| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
(IFN-) and Its Receptor IFN-R2 Fail to Switch from Early Hypoxic to Late Normotensive Development in Preeclampsia
Harris Birthright Research Center for Fetal Medicine (S.B., A.S., A.E.C., A.P., S.C., K.N.) and Department of Pathology (J.M.), King’s College Hospital Medical School, London SE5 9RS, United Kingdom
Address all correspondence and requests for reprints to: Dr. Subhasis Banerjee, Harris Birthright Research Center for Fetal Medicine, King’s College Hospital Medical School, Denmark Hill, London SE5 9RS, United Kingdom. E-mail: dr_sbanerjee{at}hotmail.com.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
(IFN-), TNF-
, and IL-12] and Th2 (IL-4 and IL-10) cytokines have opposite effects on human pregnancy. Leukemia inhibitory factor (LIF) promotes embryo implantation and sustains pregnancy, whereas IFN- and TNF- are detrimental to pregnancy. Both IFN- and LIF are produced by maternal cells and tissues at the fetal-maternal interface, whereas the IFN- receptors (IFN-R1 and IFN-R2) and LIF receptor are abundantly expressed on the surface of placental trophoblasts. The effect of IFN- on T lymphocyte activation is influenced by the relative membrane density of its two receptors, particularly IFN-R2. In this study we report that in PE (25–40 wk gestation) and PE complicated by IUGR, IFN-R2 protein expression is severely down-regulated and is similar to that observed in early placenta (7–10 wk gestation) developing under low O2 tension. IFN- production was found to be inversely related to the IFN-R2 protein expression, and LIF receptor protein expression in PE mimicked that in early placental development. These results show that in PE, placental trophoblasts fail to establish an early to late switch with respect to IFN- and IFN-R2 expression. This supports the hypothesis that trophoblasts control the polarization of maternal immune effectors and cytokine profiles at the fetal-maternal interface that could be subject to oxidative stress in PE. |
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Human pregnancy has two aspects of development, the growth and differentiation of trophoblasts for optimum uterine implantation and the ability of the placenta to prevent activation of maternal leukocytes and macrophages after recognition of placental/fetal antigen. This is achieved specifically by expression of a special class of major histocompatibility complex (MHC) antigen, such as human leukocyte antigen-G in trophoblasts (3, 4, 5, 6), and nonspecifically by altering the reactivity of decidual lymphocytes via progesterone and other nonspecific immunosuppressors (7, 8, 9, 10). Based on their reactivity to alloantigens, lymphocytes differentiate into T helper type 1 (Th1) and Th2 phenotypes. Th1 responses are usually associated with cell-mediated immunity (intracellular pathogens), whereas Th2 responses are antibody-mediated humoral responses against extracellular pathogens (11, 12, 13).
Biochemical (14, 15), immunological (11), and genetic (1) hypotheses have been proposed to explain the superficial placental invasion in PE (16). For instance, it is well documented that continued hypoxic development of the fetus and placenta in PE mimics the hypoxia/reoxygenation injury that results in superoxide-induced tissue damage and systemic toxemia (14, 17). Immunologically, a failure to switch from Th1 to Th2 cytokine profiles in response to placental alloantigen at the fetal-maternal interface could be detrimental to pregnancy. Spontaneous recurrent abortion in human (8) and mouse (18, 19) models has been linked to vascular endothelial attack (20) and trophoblast degeneration (18) due to the synergistic effects of interferon- (IFN-) and TNF-.
IFN- may be considered a polypeptide hormone primarily because of its ability to transduce signals from one cell type to the other through diverse effects. The wide range of cellular responses to IFN- include proliferation, apoptosis, leukocyte-endothelial interactions, up-regulation of inducible nitric oxide synthase production in macrophages (thereby increasing reactive oxygen species production), tryptophan metabolism, mediation of antigen presentation by class I and II MHC complexes, and regulation of numerous genes whose functions are yet to be identified (21, 22).
IFN- functions by binding to the receptor heterodimers IFN-R1 and -2; IFN-R1 has a ligand-binding site, and IFN-R2 facilitates signaling via Janus kinase-signal transducer and activator of transcription pathways (23, 24). Recent genetic and immunological experiments have established that the proliferative or cytotoxic effect of IFN- is primarily determined by the relative density of IFN-R1 and -2 on the T cell surface (25, 26, 27).
