Intrauterine inflammation plays a major role in the etiology of preterm labor and birth. We established an ex vivo model employing perfused full-thickness term gestational membranes to study membrane transport, function, and inflammatory responses. Exposure of the maternal (decidual) face of the membranes to LPS (5 μg/ml) resulted in increased accumulation of proinflammatory cytokines in the maternal compartment within 4 h, followed by a response in the fetal (amniotic) compartment. Using cytokine arrays, exposure to LPS was found to result in increased secretion of a large number of cytokines and chemokines in both compartments, most notably IL-5, IL-6, IL-7, MDC (macrophage-derived chemokine), MIG (monokine induced by IFN-γ), TARC (thymus and activation-regulated chemokine), TGF-β, and TNF-α. PGE2 accumulation also increased in response to LPS, particularly in the fetal compartment. Cotreatment with sulfasalazine, which inhibited nuclear translocation of NF-κB p65, had a rapid and marked inhibitory effect on the rate of cytokine accumulation in the maternal compartment, with lesser but significant effects observed in the fetal compartment. While membrane integrity was not discernibly impaired with LPS or sulfasalazine exposure, rates of chorionic apoptosis after 20 h were doubled in sulfasalazine-treated tissues. We conclude that the system described provides a means of accurately modeling human gestational membrane functions and inflammatory activation ex vivo. Decidual LPS exposure was shown to elicit a robust inflammatory response in both the maternal and fetal compartments. Sulfasalazine was an effective antiinflammatory agent in this model, but also exerted proapoptotic effects that raise concerns regarding its placental effects when administered in pregnancy.

Spontaneous preterm labor and birth remains a major obstetric healthcare problem worldwide, with preterm delivery rates increasing in many countries during recent decades despite advances in the understanding of its pathophysiology (1, 2). In the United States alone, which in 2005 had a preterm delivery rate of 12.5% (1), costs associated with the perinatal, neonatal, childhood, and long-term care of preterm infants have been estimated to exceed $26 billion annually (1). Ascending intrauterine infection is recognized as a major cause of preterm labor, particularly in deliveries before 30 wk of gestation (3). Bacterial invasion elicits a host response in the gestational membranes (amnion, chorion, and decidua) (4), with enhanced expression of inflammatory genes such as cytokines, chemokines, and matrix metalloproteases (MMPs)3 (4, 5, 6, 7). Subsequent recruitment and activation of leukocytes to the fetal membranes results in amplification of the inflammatory response and histologic chorioamnionitis, a diagnostic hallmark of intrauterine inflammation and a common finding following spontaneous preterm birth (4, 5, 8).

Exposure of the fetus to elevated cytokine concentrations in utero is thought to be associated with various morbidities after birth (9, 10). Fetal inflammatory response syndrome, a condition associated with elevated levels of cytokines in the fetal circulation, is a risk factor for development of cerebral palsy and white matter damage (10). In a significant proportion of preterm deliveries complicated by histologic chorioamnionitis, amniotic fluid is sterile (11, 12, 13), and although there is no evidence of fetal infection, elevated concentrations of cytokines in amniotic fluid (9) and fetal plasma (14) have been reported. Under these circumstances, it is likely that detrimental fetal responses occur primarily in response to exposure to elevated cytokine concentrations, rather than to direct exposure to the pathogen (14, 15). Since the membranes present an effective barrier to endotoxin and most cytokines (16, 17), the source of cytokines in amniotic fluid is presumed to be the fetal membranes or fetal leukocytes.

Activating molecules derived from microbial pathogens trigger the innate immune system via an array of pathogen-activated molecular pattern receptors (18, 19). The nuclear transcription factor NF-κB appears to be a nexus through which pathogen-activated molecular patterns activate inflammatory gene transcription (20). NF-κB also activates transcription of antiapoptotic genes, and its inhibition can dramatically increase rates of apoptosis, particularly in neoplastic tissues (21). NF-κB activation in fetal membranes has been studied and documented in the context of normal term labor and inflammatory activation, both in vitro and in vivo (22, 23, 24). Several studies in which pharmacologic agents were employed to inhibit LPS-stimulated inflammatory activation in choriodecidual explants have been published (25, 26). In one of these studies, the aminosalicylate drug sulfasalazine (SSZ), prescribed during pregnancy for treating inflammatory bowel disease (27), was shown to be successful in suppressing LPS-stimulated production of IL-6 and TNF-α during a relatively short time period (4 h), without evidence of increased cell death (25). Its effects on gene expression and apoptosis in gestational membranes during longer periods of time have not been investigated, however.

In the present study, we have established and validated an ex vivo model with which to study inflammatory activation and mediator production in gestational membranes in response to maternal endotoxin exposure. Based on the premise that treatment of women at risk of preterm labor with anti-inflammatory therapeutics offers a potential modality for the prevention of preterm birth (28), we investigated the effects of cotreatment with SSZ on the inflammatory response, its rate of progression, and the subsequent viability of the various cells and tissues within the gestational membranes.

M199 media (phenol red-free), BSA penicillin-streptomycin-glutamine supplement and Cy3-labeled anti-rabbit IgG Ab were obtained from Invitrogen. FITC-coupled dextran (FITC-Dx, Mr of 4400 Da), SSZ hydrochloride, LPS from Escherichia coli 055:B5, DAB staining kits, TOX-7 lactate dehydrogenase (LDH) assay kits, and bovine LDH (500 U/mg) calibration standard were obtained from Sigma-Aldrich Australia. The M30 CytoDeath Ab was sourced from Roche Diagnostics Australia, the anti-p65/relA Ab was from BioLegend, while anti-cleaved caspase-3 Ab was from Santa Cruz Biotechnology. Cytokine ELISA development kits were supplied by R&D Systems or PeproTech. The MMP-9 FRET assay kit was from AnaSpec. Human cytokine arrays were from RayBiotech.

