Estrogen Receptor α Regulates Tripartite Motif–Containing Protein 21 Expression, Contributing to Dysregulated Cytokine Production in Systemic Lupus Erythematosus

To examine the role of 17β‐estradiol in the regulation of the autoantigen tripartite motif–containing protein 21 (TRIM‐21) in patients with systemic lupus erythematosus (SLE).

Although increased levels of estrogen in SLE patients have not been demonstrated (5), elevated levels of estrogen during pregnancy have been associated with disease exacerbation and an increased incidence of flares (6). Studies in murine lupus models have shown accelerated glomerulonephritis, autoantibody formation, and overall mortality following the administration of estrogen, supporting a role of this hormone in driving disease pathogenesis (7).
Studies investigating the use of hormone-based contraceptives or hormone replacement therapy in menopausal patients, however, have described conflicting findings (8)(9)(10). For example, two large trials published concurrently in 2005 demonstrated no link between administration of exogenous hormones and the development of SLE (8,9), whereas an earlier study showed that the use of oral contraceptives by SLE patients induced flares (5). Despite these findings, links between the estrogen system and the development of disease or the occurrence of flares in SLE remains strong. More recently, the estrogen receptor (ER), a nuclear hormone receptor, has come to the fore as the potential link between estrogen and disease severity (11).
The primary effects of estrogen are mediated via activation of the intracellular receptors ER␣ and ER␤. These receptors exert their regulatory effect on gene expression via 2 classic mechanisms: by directly binding to estrogen response elements (EREs) located within the promoter of target genes or by indirectly interacting with other transcription factors, such as activator protein 1 (AP-1) (12) or NF-B (13), resulting in modulation of their binding to defined consensus sequences within the promoter regions. Interestingly, autoantibodies against ERs have been described in patients with SLE and were reported to interfere with T cell homeostasis and to correlate with disease activity (14). Although the reports vary, it also appears that ER␣ and ER␤ levels are altered between different immune cells and between SLE patients and healthy controls, with enhanced levels of ER␣ expression observed in peripheral blood mononuclear cells (PBMCs) from SLE patients (15).
Further underlining the importance of ER␣ in SLE is a study showing that lupus-prone NZM2410 mice with a deletion of ER␣ have a better prognosis than their ER␣ ϩ/ϩ counterparts, with prolonged survival, less severe renal involvement, and lower proteinuria levels (11). Importantly, immune complex deposition and autoantibody production were not altered in these mice, suggesting that ER␣ plays a role in the earlier innate response. Interestingly, a number of ER␣-regulated genes that are associated with SLE have since been identified, most noteworthy being interferon regulatory factor 5 (IRF-5), a transcription factor that is genetically and functionally associated with SLE (16), BAFF, a B cell growth-promoting cytokine (17), and IFI202, a murine lupus susceptibility gene (18). Proteins known to play a role in T cell function in SLE are also ER␣ dependent, for example, calreticulin and CREM␣ (19,20). Estrogen has also been shown to enhance Toll-like receptor 9 (TLR-9)-driven cytokine production and cell activation from B cells and dendritic cells, potentially implicating ER␣ as a key regulator of gene expression in innate immunity (21,22).
Tripartite motif-containing protein 21 (TRIM-21) was first described as an autoantigen in SLE and has subsequently been shown to regulate antiviral responses and interferon-␤ (IFN␤) production downstream of antiviral pattern-recognition receptors such as TLRs 3, 7, and 9 (23)(24)(25). With regard to its role in IFN␤ regulation, TRIM-21 was originally described as a negative regulator of IFN production through its ability to target IRF-3 and IRF-7 for ubiquitination and subsequent degradation downstream of TLR signaling (23). Consistent with this, TRIM-21 knockout mice develop a lupus-like condition following ear-tagging as a result of the overproduction of IFN␣ and interleukin-23p19 (IL-23p19) downstream of the DNA receptors TLR-3 and TLR-9 (24,25), demonstrating that TRIM-21 plays an important role as a negative regulator of IFN and IL-23 production downstream of the antiviral TLRs.
