inhibits extravillous trophoblast cell invasion by a ... - The FASEB Journal

The FASEB Journal • Research Communication

Interferon-␥ inhibits extravillous trophoblast cell invasion by a mechanism that involves both changes in apoptosis and protease levels Gendie E. Lash,*,1 Harry A. Otun,* Barbara A. Innes,* Maureen Kirkley,* Leandro De Oliveira,* Roger F. Searle,‡ Stephen C. Robson,* and Judith N. Bulmer† Schools of *Surgical and Reproductive Sciences, †Clinical and Laboratory Sciences and ‡Medical Education Development, University of Newcastle upon Tyne, Newcastle upon Tyne, UK Background: Extravillous trophoblast cell (EVT) invasion of decidua and inner third of the myometrium is critical for a successful pregnancy. Many decidual factors are likely to play a role in regulating this process, including uterine natural killer (uNK) cell-derived cytokines. Hypotheses: 1) uNK cells are a major source of IFN gamma (IFN-␥) and 2) IFN-␥ inhibits EVT invasion via an increase in EVT apoptosis and/or a decrease in active protease levels. Methods: Total decidual and uNK cells from 8 –10 wk and 12–14 wk gestational age were cultured. IFN-␥ mRNA (realtime RT-polymerase chain reaction) and protein levels (FastQuant multicytokine analysis) were determined. EVT invasion in the presence of IFN-␥ or anti-IFN-␥neutralizing antibodies was assessed. Trophoblast apoptosis and proliferation was assessed in explants by immunohistochemistry for M30 and Ki67. Substrate zymography was performed to determine levels of secreted MMP2, MMP9, and uPA. Results: mRNA and protein for IFN-␥ was detected in both total decidual and uNK cell fractions. Trophoblast invasion was inhibited by IFN-␥. The level of M30-positive EVT was increased in the presence of IFN-␥ whereas levels of secreted MMP2 were decreased. Conclusions: uNK cells are a source of IFN-␥ within early human pregnancy decidua. Mechanisms of IFN-␥ inhibition of EVT invasion include both increased EVT apoptosis and reduced levels of active proteases.—Lash, G. E., Otun, H. A., Innes, B. A., Kirkley, M., De Oliveira, L., Searle, R. F., Robson, S. C., Bulmer, J. N. Interferon-␥ inhibits extravillous trophoblast cell invasion by a mechanism that involves both changes in apoptosis and protease levels. FASEB J. 20, 2512–2518 (2006) ABSTRACT

Key Words: IFN-␥ 䡠 uNK cells The establishment of a successful pregnancy requires extravillous trophoblast cell (EVT) invasion into uterine decidua and the inner third of myometrium along with remodeling of the spiral arteries (1). EVT invasion proceeds by two different routes: interstitial whereby EVT invade through the decidual and myometrial stroma, and endovascular whereby EVT migrate in 2512

a retrograde direction up the lumen of the spiral arteries (1, 2). Trophoblast invasion and spiral artery remodeling are both complex processes requiring tight regulation. Incomplete EVT invasion and spiral artery remodeling can lead to a range of major pregnancy complications including second trimester miscarriage (3), preeclampsia (4), and fetal growth restriction (5). EVT are a naturally highly invasive cell type that, left unchecked, will proceed through the myometrium, as observed in placenta accreta (6). The focus of research into the control of EVT invasion has therefore shifted from investigation of factors that may promote this process to those that may inhibit it. Indeed, in the past few years several cytokines and growth factors, including vascular endothelial growth factor (VEGF) -A (7), TGF-␤1, 2, and 3 (8), and TNF-␣ (9), have been shown to inhibit EVT invasion by various mechanisms including altered EVT apoptosis, proliferation, and protease activity. Cellular invasion can be regulated by altering rates of apoptosis and/or proliferation leading to either increased or decreased numbers of cells available to invade. Uterine natural killer cells (uNK) are the major leukocyte cell type in the decidua during early pregnancy and are characterized by their CD56 bright/ CD16 dim phenotype (reviewed in ref. 10). In the human, uNK cell numbers are maximal during the first 20 wk of pregnancy, during the period of trophoblast invasion and spiral artery remodeling, before declining toward term (10). They have been shown to have cytotoxic, immunoregulatory, and secretory activities in in vitro studies but their in vivo role is not known (10). Among other functions, uNK cells are a major source of cytokines and growth factors; a proposed role for uNK cells in vivo is regulation of EVT invasion via these soluble products. Mouse and rat uNK cells have been shown to produce IFN gamma (IFN-␥), which has been shown to 1 Correspondence: School of Surgical and Reproductive Sciences, 3rd Floor, William Leech Bldg., University of Newcastle upon Tyne, Newcastle upon Tyne NE2 4HH, UK. E-mail: [email protected] doi: 10.1096/fj.06-6616com

