Paracrine effects of uterine leucocytes on gene expression of human uterine stromal fibroblasts

Molecular Human Reproduction, Vol.15, No.1 pp. 39–48, 2009 Advanced Access publication on December 16, 2008 doi:10.1093/molehr/gan075

ORIGINAL RESEARCH

Paracrine effects of uterine leucocytes on gene expression of human uterine stromal fibroblasts Ariane Germeyer 1,6†, Andrew Mark Sharkey 2†, Mirari Prasadajudio 1, Robert Sherwin 2, Ashley Moffett 2, Karen Bieback 3, Susanne Clausmeyer 4, Leanne Masters 2, Roxana Maria Popovici 1, Alexandra Petra Hess 5, Thomas Strowitzki 1, and Michael von Wolff 1 Department of Gynecological Endocrinology and Reproductive Medicine, University Hospital Heidelberg, Heidelberg, Germany Department of Pathology, Cambridge University, Cambridge CB2 1QP, UK 3Institute of Transfusion Medicine and Immunology, University Heidelberg, Medical Faculty Mannheim, Mannheim, Germany 4Laboratory for Endocrinology and Genetics, Heidelberg, Germany 5 Department of Obstetrics and Gynecology, University of Du¨sseldorf, Du¨sseldorf, Germany 2

6

Correspondence address. Tel: þ49-6221-5637685; Fax: þ49-6221-565356; E-mail: [email protected]

abstract: The endometrium contains a distinct population of immune cells that undergo cyclic changes during the menstrual cycle and implantation. The majority of these leucocytes are uterine NK (uNK) cells, however how these cells interact with uterine stromal fibroblasts remains unclear. We therefore investigated the paracrine effect of medium conditioned by uterine decidual leucocytes (which are enriched for uNK cells) on the gene expression profile of endometrial stromal fibroblasts in vitro using a cDNA microarray. Our results, verified by real-time PCR, ELISA and FACS analysis, reveal that soluble factors from uterine leucocytes substantially alter endometrial stromal fibroblast gene expression. The largest group of up-regulated genes found was chemokines and cytokines. These include IL-8, CCL8 and CXCL1, which have also been shown to be stimulated by contact of stromal fibroblasts with trophoblast, suggesting that uNK cells work synergistically to support trophoblast migration during implantation. The decidual leucocytes also up-regulated IL-15 and IL-15Ra in stromal fibroblasts which could produce a niche for uNK cells allowing proliferation within and recruitment into the uterus, as seen in bone marrow. Overall this study demonstrates, for the first time, the paracrine communication between uterine leucocytes and uterine stromal fibroblasts, and adds to the understanding of how the uterine immune system contributes to the changes seen within the cycling endometrium. Key words: paracrine effect / endometrium / gene expression / uterine stromal fibroblast response / uterine leukocytes

Introduction In the human endometrium, uterine leucocytes undergo cyclic changes in cell number during the menstrual cycle. The majority of these (70%) are uterine NK (uNK) cells, whose numbers increase dramatically during the secretory phase and early pregnancy. This is due in part to NK cell proliferation within the endometrium and decidua, respectively (King et al., 1989; Verma et al., 2000). However, how this proliferation is controlled is not clearly understood. uNK cells lack nuclear progesterone receptors, so paracrine signals from the surrounding stromal fibroblasts may be responsible for mediating this effect (Henderson et al., 2003). There is also evidence that hormonally regulated chemokine expression within the endometrium stimulates the recruitment of blood NK cells and †

their differentiation into uNK cells contributing to the increase in numbers (Jones et al., 2004; Carlino et al., 2008). Interleukin-15 (IL-15) has been identified as one factor controlling these effects as it is up-regulated in endometrial stromal fibroblasts by progesterone in vitro as well as in vivo during the secretory phase and early pregnancy (Okada et al., 2000). uNK cells, express the IL-15 receptor a (IL-15Ra), and respond to IL-15 with increased proliferation and cytolytic activity (Verma et al., 2000). Endometrial extracts also promote migration of specific blood NK subsets, an effect that is blocked by antibodies to IL-15 (Kitaya et al., 2005). These studies emphasize the effects of endometrial stromal fibroblasts on the proliferation, recruitment and differentiation of uNK cells. The reciprocal interactions, that is, the effects of uNK, as well as other leucocytes and their secreted products on endometrial stromal

Both authors contributed equally to the study.

& The Author 2008. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please email: [email protected]

Downloaded from http://molehr.oxfordjournals.org/ by guest on January 29, 2015

1

40

Materials and Methods Collection of material Decidual tissue was collected from healthy women undergoing uncomplicated, legal termination of pregnancy during the first trimester for personal reasons. The gestational age of the decidual tissue ranged from 6 to 12 weeks, as determined from the last menstrual period. Stromal fibroblasts

were isolated from endometrial biopsies taken during the proliferative phase of the cycle (days 8 – 12) of healthy women of reproductive age undergoing laparoscopy (Women’s University Hospital, Heidelberg, Germany) for benign disease. The average age was 29 + 4.5 years. Exclusion criteria were hormonal stimulation, endocrinopathies, cancerous lesions and irregular menstrual bleeding. Histological examination was used to confirm cycle phase by the Department of Pathology, according to Noyes’ criteria (Noyes et al., 1950). All tissue samples were obtained after informed consent, and the study was approved by the University of Heidelberg Ethical Committee on human research.

Production of uterine leucocyte-conditioned medium Decidua was used as source of uterine leucocytes as they are much more abundant in this tissue, and phenotypically they closely resemble uterine leucocytes from the mid-late secretory phase endometrium (Trundley and Moffett, 2004). Uterine leucocytes were isolated from decidua as previously described (Verma et al., 2000). Briefly, decidua was separated from trophoblast and fetal tissue and digested in an enzyme mixture containing MCDB-105, Hyaluronidase V, DNAse I and Collagenase IV for 1 h at 378C (Popovici et al., 2000). The cells were separated by filtration through 100, 70 and 40 ml filters, to obtain a single cell suspension and applied to Histopaque (Sigma, Taufkirchen, Germany) to separate mononuclear cells from granulocytes and erythrocytes. To maximize the number of uNK cells in the isolated leucocytes, the cells were then incubated on plastic for 2 h at 378C in Dulbecco’s Modified Eagle’s Medium (DMEM) with 10% fetal calf serum (FCS). Using this method most adherent cells, including contaminating stromal fibroblasts and macrophages were removed. The non-adherent cells were harvested and resuspended in RPMI, with 1% charcoal stripped FCS (Perbio, Bonn, Germany), L-glutamine (2 mM) and IL-15 (10 ng/ml) (R&D Systems GmbH, Wiesbaden, Germany), as IL-15 has been shown to be essential to maintain viability of purified uNK cells (Verma et al., 2000). Leucocyte-conditioned media (LCM) was prepared using 200 ml of the RPMI media containing 2.5  105 of these purified cells, plated in the upper chamber of a 0.4 mm pore culture plate insert (Millicel, Millipore GmbH, Schwalbach, Germany) in a 24-well tissue plate with 600 ml of the same RPMI 1640 medium without cells in the lower chamber. The aim was to allow only soluble molecules from the leucocytes to pass through the filter into the conditioned medium below. As ‘control supernatant’ for the use during later experiments, 800 ml of the same RPMI solution [containing 1% FCS with L-glutamine (2 mM) and IL-15 (10 ng/ml)] was placed in the adjacent wells, without contact to uterine leucocytes and incubated in the same manner. The amount of uNK cells contained in the leucocyte-conditioned medium was analysed by fluorescent activated cell sorting (FACS) using antibodies to CD56, CD3, Leucogate and isotype controls (Becton – Dickinson, Heidelberg, Germany). Only cell preparations that showed over 85% CD56bright uNK cells were used for the supernatant collection. The remaining leucocytes comprised mostly CD3 positive T-cells with ,2% being dendritic cells or B cells (data not shown). After incubation for 22 h at 378C, 5% CO2, the LCM and the control medium, was collected and frozen at 2808C. Meanwhile, the cells in the upper chamber were collected and the viability measured using the live/dead viability kit according to the manufacturers’ instruction (Invitrogen, Karlsruhe, Germany). Only conditioned supernatant from uterine leucocyte populations with ,30% of dead cells after overnight incubation were used for further experiments.

