Functional Effects of Transforming Growth Factor on Adhesive Properties of Porcine Trophectoderm
http://www.100md.com
《内分泌学杂志》
Department of Veterinary Integrative Biosciences, College of Veterinary Medicine and Biomedical Sciences (L.A.J., A.K.S., N.H.I., G.A.J., R.C.B.), Department of Animal Science, College of Agriculture and Life Sciences (L.A.J., N.H.I., F.W.B.), and Center for Animal Biotechnology and Genomics (L.A.J., N.H.I., G.A.J., F.W.B., R.C.B.), Texas A&M University, College Station, Texas 77843
Abstract
In pigs, expression and amounts of biologically active TGFs at the conceptus-maternal interface increase significantly as conceptuses elongate and begin the implantation process. Before their activation, secreted TGFs are noncovalently associated with their respective, isoform-specific latency-associated peptides (LAPs), which contain the Arg-Gly-Asp (RGD) amino acid sequence that serves as a ligand for numerous integrins. Objectives of this study were to determine whether TGF1 increases production of fibronectin by porcine trophectoderm, whether porcine trophectoderm adheres specifically to fibronectin and LAP, and whether functional interactions between porcine trophectoderm and the two TGF-associated proteins, fibronectin and LAP, are integrin mediated. Porcine trophectoderm cells (pTr2) were cultured in presence of TGF1, LAP, or pan-neutralizing anti-TGF antibody; TGF specifically increased (P < 0.05) fibronectin mRNA levels, as determined by Northern and slot blot analyses. Immunofluorescence microscopy demonstrated a TGF-induced increase in fibronectin in pTr2 cells. In dispersed cell adhesion assays, adhesion of pTr2 cells to fibronectin was inhibited by an RGD-containing peptide (P < 0.05) and pTr2 cells attached to recombinant LAP but not to an LAP mutant, which contained an RGE sequence rather than the RGD site (P < 0.05). Fibronectin- and LAP-coated microbeads induced integrin activation at apical surfaces of both trophectoderm and uterine luminal epithelial cells, as indicated by aggregation and transmembrane accumulation of talin detected with immunofluorescence microscopy. Cell surface biotinylation and immunoprecipitation revealed integrin subunits v and 1 on apical membranes of pTr2 cells. These results suggest multiple effects of TGF at the porcine conceptus-maternal interface, including integrin-mediated conceptus-maternal communication through LAP.
Introduction
THE TGFs AND THEIR RECEPTORS are present at the conceptus-maternal interface in numerous species (1, 2, 3), but their functional roles during early pregnancy are unclear. TGFs are widely known as multifunctional growth factors. They are of particular interest during early pregnancy because of their effects on proteases and protease inhibitors, which are involved in implantation, as well as their potential to alter expression of integrins and extracellular matrix molecules that appear to play critical roles in the implantation process (3, 4).
The three mammalian isoforms of TGF (TGF1, 2, and 3), along with TGF receptors type I and II, are present on both sides of the conceptus-maternal interface in pigs during the peri-implantation phase of pregnancy (5, 6, 7). Accordingly, the TGFs may exert both autocrine and paracrine actions on the conceptus and/or endometrium during early pregnancy. Expression of all TGF isoforms increases as conceptuses undergo their remarkable morphologic transformation from spherical to elongated filamentous forms and attach to the endometrial luminal epithelium (6, 7, 8, 9). Because TGFs are secreted in latent forms and only activated forms of TGFs can bind to and signal through type I and II receptors, activation is an essential step in controlling the availability and receptor-mediated actions of these growth factors (10, 11, 12, 13). Interestingly, in addition to the increase in absolute amounts of total TGFs, amounts of biologically active TGFs at the conceptus-maternal interface in pigs also increase significantly as conceptuses elongate and begin the attachment and implantation process, implying a role for TGFs during this critical phase of pregnancy (6, 7). The concurrent increase in expression of some integrin subunits on the uterine luminal epithelium and the expression of osteopontin, fibronectin, and other extracellular matrix molecules at the conceptus-maternal interface may be in response to TGFs (14, 15, 16). This suggests that TGFs influence integrin-mediated signaling and conceptus attachment in pigs, which have central implantation and develop a noninvasive epitheliochorial placenta.
The secreted TGFs are noncovalently associated with their respective isoform-specific latency-associated peptides (LAPs), which form linkages with latent-TGF- binding proteins; the resultant complex may further link to and become immobilized in the extracellular matrix (10, 11, 12, 13). The LAPs of TGFs 1 and 3 contain the Arg-Gly-Asp (RGD) amino acid sequence that serves as a ligand for numerous integrins. These LAPs, as free molecules or as part of latent TGF complexes, bind to and initiate signals through activation of particular integrins, and some LAP-integrin interactions have been shown to initiate functional cell signaling for cell migration, spreading, and proliferation (17, 18, 19, 20, 21, 22, 23, 24). Such cellular processes are critical to implantation and, in particular, differentiation and attachment of trophectoderm to uterine luminal epithelium, but this potential novel function for TGF during implantation has not been explored.
Our hypothesis was that TGFs function through multiple cell signaling pathways to stimulate conceptus elongation and attachment, as well as effect changes in conceptus and uterine gene expression critical to conceptus survival and successful establishment of pregnancy. We focused on TGF1 and its LAP because this isoform is abundant at the porcine conceptus-maternal interface and is the best characterized of the mammalian TGFs. The specific objectives of this study were to determine whether: 1) TGF1 increases production of fibronectin by porcine trophectoderm, 2) porcine trophectoderm adheres specifically to fibronectin and LAP, and 3) whether functional interactions between porcine trophectoderm and the two TGF-associated proteins, fibronectin and LAP, are integrin mediated.
Materials and Methods
Porcine trophectoderm cell culture
The spontaneously immortalized porcine trophectoderm cell line pTr2 was cultured as previously described (25) in phenol red-free DMEM-F12 containing 5% charcoal-stripped fetal bovine serum, 2 mM glutamine, and 0.1 U/ml bovine insulin. Cells were maintained and all culture experiments were conducted in a 5% CO2 humidified environment.
For determination of the effect of TGF on fibronectin mRNA levels, cells were plated at approximately 15,000 cells/cm2 and, after two days of culture, were changed to serum-free DMEM-F12 containing 0.1% BSA (DMEM-F12/BSA) and 2 mM glutamine. After 24 h of serum deprivation, cultures were treated with recombinant human TGF1 (Life Technologies, Inc., Gaithersburg, MD; 0, 0.1, 1.0, or 10.0 ng/ml) for 24 h in DMEM-F12/BSA (n = 3 separate experiments). Subsequent experiments were similarly conducted in which cultures were treated for 24 h in DMEM-F12/BSA (n = 3 separate experiments) with recombinant human TGF (0 or 1.0ng/ml) and either TGF pan-neutralizing antibody (0.15 μg/ml; AB-100-NA; R&D Systems, Minneapolis, MN), control rabbit Ig (IgG; 0.15 μg/ml), or recombinant simian TGF LAP (0 or 2.5 ng/ml) that was produced by Sf9 insect cells (Invitrogen, Carlsbad, CA) infected with a recombinant baculovirus (generously provided by Dr. J. S. Munger, New York University School of Medicine, New York, NY) and purified as described by Munger et al. (18).
For determination of the effect of TGF on fibronectin protein expression, pTr2 cells were similarly plated on Lab-Tek chamber slides (Nalge Nunc International, Naperville, IL), serum-deprived, and treated with TGF (0, 1.0, or 10 ng/ml) for 24 h.
Immunofluorescence
Expression of TGFs by porcine trophectoderm in vivo has been demonstrated (5, 6). To determine whether pTr2 cells maintained expression of TGF1 in culture, cells were seeded on four-well Lab-Tek glass chamber slides, cultured for 48 h, fixed with ice-cold methanol, and subjected to immunofluorescence analysis as previously reported (14, 15), and with anti-TGF1 (10 μg/ml; AB-100-NA; R&D Systems) or control rabbit IgG used at the same concentration as the primary antibody.
To further characterize the pTr2 cells, immunofluorescence microscopy analysis on ice-cold methanol-fixed cells was conducted using antibody directed against cytokeratin 7 (2 μg/ml; clone OV-TL 12/30; InnoGenex, San Ramon, CA) and, on paraformaldehyde-fixed cells, using monoclonal antibody SN1/38 (undiluted tissue culture supernatant, approximately 10 μg/ml, kindly provided by A. Whyte, The Babraham Institute, Cambridge, UK), which identifies a porcine trophectoderm-specific antigen (26), or with control mouse IgGs. Additionally, expression and distribution of fibronectin in TGF-treated cells was evaluated via immunofluorescence using antibody against fibronectin (2 μg/ml; Neomarkers Ab-2, clone HFN36.3; Labvision Corp., Fremont, CA).
Slides were covered with commercially obtained mounting medium (Prolong Gold antifade solution with or without the nuclear counterstain 4',6-diamidino-2-phenylindole; Molecular Probes, Eugene, OR) and protected by coverslips. Reactions were visualized and imaged using a Zeiss Axioplan 2 microscope (Carl Zeiss, Thornwood, NY) fitted with an Axiocam high-resolution digital camera. Digital images were captured using Axiovision 3.0 software.
For analysis of the effects of 0, 1.0, and 10 ng/ml TGF on fibronectin staining in pTr2 cells, digital fluorescence images of continuous monolayers were recorded using the same instrument settings. Images were adjusted electronically using the same brightness and contrast instrument settings to optimize visualization of extracellular fibronectin at the basal aspect of cells. These fibronectin staining patterns were then converted to binary images before quantification of the area of fibronectin staining in each field of cells. Data were analyzed in two separate experiments on different days. Values for fibronectin staining areas in 1.0 and 10 ng/ml TGF-treated cells were normalized to 0 ng/ml TGF.
Northern and slot blot analysis
Total cellular RNA was isolated from pTr2 cells using TriPure reagent (Roche Applied Science, Indianapolis, IN) as described by the manufacturer. Steady-state expression of fibronectin was determined by Northern and slot blot analyses as previously described (25, 27). Briefly, 10 μg total RNA per lane for Northern analysis and 2 μg total RNA per slot for slot blot analysis were hybridized with 32P-labeled antisense cRNA probes generated against a linearized equine fibronectin cDNA (GenBank accession no. U52107; generously provided by J. N. MacLeod, Cornell University, Ithaca, NY) and 28S ribosomal RNA (27). Blots were exposed to autoradiography film (Kodak BioMax XAR; Eastman Kodak, Rochester, NY), and radioactivity of each slot was quantitated using an InstantImager (Packard, Meridian, CT). Hybridization signals of fibronectin mRNA were normalized to those of 28S RNA to account for loading differences among samples.
Adhesion assays
Ninety-six-well suspension culture plates (Greiner Labortechnik; PGC Scientific, Frederick, MD) were coated with solutions of bovine fibronectin (no. F1141; Sigma, St. Louis, MO), LAP, or mutant LAP, which contained an RGE sequence rather than the integrin-binding RGD site (LAP-RGE; produced by Sf9 insect cells infected with a recombinant baculovirus provided by Dr. J. S. Munger) at concentrations ranging from 0–10 μg by incubating 100 μl per well of each solution overnight at 4 C on an orbital shaker. Wells were then washed three times with Dulbecco’s PBS (DPBS; Life Technologies, Inc.) blocked with 1% BSA in PBS for 1 h at 37 C, and adhesion assays (18) conducted using pTr2 cells from subconfluent cultures. Cells were rinsed three times with calcium-magnesium-free DPBS (Ca-Mg-free DPBS; Life Technologies, Inc.) then incubated 5 min at 37 C, in Ca-Mg-free DPBS containing 0.02% EDTA, followed by a 30-sec to 1-min incubation with 0.12% trypsin. Flasks were tapped gently to dislodge the cells, which were then dispersed in serum-containing medium to stop trypsin activity. Cells were washed twice by centrifugation in DMEM-F12/BSA and then suspended at 300,000 cells/ml in that same medium. In some cases, the peptide inhibitor GRGDSPL (no. G-1269; Sigma) or GRADSPL control peptide (no. G-4144; Sigma) was added at concentrations ranging from 0–100 μg/ml and incubated at 22 C for 15 min. Otherwise, 100 μl of cell suspension was immediately distributed to each well, in triplicate, and plates were incubated 1.5 h at 37 C in a humidified incubator containing 5% CO2. Wells were washed twice with DPBS to remove nonadherent cells, and cells were then fixed with 4% paraformaldehyde in PBS for 30 min, permeabilized with 20% methanol in distilled water for 20 min, and stained with 0.5% crystal violet in 20% methanol for 20 min. Excess stain was removed by repeated washing with distilled water, and plates were allowed to dry overnight. Cell binding was quantified by eluting the crystal violet with 100 mM sodium citrate, pH 4.0, in 50% ethanol and measuring absorbance at 595 nm on an E-Max Precision Plate Reader (Molecular Devices Corp., Menlo Park, CA). Baseline absorbances of control wells that were coated only with BSA and were not incubated with cells were subtracted from each measurement.
