TFH cells progressively differentiate to regulate the germinal center response.

Germinal center (GC) B cells undergo affinity selection, which depends on interactions with CD4(+) follicular helper T cells (TFH cells). We found that TFH cells progressed through transcriptionally and functionally distinct stages and provided differential signals for GC regulation. They initially localized proximally to mutating B cells, secreted interleukin 21 (IL-21), induced expression of the transcription factor Bcl-6 and selected high-affinity B cell clones. As the GC response evolved, TFH cells extinguished IL-21 production and switched to IL-4 production, showed robust expression of the co-stimulatory molecule CD40L, and promoted the development of antibody-secreting B cells via upregulation of the transcription factor Blimp-1. Thus, TFH cells in the B cell follicle progressively differentiate through stages of localization, cytokine production and surface ligand expression to 'fine tune' the GC reaction.

High-affinity antibodies and long-lived memory B cells are the hallmarks of the humoral response. Activated B cells undergo affinity maturation and differentiation in the germinal center (GC), dependent upon signals provided by CD4+ follicular helper T (TFH) cells[1], including interleukin 21 (IL-21) and costimulatory molecules such as CD40L (CD40 ligand) [2-5]. The signals provided by TFH cells include cytokines shared by other TH cell subsets, such as IL-4 and interferon-γ (IFN-γ), which promote B cell isotype switching appropriate to pathogen challenge [3,6-8]. TFH cell-derived IL-21 is a key regulator of the GC as, in its absence, B cells display defects in affinity maturation and generation of long-lived plasma cells [4,5]. IL-4 also promotes the GC response as mice deficient in this cytokine or its high affinity receptor IL-4Rα have compromised immunoglobulin IgG1 and IgE responses [7,9,10], and its deletion results in defective GC B cell expansion [7]. IL-4 secretion, together with CD40-CD40L signaling, enables TFH cells to induce the enzyme activation-induced cytidine deaminase (AID) in B cells, necessary for class switch recombination (CSR) and Ig affinity maturation [6,11]. The interplay of IL-21 and IL-4 signals shapes the humoral response, with IL-21-deficiency in mice resulting in increased IL-4-driven IgE switching, with their combined deficiency leading to an impairment in GC formation and antibody responses that exceeds that of either alone [12,13].

Interactive engagement between TFH cells and GC B cells entails repeated short-lived cellular contacts [14]. Chronological accumulation of T cell-derived signals results in the development of B cells expressing high affinity Ig receptors [15], and their differentiation into antibody secreting cells (ASCs) [16]. Conversely, repetitive cognate T-GC B cell interactions result in TCR-dependent changes in Ca+ and in cytokine expression in T cells [17], with B cell-derived ICOS signals promoting proper positioning of TFH cells within the B cell follicle and GC [18] and upregulation of CD40L on TFH cells [19], necessary for GC B cell selection [20].

Here we show that as a consequence of T-B cell interactions, TFH cell function evolved during the GC response, with these changes critical for B cell maturation. TFH cells differentiated from an IL-21+ TFH population observed proximally to the GC dark zone, the site of Ig gene hypermutation, early after immune challenge to an IL-4+ TFH cell population robustly expressing CD40L that developed later and resided more distal to the dark zone. Modulation of the TFH cell phenotype within the GC was dependent upon cell division and occurred in concert with alterations in gene expression. These distinct TFH cell populations were responsible for unique effects on B cell maturation, with the IL-21+ TFH cells enabling selection of high-affinity clones and IL-4+ TFH cells facilitating differentiation of antibody-secreting plasma cells. Thus, after entering the GC, TFH cells undergo progressive maturation to regulate GC B cell differentiation.

RESULTS

IL-4 and IL-21 expression define three populations of TFH cells

Disruption of signaling by either IL-21 or IL-4 results in defective humoral responses [4,5,7,12,21]. The non-redundant functions of IL-21 or IL-4 [22] suggest that TFH cells producing these cytokines are discrete, differing in their ability to regulate GC B cells. To explore this possibility, we generated C57BL/6 (B6) bicistronic Il21-IRES-Katuska (Kat) reporter mice (Il21Kat/+) to investigate the kinetics of IL-21 expression in TFH cells[17]. We immunized IL-21Kat/+ mice intraperitoneally (i.p.) with sheep red blood cells (SRBCs)[23]. The percentage of TFH cells expanded following immunization (), with Il21-Kat expression largely restricted to the follicle and absent from sites of nascent TFH cell-DC interactions [24] in the T cell zone (), suggesting that initial cytokine expression depended upon B cell interactions. The number of Il21-Kat+ TFH cells within the GC increased through day 12 (), indicating that TFH cells acquire expression of IL-21 after entering the GC or that a Il21-Kat+ TFH cell population entered the GC and displaced the Il21-Kat− TFH cells.

