of 1*10 6 cells/ml on 6-well plates coated with collagen type 1. In the experiments designed to examine the impact of reduced oxygen environment, trophoblasts were exposed to 2% O 2 condition (5% CO 2 /2% O 2 /93% N 2 ) in a Bio Labo Multigas Incubator (Jujifield, Tokyo, Japan) from 72 to 96 hours post-plating. Cell lines. HTR-8/SVneo, a trophoblast-derived cell line, was kindly supplied by CH Graham 38 and BeWo, a choriocarcinoma-derived cell line, was obtained from American type culture collection. HTR-8/SVneo cells were cultured in RPMI 1640 (Wako, Osaka) supplemented with 10% FBS, and BeWo cells in Ham's F12 (Wako) supplemented with 10% FBS.
Immunoassays for HCG and PlGF. HCG concentrations in culture supernatants were measured by chemiluminescence enzyme immunoassay using IMMULITE 2000 (Siemens Healthineers, Erlangen). This procedure was performed for fee by SRL, Inc. (Tokyo). PlGF concentrations in the culture supernatants were measured using a sandwich ELISA (Quantikine ELISA kit; R&D Systems, Minneapolis). Minimal detectable limits for these assays were 1.1 mIU/mL for HCG and 7 pg/ml for PlGF.
Western blot analysis. Protein was extracted from the cultured cells using the Cell Lytic ™ MT Cell Lysis Reagent (Sigma-Aldrich) and Protease Inhibitor Cocktail (Nacalai Tesque,Kyoto). The extracted protein solutions were mixed with Sample Buffer Solution without Reducing Reagent (6x) for SDS-PAGE (Nacalai Tesque) and denatured for 5 min at 95 °C. 20 µg of the total protein samples were separated on Any kD ™ Mini-PROTEAN ® TGX ™ Precast Protein Gels (BIO-RAD) and then transferred onto polyvinylidene difluoride membranes (EMD