Galectin-3 (gal-3) is a β-galactoside-binding protein which can be detected in endometrium. The study was designed to investigate synergism of gal-3 and integrinβ3 in endometrial cell proliferation and adhesion in an in vitro model of endometrial receptivity.
The RL95-2 cell line was employed as an in vitro model for receptive endometrium. Cells transfected with gal-3 siRNA or treated with exogenous gal-3 were incubated with or without function-blocking integrinβ1/3 antibody for evaluating synergism of gal-3 and integrins on cell proliferation and adhesion. Proliferation was measured by BrdU incorporation, and adhesion to fibronectin (FN) was determined by an adhesion assay. Integrin expression was analyzed by Flow Cytometry and western blots. Bewo spheroids were co-cultured with the RL95-2 monolayer to mimic the blastocyst–endometrial interaction, and colocalization of gal-3, integrinβ3 and FN at the interface was observed by confocal microscopy.
The knock-down of gal-3 inhibited RL95-2 cell proliferation and adhesion. However, a reduction of proliferation and adhesion was also observed in presence of exogenous gal-3, and this was further reduced by a functional block to integrinβ3. Moreover, gal-3 knock-down significantly increased integrinβ3 expression, however, the colocalization of integrinβ3 and FN was not increased. As expected, the colocalization of integrinβ3 was decreased with the knock-down of gal-3.
This study has provided an in vitro model for the complex interactions between gal-3 and integrinβ3 in the regulation of endometrial cell proliferation and adhesion.
Embryo implantation is the process of the developing blastocyst adhering and embedding into the receptive endometrium (Dominguez et al., 2002; Paria et al., 2002). There is only a restricted time (‘window of endometrial receptivity’) when the implantation could occur during the menstrual cycle (Psychoyos, 1986). It is clear that a successful establishment of endometrial receptivity requires endometrial development, proliferation and differentiation in preimplantation period (Guzeloglu-Kayisli et al., 2007). Endometrial cells first undergo proliferation and then become differentiated, increasing their adhesive ability to establish the receptive status. This requirement of proliferation/differentiation has to be tightly controlled. Alternations in the expression of molecules on the cell surface occur during conversion of the endometrial surface from a non-receptive to a receptive state. The cross talk of these molecules ensures changes of cell proliferation and adhesion. However, molecular interactions at the feta-maternal interface are not yet fully understood. Many molecules, especially cell adhesion molecules such as integrins, are best known to play important roles during the preimplantation period. Through the activity of tightly regulated adhesion molecules and other cytokines, the endometrium becomes receptive (Lessey, 2002).
Galectins are a family of animal carbohydrate binding proteins which have been strongly implicated in cancer metastasis. Studies have suggested that extracellular galectins cross-link cell-surface and extracellular glycoproteins and may thereby modulate cell adhesion including cell–matrix adhesion and induce intracellular signals (Bao and Hughes, 1995; Rabinovich et al., 2002a; Rubinstein et al., 2004). Most recently galectins are suggested to bind intracellular non-carbohydrate ligands and have regulatory roles in intracellular processes (Yang et al., 2008). Galectin-3 (gal-3), a β-galactoside-binding protein, is classed as a chimera galectin which only contains one carbohydrate recognition domain. It is broadly expressed in tumor cells, macrophages, activated T cells, epithelial cells and fibroblasts. It has also been shown to be involved in cell growth, adhesion, apoptosis and differentiation, mainly through binding to glycoproteins (Liu et al., 2002; Rabinovich et al., 2002b). Its expression increases from the hyperplastic to the cancerous state of endometrial tissue (Brustmann et al., 2003) and is also altered in a variety of human tumors including endometrial tumors (van den Brule et al., 1996). Gal-3 interacts with a broad spectrum of ligands, such as integrins (Hadari et al., 2000; Cooper, 2002). It binds a variety of matrix glycoproteins, including fibronectin (FN), and is implicated in a variety of biological functions including cancer metastasis (Liu and Rabinovich, 2005; Elola et al., 2007). Increasing expression of gal-3 has been found in the peri-implantation period in different species (Lee et al., 1998; Ponsuksili et al., 2002; von Wolff et al., 2005; Du et al., 2006), implicating its role in endometrial receptivity.
