Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Human amnionic WISH cells (Hayflick, 1961) have proven to be a robust in vitro model for the study of arachidonic acid release and PG formation. WISH cells have been shown to produce a variety of prostanoids in response to a wide range of stimuli (Harris et al., 1988). Production of PGE₂, in a concentration-dependent manner, was demonstrated in WISH cells with the cytokines TNF- and IL-1, the growth factors epidermal growth factor and transforming growth factor- and the phorbol esters phorbol dibutyrate and PMA (Harris et al., 1988). Subsequent investigations revealed that IL-1-induced PGE₂ formation in WISH cells is mediated through the expression of the inducible PGHS-2 (Albert et al., 1994) and a cytosolic PLA₂ (Xue et al., 1995; Xue et al., 1996). Recently, it was also observed that PGE₂ synthesis induced by IL-1 in WISH cells can be modulated by interferon and IL-4 at the PGHS-2 mRNA level (Harding et al., 1996). Because sites of inflammation contain an abundance of cytokines and growth factors (Dinarello, 1992), this implicates complex regulation of both autocrine cytokines and proinflammatory genes, such as PGHS-2, in such settings.

Several different signal transduction pathways have been reported to be involved in the induction and modulation of PGHS-2. Signals that ultimately trigger the accumulation of PGHS-2 mRNA have been shown to be initiated through tyrosine kinase, protein kinase A, protein kinase C, mitogen-activated protein kinase or Janus activated kinase-signal transduction activator of transcription (JAK-STAT) pathways, depending on the type of cell and stimulus studied (Blanco et al., 1995; Hamasaki and Eling, 1995; Rzymkiewicz et al., 1995; Herschman, 1996). Additionally, okadaic acid, a non-phorbol ester tumor promoter that acts through the inhibition of protein phosphatases 1 and 2A, has been shown to stimulate PGE₂ production in rat peritoneal macrophages (Ohuchi et al., 1989). Okadaic acid has also been shown to induce the synthesis of IL-1 (Sung and Walters, 1993) and TNF- (Sung et al., 1992) in human monocytes.

In the present study, we have studied the signal transduction mechanisms by which induction of PGHS-2 can occur within WISH cells. Additionally, we have examined the effect of okadaic acid on PGE₂ production and cytokine synthesis in these cells and have investigated whether cross-regulation of PGE₂ and cytokine production occurs in these cells. Our results indicate that, whereas several cytokines and tumor promoters induce PGHS-2 protein in WISH cells, they do so through distinct intracellular mechanisms. We have demonstrated for the first time that WISH cells are capable of producing the cytokines TNF- and IL-1 and that the signal transduction mechanisms involved in the induction of these cytokines are distinct from those involved in the induction of PGHS-2.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Reagents. The following supplies were obtained from the indicated sources: WISH cells, American Type Culture Collection (Rockville, MD); cell culture products and media, GIBCO (Grand Island, NY); recombinant human IL-1 and TNF-, UBI (Lake Placid, NY); IL-1ra and goat anti-human TNF- antibodies, R & D Systems (Minneapolis, MN); okadaic acid sodium salt, RBI (Natick, MA); okadaic acid 7,10,24,28-tetraacetate, LC Laboratories (Woburn, MA); staurosporin, BioMol (Plymouth Meeting, PA); o-phenylenediamine dihydrochloride, dexamethasone, indomethacin and PMA, Sigma Chemical Co. (St. Louis, MO); casein, BDH Laboratory Supplies (Poole, England); purified PGHS-1, PGHS-2 and rabbit polyclonal anti-human PGHS-2, Cayman Chemicals (Ann Arbor, MI); 10% polyacrylamide precast gels, nitrocellulose membrane, molecular weight standards and electrophoresis/transfer buffers and equipment, Novex (San Diego, CA); PGE₂ EIA reagents, PerSeptive Diagnostics (Cambridge, MA); streptavidin-horseradish peroxidase, enhanced chemiluminescence reagents and Hyperfilm, Amersham (Arlington Heights, IL); Immobilon 4 microtiter ELISA plates, Dynatech (Chantilly, VA); biotinylated goat anti-rabbit F(ab)₂ fragments, Biosource International (Camarillo, CA). The TNF-converting enzyme inhibitor GI 129471 and the PGHS-2 inhibitor NS-398 were synthesized at Abbott Laboratories (Abbott Park, IL). Rabbit polyclonal anti-PGHS-1 was a gift from Dr. David DeWitt, Michigan State University. Biotinylated anti-TNF- was purchased from The Binding Site (Burmingham, UK). The 1B15 cDNA probe was kindly provided by Dr. Lynn Matrisian, Vanderbilt University. All other chemicals were obtained through standard suppliers.