Despite extensive immunological studies, very little is known about the placental regulation of expression of IFN- and its receptors in human pregnancy, except that IFN-R1 is localized in placenta throughout pregnancy (28). Similar data on IFN-R2 are not available. One of the outstanding questions in human pregnancy pathology is whether the cytokine profile at the fetal-maternal interface is controlled maternally or whether the innate immunity of the trophoblasts regulates the polarization of maternal immune effectors. Interestingly, decidual infection with Listeria monocytogenes is resolved in pregnant mice by local recruitment of neutrophils. In this system the trophoblasts function as a component of the innate immunity to infection by secreting neutrophil chemoattractants and macrophage inflammatory protein-2 in response to uterine colony-stimulating factor-1 (19). Based on the hypothesis that at a given stage of pregnancy, IFN- signaling in trophoblasts is directly related to the controlled expression of these receptors in placenta, vascular endothelial cells (VEC), and smooth muscle cells (SMC), we have attempted to examine the relative chorionic expression of IFN-R1 and IFN-R2 during the course of normal and PE pregnancies.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This study was approved by the local ethics committees of St. George’s and King’s College Hospitals (London, UK), and written consent was obtained from patients before the collection of samples. Placental tissue was obtained from patients undergoing termination of pregnancy at a gestational age range of 7–12 wk. Full-term and PE placental tissues were obtained after vaginal delivery or cesarean section. PE was defined as gestational hypertension with proteinurea that developed for the first time during pregnancy. Gestational hypertension was defined as systolic blood pressure of 140 mm Hg or greater and diastolic blood pressure of 90 mm Hg or greater on at least two occasions, with the exception of patient 2, who had systolic and diastolic blood pressures of 140/89 mm Hg (Table 1
). Proteinurea was defined as 300 mg total protein or more in a 24-h collection or more than 2+ by dipstick on two consecutive measurements at least 4 h apart. IUGR was defined as a birth weight below the fifth percentile of the reference group (Table 1). Patients 13, 15, and 16 had PE with IUGR, and patient 17 had PE with hemolysis, elevated liver enzymes, and low platelets syndrome (HELLP) syndrome. Approximately 2–5 g placenta tissue were dissected and washed extensively in chilled PBS (Ca+2/Mg+2-free; Invitrogen Life Technologies, Inc., Carlsbad, CA), and villous material was isolated by additional dissection in a 60-mm sterile petri dish placed under a stereomicroscope. Samples were stored in RNALater (Ambion, Huntingdon, UK) overnight at 4 C before transfer to storage at –20 C.
|
RNA extraction was carried out by two alternative methods: 1) tissues were homogenized in TRIzol (Invitrogen Life Technologies, Inc.) using the manufacturer’s standard protocol for total RNA preparation and 2) using an RNA extraction kit (Qiagen, West Sussex, UK). After TRIzol extraction, RNA samples were stored as ethanol precipitates at –70 C. The RNA eluted from Qiagen columns was stored in H2O at –70 C. Extracted RNA was treated with deoxyribonuclease I (Sigma-Aldrich Corp., St. Louis, MO) in 30 mM Tris-Cl (pH 7.8), 50 mM NaCl, and 10 mM MgCl2, in the presence of 40 µU RNaseOUT (Invitrogen Life Technologies, Inc.) for 30 min before phenol-chloroform extraction and ethanol precipitation. First strand cDNA was synthesized from 5 µg total RNA using Superscript II reverse transcriptase (Invitrogen Life Technologies, Inc.), primed with oligo(deoxythymidine)12–18 (Invitrogen Life Technologies, Inc.) according to the manufacturer’s recommended protocol, then treated with ribonuclease H (Invitrogen Life Technologies, Inc.) for 30 min before heat inactivation (70 C for 15 min) and PCR amplification.
Quantitative PCR
Quantitative PCR was performed using the Light Cycler RNA amplification system in glass capillaries and fluorescence-based hybridization detection format (Roche, Mannheim, Germany). Two companies were used for primer design and synthesis (TIBMolBiol, Berlin, Germany; and Metabion GmbH, Martinsried, Germany). The Light Cycler PCR detection method employed here is based on two sequence-specific labeled oligonucleotide probes that hybridize to the internal region of the cDNA products during the annealing phase of the PCR cycles. One probe is labeled at the 5' end with LC-Red 640 or LC-Red 705, and the 3'-end is blocked by phosphorylation; the second probe is fluorescein-labeled at the 3'-end. The probes anneal to the opposite strand of the amplified target DNA in a head to tail fashion separated by a few base pairs; a gap is required for optimum interaction of the fluorophores. When the donor fluorescein dye attached to one probe is stimulated by the light source in the instrument, it activates the reporter dye (LC-Red 640 or LC-Red 705) on the other probe by fluorescence resonance energy transfer, which emits red light that is measured by the machine during every cycle. The criteria used to establish target DNA sequences for hybridization were 1) central sequences not adjacent to either end of the amplicon, 2) melting temperatures of at least 5 C higher (i.e. 61 C in our experiments) than the PCR primers, and 3) absence of complementary or repetitive DNA sequences.