We next crossed Il21Kat/+ to Il4-IRES-GFP (Il4GFP/+; 4get) mice [25] to obtain Il21Kat/+Il4GFP/+ double reporter mice[17] and examined Il21-Kat and Il4-GFP expression during TFH cell development. We subcutaneously (s.c.) infected Il21Kat/+Il4GFP/+ mice with the helminth Nippostrongylus brasiliensis, which induces IL-4 production in TFH cells [26]. We assessed Il21-Kat and Il4-GFP expression in splenic CD4+CD44hiCXCR5hiPD-1hi TFH and CD4+CD44hiCXCR5loPD-1lo TH2 cells, as well as in a CD4+CD44hiCXCR5intPD-1int population (), a mixture of central memory T cells, TH2 and pre-TFH cells [27]. Three days post-parasite challenge, we detected small numbers of splenic Il21-Kat+, Il21-Kat+Il4-GFP+ double positive and Il4-GFP+ TFH cells (), and a fourth population that was Il21-Kat−Il4-GFP− double negative. From days 3 to 5, the total number of TFH cells rose, driven by expansion of the Il21-Kat+ and Il21-Kat+Il4-GFP+ pools (). From day 5, Il21-Kat+ TFH cells decreased, while Il4-GFP+ TFH cells increased, becoming dominant by day 15, at which point the Il21-Kat+Il4-GFP+ double positive population had also receded (). All three populations were present in the GC at days 7-8 (). By contrast, beginning at day 3, and extending to day 15 post-infection, TH2 cells were Il21-Kat−Il4-GFP+ as expected (). The latter cells arose by day 3 and showed a gradual loss of Il21-Kat+ cells by day 15.

The immune response to N. brasiliensis begins in lymph nodes (LNs) of the mediastinum, followed by those in the mesentery, and then the spleen [28]. In the mediastinal LNs of Il21Kat/+Il4GFP/+ infected mice, the earliest TFH cells appeared at day 2 and were mostly Il21-Kat+; by days 3 and 5 most TFH cells were Il21-Kat+Il4-GFP+ double positive (). In the mesenteric nodes of infected Il21Kat/+Il4GFP/+ mice, TFH cells were Il21-Kat+ by day 3, followed by a shift towards Il4-GFP+ TFH cells by day 8 (), analogous to the spleen. Because we could detect GC B cells in the mediastinal and mesenteric LNs of uninfected mice (), presumably due to environmental antigen exposure [29], we focused the remainder of our experiments on splenic responses to minimize background.

We next assessed the cell division-dependent expression of Il21 and Il4 following transfer of CellTrace Violet® dye labeled ovalbumin (OVA)-specific Thy1.2+CD4+OT-II TCR transgenic T cells from Il21Kat/+Il4GFP/+ mice into congenic Thy1.1+ recipient mice. One day following cell transfer recipients were s.c. infected with N. brasiliensis combined with 4-hydroxy-3-nitrophenylacetyl-OVA (NP-OVA), followed by a single intravenous (i.v.) injection of NP-OVA two days post-infection, to ensure Ag persistence and enable tracking of Ag-specific T and B cells. Il21-Kat expression in the Thy1.1+CD4+CD44hiCXCR5hiPD-1hi transferred TFH cells was detected by the third division, while Il4-GFP expression was seen beginning at the fifth division, and occurred in cells that were also Il21-Kat+ (). Three days following N. brasiliensis plus NP-OVA injection we found Il21-Kat expressing Thy1.1+CD4+CD44hiCXCR5hiPD-1hi TFH cells without significant expression of Il4-GFP (. These results suggest that cytokine transcript expression in TFH cells occurred in a progressive manner, with Il21-Kat appearing first at day 3, followed by Il4-GFP five days post-infection, after which Il21-Kat expression was gradually lost with Il4-GFP expression maintained to day 15.