Cell activation. Human amnionic WISH cells (Hayflick, 1961) were maintained in Dulbecco's minimal essential medium supplemented with 10% fetal bovine serum and 1% antibiotics (final concentrations, 50 U/ml penicillin G sodium and 50 µg/ml streptomycin sulfate). Cells were passaged via trypsinization and seeded, at a density of 1 × 10⁵ cells/cm² of growth area, into 48-well culture dishes for activation studies, into six-well culture dishes for immunoblot analysis or into 162-cm² flasks for mRNA preparation. After 48 hr (90% confluence), the growth medium was decanted and the monolayers were washed twice with Gey's balanced salt solution. The cells were then treated with stimulants for specified times in Neuman-Tytell serumless medium containing 1% penicillin/streptomycin solution. For certain studies, dexamethasone, staurosporin, GI 129471 and NS-398 were added to the cultures at the time of treatment (final dimethylsulfoxide concentration, 0.1%). As indicated, IL-1ra was preincubated with the cultures for 1 hr before treatment with stimuli. After the appropriate incubation time at 37°C in a humidified CO₂ incubator, the conditioned medium was removed from the cultures and assayed for PGE₂ and/or TNF- content. Certain cultures were solubilized directly into Laemmli sample buffer and stored at 20°C for subsequent electrophoresis (Laemmli, 1970) and immunoblotting. The protein content of cell lysates was determined using the method of Bensadoun and Weinstein (1976), with bovine serum albumin as a standard. Each experiment was repeated at least three times with each data point in triplicate. Figures are representative data from individual experiments. Data analysis was performed using Student's t test.

PGE₂ assay. The PGE₂ content of the conditioned media was assayed using the PerSeptive Diagnostics EIA kit, according to the manufacturer's protocol. The anti-PGE₂ antibody has <3.5% cross-reactivity with PGA₁, PGA₂, PGB₁, PGB₂, PGF₁ and PGF₂. Sensitivity of this ELISA is 15 pg/ml, with a detection range of 0.1 to 50 ng/ml.

TNF- assay. The human TNF- content of the conditioned media was assayed using a sandwich ELISA developed at Abbott Laboratories. Immobilon 4 microtiter plates were coated with goat anti-human TNF- antibodies. TNF- standards or samples were added to each well and incubated for 2 hr at ambient temperature. The plates were washed, and biotinylated anti-human TNF- was added. After a 1-hr incubation, the plates were washed (four times) and streptavidin-labeled horseradish peroxidase was added to the wells. After 1 hr at ambient temperature, the plates were washed (four times) and o-phenylenediamine dihydrochloride substrate was added for 15 min. The reaction was stopped with 1 N H₂SO₄ and read on a Molecular Devices Thermomax microtiter plate reader at 480 nm. The anti-human TNF- antibody has no cross-reactivity with other cytokines and <5% cross-reactivity with murine TNF-. Sensitivity of this ELISA is 31 pg/ml, with a detection range of 31 to 2000 pg/ml.

