Introduction

Coordinated interaction between different immune cells is crucial for the development of protective immunity. Nowhere is this more obvious than in the emergence of high affinity antibody responses, where carefully orchestrated contacts between dendritic cells (DCs), T cells and B cells culminate in a refined and long-lived antibody response. The critical go-between in this cellular trio is the follicular helper T (TFH) cell that liaises first with the DC, before migrating to the B cell follicle for repeated interaction with B cells1. Such TFH cells bear a characteristic phenotype including expression of markers such as CXC-chemokine receptor 5 (CXCR5; which promotes homing to the B cell follicle), programmed cell death protein 1 (PD1), inducible T cell costimulator (ICOS) and the transcription factor BCL-6 (ref.2) (Box 1). TFH cells are classically found in secondary lymphoid organs with a small population of similar cells present in the blood (referred to as circulating TFH (cTFH) cells). Curiously, multiple studies have revealed that cTFH cells are present at increased frequencies in many autoimmune diseases, including systemic lupus erythematosus (SLE), rheumatoid arthritis (RA), Sjogren syndrome (SS), autoimmune thyroid diseases, myasthenia gravis, type 1 diabetes (T1D) and multiple sclerosis (MS)3. Consistent with elevations in cTFH cell numbers, autoantibodies are commonly associated with these conditions and their presence often precedes symptomatic disease.

In this Perspective, I discuss core mechanisms controlling TFH cell differentiation and highlight that these same pathways are linked to the regulation of autoimmune disease. I also discuss the potential for infection to serve as a link between TFH cells and autoimmunity. Overall, I postulate that regulation of TFH cells and regulation of autoimmunity are tightly coupled, perhaps explaining why increases in cTFH cell numbers are evident across multiple autoimmune diseases.

Discovery of cTFH-like cells

In 2005, a seminal publication from Vinuesa et al. showed that mutant mice with dysregulated TFH cell differentiation exhibited systemic autoimmunity4. The causative mutation in these animals mapped to the Roquin (Rc3h1) gene, the product of which repressed ICOS expression and negatively regulated TFH cell differentiation. These sanroque mice exhibited lupus-like pathology and robust development of T1D when crossed to a T cell receptor (TCR) transgenic mouse model. Impairing TFH cell differentiation, by rendering sanroque mice heterozygous for BCL-6 or deficient for SLAM-associated protein (SAP, an adaptor protein required for TFH cell–B cell interactions5), ameliorated the autoimmune phenotype, leading to reduced autoantibody production and decreased renal pathology6. These findings sparked interest in whether TFH cell differentiation might be connected to lupus pathology in humans, leading to the demonstration that cells with a TFH cell phenotype were elevated in the blood of patients with SLE7. Reports documenting cTFH-like cells in numerous other autoimmune diseases rapidly followed (reviewed in ref.8).

The spotlight was then turned on the true identity of these blood-borne TFH-like cells, specifically on their relationship to germinal centre (GC)-resident TFH cells. Important work from the Ueno group revealed that circulating CD4+CXCR5+ T cells shared functional properties with TFH cells from secondary lymphoid organs and could provide help to B cells via IL-21 production9. This led to the idea that blood-borne CD4+CXCR5+ T cells were a memory counterpart of lymphoid TFH cells, and a substantial body of evidence has since emerged to support this view10,11,12,13,14. Consistent with this notion, individuals with defects in GC formation due to deficiency in ICOS or CD40L, or a developmental block in B cell development (BTK deficiency), have fewer circulating CD4+CXCR5+ cells, and they are absent altogether from the cord blood of newborns15,16. Of note, many TFH cell markers are downregulated in the memory phase12,17,18,19, perhaps explaining early controversies over the existence of a TFH cell memory pool. Experiments involving the adoptive transfer of mouse TFH cells into antigen-free hosts indicate that CXCR5 expression is least affected by the phenomenon of TFH cell marker loss20, suggesting it may be the most reliable marker for tracking cTFH cells. Painstaking experiments in which cTFH cells were purified from mouse blood revealed that these cells can home to secondary lymphoid tissues upon re-challenge and participate in GC reactions, providing direct evidence of their functional capacity upon antigen re-encounter18. Thus, the consensus view is that TFH cells have a circulating memory compartment that retains expression of CXCR5 and, to a variable extent, other TFH cell markers and can be recalled to participate in the memory phase of humoral responses.