We next sought to address the secretion of IL-4 and IL-21 proteins in reporter expressing TFH cells following N. brasiliensis infection. Although we detected three TFH cell populations expressing Il4 and Il21 mRNA between days 5 and 8 during our initial time-course experiment, intracellular cytokine staining after ex vivo stimulation with phorbol 12-myristate 13-acetate and ionomycin at these time points indicated that TFH cells primarily produced either IL-4 or IL-21 (). Similar observations were made after i.p. immunization of wild type mice with NP-keyhole limpet hemocyanin (NP-KLH) in alum (). Using a dual-color ELISPOT assay we detected IL-21 or IL-4 protein secretion by individual splenic TFH cells isolated from Il21Kat/+Il4GFP/+ mice 8 days following N. brasiliensis infection (). While ELISPOT assays only detected a relatively small number of IL-21+ and IL-4+ TFH cells, as noted by others [30], analysis of Il21-Kat+, Il21-Kat+Il4-GFP+, and Il4-GFP+ sorted splenic TFH cells revealed a strong correlation between reporter expression and cytokine secretion, with Il21-Kat+ TFH cells producing IL-21 and Il4-GFP+ TFH cells producing IL-4, substantiating fidelity of the reporters, with the majority of TFH cells secreting only one or the other cytokine (). By contrast, relatively few TFH cells co-secreted IL-21 and IL-4. Thus, we designated the Il21-Kat+ cells as TFH21 and Il4-GFP+ cells as TFH4 cells. We found that the Il21-Kat+ Il4-GFP+ double positive TFH cells produced either IL-21 or IL-4 in ELISPOTs, rarely both, suggesting that secretion of these cytokines by TFH cells is largely exclusive (); thus, we designated the dual-reporter expressing Il21-Kat+Il4-GFP+ TFH cells as TFH21+4. Collectively, these results supported the idea that cytokine expression and secretion in TFH cells change over the course of an infection.

TFH21 and TFH4 cell populations are transcriptionally distinct

To determine if the three cytokine-expressing CD4+CD44hiCXCR5hiPD-1hi TFH cell populations were transcriptionally distinct, we infected Il21Kat/+Il4GFP/+ mice with N. brasiliensis and sorted splenic Il21-Kat+, Il21-Kat+Il4-GFP+, and Il4-GFP+ TFH cells, as well as Il4-GFP+ TH2 cells at day 8 post infection (). Using RNA sequencing, a total of six pairwise comparisons were made between the aforementioned four populations, resulting in 1300 genes that were differentially expressed (FDR q-value < 0.05) in at least one of the pairwise comparisons, with TH2 cells the most different compared to each population of TFH cells (). The TFH21 and TFH21+4 groups were most alike, while TFH4 cells shared similarities with the TFH21+4 and the TH2 populations (). Clustering analysis assigned each differentially expressed gene to one of four groups (), representing modules of Il21-coregulated genes, Il4-coregulated genes, TFH-defining genes including Sh2d1a and Batf, and non-TFH effector genes such as Il2ra and Tbx21. To analyze the transcriptional differences among the three TFH cell populations, we examined a curated set of genes previously described to be up- or down-regulated in TFH cells compared to other CD4+ TH subsets [31-33]. The TFH cell-associated transcription factor genes Batf and Ascl2 [33,34] were upregulated in TFH21 and TFH21+4 cells compared to TFH4 cells, while expression of the chemokine receptor gene Cxcr4 was higher in TFH cells expressing Il21-Kat compared to other two TFH cell populations (). Expression of Cxcr5 and Pdcd1 was highest in TFH21+4 cells confirmed by flow cytometry staining of CXCR5 and PD-1 (). Thus, TFH21, TFH21+4 and TFH4 cells were transcriptionally distinct from one other.

TFH21 and TFH4 cells differentially localize in the GC

CXCR4 is necessary for B cell migration between the light and dark zones (LZ and DZ) of the GC [35], and differences in its expression could affect the ability of TFH cells to interact with centrocytes and centroblasts, respectively [20]. Consistent with our transcriptome analysis, at day 8 following N. brasiliensis infection, TFH4 cells had significantly less surface expression of CXCR4 than TFH21 and TFH21+4 cells (), in agreement with its upregulation by Ascl2 [34] in the latter two (). Based upon this observation, we tested whether the TFH cell populations localized to different regions of the GC. Staining of splenic sections of Il21Kat/+Il4GFP/+ mice with CD35 detected follicular dendritic cells that mark the GC LZ [36] (). When measuring the distance from the centers of the DZ to the TFH cells of each type in that GC we observed that TFH21 cells were located on average more proximal to the DZ than TFH4 cells, and that TFH21 cells migrated more efficiently than TFH4 cells to the CXCR4 ligand CXCL12 in the dark zone, consistent with their increased CXCR4 expression, while TFH21 and TFH4 migrated similarly to the control chemokine CXCL13 () [36]. In contrast, cell-surface expression of CD40L [37], which, in combination with IL-4, promotes IgG1 switching and regulates GC B cell differentiation via CD40 signaling [2,38], was significantly enhanced on TFH4 and TFH4+21 cells compared to TFH21 cells at day 8 after N. brasiliensis infection (). Thus, TFH4 cells expressed increased CD40L and localized more distal to the LZ of the GC than TFH21 cells.

