Emerging concepts on T_FH cell dynamics in HIV infection

Abstract

Inducing cross-reactive broadly neutralizing antibody (bNAb) responses to HIV through vaccination remains an insurmountable challenge. T follicular helper cells (T_FH) are fundamental for the development of antigen-specific antibody responses and therefore critical for anti-HIV vaccine design. Here we review recent studies supporting an intricate involvement of T_FH in HIV pathogenesis and bNAb development in HIV infection. We also examine emerging data suggesting that T_FH responses may be traceable in peripheral blood, and discuss the implications of these findings in the context of vaccine design and future research in T_FH immunobiology.

Antibody-mediated immunity and HIV

In the face of 2 million new infections per year, the stagnation of progress towards an efficacious HIV vaccine is sobering. In 2009 a double-blind phase III HIV vaccine RV144 “Thai” trial that used a combination of a recombinant canarypox vector (ALVAC-HIV [vCP1521]) and two booster injections of a recombinant glycoprotein 120 subunit vaccine (AIDSVAX B/E) showed marginal, yet significant protection from HIV acquisition[1] raising considerable hope for a protective vaccine. However, the results of follow-up assessments of the correlates of protection were surprising and unexpected[2]: Assessing multiple immunological parameters, the study found that not only HIV Envelope (Env)-specific, non-neutralizing IgG antibody (Ab) responses were correlated with protection, but also that protection was lost in the presence of high levels of anti-Env IgA Ab[2]. These findings raise questions regarding whether other Ab-related effector mechanisms [such as Ab-mediated cellular cytotoxicity (ADCC)] may have played a role in the marginal levels of protection observed in RV144, in contrast to most vaccines that commonly –yet not exclusively - induce neutralizing Ab (NAb) responses[3]. These data highlights critical gaps in our understanding of what constitutes a protective vaccine-induced immune response, and more importantly, raises questions as to the mechanisms that shape the Ab response to viral challenges and how these can be tailored by vaccination.

The large genetic diversity of global HIV variants stipulates broad reactivity of NAbs. Thus the ideal Ab response against HIV infection would have broadly neutralizing activity covering the wealth of different worldwide HIV variants. Passive transfer of such broadly cross-reactive neutralizing Abs (bNAbs) has been shown to protect both prophylactically[4, 5] and therapeutically[6, 7] in both non-human primates (NHP) and mice. Yet, no immunization strategy attempted against HIV has succeeded in inducing bNAbs, and only a fraction of HIV-infected individuals naturally develop bNAbs after years of HIV infection[8–15]. The difficulty of generating such a response may, in part, lie in the particular characteristics of bNAbs. Overall, monoclonal bNAbs isolated from infected individuals are characterized by unusual properties such as long antigen-contacting sites [complementarity determining region-3 (CDR-3)] and tremendous levels of somatic hypermutation (SHM), the extent of which is uncommon in other Ab responses[10, 16, 17]. These changes are characteristic of Ab affinity maturation processes, which occur in germinal centers (GC) within lymphoid tissues. T follicular helper cells (T_FH) play a cardinal role in Ab affinity maturation by promoting immunoglobulin class switching [class-switch recombination (CSR)], SHM, and B cell differentiation and survival[18]. During this process, termed the GC reaction, T_FH cells and antigen-specific B cells undergo a tight regulatory program that drives these mechanisms[19]. Thus, insight into the generation of protective Ab responses requires understanding of these cellular interactions and the underlying mechanisms.

There have been significant advances in our understanding of T_FH immunobiology in both natural settings and during HIV infection. While perturbation of the B cell compartment in HIV infection is well established[20], a number of recent studies have shown a link to T_FH immunopathology during HIV/SIV infection, including changes in T_FH frequency, phenotype and function[21–25]. Recent studies suggest that the rare and infrequent generation of HIV-specific bNAbs may be associated with T_FH frequency[12, 26]. This is further supported by mouse studies demonstrating a dynamic exchange of T_FH cells between different GCs[27], ensuring the maximal diversification of T cell help. Using two-photon microscopy, it was demonstrated that the ability of T_FH cells to enter ongoing GCs may accommodate antigenic variation during the immune response, maximizing the humoral response[27]. Moreover, emerging data suggest that T_FH responses may be traceable in peripheral blood[26, 28–31], providing a potential tool to monitor vaccine efficacy. In this article, we therefore focus on the immunobiology of T_FH cells in natural HIV infection and their involvement in HIV pathogenesis and bNAb development. We discuss studies that examine whether peripheral blood T_FH counterparts (pT_FH) may serve as surrogates for T_FH cells in lymphoid tissues, and provide an outlook on how these and other recent findings might inform HIV vaccine design.