Integrins are one of the major families of cell adhesion receptors. They are well characterized, as these cell surface receptors interact with the extracellular matrix (ECM) ligands and mediate various intracellular signals. Integrins play a role in cell–cell attachment, cell–cell communication and invasiveness of tumor cells, through interacting with ECMs (Seales et al., 2005). Integrins are regulated spatially and temporally within the endometrium throughout the menstrual cycle and early pregnancy, and have been observed in human endometrial glandular epithelium (Bondza et al., 2008). Integrins α1β1, α4β1, ανβ3 are key molecules which frame the putative window of endometrial receptivity (Damario et al., 2001), and integrinβ1 and β3 subunits seem to be important transducers of different signals. Recognition of integrins and their ligands, such as FN and vitronectin (VN), is followed by embryo attachment and invasion (Schultz et al., 1997; Rout et al., 2004). Women with various benign gynecologic disorders, including endometriosis, polycystic ovary syndrome, hydrosalpinges and luteal phase defect, appear to exhibit decreased endometrial receptivity and abnormal expression of integrins (Garcia-Velasco et al., 1999; Donaghay and Lessey, 2007).
In mammary carcinoma cells, the gal-3 and integrinβ1 complex is internalized, accompanied by rapid cell spreading (Furtak et al., 2001), and extracellular gal-3 binding to integrinβ1 induces cell apoptosis (Fukumori et al., 2003). Collectively, the main role of gal-3 in modulation of cellular function may be triggered by its coordination with integrinβ subunits. Work done previously in our lab and in others, has shown that gal-3 is expressed in endometrium and endometrial cells during peri-implantation (von Wolff et al., 2005; Du et al., 2006), suggesting that gal-3 might be active in the process of establishment of endometrial receptivity and involved in the regulation of endometrial cellular function. However, the mechanism of how gal-3 is involved in regulating endometrial cellular function, and how the co-operative regulation of gal-3 and integrinβ1/3 contribute to the process, is still unclear. Therefore, in this study, we investigated how gal-3 and its synergism with integrinβ subunits regulate endometrial cell proliferation and adhesion.
We used monolayer-cultured endometrial epithelial RL95-2 cells as an in vitro model for the human receptive uterine epithelium. RL95-2 cells mimic important aspects of the in-vivo situation at implantation, i.e. apical adhesiveness. More importantly, various integrins (α6, β1, β4 and β6, which are the natural receptors for gal-3) are found evenly distributed along the entire plasma membrane of RL95-2 cells (Thie et al., 1995). Compared with other endometrial cell lines such as HEC-1A or AN3CA, RL95-2 cells allow us to study certain mechanisms involved in formation of cell–cell contacts between endometrial epithelial cells and trophoblast. Use of the model of multicellular spheroids of trophoblast cell lines such as Bewo or JAR adhering to the free surface of RL95-2 cell monolayer as an in vitro experiment system to study embryo implantation has been applied by many researchers (Thie et al., 1998; Hohn et al., 2000; Heneweer et al., 2003; Harduf et al., 2009), so we co-cultured Bewo spheroids with an RL95-2 monolayer to mimic the blastocyst–endometrial interaction.
Materials and Methods
Unless stated, all the reagents from abroad were ordered from their respective agencies in Shanghai, China.
Human cell lines RL95-2 (CRL1671; (Way et al., 1983)) and Bewo (CCL98; (Pattillo and Gey, 1968)) were purchased from the American Type Culture Collection, Rockville, MD, USA. RL95-2 cells were maintained in plastic flasks in a 1+1 mixture of Dulbecco's modification of Eagle's Medium and Ham's F12 (Gibco Invitrogen'Carlsbad, CA, USA) containing 10% Fetal Bovine Serum (Gibco), 2 mM/l-glutamine (Sigma-Aldrich, St. Louis, MO, USA), 10 mM HEPES (Sigma) and 0.5 µg/ml insulin (Sigma), and penicillin (100 IU/ml, Sigma) and streptomycin (100 µg/ml, Sigma) at 37°C in a humid atmosphere with 5% CO2. Bewo cells were maintained in Ham's F-12 (Gibco), containing with 10% FBS (Gibco) and 3 mM/l-glutamine, penicillin (100 IU/ml) and streptomycin (100 µg/ml). Both culture media were changed every 2–3 days, and Bewo multicellular spheroids were generated as previously described (John et al., 1993; Hohn et al., 2000). Briefly, Bewo monolayers were suspended by treatment with 0.05% trypsin/0.02% EDTA (Gibco), and about 6 × 105 cells in 6 ml of medium were aliquoted into 25 ml Erlenmeyer flasks and cultured in an atmosphere containing 5% CO2 in air at 37°C on a gyratory shaker set at 70 rpm.
