Human peripheral blood mononuclear cells (PBMCs) were isolated from whole blood obtained from consenting healthy non‐pregnant female donors as approved by the Human Investigation Committee of the Yale University Institutional Review Board (IRB) with no written consent requirement (#20000021607).

Dulbecco’s Modified Eagle’s medium (DMEM), McCoy’s 5A medium, and RPMI-1640 medium, OptiMEM medium were purchased from Thermo Fisher Scientific (Waltham, MA, USA). DMEM/F-12 from Invitrogen (Life Technologies, Inc., Carlsbad, CA), heat-inactivated fetal bovine serum (FBS) from Sigma-Aldrich (St. Louis, MO, USA), Charcoal dextran-stripped heat-inactivated FBS from Gemini Bio Products (West Sacramento, CA, USA).

Human TNF-α (Cat. No. 300-01), GM-CSF (Cat. No. 300-03), GCSF (Cat. No. 300-23), IP-10 (Cat. No. 300-12), and RANTES (Cat. No. 300-06) were purchased from PeproTech (Rocky Hill, NJ, USA). Human TNFR1 (Cat. No. L-005197-00-0005), TNFR2 (Cat. No. L-003934-00-0005), and non-targeting control (NTC; Cat. No. D-001810-10-05) ON-TARGET plus SMART pool small interfering RNA (siRNA) were obtained from GE Healthcare Dharmacon (Little Chalfont, UK). The predesigned 384-well Cytokines and Chemokines panel assay (Cat. No. 10034472) was purchased from BioRad Laboratories (Hercules, CA, USA).

After culture and treatment, the supernatant was collected, centrifuged, aliquoted, and stored until use. The samples were thawed only once immediately prior to running the assay. The samples were run on the Luminex Multiplex Assay (R&D Systems, Minneapolis, MN) for the following cytokines and chemokines: GROα, IL-1β, IL-6, IL-8, IL-10, IL-12, IL-17, G-CSF, GM-CSF, IFN-γ, CXCL-10,CCL2, MIP-1α, MIP-1β, RANTES, TNF-α, and VEGF.

The cell lines used in the experiments were the first trimester trophoblast cell line Swan 71 (55); Immortalized human endometrial stromal cells (hESC) (56–58) and endometrial epithelial cells (HEC-1A and RL95-2) obtained from ATCC. hESC were grown in DMEM medium, Swan 71 and RL95-2 were grown in DMEM-F12 medium, while HEC-1A were grown in McCoy’s 5A medium with additional 10ug/mL insulin and cultured at 37°C with 5% CO₂. All the complete culture media were supplemented with 10% FBS, 1000 U/ml penicillin, 100 ug/ml streptomycin, 10 mM HEPES, 100 nM non-essential amino acids, and 1mM sodium pyruvate.

BLS were obtained as previously described (19), In brief, first trimester trophoblast Sw.71 cells were trypsinized and then 4,000 of these cells were added to each well of a Costar ultra-low attachment 96-well microplate (Corning Incorporated, Corning, NY, USA). Cells were incubated for 48 hr until achieved a compact morphology (spheroid). Morphology was monitored using the IncuCyte Zoom (Essen Biosciences, Ann Arbor, MI, USA). The differential cellular characteristics of trophoblast cells as a 3D model or monolayer are described in details elsewhere (19).

Human peripheral blood mononuclear cells (PBMCs) were isolated from normal non-pregnant female donors via density centrifugation. Briefly, whole blood collected in EDTA-coated vacutainers was diluted 1:2 with PBS and carefully overlayed into a tube containing equal volume of Lymphoprep™ (StemCell Tech, Vancouver, CA). Cells were spun at 2000 rpm without brake and acceleration for 25 mins at room temperature with resultant mononuclear cells harvested from the interface of plasma and Lymphoprep™ and washed twice with large volumes of PBS.