0892-6638/06/0020-2512 © FASEB

inhibit trophoblast outgrowth from blastocysts in vitro (11, 12). Moreover, Ain et al. (12) demonstrated that EVT invasion of the metrial gland did not occur until numbers of uNK cells declined later in pregnancy. They also demonstrated that in uNK or IFN-␥ signalingdeficient mice, the timing of EVT invasion of the mesometrial decidua was advanced compared with wild-type (WT) animals (12). In addition, in the mouse, uNK cells and IFN-␥ signaling have been shown to be involved in spiral artery remodeling (13, 14). Several groups have investigated human uNK cell production of IFN-␥ by flow cytometry and ELISA after cell culture (and often activation by PMA stimulation), and only low levels have been reported (15–17). However, IFN-␥ is rapidly secreted and, depending on the experimental protocol, may be lost from cells and therefore not detected in vitro (18). Furthermore, immunolocalization of IFN-␥ is unsatisfactory (19). Only one previous study has investigated the effect of IFN-␥ on trophoblast invasion. Karmakar et al. (20) demonstrated inhibition of invasion of the choriocarcinoma cell line JEG3 by IFN-␥, which was associated with a decrease in matrix metalloproteinase (MMP) and urokinase plasminogen activator (uPA) activities. In the current study we hypothesized that 1) uNK cells are a source of IFN-␥ in the decidua and 2) IFN-␥ inhibits EVT invasion via a mechanism dependent on an increase in EVT apoptosis and/or a decrease in protease activity.

MATERIALS AND METHODS Sample collection Placental and decidual samples were obtained from women undergoing elective surgical termination of pregnancy at the Royal Victoria Infirmary, Newcastle on Tyne, UK. The study was approved by the Joint Ethics Committee of Newcastle and North Tyneside Health Authority and the University of Newcastle, and all women gave informed written consent. Placenta was obtained from pregnancies at 8 –10 wk gestational age and decidua was obtained from pregnancies at 8 –10 and 12–14 wk gestational age (as determined by ultrasound measurement of crown rump length or biparietal diameter). After collection, placental or decidual tissue was immediately suspended in sterile saline, transported to the laboratory, and washed two or three times in sterile PBS to remove excess blood.

CD56⫹ uNK cells were plated in a 96-well plate at a concentration of 1 ⫻ 105 cells/well in 100 ␮l RPMI 1640 (containing 1000U/ml penicillin, 1 mg/ml streptomycin, 2 mM l-glutamine, and 10% FBS [all from Sigma Chemical Co., Poole, UK]) and incubated for 24 and 48 h. Conditioned medium was removed and stored at –20°C until required for analysis. In addition, cells were harvested and frozen at – 80°C until required for RNA isolation. Cell smears of freshly isolated total decidual cell suspensions were prepared on 3⬘-aminopropyltriethoxysilane-coated slides for immunostaining with anti-IFN-␥. Using this methodology, the CD56⫹ cell-enriched isolate was consistently ⬎ 95% pure (22). Approximately 25% of the total decidual cell suspension is CD56⫹ cells. FAST Quant® microspot assays for IFN-␥ quantification A FAST Quant® human THI/THII microspot assay was performed on conditioned medium according to the manufacturer’s instructions (Schleicher & Schuell, Dassel, Germany) to determine levels of secreted IFN-␥. Briefly, slides were blocked in 70 ␮l blocking buffer for 30 min with shaking at room temperature; blocking buffer was removed and 70 ␮l samples or standards were added to the appropriate well and incubated overnight. The slides were washed three times, then 70 ␮l biotinylated detection antibody (Ab) was added and incubated for 1 h. After another three washes, 70 ␮l streptavidin-Cy5 solution was added, the slides were incubated for 45 min in the dark, washed three times, and allowed to dry. The slides were imaged using a GenePix scanner and data analysis was performed by the ArrayVision™FAST® software. Real-time reverse transcriptase polymerase chain reaction Total RNA was isolated from decidual and CD56⫹ cells (t⫽0, t⫽24, and t⫽48 h) using TRI-reagent according to the manufacturer’s instructions (TRIZOL, Sigma, Poole, UK). The quantity of RNA was assessed at 260 nm using a 96-well spectrophotometer (Molecular Devices, Workingham, UK) and UV visible 96-well plates (Greiner, Gloucestershire, UK). Total RNA (500 ng) was reverse transcribed using Superscript III according to the manufacturer’s instructions (InVitrogen, Paisley, UK). Real-time RT-polymerase chain reaction (RTPCR) was performed using an ABI7000 (Applied Biosystems, Foster City, CA, USA) PCR machine and probes were labeled with the fluorophore 6-Fam and quencher TAMRA. In each run, a standard curve was produced for each set of primers and probe based on serial dilutions of a standard amount of RNA that was reverse transcribed. For each sample, RT-PCR was performed for GAPDH (GAPDH Assay on Demand, Applied Biosystems) and IFN-␥ (IFN-␥ Assay on Demand, Applied Biosystems). Data are presented as the ratio of the amount of IFN-␥ to GAPDH. Intracellular cytokine flow cytometry

Uterine natural killer cell isolation Total decidual cell isolates and CD56⫹ cell-enriched isolates were prepared by enzymatic disaggregation and positive immunomagnetic selection (MACS) as described previously (21). Briefly, decidual tissue was minced, DNase/collagenase treated, allowed to adhere overnight, and either used as total decidual cell suspensions or subjected to anti-CD56 (Coulter, High Wycombe, UK) reactive magnetic bead separation (MidiMACS, Miltenyi Biotec., Surrey, UK) to obtain CD56⫹ cell suspensions of ⬎95% purity. After isolation, a fraction of the total decidual suspension and CD56⫹ cells was snap frozen for RNA isolation (t⫽0). Total decidual cell suspensions or IFN-␥ AND TROPHOBLAST INVASION