Given the role of TLR-7 and TLR-9 in driving the production of autoantibodies and, consequently, the deregulated cytokine profile associated with SLE, a role of TRIM-21 in protecting against autoimmunity has been suggested. In addition to its protective role, TRIM-21 has also been shown to stabilize IRF levels in resting cells. Specifically, Kim et al demonstrated that in resting cells, IRF-8 is stabilized by TRIM-21, resulting in enhanced IL-12p40 production, whereas upon stimulation, the p62 sequestosome is recruited to the TRIM-21/ IRF-8 complex, leading to the polyubiquitination and destabilization of IRF-8 (26). We have also observed that genetic ablation of TRIM-21 (either via short hairpin RNA [23] or by genetic deletion [24]) results in reduced levels of IRF-3 in resting cells, a phenotype that is reversed upon TLR stimulation of the cells (24). Thus, TRIM-21 plays a dual role in regulating IRF stability, a role that is very much dependent on context or stimulation.
Despite numerous studies into the function of this autoantigen, its regulation at a genetic level is poorly 164 SMITH ET AL understood. Interestingly, enhanced levels of TRIM-21 transcripts have been observed in the PBMCs of patients with SLE as compared with healthy controls (27), thus implicating a role for this antigen in the pathogenesis of this condition. The identification of a potential ERE in the human TRIM-21 promoter led us to hypothesize a role for estrogen in the regulation of TRIM-21 expression. Our results demonstrated that the enhanced levels of ER␣ seen in SLE patients correlate with the increased expression of TRIM-21 observed in these patients. We also demonstrated that 17␤-estradiol upregulates TRIM-21 expression in human monocytes, both at the messenger RNA (mRNA) level and the protein level, through an ER␣-dependent mechanism, a pathway that is hyperactive in SLE patients. Importantly, modulation of ER␣ signaling through the use of the selective ER␣ antagonist methyl-piperidino-pyrazole dihydrochloride (MPP) abrogated estrogen-induced TRIM-21 expression, leading to the degradation of IRF-3 and thus limiting excessive production of IL-23 and IFN␤. Overall, our results suggest that targeting estrogen and its signaling pathways could hold therapeutic potential in the treatment of SLE.

MATERIALS AND METHODS
Cell culture. PBMCs were isolated from whole blood of patients with SLE and healthy control subjects with the use of a Ficoll gradient and were cultured in phenol red-free RPMI 1640 medium supplemented with 10% charcoalstripped fetal calf serum and 100 g/ml of penicillin/ streptomycin. Monocytes were extracted from the PBMCs by positive selection using CD14ϩ beads (Miltenyi Biotec). Cells were allowed to recover overnight prior to stimulation. Sterile ethanol-soluble 17␤-estradiol solution (Sigma) was prepared fresh for each experiment and was used at a final concentration of 10 -7 M. Sterile ethanol-soluble MPP (Sigma) was also prepared fresh for each experiment and was used at a final concentration of 10 -6 M. Study subjects. All patients with SLE who were included in our study met the American College of Rheumatology classification criteria (28)(29)(30) and were recruited from Beaumont Hospital, Dublin, Belfast City Hospital, Belfast, St. James Hospital, Dublin, and St. Vincent's University Hospital, Dublin. SLE patients included in the study were female and 22-62 years of age. Age-and sex-matched healthy individuals were chosen as controls. The study protocol was approved by the institutional review boards of all involved institutions, and written informed consent was obtained from all participants.
Real-time quantitative polymerase chain reaction (qPCR). RNA was extracted from cell cultures using TRIzol reagent (Sigma) and reverse transcribed to complementary DNA using Omniscript reverse transcriptase (Qiagen) according to the manufacturer's recommendations. Real-time qPCR investigating TRIM-21, CREM␣, and ER␣ gene expression was performed using appropriate primers (Table 1) with SYBR Green Taq ReadyMix (Sigma) according to the manufacturer's recommendations. Data were analyzed using an ABI Prism 7900 system (Applied Biosystems) and were normalized to an 18S RNA reference. Real-time PCR data were analyzed using the 2 -⌬⌬Ct method (31).