Analysis of integrin activation
Polystyrene beads (6.0 μm; Polysciences Inc., Warrington, PA) were prepared and coated with 100 μg/ml of either fibronectin, LAP, or poly-L-lysine (Sigma), which allows nonintegrin-mediated adhesion, as previously reported (15). Porcine trophectoderm cells were seeded on two-well Lab-Tek coverglass chamber slides and cultured for 48 h as described previously, then washed with and changed to DMEM-F12/BSA. Likewise, to further explore the possible roles of LAP and fibronectin at the conceptus-maternal interface, primary cultures of porcine uterine luminal epithelium were established and cultured as previously reported (28) and then washed and changed to DMEM-F12/BSA for subsequent analysis of activation of integrins on the apical membrane of uterine luminal epithelium.
The matrix-coated beads were suspended in DMEM-F12/BSA (106 beads/ml) and incubated with the cells for 1 h. Cultures were then fixed, without prior rinsing, in 4% paraformaldehyde in PBS and subjected to immunofluorescence procedures using an antibody to the cytoskeletal protein talin (antitalin clone 8d4; Sigma) to evaluate integrin activation in response to apical cell membrane contact with each population of matrix-coated beads, as previously described (15, 29). Talin aggregation at focal adhesions requires both integrin ligand occupancy and integrin aggregation (29). Fluorescence imaging of the cultures was performed using a digital fluorescence imaging system consisting of a charge-coupled device camera and image capture software (CELLscan; Scanalytics, Bedford MA) integrated with a Zeiss Axiovert inverted fluorescence microscope. Previously reported methods (15, 30) were used to quantify the percentage of beads with talin immunofluorescence reactions that indicated apical focal adhesion formation in response to integrin activation. Briefly, image collection began 1 μm below the basal surface of the cell and optical slices were collected at 0.5-μm steps up through the apical surface and attached beads. Phase contrast microscopy was first used to locate cells and count the total number of beads adherent to apical aspects of the cells. Fluorescence microscopy was then used to identify transmembrane accumulation of immunoreactive talin at the apical cell membrane-matrix-coated bead interface. For each treatment group, the percentage of ligand-coated beads in contact with cells that exhibited apical "focal adhesions" was obtained by multiplying the ratio of focal adhesion-positive beads to the total number of beads in contact with the cell by 100. For each of the ligands tested, at least four separate experiments on different days were performed with at least three slides per ligand in each experiment. Data are presented as mean percentage of beads ± SD inducing apical focal adhesions.
Analyses of integrin expression
Because the integrin heterodimer v1 is a known receptor for both TGF LAP and fibronectin, cell surface expression of the individual constituent subunits v and 1 was assessed on porcine trophectoderm after cell surface biotinylation and immunopreciplitation. Confluent monolayers of pTr2 cells were washed with Ca-Mg-free DPBS, and cell surface proteins were labeled using a modification of procedures reported by Munger et al. (18) and recommendations by the manufacturer of the biotinylation reagent. Briefly, cells were rinsed with 5 mM EDTA in Ca-Mg-free DPBS at 22 C for 1 min, followed immediately by two brief rinses with DPBS containing calcium and magnesium. Each 75-cm2 flask was then treated with 4 ml of freshly prepared DPBS containing 0.25 mg/ml of membrane impermeable biotin (N-hydroxysulfosuccinimidobiotin; EZ-Link Sulfo-NHS-Biotin; Pierce Biotechnology, Inc., Rockford, IL) or DPBS only, and incubated on a rocking platform, shielded from direct light, at 22 C for 1 h. After removal of the biotin solution, cells were rinsed twice with cold DPBS, once in 0.1 M glycine in cold DPBS, and then twice in cold Tris-buffered saline (TBS) with 1 mM calcium chloride (TBS-Ca) containing 0.02% sodium azide. After complete removal of the last TBS-Ca rinse, cells were lysed with 50 mM octyl--D-thioglucopyranoside (Calbiochem, La Jolla, CA) containing 1 mM each CaCl2, MgCl2, and MnCl2 for 30 min on an orbital shaker at 4 C. Lysed cells were scraped from the flasks and passed through a 25-gauge needle four times. Lysates were centrifuged at 16,000 x g, 20 min, 4 C, and total protein in lysates was determined (Pierce Coomassie Plus Protein Assay Reagent; Pierce Biotechnology, Inc). Portions of the lysates were run on reducing SDS-PAGE (7.5%), blotted to nitrocellulose, and incubated with horseradish peroxidase avidin D (avidin-HRP; Vector Laboratories, Burlingame CA) and chemiluminescent detection reagent (SuperSignal West Pico Substrate; Pierce Biotechnology, Inc.) to confirm biotinylation. Biotinylated lysates were immunoprecipitated with polyclonal antibodies to integrin 1 (no. 1952; Chemicon International, Temecula, CA), integrin v (no. 1930; Chemicon International), or control normal rabbit serum and a protein A-protein G agarose conjugate (Protein A/G plus; Santa Cruz Biotechnology, Santa Cruz CA), using procedures recommended by the supplier of the agarose conjugate. The immunoprecipitates were run on 7% nonreducing SDS-PAGE, 10% nonreducing SDS-PAGE, or 10% reducing SDS-PAGE, blotted, and incubated with avidin-HRP as described previously, and bands were visualized by exposure to autoradiography film (BioMax; Eastman Kodak).
Statistical analyses
Dose response data were subjected to least-squares ANOVA using General Linear Models procedures of the Statistical Analyses System. All effects of treatment were determined using PROC MIXED procedure of SAS (SAS Institute Inc., Cary, NC), and significant treatment effects (P < 0.05) separated using least squares means and the PDIFF procedure of SAS. Statistical significance was set at P < 0.05 for all assays. Data charts were generated using Microsoft Excel (Microsoft Corp., Redmond, WA); for clarity of presentation, data are expressed presented as arithmetic means with SEM.
Results
Characterization of pTr2
The epithelial phenotype of the pTr2 cells was confirmed by positive immunoreactions for cytokeratin 7, and identity as trophectoderm was confirmed by positive immunoreactions for SN1/38 (Fig. 1). Reactivity for SN1/38 was predominantly apical, and presence and intensity of the immunoreaction varied markedly among cells, which is consistent with previous reports for cultured porcine trophectoderm (26, 31, 32) (Fig. 1). TGF localized to perinuclear cytoplasmic and Golgi regions and to margins of the pTr2 cells (Fig. 2).
Effect of TGF on pTr2 fibronectin expression
The specificity of the fibronectin probe was confirmed by Northern blot hybridization and stringent washing, and the same conditions were used for slot blot analyses. One major transcript of approximately 7 kb and one less prevalent, slightly larger transcript were identified on Northern blots, consistent with previously reported splice variants of fibronectin mRNA (33) (Fig. 3A). Treatment of pTr2 cells with TGF for 24 h increased steady-state levels of fibronectin mRNA (Fig. 3B). Effect of treatment was linear (P = 0.001); however, when treatment effects were separated by dose, analysis of initial experiments indicated that 1.0 and 10 ng/ml TGF increased levels of fibronectin compared with vehicle-treated controls (P = 0.003, P = 0.001, respectively) whereas effect of 0.1 ng/ml TGF was not statistically significant (P = 0.143) and effect of 10 ng/ml TGF was not greater than that of 1.0 ng/ml treatment (P = 0.395) (Fig. 3, A and B). Therefore, subsequent experiments to verify specificity of the response used the 1.0-ng/ml dose of TGF (Fig. 3C). Specificity of the increase was confirmed using LAP, which neutralizes the biological activity of TGF, and with a pan-neutralizing TGF antibody. Concurrent treatment with LAP decreased fibronectin levels compared with effects of TGF alone, whereas treatment of control cultures with LAP had no effect on fibronectin mRNA compared with control values (Fig. 3C). In immunoneutralization experiments, addition of pan-neutralizing TGF antibody to TGF-treated cultures significantly decreased fibronectin mRNA levels compared with cultures treated with TGF and control IgG, but did not differ from control cultures treated with vehicle concurrent with control IgG or pan-neutralizing TGF antibody (Fig. 3D).
Immunofluorescence microscopy using antibody specific for fibronectin revealed basal levels of fibronectin in control cultures, which increased in response to TGF (Fig. 4). Immunoreactive fibronectin was localized to basal cell attachment sites associated with focal adhesions, cytoplasmic domains, and extracellular cell surfaces. Fibronectin staining at the basal aspects of pTr2 was increased 6.7-fold by 1.0 ng/ml TGF, whereas 10 ng/ml TGF resulted in a 64.4-fold increase in the area of fibronectin staining at the basal cell surfaces. Further analysis was not conducted because of the visual magnitude of the response and the close agreement of the protein expression results with those of the fibronectin mRNA expression.
Trophectoderm adhesion of fibronectin and LAP
Adhesion of pTr2 cells to fibronectin increased with increasing concentrations of fibronectin substrate up to 5 μg/ml (P = 0.026, quadratic) (Fig. 5A). Experiments to determine mechanism of attachment to fibronectin were conducted using 5 μg/ml. Cell adhesion to fibronectin was inhibited in a linear manner (P = 0.0001) by the peptide containing the integrin-binding amino acid sequence Arg-Gly-Asp (RGD) but not by the control peptide containing an Arg-Ala-Asp (RAD) sequence (Fig. 5B).
All concentrations of LAP tested supported trophectoderm adhesion (quadratic, P = 0.01), but adhesion was not significantly different among the LAP concentrations tested. All tested concentrations of the LAP mutant, in which the RGD sequence was replaced by an Arg-Gly-Glu (RGE) sequence, failed to support adhesion (Fig. 6).
Functional activation of integrin signaling by fibronectin and LAP
The aggregation of the cytoskeletal protein talin at the interface between apical cell membranes and matrix-coated beads, as visualized using combined immunofluorescence and phase contrast microscopy, indicated integrin-mediated signaling in response to both fibronectin and LAP in both epithelial cell types of the conceptus-maternal interface (Table 1 and Fig. 7). Approximately 56% of LAP-coated beads in contact with apical membranes of adherent pTr2 cells and 22% of LAP-coated beads in contact with and porcine uterine luminal epithelial cells elicited talin aggregation at the bead-cell interface. Approximately 39 and 19% of fibronectin-coated beads elicited similar responses in pTr2 and porcine uterine luminal epithelial cells, respectively. Talin aggregation was not detected in either cell type at the interface between apical cell membranes and poly-L-lysine-coated beads (Table 1).
Expression of integrin subunits on apical pTr2 membranes
Immunoprecipitation of biotinylated pTr2 lysates with anti-integrin 1 yielded a labeled band at a relative molecular mass of approximately 110 x 103 on a nonreducing gel (Fig. 8), consistent with reported relative molecular mass of integrin 1 (18), whereas no bands were obtained with the normal rabbit serum control. In a separate experiment, immunoprecipitation with anti-integrin v yielded a major band at the expected relative molecular mass of approximately 150 x 103 and a lesser band at approximately 110 x 103, consistent with coimmunoprecipitation of a integrin subunit along with the v subunit. Use of a reducing gel to aid in identifying the labeled proteins after immunoprecipitation of biotinylated pTr2 lysates with anti-integrin v revealed two labeled bands consistent with the reported relative molecular masses of integrin v, which is posttranslationally cleaved to yield fragments with relative molecular masses of 125 x 103 and 27 x 103 when analyzed under reducing conditions (34). No labeled protein bands were detected in normal rabbit serum controls.
