Protection against collagen-induced arthritis in mice afforded by the parasitic worm product, ES-62, is associated with restoration of the levels of interleukin-10-producing B cells and reduced plasma cell infiltration of the joints

We have previously reported that ES-62, a molecule secreted by the parasitic filarial nematode Acanthocheilonema viteae, protects mice from developing collagen-induced arthritis (CIA). Together with increasing evidence that worm infection may protect against autoimmune conditions, this raises the possibility that ES-62 may have therapeutic potential in rheumatoid arthritis and hence, it is important to fully understand its mechanism of action. To this end, we have established to date that ES-62 protection in CIA is associated with suppressed T helper type 1 (Th1)/Th17 responses, reduced collagen-specific IgG2a antibodies and increased interleukin-10 (IL-10) production by splenocytes. IL-10-producing regulatory B cells have been proposed to suppress pathogenic Th1/Th17 responses in CIA: interestingly therefore, although the levels of IL-10-producing B cells were decreased in the spleens of mice with CIA, ES-62 was found to restore these to the levels found in naive mice. In addition, exposure to ES-62 decreased effector B-cell, particularly plasma cell, infiltration of the joints, and such infiltrating B cells showed dramatically reduced levels of Toll-like receptor 4 and the activation markers, CD80 and CD86. Collectively, this induction of hyporesponsiveness of effector B-cell responses, in the context of the resetting of the levels of IL-10-producing B cells, is suggestive of a modulation of the balance between effector and regulatory B-cell responses that may contribute to ES-62-mediated suppression of CIA-associated inflammation and inhibition of production of pathogenic collagen-specific IgG2a antibodies.


Introduction
The prevalence of autoimmune disease in the developing world inversely correlates with parasitic helminth infection, 1,2 an apparent serendipitous side-effect of the ability of worms to secrete anti-inflammatory molecules that modulate the host immune system and promote parasite survival. Such epidemiological evidence has understandably generated interest in the therapeutic potential of such immunomodulators and also in their use as tools to dissect the pathogenic mechanisms underlying inflammatory disorders. 3 We have previously shown that one such immunomodulator, ES-62, a phosphorylcholine-containing glycoprotein, secreted by the filarial nematode Acanthocheilonema viteae, is protective in the mouse collagen-induced arthritis (CIA) model of rheumatoid arthritis (RA) by acting to suppress pathogenic T helper type 1 (Th1)/Th17 responses. 4 The success of rituximab as a therapy for autoimmune disease has refocused interest on the pathogenic and protective roles of B cells in RA 5 with recent studies highlighting the importance of interleukin-10 (IL-10)-producing regulatory B (Breg) cells in the prevention and amelioration of CIA and also antigen-induced arthritis (AIA), via the suppression of Th1/Th17 responses and the promotion of type 1 regulatory T (Tr1) and regulatory T (Treg) cell differentiation. [5][6][7][8] Interestingly, therefore, we have found that while exposure to ES-62 results in enhanced spontaneous ex vivo production of IL-10 by splenocytes from mice with CIA, it induces hyporesponsiveness of normal and CIA-derived splenic B cells and reduces the levels of pathogenic collagen-specific IgG2a antibodies. 9 Relating to this, we now show that the protective effects of ES-62 in CIA are associated with restoration of the levels of IL-10-producing B cells and suppression of the infiltration of activated plasma cells into the joints, perhaps suggesting that ES-62 may act, at least in part, to modulate the balance between effector and regulatory B-cell responses in this mouse model of RA.

Collagen-induced arthritis
Animals were maintained in the Biological Services Units in accordance with the Home Office UK Licences PPL60/ 3580, PPL60/3791, PPL60/4300 and PIL60/9576, PIL60/ 12183 and PIL60/12950 and the Ethics Review Boards of the Universities of Glasgow and Strathclyde. Arthritis was induced in male DBA/1 mice (8-10 weeks old; Harlan Olac, Bicester, UK) by intradermal immunization with bovine type II collagen (MD Biosciences, Zurich, Switzerland) in complete Freund's adjuvant on day 0 and in PBS on day 21. Mice with CIA were treated with purified endotoxin-free ES-62 (2 lg/dose) or PBS subcutaneously on days À2, 0 and 21 and cells were recovered from joints 10 as previously described. 4,9,11 All analysis was performed at cull (day 28) and represents data from at least two independent experiments.