B cells were then purified by negative selection from the PBMCs using the MojoSort™ Human Pan B Cell Isolation Kit (BioLegend, San Diego, CA, USA) following the manufacturer’s instructions, resulting in highly purified population of CD19⁺ Cells (> 90% purities) (59). T cells were purified from isolated PBMCs by negative selection using the EasySep™ Human T cell Isolation Kit (StemCell Tech) following the manufacturer’s instructions resulting in purity of >95% CD3⁺ cells. CD14+ monocyte isolation was performed as previously described (60). Briefly, peripheral mononuclear cells were separated using the density centrifugation technique with lymphocyte separation medium. CD14+ monocytes were further isolated by magnetic affinity cell sorting using an EasySep™ human CD14 positive selection kit II (EasySep, Vancouver, CA, USA).

hESC and HEE cells were plated at 5×10^5 cells/100 mm dish with 10% FBS DMEM-F12 media and allowed to attach overnight. Media was then changed to 1% FBS DMEM-F12 media and incubated for 48 hrs. Cell supernatant was collected and spun down to remove any cellular debris. Cell free supernatant was aliquoted and stored at -80°C until use.

CD19⁺ B cells and CD3⁺ cells T cells were resuspended in RPMI-1640 media with 10% FBS (Gibco Invitrogen, Carlsbad, CA, USA) and seeded at 1 x 10⁶ cells/well in a 24-well flat-bottom plate and incubated at 37°C with 5% CO₂. After 24hr, cells were centrifuged at 1500 RPM for 5 minutes and changed to 1% FBS RPMI Medium for 24 hours. Cell supernatant was collected and spun down to remove cellular debris in the collection of cell-free supernatant.

CD14+ monocytes were resuspended in DMEM/F-12 media supplemented with 1% FBS and 10ng/ml GM-CSF and seeded at 2 x 10⁶ cells/well in a 6-well flat-bottom plate at 37°C with 5% CO₂. Media and GM-CSF were refreshed every 2 days for a total of 6 days. Cell supernatant was collected and spun down to remove cellular debris in the collection of cell-free supernatant.

hESC and HEC-1A cells (80,000 cells/well) were plated in a 24-well Image Lock Plate (Essen Bioscience). After 24 h, the 100% confluent cells were wounded using a semi-manual wound maker tool (61). Wound width was calculated by imaging plates using the Incucyte system (Essen Instruments), during around-the-clock kinetic imaging. After wound, cells were washed once with 1X PBS to remove dislodged cells and media was replaced with phenol red-free media containing 10% stripped heat-inactivated FBS supplemented as described in cell culture conditions. Conditioned Media (CM) was collected at 6 and 24 hr following scratch, and cells were lysed for RNA isolation.

A total of 200,000 hESC Cells or 200,000 HEC-1A were cultured in a 6-well plate for 24 hours in 1% FBS DMEM/F-12 medium. Single BLS was transferred into individual wells of a flat-bottom, tissue-culture treated 96-well plate (Corning Incorporated) containing 200 µL control or conditioned media from hESC or HEC-1A cells. Attachment and migration of the BLS to the well surface was monitored by live imaging using the IncuCyte Zoom (19). The extension of trophoblast migration was determined by measuring the radius of migrated trophoblast cells out of the center of the spheroid. The diameter of the original spheroids and the expansion (radius) were measured and quantified by Incucyte Zoom software (Sartorius, 2015A).

HEC-1A cells or hESC transfected with TNFR1 or NTC siRNA hESC cells were plated in flat bottom 96‐well plates (Corning) for 24 hours. Matrigel (Corning) was then diluted at a 1:1 ratio with 10% FBS DMEM growth medium and added on top of the HEC-1A or hESC monolayer. Following the addition of Matrigel, the 96-well plate was returned to the 37°C incubator for 30 minutes or until the Matrigel had solidified. Then, a single BLS was transferred to the top of the Matrigel layer in a volume of 150 μL DMEM‐F12 growth medium. Images were taken every 24 hours to monitor the invasion of trophoblast cells from BLS using the Echo Revolve Microscope (Echo Laboratories, San Diego, CA, USA) (19). The number of invading cells (BLS projections) were quantified by using Fiji Image J(GPL v2) for the analysis.

hESC cells that were plated to approximately 70% confluency were transfected with TNFR1, TNFR2, or NTC siRNA using DharmaFECT 1 transfection reagent as recommended by the manufacturer (GE Healthcare Dharmacon). The following day, cells were passaged for experimental endpoint.

