Endogenous estrogen regulation of inflammatory arthritis and cytokine expression in male mice predominantly via estrogen receptor ╨Ю┬▒.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 62, No. 4, April 2010, pp 1017–1025 DOI 10.1002/art.27330 © 2010, American College of Rheumatology Endogenous Estrogen Regulation of Inflammatory Arthritis and Cytokine Expression in Male Mice, Predominantly via Estrogen Receptor ␣ Y. H. Yang,1 D. Ngo,1 M. Jones,2 E. Simpson,2 K. H. Fritzemeier,3 and E. F. Morand1 Objective. A number of experimental observations have associated elevated estrogen levels with amelioration of inflammation. The involvement of estrogen and estrogen receptor (ER) isotypes in the regulation of inflammation in males is not well understood. In this study, we used specific ER␣ and ER␤ agonists in male mice deficient in estrogen because of a deletion of aromatase (aromatase-knockout [ArKO] mice) to investigate ER isotype utilization in estrogen regulation of inflammation. Methods. Lipopolysaccharide (LPS)-induced cytokine expression and antigen-induced arthritis (AIA) were investigated in male ArKO and WT littermate mice, as well as in response to selective agonists of ER␣ (16␣-LE2) and ER␤ (8␤-VE2). The therapeutic effect of selective ER agonists was also examined in mice with collagen-induced arthritis (CIA). Results. Estrogen deficiency in ArKO mice was associated with significant increases in LPS-induced serum interleukin-6 (IL-6), tumor necrosis factor, monocyte chemotactic protein 1, and interferon-␥ levels, which were significantly abrogated by administration of 16␣-LE2, but not 8␤-VE2. In contrast, both 16␣-LE2 and 8␤-VE2 significantly increased LPS-induced IL-10 levels. Estrogen deficiency was also associated with significant exacerbation of AIA and antigen-specific T cell proliferation, which was reversed by administration of either 16␣-LE2 or 8␤-VE2. ArKO mice showed increased antigen-specific T cell proliferation in response to immunization with type II collagen (CII). Administration of 16␣-LE2, but not 8␤-VE2, significantly reduced the severity of CIA, which was associated with inhibition of anti-CII–specific IgG. Conclusion. These data indicate that endogenous estrogen plays an essential inhibitory role in inflammation in male mice and that ER␣ is the dominant receptor that mediates these effects. Most autoimmune diseases, including rheumatoid arthritis (RA), systemic lupus erythematosus, multiple sclerosis, and Sjögren’s syndrome (SS), are more common in females than in males. A significant body of evidence implicates the influence of sex-specific factors, including sex hormones, on inflammation (1,2). For example, pregnancy has an ameliorating effect on RA, scleroderma, and SS. A shift from Th1 to Th2 immune response bias has been proposed as a mechanism underlying the improvement in putative Th1-mediated autoimmune diseases during pregnancy (2). In support of this, collagen-induced arthritis (CIA) in rats has been shown to be exacerbated by ovariectomy in females or by castration in males, and in both cases, correction of the sex hormone deficiency reversed these effects (3). Treatment with physiologic levels of 17␤-estradiol was shown to inhibit the development of CIA in DBA/1 mice (4), whereas deprivation of endogenous estrogen by ovariectomy exacerbated disease severity and induced bone loss in this model (5). Highlighting the complexity of this area, serum estradiol levels in male patients with RA are higher than those in healthy male controls (6), while some studies Supported in part by Bayer Schering Pharma AG, Germany, and by the National Health and Medical Research Council, Australia (program grant 494802). 1 Y. H. Yang, MD, PhD, D. Ngo, BSc (Hons), E. F. Morand, MD, PhD: Monash University Department of Medicine and Monash Medical Centre, Clayton, Victoria, Australia; 2M. Jones, PhD, E. Simpson, PhD: Prince Henry’s Institute, Clayton, Victoria, Australia; 3 K. H. Fritzemeier, PhD: Bayer Schering Pharma AG, Berlin, Germany. Address correspondence and reprint requests to Y. H. Yang, MD, PhD, Monash University Department of Medicine, Monash Medical Centre, Locked Bag No. 29, Clayton, Victoria 3168, Australia. E-mail: email@example.com. Submitted for publication August 24, 2009; accepted in revised form December 28, 2009. 