Chromatin immunoprecipitation (ChIP) analysis. ChIP assays were performed using a commercial ChIP assay kit (Active Motif), according to the manufacturer's instructions. Briefly, cells (1 ϫ 10 7 in total) were crosslinked for 10 minutes at room temperature with 1% formaldehyde. Nuclei were isolated into 1% sodium dodecyl sulfate (SDS) buffer containing protease inhibitors and subjected to sonication to yield 200-500-bp DNA fragments. Sonicated chromatin was immunoprecipitated overnight at 4°C with 1 g of anti-ER␣ (Santa Cruz Biotechnology) or control rabbit IgG (Santa Cruz Biotechnology). Samples were washed, and then bound chromatin was eluted and incubated overnight at 65°C for reversal of crosslinking. After proteinase K digestion for 1 hour, DNA was extracted using a DNeasy kit (Qiagen). To determine the identity of the ER␣ target genes, ChIP DNA was further analyzed by PCR using primers encompassing the regions of interest on the TRIM-21 promoter. Input DNA was used as a positive control. PCR products were analyzed on 1.5% agarose gels by electrophoresis. The band intensities of products were analyzed using GeneTools (Syngene). IgG control values were subtracted, and the sample values were normalized to those of Western blotting. To prepare whole cell lysates, cells were lysed in 1ϫ loading buffer containing 63 mM Tris HCl (pH 6.8), 10% glycerol, 2% SDS, and 0.0025% bromphenol blue. Equal quantities of whole cell lysates were resolved by electrophoresis on a denaturing SDS-polyacrylamide gel and transferred to a nitrocellulose membrane. Following immunoblotting, the membrane was developed using an enhanced chemiluminescent horseradish peroxidase (HRP) substrate (Millipore). The antibodies used for immunoblotting included anti-␣-actinin (Santa Cruz Biotechnology), anti-TRIM-21 (Santa Cruz Biotechnology), anti-␤-actin (Santa Cruz Biotechnology), and anti-IRF-3 (Active Motif).
Statistical analysis. Statistical analyses were performed using GraphPad Prism version 4.03 software. All data are presented as the mean Ϯ SD. Spearman's test was used for correlation analyses. P values less than or equal to 0.05 were considered significant.

Estrogen regulation of TRIM-21 expression in human monocytes via an ERE in the TRIM-21 promoter.
A link between estrogen and increased levels of circulating autoantibodies against TRIM-21 has been demonstrated in human keratinocytes, indicating a possible role of this hormone in the regulation of TRIM-21 expression (32). To test this hypothesis, human monocytes were stimulated with 17␤-estradiol (10 -7 M) or ethanol control for 6 hours and 24 hours, following which TRIM-21 expression was analyzed by real-time PCR and Western blotting. We observed significantly increased levels of TRIM-21 expression by human monocytes at both the mRNA ( Figure 1A) and protein ( Figure 1B) levels following 17␤-estradiol treatment, indicating that estrogen and, potentially, the estrogen receptor regulate TRIM-21 expression.
ER has been shown to enhance gene expression either via binding directly to a putative ERE site or by interacting with transcription factors already bound to a specific promoter (such as AP-1), recruiting the coactivator CREB binding protein and driving transcription. Using a bioinformatic approach with the MatInspector Search program, we analyzed the human TRIM-21 gene promoter for transcription factor-binding sites. Sequence analysis revealed that in addition to a variety of putative transcription factor-binding sites, such as IRF-3 and NF-B, there was a potential ERE site in the to the human TRIM-21 promoter, as determined by chromatin immunoprecipitation. PCR was carried out using primers encompassing the TRIM-21 promoter, and the products were subjected to electrophoresis on 1.5% agarose gels. IgG was used as a negative control, and input DNA as a positive control. Results are representative of 3 samples.

166
SMITH ET AL TRIM-21 promoter lying 2,000 bp upstream of the transcription start site ( Figure 1C). Having identified this potential binding site in the promoter region of the TRIM-21 gene, we hypothesized that estrogen may drive TRIM-21 expression through ER␣ promoter binding. To test this, we used ChIP assays to investigate whether the ERE identified in the promoter region of human TRIM-21 was functional. Monocytes were stimulated with 17␤-estradiol (10 -7 M) or ethanol control for 1 hour and 3 hours. ChIP analysis was performed using an ER␣-specific antibody, and PCR was carried out on the resulting immunoprecipitates using primers spanning the potential ERE site within the TRIM-21 promoter region. As shown in Figure 1D, ER␣ was not bound endogenously to the TRIM-21 promoter; however, following 3 hours of treatment with 17␤-estradiol, ER␣ binding to the promoter was observed. This suggests a classic mode of estrogenregulated gene expression through induction of the direct binding of ER␣ to an ERE site located within the TRIM-21 gene promoter.