Stromal fibroblast culture Stromal fibroblast culture was performed as described elsewhere (Popovici et al., 2006). Briefly, after separation of stromal fibroblasts from non-

Downloaded from http://molehr.oxfordjournals.org/ by guest on January 29, 2015

fibroblast gene expression and function, are unknown and are therefore addressed herein. During early implantation, extravillous trophoblast cells migrate into the decidua where they take part in remodelling the uterine vasculature to ensure an adequate blood supply for the developing fetus. Failure of trophoblast invasion is frequently associated with diseases of pregnancy such as intrauterine growth restriction and pre-eclampsia (Hiby et al., 2004). In comparison with peripheral NK cells, uNK cells are a unique population with a specific surface marker constellation (CD56þve CD9þve) and a different gene expression profile (Koopman et al., 2003). Furthermore, uNK cells are also very abundant at the implantation site where they interact directly with invading trophoblast cells, leading to the suggestion that uNK cells may regulate trophoblast invasion in vivo (Moffett and Loke, 2006). There is strong evidence that uNK cells can regulate human trophoblast invasion through the secretion of the cytokines IL-8 and IP-10 (Hanna et al., 2006). Also, uNK cells have been found to be essential in vascular remodelling during placentation through their production of interferon-gamma (IFN-g) in mice (Leonard et al., 2006). Although it is not clear whether the IFN-g produced by uNK cells acts directly on the vasculature or on the murine trophoblast cells, these results support the view that uNK cells have at least two functions in decidua: regulation of trophoblast invasion and spiral artery modification in decidua (Hiby et al., 2004). In the non-pregnant endometrium, there is accumulating evidence that uNK cells may play a role in regulating the process of decidualization itself. Long-standing observation has shown that decidualization of endometrial stromal fibroblasts is almost always associated with the presence of uNK cells (Ordi et al., 2006). Even in the non-pregnant menstrual cycle, localized areas of stromal fibroblasts undergo predecidual changes with many features of decidualization in the mid-late secretory phase (Acosta et al., 2000). These partially decidualized cells occur around blood vessels and the base of glands, where uNK cells and other leucocytes are particularly abundant. The fact that NK cells are present in the endometrium prior to the onset of decidualization suggests that decidualization is not essential for uNK cell localization to the endometrium. The goal of this study was to elucidate the paracrine effects of uterine leucocytes, containing mainly uNK cells (.85%), on the gene expression of undecidualized uterine stromal fibroblasts, the dominant cell type in the non-pregnant endometrium, using microarray analysis. Our results reveal that soluble factors from uterine leucocytes have substantial effects on endometrial stromal fibroblast gene expression. On one hand, this may well regulate the process of leucocyte recruitment and promote decidualization of stromal fibroblasts in preparation for trophoblast implantation. On the other hand, uterine leucocytes were found to up-regulate transcripts in stromal fibroblasts that are known to promote trophoblast migration after implantation has occurred.

Germeyer et al.

41

Uterine leukocyte effect on stromal fibroblast gene expression

pregnant endometrium by filtration and collagenase digestion, the cells were cultured until confluency in DMEM with 10% FCS medium and L-glutamine (2 mM) and passaged twice to remove contaminating cells. At the second passage, cells were placed in 4 cm petri dishes in the same medium. The cells were cultured for 6 – 8 days until fully confluent and ready to use. Contamination of the stromal fibroblasts was assessed by FACS with anti-CD45 (for leucocytes) and anti-CD31 (for endothelial cells; Sigma, Pool, UK), and found to be ,2% (data not shown).

Treatment of stromal fibroblasts with leucocyte-conditioned medium

Microarray experiments Microarray analysis was used to identify RNA transcripts that are altered in stromal fibroblasts by paracrine signals from uterine leucocytes. The expression profile of stromal fibroblasts treated with LCM or control medium was compared using a microarray containing 23 000 70mer oligonucleotides to human RefSeq cDNAs printed on a single slide. The microarrays were manufactured by the University of Cambridge Microarray Resource Centre in the Department of Pathology. The manufacture of the slides is described in Rossi et al. (2005). A full list of the cDNAs is available at: http://www.path.cam.ac.uk/resources/microarray/microarrays/ humanrefset23k.html. The RNA samples from the patients’ stromal fibroblasts treated with either LCM or control medium were labelled with Cy5 (test samples). A batch of common reference sample, produced by pooling RNA from

Statistical analysis The raw data set was normalized per spot and per chip using GeneSpring v7.0 software with intensity dependent (Lowess) normalization (percent of the data used for smoothing 10%) and per chip normalized to the 50th percentile. Low hybridization signals were removed to give an average of 10 000 different RNA transcripts expressed above background. For each cDNA spot on the array, a ratio was derived in which the signal from the test sample (stromal fibroblasts treated either with LCM or control medium and labelled with Cy5) was expressed relative to the hybridization signal in the Cy3-labelled common reference sample hybridized to the array at the same time. Transcripts that showed statistically significant changes between the two groups were identified using the Cyber-T programme, which employs a paired t-test modified using a Bayesian prior. Median fold-change ratios between the groups (i.e. cells treated with LCM or control medium) were subsequently derived for each transcript. Up- and down-regulated genes were selected on the basis that they showed a median fold change of at least two up or down and P , 0.001.