Somewhat unexpectedly, the majority of cTFH cells do not seem to derive from the GC environment itself. In fact, imaging studies have shown that, although GC TFH cells frequently move between GCs, their egress to the circulation is rare21. Instead, it appears that circulating CXCR5+ cells arise mainly from TFH cells that have not yet undergone sustained interaction with B cells. In line with this, the frequency of cTFH cells appears undiminished in mice and humans that lack SAP expression and consequently exhibit impaired T cell–B cell interactions22. The link between lymphoid tissue TFH cells and cTFH cells has been compellingly demonstrated both by mass cytometric comparison of blood and tonsillar tissue with dimensionality reduction approaches23 and by vaccination studies showing that blood CD4+CXCR5+ICOS+PD1+ cells24 or CD4+CXCR5+PD1hi cells25 are clonally related to lymph node GC TFH cells. Overall, this suggests a model in which TFH cells at the T cell–B cell border can give rise to both GC-resident TFH cells and cTFH cells that travel through efferent lymph to the blood (Fig. 1). Consistent with this, T cells with TFH cell features can be detected in thoracic duct lymph of mice and humans, and treatment with the drug FTY720, which prevents lymph node exit, dramatically decreases cTFH cells in both species18,26,27.

Fig. 1: Blood-borne cTFH cells are clonally related to their lymphoid counterparts.
figure 1

Follicular helper T (TFH) cells classically reside in secondary lymphoid organs where they support germinal centre (GC) B cell responses; however, T cells with similar markers can be detected in the lymph and blood. These circulating TFH (cTFH) cells have a memory phenotype and do not depend on GCs as they can arise in the context of SLAM-associated protein (SAP) deficiency, which abrogates sustained T cell–B cell interactions22. However, they do require B cells, as cTFH cell frequencies are greatly reduced in individuals with B cell deficiency16. Blocking lymph node exit, by treatment with FTY720, dramatically decreases cTFH cells in mice and humans18,26,27. Clonotype sharing reveals a developmental relationship between TFH cells found in the GC and the blood24,25 and those found in the lymph and the blood26, suggesting that T cell–B cell interactions at the T cell–B cell border give rise to both GC TFH cells and cTFH cells. It is possible that some GC TFH cells enter the circulation from GCs but this is likely to be rare21. Follicular regulatory T (TFR) cells also have circulating counterparts (cTFR); however, these have an immature phenotype and differentiate from regulatory T cells without the requirement for contact with B cells164.

Common pathways regulate TFH cell differentiation and autoimmunity

TFH cell differentiation and provision of help to B cells is a tightly controlled process. B cell tolerance relies heavily on restricting T cell help, in part because the developmental regulation of T cells is more stringent; nascent T cells are screened against self-antigens during their thymic development, with transcription factors, such as autoimmune regulator (AIRE), ensuring a broad representation of peripheral self-antigens. By contrast, peripheral self-antigens may be less available to B cells developing in the bone marrow, and it has been reported that around 20% of mature naive B cells exhibit low levels of self-reactivity28. B cell antigen specificity is also prone to diversification within GCs, superseding any developmental constraints on self-reactivity. Thus, denying T cell help to self-reactive B cells is necessary to prevent the initiation of potentially damaging autoimmune humoral responses. Ensuring that T cell–B cell collaboration is tightly regulated is also important for optimal protective immunity (Box 2).

In considering the mechanisms controlling TFH cell differentiation, three core and interconnected pathways emerge involving cytotoxic T lymphocyte antigen 4 (CTLA4), regulatory T (Treg) cells and IL-2. Curiously, these same pathways are well recognized for their ability to regulate autoimmunity. Below, I explore evidence indicating that these pathways control TFH cell differentiation and highlight their connection to autoimmune disease susceptibility.