siRNA transfection assay
The RL95-2 monolayers were suspended by treatment with 0.05% trypsin/0.02% EDTA (Gibco), and about 2.5 × 105 cells in 500 µl of medium were seeded into wells of a 24-well plate or 1 × 106 cells in 1 ml of medium were seeded into wells of a 6-well plate and grown for 24 h to about 50% confluent in medium containing 10% FBS without antibiotics. All transfection reagents (Lipofectamine 2000, opti-MEM reduced serum medium) and siRNAs (gal-3 stealth siRNA, Stealth RNAi Negative Control Medium GC Duplex (Negative siRNA), Block-it fluorescent oligo) were purchased from Invitrogen, Carlsbad, California, USA. Gal-3 stealth siRNA templates were as follows (GALIG HSS 149595, Invitrogen): 5′-CUU UCU UCC CAA AUC UGC UGC UGG A-3′; 5′-UCC AGC AGC AGA UUU GGG AAG AAA G-3′. The Gal-3 treatment group was transfected by 100 nM gal-3 siRNA, the negative siRNA control group was transfected by 100 nM Negative siRNA and the mock group was transfected without siRNAs using Lipofectamine 2000 following the instructions from manufacturer. Meanwhile a block-it fluorescent oligo were transfected to visualize how effectively siRNAs were being delivered to RL95-2 cells. For transfection carried out in 24-well culture plates, the plating medium was 400 µl per well. 2.5 µl of siRNA or 1.5 µl of lipofectamine 2000 was added to 50 µl of opti-MEM reduced serum medium (in separate tubes) and allowed to incubate for 5 min at room temperature. The mixtures were then combined and incubated at room temperature for an additional 15 min and then added to the well. After 4 h of incubation, cells were fixed by 95% ethanol, washed and then investigated under an Olympus BX51 fluorescence microscope. Images were acquired by the Olympus DP70 color CCD Camera and analyzed by Image-Pro Plus (Version 5.0.2). The successful proportion of transfection was representative of the cellular population. Six hours later, the transfection mixture was replaced with fresh complete culture medium. The cells were harvested 24 h later for verification of gal-3 mRNA and protein expression. Experiments were performed in duplicate and repeated three times.
Twenty-four hours after transfection, total RNA of RL95-2 cells was extracted using Trizol Reagent (Invitrogen), and the purity of RNA was determined by the ratio of absorbance readings at 260 and 280 nm (A260/A280) where 1.8–2.0 was considered proper for cDNA synthesis. Total RNA (2 µg) was reverse transcribed using a RevertAid first-strand cDNA synthesis kit (Fermentas). Reverse-transcription PCR was performed prior to quantitative real-time PCR. The levels of mRNA expression were determined by real-time PCR using SYBR Premix Ex Taq (TaKaRa, Dalian, China) with the Applied Biosystems 7000 system SDS software. Data were collected after each annealing step. GAPDH was used as an endogenous control to normalize for differences in the amount of total RNA in each sample. The primer sequences and the sizes of the amplified fragments were as follows: gal-3 (93 bp) 5′-CTT CCA CTT TAA CCC ACG CTT CAA-3′ (sense), 5′-TGT CTT TCT TCC CTT CCC CAG TTA TT-3′ (anti-sense); GAPDH (131 bp) 5′-ATG ACC CCT TCA TTG ACC-3′ (sense), 5′-GAA GAT GGT GAT GGG ATT TC-3′ (anti-sense). PCR products were analyzed by electrophoresis in 2% agarose gel. Gel photographs were taken by Peiqing JS-380 automatic gel image analysis instrument. The integrated optical density of gal-3/GAPDH was recorded. Experiments were performed in triplicate and repeated three times.
