• Autoimmune disease;
  • B cells;
  • Infection;
  • Interleukin-10;
  • Plasma cells;
  • Regulation


  1. Top of page
  2. Abstract
  3. Acknowledgments
  4. Conflict of interest
  5. References

B-cell depletion can improve disease in some patients with rheumatoid arthritis or multiple sclerosis, indicating the pathogenic contribution of B cells to autoimmunity. However, studies in mice have demonstrated that B cells have immunosuppressive functions as well, with IL-10 being a critical mediator of B-cell-mediated suppression. IL-10-secreting B cells have been shown to promote disease remission in some mouse models of autoimmune disorders. Human B cells also produce IL-10, and evidence is accumulating that human IL-10-producing B cells might inhibit immunity. There is considerable interest in identifying the phenotype of B cells providing IL-10 in a suppressive manner, which would facilitate the analysis of the molecular mechanisms controlling this B-cell property. Here, we review current knowledge on the B-cell subpopulations found to provide suppressive functions in mice, considering both the pathological context in which they were identified and the signals that control their induction. We discuss the phenotype of B cells that have IL-10-dependent regulatory activities in mice, which leads us to propose that antibody-secreting cells are, in some cases at least, the major source of B-cell-derived regulatory IL-10 in vivo. Anti-inflammatory cytokine production by antibody-secreting cells offers a novel mechanism for the coordination of innate and humoral immune responses.

The initial observations demonstrating a suppressive role for IL-10-producing B cells in autoimmune diseases were made in mouse models of ulcerative colitis (UC), EAE, and collagen-induced arthritis (CIA), which involve organ-specific inflammation in the intestine, CNS, and joint, respectively [1-3]. In the first model, mice carrying a null-mutation in the gene coding for TCRα spontaneously develop chronic intestinal inflammation with features of human UC [4]. UC pathogenesis in this model involved unconventional IL-4-producing CD4+ T cells (Th2-like cells) and was associated with a strong B-cell response [4-6]. However, the disease started earlier and was more severe in TCRα-deficient mice lacking B cells than in TCRα-deficient mice with B cells [4]. This protective effect was found to be mediated partly by antibodies, which facilitated the clearance of apoptotic cells, and possibly limited the proinflammatory effects that can be associated with an abnormal accumulation of dying-cells [7]. In addition, B cells attenuated UC progression by producing IL-10 [1]. The IL-10 produced by B cells did not affect UC onset, but reduced the progression of UC pathogenesis after disease initiation. B cells upregulated IL-10 expression in mesenteric lymph nodes of diseased TCRα-deficient mice, but not in healthy age-matched TCRα-deficient mice, indicating that the disease itself controlled the induction of this protective B-cell response [1]. Of therapeutic interest, adoptive transfer of B cells into mice with mild to moderate UC suppressed disease progression in an IL-10-dependent manner [1, 4].

Similar observations were made in EAE, the primary animal model for relapsing-remitting multiple sclerosis (RR-MS) [2]. EAE can be induced in mice by immunization with myelin antigens such as myelin oligodendrocyte glycoprotein (MOG), which leads to the activation of autoreactive CD4+ T cells and their differentiation into encephalitogenic effectors that secrete T helper 1 (Th1)- and Th17-type cytokines. Disease typically manifests as a moderate paralysis that spontaneously resolves after a few days in wild-type mice [2]. A lack of IL-10 production solely by B cells did not affect time to onset or incidence of EAE, but severely exacerbated the following phase of the disease: mice with IL-10-deficient B cells failed to recover from EAE, unlike control mice, and instead developed a severe chronic paralysis [2]. The suppressive effect of B cells therefore became apparent after EAE was established, and remarkably, this suppression drove an almost complete recovery from disease [2]. In another study, CD4+Foxp3+ regulatory Treg cells were activated normally in the absence of B cells, suggesting that B cells provide an independent mechanism of protection from autoimmunity [8]. Noteworthy, B cells isolated from wild-type mice after EAE recovery efficiently protected recipient mice from disease upon adoptive transfer, even if all B cells in the recipient mice were IL-10-deficient [2]. These findings indicate the potential of adoptive B-cell transfer for treatment of CNS autoimmunity [2]. Interestingly, B cells also play pathogenic roles in EAE, primarily via their secretion of IL-6 [9].

