Estrogen has diverse effects on inflammation and immune responses. That pregnancy is associated with remission of some autoimmune diseases and exacerbation of others suggests that physiological fluctuation in estrogen levels could affect the immune responses in humans. However, the molecular basis for these phenomena is poorly understood. We hypothesized that fluctuations of estrogen levels modulate intracellular signaling for immune responses via estrogen receptors (ERs). In reporter assays, 17β-estradiol (E2) at a physiologically high concentration increased the activity of NF-κB in Jurkat cells stimulated by PMA/ionomycin or TNF-α. Overexpression and RNA interference experiments suggested that the effects were mediated through ERβ. Immunoprecipitation assay showed that both ERα and ERβ are directly associated with NF-κB in the cell nucleus. Using chromatin immunoprecipitation assay, we confirmed that ERα and ERβ associated with NF-κB and steroid hormone coactivators at the promoter region of NF-κB regulated gene. Considering that NF-κB regulates the expression of various genes essential for cell growth and death, estrogen could regulate the fate of T cells by affecting the activity of NF-κB. To determine whether E2 alters the fate of T cells, we investigated E2 actions on T cell apoptosis, a well-known NF-κB-mediated phenomenon. E2 increased apoptosis of Jurkat cells and decreased that of human peripheral blood T cells. Our results indicate that E2 at a physiologically high concentration modulates NF-κB signaling in human T cells via ERβ and affects T cell survival, suggesting that these actions may underlie the gender differences in autoimmune diseases.
- sex hormone
- autoimmune disease
- transcription factor
many autoimmune diseases are more prevalent in women than in men (42, 44). Some autoimmune diseases, such as systemic lupus erythematosus (SLE), frequently show exacerbation during pregnancy (42), and in this regard, patients with SLE often have high serum estrogen concentrations (27). These clinical facts led us to hypothesize that fluctuation in estrogen levels could modulate autoimmune responses.
Baseline estrogen concentrations do not differ significantly between men and women. However, the serum concentration of 17β-estradiol (E2), the most potent and predominant estrogen in human serum, increases from baseline 10 pM to 40 nM during pregnancy (45). Estrogen has diverse effects on inflammatory and immune systems (8, 33), and the physiological fluctuation in estrogen level could affect the immune responses in humans (5). However, the molecular basis for these phenomena is poorly understood.
NF-κB is a key regulator of inflammatory and immune systems (22). Recent studies have revealed the role of NF-κB signaling in T cells (41). NF-κB regulates the expression of various genes that control cell cycle and cell viability (16). On the other hand, estrogen elicits various biological effects in cells by binding to two distinct estrogen receptor (ER) isoforms, ERα and β (7), members of the nuclear receptor family of ligand-dependent transcription factors. The effects of estrogen on NF-κB signaling have been mainly studied in breast cancer cells (11, 34) and studies that examined the effects of pharmacological concentrations of estrogen reported that estrogen modulates the NF-κB signaling that interacts with ERs (13, 15, 29).
The present study was designed to determine the effects of physiological fluctuations in estrogen on NF-κB signaling and the fate of immune cells. These effects may help explain the gender differences in the immune system.
MATERIALS AND METHODS
Human T cell line Jurkat T cells (E6-1) were purchased from Dainippon Sumitomo Pharma (Tokyo, Japan). Peripheral blood CD3+ T lymphocytes (PBT cells) from healthy donors' blood were obtained as follows. Peripheral blood mononuclear cells (PBMC) were isolated by Ficoll centrifugation and washed twice, and CD3+ cells were enriched by depletion of non-T cells using the IMag T lymphocyte enrichment set (BD Biosciences, San Jose, CA). Jurkat T Cells and PBT cells were cultured in phenol red free RPMI 1640 medium (Invitrogen, Carlsbad, CA) supplemented with 10% charcoal-stripped fetal bovine serum at 37°C in a 5% CO2 incubator. For NF-κB activation, Jurkat cells were aliquoted into a 96-well plate 40 h after transfection, cultured in a final volume of 100 μl of RPMI 1640 growth medium, and then stimulated at 37°C in a growth medium containing 60 ng/ml PMA (Sigma, St. Louis, MO) and 1 μg/ml ionomycin (Sigma) or 10 ng/ml human TNF-α (Wako, Osaka, Japan) and with the indicated concentrations of E2 (water-soluble form; Sigma).
The reporter plasmids, pNF-κB-Luc (Stratagene, La Jolla, CA) and pRL-TK (Promega, Madison, WI) were purchased. pOR-3, the ERα expression vector (18), was kindly supplied by Dr. P. Chambon (Institute of Genetics and Molecular and Cellular Biology, Strasbourg, France) and pSG5-ERβ, the ERβ expression vector (31) by Dr. J. A. Gustafsson (Department of Biosciences, Karolinska Institute, Huddinge, Sweden). Human p65 (NF-κB) cDNA was amplified by a conventional RT-PCR method and cloned into pcDNA3.1 (-)-myc-His expression vector (Invitrogen). ERα and ERβ siRNA expression vectors for RNA interference were purchased from Panomics (Redwood City, CA).