Correlation of enhanced TRIM-21 expression with ER␣ expression in SLE patients. We next assessed whether TRIM-21 expression was altered in SLE immune cells and whether any relationship existed between TRIM-21 and ER␣ expression. Consistent with previously published reports, enhanced expression of TRIM-21 mRNA was observed in SLE PBMCs ( Figure  2A) as compared with those from healthy controls, with a concomitant increase in TRIM-21 protein levels, as expected ( Figure 2B). Given the role of TRIM-21 in regulating IRF-3 levels, we also observed a similar increase in IRF-3 protein in PBMCs from SLE patients as compared with those from healthy controls ( Figure  2B), indicating a potential relationship between TRIM-21 and IRF-3-regulated pathways. Supporting this, we have previously reported increased expression of IL-23p19 in resting monocytes from SLE patients as compared with healthy control subjects (25).
Interestingly, ER␣ expression was also enhanced in SLE patients as compared with healthy controls ( Figure 3A). We observed a strong correlation between TRIM-21 and ER␣ gene expression levels (r 2 ϭ 0.7703, P ϭ 0.0001) (Figure 3B), suggesting a direct link between ER␣ levels and enhanced TRIM-21 expression in SLE patients. Although estrogen levels reportedly do not vary significantly between SLE patients and healthy controls (33), we hypothesized that the increased ER␣ levels we observed may render PBMCs from SLE patients hyperresponsive to the effects of estrogen, potentially contributing to the enhanced TRIM-21 expression observed. To test this hypothesis, PBMCs from SLE patients and healthy control subjects were treated with 17␤-estradiol (10 -7 M) or vehicle control (100% ethanol) and TRIM-21 expression was analyzed by real-time PCR. As expected, SLE patient PBMCs responded to estrogen treatment to a much higher extent than did their healthy counterparts, thus contributing to the enhanced TRIM-21 expression observed in these patients ( Figure 3C).
We also examined the effects of estrogen stimulation on another ER␣-regulated gene, CREM␣, which in addition to its role as a potent negative regulator of T cell responses, has also been shown to regulate antigenpresentation functions as well as the costimulatory activities of antigen-presenting cells (34). Consistent with the results observed for TRIM-21 expression, SLE patients demonstrated heightened estrogen responses, resulting in enhanced CREM␣ expression as compared with healthy controls (Figure 3D), further indicating that monocytes isolated from SLE patients exhibit hyperresponsiveness to estrogen.
Abrogated estrogen-induced TRIM-21 expression following ER␣ inhibition. Given the central role of TRIM-21 in regulating IRF-3 levels and, hence, IFN␤ and IL-23 expression, we next tested the effect of estrogen stimulation and ER␣ inhibition on these pathways. Monocytes from healthy control subjects were left untreated or were pretreated with MPP, a specific pharmacologic antagonist to ER␣, followed by stimulation with 17␤-estradiol (10 -7 M). Blocking ER␣ using this selective antagonist abrogated 17␤-estradiol-mediated TRIM-21 mRNA expression, indicating that 17␤estradiol up-regulates TRIM-21 expression in an ER␣dependent manner.
In order to fully appreciate the effects of ER␣ inhibition on TRIM-21 expression at the protein level, monocytes from SLE patients were left untreated or were pretreated with MPP, followed by stimulation with 17␤-estradiol (10 -7 M). Consistent with the results observed in the healthy controls, 17␤-estradiol upregulated TRIM-21 protein expression ( Figure 4B), the levels of which were then subsequently reduced with MPP treatment. Significantly, levels of IRF-3 protein accumulated in untreated cells, while cells pretreated with MPP demonstrated reduced levels of IRF-3 protein, which corresponded to a decrease in the levels of TRIM-21 protein, confirming a role of TRIM-21 in stabilizing this transcription factor in non-TLRstimulated cells ( Figure 4B).