Finally, B cells displayed IL-10-mediated protective functions in the CIA mouse model [3]. B cells isolated from arthritic mice, and restimulated via the B-cell receptor for antigen (BCR) and CD40 ex vivo, protected recipient mice from CIA in an IL-10-dependent manner, even when administered at time of disease onset [3]. The fact that B cells are strictly required for development of CIA further emphasizes the notion that B cells can act as drivers as well as regulators of autoimmune pathology in a single immune-mediated disease.

Collectively, these findings establish that during disease-causing immune responses, B cells can acquire the capacity to produce IL-10, and subsequently limit immunopathogenesis. An immunosuppressive role for IL-10-producing B cells has now been documented in numerous other disease contexts including allergy, infection, transplantation, and cancer, and this highlights the broad relevance of this regulatory mechanism [10-14]. The unique capacity of B-cell-derived IL-10 to limit established diseases involving Th1 cells, Th2 cells, Th17 cells, IL-6-producing B cells or autoantibodies, could make IL-10-producing B cells a valuable vehicle for treatment of immune-mediated disorders.

A comprehension of the specific role of this B-cell-mediated suppressive circuit within the general regulation of immunity would be helpful to rationally envision its therapeutic manipulation. The notion that the suppressive activities of B cells embody a natural mechanism for the self-limitation, rather than the prevention, of immune responses, is supported by the fact that mice lacking IL-10 production by B cells do not spontaneously develop any obvious autoimmune syndrome. Further supporting this notion, B cells from naïve mice do not constitutively secrete IL-10, but require activation by signals that have also been implicated in immune response activation in order to produce this cytokine. For example, the involvement of the BCR in B-cell-mediated suppression was established by showing that mice in which all B cells expressed a unique BCR of irrelevant antigen-specificity developed a chronic EAE upon immunization with MOG, similarly to what was observed in mice with IL-10-deficient B cells [2]. Mice lacking the BCR coreceptor CD19 also developed an exacerbated EAE, which was associated with reduced B cell production of IL-10 by B cells [15]. An important component of the suppressive function of B cells is therefore antigen specificity, in agreement with data from adoptive transfer experiments [10, 16]. In line with this, MOG-reactive B cells isolated from mice after recovery from EAE secreted IL-10 upon restimulation via BCR, in the presence of appropriate cosignals [2].

CD40 is another B-cell activating receptor shown to be important for the suppressive function of B cells. For instance, CD40 was critical for the protective function of B cells during EAE [2]. Interestingly, CD19 and CD40 are both required for optimal humoral immunity. In fact, null mutations in CD40 (or CD40L) are well-known causes of immunodeficiency [17, 18]. TLRs represent another class of receptors important for both B cell-mediated suppression [19] and B-cell activation and differentiation into antibody-secreting cells (ASCs) [20]. Intrinsic TLR signalling has been shown to have a unique capacity to trigger IL-10 expression by mouse naïve B cells in vitro, which could not be induced by signals propagated via BCR or CD40, either alone or in combination [19]. In contrast, B cells first activated via TLR secreted IL-10 upon restimulation via BCR and CD40 [19]. This led to the formulation of a “two-step model” proposing that TLR controls the initiation of the IL-10-mediated suppressive functions of B cells, which can then be further amplified through restimulation of TLR-primed B cells via the BCR and CD40 [21, 22]. It might also be relevant that the basal level of BCR signaling in resting B cells was recently found to modulate their subsequent responsiveness to TLR and CD40 agonists, suggesting a more complex interplay between these different signals [23].

