Vasoactive Intestinal Peptide induces glucose and neutral amino acid uptake through mTOR signalling in human cytotrophoblast cells

The transport of nutrients across the placenta involves trophoblast cell specific transporters modulated through the mammalian target of rapamycin (mTOR). The vasoactive intestinal peptide (VIP) has embryotrophic effects in mice and regulates human cytotrophoblast cell migration and invasion. Here we explored the effect of VIP on glucose and System A amino acid uptake by human trophoblast-derived cells (Swan 71 and BeWo cell lines). VIP activated D-glucose specific uptake in single cytotrophoblast cells in a concentration-dependent manner through PKA, MAPK, PI3K and mTOR signalling pathways. Glucose uptake was reduced in VIP-knocked down cytotrophoblast cells. Also, VIP stimulated System A amino acid uptake and the expression of GLUT1 glucose transporter and SNAT1 neutral amino acid transporter. VIP increased mTOR expression and mTOR/S6 phosphorylation whereas VIP silencing reduced mTOR mRNA and protein expression. Inhibition of mTOR signalling with rapamycin reduced the expression of endogenous VIP and of VIP-induced S6 phosphorylation. Our findings support a role of VIP in the transport of glucose and neutral amino acids in cytotrophoblast cells through mTOR-regulated pathways and they are instrumental for understanding the physiological regulation of nutrient sensing by endogenous VIP at the maternal-foetal interface.


Results
VIP induces specific glucose uptake in cytotrophoblast cell lines. We performed a sensitive assay for measuring glucose uptake in single, living cells based on direct incubation of cells with a fluorescent D-glucose analogue, 2-NBDG, followed by flow cytometry detection. The incorporation of the fluorescent probe in the human cytotrophoblast-derived cell line Swan 71 was analysed as the percentage of fluorescent cells (Fig. 1a) or the mean of fluorescence intensity (Fig. 1b,c) in the absence/presence of phloretin, an inhibitor of glucose transport. The assays were carried out at different times or with different concentrations of D-glucose. Figure 1a-c show rapid saturable specific glucose uptake in Swan 71 trophoblast cells. The linearity between 2 and 10 min (Fig. 1b) provided a suitable time of 10 min to evaluate the effect of different stimuli throughout. Next, we investigated the effect of VIP on glucose uptake in human cytotrophoblast cells. As shown in Figure 1d, VIP induced, in a concentration-dependent manner, a significant increase of 2-NBDG uptake in first trimester cytotrophoblast cells. Indeed, the effective concentrations of VIP ranged between 10 and 100 nM so we next assayed 50 nM VIP or 50 ng/ml leukemia inhibitory factor (LIF), a gp130 family cytokine involved in the placentation process 43 that induces glucose uptake in mouse skeletal muscle 44 . Figure 2a shows that a short incubation with VIP increased 2-NBDG uptake to a similar extent than LIF in both Swan 71 and BeWo cell lines. Consistent with the results of the flow cytometry assays, VIP also increased 2-NBDG incorporation as measured by fluorescence microscopy in both cell lines (Fig. 2b). On the basis that trophoblast cells synthesize VIP 29,35 and that VIP regulates cytotrophoblast cell function through autocrine loops 29 , loss of function experiments were next carried out to investigate the relevance of endogenous VIP in the regulation of trophoblast glucose uptake. We performed knocking down experiments using a VIP siRNA in Swan 71 and BeWo cells for 72 h, and after confirming the decrease of VIP expression to less than 50% compared to scramble-transfected cells as previously 29,31 , 2-NBDG incorporation was measured by flow cytometry. The results shown in Figure 2c indicate that VIP silencing reduced 2-NBDG uptake in both cell lines suggesting the involvement of endogenous VIP in glucose transport in trophoblast cells.

VIP increases GLUT1 and GLUT3 glucose transporter expression in cytotrophoblast cell lines.
Next, we investigated whether VIP could also regulate the mRNA and protein expression of the main glucose transporters expressed in human placenta GLUT1 (SLC2A1) and GLUT3 (SLC2A3). RT-qPCR, Western blot and flow cytometry assays were carried out in human trophoblast cell lines Swan 71 and BeWo cultured in the absence/presence of 50 or 100 nM VIP for 6 h. As shown in Figure 3a (left and right panel), VIP induced mRNA and protein expression of GLUT1 glucose transporter in both cell lines. GLUT3 mRNA was also induced by VIP, however, no changes were found in GLUT3 expression at the protein level (not shown).
