T Follicular Helper Cell–Derived IL-4 Is Required for IgE Production during Intestinal Helminth Infection

The production of IgE is a fundamental feature of the type 2 immune response that contributes to protective immunity to helminth infection and venomous challenge, but it also exacerbates pathological allergic disease (1). IL-4 secretion by CD4⁺ T cells has been shown to be indispensable in IgE production by B cells that have matured into plasma cells (PCs; broadly referring to all Ab-secreting cells, including plasmablasts) (2, 3). Understanding the nature of T cell help required for IgE production and the cellular requirements that orchestrate IgE responses could provide important insight into mechanisms regulating type 2 immunity.

CD4⁺ Th2 cells that express the lineage-specific transcription factor Gata-3 are potent producers of type 2 cytokines, including IL-4, IL-5, and IL-13 at sites of inflammation and infection (4). Given their function and important contribution to anti-helminth immunity and allergic disease, these cells have been defined as the source of IL-4 that drives IgE production (5). However, IgE class switching is initiated in the secondary lymphoid organs (SLOs), a location where few IL-4–producing Th2 cells can be identified (6, 7). Additionally, we and others have shown that the dominant IL-4–producing cell in the inflamed lymph nodes of helminth-infected mice are Tfh cells, a subset of B cell lymphoma (Bcl)-6–expressing CD4⁺ T lymphocytes that migrate deep into the B cell follicle and promote germinal center (GC) formation and select high-affinity B cell clones for long-lived humoral immune responses (6, 8, 9). These studies bring into question the critical source of IL-4 that initiates the development of IgE-producing B cells during a type 2–dominated immune response to helminth infection.

T cell–dependent Ig class switching occurs at various locations within the SLOs, including the B cell follicle, GC, and interfollicular regions (10, 11). In addition to the production of IgE during type 2 immune responses, IgG1 class switching is also promoted by IL-4 signaling and is the dominant Ig produced in mice and humans (12). Although the generation of IgG1-producing PCs generally requires the GC reaction, the B cell differentiation program required for IgE-producing cells is less clear (13, 14). Recent studies using transgenic IgE reporter mice have identified IgE⁺ cells within the GC and the subsequent detection of affinity mature IgE suggests that this reaction contributes to the IgE-secreting PC compartment (7, 13, 15, 16). However, the kinetics of IgE⁺ GC cell formation and serum IgE detection during helminth infection parallels the detection of both IgE⁺ PCs in the medullary sinuses, a region relatively distant from GCs and B cell follicles (16). These results bring into question the requirement for the GC reaction in IgE⁺ PC development and suggest that extra-GC T cell–B cell interactions may play an important role in IgE production. Indeed, PC responses that occur independent of GCs not only represent a significant source of Ab-secreting cells, but they also depend on B cell lymphoma (Bcl)-6⁺ T cells (17). With the aim of understanding the cellular requirements for the primary IgE response, we hypothesized that Tfh cells provide a nonredundant source of IL-4 required to initiate IgE class switching during infection with the parasitic helminth Heligmosomoides polygyrus bakeri. Furthermore, we posited that early IgE production would occur independent of the GC reaction. Using methods to selectively delete Tfh cells and GC B cells in mice, we demonstrate that depleting the lymph node microenvironment of IL-4–producing Tfh, but not Th2, cells results in the impaired differentiation of IgE⁺ PCs and Ab production during primary enteric helminth infection. Additionally, this early IgE response was not compromised in the absence of GC B cells. Furthermore, we also find that Tfh cell–derived IL-4 promotes the maintenance of the ongoing Th2 response. These data indicate that Tfh cells are central regulators of the type 2 adaptive immune response in the reactive lymph nodes during intestinal helminth infection.

All experiments were approved by the McGill University Animal Care Committee, and mice were used at 6–16 wk of age. 4get/KN2, KN2/KN2, CD45.1⁺Tcrb^−/−, CD4.cre, Mb1.cre, and Bcl-6fl/fl mice on a C57BL/6 background were bred and kept under specific pathogen-free conditions at McGill University. CD4cre, Mb1cre, and Bcl-6fl/fl mice were provided by Dr. J. Madrenas (McGill University, Montreal, QC, Canada), Dr. M. Reth (Max Planck Institute, Freiburg, Germany), and Dr. T. Takemori (Riken Institute, Yokohama, Japan) (18), respectively. All animals were infected with 200 L3 (H. polygyrus bakeri) larvae by gavage and were euthanized at the time points indicated.