T follicular helper cells and antibody-mediated immunity

Burnet’s proposal for the clonal selection theory of Ab formation[32] launched the modern age of cellular immunology, in which lymphocytes were established as the agents of immunity and divided into distinct humoral and cellular components. Subsequently, in the 1960s, Miller and Claman described for the first time a dependence of Ab production on the presence of T cells[33, 34], termed T-dependent responses. Then in 1986, Velardi and colleagues described a CD4+ T cell subset within the GC[35]. While this subset showed properties distinct from the known T helper 1 (Th1) and Th2 CD4 T cell subsets, these cells were formally recognized as a distinct subset only in early 2000, taking center stage as a key player for T-dependent responses[18]. Since then our understanding of the development and function of T_FH cells in both health and disease has grown rapidly in the past decade.

The transcriptional repressor B-cell lymphoma-6 (Bcl6) facilitates transcription of signature T_FH markers that mediate T_FH migration and function[36–38] while antagonizing other CD4 T cell differentiation programs[37, 38]. Expression of Bcl6 is initiated by DC-mediated priming of naïve CD4 T cells, triggering differentiation into T_FH cells in T cell zones of lymphoid follicles[39]. Bcl6 induction promotes CXCR5 expression, which in turn facilitates T_FH migration to the T–B cell border, promoting encounters with B cells specific for the same antigen (cognate) that reinforce T_FH lineage commitment[40]. Inducible T-cell costimulator (ICOS), a surface protein expressed on activated T cells, is central to all phases of T_FH development and GC formation[41, 42]. Additionally, in conjunction with IL-6, ICOS also promotes expression of IL-21, which is the cardinal cytokine for T_FH cells[43, 44]. IL-21 further augments Bcl6 expression to drive T_FH lineage differentiation[44–46] as shown by the inability of IL-21^−/− mice to develop T_FH cells with full helper activity and maintain GCs[45, 46]. These various cellular and molecular signals together contribute to B cell responses by initiating the T_FH program.

The coalescence of T_FH and B cells initiates the GC reaction, during which T_FH cells become fully polarized, highly activated GC T_FH cells[19]. GC T_FH cells are exposed to ongoing stimulation by cognate B cells presenting antigen and co-stimulatory signals resulting in heightened expression of T_FH markers, including ICOS and PD-1, in GC T_FH compared to T_FH cells[27]. ICOS signaling in GC T_FH cells induces expression of the transcription factor c-Maf[47], which synergizes with Bcl6 to orchestrate further initiation and maintenance of the T_FH program[48]. T_FH lineage differentiation is therefore tightly linked to its functional role in B cell differentiation, survival, and longevity of Ab responses.

Upon encountering antigen, B cells can differentiate directly into short-lived plasma cells (PC) or migrate into the lymphoid follicles to receive T_FH-derived help. Here B cells compete for a limiting source of T_FH signals based on the amount of MHC-II:antigen complexes presented on B cells[49]. The degree of T_FH signal is dictated by the quality and duration of B cell interactions with molecules expressed on the surface of T_FH cells[50]. These T_FH signals are critical for differentiation of GC B cells into long-lived PCs and memory B cells, as well as the ensuing maturation of Ab affinity through CSR and SHM[19]. Thus B cells with the strongest interactions with T_FH cells subsequently either become memory B cells or leave the GCs and differentiate into long-lived PCs[51], colonizing the bone marrow or tissues underlying epithelial surface. Several surface receptors govern this interaction in addition to MHC-II: First, the signaling lymphocytic activation molecule (SLAM) family of receptors and the SLAM-associated protein (SAP) family of intracellular adaptors are responsible for maintaining long-duration contacts with antigen-specific B cells[52]. SLAM-bound SAP triggers positive signaling through Src-family kinases while inhibiting negative signaling through the negative signaling receptor, LY108[50, 52]. Another T_FH surface molecule, the inhibitory receptor programmed death-1 (PD-1) is highly expressed on T_FH cells and actively involved in sustaining the GC reaction by inhibiting follicular regulatory T cells that suppress the GC response[53, 54]. Among other molecules, the interaction between CD40 on activated B cells and its ligand (CD40L) expressed on T_FH cells is one of the most crucial receptors for the development of T-dependent Ab responses and the generation of affinity-matured Abs. CD40 ligation leads to expression of activation-induced cytidine deaminase (AID), which is a critical enzyme regulating CSR, SHM and likely many other additional aspects in Ab affinity maturation[55]. Whether HIV-specific T_FH responses –including CD40L expression and cytokine secretion - differ between individuals that develop bNAbs or those that fail to do so remains unknown.