Full thickness extraplacental membranes were obtained from women at term undergoing delivery by Cesarean section; indications were previous Cesarean section or malpresentation. Women with preeclampsia, intrauterine infection, or any metabolic disease of pregnancy were excluded. Maternal consent for collection of placentae was obtained in accordance with the requirements of the Regional Ethics Committees. Within 45 min of delivery, placentas were transported to the laboratory and the membranes washed in sterile PBS to remove blood. A small section (1 × 3 cm) of each placenta was dissected and placed in 4% paraformaldehyde for histological analysis. Membranes were spread on a glass plate and four circular sections cut using a disk cutter (6.5 cm diameter). Each section was then sandwiched between two 6.5-cm semirigid mesh disks and secured within an Ussing chamber (supplemental Fig. 1).4 A small sterile magnetic stirring bar was added to the maternal compartment. Using a syringe, the maternal and fetal compartments were slowly filled with warm sterile media (M199 supplemented with 0.2% BSA, 1× penicillin-streptomycin-glutamine, and 25 mM HEPES (pH 7.4)), and the apparatus was placed onto a large magnetic stirrer within a prewarmed (37°C) perfusion cabinet. The inlet ports of the four chambers (maternal and fetal) were then connected to a 16-channel peristaltic perfusion pump via tygon tubing, and media from individual perfusate reservoirs (maintained at 37°C) were pumped through the chambers at 1.5 ml/min flow rate. The maternal perfusate only was gassed with carbogen to maintain oxygen and carbon dioxide partial pressures and was supplemented with FITC-Dx (10 μM) and either PBS vehicle (control), LPS (5 μg/ml), or LPS plus SSZ (3.5 mM). Duplicate chambers were used for each condition. Outlet ports were connected to sterile tubing allowing the exiting media to flow back into the perfusate reservoirs, thereby creating a closed loop for each compartment. The total volume of perfusion fluid contained within each compartment and tubing loop was 25 ml, with the chamber and tubing holding ∼18 ml. After allowing approximately 30 min to equilibrate the system, the experiment was started (assigned t = 0 min) and pumping continued for 20 h. Samples of perfusates (0.5 ml) were taken by syringe at 0, 1, 2, 4, 8, and 20 h and sealed for blood gas analysis, which was conducted using the Clinical Biochemistry Department’s blood gas analyzer. Additional samples of 0.5 ml were removed and frozen at −20°C for subsequent analysis. At the completion of the perfusion period the chambers were drained of media (20 h final collection); postperfusion membrane segments (1 × 3 cm) were fixed in 4% paraformaldehyde for 8 h, dehydrated, and embedded in paraffin for histological analysis. Membranes from a total of 10 different placentas were used in these studies and all were included in the analyses. The effects of LPS stimulation (vs vehicle controls) were tested in five experiments, and the effects of SSZ on LPS-stimulated membranes (vs LPS controls) were subsequently investigated using another five placentas.

SSZ concentrations in perfusates were measured by determining the absorbance at 405 nM with reference to a calibration curve (0.078–5 mM) prepared in the same media and read in a SpectraMax 750 plate reader (Molecular Devices). Glucose was measured using a modification of Trinder’s glucose oxidase/4-amino-antipyrene method (29). The calibration curve ranged from 0.078 to 5 mM; intraassay coefficient of variation (CV) was <5%. FITC-Dx was measured by fluorescence (1420 Victor2 plate reader; Wallac-Oy, Turku, Finland) at excitation/emission of 495/520 nm. The assay was linear across the standard curve range of 0.078–5 μM. LDH release was measured using a TOX-7 LDH kit in accordance with the manufacturer’s instructions (Sigma-Aldrich). The assay was calibrated against bovine heart LDH (calibration curve range, 1.4–1000 mU/ml; CV, 7%).

Concentrations of cytokines in maternal and fetal perfusates were measured as previously described (30) using ELISA development kits from either R&D Systems or PeproTech according to the manufacturers’ instructions. Sensitivity ranged from 2 to 25 pg/ml; CVs were <12%. PGE2 was measured by RIA as previously described (30, 31); the assay has a sensitivity of 7 pg/ml and a CV of <10%. The major PGE2 metabolite PGEM-II was measured by enzyme immunoassay using kits from Cayman Chemicals according to the manufacturer’s recommendations; sensitivity was <2 pg/ml and the CV was <15%. MMP-9 activity was measured by fluorescence resonance energy transfer assay using kits from AnaSpec, as detailed by the manufacturer. The assay was calibrated against a substrate concentration curve ranging from 0.078 to 2.5 μM; CV was <5%.

Cytokine protein arrays were performed as recommended by the manufacturer (RayBiotech) on maternal and fetal samples at the 20 h time point pooled from n = 3 representative experiments. Each array contained duplicate spots for 42 cytokines, chemokines, and growth factors, plus positive and negative controls. Eight arrays (two sets of four) were probed in total, with two pairs of maternal perfusate samples and two pairs of fetal samples representing either control and LPS (set 1) or LPS and LPS plus SSZ (set 2). Arrays were developed using ECL and visualized by exposure to x-ray film. Spot intensity was quantitated by densitometry using QuantityOne software (Bio-Rad Laboratories) and normalized to a positive control spot on each array. CV between duplicate spots was ∼15%. After normalization, data from the LPS samples from sets 1 and 2 differed on average by <30%, and therefore they were combined and treated as a single sample for subsequent analysis.

Sections of pre- and postperfusion membranes were stained with H&E to exclude the presence of infiltrating leukocytes, and changes in structural integrity and morphology following perfusion were assessed. Apoptosis was assessed after Ag retrieval using either anti-M30 immunostaining or immunostaining for cleaved caspase-3. Primary Ab was omitted in negative controls, which were devoid of staining. Peroxidase staining was visualized using diaminobenzidine, and slides were counterstained with Gill’s hematoxylin before dehydration and coverslipping. Slides were viewed using a Nikon Eclipse E800 microscope equipped with a TK-C1381 color video camera (JVC). Positive staining was quantified by counting the percentage of stained cells per high-power (×400) field; five fields per slide were selected at random for counting (∼500 cells).

Membranes were exposed to LPS with/without SSZ for 3 h, then were isolated, fixed in 4% paraformaldehyde, embedded in OCT, and frozen on dry ice. Sections (10 μm) were cut using a CM1900 cryostat (Leica Microsystems), subjected to Ag retrieval in 0.05% citraconic anhydride (32), and immunostained using anti-p65/relA Ab (1/400). Staining was developed by incubation with Cy3-labeled anti-rabbit IgG (1/600) and visualized at ×200 magnification by fluorescence using a Nikon Eclipse Ti inverted microscope. Digital images were processed using NIS-Elements software version 3.0 (Nikon).

To correct for small changes in perfusate volume during perfusion, concentration data were multiplied by perfusion volume at each time point to derive an estimate of amount per compartment. Data from duplicate chambers were averaged, and the data from multiple experiments were analyzed collectively by ANOVA with a Student-Newman-Keuls post hoc test or with a paired t test. No statistical analysis was made of the protein array data due to the absence of experimental replicates.

