Modulation of matrix metalloproteases by ciliary neurotrophic factor in human placental development

Ciliary neurotrophic factor (CNTF) is a pleiotropic cytokine that signals through a receptor complex containing a specific subunit, CNTF receptor α (CNTFRα). The two molecules are constitutively expressed in key structures for human placental growth and differentiation. The possible role of CNTF in enhancing cell proliferation and/or invasion during placental development and remodelling was investigated using HTR-8/SVneo and BeWo cells, taken respectively as cytotrophoblast and syncytiotrophoblast models. In both cell lines, treatment with human recombinant (hr) CNTF activated JAK2/STAT3 signalling and inhibited the ERK pathway. Interestingly, in HTR-8/SVneo cells, 50 ng hrCNTF induced significant downregulation of matrix metalloprotease (MMP)-1 and significant upregulation of MMP-9. Moreover, pharmacological inhibition of JAK2/STAT3 signalling by AG490 and curcumin resulted in MMP-9 downregulation; it activated the ERK signalling pathway and upregulated MMP-1 expression. Collectively, these data suggest a role for CNTF signalling in extravillous cytotrophoblast invasion through the modulation of specific MMPs.

The role of the IL-6 cytokine family in pregnancy has been extensively investigated (Ge et al. 2018;Horita et al. 2007;Markert et al. 2011). Interestingly, IL-6 and/or IL-6R dysregulation has been associated with impaired embryo implantation and abortion (Markert et al. 2011). In addition, IL-6 plays a pivotal role in regulating the proliferation and invasiveness of cytotrophoblast cells by increasing MMP-9 and MMP-2 activity (Meisser et al. 1999).
Very few studies have explored a role for CNTF in pregnancy. A study of the circulating levels of CNTF in nonpregnant and pregnant women and in women with pregnancies complicated by PE has found that CNTF levels were lower in pregnant compared with non-pregnant women, that maternal CNTF blood levels tended to decline with the progression of normal pregnancy and that they were lower in pregnancies complicated by PE than in normal pregnancies (Akahori et al. 2010). In contrast, Bienertova-Vasku and colleagues showed no difference in maternal CNTF blood levels between normal and PE pregnancies at term, but described higher CNTF levels in umbilical cord foetal blood of PE compared with normal pregnancies (Bienertova-Vasku et al. 2013). Unfortunately, in the latter study, the samples were not matched for gestational age, thus preventing all comparison. We hypothesize that the lower CNTF levels found in maternal plasma during gestation and the altered CNTF levels described in pregnancies complicated by PE can be related to placental development during gestation. This study was devised to assess the presence of CNTF in maternal blood and understand its action on placental development. First we identified the placental structures expressing CNTF and CNTFRα in first-and third-trimester human placentas and then investigated CNTF involvement in placental development by testing its possible modulation of MMP-1, MMP-2 and MMP-9 (Chen and Khalil 2017).

Tissue collection
We examined 24 placentas, 12 from pregnant women undergoing voluntary pregnancy termination at 9-12 weeks of gestation (Obstetrics and Gynaecology of San Severino Hospital, Macerata, Italy) and 12 from third-trimester pregnancies that were terminated by caesarean section (Department of Woman and Child Health, A. Gemelli Hospital, Università Cattolica del Sacro Cuore Rome, Italy and Obstetrics and Gynaecology, Department of Clinical Sciences, Polytechnic University of Marche, Ancona, Italy). In parallel, we also analysed first-trimester placentas from clinically normal pregnancies interrupted by aspiration curettage for psychosocial or medical reasons that were unlikely to affect placental structure or function. All women gave their written informed consent to the collection of placental specimens. The collection procedures were in accordance with the Helsinki Declaration of 1975, as revised in 2013. The permission of the Human Investigation Committee of the Marche Region (Italy) was obtained (protocol no. 2019.172;study ID 980;CERM no. 172).
Immediately after delivery and gross examination, three zones were identified in each placenta: a central zone near the umbilical cord insertion, a peripheral zone, i.e. the most distal from the umbilical cord, and an intermediate zone.
Two tissue samples were collected from each zone. The samples for immunohistochemistry were fixed in 4% buffered formalin at 4 °C for 12 h and routinely processed for paraffin embedding at 56° C, whereas the samples for biochemical and molecular analysis were frozen in liquid nitrogen as described previously (Fantone et al. 2020a;Tossetta et al. 2019).