Another prominent factor for Ab affinity maturation is the cardinal T_FH cytokine, Interleukin-21 (IL-21). In addition to its central role in T_FH development, IL-21 is indispensable for Ab affinity maturation. IL-21 and Interleukin-4 have been shown to cooperatively direct Ab responses in both mice and humans[56], and further instructs the differentiation of B cells by stimulating expression of B cell master transcription factors[57]. Together, data suggest T_FH cells provide critical help to GC B cells and promote Ab affinity maturation by providing IL-21 in the presence of CD40 ligation. Additionally, SAP/SLAM-dependent IL-4 production provides further survival signaling[58]. Interestingly, IL-21-secreting CD4 T cells have also been shown to help sustain CD8 T cell effector function during chronic viral infections[59, 60] and are associated with HIV control[61]. However, whether T_FH cells, or a distinct IL-21-secreting CD4 subset, mediate this IL-21-dependent CD8 T cell “help” remains to be established.

T_FH cells in HIV pathogenesis

A fundamental characteristic of HIV pathogenesis is widespread immune activation and perturbation of the B cell compartment[20]. That CD4 T cells are the primary targets of HIV[21] and the specific induction of T_FH cells through vaccination may (temporarily) increase the pool of HIV target cells represents a duality in harnessing T_FH cells for HIV vaccinology. Thus understanding the immunobiology of T_FH cells in HIV infection and their involvement in the generation of bNAbs during natural HIV infection –in spite of a preferential depletion - is a fundamental question for HIV vaccine development. Currently, several studies have begun to explore T_FH cells in the context of HIV/SIV.

All studies of these cells to date have consistently demonstrated an unexpected expansion of both general and virus-specific T_FH frequency in both chronic HIV[21, 23, 24] and SIV[22, 25] infection, despite potential preferential depletion of antigen-specific CD4 T cells[62]. While no direct relationship with viral loads was observed[22, 23, 25], T_FH frequencies substantially decreased with ART treatment, suggesting a role for antigen levels in driving T_FH differentiation[21, 23]. Similarly, T_FH expansion has been previously reported in the lymphocytic choriomeningitis virus (LCMV) mouse model where it has been suggested that viral persistence and prolonged T cell receptor stimulation may progressively direct T cells towards a T_FH program[63]. However, another study of LCMV infection demonstrated a biphasic secretion of IL-6, the strongest driver of T_FH development, during acute infection and again during the chronic phase. Yet, only the chronic phase of IL-6 secretion seems to be responsible for driving subsequent T_FH expansion[64]. Interestingly, a similar biphasic IL-6 response has been demonstrated in acute HIV infection[65], but an association with human T_FH cell expansion in HIV remains to be shown. Nevertheless, studies of SIV infection have shown that T_FH expansion is strongly linked to a skewing in the IL-6 signaling axis[22]. Likewise, they have also reported elevated T_FH levels in chronic, compared to acute, infection suggesting that immune activation, rather than antigen levels may be driving T_FH expansion.

Expansion of T_FH cells has been associated with perturbation of the B cell compartment. In particular, GC B cell and PC frequency was positively correlated with T_FH frequency, which is not surprising given T_FH cells’ involvement in the development of these cells. Moreover, memory B cells in HIV⁺ Lymph Nodes (LN) were negatively correlated to T_FH frequency[23, 24]. In addition, aside from a general expansion of T_FH frequency, T_FH function seems to also be altered in HIV/SIV infection. Cubas et al. demonstrated that T_FH cells from HIV+ LNs less-efficiently promoted B cell survival and differentiation, likely due to deficient IL-21 secretion. However, T_FH dysfunction was primarily attributed to B cell dysregulation in that excess PD-1–PD-L1 ligation by dysregulated B cells impaired the overall T_FH cell response[24].

Despite compromised T_FH-derived help, T_FH expansion correlates with enhanced circulating SIV-specific IgG in NHP [22, 25] and hypergammaglobulinemia in humans[23]. Petrovas et al. described a positive association between T_FH frequency, expansion of the B cell compartment and increased circulating high-avidity SIV-specific Abs in chronic SIV infection; this finding suggests that chronic SIV infection may not disrupt T_FH-derived Ab affinity maturation and SIV-specific Ab production[22], but Ab affinity maturation may be impaired overall. Indeed, while higher T_FH frequency drives B cell proliferation, constant antigenic stimulation may lead to the selection of low affinity B cells[25]. Thus, it is likely that the low-quality Ab production described in HIV/SIV infection may be induced by a combination of factors, including T_FH /B cell dysfunction, high levels of antigenemia, and immune activation.

