Inhibition of Cytotrophoblastic (JEG-3) Cell Invasion by Interleukin 12 Involves an Interferon γ-mediated Pathway*

Trophoblast invasion, like tumor invasion, shares common biochemical mechanisms. However, in contrast to tumor invasion of a host tissue, trophoblastic invasion during implantation is strictly regulated, temporospatially. Factors responsible for these important regulatory processes are presently unknown; however, studies indicate that cytokines and growth factors represent in the peri-implantation uterine milieu as the possible candidates. In this study we investigated the role of interleukin (IL) 12 in regulating trophoblast invasion and the expression of trophoblast proteases (matrix metalloprotease (MMP)-2, MMP-9, and urokinase-type plasminogen activators) and their inhibitors (tissue inhibitors of metalloprotease (TIMP) 1, TIMP-2, and plasminogen activator inhibitor (PAI)-1) using an in vitro tissue culture system of human choriocarcinoma cell line JEG-3. Our major findings show an anti-invasive role of IL-12, associated with an inhibitory effect on the proteases but with an opposite up-regulating influence on the protease inhibitor, TIMP-1, whereas TIMP-2 and plasminogen activator inhibitor 1 remained unaltered. Stimulation of JEG-3 cells with IL-12 also induced interferon (IFN)-γ production, which when neutralized using a monoclonal anti-IFN-γ antibody, F12, abrogates its ability to down-regulate the MMPs. IL-12 also mediates an IFN-γ-dependent up-regulation of E-cadherin, thereby implying that alteration in cell-cell adhesion besides regulating the proteases and the inhibitors possibly contributes to the observed anti-invasive role of this cytokine. TIMP-1, although stimulated by IL-12, was found to be unaltered by antibody F12, thereby implying a possibility of an IL-12-dependent-IFN-γ independent regulation. These findings thereby suggest an important role of IL-12 in modulation of trophoblast proteases and their inhibitors besides regulating cell-cell interactions and invasion during implantation, with far reaching possibilities for understanding the mechanism(s) and regulations of invasion and metastasis.

Implantation is an excellent example of successive interactions between two dissimilar tissues, the receptive uterus and the developing blastocyst, each genetically distinct from the other. This two-way interaction is largely mediated by the intimate contact between the uterine luminal epithelium and the outermost polarized epithelial cells of the blastocyst, the trophoblast. Trophoblast cells are instrumental during implantation and placentation both in terms of molecular recognition and cross-talk at the feto-maternal interface as well as a repository of variety of cytokines and growth factors that influence both the conceptus and the maternal physiology in an autocrine, paracrine, or juxtacrine manner (1,2). This represents a highly complex but coordinated process involving the participation of different trophoblast cell populations with specific functions (3,4). The interaction begins as the blastocyst enters the uterine lumen and becomes apposed to the uterine epithelium with its trophoblast cells penetrating deeper into the endometrium. The basic hurdle encountered by the invading trophoblasts is the maternal extracellular matrix through which it has to migrate. This hurdle is overcome by the spectacular process of trophoblast invasion, similar to that of cancer invasion (5,6) except that it is highly regulated and, thus, is often called pseudo-malignant. A number of metalloproteases and serine proteases has been associated with this invasive process, the proteolytic activity of which eventually contributes to the digestion of the maternal extracellular matrix by the trophoblast cells (7)(8)(9). Matrix metalloproteases (MMP) 1 are an important group of zinc-containing enzymes responsible for degradation of the extracellular matrix components such as collagen and proteoglycans and are believed to be crucial during the process of tumor invasion, metastasis, and angiogenesis (10 -12). The tissue inhibitors of metalloproteases (TIMPs), TIMP-1 for MMP-9 and TIMP-2 for MMP-2, act as natural inhibitors, counteracting these metalloproteases (13), whereas the plasminogen activator inhibitors, PAI-1 and PAI-2, inhibit uPA (14). Trophoblast invasion is spectacular, as it is tempo spatially regulated and, thus, clearly requires the participation of both the proteases as well their inhibitors (15) in an appropriate stoichiometric amount, as too little or an excess of invasion can both result into pregnancy abnormalities like preeclampsia or choriocarcinoma.
As mentioned before, successful implantation is the final outcome of an innumerable sequel of events orchestrated by the synchronized co-ordination between the fetal and maternal components involving the participation of a large number of cytokines, growth factors, hormones, lipid mediators, and local acting substances (16).
