Regulation of urokinase-type plasminogen activator production by cultured human cytotrophoblasts.

Cultured human cytotrophoblasts synthesize and secrete urokinase-type plasminogen activator (uPA) during the first 24 h of culture, but secretion declines during the subsequent day. In contrast, synthesis and secretion of fibronectin increases during the 2 days of culture. The levels of uPA mRNA parallel the changes in synthesis and secretion of uPA. Treatment of cytotrophoblasts with 8-bromo-cAMP (1.5 mM) transiently raises uPA mRNA levels and uPA secretion. This treatment reduces fibronectin mRNA levels and causes a sustained increase in beta chorionic gonadotropin mRNA content and chorionic gonadotropin secretion. We conclude that a cAMP-mediated process up-regulates uPA expression in cytotrophoblasts. However, the stimulatory effect of the cyclic nucleotide analog on uPA is transient.

4 To whom correspondence should be addressed: Dept. of Obstetrics and Gynecology, Hospital of the University of Pennsylvania, Philadelphia, P A 19104.
The abbreviations used are: uPA, urokinase-type plasminogen activator; hCG, human chorionic gonadotropin; HEPES, 4-(2-hy-droxyethy1)-1-piperazineethanesulfonic acid; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; tPA, tissue plasminogen activator. mouse (2), but little is known regarding PA production by human trophoblast. We have developed a method to isolate cytotrophoblasts, the cells which invade the uterus, from human placenta (4). These cytotrophoblasts respond to the analog of CAMP, 8-bromo-cAMP, by increasing production of chorionic gonadotropin (hCG) and progesterone (5, 6). Here we report that the cultured cytotrophoblasts synthesize and secrete uPA in an evanescent fashion, and that 8-bromo-cAMP acutely increases the synthesis and secretion of uPA by elevating uPA mRNA.

