**1. Introduction**

Rheumatoid arthritis (RA) is the most frequent systemic autoimmune disease in the Western-European and North-American countries [1]. The disease primarily affects the small joints, where a chronic, progressive inflammation leads to cartilage and bone destruction, associated with severe pain and disability [1]. Despite intensive research, no definitive cause(s) of the disease are known, and therefore, unfortunately, no curative treatment is available to date [2]. So, finding the potential pathogenic factors and mechanisms in the background of RA is of utmost importance because they might provide a basis for future therapies. In this regard, mouse models of RA are extremely useful because several aspects of the disease can be studied more efficiently than in humans [3]. Especially those

models are beneficial, which share many features of RA, like proteoglycan-aggrecan-induced arthritis (PGIA) [4] and its refined version, recombinant human G1 domain-induced arthritis (GIA) [5]. (P)GIA is similar to RA in many respect: (i) clinical picture [4,5], (ii) histological changes [4,5], (iii) radiological changes [4,6], (iv) autoreactive T cell activation [7], (v) Th1 and Th17 differentiation [8], (vi) production of autoantibodies (both against the mouse cartilage aggrecan and citrullinated antigens) [5] and (vii) proinflammatory cytokines was described [5].

RA is a chronic, progressive disease, usually lasting for decades [2,9,10]. According to our present view on RA pathogenesis, the patients are diagnosed only in the final inflammatory/destructive phase of the disease, when the typical symptoms (pain, swollen joints) appear [10]. However, the loss of tolerance and the development of symptomless autoimmunity precedes this usually by several years [10]. The dysregulation of the immune system might be the result of the interplay between genetic (*MHC* and *non-MHC* genes), epigenetic and environmental factors (infections, diet, smoking) [9,10]. The development of autoreactive T cells and starting of autoantibody production might be the key elements of this preclinical/latent phase of RA [10] and likewise, during the initiation period of its model PGIA [8]. These processes most likely take place not only locally, in the joints, but, importantly, in the lymphatic tissues like the lymph nodes and the spleen [2,8].

The spleen is well known for its function in the degradation of red blood cells and the immune response against blood-borne antigens, especially those of encapsulated bacteria [11,12]. Moreover, the spleen is critical for B cell development and, uniquely, all major peripheral B cell populations (B1a-, B1b-, B2-, marginal zone (MZ) B cells) can be found here. Not much is known about the exact role of the spleen in RA, however, based on mouse models of GIA and collagen-induced arthritis (CIA), we might suspect a potential involvement: increased size and more activated cells can be detected in spleens from both GIA (own unpublished observation) or CIA mice [13].

Nirenberg-Kim (NK) 2 homeobox 3 (Nkx2-3) is a homeodomain transcription factor, which is essential for the normal development of the spleen, Peyer's patches and small intestine [14–17]. Along with its role in the development of intestinal lymphoid tissues. Nkx2-3 is crucial for the expression and regulation of the mucosal addressin cell adhesion molecule-1 (MADCAM-1) on the spleen sinus lining cells and on high endothelial venules of the mesenteric lymph nodes and Peyer's patches [17–20]. Nkx2-3 has an important role in spleen organization and function, since it controls the correct micro-environment for B cell maturation and T-cell-dependent (TD) immune reaction [14,21]. Its absence results in disorganized germinal center (GC) formation leading to abnormal secondary B cell differentiation and decreased antibody response with minimal affinity maturation [21,22]. Nkx2-3-deficient mice (Nkx2-3−/−) are either asplenic or have a significantly reduced spleen size with a lack of the marginal zone [21]. In response to the TD antigen, the number of circulating lymphocytes of the Nkx2-3−/<sup>−</sup> mice was found to be increased compared to both Nkx2-3+/<sup>−</sup> and Nkx2-3+/+ [21] indicating their altered distribution between peripheral lymphoid tissues. Moreover, Nkx2-3−/<sup>−</sup> mice showed an elevation in the number of the B cells in mesenteric lymph nodes which may be due to the abnormal development of the small intestine and the Peyer's patches of these mice, and also a significant increase in the number of the IgM<sup>+</sup> B cells in the bone marrow (BM) [21].

In humans, overexpression of Nkx2-3 was found to be associated with both Crohn's disease and ulcerative colitis through its effect on the regulation of PTPN2 expression, VEGF and MADCAM-1 signaling, and the production of endothelin-1 [16,18,23,24]. Additionally, Robles and colleagues reported that chromosomal translocation of *Nkx2-3* gene alongside with immunoglobulin heavy chain gene (*IGH*), resulted in irregular B cell receptor signaling leading to the MZ B cell lymphomagenesis, through the activation of the NF-KB and PI3K-AKT pathways [25].

Our aim in this study was to investigate the effect of Nkx2-3 deficiency in GIA, a mouse model of autoimmune arthritis, and study the effect of Nkx2-3 absence on B cell signaling and activation. Here, we report for the first time that GIA can be induced in Nkx2-3−/<sup>−</sup> mice, although with lower incidence, decreased severity and less joint destruction. We measured decreased T cell proliferation and cytokine production in spleen cultures. We found less anti-CCP-IgG2a, IL-17 and IFNγ, but more IL-1β, IL-4

and IL-6 in the sera. Finally, B cells of Nkx2-3−/<sup>−</sup> mice showed decreased intracellular Ca2<sup>+</sup> signaling compared to those isolated from BALB/c mice. Collectively, these data indicate that Nkx2-3 mice are relatively resistant to GIA-induction which correlates with their impaired in vitro B cell responsiveness.
