DOCK8 and STAT3 cooperate to restrain IgE-inducing T follicular helper cells

Abstract

Patients with loss-of-function DOCK8 or dominant-negative STAT3 variants have hyper-IgE syndrome, although only DOCK8 deficiency consistently presents with elevated food-specific IgE and symptomatic allergy. We previously found in mice that DOCK8 restricts the differentiation of IL-13+ T follicular helper (Tfh13) cells that drive anaphylactic IgE, although the mechanisms were unclear. Here, we show that DOCK8 promotes STAT3 activity, which inhibits GATA3 in T cells. However, only patients with DOCK8, but not STAT3, deficiency had elevated Tfh13 cells. T cell-specific deletion of either Dock8 (T-Dock8^−/−) or Stat3 (T-Stat3^−/−) augmented peanut-specific IgE and Tfh13s when oral sensitization was promoted by adjuvants. However, the phenotypes diverged during adjuvant-free oral peanut exposure: only T-Dock8^−/− mice developed Tfh13 cells and peanut-specific IgE, accompanied by reduced Foxp3+ Tregs. Treg depletion in T-Stat3^−/− mice unmasked Tfh13 induction to oral antigen alone. Thus, DOCK8 and STAT3 cooperate to restrain Tfh13 differentiation to food allergens, and additional Treg impairment in DOCK8 deficiency allows for Tfh13 cell induction and allergy.

Introduction

IgE antibody-mediated food allergy is one of the most common forms of food allergy (Gowthaman et al., 2020; Tordesillas et al., 2017). A potentially life-threatening complication of allergy is anaphylaxis, a severe form of allergic reaction to otherwise innocuous food antigens, such as peanut (Bunyavanich and Berin, 2019). Results from a recent survey of more than 38,000 children in the U.S. indicate that nearly eight percent, which amounts to 5.6 million children nationwide, have food allergies (Gupta et al., 2018). Multiple studies including ours have shown that T follicular helper (Tfh) cells, but not Th2 effector cells, promote in vivo generation of allergen-specific IgE antibodies (Dolence et al., 2018; Gowthaman et al., 2019; Kobayashi et al., 2017; Lee et al., 2022; Meli et al., 2017; Zhang et al., 2020). We recently discovered that anaphylactic IgE induction to allergens depends on a novel subset of GATA3-expressing CD4⁺ Tfh cells, Tfh13 cells, which co-produce IL-4 and IL-13 and induce the production of pathogenic high-affinity, anaphylactic IgE antibodies to allergens (Gowthaman et al., 2019). Genetic ablation of Tfh13 cells abrogated anaphylactic IgE induction. Subsequently, many studies have reported the presence of Tfh13 cells in mice (Clement et al., 2019; Gonzalez-Figueroa et al., 2021; Haque et al., 2024; Immormino et al., 2024; Xie et al., 2020) and humans (Berin et al., 2022; Monian et al., 2022; Munemura et al., 2022; Tong et al., 2022; Wang et al., 2020). Importantly, peanut-specific Tfh13 cells were present in peanut-allergic donors but not in healthy subjects suggesting that Tfh13 cell induction is a critical checkpoint driving anaphylactic IgE and the allergic disease state (Gowthaman et al., 2019; Tong et al., 2022). Thus, understanding the molecular determinants of Tfh13 cell differentiation has important clinical implications.

Key factors influencing IgE-mediated allergies to food antigens include host genetics and environmental factors (Sampson et al., 2018). While inherited monogenic diseases leading to dysregulated hyper-IgE responses and allergies are rare, they can provide useful insights into the biology of immune cells, including Tfh13 cells. One such disease is Dedicator of Cytokinesis 8 (DOCK8) deficiency. Biallelic mutations resulting in the loss of function of the DOCK8 gene in humans lead to a combined immunodeficiency with hyper-IgE syndrome (HIES) (Su, 2023). Patients with DOCK8 deficiency exhibit not only increased susceptibility to infections but also allergic diseases including food allergies (Boos et al., 2014; Engelhardt et al., 2015; Haskologlu et al., 2020; Su et al., 2019). Using a mouse model of DOCK8 deficiency in T cells (T-Dock8^−/− mice), we previously demonstrated that the loss of DOCK8 in T cells causes aberrant induction of Tfh13 cells and IgE in mice administered non-allergic stimuli (such as lipopolysaccharide), suggesting that DOCK8 acts as a negative regulator of the Tfh13 cell differentiation program. However, the precise mechanisms by which DOCK8 regulates Tfh13 cells remain unknown. Additionally, it is unclear if the aberrant allergic responses to food antigens observed in patients with DOCK8 deficiency are primarily due to loss of DOCK8 in T cells.

Originally described as a guanine nucleotide exchange factor that regulates the actin cytoskeleton, recent evidence suggests that DOCK8 also amplifies cytokine-dependent Signal Transducer and Activator of Transcription 3 (STAT3) activation in human T cells(Keles et al., 2016). Experiments with primary human CD4⁺ T cells have demonstrated a direct interaction between DOCK8 and STAT3 (Keles et al., 2016), suggesting that DOCK8 may regulate T cell differentiation through modulating STAT3 signaling. Loss of DOCK8 in human T cells is associated with defects in STAT3 activation and decreased Th17 polarization (Keles et al., 2016; Tangye et al., 2017). STAT3 is a transcription factor activated downstream of cytokines such as IL-6, IL-21 and IL-23, which in T cells, synergize with T cell receptor signaling to promote Th17 differentiation (Mathur et al., 2007; Qin et al., 2024; Wei et al., 2007; Yang et al., 2007). Given that STAT3 dominant negative (DN) mutations also result in HIES (Tsilifis et al., 2021), we hypothesized that a common immune pathway in these disease-states might influence Tfh cell-driven IgE responses.

Using an oral peanut exposure model without additional adjuvants, we found that mice with isolated loss of DOCK8, but not STAT3, in T cells developed aberrant induction of Tfh13 cells and peanut-specific anaphylactic IgE. Congruent with these observations, humans with impaired DOCK8, but not STAT3, showed higher frequencies of circulating Tfh13 cells. Mechanistically, two checkpoints of Tfh13 cell suppression were impaired in T-Dock8^−/− mice. We discovered that DOCK8-dependent STAT3 activation is required to intrinsically suppress aberrant GATA3 expression in Tfh cells. Additionally, a secondary Tfh cell-extrinsic suppression mechanism mediated by regulatory T cells is also impaired in DOCK8 deficiency.

Results

Loss of DOCK8 in T cells enhances Tfh13 differentiation

We originally identified Tfh13 cells, which express the Tfh markers PD-1 and CXCR5 and co-produce IL-4 and IL-13 (Fig. S1), in mice lacking DOCK8 in T cells (Dock8^f/fCD4^Cre referred to as T-Dock8^−/−)(Gowthaman et al., 2019). Like patients with DOCK8 deficiency, T-Dock8^−/− mice exhibited heightened type 2 cytokines IL-4 and IL-13, Tfh13 cells and increased IgE production (Gowthaman et al., 2019; Su, 2023; Tangye et al., 2017). This allowed for the original observation of these cells uniquely involved in promoting allergic humoral responses in both mice and humans and suggested that DOCK8 plays an active, T cell-intrinsic role in suppressing Tfh13 differentiation.

To investigate further the mechanisms through which DOCK8 suppresses Tfh13 cells and IgE induction and to understand the mechanisms of heightened food allergy seen in DOCK8 deficient patients, we tested whether an adjuvant is necessary for T-Dock8^−/− mice to generate Tfh13 cells and allergen-specific IgE to food antigens. Previous experiments in wild type (WT) mice demonstrate that Tfh13 cells to food antigens, such as peanut, could be induced in the mesenteric lymph nodes (mesLN) only in the presence of the mucosal adjuvant cholera toxin (CT) but not when given oral peanut without an adjuvant (Gowthaman et al., 2019; Zhang et al., 2020). Indeed, the dominant response to ingested food proteins, in the absence of an adjuvant such as CT, is a tolerogenic response, which is achieved by mechanisms including induction of regulatory T cells (Tregs) and the generation of anergic antigen-specific T cells (Curotto de Lafaille et al., 2008; Hong et al., 2022; Mowat, 2018; Sampson et al., 2018). However, it is unclear if genetic predispositions for food allergy such as DOCK8 deficiency conform to these norms.

