Development of allergen-specific IgE in a food-allergy model requires precisely timed B cell stimulation and is inhibited by Fgl2.

Immunoglobulin E (IgE) responses are a central feature of allergic disease. Using a well-established food-allergy model in mice, we show that two sensitizations with cognate B cell antigen (Ag) and adjuvant 7 days apart promotes optimal development of IgE+ germinal center (GC) B cells and high-affinity IgE production. Intervals of 3 or 14 days between Ag sensitizations lead to loss of IgE+ GC B cells and an undetectable IgE response. The immunosuppressive factors Fgl2 and CD39 are down-regulated in T follicular helper (TFH) cells under optimal IgE-sensitization conditions. Deletion of Fgl2 in TFH and T follicular regulatory (TFR) cells, but not from TFR cells alone, increase Ag-specific IgE levels and IgE-mediated anaphylactic responses. Overall, we find that Ag-specific IgE responses require precisely timed stimulation of IgE+ GC B cells by Ag. Furthermore, we show that Fgl2 is expressed by TFH cells and represses IgE. This work has implications for the development and treatment of food allergies.

INTRODUCTION

Food allergy is a pathologic response of the immune system triggered by food antigens (Ags). The prevalence of food allergy has been increasing since the 1990s as part of the second wave of the allergy epidemic (Asher et al., 2006; Yu et al., 2016), and it is estimated that up to 8% of children are affected by a food allergy in the United States (Tordesillas et al., 2017). One of the most important categories of food allergy is immunoglobulin E (IgE)-mediated immediate hypersensitivity reactions, as cross-linking of allergen-specific IgE bound to mast cells induces severe immune responses including anaphylaxis (Gould and Ramadani, 2018; Gould et al., 2003).

Despite its critical role in mediating allergic disease, we are just beginning to understand the mechanism of IgE regulation in vivo. Like the other Ig isotypes, IgE is regulated by several cytokines as well as signaling receptors. Interleukin-4 (IL-4) activation of STAT6 promotes IgE class switching in B cells (Finkelman et al., 1988; Kopf et al., 1993; Shimoda et al., 1996). IL-21 is a potent suppressor of IgE responses, while CD40 signaling promotes IgE (Yang et al., 2020). IgE has several unique features compared with other Ig isotypes including lower surface expression on B cells and increased differentiation of IgE+ B cells into plasma cells (Haniuda et al., 2016).

The germinal center (GC) micro-environment is required for antibody (Ab) affinity selection and high-affinity Ab production (Klein and Dalla-Favera, 2008; Victora and Nussenzweig, 2012), while high-affinity IgE is required for pathogenesis and anaphylaxis of allergic diseases (Gowthaman et al., 2019; Xiong et al., 2012). The GC is not a favorable environment for IgE+ B cells; IgE+ GC B cells have altered signaling, are more likely to undergo apoptosis, and are unlikely to develop into memory cells (Haniuda et al., 2016; Harris et al., 2005; He et al., 2013; Newman and Tolar, 2021; Saunders et al., 2019). It has been reported that low-affinity IgE is produced by IgE+ B cells through direct class switching, while indirect class switching from IgG1+ B cells accounts for high-affinity IgE (Erazo et al., 2007; He et al., 2017; Xiong et al., 2012; Yang et al., 2014). However, it is still controversial whether pathogenic high-affinity IgE can develop from IgE+ GC B cells (He et al., 2013; Talay et al., 2012; Yang et al., 2012). Overall, little is still known about the pathways regulating the initial class switching to IgE in vivo and expansion of IgE+ B cells in GC.

Several papers have suggested that the allergic response is dependent on T follicular helper (TFH) cells and that TFH-derived IL-4 is essential for the development of Ag-specific IgE (Ballesteros-Tato et al., 2016; Dolence et al., 2018; Kobayashi et al., 2017; Meli et al., 2017; Noble and Zhao, 2016; Zhang et al., 2020). Furthermore, a subset of TFH cells producing IL-13 (TFH13) is required for the development of high-affinity IgE (Gowthaman et al., 2019). T follicular regulatory (TFR) cells also regulate the development of IgE, either positively or negatively depending on the model used (Clement et al., 2019; Koh et al., 2020; Xie et al., 2020). Overall, much is still unknown about the precise role of TFH and TFR cells in allergen-specific IgE class switching and production.

In this study, we utilized an intragastric-sensitization food-allergy model and found that the timing of sensitization by the allergen is a critical factor for allergen-specific IgE production, which affects the expansion of Ag-specific B cells and the activity of TFH cells. In addition, we identified two molecules (fibrinogen-like protein 2 [Fgl2] and the ectonucleotidase CD39 [Entpd1]) produced by TFH cells that play an inhibitory role in regulating high-affinity IgE production.

RESULTS

Development of food-allergen-specific IgE requires two Ag sensitization steps

To better characterize the development of allergen-specific IgE in food allergy, we used a widely used method involving repeated intragastric priming or sensitization steps with peanut protein plus cholera toxin (PCT) (Li et al., 2000; Orgel and Kulis, 2018). Our first goal was to define the minimum sensitization steps required for the development of allergen-specific IgE (Figure 1A). We found that two PCT sensitization steps produced robust peanut (PN)-specific IgE and PN-specific IgG1 responses, while no PN-specific IgE or PN-specific IgG1 was detected following only one sensitization (Figure 1B). In contrast to PN-specific IgE, total IgE is seen with just one sensitization (Figure 1B) and is higher after one PCT sensitization compared with naive mice (Figure 1C). In addition, the IgE produced by day 15 of this two-step sensitization model can cause anaphylaxis (Figure 1D). Thus, the development of allergen-specific IgE, but not total IgE, is dependent on a second sensitization step in this food-allergy model.

Figure 1.

Development of Ag-specific IgE, but not total IgE, requires a second peanut and cholera toxin (PCT) sensitization

(A) Schematic figure of the PCT sensitization model.

(B) PN-specific IgE, PN-specific IgG1, and total IgE at the indicated time points after PCT sensitizations. Mice were sensitized with PCT at day 1 or at days 1 and 8, then sera were collected at day 15. n = 4.

(C) Total IgE in naive mice and mice after PCT sensitization. n = 5.

(D) Anaphylaxis reaction after PCT sensitizations. Naive and sensitized mice were injected with 2 mg peanut protein (intraperitoneally [i.p.]) at day 15. n = 7.

*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 by two-way ANOVA. Data are representative of two independent experiments. Data are mean ± SEM.

