Pathophysiology of T follicular helper cells in humans and mice

Abstract

Follicular helper T cells (T_FH cells) compose a heterogeneous subset of CD4⁺ T cells that induce the differentiation of B cells into plasma cells and memory cells. They are found within and in proximity to germinal centers in secondary lymphoid organs, and their memory compartment also circulates in the blood. Our knowledge on the biology of T_FH cells has increased significantly during the past decade, largely as a result of mouse studies. However, recent studies on human T_FH cells isolated from lymphoid organ and blood samples and recent observations on the developmental mechanism of human T_FH cells have revealed both similarities and differences between human and mouse T_FH cells. Here we present the similarities and differences between mouse and human lymphoid organ–resident T_FH cells and discuss the role of T_FH cells in response to vaccines and in disease pathogenesis.

A number of seminal discoveries made in mice and humans led to the description of B follicular helper T (T_FH) cells in the early 2000s. The requirement of T cell help for the development of antibody responses was first described in the 1960s (ref. 1). CD4⁺ helper T cells (T_H cells) were then found to be necessary for the development of germinal centers, discrete structures in secondary lymphoid organs where the selection of high-affinity B cells and the development of B cell memory occur^2–4. In vitro studies in the 1980s, mostly involving CD4⁺ T cell clones and recombinant cytokines, showed that T_H2 cells are the major T_H subset engaged in helping B cells by secreting interleukin 4 (IL-4) and IL-10 (refs. 5,6). In mouse, T_H1 cells also contribute to the regulation of antibody responses by inducing B cell class switching toward IgG2a. However, for almost two decades it was unclear how the T_H1 and T_H2 cells engaged in B cell help in lymphoid organs were biologically and developmentally distinct from those that exit lymphoid organs and migrate into peripheral tissues. The chemokine receptor CXCR5 was discovered in 1993 as a G protein–coupled receptor expressed primarily by B cells⁷, and in 1996 it was shown to be critical for the migration of B cells into follicles in lymphoid organs in mice⁸. In 1999, CD4⁺ T cells activated in lymphoid organs of immunized mice were found to express CXCR5, which was required for the cells’ migration into follicles⁹. In the early 2000s, studies on CD4⁺ T cells in human tonsils showed that cells expressing CXCR5 have a superior capacity to induce immunoglobulin production in B cells in vitro relative to CD4⁺ T cells lacking CXCR5 expression. On the basis of their localization and functions, tonsillar CXCR5⁺ CD4⁺ T cells were designated as T_FH cells^10–12. A similar CD4⁺ T cell subset was found in mouse lymph nodes¹³. Profiling of cytokine production and gene expression in human and mouse T_FH cells showed that these cells are distinct from T_H1 and T_H2 cells^14–16 and help B cells mainly by delivering activating signals with the TNF family molecule CD40L and the cytokine IL-21 (refs. 14,17–20). In 2009, the transcription repressor Bcl-6 was discovered to be an essential factor for T_FH cell generation in vivo in mice^21–23, and since then T_FH cells have been recognized as an independent T_H subset distinct from T_H1, T_H2 and T_H17 cells.

Our knowledge of the biology of T_FH cells has increased significantly during the past decade (reviewed in refs. 24,25). Like in other fields of immunology, important biological features of T_FH cells have been learned of from studies in mouse models, whereas studies of the ontogeny and function of T_FH cells in humans have remained relatively limited, mainly because of difficulties in investigating and manipulating T_FH cells from human secondary lymphoid organs. Furthermore, there are only two main sources of human T_FH cells for research: tonsils from children who have experienced recurrent throat infections but are otherwise healthy, and spleens, generally from cadaveric organ donors. This poses a challenge in investigations of human T_FH cells’ association with human diseases such as cancer and autoimmunity. Over 60 million years of independent evolution have introduced significant differences in the immune systems of humans and mice. Thus, it is important to address whether conclusions drawn in mouse T_FH studies also hold true for human T_FH cells. Recent progress in our understanding of the biology of blood-circulating T_FH cells in humans has provided clues on how to determine whether alteration of T_FH responses contributes to human diseases. Furthermore, analyses of blood memory T_FH cells (and also lymph node cells in some instances) from patients with primary or acquired immunodeficiencies have also provided important insights regarding the development and/or maintenance of T_FH cells in humans. Together with in vitro studies aiming at determining the developmental mechanisms of human T_FH cells, these studies have started revealing similarities as well as differences between humans and mice.

In this review, we first summarize our current knowledge of the biology of lymphoid organ T_FH cells and discuss the similarities and differences between these cells in mice and humans. Then we discuss recent findings on the circulating memory compartment of T_FH cells in human blood (hereinafter called blood memory T_FH cells). Last, we summarize recent insights into the role of T_FH cells in disease pathogenesis and discuss how T_FH cells participate in or contribute to both beneficial and aberrant immune responses observed in various human diseases.

T_FH subsets and dynamics in mice

Recent studies in mice and humans show that T_FH lineage cells in lymphoid organs are composed of subsets that differ in their localization, phenotype and function. The circulating memory compartment of T_FH cells in human blood also contains subsets that differ in phenotype and function.

Studies in mice have shown that after interaction with dendritic cells (DCs) in the T cell zones of secondary lymphoid organs, a fraction of activated CD4⁺ T cells migrate toward B cell follicles by upregulating the chemokine receptor CXCR5 (refs. 9,26) (mediated by increased expression of the transcription factors Bcl-6 (refs. 27,28) and Ascl2 (ref. 29)) while downregulating the chemokine receptor CCR7 (ref. 30) and the cell adhesion molecule PSGL-1 (ref. 31). These CXCR5⁺Bcl-6⁺CD4⁺ T cells (hereinafter called T_FH precursors) interact with antigen-presenting B cells at the border of the B cell follicle and T cell zone^9,32, a required process for the generation of germinal center (GC)-resident T_FH cells (GC T_FH cells) and the differentiation of primed B cells along both GC and extrafollicular pathways³³. Bcl-6 expression in CD4⁺ T cells is a prerequisite for GC formation^21–23. The origin of T_FH cells is not restricted to naive cells, and there is some evidence suggesting that other T_H subsets including T_H1, T_H2, T_H17 and regulatory T (T_reg) cells may become T_FH cells in GCs (reviewed in ref. 25). This is consistent with the heterogeneity in cytokine expression patterns among GC T_FH cells developed under different immunization protocols and by different types of infectious agents (reviewed in ref. 34).

