DOCK8 in T cells promotes Th17 and Treg cell functionality to restrain mucosal mast cells and limit susceptibility to oral anaphylaxis

SUMMARY

Immunoglobulin E (IgE)-mediated release of mediators from mast cells (MCs) drives food allergy, and intestinal MC load is an important determinant of disease severity. Dedicator of cytokinesis 8 (DOCK8)-deficient patients are highly susceptible to food allergy. We found that they exhibited elevated serum MC tryptase levels, suggesting increased MC load. Dock8^−/− mice also had exaggerated IgE-mediated oral anaphylaxis, expansion of jejunal mucosal MCs (MMCs), and elevated serum levels of MMC-derived tryptase. This resulted in increased intestinal permeability, which promoted antigen absorption and thereby oral anaphylaxis. Mechanistically, these events were driven by an intestinal cascade in which reduced interleukin (IL)-17 cytokines led to dysbiosis, which drove IL-25 production. Increased IL-25 enhanced T helper (Th)2 production of IL-4 that expanded MMCs and exaggerated oral anaphylaxis. Furthermore, the failure of DOCK8-deficient T regulatory (Treg) cells to suppress intestinal IL-4 production and MC expansion left the exaggerated anaphylaxis unrestrained. These results suggest multi-faceted coordination between the microbiome, mucosal T cells, and MCs to restrict oral anaphylaxis.

In brief

DOCK8-deficient patients and mice are prone to food allergy and oral anaphylaxis. Here, Janssen et al. demonstrate that DOCK8 deficiency yields expanded mucosal mast cells and elevated circulating tryptase concentrations. Loss of DOCK8 in T cells impairs Th17 and Treg cells, resulting in dysbiosis and unrestrained IL-4-driven mast cell expansion. Disruption of this cascade attenuates oral anaphylaxis.

Graphical Abstract

INTRODUCTION

Food allergy affects 6%–8% of children and 3% of adults in the US.^1–3 Exposure to food allergens through a disrupted skin barrier or a dysbiotic gut drives immunoglobulin E (IgE) sensitization to foods.^4–6 The effector phase in food allergy depends on mast cell (MC) degranulation following antigen recognition by IgE antibodies (Abs) bound to high-affinity immunoglobulin epsilon receptor I (FcεRI). An important determinant of the severity of food anaphylaxis is intestinal MC load.^7–9 MC mediator release in the gut increases intestinal permeability and thereby promotes intestinal absorption of antigen and subsequent increased systemic MC activation and anaphylaxis.¹⁰ An important determinant of the severity of food anaphylaxis is intestinal MC load.^7–9 Determining the factors that drive intestinal MC load and activity is key to better understanding and treatment of food allergy.

The intestinal microbiome plays a role in food allergy.^11–13 Specific taxa of gut bacteria are associated with food allergy in human infants. Administration of such bacteria promotes IgE Ab production and anaphylaxis in mouse models of food allergy, whereas administration of bacteria from healthy infants is protective.^14,15 Cytokines such as interleukin (IL)-17A and IL-22 in the gut are induced by specific members of the microbiome.^16–18 They, in turn, act to promote antimicrobial gene expression by intestinal epithelial cells (IECs).¹⁹ Lack of IL-22 or IL-17A is associated with compositional changes in the microbiome i.e., dysbiosis.^20–22

DOCK8 is a nucleotide guanine exchange factor (GEF) for the small GTPase cell division control protein 42 (CDC42), which activates Wiskott-Aldrich syndrome protein (WASP) to cause actin polymerization, and subsequent cell motility, positioning, and function.^23–25 DOCK8 promotes signal transducer and activator of transcription 3 (STAT3) activation and is important for retinoic acid-related orphan receptor gamma t (RORγt) expression required for the development and survival of RORγt⁺ group 3 innate lymphoid cells (ILC3s) and T helper (Th)17 cells.^26,27 DOCK8 also promotes IL-2-driven STAT5 activation and thereby T regulatory (Treg) cell function and stability.^23,28

DOCK8 deficiency is characterized by recurrent infections, eczema, CD4⁺ T cell lymphopenia and decreased numbers of circulating Treg cells.^29,30 The vast majority of DOCK8-deficient patients (75%–85%) have IgE sensitization to food,^31,32 and as many as 50% experience food-induced anaphylaxis.³³ This occurs at a substantially higher incidence than in other individuals who have evidence of food sensitization, including patients with atopic dermatitis.³⁴