To test whether DOCK8 suppresses Tfh13 differentiation in the mesLN during unadjuvanted peanut immunization, we delivered ground peanut by oral gavage to T-Dock8^−/− mice or littermate controls. We assessed cytokine production and transcription factor expression in mesLN T cells at 8 days post-immunization (Fig. 1A and S1). In accordance with previous data (Zhang et al., 2020), Tfh cells in control mice given peanut alone did not co-produce IL-4 and IL-13; however, T-Dock8^−/− mice given peanut alone, in the absence of an adjuvant, generated IL-4⁺IL-13⁺ Tfh13 cells (Fig. 1B). In addition to increased type 2 cytokine production, Tfh cells in T-Dock8^−/− mice also expressed increased amounts of the Th2 lineage-defining transcription factor GATA3, but similar levels of Tfh master transcription factor BCL6, compared to littermate controls (Fig. 1C and S2A). Further, they also had increased frequencies of IL-4⁺IL-13⁺ Th2 effector cells (Fig. S2 B). Using this adjuvant-free model, we also assayed serum for peanut-specific IgE following six weeks of weekly oral administration of peanut. Since Tfh13 cells are necessary to produce high-affinity, anaphylactic IgE post-immunization, control mice predictably made no detectable peanut-specific IgE, but T-Dock8^−/− mice exhibited a robust, antigen-specific IgE response (Fig. 1D) and an increase in overall IgE production (Fig. 1E). Using a passive cutaneous anaphylaxis model, we found that the IgE generated to peanut could induce an anaphylaxis response (Fig. 1F). Accordingly, these antibody responses aligned with increased proportions of IgE⁺ plasmablasts in the mesLNs of T-Dock8^−/− mice, although germinal center (GC) B cell frequencies were similar between groups (Fig. S2 C–E). These data underscore the critical role of DOCK8 in restraining Tfh13 cell differentiation and IgE responses, as DOCK8 loss leads to aberrant Tfh13 differentiation and increased IgE production even in the absence of innate stimuli typically required for such responses. Furthermore, these results emphasize that DOCK8-mediated suppression of Tfh13 differentiation is operative across multiple tissue sites including the lung, as previously reported (Gowthaman et al., 2019), and as well as the gut.

Figure 1: DOCK8 restrains Tfh13 differentiation and peanut-specific IgE in mice given oral peanut in absence of adjuvant.

(A) Experimental workflow: T-Dock8^−/− (Dock8^f/fCD4^Cre/+) or floxed littermate controls were orally gavaged with 100 mg of ground peanut. For analysis of Tfh13 cells, mice received one peanut dose and were harvested 8 days post-immunization. For evaluation of antibody titers, mice were given 6 weekly peanut doses and sera were collected one week after the final immunization. (B) Frequency of Tfh2 (IL-4⁺) and Tfh13 (IL-4⁺IL13⁺) cells in the indicated parent population. (C) Mean Fluorescence Intensity of GATA3 in Tfh cells 8 days after a single peanut immunization. (D-E) Sera from 6X peanut-immunized mice were collected one-week post-immunization and assayed by ELISA for peanut-specific IgE (D) and total IgE (E). (F) Passive cutaneous anaphylaxis assay performed by transferring day 7 post-boost sera into naïve recipients and challenging with 100ug peanut extract and 1% Evans blue. Dye extravasation quantification is shown. In (B-F) each symbol indicates an individual mouse. Numbers in flow plots indicate percentages. Error bars indicate SD. Statistical tests: Unpaired t test (B-C); Mann-Whitney U test (D-F). For all data, *P<0.05, **P<0.01, ***P<0.001 and n.s is not significant. Data representative of three independent experiments. LOD; limit of detection.

DOCK8 and STAT3 cooperate to inhibit Tfh13 differentiation

Prior studies with primary human CD4⁺ T cells have demonstrated a direct interaction between DOCK8 and STAT3 (Keles et al., 2016). Furthermore, humans with STAT3 DN mutations (causing the most common form of autosomal dominant HIES) have defective induction of Th17 cells, but elevated type 2 cytokines and serum IgE (Ma et al., 2008; Mackie et al., 2023; Milner et al., 2008; Minegishi et al., 2007). Thus, these observations suggest that DOCK8 and STAT3 may work together to suppress the differentiation and/or function of proallergic T cells. To test whether murine DOCK8-deficient T cells exhibit defects in STAT3 signaling, we stimulated either control or T-Dock8^−/− T cells in vitro with IL-6 or IL-21 and assessed phosphorylation of STAT3 downstream of canonical cytokine signals. Consistent with previously published human data(Keles et al., 2016), DOCK8-deficient T cells failed to activate STAT3 to WT levels in response to either IL-6 or IL-21 (Fig. 2A–B). While STAT3 is the primary signaling pathway for both cytokines, IL-6 and IL-21 can also activate STAT1 and STAT5, albeit to a lesser extent (Heinrich et al., 2003; Leonard et al., 2008; Tormo et al., 2012; Wan et al., 2015). Notably, DOCK8 deficiency did not impair STAT1 or STAT5 phosphorylation in response to either cytokine (Fig. S3 A–D), indicating that DOCK8 specifically promotes STAT3 activation to IL-21 and IL-6 in primary CD4 T cells. We next examined if DOCK8 and STAT3 proteins physically interact with each other in murine T cells. Immunoprecipitation of DOCK8 from lysates of WT CD4⁺ T cells with an anti-DOCK8 antibody resulted in the co-immunoprecipitation of STAT3. Furthermore, a reciprocal immunoprecipitation with anti-STAT3 antibodies resulted in co-immunoprecipitation of DOCK8 validating DOCK8 protein as an interacting partner of STAT3 (Fig. 2C). Thus, like in humans (Keles et al., 2016), DOCK8 directly interacts with STAT3 to promote STAT3 signaling in murine T cells. As loss of STAT3 can lead to increased GATA3 expression in mesLN Tfh cells in mice (Wu et al., 2015), we reasoned that the type 2 bias observed in DOCK8-deficient T cells could be due to the loss of GATA3 inhibition mediated by the DOCK8-STAT3 axis. To address the downstream effects of STAT3 signaling on T cell differentiation and cytokine production, we cultured T cells from mice lacking T cell-intrinsic STAT3 (Stat3^f/fCD4^Cre referred to as T-Stat3^−/− mice) or floxed control mice in vitro under Tfh-like skewing conditions (Eto et al., 2011; Lu et al., 2011) with limiting amounts of IL-4. We observed that T cells lacking STAT3 produced significantly more IL-4 and IL-13 than WT T cells (Fig. 2D) and expressed higher levels of GATA3 (Fig. 2E) mirroring the phenotype we observed in T-Dock8^−/− mice. Together these data suggest that the predominant pathway by which DOCK8 suppresses Tfh13 cell generation is by promoting STAT3 activation.

Figure 2: DOCK8 promotes STAT3 phosphorylation in murine T cells to reduce type 2 Tfh differentiation in vitro.

(A-B) Phosphorylation of STAT3 in response to the indicated doses of IL-6 (A) or IL-21 (B) in CD4 T cells from T-Dock8^−/− (Dock8^f/fCD4^Cre/+) mice or littermate controls was assessed by intracellular flow cytometry and quantified by mean fluorescence intensity (MFI). (C) DOCK8 and STAT3 are constitutively associated in murine CD4⁺ T cells. Upper panel: DOCK8 (left) and STAT3 (right) immunoprecipitates derived from CD4⁺ T cell lysates from WT mice were blotted with anti-STAT3 or anti-DOCK8 antibodies. Lower panel: shows immunoblots for DOCK8, STAT3 and β-actin from input controls. (D-E) Naïve CD4 T cells were cultured from T-Stat3^−/− (Stat3^f/fCD4^Cre/+) or Stat3^f/f control mice in a “Tfh2 like” in vitro differentiation condition for 5 days. Intracellular expression of (D) IL-4 and IL-13 and (E) GATA3 expression is shown as flow cytometry histograms and bar graphs. Error bars represent S.D. Statistical tests: analysis of variance (ANOVA) followed by Šídák’s multiple comparisons test with a single pooled variance (A-B); Unpaired t test (D-E). For all data, *P<0.05, **P<0.01, ***P<0.001 and n.s is not significant. Data representative of three independent experiments.

