Here, there and everywhere: T follicular helper cells on the move

Summary

T follicular helper (Tfh) cells have the important function of providing B‐cell help for the induction of antigen‐specific antibody production. As such, it is important to determine the factors that regulate the development, differentiation and function of Tfh cells. This review highlights some of the recent advances in our understanding of Tfh cell migration, Tfh cell memory and the origins and fate of circulating Tfh cells in the blood, that have been revealed from studies in humans and mice.

Introduction

They seek him here, they seek him there

Those Frenchies seek him everywhere

Is he in heaven? or is he in hell?

That demned elusive Pimpernel

– Baroness Emmuska Orczy, The Scarlet Pimpernel

Like the Scarlet Pimpernel, T follicular helper (Tfh) cells have been the subject of intense interest and study, and yet their true identity remains frustratingly elusive. The conundrum revolves around the fact that the defining features of Tfh cells, i.e. their anatomical localization within the B‐cell follicle and their functional capacity to provide help to B cells, do not provide easy handles with which to identify them, particularly when working with human blood samples. As a result, a number of cell surface molecules, such as CXCR5, programmed death protein 1 (PD‐1) and inducible T‐cell co‐stimulator (ICOS), have been used as surrogate markers of Tfh cells. However, expression of these markers is not restricted to Tfh cells. In addition, memory T cells expressing these markers have been identified in human peripheral blood, raising the possibility that these ‘circulating’ Tfh (cTfh) cells may somehow be related to the Tfh cells within secondary lymphoid organs. Consistent with this, cTfh cell defects have been described in a number of primary immunodeficiency diseases characterized by defective antibody production. Not surprisingly, cTfh cells are now becoming increasingly promoted as biomarkers of vaccine responses and disease activity in autoimmune diseases (such as systemic lupus erythematosus) and infectious diseases (such as HIV infection). Hence, also like the Scarlet Pimpernel, Tfh cells are now becoming ubiquitous. This review will discuss recent advances in our understanding of Tfh cell migration, Tfh cell memory and the origins and fate of cTfh cells in the blood.

From little things big things grow

During the course of a natural infection, the elimination of invading pathogens requires the generation of antigen‐specific high‐affinity antibodies. Although this important function is ultimately carried out by B cells, it was recognized quite early on that this could not occur in the absence of thymus‐derived T cells.1, 2, 3, 4 Hence, the notion of the CD4⁺ T‐cell ’help’ to B cells to induce antibody production was conceived. Further studies began to reveal the intricacies of this response, which occur in specialized structures called germinal centres (GCs) within the B‐cell follicle of secondary lymphoid organs.5 Moreover, it was quickly realized that CD4⁺ T helper cells came in lots of different flavours and had specialized functions, depending on the spectrum of cytokines they produced. In terms of B‐cell helper function, the first B‐cell growth and differentiation factor was found to be interleukin‐4 (IL‐4).6, 7 Along with this finding came the notion that IL‐4‐secreting ’T helper type 2’ (Th2) cells were the CD4⁺ T‐cell subset that induced antibody production from B cells.8 However, mouse studies revealed the presence of intact GCs and T‐cell‐dependent antibody responses even when Th2 cell development had been disrupted,9 suggesting that these cells were not the primary T helper cell subset involved in providing B‐cell help. It was not until the year 2000 that another subset of CD4⁺ T helper cells with specific B helper capabilities was described. These so called Tfh cells were CD4⁺ T cells that localized in the B‐cell follicle of human secondary lymphoid tissues.10, 11, 12 The defining feature of these Tfh cells was their expression of the B‐cell zone homing chemokine receptor CXCR5, which facilitated their positioning in the B‐cell follicle where they are placed to interact with antigen‐specific B cells.13 However, it was quickly revealed that in addition to their anatomical location, Tfh cells possessed unique features that gave them specialized B‐cell helper capabilities.14 This included expression of the cytokine IL‐21, a potent inducer of B‐cell activation, proliferation and differentiation,15, 16, 17, 18 and expression of co‐stimulatory molecules such as CD40 ligand (CD40L) and ICOS.19, 20, 21, 22

