| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Fearing Research Laboratory, Department of Obstetrics, Gynecology, and Reproductive Biology, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts 02115
Address all correspondence and requests for reprints to: Danny J. Schust, M.D., Division of Reproductive Biology, Department of Obstetrics and Gynecology, Boston Medical Center, Boston University School of Medicine, 715 Albany Street, Evans 222, Boston, Massachusetts 02118. E-mail: danny.schust{at}bmc.org.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
. T-bet is here shown to be present in epithelial cells of the human female reproductive tract. Regulation of T-bet expression was modulated by cytokines and the female reproductive steroids, estrogen, and progesterone. The mechanisms of T-bet regulation in epithelia differ from those in conventional immune cells. During a 15-d exposure to progesterone, T-bet levels in endometrial epithelial cells (EECs) undulated. Prior exposure to estrogen enhanced these effects. More prolonged exposure of EECs to these hormones, singly or in combination, suppressed T-bet production. Stat1 and Stat5 bound to the EEC T-bet regulatory region (TRR) at the IFN--activated sequence site, but Stat3 and Stat4 did not. Binding of Stat1 and Stat5 to the TRR were modified by progesterone in distinct ways. Estrogen suppressed the binding of Stat1 and Stat5 to the TRR. Mutation of -activated sequence element reduced T-bet promoter activity, binding of Stat proteins to the TRR and regulation of the promoter by cytokines and hormones. In EECs, cytokine exposure caused phosphorylation of Janus kinase 2 and TRR-bound Stat proteins; female steroid hormones altered only phosphorylation of TRR-bound Stat5. Although there is no autocrine IFN- feedback loop in reproductive tract epithelial cells, an IL-15/T-bet positive feedback loop may exist. The implications of hormonally regulated T-bet expression are discussed. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Progesterone suppresses TH1 responses, decreases cytotoxic T lymphocyte activity, increases antibody secretion, and increases susceptibility to microbial infections in mice or human exposed to progesterone (1, 2, 3, 6). The molecular mechanisms for the immunological alterations induced by female steroid hormones remain undefined. For instance, it is unclear whether progesterone produces its immunosuppressive effects by direction action on T cells because it is not clear that these cells express progesterone receptors (PRs) (1, 7, 8).
Estrogen and progesterone are both elevated during human pregnancy, and their combined effects on immune status is thought to support an environment favoring pregnancy maintenance. Such an environment favors differentiation of TH2 cells over that of TH1 cells. Differentiation of naive T helper cells along the TH1 pathway is directed by a transcription factor called T-bet (9, 10). Although T-bet was initially described in T and NK cells, studies have demonstrated T-bet expression in other immune cells, including a diverse array of antigen-presenting cells (APCs), such as monocytes, macrophage cells, dendritic cells, and myeloid cells. The mechanisms of tissue-specific expression of T-bet have not been well explained and little is known of its expression in mucosal immune cells, including the epithelial cells of the female reproductive tract. Recently, roles for T-bet in asthma and Crohn’s disease have been described (11, 12).
The signal transducers and activators of transcription (Stat) proteins are latent cytoplasmic transcription factors that are activated via receptors for cytokines, growth factors, and hormones. Stat1 is central to the regulation of T-bet expression in lymphoid and nonlymphoid cells through interferon (IFN)- receptor/Janus kinases (JAK) signaling (13, 14, 15). Stat4, synergistic with IL-12 receptor-ß2 (IL12Rß2), is essential for Th1 development in naive T cells or Th1 cells (reviewed in Ref. 14). T-bet is an inducer of the IL12Rß2/Stat4 pathway in Th1 cells resulting in IFN- induction (15). The induction of chromatin remodeling of the interferon (IFN)- locus by T-bet establishes TH1 effector cells (9) and results in the induction of IFN- and repression of IL-4 and IL-5 expression (10). IFN-, in turn, induces T-bet expression in these cells forming a positive feedback loop (10). However, the IL12Rß/Stat4 pathway is not essential for induction of T-bet expression in either T cells or APCs because T-bet induction of IFN- is not altered in Stat4-deficient T cells and APCs (13, 14).
Female steroid hormones, signaling through Stat1 and Stat5 pathways in diverse cell types, including endometrial epithelial cells (EECs) and stromal cells, modulate gene expression (16, 17, 18, 19, 20, 21, 22). Estrogen suppresses Stat5-mediated gene transcription (17, 19). In contrast, estrogen induces phosphorylation of Stat1 and 5 in a rapid, but transient manner (18, 22). Progesterone activates the Stat5 pathway while suppressing the Stat1 pathway via the protein inhibitor of activated signal transducer and activator of transcription (PIAS)-y (16, 17, 20, 21). These in vitro data are consistent with data from human endometrial materials (23).
In humans, Stat1 and Stat5 are expressed in normal endometrial glandular epithelia. The expression level of Stat5 is negligible during the proliferative phase but increases during the secretory phase of the menstrual cycle. Levels of Stat1 in endometrial glandular epithelia are stable throughout the menstrual cycle (23). Additionally, studies on Stat5-deficient mice indicate that Stat5 plays critical roles in a variety of reproductive processes. One well-defined mechanism for Stat5 action involves the Stat5b-mediated suppression of 20
-hydroxysteroid dehydrogenase. This, in turn suppresses progesterone production (24, 25).