Here, we report elevated serum levels of MC tryptase in DOCK8-deficient patients. Furthermore, Dock8^−/− mice exhibited intestinal MC expansion, elevated serum levels of the mucosal MC (MMC)-derived tryptase, mast cell protease-1 (MCPT-1), increased intestinal permeability, and exaggerated oral anaphylaxis. These events were driven by an intestinal cascade in which dysbiosis with reduced intestinal IL-17 cytokines drove IL-25 production and enhanced Th2 production of IL-4. This shift from Th17 to Th2 cytokine production expanded intestinal MC and exaggerated oral anaphylaxis. DOCK8-deficient Treg cells failed to suppress intestinal IL-4 production and MC expansion, thus allowing unrestrained inflammation. Intervention at several points in the cascade attenuated oral anaphylaxis in Dock8^−/− mice and have implications for the treatment of food allergy in DOCK8-deficient patients.

RESULTS

Elevated serum MC tryptase levels and increased susceptibility to food anaphylaxis in DOCK8 deficiency

Beta-tryptase (tryptase) is the most abundant mediator stored in the granules of human MCs.³⁵ Serum tryptase levels were examined in 10 patients with DOCK8 deficiency including 6 with documented food allergy. The patients’ age, sex, and DOCK8 mutations are described in Table S1. DOCK8 expression was undetectable in all patients. For at least 1 month prior to the study, none were exposed to offending food allergens or presented food allergy symptoms. Serum tryptase levels were elevated in DOCK8-deficient patients compared with 14 healthy controls (Figure 1A), suggesting that DOCK8-deficient patients have increased MC load and/or activation.

Figure 1. Increased serum levels of MC tryptase in DOCK8-deficient patients and mice and anaphylactic response to oral challenge in adult Dock8^−/− mice.

(A) Serum tryptase concentrations in DOCK8-deficient patients (n = 10) and healthy controls (n = 14).

(B) Serum concentrations of MCPT-1 (left) and MCPT-4 (right) in 8-week-old Dock8^−/− mice and WT controls, n = 8 mice/group.

(C–E) Response of 6–8-week-old EC sensitized Dock8^−/− mice and WT controls to oral antigen challenge. Experimental protocol (C). Change in body temperature 0–60 min after OVA challenge and serum concentrations of MCPT-1 pre-challenge and 60 min after challenge (D). Serum concentrations of IgE anti-OVA (E), n = 5 mice/group.

(F) Response of passively sensitized 6–8-week-old Dock8^−/− mice and WT controls to oral antigen challenge. Top: experimental protocol. Bottom: change in body temperature after challenge and serum MCPT-1 pre-challenge and 60 min after, n = 5 mice/group.

(G) Response of passively sensitized 6–8-week-old Dock8^−/− mice and WT controls to intraperitoneal (i.p.) antigen challenge. Top: experimental protocol. Bottom: change in body temperature after challenge and serum MCPT-1 pre-challenge and 60 min after challenge, n = 5 mice/group. (C–E) Representative experiment out of 2. (F–G) Representative experiment out of 3. Data presented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001 by t test in (A), (B), (D)–(G) and two-way ANOVA in (D, left) and (F, right).

Differences in genetic background, infections, and exposure to allergens may have contributed to the elevation of serum tryptase levels in the patients. To circumvent these limitations, we examined Dock8^−/− mice. The mice had no evidence of infections, skin inflammation, or elevated serum IgE levels.³⁶ Serum levels of MCPT-1, selectively expressed by MMCs, were elevated in 7–9-week-old Dock8^−/− mice compared with age- and sex-matched C57BL/6 wild-type (WT) controls (Figure 1B). In contrast, serum levels of mast cell protease-4 (MCPT-4), selectively expressed by connective tissue MCs (CTMCs), were not elevated (Figure 1B).

Unlike WT BALB/c mice,³⁷ WT C57BL/6 mice epicutaneously (EC) sensitized with ovalbumin (OVA) demonstrated little, if any, change in body temperature and a small rise in serum MCPT-1 levels post-oral OVA challenge (Figures 1C and 1D). In contrast, C57BL/6 Dock8^−/− mice EC sensitized with OVA had a sustained decrease in body temperature and higher serum MCPT-1 levels post-oral OVA challenge compared with controls (Figure 1D).