Total RNA was harvested from hESC and HEC-1A cells using the QIAGEN RNeasy mini kit and RNase-Free DNase Kit (Qiagen, Valencia, CA, USA) according to the manufacturer’s protocol. RNA samples were quantified using the Nanodrop One spectrophotometer (ThermoFisher). The purity of the RNA was assessed through the A260/A280 and A260/A230 levels. Individual mRNA abundance was determined using TaqMan one-step RT-PCR procedures on the CFX Connect Real-Time System (Bio-Rad Laboratories). Primer-probe sets for TaqMan assays were purchased from Applied Biosystems (ThermoFisher). Relative expression values for each gene were calculated using a standard curve and the reference gene peptidylprolyl-isomerase B (PPIB). Transcript levels of specific cytokines and chemokines were measured in hESC transfected with TNFR1 or NTC siRNA and treated for 6 hr with TNF-α using a predeveloped human cytokine and chemokine 384-well panel assay from Bio-Rad Laboratories. Assay data was analyzed using the delta-delta Ct method with Bio-Rad CFX Manager 3.1 software.

200,000 hESC cells were cultured in each well of 6-well plate until reach 70% confluence and treated with vehicle or 25 ng/ml TNF-α in OptiMEM for 24 hours. Cells were collected for RNA extraction, and conditional medium were collected for the BLS treatment. BLS were obtained by plating 200,000 first trimester trophoblast Sw.71 cells with 10% FBS DMEM/F12 into each well of Costar ultra-low attachment 6-well microplate (Corning Incorporated, Corning, NY, USA) for 48 hours. 50% TNF-α Stroma Conditional medium was obtained by diluting collected conditional stroma medium with 1% FBS DMEM/F12 medium in a ratio of 1:1. As control, OptiMEM medium was diluted with 1% FBS DMEM/F12 medium in a ratio of 1:1. BLS was treated by 50% OptiMEM Medium or 50% TNF-α stromal conditional medium for 6 hours. Total RNA was harvested using the QIAGEN RNeasy mini kit (Qiagen) with an on-column deoxyribonuclease (DNase) treatment. RNA purity and yield were assessed based on the absorbance ratios of 260 to 280 nM and 260 to 230 nM. RNA sequencing was performed on a total of 6 samples, n=3 for each treatment. Samples were sequenced at the Yale Center for Genome Analysis using 1 X 75 bp strand-specific sequencing with the Illumina HiSeq 2500 sequencing system according to the manufacturer’s recommended protocol. Sequence reads were aligned to the human reference genome GRCh38 using the splice junction mapper for RNA-sequencing reads in the TopHat 2 software package (62). Sequencing depth for RNA-sequencing samples averaged 30 million reads per biological sample with 95% overall alignment rate. After alignment, read counts were determined with HTSeq-count and relative RNA abundance was measured as Transcripts Per Million Transcripts (TPM), as recommended. TPM data were analyzed by the EBSeq/Bioconductor software, based on empirical Bayesian methods, to identify differentially expressed genes (63). Differentially expressed genes were imported to the Ingenuity Pathway Analysis software program (IPA; Qiagen) and iPathwayGuide (Advaita) to identify gene ontology terms and KEGG pathways. Gene set enrichment p values were determined in IPA using the Fisher’s exact test with a cutoff of p<0.05.

All the Data for qPCR were generated by Bio-Rad CFX manager 3.1 software and calculated the Fold Change of gene expression with 2^(–ΔΔCt). Statistics were performed using the Prism 9 Software.