IL-12 is a heterodimeric cytokine produced by antigen-pre- 1 The abbreviations used are: MMP, matrix metalloprotease; TIMP, tissue inhibitor of MMP; uPA, urokinase-type plasminogen activator; IL, interleukin; IFN-␥, interferon ␥; rhIFN-␥, recombinant human IFN-␥; SEM, scanning electron microscopy; Ab, antibody; RT, reverse transcription; ELISA, enzyme-linked immunosorbent assay; CREB, cAMP-response element-binding protein; STAT, signal transducers and activators of transcription; PAI, plasminogen activator inhibitor. senting cells, phagocytes, and granulocytes (17)(18) and is recently immunolocalized in mice feto-placental unit (19,20). Despite few recognized direct effects on T-lymphocytes and NK cells, where it acts as a growth factor and an enhancer of cytotoxicity and activator of other cytokines, its in vivo action is possibly mediated through (interferon ␥) IFN-␥ (21). IL-12 has been tested in clinical trials in curing neoplasias and has been found to improve the survival of mice bearing a great variety of tumors (22)(23)(24). Based upon these observations, we aimed to investigate the role of IL-12 on trophoblast invasion in terms of regulation of proteases (MMPs and uPA), protease inhibitors (TIMPs and PAI-1), and cell-matrix interactions using an in vitro tissue culture system of JEG-3 choriocarcinoma cells due to its similarity in the expression profile of these proteases and their inhibitors to that of the first trimester extra villous trophoblast cells (7). Use of JEG-3 cells in this study is also justified by the fact that several investigators have similarly used the same model system to study the various aspects of trophoblast invasion (25)(26)(27)(28)(29).
Trophoblast invasion, unlike tumor invasion, is a highly controlled and coordinated process, deciphering the molecular basis of which is of paramount importance both in view of understanding this magnificent process as well as in delineating the biology of cancer invasion and metastasis.

EXPERIMENTAL PROCEDURES
Materials-Recombinant human IL-12 and IFN-␥ were purchased from R&D Systems Inc, Minneapolis, MN. Mouse monoclonal antihuman IFN-␥ neutralizing antibody, F12, was procured from Abcam Ltd., Cambridge Science, Cambridge, UK. Polyclonal goat anti-human MMP-2, MMP-9, uPA, and PAI-1 and polyclonal rabbit antihuman TIMP-1 and -2 and E-cadherin were all purchased from Santa Cruz Biotechnology, CA. A secondary antibody system used for immunocytochemistry, the Universal LSAB kit, was procured from DAKO Laboratories, Glostrup, Denmark; 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide cell proliferation assay kit was from Roche Applied Science. Dual luciferase reporter assay system, TransFast transfection reagent (Lipofectamine), and pGL-2 basic vector were all obtained from Promega, Madison, WI. Matrigel invasion chamber was from Biocoat, BD Biosciences. TRIzol reagent for RNA extraction was from Invitrogen. PCR primers were obtained from Microsynth, Balgach, Switzerland with the remaining PCR chemicals from Promega. The gel purification kit was purchased from Macherey-Nagel, Duren, Germany. Tissue culture materials (fetal calf serum and Ham's F-12 media) were obtained from Invitrogen, and sterile plastic ware was from Corning-Costar Corp., Cambridge, MA. All other chemicals were from Sigma and were all of analytical grade.
Cell Line-Human choriocarcinoma cell line, JEG-3, was procured from American Type Culture Collection (Parkville, MA). These were expanded in tissue culture flasks in F-12 media containing 10% fetal calf serum and the antibiotic/antimycotic solutions and maintained at 37°C in a sterile humid atmosphere under 5% CO 2 and 95% O 2 .
Cytokine Challenge-Recombinant human IL-12 was used at concentrations of 200, 500, and 1000 pg/ml for 12 h. Recombinant human IFN-␥ (rhIFN-␥) was used at a concentration of 500 units/ml. 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay confirmed that the two cytokines used at the above-mentioned concentrations did not compromise with the viability of the cell (data not shown).
Expression Vectors and Reporters-A 1716-bp (Ϫ1659 to ϩ57 bp) promoter fragment of human MMP-2 (a gift from Sun Yi, Department of Molecular Biology, Parke-Davis Pharmaceutical Research, Ann Arbor, Michigan, MI) was excised from the pGEM-3Z vector by cleaving with Hind III and Kpn1 restriction enzymes. This was then gel-purified (Nucleo Spin extract) and cloned upstream to the pGL2 basic promoter less vector. A human MMP-9 promoter region of 689 bp (Ϫ670 to ϩ19) and a human TIMP-1 promoter region (1718 bp) was also PCR-amplified and inserted upstream to the pGL2 basic vector.