SDS-polyacrylamide gel electrophoresis and immunoblot analysis. Samples (50 µg protein) of whole-cell homogenates in Laemmli buffer were subjected to electrophoresis on 1.5-mm-thick 10% polyacrylamide gels. The separated proteins were transferred to nitrocellulose filters using half-strength Towbin buffer (Towbin et al., 1979) with 20% (v/v) methanol. The filters were blocked for 1 hr at room temperature with 2.5% casein in 50 mM Tris-HCl (pH 7.6), 154 mM NaCl, 0.2 mM thimerosal, and then incubated with anti-PGHS antibody (1:1000 dilution) in blot buffer (50 mM Tris-HCl, 200 mM NaCl, 0.05% Tween-20, 1% casein, pH 7.5) overnight at 4°C. After three 5-min washes in TBS/Tween (50 mM Tris-HCl, 200 mM NaCl, pH 7.5, containing 0.05% Tween-20), the filters were incubated with biotinylated goat anti-rabbit F(ab)₂ fragments (1:2000 dilution) in blot buffer for 30 min at room temperature and then washed (3 × 5 min) in TBS/Tween. The filters were then incubated with streptavidin-horseradish peroxidase (1:5000 dilution) in blot buffer for 30 min at room temperature and washed (3 × 5 min) in TBS/Tween, and the immunoreactive bands were visualized with the enhanced chemiluminescence reagents on Hyperfilm (Amersham), according to the manufacturer's protocol.

Preparation of mRNA and Northern analysis. After treatment, the cells were washed once with Gey's balanced salt solution and harvested using trypsin-EDTA. The cells were resuspended in Gey's balanced salt solution, counted, pelleted and stored at 70°C until RNA isolation. Poly(A)⁺ RNA was isolated using a Mini RiboSep Ultra mRNA isolation kit, according to the manufacturer's instructions (Collaborative Biomedical Products, Bedford, MA), and stored as an ethanol precipitate at 70°C until gel electrophoresis.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Stimulus-induced PGE₂ production by WISH cells. WISH cells responded to multiple stimuli, such as PMA, IL-1, TNF- and okadaic acid, alone or in combination, to produce PGE₂ (fig. 1A). All of these stimuli culminated in the induction of PGHS-2 (fig. 1B). In agreement with previous data (Harris et al., 1988), PGE₂ production in response to the cytokines IL-1 and TNF- was found to be dose-dependent (fig. 2). Interestingly, the kinetics of PGE₂ production in WISH cells differed widely depending upon the stimulus used (fig. 1A). The combination of PMA and IL-1 yielded the most rapid induction of PGE₂, with measurable product at 1 hr. This was in contrast to PMA alone, which produced less PGE₂ than PMA and IL-1 and had a substantial lag phase (4-6 hr). Although the amount of PGE₂ induced by IL-1 alone was substantially less than that with PMA and IL-1, IL-1 alone stimulated the production of PGE₂ rapidly (within 2-4 hr). In contrast, okadaic acid appeared to require a significant amount of time (12-18 hr) before PGE₂ production became maximal. However, in all cases, the level of PGE₂ produced in response to a given agent or combination of agents in WISH cells was temporally correlated with PGHS-2 induction by the agent (fig. 1B). For example, PGHS-2 protein was induced as early as 1 to 2 hr in cells treated with the combination of PMA and IL-1, whereas such induction in response to okadaic acid did not occur until 12 hr after treatment. The levels of PGE₂ and PGHS-2 protein produced in response to either TNF- or IL-1 were significantly less than that reached by treatment with the combination of PMA and IL-1 (fig. 1). In contrast, levels of PGHS-1 protein remained unchanged in response to all stimuli (data not shown).

View larger version (37K): [in this window] [in a new window]   Fig. 1.   Effect of stimulus and treatment time on PGE2 production and PGHS-2 induction by WISH cells. A, WISH cells were treated, as noted, with 30 nM okadaic acid, 10 ng/ml PMA, 10 ng/ml IL-1, 10 ng/ml TNF- or combinations of PMA with IL-1 or TNF- for 1 to 24 hr, and the amount of PGE2 released into the conditioned medium was measured as described in "Materials and Methods." Values are the mean ± S.E.M. of three determinations. For IL-1, stimulation was considered significantly above background from 4 to 24 hr (P  .05). PMA was considered significantly different from 6 to 24 hr (P  .05). The combination of IL-1 and PMA was considered significant from 4 to 24 hr (P  .01). TNF- was considered significant from 18 to 24 hr (P  .06). Okadaic acid was considered significant from 12 to 24 hr (P  .05). B, immunoblot analysis was performed using a 1:1000 dilution of anti-human PGHS-2 IgG with the whole-cell homogenates (50 µg protein) from cultures treated from 1 to 24 hr as described in A. Lanes S1 and S2, purified PGHS-1 and PGHS-2 proteins run as standards. For the control group, lanes A, B and C contain the 8-, 18- and 24-hr IL-1 samples, respectively, as positive controls for the induction of PGHS-2; for all other groups, lanes A, B and C contain the 8-, 18- and 24-hr control samples, respectively, as negative controls for PGHS-2 induction. Exposure to film after enhanced chemiluminescence detection was for 20 sec for each of the blots. Levels of PGHS-1 protein were unchanged over time in all treatment groups (data not shown).