Loss of T cell intrinsic STAT3 does not impair Tfh responses

Tfh cell differentiation is challenging to accurately replicate in vitro, so we tested whether loss of STAT3 signaling is permissive for in vivo Tfh13 cell differentiation in T-Stat3^−/− mice. STAT3 is important for several aspects of T cell differentiation, most notably that of Tfh and Th17 cells, both of which integrate IL-6 and IL-21 signals into their developmental programs(Eto et al., 2011; Zhou et al., 2007). Indeed, T-Stat3^−/− mice have reduced Th17 cells (RORγt⁺IL-17⁺) (Liu et al., 2008) and a decrease in RORγt⁺Tregs (RORγt⁺FOXP3⁺) at both steady state and after immunization (data not shown). As Tfh cell differentiation is also influenced by cytokines that signal through STAT3 (Crotty, 2019), we characterized Tfh cells in T-Stat3^−/− mice. Tfh cells in the mesLN were present in T-Stat3^−/− mice at normal frequencies at steady state in naïve mice and after immunization with peanut (Fig. S3E–G). This aligns with previous work showing that the requirement of STAT3 for Tfh cell differentiation appears to be model and site dependent (Choi et al., 2013; McIlwain et al., 2015; Wu et al., 2016; Wu et al., 2015). Loss of STAT3 in T cells does not impair gut Tfh populations, despite being necessary for Tfh development in the spleen following certain immunization models such as sheep red blood cells or lymphocytic choriomeningitis virus challenges as observed in some studies(McIlwain et al., 2015; Ray et al., 2014; Wu et al., 2016). Furthermore, the Tfh cells present in T-Stat3^−/− mice can support productive humoral responses in the gut, as T-Stat3^−/− mice immunized with peanut and CT showed no defects in peanut-specific IgG1 production, which is a Tfh dependent response (Fig. S3H) (Zhang et al., 2020). In addition, we observed that BCL6 expression in Stat3-deficient Tfh cells, though modestly reduced compared to Stat3-sufficient counterparts, remained substantially higher than in non-Tfh populations, indicating that their core Tfh identity is largely preserved (Fig. S3I). As such, we can effectively use this model to assess Tfh cytokine production and antibody outcome after oral administration of antigens.

Loss of STAT3 in vivo is insufficient to drive Tfh13 responses to oral antigens without adjuvant.

To test whether loss of T cell-intrinsic STAT3 can predispose cells to Tfh13 differentiation and phenocopy T-Dock8^−/− mice, we orally administered peanut in absence of any adjuvants to T-Stat3^−/− mice and littermate controls and assessed Tfh cytokine production (Fig. 3A). In contrast to our observations in T-Dock8^−/− mice (Fig. 1), loss of T cell-intrinsic STAT3 did not induce Tfh13 differentiation as measured by cytokine production (Fig. 3B). This discordance between the phenotypes that we observed in T-Dock8^−/− mice and T-Stat3^−/− mice suggests that there are additional factors independent of STAT3 that could inhibit Tfh13 differentiation in vivo. Although STAT3 loss was not sufficient to drive IL-4 or IL-13 expression, Tfh cells in T-Stat3^−/− mice given peanut expressed more GATA3 than their STAT3-sufficient counterparts (Fig. 3C), concordant with prior literature suggesting that STAT3 negatively regulates GATA3 in T cells (Hercor et al., 2017; Wu et al., 2015). Furthermore, T-Stat3^−/− mice also had increased frequency of Th2 effector cells (Fig. S3J), suggesting that any additional inhibitory factors suppressing Tfh13 differentiation are likely unique to the Tfh compartment and may not influence type 2 effector cell differentiation. Despite this, T-Stat3^−/− mice did not produce peanut-specific IgE following immunization, although the levels of total IgE in the serum from these mice were elevated (Fig. 3D–E). Taken together, these data support a partially redundant role for STAT3 in suppressing type 2 immune responses. Moreover, the contrasting elicitation of Tfh13 cells and peanut IgE responses in this model between the T-Dock8^−/− and the T-Stat3^−/− mice indicates that DOCK8-mediated suppression of Tfh13 differentiation may involve multiple mechanisms, only some of which operate through STAT3.

Figure 3: T cell specific loss of STAT3 is insufficient to drive Tfh13 responses in vivo to oral antigens without adjuvant.

(A) Experimental workflow: T-Stat3^−/− (Stat3^f/fCD4^Cre/+) or Stat3^f/f littermate controls were orally gavaged with 100 mg of ground peanut. For analysis of Tfh13 cells, mice received one peanut dose and were harvested 8 days post-immunization. For evaluation of antibody titers, mice were given 6 weekly peanut doses and sera were collected one week after the final immunization. (B) Frequency of Tfh2 (IL-4⁺) and Tfh13 (IL-4⁺IL13⁺) cells in the indicated parent population. (C) Mean Fluorescence Intensity of GATA3 in pregated Tfh cells 8 days after a single peanut immunization. (D-E) Sera from 6X peanut-immunized mice were collected one-week post-immunization and assayed by ELISA for peanut-specific IgE (D) and total IgE (E) In (B-E) each symbol indicates an individual mouse. Dotted line represents limit of detection (LOD). Numbers in flow plots indicate percentages. Error bars indicate SD. Statistical tests: Unpaired t test (B-C); Mann-Whitney U test (D-E). For all data, *P<0.05, **P<0.01, ***P<0.001 and n.s is not significant. Data representative of at least three independent experiments.

Donors with DOCK8 loss of function mutations but not STAT3 dominant negative mutations have increased Tfh13 cells.

The observation that mice lacking T cell-intrinsic DOCK8 or STAT3 have different propensities for Tfh13 differentiation and antigen-specific IgE production is not entirely surprising from a clinical viewpoint (Su, 2023; Su et al., 2019). Patients with DOCK8 deficiency or DN STAT3 variants are both well-known to have elevated levels of IgE, but their clinical phenotypes differ, particularly with respect to allergy and anaphylaxis(Boos et al., 2014; Su, 2023; Su et al., 2019; Tangye et al., 2017). Patients with DOCK8 deficiency present with recurrent sinopulmonary infections and viral skin infections such as human papillomavirus (HPV), poor responses to vaccines, and, in most patients, severe allergies, particularly food allergies (Aydin et al., 2015; Boos et al., 2014; Haskologlu et al., 2020; Su, 2023; Zhang et al., 2016). In contrast, patients with STAT3 DN mutations exhibit only modestly elevated levels of food allergy compared to the general population but are highly susceptible to bacterial and fungal infections due to their lack of Th17-driven immunity (Ma et al., 2008; Ma et al., 2015; Milner et al., 2008; Siegel et al., 2013). Prior analyses of peripheral blood T cells from donors with DOCK8 or STAT3 deficiency show increased production of Th2 cytokines and/or increased presence of circulating Th2 cells concomitant with decreased Th17 frequencies (Ma et al., 2015; Mackie et al., 2023; Su, 2023; Tangye et al., 2017). However, whether these patients have increased Tfh13 cells is unknown.

In our previous study, we observed that individuals with peanut-allergy with no known mutations in DOCK8 or STAT3 had increased circulating Tfh13 populations in the blood compared to non-allergic control subjects (Gowthaman et al., 2019). Given our observations that DOCK8 influences Tfh13 differentiation, at least partially, through STAT3, we sought to assess Tfh13 populations in patients with pathogenic mutations in either DOCK8 or STAT3. We obtained peripheral blood samples from donors with loss of function mutations in DOCK8, DN mutations in STAT3, and healthy controls and performed flow cytometry to assess T cell populations (Fig. 4A). The clinical characteristics of donors with DOCK8 deficiency or STAT3 DN mutations are summarized in Table 1 and detailed in Supplementary Table 1. We found a notable increase (~8 fold) in Tfh13 cells within the circulating Tfh cell population (CD3⁺ CD4⁺ CD45RA⁻ CXCR5⁺) in donors with DOCK8 deficiency compared to healthy, non-allergic controls (Fig. 4B–D). In contrast, donors with STAT3 DN mutations did not show a significant increase in Tfh13 cell frequency; however, there was a slight (~1.5 fold) elevation compared to healthy controls, which did not reach statistical significance (Fig. 4B–D). In comparison, our analysis of Th2 cells within the T effector cell population (CD3⁺ CD4⁺ CD45RA⁻ CXCR5⁻) showed that both donors with STAT3 DN mutations and donors with DOCK8 deficiency had significantly higher Th2 cell frequencies than healthy controls, affirming previous studies (Mackie et al., 2023; Tangye et al., 2017) (Fig. 4E–G). These results offer a mechanistic basis for the known phenotypes of DOCK8 and STAT3 deficiency: donors with DOCK8 deficiency have elevated Tfh13 populations, correlating with high rates of food allergy, while donors with STAT3 dominant negative mutations have low Tfh13 populations, correlating with lower rates of food allergy despite having high serum total IgE titers (Fig. 4B–D, Tables 1 and S1). Furthermore, the lack of Tfh13 cells in patients with STAT3 DN mutations support the conclusions from the previous mouse experiments suggesting that DOCK8-deficient T cells may have multiple defects resulting in aberrant Tfh13 differentiation.

Figure 4: Tfh13 cells are abundant in donors with DOCK8 loss of function mutations but not in donors with STAT3 dominant negative mutations, or healthy controls.