Immunohistochemical analysis of firstand third-trimester placentas
Immunohistochemical staining was performed as described previously (Altobelli et al. 2017;Avellini et al. 2017). Briefly, after dewaxing, paraffin sections were rinsed in phosphate buffered saline (PBS) and incubated with 3% hydrogen peroxide for 40 min to block endogenous peroxidase. They were then subjected to high-temperature pretreatment in 10 mM citrate buffer, pH 6.0, for 5 min (CNTF) or in 100 ng/ml proteinase K (Sigma-Aldrich, St. Louis, MO, USA) for 5 min at 37 °C (CNTFRα). Sections were subsequently rinsed with PBS and incubated with normal horse serum (Vector Laboratories, Burlingame, CA, USA) diluted 1:75 in PBS for 1 h at room temperature (RT). This was followed by incubation with the primary antibodies (Table 1) diluted in PBS and then overnight incubation at 4 °C. After a thorough rinse in PBS, sections were incubated in a 1:200 biotinylated secondary antibody (Vector Laboratories) solution for 30 min at RT. Histochemical reactions were performed using Vectastain ABC Kit (Vector Laboratories) for 1 h at RT with 3′,3′-diaminobenzidine hydrochloride (Sigma-Aldrich) as the chromogen. Sections were counterstained with Mayer's haematoxylin, dehydrated and mounted in Eukitt (Kindler GmbH and Co., Freiburg, Germany). Negative controls were performed by replacing the primary or the secondary antibody with an isotype antibody: monoclonal rabbit IgG (ab172730; Abcam, Cambridge, UK) for CNTF and monoclonal mouse IgG2a (ab18415; Abcam) for CNTFRα.

Culture of HTR-8/SVneo and BeWo cell lines
To test the involvement of CNTF in proliferation and invasion processes, we used the cytotrophoblast cell line HTR-8/ SVneo (kindly provided by C. H. Graham, Queen's University, Kingston, ON, Canada) and the choriocarcinoma derived BeWo human placental cell line (kindly provided by S. Alberti, Laboratory of Cancer Pathology, CeSI-MeT, University 'G. d'Annunzio', Chieti, Italy). BeWo cells express an epithelial proteome characteristic of villous trophoblasts, whereas HTR-8/SVneo cells exhibit a phenotype characteristic of extravillous trophoblasts (Szklanna et al. 2017). Notably, BeWo cells also show a spontaneous ability to form syncytial units mimicking the syncytiotrophoblast, whereas a marked self-renewal ability makes HTR-8/ SVneo cells a good model of cytotrophoblast cells (Weber et al. 2013). Although freshly isolated cytotrophoblast can differentiate into multinucleated syncytiotrophoblast, their inability to proliferate in vitro hampers experiments and prevents their use King et al. 2000;Kohli et al. 2016;Mannelli et al. 2015).
HTR-8/SVneo cells (Graham et al. 1993) were cultured in RPMI 1640 medium (Life Technologies, Monza, Italy), whereas BeWo cells were cultured in DMEM/F12 medium. Both media were supplemented with 10% foetal bovine serum (FBS) and 100 U/ml penicillin and streptomycin (all from Gibco, Thermo Fisher Scientific, Waltham, MA, USA). All cultures were incubated at 37 °C with 95% humidity and 5% CO 2 . The culture media were changed twice a week.