CTLA4-mediated regulation of TFH cells

CTLA4 controls T cell CD28 costimulation and is a crucial regulator of T cell responses (Box 3). Multiple lines of evidence pinpoint the CTLA4–CD28 axis as a key modulator of TFH cell responses (Box 3). In mice, genetic deficiency or antibody-mediated blockade of CTLA4 triggers spontaneous TFH cell differentiation and GC formation29. GCs that form following CTLA4 blockade are dependent on CD28 (ref.29), consistent with the established role of CTLA4 in regulating CD28 engagement and the importance of the CD28 pathway for TFH cells and GCs30,31. Experiments with CD28 heterozygous mice revealed a relationship between the amount of CD28 engagement and propensity for TFH cell differentiation29, suggesting that CTLA4 may regulate this fate by policing access of CD28 to its ligands. CD28 likely promotes TFH cell differentiation in multiple ways, including increasing ICOS expression, which drives strong phosphoinositide 3-kinase (PI3K) activation important for TFH cell formation32, as well as regulating microRNAs implicated in TFH cell fate29,33.

There is also evidence that the CTLA4 pathway regulates TFH cells in humans. Individuals deficient in LRBA, that show defective CTLA4 trafficking and function, exhibit increases in circulating T cells expressing TFH cell markers (CXCR5 and PD1)34. In addition, patients with cancer receiving anti-CTLA4 antibody immunotherapy show an increase in circulating T cells with TFH cell markers (E. Ntavli, N. M. Edner and L.S.K. Walker, unpublished observations). Conversely, cTFH cells are decreased following treatment with soluble CTLA4 molecules, such as the CTLA4–immunoglobulin fusion protein abatacept, in individuals with SS35, RA36, MS37 or T1D38. Thus, the CTLA4 pathway negatively regulates TFH cell homeostasis in mice and humans, likely by restricting CD28 engagement.

CTLA4 and autoimmunity

The association between CTLA4 and autoimmunity is well documented. Genetic variation at the CTLA4 locus is linked to numerous autoimmune diseases, including T1D, RA, SLE, myasthenia gravis, autoimmune thyroid diseases, coeliac disease, alopecia areata and vitiligo (see GWAS Catalogue). Mice genetically deficient for Ctla4 develop lethal lymphoproliferation and multiorgan immune cell infiltration39,40, and heterozygous CTLA4 mutations in humans are associated with an immune dysregulation syndrome with multiple autoimmune manifestations41,42. Targeting the CTLA4 pathway by immunotherapy in patients with cancer can also elicit autoimmune side effects. CTLA4 function may be altered indirectly by mutations in genes encoding CTLA4 pathway regulators. For example, mutations in LRBA lead to reduced CTLA4 expression and autoimmune outcomes43.

Treg cell-mediated control of TFH cells

Treg cells express the transcription factor FOXP3 and play a crucial role in the maintenance of immune homeostasis. Scurfy mice, which lack functional Treg cells owing to a frameshift mutation disrupting Foxp3, exhibit a marked expansion of BCL-6+CXCR5+ TFH cells in secondary lymphoid tissues44. Consistent with this, in mice expressing diphtheria toxin receptor under the control of the Foxp3 promoter, short-term depletion of Treg cells enhances the generation of antigen-specific TFH cells in response to immunization45,46. Similar to mice, patients with IPEX (immune dysregulation, polyendocrinopathy, enteropathy, X-linked) that have mutations in FOXP3 also exhibit an increased frequency of CXCR5+PD1+ cTFH cells47. Thus, FOXP3+ Treg cells appear to control TFH cell numbers in both mice and humans.

Treg cells constitutively express CTLA4. Interestingly, the enhanced TFH cell differentiation associated with CTLA4 deficiency29 can be recapitulated by loss of CTLA4 expression in Treg cells alone45,48. To avoid widespread immune dysregulation, Sage et al.48 used mice in which tamoxifen-inducible Foxp3-Cre was used to excise the floxed Ctla4 gene in Treg cells immediately prior to immunization, whereas Wing et al.45 probed the impact of partial loss of CTLA4 expression using heterozygous Ctla4-flox/wt mice expressing Foxp3-Cre. In both settings, increases in TFH cells were observed after immunization. Collectively, these findings illustrate that Treg cells are a non-redundant population mediating CTLA4-dependent regulation of TFH cells.

