Hypotonic Stress Increases Cyclooxygenase-2 Expression and Prostaglandin Release from Amnion-derived WISH Cells*

This report examines the effect of cell volume expansion on cyclooxygenase-2 (COX-2) mRNA expression, COX-2 protein expression, and prostaglandin E2 release from human amnion-derived WISH cells. Earle’s balanced salts solution (EBSS) with limited NaCl concentration was utilized as the induction medium. COX-2 mRNA was elevated 6-fold in cells incubated for 1 h in hypotonic EBSS. COX-2 mRNA expression was not increased when raffinose or sucrose were used to reconstitute low NaCl. Actinomycin D blocked COX-2 mRNA increase by hypotonic stress, while cycloheximide enhanced COX-2 mRNA expression. COX-2 mRNA and protein concentrations increased as a function of decreasing media osmolarity and incubation time in hypotonic EBSS. Hypotonic EBSS induced a 3-fold increase in prostaglandin E2 release. WISH cells transiently transfected with a luciferase expression vector driven by the human COX-2 promoter for the COX-2 gene show a 3-fold increase in luciferase activity when incubated in hypotonic EBSS. COX-2 mRNA levels in primary human amnion cells were also increased by hypotonic stress. This study suggests that amnion cell COX-2 gene expression is regulated by cell volume expansion and/or increased plasma membrane tension.

Prostaglandins have a central role in regulating human parturition. Prostaglandin E 2 (PGE 2 ), 1 the primary prostaglandin produced by fetal membranes during labor, may be directly involved in the initiation and maintenance of uterine contractions (1,2). Untimely increases in prostaglandin biosynthesis early in gestation may be responsible for inducing preterm labor in some individuals (3). Although many hormones, growth factors, and cytokines have been reported to increase or decrease the synthesis and/or release of PGE 2 in fetal tissues, the physiological factor(s) that up-regulates prostaglandin biosynthesis during parturition has not been identified.
Metabolism of arachidonic acid to prostaglandin H 2 is a key and rate-limiting step in prostaglandin biosynthesis. The reaction is catalyzed by cyclooxygenase. Two isoforms of cyclooxygenase have been identified, designated as COX-1 and COX-2. Both enzymes have been characterized in considerable detail (for review, see Refs. 4 -8). The genes for COX-1 and COX-2 are encoded on different chromosomes and have been sequenced. The gene for COX-1 is constitutively expressed, present in most tissues at low to nondeductible levels, and is generally considered to have "housekeeping" functions. In contrast, the gene for COX-2 has been characterized as an immediate early response gene. A wide range of mitogens, hormones, cytokines, and endotoxins increase the rates of COX-2 gene transcription and prostaglandin biosynthesis (4 -8). COX-2 mRNA and COX-2 protein have been reported to increase near the onset of labor. A major site of PGE 2 synthesis during labor occurs in the amnion (9,10). COX-2 protein concentrations (11) and mRNA levels (10) are higher in amnion from women in labor versus patients not in labor. The underlying mechanism(s) leading to enhanced COX-2 mRNA expression in amnion during labor remains to be identified.