Transient transfection and reporter assay.
Jurkat cells were cotransfected with pNFκB-Luc and the indicated expression plasmids using the method as explained by the manufacturer (Amaxa Biosystems, Cologne, Germany). pRL-TK, which expresses Renilla luciferase, was also cotransfected as an internal control. After 6-h stimulation with PMA/ionomycin or TNF-α, cells were lysed and assayed using Dual-Glo Luciferase Assay System (Promega). Luciferase activity was read on the luminometer (Luminoskan Ascent; Thermo Electron, Waltham, MA) and determined at least three times for each experimental condition. The fold induction of luciferase activity (firefly luciferase/Renilla luciferase) was expressed relative to the baseline activity in untreated cells.
Jurkat cells and PBT cells were treated with 10 ng/ml TNF-α and the indicated concentrations of E2 for 6 h. Cells were washed twice and lysed with radioimmunoprecipitation (RIPA) buffer (1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, pH 7.4, in PBS). The level of activated NF-κB was quantified in cell lysate by using the NF-κB-p65 ELISA Kit (Nventa Biopharmaceuticals, San Diego, CA).
Immunoprecipitation and Western blot analysis.
For immunoprecipitation assay, p65 and ERα or ERβ were overexpressed in cells. Forty hours after transfection, cells were treated for 6 h with 10 nM E2 and 10 ng/ml TNF-α. Nuclear proteins were obtained using Nuclear Complex Co-IP Kit (Active Motif, Carlsbad, CA). Using agarose-conjugated anti-p65 (Santa Cruz Biotechnology, Santa Cruz, CA), anti-ERα (Lab Vision, Fremont, CA), anti-ERβ (Santa Cruz) or control rabbit IgG antibodies, we examined the status of coprecipitation of ERα or ERβ with p65. Proteins were separated by SDS-PAGE on 10% gels and transferred onto a polyvinylidene difluoride membrane (Millipore, Bedford, MA). The membrane was blocked overnight with 5% skim milk, followed by incubation with anti-ERα (Lab Vision) or anti-ERβ antibody (Lab Vision) for 2 h at room temperature. After washing, the horseradish peroxidase-labeled anti-mouse secondary antibody (Chemicon International, Temecula, CA) was added for 2 h. The proteins were visualized with the ECL substrate kit (Amersham Life Science, Buckinghamshire, UK).
Chromatin Immunoprecipitation assays.
Jurkat cells were stimulated with 10 nM E2 and 10 ng/ml TNF-α for 6 h. DNA was purified with chromatin immunoprecipitation (ChIP)-IT Kit (Active Motif). Protein/DNA complexes were immunoprecipitated with control mouse IgG (included in the ChIP-IT Kit), anti-ERα antibody (Lab Vision), anti-ERβ antibody, anti-p65 antibody, anti-SRC1 antibody, and anti-GRIP1 antibody (Santa Cruz). The primers used for the IL-4 promoter were (-336)-5′-CAACAAATTCGGACACCTG-3′ and (-40)-5′-GCTGAAACCGAGGGAAAAT-3′ (a 296-bp product) (23).
Jurkat cells and PBT cells were pretreated with or without 10 nM E2 or 10 mM ICI 182780 (Tocris Cookson, Ellisville, MO) for 6 h. Jurkat cells were then plated in anti-human CD3 plate (BD Biosciences) for 36 h to induce the activation-induced cell death (AICD) (6, 14, 25). On the other hand, PBT cells were cultured for 36 h in a medium containing 4 μg/ml Concanavalin A (Sigma) (43). Cells were analyzed with Annexin V Fluos kit (Roche, Basel, Switzerland) by a flow cytometer (FACS Aria; BD Biosciences). Cells were stained with Annexin V-Fluos and propidium iodide (PI). Apoptotic cells were defined as Annexin V-positive and PI-negative cells.
Data are expressed as means ± SD of at least three independent experiments with triplicate estimations per data point. P values were calculated by ANOVA.
E2 enhances NF-κB reporter activity in Jurkat cells.
To investigate the actions of estrogen on intracellular signaling, we screened changes in activities of various transcription factors with reporter plasmids in Jurkat cells (data not shown), and we then focused on NF-κB signaling. First, we examined the NF-κB activity in Jurkat cells with reporter plasmid. The results showed E2 alone did not affect the NF-κB signaling in these cells (Fig. 1). However, when the cells were stimulated in the presence of phorbol ester PMA and ionomycin, known as strong activators of NF-κB in T cells (24, 38), physiologically high concentrations (10 nM) of E2, representing the levels in pregnancy, enhanced the activated NF-κB signaling (Fig. 2A). After stimulation by TNF-α, another activator of NF-κB in T cells, E2 also showed a similar effect on Jurkat cells (Fig. 2B).
