The delicate balance between embryo invasion and suppression of maternal immune rejection requires a fully functional decidua in species with haemochorial placenta. Our understanding of the decidual function is very limited due to the molecular and cellular complexity involved in decidualisation. The cell adhesion molecule α_vβ₃ integrin and its ligand vitronectin are upregulated in the mouse decidua during mid-pregnancy. The implications of interactions between α_vβ₃ and vitronectin in regulating decidual function are not known. In the present study, interactions between α_vβ₃ and vitronectin in the decidual cells of the mouse were blocked in vitro and effects on cell fate were evaluated by studying the differentially regulated genes by cDNA array and real-time polymerase chain reaction (PCR). The results indicate that expression of various genes involved in apoptotic and cell cycle pathways, as well as cytokine receptors, was deranged. Signalling through α_vβ₃ seems to be important to maintain a balance between cell proliferation and apoptosis, along with the modulation of inflammatory responses of decidual cells.

Decidualisation is characterised by proliferation and differentiation of endometrial stromal cells into specialised polyploid cells (decidual cells) and finally the programmed death of decidual cells to accommodate the growing embryo. A fully developed decidua is fundamental to continued survival of the embryo and successful pregnancy in species with haemochorial placentation, because it regulates trophoblast invasion and prevents maternal allograft rejection. Common obstetric diseases, such as pre-eclampsia, intrauterine growth retardation, preterm birth and recurrent pregnancy loss, are known to be associated with abnormal or impaired placental and/or decidual function (Dekker and Sibai 1998; Shih and Kurman 2002). Although there is considerable information regarding decidual morphology, the definitive molecular mechanisms underlying decidual function remain to be resolved. Identification of molecules associated with decidualisation may provide valuable insights into decidual function and the maintenance of pregnancy.

Several studies have suggested a role for integrins and their ligands in decidual function. It has been proposed that altering integrin–matrix interactions may hamper decidualisation (Strakova et al. 2003). Endometrial matrix shows enhanced expression of several extracellular matrix (ECM) molecules, such as fibronectin, laminin and collagen IV, during decidualisation, which is also accompanied by switching on of the respective receptor integrins in the transforming decidual cells (Qin et al. 2003). Among these, integrin α_vβ₃ and its ligands vitronectin and osteopontin have generated special interest. This receptor–ligand group is upregulated in the decidualised stroma during pregnancy (Fazleabas et al. 1997; Mangale and Reddy 2007). Knocking out the β₃-subunit in mice results in intrauterine growth retardation, placental defects and reduced embryo survival (Hodivala-Dilke et al. 1999), thus highlighting the role of α_vβ₃ in the maintenance of pregnancy.

Our previous studies demonstrated increased expression of α_vβ₃ integrin and its receptor vitronectin in the mouse decidua capsularis at 8.5 days post coitus (d.p.c.), when decidualisation is at its peak (Mangale and Reddy 2007). These in vivo observations were indicative of some functional relevance of the α_vβ₃–vitronectin interaction in decidualisation. On the basis of results from our previous studies and reports demonstrating altered gene transcription and activation of various signalling events in response to binding of α_vβ₃ to its ligands, such as vitronectin and osteopontin, the present study was undertaken to identify the genes that are activated by an interaction between α_vβ₃ and vitronectin in mouse decidual cells. To this end, the interaction between α_vβ₃ and vitronectin was blocked and transcriptional changes were investigated using cDNA expression arrays spotted with 588 genes for various functional pathways, such as apoptosis, cell cycle regulation, intracellular signalling, tumour suppressor pathway, cytokine pathways and transcription factors.

Animals

The study was approved by the Institutional Ethics Committee for Animal Experimentation. Mature (6–8-week-old) Swiss mice were maintained under a constant photoperiod (14 h light and 10 h dark) and temperature (22–24°C) in the animal house facility of the National Institute for Research in Reproductive Health. Mice were paired for mating overnight. The next morning, mice were checked for vaginal plugs. The day a vaginal plug was found was considered 0.5 d.p.c. of pregnancy. Mice (six in each group) were killed on 6.5, 8.5 and 13.5 d.p.c. to obtain decidua.