VIP-induced glucose uptake is mediated through PKA, MAPK, PI3K and mTOR signalling. In order to elucidate the mechanisms involved in the effect of VIP for the regulation of D-glucose transport, we studied the main signalling pathways described downstream VIP activation of its specific receptors VPAC1 and VPAC2: protein kinase A (PKA), protein kinase C (PKC) and MAP kinases (MAPK). 2-NBDG incorporation assays showed that the pre-treatment of trophoblast cells with the PKA inhibitor H89 (10 µM) or the MEK inhibitor PD 98059 (50 µM) prevented the increase of 2-NBDG uptake induced by 50 nM VIP, whereas the PKC inhibitor STP did not prevent VIP effect (Fig. 4a). It has been reported that PI3K induces mTOR complex activation 19,45 and that mTOR complex is activated by cAMP and would induce the GLUT1 translocation to the plasma www.nature.com/scientificreports www.nature.com/scientificreports/ membrane 46,47 . Based on this evidence, we analysed the role of PI3K and mTOR activation in the regulation of glucose uptake induced by VIP. As shown in Figure 4b, both the PI3K inhibitor Ly294502 and the mTOR inhibitor rapamycin prevented the increase of 2-NBDG uptake induced by VIP.
VIP induces amino acid uptake by System A sodium-coupled neutral amino acid transporters (SNAT) in human cytotrophoblast cells. Based on the effect of VIP on glucose uptake and the involvement of mTOR, we next evaluated the effect of VIP on the activity of System A amino acid transporters in Swan 71 and BeWo cell lines. The incorporation of 14 C-MeAIB, a specific substrate of System A, was measured in sodium-containing and sodium-free (for non-specific transport) Tyrode's solution (Fig. 5a). System A activity was calculated in all the next experiments by subtracting 14 C-MeAIB uptake in both conditions. As shown in Figure 5b, VIP induced a significant increase of System A activity in Swan 71 and BeWo cells. Also, VIP increased the expression of SNAT1 but not of SNAT2 mRNA, the two main System A transporters present in human placenta (Fig. 5c,d).

Cross-regulation between VIP and mTOR in human trophoblast cell lines.
On the basis that the effects of VIP on glucose uptake involved the activity of mTOR and that VIP-knocked down trophoblast cells incorporated less glucose than control scramble-transfected cells, we explored the link between mTOR and VIP in these cells. VIP induced mTOR expression (Fig. 6a) and phosphorylation (Fig. 6b). In line with the involvement of VIP-activated mTOR phosphorylation pathways in cytotrophoblast cells, VIP also induced the phosphorylation of S6, a substrate downstream mTOR activation, and the effect was inhibited by rapamycin (Fig. 6c). A role of endogenous VIP in cross-regulatory loops with mTOR was also tested. VIP-silenced cells presented lower expression levels of mTOR mRNA and protein (Fig. 6d) and conversely, the inhibition of mTOR activity with rapamycin resulted in a reduced expression of VIP (Fig. 6e).

Discussion
Our results point to a novel role of VIP in the regulation of nutrient transport in cytotrophoblast cells through a mechanism that involves mTOR. The following observations support this proposal: First, VIP increased specific phloretin-sensitive glucose uptake in two human cytotrophoblast cell lines. Deficiency of VIP expression in trophoblast cells resulted in a reduced uptake of glucose whereas VIP induced GLUT1 mRNA and protein expression and GLUT3 mRNA expression in both cells lines. Rapamycin, an mTOR inhibitor, prevented VIP-induced glucose uptake that was also dependent on PKA, MAPK and PI3K but not PKC signalling. Second, VIP increased Na + -dependent amino acid transport as well as SNAT1 mRNA and protein expression in cytotrophoblast cells. Third, VIP induced mTOR expression, phosphorylation, and mTOR downstream substrate S6 phosphorylation www.nature.com/scientificreports www.nature.com/scientificreports/ as well, which was blocked by rapamycin. In parallel, cells silenced in VIP expression presented reduced mRNA and protein expression of mTOR. Moreover, rapamycin reduced VIP expression suggesting that the regulation of VIP levels and VIP-activated signalling is downstream mTOR activation in cytotrophoblast cells.