1017 1018 suggest a protective effect of estrogen replacement therapy on the incidence of RA in postmenopausal women (7), although this was not confirmed in other studies (8). In a study of complement receptor deficiency in arthritis, female mice exhibited more severe disease, which was further exacerbated by ovariectomy (9). Notwithstanding the unresolved issues with respect to the immunomodulatory effects of estrogens in females, even less is understood about these effects in males. Male humans produce estrogen, and aromatase, the enzyme required for conversion of androgens to estrogen, is found in the testis (10). Subramanian et al (11) recently reported suppressive effects of estrogen in CIA in a study of male DBA/1 mice. Similarly, inhibitory effects of 2-methoxyestradiol, an estrogen metabolite, were recently reported in rats with adjuvant-induced arthritis in a study of male rodents (12). Interestingly, an estrogen receptor polymorphism has been associated with the risk of gouty arthritis in humans in a study predominantly of men (13). Together, these and other findings indicate a major, but poorly understood, role of estrogen in the modulation of immune and inflammatory responses in males. Estradiol (17␤-estradiol) acts through 2 known hormone receptors, estrogen receptors (ERs) ␣ and ␤, which are expressed in distinct tissues and immune cells. The ER isotype(s) responsible for the apparent inhibitory effects of endogenous estrogen in males are not known. ER␣ has been reported to mediate the inhibitory effects of estrogens on experimental autoimmune encephalomyelitis (14,15) and on adjuvant-induced arthritis in Lewis rats (16), and reduced local expression of ER␣ in synovium samples from mice with CIA has been reported (17), but these studies were limited to female rodents. Tools for the investigation of estrogen function in inflammatory immune regulation have recently been developed. Highly potent and selective ER ligands permit the examination of ER subtype effects in a variety of settings (18). The aromatase-knockout (ArKO) mouse lacks a functional Cyp19 gene that encodes aromatase and, hence, cannot synthesize endogenous estrogens (19). To determine the functional contribution of ER isotypes to the regulation of inflammation by endogenous estrogen in male mice, models of innate and adaptive immunity and arthritis were investigated in male ArKO mice and wild-type (WT) littermate control mice, and the responses to selective ER␣ and ER␤ agonists were determined. YANG ET AL MATERIALS AND METHODS Animals. ArKO mice (129SV/J ⫻ C57BL/6J) were generated by deleting 90% of exon 9 of the Cyp19 gene as described by Fisher et al (20). WT and homozygous-null offspring were generated by heterozygous matings. The genotype of the offspring was determined by polymerase chain reaction analysis, as described by Robertson et al (21). The animals were housed under specific pathogen–free conditions and had unlimited access to drinking water and a commercial mouse diet containing 15% of calories as fat, 20% of calories as protein, and 65% of calories as carbohydrate (Glen Forrest Stockfeeders, Glen Forrest, Western Australia, Australia). Studies were performed in male mice between the ages of 8 and 10 weeks. Male DBA/1J (DBA) mice were obtained from the Animal Resources Centre (Perth, Western Australia, Australia). These studies were approved by the Monash Medical Centre Animal Ethics Committee. Endotoxin-induced cytokines and treatment with ER agonists. Male ArKO and WT littermate mice ages 8–10 weeks were injected intraperitoneally with 10 mg/kg of lipopolysaccharide (LPS) from Escherichia coli O111:B4 (Sigma-Aldrich, Castle Hill, New South Wales, Australia). Blood was collected at the indicated times. In ER-replacement experiments, the selective ER␣ agonist 16␣-LE2 (3 g/kg), the selective ER␤ agonist 8␤-VE2 (100 g/kg), 17␤-estradiol (3 g/kg), or vehicle was administered to ArKO and WT littermate mice 2 hours after LPS challenge. Serum cytokine levels were measure by cytometric bead array (see below). Antigen-induced arthritis (AIA) and ER replacement in ArKO mice. AIA was established as described previously (22,23). Briefly, male WT or ArKO C57BL/6 mice were immunized with 200 g of methylated bovine serum albumin (mBSA; Sigma-Aldrich) in 200 l of Freund’s complete adjuvant, which was injected subcutaneously into the flank. On day 7, mice received 100 g of mBSA/0.1 ml of Freund’s complete adjuvant injected intradermally at the base of the tail. Arthritis was induced on day 21 by injection of 30 g of mBSA in 10 l of sterile saline or sterile saline alone into knee joints. To investigate the effects of ER␣ and ER␤ agonists, mice were treated daily with 16␣-LE2 (3 g/kg), 8␤-VE2 (100 g/ kg) (both donated by Schering; Bayer Schering Pharma, Berlin, Germany), 17␤-estradiol (3 g/kg), or vehicle for 7 days, beginning on day 21. Arthritis was analyzed histologically on day 28 after the first immunization, as described elsewhere (23). Briefly, sagittal sections (5 m) of fixed and decalcified knee joints were stained with Safranin O and counterstained with fast green/iron hematoxylin. Histologic sections were scored on a scale of 0–3 for each of the following 5 parameters: synovitis, joint space exudate, soft tissue inflammation, cartilage degradation, and