The second sensitization step promotes the expansion of Ag-specific GC B cells and IgE+ GC B cells

To define exactly when allergen-specific IgE develops in the PCT food-allergy model, we measured the levels of IgE Abs at several time points before and after the second sensitization (Figure 2A). PN-specific IgE only emerged at day 12, 4 days after the second sensitization. PN-specific IgG1 could be detected at day 10 at a very low level. In contrast, total IgE could be detected at early as day 7 with titers increasing over the course of the experiment (Figures 1C and 2A). When ovalbumin (OVA) was used to replace PN, similar results were obtained in which OVA-specific IgE was first detected 4 days after the second sensitization (Figure S1A). These data indicate that Ag-specific IgE can develop early after the sensitization steps and likely does not require the formation of memory B cells. To characterize Ag-specific B cells in our model, an 4-hydroxy-3-nitrophenylacetyl (NP)-OVA hapten-protein conjugate was used in place of PN protein extract to allow for the detection of Ag-specific B cells (Figure S1B). Compared with naive mice, both the IgE+ GC B cells and NP-specific GC B cells increased significantly after the second sensitization (Figure 2B). Though at a very low level, NP-specific IgE+ GC B cells could be detected, and these cells increased significantly after the second sensitization (Figure 2B). IgE+ plasma cells and NP-specific IgE+ plasma cells also increased after the second sensitization (Figure S1C).

Figure 2.

Second sensitization maintains the expansion of Ag-specific GC B cells and IgE+ GC B cells and inhibits apoptosis

(A) Development of PN-specific IgE after days 1 and 8 sensitizations. Mice were sensitized with PCT at days 1 and 8. PN-specific IgE, PN-specific IgG1, and total IgE were determined at the indicated time points. n = 4.

(B) NP-specific IgE+ GC B cells before and after second sensitization. Verigem mice were sensitized with NP-OVA + CT at day 1 or days 1 and 8. IgE+ GC B cells and NP-specific GC B cells were analyzed at the indicated time points. Cell number was calculated per mesenteric lymph node (mLN). n = 4 or 8.

(C) NP-specific IgG1+ GC B cells before and after second sensitization. Verigem mice were sensitized with NP-OVA + CT at day 1 or days 1 and 8. IgG1+ GC B cells and NP-specific GC B cells were analyzed at the indicated time points. Cell number was calculated per mLN. n = 4 or 8.

(D) Apoptosis of GC B cells and IgE+ GC B cells. Mice were sensitized with PCT at day 1 or days 1 and 8. Annexin V-positive GC B cells and IgE+ GC B cells were analyzed. n = 5.

(E) GC B cells and NP-specific GC B cells with and without second sensitization. Mice were sensitized with NP-OVA + CT at day 1 or days 1 and 8. GC B cells and NP-specific GC B cells were analyzed at the indicated time points. Cell number was calculated per mLN. n = 5.

*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 by one-way ANOVA. ns, not significant. Data are representative of two independent experiments. Data are mean ± SEM.

The IgG1+ B cell response showed similar kinetics to the IgE+ B cell response, with increases of IgG1+ GC B cells, NP-specific IgG1+ GC B cells, IgG1+ plasma cells, and NP-specific IgG1+ plasma cells after the second sensitization (Figures 2C and S1C). We hypothesized that the second sensitization may maintain the GC B cell response and inhibit apoptosis of GC B cells. We found that the percentage of both annexin V+ total GC B cells and annexin V+ IgE+ GC B cells decreased at least 2-fold after the second sensitization (Figure 2D). To better characterize the role of the second sensitization in the GC B cell response, we compared GC B cells and NP-specific GC B cells at day 12 following one or two sensitizations. We observed that, from day 8 to 12, NP-specific GC B cells decreased significantly without a second sensitization at day 8 (Figure 2E). The second sensitization rescued the NP-specific GC B cells and promoted the expansion of GC B cells and NP-specific GC B cells (Figure 2E). The expression of IgE germline transcript (ε GLT), a marker for IgE switching, mainly increased after the second sensitization, while the expression of GLT of IgG1 (γ1 GLT) and Aicda increased steadily (Figure S1D).

Development of Ag-specific IgE, but not IgG1, is dependent on T cells and stimulation of Ag-specific B cells at second sensitization

Since the second sensitization is critical for the IgE response, we compared the role of Ag versus cholera toxin (CT) for the second sensitization. There was no significant difference for the level of PN-specific IgE between groups in which PN or PN + CT were used at the second sensitization, while CT alone didn’t induce PN-specific IgE at all (Figures 3A and S2A). These results suggest that the specific Ags are critical at the second sensitization. To determine whether the development of Ag-specific IgE is T helper cell dependent, we used NP-OVA, a T cell-dependent Ag, and NP-Ficoll, a T cell-independent Ag, in our food-allergy model. NP-specific IgE was detected in mice sensitized twice with NP-OVA but was absent in mice when NP-Ficoll was used (Figures 3B and S2B). In contrast, NP-specific IgG1 could be detected in all three groups (Figures 3B and S2B). This indicates that Ag-specific IgG1 can develop by a T cell-independent pathway. NP-specific GC B cell numbers were significantly lower when NP-Ficoll was used, suggesting that activation of T helper cells is critical to promote NP-specific GC B cells (Figure S2C). To study how B cells are involved in supporting Ag-specific IgE at second sensitization, we compared Ags only versus Ag-hapten conjugates (OVA versus NP-OVA) at the second sensitization after NP-OVA at the first sensitization. NP-IgE could only be produced with NP-OVA and not OVA alone, indicating that activation of existing NP-specific B cells at the second sensitization is critical (Figures 3C and S2D). To further study how T helper cells support Ag-specific IgE at the second sensitization, we used two different T cell Ags within the Ag-hapten conjugates (NP-OVA versus NP-KLH) at the second sensitization after NP-OVA sensitization at the first sensitization. NP-specific IgE was detected in mice receiving NP-OVA or NP-KLH sensitization at the second sensitization (Figures 3C and S2D). The similar level of NP-specific IgE regardless of the carrier protein used for the second sensitization indicates that the specificity of TFH cells is not critical at the second sensitization. At the same time, high levels of NP-specific IgG1 were produced in all conditions, again indicating a higher stringency in the development of Ag-specific IgE responses compared with Ag-specific IgG1 responses (Figure 3C).

Figure 3.

Development of PN-specific IgE, but not IgG1, is dependent on T cells and stimulation of Ag-specific B cells at second sensitization

(A) Mice were sensitized with PCT at day 1. These mice were secondarily sensitized with nothing, PN, CT, or PCT at day 8. PN-specific IgE and PN-specific IgG1 were analyzed. n = 4.

(B) Mice were sensitized with different recombination of NP-OVA and NP-Ficoll plus CT at days 1 and 8. NP-specific IgE and NP-specific IgG1 were analyzed. n = 4.