The integrity and viability of the model was evaluated using a number of approaches. FITC-Dx was added to the maternal compartment, and its accumulation in the fetal compartment (as an index of membrane permeability or leakage) was measured by fluorescence spectroscopy. None of the chambers used (n = 10 experiments, n = 4 chambers per experiment) showed evidence of rupture, with maternal-fetal dextran transfer averaging 0.5% per hour and never exceeding 1% per hour (Fig. 1,A). Exposure to LPS did not increase FITC-Dx permeability. Glucose utilization rates on both sides of the chambers were within accepted values (∼3.75% per hour) and remained relatively constant throughout 20 h irrespective of treatment (Fig. 1,B). Maternal pO2 values were maintained at 25–40 kPa (<98% saturation), pCO2 at 2.6–4.0 kPa, and pH at 7.0–7.6. Fetal pO2 and pCO2 levels were only marginally lower (20–27 and 1.2–3.1 kPa, respectively) despite the absence of carbogenation. LDH accumulation showed a modest increase with time (Fig. 1 C) and was unaffected by treatment (data not shown). Histological examination of H&E-stained membranes postperfusion revealed no differences in structural morphology compared with sections from freshly delivered membranes.

FIGURE 1.

Ussing chamber validation (mean ± SD). A, Fluorescein-dextran accumulation in control and LPS-treated fetal compartments expressed as percentage of total (maternal plus fetal) (n = 3) experiments. B, Glucose uptake in maternal and fetal control compartments (n = 3 experiments). C, LDH accumulation in maternal and fetal compartments (n = 3 experiments).

FIGURE 1.

Ussing chamber validation (mean ± SD). A, Fluorescein-dextran accumulation in control and LPS-treated fetal compartments expressed as percentage of total (maternal plus fetal) (n = 3) experiments. B, Glucose uptake in maternal and fetal control compartments (n = 3 experiments). C, LDH accumulation in maternal and fetal compartments (n = 3 experiments).

Close modal

To evaluate the effects of LPS on secretion of a wide range of cytokines and chemokines, protein expression arrays were used to analyze pooled perfusates (20 h) from control and LPS-treated chambers (Table I). Overall, cytokine accumulation was lower in the fetal compared with maternal perfusates. Robust responses to LPS stimulation in the maternal compartment were observed for IL-5 (16-fold), IL-6 (8-fold), IL-7 (11-fold), MDC (macrophage-derived chemokine; 20-fold), MIG (monokine induced by IFN-γ 4-fold), M-CSF (3-fold), TGF-β (3-fold), TARC (thymus and activation-regulated chemokine; 4.5 fold), TNF-α (7.5-fold), and MCP-1, whereas IL-1β, G-CSF, GM-CSF, GRO (growth-related oncogene), IL-8, IL-13, I-309, MCP-2, OSM (oncostatin-M), and RANTES were unresponsive. Stem cell factor levels actually decreased by ∼50% in LPS-treated perfusates. In the fetal compartment, robust responses were exhibited by IL-5 (6.5-fold), IL-6 (3-fold), IL-7 (4-fold), IL-8 (4-fold), MIG (11-fold), SDF-1 (stromal cell-derived factor-1; 4.5-fold), TGF-β (2-fold), and TARC (5-fold), with the remainder giving a weak or absent response.

Table I.

Cytokine protein array dataa

CytokinesMaternal CompartmentFetal CompartmentChemokinesMaternal CompartmentFetal Compartment
ControlLPSLPS + SSZControlLPSLPS + SSZControlLPSLPS + SSZControlLPSLPS + SSZ
IL-1β 4.2 2.1 2.3 0.5 0.9 0.7 GRO 4.7 5.0 1.3 0.1 0.1 0.1 
IL-3 1.5 2.6 0.7 1.2 1.6 2.7 I-309 3.6 3.2 0.5 0.6 1.0 0.4 
IL-4 1.6 3.1 0.7 0.9 2.2 2.0 IL-8 31.8 36.2 2.8 5.8 25.6 11.0 
IL-5 1.8 29.8 1.2 1.9 12.1 8.2 MCP-1 0.1 6.3 0.4 1.4 1.0 1.5 
IL-6 172.0 1392.1 621.2 270.7 733.3 646.1 MCP-2 1.2 1.3 2.3 1.1 0.7 1.1 
IL-7 6.3 69.6 8.5 2.7 10.8 3.9 MCP-3 2.4 4.8 1.2 1.5 2.3 1.4 
GCSF 7.4 8.3 3.9 1.1 1.9 1.1 MCSF 1.3 4.1 0.5 1.1 1.3 1.1 
GM-CSF 3.6 4.8 1.7 0.9 1.2 0.4 MDC 16.3 335.0 11.2 2.7 5.8 5.6 
IL-12 100.2 261.1 124.1 31.2 76.1 56.6 MIG 119.7 467.9 3.9 3.8 41.9 4.7 
IL-13 6.7 6.4 1.7 1.6 1.7 1.4 MIP-1δ 9.6 23.2 2.4 1.4 2.2 0.7 
IL-15 3.1 5.4 0.8 0.7 0.2 0.2 RANTES 61.7 46.5 10.4 22.0 35.6 34.3 
OSM 3.3 5.5 0.6 0.8 0.5 0.2 SCF 7.5 3.7 2.2 0.5 1.6 1.7 
TGF-β 4.7 13.1 1.1 10.4 23.7 12.8 SDF-1 33.0 62.7 20.3 2.7 11.6 10.3 
TNF-α 18.1 136.4 11.9 13.2 12.0 7.3 TARC 186.0 815.4 558.6 80.4 390.5 376.0 
CytokinesMaternal CompartmentFetal CompartmentChemokinesMaternal CompartmentFetal Compartment
ControlLPSLPS + SSZControlLPSLPS + SSZControlLPSLPS + SSZControlLPSLPS + SSZ
IL-1β 4.2 2.1 2.3 0.5 0.9 0.7 GRO 4.7 5.0 1.3 0.1 0.1 0.1 
IL-3 1.5 2.6 0.7 1.2 1.6 2.7 I-309 3.6 3.2 0.5 0.6 1.0 0.4 
IL-4 1.6 3.1 0.7 0.9 2.2 2.0 IL-8 31.8 36.2 2.8 5.8 25.6 11.0 
IL-5 1.8 29.8 1.2 1.9 12.1 8.2 MCP-1 0.1 6.3 0.4 1.4 1.0 1.5 
IL-6 172.0 1392.1 621.2 270.7 733.3 646.1 MCP-2 1.2 1.3 2.3 1.1 0.7 1.1 
IL-7 6.3 69.6 8.5 2.7 10.8 3.9 MCP-3 2.4 4.8 1.2 1.5 2.3 1.4 
GCSF 7.4 8.3 3.9 1.1 1.9 1.1 MCSF 1.3 4.1 0.5 1.1 1.3 1.1 
GM-CSF 3.6 4.8 1.7 0.9 1.2 0.4 MDC 16.3 335.0 11.2 2.7 5.8 5.6 
IL-12 100.2 261.1 124.1 31.2 76.1 56.6 MIG 119.7 467.9 3.9 3.8 41.9 4.7 
IL-13 6.7 6.4 1.7 1.6 1.7 1.4 MIP-1δ 9.6 23.2 2.4 1.4 2.2 0.7 
IL-15 3.1 5.4 0.8 0.7 0.2 0.2 RANTES 61.7 46.5 10.4 22.0 35.6 34.3 
OSM 3.3 5.5 0.6 0.8 0.5 0.2 SCF 7.5 3.7 2.2 0.5 1.6 1.7 
TGF-β 4.7 13.1 1.1 10.4 23.7 12.8 SDF-1 33.0 62.7 20.3 2.7 11.6 10.3 
TNF-α 18.1 136.4 11.9 13.2 12.0 7.3 TARC 186.0 815.4 558.6 80.4 390.5 376.0 
a