In mice, GC T_FH cells and T_FH precursors have been defined largely on the basis of differences in cell surface markers, particularly the expression of CXCR5 and PD-1, which are more highly expressed in GC T_FH cells than in their precursors outside GCs. During the acute or early phase following immunization, IL-21 is expressed almost exclusively by CXCR5^hiPD-1^hi GC T_FH cells in lymphoid organs³⁵. Investigation of T_FH cells is, however, no longer restricted to their identification on the basis of phenotypic markers and cytokine secretion; it is now possible to visualize and track GC T_FH cells directly with intravital imaging. Such imaging studies have revealed that GC T_FH cells display unique cell dynamics: GC T_FH cells continuously immigrate and redistribute to other follicles and neighboring GCs³⁶, in a manner dependent on high expression of CXCR5 and the trafficking molecule sphingosine-1-phosphate receptor 2 (S1PR2)³⁷. Signals derived from interacting B cells induce GC T_FH cells to increase intracellular Ca²⁺ and their expression of IL-4 and IL-21, which enhance their capacity to promote the growth and differentiation of B cells³⁸. A dynamic exchange of GC T_FH cells among multiple GCs and bidirectional signals between GC T_FH cells and B cells during their cognate interactions likely represent important mechanisms associated with the efficient selection and expansion of high-affinity B cells during GC responses. GC T_FH cells exiting GCs rarely remain in the T cell zone or enter the bloodstream (at least while GC responses are actively in progress)³⁶, suggesting that GC T_FH cells and their precursors display different cell dynamics.

Mouse studies also identified cell subsets important for the suppression of GC response³⁴. These suppressors contain Foxp3⁺ T follicular regulatory (T_FR) cells^39–41. Current evidence shows that, at least in mice, T_FR cells are differentiated from thymus-derived Foxp3⁺ regulatory T cells. T_FR cells dampen GC responses by limiting the numbers of both T_FH and B cells in GCs^39–41. Given that mice lacking functional T_FR cells favor the accumulation of non-antigen-specific B cells⁴⁰, T_FR cells might be specialized in repressing self-reactive B cells in GCs. T_FR cells are also likely to be responsible for terminating the GC response³⁹. In either case, the balance between T_FH and T_FR cells in the GC environment likely represents a key factor in the generation of both high-affinity protective antibodies and pathogenic autoantibodies.

T_FH subsets in human tonsils

Studies in the early 1980s showed that CD4⁺ T cells in human tonsillar GCs express CD57 (ref. 42) and that these CD57⁺ T cells display a limited ability to express IL-2 (ref. 43). Although early studies suggested that CD57 is a marker for functionally mature GC T_FH cells in humans, CD57 is expressed by only approximately 30% of GC T_FH cells, and it is also expressed by a fraction (~10%) of CD4⁺ T cells localized outside GCs^15,18. T_FH cells in human GCs are currently defined by their high expression of CXCR5, inducible costimulator (ICOS) and T cell inhibitory receptor PD-1 (refs. 15,18,44). Of note, ICOS is not a useful marker for defining GC T_FH cells in mice, as ICOS expression is largely similar between GC T_FH cells and T_FH precursors³⁶. In human tonsils, Foxp3⁺ T_FR cells within GCs are much rarer than in mice (unpublished observations), and current knowledge on human T_FR cells is very limited.

While Bcl-6 is well recognized as a transcription factor defining the T_FH lineage, mouse studies show that GC T_FH cells contain cells coexpressing Bcl-6 and T-bet (the transcription factor typically expressed by T_H1 cells)^35,45. In humans, GC T_FH cells in tonsils also contain a subset coexpressing Bcl-6 and RORγt (the transcription factor typically expressed by T_H17 cells), in addition to a subset coexpressing Bcl-6 and T-bet⁴⁶. It remains to be established whether these GC T_FH cell subsets and GC T_FH cells lacking expression of T-bet or RORγt have distinct functions; nevertheless, this observation suggests that other T_H subsets might be able to differentiate into T_FH cells in humans, similar to cells in mice. An alternative explanation is that there are developmental paths that are shared between T_H1 and T_FH cells and between T_H17 and T_FH cells in humans, as also proposed in mice^47,48.

Recent data point to the possibility that the development of T_FH cells might differ between mice and humans. In humans, the cytokine TGF-β acts with IL-12 and IL-23 to promote the expression of multiple T_FH molecules, including CXCR5, IL-21 and Bcl-6 (ref. 46), on activated naive CD4⁺ T cells (Fig. 1). Furthermore, human CD4⁺ T cells cultured under conditions used to generate T_H17 cells (for example, a combination of TGF-β, IL-23, IL-6 and IL-1β) coexpress T_FH molecules and T_H17 molecules⁴⁶. This is in stark contrast to mouse CD4⁺ T cells, in which TGF-β signals suppress the expression of T_FH molecules including IL-21, ICOS and Bcl-6 (refs. 21,46,49,50).

Figure 1.

Potential mechanism in the generation of human T_FH subsets. As with other T_H subsets, signals derived from antigen-presenting cells (including DCs) and the microenvironment instruct naive CD4⁺ T cells to differentiate into the T_FH lineage. The major cytokines driving the early T_FH differentiation process in humans are IL-12 and IL-23, and TGF-β synergizes with these cytokines. Other STAT3-activating cytokines, including IL-6, IL-21 and IL-1β, also support this process in the presence of IL-12, IL-23 and TGF-β. The differentiation of human naive CD4⁺ T cells is regulated by the balance of signals derived from these cytokines, and activated CD4⁺ T cells differentiate into precursors of variable T_H subsets such as T_H1, T_H17 and T_FH cells. Some T_FH precursors share properties of T_H1 and T_H17 cells (dotted rectangle in middle panel); interactions with B cells promote their differentiation into mature T_FH cells, including T_FH1 and T_FH17 cells. The mechanism associated with the generation of T_FH2 cells is currently unknown.

