Dysregulation of T Follicular Helper Cells in Lupus

Abstract

Although multiple and overlapping mechanisms are ultimately responsible for the immunopathology observed in patients with Systemic Lupus Erythematosus (SLE), autoreactive antibodies (Ab) secreted by autoreactive plasma cells (PCs), are considered to play a critical role in disease progression and immunopathology. Given that PCs derive from the germinal centers (GC), long-term dysregulated GCs reactions are often associated with the development of spontaneous auto-Ab responses and immunopathology in SLE patients. In this review, we summarize the emerging evidence concerning the roles of T-follicular helper (Tfh) cells in regulating pathogenic GC and autoreactive PC responses in lupus.

Introduction

Multiple factors, including genetic and environmental determinants, contribute to SLE susceptibility and pathogenesis. Despite that SLE is a multifactorial disorder and multiple mechanisms contribute to disease severity, Auto-Ab against nuclear and cytoplasmic antigens are thought to play a critical role in disease progression and damage (1, 2). More commonly, the auto-Ab-associated pathology is mediated through the formation of immune complexes (IC) with circulating apoptotic cell debris (3, 4). These IC can directly initiate inflammation through the ligation of activating-Fc receptors (FCRs) or TLR signal induction (5). In addition, trapping of large ICs in blood vessels, kidney and lungs contribute to the vasculitis, renal disease and lung manifestations that are characteristic of SLE. Thus, in addition to the primary defect of loss of tolerance, the spectrum of auto-Ab specificities, their affinity, particular isotype, and titers influence the clinical manifestations of SLE.

The primary sites of peripheral regulation of antibody production are the GCs. In normal conditions, the GCs play key roles in the maintenance of tolerance, as well as class switching, affinity maturation and the development of plasmablasts and memory B cells. AutoAbs are present up to 10 years before the onset of clinical disease in patients with SLE (6, 7), and in RA (8, 9). Moreover, at the time of onset of chronic clinical disease, the levels of certain autoAbs and their pathogenic potential, in terms of specificity and affinity maturation, appear to have maximized (7, 9–11). Superimposed on the association of the autoantibodies with the chronic immune disease, is the periodic occurrence of disease “flares”, which appear to be associated with an acute rise in the titers of certain autoAbs together with enhanced affinity and pathogenicity (12–15). Although the life-span of autoAb-producing GCs is not known, it is unlikely that the GCs that produce pathogenic autoAbs persist for years. It has therefore been proposed that several cycles of GC reactions in different GCs lead to the development of the pathogenic autoAbs associated with the onset of disease (16).

Central questions in autoimmunity are therefore what gives rise to spontaneous autoreactive GCs and how do they differ mechanistically from “healthy” GCs? There is now a considerable body of evidence that Tfh cells, which are essential for the formation and maintenance of GCs, play a central role in the development of autoimmunity by helping deregulated GC responses. Tfh cells are a specialized subset of CD4⁺ T cells that, after being primed by dendritic cells (DCs) in the outer zone of the B cell follicles, home into the B cell follicles where they produce survival and differentiation signals that are essential for the development and maintenance of the GCs (17, 18). Phenotypically, Tfh cells are characterized by the expression of the transcription factor BCL6, which is fundamental for the establishment of the Tfh program; the chemokine receptor CXCR5, which enables the migration into the B cell follicles; and inhibitory molecule PD1, that stabilizes the Tfh phenotype, contribute to the GC localization (19), and prevents excessive proliferation in the GCs (20–23). Other essential pathways in Tfh differentiation include PI3K-mTORC (24, 25) ICOS-Foxo1 (26) and OX40L (27).