(C) Mice were sensitized with NP-OVA + CT at day 1. These mice were secondarily sensitized with different antigens plus CT as indicated at day 8. NP-specific IgE and NP-specific IgG1 were analyzed by ELISA. Sera dilutions for optical densities (ODs) in graphs: PN-IgE at 1:10; PN-IgG1 at 1:50. n = 4.

*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 by one-way ANOVA. ns, not significant. Data are representative of two independent experiments. Data are mean ± SEM.

To exclude that NP-KLH may be more potent than NP-OVA in inducing Ag-specific IgE, we compared different groups using a combination of NP-OVA and NP-KLH and found that there is no difference between these groups for the NP-specific IgE (Figure S2E). To study the effects on TFH cells when different combinations of Ags were used (Figure 3C), we characterized the level and activation status of TFH cells in GC (Figure S2F). We found that there were significantly more TFH cells, CD137-positive TFH cells, and CD90-negative/low TFH cells with NP-OVA at the second sensitization than the other Ags but not more CD69+ TFH cells (Figure S2G), suggesting that TFH cell activation and GC residency (Yeh et al., 2022) may be affected by the specific Ag presented by GC B cells. Overall, these data showed that the development of Ag-specific IgE in our food-allergy model is dependent on both T helper cell activation and the stimulation of existing Ag-specific B cells at the second sensitization. However, the specificity of T helper cells is not critical at the second sensitization for the development of Ag-specific IgE.

Development of allergen-specific IgE, but not allergen-specific IgG1, requires proper timing of sensitization by allergen

The requirement of second sensitization in food allergy and the critical stimulation of Ag-specific B cells made us wonder whether the timing of the second sensitization could affect the development of Ag-specific IgE. We compared the PN-specific IgE and PN-specific IgG1 in mice receiving their second sensitization at day 3, 8, or 15 after the first sensitization (Figure 4A).

Figure 4.

Development of Ag-specific IgE, but not IgG1, requires proper timing of second sensitization

(A) Levels of PN-specific IgE, PN-specific IgG1, and total IgE at the indicated time points after different timing of PCT. n = 4. Asterisk (*) represents comparison between days 1 and 8 group and days 1 and 3 group; hash (#) represents comparison between days 1 and 8 group and days 1 and 15 group.

(B) Expression of GLTs and Aicda after PCT sensitizations at days 1 and 3 and days 1 and 8 in total B cells. n = 4.

(C) Numbers of TFH and TFR cells in spleen and mLN after PCT at days 1 and 3 or days 1 and 8. n = 5. Cell number was calculated per spleen or mLN.

(D) GC B cells in spleen and mLN after PCT at days 1 and 3 or days 1 and 8. Cell number was calculated per spleen or mLN. n = 5.

*p < 0.05; **p < 0.01; ***p < 0.001 by two-way ANOVA (A) or t test (B–D). ns, not significant. Data are representative of two independent experiments. Data are mean ± SEM.

Strikingly, the PN-specific IgE was only detectable in mice with sensitization at days 1 and 8, which contrasts with the development of PN-specific IgG1 that occurred following the secondary sensitization at all three time points (Figure 4A). The expression of ε-GLT and Aicda, but not γ1-GLT, were significantly higher in the days 1 and 8 group than the days 1 and 3 group (Figure 4B), further suggesting that Ag-specific IgE, but not IgG1, requires precise timing for the second sensitization in the food-allergy model. We measured the numbers of TFH, TFR, and GC B cells between days 1 and 3 and days 1 and 8 and didn’t find any significant differences (Figures 4C and 4D), indicating a specific effect of timing on IgE regulation.

Properly timed expansion of GC B cells is required for the development of Ag-specific IgE

We then investigated why the interval between the two sensitization steps affected the development of Ag-specific IgE. One hypothesis is that 3 days after the first sensitization, GC B cells cannot be activated at the second time, perhaps due to a refractory state from the first stimulation. Another hypothesis is that there are too few Ag-specific GC B cells at day 3 to stimulate with a second sensitization that will give a robust response. To test these hypotheses, mice were sensitized twice at days 1 and 8 then given a third sensitization at 3 days after the second sensitization (days 1, 8, and 11) or 7 days after the second sensitization (days 1, 8, and 15). Here, there is a large starting population of Ag-specific GCB to stimulate again, in contrast to the stimulation 3 days after just one sensitization. We observed that the titer of Ag-specific IgE with sensitization at days 1, 8, and 15 was significantly higher than the days 1, 8, and 11 titer (Figure 5A). To exclude the possibility that the difference observed is caused by the prolonged interval between analysis and last sensitization, we measured PN-specific IgE at similar days after the last sensitization, and the PN-specific IgE was still significantly higher with sensitization at days 1, 8, and 15 than at days 1, 8, and 11 (Figure 5B). NP-specific GC B cell numbers were not significantly affected by the differently timed sensitizations (Figure 5C). These data suggested that an Ag-specific IgE response requires an optimal period of time between sensitizations, past an initial refractory period where GCB cells cannot be productively stimulated again. This refractory period appears to particularly critical for IgE+ GC B cells. To understand why a longer time between sensitizations doesn’t lead to an IgE response, we measured GC B cell numbers over time after one sensitization. GC B cell and Ag-specific GC B cell numbers were higher at day 8 than that at days 3 and 15 (Figure 5D), indicating that these cells decay after the day 8 peak. Additionally, IgG1+ GC B cells and CD138+ plasma cells were consistent with a cellular response that is highest at day 8 (Figure S3A). These results suggested that the optimal expansion of GC B cells relies on the second sensitization of Ag-specific GC B cells at day 8, which leads to the development of Ag-specific IgE in the food-allergy model. However, the apoptosis of GC B cells was not significantly affected by different timings of sensitizations (Figure S3B). Taking these data together, we can construct a model where an early second sensitization cannot stimulate Ag-specific GC B cells to produce an Ag-specific IgE response, whereas too late of a second sensitization leads to apoptotic loss of the Ag-specific GC B cells before they can generate an Ag-specific IgE response.

Figure 5.

Properly timed expansion of GC B cells is required for the development of anti-specific IgE

(A) Levels of PN-specific IgE at day 22 after different timing of PCT. n = 5 or 6.

(B) Levels of PN-specific IgE at 4, 7, and 10 days (separate titration graphs) after different timings of PCT. n = 4 or 5.

(C) GCB and NP-specific GC B cells at day 22 after different timing of PCT. Cell number was calculated per mLN. n = 5.

(D) GC B cells and NP-specific GC B cells at different time points after just one sensitization. Cell number was calculated per mLN. n = 6.

*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 by t test (A, C, and D) or two-way ANOVA (A and B). Data are representative of two independent experiments. ns, not significant. Data are mean ± SEM.

