T cell-dependent plasmablasts form in the absence of single differentiated CD4⁺ T cell subsets

Abstract

The T follicular subset (Tfh) of CD4⁺ helper T (Th) cells promotes affinity maturation by B cells in germinal centers (GCs). The contribution of other Th cell subsets to B cell responses has not been fully explored in vivo. We addressed this issue by analyzing the T cell-dependent B cell response to the protein antigen phycoerythrin (PE) in mice lacking specific Th cell subsets. As expected, PE-specific GC B cell production required Tfh cells. However, Tfh-, Th1-, or Th17-deficient mice produced as many PE-specific isotype-switched plasmablasts as wild-type mice. This response depended on Th cell expression of CD154 and antigen presentation by B cells. These results indicate that many Th subsets can promote plasmablast formation by providing CD40 signals to naïve B cells.

Introduction

Maximal production of high affinity isotype-switched Abs by B cells depends on signals from CD4⁺ helper T (Th) cells. Antigen binding to surface Ig molecules (BCRs) causes naïve B cells to migrate to the border between the follicle and T cell area and present MHCII-bound antigen-derived peptides (p:MHCII) to Th cells (1–3) that were previously activated by the same p:MHCII complex on dendritic cells. The antigen-specific B cells then receive signals from the Th cells, proliferate, and undergo isotype switching (4–6). Some of the activated B cell progeny become extrafollicular Ab-secreting plasmablasts (PBs) while others enter germinal centers (GCs) along with specialized follicular helper (Tfh) cells that express the Bcl-6 transcription factor and the follicle-homing chemokine receptor CXCR5 (7). Tfh cells then engage in p:MHCII-dependent interactions with the GC B cells and drive somatic mutations and formation of high-affinity memory B cells and long-lived plasma cells (8, 9).

Early PB formation also depends on CD4⁺ T cells, (10, 11), but it is uncertain whether Tfh cells are required (12–14). Although isotype switched Abs were severely impaired in the absence of Tfh cells after NP-OVA in alum immunization (15) and Salmonella infection (16), another study showed that Th1 cells played a critical role in generating influenza-specific IgG2 Abs independently of Tfh cells (17). Furthermore, little is known regarding the contribution of other Th subsets such as Th17 cells to the B cell response in vivo. We therefore examined the contribution of Tfh, Th1, and Th17 cells to PB formation and isotype switching in response to a stimulus that primes all three Th subsets. The results show that although GC B cell formation was defective in mice lacking Tfh cells, formation of isotype-switched PBs occurred normally in mice lacking, Tfh, Th1, or Th17 cells. Isotype-switched PB formation was defective when CD154-CD40 interactions were absent or B cells could not present p:MHCII complexes. These results indicate that isotype-switched PB production requires a CD40-dependent form of cognate T cell help that does not depend on a single differentiated Th cell subset.

Materials and methods

Mice.

Six-12 week old male and female mice were used. C57BL/6 (B6) and CD45.1⁺ (B6.SJL-Ptprca Pep3b/BoyJ) mice were purchased from the National Cancer Institute (Frederick, MD). Bcl6^fl/fl (B6.129S(FVB)-Bcl6^tm1.1Dent/J) (18), Tbx21^fl/fl (B6.129-Tbx21^tm2Srnr/J) (19), Rorc^fl/fl (B6(Cg)-Rorc^tm3Litt/J), Lck-cre (B6.Cg-Tg(Lck-icre)3779Nik/J), Tcra^−/− (B6.129S2-Tcratm1Mom/J), CD154^−/− (B6.129S2-Cd40lgtm1Imx/J) (20), H2^dlAb1-Ea (MHCII-deficient; B6.129S2-H2dlAb1-Ea/J) (21), and B cell-deficient (μMT; B6.129S2-Ighm^tm1Cgn/J) (22) mice were purchased from The Jackson Laboratory. Mice with floxed alleles were crossed to Lck-cre mice to obtain mice with two floxed alleles and one Lck^cre allele. Mice with floxed alleles but lacking the Lck^cre allele served as controls. A.L. Dent (Indiana University) provided WT and Bcl6^−/− mice (23). All mice were housed in specific pathogen–free conditions at the University of Minnesota. Experimental protocols were performed in accordance with guidelines of the University of Minnesota Institutional Animal Care and Use Committee and National Institutes of Health.