bone damage. A total score was generated from the sum of the scores for these parameters (maximum total score 15). T cell activation. Antigen-specific T cell activation was analyzed as described previously (23). Spleens and lymph nodes were removed on day 28, and single-cell suspensions were prepared in RPMI 1640 containing 5% fetal calf serum (FCS) and 50 M 2-mercaptoethanol. Cells (105) were cultured in triplicate in the presence or absence of mBSA (10 g/ml) or anti-CD3 (1 g/ml) for 48 hours (37°C in an ESTROGEN REGULATION OF ARTHRITIS AND CYTOKINES IN MICE VIA ER␣ atmosphere containing 5% CO2). T cell proliferative responses were determined by measuring the incorporation of 3Hthymidine (0.5 Ci/well) during the final 18 hours of culture. Anti-mBSA antibody response. Serum concentrations of anti-mBSA IgG on day 28 were determined by enzymelinked immunosorbent assay (ELISA) as previously described (24). Briefly, polyvinyl microtiter plates were coated with 100 l of mBSA (100 g/ml) for 24 hours at 4°C. The plates were blocked for 1 hour with 2% casein (Sigma-Aldrich) in phosphate buffered saline with 0.05% Tween 20. Then, 100 l of diluted serum samples was added and incubated for 24 hours at 4°C. Biotinylated rabbit anti-mouse IgG (1:2,000 dilution) or anti–isotype-specific (IgG1 or IgG2a) antibodies (all from Dako, Carpinteria, CA) were incubated for 2 hours. The streptavidin–horseradish peroxidase conjugate (1:2,000 dilution) was incubated for 30 minutes and detected using the tetramethylbenzidine peroxidase substrate with hydrogen peroxide (optical density determined at 450 nm). Measurement of cytokines. Concentrations of cytokines in serum and culture supernatants were measured using a commercially available cytometric bead array (CBA; BD Biosciences, San Jose, CA). A mouse inflammation CBA kit was used to detect IL-6, IL-10, monocyte chemotactic protein 1 (MCP-1), interferon-␥ (IFN␥), tumor necrosis factor (TNF), and IL-12p40, as described in the manufacturer’s instructions. The samples were analyzed by flow cytometry using a MoFlo flow cytometer from Cytomation (Fort Collins, CO). The amount of each cytokine in the supernatant or serum was interpolated from a standard curve. Results are expressed as picograms per milliliter. Collagen-induced arthritis. A total of 100 g of native bovine type II collagen (CII; Chondrex, Redmond, WA) dissolved in 0.05M acetic acid was emulsified with Freund’s complete adjuvant and injected subcutaneously into the base of the tail of 7–9-week-old DBA/1, ArKO, and WT mice. Mice were given subcutaneous boosters with collagen (100 g) in Freund’s incomplete adjuvant on day 21. Mice were monitored daily for clinical features of arthritis beginning after day 22. Once arthritis was detected, DBA mice were treated daily for 14 days with 16␣-LE2 (3 or 30 g/kg), 8␤-VE2 (100 or 1,000 g/kg), 17␤-estradiol (3 g/kg), or vehicle. The clinical severity of arthritis was graded as follows: each limb was scored on a scale of 0–3, where 0 ⫽ no erythema or swelling, 1 ⫽ slight swelling and erythema in at least some digits, 2 ⫽ moderate swelling and erythema involving the entire limb or multiple limbs, and 3 ⫽ pronounced swelling leading to incapacitated limbs. Clinical evaluation was performed in a blinded manner by 2 investigators (YHY and DN). CII-specific T cell activation was analyzed in ArKO mice. Spleens and lymph nodes were removed on day 41, and single-cell suspensions were prepared in RPMI 1640 containing 5% FCS and 50 M 2-mercaptoethanol. Cells (105) were cultured in triplicate in the presence or absence of CII (50 g/ml) or anti-CD3 (1 g/ml) for 72 hours (37°C in an atmosphere containing 5% CO2). The T cell proliferative response was determined by measuring the incorporation of 3 H-thymidine (0.5 Ci/well) during the final 18 hours. Serum was collected from mice developing CIA, and IgG and IgM antibody levels were measured by ELISA. The 96-well plates were coated overnight at 4°C with 10 g/ml of 1019 Figure 1. Endotoxin-induced cytokines in vivo in male wild-type (WT) and aromatase-knockout (ArKO) mice. WT and ArKO littermates ages 8–10 weeks were injected with 10 mg/kg of lipopolysaccharide (LPS), and cytokine levels were examined in serial blood samples collected at the indicated times from time 0 to 4 hours (n ⫽ 9 per group) (A–D) as well as at 24 hours (n ⫽ 5 per group) (E–H). Concentrations of tumor necrosis factor (TNF), interleukin-6 (IL-6), monocyte chemotactic protein 1 (MCP-1), and interferon-␥ (IFN␥) were determined by cytometric bead array. Values are the mean and SEM. ⴱ ⫽ P ⬍ 0.05; ⴱⴱ ⫽ P ⬍ 0.01 versus WT mice. bovine CII and blocked with blocking buffer as described above. Following overnight incubation of samples, peroxidaseconjugated goat anti-mouse antibody specific for IgG and IgM was added, and color was developed as described above. Statistical analysis. Data were analyzed using the Mann-Whitney 2-sample rank test to determine the level of significance of between-group differences in mean histologic scores, or Student’s t-test was used for comparison of continuous variables. Results are expressed as the mean ⫾ SEM. P values less than 0.05 were considered statistically significant. 