Cumulative evidence supports the role of VIP as an endogenous regulatory peptide during early pregnancy. In human pregnancy, cytotrophoblast and syncytiotrophoblast cells of first and third trimester placenta express VIP 35 . VIP induces hCG and progesterone release in the human trophoblast JEG-3 cell line and in human trophoblast primary cultures 35 . In the Swan 71 and the HTR8/SVneo human cytotrophoblast cell lines, VIP promotes trophoblast cell migration and invasion through CRE-PKA pathways 29 . In line with an active role of endogenous VIP in cytotrophoblast cell regulation, VIP knocked-down cells presented lower expression of metalloproteases 31 , reduced basal and LIF-mediated migration 29 as well as they failed to regulate the functional phenotype of  www.nature.com/scientificreports www.nature.com/scientificreports/ maternal leukocytes 32 . Regarding trophic effects of VIP, it displayed an embryotrophic effect on mouse embryos explanted with their yolk sacs at day 9.5 39 whereas lower birth-weight pups were born from mothers deficient in VIP compared to wild type mice 48 . Also, an association of reduced levels of VIP expressed by trophoblast cells and reduced foetal growth has been recently demonstrated in a VIP deficient murine model 42 . Reduced foetal weight at gestational day 14.5 along with several structural abnormalities of the placenta was reported in pregnancies with VIP-deficient trophoblast cells 42 . Interestingly, the reduced foetal weight observed was not associated with lower placental weight suggesting a metabolic rather than trophic effect of the polypeptide in the placenta 42 .
Based on the multiple energy-requiring functions of the cytotrophoblast cells to support the syncytium, here we tested the hypothesis that VIP contributes to placental metabolism through the modulation of glucose and amino acid uptake by cytotrophoblast cells. Our results clearly support this proposal since VIP increased the transport of glucose and amino acids in two human trophoblast-derived cells lines (Swan 71 and BeWo) that share many characteristics of cytotrophoblast cells and having the BeWo cells the ability to syncytialize. The concentration range of VIP to modulate glucose and amino acid uptake shown here is the same as that recently reported for the migration and invasion of human trophoblast cell lines 29 as well as for trophoblast migration in primary cultures of human first trimester placental explants 31 . Accordingly, VIP concentrations used here are consistent with VPAC1 and VPAC2 receptor-mediated effects, both subtypes expressed on Swan 71 and BeWo cytotrophoblast cells 31,37 . Moreover, VIP not only induced glucose and amino acid transport but also increased www.nature.com/scientificreports www.nature.com/scientificreports/ the expression of GLUT1 mRNA and protein and GLUT3 mRNA glucose transporters. Similarly, VIP increased the expression of SNAT1 amino acid transporter isoform. Maternal circulating glucose is the primary energy substrate for foetal and placental growth 3 . Three isoforms of the GLUT family, GLUT1, GLUT3 and GLUT4, are involved in trophoblast glucose uptake, with a higher relevance of GLUT1 and GLUT3 as they are constitutively expressed on trophoblast membranes 24,49 . GLUT1 is the only isoform abundantly expressed in early pregnancy and at term 49 . On the other hand, SNAT1 and SNAT2 activity increases throughout gestation in animal models and in human placenta from first trimester to term 50,51 . Both insulin and Insulin like growth factor 1 (IGF1) have been appointed as the major extracellular signals that promote foetal growth through increasing the transport of glucose and amino acids in trophoblast cells 8,[52][53][54] . The rapid effect of VIP on glucose uptake seems to be direct and not mediated by insulin or IGF1 release since VIP is added only for a 20 minutes time lapse. However, the fact that VIP induced GLUT1 and SNAT1 expression in cytotrophoblast cells as shown here strongly suggests that VIP acting as a growth factor might enable further long lasting adaptive activation of glucose and amino acid transport in trophoblast cells. On the basis that VIP is detected at 2-15 pM concentrations in the serum of pregnant women and in cord blood 55 , it is likely that VIP targets placental cells throughout pregnancy and, despite its short half-life, it might rapidly trigger receptor-mediated signalling that sustain metabolic and trophic functions at the maternal-placental interface.