(A) Gating strategy to identify circulating Tfh13 and Th2 effector cells in donors with DOCK8 deficiency, STAT3 dominant negative mutations, or healthy control PBMCs. Rested CD4⁺ T cells were stimulated with or without PMA/ionomycin in the presence of GolgiPlug. Surface markers were stained followed by fixation, permeabilization and intracellular staining of cytokines. A representative sample from a donor with DOCK8 deficiency is shown. Unstimulated cells show background staining for IL-4 and IL-13 (far right panel). (B-D) Frequencies of circulating Tfh13 cells in healthy controls, donors with STAT3 dominant negative mutations (STAT3-DN) and donors with DOCK8 deficiency were enumerated based on intracellular IL-4 and IL-13 expression in circulating Tfh cells (cTfh cells; CD3⁺CD4⁺CD45RA⁻CXCR5⁺) cells. (B) Representative flow cytometry contour plots showing IL-4 and IL-13 co-expressing cTFh13 cells (C-D) Absolute frequencies of Tfh13 cells (C) and frequencies of Tfh13 cells normalized to healthy controls is shown as fold change (D). (E-G) Frequencies of circulating Th2 effector cells in healthy controls, donors with STAT3 dominant negative mutations (STAT3-DN) and donors with DOCK8 deficiency were enumerated based on intracellular IL-4 and IL-13 expression in circulating CD4 effector cells (CD3⁺CD4⁺CD45RA⁻CXCR5⁻) cells. (E) Representative flow cytometry contour plots showing IL-4 and IL-13 co-expressing Th2 effector cells. (F-G) Frequencies of Th2 within activated/memory non-Tfh cells (F) and frequencies of Th2 effectors normalized to healthy controls (G) are shown as bar graphs. Each symbol indicates an individual donor. Numbers in flow plots indicate percentages. Error bars indicate SD. Statistical tests: analysis of variance (ANOVA) followed by Dunnett’s multiple comparisons test with a single pooled variance. For all data, *P<0.05, **P<0.01, ***P<0.001 and n.s is not significant.

Table 1:

Clinical Characteristics of Study Participants

	DOCK8 Deficiency (N=6)	STAT3 Dominant Negative (N=12)	Healthy (N=13)
Age-Years mean (SD)	8.3 (5.2)	17.4 (11.9)	19.9 (17.1)
Sex Female N (%)	2 (33)	3 (25)	3 (40)
Total IgE (kU/L) mean (SD)	4256.5 (4287.8)	8682.2 (10,111.2)	44.3 (71.3)
Total eosinophils (cells/uL) mean (SD)	5606 (6831.8)	855.6 (514.8)	81.8 (40.1)
Food Allergy N (%)	4 (66.7)	4 (33)	0 (0)
Drug Allergy N (%)	0 (0)	2 (15.4)	0 (0)

Statistics reflect total N in each group unless otherwise specified. One participant with STAT3 DN was excluded in the total eosinophils calculation as labs were drawn after they had begun dupilumab.

Healthy control summary: Data from healthy participants was incomplete in some categories (see Table S1 for details); Data was summarized based on available information. N=10 for sex, N=7 for total IgE, N=6 for total eosinophils, and N=13 for food and drug allergy.

T-Dock8^−/− but not T-Stat3^−/− deficient mice have decreased Treg and Tfr frequencies.

Our data suggest that DOCK8 inhibits Tfh13 differentiation in multiple ways. One of these pathways appears to involve STAT3-mediated repression of GATA3, as loss of T cell-intrinsic STAT3 signaling results in increased GATA3 expression in Tfh-like cells both in vitro and in vivo (Fig. 2E, 3C). However, GATA3 upregulation was not sufficient for functional Tfh13 induction, as T-Stat3^−/− mice immunized with peanut do not have increased Tfh13 populations (Fig. 3B). These data suggest that Tfh13 differentiation is being repressed in T-Stat3^−/− mice, likely through a Tfh-extrinsic but T cell-intrinsic mechanism as Dock8 deletion in all T cells is sufficient to drive aberrant Tfh13 generation (Fig. 1B). Likely candidates for this T cell-mediated suppression of Tfh13 differentiation are FOXP3⁺ regulatory T cells (Tregs) and T follicular regulatory cells (Tfrs), which suppress inappropriate cellular and humoral responses, including Tfh13 cells (Berin, 2023; Clement et al., 2019; Gonzalez-Figueroa et al., 2021). Both cell types are important for tolerance to ingested antigens, and loss of these populations in various models is associated with increased rates of food allergy, IgE, and anaphylaxis (Graca et al., 2023; Kazmi and Berin, 2023; Noval Rivas et al., 2015; Tordesillas et al., 2017; Turner et al., 2020). Importantly, FOXP3⁺ T cells appear to be distinctly regulated by STAT3 and DOCK8. Indeed, in humans and mice lacking DOCK8, Treg populations are decreased in both prevalence and functionality (Janssen et al., 2017; Janssen et al., 2014; Singh et al., 2017). In contrast to DOCK8, decreased STAT3 function in humans or mice does not result in reduced Treg frequency, differentiation or function (Ma et al., 2008; Ma et al., 2015; Saito et al., 2011). Our analysis of Tregs in T-Stat3^−/− mice showed no defects in Tregs at homeostasis or after oral administration of peanut (Fig. 5A, B). In contrast, mesLN Treg frequencies in T-Dock8^−/− mice were reduced by about 30% in both cases (Fig. 5C, D). We also examined Tfr populations for defects since mice lacking Tfr cells frequently display increased susceptibility to allergic sensitization and anaphylaxis (Clement et al., 2019; Gonzalez-Figueroa et al., 2021; Koh et al., 2020). Like Tregs, Tfr populations were unperturbed in T-Stat3^−/− mice (Fig. 5E, F); however, T-Dock8^−/− mice had decreased Tfr cell frequencies compared to controls both at steady state and after oral peanut administration (Fig. 5G, H). The observed Treg and Tfr defects in mice lacking T cell-intrinsic DOCK8 but not STAT3 supports our hypothesis that tolerogenic regulatory cells form a Tfh cell-extrinsic brake on Tfh13 differentiation.

Figure 5: Regulatory T cells and T follicular regulatory cells are impaired by the loss of DOCK8 but not STAT3.

(A-D) Frequencies of FOXP3 expressing Tregs in mesLNs of (A) naïve T-Stat3^−/− (Stat3^f/fCD4^Cre/+) or Stat3^f/f littermate control mice, (B) peanut immunized T-Stat3^−/− or control mice, (C) naïve T-Dock8^−/− (Dock8^f/fCD4^Cre/+) or Dock8^f/f littermate control mice and (D) peanut immunized T-Dock8^−/− or control mice are shown as representative flow cytometry plots and bar graphs. (E-H) Frequencies of FOXP3 expressing T follicular regulatory cells in mesenteric lymph nodes of (E) naïve T-Stat3^−/− or control mice, (F) peanut immunized T-Stat3^−/− or littermate control mice, (G) naïve T-Dock8^−/− or littermate control mice and (H) peanut immunized T-Dock8^−/− or control mice are shown as representative flow cytometry plots and bar graphs. Each symbol indicates an individual mouse. Numbers in flow plots indicate percentages. Error bars indicate SD. Statistical tests: Unpaired t test. For all data, *P<0.05, **P<0.01, ***P<0.001 and n.s is not significant. Data representative of at least three independent experiments.

T-Stat3^−/− mice exhibit increased Tfh13 differentiation following adjuvanted immunization

To test whether overcoming oral tolerance is sufficient to drive exacerbated Tfh13 differentiation in T-Stat3^−/− mice, we utilized the mucosal adjuvant, CT (Fig. 6A). In our past studies, immunizing mice orally with peanut and CT can induce allergic sensitization in WT mice, something that is not achieved when immunizing with peanut alone (Berin and Mayer, 2009; Gowthaman et al., 2019; Li et al., 2000; Zhang et al., 2020). When we immunized T-Stat3^−/− mice with peanut and CT, we observed a significant increase in Tfh13 differentiation compared to littermate controls as measured by IL-4 and IL-13 expression as well as GATA3 expression (Fig. 6B, C). Additionally, peanut + CT-immunized T-Stat3^−/− mice had increased peanut-specific and total IgE production compared to controls (Fig. 6D, E). This is in stark contrast to our observation with unadjuvanted peanut immunization where we detected very few Tfh13 cells and no peanut-specific IgE (Fig. 3B, D). Notably, peanut + CT immunized T-Stat3^−/− mice had similar Treg and Tfr frequencies to controls (Fig S4A and B). However, mice immunized with PN+ CT had reduced Treg frequencies (WT 32.6±7.3; T-Stat3^−/− 34.7±6.5) compared to peanut alone (WT 38.3±1.8; T-Stat3^−/− 43.4±4.1) regardless of genotype (Fig. S4A). This suggests that CT can suppress regulatory T cell populations and aligns with our hypothesis that impaired Treg populations could permit Tfh13 differentiation. Unsurprisingly, when we immunized T-Dock8^−/− mice and littermate controls with peanut and CT, we also observed Tfh13 generation and IgE production in both control and T-Dock8^−/− mice, with a trend towards increased Tfh13 cells, GATA3 expression, and peanut-specific IgE production in T-Dock8^−/− mice (Fig. S4 C–G). In addition, Tfh cells from both T-Dock8^−/− and T-Stat3^−/− mice immunized with peanut and CT expressed higher BCL6 than non-Tfh populations, with only a slight reduction compared to WT Tfh cells (Fig. S4 H, I).