Mesenteric lymph node (mesLN) single-cell suspensions were processed and analyzed as previously described (19). Abs used included GL7 (GL7) and CD138 (281-2) from BD Biosciences, IL-5 (TRFK5) from BioLegend, IgE (23G3) from SouthernBiotech, and IgD (11-26c), CD44 (IM7), CD62L (MEL-14), human (hu)CD2 (RPA-2.10), Gata-3 (TWAJ), IgG1 (M1-14D12), IL-13 (eBio13A), and streptavidin–eFluor 450 from eBio/Affymetrix. Fixable viability dye (eBio/Affymetrix) or propidium iodide was used to exclude dead cells. Intracellular staining for IL-5, IL-13, IgG1, and IgE was performed as previously described (19). Data were acquired with either a FACSCanto II or LSRFortessa (BD Biosciences) and analyzed using FlowJo software.

MesLN cells (1.5 × 10⁶) were stimulated in complete RPMI 1640 in the presence of 50 ng/ml PMA, 1 μg/ml ionomycin, and 0.67 μl/ml BD GolgiStop for 4 h. After stimulation, cells were stained as described above. Total serum IgE was quantified using a mouse IgE ELISA Ready SET-Go! kit (eBio/Affymetrix).

Bone marrow cells from 4get/KN2.CD4cre-Bcl-6fl/fl mice were mixed with either 4get/KN2 or KN2/KN2 bone marrow cells at a 1:1 ratio and 5 × 10⁶ total cells were injected i.v. into lethally irradiated (900 rad split into two doses) CD45.1⁺Tcrb^−/− hosts to generate 4get/KN2.Tfh^WT or 4get/KN2.Tfh^ΔIl4 mice, respectively. Following 8 wk of hematopoietic reconstitution, chimeric mice were infected with H. polygyrus bakeri larvae and processed as described in Fig. 6.

Data are expressed as mean ± SD. Data were analyzed by a two-tailed Student t test or one-way ANOVA as appropriate using the GraphPad Prism program (version 6). A p value <0.05 was considered significant.

Infection of mice with the enteric nematode H. polygyrus bakeri induces a robust type 2 response characterized by an early IgE⁺ extrafollicular PC and GC response within the mesLNs that drain the active site of infection and contributes to protective immunity upon reinfection (20, 21). Consistent with prior studies, H. polygyrus bakeri infection led to the induction of a distinct B220^lowCD138⁺ PC population within the mesLNs at 2 wk postinfection (Fig. 1A) (22). Further analysis of the PC population confirmed that most plasmablasts had class switched by this time point. Although most of the PCs were IgG1⁺, a small, but distinct population of IgG1⁻IgE⁺ cells was readily detectable at this time point (Fig. 1A). To determine the optimal time point to investigate the IgE response, the kinetics of PC development were assessed over time. Both IgG1⁺ and IgE⁺ PCs peaked at 2 wk postinfection and paralleled serum IgE titers (Fig. 1B–D). This kinetic is consistent with previous reports of infection with Nippostrongylus brasiliensis, another parasitic helminth, that demonstrated an IgE response that rapidly contracted (13, 14, 16).

As IL-4 producing CD4⁺ T cells are essential for the production of IgE, we hypothesized that the kinetics of IL-4 production in the mesLN would coincide with that of the PC response (3). Using IL-4 dual reporter mice, where GFP denotes cells actively transcribing IL-4 and membrane huCD2 indicating cells actively secreting IL-4 protein, we found that the number of IL-4–expressing and –producing CD4⁺ T cells peaked at 2 wk postinfection and subsided thereafter (Fig. 1E) (23). To confirm that IL-4 is required for this early IgE⁺ PC response, IL-4–deficient (KN2/KN2) mice were infected with H. polygyrus bakeri and the PC response was assessed at 2 wk postinfection. Not only was the number of total plasmablasts decreased, but the IgG1 response was significantly impaired in the absence of IL-4 (Fig. 1F–H). Indeed, the number of IgE⁺ PCs was also significantly attenuated (Fig. 1I). Consistent with previous reports (24), serum IgE was also dramatically diminished (or undetectable in three out of eight animals) in H. polygyrus bakeri–infected IL-4–deficient animals (Fig. 1J). Taken together, these results indicate that the peak of the IgE⁺ PC response within the reactive mesLNs during H. polygyrus bakeri infection occurs in tandem with and is dependent on IL-4 production.

