Endocrinology and Metabolism

17β-Estradiol modulates the macrophage migration inhibitory factor secretory pathway by regulating ABCA1 expression in human first-trimester placenta

Francesca Ietta, Nicoletta Bechi, Roberta Romagnoli, Jayonta Bhattacharjee, Massimo Realacci, Maura Di Vito, Cristina Ferretti, Luana Paulesu

Abstract

Successful pregnancy involves a series of events, most of them mediated by hormones and cytokines. Estrogens, besides being important for placental growth and embryo development, have a marked effect on the immune system exerting either pro- or anti-inflammatory properties. Numerous studies suggest that estrogens directly affect cellular function, including cytokine production. Macrophage migration inhibitory factor (MIF) is a proinflammatory cytokine involved in pregnancy, particularly during the earlier stages of placentation. Since reports on mice have shown that estrogens modulate MIF, herein we investigated the effect of estrogens on human placental MIF. By using an in vitro model of first-trimester chorionic villous explants, we found that 17β-estradiol (E2) was able to modulate the release of MIF in a dose-dependent manner (10−12 vs. 10−9 M, P < 0.05; 10−9 vs. 10−5 M, P < 0.05; 10−12 vs. 10−5 M, P < 0.001). Unlike MIF release, no significant change in tissue MIF protein or MIF mRNA was observed. We showed evidence that E2 concentrations (10−9 and 10−5 M) act on placental tissue downregulating the mRNA and protein expression of the ATP-binding cassette transporter protein A1, a membrane transporter involved in MIF secretion. These findings emphasize the mutual cooperation between hormones and cytokines and suggest that increasing estrogen levels with advancing gestation may have a major role in regulating placental MIF secretion.

  • ATP-binding cassette transporter protein A1
  • steroid hormones
  • placental cytokines
  • chorionic villous explants

endocrine, paracrine, and autocrine factors are believed to play a major role in creating the environment essential for placental establishment and development. In particular, hormones and cytokines appear to be very important in the complex interaction between the immunoendocrine and reproductive systems in the fetomaternal unit (31, 33, 41, 50).

17β-Estradiol (E2), the most potent and predominant estrogen in human serum, increases gradually during pregnancy until term (42). E2 has a marked effect on the immune system, having either pro- or anti-inflammatory properties in a concentration- and tissue-dependent manner (51). Although numerous studies have demonstrated the influence of estrogens on immune functions (12, 35), the molecular mechanisms involved are poorly understood. Several studies suggest that estrogens directly affect cellular function, including cytokine production (29, 34).

The cytokine macrophage migration inhibitory factor (MIF) is a key regulator of the inflammatory and immune responses (11, 19). MIF is unique among cytokines in overcoming glucocorticoid anti-inflammatory actions by inducing monocyte production of IL-6, IL-1, TNFα, and IL-8 (14). MIF also plays an important role in the innate and adaptive immune responses (7, 15), including generation of matrix metalloproteinases, arachidonic acid products, nitric oxide induction, and T cell cytokines (37, 46, 49, 59). Recent evidence has shown that MIF is one of the cytokines involved in pregnancy, particularly during the earlier stages of placentation (28). MIF mRNA and protein are highly expressed in human trophoblast mostly at 7–10 wk of gestation (4, 28). High expression of MIF mRNA and protein has also been found in human decidua at the site of implantation and in the endometrium with a cycle-dependent expression, suggesting a potential hormone regulation of MIF expression (5, 32).

Reports in mice have highlighted a direct interaction between estrogens and MIF (6, 22, 47). It has been demonstrated that MIF triggers delayed healing in ovariectomized mice. Indeed, MIF-null mice heal normally in the absence of estrogens, and estrogens modulate cutaneous wound healing almost exclusively via MIF downregulation (6). More recently, sex divergence has been shown in wound healing, and wound areas are in fact greater in ovariectomized female mice than in castrated males (22). Oshima et al. (47) used ovariectomized mice as a model for postmenopausal osteoporosis and demonstrated that ovariectomy rapidly induces the secretion of MIF. Association between estrogens and MIF has been shown in the susceptibility to develop colitis (25) and in the attenuation of lung injury (26). In the last two studies, estrogens affected the early acute phase of inflammation, decreasing MIF production in the female rat colon and in mouse lung tissue damaged by trauma-hemorrhage (25, 26). In a previous study regarding the sex differences in the chronic pain conditions, we showed a negative correlation in the female control group between MIF and estradiol plasma level in humans (2).