Immunohistochemistry

Immunolocalisation of β₃ integrin and vitronectin was performed as described previously (Reddy and Meherji 1999). Sections were deparaffinised in xylene, rehydrated through an alcohol series in distilled water, quenched in 0.3% H₂O₂ in 0.01 m phosphate-buffered saline (PBS), blocked in 1% lamb serum (Bangalore Genei, Bangalore, India) for 1 h and then incubated with rabbit polyclonal antibodies against the β₃-subunit of integrin (1 : 30) or vitronectin (1 : 30; Santa Cruz Biotechnology, Santa Cruz, CA, USA) overnight at 4°C. Negative controls were incubated with 1% lamb serum. Sections were incubated at room temperature for 1.5 h with goat anti-rabbit–horseradish peroxidase (HRP; 1 : 100; Sigma, St Louis, MO, USA). Detection with diaminobenzidine (DAB) substrate was followed by counterstaining with Delafield’s haematoxylin. Sections were observed under an Axioskop 2 plus microscope (Carl Zeiss, Oberkochen, Germany) after mounting in di-butyl phthalate in xylene (DPX) and photographed using a digital still camera (DSC-S75; Sony, Tokyo, Japan). Images were digitised using Photoshop 6.0 software (Adobe Systems, San Jose, CA, USA) for documentation purposes. All chemicals were of analytical grade and were purchased from Merck (Mumbai, India) unless stated otherwise.

Preparation of decidual cells

Decidual capsules were dissected out of the uterus and placed in Dulbecco’s Modified Eagles Medium (DMEM; GIBCO BRL, Grand Island, NY, USA). Each capsule was checked for the presence of a normally developed embryo and the embryonic tissue was removed as completely as possible under a dissecting microscope. Only decidual tissue from a healthy embryonic bead was used further in the preparation of decidual cells. The tissue was minced with scissors in DMEM containing 0.1% collagenase (Sigma Chemical) and 0.1% RNAse-free–DNAse I (Sigma Chemical). Pieces of decidua were incubated in collagenase for 30 min at 37°C, then washed twice in medium and incubated in 0.1% trypsin (Sigma Chemical) in DMEM containing 0.1% DNAse. Incubation in trypsin was followed by gentle pipetting for 3 min in siliconised Pasteur pipettes. The tissue digests were filtered through a 40-μm nylon gauze mesh into DMEM. Cells were centrifuged at 300g and resuspended in DMEM with 0.1% DNAse. The purity of the cells was checked by staining for vimentin and was found to be approximately 99%, with less than 1% non-decidual component. Cell viability was assessed by Trypan blue exclusion and was found to be approximately 97% before culture and approximately 90% after culture. Viable cells were used for further experiments.

In vitro blocking of α_vβ₃ integrin

Decidual cells were pre-incubated in DMEM with 10% fetal calf serum (FCS) for 1 h at 37°C and then divided into two groups. Cells in the control group were incubated with DMEM alone, but cells in the treated group were incubated with anti-β₃-integrin antibody (20 μg mL^–1; subclone 2C9.G2; BD PharMingen, San Diego, CA, USA) at 37°C for 1 h in DMEM. Both groups were seeded on vitronectin (10 μg mL^–1; BD Biosciences, Franklin Lakes, NJ, USA)-coated 24-well culture plates at a concentration of 10⁶ cells per well and cultured for 12 h in DMEM with 10% fetal bovine serum (GIBCO BRL). In the end, the cells were processed for RNA extraction and cDNA array analysis.

Adhesion assays

Adhesion assays were performed to determine whether the β₃ antibody prevents the binding of α_vβ₃-expressing cells to vitronectin. Decidual cells (4 × 10⁵) pre-incubated with either PBS, anti-β₃ antibody (20 μg mL^–1) or α₅-integrin antibody (20 μg mL^–1) (BD PharMingen) were added to the vitronectin-coated plates. After 1 h, plates were washed twice with washing buffer. Cells were fixed with 4% paraformaldehyde (Merck, Mumbai, India) and stained with Crystal Violet (Qualigens Fine Chemicals, Mumbai, India) for 10 min. Plates were measured photometrically at 550 nm after dye solubilisation with Triton-X-100. Differences in binding were considered significant when P ≤ 0.05 (Student’s t-test).

