Targeting CD22 with the monoclonal antibody epratuzumab modulates human B-cell maturation and cytokine production in response to Toll-like receptor 7 (TLR7) and B-cell receptor (BCR) signaling

B cells are central mediators of humoral immunity and play a key role in the protection against pathogens. However, aberrant B-cell activation is a common feature of many autoimmune diseases including systemic lupus erythematosus (SLE), primary Sjögren’s syndrome (pSS), and rheumatoid arthritis (RA). B cells may contribute to the pathogenesis of autoimmune diseases through several mechanisms, including both antibody (Ab)-dependent functions (e.g., secretion of autoantibodies (auto-Abs)) and Ab-independent functions, such as antigen (Ag) presentation to T cells and production of pro-inflammatory cytokines [1, 2].

SLE is a complex, systemic autoimmune disease with highly diverse clinical manifestations [3]. The hallmarks of the disease are polyclonal B-cell activation, production of auto-Abs against nuclear-containing Ags, and organ damage due to immune complex deposition. Abnormalities in various B-cell compartments have been described in SLE patients, including expansion of newly formed (transitional) cells, as well as an expansion of Ag-experienced CD27⁺/IgD^– switched memory B cells and CD27^–IgD^– double-negative (DN) memory B cells [2, 4‐6]. Previous studies have shown that the frequency of circulating CD20^–CD27⁺⁺ plasma cells correlates with SLE disease activity and auto-Ab titers [7]; however, the cellular origins of Ab-producing cells in SLE as well as the signals that drive their activation are not well defined. Alterations in the balance of signals downstream of the B-cell receptor (BCR) and pro-survival signals may contribute to the loss of immune tolerance and to the survival and activation of autoreactive B cells [8, 9]. Dysregulation of BCR signaling, including increased phosphorylation of Syk and decreased phosphorylation of phosphatase activities, has also been described in SLE patients [10, 11].

B-cell activation via innate immune receptors, including endosomal Toll-like receptors (TLRs) expressed in B cells, may play a key role in driving autoreactive B cells in SLE [9, 12‐15]. In mice, RNA-associated autoantigens activate autoreactive B cells by engaging BCRs and TLR7, an endosomal TLR specialized in the recognition of viral ssRNA, and induce the production of auto-Abs [12, 16]. Data from murine lupus models further support a role for TLR7 in SLE pathogenesis. Extra copies of the Tlr7 gene drive lupus-like disease [17‐19]; whereas lupus-prone Tlr7-deficient mice develop attenuated disease symptoms [20, 21]. Importantly, B-cell-intrinsic TLR7 signals can drive cell proliferation and differentiation and amplify auto-Ab production to further exacerbate SLE disease in mice [22]. In humans, genetic studies have demonstrated a correlation between genetic variations associated with dysregulation of TLR7 expression and SLE susceptibility [23‐25].

Despite significant evidence implicating TLR7 in SLE, the effects of TLR7 on human B cells have not been explored fully. Little is known about the signals that can regulate TLR7-mediated activation of human B cells. IFN-α can promote human B-cell responses to TLR7 ligation, most likely through its ability to upregulate TLR7 receptor levels [26]. In-vitro stimulation of CD19⁺CD27^– blood B cells with a synthetic TLR7 ligand induces IgM and IgG production as well as secretion of IL-6 and IL-10 [27]. TLR7 activation also expands IgM⁺CD27⁺ memory B cells and CD27^hi B cells, and a combination of TLR7 plus IFN-α promotes the production of auto-Abs [28]. Thus, TLR7-mediated activation of human B cells, as with mouse B cells, can induce the production of auto-Abs.