Since human placenta is estrogen-responsive tissue expressing both estrogen receptor (ER)α and ERβ (10, 13) and is physiologically exposed to increasing concentration of estrogens during pregnancy, in the present work we tested the possibility that placental MIF might be modulated by estrogens.

MATERIALS AND METHODS

Tissue Collection

First-trimester human placentas (n = 26) were obtained after elective termination of pregnancy at 7–9 wk of gestation, with the consent of the patients and approval of the local ethics committee (Siena, Italy, 2004) in accordance with the Helsinki Declaration guidelines. Tissues were rinsed in cold phosphate-buffered saline (PBS) to remove excess blood and were processed for explant cultures within 2 h.

Chorionic Villous Explants and Hormone Treatments

Chorionic villous explants were dissected as described by Caniggia et al. (16). Briefly, small fragments of villous tips (15–20 mg wet wt) were placed on Millicell CM culture dish inserts (Millipore, Billerica, MA) previously coated with 180 μl of undiluted matrigel (Collaborative Research, Bedford, MA) and inserted in 24-well plates. Explants were cultured in serum-free DMEM-F-12 (Invitrogen, Carlsbad, CA) without phenol red and supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin, and 2 mM l-glutamine (Sigma Chemical, St. Louis, MO) and maintained at 37°C in 5% CO2-95% air. Explant cultures were incubated overnight to allow attachment to the matrigel. Culture medium was then replaced with medium supplemented as above plus E2 (Sigma Chemical) at the final concentration of 10−12, 10−9, or 10−5 M in 0.1% ethanol. Parallel cultures were exposed to medium containing 0.1% ethanol and used as controls. Explants were incubated further for 4, 24, and 48 h. At the end of each incubation, culture medium was centrifuged and stored at −80°C until the assay for MIF and β-human chorionic gonadotropin (hCG) concentration. Tissue explants were removed from matrigel and immediately frozen in liquid nitrogen and stored at −80°C until processing for protein and mRNA expression. In some experiments, tissue explants were fixed in 10% buffered formalin and then embedded in paraffin for MIF and ATP-binding cassette transporter A1 (ABCA1) immunohistochemistry.

A single placenta was used for each experiment, and at least three explant cultures were set up for each experimental condition (E2 treatments and control cultures). Tissue samples and supernatants, from separate explant cultures, were pooled at the end of incubation before analyses.

Protein Extraction

Total proteins from chorionic villous explant cultures (n = 11 separate experiments) were obtained by tissue homogenization in ice-cold lysing buffer (50 mM Tris · HCl, 50 mM NaCl, 1% Triton X-100, 1% Na deoxycholate, 0.1% SDS) 100 mM containing sodium orthovanadate and protease inhibitor cocktail, including 4-(2-aminoethyl)benzenesulfonyl fluoride, pepstatin A, E-64, bestatin, leupeptin, and aprotinin (Sigma Chemical). Protein lysates were clarified by centrifugation at 13,000 g for 15 min at 4°C.

Alternatively, chorionic villous explants (n = 3 separate experiments) were processed for native membrane protein extraction with the ProteoExtract kit (Calbiochem-Novabiochem, San Diego, CA) according to the manufacturer's instructions.

Total and membrane protein concentrations were determined by Quick Start Bradford Protein Assay (Bio-Rad Laboratories, Hercules, CA).