Each array was exposed to a sample comprised of pooled perfusate from the 20 h time point of n= 3 experiments (either maternal or fetal compartments). Data are shown as mean background-corrected densitometry of duplicate spots normalized to an internal positive control. The LPS column represents the mean signal of two arrays each probed with perfusate from LPS-treated chambers from two different sets of experiments. Data from low signal/inconsistent spots are not shown. OSM indicates oncostatin-M; GRO, growth-related oncogene; SCF, stem cell factor; SDF-1, stromal cell-derived factor-1.

The individual 20-h perfusates from all Ussing chambers were assayed by ELISA to assess consistency and accuracy of the protein array data (Fig. 2). Consistent with the array data, IL-6 accumulation was responsive to LPS stimulation in both compartments, while IL-8 showed little response. TNF-α accumulation was similar in the maternal and fetal compartments and showed the greatest response to LPS of the cytokines measured. IL-1β was only detectable in the maternal compartment and was modestly increased by LPS treatment. IL-10, TARC, and MDC, which were abundant on the arrays, were not consistently detectable by ELISA. In contrast to the cytokines, PGE2 was detected in (∼4-fold) higher amounts in the fetal compartment; decidual LPS exposure resulted in significant increases in PGE2 accumulation in both compartments (Fig. 2). The terminal PGE2 metabolite PGEM-II was present in similar amounts in the fetal and maternal compartments; levels were significantly increased by LPS treatment in both compartments. LPS treatment did not significantly change the PGE2-to-PGEM ratio in maternal compartments, but it effected a decrease in the ratio in fetal compartments.

FIGURE 2.

Accumulation of cytokines, chemokines, and prostaglandins in maternal and fetal compartments under basal conditions (control) or stimulated with LPS (5 μg/ml) over 20 h. Data are means ± SEM from n = 5 experiments performed in duplicate. ∗, p < 0.05 vs control by ANOVA with Dunnett’s post hoc test.

FIGURE 2.

Accumulation of cytokines, chemokines, and prostaglandins in maternal and fetal compartments under basal conditions (control) or stimulated with LPS (5 μg/ml) over 20 h. Data are means ± SEM from n = 5 experiments performed in duplicate. ∗, p < 0.05 vs control by ANOVA with Dunnett’s post hoc test.

Close modal

Samples taken at 0, 2, 4, 8, and 20 h perfusion were then assayed to determine rates of accumulation of IL-6, TNF-α, IL-8, and PGE2 in maternal and fetal compartments under control or LPS-stimulated conditions (Fig. 3). IL-6 accumulation increased steadily with time in both compartments (Fig. 3,A). LPS exposure significantly increased IL-6 production at the 4 h time point and this stimulation was maintained through to the 20 h time point in the maternal compartment. IL-8 levels were insensitive to LPS stimulation at all time points (Fig. 3,B). TNF-α production was significantly stimulated by LPS from the 4 h time point in both compartments (Fig. 3,C). In contrast, PGE2 accumulation increased in both compartments relatively slowly (Fig. 3 D). From the 4 h time point onward, LPS-treated chambers accumulated significantly greater amounts of PGE2, but the differences were modest (∼2-fold above basal) and did not increase with time.

FIGURE 3.

Time course of accumulation of inflammatory mediators in control and LPS-treated chambers (maternal and fetal). IL-6 (A), IL-8 (B), TNF-α (C), PGE2 (D). Data are shown as means ± SEM from n = 5 experiments performed in duplicate.

FIGURE 3.

Time course of accumulation of inflammatory mediators in control and LPS-treated chambers (maternal and fetal). IL-6 (A), IL-8 (B), TNF-α (C), PGE2 (D). Data are shown as means ± SEM from n = 5 experiments performed in duplicate.

Close modal

To explore the effects of pharmacological inhibition of NF-κB activation in this model, maternal compartments were exposed either to LPS alone (5 μg/ml) or to LPS plus SSZ (3.5 mM). SSZ levels steadily declined in the maternal compartment and increased in the fetal compartment, consistent with a slow rate of transfer across the membranes (supplemental Fig. 2). However, even after 20 h a differential of ∼30% between maternal and fetal concentrations remained.

SSZ exerted rapid and significant inhibition of IL-6 and TNF-α accumulation in both fetal and maternal compartments (Fig. 4). In the case of IL-6, the presence of SSZ significantly reduced production at the initial time point, indicating an almost immediate effect exerted during the initial equilibration phase of the experiment (Fig. 4, A and C). The effects of SSZ on TNF-α production were less rapid, becoming significant only after 8 h (Fig. 4, B and D). For both cytokines, SSZ exerted a greater inhibitory effect on maternal cytokine production compared with fetal production. These trends are evident in the data from the final time point shown in Fig. 5. SSZ treatment reduced LPS-stimulated production of IL-6, TNF-α, IL-8, and PGE2, with maternal production of IL-8, TNF-α, and PGE2 being the most responsive (>90% inhibition).

FIGURE 4.

Time course of effects of SSZ treatment (3.5 mM) on LPS-stimulated accumulation of IL-6 and TNF-α in maternal (A and C) and fetal (B and D) compartments over 20 perfusions. Data represent means ± SD of n = 5 experiments performed in duplicate. ∗, p < 0.05 vs control (LPS only) by paired t test.

