|Home | About | Journals | Submit | Contact Us | Français|
Unlike other immune cells, B cells express both an antigen-specific B cell receptor (BCR) and Toll-like receptors (TLRs). Dual BCR and TLR engagement can fine-tune functional B cell responses, directly linking cell-intrinsic innate and adaptive immune programs. While most data regarding B cell-specific functions of the TLR signaling pathway has been obtained in mice, the discovery of patients with a deficiency in this pathway has recently provided insight into human B cell responses. Here, we highlight the importance of the integration of signals downstream of BCR and TLR activation in modulating B cell function, focusing when possible on cell intrinsic roles.
Over the last decade, the unexpected success of B cell depletion therapies in human autoimmunity, combined with a growing recognition of the importance of neutralizing antibody responses in host defense, has led to an increased focus on understanding the role(s) for B cells in human immune function. B cells do not merely produce immunoglobulin, but can also secrete cytokines and serve as antigen-presenting cells and therefore B cells have a multi-faceted involvement in distinct immune responses.
A striking characteristic of B cells is expression of a clonally-rearranged, antigen-specific B cell receptor (BCR) in conjunction with expression of one or more members of a family of germline-encoded receptors termed Toll-like receptors (TLRs), capable of recognizing discrete microbiological ligands. This dual expression program permits B cells to uniquely integrate both antigen-specific and ‘danger’ signals via these key receptor systems. Although both B cell development and survival appear phenotypically unperturbed in the absence of TLR signals1, patients with IRAK-4 or MyD88 deficiency possess an altered BCR repertoire with an increased proportion of autoreactive cells, presumably due to alterations in B cell selection processes2. Different B cell subsets exhibit variations in TLR expression patterns, and signaling via TLRs can modify B cell responses such as antibody production, antigen presentation and cytokine secretion. Therefore, individual TLR expression profiles permit various effector B cell populations to manifest specific response profiles following TLR engagement3,4.
Notably, based upon their functional responses as well as their BCR repertoire, naïve mature B cell populations have been defined as either innate-like or adaptive cells (Box 1). Innate, B-1 or marginal zone, B cells generate rapid antibody responses independent of T cell help. In contrast adaptive, follicular B cells primarily participate in T-dependent responses leading to generation of high-affinity antibodies and long-term memory. Importantly, expression of a distinct profile of TLRs and a specific BCR profile likely helps to specify the differentiation and function of these key innate vs. adaptive B cell populations. During T-independent immune responses, dual BCR and TLR signaling rapidly induce marginal zone cells and B-1 B cell migration and antibody production. Additionally, upon triggering of T-dependent immune responses, TLR responsiveness is directly modulated in activated follicular B cells thereby impacting germinal centre responses. TLR engagement, in conjunction with BCR ligation, also provides a bridge between the innate and the adaptive immune system that may impact on antigen presentation, primary antibody responses, class-switch recombination and subsequent memory responses.
Based on phenotypic, functional and topographical characteristics, B cells can be divided into innate-like and adaptive immune cells109. Follicular B cells are the main players during T-dependent immune responses and belong to the adaptive arm of the immune system. They generate a clonally rearranged antigen-specific B cell receptor (BCR) and form memory responses that are dependent on T cell help. In contrast, B-1 and marginal zone B cells are usually considered innate-like immune cells and generate rapid but lower affinity antibody responses that are independent of T cell help.
The term ‘B-1’ refers to the idea that this populations develops earlier during ontogeny than conventional B-2 cells110. B-1 cells are enriched in the peritoneal and pleural cavity but can also be found in the spleen. CD5 expression further subdivides mouse B-1 cells into CD5+ B-1a and CD5− B-1b cells. Recently, a B-1 cell progenitor was identified in the bone marrow of adult mice111. The term ‘B-2’ has traditionally been used to describe the main population of mature B cells that develop from common bone marrow precursors and are located in the bone marrow, spleen and lymph nodes; B-2 cells therefore include both follicular and marginal zone subsets, which presently are referred to as separate populations because of their distinct phenotypic and functional characteristics.
Recent work has defined a B cell subset in human peripheral blood with functional responses similar to mouse B-1 cells112. Consistent with their innate-like immune cell phenotype, marginal zone and B-1 B cells mainly express germ-line encoded antigen receptors that have limited diversity and are enriched for specificities that recognize microbial and self-antigens. In addition to BCR ligation, activation of pattern-recognition receptors including Toll-like receptors (TLRs) on these cells is important for their immune responses. Moreover, B-1 and marginal zone B cells are the primary producers of natural IgM antibody113. Through these characteristics, both subsets are crucial in the early phase of T-independent immune responses10 linking innate and adaptive immune mechanisms. Moreover, due to their polyspecific BCR repertoire, B-1 and marginal zone B cells have been implicated in driving autoimmune processes109.
All TLRs, with the exception of TLR3, require the signaling adaptor myeloid differentiation primary-response protein 88 (MyD88) to mediate activation signals, although TLR4 can uniquely signal through both MyD88-dependent and –independent pathways5. Therefore, analyses of animal models and humans with deficient function of this adaptor have begun to provide important new insight into how signals via TLRs impact B cell function and immune responses. Further, recent work indicates that MyD88 also orchestrates TACI (transmembrane activator and calcium-modulator and cyclophilin ligand interactor)-mediated, BAFF-driven signals.