The central role of spontaneous GCs in autoimmunity

The GCs that develop normally in response to immunization or after infection with a pathogen are well characterized. They have a well organized light zone and dark zone, arise rapidly, are self-limiting and leading to the development of plasmablasts within weeks (28). The origin and organization of germinal centers in autoimmune mice (and especially in humans) have been more difficult to analyze since they arise spontaneously without a known immunization (29). Extra follicular aggregates of T and B cells that form GC-like structures have been observed in synovial tissue in rheumatoid arthritis, AICDA expression has been described, and are thought to be a key source of highly mutated anti-CCP autoantibodies (30–33). However such GC-like structures are rare and may require CD8 T cells for their development (34). Studies of GC organization in the Fas-mutant MRL-Fas^lpr/lpr mouse model have led investigators to the conclusion that spontaneous GCs arise in a non-classical extra-follicular location outside the follicle such as in the marginal zone (35) where TLR signaling but not Tfh may play an important role in development (36). However the significance of these mis-localized GCs in Fas mutant mice is unclear since the splenic architecture is greatly disrupted by hyperplasia of abnormal CD4-CD8-B220+ T cells (37–40). Furthermore SLE is a complex, multi-genetic disease that is not well represented by single gene mutant Fas^lpr/lpr, FasL^gld/gld, BXSB-Yaa or Ptpn6^me mice, which are prominent in the literature primarily because generation of transgenic or knockout mice bearing these single mutation is straightforward (41–43). GC development and architecture has been studied in two multi-genetic strains of mice, the BXD2 (44–49) and the NZB/W and related models which are congenic at loci containing the NZB/W gene defects (50–52). Similar to induced GCs, spontaneous germinal centers in the BXD2 lupus mouse model are well organized into a light zone and dark zone, with CXCR5⁺ CD4⁺ Tfh localized in the light zone where CD4 T cells and GC B cells are adjacent to CXCL13 producing follicular dendritic cells (FDCs) (Figure 1). In conclusion, the overall architecture of spontaneous GCs that produce pathogenic autoantibodies in lupus mice that highly resemble SLE patients may be very similar to what has been observed for induced GCs, necessitating the detailed analysis of abnormal Tfh cells and cytokines that drive their development at the cellular and molecular level.

Figure 1. Classical Tfh cells in the light zone of a spontaneous GC in BXD2 mice.

Expression of CXCR5 on both Tfhs and B cells assist their co-localization near CXCL13 expressing CD35⁺ FDCs in the light zone of the GC. The representative PNA⁺ GC (blue) GC was derived from the spleen of a BXD2 mouse. The structure of this spontaneous GC is similar to immunized GCs in which CXCR5⁺ (red) CD4⁺ (green) Tfh cells (yellow) are polarized to the light zone (LZ) end. In contrast, CXCR5⁻ cells are polarized to the dark zone (DZ) end (44) (Ding et al, J Immunol, 2013).

The role of the cytokine milieu in abnormal Tfh cell responses in SLE

Abnormal differentiation and expansion of Tfh cells is observed in patients with SLE and in lupus-prone mice. Although the exact mechanisms that drive Tfh cell differentiation remain unclear, it is well established that the relative balance between the levels of expression of the transcription factors Bcl6 and Blimp-1 greatly influence Tfh differentiation. In agreement with this, Bcl6 promotes the Tfh cell differentiation program, whereas Blimp-1 represses it (53–55). Importantly, the Bcl6/Blimp-1 ratio is known to be tightly regulated by certain cytokine signaling pathways, suggesting that an abnormal cytokine milieu has the potential to contribute to the development of spontaneous Tfh cell responses, thereby inducing spontaneous GC reactions in SLE. In agreement with this idea, deregulated cytokine production contributes to immune dysfunction in SLE patients and is often associated with abnormal expansion of Tfh cells in patients and lupus-prone mice. Herein, we will focus on deregulated cytokine pathways with the potential to drive the development of spontaneous Tfh cell responses in SLE.

Interleukin-2 (IL-2)

IL-2 is a member of the cytokine receptor γ-chain family of cytokines that is central to the maintenance of immune tolerance (56). IL-2 and IL-2 receptor-deficient mice develop a catastrophic autoimmune syndrome characterized by the aberrant expansion of self-reactive T and B cells (57–59). There is a strong association of single nucleotide polymorphisms (SNPs) in the IL-2 and the IL-2R genes and autoimmune pathogenesis in preclinical murine models (60, 61) and in patients with autoimmune disease (62–67). T cells from lupus-prone mice (60, 61) and patients with SLE (62–67) characteristically produce low levels of IL-2, which further suggests a potential relationship between deficiencies in IL-2 signaling and SLE pathology (66).