Interestingly, cytokines involved in autoimmune pathogenesis might also contribute to the suppressive functions of B cells. For example, IL-21 increased IL-10 secretion by activated mouse B cells, and IL-21 receptor-deficient B cells were unable to provide protection from disease upon adoptive transfer in recipient mice during EAE [24]. Another function of IL-21 is to induce the expression of PR domain zing finger gene 1 (Prdm1) in B cells via signal transducer and activator of transcription 3 (STAT3) and interferon regulatory factor 4 (IRF4) signaling. PRDM1 (also known as BLIMP1) controls terminal plasma cell differentiation [25-27]. These data illustrate how B cells utilize key stimulatory pathways to provide regulatory functions. Collectively, the signals promoting the suppressive functions of B cells largely overlap with those controlling humoral immunity, suggesting a linkage between the molecular B-cell programs of plasma cell differentiation, and acquisition of IL-10-mediated suppressive functions. A remarkable feature concerning the signals supporting the suppressive functions of B cells is that the absence of any of them is sufficient to severely impair the regulatory effect of B cells on the course of an autoimmune disease such as EAE, while having a milder consequence on humoral immunity. A possible interpretation for these findings is that B cells continuously need stimulatory signals to maintain their suppressive activity, whereas humoral immunity can be maintained independently of continuous restimulation via the persistence of long-lived plasma cells [28].

What might be the role of such suppressive functions provided by activated B cells in the general regulation of immunity? Activated B cells might contribute through the provision of suppressive cytokines a counter-regulatory process important for an optimal control of immune dynamics combined with protection of immunopathology [21]. It is tempting to speculate that increasing this negative feedback circuit beyond its physiological level might permit to interrupt ongoing immune responses.

Identifying the phenotype of IL-10-secreting B cells with immunosuppressive activity might lead to the identification of a “regulatory B-cell” subpopulation. The quest for the phenotype of IL-10-producing B cells in mice has been addressed using two strategies. A first approach has been to determine the phenotype of the B cells that could, after isolation from mice and short term in vitro stimulation, express IL-10. This strategy allowed for identification of the cells that have the competence to produce IL-10. Using such a protocol Yanaba et al. demonstrated that some of the splenic B cells competent for IL-10 production have a CD1dhighCD5+ phenotype [16]. These cells also express IgM and CD24, and lack CD23 and CD93 [16, 29]. Remarkably, CD1dhighCD5+ B cells from naïve mice could suppress immunity in recipient mice in an IL-10-dependent manner upon adoptive transfer in multiple disease models [30]. These data demonstrated that CD1dhighCD5+ B cells have the capacity to develop into IL-10-producing B cells with suppressive functions upon adoptive transfer in vivo. It is unknown whether the cells that actually produced IL-10 in the recipient mice in these adoptive transfers retained their initial CD1dhiCD5+ phenotype, or differentiated toward a distinct phenotype. Noteworthy, CD1dhigh B cells have been shown to produce the highest amounts of IL-6 among B-cell subsets in vitro and in vivo during EAE [9]. Given that IL-6 production is the major mechanism of B-cell-mediated pathogenesis in EAE, CD1dhigh B cells might also provide some pathogenic functions [9]. Thus, it would be more appropriate to consider CD1dhi B cells as precursors with multifunctional potential rather than as “regulatory B cells.” A recent study provided further support to this concept by demonstrating that the CD1dhi B cells expressing IL-10 after in vitro stimulation via TLR-4 expressed comparable IL-6 to the CD1hi B cells from the same culture that did not express IL-10 [31]. These considerations emphasize the importance of assessing more factors than IL-10 when analyzing the expression of cytokines by B cells. In fact, it is also important that ectopic expression of IL-10 in naïve B cells is insufficient to achieve suppressive functions [32], indicating that additional features (possibly the production of other cytokines) are therefore required, in addition to IL-10 expression, for B cells to provide regulatory effects.