BAFF and the related cytokine APRIL play a major role in peripheral B cell homeostasis and survival6. These cytokines bind to a family of receptors expressed primarily on B cells including the BAFF receptor (BAFFR), B cell maturation antigen (BCMA), and TACI. BAFF levels are modulated by a complex interplay of factors that includes the number of B cells competing for ligand, the relative level of BAFF-binding receptors on specific B cell subsets, and inflammatory (including TLR-driven) signals that modulate BAFF production by myeloid and/or stromal cell populations. BAFF- or BAFFR-deficient mice have severely decreased numbers of B cells, whereas overexpression of BAFF increases B cell numbers and promotes systemic lupus erythematosus (SLE)-like disease. Both BCR and TLR-mediated stimulation of mouse marginal zone B cells7,8 or B-1 cells9 up-regulates BAFFR and TACI expression, increasing the sensitivity of these cells to BAFF and APRIL. As a result, B cell activation during immune responses directly modulates downstream outcomes of BAFF signaling by regulating BAFF-family receptor expression. Finally, an additional level of control is provided via direct participation of MyD88 in BAFF-triggered TACI signals. This dual role for MyD88 in TLR and TACI signaling thereby permits precise regulation of peripheral B cell responses to infectious challenge and other microenvironmental cues.
This review focuses on the role dual BCR and TLR (as well as other MyD88-dependent signals) in normal and dysregulated B cell immune function, with a particular emphasis on B cell-intrinsic events. We first summarize the specific roles for MyD88 signals in T-independent versus T-dependent B cell immune responses. We will highlight work that demonstrates a non-redundant role for B cell-intrinsic MyD88 signals in antiviral immune responses; describe how MyD88 controls a B cell-intrinsic, TLR-independent, pathway for immunoglobulin diversification; and review new information regarding the role for TLR signals in human immune function gained via analysis of patients with MyD88 or IL-1R-associated kinase 4 (IRAK4) deficiencies. Next, we describe how dual BCR and TLR signals may potentiate the risk for autoimmunity and discuss recent findings regarding how regulatory B cell function also requires TLR/MyD88 signals. Finally, we will discuss emerging data indicating that disregulated TLR and BCR signals collaborate during the development of B cell malignancies and, possibly, in other pathogenic B cell states.
As innate-like cells, marginal zone and B-1 B cells are the main players during T-independent immune responses. Innate B cells rapidly develop into IgM-producing plasmablasts during the early primary immune response10 and the importance of these subsets in T-independent responses to encapsulated bacteria has been demonstrated in infectious models most notably including Streptococcus pneumoniae11. Both mice and humans with deficiencies in B-1 and marginal zone B cells, including patients with hypo- or agammaglobulinemia12 and Wiskott-Aldrich Syndrome (WAS)13, also exhibit increased susceptibility to infections with encapsulated bacteria; and both young children, in whom formation of the marginal zone B cell compartment is physiologically delayed until about two years of life, and splenectomized individuals are at very high risk for such infections14. Figure 1 provides an overview how BCR and MyD88 signal function in concert to modulate T-independent responses based on the in vitro and in vivo studies described below.
Stimulation of B cells via TLR ligands has been used extensively as an in vitro model of T-independent immune responses. In vitro, both B-1 and marginal zone B cells rapidly proliferate and secrete antibody in response to TLR engagement, including TLR1–TLR2, TLR4, TLR6, TLR7 and TLR93,15–18. Interestingly, B-1 cells (predominantly B-1b cells) stimulated in this manner secrete large amounts of IgA, whereas marginal zone B cells primarily produce IgM4,15; a feature that correlates with their in vivo functional roles in mucosal vs. blood-borne infections. Notably, both marginal zone and B-1 B cells exhibit stronger functional responses to TLR ligands than follicular B cells, as measured by up-regulation of activation markers15,18 and production of IL1019,20. Marginal zone B cells have also been shown to exhibit greater potential to act as antigen-presenting cells than follicular B cells in response to TLR stimulation18, allowing this unique subset to efficiently bridge T-independent and T-dependent responses via shuttling antigen from the marginal zone and presenting it within the T cell zones.
Our group investigated the mechanism(s) accounting for the differential responsiveness of adaptive (follicular) vs. innate (marginal zone) B cells to the classically defined, T-independent type 1 stimulus and TLR4 ligand, lipopolysaccharide (LPS)3. We confirmed a delay in cell cycle entry in follicular compared with marginal zone B cells, but could not identify significant differences in proximal TLR-driven biochemical signaling events including activation of nuclear factor-κB (NF-κB) and mammalian target of rapamycin (mTOR) pathways. Notably, follicular B cells exhibited reduced basal and inducible levels of the cell cycle and growth regulator, c-MYC. Consistent with a key role for c-MYC in modulating LPS-responsiveness, enforced expression of c-MYC in follicular B cells eliminated the delay in cell cycle entry and promoted increased immune responses that mimicked functional responses of MZ B cells
Thus, consistent with their characterization as innate immune cells, B-1 and marginal zone B cells generally exhibit stronger in vitro responses to TLR signaling compared with follicular mature B cells and regulation of c-MYC expression levels contribute, in part, to this differential response profile.
Signaling via both MyD88 and Bruton’s tyrosine kinase (BTK) is required for T-independent pathogen-specific IgM production in a mouse model of Borrelia hermsii infection. Mice deficient in both BTK (a kinase essential for BCR signalling) and MyD88 failed to generate pathogen-specific IgM21, whereas mice deficient in BTK alone generated specific antibodies and resolved bacteremia. In contrast, mice deficient in TLR1, TLR2 or MyD88 generated pathogen-specific antibodies with delayed kinetics and suffered more severe episodes of bacteremia, suggesting that MyD88 specifically synergizes with BCR signals and that in Btk−/− B cells, TLR-mediated stimulation is sufficient to rescue the defective BCR signal.