Decidual samples were disaggregated with enzyme as described above, filtered to give a single cell suspension, and cultured overnight. Each sample was divided in two; one part was stimulated with phorbol 12-myristate 13-acetate (PMA, 25 ng, Sigma Chemical Co.) and ionomycin (1 ␮g, Sigma Chemical Co.) for 4 h at 37°C and the other was left unstimulated. Both samples contained brefeldin A (10 ␮g, Sigma Chemical Co.) to inhibit Golgi transport of IFN-␥ (23). The stimulated and unstimulated samples were split in two and stained with anti-CD56-FITC (1:10, BD Biosciences, Cowly, UK), anti-CD45-PerCP (1:10, BD Biosciences), and anti-CD3-APC (1:10, BD Biosciences) for 30 min, fixed in 1% 2513

paraformaldehyde for 30 min, and perforated in 0.05% saponin for 30 min before intracellular staining with either anti-IFN-␥-PE (1:10, BD Biosciences) or control IgG-PE for 30 min. Cells were washed between each step in PBS. Samples were analyzed on a FACScan (BD Biosciences).

(1:30 dilution, 1 h room temperature, Santa Cruz Biotech, Santa Cruz, CA, USA) using an alkaline phosphatase antialkaline phosphatase technique (Dako, Ely, Cambs., UK). The reaction was developed with alkaline phosphatase substrate kit III (Vector Laboratories) to give a blue reaction product.

Invasion assay

Substrate gel zymography

Placental explants were prepared as described previously (8). Briefly, chorionic villous tips were dissected, minced to ⬃0.5 mm (3), and resuspended in culture medium (Dulbecco’s modified Eagle medium:F12 containing 10% FBS [FBS], penicillin/streptomycin, and amphoteracin B [all from Sigma Chemical Co.]) such that 15 ␮l of the suspension constituted ⬃10 mg of tissue. Matrigel invasion assays were performed as described previously (8). To investigate the effects of IFN-␥ on trophoblast invasion, exogenous IFN-␥ (0.1, 1, and 10 ng/ml; R&D Systems, Abingdon, UK) was added to both the Matrigel and the culture medium. Specific anti-IFN-␥-neutralizing Ab (mouse anti-human IFN-␥, 10 ␮g/ml, R&D Systems) was added to inhibit the effect of any endogenous cytokine, with nonimmune mouse IgG as a control (10 ␮g/ml, R&D Systems). To verify the specificity of the neutralizing Ab, a combination of cytokine and neutralizing Ab was added (IFN-␥, 10 ng/ ml⫹mouse anti-human IFN-␥, 10 ␮g/ml). Neutralizing Ab concentrations were chosen based on the manufacturer’s specifications to give a minimum of 50% neutralization. Each experiment was performed in triplicate and repeated on at least five separate occasions. Data are expressed as the mean invasion index (⫾se), where the level of invasion was normalized to the control within each experiment.

Zymographic analysis was performed on conditioned medium collected from placental explant cultures as described previously (8, 25). Briefly, 20 ␮g total protein was resolved in a 12% SDS-PAGE containing either 2 mg/ml gelatin (for MMP2 and MMP9) or 2 mg/ml casein and 0.025 U/ml plasminogen (for uPA) (American Diagnostica Inc., Greenwich, CT, USA). The gels were washed to remove SDS and incubated overnight at 37°C to allow digestion of substrate. Reverse zymography was performed to determine levels of secreted TIMPs. Total protein (50 ␮g) was resolved in a 15% SDS-PAGE containing 2 mg/ml gelatin and a mixture of MMP9 and MMP2 (kind gift from Dr. Dylan Edwards, University of East Anglia, Norwich, UK). After electrophoresis, the gels were incubated overnight on a shaking table, washed, and incubated for a further 16 –18 h at 37°C. The gels were stained with Coomassie Brilliant Blue R250, destained, preserved, and dried. Dried gels were then scanned and densitometry was performed (UnScan-It, Silk Scientific Co., Orem, UT, USA). To remove intersubject variability, all results are normalized to their respective controls.

Immunohistochemistry To determine whether IFN-␥ (10 ng/ml) altered trophoblast cell viability and numbers (apoptosis and proliferation), at the end of the invasion assay explants were fixed in 10% neutral-buffered formalin for 24 h and processed into paraffin wax (n⫽5). Serial 3 ␮m sections were immunostained using an avidin-biotin peroxidase method (Vectastain Elite mouse kit, Vector Laboratories, Peterborough, UK) with M30, Ki67, human leukocyte antigen (HLA) -G, and cytokeratin 7 as described previously (8). The reaction was developed with Fast™ diaminobenzidine (3,3⬘-diaminobenzidine, Sigma Chemical Co.) tablets. Washes between each step were performed in TBS (0.15 M TRIS-buffered 0.05 M saline, pH 7.6). Sections were counterstained in Mayer’s hematoxylin (BDH, Poole, UK) and mounted in DPX synthetic resin (Raymond Lamb, London, UK). Omission of primary Ab or substitution with nonimmune mouse serum for the primary Ab was included as controls. Sections were scored for the approximate percentage of immunopositive cells (1⫽⬍10% immunopositive cells, 2⫽11–24% immunopositive cells, 3⫽25–74% immunopositive cells, 4⫽⬎75% immunopositive cells) by one investigator (G.E.L.) who was blinded to the identity of the sample (8, 24). Data are shown as the mean ⫾ se score for each Ab. To demonstrate that freshly isolated CD56⫹ uNK cells produced IFN-␥ protein, cell smears were immunostained using a double labeling technique for CD56 and IFN-␥. Briefly, slides were fixed in acetone and incubated with anti-CD56 (1:100 dilution, 1 h room temperature, Novocastra Laboratories, Newcastle on Tyne, UK). The anti-CD56 reaction was developed with an avidin-biotin peroxidase method (Vectastain Elite mouse kit) and 3,3⬘-diaminobenzidine to give a brown reaction product. After blocking in rabbit serum, smears were then immunostained with anti-IFN-␥ 2514

