With a Little Help from T Follicular Helper Friends: Humoral Immunity to Influenza Vaccination

Although 2018 marks the 100-y anniversary of the 1918 catastrophic influenza pandemic, which killed ∼50 million people worldwide, seasonal, pandemic, and avian influenza viruses are a constant global health threat (1). The annual seasonal circulation of influenza A virus (IAV) and influenza B virus (IBV) cause 3–5 million cases of severe illness and ∼500,000 deaths each year (1). Influenza disease can be particularly severe in young children, pregnant women, and the elderly for IAV (1) and mainly in the young during IBV infections (2). Additionally, avian IAVs, such as H5N1 and H7N9, can cause severe and often fatal human influenza disease (3) and could potentially cause devastating pandemics if these viruses ever acquire the ability of sustained human-to-human transmission.

Immune protection against influenza viruses and disease is conferred by multiple modalities of the human adaptive immune system, including Abs and B cells as well as CD8⁺ and CD4⁺ T cells (1, 4). Humoral immunity is mediated by 1) sterilizing Abs, which are most often strain-specific; or 2) broadly neutralizing Abs, which target the highly conserved viral hemagglutinin (HA) stalk region and contribute to protection via a plethora of mechanisms. Following infection with an antigenically novel strain, pre-existing cross-reactive memory B cells become reactivated and provide a rapid source of Abs as well as undergo adaptive evolution to the new Ag. Thus, the processes controlling the development and recall of memory B cells and Ab production are central for effective humoral immunity against influenza.

A considerable component of immune memory to influenza viruses is provided by high-affinity Ag-specific Abs, which can efficiently neutralize invading pathogens. The generation of such high-affinity Abs as well as their memory B cell counterparts results from a well-orchestrated process termed the germinal center (GC) reaction (5, 6) (Fig. 1). The GC forms in B cell follicles in secondary lymphoid organs (SLOs). Ag presentation by specialized dendritic cells (DCs) results in the expansion of cognate CD4⁺ T cells, which, under the appropriate polarizing environment, differentiate toward the T follicular helper (Tfh) lineage. The incoming Ag also activate B cells, which capture and process the Ag for presentation on the cell surface in the context of MHC class II. Tfh cells recognize cognate peptide MHC class II complexes on activated B cells, resulting in a two-way interaction that 1) re-enforces the Tfh lineage on the CD4⁺ T cells and 2) promotes B cell differentiation. B cells will either differentiate into extrafollicular short-lived Ab-secreting cells (ASCs) or enter the GC reaction. In the GC, B cells proliferate in the dark zone, where they diversify their BCR sequence via somatic hypermutation (SHM) to generate a heterogeneous pool of B cell clones with variable affinities for their Ag. Following exit from the dark zone and migration to the light zone, B cells compete for Ag deposited on follicular DCs. B cells bearing high-affinity BCRs capture and internalize more Ag than lower-affinity B cells, resulting in more presentation of Ags to Tfh cells. The interaction with Tfh cells can lead to two distinct outcomes: 1) re-entry in the GC reaction for further SMH, or 2) exit from the GC and differentiation in either memory B cells or long-lived plasma cells (LLPCs). It is thought that the nature of the interaction between GC B cells and Tfh cells determines the fate of GC B cells. This is supported by mouse studies showing that changes in BCR affinity for Ag and/or changes in the Tfh help available can affect the LLPC or memory B cell output of the GC (5). According to this model, B cells that receive little Tfh help differentiate into memory B cells, whereas those with intermediate levels of help re-enter the GC reaction. Furthermore, B cells receiving high levels of Tfh help differentiate into LLPCs (5). Memory B cells primarily reside in SLOs, with some circulating in the periphery, whereas LLPCs migrate to the bone marrow, from which they provide long-lasting serological memory. As the density of Ag presented on B cells is determined by the amount of Ag taken up by the B cell, Tfh cells indirectly sense BCR affinity and have a pivotal role in directing the outcome of the GC reaction, so that high-affinity LLPCs provide serological memory, whereas memory B cells of lower affinity but broader reactivity can provide recall responses upon Ag re-encounter (5, 6). Thus, Tfh cells have a key role in the entry and exit of B cells into the GC reaction, which in turn has a substantial effect on the quality of the humoral response.