Two groups have analyzed in vivo B cell responses to Salmonella typhimurium in the setting of MyD88 deficiency restricted to the B cell compartment22,23. Salmonella typhimurium induces both T-independent and T-dependent immune responses. Following S. typhimurium infection, S. typhimurium-specific IgM and IgG responses were initially lower in mice with Myd88−/− B cells, but no difference was observed at later time points (>4 weeks) after infection. Moreover, MyD88 signaling in B cells significantly inhibited both neutrophil and natural killer (NK) cell responses to S. typhimurium23. Based on these limited in vivo data, MyD88 signaling in B cells, in conjunction with BCR engagement, appears to play an important role during early T-independent immune responses to bacteria leading to both rapid production of protective IgM and IgG, and modulation of other innate effector cell populations (presumably via cytokine and/or chemokine production).
TLR signals play an important role in the regulation of localization and migration of B-1 and marginal zone B cells. While enriched in the splenic marginal zone, marginal zone B cells continuously circulate between the marginal zone and splenic follicles, a property that allows this population to capture and shuttle blood borne antigens to follicular dendritic cells. In vivo treatment with LPS promotes near complete relocalization of marginal zone B cells into the splenic white pulp24–26. Sphingosine-1-phosphate receptor 1 (S1P1) represents a important regulator marginal zone B cell retention and blocking S1P1 with the inhibitor FTY720 in vivo leads to migration of marginal zone B cells into follicles27. Interestingly, LPS administration results in downregulation of S1P1 expression by marginal zone B cells and reduced chemotactic responsiveness to S1P possibly explaining this observation. Of note, TLR2, TLR3 and TLR7 ligands also promote marginal zone B cell migration28. Interestingly, whereas in vivo treatment with TLR2, TLR4 and TLR7 ligands resulted in downregulation of S1p1 mRNA levels, no change was seen after stimulation via TLR3, implying that at least two distinct TLR dependent signalling pathways can promote migration of marginal zone B cells.
Similarly, B-1 cells express very high levels of integrins and stimulation via TLRs induces a massive egress of B-1 cells from the peritoneal cavity associated with coordinated down-regulation of integrins and CD9 expression29. Thus, B cell intrinsic TLR signaling in vivo can directly alter B-1 B cell responsiveness, thereby directing their migration to sites where rapid local antibody responses help to limit pathogen growth. This likely also allows innate B cells to modulate the local effector functions other immune cell types.
While it is difficult to judge how findings generated using mouse models can be translated into the human system, some insight into the role of B cell-intrinsic TLR signals can be obtained from the analysis of patients with inborn errors in the downstream TLR signaling effectors, MyD88 or IRAK4 (Box 2)30–32. In these patients, S. pneumoniae is the most common invasive pathogen, followed by S. aureus and P. aeruginosa — all of which trigger T-independent responses31. In contrast, antibody responses to protein antigens are normal in these patients and no severe viral, parasitic or fungal diseases have been observed. The predominance of infections with encapsulated bacteria suggests that TLR signaling in human B cells is likely crucial for triggering T-independent immune responses that require B-1 and/or marginal zone B cells. A detailed analysis of the in vivo responses to polysaccharide immunization in such individuals would be a helpful means to begin to directly test this interpretation.
Over the last decade several monogenic primary immunodeficiencies affecting the Toll-like receptor (TLR) signaling pathway have been identified. Inborn errors of the TLR and interleukin-1 receptor (IL-1R) pathway include IL-1R-associated kinase 4 (IRAK4) and myeloid differentiation primary-response protein 88 (MyD88) deficiency identified in patients in 200332 and 200830, respectively. The serine/threonine kinase IRAK4 and MyD88 are key adaptor molecules that function downstream of TLRs and IL-1R. Both inborn errors are inherited in an autosomal recessive manner and result in indistinguishable clinical disease31. Patients with these inborn errors are typically predisposed during infancy and early childhood to recurrent life-threatening bacterial infections, most commonly involving invasive pneumococcal disease with only limited signs of systemic inflammation. Interestingly, no deaths or invasive infectious episodes have been reported in patients beyond age 14 years. Moreover, no severe viral, parasitic and fungal disease have been reported and the range of bacterial pathogens that theses patients are susceptible to seems to be restricted to Streptococcus pneumococcus, Pseudomonas aeruginosa and Staphylococcus aureus. Immune responses to protein antigen were also mainly normal. Treatment of these patients with prophylactic antibiotics, anti-pneumococcal vaccination and/or intravenous immunoglobulin seemed to have a beneficial impact on the disease. The absence of severe infections in older patients may result from maturation of other signaling cascades and/or adaptive immune function that gradually compensates for the lack of TLR signaling.
Another group of inborn errors affects the TLR3–interferon β (IFNβ) and –IFNλ pathway, independent of MyD88 signaling and results from autosomal recessive UNC93B deficiency (an ER-resident transmembrane protein, deficiency in which impacts on TLR3, TLR7, TLR8 and TLR9 signaling)114 and autosomal dominant TLR3 deficiency115. Both deficiencies are associated with the occurrence of Herpes simplex virus-mediated encephalitis. Overall, these data suggest that deficits in TLR signaling increase susceptibility for specific pathogens without obviously impacting responses to T-dependent immunization or immune cell development. Further studies, however, are required to fully understand the precise contributions of TLRs and IL-1R to human host defense.