Fetal membranes clearly undergo mechanical stretching as a result of several processes during gestation, including fetal growth, increased amniotic fluid volume, and labor. Increased membrane tension of fetal cells may also occur as a direct result of cell volume expansion; human amniotic fluid osmolarity decreases as a function of advancing gestational age (12), raising the possibility that increased amnion cell plasma membrane tension might occur, in part, by changes in cell volume. Furthermore, as initially observed by Danforth and Hull (13) and as summarized by Alger and Pupkin (14), little if any amnion cell mitotic activity is observed in the latter part of gestation, and therefore, to accommodate the growing fetus, existing amnion cells must increase their size by stretching and hypertrophy. It has been recognized for some time that mechanical stretching of human cultured amnion cells increases the release of PGE 2 (15). Although the biochemical mechanism by which mechanical stretching increases prostaglandin release from these tissues is largely unknown, it was recently reported that mechanical stretching of cultured rat glomerular mesangial cells induces a number of immediate early response genes including the gene for COX-2 (16). Based on this knowledge, we hypothesize that an increase in the volume of amnion cells and the resulting increase in plasma membrane tension induce COX-2 gene transcription and promote PGE 2 release from fetal tissue. This study provides evidence that an increase in cell volume up-regulates COX-2 mRNA expression and elevates prostaglandin biosynthesis in amnion cells.  1 The abbreviations used are: PGE 2 , prostaglandin E 2 ; COX-1 and -2, cyclooxygenase-1 and -2, respectively; EBSS, Earle's balanced salts solutions; MOPS, 3-(N-morpholino)propanesulfonic acid; SSC, sodium chloride/sodium citrate buffer; EGF, epidermal growth factor; BSA, bovine serum albumin; kb, kilobase(s). culture flask into 60 ϫ 15-mm Falcon culture dishes (1.5 ϫ 10 6 cells/ dish) containing 5.0 ml of DF10F. Unless otherwise stated, cultures were incubated at 37°C in humidified air containing 5% CO 2 , fed on day 3 with DF10F, and used for experiments on day 5. Under these growth conditions, WISH cells reach confluence between the third and fourth day of culture.

Culture and Treatment of WISH
Cells cultured for 5 days in DF10F were washed with 2.0 ml of Earle's balanced salts solution (EBSS). Induction medium (5.0 ml, equilibrated to 37°C), as defined under "Results," was added to cultures that were then incubated at 37°C in humidified air containing 5% CO 2 . As described below, cells were then harvested for preparation of mRNA or COX-2 protein. In some experiments, spent incubation media were removed, frozen at Ϫ70°C, and, as described below, analyzed for prostaglandin concentrations.
Preparation and Culture of Primary Human Amnion Cells-Human amnion cells were cultured using a modification of the method of Okita et al. (17). Placentae and associated membranes were obtained from women undergoing repeat cesarean section. All tissue manipulations were performed using aseptic technique and sterile (0.2 M filtered) solutions. Amnion was stripped from choriodecidua and washed successively in 200-ml changes of ice-cold Ca 2ϩ -and Mg 2ϩ -free phosphatebuffered saline containing gentamicin (50 g/ml) until clear of blood. Tissue was minced in 50 ml of TEP buffer (0.05% trypsin (SIGMA, Type II), 0.02% EDTA in phosphate-buffered saline) and then incubated at 37°C for 20 min in a shaking water bath. The preparation was filtered through 2-mm stainless steel wire gauze. The filtrate was discarded, and the tissue fragments were recombined with 150 ml of TEP and incubated for 90 min. Following incubation, the cell suspension was combined with 450 ml of ice-cold phosphate-buffered saline, shaken for 15 s, and filtered sequentially through 2-and 0.7-mm stainless steel wire gauze. The filtrate was centrifuged at 1000 ϫ g for 8 min at 4°C. Pelleted cells were resuspended in Earle's minimal essential medium, plated in 150-mm tissue culture dishes, and then incubated at 37°C in 5% CO 2 at 95% relative humidity. Medium was changed initially at 24 h and subsequently at 48-h intervals. Cells were grown to confluence and used directly.
Preparation of RNA-Induction media were removed from dishes, and 2.5 ml of RNA lysis buffer (4.0 M guanidine isothiocyanate, 25 mM sodium citrate, pH 7.0, 0.5% sarcosyl, and 0.1 M 2-mercaptoethanol) was added to cell cultures. Cultures were allowed to stand in the lysis buffer for 5 min at room temperature and then mixed repeatedly by pipette. Cell lysates were transferred to Falcon number 2059 test tubes and frozen at Ϫ70°C until preparation of RNA. Frozen cell lysates were thawed and passed three times through an 18-gauge needle. The sample (2.1 ml) was layered on a 2.5-ml pad of 5.7 M CsCl containing 0.1 M EDTA and then centrifuged at 35,000 rpm for 18 h at 18°C employing a Beckman SW 50.1 rotor.