Previous studies indicate that the dominant producers of IL-4 in the mesLNs of H. polygyrus bakeri–infected mice are Tfh cells (6). Additionally, the kinetics of IL-4–producing Tfh cell development matched the PC response, and CXCR5⁺Bcl-6⁺ Tfh cells produced significantly more IL-4 compared with CXCR5⁻Gata-3⁺ Th2 cells in the mesLN during H. polygyrus bakeri infection at all time points examined (Fig. 2A). Although previous studies indicate that IL-4–expressing Th2 cells can give rise to Tfh cells during helminth infection (8), Tfh cells uniquely express high levels of Bcl-6 and have been shown to require this transcriptional repressor for their development (25–27). To test whether Tfh cell differentiation requires Bcl-6 during a type 2–dominated immune response to helminth infection, we generated 4get/KN2 mice with a selective deletion of Bcl-6 in T lymphocytes (4get/KN2.CD4cre-Bcl-6fl/fl; 4get/KN2.T^ΔBcl6 mice) and compared these animals with their Cre⁻ littermate controls (4get/KN2.T^WT). Consistent with previously published reports, CXCR5⁺CD4⁺ T cells produced the most IL-4 (6), and Bcl-6 deletion resulted in a complete loss of CXCR5⁺ cells expressing either PD-1 or huCD2 (Fig. 2B–D). Therefore, Bcl-6 is required for the generation of Tfh cells in the mesLN during helminth infection.

The transcription factor Gata-3 is highly expressed by Th2 cells and required for their differentiation and production of type 2 cytokines (4, 28, 29). Previous studies have implicated Bcl-6 in limiting the development of this effector lineage in the context of immunization and models of allergy (4, 28, 29). To assess the intrinsic role of Bcl-6 in Th2 cell development during H. polygyrus bakeri infection, the number of mesLN CD4⁺CD44⁺Gata-3⁺ T cells was assessed in H. polygyrus bakeri–infected 4get/KN2.T^ΔBcl6 mice and littermate controls. H. polygyrus bakeri infection induced a robust and similar expansion of Th2 cells at 2 wk postinfection in both groups of animals (Fig. 3A, 3B). Whereas Th2 development was not compromised after intrinsic deletion of Bcl-6 in CD4⁺ T cells, their ability to execute their effector function and produce cytokines could still be affected. However, on a per cell basis, Bcl-6 deletion did not alter the ability of Gata-3⁺CD44⁺CD4⁺ Th2 cells to produce either IL-5 or IL-13 after restimulation (Fig. 3C, 3D). Similarly, ex vivo analysis demonstrated that the frequency of IL-4–producing huCD2⁺Gata-3⁺ Th2 cells was not compromised in the absence of T cell–expressed Bcl-6 (Fig. 3C, 3D). In contrast, the Gata-3⁻huCD2⁺ population, confirmed to be Tfh cells by their elevated expression of CXCR5 and Bcl-6 (Fig. 3E), was severely diminished as previously noted (Figs. 2, 3F). Taken together, these data demonstrate that Bcl-6 deletion in CD4⁺ T cells does not compromise the generation of cytokine-producing Th2 cells during helminth infection.

The ability to produce IgE has been classically attributed to IL-4–producing Th2 cells (5). However, given that the IgE⁺ PC response is initiated in the SLOs and that Tfh cells are the dominant source of IL-4 in the mesLN following H. polygyrus bakeri infection, we hypothesized that Tfh, not Th2, cells would be the primary source of IL-4 that drives the IgE PC response. Consistent with the importance of Tfh cells in GC formation, the number of IgD⁻GL7⁺ GC B cells was significantly decreased in 4get/KN2.T^ΔBcl6 mice compared with Cre⁻ controls (Fig. 4A, 4B). Upon assessment of the PC response, the frequency and cell counts of IgG1⁺ PCs were similar between groups, accounting for the minimal difference in total PC numbers (Fig. 4C–E). In contrast, class switching to IgE was significantly compromised as exhibited by fewer IgE⁺ PCs in 4get/KN2.T^ΔBcl6 mice compared to littermate controls (Fig. 4D, 4F). Consistent with the defective IgE⁺ PC response, serum IgE titers were also significantly diminished in the absence of Tfh cells (Fig. 4G). Collectively, our results indicate that even in the setting of robust Th2 cell differentiation, Tfh cells are critical for driving the primary IgE response during helminth infection.