VIP receptors are coupled primarily to Gs increasing cAMP and activating PKA, with crosstalk signalling in parallel or downstream cAMP that involves NOS 56 , PKC 57 , phosphatidylinositol 3-kinase 58 , MAPK 59-63 , JAK/STAT and NF-kB 64,65 . Our results in cytotrophoblast cells indicate that phosphorylation pathways mediated by PI3K, MAPK, PKA, mTOR but not by PKC activation are involved in VIP-induced glucose uptake. Phosphorylation mediated by PI3K and mTOR is implicated in glucose uptake by brown fat cells and lung adenocarcinoma cells 46,47 . In turn, in BeWo trophoblast cells the transport of glucose induced by resistin involves MAPK-mediated phosphorylation 66 . mTOR integrates multiple hormonal, stress and energy signals involved in foetal and placental growth. Both glucose and amino acid transport are regulated by the mTOR pathway in response to maternal signals like IGF1, insulin and leptin 14,67-69 , all displaying a stimulatory effect, or by inhibiting signals like adiponectin and hypoxia 17,68,70 . In cultured primary human trophoblast cells, the stimulation of System A activity by insulin and IGF1 was shown to depend on mTOR signalling 8 , similar to the stimulatory effect of VIP on glucose uptake shown here for two human trophoblast-derived cell lines.
Finally, a noteworthy observation is that VIP and mTOR signalling appeared mutually regulated in cytotrophoblast cells: VIP induced mTOR expression and mTOR/S6 phosphorylation. Silencing VIP expression down-regulated the levels of mTOR mRNA and protein and conversely, inhibiting mTOR activity with rapamycin reduced VIP protein expression and VIP-induced phosphorylation of S6, a substrate downstream mTOR activation. These observations along with the reduced glucose uptake in VIP silenced cells or upon blocking mTOR and the main VIP signalling cascades strongly support a cross-talk of mTOR and VIP signalling pathways in cytotrophoblast cells. Thus, the levels of VIP expressed by cytotrophoblast cells as well as their responsiveness to VIP stimulation seem to fine-tune the ability of mTOR to integrate multiple signals of the milieu sensed by the cytotrophoblast. In conditions where VIP levels declined or trophoblast cells were less responsive to VIP stimulation, the uptake of glucose and amino acids would be down-regulated whereas mTOR-mediated nutrient uptake would be impaired as well, a scenario observed in several models of IUGR. A link between mTOR and VIP expression has been proposed in suprachiasmatic neurons 71 . Mice deficient in mTOR expression showed a lower expression of VIP and altered susceptibility to constant light stimulation 71 . Likewise, a functional correlation between VIP expression and mTOR signalling was demonstrated in a conditional mTOR knockout mouse model where the mTOR gene was knocked out specifically in VIP-expressing cells 72 . An erratic circadian behaviour, weakened synchronization among cells in the suprachiasmatic nucleus together with reduced olfactory sensitivity was observed in these animals. There are no reports on a relationship between VIP and mTOR in human cytotrophoblast cells, but conclusive evidence indicates that mTOR inhibition impairs trophoblast cell invasion and migration in first trimester Swan 71 trophoblast cells 12 . Decreased migration and invasion is a feature of VIP-knocked down Swan 71 cytotrophoblast cells 29 , the same cells displaying decreased mTOR expression in the present data. Taken together, these observations strongly suggest that VIP regulates cytotrophoblast cell function and nutrient transport through mTOR activation.
To our knowledge, this is the first report on the effect of VIP on nutrient uptake in cytotrophoblast cells. These effects seem to be relevant at all stages of pregnancy considering the various functions of cytotrophoblast cells that require rapid available energy sources early at placentation and beyond as a support for the syncytium. Moreover, the mutual regulation of VIP and mTOR signalling for glucose transport in cytotrophoblast cells shown here supports a central role of the cytotrophoblast to sense maternal nutrient availability and placental demands throughout pregnancy and is in line with recent reported evidence that cytotrophoblast rather than syncytiotrophoblast cells dominate glycolysis and oxidative phosphorylation in human term placenta 28 . Although this is an in vitro design and we cannot ascertain that these mechanisms operate in vivo, our present observations are instrumental for understanding the role of VIP as one of the molecules implicated in placental metabolism. www.nature.com/scientificreports www.nature.com/scientificreports/ Vip silencing. Swan 71 and BeWo cells were transfected with a VIP siRNA (Santa Cruz Biotechnology, Dallas, TX, US) as previously described 29 . Briefly, cells were grown at 60% of confluence and arrested for 3 h in Optimem ® . 100 nM VIP siRNA: Lipofectamine RNAimax (Life Technologies, Grand Island, NY, US) complex were made in Optimem and incubated for 15 min prior to addition to the cells in a drop wise manner. siRNA with a scramble sequence was used as a negative control (Scrbl).
