Defective Production of Interleukin11 by Decidua and Chorionic Villi in Human Anembryonic Pregnancy

Defective Production of Interleukin-11 by Decidua andChorionic Villi in Human Anembryonic Pregnancy

HSIN-FU CHEN, CHING-YIN LIN, KUANG-HAN CHAO, MING-YIH WU, YU-SHIH YANG, AND

HONG-NERNG HO

Division of Reproductive Endocrinology and Infertility (H.-F.C., K.-H.C., M.-Y.W., Y.-S.Y., H.-N.H.), Department ofObstetrics and Gynecology, College of Medicine and the Hospital, National Taiwan University; and Graduate Institute ofImmunology (C.-Y.L., H.-N.H.), College of Medicine, National Taiwan University, Taipei 100, Taiwan

Previous study demonstrated that IL-11 receptor � knockoutfemale mice (IL-11R��/�) were phenotypically normal but in-fertile due to defective decidualization. However, the role ofIL-11 signaling in human reproduction remains unclear. Thisstudy examined the expression of IL-11, IL-11R�, and signaltransduction factor glycoprotein 130 in different phases ofendometrium (six in proliferative phase and four in secretoryphase), and the decidua and villi of normal pregnancy (NP; n �25) and anembryonic pregnancy (AP; n � 25) in the first tri-mester (gestational week 7–9). RT-PCR showed IL-11, IL-11R�,and glycoprotein 130 mRNA expression in all samples, exceptthe absence of IL-11 signal in the unstimulated MRC-5 cell andthe proliferative phase endometrium. Real-time quantitativePCR showed that the relative level of IL-11R� mRNA was notsignificantly different among proliferative phase endome-trium (relative level; mean � SEM, 1.4 � 0.5), secretory phaseendometrium (1.3 � 0.1), or decidua from NP or AP (1.7 � 0.3and 1.9 � 0.4, respectively), but was significantly greater inchorionic villi either from NP or AP (7.6 � 1.3 and 10.6 � 1.9,respectively; both P < 0.05, compared with decidua or endo-metrium). No difference of IL-11R� mRNA level was foundbetween NP and AP (1.7 � 0.3 vs. 1.9 � 0.4 in deciduas; 7.6 � 1.3vs. 10.6 � 1.9 in villi; both P > 0.05). In situ hybridizationlocalized IL-11R� mRNA expression in proliferative phase en-dometrium (stroma only), secretory phase endometrium (stro-ma and gland), decidua (stroma and gland), and villi (tropho-blast and stroma). The staining intensities were not distinctly

different between different groups of samples or between dif-ferent cell types in a sample. No difference in IL-11R� expres-sion was found between NP and AP when either decidua orchorionic villi was analyzed. IL-11 mRNA level was not de-tected in the proliferative phase (relative level, 0.0 � 0.0), wasbarely detectable in the secretory phase (0.03 � 0.02), and wassignificantly increased in decidua (1.7 � 0.2 and 0.1 � 0.1,respectively, for NP and AP) and chorionic villi (13.0 � 2.2 and0.2 � 0.1). In addition, IL-11 mRNA level was higher in NP thanin AP both in decidua (1.7 � 0.2 vs. 0.1 � 0.1; P � 0.03) and invilli (13.0 � 2.2 vs. 0.2 � 0.1; P < 0.001). Immunohistochemistrystudy showed that IL-11 was nearly absent in endometrium inboth phases, but clearly detectable in decidua and villi. Con-sistent with the results of quantitative PCR, the staining in-tensity was stronger in villi and decidua from NP than thosefrom AP. The spatial and temporal changes in IL-11 and itsreceptor observed in this study suggest that IL-11 may beproduced both by the embryo (predominantly) and the decid-ual cells and exerts its action on chorionic villi and decidua inan autocrine or paracrine manner. In the presence of a base-line level of IL-11R�, IL-11 may subsequently regulate placen-tation and decidualization for the maintenance of a NP. Thefinding of decreased IL-11 expression in the absence of anychange in IL-11R� in AP suggests that defective expression ofIL-11 but not IL-11R� may account for certain cases of AP.(J Clin Endocrinol Metab 87: 2320–2328, 2002)

IMPLANTATION AND PLACENTATION are major stepsin the development of mammalian embryos (1). The

mechanisms underlying failures of implantation or placentaldevelopment remain mostly unknown. Recently, a numberof growth factors and cytokines, including colony-stimulat-ing factor, leukemia inhibitory factor, and IL-11, have beensuggested to play a role in these processes (2–7). A temporalanalysis revealed that IL-11 expression was maximal in thenormal pregnant uterus at the time of decidualization in mice(4). Subsequently, mating of female mice with a null muta-tion of the IL-11 receptor � (IL-11R��/�) with fertile males ofany genotype for periods up to 12 months never resulted inclinical pregnancy (4). Histological study showed that thesefemale mice were infertile due to defective decidualization(4). This observation identified a previously unrecognized

critical role of IL-11 in female reproduction of mice duringthe period of decidualization. However, the role of IL-11signaling in humans has not been established. Further stud-ies are therefore needed to evaluate the expression of IL-11and IL-11R� in pathological conditions associated with hu-man implantation failure or recurrent abortion.

IL-11 was first isolated as a soluble factor produced by thePU-34 primate bone marrow stromal cell line (8), and sub-sequently human IL-11 was cloned (9, 10). The actions ofIL-11 are mediated through a receptor complex composed ofIL-11R� and signal transduction factor, glycoprotein 130(gp130) (11). Recently, IL-11 has been shown to have pleio-tropic effects in many tissues, including the hematopoieticsystem, skin, bone, central nervous system, thymus, and lung(12–14). Studies have also suggested that IL-11 plays variousroles in the urogenital system (4, 12, 14, 15). For example,immunoreactive IL-11 was detected in human follicularfluid, although its action in the ovary has not been estab-lished (15).