MIF Colorimetric Sandwich Assay (ELISA)

MIF concentrations in the lysates of the n = 11 explant cultures and the corresponding culture medium were measured using an enzyme-linked immunosorbent assay (ELISA), as described previously (28). One-hundred μl/well of anti-human MIF monoclonal antibody (2.0 μg/ml; R & D Systems, Abingdon, UK) was used for dish coating and 100 μl/well of biotinylated goat anti-human MIF polyclonal antibody (200 ng/ml; R & D Systems) for MIF detection. The MIF concentration was calculated by extrapolation from a standard curve (range 25–2,000 pg/ml) using bacterially expressed recombinant human MIF (R & D Systems). The sensitivity limit was 18 pg/ml. Intra- and interassay coefficients of variation were 3.86 (±0.95) and 9.14 (±0.47)%, respectively.

β-hCG Immunoenzymometric Assay

The endocrine integrity and viability of villous explants under the various experimental conditions were monitored by measuring β-hCG in the chorionic villous explant culture medium using a commercial immunoenzymometric assay (Radim, Pomezia, Italy).

MIF and ABCA1 Western Blot Analysis

Ten micrograms of chorionic villous explant total proteins or membrane proteins was separated on 12 or 7% polyacrylamide gels in the presence of SDS and β-mercaptoethanol to detect MIF and ABCA1, respectively. After electrophoresis, gels were removed and equilibrated in transfer buffer [20 mM Tris, 190 mM glycine, and 20% (vol/vol) methanol, pH 8.3] for 5 min at room temperature. Proteins were transferred to nitrocellulose membranes (Amersham International, Little Chalfont, UK) for 1.5 h. The blots were incubated in blocking solution [7% (wt/vol) powdered milk in 10 mM PBS, 0.1% Tween-20] for 4 h and exposed to mouse anti-human monoclonal MIF (R & D Systems) or ABCA1 antibodies (Abcam, Cambridge, UK), both of which were diluted 1:1,000 overnight at 4°C. The membranes were washed three times with PBST (0.1% Tween-20 in PBS 10 mM) and exposed for 1 h to the goat anti-mouse antibody (dilution 1:3,000) labeled with peroxidase (Bio-Rad Laboratories) at room temperature. The blots were washed three times with PBST and visualized using a chemiluminescence kit (Bio-Rad Laboratories) according to the manufacturer's instructions. All blots were confirmed for equal protein loading and transfer using Ponceau staining (44).

MIF and ABCA1 qRT-PCR Analysis

Total RNA was extracted from placental villous explants (n = 9 separate experiments) using the RNeasy mini kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. One microgram of total RNA was reverse transcribed using random hexamer and MultiScribe enzyme (Applied Biosystems Group, Foster City, CA).

Quantitative PCR (qRT-PCR) reactions were run in the StepOne Real-Time PCR System instrument (Applied Biosystems Group) using TaqMan chemistry. Two microliters of cDNA in a final volume of 20 μl was amplified using the 20× Assays-on-Demand gene expression assay mix (Applied Biosystems Group). TaqMan probes and specific primers for MIF, ABCA1, and ribosomal 18S, selected as housekeeping gene, were purchased from Applied Biosystems Group.

The MIF and ABCA1 mRNA levels were normalized to those of 18S, and the relative mRNA levels after E2 treatment were calculated using the ΔΔCT method (39). The fold change of MIF and ABCA1 mRNA levels in E2-treated cultures vs. controls were expressed as 2−ΔΔCT.

Immunohistochemistry for MIF and ABCA1

Immunohistochemical staining was performed on formalin-fixed and paraffin-embedded villous explant cultures (n = 3 separate experiments). Tissue sections (4 μm) were processed by the streptavidin-biotin method. Nonspecific binding sites were blocked using 2% normal rabbit serum diluted in 0.05 M Tris-buffered saline (TBS; pH 7.4) for 30 min at 37°C. Slides were incubated overnight at 4°C with 1:200 TBS dilution of mouse monoclonal anti-human ABCA1 (Abcam) or anti-human MIF (R & D System) antibodies. The preparations were then rinsed in TBS and incubated with biotinylated rabbit anti-mouse antibody (DAKO, Glostrup, Denmark) at a dilution of 1:500. The reaction signal was amplified using the StreptABComplex/AP (DAKO) developed with fast red-naphthol (Sigma Chemical) and counterstained with Mayer's hematoxylin. For each staining, a negative control was carried out by replacing the primary antibody with mouse nonimmune serum immunoglobulins at the same concentration as the primary antibody.