A large body of evidence indicates that IL-2 signaling is required for the development, function, and survival of FoxP3⁺-regulatory T cells (Tregs), specialized CD4⁺ T cells that suppress self-reactive T and B cell responses and are fundamental for the maintenance of immune tolerance (68). Thus, abnormal Treg cell function due to dismissed IL-2 signaling is generally considered the main mechanism underlying the close association between limited IL-2 production and autoimmune disease development in SLE patients (65). While it is clear that IL-2 is important in Treg cell homeostasis, it has not been demonstrated formally that this is the only mechanism by which IL-2 deficiencies contribute to autoimmune pathogenesis in SLE. Recent data show that IL-2 is a potent inhibitor of Tfh cell differentiation (69–73). Data from us demonstrate that excessive IL-2 signaling following treatment with recombinant-IL-2 directly suppresses Tfh cell responses in an in vivo model of influenza infection, thereby hindering the development of influenza-specific GCs and the subsequent long-term Ab response to influenza (69). Notably, in this model, the capacity of IL-2 to suppress Tfh cell responses is independent of its role in promoting Treg cell-mediated suppression, as selective depletion of Treg cells does not preclude the capacity of IL-2 to inhibit Tfh cell responses (69, 70). These studies reveal an immunosuppressive function of IL-2 that is independent of its role on Treg cells. This concept is supported by the multiple and overlapping mechanisms by which IL-2 can prevent Tfh cell differentiation. For example, IL-2 signaling directly promotes Blimp-1 up-regulation via STAT5, which in turn represses Bcl6 expression and Tfh cell differentiation (71, 72). STAT5 also can bind directly to the Bcl6 promoter (73, 74), thereby recruiting repressing chromatin modifiers to the Bcl6 locus that prevent Bcl6 upregulation (73–75). In agreement with the negative effect of the IL-2/STAT5 axis in the initiation of the Tfh cell program, STAT5 and IL-2Rα (CD25) deficiency in CD4⁺ T cells has been shown to result in enhanced Tfh cell responses during T cell priming in vivo (69, 71, 72). Conversely, expression of a constitutively active form of STAT5 selectively prevented Tfh cell formation (71, 72). Besides directly suppressing Bcl6 expression, IL-2 signaling also negatively regulates the biological activity of Bcl6 through the activity of the transcription factor, T-bet (73). Specifically, elegant data from Weinmann and colleagues demonstrate that elevated IL-2 signaling in developing Th1 cells favors the formation of T-bet/Bcl6 complexes that mask the DNA-binding domain of Bcl6, thus preventing it from binding to its target genes (73). While these mechanistic findings are largely based on murine studies, a recent study by Crotty and colleagues demonstrate that IL-2 is a negative regulator of Tfh cell differentiation in humans (76). In agreement with the negative effect of IL-2 in Tfh cell development, it has been reported that treatment of SLE patients with low-dose IL-2 selectively prevents Tfh cell expansion (77). This not only provides further evidence of a causative correlation between IL-2 signaling deficiencies and Tfh cell development but is clinically relevant in that it suggests that the shortage can be corrected by administration of exogenous IL-2. Collectively, these studies indicate that the physiological availability of IL-2 is an important factor in the control of Tfh cell differentiation in vivo. They suggest that while excessive IL-2 signaling inhibits Tfh cell responses, deficiencies in IL-2 signaling skew CD4⁺ T cell responses towards the Tfh cell differentiation pathway (69, 71, 72). Based on these data, it is possible that, in addition to perturbing Treg cell function, deficiencies in IL-2 signaling in patients with SLE contribute to the autoimmune pathology by favoring the development of a self-reactive Tfh cell response.

Interleukin-6 (IL-6), Interleukin 21 (IL-21) and other STAT3-activating cytokines.

Polymorphisms in the IL-6 gene are associated with the risk of SLE susceptibility (78, 79), suggesting a role for IL-6 in the pathogenesis of lupus. Higher levels of IL-6 are found in the serum of patients with SLE than healthy individuals (80, 81) and IL-6 deficiency has been shown to limit Ab-mediated pathology and delay lupus nephritis in lupus-prone mice (82–84). IL-6 is known to positively regulate Tfh cell differentiation. For example, during the early stages of T cell priming in mice, IL-6 signaling through the IL-6R induces activation of STAT1 and STAT3, which in turn favor upregulation of Bcl6 in the responding T cells (85, 86). IL-6 signaling also is required for the long-term maintenance of Tfh cell responses during chronic viral infections in mice (87). In humans, the importance of the STAT3 axis in Tfh cell differentiation is highlighted by the fact that patients with dominant-negative mutations in the Stat3 gene have diminished Tfh cell responses (88).

Evidence of an association between IL-6 and Tfh differentiation includes the demonstration that IL-6 deficiency in mouse models of lupus prevents the expansion of self-reactive Tfh cells. Furthermore, in a Wiskott–Aldrich syndrome (WAS) protein deficient-mouse model of lupus, it has been shown that B cell-specific induction of IL-6 deficiency precludes spontaneous Tfh cell expansion as well as preventing the development of self-reactive GC reactions and immunopathology (84). The positive association between increased IL-6 production, abnormal expansion of Tfh cells and autoimmune disease pathology suggests that one of the potential mechanisms by which higher levels of IL-6 production in patients with SLE could contribute to disease development is by favoring the development of spontaneous self-reactive Tfh cell responses.