Twenty-four hours after transfection, protein samples were extracted from RL95-2 cells using eukaryotic cell lysis buffer (Biocolor, Shanghai, China). The concentration of protein was measured by BCA method. Samples were run at SDS-PAGE gel, and then transferred to nitrocellulose membranes, and immunoblotted overnight at 4°C with primary antibody, monoclonal mouse anti-human gal-3 (mab1154, R&D Systems, Inc., Minneapolis, MN, USA) antibody (1:2500), monoclonal mouse anti-human integrinβ1, clone P5D2 (MAB1959, Millipore, Boston, MA, USA) antibody (1:1000), rabbit monoclonal integrinβ3 antibody (ab75872, Abcam, Cambridge, MA, USA, 1:1000) and monoclonal mouse anti-human GAPDH (Kangchen, Shanghai, China) antibody (1:5000) with gentle shaking. After washing, the membrane was incubated with horseradish peroxidase-conjugated secondary antibodies and detected by enhanced chemiluminescence assay. GAPDH was used as an endogenous control to normalize for differences. Experiments were performed in triplicate and repeated three times.
Fluorescent distribution of gal-3
The RL95-2 monolayer was grown on coverslips and then was transfected with siRNA of gal-3 for 24 h. The coverslips were then fixed with 4% paraformaldehyde for 15 min, incubated with 0.02% Triton X-100 for 15 min and blocked in 1% BSA for 30 min. Then the coverslips were incubated with monoclonal mouse anti-human gal-3 antibody (R&D Systems, Inc., 1:100) for 1 h followed by secondary Alexa Fluor 555 conjugated goat anti mouse IgG (Invitrogen, 1:100) for a further 1 h. The coverslips were washed and soaked in ddH2O for 5 s, mounted onto glass microscope slides with Vectashield mounting medium with DAPI (H-1200, Vector, Burlingame, VA, USA), and examined by Olympus Fluo View FV1000 confocal microscopy. For DAPI, excitation was carried out with a blue diode laser at 405 nm, and emission was collected at 425–475 nm. For Alexa Fluour 555, excitation was provided by a HeNe laser emitting at 543 nm, and emission was collected at 555–625 nm. Data analysis of the images was carried out using the software package provided by the Olympus instruments. Experiments were performed in duplicate and repeated at least three times.
The proliferation of RL95-2 was examined by a BrdU cell proliferation kit following the manufactory's instruction (2750, Millipore). In brief, 2 × 104 RL95-2 cells in 200 µl culture medium were seeded into a well of a 96 well culture plate. Cells were then transfected with 100 nM siRNA for 24 h. BrdU was added into culture 6 h before the end of the experiment incubation. After aspirating the media, 200 µl fixing solution was added into each well and incubated at room temperature for 30 min. Then 100 µl anti-BrdU monoclonal antibody (1:200) was added and incubated for another 1 h followed by incubation with 100 µl peroxidase conjugated goat anti-mouse IgG (1:2000) for 30 min at room temperature. After washing, 100 µl TMB peroxidase substrate was added and incubated for 30 min at room temperature in the dark. The reaction was then stopped, and plates were read at 450 nm using a microplate reader (BIO-RAD Model 3550). The proliferation absorbance was calculated by following formula: the proliferation absorbance = actual absorbance − background absorbance. Experiments were performed in triplicate and repeated three times.
RL95-2 cells adhesion assay
Twenty-four hours after transfection, adhesion of RL95-2 to FN was examined by FN strips following the instruction from the manufacturers (ECM101, Millipore). The cytomatrix cell adhesion strips were provided as 12 8-well removable strips in a plate frame. The wells in rows A-G had been coated with human FN. Row H of each strips was coated with BSA as a negative control. About 2 × 105 treated RL95-2 cells were added to each well. The plate was incubated at 37°C for 1 h with 5% CO2, and then washed three times with Hank's buffer. After 100 µl 0.2% crystal violet in 10% ethanol was added to each well, the strips were gently washed three times with PBS before 100 µl solubilization buffer (a 50/50 mixture of 0.1 M NaH2PO4, pH 4.5 and 50% ethanol) was added. The strips were incubated for 5 min at room temperature until the cell-bound stain was completely solubilized. The strips were then read at 562 nm by a microplate reader. Adhesion rate = Absorbance of test group/Absorbance of BSA × 100%. Experiments were repeated at least three times.
Integrinβ function blocking protocol
RL95-2 cells were treated with function-blocking monoclonal integrinβ1 antibody, clone P5D2 (Millipore, 1 ng/ml) (Balasubramanian and Kuppuswamy, 2003) or integrinβ3 antibody, clone 25E11 (MAB1957, Millipore, 1 ng/ml) (Loike et al., 1999) in presence of increasing concentrations of recombinant gal-3 (1154-GA, R&D Systems, Inc., 0, 5, 10, 20, 40 ng/ml) for 24 h, and then analyzed for rates of proliferation and adhesion.