TFH and TFR gene expression is altered by timing of Ag sensitization

In Figure 3, we show that the development of Ag-specific IgE is T cell-dependent at both sensitization steps, and our earlier studies showed that both TFH and TFR cells are required for IgE production (Xie et al., 2020). We then wondered if the sensitization timing affected the nature of the help from TFH cells and also the suppression capacity of TFR cells. We therefore used RNA sequencing (RNA-seq) to profile gene expression of TFH and TFR cells from mice sensitized at days 1 and 3 and days 1 and 8 (Table S1). In TFH cells, the expression of 1,483 out of 12,275 genes increased by at least 1.5-fold, while the expression of 1,787 genes out of 12,275 genes decreased by at least 1.5-fold in the days and 8 group compared with the days 1 and 3 group. In TFR cells, 1,263 genes out of 12,162 genes increased expression by at least 1.5-fold, while 3,213 genes out of 12,162 genes decreased in expression by at least 1.5-fold in the days 1 and 8 group compared with the days 1 and 3 group. We focused specifically on the known regulatory factors expressed by TFH and TFR cells that might regulate IgE switching and the survival and proliferation of IgE+ GC B cells. The expression of the majority of genes that play important roles in the function of TFH and TFR cells (signaling receptors Cxcr5, Cd40lg, Icos, Pdcd1; cytokines Il4, Il10, Il21, Tgfb1, Ifng; cytokine receptors Il1r2, Il1rn, Il2ra; inhibitory molecules Tigit, Fasl, Tnfsf10, Lag3, Ctla4, Gzmb) were unchanged between TFH and TFR cells from mice sensitized on days 1 and 3 and days 1 and 8 (Figures 6A-6C, S4A, and S4B). Analysis of IL-4-producing TFH cells using Il4 reporter mice (4get) showed separately that there is no difference in TFH-derived IL-4 between sensitizations at days 1 and 3 and days 1 and 8 (Figure S4C). It has been reported that TFR cells produce the inhibitory factor Neuritin (Nrn1) to repress IgE (Gonzalez-Figueroa et al., 2021). In our food-allergy model, we observed a 2-fold decrease in Nrn1 in TFR cells after sensitization at days 1 and 8 compared with at days 1 and 3 (Figures 6A-6C and S4D). Unexpectedly, we also observed a 2-fold decrease in Nrn1 in TFH cells after sensitization at days 1 and 8 compared with at days 1 and 3 (Figures 6A-6C and S4D). We therefore examined both TFH and TFR cells for changes in the expression of other suppressor genes. Fgl2 and Entpd1 are immune-suppressive factors typically associated with Tregs, but their function in regulating Ab production has not been extensively studied (Liu et al., 2017; Shalev et al., 2008; Wang et al., 2015). We observed that Fgl2 and Entpd1 had a marked decrease in expression in days 1 and 8 TFH cells compared with days 1 and 3 TFH cells (Figures 6A-6C). Fgl2 showed a 13.6-fold decrease in TFH cells and a 5.5-fold decrease in TFR cells from days 1 and 3 to days 1 and 8. (Figures 6A-6C). Independent qPCR data validated that the expression of Fgl2 and Entpd1 in TFH cells was significantly lower in the days 1 and 8 group than the days 1 and 3 group (Figure 6B). While the expression of Fgl2 was significantly lower in TFR cells from the days 1 and 8 group, Entpd1 was not significantly altered in TFR cells (Figures 6C and 6D). As expected, both Fgl2 and Entpd1 were more highly expressed in TFR cells compared to TFH cells (Figure 6E). Overall, the expression pattern we observed for Fgl2 and Entpd1 implied that they may play a suppressive role in regulating Ag-specific IgE in this food-allergy model.

Figure 6.

TFH and TFR gene expression is altered by the timing of antigen sensitization

(A) Heatmap of selected genes from bulk RNA-seq. TFH cells were sorted 4 days after PCT sensitizations at days 1 and 3 and days 1 and 8. Total RNA was extracted and subjected to bulk RNA-seq and analysis. The heatmap represents the values of log2FPKM.

(B) mRNA Expression of Fgl2 and Entpd1 in TFH cells after PCT at days 1 and 3 and days 1 and 8. Mice were sensitized with PCT at days 1 and 3 and days 1 and 8. TFH cells were sorted at the indicated time points. n = 12.

(C) TFR cells were sorted and used for bulk RNA-seq after PCT sensitizations at days 1 and 3 and days 1 and 8. Total RNA was extracted and subjected to bulk RNA-seq and analysis. The heatmap represents the values of log2FPKM.

(D) mRNA Expression of Fgl2 and Entpd1 in TFR cells after PCT at days 1 and 3 and days 1 and 8. Mice were sensitized with PCT at days 1 and 3 and days 1 and 8. TFR cells were sorted at the indicated time points. n = 10.

(E) Comparison of Fgl2 and Entpd1 mRNA expression in TFH and TFR cells. Expression is relative to beta-5 tubulin (Tubb5). Mice were sensitized with PCT at days 1 and 3 and days 1 and 8. TFR cells were sorted at the indicated time points. n = 4.

*p < .05 by t test. ns, not significant. Data are combined from three independent experiments (B–D) or representative of two or three independent experiments (E). Data are mean ± SEM.

CD39 and TFH-derived FGL2 inhibit Ag-specific IgE

We then tested whether the expression of Entpd1 was affected by the second sensitization. qRT-PCR data showed that Entpd1 gene expression is significantly down-regulated after the second sensitization in TFH cells but not in TFR cells (Figure S5A). To test whether CD39 regulates the development of Ag-specific IgE in our food-allergy model, a specific CD39 inhibitor (POM-1) was used. Flow cytometry showed that POM-1 treatment increased the percent of IgE+ GC B cells compared with control mice (Figure S5B), and we found that after PCT sensitization at days 1 and 8, POM-1 treatment led to slightly more PN-specific IgE (Figure S5C). PN-specific IgG1 was not different between control and POM-1-treated mice (Figure S5C).

