What are regulatory B cells?


  • David Gray,

    Corresponding author
    1. Institute of Immunology and Infection Research, School of Biological Sciences, University of Edinburgh, Ashworth Labs, Edinburgh, UK
    • Institute of Immunology and Infection Research, School of Biological Sciences, University of Edinburgh, Ashworth Labs, King's Buildings, West Mains Road, Edinburgh EH9 3JT, UK Fax: +44-131-650-7322
    Search for more papers by this author
  • Mohini Gray

    Corresponding author
    1. Centre for Inflammation Research, University of Edinburgh, Queen's Medical Research Institute, Edinburgh, UK
    • Centre for Inflammation Research, University of Edinburgh, Queen's Medical Research Institute, Edinburgh, UK Fax:+44-131-650-7322
    Search for more papers by this author


B cells are now acknowledged to play multiple roles in the immune response, in addition to making antibodies. Their role in regulating T-cell responses during inflammation has come into focus recently. Thus, IL-10 production by B cells has been shown to be important in limiting auto-reactive and pathogen-driven immune pathology; however, the exact identity of the regulatory B cells remains elusive; do they belong to a committed subset or can all B cells regulate given the appropriate inducing stimuli? A study in this issue of the European Journal of Immunology provides insight into the IL-10-producing B cells in humans, suggesting that many B cells have the capacity to make IL-10 when optimally stimulated via the BCR and TLR9. Despite producing significant amounts of inflammatory cytokines as well, these B cells suppress T-cell proliferation. This Commentary places this study in the context of what we think we know about regulatory B cells and more importantly highlights the questions we still need to answer.

B cells are perhaps best known for their production of antibodies, providing protection against infection; however, they are also central to the pathogenesis of many autoimmune diseases. Often, this is through the secretion of autoantibodies, but it can also result from their Ag-presenting activity 1. This central role in inciting inflammation in autoimmune diseases has made B cells attractive therapeutic targets through the depletion of these cells. Indeed, this has proven efficacious in rheumatoid arthritis where Rituximab is now used routinely in seropositive patients when anti-TNF therapies have failed. In the past decade, however, data from a number of laboratories have revealed that B cells can also dampen inflammation through their interaction with effector T cells and other innate cells. In all the models studied so far, this suppressive or regulatory effect of B cells is mediated by the production of the cytokine IL-10, which inhibits both Th1 and Th2 polarization, Ag presentation and pro-inflammatory cytokine production by myeloid cells.

A suppressive role for B cells had been suspected for many years and in 1996 Janeway's group 2 showed that in the mouse model of multiple sclerosis, EAE, a genetic lack of B cells led to exacerbated disease. Subsequent to this, Fillatreau et al. 3 showed that a B-cell-restricted IL-10 deficiency had a similar exacerbating effect on EAE, suggesting that activated B cells exerted regulatory activity that resolved the inflammation. At around the same time, similar observations were made in inflammatory bowel disease (4 and rheumatoid arthritis 5). From this, arose the idea of the “regulatory B cell”; recent reviews of this field 6–8 give more information on their patho-biology.

Generally, in immunology, new subsets of immune cells do not enter mainstream consciousness unless they have a unique set of cell surface markers that define functionality or a defined transcription factor that controls their development and/or function. Currently, regulatory B cells have neither attributes. Work has been done on the markers expressed by B cells that produce IL-10; however, a consensus is lacking. In mice, Tedder and colleagues 9 have described a “B10” subset contained within the splenic CD5+CD1d+ B-cell population that uniquely secretes IL-10 in response to LPS activation, whereas Mauri and coworkers 10 have consistently found that splenic B cells with the transitional T2 phenotype (CD21hiCD23+) secrete IL-10 in response to CD40 stimulation. Regulatory B-cell function has been observed in humans 11, but even less is known about their identity. Mauri's group recently provided further support to their mouse data by showing that human IL-10-producing B cells were found within the transitional CD38+CD24+ B-cell population 12, whereas others have assigned them to the CD27+ memory cell pool 12. A study in this issue of the European Journal of Immunology by Bouaziz et al.13 addresses this and several other important questions surrounding regulatory B cells. Bouaziz et al.13 show that when analyzing B cells directly from the blood, the frequency of IL-10 expression (by intracellular staining) was higher in the CD27+ (memory) or the CD38+CD24+ (transitional) subsets. However, crucially, when asking the reverse question (which markers are found on IL-10-expressing B cells?), they show that the IL-10 producers had diverse phenotypes. What does this mean? It could be that the memory and transitional subsets are simply at a higher level of activation and a significant proportion of activated B cells make IL-10. In other words, sampling the peripheral blood followed by short-term stimulation gives a read out of recent activation history. Alternatively, it could be that these subsets express the receptors that efficiently stimulate IL-10 secretion, whereas other subsets have less and need to acquire them via activation. In other words, all B cells can make IL-10, given the correct activatory context and time. This notion is supported by the fact that the optimal time for IL-10 production after optimal stimulation is 48 h, a longer time than might be expected if a primed or dedicated IL-10-secreting population was the major contributor.