FIGURE 4.

Time course of effects of SSZ treatment (3.5 mM) on LPS-stimulated accumulation of IL-6 and TNF-α in maternal (A and C) and fetal (B and D) compartments over 20 perfusions. Data represent means ± SD of n = 5 experiments performed in duplicate. ∗, p < 0.05 vs control (LPS only) by paired t test.

Close modal
FIGURE 5.

Comparison of inflammatory mediator accumulation in maternal and fetal compartments after 20 h of exposure to LPS with/without SSZ (3.5 mM). Data are means ± SD from n = 5 experiments performed in duplicate. ∗, p < 0.05 vs control (zero dose SSZ) by ANOVA with Dunnett’s post hoc test.

FIGURE 5.

Comparison of inflammatory mediator accumulation in maternal and fetal compartments after 20 h of exposure to LPS with/without SSZ (3.5 mM). Data are means ± SD from n = 5 experiments performed in duplicate. ∗, p < 0.05 vs control (zero dose SSZ) by ANOVA with Dunnett’s post hoc test.

Close modal

Data from the cytokine arrays reinforced these trends, with most cytokines showing a good response to SSZ-mediated suppression, even those that had not been particularly responsive to LPS (Table I). In the maternal compartment perfusate pool, MDC, MIG, and IL-5 concentrations were all reduced by >95% by SSZ treatment, while IL-8, MCP-1, TGF-β, and TNF-α levels were reduced by >90%. In the fetal compartments, however, only MIG concentrations were inhibited to a similar magnitude (89%), with another seven cytokines showing >50% inhibition. TNF-α and IL-6 levels in the fetal perfusate pool were reduced by 39 and 12%, respectively. MMP-9 was also measured in the final 20-h perfusates (supplemental Fig. 3). MMP-9 activity was not responsive to LPS stimulation, and only in the maternal compartment was SSZ treatment associated with a significant decline (∼50%) in MMP-9 activity.

To confirm that SSZ was exerting its antiinflammatory actions through inhibition of NF-κB activation, we studied the cellular localization of the NF-κB subunit p65/RelA in perfused full thickness membranes by immunofluorescence microscopy. Staining for p65 was intensely cytoplasmic in the amnion and chorion of control membranes, with weaker cytoplasmic staining evident in the decidua (Fig. 6,A). After 3 h of stimulation with LPS (decidual face), the strong cytoplasmic p65 staining in the chorion was diminished and became more evenly distributed throughout the cells (Fig. 6,B); less of a response to LPS was observed in the decidua and amnion. Cotreatment with SSZ prevented the nuclear localization in the chorion and preserved the distinctive cytoplasmic staining pattern similar to controls (Fig. 6 C). More equivocal results were observed in amnion and decidua.

FIGURE 6.

Effects of LPS and SSZ on p65/RelA localization. Full-thickness membranes were perfused for 3 h with either vehicle control (A), LPS (B), or LPS plus SSZ (C) in the maternal compartment and then fixed, embedded, and cryosectioned. Localization of p65/RelA was visualized by immunofluorescence. Images were taken at ×200 magnification. Nuclear translocation was observed in response to LPS, mainly in the chorion, which was abrogated by SSZ cotreatment. EA, amniotic epithelium; CT, chorionic trophoblast; DS, decidual stromal cells.

FIGURE 6.

Effects of LPS and SSZ on p65/RelA localization. Full-thickness membranes were perfused for 3 h with either vehicle control (A), LPS (B), or LPS plus SSZ (C) in the maternal compartment and then fixed, embedded, and cryosectioned. Localization of p65/RelA was visualized by immunofluorescence. Images were taken at ×200 magnification. Nuclear translocation was observed in response to LPS, mainly in the chorion, which was abrogated by SSZ cotreatment. EA, amniotic epithelium; CT, chorionic trophoblast; DS, decidual stromal cells.

Close modal

Because of the propensity for NF-κB inhibition to promote apoptosis, we examined cell viability and tissue integrity in the membranes during and after perfusion with LPS and SSZ (Fig. 7). Histological analysis using M30 immunostaining revealed a significant increase (∼2-fold) in the number of proapoptotic cells in the chorionic membrane with SSZ plus LPS vs LPS alone (Fig. 7,A). Staining for cleaved caspase-3 gave similar results (data not shown). No M30 staining was detected in either the amnion or decidua. There was also no significant change in membrane integrity, as judged by the passage of FITC-Dx from the maternal to fetal compartment (Fig. 7,B), or LDH accumulation (Fig. 7 C), after LPS or SSZ exposure.

FIGURE 7.

Assessment of tissue death and integrity in response to LPS and SSZ treatment after 20 h of perfusion. A, Immunohistochemical staining of apoptotic cells in the chorion (as percentage total) after exposure LPS or LPS plus SSZ. B, FITC-Dx accumulation in fetal compartments treated with LPS or LPS plus SSZ. C, LDH release in maternal and fetal compartments. Data are from n = 3 experiments performed in duplicate. ∗, p < 0.05 vs control (no SSZ) by ANOVA with Dunnett’s post hoc test.

FIGURE 7.

Assessment of tissue death and integrity in response to LPS and SSZ treatment after 20 h of perfusion. A, Immunohistochemical staining of apoptotic cells in the chorion (as percentage total) after exposure LPS or LPS plus SSZ. B, FITC-Dx accumulation in fetal compartments treated with LPS or LPS plus SSZ. C, LDH release in maternal and fetal compartments. Data are from n = 3 experiments performed in duplicate. ∗, p < 0.05 vs control (no SSZ) by ANOVA with Dunnett’s post hoc test.

Close modal

In this study we have developed and validated an ex vivo model for evaluating the secretory, barrier, transfer functions and inflammatory responses of human fetal membranes. While rodent and nonhuman primate models have been used recently to investigate potential modalities for preventing preterm labor and birth, similar studies in humans first require that candidate agents have been fully tested using human in vitro models. Our approach was aimed at approximating conditions in vivo, and it overcomes many of the limitations of cell culture or explant-based methods related to issues of structural integrity, cellular composition, and architecture. While other investigators have utilized a variety of simpler membrane perfusion models in the past (7, 17, 33, 34), ours provides an additional level of sophistication in terms of the ability to independently regulate fetal and maternal oxygenation, media composition, and stimuli exposure, while sampling both maternal and fetal compartments in real time and maintaining tissue integrity with no loss of viability over a minimum of 20 h. By using this system, studies of membrane transport and permeability are possible, both for the short and long terms, providing that adequate replacement of perfusion media is conducted, as little evidence of tissue necrosis, apoptosis, or structural remodeling was observed over a 20-h period under basal conditions. Overall, our results are consistent with findings from clinical studies in which inflammatory substances have been measured in amniotic fluid from preterm deliveries complicated by intraamniotic infection, supporting the validity and accuracy of the model (3, 5).