The observed loss of IgE⁺ PC generation in 4get/KN2.T^ΔBcl6 mice may be due to a compromised GC reaction, an indirect consequence of preventing Tfh cell differentiation. However, most IgE-producing PCs have been previously described to be located within extrafollicular medullary regions (16). Additionally, Bcl-6⁺ T effector cells are known to promote both GC and extrafollicular-derived Ab responses (17, 25–27). To determine whether the GC response is required for initiating IgE⁺ B cell class switching and PC formation, we crossed 4get/KN2 mice with animals lacking Bcl-6 in the B cell compartment (Mb1cre-Bcl-6fl/fl; referred to in this study as 4get/KN2.B^ΔBcl6 mice). Consistent with a previous immunization study, 4get/KN2.B^ΔBcl6 mice failed to generate a detectable GC B cell response following H. polygyrus bakeri infection (18) (Fig. 5A, 5B) compared with their Cre⁻ littermate controls (4get/KN2.B^WT). Compared to 4get/KN2.T^ΔBcl6 mice, 4get/KN2.B^ΔBcl6 animals generated CXCR5⁺ IL-4–producing Tfh cells, albeit to a lesser extent than did controls, whereas IL-4 production by Th2 cells was uncompromised (Fig. 5C, 5D). Additionally, loss of Bcl-6 in the B cell compartment resulted in a significant decrease in the number of total CD138⁺ PCs (Fig. 5E). However, intracellular staining of PCs from H. polygyrus bakeri–infected 4get/KN2.B^ΔBcl6 mice revealed a change in the partitioning of the PC population compared with control mice whereby the frequency of IgG1⁺ PCs decreased whereas IgE⁺ PCs increased (Fig. 5F). Consistently, the number of class-switched IgG1⁺ PCs and serum IgG1 levels were both decreased in mice lacking expression of Bcl-6 in the B cell compartment (Fig. 5G, 5H). Whereas the frequency of IgG1⁻IgE⁺ cells was significantly increased, there was no difference in the total number of class-switched IgE⁺ PCs nor serum IgE at this point during infection (Fig. 5I–K). These results indicate that, as opposed to the IgG1 response, initiation of the IgE⁺ PC response does not require the GC stage of B cell differentiation.

Our results indicated the importance of Tfh cells for initiating IgE⁺ PC differentiation during H. polygyrus bakeri infection. However, our studies thus far did not discern the relative importance of IL-4 produced by Tfh cells versus Tfh cell differentiation itself for the humoral immune response. Furthermore, we have previously shown that IL-4R signaling is required to maintain, but not initiate, the Th2 response during H. polygyrus bakeri infection (6). Therefore, we were surprised to find that eliminating the dominant source of IL-4 (i.e., Tfh cells) had no effect on Th2 cell numbers (Fig. 3). To provide insight into these questions, bone marrow chimeras were generated to selectively eliminate IL-4 production by Tfh cells. To this end, 4get/KN2.T^ΔBcl6 bone marrow cells were mixed at a 1:1 ratio with either 4get/KN2 or KN2/KN2 cells and transferred to lethally irradiated congenic Tcrb^−/− hosts to generate either 4get/KN2.Tfh^WT mice or 4get/KN2.Tfh^ΔIl4 mice (Fig. 6A). H. polygyrus bakeri infection of both groups of mice led to the generation of a similar number of Tfh cells (Fig. 6B). In contrast, a relative decrease in the number of Th2 cells (as determined by a CD4⁺CD44⁺Gata-3⁺ phenotype) was observed in 4get/KN2.Tfh^ΔIl4 mice compared with 4get/KN2.Tfh^WT mice (Fig. 6C). In addition to diminished CD4⁺ effector T cell responses, the total number of PCs was also significantly decreased in 4get/KN2.Tfh^ΔIl4 mice compared with control mice (Fig. 6D). Further analysis of the PC population confirmed limited class switching to both IgG1 and IgE isotypes (Fig. 6E–G). Consistent with these data, IgE Ab titers were significantly lower or undetectable in the absence of Tfh cell–derived IL-4 (Fig. 6H).