In general, assays were carried out in duplex where both the experimental sample and an internal control (ß-actin (ACTB) and HPRT) were run in the same reaction. The reporters LC-Red 640 and LC-Red 705 were employed to generate hybridization probes for experimental and internal controls, and the amplification for each cDNA was recorded by dual color detection. A color compensation file was created according to the manufacturer’s instructions (Roche) and was compensated for PCR cycles in duplex during each run. In some experiments the control reactions were run at the same time as the test samples under the same reaction conditions, but as single reactions. The method was highly reproducible when different external controls were used. To establish that there was no cross-contamination, negative controls (a full reaction without cDNA) were run in each experiment.
Standard DNA was synthesized in bulk from cDNA by conventional PCR, gel-purified, and quantified before serial dilutions. Where duplex reactions were used, both the test and control gene fragments were included in one standard dilution series. A standard series was included in each reaction and used to generate a standard curve for quantification of test samples synthesized in the PCRs.
Cross-contamination was avoided by sequentially adding in the following order: water, reaction Master-Mix (containing enzyme, Mg2+ and PCR buffer; Light Cycler RNA amplification kit, Roche), and cDNA to a final volume of 10 µl. Melting curve analysis was employed in the initial optimization stages of assay development, such as the determination of optimum Mg2+ concentrations. The analysis mode set at quantification included initial denaturation at 95 C for 10 min, followed by 40 cycles consisting of the following parameters for segments 1–3: target temperature, 95, 55, and 72 C, respectively; incubation time, 10, 7, and 12 sec, respectively; transition rate, 20, 20, and 10 C/sec, respectively; and a single acquisition mode at segment 2. The PCR primers, hybridization probes, and amplicon lengths are shown in Table 2. Data were automatically collected and filtered to remove background by the Light Cycler software, which set the crossing point for all of the different reactions against the standard curve. Data were transferred to Microsoft Excel (Microsoft Corp., Redmond, WA) for additional analysis.
|
Proteins from placental villous tissue, dissected as described above, were extracted in T-PER (Perbio, Helsinborg, Sweden) following the manufacturer’s recommended protocol. Protein concentrations were measured by Bradford assay (Bio-Rad Laboratories, Hemel Hempstead, UK), and 15–20 µg extract were resolved by denaturing electrophoresis through 8–10% SDS-PAGE under reducing conditions and transferred to polyvinylidene difluoride membrane (Immobilon P, Millipore Corp., Watford, UK). The blots were preincubated with one of three different blocking agents: 1% (wt/vol) nonfat skimmed milk, 1% (wt/vol) BSA (fraction V), or 1% (wt/vol) casein-based blocking buffer (Sigma-Aldrich Corp.). Immunoreactions were carried out using optimized dilutions of antibody, and the reactions were detected by chemiluminescence using ECL Plus reagents (Roche). Blots were stripped before reprobing with ß-actin antibody as previously described (29, 30). Some blots were stained with 0.25% (wt/vol) Coomassie Brilliant Blue R (Sigma-Aldrich Corp.) in 40% (vol/vol) methanol, destained, and photographed to ensure complete protein transfer to membranes. The primary and secondary antibodies used were as follows: affinity-purified goat antihuman IFN-R1; AF 773 and IFN-R2; AF 673, affinity-purified goat antihuman leukemia inhibitory factor receptor (LIFR); AF-249-NA, goat antihuman IFN-; AF-285-NA (all from R&D Systems, Oxford, UK); goat antimouse IgG (H+L) horseradish peroxidase-conjugated (Chemicon International, Inc., Temecula, CA); anti-ß-actin, clone AC-15 (Sigma-Aldrich Corp.); and affinity-purified antigoat IgG-horseradish peroxidase-conjugated; ab6741 (Abcam, Cambridge, UK). IFN-R1 and -R2 are 472- and 226-amino acid proteins, respectively. The anti-IFN-R1 recognizes fully glycosylated 85–110K molecular mass (Mr) and deglycosylated 49.8K Mr bands, whereas anti-IFN-R2 recognizes fully glycosylated 61–67K Mr and deglycosylated 34.8K Mr bands. In placenta, the major bands detected were unglycosylated 49.8K Mr by anti-IFN-R1, glycosylated 61–67K Mr by IFN-R2, 110–150K Mr by anti-LIFR (29), and 17.2K to approximately 21K (glycosylated) Mr by anti-IFN.