Data represent the average of at least three biological replicates and are presented as means ± SEM. Normality was determined by the Shapiro-wilk method. Statistical significance was determined by student’s t-test or ANOVA with Tukey’s post-hoc analysis. Statistical significance is defined as p<0.05 (*),p<0.01 (**), p<0.001 (***), or p<0.0001 (****).

Our first objective was to investigate the individual contribution of the endometrial epithelium (HEC) and stroma (hESC) cells in the regulation of trophoblast attachment and trophoblast migration. In order to achieve this objective, we used our previously described 3D in vitro model using blastocyst-like spheroids (BLS) formed by first trimester trophoblast Swan.71 cell line (Sw.71) (19) and the conditioned media (CM) from epithelial HEC-1A and stromal hESC cells (19, 56). Trophoblast cells when form the spheroid are characterized by a mesenchymal phenotype (19) which recapitulate the characteristics of the trophoectoderm. Indeed, it is well established that the trophectoderm at the time of implantation express mesenchymal markers (19, 64, 65). During the adhesion of the embryo to the endometrium, the embryonic trophectoderm upregulates the expression of genes characteristic of an epithelial to mesenchymal transition (EMT) in order to be able to migrate and invade the uterine cavity (64, 66–68). Thus, BLS were transferred into individual wells of a 96-well plate containing CM from HEC-1A, hESC, or DMEM/F-12 growth medium supplemented with low serum (1% FBS) as a control (Conditioned Media for all the cells was prepared using 1% FBS). Attachment of BLS to the plate and trophoblast migration was monitored using live cell imaging ( Figure 1A ). BLS transferred into wells containing hESC CM attached to the bottom of the plate within 2 hours, after which, the trophoblasts equally and symmetrically migrated out of the spheroid after 24h ( Figure 1A ). Importantly, attachment and trophoblast migration were not evident in wells containing low serum or HEC-1A CM. To further investigate the differential effects between stromal and endometrial epithelial cells, we performed an attachment and migration assay with RL95-2 cell line: an alternative in vitro model of endometrial epithelial cells (69). Unlike the hESC CM, the CM from RL95-2 cells did not induce BLS attachment or trophoblast migration ( Figure S1A ). To determine whether specific immune cell populations were required for the process of attachment and trophoblast migration, BLS were transferred into wells containing CM from B-cells, T-cells, or macrophages ( Figure S1B ). While some trophoblast migration was seen in BLS treated with B-cell CM, the number of migrating cells was considerably lower than that observed within the hESC CM group. Furthermore, no migration was found in the models of BLS exposed to T-cell or macrophage CM ( Figure S1B ). These results imply that endometrial stroma cells secrete specific factors with the capacity to promote trophoblast migration. Indeed, evaluation of the cytokine/chemokine content released by stroma cells at basal conditions showed the presence of several cytokines and chemokines potentially associated with trophoblast migration and invasion such as IL-6, IL-8, IL-17, GM-CSF and MCP-1 ( Table 1 ).

In order to quantify the process of trophoblast migration, we established a standardized method that could determine the rate and distance of trophoblast migration by measuring the diameter of migrated cells over time ( Figure 1B ). We observed a time-dependent migration of trophoblast cells when BLS were exposed to hESC CM. At 12 and 24 hours, the diameter of migrated trophoblasts was significantly larger when BLS were incubated with hESC CM compared to BLS incubated with HEC-1A CM ( Figure 1C ). This suggests that factors derived from stromal cells are involved in trophoblast migration.