One of the major findings of this study is that a vigorous cytokine response in the fetal compartment can be elicited by exposure of maternal decidual tissues to LPS. This response occurs in the constituent cells that comprise the tissues without the requirement for leukocyte recruitment and activation. Since LPS does not cross the membranes to any significant extent (16), this implies that amniotic fluid cytokine concentrations can be increased in response to decidual activation, presumably via a paracrine cascade of mediators released within the various layers of the membranes leading to secretion from the amnion into the amniotic cavity. This finding may illuminate how elevated fetal cytokine levels and fetal inflammatory response syndrome can occur even in the absence of any identifiable pathogen in fetal fluids. Fetal exposure to amniotic fluid proinflammatory cytokines can exert inflammatory effects on fetal tissues and elevate circulating cytokine levels with potentially deleterious consequences (10, 35, 36, 37).

We observed an increase in a secretion of a wide range of mediators in response to LPS, several of which have also been documented in amniotic fluid from pregnancies with intrauterine inflammation (3, 5). The protein array and ELISA data were consistent in general, although there were some anomalies that may reflect differences in Ab specificity. While cytokine levels in the fetal compartment were somewhat lower than maternal levels, the reverse was true for PGE2, in accordance with previous reports that the amnion is the major source of prostaglandins in the membranes (38) and that prostaglandins cross the gestational membranes only poorly (39). Proportionally more PGE2 in the maternal compartment was metabolized to PGEM-II than in the fetal compartment, consistent with the expectation that most amnion-derived PGE2 crossing the membranes would be metabolized by chorionic 15-PGDH (15-hydroxyprostaglandin dehydrogenase) during passage. We and others have previously reported that cytokine exposure decreases 15-PGDH expression (40) and therefore anticipated that LPS-treated chambers would have decreased prostaglandin metabolism resulting in an increased PGE2-to-PGEM ratio. However, an apparent decrease in the ratio was observed. It has recently been reported that 15-PGDH abundance in gestational membranes is increased in premature rupture of membranes and unaffected in intact membranes with chorioamnionitis (41), consistent with our data.

SSZ, a salicylate precursor that prevents activation of NF-κB (42, 43), has been used for decades clinically as a well-tolerated and effective antiinflammatory drug in the treatment of rheumatoid arthritis and Crohn’s disease (44). Our findings that SSZ is an effective antiinflammatory agent in gestational membranes support those of Lappas et al. in a short-term (4 h) explant model (25). However, while Lappas and coworkers failed to find evidence of SSZ-induced apoptosis, in our model with longer exposure times we observed a marked increase in apoptotic cells in the chorionic membrane in response to 20-h treatment with SSZ. Although this did not result in detectable changes in LDH release or membrane permeability, degradative changes may develop with extended exposure times. This issue is of particular concern if SSZ or related drugs were to be employed prophylactically to prevent inflammation-driven preterm labor in high-risk women (28). Evidence from earlier retrospective studies suggested that SSZ administration in pregnancy is not associated with a discernable increase in risk of fetal congenital defects, morbidity, or mortality (44, 45, 46), findings supported by a recent metaanalysis (47). However, a Scandinavian cohort study of 900 pregnancies reported evidence of increased risk of a number of adverse outcomes associated with SSZ treatment, although causality was not proven (48).

The effects of SSZ in our model were rapid, an important pharmacological characteristic should the drug be used to treat women with inflammation-driven preterm labor. The drug was also effective: most cytokines and chemokines that were measured showed significant inhibition by SSZ, even those that had not responded well to LPS stimulation. The most likely explanation for this latter observation is that inflammatory pathways were stimulated under basal conditions due to the trauma of isolation and manipulation, and hence were amenable to SSZ-mediated inhibition. Our studies suggest that although the membranes acted as a partial exclusion barrier, significant amounts of SSZ could enter amniotic fluid. We observed that SSZ inhibited nuclear p65 translocation in the chorion after 3 h of exposure, implying exposure of these tissues to adequate concentrations of the drug, whereas in the decidua and amnion changes in staining patterns were more equivocal. However, it is possible that p65 translocation/inhibition may have become more evident in amnion and decidua if longer or shorter incubation periods had been employed. SSZ has recently been shown to be a substrate for the ABC efflux protein breast cancer resistance protein (BCRP)/ABCG2 (49), which is expressed both in the placenta and fetal membranes (50). It remains to be determined whether BCRP (or any other transporter) is responsible for maintaining the apparent concentration differential between maternal and fetal compartments.

In conclusion, we have employed an ex vivo model to study the inflammatory response to LPS in human full-thickness gestational membranes, demonstrating that decidual exposure is sufficient to evoke a vigorous response on the fetal side of the membranes, a finding with important implications for pregnancies complicated by maternal infection. The antiinflammatory drug SSZ exerted marked inhibitory effects on both maternal and fetal cytokine, chemokine, and prostaglandin production, suggesting it may have potential for the prevention and treatment of inflammation-driven preterm labor. However, further investigation is required to explore the potential effects of prolonged treatment with SSZ on fetal membrane integrity and function.

We gratefully acknowledge the assistance of the theater staff at National Women’s Hospital, Auckland, New Zealand, for their help with the collection of tissues, Dr. Roger Lins for providing valuable comments on this manuscript, and Irving Aye for his assistance with the cryosectioning of tissues for immunofluorescence staining.

The authors have no financial conflicts of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported in part by a Prematurity Initiative Program Grant No. 21-FY05-1247 from the March of Dimes Foundation, the Health Research Council of New Zealand, the Maurice and Phyllis Paykel Trust, and the National Research Centre for Growth and Development (New Zealand). J.A.K. is supported by the Women and Infants Research Foundation (Western Australia); M.D.M. is supported by a James Cook Research Fellowship funded by the Royal Society of New Zealand.

3

Abbreviations used in this paper: MMP, matrix metalloprotease; CV, coefficient of variation; FITC-Dx, FITC-coupled dextran; LDH, lactate dehydrogenase; MDC, macrophage-derived chemokine; MIG, monokine induced by IFN-γ; SSZ, sulfasalazine; TARC, thymus and activation-regulated chemokine.