An aliquot from each RNA sample was diluted into TE buffer (10 mM Tris plus 0.1 mM EDTA, pH 7.4), and the concentration of RNA was estimated spectrophotometrically as described previously (18). Specific amounts of RNA (10 g/well for Northern analysis and 5 g/well for dot blots) from each sample were aliquoted into microcentrifuge tubes. Sufficient diethylpyrocarbonate-treated water was added to each sample to provide the same RNA concentrations throughout all tubes. For Northern analyses, RNA precipitation solution (0.1 volume of 3 M sodium acetate, pH 5.0, followed by 2.5 volumes of 100% ethanol) was added to the samples, which were then stored at Ϫ70°C until used. For dot blot assays, RNA samples were processed as described below.
Northern Blots and Dot Blots-For Northern analysis, RNA samples were pelleted from the precipitation solution by centrifugation (12,000 ϫ g for 15 min at 5°C) and then dried by vacuum. Pellets were dissolved into 10 l of denaturing solution (5 l of formaldehyde, 2 l of formamide, 1 l of 10 ϫ MOPS buffer (0.2 M MOPS), 0.1 M sodium acetate, 10 mM EDTA, pH 7.0), and 2 l of H 2 O) plus 1 l of ethidium bromide (400 g/ml H 2 O). The samples were heated for 10 min at 65°C and then chilled on ice. Samples were electrophoresed in formaldehydeagarose gels (1.0% agarose (w/v), 6.6% formaldehyde (v/v) in 1 ϫ MOPS buffer). Electrophoresis was carried out in 1 ϫ MOPS buffer overnight (0.15 V/cm 2 ). RNA was transferred by capillary action to Nytran filters overnight employing 10 ϫ SSC (1.5 M sodium chloride, 150 mM sodium citrate, pH 7.0) as the transfer buffer. The Nytran filters were air-dried and then baked at 80°C for 1.5 h. As previously demonstrated (19), utilization of ethidium bromide under the above denaturing conditions provided direct evidence that (a) approximately equal amounts of RNA from different cell cultures were applied to gels for electrophoresis, (b) RNA integrity was maintained during sample preparation and electrophoresis, and (c) the transfer of RNA from gel to filter was complete.
RNA dot blots were performed with a VacuDot-VS manifold (Amer-ican Bionetics). Nytran filters were prewet with 6 ϫ SSC prior to applying samples. RNA solutions were diluted with 4 volumes of denaturing solution and then heated at 65°C for 15 min. Solutions of denatured RNA were chilled on ice, diluted with 1.5 volumes of 6 ϫ SSC, and then loaded into manifold wells. The samples were allowed to drain by gravity for 30 min prior to applying a vacuum. Each well was vacuum-washed twice with 400 l of 6 ϫ SSC. Filters were air-dried and then heated at 80°C for 1.5 h. Filters from both Northern blots and dot blots were then incubated for 2 h at 42°C in prehybridization buffer as described previously (20). The 32 P-labeled probe for COX-2 mRNA was generated with a random primed labeling kit (Amersham International plc, Buckinghamshire, United Kingdom). The substrate was the 1.2-kilobase COX-2 cDNA insert purified from plasmid pcDNAhCOX-2. After prehybridization, the 32 P-labeled COX-2 cDNA probe was denatured and then added directly to the prehybridization buffer (1.0 ϫ 10 6 dpm/ml). Filters were incubated overnight at 42°C. They were then incubated twice in 6 ϫ SSC buffer plus 0.5% SDS and once in 1 ϫ SSC plus 0.1% SDS for 15 min at room temperature. The final wash was carried out in 1 ϫ SSC plus 0.1% SDS for 30 min at 56°C.
Autoradiography and COX-2 mRNA Analysis-Nytran filters from Northern blots and dot blots were exposed to Kodak X-Omat RP film at Ϫ70°C employing Lighting Plus intensifying screens. Autoradiographs of 32 P-hybridized RNA dot blots were scanned with a Microtek MSF-300GS image scanner (Microtek, Torrance, CA) that was linked to a Macintosh IIsi computer. The image generated by the Microtek grayscale scanner was captured by Image Studio software (Letraset, Paramus, NJ) and then analyzed for intensity of grays relative to background by Scan Analysis software (Biosoft, Milltown, NJ). Quantitation of image intensities on film was carried out at less than maximal densities as described previously (20). When an image from a given dot blot was determined to be overexposed, autoradiography was repeated employing a shorter exposure time.