It will now be important to identify the molecular switches controlling the production of pro- versus anti-inflammatory cytokines by B cells, as well as the factor that can stabilize such cytokine expression profiles. The differential engagement of stimulatory receptors might play a role in this differential cytokine expression, because CD1dhigh B cells secreted large amounts of IL-10 and little IL-6 upon activation with LPS, whereas they produced less IL-10 and more IL-6 following concomitant activation via TLR4 and CD40 [9]. Importantly, some non-CD1dhighCD5+ B cells have also been shown to produce IL-10 in vitro [10], and B1 cells as well as transitional T2-like B cells inhibited immunity in recipient mice upon their adoptive transfer [33-35]. Thus, the competence to secrete IL-10 upon activation and provide regulatory functions upon adoptive transfer is not restricted to a particular B-cell subset.

It would be of great interest to identify functional and phenotypic properties that distinguish IL-10-competent B-cell subsets from noncompetent ones. A shared functional feature might be that CD1dhigh B cells, B1 cells, and T2-like B cells can all respond with fast kinetics to TLR stimulation, and subsequently develop into plasma cells [36-38]. Accordingly, the activation status of a B cell might be an important parameter for its competence to produce IL-10. The recent finding that T cell Ig domain and mucin domain 1 (TIM-1) expression is associated with the competence of B cells to produce IL-10, independently of their CD1d expression level, argues in favor of this concept [10]. Indeed, TIM-1 expression reflects a particular B-cell activation status, as it can be induced on Tim-1 cells upon B-cell stimulation in vitro, and Tim-1 B cells can upregulate TIM-1 expression in vivo in recipient mice upon adoptive transfer [10]. These data suggest that TIM-1 B cells can acquire the capacity to express TIM-1 and IL-10 upon appropriate stimulation, and consequently argue against the existence of a unique subset of regulatory B cells.

Importantly, TIM-1 expression might directly contribute to the suppressive functions of B cells. TIM1 engagement has been shown to increase IL-10 production by B cells [10], and B cells from mice with a mutation in the extracellular domain of TIM-1 displayed a reduced secretion of IL-10 compared to controls [39]. TIM-1 is a pattern recognition receptor specialized in the recognition of phosphatidyl serine, a phospholipid that gets exposed on the surface of apoptotic cells [40, 41]. B cells produced IL-10, and acquired suppressive functions upon exposure to apoptotic cells [42]. Thus, B cells activated in a suitable way might then be prepared for secreting IL-10 upon exposure to apoptotic cells via a mechanism involving TIM-1.

A second approach to identify “regulatory B cells” has been to determine the phenotype of B cells that actually provide IL-10 in a regulatory manner in vivo, in absence of any ex vivo restimulation. Such analysis was first made in mice in the context of infection by the Gram-negative bacterial pathogen Salmonella enterica Typhimurium, a disease in which IL-10 from B cells also had a suppressive role [11]. Indeed, mice lacking IL-10 expression in B cells displayed an increased resistance to virulent Salmonella infection and prolonged survival compared with mice with wild-type B cells [11]. Using the B-Green reporter mouse strain — a knock-in mutant with an enhanced GFP (eGFP) reporter coding sequence inserted after the stop codon of the Il10 gene — it was found that B cells did not express IL-10 in unchallenged mice. Yet, IL-10-expressing B cells were found in spleen of challenged mice already on day 1 after infection. Remarkably, all eGFP-positive cells carried high levels of the plasma-cell marker CD138 on their surface, and this indicated that IL-10-producing cells were plasma cells [11]. The fact that CD19+CD138high cells were the major source of B-cell-derived IL-10 in vivo was confirmed by quantifying Il-10 mRNA expression in CD19+CD138+ and CD19+CD138 B cells that had been isolated from wild-type mice on days 1 and 3 after infection [11]. The percentage of CD19+CD138+ cells expressing IL-10.eGFP increased proportionally to the amount of bacteria administered, with up to 50% of CD19+CD138+ expressing the reporter gene, suggesting that direct sensing of microbial products, for instance via TLR, regulates the magnitude of IL-10 expression [11]. Remarkably, no GFP expression was detected in CD19 or CD19+CD138 cells immediately after S. enterica Typhimurium infection, indicating that CD19+CD138+ plasma cells were a unique source of IL-10 the first day after challenge [11]. This unique role of plasma cells as early providers of IL-10 was further characterized in a subsequent study [31]. It was found that after their isolation from spleens of S. enterica Typhimurium-infected mice, and their in vitro restimulation, only CD138hi plasma cells, but not CD19+CD138 B cells, secreted IL-10 [31]. The link between IL-10 expression and plasma cell differentiation was further examined by separating plasma cells into different fractions according to their maturation stages. Remarkably, the most mature plasma cell subset, that is plasma cells that contained the highest mRNA levels for the transcription factors BLIMP1 and IRF4 and were the most efficient antibody-producers, also expressed the highest level of IL-10 compared with less differentiated subsets [31]. Moreover, single cell analyses demonstrated that all IL-10-expressing cells were also positive for the Prdm1 mRNA, as expected for plasma cells. Taken together, these data clearly demonstrated that plasma cells are the major source of B-cell-derived IL-10 during S. enterica Typhimurium infection. Extending these findings, it was also found that CD138hi plasma cells are the major IL-10-producing B-cell subset in mice during EAE [31]. Remarkably, plasma cells did not express any IL-6 upon restimulation, in contrast to CD1dhi B cells, yet secreted another suppressive factor, namely IL-35 [31].