Treatment of HTR-8/SVneo and BeWo cell lines with rhCNTF
Both placental cell lines express CNTFRα. After seeding at confluence (1.5 10 4 /cm 2 ), they were incubated with 20, 50 or 100 ng/ml recombinant human (rh) CNTF (cat. #257-NT-010, R&D Systems, Minneapolis, MN, USA) for 15 min, to investigate the possible signalling pathways triggered by CNTF. We focused on 3 possible pathways where the IL-6 cytokine family is commonly involved, the JAK2/STAT3, PI3K/AKT and MAPK/ERK pathways. The two cell lines were also treated with two pSTAT3 inhibitors, AG490 and curcumin (Merck, Darmstadt, Germany), to assess the specificity of CNTF action. To do so, HTR-8/SVneo cells were incubated with 100 µM AG490 or 15 µM curcumin with/without 50 ng/ml rhCNTF for 24 h in suitable media. The doses were chosen after evaluation of pSTAT3 inhibition in a dose-response curve.
Viable counts were routinely performed using the trypan blue dye exclusion test. All experiments were conducted in triplicate and were repeated at least 3 times.

Analysis of HTR-8/SVneo cell cycle after hrCNTF treatment
To examine the role of rhCNTF in cell proliferation, HTR-8/ SVneo cells were treated with 50 ng/ml hrCNTF for 24 h, suspended at 1 × 10 6 cells/ml in 1 ml PBS and centrifuged at 200 g for 5 min at RT. PBS was then removed and cells were fixed by adding 4.5 ml 70% cold ethanol (− 20 °C) overnight at 4 °C. After fixation, cells were centrifuged at 400 g for 5 min; pellets were washed 3 times with 5 ml PBS and centrifuged at 400 g for 5 min. After removing the supernatant, cells were resuspended in 1 ml propidium iodide solution (Miltenyi Biotec, Bergisch Gladbach, Germany) and incubated for 30 min at RT in the dark. Approximately 20,000 cells/sample were analysed using FacsCanto II (DB Biosciences, San Jose, CA, USA) with excitation at 488 nm and emission at 617 nm. The percentage of cells in each phase of the cell cycle was determined using FacsDiva Software (DB Biosciences).

Wound healing assay
To investigate the role of CNTF in cell motility, HTR-8/ SVneo cells grown to confluence were scratch-wounded with a sterile plastic micropipette tip. Cells were rinsed 3 times with warm medium to remove any scraped off cells in the wound and then treated/not treated with 50 ng/ml rhCNTF for 24 h. Digital images were taken at 0, 4 and 24 h. The area of the wound not occupied by cells was measured using 1 3 ImageJ software. The experiment was repeated at least 3 times.

Transwell invasion assay
Matrigel invasion assays were performed using 24-well transwell inserts with 8.0 μm pores (cat. no. 353097) coated with 100 µl of 250 µg/ml LDEV-free Matrigel (cat. no. 356234; all from Corning, NY, USA) diluted in serumfree RPMI 1640 medium according to the manufacturer's instructions. The effects of CNTF on invasiveness were assessed by pretreating HTR-8/SVneo cells with 50 ng/ml CNTF for 24 h before seeding in the transwell chambers. In the upper chambers, 5 × 10 4 cells were seeded on top of the Matrigel and supplemented with 200 µl serum-free RPMI 1640, with/without 50 ng/ml CNTF. The lower chambers contained 750 µl of the medium with a chemoattractant (10% FBS). After 24 h, non-invading cells and Matrigel were gently removed from the apical side of the membrane by scrubbing with a cotton swab, whereas the cells remaining on the lower surface were fixed with 100% methanol for 20 min and stained with 0.5% crystal violet for 15 min. Four non-overlapping fields of view per sample were captured at 10 × magnification and cells were counted using ImageJ software. Data were expressed as mean cell number in each of the four fields. The experiments were repeated 3 times in triplicate.

Statistical analysis
Data are reported as mean ± standard deviation (SD). Statistical significance was analysed using Student's t-test (p < 0.05). Statistical analyses were performed with Graph-Pad Prism (ver. 8) software.