While signals from BCR and TLR7 can synergize and promote inappropriate activation of autoreactive B cells, the engagement of other surface receptors, such as CD19 and CD22, have been proposed to inhibit their activation [29, 30]. The CD22 Siglec receptor family member is expressed predominantly on B cells and binds via its extracellular ligand-binding domain to α2-6-linked sialic acids on glycoproteins expressed on the same cell (cis interactions) or on opposing cells and/or soluble proteins (trans interactions) [31, 32]. CD22 acts as an adhesion receptor and functions to regulate B-cell migration [33‐35]. Crosslinking of CD22 and the BCR triggers phosphorylation of the CD22 cytoplasmic tail, leading to the activation of a number of signaling molecules, known to either inhibit the BCR signaling or to promote the activation of JNK/SAPK and mitogen activated protein kinase ERK2 [30, 36, 37]. In addition to its function in regulating BCR signaling, CD22 has been implicated in the regulation of TLR-mediated signaling in B cells [38]. CD22^–/– B cells have hyperactive responses to TLR stimulation compared to wild-type (WT) B cells [38, 39]. Furthermore, studies have shown that LPS-induced activation of nuclear factor-κB (NF-κB) downstream of TLR4 is inhibited by the expression of CD22 [38].

The expression of both CD22 and its ligands vary according to the B-cell maturation/activation state. In the periphery, CD22 is expressed at maximum density on human CD27^– naïve and transitional B cells, while it is downregulated by plasma cells [40, 41]. CD22 availability on the cell surface is also dependent on masking or unmasking of CD22 by endogenous (cis) ligand interactions [42]. The expression of CD22 ligands on human B cells is less well explored, but recent studies have shown that, due to changes in glycosylation, germinal center (GC) B cells lose the expression of high-affinity CD22 ligands, leading to CD22 “unmasking” [43].

The development of monoclonal Abs designed to target human CD22 [44] as well recent advances in human B-cell phenotyping [45] provide new opportunities to explore the effects of CD22 engagement on different subsets of human B-cell subsets. Epratuzumab (Emab), a humanized anti-human CD22 IgG1 Ab, has previously shown promising clinical activity both as a single agent and in combination with rituximab in patients with non-Hodgkin’s lymphoma (NHL) [46]. Unlike rituximab, which depletes circulating B cells, Emab does not induce complement-dependent cytotoxicity or Ab-dependent cellular cytotoxicity [47]. CD22 ligation by Emab, however, provokes rapid internalization and phosphorylation of CD22, inhibits the phosphorylation of Syk and PLCγ2, and reduces intracellular Ca²⁺ mobilization after BCR stimulation in vitro [44, 48, 49]. Given the role of CD22 in modulating both BCR and TLR signaling, targeting CD22 with Emab has also been explored as a therapy for autoimmune diseases, including SLE and pSS [50, 51]. Phase I and IIb clinical trials have demonstrated clinically relevant, sustained improvements in patients with moderate-to-severe SLE and a good safety profile of Emab [52, 53]. Emab treatment was found to induce a partial reduction of circulating B cells in SLE patients affecting primarily CD27^– cells [54], a phenomenon that later led to an exploration of the in-vitro effects of Emab on the expression of the adhesion molecules CD62L, β7 integrin, and β1 integrin and on B-cell migration toward CXCL12 [41]. CD22 binding by Emab also induces a reduction of CD19, CD21, and CD79b expression through a process known as trogocytosis (i.e., Fc-mediated receptor “shaving” to other effector cells). In line with these findings, a decrease in CD19 expression was observed in SLE patients treated with Emab [55]. Recent studies have also investigated the effects of Emab on cytokine production. Emab inhibited IL-6 and tumor necrosis factor alpha (TNF-α) production of blood B cells isolated from healthy donors and SLE patients in response to BCR crosslinking alone or in combination with TLR9 ligand CpG [56]. Overall, the available data led to the hypothesis that the primary mode of action of Emab is to enhance the normal inhibitory role of CD22 on B-cell activation [57].