Transient Transfection and Reporter Gene Assay-JEG-3 cells plated in 12-well Transwells were incubated at 37°C till 60 -70% confluency. 2 g each of the reporter vectors, MMP-2 promoter reporter construct (pMMP2-Luc), MMP-9 promoter reporter construct (pMMP9-Luc), or TIMP-1 promoter reporter construct (pTIMP1-Luc) were transiently transfected using Lipofectamine reagent according to the manufacturer's instruction. Briefly, cells were incubated in serum-and antibiotic-free F-12 media with or without IL-12 or IFN-␥ for 48 h and harvested, and a luciferase assay performed using a Sirius Luminometer V2.2 (Sirius Berthold detection system) with a measurement time of 10 s. Renilla luciferase activity was measured after quenching the light with Stop and Glow reagent. Luciferase values are the mean Ϯ S.E. of four replicate wells represented by three independent experiments. The results were expressed as fold luciferase activity (relative light units) (0.5 g of pRLSV40 control vector was also co transfected for normalization of the transfection efficiency).
RNA Isolation and Semiquantitative RT-PCR-Cells from culture wells were harvested in TRIzol reagent following the manufacturer's instructions. Complementary DNA (cDNA) was prepared when 5 g of the total RNA was reverse-transcribed as described earlier (7). 2 l of cDNA prepared were then processed for PCR using deoxynucleotide triphosphates, 10ϫ PCR buffer containing 1.5 mM MgCl 2, Taq DNA polymerase (5 units/l) and molecule-specific primers. IFN-␥ (forward primer, 5Ј-TGA CCA GAG CAT CCA AAG A-3Ј; reverse primer, 5Ј-CTG ACT CCT TGT TTC-GCT TCC-3Ј; amplicon size, 214 bp; annealing temperature, 60°C). The primer sequence and PCR conditions for MMP-2, MMP-9, uPA, TIMP-1, TIMP-2, PAI-1 and -2, E-cadherin, and glyceraldehyde-3-phosphate dehydrogenase were followed as described earlier (7,30) in an Eppendorf Master cycler gradient PCR machine (Eppendorf, Hamburg, Germany). The cycle number (n ϭ 30) was adjusted so that all reactions fell within the linear range of amplification. Glyceraldehyde-3-phosphate dehydrogenase is used as an internal standard as it is ubiquitously expressed in most tissues, and its expression was found to be unaltered after the cytokine challenge (data not shown). PCR products (amplicons) were analyzed using the Bio-Rad Quantity One program, and the specificity of bands was verified by sequencing (data not shown). To further establish the RT-PCR results, protein expression and localization studies were undertaken to get insight into the cytokine-mediated translational regulation of the above molecules.
Immunocytochemistry-JEG-3 cells plated on coverslips at an initial density of 3 ϫ 10 5 and grown till 70% confluence were challenged with IL-12 (500 pg/ml) for 12 h and processed for immunolocalization of MMPs and TIMPs after paraformaldehyde fixation (4% for 2 h at 4°C) following the procedure as described earlier (31) with slight modifications. Primary antibodies, goat anti-human 1:200 (MMP-2, MMP-9, uPA, and PAI-1) and rabbit antihuman (TIMP-1 and -2) were used, and non-immune goat or rabbit serum used in the place of primary antibody served as the negative control. Proteins were immunolocalized after incubation with secondary antibody for 1 h at room temperature and were then visualized using 3,3Ј-diaminobenzidine hydrochloride, counterstained with Meyers hematoxylin (Sigma).
Flow Cytometric Assessment-Fluorescence-activated cell sorter was used to quantitate the intracellular level of proteases and their inhibitors after IL-12 (500 pg/ml) challenge. Fluorescein isothiocyanate-conjugated secondary Ab was used to label the proteins under study, the mean fluorescence was recorded using Coulter Counter, and the results were analyzed using Win MDI 2.8 Software.
Zymography-Substrate gel zymography for the gelatinases (gelatin) and uPA (casein) were carried out to demonstrate the presence of gelatinolytic metalloproteases and caseinolytic serine protease uPA from JEG-3-conditioned media (7). To further confirm that the gelatindegrading activities in the conditioned media were indeed due to metalloproteases, gels were incubated in reacting buffer containing metalloprotease inhibitor, EDTA (1-5 mM). For casein zymography, gels were incubated with serine protease inhibitor, phenylmethylsulfonyl fluoride, for 1 h at 37°C, and electrophoresis was performed as described above (data not shown).