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Effect of stimulus dose on PGE2 production by WISH cells. WISH cells were treated for 18 hr, and the amount of PGE2 released into the conditioned medium was measured by EIA. Values are the mean ± S.E.M. of three determinations. A, IL-1. B, TNF-. *, significance above background of P  .05; **, significance of P  .01.

View larger version (54K): [in this window] [in a new window]   Fig. 3.   PGHS-2 mRNA induction in response to treatment with various stimuli. A, cells were treated with 30 nM okadaic acid (OKA), 10 ng/ml PMA, 10 ng/ml IL-1, 10 ng/ml TNF- or combinations of PMA with IL-1 or TNF- for 18 hr, and poly(A)+ mRNA was extracted from each culture and subjected to Northern analysis as described in "Materials and Methods." B, the graph depicts densitometric analysis of the treatment groups and normalization to 1B15 mRNA content.

Effect of IL-1ra on stimulus-induced PGE₂ production. One possible explanation for the lag periods observed in PGE₂ biosynthesis with certain stimuli, such as okadaic acid, could be that the induction of PGE₂ is secondary to the production of another cytokine, such as IL-1. Once induced, the IL-1 or other cytokine would then induce PGHS-2. To examine this possibility, we examined the effect of IL-1ra on PGE₂ production induced by a range of stimuli (fig. 4). As expected, pretreatment of the cells with a 1000-fold molar excess of IL-1ra (3 µg/ml) inhibited the IL-1 (3 ng/ml)-induced production of PGE₂ by 90%. Interestingly, IL-1ra pretreatment also inhibited PGE₂ production in response to 30 nM okadaic acid, PMA and TNF- (50-60%). IL-1ra pretreatment blocked PGE₂ production in response to the combination of PMA and IL-1 by 80%, but that induced by the combination of PMA and TNF- was reduced by only 20% (fig. 4). It therefore seems likely that at least some of the PGE₂ produced in response to okadaic acid, PMA and TNF- may be attributed to the induction of IL-1 by these agents, because IL-1ra has no known function other than to block signal transduction through the IL-1 receptor. We were unable to detect IL-1 in the conditioned medium from WISH cells treated for 18 hr either with PMA and TNF- or with okadaic acid (data not shown). It is possible that production of this cytokine is below the limits of detection (i.e., <10 pg/ml) of commercially available EIA kits.

View larger version (33K): [in this window] [in a new window]   Fig. 4.   Effect of the IL-1ra on PGE2 production by various stimuli in WISH cells. Cultures were preincubated for 1 hr with 3 µg/ml IL-1ra or with medium only before treatment for 18 hr with 30 nM okadaic acid (OKA), 10 ng/ml PMA, 3 ng/ml IL-1, 10 ng/ml TNF- or combinations of PMA and IL-1 (3 ng/ml) or TNF-. The amount of PGE2 released into the conditioned medium was measured by EIA. Values are the mean ± S.E.M. of three determinations. All stimuli were significantly greater than control (P  .05). *, significance of P  .05 for specific stimuli vs. IL-1ra treatment; **, significance of P  .01.