TFH21 cells become IL-4 producers upon helminth challenge

We next used an adoptive-transfer model to assess the effect of TFH21, TFH21+4 and TFH4 cells on the humoral immune response. Previous studies have shown that following adoptive transfer, CD4+ T cells maintain expression of either Il21 or Il4 for 7 to 10 days [6,39]. TFH21, TFH4+21 or TFH4 cells were sorted from Thy1.2+ Il21Kat/+Il4GFP/+ mice 8 days following N. brasiliensis infection and 300,000 cells from each group were transferred into Thy1.1+ B6 recipients bearing an irrelevant TCR transgene specific for the GP66-77 epitope of lymphocytic choriomeningitis virus (LCMV; Smarta TCR transgenics; Stg). Recipients of cell transfers were then re-infected and spleens were harvested 13 days after the secondary infection. Following transfer and re-exposure to N. brasiliensis we examined cytokine reporter expression by flow cytometry and protein expression by ELISPOT assays in Thy1.2+CD4+CXCR5hiPD-1hi TFH cells (). At 13 days post-secondary infection approximately 15% of the transferred Il21-Kat+ TFH cells retained sole expression of this transcript, while about 50% acquired Il4-GFP expression (). Approximately 40% of the transferred Il21-Kat+Il4-GFP+ double positive TFH cells became Il4-GFP-expressing TFH4 cells, while about 10-15% retained expression of both Il21-Kat and Il4-GFP (). By contrast, virtually all of the transferred Il4-GFP+ TFH cells that were reporter positive were Il4-GFP+, with around 15% also upregulating II21-Kat (). Transferred Il21-Kat− Il4-GFP− double negative TFH cells did not acquire expression of either cytokine at the day 13 time-point (data not shown). Transferred Il4-GFP+ expressing cells were localized more distal to the DZ compared to the Il21-Kat+ cells (), consistent with their higher CXCR4 expression. Transferred TFH cell subsets maintained their surface phenotype post-transfer, albeit with fewer donor TFH4 than TFH21 and TFH21+4 cells in recipient spleens (). Approximately 10% of the transferred TH2 cells were able to acquire TFH cell markers, though this resulted in fewer total TFH cells than when TFH21 or TFH21+4 were transferred [17] (). In a like manner, donor TFH cells was maintained similar phenotypes when examined at 7 days following transfer ().

We sorted each of the three TFH cell populations at 13 days post transfer and determined their IL-21 and IL-4 production by dual color ELISPOT assays. Progeny of transferred Il21-Kat+ TFH cells contained both IL-21 and IL-4 secretors with rare cells secreting both (. Transferred Il21-Kat+Il4-GFP+ double positive TFH cells robustly became IL-4-secreting cells, while transferred Il4-GFP+ TFH cells retained the IL-4 single-positive phenotype (). There was a strong correlation between cytokine reporter expression and protein secretion in all transferred TFH cells, as well as a near-absence of cells secreting IL-21 and IL-4 simultaneously ().

To determine if TFH cells isolated at an early time point following pathogen challenge exhibited similar kinetics of cytokine expression, we modified our transfer model () to obtain Il21-Kat+ and Il21-Kat+Il4-GFP+ TFH cells from five days post-infection with N. brasiliensis instead of day 8. There were insufficient splenic TFH4 cells to transfer at this time point, precluding their analysis (). Il21-Kat+ and Il21-Kat+Il4-GFP+ TFH cells isolated from infected donors at day 5 maintained CXCR5 and PD-1 expression in recipients after adoptive transfer (). However, in contrast to Il21-Kat+ cells isolated 8 days after infection, which largely remained Il21-Kat+ after transfer, 40% of such cells isolated at day 5 post-infection converted into Il4-GFP+ expressers following transfer (). Collectively, these data indicate a model of progressive differentiation of TFH cells from Il21-Kat-expressing cells to Il4-GFP-expressing cells.