MMP and uPA Activity Assay-Activity assay were performed to further quantitate the above zymography findings. MMP activity assay was performed using the MMP substrate (7-methylcoumarin-4 acetyl-Pro-Leu-␤-(2,4-dinitrophenylamino)-Ala-Ala-Arg amide). Briefly, culture spent media with or without IL-12 challenge (500 and 1000 pg/ml) was mixed with the MMP substrate and incubated at 37°C followed by measurement of fluorescence intensity at 325-nm excitation and 395-nm emission.
Urokinase activity was similarly estimated in conditioned media by a direct chromogenic assay using substrate S-2444 (Chromogenix, MoIndal, Sweden) in a reaction that does not involve plasmin formation from plasminogen (32), following the manufacturer's instruction. Final absorbance was measured at 405 nm. Furthermore, Matrigel invasion assay was performed to correlate the functional implications of MMP and uPA activities in terms of cellular invasion.
Matrigel Invasion Assay; Scanning Electron Microscopy (SEM), Crystal Violet Staining, and Cell Migration Assay-Matrigel-coated porous filters (8-m pore size) in a 12-well format were used as a barrier in a Boyden chamber (33,29) to assess the extent of invasion by JEG-3 cells. Cells at an initial density of 2 ϫ 10 5 (in F-12 media supplemented with 10% fetal bovine serum) were plated into the inserts and maintained in culture in the presence of IL-12 (500 pg/ml) or IFN-␥ (500 units/ml) for 12 h. Membranes were then cut and removed from their insert housings and fixed in Karnovsky's fixative (4% paraformaldehyde and 2% glutaraldehyde for 6 h) along with the removal of non-invasive cells from the upper chamber using a cotton swab. The bottom surface containing the invaded cells were then processed by SEM (LEO 435 VP; Leo Electron Microscopy Ltd., Cambridge, England) or stained with crystal violet for calculating the invasion index.
For crystal violet staining (34), membranes were stained with 0.5% crystal violet (in 20% methanol) for 5 min, washed with phosphatebuffered saline, and visualized under an optical microscope using a drop of immersion oil on the membrane at 100ϫ magnification, or the migratory cells attached to the bottom of the membrane were stained with 0.1% crystal violet in 0.1 M borate, pH 9, and 2% ethanol for 15 min at room temperature. The stained cells were extracted using extraction buffer (Dispase, BD Biosciences), and the number of invaded cells per membrane was determined by absorbance at 550 nm.
Invasion index was expressed as the number of cells invading through the Matrigel "test membrane" relative to the invasion through the "control membrane." Cellular invasion is the final outcome of large number of processes, involving broadly, cell-cell, and cell-matrix interactions, associated with cellular motility. Wound healing and cell-cell adhesion assays were further performed to study the intercellular interactions and cellular motility involved during invasion.
Cell Motility Assay; Wound Healing Assay-Wound healing assay was performed to study the effect of IL-12 (500 pg/ml) on JEG-3 cell motility, following the procedure as mentioned earlier (35) with slight modifications. Cells seeded in equal number into six well plates were grown till 70% confluence, scrapped with a sterile tip to create an artificial wound, and allowed to heal for the next 36 h. The number of individual cells in the wound was quantitated as an average from six independent fields under the microscope (Nikon Microphot FXA) at 100ϫ magnification and Cell-Cell Adhesion Assay-JEG-3 cells in 24-well plates were grown till 100% confluency. Another lot of JEG-3 cells were labeled with [H 3 ]thymidine overnight and trypsinized. Radiolabeled cells were then re-suspended in F-12 media supplemented with 10% fetal bovine serum and added to the unlabeled, attached confluent cells. After 2 h on incubation, non-adherent cells were collected, and plates were washed with phosphate-buffered saline and collected in the same container. Bound cells were then trypsinized and collected in a separate container. Radioactivity was measured by a ␤-counter (Wallac), and the percentage of adherent cells was counted. A Western blot was further performed to elucidate the role of E-cadherin as a possible mediator of intercellular adhesion during invasion and motility.