Statistics
Parametric data were analysed by the Student's t-test or by one-way analysis of variance (ANOVA) while non-parametric data were analysed by Mann-Whitney and Kruskal-Wallis tests where *P < 0Á05, **P < 0Á01 and ***P < 0Á001.

ES-62 modulates the recruitment of B cells to the joints of mice with CIA
As pathogenic B cells migrate to the joints and even form ectopic GCs in response to B-cell recruitment and survival factors such as BAFF, CXCL12 and CXCL13 secreted by synovial fibroblasts, 5 we next analysed whether the above effects of in vivo exposure to ES-62 on the profile of B cells were reflected in the arthritic joint. This revealed that both the proportion (Fig. 3a,b) and absolute numbers ( Fig. 3c) of CD19 + B cells found in the joints were significantly reduced by ES-62 treatment. This reduction was reflected in a CD19 + CD23 + B-cell population (Fig. 3d,e), which further analysis revealed to be Fo1 B cells (Table 1). There was also a clear decrease in CD19 À CD138 + (from 9Á27 to 2Á45% live cells) and CD19 + CD138 + (from 15Á6 to 4Á51% live cells) cells infiltrating the joints of mice treated with ES-62 (Fig. 3f,g), which suggested a reduction in plasma cells. Consistent with this, further analysis, excluding the myeloid and T-cell lineages expressing CD138 (Fig. 3h), revealed that exposure to ES-62 indeed suppressed the proportions (Fig. 3i, j) and numbers (Table 1) of CD19 À B220 À CD138 + (from 8Á31 to 3Á69% live cells) and CD19 + B220 low/À CD138 + (from 1Á37 to 0Á72% live cells) plasma cells, which respectively are phenotypically similar to the long-lived plasma cell and short-lived plasma cell/plasmablast functional populations, reported previously. [16][17][18] This presumably reflects reduced development and/or migration of such cells, as suggested by the significant increases in the levels of Fo1 (Fig. 2e) and CD19 À B220 À CD138 + plasma cells (numbers (9 10 6 ) AE SEM: Naive, 0Á75 AE 0Á22; PBS, 1Á28 AE 0Á31; ES-62, 1Á58 AE 0Á26) found in the spleen, as ES-62 did not modulate the levels of early CD19 + B220 + CD138 + 'preplasma cells', which have been reported as being subject to a tolerance checkpoint that is defective in the autoimmune-prone MRL/Lpr mouse 19 (results not shown).
In addition to investigation of modulation of the levels of B cells found in the joint, ES-62 was assessed for effects on the functional phenotype of such infiltrating cells. Hence, although ES-62 did not significantly modulate the expression of BAFF-R, CD80, CD86 or TLR4 (either in terms of percentage positive cells or levels of expression) on splenic CD19 + B cells, expression of CD80, TLR4 and, to a lesser extent, CD86, but not BAFF-R, by CD19 + B cells in the joint was down-regulated, mirroring a tendency towards a decrease in expression of both CD80 and TLR4 by CD19 + DLN cells (results not shown). Further analysis revealed that this reflected selective down-regulation of TLR4, CD80 and CD86 expression on CD19 + CD23 high CD21 low Fo B cells but not CD19 + CD23 À CD21 high or CD19 + CD23 high CD21 high B cells ( Fig. 4 and results not shown). Moreover, although B1, GC and plasma cells also expressed TLR4, such expression was not modulated by exposure to ES-62 (results not shown). Collectively, therefore, these data suggest that ES-62 may act to prevent development and migration of pathogenic B cells to the site of inflammation, with the residual B cells that infiltrate the joint being rendered functionally hyporesponsive.