Discussion
This is the first known report that demonstrates a direct functional role for the LAP of TGF in the reproductive system. Results of the present study indicate that trophectoderm adheres to LAP in an integrin-dependent manner and that LAP contact with apical surfaces of both trophectoderm and uterine luminal epithelium elicits a functional response (i.e. rapid induction of focal adhesion formation). Furthermore, results demonstrated that TGF stimulates production of fibronectin by porcine trophectoderm in early stages of pregnancy, that fibronectin support of trophectoderm adhesion is via a mechanism partially dependent on integrins, and that apical contact of both trophectoderm and uterine luminal epithelium results in cytoskeletal rearrangement indicative of integrin-mediated signaling.
We previously confirmed the epithelial phenotype of the pTr2 cells based on immunoreactivity with an antibody that recognizes a broad spectrum of epithelial cytokeratins and verified the trophectoderm origin of the cells by demonstrating that they express interferon mRNA, an interferon unique to porcine trophoblast (25). In this study, we also report in pTr2 cell the presence of cytokeratin 7, a preferred marker for human trophoblast cells (35, 36). Identity of the cells as trophectoderm was further verified by positive immunoreactivity with the SN1/38 antibody. The microvillar localization and the cell-to-cell variations in intensity of the SN1/38 immunofluorescence is virtually identical to that previously reported for porcine trophectoderm cells in culture (31, 32). Although the specific identity of the molecule recognized by SN1/38 monoclonal antibody is not known, it is porcine trophectoderm specific (26). These findings support use of the pTr2 line as an in vitro model for studying the functional biology of pig trophectoderm, which is an early and critical participant in establishing blastocyst interactions with the uterine luminal epithelium.
The TGF antibody used for neutralization and immunofluorescence recognizes all mammalian forms of TGF and does not distinguish between latent and active forms. The pattern of TGF localization suggests that once secreted, TGF is bound to trophectoderm cell membranes. The current experiments did not address whether the extracellular immunoreactive TGF was bound to TGF receptors, integrins, or other binding proteins expressed by the pTr2 cells. These results raise several possibilities for autocrine actions of TGF produced by trophectoderm cells, including direct integrin-mediated effects on their migration and/or adhesion.
Integrin heterodimers containing the v subunits bind LAP (18, 19, 20, 24), and some of the constituent subunits, including v1 and v3, are present on uterine luminal epithelium and conceptus trophectoderm during early pregnancy (14). Our characterization of integrins on the cell surface of porcine trophectoderm cells is not complete, but based on the relative molecular masses of the labeled proteins, current results indicate that integrins v and 1 are expressed on apical membranes of the pTr2 cells and could be responsible, in part, for adhesion and functional response to LAP. Definitive characterization via Western blot analysis of the immunoprecipitates was hampered by the lack of multiple commercially available antibodies that recognize porcine integrins and were generated from different host species and/or would recognize integrin subunits under reducing conditions. Nevertheless, results from both immunoprecipitations are consistent and support expression of integrins v and 1 on the cultured apical trophectoderm cells, and are also consistent with in vivo evidence for integrin expression on porcine conceptuses (14).
Trophectoderm cells are critical to implantation because they must communicate with and adhere to maternal uterine epithelium. The retention of TGF production and responsiveness by the pTr2 cells, in combination with expression of some of the known integrins and matrix proteins present in porcine trophectoderm in vivo suggest that this model system is valid for further studies to determine mechanisms of implantation in pigs. Indeed, because TGFs are present at the conceptus-maternal interface in numerous species (1, 3), and because implantation in all species requires interaction of the trophectoderm with the uterine epithelium (4, 37), this model system may also prove useful to provide mechanistic insight into factors controlling implantation in other species, including humans.
The ability of TGF to stimulate expression of the fibronectin gene has been demonstrated in numerous cell systems (38, 39, 40, 41, 42, 43, 44, 45). The TGF-stimulated increase in fibronectin expression by porcine trophectoderm resembles that reported for cultured human cytotrophoblast isolated from term pregnancies, in which treatments with TGF dose-dependently increased production of oncofetal fibronectin, a variant of fibronectin characterized by O-linked glycosylation in the IIICS portion of the molecule, and also increased apparent secretion and deposition of extracellular fibrillar oncofetal fibronectin (46). Although we did not characterize the isoforms or variants of fibronectin produced by porcine trophectoderm in culture, our results indicate both an increase in fibronectin mRNA levels and an increase in production and deposition of fibronectin protein in response to treatment with TGF1. In humans, oncofetal fibronectin is found in the extracellular matrix connecting the trophoblast with the decidua, prompting the hypothesis that this particular matrix protein functions as "trophoblast glue" to mediate implantation and placental attachment to the uterus (47). This potential role of fibronectin may be particularly important in the noninvasive epitheliochorial placentation of pigs. Fibronectin, including the oncofetal fibronectin variant, is present at the peri-implantation interface in pigs (14, 16), and at least some of the possible integrin subunits that may bind fibronectin are up-regulated during this time (14). Indeed, expression of TGFs by both uterine and conceptus cells increases during the peri-implantation phase of pregnancy and TGFs are activated as the conceptus prepares for and begins attachment to the uterine luminal epithelium (6, 7). Doses of TGF used in this study were selected to roughly correspond to the physiologic range of total TGFs present at the peri-implantation conceptus-maternal interface based on estimates obtained from previous bioassay of porcine uterine fluids (7). In addition to expressing TGFs, trophectoderm of porcine conceptuses also expresses type I and type II TGF receptors (5). Thus, results of the present study suggest that one role of active TGF during the peri-implantation phase of pregnancy is to increase production of fibronectin in trophectoderm through autocrine and/or paracrine pathways. Fibronectin then, by binding maternal and conceptus integrins, may provide an element of the complex maternal-conceptus communication required for successful establishment of pregnancy. Additionally, because fibronectin can exist in polymeric, fibrillar forms (48) and potential integrin receptors are present on trophectoderm and uterine luminal epithelial cells (14), this matrix molecule may also serve as a molecular "bridge" between the trophectoderm and integrins on the uterine luminal epithelium to facilitate blastocyst adhesion through mechanisms that may be similar to those proposed for osteopontin (30), which is another matrix molecule present at the maternal-conceptus interface in pigs (15). Finally, it is also possible that TGF may increase production of fibronectin by cells of the endometrium; however, such investigation was beyond the scope of the current study.
The TGFs signal through type I and II receptors; however, type III receptors, often considered "accessory" receptors, can regulate the interaction of TGFs with the signaling receptors (49, 50). We do not know whether type III receptors such as betaglycan and endoglin are present in porcine trophectoderm and influenced the response of the pTr2 cells to TGF or would do so in vivo. Betaglycan is the most abundant TGF-binding protein on the surface of many cells and can function to promote TGF binding to the signaling receptors; however, betaglycan is also found in soluble forms in serum and extracellular matrices (49, 50). Under in vitro conditions, soluble forms of betaglycan bind to all three mammalian isoforms of TGF (50), and soluble betaglycan inhibits TGF-induced increases in fibronectin in mesangial cells, as does LAP (51). Recently, betaglycan was identified in human syncytiotrophoblast, as well as in decidual cells and chorionic connective tissue, in expression patterns that overlapped with those of type I and type II receptors (52). Further study will be required to determine whether betaglycan is also present in trophectoderm and placental tissues of species, such as pigs, that use noninvasive implantation strategies and also to determine whether betaglycan at the conceptus-maternal interface of all species functions to stimulate or inhibit actions of TGFs.
Both fibronectin and LAP support adhesion of pTr2 cells and both contain an RGD amino acid sequence recognized by many integrins; therefore, we hypothesize that the pTr2 adhesion is accomplished via integrin binding to those amino acids. Fibronectin, in particular, has an RGD site in the III10 region of the molecule, which is the recognition site for 51, the prototype fibronectin receptor, and several v-containing integrin heterodimers (48). Because a mutant fibronectin lacking the RGD site was not readily available, possible integrin dependence for cell adhesion was tested by addition of pentapeptides containing either the RGD or RAD sequence. The competitive inhibition by the RGD-containing peptide supports the hypothesis that integrin-mediated adhesion of pTr2 cells is to the RGD site in fibronectin. However, it does not rule out additional mechanisms of adhesion that could supplement this mechanism of binding in vivo, such as involvement of the PHSRN synergy site, 4 integrin binding to LDV amino acids in the CS1 site (48), or cellular interactions with the heparin binding domain of fibronectin (53). Each dimer of LAP contains two RGD sites. The recombinant mutant LAP used in this study is identical to recombinant LAP except for the substitution of a glutamate residue for an aspartate residue, which eliminates the potential integrin-binding RGD site from the protein (18). The inability of the LAP-RGE mutant to support pTr2 cell was striking, and strongly supports an integrin-mediated mechanism of adhesion. We are unaware of other known mechanisms of cell binding to LAP, and suggest that this mechanism represents a probable means of LAP interaction with conceptus trophectoderm in vivo.
It should be noted that dispersed cells used in the plate adhesion assays are nonpolarized when placed in the matrix-coated wells of the plate, and these assays cannot address whether or not the binding in vitro is directly relevant to the apical adhesion and signaling events of conceptus attachment and implantation. However, the results of the coated bead assays clearly demonstrated specific interactions between apical membranes of trophectoderm and uterine epithelial cells and both fibronectin and LAP. Further, the aggregation of talin at the bead-apical membrane interface supports the transmission of a functional signal characteristic of an integrin-mediated response (29). Indeed, the association of the cytoskeletal protein talin with integrin cytoplasmic domains is a critical step during integrin activation, and regulation of this step may be a final common element in the signaling pathways that control integrin activation (54). Although similar cytoskeletal aggregates at implantation sites in pigs have not been reported, this type of focal adhesion response is present in placentation sites in sheep in association with integrin subunits v and 5 and the extracellular matrix protein osteopontin (55). Whereas the precise down-stream signals and resultant functions of fibronectin-integrin and LAP-integrin interactions at the porcine conceptus-maternal interface remain to be determined, evidence suggests that both extracellular matrix molecules bind to integrin receptors on apical aspects on trophectoderm and/or uterine luminal epithelium and transduce signals that contribute to conceptus-maternal communication and the implantation cascade. Because of the marked remodeling and morphologic changes of the conceptus that accompany implantation in pigs (8, 9, 56), potential functions of fibronectin and LAP in controlling trophectoderm cell shape, differentiation, migration, and conceptus elongation may be equally as important as are roles for these molecules in directly mediating conceptus adhesion to the uterine luminal epithelium.
Results of the present study raise the possibility that integrins on the basolateral aspects of the trophectoderm cells are interacting with fibronectin or LAP. Fibronectin is present in basal laminae and at the interface of trophectoderm and endoderm in the porcine blastocyst (57). Similarly, TGFs, and thus LAP, are present in mesodermal cells underlying trophectoderm in porcine blastocysts and are also present in the stroma underlying uterine luminal epithelium (5, 7). It is reasonable to assume that interactions between basally located integrins and these matrix proteins may affect cellular differentiation and function of both conceptus and maternal cells.
Although initially recognized as a portion of the TGF molecule that confers latency to the secreted growth factor, LAP was first described as an "atypical" integrin ligand for v1 and v5 (18) and is now known to bind to several other v-containing integrins, such as v3, v6, and v8, as well as to integrin 81 (19, 21, 22, 24). Consequences of LAP interactions with integrins vary depending upon the specific integrin heterodimer to which LAP binds. Although the physiologic relevance of these interactions is not yet understood, LAP-integrin interactions in vitro lead to phosphorylation of signaling and cytoskeletal molecules (19, 21) and changes in cell behavior that include increased cell spreading, migration (18, 21, 23), and proliferation (21). Because both complexed LAP and free LAP can interact with integrins (18, 19, 22), multiple possibilities for LAP-mediated signaling are present at the conceptus maternal interface. In fact, binding of the latent TGF complex to integrins may provide a mechanism for TGF activation, either with or without contributions from metalloproteases (19, 20, 22). These interactions have not been examined at the conceptus-maternal interface in vivo; however, they present attractive possibilities for localized conceptus-maternal communications necessary for the implantation process.
The roles of TGFs in early pregnancy are not totally understood, but it is likely that this family of growth factors has multiple, well-integrated functions that contribute to successful establishment of pregnancy. Evidence suggests that effects of TGF at the maternal-conceptus interface may extend beyond the transcriptional and translational responses elicited by receptor-mediated active TGFs to include integrin-mediated conceptus-maternal communication transmitted through LAP, either before, concurrent with, or after TGF activation.