From these humble beginnings, Tfh cells came of age with the landmark finding by three separate groups that Bcl6 was a lineage‐specific transcription factor required for the differentiation of naive murine CD4⁺ cells into Tfh cells.23, 24, 25 This unequivocally established Tfh cells as a separate CD4⁺ T helper cell population. However, while Bcl6 is required for Tfh cell development, it should be pointed out that its expression is not Tfh cell‐specific because it is up‐regulated in all dividing CD4⁺ T cells following activation by dendritic cells.26, 27 Nevertheless, in addition to Bcl6, numerous other transcription factors were subsequently found to be involved in the transcriptional regulation of Tfh cell differentiation. Moreover, it was quickly realized that there exists an intricate interplay between transcription factors involved in promoting (c‐MAF, BATF, ASCL2, IRF4, STAT1, STAT3, STAT4, NOTCH1/2, EGR2/3, TCF‐1, LEF‐1, NFAT) as well as inhibiting (BLIMP‐1, FOXO1, FOXP1, KLF2) Tfh cell formation28 (reviewed in ref. 21). Furthermore, other factors such as microRNAs (miRNA) have also been shown to be involved in positively and negatively regulating Tfh cell development28 (reviewed in ref. 21). Concomitantly, our rapidly evolving understanding of Tfh cells has been extended by recent insights from multiple intravital two‐photon microscopy studies investigating the migration pattern and interactions between Tfh cells and GC B cells (Fig. 1). Tfh cells are dependent on interactions with B cells and B‐cell‐derived signals such as the signalling lymphocyte activation molecule (SLAM) family of immune cell receptors29 and ICOS ligand (ICOSL)30, 31, 32 for induction of the Tfh cell transcriptional programme and follicular localization. Remarkably, within the GC, Tfh cells engage in only brief ‘entanglements’ with GC B cells that involve extensive cell surface contact between the two cells.33, 34 These short‐lived interactions are dependent on ICOSL expression on B cells, and result in increased delivery of CD40 signals to promote positive selection of high‐affinity clones.33

Figure 1.

T follicular helper (Tfh) cells on the move. Schematic showing possible relationship between lymphoid Tfh and circulating Tfh (cTfh) cells. In the primary antibody response (left panel), CD4⁺ T cells activated by dendritic cells in the T‐cell zone (1) migrate to the T‐cell–B‐cell border to interact with activated B cells and subsequently enter the B‐cell follicle. Some of the precursor Tfh cells may leave and enter the circulation as cTfh rather than migrate to the T‐cell–B‐cell border. Within the follicle, two distinct populations of Tfh cells are localized in the follicular mantle (FM) (2) and germinal centre (GC) (3). FM and GC Tfh cells are spatially confined to their respective microanatomical compartments. In the secondary antibody response (right panel), FM Tfh cells (5) and GC Tfh cells (6) are able to freely migrate from one compartment to another. Secondary Tfh cells are also able to enter the lymphatic sinus (7) to ‘surf’ the lymphatic system to downstream follicles and ultimately egress from the lymph node as a cTfh cell (8).

Tfh cells on the move

In the primary antibody response, Tfh cells have been shown to comprise two spatially and molecularly distinct subpopulations, one localized to the GC (GC Tfh cells) and the other to the follicular mantle (FM) zone (FM Tfh cells; Fig. 1).35 GC Tfh cells expressed higher levels of Cxcr5 and Pdcd1 and were the only cells that expressed Il21 mRNA. This, and other imaging studies, showed that GC Tfh cells were spatially confined to the GC and did not exchange with FM Tfh cells.29, 35, 36, 37 The spatial segregation was maintained in part by GC Tfh cells down‐regulating expression of the chemoattractant receptor EBI2,35 which attracts cells to the outer follicle,38 and increasing expression of the chemorepulsive receptor S1PR2,37 which repels cells from the S1P‐rich lymph in the subcapsular sinus towards the follicle centre.39 In addition to differential expression of chemokine receptors, the retention of Tfh cells in the GC is regulated by contact‐dependent repulsion of Tfh cells by Ephrin B1‐expressing GC B cells.36 Paradoxically this repulsion of Tfh cells appears to be important for their capacity to secrete IL‐21, possibly by preventing Tfh cell exhaustion by providing respite from interacting with GC B cells.

