CCN1 (CYR61) and CCN3 (NOV) signaling drives human trophoblast cells into senescence and stimulates migration properties

ABSTRACT

During placental development, continuous invasion of trophoblasts into the maternal compartment depends on the support of proliferating extravillous trophoblasts (EVTs). Unlike tumor cells, EVTs escape from the cell cycle before invasion into the decidua and spiral arteries. This study focused on the regulation properties of glycosylated and non-glycosylated matricellular CCN1 and CCN3, primarily for proliferation control in the benign SGHPL-5 trophoblast cell line, which originates from the first-trimester placenta. Treating SGHPL-5 trophoblast cells with the glycosylated forms of recombinant CCN1 and CCN3 decreased cell proliferation by bringing about G0/G1 cell cycle arrest, which was accompanied by the upregulation of activated Notch-1 and its target gene p21. Interestingly, both CCN proteins increased senescence-associated β-galactosidase activity and the expression of the senescence marker p16. The migration capability of SGHPL-5 cells was mostly enhanced in response to CCN1 and CCN3, by the activation of FAK and Akt kinase but not by the activation of ERK1/2. In summary, both CCN proteins play a key role in regulating trophoblast cell differentiation by inducing senescence and enhancing migration properties. Reduced levels of CCN1 and CCN3, as found in early-onset preeclampsia, could contribute to a shift from invasive to proliferative EVTs and may explain their shallow invasion properties in this disease.

Introduction

In mammalian species, the formation of a functional placenta is essential for normal fetal growth and development. Appropriate placentation in humans relies on the ability of extravillous cytotrophoblasts (EVTs) to proliferate and then to invade the maternal tissue. Diploid EVTs located in the proximal cell column continuously proliferate to provide a constant supply of invading EVTs during the first trimester of pregnancy.^1-4 To establish a connection between the placenta and the maternal vasculature, terminal differentiated trophoblast giant cells, characterized by endoreduplication up to 1000N of DNA,⁵ invade and transform the maternal vessels, and this transformation in turn guarantees nutrition and oxygen support to the placenta and the fetus.^6,7

Before differentiating into the invasive pathway, the proliferative trophoblast cells of the cell column escape from the cell cycle when they come into contact with the maternal decidua. To prevent tumor-like behavior, such as that occurring in choriocarcinoma, the proliferation and invasion properties are temporally and spatially separated in EVTs. In humans, these two processes are tightly controlled by a plethora of multiple and complex signaling factors, such as growth factors, hormones, and chemokines.^8-11 Preeclampsia, a complication in pregnancy, is known to coincide with an insufficient invasion of trophoblast cells into the decidua and the maternal spiral arteries. Such placentas lack sufficient maternal vascular remodeling, and this characteristic, combined with a restricted supply of oxygen and nutrition for the embryo, may result in intrauterine growth restriction. Therefore, deciphering this defined regulation process is important for understanding the pathogenesis of preeclampsia.

Previous studies have shown that the matricellular CCN protein family members CCN1 (CYR61) and CCN3 (NOV) play an important role in these regulatory processes.^12-15 CCN proteins are known to regulate pivotal cellular processes, such as differentiation, proliferation, migration, and angiogenesis.^16,17 Downstream signaling events are mediated by integrins, bone morphogenetic proteins (BMPs), vascular endothelial growth factor (VEGF), Wnt proteins, and Notch.¹⁸

CCN1 and CCN3 proteins occur in a secreted glycosylated form (g-CCN1 and g-CCN3) or in an intracellular non-glycosylated form (ng-CCN1 and ng-CCN3).^19,20 As shown in earlier studies, these two types of proteins function differently in regulating trophoblast proliferation and migration.^14,21 In the human placenta, CCN1 and CCN3 are expressed in endothelial cells of placental vessels, stromal cells, and interstitial EVT giant cells, and their expression levels increase during pregnancy.²² In the placentas of women with early-onset preeclampsia, CCN1 and CCN3 protein levels are significantly lower than in gestation-matched control placentas.²³ We have already demonstrated that in the malignant trophoblast cell line Jeg3, a model of invasive EVT, CCN3 reduces cell proliferation and enhances migration properties.^12-14 These studies showed that CCN3 acts by inducing the mitogen-activated protein kinase/extracellular signal-related kinase (MAPK/ERK), phosphatidylinositol 3-kinase/protein kinase B (PI3K/Akt), and Notch/p21 pathways, mediating these separate functions for proliferation and migration/invasion in Jeg3 cells.¹⁴

In the study reported here, we investigated the proliferation control of CCN1 and CCN3 in benign SGHPL-5 trophoblast cells, which are more similar to the in vivo situation than previous models. We confirmed that the proliferation of the SGHPL-5 cell line is reduced by CCN1 and CCN3, whereas the migration is mostly enhanced by these proteins. We found that the CCN1 and CCN3 proteins induce senescence of the trophoblast cells, which is accompanied by cell cycle arrest at G0/G1. Simultaneously, CCN1 and CCN3 seem to promote migration capability by activating focal adhesion kinase (FAK) and Akt kinase (protein kinase B), a finding suggesting that the CCNs play a regulatory role in controlling proliferation and stopping differentiation, inducing senescence and the onset of migration in EVTs.

Materials and methods

Cell culture and treatment of SGHPL-5 trophoblast cells

The cytotrophoblast cell line SGHPL-5 (kindly provided by G. Whitley, Division of Basic Medical Sciences, St George's University of London, UK) was routinely cultivated in Ham's F10 nutrient mixture (Biochrom AG, Berlin, Germany) supplemented with 10% fetal calf serum (FCS; Biochrom AG), 2 mM L-glutamine, and 1% penicillin/streptomycin (10,000 U/ml, 100x; Live Technologies, Carlsbad, CA, USA). Cells were seeded as specified in the following sections and allowed to attach for 24 h in normal culture medium. Synchronization in cell cycle phase distribution was achieved by serum starvation for another 24 h.

