|Home | About | Journals | Submit | Contact Us | Français|
The amnion plays an important role during pregnancy and parturition. Though referred to as a single structure, this fetal tissue is regionally divided into placental amnion, reflected amnion, and umbilical amnion. Histological differences between placental amnion and reflected amnion led us to hypothesize that the amnion is biologically heterogeneous. The gene expression profiles of placental amnion and reflected amnion were compared in patients at term with no labor (TNL; n = 10) and in labor (TIL; n = 10). Real-time quantitative RT-PCR revealed a higher expression of IL1B mRNA in reflected amnion than in placental amnion in TNL cases but not in TIL cases. Extended screening using microarrays showed differential expression of 17 genes in labor, regardless of the region. Interestingly, 839 genes were differentially expressed between placental amnion and reflected amnion. Pathway analysis identified 19 signaling pathways, such as mitogen-activated protein kinase and transforming growth factor beta pathways, associated with region. Lipopolysaccharide (LPS) treatment of the amnion explants showed more robust activation of mitogen-activated protein kinase 3/1 (extracellular signal-regulated kinase 1/2) in placental amnion of TNL but not in TIL cases. Placental amnion from TNL and TIL cases showed a significant difference in the amplitude of IL1B mRNA induction by LPS. We report that the anatomical region has a substantial impact on the transcriptional program and the biological properties of the amnion. Labor-associated switching to a proinflammatory signature is a feature particular to placental amnion. The novel observations herein strongly suggest that the seemingly homogeneous amnion is biologically heterogeneous and compartmentalized, with implications for the physiology of pregnancy and parturition..
The human amnion is the inner layer of the fetal membranes that line the intrauterine surface. The surface area of the fetal membranes reaches up to 1876 ± 307 cm2 at term . Although human amnion is often referred to as a single, continuous structure, it is anatomically divided into three regions: placental amnion (amnion on the chorionic plate), reflected amnion (amnion of the extraplacental fetal membranes), and umbilical amnion (amnion covering the surface of the umbilical cord) . Not only is the amnion a strong, collagen-rich physical sheath encasing the amniotic cavity, it also plays a key role in the maintenance of pregnancy and parturition [3–5]. Several critical biological responses, including prostaglandin synthesis and release of proinflammatory cytokines, are enhanced in amnion during labor [6, 7]. However, whether the biological responses occurring across this substantially large mucosal surface are synchronized is yet undetermined.
The amniotic epithelial cells of the placental amnion tend to be more columnar compared to those of the reflected amnion . In addition, meconium exposure leads to variable degrees of reactive histological changes of the amniotic epithelial cells . Our empirical observations of the regional histological differences, especially between placental amnion and reflected amnion, in meconium-exposed placentas led us to hypothesize that human amnion is biologically heterogeneous depending on the anatomical region in the placenta. To test this hypothesis, we compared IL1B mRNA expression between placental amnion and reflected amnion and further analyzed the impact of spontaneous labor (term with no labor vs. term in labor) and region (placental amnion vs. reflected amnion) on the global gene expression profiles of the amnion at term.
Hematoxylin-eosin–stained histologic sections from 10 archival placentas with overt meconium-induced histologic changes of the amnion were retrieved from the bank of biological materials at the Perinatology Research Branch, NICHD/NIH. All the cases were evaluated for differences in meconium-induced reactive histologic changes between placental amnion and reflected amnion.
For microarray analysis, independent sets of placental amnion and reflected amnion were obtained from fresh singleton placentas of patients at term with no labor (TNL; n = 10) and in labor (TIL; n = 10). Approximately a quadrant of placental amnion was sampled in each placenta by blunt dissection. Reflected amnion samples were taken from regions at least 2 cm apart from the rupture sites and the placental margin, also by blunt dissection (Fig. 1A). Tissue samples were snap frozen and kept at −80°C until total RNA or protein isolation. Histology of the fetal membranes was confirmed for the absence of overt pathological findings such as inflammation and meconium staining. All patients provided written informed consent prior to sample collection. The utilization of tissue for research purposes was approved by the institutional review boards of the participating institutions.