TNF- production by WISH cells. Because our results suggested that WISH cells are capable of producing IL-1, we next examined whether these cells had the capacity to produce TNF-. IL-1 and PMA by themselves failed to produce detectable levels of TNF- after 18 hr of treatment (data not shown). However, when the cells were treated with the combination of PMA and IL-1 or with okadaic acid, TNF- could be measured in the conditioned medium. Although the production of TNF- was rapid with the combination of PMA and IL-1 (i.e., within 2 hr after treatment), levels of maximal production were modest (200 pg/10⁶ cells; data not shown). In contrast, TNF- production in response to okadaic acid was delayed, with no appreciable production measured until 18 hr after treatment. However, the magnitude of TNF- produced in response to okadaic acid was much greater (15-fold) than that observed with the combination of PMA and IL-1, reaching 3 ng/10⁶ cells (fig. 5A). It therefore appears that the kinetics of TNF- production in response to either okadaic acid or the combination of PMA and IL-1 parallel those of the induction of PGHS-2 by these same agents.

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Effect of okadaic acid on WISH cells. A, effect of treatment time with 30 nM okadaic acid on TNF- production. , okadaic acid-treated cultures; , cultures treated with vehicle only. *, significance of P  .05 vs. control. B, dose-response relationship for okadaic acid effects on PGE2 and TNF- production after 18-hr treatments. The amount of PGE2 and TNF- released into the conditioned medium was measured by EIA. Values are the mean ± S.E.M. of three determinations. *, significance of P  .05; **, significance of P  .01 from control.

Effect of signal transduction inhibitors on PGE₂ and TNF- production. To examine whether the inductions of PGHS-2 and TNF- in WISH cells are regulated through a common set of intracellular signaling pathways, we investigated the effects of dexamethasone and staurosporin on the cellular responses to various stimuli. Dexamethasone (100 nM) completely inhibited the induction of PGE₂ in response to PMA, IL-1 and TNF-, whereas it was less effective in inhibiting PGE₂ production in response to okadaic acid or the combination of PMA with either IL-1 or TNF- (fig. 6A). Dexamethasone was also examined for its effect on levels of PGHS-2 protein induced by the various stimuli. Dexamethasone by itself did not cause the induction of PGHS-2 protein, whereas it completely blocked the induction of PGHS-2 in response to PMA, IL-1 and TNF-. In addition, levels of PGHS-2 protein induced by okadaic acid as well as by the combinations of PMA with either IL-1 or TNF- were markedly reduced (70-80%) by dexamethasone (fig. 6B). Thus, the effects of dexamethasone on the induction of PGHS-2 protein correlated directly with its effects on PGE₂ production (fig. 6B).

View larger version (35K): [in this window] [in a new window]   Fig. 6.   Effects of dexamethasone and staurosporin on the induction of PGE2, PGHS-2 and TNF- in WISH cells. Cells were treated for 18 hr with okadaic acid, PMA, IL-1 or TNF-, as described for figure 1, in the presence and absence of either 100 nM dexamethasone or 100 nM staurosporin. A, PGE2 was measured by EIA in conditioned medium from treated cultures. Values are the mean ± S.E.M. of three determinations. All stimuli were significantly greater than control (P  .01). *, significance of P  .05 for specific stimuli vs. dexamethasone or staurosporin treatment; **, significance of P  .01. B, immunoblot analysis was performed using a 1:1000 dilution of anti-human PGHS-2 IgG with the whole-cell homogenates (50 µg protein) from cultures treated as described for A. Lanes S1 and S2, purified PGHS-1 and PGHS-2 proteins run as standards. Lanes A, treatments in the absence of either dexamethasone or staurosporin. Lanes B, treatments in the presence of 100 nM dexamethasone. Lanes C, treatments in the presence of 100 nM staurosporin. In all three blots, lane group 1 contains the control treatments. Blot I contains PMA (lane group 2), IL-1 (lane group 3) and PMA and IL-1 (lane group 4). Blot II contains PMA (lane group 2), TNF- (lane group 3) and PMA and TNF- (lane group 4). Blot III contains okadaic acid (lane group 2), IL-1 (lane group 3) and PMA (lane group 4). Exposure to film after enhanced chemiluminescence detection was for 20 sec for each blot. C, TNF- was measured by EIA in conditioned medium from treated cultures. Values are the mean ± S.E.M. of three determinations. Okadaic acid treatment vs. control was considered significant at P  .01. **, significance of P  .01 vs. okadaic acid treatment only.