Western Blot-JEG-3 cells were gently removed from the culture flask by scrapping and washed with 0.01 M phosphate-buffered saline, and proteins were extracted using radioimmune precipitation assay lysis buffer containing protease inhibitor mixture. Western blot was performed with the supernatant (50 g of total protein) obtained by centrifuging the lysate at 10,000 ϫ g for 15 min at 4°C following the procedure as described before (31) using a 1:200-diluted primary antibody for E-cadherin (200 g/ml IgG). Bands were visualized using 3,3Ј-diaminobenzidine hydrochloride as chromogen (enhanced by nickel). Although E-cadherin protein expression was quantitated by Western blot, its actual cellular localization was only confirmed using confocal laser microscopy.
Confocal Laser Microscopy-JEG-3 cells grown on coverslips till 70% confluency were challenged with IL-12 (500 pg/ml) or IFN-␥ (500 units/ ml) and incubated for the next 12 h under culture conditions. After incubation, cells were washed with 0.01 M phosphate-buffered saline and processed for confocal microscopy (Bio-Rad) for immunolocalization of E-cadherin using biotinylated secondary antibody and fluorescein isothiocyanate-conjugated avidin (Calbiochem-Novabiochem). Nuclei were counterstained with propidium iodide.
IFN-␥ Enzyme-linked Immunosorbent Assay (ELISA)-IFN-␥-specific purified coating antibody (anti-human IFN-␥ polyclonal anti-sera (Endogen Inc.) from sheep serum was diluted (0.5-1.5 g/ml) in the coating buffer. 100 l of the diluted antibody was added to each well of 96-well Immulon-II plates, incubated overnight at 25-28°C, and further processed following the manufacturer's instruction. Final color was developed using horseradish peroxide-streptavidin conjugate and tetramethyl benzidine as chromogen. Color was developed for 10 -15 min, the reaction was stopped with H 2 SO 4 , and absorbance was read at 450 nm. All samples were analyzed in duplicate. The standard curve was linear from 15-10,000 pg/ml, and IFN-␥ above this range were diluted. The inter-assay and intra-assay coefficient of variation was Ͻ10%.
IFN-␥ Neutralization Assay-To study the effect of IFN-␥ on invasion, JEG-3 cells stimulated with IL-12 (500 pg/ml) were incubated with human IFN-␥ neutralizing antibody (F12, Ab 8096, mouse monoclonal). ELISA was previously performed to monitor the antibody titration and determine the dose of the neutralizing Ab (data not shown). 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay was carried out to ascertain the safety dose of F12 monoclonal Ab (data not shown). Species-specific random IgGs were also used simultaneously to rule out the possibility of a nonspecific antibody-mediated response.
Statistical Analysis-Each experiment was repeated three times, and statistical comparisons were made either by Student's paired t test or by one-way analysis of variance using Microcal Origin software (Microcal Software Inc.). A value of p Ͻ 0.05 was considered to be significant.

Effect of IL-12 on JEG-3 Cell Proteases and Inhibitors-To
evaluate the role of IL-12 on the protease expression and activity, JEG-3 cells stimulated with the above cytokine showed a dose-dependent reduction in the mRNA expression of both MMP-2 (Fig. 1A) and MMP-9 (Fig. 1B). Expression of the serine protease, uPA, was also found to be inhibited but only at a higher concentration of IL-12 even though it remains fairly constant at a lower dose (200 pg/ml) (Fig. 1C). TIMP-1, the natural in vivo inhibitor for MMP-9, was significantly up-regulated (Fig. 1D) after IL-12 challenge, although TIMP-2 and PAI-1 were both found to be unaltered (Fig. 1, E and F). Furthermore, the IL-12-mediated transcriptional regulation of MMP-2, MMP-9, and uPA were confirmed by luciferase reporter assay (Fig. 1G).

Effect of IL-12 on JEG-3 Cell Invasion (Matrigel Invasion Assay) and
Motility-To understand the actual role of IL-12 on JEG-3 cell invasion besides its effect on the proteases as observed before, Matrigel invasion assay was performed because it provides an artificial basement membrane and is often used to assess the invasion of cells with metastatic properties. JEG-3 cells grown over Matrigel and challenged with IL-12 (500 pg/ml) on observation under SEM showed negligible invasion with associated large intercellular aggregations and cellular clumps, implying an enhanced cell-cell interaction (Fig. 4A), as compared with the unchallenged control, which displayed an extensive array of cell-to-matrix interaction as evident by cell spreading. SEM findings after Matrigel invasion assay was further confirmed by crystal violet staining (Fig. 4B), which quantitatively depicted a dose-dependent reduction in the number of cells invading through the Matrigel after the cytokine challenge as compared with the unchallenged control. Fig.  4C shows the % reduction in the number of cells invading through the Matrigel membrane. Table I summarizes the results of Matrigel invasion assay along with corresponding zymography findings.