The above observations could potentially reflect the role of CTLA4 in follicular regulatory T (TFR) cells, the subset of Treg cells that exhibit TFH cell features and enter GCs44,46,49. For Treg cells to acquire the TFR cell programme they downregulate CD25 (IL-2Rα) expression, permitting them to express BCL-6, which would otherwise be antagonized by IL-2-driven BLIMP1 expression50,51. TFR cells are thought to optimize protective antibody responses while suppressing the generation of antibodies to self-antigens and allergens44,46,49,52,53,54 and, like TFH cells, TFR cells have a circulating counterpart that can be altered in autoimmunity55,56 (Fig. 1). Interestingly, selective depletion of TFR cells, either in mice with floxed Bcl6 and Foxp3-driven Cre expression54,57,58 or in mice with STOP-floxed Cxcr5-DTR and Foxp3-Cre53, results in only minor or transient increases in TFH cells despite marked effects on B cell differentiation. As loss of CTLA4 expression in FOXP3+ cells increases TFH cells but loss of TFR cells does not, it is possible that substantial CTLA4 activity occurs outside of the B cell follicle during the initial encounter between T cells and DCs. At the T cell–B cell border and within the follicle, TFR cells may employ additional suppressive mechanisms such as expression of the neuropeptide neuritin57 or targeting of B cell metabolic pathways59.

In a fascinating twist, it appears that FOXP3-based regulation of TFH cells can also operate in a cell-intrinsic manner. TFH cells themselves upregulate FOXP3 expression in late-stage GCs, and this is associated with loss of expression of the T cell help-associated genes IL21 and CD40L and GC collapse60. These FOXP3+ TFH cells express high levels of CTLA4 and are reminiscent of the CD25 TFR cells described by Wing et al.61, the transcriptional profiles of which place them equidistant between TFH cells and activated Treg cells. The division of labour between Treg cells, TFR cells and FOXP3+ TFH cells will need to be dissected by further experimentation. Taken together, Treg cell populations play a key role in controlling TFH cell numbers in both mice and humans, potentially via the CTLA4 pathway.

Treg cells and autoimmunity

Many of the genes associated with susceptibility to autoimmunity are expressed in Treg cells62 and the pre-eminent role for Treg cells in regulating autoimmunity is well recognized. Mice lacking Treg cells develop lethal autoimmunity63 and humans with an impaired Treg cell compartment as a result of mutations in FOXP3 develop the aggressive early-onset immune dysregulation syndrome IPEX64. Interestingly, deficits in Treg cells can interfere with normal costimulatory control of T cell immunity — the unexpected exacerbation of disease in CD28-deficient non-obese diabetic mice was reconciled by the discovery of the role of CD28 in Treg cell development65, and recent findings suggest CD28 also contributes to Treg cell homeostasis in humans66. A replete Treg cell compartment is therefore key to the normal regulation of immune responses, and strategies aimed at augmenting Treg cell numbers, by low-dose IL-2 treatment or Treg cell therapy, are being actively pursued in settings of autoimmunity.

IL-2-mediated regulation of TFH cells

The IL-2 pathway is recognized as a major regulator of TFH cell differentiation (Fig. 2). In mice, exogenous provision of IL-2 has been shown to suppress TFH cell differentiation both in the context of viral infection67 and autoimmunity68. In humans, IL-2 is also a known regulator of TFH cell differentiation69, and low-dose IL-2 therapy can decrease numbers of cTFH cells in individuals with autoimmune disease70.

Fig. 2: IL-2-based inhibition of TFH cells and adaptations to subvert this.
figure 2

a | IL-2 plays an important role in promoting FOXP3+ regulatory T (Treg) cell homeostasis; Treg cell-deficient mice44 and humans47 exhibit increased follicular helper T (TFH) cell numbers. b | During T cell priming, IL-2-induced signalling through the IL-2 receptor (IL-2R) upregulates BLIMP1 expression, skewing the BLIMP1 to BCL-6 ratio and inhibiting TFH cell differentiation74,75,76,77. Transforming growth factor-β (TGFβ) decreases IL-2R signalling by suppressing CD25 (IL-2Rα) expression80. c | CD25-expressing dendritic cells (DCs) in the outer T cell zone can capture local IL-2, favouring TFH cell differentiation82. d | Within the germinal centre (GC), IL-6 can desensitize IL-2R signalling in mature TFH cells by downregulating the IL-2R subunit CD122 (IL-2Rβ)83. APC, antigen-presenting cell; CTLA4, cytotoxic T lymphocyte antigen 4; CXCR5, CXC-chemokine receptor 5; FDC, follicular dendritic cell; ICOS, inducible T cell costimulator; ICOSL, ICOS ligand; PD1, programmed cell death protein 1; TCR, T cell receptor.