Expansion of virus-specific T_FH cells seems to conflict with the profound depletion of activated antigen-specific CD4 T cells seen in HIV infection, raising the question of whether T_FH cells are somewhat protected from infection/depletion. However, it has been demonstrated that SIV readily infects T_FH cells in GCs[22] and the T cell population within the lymph node that corresponds to T_FH cells (CXCR5⁺PD-1⁺Bcl6⁺) serves as the major T cell compartment for HIV infection, replication, and production[21]. Interestingly, a study of pigtail macaques found infection of T_FH cells in the absence of CCR5 or other HIV/SIV co-receptor expression[66]; however, whether infection of T_FH cells occurs through a different co-receptor or whether T_FH cells are infected prior to differentiation and subsequent CCR5 downregulation remains to be seen. Overall, the current data suggest that T_FH cells may represent a continually replenishing pool of target cells for SIV[22, 25, 66] and HIV[21]. Taken together, these data indicate a model in which HIV promotes B cell dysfunction while securing a viral reservoir in lymphoid tissues that are relatively devoid of CD8 CTLs[66] (Figure 1).

Figure 1. Development of T_FH cells and the germinal center reaction.

(1) Dendritic cells present antigen and activate naïve CD4 T cells via TCR:MHC-II, ICOS-ICOS-L, and CD40:CD40L interactions to initiate the T_FH program. DC- and B cell-derived IL-6 and IL-27 cytokines maintain and reinforce this lineage, leading to upregulation of the T_FH transcription factors Bcl6, cMaf, and CXCR5. (2) Expression of CXCR5 facilitates the migration of pre-T_FH cells toward the T—B cell interphase. Here, cognate pre-T_FH and B cells exchange molecular signals including ICOS:ICOS-L, CD40:CD40L, TCR:MHC-II, and several SLAM receptor family members, as well as several cytokines including the cardinal cytokine IL-21, that reinforce expression of Bcl6 and c-Maf that further increase expression of T_FH surface receptors. B cells simultaneously undergo diverse signals that instruct their maturation pathway: either extrafollicular or GC. (3) Extrafollicular activated B cells then convert to short-lived plasma cells and secrete Ab with or without extrafollicular T_FH help. (4) In the GC, prolonged interactions between GC T_FH and GC B cells further confirm the T_FH lineage commitment, resulting in mutual exchange of survival signals that result in GC maintenance. GC T_FH cells provide a limiting source of antigen for which B cells compete for survival signals based on BCR affinity. Furthermore, here GC T_FH cells instruct GC B cells to phosphorylate and express Activation-induced Deaminase (AID) that mediates SHM, CSR, memory cell and long-lived plasma cell formation.

While significant progress has been made in understanding the immunbiology of T_FH cells in HIV infection, several outstanding questions remain regarding how T_FH cells are involved in B cell dysfunction and/or the rare generation of bNAbs that occur after years of infection. Of particular interest is the causality of these observations: that is, it is unknown whether T_FH expansion directly promotes the perturbation of the B cell compartment or whether these are two simultaneously occurring infection-mediated mechanisms. Moreover, it remains to be shown whether the described alterations are overall beneficial or disadvantageous overall for the generation of bNAbs. As all HIV-infected individuals develop NAbs but only a fraction develops bNAbs, it will be critical to understand whether antigen-specific T_FH function, HIV antigenemia, immune activation, other unknown factors, or a combination of these drives the extensive levels of SHM observed in bNAbs. It is likely that multiple factors synergize to dampen or delay bNAb generation; for example, low-affinity Ab responses to linear epitopes may arise from lowered selection and competition threshold for T_FH help (from T_FH expansion) between B cells (with low-affinity BCR) specific for different viral quasispecies[49]; this may divert the Ab response away from neotopes. Coincidentally, HIV- or Env-derived B cell deregulation may impair T_FH-derived B cell help, disrupting affinity maturation, and further promoting low-affinity Ab production. However, this hypothesis is challenged by findings describing associations of T_FH-like counterparts circulating in peripheral blood with bNAb development (see below). Overall, dissecting the signals involved in promoting SHM for the generation of bNAbs and understanding which T_FH functions and B cell perturbations are unique to HIV (or other chronic infections) will be critical hurdles to overcome for effective vaccine design.

Broadly cross-reactive NAbs and T_FH cells

In recent years, there has been substantial progress in the identification and isolation of bNAb-producing B cells from chronically infected individuals. The profound SHM displayed by most bNAbs[67] suggests extended affinity maturation signals, likely mediated by T_FH cells, thus implicating a prominent role of T_FH cells in bNAb development.