Cellular invasion is always associated with enhanced cell motility, a feature commonly observed during malignancy. Cell motility based upon wound healing assay also showed a significant delay by the IL-12-stimulated cells to fill up the wounded area as compared with their unchallenged counterparts (Fig.  4D), represented graphically in Fig. 4E, with ϳ30% reduction at 500 pg/ml and ϳ70% reduction at 1000 pg/ml. The number of cells within the wound also revealed a significant difference between those of IL-12 challenge as compared with the control (Table II).

Regulation of IFN-␥ and E-cadherin by IL-12; Effect of IFN-␥-neutralizing Ab F12 on JEG-3 Cell
Invasion-To investigate the downstream effectors and explore the possible mechanisms of IL-12 during invasion, JEG-3 cells, when challenged with this cytokine, showed a severalfold induction in IFN-␥ production as detected by both RT-PCR (Fig. 5A) and ELISA (Fig. 5B).  Furthermore, an IFN-␥-neutralizing Ab, F12, almost completely neutralized the IL-12-induced production of IFN-␥ when detected by ELISA (Fig. 5B), usage of which on JEG-3 cells abrogated the IL-12-mediated inhibition of the protease activity as seen by zymography for MMPs (Fig. 5C) and uPA (Fig. 5D). Activity assay (Fig. 5E) and RT-PCR (Fig. 5, F and G) further confirmed the above findings. Although transcriptional analysis by luciferase reporter assay after IL-12 stimulation showed a significant reduction in MMP promoter activity (Fig.  1G), it failed to do so in presence of Ab F12 (Fig. 6A). However, although TIMP-1 promoter activity showed an enhancement after IL-12 challenge, it remained unaltered in the presence of Ab F12 (Fig. 6A), thereby implying an IL-12-mediated IFN-␥ independent regulation. Furthermore, to conclusively prove our observation that IL-12-stimulated reduction in the MMP promoter activities was actually mediated by the IFN-␥ it produces, JEG-3 cells when challenged with rhIFN-␥ (500 units/ ml) showed an inhibition in MMP-2 and MMP-9 promoter activity (Fig. 6A), similar to those observed earlier with IL-12, thereby suggesting an involvement of IFN-␥ in this process. rhIFN-␥, however, failed to influence TIMP-1 promoter activity. rhIFN-␥ alone also inhibited JEG-3 cell invasion, similar to that observed in the presence of IL-12, further proving the fact that the anti-invasive effect of IL-12 is most likely executed by the IFN-␥ it produces (Fig. 6B), which in turn was established by our observation that neutralization of IFN-␥ by Ab F12 failed to inhibit JEG-3 cell invasion, showing an invasion index greater than those of IL-12 challenged alone (summarized in Table III). Crystal violet staining after Matrigel invasion assay and absorbance at 550 nm (Fig. 6C) further confirmed the above facts.