Viewing these data, we hypothesized that female steroid hormones regulate T-bet expression in reproductive epithelial cells via Stat proteins and that this results in specific alterations of innate and adaptive immunity in hormone-responsive tissues. To examine the hypothesis, we investigated the effects of female steroid hormones on T-bet expression in epithelial cells of the reproductive tract. We focused particularly on EECs, which are exquisitely responsive to the cyclic changes in female steroid hormone exposure in vivo and central to pregnancy maintenance. Although several TH1-related cytokines, including IFN- and IL-15, enhance the binding of Stat proteins to the IFN- promoter (26, 27), there has been no report of the binding of Stat proteins to the T-bet regulatory region (TRR). We have examined the binding of Stat proteins to the TRR in EECs and the actions of female steroid hormones and TH1-relevant cytokines in this process.
IFN- secretion causes a positive feedback loop with T-bet expression in lymphoid cells. Epithelial cells lining the human female reproductive tract have IFN- receptors and secrete TH1-cytokines in response to IFN- secreted from dendritic cells, macrophage, and T cells homing to these epithelia and their supporting stroma (reviewed in Refs. 1 , 28 , and 29). However, IFN- is not expressed by mucosal epithelial cells of the reproductive tract, including EECs (28). Given the differences in immunological characteristics between lymphoid and epithelial cells, we investigated whether T-bet has a role in reproductive epithelial cells that is distinct from that in lymphoid cells.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
). In mouse spleen (positive control), scattered small lymphocytes in the periarteriolar T-cell zone were immunoreactive to the 4B10 antibody (30). In vaginal squamous epithelia, nuclei of the basal cells were clearly immunoreactive to the anti-T-bet monoclonal antibody, although the immunoreactivity in more superficial layers was weaker. T-bet immunoreactivity was observed in endometrial glandular cells isolated in the proliferative phase, whereas T-bet was barely detectable in endocervical glandular cells. In all tissues, stromal cells were not immunoreactive to the 4B10 antibody.
|
) and Western immunoblotting (Fig. 2B) of immortalized epithelial cell lines were performed. These experiments revealed that T-bet mRNA and protein were present at varying levels in cells of endometrial, endocervical, and vaginal origin. Basal levels of T-bet mRNA and protein were high in vaginal cells but lower in endometrial and cervical cells. Exposure of cells to IFN- (10 ng/ml) for 9 h before harvesting led to increased levels of T-bet mRNA in all cell types tested with the most dramatic increase observed in vaginal cells. T-bet protein levels also increased upon exposure to IFN-. The largest increment was observed in vaginal cells (59%). The changes in endometrial and cervical cells were less dramatic (20 or 25% increase, respectively).
|
). Luciferase assays using cells transfected with the T-bet-luc vector indicated that T-bet promoter activity was strong in vaginal cells and weaker in endometrial and cervical cells (Fig. 3B), consistent with the T-bet mRNA and protein expression data (Fig. 2
).
|
). T-bet promoter activity was enhanced by IFN- in all epithelial cells types tested. IL-4, a cytokine critical for determination of the TH2 lineage, suppressed T-bet promoter activity in endometrial, cervical, and vaginal cell lines. Progesterone at physiologic concentrations enhanced T-bet promoter activity in EECs, but not in the other cells. Nine hours of exposure to estrogen did not significantly affect promoter activity in any of the cells tested.
In the female reproductive tract, epithelia are exposed to elevated levels of female sex hormones over weeks at a time. Therefore, the effect of varying lengths of exposure to sex hormones on T-bet mRNA levels in EECs was determined. RT-PCR of progesterone-exposed EECs revealed that T-bet mRNA levels increased after 2 d of exposure, decreased at d 4, rose again by d 6, and decreased again by d 8 (Fig. 4A, left panel). Quantitative real-time RT-PCR demonstrated that the up-regulation of T-bet transcription on d 2 of progesterone exposure was dose dependent over a range of physiological concentrations (Fig. 4A, right panel). Both semiquantitative and quantitative analysis showed that the maximum increase in T-bet mRNA levels upon progesterone (10–6 M) exposure was 2-fold. Alterations in T-bet protein levels upon progesterone exposure were also cyclic with a maximum variation of 1.5-fold. These alterations lagged behind changes in mRNA levels, likely due to accumulation of the protein in nucleus (10) (Fig. 4B). In contrast to progesterone, estrogen induced an increase in T-bet mRNA levels at 6 d of exposure followed by a decrease in these levels at d 8 (Fig. 4A, left panel). Levels of T-bet protein fluctuate similarly in estrogen-exposed EECs, but these changes lag behind changes in mRNA levels. In longer experiments, 15 d of exposure to estrogen or progesterone reduced T-bet protein levels in EECs (estrogen, 34% reduction; progesterone, 68% reduction) (Fig. 4C).