Real-time polymerase chain reaction Reverse transcription (RT) was performed with 1 mg of total RNA per 20 ml reaction using the first Strand cDNA Synthesis Kit (Roche Diagnostics GMBH) according to the manufacturers’ instructions. The real-time PCR primer sequences for target and reference genes were either taken from prior publications (ICAM-1 and CXCL1) (Hess et al., 2007) or selfdesigned to span an intron using public databases and were synthesized by MWG Biotech AG (Ebersberg, Germany). Primer sequences used are shown in Table I. Real-time PCRs were performed in triplicate to compare cDNA from stromal fibroblasts treated either with LCM or control medium using the LightCycler FastStart DNA Master SYBR Green I Kit (Roche Diagnostics) following the manufacturers’ instructions.

Table I Sequences of real-time primers, including their annealing temp, amplification length [in base pairs (bp)], as well as the magnesium chloride concentration. Gene

Forward primer

Reverse primer

Amplicon (bp)

Annealing temp (88 C)

Magnesium chloride conc ( mM)

............................................................................................................................................................................................... hRPL-19

GTAAGCGGAAGGGTACAGCCA

TTGTCTGCCTTCAGCTTGTG

211

58

3.0

ICAM-1

CCCACCATGAGGACATACAAC

GGCCTTTGTGTTTTGATGCTA

260

59

3.0

CXCL1

ATAGCCACACTCAAGAATG

TCTGCAGCTGTGTCTCTCTT

194

55

2.5

LIF

CCA ATG CCC TCT TTA TTC TCT

CAT AGC TTG TCC AGG TTG TTG

70

60

1.5

IL-8

ATCACTTCCAAGCTGGCCGTGGCT

TCTCAGCCCTCTTCAAAAACTTCTC

288

59

3.0

NNMT

CAGGAGCTGGAGAAGTGGCT

GGACCCTTGACTCTGTTCCCT

100

59

2.5

SCL11A3

CTGTTTGCAGGCGTCATTG

GAGCCAGGATGACCATGA

172

51

2.5

Downloaded from http://molehr.oxfordjournals.org/ by guest on January 29, 2015

LCM collected as described above was pooled from several batches for further experiments, to decrease interassay variability. Confluent stromal fibroblasts were placed in serum-free DMEM containing glutamine (2 mM) for 22 h prior to the experiment. The DMEM was then replaced with 80% LCM diluted with 20% serum free DMEM þ glutamine. This dilution was used following preliminary experiments using a range of concentrations from 5 to 80%. Eighty percent LCM gave the highest stromal fibroblast viability. Control cells were treated with 80% control medium (containing the same amount of IL-15), prepared as stated above, with 20% serum free DMEM þ glutamine. After a 6 h incubation in the conditioned or control media, the supernatants were collected and frozen at 2808C. RNA isolation from the cells was performed as described previously (Germeyer et al., 2007) using TRIZOL (Invitrogen, Carlsbad, USA), followed by digestion of contaminating DNA and purification with the High Pure RNA Isolation Kits (Roche Diagnostics, Mannheim, Germany). RNA quality was assessed by loading 300 ng of total RNA onto an RNA Labchip and analysed on an A2100 Bioanalyser (Agilent Technologies, Germany).

all stromal fibroblasts treated with control medium, was labelled in the same way with Cy3 (common reference sample). Similar amounts of each red test sample and the green-labelled reference sample were mixed and then co-hybridized to the same microarray to allow comparison of the hybridization intensity produced for each transcript against the common reference. The generation of the amplified labelled cDNA targets and the chip hybridization was performed using the method of Petalidis et al. (2003). The fluorescence signal on the microarrays was acquired by using a Genepix 4100 microarray scanner (Axon Instruments, Foster City, CA). The scanned images were processed using GenePix Pro 3.0 software (Axon Instruments).

42 All assays were optimized for primer concentration and magnesium chloride concentrations, based on the individual melting curve analysis and gel electrophoresis. For each real-time PCR, typical thermal cycling conditions included an initial activation step at 958C for 10 min, 40 cycles of amplification and a final melting curve (55– 958C) with continuous fluorescence measurement. A no-template (NTC) and no-reverse transcriptase (no enzyme added) control assured lack of DNA contamination and primer specificity. PCR efficiency was tested with each primer pair in a serial dilution. Only primers that had an efficiency coefficient between 1.9 and 2 were used. Threshold cycle (Ct) values were calculated using the LightCycler software based on the second derivative maximum method. The identity of the respective PCR products was confirmed by sequencing. Expression values of selected transcripts in control and treated cells were compared using the Ct values obtained by real-time RT– PCR after correction for variation between samples and normalization to the housekeeping gene hRPL19. Statistical significance of differences between the two groups was determined using a paired t-test with a significance level of P , 0.01.

IL-8 was analysed in LCM prior to stromal fibroblast treatment and in supernatants of stromal fibroblast culture treated either with control media or with LCM, using a commercially available enzyme-linked immunoabsorbent assay (ELISA) kit (R&D Systems GmbH) according to the manufacturer’s instructions. To determine the amount of IL-8 concentration secreted by the stimulated stromal fibroblasts, the starting IL-8 content of the leucocyte supernatant used for stimulation of stromal fibroblasts was subtracted. Paired t-test was used for statistical evaluation and the significance level was established with P , 0.01.

FACS staining for CD54 Following incubation with control or LCM, stromal fibroblasts were stained using mouse monoclonal antibody to CD54 (ICAM-1), to monitor how expression changed following treatment. A total of 5  105 cells were incubated with anti-CD54 antibody as recommended by the manufacturer (Zymed, San Francisco, USA), followed by goat anti-mouse-PE secondary antibody (Sigma, Pool, UK). An isotype matched antibody (mouse IgG1, Sigma, Pool, UK) was used as a negative control. After staining, cells were fixed for 10 min in 2% paraformaldehyde before analysis. The intensity of CD54 staining on each sample of stromal fibroblasts (n ¼ 4) treated with control or conditioned medium was then compared using a paired t-test, and the average fold increase was calculated.