ES-62 restores the levels of IL-10-producing B cells in the spleen of mice with CIA Interleukin-10-producing B cells that exhibit regulatory activity (Breg cells) have been reported to curb pathogenic Th1/Th17 responses in CIA and it has been proposed that these cells protect against disease by promoting the development of Tr1 cells. 6 Although the ES-62mediated suppression of B-cell development and migration (Figs 2 and 3) and consequent inhibition of generation of pathogenic IgG2a responses observed in CIA 11 might reflect the ability of the parasitic worm product to induce hyporesponsiveness of B2 cells by uncoupling the B-cell receptor from extracellular signal regulated kinase/mitogen-activated protein kinase signalling, both in vitro and in vivo, 20,21 it is also interesting to note that ES-62 has previously been shown to induce the production of IL-10 by peritoneal B1 B cells. 22 As B1 cells are not thought to play a central role in systemic autoimmunity and autoantibody production in CIA, 23 it is therefore possible that the reduced levels of CD19 + IgM + CD43 high CD5 + and CD19 + IgM + CD43 high CD5 À cells, which have been reported to be B1a and B1b cells, respectively, 24-28 found in the spleens (Fig. 5a-c) and DLNs (Fig. 5d) of ES-62-treated mice with CIA could reflect their egress and migration Data (number of cells 9 10 6 ) are presented from two independent experiments where CD19 + B220 low/À CD138 + and CD19 À B220 À CD138 + plasma cell and CD19 + CD21 low CD23 + AA4Á1 À IgM low IgD high Fol B-cell populations infiltrating the joints were analysed by flow cytometric analysis of joint cells harvested from six or seven mice from each group. Live cells Live cells Figure 3. ES-62 modulates the recruitment of B cells to the joints of mice with collagen-induced arthritis (CIA). Cells extracted from the joints of mice with CIA were analysed for the proportion (a,b) and number (c) of infiltrating CD19 + B cells (data in c are presented as the means AE SEM of four biological replicates pooled from two independent experiments) and consequently for the relative proportions of CD19 + CD23 + (d,e) and plasma cells on the basis of CD19 CD138 expression (f,g). Exclusion of myeloid and T-cell-expressing CD138 cells by use of Dump channel (CD4 + CD8 + GR1 + F4/80 + CD11b + CD11c + ; h) allowed analysis of Dump À CD19 À B220 À CD138 + and Dump À CD19 + B220 low/À CD138 + plasma cells (i,j).
to the joints to mediate IL-10-dependent anti-inflammatory effects. Perhaps consistent with this, although the levels of B cells infiltrating the joint overall were reduced following treatment with ES-62, the proportion of such B1like cells was slightly increased (PBS: 1Á42%, 1Á3%; ES-62: 1Á58%, 1Á51%, where data are expressed as the proportion (%) of live cells, that are CD43 + CD19 + IgM + B1-like cells, harvested from the joints of six or seven mice/group in two independent experiments). Analysis of IL-10-producing B cells (Fig. 6a,b) revealed that although induction of CIA resulted in a significant reduction in the levels of IL-10-producing B cells, relative to naive mice, this was not the case for those treated with ES-62 (Fig. 6c). Indeed, exposure to ES-62 led to a significant enhancement in the level of IL-10-producing B cells in mice with CIA, restoring them to levels comparable with those found in healthy naive mice [ Fig. 6c; numbers (9 10 6 ) AE SEM: Naive, 1Á81 AE 0Á22; PBS, 1Á48 AE 0Á12; ES-62, 1Á99 AE 0Á26]. CD19 + IL-10 + B cells in the spleens of naive and CIA mice reflected a mixed population comprising phenotypes consistent with marginal zone precursor, marginal zone, follicular B cells and CD19 + CD21 À CD23 À B cells (Fig. 6b): almost all of these cells expressed CD1d whereas some 20-30% of the follicular and CD19 + CD21 À CD23 À B cells expressed CD5 (data not shown), the latter a marker previously associated with certain IL-10-producing B cells. 