Acknowledgments
We thank Dr. J. S. Munger for providing recombinant baculovirus, Dr. J. N. MacLeod for providing the fibronectin plasmid, Dr. A. Whyte for providing SN1/38 antibody, and Dr. Frankie White for assistance with statistical analyses.
Footnotes
This work was supported by the United States Department of Agriculture National Research Initiative Competitive Grants Programs Grant 2000-02290 (to F.W.B and L.A.J.) and National Institutes of Health Grant P30ES09106.
Abbreviations: DPBS, Dulbecco’s PBS; LAP, latency-associated peptide; TBS, Tris-buffered saline.
References
Rider V, Piva M 1998 Role of growth factors of uterine and fetal-placental origin during pregnancy. In: Bazer FW, ed. Endocrinology of pregnancy. Totowa, NJ: Humana Press; 83–124
Godkin JD, Doré JJE 1998 Transforming growth factor and the endometrium. Rev Reprod 3:1–6
Jaeger LA, Johnson GA, Ka H, Garlow JG, Burghardt RC, Spencer TE, Bazer FW 2001 Functional analysis of autocrine and paracrine signalling at the uterine-conceptus interface in pigs. Reprod Suppl 58:191–207
Burghardt RC, Johnson GA, Jaeger LA, Ka H, Garlow JG, Spencer TE, Bazer FW 2002 Integrins and extracellular matrix proteins at the maternal-fetal interface in domestic animals. Cells Tissues Organs 172:202–217
Gupta A, Bazer FW, Jaeger LA 1996 Differential expression of TGFs (TGF1, TGF2, TGF3) and their receptors (type I and type II) in peri-implantation porcine conceptuses. Biol Reprod 55:796–802
Gupta A, Ing NH, Bazer FW, Bustamante LS, Jaeger LA 1998 Transforming growth factors (TGFs) at the porcine conceptus-maternal interface. Part I: expression of TGF1, TGF2, and TGF3 messenger ribonucleic acids. Biol Reprod 59:905–910
Gupta A, Dekaney CM, Bazer FW Madrigal MM, Jaeger LA 1998 Transforming growth factors (TGFs) at the porcine conceptus-maternal interface. Part II: uterine TGF bioactivity and expression of immunoreactive TGFs (TGF1, TGF2, and TGF 3) and their receptors (type I and type II). Biol Reprod 59:911–917
Bazer FW, Geisert RD, Zavy MT 1987 Fertilization, cleavage, and implantation. In: Hafez ESE, ed. Reproduction in farm animals. 5th ed. Philadelphia: Lea and Febiger; 210–228
Geisert RD, Brookbank JW, Roberts RM, Bazer FW 1982 Establishment of pregnancy in the pig. II. Cellular remodeling of the porcine blastocyst during elongation on day 12 of pregnancy. Biol Reprod 27:941–955
Munger JS, Harpel JG, Gleizes PE, Mazzieri R, Nunes I, Rifkin DB 1997 Latent transforming growth factor -structural features and mechanisms of activation. Kidney Intl 51:1376–1382
Rifkin DB, Mazzieris R, Munger JS, Noguera I, Sung J 1999 Proteolytic control of growth factor availability. APMIS 107:80–85
Justin P. Annes JP, Munger JS, Rifkin DB 2003 Making sense of latent TGF activation. J Cell Sci 116:217–224
Hyytiinen M, Penttinen C, Keski-Oja J 2004 Latent TGF- binding proteins: extracellular matrix association and roles in TGF- activation. Crit Rev Clin Lab Sci 41:233–264
Bowen JA, Bazer FW, Burghardt RC 1996 Spatial and temporal analyses of integrin and Muc-1 expression in porcine uterine epithelium and trophectoderm in vivo. Biol Reprod 55:1098–1106
Garlow, JE, Ka H, Johnson GA, Burghardt RC, Jaeger LA, Bazer FW 2002 Analysis of osteopontin at the maternal-placental interface in pigs. Biol Reprod 66:718–725
Tuo W, Bazer FW 1996 Expression of oncofetal fibronectin in porcine conceptuses and uterus throughout gestation. Reprod Fertil Dev 8:1207–1213
Derynck R, Rhee L 1987 Sequence of the porcine transforming growth factor- precursor. Nucleic Acids Res 15:3187
Munger JS, Harpel JG, Giancotti FG, Rifkin DB 1998 Interactions between growth factors and integrins: latent forms of transforming growth factor- are ligands for integrin v1. Mol Biol Cell 9:2627–2638
Munger JS, Huang X, Kawakatsu H, Griffiths MJD, Dalton SL, Wu J, Pittet J-F, Kaminski N, Garat C, Matthay MA, Rifkin DB, Sheppard D 1999 The integrin v6 binds and activates latent TGF1: a mechanism for regulating pulmonary inflammation and fibrosis. Cell 96:319–328
Annes JP, Rifkin DB, Munger JS 2002 The integrin v6 binds and activates latent TGF3. FEBS Lett 511:65–68
Lu M, Munger JS, Steadele M, Busald C, Tellier M, Schnapp LM 2002 Integrin 81 mediates adhesion to LAP-TGF1. J Cell Sci 115:4641–4648
Mu D, Cambier S, Fjellbirkeland L, Baron JL, Munger JS, Kawakatsu H, Sheppard D, Broaddus VC, Nishimura SL 2002 The integrin v8 mediates epithelial homeostasis through MT1-MMP-dependent activation of TGF-1. J Cell Biol 157:493–507
Thomas GJ, Hart IR, Speight PM, Marshall JF 2002 Binding of TGF-1 latency-associated peptide (LAP) to v6 integrin modulates behaviour of squamous carcinoma cells. Br J Cancer 87:859–867
Ludbrook SB, Barry ST, Delves CJ, Horgan CMT 2003 The integrin v3 is a receptor for the latency associated peptides of transforming growth factors 1 and 3. Biochem J 369:311–318
Ka H, Jaeger LA, Johnson GA, Spencer TE, Bazer FW 2001 Keratinocyte growth factor is up-regulated by estrogen in the porcine uterine endometrium and functions in trophectoderm cell proliferation and differentiation. Endocrinology 142:2303–2310
Whyte A, Bacon M, Ellis S 1984 A monoclonal antibody to an antigen present on the microvillus membrane of the trophectoderm of the preimplantation blastocyst of the pig. J Reprod Fertil 71:599–605
Ing NH, Spencer TE, Bazer FW 1996 Estrogen enhances endometrial estrogen receptor gene expression by a posttranscriptional mechanism in the ovariectomized ewe. Biol Reprod 54:591–599
Bowen JA, Bazer FW, Burghardt RC 1997 Spatial and temporal analysis of integrin and Muc-1 expression in porcine uterine epithelium and trophectoderm in vitro. Biol Reprod 56:409–415
Miyamoto S, Teramoto H, Coso OA, Gutkind JS, Burbelo PD, Akiyama SK, Yamada KM 1995 Integrin function: molecular hierarchies of cytoskeletal and signaling molecules. J Cell Biol 131:791–805
Johnson GA, Bazer FW, Jaeger LA, Ka H, Garlow JE, Pfarrer C, Spencer TE, and Burghardt RC 2001 Muc-1, integrin and osteopontin expression during the implantation cascade in sheep. Biol Reprod 65:820–828
Fléchon J-E, Laurie S, Notarianni E 1995 Isolation and characterization of a feeder-dependent, porcine trophectoderm cell line obtained from a 9-day blastocyst. Placenta 16:643–658
Whyte A, Wooding FB, Saunders PT 1986–87 Characteristics of cellular structures derived by tissue culture of pre-implantation blastocysts of the pig. Arch Anat Microsc Morphol Exp 75:91–109
Macleod JN, Burton-Wurster N, Gu DN, Lust G 1996 Fibronectin mRNA splice variant in articular cartilage lacks bases encoding the V, III-15, and I-10 protein segments. J Biol Chem 271:18954–18960
Schneller M, Arap W, Pasqualini 1999 Immunoblotting of integrins. In: Howlett AR, ed. Methods in molecular biology. Vol. 129. Integrin protocols. Totowa, NJ: Humana Press Inc.; 63–78
Haigh T, Chen C-P, Jones CJP, Aplin JD 1999 Studies of mesenchymal cells from 1st trimester human placenta: expression of cytokeratin outside the trophoblast lineage. Placenta 20:615–625
King A, Thomas L, Bischof P 2000 Cell culture models of trophoblast II: trophoblast cell lines—a workshop report. Placenta 21(Suppl A):S113–S119
Carson DD, Bagchi I, Dey SK, Enders AC, Fazleabas AT, Lessey BA, Yoshinaga 2000 Embryo implantation. Develop Biol 223:217–237
Ignotz RA, Massagúe J 1986 Transforming growth factor- stimulates the expression of fibronectin and collagen and their incorporation into the extracellular matrix. J Biol Chem 261:4337–4345
Dean DC, Newby RF, Bourgeois S 1988 Regulation of fibronectin biosynthesis by dexamethasone, transforming growth factor and cAMP in human cell lines. J Cell Biol 106:2159–2170
Roberts C, Birkenmeier J, McQuillan S, Akiyama S, Yamada S, Chen W-T, Yamada K, McDonald J 1988 Transforming growth factor stimulates the expression of fibronectin and of both subunits of the human fibronectin receptor by cultured lung fibroblasts. J Biol Chem 263:4586–4592
Chakrabarty S, Tobon A, Varani J, Brattain MG 1988 Induction of carcinoembryonic antigen secretion and modulation of protein secretion/expression and fibronectin/laminin expression in human colon carcinoma cells by transforming growth factor-. Cancer Res 48:4059–4064
Huang S, Chakrabarty S 1994 Regulation of fibronectin and laminin receptor expression, fibronectin and laminin secretion in human colon cancer cells by transforming growth factor-1. Int J Cancer 57:742–746
Linnala A, Kinnula V, Laitinen LA, Lehto V-P, Virtaben I 1995 Transforming growth factor- regulates the expression of fibronectin and tenascin in BEAS 2B human bronchial epithelial cells. Am J Respir Cell Mol Biol 13:578–585
Hocevar BA, Brown TL, Howe PH 1999 TGF- induces fibronectin syntheis through a c-Jun N-terminal kinase-dependent, Smad4-independent pathway. EMBO J 18:1345–1356
Luo X, Ding L, Chegin N 2004 Gonadotropin-releasing hormone and TGF- activate MAP kinase and differentially regulate fibronectin expression in endometrial epithelial and stromal cells. Am J Physiol Endocrinol Metab 287:E991–E1001
Feinberg RF, Kliman HJ, Wang CL 1994 Transforming growth factor- stimulates trophoblast oncofetal fibronectin synthesis in vitro: implications for trophoblast implantation in vivo. J Clin Endocrinol Metab 78:1241–1248
Feinberg RF, Kliman HJ, Lockwood CJ 1991 Is oncofetal fibronectin a trophoblast glue for human implantation Am J Pathol 138:537–543
Johansson S, Svineng G, Wennerberg K, Armulik A, Lohikangas L 1997 Fibronectin-integrin interactions. Front Biosci 2:d126–d146
De Caestecker M 2004 The transforming growth factor- superfamily of receptors. Cyto Growth Factor Rev 15:1–11
Vilchis-Landeros MM, Montiel JL, Mendoza V, Mendoza-Hernández G, López-Casillas 2001 Recombinant soluble betaglycan is a potent and isoform selective transforming growth factor- neutralizing agent. Biochem J 355:215–222.