In human tonsils, CD4⁺ T cells expressing low amounts of CXCR5 and ICOS (CXCR5^loICOS^lo cells) are exclusively localized outside GCs¹⁸. CXCR5^loICOS^loCD4⁺ T cells in tonsils appear to be T_FH precursors (or extrafollicular helper T cells), as these cells express multiple T_FH molecules such as CD40L, IL-21 and CXCL13 but lack the expression of Bcl-6 protein¹⁸. Functionally, isolated CXCR5^loICOS^loCD4⁺ T cells are more effective than T_FH cells in vitro at helping naive B cells to become immunoglobulin-producing cells, possibly because they produce large amounts of IL-21. In contrast, GC T_FH cells but not CXCR5^loICOS^loCD4⁺ T cells provide help to GC B cells and promote their survival, proliferation and differentiation into immunoglobulin-producing cells in vitro. The inability of CXCR5^loICOS^loCD4⁺ T cells to help GC B cells is due to their expression of Fas ligand, which induces the death of Fas-expressing GC B cells¹⁸. Thus, T_FH cells at different maturation stages differ in their location and biological functions in human tonsils¹⁸. The observation that human T_FH precursors help naive B cells is consistent with the presence of mouse T_FH precursors with low expression of PD-1 and Bcl-6 at the T-B border shortly after T cell priming and before the induction of GCs³³. These T cells likely induce the differentiation of B cells that have recently engaged their antigen receptors.

Memory T_FH subsets in human blood

CD4⁺ T cells expressing the chemokine receptor CXCR5 were first described in human blood in 1994 (ref. 51) and were considered to represent recently activated T cells^10,12,52. More recent studies indicate that blood CXCR5⁺CD4⁺ T cells contain long-lived memory cells that share functional properties with T_FH cells⁵³. Accordingly, blood CXCR5⁺CD4⁺ T cells are currently termed blood (or peripheral) memory T_FH cells. Unlike GC T_FH cells, blood memory T_FH cells—even those expressing ICOS and the proliferation marker Ki67—do not express the Bcl-6 protein^54–58, indicating that Bcl-6 is dispensable for their maintenance. The molecular mechanisms by which blood memory T_FH cells maintain their T_FH characteristics remain largely unknown and thus represent an important research topic.

Human blood memory T_FH cells actually include several populations with unique phenotypes and functions (see ref. 53 for a detailed review). Although staining strategies and markers selected to define blood T_FH subsets differ among laboratories, we propose that blood memory T_FH cells can be assessed on the basis of the following three sets of parameters: the presence of the chemokine receptors CXCR3 and CCR6; of the immunoregulatory molecule PD-1 and the chemokine receptor CCR7; and of the costimulatory molecule ICOS (Fig. 2a).

Figure 2.

Alteration of blood memory T_FH subsets in human autoimmune diseases. (a) Three parameters—CXCR3 versus CCR6, PD-1 and CCR7, and ICOS—can subdivide human blood memory T_FH cells (CD4⁺CD45RA⁻CXCR5⁺) into at least three T_FH1, T_FH2 and T_FH17 subsets (nine T_FH subsets in total). PD-1 and CCR7 define two quiescent subpopulations, PD-1⁻CCR7^hi and PD-1⁺CCR7^int, and ICOS defines the ICOS⁺PD-1²⁺ activated population within the blood memory T_FH1, T_FH2 and T_FH17 subsets. The nine blood memory T_FH subsets are indicated in a three-dimensional scale. (b) T_FH2 and T_FH17 subsets represent efficient helpers among blood memory T_FH subsets. In the blood of patients with active autoimmune disease (such as SLE, juvenile dermatomyositis, Sjögren’s syndrome or multiple sclerosis), the frequency of active (ICOS⁺PD-1²⁺) T_FH2 and T_FH17 subsets increases, whereas the frequency of T_FH1 subsets decreases. This alteration likely reflects the increase of functional T_FH cells in lymphoid organs and/or inflamed tissues.

The first set of parameters defines three major subsets: CXCR3⁺CCR6⁻ cells that share properties with T_H1 cells (hereinafter called blood memory T_FH1 cells), CXCR3⁻CCR6⁻ cells resembling T_H2 cells (hereinafter called blood memory T_FH2 cells) and CXCR3⁻CCR6⁺ cells resembling T_H17 cells (hereinafter called blood memory T_FH17 cells)⁵⁵. Experiments in which isolated T cell subsets were cultured with B cells in the presence of superantigen (such as staphylococcus enterotoxin B to induce cognate interactions between T and B cells) indicated that blood memory T_FH2 and T_FH17 cells can induce naive B cells to produce immunoglobulins and to switch isotypes through IL-21 secretion. In contrast, blood memory T_FH1 cells lack the capacity to help naive B cells^55,59,60. Furthermore, whereas blood memory T_FH2 cells promote IgG and IgE secretion, blood memory T_FH17 cells induce IgG and IgA secretion⁵⁵. Thus, T_FH2 and T_FH17 cells represent efficient B cell helper cells with a distinct capacity to regulate immunoglobulin isotype switching.

The second (PD-1 and CCR7) and third (ICOS) sets of parameters define three other subsets: one activated subset (ICOS⁺PD-1²⁺CCR7^lo) and two quiescent subsets (ICOS⁻PD-1⁺CCR7^int and ICOS⁻PD-1⁻CCR7^hi). In healthy subjects ICOS expression is limited to a small population of blood T_FH cells; this population substantially increases after vaccination, and thus defines activated memory T_FH cells. In mice these cells appear before the formation of GC T_FH cells, because deletion of the Sh2d1a gene, encoding the SAP, an adaptor protein essential for the functions of SLAM family receptors, does not affect the formation of this circulating subset but prevents the appearance of terminally differentiated GC T_FH cells⁵⁸. ICOS⁻ cells are further divided into PD-1⁺ cells (~30% of blood memory T_FH cells) and PD-1⁻ cells, both of which lack the expression of Ki67 and thus are in a quiescent state^58,59. CCR7 expression by blood memory T_FH cells negatively correlates with that of PD-1 (refs. 52,53).

The matrix combination of the three parameters defines nine blood memory T_FH subsets (Fig. 2a). The activated ICOS⁺PD-1²⁺CCR7^lo populations within blood memory T_FH2 and T_FH17 cells might represent the most efficient helpers. Among quiescent T_FH2 and T_FH17 subsets, the ICOS⁻PD-1⁺CCR7^int population provides help to memory B cells more promptly than the ICOS⁻PD-1⁻CCR7^hi population^59,60, but the ICOS⁻PD-1⁺CCR7^int and ICOS⁻PD-1⁻CCR7^hi T_FH2 and T_FH17 subsets are equally capable of helping naive B cells.