The present study was designed to evaluate the role of the

Abbreviations: AP, Anembryonic pregnancy; BCIP, 5-bromo-4-chloro-3-indoxyl-phosphate; CT, threshold cycle; DIG, digoxigenin;gp130, glycoprotein 130; IL-11R, IL-11 receptor; NP, normal pregnancy;NBT, nitro-blue tetrazolium chloride; PBMC, peripheral blood mono-nuclear cells; RT, room temperature.

0013-7227/02/$15.00/0 The Journal of Clinical Endocrinology & Metabolism 87(5):2320–2328Printed in U.S.A. Copyright © 2002 by The Endocrine Society

2320

IL-11 system in early human pregnancy, especially in de-cidualization and placental formation. We hypothesized thatgestational tissues from patients with spontaneous abortionmight be characterized by differences in traits of IL-11 ex-pression compared with those in normal pregnancy (NP). Totest this hypothesis, we analyzed and compared IL-11, IL-11R�, and gp130 expression in NP and anembryonic preg-nancy (AP; blighted ovum).

Subjects and MethodsSubjects and materials

After confirmation by ultrasonography, 25 patients with the diag-nosis of AP at the gestational age of 7–10 wk were recruited into thestudy group. Twenty-five age-matched patients with NP of gestationalage 7–10 wk who had requested pregnancy termination due to multi-parity were recruited as control. The maternal age (mean � sem, 32.3 �1.2 vs. 33.6 � 2.1 yr for AP and NP, respectively) and gestational age(mean � sem, 8.8 � 0.5 vs. 8.2 � 0.6 wk) were comparable. Four of the25 patients in the AP group had history of repeated spontaneous abor-tion of three or more times, whereas one in the NP group had this kindof history. Fetal heart motion was documented by ultrasound. Samplesof decidua and chorionic villi were obtained during dilatation andcurettage. Endometrial tissues were obtained from 10 patients withnormal menstruation (6 in the proliferative phase and 4 in the secretoryphase) who had undergone hysterectomy due to uterine fibroids. Pe-ripheral blood mononuclear cells (PBMC) were obtained from ninenormal, nonpregnant women by venous puncture. Cell lines for use aspositive controls, including MRC-5, a human embryonal lung fibroblastcell line (served as a positive control for IL-11 expression); K562, a humanchorionic myelogenous leukemia lymphoblast cell line (served as apositive control for IL-11R� expression); and HeLa, a human cervicalepitheloid carcinoma cell line, were obtained from the Food IndustryResearch and Development Institute of Taiwan. Informed consent wasobtained from each patient who participated in this study. This studywas approved by the Committee for Human Body Experiments of Na-tional Taiwan University Hospital in 1998.

Tissue and cell processing

The endometrial and gestational tissues were collected aseptically,put into test tubes containing normal saline, and immediately taken tothe laboratory for further processing. The preparation of decidual andvillus tissues was performed as previously described (16). Briefly, tissuemixtures were macroscopically separated into decidua and chorionicvilli, washed twice with HBSS, and divided into portions for further use.After collection, the endometrial tissues were prepared using methodssimilar to those used for gestational tissues. Small pieces of endome-trium, decidua, and villi were also stained with hematoxylin and eosinand examined by a pathologist to exclude any possible pathology. Ve-nous blood was collected into a syringe prerinsed with heparin and sentto the laboratory for immediate preparation of PBMC. The blood sam-ples were diluted 1:1 in HBSS and centrifuged on a 100% Ficol-Hypaquegradient at 1800 rpm for 20 min without the use of brake at completion.Cells at the interface were aspirated, diluted 1:9 in HBSS, and centrifugedtwice at 1400 rpm for 5 min. Finally, the cell pellet was collected forfurther study. Cell lines were thawed, washed, and cultured in T75 tissueculture dishes (Corning, Inc., Corning, NY) using respective medium

(�-MEM plus 10% FCS for MRC-5; RPMI-1640 plus 10% FCS for K562and HeLa cells). When cells covered approximately 90% of the dish floor,the culture medium was removed, and the cells (except K562 cells, whichare floating cells) were washed with HBSS and trypsinized. The cellsuspension was collected, washed, and centrifuged, and the cell pelletwas harvested for subsequent subculture or further experiments.