In some experimental models, Tfh cells can differentiate normally in the absence of IL-6 (89–91), which suggests that IL-6 is not absolutely necessary for Tfh cell development. This is likely due to the capacity of other STAT3/1-activating cytokines, particularly IL-21 (90, 92), IL-23 (93) and IL-27 (94, 95), to activate STAT3/1. In fact, similar to IL-6, IL-21 also activates STAT3 and promotes Bcl6 expression and Tfh cell differentiation (20, 90, 96). The role of IL-21 in Tfh cell development is, however, controversial since Tfh cells normally differentiate in the absence of IL-21 in some studies (97), most likely because STAT3 activation via IL-6 compensates for the lack of IL-21(98). Importantly, however, single-nucleotide polymorphisms (SNPs) within the Il21 and Il21r genes associate with increased susceptibility to SLE (98), and increased levels of IL-21 in the serum of lupus patients correlates with disease severity (99, 100), thus suggesting a cause-relationship between deregulated IL-21 responses and SLE.

Importantly, recent studies demonstrate that, following infection or protein immunization, some Foxp3⁺ FoxP3⁺CD4⁺ Treg cells up-regulate Bcl6 and CXCR5, and home into the B cell follicles where they suppress Tfh and GC B cell responses. This particular subset of Bcl6⁺CXCR5⁺Treg cells is known as T follicular regulatory (Tfr) cells (101). Importantly, the relative balance between Tfh and Tfr cells, which respectively promote and repress GC responses, may play an important role in the enforcement of tolerance in the GC (102, 103). Supporting this view, the Tfh/Tfr ratio is increased in SLE patients (104). Interestingly, while IL-21 favor Tfh cell development, it prevents Tfr responses(105, 106). In agreement with this, we have recently found that IL-21 skews the balance from Tfr cells to Tfh cells, thereby promoting autoreactive GC reactions in BXD2 mice (45). Thus, in addition to directly favoring GC and Tfh cell responses, IL-21 signaling may also contribute to promote autoreactive GC responses by limiting Tfr cell development.

Increased serum levels of IL-23, another important regulator of STAT3 and Tfh cell differentiation in humans (93), also correlates with disease activity in SLE patients (107). Furthermore, IL-23 deficiency prevents the accumulation of self-reactive Tfh cells, leading to decreased production of anti-dsDNA Abs and glomerulonephritis in Il23r^−/− MRL-Fas^lpr/lpr mice. Taken together, these studies suggest that therapeutic approaches aimed at targeting STAT3-activating cytokines may represent an effective strategy for preventing self-reactive Tfh cell responses and treating Ab-mediated immunopathology in patients with SLE.

Interleukin-17 (IL-17)

IL-17 is thought to play an important role in SLE pathology. For example, while elevated serum levels of IL-17 correlate with disease activity in SLE patients (108), IL-17 deficiency prevents autoAb production and subsequent disease progression in lupus-prone mice (107, 109, 110). Similarly, lack of IL-23 signaling, which is essential for the differentiation of IL-17-producing T cells, precludes the development of lupus nephritis in C57BL/6-Fas^lpr/lpr mice (110).

Elegant in vitro studies show that IL-17 synergizes with BAFF to protect B cells from BCR-mediated apoptosis (111), suggesting a direct effect of IL-17 in B-cell survival. In addition, Th17 cells, secrete large amounts of IL-21 and are efficient B cell helpers that enforce the development of autoreactive GC B cells and extrafollicular plasma cells (44, 46, 112–116). Thus, IL-17-producing cells have the potential to promote autoreactive B cell responses by both direct and indirect mechanisms. Interestingly, there is a significant overlap between the Tfh and Th17 cell developmental pathways. For example, similar to Th17 cells, the IL-2/STAT5 axis prevents, and the IL-6/STAT3 axis promotes, Tfh cell differentiation. (117, 118). ICOS is important for the development of both Tfh and Th17 cells (119). Similarly, TGFβ and IL-23, two cytokines that positively favor human Th17 cell differentiation, are implicated in the acquisition of the Tfh cell differentiation program (120, 121). These data suggest that some of the critical regulators of the Th17 cell differentiation program are also important regulators of the Tfh differentiation program. Normally, a distinguishing feature of Tfh cells is their expression of Bcl6 that represses the Teff program associated with Th1, Th2 and Th17 differentiation (53–55) and prevents production of Teff cytokines, including IL-17. Recent studies demonstrate, however, that Bcl6-mediated repression can be overcome in some circumstances, such as acute viral infections or autoimmune disease. This allows the initiation of secondary differentiation programs in response to a highly polarizing cytokine environment (122). Supporting this view, Tfh cells produce significant amounts of IL-17 in lupus-prone BXD2 mice (44, 46). In this mouse model, IL-17 production by the Tfh cells was shown to affect GC functional organization. It was found to prevent the chemotactic response of GC B cells to CXCL12 and CXCL13 thereby leading to prolonged retention of responding B cells in the GCs that promoted the production of autoAbs. In addition, intrinsic IL-17 signaling was required for the localization of Tfh cells in the light zone (44). Thus, while it is generally believed that Th17 cells contribute to disease pathogenesis by promoting tissue inflammation, these studies provide evidence of a critical role for IL-17-producing Tfh (Tfh-1L-17) cells in promoting self-reactive B and Tfh cell responses.