In contrast, blood memory T_FH1 cells lack the capacity to help naive or memory B cells^54,55,59,60. Therefore, whether blood memory T_FH1 cells represent a subset of memory T_FH cells remains controversial. First, upon T cell receptor stimulation, blood memory T_FH1 cells produce CXCL13, a chemokine that is highly expressed by GC T_FH cells and whose receptor, CXCR5, is expressed at high levels by B cells^18,61. Second, blood memory T_FH1 cells contain antigen-specific memory cells⁵⁹. Third, upon polyclonal stimulation with the phorbol ester PMA and ionomycin, blood memory T_FH1 cells produce IL-21 in amounts equivalent to those of blood memory T_FH2 and T_FH17 cells⁵⁹. Finally, influenza vaccination transiently induces ICOS expression exclusively on blood memory T_FH1 cells⁵⁴. The increase in ICOS⁺ T_FH1 cells in blood is positively correlated with the generation of protective antibody responses⁵⁴. Mechanistically, ICOS⁺ blood memory T_FH1 cells help memory, but not naive, B cell differentiation into plasma cells via secretion of IL-21 and IL-10⁵⁴. This observation suggests that blood memory T_FH1 cells also contribute to antibody responses, but only when they become ICOS⁺PD-1²⁺CCR7^lo activated cells. Collectively, blood memory T_FH cells can be subdivided into nonefficient helpers (T_FH1) and efficient helpers (T_FH2 and T_FH17). The differential expression of ICOS, PD-1 and CCR7 further defines developmentally and functionally distinct subpopulations within the subsets.

T_FH cells in primary immunodeficiencies

Primary immunodeficiencies are caused by defects in the expression of molecules involved in immune system development and/or function. Analyses of lymphoid tissue and blood samples from patients with primary immunodeficiencies have helped clarify the requirements for the development of normal immune responses in humans, including the generation of T_FH cells and GCs. With the success of whole exome sequencing in identifying gene variants responsible for primary immunodeficiencies (at least 34 new gene defects were identified in the past 4 years⁶²), investigation of immune responses in affected patients is likely to accelerate understanding of human immunity in the near future.

Patients with ICOS deficiency suffer from common variable immunodeficiency and show severely impaired GC formation in lymphoid tissues and severely decreased blood memory T_FH cells, accompanied by a severe deficiency of memory B cells⁶³. This demonstrates that ICOS is essential for the generation of T_FH and GC responses in humans, as shown in mice⁶⁴.

Multiple mouse models show that B cells play a fundamental role in the generation of GC T_FH cells by interacting with T_FH precursors²⁵. This is also the case in humans, as patients with X-linked agammaglobulinemia, who lack mature B cells because of a deficiency of the tyrosine kinase Btk, and those with CVID, who have a significantly reduced number of B cells (<2% in lymphocytes), display significantly reduced frequencies of blood memory T_FH cells⁶⁵. Patients with hyper IgM syndrome caused by a deficiency of functional CD40L also show severely impaired GC formation, as well as reduced blood memory T_FH cells⁶³ and memory B cells⁶⁶, which indicates the importance of CD40-CD40L interactions for GC and T_FH development in humans, as shown in mice⁶⁷.

SAP-deficient mice show profoundly altered GC reactions and generation of T_FH cells⁶⁸, as SAP is required for durable interactions between helper T cells and B cells, a process essential for the maturation of T_FH cells⁶⁹. Consistently, patients with X-linked lymphoproliferative disease caused by SAP deficiency show impaired formation of GCs and impaired development of memory B cells⁷⁰. Interestingly, both SAP-deficient mice and human patients with X-linked lymphoproliferative disease show normal frequencies of blood memory T_FH cells⁵⁸, suggesting that at least a subset of circulating blood T_FH cells develop before GC T_FH formation. Furthermore, migration of GC T_FH cells into the circulation is limited during the acute phase of the GC response^36,38. It remains possible that GC T_FH cells and their descendent memory T_FH cells in lymphoid organs^39,40 might start circulating during the termination of the GC response.

In vitro studies have suggested that IL-12 promotes antibody responses in humans via two mechanisms. IL-12 produced by DCs (CD14⁺ dermal DCs in human skin⁷¹) can induce the in vitro differentiation of CD40-activated naive B cells into IgM-producing plasma cells^72,73 and of naive CD4⁺ T cells into T_FH-like cells^74,75. Consistently, children deficient for the IL-12 receptor β1 (IL-12Rβ1) chain, which is shared by the IL-12 and IL-23 receptors, have reduced T_FH and GC responses⁷⁴. Thus, in addition to the recent in vitro observation that TGF-β acts as a critical co-factor of IL-12 and IL-23 for the early T_FH cell differentiation process in humans⁴⁶, impaired T_FH response in IL-12Rβ1 deficiency in vivo provides evidence that IL-12 and IL-23 contribute to the development of T_FH responses in humans. Although the frequency of blood memory T_FH cells is significantly reduced in subjects with IL-12Rβ1 deficiency during childhood (less than 10 years old), it gradually increases with age and becomes normal in adults. This indicates that the development and/or maintenance of the T_FH cell response can be compensated through other pathways and/or cytokines, possibly IL-6, IL-21 and IL-27.

Mutations in the intracellular signaling molecule STAT3 cause most cases of hyper IgE syndrome. People with this syndrome are susceptible to a narrow spectrum of infections linked to defective T_H17 responses and show altered antibody responses with reduced blood memory T_FH cells^76,77, indicating that STAT3 is required for optimal T_FH cell generation in humans. Given that B cell responses to IL-10 and IL-21, cytokines delivering activation signals via STAT3, are severely altered in STAT3-deficient subjects⁷⁸, the reduced number of blood memory T_FH cells in STAT3-mutant subjects might be dependent on both B cell–intrinsic and T cell–intrinsic defects. Notably, upon exposure to IL-12, human STAT3 mutant naive CD4⁺ T cells are induced to express high amounts of several T_FH molecules, such as CXCR5, ICOS and Bcl-6, despite lower expression of IL-21 (ref. 76). Thus, other molecules such as STAT4 might functionally compensate for the lack of STAT3. This concept is supported by the finding that STAT3 and STAT4 play largely redundant roles in the expression of T_FH molecules by human naive CD4⁺ T cells exposed to TGF-β⁴⁶.

T_FH cells and autoimmunity

The generation of autoantibodies is a hallmark of autoimmune disease. Autoantibodies target a broad range of self-antigens, including nuclear components (such as double-stranded DNA), organ-specific antigens and soluble factors. These autoantibodies can profoundly dysregulate the function of multiple organs or systems through a variety of mechanisms. Autoantibodies result from a breakdown of tolerance mechanisms during B cell development⁷⁹, from T cell–dependent or –independent B cell activation, or as a consequence of somatic mutation and rogue selection within GCs^80,81. Yet GCs can in some cases redeem self-reactive B cells⁸². There is now compelling evidence in mice and in humans that aberrant generation and/or activation of T_FH cells and extrafollicular helper T cells contributes to the pathogenesis of autoimmune diseases. Here we first describe evidence of aberrant T_FH responses in mouse and human autoimmunity, and then discuss how genetic factors are potentially associated with aberrant T_FH responses in human autoimmune diseases.