Vol. 20

December 2006

Statistical analysis Data are presented as mean with se. Statistical calculations were performed using the StatView statistical software package (Abacus Concepts Inc., Berkley, CA, USA). Statistical significance was determined by use of 1-way ANOVA, followed by Fisher’s post hoc analysis, unless stated otherwise. Student’s t test was used when only two sets of data were compared. All statistical tests were two-sided, and differences were considered statistically significant at P ⬍ 0.05.

RESULTS Expression of IFN-␥ by uNK cells Intracellular flow cytometry was not sufficiently sensitive to detect IFN-␥ in freshly isolated total decidual cell suspensions. However, a significant proportion of CD56⫹ cells were positive for intracellular IFN-␥ after stimulation with PMA and ionomycin for 4 h (data not shown). In freshly isolated total decidual cell smears, a portion of the CD56⫹ cells (brown staining) demonstrated immunostaining for IFN-␥ (blue staining) (Fig. 1A). CD56-positive IFN-␥-negative and CD56 negative IFN-␥ immunopositive cells were also detected (Fig. 1A). mRNA for IFN-␥ was detected in all freshly isolated (t⫽0) total decidual and uNK cell fractions at 8 –10 and 12–14 wk gestational age (n⫽5 both groups). There was no difference in levels of IFN-␥ mRNA detected between total decidual and uNK cell fractions at either gestational age (Fig. 1B) or between the two gestational age groups for each individual cell fraction (Fig. 1B). After 24 or 48 h in culture, mRNA for IFN-␥ was detected in two of five total decidual samples but in no

The FASEB Journal

LASH ET AL.

uNK cell samples from 8 –10 wk gestational age. After 24 or 48 h in culture, mRNA for IFN-␥ was detected in one of five total decidual and in one uNK cell sample from 12–14 wk gestational age. IFN-␥ protein was detected in culture supernatants after 24 and 48 h. There was no difference in protein levels after 48 h compared with 24 h culture, and therefore only data from 24 h culture are shown (Fig. 1C). There was significantly less (P⫽0.04) IFN-␥ protein secreted by the uNK cell fraction from 8 –10 wk gestation compared with the total decidual cell fraction. However, at 12–14 wk gestation there was no

Figure 2. The effect of addition of exogenous IFN-␥ and/or neutralizing Ab on invasion of extravillous trophoblast from isolated villous explants in a Matrigel invasion assay. Data are expressed as invasion index (mean⫾se for each experiment). *Statistical significance compared with control (P⬍0.05).

difference in the levels of protein secreted by total decidual and uNK cell fractions, suggesting that uNK cells are a major source of IFN-␥ in the placental bed at this gestational age. There were no gestational age differences in IFN-␥ levels for either cell fraction. Effect of IFN-␥ on extravillous trophoblast cell invasion Addition of exogenous IFN-␥ inhibited invasion of extravillous trophoblast cells by ⬃35% at the highest concentration used (10 ng/ml, P⬍0.001, n⫽13) (Fig. 2). Although addition of a neutralizing Ab alone had no effect on extravillous trophoblast invasion, addition of the specific neutralizing Ab with IFN-␥ (10 ng/ml) removed its inhibitory effect (Fig. 2). Nonimmune mouse IgG at the same concentration used for antibodies had no effect on invasion compared with untreated controls (Fig. 2). Effect of IFN-␥ on placental explant apoptosis and proliferation

Figure 1. A) Double immunostaining for CD56 (brown) and IFN-␥ (blue). Filled arrows denote double-labeled CD56/ IFN-␥ cells; open arrows denote cells immunostaining for IFN-␥ alone. Single staining brown cells are CD56 cells negative for IFN-␥. B) Real-time RT-PCR data for IFN-␥ mRNA in total decidual and uNK cell fractions from 8 –10 and 12–14 wk gestation at t ⫽ 0, 24, and 48 h. Data are shown as a ratio of IFN-␥:GAPDH (mean⫾se arbitrary units). C) Levels of IFN-␥ protein secreted by total decidual and uNK cell fractions from 8 –10 and 12–14 wk gestation after 24 h culture. Data are shown as mean ⫾ se pg/ml. *P ⫽ 0.04. IFN-␥ AND TROPHOBLAST INVASION