We next tested whether the expression of Fgl2 was affected by the second sensitization. qRT-PCR data showed that Fgl2 gene expression is significantly down-regulated after the second sensitization in TFH cells but not in TFR cells (Figure 7A). To further investigate the role of Fgl2 in the development of Ag-specific IgE in our food-allergy model, we initially tested Fgl2 germline knockout (−/−) mice. Fgl2 mice showed higher titers of PN-specific IgE than wild-type (WT) mice but unchanged titers of PN-specific IgG1 (Figure 7B). Fgl2 mice also showed more TFH cells and GC B cells, but not TFR cells, than WT mice (Figure S6A). Fgl2 mice also showed decreased apoptosis of GC B cells than WT mice (Figure S6B). To see whether the deletion of Fgl2 and inhibition of CD39 could rescue IgE in timed sensitizations at days 1 and 3, we compared PN-specific IgE in WT mice, Flg2 mice, and Fgl2 with POM-1 treatment in mice that were sensitized with PCT at days 1 and 3. It was found that PN-specific IgE significantly increased in Fgl2 mice treated with POM-1 (Figure 7C), suggesting that combined Fgl2 deletion and CD39 inhibition partly reversed the Ag-specific IgE deficiency with sensitizations at days 1 and 3. A bone marrow (BM) chimera experiment was then designed to compare the role of Fgl2 from TFH cells and TFR cells versus Fgl2 from TFR cells alone. BM chimeras with Fgl2 BM plus Cd4-Cre Bcl6-flox BM allowed us to test the effect of loss of Fgl2 in both TFH and TFR cells, whereas BM chimeras with Fgl2 BM plus Foxp3-Cre Bcl6-flox BM allowed us to test the effect of loss of Fgl2 in TFR cells alone (Figure 7D). We found that mice with Fgl2-deficient TFH and TFR cells developed higher PN-specific IgE than control mice or mice with Fgl2-deficient TFR cells (Figure 7E). Ag-specific IgG1 and total IgE was also higher in mice with Fgl2-deficient TFH and TFR cells than in mice with Fgl2-deficient TFR cells (Figure S6C). These data imply that Fgl2 expressed by TFH cells is negatively regulating Ag-specific IgE. To study whether the pathogenic function of IgE was affected by Fgl2 deficiency in TFH cells, we tested the anaphylactic response in the BM chimeric mice. We found that anaphylaxis is much more severe, measured by temperature drop, in mice with Fgl2-deficient TFH and TFR cells than in control mice and mice with Fgl2-deficient TFR cells alone (Figure 7F). GC B cells were also higher in mice with Fgl2-deficient TFH and TFR cells than control mice or mice with Fgl2-deficient TFR cells alone (Figure 7G). Thus, FGL2 derived from TFH cells regulates the production of anaphylactic IgE in this food-allergy model. These data show that a regulatory factor expressed by TFH cells can specifically repress Ag-specific IgE production and that the expression of these factors is regulated by the timing of allergen sensitization steps.

Figure 7.

TFH-cell-derived FGL2 inhibits antigen-specific IgE

(A) mRNA expression of Fgl2 in TFH and TFR cells after PCT at the indicated time points. Mice were sensitized with one or two PCT. TFH and TFR cells were sorted at the indicated time points. n = 8 or 11.

(B) Peanut-specific IgE and peanut-specific IgG1 after PCT at days 1 and 8 in WT and Fgl2−/− mice. n = 4.

(C) PN-specific IgE in WT mice or Fgl2−/− mice with or without POM-1 treatment after PCT at days 1 and 3. n = 2–5.

(D) Scheme of bone marrow chimera mice.

(E) Peanut-specific IgE in bone marrow chimera mice after two PCT sensitizations. Rag1−/− recipient mice received bone marrow from WT or Fgl2−/− mice in combination with Cd4-Cre Bcl6-fl/fl mice or Foxp3-Cre Bcl6-fl/fl (Bcl6FC) mice. Bone marrow chimera mice were sensitized with PCT at days 1 and 8. PN-specific IgE was determined. n = 5 or 6.

(F) Anaphylaxis reaction in bone marrow chimera mice. Bone marrow chimera mice were injected with 2 mg peanut protein (i.p.) at day 15 after two PCT sensitizations. Body temperature was monitored for 40 min post peanut challenge. n = 5 or 6.

(G) GC B cells in bone marrow chimera mice after two PCT sensitizations. Cell number was calculated per mLN. n = 5 or 6.

*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 by t test (A), one-way ANOVA (C and E–G), or two-way ANOVA (B). ns, not significant. Data are combined from two independent experiments (A) or representative of two or three independent experiments. Data are mean ± SEM.

DISCUSSION

Here, we used a food-allergy model in mice to reveal several aspects of how Ag-specific IgE responses develop within the GC reaction and are regulated by TFH cells. We found that both the timing of allergen sensitization and the specific form of the Ag required to stimulate GC B cells were critical for the development of Ag-specific IgE but not for Ag-specific IgG1. The allergen-sensitization timing affects the expansion of Ag-specific GC B cells and affects the nature of the help from TFH cells. We further discovered that two genes, Fgl2 and Entpd1, which are regulated by the timing of allergen sensitization, play inhibitory roles in the development of Ag-specific IgE. Finally, we found that expression of Fgl2 specifically from TFH cells suppresses the Ag-specific IgE response and anaphylaxis.

Though IgE responses have been studied for decades, details about how the Ag-specific IgE response is initiated in vivo, particularly in GC, remains unknown. It is still not fully understood at what stage IgE class switching of Ag-specific B cells occurs in vivo and the complete sequence of signals involved (Satitsuksanoa et al., 2021). Here, we have described a detailed pathway for allergen-specific IgE production where a minimum of two allergen sensitization steps are required, in a very specific time frame. We showed that the two sensitization steps are necessary to expand significant numbers of Ag-specific GC B cells, including IgE+ Ag-specific GC B cells. Ig switching to IgE in GC occurs at a low level compared with switching to IgG1, which reflects the fact that IgE class switching is tightly controlled (Newman and Tolar, 2021; Wade-Vallance and Allen, 2021). One model for high-affinity IgE responses is that there is competition between IgG1 and IgE class switching by B cells within the GC. IgE promoting signals such as IL-4 are limiting factors, while IL-21 inhibits IgE switching (Wesemann et al., 2011; Wu et al., 2017; Yang et al., 2020). Our data support the model that the GC reaction plays a critical role in the development of Ag-specific IgE and that repeated sensitization promotes more IgE production in the GC. The consensus in the IgE field is that IgE+ B cells are not efficiently expanded and selected in the GC and thus do not generate high-affinity IgE production. In this model, high-affinity IgE is typically produced through sequential switching from IgG1+ B cells that have previously gone through affinity selection in the GC (Gould and Ramadani, 2015; He et al., 2015; Turqueti-Neves et al., 2015). The stage at which IgG1+ B cells switch to IgE is not well characterized, but some studies indicate that this occurs at a memory B cell stage (Jimenez-Saiz et al., 2018, 2019). Our findings with this food-allergy model do not fit with sequential switching to IgE from IgG1+ memory B cells, since a 7 day spacing in sensitization produced Ag-specific IgE, but a 15 day spacing did not produce Ag-specific IgE. If memory B cells were involved, the 15 day spacing should have worked as well or better to produce IgE than the 7 day spacing. Thus, our data suggest that Ag-specific IgE develops specifically within the GC in the food-allergy model. In addition, we found that Ag-specific IgE emerged as early as day 12, which is only 2 days after the Ag-specific IgG1, suggesting that IgE and IgG1 response develop nearly simultaneously within the GC. We have also found in this food-allergy model that high-affinity IgE, which can promote anaphylaxis, is produced as early as day 15 (Figure 1D). Whether sequential class switching of IgE through an IgG1 intermediate can occur efficiently within this time frame is not clear. We are currently studying whether high-affinity IgE can develop by direct switching in the GC or whether sequential class switching from IgG1 to IgE occurs within the GC.