Total RNA was isolated using Trizol. Isolated RNA samples were treated with DNase, and reverse transcription was performed using a SuperScript III reverse transcriptase (Invitrogen, Carlsbad, CA) and oligodT primers using 100 ng of total RNA. PCR analyses were performed with TaqMan Gene Expression Assays (IL1B: Hs00174097_m1, TGFBR1: Hs00610318_m1, PTGS2: Hs00153133_m1, BNC2: Hs00214187_m1; Applied Biosystems, Foster City, CA). RPLP0 was used for normalization. ABI 7500 FAST Real Time PCR system was used for PCR reaction. For BNC2 mRNA expression analysis in amnion, 5-μm-thick frozen sections on foil-covered glass slides (MicroDissect GmbH, Herborn, Germany) were used. Either amnion epithelium or mesodermal layer was selectively dissected using a Leica LMD6000 laser microdissection system (Leica Microsystems, Wetzlar, Germany). Total RNA was isolated using an RNeasy Mini kit (Qiagen, Valencia, CA).
For microarray analysis, 10 total RNA samples from each condition were randomly divided into three sets (three, three, and four samples each). The Affymetrix HGU-133 Plus 2.0 array platform was used to measure the gene expression levels in four groups, each with three pooled samples: TIL-placental amnion, TIL-reflected amnion, TNL-placental amnion, and TNL-reflected amnion. There were 47650 annotated probe sets on the array targeting 19886 unique genes. RNA quality and amount were analyzed using Agilent 2100 Bioanalyzer (Agilent Technologies, Foster City, CA). Synthesis of cDNA was performed using the Affymetrix GeneChip Expression 3′-Amplification One-Cycle cDNA Synthesis Kit (Affymetrix, Santa Clara, CA) with 3 μg of pooled total RNA samples. Double-stranded cDNA was purified using the Affymetrix GeneChip Sample Clean-up Module (Affymetrix). cDNA product was used for the synthesis of biotin labeled cRNA using the GeneChip IVT Labeling Kit (Affymetrix). The reaction was performed at 37°C for 16 h. Fifteen micrograms of labeled cRNA were fragmented at 94°C for 35 min using 5× Fragmentation Buffer included in the GeneChip Sample Clean-up Module (Affymetrix). Following the procedure, 10 μg of cRNA in 200 μl of hybridization cocktail were hybridized to an Affymetrix GeneChip Human Genome U133 Plus 2.0 array for 16 h at 45°C. The samples were washed and stained on the Genechip Fluidics Station 450 (Affymetrix). Arrays were scanned with the GeneChip Scanner 3000 (Affymetrix).
Affymetrix gene expression data was preprocessed using the RMA algorithm . This preprocessing method involves the following steps: 1) background correction of the probe intensities for nonspecific hybridization, 2) normalization of log-transformed data using a quantile-based approach, and 3) summarization of the expression values of all the probes for the same probe set into a single value per array. Data visualization was performed using principal components analysis to observe natural groupings (clusters) of the samples based on the expression profiles as previously described . Differential gene expression between TIL and TNL samples (regardless of region) and between the placental amnion and reflected amnion (regardless of labor condition) was assessed using a moderated t-test . The resulting P values, indicating the amount of evidence for differential expression between the groups, were adjusted using the false discovery rate (FDR) method  to allow for multiple hypothesis testing. Genes were called differentially expressed provided that they had a corrected P value less than 0.05 and changed in expression more than 2-fold.