29,30 Rather surprisingly, the ability of ES-62 to return the levels of IL-10-producing B cells towards that existing in naive, non-arthritic mice, did not appear to involve a preferential modulation of any of these phenotypes, perhaps suggesting that it acts rather to regulate B-cell responses in a 'homeostatic' manner. Although the protective effects of IL-10-producing B cells in CIA have previously been reported to be associated with the induction of Tr1 cells, 6 our analysis showed the IL-10 production by splenocytes to be predominantly B-cell-derived (Fig. 6a) and that the CD19 À IL-10 + population was not increased by treatment with ES-62 (Fig. 6d), suggesting that Tr1 cells were not being induced in this case. Likewise, and consistent with the previously reported lack of Breg-mediated induction of Treg cells in the CIA model, 6

Discussion
The successful clinical implementation of B-cell depletion therapies in recent years has reignited interest in the pathogenic and protective roles of B cells in RA. In particular, much interest has focused on the ability of IL-10producing Breg cells to prevent development of pathogenic Th1/Th17 responses and to induce Treg cells that result in the suppression of disease in the AIA (induced by methylated BSA) and CIA mouse models of RA. 6,8,31 Interestingly, therefore it has recently emerged that one of the strategies exploited by parasitic helminths to dampen host immune responses and hence promote their survival is the induction of IL-10-producing Breg cells: moreover, the protection afforded against allergic inflammatory disease by such worms has also been associated, at least in part, with their induction of such Breg cells. 28,[32][33][34][35] Hence, our current findings that exposure to ES-62 results in elevated levels of IL-10-producing B cells may provide a rationale for our previous findings that the parasite product exerts its protective effects in CIA via suppression of Th1, Th17 and IL-17-producing cd T-cell responses, as well as being consistent with our earlier observation that it promotes spontaneous IL-10 production by splenocytes from mice with CIA. 4,9,11 Interestingly, although regulatory B cells have been proposed to mediate at least some of their protective effects in experimental arthritis via the generation of natural Treg cells (AIA) and/or induced Tr1 cells (AIA and CIA), 6,8,31 we have found no evidence that ES-62 induces any IL-10-producing Treg cells in DBA/1 mice with CIA. Although perhaps surprising, this failure to induce Treg cells is supported by our preliminary data from the C57BL/6 model of chronic CIA, 36 in which in vivo exposure to ES-62 does not increase the levels of either FoxP3 + or IL-10-producing CD4 + T cells in the DLN of such mice (results not shown) and is consistent with their lack of induction in our previous studies investigating ES-62-mediated hyporesponsiveness to the model antigen ovalbumin in both the DO.11.10 transfer model 37 and the ovalbumin-induced airway inflammation model of asthma. 38 Nevertheless, we cannot rule out the possibility that while ES-62 does not increase the levels of Treg cells in mice with CIA, it may act to reverse/overcome the impaired (cell contact-mediated) suppressive ability of Treg cells reported in RA. 39 Perhaps also surprisingly, the restoration of IL-10-producing B cells resulting from exposure to ES-62 did not There is no unambiguous phenotype for B1 cells in the spleen but they have been described as CD19 high CD23 À CD43 + IgM high IgD low/À CD5 AE cells where CD5 + B1a and CD5 À B1b comprise~2% and 1% of cells in the spleen, respectively. 24 However, following gating on CD19 + IgM + (a), analysis of CD43 + CD5 AE cells has been widely used to describe B1 cells 25,26 whereas CD19 + CD5 + gating has been used to describe B1a cells. 