Results Microarray data Gene expression profiling using an oligonucleotide microarray was used to compare transcript expression in stromal fibroblasts (n ¼ 7) treated with control medium or medium conditioned by uterine leucocytes. A detailed result list is available under the Accession number GSE9718 in the GEO database (www.ncbi.nlm.nih.gov/geo). This analysis identified 54 transcripts that exhibited a statistically significant change in median expression level of at least 2-fold up or down in response to LCM with P , 0.001. These transcripts are listed in Table II. Although 45 genes were up-regulated, only nine were downregulated. Using the Gene Ontology tree machine (http://bioinfo.vanderbilt.edu/gotm), the altered transcripts were ordered into their main functions within tissues. The largest group of up-regulated

genes were immunomodulatory gene regulators, and included a number of chemokines/cytokines, such as IL-8, IL-15, chemokine (C-X-C motif) ligand-1 (CXCL1), chemokine (C– C motif) ligand-8 (CCL-8), and IFN-induced protein 35 (IFI35), as well as cytokine receptors, including IL-7 receptor (IL-7R) and IL-15 receptor alpha (IL-15Ra). Also up-regulated were molecules responsible for regulating signal transduction, namely the chloride intracellular channel 2 (CLIC2), syndecan 4 (SDC4) and endothelial differentiation lysophosphatidic acid G-protein-coupled receptor 2 (EDG2). Adhesion molecules that may be involved in mediating cell chemotaxis, stimulated by the cytokines above, were also altered, such as the Intercellular adhesion molecule 1 (ICAM-1) and Pannexin-1 (PANX1). ICAM-1 (CD54) binds to the leucocyte integrins LFA-1, Mac-1 and p150,95, which are expressed on NK cells, and so would be expected to increase localized interactions between uNK cells and stromal fibroblasts. A number of transcripts encoding transcription factors were altered such as the NK6 transcription factor-related gene (NKX6-1), nuclear factor of kappa light polypeptide gene enhancer in B-cells (NFkB1), signal transducer and activator of transcription 3 (STAT3), and activating transcription factor 5 (ATF5), as well as several genes regulating apoptosis and anti-apoptosis, including SERPINB2, TNF receptorassociated factor 1 (TRAF1), receptor-interacting serine-threonine kinase 2 (RIPK2) and caspase 7 (CASP7) were up-regulated. Besides these, genes with specific functions, like the enzyme NNMT (nicotinamide N-methyltransferase) and tryptophan 2,3-dioxygenase (TDO2), the ion-binding S100 calcium-binding protein A3 (S100A3), and several genes of unknown function were found to be up-regulated. Down-regulated transcripts included only one gene of known function that was down-regulated .3-fold, namely the solute carrier family 11 member 3 (SLC11A3), an ion transporter, also called ferroportin. More than 2-fold down-regulated were the signal transducer RGS2 (Regulator of G-protein signalling 2) and transforming growth factor beta 2 (TGFB2), as well as the transcription factor MITF (microphthalmia-associated transcription factor) and the adhesion protein KANK (Kidney ankyrin repeat-containing protein).

Real-time RT – PCR In order to verify these changes, transcript levels for several genes were measured by real-time RT –PCR (n ¼ 6). These included CXCL1, IL-8, LIF, ICAM-1 and NNMT for the up-regulated genes and SLC11A for the down-regulated genes. The PCR results confirmed as statistically significant, the substantial changes in expression indicated by the microarray analysis (P , 0.01). Stromal fibroblasts treated with LCM showed increased expression of CXCL1 (5648.1-fold), IL-8 (1631.7-fold), LIF (251.1-fold), ICAM-1 (53.9-fold) and NNMT (6.3-fold). We also confirmed the downregulation of SLCA11 by 8.3-fold (Fig. 1).

ELISA To confirm these effects on the protein level, the amount of IL-8 was analysed by ELISA in the supernatant of the control and leucocyte-stimulated stromal fibroblasts. The median level of IL-8 in stromal fibroblast medium from cells incubated with control supernatant was 75.4 + 25.9 pg/ml (SEM). Stromal fibroblasts incubated with LCM had a median level of IL-8 of 64 265 + 15 679 pg/ml

Downloaded from http://molehr.oxfordjournals.org/ by guest on January 29, 2015

Enzyme-linked immunoabsorbent assay

Germeyer et al.

43

Uterine leukocyte effect on stromal fibroblast gene expression

Table II Transcripts altered more than two fold in stromal fibroblasts by stimulation with leucocyte-conditioned medium (P < 0.001) using a 23 000 gene cDNA microarray chip (n 5 7). Upregulated genes: 45 transcripts Gene symbol

Fold change

Genebank ID

Description

Cytokines/immune response/chemotaxis IL-8

90.0

GI_10834977

Interleukin 8

CXCL1

29.7

GI_4504152

Chemokine (C-X-C motif) ligand 1

LIF

7.9

GI_6006018

Leukemia inhibitory factor

CCL8

4.8

GI_22538815

Chemokine (C-C motif) ligand 8

IL7R

4.4

GI_4504678

Interleukin 7 receptor

IL15RA

3.6

GI_4504648

Interleukin 15 receptor alpha

IFI35

3.5

GI_24307900

Interferon-induced protein 35

CD83

3.3

GI_24475618

CD83 antigen

2.4

GI_5031782

Interferon gamma receptor 2

2.4

GI_23312365

Tumour necrosis factor receptor superfamily, 1B

IL15

2.2

GI_10835152

Interleukin 15

CXCL6

2.1

GI_4506850

Chemokine (C-X-C motif) ligand 6

Signal transduction CLIC2

3.9

GI_17105380

Chloride intracellular channel 2

SDC4

2.1

GI_4506860

Syndecan 4

EDG2

2.1

GI_16950637

Endothelial differentiation, lysophosphatidic acid G-protein-coupled receptor 2

Transcription NKX6-1

3.4

GI_5453787

NK6 transcription factor related, locus 1

NFKB1

2.8

GI_10835176

Nuclear factor of kappa light polypeptide gene enhancer in B-cells 1

STAT3

2.1

GI_21618339

Signal transducer and activator of transcription 3

ATF5

2.1

GI_12597624

Activating transcription factor 5

Apoptosis/anti-apoptosis SERPINB2

4.5

GI_4505594

Serine or cysteine proteinase inhibitor, clade B, 2

TRAF1

3.1

GI_22027610

TNF receptor-associated factor 1

RIPK2

2.3

GI_20127435

Receptor-interacting serine-threonine kinase 2

CASP7

2.0

GI_21536272

Caspase 7

ICAM1

23.4

GI_4557877

Intercellular adhesion molecule 1

PANX1

2.4

GI_7662507

Pannexin 1

2.8

GI_4506762

S100 calcium binding protein A3

2.3

GI_5032164

Tryptophan 2,3-dioxygenase

3.8

GI_5453789

Nicotinamide N-methyltransferase

2.4

GI_4502078

Adenosine monophosphate deaminase

2.0

GI_19923123

Colony stimulating factor 1

Cell– cell adhesion

Ion binding S100A3 Tryptophan metabolism TDO2 Enzyme activity NNMT Nucleotide metabolism AMPD3 Cell proliferation/cell differentation CSF1 Others NMES1

6.5

GI_14165279

Normal mucosa of oesophagus specific 1

KIAA0062

5.3

GI_22050863

KIAA0062 protein

SNFT

3.7

GI_8924245

Jun dimerization protein p21SNFT

FLJ11 125

3.5

GI_22050849

Hypothetical protein

Continued

Downloaded from http://molehr.oxfordjournals.org/ by guest on January 29, 2015

IFNGR2 TNFRSF1B

44

Germeyer et al.