WT C57BL/6 mice immunized by oral gavage (o.g.) with OVA and cholera toxin exhibited no changes in body temperature or serum levels of MCPT-1 post-oral OVA challenge (Figures S1A and S1B). In contrast, Dock8^−/− mice enterally immunized with OVA had a sustained decrease in body temperature and higher serum MCPT-1 levels post-oral OVA challenge compared with controls (Figure S1B).

Following cutaneous or oral immunization, serum levels of OVA-specific IgE were higher in Dock8^−/− mice compared with WT controls (Figures 1E and S1C). Moreover, Dock8^−/− mice generate higher-affinity IgE Abs than WT controls following mucosal immunization.³⁸ To determine whether Dock8^−/− mice are more susceptible to anaphylaxis following oral antigen challenge independent of the level and affinity of antigen-specific IgE, and differences in the response to active immunization, we examined their response to oral trinitrophenyl-bovine serum albumin (TNP-BSA) challenge following passive sensitization with IgE anti-TNP monoclonal Ab (Figure 1F). Dock8^−/− mice had a greater drop in body temperature and higher MCPT-1 serum levels compared with controls (Figure 1F). Passively sensitized Dock8^−/− mice and WT controls had comparable decreases in body temperature and serum levels of MCPT-1 following intraperitoneal (i.p.) challenge with TNP-BSA (Figure 1G). We used passive oral anaphylaxis (POA) to dissect the mechanism of increased susceptibility of DOCK8-deficient mice to food allergy.

Dock8^−/− mice have increased MC load in the small intestine; however, this expansion and susceptibility to oral anaphylaxis is not due to intrinsic DOCK8 deficiency in MCs

Intestinal MCs play a key role in oral anaphylaxis.⁸ MMCs reside in the epithelial layer and lamina propria (LP) of the intestine and are enriched in the jejunum, whereas CTMCs reside in the intestinal submucosa.^39,40 Staining with chloroacetate esterase revealed very few CTMCs in the jejunal submucosa of Dock8^−/− mice and WT controls. MMCs were more abundant in the epithelial layer and the underlying LP in Dock8^−/− mice compared with controls (Figure 2A). Transcriptomic analysis of jejunal LP cells revealed increased expression of several MC genes including Mcpt1, Mcpt2, Kit, Fcer1a, Ms4a2, Cpa3, and Hpgds (Figure 2B). Increased Mcpt1 expression was verified by qPCR analysis (Figure 2B). The accumulation of MCs in the jejunum of Dock8^−/− mice was also assessed by flow cytometry analysis of cells prepared from isolated jejunal LP and jejunal epithelium gating on CD45⁺CD3⁻B220⁻CD11b⁻CD11c⁻F4/80⁻Gr-1⁻NKp46⁻c-Kit⁺IgE⁺ MCs. The numbers of both LP MCs and intraepithelial MCs were higher in Dock8^−/− mice compared with controls (Figures 2C and 2D). The numbers of MCs were not increased in the skin or lungs of Dock8^−/− mice (Figure 2E). IgE promotes MC survival.⁴¹ Dock8^−/−Igh7^−/− mice, double deficient in DOCK8 and IgE, demonstrated increased MC load and elevated serum levels of MCPT-1 comparable to Dock8^−/− mice (Figure S2A). Thus, MCs selectively expand in the small intestine of Dock8^−/− mice independent of IgE.

Figure 2. T cell-driven expansion of jejunal MCs in adult Dock8^−/− mice.

(A) Left, representative photomicrographs at 40× magnification of jejunal sections from Dock8^−/− mice and WT controls stained with H&E and CAE. MCs stain red (arrows). Scale bar, 150 μm. Right, number of MCs per villus in the jejunum of Dock8^−/− mice and controls, n = 7–8 mice/group.

(B) Left, gene expression heatmap of MC genes differentially expressed (p < 0.05, one-way ANOVA) by jejunal LP cells from Dock8^−/− mice and controls. Heatmap colors indicate log₂ fold difference relative to the geometric mean of each row determined by an AmpliSeqRNA transcriptome panel, n = 3 mice/group. Right, qPCR analysis of Mcpt1 in jejunal LP cells. Fold increase of Mcpt1:B2m mRNA relative to mean of controls, n = 4 mice/group.

(C and D) Left, representative flow cytometry plots of surface bound IgE and c-Kit/CD117 expression gating on live CD45⁺CD3⁻B220⁻CD11b⁻CD11c⁻F4/80⁻Gr-1⁻NKp46⁻ jejunal LP MCs (C) and intraepithelial MCs (D). Right, number of LP MCs (C) and intraepithelial MCs (D) per jejunum from Dock8^−/− mice and controls, n = 4–6 mice/group.