Immunizations.

Biotinylated 2W peptide (GenScript) was mixed with streptavidin- phycoerythrin (PE) (ProZyme) in PBS at a 4:1 molar ratio to form 2W-PE. Each mouse was injected i.p. with 100 μl of 0.6 μg 2W peptide conjugated to 25 μg PE emulsified in 100 ul CFA (Sigma-Aldrich).

Bone marrow chimeras and cell transfer experiments.

Recipient CD45.1⁺ mice were irradiated twice with 500 rad with 6 hours between doses. After the second irradiation, 1 × 10⁶ CD45.2⁺ Bcl6^−/− or WT bone marrow cells were injected into recipients. After 8–10 weeks, CD45.2⁺ Bcl6^−/− or WT T cells T cells were isolated to 90–99% purity by negative selection using the EasySep™ Mouse T Cell Isolation Kit (Stemcell) with added biotin-conjugated CD45.1 Ab (eBioscience). Thirteen-16 × 10⁶ CD45.2⁺ Bcl6^−/− or WT CD4⁺ T cells were injected into Tcra^−/− T cell-deficient mice before immunization with 2W-PE in CFA the next day. B cells were isolated from WT and MHCII-deficient mice using a negative selection kit (Miltenyi Biotec) and 88 × 10⁶ B cells were injected separately into μMT mice before immunization with 2W-PE in CFA.

Cell enrichment and flow cytometry.

Single cell suspensions of spleens and lymph nodes were split equally for Th and B cell analyses. For 2W:I-A^b-specific T cell analysis, cells were stained with fluorochrome-conjugated CXCR5 (2G8; BD) Ab and allophycocyanin-conjugated I-A^b tetramer containing 2W peptide (EAWGALANWAVDSA) for one hr at room temperature. Tetramer-bound cells were positively enriched using allophycocyanin-specific magnetic isolation (Stemcell). Tetramer-enriched cells were stained with fluorochrome-labeled Abs specific for B220 (RA3–6B2; all Abs from eBioscience unless otherwise indicated), CD11b (M1/70), CD11c (N418), CD44 (IM7), PD-1 (J43), CD90.2 (53–2.1), or CD4 (GK1.5; BD). Cells were incubated in fixation/permeabilization buffer (eBioscience) and the fluorochrome-labeled Bcl-6 (K112–91; BD), T-bet (4B10; BioLegend), and RORγt (Q31–378; BD) Abs in permeabilization buffer (eBioscience).

For PE-specific B cell analysis, spleens and lymph node fragments were incubated with Dispase (Invitrogen), collagenase P (Roche), and DNase I (Roche) at 37°C for 20 min. The released cells were mixed with unlabeled CD16/CD32 (2.4G2) Ab (Tonbo) and fluorochrome-labeled Abs specific for IgG1 (A85–1; BD), IgG2b (polyclonal; Life Technologies), IgG3 (polyclonal; Life Technologies), or IgA (C10–3; BD). Cells were then incubated with 1 μg of PE (ProZyme) for 30 min at 4° C (15). PE-bound B cells were positively enriched using PE-specific magnetic isolation (Stemcell). PE-enriched B cells were stained with fluorochrome-labeled Abs against CD90.2 (53–2.1), CD11c (N418), F4/80 (BM8), GR1 (RB6–8C5), CD38 (90), IgM (II/41), GL7 (GL-7), IgD (11–26c.2a; BD), and B220 (RA3–6B2; BD). Cells were fixed with Fixation/Permeabilization buffer (BD) and stained with fluorochrome-labeled Abs against IgG1, IgG2b, IgG3, and IgA followed by biotin-labeled IgG2c Ab (IgG2a [b]; 5.7; BD) and then IgG [H+L] Ab (Life Technologies). Cells were analyzed on a Fortessa (Becton Dickinson) flow cytometer and analyzed with FlowJo (TreeStar).

Statistical analysis.

Statistical tests were performed using Prism (Graphpad) software. Data were log transformed and p values were obtained from a one-way ANOVA and Dunnett post-test comparing all groups to the WT control group with a 95% confidence interval. In cases where only two groups are compared to one another, p values were obtained from two-tailed unpaired t tests with a 95% confidence interval.