Tfh cells were first identified in human tonsil samples as a CXCR5-expressing subset of memory CD4⁺CD45RO⁺ T cells (7, 8). Expression of the migratory chemokine receptor CXCR5 and concomitant lack of expression of the lymph node–homing receptor, CCR7, allows these CD4⁺ T cells to migrate toward the B cell follicles. These cells have an activated CD69⁺HLA-DR⁺PD-1⁺ phenotype, express high levels of the costimulatory molecules CD40L and ICOS, and support Ig production by B cells (7, 8). Human Tfh cells express transcription factor Bcl-6, a signature of the Tfh lineage (9, 10). CD57 is also expressed on approximately one third of human tonsillar Tfh cells, and its expression can identify GC Tfh cells with superior helper capacity (10, 11). A hallmark feature of Tfh cells is the production of IL-21, which acts both on B cells to promote survival and differentiation and on Tfh cells to promote their survival in an autocrine manner (12). Further Tfh help to B cells is provided by secretion of IL-10 and costimulation through ICOS–ICOSL and CD40L–CD40 T–B cell interactions (10, 13).

Human lymphoid tissues have been instrumental in providing insights into development of Tfh immunity; however, tissue specimens are often difficult to obtain. A small analogous subset of circulating CD4⁺ T cells expressing CXCR5 (∼10% of total CD4⁺ T cells) (14) has been described in human peripheral blood within memory CD45RO⁺ T cells (7, 8, 15). These circulating CD4⁺CXCR5⁺ T cells display a superior capacity for IL-21 and IL-10 production, as compared with their CXCR5-negative counterparts, thus exhibiting superior B cell helper capacity (14). CD4⁺CXCR5⁺ T cells were actually identified in the blood over 30 y ago (15), and most studies described them as a genuine subset of Tfh cells, called circulating Tfh (cTfh) cells. However, it is important to note that blood CD4⁺CXCR5⁺ T cells exhibit key differences when compared with GC Tfh cells (Fig. 2A). First, cTfh cells express CCR7, allowing them to migrate to lymph nodes, whereas GC Tfh cells downregulate their CCR7 cell surface expression to exit the T cell zone and enter B cell follicles (16). Second, cTfh cells from healthy donors lack expression of the activation markers ICOS and CD69 (16). Consistent with a quiescent phenotype, cTfh cells need to be activated before exerting any B cell helper function (16). Last, cTfh cells lack expression of the Tfh master regulator Bcl-6 (16), although a modest upregulation of Bcl-6 has been reported in activated cTfh cells (17). However, cTfh cells express the Tfh transcriptional regulator c-Maf (18), which, in the absence of Bcl-6, can drive the expression of the Tfh signature molecules CXCR5 and IL-21 (19).

Despite key differences demarcating cTfh from GC Tfh cells, a shared transcriptional program was identified between human tonsil-derived GC Tfh cells and circulating CD4⁺CXCR5⁺(PD1⁺CXCR3⁻) T cells (18), suggesting that cTfh cells and GC Tfh cells are developmentally linked. Similarly, a recent analysis of TCRβ repertoire via deep sequencing from paired tonsil and peripheral blood revealed clonal relationships between GC Tfh cells and PD-1⁺ICOS^+/− cTfh cells (20). The authors provided evidence for a high degree of TCRβ repertoire sharing between tonsillar GC Tfh cells and blood CD4⁺CXCR5⁺PD1⁺ T cells. Although the exact relationship between cTfh and GC Tfh cells is still not entirely clear, it is postulated that cTfh cells arise from early Tfh cells, which have not fully matured or participated in a GC reaction (21). This is supported by studies in mice and humans deficient in the signaling adaptor SLAM-associated protein (SAP). In mice, SAP deficiency results in abnormal GC formation and Tfh maturation, yet the frequency of cTfh cells is maintained. Similarly, cTfh cells can be found at equal frequencies in female carriers of X-linked lymphoproliferative disease, characterized by SAP deficiency (21). It is worth noting that GC Tfh cells can downregulate Tfh molecules like Bcl-6 and upregulate CCR7 in the late stages of a GC response (>1 mo) in mice to acquire a memory phenotype resembling cTfh cells. However, the contribution of these GC Tfh cells to the circulating cTfh pools is unknown. Thus, it has been proposed that a subset of developing Tfh cells exit the draining SLOs after being primed and disperse to other nondraining SLOs, where they can generate new GC reactions following Ag re-encounter (21). Therefore, cTfh activity reflects Tfh differentiation and activity in SLOs.