Recent work has identified an unexpected role for MyD88 in signal transduction via the BAFF- and APRIL-binding cell surface receptor TACI (also known as TNFRSF13B) (Figure 1)33. Following BAFF-binding, TACI directly interacts with MyD8833 resulting in activation of NF-κB via a signaling cascade dependent on IRAK1, IRAK4, TNFR-associated factor 6 (TRAF6), TGFβ-activated kinase 1 (TAK1) and IκB kinase (IKK). Strikingly, TACI engagement, in conjunction with cytokines or TLR ligands, markedly facilitates class-switch recombination (CSR) 33; suggesting that TACI signaling via MyD88 may help promote extrafollicular B cell responses that lead to rapid and sustained IgG and IgA responses to T-independent antigens, including encapsulated bacteria polysaccharides. TACI–MyD88 interactions may also enhance antibody responses by delivering survival signals to activated extrafollicular B cells, including class-switched plasmablasts. This mechanism may help to explain the clinical observations in patients with MyD88 or IRAK4 deficiency, as well as those with common variable immune deficiency (CVID) syndrome and mutations in TACI, who present with hypogammaglobulinemia and impaired IgG responses to T-independent antigens34,35.
In summary, MyD88 signaling in B cells is crucial for T-independent responses by activating and positioning B cells, thereby resulting in rapid and appropriately localized, pathogen-specific, antibody production. In addition, B intrinsic, TACI-triggered, MyD88 signals appear to promote extrafollicular class-switch recombination facilitating improved primary, protective, T-independent, antibody responses.
B cells are recruited into T-dependent immune responses primarily from the FM subset, after which they enter a germinal center and subsequently differentiate into either antibody-secreting plasma cells or memory B cells. Although innate-like B cells generally display stronger responses to TLR engagement alone, antigen-mediated B cell responses can be modulated by simultaneous BCR and TLR engagement (Figure 2). In vitro, follicular B cells proliferate, produce cytokines and secrete antibody following stimulation with various TLR ligands3,4,15,36. Purification of mouse germinal center B cells revealed enhanced TLR ligand-responsiveness and increased expression of Myd88 mRNA compared with follicular B cells, suggesting that follicular B cells may become more sensitive to TLR engagement during the course of a T-dependent immune response16. Additionally, TLR4 signaling in B cells has been proposed to increase migration to lymph nodes and accumulation in germinal center dark zones, perhaps helping to sustain ongoing germinal center reactions37.
A role for TLRs in promoting plasma cell differentiation has been proposed based on immunoglobulin secretion and expression of the differentiation factors B lymphocyte-induced maturation protein 1 (BLIMP1) and X-box-binding protein 1 spliced isoform (XBP1s) in response to TLR ligands in follicular B cells, although in the absence of co-stimulation these responses were primarily elicited in B-1 and marginal zone B cells15. In humans, naïve B cells are minimally responsive to TLR ligands, but memory B cells proliferate and differentiate into antibody-secreting plasma cells in response to the TLR9 ligand CpG DNA38. These data led to the suggestion that stimulation of human memory B cells through TLRs may contribute to long-term maintenance of serological memory38. A later study showed that maximal activation of naïve human B cells required a combination of BCR engagement, T cell help via CD40 signaling and TLR stimulation39.
Together, the above data suggested that, in both mice and humans, B cell intrinsic TLR signals can promote events associated with T-dependent immune responses. This has led to significant interest in determining the in vivo role of MyD88 in B cells in the response to protein antigens. This question was first addressed by transferring either wild-type or MyD88-deficient B cells into μMT mice and analyzing the response to intraperitoneal immunization with human serum albumin (HSA) and LPS co-adsorbed on the adjuvant alum40. In this setting, production of HSA-specific IgM and IgG1 antibodies was severely impaired in mice with MyD88-deficient B cells. Similar results were obtained after immunization with flagellin, and the authors provided evidence that MyD88 signaling promotes enhanced antigen presentation by B cells, as well as differentiation of germinal center B cells and plasma cells. This study concluded that activation of TLRs in B cells is necessary for antibody responses to T-dependent antigens.
A subsequent study further tested this idea by immunizing wild-type and MyD88- and TIR-domain-containing adaptor protein inducing IFNβ (TRIF)-double deficient mice with trinitrophenol-hemocyanin (TNP–Hy) or TNP–keyhole limpet hemocyanin (TNP–KLH) as antigens in combination with various adjuvants1. TNP-specific antibody responses were largely intact in double deficient mice, although IgG2b and IgG2c titers were slightly lower in response to TNP–KLH in Ribi adjuvant. In a third study, wild-type or MyD88-deficient B cells were transferred to μMT mice, which were then immunized with the antigen 4-hydroxy-3-nitrophenylacetyl (NP)–chicken γ-globulin (CGG) in alum16. During both the primary immune response and following re-challenge with NP–CGG 4 months after initial immunization, NP-specific IgM and IgG levels were unaffected in recipients of MyD88-deficient B cells. However, addition of LPS during the primary immunization greatly increased the antibody response to NP–CGG, and increased NP-specific IgM and IgG2a levels required MyD88 expression in B cells, whereas increased NP-specific IgG levels did not. The ability of TLR signals to activate memory B cells was also tested by injecting μMT mice that had received wild-type or MyD88-deficient B cells with LPS several months after a series of NP–CGG immunizations. LPS induced a transient increase in both total and NP-specific IgM and IgG levels. Together, these latter two studies support a model in which B cell MyD88 signaling is not required to generate T-dependent antigen-specific antibody responses, but such signals can augment early antibody production, influence CSR and promote differentiation of memory B cells into plasma cells.