CNTF and CNTFRα were expressed in normal human placenta and in human placenta cell models (HTR-8/ SVneo and BeWo)
In first-trimester placentas, CNTF (Fig. 1a, b) and CNTFRα (Fig. 1e, f) immunohistochemistry showed that both molecules were expressed in the cytotrophoblast layer (Fig. 1a, f: arrows), in the foetal vessels (see, for example, CNTFRα staining in Fig. 1e: V) and in the extravillous cytotrophoblast of cell columns and islands (see, for example, CNTF staining in Fig. 1b, where a column is indicated by asterisks). The syncytiotrophoblast of first-trimester placentas was negative for both CNTF and CNTFRα. In striking contrast, in third-trimester placentas, CNTF (Fig. 1c, d) and CNTFRα (Fig. 1g, h) were detected in the syncytiotrophoblast layer ( Fig. 1d, h: arrowheads), whereas cytotrophoblast and endothelial cells were negative for both molecules. CNTF (Fig. 2a) and CNTFRα (Fig. 2b) were identified in the Western blots of first-and third-trimester placentas, but showed no significant differences in expression (a', b'). By immunofluorescence analysis, HTR-8/SVneo and BeWo cells (respectively in vitro villous cytotrophoblast and syncytiotrophoblast models) were positive for both CNTF (Fig. 3 green staining) and CNTFRα (Fig. 3 red staining). In particular, CNTF was expressed in the nuclei and the cytoplasm of both cell lines (HTR-8/SVneo cells, Fig. 3a-c; BeWo cells, Fig. 3g-i), whereas CNTFRα staining was limited to the cytoplasm, especially the perinuclear region (HTR-8/ SVneo cells, Fig. 3d-f; BeWo cells, Fig. 3k-l).

CNTF did not modulate HTR-8/SVneo cell proliferation
The cell cycle phases did not exhibit significant changes between cells treated with 50 ng/ml hrCNTF for 24 h and untreated cells (Fig. 7a). This finding was further confirmed by the analysis of PCNA expression in the Western blots, Fig. 1 Immunohistochemistry of CNTF (a-d) and CNTFRa (e-h) in normal first-trimester (a, b, e, f) and third-trimester (c, d, g, h) placentas. a In first-trimester placental villi, CNTF is expressed in villous cytotrophoblast cells (arrows) whereas the syncytiotrophoblast is negative. b The extravillous cytotrophoblast cells are CNTF-positive, showing intense staining in the proximal portion of cell columns, near the villus (*). In third-trimester placental villi, only the syncytiotrophoblast (arrowheads) shows immunostaining (c, d). In (d), a terminal villus shows that the syncytiotrophoblast is CNTF-positive. In first-trimester placental villi, CNTFRa (e, f) is expressed in villous cytotrophoblast cells (arrows) whereas the syncytiotrophoblast is negative. Foetal vessels are CNTFRa-positive (e: V). g, h In third-trimester placental villi, only the syncytiotrophoblast (arrowheads) shows CNTFRa immunostaining. i, j Negative controls. b Bar = 50 mm; c, e, g, i, j bars = 100 mm; a, f, h bars = 30 mm; d bar = 25 mm which was not significantly different in treated and untreated cells (Fig. 7b).

CNTF did not affect the motility and invasion of HTR-8/SVneo cells
To assess the effect of CNTF on cell migration, HTR-8/ SVneo cells were treated with 50 ng/ml hrCNTF for 0, 4, 8 or 24 h (Fig. 8). The lack of significant differences in cell migration (Fig. 8e) suggested that CNTF is not involved in modulating villous cytotrophoblast migration. HTR-8/ SVneo cells were also pretreated with 50 ng/ml hrCNTF for 24 h, to establish whether CNTF affects invasion properties. However, there was no significant difference in invasion ability between treated and untreated cells (Fig. 9). These data suggest that MMP-1 and MMP-9 modulation by CNTF did not alter cell motility and invasion in our in vitro model.

