IntroductionThe immune modulatory role of estrogens in inflammation is complex. Both pro- and anti-inflammatory effects of estrogens have been described. Estrogens bind both estrogen receptor (ER)α and β. The contribution of ERα and ERβ to ER-mediated immune modulation was studied in delayed type hypersensitivity (DTH) and in experimental arthritisMethodsER-mediated suppression of rat adjuvant arthritis (AA) was studied using ethinyl-estradiol (EE) and a selective ERβ agonist (ERB-79). Arthritis was followed for 2 weeks. Next, effects of ER agonists (ethinyl-estradiol, an ERα selective agonist (ERA-63) and a selective ERβ agonist (ERB-79) on the development of a tetanus toxoid (TT)-specific delayed type hypersensitivity response in wild type (WT) and in ERα - or ERβ-deficient mice were investigated. Finally, EE and ERA-63 were tested for their immune modulating potential in established collagen induced arthritis in DBA/1J mice. Arthritis was followed for three weeks. Joint pathology was examined by histology and radiology. Local synovial cytokine production was analyzed using Luminex technology. Sera were assessed for COMP as a biomarker of cartilage destruction.ResultsEE was found to suppress clinical signs and symptoms in rat AA. The selective ERβ agonist ERB-79 had no effect on arthritis symptoms in this model. In the TT-specific DTH model, EE and the selective ERα agonist ERA-63 suppressed the TT-specific swelling response in WT and ERβKO mice but not in ERαKO mice. As seen in the AA model, the selective ERβ agonist ERB-79 did not suppress inflammation. Treatment with EE or ERA-63 suppressed clinical signs in collagen induced arthritis (CIA) in WT mice. This was associated with reduced inflammatory infiltrates and decreased levels of proinflammatory cytokines in CIA joints.ConclusionsERα, but not ERβ, is key in ER-mediated suppression of experimental arthritis. It remains to be investigated how these findings translate to human autoimmune disease.

We sought to identify an altered peptide ligand (APL) based on the endogenously expressed synovial auto-epitope of human cartilage glycoprotein-39 (HC gp-39) for modulation of cognate, HLA-DR4-restricted T cells. For this purpose we employed a panel of well-characterized T cell hybridomas generated from HC gp-39-immunized HLA-DR4 transgenic mice. The hybridomas all respond to the HC gp-39(263-275) epitope when bound to HLA-DR4(B1*0401) but differ in their fine specificities. First, the major histocompatibility complex (MHC) and T-cell receptor (TCR) contact residues were identified by analysis of single site substituted analogue peptides for HLA-DR4 binding and cognate T cell recognition using both T hybridomas and polyclonal T cells from peptide-immunized HLA-DR4 transgenic mice. Analysis of single site substituted APL by cognate T cells led to identification of Phe265 as the dominant MHC anchor. The amino acids Ala268, Ser269, Glu271 and Thr272 constituted the major TCR contact residues, as substitution at these positions did not affect HLA-DR4(B1*0401) binding but abrogated T cell responses. A structural model for visualisation of TCR recognition was derived. Second, a set of non-classical APLs, modified at the MHC key anchor position but with unaltered TCR contacts, was developed. When these APLs were analysed, a partial TCR agonist was identified and found to modulate the HC gp-39(263-275)-specific, pro-inflammatory response in HLA-DR4 transgenic mice. We identified a non-classical APL by modification of the p1 MHC anchor in a synovial auto-epitope. This APL may qualify for rheumatoid arthritis immunotherapy.

Prednisolone and other glucocorticoids (GCs) are potent anti-inflammatory drugs, but chronic use is hampered by metabolic side effects. Therefore, there is an urgent medical need for improved GCs that are as effective as classical GCs but have a better safety profile. A well-established model to assess anti-inflammatory efficacy is the chronic collagen-induced arthritis (CIA) model in mice, a model with features resembling rheumatoid arthritis. Models to quantify undesired effects of glucocorticoids on glucose kinetics are less well-established. Recently, we have described a model to quantify basal blood glucose kinetics using stably-labeled glucose. In the present study, we have integrated this blood glucose kinetic model in the CIA model to enable quantification of both efficacy and adverse effects in one animal model. Arthritis scores were decreased after treatment with prednisolone, confirming the anti-inflammatory properties of GCs. Both inflammation and prednisolone induced insulin resistance as insulin secretion was strongly increased whereas blood glucose concentrations and hepatic glucose production were only slightly decreased. This insulin resistance did not directly resulted in hyperglycemia, indicating a highly adaptive compensatory mechanism in these mice. In conclusion, this ‘all-in-one’ model allows for studying effects of (novel) GC compounds on the development of arthritis and glucose kinetics in a single animal. This integrative model provides a valuable tool for investigating (drug-induced) metabolic dysregulation in an inflammatory setting.