|
). As expected, the initial 14-d exposure to estrogen alone resulted in decreased levels of T-bet mRNA when measured by quantitative real-time RT-PCR. Subsequent addition of progesterone, in the continued presence of estrogen, resulted in an undulating response in T-bet mRNA levels over the following 15 d. When compared with earlier experiments, estrogen priming altered the time-course of the changes in T-bet mRNA levels induced by progesterone. In the absence of estrogen priming, peaks in T-bet mRNA levels occurred on d 2 and 6 of progesterone exposure whereas, after estrogen priming, peaks were observed on d 2 and 12 of progesterone exposure. The magnitude of the latter increase in T-bet mRNA levels was enhanced after estrogen priming when compared with that observed after progesterone exposure alone (Figs. 4A and 5). When dual hormonal stimulation was continued for an additional 4 d (32 d of total hormone exposure), T-bet mRNA levels became nearly undetectable. In contrast, if hormonal stimulation was withdrawn for these final 4 d, T-bet mRNA production recovered (data not shown).
|
-activated sequence (GAS) element at –1200 bp from the transcriptional start site of the T-bet gene. Binding of Stat proteins to GAS can be markedly diminished by a mutation of the GAS element, TTC(N)2–4GAA (24, 26). To evaluate the binding of Stat proteins to the putative TRR GAS element in EECs, three synthetic oligonucleotides were used to precipitate proteins present within the total cell extracts of EEC: a 50-oligomer oligonucleotide corresponding to the TRR containing the putative GAS element, an oligonucleotide mutated in the putative TRR GAS element [mutated TRR (mutTRR); specificity control] and an oligonucleotide containing the GAS element from the IFN- promoter (IFN--pro; positive control) (Fig. 6A). Experiments were performed on untreated cells and on those treated with IFN-, IL-15, progesterone, or estrogen. Western immunoblotting of precipitated proteins from control EECs (no hormone or cytokine treatment) using anti-Stat1, 3, 4, and 5 antibodies indicated that each of these Stat proteins bound to the IFN- promoter. Stat1 and Stat5 bound to the TRR but Stat3 and Stat4 did not (Fig. 6B). The anti-Stat5 antibody used can recognize both Stat5a (84 kDa) and 5b (90–92 kDa). Both the IFN- promoter and the TRR bound predominantly to the p91 form of the Stat1 protein that, unlike the p84 form, is essential for Stat1 transcriptional activity. Mutation of the GAS element (mutTRR) eliminated the binding of Stat1 and Stat5 to the TRR, indicating that the binding was sequence specific (26). IFN- and IL-15 induced the binding of Stat1, but not Stat5, to the TRR and the positive control (IFN--promoter) in EECs. Short-term estrogen exposure suppressed the binding of Stat1 and Stat5 to the TRR at 9 h, although binding of each transiently increased during the first 3–6 h after exposure (Fig. 6, B and C). Short-term progesterone exposure induced a biphasic pattern in Stat1 binding, whereas Stat5 binding peaked after 3 h of progesterone exposure (Fig. 6D). Stat1 and Stat5 bound the TRR with markedly different characteristics upon exposure to cytokines and female steroid hormones.
|
). Either wild-type T-bet-luciferase vector (wtT-bet-luc) or mutT-bet-luc was transfected into EECs, and promoter activities were assessed using luciferase assays. The amounts of DNA transfected into cells was estimated by amplifying the luciferase gene included in cell-associated DNA extracted from deoxyribonuclease I-treated cells. Results revealed no differences in the DNA content transfected into cells by each vector (data not shown). Nevertheless, the promoter activity of the TRR with the mutated GAS element decreased to approximately one fourth of that of the normal TRR in transfected cells (Fig. 7B). Although modulation of T-bet promoter activity after 9 h exposure to IFN-, IL-15, and female steroid hormones were similar in these experiments to those mentioned previously (Fig. 3), all modulations were lost with GAS mutation (Fig. 7B). Taken in their entirety, our data indicate that Stat family proteins play a critical role in T-bet expression and its regulation by cytokines and female steroid hormones in EEC.
|
; FLAG-T-bet fusion protein is 17 amino acids larger than endogenous T-bet). Using semiquantitative RT-PCR, we evaluated the expression of several TH1-relevant cytokines or cytokine receptors in FLAG-T-bet-transfected EECs and vector-transfected controls. Overexpression of the T-bet gene resulted in a significant induction in human IL-15 expression by EECs; human IFN-, IL-8, IL-12, IL-18, and IL12Rß2 were unaffected (Fig. 8B). IFN- and IL12Rß2 transcription by EECs were nearly undetectable. In lymphoid and dendritic cells, T-bet-induced IFN- production and IFN--induced T-bet expression form a positive feedback loop, resulting in rapid amplification of T-bet expression (13). In epithelial cells of reproductive tract, IFN- receptor is present, but IFN- secretion from the epithelial cells is absent (28). Given these characteristics, we examined whether IL-15 might serve a function in epithelial cells similar to that of IFN- in lymphoid cells. Over the first 24 h of IL-15 exposure, T-bet and IL-15 mRNA levels in EECs were altered in a parallel and biphasic manner. T-bet mRNA levels rapidly increased at 1 h after stimulation by IL-15, were below basal levels after 3 h, and began to increase again after 5–6 h of IL-15 exposure. This second increase peaked at approximately 18 h of exposure. IL-15 mRNA levels also increased rapidly at 1 h after IL-15 exposure, bottomed at 3 h, and rose again after 9 h. The latter increase in IL-15 mRNA levels in response to IL-15 lagged temporally behind the increase noted in T-bet mRNA levels. These results are consistent with a positive feedback interaction between T-bet and IL-15 in EECs (Fig. 9). To assess the effects of combined estrogen and progesterone treatment of EECs on IL-15 secretion, we used quantitative RT-PCR to examine IL-15 mRNA levels in hormone-treated EECs (Fig. 10). IL-15 mRNA levels increased most dramatically at d 12 of progesterone exposure (overall, d 26), corresponding to the second late elevation in T-bet mRNA levels demonstrated in Fig. 5B and to the timing of the mid-to-late menstrual secretory phase in vivo.
