MIF IS AMONG THE PROINFLAMMATORY CYTOKINES INCREASED BY LPS IN THE HUMAN TROPHOBLAST LINE

Infection is increasingly considered to contribute to pathological conditions in pregnancy. The placenta acts as a protective immunological fetomaternal barrier which recognizes microbes by pattern recognition receptors on the trophoblast. Lipopolysaccharide (LPS) is a cell wall constituent of Gram-negative bacteria that elicits a strong immune response. In this study, LPS from E. coli was used to treat the HTR-8/SVneo trophoblast cell line and examine its influence on cytokines IL-6, IL-8 and MIF using real-time PCR, metalloproteinases (MMP)-2 and -9 by gelatin zymography, and Western analysis of integrin subunits α1 and β1, all known to contribute to migration of human trophoblasts in vitro. The results described herein for the first time, show that MIF mRNA and secreted MIF protein were significantly elevated (2.5-3and 2-fold, respectively) in LPS-treated cells. MMP-2 and MMP-9 levels were increased, as well as cell migration, as judged by a wound-healing test, however, no changes in the studied integrin subunits, cell viability or cell numbers were observed. The data obtained furthers our understanding of LPS actions on the trophoblast in vitro, additionally implicate MIF, and suggest that infection in vivo could indeed alter the functional characteristics of the trophoblast.


INTRODUCTION
Trophoblast invasion into the uterine stroma and spiral arteries is a crucial step for implantation and the establishment of normal pregnancy [1].During differentiation along the invasive pathway, extravillous trophoblast cells undergo a change in phenotype [2].Trophoblast cells invading from the anchoring chorionic villi of the placenta into the uterus lose the adhesion molecules specific for the epithelia and acquire others, such as integrin α 5 β 1 and α 1 β 1 [3].The process of trophoblast invasion is achieved, among other mechanisms, through degradation of the extracellular matrix by proteolytic enzymes, including matrix metalloproteinases (MMP) -2 and -9 [4,5].Various autocrine and paracrine factors, including cytokines IL-6, IL-8 and MIF, have been suggested to modulate trophoblast function and control implantation [6][7][8][9][10].
With the establishment of pregnancy there is a marked influx into the uterine endometrium of immune cells that participate in the establishment of pregnancy and immunotolerance of the fetus.In addition to endocrine and transport functions, the placenta acts as a protective maternal-fetal barrier where the trophoblast recognizes microbes through pattern recognition receptors which participate in the innate immune response against infections [11,12].Infection is considered to contribute to pathological conditions in pregnancy, including abnormal placentation in preeclampsia.The mechanisms linking infection to placental dysfunction are not sufficiently understood.Bacterial lipopolysaccharide (LPS) is a cell wall constituent of most Gram-negative bacteria that elicits strong immune responses in animals.LPS typically consists of a hydrophobic domain known as lipid A (or endotoxin), a core oligosaccharide and a distal polysaccharide (or O-antigen) [13].LPS is classically recognized by immune cells via the pattern recognition receptor Toll-like receptor 4 (TLR4).In macrophages, TLR4 activation triggers the biosynthesis of different mediators of inflammation, such as TNF-α and IL1-β [14,15].In trophoblast cells, a strong expression of TLR 4 has been reported in villous, intermediate and extravillous trophoblasts [16,17].Both isolated first trimester trophoblast cells and HTR-8 cell line express TLR-2 and TLR-4 [18].LPS stimulation via TLR-4 led to the activation of extracellular signal-regulated kinase (ERK), c-Jun NH 2 -terminal kinase (JNK) and mitogenactivated protein kinase (MAPK) [19].Increased levels of TLR-4 were observed in placental bed extravillous trophoblasts of patients with preeclampsia [20], as well as cytokines such as TNF-α, IL-6 and monocyte chemoattractant protein-1, increased both systemically and locally in the placenta [21,22].LPS was shown to increase IL-6 and IL-8 by isolated extravillous trophoblasts through the activation of MAPK signaling [23].Differential inflammatory responses in trophoblast cell models, JEG-3 and BeWo human choriocarcinoma cell lines to LPS were demonstrated [24].In the immortalized normal trophoblast cell line HTR-8/SVneo, LPS from Porphyromonas gingivalis was shown to increase IL-1β and IL-8 [25].
This study was designed to investigate the effect of E. coli-derived LPS on proinflammatory cytokine MIF production in the extravillous trophoblast cell line, HTR-8/SVneo, which could influence functional characteristics, such as viability, cell migration and the expression of markers associated with trophoblast invasive properties.