Nomura K, Tada H, Kuboki K, Inokuchi T 2002 Transforming growth factor--1 latency-associated peptide and soluble betaglycan prevent a glucose-induced increase in fibronectin production in cultured human mesangial cells. Nephron 91:606–611
Ciarmela P, Florio P, Toti P, Grasso D, Santopietro R, Tosi P, Petraglia F 2003 Expression of betaglycan in pregnant tissues throughout gestation. Eur J Endocrinol 149:433–437
Wierzbicka-Patynowski I, Shcwarzbauer JE 2003 The ins and outs of fibronectin matrix assembly. J Cell Sci 116:3269–3276
Calderwood DA 2004 Talin controls integrin activation. Biochem Soc Trans 32(Pt 3):434–437
Johnson GA, Burghardt RC, Bazer FW, Spencer TE 2003 Osteopontin: roles in implantation and placentation. Biol Reprod 69:1458–1471
Mattson BA, Overstrom EW, Albertini DF 1990 Transitions in trophectoderm cellular shape and cytoskeletal organization in the elongating pig blastocyst. Biol Reprod 42:195–205
Richouxs V, Darribère T, Boucaut J-C, Flèchon J-E, Thiery J-P 1989 Distribution of fibronectins and laminin in the early pig embryo. Anat Record 223:72–81(Laurie A. Jaeger, Aaron K)
Abstract
In pigs, expression and amounts of biologically active TGFs at the conceptus-maternal interface increase significantly as conceptuses elongate and begin the implantation process. Before their activation, secreted TGFs are noncovalently associated with their respective, isoform-specific latency-associated peptides (LAPs), which contain the Arg-Gly-Asp (RGD) amino acid sequence that serves as a ligand for numerous integrins. Objectives of this study were to determine whether TGF1 increases production of fibronectin by porcine trophectoderm, whether porcine trophectoderm adheres specifically to fibronectin and LAP, and whether functional interactions between porcine trophectoderm and the two TGF-associated proteins, fibronectin and LAP, are integrin mediated. Porcine trophectoderm cells (pTr2) were cultured in presence of TGF1, LAP, or pan-neutralizing anti-TGF antibody; TGF specifically increased (P < 0.05) fibronectin mRNA levels, as determined by Northern and slot blot analyses. Immunofluorescence microscopy demonstrated a TGF-induced increase in fibronectin in pTr2 cells. In dispersed cell adhesion assays, adhesion of pTr2 cells to fibronectin was inhibited by an RGD-containing peptide (P < 0.05) and pTr2 cells attached to recombinant LAP but not to an LAP mutant, which contained an RGE sequence rather than the RGD site (P < 0.05). Fibronectin- and LAP-coated microbeads induced integrin activation at apical surfaces of both trophectoderm and uterine luminal epithelial cells, as indicated by aggregation and transmembrane accumulation of talin detected with immunofluorescence microscopy. Cell surface biotinylation and immunoprecipitation revealed integrin subunits v and 1 on apical membranes of pTr2 cells. These results suggest multiple effects of TGF at the porcine conceptus-maternal interface, including integrin-mediated conceptus-maternal communication through LAP.
Introduction
THE TGFs AND THEIR RECEPTORS are present at the conceptus-maternal interface in numerous species (1, 2, 3), but their functional roles during early pregnancy are unclear. TGFs are widely known as multifunctional growth factors. They are of particular interest during early pregnancy because of their effects on proteases and protease inhibitors, which are involved in implantation, as well as their potential to alter expression of integrins and extracellular matrix molecules that appear to play critical roles in the implantation process (3, 4).
The three mammalian isoforms of TGF (TGF1, 2, and 3), along with TGF receptors type I and II, are present on both sides of the conceptus-maternal interface in pigs during the peri-implantation phase of pregnancy (5, 6, 7). Accordingly, the TGFs may exert both autocrine and paracrine actions on the conceptus and/or endometrium during early pregnancy. Expression of all TGF isoforms increases as conceptuses undergo their remarkable morphologic transformation from spherical to elongated filamentous forms and attach to the endometrial luminal epithelium (6, 7, 8, 9). Because TGFs are secreted in latent forms and only activated forms of TGFs can bind to and signal through type I and II receptors, activation is an essential step in controlling the availability and receptor-mediated actions of these growth factors (10, 11, 12, 13). Interestingly, in addition to the increase in absolute amounts of total TGFs, amounts of biologically active TGFs at the conceptus-maternal interface in pigs also increase significantly as conceptuses elongate and begin the attachment and implantation process, implying a role for TGFs during this critical phase of pregnancy (6, 7). The concurrent increase in expression of some integrin subunits on the uterine luminal epithelium and the expression of osteopontin, fibronectin, and other extracellular matrix molecules at the conceptus-maternal interface may be in response to TGFs (14, 15, 16). This suggests that TGFs influence integrin-mediated signaling and conceptus attachment in pigs, which have central implantation and develop a noninvasive epitheliochorial placenta.
The secreted TGFs are noncovalently associated with their respective isoform-specific latency-associated peptides (LAPs), which form linkages with latent-TGF- binding proteins; the resultant complex may further link to and become immobilized in the extracellular matrix (10, 11, 12, 13). The LAPs of TGFs 1 and 3 contain the Arg-Gly-Asp (RGD) amino acid sequence that serves as a ligand for numerous integrins. These LAPs, as free molecules or as part of latent TGF complexes, bind to and initiate signals through activation of particular integrins, and some LAP-integrin interactions have been shown to initiate functional cell signaling for cell migration, spreading, and proliferation (17, 18, 19, 20, 21, 22, 23, 24). Such cellular processes are critical to implantation and, in particular, differentiation and attachment of trophectoderm to uterine luminal epithelium, but this potential novel function for TGF during implantation has not been explored.
Our hypothesis was that TGFs function through multiple cell signaling pathways to stimulate conceptus elongation and attachment, as well as effect changes in conceptus and uterine gene expression critical to conceptus survival and successful establishment of pregnancy. We focused on TGF1 and its LAP because this isoform is abundant at the porcine conceptus-maternal interface and is the best characterized of the mammalian TGFs. The specific objectives of this study were to determine whether: 1) TGF1 increases production of fibronectin by porcine trophectoderm, 2) porcine trophectoderm adheres specifically to fibronectin and LAP, and 3) whether functional interactions between porcine trophectoderm and the two TGF-associated proteins, fibronectin and LAP, are integrin mediated.
Materials and Methods
Porcine trophectoderm cell culture
The spontaneously immortalized porcine trophectoderm cell line pTr2 was cultured as previously described (25) in phenol red-free DMEM-F12 containing 5% charcoal-stripped fetal bovine serum, 2 mM glutamine, and 0.1 U/ml bovine insulin. Cells were maintained and all culture experiments were conducted in a 5% CO2 humidified environment.
For determination of the effect of TGF on fibronectin mRNA levels, cells were plated at approximately 15,000 cells/cm2 and, after two days of culture, were changed to serum-free DMEM-F12 containing 0.1% BSA (DMEM-F12/BSA) and 2 mM glutamine. After 24 h of serum deprivation, cultures were treated with recombinant human TGF1 (Life Technologies, Inc., Gaithersburg, MD; 0, 0.1, 1.0, or 10.0 ng/ml) for 24 h in DMEM-F12/BSA (n = 3 separate experiments). Subsequent experiments were similarly conducted in which cultures were treated for 24 h in DMEM-F12/BSA (n = 3 separate experiments) with recombinant human TGF (0 or 1.0ng/ml) and either TGF pan-neutralizing antibody (0.15 μg/ml; AB-100-NA; R&D Systems, Minneapolis, MN), control rabbit Ig (IgG; 0.15 μg/ml), or recombinant simian TGF LAP (0 or 2.5 ng/ml) that was produced by Sf9 insect cells (Invitrogen, Carlsbad, CA) infected with a recombinant baculovirus (generously provided by Dr. J. S. Munger, New York University School of Medicine, New York, NY) and purified as described by Munger et al. (18).
For determination of the effect of TGF on fibronectin protein expression, pTr2 cells were similarly plated on Lab-Tek chamber slides (Nalge Nunc International, Naperville, IL), serum-deprived, and treated with TGF (0, 1.0, or 10 ng/ml) for 24 h.
Immunofluorescence
Expression of TGFs by porcine trophectoderm in vivo has been demonstrated (5, 6). To determine whether pTr2 cells maintained expression of TGF1 in culture, cells were seeded on four-well Lab-Tek glass chamber slides, cultured for 48 h, fixed with ice-cold methanol, and subjected to immunofluorescence analysis as previously reported (14, 15), and with anti-TGF1 (10 μg/ml; AB-100-NA; R&D Systems) or control rabbit IgG used at the same concentration as the primary antibody.
To further characterize the pTr2 cells, immunofluorescence microscopy analysis on ice-cold methanol-fixed cells was conducted using antibody directed against cytokeratin 7 (2 μg/ml; clone OV-TL 12/30; InnoGenex, San Ramon, CA) and, on paraformaldehyde-fixed cells, using monoclonal antibody SN1/38 (undiluted tissue culture supernatant, approximately 10 μg/ml, kindly provided by A. Whyte, The Babraham Institute, Cambridge, UK), which identifies a porcine trophectoderm-specific antigen (26), or with control mouse IgGs. Additionally, expression and distribution of fibronectin in TGF-treated cells was evaluated via immunofluorescence using antibody against fibronectin (2 μg/ml; Neomarkers Ab-2, clone HFN36.3; Labvision Corp., Fremont, CA).
Slides were covered with commercially obtained mounting medium (Prolong Gold antifade solution with or without the nuclear counterstain 4',6-diamidino-2-phenylindole; Molecular Probes, Eugene, OR) and protected by coverslips. Reactions were visualized and imaged using a Zeiss Axioplan 2 microscope (Carl Zeiss, Thornwood, NY) fitted with an Axiocam high-resolution digital camera. Digital images were captured using Axiovision 3.0 software.
For analysis of the effects of 0, 1.0, and 10 ng/ml TGF on fibronectin staining in pTr2 cells, digital fluorescence images of continuous monolayers were recorded using the same instrument settings. Images were adjusted electronically using the same brightness and contrast instrument settings to optimize visualization of extracellular fibronectin at the basal aspect of cells. These fibronectin staining patterns were then converted to binary images before quantification of the area of fibronectin staining in each field of cells. Data were analyzed in two separate experiments on different days. Values for fibronectin staining areas in 1.0 and 10 ng/ml TGF-treated cells were normalized to 0 ng/ml TGF.
Northern and slot blot analysis
Total cellular RNA was isolated from pTr2 cells using TriPure reagent (Roche Applied Science, Indianapolis, IN) as described by the manufacturer. Steady-state expression of fibronectin was determined by Northern and slot blot analyses as previously described (25, 27). Briefly, 10 μg total RNA per lane for Northern analysis and 2 μg total RNA per slot for slot blot analysis were hybridized with 32P-labeled antisense cRNA probes generated against a linearized equine fibronectin cDNA (GenBank accession no. U52107; generously provided by J. N. MacLeod, Cornell University, Ithaca, NY) and 28S ribosomal RNA (27). Blots were exposed to autoradiography film (Kodak BioMax XAR; Eastman Kodak, Rochester, NY), and radioactivity of each slot was quantitated using an InstantImager (Packard, Meridian, CT). Hybridization signals of fibronectin mRNA were normalized to those of 28S RNA to account for loading differences among samples.
Adhesion assays
Ninety-six-well suspension culture plates (Greiner Labortechnik; PGC Scientific, Frederick, MD) were coated with solutions of bovine fibronectin (no. F1141; Sigma, St. Louis, MO), LAP, or mutant LAP, which contained an RGE sequence rather than the integrin-binding RGD site (LAP-RGE; produced by Sf9 insect cells infected with a recombinant baculovirus provided by Dr. J. S. Munger) at concentrations ranging from 0–10 μg by incubating 100 μl per well of each solution overnight at 4 C on an orbital shaker. Wells were then washed three times with Dulbecco’s PBS (DPBS; Life Technologies, Inc.) blocked with 1% BSA in PBS for 1 h at 37 C, and adhesion assays (18) conducted using pTr2 cells from subconfluent cultures. Cells were rinsed three times with calcium-magnesium-free DPBS (Ca-Mg-free DPBS; Life Technologies, Inc.) then incubated 5 min at 37 C, in Ca-Mg-free DPBS containing 0.02% EDTA, followed by a 30-sec to 1-min incubation with 0.12% trypsin. Flasks were tapped gently to dislodge the cells, which were then dispersed in serum-containing medium to stop trypsin activity. Cells were washed twice by centrifugation in DMEM-F12/BSA and then suspended at 300,000 cells/ml in that same medium. In some cases, the peptide inhibitor GRGDSPL (no. G-1269; Sigma) or GRADSPL control peptide (no. G-4144; Sigma) was added at concentrations ranging from 0–100 μg/ml and incubated at 22 C for 15 min. Otherwise, 100 μl of cell suspension was immediately distributed to each well, in triplicate, and plates were incubated 1.5 h at 37 C in a humidified incubator containing 5% CO2. Wells were washed twice with DPBS to remove nonadherent cells, and cells were then fixed with 4% paraformaldehyde in PBS for 30 min, permeabilized with 20% methanol in distilled water for 20 min, and stained with 0.5% crystal violet in 20% methanol for 20 min. Excess stain was removed by repeated washing with distilled water, and plates were allowed to dry overnight. Cell binding was quantified by eluting the crystal violet with 100 mM sodium citrate, pH 4.0, in 50% ethanol and measuring absorbance at 595 nm on an E-Max Precision Plate Reader (Molecular Devices Corp., Menlo Park, CA). Baseline absorbances of control wells that were coated only with BSA and were not incubated with cells were subtracted from each measurement.