Immunohistochemistry
Placental tissues from 7-, 8-, and 12-wk gestation and full-term pregnancies, PE and PE with IUGR were fixed in 4% (vol/vol) freshly prepared paraformaldehyde (Sigma-Aldrich Corp.) in PBS (Ca+2/Mg+2-free) for 4 h at room temperature. The samples were cryoprotected, embedded in paraffin, and frozen in liquid nitrogen. Immunostaining for IFN-R1 and -2 was carried out on 7-µm paraffin sections. Antihuman IFN-R2 antibody was a mouse monoclonal 17662-17G clone 9.PL.3 (U.S. Biologicals, Swampscott, MA) and received no pretreatment before staining. Antihuman IFN-R1, an affinity-purified goat polyclonal antibody (AF 673, R&D Systems, Oxford, UK), required pressure cooking for 3 min in 10 mM citrate buffer, pH 6.0, before staining. Sections were then incubated for 60 min in primary antibody at room temperature. This was followed by incubation in either rabbit antigoat antibody (DakoCytomation, Copenhagen, Denmark) or rabbit antimouse (DakoCytomation) biotinylated antibody. The sections were incubated in either goat peroxidase anti peroxidase (DakoCytomation) or streptavidin-biotin complex (DakoCytomation), both of which were conjugated to horseradish peroxidase. The reaction was visualized in both cases using 3,3'-diaminobenzidine (DakoCytomation) as a chromogen. In all cases the washes were carried out in Tris-Cl-buffered saline, pH 7.6. Sections were counterstained with freshly prepared and filtered Mayer’s hematoxylin (Merck & Co., Poole, UK), followed by rinsing in 1% (vol/vol) acid-alcohol. The slides were dehydrated in an ascending ethanol series before mounting. For negative controls, the primary antibody was replaced with nonspecific isotype controls (negative control reagents, IgG1, monoclonal; DakoCytomation). Sections were visualized in a Nikon E400 microscope (Melville, NY), and images were captured using the Lucia version 4.21 system (Laboratory Imaging, Ltd., Hostivar, Czechoslovakia) for processing and analysis.
Densitometry and data analysis
Densitometry of autoradiograms was performed using a 1D-Multi Lane densitometry program in an AlphaImager (1220v5.5, Alpha Innotech Corp., San Leandro, CA). For autodetection of pixel density in a peak, the scan width was maintained constant, and the baseline values were set by background subtraction for each exposure. Scan data (experimental and ß-actin control) were transferred to Microsoft Excel Worksheet, where the pixel density of each experimental lane was normalized to its corresponding ß-actin value. Each experiment was repeated at least twice, and average values for each data point were plotted. The means, SDs, and variance (ANOVA) for each dataset were computed using Analysis ToolPak software (Excel, version 2000; Microsoft Corp., Seattle, WA). Values are shown as the mean ± SEM. To compare gene expression (RNA and protein), the average values were compared. A value for the level of significance (P value) was calculated using the Poisson statistic. P < 0.05 was considered significant.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
R and LIFR in normal and PE pregnancies
Previous immunohistochemical studies demonstrated that IFN-Rs were expressed in early and late human placenta, whereas the expression of the ligand IFN- was restricted to early pregnancy (28). The LIFR has been shown to be expressed by trophoblasts throughout pregnancy (29). To examine the transcriptional regulation of these receptors in normal pregnancies and in PE with or without IUGR, mRNA synthesis from their corresponding genes was measured by quantitative PCR (Fig. 1A). IFN-R1 transcription was highest in normal early pregnancy and was reduced by 33.6% and 25.7% in normal late (P = 0.06) and PE (P = 0.09) pregnancies, respectively. However, transcription from the IFNGR2 and LIFR genes showed a different pattern when early and late normal and PE expression were compared. Transcriptional expression from the IFNGR2 gene in normal late pregnancy was increased (34.6%; P = 0.072) compared with that in normal early pregnancy, but in PE with or without IUGR, it was only marginally elevated (21.4%; P = 0.095) compared with that in normal early pregnancy. Transcriptional expression from the LIFR gene appeared unchanged in a comparison of early and late normal pregnancy and showed a very slight reduction (9%) in PE that was not significant (P = 0.125). Thus, although levels of transcription from the IFNGR1 gene in PE resembled those found in late normal pregnancy, transcriptional expression from the IFNGR2 gene in PE more closely resembled levels found in early normal pregnancy, and transcription from the LIFR gene did not appear to be affected by either pregnancy stage or PE. The ratio of IFN-R1 to IFN-R2 transcription (IFN-R1:IFN-R2) was calculated for each normal pregnancy stage and for PE. In early normal pregnancy this ratio was 1:0.5, which is significantly different from that of late (1:1; P = 0.037) normal, but not PE (1:0.8; P = 0.063), pregnancies. These studies suggest that the placental regulation of IFNGR and LIFR genes in PE (25–40 wk gestation) is comparable to that in late human pregnancies.