Statistical Analysis

All data are presented as means ± SE of n = 3–5 separate experiments for the MIF concentration in the culture medium and MIF protein intratissutal content: n = 9 for MIF mRNA, n = 4 for ABCA1 mRNA, and n = 3 for ABCA1 membrane protein expression. The statistical significance of the data was determined with one-way ANOVA or Student's t-test when appropriate. qRT-PCR statistical analysis was performed with the relative expression software tool, a pair-wise fixed reallocation randomization test (39). P values <0.05 were considered statistically significant.

RESULTS

Effect of E2 on MIF Production

MIF release in culture medium.

The specific sandwich ELISA showed that MIF release (normalized to the total explant tissue proteins) in supernatants of explant cultures was modulated by treatment with E2 at various concentrations (10−12, 10−9, or 10−5 M), reaching significant differences at 24 and 48 h of incubation (Fig. 1). At 24 h, comparison between control and treated cultures showed that E2 at 10−12 M was significantly increasing (P < 0.05), E2 at 10−9 M was ineffective, and E2 concentrations as high as 10−5 M were significantly decreasing (P < 0.05) the secretion of MIF. At 48 h, MIF levels reached a peak in control cultures, and a decrease in MIF secretion was found at each E2 concentration tested (10−12, 10−9, or 10−5 M). As shown at 24-h incubation, a dose-dependent reduction in MIF secretion was observed with increasing E2 concentrations (10−12 vs. 10−9 M, P < 0.05; 10−9 vs. 10−5 M, P < 0.05; 10−12 vs. 10−5 M, P < 0.001). Tissue viability was established by the continuous hormone (β-hCG) secretion throughout incubation. According to previous studies (10), levels of β-hCG in the culture medium were growing more in E2-treated cultures than in control with increasing time of incubation, becoming significant at 48 h (data not shown). This led us to select 24-h incubation as the experimental end point for further analyses.

Fig. 1.

Macrophage migration inhibitory factor (MIF) release in the culture medium by first-trimester chorionic villous explants. Explants were incubated for 4, 24, and 48 h with different 17β-estradiol (E2) concentrations (10−12, 10−9, or 10−5 M) or medium plus 0.1% ethanol (control). MIF secretion in the culture medium was quantified by an ELISA assay and expressed as pg/mg of total explant tissue proteins. Data are means ± SE of n = 3–5 separate experiments (n = 3 for the 4 h of incubation, n = 5 for the 24 h of incubation, and n = 3 for the 48 h of incubation). At 24 h, MIF secretion increased significantly in explant cultures treated with 10−12 M E2 and decreased in explants treated with 10−5 M E2 compared with control cultures using one-way ANOVA analysis. A significant decrease in MIF secretion was observed when 10−12 M E2-treated cultures were compared with 10−9 or 10−5 M E2 using Student's t-test. A significant reduction was also observed when 10−9 M E2-treated cultures were compared with 10−5 M E2. At 48 h, MIF secretion decreased significantly in explants treated with 10−9 or 10−5 M E2 compared with control cultures using one-way ANOVA analysis. Significant decrease in MIF secretion was observed when 10−12 or 10−9 M E2-treated cultures were compared with 10−5 M using Student's t-test. No significant differences were observed at 4 h of incubation. *P < 0.05; **P < 0.001.

MIF protein in tissue explants.

Western blot analysis of total protein extract lysates showed a specific band of 12.5 kDa, corresponding to the molecular weight of MIF, in cultures treated with 10−12, 10−9, and 10−5 M E2 and in untreated control cultures (Fig. 2A) after 24 h of incubation.

Fig. 2.