Figure 6: Loss of DOCK8 or STAT3 promotes Tfh13 differentiation and peanut-specific IgE production in mice given oral peanut and cholera toxin.

(A) Experimental workflow: T-Stat3^−/− (Stat3^f/fCD4^Cre/+) or floxed littermate controls (Stat3^f/f) were orally gavaged with 5 mg of ground peanut and 10μg of cholera toxin (CT). For analysis of Tfh13 cells, mice received one peanut+CT dose and were harvested 8 days post-immunization. For evaluation of antibody titers, mice were given 6 weekly peanut+CT doses and sera were collected one week after the final immunization. (B) Frequencies of Tfh2 (IL-4⁺) and Tfh13 (IL-4⁺IL13⁺) cells in T-Stat3^−/− mice in the indicated parent population. (C) Mean Fluorescence Intensity of GATA3 in pregated Tfh cells in T-Stat3^−/− mice 8 days after a single peanut immunization. (D-E) Sera from 6X peanut-immunized T-Stat3^−/− mice and floxed littermate controls were collected one-week post-immunization and assayed by ELISA for peanut-specific IgE (D) and total IgE (E). (F) Experimental workflow: mice lacking T cell-intrinsic STAT3 (Stat3^f/fCD4^Cre/+), Treg cells (Stat3^f/f Foxp3^DTR), or both (Stat3^f/fCD4^Cre/+ Foxp3^DTR) along with littermate controls (Stat3^f/f) were orally gavaged with 100 mg of ground peanut weekly for six weeks and received intraperitoneal doses of 1μg diphtheria toxin (DT) in each of the two days prior to the first and fourth peanut immunizations. For analysis of Tfh13 cells, mice received one peanut dose and were harvested 8 days post-immunization. For evaluation of antibody titers, mice were given 6 weekly peanut doses and sera were collected one week after the final immunization. (G) Frequencies of Tfh2 (IL-4⁺) and Tfh13 (IL-4⁺IL13⁺) cells in mice of the noted groups in the indicated parent population. (H) Sera from peanut-immunized mice were collected one week post final immunization and assayed by ELISA for peanut-specific IgE. (I) Passive cutaneous anaphylaxis assay performed by transferring day 7 post-boost sera into naïve recipients and challenging with 100ug peanut extract and 1% Evans blue. Dye extravasation quantification is shown. Each symbol on the graph indicates an individual mouse. Numbers in flow plots indicate percentages. Error bars indicate SD. Statistical tests: Unpaired t test (B, C); Mann-Whitney U test (D, E, I); analysis of variance (ANOVA) followed by Šídák’s multiple comparisons test with a single pooled variance (G); Kruskal-Wallis test followed by Dunn’s multiple comparisons test with the Stat3^f/f group as the reference (H). For all data, *P<0.05, **P<0.01, ***P<0.001 and n.s is not significant. Data representative of three (A-E) or two (F-I) independent experiments.

Taken together, our results reveal a Tfh-intrinsic checkpoint for preventing aberrant Tfh13 differentiation in the form of STAT3 signaling, which represses GATA3 upregulation in Tfh cells. However, a Tfh-extrinsic regulatory checkpoint in the form of Tregs and/or Tfr cells likely prevents Tfh13 differentiation even when STAT3 is lost. Only when both inhibitory mechanisms are impaired, as we observed in the T-Dock8^−/− mice or in mice where we break oral tolerance using an adjuvant, can Tfh13 cells fully differentiate, produce cytokines, and induce antigen-specific IgE responses. This led us to hypothesize that regulatory T cells (Tregs), including T follicular regulatory (Tfr) cells, may provide an extrinsic layer of suppression that restrains Tfh13 differentiation under non-inflammatory conditions in T-Stat3^−/− mice

Treg cells and Tfh-intrinsic STAT3 signaling redundantly suppress aberrant Tfh13 differentiation and IgE production

To directly test the role of Foxp3+ regulatory T cells in masking the intrinsic Tfh13 bias of STAT3-deficient T cells, we generated a model allowing acute and selective Treg depletion in a T-Stat3-deficient background. We crossed CD4-Cre Stat3^fl/fl with Foxp3-DTR lines to produce T-Stat3^−/− Foxp3-DTR mice, where diphtheria toxin (DT) administration acutely depletes Foxp3+ T cells (Fig. S4J). We treated T-Stat3^−/− Foxp3-DTR mice as well T-Stat3^−/−, Foxp3-DTR, and floxed controls with DT and then orally immunized mice with unadjuvanted peanut (Fig. 6F). As expected, neither the Stat3^fl/fl controls nor the T-Stat3^−/− mice exhibited increased Tfh13 differentiation (Fig. 6G). However, T-Stat3^−/− Foxp3-DTR mice exhibited a robust induction of both Tfh2 and Tfh13 cells, significantly exceeding the modest increase seen in Stat3^fl/fl DT-treated Foxp3-DTR controls (Fig. 6G). Furthermore, these mice mounted a strong peanut-specific IgE response, and using a passive cutaneous anaphylaxis assay, we confirmed that the IgE produced was capable of inducing anaphylaxis (Fig. 6H, I).

These data highlight the robust nature of immune tolerance mechanisms and underscore the complementary roles of intrinsic T cell signaling and regulatory networks in suppressing aberrant type 2 immunity to dietary antigens and restraining inappropriate Tfh13 differentiation. The DOCK8-STAT3 axis functions cell-intrinsically within T cells to limit GATA3 expression and Tfh13differentiation (Checkpoint 1), while Tregs provide extrinsic suppression to maintain immune tolerance to food in mucosal tissues (Checkpoint 2) (Model; Figure S5). The presence of either checkpoint is sufficient to block Tfh13 differentiation and IgE generation to food. Altogether, these results reinforce our model that both T cell–intrinsic STAT3 signaling and T cell–extrinsic regulatory mechanisms must be disrupted to unleash full Tfh13 differentiation and robust IgE sensitization to innocuous food antigens. However, adjuvants such as CT likely overcome these regulatory restraints by introducing inflammatory signals and transiently suppressing Treg function.

Discussion

Tfh13 cells are a type of group 2 T follicular helper cells that play a non-redundant role in eliciting anaphylactic IgE to allergens (Eisenbarth et al., 2021; Gowthaman et al., 2019; Lee et al., 2022). Immune mechanisms of Tfh13 cell regulation remain poorly understood. Inherited monogenic diseases like DOCK8 and STAT3 deficiency can cause dysregulated, hyper-IgE responses as well as food allergies, and they can provide significant insights into the biology of immune cells, including Tfh13 cells. Our study sheds light on the multiple mechanisms preventing inappropriate Tfh13 differentiation and the subsequent allergic responses, particularly focusing on the roles of DOCK8 and STAT3 in T cells.

Here we show that DOCK8 plays a crucial role in suppressing the differentiation Tfh13 cells to food antigens. This suppression is partially mediated through the activation of STAT3, which represses the expression of the transcription factor GATA3 in Tfh cells. In our experiments, we observed that the loss of DOCK8 in all T cells is sufficient for the aberrant induction of Tfh13 cells and subsequent IgE production to food antigens, even in the absence of adjuvants. This contrasts with WT mice, where additional adjuvants are necessary for Tfh13 differentiation and IgE induction (Gowthaman et al., 2019; Zhang et al., 2020). Mouse models of food allergic sensitization in the absence of adjuvants such as CT or staphylococcal enterotoxin B are rare (Schulke and Albrecht, 2019). While mouse models such as the collaborative cross mice strain CC027 do exist (Orgel et al., 2019), the T-Dock8^−/− mice (C57BL/6 background) are compatible with the extensive array of genetic tools available for the C57BL/6 strain, making them particularly advantageous for further research into allergic sensitization and immune response mechanisms.