Results and Discussion

Analysis of mice deficient in Tfh, Th1, and Th17 cells

We used mice lacking functional exons of Th lineage-defining transcription factors (15, 24, 25) to study the contribution of Th cell subsets to B cell responses. WT, Lck-cre⁺Bcl6^fl/fl (Tfh-deficient), Lck-cre⁺Tbx21^fl/fl (Th1-deficient), Lck-cre⁺Rorc^fl/fl (Th17-deficient), and Tcra^−/− (T cell-deficient) C57BL/6 (B6) mice, which express the I-A^b MHCII molecule, were immunized with a CFA emulsion containing an immunogenic I-A^b binding peptide called 2W (26) linked to PE (2W-PE). This immunization strategy allowed for analysis of 2W:I-A^b-specific CD4⁺ T cell and PE-specific B cell responses in the same host (27). 2W:I-A^b-specific CD4⁺ T cells in secondary lymphoid organs were identified by flow cytometry after 2W:I-A^b tetramer staining and magnetic bead enrichment (28) (Fig. 1A). WT mice, which have about 300 2W:I-A^b-specific CD4⁺ naïve T cells (28), contained about 6,000 2W:I-A^b-specific CD4⁺ CD44^high effector T cells in secondary lymphoid organs on day 11 after immunization with 2W-PE in CFA (Fig. 1B). The expanded effector cell population in WT mice contained on average 1,300 CXCR5⁺ PD-1⁻ Tfh and 1,000 CXCR5⁺ PD-1⁺ GC-Tfh, 600 CXCR5⁻ T-bet⁺ Th1, and 500 RORγt⁺ Th17 cells (Fig. 1C, D). The Bcl-6-deficient effector T cell population contained 10–30-fold fewer Tfh and GC-Tfh cells than the WT population but had normal numbers of Th1 and Th17 cells (Fig. 1D). Mice with T-bet-deficiency generated 10-folder fewer 2W:I-A^b-specific Th1 cells than WT mice but had normal numbers of Tfh, GC-Tfh, and Th17 cells, whereas RORγt-deficient mice generated 10-fold fewer Th17 and 4-fold fewer Tfh and GC-Tfh cells than WT mice but had normal numbers of Th1 cells. These results demonstrate that the expected T cell subsets were missing in mice lacking Bcl-6, T-bet, or RORγt with exception that RORγt-deficiency created a small reduction in Tfh and GC-Tfh cells in addition to Th17 cells.

FIGURE 1.

Analysis of 2W:I-A^b-specific T cells in Tfh-, Th1- and Th17-deficient mice. (A) CD44⁺ 2W:I-A^b tetramer⁺ cells gated from B220⁻ CD11b⁻ CD11c⁻ CD4⁺ T cells from 2W:I-A^b tetramer-enriched spleen and lymph node samples from the indicated mice immunized with 2W-PE in CFA 11 days earlier. (B) Numbers of 2W:I-A^b-specific cells from groups in (A). (C) Identification and (D) numbers of Tfh, GC-Tfh, Th1, and Th17 cells (from gate in (A)), based on CXCR5, PD-1, T-bet, and RORγt expression. Data are expressed as mean values and representative of two independent experiments (n = 3–6 mice/group). P values were obtained from one-way ANOVA and Dunnett post-test that compares all groups to the WT control group (* p < 0.05, *** p < 0.001).