Western Blot Analysis-Monolayers of WISH cells were suspended in 0.5 ml of lysis buffer (20 mM Hepes (pH 7.2), 120 mM NaCl, 1% Triton X-100, 5 g/ml aprotinin, 10 g/ml antipain, 5 g/ml leupeptin, 5 g/ml pepstatin A, and 0.1 mM phenylmethylsulfonyl fluoride) and frozen at Ϫ70°C. Cell homogenates were thawed on ice and maintained at 4°C throughout manipulation. Homogenates were sonicated for 10 s (Heat Systems-Ultrasonics, Inc., model W225R with microtip), setting 4, 100% duty cycle, and then cleared by centrifugation at 12,000 ϫ g for 5 min at 4°C. Aliquots (25 l) of homogenates were subjected to polyacrylamide gel electrophoresis on 8% gels. Gels were equilibrated in 20 mM Tris, 190 mM glycine, 20% methanol, pH 8.3 overnight at 10°C. Resolved proteins were semi-dry transferred for 1 h at 10 V (Bio-Rad, Melville, NY) to Hybond Nϩ (Amersham Corp.). Western blots were blocked in wash buffer plus bovine serum albumin (BSA) (10 mM Tris, pH 7.2, 100 mM NaCl, 0.2% Tween-20, containing 1% fatty-acid free BSA) overnight at 4°C prior to incubation with COX-2 antibody (1:500 dilution in wash buffer plus BSA) for 1 h at 25°C. Blots were washed in three changes of wash buffer at 10-min intervals and then incubated with a 1:2000 dilution of anti-mouse IgG-horseradish peroxidase conjugate in wash buffer plus BSA overnight at 4°C. Following incubation, blots were washed as before prior to signal generation by ECL (Amersham) and visualized by a 10-s exposure to Kodak X-Omat AR film.
Transfection-WISH cells were plated in 3.8-cm 2 wells at a density of 5.0 ϫ 10 5 cells, 2.0 ml of DF10F. After 24 h in culture the medium was removed, and cells were liposome-mediated transfected employing the following: 50 g of LipofectAMINE (Life Technologies, Inc.), 10 g of the COX-2 pXP1 luciferase reporter vector, 2 g of the secreted alkaline phosphatase genetic reporter system (pSEAP-Control vector; CLON-TECH, Palo Alto, CA), and 1.0 ml of DF10F lacking serum and antibiotics. Cells were incubated for 16 h in the initial transfection medium, after which an additional 1.0 ml of DF10F containing 10% fetal calf serum was added to each plate. The transfection medium was removed after 24 h, and cells were then incubated for an additional 48 h in DF10F plus 10% fetal calf serum minus antibiotics. Induction media, as defined under "Results," were added to plates, and cells were incubated for designated times at 37°C in 5% CO 2 . For alkaline phosphatase assays, induction media (500 l) were removed from each plate and then centrifuged at 12,000 ϫ g for 5 min at 10°C. The supernatants were frozen at Ϫ70°C until analysis by chemiluminescent detection using a commercial system (CLONTECH) and a 96-well format. Samples were exposed to x-ray film for various time periods to assure that detection was within the linear range of the film. For luciferase activity, cells were scraped into 200 l of luciferase assay lysis buffer (Promega) and then sonicated (Heat Systems-ultrasonics Inc.) for 5 s on ice at 50% output. Samples were centrifuged at 12,000 ϫ g for 5 min at 10°C.
Luciferase activity was determined using a commercial kit (Promega) by scintillation counting per the manufacturer's instructions.
Reagents, Media Osmolarities, and cDNA Probe-Prostaglandin antibodies and anti-mouse IgG-horseradish peroxidase conjugate were from Sigma. COX-2 antibodies were from Transduction Laboratories (Lexington, KY). PGE 2 was determined by radioimmunoassay assay using a commercial antibody and following the supplier's protocol (Sigma). Media osmolalities were determined by vapor pressure employing a Wescor 5500 Vapor Pressure Osmometer (Wescor Inc., Logan, Utah). The plasmid pcDNAhCOX-2, which contains mouse COX-2 cDNA (21), was a generous gift from Dr. Timothy Hla (American Red Cross, Rockville, MD). The vector containing the human COX-2 promoter-luciferase reporter construct (22) was a generous gift from Dr. Lee-Ho Wang (University of Texas, Houston).