T_FH cells in autoimmune mouse models

The first evidence linking aberrant T_FH responses and autoimmunity came from studies of the sanroque mouse model, which carries a single-amino-acid mutation known as ‘san’ in the RNA-binding protein Roquin-1 (ref. 16). Roquin^san/san (or Rc3h1^san/san) mice spontaneously develop lupus-like clinical symptoms and, in a genetically susceptible background, type 1 diabetes (T1D). The development of these diseases is accompanied by high amounts of antinuclear and anti-islet antibodies^16,83. Mechanistically, the mutated Roquin in Roquin^san/san mice displays an impaired ability to repress the expression of Icos¹⁶ and Ifng⁸⁴ in activated helper T cells and promotes the generation of T_FH cells. Ox40 (also known as Tnfrsf4), a co-stimulatory molecule highly expressed by T_FH cells²⁵, and Tnf are additional targets of Roquin and its paralog Roquin-2 (refs. 85–87). T_FH cells and myeloid cells accumulate and are overactive in Roquin^san/san mice and in mice doubly deficient in Roquin and Roquin-2 (refs. 86,87). Genetic ablation of Sh2d1a (encoding SAP) in Roquin^san/san mice prevents lupus pathology, whereas transfer of sanroque T_FH cells into wild-type mice promotes spontaneous GC formation, supporting a causal role for T_FH cell accumulation in the lupus-like disease⁸⁸. These findings suggest that excessive T_FH cells in GCs corrupt positive selection, such as diminished competition for T cell help, and a lower threshold for selection allows the emergence of self-reactive clones. Excessive expression of ICOS⁸⁹ and, more prominently, of IFN-γ⁸⁴ was found to be an important contributor to Roquin^san-mediated T_FH cell accumulation. Of note, IFN-γ signaling blockade has been found not only to alleviate T_FH and GC B cell accumulation but also to virtually prevent all clinical manifestations associated with Roquin^san-mediated disease, unlike SAP deficiency, which did not correct the splenomegaly or hypergammaglobulinemia⁸⁸, or ICOS deficiency, which did not eliminate autoantibody formation⁸⁴. Other mouse models have also revealed additional mechanisms that cause aberrant T_FH responses in autoimmune disease. In the BXSB-Yaa mouse model, which displays a duplication of the Tlr7 gene, aberrant T_FH responses and the development of glomerulonephritis are completely dependent on IL-21 signals⁹⁰, in contrast to what occurs in Roquin^san/san mice, in which IL-21 signals are not essential for pathology⁹¹. The generation of autoantibodies in the MRL/lpr lupus mouse model, which is characterized by deficiency of the proapoptotic molecule Fas, is dependent on the T and B cell interactions at extrafollicular sites³¹. Extrafollicular T_H cells that develop in MRL/lpr mice share features with GC T_FH cells: their development is dependent on ICOS and Bcl-6, and their helper function is dependent on IL-21 and CD40L³¹. These studies demonstrate that inhibiting the generation of pathogenic T_FH and extrafollicular helper T cells by blocking ICOS, CD40L, IFN-γ or IL-21 ameliorates the generation of autoantibodies and/or the development of glomerulonephritis, thereby providing a rationale for targeting these molecules for therapeutics.

Although T1D has long been thought to be a T cell–driven and organ-specific autoimmune disease, there is now evidence for a pathogenic role for both B cells and antibodies, as well as for T_FH cell dysregulation, in T1D etiopathogenesis. B cells are required for diabetes development in the nonobese diabetic (NOD) mouse model of T1D^92–94 and in the Roquin^san/san model⁸³. Direct evidence for a pathogenic role of autoantibodies comes from experiments demonstrating protection from diabetes in NOD mice made deficient for activating Fcγ receptors^95,96 and, conversely, diabetes induction upon passive transfer of antibodies against islet-expressed neo-self-antigens^83,96. In line with these observations, it is not surprising that correlations are being described between T_FH cells and autoimmune diabetes. In the Roquin^san/san model, aberrant T_FH response was directly linked with the development of high anti-islet autoantibody titers and T1D⁸³. T_FH cells were increased in the pancreatic draining lymph nodes of mice that developed autoimmune diabetes⁸³, and cells sharing characteristics of T_FH cells, but expressing CCR9 instead of CXCR5, have been found in the pancreas of diabetic-prone NOD mice⁹⁷.

T_FH cells in human autoimmune diseases

Increased GC response in patients with systemic lupus erythematosus (SLE) was suggested in early studies by an increased frequency of CD27⁺CD38^hi somatically mutated antibody-producing plasmablasts in peripheral blood^98,99. Because SLE patients display higher frequencies of self-reactive mature naive B cells than healthy controls owing to a defect in the early checkpoint of B cell repertoires^100,101, SLE patients seem predisposed to the development of a broader range of autoantibodies than healthy subjects. Whether and how T_FH cells contribute to the pathogenesis of human autoimmunity has been unclear, but recent progress in understanding the biology of blood memory T_FH cells has rendered the analysis of human T_FH responses in the context of autoimmunity feasible.

Since the description of an association between the frequency of CXCR5⁺ICOS⁺ and/or CXCR5⁺PD1⁺ circulating T_FH cells and the severity of SLE and Sjögren’s syndrome⁵⁷, multiple studies have confirmed¹⁰² or refined these findings, providing improved phenotypic characterization (for example, a reproducible association with the CCR7^loPD-1^+/hi subset⁵⁸), and extended them to additional autoimmune diseases such as myasthenia gravis¹⁰³, rheumatoid arthritis^104,105, autoimmune thyroid diseases¹⁰⁶ and T1D¹⁰⁷. An increase of ICOS⁺ blood memory T_FH cells showed a positive correlation with serum autoantibody titers and disease activity and/or severity in these diseases^{57,58,104,106–108}. These observations suggest that patients with active autoimmune disease display aberrant T_FH responses; this can be monitored through assessment of the increase of activated blood memory T_FH cells.