Antigen-specific B cell analysis in mice deficient in Tfh, Th1, and Th17 cells

We then assessed the roles of individual CD4⁺ T cell subsets in the PE-specific B cell response. PE-specific B cells, PBs, GC B cells (Fig. 2A) and isotype-switched PBs (Fig. 2C) were detected in secondary lymphoid organs by flow cytometry after PE staining and magnetic bead enrichment (11). The ~20,000 naive PE-specific B cells in WT mice (11) produced about 300,000 activated B cells by day 11 after immunization with 2W-PE in CFA (Fig. 2B). The activated B cell population contained 70,000 CD38⁻ GL7⁺ GC B cells (Fig. 2B) and 40,000 intracellular Ig^high isotype-switched (IgM⁻ IgD⁻) PBs (Fig. 2D) including cells with the IgG1, IgG2b, IgG2c, IgG3, or IgA isotypes (Supplemental Fig. 1). In contrast, only 40,000 activated PE-specific B cells, including less than 100 GC B cells and 400 isotype-switched PBs were generated in T cell-deficient Tcra^−/− mice (Fig. 2B, D), confirming the T cell dependence of the PE-specific response (11). Mice with T cell-specific deficiency in Bcl-6 and lacking Tfh cells produced 2-fold fewer PE-specific activated B cells than WT mice. This reduction was due to a 15-fold drop in the number of PE-specific GC B cells, confirming that GC B cell formation depends on Tfh cells (7). Surprisingly, however, mice lacking Tfh cells produced on average 22,000 isotype-switched PBs, a number that was not significantly different from the 37,000 produced in WT mice. The isotype switched PE-specific PB population that formed in mice with T cell-specific Bcl-6 deficiency also had the same composition of isotypes as WT mice (Supplemental Fig. 1). Similarly, mice with T cell-specific deficiencies in T-bet or RORγt and lacking Th1 or Th17 cells, respectively, produced the same number of PE-specific total and GC B cells, and isotype-switched PBs with the same isotypes as WT mice. These results indicate that although formation of isotype-switched PBs is T cell-dependent, Tfh, Th1 or Th17 cells are not uniquely required for this function.

FIGURE 2.

Early T cell dependent PBs form in the absence of single differentiated CD4⁺ T cell subsets. (A) Identification of PE⁺ cells among CD90.2⁻ CD11c⁻ F4/80⁻ Gr-1⁻ cells (left panel) from PE-enriched spleen and lymph node samples from WT mice immunized with 2W-PE in CFA 11 days earlier. The middle and right panels show gates used to identify PE-specific PBs (B220^low IgG [H+L]^high) or GC B cells (B220^high CD38⁻ GL7⁺). (B) Numbers of PE-specific total (left panel) and GC B cells (right panel). (C) Gating and (D) numbers of IgM⁻ IgD⁻ PBs. Data are expressed as the mean value and representative of two independent experiments (n = 2–6 mice/group). P values were obtained from a one-way ANOVA and Dunnett post-test that compares all groups to the WT control group (* p < 0.05, *** p < 0.001).

Confirmation that early switched PBs form in the absence of Tfh cells

Recent work in an influenza infection model showed that production of IgG1 Abs was reduced in mice without Tfh cells (17). Thus, it was surprising to find that PE-specific IgG1⁺ PBs formed normally in Lck-cre⁺Bcl6^fl/fl Tfh-deficient mice (Supplemental Fig. 1). It was concerning, however, that these mice still generated some PE-specific GC B cells (Fig. 2B). Lck-cre⁺Bcl6^fl/fl mice have a deletion in exons 7–9 (18) of the Bcl6 gene and therefore could express a truncated but partially functional version of the Bcl-6 protein that could have supported weak Tfh formation in the experiment shown in Fig. 1. Residual Tfh cells may have been sufficient for the weak GC B cell and normal isotype-switched PB formation observed in these mice (Fig. 2). This concern was ameliorated by transferring Bcl6^−/− T cells, which completely lack the Bcl6 gene, into T cell-deficient mice prior to immunization with 2W-PE in CFA. Recipients of Bcl6^−/− T cells made no more PE-specific GC B cells than mice that did not contain T cells, and many fewer than T cell-deficient mice that received WT T cells (Fig. 3A, C). T cell-deficient mice that received Bcl6^−/− T cells, however, generated the same number of isotype-switched PBs that had the same distribution of isotypes as T cell-deficient mice that received WT T cells (Fig. 3B, C, Supplemental Fig. 2). Thus, T cell-dependent isotype-switched PB formation does not require Tfh cells in this system.

FIGURE 3.

