Collagen-induced arthritis as an animal model for rheumatoid arthritis: role of IFN-gamma in immunomodulation by mesenchymal stem cells and in extra-articular manifestations.

Over an extended period of time, work in our laboratory focused on unravelling the role of interferon-γ (IFN-γ) in collagen-induced arthritis (CIA), a well characterised animal model for human rheumatoid arthritis (RA). IFN-γ was found to display anti-inflammatory properties in CIA since IFN-γ receptor knock-out (IFN-γR KO) mice developed CIA more rapidly and with a higher severity as compared to their wild-type counterparts. The protective effect of IFN-γ could be explained by its suppressive effects on neutrophil mobilisation, activation of macrophages/osteoclasts and its role in the activation of regulatory T cells. The dogma that IFN-γ acts as a proinflammatory cytokine was questioned by these findings. Moreover, IFN-γ could be considered a novel therapy for RA, a perception confirmed by numerous clinical trials concerning IFN-γ in RA. In the quest for novel therapies for RA, mesenchymal stem cells (MSCs) have recently been proposed. They are multipotent cells that combine interesting characteristics such as tissue regeneration and immunosuppression and might therefore prove to be ideal therapeutics for RA. Since IFN-γ has been identified as an activator of the immuosuppressive capacities of MSCs, in this thesis we investigated the potential of MSCs to suppress the joint pathology in CIA and the effect of IFN-γ on this suppression. Although the joint pathology in CIA resembles that of RA remarkably well, RA is associated with extra-articular manifestations in for instance lung tissue. In CIA no reports of pulmonary involvement were reported. In this thesis we therefore investigated whether mice with CIA would develop pulmonary manifestations and analysed the role of IFN-γ on this systemic pathology.MSCs were isolated from the bone marrow of wild-type and IFN-γR KO mice and after several passages homogenous populations of MSCs were obtained. By culturing anti-CD3-stimulated T cells in vitro in the presence of MSCs we could demonstrate that T cell responses were suppressed by MSCs in a dose-dependent manner. IFN-γR KO MSCs were less suppressive than wild-type MSCs, indicating an IFN-γ-dependent mechanism was involved in the T cell suppression. By stimulating MSCs with IFN-γ we could demonstrate that the production of programmed death ligand-1, inducible nitric oxide synthase and cyclo-oxygenase-2, but not indoleamine 2,3-dioxygenase was involved in the immunosuppression mediated by MSCs. Despite their potent in vitro suppressive effects, MSCs were unable to counteract in vivo anti-CD3-induced T cell proliferation in mice. MSCs also failed to suppress CIA, even when they were administered at multiple time points or via different routes. This was reflected in unchanged humoral and cellular responses towards collagen type II. Together these data demonstrate that the immunosuppressive potential of MSCs in vitro can not be extrapolated to an in vivo situation.When MSCs are considered as a therapeutic in RA, thorough investigation of the effect of MSCs on all cell types involved in the pathogenesis of RA is necessary. Therefore, we investigated the effect of MSCs on osteoclasts, the cell type responsible for bone degradation in the joints of RA patients. Using mouse embryo fibroblasts (MEFs) as a source of fibroblasts, we developed an in vitro co-culture system between MEFs and splenocytes from mice with CIA. We could demonstrate that MEFs could stimulate the development of osteoclasts from splenocytes even without addition of osteoclast growth factors. When MEFs were substituted by MSCs, no osteoclasts could be detected. Only when receptor activator of nuclear factor-kB ligand (RANKL), a potent osteoclast inducer, was added osteoclasts developed. In the presence of tumor necrosis factor-α (TNF-α), osteoclasts were also not formed. Thus, MSCs can only support osteoclast differentiation when RANKL is present. As a consequence, whether MSCs will induce osteoclast differentiation in the joint of RA patients will probably depend on the cytokine balance present in the joint and administration of these cels in inflammatory conditions (containing RANKL and TNF-α) possibly has detrimental effects.As a second main goal in this thesis, we investigated whether pulmonary manifestations, which regularly complicate RA, were also present in CIA in mice. On macroscopic examination of the lungs 21 days after the induction of CIA, lesions could be detected on the surface of the lungs in IFN-γR KO mice but not in wild-type mice. Elevated numbers of macrophages and neutrophils were detected in bronchoalveolar lavage fluid of IFN-γR KO mice. Upon histological examination of the lungs, perivascular and peribronchial lymphocytic infiltrates as well as subpleural nodular accumulations of neutrophils and histiocytes could be visualised in lungs of arthritic IFN-γR KO mice but not wild-type mice. In IFN-γR KO mice, this pulmonary infiltration was accompanied by elevated mRNA levels of proinflammatory cytokines and chemokines. Upon functional assessment of the lungs, impaired lung function was ascertained which presented as a stiffening of the lungs. Treating the mice with anti-TNF-α therapy resulted in a complete prevention of joint pathology and a partial but significant reduction of the pulmonary complications. These data indicate that extra-articular manifestations in CIA can be provoked in CIA by interfering with one cytokine signaling pathway i.e. IFN-γ.Previous doctoral research in our laboratory demonstrated that upon immunisation with complete Freunds adjuvant (CFA) without the cartilage antigen collagen type II, IFN-γR KO mice develop arthritic symptoms which are identical to the joint pathology observed in CIA. Upon examination of the lungs, we could demonstrate that IFN-γR KO mice with CFA-induced arthritis displayed similar pulmonary pathology as IFN-γR KO mice with CIA. An important component of CFA is heat-killed Mycobacterium butyricum. Since this M. butyricum is the common factor in both arthritis models, we focused our attention to the M. butyricum to find a possible cause of the lung involvement. After fluorescent labelling of the M. butyricum, we could determine that the M. butyricum was present in the lungs of immunised animals. Thus, from these data we can conclude that M. butyricum used for the induction of CIA and CFA-induced arthritis translocates to the lungs after immunisation of IFN-γR KO mice and might locally cause production of proinflammatory cytokines and chemokines, followed by pulmonary inflammation.Summarising our investigations, we can state that although MSCs display potent immunosuppressive potential in vitro, they were unable to counteract in vivo T cell proliferation or CIA. Furthermore, MSCs could, depending on the local cytokine balance, possibly stimulate the development of osteoclasts. These findings are important to keep in mind when considering MSCs as potential therapy for RA. Although CIA is generally considered not to be associated with extra-articular manifestations, we could demonstrate that IFN-γR KO mice with CIA did present with pulmonary pathology. This finding adds relevance to the animal model and offers clear future perspectives both on fundamental and clinical level. We feel that on fundamental level, the model is suited to investigate the association between environmental factors, such as smoking, citrullination in the lungs and the development of arthritis. The fact that these extra-articular manifestations mainly occur in mice that have a mutation in the IFN-γ receptor, might offer new perspectives to the clinic such as treatment of RA patients with IFN-γ, possibly combined with current therapies.

Publication year:2011

Accessibility:Closed