MIF protein expression by first-trimester chorionic villous explants. Explants were incubated for 24 h with different E2 concentrations (10−12, 10−9, or 10−5 M) or medium plus 0.1% ethanol (control). A: representative Western blot of villous explant lysates analyzed by SDS-PAGE followed by detection with anti-MIF monoclonal antibodies. A specific band of 12.5 kDa, corresponding to the molecular weight of MIF, was revealed in all samples. Equal protein loading was assessed by Ponceau staining. B: quantification of MIF in explant lysates by ELISA assay. MIF is expressed as pg/mg of total tissue proteins. Data are means ± SE of n = 5 separate experiments. No difference in MIF concentration was detected at any E2 concentration with respect to control cultures. C: total RNA was extracted from n = 9 separate experiments, and MIF mRNA expression was evaluated by qRT-PCR. 18S was used as housekeeping gene. MIF mRNA levels were normalized to those of 18S and expressed as MIF fold increase relative to control cultures. No significant difference in MIF mRNA levels was detected at any E2 concentration with respect to control cultures.

Quantification by ELISA revealed an extremely high MIF concentration in the chorionic tissues, 100 times higher than that found in the culture medium (Fig. 2B). Unlike MIF release, intratissutal MIF protein content was not significantly affected by E2 treatment at any of the concentrations used (Fig. 2B).

MIF mRNA in tissue explants.

Relative quantification of MIF mRNA was performed with qRT-PCR (Fig. 2C) in explants culture after 24 h of incubation. Endogenous expression of MIF was readily detected at 23 ± 1.8 cycles (data not shown). Similarly to MIF protein concentration, we found that the different concentrations of E2 (10−12, 10−9, or 10−5 M) did not significantly affect MIF transcript expression levels with respect to untreated control cultures. Although not significant, decreasing MIF transcript levels were observed with 10−5 M E2 compared with control cultures.

Effect of E2 on ABCA1 Expression

A recent study showed that ABCA1 transporter protein is involved in MIF secretion (21). To determine whether differences in MIF release in culture medium of chorionic villous explants after treatment with different concentrations of E2 were due to a variation in ABCA1, we examined ABCA1 expression in E2-treated and untreated cultures (controls). Separate experiments (n = 3) were processed for isolation of membrane proteins for ABCA1 expression, and n = 4 of the n = 9 experiments used for MIF mRNA analysis were also used for ABCA1 mRNA analysis.

ABCA1 protein expression.

Protein expression in purified membranes was analyzed by Western blotting using specific anti-human ABCA1 monoclonal antibody. Figure 3A shows a specific band of 254 kDa, corresponding to the predicted molecular weight of ABCA1 transporter protein, in E2-treated and untreated cultures. Densitometric analysis showed that treatment with 10−5 M E2 resulted in a significant reduction of ABCA1 transporter protein expression compared with untreated control cultures (Fig. 3B).

Fig. 3.

ATP-binding cassette transporter protein A1 (ABCA1) protein expression by chorionic villous explants. Explants were incubated for 24 h with different E2 concentrations (10−12, 10−9, or 10−5 M) or medium plus 0.1% ethanol [control (C)]. A: representative Western blot performed on membrane-purified total proteins by SDS-PAGE followed by detection with anti-ABCA1 monoclonal antibodies. A specific band of 254 kDa, corresponding to the molecular weight of ABCA1, was detected in all samples. A′: equal loading was assessed by Ponceau staining. St, standard markers. B: densitometric analysis performed on n = 3 separate experiments showed a significant (*P < 0.05) reduction of ABCA1 expression in explant cultures treated with 10−5 M E2 compared with controls. C: total RNA from n = 4 separate experiments was analyzed for ABCA1 mRNA expression by qRT-PCR. 18S was used as housekeeping gene. ABCA1 mRNA levels were normalized to those of 18S and expressed as ABCA1 fold increase relative to control cultures. Significant decrease in ABCA1 mRNA levels was detected at 10−9 (P = 0.027) and 10−5 M (P = 0.020) E2 concentrations with respect to control cultures.

ABCA1 mRNA expression.

