8-Iso-Prostaglandin F_2α Reduces Trophoblast Invasion and Matrix Metalloproteinase Activity

Materials

Reagents, when not otherwise specified, were purchased from Sigma Chemical Co.

Cell Cultures

The choriocarcinoma cell line JAR (ATTC HTB-144) was grown as monolayers in 75-cm² plastic flasks (Costar) at 37°C in 95% air and 5% CO₂ in RPMI-1640 supplemented with 10% heat-inactivated FBS, l-glutamine (2 mmol/L), penicillin (50 IU/mL), and streptomycin (50 μg/mL) (Bio Whittaker). For subculture, the cells were detached with Accutase (Innovative Cell Technologies).

Transwell Invasion Assay

The effect of 8-iso-PGF_2α on JAR cell invasiveness was examined by a Transwell invasion assay with MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide) from Fluka (Buchs) to stain viable cells.¹⁵ To the upper wells of the invasion system (Transwell plates, 8-μm pore size, from Costar), one of the following test substances was added: 0.009% ethanol (control), 8-iso-PGF₂ (1 pmol/L-100 μmol/L) (Cayman Chemical), 1,10-phenanthroline (50 ng/mL), a metalloproteinase inhibitor,¹⁶ or 8-Bromo-cAMP (8-Br-cAMP) (1.5 mmol/L), a nonhydrolyzable cAMP analogue that induces differentiation of cytotrophoblasts into syncytiotrophoblasts.¹⁷ The invasion was calculated as filter absorbance in percentage of control. The absorbance reflects the number of trophoblasts that invaded the filter.

Transwell Invasion With [³H]Thymidine-Labeled JAR Cells

The trophoblast invasion was also estimated with the use of JAR cells labeled with [6-³H]thymidine (du Pont-NEN), as described by Graham et al,¹⁸ omitting Matrigel coating of the Transwell filters. Subconfluent cultures of JAR cells were labeled for 72 hours with 10 μCi/mL [³H]thymidine in growth medium and resuspended in serum-free RPMI-1640. Then, 1×10⁵ cells in 250 μL RPMI-1640 were added in quadruplicate to the upper wells of the Transwell system, and 500 μL growth medium was added to the lower wells. After 24-hour incubation with either 0.009% ethanol (control), 8-iso-PGF_2α (100 nmol/L or 10 μmol/L), 8-Br-cAMP (3 mmol/L), or 1,10-phenanthroline (100 ng/mL) in the upper wells, quantification of invaded cells was done as for the colorimetric method, except that the radioactivity was counted in a Packard TRICARB 1900 Scintillation Counter.

[³H]Leucine and [³H]Thymidine Incorporation

The JAR cell viability after 8-iso-PGF_2α exposure was examined by incorporation of dl-[4,5-³H]leucine (Amersham) into cell-associated proteins to measure protein synthesis and by incorporation of [6-³H]thymidine (du Pont-NEN) into DNA to analyze DNA synthesis, as a measure of cell proliferation. At 70% confluence, the cells were preincubated for 22 hours with growth medium supplemented with either 8-iso-PGF_2α (1 or 10 μmol/L), 8-Br-cAMP (1.5 mmol/L), or 1,10-phenanthroline (100 ng/mL). The control cells were exposed to 0.009% ethanol. The cells were then incubated for additional 2 hours in fresh growth medium supplied with 5 μCi/mL [³H]leucine or 5 μCi/mL [³H]thymidine. The cells were harvested and analyzed as described previously.¹⁹

Type IV Collagenase Activity

A modified procedure of Emonard et al²⁰ was performed to investigate the activity of type IV collagen–degrading enzymes in the conditioned medium (CM) from JAR cells. Briefly, 1×10⁵ JAR cells were seeded in RPMI-1640 into a 24-well plate precoated with 300 μL of an extracellular matrix gel diluted 1:1 (vol/vol) in DMEM (Bio Whittaker). The cultures received either 0.009% ethanol, 10 μmol/L 8-iso-PGF_2α, or 100 ng/mL 1,10-phenanthroline. The experiments were performed in quadruplicate in 2 separate experiments. Specific type IV collagenolytic activity was calculated as total type IV enzymatic activity of the CM from 1×10⁵ cells minus the EDTA-insensitive activity.