|
|
|
). Many membrane receptors for cytokines, growth factors, and protein hormones use pathways involving receptor-mediated phosphorylation of JAK family members. Upon phosphorylation, JAK proteins, in turn, phosphorylate a conserved tyrosine residue in the C-terminal region of Stat proteins (14, 24, 31). We compared these phosphorylation patterns with T-bet transcription levels. IFN- induced phosphorylations of JAK2 and Stat proteins. Although IL-15 did not alter Stat5 binding to the TRR (Fig. 6B), the phosphorylation of both JAK2 and Stat5 were induced by IL-15 in EECs. Although neither progesterone nor estrogen caused phosphorylation of JAK2, binding of phosphorylated Stat5 to the TRR was observed upon exposure to either female steroid hormone for 6 h, suggesting that these steroid hormones do not use JAK2-mediated Stat5 phosphorylation. The level of TRR-bound phosphorylated Stat1 in cells exposed to progesterone or estrogen was similar to that of nonexposed cells and consistent with data on the binding of Stat1 to the TRR upon 6 h exposure to either steroid hormone (Fig. 6, C and D).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
enhances T-bet expression in reproductive tract epithelial cell lines consistent with its previously demonstrated role in the control of T-bet expression in lymphoid and antigen presenting cells (10, 13, 14). IFN--induced T-bet expression in EECs and cervical epithelial cells is relatively weak compared with that in vaginal epithelial cells. In EECs, steroid hormones alter T-bet expression more dramatically than does IFN-. These distinct patterns of T-bet regulation may be linked to distinct roles for T-bet in EECs, vaginal, and cervical epithelia. In addition, we show that Stat1 and Stat5 are critical to the regulation of T-bet expression by cytokines and female steroid hormones in EECs. This is the first report to demonstrate that Stat1 and Stat5 recognize a GAS element in the TRR and bind to it. Stat protein binding to TRR regulates T-bet expression in these cells. In EECs, 0–8 d of progesterone exposure induced an undulating pattern of T-bet expression. By 15 d of exposure, T-bet levels were substantially reduced. In estrogen-primed cells, the stimulatory effects of progesterone on T-bet expression were enhanced. This may be due to up-regulation of PR levels by estrogen (32). We reasoned that the progesterone might be acting through Stat5 to regulate T-bet expression in EECs. PR-B signaling stimulates binding of Stat5 to target gene promoters and acts in synergy with growth factor and cytokine signaling to enhance transcriptional activation of Stat5 (16, 17, 21). Richer et al. (16) have demonstrated that PR and Stat5 can be coprecipitated and that progesterone promotes nuclear translocation of Stat5. They suggest that Stat5 may accompany PR molecules as they shuttle between the cytoplasm and nucleus. Consistent with these studies, exposure of EECs to progesterone modulated the binding of Stat1 and Stat5 to the TRR, in a time-dependent fashion. Stat1 binding was modulated in an undulating fashion, whereas Stat5 binding peaked after 3 h of exposure and decreased subsequently.
The Stat1 pathway regulates T-bet expression in lymphoid and dendritic cells (10, 13, 14). The observed progesterone-induced phosphorylation of TRR-precipitable Stat5, but not Stat1, in EECs is consistent with regulation of Stat1 in endometrial stromal cells (16, 20, 21, 23) and suggests that PR activation stimulates Stat5 regulated T-bet transcription and suppresses the Stat1 pathway. Zoumpoulidou et al. (20) recently demonstrated that PR interacts with PIAS-y to inhibit IFN-/Stat1 signaling pathways in human endometrial stromal cells. Others expanded on this to show that PIAS1 binds directly to Stat1 and interferes with the recruitment of Stat1, but not other Stat proteins, to promoters of IFN-inducible genes (33, 34). The PR/PIAS-y complex may inhibit Stat1 signaling in EECs, thereby contributing to the decrease in T-bet protein observed after 15 d of progesterone exposure.