E2 enhances activation of NF-κB in human T cells.
The major form of NF-κB is known as a heterodimer of p65 and p50 (22). As indicated above, E2 enhanced activation of NF-κB signaling in Jurkat cells with the reporter assay system. To confirm the results, we next quantified the activated p65 in Jurkat and PBT cells with ELISA. The assay uses NF-κB consensus sequence to capture only the active forms of NF-κB, because only active forms of NF-κB can translocate from the cytoplasm to nucleus and bind the genomic DNA. In both Jurkat and PBT cells, TNF-α induced active forms of p65 were increased in the presence of physiologically high concentration of E2 (10 nM) but not with physiologically low concentration (10 pM) or absence of E2 (Fig. 3).
E2 enhances activation of NF-κB in Jurkat cells via ERβ.
Several studies have reported ER expression in immune cells (32, 37), and a recent study suggested that ERβ is predominantly expressed in human T cells (37). Our results also showed the mRNA expression of both ERα and ERβ in both Jurkat cells and human PBT cells and the predominance of ERβ (data not shown). Since the effects of estrogen are mediated mainly by binding to ERs, we transiently transfected Jurkat cells with ER expression vectors or ER siRNA vectors and determined the ER subtype that could affect the enhancement of activation of NF-κB signaling in human T cells. In these experiments, Jurkat cells were stimulated with PMA and ionomycin followed by analysis of luciferase activity. As shown in Figs. 2 and 3, 10 nM E2 enhanced activated NF-κB signaling with endogenous ERs. Under physiologically high E2 condition, overexpression of ERβ enhanced NF-κB activation (Fig. 4B), while overexpression of ERα inhibited NF-κB activation (Fig. 4A). On the other hand, the results of RNA interference experiments for endogenous ERs (Fig. 4, C and D) showed that physiologically high E2 did not enhance NF-κB activation in ERβ-knockdown cells. These results suggest that the effect of E2 on NF-κB signaling is mediated mainly through ERβ and not ERα.
NF-κB interacts with ERs and steroid hormone coactivators on the promoter region of NF-κB target gene in T cell nucleus.
To investigate the mechanisms of estrogen effects on NF-κB signaling, we examined the direct interaction between ER and NF-κB under stimulation with TNF-α and E2. After cotransfection of p65 expression vector and ERα or ER-β expression vector, nuclear extracts of cells were immunoprecipitated with p65 antibody or control antibodies and the proteins were immunoblotted with anti-ERα or ERβ antibody (Fig. 5, A and B). Activated p65 of NF-κB interacted with both ERα and ERβ in cell nucleus. To further investigate the association between ERs and NF-κB, we conducted ChIP assays in Jurkat cells. ChIP assays are used to identify the in vivo position of transcription factors on cognate DNA sites (19). We examined the binding of NF-κB/ER complex on IL-4 promoter, a NF-κB regulated gene (23), under TNF-α- and E2-stimulated conditions. Both endogenous ERα and ERβ bound to the IL-4 promoter in vivo (Fig. 5C, lanes 3 and 4). We also examined steroid hormone coactivators, steroid receptor coactivator1 (SRC1/NCOA1) and glucocorticoid receptor interacting protein1 (GRIP1/NCOA2) (9, 39). Since both coactivators are known to interact with nuclear receptors (26), we confirmed that there are no steroid hormone response elements (including estrogen response elements) on the IL-4 promoter region with MATCH software public version 1.0 (46). SRC1 and GRIP1 were precipitated concurrently with NF-κB and ERs on IL-4 promoter (Fig. 5C, lanes 5 and 6), suggesting that they form a large transcriptional complex on the promoter region.
E2 modulates apoptosis of T cells.
To examine whether estrogen affects the T cell fate through NF-κB, we analyzed AICD in Jurkat cells, which is mediated by NF-κB (25). E2 is known to induce apoptosis in Jurkat cells (12). Treatment with a physiologically high concentration of E2 (10 nM) increased the proportion of apoptotic Jurkat cells (Fig. 6A). On the other hand, treatment of PBT cells with the same concentration of E2 rescued these cells from apoptosis (Fig. 6B), in agreement with a previous study that showed inhibition of NF-κB activity resulted in increased PBT cell apoptosis through concanavalin A stimulation (43). To investigate whether these effects are mediated via ERs, we next examined effects of ICI-182780, an ER antagonist (1). Inhibition of endogenous ER action with ICI-182780 decreased apoptosis by 10% in Jurkat and increased by 31% in PBT cells. These results indicate that the estrogen receptor signaling affects the NF-κB signaling in human T cells and consequently affects the T cell fate.
