This study was approved by the University of South Florida Institutional Review Boards (Pro00019480). Human first trimester immortalized extravillous trophoblast cells (HTR8/SV^neo) and choriocarcinoma trophoblast cells (JEG3) (ATCC, Manassas, VA) were cultured in phenol-free basal medium (DMEM/F12, Thermo Fisher Scientific, Waltham, MA) with 10% fetal bovine serum (Thermo Fisher Scientific) and 1% antibiotic-antimycotic complex (Gibco, Thermo Fisher Scientific). We chose human placental microvascular endothelial cell (PMVEC), which is an excellent in vitro model to study vascularization in the placenta (23), to evaluate fetal microvascular endothelial responses against SARS-CoV-2 S-protein, whereas human endometrial endothelial cells (HEECs) were chosen to evaluate maternal microvascular endothelial responses against SARS-CoV-2 S-protein. Frozen PMVECs is a kindly gift from Dr. Hana Totary-Jain (USF), purchased from ScienCell Research Laboratories (Carlsbad, CA). According to the manufacturer’s instruction, PMVECs are obtained from healthy pregnant women and characterized by immunofluorescence with antibodies specific to vWF/Factor VIII. As characterized previously (24), frozen HEECs were isolated from endometrial biopsies obtained from healthy women, who were not under hormonal treatment. Frozen PMVECs and HEECs were thawed and cultured in EGM-2 medium supplemented with low serum growth supplement (Gibco, Thermo Fisher Scientific) with 1% antibiotic-antimycotic complex. SARS-CoV-2 rh-S-protein was provided by BEI Resources (Manassas, VA). Human recombinant interferon gamma (rh-IFNγ) was purchased from R&D systems (Minneapolis, MN).

Confluent HTR8, JEG3, HEECs and PMVECs cultures were trypsinized and seeded in 6-well culture plates (1×10⁵ cells/well). The next day, the cells were exposed to either mock (control) or rh-S-protein at concentrations of 10, 100 and 1000 ng/ml, or 10ng/ml rh-IFNγ ± 10 ng/ml rh-S-protein in 500 µl serum free media and then shaken every 15 min to enhance rh-S-protein binding to cells at 37°C for 1 hour. Thereafter, 1500 µl fresh media with serum was added into cells. After 24 hours, plates were washed 3 times with ice-cold PBS and stored at −80°C for further RNA extraction.

Total RNAs from HTR8, JEG3, PMVECs and HEECs cultures were isolated using RNeasy Mini Kit (Qiagen, Germantown, MD) followed by DNase I treatment (Qiagen) to eliminate genomic DNA contamination. To compare endogenous expression of SARS-CoV-2 entry molecules, ACE2 and TMPRSS2 in fetal and maternal cells, previously isolated RNAs from primary cultured term cytotrophoblasts, syncytiotrophoblasts, term decidual cells, first trimester decidual cells and human endometrial stromal cells were employed (25). Reverse transcription using RETROscript kit (Ambion, Austin, TX) was performed as described (26) and qPCR performed using TaqMan gene expression assays to detect gene expression levels of: 1) pro-inflammatory cytokines interleukin (IL)-1β, IL-6, and IL-8; 2) chemokines C-C motif chemokine ligand (CCL)2 -5 as well as C-X-C motif chemokine ligand (CXCL) 9 and 10; and 3) tissue factor (F3) (Applied Biosystems, Grand Island, NY, TaqMan assay ID given in Supplementary Table 1). All reactions were performed in duplicate. Expression of the target genes was normalized to β-actin levels, and the 2^−ΔΔCT method was used to calculate relative expression levels (27).