Accumulating evidence also indicates that patients with autoimmune disease display an alteration in the balance of blood memory T_FH1, T_FH2 and T_FH17 cells. In patients with juvenile dermatomyositis⁵⁵, adult SLE¹⁰⁹ and Sjögren’s syndrome¹⁰², T_FH1 cells are underrepresented among blood memory T_FH cells, whereas T_FH2 and/or T_FH17 cells are overrepresented. Such alterations were found to correlate with disease activity, serum autoantibody titers and/or the frequency of blood plasmablasts^55,102,109. Furthermore, although multiple sclerosis (MS) is not generally considered an autoantibody-mediated autoimmune disease, patients with MS also show the same alteration (low T_FH1 and high T_FH17) in the composition of blood memory T_FH subsets¹¹⁰. Two additional lines of evidence also support a pathogenic role of T_FH cells in MS. First, B cells are now thought to play a major pathogenic role in MS, as depletion of B cells with anti-CD20 significantly reduces the number of brain inflammatory lesions and halts the development of new lesions¹¹¹. Second, ectopic B cell follicles are formed in the brain lesions of more than 40% of patients with secondary progressive MS, and the development of these structures correlates with disease severity¹¹².

Collectively, although definitive evidence is yet to be found, these observations suggest that an increase in activated T_FH2 and/or T_FH17 cell subsets and a decrease of T_FH1 cells within blood memory T_FH cells might be common across multiple autoimmune diseases (Fig. 2b). Such alterations in blood memory T_FH cells might reflect an overall increase of efficient helpers that promote the generation of antibodies in lymphoid organs and/or inflammatory sites in patients with autoimmune diseases. However, to date, studies linking human blood memory T_FH cells and T_FH cells in lymphoid organs and/or ectopic GCs in inflammatory tissues are lacking. T_FH cell responses in ectopic GCs might be of considerable importance, as discussed below.

T_FH cells in inflamed tissues

Inflammatory sites in autoimmune diseases often develop lymphoid cell aggregations including helper T cells and B cells, which leads to the formation of well-structured GCs (hereinafter called ectopic GCs). The mechanisms that control the initial development, cellular composition and functional maintenance of ectopic GCs seem to be largely shared with GCs in lymphoid organs¹¹³. For example, in lupus tubulointerstitial nephritis lesions, T_H cells found in lymphoid T and B cell aggregates are phenotypically similar to T_FH cells in lymphoid organs¹¹⁴. Although the precise mechanisms by which T_FH cells accumulate in inflammatory sites in humans remain largely unknown, studies in mouse models suggest the involvement of T_H17 cells¹¹³. In the experimental autoimmune encephalomyelitis mouse model, T_H17 cells induce the formation of ectopic lymphoid follicles in inflammatory brain lesions via IL-17 and the cell surface molecule podoplanin and show features of T_FH cells in these tertiary lymphoid structures¹¹⁵. Given that human patients with autoimmune diseases have increased activated memory T_FH17 cells in blood (as discussed earlier), that early developmental pathways for T_FH and T_H17 cells in inflammatory environments are shared in humans⁴⁶, and that in mice T_H17-derived IL-17 directly promotes B cell differentiation into GC B cells, which are thought to be the source of pathogenic autoantibodies¹¹⁶, it is tempting to speculate that T_FH17 cells might be involved in the formation of ectopic GCs in human autoimmune diseases.

Among patients with autoimmune disease, those with ectopic GCs in inflammatory lesions often have higher disease activity and are refractory to treatment. Yet the formation of ectopic GCs does not simply reflect the extent or the duration of inflammation in the lesions, as only 20% to 40% of patients with chronic inflammation develop ectopic GCs¹¹³. The formation of ectopic GCs is preceded by aggregates of T and B cells in inflammatory sites. Therefore, at variance with B cell follicles in lymphoid tissues, where autoreactive B cells are excluded because of their reduced expression of CXCR5 (ref. 117), lymphoid aggregates in inflammatory tissues are likely to be permissive to the entry of autoreactive B cells. Encounters with T_FH-like cells at these sites¹¹⁴ might induce their cell growth and/or differentiation into antibody-producing cells. This possibility is directly supported by the observation that B cells in lymphoid aggregates and/or ectopic GCs produce autoantibodies¹¹³. Moreover, organ-specific self-antigens are easily accessible by antigen-presenting cells, including B cells, because of the abundance of damaged and/or apoptotic cells caused by the inflammatory process. Therefore, ectopic GCs might represent a major site of autoantibody production in autoimmune diseases, and thus a potentially good target for therapeutics. The nature of T_FH cells in inflamed tissues requires further studies. Questions regarding subjects such as the developmental mechanism of T_FH cells at inflammatory lesions and how lymphoid aggregates develop into ectopic GCs can be addressed in mouse models, but an effort to determine whether the observations apply to humans will be required. Another important question is whether activated blood memory T_FH cells present in active autoimmune disease patients originate from inflammatory lesions. It will be critical to answer these questions in order to directly link the information obtained from the analysis of blood T_FH subsets with autoimmune disease pathogenesis.