|
R2 protein is severely inhibited in PE and resembles that in early normal pregnancy
The density of functional receptor proteins is often determined by the stability of their mRNA transcripts, posttranslational modifications, and transport to the surface of the cells (31). In addition, one of the characteristic features of cytokine receptors is that their function, which is dependent on their membrane distribution, is often independent of their transcriptional regulation and cytoplasmic accumulation (32, 33, 34). To test whether transcription and protein synthesis are coupled, the expression of IFN-R1 and IFN-R2 proteins was measured. Proteins from early and late normal pregnancies and from PE with or without IUGR were extracted from chorionic villous tissues, electrophoretically separated, blotted, and immunoreacted with antibodies specific to each receptor. To directly compare the data, proteins from normal pregnancies and PE were electrophoretically separated simultaneously, and protein transfer conditions to Immobilon-P membrane (Millipore, Billerica, MA), antibody concentrations, immunoreaction time, and autoradiographic exposure for the detection of chemiluminescence signals were maintained constant. This procedure was followed for immunostaining of stripped blots with ß-actin antibody. The ß-actin signals acted as internal loading controls and were used to normalize IFN-R1 and IFN-R2 protein expression signals in the corresponding lanes. In contrast to the reduction in transcription from the IFNGR1 gene noted in late normal compared with early normal pregnancy (Fig. 1A
), the protein expressions of IFN-R1 in early and late normal pregnancies were comparable (Fig. 1, B and D; P = 0.136). However, the protein expression profile of the IFNGR2 gene for both early and late normal pregnancies (Fig. 1, B and D) reflected transcription from this gene (Fig. 1A) and showed a significant elevation of protein expression in late normal pregnancy compared with early normal pregnancy (Fig. 1, B and D). Comparing the two receptors, IFN-R2 protein expression was approximately 46.7% that of IFN-R1 (P = 0.018) in early pregnancy and 72.5% that of IFN-R1 (P = 0.09) in late pregnancy. In normal pregnancy, the protein expression results were comparable to the mRNA data (Fig. 1A and below), but in PE with or without IUGR, protein expression did not reflect changes at the transcriptional level. In PE, IFN-R1 protein expression was markedly elevated (Fig. 1, C and D) and failed to correlate with transcriptional activation (Fig. 1A). In addition, the up-regulation of IFN-R2 protein expression observed in late normal pregnancy was severely inhibited in PE and resembled that in early pregnancy (Fig. 1, B–D). These results show that the early to late transition in chorionic expression of IFN-R2 proteins does not occur in pregnancies with PE with or without IUGR.
The quantification of IFN-R1 and IFN-R2 proteins was carried out by immunostaining blots with individual antibodies (Fig. 1, B–D). Although these experiments were internally controlled using ß-actin expression levels as a standard, it could be argued that unknown variables might have compromised the results. Because the antibodies against human IFN-R1 and IFN-R2 do not cross-react and recognize protein species of relative Mr 49K and 62–67K, which are separable (35), we performed an additional control for the experiments shown in Fig. 1, B–D, by probing the same blot with an equimolar mixture of anti-IFN-R1 and anti-IFN-R2 antibodies at two concentrations (Fig. 2, A and B). The expression levels of IFN-R1 and IFN-R2 proteins were measured from individually (Fig. 1) and simultaneously (Fig. 2, A and B) stained experiments, and the ratios of IFN-R2 to IFN-R1 protein expression were plotted for early and late normal pregnancies and for PE with or without IUGR (Fig. 2, C and D). These results confirmed that the IFN-R2:IFN-R1 ratio in early pregnancy was consistently lower (R1 high:R2 low) compared with that in late pregnancy (R1 high:R2 high) and that the ratio in PE with or without IUGR (R1 high:R2 low) was significantly lower than that observed in late control pregnancy. In normal pregnancy, the difference in the ratio between early and late stages was largely produced by an elevation of IFN-R2 protein expression in late pregnancy. By contrast, in PE with or without IUGR, the drop in the ratio was caused by both a reduction in the level of IFN-R2 protein expression and an elevation in the level of IFN-R1 protein expression.