Cells were treated with 1 µg/ml recombinant human glycosylated CCN1 and CCN3 (g-rhCCN1, g-rhCCN3) from mouse myeloma cells (R&D Systems, Minneapolis, MN, USA); with 1 µg/ml non-glycosylated CCN1 and CCN3 (ng-rhCCN1, ng-rhCCN3) from E. coli (PeproTech, Hamburg, Germany); or with 1 µg/ml solvent control (0.1% bovine serum albumin [BSA] in phosphate-buffered saline [PBS]).

In vitro proliferation assay

Cells were seeded at a density of 5×10⁴ cells per well in 12-well plates in triplicate. After 24 h of serum starvation, the cells were treated with 5% FCS and 1 µg/ml g-rhCCN1, ng-rhCCN1, g-rhCCN3, ng-rhCCN3, or PBS/0.1% BSA as a solvent control. An electronic cell counter (CASY-I; Schärfe Systems, Reutlingen, Germany) was used to count the cells 24 h and 48 h after plating, as previously described.^13,24

Analysis of cell cycle distribution

Cells were seeded at a density of 7×10⁵ cells per well in 25-cm² cell culture flasks. After 24 h of serum starvation, cells were treated with 5% FCS and 1 µg/ml g-rhCCN1, ng-rhCCN1, g-rhCCN3, ng-rhCCN3, or PBS/0.1% BSA as a solvent control for 0 h, 4 h, or 24 h. Bromodeoxyuridine (BrdU) was added to the culture for the last two hours of the incubation period. Cells were then fixed and stained for newly synthesized DNA as marked by incorporated BrdU using a specific fluorescein isothiocyanate (FITC)-conjugated anti-BrdU antibody as well as total DNA by 7-amino-actinomycin D (7-AAD) according to the manufacturer's protocol (FITC BrdU Flow Kit; BD Pharmingen, San Jose, CA, USA). Two-color flow cytometric analysis was used to detect cells actively synthesising DNA (Fl-1, FACSCalibur; Becton Dickinson, Heidelberg, Germany) and total DNA (Fl-3). Positions in the G0/G1, S, and G2/M phases of the cell cycle were quantified with a classical DNA profile (FL-3; histogram plot of DNA content against cell numbers).

Annexin V apoptosis assay

Cells were seeded at a density of 9×10⁴ cells per well in 6-well plates. After 24 h of serum starvation, the cells were treated with 1 µg/ml g-rhCCN1, g-rhCCN3, or PBS/0.1% BSA as a solvent control for 24 h. Annexin V apoptosis assays were performed as described by Koch et al.²⁵ using flow cytometry (FACSCalibur, Becton Dickinson) in combination with FITC-coupled annexin and propidium iodide (PI; BD Pharmingen).

Senescence-associated β-galactosidase staining

SGHPL-5 cells were seeded in 6-well plates (3×10⁵ cells per well), and experiments were performed with 1 µg/ml rhCCN1, rhCCN3, or PBS/0.1% BSA as a solvent control for 24 h or 48 h. Cells were washed with PBS and were then fixed for 15 min in 0.2% glutaraldehyde in PBS. After two washes with PBS, fixed cells were incubated in freshly prepared senescence-associated β-galactosidase (SA-β-Gal) staining solution (1 mg/ml X-Gal, 5 mM potassium ferricyanide, 5 mM potassium ferrocyanide, and 2 mM MgCl₂ in PBS at pH 6.0) for 24 h at 37°C. At least three random fields were digitally photographed with a phase-contrast microscope (10× magnification). The numbers of total cells and of positive blue-stained cells were counted and depicted as SA-β-gal–positive cells per 100 cells.

Analysis of migration

Wound healing migration assays for analyzing horizontal migration properties were performed with co-culture inserts (ibidi GmbH, Martinsried, Germany). We seeded 2 × 10⁴ cells into each chamber of the insert and allowed to attach in regular culture medium. After 24 h, the cells were pretreated with 1 µg/ml rhCCN1, rhCCN3, or PBS/0.1% BSA as a vehicle control in serum-free culture medium for another 24 h. Then the culture insert was removed, leaving a defined cell-free gap of 500 µm. Cell patches were overlaid with serum-free culture medium and 1 µg/ml ng-rhCCN1, g-rhCCN1, ng-rhCCN3, g-rhCCN3, or PBS/0.1% BSA as a vehicle control. Cells were allowed to migrate, and phase contrast images were collected with a Zeiss Axiovert 25 microscope (Carl Zeiss AG, Oberkochen, Germany) at 0 h and 24 h. For documentation, an AxioCamICc1 camera and the AxioVision LE Release 4.8.2 image analysis software (Carl Zeiss AG) were used. Photographs of wound healing migration assays were analyzed with the WimScratch quantitative image analysis software (www.wimasis.com).

To analyze vertical migration with transwell assays, we seeded SGHPL-5 cells were seeded at a concentration of 3×10⁴ cells per well on uncoated Transwell chambers (Falcon cell culture inserts for 24-well plates, 8-µm pore size; Thermo Fisher Scientific, Waltham, MA, USA). After attachment, cells were incubated with 1 µg/ml g-rhCCN1, ng-rhCCN1, g-rhCCN3, ng-rhCCN3, or PBS as a vehicle control in both chambers. After 6 h, non-invaded cells on the upper side of the inserts were removed with a cotton swab. Cells on the lower surface were fixed in ice-cold methanol and stained with hematoxylin. Membranes were cut out with a scalpel blade, placed on glass slides, and covered with Shandon Xylene Substitute mountant (Thermo Fisher Scientific). For evaluation, we took 5 non-overlapping pictures of each membrane in three independent experiments prepared as duplicates (20× magnification; Axiophot microscope, Carl Zeiss AG; Digital Sight DS-U1 camera, Nikon, Düsseldorf, Germany) and analyzed with Image J software (http://imagej.nih.gov/ij/).