Analysis of integrin activation
Polystyrene beads (6.0 μm; Polysciences Inc., Warrington, PA) were prepared and coated with 100 μg/ml of either fibronectin, LAP, or poly-L-lysine (Sigma), which allows nonintegrin-mediated adhesion, as previously reported (15). Porcine trophectoderm cells were seeded on two-well Lab-Tek coverglass chamber slides and cultured for 48 h as described previously, then washed with and changed to DMEM-F12/BSA. Likewise, to further explore the possible roles of LAP and fibronectin at the conceptus-maternal interface, primary cultures of porcine uterine luminal epithelium were established and cultured as previously reported (28) and then washed and changed to DMEM-F12/BSA for subsequent analysis of activation of integrins on the apical membrane of uterine luminal epithelium.
The matrix-coated beads were suspended in DMEM-F12/BSA (106 beads/ml) and incubated with the cells for 1 h. Cultures were then fixed, without prior rinsing, in 4% paraformaldehyde in PBS and subjected to immunofluorescence procedures using an antibody to the cytoskeletal protein talin (antitalin clone 8d4; Sigma) to evaluate integrin activation in response to apical cell membrane contact with each population of matrix-coated beads, as previously described (15, 29). Talin aggregation at focal adhesions requires both integrin ligand occupancy and integrin aggregation (29). Fluorescence imaging of the cultures was performed using a digital fluorescence imaging system consisting of a charge-coupled device camera and image capture software (CELLscan; Scanalytics, Bedford MA) integrated with a Zeiss Axiovert inverted fluorescence microscope. Previously reported methods (15, 30) were used to quantify the percentage of beads with talin immunofluorescence reactions that indicated apical focal adhesion formation in response to integrin activation. Briefly, image collection began 1 μm below the basal surface of the cell and optical slices were collected at 0.5-μm steps up through the apical surface and attached beads. Phase contrast microscopy was first used to locate cells and count the total number of beads adherent to apical aspects of the cells. Fluorescence microscopy was then used to identify transmembrane accumulation of immunoreactive talin at the apical cell membrane-matrix-coated bead interface. For each treatment group, the percentage of ligand-coated beads in contact with cells that exhibited apical "focal adhesions" was obtained by multiplying the ratio of focal adhesion-positive beads to the total number of beads in contact with the cell by 100. For each of the ligands tested, at least four separate experiments on different days were performed with at least three slides per ligand in each experiment. Data are presented as mean percentage of beads ± SD inducing apical focal adhesions.
Analyses of integrin expression
Because the integrin heterodimer v1 is a known receptor for both TGF LAP and fibronectin, cell surface expression of the individual constituent subunits v and 1 was assessed on porcine trophectoderm after cell surface biotinylation and immunopreciplitation. Confluent monolayers of pTr2 cells were washed with Ca-Mg-free DPBS, and cell surface proteins were labeled using a modification of procedures reported by Munger et al. (18) and recommendations by the manufacturer of the biotinylation reagent. Briefly, cells were rinsed with 5 mM EDTA in Ca-Mg-free DPBS at 22 C for 1 min, followed immediately by two brief rinses with DPBS containing calcium and magnesium. Each 75-cm2 flask was then treated with 4 ml of freshly prepared DPBS containing 0.25 mg/ml of membrane impermeable biotin (N-hydroxysulfosuccinimidobiotin; EZ-Link Sulfo-NHS-Biotin; Pierce Biotechnology, Inc., Rockford, IL) or DPBS only, and incubated on a rocking platform, shielded from direct light, at 22 C for 1 h. After removal of the biotin solution, cells were rinsed twice with cold DPBS, once in 0.1 M glycine in cold DPBS, and then twice in cold Tris-buffered saline (TBS) with 1 mM calcium chloride (TBS-Ca) containing 0.02% sodium azide. After complete removal of the last TBS-Ca rinse, cells were lysed with 50 mM octyl--D-thioglucopyranoside (Calbiochem, La Jolla, CA) containing 1 mM each CaCl2, MgCl2, and MnCl2 for 30 min on an orbital shaker at 4 C. Lysed cells were scraped from the flasks and passed through a 25-gauge needle four times. Lysates were centrifuged at 16,000 x g, 20 min, 4 C, and total protein in lysates was determined (Pierce Coomassie Plus Protein Assay Reagent; Pierce Biotechnology, Inc). Portions of the lysates were run on reducing SDS-PAGE (7.5%), blotted to nitrocellulose, and incubated with horseradish peroxidase avidin D (avidin-HRP; Vector Laboratories, Burlingame CA) and chemiluminescent detection reagent (SuperSignal West Pico Substrate; Pierce Biotechnology, Inc.) to confirm biotinylation. Biotinylated lysates were immunoprecipitated with polyclonal antibodies to integrin 1 (no. 1952; Chemicon International, Temecula, CA), integrin v (no. 1930; Chemicon International), or control normal rabbit serum and a protein A-protein G agarose conjugate (Protein A/G plus; Santa Cruz Biotechnology, Santa Cruz CA), using procedures recommended by the supplier of the agarose conjugate. The immunoprecipitates were run on 7% nonreducing SDS-PAGE, 10% nonreducing SDS-PAGE, or 10% reducing SDS-PAGE, blotted, and incubated with avidin-HRP as described previously, and bands were visualized by exposure to autoradiography film (BioMax; Eastman Kodak).
Statistical analyses
Dose response data were subjected to least-squares ANOVA using General Linear Models procedures of the Statistical Analyses System. All effects of treatment were determined using PROC MIXED procedure of SAS (SAS Institute Inc., Cary, NC), and significant treatment effects (P < 0.05) separated using least squares means and the PDIFF procedure of SAS. Statistical significance was set at P < 0.05 for all assays. Data charts were generated using Microsoft Excel (Microsoft Corp., Redmond, WA); for clarity of presentation, data are expressed presented as arithmetic means with SEM.
Results
Characterization of pTr2
The epithelial phenotype of the pTr2 cells was confirmed by positive immunoreactions for cytokeratin 7, and identity as trophectoderm was confirmed by positive immunoreactions for SN1/38 (Fig. 1). Reactivity for SN1/38 was predominantly apical, and presence and intensity of the immunoreaction varied markedly among cells, which is consistent with previous reports for cultured porcine trophectoderm (26, 31, 32) (Fig. 1). TGF localized to perinuclear cytoplasmic and Golgi regions and to margins of the pTr2 cells (Fig. 2).
Effect of TGF on pTr2 fibronectin expression
The specificity of the fibronectin probe was confirmed by Northern blot hybridization and stringent washing, and the same conditions were used for slot blot analyses. One major transcript of approximately 7 kb and one less prevalent, slightly larger transcript were identified on Northern blots, consistent with previously reported splice variants of fibronectin mRNA (33) (Fig. 3A). Treatment of pTr2 cells with TGF for 24 h increased steady-state levels of fibronectin mRNA (Fig. 3B). Effect of treatment was linear (P = 0.001); however, when treatment effects were separated by dose, analysis of initial experiments indicated that 1.0 and 10 ng/ml TGF increased levels of fibronectin compared with vehicle-treated controls (P = 0.003, P = 0.001, respectively) whereas effect of 0.1 ng/ml TGF was not statistically significant (P = 0.143) and effect of 10 ng/ml TGF was not greater than that of 1.0 ng/ml treatment (P = 0.395) (Fig. 3, A and B). Therefore, subsequent experiments to verify specificity of the response used the 1.0-ng/ml dose of TGF (Fig. 3C). Specificity of the increase was confirmed using LAP, which neutralizes the biological activity of TGF, and with a pan-neutralizing TGF antibody. Concurrent treatment with LAP decreased fibronectin levels compared with effects of TGF alone, whereas treatment of control cultures with LAP had no effect on fibronectin mRNA compared with control values (Fig. 3C). In immunoneutralization experiments, addition of pan-neutralizing TGF antibody to TGF-treated cultures significantly decreased fibronectin mRNA levels compared with cultures treated with TGF and control IgG, but did not differ from control cultures treated with vehicle concurrent with control IgG or pan-neutralizing TGF antibody (Fig. 3D).
Immunofluorescence microscopy using antibody specific for fibronectin revealed basal levels of fibronectin in control cultures, which increased in response to TGF (Fig. 4). Immunoreactive fibronectin was localized to basal cell attachment sites associated with focal adhesions, cytoplasmic domains, and extracellular cell surfaces. Fibronectin staining at the basal aspects of pTr2 was increased 6.7-fold by 1.0 ng/ml TGF, whereas 10 ng/ml TGF resulted in a 64.4-fold increase in the area of fibronectin staining at the basal cell surfaces. Further analysis was not conducted because of the visual magnitude of the response and the close agreement of the protein expression results with those of the fibronectin mRNA expression.
Trophectoderm adhesion of fibronectin and LAP
Adhesion of pTr2 cells to fibronectin increased with increasing concentrations of fibronectin substrate up to 5 μg/ml (P = 0.026, quadratic) (Fig. 5A). Experiments to determine mechanism of attachment to fibronectin were conducted using 5 μg/ml. Cell adhesion to fibronectin was inhibited in a linear manner (P = 0.0001) by the peptide containing the integrin-binding amino acid sequence Arg-Gly-Asp (RGD) but not by the control peptide containing an Arg-Ala-Asp (RAD) sequence (Fig. 5B).
All concentrations of LAP tested supported trophectoderm adhesion (quadratic, P = 0.01), but adhesion was not significantly different among the LAP concentrations tested. All tested concentrations of the LAP mutant, in which the RGD sequence was replaced by an Arg-Gly-Glu (RGE) sequence, failed to support adhesion (Fig. 6).
Functional activation of integrin signaling by fibronectin and LAP
The aggregation of the cytoskeletal protein talin at the interface between apical cell membranes and matrix-coated beads, as visualized using combined immunofluorescence and phase contrast microscopy, indicated integrin-mediated signaling in response to both fibronectin and LAP in both epithelial cell types of the conceptus-maternal interface (Table 1 and Fig. 7). Approximately 56% of LAP-coated beads in contact with apical membranes of adherent pTr2 cells and 22% of LAP-coated beads in contact with and porcine uterine luminal epithelial cells elicited talin aggregation at the bead-cell interface. Approximately 39 and 19% of fibronectin-coated beads elicited similar responses in pTr2 and porcine uterine luminal epithelial cells, respectively. Talin aggregation was not detected in either cell type at the interface between apical cell membranes and poly-L-lysine-coated beads (Table 1).
Expression of integrin subunits on apical pTr2 membranes
Immunoprecipitation of biotinylated pTr2 lysates with anti-integrin 1 yielded a labeled band at a relative molecular mass of approximately 110 x 103 on a nonreducing gel (Fig. 8), consistent with reported relative molecular mass of integrin 1 (18), whereas no bands were obtained with the normal rabbit serum control. In a separate experiment, immunoprecipitation with anti-integrin v yielded a major band at the expected relative molecular mass of approximately 150 x 103 and a lesser band at approximately 110 x 103, consistent with coimmunoprecipitation of a integrin subunit along with the v subunit. Use of a reducing gel to aid in identifying the labeled proteins after immunoprecipitation of biotinylated pTr2 lysates with anti-integrin v revealed two labeled bands consistent with the reported relative molecular masses of integrin v, which is posttranslationally cleaved to yield fragments with relative molecular masses of 125 x 103 and 27 x 103 when analyzed under reducing conditions (34). No labeled protein bands were detected in normal rabbit serum controls.
Discussion
This is the first known report that demonstrates a direct functional role for the LAP of TGF in the reproductive system. Results of the present study indicate that trophectoderm adheres to LAP in an integrin-dependent manner and that LAP contact with apical surfaces of both trophectoderm and uterine luminal epithelium elicits a functional response (i.e. rapid induction of focal adhesion formation). Furthermore, results demonstrated that TGF stimulates production of fibronectin by porcine trophectoderm in early stages of pregnancy, that fibronectin support of trophectoderm adhesion is via a mechanism partially dependent on integrins, and that apical contact of both trophectoderm and uterine luminal epithelium results in cytoskeletal rearrangement indicative of integrin-mediated signaling.