Given the importance of TLR7 signaling in activating autoreactive B cells in SLE, we investigated whether CD22 crosslinking by Emab might affect B-cell activation in response to BCR and/or TLR7 stimulation. We further aimed to identify which subpopulation of human B cells might be affected to the greatest extent by Emab in the context of BCR/TLR7 stimulation. Using human tonsillar CD10^–CD27^– B cells, we found that Emab modulates cytokine production by inhibiting IL-6, while at the same time enhancing IL-10 production. Emab dramatically and selectively inhibited levels of PRDM1, the gene encoding Blimp-1 in CD10^–CD27^– tonsillar B cells, activated by either TLR7 signaling alone or in combination with BCR stimulation. Tonsillar B-cell subsets responded differently to BCR/TLR7 stimulation; among them CD10^–CD27^–IgD^– cells, which correspond to DN memory B cells found in the periphery, were most responsive to TLR7 stimulation, as evidenced by the appearance of CD27^hiCD38^hiBlimp1⁺ plasmablasts. In particular, Emab inhibited the in-vitro differentiation of CD10^–CD27^–IgD^– and CD10⁺CD27^–/+ B cells and reduced the survival of CD10⁺CD27^–/+ B cells but not other B-cell subsets. Thus, Emab can both enhance and inhibit TLR7-driven B-cell responses: depending on the B-cell developmental/maturation state, CD22 crosslinking by Emab affects B-cell survival, activation, and differentiation differently, which may have important implications for the clinical use of Emab and monitoring of CD22-based therapies.

A number of recent studies have demonstrated that TLR signaling in B cells plays an important role in modulation of B-cell effector functions such as cytokine production and ability to produce auto-Abs [13]. Activation of B cells by TLR7 in particular has been proposed to play an important role in SLE [17, 20, 63]. Because of the significance of TLR7 in the pathogenesis of SLE, we assessed how targeting CD22 with Emab affected B-cell responses after TLR7 and/or BCR stimulation. We used human tonsillar B cells, which provided the means to investigate the responses of B cells from different developmental/activation stages.

In our initial studies, we used tonsillar CD20⁺CD10^–CD27^– enriched B cells, which in many respects resemble CD27^– naïve blood B cells [64]. Because Emab has previously been shown to modulate signaling downstream of the BCR [48, 49, 57], we investigated whether Emab might also affect the expression of genes known to be activated through the BCR. While BCR or TLR7 stimulation or a combination of the two stimuli induced upregulation of a number of genes, including cMYC and BCL _XL, Emab did not affect their expression. Similarly, Emab showed no significant effect on the expression of TLR7 and TLR9 or of MyD88 and IRF7 genes that encode proteins downstream of TLR7 signaling. In another set of experiments, naïve tonsillar B cells were incubated with IFN-α or a combination of IFN-α and anti-IgM. While this stimulation led to a significant upregulation of multiple genes, and in particular those involved in TLR signaling, the expression of these genes was not affected by Emab (data not shown).

However, TLR7/BCR-driven IL-6 and IL-10 cytokine production by CD20⁺CD10^–CD27^– B cells were differentially modulated by Emab with a significant increase in IL-10 production and an overall inhibition of IL-6. Analysis of IL-6 and IL-10 transcripts 6, 12, or 24 hours after TLR7/BCR stimulation showed no significant effect of Emab on gene expression (data not shown); IL-10 mRNA expression was only weakly increased after TLR7/BCR stimulation and again was not affected by Emab. Thus, the exact mechanism for Emab-induced modulation of cytokine production remains to be determined. Recently, Fleischer et al. [56] analyzed the effects of Emab on the cytokine production by purified total blood B cells from healthy donors and SLE patients. They found that Emab significantly decreased the expression of TNF-α and IL-6 in response to anti-BCR and/or anti-BCR/CpG stimulation. Although anti-BCR/CpG stimulation also induced significant increase of IL-10 production, the authors found that IL-10 levels were not affected by Emab.

In contrast, we found that Emab significantly enhanced IL-10 production by naïve tonsillar B cells in response to anti-IgM/TLR7 stimulation. Interestingly, the capacity of B cells to produce IL-10 correlated positively with the expression levels of TLR7 at the time of stimulation. In this regard, we found that prestimulation of the cells with IFN-α, a known inducer of TLR7 [26], was able to boost IL-10 production, which was further enhanced in the presence of Emab. The disparity between our results and those of Fleisher et al. [56] may be due to different sources of B cells being tested, use of a TLR7 vs a TLR9 agonist, and/or the levels of TLR9 or TLR7 in the B cells that were examined. We observed that IL-10 production was largely dependent on TLR7 activation by R848 and less dependent on BCR activation, a similar conclusion reached by Fleischer et al. although that they used a TLR9 ligand. In summary, in the majority of our experiments, Emab displayed a differential effect on cytokine production by suppressing IL-6 and inducing IL-10.