Several factors may help explain the discrepancy in the results from these studies. TLR engagement on B cells might be required in settings where BCR engagement or perhaps T cell co-stimulation is limited. Thus, it is possible that decreased levels of BCR crosslinking by HSA compared with TNP–KLH (or NP–CGG) may affect the requirement for MyD88 signals in B cells41. Additionally, a comparison of responses to HSA with or without the hapten dinitrophenol (DNP) indicated that antibody response to HSA with LPS required MyD88 signaling, whereas DNP–HSA in incomplete Freund’s adjuvant elicited equivalent responses in wild-type versus MyD88- and TRIF-double deficient mice, leading to the suggestion that haptenated antigens are inherently immunogenic, thereby limiting the requirement for MyD88 signals42.
In addition to variability in the antigens used in these studies, recent reports suggest an important role for the physical form of the TLR ligands during adaptive immune responses. Using either myeloid cell- or B cell-specific deletion of MyD88, DeFranco and colleagues investigated how various forms of a TLR ligand influence T-dependent immune responses43. Ovalbumin (OVA)-specific antibody responses to OVA with soluble CpG DNA, OVA with an aggregated form of CpG DNA, and OVA covalently linked to CpG DNA depended on MyD88 expression in DCs but not B cells. In contrast, antigen-specific IgG responses to immunization with CpG DNA incorporated in proteinacious virus-like particles (VLPs) that were derived from the Qβ bacteriophage depended largely on MyD88 expression in B cells but not DCs. B cell-specific MyD88-deficent mice were deficient in IgG2b and IgG2c Qβ-specific antibodies and this correlated with a dramatic defect in differentiation of germinal center B cells. Consistent with a critical role for B cell intrinsic MyD88 signals in the humoral response to viruses, the IgG influenza virus-specific response generated following immunization with inactivated H1N1 virus was significantly reduced in B cell-specific MyD88-deficient mice43.
These data are consistent with the previous demonstration that human papillomavirus type 16 (HPV16) major caspid protein L1-containing VLPs can directly induce CSR in B cells, dependent on intrinsic MyD88 expression44. In another study, μMT mice reconstituted with TLR9-deficient B cells and challenged with Qβ VLPs containing CpG DNA showed a dramatic reduction in VLP-specific IgG2a titers, further supporting a role for B cell intrinsic MyD88 signaling in driving CSR45. Of note, major T-dependent immune deficits have not been identified in MyD88 or IRAK4 deficient patients: however, published data are largely limited to total IgG responses to robust T-dependent antigens (for example, tetanus toxoid- and diptheria toxin-derived antigens)31.
Infection of mice with live pathogens has supported a role for MyD88 in modulating humoral immunity, although this has not always been shown to be B cell intrinsic. Both TLR7 and MyD88 were shown to be important for CSR to IgG2a and IgG2c in response to influenza virus infection46. Alternatively, B cell-specific MyD88 expression was shown to be dispensable for the early T-dependent antibody response to mouse polyoma virus infection but was required for the maintenance of long-term antibody production47. A recent study examining the failed response to respiratory syncytial virus (RSV) vaccines in infants concluded that insufficient TLR signaling in B cells resulted in low affinity, non-protective antibody production48. Further, infection with live RSV failed to elicit virus-specific IgG in B cell-specific MyD88-deficient μMT mice48.
These findings are not limited to viral infections; one report described a B cell intrinsic role for MyD88 signaling to generate an appropriate primary Th1 cell response to Salmonella enterica36. In that study, mixed bone marrow chimeras were used in which the B cell compartment was deficient in MyD88. At one week post-infection, T cells isolated from these chimeras exhibited deficient IFNγ and IL-10 secretion following challenge with S. enterica antigen. Using similar mixed bone marrow chimeras in which B cells lacked expression of specific cytokines, a supporting role for B cell secretion of IL-6 and IFNγ for the development of Th17 and Th1 responses, respectively, was identified36.
Collectively, these data demonstrate that MyD88 signaling in B cells can make important contributions to T-dependent antibody responses, and the dependence on B cell MyD88 varies depending on the nature of both the protein antigen and the TLR ligand. In particular, B cell MyD88 signaling appears to contribute to: driving CSR to IgG2a (or its equivalent, e.g. IgG2c, in C57BL6 mice) during primary T-dependent responses; promoting the differentiation of germinal center and memory B cells into antibody-secreting cells; and supporting effector T cell differentiation through cytokine secretion (Figure 2).
Recent work demonstrates that MyD88 signals and most notably, dual TLR and BCR engagement, influence mechanisms of B cell tolerance (Figure 3). Further, new data suggest that genetic changes that alter BCR and/or TLR signaling thresholds likely promotes these events; and autoreactive B cells triggered in this way may directly break T cell tolerance facilitating germinal centre reactions that lead to pathogenic autoantibody production.
Despite a diverse repertoire of potential autoantigens, many autoimmune diseases are characterized by a restricted autoantibody repertoire. In addition to pathogens, TLRs can also recognize self-ligands, in particular nuclear antigens released from apoptotic cells. Further, dual TLR and BCR engagement in B cells has been implicated in the initial activation of autoreactive B cells, helping to explain the propensity of these cells to develop antinuclear antibodies (ANAs)49–51.