Effect of nonsteroidal anti-inflammatory drugs on okadaic acid-induced PGE₂ and TNF- production. To examine whether cross-regulation exists between the production of TNF- and PGE₂ in WISH cells, we tested a PGHS-2-selective nonsteroidal anti-inflammatory drug, NS-398 (Futaki et al., 1994), for its ability to inhibit okadaic acid-induced PGE₂ and TNF- production. NS-398 potently inhibited the production of PGE₂ induced by okadaic acid, with an IC₅₀ of 47 nM (95% confidence limits, 36-60 nM), whereas it failed to inhibit TNF- production by >20% at a concentration of 1 µM (fig. 7A). Conversely, the TNF--converting enzyme inhibitor GI 129471 (McGeehan et al., 1994) inhibited the production of TNF- with an IC₅₀ of 0.94 µM (95% confidence limits, 0.8-1.1 µM), whereas it failed to inhibit PGE₂ production by >20% at a 3 µM concentration (fig. 7B).

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Effects of anti-inflammatory compounds on okadaic acid-induced PGE2 and TNF- production in WISH cells. Cells were coincubated for 18 hr with 30 nM okadaic acid and various concentrations of compounds as indicated. PGE2 and TNF- were measured in the conditioned medium by EIA, and percent inhibition was calculated for the compounds. Control values for PGE2 and TNF- were 1540 ng/106 cells and 4350 pg/106 cells, respectively. Values are the mean ± S.E.M. of three experiments, each in triplicate (n = 9). A, NS-398 (PGHS-2 inhibitor); B, GI 129471 (TNF-converting enzyme inhibitor).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The human amnionic epithelial cells (WISH cells) have the potential to produce copious amounts of prostanoids (100-700 ng/10⁶ cells) in response to a wide variety of stimuli, including growth factors, phorbol esters and inflammatory cytokines such as IL-1 and TNF-. Table 1 summarizes our findings from observations made using alternative stimuli to induce proinflammatory mediators in WISH cells. Our studies have now clearly demonstrated that both the time of induction and the magnitude of PGE₂ produced by WISH cells can be regulated differentially, depending upon the stimulus used. The data also distinctly illustrate that the magnitude of PGE₂ induction by a given stimulus is tightly linked to levels of both PGHS-2 mRNA and protein in these cells. The combination of PMA and IL-1 produces a more than additive effect on PGHS-2 protein and subsequent PGE₂ production, compared with effects observed with these stimuli alone. Previous investigators have shown that IL-1 can serve to stabilize the mRNA induced by PMA (Ristimaki et al., 1994; Srivastava et al., 1994). Our finding that PGHS-2 mRNA was still abundant after 18 hr of treatment with the combination of PMA and IL-1, but not with these activators individually, is in agreement with those observations.

View this table: [in this window] [in a new window]   TABLE 1 Summary of inhibitor effects on PGE2 and TNF- production in WISH cells

We have demonstrated herein that okadaic acid induces both PGE₂ and TNF- production in WISH cells. Cultures of WISH cells treated with okadaic acid showed a longer lag phase (12-18 hr) to induce PGE₂ production than did those treated with PMA (2-4 hr), which correlated with the time required to induce PGHS-2 protein (fig. 1). Our observations are consistent with those of Ohuchi et al. (1989), who observed an 8- to 12-hr lag in PGE₂ synthesis in okadaic acid-treated rat peritoneal macrophages, whereas those treated with PMA produced PGE₂ within 2 to 4 hr. TNF- production also required a 12- to 18-hr exposure to okadaic acid (fig. 5A). Our studies suggest that the effects of okadaic acid on WISH cells can be specifically linked to its activity as an inhibitor of protein phosphatases 1 and 2A, for several reasons. Firstly, okadaic 7,10,24,28-tetraacetate, a structural analog of okadaic acid reported not to inhibit protein phosphatases, did not induce either TNF- or PGE₂ production in WISH cells at concentrations as high as 100 nM (maximum noncytotoxic dose for both okadaic acid and okadaic acid 7, 10, 24, 28-tetraacetate). Secondly, FK506, a known immunosuppressant that inhibits protein phosphatase 2B (calcineurin) (Fruman et al., 1992; Kincaid, 1995), was without effect on PGE₂ and TNF- production at concentrations up to 1 µM (K. I. Hulkower, E. R. Otis, R. L. Bell and K. B. Glaser, unpublished observations). Calyculin A, another reported inhibitor of protein phosphatases 1 and 2A that is structurally unrelated to okadaic acid (Ishihara et al., 1989), was able to induce PGE₂ but not TNF- production in WISH cells. However, that TNF- production is not induced by calyculin A may be explained by the cytotoxic effects of this compound observed at lower concentrations than with okadaic acid (K. I. Hulkower, E. R. Otis, R. L. Bell and K. B. Glaser, unpublished observations).