|
expression are elevated in PE
LIF, a member of the IL-6 cytokine family (which includes IL-6, IL-11, oncostatin M, cardiotropin, and neurotropin), is the specific ligand for LIFR, and its high affinity binding is mediated by heterodimerization with gp130 (36). In humans, LIF is predominantly secreted by the uterine endometrium (37, 38, 39) and is essential for implantation and the establishment of pregnancy (40). Therefore, the high expression of LIFR in villous tissues suggests that it has a critical role in placental development (29). The superficial placental implantation (2, 16) and down-regulation of IFN-R2 in PE (Figs. 1 and 2) prompted our examination of LIFR expression in PE. As expected, LIFR is extensively glycosylated in placenta, producing various species up to Mr 150–200K (Fig. 3A). Unlike its transcriptional expression (Fig. 1A), LIFR protein accumulation was dramatically reduced in late pregnancy. However, in PE, LIFR protein expression was significantly higher (P < 0.05) than that in age-matched controls (Fig. 3, A and B).
|
strongly up-regulates its own expression in natural killer (NK) cells or stimulated T cells (41). However, it down-regulates IFN-R1 expression by preventing receptor internalization and IFN-R2 expression in Th1 cells, leading to cellular desensitization (42, 43). Additionally, unlike that in the mouse, the cell surface expression of IFN-R1 and -2 in humans is independently regulated by intracellular trafficking (44). The experiments presented above (Figs. 1 and 2) established that the relative receptor protein density for IFN- signaling in trophoblasts changes dramatically from early to late human pregnancy. This led us to examine whether altered receptor expression influences IFN- production in villous trophoblasts. In placenta, glycosylation of IFN- produces molecules of Mr up to 22.5K (Fig. 3C). Quantitative analysis of the results shown in Fig. 3D revealed that in early placenta IFN- production is significantly higher than that in late pregnancy and that a high level of expression is maintained in PE (Fig. 3D).
IFN-R2 expression is most intense in extravillous invasive CT
Real-time PCR (Fig. 1A) and Western blot analyses (Figs. 1–3) provided a quantitative estimate of the receptor-specific RNA and protein levels in early and late normal pregnancy and in PE. However, these studies did not examine the cellular distribution of IFN-R1 and IFN-R2 receptors in placental tissues. Although previous reports have established that IFN-R1 expression occurs in villous syncytiotrophoblasts (ST) as well as in extravillous CT (28), the localization of IFN-R2 in human placenta has never been established. To address this, placental tissues from various pregnancy stages and conditions were reacted with anti-IFN-R1 and IFN-R2 antibodies. IFN-R1 antibody (1:50 dilution) staining was dark brown at all dilutions tested, whereas IFN-R2 antibody (1:200 dilution) stained light brown (Fig. 4). As expected, in early placenta, the IFN-R1 antibody stained villous stem CT (sCT), outer ST layer, and extravillous CT almost uniformly. IFN-R1 was also localized in villous stroma, Hofbouer cells, and fetal blood vessels. Unlike IFN-R1, IFN-R2 barely reacted with villous sCT, and the staining of the outer ST layer was uneven; intense staining was interspersed with poorly stained areas, whereas the floating villi, extravillous trophoblasts in the cell column, and invasive CT (iCT) reacted strongly with the antibody (Fig. 4). These results suggest that IFN-R2 expression in early pregnancy is highest in extravillous trophoblasts, but is uneven on the anchoring villous ST layer. Similar experiments with placental sections from late pregnancy revealed that the antibodies against both receptors reacted strongly with the villous tissue. However, the IFN-R2 antibody poorly stained placental sections from PE and PE with IUGR compared with age-matched control placenta (Fig. 5). These results confirm our Western blot experiments (Figs. 1 and 2), which showed that IFN-R2 protein in PE was significantly reduced compared with that in late normal pregnancies.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
R2 in the placenta is severely inhibited. Most importantly, the protein expression of IFN-R1 and -2 in late PE mimics that of normal early (7–10 wk gestation) hypoxic placenta. Paradoxically, placental LIF receptor expression, which is essential for implantation, is enhanced in PE compared with control pregnancies.
The transcriptional regulation of the three genes did not always correspond to the protein expression data. For example, although LIFR protein synthesis in placenta was dramatically reduced in late pregnancy compared with that in early pregnancy and was significantly high in PE, the mRNA levels were comparable in all pregnancy conditions. Similarly, IFN-R2 transcription in late pregnancy and that in PE were not significantly different. This suggests that the primary regulation of these cytokine receptors is posttranscriptional, and that the stability of mRNA, the rate of its translation, the level of protein stability and degradation, and/or the degree of posttranslational modification may be key points of modulation that merit additional study. It is interesting to note that unlike IFN-R1, both IFN-R2 and LIFR are heavily glycosylated in human trophoblasts.