In most patients, bNAbs develop only after years of chronic infection[9, 11, 12] and show remarkable properties including extremely high levels of SHM[10, 16, 68, 69], long CDR3 regions[70], and, in certain cases, polyreactivity[71]. For example, Klein et al. showed that monoclonal bNAbs undergo SHM in Ab framework regions that control molecular structure. Furthermore, these mutations were essential for bNAb function, as removing point mutation occurring in these regions ablated neutralizing activity[67]. Similarly, restoring monoclonal bNAbs to their germline sequences removes neutralization potency and Env specificity[72]. These high levels of Ab rearrangements (up to 42%) are not commonly found in other Abs[10, 16, 17], suggesting a constant and prolonged activation of AID. Activation and phosphorylation processes of AID occur through CD40-CD40L signaling[55] provided by T_FH cells suggesting an intimate role for T_FH cells in bNAb development as they undergo multiple rounds of selection in response to a continually evolving pathogen[73]. While these observations provide a glimpse into the difficulty of bNAb development, the mechanisms behind this process, the extent of T_FH involvement and the specific signals provided by T_FH cells remain unknown.

In this regard, optimal antigen recognition and presentation mechanisms would likely optimize T_FH B helper activity[51]. Interestingly, a retrospective study conducted on uninfected RV144 vaccine recipients demonstrated that patients carrying HLA-DQA1*05:01 and DQB1*03:01 alleles were less likely to produce NAbs[74] suggesting a direct involvement of antigen presentation to the induction of NAbs in vaccine recipients. No association between bNAb generation and HLA-II expression has yet been described in natural HIV infection, however. The extent to which HLA-II alleles influence B cell selection and whether HLA-DR or HLA-DP plays a more predominant role over HLA-DQ remains unknown. Moreover, it is also unknown whether different epitope-specificities of CD4 T cell responses are involved in NAb development and whether the abundance of Gag-specific responses over Env-specific T_FH responses impacts the development of NAbs in natural HIV infection[23].

Recent landmark studies reporting the isolation of potent bNAbs from HIV-infected patients with broad neutralization breadth and advances in deep sequencing of single-cell bNAb-secreting memory B cells and PCs have provided important insights into their maturation pathways[16, 75, 76] as well as bNAb binding mechanisms[10]. While these illustrate the profound intricacies of bNAb development, the hope is that they will inform rational vaccine design to harness T_FH cells to mediate targeted affinity maturation, PC differentiation and stimulation of bNAb responses. However, many questions surround this approach, particularly regarding what signals are required for optimized T_FH generation, and how to induce them through vaccination (Box 1).

Box 1: Critical questions regarding T_FH cells in HIV vaccine design.

Translating lessons from natural HIV infection to vaccination-induced NAb development is a hurdle that over 30 years of active research has yet to overcome. The resolution of several key issues, particularly regarding the duration of NAb development, is imperative to achieve this feat: (1) What is the relationship between T_FH cells, B cell dysfunction, and NAb development in natural infection; (2) Does viral quasispecies evolution promote continuous priming of Ab responses to linear epitopes, leading to constant production of low affinity antibodies rather than affinity maturation to conserved regions (or neotopes), and is this a result of B cell competition; (3) Does the number and location of SHM that are required to convey bNAb activity solely explain the delay in NAb development or is the SHM/AID machinery impaired by HIV pathogenesis, thus affecting the efficiency of T_FH:B cell help and the ensuing B cell proliferation, differentiation, and Ab affinity maturation; (4) Is the trimeric, highly glycosylated nature of HIV Env detrimental to affinity maturation and SHM by activating T-independent Ab responses or short-lived plasma cells; (5) Are certain Env variants more elusive than others at inducing beneficial T_FH responses; (6) Is this an issue of epitope miscommunication between T_FH cell and B cells or is it an issue of immunogenic silence; (7) Does antigenemia promote SHM in an individual with a healthier B cell compartment and how can this be achieved through vaccination; (8) Are T_FH cells a major latent reservoir during HIV infection; (9) Is T_FH cell dysfunction/expansion influencing the lack of bNAb development in the majority of patients and can T_FH functionality/dynamics account for the rapid bNAb development in certain subjects; (10) Can the induction of specific T_FH functions drive bNAb development; and (11) Can Gag-specific T_FH responses provide help for Env-specific Abs.

