Objective
Inflammation/infection and abruption are leading causes of preterm premature rupture of the membranes. Recently, we identified granulocyte-macrophage colony-stimulating factor (GM-CSF) as a critical mediator of both tumor necrosis factor-α– (TNF; modeling inflammation) and thrombin-induced (modeling abruption) weakening of the fetal membranes. We found that (1) TNF and thrombin both induced GM-CSF in the choriodecidua, (2) blockade of GM-CSF action with neutralizing antibodies inhibited both TNF- and thrombin-induced fetal membrane weakening, and (3) GM-CSF alone induced fetal membrane weakening. GM-CSF is thus part of an overlap of the inflammation and abruption-induced fetal membrane weakening pathways. The effects of progesterone analogs on the pathways by which fetal membranes are weakened have not been investigated. We examined the effects of progesterone, medroxyprogesterone acetate (MPA) and 17α-hydroxyprogesterone (HP) on TNF- and thrombin-induced fetal membrane weakening.
Study Design
Full-thickness fetal membranes from uncomplicated term repeat cesarean deliveries were mounted in Transwell inserts in Minimum Essential Medium alpha and incubated at 37°C in 5% CO 2 . The choriodecidua side of the fetal membrane fragments were preincubated with progesterone, MPA, HP, or vehicle for 24 hours. Fetal membranes were then exposed to TNF, thrombin, or GM-CSF on the choriodecidua side for an additional 48 hours. The fetal membrane tissues were then strength tested, and medium from the choriodecidua and amnion compartments was assayed for GM-CSF content.
Results
TNF and thrombin both weakened fetal membranes and elevated media GM-CSF levels on the choriodecidua side of the fetal membrane. Pretreatment with progesterone, MPA, or HP inhibited both TNF- and thrombin-induced fetal membrane weakening and also inhibited the induced increase in GM-CSF. GM-CSF decreased fetal membrane rupture strength by 68%, which was inhibited by progestogen pretreatment with a potency order: progesterone <MPA <HP.
Conclusion
Progestogen pretreatment blocks TNF- and thrombin-induced fetal membrane weakening by inhibiting both the production and action of GM-CSF. These findings are consistent with the administration of progestogens in the prevention of preterm premature rupture of the membranes.
See related editorial, page 447
Preterm premature rupture of the fetal membranes (PPROM) and resultant premature birth is a major cause of infant morbidity and death. Infection/inflammation with cytokine production and decidual bleeding/abruption with thrombin production, respectively, increase the risk of PPROM; however, the mechanisms by which these conditions affect PPROM remain obscure.
Research into the physiologic condition of PPROM is limited because animal models fail to mimic the human condition and because cell culture studies do not address adequately the major tissue changes that are associated with fetal membrane weakening and rupture. To address these shortcomings, we developed an in vitro explant model system in which human fetal membrane rupture strength, which is the major parameter of clinical interest, and concomitant (and presumably causative) biochemical changes within the fetal membrane can be measured accurately. Using this system, we found that cytokines such as tumor necrosis factor-α (TNF) and interleukin 1ß, which induce tissue level inflammation, and thrombin, which causes changes that are associated with abruption and bleeding, markedly weaken full-thickness fetal membrane. Concomitant with weakening, these agents induce biochemical and histologic tissue changes that mimic those seen in the physiologic weak zone (fetal membrane rupture initiation site) in the fetal membrane region overlying the cervix.
Studies of the pathways of TNF/interleukin-1ß– and thrombin-induced fetal membrane weakening have demonstrated that the initial tissue and cellular targets of these agents are in the choriodecidua component of the fetal membrane rather than the amnion. This is consistent with the sources of inflammation or bleeding that originate from the maternal side of the fetal membrane. Using an enhanced version of the model system that mimics the directional aspect of the fetal membrane weakening pathways, we recently demonstrated that TNF or thrombin, when applied to only the choriodecidua side of fetal membrane, caused weakening in the same manner as when both sides of the fetal membrane were exposed. We also identified granulocyte-macrophage colony-stimulating factor (GM-CSF) as a critical intermediate in the fetal membrane weakening pathways that are induced by both TNF and thrombin. As documented in this previous report, GM-CSF is induced in the choriodecidua by TNF and thrombin in a concentration-dependent manner that is concomitant with TNF- and thrombin-induced fetal membrane weakening. Importantly, GM-CSF alone also causes dose-dependent fetal membrane weakening and blockade of GM-CSF with neutralizing antibody prevents fetal membrane weakening by TNF or thrombin. These observations suggest that GM-CSF mediates fetal membrane weakening that is induced by both inflammation and bleeding. It is therefore possible to reduce the fetal membrane weakening process into those events that are involved in the generation of GM-CSF and those events caused by the action of GM-CSF. Similarly, inhibitors of the weakening process can now be categorized as acting to prevent GM-CSF generation, acting to inhibit GM-CSF action, or both. This categorization facilitates the identification of the point of action of specific inhibitors and thus provides information that may allow their use in inhibition of PPROM.