LIFR expression in the human utero-placental unit has been reported in villous and extravillous trophoblasts, fetal VEC (29), and uterus (37, 38, 39). Although LIFR mRNA is uniformly expressed throughout human pregnancy (29), receptor protein expression was dramatically reduced in late pregnancy and was significantly higher in PE than in age-matched control pregnancies. This increased placental synthesis of LIFR in PE was unexpected given that LIFR–/– mouse exhibits disrupted placental structure, severe fetal growth restriction, and neuronal loss (40). However, in light of the recent discovery of dual nonredundant functions for LIF, the overexpression of LIFR in PE could be linked to the LIF-induced activation of proapoptotic signal transducer and activator of transcription 3 and the simultaneous inhibition of the Src homology protein tyrosine phosphatase-2/Ras/extracellular signal regulated kinase pathway (45, 46).
IFN- is one of the key cytokines that promotes innate and adaptive immune responses at the fetal-maternal interface (8, 18, 19). The uterine vascular abnormalities and decidual basalis pathology in genetically modified mice depleted of uterine NK (uNK) cells are very similar to those in IFN--deficient and IFN-R-deficient pregnant mice (47). Moreover, the differentiation and proliferation of uNK cells in mice are regulated by IL-15 (48, 49). Although IFN- production in chorionic villous is primarily restricted to early pregnancy (28), both IFN-R1 and -R2 proteins are expressed in placental trophoblasts throughout human pregnancy. Nevertheless, IFN-R2 protein expression is significantly increased in late compared with early pregnancy. This early to late transition in relative cell surface expression of the two receptors in placental trophoblasts is remarkably similar to that of the activated Th1 and Th2 T cells with respect to IFN-R expression (11, 25, 26, 27). This parallel together with the results reported in this paper suggest a scheme (Fig. 6) that might facilitate understanding the role of IFN- in the developmental and immune regulation of normal human pregnancy as well as in PE.
|
R1 and -R2 expression in sCT as well as in ST are high and low (R1hiR2low), respectively. This facilitates sensitization of ST to the autocrine regulation by IFN- (50), triggering proliferation, which leads to the secretion of a variety of growth factors, cytokines, and hormones. The reproductive hormone human chorionic gonadotropin released from ST regulates the differentiation of epithelial sCT to endothelial-type iCT (51). Unlike ST, the mononucleated iCT have a high membrane density of both IFN-R1 and -R2 (R1hiR2hi), a phenotype that favors an antiproliferative/apoptotic effect, and are subject to paracrine regulation by IFN-. The iCT escape apoptosis by up-regulation of nonclassical MHC-1 expression (human leukocyte antigen-G), but retain the capacity to induce an apoptotic effect on VEC and SMC after maternal arteriole invasion. In early pregnancy, VEC and SMC could also be subject to apoptosis by coordinated paracrine regulation via IFN- release from uterine T and NK (CD56bright) cells and macrophages at the fetal-maternal interface (8, 52).
Once pregnancy is established and progresses into the second and third trimester, IFN- production subsides, thereby down-regulating iCT (R1Hi R2Hi) function and uterine invasion. Therefore, reduced IFN- secretion by late placental trophoblasts is a physiological necessity to circumvent uncontrolled invasion, which is rarely observed in previa percreta (53). This early to late switch of IFN- and its receptor regulation coincide with the developmental transition from hypoxic to normotensive in human placenta. Indeed, the IFN-R2 density in T cells can be modulated by environmental factors such as nitric oxide and low O2 tension (54, 55).
In PE, IFN- production persists until late pregnancy, when placental receptor expression (R1hiR2low) mimics that of early hypoxic development. If IFN- acts as an inflammatory mediator in establishing endothelial invasion in early pregnancy, one would expect that the abundance of this cytokine might counter the superficial uterine implantation most commonly associated with PE. This apparent contradiction could be reconciled, considering that the differentiation of iCT, most likely controlled by human chorionic gonadotropin and other hormones released from ST, is relatively independent of IFN- signaling. Therefore, the lack of an adequate mass of iCT in PE (14) impairs the uterine invasion despite the high abundance of IFN-. IFN-, instead of exhibiting its physiological paracrine effect on iCT, activates uNK cells, which activate macrophages that induce proapoptotic pathways in PE. This scheme is consistent with the hypothesis (19) that the innate immunity of placental trophoblasts regulates the cytokine profile of immune effector cells at the fetal-maternal interface.