Peripheral T_FH cells in infection and vaccination

One inherent problem to studying co-dynamics of T_FH and Ab development in infection and vaccines in human patients is the paucity of lymphoid tissue available for experimental analyses. Thus, the ability to define a population of memory T_FH cells in peripheral blood (pT_FH) would facilitate our understanding of CD4 T cell dynamics within lymphoid tissue during vaccination and infection. Initially, Mikell et al. reported higher frequencies of circulating CD4 T cells expressing PD-1 in individuals with broad neutralizing serum activity[12], thus potentially associating the emergence of a T_FH-like population in peripheral blood with bNAb development, as PD-1 is highly expressed on T_FH cells[53, 54]. While this study was not designed to investigate pT_FH cell immunobiology, it was the first study to shine a spotlight on the potential associations between T_FH (and pT_FH) cells and bNAb development. Given the importance of such findings for the design of potential vaccines to specifically promote bNAb development, several groups have subsequently focused on characterizing the association between GC T_FH and peripheral counterparts during HIV-specific and other types of Ab responses[26, 28–31, 77].

While most studies agree that CXCR5⁺CD4 T cells in peripheral blood represent a counterpart of T_FH cells, their relationship to GC T_FH cells has been challenged based on differences in surface marker phenotypes and transcription factor expression levels [18]. Indeed, a large heterogeneity of circulating CXCR5⁺CD4 T cell subsets has been described in terms of gene expression[26, 31, 77], surface marker expression, and function[26, 28, 29, 31, 77], complicating clear conclusions on this issue. Indeed, emerging studies describe biologically relevant B helper activity in various subsets of CXCR5⁺CD4 pT_FH cells; moreover, these subsets have been defined using different groups of cell surface markers, further complicating comparisons between these studies (see below and Table 1). Emerging evidence suggests that T_FH memory is generated along with B cell memory, and combinations of phenotypic markers likely represent different layers of T_FH differentiation and effector function that is reflected in peripheral blood[28, 29]. Indeed, evidence shows that pT_FH cells (CXCR5⁺CD4⁺) are central memory cells comprised of Th1-, 2-, and 17-like subsets (based on CXCR3 and CCR6 expression) in healthy adults, each with diverse B helper functions in co-culture and in vivo[29, 30]. Of these, the Th1-like, CXCR3⁺CCR6⁻CXCR5⁺CD4 subset lacks naïve or memory B helper capacity in mice[29]. However, upon activation by trivalent influenza vaccine in humans, activated Th1-like ICOS⁺CXCR3⁺CCR6⁻CXCR5⁺CD4 subset efficiently induced memory, but not naïve, antigen-specific B cells to proliferate, differentiate, and produce Igs[30]. In contrast, the Th2- and −17-like pT_FH (CXCR3⁻CCR6^+/−) supported naïve B cell differentiation in mice[29] and humans[30]. Supporting these observations, evidence from studies of healthy and HIV-infected human specimens suggests CCR6^highCXCR3^+/−CXCR5⁺CD4 pT_FH induce Ab secretion from naïve B cells, although helper capacity is impaired in HIV-infected samples[77]. Indeed, this pT_FH subset (CCR7^highCXCR5^highCCR6^highPD-1^high) displayed the highest IL-21 production and naïve B cell help, inducing the greatest production of IgG1, IgG3, and IgA compared to other subsets in co-culture with naïve B cells[77]. These data suggest that perhaps CXCR3⁻CXCR5⁺CD4 pT_FH cells are potent inducers of naïve B cell differentiation while CXCR3⁺CXCR5⁺CD4 pT_FH primarily induce memory B cell differentiation.

Table 1.

Divergent surface marker profile of pT_FH cells with B helper activity.

	Morita (2011) pTfh	Mikell (2011)*-	Chevalier (2011) pTfh	Bentebibel (2013) pTfh	Locci (2013) pTfh	He (2013) pTfh & GC Tfh	Boswell (2014) pTfh & GC Tfh
CXCR5	+		+	+	+	+	hi**#
CXCR3	−			+	−		±/−
PD-1	+	+			+	hi	hi
ICOS				+
CCR6	+					hi	hi
CCR7					low	low	hi

^*study was not designed to interogate Tfh phenotype

^**CCR7^high CXCR5^low cells had limited naïve B helper activity

^#CXCRS^low cells in HIV viremic subjects displayed limited naïve B helper activity