TFH21 cells and TFH4 cells differentially regulate GC responses

We next tested the idea that TFH cell populations with differential expression of IL-21, IL-4, CD40L and CXCR4 contribute to distinctive GC B cell responses. We transferred 300,000 TFH21, TFH4+21 or TFH4 cells sorted from Thy1.2+ Il21Kat/+Il4GFP/+ mice 8 days following N. brasiliensis infection into Thy1.1+ Stg B6 recipients, with the latter re-infected and spleens harvested 13 days after the secondary infection, a time point coinciding with the peak of the GC response. Mice that received TFH cells had a significant increase in the percentages and numbers of B220+IgD−CD95hiGL-7hi splenic GC B cells compared to N. brasiliensis-infected mice that were not recipients of adoptively transferred cells, with those receiving TFH21+4 cells showing the greatest increase in percentages and numbers of GC B cells and average GC size, compared to recipients of TFH21 or TFH4 cells (. This suggested that even though the TFH21+4 cell population consisted of TFH cells that secreted either IL-4 or IL-21, they functioned synergistically in promoting the GC response, as predicted by in vitro experiments [40].

Compared to TFH21 cells, adoptively transferred TFH21+4 and TFH4 cells promoted IgG1 synthesis by GC B cells, with transferred TFH21+4 cells inducing the greatest numbers of IgG1+ GC B cells (). Additionally, IgG1 production by GC B cells in mice into which we adoptively transferred IL-4−/− and IL-4+/+ TFH21 cells were similar (), suggesting that the minimal amount of IL-4 produced by TFH21 cells had a negligible effect upon IgG1 synthesis, as expected. Moreover, adoptively transferred TFH21+4 and TFH4 cells drove more splenic ASCs to secrete IgG1 and IgE than TFH21 cells (), consistent with their increased expression of IL-4 () and CD40L () [38,41]. TFH21+4 cells also promoted the development of more B220loCD138+ plasma cells than did TFH21 cells (). Using an ELISA specific for N. brasiliensis antigens, we found no difference in the amounts of serum IgM, but significantly higher amounts of IgG1 in recipients of TFH21+4 and TFH4 cells than in mice transferred with TFH21 cells, and a trend toward increased IgE production (. The differences in B cell and plasma cell phenotypes in recipient mice were not the result of enhanced proliferation or diminished B cell apoptosis, because BrdU uptake in GC B cells was similar among mice transferred will all the TFH cell subsets (), while caspase 3 activity was similar in mice that received TFH21 or TFH4 cells (.

To assess the quality of Ag-specific GC B cells generated by the TFH cell subsets, we adoptively transferred TFH21, TFH21+4 and TFH4 cells isolated from Thy1.2+Il21Kat/+Il4GFP/+ OT-II TCR transgenic mice at day 8 post N. brasiliensis infection into Stg B6 recipients. Upon transfer, these mice were infected with N. brasiliensis and serially administered NP-OVA (). Thirteen days post infection we then sequenced the Vh186.2 gene in splenic GC B cells isolated from host mice, in order to determine the occurrence of the high-affinity tryptophan to leucine point mutation at position 33 (W33L) of the VH186.2 Ig heavy chain [42]. Although GC B cells from all three recipient groups had a similar overall frequency of mutations in their Vh186.2 genes (), the W33L point mutation occurred significantly more often in GC B cells taken from recipients of TFH21 cells compared to TFH4 cells (), indicating that TFH21 cells positively promote selection, rather than merely somatic mutation. While transferred TFH21 cells promoted the formation of fewer low- as well as high-affinity anti-NP ASCs than did TFH21+4 and TFH4 cells (), this was likely due to impairment in CSR () rather than in affinity maturation. Conversely, recipients of TFH21+4 and TFH4 cells, which had fewer GC B cell clones with mutations that confer increased affinity for NP, had higher numbers of high-affinity anti-NP IgG1 and IgE ASCs than recipients of TFH21 cells (, a finding consistent with their enhanced promotion of IgG1- and IgE-secreting ASCs. Although transferred TFH21 cells had differentiated into TFH21+4 and TFH4 cells and began producing IL-4 over the 13 day course of the experiment (), they were not able to induce the phenotypic changes in GC B cells expected of IL-4+ TFH cells, such as class-switching to IgG1 (, potentially due to the insufficient or delayed onset of IL-4 production ().