The finding that plasma cells were a major source of IL-10 suggests that ASCs could regulate immunity via antibody-independent mechanisms. How does IL-10 produced by ASCs impact on immunity? During S. enterica Typhimurium infection, IL-10 secretion by plasma cells resulted in suppression of the innate immune response mediated by neutrophils and NK cells [11]. Interestingly, plasma cells and neutrophils can colocalize in the spleen [43]. Thus, plasma cells might provide immunosuppressive signals to innate immune cells through the production of anti-inflammatory cytokines such as IL-10 and IL-35 (Fig. 1). Such a connection between plasma cells and innate immune cells might be important for the transition from a nonspecific system of host defense, which poses a high risk for immunopathology, to a system involving the adaptive immune system, which is more specific. In that respect, it is of interest that IL-10 had been previously identified as a potent stimulatory factor for the growth of activated human B cells and their differentiation into high-rate antibody-secreting plasma cells [44]. As such, IL-10 production by plasma cells might have positive or negative consequences depending on the respective roles of innate versus humoral immunity in the disease considered. In line with this, CD19+CD138+ cells have been reported to secrete IL-10 in a model of systemic lupus erythematosus, yet a lack of IL-10 production by B cells did not affect the disease course [45]. In this case, it might be pertinent that IL-10 can act in a pathogenic manner in systemic lupus erythematosus, possibly by catalysing humoral immunity [46-48]. Of note, cells with a plasma cell phenotype were also the main type of B cells expressing IL-10 in mice after LPS administration, extending the link between IL-10 and plasma cells further to different types of immune responses [49].


Figure 1. Upon infection of mice with Salmonella enterica Typhimurium, engagement of TLR leads to rapid activation of innate immune cells, which provide a first line of defense against the pathogen. In parallel, the adaptive immune response progressively develops, leading to the differentiation of B cells into antibody-secreting cells (ASCs). A fraction of these plasma cells secrete IL-10 or IL-35, which subsequently results in an attenuation of innate immune cells, particularly of NK cells and neutrophils, as well as of macrophages. A function of ASC-derived IL-10 and IL-35 might be to avoid the superposition of full-blown innate and humoral responses, which could lead to immunopathology. As such, the provision of IL-10 and IL-35 by ASCs might provide an ingenious way to orchestrate the transition from a nonspecific innate immune response to a more specific mechanism of immunity involving antibodies. In this perspective, it might be relevant that IL-10 had been previously identified as a potent stimulatory factor for the growth and differentiation of activated human B cells into ASCs.