One oddity we observed is the increase in total IgE level after just one sensitization, while ε GLT expression only increases after two sensitizations. Possible reasons for the discrepancy are that either the signal to induce the ε GLT is stronger at day 12 than day 8 and/or that an increase of IgE+ plasma cells giving rise to increased serum IgE is not captured in the total B cell population analyzed.

While we have increased our understanding of the role of the second allergen sensitization step in inducing IgE responses, some questions remain. We showed that CT is not required at the second sensitization, suggesting that tolerance is broken at the first sensitization and that only Ag is needed to stimulate B and T cells at the second sensitization. Unexpectedly, using B cell haptens coupled to different T cell Ags, we found that TFH cells at the second sensitization do not require the same Ag used in the first sensitization to help Ag-specific B cells. However, for both sensitization steps, the cognate B cell Ag needs to be linked to the cognate T cell Ag. The most likely explanation for these results is that newly generated TFH cells can enter pre-existing GCs to provide help to IgE+ B cells after the second sensitization. Previous work showed that newly generated TFH cells could seed ongoing GCs and provide help (Shulman et al., 2013), but here we show that this can occur in the context of an IgE response. A second explanation is that rather than newly generated TFH cells entering the GC, non-Ag-specific bystander TFH cells outside the GC can provide help to B cells if given the proper Ag stimulation (Ritvo et al., 2018; Wan et al., 2019). Future work is needed to understand if the help to IgE+ GC B cells is delivered by cognate TFH cells or from bystander TFH cells.

Another striking finding from our study is the differential regulation of Ag-specific IgE and Ag-specific IgG1. We observed that Ag-specific IgG1 is not sensitive to the timing of sensitization steps, unlike IgE. It has been reported that Ag-specific IgE, but not IgG1, is extremely sensitive to IL-4 levels (Robinson et al., 2017). Ag-specific IgE+ GC B cells may also be more sensitive to the levels of both positive and negative regulatory factors than IgG1. This hypothesis is consistent with our data showing that Ag-specific IgE, but not IgG1, is affected in Fgl2-deficient mice and CD39 inhibitor treatment. Additionally, Ag-specific IgE, but not IgG1, requires stimulation of pre-existing GC B cells at the second sensitization. Several papers have shown that IgE+ GC B cells display distinct features from IgG1+ GC B cells, including less surface immunoglobin receptor expression and greater propensity to undergo apoptosis or differentiate into short-lived plasma cells (Haniuda et al., 2016). The differential survival and differentiation capabilities of switched IgE+ GC B cells and IgG1+ GC B cells may contribute to different outcomes of Ag-specific IgE and IgG1 in our food-allergy model.

Several papers have shown that the Ag-specific IgE response is dependent on TFH and TFR cells (Ballesteros-Tato et al., 2016; Clement et al., 2019; Kobayashi et al., 2017; Noble and Zhao, 2016). The mechanism for how TFH and TFR cells regulate the IgE response is just starting to be understood and includes the IL-4, IL-13, IL-10, IL-21, and IL-2 signaling pathways (Dolence et al., 2018; Gowthaman et al., 2019; Meli et al., 2017; Xie et al., 2020; Yang et al., 2020). New molecules have recently been discovered that regulate the GC and IgE response. TFR cells produce the inhibitory factor Nrn1 to regulate IgE (Gonzalez-Figueroa et al., 2021). Interestingly, we also observed lower expression of Nrn1 in TFH and TFR cells after sensitization at days 1 and 8 compared with at days 1 and 3, supporting an inhibitory role of Nrn1 in regulating IgE. Here, we have identified two factors with differential expression between PCT sensitization at days 1 and 3 and days 1 and 8, Fgl2 and CD39, and their roles in regulating Ag-specific IgE were characterized. Previous work on the immunoregulatory role of Fgl2 has focused on its role in regulatory T cells (Tregs) and suppressing immune responses (Liu et al., 2017; Shalev et al., 2008; Wang et al., 2015). Most recently, Fgl2-derived from TFR cells was found to regulate the Ab response (Sungnak et al., 2020). Interestingly, these researchers found that the expression of Fgl2 is not detectable in TFH cells but is quite high in TFR cells (Sungnak et al., 2020). We have analyzed the GC and Ag-specific IgE response in Fgl2-deficient mice and conclude that TFH cell-derived, instead of TFR cell-derived, Fgl2 inhibits Ag-specific IgE production and the anaphylactic reaction in our food-allergy model. Unlike Ag-specific IgE, Ag-specific IgG1 was not affected by Fgl2 deficiency in mice in our food-allergy model. Whether Fgl2 regulates the GC response generally and affects IgE+ B cells more intensely, as these cells are more sensitive to regulatory signals, or whether Fgl2 more specifically regulates IgE+ GC B cells, needs future study. We note that the expression of Fgl2 and Entpd1 were also higher in TFH cells with an aberrant granzyme B+ phenotype, which correlated with poor Ag-specific IgE production (Xie et al., 2019). These data further suggest an inhibitory role of Fgl2 and Entpd1 in TFH cells. However, since we used a pharmaceutical CD39 inhibitor, we were unable to study the role of CD39 specifically in TFH or TFR cells. A mouse model with a specific deletion of Entpd1 or Fgl2 in TFH cells is not currently available. The key question of how the expression of Fgl2, Entpd1, and other inhibitory factors are regulated in TFH cells during the IgE response needs more investigation.

In this study, we observed that the timing of allergen sensitization determined whether specific IgE was produced or not, with too short of a window or too long of a window between allergen sensitizations leading to a failed IgE response. We showed that a short period of timing between allergen sensitization steps blocked development of IgE+ B cells and relates in part to expression of inhibitory factors, such as Fgl2 and Entpd1, by TFH cells. Whether human food allergy depends on a similar pattern of exposure to allergens is not clear. Overall, this study adds important details to our knowledge of how Ag-specific IgE responses are regulated in the development of food allergy, which may inform the human IgE response.