Our in vitro studies further elucidated that DOCK8-deficient T cells exhibit impaired STAT3 activation in response to the cytokines IL-6 and IL-21, and that DOCK8 interacts with STAT3. Furthermore, T cells lacking STAT3 also showed increased IL-4 and IL-13 production and higher GATA3 expression in vitro, mirroring the T cell phenotype of T-Dock8^−/− mice and supporting the hypothesis that DOCK8 restrains Tfh13 cell differentiation by augmenting STAT3 signaling. Indeed, in human T cells, DOCK8 has been shown to stabilize STAT3 activation by likely shielding it from negative regulatory phosphatases (Keles et al., 2016). While our study shows that DOCK8 regulation of optimal STAT3 activation is critical to suppress GATA3 expression, how STAT3 achieves this inhibition is unclear. The role of STAT3 in regulating the type 2 program in Tfh and Th2 cells is an area of active investigation. Early reports suggest that STAT3 is required for optimal tissue Th2 effector function, particularly in a lung allergic inflammation model (Stritesky et al., 2011). Mechanistically, IL-10 driven STAT3 signaling upregulates BLIMP-1 expression, which is absolutely required for Th2 effector function in the lung to aeroallergens (He et al., 2020). However, this pathway is likely tissue specific, as loss of STAT3 or BLIMP-1 does not inhibit but rather augments Th2 effector cells and type 2 cytokines in the skin draining lymph nodes(He et al., 2020; Poholek et al., 2016). In line with this, we observe an increase in Th2 effector cells in the in the blood of STAT3 DN donors and in the mesLNs of T-Stat3^−/− or T-Dock8^−/− mice (Fig. 4E–G; S2B and S3J). Therefore, it is likely that additional microenvironmental cues dictate the requirement of STAT3 for Th2 function and may explain why patients with STAT3 DN mutations are prone to develop type 2 driven atopic dermatitis but not asthma as frequently (Mackie et al., 2023). In contrast to Th2 effectors, the role of STAT3 on the type 2 program for Tfh cells is less ambiguous. Like published reports (Hercor et al., 2017; Wu et al., 2015), we observed that the loss of STAT3 augments the expression of GATA3 and type 2 cytokines in Tfh cells. Given that Tfh cells are a separate lineage from Th2 cells, the former dependent on BCL-6 (Johnston et al., 2009) and the latter on BLIMP-1 (He et al., 2020), the role of STAT3 in modulating the type 2 program in these cell types could be lineage-dependent; contingent on how the STAT3 pathway intersects with BCL-6 and BLIMP-1 driven targets. Potential mechanisms by which STAT3 suppresses the type 2 program in Tfh cells include the direct regulation of Gata3 expression. For instance, in human T cells, STAT3 has been shown by ChIP-seq analyses to bind to the Gata3 locus (Tripathi et al., 2017). Another possible mechanism is the induction of Suppressor of Cytokine Signaling (SOCS) proteins, particularly SOCS1 and SOCS3. These proteins are actively induced downstream of STAT3 signaling and, in turn, inhibit Gata3 expression (Bachus et al., 2023; Zimmer et al., 2016).

Despite the effects of DOCK8 on STAT3 signaling, loss of STAT3 in T cells was not sufficient for Tfh13 cell differentiation or peanut-specific IgE production in vivo following oral peanut administration without an adjuvant. This suggests that DOCK8-mediated suppression of Tfh13 cell differentiation involves additional mechanisms beyond STAT3 signaling within Tfh cells. Our clinical data complement these findings, as donors with DOCK8 deficiency, but not STAT3 DN variants, showed significantly higher frequencies of circulating Tfh13 cells. This disparity aligns with the clinical observation that patients with DOCK8 deficiency have more severe allergic phenotypes, including food allergies, compared to patients with STAT3 deficiency (Avery et al., 2008; Boos et al., 2014; Ma et al., 2015; Milner et al., 2008; Siegel et al., 2013; Su, 2023; Su et al., 2019; Tangye et al., 2017). Conventional Tregs and Tfr cells play a key role in suppressing allergic responses and in the maintenance of oral tolerance (Gonzalez-Figueroa et al., 2021; Kazmi and Berin, 2023; Noval Rivas et al., 2015; Turner et al., 2020). Previous studies have noted that Treg development and function is impaired in DOCK8 deficiency (Janssen et al., 2017; Janssen et al., 2014; Singh et al., 2017). Indeed, DOCK8 positively regulates IL-2-induced STAT5 signaling in Tregs, which is important for their development, homeostasis, and function (Janssen et al., 2017). In contrast to DOCK8, STAT3 deficiency does not impair Treg differentiation(Ma et al., 2008; Ma et al., 2015; Mackie et al., 2023). Intriguingly, STAT3 and STAT5 are known to act as reciprocal inhibitors of each other (Olson et al., 2016; Yang et al., 2011). This could explain why patients with STAT3 gain-of-function mutations in fact show a reduction in Treg frequencies (Flanagan et al., 2014; Mackie et al., 2023; Milner et al., 2015; Tanita et al., 2021). In line with these findings, our analysis of T-Dock8^−/− mice at steady state or when administered unadjuvanted peanut extract, exhibited reduced frequencies of both Tregs and Tfr cells, whereas T-Stat3^−/− mice did not show such defects. Collectively, our findings suggest that STAT3 acts as a context-dependent, cell-intrinsic regulator of Tfh13 fate. Although STAT3-deficient CD4+ T cells exhibit a bias toward Tfh13 differentiation, this potential is restrained by intact regulatory T cells. Inflammatory cues—such as those elicited by CT—or transient Treg depletion are sufficient to unmask this Tfh13 potential and drive IgE production. In contrast, DOCK8 deficiency leads to spontaneous Tfh13 skewing, even in the absence of inflammation, likely due to combined effects on both effector T cells and the regulatory T cell compartment. Together, these data highlight that while both STAT3 and DOCK8 cooperate to prevent inappropriate type 2 immunity, they do so via complementary, but only partially overlapping mechanisms—STAT3 by restraining Tfh13 programming within CD4+ T cells, and DOCK8 by also preserving the regulatory cell networks that enforce peripheral tolerance.

Our experiments involving oral sensitization with adjuvants revealed that overcoming oral tolerance using CT can exacerbate Tfh13 cell differentiation in both T-Stat3^−/− and T-Dock8^−/− mice, leading to increased peanut-specific and total IgE titers. This underscores the critical role of oral tolerance and its disruption in allergic sensitization and anaphylaxis(Berin, 2023; Berin et al., 2022; Kazmi and Berin, 2023; Sampson et al., 2018). In addition to the induction of proinflammatory mediators (Datta et al., 2010; Kim et al., 2019; Stratmann, 2015; Terrinoni et al., 2019; Wakabayashi et al., 2018) (such as IL-1, ATP, HMGB-1, TSLP, IL-33 etc), a mechanism by which oral adjuvants could temporarily break tolerance is by Treg inhibition (Mattsson et al., 2015; Zhang et al., 2020). Indeed, HMGB-1, which is elicited by CT(Stratmann, 2015; Wakabayashi et al., 2018), suppresses Foxp3 expression and Treg function in a PTEN dependent manner(Ge et al., 2021; Zhou et al., 2019). Further, alarmins such as IL-33 and TSLP, which are elicited by adjuvants such as CT, have also shown to have a suppressive effect on Foxp3⁺ cells in certain allergic contexts (Chen et al., 2017; Chu et al., 2013; Kobari et al., 2020; Nguyen et al., 2010). Supporting this premise, we previously observed that oral peanut sensitization in presence of CT reduces Treg and Tfr frequencies in the mesLN (Zhang et al., 2020).

Thus, our work offers mechanistic insights into DOCK8’s function in restraining Tfh13 differentiation and helps explain the disparate clinical manifestations with respect to food allergies in STAT3 and DOCK8 loss-of-function patients. We show that DOCK8 and STAT3 act as part of the same molecular circuit and have similar roles for negatively regulating type 2 transcriptional programs in Tfh cells but have divergent roles in Treg function. These findings highlight the redundant regulatory mechanisms involving DOCK8 and STAT3 in maintaining immune tolerance and preventing aberrant Tfh13 cell differentiation. The synergistic roles of DOCK8 and STAT3 in regulating Tfh13 cells and IgE responses suggest potential therapeutic targets for managing food allergies and preventing anaphylactic reactions.

Methods

Human Samples

Heathy Donors: blood samples form healthy donors, without allergies and asthma, were obtained from pediatric pulmonary and pediatric primary care clinics at Connecticut Children’s Medical Center (CCMC) or Australian Red Cross Blood Service. The non-allergic phenotype was confirmed with a negative ImmunoCAP clinical laboratory test or negative skin prick testing to environmental allergens (Dermatophagoides farina, Dermatophagoides pteronyssinus, Alternaria alternata, Cladosporium herbarum, dog, cat, mugwort, mouse, rat, or cockroach). Participants with self-reported food allergies were also excluded from analysis.