1020 YANG ET AL To examine the role of ER isotypes in the inhibitory effects of endogenous estrogen, selective ER agonists were administered to ArKO mice 2 hours after LPS challenge, and serum cytokine levels were measured at 24 hours. Administration of 16␣-LE2 (ER␣ agonist) significantly suppressed the LPS-induced levels of serum TNF and MCP-1 (Figures 2A and C), but significantly increased the LPS-induced serum levels of the antiinflammatory cytokine IL-10 (Figure 2E). There was also a trend toward suppression of LPS-induced IL-6 (P ⫽ 0.057) and IFN␥ by treatment with 16␣-LE2, although these differences were not significant. In contrast, treatment of ArKO mice with 8␤-VE2 (ER␤ agonist) had no effect on the levels of LPS-induced proinflammatory cytokines but was associated with significantly increased levels of IL-10 (Figure 2E). Effects of endogenous estrogen on AIA. Exacerbation of AIA. We next examined the effects of endogenous estrogen deficiency on the expression of arthritis in male mice. In comparison with saline-injected joints, which were histologically normal (Figure 3A), the joints Figure 2. Therapeutic effect of estrogen receptors (ERs) on lipopolysaccharide (LPS)-induced cytokines in male wild-type (WT) and aromatase-knockout (ArKO) mice. Treatment with 16␣-LE2 (ER␣ agonist; 3 g/kg) or 8␤-VE2 (ER␤ agonist; 100 g/kg) was administered intraperitoneally 2 hours after LPS (10 mg/kg) administration. Blood was collected at 24 hours, and concentrations of A, tumor necrosis factor (TNF), B, interleukin-6 (IL-6), C, monocyte chemotactic protein 1 (MCP-1), D, interferon-␥ (IFN␥), and E, IL-10 were determined by cytometric bead array. Values are the mean and SEM of 5–8 mice per group. ⴱ ⫽ P ⬍ 0.05; ⴱⴱ ⫽ P ⬍ 0.01 versus WT mice. ∧ ⫽ P ⬍ 0.05; ∧∧ ⫽ P ⬍ 0.01 versus vehicle-treated ArKO mice. RESULTS Effects of endogenous estrogen on LPS-induced cytokines. The effects of endogenous estrogen on the innate immune response in male mice were examined by evaluating LPS-induced serum cytokines in male WT and ArKO mice. In WT mice, LPS treatment for 4 hours or 24 hours induced significant increases in circulating levels of IL-6, TNF, MCP-1, and IFN␥ (Figure 1). In ArKO mice, LPS-induced increases in serum cytokine levels were also significant. There was a significant increase in serum TNF levels at 4 hours and 24 hours (Figures 1A and E). Serum levels of IL-6, MCP-1, and IFN␥ were also significantly higher in ArKO mice at 24 hours (Figures 1F–H). LPS-induced levels of IL-10 were not significantly affected by estrogen deficiency (data not shown). Figure 3. Histologic manifestations of antigen-induced arthritis (AIA) in male wild-type (WT) and aromatase-knockout (ArKO) mice. Arthritis severity in the knee joints was assessed in Safranin O–stained sections obtained on day 28. Shown are the histopathologic features of AIA in joint sections from A, a normal mouse, B, a wild-type (WT) arthritic mouse, and C, an ArKO arthritic mouse. Two regions from representative joints are shown. S ⫽ synovium; C ⫽ articular cartilage; J ⫽ joint space; P ⫽ pannus; E ⫽ exudate. (Original magnification ⫻ 50.) ESTROGEN REGULATION OF ARTHRITIS AND CYTOKINES IN MICE VIA ER␣ Figure 4. Total arthritis scores and antigen-specific lymphocyte proliferation in male wild-type (WT) and aromatase-knockout (ArKO) mice with antigen-induced arthritis (AIA). A, Total arthritis scores in WT and ArKO mice on day 28. Arthritis was assessed using a 0–3 scale for each of 5 histopathologic features, and the results were summed (maximum possible score of 15 per mouse). Each data point represents the total score in a single animal; horizontal lines show the mean. B, Arthritis scores for each of the 5 individual histopathologic features of AIA, as determined on day 28. C, Methylated bovine serum albumin (mBSA)–induced antigen-specific T cell proliferation in lymph node cells from WT and ArKO mice. D, Anti-CD3–induced T cell proliferation in lymph node cells from WT (open bars) and ArKO (solid bars) mice. NS ⫽ not significant. E, Serum levels of anti-mBSA total IgG, IgG1, and IgG2a in WT and ArKO mice. Values in B–E are the mean and SEM of 13–16 mice per group in B–D and 7 mice per group in E. ⴱ ⫽ P ⬍ 0.05 versus WT mice. ∧ ⫽ P ⬍ 0.05; ∧∧ ⫽ P ⬍ 0.01 versus vehicle-treated controls (CT) in C and D. of mBSA-injected WT control mice showed extensive synovial lining hypercellularity, soft tissue inflammation, joint space