To understand the mechanism(s) involved in the long lag time after treatment of the cells with okadaic acid before the onset of PGE₂ production, we investigated whether the induction of PGE₂ could be secondary to the production of another cytokine by okadaic acid. IL-1ra (3 µg/ml) inhibited the production of PGE₂ in response to 30 nM okadaic acid by 50%. Similarly, we found that PGE₂ induced by either PMA or TNF- alone was also partially (60%) inhibited by IL-1ra. Although we were unable to detect IL-1 by conventional EIA in the conditioned medium of WISH cells after 18 hr of treatment with okadaic acid, it is possible that this cytokine is produced at a level below the detection limit of these types of assays or remains associated with the cell membrane. However, because the only currently known activity of IL-1ra is the antagonism of signal transduction through the type I IL-1 receptor (Dripps et al., 1991; Granowitz et al., 1992a,b), it is likely that at least some of the PGE₂ produced in response to okadaic acid, PMA or TNF- may be attributed to the production of IL-1 by these agents. Even minute amounts of IL-1 have been shown to act synergistically with other activators to cause heightened cellular responses, such as increased PGE₂ production (Dinarello, 1996). It is therefore possible that the small amount of IL-1 produced in response to okadaic acid could vastly augment PGE₂ production, which is also induced by this agent. The autocrine effects of inflammatory cytokines have been well described, e.g., TNF has been shown previously to induce the production of IL-1 in human fibroblasts (Le et al., 1987). More specifically, in recent studies with human monocytes, it has been reported that lipopolysaccharide induction of PGHS-2 is inhibited by pretreatment with IL-1ra, thereby implicating the induction of IL-1 by lipopolysaccharide in these cells as being responsible for driving PGHS-2 induction and PGE₂ production (Glaser and Lock, 1995). It is also plausible that an abundance of IL-1 receptors may be present on WISH cells and need to be saturated with IL-1ra before inhibition of PGE₂ production is observed.

Although we can observe the production of both PGE₂ and TNF- in WISH cells in response to okadaic acid, with similar time courses, our studies suggest that distinct intracellular processes may govern the production of each of these mediators. Not all agents that stimulate the production of PGE₂ in WISH cells are able to induce TNF-. Although we have shown that PGE₂ production in response to okadaic acid is partially mediated by the production of IL-1 in these cells (inhibitable by IL-1ra), it should be noted that this is not the case for TNF- production, because it is completely unaffected by IL-1ra. NS-398, a PGHS-2-specific inhibitor, was able to completely block PGE₂ production induced by okadaic acid but had no effect on concomitant TNF- production, implying that PGE₂ is not required for TNF production. Likewise, the TNF-converting enzyme inhibitor GI 129471 (McGeehan et al., 1994) was able to block the release of TNF into the medium of okadaic acid-treated cells, whereas it was without effect on PGE₂ production. Also indicative of separate routes of signal transduction for the production of PGE₂ and TNF- is our finding that dexamethasone, which inhibits PGE₂ production by okadaic acid, has no effect on TNF- secretion in the same cultures. This indicates that induction of PGHS-2 by okadaic acid in WISH cells occurs under transcriptional control of glucocorticoid-sensitive response elements, whereas TNF- production does not. However, because both PGE₂ and TNF- production are inhibited to some extent by staurosporin, it is likely that serine/threonine kinases are involved in both processes.