Type I Interferon (IFN)

It is well established that patients with SLE have a type I IFN signature. Multiple studies suggest that type I IFNs support Tfh cell differentiation. For example, data from John O’Shea’s laboratory indicate that IFN-α/β signaling in activated T cells directly induces Bcl6 expression through a STAT1-dependent mechanism (123). Using mixed bone marrow chimeras, Riteau and colleagues demonstrated that T-cell intrinsic type I IFN signaling is required for optimal Tfh cell responses following protein or peptide immunization (124). In another study, type I IFN signaling enhanced class-switched Ab responses to protein immunization, but this effect was prevented when T cells were IFNAR -deficient (125), indicating that IFN-α/β signaling in CD4⁺ T cells directly promotes Tfh cell differentiation. Interestingly, conditional depletion of IFNAR in DCs abrogates Tfh cell responses in an experimental model of protein immunization. Mechanistically, IFN-α/β signaling in conventional dendritic cells (DCs) synergizes with LPS and CD40 stimulation to maximize their capacity to produce IL-6, which in turns favors Tfh cell formation as described above (126). IFNα/β also augments IL-6 production by activated B cells, which contributes to the development of autoimmune pathology in several experimental models (83, 91, 127, 128). Thus, in addition to intrinsically promoting Bcl6 expression and Tfh cell differentiation, IFN-α/β signaling seems to indirectly promote Tfh cell responses by maximizing the capacity of antigen presenting cells to induce the Tfh cell program. Taken together, these studies open the possibility that IFN-α/β could play an important role in SLE by driving the spontaneous expansion of self-reactive Tfh cell responses, thereby promoting Ab-mediated pathology.

Stage-specific roles of Tfh cells in the development of GCs.

In autoimmune GCs, Tfh programs may be altered that further enable spontaneous GCs to form. Bcl6 regulates key aspects of their role in GC programs, such as their localization, the differentiation of Tfh cells and their expression of cytokines, such as IL-21 (55), as described above. Among its numerous roles, perhaps the most fundamental is its stabilization of the entire Tfh developmental program and the prevention of its diversion into other Th programs. In lupus, the lack of a strong immunogenic stimuli might preclude development of an organized wave of Tfh transcription factors and cytokines but rather enable development of a mixed population of Tfh cells. One model for the development of different Tfh cells in the same GC is suggested by the recent finding that Tfh development and expression of cytokines is highly asynchronous (129). Through the use of a double-reporter mouse (IL21-IRES-KAT and IL4-IRES-GFP), the investigators could simultaneously track expression of IL-21 and IL-4 during GC-Tfh cell development. The earliest GC-Tfh cells expressed KAT but not GFP suggesting early expression of IL-21 in the absence of IL-4. At later stages, the Tfh cells expressed both IL-21 and IL-4, whereas in the final stages, they expressed only IL-4. It is possible that Bcl6 repression of a Th2 program in the early Tfh cells led to a more classical IL-21 producing Tfh, but later, Bcl6 was down regulated enabling other programs and expression of IL-4 to emerge. This sequence of Tfh evolution is consistent with what is known regarding the effects of the cytokines on GC B cells. IL-21 is the most potent stimulus for survival, proliferation and especially, plasma cell differentiation, whereas IL-4, and to a lesser extent IL-21, promotes upregulation of AID, SHM and class switching, which would occur at the later stages of the GC B-cell response. These finding also suggest a Tfh development mechanism that is protective from SLE, since IL-4 suppressed responses to type I IFNs and is deceased in SLE patients (130–132).