RNA isolation and quantitative reverse-transcriptase polymerase chain reaction

Total RNA was isolated from cells with the E.Z.N.A RNA extraction kit (Omega-Biotek, Norcross, GA, USA) and was reversely transcribed as previously described.²⁴ Gene expression was quantified with the quantitative PCR Master Mix and SYBR green (Applied Biosystems, Darmstadt, Germany) and an ABI Prism 7300 sequence detector (Applied Biosystems). PCR reactions were carried out in triplicate with a final volume of 20 µl, with 1 µl (40 ng) cDNA, 1x reaction buffer containing SYBR green, and 10 pmol sense and anti-sense primers (for sequences, see Table 1). PCR was performed for 10 min at 95°C, followed by 40 cycles consisting of 10 sec denaturation at 95°C and 1 min annealing at 60°C. Specificity of the amplification products was confirmed by melting curve analysis. The use of 10-fold series dilutions of purified PCR products from 1 pg to 0.1 pg as standards provided a relative quantification of the unknown samples. The quantity of cDNA in each sample was normalized to the β-actin content.

Table 1.

Sequences and National Center for Biotechnology Information (NCBI) accession numbers for polymerase chain reaction primers.

Gene	NCBI accession number	Primer sequences (5′ → 3′)		Product length (bp)
β-actin	NM_001101	sense: ACC AAC TGG GAC GAC ATG GAG AAA A	anti-sense: TAC GGC CAG AGG CGT ACA GGG ATA	213
p15INK4B	NM_078487	sense: TGC TAG GAT GCG GAA ATC CC	anti-sense: AGG CGT TTG GAC TGA GTT TG	199
p16INK4A	NM_000077.4	sense: CAT GGA GCC TTC GGC TGA C	anti-sense: GGC CTC CGA CCG TAA CTA TT	120
p27Kip1	NM_004064.3	sense: CAG CTT GCC CGA GTT CTA CT	anti-sense: AAG AAT CGT CGG TTG CAG GT	236
MMP-2	NM_004530.2	sense: ATG ACA GCT GCA CCA CTG AG	anti-sense: ATT TGT TGC CCA GGA AAG TG	174
MMP-9	NM_004994.2	sense: TTG ACA GCG ACA AGA AGT GG	anti-sense: GCC ATT CAC GTC GTC CTT AT	179
INT-α5	NM_002205	sense: CTA CAA TGA TGT GGC CT CG	anti-sense: GGA TAT CCA TTG CCA TCC AG	198
INT-β1	NM_002211.3	sense: TGG CCT TGC ATT ACT GCT GA	anti-sense: GCG TGT CCC ATT TGG CAT TC	104

bp, base pairs; MMP, matrix metalloproteinase; INT, integrin.

Immunofluorescence and microscopy

SGHPL-5 cells were seeded onto sterile glass coverslips and fixed with ice-cold methanol for 10 min. Blocking was achieved with 1% BSA in PBS for 10 min at room temperature (RT). Cells were incubated with rabbit anti–Ki-67 antibody (Abcam, Cambridge, UK; ab66155, 1:100) for 1.5 h and with secondary Cy3-conjugated donkey anti-rabbit antibody (Dianova, Munich, Germany) for 45 min at RT. Cell nuclei were counterstained with 4′,6-diamidino-2-phenylindole hydrochloride (DAPI; 0.1 µg/ml) for 15 min at RT. Mowiol (Sigma Aldrich, Munich, Germany) was used as a mounting medium. Microscopy was carried out with a confocal laser-scanning microscope (Leica TCS SP5; Leica, Wetzlar, Germany).

Protein preparation and western blotting

Cells were harvested in NETN lysis buffer (175 mM Tris-Base, pH 8; 100 mM NaCl; 1 mM EDTA, 0.4% Igepal CA-630) supplemented with ethylenediaminetetraacetic acid (EDTA)-free complete protease inhibitors and PhosSTOP phosphatase inhibitor cocktail (Roche, Penzberg, Germany) and homogenized by 5 passes through a 20G needle syringe. Cell lysates were centrifuged for 10 min at 13,000 rpm at RT, and cell debris were removed. Next, 30 to 50 µg of total protein was separated on a 12% polyacrylamide gel and transferred to a nitrocellulose membrane (Amersham Bioscences, Piscataway, NJ, USA). The following primary antibodies were used: polyclonal rabbit anti–β-actin (#A2066; Sigma Aldrich), polyclonal rabbit anti-p42/44 (#9154), polyclonal rabbit anti-phospho p42/44 (#4370), polyclonal rabbit anti-Akt (#4691), polyclonal rabbit anti–phospho-Akt (#3787), mouse anti-human cyclin D1 (#sc-8396; Santa Cruz Biotechnology, Dallas, TX, USA), mouse anti-human p21^Waf1(Cip1 (#S2946), rabbit anti-human cleaved Notch 1 (#2421) (all from Cell Signaling Technologies, Danvers, MA, USA); mouse anti-human p16 (F-9) (sc-55600), monoclonal mouse α-Tubulin (#sc-8035) (both from Santa Cruz Biotechnology), and mouse anti-human glyceraldehyde 3-phosphate dehydrogenase (GAPDH; #5G4 Mab 6C5; HyTest, Turku, Finland). Secondary horseradish peroxidase (HRP)-conjugated antibodies were purchased from Santa Cruz Biotechnology.

Protein expression was analyzed as described by Yang et al.¹³ Detection was accomplished on X-ray films with Supersignal West Dura Extended Duration Substrate (Thermo Scientific Pierce, Rockford, IL, USA) according to the manufacturer's protocol (Kodak, Stuttgart, Germany). For normalization of protein expression, rabbit anti–β-actin, rabbit anti-GAPDH, or mouse anti-α-Tubulin was used.