Table II Continued SQRDL

2.6

GI_10864010

Sulfide dehydrogenase like (yeast)

MAIL

2.4

GI_13899228

Molecule possessing ankyrin repeats induced by lipopolysaccharide

LOC199695

2.4

GI_20486277

LOC199695

KIAA1404

2.4

GI_14786303

KIAA1404 protein

ZAP

2.3

GI_24762227

Likely ortholog of rat zinc-finger antiviral protein

FLJ21 709

2.2

GI_14333696

Hypothetical protein

KIAA1726

2.2

GI_22065899

KIAA1726 protein

LOC221228

2.1

GI_20562873

LOC221228

MSCP

2.1

GI_8924027

Likely ortholog of mouse mitochondrial solute carrier prot.

DKFZP667O116

2.0

GI_20542967

Hypothetical protein

LOC169308

2.0

GI_18570971

LOC169308

Genebank ID

description

0.14

GI_19923794

Solute carrier family 11 member 3

0.41

GI_4557754

Microphthalmia-associated transcription factor

Downregulated genes: 9 transcripts Gene symbol

Fold change

SLC11A3 Transcription factor MITF Signal transduction TGFB2

0.42

GI_4507462

Transforming growth factor, beta 2

RGS2

0.47

GI_4506516

Regulator of G-protein signalling 2

0.45

GI_23510376

Kidney ankyrin repeat-containing protein, transcript variant 2

Cell adhesion KANK Others LOC143909

0.31

GI_18604592

LOC143909

DKFZP566J091

0.33

GI_13569871

Hypothetical protein

MGC11 324

0.38

GI_21362091

Hypothetical protein

KIAA1913

0.39

GI_16172557

KIAA1913 protein

Genes are listed with their function determined by the gene ontology tree machine.

(Fig. 2). There was an average up-regulation of 843.4-fold in IL-8 secretion after leucocyte stimulation (n ¼ 6). Since IL-8 levels in the control or LCM medium prior to treatment of the stromal fibroblasts could contribute to the final level of IL-8, the amount of IL-8 in each was measured before addition to the fibroblasts. Control medium contained negligible amounts of IL-8, whereas LCM contained an average of (671.7 + 270.3 pg/ml) of IL-8 before it was used to treat the stromal fibroblasts. This value was subtracted from the final IL-8 concentration obtained after treatment of the fibroblasts with the corresponding LCM. Therefore, the amount of IL-8 in the LCM prior to treatment of the fibroblasts does not contribute significantly to the total IL-8 concentration obtained after incubation of fibroblasts with LCM. This effect was consistent in all samples examined and was statistically significant (P , 0.01).

FACS staining for CD54 (ICAM-1) For further confirmation at the protein level, treated stromal fibroblasts were stained with a monoclonal antibody specific for CD54 (ICAM-1) to determine whether expression of this molecule on the cell surface was altered by incubation of stromal fibroblasts with LCM (n ¼ 4). Stromal fibroblasts were gated using forward and side

scatter and all were found to be positively stained for CD54. The mean fluorescence intensity of CD54 staining of stromal fibroblasts after 6 h treatment with control or LCM medium was 2.3-fold higher (n ¼ 4, range 1.6 –4.0; P , 0.05) (Fig. 3, Panel B).

Discussion This study investigates for the first time the paracrine communication between uterine leucocytes and endometrial stromal fibroblasts. The changes in gene expression seen are likely to be principally due to secreted factors from uNK cells since these constitute at least 85% of our decidual leucocyte preparations. However, a contribution from the small number of T- and dendritic cells in the leucocyte preparation cannot be excluded. The effect of factors derived exclusively from decidual T-cells on stromal fibroblast gene expression is therefore unclear. The increase in uNK cells in the endometrium that is found during the secretory phase has previously been ascribed to the progesterone-induced expression of chemokines capable of attracting NK cells, such as CCL4 (MIP1b) and CCL21 (6Ckine) (Jones et al., 2004; Carlino et al., 2008). In addition, vigorous proliferation in response to IL-15 is thought to contribute to the rise in cell

Downloaded from http://molehr.oxfordjournals.org/ by guest on January 29, 2015

Channels/ transport molecule

Uterine leukocyte effect on stromal fibroblast gene expression

45

numbers (Verma et al., 2000). These uNK cells are known to participate in angiogenesis and to regulate trophoblast migration during pregnancy through secretion of factors such as IL-8 and VEGF (Li et al., 2001; Hanna et al., 2006; Lash et al., 2006). The data of the present study demonstrate now that these cells also have a profound effect on stromal fibroblast gene expression in the non-pregnant state. The results suggest that they influence leucocyte migration into the endometrium, as well as stromal fibroblast decidualization and several new processes, such as the regulation of the tryptophan catabolism.

Figure 2 IL-8 levels measured by ELISA in supernatants from stromal fibroblasts treated for 6 h with conditioned medium (nk supernatent) from decidual leucocytes or with control medium (value + SEM), depicted using a logscale.Conditioned medium induced an average increase (of 843-fold) in IL-8 secretion compared with control medium (n ¼ 6, P , 0.01). Values were corrected for the amount of IL-8 detected in the control and conditioned medium prior to addition to the stromal fibroblasts, then plotted on a log10 \scale with SEM represented by error bars.

Figure 3 Surface expression of CD54 (ICAM-1) is increased on endometrial stromal fibroblasts treated with leucocyte-conditioned medium. The CD54 expression of stromal fibroblasts, incubated with either control or leucocyte-conditioned medium, was analysed by FACS after gating with forward and side scatter (panel A). CD54 expression was significantly increased in the cells treated with the leucocyte-conditioned medium (filled graph), compared with the control-treated fibroblasts (pale line). There was negligible staining with the isotype control antibody (black line).

The largest group of genes, induced in stromal fibroblasts by soluble factors derived from uterine leucocytes, were cytokines and chemokines, including CXCL1 and IL-8. Expression of IL-8 in the endometrium increases in the late secretory phase particularly in perivascular stromal fibroblasts (Milne et al., 1999; Jones et al., 2004). Since uNK cells are also abundant around blood vessels, our data suggest that they may contribute to this perivascular rise in IL-8 and CXCL1. By up-regulating these two factors, uterine leucocytes therefore activate a powerful autocrine loop since endometrial stromal fibroblasts also express the corresponding receptors CXCR1 and CXCR2 (Mulayim et al., 2003). IL-8 and CXCL1 can also activate neutrophils and influence angiogenesis (Scapini et al., 2004). A similar autocrine effect is induced through the simultaneous up-regulation of IL-15 and IL-15Ra that was observed on stromal fibroblasts. IL-15 expression in endometrium increases during the secretory phase and, like IL-8, is also found in perivascular areas (Milne et al., 1999). Soluble protein extracts of secretory phase endometrium stimulate selective migration of peripheral blood CD56bright,

Downloaded from http://molehr.oxfordjournals.org/ by guest on January 29, 2015

Figure 1 Real-time RT – PCR verification (n ¼ 6) of transcripts altered by microarray analysis: theshhold cycle (Ct) values of the genes of interest (CXCL1, IL-8, LIF, ICAM-1, NNMT and SLC11A) were normalized to Ct values of the housekeeping gene RPL19.Fold changes are the (up- or down-regulation of gene expression in stromal fibroblasts treated with leucocyte-conditioned medium (nk supernatent) compared to incubation with control medium. Values are plotted on a log10scale with SEM represented by error bars. Significance using a paired t-test is denoted as *P , 0.05; **P , 0.005.