FCM analysis of integrins expression
SiRNA transfected RL95-2 cells were harvested after 24 h incubation, washed twice with cold PBS, and then resuspended at the density 1 × 106 per 100 µl PBS. Then cells were incubated with fluorescein-conjugated mouse anti-human integrinβ1 monoclonal antibody (MAB1951F-100, Millipore, 1:100) or with fluorescein-conjugated mouse IgG1, K isotype (eBioscience, 1:100) as a control antibody, or with Cy5 Conjugated Mouse monoclonal anti-human integrinβ3 antibody (MAB1957S-100, Millipore, 1:100) or with Cy5 Conjugated Mouse IgG2a isotype (SouthernBiotech, 1:100) as a control antibody, at room temperature for 30 min. Cells were then analyzed by FASCalibur by counting 10 000 cells. Experiments were performed in duplicate and repeated three times.
Colocalization of gal-3, integrinβ3 and FN
The adhesion of Bewo spheroids to monolayer of RL95-2 was performed as previously described (Hohn et al., 2000). Briefly, an RL95-2 monolayer was grown on coverslips and then was transfected with siRNA of gal-3 for 24 h. Bewo spheroids were then added onto the monolayer of RL95-2 for an hour. Non-adherent spheroids were removed by centrifugation. The coverslips were then fixed with 4% paraformaldehyde for 15 min, washed and then incubated with 0.02% Triton X-100 for 15 min, with 1% BSA as a blocking agent for 30 min. The coverslips were then incubated with purified rabbit anti-human FN antibody (CL54911AP, Cedarlane laboratories limited, 1:50) or mouse anti-human gal-3 antibody (1:100) for 1 h. After washing out primary antibody, the coverslips were incubated with a mixture of Cy5 conjugated mouse anti-integrinβ3 antibody (Millipore, 1:50), Alexa Fluor 594 conjugated goat anti rabbit IgG (Invitrogen, 1:100) or Alexa Fluor 555 conjugated goat anti mouse IgG (1:100) for a further 1 h. The coverslips were washed three times, soaked in ddH2O for 5 s, mounted onto glass microscope slides with Vectashield mounting medium with DAPI (Vector), and examined by Olympus Fluo View FV1000 confocal microscopy. For Alexa Fluour 594, excitation was provided by a HeNe laser emitting at 543 nm, and emission was collected at 555–625 nm. For Cy5, excitation was provided by a HeNe laser emitting at 633 nm, and emission was collected at 655–755 nm. Data analysis of the images was carried out using the software package provided by the Olympus instruments. Experiments were performed in duplicate and repeated at least three times.
All data were presented as mean ± SE Statistical analyses were performed using One-Way ANOVA and the LSD was applied for post hoc testing using SPSS software 15.0 for windows with P < 0.05 being considered significant. All experiments were repeated three times, and one of representatives from at least three similar results was presented.
Knock-down of gal-3 expression by siRNA in RL95-2 cells
In order to detect the transfection efficiency, a 100 nM concentration of fluorescent oligo was used for the analysis. As seen in Fig. 1A, more than 80% of cells were filled with fluorescent oligo after 4 h incubation, which means the siRNA was successfully transfected into cells. Real-time PCR and western blot analysis were applied to verify the expression of gal-3 mRNA and protein expression. As shown in Fig. 1B, gal-3 mRNA expression was inhibited significantly 24 h after transfection in the gal-3 siRNA transfected group compared with the negative siRNA transfected group or mock group. The expression of gal-3 mRNA was reduced by 90%, which suggests that gal-3 mRNA was highly inhibited by gal-3 siRNA. The similar result was found at the protein level with western blot analysis (Fig. 1C). Considering the significant change in mRNA and protein level, gal-3 expression was successfully knocked down. The distribution of gal-3 in each group was examined by confocal microscope. Figure 1D shows that the cell surface, cytoplasmic, and nuclear staining of gal-3 was decreased in gal-3 silenced cells compared with the negative control or mock groups. This also indicates the successfully knock-down of gal-3.