Multiple Tfh subpopulations in autoimmune pathologic GCs

A non-canonical Tfh development scheme has been observed in the spontaneous GCs of BXD2 mice (44). Classically, Bcl6 suppresses T-bet and STAT4 (133), thereby inhibiting their acquisition of a Th1 cell phenotype and suppresses STAT3 and RORγt, thereby suppressing acquisition of a Th17 phenotype (53). However, these suppression functions are not absolute and Tfh production of IL-4, interferon-γ, and IL-17 has been described. Consistent with the results generated by Craft and co-workers described above (129), our laboratory has observed that both conventional Bcl6⁺CD4⁺ IL-17^- Tfh and CD4⁺ IL-17⁺ Tfh were localized in the light zone of the same spontaneous GC of BXD2 mice (Fig. 2A). Also, the CXCR5⁺ICOS⁺ Tfh cells localized in the spontaneous GC of BXD2 mice could express high levels of IL-21 or IL-17, but not both (44). FACS analysis and sorting showed that in BXD2 mice, Tfh cells that had higher expression of IL-17 were CXCR5^- but ICOS⁺ whereas the more conventional Tfh cells with higher expression of IL-21 were CXCR5⁺, suggesting greater localization with the FDCs and earlier evolution that the IL-17 expressing Tfh (44). In terms of the canonical transcription factors, FACS analysis and sorting showed that in BXD2 mice, Bcl6 was strongly associated with CXCR5⁺ICOS⁺, whereas RORγt was prominent in the ICOS⁺CXCR5⁻ and the ICOS⁺CXCR5⁺ cells (44). The Tfh-IL-17 cells were less dependent upon colocalization with FDCs but more dependent on activation in B cell interaction, suggested by Tfh-IL-17 as predominately CXCR5^loICOS⁺. We propose a model of autoreactive GCs in which CXCR5⁺ ICOS⁺Bcl6⁺, IL-21⁺ CD4 T cells initially enter the GC. However, since Bcl6 suppression of other T cell programs is not absolute, some of these Tfh cells undergo reprogramming to downregulate CXCR5 and upregulate RORγt and IL-17 resulting in relocation of the Tfh cells such that they are in close association with ICOSL-expressing B cells. This interaction is further maintained through IL-17 regulation through RGS signaling in both T and B cells.

Figure 2. Autoreactive GC development require cooperation between multiple Tfh subpopulations.

(A). Confocal image of a spontaneous GCs result from interactions of CD4⁺ Tfh B cells that express Bcl6 and a separate population of CD4 T cells that express IL-17 in the LZ of the GC. The representative BCL6⁺ GC (red) is derived from the spleen of a BXD2 mouse. The GC is simultaneously stained with an anti-mIL-17 (white) and an anti-mCD4 (green) to illustrate the distribution of Tfh-IL-17⁺ cells in the GC (objective lens 20×). (B) In the spleen, development of mixed populations of Tfh occur at the late naïve stage (CD62L⁺CD44⁺, Population C) when there is high expression of IL-21, IL-17, CXCR5 and ICOS. This population has undergone approximately 4 cell divisions as indicated by a quantitative real-time PCR-based TREC assay(148), and expresses the highest levels of IL-21 or IL-17 as analyzed by intracellular flow cytometry(44).

One question is whether accessory Tfh cells are required to initiate spontaneous GCs or whether they develop secondarily to secure spontaneous GC development. In the early stages after formation of a normal GC in response to an antigen, the antigen presentation by FDC and B cells is limiting. The B cells with the highest levels of bound-surface antigen out-compete B cells with lower affinity and their interactions with T cells at the Tfh at the initial T-B border and gives such T-B cell interactions a competitive advantage. High-antigen binding B cells can divide, expand and survive, whereas their Tfh partners are restrained by their expression of PD1. In the context of autoimmunity, this initial encounter is a low -affinity interaction that does not lead to a strong Tfh - GC B developmental program. Thus, it is reasonable to speculate that the most important role for IL-17 lies in enhancement of the T-B cells interactions at this early stage of GC development simultaneous with development of conventional Tfh cells. Consistent with this, FACS analysis and sorting of CD4 T cells from BXD2 mice indicate that in the spleen, both conventional CD4⁺ Tfh and Tfh-IL-17-develop at the same developmental stage marked by TCR excision circles (TRECs) and the expression of CD62L and CD44 (Fig 2B, Population C). This population has undergone approximately 4 cell divisions as indicated by the TREC assay, and expresses the highest levels of IL-21 or IL-17 by FACS.

Another important accessary Tfh cells that play an important role to facilitate spontaneous GCs is Th1 cytokine IFN-ᵧ which plays a prominent role in lupus (134).