There was no difference in M30 immunostaining of villous trophoblast in explants cultured under any of the experimental conditions (control, 2.15⫾0.9; IFN-␥ 10 ng/ml, 2.62⫾1.19; n⫽5). In contrast, the proportion of M30 immunopositive extravillous trophoblast cells was increased in explants cultured in IFN-␥ (control, 1.77⫾1.09; IFN-␥ 10 ng/ml, 2.46⫾1.20; P⫽0.03; n⫽5). There was no difference in Ki67 immunostaining of villous trophoblast in the presence of IFN-␥ alone (control, 3.0⫾1.08; IFN-␥ 10 ng/ml, 2.9⫾1.32; n⫽5). EVT is a nonproliferative cell type and hence was Ki67 negative. Effect of IFN-␥ on the level of secreted proteases Levels of secreted MMP9 in conditioned medium from explant cultures did not differ in the presence of IFN-␥ compared with controls (Fig. 3). Levels of MMP2 secreted by placental explants were decreased in the presence of IFN-␥ (P⫽0.04, Fig. 3). TIMP-1, -2, and -3 2515

Figure 3. Upper left: the effect of IFN-␥ on levels of secreted MMP2 and MMP9 in villous explant culture supernatants. Data are expressed as mean ⫾ se relative light units compared with control (no added cytokine). *Statistical significance compared with control (P⬍0.05). Upper right: the effect of IFN-␥ on levels of uPA secreted by villous explant cultures. Data are expressed as mean ⫾ se relative light units compared with control (no added cytokine).

were secreted by placental explants in all experimental conditions tested but levels of secreted TIMPs were not altered by any of the experimental conditions tested (data not shown). There was no difference in levels of secreted uPA in the presence of IFN-␥ compared with controls (Fig. 3).

DISCUSSION In the current study we have demonstrated that uNK cells are a source of IFN-␥ in early human pregnancy decidua. Furthermore, we have shown that high concentrations of IFN-␥ can inhibit trophoblast invasion due, at least in part, to increased EVT apoptosis and decreased secreted levels of MMP2. IFN-␥-neutralizing antibodies had no effect on the basal level of EVT invasion, suggesting there was no autocrine effect of IFN-␥ on EVT invasion. uNK cells have been shown to be a major source of IFN-␥ in the mouse (14). However, evidence for production of IFN-␥ in unstimulated human uNK cells is limited. von Rango et al. (17) demonstrated that freshly isolated decidual leukocytes were the major source of IFN-␥ mRNA in the decidua compared with decidual epithelial cells and fibroblasts, and also demonstrated higher levels of IFN-␥ mRNA and protein levels in the decidua basalis than decidua parietalis (17). uNK cell production of IFN-␥ can be stimulated by contact with HLA-G-expressing cells (26). Therefore, association of uNK cells with HLA-G-positive EVT may stimulate uNK cells to secrete IFN-␥ in a localized manner. In the present study we were able to demonstrate that uNK cells were a major source of IFN-␥ mRNA in the decidua of freshly isolated cells. However, expression of IFN-␥ mRNA from both cell populations was lost after 24 h in culture, although we did measure large amounts of IFN-␥ protein in both the total decidual and uNK cell culture supernatants after 24 and 48 h of 2516

Vol. 20

December 2006

culture. This suggests that uNK and total decidual cells rapidly lose their capacity to express mRNA for IFN-␥ in culture. There was no difference in IFN-␥ protein levels after 24 or 48 h in culture, indicating that the cultured cells were no longer producing this protein. IFN-␥ is rapidly secreted by cells, and levels of this protein detected likely reflect expression of IFN-␥ soon after cell isolation. At 8 –10 wk gestation, there was less IFN-␥ protein secreted by the uNK cells than the total decidual cell suspension. This was no longer the case at 12–14 wk gestation, suggesting that uNK cells are a major source of IFN-␥ in the decidua by the end of the first trimester of pregnancy. IFN-␥ levels observed in the current study are higher than those previously reported (15, 16, 27, 28). These differences may arise from differences in methodology, isolation method, amount of cell manipulation, and length of time in culture before measurement. As demonstrated in the current study, mRNA for IFN-␥ is lost by 24 h in culture and therefore uNK cells in culture lose the ability to produce and secrete IFN-␥. Li et al. (29) also demonstrated loss of mRNA for angiopoietin 2 in uNK cells after 24 h culture. In addition, IFN-␥ is quickly secreted from cells after it is produced, so if cell culture media was changed after isolation and prior to collection, all secreted IFN-␥ would be lost. This is the first report on the role of IFN-␥ in regulating a primary source of EVT. However, the finding that IFN-␥ inhibits trophoblast invasion is in accordance with a study by Karmakar et al. (20), who demonstrated that IFN-␥ inhibited invasion by the JEG3 choriocarcinoma cell line. In addition, Ain et al. (12) reported an earlier EVT invasion into mesometrial decidua in IFN-␥ signaling-deficient mice, and exogenous IFN-␥ inhibits trophoblast outgrowth from both rat (12) and mouse (11) blastocysts. IFN-␥ also inhibits the invasiveness of several different human cancer cell lines such as U251-MG and CRT astrogliomas (30), renal cell carcinoma KG2 (31, 32), and melanoma

The FASEB Journal

LASH ET AL.