Gal-3 is involved in proliferation of RL95-2 cells, and may coordinate with integrinβ3
In order to investigate whether gal-3 is involved in proliferation in human endometrial cells, we first silenced gal-3 on RL95-2 cells, and then measured the proliferation after 24 h. The proliferation of RL95-2 cells was significantly inhibited in gal-3 silenced cells when compared with the negative transfection or mock groups after 24 h (Fig. 2A, *: P < 0.05). There was no significance between negative transfection and mock group (P = 0.100). Then we investigated whether cell surface receptors integrins β1 or β3 are involved in gal-3 regulation of RL95-2 cell proliferation. We found that exogenous gal-3 reduced the proliferation of RL95-2 cells and that blockage of integrinβ3 reduced their proliferation even further (*: P < 0.05). However, treatment of integrinβ1 antibody on RL95-2 cells did not significantly influence their proliferation in the presence of exogenous gal-3 (Fig. 2B).
Gal-3 is involved in adhesion of RL95-2 cells to Fn, and integrinβ3, but not β1, is involved in the regulation
In order to investigate whether gal-3 played a role in regulation of endometrial cell adhesion to FN, gal-3 was silenced on RL95-2 cells and the adhesion rate was measured. As shown in Fig. 3A, adhesion of gal-3 silenced RL95-2 cells to FN was significantly reduced compared with negative siRNA transfection and mock groups (*: P < 0.05). There was no significance between negative transfection and mock group (P = 0.864). Then we analyzed whether integrinβ1 or β3 was involved in gal-3 regulation of RL95-2 cell adhesion to FN. Function blocking antibodies of integrins β1 or β3 (1 ng/ml) was added to cell supernatant in the presence of increasing concentrations of recombinant gal-3 (0, 5, 10, 20, 40 ng/ml) for 24 h and the adhesion rate was measured. The adhesion rate was significantly decreased when RL95-2 cells were treated with recombinant gal-3, and the addition of an integrinβ3 blocking antibody reduced their adhesion even further (Fig. 3B). However, there was no further reduction in adhesion in the integrinβ1 antibody treatment group (Fig. 3B).
Expression of integrinβ3, but not β1, is increased on gal-3 silenced RL95-2 cells
It was a paradox that RL95-2 cells treated with integrinβ1/3 function blocking antibodies played different roles in cell proliferative and adhesive ability regulated by gal-3. We hypothesized that a close relationship existed in integrins and gal-3. Then expression of integrinβ1/3 was measured on gal-3 silenced RL95-2 cells by western blot and FCM analysis. As shown in Fig. 4A, the protein expression of integrinβ1 subunit was not significantly changed in gal-3 silenced cells, and this had also been confirmed in FCM analysis (Fig. 4B and C). However, the expression of integrinβ3 was significantly increased in gal-3 silenced RL95-2 cells (Fig. 4C, *: P < 0.05).
Colocalization of integrinβ3 and FN at the interface of trophoblast and endometrial cell is not up-regulated on gal-3 silenced RL95-2 cells
Integrinανβ3 is most important endometrial receptivity marker, and decreased expression of integrinανβ3 in the receptivity window results to failure of implantation. Previous research has suggested that in vivo, endometrial cells up-regulate both gal-3 and integrinανβ3 (Lessey et al., 1996; von Wolff et al., 2005; Du et al., 2006). However, our data showed the expression of integrinβ3 was significantly increased in gal-3 silenced RL95-2 cells. Gal-3 is thought to modulate cell–extracellular matrix interactions in cell adhesion and migration; therefore, we established an in vitro model of embryo adhesion to endometrium to investigate colocalization of these proteins. Bewo spheroids were delivered to monolayer of RL95-2 cells for 1 h to mimic the process of embryo adhesion to the endometrium, and localization was measured by confocal microscopy. First, we investigated whether the colocalization of integrinβ3 and FN on the interface of Bewo spheroid and monolayer of RL95-2 cells was up-regulated when integrinβ3 was significantly increased in gal-3 silenced cells in vitro. Localization of integrinβ3 or FN on gal-3 silenced RL95-2 cells was measured by confocal microscopy. As shown in Fig. 5b, the intensity of fluorescent of integrinβ3 on silenced RL95-2 cells was increased, but the colocalization of integrinβ3 and FN on Bewo spheroid and monolayer of RL95-2 cells was not altered on gal-3 silenced RL95-3 cells (Fig. 5c) compared with negative transfected RL95-2 cells (Fig. 5f) and mock group (Fig. 5i). This implied that the extra integrinβ3 induced by gal-3 knock-down does not bind with FN.