Cell culture
The HTR-8/SVneo trophoblast cell line was kindly provided by Dr. Charles Graham (Queen's University, Kingston, Ontario, Canada).This cell line was established from human first trimester placenta explant cultures immortalized by SV40 large T antigen [26,27].Cells were cultured at 37°C, 5% CO 2 , in RPMI 1640 supplemented with 5% FBS with an antibiotic/ antimycotic solution (all from Lonza Group Ltd, Basel, Switzerland).

Determination of viable and adherent cell numbers
The viability and adherent cell numbers of HTR-8/ SVneo cells were assessed using the MTT test [28] or crystal violet staining, as described previously [29].HTR-8/SVneo cells were seeded in 96-well plates in 100 µl of medium (2 x 10 4 /well) and allowed to adhere overnight.Cells were rinsed with PBS and incubated for 6 or 24 h with LPS (Escherichia coli 055:B5, Sigma-Aldrich Co., Saint Louis, MO, USA) at 0.1, 1 or 10 µg/ ml in complete or serum free RPMI.For the MTT test, 100 µl of MTT (1 mg/ml, Sigma) in 10% FBS/ PBS was added to each well upon treatment.After incubation for 2 h at 37°C, the medium was replaced by 1-propanol (100 µl/well) and the plates were shaken to ensure complete solubilization of the blue formazan.Absorbance was measured at 540 nm using a microplate reader (LKB, Vienna, Austria).For determination of adherent cell number, the cells were dried and fixed with ice-cold acetone-methanol for 5 min.Then, 50 µl/well of 0.05% crystal violet in 25% methanol was added.After 5 min incubation, the excess of dye was removed by immersing the plates in water and drying at room temperature.The incorporated dye was dissolved in 0.1 M sodium citrate in 50% ethanol at 100 µl/well.Optical density was read at 540 nm.The results are presented as a percentage of control values obtained for untreated cultures.The experiments were carried out three times, n=6.

Cell migration test
The effect of LPS on cell migration was investigated using a cell-wounding test.Cells at 2.5 x 10 5 per well were plated in 24-well plates in 300 µl of complete medium and incubated at 37°C in 5% CO 2 until confluence.Cells were then scraped off with a sterile pipette tip, rinsed with PBS, and cultured in medium with or without LPS at 0.1, 1 or 10 µg/ml for 24 h.Preselected fields were photographed at zero point and after 24 h.The distances crossed by the cells were measured electronically.Experimental data were expressed as a percentage of control values.The experiments were carried out three times in duplicate.

Gelatin zymography
HTR-8/SVneo cell matrix metalloproteinase gelatinolytic activity was studied using SDS-PAGE gelatin zymography.Cells were incubated in complete medium in 24-well plates until confluence, rinsed with PBS and treated with LPS at 0.1, 1 or 10 µg/ml in serum-free medium for 24 h.Media were collected and centrifuged at 300xg.Protein concentrations were determined using BCA kit (Pierce).Gelatinase activities in conditioned media were determined as described previously [30].Samples were separated using 11% SDS-polyacrylamide gels containing 1 mg/ml of gelatin under non-reducing conditions with protein loading of 20 µg per lane.Following electrophoresis, gels were washed twice for 15 min in 2.5% Triton X-100 (v/v) to remove SDS, and incubated overnight in reaction buffer (50 mmol/l Tris-HCl, pH 7, containing 5 mmol/l CaCl 2 ) at 37°C.Gels were stained with Coomassie brilliant blue G-250 for 30 min at RT and destained in 30% methanol and 10% glacial acetic acid (v/v).Proteinase activity was observed as a clear band of digested gelatin at the designated molecular mass.Gelatinase levels were semi-quantitated by densitometric analysis using the ImageMasterTotalLabv 2.01 program (Amersham Biosciences).Experiments were carried out four times in duplicate.