Statistical analysis

Statistical analysis of densitometric data from Western blot analyses and of the qRT-PCR results was performed with GraphPad Prism 5 (Graphpad Software Inc., La Jolla, CA, USA). Statistical significance was determined with PASW Statistics 18 (IBM, Duesseldorf, Germany) using the Mann-Whitney U test for nonparametric independent two-group comparisons. For comparing CCN-treated samples with untreated controls, statistical significance was set at the level of P ≤ 0.05.

Results

CCN1 and CCN3 decrease proliferation and induce G0/G1 cell cycle arrest in SGHPL-5 trophoblast cells

As previously reported, both g-CCN3 and ng-CCN3 recombinant proteins, decrease proliferation of the malignant trophoblast cell line Jeg3.^13,14 Because of the tumor characteristics of Jeg3 cells concerning proliferation control we investigated the influence of recombinant human CCN1 and CCN3 proteins on proliferation control in benign SGHPL-5 cells. SGHPL-5 cells endogenously express ng-CCN1 but not CCN3 protein (data not shown). When g-rhCCN1, ng-rhCCN3 or g-rhCCN3 was added to the cell culture medium, the numbers of SGHPL-5 trophoblast cells were significantly lower within 48 h than in control cultures (Fig. 1A), and g-rhCCN1 and ng-rhCCN3 were the most effective proteins. The expression of the proliferation marker Ki-67 was clearly lower after treatment with both glycosylation forms of CCN1 and CCN3 than in control cells, as determined by immunocytochemistry (Fig. S1)

Figure 1.

CCN1 and CCN3 significantly reduce the proliferation of SGHPL-5 trophoblast cells. (A) Total cell numbers were determined after treatment of SGHPL-5 cells with 1 µg/ml of non-glycosylated recombinant human CCN1 (ng-rhCCN1), ng-rhCCN3, glycosylated (g)-rhCCN1 and g-rhCCN3. Proliferation was significantly lower in response to g-rhCCN1, ng-CCN3, and g-rhCCN3 than in control cultures (ctrl); proliferation was slightly reduced after treatment with ng-rhCCN1. N = 3. *P ≤ 0.05. (B) Staining of SHGPL-5 cells treated with g-rhCCN1 or g-rhCCN3 alone, preincubated with anti-CCN1 or anti-CCN3 and anti-CCN1 or anti CCN3 alone, compared with staining of control cells for the early apoptosis marker Annexin V (AnnV⁺/propidium iodide [PI]⁻). Numbers of stained cells as a percentage were determined by fluorescence-activated cell sorting. Apoptosis was significantly increased only after treatment with g-rhCCN1. *P ≤ 0.05. (C) Determination of the number of polyploid SGHP-5 cells as a percentage (percentage of cells with n > 4) after treatment for 24 h with CCN1 and CCN3 recombinant proteins. No significant changes in polyploidy were found.

To identify the reasons for the reduction in cell numbers after the addition of CCNs, we investigated apoptosis and cell cycle arrest. Analysis of apoptosis with the Annexin V assay showed a significant increase in Annexin V staining after treatment with g-rhCCN1 but not with g-rhCCN3 (Fig. 1B). Antibodies against CCN1 alone or rh-CCN1 preincubated with anti-CCN1 did not increase Annexin V staining. Other markers of apoptosis, such as caspase-3 and p53, however, were not altered after the addition of either of the rhCCN proteins (data not shown). Moreover, we did not find that CCN1 and CCN3 induced any increase in the numbers of polyploid cells (n > 4) in SGHPL-5 cells as a marker of endoreduplication (Fig. 1C).

Cell cycle analysis of BrdU-labeled cells by fluorescence-activated cell sorting (FACS) showed that the number of cells arrested in the G0/G1 phase after treatment with ng-rhCCN1 (52.85% ± 0.55%), g-rhCCN1 (52.45% ± 2.69%), ng-rhCCN3 (53.06% ± 2.57%), or g-rhCCN3 (51.49% ± 2.17%) was significantly higher than that of control cells (40.05% ± 1.46%) or vehicle control (42.74% ± 0.86%) (Fig. 2). Accordingly, the fraction of cells in G2/M phase was significantly lower after the addition of ng-rhCCN1 (29.52% ± 1.40%), g-rhCCN1 (28.67% ± 1.86%), ng-rhCCN3 (31.51% ± 2.20%), or g-rhCCN3 (31.47% ± 1.39%).

Figure 2.

Treatment with CCN1 and CCN3 results in a G0/G1 cell cycle arrest of SGHPL-5 cells. Analysis of cell cycle phase distribution (G0/G1, S, G2/M, apoptotic) in SGHPL-5 cells after 24 h treatment with glycosylated recombinant human CCN1 (g-rhCCN1), g-rhCCN3, non-glycosylated rhCCN1 (ng-rhCCN1), or ng-rhCCN3 with the fluorescein isothiocyanate (FITC) bromodeoxyuridine (BrdU) Flow Kit and fluorescence-activated cell sorting (FACS) showed that both glycosylation forms of CCN1 and CCN3 induced a G0/G1 cell cycle arrest. (A) Analysis was performed according to the manufactures instruction (R9: G0/G1, R10: G2/M, R11: S-phase and R12: apoptotic cells). Representative FACS blots of each treatment condition are shown. N = 3. *P ≤ 0.05. (B) Analysis of cell cycle distribution. The proportion of SGHPL-5 cells remaining in G0/G1 after treatment with CCN1 and CCN3 for 24 h was significantly higher than that in control cells. The number of cells in G2/M phase was significantly reduced by treatment with ng-rhCCN1, g-rhCCN1, or g-rhCCN3. N = 3. *P ≤ 0.05.