We then examined whether the effects of E2 on ABCA1 protein expression were due to changes in ABCA1 mRNA levels. Endogenous ABCA1 mRNA in explant cultures could be detected by real-time quantitative PCR at 28 ± 1.8 cycles (data not shown). RT-PCR analyses revealed that treatment of explant culture with E2 at 10−9 and 10−5 M significantly decreased ABCA1 mRNA ∼1.4- and 1.7-fold, respectively, compared with the vehicle control cultures (Fig. 3C). Treatment with E2 at 10−12 M did not affect ABCA1 mRNA expression level with respect to control cultures.

MIF/ABCA1 Correlation

We next correlated the MIF levels in the culture media of explants vs. ABCA1 tissue expression at 24 h of incubation with 10−12, 10−9, or 10−5 M E2. We found significant correlations between MIF and ABCA1 protein (P < 0.0001; Fig. 4A) and MIF and ABCA1 mRNA (P < 0.05; Fig. 4B). Also noteworthy was that a strong positive linear relationship (Pearson's correlation coefficient = 0.83) was found between MIF and ABCA1 protein expression.

Fig. 4.

MIF/ABCA1 correlation. Correlation of MIF levels in the culture medium vs. ABCA1 tissue expression at 24 h of incubation with 10−12, 10−9, or 10−5 M E2. MIF concentration correlates significantly with ABCA1 protein (y = 0.0099x + 0.91, r2 = 0.69, P < 0.0001, Pearson r = 0.83; A) and ABCA1 mRNA (y = 0.0028x + 0.46, r2 = 0.27, P = 0.026, Pearson r = 0.52; B) expression.

Immunohistochemistry for MIF and ABCA1

In subsequent experiments (n = 3), we performed MIF and ABCA1 immunohistochemistry in E2-treated and untreated control cultures. Results showed high MIF immunoreactivity in the villous and extravillous cytotrophoblast, with no detectable differences between E2-treated and untreated cultures (Fig. 5, A, C, and E). In agreement with our previous study (23), moderate staining for MIF was also observed in some elements of villous stroma, whereas the syncytiotrophoblast was negative (Fig. 5, A, C, and E). ABCA1 was expressed mainly by the same cell types expressing MIF, including the villous and extravillous cytotrophoblast, as well as some cellular elements of the villous stroma (Fig. 5, B, D, and F). Just as was found for MIF, no staining was observed in the syncytiotrophoblast. According to Western blot analysis, a generalized reduction of ABCA1 immunostaining was observed after treatment with 10−5 M E2, and no difference was observed in 10−9 (data not shown) and 10−12 M E2 relative to control cultures (Fig. 5, D and F).

Fig. 5.

Immunohistochemistry for MIF and ABCA1 in chorionic villous explants. Representative immunostaining performed with the streptavidin-biotin method and anti-human MIF (A, C, and E) and anti-human ABCA1 (B, D, and F) monoclonal antibodies. A and B: explants treated with 0.1% ethanol (controls). C and D: explants treated with 10−12 M E2. E and F: explants treated with 10−5 M E2. Reddish staining represents immunoreactivity. MIF expression was localized in the villous cytotrophoblast (arrows), in the extravillous cytotrophoblast invading the matrigel (M; double arrows), and in some stromal elements (arrowheads) of both E2-treated and control cultures. The syncytiotrophoblast (*) was negative. No remarkable MIF difference was observed between 10−12 (C) and 10−5 M (E) E2 concentrations. Similarly to MIF, ABCA1 was expressed mainly in the villous cytotrophoblast (arrows), in the extravillous cytotrophoblast invading the matrigel (M; double arrows), and in some stromal cells (arrowheads) both in E2-treated (D and F) and control (B) cultures. The syncytiotrophoblast (*) was negative. A generalized reduction of ABCA1 immunostaining was observed after treatment with 10−5 M E2 (F) with respect to 10−12 M E2 (D) and control (B). Negative controls were performed by substituting the primary antibodies with nonimmune mouse serum immunoglobulins at equimolar concentrations (in E and F, insets). Bar = 50 μm.