Statistical analysis
The data were analyzed by Statistical Software Program, version 5.0 (Primer of Biostatistic, McGraw-Hill Companies, Inc., New York, NY, USA) using the non-parametric Mann-Whitney rank sum test and Student's t-test as appropriate.Values were considered significantly different when p<0.05.

Effect of LPS on HTR-8/SVneo cell migration and expression of MMPs and integrin subunits
Since each of the increased cytokines stimulates trophoblast migration, a wound-healing assay was used to determine whether E. coli LPS influenced HTR-8/ SVneo cell migration.A moderate increase was observed to 123% of control for 0.1 µg/ml, to 129% of control for 1 µg/ml and to 118% of control for 10 µg/ ml (Fig. 2).

Influence of LPS on HTR-8/SVneo cell proliferation
The possibility that LPS may impact HTR-8/SVneo cell viability and adherent cell numbers was tested by MTT and crystal violet staining.None of the concentrations (0.1, 1 and 10 µg/ml) influenced HTR-8/ SVneo adherent cell number after 6 h (Fig. 4A) or  24 h of culture (Fig. 4C) in serum-free, or complete medium after 24 h incubation (Fig. 4E).No change in cell viability by MTT test was observed under the conditions studied (Fig. 4B, D, F).Therefore, none of the observed changes were induced by altered cell numbers or viability.

DISCUSSION
In human pregnancy, infection and inflammation at term can cause premature rupture of the membranes and delivery [32].In early pregnancy, infection has been particularly studied with regard to the involvement of diverse immune cells of the placental bed; however, there is still insufficient knowledge of its effect on first trimester of pregnancy trophoblast function.
LPS impact on trophoblast cell survival has been previously reported as proapoptotic [33,34] or absent [18,35].In this study, neither HTR-8/SVneo cell viability nor adherent cell numbers were changed with LPS from E. coli, which confirms previous results on HTR-8 cells [18] and other cell types, such as Wharton's jelly-derived mesenchymal stem cells and hepatocytes [36,37].
Various proinflammatory cytokines have been reported to influence trophoblast cell migration and invasion, some of which having an inhibitory ef-  fect, such as TNF α [38] and IFN γ [39].Stimulatory cytokines include IL-1β [40], IL-6 [6], IL-8 [41,7] and recently reported MIF [10].LPS was shown to increase several inflammatory cytokines in the trophoblast, including IL-1β [42], IL-6 [33,23], IL-8 [43] and RANTES [44].In this study, the stimulatory effect of LPS on IL-6 and IL-8 expression by HTR-8/SVneo cells was confirmed.The data obtained introduce the important regulator of innate immunity, MIF, to the list of LPS-induced cytokines in the human trophoblast.Systemic MIF increase has been implicated in infection and severe sepsis [45].Specifically, MIF was identified as a major secreted protein released by anterior pituitary cells in response to LPS stimulation [46], contributing to circulating MIF present in the postacute phase of endotoxemia.The cytokines increased with LPS, IL-6, IL-8 and MIF, could be assumed to act in an autocrine fashion to stimulate HTR-8/SVneo migration in vitro.Indeed, an increase in HTR-8/ SVneo cell migration was observed in this study in cultures treated with LPS.In previous studies, each of these cytokines was shown to increase cell migration when added individually to cell culture media [6,7,10].Interestingly, two commonly used trophoblast cell models, choriocarcinoma cell lines JEG-3 and BeWo, were shown to respond differently to LPS [24].In JEG-3 cells, a dose-and time-dependent stimulation of IL-6 and TNF-α and increased NF-κB gene expression were observed, which was not replicated in BeWo cells.
Trophoblast migration and invasion are prerequisites for the establishment of a healthy pregnancy.Our finding that E. coli LPS induced a small but significant stimulatory effect on HTR-8/SVneo cell migration differs from the previous study [23], where extravillous trophoblast invasion was inhibited.In addition to the intrinsic discrepancy between these two test models, the obtained result could be due to different LPS exposure times (24 h here as opposed to 72 h in the previous study).LPS from Porphyromonas gingivalis had no effect on HTR-8/SVneo migration, but when coadministered with nicotine cell migration was inhibited [47].Nevertheless, LPS stimulated the migration and/or invasion of other cell types such as dendritic cells [48], colorectal cancer cells [49] and vascular smooth muscle cells [50], while it has been reported to have a dose-dependent differential effect on human dental pulp stem cells [51].
Trophoblast cell migration and invasion are known to depend on integrin levels and MMP secretion and activity [52].MMP-2 and -9, as significant mediators of trophoblast invasion, were stimulated here by LPS.This is in accordance with previous studies on the trophoblast where MMP-2 [33] and MMP-9 [53] were also stimulated.LPS had a similar effect in other cell types as well, increasing these MMPs in dental pulp [54], osteosarcoma [55] and colorectal cancer cells [49].No information regarding the modulation of integrins by LPS in trophoblast cells is available.The integrin subunits most relevant for trophoblast migration/invasion, α 1 and β 1 , were investigated here and found to be unaffected by LPS from E. coli.
Based on our results and previous reports, LPS could have a differential effect on trophoblast migration/invasion in vitro, depending on the concentration, type of LPS and the exposure time.In vivo, the situation is far more complex, since LPS affects multiple immune and non-immune cell types present in the placental bed.For instance, LPS activated macrophages and their products could modulate direct LPS effects on the trophoblast, as observed in vitro [56].On the other hand, trophoblast cells were shown to secrete chemokines able to recruit maternal macrophages and alter their cytokine profile [57], and also to modulate activation profile and migration of monocytes in vitro, which was restricted in the presence of LPS [58].In addition to extravillous trophoblasts, syncytiotrophoblasts were also shown to respond to LPS by a large increase in IL-8 [59].Thus, multiple signaling circuits are likely affected in infection with Gram-negative bacteria, and our finding that MIF (a potent proinflammatory cytokine) is also stimulated in the trophoblast by LPS adds yet another element to this complexity.