Following migration from the trophectoderm, trophoblast cells invade the endometrium. This stage is critical for the establishment of the placenta (20, 70, 71). In order to test the hypothesis that factors secreted by endometrial stroma cells have the potential to regulate trophoblast invasion, we employed the 3D assay that evaluates trophoblast active movement through ECM (19). For this assay, a monolayer of either hESC or HEC-1A cells was seeded into individual wells of a 96-well plate and Matrigel mixed with growth medium at a 1:1 ratio was added on top of the cells ( Figure 2A ). Wells without cells below the Matrigel layer served as controls. BLS were transferred to individual wells, and trophoblast invasion was monitored by either live imaging (IncuCyte) or microscopy (ECHO Revolve). We showed that trophoblasts migrated out from BLS and invaded the Matrigel only if wells containing hESC at the bottom ( Figure 2B ). Moreover, the number and length of the trophoblast projections increased over time when hESC were present. In the absence of stromal cells, there was no migration or invasion of trophoblast cells ( Figure 2C ). In agreement with this finding, trophoblast invasion was not present in wells seeded with RL95-2 cells ( Figure S2 ). We quantified the number of invading cells under all three culture systems, confirming that only the presence of stromal cells promoted trophoblast invasion ( Figure 2D ). These data suggest that hESC is responsible for the migration and invasion of trophoblast cells through the secretion of specific factors.

To determine whether there are differences between stroma cells and differentiated decidual cells, we exposed BLS to condition media to in vitro differentiated decidual cells (72). Interestingly, we did not observed differences on the capacity to promote trophoblast migration between the stroma and decidual cells ( Figure S3 ).

To further examine whether a wound/repair process induces the expression of inflammatory cytokines that can promote trophoblast migration, we utilized an in vitro “wound-repair” model (13, 73) where we induce a “wound” in monolayer cultures of either hESC or HEC-1A cells. Cell wounding was achieved by making a grid of scratches across confluent cells growing in a 6-well plate, and followed by mRNA extraction at 0, 6, and 24 hours later. We found that transcript levels of IL-8, RANTES, CCL4, CCL2, and Csf2 were significantly upregulated in hESC 6 hours after wound induction ( Figure 3A ). The expression of IL-8, CCL4, CCL2, and CSF2 remained significantly higher at 24 hours in hESC compared to baseline levels (0 hours) or to HEC-1A cells. The wound in HEC-1A cells induced a transient up-regulation of only IL-8 at 6 hours. Interestingly, cell wounding repressed CCL2 expression in HEC-1A cells ( Figure 3A ). In addition to the above chemokines, we observed increased levels of inflammatory signaling genes, such as TNFR1 and TNFR2, in hESC at 6- and 24-hours post-wounding. However, increased expression of TNFr2 was only observed in HEC-1A cells ( Figure 3B ).

Since we observed upregulation in the expression of TNFR1 and TNFR2, in hESC, we next determined whether TNF-α signaling could be a factor regulating the expression of pro-inflammatory cytokines in hESC. Thus, hESC cells were treated for 6 hours with 25 ng/mL TNF-α or vehicle (Veh) and the mRNA expression of IL-8 and CCL4 was evaluated by qPCR. As shown in Figure S4 , transcriptional levels of CXCL8 (IL-8) and CCL4 were significantly higher following treatment with TNF-α ( Figure S4 ).

We next performed RNA-sequencing analysis of hESC treated with vehicle or 25 ng/ml TNF-α and observed 652 genes being upregulated and 184 being downregulate ( Figure 4A ). As expected, the top overrepresented biological pathways induced by TNFα were related to inflammatory signals, including cytokine-cytokine receptor interactions, NOD-like receptors, and TNF signaling pathways ( Figure 4B ). The top upregulated genes included chemokines, such as CXCL9, CXCL10, CXCL11, and CCL5 ( Figure 4C ). We also observed that genes related to dendritic cell maturation, such as RSAD2, amongst the top upregulated genes ( Figure 4C ).

For validation of the RNA-sequencing results, we performed qRT-PCR on an independent set of hESC treated for 6 or 24 hours with either vehicle or 25 ng/mL TNF-α. In agreement with our RNA-sequencing data, TNF-α treatment enhanced the transcript levels of TNFSF18 and INHBA at 6 and 24 hours ( Figure 4D ).