Based upon the SEM and Matrigel invasion assay findings as well as our previous observations reported earlier (29), we speculated whether the retardation in the invasion and motility of JEG-3 cells as observed above in the presence of IL-12 was an outcome of an increment in the cell-cell adhesion mediated possibly by E-cadherin, a cell surface molecule, critical for maintaining homophilic interactions. To prove this, a cellcell adhesion assay was performed that showed an IL-12-mediated significant enhancement in the adhesion (compared with control) (Fig. 7A) along with an up-regulation of E-cadherin as observed by RT-PCR (Fig. 7B) and Western blot (Fig.  7C, lane 2). Results showed a dose-dependent increment in E-cadherin mRNA and protein expression. To investigate the outcome of these observations on E-cadherin expression and to correlate the IL-12-stimulated production of IFN-␥ with invasion, confocal microscopy was performed for immunolocalization of E-cadherin on JEG-3 cells, which revealed a strong localization (as compared with control) at the cell-cell junction (Fig. 7D, P2) and was inhibited upon the addition of Ab F12 (Fig. 7D, P3). Challenge with rhIFN-␥ showed a similar enhancement of E-cadherin as observed with IL-12 (Fig. 7D, P5), which was abrogated by the IFN-␥-neutralizing Ab F12 (Fig. 7D, P6), thereby implying that IL-12 possibly executes its anti-invasive effect by an IFN-␥-mediated up-regulation of E-cadherin DISCUSSION Trophoblast invasion forms an integral part during the process of blastocyst implantation, resulting in the proper anchorage of the fetus to the endometrium. This invasion differs largely from tumor invasion both in terms of its stringent regulation and temporo-spatial restriction, executed through a complex paracrine-juxtacrine, cytokine-hormonal network (16, 36 -38). Trophoblasts-derived proteases belong-ing to the class of MMPs, serine proteases, and cysteine proteases like cathepsins are believed to be responsible for the execution of these invasive programs (39 -41). Although invasion normally results from the combined action of all these proteases, MMPs seem to contribute substantially during trophoblast invasion, digesting the chemically complex extracellular matrix collagens (42). Regulation of MMPs and uPA, thus, appear to form the crucial step in moderating invasion, which also involves the participation of metalloprotease inhibitors like tissue inhibitors of metalloproteases, TIMPs (4,41,43,44), and plasminogen activator inhibitors, PAIs (45,46). These inhibitors, thus, counteract the action of the respective proteases. TIMP-2 is likely to be the physiologic inhibitor of 72-and 92-kDA gelatinases, whereas TIMP-1 is the primary collagenase inhibitor (47).
Modulation of the proteases could be executed at the level of its synthesis or release, influenced by different factors, or by changes in the synthesis of their inhibitors. Although IL-1␤ and transforming growth factor ␤1 were reported in recent years to execute a pivotal role during trophoblast protease modulation and invasion (7,(47)(48)(49), the role of IL-12 in this process is largely unknown. Uterine fluid rich in growth factors, cytokines, and hormones are, thus, likely to modulate trophoblast invasion (50 -52).
IL-12 in our present study was found to exhibit an antitumor and anti-metastatic behavior, as evidenced by its effect  to reduce the proteases, cell invasion, and motility. Our findings are in agreement with reports by several authors about a similar anti-tumor role of IL-12 mediated by the activation of NK cells (53) or inhibition of angiogenesis and the MMPs (54 -56). Furthermore, electroporation of the IL-12 gene in an expression plasmid seems to prevent subcutaneous and metastatic murine tumor, or its targeted delivery to neo-vasculature enhances its anti-tumor activity (54). Dias et al. (57,58) reported the antitumor activity of IL-12 in a murine breast cancer model, which was mediated by reducing vascular endothelial growth factor, MMP-9, and by increasing the level of inhibitor, TIMP-1. Similar findings were also obtained in our present study with JEG-3 cells, wherein IL-12 was found to increase both the mRNA transcript and protein of TIMP-1, as detected by RT-PCR and immunocytochemistry. Based upon the above findings, we may thus conclude that IL-12 possibly moderates trophoblast invasion by inhibiting the proteases (both MMPs and uPA) and up-regulating the protease inhibitor (TIMP-1), thereby biasing the ratio (protease/protease inhibitor) in favor of the inhibitors. Wound healing assay revealed the functionality of these findings, with IL-12 significantly compromising (in respect to the control cells) the capacity of cells to fill up the wounded area, possibly due to an IL-12mediated inhibition in the cellular motility, although IL-12 in our study marginally enhanced JEG-3 cell proliferation, as detected by flow cytometry and thymidine incorporation assay (data not shown). SEM showed a lack of cell-to-matrix interac-tion and an enhanced intercellular adhesion, observed by rounded aggregated cells after IL-12 stimulation, which in turn contributes to the anti-invasive and anti-tumor role of this cytokine since a reduced cell-matrix and an enhanced cell-tocell interaction essentially leads to a non-motile, non-invasive phenotype (59, 60). These facts were further substantiated by our observation that IL-12 also up-regulated E-cadherin expression. E-cadherin is a calcium-dependent, cell-to-cell adhesion molecule expressed mostly by epithelial cells with its down-regulation or loss associated strongly with an invasive phenotype (61,62). Hence, up-regulation of E-cadherin implies that influence on cell-cell adherence in addition to down-regulation of the proteases may be the plausible mechanism by which IL-12 executes its anti-invasive role. We may, thus, deduce the essential role of this cytokine in controlling trophoblast invasion, as justified by its localization at the feto-maternal interface of the murine placenta and the uterine-infiltrating immune cells (19). IL-12, by virtue of its anti-invasive role, possibly plays an important role during implantation and to the best of our knowledge is perhaps the first report of this kind, depicting the reciprocal relationship between IL-12 and trophoblast invasion. We speculated whether the IL-12-mediated enhancement in IFN-␥ plays a role in inhibiting the MMPs, since similar behavior of interferons were observed in endothelial and other cell types (63,64). Such a possibility is explored using IFN-␥-neutralizing Ab, F12. Results clearly showed a dramatic effect in terms of MMPs and uPA expres- sion, as the usage of this neutralizing F12 Ab abrogated the IL-12-mediated inhibition of the gelatinases and uPA, although TIMP-1 remained unaltered. Furthermore, Ab F12 inhibited the anti-invasive effect of IL-12, as observed in Matrigel invasion assay. However, failure of Ab F12 to alter TIMP-1 despite being inducible under IL-12 stimulation implies additional regulatory pathways, possibly independent of IFN-␥. We may, thus, conclude that the anti-invasive function of IL-12 is mediated by an IFN-␥ dependent down-regulation of MMPs and up-regulation of E-cadherin.