RT-PCR

Portions of the tissues (endometrium, decidua, and chorionic villi)subjected to RT-PCR were first cut into small pieces, washed with PBS(diethyl pyrocarbonate-treated), and homogenized in liquid nitrogen.Subsequently, RNA was extracted using the Trizol reagent (Life Tech-nologies, Inc., Grand Island, NY), chloroform-isopropanol-ethanolmethod as previously described (17). The concentration of RNA wasanalyzed by UV spectrophotometer to measure the absorbance at wave-lengths of 260 and 280 nm (OD260 and OD280). Purity of RNA wasconsidered adequate when an OD260/OD280 ratio of greater than 1.6 wasreached. The integrity of the RNA sample was further verified both byagarose gel electrophoresis and examination of the RT-PCR product of�-actin, a housekeeping gene. RNA samples (5 �g) were subsequentlysubjected to first strand cDNA synthesis using oligo(dT) primers andSuperscript II reverse transcriptase following the manufacturer’s pro-tocol (First Strand cDNA Synthesis kit, Pharmacia Biotech, Uppsala,Sweden). In a 50-�l reaction mixture, PCR was performed in the pres-ence of 25 �l Taq Master mix (QIAGEN, Valencia, CA), 3 �l first strandcDNA template, 200 ng/�l 5�-primer, 200 ng/�l 3�-primer, and 20 �ldouble distilled H2O. The primer sets were obtained or designed ac-cording to either the published oligonucleotide primer sets or the pub-lished cDNA sequences for IL-11, IL-11R�, gp130, and �-actin (18–21)(Table 1). All of the primers were designed to span at least two exonsto exclude possible genomic DNA contamination. The �-actin mRNAlevel was also measured to control for the mRNA level in the sample andfor possible semiquantitation. A number of preliminary experimentswere performed to determine the PCR conditions and the cycle numbersthat could ensure the PCR production in an exponential phase. The PCRwas carried out in a thermal cycler (GeneAmp PCR System 2400; Perkin-Elmer Corp., Wellesley, MA) in the following sequence: denaturation at94 C for 30–60 sec, annealing at 62–65 C for 30–60 sec, extension at 72C for 1–2 min, and a final extension at 72 C for 10 min. The PCR wasrepeated for 30–38 cycles according to the gene of interest, and the PCRproduct was examined using 1.5% agarose gel electrophoresis andstained with ethidium bromide. Several positive and negative controlswere used in the PCR. MRC-5 and K562 cells were used as positivecontrols for IL-11 and IL-11R�, respectively. A reverse transcriptionreaction was carried out without the addition of reverse transcriptase,and the resulting product was subjected to PCR to exclude the possibilityof genomic DNA contamination. PCR was also performed without thepresence of template DNA to test for cross-contamination of samples.

Real-time quantitative PCR

To accurately quantify the mRNA levels in the samples, a real-timequantitative PCR system (ABI PRISM 7700 Sequence Detection System,PE Applied Biosystems, Foster City, CA) was used as previously de-scribed (22) and with the manufacturer’s manual. Briefly, RNA wasprepared, and cDNA was obtained by random hexamer priming. Theprimers and probes used in the PCR were designed according to theTaqMan primer and probe design system (PE Applied Biosystems), andtheir sequences are listed in Table 2. FAM (6-carboxyfluorescein) and

TABLE 1. Sequences of primer sets for RT-PCR

Gene Sequence Product size

IL-11 Sense 5�-CTGTGTTGTCCACGGGGCTGAGTTC-3� 322 bpAntisense 5�-ATGCTGGTATCTCCTTTGGTATGGT-3�

IL-11R� Sense 5�-ACTGCTGCTGCTGAAGACTCGGCTGTGA-3� 443 bpAntisense 5�-ATGGGGAAGAGCCAGGGCAGAAGTCTGT-3�

gp130 Sense 5�-TAAAGGCATACCTTAAACAAGC-3� 292 bpAntisense 5�-GTGAATTCTGGACCATCCTTCC-3�

�-actin Sense 5�-CCTCGCCTTTGCCGATCC-3� 626 bpAntisense 5�-GGATCTTCATGAGGTAGTAGTCAGTC-3�

Chen et al. • IL-11 in Decidua and Villi J Clin Endocrinol Metab, May 2002, 87(5):2320–2328 2321

VIC were the reporter dyes, and TAMRA (6-carboxy-tetramethyl-rhodamine) was the quencher dye. The probes were labeled with re-porter dye and quencher dye on the 5� and 3� ends, respectively. DuringPCR, the reporter dye was released, and the resulting fluorescence wasdetected and quantified. A number of preliminary experiments wereperformed to determine the PCR conditions. The PCR was carried outin a thermal cycler (ABI PRISM 7700 Sequence Detection System) in thefollowing sequence: reaction at 50 C for 2 min and 95 C for 10 min;subsequently, the PCR was repeated for 40 cycles at denaturation at 94C for 15 sec and annealing and extension at 60 C for 1 min. The au-thenticity of PCR products was verified by 2% agarose gel electrophore-sis and by sequencing. The relative concentration of each mRNA wassubsequently calculated according to the manufacturer’s user manual.Briefly, the threshold cycle (CT) values of the target gene mRNA (IL-11or IL-11R�) and the internal control (�-actin) in the studied sample werefirst measured. The �CT value of the studied sample was calculated bythe following formula: �CT � CT of target gene � CT of �-actin (this �CTis designated as sample �CT). Similarly, the CT values of the target geneand its respective �-actin of positive control (proliferative phase endo-metrium) were also obtained, and the �CT was calculated (designatedas calibrator �CT). ��CT was then calculated using the following for-mula: ��CT � �CT (sample) � �CT (calibrator); and finally the relativevalue of each mRNA was calculated by the formula: 2��CT. PCR with-out template was used as a negative control (called no template control)to verify experimental results.

In situ hybridization

Protocols for the synthesis of riboprobes and in situ hybridizationwere modified according to a previous report (23). Briefly, antisense andsense digoxigenin (DIG)-labeled (by DIG RNA labeling kit, BoehringerMannheim Biochemica, Mannheim, Germany) riboprobes against IL-11R� were synthesized by in vitro transcription using a plasmid (anexpression vector, pED, Genetics Institute; kindly supplied by Y.-C.Yang, Indiana University, Indianapolis, IN) containing IL-11R� cDNAas the template. The specificity of this probe was verified by Northernblot analysis, and an expected 1.7-kb band indicating the hybridizationof the probe with IL-11R mRNA was found (19). Further examination ofthe endometrium, decidua, and chorionic villi with Northern blot anal-ysis also confirmed the efficiency of this probe (data not shown).