One potential mechanism by which IL-2 inhibits TFH cells is by enhancing Treg cell homeostasis as IL-2 is a key cytokine for Treg cell expansion and maintenance71,72,73. However, it is clear that IL-2 can also act directly on conventional T cells to inhibit TFH cell differentiation. During T cell priming, IL-2 signalling can alter the balance between TFH cell and non-TFH cell effector differentiation74,75 via STAT5-mediated skewing of the BLIMP1 to BCL-6 ratio76,77. In an interesting twist, T cells induced to make IL-2 are not those that respond to it — instead, the producers become TFH cells while using IL-2 in a paracrine manner to preclude their neighbours from this fate78. This finding establishes a biological role for the synaptic-based delivery of IL-2 between adjacent T cells undergoing activation79.

Such is the threat to TFH cell differentiation posed by IL-2 that a variety of mechanisms exists to subvert this inhibition. Transforming growth factor-β (TGFβ) insulates T cells from IL-2 signals by suppressing CD25 (IL-2Rα) expression80, which may contribute to its ability to promote TFH cell differentiation81. Activated DCs in the outer T cell zone upregulate CD25, allowing developing TFH cells to approach the B cell follicle in an environment quenched of local IL-2 (ref.82). Mature TFH cells can be shielded from IL-2 in a different manner, becoming desensitized to IL-2 signalling via IL-6-mediated downregulation of the IL-2 receptor subunit CD122 (IL-2Rβ)83. The local availability of IL-2 and other cytokines is therefore an important factor in controlling TFH cell numbers. Intriguingly, in conditions of limiting IL-2, T helper 1 cells can increase their BCL-6 to T-bet ratio and assume a partial TFH cell phenotype84. Thus, some cells bearing TFH cell markers in autoimmune settings could conceivably have arisen via this route.

The intersection between IL-2-mediated and Treg cell-mediated control of TFH cell differentiation is a complex one. Treg cells might be expected to promote TFH cell differentiation by serving as an IL-2 sink85,86 but they can also limit TFH cell formation by CTLA4-dependent regulation of costimulation29,45,48. One way to view this is that avoiding IL-2 signals may be necessary but not sufficient for TFH cell differentiation and, as highlighted above, this can be achieved through a variety of means. By contrast, there seems to be little redundancy in control of CD28 costimulation, with CTLA4-expressing Treg cells being required to restrict TFH cell numbers45,48. Thus, in the absence of Treg cells, the effects of dysregulated CD28 costimulation may dominate over the lack of IL-2 consumption as other populations can compensate for the latter.

IL-2 and autoimmunity

The IL-2 pathway is strongly implicated in susceptibility to autoimmune disease. Single nucleotide polymorphisms (SNPs) in IL2RA are associated with multiple autoimmune diseases, including T1D, RA, IBD, MS, Crohn’s disease, alopecia areata and vitiligo (see GWAS catalogue). In addition to Il2RA SNPs87, autoimmune-associated variants have been identified in IL2, IL2RB and PTPN2. PTPN2 is a phosphatase involved in many signalling pathways, including IL-2 receptor signalling. T cell-specific deficiency in Ptpn2 increased frequencies of TFH cells and GC B cells in non-obese diabetic mice and was associated with exacerbated diabetes88. It is likely that multiple pathways control IL-2 receptor signalling as defects are evident in T1D and MS even after controlling for IL2RA and PTPN2 genotypes89. Mice deficient in IL-2 or the IL-2 receptor subunits CD122 (IL-2Rβ) or CD25 (IL-2Rα) develop lethal autoimmunity90,91,92 and mutations in IL-2Rβ cause life-threatening immune dysregulation in humans93. Importantly, although IL-2 pathway genes can clearly modulate autoimmunity via effects on Treg cell homeostasis, they can also act in conventional T cells: for example, a SNP at the Il2RA locus associated with protection from T1D and MS was linked to higher levels of CD25 expression on conventional memory CD4+ T cells94. Consistent with the role of the IL-2 pathway in the regulation of autoimmunity, low-dose IL-2 can be used therapeutically across a wide range of autoimmune diseases95.