The importance of dual BCR and TLR activation was initially demonstrated in vitro using B cells expressing the AM14 BCR, which is a low-affinity BCR for autologous IgG2a52. Stimulation of AM14 B cells with DNA–IgG2a or DNA-associated protein–IgG2a immune complexes resulted in B cell proliferation, while protein-specific immune complexes did not. MyD88 signaling downstream of TLR9 was critical for activation of AM14 B cells suggesting involvement of dual BCR and TLR signals52. Similarly, RNA- or RNA-associated protein-containing immune complexes activated AM14 B cells via TLR753. Furthermore, a critical role for MyD88 signaling in the generation of RNA- and DNA-specific autoantibodies in vivo was demonstrated in MyD88-deficient MRL-lpr and MRL-gld mice. In contrast to littermate controls, these mice do not have detectable levels of ANAs, with significantly reduced RNA- and DNA-specific autoantibodies and decreased glomerulonephritis53,54.
Importantly, two alternative mechanisms could explain the importance of MyD88-dependent signaling for autoantibody development in vivo in these autoimmune models: e.g. direct B cell intrinsic signals mediated via dual BCR and TLR activation vs. an indirect effect of TLR- and immune complex-mediated activation of plasmacytoid dendritic cells leading to increased type I interferon production55. To date, the requirement for B cell-intrinsic MyD88 signals has only been investigated in a limited number of autoimmune models. Using a mixed bone marrow chimera strategy, Groom et al. demonstrated that deficiency of MyD88 specifically in B cells decreases autoantibody production and glomerular immunoglobulin and complement deposition in BAFF-transgenic mice, stressing the importance of B cell-intrinsic MyD88 signaling in BAFF-triggered autoimmunity56. MyD88 deletion also prevented CSR to pathogenic IgG2a and IgG2b in FcγRIIb-deficient B cells expressing an anti-DNA specific BCR heavy chain, suggesting that the role of MyD88 is B cell-intrinsic in this model57.
Our group recently characterized a model of autoimmunity in which B cells, but not other hematopoietic lineages, lack Wiskott–Aldrich syndrome protein (WASP)49. In the absence of WASP, peripheral B cells are mildly hyperresponsive to BCR and TLR ligands, and mice with WASP-deficient B cells develop SLE-like autoimmunity characterized by spontaneous germinal center formation, generation of pathogenic IgG2b and IgG2c autoantibodies, glomerulonephritis and early mortality. Chimeric mice with B cells deficient in both WASP and MyD88, however, fail to develop any autoantibodies or signs of autoimmune disease, demonstrating a critical role for B cell-intrinsic MyD88 signaling in this model. Disease development also required wild-type T cells, suggesting that enhanced BCR and TLR signals are sufficient to drive a B cell-intrinsic, MyD88-dependent loss of T cell tolerance.
The relative contribution of specific TLR signals to MyD88-dependent autoimmune disease has been addressed by several groups50,51. Notably, TLR7 and TLR9 exhibit divergent roles in mouse autoimmune disease. Briefly, TLR7-deficient MRL–lpr mice54 have decreased RNA-specific autoantibody titers but preserved autoimmune responses to double-stranded DNA (dsDNA). In contrast, TLR9-deficient MRL–lpr mice have markedly decreased levels of autoantibodies specific for dsDNA and chromatin, but elevated titers of autoantibody specific for RNA and RNA-associated protein58,59. Tlr7−/− autoimmune-prone mice are protected from immune complex-mediated glomerulonephritis, whereas glomerulonephritis and mortality is exacerbated in the absence of TLR958,60. While the events responsible for this TLR9-mediated protective effect remain unclear, it is interesting that accelerated autoimmunity in Tlr9−/− mice depends on intact TLR7 signaling54,60. Importantly, whether TLR7 and TLR9 signaling influences autoimmunity in a B cell intrinsic manner has not been addressed.
In addition, MyD88 signaling may impact autoimmunity independently of TLR7 and TLR9 via TACI signaling33 or, alternatively, via its role in IL1 or IL18 signaling. TACI-deficient mice develop spontaneous lymphoproliferative and autoimmune disease and CVID patients with TACI mutations exhibit a higher risk for autoimmunity61, suggesting a protective role for TACI. Overall, both B cell-intrinsic and B cell–extrinsic MYD88-dependent TLR signals likely promote autoimmune disease, with TLR7 and TLR9 playing opposing roles. However, the cell-intrinsic requirements for individual TLR signals remain to be defined.
Direct evidence for B cell-intrinsic MyD88 signaling in the pathogenesis of human autoimmune disease is lacking. However, genome-wide association studies have implicated genetic polymorphisms that affect TLR signaling pathways in susceptibility to SLE. Notable among these are mutations in IFN-regulatory factor 5 (IRF5)62, a transcription factor involved in production of IFNα following TLR ligation, and in IRAK163. Two genes involved in ubiquitin-mediated downregulation of NF-κB signaling, TNF-induced-protein 3 (TNFAIP3; also known as A20) and its ligation partner TNFAIP3-interacting protein 1 (TNIP1; also known as ABIN1), have also been associated with human SLE, possibly implicating TLR-mediated NF-κB activation in disease pathogenesis64,65. In murine studies, B cells from B cell-intrinsic A20−/− mutant66 and polyubiquitin-binding defective Abin1[D485N] knockin mice67 demonstrate enhanced TLR-mediated NF-κB activation and both B cell-restricted A20−/− and Abin1[D485N] mice develop lupus-like systemic autoimmunity. Disease development in Abin1[D485N] mice is MyD88-dependent, thus implicating TLR-mediated NF-κB activation in the pathogenesis of human SLE. Although not validated via genome-wide association studies, candidate-gene approaches have suggested associations between TLR7 or TLR9 polymorphisms and autoimmune disease68–70.