cTfh cells are generally divided into three subsets based on their expression of the inflammatory and skin-homing chemokine receptors CXCR3 and CCR6, respectively: cTfh1 (CXCR3⁺CCR6⁻), cTfh2 (CXCR3⁻CCR6⁻), and cTfh17 (CXCR3⁻CCR6⁺) (Fig. 2B) (14). These three subsets resemble the conventional CD4⁺ T cell subsets not only in their CXCR3/CCR6 phenotype but also in their transcription factor expression profile, whereby cTfh1 cells express T-bet, cTfh2 cells express GATA3, and cTfh17 cells express RORγt (14). In terms of their cytokine function, cTfh1 cells mainly produce IFN-γ, cTfh2 cells produce IL-4, and cTfh17 cells produce IL-17A (14). All three subsets produce the prototypical Tfh cytokine IL-21, albeit at varying levels, with cTfh17 and cTfh2 cells being superior IL-21 producers compared with cTfh1 cells (14). This corresponds to a superior ability of cTfh17 and cTfh2 cells to provide help to B cells, especially naive B cells, whereas activated cTfh1 cells can provide limited help to memory B cells but not naive B cells (14, 22). For all three cTfh subsets, provision of B cell help is limited to activated cTfh cells, which upregulate activation markers PD-1, ICOS, and CD38 and downregulate CCR7 molecules on the cell surface (Fig. 2C). Activated CD38⁺ICOS⁺ cTfh cells also express the intracellular proliferation marker Ki-67. Provision of B cell help is also provided by secretion of the anti-inflammatory cytokine IL-10 and expression of ICOS (14, 16).

cTfh subsets can also be differentially involved in immune responses toward human infections, immunizations, and autoimmune diseases. Activated PD-1⁺ and/or ICOS⁺ cTfh1 cells appear in the circulation following influenza (22, 23) and human papillomavirus (24) vaccinations as well as malaria infection (25). Conversely, cTfh17 cells were increased following vaccination with a replication-competent recombinant vesicular stomatitis virus expressing Ebola virus Ags (26), whereas CXCR3⁻ cTfh2/17 cells were associated with the development of broadly neutralizing HIV Ab responses (18) and autoimmune diseases, such as juvenile dermatomyositis (14), systemic lupus erythematosus (27), and Sjogren syndrome (28). What drives such differential activation of distinct cTfh subsets during different human diseases and vaccination protocols to provide B cell help and drive Ab production is not fully elucidated, although the cytokine environment plays an important role (9).

Taken together, Tfh cells are a subset of Ag-experienced CD4⁺ T cells specialized in providing B cell help by promoting B cell differentiation and Ab production. Moreover, Tfh cells are very heterogeneous, comprising of GC Tfh cells in the lymphoid tissues and a circulating pool of diverse memory and quiescent cTfh cells. The role of activated cTfh cells and their impact on promoting Ab responses following influenza vaccination is discussed in depth in the remaining sections of this review.

Current inactivated influenza vaccine (IIV) protocols are our best way to combat the annual influenza epidemics and prevent IAV/IBV infections and severe illness. Annual IIVs target predominantly Ab-mediated responses to the external HA influenza protein. Given the ability of Tfh cells to promote high-quality Ab responses, the role of Tfh cells in the immune response to IIVs has been of great interest. Much advancement in the field of Tfh cells and influenza vaccination has been based on the identification of circulating CD4⁺CXCR5⁺ T cells (14, 22), providing novel insights into the ontogeny of human Tfh cells and an in-depth characterization of human Tfh responses and their role in generating high-quality Ab responses.