exudation, cartilage degradation, and bone damage (Figure 3B). In comparison with WT animals, ArKO mice exhibited significant exacerbation of arthritis severity, as reflected in a significantly increased total histology score (P ⬍ 0.05) (Figures 3C and 4A). Examination of individual aspects of the synovial pathology revealed significantly increased soft tissue inflammation 1021 and synovitis in ArKO mice (P ⬍ 0.05 for each comparison) and trends toward exacerbation of the joint space exudate, cartilage degradation, and bone damage (Figure 4B). T cell activation and antibody response. Significant antigen-induced proliferation was observed in lymph node cells from WT (P ⬍ 0.05) and ArKO (P ⬍ 0.01) mice. A significantly greater increase in antigen-induced proliferation was observed in lymphocytes from ArKO mice (P ⬍ 0.05) (Figure 4C). Levels of anti-CD3– induced T cell proliferation were comparable in lymphocytes from ArKO mice and those from WT mice (Figure 4D). In naive spleen cells from both WT and ArKO mice, anti-CD3 significantly induced IFN␥, TNF, and IL-6 (data not shown), and there was no significant difference in the cytokine responses to anti-CD3 treatment between cells from WT mice and cells from ArKO mice. IL-10, MCP-1, and IL-12p40 were not detectable in response to anti-CD3 (data not shown). There were no significant differences in serum levels of anti-mBSA total IgG, IgG1, or IgG2a between the ArKO mice and the WT mice (Figure 4E). Effect of ER agonist replacement in ArKO mice with AIA. We next investigated the effect of selective ER isotype agonist replacement in estrogen-deficient male mice. The increased arthritis severity in ArKO mice was significantly reduced in response to treatment with 16␣LE2 (Figure 5A), with inhibition of most histologic features of AIA, including synovitis, soft tissue inflammation, exudate, and cartilage degradation (Figure 5B). Similarly, treatment with 8␤-VE2 significantly reduced the overall arthritis severity and individual aspects of synovial pathology in ArKO mice (Figures 5A and C). Unexpectedly, treatment with 16␣-LE2 was associated with increased severity of clinical arthritis in WT mice. Antigen-specific T cell proliferation was examined ex vivo in spleen cells from all treatment groups. The significantly increased antigen-specific T cell proliferation in cells from ArKO mice was significantly suppressed by in vivo administration of 16␣-LE2, but not 8␤-VE2 (Figure 5D). In contrast, treatment of ArKO mice with 16␣-LE2 or 8␤-VE2 had no effect on antigenspecific total IgG or IgG isotypes (data not shown), but this was not unexpected, since the antigen-specific titers of IgG were not affected by estrogen deficiency per se. Effects of estrogen on CIA. CIA is characterized by activation of Th1 adaptive immunity, autoantibody production, and monocyte cytokine expression, and it is considered the murine model that is most similar to human RA. ArKO mice are generated on the C57BL/6 1022 YANG ET AL Figure 5. Role of estrogen receptor (ER) agonists in antigen-induced arthritis (AIA) in male wild-type (WT) and aromatase-knockout (ArKO) mice. Treatment with 17␤-estradiol (E2; 3 g/kg), 16␣-LE2 (ER␣ agonist; 3 g/kg), 8␤-VE2 (ER␤ agonist; 100 g/kg), or vehicle was administered daily from day 21 to day 27. A, Total arthritis scores on day 28. B, Arthritis scores for each of the 5 individual histopathologic features of AIA, as determined on day 28, in mice treated with 16␣-LE2 versus vehicle. C, Arthritis scores for each of the individual histopathologic features of AIA, as determined on day 28, in mice treated with 8␤-VE2 versus vehicle. D, Antigen-specific T cell proliferation in spleen cells from WT and ArKO mice treated with vehicle (controls [CT]) or methylated bovine serum albumin (mBSA) and in ArKO mice treated with the ER␣ and ER␤ agonists. NS ⫽ not significant. Values are the mean and SEM of 13–19 mice per group. ⴱ ⫽ P ⬍ 0.05; ⴱⴱ ⫽ P ⬍ 0.01 versus WT mice in B and C. ∧ ⫽ P ⬍ 0.05 versus vehicle-treated controls (CT) in D. ∧∧ ⫽ P ⬍ 0.01 versus ArKO mice in A. background, a strain which is relatively resistant to CIA. In response to collagen immunization, a 33% incidence of clinically detectable arthritis was observed in both the ArKO and the WT mice. However, a trend toward increased disease severity was observed in ArKO mice as compared with WT mice (Figure 6A). Consistent with this, lymph node cells from ArKO mice exhibited significantly greater antigen-specific proliferation in response to CII as compared with lymph node cells from WT mice (Figure 6B). Having demonstrated the effects of endogenous estrogen on immune responses to collagen, we then investigated the effects of selective ER agonists on CIA in disease-susceptible male DBA/1J mice. Administration of 16␣-LE2 significantly and dose-dependently