Gelatin Zymography of CM

Gelatinase activity was detected in serum-free CM from JAR cells grown on Matrigel-coated Transwell plates (Biomatrix EHS from Boehringer Ingelheim, diluted 1:1 [vol/vol] in RPMI-1640), modified from Sharma et al.¹⁶ Either vehicle (0.009% ethanol), 8-iso-PGF_2α (10 μmol/L), 8-Br-cAMP (3 mmol/L), or 1,10-phenanthroline (100 ng/mL) was added to the upper wells; and the cells were incubated for 72 hours. From the CM above the filters, 20 μg protein was incubated at room temperature for 10 minutes with nonreducing loading buffer containing 4% SDS (wt/vol), then subjected to electrophoresis on SDS-polyacrylamide gels containing 0.1% (wt/vol) gelatin (Novex). Both human MMP-2 and MMP-9 standards (Chemicon) were run to ensure identification of MMP activity. After gel wash with 2.5% Triton X-100 (wt/vol) and incubation in developing buffer for 40 hours, the gels were fixed, then stained with 0.1% Coomassie blue R-250 (Biorad) and destained in water. The gels were scanned by a laser scanner densitometer (Molecular Dynamics Inc) and quantified with the use of the software Image Quant.

Western Immunoblotting

Western blotting was performed as described elsewhere²¹ on JAR cells incubated in growth medium supplemented with either 0.009% ethanol (control), 8-iso-PGF_2α (10 μmol/L), or 8-Br-cAMP (3 mmol/L) for 24 hours. The filters were incubated with monoclonal anti-human MMP-9 or MMP-2 (R & D Systems) or monoclonal anti-human TIMP-2 (Calbiochem). Proteins were detected by enhanced chemiluminescence with horseradish peroxidase–labeled anti-mouse IgG (Vector Laboratories).

Northern Blot Analysis

mRNA isolation from JAR cells incubated in growth medium supplemented with either 0.009% ethanol (control), 8-iso-PGF_2α (10 μmol/L), or 8-Br-cAMP (3 mmol/L) and Northern blotting was performed as previously described.¹⁹ The TIMP-2, MMP-2, and MMP-9 cDNAs were a gift from Drs Andrew Baker and Gill Murphy.^{22 23} mRNA levels were quantified by laser densitometry and corrected to human β-actin (ATTC).

Statistical Analysis

The results are presented as mean value±SEM, and the differences between control and treatment groups were tested with the Mann-Whitney U test. A level of P<0.05 was considered statistically significant.

Transwell Studies

The MTT colorimetric Transwell assay showed a statistically significant reduction in trophoblast invasion into the filters (Table 1⇓) when the cells were incubated for 24 hours with 8-iso-PGF_2α in concentrations from 100 pmol/L (11% reduction of filters relative to control, P=0.04) to 100 μmol/L (16% reduction, P=0.01). JAR cells incubated with 10 μmol/L 8-iso-PGF_2α showed a 20% reduction in filter invasion compared with control cells (P=0.001). No statistically significant effect on the invasion was found at 8-iso-PGF_2α concentrations <100 pmol/L (10 pmol/L: 7% reduction, P=0.07)). As expected, 1.5 mmol/L 8-Br-cAMP resulted in a 17% reduction of trophoblast invasion into the filters (n=20 filters; P<0.001). Similarly, 50 ng/mL 1,10-phenanthroline showed a 22% reduction in JAR cell invasion into filters compared with control (n=9; P=0.001). Inhibition of trophoblast and JAR cell invasion by 1,10-phenanthroline has previously been demonstrated.^{16 24} In our study, 1,10-phenanthroline was used as a control for reduced invasion by reducing the MMP activity.