CCN1 and CCN3 decrease proliferation via Notch-1 receptor/p21 signaling

The Notch-1 receptor is known to be expressed in cytotrophoblast cells of the human placenta²⁶ and to regulate the cyclin/CDK inhibitor p21 as a downstream target of Notch-1 activation.²⁷ Our recent studies using Jeg3 malignant trophoblast cells also showed a link between the CCN-induced Notch-1 signaling pathway and the decrease in cell proliferation.¹⁴

Treating SGHPL-5 cells with g-rhCCN1, ng-rhCCN1, or g-rhCCN3 significantly enhanced the cleavage of the Notch-1 receptor (Fig. 3A). After 2 h, the expression of p21 protein was significantly upregulated by both glycosylation states of CCN1, whereas stimulation with glycosylated or non-glycosylated CCN3 enhanced p21 protein expression only slightly but not significantly (Fig. 3B). Interestingly, expression of cyclin D1, a positive regulator for the transition from the G1 to the S phase,²⁸ was slightly but not significantly upregulated after the addition of both glycosylation states of CCN1 and CCN3 (data not shown).

Figure 3.

CCN1 and CCN3 activate Notch-1/p21 signaling in SGHPL-5 cells. (A) Exemplary Western blots of the cleaved Notch-1 receptor and (B) the expression of its target gene p21 expression in SGHPL-5 cells after treatment with CCN1 or CCN3 for 2 h or 8 h are shown. Levels of protein expression are normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) or β-actin. Relative expression values of protein quantification are shown as mean expression levels normalized to GAPDH or β-actin above the Western blots. Columns represent the means of three independent measurements; error bars represent SEM. *, P ≤ 0.05 versus control. Cleavage of the Notch-1 receptor and, thus, activation of the Notch pathway is enhanced in cells treated with CCN1 or CCN3 than in control cells. The activated Notch pathway in turn upregulates the expression of the cell cycle regulator p21. *P ≤ 0.05.

CCN1 and CCN3 induce cellular senescence via upregulation of the senescence marker β-galactosidase and p16 expression

Since it is known that a G0/G1 cell cycle arrest via upregulation of the p21 pathway upon CCN1 is associated with cellular senescence,²⁹ we performed senescence-associated β-galactosidase staining of CCN1- and CCN3-treated SGHPL-5 trophoblast cells. The number of cells with SA-β-Gal staining was significantly higher after 48 h of treatment with g-rhCCN1 and both glycosylation forms of CCN3 than in control cultures (Fig. 4A-B). Expression of the cyclin-dependent kinase inhibitors p15^INK4B, p27^Kip1, and p57^Kip2, as well as p16^INK4A, on mRNA level did not differ significantly between treated and untreated SGHPL-5 cells (Fig. S2). However, the increase in SA-β-Gal staining upon CCNs was preceded by a significant upregulation of the cell cycle regulator and senescence marker protein p16 upon 2 h treatment with g- and ng-CCN1 and g-CCN3 (Fig. 4C). Interestingly, p16 protein levels decreased upon g-CCN1 after 48 hours maybe due to an increase in turnover rate.

Figure 4.

Senescence associated β-Gal staining in SGHPL-5 cells is enhanced by treatment with CCN1 and CCN3. (A) Analysis of cellular senescence in SGHPL-5 cells after treatment with CCN1 and CCN3 for 48 h as determined by SA-β-gal staining (shown as blue staining). Scale bar, 50 µM. (B) Numbers of SA-β-gal positive cells per 100 cells were determined by microscopic observation. The results showed that the occurrence of senescent cells after 48 h of treatment with glycosylated recombinant human CCN1 (g-rhCCN1), g-rhCCN3, or non-glycosylated (ng)-rhCCN3 treatment was significantly higher than that in untreated control cells (ctrl). (C) Exemplary Western blot of protein expression of the cell cycle regulator and senescence marker p16^INK4A in SGHPL-5 cells after treatment with CCN1 or CCN3 for 2, 24, or 48 h. Levels of protein expression are normalized to α-tubulin. Densitometric analysis of relative expression values of protein quantification is shown as mean expression levels normalized to α-tubulin (N = 3) above the Western blot. Columns represent the means of three independent measurements; error bars represent SEM. *, P ≤ 0.05 vs. control. The expression of p16 is upregulated after 2 h of treatment with CCN1 and CCN3. *P ≤ 0.05.

CCN1 and CCN3 increase the migration of SGHPL-5 trophoblast cells via phosphorylation of FAK and Akt kinases

In our recent studies using the choriocarcinoma cell line Jeg3, we found increased migration of Jeg3 cells after treatment with ng-CCN3 but not with g-CCN3. This increased migration was mediated by the activation of Akt and MAP kinases.^13,14

We investigated the cell migration behavior of SGHPL-5 cells after treatment with g-rhCCN1, ng-rhCCN1, g-rhCCN3 or ng-rhCCN3 by using two separate assays to analyze horizontal and vertical migration properties. Horizontal migration of SGHPL-5 cells was performed as a wound healing assay using co-culture chambers and was mostly enhanced by the glycosylated forms of both CCN1 and CCN3 and by non-glycosylated CCN1 (Fig. 5A, C ). In uncoated transwell migration assays analyzing vertical migration, we observed an increase in migration, but not significantly after treatment with ng-rhCCN1 or ng-rhCCN3, and a significantly decreased migration after treatment with g-rhCCN3 as well as no obvious change in migration upon g-CCN1 (Fig. 5B, D )

Figure 5.