We failed to demonstrate binding of Stat3 or Stat4 to the TRR GAS. Stat4 is a critical mediator in the development of Th1 cells from naive T cells via IL-12Rß2 signaling and IFN- induction. Consistent with our inability to demonstrate Stat4/TRR binding, EECs do not express IL-12Rß2 and IFN-. Additionally, the IFN--promoter and the TRR precipitated only Stat5b (90–92 kDa). This may reflect the fact that Stat5b is the predominant Stat5 protein in endometrium (21, 24)
Upon exposure of EECs to estrogen, there was a transient increase in Stat1 and Stat5 binding to the TRR. TRR bound Stat1 levels peaked after 3 h of estrogen exposure, whereas maximum TRR-bound Stat5 was observed after 6 h. At this time, there was increased TRR-precipitable phosphorylated Stat5, whereas TRR-precipitable phosphorylated Stat1 levels were unchanged compared with controls. In support of a stimulatory role for Stat proteins in ER signaling, others have demonstrated that 17ß-estradiol induces rapid activation of Stat1 transcriptional activity in an ER-independent fashion (22). 17ß-Estradiol induces the phosphorylation of Stat5 and Stat3 via MAPK and Src kinase. This results in very rapid, but transient, activation of Stat5-mediated transcription in fibroblast cells (18)
Longer than 6 h of exposure to estrogen prevented Stat1 and Stat5 from binding to the TRR in EECs. ER signaling suppresses Stat5-mediated transcription in fibroblast cells (17), probably via a direct interaction between Stat5 and the ER DNA binding domain (19). Similarly, in these studies, 9–14 d of estrogen treatment resulted in T-bet protein levels about half those in untreated cells. These data suggest that Stat-dependent transcriptional activity may be regulated differently depending on the duration of ER stimulation.
A variety of studies have demonstrated that IL-15 enhances the phosphorylation of Stat5 and its binding to GAS sites in lymphoid cells (26, 31, 35, 36, 37). IL-15 signaling induces phosphorylation of JAK2 and causes phosphorylation of Stat5 in some cell lines. In mast cells, IL-15 signaling involves JAK2/Stat5 activation rather than the JAK1/JAK3 and Stat5/Stat3 systems (31). Additionally, IL-15 has been shown to up-regulate the transcription of T-bet in lymphoid cells (27). Although we demonstrated here that IL-15 promoted T-bet expression in ECC, it did not alter the binding of Stat5 to the TRR. Instead, phosphorylation of JAK2, Stat1, and Stat5 was observed in EECs upon 6 h exposure to IL-15. This differs from the effects of exposure to female steroid hormones observed in these studies. Our data indicate that female steroid hormones and IL-15 regulate T-bet through Stat1 and Stat5 but use distinct phosphorylation mechanisms.
In lymphoid and dendritic cells, T-bet-induced IFN- production and IFN--induced T-bet expression form a positive feedback loop, resulting in rapid amplification of T-bet expression (13, 38). We have demonstrated here that, whereas EECs did respond to IFN- by activating T-bet transcription, EECs did not produce IFN-.
IL-15 can be produced by a diverse array of cells, including epithelial cells, and it plays a crucial role in CD8+ T, NK, and NKT cell maturation (and in adaptive immunity, including TH1-lineage development among naive T cells (31, 39). In EECs, T-bet induced IL-15 expression and IL-15, in turn, induced T-bet expression. This suggests a possible positive feedback loop such as regulates T-bet expression in IFN-responsive cells.
IL-15 also plays a key role in the survival, differentiation, and activation of NK cells, including uNK cells (reviewed in Ref. 31). uNK cells are present in the human endometrium in very large numbers after mid-to-late secretory phase (4, 40). IL-15, through its effects on uNK cells, is necessary for normal decidual formation (4, 41, 42). In EECs treated with estrogen and progesterone to mimic the hormonal milieu during a normal human menstrual cycle, the expression of T-bet and IL-15 was elevated for the brief period of time corresponding to mid-to-late secretory phase, a time when adequate decidualization is essential. Stat5 and JAK2 expression levels in endometrial glandular epithelia are also highest during the secretory phase, whereas expression of Stat1 is constant throughout the normal menstrual cycle (23). All of these data are consistent with an important role for T-bet in the actions of female steroid hormones on endometrial physiology.
We propose that female steroid hormones induce T-bet in EECs during the mid-to-late secretory phase mainly via activation of Stat5. This results in an induction of IL-15 secretion from EECs. In turn, IL-15 secreted by EECs activates uNK cells, thereby supporting the decidua. We have observed that T-bet expression declines and remains suppressed with prolonged exposure to both estrogen and progesterone. After late secretory phase in vivo, suppression of T-bet expression may protect the embryo from maternal TH1 or innate immune responses.
Taken together, these data support the existence of a positive feedback loop in which IL-15 acts upon EECs in an autocrine fashion to induce T-bet expression. T-bet in turn promotes IL-15 secretion by these cells. In the endometrium, where intraepithelial and submucosal lymphocytes (T cells, NKT cells, and NK cells) are present, IL-15 produced by positive feedback in EECs may promote T helper cell differentiation and NK and NKT cell maturation. Because EECs do not produce IFN-, this feedback loop may subserve the same function as the IFN-/T-bet feedback loop in lymphoid cells (10, 13), which is thought to lower the threshold of APC signaling and trigger T helper cell differentiation (38).
The weak expression of T-bet observed in the human cervix may protect sperm residing within the reservoir of the cervix from the development of a TH1-type inflammatory environment. In contrast, T-bet was expressed strongly in vaginal epithelia where a strong inflammatory response may protect against pathogen exposure. The different levels of T-bet observed in each of these tissues may reflect their distinct roles in reproduction.