We previously confirmed the epithelial phenotype of the pTr2 cells based on immunoreactivity with an antibody that recognizes a broad spectrum of epithelial cytokeratins and verified the trophectoderm origin of the cells by demonstrating that they express interferon mRNA, an interferon unique to porcine trophoblast (25). In this study, we also report in pTr2 cell the presence of cytokeratin 7, a preferred marker for human trophoblast cells (35, 36). Identity of the cells as trophectoderm was further verified by positive immunoreactivity with the SN1/38 antibody. The microvillar localization and the cell-to-cell variations in intensity of the SN1/38 immunofluorescence is virtually identical to that previously reported for porcine trophectoderm cells in culture (31, 32). Although the specific identity of the molecule recognized by SN1/38 monoclonal antibody is not known, it is porcine trophectoderm specific (26). These findings support use of the pTr2 line as an in vitro model for studying the functional biology of pig trophectoderm, which is an early and critical participant in establishing blastocyst interactions with the uterine luminal epithelium.
The TGF antibody used for neutralization and immunofluorescence recognizes all mammalian forms of TGF and does not distinguish between latent and active forms. The pattern of TGF localization suggests that once secreted, TGF is bound to trophectoderm cell membranes. The current experiments did not address whether the extracellular immunoreactive TGF was bound to TGF receptors, integrins, or other binding proteins expressed by the pTr2 cells. These results raise several possibilities for autocrine actions of TGF produced by trophectoderm cells, including direct integrin-mediated effects on their migration and/or adhesion.
Integrin heterodimers containing the v subunits bind LAP (18, 19, 20, 24), and some of the constituent subunits, including v1 and v3, are present on uterine luminal epithelium and conceptus trophectoderm during early pregnancy (14). Our characterization of integrins on the cell surface of porcine trophectoderm cells is not complete, but based on the relative molecular masses of the labeled proteins, current results indicate that integrins v and 1 are expressed on apical membranes of the pTr2 cells and could be responsible, in part, for adhesion and functional response to LAP. Definitive characterization via Western blot analysis of the immunoprecipitates was hampered by the lack of multiple commercially available antibodies that recognize porcine integrins and were generated from different host species and/or would recognize integrin subunits under reducing conditions. Nevertheless, results from both immunoprecipitations are consistent and support expression of integrins v and 1 on the cultured apical trophectoderm cells, and are also consistent with in vivo evidence for integrin expression on porcine conceptuses (14).
Trophectoderm cells are critical to implantation because they must communicate with and adhere to maternal uterine epithelium. The retention of TGF production and responsiveness by the pTr2 cells, in combination with expression of some of the known integrins and matrix proteins present in porcine trophectoderm in vivo suggest that this model system is valid for further studies to determine mechanisms of implantation in pigs. Indeed, because TGFs are present at the conceptus-maternal interface in numerous species (1, 3), and because implantation in all species requires interaction of the trophectoderm with the uterine epithelium (4, 37), this model system may also prove useful to provide mechanistic insight into factors controlling implantation in other species, including humans.
The ability of TGF to stimulate expression of the fibronectin gene has been demonstrated in numerous cell systems (38, 39, 40, 41, 42, 43, 44, 45). The TGF-stimulated increase in fibronectin expression by porcine trophectoderm resembles that reported for cultured human cytotrophoblast isolated from term pregnancies, in which treatments with TGF dose-dependently increased production of oncofetal fibronectin, a variant of fibronectin characterized by O-linked glycosylation in the IIICS portion of the molecule, and also increased apparent secretion and deposition of extracellular fibrillar oncofetal fibronectin (46). Although we did not characterize the isoforms or variants of fibronectin produced by porcine trophectoderm in culture, our results indicate both an increase in fibronectin mRNA levels and an increase in production and deposition of fibronectin protein in response to treatment with TGF1. In humans, oncofetal fibronectin is found in the extracellular matrix connecting the trophoblast with the decidua, prompting the hypothesis that this particular matrix protein functions as "trophoblast glue" to mediate implantation and placental attachment to the uterus (47). This potential role of fibronectin may be particularly important in the noninvasive epitheliochorial placentation of pigs. Fibronectin, including the oncofetal fibronectin variant, is present at the peri-implantation interface in pigs (14, 16), and at least some of the possible integrin subunits that may bind fibronectin are up-regulated during this time (14). Indeed, expression of TGFs by both uterine and conceptus cells increases during the peri-implantation phase of pregnancy and TGFs are activated as the conceptus prepares for and begins attachment to the uterine luminal epithelium (6, 7). Doses of TGF used in this study were selected to roughly correspond to the physiologic range of total TGFs present at the peri-implantation conceptus-maternal interface based on estimates obtained from previous bioassay of porcine uterine fluids (7). In addition to expressing TGFs, trophectoderm of porcine conceptuses also expresses type I and type II TGF receptors (5). Thus, results of the present study suggest that one role of active TGF during the peri-implantation phase of pregnancy is to increase production of fibronectin in trophectoderm through autocrine and/or paracrine pathways. Fibronectin then, by binding maternal and conceptus integrins, may provide an element of the complex maternal-conceptus communication required for successful establishment of pregnancy. Additionally, because fibronectin can exist in polymeric, fibrillar forms (48) and potential integrin receptors are present on trophectoderm and uterine luminal epithelial cells (14), this matrix molecule may also serve as a molecular "bridge" between the trophectoderm and integrins on the uterine luminal epithelium to facilitate blastocyst adhesion through mechanisms that may be similar to those proposed for osteopontin (30), which is another matrix molecule present at the maternal-conceptus interface in pigs (15). Finally, it is also possible that TGF may increase production of fibronectin by cells of the endometrium; however, such investigation was beyond the scope of the current study.
The TGFs signal through type I and II receptors; however, type III receptors, often considered "accessory" receptors, can regulate the interaction of TGFs with the signaling receptors (49, 50). We do not know whether type III receptors such as betaglycan and endoglin are present in porcine trophectoderm and influenced the response of the pTr2 cells to TGF or would do so in vivo. Betaglycan is the most abundant TGF-binding protein on the surface of many cells and can function to promote TGF binding to the signaling receptors; however, betaglycan is also found in soluble forms in serum and extracellular matrices (49, 50). Under in vitro conditions, soluble forms of betaglycan bind to all three mammalian isoforms of TGF (50), and soluble betaglycan inhibits TGF-induced increases in fibronectin in mesangial cells, as does LAP (51). Recently, betaglycan was identified in human syncytiotrophoblast, as well as in decidual cells and chorionic connective tissue, in expression patterns that overlapped with those of type I and type II receptors (52). Further study will be required to determine whether betaglycan is also present in trophectoderm and placental tissues of species, such as pigs, that use noninvasive implantation strategies and also to determine whether betaglycan at the conceptus-maternal interface of all species functions to stimulate or inhibit actions of TGFs.
Both fibronectin and LAP support adhesion of pTr2 cells and both contain an RGD amino acid sequence recognized by many integrins; therefore, we hypothesize that the pTr2 adhesion is accomplished via integrin binding to those amino acids. Fibronectin, in particular, has an RGD site in the III10 region of the molecule, which is the recognition site for 51, the prototype fibronectin receptor, and several v-containing integrin heterodimers (48). Because a mutant fibronectin lacking the RGD site was not readily available, possible integrin dependence for cell adhesion was tested by addition of pentapeptides containing either the RGD or RAD sequence. The competitive inhibition by the RGD-containing peptide supports the hypothesis that integrin-mediated adhesion of pTr2 cells is to the RGD site in fibronectin. However, it does not rule out additional mechanisms of adhesion that could supplement this mechanism of binding in vivo, such as involvement of the PHSRN synergy site, 4 integrin binding to LDV amino acids in the CS1 site (48), or cellular interactions with the heparin binding domain of fibronectin (53). Each dimer of LAP contains two RGD sites. The recombinant mutant LAP used in this study is identical to recombinant LAP except for the substitution of a glutamate residue for an aspartate residue, which eliminates the potential integrin-binding RGD site from the protein (18). The inability of the LAP-RGE mutant to support pTr2 cell was striking, and strongly supports an integrin-mediated mechanism of adhesion. We are unaware of other known mechanisms of cell binding to LAP, and suggest that this mechanism represents a probable means of LAP interaction with conceptus trophectoderm in vivo.
It should be noted that dispersed cells used in the plate adhesion assays are nonpolarized when placed in the matrix-coated wells of the plate, and these assays cannot address whether or not the binding in vitro is directly relevant to the apical adhesion and signaling events of conceptus attachment and implantation. However, the results of the coated bead assays clearly demonstrated specific interactions between apical membranes of trophectoderm and uterine epithelial cells and both fibronectin and LAP. Further, the aggregation of talin at the bead-apical membrane interface supports the transmission of a functional signal characteristic of an integrin-mediated response (29). Indeed, the association of the cytoskeletal protein talin with integrin cytoplasmic domains is a critical step during integrin activation, and regulation of this step may be a final common element in the signaling pathways that control integrin activation (54). Although similar cytoskeletal aggregates at implantation sites in pigs have not been reported, this type of focal adhesion response is present in placentation sites in sheep in association with integrin subunits v and 5 and the extracellular matrix protein osteopontin (55). Whereas the precise down-stream signals and resultant functions of fibronectin-integrin and LAP-integrin interactions at the porcine conceptus-maternal interface remain to be determined, evidence suggests that both extracellular matrix molecules bind to integrin receptors on apical aspects on trophectoderm and/or uterine luminal epithelium and transduce signals that contribute to conceptus-maternal communication and the implantation cascade. Because of the marked remodeling and morphologic changes of the conceptus that accompany implantation in pigs (8, 9, 56), potential functions of fibronectin and LAP in controlling trophectoderm cell shape, differentiation, migration, and conceptus elongation may be equally as important as are roles for these molecules in directly mediating conceptus adhesion to the uterine luminal epithelium.
Results of the present study raise the possibility that integrins on the basolateral aspects of the trophectoderm cells are interacting with fibronectin or LAP. Fibronectin is present in basal laminae and at the interface of trophectoderm and endoderm in the porcine blastocyst (57). Similarly, TGFs, and thus LAP, are present in mesodermal cells underlying trophectoderm in porcine blastocysts and are also present in the stroma underlying uterine luminal epithelium (5, 7). It is reasonable to assume that interactions between basally located integrins and these matrix proteins may affect cellular differentiation and function of both conceptus and maternal cells.
Although initially recognized as a portion of the TGF molecule that confers latency to the secreted growth factor, LAP was first described as an "atypical" integrin ligand for v1 and v5 (18) and is now known to bind to several other v-containing integrins, such as v3, v6, and v8, as well as to integrin 81 (19, 21, 22, 24). Consequences of LAP interactions with integrins vary depending upon the specific integrin heterodimer to which LAP binds. Although the physiologic relevance of these interactions is not yet understood, LAP-integrin interactions in vitro lead to phosphorylation of signaling and cytoskeletal molecules (19, 21) and changes in cell behavior that include increased cell spreading, migration (18, 21, 23), and proliferation (21). Because both complexed LAP and free LAP can interact with integrins (18, 19, 22), multiple possibilities for LAP-mediated signaling are present at the conceptus maternal interface. In fact, binding of the latent TGF complex to integrins may provide a mechanism for TGF activation, either with or without contributions from metalloproteases (19, 20, 22). These interactions have not been examined at the conceptus-maternal interface in vivo; however, they present attractive possibilities for localized conceptus-maternal communications necessary for the implantation process.
The roles of TGFs in early pregnancy are not totally understood, but it is likely that this family of growth factors has multiple, well-integrated functions that contribute to successful establishment of pregnancy. Evidence suggests that effects of TGF at the maternal-conceptus interface may extend beyond the transcriptional and translational responses elicited by receptor-mediated active TGFs to include integrin-mediated conceptus-maternal communication transmitted through LAP, either before, concurrent with, or after TGF activation.
Acknowledgments
We thank Dr. J. S. Munger for providing recombinant baculovirus, Dr. J. N. MacLeod for providing the fibronectin plasmid, Dr. A. Whyte for providing SN1/38 antibody, and Dr. Frankie White for assistance with statistical analyses.
Footnotes
This work was supported by the United States Department of Agriculture National Research Initiative Competitive Grants Programs Grant 2000-02290 (to F.W.B and L.A.J.) and National Institutes of Health Grant P30ES09106.