In addition, GC B cells in mice transferred with TFH21 and TFH21+4 cells expressed significantly more Bcl-6 than those receiving TFH4 cells (). In contrast, expression of Prdm1, the gene that encodes Blimp-1, was higher in GC B cells from mice transferred with TFH21+4 and TFH4 cells compared to those that received TFH21 cells (), a finding consistent with the decrease in Ig secretion in ASCs isolated from the latter. Thus, GC B cells in mice that received TFH21 cells had higher affinity mutations, increased Bcl6 but decreased Ig secretion compared to mice, which received TFH21+4 and TFH4 cells.

DISCUSSION

We propose a model of TFH cell cytokine production in type 2 immune responses, in which discrete populations of IL-21+ TFH21 and IL-4+ TFH4 cells play unique roles in determining the antibody response -- B cell selection and antibody secretion, respectively, with TFH21 cells developing into TFH4 cells after exposure to antigen presented by GC B cells. In support of this model, previous work using Il21Kat/+Il4GFP/+ mice demonstrated that highly Ag-loaded GC B cells interact with TFH cells to promote TCR signaling and result in changes in TFH cell expression from Il21-Kat to Il4-GFP expression over the course of several hours [17]. Here we show that these transient TCR signals are mirrored by durable changes in transcriptional program, localization and surface ligand presentation that occur over several days.

One important difference between the TFH21 and TFH4 cell transcriptional programs seemed to be migratory regulation. GC B cells deficient in CXCR4 fail to localize to the DZ, resulting in decreased rates of somatic mutation compared to CXCR4-sufficient cells [36]. GC B cells that more robustly engage Ag, enabling its presentation to TFH cells, reside longer in the DZ than the LZ, acquiring more Ig gene mutations with increased Ag affinity [20]. Our results suggest that TFH cells were not evenly distributed within the GC; rather, those cells nearest the DZ were enriched for TFH21 cells, which efficiently select high-affinity mutant B cells. CXCR4hi centroblasts lose CXCR4 and migrate out into the light zone [43]. In a like manner, TFH4 cells that developed later in the course of the helminth-induced GC response had diminished CXCR4 expression with decreased ability to migrate efficiently to CXCL12.

Our results support the critical role of TFH-cell derived IL-21 in an effective humoral response [4,5,12]. In its absence, such as in mice receiving transfers of TFH4 cells, GC B cells expressed less Bcl-6, and generated fewer CD138+ plasma cells, while showing a trend in reduction of antigen-specific ASCs and serum helminth-specific IgG1 compared to recipients of TFH21+4 cells, in which both cytokines were present. These data demonstrated that TFH21 and TFH21+4 cells are essential sources of IL-21, which is required for the maintenance of Bcl-6 expression in GC B cells, GC longevity and production of plasma cells [4,5].

While IL-21 is necessary factor, TFH21 cells on their own were not fully capable of orchestrating a productive humoral response, concordant with observations that IL-21 expression is not sufficient for ASC formation [4,5,22]. TFH21 cells were not able to mobilize high levels of CD40L, a molecule that is critical for class-switching [44] and plasma cell differentiation [38], and also mediates Bcl-6 transcription [45,46]. Furthermore, compared to TFH4 cells, TFH21 cells failed to induce the expression of Prdm1 (Blimp-1) in GC B cells, which is critical for plasma cell development [2,44,47,48], and is antagonized by Bcl-6 [49]. GC B cells from recipients of TFH21+4 cells shared traits with those from the TFH21 group, such as more frequent high-affinity mutations and increased Bcl6 expression, and the TFH4 group, including increased IgG1 and Prdm1 expression, but had reduced apoptosis compared either group of single-cytokine producers, highlighting significant synergy between IL-21 and IL-4 in promoting GC responses [12,40].

We have shown that TFH cells mature phenotypically and transcriptionally over the course of the GC response. These cells exist as discrete subsets, which progressively differentiate with modulation of cytokine production [17] and of their localization and surface ligand expression to fine-tune the GC reaction.

METHODS

Mice

Mice were housed in pathogen-free conditions at the Yale School of Medicine (New Haven, CT). C57BL/6J (B6), C.129-Il4/J (4get), B6.PL-Thy1/CyJ (Thy1.1) B6.SJL-Ptprc/BoyJ (CD45.1), and B6.Cg-Tg(TcraTcrb)425Cbn/J (OT-II) B6.Il4/J animals were purchased from The Jackson Laboratory. B6.Tg(TcrLCMV)1Aox (Smarta; Stg) mice[50] were bred in-house. To examine the expression of Il21 in TFH cells, we generated B6 bicistronic Il21-IRES-Katuska (Kat)reporter (Il21Kat/+) mice [17]. All animals were used at 6-8 weeks of age, with approval for all procedures given by The Institutional Animal Care and Use Committee of Yale University.