In summary, we have demonstrated that T-bet is expressed in EECs under the regulation of female steroid hormones. This expression involves binding of phosphorylated Stat protein to the TRR where they lead to T-bet expression. T-bet, in turn, plays a role in the induction of IL-15 expression. This forms a positive feedback loop between T-bet and IL-15. The importance of the novel pathway shown here for the human endometrium may be 2-fold. The female steroid hormone/Stat1 and Stat5/T-bet/IL-15 pathway may play a key role in promoting an immune microenvironment favoring implantation through uNK cell-regulated endometrial decidualization. In conjunction, the IL-15/Stat1 and Stat5/T-bet/IL-15 pathway may promote the first-line innate and/or adaptive immune responses to pathogenic invaders that are essential to the protection of the epithelia of the reproductive tract. It is certain that T-bet is a central to both pathways. However, unlike the T-bet pathway in lymphoid cells, which is regulated by Stat1/4 and IFN-, Stat1/5 and IL-15 are exclusively involved in both pathways in EECs. This represents a novel molecular mechanism by which female steroid hormones alter the immunological microenvironment of the endometrium.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Cell Culture and Cytokine/Hormone Exposure
Cervical and vaginal epithelial cells were established by immortalization of the primary cultured cells using oncogenic E6 and E7 genes of human papillomavirus type 16 (29, 43). Immortalized epithelial cells derived from endocervix and vagina were cultured in serum-free media supplemented with bovine pituitary extract, recombinant epidermal growth factor and calcium chloride (Keratinocyte-serum-free medium; Invitrogen Corp., Carlsbad, CA). Ishikawa cells are a stable cell line derived from a well-differentiated human endometrial adenocarcinoma (EECs). EECs were cultured in RPMI 1640 (Invitrogen) supplemented with 10% fetal calf serum and penicillin/streptomycin (Invitrogen). All cells were exposed to IFN- (10 ng/ml) (Sigma-Aldrich Inc., St. Louis, MO), IL-4 (10 ng/ml) (R&D Systems, Inc., Minneapolis, MN), IL-15 (5 ng/ml) (R&D Systems, Inc.), estrogen (10–8 M), or progesterone (10–6 M) (Sigma-Aldrich Inc.) for 6–9 h. For long-term exposures, EECs were exposed continuously to either progesterone (10–6 M) or estrogen (10–8 M) for 2, 4, 6, 8, and 15 d, with passage every 3 d. In some experiments, EECs were exposed to progesterone at several physiological concentrations (10–5, 10–6, or 10–7 M) for 2 d. To mimic normal human menstrual cycle hormonal exposures, EECs were first primed by estrogen exposure (10–8 M) for 2 wk and then exposed to both estrogen (10–8 M) and progesterone (10–6 M) for an additional 15 d. These cells were harvested at d 2, 6, 8, 12, and 14 of estrogen exposure and d 2, 4, 8, 12, and 14 of combined estrogen and progesterone exposure
T-bet Transfection into EEC
The murine T-bet gene (1605 bp) was subcloned into pCMV-tag2B (Stratagene, La Jolla, CA) downstream of the cytomegalovirus immediate early promoter using EcoRI-XhoI sites. The resulting construct drives the expression of T-bet with an N-terminal FLAG-tag fusion (a kind gift from Dr. Eun Sook Wang, Harvard School of Public Health, Boston, MA). Either 15 µg of pCMV/T-bet or empty pCMV vector plasmid was transiently transfected into EECs (3 x 106 cells) using Lipofectamine (Invitrogen). The transfected EECs were incubated for 48 h after transfection and harvested for RNA extraction.
RT-PCR and Real-Time PCR
Cytoplasmic RNA extracted from epithelial cells. Two micrograms of RNA were subjected to reverse transcription (RT) to generate cDNA. cDNA was then amplified by PCR using primer pairs set in the human T-bet gene (406 bp) (forward, 5'-GGGCGTCCAACAATGTGACCC-3'; reverse, 5'-CCTGGGGAACCACATCCTTCG-3') and in the human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene (889 bp) (forward, 5'-GGAAGGTGAAGGTCGGAGTC-3'; reverse, 5'-AAGGTGGAGGAGTGGGTGTC-3'). Equivalent RNA samples were amplified by PCR without RT as a negative control. To assess cytokine expression, cDNA was produced via RT of 1 µg of total RNA and amplified by PCR for human IFN-, IL-8, -12, -15, -18, IL12Rß2, and GAPDH. PCR for IFN-, IL-8 and IL-15 were performed using primer pairs that amplified products of 424, 227, and 342 bp, respectively (Biosource International Inc., Camarillo, CA). The following primer pairs were synthesized to detect the IL-12, -18, and IL12Rß2 genes: IL-12 (460 bp): forward, 5'-GGGTCAGCCTCCCAGCCA-3'; reverse, 5'-GGCACAGGGCCATCATA-3'; IL-18 (308 bp): forward, 5'-GGCTGCTGAACCAGTAG-3'; reverse, 5'-GCCATACCTCTAGGCTG-3'; and IL12Rß2 (546 bp): forward, 5'-AGGATGCAGCTCCAT-3'; reverse, 5'-GGGCAGCTGTGTCTTCT-3'. GAPDH was used as an internal control for cDNA synthesis. For semiquantitative analyses, PCR products were separated over a 1.5% agarose gel and stained with ethidium bromide. Ethidium-stained bands were analyzed using image analysis software (Scion Image, Frederick, MD) and normalized to GAPDH.