In situ hybridization was carried out as follows. After cleansing, tissuesamples were fixed overnight at 4 C with 4% paraformaldehyde, incu-bated overnight at 4 C with 30% sucrose, and then embedded in OCT(Sakura Finetek, Torrance, CA). During or before experiments, tissuesembedded in OCT were sectioned at 6-�m thickness and dried. Thesecryosections were treated with RNase-free proteinase K (1 �g/ml) at 37C for 30 min, postfixed with 4% paraformaldehyde/0.2% glutaralde-hyde at 4 C for 5 min, subjected to prehybridization at 37 C for at least10 min, and then hybridized overnight at 55 C in hybridization buffer(40% deionized formamide, 10% dextran sulfate, 1� Denhardt’s solu-tion, 4� SSC, 10 mm dithiothreitol, 1 mg/ml yeast tRNA, 1 mg/mldenatured and sheared salmon sperm DNA, and 5–10 ng/30 �l DIG-labeled riboprobes). On the next day, the tissue sections were washedsequentially with 2� SSC and 1� SSC, incubated at 37 C for 30 min inNTE buffer [500 mm NaCl, 10 mm Tris, and 1 mm EDTA (pH 8.0),containing 20 �g/ml RNase A], and then incubated with 0.1� SSC twicefor 30 min each time. Then the sections were blocked in blocking solution(containing 2% normal sheep serum and 0.1% Triton X-100) at roomtemperature (RT) for 1 h, followed by incubation with sheep anti-DIGantibody (1:300 dilution)-alkaline phosphatase (Fab fragment, Boehr-inger Mannheim) (in Tri buffer with 2% normal sheep serum and 0.1%

Triton X-100) overnight at 4 C in a humidified chamber. Finally, the colorreaction was developed with 5-bromo-4-chloro-3-indoxyl-phosphate(BCIP)/nitro-blue tetrazolium chloride (NBT) (DAKO Corp., Carpinte-ria, CA) for 2–24 h in the dark in a humidified chamber, and counter-stained with methyl green (Vector Laboratories, Burlingame, CA). Twonegative controls were also used, one using tissue sections hybridizedwith sense probes and the other using pretreatment with RNase beforehybridization with antisense probes. The staining intensities were readand graded anonymously by the same pathologist.

Immunohistochemistry

The immunohistochemistry study was performed according to pre-vious reports (24, 25) and using the Streptavidin-Biotin Universal De-tection System according to the protocol provided by the manufacturer(Immunotech, Marseille, France). Briefly, formalin-fixed, paraffin-embedded tissue sections were deparaffinized with xylene and hydratedwith step-down concentrations of ethanol. The sections were then in-cubated with 30% H2O2/70% methanol solution at RT for 5 min, treatedwith protein blocking agent (goat serum) at RT for 5 min, and incubatedovernight at 4 C with anti-IL-11 antibody (1:20 dilution; R&D SystemsInc., Minneapolis, MN). The next day, the sections were washed withPBS, treated with biotinylated secondary antibody at RT for 10 min, andtreated with streptavidin-peroxidase reagent at RT for 10 min. The colorwas developed using 3,3�diaminobenzidine chromogen at RT for 15 minand counterstained with hematoxylin for 1 min.

Statistical analysis

Because the number of cases in each study group is relatively lower(25 cases in each group), a nonparametric testing would be more ap-propriate for the statistical analysis. Therefore, the difference betweengroups was analyzed by the Mann-Whitney U test. A P value of less than0.05 was considered statistically significant.

ResultsThe expression of IL-11, IL-11Rd, and gp130 genes

PCR revealed that all of the samples except unstimulatedMRC-5 cells and human endometrium in the proliferativephase (data not shown) expressed IL-11 mRNA. These in-cluded K562 cells, MRC-5 cells stimulated by IL-1� (1 ng/ml), and phorbol 12-myristate 13-acetate (10 ng/ml) (servedas positive control), HeLa cells, human PBMC, endometriumin the secretory phase, and decidua and chorionic villi sam-ples from all patients (Fig. 1). In addition, IL-11R� and gp130mRNA were detectable in all samples (Fig. 1).

Quantitation of IL-11 and IL-11R� mRNA byquantitative PCR

Real-time quantitative PCR was performed to compare thelevels of expression of IL-11 system genes in different sam-ples. Using the proliferative phase endometrium as a refer-ence tissue, it was found that the relative levels of IL-11R�mRNA were 1.4 � 0.5, 1.3 � 0.1, 1.7 � 0.3, 1.9 � 0.4, 7.6 �1.3, and 10.6 � 1.9 (mean � sem), respectively, in prolifer-

TABLE 2. The sequences of TaqMan primers and probes used in this study

Gene Sequence Position

IL-11 Forward primer 5�-TCTCTCCTGGCGGACACG-3� 278–295Reverse primer 5�-AATCCAGGTTGTGGTCCCC-3� 338–356Probe 5�-VICa-AATTTGTCCCTCAGCTGTGCAGCCAG-TAMRAa-3� 302–327

IL-11R� Forward primer 5�-CACACCCTCGGCTACTTGAT-3� 1116–1135Reverse primer 5�-AAGAAAGGATTCCCAAAGACG-3� 1169–1189Probe 5�-FAMa-ACAGCTACCTGCTCCACAGAGTCCCTGT-TAMRAa-3� 1137–1164

a FAM and VIC are the reporter dyes and TAMRA is the quencher dye as detailed in Subjects and Methods.