Using the algorithm described  and implemented in the GOstats package  under R ( http://www.r-project.org ), gene ontology analysis was performed to identify significantly enriched biological processes, molecular functions, and cellular components. Pathway analysis was performed using an overrepresentation approach  on all metabolic and signaling KEGG pathways. For the KEGG signaling pathways, an impact analysis  using Pathway Express software ( http://vortex.cs.wayne.edu/projects.htm ) was also conducted. Unlike the overrepresentation approach, the impact analysis takes into account the gene-gene signaling interactions as well as the magnitude and direction of gene expression changes. Both gene ontology and pathway analysis results included correction for multiple testing by controlling the rate of positives among all positive findings.
Real-time PCR-based Human NFκB Signaling Pathway RT2Profiler PCR Array (Superarray Bioscience Corporation, Frederick, MD) was used to screen the expression of 84 key related genes according to the manufacturer's instructions. Briefly, 1 μg of pooled total RNA after genomic DNA elimination was reverse transcribed using a RT cocktail containing primer and reverse transcriptase. RT reaction was performed at 42°C for 15 min. PCR reaction was composed of initial denaturation at 95°C for 10 min followed by 40 cycles composed of 15 sec at 95°C and 1 min at 60°C. The list of individual genes in the array included IL1B, IL8, MYD88, TLR1, TLR2, TLR3, TLR4, TLR6, TLR8, TNF, TNFRSF10A, TNFRSF10B, IRAK1, IRAK2, CHUK, IKBKG, EDARADD, STAT1, BCL3, IL10, NLRP12, NFKBIA, CD27, TNFSF14, IKBKB, NFKB1, RELA, CD40, F2R, FASLG, HTR2B, LTBR, SLC20A1, TICAM2, TNFRSF1A, TNFSF10, TICAM1, EDG2, GJA1, HMOX1, RHOA, IKBKE, RIPK1, TBK1, BCL10, BIRC2, NOD1, CASP1, CASP8, CFLAR, SLC44A2, FADD, MALT1, PPM1A, REL, TRIM13, TMED4, TRADD, AGT, CFB, ICAM1, CSF2, CSF3, IFNA1, IFNB1, IFNG, LTA, CCL2, IL6, IL1A, IL1R1, RAF1, TLR7, TLR9, AKT1, MAP3K1, ATF1, EGR1, ELK1, FOS, JUN, NFKB2, RELB, and TNFAIP3.
Amnion explants were obtained from term placentas of TNL and TIL cases (n = 3 for each group). Pieces of the amnion measuring 2.5 cm2 were kept for an hour in 5A McCoy media containing 0.5% fetal bovine serum and antibiotics. These explants were then treated with lipopolysaccharide (LPS; Escherichia coli 0111:B4, L2630; Sigma, St. Louis, MO) at a concentration of 1 μg/ml for a given period of time (0, 0.5, 2, 4 h) and harvested for total RNA or protein isolation. Placental amnion explants from TNL and TIL cases (n = 3 for each group) were also independently treated with LPS for 4 h to compare increases in IL1B mRNA expression.
Total protein was isolated using RIPA lysis buffer containing both proteinase inhibitor and phosphatase inhibitor cocktails. A total of 40 μg of protein were electrophoresed in a 12% SDS-polyacrylamide gel and electroblotted onto a PVDF membrane. Blotted proteins were probed with different antibodies to phospho-p44/42 MAPK (Thr202/Tyr204; mouse monoclonal, E10; Cell Signaling Technology, Inc., Danvers, MA), TGFB1/TGFB2/TGFB3 (rabbit polyclonal; Cell Signaling Technology), TGFBR1 (rabbit polyclonal, T-19; Santa Cruz Biotechnology, Inc., Santa Cruz, CA), TGFBR2 (mouse monoclonal, C-4; Santa Cruz Biotechnology), and phospho-SMAD1/SMAD5/SMAD9 (rabbit polyclonal; Cell Signaling Technology).