Complementary DNA array

Total RNA was extracted from decidual cells using Tripure reagent (Roche Molecular Biochemicals, Mannheim, Germany) according to the manufacturer’s instructions. Total RNA was converted into cDNA using modified oligo (dT) primer and amplified using the BD SMART PCR cDNA Synthesis Kit (BD Biosciences Clontech, Palo Alto, CA, USA). The amplified cDNA was column purified with a NucleoSpin Extraction kit (BD Biosciences Clontech) and labelled using the BD Atlas SMART probe amplification kit (BD Biosciences Clontech) in the presence of 1.85 × 10⁶ Bq [³²P]-dATP (Board of Radiation and Isotope Technology, Government of India, Hyderabad, India), cDNA synthesis primers (complimentary to the array genes; BD Biosciences Clontech) and 2 units of Klenow fragment at 50°C for 30 min. An equal amount of control and experimental labelled probes were used to hybridise to the Atlas mouse cDNA expression array membrane at 68°C (BD Biosciences Clontech). The next day, washings were performed as per the Atlas protocol and membranes were exposed to a phosphor screen (Fuji-film, Tokyo, Japan) for 12 h. Hybridisation signals were quantified using Atlas Image 2.7 software. Hybridisation signal intensities for genes were subtracted from the background and total signal intensities for the arrays were globally normalised. Signal intensities from control and treated groups were compared and converted into ratios by the software. The hybridisations were performed three times in duplicate. Each time, RNA obtained from a separate blocking experiment was used. Genes showing a greater than twofold change in expression in all three samples were considered as differentially expressed.

Real-time PCR

In order to validate the results of the cDNA array, the expression of six genes (> 10%) that showed differential expression by cDNA array was validated using real-time PCR in three separate experiments. The RNA from control and treated cells was extracted using Tripure reagent (Roche Molecular Biochemicals) and converted into cDNA using AMV Reverse Transcriptase (New England BioLabs, Beverly, MA, USA). The cDNA was further subjected to PCR amplification by the iQ^TMSYBR green supermix and iCycler real-time PCR detection system (BioRad, Hercules, CA, USA). The amplification conditions were as follows: initial denaturation at 95°C for 3 min followed by 40 cycles of denaturation at 95°C for 30 s, annealing at specific temperatures for 30 s (Table 1) and extension at 72°C for 30 s. Final extension was performed at 72°C for 5 min. The fluorescence emitted for each cycle was collected for the entire 30-s period of the extension step in each cycle. The primers used in the PCR (Table 1) were designed using primer3 output online software (http://frodo.wi.mit.edu/). For each primer pair, the reaction efficiency was estimated by amplification of a serial dilution of a mouse decidual cDNA pool over a 10-fold range. The relative levels of cyclin D₃, cyclin F, cyclin G, p53, Rb1 and E2F5 were estimated in relation to GAPDH (a housekeeping gene).

The homogeneity of the PCR amplicons was verified by running the products on 2% agarose gels and also by the melting curve method. All PCR amplifications were performed in duplicate and each experiment was repeated three times to test its reproducibility. Mean threshold cycle values generated in each experiment using the iCycler software (BioRad) were computed and normalised to the housekeeping gene (GAPDH) and were used to calculate cDNA concentrations in the samples. Relative expression ratios were calculated manually using the method of Pfaffl (2001) and considered statistically significant at P ≤ 0.05 (unpaired t-test).

The profiles of differentially expressed genes (cyclins D₃, F and G, p53, E2F5 and Rb1) were evaluated in the developing decidua by real-time PCR analysis. Total RNA was extracted from the decidual tissue on different days (6.5, 8.5 and 13.5 d.p.c.) of pregnancy and subjected to real-time PCR as described previously in the Materials and methods section. The fold change difference was considered to be statistically significant when P ≤ 0.05 (ANOVA).

Functional analysis

In order to assess the functional relevance of differentially expressed genes, downregulated and upregulated groups were subjected to functional analysis using DAVID Functional Annotation Clustering tool (http://david.abcc.ncifcrf.gov). The analysis was performed based on the GO_bp database (Gene Ontology for biological process) and KEGG pathways by selecting appropriate parameters in the tool. The annotation tool was used at highest stringency setting to generate clusters of closely related genes. The statistical significance of each cluster was calculated in terms of P values using a modified Fisher’s exact test. Each cluster was assigned a group Enrichment Score (E-Score), the geometric mean (on a log scale) of the member’s P values in a corresponding annotation cluster, to rank its biological significance. Thus, the top-ranked annotation cluster had an E-score (E ≥ 1) that was consistently higher for its annotation members.