We also demonstrated reduced trophoblast invasion into Transwell filters by using [³H]thymidine-radiolabeled JAR cells (Figure 1⇓). Incubation with 10 μmol/L 8-iso-PGF_2α reduced the filter invasion by 26.8% (P=0.009, n=20), whereas incubation with 100 nmol/L 8-iso-PGF_2α resulted in a reduced invasion that was not statistically significant (7.5% reduction, P=0.5, n=22). There was a 20% reduction of trophoblast invasion when the cells were subjected to 3 mmol/L 8-Br-cAMP (P=0.047, n=17), whereas the 11.6% reduction of invasion after exposure to 100 ng/mL 1,10-phenanthroline did not reach statistical significance (P=0.4, n=12). The cotton swabs, representing JAR cells that had not invaded the filters but were adherent to the filters, showed a statistically significant rise in radioactivity in wells subject to 100 nmol/L and 10 μmol/L 8-iso-PGF_2α (not shown), reflecting that fewer JAR cells had invaded the filters.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Effects of 100 nmol/L 8-iso-PGF_2α (gray filled bar), 10 μmol/L 8-iso-PGF_2α (hatched bar), 3 mmol/L 8-Br-cAMP (open bar), and 100 ng/mL 1,10-phenanthroline (double hatched bar) on invasion of [³H]thymidine-labeled JAR cells into Transwell filters after 24-hour incubation. Control cells (black filled bar) were given ethanol (0.009%). Cells were radiolabeled and invasion was measured as described in Methods section. Results are given as percentage of control filters; each bar graph represents mean value±SEM from 12 to 27 filters, performed in 8 separate experiments. *P<0.05.

[³H]Leucine and [³H]Thymidine Incorporation

We found no indications of the 8-iso-PGF_2α–mediated reduction in JAR invasion being due to toxicity because neither [³H]thymidine incorporation into cell-associated DNA nor [³H]leucine incorporation into cell-associated proteins were affected by 1 or 10 μmol/L 8-iso-PGF_2α incubation (Table 2⇓). However, 100 μmol/L 8-iso-PGF_2α did significantly reduce [³H]thymidine incorporation by 30% (P<0.001), whereas the [³H]leucine incorporation was not reduced. This indicates that this concentration of 8-iso-PGF_2α is not toxic to the JAR cells but may impair cell proliferation. This could be the case if there was a differentiation toward syncytiotrophoblast (multinucleated, fused trophoblasts), which would result in an overall reduced cellular invasion. The concentrations ≤10 μmol/L of 8-iso-PGF_2α did not result in reduced JAR cell proliferation. 8-Br-cAMP (1.5 mmol/L) reduced the mean cell-associated incorporation of radiolabeled thymidine to 68.4% of control wells (n=27; SEM=3.4, P<0.001), whereas the [³H]leucine incorporation was not affected. 1,10-Phenanthroline (100 ng/mL) similarly reduced the [³H]thymidine incorporation to 83.3% of control (n=19, SEM=1.8, P<0.001), whereas the leucine incorporation was not reduced.

Type IV Collagenase Activity

The specific type IV collagenolytic activity in the CM of JAR cells incubated for 24 hours with 10 μmol/L 8-iso-PGF_2α was statistically significantly reduced by 30% (n=8, SEM=11, P=0.04). There was also a 50% reduction of the specific type IV collagenolytic activity in the CM of JAR cells incubated with 100 ng/mL 1,10-phenanthroline as compared with the CM of the control cells (n=8, SEM=4).