RESULTS
Hypotonic Stress and COX-2 mRNA Expression-COX-2 mRNA expression was initially examined in WISH cells incubated for 1 h in hypotonic versus isotonic EBSS. Relative to the concentration of COX-2 mRNA from cells incubated in isotonic EBSS, COX-2 mRNA expression was markedly elevated in hypotonically stressed cells (Fig. 1, lanes 2, 5, and 8 versus  lanes 3, 6, and 9, respectively). The 32 P-labeled COX-2 cDNA probe hybridized to three mRNA species of approximately 5.8, 4.8, and 3.4 kb. All three mRNA species were elevated in hypotonically stressed cells. In contrast to COX-2 mRNA expression, no change in the level of COX-1 mRNA was detectable by Northern analysis employing a cDNA-specific probe for COX-1 (not shown).
The relative concentration of basal level COX-2 mRNA from untreated cells (i.e. cells harvested directly from spent tissue culture growth medium, DF10F) was also determined (Fig. 1,  lanes 1, 4, and 7). Basal COX-2 mRNA levels were consistently less than levels from WISH cells incubated in isotonic EBSS (compare lanes 1, 4, and 7 to lanes 2, 5, and 8, respectively). During the course of these studies it was determined that, due to evaporation, the osmolarity of fresh DF10F tissue culture medium increased from 290 -300 mosmol/liter to 315-320 mosmol/liter after 48 h of incubation. The osmolarity of standard isotonic EBSS routinely assayed between 280 and 290 mosmol/liter and, therefore, cells cultured for 48 h in DF10F medium and then incubated in isotonic EBSS were subjected to a decrease in osmolarity of between 25 and 40 mosmol/liter. Thus, shifting cells from 48-h spent tissue culture media that has become moderately hypertonic due to evaporation to standard isotonic EBSS is sufficient to increase COX-2 mRNA expression. As shown in Fig. 3, very little if any difference in COX-2 mRNA concentrations from untreated cells versus isotonic EBSS-treated cells were observed when the osmolarity of EBSS was increased by 30 mosmol/liter with 30 mM raffinose.
As determined by autoradiography and grayscale scanning of RNA dot blots hybridized to the 32 P-labeled COX-2 cDNA probe, the relative concentration of COX-2 mRNA from WISH cells incubated for 1 h in hypotonic EBSS was elevated approximately 6-fold above the COX-2 mRNA concentration from cells incubated in isotonic EBSS (Table I). Numerous reports have shown that EGF is a potent inducer of COX-2 gene transcription in a variety of biological systems, including human amnion cells. As a positive control in this study, and to compare the relative potency of hypotonic stress to a recognized inducer of COX-2 gene expression, the effect of EGF was also determined on WISH COX-2 mRNA expression. The relative concentration of COX-2 mRNA in cells incubated in isotonic EBSS plus EGF was elevated approximately 10-fold higher than that in cells incubated in isotonic EBSS (Table I).
Effects of Increased Cell Volume versus Reduced NaCl on COX-2 mRNA Expression-Hypotonic media in Fig. 1 and Table I were formulated by decreasing the amount of NaCl in EBSS. A previous study, employing rat papillary collecting tubule cells, reported that PGE 2 was induced in this system by a reduction in extracellular Na ϩ (23). To determine if elevated WISH cell COX-2 mRNA expression is related to cell volume expansion or, alternatively, related to a reduction in extracellular Na ϩ and/or Cl Ϫ , cells were incubated in isotonic EBSS in which 116 mosM sucrose or raffinose was substituted for 116 mosM NaCl (Fig. 2A). Only cells incubated in hypotonic medium were characterized by increased COX-2 mRNA expression. In the same experiment, the concentration of released PGE 2 was determined in induction media from WISH cells (Fig. 2B). PGE 2 was also elevated only in cells incubated in hypotonic EBSS.