Limitations of the study

This study focused on how high-affinity IgE responses develop after gut sensitization using protein Ags plus CT as an adjuvant. This system is a widely used mouse model for food allergy but differs from the development of food allergy in humans, which is not associated with CT. We showed that in the optimal sensitization sequence for strong IgE responses, the immune-suppressive genes Fgl2 and Entpd1 are down-regulated in TFH cells. However, these are only a subset of genes and factors altered by the timing of Ag sensitization, and we have not characterized all factors regulated by the timing of Ag sensitization. We observed that genetic ablation of Fgl2 in both TFH and TFR cells, but not in TFR cells alone, increased Ag-specific IgE levels and IgE-mediated anaphylactic responses. Since we are unable to delete genes specifically in TFH cells without affecting TFR cells, we cannot rule out that TFR cells produce Fgl2 that influences IgE responses; however, loss of Fgl2 by TFR cells alone is not sufficient to significantly affect IgE responses. Furthermore, we do not have a genetic model to delete Entpd1 in TFH and TFR cells and had to rely on an inhibitor compound that may affect other cell types. Additionally, we do not know how applicable our results are to other model systems where IgE responses are induced. Finally, we do not yet know how our results relate to the human IgE response.

STAR★METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resource/reagents should be directed to and will be fulfilled by the lead contact, Alexander Dent (adent2@iupui.edu).

Materials availability

This study did not generate new unique reagents.

Data and code availability

Data availability. The RNA-seq datasets generated in this paper are available at the NCBI GEO database (GEO: GSE202713).

Code availability. This paper does not report original code.

Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Mouse strains

All mutant mice were on a C57BL/6 background. Foxp3-Cre Bcl6-flox (Bcl6FC), and Cd4-Cre Bcl6-flox (Bcl6 cKO) mice were described previously (Hollister et al., 2013; Wu et al., 2016). C.129-Il4tm1Lky/J (4get) were obtained from The Jackson Laboratory. IgE reporter mice Igh-7tm1.2Cdca (Verigem) were obtained from Dr. Chris Allen (UCSF) (Yang et al., 2012). Fgl2 knockout (Fgl2 −/−) mice were obtained from Dr. Gary A. Levy (University of Toronto) (Marsden et al., 2003). Six-to ten-week-old male and female mice were used for most experiments. Mouse littermate comparisons were used whenever possible. Control and experimental mouse cohorts were age and sex matched. Mice were bred under specific pathogen–free conditions at the laboratory animal facility of the Indiana University School of Medicine. All experiments and handling of animals were conducted according to protocols approved by the IACUC of the Indiana University School of Medicine.

METHOD DETAILS

Gut sensitization

Mice were deprived of food for 2 h, and then each mouse was fed 300 μL 1.5% NaHCO3 water (i.g.). Thirty minutes later, each mouse was given 1 mg peanut protein extract (Greer Laboratories), OVA (Sigma), KLH (Sigma), NP16-OVA (Biosearch), or NP18-KLH (Biosearch) together with 10 μg cholera toxin (Sigma) according to the setting of experiments (Li et al., 2000; Orgel and Kulis, 2018). The mice were sacrificed on the indicated days and the mesenteric lymph nodes (mLN) and spleen were harvested. Serum was collected.

Assessment of anaphylaxis

To assess anaphylaxis, 2 mg peanut (PN) without cholera toxin (CT) was injected i.p. per mouse at the indicated time points after sensitization. Mice were monitored for 40 min after challenge for the rectal (core) body temperature (Braintree Scientific).

Total B cell isolation

Total cells were harvested from mLN and spleen. B cells were isolated using B Cell Isolation Kit, mouse (Miltenyi) according to manufacturer’s instruction.

ELISA

For the measurement of PN-specific IgE, 96 well Nunc-Immuno plates (Sigma) were coated with 5 μg/mL IgE Ab (Clone: LO-ME-3, BIO-RAD) in 0.1 M Carbonate buffer (pH 9.5) overnight at 4°C. Wells were blocked with 1% BSA for at least 1 h at room temperature and diluted serum was added and incubated at room temperature for 2 h. Peanut protein extract was labeled with biotin (Sigma) and added into wells for one hour (2 μg/mL) followed by adding poly-HRP streptavidin (Pierce Endogen) for 0.5 h (1:2000). For PN-specific IgG1, 96 well Nunc-Immuno plates were coated with 5 ug/mL peanut protein extract in 0.1 M Carbonate buffer (pH 9.5) overnight at 4°C. Wells were blocked with 1% BSA for at least 1 h at room temperature and diluted serum was added and incubated at room temperature for 2 h. An anti-mouse IgG1 (Clone: A85-1, BD Pharmingen) was used as secondary Ab (2 μg/mL) followed by adding avidin-HRP (Invitrogen) for 0.5 h (1:2000). For total IgE, 96 well Nunc-Immuno plates were coated with 2 ug/mL anti-mouse IgE (Clone: R35-72, BD Pharmingen) overnight at 4°C. Wells were blocked with 1% BSA for at least 1 h at room temperature and diluted serum was added and incubated at room temperature for 2 h. An anti-mouse IgE (Clone: R35-118, BD Pharmingen) was used as secondary Ab (2 μg/mL) followed by adding avidin-HRP (Invitrogen) for 0.5 h (1:2000). After the incubation with HRP, TMB Substrate Reagent Set (BD Pharmingen) was added for the reaction development.

Flow cytometry

Cell suspensions from mLNs were prepared and filtered through a 40-μm cell strainer (Fisherbrand). Cells were washed and diluted in PBS with 1% FBS and were stained with Fc block (Biolegend) for 5 min, followed by surface staining for the indicated markers. Following labeled Abs were used: anti-CXCR5 (L138D7), anti-PD-1 (29F.1A12), anti-CD4 (RM4-5), anti-CD38 (90) and anti-B220 (RA3-6B2), anti-GL7 (GL7), anti-IgG1 (RMG1-1), anti-Foxp3 (MF-14) were obtained from BioLegend. For Annexin V staining, PE Annexin V (BioLegend) was added after surface staining using Annexin V binding buffer (BioLegend) according to the instruction. For NP-specific B cells staining, NP16-PE was added at the concentration of 5 nM and incubated on ice for 30 min. Fixation/Permeablization Kit (BD Pharmingen) was used for intracellular staining following surface staining. All samples were acquired on an LSR2 flow cytometer (Becton Dickinson) and analyzed with FlowJo V10.6 (TreeStar).