Given the role of TRIM-21 and IRF-3 in regulating both IL-23 and IFN␤ production, we observed that estrogen stimulation drove the expression of both cytokines and that MPP pretreatment resulted in a loss of expression of both IL-23 and IFN␤, an effect mirroring that observed for TRIM-21 ( Figures 4C and D, respectively). Our study strongly suggests that in resting cells, the role of TRIM-21 in regulating IRF-3 levels has a direct impact on the levels of cytokines being produced. In cases where TRIM-21 levels are elevated, the resulting increase in basal IRF-3 results in a corresponding increase in both IFN␤ and IL-23. Given the central role of type I IFNs and IL-23 in the pathology of SLE and the published role of TRIM-21 in regulating these cytokines, our study has important implications concerning the role of estrogen in the etiology and pathogenesis of SLE as well as the potential for ER inhibitors in alleviating this.

DISCUSSION
The central role of TRIM-21 in regulating inflammatory cytokine and type I IFN production has led to the suggestion that TRIM-21 may be a central player in the establishment of autoimmunity, in particular, SLE (35,36). SLE is an autoimmune condition with a significant female sex bias, implicating the importance of sex and sex hormones in the development of this disease. This genetic element can be seen in males with Klinefelter's syndrome, an XXY condition characterized by an imbalanced estrogen/androgen pattern, who have a higher prevalence of SLE compared with XY males (33). Importantly, disease activity in female SLE patients appears to heighten during the early reproductive years (37), while flares decrease following menopause (38), supporting a role of estrogen in both the development and severity of this disease.
A number of SLE-associated genes have recently been shown to be estrogen dependent, including IRF-5, BAFF, and CREM␣ (16,17,20). Our results identify TRIM-21 as a novel estrogen-regulated gene, with ER␣ demonstrated to bind directly to the TRIM-21 promoter and drive gene expression. Importantly, an ER␣selective antagonist MPP was shown to reverse the effect of estrogen on TRIM-21 expression, with a concomitant decrease in inflammatory cytokine and type I IFN levels in SLE patient cells. In addition, a strong correlation was observed between TRIM-21 and ER␣ mRNA levels in SLE patients, thus underlining the importance of estrogen in regulating TRIM-21 expression, particularly in the context of SLE, in order to maintain the appropriate cytokine balance in the immune system.
The antiviral pattern-recognition receptors TLR-7 and TLR-9 have a well-established role in the development of SLE via their ability to recognize self RNA and DNA and thus drive B cell and dendritic cell activation, autoantibody production, and type I IFN production, key cytokines associated with the pathology of SLE (39,40). Estrogen has been shown in a number of studies to enhance TLR-9-driven B cell and dendritic cell activation, including the production of proinflammatory cytokines, which suggests that estrogen positively regulates cytokine production and immune cell activation, potentially via direct transcriptional effects, as has been shown for IL-2 in T cells (20), but also perhaps via its ability to positively regulate TRIM-21 expression, as we have shown in this study.
The link between aberrant TRIM-21 activity and SLE is growing, with initial studies in the TRIM-21 knockout mouse demonstrating that deregulated signaling through the TLR-7 and TLR-9 signaling pathways resulted in the development of lupus-like symptoms in the mice, including glomerular nephritis, autoantibody production, and enhanced production of proinflammatory cytokines and type I IFNs (24). More recently, TRIM-21 has been shown to negatively regulate the intracellular DNA receptor DEAD box polypeptide 41 and, in doing so, limit cytokine production in response to DNA detection (41). In both cases, the role of TRIM-21 in the context of DNA or RNA recognition was examined, where loss of TRIM-21 positively contributes to the autoimmune phenotype via increased stability of transcription factors, such as IRF-3 or IRF-7, and enhanced cytokine production as a result.
Interestingly, TRIM-21 has also been reported to be a high-affinity Fc receptor (42) that can bind antibody-self-antigen complexes, thus facilitating immune signaling and activation (43). Consistent with this, we report that TRIM-21, and as a consequence IRF-3, levels were elevated in resting SLE immune cells, which may suggest that enhanced TRIM-21 expression in SLE patients could enhance the detection of antibody-selfantigen complexes present in these patients, thereby exacerbating inflammation. Critically for this study, TRIM-21 also appeared to stabilize members of this transcription factor family, namely, IRF-3 and IRF-8, in resting immune cells, as evidenced by the decreased levels of IRF-3 and IRF-8 observed in the absence of TRIM-21 (23,24,26).