The effects of Emab on cytokine production may contribute to the understanding of the mode of action of Emab in vivo, particularly in autoimmune diseases, where B-cell-produced cytokines are believed to play an important role in driving or suppressing the disease [1]. For example, IL-6 has been implicated in B-cell differentiation and Ab production [65, 66], and is also known to cooperate with IL-21 to promote the differentiation of CD4⁺ T-follicular helper cells [67‐70]. Furthermore, increased IL-6 levels have been reported in SLE patients with active disease and more recently IL-6 has been identified as a major genetic risk factor for SLE [71]. Emab-mediated IL-6-inhibition could possibly suppress inflammation associated with SLE. IL-10, on the contrary, is known to be produced by regulatory B cells [72] and has been proposed to suppress effector Th1 and Th17 cell responses [73‐75]. Recent studies have suggested that IL-10 production by B cells might be defective in SLE patients [74, 76]. Based on our findings, CD22 engagement by Emab may very well help to reprogram B cells in SLE patients to restore IL-10 production.

Further studies are required to determine how CD22 signaling promotes IL-10 production by TLR7-driven B cells. Liu et al. [77] have recently shown that activation of STAT3 and ERK is required for TLR-induced IL-10 production by human B cells. The authors also found that IFN-α enhanced TLR7/8-induced but not TLR9-induced IL-10 production. Although it is well known that TLR stimulation elicits IL-10 production, to our knowledge this study is the first to show a role for CD22 in promoting IFN-α/TLR7-induced IL-10 production. Whereas CD22 is often described as a negative regulator of BCR signaling, it should be noted that previous studies have shown that CD22 associates with a number of signaling molecules, such as Syk, PI-3 kinase, Grb2, and phospholipase-Cγ2 [78‐80], and that direct CD22 engagement can induce activation of ERK2 [81]. In line with this finding, we have found that Emab induces increased ERK phosphorylation in human B cells [82]. In light of Liu et al.’s findings [77], this might provide a mechanistic explanation of how CD22 crosslinking by Emab promotes IL-10 production.

Interestingly, Emab also modestly increases B-cell proliferation in response to BCR/TLR7 stimulation. This effect of Emab might be selective to the CD27^–CD10^– (naïve and DN) cells and also dependent on the particular signals used to activate the cells. In contrast to CD27^– (naïve/DN) B cells, Emab did not affect cell proliferation of CD27⁺ (classical memory) cells; however, the rates of cell proliferation were variable between different donors (data not shown). Previous studies have shown that Emab can inhibit the proliferation of CD27^– and CD27⁺ blood B cells in response to IL-2, IL-10, F(ab′)₂, and/or CD40L and CpG [54]. Thus, the effect of Emab on cell proliferation among various B-cell subsets needs further investigation. In the context of BCR/TLR7 stimulation of the CD27^–CD10^– subset, the small increase of cell proliferation may be related to our finding that Emab inhibits B-cell differentiation. Previous studies have shown that B cells exit the cell cycle once they start to differentiate into Ab-producing plasma cells and a change in the expression of several transcription factors (e.g., a decrease of BCL6, PAX5, and c-Myc levels and an increase of Blimp1 expression) has been implicated in the transition/differentiation of naïve B cells into plasma cells [62]. In CD10^–CD27^– B-cell cultures, we found that Emab dramatically inhibited the levels of PRDM1, the gene encoding Blimp1, which was induced in response to TLR7 and/or BCR/TLR7 stimulation. This effect of Emab was highly specific to PRDM1 because this Ab did not affect the expression of other genes involved in B-cell differentiation or immunoglobulin class-switching, including MITF, BACH2, BCL6, PAX5, IRF4, XBP1, AICDA, or TBX21. We did see a trend toward reduction in the increase TBX21 in response to anti-IgM/TLR7 in the presence of Emab, but the combined results from different experiments did not show statistical significance.