46

stromal fibroblasts increase LIF expression upon immune cell stimulation may also explain the LIF increase previously reported in stromal fibroblasts of fallopian tube tissue sections during ectopic implantation (Senturk and Arici, 1998). The transcript encoding Tryptophan 2,3-dioxygenase (TDO) was also increased in stromal fibroblasts by LCM. TDO is a key enzyme in the catabolism of tryptophan that is induced in decidualized stromal fibroblasts around the implanting embryo in mice (Tatsumi et al., 2000). Furthermore, localized depletion of tryptophan at the implantation site by the related enzyme indoleamine 2,3-dioxygenase (IDO) has been shown to be critical in preventing immunological rejection of the fetal allograft in mice (Munn et al., 1998). In humans, IDO expression increases in endometrial stromal fibroblasts following decidualization. This induction of IDO expression appears to be in response to IFN-g secreted by infiltrating leucocytes, while progesterone suppesses IDO expression (Kudo et al., 2004). To our knowledge, there are no published studies of TDO in human endometrium, but these results suggest that TDO may also regulate tryptophan levels in stromal fibroblasts and its expression may also be stimulated by leucocytes. While the origin of uNK cells is still not clear, they proliferate and differentiate in the stromal microenvironment of the endometrium. Several of the molecules altered by uterine leucocytes in our study are known to have important functions in NK development, including IL-15 and IL-15Ra. There is also evidence that the endometrium may harbour haematopoietic stem cell precursors that can differentiate into NK cells (Keskin et al., 2007). The reciprocal interactions between stromal fibroblasts and uterine leucocytes found in this study are reminiscent of those seen in bone marrow during NK development. NK cells up-regulate IL-15 in bone marrow stromal cells, which is then bound and presented by IL-15Ra on the stromal cell surface, promoting increased local NK cell proliferation (Iizuka et al., 1999; Sandau et al., 2004). Our results suggest that a similar mechanism may occur in endometrium since both IL-15 and IL-15Ra are up-regulated in endometrial stromal fibroblasts by uterine leucocyte supernatant. In support of this, IL-15Ra expression by endometrial stromal fibroblasts in vivo has previously been shown to increase during the secretory phase (Lobo et al., 2004). Roberts et al. (2005) have also shown that IL-15 is up-regulated in stromal fibroblasts not only by progesterone, but also by TNFa a cytokine released by uNK cells. These results suggest a mechanism by which uNK cells and non-decidualised stromal fibroblasts may cooperate to maintain immune cell homeostasis in the endometrial microenvironment. Overall, we demonstrate for the first time the paracrine effects of decidual leucocytes on endometrial stromal fibroblasts. Decidual leucocytes were used as it is difficult to isolate sufficient leucocytes from endometrium to prepare conditioned medium. The phenotype of the dominant leucocyte (NK cells) in decidua and non-pregnant endometrium appear broadly similar, in that they are poorly cytotoxic and, unlike blood NK cells, they both secrete angiogenic cytokines. However, it is important to note that the phenotype of NK cells from endometrium is poorly characterized compared with the decidua and further work will be required to confirm whether the effects of leucocytes isolated from endometrium are the same as reported here (Manaster and Mandelboim, 2008). The up-regulation of factors like IL-8, ICAM-1 and CXCL1 by uterine decidual leucocytes in the endometrium, prior to the contact of trophoblast, with endometrial tissue suggests that uterine leucocytes influence the early

Downloaded from http://molehr.oxfordjournals.org/ by guest on January 29, 2015

CD162ve NK cells in vitro. This effect is abolished by a neutralizing antibody to IL-15 (Okada et al., 2000). Taken together, these data support the opinion that uterine leucocytes in situ can promote the perivascular production of IL-8 and IL-15 by endometrial stromal fibroblasts, which could contribute to leucocyte migration into the endometrium in a self-reinforcing loop. Two recent microarray studies have examined the effects of trophoblast-derived factors on endometrial stromal fibroblasts. Popovici et al. (2006) examined the effects of incubating trophoblast explants with non-decidualized stromal fibroblasts, while Hess et al. (2007) incubated decidualized stromal fibroblasts with medium conditioned by trophoblast cells. Remarkably, many of the same transcripts that we have found to be up-regulated by uterine leucocytes are also up-regulated by trophoblast. Indeed, IL-8, CXCL1 and CCL8 were among the most abundant up-regulated transcripts in all three studies. This remarkable concordance strongly suggests that the same soluble mediators may be produced by both decidual leucocytes and trophoblast, as they bring about similar changes on both decidualized and non-decidualized stromal fibroblasts. The local up-regulation of factors such as IL-8 and CXCL1 in the stroma, by leucocytes accumulating at the site of trophoblast invasion, may thus help to promote trophoblast invasion, since the relevant receptors are expressed in trophoblast (Drake et al., 2004; Hanna et al., 2006). IL-1b produced by leucocytes may be one of the responsible mediators, since it upregulates IL-8 and CXCL1 in endometrial stromal fibroblasts (Rossi et al., 2005). Furthermore, trophoblast and uNK cells both secrete significant amounts of IL-1b (Bennett et al., 1999). Whether IL-1b is indeed the relevant factor could be tested by prior treatment of uNK and trophoblast supernatants with neutralizing antibodies to this cytokine. Nevertheless, implantation of the embryo is not absolutely dependant on the presence of leucocytes such as uNK cells within the implantation site (Ordi et al., 2006). IL-15 knockout mice lacking uNK cells can reproduce although they show abnormalities in decidualization and reduced litter weights, indicating that in mice, uNK cells may influence the process of decidualization itself (Barber and Pollard, 2003). We found some evidence that uterine leucocytes promote changes in human stromal fibroblasts characteristic of decidualization. IL-8 and CXCL1, which were up-regulated in stromal fibroblasts by LCM, also increase dramatically during decidualization of stromal fibroblasts (Popovici et al., 2000; Nasu et al., 2001). Our data suggest that leucocytes may prime some of these changes in the stroma prior to the onset of full decidualization. Nevertheless, since purified endometrial stromal fibroblasts can decidualize when cultured with estrogen and progesterone or CAMP without leucocytes, leucocytes are not essential. We are currently investigating whether conditioned medium from decidual leucocytes alters the rate of in vitro decidualization stimulated by progesterone, using known markers for in vitro decidualization (Telgmann and Gellersen, 1998). This global approach has also identified several previously unsuspected effects of leucocytes on stromal fibroblasts. Expression of LIF, a gene essential for implantation in mice, increases during the receptive phase in both mice and humans (Stewart et al., 1992). While the glandular epithelium is the primary source for LIF in endometrium, endometrial stroma also exhibits some LIF immunostaining in the secretory phase (Dimitriadis et al., 2006). Our finding that

Germeyer et al.