In terms of immune stimulation, most likely through recruitment of additional factors such as the p62 sequestosome, as has been shown for IRF-8 (26), the role of TRIM-21 is to ubiquitinate and degrade IRF-3, IRF-7, and IRF-8 (23,26,44). Consistent with this, we report that TRIM-21 levels were elevated in resting PBMCs and monocytes from SLE patients and that the increased expression of TRIM-21 observed at both the mRNA and protein levels was accompanied by increased levels of IRF-3.
We observed enhanced levels of TRIM-21 expression as a result of estrogen regulation, which corresponds to IRF-3 levels in monocytes. Although we cannot definitively rule out a direct effect of estrogen on IL-23p19 or IFN␤ gene expression, our results suggest that the ability of estrogen to drive TRIM-21 expression and the role of TRIM-21 in regulating IRF-3 stability potentially contribute to the enhanced basal levels of type I IFNs and IL-23 observed in SLE patients. Upon investigation into the mechanism of this estrogeninduced TRIM-21 expression, we identified a potential ERE site in the TRIM-21 promoter. This GGTTA-ANNNTGACC region differs from the ERE consensus sequence GGTCANNNTGACC by 1 nucleotide. Although base mutations from this consensus sequence can decrease the binding affinity between the ERs and their target DNA, previous studies have shown such sites to be functional (12). Similarly, we found the potential ERE site located in the TRIM-21 promoter to be functional, with estrogen inducing ER␣ binding to this region and, hence, regulating gene expression.
Estrogen antagonists have previously been investigated for their ability to ameliorate SLE symptoms and pathology, both in murine studies and in clinical trials. Modulation of estrogen signaling using the wellcharacterized ER␣ antagonist tamoxifen in female NZB/W mice has proved highly successful, with significantly reduced anti-DNA antibody production, improved nephritis, and improved overall survival (7). Similarly, in lupus-prone MRL-lpr/lpr mice, estrogendriven enhancement of inflammatory cytokines, proteinuria, and autoantibody levels were reversed by coadministration of tamoxifen (45).
Unfortunately, these promising results observed in murine studies have yet to be realized in humans, with a small study in 11 SLE patients demonstrating no beneficial effects of tamoxifen on disease activity or serologic markers (46). However, in a double-blind, placebo-controlled clinical trial of fulvestrant, an estrogen receptor-selective down-regulator, patients receiving fulvestrant for 12 months showed improved scores on the Systemic Lupus Erythematosus Disease Activity Index and a reduction in the level of conventional medication required to manage disease symptoms, thus highlighting the therapeutic potential for estrogen receptor down-regulators (47). Our current findings indicate that TRIM-21 levels are intricately linked with IL-23p19 and IFN␤ levels, with increased TRIM-21 expression as a result of estrogen stimulation, yielding a corresponding increase in these cytokines, whereas cotreatment with the ER␣ antagonist MPP completely blocked the estrogen-driven responses and, importantly, restored IL-23p19 and IFN␤ levels to baseline.
While studies of estradiol levels in healthy females as compared with females with SLE demonstrated no difference in estrogen levels, abnormal levels of estrogen metabolites, such as 2-and 16-hydroxyestrone, were noted in the SLE patients (48). Furthermore, it has been suggested that estrogen sensitivity may be enhanced in SLE patients due to variations in ER expression in these patients. In the present study, we observed significantly enhanced levels of ER␣ expression in PBMCs from SLE patients that correlated positively with their TRIM-21 levels. We also demonstrated enhanced estrogen-induced TRIM-21 expression in PBMCs from SLE patients as compared with those from healthy controls, thus supporting the idea that an overactive ER␣/TRIM-21 pathway may hold potential as a therapeutic target for use in SLE. Overall, our results identify a novel link between estrogen and the deregulated immune system seen in SLE and further support estrogen and ER signaling as important therapeutic targets for the treatment of SLE.