The pathway analysis suggested the involvement of TNFR1/2 in the regulation of cytokines/chemokines expression in hESC. In order to elucidate the specific function of these receptors, we knocked down TNFR1 and TNFR2 via siRNA. Transfection of TNFR1 siRNA significantly reduced expression of TNFR1 by 98% when compared to cells transfected with non-targeting control (NTC siRNA) but did not alter the expression of the TNFR2. Knockdown of TNFR2 was efficient and reduced levels of TNFR2 by 92%, although TNFR2 siRNA had a modest effect on the transcript levels of TNFR1 ( Figure 5A ).

We next evaluated the differential response to TNFα in hESC depleted of TNFR1 (hESC-TNFR1KD) and TNFR2 (hESC-TNFR2KD). Treatment with TNF-α resulted in a significant up-regulation of CXCL8 and CCL4 in both wt-hESC and hESC-TNFR2KD ( Figure 5B ) but not in hESC-TNFR1KD demonstrating that only TNFR1 is required for TNF-α mediated upregulation of CXCL8 and CCL4 ( Figure 5B ). We further validated the cytokine/chemokine profile regulated by TNFα/TNFR1 by using a cytokines/chemokines qRT-PCR panel array. In wt-hESC cells transfected with scrambled siRNA, TNF-α treatment induced the expression of 21 pro-inflammatory cytokines (21 of 88 cytokines/chemokines that were included in the panel array) ( Figure 5C ). However, TNF-α had no effect on the expression of these 21 cytokines in hESC-TNFR1KD. These data established the specificity of TNFα/TNFR1 cytokine/chemokine regulation in hESCs.

To further confirm the results from the cytokine/chemokine array, we determined the protein concentrations of inflammatory cytokines in the supernatant of wt-hESC treated with vehicle (Veh) or 25 ng/mL TNF-α for 24 hours. As shown in Table 1 , TNFα treatment induced a significant increase on the protein expression of cytokines and chemokines, including IL-17, CXCL8, GM-CSF, CXCL-10, TNF-α, G-CSF, VEGF, CCL5, IL-6, and CCL2 ( Table 1 ).

To determine the specificity of the TNFα receptor, we evaluated CCL2 and CCL5 in the wt-hESC and hESC-TNFR1KO treated with vehicle (Veh) or 25 ng/mL TNF-α for 24 hours. TNF-α treatment enhanced the secretion of CCL2 (538.7 ± 39.9 pg/mL vs 45. 3± 3.4 pg/mL in vehicle) and CCL5 (225.5 ± 39.6 pg/mL vs 0 pg/mL in vehicle) in the wt-hESC ( Table 1 and Figure 6 ). In contrast, the TNF-α-dependent expression of CCL2 and CCL5 was abolished in cells lacking TNFR1 ( Figure 6 ).

Next, we tested whether factors secreted by hESC treated with TNF-α may enhance trophoblast activity. Again, we utilized the 3D invasion assay described above (19). BLS were transferred individually to a 96-well plate containing Matrigel alone, a monolayer of hESC seeded below the Matrigel matrix, or a 24 hour TNF-α treated (20 ng/ml) hESC seeded below the Matrigel matrix ( Figure 7A ). The presence of hESC below the Matrigel matrix enhanced invasion of trophoblast cells from the BLS on Day 2 and 4 when compared to Matrigel alone and this effect was further enhanced in hESC treated with TNF-α ( Figure 7A ). Quantification of the invasion process reveled that the addition of TNF-α to hESC doubled the number of invading projections on day 2 when compared to vehicle treated hESC ( Figure 7B ). On Day 4, the average number of trophoblasts invading projections was significantly higher in the TNF-α treated hESC wells (131.2 ± 11.7) compared to the Matrigel alone wells (1.5 ± 2.14) or the vehicle treated hESC wells (69.8 ± 6.97) ( Figure 7B ).