Uterine leukocyte effect on stromal fibroblast gene expression

Acknowledgements Many thanks to Mrs J. Jauckus for her outstanding contribution in the performance of the experiments. Thanks too to the patients without whom this study would not have been possible. Microarrays were kindly provided by the BBSRC microarray group in the Department of Pathology, Cambridge (BBSRC grant No. 8/EGH16 106). AS was supported by The Wellcome Trust (GRO76 850).

Funding Grants: Department of Pathology, Cambridge (BBSRC grant No. 8/EGH16 106), AS: the Wellcome Trust (GRO76 850).

References Acosta AA, Elberger L, Borghi M, Calamera JC, Chemes H, Doncel GF, Kliman H, Lema B, Lustig L, Papier S. Endometrial dating and determination of the window of implantation in healthy fertile women. Fertil Steril 2000;73:788– 798. Barber EM, Pollard JW. The uterine NK cell population requires IL-15 but these cells are not required for pregnancy nor the resolution of a Listeria monocytogenes infection. J Immunol 2003;171:37 – 46. Bennett WA, Lagoo-Deenadayalan S, Whitworth NS, Stopple JA, Barber WH, Hale E, Brackin MN, Cowan BD. First-trimester human chorionic villi express both immunoregulatory and inflammatory cytokines: a role for interleukin-10 in regulating the cytokine network of pregnancy. Am J Reprod Immunol 1999;41:70 – 78. Carlino C, Stabile H, Morrone S, Bulla R, Soriani A, Agostinis C, Bossi F, Mocci C, Sarazani F, Tedesco F et al. Recruitment of circulating NK cells through decidual tissues: a possible mechanism controlling NK cell accumulation in the uterus during early pregnancy. Blood 2008;111: 3108 – 3115. Dimitriadis E, Stoikos C, Stafford-Bell M, Clark I, Paiva P, Kovacs G, Salamonsen LA. Interleukin-11, IL-11 receptoralpha and leukemia

inhibitory factor are dysregulated in endometrium of infertile women with endometriosis during the implantation window. J Reprod Immunol 2006;69:53 – 64. Drake PM, Red-Horse K, Fisher SJ. Reciprocal chemokine receptor and ligand expression in the human placenta: implications for cytotrophoblast differentiation. Dev Dyn 2004;229:877 – 885. Germeyer A, Klinkert MS, Huppertz AG, Clausmeyer S, Popovici RM, Strowitzki T, von Wolff M. Expression of syndecans, cell-cell interaction regulating heparan sulfate proteoglycans, within the human endometrium and their regulation throughout the menstrual cycle. Fertil Steril 2007;87:657– 663. Hanna J, Goldman-Wohl D, Hamani Y, Avraham I, Greenfield C, Natanson-Yaron S, Prus D, Cohen-Daniel L, Arnon TI, Manaster I et al. Decidual NK cells regulate key developmental processes at the human fetal-maternal interface. Nat Med 2006;12:1065 – 1074. Henderson TA, Saunders PT, Moffett-King A, Groome NP, Critchley HO. Steroid receptor expression in uterine natural killer cells. J Clin Endocrinol Metab 2003;88:440 – 449. Hess AP, Hamilton AE, Talbi S, Dosiou C, Nyegaard M, Nayak N, Genbecev-Krtolica O, Mavrogianis P, Ferrer K, Kruessel J et al. Decidual stromal cell response to paracrine signals from the trophoblast: amplification of immune and angiogenic modulators. Biol Reprod 2007;76:102– 117. Hiby SE, Walker JJ, O’shaughnessy KM, Redman CW, Carrington M, Trowsdale J, Moffett A. Combinations of maternal KIR and fetal HLA-C genes influence the risk of preeclampsia and reproductive success. J Exp Med 2004;200:957 – 965. Iizuka K, Chaplin DD, Wang Y, Wu Q, Pegg LE, Yokoyama WM, Fu YX. Requirement for membrane lymphotoxin in natural killer cell development. Proc Natl Acad Sci USA 1999;96:6336 – 6340. Jones RL, Hannan NJ, Kaitu’u TJ, Zhang J, Salamonsen LA. Identification of chemokines important for leucocyte recruitment to the human endometrium at the times of embryo implantation and menstruation. J Clin Endocrinol Metab 2004;89:6155 – 6167. Keskin DB, Allan DS, Rybalov B, Andzelm MM, Stern JN, Kopcow HD, Koopman LA, Strominger JL. TGFbeta promotes conversion of CD16þ peripheral blood NK cells into CD16- NK cells with similarities to decidual NK cells. Proc Natl Acad Sci USA 2007; 104:3378– 3383. King A, Wellings V, Gardner L, Loke YW. Immunocytochemical characterization of the unusual large granular lymphocytes in human endometrium throughout the menstrual cycle. Hum Immunol 1989; 24:195– 205. Kitaya K, Yamaguchi T, Honjo H. Central role of interleukin-15 in postovulatory recruitment of peripheral blood CD16(2) natural killer cells into human endometrium. J Clin Endocrinol Metab 2005;90:2932 – 2940. Koopman LA, Kopcow HD, Rybalov B, Boyson JE, Orange JS, Schatz F, Masch R, Lockwood CJ, Schachter AD, Park PJ et al. Human Decidual Natural Killer Cells Are a Unique NK Cell Subset with Immunomodulatory Potential. J Exp Med 2003;8:1201 – 1212. Kudo Y, Hara T, Katsuki T, Toyofuku A, Katsura Y, Takikawa O, Fujii T, Ohama K. Mechanisms regulating the expression of indoleamine 2,3-dioxygenase during decidualization of human endometrium. Hum Reprod. 2004;19:1222– 1230. Lash GE, Schiessl B, Kirkley M, Innes BA, Cooper A, Searle RF, Robson SC, Bulmer JN. Expression of angiogenic growth factors by uterine natural killer cells during early pregnancy. J Leukoc Biol 2006;80:572 – 580. Leonard S, Murrant C, Tayade C, van den Heuvel M, Watering R, Croy BA. Mechanisms regulating immune cell contributions to spiral artery modification – facts and hypotheses – a review. Placenta 2006; 27(Suppl. A):S40 – S46.