Currently, progestogens are the only agents recommended by the American College of Obstetricians and Gynecologists for the prevention of preterm delivery. Weekly intramuscular administration of 17 (OH) progesterone-caproate or daily vaginal administration of progesterone have been recommended for the prevention of recurrent preterm birth or for short cervical length, respectively. However, there is no direct evidence that progestogens inhibit fetal membrane weakening. The cellular targets for progesterone action in fetal membrane are thought to be mainly decidual cells that express the nuclear progesterone receptor (nPR) isoforms (PR-A and PR-B). In a previous study, we found that the progesterone analog medroxyprogesterone acetate (MPA) inhibited TNF-induced GM-CSF induction in decidual cells. It thus seemed appropriate to hypothesize that progesterone would inhibit fetal membrane weakening by blocking GM-CSF production. However, GM-CSF is proposed to act, in part, by the activation of resident decidual mononuclear cells, which express membrane progesterone receptors. Thus, it is also possible that progesterone could block fetal membrane weakening by inhibiting GM-CSF action. The studies herein were undertaken to determine whether progestogens inhibit TNF- and thrombin-induced fetal membrane weakening.
Materials and Methods
Materials
Humankine TNF, Humankine GM-CSF (both produced in HEK 293 cells), progesterone, MPA, 17α-hydroxyprogesterone (HP), and other miscellaneous reagents, unless otherwise stated, were obtained from Sigma-Aldrich, St. Louis, MO. Thrombin (from bovine plasma, 1500NIH U/mg protein) was obtained from Thermo Fisher Scientific, Pittsburgh, PA. RU486 (mifepristone) was obtained from Cayman Chemical Co, Ann Arbor, MI.
Fetal membrane collection and preparation
Full-thickness fetal membrane fragments from term uncomplicated repeat cesarean deliveries were collected after patient consent and approval by MetroHealth Medical Center’s Institutional Review Board (# IRB10-00861; Cleveland, OH). None of the patients had a history of preterm birth. Fetal membrane tissue was taken from areas remote from the weak zone region, washed in 2 × 250–mL changes of phosphate-buffered saline solution (pH 7.2) and mounted, choriodecidua side down, in 24-mm Transwell (Costar, Corning, NY) inserts that were secured with an O-ring to separate the choriodecidua and amnion sides. The inserts were placed in 6-well culture plates and 2 mL MEM+ (Minimum Essential Medium alpha with Earle’s salts, supplemented with 1 mmol/L L-glutamine, 2.24 g/L sodium bicarbonate [Mediatech, Manassas, VA], 10 ml/L Antibiotic-Antimycotic (Sigma Chemical Co, St. Louis, MO), and 50 mg/L gentamicin sulfate) was added to the upper (amnion side) and the lower (choriodecidua side) chambers. Progestogens (progesterone, HP, MPA [each 10 –7 mmol]) or vehicle (0.01% ethanol final concentration) were added to the choriodecidua side 24 hours before the addition of weakening agents. Where indicated, RU486 (10 –8 mmol) was also added to the choriodecidua side medium 1 hour before the addition of progestogens. Weakening agents (TNF [20 ng/mL], thrombin [10 U/mL], GM-CSF [200 ng/mL]) were then added to the choriodecidua side, and the cultures were incubated at 37°C in 5% CO 2 and 100% relative humidity for an additional 48 hours. After culture, medium from each compartment was collected, centrifuged for 15 minutes at 12,000 g /10°C, and the supernatants were stored at –70°C.
Doses for TNF, thrombin, and GM-CSF that were used in these studies were determined from dose-response studies in our previous publication, and they were added to the choriodecidua side only, because our previous studies have shown that the choriodecidua (rather than the amnion) is the initial target tissue or the tissue of origin of each.
Progestogens were chosen for study with the following rationale: progesterone was selected as the natural progestogen of major focus. MPA was chosen because it is known to be an effective inhibitor of TNF-induced GM-CSF production. It also metabolizes less rapidly in tissue culture than does progesterone. HP was selected as a natural progestogen that has a strong affinity for the membranous progesterone receptor (mPR), but not the nPR. 17-OH progesterone caproate was not used in this study. The concentration of progestogens used (10 –7 M) was within the range that is seen in pregnant serum at term. Progestogens were added to the choriodecidua side to ensure that they reached the target tissue concomitant with the weakening agents.