(E) Number of MCs in the lung (left) and skin (right) of 8-week-old Dock8^−/− mice relative to controls, n = 5–7 mice/group.

(F) Serum concentration of HRP in 8-week-old Dock8^−/− mice relative to controls 20 min after oral administration of HRP, n = 7 mice/group.

(G and H) Number of jejunal MCs (G, left), baseline serum MCPT-1 (G, right), and POA with drop in body temperature (H, left) and serum MCPT-1 concentrations 60 min post-challenge (H, right) in 8-week-old Mcpt5-Cre^TgDock8^flox/flox mice and controls, n = 5–7 mice/group.

(I and J) Number of jejunal MCs (I, left), baseline serum MCPT-1 (I, right), and POA with drop in body temperature (J, left) and serum MCPT-1 concentrations 60 min post-challenge (J, right) in 8-week-old Cd4-Cre^TgDock8^flox/flox mice and controls, n = 4–6 mice/group. Representative experiment out of three is shown for (G)–(J).

Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001 by t test in (A)–(J) and by two-way ANOVA in the left panels of (H) and (J).

Increased intestinal MC load and maturation are associated with increased intestinal permeability to antigen.⁴² Intestinal permeability was assessed by administering horseradish peroxidase (HRP) to mice by o.g. and measuring serum HRP concentrations 20 min later. Serum HRP levels were higher in Dock8^−/− mice compared with WT controls (Figure 2F).

Dock8^pri/pri mice express a DOCK8 mutant protein with impaired GEF activity for CDC42.^24,43 Dock8^pri/pri mice demonstrated increased jejunal MC load and elevated serum MCPT-1 levels, compared with WT controls (Figure S2B). Thus, DOCK8 GEF activity for CDC42 is important for intestinal MC homeostasis.

MC progenitors (MCps), which give rise to CTMCs and MMCs, express Mcpt5.^44,45 Dock8 mRNA expression in jejunal LP MCs and intraepithelial MCs was readily detectable in Mcpt5-Cre^Tg mice, but not Mcpt5-Cre^Tg Dock8^flox/flox mice (Figure S2C). Jejunal MC numbers, baseline serum MCPT-1 levels, and POA were comparable in 8-week-old Mcpt5-Cre^Tg Dock8^flox/flox mice and Mcpt5-Cre^Tg controls (Figures 2G and 2H). In addition, bone marrow-derived MCs (BMMCs) from Dock8^−/− mice degranulated normally following FcεRI cross-linking, as measured by surface mobilization of the granule protein, lysosomal-associated membrane protein-1 (LAMP-1) (Figure S2D). These results indicate that intestinal MC expansion and susceptibility to oral anaphylaxis of Dock8^−/− mice are not due to MC-intrinsic DOCK8-deficiency. In all subsequent experiments, like in this one, baseline and pre-challenge serum MCPT-1 level was measured on the same sample and the value is shown as baseline MCPT-1.

Rag2^−/−Dock8^−/− mice, which lack mature T and B lymphocytes, had jejunal MCs numbers and serum levels of MCPT-1 comparable to Rag2^−/− controls (Figure S2E). Cd4-Cre^Tg Dock8^flox/flox mice which selectively lack DOCK8 in T cells, like Dock8^−/− mice, demonstrated intestinal MC expansion, increased serum levels of MCPT-1, and enhanced POA compared with Cd4-Cre^Tg controls (Figures 2I and 2J). These results implicate T cells in the intestinal MC expansion and susceptibility to oral anaphylaxis of Dock8^−/− mice.

Dock8^−/− mice have dysbiosis along with increased Th2 cells, ILC2s, and tuft cells, as well as decreased Th17 cells, ILC3s, and expression of antimicrobial genes in the small intestine