Zymography

In general, the MMP activity was higher in the CM from the upper wells compared with the CM from the lower wells of the Transwell assay, and the results presented are from the CM of the upper wells. Several lytic bands, seen as clear areas of lysis in the uniformly stained blue gel, were observed in the zymograms from the CM of the upper wells. The major gelatinolytic bands were activated MMP-2 and MMP-9 (both identified with protein weight markers, mixed MMPs, and a selectively activated MMP-2 marker). The APMA-activated MMP-2 marker comigrated with the MMP-2 band from the mixed MMPs marker, indicating that the SDS that was added to the samples activated the MMPs. MMP-2 was the predominant of these 2 enzymes. Incubation with 10 μmol/L 8-iso-PGF_2α for 72 hours resulted in a reduction in both MMP-2 and MMP-9 activity (42% and 33% reduction, respectively, P=0.003 and P=0.019) (Figure 2⇓). As expected, the MMP inhibitor 1,10-phenanthroline (100 ng/mL) reduced MMP activity compared with control in the CM, both for MMP-2 and MMP-9 activity (36% and 32% reduction, respectively). Incubation with 8-Br-cAMP (3 mmol/L) resulted in a 2-fold–augmented MMP-2 activity (P=0.006), whereas the MMP-9 gelatinolytic activity was unaltered. Identical gels were incubated in parallel in the presence of EDTA (10 mmol/L), which is a metal chelator. The disappearance of lytic bands in these gels after incubation with EDTA confirmed that the gelatinolytic activity observed was metal dependent, which is characteristic of MMPs.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Zymographic analyses of gelatinolytic activity of MMP-9 (left) and MMP-2 (right) in CM from upper wells of Transwell system of JAR cells treated for 72 hours with 10 μmol/L 8-iso-PGF_2α (hatched bars), 100 ng/mL 1,10-phenanthroline (double-hatched bars), or 1.5 mmol/L 8-Br-cAMP (open bars). Control cells (filled bars) were given ethanol (0.009%). Twenty micrograms of protein was applied to gel from CM of upper wells as described in Methods section. Results are given as percentage of control; each bar graph (bottom) represents mean value±SEM of 4 experiments, each experiment consisting of CM from 4 separate wells. Representative zymographies from 1 experiment are presented at top. *P<0.05.

Western Blot

Figure 3⇓ demonstrates a 31% reduction in protein level of MMP-9 and a 42% reduction in protein level of MMP-2 when JAR cells were subjected to 8-iso-PGF_2α (10 μmol/L) compared with controls (P=0.003 and P=0.019, respectively). Incubation with 3 mmol/L 8-Br-cAMP resulted in a significant 1.8-fold augmentation of MMP-9 and a 1.6-fold augmentation of MMP-2 protein levels compared with controls (P=0.004 and P=0.028). To assess whether an elevated protein level of TIMP-2 could be involved in a downregulation of MMP activity, we investigated TIMP-2 by Western blots (Figure 4⇓). We found a statistically significant 1.5-fold elevation of TIMP-2 protein level in JAR cells subjected to 10 μmol/L 8-iso-PGF_2α (P=0.014). On the other hand, incubation with 3 mmol/L 8-Br-cAMP did not alter TIMP-2 protein levels (P=0.2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Protein expression of MMP-9 (left) and MMP-2 (right) in JAR cells treated with vehicle (0.009% ethanol) (filled bars), 10 μmol/L 8-iso-PGF_2α (hatched bars), or 3.0 mmol/L 8-Br-cAMP (open bars) for 24 hours. Cell proteins (20 μg) were separated by reducing SDS-PAGE, electrotransferred, and incubated with either monoclonal anti-human MMP-9 or anti-human MMP-2 followed by horseradish peroxidase–labeled anti-mouse IgG and detected by enhanced chemiluminescence, as described in Methods section. Results are given as percentage of control; each bar graph (bottom) represents mean value±SEM of 3 experiments, each experiment was performed in duplicate. Representative Western immunoblots from 1 experiment are presented at top. *P<0.05.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Protein expression of TIMP-2 in JAR cells treated with 10 μmol/L 8-iso-PGF_2α (hatched bar) or 3.0 mmol/L 8-Br-cAMP (open bar) for 24 hours. Control cells (filled bar) were given ethanol (0.009%). Experiments were performed and data are presented as described in Figure 3⇑. Representative Western immunoblot exposure from 1 experiment is presented at top. *P<0.05.