Patients with loss-of-function variants in DOCK8 or STAT3: Blood was obtained from patients after receiving informed consent from patients and/or guardians. This study was approved by NIAID/NIH, USA IRB protocols (NCT00001467, NCT0000610, NCT00128973), the Sydney Local Health District RPAH Zone Human Research Ethics Committee and Research Governance Office, Royal Prince Alfred Hospital, Camperdown, NSW, Australia (Protocols X16–0210/LNR/16/RPAH/257, X16–0210 & 2019/ETH06359 and X20–0177 & 2020/ETH00998) and Connecticut Children’s Medical Center, USA (Protocols CCMC IRB# 15–007). Charts were reviewed for clinical allergy and laboratory assessments.

Mice:

Age and sex-matched 6–12 week-old mice were kept on peanut- and egg-free chow in specific pathogen-free conditions. C57BL/6N or C57BL/6J were purchased from Charles River Laboratories and Jackson Laboratories respectively. CD4^Cre [Tg(Cd4-cre)1Cwi/BfluJ] and Stat3^f/f [B6.129S1-Stat3tm1Xyfu/J] were purchased from Jackson Laboratories (Bar Harbor, ME). Dock8^f/f mice were generated and validated as described previously(Krishnaswamy et al., 2017). To generate T-Dock8^−/− or T-Stat3^−/− mice, CD4^Cre mice were crossed with Dock8^f/f or Stat3^f/f mice, respectively. All mice used were on the C56BL/6 background. All protocols used in this study were approved by the Institutional Animal Care and Use Committee at the Yale University School of Medicine, Northwestern University and University of Massachusetts Chan medical school.

Immunizations and depletions:

For peanut-only immunizations, mice were immunized by oral gavage either once or weekly with 100 mg of ground raw blanched peanut (Aiva Products) in 400μl of 0.2M sodium bicarbonate buffer per mouse. Peanut and CT immunizations were performed as described previously(Zhang et al., 2020). Briefly, mice were immunized by oral gavage either once or weekly with 5 mg of ground raw blanched peanut and 10μg of cholera toxin (List Biologicals) in 200 μl of sodium bicarbonate buffer per mouse. To deplete FOXP3 expressing cells, 1μg of diphtheria toxin was intraperitoneally injected into mice on day −2 and day −1 of experiments where immunization begins on day 0. For mesLN flow cytometry, mice were sacrificed at day 8. For antibody analysis, mice were treated a second time with the same dose of diphtheria toxin at week three and two days prior to the fourth immunization as FOXP3 expressing cells start to reemerge around 2–3 weeks of post diphtheria toxin treatment (Data not shown and (de Goer de Herve et al., 2012; Nystrom et al., 2014)). Sera were collected at week 6. In Treg cell depletion experiments, all mice, regardless of DTR expression, received diphtheria toxin injections to control for receptor-independent toxicity.

Flow Cytometry:

Mesenteric lymph nodes were extracted and ground between frosted glass slides to prepare single cell suspensions. For murine Tfh13 analysis, approximately ¼ of the cell suspension was stimulated and stained for intracellular cytokines as described previously(Chen et al., 2021; Gowthaman et al., 2019). For human Tfh13 analysis, PBMCs were isolated, enriched for CD4⁺ T cells, stimulated, and stained. A detailed protocol for mouse and human Tfh13 staining can be found here(Chen et al., 2021). Cell fixation and permeabilization for transcription factor staining was performed using the Foxp3 staining kit (eBiosciences) according to manufacturer instructions, following which cells were stained for transcription factors either at room temperature (RT) for 40 min or at 4°C overnight. IgE B cell staining was performed as described previously(Gowthaman et al., 2019). For the analysis of phosphorylation of STAT3, splenocytes from naïve mice were rested in 1%FCS RPMI media for 1hr at 37°C. Following which they were treated with IL-6 or IL-21 at indicated concentrations for 30 min. Cells were washed with ice-cold flow cytometry buffer and fixed immediately with cell fixation buffer (Biolegend). Cells were permeabilized with ice-cold methanol for 30 min and washed 3 times with flow cytometry buffer. Cells were then stained with pSTAT3 (Y705), pSTAT5 (Y694), pSTAT1 (Y701) and lineage markers including TCRβ, CD4, and B220 for 30 min in RT. All flow cytometry samples were acquired on Cytoflex (Beckman Coulter) or Symphony (BD Biosciences) flow cytometers and analyzed using FlowJo software (Version 10.10.0, TreeStar). Antibodies and dyes used in our experiments are listed in Supplementary Table 2.

In vitro type 2 Tfh-like cell culture:

Culture conditions for in vitro Tfh2-like T cell differentiation were adopted from the culture conditions for Tfh like cells(Eto et al., 2011; Lu et al., 2011) with the addition of IL-4. Briefly, 4×10⁵ isolated CD4 T cells/mL were cultured on anti-CD3 (5 μg /mL) coated 24 well plates with soluble anti-CD28 (Clone 37.51; 2μg /mL), anti-IFNγ (XMG1.2; 4μg /mL), Anti-IL-12 (C17.8; 2μg/mL), anti TGFβ (TW7–28G11; 2μg/mL), recombinant IL-2 (25u/mL), IL-4 (2ng/mL), IL-6 (20ng/mL) and IL-21 (20ng/mL) for three days. After three days, the cells were transferred with the culture media to a non-CD3 coated plate for an additional two to three days and then restimulated with PMA and Ionomycin for cytokine analysis. For transcription factor analysis, unstimulated cells were used after 5 days of culture.

Enzyme linked immunosorbent assay (ELISA):

Sera from mice immunized 6 times was collected one week following the final immunization. Briefly, 20μg/mL crude peanut extract (Greer Laboratories) or anti-mouse IgE (BD Pharmingen 553413) were used to coat 96-well Maxisorp plates (Thermo Fisher Scientific) for one hour at 37°C. Plates were blocked in 1% BSA in PBS for one hour at 37°C, after which serially diluted serum samples were added to the plate and incubated overnight at 4°C. After washing in PBS/0.5%Tween, bound antibodies were incubated with HRP-conjugated goat-anti-mouse antibodies targeting either IgG1 or IgE (SouthernBiotech) for one hour at 37°C and developed with stabilized chromogen tetramethylbenzidine (Life Sciences). Development was stopped with 3N hydrochloric acid before reading at 450 nm on a microplate reader (Molecular Devices). Quantification of Total IgE was performed by comparing samples to a Purified Mouse IgE standard (BD Pharmingen 557079) serially diluted on the ELISA plate with the top dilution at 100 ng/mL. Peanut-specific antibodies of each isotype were obtained from pooled serum or stool supernatant from hyperimmunized mice and used as standards for peanut-specific ELISAs to generate Arbitrary Unit (AU) values.

Passive cutaneous anaphylaxis:

Passive cutaneous anaphylaxis assays were performed as described earlier(Gowthaman et al., 2019; Xiong et al., 2012). Briefly, 20 μL serum from immunized and boosted mice were injected into ear pinnae of naïve B6 recipients and 24 hours later, the recipient mice were challenged with 100ug crude peanut extract (Greer Laboratories) together with 1% Evans blue (i.v.). Thirty minutes later, the mice were euthanized, and ear pinnae were harvested and incubated in formamide for 72 hours at 56°C to release the dye. The extent of vascular leakage was measured spectrophotometrically at 650 nm using Evans blue as standard.

Immunoprecipitation and Immunoblotting:

Total CD4+ T cells were isolated from the spleens of 4-week-old C57BL/6J mice using the Dynabeads Untouched Mouse CD4 cells kit (Thermo Scientific). Then cells were lysed with ice-cold RIPA buffer containing protease inhibitors (Thermo Scientific). Lysate supernatant was incubated overnight at 4°C with anti-DOCK8, anti-STAT3, or isotype control antibodies. Immunoprecipitation was performed by adding 20 μL of Protein G agarose beads (Active Motif) for 2 hours at 4°C. The bound proteins were eluted and denatured with Laemmli buffer at 98°C for 10 minutes. Proteins were then separated by SDS-PAGE with protein ladder, transferred to a PVDF membrane, and blocked with 1X TBS, 0.1% Tween-20, and 5% non-fat dry milk for 1 hour at room temperature. The membrane was incubated overnight at 4°C with anti-DOCK8 or anti-STAT3 antibodies, followed by washing with TBST and probing with HRP-conjugated anti-mouse IgG secondary antibody (Cell Signaling). Protein bands were visualized using Pierce^™ ECL Western Blotting Substrate (Thermo Scientific).

Intracellular detection of DOCK8 for determination of reversion status:

Thawed PBMCs were stained with fluorochrome-conjugated antibodies to a panel of lymphocyte subset markers including BV-510 anti-CD3 (clone OKT3 and BUV737 anti-CD19 (clone SJ25C1) on ice for 30 minutes. Stained cells were fixed with 4% paraformaldehyde for 15 minutes at room temperature. Fixed cells were resuspended in Permeabilization Medium (Medium B, Invitrogen) containing Alexa Fluor488 conjugated anti-DOCK8 monoclonal antibody (G2, Santa Cruz Biotechnology) and incubated 1 hour at room temperature at dark. Cells were washed and acquired on a BD LSRFortessa using FACSDiva software. Analyses were performed using the FlowJo software (version 10, BD).