Both Th2 cells and ILC2s promote MC proliferation.⁴⁶ Moreover, administration of IL-4 causes intestinal MC expansion.⁴⁷ Flow cytometry analysis of jejunal LP cells from 8-week-old mice revealed comparable numbers of CD45⁺ cells but increased numbers of CD4⁺ T cells in Dock8^−/− mice compared with WT controls (Figures 3A and S3A). The percentage of CD4⁺Foxp3⁻GATA-3⁺ Th2 cells was higher, while the percentage of CD4⁺Foxp3⁻RORγt⁺ Th17 cells was lower in Dock8^−/− mice compared with controls (Figure 3A). The percentage CD4⁺Foxp3⁻Tbet⁺ Th1 cells was comparable between the two strains (Figure S3B). As previously reported,²⁷ the numbers of CD45⁺CD3⁻CD90.2⁺ ILCs as well as the percentage of CD45⁺CD3⁻CD90.2⁺RORγt⁺ ILC3s were markedly lower (Figure 3B). Meanwhile, the percentages of CD45⁺CD3⁻CD90.2⁺Tbet⁺ ILC1s and CD45⁺CD3⁻CD90.2⁺GATA-3⁺ ILC2s were higher in Dock8^−/− mice compared with controls (Figures 3B and S3C). Transcriptomic analysis of jejunal LP cells revealed increased expression of Il4 and Il13 and decreased expression of Il17a, Il17f, and Il22 in Dock8^−/− mice compared to WT controls, while expression of Ifng was comparable (Figure 3C). These results were verified by qPCR (Figure 3C).

Figure 3. Reduced jejunal Th17 and ILC3s, expanded jejunal Th2 cells and ILC2s, and intestinal dysbiosis in adult Dock8^−/− mice.

(A and B) Quantification by flow cytometry of cell numbers and percentages in the jejunum of 8-week-old Dock8^−/− mice and controls. Number of CD4⁺ cells and percentage of Th2 and Th17 cells out of CD4⁺FOXP3⁻ T cells (A). Number of total ILCs and percentages of ILC2s and ILC3s out of total ILCs (B), n = 4 mice/group.

(C) Top, heatmap of cytokine gene expression in the jejunal LP cells of 8-week-old Dock8^−/− mice and WT controls, n = 3–4 mice/group. Bottom, cytokine gene expression by qPCR in the LP from Dock8^−/− mice and WT controls, n = 3–5 mice/group.

(D) Number of CD45⁻CD31⁻EpCam⁺SiglecF⁺ tuft cells determined by flow cytometry in the jejunum of 8-week-old Dock8^−/− mice and controls. n = 7 mice/group.

(E) Top, heatmap of tuft cell gene expression in the jejunal epithelial layer of 8-week-old Dock8^−/− mice and WT controls, n = 3 mice/group. Bottom, Dclk1 and Il25 gene expression by qPCR in jejunal epithelial layer, n = 5–6 mice/group.

(F) Top, heatmap of antimicrobial genes in the jejunal epithelial layer, n = 3 mice/group. Bottom, qPCR analysis of Reg3g1 and Reg3b in the epithelial layer, n = 4–5 mice/group.

(G–I) Gut microbiome analysis in 7–9-week-old Dock8^−/− mice and WT controls. Shannon index measurement of alpha-diversity of intestinal microbiome communities (G). Principal-coordinate analysis of beta diversity of microbiome communities measured in Bray-Curtis index, PERMANOVA test, F value = 4.2053, p < 0.003 (H). LEfSe analysis of genus-level taxa significantly enriched in 7–9-week-old Dock8^−/− mice and controls, false discovery rate (FDR)-adjusted p < 0.05 for each of the listed taxa (I).

In (C), (E), and (F), heatmap colors correspond to the log₂ fold difference relative to the geometric mean of each row. Gene expression, results are expressed as fold change of the gene of interest:B2m mRNA ratio relative to the mean of WT controls. Representative experiment out of 3 is shown for (A), (B), and (D). Representative experiment of out 2 is shown for (G)–(I). Data presented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001 by t test in (A)–(G).

Intestinal tuft cells are the sole IECs that produce IL-25 (30). They proliferate and produce IL-25 in response to microbial products that include succinate^48,49 as well as IL-4 and IL-13.^50–52 Flow cytometry analysis of cells from the epithelial layer of the jejunum revealed that the numbers of CD45⁻CD31⁻EpCam⁺SiglecF⁺ tuft cells were higher in Dock8^−/− mice compared with controls (Figure 3D). Transcriptomic analysis of the epithelial cell layer revealed increased expression of the tuft cell genes Siglecf, Dclk1, Pou2f3, Sucnr1, and Il25,⁵³ and qPCR confirmed increased expression of Dclk1 and Il25 (Figure 3E).