Expression of mRNA

Figure 5⇓ shows unaltered mRNA signal density of MMP-2 in JAR cells subjected to 10 μmol/L 8-iso-PGF_2α (incubation time up to 24 hours). On the other hand, the signal density of JAR cells exposed to 3 mmol/L 8-Br-cAMP shows a time-dependent transcriptional induction of MMP-2, being 1.5-fold induced at 2 hours and 6-fold induced at 24 hours, compared with control cells (both P=0.014). Similar results were obtained for MMP-9, showing unaltered mRNA level after incubation of JAR cells with 10 μmol/L 8-iso-PGF_2α up to 24 hours. Incubation with 3 mmol/L 8-Br-cAMP gave a time-dependent induction of MMP-9 mRNA, being 1.5-fold induced compared with control at 2 hours and 8-fold induced after 24-hour incubation (both P=0.014). We also investigated whether there was an elevated transcription of the TIMP-2 gene after JAR cell incubation with 8-iso-PGF_2α (10 μmol/L). Analogous to the study by Graham et al,¹⁸ our study shows that JAR trophoblasts exhibit the 1.0-kb as well as the 3.5-kb TIMP-2 transcript. There were no differences between the signal densities of the 1.0- and 3.5-kb mRNAs relative to β-actin. Therefore, only the results from the 3.5-kb transcript is presented in Figure 6⇓. The TIMP-2 mRNA level was unaltered after incubation of JAR cells with either 10 μmol/L 8-iso-PGF_2α or 3 mmol/L 8-Br-cAMP up to 24 hours.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

mRNA signal density of MMP-9 and MMP-2 in JAR cells treated up to 24 hours with either 10 μmol/L 8-iso-PGF_2α (hatched bars) or 3.0 mmol/L 8-Br-cAMP (open bars). Control cells (filled bars) were given ethanol (0.009%). Cells were harvested and Northern blot analysis was performed as described in Methods section. Results are given as percentage of control; each bar graph (bottom) represents mean value±SEM of 4 different experiments, each experiment consisting of cells pooled from 2 separate culture wells. Representative autoradiograms from 1 experiment are presented at top. *P<0.05.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

mRNA signal density of TIMP-2 in JAR cells treated up to 24 hours with either 10 μmol/L 8-iso-PGF_2α (hatched bars) or 3.0 mmol/L 8-Br cAMP (open bars). Control cells (filled bars) were given ethanol (0.009%). Experiments were performed and data are presented as described in Figure 5⇑. Representative autoradiograms from 1 experiment are presented at top.

Statistical Analysis

When statistical analyses were performed comparing more than 2 groups, we used the Bonferroni method to control the overall type I error at a level ≤5% of statistical significance.²⁵ Where Mann-Whitney U tests were <0.05, the Bonferroni method still resulted in a corrected probability value <0.05 (not shown). An exception to this is the reduced invasion of 20% (Figure 1⇑) in JAR cells incubated with 3 mmol/L 8-Br-cAMP compared with control cells incubated with [³H]thymidine-labeled JAR cells. This difference does not reach statistical significance with the use of the Bonferroni method (corrected P=0.188). In addition, the effect of 3 mmol/L 8-Br-cAMP on augmenting the MMP-2 protein levels in Western blots with 60% (Figure 3⇑) compared with control cells is of borderline statistical significance when corrected for multiple comparisons (corrected P=0.056). These corrections do not affect the results regarding the effects of 8-iso-PGF_2α on invasion of JAR cells and possible MMP mechanisms.