Strength testing of fetal membrane
Fetal membrane fragments were strength-tested within the Transwell inserts with the use of our rupture testing apparatus, as previously reported. Briefly, Transwell-mounted fetal membrane were secured in a 2.5-cm diameter fixture between the aligned horizontal plates of the rupture testing equipment. A motor-driven 1-cm diameter spherical-head plunger that was aligned perpendicular to the fetal membrane was then forced against the amnion side. Fetal membrane displacement and concomitant plunger force were recorded continuously until rupture. From these data, force (rupture strength in newtons) and maximum displacement (centimeters) that were needed to cause fetal membrane rupture were determined.
GM-CSF determination
GM-CSF levels in thawed supernatants from the fluid compartments adjacent to the choriodecidua and amnion sides of Transwell cultured fetal membrane were determined with the use of the Human GM-CSF Quantikine ELISA Kit (R&D Systems, Minneapolis, MN), according to the manufacturer’s protocol. Intra- and interassay precision of the assay was 2.7% and 5.3% coefficient of variation, respectively, for cell culture supernatants at a sensitivity level of 3 pg/mL with Escherichia coli expressed, recombinant human GM-CSF as standard.
Statistical analysis
All experiments were performed at least in triplicate. Data were analyzed by analysis of variance followed by post-hoc pair-wise comparisons (Holm-Sidak method) with the use of Sigmaplot software (Systat Software, Inc, Chicago, IL). Differences were considered significant when at a probability value of < .05.
Results
Progestogens inhibit thrombin- and TNF-induced fetal membrane weakening
Thrombin or TNF, with or without progestogens, were each added only to the choriodecidua side of the fetal membrane to mimic the physiologic conditions of pregnancy in which the choriodecidua is exposed to the maternal circulation and the amnion is in contact with the amniotic fluid. Compared with control membrane strength, thrombin or TNF each markedly decreased fetal membrane rupture strength in all studies (both P < .01). Preincubation with progesterone (10 –7 mmol) inhibited both thrombin- and TNF-induced fetal membrane weakening (both P < .05; Figure 1 , A). MPA (10 –7 mmol) preincubation also inhibited fetal membrane weakening that was induced by thrombin ( P = .01) and by TNF ( P = .02; Figure 1 , B). Finally, HP (10 –7 mmol) preincubation inhibited fetal membrane weakening that was induced by thrombin and TNF (both P < .01; Figure 1 , C). Preincubation with progesterone, MPA, or HP alone did not affect rupture strength ( Figure 1 ).
Effect of progestogens on GM-CSF production
Because GM-CSF is a critical intermediate in both the TNF- and thrombin-induced fetal membrane weakening pathways, further studies were done to determine whether progestogens inhibit TNF- or thrombin-induced fetal membrane weakening by blocking GM-CSF production. Consistent with our previous report, TNF and thrombin each increased GM-CSF release on the choriodecidua side concomitant with fetal membrane weakening ( Figure 2 ; P < .01). MPA markedly inhibited GM-CSF production by both TNF and thrombin to control levels ( P < .01). HP markedly inhibited the GM-CSF increase that was produced by TNF ( P < .01) but only partially inhibited the GM-CSF increase that was produced by thrombin ( P < .05). Finally, progesterone partially inhibited the GM-CSF increase produced by TNF ( P < .05). The progesterone effect on thrombin-induced GM-CSF production was not significant. Only small amounts of GM-CSF were detected on the amnion side of the fetal membrane ( Figure 2 ).
Progestogens inhibit GM-CSF–induced fetal membrane weakening
Studies were then performed to determine whether progestogens inhibit TNF- and thrombin-induced fetal membrane weakening by inhibiting the action of GM-CSF. Consistent with our previous report, GM-CSF incubation on the choriodecidua side of fetal membrane induced significant ( P < .01) fetal membrane weakening ( Figure 3 ). This GM-CSF–induced fetal membrane weakening was almost completely inhibited by preincubation with either MPA (10 –7 mmol) or HP (10 –7 mmol; both P < .01). Although preincubation with progesterone (10 –7 mmol) also blunted GM-CSF–induced fetal membrane weakening, its effect was not statistically significant. Rupture strength after incubation with progesterone, HP, or MPA alone was not different from control membranes ( Figure 3 ).
RU486 blocks HP inhibition of GM-CSF–induced fetal membrane weakening
In the experiment with RU486, consistent with the previously presented experiments, GM-CSF significantly decreased fetal membrane rupture strength, and preincubation with HP inhibited the GM-CSF–induced fetal membrane weakening. RU486 (10 –8 mmol) that was applied 1 hour before HP blocked HP inhibition of GM-CSF–induced weakening ( P < .01; Figure 4 ).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