4

The online version of this article contains supplemental material.

1
Behrman, R. A., A. Stitch Butler.
2006
.
Preterm Birth: Causes, Consequences and Prevention
Natl. Acad. Press, Washington, DC.
2
Goldenberg, R. L., J. F. Culhane, J. D. Iams, R. Romero.
2008
. Epidemiology and causes of preterm birth.
Lancet
371
:
75
-84.
3
Romero, R., J. Espinoza, L. F. Goncalves, J. P. Kusanovic, L. A. Friel, J. K. Nien.
2006
. Inflammation in preterm and term labour and delivery.
Semin. Fetal Neonatal Med.
11
:
317
-326.
4
Goldenberg, R. L., W. W. Andrews, J. C. Hauth.
2002
. Choriodecidual infection and preterm birth.
Nutr. Rev.
60
:
S19
-S25.
5
Romero, R., O. Erez, J. Espinoza.
2005
. Intrauterine infection, preterm labor, and cytokines.
J. Soc. Gynecol. Investig.
12
:
463
-465.
6
Keelan, J. A., M. Blumenstein, R. J. Helliwell, T. A. Sato, K. W. Marvin, M. D. Mitchell.
2003
. Cytokines, prostaglandins and parturition: a review.
Placenta
24
:
S33
-S46.
7
Zaga-Clavellina, V., H. Merchant-Larios, G. Garcia-Lopez, R. Maida-Claros, F. Vadillo-Ortega.
2006
. Differential secretion of matrix metalloproteinase-2 and -9 after selective infection with group B streptococci in human fetal membranes.
J. Soc. Gynecol. Investig.
13
:
271
-279.
8
Andrews, W. W., R. L. Goldenberg, O. Faye-Petersen, S. Cliver, A. R. Goepfert, J. C. Hauth.
2006
. The Alabama Preterm Birth study: polymorphonuclear and mononuclear cell placental infiltrations, other markers of inflammation, and outcomes in 23- to 32-week preterm newborn infants.
Am. J. Obstet. Gynecol.
195
:
803
-808.
9
Romero, R., R. Gomez, F. Ghezzi, B. H. Yoon, M. Mazor, S. S. Edwin, S. M. Berry.
1998
. A fetal systemic inflammatory response is followed by the spontaneous onset of preterm parturition.
Am. J. Obstet. Gynecol.
179
:
186
-193.
10
Yoon, B. H., C. W. Park, T. Chaiworapongsa.
2003
. Intrauterine infection and the development of cerebral palsy.
BJOG
110
: (Suppl. 20):
124
-127.
11
Seong, H. S., Y. D. Kang, S. I. E. Lee, S.-S. Shim, R. Romero, B. H. Yoon.
2006
. Lack of evidence for microorganisms in most women with clinical chorioamnionitis: a need to revisit the clinical and microbiologic criteria for one of the most important obstetrical complications.
Am. J. Obstet. Gynecol.
195
:
S73
12
Lee, S. E., R. Romero, C. W. Park, J. K. Jun, B. H. Yoon.
2008
. The frequency and significance of intraamniotic inflammation in patients with cervical insufficiency.
Am. J. Obstet. Gynecol.
198
:
633.e1
-8.
13
Smulian, J. C., S. Shen-Schwarz, A. M. Vintzileos, M. F. Lake, C. V. Ananth.
1999
. Clinical chorioamnionitis and histologic placental inflammation.
Obstet. Gynecol.
94
:
1000
-1005.
14
Lee, S. E., R. Romero, H. Jung, C. W. Park, J. S. Park, B. H. Yoon.
2007
. The intensity of the fetal inflammatory response in intraamniotic inflammation with and without microbial invasion of the amniotic cavity.
Am. J. Obstet. Gynecol.
197
:
294.e1
-6.
15
Romero, R., F. Gotsch, B. Pineles, J. P. Kusanovic.
2007
. Inflammation in pregnancy: its roles in reproductive physiology, obstetrical complications, and fetal injury.
Nutr. Rev.
65
:
S194
-S202.
16
Romero, R., D. Lafreniere, G. W. Duff, N. Kadar, S. Durum, J. C. Hobbins.
1987
. Failure of endotoxin to cross the chorioamniotic membranes in vitro.
Am. J. Perinatol.
4
:
360
-362.
17
Kent, A. S., M. H. Sullivan, M. G. Elder.
1994
. Transfer of cytokines through human fetal membranes.
J. Reprod. Fertil.
100
:
81
-84.
18
Martinon, F., J. Tschopp.
2005
. NLRs join TLRs as innate sensors of pathogens.
Trends Immunol.
26
:
447
-454.
19
Takada, H., A. Uehara.
2006
. Enhancement of TLR-mediated innate immune responses by peptidoglycans through NOD signaling.
Curr. Pharm. Des.
12
:
4163
-4172.
20
Lu, Y. C., W. C. Yeh, P. S. Ohashi.
2008
. LPS/TLR4 signal transduction pathway.
Cytokine
42
:
145
-151.
21
Greten, F. R., M. Karin.
2004
. The IKK/NF-κB activation pathway: a target for prevention and treatment of cancer.
Cancer Lett.
206
:
193
-199.
22
Lindstrom, T. M., P. R. Bennett.
2005
. The role of nuclear factor κ B in human labour.
Reproduction
130
:
569
-581.
23
Lappas, M., M. Permezel, H. M. Georgiou, G. E. Rice.
2002
. Nuclear factor κB regulation of proinflammatory cytokines in human gestational tissues in vitro.
Biol. Reprod.
67
:
668
-673.
24
Koga, K., G. Mor.
2007
. Toll-like receptors and pregnancy.
Reprod. Sci.
14
:
297
-299.
25
Lappas, M., K. Yee, M. Permezel, G. E. Rice.
2005
. Sulfasalazine and BAY 11-7082 interfere with the nuclear factor-κB and IκB kinase pathway to regulate the release of proinflammatory cytokines from human adipose tissue and skeletal muscle in vitro.
Endocrinology
146
:
1491
-1497.
26
Lappas, M., M. Permezel, S. J. Holdsworth, G. Zanoni, A. Porta, G. E. Rice.
2007
. Antiinflammatory effects of the cyclopentenone isoprostane 15-A2-IsoP in human gestational tissues.
Free Radic. Biol. Med.
42
:
1791
-1796.
27
Vroom, F., E. N. van Roon, P. B. van den Berg, J. R. Brouwers, L. T. de Jong-van den Berg.
2008