A2058 (33). The IFN-␥ receptor is made up of dimers of the two IFN-␥ receptors, IFN-␥R1 and IFN-␥R2 (18). Expression of IFN-␥ receptors is ubiquitous throughout the body (18), including trophoblast (34). IFN-␥R1 is involved in ligand binding whereas IFN-␥R2 is required for signal transduction, which is primarily through the janus-activated kinase (JAK):STAT pathway involving JAK1, JAK2, and STAT1 (18). STAT1 is translocated to the nucleus and acts on gamma activation sequences to influence transcription in response to IFN-␥ (18). IFN␥-regulated genes include MMP-1, -9, -13, stromelysin, and type II collagen (35). IFN-␥ has been shown to induce apoptosis of primary isolates of cytotrophoblast cells, although the effect of IFN-␥ on apoptosis in placental explants has not been studied before (36). In the current study we demonstrated an increased percentage of M30-positive EVT (as defined by HLA-G expression) in response to IFN-␥. This suggests that IFN-␥ induces EVT apoptosis; since fewer cells would then be available to invade, this represents a potential mechanism for inhibition of EVT invasion by IFN-␥. No effect of IFN-␥ on the level of apoptosis in villous cytotrophoblast was observed. Addition of IFN-␥ also did not affect proliferation of villous cytotrophoblast cells. In a previous study, Knupfer et al. (37) demonstrated that IFN-␥ induced apoptosis of the glioblastoma cell line, A172, while inhibiting proliferation. MMP2 and MMP9 have been suggested to be the major proteases used by EVT during the invasive process (38). IFN-␥ inhibits the invasiveness of several different cancer cell lines, including astrogliomas U251-MG and CRT (30), renal cell carcinoma KG2 (31, 32), and melanoma A2058 (33), via a decrease in MMP2 activity. Furthermore, Karmakar et al. (20) have demonstrated that IFN-␥ inhibition of JEG3 invasion was associated with a decrease in MMP2, MMP9, and uPA activity. In the current study we have demonstrated a decrease in MMP2 levels secreted by placental explants in the presence of IFN-␥, although there was no difference in levels of secreted MMP9, TIMPs, or uPA. This difference may reflect the use of primary first trimester EVT as opposed to a choriocarcinoma cell line: JEG3 cells respond differently from primary sources of EVT to hypoxia and other external stimuli such as TGF-␤1 (39, 40). uNK cell-derived IFN-␥ has been shown to play a pivotal role in spiral artery remodeling in mouse pregnancy (13, 14). The role of uNK cells and IFN-␥ in human spiral artery remodeling, however, remains unclear. Whereas spiral artery remodeling is independent of trophoblast in the mouse, in the human placental bed EVT interactions with spiral arteries are thought to be essential for successful spiral artery remodeling. Craven et al. (41), however, demonstrated intimal separation and vessel dilation in the absence of EVT in 80% of spiral arteries examined, raising the possibility that a trophoblast-independent mechanism initiates human spiral artery remodeling. These observations were later confirmed by Kam et al. (42). Indeed, IFN-␥ AND TROPHOBLAST INVASION

in the human placental bed, spiral arteries are surrounded by uNK cells in the absence of EVT (10) and produce a range of angiogenic growth factors implicated in vascular smooth muscle cell remodeling (43). A role for decidual IFN-␥ in spiral artery remodeling, however, does not exclude an additional role in regulating trophoblast invasion. Indeed, EVT are a naturally highly invasive cell type that would continue to invade if left unchecked. Therefore, in normal pregnancy a fine balance of cytokines and growth factors present in the decidua and myometrium is required to allow EVT to reach the inner third of the myometrium, but no further. An imbalance in decidual and myometrial cytokines could lead to placenta accreta (6), which is associated with excess invasion of EVT and preeclampsia (4), fetal growth restriction (5), and second trimester miscarriage (3), all of which are associated with deficient trophoblast invasion. Indeed, elevated serum and decidual levels of IFN-␥ have been observed in women with preeclampsia, although the source of this IFN-␥ is not clear (44, 45). To summarize, in the present study we have demonstrated that uNK cells are a source of IFN-␥ in first and early second trimester decidua in normal human pregnancy. In addition, we have demonstrated that IFN-␥ inhibits EVT invasion via both an increase in EVT apoptosis and a decrease in secreted MMP2 levels. The authors wish to acknowledge the staff at the Royal Victoria Infirmary, Newcastle upon Tyne, for their assistance in sample collection. This project was supported by funding from BBSRC (S19967).

REFERENCES 1. 2. 3.

4.

5.

6.

7.

8.

Pijnenborg, R., Dixon, G., Robertson, W. B., and Brosens, I. (1980) Trophoblastic invasion of human decidua from 8 to 18 weeks of pregnancy. Placenta 1, 3–19 Brosens, I., Robertson, W. B., and Dixon, H. G. (1967) The physiological response of the vessels of the placental bed to normal pregnancy. J. Pathol. Bacteriol. 93, 569 –579 Ball, E., Bulmer, J., Ayis, S., Lyall, F., and Robson, S. (2006) Late sporadic miscarriage is associated with abnormalities in spiral artery transformation and trophoblast invasion. J. Pathol. 208, 535–542 Pijnenborg, R., Anthony, J., Davey, D. A., Rees, A., Tiltman, A., Vercruysse, L., and van Assche, A. (1991) Placental bed spiral arteries in the hypertensive disorders of pregnancy. Br. J. Obstet. Gynaecol. 98, 648 – 655 Khong, T. Y., De Wolf, F., Robertson, W. B., and Brosens, I. (1986) Inadequate maternal vascular response to placentation in pregnancies complicated by pre-eclampsia and by small for gestational age infants. Br. J. Obstet. Gynaecol. 93, 1049 –1059 Goldman-Wohl, D. S., Ariel, I., Greenfield, C., Hanoch, J., and Yagel, S. (2000) HLA-G expression in extravillous trophoblasts is an intrinsic property of cell differentiation: a lesson learned from ectopic pregnancies. Mol. Hum. Reprod. 6, 535–540 Lash, G. E., Cartwright, J. E., Whitley, G. St. J., Trew, A. J., and Baker, P. N. (1999) The effects of angiogenic growth factors on extravillous trophoblast invasion and motility. Placenta 20, 661– 667 Lash, G. E., Otun, H. A., Innes, B. A., Bulmer, J. N., Searle, R. F., and Robson, S. C. (2005) Inhibition of trophoblast cell invasion by TGFB1, 2, and 3 is associated with a decrease in active proteases. Biol. Reprod. 73, 374 –381

2517

9.