28 Moreover, although CD43 can be up-regulated on B2 cells, this is usually expressed at a lower level than on B1 cells 27 and we have therefore chosen to gate only on CD43 high cells (b) so as to exclude any potential CD43 + B2 cells. We have therefore phenotyped CD19 high CD43 + IgM high B cells as reflect induction of a particular phenotype of B cells associated with regulatory function: although modest, these statistically significant increases in IL-10-producing B cells are consistent with the numbers seen in other studies in the absence of enrichment by anti-CD40 or lipopolysaccharide/IL-21 stimulation 12,40,41 and suggest that ES-62 may be acting in a homeostatic manner to reset the balance of effector and regulatory B cells back towards that observed in healthy DBA/1 mice. However, although the ES-62-mediated suppression of the level of GC B cells was associated with a reduction of follicular helper T cells, our preliminary data suggest that the residual follicular helper T cells produced slightly higher levels of the cytokine IL-21, which appears to be critical for the generation of functional Breg cells that combat autoimmunity. 41 Hence, as we have not formally demonstrated the regulatory function of these IL-10-producing B cells, we cannot rule out the possibility at this stage that ES-62 is also reversing the defective regulatory function of one or more Breg phenotypes observed in CIA and also in RA patients. 42,43 Nevertheless, we have shown previously that whereas ES-62 induces high levels of IgG1, but not IgG2a, antibodies in naive wild-type BALB/c mice, in IL-10 À/À mice the helminth product induces both IgG1 and IgG2a antibodies. This suggests that IL-10 plays an important role in the suppression of IgG2a antibodies directed against ES-62. 44 Interestingly, therefore, we have shown that neither the ES-62-mediated suppression of DC-priming of OVA-specific Th1 responses 45 nor the suppression of TLR-mediated IL-12 responses of macrophages exposed to ES-62 either in vitro or in vivo 46 is due to autocrine production of IL-10 by APC. By contrast, although we have shown that exposure to ES-62 in vivo induces hyporesponsiveness of splenic B2 cells, peritoneal B cells from such mice produced enhanced spontaneous and B-cell receptor-stimulated IL-10 responses 21,22 and although these peritoneal cells will predominantly comprise B1 cells, there is increasing evidence that they probably also include some B2 cells. 27 Collectively, therefore, these data suggest that exposure to ES-62 leads to hyporesponsiveness of effector B2-cell responses and restoration of IL-10-producing B cells that in CIA, given the complex interplay between B-cell-derived IL-10 and pathogenic IL-17 responses in the regulation of inflammation and autoantibody responses, 6,12,31,41,[47][48][49] results in the reduction of plasma cells that may contribute to the suppression of pathogenic autoantibodies and inflammation associated with the protection against CIA. For example, the ability of B-cellderived IL-10 to impact on antigen-presenting cells such as dendritic cells in vivo may contribute to the effects of ES-62 on the DC-dependent priming of Th1/Th17 and IL-17-producing cd T cells. 4,38,45 Alternatively, such IL-10 may suppress effector B-cell activation with consequent induction of T-cell hyporesponsiveness, as the reduction of GC B cells and follicular helper T cells observed in the spleens of ES-62-treated mice is reminiscent of that reported for mast cell-derived IL-10-mediated suppression of follicular helper T-cell function. 50 This could be particularly important at the site of inflammation with the profound down-regulation of CD80 and CD86 expression on follicular B cells impacting on the functionality of ectopic GCs given that expression of CD80 by B cells has been shown to be important in the regulation of follicular helper T-cell development, and consequent GC B-cell survival and plasma cell production, 51 the latter population being clearly reduced in the joints of ES-62-treated mice. Interestingly therefore, blocking of B7-CD28 interactions has been reported to be sufficient to prevent development of CIA. 52 Finally, the strong down-regulation of TLR4 is similarly likely to suppress B-cell activation and plasma cell generation at this site, 53 and in this way, disrupt the destructive chronic inflammation resulting from cells in the joint expressing up-regulated levels of TLRs, including TLR4 54 and responding to damage-associated molecular pattern molecules, such as heat-shock protein 22 and tenascin-C 55 found in the synovium of RA patients.