Other studies have explored the relatedness of pT_FH with GC T_FH in the lymphoid follicle, based on other combinations of surface marker and gene expression profiles. These studies demonstrate that CXCR5⁺CD4 pT_FH cells are either activated (PD-1⁺ICOS⁺[30] or PD1⁺CCR7^lo[28]), or quiescent (PD-1^+/−ICOS⁻[26, 28]) cells. In this regard, CXCR3⁻ (PD-1⁺CXCR3⁻CXCR5⁺CD4⁺) pT_FH were described as memory cells with highly functional B helper activity in co-culture that share a transcriptional profile signature with GC T_FH cells, including expression of MAF and SLAMF6[26]. On the other hand, a slightly different pT_FH subset (CXCR5^highCCR6^highPD-1^high) was concluded to more closely resemble a memory, non-T_FH CD4 T cell population from the tonsil compared to non-GC and GC T_FH cells (based on gene expression analyses), as transcript levels of MAF, BCL6, IL-21 were lower in pT_FH than tonsillar T_FH cells[77]. While Boswell et al. suggested that the pT_FH subset they identified was predominantly CXCR3⁻CCR6^high, Locci et al. did not incorporate CCR6 in their phenotypic panel[26, 77]. Yet a separate study using both mouse and human specimens found high CCR6 expression in the PD-1^hiCCR7^lo subset of CXCR5⁺CD4 pT_FH; this subset, but not PD-1^loCCR7^hi reportedly indicates active GC T_FH differentiation[28]. Further studies are required to reconcile the various surface markers used to derive the conclusions described above.

Interestingly, in vivo functional relevance was shown for each of the particular pT_FH subsets identified. Both subsets of activated pT_FH (PD-1⁺ICOS⁺ and PD-1⁺CCR7^lo) cells were associated with Ab responses to influenza vaccination[28, 30] and autoimmune Ab production[28]. In co-culture assays, PD-1^hiCCR7^lo pT_FH potently induced plasmablast and PC differentiation as well as total antigen-specific IgG production to Influenza[28, 30] and doublestranded DNA[28]. Similarly, the ratio of Th1-, 2-, and 17-like pT_FH subsets were reportedly skewed in patients with dermatomyositis, an autoantibody-mediated autoimmune disease, compared to healthy controls. This resulted in increased frequency of naïve B cell helper vs. non-helper pT_FH cells that further correlated with disease severity and circulating plasmablasts[29].

Establishing a significant association between HIV-specific NAb and bNAb development, however, is much more convoluted. On one hand, higher frequencies of quiescent pT_FH (PD-1⁺CXCR3⁻CXCR5⁺CD4⁺)[26] and PD-1⁺CD4 T cells[12] were described in HIV-infected donors exhibiting broad and potent serum neutralization activity[12, 26], while no association was observed between pT_FH frequency (irrespective of phenotype) and HIV Env-specific Ab titers, total IgG levels, or HIV-specific serum neutralizing activity in HIV-infected individuals exhibiting normal serum neutralization activity[77]. The difference in breadth and potency between the donors used in these studies may contribute to the discordant results. However, similarly discordant associations exist even in analyses using only samples from HIV-infected donors with broad and potent neutralizing activity, further complicating the situation. Indeed, while Mikell et al. found higher frequencies of PD-1⁺CD4 T cells in these patients[12], subsequent analyses failed to corroborate this association using a slightly further delineated PD-1⁺ pT_FH subset (PD-1^hiICOS⁺CXCR5⁺CD4⁺)[26]. The complex nature of bNAb development is confounded by classification of “good neutralizers” and subjects with broad and potent neutralizing serum activity[12, 26, 77]. Indeed, while all HIV-infected individuals develop patient-specific NAb, only a minority develops broad and potent neutralization activity[73]. Furthermore, an important caveat to consider is that broad and potent neutralizing activity appears in serum of chronically infected subjects, meaning bNAb development likely takes years to develop under normal conditions; thus, perhaps pT_FH:B cell dynamics should be monitored during the acute, rather than chronic, phase of HIV infection. In this regard, Locci et al. observed a correlation between the frequency of PD-1⁺CXCR3⁻CXCR5⁺ pT_FH cells at approximately 4 months post infection and the capacity of individuals to subsequently develop bNAbs during the chronic phase[26]. Although this association remained during the time of bNAb development months later, examining the pT_FH:B cell dynamics may elucidate early events that facilitate the ensuing development of bNAbs, and explain why this occurs in only a minority of individuals. While the evidence presented thus far is quite encouraging, the novelty of using pT_FH cells to track bNAb responses necessitates further experimentation to reconcile the different conclusions described. It may be possible, however, that all these subsets play varying roles in different facets of Ab responses. Indeed, recent evidence suggests that surface expression of certain T_FH markers is downregulated in memory T_FH cells after antigen clearance, while the T_FH lineage of these cells remains intact[78].