N. brasiliensis Infection and Detection of Serum Antibodies

Mice were injected s.c. at the base of the tail with third-stage larvae of N. brasiliensis (500 viable third-stage larvae/mouse) in 0.2 mL sterile PBS with 1X gentamicin (Gibco). Animals were sacrificed at time points described in the text. Antibody ELISAs for detecting anti-L5 serum antibodies were performed as previously described [21]; briefly, stage 5 adult worms (L5) were isolated from intestines of mice infected with L3 larvae 7 days earlier and homogenized. Five micrograms of lysates were used to coat microtiter plates, which were then incubated with sera, followed by anti-mouse IgG1 and IgE secondary antibodies ().

Cell Transfers and Immunizations

2 × 105 TFH or TH2 cells, sorted by flow cytometry, or 5 × 105 CD4+ ovalbumin (OVA)-specific OTII TCR transgenic T cells, were transferred to recipient mice via retro-orbital injection. Cells were labeled with CellTrace™ Violet dye according to the protocol provided by the manufacturer (Life Technologies). In certain experiments, mice were immunized i.p. with 10% SRBC (Colorado Serum) () or 100μg NP14-KLH (Biosearch Technologies) emulsified in alum (). In others, mice were injected with 100 μg NP15-OVA (Biosearch Technologies) mixed with N. brasiliensis (400 third-stage larvae/mouse) s.c. 24 hours post-transfer of cells (). Mice subsequently received an additional 100μg NP15-OVA in PBS every 48 hours by i.v. injection for up to 10 days. Animals were sacrificed at different time points post infection and harvested spleens were processed for flow cytometry.

RNA and cDNA Synthesis

Sorted cell populations were processed for RNA isolation and conversion into cDNA as described previously [51]. RNA from sorted cells was extracted and purified with RNeasy Mini kit (Qiagen), according to the manufacturer's instructions. RNA purity was verified by OD 260:280 and OD 260:230 ratio analysis on a NanoDrop® ND-1000 Spectrophotometer (Thermo Scientific). cDNA was synthesized with the iScript™ cDNA Synthesis kit (Bio-Rad) according to the manufacturer's instructions.

Quantitative PCR

Real-Time PCR was set up using Brilliant II SYBR Green Master Mix™ and performed on an MX4005P Thermal Cycler™ according to manufacturer's protocols (Agilent Technologies), using primers as noted (). Expression was calculated with the ΔΔxp method normalized to Hprt, and all measurements were performed in triplicate.

Flow Cytometry and Cell Sorting

Tissues were homogenized by crushing with the head of a 1mL syringe in a petri dish followed by straining through a 40μm nylon filter. ACK buffer was used for red cell lysis and remaining cells were counted. Antibodies used for flow cytometry staining are listed in . BrdU was administered to recipient mice (1.5 mg BrdU) i.p. 4 h prior to sacrifice or added daily (0.8mg/mL) in 20% sucrose drinking water for 7 days. Isolated splenocytes were stained using the BrdU staining kit (BD Biosciences) following the manufacturer's protocol. Staining for caspase 3 was performed using Caspase 3 staining kit (AAT Bioquest, Inc) following the manufacturer's protocol. Staining for CXCR5 was performed at room temperature (25°C) with 30 min incubation. Gates for TFH cells (CXCR5 and PD-1) were determined in all instances based upon staining of these markers in CD44lo naïve CD4 T cells; i.e., the top of the CXCR5 and PD-1 gates for naïve T cells were set as the bottom of that for TFH cells. Intracellular staining for cytokines was performed using Cytofix/Cytoperm™ kits (BD Biosciences) following the manufacturer's protocol. Stained and rinsed cells were analyzed using an LSRII Multilaser Cytometer (BD Biosciences). For experiments requiring sorting, CD4+ T cells were enriched using a biotin-based magnetic separation kit (EasySep™, StemCell Technologies) prior to cell surface staining, with specific populations sorted using a FACSAria™ (BD Biosciences). CD40L surface mobilization we performed as previously described [37]. Briefly, cells were blocked with anti-FcγRII/III antibody, and 10 μg/mL APC-labeled anti-CD40L was introduced into the cell culture immediately before stimulation PMA (50ng/mL) and ionomycin (1μg/mL) (Sigma-Aldrich), and then incubated for 30 minutes at 37°C. After 3 washes, cells were surface stained to identify subsets of CD4+ T cells.