Interestingly, despite the role for TLR–MyD88 signaling in activation of autoreactive B cells, an unanticipated increase in autoreactive naive B cells was observed in patients with MyD88 or IRAK4 deficiency2. A high number of immature B cells from these patients exhibit BCRs that bind nuclear antigens and deletion of these autoreactive B cells prior to entry into the mature peripheral compartment is defective. Together, these data suggest that the normal mechanisms of central and peripheral B cell tolerance may depend on intact MyD88 signaling. However, despite the increase in autoreactive B cells, patients with MyD88 or IRAK4 mutations did not display increased serum ANA titers, perhaps consistent with a requirement for MyD88 also in the activation of such autoreactive B cells.
Although B cell-intrinsic TLR signals can potentially break tolerance and trigger autoimmune disease, these same signals also play a role in regulatory B cell function, helping to suppress immune responses and maintaining self-tolerance (Figure 4). It has been long-recognized that both murine and human B cells produce IL-1071,72. Recently, this capacity to generate high levels of IL-10 has defined populations of B cells with regulatory function. Two distinct phenotypic definitions for regulatory B cells have been described in mice- CD19+CD21hiCD23hiCD1dhi marginal zone-precursor B cells73 and CD1dhiCD5+ B cells74. Based on their function, these populations have been termed regulatory B cells (Breg cells)75,76 and B10 B cells74, respectively. As IL-10 production can be triggered via a variety of stimuli, it remains unlikely that these cells comprise a distinct B cell developmental subset. Also, although not discussed here, B cells may also utilize transforming growth factor-β (TGFβ) secretion to exert their regulatory activity77,78.
The importance of IL-10-producing B cells has been demonstrated in a range of mouse models of autoimmune disease including inflammatory bowel disease75, collagen-induced arthritis73, experimental autoimmune encephalomyelitis (EAE)79, and SLE20. Additional data indicate a regulatory function for B cells in infectious models with Leishmania major80, Schistosoma mansoni81, Brugia pahangi82, Salmonella typhimurium23 and Helicobacter pylori83.
Intrinsic TLR stimulation appears to be crucial for modulating the activity of IL-10-producing B cells, predominantly via promoting their differentiation and expansion. In vitro, stimulation of mouse splenic B cells via TLR9 results in high levels of IL-10 production and, interestingly, the amount of IL-10 is significantly higher in B cells from lupus-prone than wild-type mice84. In splenic marginal zone and marginal zone-precursor B cells, IL-10 production can be triggered via TLR4 or TLR9 engagement20. TLR4 ligation, in addition to polyclonal B cell stimulation, induced cytoplasmic IL-10 expression and rapid clonal expansion of splenic CD1dhiCD5+ B cells74. While the development of IL-10-producing CD1dhiCD5+ B cells was normal in Myd88−/− mice, direct LPS-induced IL-10 production and secretion was significantly reduced. In a mouse model of EAE, B cell-intrinsic TLR signals resulted in the suppression of both Th1 and Th17 cell-mediated inflammatory T cell responses and facilitated recovery from disease85.
MyD88-dependent, IL-10-producing B cells can also modulate the response to infectious challenge. B cells exert regulatory function in S. typhimurium infection that requires both B cell intrinsic MyD88 signalling and IL-10 secretion23. MyD88-deficient mice infected intravenously with virulent S. typhimurium exhibit increased mortality, whereas mice with B cell-specific MyD88 deficiency exhibit prolonged survival. Notably, B cell-intrinsic MyD88 signals appear to drive two distinct programs in this setting. First, B cell-specific Myd88−/− mice exhibit lower basal levels of total and S. typhimurium-specific natural IgM and delayed S. typhimurium-specific IgM and IgG responses after infection, indicating that MyD88 signals accelerate humoral immunity. In contrast, bone marrow chimeras with IL-10 deficiency restricted to B cells exhibit increased survival23, suggesting that B cell-derived IL-10 limits immunity to S. typhimurium.
In a model of H. pylori-induced gastric immunopathology, infection activates B cells in a MyD88- and TLR2-dependent manner83. Activated B cells subsequently promote IL-10 production by CD4+CD25+ T regulatory type 1 cells in vitro and in vivo, suppressing excessive gastric H. pylori-associated pathology. In contrast to the S. typhimurium infection model, however, these events do not require B cell intrinsic IL-10 expression.
IL-10-producing B cells with a regulatory function have also been identified in humans by two independent groups86,87. IL-10-expression was identified in B cells derived from peripheral blood, spleen and tonsils following stimulation with either LPS or CpG DNA plus CD40 ligation86,87. Notably, suppressive function and IL-10 production was impaired in peripheral B cells derived from patients with SLE87.
In summary, B cells with regulatory function differentiate and exert their suppressor function mainly by IL-10 production upon TLR signaling. Most data to date have been obtained from murine studies and the importance of MyD88 signals in initiating suppressive B cell function has been demonstrated in at least two infectious models. Additional studies are required to determine whether MyD88 signaling is required for the generation or function of IL-10-producing B cells in humans.