Collectively, most recent studies reinforce the notion that a circulating pool of memory T_FH cells exists in both humans and mice. Indeed, in an elegant study using both mice and human specimens, He et al. demonstrated that PD-1^hi(ICOS⁺CCR7^lo) pT_FH are activated cells that identify active GC T_FH differentiation[28]. Furthermore, these authors demonstrated that pT_FH cells are antigen-experienced precursors of GC T_FH cells, as pT_FH generation occurred independent of GC-forming factors (SAP) but still required T_FH differentiating factors (ICOS and Bcl6)[28]. While the relatedness of pT_FH and T_FH cells in lymphoid follicles was challenged by another study based on naïve B cell helper capacity, CCR6 expression, and gene expression profiles[77], it did not directly address the relationship between GC T_FH in LN and pT_FH from matched donors as was done by He et al. It may be possible that T_FH marker expression and B helper function is affected by phase of infection or nature of immunogen. Indeed, Bentebibel et al. suggested that different vaccination modalities and/or adjuvants might stimulate divergent subsets of pT_FH, skewing the Ab response to the vaccine[30]. Thus future studies investigating the pT_FH profiles induced by various previously tested vaccines would be informative in this regard. Importantly, by showing that CCR7^loPD-1^hi pT_FH are precursor T_FH cells with prominent helper capacity that circulate through nondraining LN, He et al. proposed that pT_FH cells can rapidly differentiate into mature T_FH cells upon antigen re-exposure to support GC formation and mediate B cell help, thus implying a mechanism by which T_FH cells exit the secondary LN and are poised to accelerate Ab responses in non-draining LN, likely to contain a spreading infectious agent[28]. These studies provide a better-defined phenotype with which to monitor the status and specificities of T_FH cells and provide critical support to using pT_FH to study the development of affinity matured Abs, such as bNAbs.

A key question that remains is the role of antigen persistence in vaccine responses and whether and how this contributes to divergent pT_FH phenotypes, function, Ab potency and duration. Moreover, it may be possible that all identified potential pT_FH characteristics are correct in different settings of diseases, vaccination or populations. Indeed, while antigen specificity of T_FH cells in bNAb development remains to be established[26], a possible explanation for the divergent results between in vivo HIV bNAb and influenza-specific Ab development is that the CXCR3⁺CXCR5⁺CD4⁺ T cell population induced by influenza vaccination may provision suboptimal B cell help by only promoting memory responses[29, 30], thus potentially explaining the low efficiency and duration of seasonal influenza vaccine campaigns. Future studies exploring the interplay between pT_FH cells, B cells and Ab production will likely provide insight into how to best design more effective vaccines.

Concluding remarks

Current research highlights the tight involvement of T_FH cells in the development of bNAbs during natural HIV/SIV infection despite evidence of dysregulated T_FH-mediated B cell help. Exploring pT_FH dynamics may facilitate the dissection of T_FH:B cell interactions that culminate in bNAb development during chronic infection, providing insight into how T_FH responses can be manipulated to optimize AID expression, SHM, and CSR and long-lived B cell responses. Knowledge of how to tailor T_FH responses through vaccination will inform vaccine development for HIV and others diseases.

Figure 2. HIV/SIV-mediated T_FH dysfunction and immunopathogenesis.

(a.) HIV/SIV mediated immune activation induces high IL-6 production found within infected lymph nodes. (b.) IL-6 induces expansion of (potentially dysfunctional) T_FH cells expressing high levels of Bcl6 and IL-6Rα. T_FH expansion is further associated with increased numbers of GC B cells. (c.) GC T_FH expansion increases contact with PD-L1-expressing GC B cells, resulting in deregulated GC T_FH cells and inadequate help provisioned to GC B cells, likely by lowering the selection threshold and decreasing IL-21 signaling (and other T_FH cytokines such as IFN γ and IL-10) that lead B cell differentiation into short-lived PC formation and increased polyclonal and HIV/SIV-specific (primarily targeting Gag) IgG production. (d.) Similarly, direct IL-6 signaling may also mediate spontaneous terminal differentiation of memory B cell into plasma cells, resulting in observed decrease of memory B cells in HIV/SIV infection. (e.) High antigenic availability within the lymph node likely also enhances plasma cell survival. (f.) High antigenemia likely also contributes to B cell exhaustion, apoptosis, and the subsequent aberrant B cell phenotypes. (g.) HIV/SIV infection of T_FH cells (enhanced by IL-6) likely maintains the viral reservoir, as infected T_FH cells, may be resistant to apoptosis; this likely constitutes a latent reservoir within a privileged tissue, as HIV/SIV-specific CD8 CTL relatively seldom enters the lymphoid tissue.

Highlights.

• -T follicular helper cells present a latent HIV viral reservoir
• -Tfh cells may be intricately involved in the generation of HIV-specific broadly neutralizing Abs
• -Insights into Tfh immunobiology may provide inroads into effective vaccine design
• -Peripheral Tfh cell counterparts may facilitate studies of Tfh in lymphoid tissues