Microscopy

Spleens were snap frozen in OCT tissue-freezing solution and stored at −80°C. Tissues were cut into 8μm sections and processed as described previously [52]. Reagents used to stain sections are listed in . Images were obtained from a laser-scanning confocal microscope (Model 510 META; Carl Zeiss) at 25x magnification. ImageJ software from the National Institutes of Health was used for the measurement of GC and B cell follicle size, distance measurements, and for T cell counting. The latter analyses were performed by in a blinded manner. The distance from GC TFH cells to the dark zone (DZ) centers was normalized to the equivalent radius of the DZ.

Transwell Migration Assays

Chemotaxis assays were performed as described previously [53]. 5×105 enriched CD4 splenic T cells were incubated for 1 h with 1× DMEM containing 0.5% fatty acid–free BSA (EMD Biosciences), 5% antibiotics, L-glutamine (Cellgro), and Hepes. Cells were then allowed to migrate through 5-μm-pore–sized transwells (Corning) toward soluble CXCL12, CXLC13 (R&D Systems) or media alone, for 3 h at 37°C. Cells were collected, stained, and resuspended in 45 μl of staining buffer and analyzed by flow cytometry.

ELISPOT for Total Immunoglobulins

Splenocytes from mice infected with N. brasiliensis were harvested, analyzed by flow cytometry to determine B cell numbers (anti-CD19), and plated (1×105 cells/well) in quadruplicate on MultiScreen HTS plates (Millipore) coated with rat IgG anti-mouse light chain Abs, NP6-BSA, or NP33-BSA (Biosearch Technologies, 20 μg/ml). Antibodies used are listed in . Spots were developed with Vector Blue (Vector Laboratories), and quantified using an ImmunoSpot analyzer (Cellular Technology Limited).

Cytokine ELISPOT Assays

MultiScreen HTS plates were coated overnight at 4°C with anti-IL-21 and/or anti-IL-4 (eBioscience). Sorted cells were cultured (0.25 × 105 cells/well) with PMA (50ng/mL) and ionomycin (1μg/mL) for 36 h at 37°C followed by adding primary then secondary detection antibodies (. Spots were developed with Vector Blue (Vector Laboratories) and AEC (BD Biosciences). Spots were quantified using an ImmunoSpot analyzer (Cellular Technology Limited).

Sequence Analysis of the Vh186.2 Heavy Chain Ig Gene of NP-specific GC B Cells

GC B cells were harvested from spleens of recipient mice 13 days post infection. PCR was performed as previously described [54]. Briefly, the cDNA mixture was processed through two separate sets of the PCR reactions using Phusion high-fidelity Taq (0.02 U/μl/reaction) (Finnzymes). PCR products were cloned by zero blunt TOPO cloning (Invitrogen) and DNA sequence analysis was carried out on an Applied Biosystems 3730 capillary instrument.

RNA-Seq and Analysis

TFH populations were sorted by flow cytometry. Two separate sorts were performed on different days, pooling spleens of 25-30 mice each day. Quality verification, library preparation, and sequencing were performed at the Yale Center for Genomic Analysis. Samples were sequenced on an Illumina HiSeq 2000 using 75-bp paired end reads. Reads were aligned to the mm9 mouse genome using TopHat version 2.0.6 and Bowtie version 1.3.0, and differential expression computed with Cufflinks version 2.0.2 [55]. Analysis of RNA-seq data was done using the CummeRbund package version 2.0.0 in R. Differentially expressed genes were filtered by keeping transcripts with at least 1 read from each population significant at FDR-adjusted p-value < 0.05 between 2 of the populations, and Fragments Per Kilobase of transcript per Million mapped reads (FPKM) were summed across isoforms to obtain data for 1300 significant genes. Clustering was done using either hierarchical clustering in R by the “average” method or the partitioning around medoids method with an arbitrary value of k=4, and Jensen-Shannon distance metrics. For heatmaps, expression values were normalized per gene.

Statistics

Data were analyzed using the Student's t-test or Fisher's exact test with Prism 6 (GraphPad Software). The number of asterisks represents the degree of significance with respect to p value, with the exact value presented within each figure legend.