Discussion
The placenta is a vital organ with key roles in foetal development. In the first stages of gestation, the cytotrophoblast differentiates into extravillous and villous. The villous cytotrophoblast gives rise to the syncytiotrophoblast (Huppertz 2008). In differentiated placental villi at the end of gestation, the trophoblast surface consists of a continuous outer syncytiotrophoblast layer beneath which lies a discontinuous layer of villous cytotrophoblast cells. The villous cytotrophoblast is the proliferating stem cell that can differentiate into the syncytiotrophoblast, forming the maternal-foetal transport barrier. In contrast, extravillous trophoblast cells lie outside the villi and form cell columns and islands, which are the result of different stages of proliferation and differentiation and whose main function is to anchor the placenta to the uterine wall (Moser et al. 2018). Thus, the villous and extravillous cytotrophoblast is a key structure for the development of the placental villous tree and play a central role in growth and invasion processes during gestation (Ferretti et al. 2007;Muhlhauser et al. 1995), whereas the syncytiotrophoblast is responsible for foetalmaternal exchanges (Huppertz 2008). In this study, we found that in the first trimester of gestation CNTF and CNTFRα are expressed in both villous and extravillous cytotrophoblast cells, whereas in the third trimester their expression is confined to the syncytiotrophoblast. We propose that such consistent expression in structures vital for placental growth makes CNTF a constitutive protein with a potential role in the maintenance of placental morphology. A paracrine role for this cytokine  ', b'). CNTF and CNTFRa quantities were normalized using β-actin expression. 5 ng of human hrCNTF was used as positive control (not shown). There were no significant differences in CNTF and CNTFRa protein expression between the first (I Trim) and third (III Trim) trimester (n = 12 for I Trim; n = 12 for III Trim) linked to placental tissue remodelling during gestation cannot be excluded. In particular, the presence of CNTFRα in the foetal vessels suggests a possible role of CNTF in vascular remodelling as demonstrated in other tissues including the retina, where CNTF has been shown to exert an anti-angiogenic effect on endothelial cell sprouting through activation of the JAK/STAT3/SOCS3 signalling (Bucher et al. 2016).
In addition, it is also conceivable that in some conditions, CNTF may be secreted into the maternal circulation, thus acting as a hormone, as described previously (Akahori et al. 2010;Bienertova-Vasku et al. 2013).
It has clearly been documented that CNTF can modulate MMPs in human tissues (Chen et al. 2016) and that the invasive capacity of the trophoblast is strongly associated with pSTAT3, pAKT and pERK1/2 quantities were normalized using total STAT3, AKT and ERK1/2, respectively. Data are mean ± SD (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 c' significant MMP-9 upregulation; d' lack of significant differences in MMP-2 protein levels; e' unchanged MMP-9 protein expression. MMP-1, MMP-2 and MMP-9 quantities were normalized using β-actin. Results are expressed in arbitrary units (A.U.) and reported as histogram bars. Data are mean ± SD (n = 3). *p < 0.05 its ability to produce and secrete MMPs (Fisher et al. 1989;Librach et al. 1991;Vegh et al. 1999). Here, we suggest that the presence of CNTFRα in the villous and extravillous cytotrophoblast may be related to the modulation of MMP-1 and MMP-9 through the action of CNTF. Interestingly, CNTF appears to regulate their expression through two different pathways. Our findings provide the first evidence of endogenous CNTF production, whose function seem to be the maintenance of placental tissue homeostasis. Notably, we demonstrated that MMP-1 downregulation was mediated MMP-1 (c', g') upregulation. pSTAT3 and pERK quantities were normalized using total STAT3 and ERK1/2, respectively. MMP-1 and MMP-9 quantities were normalized using β-actin. Results are expressed in arbitrary units (A.U.) and reported as histogram bars. Data are mean ± SD (n = 3). *p < 0.05, **p < 0.01 Fig. 7 The HTR-8/SVneo cell cycle was not changed by CNTF treatment. Graphs representing untreated (a) and CNTF-treated (a') HTR-8/SVneo cells. The