A seminal and elegant study by Bentebibel et al. (22) was the first to demonstrate a transient emergence of activated PD1⁺ICOS⁺ cTFh1 cells, but not cTfh2 or cTfh17, in the peripheral blood of individuals vaccinated with a trivalent influenza vaccine. These PD1⁺ICOS⁺ cTFh1 cells peaked at day 7 following IIV before returning to baseline by day 14. This observation was subsequently verified by several other groups in the settings of IIV immunization (17, 23, 29, 30). It is now well known that IIV-induced cTfh cells express activation markers PD-1, ICOS, CD38, and Ki-67 (17, 22, 23, 30), and most importantly, the emergence of activated cTfh1 cells in peripheral blood on day 7 after IIV correlates with the emergence of ASCs and increased levels of serum Ab titers (17, 22, 23, 29).

The specificity of cTfh cells toward IIV was confirmed by in vitro experiments in which day 7 PBMCs were cultured with influenza Ags or killed H1N1-PR8 influenza virus. Following in vitro stimulation, ∼70% of CD4⁺ T cells expressing CXCR5⁺ICOS⁺CD154⁺ produced IFN-γ, whereas ∼50% and ∼40% produced IL-2 and IL-21, respectively (22). cTfh1 cells from day 7 post-IIV also promoted the induction of CD38^hiCD138⁺ ASCs, in a manner dependent on IL-21 and IL-10 (22). Further experiments with CXCR5⁺CXCR3⁺ICOS⁺CD4⁺ T cells isolated from day 7 post-IIV clearly demonstrated that these cTfh cells efficiently stimulated memory, but not naive, B cells to differentiate into plasma cells (22). Interestingly, in a cohort of IIV-immunized children, activation of cTfh1 cells correlated only with Ab responses against the HA components of the seasonal A/H3N2 and B influenza viruses but not from the A/H1N1 virus (22). Because these children (mean age of 11) were vaccinated with the pandemic H1N1 A/California/7/2009 strain in the 2010/2011 season, they may have been previously unexposed to the pandemic H1N1-HA vaccine component, further supporting the notion that cTfh1 activation stimulates memory B cell–derived Ab responses but not naive B cell–derived Abs (22).

In a follow-up study by Bentebibel et al. (31), activated cTfh1 cells correlated with increased affinity of IIV-induced Abs, supporting the idea that activated cTfh1 cells reflect GC activity in SLOs. Recently, we demonstrated that activated cTfh1 cells also correlated with the magnitude of the memory B cell responses, particularly circulating CD21^hi and CD21^lo CD27⁺ memory B cells (23). Given that CD21^lo CD27⁺ B cells are clonally related to memory B cells and ASCs, and that CD21^lo CD27^hi B cells have been identified as recent GC emigrants poised toward the LLPCs lineage (32), the correlation between cTfh1 cells and CD21^lo memory B cells further supports the idea that cTfh activity reflects GC activity.

It is thus well established that activated cTfh1 cells play an important role in humoral responses to influenza Ags following IIV vaccination. However, reports also emerge on the importance of cTfh cells following immunization with MF59 vaccines. Studies from Spensieri and colleagues (33, 34) show that the magnitude of IL-21–producing CD4⁺PD-1⁺ICOS⁺ T cells positively correlated with the Ab responses in individuals vaccinated with a MF59-adjuvanted H5N1 inactivated vaccine and following overnight in vitro stimulation with influenza virus Ags, although these CD4⁺ T cells did not express the Tfh marker, CXCR5, in one study (33).

Important insights into the ontogeny of activated cTfh cells emerging in the peripheral blood after influenza vaccination were provided by Herati and colleagues (17). TCRβ repertoires of activated CD38⁺ICOS⁺ cTfh cells were dissected over multiple years of consecutive IIV vaccination and showed a recruitment of recurrent TCRβ clonotypes across different years, suggesting that cTfh cells can be recalled from a pool of resting CD38⁻ICOS⁻ cTfh cells (17).