IL-17 and IL-22 drive the expression by IECs of antimicrobial peptide (AMP) genes that influence the composition of the microbiome and the inflammatory state of the intestine.^19,54 Consistent with the decreased expression of Il17 and Il22 genes in the LP, transcriptome analysis of the epithelial cell layer revealed decreased expression of the antimicrobial genes Reg3b, Reg3g1, Lyz2, and Nos2^55–57 in Dock8^−/− mice compared with WT controls, and the decrease in Reg3b and Reg3g1 was verified by qPCR (Figure 3F).

We performed 16S rRNA composition analysis on stools from 7–9-week-old Dock8^−/− mice and WT littermates. The intestinal microbiome in Dock8^−/− mice showed higher alpha-diversity measured by the Shannon index compared with controls (Figure 3G). Dock8^−/− mice and WT controls harbored distinct microbiome compositions, as measured by the Bray-Curtis dissimilarity index with different bacterial genera identified by Linear discriminant Effect Size (LEfSe) analysis (Figure 3H). Dock8^−/− mice showed enrichment of Candidatus arthromitus (i.e., segmented filamentous bacteria [SFB]), Blautia, Staphylococcus, Coprococcus, and Streptococcus, compared with enrichment of Lactobacillus, Alloprevotella, Parabacteroides, and Turicibacter in WT controls (Figure 3I). Cd4-Cre^Tg Dock8^flox/flox mice also demonstrated enrichment of SFB, Coprococcus, Marvinbryantia, and Odoribacter, compared with enrichment of Lactobacillus, Alloprevotella, and Turicibacter in Cd4-Cre^Tg controls (Figure S3D), indicating that DOCK8 deficiency in T cells is likely responsible for alterations in the gut microbiome in Dock8^−/− mice.

Decreased intestinal expression of IL-17 cytokines precedes the development of dysbiosis, intestinal MC expansion, and increased type 2 cytokine expression, while reconstitution with type 17 cytokines ameliorates intestinal MC expansion and susceptibility to oral anaphylaxis in Dock8^−/− mice

To define the relationship between altered intestinal cytokine expression, intestinal dysbiosis, and intestinal MC expansion in Dock8^−/− mice, we examined mice shortly after weaning. Four-week-old Dock8^−/− mice demonstrated no differences in the numbers of intestinal LP MCs or serum levels of MCPT-1 compared with WT controls (Figure 4A). Intestinal expression of Il4, Il13, Ifng, and Il25 was comparable, but intestinal expression of Il17a, Il17f, and Il22 was drastically reduced (Figure 4B). There were no differences in the gut microbiome composition between 4-week-old Dock8^−/− mice and WT controls as measured by the Shannon index of alpha-diversity, the Bray-Curtis dissimilarity index, or discriminating taxa by LEfSe (Figures S4A–S4C). Thus, type 17 cytokine deficiency precedes intestinal dysbiosis and MC expansion in Dock8^−/− mice.

Figure 4. Type 17 cytokine reconstitution or treatment with oral antibiotics starting at 3 weeks of age reverses jejunal MC expansion and susceptibility to oral anaphylaxis in Dock8^−/− mice.

(A) Jejunal MC numbers (left) and serum MCPT-1 (right) in 4-week-old Dock8^−/− mice and WT controls, n = 3–6 mice/group.

(B) Cytokine gene expression in the LP (left) and Il25 gene expression in the epithelial layer (right) from the jejunum of 4-week-old Dock8^−/− mice and WT controls, n = 3–6 mice/group.

(C–E) Effect of treating 3-week-old Dock8^−/− mice with IL-22+IL17A or IgG-Fc for 4 weeks. Number of jejunal MCs determined (C, left), baseline serum MCPT-1 (C, right), POA with drop in body temperature (D, left) and serum MCPT-1 concentrations 60 min post-challenge (D, right), and expression of Il25 by the jejunal epithelial layer and Il4 by LP cells (E), n = 4 mice/group.

(F–H) Effect of treating 3.5-week-old Dock8^−/− mice with oral antibiotics for 4 weeks versus no treatment. Number of jejunal MCs (F, left), baseline serum MCPT-1 (F, right), POA with drop in body temperature (G, left) and serum MCPT-1 concentrations 60 min post-challenge (G, right), and expression of Il25 by the jejunal epithelial layer and Il4 by LP cells (H), n = 3–4 mice/group.