Collectively, these studies highlight the connection between control of TFH cell differentiation and the genetic regulation of autoimmunity. Notably, genes in the CTLA4 and IL-2 pathways are consistently highlighted in genetic analyses across a broad range of common autoimmune conditions96,97, and SNPs associated with autoimmunity are enriched in CpG demethylated regions specifically found in Treg cells62. This places CTLA4, IL-2 and Treg cells at the heart of the shared genetic susceptibility to autoimmune disease that underpins heritability. Importantly, any defects in the CTLA4 or IL-2 pathways or deficits in Treg cell homeostasis or function would lead to a dysregulated TFH cell compartment as well as to the propensity to autoimmune disease (Fig. 3). This could be one reason for increases in cTFH cells being seen across multiple autoimmune disease settings.

Fig. 3: Common pathways control autoimmune susceptibility and TFH cell homeostasis.
figure 3

In healthy individuals, follicular helper T (TFH) cell homeostasis is maintained by the CTLA4 pathway and regulatory T (Treg) cells as well as by IL-2. Defects in the IL-2–IL-2R pathway can increase TFH cell numbers, and single nucleotide polymorphisms (SNPs) in the IL-2 pathway are associated with autoimmune disease. Altered CTLA4 expression or trafficking (for example, owing to LRBA mutations) or defects in Treg cells can increase TFH cell numbers and are also associated with autoimmune disease susceptibility. CTLA4, cytotoxic T lymphocyte antigen 4; CXCR5, CXC-chemokine receptor 5; ICOS, inducible T cell costimulatory; PD1, programmed cell death protein 1; TCR, T cell receptor.

Infection as a link between TFH cell differentiation and autoimmunity

The principal biological role for TFH cells lies in protection from infectious disease. TFH cell differentiation in response to viral, bacterial, parasitic or fungal antigens is key for the generation of protective antibody responses, in particular affinity-matured neutralizing antibodies98. Mimicking infection by vaccination also induces TFH cells, with transient increases in cTFH cells being detectable in the blood99.

Notably, infectious triggers have been postulated for many autoimmune diseases100,101,102,103,104. Lyme disease is the quintessential example whereby immune responses to tick-borne Borrelia burgdorferi can give rise to Lyme arthritis with autoimmune T cell and B cell responses105. In most of the common autoimmune diseases, a clear link to an individual pathogen is lacking; however, circumstantial evidence is often strong. For example, there is a known association between enteroviral infections and T1D106, and enteroviral capsid protein107 and an antiviral signature108 have been detected in the pancreatic islets of people with T1D. Furthermore, recent data show a clear link between Epstein–Barr virus infection and the development of MS109,110.

Persistent TFH cells in chronic infection: helpers for autoreactive B cells

Although TFH cell increases are short-lived in acute infection, chronic viral infections elicit persistent TFH cell responses111,112. Such increases in TFH cells could interfere with competitive selection within the GC, potentially allowing the survival of self-reactive B cells. Indeed, persistent viral infection can elicit polyclonal B cell activation and production of autoantibodies113. In light of this, it is interesting that prolonged enteroviral infection, rather than multiple short-duration infections, is associated with islet autoimmunity in T1D114.

The ability of self-reactive B cells to take up viral antigens, via pinocytosis, Fc receptors or complement receptors, may allow them to present viral peptides and solicit help from virus-specific TFH cells113. In experimental systems, B cells specific for the central nervous system self-antigen myelin oligodendrocyte glycoprotein (MOG) can co-capture influenza virus haemagglutinin and MOG from cell membranes and obtain help from haemagglutinin-specific T cells to produce anti-MOG antibodies115. Interestingly, the ability of TFH cells to provide help to bystander B cells and elicit autoantibody production appears to be enhanced by lymphopenia20, a state long associated with autoimmunity. Conceivably, lymphopenia in the context of SARS-CoV-2 infection could contribute to the generation of autoantibodies documented in patients with COVID-19 (ref.116).

Potential for infection to unleash self-reactive TFH cells

Chronic infections may alter T cell activation thresholds via the upregulation of costimulatory ligands in response to Toll-like receptor engagement or pro-inflammatory cytokines. This costimulatory ligand upregulation may outpace CTLA4-dependent ligand downregulation, permitting the activation of self-reactive T cells normally censored by insufficient CD28 engagement117. Prolonged increases in TFH cells could exacerbate this by increasing levels of IL-21, the archetypal TFH cell-associated cytokine, which can counteract Treg cell suppression118,119 and may reinforce TFH cell differentiation120. IL-21 is overexpressed in several autoimmune diseases, including SLE, RA, SS and T1D121, and although it can also derive from other cells, significant correlations between IL-21 production and TFH cell frequencies have been noted122,123. Thus, chronic infectious settings associated with elevated IL-21, blunted Treg cell function and increased costimulation could permit self-reactive T cells to acquire a TFH cell phenotype.