10. 11. 12.

13.

14.

15. 16.

17.

18. 19.

20.

21.

22. 23.

24. 25.

26.

27.

28.

2518

Otun, H. A., Lash, G. E., Bulmer, J. N., Searle, R. F., and Robson, S. C. (2003) Inhibition of first trimester human extravillous trophoblast invasion by TNF-␣ and IFN-␥. Placenta 24, A67 (abstr.) Bulmer, J. N., and Lash, G. E. (2005) Human uterine natural killer cells: a reappraisal. Mol. Immunol. 42, 511–521 Haimovici, F., Hill, J. A., and Anderson, D. J. (1991) The effects of soluble products of activated lymphocytes and macrophages on blastocyst implantation events in vitro. Biol. Reprod. 44, 69 –75 Ain, R., Canham, L. N., and Soares, M. J. (2003) Gestation stage-dependent intrauterine trophoblast cell invasion in the rat and mouse: novel endocrine phenotype and regulation. Dev. Biol. 260, 176 –190 Guimond, M. J., Wang, B., and Croy, B. A. (1998) Engraftment of bone marrow from severe combined immunodeficient (SCID) mice reverses the reproductive deficits in natural killer cell-deficient tg epsilon 26 mice. J. Exp. Med. 187, 217–223 Ashkar, A. A, Di Santo, J. P., and Croy, B. A. (2000) Interferon ␥ contributes to initiation of uterine vascular modification, decidual integrity, and uterine natural killer cell maturation during normal murine pregnancy. J. Exp. Med. 192, 259 –270 Jokhi, P. P., King, A., and Loke, Y. W. (1997) Cytokine production and cytokine receptor expression by cells of the human first trimester placental-uterine interface. Cytokine 9, 126 –137 Marzusch, K., Buchholz, F., Ruck, P., Handgretinger, R., Geiselhart, A., Engelmann, L., and Dietl, J. (1997) Interleukin-12- and interleukin-2-stimulated release of interferon-gamma by uterine CD56⫹⫹ large granular lymphocytes is amplified by decidual macrophages. Hum. Reprod. 12, 921–924 Von Rango, U., Classen-Linke, I., Raven, G., Bocken, F., and Beier, H. M. (2003) Cytokine microenvironments in human first trimester decidua are dependent on trophoblast cells. Fertil. Steril. 79, 1176 –1186 Pestka, S., Krause, C. D., and Walter, M. R. (2004) Interferons, interferon-like cytokines, and their receptors. Immunol. Rev. 202, 8 –32 Van der Loos, C. M., Houtkamp, M. A., de Boer, O. J., van der Wal, A. C., and Becker, A. E. (2001) Immunohistochemical detection of interferon-␥: fake or fact? J. Histochem. Cytochem. 49, 699 –710 Karmakar, S., Dhar, R., and Das, C. (2004) Inhibition of cytotrophoblastic (JEG-3) cell invasion by interleukin 12 involves an interferon gamma-mediated pathway. J. Biol. Chem. 279, 55297–55307 Vassiliadou, N., and Bulmer, J. N. (1998) Functional studies of human decidua in spontaneous early pregnancy loss: effect of soluble factors and purified CD56⫹ lymphocytes on killing of natural killer- and lymphokine-activated killer-sensitive targets. Biol. Reprod. 58, 982–987 Jones, R. K., Bulmer, J. N., and Searle, R. F. (1997) Cytotoxic activity of endometrial granulated lymphocytes during the menstrual cycle in humans. Biol. Reprod. 57, 1217–1222 Picker, L. J., Singh, M. K., Zdraveski, Z., Treer, J. R., Waldrop, S. L., Bergstresser, P. R., and Maino, V. C. (1995) Direct demonstration of cytokine synthesis heterogeneity among human memory/effector T cells by flow cytometry. Blood 86, 1408 –1419 Adams, E. J., Green, J. A., Clark, A. H., and Youngson, J. H. (1999) Comparison of different scoring systems for immunohistochemical staining. J. Clin. Pathol. 52, 75–77 Graham, C. H., and McCrae, K. R. (1996) Altered expression of gelatinase and surface-associated plasminogen activator activity by trophoblast cells isolated from placentae of preeclamptic patients. Am. J. Obstet. Gynecol. 175, 555–562 Van der Meer, A., Lukassen, H. G., van Lierop, M. J., Wijnands, F., Mosselman, S., Braat, D. D., and Joosten, I. (2004) Membrane-bound HLA-G activates proliferation and interferongamma production by uterine natural killer cells. Mol. Hum. Reprod. 10, 189 –195 Vigano, P., Gaffuri, B., Somigliana, E., Infantino, M., Vignali, M., and Di Blasio, A. M. (2001) Interleukin-10 is produced by human uterine natural killer cells but does not affect their production of interferon-gamma. Mol. Hum. Reprod. 7, 971–977 Kitaya, K., Yasuda, J., Nakayama, T., Fushiki, S., and Honjo, H. (2003) Effect of female sex steroids on human endometrial

Vol. 20

December 2006

29.