(I–K) Effect of weekly administration for 2 weeks of fecal material from Dock8^−/− mice or WT donors to 8-week-old GF WT recipient mice, with analysis performed 1 week after the second transfer. Number of jejunal MCs (I, left), baseline serum MCPT-1 (I, right), POA with drop in body temperature (J, left) and serum MCPT-1 concentrations 60 min post-challenge (J, right), and expression of Il25 by the jejunal epithelial layer and Il4 by LP cells (K), n = 4–5 mice/group.

Data shown in (C)–(E) are representative of 4 independent experiments and (F)–(H) are representative of 3 independent experiments. Data shown in (I)–(K) are representative of 2 independent experiments. In (B), (E), (H), and (K) gene expression is shown as fold increase of the cytokine gene:B2m mRNA ratio relative to the mean of controls. Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001 by t test in (A)–(K) and two-way ANOVA in left panels of (D), (G), and (J).

To determine whether deficiency in type 17 cytokines results in intestinal MC expansion and susceptibility to oral anaphylaxis in Dock8^−/− mice, 15 μg each of recombinant IL-22:IgG-Fc fusion protein (IL-22) and rIL-17A were co-administered i.p. weekly starting at 3–4 weeks of age, and the mice were examined 4 weeks later. Dock8^−/− mice treated with IL-22+IL-17A had reduced intestinal MC expansion, serum MCPT-1, and POA compared with Dock8^−/− controls treated with immunoglobulin G (IgG)-Fc (Figures 4C and 4D). In addition, they had decreased intestinal expression of Il25 and Il4 (Figure 4E). Three taxa, Turicibacter, Parabacteroides, and Romboutsia, were enriched in the intestinal microbiome of IL-22+IL-17A-treated mice (Figure S4C). As Turicibacter and Parabacteroides were also enriched in WT mice compared with Dock8^−/− mice, this suggests protection by these taxa from oral anaphylaxis.

To examine the role of intestinal dysbiosis in intestinal MC expansion in Dock8^−/− mice, mice were treated starting at 3–4 weeks of age with antibiotics (ampicillin, vancomycin, neomycin, and metronidazole) added to the drinking water, then examined 4 weeks later. Dock8^−/− mice treated with antibiotics had a reduction in jejunal MC load, serum MCPT-1 levels, and POA compared with untreated Dock8^−/− controls (Figures 4F and 4G). They also had decreased intestinal Il25 and Il4 expression (Figure 4H). Thus, intestinal dysbiosis drives intestinal MC expansion and susceptibility to oral anaphylaxis in Dock8^−/− mice.

To examine whether intestinal dysbiosis is sufficient to drive intestinal MC expansion in the absence of DOCK8 deficiency, we investigated the impact of fecal material transfer (FMT) from Dock8^−/− or WT donors germ-free (GF) WT recipients. Consistent with their previously reported increase in tissue MC numbers and heightened susceptibility to anaphylaxis,⁵⁸ 8-week-old GF WT mice had increased jejunal MCs and a more severe POA compared with WT mice housed under specific-pathogen-free (SPF) conditions (Figures S4D and S4E). Two weeks after FMT from Dock8^−/− or WT donors to 8-week-old GF WT recipients, jejunal MC load, serum MCPT-1 levels, and POA were comparable in the two groups (Figures 4I and 4J), Intestinal expression of Il25, but not Il4, was higher in recipients of FMT from Dock8^−/− donors (Figure 4K).

IL-25 and T cell-derived IL-4 drive intestinal MC expansion and susceptibility to oral anaphylaxis in Dock8^−/− mice

To determine the role of IL-25 in intestinal MC expansion and susceptibility to oral anaphylaxis, 50 μg of anti-IL-25 or IgG1 isotype control was administered i.p. weekly to Dock8^−/− mice, starting at 3–4 weeks old for 4 weeks. IL-25 blockade reduced jejunal MC load and serum levels of MCPT-1 and attenuated POA in compared with IgG isotype control (Figures 5A and 5B). IL-25 blockade also reduced jejunal Il4 expression (Figure 5C), suggesting that IL-25 may drive increased intestinal Il4 expression in Dock8^−/− mice.

Figure 5. Blockade of IL-25 or IL-4 reverses jejunal MC expansion and susceptibility to oral anaphylaxis in Dock8^−/− mice.