Immunolocalisation of α_vβ₃ integrin and vitronectin in mouse decidual cells

α_vβ₃ Integrin and vitronectin were immunolocalised in decidua at 8.5 d.p.c., reconfirming our previous finding (Mangale and Reddy 2007). The decidua revealed characteristic polyploidal decidual cells with localisation of β₃-integrin and its ligand vitronectin on the surface and in the cytoplasm of decidual cells (Fig. 1).

Adhesion assay

The adhesion assay revealed that pre-incubation of cells with the anti-β₃ antibody inhibited binding to vitronectin-coated plates. In three independent experiments, the anti-β₃ antibody consistently induced a greater than twofold inhibition compared with control. However, the cells retained their binding capacity when they were incubated with anti-α₅ integrin antibody (Fig. 2).

Complementary DNA array analysis

Expression levels of 588 genes were estimated in decidual cells in which the α_vβ₃–vitronectin interaction was inhibited. Of the 588 genes spotted on the array, 53 showed differential expression of greater than or equal to twofold between the two groups: 34 genes were upregulated and 19 genes were downregulated in the antibody treated group compared with the control group. The observed differential expression of 53 genes was consistent in three independent experiments.

Upregulated genes

Genes from various families were upregulated when the α_vβ₃–vitronectin interaction was blocked using a specific antibody (Table 2). The major group comprised the pro-apoptotic genes, TNF-α receptor, APO-1 antigen (FAS) receptor (a member of the tumour necrosis factor (TNF)-α receptor family), interleukin (IL)-converting enzyme (caspase 1), FAS-associated factor 1 (FAF1), UV-excision repair protein (Ras associated with diabetes) (RAD) 21A, RAD 23 and nucleotide diphosphate (NDP) kinase b. Genes like defender against death 1 (Dad1), Retinoblastoma 1 (Rb1) and c-myc from the oncogene and tumour suppressor family were also upregulated. This was accompanied by upregulation of receptors of the cytokine family, such as granulocyte–macrophage colony stimulating factor receptor (GM-CSFR), transforming growth factor (TGF)-β receptor 1 (TGFβR1), and interferon-γ receptor (IFN-γR). A large number of transcription factors was upregulated, the most notable being E2F5, a transcriptional repressor involved in the cell cycle. Genes from the heat shock protein family and motor proteins were also upregulated. P^27kip1, an inhibitor of the cell cycle, was also upregulated.

Downregulated genes

The most remarkable downregulation was that of Shc, an adaptor protein that falls under the α_vβ₃–focal adhesion kinase (FAK)-mediated signalling pathway. Also downregulated was ets-related Proto-oncogene (ELK1) (a member of the erythroblastosis virus E26 oncogene homolog oncogene family), a member of the Shc signalling pathway. Levels of p53, a potent regulator of the cell cycle, were also decreased. Interestingly, genes belonging to the cell cycle regulator family cyclin D₃, F, G1 and G2 were downregulated, along with cdc25b (Table 2).

Validation of array results by real-time PCR

Real-time PCR profiles of the cyclins, p53, Rb1 and E2F5 in control and antibody treated decidual cells corroborated cDNA array data. Expression levels of cell cycle molecules cyclin D₃, F and G and p53 were less than half in β₃ antibody blocked decidual cells compared with the control group. The expression of Rb1 and E2F5 was significantly upregulated in antibody treated decidual cells compared with control (Fig. 3).

Real-time PCR analysis of cyclins, p53, Rb1 and E2F5 in decidua at various embryonic ages

There were dynamic changes in the expression of genes for the cyclins, p53, Rb1 and E2F5 in mouse decidua at different ages. Expression of cyclin D₃, F and G and p53 was maximal and significantly higher at 8.5 d.p.c. compared with that on 6.5 and 13.5 d.p.c. Conversely, expression of Rb1 and E2F5 was significantly lower at 8.5 d.p.c. compared with that on 6.5 and 13.5 d.p.c. (Fig. 4).