For quantitative analyses, mRNA were subjected to real-time RT-PCR using 1 µg of cytoplasmic RNA. Human T-bet, IL-15, and ß-actin expression levels were assessed using SYBR Green (QIAGEN, Inc., Valencia, CA) methodology. In these experiments, the following primer pairs amplified a single PCR product for each gene: T-bet (67 bp): forward, 5'-AACCGCCTGTACGTCCACC-3'; reverse, 5'-ATGAAACTTCCTGGCGCATC-3'; IL-15 (106 bp): forward, 5'-TTCACTTGAGTCCGGAGATGC-3'; reverse, 5'-GCATCCAGATTCTGTTACATTCCC-3'; ß-actin (113 bp): forward, 5'-GAAATCGTGCGTGACATTAAGG-3'; reverse, 5'-TCAGGCAGCTCGTAGCTTCT-3'. The mRNA levels of T-bet were normalized to ß-actin.
Western Immunoblotting
Fifty micrograms of the nuclear extracts from epithelial and control cells were harvested and used for Western immunoblotting. Nuclear extracts were prepared according to the methods described in Ref. 16 . The 4B10 MAb (Santa Cruz Biotechnology Inc.) (1:1000) was used for Western immunoblotting. A T-bet-specific band was observed at 62 kDa on an 8% polyacrylamide gel. Nuclear extracts from T-bet-knockout mouse T cells transduced with the murine T-bet gene (kind gift from Dr. L. H. Glimcher, Harvard School of Public Health) were used as a positive control. As loading controls for nuclear extracts, nuclear ß-actin proteins were detected by Western immunoblotting using a rabbit anti-ß-actin polyclonal antibody (Abcam Inc., Cambridge, MA) (1:1000).
For experiments involving the detection of phosphorylated JAK2, EECs were exposed to IFN- (10 ng/ml), IL-15 (5 ng/ml), estrogen (10–8 M), or progesterone (10–6 M) (Sigma-Aldrich Inc.) for 6 h and then lysed in 5 packed-cell volumes of lysis buffer (1% Nonidet P-40, 10% glycerol, 10 mM HEPES, 150 mM KCl, 2 mM EDTA, 1 mM EGTA, 0.5 mM dithiothreitol, 1 mM Na3VO4 and protease inhibitors (Sigma-Aldrich Inc.). The total protein concentrations of cell lysates were measured using standard protein assay kits (Pierce Biotechnology, Rockford, IL). Fifty micrograms of cellular proteins were separated on a 7.5% polyacrylamide gel and immunoblotted with an antiphosphorylated JAK2 (pYpY1007/1008) antibody (Biosource International, Inc.) (1:500). The molecular weight of the band corresponding to phosphorylated JAK2 was confirmed by comparison to standard size markers and molecular weight analysis (FluorChem SP; Alpha Innotech, San Leandro, CA).
Luciferase Assays
The TRR containing bases –2906 to +138 of the T-bet gene (kind gift from Dr. Stanford Peng, Washington University School of Medicine, St. Louis, MO) was subcloned into the pGL-basic vector (Promega Corp., Madison, WI) adjusting the reading frame of the T-bet gene to that of the luciferase gene (T-bet-luc). The pGL-basic vector lacking a promoter (Promega) served as a negative control (Vector). These vectors were transfected into 2x105 epithelial cells using Lipofectamine. Transfected cells were harvested 24 h after transfection. Differences in transfection efficiency were accounted for by normalizing the relative luminescence units (RLUs) obtained from cells transfected with T-bet-luc to the RLUs obtained from the corresponding cell line transfected with a luciferase vector containing a standard simian virus 40 (SV40) promoter (Promega). Cells transfected with T-bet-luc were exposed to IFN- (10 ng/ml), IL-4 (10 ng/ml), estrogen (10–8 M), or progesterone (10–6 M) (Sigma-Aldrich Inc.) for 9 h and harvested for luciferase activity assessment (Promega). A mutation was induced into the TRR of the T-bet-luc vector (mutT-bet-luc) by amplifying a 260-bp fragment within the TRR using the primers, 5'-CAGACCCCGGGGATGCTTTTTATTaCAAActAA-3' and 5'-GGTAGGGGCTGTGTACCT-3' (shown by small letters in Fig. 7A). The DNA sequence of the amplified fragment was confirmed using standard sequencing. This fragment was cloned into T-bet-luc vector using SmaI-AgeI sites. Luciferase reporter vectors with wild-type TRR or mutTRR were transfected into EECs using Lipofectamine. Transfection efficiency for each vector was compared by amplifying the luciferase gene included in the cell-associated DNA extracted from deoxyribonuclease I-treated transfectants. Transfectants were exposed to IFN- (10 ng/ml), IL-15 (5 ng/ml) (R&D Systems, Inc.), estrogen (10–8 M), or progesterone (10–6 M) (Sigma-Aldrich Inc.) for 9 h before assessment in the luciferase assays.