IFN-␥ is a type II interferon and executes its function by binding to the cell surface receptor heterodimer, IFN-␥R1 and IFN-␥R2, which functions through Janus kinases, JAK1 and JAK2. JAK1 and JAK2 kinases are used in turn to phosphorylate the downstream STAT-1␣ protein, which dimerizes, translocates to the nucleus, and induces target gene transcription by binding to ␥-activated sequences in the promoters of IFN-␥-responsive genes. IL-12 has also been proposed to be an activator of STAT 4 in T-lymphocytes (65). We searched the upstream regulatory sequences of MMP-2 and MMP-9 for the presence of cis-acting ␥-activated sequences or interferon-stimulated response element using varieties of available software programs (TFSEARCH, TESS, etc.) but failed to locate one within the promoter of MMP-2 or MMP-9 gene, thereby ruling out the possibility of a direct IFN-␥-activated STAT-1␣ binding to these regulatory sequences to reduce gelatinase expression. MMP-9, with its ϳ2.5-kilobase upstream promoter, houses the binding of an array of transcription factors like AP-1, Sp-1, GT box, etc. within its proximal promoter region. MMP-2 promoter analysis has been carried out in several cell types of human origin. The MMP-2 promoter has no TATA or CAAT boxes and is considered to be under constitutive control via GC-rich promoter regions. The presence of the GATA-2 binding site within the MMP-2 promoter seems to be necessary to drive the gene under extracellular matrix stimulation (66 -67). Although the MMP-2 gene has been considered refractory to transcriptional modulation, identification of several potential transcription factor binding sites in recent times within its promoter like Sp-1, AP-1, CREB, and p53 opened up newer possibilities of regulation by these cis-acting elements (68). Analysis of the 5Ј-untranslated region of the human TIMP-1 gene also showed the presence of several transcription factor (Sp1, AP-1, and polyoma enhancer A3) binding sites (69), implying a complex regulation.
IL-12-mediated negative regulation of MMPs might also be achieved as an outcome of complex cross-talks emerging from STAT4 as well as IFN-␥-mediated STAT-1␣ signaling, which may in turn affect the function of various co-activators and components of the basal transcriptional machinery. Among all the identified co-activators, CREB-binding protein (CBP)/p300 and the other co-activators of transcription factors, including c-Fos, c-Jun, p53, CREB, YY1, STAT-1␣, and STAT-2 may be involved (70 -73). CBP/p300 by its intrinsic histone acetyltransferases activity, can affect transcription of a wide range of genes upon interaction with these transcription factors (74,75), thereby implying that coordination of cell signaling, chromatin remodeling, and stepwise recruitment of transcription regulators is critical to precisely regulate MMP-9 gene transcription in a temporally and spatially dependent manner (76).
Future work in this area will not only shed light on the molecular details of the processes, delineating the essential components down stream to IL-12 signaling, but will in turn also open potential avenues for clinico-pharmacological interventions in view of its anti-invasive, anti-metastatic role. Present work in our laboratory is directed in developing STAT-1␣ and STAT-4 negative trophoblasts cell lines to delineate the essential role of these cis-acting elements during IL-12 signaling. The effect of E-cadherin antisense after IL-12 or IFN-␥ stimulation is under investigation. Developmental biologists will be specifically benefited from these in terms of understanding the regulation of trophoblast invasion and etiology of various gestational trophoblastic disorders and trophoblastic malignancies.