2322 J Clin Endocrinol Metab, May 2002, 87(5):2320–2328 Chen et al. • IL-11 in Decidua and Villi

ative phase endometrium, secretory phase endometrium, de-cidua (from NP), decidua (from AP), chorionic villi (fromNP), and chorionic villi (from AP) (Fig. 2). These findingsindicate that IL-11R� mRNA levels did not change in theendometrium during each phase of the menstrual cycle andin the decidua after pregnancy was established. However,

chorionic villi contained a distinctly higher level of IL-11R�mRNA than decidua or other phases of endometrium (P �0.05 comparing villi from NP or AP with all other tissues)(Fig. 2). There was, however, no difference in IL-11R� mRNAlevel between NP and AP in either decidua or chorionic villi(in decidua, 1.7 � 0.3 vs. 1.9 � 0.4, respectively, for NP andAP, P � 0.05; in villi, 7.6 � 1.3 vs. 10.6 � 1.9, P � 0.05).

Because IL-11 mRNA was not detectable in proliferativephase endometrium, the level in activated MRC-5 cells wasused as the reference. The relative levels of IL-11 mRNA were0.0 � 0.0, 0.03 � 0.02, 1.7 � 0.2, 0.1 � 0.1, 13.0 � 2.2, and 0.2 �0.1 (mean � sem), respectively, for proliferative phase en-dometrium, secretory phase endometrium, decidua (fromNP), decidua (from AP), chorionic villi (from NP), and cho-rionic villi (from AP) (Fig. 3). These data indicate that theIL-11 mRNA level was low in endometrium during the pro-liferative phase and increased during the secretory phase(relative level, 0.0 � 0.0 vs. 0.03 � 0.02, respectively, forproliferative and secretory phases; P � 0.004) (Fig. 3) andafter NP (1.7 � 0.2 and 13.0 � 2.2 in decidua and chorionicvilli, respectively; P � 0.05 compared with endometriumduring the proliferative phase or the secretory phase). In AP,the IL-11 mRNA level also increased (0.1 � 0.1 and 0.2 � 0.1in decidua and chorionic villi, respectively) compared withthat in endometrium of both phases, although it was not ashigh as in NP. The IL-11 mRNA levels were significantlyhigher in NP than in AP, both in decidua and chorionic villi(in decidua, 1.7 � 0.2 vs. 0.1 � 0.1, respectively, for NP andAP, P � 0.0001; in chorionic villi, 13.0 � 2.2 vs. 0.2 � 0.1, P �0.03) (Fig. 3).

FIG. 1. Detection of IL-11 (322 bp), IL-11R� (443 bp), gp130 (292 bp),and �-actin (626 bp) mRNA expression in various samples by con-ventional RT-PCR using primer sets shown in Table 1. The PCRproduct was resolved in 1.5% agarose gel electrophoresis and stainedwith ethidium bromide. �-actin was used as an internal control. M,DNA marker; 1, K562 cells; 2, unstimulated MRC-5 cells; 3, MRC-5cells stimulated with 10 ng/ml phorbol 12-myristate 13-acetate and 1ng/ml IL-1� for 24 h; 4, HeLa cells; 5, human peripheral mononuclearcells; 6, human secretory phase endometrium; 7, decidua from NP; 8,chorionic villi from NP; 9, decidua from AP; and 10, chorionic villi fromAP. IL-11, IL-11R�, gp130, and �-actin mRNA were detected in allsamples, except for the absence of IL-11 mRNA in unstimulatedMRC-5 cells and proliferative phase endometrium (not shown in thisfigure).

FIG. 2. Comparison of the level of IL-11R� mRNA in various samplesby real-time quantitative PCR. Longitudinal axis indicates the rel-ative values of IL-11R� mRNA, which were calculated by the formula:2��CT, as described in Subjects and Methods. PE, Proliferative phaseendometrium; SE, secretory phase endometrium; ND, decidua fromNP; AD, decidua from AP; NV, chorionic villi from NP; and AV,chorionic villi from AP. *, P � 0.05 compared with PE, SE, ND, or AD.The bars represent the mean � SEM.

FIG. 3. Comparison of the level of IL-11 mRNA in various samples byreal-time quantitative PCR. Longitudinal axis indicates the relativevalues of IL-11 mRNA, which were calculated by the formula: 2��CT,as described in Subjects and Methods. The level of stimulated MRC-5cell mRNA was used as the reference. PE, Proliferative phase endo-metrium; SE, secretory phase endometrium; ND, decidua from NP;AD, decidua from AP; NV, chorionic villi from NP; and AV, chorionicvilli from AP. *, P � 0.004, comparison between PE and SE. **, P �0.001, comparison between ND and AD. ***, P � 0.03, comparisonbetween NV and AV.


Immunohistochemical study for IL-11

Immunohistochemical study for IL-11 production foundthat the immunoreactive IL-11 was barely detectable in theglandular epithelium of secretory phase endometrium andwas undetectable in proliferative phase endometrium (rep-resentative photographs are shown in Fig. 4, C and D, andthe relative staining intensities in different tissues are listedin Table 3). After pregnancy, however, IL-11 immunoreac-tivity became detectable both in the stromal and glandularcells of decidua and syncytiotrophoblast and cytotrophoblastcells of the chorionic villi (Fig. 4, E–H, and Table 3). Strongerimmunoreactivity intensity was found in the glandular ep-ithelium than in the stromal cells of decidua (Table 3). Incontrast, the intensity was stronger in trophoblast than instroma of chorionic villi (Table 3). Comparison of tissuesamples from different groups of patients showed that cho-rionic villi from NP produced a higher level of IL-11 than thatfrom AP (Table 3). Similarly, decidua from NP expressedstronger IL-11 immunoreactivity than those from AP.