MATERIALS AND METHODS
Preparation and Culture of Cytotrophoblnsts-Cytotrophoblasts were isolated from term placentas obtained following normal spontaneous vaginal delivery or uncomplicated cesarean section, as previously described by Kliman et al. (4). This procedure yields a highly purified (-95% pure) preparation of cytotrophoblasts with >90% viability. Briefly, villous tissue was subjected to three 30-min digestions with 0.125% trypsin and 0.2 mg/ml DNase I (Sigma). Collected cells were applied to a 5-70% Percoll gradient. After centrifugation at 1200 X g, the middle band (density 1.048-1.062 g/ml) containing the cytotrophoblasts was removed, washed, and diluted to a concentration of 1 X 10' cells/ml with Dulbecco's modified Eagle's medium containing 25 mM glucose, 25 mM HEPES, and 50 pg/ml gentamycin (DMEM-HG). In experiments in which uPA synthesis and mRNA were quantitated, the medium was supplemented with 20% heatinactivated fetal calf serum, whereas in experiments in which uPA activity was determined, cells were cultured in serum-free medium on dishes coated with type I collagen (Sigma). Cell suspensions were plated into 35-mm Nunclon (Nunc, Roskilde, Denmark) culture dishes (1 X lo6 cells in 2 ml of medium) and incubated in humidified 5% Con, 95% air at 37 "C. 8-Bromo-CAMP (1.5 mM, Sigma) was added at the time of plating to some cultures. This concentration of 8-bromo-CAMP consistently stimulates hCG and progesterone secretion by the cytotrophoblasts (5,6). Media were changed after 24 h and experiments were terminated after the second 24-h period. At termination, cells were harvested by scraping with a plastic spatula. Each experiment was repeated on at least three separate occasions using different cell preparations.
Detection of uPA Activity-serum-free media from control and 8bromo-CAMP-treated cell cultures were analyzed by zymography using the method of Granelli-Piperno and Reich (7) as modified by Heussen and Dowdle (8). Media were mixed with an equal volume of 2 X Laemmli (9) sample buffer and electrophoresed in 10% polyacrylamide slab mini-gels (Idea Scientific, Corvallis, OR). Each experiment involved two gels, the first contained 0.1% gelatin to detect non-plasminogen-dependent proteases, and the second contained gelatin and purified human plasminogen (Sigma) to detect plasminogendependent proteases. Molecular weight standards (Bethesda Research Laboratories, Bethesda, MD), purified uPA (Sigma), and recombinant tissue plasminogen activator (tPA), generously provided by Genentech Corp. (San Francisco, CAI, were applied to each gel. Electrophoresis was carried out at 4 "C for 2% h. Following SDS-PAGE, gels were treated for 1 h with 2.5% Triton X-100 in 50 mM Tris-HC1, pH 8.0, at room temperature to remove the SDS. The gels were washed three times for 1 h with 50 mM Tris-HC1, pH 8.0, and then incubated for 16 h at 37 "C, fixed, and stained with 0.1% Coomassie Brilliant Blue R-250. Metabolic Labeling of Secreted uPA and Fibronectin with P S I -Methionine-Two hours before labeling with [%]methionine, the serum-containing media were removed and replaced with an equal volume of serum-and methionine-free DMEM-HG. [%]Methionine (100 pCi/ml, Du Pont-New England Nuclear) was then added and the cells were incubated for 2 h. Media were collected and cells harvested.
Equal volumes of culture medium (100 pl) from control and 8bromo-CAMP-treated cultures were employed for immunoprecipitation of uPA using 250 pg of an I g G fraction of an antiserum raised in 10903 rabbits against human uPA (Green Cross, Osaka, Japan). As a control, immunoprecipitations with an IgG fraction of an antiserum raised in rabbits against recombinant human tPA were performed. The immunoisolation procedure employed Pansorbin (Behring Diagnostics) to precipitate the immunoglobulin (10). The specificity of the uPA antibody was assessed by performing immunoisolations in the presence of 20 pg of purified uPA (Abhokinase, Abbott). Fibronectin was immunoisolated from the medium using 20 pl of a rabbit anti-human fihronectin antibody (Behring Diagnostics) and protein A-Sepharose CL-4B to precipitate the IgG-bound fibronectin (11). Immunoprecipitation of labeled fibronectin could be prevented by the addition of 200 pg of plasma fibronectin (Biomedical Technologies, Inc., Cambridge, MA) to the immunoprecipitation mixture. The immunoisolates were solubilized in Laemmli buffer (9), heated for 10 min a t 90 "C, and subjected to SDS-PAGE using 7.5% (uPA) or 5% (fibronectin) polyacrylamide gels in the presence of 2-mercaptcetha-no1 (9). "C-Labeled molecular weight markers were run on each gel. Following electrophoresis, gels were fixed for 1 h, impregnated with 1.0 M sodium salicylate, 1% glycerol, dried, and stored in contact with Kodak X-Omat film (Eastman Kodak, Rochester, NY) at -70 "C for autoradiography.
RNA Isolation, Electrophoresis, and Blot Hybridization-RNA was isolated from cells using the guanidine isothiocyanate-cesium chloride method (12). Equal amounts of RNA (10 pg), quantified by absorbance a t 260 nm, were denatured in MOPS/formaldehyde and electrophoresed in 0.8% agarose/formaldehyde gels and transferred to Nytran paper (Schleicher & Schuell) by standard procedures (12). A Hind111 digest of bacteriophage X DNA was also electrophoresed in these gels to estimate nucleic acid size.
Hybridizations with nick-translated cDNA probe were performed according to the method of Berent et al. (13). Dried filters were first moistened with 6 X SSC (1 X SSC = 0.15 M NaCl and 0.015 M sodium citrate), and then placed in bags with 5-8 ml of prehybridization solution (50% formamide, 5 X SSC, 0.1% each of Ficoll, polyvinylpyrollidone, bovine serum albumin, SDS, and 250 mg/ml denatured salmon sperm DNA) and incubated a t 42 "C for 4 h. This solution was replaced with fresh solution to which was added nick-translated cDNA (5-10 X lofi cpm/ml) and nick-translated X DNA. Hybridizations were continued for 18 h. Following hybridization, the filters were washed twice for 30 min a t room temperature in 2 X SSC and 0.1% SDS, followed by two 45-min washes at 65 "C in 1 X SSC and 0.1% SDS. After the washes the filters were blotted, wrapped in plastic, and placed with x-ray film (Kodak X-Omat) for autoradiography at -20 "C. Autoradiograms were scanned with a Kontes densitometer (Vineland, NJ) to determine relative changes in mRNA levels.
Plasmid Preparation, Isolation, and Nick Translation-A plasmid containing a 609-base pair cDNA for human uPA from pHUK-1 (14) was generously provided by Dr. Francesco Blasi (International Institute of Genetics and Biophysics, Naples, Italy). The cDNA was inserted into vector SP64 at the PstI/EcoRI site. A plasmid haboring a 579-base pair cDNA for hCG 8-mRNA (15) was kindly provided by Dr. John Fiddes (California Biotechnology, Mountain View, CA). Another plasmid containing a 1300-base pair cDNA corresponding to a portion of the 3"coding sequence and 3"noncoding sequence of human fibronectin (16)  hCG Assay-hCG secreted into the incubation medium was quantitated with a commercially available radioimrnunoasssay (Serono Diagnostics, Braintree, MA). This assay is calibrated to the First International Reference Preparation.