Media from HTR8 and PMVEC cultures treated with vehicle 10 or 1000 ng/ml rh-S-protein or 10 ng/ml IFNΥ ± 10 ng/ml rh-S-protein for 24 hours were collected, centrifuged and the resultant supernatants were stored at -80⁰C. Secreted IL-6 and IL-8 levels were measured using specific enzyme-linked immunosorbent assay (ELISA) kits (R&D Systems; Minneapolis, MN). Briefly, 96-well ELISA microplates were coated with a capture antibody; after blocking with 5% BSA, 1:4 diluted samples were added to the coated plates for 2 h, followed by a biotin-conjugated detection antibody. Antibody binding was measured with horseradish peroxidase-conjugated avidin along with a soluble colorimetric substrate. The absorbance was read at 450 nm using a microplate reader (Bio-Rad). Each sample was measured in duplicate

After receiving IRB approval, placental specimens from SARS-Cov-2 infected mothers (n=3) who tested positive for COVID-19 in the third trimester and gestational age-matched normal pregnancies (n=3) were obtained from Clinical Pathology Laboratories at Tampa General Hospital. 5 μm formalin-fixed paraffin embedded placental sections were processed for immunohistochemistry as described previously (28). Briefly, after deparaffinization and rehydration, paraffin-embedded sections were boiled in 10 mM citric acid solution (pH: 6.0) for antigen retrieval for 20 min and incubated in 3% H₂O₂ for endogenous peroxidase quenching for 10 min. The slides were incubated with 10% goat serum (Vector Labs, Burlingame, CA) for 30 min at room temperature, then overnight with mouse monoclonal anti-SARS-Cov-2 Spike RBD (monoclonal mouse IgG2A, clone no. 1035423, 10µg/ml dilution; R&D Systems, Minneapolis, MN). For negative control, placental sections were incubated with non-immune mouse IgG_2a in place of primary antibody at the same concentration. All sections were washed in PBS containing 0.1% Tween-20 (PBS-T) and incubated with biotinylated anti-mouse IgG antibody (1/400 dilution; Vector Labs) for 30 min. Following several rinses in PBS-T, the sections were incubated in streptavidin–peroxidase complex (Elite ABC Kit, Vector Labs) for 30 min. After washing, slides were exposed to diaminobenzidine tetrahydrochloride dehydrate (Vector Labs) as a chromogen for 3 min and counterstained with hematoxylin before permanent mounting.

Results were analyzed by One-Way ANOVA followed with a post-hoc Tukey test if normally distributed or using the Kruskal-Wallis test followed by the post-hoc Student-Newman-Keuls’ test if non-parametrically distributed using SigmaStat version 3.0 software (Systat Software, San Jose, CA), P<0.05 was considered statistically significant.

To elucidate SARS-CoV-2 cell tropism in the placenta, we first investigated the expression levels of the cell entry receptor, ACE2, and priming protease, TMPRSS2 at the maternal-fetal interface including primary cultured term cytotrophoblasts, syncytiotrophoblasts, term decidual cells, and first trimester decidual cells as well as human endometrial stromal cells obtained from non-pregnant women. qPCR analysis revealed that ACE2 mRNA levels are significantly higher in both cytotrophoblasts and syncytiotrophoblasts (Ct < 30) vs. term or first trimester decidual cells as well as endometrial stromal cells (Ct > 33) (Figure 1A). While both trophoblastic cells displayed weak TMPRSS2 expression (Ct > 33), maternal stromal decidual cells did not express TMPRSS2 (Figure 1B). Subsequently, we compared ACE2 and TMPRSS2 levels in trophoblastic cell lines HTR8, and JEG3, and detected significantly higher ACE2 (~ 4.9-fold) and TMPRSS2 (~9-fold) levels in JEG3 than HTR8 cells (Figure 1C). In addition to trophoblast cells, we also compared ACE2 and TMPRSS2 levels in fetal and maternal endothelial cell types PMVECs from placental specimens and HEECs, respectively to explain COVID-19 severity in pregnancy. qPCR results revealed weak ACE2 mRNA levels in both cell types and a slightly higher in HEECs (Mean ± SEM; 1.02 ± 0.13) compared to PMVECs (0.52 ± 0.08) (Figure 1D). In contrast, TMPRSS2 levels were undetectable in both endothelial cell types.

