*2.4. Comparison of the Ca2*+*-Signaling in B and T Cells of Nkx2-3*−/<sup>−</sup> *and BALB*/*c Mice*

Since we found significant differences in the rhG1-induced immune responses of the Nkx2-3−/<sup>−</sup> and BALB/c mice, we were curious about what could be in the background of such variations. In the case of GIA, similarly to RA, close cooperation between T and B cells is necessary for the development of autoimmunity [7]. Since the Nkx2-3 mutation affects mostly the B lymphocyte development and recirculation, we hypothesized that the activation of B cells might be impaired, which, in turn, led to ameliorated arthritis. To characterize the activation of the B cells, we isolated the inguinal and mesenteric lymph nodes from Nkx2-3−/<sup>−</sup> or BALB/c control mice and loaded the lymphocytes with the Ca2+-specific indicator Fluo-3. Then, we activated the B cells with cross-linking of the BcR with anti-IgM or IgG antibodies and followed the intracellular Ca2<sup>+</sup> signals (Figure 4).

We found significantly lower anti-IgM or anti-IgG-induced Ca2+-signal in the B cells isolated from the inguinal lymph nodes of Nkx2-3−/<sup>−</sup> mice than those from BALB/c (Figure 4A,C). Similarly, B cells isolated from the mesenteric lymph nodes, showed decreased Ca2<sup>+</sup>-signal both after anti-IgM or anti-IgG-activation (Figure 4B,D) when compared to BALB/c controls, however, these differences did not reach statistical significance. Finally, although we were not expecting any changes in the T cell activation in Nkx2-3−/<sup>−</sup> mice, we tested the Ca2+-signal of T cells after anti-CD3 cross-linking. As we expected, the Ca2+-signals of T cells isolated from both the inguinal and mesenteric lymph nodes were similar in Nkx2-3−/<sup>−</sup> and BALB/c mice indicating that, indeed, the Nkx2-3 mutation caused an activation perturbance specifically in B cells.

**Figure 4.** The comparison of the Ca2<sup>+</sup>-signals of B and T cells in Nkx2-3−/<sup>−</sup> and control BALB/c mice. Cells were isolated from the inguinal (**A**,**C**,**E**) or mesenteric (**B**,**D**,**F**) lymph nodes of Nkx2-3−/<sup>−</sup> and BALB/c mice and loaded with the Ca2+-specific indicator Fluo-3. Activation of B cells was induced by anti-IgM (**A**,**B**) or anti-IgG (**C**,**D**), activation of T cells was induced by anti-CD3 cross-linking (**E**,**F**). The changes in the intracellular Ca2<sup>+</sup> levels of B or T cells were measured with a flow cytometer in the FL1 channel for five and a half minutes. Graphs show the time-dependent changes in the FL1 fluorescence (ratiometric with the intracellular Ca2<sup>+</sup> level) as mean <sup>±</sup> SEM values calculated from the data of *n* = 3 Nkx2-3−/<sup>−</sup> and *n* = 3 BALB/c mice. \* *p* < 0.05.

#### **3. Discussion**

In the present study, we set out to study how Nkx2-3-deficiency impacted the development of autoimmune arthritis. To answer this question, we used a mouse RA model, GIA, whereby we immunized Nkx2-3−/<sup>−</sup> and BALB/c control mice side-by-side with the rhG1 antigen. While the spleen plays important roles in the T-dependent immune responses [11,29], and in GIA it also serves as an activation niche for autoreactive lymphocytes [8,30], in Nkx2-3−/<sup>−</sup> mice, with severely damaged spleen, GIA could still be induced. This result shows that in GIA the spleen's role is not exclusive in the activation of autoreactive lymphocytes, and confirms that the lymph nodes and perhaps other lymphatic or extralymphatic tissues [8] may play an equally important role in the disease induction. On the other hand, we observed lower incidence and decreased arthritis severity in the Nkx2-3−/<sup>−</sup> mice compared to the wild-type controls, thus, the splenic defects had a significant impact on the rhG1-induced immune reaction. During the induction of (P)GIA, we immunized the mice with PG extracts/rhG1 and DDA intraperitoneally which led to the local activation of T lymphocytes both in the peritoneal cavity and the mesenteric lymph nodes followed by the systemic immune response in which the spleen is involved [8]. The local activation of Th1 and Th17 cells in the peritoneal cavity is of special importance because it was a specific feature of BALB/c mice which are the only susceptible mouse strain for (P)GIA [8]. The present results in Nkx2-3−/<sup>−</sup> mice might also support this: the local