CCN1 and CCN3 change the migration properties of SGHPL-5 trophoblast cells. (A) Exemplary phase micrographs of wound healing horizontal migration assays using ibidi co-culture chambers. SGHPL-5 cells were treated with glycosylated or non-glycosylated CCNs for 24 h. Untreated cells were used as controls (controls and vehicle controls). Each micrograph is representative of three independent experiments. Glycosylated and non-glycosylated CCNs stimulated the horizontal migration of SGHPL-5 cells; the most pronounced effect was achieved with glycosylated CCNs compared to controls. Scale bar, 500 µm. (B) Exemplary micrographs of migration assays using transwell chambers are shown. The cells were treated with glycosylated or non-glycosylated CCNs for 6 h. Five random fields per condition were photographed at 20× magnification. Interestingly, non-glycosylated CCN1 (ng-CCN1) and CCN3 (ng-CCN3) slightly induced migration, whereas glycosylated (g)-CCN3 significantly reduced the migration capability of SGHPL-5 cells. Scale bar, 50 µm. (C) Quantification of horizontal migration assays. Bars represent mean values of three independent experiments; error bars indicate SD. Horizontal migration was significantly enhanced after stimulation with glycosylated CCNs and non-glycosylated CCN1 compared to controls. N = 3 *P ≤ 0.05 (D) Quantification of vertical migration assays using transwell chambers. Bars represent mean values of three separate (glycosylated and non-glycosylated) experiments performed in duplicate; error bars indicate SD. Mean value of untreated cultures (control) was arbitrarily set at 100%. Vertical migration of SGHPL-5 cells was moderately higher after treatment with non-glycosylated (ng)-CCN3 and with ng-CCN1, not changed upon g-CCN1, but it was significantly reduced by treatment with g-CCN3. N = 3. *P ≤ 0.05.

Our analysis of matrix metalloproteinases (MMPs) showed that treatment with CCN1 or CCN3 does not alter transcript expression of MMP-2 by SGHPL-5 trophoblast cells (Fig. 6A). However, the glycosylated forms of both CCN1 and CCN3 significantly enhance MMP-9 mRNA expression in SGHPL-5 cells (Fig. 6B), whereas ng-rhCCN1 seems to exert no substantial influence on the expression of MMP-9 transcript.

Figure 6.

Analysis of matrix metalloproteinase (MMP)-2 and MMP-9 mRNA in SGHPL-5 trophoblast cells after treatment with CCN1 or CCN3. Relative transcript expression of MMP-2 and MMP-9 in SGHPL-5 cells after treatment with CCN1 or CCN3 for 2, 4, or 8 h, as determined by quantitative polymerase chain reaction (q-PCR). The glycosylated forms of CCN1 and CCN3 enhanced mRNA expression of MMP-9 (A), whereas the expression of MMP-2 mRNA (B) was not different from that of controls after treatment with CCN1 or CCN3. Levels of mRNA expression are normalized to β-actin. Relative expression values are shown as mean expression levels normalized to β-actin (N = 3). Columns represent the means of three independent measurements; error bars represent SEM. *P ≤ 0.05 versus controls.

FAK/Akt signaling is known to be activated by secreted MMP-2 and MMP-9.^30-33 Our studies showed that the phosphorylation of FAK is significantly affected only by g-CCN3 and is slightly increased by ng-CCN3 after 8 h of CCN3 treatment (Fig. 7A). Both glycosylation forms of CCN1 and the glycosylated form of CCN3 significantly promote the phosphorylation of Akt after 2 h and 8 h of CCN treatment (Fig. 7B). The phosphorylation of ERK1/2 was not changed by treatment with CCN1 or CCN3 (Fig. 7C).

Figure 7.

Analysis of activation of focal adhesion kinase (FAK), Akt, and extracellular signal-related kinase (ERK) in SGHPL-5 trophoblast cells after treatment with CCN1 or CCN3. Densitometric analysis of the expression of phosphorylated FAK (A), Akt (B), and ERK1/2 (C) compared to total expression of FAK, Akt, and ERK in SGHPL-5 cells after treatment with CCN1 or CCN3 for 2 or 8 h. Columns represent the means of three independent measurements; error bars represent SEM. *, P ≤ 0.05 vs. controls. An exemplary Western blot is shown below the graphs. Phosphorylation of FAK is significantly enhanced only by the glycosylated form of CCN3. (B) Phosphorylation of Akt is significantly increased by both glycosylation forms of CCN1 and by glycosylated recombinant human CCN3 (g-rhCCN3). (C) Phosphorylation of ERK1/2 in SGHPL-5 trophoblast cells is not affected by CCN1 or CCN3. *P ≤ 0.05.

CCN proteins are known to mediate the regulation of cell migration and invasion through diverse integrin receptors.³⁰ Using Jeg3 trophoblast cells we confirmed that integrin α5β1 is the receptor for CCN3-promoted trophoblast migration.¹⁴ Both subunits of the integrin α5β1 are expressed by SGHPL-5 trophoblast cells (Fig. S3). Whether integrin α5β1 mediates the increased migration behavior of SGHPL-5 like the Jeg3 cells must be elucidated by future experiments.

The schematic overview (Fig. 8) summarizes the identified signaling pathways of both CCN proteins in SGHPL-5 cells. CCN1 and CCN3 decrease proliferation by inducing cell cycle arrest and bringing about senescence; they also activate Notch/p21 signaling and simultaneously increase migration by activating FAK and Akt, probably via integrins α5β1, which are expressed in SGHPL-5 cells.

Figure 8.

Molecular mechanisms of CCN-mediated signaling in the human trophoblast leading to the switch between proliferation and invasion. Treatment with CCN1 or CCN3 decreased cell proliferation via Notch-1, accompanied by an upregulation of activated Notch-1 and its target gene p21, causing a cell cycle arrest. Both CCN proteins increased cellular senescence in SGHPL-5 cells, as characterized by an increase in senescence-associated β-galactosidase activity and p16 expression. In parallel, the migration capability of SGHPL-5 cells was mostly enhanced in response to non-glycosylated CCN1 and CCN3 by the activation of FAK and Akt kinase but not ERK1/2, probably via integrin α5β1 as the receptor. These results are transferable to the placenta. Here the Notch-1 receptor is expressed proximally in the placental cell column, whereas the integrin α5β1 receptor is expressed distally in the invading extravillous cytotrophoblasts (EVTs), as seen in the upper right corner (modified scheme from ref. 9).