Effects of Actinomycin D and Cycloheximide on COX-2 mRNA Expression-The effects of actinomycin D and cycloheximide on elevated COX-2 mRNA resulting from hypotonic stress were determined by dot blot analysis (Fig. 3). The addition of actinomycin D to hypotonic EBSS completely suppressed COX-2 mRNA concentrations to levels observed for untreated cells and cells incubated in isotonic EBSS. In contrast, the relative concentration of COX-2 mRNA in hypotoni-  cally stressed cells was not reduced by the addition of cycloheximide to hypotonic EBSS. Kinetics of COX-2 mRNA Expression-COX-2 mRNA expression was examined as a function of decreasing medium osmolarity (Fig. 4A). The relative concentration of COX-2 mRNA was characterized by a gradual increase as the osmolarity was reduced from 279 mosM to 200 mosM. A more pronounced increase in COX-2 mRNA expression was evident when the osmolarity was shifted from 200 to 178 mosM, the lowest osmolarity tested.
The relative concentration of COX-2 mRNA was also examined as a function of incubation time in hypotonic medium (Fig.  4B, open circles). A change in WISH cell COX-2 mRNA levels resulting from hypotonic induction medium (161 mosM EBSS) was detected within 30 min, the earliest time period examined. COX-2 mRNA levels remained elevated for an additional hour and then declined to near basal levels between 2.5 and 3.0 h.
Hypotonic Stress and COX-2 Protein Expression-Relative to cells incubated in isotonic EBSS, cells subjected to hypotonic EBSS (176 mosM and 218 mosM) were characterized by increased COX-2 protein concentration (Fig. 5). Two COX-2 protein bands of 72 and 74 kDa were detected, an observation that is consistent with the sizes of COX-2 protein reported for other systems (24). At 176 mosM, COX-2 protein appears to be maximally expressed between 1 and 2 h, and then it falls to basal levels between 3 and 6 h, a pattern that closely parallels the increase in COX-2 mRNA expression when cells are incubated in 161 mosM EBSS (Fig. 4B). When incubated in a less pronounced hypotonic EBSS medium (218 mosM), the increase in COX-2 protein expression was delayed for at least 2 h, increasing between 3 and 6 h, and remained elevated for several hours prior to returning to basal levels between 6 and 18 h (Fig. 5).
Response of COX-2 Promoter-Luciferase Construct to Hypotonic Stress-To ascertain the effect of hypotonic stress on COX-2 gene expression, WISH cells were transiently transfected with a human COX-2 promoter-luciferase construct. This reporter construct contains an 899-base pair fragment of the human promoter for the COX-2 gene, progressing 5Ј from the ATG start site, and coupled to the Pxp1 luciferase expression vector (22). Relative to cells incubated in isotonic EBSS, cells incubated in hypotonic EBSS were characterized by approximately a 3-fold increase in luciferase activity (Fig. 6). The level of luciferase induction observed for cells incubated in isotonic EBSS plus EGF was approximately 6-fold above cells incubated in isotonic EBSS (Fig. 6).
Hypotonic Stress and COX-2 Expression in Primary Human Amnion Cells-Because WISH cells are immortal and therefore transformed, it cannot be assumed that the activation of a given signal transduction pathway in this cell line is applicable to normal human amnion cells. Previous studies of primary human amnion cells in culture have show that they retain many of the biochemical and morphological properties of amnion cells studied immediately after parturition (17). For this reason the effect of cell volume expansion on COX-2 mRNA expression was determined on primary human amnion cells prepared from placentae of women undergoing repeat cesarean section (Fig. 7). Relative to cells incubated for 1 h in isotonic EBSS, COX-2 mRNA was markedly elevated in cells incubated in hypotonic EBSS. The increased concentration of COX-2 mRNA in primary human amnion cells resulting from hypotonic stress was approximately 50% of that induced by EGF (Fig. 7), a pattern that is consistent with WISH cell COX-2

FIG. 2. Effect of increased cell volume versus decreased NaCl on COX-2 mRNA expression (A) and PGE 2 release (B).