peritoneal activation could remain unchanged, however, the systemic response, developing in the spleen is missing, which could be responsible for the weaker arthritis.

As it was previously described there is a significant B cell trafficking between the peritoneal cavity and the spleen [26]. Although there was a preferential homing of the peritoneal B1 and B2 cells towards other serosa surfaces like the pleura, however, some B1 and B2 cells migrated from the peritoneal cavity into the spleen [26]. Since in Nkx2-3−/<sup>−</sup> mice, the spleen microarchitecture is seriously defective, with abnormal adhesion molecule expression and vessel formation [29] most likely, the above-mentioned B cell trafficking is defective [29], in addition to the absence of MZ B cells involved in antigen delivery for GC initiation [21]. B lymphocytes not only play a role as precursors for antibody-producing plasma cells in autoimmune arthritis but have an equally important role as antigen-presenting cells [31]. Based on the above-mentioned works [26,29], we propose that the local, peritoneal antigen presentation by B cells could remain unchanged, however, due to the potential lack of splenic homing (and the severely reduced GC formation in the spleen upon T-dependent antigen challenge [21]), the systemic activation could be impaired. The subsequently weaker immune response is mirrored in the decreased splenic proliferation and cytokine production. This was particularly pronounced in the case of lymphocyte-derived cytokines (IL-4, IL-6, IL-17 and IFNγ), but, in the case of TNFα, we measured approximately equal amounts in Nkx2-3−/<sup>−</sup> and control BALB/c spleen cell cultures, showing that the macrophages were not affected.

In humans, the potential pathogenic role of the Nkx2-3 transcription factor was found in inflammatory bowel diseases (Crohn's and ulcerative colitis) so far [23,24] and also suggested in spondylarthritis more recently [32]. In mice, the absence of Nkx2-3 proved to be protective in DSS-induced colitis through an IL-22-independent mechanism [18]. To our knowledge, this was the first experiment where the Nkx2-3−/<sup>−</sup> mutation was investigated in the context of autoimmune arthritis. Based on the data presented in this study, Nkx2-3 is not only involved in intestinal inflammatory diseases, but also affects autoimmune arthritis.

An important finding of the present study was that not all Nkx2-3−/<sup>−</sup> mice developed arthritis upon rhG1 immunization: in all experiments, 20–40% of Nkx2-3−/<sup>−</sup> mice remained healthy contrary to BALB/c mice (>90% incidence, as seen here and in previous studies [5,27]). To decipher what could be the reason why certain immunized mice did not develop arthritis, we analyzed their immune response parameters separately. We measured considerably lower anti-rhG1 and anti-CCP IgG1 antibody levels in the sera of nonarthritic than in the sera of arthritic Nkx2-3−/<sup>−</sup> mice, respectively. This alone could explain the differences in the arthritis, since it has been shown in several earlier studies that the serum antibody levels against the proteoglycan aggrecan (and its immunodominant region: G1 domain) and CCP show the strongest correlation with the severity of GIA [33,34]. Furthermore, in nonarthritic Nkx2-3−/<sup>−</sup> mice, the serum concentrations of IL-1β, IL-4 and IL-6 were lower than in arthritic Nkx2-3−/<sup>−</sup> mice. We hypothesize that in those Nkx2-3−/<sup>−</sup> mice, which did not develop arthritis, the B cell activation and/or antigen presentation was inadequate to induce sufficient antibody production and autoreactive T cell activation. However, further and more detailed investigation would be needed to adequately answer this question.