The lack of effect of dexamethasone on TNF- production in WISH cells may be explained by the mechanisms by which okadaic acid and dexamethasone regulate IB and NFB (Menon et al., 1995; Verma et al., 1995). The inhibition of phosphatase activity causes IB to remain phosphorylated and thus become vulnerable to degradation by a proteasome-dependent pathway, which thereby allows gene activation via translocation of the nuclear transcription factor NFB (Palombella et al., 1994). Okadaic acid has been shown to inhibit the phosphatase activity that removes the phosphate from IB, leading to its degradation and subsequent activation of NFB-dependent TNF- production in U937 cells (Menon et al., 1995). Recently, one possible mechanism of steroid action was proposed that suggested that inhibition of NFB by steroids was due to increased de novo synthesis of IB (Auphan et al., 1995; Scheinman et al., 1995). Because okadaic acid allows IB to be phosphorylated without subsequent dephosphorylation, the newly synthesized IB is also degraded and thus does not inhibit TNF- production in WISH cells. Alternatively, okadaic acid could stimulate both NFB and AP-1 (jun/fos) transcriptional machinery in WISH cells. Different genes, e.g., PGHS-2 and TNF-, may use these different activators independently, and thus one may be inhibitable by dexamethasone and the other not. Further experiments on the regulation of IB and jun/fos are in progress, to make this distinction between steroid mechanisms in WISH cells.

It is noteworthy that such a diverse group of agents that signal through very different mechanisms are unified in their ability to induce PGHS-2 and cause the production of PGE₂ in WISH cells. This is likely to occur at a convergent point distal to the initial activation of their separate pathways, at the level of phosphorylation events. PMA is noted for its ability to stimulate the protein kinase C pathway (Nishizuka, 1984) and increase serine/threonine phosphorylation, whereas the actions of okadaic acid are mediated by its inhibition of protein phosphatases 1 and 2A (Bialojan and Takai, 1988), indirectly elevating levels of serine/threonine and tyrosine phosphorylation. A novel IL-1-responsive serine/threonine kinase has recently been purified and cloned; it is claimed to be part of a key pathway for the signal transduction process of this cytokine, leading to the activation of NF-B (Cao et al., 1996). TNF- has been described to work through a ceramide-activated protein kinase pathway, leading to the activation of NF-B (Schutze et al., 1992). It is noteworthy that staurosporin, an inhibitor of protein kinases, induces PGE₂ production by WISH cells independently of PGHS-2 induction. Staurosporin has previously been reported to induce PGE₂ production in rabbit articular chondrocytes (Hulkower et al., 1991) and human decidual cells (Cole et al., 1995). A possible explanation for this finding is that staurosporin may modulate levels of arachidonic acid by regulating the activity of a PLA₂ in these cells, thereby operating at a step proximal to cyclooxygenase enzyme activity. It is therefore likely that the formation of PGE₂ through staurosporin treatment in WISH cells may be primarily mediated through the release of arachidonic acid via a PLA₂ and only secondarily by a cyclooxygenase step. It is not unusual for PGE₂ production in cells to be regulated at the level of both PLA₂ and cyclooxygenase activities, because this phenomenon has been observed previously in rheumatoid synoviocytes (Hulkower et al., 1994), as well as in WISH cells (Xue et al., 1996).

In WISH cells, a diverse group of stimuli acting through different signaling pathways induce the de novo synthesis of PGHS-2 (fig. 8). The ability of these cells to make TNF- in response to various stimuli may also be part of their role in parturition (Terranova et al., 1995). Indeed, the normal physiological process of parturition shares many features with the inflammatory process, including production by the decidual cells of prostanoids and cytokines, including IL-1 and TNF- (MacDonald et al., 1991; Romero et al., 1991). As a result of the production of these bioactive mediators, WISH cells make an excellent model to study the regulatory process of proinflammatory mediator release.

View larger version (53K): [in this window] [in a new window]   Fig. 8.   Proposed scheme of PGE2 and cytokine production by proinflammatory stimuli in WISH cells. P-Pase, protein phosphatase; PKC, protein kinase C; AA, arachidonic acid; DEX, dexamethasone; OKA, okadaic acid; TCE, TNF-converting enzyme; NSAIDs, nonsteroidal anti-inflammatory drugs; Ser/Thr, serine/threonine.