IFN-ᵧ producing Tfh-like cells have been observed in the choroid plexus of MRL-Fas^lpr/lpr mice (135). In BXD2 mice with altered environment, an accessory IFN-ᵧ-producing Tfh (Tfh-IFN-ᵧ ) can replace the need for an accessory Tfh -IL-17 (136). Tfh-IFN-ᵧ and Tfh -IL-17, however, act through different mechanisms to potentiate the effects of canonical Tfh (44, 45, 136). The mechanism for development of Tfh-IFN-ᵧ has been analyzed during viral infection where upregulation of T-bet and STAT4 initiate expression of co-expression of IFN-ᵧ and IL-21 for the duration of the infection (133, 137). As seen with LCMV viral infection, a chronic inflammatory state such as lupus may result in increased STAT4 and continued expression of IL-21 and interferon-ᵧ during active disease, despite reduced T-bet expression (137). Increased IL-21 and interferon-ᵧ promotes development of abnormal CD11c T-bet⁺ B cell population that produces pathogenic autoantibodies, and these B cells reciprocally promote development of pathogenic Tfh or Th1 T cells (138). IL-12 signaling through STAT4 further stabilizes the T-bet and Th1 phenotype for development of Tfh-IFN-ᵧ, signaling through the inducible T cell co-stimulator molecule (ICOS) (139). ICOS-induced up-regulation of Bcl6 and interactions with B cells at the T-B border reinforces the Tfh-IFN-ᵧ development. At this point, Tfh-IFN-ᵧ T cells enter the follicle and drive a GC response through production of cytokines IL-21 and either IL-4 or IFN-ᵧ (90, 129). The preference to develop into Th1 and Tfh-IFN-ᵧ is regulated by factors in co-stimulator molecules outside the T cells. Continued IL-12 and STAT4 signaling down-regulates Bcl6 and IL-21 expression leading to chromatin remodeling that secures T-bet driven Tfh-IFN-ᵧ cell differentiation over Tfh. Conversely, signaling by ICOS, IL-6 or TGF-β preferentially promotes Tfh development compared to Tfh-IFN-ᵧ, which is seen during a virus infection (86). The two predominant pathways for development of pathogenic GCs including mixed populations of Tfh is shown in Figure 3.

Fig 3. Accessory Tfh cells in a spontaneous pathologic autoreactive GC.

Diagram of spontaneous GCs result from interactions of T cells with B cells and FDCs at the T-B border similar to those associated with induced GCs except that the affinity of the T-cell interaction is lower and autoantigen abundance may be higher. Autoreactive GCs that arise spontaneously and is enabled by incorporation of mixed populations of Tfh cells. The primary pathogenic Tfh are development of Tfh-IL17 from RORᵧt expressing Tfhs or Tfh-IFNɣ from Tbet. Such accessory Tfh cells in the GCs can lead to increased expression and engagement of co-stimulatory molecule interactions, such as CD28-CD86, ICOS-ICOS-ligand, or CD40 ligand-CD40.

Circulating Tfh in SLE

As mentioned above, Tfh cells are not confined to the GC and can migrate into the peripheral blood (22). Due to the difficulty in obtaining lymphoid organs from patients with SLE, characterization of Tfh cells in SLE and other autoimmune diseases has relied largely on analysis of Tfh cells isolated from the peripheral blood, known as circulating Tfh cells. The phenotype of the circulating Tfh does not meet all the criteria required for classical GC-associated Tfh cells, however, and so they are commonly referred to as “Tfh-like” cells or “activated Tfh-like” cells due to their expression of high levels of PD1. Notably, several investigators have reported that the expression of Bcl6 is not an essential feature of circulating Tfh-like cells in SLE, although there is one report of high expression of Bcl6 in circulating Tfh-like T cells in SLE (140). As ICOS is expressed on several subpopulations of activated CD4 T cells, it is not useful for identification of bone-fide circulating Tfh. The presence of CXCR5^hi and PD1^hi circulating Tfh-like CD4 T cells in SLE is associated with the SLEDAI, and the degree of PD1 expression in human Tfh cells is the strongest indicator for pathogenic B cell responses in disease activity (141). In this context, the expression of high levels of PD1 is more important than the levels of expression of CXCR5 since PD1⁺CXCR5^lo CD4⁺ T cells correlate with disease flares in SLE (142). Moreover, the highest levels of IL-21 are found in PD1^hi and CCR7^lo Tfh-like cells. The numbers of Tfh-like cells have been correlated with the numbers of circulating plasmablasts in SLE. Other investigators have found that CXCR5⁺ CD4 helper T cells or Tfh-like cells correlate with plasmablasts and autoAb production. (143)

Therapeutic directions

Given that long-lived PCs (69–71) and extrafollicular PC differentiation (72, 73) are dependent on help provided by Tfh cells, therapeutic interventions aimed at specifically targeting Tfh are likely to be an effective alternative for treating Ab-mediated pathology in SLE patients (74). Supporting this notion, blockade of Tfh cell activity prevents autoAb production and disease progression in lupus-prone mice (75). Thus, blockade of Tfh cell products such as IL-21 or CD40, or blockade of cytokine and/or signaling pathways that positively regulate Tfh cell differentiation might be a good strategy to prevent Tfh cell activity and intervene in the case of Ab-mediated disorders. Supporting this view, treatment of patients with RA with Tocilizumab, a humanized anti-IL-6 receptor monoclonal antibody, led to a significant reduction in circulating Tfh cell numbers and IL-21 production, which correlated with reduced PC responses (76). Similarly, Abatacept (CTLA4-Ig) treatment selectively reduced the frequency of circulating follicular Tfh cells in primary Sjögrens Syndrome (77).