Overall, based on the above-mentioned studies, we propose a model in which following influenza vaccination, memory cTfh1 cells become activated and provide help to memory B cells through cytokine production and ICOS costimulation, thereby promoting memory B cell differentiation into ASCs and potential re-entry into the GC for affinity maturation and differentiation into CD21^hi and CD21^lo memory B cells (Fig. 3).

Elderly adults ≥65 y of age are particularly susceptible to severe influenza disease (1). Elderly individuals have higher attack rates, more severe disease, and reduced vaccine effectiveness, attributed to declining immunity and immunosenescence. The age-dependent decline in IAV-specific T cell and B cell responsiveness can reflect both immune dysfunction and “repertoire holes” from the loss of IAV-specific lymphocytes (35, 36). Recent reports provide consistent data that the elderly adults also have reduced frequency of cTfh cells at steady-state but higher expression of ICOS, but not CD38 or Ki-67, as compared with young adults (29, 30). Importantly, cTfh cells from elderly adults have decreased helper capacity toward naive B cells, as measured by IgG and IgM production after T and B cell cocultures (29).

At day 7 after IIV, ICOS expression increased in young adults but not in the elderly individuals, although CD38 and Ki-67 expression increased in ICOS⁺ cTfh cells across both cohorts, suggesting that cTfh cells do indeed become activated in the elderly following vaccination. However, activation of cTfh cells in the elderly did not correlate with influenza vaccine–specific serum IgM and IgG responses, although cTfh cells in adults predicted serum Ab levels (29). Pilkinton et al. (30) also independently demonstrated that IIV induces cTfh activation (CD38, ICOS, Ki-67) in elderly adults. However, in their cohort, the changes in activated cTfh cells correlated with changes in ASCs as well as Ab serum titers against H1 and H3 but not B components of the vaccine. Importantly, high-dose (60 μg) IIV induced greater cTfh cell activation compared with the standard (15 μg) IIV dose.

Overall, aging impacts the cTfh cells in a steady-state as well as cTfh-mediated responses to IIV, which might be overcome by alternative vaccination regimens.

Immunosuppressed individuals, including HIV patients, are also a high-risk group from severe influenza illness. Tfh cells have a central role in HIV infection, both as a viral reservoir and as mediators of B cell and Ab responses (13). HIV infection increases the frequency of Tfh cells in lymph nodes (37) but reduces cTfh cells in the periphery, although their numbers are partially restored following antiretroviral therapy treatment (38). Vaccination of HIV-positive adults with a monovalent H1N1 formulation results in the activation of Ki-67⁺ cTfh cells in vaccine responders. Similarly, the frequency of ICOS⁺ cTfh cells positively correlates with the Ab serum titers following vaccination. cTfh cells from HIV-positive responders, compared with nonresponders, had a superior B cell helper capacity toward autologous B cells (39). Similar findings were observed by de Armas et al. (40) in a cohort of HIV-positive children, with comparable frequencies of cTfh cells found in both HIV-positive and HIV-negative children. Within the HIV-positive children, responsiveness to IIV (as measured by a 4-fold increase in Ab serum titers via HA inhibition assay) was associated with transcriptional changes related to Tfh function, including upregulation of IL-21 and CXCR5. Conversely, a transcriptomic signature, including higher expression of IL-2 and STAT5, which inhibit Tfh differentiation, was observed in nonresponders (40). In a study using lymph nodes from HIV-positive donors, the frequency of Tfh cells, particularly CD57⁺ Tfh cells, was predictive of the HA inhibition titer toward the IBV component of IIV but intriguingly not from the IAV-derived H1 or H3 proteins (37).

Aging in combination with HIV infection appears to have an additive effect, as HIV-positive elderly (>55 y) individuals exhibited the weakest response to IIV in terms of cTfh cell activation and IL-21 production (41). Interestingly, this was associated with a heightened baseline of activated CD4⁺ T cells and TNF levels in the HIV-positive elderly group, compared with HIV-negative elderly and HIV-positive or HIV-negative young adults (41).

Overall, impaired Tfh cell responses, due to aging and/or immunosuppressive diseases such as HIV, result in diminished B cell responses and Ab production. A summary of the key studies highlighting the role of Tfh cells following different influenza vaccination strategies in a range of healthy and high-risk groups is outlined in Table I.