The Th2 cell has been defined as the relevant source of IL-4 that mediates class switching to the IgE isotype (5). However, in vivo evidence for this paradigm of humoral immunity is circumstantial and, given the discovery of Tfh cells as an important source of IL-4, is now being revisited (6). Nevertheless, the role of the Tfh cell in IgE production has remained controversial. For example, it has been reported that IgE production is uncompromised following helminth infection of mice bearing IL-4–deficient CXCR5⁺ T cells, suggesting that Th2 cell–derived IL-4 is sufficient for IgE production (30). However, CXCR5 deficiency in CD4⁺ T cells does not impair initial T cell–B cell interactions, PC formation, or Bcl-6 upregulation by activated CD4⁺ T cells (31, 32). In contrast, recent studies using a model of airway allergy demonstrated impaired serum IgE titers in Tfh, but not Th2, cell–deficient hosts (29, 33). Using methods to definitively ablate Tfh cell differentiation or IL-4 specifically derived from Tfh cells, our results support and extend the latter findings by demonstrating that, in a setting of natural parasitic helminth infection, Tfh cells provide a critical source of IL-4 to support IgE PC development as well as orchestrate a robust Th2 response in the inflamed lymph nodes.

As Bcl-6 has been shown to be intrinsically required for the generation of both Tfh and GC B cells, we compared mice lacking this transcriptional repressor in either the T or B cell compartments to delineate the relative importance of these lymphocyte subsets in the IgE⁺ PC response during helminth infection. Our results indicated that Bcl-6–deficient T cells completely failed to generate IL-4–producing CXCR5⁺ Tfh cells, support GC B cell formation, or elicit IgE production. However, an early IgG1⁺ PC response was evident, indicating that Tfh cells are indispensable for the IgE response. In comparison, mice containing Bcl-6–deficient B cells mounted a significantly diminished GC and IgG1⁺ PC response, but they developed IgE⁺ PCs to a similar extent as did control animals. The latter result is consistent with previous studies describing IgE⁺ cells arising early during H. polygyrus bakeri infection in extra-GC regions of the draining mesLN as well as rapidly accumulating in the lymph node medullary regions following immunization (16, 20). Although our results contrast with prior in vitro studies demonstrating that Bcl-6 directly represses IgE production by B cells, the in vivo requirements to generate extra-GC IgE⁺ PCs may not be significantly influenced by this transcriptional repressor (34). Given the fact that substantial populations of IL-4–producing Tfh cells were still generated during H. polygyrus bakeri infection of GC-less (4get/KN2.B^ΔBcl6) animals suggests they were likely sufficient for supporting IgE production in these animals. Combining these results with our data showing that germline deletion of IL-4 compromised either the generation and/or survival of both IgG1 and IgE responses suggests that the IgG1 response may require a lower magnitude of IL-4 signals (e.g., Th2 cell–derived) than the IgE response that relies on abundant Tfh cell–derived IL-4. Consistently, induction of germline Cε transcription is more sensitive to the loss of IL-4 signaling than Cγ1, a locus that can be activated by other cytokines (35). For example, IL-21 derived from CD4⁺ICOS⁺CXCR4⁺ effector T cells that localize to extrafollicular regions may shift the balance to promote IgG1⁺ PC differentiation when IL-4 signals are limiting (36). Another possibility for the decrease of IgG1⁺ PCs in H. polygyrus bakeri–infected GC-less mice may be that deletion of Bcl-6, an antiapoptotic factor, in B lymphocytes may lead to premature death of IgG1 class-switched B cells, an event that may not be relevant for IgE⁺ PCs, which rarely initiate GC differentiation and exhibit a shorter half-life (7, 13, 14, 37). Although we have not examined the affinity maturation status of these GC-independent IgE⁺ PCs, numerous studies indicate that these are likely low-affinity clones derived from a direct μ to ε class-switching event (13, 15, 16). One potential role for these cells is that some low-affinity IgE⁺ PCs would remain in the SLOs and contribute to the secondary GC reaction upon reinfection where they would undergo additional rounds of selection to become high-affinity and long-lived IgE-secreting PCs (38). This would not be inconceivable considering the high levels of BCRs uniquely retained on these cells (13, 16).