. Prescribing of sulfasalazine, azathioprine and methotrexate round pregnancy: a descriptive study.
Pharmacoepidemiol. Drug Saf.
17
:
52
-61.
28
Elovitz, M. A..
2006
. Anti-inflammatory interventions in pregnancy: now and the future.
Semin. Fetal Neonatal Med.
11
:
327
-332.
29
Barham, D., P. Trinder.
1972
. An improved colour reagent for the determination of blood glucose by the oxidase system.
Analyst
97
:
142
-145.
30
Keelan, J. A., R. L. Zhou, M. D. Mitchell.
2000
. Activin A exerts both pro- and anti-inflammatory effects on human term gestational tissues.
Placenta
21
:
38
-43.
31
Simpson, K. L., J. A. Keelan, M. D. Mitchell.
1999
. Labour-associated changes in the regulation of production of immunomodulators in human amnion by glucocorticoids, bacterial lipopolysaccharide and pro-inflammatory cytokines.
J. Reprod. Fertil.
116
:
321
-327.
32
Namimatsu, S., M. Ghazizadeh, Y. Sugisaki.
2005
. Reversing the effects of formalin fixation with citraconic anhydride and heat: a universal antigen retrieval method.
J. Histochem. Cytochem.
53
:
3
-11.
33
Bennett, P. R., G. V. Chamberlain, L. Patel, M. G. Elder, L. Myatt.
1990
. Mechanisms of parturition: the transfer of prostaglandin E2 and 5-hydroxyeicosatetraenoic acid across fetal membranes.
Am. J. Obstet. Gynecol.
162
:
683
-687.
34
Nakla, S., K. Skinner, B. F. Mitchell, J. R. Challis.
1986
. Changes in prostaglandin transfer across human fetal membranes obtained after spontaneous labor.
Am. J. Obstet. Gynecol.
155
:
1337
-1341.
35
Viscardi, R. M., C. K. Muhumuza, A. Rodriguez, K. D. Fairchild, C. C. Sun, G. W. Gross, A. B. Campbell, P. D. Wilson, L. Hester, J. D. Hasday.
2004
. Inflammatory markers in intrauterine and fetal blood and cerebrospinal fluid compartments are associated with adverse pulmonary and neurologic outcomes in preterm infants.
Pediatr. Res.
55
:
1009
-1017.
36
Murthy, V., N. L. Kennea.
2007
. Antenatal infection/inflammation and fetal tissue injury.
Best Pract. Res. Clin. Obstet. Gynaecol.
21
:
479
-489.
37
Sadowsky, D. W., K. M. Adams, M. G. Gravett, S. S. Witkin, M. J. Novy.
2006
. Preterm labor is induced by intraamniotic infusions of interleukin-1β and tumor necrosis factor-α but not by interleukin-6 or interleukin-8 in a nonhuman primate model.
Am. J. Obstet. Gynecol.
195
:
1578
-1589.
38
Brennand, J. E., R. Leask, R. W. Kelly, I. A. Greer, A. A. Calder.
1995
. Changes in prostaglandin synthesis and metabolism associated with labour, and the influence of dexamethasone, RU 486 and progesterone.
Eur. J. Endocrinol.
133
:
527
-533.
39
Roseblade, C. K., M. H. Sullivan, H. Khan, M. R. Lumb, M. G. Elder.
1990
. Limited transfer of prostaglandin E2 across the fetal membrane before and after labor.
Acta Obstet. Gynecol. Scand.
69
:
399
-403.
40
Mitchell, M. D., V. Goodwin, S. Mesnage, J. A. Keelan.
2000
. Cytokine-induced coordinate expression of enzymes of prostaglandin biosynthesis and metabolism: 15-hydroxyprostaglandin dehydrogenase.
Prostaglandins Leukotrienes Essent. Fatty Acids
62
:
1
-5.
41
Rizek, R. M., C. S. Watson, S. Keating, H. H. Tai, J. R. Challis, A. D. Bocking.
2007
. 15-Hydroxyprostaglandin dehydrogenase protein expression in human fetal membranes with and without subclinical inflammation.
Reprod. Sci.
14
:
260
-269.
42
Weber, C. K., S. Liptay, T. Wirth, G. Adler, R. M. Schmid.
2000
. Suppression of NF-κB activity by sulfasalazine is mediated by direct inhibition of IκB kinases α and β.
Gastroenterology
119
:
1209
-1218.
43
Wahl, C., S. Liptay, G. Adler, R. M. Schmid.
1998
. Sulfasalazine: a potent and specific inhibitor of nuclear factor κB.
J. Clin. Invest.
101
:
1163
-1174.
44
Connell, W., A. Miller.
1999
. Treating inflammatory bowel disease during pregnancy: risks and safety of drug therapy.
Drug Saf.
21
:
311
-323.
45
Ishaq, S., J. R. Green.
2001
. Tolerability of aminosalicylates in inflammatory bowel disease.
BioDrugs
15
:
339
-349.
46
Norgard, B., A. E. Czeizel, M. Rockenbauer, J. Olsen, H. T. Sorensen.
2001
. Population-based case control study of the safety of sulfasalazine use during pregnancy.
Aliment. Pharmacol. Ther.
15
:
483
-486.
47
Rahimi, R., S. Nikfar, A. Rezaie, M. Abdollahi.
2008
. Pregnancy outcome in women with inflammatory bowel disease following exposure to 5-aminosalicylic acid drugs: a meta-analysis.
Reprod. Toxicol.
25
:
271
-275.
48
Norgard, B., L. Pedersen, L. A. Christensen, H. T. Sorensen.
2007
. Therapeutic drug use in women with Crohn’s disease and birth outcomes: a Danish nationwide cohort study.
Am. J. Gastroenterol.
102
:
1406
-1413.
49
Urquhart, B. L., J. A. Ware, R. G. Tirona, R. H. Ho, B. F. Leake, U. I. Schwarz, H. Zaher, J. Palandra, J. C. Gregor, G. K. Dresser, R. B. Kim.
2008
. Breast cancer resistance protein (ABCG2) and drug disposition: intestinal expression, polymorphisms and sulfasalazine as an in vivo probe.
Pharmacogenet. Genomics
18
:
439
-448.
50
Aye, I. L., J. W. Paxton, D. A. Evseenko, J. A. Keelan.
2007
. Expression, localisation and activity of ATP binding cassette (ABC) family of drug transporters in human amnion membranes.
Placenta
28
:
868
-877.