Although a definitive “Breg” subset may not exist, what is emerging is the importance of the stimulus. Among the strongest stimuli for IL-10 production are TLR ligands 14. In EAE, TLR2 and TLR4 have been shown to be critical for the development of B-cell regulatory function 15 and there is known to be a protective effect in EAE 16 and diabetes in NOD mice 17 of transferring LPS-activated B cells. Human B cells do not express TLR4 but they do express TLR9 and, as Bouaziz et al.13 show (as others before) TLR9 is a potent inducer of IL-10. The optimal stimulus, in fact, is a combination of signals via TLR9 and BCR, thus, CpG and anti-Ig act synergistically. Interestingly, CD40L stimulates minimal IL-10 production and indeed, when added to the CpG+anti-Ig stimulus, it inhibits IL-10 production. This is in apparent contrast to data from the Mauri lab 12 and also with some of the mouse studies 3, but might be reconciled by the proposal that different B-cell subsets respond optimally to different stimuli at different stages of their activation and maturation. Thus, some B cells (CD27+ or CD38+) might respond early in an innate fashion to temper the inflammatory response (and/or induce subsequent regulatory populations), whereas the Ag-specific regulatory B cells appear later as part of the adaptive (CD40L dependent) immune response.

If B cells make IL-10 and become regulatory following BCR and TLR9 activation, then several questions arise. During many infections, both BCR and TLR9 receptors in B cells are likely to be engaged, does this mean that B cells activated in this way become regulatory by default? The fact that B-cell regulation develops as part of the normal activation process suggests that regulatory activity is generated concurrently with the inflammatory response. This fits with much of the data from mouse autoimmune models, especially the observation that anti-CD20-mediated B-cell depletion prior to EAE induction increases disease severity, whereas depletion during the disease alleviates disease symptoms 18. Does this mean that populations of regulatory B cells exist that respond to endogenous TLR ligands 19, and if so what are these ligands and do they become the source of auto-Ag in the breakdown of tolerance that precedes autoimmunity? More intriguing would be to speculate that a single ligand serves both to stimulate the BCR and to activate the TLR receptor. In the case of nuclear components (e.g. DNA and RNA), the BCR might be the only way for them to be delivered to the intracellular compartments where the appropriate TLR are expressed 20.

Importantly, Bouaziz et al. go on to demonstrate that the IL-10-producing B cells generated by CpG and anti-Ig have significant regulatory function as shown by their ability to suppress CD4 T-cell proliferation. This is only partially inhibited by anti-IL-10, possibly because of technical issues or may be because strong cognate interaction and synaptic release of IL-10 is not easily interrupted. Interestingly, the suppressive action of human B cells has been shown to be dependent on CD80 and CD86 12. What is not clear from this study or a similar recent study is whether the IL-10-secreting B cells are able to induce T cells to secrete IL-10 to become Tr1-type regulatory T cells.

Finally and importantly, this article does not ignore the fact that the stimulus they use is likely to elicit production of other cytokines and hence they find that their regulatory B cells also make IL-4, IL-6, IL-12 and IFN-γ (but not TGF-β). Given their demonstrable suppressive activity, the effect of IL-10 appears to be dominant, but, as mentioned above, the experiments are not extended to show the outcome of T-cell differentiation in these cultures. The production of many of these B-cell-derived cytokines may well have an impact on T-cell polarization as indicated by recent studies 21, 22. We do not yet understand the temporal and contextual constraints controlling the expression of these cytokines by B cells in vivo during inflammation or infection. In this, B cells are no different from other APC.


Work in the authors' laboratories is funded by the Wellcome Trust, the Medical Research Council (UK) and Arthritis Research, UK (M. G.).

Conflict of interest: The authors declare no financial or commercial conflict of interest.