Previous reports have suggested that Bcl-6 limits type 2 immune responses (39, 40). In these experiments, germline deletion of Bcl-6 led to a fatal inflammatory disease with hallmarks of type 2 inflammation, including elevated Th2 cytokines, tissue eosinophilia, and an increase in IgE-bearing cells. However, whether this phenotype was due to an intrinsic dysregulation in the CD4⁺ T cell compartment was not determined. Subsequent studies using mice with a selective deletion of Bcl-6 in the T cell compartment produced less IL-4 after immunization with a complex Ag (28). Using the same mouse line in the context of a type 2–biased allergic airway model, Kobayashi et al. (29) found that Th2 cell differentiation was similar to controls in the lung, but showed enhanced Th2 cytokine production in the lung-draining lymph nodes of sensitized mice. In contrast, we found no difference in Th2 cell differentiation in Bcl-6–deficient T cells from the gut-draining lymph nodes compared with controls during H. polygyrus bakeri infection. Collectively, these results suggest that the impact of Bcl-6 on Th2 cell differentiation is likely context and potentially tissue specific. Determining the additional factors that regulate Th2 cell differentiation in these diverse settings is an important area of future investigation.

Given that the primary IgE response, but not Th2 cell differentiation or early IgG1 switching, was impaired in H. polygyrus bakeri–infected 4get/KN2.T^ΔBcl6 mice, we hypothesized that selectively abrogating IL-4 production by Tfh cells, but not Tfh cell differentiation itself, would selectively compromise the primary IgE response. However, depleting the lymph node microenvironment of Tfh cell–derived IL-4 but maintaining Tfh cell differentiation impaired both IgG1 and IgE responses. Interestingly, abrogating IL-4 production by Tfh cells also led to a compromised Th2 response. Consistent with this result, Tfh cells have been shown to contribute to the Th2 response upon secondary challenge in a house dust mite model of allergic inflammation, and IL-4 is required for the maintenance of the Th2 cell population (6, 41). Combining these results with our studies involving 4get/KN2.T^ΔBcl6 mice where IgG1 class switching and Th2 responses were intact despite a loss of IL-4–producing Tfh cells and GC B cells, we speculate that the substantial population of GC B cells generated following H. polygyrus bakeri infection in mice that lack IL-4 production by Tfh cells (unpublished observations) may compete with Th2 cells for IL-4 signals when this cytokine is limiting, thereby compromising Th2 development or maintenance (30). Indeed, B cells undergo an IL-4–dependent expansion in the mesLN of H. polygyrus bakeri–infected mice and require STAT6 signaling for their differentiation into proliferating GC cells (30, 42). Although our results indicate that Tfh cells are the dominant source of IL-4 in the draining lymph node of H. polygyrus bakeri–infected mice and are necessary for IgE production, our studies do not definitively rule out a contribution of Th2 cell–derived IL-4 in this response and therefore warrant further investigation. In summary, we propose that under physiological conditions, IL-4 production by Tfh cells in the reactive lymph node provides IL-4R signals sufficient to elicit IgG1 and IgE class switching, GC formation, Ab production, and maintenance and/or amplification of the Th2 response, all components of a protective humoral immune response to intestinal helminth infection. Identifying Tfh cells as a central orchestrator of an adaptive immune response in the reactive lymph node may provide an important point of intervention for controlling type 2–mediated immunity in health and disease.

We thank Camille Stegen for management of the MIMM flow cytometry facility and Dr. M. Reth (Max Planck Institute, Freiburg, Germany) and Dr. T. Takemori (RIKEN Center for Integrative Medical Sciences, Yokohama, Japan) for providing the Mb1-cre and Bcl-6fl/fl mice, respectively. We also acknowledge members of the King laboratory for their technical input.

The authors have no financial conflicts of interest.