Interestingly, the differential expression of EBI2 and S1PR2 between FM and GC Tfh cells was not evident in the secondary antibody response, and a different cell migration pattern emerged with free exchange of cells between the two compartments.35 Notably, GC Tfh cells were observed not only to leave the GC and enter the FM, but also to migrate from the follicle into the subcapsular sinus where they were carried away by the lymphatic flow, potentially to ‘land’ on other ‘downstream’ follicles and GCs (Fig. 1).35 Indeed, the migration of secondary Tfh cells from one GC to another was also reported in another study that used a prime‐boost vaccination strategy.34 The dissemination of secondary Tfh cells via the lymphatic channels in the lymph node suggests that in the secondary antibody response Tfh cells may also eventually return to the bloodstream. Hence, although T follicular memory cells were found to reside in a niche in the subcapsular region of the lymph node,35 it is possible that some reactivated secondary Tfh cells leave the lymph node during an active immune response (Fig. 1). This notion that Tfh cells or their precursors may enter the circulation is not new. It was previously shown that circulating CXCR5⁺ CCR7^lo PD‐1^lo CD4⁺ T cells with B‐cell helper activity were generated before GC formation.40 Indeed, these cells were also present in the blood of patients with X‐linked lymphoproliferative disorder, and so are formed independently of SLAM‐associated protein (SAP), and are postulated to be an ‘early memory’ cell.40

There and back again – from humans to mice and back again

These studies of Tfh cell dynamics and migration in murine models support the notion that Tfh cells may egress from the lymph node and enter the circulation, possibly as memory T cells (Fig. 1). Tfh cells were originally defined by their anatomical location in the B‐cell follicle of human tonsils, which made studying Tfh cells in humans extremely difficult, as secondary lymphoid tissues are not readily accessible. Therefore, in order to track human Tfh cells during vaccine responses and in disease settings, there has been a concerted effort to identify the circulating memory counterparts of bona fide Tfh cells in peripheral blood. To this end multiple groups were able to show that CXCR5⁺ CD4⁺ T cells in human peripheral blood produced the cytokines IL‐21 and IL‐10, expressed cell surface molecules characteristic of lymphoid Tfh cells (i.e. ICOS, CD57, PD‐1, CD40L) and importantly were able to provide more help to B cells than CXCR5⁻ CD4⁺ T cells in in vitro co‐culture assays.41, 42, 43, 44 However, compared with lymphoid Tfh cells, cTfh cells expressed much lower levels of characteristic cell surface molecules (i.e. ICOS, PD‐1, CD40L) and did not express the Tfh lineage transcription factor BCL‐6.40, 41, 42, 43, 44, 45, 46, 47 Furthermore, it was quickly realized that there was substantial heterogeneity within the cTfh cell population, with the identification of Th1‐like (CXCR5⁺CXCR3⁺ CCR6⁻), Th2‐like (CXCR5⁺ CXCR3⁻ CCR6⁻) and Th17‐like (CXCR5⁺ CXCR3⁻ CCR6⁺) Tfh cells.43, 44 In addition to this, it was evident that not all types of cTfh cells were equal. For instance CXCR5⁺ CXCR3⁺ Th1‐like Tfh cells were found to be poor B‐cell helpers as their high expression of interferon‐γ and PD‐1 is likely to inhibit B‐cell activation and antibody production.43, 44 Hence, it was clear that the quality of cTfh cells generated is just as important as the quantity and there is a need to assess cTfh cells in terms of additional phenotypes and function and not solely based on the expression of CXCR5 by CD4⁺ T cells

cTfh cells as blood biomarkers of disease

As it is generally accepted that CXCR5⁺ cTfh cells were reflective of lymphoid Tfh cells, numerous studies focused on tracking cTfh cells in different disease settings and during vaccine responses. As such, reduced frequencies of cTfh cells have been revealed in primary immunodeficiencies due to monogenic loss‐of‐function mutations in ICOS, STAT3, IL10R, CD40LG, NEMO/IKBKG, BTK and TCF3.43, 48, 49 However, as highlighted above, not all cTfh cells are equal and although loss‐of function mutations in IL21 or IL21R or gain‐of function mutations in STAT1 had no effect on the frequencies of cTfh cells, cTfh cells in these individuals were primarily of the non‐B‐cell helper Th1‐type.43, 50 Other mutations in genes such as SH2D1A, which encodes SAP, do not appear to affect the frequency of cTfh cells but rather impact their function as they are unable to provide help to B cells due to a defect in forming stable contacts with antigen‐specific B cells.29, 40, 51, 52, 53 On the flip side, an increase in cTfh cells has been described in numerous autoimmune diseases. This includes Sjögren syndrome, rheumatoid arthritis, systemic lupus erythematosus, juvenile dermatomyositis, autoimmune myasthenia gravis, autoimmune thyroid disease and multiple sclerosis (reviewed in refs 19, 50, 54, 55).