Download figure to PowerPoint

The observation that some ASCs are a major source of B-cell-derived IL-10 in vivo in various models is somehow a surprise because plasma cells have so far been considered exclusively for their roles in antibody production. A better characterization of plasma cell subsets is needed to confirm the existence of “regulatory plasma cells”. For a regulatory subset to be identified, it would require that the vast majority of the cells in this subset produce anti-inflammatory cytokines such as IL-10, and no proinflammatory cytokines such as IL-6. This cell subset should also show stability (or at least resilience) of such a cytokine expression profile upon exposure to various environments. The finding that plasma cells can also secrete the suppressive cytokine IL-35, and are unable to produce IL-6, is consistent with the notion that they can mediate regulatory functions. However, after the discovery that ASCs are the main source of IL-10 in S. enterica Typhimurium-infected mice, it has been reported that distinct subsets of ASCs can produce other cytokines. B1 cells have been shown to differentiate into GM-CSF-producing CD19+CD138+ plasma cells in a mouse model of sepsis induced by caecal ligation puncture [50]. In addition, plasma cells have been found to express both iNOS and TNF in the lamina propria of the small intestine in mice [51]. These data collectively ascertain that ASCs do more than produce antibody, in particular secrete cytokines, and subsequently can mediate a wealth of effects. This might confer novel functions to ASCs, which could operate independently of their antigen specificity. We anticipate that the exploration of the cytokine- and chemokine-mediated functions of ASCs will be a fruitful area of future investigation, particularly given the fact that only a fraction of these cells produce autoreactive or neutralizing antibodies in autoimmune or infectious diseases, respectively. Could cytokine/chemokine secretion be a raison d’être for such ASCs, and lead to the definition of novel ASC subsets?

There is now clear evidence that B cells can inhibit immunity through the production of IL-10 in mice. Human B cells might also have regulatory functions that similarly depend on secretion of IL-10. Ex vivo-activated B cells from patients with RR-MS or type 1 diabetes have been reported to produce less IL-10 than B cells from healthy individuals, and this suggest that defective IL-10 production by B cells might facilitate autoimmune pathogenesis in humans [52, 53]. Further evidence that IL-10-secreting B cells might protect from deleterious inflammation in humans comes from case studies showing that rituximab treatment was followed by a severe worsening of disease in patients with UC, which temporally correlated with depletion of both B cells and IL-10 in the intestinal mucosa [54]. Moreover, increased production of IL-10 by B cells in patients with autoimmune diseases has been associated with less severe disease course. For instance, patients with RR-MS that had also been infected with helminth parasites displayed enhanced production of IL-10 by B cells compared with noninfected RR-MS patients, and this increased IL-10 production by B cells correlated with a lower frequency of RR-MS flares [55, 56]. Noteworthy, the data obtained in mice showing that ASCs can regulate immunity via antibody-independent mechanisms, most notably the secretion of IL-10 and/or IL-35 might explain why Atacicept, a drug that reduces the numbers of short- and long-lived ASCs in mice [57, 58], led to an aggravation of disease in patients with RR-MS [59]. Indeed, a possible explanation for this effect could be that some ASCs provided protection from disease via the secretion of IL-10 (or IL-35) in these patients. Further studies are required to assess cytokine production by plasma cells in humans.


  1. Top of page
  2. Abstract
  3. Acknowledgments
  4. Conflict of interest
  5. References

S.F. is supported by grants from the Deutsche Forschungsgemeinschaft (SFB-650, TRR-36, TRR-130, FI-1238/02), Hertie Stiftung, and an advanced grant from the Merieux Institute.

Conflict of interest

  1. Top of page
  2. Abstract
  3. Acknowledgments
  4. Conflict of interest
  5. References

The authors declare no financial or commercial conflict of interest.


  1. Top of page
  2. Abstract
  3. Acknowledgments
  4. Conflict of interest
  5. References

antibody-secreting cell


B-cell receptor


collagen-induced arthritis


enhanced GFP


myelin oligodendrocyte glycoprotein


relapsing-remitting multiple sclerosis


ulcerative colitis