reduced arthritis severity in mice with CIA (Figure 6C). High-dose 16␣-LE2 significantly inhibited arthritis severity from an early time point, whereas the inhibitory effect of low-dose 16␣-LE2, while apparent earlier, became statistically significant only after 15 days of treatment. Consistent with previous reports, estrogen treatment markedly and significantly inhibited CIA severity (Figure 6D). In contrast, there was no significant therapeutic effect on CIA in response to 8␤-VE2 (Figure 6D). Serum anti-CII IgG, but not IgM, antibody levels were significantly inhibited by administration of ESTROGEN REGULATION OF ARTHRITIS AND CYTOKINES IN MICE VIA ER␣ 1023 Figure 6. Effects of endogenous estrogen and estrogen receptor (ER) agonists on collagen-induced arthritis (CIA) in male wild-type (WT) and aromatase-knockout (ArKO) mice. A and B, CIA was induced in WT and ArKO mice with type II collagen (CII) as described in Materials and Methods, and the clinical severity of CIA (n ⫽ 6 per group) (A) and CII-specific T cell proliferation in lymph nodes (B) were determined. C–H, CIA was induced in DBA/1 mice as described in Materials and Methods. DBA/1 mice were treated daily with a low dose (3 g/kg; n ⫽ 12 mice) or a high dose (30 g/kg; n ⫽ 7 mice) of ER␣ agonist, with a low dose (100 g/kg; n ⫽ 11 mice) or a high dose (1 mg/kg; n ⫽ 7 mice) of ER␤ agonist, with 17␤-estradiol (E2) (3 g/kg; n ⫽ 7 mice), or with vehicle (n ⫽ 15 mice) for 14 days beginning at the onset of arthritis. The effects of treatment with the ER␣ agonist (C) or with the ER␤ agonist or E2 (D) on the arthritis scores for CIA were determined at the indicated times. Serum levels of anti-CII IgG (E) or IgM (F) were measured in mice with CIA treated with each dose of the ER␣ agonist or the ER␤ agonist. Serum levels of interleukin-10 (IL-10) were measured in mice with CIA treated with E2, high-dose ER␣ agonist, or high-dose ER␤ agonist (G). Levels of interferon-␥ (IFN␥) in culture supernatants of splenocytes cultured ex vivo and treated with E2, high-dose ER␣ agonist, or high-dose ER␤ agonist (H) were also determined. Values are mean and SEM. ⴱ ⫽ P ⬍ 0.05; ⴱⴱ ⫽ P ⬍ 0.01 versus WT mice (B) or versus vehicle treatment (C–H). high-dose ER␣ agonist (Figures 6E and F). No significant effect on anti-CII antibodies was observed in response to 8␤-VE2 administration. Serum concentrations of IL-6, TNF, and IFN␥ were undetectable in mice with CIA regardless of treatment. Interestingly, circulating IL-10 was detectable and the levels were increased in mice treated with estrogen or with 16␣-LE2, although there was no significant difference in comparison with controls (Figure 6G). In contrast, 8␤-VE2 had no effect on the induction of IL-10 in mice with CIA. IFN␥ concentrations were examined in supernatants of spleen cells from all treatment groups cultured ex vivo with CII. Although increased levels of antigen-induced IFN␥ were not observed in any group, significantly increased basal concentrations of IFN␥ were observed in supernatants of spleen cells from mice treated in vivo with 8␤-VE2 (P ⬍ 0.05) (Figure 6H). DISCUSSION Considerable evidence suggests that endogenous estrogen plays a significant role as a modulator of immune responses and autoimmune disease. Although the reported effects of estrogen are variable, when taken together, the available data suggest that endogenous estrogen has an inhibitory immunoregulatory effect on autoimmune diseases such as RA. Understandably, most previous studies have focused on the effects of endogenous estrogen in females. However, one-third of patients with RA are male, and as noted above, limited previous studies suggest that estrogen has inhibitory effects in male mice. Our data support an inhibitory effect of endogenous estrogen on the inflammatory response in male mice and, moreover, demonstrate significant differences in the roles of ER␣ and ER␤ in these effects. In the current study, estrogen deficiency in male mice was associated with increased and/or prolonged expression of serum proinflammatory cytokines TNF, IL-6, MCP-1, and IFN␥ in response to in vivo administration of LPS, demonstrating an inhibitory effect of endogenous estrogen on cytokine expression in male mice. Significant reversal of increased LPS-induced se- 1024 rum cytokine concentrations in estrogen-deficient mice was only observed in response to administration of 16␣-LE2, an ER␣ agonist. These observations are consistent with a previous report of significantly increased LPS-induced TNF release in macrophages deficient in ER␣ (25) and suggest that ER␣, not ER␤, mediates the inhibitory effects of endogenous estrogen on proinflammatory cytokine production in innate immune responses. The possibility that the ER␣-dependent inhibitory effect of estrogen on levels of LPS-induced proinflammatory cytokines