To mimic inflammatory changes induced by SARS-CoV-2 in the placenta, and to explore virus-induced pregnancy outcomes, HTR8 and JEG3 cell lines as well as PMVECs and HEECs were treated with 10, 100, 1000 ng/ml rh-S-protein for 24 hours and the expression of pro-inflammatory cytokine genes IL-1β, IL-6, and IL-8 measured by qPCR. In HTR8 cells, rh-S-protein treatment significantly increased levels of IL-1β and IL-6 vs. controls, displaying a clear dose-response effect to increasing concentrations of rh-S-protein, whereas only 1000 ng/ml rh-S-protein significantly induced mRNA expression of IL-8 (Figure 2A). In contrast, in JEG3 cells, IL-1β and IL-8 mRNA levels were undetectable in all groups (not shown); additionally, no rh-S-protein concentration altered basal IL-6 levels (Figure 2B). In PMVECs, IL-1β and IL-6 levels were significantly elevated by rh-S-protein again with a clear dose response evident, while only highest concentration of rh-S-protein significantly induced IL-8 levels (Figure 2C), similar to the pattern seen with HTR8. However, in HEECs, IL-1β levels increases did not attain statistical significance; while IL-6 levels were significantly induced by 100 and 1000ng/ml of rh-S-protein, and IL-8 levels were only elevated by the highest concentration of rh-S-protein (Figure 2D). Similarly, ELISA analysis revealed significantly higher levels of IL-6 and IL-8 secretion in culture supernatants of HTR8 (Figure 2E) and PMVECs (Figure 2F) treated with 1000 ng/ml rh-S-protein vs. control, validating S-protein mediated increase in IL-6 and IL-8 transcription in HTR8 and PMVECs. However, 10 ng/ml rh-S-protein treatment did not induce secretion levels of either cytokine in these cell types. These findings indicate: 1) low concentrations of S-protein appear sufficient to induce of IL-1β and IL-6 levels, but a higher concentration is required to induce IL-8 levels in HTR8, PMVECs and HEECs; 2) only higher concentration of rh-S-protein induces secretion of IL-6 and IL-8 levels in HTR8 and PMVECs; and 3) there is a clear inflammatory response to COVID-19 in vascular endothelial cells, potentially contributing to viral pathogenesis in pregnant women.

Following confirmation that rh-S-protein treatment resulted in an alteration of pro-inflammatory gene expression, we compared mRNA expression levels of chemokines CXCL9, CXCL10, CCL2, and CCL5 in HTR8, JEG3, PMVECs, and HEECs cultures. The qPCR analysis revealed that compared to mock-treated cells, levels of CCL2 and CCL5 in both HTR8 (Figure 3A) and JEG3 cells (Figure 3B) were not altered by any rh-S-protein exposure. Moreover, levels of CXCL9 and CXCL10 were undetectable and not induced by any rh-S-protein concentration in both HTR8 and JEG3 cells. In contrast, in PMVECs cultures, CCL2 levels were significantly induced by all rh-S-protein concentrations in a dose-response fashion, while CCL5 levels were only significantly increased at concentrations of rh-S-protein of 100 or 1000 ng/ml and whereas basal CXCL9 and CXCL10 levels were unchanged (Figure 3C). Again, in contrast to PMVECs, rh-S-protein did not affect expression of these cytokines in HEECs (Figure 3D).

Thrombotic complications are frequent in COVID-19 patients and are associated with disease severity and mortality (29). F3 that is the primary initiator of coagulation is not normally expressed by endothelial cells or trophoblast, though its expression can be induced by proinflammatory cytokines. Thus, we evaluated expression levels of F3 in cells treated with 10, 100, and 1000 ng/ml rh-S-protein to explore potential etiologies of placental thrombosis in SARS-CoV-2 infected pregnant women. After 24 hours treatment, qPCR results displayed no significant difference in mRNA expression levels of F3 in either HTR8 (Figure 4A) or JEG3 (Figure 4B). However, 1000 ng/ml of rh-S-protein significantly increased F3 levels in PMVECs (Figure 4C), whereas all concentrations induced F3 levels in HEECs compared to control groups (Figure 4D).