Recent studies show that low-dose IL-2-based immunotherapy has potent immunosuppressive effects in patients with autoimmune disorders (25, 78–81) and reduces disease activity in patients with active and refractory SLE (25). Based on these results, low-dose IL-2-based therapy is considered a promising new approach for the treatment of SLE and clinical investigations are currently underway (82–84). Although the exact mechanism by which IL-2 induces immunosuppression remains unclear, it is currently believed that IL-2 prevents immunopathology by promoting expansion of Treg cells (82, 85). This has led to the investment of considerable effort in the development of new therapeutic approaches that selectively target IL-2 to Treg cells (86–92). However, based on the putative role of Tfh cells in supporting Ab-mediated pathology and the inhibitory role of the IL-2/STAT5 axis in Tfh cell development (93, 94), it is possible that Tfh cell suppression represents a major mechanism underlying the immunosuppressive effects of the low-dose rIL-2 immunotherapies. Thus, the potential effects of IL-2 on preventing Tfh cell responses should now be considered when evaluating the immunosuppressive effects of the low-dose IL-2 immunotherapies.

While the IL-2/STA5 axis inhibits Tfh cell differentiation, STAT3-activating cytokines promote Tfh cell responses by favoring Bcl6 upregulation. Thus, it is tempting to speculate that blockade of the JAK-STAT3 pathway (or blockade of cytokines that promote STAT3 activation) will render Tfh cells more sensitive to the low-dose IL-2 therapies by lowering the threshold of IL-2 required to prevent Bcl6 expression.

The potential utility of CTLA4 as an inhibitor of GC development has been demonstrated in the BXD2 lupus mouse model, in which autoimmune disease depends on very large and active GCs. Treatment with Ad-CTLA4 Ig, which produces high levels of circulating CTLA4 that can block CD28-CD86 interactions, greatly inhibited the development of autoAbs and autoimmune disease, including nephritis, in the BXD2 mice (144). Furthermore, the development of GCs, autoAbs and autoimmune disease was not dependent upon prolonged circulation of CTLA4-Ig, which lasted 10–14 days. When young BXD2 mice were administered Ad-CTLA4-Ig at 2 months of age, before significant autoAbs and autoimmune disease was apparent, they remained autoAb and disease free for up to 8 months of age. Furthermore, the treatment induced long-lived TFRs that persisted for months after the administration of the CTLA4-Ig. These results are consistent with several reports that CTLA4-Ig promotes the function of Treg cells in patients with RA (145, 146). It also has been reported that CTLA4-Ig may enhance IL-10 production by Treg cells and 1,25(OH)2D3 promotes the efficacy of CD28 costimulation blockade by abetacept (147). Since there are several types of Treg cells, the role of CTLA4-Ig in the development and function of Tfr cells, and Tfh cells, needs further study especially as CTLA4-Ig may be most useful in promoting Tfrs specifically and not Treg cells in general.

Conclusion

In conclusion, the current data indicate that spontaneous autoreactive GCs that arise in lupus exhibit share core regulatory interactions with normally induced GCs. These similarities include expression of key transcription factors, cell-surface signaling molecules and cytokine production. In common with normally induced GCs, Tfh cells are required for autoreactive GCs and their localized development in the light zone of GCs plays an essential role in promoting somatic hypermutation and class switching in the GC B cells. The critical differences in the spontaneous autoreactive GCs appear to be associated with deviations that redirect the overall developmental program. The current data indicate that abnormal levels of key cytokines, including IL-2, IL-6, IL-17 and/or type 1 interferon, can be sufficient to promote development of autoreactive GCs. Dysregulated cytokine signaling as well as other factors can initiate a second mechanism that is associated with compromised regulation of the Tfh cell differentiation program. That is, the autoreactive GCs are unique in terms of the occurrence of multiple, functionally diverse subpopulations of Tfh cells, which appears to be due primarily to abnormal regulation of the Bcl6 transcription factor. This emerging view that Tfh programs in autoreactive GCs are heterogeneous is consistent with the observation that SLE is not a single disease, but exhibits distinct clinical features often associated with distinct autoantibody profiles. While much of the work summarized in this review is drawn from animal models, an improved mechanistic understanding of the molecular and cellular dynamics of Tfh cells among patients with SLE should promote development of improved precision medicine-guided therapeutic approaches.