area under the green curve represents the G0/G1 phases, the red curve the S phase and the blue curve the G2/M phases of the cell cycle. a'' Histograms represent the percentage of the cells in each cell cycle phase and show that there were no significant differences in the cell cycle phases between treated and untreated cells. b PCNA Western blots of HTR-8/SVneo cells, untreated and CNTFtreated. Right, b' histograms showing band quantification. There were no significant differences between untreated and CNTF-treated cells. Bands were normalized using β-actin. Results are expressed in arbitrary units (A.U.) and reported as histogram bars. Data are mean ± SD (n = 3) Fig. 8 Wound closure experiments. CNTF did not affect HTR-8/SVneo cell mobility. ad, a'-d' Representative photographs show the effect of CNTF treatment on cell migration at different time points (0, 4, 8, 24 h). e The graph, representing the cell-free area, shows that CNTF treatment did not affect wound closure. In treated and untreated cells, wound closure was completed in 24 h. Results are expressed in arbitrary units (A.U.) and reported as histogram bars. Bars = 100 mm. Data are mean ± SD (n = 3) by inhibition of MAPK/ERK signalling, whereas MMP-9 upregulation was mediated by activation of the JAK2/STAT3 pathway, in line with a recent report showing MMP-9 modulation by JAK2/STAT3 signalling in HTR-8/SVneo cells (Zong et al. 2016). We also demonstrated that CNTF did not affect cell proliferation, mobility and invasion, but only the balance of MMPs secretion.
So, MMPs effect was not detected in mobility and invasion processes probably due to the compensatory action of MMP-9 and MMP-1.
Collagen is a major substrate for MMPs and type I, III, IV and V collagens are expressed in human placenta (Oefner et al. 2015;Sati et al. 2008). In particular, type IV is chiefly expressed in the basal membrane of the placental villus, under the villous cytotrophoblast and around foetal vessels, whereas types I and III are localized in the interstitial stroma of foetal villi and in the extravillous trophoblast (Oefner et al. 2015;Sati et al. 2008). MMP-1 is an interstitial collagenase that degrades various interstitial collagens, including types I and III, thus playing a role in extravillous trophoblast  (a, a'). b The graph shows that there were no differences in invasion between treated and untreated HTR-8/ SVneo cells. Invading cells are expressed as mean ± SD (n = 3) of invading cells/field and reported as histogram bars. Bars = 100 µm invasion of the uterine wall. In contrast, MMP-9 is a type IV collagenase that principally degrades collagen type IV, in this respect it being similar to MMP-2 (Klein and Bischoff 2011). Notably, complete inhibition of in vitro trophoblast invasion by an MMP-9 antibody has suggested a key role for it in cell invasion processes (Librach et al. 1991). The proliferative ability of the villous cytotrophoblast and the extravillous cytotrophoblast are ensured by closeness to the basal membrane, where they form a compact cluster of cells. When villous cytotrophoblast cells leave the basal membrane, they lose their polarity and acquire an invasive phenotype that allows interaction with the surrounding extracellular matrix molecules, such as MMPs (Benirschke et al. 2006). Altogether, these data suggest that CNTF is involved in controlling the first phases of villous and extravillous cytotrophoblast migration by promoting collagen IV degradation by MMP-9 and by inhibiting cell mobility via MMP-1 downregulation. It is conceivable that MMP-9 upregulation by CNTF induces collagen type IV degradation which would be followed by a breach in the basal membrane and the switch of trophoblast cells to the invasive phenotype. The hypothesis is consistent with studies demonstrating that MMP-9 downregulation in the trophoblast plays an important role in PE onset, since a shallow extravillous trophoblast invasion hampers the transformation of uterine spiral arteries in the placental bed (Chen and Khalil 2017). In conclusion, we demonstrate that CNTF should be considered a constitutive placental cytokine acting in structures which are essential for human placental development and function, and that it plays a central role in modulating MMP-1 and MMP-9 expression by two different signalling pathways.