Oligonucleotide Precipitation
A putative binding site for Stat family members was identified in the TRR by the presence of a GAS element, TTC(N)4GAA, at base pair –1200 from the T-bet gene. Three nucleotides in the core motif of this GAS element were replaced with alternative nucleotides [shown by small letters as a mutTRR in Fig. 6A]. Biotinylated 50-oligomer oligonucleotides with sequences derived from the TRR (5'-biotin-TTGAACTATATCCCAGACCCCGGGGATGCTTTTTATTTCAAAAGAAAACT-3') or mutTRR (5'-biotin-TTGAACTATATCCCAGACCCCGGGGATGCTTTTTATTaCAAActAAAACT-3') were synthesized (Invitrogen) and streptavidin bead-conjugated (Pierce Biotechnology) for precipitation. Total cell extracts were prepared in lysis buffer with 2 mM EDTA, 1 mM EGTA, 0.5 mM dithiothreitol, 1 mM Na3VO4 and protease inhibitors (Sigma-Aldrich Inc.) (26). Total protein concentrations in these cell lysates were measured using standard protein assay kits. Conjugated beads carrying 25 pmol of oligonucleotides with the sequences of either the TRR, the mutTRR, or the IFN--promoter (positive control) were incubated with 100 µg of total cell extracts for 4 h at 4 C in binding buffer (26). Precipitants were boiled in sample buffer for 5 min and separated over a 7.5% polyacrylamide gel. Anti-Stat1, 3, 4, or 5 (Santa Cruz Biotechnology Inc.), antiphosphorylated Stat1 (pY 701) (Biosource International, Inc.) or antiphosphorylated Stat5 (pY 694) antibodies (Santa Cruz Biotechnology Inc.) were used at 1:500 working concentration for Western immunoblotting of precipitants. One hundred micrograms of total cell extracts from EECs exposed to estrogen (10–8 M) or progesterone (10–6 M) for 0, 3, 6, or 9 h were used for oligonucleotide precipitation followed by Western analysis for Stat1 or 5. ß-Actin proteins were detected by Western immunoblotting using a rabbit anti-ß-actin polyclonal antibody (Abcam Inc.) (1:1000) and served as an internal control. The molecular weight of each band corresponding to a Stat protein was confirmed by comparison to standard molecular size markers using molecular weight analysis (FluorChem SP).
Statistical Analysis
Luminescence and cytokine PCR data are presented as means ± SD. Both types of experiments were performed independently four to six times. RLUs obtained from cytokine- or hormone-treated T-bet-luc-transfected cells were compared with those without treatment using paired, two-tailed Student’s t tests. GAPDH-normalized band intensities obtained from T-bet expression vector-transfected EECs were compared with those from cells transfected with empty vectors. A P value less than 0.05 was considered significant.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
First Published Online April 28, 2005
Abbreviations: APCs, Antigen-presenting cells; EEC, endometrial epithelial cells; ER, estrogen receptor; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GAS, -activated sequence; IFN, interferon; IL12Rß2, IL-12 receptor-ß2; JAK, Janus kinase; mutTRR, mutated TRR; NK, natural killer; PIAS, protein inhibitor of activated signal transducer and activator of transcription; PR, progesterone receptor; RLU, relative luminescence unit; RT, reverse transcription; Stat, signal transducers and activators of transcription; SV40, simian virus 40; TRR, T-bet regulatory region; uNK, uterine NK.
Received for publication December 6, 2004. Accepted for publication April 18, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
in lymphoid and myeloid cells. Proc Natl Acad Sci USA 98:15137–15142 and progesterone signaling pathways in human endometrium: role of PIASy (protein inhibitor of activated signal transducer and activator of transcription-y). Mol Endocrinol 18:1988–1999-hydroxysteroid dehydrogenase. Mol Endocrinol 19:431–440 production in human NK and T cells. J Immunol 170:5464–54691 integrin. J Steroid Biochem Mol Biol 59:31–39[CrossRef][Medline]
NURSA Molecule Pages Link:
This article has been cited by other articles:
 |
|
 |
|
  K. Kawana, J. Matsumoto, S. Miura, L. Shen, Y. Kawana, T. Nagamatsu, T. Yasugi, T. Fujii, H. Yang, A. J. Quayle, et al. Expression of CD1d and Ligand-Induced Cytokine Production Are Tissue Specific in Mucosal Epithelia of the Human Lower Reproductive Tract Infect. Immun., July 1, 2008; 76(7): 3011 - 3018. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Inman, K. Kawana, D. Schust, R. Lininger, and S. Young Cyclic Regulation of T-Bet and GATA-3 in Human Endometrium Reproductive Sciences, January 1, 2008; 15(1): 83 - 90. [Abstract] [PDF]
|
|
|
|
 |
|
 M. Bhanoori, M. Deenadayal, S. Kennedy, and S. Shivaji The G2964A 3'-untranslated region polymorphism of the signal transducer and activator of transcription 6 gene is associated with endometriosis in South Indian women Hum. Reprod., April 1, 2007; 22(4): 1026 - 1030. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J. Jasper, K. P. Tremellen, and S. A. Robertson Primary unexplained infertility is associated with reduced expression of the T-regulatory cell transcription factor Foxp3 in endometrial tissue Mol. Hum. Reprod., May 1, 2006; 12(5): 301 - 308. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