Supplementary Material

ALL Supplementary Figures (Supplementary Fig. S1-S5) and Tables (table S1-S2)

Figure S1: Representative gating strategy for identifying Tfh2 and Tfh13 cells in mesLN

(A) Mice were immunized with peanut+CT and Tfh subsets were identified in mesLN post restimulation (PMA+ionomycin) by flow cytometry. See Supplementary Table 2 for flow cytometry antibody panels.

Figure S2: B cell responses in T-Dock8^−/− mice post oral peanut sensitization in the absence of adjuvant

T-Dock8^−/− (Dock8^f/fCD4^Cre/+) or floxed littermate controls were orally gavaged with 100 mg of ground peanut and mesLN were harvested 8 days post-immunization. (A) Mean Fluorescence Intensity of BCL6 in pregated nonTfh cells (Live TCRβ⁺CD4⁺CD44⁺PD1⁻CXCR5⁻) and Tfh cells (Live TCRβ⁺CD4⁺CD44⁺PD1⁺CXCR5⁺) in T-Dock8^−/− mice 8 days after a single peanut immunization. (B) IL-4 and IL-13 expression in non-Tfh effector T cells (Th2) from mesLNs of immunized T-Dock8^−/− mice. (C) Germinal center B cell frequencies, (D) IgE⁺ germinal center B cells, and (E) IgE⁺CD138⁺ plasmablasts in the mesLN were identified by flow cytometry. Each symbol indicates an individual mouse. Numbers in flow plots indicate percentages. Error bars indicate SD. Statistical tests: Analysis of variance (ANOVA) followed by Šídák’s multiple comparisons test with a single pooled variance (A), Unpaired t test (B-E). For all data, *P<0.05, **P<0.01, ***P<0.001 and n.s is not significant. Data representative of three independent experiments.

Figure S3: DOCK8 regulation of STAT signaling and Gut Tfh cell differentiation and antibody responses in T-Stat3^−/− mice. (A-D) Phosphorylation of STAT5 (A,C) or STAT1 (B,D) in response to the indicated doses of IL-6 (A-B) or IL-21 (C-D) in CD4 T cells from T-Dock8^−/− (Dock8^f/fCD4^Cre/+) mice or littermate controls was assessed by intracellular flow cytometry and quantified by mean fluorescence intensity (MFI). (E-J) Tfh cell frequencies in the mesLN of T-Stat3^−/− (Stat3^f/fCD4^Cre/+) or floxed littermate control mice: (E) at steady state; (F) day 8 post oral peanut immunization; (G) day 8 post peanut + CT immunization. (H) Analysis of serum peanut-specific IgG antibodies after 6 immunizations with peanut + CT. (I) Mean Fluorescence Intensity of BCL6 in pregated nonTfh cells (Live TCRβ⁺CD4⁺CD44⁺PD1⁻CXCR5⁻) and Tfh cells (Live TCRβ⁺CD4⁺CD44⁺PD1⁺CXCR5⁺) in T-Stat3^−/− mice 8 days after a single peanut immunization. (J) IL-4 and IL-13 expression in Th2 cells from mesLNs of peanut immunized T-Stat3^−/− mice. Each symbol indicates an individual mouse. Numbers in flow plots indicate percentages. Error bars indicate SD. Statistical tests: Unpaired t test (A-G, J); Mann-Whitney U test (H), Analysis of variance (ANOVA) followed by Šídák’s multiple comparisons test with a single pooled variance (I). Data representative of three independent experiments. For all data, *P<0.05, **P<0.01, ***P<0.001 and n.s is not significant. Data representative of two (A-D) or three (E-J) independent experiments.

Figure S4: Loss of DOCK8 or STAT3 promotes Tfh13 differentiation and peanut-specific IgE production in mice given oral peanut and cholera toxin. (A) Frequencies of FOXP3 expressing Tregs in mesLNs of T-Stat3^−/− (Stat3^f/fCD4^Cre/+) mice or littermate controls after a single peanut only or peanut + CT immunization. (B) Frequencies of FOXP3 expressing T follicular regulatory cells in mesenteric lymph nodes of peanut + CT immunized T-Stat3^−/− or control mice. (C) Experimental workflow: T-Dock8^−/− (Dock8^f/fCD4^Cre/+) or floxed littermate controls (Dock8^f/f) were orally gavaged with 5 mg of ground peanut and 10μg of cholera toxin (CT). For analysis of Tfh13 cells, mice received one peanut+CT dose and were harvested 8 days post-immunization. For evaluation of antibody titers, mice were given 6 weekly peanut+CT doses and sera were collected one week after the final immunization. (D) Frequencies of Tfh2 (IL-4⁺) and Tfh13 (IL-4⁺IL13⁺) cells in T-Dock8^−/− mice in the indicated parent population. (E) Mean Fluorescence Intensity of GATA3 in pregated Tfh cells in T-Dock8^−/− mice 8 days after a single dose of peanut+CT immunization. (F, G) Sera from 6X peanut-immunized T-Dock8^−/− mice and floxed littermate controls were collected one-week post-immunization and assayed by ELISA for peanut-specific IgE (F) and total IgE (G). (H, I) Mean Fluorescence Intensity of BCL6 in pregated non-Tfh and Tfh cells in T-Dock8^−/− (H) and T-Stat3^−/− (I) mice 8 days after a single dose of peanut + CT immunization. (J) Frequencies of FOXP3 expressing Tregs in mesLNs of Stat3^f/fCD4^Cre/+, Stat3^f/f Foxp3^DTR, Stat3^f/fCD4^Cre/+ Foxp3^DTR, and Stat3^f/f mice lines treated with diphtheria toxin (DT). Mice were dosed intraperitoneally with diphtheria toxin (DT) as in Figure 6F and were assessed 6 days after the last DT injection for the presence of Treg cells. Each symbol indicates an individual mouse. Numbers in flow plots indicate percentages. Error bars indicate SD. Statistical tests: Unpaired t test (A,B, D, E); Mann-Whitney U test (F,G); Analysis of variance (ANOVA) followed by Šídák’s multiple comparisons test with a single pooled variance (H-J). For all data, *P<0.05, **P<0.01, ***P<0.001 and n.s is not significant. Data pooled from three experiments (A) or representative of three (B, D-I) or two (J) independent experiments.

Figure S5: A two-hit model for anaphylactic IgE induction to food antigens in DOCK8 deficiency

Two independent mechanisms restrain aberrant induction of Tfh13 cell induction to nominal food antigens. As a Tfh-intrinsic checkpoint, a DOCK8-STAT3 axis plays a crucial role in restraining aberrant Tfh13 differentiation by suppressing GATA3 upregulation. In addition, a DOCK8-dependent, Tfh-extrinsic regulatory checkpoint involving Tregs and/or Tfr cells prevents Tfh13 differentiation, even in the absence of STAT3. Full differentiation of Tfh13 cells, production of cytokines, and induction of antigen-specific IgE to nominal food antigens occur when both inhibitory mechanisms are compromised.

The manuscript findings are supported by supplementary figures and tables. Fig. S1 shows the representative gating strategy for identifying Tfh2 and Tfh13 cells in mesLN. Fig. S2 shows B cell responses in T-Dock8^−/− mice post oral peanut sensitization in the absence of adjuvant. Fig. S3 shows DOCK8 regulation of STAT signaling and Gut Tfh and Th2 cell differentiation and antibody responses in T-Stat3^−/− mice. Fig. S4 shows Treg and TFR cell frequency in T-Stat3^−/− mice, Tfh cell and antibody responses in T-Dock8^−/− mice with peanut + cholera toxin oral sensitization and Treg depletion efficacy in Foxp3 DTR mice. Fig. S5 is a schematic representation of our proposed two-hit model for anaphylactic IgE induction to food antigens in DOCK8 deficiency. Table S1 contains detailed clinical characteristics of donors with DOCK8 deficiency or STAT3 DN mutations. Table S2 contains the details of all the antibodies and dyes used in the manuscript.

Non-standard abbreviations:

DOCK8

Dedicator of cytokinesis 8

STAT3

Signal transducer and activator of transcription 3

GATA3

GATA-binding protein 3

Tfh

T follicular helper Cell

Tfh13

T follicular helper 13 Cell

Tregs

Regulatory T cells

CT

Cholera Toxin

PCA

Passive Cutaneous Anaphylaxis

Data availability:

All the data presented in the manuscript are available in the manuscript itself, its supplementary materials, and the source data files.