Even in those Nkx2-3−/<sup>−</sup> mice which did develop GIA, there were significantly milder symptoms (lower clinical scores, lesser edema, radiologically decreased cartilage and bone destruction). This was in line with those serum parameters which were clearly different from the control BALB/c mice. Specifically, although the anti-rhG1 and anti-CCP-IgG1 antibody levels were similar in the arthritic Nkx2-3−/<sup>−</sup> mice and the controls, the anti-CCP-IgG2a antibodies were produced in lesser amounts. Additionally, the concentrations of signature cytokines of GIA [5], and likewise, RA [2], IFNγ and IL-17 were markedly lower, whereas the concentrations of IL-1β, IL-4 and IL-6 were markedly higher in Nkx2-3−/<sup>−</sup> than in BALB/c mice. Overall, these markers suggest a stronger Th2 activation (primarily indicated by IL-4 and anti-CCP IgG2a) in Nkx2-3−/<sup>−</sup> mice instead of the characteristic Th1/Th17 dominated immune response seen in GIA of BALB/c mice [5,8,28]. Since Th2 cytokines have primarily

anti-inflammatory effects [11] and have been shown to ameliorate PGIA [35], we suggest that this slight shift towards Th2 measured in Nkx2-3−/<sup>−</sup> mice might explain the milder arthritis.

As described here, and earlier [5], similarly to RA, in GIA mice significant anti-CCP antibody production can be detected. The process of citrullination and the role of autoantibodies against these modified antigens is of special interest in RA. ACPA production is promoted by environmental factors like smoking and genetic predispositions such as *HLA-DRB1* [36]. Several studies reported that ACPA positive patients are prone to have more joint erosions [36,37]. ACPA form immune complexes with the citrullinated peptides leading to the activation of macrophages and proinflammatory cytokine production, as well as osteoclastogenesis [36,37]. ACPA is highly specific to RA and can be early detected, even before the onset of RA. Therefore, it is used as a specific diagnostic marker for RA [38]. Recently, it was also found that the pathogenic ACPA are hyperglycosylated which might be regulated by Th17 cells [37]. Since in GIA Th17 activation was observed it is tempting to speculate that altered glycosylation might also occur which could contribute to the pathologic immune reaction against the cartilage antigen components.

Finally, to find a cellular mechanism in the background of the above-detailed immune response differences, we turned our attention to the activation of B cells. B cell activation starts with the engagement of the BcR by the antigen and followed by a well-characterized line of biochemical events including the phosphorylation of cytoplasmic signaling proteins and the transient elevation of the cytoplasmic Ca2<sup>+</sup>-level [11]. We studied the latter and found that in mesenteric and inguinal lymph node B cells of Nkx2-3−/<sup>−</sup> mice, in vitro stimulation with both anti-mouse IgM and anti-mouse IgG caused a weaker Ca2+-signal compared to BALB/c controls. At the same time, the Ca2<sup>+</sup>-signal in Nkx2-3−/<sup>−</sup> T cells remained unchanged, showing that this was a B-cell-specific alteration. This decreased B cell activation capacity of Nkx2-3−/<sup>−</sup> mice was also seen on phospho-blots from some preliminary experiments (data not shown), however, further analysis is needed to clarify which protein(s) could be involved in the signaling changes. These data are in harmony with those described earlier: increased expression of Nkx2-3 in B cells led to Syk and Lyn phosphorylation, elevated basal Ca2+-level and anti-IgM-induced Ca2+-signal [25], the exact opposite to what we found here, in the absence of Nkx2-3. The decreased antigen-driven B cell activation could lead to weaker B cell proliferation and differentiation which, in turn, could explain the less pronounced immune response and the consequently milder autoimmune arthritis.

In conclusion, the complex immune response changes in Nkx2-3−/<sup>−</sup> mice due to the defective spleen structure and function could explain the reduced severity and lower incidence of autoimmune arthritis. At the cellular level, we found weaker B cell activation which might, at least in part, be responsible for the altered immune response. These data add to our knowledge about the significance of the spleen in the development of autoimmunity, and hopefully serve as a starting point for future studies.