DISCUSSION

MIF mRNA and protein are expressed in hormonally responsive reproductive tissues, including the ovary, endometrium, and placenta (4, 5, 56) and contribute to reproductive processes such as ovulation (56), menstrual cycle (5), early pregnancy (4, 28), and parturition (27). This study was designed to explore the link between estrogens and MIF in first-trimester human placenta. Using an in vitro model of chorionic villous explants, we demonstrated that E2 is able to modulate the release of MIF in a dose-dependent manner. After 24 h of exposure, E2 at 10−12 M significantly increased MIF secretion in the culture medium, whereas E2 at 10−9 M was ineffective and E2 at 10−5 M significantly reduced MIF release with respect to untreated control cultures. Hence, there was a dose-dependent reduction of MIF secretion with increasing E2 concentration from 10−12 to 10−5 M. It is noteworthy that there was no significant change in tissue MIF protein levels or MIF mRNA expression at any of the E2 concentrations used. We also found further evidence that E2 concentrations (10−9 and 10−5 M) act on placental explants by downregulating the mRNA and protein expression of ABCA1, which was recently shown to be involved in MIF secretion (21).

Therefore, our results suggest that E2 concentrations modulate placenta MIF release, regulating the expression of ABCA1 transporter protein.

It is noteworthy that MIF tissue protein content did not result in being affected despite the difference in tissue release. Therefore, concentration of MIF in the culture medium is only 1–3% of the tissue MIF protein under the coefficients of variation (intra-assay: 3.86 ± 0.95%; interassay: 9.14 ± 0.47%) of MIF ELISA immunoassays. This makes differences in MIF protein content between E2-treated and control cultures undetectable. These findings also reflect the unique MIF production pattern and mechanism of its release by the producing cells. Unlike other cytokines, MIF is in fact constitutively stored, and it is released upon stimulation from preformed and the de novo synthesized pools (9, 11).

During development, the placenta acquires endocrine functions and is able to produce various hormones, including estrogens and progesterone, hCG, and placental lactogen, many of which have immunoregulatory roles (12).

The circulating levels of estrogens increase markedly with advancing gestation from 10−10 M at mid-trimester to around 10−7 M at term (8). Thus the placenta is constitutively exposed to increasing E2 concentrations. So far, no data report E2 concentrations in the intervillous space. Since the placenta itself produces estrogens, local levels in the intervillous space are likely to become much higher than those in the peripheral circulation with advancing of gestation. On the other hand, local estrogen levels can be lower than in peripheral circulation until the placenta is not be able to produce its own estrogens (53). In the early stages of placentation, trophoblast aggregates also occlude the uterine spiral artery limiting the blood flow into the intervillous space (30).

In our in vitro experiments, E2 at 10−9 and 10−5 M, representative of those observed in pregnant women, decreased MIF release modulating ABCA1 expression. By measuring the release of the β-hCG, a marker of the continuous endocrine activity and therefore of the integrity of placental barrier in explant cultures (43), we demonstrated that culture functionality was preserved even at 10−5 M, the highest E2 concentration used.

Increasing evidence indicates that E2 has diverse effects on inflammation and immune system functions; in particular, fluctuations of E2 concentration seem to be the key elements determining its immunomodulatory properties (24, 51). In addition, the ability of estrogens to have either pro- or anti-inflammatory properties is principally tissue and cell dependent (51). This is particularly evident when the ability of estrogens to induce the production of immunoregulatory mediators such as cytokines is considered (17, 54).

MIF secretion relied on an atypical protein-secretion pathway that does not use an amino-terminal leader sequence (45). It has been demonstrated in human acute monocytic leukemia cell lines that MIF secretion occurs via a nonclassical export route, a pathway that involves ABCA1 transporter protein (21).

ABCA1 is one of the major players in mediating cellular lipid efflux. Unlike other members of the family that have a more restricted pattern of expression, it is broadly expressed at high levels in macrophages, endothelial cells, liver and intestinal cells in mice, and human trophoblast at term (1, 36, 38). We have shown here that, in first-trimester human placenta, ABCA1 is expressed mainly in the villous and extravillous cytotrophoblasts, the same cell types expressing MIF, and that ABCA1 expression is downregulated by E2 concentrations (10−9 and 10−5 M).