Transfer of Bcl6^−/− CD4⁺ T cells into TCRα knockout mice confirms that PBs form in the absence of Tfh cells. CD45.2⁺ WT or Bcl6^−/− T cells were transferred into T cell-deficient mice and immunized with 2W-PE in CFA. Immunized B6 and T cell-deficient mice with no transferred T cells were used as controls. After 11 days, PE-specific B cells were enriched from spleen and lymph nodes. (A) CD90.2⁻ CD11c⁻ F4/80⁻ Gr-1⁻ PE-specific B cells from the indicated groups with a gate on B220^high CD38⁻ GL7⁺ GC B cells or (B) B220^low IgG [H+L]^high PBs. (C) Numbers of PE-specific GC B cells (top panel) and IgM⁻ IgD⁻ PBs (bottom panel). Data are expressed as mean values and representative of two independent experiments (n = 2–5 mice/group). P values were obtained from a one-way ANOVA and Dunnett post-test comparing all groups to the WT T cell transfer into T cell-deficient group (* p < 0.05, *** p < 0.001).

Early switched PBs require interactions with CD4⁺ T cells and CD40 signaling

Previous work showed that T cells other than CD4⁺ T cells can contribute to humoral responses. For example, γδ and NK T cells can influence isotype switching (29, 30). We therefore explored the nature of the T cell-dependence of isotype-switched PB formation by determining whether it depended on p:MHCII presentation by B cells as expected if cognate interaction between Th cells and B cells is required. B cells from WT or H2^dlAb1-Ea (MHCII-deficient) mice (21) were transferred into B cell-deficient μMT mice (22) before immunization with 2W-PE to test this possibility. μMT recipients of MHCII-deficient B cells formed 200-fold fewer PE-specific isotype-switched PBs than recipients of WT B cells (Fig. 4A, B) indicating that cognate interactions with Th cells were required. We then determined whether CD40 signaling in the B cells, which occurs during cognate interactions with CD154⁺ Th cells (31), was also required for PE-specific isotype-switched PB formation. Indeed, PE-specific isotype-switched PB formation was as defective in Cd154^−/− mice immunized with 2W-PE in CFA as in mice lacking all Th cells (Fig 4C, D).

FIGURE 4.

Early PBs require cognate interactions with CD4⁺ T cells and CD40-CD154 signaling. (A) CD90.2⁻ CD11c⁻ F4/80⁻ Gr-1⁻ PE-specific WT or MHCII-deficient B cells from μMT recipient mice on day 11 after 2W-PE in CFA immunization with a gate on B220^low IgG [H+L]^high PBs. (B) Numbers of PE-specific total B cells (left panel) and IgM^- IgD^- PBs of the indicated types (right panel). (C) CD90.2⁻ CD11c⁻ F4/80⁻ Gr-1⁻ PE-specific B cells from T cell-deficient, WT, or CD154-deficient mice on day 11 after immunization with 2W-PE in CFA with a gate on B220^low IgG [H+L]^hi PBs. (D) Numbers of PE-specific total B cells (left panel) or IgM⁻ IgD⁻ PBs of the indicated types (right panel). Data are expressed as mean values and representative of two independent experiments (n = 2–5 mice/group). In (B), the p value was obtained from a two-tailed unpaired t test (*** p < 0.001). In (D), p values were obtained from a one-way ANOVA and Dunnett post-test that compares all groups to the WT control group (** p < 0.01, *** p < 0.001).

Our results suggest a model in which early cognate interactions between undifferentiated Th cells and B cells are sufficient for the generation of isotype-switched PBs. The independence of this response from Tfh cells, which use CXCR5 to exert their actions in follicles (32–34) indicates that it could occur outside of this location. These results are in line with previous work showing that isotype switching can be achieved by extrafollicular PBs, outside of the Tfh-rich GC environment (35, 36). Our work provides the additional insight that this process does not rely on other Th subsets like Th1 and Th17 cells in the case of immunization with CFA. It remains possible, however, that the PB response has a greater dependence on a given Th subset during other types of immune responses.

Recently, it has been shown that antigen-stimulated Th cells rapidly induce the G-protein-coupled receptor EBI2 (GPR183) (37) and migrate to the outer T cell zone to meet antigen-stimulated B cells (1, 2). Because this migration does not require CXCR5 it could be achievable by as yet undifferentiated Th cells. Once in the outer T cell zone, these Th cells could deliver CD40 signals to their B cell partners, driving them to proliferate and become extrafollicular PBs.

Abbreviations used in this article:

PE

Phycoerythrin

PB

plasmablast