Quantitative RT-PCR data were analyzed using a Mann-Whitney U-test and a Wilcoxon signed ranks test. The SPSS version 12.0 (SPSS, Inc., Chicago, IL) statistical package was used, and a P value of less than 0.05 was considered significant. PCR array data were analyzed using a Student t-test.
Histological evaluation of sections obtained from 10 formalin-fixed, archival placentas with meconium-associated reactive changes revealed rather prominent changes in placental amnion, while only minimal changes in reflected amnion of all cases. Placental amnion epithelial cells generally showed relocation of the nuclei to the apical region, elongation, and increased eosinophilia of the cytoplasm and pseudostratification of variable degrees (Fig. 1B). On the other hand, reflected amnion epithelial cells were flat and devoid of recognizable reactive changes (Fig. 1B).
To further address biological differences between placental amnion and reflected amnion, total RNA samples from 10 TNL and TIL cases were analyzed. All the cases were histologically confirmed for the absence of pathological lesions, such as inflammation and meconium staining. Quantitative RT-PCR analysis of IL1B mRNA revealed a 23.7-fold higher expression in placental amnion of TIL cases compared to that of TNL cases (P < 0.005; Fig. 1C). However, there was no difference in IL1B mRNA expression in reflected amnion between the two groups (Fig. 1C). Similarly, IL1B mRNA expression in reflected amnion was higher than that of placental amnion in TNL cases (P < 0.01), while no difference was noted between placental amnion and reflected amnion in the TIL group. The overall findings indicated that placental amnion showed a more dramatic increase in IL1B mRNA expression in labor at term than in reflected amnion.
Based on the clear differences in the reactive changes to meconium exposure and IL1B mRNA expression with the presence of labor and notably with region (placental amnion vs. reflected amnion), we performed a microarray study to assess the impact of labor and region on the global gene expression of the amnion. Principal components analysis of gene expression profiles showed that the main direction of variability with the gene expression data set (the first principal component) was related to regional differences (Fig. 2A). The differences associated with labor were captured by the second principal component (Fig. 2A). The “volcano plots” in Figure 2B show the differential expression of all the annotated probe sets on the Affymetrix HGU-133 Plus 2.0 array. A heat map in Fig. 2C shows clustering of both genes and samples. A comparison of the placental amnion and reflected amnion obtained from TNL and TIL cases revealed subsets of differentially expressed genes associated with the presence of labor and region. For the placental amnion and reflected amnion comparison, there were 1124 probe sets with adjusted P values less than 0.05 and a fold change greater than 2, which corresponded to 839 unique genes (Table 1). Gene ontology analysis revealed 56 enriched biological processes, including development and biological adhesion. Pathway Express analysis identified 19 signaling pathways associated with regional differences, such as cell adhesion molecules, JAK/STAT signaling, mitogen-associated protein kinase signaling, and transforming growth factor-β signaling pathways (Table 1). In a TIL-TNL comparison, there were 21 probe sets corresponding to 17 unique genes with differential expression. Among those 17 unique genes, seven (IL1A, IL1B, IL6, TNF, CXCL1, CXCL2, CXCL3) were cytokine/chemokine genes (Table 2). Gene ontology analysis showed enrichment of 106 biological processes, including immune response and inflammatory response associated with differential gene expression with labor, and the pathway analysis using Pathway Express identified five significant pathways, including cytokine-cytokine receptor interaction and apoptosis (Table 2).