To further demonstrate the role of TNFα in the above results, wt-hESC or hESC-TNFR1KD were treated with either vehicle or TNF-α (25 ng/mL) prior to seeding below the Matrigel matrix ( Figure 8A ). BLS were then transferred to wells containing wt-hESC treated with either vehicle or TNF-α and to hESC-TNFR1KO treated with either vehicle or TNF-α. Trophoblast invasion was monitored, and number and length of projections was quantified by life imaging after 2 and 4 days. At both time points, the number of invasive trophoblast cells was greater in wells plated with TNF-α treated wt-hESC compared to vehicle treatment ( Figures 8B, C ). However, in cells lacking TNFR1 (hESC-TNFR1KD), trophoblast invasion was significantly reduced on Day 2 and 4 ( Figures 8B, C ). While the addition of TNF-α to wt-hESC significantly increased the migration of trophoblast cells ( Figures 8B, C ), it did not have a comparable effect in wells containing hESC-TNFR1KO. These results suggest that TNF-α enhances the expression of hESC-secreted factors responsible for trophoblast invasion in a TNFR1-dependent manner.

To further determine that the signals necessary for the promotion of trophoblast invasion are factors secreted from hESC in response to TNF-α, we used the same 3D invasion model as described above. However, we substituted the hESC cells with conditioned media (CM). BLS were transferred to wells containing Matrigel alone, Matrigel mixed in a 1:1 ratio with hESC-CM, or Matrigel mixed in a 1:1 ratio with CM obtained from TNF-α (20 ng/mL)-treated hESC (24 hours) ( Figure 9A ). While the addition of hESC CM enhanced trophoblast outgrowth compared to that of Matrigel alone, the CM from hESC exposed to TNF-α increased the number of invading cells by approximately 3-fold on day 2 as compared to that of Matrigel alone ( Figures 9B, C ). Furthermore, at each time point, the factors secreted from hESC in response to TNF-α enhanced trophoblast invasion by 2.8-fold compared to the factors secreted from hESC under basal conditions ( Figure 9C ). We did not observe trophoblast migration in wells containing media with TNF-α alone (data not showed).

To understand the potential molecular mechanics by which hESC-derived factors enhance trophoblast migration and invasion, we performed RNAseq on BLS exposed to CM from TNF-α treated hESC. Our data shows that 229 genes were upregulated, and 322 genes were downregulated when BLS were exposed to CM produced by TNF-α treated hESC ( Figure 10A ). Signal receptor activity, cell signaling, communication, and cell receptor activities and cell motility were the main biological processes enriched by CM of TNF-α treated hESC ( Figure 10B ). KEGG pathway analysis showed an association with activation of inflammatory pathways, such as cytokine-cytokine receptor interaction, TNF-α, IL-17, and NF-κ signaling pathways ( Figure 10C ).

Since we observed IL-17 expression by hESC treated with TNFα ( Table 1 ) and IL-17 pathway as one of the KEGG pathways enriched in the trophoblast, we investigated whether IL-17 could be one of the factors produced by hESC that is necessary to promote trophoblast migration and invasion. Using gene pathway analysis, we investigated the downstream genes induced by IL-17 that could promote trophoblast migration. We identified several chemokines and cytokines known to induce cell motility that were upregulated in trophoblast cells exposed to hESC condition media, including CCL2, CXCL6/IL6, CXCL1/GROα, IL-1β, CSF2/GMCSF, CXCL8/IL8, MMP3, and MMP9 ( Figure 11A ). These genes are all involved in cell motility function ( Figure 11B ).

To test the hypothesis that IL-17 has a regulatory function on trophoblast gene expression and function, we treated trophoblast cells with IL-17 (200 ng/ml) and determined the cytokines/chemokines mRNA expression levels by qPCR. Our data showed that IL-17 significantly enhanced the mRNA expression of IL6, CXCL8, PDL1, CXCL1, CSF2/GMCSF, MMP3, and MMP9 in trophoblast cells ( Figure 12 ). Finally, we tested whether IL-17 could impact trophoblast migration and invasion. We exposed BLS to IL-17 (200 ng/ml) and monitored attachment and migration. Interestingly, addition of IL-17 to the medium promoted trophoblast attachment and migration ( Figure 13 ); a process that is not observed in the absence of IL-17 ( Figure 13 ).