Functional annotation

It was evident from functional annotation clustering that most of the upregulated genes clustered into 12 functional groups, which had a significant number of genes (E ≥ 1). These genes are known to play a role in cellular metabolism, transcriptional regulation or apoptosis (Table 3). Conversely, among the downregulated genes, most significant clusters comprised genes involved in cell cycle regulation and cellular physiological processes.

The α_vβ₃ integrin has well-documented roles in various cellular processes, including proliferation, differentiation and apoptosis (Stupack et al. 2001; Faccio et al. 2003; Sahni and Francis 2004). Because α_vβ₃ and its ligand vitronectin are coexpressed in mouse decidual cells on the cell surface as well as in the cytoplasm (Fig. 1), it is conceivable that this receptor–ligand interaction may be playing a similar role in decidual function. To further our understanding of this receptor–ligand pair in decidual cell functioning, changes in gene expression profiles in response to α_vβ₃–vitronectin interactions in mid-gestation murine decidua were investigated. To this end, the α_vβ₃–vitronectin interaction in decidual cells was blocked by treating cells with an α_vβ₃ integrin-blocking antibody in vitro and the expression of 588 genes was assessed by cDNA expression array. Adhesion assays confirmed that α_vβ₃ integrin was blocked and our experimental set up was valid (Fig. 2). Complementary DNA array results revealed that disruption of the vitronectin–α_vβ₃ interaction in mid-trimester decidual cells in vitro results in marked changes in expression profiles of genes belonging to multiple pathways, mainly cellular proliferation and apoptosis.

Using the above assay system, we identified several key molecules whose expression was affected by the disruption of the α_vβ₃–vitronectin interaction. Fifty-three genes (approximately 9%) showed differential expression when the interaction between α_vβ₃ and vitronectin was inhibited (Table 2). Among these, the expression of 34 genes was consistently upregulated and that of 19 genes was consistently downregulated in three independent experiments. The results of cDNA arrays were also validated by real-time PCR, in which the expression of cyclins D₃, F and G and p53 was found to be downregulated, whereas that of E2F5 and Rb1 was found to be upregulated after α_vβ₃ blockade. Although in quantitative terms the fold change observed was not in complete agreement, the pattern of differential expression was found to be similar between the cDNA array and real-time PCR. Similar quantitative discrepancies between the data derived by real-time PCR and cDNA arrays have been reported previously (Rajeevan et al. 2001; Andersen et al. 2004). These discrepancies have been explained by differences in the sensitivity of the two methods, the dynamic range of detectors and the mode of assessment (hybridisation v. amplification).

The biological significance of the differentially expressed genes was assessed using DAVID functional annotation software. It was evident that most of the upregulated genes belonged to clusters involved in cellular metabolism, transcriptional regulation and apoptosis, whereas the downregulated genes fell in two significant clusters that included genes involved in cell cycle regulation and cellular physiological processes. Thus, it appears that α_vβ₃–vitronectin signalling directly or indirectly regulates the expression of some molecules involved in these pathways to modulate decidual functions. Indeed, among the downregulated genes, Shc1 and Elk-1 are known to be involved in α_vβ₃-mediated signalling pathway for the regulation of cell proliferation (Wary et al. 1996; Aplin et al. 2001). These genes, along with those for cyclins D₃, F, G₁ and G₂, were clustered under the group of cell cycle regulators. Cyclin D₃, cyclin G₁ and G₂ are known to be present in high levels in proliferating decidual cells during mid-pregnancy (Das et al. 1999; Yue et al. 2005), when decidual proliferation is at its peak (Abrahamsohn and Zorn 1993; Correia-da-Silva et al. 2004), whereas cyclin F-null mice fail to survive owing to defects in placentation (Tetzlaff et al. 2004). Taken together, these data suggest that, during mid-pregnancy, proliferation of decidual cells requires the coordinated activity of cyclins, which are possibly regulated by the α_vβ₃–vitronectin interaction in decidual tissue.