Colocalization of integrinβ3 and gal-3 at the interface of trophoblast and endometrial cell is decreased on gal-3 silenced RL95-2 cells
We then investigated colocalization of integrinβ3 and gal-3 at the interface. As shown in Fig. 6a, in the gal-3 knock-down cells, the fluorescent intensity of gal-3 was decreased at the interface of the two cells compared with negative siRNA control (Fig. 6d) and mock group (Fig. 6g). The intensity of fluorescence of integrinβ3 on the gal-3 silenced RL95-2 cells was increased (Fig. 6b), but as expected colocalization of gal-3 and integrinβ3 was decreased (Fig. 6c).
The present study provided comprehensive information on the gal-3 regulation of endometrial cell proliferation and adhesion. In the preimplantation period, the uterine environment must be suitable for endometrial differentiation to take place and produce the receptive state which is crucial to successful embryo implantation. This environment is highly controlled by the synergism of the steroid hormones: estrogen and progesterone. By autocrine, paracrine or juxtacrine routes, growth factors or cytokines are involved in this steroidal regulated environment (Cai et al., 2003). Gal-3 is a unique mammalian beta-galactoside-binding protein with wide distribution in the cell nucleus, cytoplasm, on the cell surface or in the extracellular environment. Numerous studies have suggested that gal-3 is expressed in uterine epithelial cells adjacent to implanting blastocysts as well as in the decidualized endometrium of implantation sites (Knisley and Weitlauf, 1993; Phillips et al., 1996; Lee et al., 1998). Along with the additional findings that gal-3 is absent from non-decidualized endometrium and is not present in the uterus of non-pregnant females (Phillips et al., 1996), this strongly suggests that gal-3 has a pregnancy related function in the uterine environment. Increased expression of gal-3 in the mid-secretary phase of menstrual cycle particularly suggests that steroid hormones can regulate gal-3 expression and that gal-3 is involved in the establishment of endometrial receptivity. Secretion of gal-3 is observed in many cell types including trophoblast cells (Bao and Hughes, 1995), so gal-3 may exist as a soluble protein in the endometrial environment. Although the molecular mechanism of gal-3 activity during this period remains to be elucidated, much data suggests it may have a role in regulation of cellular functions such as cell proliferation and adhesion, and these functional changes are critical events during establishment of endometrial receptivity.
Our data showed, for the first time, that deficits of gal-3 could reduce endometrial cell proliferation and adhesion to FN. Furthermore, we also found that endometrial cell proliferation and adhesion to FN were inhibited in presence of exogenous gal-3. This phenomenon was a paradox, since intracellular gal-3 appeared to promote proliferation and adhesion and exogenous gal-3 inhibited these functions. Significant evidence has accumulated over the years and supported the notion that intra- and extracellular gal-3 affects numerous biological processes and seems to be involved in different physiological and pathophysiological conditions. Current evidence suggests that intracellular gal-3, in particular, containing the NWGR anti-death motif of the Bcl-2 family, inhibits cell apoptosis induced by chemotherapeutic agents such as cisplatin and etoposide in cancer cells (Fukumori et al., 2007). However, exogenous addition of recombinant gal-3 leads to reduced proliferation of T cells (Muller et al., 2006) which is consistent with our data, and induces apoptosis of many cell types (Nakahara et al., 2005; Zhuo et al., 2008). Thus varying effects may due to multivalent binding properties of gal-3 although it possesses only one carbohydrate-recognition domain. There are numerous biological ligands of gal-3 in the nucleus, cytoplasm, on the cell surface and in the extracellular environment which are structurally and functionally very diverse. Although less information indicates which cell surface receptors are responsible for gal-3 signaling in this manner, the literature shows that integrins associate with traditional ECM ligands such as FN. There are also a number of studies suggesting that integrins also bind to galectins such as gal-3 through interactions with laminin (LN) and FN (Ochieng et al., 2004; Elola et al., 2007).