In situ hybridization for IL-11R� mRNA

Because IL-11R� monoclonal antibody was not available atthe time of study, in situ hybridization technique was usedto examine the expression of IL-11R� gene in the endome-trium and gestational tissues. Proliferative phase endome-trium expressed IL-11R� mRNA in its stromal cells but notglandular cells (Fig. 5 and Table 3). In the secretory phase,however, IL-11R� expression was evident in both stromaland glandular cells (Fig. 5). During pregnancy, both stromaland glandular cells of the decidua and trophoblasts of thechorionic villi continuously expressed this transcript (Fig. 6).However, no distinct difference of staining intensity wasfound among samples from secretory phase endometrium,decidua, or chorionic villi (Figs. 5 and 6 and Table 3). Therewas no significant difference of IL-11R� mRNA levels be-tween NP and AP in either decidua or villi (Table 3).

Discussion

Implantation and placentation are complex processes thatrequire the synchronized growth of an embryo and a con-tinuous and adequate decidual development in the endo-metrium (1). It is conceivable that these processes are sus-ceptible to endogenous or exogenous insults. Among them,defective IL-11R� action has been shown to be the mecha-nism for abnormal decidualization in mice (4). That studyreasonably suggests a potential role of the IL-11 system genesin abnormal decidualization in humans. The present studydemonstrated the expression of IL-11R� mRNA in the en-dometrium throughout the menstrual cycle and in the de-cidua of early pregnancy. Although the level of IL-11R�mRNA in chorionic villi was higher than in decidua, nosignificant difference in IL-11R� mRNA levels was foundbetween decidua and chorionic villi from NP and AP. Incontrast, the IL-11 mRNA level in endometrium progres-sively increased from the secretory phase and reached peaklevels in the decidua and chorionic villi after pregnancy.IL-11 mRNA levels in both decidua and villi of AP weresignificantly lower than those of NP. Gp130 mRNA was also

detectable in all of the samples examined. These data supportthe hypothesis that IL-11 signaling in human endometriumand gestational tissues may be critical in the maintenance ofa NP.

A previous study by Dimitriadis et al. (26) showed thatIL-11 expression varied with different phases of the men-strual cycle. The results of our study support their findingsand also showed that both IL-11 and IL-11R� mRNA werepresent in the endometrium throughout the menstrual cycleand during early pregnancy. In addition, our study clearlydemonstrated that IL-11R� levels remained stable duringthese stages with minimal variation between levels in theendometrium and the decidua. These data are basically con-sistent with those reported in mice, in which constant levelsof IL-11R and gp130 were detected since before pregnancyuntil 9.5 d post coitum (4). However, the real-time quanti-tative PCR results of our study also showed that IL-11R� andIL-11 expression were augmented in the chorionic villi com-pared with the decidua. This suggests that IL-11 may beproduced both in the embryo and in decidual cells, althoughthe former may be the predominant source in humans dueto its greater level. Subsequently, through binding to IL-11R�and gp130, IL-11 may exert its action on the embryo and thedecidua to regulate chorionic villus development and de-cidualization, respectively. This potential dual action of IL-11, however, was not clearly demonstrated in a previousstudy in mice by Robb et al. (4). In that IL-11R� knockoutmodel, the number of secondary trophoblast giant cellspresent near the defective decidua was markedly increased,and the size was enlarged. However, they found that IL-11R��/� embryos could still survive normally in wild-typefoster female but that wild-type embryos were unable tosurvive in the IL-11R��/� female, which does not unani-mously support the role of IL-11 action in trophoblast de-velopment. Therefore, further study is needed to explainwhether the role of IL-11 differs between species. Irrespectiveof all these unanswered issues, the results of the presentstudy suggest that IL-11 may play a crucial role in the de-velopment of human chorionic villi and decidua.

There are inconsistent results obtained by different exper-imental methods that need to be noted. As detailed previ-ously in Results, in situ hybridization did not show a distinctdifference of staining intensity between samples from secre-tory phase endometrium, decidua, or chorionic villi (Table 3).These data are not consistent with the result of real-timequantitative PCR, which showed that chorionic villi ex-pressed a significantly higher level of IL-11R� mRNA thandecidua (Fig. 2). The reason for the divergence of these resultsis not clear, but may involve varied sensitivity of each of theexperimental methods.

The presence of IL-11R� mRNA with minimal variation inamount despite changing IL-11 levels suggests divergentactions between IL-11 and IL-11R�. One possible explanationfor this finding is that the level of IL-11 rather than IL-11R�may regulate normal placental formation and decidualiza-tion. Although previous data indicate that IL-11R� knockoutleads to infertility in mice (4), complete IL-11R� depletion inhumans has not been reported. Therefore, a baseline IL-11R�level most likely is more than enough to fulfill its role in IL-11function. This argument is further supported by the lack of


FIG. 4. Immunohistochemical study of IL-11 in various tissues. Formalin-fixed, paraffin-embedded tissue samples were treated with anti-IL-11antibody (1:20 dilution; R&D Systems Inc.) using the Streptavidin-Biotin Universal Detection System (Immunotech) as described in Subjectsand Methods. The color was developed using 3,3� diaminobenzidine chromogen and counterstained by hematoxylin as described in Subjects andMethods. A, Inflammed gum tissue; B, stimulated MRC-5 cell; C, proliferative phase endometrium; D, secretory phase endometrium; E, deciduafrom NP; F, chorionic villi from NP; G, decidua from AP; and H, chorionic villi from AP. Brown color indicates positive staining (arrowhead).Magnification, �200. GL, Gland; ST, stroma; TR, trophoblast.


difference of IL-11R� expression between NP and AP. How-ever, the potential for defects in IL-11R� action due to eithermutation or deletion in the IL-11R� gene cannot be excludedat present. It would therefore be interesting to investigateIL-11 and IL-11R� at the genomic level in patients withrecurrent spontaneous abortion to identify their possible de-fects in action.