Q-PCR

The mRNA expression of cytokines, Fgl2, and Entpd1 were measured using Taqman Fast Advanced Master Mix (Thermo Fisher) in QuantStudio 6 and 7 Flex Real-Time PCR System (Thermo Fisher). The probes for Il4 (Mm00445259), Il21 (Mm00517640), Il10 (Mm00439614), Ifng (Mm01168134), Cd40lg (Mm00441911), Fgl2 (Mm00433327), Entpd1 (Mm00515448), Tubb5 (Mm00495806), Aicda (Mm01184115) were purchased from Thermo Fisher. Taqman probes for γ1-GLT (Probe: 5′-FAM-TGGTTCTCTCAACCTGTAGTC CATGCCA-3’, Forward: 5′-CGAGAAGCCTGAGGAATGTGT-3′, Reverse: 5′-GGAGTTAGTTTGGGCAGCAGAT-3’) and ε-GLT (Probe: 5′-FAM-CAGCCACTCACTTATCAGAG-3′, Forward: 5′-GCAGAAGATGGCTTCGAATAAGAACAGT-3’, Reverse: 5′-TCGTTGAATGATGGAGGATGTGTCACGT-3′) were purchased from Custom TaqMan Probes (Roco et al., 2019).

Bone marrow chimera

Bone marrow cells (2.5 × 106 each type) from donor mice were mixed and then injected i.v. into Rag1 −/− recipient mice which were sublethally irradiated (350 Gy). The lymphoid compartment in the recipients was allowed to constitute for 2 months before PCT sensitization.

QUANTIFICATION AND STATISTICAL ANALYSIS

RNA-seq and analysis

RNA-seq was performed by the Indiana University School of Medicine Center for Medical Genomics. Uniquely mapped sequencing reads were assigned to mm10 refGene genes. Quality control of sequencing and mapping results was summarized using MultiQC. Genes with read count per million (CPM) < 0.5 in more than 4 of the samples were removed. The data was normalized using TMM (trimmed mean of M values) method. Differential expression analysis was performed using edgeR. False discovery rate (FDR) was computed from p values using the Benjamini-Hochberg procedure. Differentially expressed genes (DEGs) were determined if their p-values were less than 0.05 after multiple-test correction with FDR-adjustment and the amplitude of fold changes (FCs) were larger than 1.8. The RNA-seq data can be accessed via NCBI GEO database (GEO: GSE202713).

Statistical analysis

All data analysis was performed using GraphPad Prism software (GraphPad Software). Graphs show the mean ± SEM. Unless otherwise stated, Two-tailed Student’s t test or One-way ANOVA with Tukey’s post hoc analysis was used. All ELISAs were analyzed using two-way ANOVA with Holm-Šidák multiple comparisons test. Significant differences (p < .05) and some non-significant differences are indicated in the figures. All the statistical details of experiments can be found in figure legends. The investigators were not blinded for the analyses.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
PE anti-mouse CD185 (CXCR5) Antibody	Biolegend	Cat#145504; RRID: AB_2561968
APC anti-mouse CD279 (PD-1) Antibody	Biolegend	Cat#135210; RRID: AB_2159183
PerCP/Cyanine5.5 anti-mouse CD4 Antibody	Biolegend	Cat#100540; RRID: AB_893326
Pacific Blue™ anti-mouse CD38 Antibody	Biolegend	Cat#102720; RRID: AB_10613468
PerCP/Cyanine5.5 anti-mouse/human CD45R/B220 Antibody	Biolegend	Cat#103236; RRID: AB_893354
APC anti-mouse/human GL7 Antigen (T and B cell Activation Marker) Antibody	Biolegend	Cat#144618; RRID: AB_2800675
FITC anti-mouse IgG1 Antibody	Biolegend	Cat#406606; RRID: AB_493293
Pacific Blue™ anti-mouse FOXP3 Antibody	Biolegend	Cat#126410; RRID: AB_2105047
PE Annexin V	Biolegend	Cat#640908; RRID: AB_2561298
Rat anti Mouse IgE Heavy Chain	BIO-RAD	Cat#MCA419
Biotin Rat Anti-Mouse IgG1	BD	Cat#553441; RRID: AB_394861
Purified Rat Anti-Mouse IgE	BD	Cat#553413; RRID: AB_394846
Biotin Rat Anti-Mouse IgE	BD	Cat#553419; RRID: AB_394850
Chemicals, peptides, and recombinant proteins
Peanut protein extract	Greer Laboratories	Patent: XPF171D3A25
OVA	Sigma	Cat#A5503
KLH	Sigma	Cat#H7017-20MG
NP-OVA	BIOSEARCH	Cat#N-5051-100
NP-KLH	BIOSEARCH	Cat#N-5060-25
NP-PE	BIOSEARCH	Cat#N-5070-1
Cholera toxin	Sigma	Cat#C8052-2MG
Critical commercial assays
B Cell Isolation Kit, mouse	Miltenyi	Cat#130-090-862
CD4+ T Cell Isolation Kit, mouse	Miltenyi	Cat#130-104-454
Nunc-immuno MicroWell 96 well solid plates	Sigma	Cat#M9410
Pierce Streptavidin Poly-HRP	Thermo Fisher Scientific	Cat#21140
Avidin HRP	Invitrogen	Cat#18-4100-94
TMB Substrate Reagent Set	BD	Cat#555214
Fixation/Permeabilization Solution Kit	BD	Cat#555028
Annexin V Binding Buffer	Biolegend	Cat#422201
TaqMan Fast Advanced Master Mix	Thermo Fisher Scientific	Cat#4444963
Deposited data
RNA-sequencing	This study	GSE202713
Experimental models: Organisms/strains
Mouse Foxp3-Cre Bcl6-flox (Bcl6FC)	Alexander Dent	Indiana University School of Medicine
Cd4-Cre Bcl6-flox (Bcl6 cKO)	Alexander Dent	Indiana University School of Medicine
Igh-7tm1.2Cdca (Verigem)	Chris Allen	University of California, San Francisco
Fgl2 knockout (Fgl2 −/−)	Gary A. Levy	University of Toronto
C.129-Il4tm1Lky/J(4get)	The Jackson Laboratory	Stock #004190
Oligonucleotides
Il4	Mm00445259	Thermo Fisher Scientific
Il21	Mm00517640	Thermo Fisher Scientific
Il10	Mm00439614	Thermo Fisher Scientific
Ifng	Mm01168134	Thermo Fisher Scientific
Cd40lg	Mm00441911	Thermo Fisher Scientific
Fgl2	Mm00433327	Thermo Fisher Scientific
Entpd1	Mm00515448	Thermo Fisher Scientific
Aicda	Mm01184115	Thermo Fisher Scientific
Tubb5	Mm00495806	Thermo Fisher Scientific
Software and algorithms
FlowJo (version 10.6)	TreeStar	https://www.flowjo.com/
Graphpad Prism (version 9)	GraphPad	https://www.graphpad.com/