The responses of human B-cell subsets to TLR7 or anti-BCR/TLR7 stimulation have not been studied in detail. Recently, Simchoni et al. [28] showed that TLR7 stimulation expands IgM⁺CD27⁺ memory B cells and promotes the generation of CD27^hi B cells. In the current study, by comparing four different B-cell populations based on the appearance of CD38^hiCD27^hi cells, we discovered that the CD10^–CD27^–IgD^– subset was the most responsive to TLR7 stimulation compared to the other B-cell subsets. Notably, the generation of CD38^hiCD27^hi cells was inhibited in the presence of Emab. A portion of CD10⁺CD27⁺ and CD10⁺CD27^–/+ B cells also differentiated into CD38^hiCD27^hi cells, but their frequencies were lower compared to those generated by CD10^–CD27^–IgD^– B cells. Emab clearly decreased CD38^hiCD27^hi cell frequencies, particularly within the CD10⁺CD27^–/+ cells. CD10^–CD27^–IgD⁺ (naïve) B cells did not generate CD38^hiCD27^hi cells in response to TLR7 stimulation. Of note, there were no significant differences in TLR7 expression between CD10^–CD27^–IgD^– and CD10^–CD27^–IgD⁺ B cells, suggesting that their differential responsiveness to TLR7 ligation cannot be simply attributed to increased TLR7 levels. Furthermore, a comparison between these two cell populations showed no differences in their CD22 expression and/or their ability to internalize Emab upon CD22 binding (NV Giltiay, unpublished data). Data from mouse and human studies have indicated that TLR7 can promote the production of auto-Abs [13, 28]; because Blimp1 is required by B cells to mature into Ab-producing cells, we propose that Emab-mediated inhibition of Blimp1 may reduce Ab/auto-Ab production. While the mechanisms for Emab-induced inhibition of B-cell differentiation need further elucidation, our data suggest that one therapeutic effect of Emab may be via inhibiting the expression of PRDM1 (Blimp1). In this respect, it is highly relevant that there is elevated expression of Blimp1 in SLE patients and this is correlated with increases in plasma cells, auto-Abs, and disease activity [83].

Interestingly, Emab also affected cell survival, although this effect was only evident within the (pre-GC/GC) CD10⁺CD27^+/– cell population. A possible explanation for these data is that cells within this mixed cell population might be more sensitive to CD22-induced apoptosis. While apoptosis induction is not considered the primary mode of action of Emab, previous in-vitro data have demonstrated that crosslinking of CD22 by mAbs including Emab induces apoptosis in human lymphoma cells [84, 85]. Recently, Macauley et al. [43] demonstrated that changes in the glycosylation patterns due to altered enzyme activity in the GC leads to “unmasking” of CD22 binding site on GC B cells, relative to naïve and memory B cells. Such unmasking could very well alter the effects of CD22 binding by Emab. It should be noted that the effects of Emab on GC cell survival in culture were independent of BCR/TLR7 stimulation. These data, however, may be relevant to the clinical effects of Emab in patients with NHL, because GC B cells are considered a major source (i.e., cell of origin) for lymphoma cells [46, 86].

Despite initial promising data in phase II clinical studies [52, 53], results reported recently from a phase III clinical trial in SLE patients with moderate to severe disease showed that Emab failed to reach the primary clinical endpoint [87]. It has been suggested that a high placebo response and early rescue of nonresponders with increased doses of glucocorticoids might have confounded the data from the trial [87, 88]. Although disappointing, these results reflect to a large extent the complexity and diversity of the pathology of SLE and suggest that perhaps some, but not all, SLE patients would benefit from Emab therapy. In our study, we have found that Emab affects the production of cytokines in response to BCR/TLR7 stimulation, by skewing B cells to produce immunoregulatory cytokines such as IL-10. Thus, we can predict that targeting CD22 with Emab might be able to restore IL-10 production by CD27^–CD24^hiCD38^hi transitional B cells in SLE patients. Transitional B cells are expanded in some SLE patients and our data showed that these cells express relatively high levels of CD22, suggesting that this cell population might be a good target for Emab therapy.