A significant decrease of IL-11 mRNA was found in APcompared with NP. In the presence of IL-11R�, the level ofwhich might only be rarely under the threshold level inhumans, the resulting decreased level of IL-11 might lead todefective placental formation and subsequently result in ablighted ovum (AP). However, we cannot completely ruleout the possibility that decreased IL-11 level could be sec-ondary to fetal demise, rather than itself being the primaryphenomenon. Consequently, more work is needed to clarifythis cause-and-effect relationship.

IL-11, IL-6, ciliary neurotrophic factor, leukemia inhibitoryfactor, and oncostatin M belong to a specific group of cyto-kines that share the common signal-transduction subunitgp130 (12). Overlapping of the biological function of thesecytokines may at least partly be explained by the sharing ofthis common signal-transduction component in formation oftheir high-affinity receptor complexes (20). As expected, thisstudy identified the expression of gp130 mRNA in phases ofendometrium and in gestational tissues. These data providean indirect evidence of the role of gp130 in the completeaction of IL-11R�. However, further detailed spatial andtemporal analyses of this gene will be needed to clarify itsrole in human disease states.

In conclusion, the present study suggests that IL-11 mayplay a crucial role in the maintenance of early pregnancy. Itis likely that IL-11 has dual actions, both on placentation anddecidualization, although in humans the major focus might

TABLE 3. Results of immunohistochemistry (IL-11 protein) and in situ hybridization (IL-11R� mRNA)

Gene

Proliferativephase

endometrium (n)

Secretory phaseendometrium (n) NP (n) AP (n)

Gland(6)

Stroma(6)

Gland(4)

Stroma(4)

Villi (5) Decidua (5) Villi (5) Decidua (5)

Trophoblast Stroma Gland Stroma Trophoblast Stroma Gland Stroma

IL-11 � � � � � � IL-11R� �

Relative staining intensities are represented by �, , , and .

FIG. 5. Expression of IL-11R� mRNA in human endometrium was examined by in situ hybridization, using DIG-labeled anti-IL-11R�riboprobes and color development by NBT/BCIP as described in Subjects and Methods. A, Proliferative phase endometrium (antisense riboprobe);B, proliferative phase endometrium (sense riboprobe); C, secretory phase endometrium (antisense riboprobe); D, secretory phase endometrium(sense riboprobe). Deep purple color indicates positive staining (arrowhead). Magnification, �200. GL, Gland; ST, stroma.


FIG. 6. Expression of IL-11R� mRNA in human decidua and chorionic villi was examined by in situ hybridization, using DIG-labeledanti-IL-11R� riboprobes and color development by NBT/BCIP as described in Subjects and Methods. A and B, Decidua in NP; C and D, chorionicvilli in NP; E and F, decidua in AP; G and H, chorionic villi in AP. Both antisense (A, C, E, and G) and sense (B, D, F, and H) riboprobes wereused. Deep purple color indicates positive staining (arrowhead). Magnification, �200. GL, Gland; ST, stroma; TR, trophoblast.


be on the former. Decreased expression of IL-11 during orafter implantation probably can account for some cases withimplantation failure or recurrent abortion in early preg-nancy. This conclusion however needs further confirmationbecause the cause vs. effect relation cannot be determined atpresent. In addition to the evident academic interest, furtherstudies of IL-11 action will have important clinical implica-tions. Medical use of IL-11 has been shown to effectivelyenhance a number of biological functions including bonemarrow recovery and platelet formation (13). Extrapolationfrom these experiences provides strong support for its po-tential use in treating diseases of the genital organs, whereIL-11 action has previously been identified. In addition, themedical use of IL-11 is relatively nontoxic compared withother cytokines. Therefore, after further confirmation, it islikely that IL-11 may eventually be shown to have clinicalapplications in the treatment of repeated implantation failureand/or recurrent spontaneous abortion.

Acknowledgments

We thank Dr. Yu-Chung Yang (Walther Oncology Center, IndianaUniversity School of Medicine, Indianapolis, IN) for her generosity inproviding the cDNA clones for this study.

Received November 19, 2001. Accepted February 5, 2002.Address all correspondence and requests for reprints to: Hong-Nerng

Ho, M.D., National Taiwan University Hospital, Department of Obstet-rics and Gynecology, 7 Chung-Shan Street Road, Taipei 100, Taiwan.E-mail: [email protected].

This work was supported by grants from the National Science Councilof the Republic of China (NSC89-2314-B-002-582 and NSC89-2314-B-002-289) and the National Health Research Institute (DOH89-DC-1005).