Our study demonstrated that E2 enhances NF-κB activity in human T cells. E2 modulated NF-κB signaling in both Jurkat cells and human PBT cells when stimulated. In T cells, E2 seems to act through ERβ, and the direct interaction among ERs, NF-κB, and steroid hormone coactivators could modulate the NF-κB signaling.
To our knowledge, this is the first report on the functional relationship between estrogen and NF-κB in human T cells and on the association among ERs, NF-κB, and steroid hormone coactivators. Our results suggest that the effect of E2 is mediated via ERβ in T cells. Several studies have examined estrogen actions on NF-κB activity, mainly in breast cancer cells (11, 34, 36). However, the results of these studies are controversial. Liganded ERα is reported to inhibit NF-κB activity in an estrogen-dependent manner at nanomolar concentrations of estrogen in various cell lines, such as U937 and MCF-7 cells, and ERβ is also known to have an inhibitory effect on NF-κB activity (15). On the other hand, no antagonism between ER and NF-κB was reported in a study of other breast cancer cells, T47D and HC11 cells treated with TNF-α and E2 (36). Furthermore, another study showed that E2 enhanced NF-κB transactivation via ERβ in human embryonic kidney HEK-293 cells and prostate cancer DU145 cells (21). These results suggest that estrogen actions on NF-κB activity may depend on ER and cell subtypes. Although there are also many controversial results about expression of ER genes in immune cells (32, 37), the latest report indicates that ERβ is the predominant ER in human peripheral blood leucocytes (37). Our results are consistent with that study, i.e., estrogen could affect the NF-κB activity in human T cells via ERβ.
Our results showed E2 alone did not change NF-κB signaling in Jurkat cells and human PBT cells, however; it enhanced the NF-κB signaling when activated. Though the effects between two different concentrations of E2 were modest, our report focused on the physiological fluctuations and that slight changes in physiological circumstances could affect or trigger subsequent events.
In addition, we obtained similar results in experiments of both PMA/ionomycin-activated and TNF-α-activated NF-κB signaling. In the NF-κB activation pathway, it was reported that PMA and ionomycin act on protein kinase C (30), and TNF-α acts on caspase-8 (28). Therefore, we suggest that E2 modulates the NF-κB signaling downstream of these points, i.e., at the nucleus. Using immunoprecipitation assay, we demonstrated a direct interaction between ERs and p65 in cell nucleus. With ChIP assays, liganded ERα and ERβ associated with NF-κB on the NF-κB regulated gene promoter. Moreover, steroid coactivators SRC1 and GRIP1 were confirmed to associate with NF-κB under physiologically high E2 conditions. These results suggest that the E2-liganded ERs associate with steroid coactivators and simultaneously interact with NF-κB on the NF-κB regulated gene promoters, and that recruitment of steroid hormone coactivators could affect the transcriptional activity of NF-κB (Fig. 7).
Physiologically high concentrations of E2, equivalent to those in pregnancy, altered apoptosis of human T cells. Using Jurkat T cells and PBT cells, we showed here that E2 increased apoptosis of Jurkat cells but rescued PBT cells from apoptosis. NF-κB has been reported as a contextual regulator of apoptosis (24). It was reported also that NF-κB contributed to AICD by upregulation of Fas ligand expression in Jurkat T cells (17), whereas NF-κB in PBT cells is required for T cell survival (43). We believe that distinct signaling pathways mediate T cell apoptosis under different contexts in these two types of T cells.
The significance of estrogen in human T cells is still not clear. There are several reports on the contribution of the Rel/NF-κB family on the development of Th1 and Th2 lymphocytes (4, 10). Pregnancy is associated with exacerbation of SLE, a Th2 disease, but with remission of multiple sclerosis (MS) and rheumatoid arthritis, known as Th1 diseases (20, 44). Previous reports suggested that E2 reduces pathological Th1 responses (20, 33). Moreover, several studies reported that E2 administration accelerated disease activities in murine models of SLE (3, 35, 40), but inhibited experimental autoimmune encephalomyelitis, an animal model of MS (2). Autoreactive cells that escape immunosurveillance can result in many autoimmune diseases. In the present study, we demonstrated that E2 enhanced activation of NF-κB, a key regulator of cytokine production. Although further studies are needed to address the underlying mechanisms of gender differences in the immune system, these results could help explain inappropriate homeostasis of the immune system.
This work was supported by a grant from the Osaka Medical Research Foundation for Incurable Diseases and the Japan Society for the Promotion of Science (to D. Furutama).
The authors thank Masa-aki Shibata from the Department of Anatomy and Cell Biology, Osaka Medical College, for his help on Jurkat transfection. We acknowledge the helpful advice of Shigeki Makino.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society