Western blotting. Cells were seeded until subconfluence. For protein expression, 50/100 nM VIP were added for 6 h in DMEM-F12 2% FBS and cells were harvested with RIPA buffer supplemented with protease inhibitor cocktail for protein extraction and quantification. For mTOR and S6 phosphorylation assays, 50 nM VIP were added for 20 min in DMEM-F12 2% or 0% FBS. 100 µM rapamycin were added 20 min before the stimuli. Cells were harvested with RIPA buffer supplemented with protease inhibitor cocktail and 1 mM sodium orthovanadate. Western Blotting was performed as previously described 29  After washing with TBS-0.1% Tween 20, the NC was incubated with anti mouse or anti rabbit horseradish peroxidase-conjugated antibody (ThermoFisher Scientific, Massachusetts, US) for 1 h at room temperature. After washing, specific antibody signals were detected using the enhanced Chemiluminescence system ECL Plus kit and Luminescent Image Analyzer Amersham Imager 600 (GE Healthcare, UK). For stripping, NC was incubated for 15 min at room temperature in Restore Western Blot Stripping Buffer (ThermoFisher Scientific, Massachusetts, US). After washed, NC was blocked again and protocol was carried out as described above. Images were analysed using ImageJ software (NIH, Maryland, US). flow cytometry. 4 × 10 4 Swan 71 or BeWo cells were seeded until subconfluence. 50 nM VIP/100 µM rapamycin were added for 6 h or 24 h respectively in DMEM-F12 2% FBS. For VIP detection, protein secretion was inhibited by Stop Golgi incubation for 4 h. Cells were trypsinized, fixed and permeabilized with the Cytofix/ Cytoperm TM plus kit according to manufacturer's instructions (Becton Dickinson, San José, CA, US). Cells were washed twice and then incubated for 1 h with mouse anti SNAT1 or rabbit anti VIP (Abcam, Cambridge, UK), mouse anti GLUT3 (Santa Cruz Biotechnology, Dallas, TX, US) or rabbit anti mTOR (Cell Signaling, Massachusetts, US) monoclonal antibodies at room temperature. Cells were washed twice and then incubated with secondary antibody anti mouse-Alexa 488/anti rabbit-Alexa 647 (ThermoFisher Scientific, Massachusetts, US) for 40 min at room temperature. After washing twice with permwash buffer and resuspended in FACS solution, ten thousand events were acquired in a FACS Aria II cytometer ® (Becton Dickinson, San José, CA, US) and the data was analysed using the FlowJo software (http://www.flowjo.com/). qRT-PCR. mRNA expression of nutrient transporters was analysed by quantitative reverse transcription polymerase chain reaction (qRT-PCR) as previously described 29 . Briefly, 4 × 10 4 Swan 71 or BeWo cells were seeded until sub confluence. Stimuli were added for 6 h in DMEM-F12 2% FBS, then the cells were harvested with TriReagent (Molecular Research Centre, Ohio, US) and total RNA was obtained. 1 µg RNA was treated with DNAasa I following manufacturer's instructions (Sigma-Aldrich, San Luis, MO, US) to avoid DNA contamination and the samples were reverse transcribed using a MMLV reverse transcriptase, RNAse inhibitor and oligodT kit (Promega Corporation, Madison, WI, US). cDNA was stored at −20 °C for batch analysis. Samples were incubated with SYBR Green PCR Master Mix (Roche, Basilea, Switzerland) and forward and reverse primers (Table 1) in Bio-Rad iQ5 Real-time PCR system. Data was analysed using the threshold cycle (CT) method (2 −ΔΔCT method) normalized to the endogenous GAPDH gene control. System A amino acid transport activity. 14