A, relative COX-2 mRNA concentrations in cells incubated for 1 h in the following media: standard isotonic EBSS (I); EBSS made hypotonic by reducing the concentration of NaCl from 116 mM to 58 mM (H); isotonic EBSS containing 116 mM sucrose in place of 58 mM NaCl (S); and isotonic EBSS containing 116 mM raffinose in place of 58 mM NaCl (R). Total RNA (5 g) was dot-blotted and probed for COX-2 mRNA as described under "Materials and Methods." An autoradiogram of the dot blot was then analyzed by grayscale scanning as described under "Materials and Methods." B, PGE 2 concentrations released into media from cells described in A. PGE 2 concentrations were determined by radioimmunoassay as described under "Materials and Methods." Each treatment represents the mean Ϯ S.D. of three replicate cultures.  (Table I). DISCUSSION The central finding in this study is that hypotonic stress increases prostaglandin biosynthesis in amnion cells. Results in this report demonstrate that COX-2 mRNA expression, the amount of COX-2 protein, and the release of PGE 2 from WISH cells originally derived from amnion are all enhanced by cell volume expansion.
The expression of three apparent COX-2 mRNA species in hypotonically stressed WISH cells was an unexpected finding (Fig. 1). Most earlier studies of COX-2 mRNA expression, employing a wide variety of biological systems, reported the presence of a single COX-2 mRNA species ranging in size from 4.2 to 4.8 kb (4 -8). Recently, in an extensive study of COX-2 mRNA isoforms induced by cytokines in human lung and kidney cells (25), three isoforms of COX-2 mRNA were detected: two major isoforms of 4.6 and 2.8 kb and a minor species of 4.1 kb. In the same study it was shown that the different isoforms are a product of alternative polyadenylation and, surprisingly, that the 2.8-kb COX-2 mRNA had a significantly longer halflife than the 4.6-kb species. In the current study, as determined by Northern analysis, two major COX-2 bands of 5.8 and 3.4 kb were detected, with a less pronounced band of 4.8 kb (Fig. 1). In a previous communication, interleukin-18 induced a single COX-2 mRNA species of 5.5 kb in primary human amnion cells from term placenta (26). The physiological significance of several COX-2 mRNA bands in WISH cells is unclear. To our knowledge, the large 5.5-5.8-kb species of COX-2 mRNA has and isotonic EBSS plus EGF (10 ng/ml) (IϩE). Untreated cells (U) were harvested directly from plates without prior treatment (basal luciferase activity). Results are expressed as the ratio of luciferase activity/plate to alkaline phosphatase/plate. Hypotonic media were formulated by decreasing the concentration of NaCl in EBSS from 116 mM to 58 mM. Cells were harvested, and the relative levels of luciferase and alkaline phosphatase activities were determined as described under "Materials and Methods." Cells were co-transfected with the human COX-2 promoter-luciferase vector and the SV40 alkaline phosphatase vector as described under "Materials and Methods." Values represents the mean Ϯ S.E. of three replicate cultures. only been detected in WISH cells and primary amnion cells. It remains to be established if the various COX-2 mRNA species in hypotonically stressed WISH cells are a result of alternative polyadenylation or alternative splicing. Knowledge of potential changes in the turnover rate of COX-2 mRNA, as a result of increased cell volume and/or increased membrane tension, may provide important information concerning the regulation of prostaglandin metabolism in amnion cells.
The observation that hypotonic stress induces luciferase in WISH cells transfected with a human COX-2 promoter-luciferase vector (Fig. 6), coupled with the knowledge that actinomycin D prevents hypotonic stress from increasing COX-2 mRNA levels (Fig. 3), suggests that changes in prostaglandin biosynthesis resulting from cell volume expansion are due to an increase in the rate of COX-2 gene transcription. Hypotonic stress induced a large increase in COX-2 mRNA levels within 30 min after treatment (Fig. 4B). The increase in COX-2 mRNA associated with cell volume expansion was not suppressed by cycloheximide (Fig. 3). Collectively, results in this study demonstrate that the gene for COX-2 acts as an immediate early response gene when stimulated by cell volume expansion. This suggestion is consistent with previous reports, employing a variety of biological systems, that classified the gene for COX-2 as an immediate early response gene following treatment with different growth factors, mitogens, and cytokines (4 -8).