Downloaded from http://molehr.oxfordjournals.org/ by guest on January 29, 2015

recruitment of immune cells into the endometrium during decidualization. Many of these factors such as IL-8 may also play a role in endometrial angiogenesis during the menstrual cycle. In addition, several of the identified factors (e.g. IL-8, CCL6) are regulated in a similar fashion by the action of trophoblast supernatant on decidualized stromal fibroblasts (Hess et al., 2007). This indicates that decidual leucocytes may prime the stroma prior to trophoblast implantation and soluble factors from trophoblast then amplify these changes. Both these mechanisms contribute to a dramatic rise in IL-8 secretion in stromal fibroblasts, and could therefore facilitate implantation as IL-8 has been shown to promote trophoblast migration into the decidua (Hanna et al., 2006). Finally, we have shown that uterine leucocytes up-regulate both IL-15 and IL-15Ra transcripts in uterine stromal fibroblasts. This may contribute to the creation of a ‘niche’ in the endometrium in which IL-15 is presented by IL-15Ra on stromal fibroblasts to promote uNK proliferation, as seen in bone marrow, contributing to the dramatic increase in uNK cell numbers during the secretory phase. This is the first detailed study of the paracrine interaction between immune cells and stromal fibroblasts in the endometrial environment. The next step will be to identify which secreted factors mediate these interactions between uterine immune cells and stromal fibroblasts in the endometrium, as well as in the decidua, to assess the paracrine effect of immune cells on the process of implantation.

47

48

Popovici RM, Betzler NK, Krause MS, Luo M, Jauckus J, Germeyer A, Bloethner S, Schlotterer A, Kumar R, Strowitzki T et al. Gene expression profiling of human endometrial-trophoblast interaction in a coculture model. Endocrinology 2006;147:5662 – 5675. Roberts M, Luo X, Chegini N. Differential regulation of interleukins IL-13 and IL-15 by ovarian steroids, TNF-alpha and TGF-beta in human endometrial epithelial and stromal cells. Mol Hum Reprod 2005; 11:751 – 760. Rossi M, Sharkey AM, Vigano P, Fiore G, Furlong R, Florio P, Ambrosini G, Smith SK, Petraglia F. Identification of genes regulated by interleukin-1beta in human endometrial stromal cells. Reproduction 2005;130:721 – 729. Sandau MM, Schluns KS, Lefrancois L, Jameson SC. Cutting edge: transpresentation of IL-15 by bone marrow-derived cells necessitates expression of IL-15 and IL-15R alpha by the same cells. J Immunol 2004;173:6537 – 6541. Scapini P, Morini M, Tecchio C, Minghelli S, Di Carlo E, Tanghetti E, Albini A, Lowell C, Berton G, Noonan DM et al. CXCL1/ macrophage inflammatory protein-2-induced angiogenesis in vivo is mediated by neutrophil-derived vascular endothelial growth factor-A. J Immunol 2004;172:5034– 5040. Senturk LM, Arici A. Leukemia inhibitory factor in human reproduction. Am J Reprod Immunol 1998;39:144 – 151. Stewart CL, Kaspar P, Brunet LJ, Bhatt H, Gadi I, Kontgen F, Abbondanzo SJ. Blastocyst implantation depends on maternal expression of leukaemia inhibitory factor. Nature 1992;359:76 – 79. Tatsumi K, Higuchi T, Fujiwara H, Nakayama T, Egawa H, Itoh K, Fujii S, Fujita J. Induction of tryptophan 2,3-dioxygenase in the mouse endometrium during implantation. Biochem Biophys Res Commun 2000; 274:166– 170. Telgmann R, Gellersen B. Marker genes of decidualization: activation of the decidual prolactin gene. Hum Reprod Update 1998;4:472– 479. Trundley A, Moffett A. Human uterine leucocytes and pregnancy. Tissue Antigens 2004;63:1 – 12. Verma S, Hiby SE, Loke YW, King A. Human decidual natural killer cells express the receptor for and respond to the cytokine interleukin 15. Biol Reprod 2000;62:959 – 968. Submitted on December 13, 2008; resubmitted on November 24, 2008; accepted on December 1, 2008

Downloaded from http://molehr.oxfordjournals.org/ by guest on January 29, 2015

Li XF, Charnock-Jones DS, Zhang E, Hiby S, Malik S, Day K, Licence D, Bowen JM, Gardner L, King A et al. Angiogenic growth factor messenger ribonucleic acids in uterine natural killer cells. J Clin Endocrinol Metab 2001;86:1823 – 1834. Lobo SC, Huang ST, Germeyer A, Dosiou C, Vo KC, Tulac S, Nayak NR, Giudice LC. The immune environment in human endometrium during the window of implantation. Am J Reprod Immunol 2004; 52:244 – 251. Manaster I, Mandelboim O. The unique properties of human NK cells in the uterine mucosa. Placenta 2008;29(Suppl. A):S60 – S66. Review. Milne SA, Critchley HO, Drudy TA, Kelly RW, Baird DT. Perivascular interleukin-8 messenger ribonucleic acid expression in human endometrium varies across the menstrual cycle and in early pregnancy decidua. J Clin Endocrinol Metab 1999;84:2563 – 2567. Moffett A, Loke C. Immunology of placentation in eutherian mammals. Nat Rev Immunol 2006;6:584– 594. Mulayim N, Palter SF, Kayisli UA, Senturk L, Arici A. Chemokine receptor expression in human endometrium. Biol Reprod 2003;68:1491– 1495. Munn DH, Zhou M, Attwood JT, Bondarev I, Conway SJ, Marshall B, Brown C, Mellor AL. Prevention of allogeneic fetal rejection by tryptophan catabolism. Science 1998;281:1191 – 1193. Nasu K, Fujisawa K, Arima K, Kari K, Sugano T, Miyakawa I. Expression and regulation of growth-regulated oncogene alpha in human endometrial stromal cells. Mol Hum Reprod 2001;7:741 – 746. Noyes RW, Hertig AT, Rock J. Dating the endometrial biopsy. Fertil Steril 1950;1:3 – 25. Okada H, Nakajima T, Sanezumi M, Ikuta A, Yasuda K, Kanzaki H. Progesterone enhances interleukin-15 production in human endometrial stromal cells in vitro. J Clin Endocrinol Metab 2000;85:4765–4770. Ordi J, Casals G, Ferrer B, Creus M, Guix C, Palacin A, Campo E, Balasch J. Uterine (CD56þ) natural killer cells recruitment: association with decidual reaction rather than embryo implantation. Am J Reprod Immunol 2006;55:369 – 377. Petalidis L, Bhattacharyya S, Morris GA, Collins VP, Freeman TC, Lyons PA. Global amplification of mRNA by template-switching PCR: linearity and application to microarray analysis. Nucleic Acids Res 2003; 31:e142. Popovici RM, Kao LC, Giudice LC. Discovery of new inducible genes in in vitro decidualized human endometrial stromal cells using microarray technology. Endocrinology 2000;141:3510 – 3513.

Germeyer et al.