Integrins are expressed in all kinds of cells and associate with cell proliferation, tumor growth and metastasis. Members of the integrin family work as receptors for FN, LN, VN and other cell membrane protein. Integrins also work as signal transducers regulating cell proliferation and survival. A mount of studies have shown that extracellular gal-3 promotes apoptosis of cells (Liu and Rabinovich, 2005; Nakahara et al., 2005; Dumic et al., 2006), through binding to integrinβ1, showing that extracellular gal-3 can transmit intracellular signals through integrinβ1 subunits. A study by Zhuo et al. (2008) demonstrated that a subunit of integrinβ, α 2-6 sialylation, blocked human colon tumor cell apoptosis induced by extracellular gal-3 and blocked adhesion of integrinβ1 to galectin. Gal-3 influences integrinα2β1-mediated adhesion complex formation by altering receptor clustering in Madin-Darby canine kidney cells (Friedrichs et al., 2008). The reduction of endometrial cell proliferation in the presence of exogenous gal-3 might be partly, due to apoptosis induced by recognition of gal-3 and integrinβ1, however, it cannot account for it fully or else the deficiency of integrinβ1 would have reversed the reduction of proliferation caused by exogenous gal-3. However, downward trend in proliferation after treatment of recombinant gal-3 with integrinβ3 antibody was more apparent than with the integrinβ1 antibody, suggesting that recognition of integrinβ3 by exogenous gal-3 may induce endometrial cell proliferation partially counteracting the apoptotic effect of gal-3, and when this recognition is prevented, the rate of cell proliferation further decreased.
The present study also demonstrated that a functional block of integrinβ3 further inhibited endometrial cell adhesion to FN in the presence of different concentrations of recombinant gal-3. The reduced cell adhesion caused by exogenous gal-3 may thus be counteracted by the presence of integrinβ3. These results indicate that integrinβ3 may be involved in gal-3 regulated endometrial cell adhesion. In addition, the expression of integrinβ3 was increased in gal-3 knock-down endometrial cells in our study. This result implies that gal-3 normally acts a negative regulator of integrinβ3, although how this affects its functional capacity is presently unclear. Colocalization data of integrinβ3 and FN at the interface of trophoblasts and endometrial cells further demonstrated that despite the increased levels of integrinβ3 in the absence of gal-3, there was no increase its interaction with FN. Other studies have suggested that integrins bind to galectins, as they are a family of lectins that can associate with the matrix through interaction with LN and FN (Hughes, 1999, 2001; Dumic et al., 2006). Expression of integrins α1β1, α4β1, ανβ3 frame the window of endometrial receptivity that is important for embryo attachment (Lessey, 2002), and gal-3 is also increased during this period. Taken together, not only the expression of gal-3 and its distribution may determine the binding specificity to integrins in the endometrial receptivity window, but also the coordination of integrin expression in time and space would be important.
Endometrial receptivity is established during the mid-secretory phase. Some pregnancy failures are due to a lack of expression of specific critical participating proteins such as integrins which are expressed by endometrium and appear to be necessary for the successful interaction of the embryo with the endometrium. Our study found that in vitro integrinβ3 increased whereas gal-3 expression was knocked down, however, previous research has suggested that in vivo endometrial cells can up-regulate both gal-3 and integrinβ3 at the same time (Lessey et al., 1996; von Wolff et al., 2005; Du et al., 2006). One explanation is the limitation of the in vitro cell model, as RL95-2 is well known to be a carcinoma derived cell line. Though a monolayer of RL95-2 cells mimics much important aspects of the in-vivo situation of a receptive endometrium, it may not well represent the exact in-vivo physiologic consequences. In vivo, integrin ανβ3 expression increases significantly from day LH +6 to day LH +8, and has been proposed as a means of distinguishing receptive endometrium from non-receptive in clinical practice (Ordi et al., 2003). The β3 integrin subunit serves as the rate limiting step in ανβ3 expression, is inhibited by estrogen and stimulated by epidermal growth factor (EGF) (Lessey et al., 2002), suggesting that β3 appears to normally be suppressed in the endometrium by the sex steroids. Regulation of gal-3 and integrinβ3 is likely the same (Somkuti et al., 1997; Deschildre et al., 2007), as EGF seems to increase gal-3 expression (Leach et al., 1999; Chobotova et al., 2002), which suggests that its role may be related to regulation of gal-3 and integrinβ3 expression. Another explanation for increased integrinβ3 expression in gal-3 silenced cells in the present study may be that gal-3 can de-stabilize the integrin or cause internalization or clustering, as opposed to directly regulating its expression (Furtak et al., 2001; Friedrichs et al., 2008).
In summary, our observations have provided an experimental model for the interaction of gal-3 and integrinβ3 in regulating cell proliferation and adhesion to FN.
This work was supported by grants from the Natural Science Foundation of China (NSFC, 30271354) and Shanghai Leading Academic Discipline Project (B117).
We thank Mengxiao Yu and Fuyou Li of the Department of Chemistry & Laboratory of Advanced at Fudan University for their excellent technical assistance in producing confocal images.