Infection may play a role in epitope spreading, which is a feature of many autoimmune diseases and is frequently associated with disease progression. One mechanism that may contribute to this phenomenon is the invasion and reuse of existing GCs by B cells bearing a different specificity, particularly in the context of shared T cell help (for example, when B cells specific for distinct regions of a protein interact with the same TFH cell owing to presentation of a common peptide)124. Cross-reactive recognition of bacterial antigens by self-reactive T cells could conceivably form the basis of shared T cell help in autoimmune settings125. Notably, the presence of adjuvants, including the bacterial molecule lipopolysaccharide, can enhance GC invasion and reuse124, suggesting that, in addition to unleashing new cohorts of TFH cells, infection may also enhance the sharing of T cell help between B cells of different specificities.

The type 1 interferon connection

Infection and autoimmunity may also intersect mechanistically at involvement of the type 1 interferon pathway. Type 1 interferons provide the key ‘early warning’ signal of viral infection and play complex roles in protection or pathology following infection with viruses, bacteria, parasites and fungi126. The type 1 interferon pathway has been widely associated with autoimmune diseases, most notably in the case of SLE but also in T1D, MS, RA and others127,128,129,130. There are conflicting reports on the impact of type 1 interferons on TFH cell differentiation. There is evidence that type 1 interferons promote acquisition of at least some aspects of the TFH cell phenotype131, although other data suggest that they support T helper 1 cells at the expense of TFH cells132. Studies focusing on DCs as the target cells for type 1 interferons point to a positive role in TFH cell differentiation following immunization or vaccination133,134 but a negative role in the setting of Plasmodium infection135. As might be expected for the type 1 interferon pathway, timing, location and context are likely to be key in determining outcome136.

Future directions

Different autoimmune diseases are highly distinct in their presentation, with unique tissue-specific features: for example, bone erosion in RA, renal pathology in SLE and metabolic abnormalities in T1D. Given the striking differences in disease processes and clinical presentation, it is remarkable that increases in cTFH cells appear to be a unifying theme across a large number of autoimmune diseases. In this article, I highlight that CTLA4, IL-2 and Treg cells are key modulators of TFH cell differentiation and are also major players in the regulation of autoimmunity. This fundamental connection may go some way to explaining why dysregulated TFH cell homeostasis is a feature of so many different autoimmune diseases. Reinforcing this connection, several other TFH cell-associated genes (including CXCR5, CCR7, ICOSL, PD1, IL4R, IL21R and CD40) are linked to autoimmune disease susceptibility. The notion that genetic predisposition to autoimmunity may be coupled to genetic predisposition to TFH cell formation provides a new perspective on the cTFH cell changes widely reported in autoimmune settings.

As infection can lead to TFH cell differentiation, it is conceivable that infectious triggers occurring on autoimmune-susceptible backgrounds may cause discernible changes to TFH cell populations that precede autoimmunity. Thus, TFH cells may represent the consummate biomarker in autoimmunity, neatly integrating genetic and environmental risk. Accordingly, TFH cell profiling may be increasingly important in the prediction of autoimmune disease development and in monitoring and predicting clinical response to immunotherapies38.

The precise contribution of TFH cells, and their circulating counterparts, to pathology likely differs between autoimmune diseases. Dysregulated TFH cell homeostasis may allow the production of autoantibodies that mediate diverse effector functions — for example, enhancing cross-presentation of islet antigens in T1D137, driving renal pathology in SLE138, and promoting osteoclastogenesis and bone loss in RA139,140. As well as acting on B cells, the capacity for TFH cells to produce IL-21 and CXCL13 (ref.8) may influence CD8+ T cell activation and ectopic lymphoid structure formation, respectively, whereas dysregulated IFNγ production by TFH cells141 has been shown to drive autoimmune GCs and systemic autoimmune disease142.

Altogether, closer study of TFH cell populations in blood, lymphoid organs and tissues promises to yield new insight into autoimmune pathogenesis as well as allowing us to harness the full biomarker potential of this population.