30.

31.

32.

33. 34.

35.

36.

37.

38. 39.

40.

41. 42. 43.

44.

45.

CD16neg CD56bright natural killer cells. Fertil. Steril. 79, 730 – 734 Li, X. F., Charnock-Jones, D. S., Zhang, E., Hiby, S., Malik, S., Day, K., Licence, D., Bowen, J. M., Gardner, L., King, A., et al. (2001) Angiogenic growth factor messenger ribonucleic acids in uterine natural killer cells. J. Clin. Endocrinol. Metab. 86, 1823– 1834 Qin, H., Moellinger, J. D., Wells, A., Windsor, L. J., Sun, Y., and Benveniste, E. N. (1998) Transcriptional suppression of matrix metalloproteinase-2 gene expression in human astroglioma cells by TNF-alpha and IFN-gamma. J. Immunol. 161, 6664 – 6673 Gohji, K., Fidler, I. J., Tsan, R., Radinsky, R., von Eschenbach, A.C., Tsuruo, T., and Nakajima, M. (1994) Human recombinant interferons-beta and -gamma decrease gelatinase production and invasion by human KG-2 renal-carcinoma cells. Int. J. Cancer 58, 380 –384 Gohji, K., Nomi, M., Hara, I., Arakawa, S., and Kamidono, S. (1998) Influence of cytokines and growth factors on matrix metalloproteinase-2 production and invasion of human renal cancer. Urol. Res. 26, 33–37 Hujanen, E. S., and Turpeenniemi-Hujanen, T. (1991) Recombinant interferon alpha and gamma modulate the invasive potential of human melanoma in vitro. Int. J. Cancer 47, 576 –581 Paulesu, L., Romagnoli, R., Cintorino, M., Ricci, M. G., and Garotta, G. (1994) First trimester human trophoblast expresses both interferon-gamma and interferon-gamma-receptor. J. Reprod. Immunol. 27, 37– 48 Ma, Z., Qin, H., and Benveniste, E. N. (2001) Transcriptional suppression of matrix metalloproteinase-9 gene expression by IFN-gamma and IFN-beta: critical role of STAT-1alpha. J. Immunol. 167, 5150 –5159 Pijnenborg, R., Luyten, C., Vercruysse, L., Keith, J. C., Jr., and Van Assche, F. A. (2000) Cytotoxic effects of tumour necrosis factor (TNF)-alpha and interferon-gamma on cultured human trophoblast are modulated by fibronectin. Mol. Hum. Reprod. 6, 635– 641 Knupfer, M. M., Knupfer, H., Jendrossek, V., Van Gool, S., Wolff, J. E., and Keller, E. (2001) Interferon-gamma inhibits growth and migration of A172 human glioblastoma cells. Anticancer Res. 21, 3989 –3994 Cohen, M., Meisser, A., and Bischof, P. (2005) Metalloproteinases and human placental invasiveness. Placenta 27, 783–793 Graham, C. H., Connelly, I., MacDougall, J. R., Kerbel, R. S., Stetler-Stevenson, W. G., and Lala, P. K. (1994) Resistance of malignant trophoblast cells to both the anti-proliferative and anti-invasive effects of transforming growth factor-beta. Exp. Cell Res. 214, 93–99 Lash, G. E., Hornbuckle, J., Brunt, A., Kirkley, M., Searle, R. F., Robson, S. C., and Bulmer, J. N. (2006) Effect of low oxygen concentrations on trophoblast-like cell line invasion. Placenta In press Craven, C. M., Morgan, T., and Ward, K. (1998) Decidual spiral artery remodelling begins before cellular interaction with cytotrophoblasts. Placenta 19, 241–252 Kam, E. P., Gardner, L., Loke, Y. W., and King, A. (1999) The role of trophoblast in the physiological change in decidual spiral arteries. Hum. Reprod. 14, 2131–2138 Lash, G. E., Schiessl, B., Kirkley, M., Innes, B. A., Cooper, A., Searle, R. F., Robson, S. C., and Bulmer, J. N. (2006) Expression of angiogenic growth factors by uterine natural killer cells during early pregnancy. J. Leuk. Biol. In press Saito, S., Sakai, M., Sasaki, Y., Tanebe, K., Tsuda, H., and Michimata, T. (1999) Quantitative analysis of peripheral blood Th0, Th1, Th2 and the Th1:Th2 cell ratio during normal human pregnancy and preeclampsia. Clin. Exp. Immunol. 117, 550 –555 Wilczynski, J. R., Tchorzewski, H., Glowacka, E., Banasik, M., Lewkowicz, P., Szpakowski, M., Zeman, K, and Wilczynski, J. (2002) Cytokine secretion by decidual lymphocytes in transient hypertension of pregnancy and pre-eclampsia. Mediators Inflamm. 11, 105–111

The FASEB Journal

Received for publication June 1, 2006. Accepted for publication July 17, 2006.

LASH ET AL.