Abbreviations: DPBS, Dulbecco’s PBS; LAP, latency-associated peptide; TBS, Tris-buffered saline.
References
Rider V, Piva M 1998 Role of growth factors of uterine and fetal-placental origin during pregnancy. In: Bazer FW, ed. Endocrinology of pregnancy. Totowa, NJ: Humana Press; 83–124
Godkin JD, Doré JJE 1998 Transforming growth factor and the endometrium. Rev Reprod 3:1–6
Jaeger LA, Johnson GA, Ka H, Garlow JG, Burghardt RC, Spencer TE, Bazer FW 2001 Functional analysis of autocrine and paracrine signalling at the uterine-conceptus interface in pigs. Reprod Suppl 58:191–207
Burghardt RC, Johnson GA, Jaeger LA, Ka H, Garlow JG, Spencer TE, Bazer FW 2002 Integrins and extracellular matrix proteins at the maternal-fetal interface in domestic animals. Cells Tissues Organs 172:202–217
Gupta A, Bazer FW, Jaeger LA 1996 Differential expression of TGFs (TGF1, TGF2, TGF3) and their receptors (type I and type II) in peri-implantation porcine conceptuses. Biol Reprod 55:796–802
Gupta A, Ing NH, Bazer FW, Bustamante LS, Jaeger LA 1998 Transforming growth factors (TGFs) at the porcine conceptus-maternal interface. Part I: expression of TGF1, TGF2, and TGF3 messenger ribonucleic acids. Biol Reprod 59:905–910
Gupta A, Dekaney CM, Bazer FW Madrigal MM, Jaeger LA 1998 Transforming growth factors (TGFs) at the porcine conceptus-maternal interface. Part II: uterine TGF bioactivity and expression of immunoreactive TGFs (TGF1, TGF2, and TGF 3) and their receptors (type I and type II). Biol Reprod 59:911–917
Bazer FW, Geisert RD, Zavy MT 1987 Fertilization, cleavage, and implantation. In: Hafez ESE, ed. Reproduction in farm animals. 5th ed. Philadelphia: Lea and Febiger; 210–228
Geisert RD, Brookbank JW, Roberts RM, Bazer FW 1982 Establishment of pregnancy in the pig. II. Cellular remodeling of the porcine blastocyst during elongation on day 12 of pregnancy. Biol Reprod 27:941–955
Munger JS, Harpel JG, Gleizes PE, Mazzieri R, Nunes I, Rifkin DB 1997 Latent transforming growth factor -structural features and mechanisms of activation. Kidney Intl 51:1376–1382
Rifkin DB, Mazzieris R, Munger JS, Noguera I, Sung J 1999 Proteolytic control of growth factor availability. APMIS 107:80–85
Justin P. Annes JP, Munger JS, Rifkin DB 2003 Making sense of latent TGF activation. J Cell Sci 116:217–224
Hyytiinen M, Penttinen C, Keski-Oja J 2004 Latent TGF- binding proteins: extracellular matrix association and roles in TGF- activation. Crit Rev Clin Lab Sci 41:233–264
Bowen JA, Bazer FW, Burghardt RC 1996 Spatial and temporal analyses of integrin and Muc-1 expression in porcine uterine epithelium and trophectoderm in vivo. Biol Reprod 55:1098–1106
Garlow, JE, Ka H, Johnson GA, Burghardt RC, Jaeger LA, Bazer FW 2002 Analysis of osteopontin at the maternal-placental interface in pigs. Biol Reprod 66:718–725
Tuo W, Bazer FW 1996 Expression of oncofetal fibronectin in porcine conceptuses and uterus throughout gestation. Reprod Fertil Dev 8:1207–1213
Derynck R, Rhee L 1987 Sequence of the porcine transforming growth factor- precursor. Nucleic Acids Res 15:3187
Munger JS, Harpel JG, Giancotti FG, Rifkin DB 1998 Interactions between growth factors and integrins: latent forms of transforming growth factor- are ligands for integrin v1. Mol Biol Cell 9:2627–2638
Munger JS, Huang X, Kawakatsu H, Griffiths MJD, Dalton SL, Wu J, Pittet J-F, Kaminski N, Garat C, Matthay MA, Rifkin DB, Sheppard D 1999 The integrin v6 binds and activates latent TGF1: a mechanism for regulating pulmonary inflammation and fibrosis. Cell 96:319–328
Annes JP, Rifkin DB, Munger JS 2002 The integrin v6 binds and activates latent TGF3. FEBS Lett 511:65–68
Lu M, Munger JS, Steadele M, Busald C, Tellier M, Schnapp LM 2002 Integrin 81 mediates adhesion to LAP-TGF1. J Cell Sci 115:4641–4648
Mu D, Cambier S, Fjellbirkeland L, Baron JL, Munger JS, Kawakatsu H, Sheppard D, Broaddus VC, Nishimura SL 2002 The integrin v8 mediates epithelial homeostasis through MT1-MMP-dependent activation of TGF-1. J Cell Biol 157:493–507
Thomas GJ, Hart IR, Speight PM, Marshall JF 2002 Binding of TGF-1 latency-associated peptide (LAP) to v6 integrin modulates behaviour of squamous carcinoma cells. Br J Cancer 87:859–867
Ludbrook SB, Barry ST, Delves CJ, Horgan CMT 2003 The integrin v3 is a receptor for the latency associated peptides of transforming growth factors 1 and 3. Biochem J 369:311–318
Ka H, Jaeger LA, Johnson GA, Spencer TE, Bazer FW 2001 Keratinocyte growth factor is up-regulated by estrogen in the porcine uterine endometrium and functions in trophectoderm cell proliferation and differentiation. Endocrinology 142:2303–2310
Whyte A, Bacon M, Ellis S 1984 A monoclonal antibody to an antigen present on the microvillus membrane of the trophectoderm of the preimplantation blastocyst of the pig. J Reprod Fertil 71:599–605
Ing NH, Spencer TE, Bazer FW 1996 Estrogen enhances endometrial estrogen receptor gene expression by a posttranscriptional mechanism in the ovariectomized ewe. Biol Reprod 54:591–599
Bowen JA, Bazer FW, Burghardt RC 1997 Spatial and temporal analysis of integrin and Muc-1 expression in porcine uterine epithelium and trophectoderm in vitro. Biol Reprod 56:409–415
Miyamoto S, Teramoto H, Coso OA, Gutkind JS, Burbelo PD, Akiyama SK, Yamada KM 1995 Integrin function: molecular hierarchies of cytoskeletal and signaling molecules. J Cell Biol 131:791–805
Johnson GA, Bazer FW, Jaeger LA, Ka H, Garlow JE, Pfarrer C, Spencer TE, and Burghardt RC 2001 Muc-1, integrin and osteopontin expression during the implantation cascade in sheep. Biol Reprod 65:820–828
Fléchon J-E, Laurie S, Notarianni E 1995 Isolation and characterization of a feeder-dependent, porcine trophectoderm cell line obtained from a 9-day blastocyst. Placenta 16:643–658
Whyte A, Wooding FB, Saunders PT 1986–87 Characteristics of cellular structures derived by tissue culture of pre-implantation blastocysts of the pig. Arch Anat Microsc Morphol Exp 75:91–109
Macleod JN, Burton-Wurster N, Gu DN, Lust G 1996 Fibronectin mRNA splice variant in articular cartilage lacks bases encoding the V, III-15, and I-10 protein segments. J Biol Chem 271:18954–18960
Schneller M, Arap W, Pasqualini 1999 Immunoblotting of integrins. In: Howlett AR, ed. Methods in molecular biology. Vol. 129. Integrin protocols. Totowa, NJ: Humana Press Inc.; 63–78
Haigh T, Chen C-P, Jones CJP, Aplin JD 1999 Studies of mesenchymal cells from 1st trimester human placenta: expression of cytokeratin outside the trophoblast lineage. Placenta 20:615–625
King A, Thomas L, Bischof P 2000 Cell culture models of trophoblast II: trophoblast cell lines—a workshop report. Placenta 21(Suppl A):S113–S119
Carson DD, Bagchi I, Dey SK, Enders AC, Fazleabas AT, Lessey BA, Yoshinaga 2000 Embryo implantation. Develop Biol 223:217–237
Ignotz RA, Massagúe J 1986 Transforming growth factor- stimulates the expression of fibronectin and collagen and their incorporation into the extracellular matrix. J Biol Chem 261:4337–4345
Dean DC, Newby RF, Bourgeois S 1988 Regulation of fibronectin biosynthesis by dexamethasone, transforming growth factor and cAMP in human cell lines. J Cell Biol 106:2159–2170
Roberts C, Birkenmeier J, McQuillan S, Akiyama S, Yamada S, Chen W-T, Yamada K, McDonald J 1988 Transforming growth factor stimulates the expression of fibronectin and of both subunits of the human fibronectin receptor by cultured lung fibroblasts. J Biol Chem 263:4586–4592
Chakrabarty S, Tobon A, Varani J, Brattain MG 1988 Induction of carcinoembryonic antigen secretion and modulation of protein secretion/expression and fibronectin/laminin expression in human colon carcinoma cells by transforming growth factor-. Cancer Res 48:4059–4064
Huang S, Chakrabarty S 1994 Regulation of fibronectin and laminin receptor expression, fibronectin and laminin secretion in human colon cancer cells by transforming growth factor-1. Int J Cancer 57:742–746
Linnala A, Kinnula V, Laitinen LA, Lehto V-P, Virtaben I 1995 Transforming growth factor- regulates the expression of fibronectin and tenascin in BEAS 2B human bronchial epithelial cells. Am J Respir Cell Mol Biol 13:578–585
Hocevar BA, Brown TL, Howe PH 1999 TGF- induces fibronectin syntheis through a c-Jun N-terminal kinase-dependent, Smad4-independent pathway. EMBO J 18:1345–1356
Luo X, Ding L, Chegin N 2004 Gonadotropin-releasing hormone and TGF- activate MAP kinase and differentially regulate fibronectin expression in endometrial epithelial and stromal cells. Am J Physiol Endocrinol Metab 287:E991–E1001
Feinberg RF, Kliman HJ, Wang CL 1994 Transforming growth factor- stimulates trophoblast oncofetal fibronectin synthesis in vitro: implications for trophoblast implantation in vivo. J Clin Endocrinol Metab 78:1241–1248
Feinberg RF, Kliman HJ, Lockwood CJ 1991 Is oncofetal fibronectin a trophoblast glue for human implantation Am J Pathol 138:537–543
Johansson S, Svineng G, Wennerberg K, Armulik A, Lohikangas L 1997 Fibronectin-integrin interactions. Front Biosci 2:d126–d146
De Caestecker M 2004 The transforming growth factor- superfamily of receptors. Cyto Growth Factor Rev 15:1–11
Vilchis-Landeros MM, Montiel JL, Mendoza V, Mendoza-Hernández G, López-Casillas 2001 Recombinant soluble betaglycan is a potent and isoform selective transforming growth factor- neutralizing agent. Biochem J 355:215–222.
Nomura K, Tada H, Kuboki K, Inokuchi T 2002 Transforming growth factor--1 latency-associated peptide and soluble betaglycan prevent a glucose-induced increase in fibronectin production in cultured human mesangial cells. Nephron 91:606–611
Ciarmela P, Florio P, Toti P, Grasso D, Santopietro R, Tosi P, Petraglia F 2003 Expression of betaglycan in pregnant tissues throughout gestation. Eur J Endocrinol 149:433–437
Wierzbicka-Patynowski I, Shcwarzbauer JE 2003 The ins and outs of fibronectin matrix assembly. J Cell Sci 116:3269–3276
Calderwood DA 2004 Talin controls integrin activation. Biochem Soc Trans 32(Pt 3):434–437
Johnson GA, Burghardt RC, Bazer FW, Spencer TE 2003 Osteopontin: roles in implantation and placentation. Biol Reprod 69:1458–1471
Mattson BA, Overstrom EW, Albertini DF 1990 Transitions in trophectoderm cellular shape and cytoskeletal organization in the elongating pig blastocyst. Biol Reprod 42:195–205
Richouxs V, Darribère T, Boucaut J-C, Flèchon J-E, Thiery J-P 1989 Distribution of fibronectins and laminin in the early pig embryo. Anat Record 223:72–81(Laurie A. Jaeger, Aaron K)