In conclusion, this report provides the first experimental evidence demonstrating a novel developmental placental expression of two cytokine receptors that critically influence placental and fetal growth in human pregnancy. These observations raise the possibility of a link between oxidative stress and the regulation of these cytokine receptors in CT, VEC, and SMC. However, the data presented here are insufficient to distinguish whether increased oxidative stress affects the posttranslational modification of cytokine receptors at the fetal-maternal interface or whether the aberrant expression of these receptors causes a failure to stimulate growth and differentiation, resulting in hyperactivation of reactive oxygen species in PE (17). Synthetic superoxide dismutase/catalase mimetics, which target mitochondria (56), may help in testing these alternative explanations.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: CT, Cytotrophoblast; HELLP, hemolysis, elevated liver enzymes, and low platelets syndrome; iCT, invasive cytotrophoblast; IFN-, interferon-; IFN-R, interferon- receptor; IUGR, intrauterine growth restriction; LIF, leukemia inhibitory factor; LIFR, leukemia inhibitory factor receptor; MHC, major histocompatibility complex; Mr, molecular mass; NK, natural killer; PE, preeclampsia; sCT, stem cytotrophoblast; SMC, smooth muscle cell; ST, syncytiotrophoblast; Th, T helper cell; uNK, uterine natural killer; VEC, vascular endothelial cell.
This work was supported by the Fetal Medicine Foundation, United Kingdom.
Received June 11, 2004.
Accepted November 23, 2004.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
14 NKT cells in abortion. Proc Natl Acad Sci USA 97:740–744. Annu Rev Immunol 15:749–795[CrossRef][Medline]
, ß, or using oligonucleotide arrays. Proc Natl Acad Sci USA 95:15623–15628 receptor 2 expression as the deciding factor in human T, B, and myeloid cell proliferation or death. J Leukocyte Biol 70:950–960 signaling pathways in optimal IFN- production by human T cells. Eur J Immunol 32:693–700[CrossRef][Medline]
R2 chain surface expression and desensitizes IFN-/STAT-1 signaling in human T lymphocytes. Blood 102:2933–2939 and interferon--receptor. J Reprod Immunol 27:37–48[CrossRef][Medline]
interferon gene expression by interleukin 12 (IL-12) and IL-2 represents a novel form of posttranscriptional control. Mol Cell Biol 22:1742–1753 production in IL-18 treated NK cell line LNK5E6. Biochem Biophys Res Commun 300:980–985[CrossRef][Medline]
and IL-4 through mRNA stabilization: involvement of the RNA-binding protein HuR. J Immunol 171:4369–4378 inhibits human T lymphocyte and natural killer cell activation. J Immunol 147:1445–1452[Abstract]
receptor ß chain expression in T helper cell subsets. Science 270:1215–1218 of STAT factors following T cell activation. Eur J Immunol 26:704–709[Medline]
R2 chain is regulated by intracellular trafficking in human T lymphocytes. J Immunol 164:201–207 contributes to initiation of uterine vascular modification, decidual integrity, and uterine natural killer cell maturation during normal murine pregnancy. J Exp Med 192:259–270. J Exp Med 193:1199–1212-treated murine macrophages cultured under hypoxic conditions. J Immunol 157:2638–2644[Abstract]
This article has been cited by other articles:
 |
|
 |
|
  S. P. Murphy, C. Tayade, A. A. Ashkar, K. Hatta, J. Zhang, and B. A. Croy Interferon Gamma in Successful Pregnancies Biol Reprod, May 1, 2009; 80(5): 848 - 859. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. R Dietert and M. S Piepenbrink The managed immune system: protecting the womb to delay the tomb Human and Experimental Toxicology, February 1, 2008; 27(2): 129 - 134. [Abstract] [PDF]
|
|
|
|
 |
|
 M. Cudmore, S. Ahmad, B. Al-Ani, T. Fujisawa, H. Coxall, K. Chudasama, L. R. Devey, S. J. Wigmore, A. Abbas, P. W. Hewett, et al. Negative Regulation of Soluble Flt-1 and Soluble Endoglin Release by Heme Oxygenase-1 Circulation, April 3, 2007; 115(13): 1789 - 1797. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Francis, R. Rai, N. J. Sebire, S. El-Gaddal, M. S. Fernandes, P. Jindal, A. Lokugamage, L. Regan, and J. J. Brosens Impaired expression of endometrial differentiation markers and complement regulatory proteins in patients with recurrent pregnancy loss associated with antiphospholipid syndrome Mol. Hum. Reprod., July 1, 2006; 12(7): 435 - 442. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. A. Hills, V. M. Abrahams, B. Gonzalez-Timon, J. Francis, B. Cloke, L. Hinkson, R. Rai, G. Mor, L. Regan, M. Sullivan, et al. Heparin prevents programmed cell death in human trophoblast Mol. Hum. Reprod., April 1, 2006; 12(4): 237 - 243. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
