Studies of the immunopathogenesis of arthritis, with an emphasis on the alarmin HMGB1

Abstract

Dysregulated inflammatory responses are characterized by the excessive release of endogenous inflammatory molecules, which initiate a chain of reactions, including immune cell infiltration, activation, polarization, necrosis, apoptosis and pyroptosis. These changes in cell distribution and status can lead to tissue damage, represented by swelling, pain, redness, heat and loss of function. Rheumatoid arthritis (RA) and juvenile idiopathic arthritis (JIA) are diseases involving dysregulated inflammatory responses and are the most common rheumatic diseases in adults and children, respectively. Although there have been many studies investigating the prognosis, diagnosis and treatment of RA and JIA, there are still unmet clinical needs in this regard. A better understanding of the pathogenesis of each disease is essential to address these unmet needs.

High mobility group box 1 (HMGB1), a prototypical damage-associated molecular pattern (DAMP), is expressed in all nucleated animal cells and platelets. The role of HMGB1 depends on its localization, post-translational modification and redox modification. HMGB1 is actively secreted or passively released into the extracellular region during cell activation or cell death. Extracellular HMGB1 acts as an alarmin that can initiate the immune response alone or combined with other molecules, such as nucleic acid, to participate in multiple biological processes. It has been evident that HMGB1 is involved in various inflammatory responses and autoimmunity, including RA and JIA. Increased levels of HMGB1 have been recorded in arthritic joints, and HMGB1 blockade ameliorates experimental arthritis.

The overall aim of this thesis is to investigate the pathogenesis of arthritic diseases, including RA and JIA, with a special emphasis on the involvement of HMGB1.

In papers I and II, we used proximity extension assay (PEA) to measure 92 inflammation-related proteins in plasma and synovial fluid (SF) from patients with systemic JIA (sJIA) and oligoarticular JIA (oligoJIA). By performing a cross-sectional comparison with age- and sex-matched healthy controls, we were able to not only confirm the previously reported sJIA biomarkers (IL6, IL18 and S100A12), but also find novel markers which could distinguish active sJIA from inactive sJIA or controls. The level of HMGB1 was significantly higher in active sJIA than inactive sJIA in both paired and cross-sectional analysis. In contrast, plasma proteomics did not distinguish patients with oligoJIA from healthy controls. This finding corresponds to the clinical definition of oligoJIA, where patients have local joint inflammation but lack significant systemic inflammation. Longitudinal analysis of twenty SF and ten plasma samples from an individual patient revealed the immunosuppression effects of methotrexate (MTX). Finally, the paired analysis of SF indicated that, compared to the persistent phase, cell chemotaxis was the main character defining the early phase of oligoJIA.

Macrophages are versatile myeloid cells that play an important role in tissue homeostasis, immune defense, and inflammatory progression and resolution. Macrophage polarization plays a role in defining the outcomes of different diseases, including arthritis. In papers III and IV, we investigated the effects of HMGB1 on the in vitro polarization of murine bone marrow-derived macrophages (BMDMs). Compared to proinflammatory phenotype M1 and alternatively activated anti-inflammatory phenotype M2, disulfide HMGB1 (dsHMGB1) induced a unique macrophage phenotype that secretes proinflammatory cytokines, rather than inducing metabolic changes leading to nitric oxide production; while fully reduced HMGB1 (frHMGB1) did not trigger any significant change in cytokine release or gene expression. Both dsHMGB1 and frHMGB1 could induce cell migration. Moreover, RNA sequencing (RNA-Seq) was performed to generate overall transcriptomic profiles of HMGB1-stimulated BMDMs. The results further confirmed the initial findings that dsHMGB1 induced a distinct BMDM polarization phenotype compared to LPS/IFNγ- (M1) and LPS-induced phenotypes, while frHMGB1 failed in inducing a significant transcriptomic profile shift compared to controls.

In paper V, we investigated the occurrence of neuroinflammation in experimental arthritis. Two mouse arthritis models, collagen antibody-induced arthritis (CAIA) and KRN T-cell arthritis (KRN), were established, and brain tissues were analyzed, with a focus on the hippocampus area. Our data suggested that arthritis could lead to neuroinflammation, significantly upregulating proinflammatory gene expression and interfering with hippocampal neurogenesis and proliferation.

Taken together, the findings presented in this thesis contributed new knowledge about arthritis and inflammation. We discovered novel protein biomarkers for JIA, which could benefit the prognosis, diagnosis and classification of JIA. We compared HMGB1, a DAMP, with LPS, a pathogen-associated molecular pattern (PAMP), regarding their effects on macrophage polarization. We also carried out RNA-Seq analysis to obtain the transcriptomic profiles of frHMGB1 and dsHMGB1 stimulated BMDMs, which may provide good references for other researchers interested in the inflammatory properties of HMGB1 and role of HMGB1 as a mediator of arthritic inflammation. Finally, we explored the occurrence of neuroinflammation in two arthritis mouse models, revealing the potential mechanisms behind arthritis-induced cognitive disorders.