With regards to vaccine responses, tetanus‐, cytomegalovirus‐ and influenza‐specific cells can be detected in the CXCR5⁺ cTfh cell population in healthy adults.40, 44, 47, 56, 57 These so‐called ‘memory’ cTfh cells expressed PD‐1, but neither ICOS nor CD38. In contrast, after influenza vaccination there was an increase in ‘activated’ cTfh cells, defined as proliferating cells expressing high levels of PD‐1, ICOS and/or CD38.42, 47, 56, 58 It was recently revealed that following seasonal influenza vaccination there is an increase in multiple clonotypes within the activated CXCR5⁺ ICOS⁺ CD38⁺ cTfh population, but these cells did not persist and were quickly turned over. Instead, antigen‐specific memory cells remained in the CXCR5⁺ ICOS⁻ CD38⁻ cTfh cell population and there was some evidence of a low level of conversion between ICOS⁻ CD38⁻ and ICOS⁺ CD38⁺ cTfh cells.56 These studies have provided precedence for monitoring CXCR5⁺ activated ICOS⁺ CD38⁺ and memory ICOS⁻ CD38⁻ cTfh cells during natural infection or following vaccination to determine the quality of an immune response and the persistence of a memory response, respectively. Interestingly, while efforts to generate an effective HIV vaccine have been unsuccessful, the frequency of CXCR5⁺ PD1⁺ cTfh cells has been associated with the presence of broadly neutralizing antibodies in HIV‐infected individuals.57

Tfh cells as an HIV reservoir

The primary role for a Tfh cell is to provide help to B cells for the generation of protective antibodies. However, more recently there has been evidence to suggest that Tfh cells also act as a reservoir for infection and replication of HIV.59, 60 Tfh cells have been shown to be increased in the setting of HIV infection, and HIV‐infected individuals often develop hypergammaglobulinaemia. However, Tfh cell function is likely to be defective as HIV‐infected individuals have dysregulated antibody production and the HIV vaccine is an ineffective inducer of antibody responses.59, 60, 61 Indeed, compared with cTfh cells from HIV‐negative controls, cTfh cells from HIV‐infected individuals were less capable of providing B‐cell help in vitro.61 This was in part due to interactions between PD‐1 on Tfh cells and PDL‐1, which is aberrantly expressed at high levels on GC B cells from HIV‐infected individuals. Ligation of PD‐1 on Tfh cells by PDL‐1 resulted in increased cell death and decreased IL‐21 expression.61 Instead, the Tfh cell compartment in HIV‐infected individuals contained the highest number of copies of HIV DNA, supporting the theory that it is a reservoir for HIV.60 Interestingly, both Tfh cell frequencies and HIV DNA decreased following treatment with combined antiretroviral therapy.59, 60 This suggests that the virus is driving the expansion of cTfh cells. It has been hypothesized that HIV has adapted to using Tfh cells as a reservoir to evade detection and subsequent clearance by CD8⁺ T cells, which generally do not express CXCR5 and so are rarely observed in the B‐cell follicle or in close proximity to Tfh cells. However, this theory has been recently questioned with the discovery of a small population of CXCR5⁺ CD8⁺ T cells in secondary lymphoid tissues, which are believed to be involved in controlling chronic viral infections.62, 63, 64, 65, 66

Outstanding questions

Even after his unmasking, the Scarlet Pimpernel was able to evade his pursuers and save the day. In the same way, although we have made major advances in the understanding of Tfh cells, their true nature remains elusive and many questions remain unanswered. For example, we still do not know the function of FM Tfh cells, or whether T follicular memory cells arise from FM or GC Tfh cells (Fig. 1). Importantly, it is still unclear how cTfh cells in the blood relate to bona fide Tfh cells that are found in secondary lymphoid tissues. Indeed, microarray analyses have revealed that cTfh cells are very similar, but still differ significantly from those found in secondary lymphoid tissues.45, 57 Identification of the true nature of cTfh populations will be required if they are to be an important biomarker for disease settings or the development of an intact immune response following vaccination.

Disclosures

The authors have no conflicts of interest to disclose.