To explore whether the immunological state of pregnancy can promote adverse pregnancy outcomes in SARS-CoV-2 infected pregnant women, HTR8, JEG3, PMVECs and HEECs cultures were treated with 10 ng/ml rh-IFNγ ± 10 ng/ml rh-S-protein since a significant positive correlation was reported between IFNγ levels and disease severity in pregnant women (30). The expression of pro-inflammatory cytokines and chemokines levels were then measured. The qPCR analysis revealed that compared to mock-treated controls: 1) in HTR8 cells, rh-IFNγ alone significantly increased mRNA levels of the pro-inflammatory cytokines IL-1β, IL-6, IL-8. However, the combination of rh-IFNγ with rh-S-protein further elevated IL-8 levels, but not IL-1β or IL-6 levels (Figure 5A); 2) in JEG3 cells, rh-IFNγ alone or in combination with rh-S-protein did not alter IL-6 levels, whereas IL-1β and IL-8 levels were undetectable (Figure 5B); 3) in PMEVCs, rh-IFNγ alone significantly increased IL-1β, IL-6 and IL-8 levels, which are further induced by the combination of rh-IFNγ and rh-S-protein (Figure 5C); and 4) rh-IFNγ alone enhanced IL-6 levels, and addition of rh-S-protein did not further induce cytokine mRNA levels in HEECs (Figure 5D). Further analysis by ELISA revealed that HTR8 (Figure 5E) and PMVEC cultures (Figure 5F) treated with 10 ng/ml IFNΥ displayed significantly higher IL-6 and IL-8 secretion levels, which are the further increased by the addition of rh-S-protein (Figures 5E, F).

Following the same protocol with Figure 5, we evaluated the impact of rh-IFNγ on chemokine expression in these cells by qPCR and noted that mRNA levels of CXCL9, CXCL10 (Figure 6A), and CCL2 and CCL5 (Figure 6B) were significantly enhanced by rh-IFNγ, but not further altered by addition of rh-S-protein in HTR8. In JEG3 cells, only CCL5 levels were induced by rh-IFNγ, but again not further increased by adding rh-S-protein, whereas CCL2 levels did not attain significance in any incubation condition (Figure 6C). Interestingly, in JEG3 cells, CXCL9 and CXCL10 levels were undetectable in both the control and rh-IFNγ treatment groups. Finally, rh-IFNγ significantly induced expression of CXCL9, CXCL10 (Figures 6D, F), and CCL2 and CCL5 (Figures 6E, G) in both PMVECs and HEECs, and rh-S-protein further elevated their expression in PMVECs, but not HEECs (Figures 6D–G).

We next investigated whether enhanced rh-IFNγ contributes to the risk of thrombosis by measuring F3 expression in trophoblastic HTR8 and JEG3 as well as endothelial PMVECs and HEECs cultures. Figure 7 shows that F3 mRNA levels are not induced by either rh-IFNγ alone or in combination with rh-S-protein in both HTR8 (Figure 7A) and JEG3 cells (Figure 7B). However, compared to control, F3 mRNA levels were significantly higher in PMVECs treated with 10ng/ml rh-IFNγ and the combination of rh-IFNγ +10 ng/ml rh-S-protein further increased F3 expression in PMVECs (Figure 7C). Interestingly, F3 expression was not induced by either rh-IFNγ alone or in combination with rh-S-protein in HEECs, suggesting that interferon blocked spike protein induction of tissue factor (Figure 7D).

Analysis of placental sections immunostained with anti-SARS-CoV-2 spike RBD revealed that endothelial cells as well as trophoblast layer displayed immunoreactivity in placental villi obtained from mothers who tested COVID-19 positive in the third trimester (Figure 8), whereas no reaction was detected in either endothelial cells or other cells in the placental villi obtained from gestational age-matched normal pregnancies (Figure 8).