Genetics of altered T_FH cell response

Autoimmune diseases result from complex interplay between genetic and environmental factors. Genome-wide association studies (GWAS) on multiple autoimmune diseases have yielded risk-associated loci¹¹⁸ (summary at http://www.genome.gov/gwastudies). Variants in HLA class II alleles are at the top of the list of susceptibility loci in most autoantibody-mediated autoimmune diseases, supporting a requirement of CD4⁺ helper T cells in disease pathogenesis. Importantly, multiple risk-loci identified in autoimmune diseases are potentially associated with the regulation of the development and/or the function of human T_FH cells (Fig. 3). The list includes genes associated with IL-12 and IL-23, as well as cytokines implicated in human T_FH cell differentiation^44,46,74,75, such as IL12A, IL12B (both in multiple sclerosis), IL23R (rheumatoid arthritis), STAT3 (multiple sclerosis), STAT4 (SLE, rheumatoid arthritis, Sjögren’s syndrome), IRF5 (SLE, rheumatoid arthritis) and IRF8 (SLE, rheumatoid arthritis, multiple sclerosis). The transcription factors IRF5 and IRF8 positively regulate the production of IL-12 and IL-23 by macrophages and DCs^119,120. IRF8 is essential for IL-12 production in humans. IRF8-deficient human subjects experience disseminated infection caused by bacille Calmette-Guérin vaccines, a clinical manifestation also seen in IL-12 receptor deficiency, due to a profound reduction of IL-12 production by monocytes¹¹⁹. IRF8 also contributes to signaling in TGF-β, another cytokine important for the generation of human T_FH cells⁴⁶, by promoting the expression of TGF-β-activating integrin α_vβ_⁸ on the surface of DCs¹²¹. Studies in SLE and Sjögren’s syndrome suggest that the OX40 ligand (OX40L)-coding TNFSF4 might be involved in the generation and/or homeostasis of T_FH cells: overexpression of OX40L causes the accumulation of T cells in B cell follicles¹²², and OX40 receptor signals induce CD4⁺ T cells to express CXCR5 in both mice^26,123,124 and humans (unpublished observations). Furthermore, GWAS identified multiple risk-loci encoding molecules expressed by T_FH cells, including IL21 (SLE, rheumatoid arthritis) and CXCR5 (multiple sclerosis, Sjögren’s syndrome). A recent study showed that in addition to Blimp-1-encoding PRDM1 (ref. 22), PTPN22, a risk locus identified in GWAS of SLE, rheumatoid arthritis, myasthenia gravis and T1D, negatively regulates the generation of T_FH and GC responses in mice¹²⁵. However, whether and how most of these gene variants contribute to aberrant T_FH responses in autoimmune diseases remains unknown. The majority of risk loci consist of single-nucleotide polymorphisms within noncoding, putatively regulatory DNA in the proximity of genes¹²⁶, suggesting either that there is a disease-causing coding mutation in linkage disequilibrium with the GWAS-identified single-nucleotide polymorphism, or that alteration in pre- and/or post-transcriptional regulation is central to dysregulated immune responses in subjects with autoimmune disease traits. Recent studies aimed at mapping genetic variation contributing to transcriptional variation (termed expression quantitative trait locus mapping studies) through the use of purified immune cell subsets have started to reveal how gene variants regulate immune responses in different cell types^126–128. The integration of expression quantitative trait locus mapping with blood memory T_FH cells and/or T_FH cells within inflammatory tissues and antigen-presenting cells obtained from patients with autoimmune disease might reveal how gene variants identified in GWAS contribute to the aberrant T_FH and GC responses.

Figure 3.

Risk loci of human autoimmune diseases associated with the T_FH developmental pathway. Multiple risk loci identified in GWAS in autoimmune diseases (indicated in red) are potentially associated with the regulation of the development and/or the function of human T_FH cells. At least seven risk loci—IL12A, IL12B, IL23R, STAT3, STAT4, IRF5 and IRF8—are associated with IL-12 and IL-23. IRF8 also might contribute to TGF-β signaling by promoting the expression of TGF-β-activating integrin α_vβ₈ on the surface of DCs. Risk loci contain genes encoding T_FH-specific molecules (such as IL21 and CXCR5), as well as genes associated with the inhibition of T_FH cell development (such as PRDM1 and PTPN22). Whether and how these gene variants are associated with aberrant T_FH responses in autoimmune diseases remains to be established.

T_FH cells in cancer

Given the demonstrated mutual dependence of T and B cells for growth and survival, it is not surprising that T_FH cells play important roles in supporting the growth and survival of follicular B cell tumors. Furthermore, T_FH cells themselves can give rise to a peripheral T cell tumor known as angioimmunoblastic T cell lymphoma (AITL). Other peripheral T cell tumors also present with phenotypic features and genetic abnormalities that suggest a T_FH origin. Recently, T_FH cells were found to infiltrate solid-organ tumors, where they might play both protective and pathogenic roles.

AITL, which accounts for ~20% of peripheral T cell lymphomas (PTCLs), is an aggressive tumor associated with a poor survival rate (33% 5-year survival)¹²⁹. Most patients present with systemic disease associated with lymphadenopathy, hepatosplenomegaly, anemia and hypergammaglobulinemia and suffer from systemic illness. The neoplastic T cells account for only a small fraction of the lymphoid infiltrate and are mixed with a large number of reactive immune cell types, including small lymphocytes, eosinophils and plasma cells, and expansion of follicular dendritic cell networks. Genetic profiling of AITL and immunohistochemical analysis of neoplastic T cells have provided strong evidence that T_FH cells are the normal cellular counterpart of the neoplastic cells in AITL, which typically express BCL6, CD10 and other T_FH markers, including CXCL13 and PD-1 (ref. 130). In mice, heterozygosity for a Roquin-1 mutation that causes T cell–autonomous T_FH expansion also leads to an AITL-like disease¹³¹. These mutations have not been found in humans, but mutations in TET2, IDH2, DNMT3A and RHOA are commonly found in AITL^132–134. Although the first three genes are known to play a role in DNA methylation and epigenetic modification of gene expression, how these mutations selectively affect T_FH cells remains to be understood.

Up to 40% of other PTCLs (PTCLs not otherwise specified (NOS)) also express T_FH markers, and some share genetic mutations typically found in AITL, including in TET2, IDH2 and RHOA. For this reason, these tumors are now being referred to as T_FH-like PTCL-NOS¹³⁵. The neoplastic T cells of various types of primary cutaneous cell lymphomas also express various T_FH cell markers, with the exception of CD10, and in some cases also form rosettes with B cells, which typically occurs in AITL^136–139.

Evidence for a pathogenic role of T_FH cells in B cell lymphoma first came from studies demonstrating that the number of T cells infiltrating B cell tumors was an important predictor of outcome¹⁴⁰. Follicular lymphoma, the most frequent indolent non-Hodgkin lymphoma, is thought to originate from GC B cells¹⁴¹. The gene expression profiles of tumor-infiltrating T cells and myeloid cells, rather than that of malignant B cells, determine the prognosis for follicular lymphoma¹⁴². Furthermore, that prognosis worsens if T cells are localized within neoplastic follicles¹⁴³. Tumor-infiltrating T cells have the phenotypic features of T_FH cells and overexpress IL-4 (ref. 144), TNF, IFN-γ and LT-α¹⁴⁵. In particular, IL-4 produced by tumor-infiltrating T_FH-like cells causes follicular lymphoma tumor cells to secrete the chemokines CCL17 and CCL22, which attract the migration of T_reg cells and T_H2 cells¹⁴⁵. Foxp3⁺ T cells that resemble T_FR cells described in mice are also present in neoplastic follicles and are expanded during lymphomagenesis¹⁴⁶, although their significance is still unclear. While overall numbers of tumor-infiltrating Foxp3⁺ T cells may be associated with improved survival¹⁴⁷, their follicular localization may be associated with worse survival and increased risk of transformation¹⁴⁸, perhaps through the dampening of the cytotoxic T lymphocyte–driven anti-tumor response¹⁴⁹. Recent reports have described increased circulating T_FH-like cells in patients with chronic lymphocytic leukemia^150,151, particularly during the more advanced stages. These expanded T_FH cells may be of pathological relevance, given that the combination of IL-21 and CD40L induces robust chronic lymphocytic leukemia cell proliferation^150,152.