Approximately 90% of human lymphomas arise from B cells, and diffuse large B cell lymphoma (DLBCL) comprises the most common type of non-Hodgkin’s lymphoma. Several distinct DLBCL subtypes have been defined based on genetic signatures reflecting their putative origins, including (treatment-resistant or poor-prognosis) activated B cell-like (ABC) DLBCL88. ABC DLBCL probably originate from early stages of a germinal center reaction and this tumor type is characterized by high-level constitutive NF-κB activation89 and coordinate high level expression of PKCβ, the BCR-dependent upstream activator of this signaling cascade90,91.
RNA interference screens of ABC DLBCLs identified the importance of the CARD11–BCL-10–MALT1 complex downstream of the BCR/PKCβ in driving NF-κB activation92. Consistent with this, somatic mutations in the scaffold protein caspase recruitment domain-containing protein 11 (CARD11; also known as CARMA1)93 and in BCR-associated signaling effector molecules CD79A and CD79B94 are present in ~10% and 20% of ABC DLBCLs, respectively. Recently, MyD88 mutations have also been identified in 39% of ABC DLBCLs, thereby also implicating TLR-mediated activation of NF-κB in lymphomagenesis95. Strikingly, a single substitution (Leu256Pro) within the Toll/IL-1-receptor (TIR) domain of MyD88 accounted for the majority of these mutations. Leu256Pro MyD88 was noted to be a gain-of-function mutant resulting in spontaneous assembly of a signaling complex with IRAK1 and IRAK4 and enhanced NF-κB activation. Interestingly, Leu256Pro MyD88 also activated JAK–STAT3 (Janus kinase–signal transducer and activator of transcription 3) signaling resulting in enhanced production of IL-6 and IL-10, likely promoting autocrine survival signals 96,97.
Most remarkably, MyD88 mutations were noted in ABC DLBCLs that also contained CARD11 or CD79A and CD79B mutations that impact on BCR signaling95. Thus, in a manner analogous to both viral-driven B cell activation and B intrinsic autoimmunity, dual BCR and TLR signals appear to promote the survival of a subset of refractory B cell lymphomas. Finally, MyD88 mutations are not limited to ABC DLBCLs. The same Leu256Pro mutation was also observed in 9% of gastric mucosa-associated lymphoid tissue lymphomas and 3% of chronic lymphocytic leukaemia98 highlighting the importance of dysregulated TLR signaling in B cell-derived malignancies.
Atherosclerosis is a chronic inflammatory disease that is impacted by TLR responses. TLRs, in particular TLR4, can be activated by endogenous oxidized lipid epitopes and apolipoprotein E (Apoe)−/− mice that lack express of MyD88 or TLR4 have decreased levels of atherosclerosis compared with Apoe−/− control mice99,100. Furthermore, a human Asp299Gly TLR4 polymorphism that decreases TLR4 signaling is associated with decreased risk of atherosclerosis101.
Although the atheroprotective effects of MyD88 and TLR4 deficiency have been hypothesized to involve an absence of TLR signals in tissue macrophages or foam cells within atherosclerotic plaques, there is now growing interest in the atheroprotective and atherogenic roles for B cells in disease. While B cell-deficient102 and serum IgM-deficient103 low-density lipoprotein receptor (LDLR)-deficient mice develop increased aortic atherosclerotic lesions, an effect attributed to the loss of protective natural IgM antibodies, B cells can also actively promote atherosclerosis, particularly in autoimmune settings104. Further, B cell depletion with CD20-specific antibody decreases the development of atherosclerosis in both Ldlr−/− and Apoe−/− mouse models of atherosclerosis105,106 and protection correlates with a decline in oxidized low density lipid (oxLDL)-specific IgG and coordinate preservation of oxLDL-specific IgM titers106. Patients with SLE have a markedly increased risk of atherosclerotic cardiovascular disease107, and those with a history of cardiovascular events have elevated oxLDL-specific and malondialdehyde-modified LDL-specific IgG antibody titers compared with other SLE patients or population controls108, suggesting that these autoantibodies may promote atherogenesis. While not yet directly addressed, given the role of B cell-intrinsic MyD88 signals in the pathogenesis of autoimmunity, it will be interesting to determine whether B cell MyD88-TLR signals promote accelerated atherosclerosis in animal models and patients with autoimmune disease.
B cell-intrinsic MyD88 signals and, most notably, integration of dual TLR and BCR signals play a crucial role in B cell responses to pathogen challenge. This unique functional program is crucial for protective immune responses to specific viruses and, most likely, also to encapsulated bacteria. Strikingly, new findings suggest that this hard-wired program also predisposes to pathogenic B cell responses including humoral autoimmunity and B cell malignancies. Key open questions include how developmental events as well as local microenvonmental and inflammatory cues (or other cellular co-stimuli) function to specifically modulate these events; and most importantly, whether is will be possible to manipulate such signals to develop novel immunization strategies. Finally, while most of our knowledge to date is derived from mouse studies, it will be important to progressively extend investigation to human B cells to optimally leverage or curtail these responses for translational applications.
This work was supported by NIH grants HD037091, HL075453, AI084457 and AI071163 (to D.J.R.); a Cancer Research Institute Pre-doctoral Training Grant (to M.A.S.); a Rheumatology T32 Postdoctoral Training grant 5T32AR007108 (to S.W.J.) and a German Research Foundation (Deutsche Froschungsgemeinschaft) grant ME2709/2-1 (to A.M.B.).