Discussion

Cell cycle exit and subsequent differentiation into the invading cell type of trophoblast cells are central processes of placentation and are coordinated by an exact interplay between proliferation, differentiation, and invasion capabilities.³⁴ This study focused on the molecular regulatory mechanisms, such as proliferation and migration, that are mediated by both glycosylation forms of the matricellular proteins CCN1 and CCN3. We have previously shown that these proteins control the proliferation process in Jeg3 cells as a model of EVTs. EVTs detach from the cell column and differentiate into the invasive phenotype; they then deeply invade the maternal decidua and maternal spiral arteries.

Previous studies showed that CCN1 and CCN3 are expressed at high levels in the human placenta during pregnancy, with expression in interstitial EVT cells, in endothelial cells of vessels, and in stromal cells.²² The levels of both CCN proteins are consistently high in the sera of non-pregnant and pregnant women. However, lower levels of CCN1 and CCN3 were detected in the sera of pregnant women with early-onset preeclampsia, a disease that is associated with insufficient trophoblast invasion.^22,23 These findings indicate that CCNs are involved in the regulation of cell biological events at the feto-maternal interface.

More detailed analyses in previous studies showed that CCN3-mediated migration was induced by integrin α5β1 as the receptor and activator of Akt kinase, whereas Notch-1 and p21 are involved in antiproliferative capabilities of CCN3.¹⁴ In the present study we focused mainly on the role of CCN proteins in proliferation control, using the benign cytotrophoblast cell line SGHPL-5 as a model system for the in vivo situation.

CCN1 and CCN3 decrease proliferation of SGHPL-5 cells by inducing a G0/G1 cell cycle arrest, and then differentiate into a cellular senescent state

During trophoblast differentiation, some of the cytotrophoblasts (CTBs) underlining the syncytiotrophoblast layer maintain their undifferentiated phenotype throughout pregnancy, thereby providing a reservoir of placental stem cells. The remainder of the CTBs differentiate into two subpopulations of trophoblast cells: syncytiotrophoblast cells (STBs) and invasive extravillous interstitial cytotrophoblasts (EVTs).³ Until now, little has been known about the interplay of cell cycle regulators, and it has been impossible to determine whether trophoblast cells proliferate or exit from the cell cycle to allow further differentiation.^35,36 The results of previous experiments using the choriocarcinoma cell line Jeg3 suggested that CCN3 causes an imbalance between the proliferation and migration of human trophoblast cells.^13,14,22

In the present study we found that both glycosylation forms of CCN1 and CCN3 proteins reduce the numbers of benign SGHPL-5 trophoblast cells, whereas in Jeg3 cells only CCN3 seems to regulate proliferation. Comparing the effect of CCNs on migration properties in both cell lines showed that Jeg3 trophoblast cells and SGHPL-5 cells are mostly stimulated by non-glycosylated CCN1 and CCN3.^12-14 Thus, the regulation properties of CCNs on proliferation differ between the malignant and the benign trophoblast cell lines, and migration seems to be similarly regulated.

The reduced number of SGHPL-5 cells after treatment with CCN1 and CCN3 is based on cell cycle control and not on apoptosis. The analysis of cell cycle phase distribution found that reduced proliferation after treatment with CCN1 or CCN3 is associated with a G0/G1 cell cycle arrest characterized by an increased number of cells in the G0/G1 phase. The proportion of cells in the G2/M phase was significantly reduced by both glycosylation forms of CCN1 and CCN3. Arrest of or exit from the cell cycle is a precondition for a cell to pass into postmitotic states, such as quiescence, senescence, or terminal differentiation.³⁷

Studies of murine trophoblast giant cells have shown that terminal differentiation is marked by endoreduplication.³⁸ However we did not detect an increase in the number of polyploid SGHPL-5 cells after treatment with either CCN. Instead, our results clearly showed that both CCN proteins induced cellular senescence in SGHPL-5 cells, as demonstrated by an increased expression of SA-β-gal, and the increased expression of p16, both are well-established markers of cellular senescence.³⁹ Meanwhile, separate signaling pathways are mediated by CCNs and lead to alterations in proliferation and migration properties.

It is known that the Notch-1 receptor is expressed in CTBs of the human placenta²⁶ and that this receptor regulates the cyclin/CDK inhibitor p21.²⁷ In small cell lung cancer cells, Notch-1 signaling induces a p21-mediated cell cycle arrest.⁴⁰ Furthermore, Notch signaling plays an important role in the regulation of proliferation in the placental cell column and of trophoblast invasion and differentiation of EVTs. Inhibiting the Notch signaling pathway in primary EVTs and SGHPL-5 cells enhanced proliferation in the placental cell column, invasion capability, and expression of EVT markers, as shown by Haider et al.⁴¹ In the present study we found that Notch-1 expression is associated with the proliferative capability of CTB cell column progenitor cells, which is highest during the first trimester of pregnancy. This finding strongly corroborates our findings that CCN1 and CCN3 activate Notch-1 signaling and thereby reduce proliferation of the cytotrophoblast cell line SGHPL-5, as reported by Haider et al.⁴¹

CCN proteins are known to act via Notch-1 in other systems, such as myoblasts.^42-45 Our recent studies showed that Notch/p21 signaling also seems to mediate the proliferation-reducing activity of CCN3 in malignant Jeg3 cells.¹⁴ In the present study we found that CCN1 and CCN3 cause a G0/G1 cell cycle arrest and induce cellular senescence in SGHPL-5 trophoblast cells; this effect is presumably mediated by activation of the Notch-1 receptor after upregulation of p21.