PCR array analysis focused on 84 genes involved in NF-κB signaling pathway revealed consistent results with microarray analysis: a higher expression of IL1B mRNA in reflected amnion than in placental amnion in the TNL group but not in the TIL group, and up-regulation of IL1B, IL6, and IL8 mRNA expression with labor. The expression of IRAK2 (interleukin-1 receptor-associated kinase 2) mRNA in placental amnion and reflected amnion, and TLR4 mRNA in reflected amnion were additionally found to increase with labor (Table 3). For validation of the microarray results, we analyzed the expression of TGFBR1 mRNA in placental amnion and reflected amnion. Protein expressions of TGFB1/TGFB2/TGFB3, TGFBR1, TGFBR2, and phosphorylated SMAD1/SMAD5/SMAD9 were also assessed. Expression of TGFBR1 mRNA determined by qRT-PCR was significantly higher in placental amnion of TNL cases than in reflected amnion, as was found by microarray analysis, and TIL cases showed a similar tendency (P = 0.066). However, TGFB1/TGFB2/TGFB3, TGFBR1, TGFBR2, and phosphorylated SMAD1/SMAD5/SMAD9 protein expressions were higher in reflected amnion than in placental amnion. Therefore, TGFBR1 mRNA and protein expressions were found to show paradoxical expression patterns (Fig. 3, A and B). Transcription factor BNC2 (Basonuclin-2) was also among the differentially expressed genes, and qRT-PCR of laser-microdissected samples revealed BNC2 mRNA expression in the amnion epithelial layer, while it was undetected in the mesodermal layer (Fig. 3C). Review of PTGS2 mRNA, which was not among the differentially regulated genes associated with labor but was expected to increase, revealed a nominal P value of less than 0.05 but not after FDR correction (data not shown). Confirmative qRT-PCR demonstrated increased expression in both placental amnion (P = 0.054) and reflected amnion (P = 0.001) with labor. Interestingly, PTGS2 mRNA expression was higher in placental amnion compared to reflected amnion in TNL but not TIL cases (P = 0.017) (Fig. 3D).
Based on the evidence that gene expression profiles of placental amnion and reflected amnion are different, we compared the proinflammatory response of placental amnion and reflected amnion in TNL and TIL cases. Treatment with LPS induced phosphorylation of MAPK3/MAPK1 (ERK 1/2) in placental amnion and reflected amnion of all cases tested. Interestingly, placental amnion of TNL cases but not TIL cases showed more robust and sustained activation of MAPK3/MAPK1 when compared to corresponding reflected amnion (Fig. 4A). Explant cultures of placental amnion from TNL and TIL cases revealed 4.9-fold and 1.7-fold increases in IL1B mRNA expression after 4 h following LPS treatment (Fig. 4B), indicating that the biological responsiveness of placental amnion differs before and after labor at term.
The principal findings of this study are that 1) human placental amnion and reflected amnion are two biologically heterogeneous regions based on the gene expression profile patterns, 2) biological responses of the placental amnion were stronger before rather than after labor at term, and 3) the cytokine-cytokine receptor interaction is a key biological circuit in the amnion during spontaneous labor at term.
Our observation that placental amnion and reflected amnion transcriptomes differ substantially is novel, and many of the differentially expressed genes between placental amnion and reflected amnion are functionally relevant. For example, HLA-G mRNA expression is higher in reflected amnion. Amniotic epithelial cells express HLA-G, and amniotic substrate regulates HLA-G expression in corneal and limbal epithelial cells in vitro [16, 17]. HLA-G, a less polymorphic MHC Class I molecule, exerts immunomodulatory functions such as apoptosis of CD8+ T cells and promotion of Th2 response [18, 19], and reflected amnion has a higher chance of exposure to maternal immunocytes in the decidua. Transcription factor BNC2 mRNA expression was higher in placental amnion. Basonuclin expression is restricted to the stratified squamous epithelium (e.g., epidermal, corneal, esophageal, and vaginal epithelia) and the germ cells of the gonads [20, 21]. Expression of BNC2 mRNA in the amnion epithelial cells and its higher expression in placental amnion are also novel findings, most likely related to the simple squamous nature of amniotic epithelium and the proximity of placental amnion to the fetal skin. This suggests that differential expression of certain genes is related to the difference in developmental characteristics between placental amnion and reflected amnion. Other anatomical factors that might explain the differential expression of the genes include the influence of chorion laeve and decidua on reflected amnion and, likewise, the influence of the chorionic plate and fetal vessels on placental amnion . At the cellular level, amnion is composed of different types of cells, such as amnion epithelial cells, fibroblasts/myofibroblasts, and resident macrophages . It was shown that PGHS-2 (PTGS2) is differentially regulated by dexamethasone in amnion epithelial cells and fibroblasts . Therefore, potential differences in cellular composition between placental amnion and reflected amnion might also significantly contribute to the differential gene expressions.