Our study identified CD10^–CD27^–IgD^– as another important cell population, whose responses to BCR/TLR7 stimulation were affected by Emab. We believe this cell population closely resembles the DN memory B cells found in the blood. We found a significant variation in the frequency of this B-cell population in tonsils, which we think might be reflective of environmental factors, such as recent viral exposure. An increased frequency of DN memory B cells has been described previously in SLE patients [4, 5]. Although there has been no direct evidence for their contribution to disease pathology, DN memory B cells have been linked to SLE autoimmunity [4, 5]. For example, a significant portion of DN memory cells found in SLE patients produce VH4-34-encoded 9G4 Abs, known to be a source of SLE-associated autoreactivity [89]. Frequencies of DN memory B cells were found to be associated with higher disease activity, history of nephritis, and presence of auto-Abs [4, 5]. The cellular origin of DN memory B cells remains somewhat elusive; because these cells express switched isotypes, yet at the same time show a reduced rate of somatic hypermutations compared to post-switched memory B cells, it has been proposed that they might be of extra-GC origin [5]. Recent studies have shown that a substantial fraction of Ab-secreting cell clones found during SLE flares contained auto-Abs without (or with very few) mutations, consistent with differentiation outside GCs [90].

Whether TLR-mediated activation of DN B cells might contribute to the generation of Ab-secreting clones has not been established. One study has shown that DN memory B cells are also highly responsive to stimulation through TLR9 by CpG [4]. Here, we reported that these cells are also highly responsive to stimulation through TLR7 by R848 or a combination of anti-IgM and TLR7. Importantly, we demonstrated a new role for CD22 in regulating the activation of these cells and their differentiation into CD27^hiCD38^hiBlimp1⁺ plasmablasts. In this study, we did not address in detail whether DN cells produce cytokines. In a limited set of experiments, we found that DN cells produced lower levels of IL-6 and IL-10 in response to TLR7 and or BCR/TLR7 stimulation, compared to naïve B cells; however, this needs to be investigated further. Recent reports describe an increase in DN memory B cells in RA patients and older people [91, 92]. Although the link between DN B cells and SLE disease activity seems well established, we have found a significant variation of DN memory B-cell frequencies in SLE patients, even those with active disease (NV Giltiay, unpublished data). In light of these findings, one can predict that patients with high frequencies of DN memory B cells might be better candidates for CD22 targeting by Emab.The fact that Emab blocks B cell differentiation in a distinct subset of memory B cells is also consistent with our studies in mice showing that CD22 is required for normal memory B cell formation [93].

Recent data suggest that Blimp1 levels positively correlate with the levels of pathogenic auto-Abs in SLE patients and can be used as a potential biomarker for monitoring disease activity [83]. Although results from the EMBODY1 and EMBODY2 trials showed no significant reduction of the total IgG levels and/or anti-DNA titers in Emab-treated patients [87], it would be of interest in the future to test whether decreases in Blimp1 levels and auto-Ab titers are associated with responses to Emab in vivo.

Together, our data show that binding of CD22 by Emab can both regulate cytokine expression and block Blimp1-dependent B-cell differentiation in response to BCR/TLR7 stimulation. The effects of Emab depend on the stage of B-cell development or activation. A better understanding of the differential effects of Emab on different B-cell populations, including its effects on cell survival and cell activation, would provide important information on how best to use Emab in a clinical setting. We found that Emab inhibits the activation of CD27^–IgD^– (DN) cells which are highly responsive to stimulation via TLR7 and which are known to be expanded in some but not all SLE patients. The therapeutic effect of Emab in SLE might be via modulating the production of pro-inflammatory and anti-inflammatory cytokines by B cells, or by inhibiting the survival and/or differentiation of specific B-cell populations. Because Blimp1 is required by B cells to mature into Ab-producing cells, CD22-mediated inhibition of Blimp1 may also reduce auto-Ab production. Intriguingly, given the failure of Emab to demonstrate efficacy in a general SLE patient population, levels of Blimp1 expression could possibly be used as a biomarker to predict clinical responses to Emab in SLE and in patients with other autoimmune diseases.