Normally, cellular senescence is a characteristic feature of aging. It protects against tumourigenesis by limiting the proliferation of potentially detrimental cells and restricts tissue damage.⁴⁶ Recent studies by Krizhanovsky and colleagues⁴⁶ clearly showed that, in the placenta, the fusion of CTBs to STBs induces cellular senescence and that this action may be necessary for proper STB function during embryonic development.⁴⁷ The same finding has been reported in studies of mouse placentas. Zhang et al.⁴⁸ showed that throughout gestational days 14.5 to 18.5 the labyrinthine trophoblast cells strongly express SA-β-gal, p53, and p21 and therefore induce cellular senescence. The induction of senescence by CCN1 has already been described in other cell types, such as fibroblasts,⁴⁹ hepatic myofibroblasts,⁵⁰ cell lines of non-small-cell lung carcinoma,²⁹ and aging muscle cells.⁵¹

CCN1 and CCN3 induce the migration properties of SGHPL-5 cells by FAK and Akt signaling

In addition to the inhibition of proliferation, the non-glycosylated forms of CCN1 and CCN3 in particular tend to enhance the vertical migration properties of SGHPL-5 trophoblast cells, whereas for horizontal migration properties the glycosylated CCNs exert the strongest effect. Interestingly, the application of glycosylated CCN3 results in less vertical migration of SGHPL-5 cells. So far we had no proven explanation for these separate effects of the various CCN forms on migration directions. However, it is already known that glycosylation controls diverse protein functions such as migration and invasion properties of the extravillous trophoblast.⁵² Thus, here the different glycosylated CCN proteins may differ in the modulation of focal adhesion structures. The results in horizontal migration between the different CCN isoforms regarding its glycosylation may be explained by the fact that the glycosylated CCN proteins are the secreted isoforms which could act from outside on migration behavior of the cells in a paracrine manner and may therefore more efficiently increase migration compared to the non-glycosylated intracellular form of the CCNs which is located intracellularly and could only act in a autocrine manner.

Future investigations will focus on other potential signaling pathways that differ between the glycosylation forms of the CCN proteins. FAK is involved in integrin-mediated signal transduction pathways of the extracellular matrix and plays an important role in the regulation of cell proliferation, migration, and invasion.⁵³ It is known that the activation of Akt and ERK1/2 is related to cell migration and the activation of FAK^54-57 and that these kinases are involved in trophoblast migration and invasion (reviewed by Chakraborty et al.⁸). In SGHPL-5 cells, phosphorylation and thereby activation of FAK occurs only after treatment with glycosylated CCN3. The phosphorylation status of FAK does not change after treatment with CCN1. The phosphorylation and activation of Akt are induced by both CCN1 and CCN3. This finding has been verified for CCN3 in renal carcinoma cells.⁵⁸

Haslinger et al.⁵⁹ verified the increase in SGHPL-5 migration after the application of epidermal growth factor by Akt signaling, in particular the Akt 1 and Akt 3 isoforms. In Jeg3 cells, epidermal growth factor–like domain 7 promotes migration and invasion by activating the MAPK, PI3K, and Notch pathways.⁶⁰ In contrast here, phosphorylation and activation of ERK1/2 do not seem to play a role in CCN1/3-mediated regulation of migration in SGHPL-5 cells, because the phosphorylation status remains unchanged after treatment with both CCN proteins.

An important aspect of cell migration and invasion is the FAK/Akt-mediated enhanced expression and activity of MMP-2 and MMP-9.^31-33 SGHPL-5 cells treated with g-CCN1 or g-CCN3 exhibit significantly higher mRNA expression of MMP-9. If the protein level or activity of MMP-9 is also increased upon g-CCNs is unknown up to now and has to be investigated in future experiments. However, we assume that the involvement of CCN1 and CCN3 in invasion is obvious, and the activation of these signaling cascades and the resulting changes in cell physiology seem to depend on the inducing factor and the trophoblast cell line.

Thus, in a receptor-dependent manner, CCN1 and CCN3 support the inhibition of trophoblast proliferation and promote the migration of invasive trophoblasts into the maternal decidua. This conclusion is easily transferable to the placenta in vivo, because the Notch-1 receptor is expressed proximally in the placental cell column, whereas the integrin α5β1 receptor is expressed distally in the invading EVTs,⁴¹ thereby providing a spatially distributed spectrum of action (Fig. 8).

Taken together, the results of this study show that CCN1 and CCN3 are key regulatory proteins of the EVTs that control proliferation and invasion. They could support the cell cycle exit of trophoblast cells located at the proximal column and simultaneously enhance the migration properties of the invasive trophoblasts detaching from the column. Thus, we assume that both CCN proteins regulate the switch of EVTs from the proliferating to the non-proliferating senescence phenotype but not endoreduplication. We further assume that the reduced levels of CCN1 and CCN3 observed in early-onset preeclampsia could lead to the increased proliferation and thereby the reduced invasion capability of EVT cells. Our findings regarding the coordinated multifunctional properties of both CCN proteins in the human placenta and their defined signaling cascades may inspire efforts aimed at correcting impaired pathways in reproductive diseases by interfering with the CCN molecules.

Disclosure of potential conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We would like to thank Guy S. Whitley, London, for providing the SGHPL-5 cell line. The authors thank Claudine Kühn, Gabriele Sehn, Ursula Schmücker, Kathrin Kazuschke, and Dagmar Thyssen for their excellent technical assistance. We are grateful to Dr. Florence Witte for English editing and critical reading of the manuscript.

Funding

This study was funded by the German Research Foundation (DFG) with the contract numbers WI 774/22-2 to Elke Winterhager/Alexandra Gellhaus and GE 2223/2-1 to Alexandra Gellhaus.