is mediated via induction of IL-10 is not supported by the data from the current study, since IL-10 was significantly increased by both ER␣ and ER␤ agonists. The deficiency of endogenous estrogen in male ArKO mice resulted in exacerbation of joint inflammation and antigen-specific T cell activation in the presence of AIA. These are the first observations of a suppressive effect of endogenous estrogen on arthritis in male mice. Estrogen deficiency had no effect on serum antigenspecific total IgG, IgG1, and IgG2a, suggesting that the exacerbation of AIA in estrogen-deficient male mice is independent of humoral immune responses. In contrast, evidence presented here that endogenous estrogen inhibited antigen-specific T cell proliferation suggests that estrogen plays a critical role in the regulation of cellmediated immunity in male mice and in AIA. These findings are consistent with a previous report that in vivo administration of 2-methoxyestradiol suppressed adjuvant-induced arthritis in male rats, accompanied by inhibition of T cell responses to recall antigens and mitogens ex vivo (12). The receptors used by estrogen in the mediation of these effects on immune responses in male mice have not previously been studied. Administration of either the ER␣ or the ER␤ agonist was able to significantly reverse the increase in AIA severity, but only the ER␣ agonist inhibited antigen-induced T cell activation in ArKO mice. These results suggest that the inhibitory effects of endogenous estrogen on antigen-specific T cell activation is mediated through ER␣ but that additional effects contribute to the antiinflammatory effects of estrogen on arthritis in male mice with AIA. Endogenous estrogen was observed in the present study to inhibit aspects of both innate and adaptive immune responses, as measured by LPSinduced cytokines and antigen-induced T cell proliferation, but only restoration with an ER␣ agonist was able to reverse both of these responses. The broader range of effects of ER␣ suggested the possibility that ER␣ mediates the reported effects of estrogen in the YANG ET AL CIA model of arthritis in male mice, since aspects of both these pathways are operative in CIA. A trend toward increased CIA severity in male ArKO mice suggested that endogenous estrogen is an inhibitory factor in immune responses in CIA in male mice. This was supported by the increased CII-induced T cell proliferative responses in estrogen-deficient male mice, which was also consistent with the results we observed in AIA, and these data further support the hypothesis that endogenous estrogen modulates adaptive immune responses in male mice. The severity of CIA in ArKO mice, which are generated on a C57BL/6 background, was insufficient to allow the study of ER␣ and ER␤ agonists, so we elected to study the effects of ER␣ and ER␤ agonists as compared with the effects of estrogen on CIA in male DBA/1 mice. A significant inhibitory effect of ER␣ agonist administration on arthritis severity was observed, which was comparable to the effects of estrogen administration. Conversely, administration of ER␣ agonist increased the severity of AIA in male WT C57BL/6 mice. A significant inhibitory effect of ER␣ agonist on anti-CII autoantibody production was also observed, accompanied by an increase in serum levels of IL-10. In contrast, no significant effect of ER␤ agonist administration on CIA or anti-CII autoantibody production was observed. The lack of effect of 8␤-VE2 on the expression of monocyte cytokines such as TNF and MCP-1 may contribute to its lack of effect in CIA. In conclusion, the current data indicate that endogenous estrogen mediates an inhibitory effect on innate and adaptive immune responses and on the development of experimental arthritis in male mice. Although both ER␣ and ER␤ appear to be able to transduce inhibitory signals of estrogen on antigenspecific T cell responses and AIA, ER␣ mediates endogenous estrogen effects on a broader range of events, including proinflammatory cytokine production, and in association, is responsible for estrogen effects on inflammation in CIA in male mice. This suggests that ER␣ is the dominant receptor that mediates the inhibitory effects of estrogen on inflammation in male mice. AUTHOR CONTRIBUTIONS All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Yang had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study conception and design. Yang, Jones, Simpson, Fritzemeier, Morand. ESTROGEN REGULATION OF ARTHRITIS AND CYTOKINES IN MICE VIA ER␣ Acquisition of data. Yang, Ngo. Analysis and interpretation of data. Yang, Ngo, Morand. REFERENCES 1. Straub RH. The complex role of estrogens in inflammation. Endocr Rev 2007;28:521–74. 2. Whitacre CC, Reingold SC, O’Looney PA. A gender gap in autoimmunity. Science 1999;283:1277–8. 