Fig. 1 .
Fig.1.LPS impact on proinflammatory cytokines in HTR-8/SVneo cells.The effect of LPS on the expression of IL-6 (A), IL-8 (B) and MIF (C) as assessed by qPCR.The data are expressed as fold of change±SEM from the untreated control.D. Western blot analysis of the effect of LPS on secreted MIF in HTR-8/ SVneo cell culture media.Results of densitometric analysis of the corresponding bands are shown in the chart.The data are expressed as percent of the untreated control±SEM.* p<0.05; ** p<0.01; *** p<0.001.

Fig. 2 .
Fig. 2. The effect of LPS on HTR-8/SVneo migration.The distances crossed by cells over a period of 24 h were determined electronically and are given in the chart (A) as percentages of the untreated control±SEM (A).* p<0.05; *** p<0.001.The representative experiment is shown in B.

Fig. 4 .
Fig. 4. The effect of LPS on HTR-8/SVneo cell viability and adherent cell number.The HTR-8/SVneo adherent cell number was assessed by crystal violet staining (A, C, D) and cell viability by MTT test (B, D, F) in serum free medium (A, B, C, D) or complete medium (E, F).The data are expressed as percent of the untreated control±SEM.

Fig. 3 .
Fig. 3.The effect of LPS on the markers of trophoblast invasive phenotype in HTR-8/SVneo cells.MMP-2 and -9 (A, B), and integrin subunits α 1 and β 1 (C) in LPS-treated HTR-8/SVneo cells, using culture media gelatin zymography and cell lysate Western blot analysis respectively.Results of densitometric analysis of the corresponding bands for MMP-9 (A) and MMP-2 (B) and the integrins α 1 and β 1 (C) are shown in the charts.The data are expressed as percent of the untreated control±SEM (D).** p<0.01; *** p<0.001.