The antigenic specificity of CD4⁺ T cells, which includes the Tfh cell population, is an important parameter that needs to be considered in the design of protective vaccines, as it can affect both the breadth and functionality of the CD4⁺ T cell responses (42, 43). Following IIV immunization, the specificity of activated cTfh cells against influenza Ag has been confirmed by stimulation with heat-killed IAV or IIV or influenza-derived peptide pools and measuring cytokine (IL-21, IL-2, IFN-γ) production and/or upregulation of the T helper activation marker CD40L (also referred to as CD154) or activation markers (CD200, CD69) (17, 22, 31). Activated cTfh cells specifically respond to peptide pools derived from the influenza proteins HA, nucleoprotein (NP), and matrix protein 1 (MI) (17, 31). Additionally, HLA-DRB1*04:01/HA_306–318 and HLA-DRB1*04:01/HA_398–410 tetramer-specific CD4⁺ T cells observed on day 7 after vaccination exhibited varying levels of activated CD38⁺ICOS⁺ and CXR5⁺PD-1⁺ cTfh phenotypes (17). Interestingly, Leddon et al. (44) reported that memory (CD45RA⁻) CD4⁺CXCR5⁺ cTfh cells from healthy adults preferentially recognized external HA-derived peptides, compared with the CXCR5⁻CD4⁺ T cell population, displaying preferential reactivity toward Ags derived from the internal influenza NP. Although it is important to note that these results are based on healthy donors and that cTfh cells emerging after vaccination can respond to internal M1 and NP Ags as well as HA, such bias toward HA-derived peptides can have important implications for a design of the vaccine content, as the inclusion of Tfh-specific Ags may promote more efficient Tfh responses and subsequent B cell immunity. Alternatively, a higher HA-specific Tfh response may be skewed by higher levels of HA in the vaccine content, whereby higher amounts of the HA protein can result in preferential presentation of HA peptides on B cells, favoring activation and boosting of HA-reactive Tfh cells. It would thus be pertinent to further understand the reactivity and specificity of Tfh cells during human vaccination and influenza virus infection.

Taken together, the role of human Tfh cells during the B cell response to influenza vaccination is well established. Human Tfh cells provide help to B cells through costimulatory molecules and the production of cytokines to promote effective B cell responses. Specifically, they promote production of ASCs, correlate with serum Ab levels, as well as boost memory CD21^hi and CD21^lo influenza-specific B cells. However, despite immense efforts to characterize Tfh responses during human vaccination and influenza virus infection, important questions remain. Specifically, where exactly are these cTfh cells reactivated and where do they exert their help function? Given that much evidence surrounding IIV influenza vaccination has focused on the CXCR5⁺CXCR3⁺ cTfh1 subset, should future vaccines aim to induce CXCR3⁻ cTfh2/17 cells, displaying greater helper capacity and being implicated in generating HIV-specific broadly neutralizing Abs? Could the induction of these cells improve the recruitment of naive B cells into the response? How does the Tfh response to vaccination compare with Tfh immunity following an influenza virus infection and whether Tfh cells correlate with protection and/or recovery from disease? It is conceivable that, given their ability to promote ASC and Ab responses following influenza vaccination, activation of Tfh cells may promote recovery from infection. Further questions remain about the ontogeny and origin of cTfh cells, their relation to GC Tfh cells, as well as their developmental relationship to conventional Th1 CD4⁺ T cells. Last, although current IIV induces cTfh1 responses, vaccination regimes such as a high dose IIVs, which further boost cTfh1 responses, quantitatively or qualitatively, or engage CXCR3⁻ cTfh cells, may provide novel strategies to boost humoral immunity toward seasonal and the potentially pandemic avian influenza viruses, especially in high-risk groups with diminished immune responses. Thus, novel vaccination strategies that aim to establish broadly protective immunity to IAV and IBV infections should be rationally designed to promote effective Tfh responses and thereby enhance B cell responses to prevent severe epidemic, pandemic, and avian influenza disease.

The authors have no financial conflicts of interest.