3. Ganesan K, Tiwari M, Balachandran C, Manohar BM, Puvanakrishnan R. Estrogen and testosterone attenuate extracellular matrix loss in collagen-induced arthritis in rats. Calcif Tissue Int 2008;83:354–64. 4. Latham KA, Zamora A, Drought H, Subramanian S, Matejuk A, Offner H, et al. Estradiol treatment redirects the isotype of the autoantibody response and prevents the development of autoimmune arthritis. J Immunol 2003;171:5820–7. 5. Yamasaki D, Enokida M, Okano T, Hagino H, Teshima R. Effects of ovariectomy and estrogen replacement therapy on arthritis and bone mineral density in rats with collagen-induced arthritis. Bone 2001;28:634–40. 6. Tengstrand B, Carlstrom K, Fellander-Tsai L, Hafstrom I. Abnormal levels of serum dehydroepiandrosterone, estrone, and estradiol in men with rheumatoid arthritis: high correlation between serum estradiol and current degree of inflammation. J Rheumatol 2003;30:2338–43. 7. Vandenbroucke JP, Witteman JC, Valkenburg HA, Boersma JW, Cats A, Festen JJ, et al. Noncontraceptive hormones and rheumatoid arthritis in perimenopausal and postmenopausal women. JAMA 1986;255:1299–303. 8. Spector TD, Brennan P, Harris P, Studd JW, Silman AJ. Does estrogen replacement therapy protect against rheumatoid arthritis? J Rheumatol 1991;18:1473–6. 9. Nilsson KE, Andren M, Diaz de Stahl T, Kleinau S. Enhanced susceptibility to low-dose collagen-induced arthritis in CR1/2deficient female mice—possible role of estrogen on CR1 expression. FASEB J 2009;23:2450–8. 10. Dorrington JH, Fritz IB, Armstrong DT. Control of testicular estrogen synthesis. Biol Reprod 1978;18:55–64. 11. Subramanian S, Tovey M, Afentoulis M, Krogstad A, Vandenbark AA, Offner H. Ethinyl estradiol treats collagen-induced arthritis in DBA/1LacJ mice by inhibiting the production of TNF-␣ and IL-1␤. Clin Immunol 2005;115:162–72. 12. Issekutz AC, Sapru K. Modulation of adjuvant arthritis in the rat by 2-methoxyestradiol: an effect independent of an anti-angiogenic action. Int Immunopharmacol 2008;8:708–16. 13. Huang CM, Lo SF, Lin HC, Chen ML, Tsai CH, Tsai FJ. Correlation of oestrogen receptor gene polymorphism with gouty arthritis. Ann Rheum Dis 2006;65:1673–4. 1025 14. Tiwari-Woodruff S, Morales LB, Lee R, Voskuhl RR. Differential neuroprotective and antiinflammatory effects of estrogen receptor (ER) ␣ and ER␤ ligand treatment. Proc Natl Acad Sci U S A 2007;104:14813–8. 15. De Coupade C, Solito E, Levine JD. Dexamethasone enhances interaction of endogenous annexin 1 with L-selectin and triggers shedding of L-selectin in the monocytic cell line U-937. Br J Pharmacol 2003;140:133–45. 16. Caggiano TJ, Brazzale A, Ho DM, Kraml CM, Trybulski E, Chadwick CC, et al. Estrogen receptor dependent inhibitors of NF-B transcriptional activation-1 synthesis and biological evaluation of substituted 2-cyanopropanoic acid derivatives: pathway selective inhibitors of NF-B, a potential treatment for rheumatoid arthritis. J Med Chem 2007;50:5245–8. 17. Bonnelye E, Laurin N, Jurdic P, Hart DA, Aubin JE. Estrogen receptor-related receptor-␣ (ERR-␣) is dysregulated in inflammatory arthritis. Rheumatology (Oxford) 2008;47:1785–91. 18. Hertrampf T, Seibel J, Laudenbach U, Fritzemeier KH, Diel P. Analysis of the effects of oestrogen receptor ␣ (ER␣) and ER␤selective ligands given in combination to ovariectomized rats. Br J Pharmacol 2008;153:1432–7. 19. Aeberli D, Yang Y, Mansell A, Santos L, Leech M, Morand EF. Endogenous macrophage migration inhibitory factor modulates glucocorticoid sensitivity in macrophages via effects on MAP kinase phosphatase-1 and p38 MAP kinase. FEBS Lett 2006;580: 974–81. 20. Fisher CR, Graves KH, Parlow AF, Simpson ER. Characterization of mice deficient in aromatase (ArKO) because of targeted disruption of the cyp19 gene. Proc Natl Acad Sci U S A 1998;95: 6965–70. 21. Robertson KM, O’Donnell L, Jones ME, Meachem SJ, Boon WC, Fisher CR, et al. Impairment of spermatogenesis in mice lacking a functional aromatase (cyp 19) gene. Proc Natl Acad Sci U S A 1999;96:7986–91. 22. Brackertz D, Mitchell GF, Mackay IR. Antigen-induced arthritis in mice. I. Induction of arthritis in various strains of mice. Arthritis Rheum 1977;20:841–50. 23. Gregory JL, Leech MT, David JR, Yang YH, Dacumos A, Hickey MJ. Reduced leukocyte–endothelial cell interactions in the inflamed microcirculation of macrophage migration inhibitory factor–deficient mice. Arthritis Rheum 2004;50:3023–34. 24. Ohshima S, Saeki Y, Mima T, Sasai M, Nishioka K, Nomura S, et al. Interleukin 6 plays a key role in the development of antigeninduced arthritis. Proc Natl Acad Sci U S A 1998;95:8222–6. 25. Lambert KC, Curran EM, Judy BM, Lubahn DB, Estes DM. Estrogen receptor-␣ deficiency promotes increased TNF-␣ secretion and bacterial killing by murine macrophages in response to microbial stimuli in vitro. J Leukoc Biol 2004;75:1166–72.