References

1. Cross JC, Werb Z, Fisher SJ 1994 Implantation and the placenta: key piecesof the development puzzle. Science 266:1508–1518

2. Bhatt H, Brunet LJ, Stewart CL 1991 Uterine expression of leukemia inhibitoryfactor coincides with the onset of blastocyst implantation. Proc Natl Acad SciUSA 88:11408–11412

3. Chen JR, Cheng JG, Shatzer T, Sewell L, Hernandez L, Stewart CL 2000Leukemia inhibitory factor can substitute for nidatory estrogen and is essentialto inducing a receptive uterus for implantation but is not essential for sub-sequent embryogenesis. Endocrinology 141:4365–4372

4. Robb L, Li R, Hartley L, Nandurkar HH, Koentgen F, Begley CG 1998Infertility in female mice lacking the receptor for interleukin 11 is due to adefective uterine response to implantation. Nat Med 4:303–308

5. Pollard JW, Hunt JS, Wiktor-Jedrzejczak W, Stanley ER 1991 A pregnancydefect in the osteopetrotic (op/op) mouse demonstrates the requirement forCSF-1 in female fertility. Dev Biol 148:273–283

6. Chen HF, Shew JY, Ho HN, Hsu WL, Yang YS 1999 Expression of leukemia

inhibitory factor and its receptor in preimplantation embryos. Fertil Steril72:713–719

7. Stewart CL, Kaspar P, Brunet LJ, Bhatt H, Gadi I, Kontgen F, AbbondanzoSJ 1992 Blastocyst implantation depends on maternal expression of leukaemiainhibitory factor. Nature 359:76–79

8. Paul SR, Bennett F, Calvetti JA, Kelleher K, Wood CR, O’Hara Jr RM, LearyAC, Sibley B, Clark SC, Williams DA 1990 Molecular cloning of a cDNAencoding interleukin 11, a stromal cell-derived lymphopoietic and hemato-poietic cytokine. Proc Natl Acad Sci USA 87:7512–7516

9. McKinley D, Wu Q, Yang-Feng T, Yang YC 1992 Genomic sequence andchromosomal location of human interleukin-11 gene (IL11). Genomics 13:814–819

10. Paul SR, Schendel P 1992 The cloning and biological characterization ofrecombinant human interleukin 11. Int J Cell Cloning 10:135–143

11. Hilton DJ, Hilton AA, Raicevic A, Rakar S, Harrison-Smith M, Gough NM,Begley CG, Metcalf D, Nicola NA, Willson TA 1994 Cloning of a murine IL-11receptor �-chain; requirement for gp130 for high affinity binding and signaltransduction. EMBO J 13:4765–4775

12. Davidson AJ, Freeman SA, Crosier KE, Wood CR, Crosier, PS 1997 Expres-sion of murine interleukin 11 and its receptor �-chain in adult and embryonictissues. Stem Cells 15:119–124

13. Du X, Williams DA 1997 Interleukin-11: review of molecular, cell biology, andclinical use. Blood 89:3897–3908

14. Du XX, Williams DA 1994 Interleukin-11: a multifunctional growth factorderived from the hematopoietic microenvironment. Blood 83:2023–2030

15. Branisteanu I, Pijnenborg R, Spiessens C, Van der Auwera I, Keith Jr JC, VanAssche FA 1997 Detection of immunoreactive interleukin-11 in human follic-ular fluid: correlations with ovarian steroid, insulin-like growth factor I levels,and follicular maturity. Fertil Steril 67:1054–1058

16. Ho HN, Chao KH, Chen CK, Yang YS, Huang SC 1996 Activation status ofT and NK cells in the endometrium throughout menstrual cycle and normaland abnormal early pregnancy. Hum Immunol 49:130–136

17. Chadderton T, Wilson C, Bewick M, Gluck S 1997 Evaluation of three rapidRNA extraction reagents: relevance for use in RT-PCR’s and measurement oflow level gene expression in clinical samples. Cell Mol Biol 43:1227–1234

18. Nandurkar HH, Hilton DJ, Nathan P, Willson T, Nicola N, Begley CG 1996The human IL-11 receptor requires gp130 for signalling: demonstration bymolecular cloning of the receptor. Oncogene 12:585–593

19. Yang YC, Yin T 1992 Interleukin-11 and its receptor. Biofactors 4:15–2120. Yin T, Taga T, Tsang ML, Yasukawa K, Kishimoto T, Yang YC 1993 In-

volvement of IL-6 signal transducer gp130 in IL-11-mediated signal transduc-tion. J Immunol 151:2555–2561

21. Leavitt J, Gunning P, Porreca P, Ng SY, Lin CS, Kedes L 1984 Molecularcloning and characterization of mutant and wild-type human �-actin genes.Mol Cell Biol 4:1961–1969

22. Gibson UE, Heid CA, Williams PM 1996 A novel method for real timequantitative RT-PCR. Genome Res 6:995–1001

23. Wilkinson DG, Nieto MA 1993 Detection of messenger RNA by in situhybridization to tissue sections and whole mounts. Methods Enzymol 225:361–373

24. Hsu SM, Raine L, Fanger H 1981 A comparative study of the peroxidase-antiperoxidase method and an avidin-biotin complex method for studyingpolypeptide hormones with radioimmunoassay antibodies. Am J Clin Pathol75:734–738

25. Lessey BA, Castelbaum AJ, Buck CA, Lei Y, Yowell CW, Sun J 1994 Furthercharacterization of endometrial integrins during the menstrual cycle and inpregnancy. Fertil Steril 62:497–506

26. Dimitriadis E, Salamonsen LA, Robb L 2000 Expression of interleukin-11during the human menstrual cycle: coincidence with stromal cell decidual-ization and relationship to leukaemia inhibitory factor and prolactin. Mol HumReprod 6:907–914


Documents

Defective Production of Interleukin11 by Decidua and Chorionic Villi in Human Anembryonic Pregnancy