Cyclins are proximal members of the cell cycle regulation assembly, which, in turn, activates or suppresses downstream pathways that comprise molecules like Rb1, E2F5 and p53 (Sears and Nevins 2002). Functional annotation clustering of the upregulated genes identified a cluster that was comprised of the genes Rb1 and E2F5, which are involved in the negative regulation of the cell cycle (Table 3). Rb1 is a well-studied anti-proliferative molecule involved in cell cycle regulation (Kato 1999), whereas E2F transcription factors are placed furthest downstream in the cell cycle regulatory mechanism. Of the E2F family, E2F5 is known for transcriptional repression (Sears and Nevins 2002). Upregulation of Rb1 and E2F5, along with cell cycle inhibitor p27^Kip1, as indicated by cDNA array analysis further strengthens our hypothesis that blocking of α_vβ₃ leads to inhibition of cell cycle events in the decidua. Complementary DNA array and real-time PCR analysis also indicated downregulation of p53. Increased levels of p53 have been reported in in vitro decidualised human endometrial stromal cells and are speculated to be of relevance in the decidua (Pohnke et al. 2004). However, it remains to be investigated whether this change in p53 expression is a result of direct transcriptional regulation by α_vβ₃ or through cyclins.

Because inhibition of the cell cycle was the most striking effect observed after interactions between α_vβ₃ and vitronectin had been blocked and the genes for cyclins D₃, F and G₁, p53, Rb1 and E2F5 occupy different niches in the cell cycle regulatory machinery, the expression profiles of these genes were validated by real-time PCR. The results of the PCR were in agreement with those from cDNA array (Fig. 3). To gain further insight into the regulation of these molecules, we investigated the expression profile of these genes (cyclin D₃, cyclin F, cyclin G₁, p53, Rb1 and E2F5) during the proliferative (6.5 and 8.5 d.p.c.) and regressive (13.5 d.p.c.) phases in the decidua (Fig. 4). The results show that, at 8.5 d.p.c., when decidual proliferation has reached its peak, the expression of cyclins D₃, F and G₁ and p53 was also highest, whereas the expression of the cell cycle inhibitory molecules Rb1 and E2F5 was lowest. The expression levels coincided with those of α_vβ₃ and vitronectin. Conversely, in the regressive phase, at 13.5 d.p.c., when the levels of α_vβ₃ and its ligands have diminished, the expression of the cyclins and p53 dropped, with a concomitant increase in the mRNA levels for Rb1 and E2F5. These observations collectively support our hypothesis that cell proliferation in the decidua may be regulated, in part, by interactions between α_vβ₃ and vitronectin.

Functional annotational clustering placed the pro-apoptotic genes FAS, caspase-1, c-myc, RAD 21, RAD 23 and NDP kinase b in a single group with high statistical significance (Table 3), indicating a shift towards apoptosis. This was further supported by the upregulation of genes representing receptors for the inflammatory cytokines TGF-β1 and IFN-γ, which promote apoptosis in endometrial stromal cells in vitro (Moulton 1994; Christian et al. 2001; Chatzaki et al. 2003). Together, these data leads us to postulate that the α_vβ₃–vitronectin interaction may be crucial for maintaining the balance between cell proliferation and apoptosis in decidua and, when this interaction is perturbed, pro-apoptotic pathways predominate in decidual cells.

The decidua is known to undergo sequential proliferation, differentiation and death from implantation to the end of pregnancy. Initial attachment of the embryo to the uterus triggers proliferation of endometrial stromal cells, followed by their transformation into decidual cells. During the course of embryonic invasion, the matrix metalloproteases secreated by the trophoblast cells degrade the ECM, thereby perturbing integrin–ECM interactions. On the basis of the present findings, it may be envisaged that the complex interplay between α_vβ₃ integrin and vitronectin, or a perturbation in their interaction caused by matrix degrading enzymes produced by the embryo, is crucial for the dynamic balance between proliferation and death in the decidua. This may be crucial for controlling embryonic invasion, a disturbance that may lead to overinvasion or termination of the pregnancy. To the best of our knowledge, this is the first study that identifies the putative pathways driven by ECM–integrin interactions in murine decidual cells. Establishing the role of the downstream targets of the α_vβ₃–vitronectin interaction under in vivo conditions may further unravel the intricacies of decidual function.

The present study was funded by the Indian Council of Medical Research. The authors are grateful to Dr Chander P. Puri (Director, NIRRH) for support and encouragement. SSM was supported by the Council for Scientific and Industrial Research (India) with a Senior Research Fellowship.