Whereas the effects of E2 on MIF release did not find any correlation with MIF transcriptional or protein level, the secretion of MIF in the culture medium significantly correlated with ABCA1 protein and mRNA expression. In particular, the highly significant association between MIF secretory levels and ABCA1 protein expression as well as their strong positive linear relationship suggested a direct action of E2 on ABCA1 membrane expression rather than on its activity.

Estrogens elicit various biological effects by binding to the two distinct estrogen receptors, ERα and ERβ, members of the nuclear receptor family of ligand-dependent transcription factors (20). We demonstrated recently that first-trimester human placenta expresses both ERα and ERβ and that it is responsive to the action of estrogens by increasing cell differentiation and apoptosis (10).

The inhibitory effect of E2 observed on ABCA1 gene expression appears to be the result of hormone-activated ERs, leading ultimately to transrepression mechanism. Interactions of activated ERs with other transcriptional factors as well as competition for limiting amounts of general coactivators could be implicated in transrepression of ABCA1 in the placental tissues (23). However, the possibility of a posttranscriptional mechanism cannot be excluded. Several protein kinases are in fact involved in the regulation of the expression level and activity state of the ABCA1 transporter. E2 via the nongenomic pathway can also activate several protein kinases, including PKA and PKC, both of which are involved in the degradation of the transporter by calpain protease and thereby reduce ABCA1 cell surface expression (48, 57).

In humans, placental MIF was first described by Zeng et al. (60), who found a MIF-like protein at term. In early pregnancy, MIF has been detected at the site of implantation in both maternal decidua and trophoblast (4). It is noteworthy that MIF mRNA and protein levels are higher in the very early gestational stages (6–10 wk of gestation) and decline in the late first trimester (12 wk of gestation) (28). Moreover, by using chorionic villous explant cultures from first-trimester placenta, we showed that MIF protein and mRNA are upregulated by low oxygen tension (28).

Thus far, few studies have shown the importance of MIF in early gestation. Arcuri et al. (3) showed that trophoblast MIF reduces the cytotoxicity of decidual natural killer cells. In vivo studies have shown that decreasing maternal serum MIF levels are associated with first-trimester miscarriages (58). Events occurring in the earlier phases of gestation, including implantation and placenta establishment, are characterized by inflammatory-like processes, and the role of proinflammatory cytokines, i.e., MIF, has been emphasized (18, 27, 28). Several studies indicate that MIF is a mediator of excessive inflammation (40). Concerning pregnancy, high serum levels of MIF have been found in women affected by preeclampsia, a multifactorial disorder involving an excessive inflammatory response (52).

As a potent immunomodulator, MIF expression must be closely regulated. Recently, Verjans et al. (55) showed that exogenous MIF triggers a dramatic upregulation of MIF secretion, leading to cell proliferation and invasion in breast cancer cells. These authors also suggested that intracellular MIF has a protective function, whereas extracellular MIF plays a major role as pro-oncogenic factor (55). Because intratissutal MIF shows a positive correlation with ER expression marker of favorable prognosis (55), we could speculate that estrogens act on these tissues by modulating MIF secretion, as shown here for trophoblast.

In conclusion, the data reported herein show that E2 plays a major role in regulating MIF secretion by trophoblast. Since E2 concentrations vary throughout pregnancy, the hormone-cytokine interaction appears to be a key element for regulating concentration of MIF at the maternal-fetal interface.

GRANTS

This research was supported by the Sixth European Union Framework Program, Integrated Project ReProTect.

DISCLOSURES

No conflicts of interest are declared by the author(s).

ACKNOWLEDGMENTS

We thank the Obstetrics and Gynecology Division, USL7, Hospital Campostaggia, Siena, Italy, for providing placental tissues. We are also grateful to Federica Paulesu for English revision.

REFERENCES

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
View Abstract