Previous comparisons of local gene expression patterns of the amnion have focused primarily on differences between the regions in reflected amnion. Regional differences in IL1B, IL6, and IL8 mRNA expression in samples from three areas—the zone of altered morphology, midzone, and periplacental regions—were not found within reflected amnion . On the other hand, a site-specific augmentation pattern of prostaglandin synthesis in the fetal membranes during labor was identified between the upper and lower part (adjacent to cervical canal) of the amnion . Few studies have considered differences in selective biochemical properties between placental amnion and reflected amnion. The level of vasoactive PTHLH (parathyroid hormone-like hormone) mRNA in placental amnion was found to be higher than in reflected amnion, and placental amnion released a significantly larger amount of PTHLH than reflected amnion in vitro [22, 27]. Placental amnion directly covers fetal chorionic vessels in the chorionic plate. Therefore, placental amnion-derived PTHLH has a higher probability of effectiveness as a vasorelaxant of chorionic vessels, which may be biologically relevant. This differential expression pattern of PTHLH mRNA was also found in our microarray analysis.
Amnion has been used for the reconstruction of ocular surface disorders [28, 29]. Its major wound healing properties are, in part, attributed to anti-inflammatory TGFB content, and a comparison of the expression of TGFB isoforms in the amnion revealed that TGFB1 is expressed most abundantly . We show that the expression and intensity of TGFB signaling proteins (TGFB1/TGFB2/TGFB3, TGFBR1, TGFBR2, SMAD1/SMAD5/SMAD9) reveal differences between placental amnion and reflected amnion. SMAD1/SMAD5/SMAD9 proteins are activated by bone morphogenetic proteins (BMPs), and, as we previously reported, BMP2 is expressed in the fetal membranes . It is interesting that the levels of expression of TGFBR1 mRNA are paradoxically higher in placental amnion.
Involvement of both physiologic and pathologic pro-inflammatory cytokine response in the fetal membranes in both term and preterm parturition is well known [32–34]. The pattern of IL1B and other cytokine/chemokine mRNA expression in this study indicates that placental amnion is the region where their expression increases most substantially with labor. A recent study reported increased expression of the genes involved in oxidative stress and NF-κB pathway activation, including TNF and IL1B mRNA in the placental villous tissues after labor . Up-regulation of proinflammatory cytokine expression is a common feature of the amnion and the villous placenta, while that of oxidative stress response seems to be more pertinent to the villous placenta. Of particular interest is the fact that placental amnion obtained from TNL cases behaves quite differently from corresponding reflected amnion or placental amnion obtained from TIL cases. Placental amnion of TNL cases showed a stronger LPS-induced response (MAPK3/MAPK1 activation and IL1B mRNA expression). The robust nature of this proinflammatory response in placental amnion might explain the more prominent meconium-induced reactive histological alterations and earlier involvement of subchorionic space of the chorionic plate in acute chorioamnionitis . It is also of note that the IL1B mRNA expression data clearly indicate that response to LPS in placental amnion markedly decreases after spontaneous labor at term.
We report herein that human amnion is heterogeneous. The findings in the present study have many implications and identify many potential target genes and pathways worthy of further investigation. Meticulous molecular dissection of placental amnion and reflected amnion would provide more valuable information on the role for this crucial fetal structure in human pregnancy and parturition. Our findings suggest that placental amnion is an important subject for further research in the biology of the fetal membranes.
1Supported by the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development, NIH, DHHS.