RESULTS
uPA Secretion and Synthesis by Culture Cytotrophoblasts-Cultured cytotrophoblasts secreted a plasminogen-dependent protease during the initial 24 h of culture, but enzymic activity declined during the subsequent 24 h of incubation to approximately one-third of that released during the 1st day (Fig. 1). The major plasminogen-dependent activity migrated with a M , value of -50,000. No proteolytic activity with M , values corresponding to M , -50,000 uPA or M , -70,000 tPA was detected when plasminogen was omitted from the polyacryl- amide gels. However, hydrolysis of gelatin by activities at higher and lower apparent M, values was observed in both the plasminogen-containing and the control gels. The nature of these protease activities is not known. The fibrin film overlay method (17) was employed in some analyses and revealed a single band of lysis which corresponded to the mobility of the uPA standard. Plasminogen activator activity in the media was also detected using the Spectrolyse assay (American Diagnostics, Inc., Greenwich, CN). Enzyme activity revealed by this method was completely abrogated by addition of an antibody to uPA, but not an antibody against tPA. Therefore, the plasminogen activator secreted by the cultured cytotrophoblasts is primarily uPA.
The de novo synthesis of uPA by the cultured cytotrophoblasts was demonstrated by metabolically labeling the cells with [35S]methionine followed by immunoprecipitation of labeled uPA from the culture medium. A major labeled protein with a M , -50,000 was isolated using an I& fraction of an antiserum to uPA (Fig. 2 4 ) . A smaller amount of label was incorporated into a lower molecular weight protein (M, -48,000). These proteins were not immunoprecipitated with an antibody to tPA, nor were they detected when immunoprecipitations were performed with the anti-uPA IgG in the presence of added human uPA. Since the SDS gels of the immunoprecipitates were run under reducing conditions, the M , -50,000 protein which was immunoisolated represents the single chain enzyme, since under reducing conditions the two chains of activated uPA would have been separated by reduction of the disulfide bond which bridges them, yielding labeled proteins of approximately M , 30,000 and 20,000 (1).
The synthesis of uPA by cultured cytotrophoblasts as assessed by incorporation of [35S]methionine was greatest during the first 24 h of culture (Fig. 2B). The amount of radioactivity incorporated into uPA was substantially reduced after the 2nd day of culture. In contrast, the synthesis and secretion of fibronectin increased appreciably during the 48 h of incubation (Fig. 2B). Therefore, the decline in synthesis and secretion of uPA was not due to a general reduction in protein synthesis and secretion. uPA mRNA Leuels in Cultured Cytotrophoblasts-Blot hy-  3. Hybridization blot analyses of uPA, fibronectin, and &hCG mRNAs extracted from cultured cytotrophoblasts. Total RNA (10 pg) extracted from cells exposed to the indicated treatments were electrophoresed in agarose gels, transferred to Nytran, and probed with specific "P-labeled cDNAs. The representative filter shown was probed sequentially with cDNAs for uPA, fibronectin, and B-hCG. Nucleic acid size markers are presented in kilobases.

DISCUSSION
Human trophoblast has been shown to degrade extracellular matrix (19). The present study reveals for the first time that human cytotrophoblasts synthesize and secrete uPA, which may well account, in part, for the previously described capacity of placental cells to break down extracellular matrix proteins. It is intriguing that the production of uPA by cultured trophoblasts is evanescent. Activity is expressed primarily during the first 24 h of culture and then declines. The fall in uPA occurs in concert with a rise in the level of mRNA encoding the extracellular matrix protein, fibronectin, under our basal culture conditions. This temporal pattern of initial production of a protease which degrades extracellular matrix followed by increased elaboration of an extracellular matrix protein raises the possibility that these processes might be coupled in some way.
The factors which stimulate uPA synthesis and secretion by trophoblasts and then suppress its expression remain to be determined. cAMP seems to have some role in the upregulation of uPA in trophoblasts as it has in other cell systems (3). Preliminary studies have revealed that uPA mRNA levels in the cytotrophoblasts are increased within the initial 4 h after addition of 8-bromo-cAMP, demonstrating a relatively rapid response. However, the stimulatory effect of the cAMP analog is short lived, as uPA mRNA levels are markedly reduced after 48 h of culture in the presence of the cyclic nucleotide. The inability of 8-bromo-CAMP to sustain trophoblast uPA secretion is not due to development of a refractory state since P-hCG mRNA and hCG secretion are increased throughout the exposure to the cyclic nucleotide analog (5, 6). The transient stimulation of uPA by 8-bromo-cAMP in our system is very similar to the transient stimula- The factors responsible for the circumscribed stimulation of plasminogen activator expression in these systems remain to be explored.