In nonlymphoid malignancies, T_FH cells are a component of the inflammatory infiltrate of breast cancer, and their presence, as assessed by an eight–T_FH gene signature, was associated with increased survival¹⁵³. Moreover, a single T_FH-associated gene, CXCL13, conferred the dominant prognostic value. Extensively infiltrated tumors were found to contain tertiary lymphoid structures with visible GCs and CXCL13-producing T_FH cells. These findings suggest that T_FH cells, by virtue of their ability to secrete CXCL13 and organize ectopic lymphoid filtrates, may be important in coordinating the recruitment of immune cells that mediate the anti-tumor response.

In other tumor types, a high proportion and altered distribution of T_FH cells has been reported in colorectal cancer¹⁵⁴ and non–small cell lung cancer¹⁵⁵. T_FH cells infiltrate thymomas, correlating with the severity of myasthenia gravis¹⁵⁶. A recent comprehensive gene analysis study on various immune cells isolated from human colorectal tumors revealed that the infiltration of B cells correlates with the infiltration of T_FH cells at tumor sites¹⁵⁷ and that the infiltration of B cells and T_FH cells, which is associated with the expression of CXCL13 and IL21, positively correlates with patient survival. Whether and how T_FH cells might promote the development of colorectal tumors remains to be established.

T_FH cells in HIV and vaccine design

In the past couple of years an intriguing relationship was revealed between T_FH cells and both human and simian immunodeficiency virus (HIV and SIV, respectively). HIV eliminates the circulating memory CD4⁺ T cells it infects, but several groups showed that T_FH cells accumulate in lymph nodes during chronic SIV^158,159 and HIV^160–163 infection. Furthermore, T_FH cells expand, despite old^164–166 and new^158,160 evidence that HIV infects T_FH cells, in which viral replication occurs. Reports from the late 1980s demonstrated that follicular dendritic cell networks in GCs represent a large reservoir of HIV virions¹⁶⁷, retained as immune complexes¹⁶³. Thus, T_FH cells are constantly exposed to the virus during chronic infection. In contrast to the T_FH cell accumulation seen in lymph nodes, blood memory PD-1⁺ T_FH17 cells capable of providing help to B cells are decreased in chronic HIV infection, but they recover after antiretroviral therapy⁶⁰.

The relationships among viral load, chronicity and numbers of T_FH cells are complex. Although most studies have failed to identify a correlation between lymph node T_FH cell numbers and viral load^158,159,161, there are possible explanations for the reduction in T_FH cells after antiretroviral therapy^160,162: the antigen load might control the total number of T_FH cells, the compartmentalization of T_FH cells between secondary lymphoid tissues and the circulation, or the survival of the blood memory T_FH cells. Further work is clearly needed to explain the mechanisms behind these findings.

Chronic HIV infection is associated with an increased frequency of T_FH cells in lymph nodes together with increased expression of Bcl-6 (ref. 158). The expression of IL-6 receptor on T_FH cells is increased, which might mediate increased responsiveness to IL-6 (ref. 158). The importance of IL-6 signals for the maintenance of the T_FH response in chronic infection is also highlighted in a mouse model of chronic viral infection with LCMV clone 13 (ref. 168). Increased T_FH generation during chronic HIV infection may in turn affect the host’s immune response not only against HIV, but also against other, unrelated viral and bacterial infections¹⁶⁹. High T_FH numbers in HIV-infected individuals correlate with B cell dysregulation, including hypergammaglobulinemia, loss of memory B cells and, occasionally, production of autoantibodies and development of autoimmunity¹⁷⁰. It is possible that the lowered threshold of B cell selection in GCs due to an excess of T_FH cells¹⁷¹ might lead to the generation of antibodies with low affinities. Alternatively, T_FH cell function might be dampened by either T cell–intrinsic factors (e.g., increased sensitivity to IL-6 signaling)¹⁵⁸ or the microenvironment (e.g., increased PD-L1 expression by GC B cells that dampens IL-21 production by T_FH cells¹⁶²).

Designing potent HIV vaccines remains a major challenge. It is now established that antibodies can be protective and prevent infection, as passive transfer of broadly neutralizing antibodies confers protection to HIV challenge in humanized mice infected with HIV^172,173. However, most vaccine trials have failed to protect vaccinated individuals to any significant degree. A number of factors associated with HIV make the creation of a decently protective antibody-based vaccine difficult, such as fast mutation rates, structural properties of the envelope complex that make conserved epitopes relatively inaccessible to antibodies, and the need for B cells to undergo extensive somatic mutation to generate broadly neutralizing antibodies. The recent HIV vaccine RV144 “Thai” trial showed that binding of IgG antibodies to envelope proteins is associated with protection, whereas binding of IgA antibodies to envelope proteins correlates directly with the rate of infection¹⁷⁴. These data suggest a need to better understand the nature and isotype of the antibody response to be elicited by vaccines for protection. There is encouraging evidence that numbers of blood ICOS⁻PD1⁺ T_FH2 and T_FH17 cells (but not T_FH1 cells) may serve as a useful biomarker for patients producing broadly neutralizing antibodies to HIV¹⁷⁵. However, the significant overlap in measurements between high- and low-affinity groups calls for additional studies to further delineate the best correlates of protection. Although numerous studies have concluded that a potent T_FH response correlates with antibody titers and protective responses^54,58,59, there is also evidence that an excessive response might be deleterious. Thus, limiting of T_FH cells is required for optimal affinity maturation¹⁷⁶, and an excessive number of T_FH cells lowers the threshold for positive selection, allowing survival of low-affinity and self-reactive clones⁸⁸. Furthermore, the balance between T_FH and T_FR cells is also likely to be important for the overall duration and quality control of GC responses³⁴. So far, very little is known about how different adjuvants and prime-boost regime strategies influence the magnitude, longevity and quality of antibody responses. This knowledge will be important for the improved and rational design of protective vaccines.

Conclusions

In the past 5 years considerable progress has been made in the understanding of T_FH cells, particularly in humans. We predict that this progress will lead to improved vaccine designs, better management of autoimmune diseases and novel prognostic biomarkers for lymphoid and solid tumors.