The signal transduction pathway(s) that is activated in higher eukaryotes by cell volume expansion and leads to increased gene expression has not been identified. Nevertheless, recent studies employing yeast mutants have demonstrated that hypotonic stress activates the PKC1 pathway (27), one of four recognized yeast mitogen-activated protein kinase pathways (MAP kinase pathway). These studies demonstrated that a functional PKC1 pathway is required for the survival of yeast in a hypotonic environment. In contrast to hypotonic stress, hypertonic stress activates a different MAP kinase pathway in yeast, designated as the HOG pathway. This pathway is required for survival of yeast in a hypertonic environment. Interestingly, hypertonic stress of Chinese hamster ovary cells produces a marked increase in Jnk 1 (28), an enzyme similar to the yeast protein kinase HOG1. Furthermore, Jnk 1 was able to rescue yeast mutants lacking functional HOG 1 from hypertonic shock. Since yeast and mammalian systems appear to have similar MAP kinase pathways and these pathways may be important in cell volume regulation (27)(28)(29), it is possible that a PKC1-like pathway exists in amnion cells and that this pathway is activated by cell volume expansion. The end product of this pathway may, in turn, modify a factor that increases the rate of COX-2 gene transcription.
Cell volume expansion and membrane stretch have been shown to modify the expression of a wide variety of genes in a number of diverse biology systems. Previous studies have reported that increased cell volume resulting from hypotonic stress modifies the concentration or activity of a number of factors that impact, directly or indirectly, on the transcription rate of specific genes (for a review, see Ref. 30) including, for example, intracellular calcium, cyclic AMP, tyrosine kinases, inositol 1,4,5-trisphosphate, and protein kinase C. Although the mechanism by which cell volume expansion increases amnion cell COX-2 gene transcription remains to be characterized, the promotor region for the human COX-2 gene has been shown to contain a wide variety of potential regulatory elements, including CRE, NF-B, Sp1, and AP2 sites (22). Recently, what appears to be a novel putative cis-element in the promoter region of rabbit aldose reductase has been shown to be necessary for increased expression of this gene during hypertonic stress (31). Perhaps a novel regulatory element, yet to be identified, is essential for regulating gene expression as a function of cell volume expansion and/or increased membrane tension.
It has been recognized for more than 35 years that prostaglandins provoke uterine contractions (for a historical review of prostaglandins and uterine contraction, see Ref. 32). Within the past 10 years, a vast array of studies have documented that prostaglandins have an important role in human parturition. It is well established that amnion is a major site of prostaglandin biosynthesis. Approximately 10 years ago it was documented that mechanical stretching of cultured human primary amnion cells increases the release of PGE 2 (15). Recent studies have shown that prostaglandin biosynthesis increases in amnion prior to labor, and the increase is associated with elevated cyclooxygenase expression (10,11,33), a rate-limiting step in prostaglandin biosynthesis (4 -8). Although more than 30 hormones, mitogens, and cytokines have been shown to increase prostaglandin production, the mechanism by which increased PGE 2 biosynthesis occurs in fetal tissues remains unclear. Studies in this report raise the possibility that an increase in cell membrane tension induced by cell volume expansion and perhaps cell membrane stretching up-regulates the rate of COX-2 gene transcription, increasing PGE 2 biosynthesis and release in cells derived from human amnion. These observations raise the additional possibility that increasing mechanical forces resulting from increased cell volume and/or membrane stretch during gestation may be a critical factor in the initiation of labor. FIG. 7. COX-2 mRNA levels in hypotonically stressed and EGFtreated human amnion primary cultured cells. Cell were incubated for 1 h in 5.0 ml of the following media: isotonic EBSS (I); hypotonic EBSS (H); and EBSS plus EGF (10 ng/ml) (E). EBSS was made hypotonic by reducing the concentration of NaCl in EBSS. Primary human amnion cells were prepared and cultured as described under "Materials and Methods." Total RNA (5 g) was dot-blotted and processed for COX-2 mRNA as described under "Materials and Methods." An autoradiogram of the dot blot was then analyzed by grayscale scanning as described under "Materials and Methods." Each treatment represents the mean Ϯ S.D. of three replicate cultures.