(A–C) Effect of treating 3-week-old Dock8^−/− mice with 50 μg anti-IL-25 or IgG1 isotype control weekly for 4 weeks. Number of jejunal MCs (A, left), baseline serum MCPT-1 (A, right), POA with drop in body temperature (B, left) and serum MCPT-1 concentrations 60 min post-challenge (B, right), and expression of Il4 by LP cells (C), n = 5 mice/group.

(D–F) Effect of treating 3-week-old Dock8^−/− mice with 100 μg of anti-IL-4 or IgG1 isotype control three times a week for 4 weeks. Number of jejunal MCs (D, left), baseline serum MCPT-1 (D, right), POA with drop in body temperature (E, left) and serum MCPT-1 concentrations 60 min post-challenge (E, right), and Il4 expression by LP cells (F), n = 3–6 mice/group.

(G–I) Number of jejunal MCs determined by flow cytometry (G, left), baseline serum MCPT-1 (G, right), POA with drop in body temperature (H, left) and serum MCPT-1 concentrations 60 min post-challenge (H, right), and Il4 expression by LP cells (I) in 8-week-old Cd4-Cre^TgIl4-Il13^flox/floxDock8^−/− mice and Cd4-Cre^TgIl4-Il13^flox/+Dock8^−/− controls, n = 3–5 mice/group.

Data in (A)–(C) are representative of 3 independent experiments. Data in (D)–(I) are representative of 2 independent experiments. Gene expression in (C), (F), and (I), results are expressed as fold increase of the Il4:B2m mRNA ratio relative to control. Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001 by t test in (A)–(I) and by two-way ANOVA in left panels of (B), (E), and (H).

To determine the role of IL-4, 100 μg of IgG anti-IL-4 or IgG1 isotype control was administered i.p. three times weekly to Dock8^−/− mice starting at 3–4 weeks of age for 4 weeks. IL-4 blockade reduced jejunal MC load and serum levels of MCPT-1 and attenuated POA compared with IgG isotype control (Figures 5D and 5E). IL-4 blockade in Dock8^−/− mice starting at 8–10 weeks of age for 4 weeks also decreased jejunal MCs, serum MCPT-1 levels, and POA (Figures S5A and S5B), suggesting that IL-4 blockade may be useful in patients with DOCK8 deficiency.

We determined the individual contribution of ILC2s and Th2 cells to the intestinal MC expansion and enhanced oral anaphylaxis in Dock8^−/− mice. There was no difference in jejunal MC load, serum MCPT-1 levels, POA, or jejunal Il4 expression between Rora^CreIl4–13^flox/floxDock8^−/− mice, which selectively lack IL-4 and IL-13 in ILCs, and Rora^CreIl4–13^flox/+Dock8^−/− controls (Figures S5C–S5E). In contrast, jejunal MC load, serum MCPT-1 levels, and POA were reduced in Cd4-Cre^TgIl4–13^flox/floxDock8^−/− mice, which selectively lack IL-4 and IL-13 in T cells, compared with Cre^TgIl4–13^flox/+Dock8^−/− controls (Figures 5G–5I). Furthermore, jejunal Il4 expression was lower in Cd4-Cre^TgIl4–13^flox/floxDock8^−/− mice compared with controls (Figure 5I). Thus, T cell-derived IL-4 drives intestinal MC expansion and exaggerated anaphylaxis in Dock8^−/− mice.

IL-4 targets MCs to drive intestinal MC expansion to promote oral anaphylaxis in Dock8^−/− mice

To determine whether IL-4 targets MCs directly to drive their expansion and promote food anaphylaxis in Dock8^−/− mice, we examined Mcpt5-Cre/Il4ra^flox/flox/Dock8^−/− mice, which selectively lack interleukin-4 receptor alpha (IL-4Rα) and thus both type I and type II IL-4 receptors in MCs. IL-4Rα surface expression was not detectable on MCs in the intestinal LP of these mice (Figure S6A). Mcpt5-Cre/Il4ra^flox/flox/Dock8^−/− mice had reduced jejunal MC load, serum MCPT-1 levels, and POA compared with Mcpt5-Cre/Il4ra^flox/+/Dock8^−/− controls (Figures 6A and 6B). Lack of IL-4Rα in the MCs of Dock8^−/− mice had no effect on intestinal Il4 expression (Figure 6C). However, it reduced intestinal permeability (Figure 6D), indicating that IL-4 driven MC expansion/activation drives increased intestinal permeability in Dock8^−/− mice.

Figure 6. Selective deficiency of IL-4Rα in MCs reverses jejunal MC expansion and susceptibility to oral anaphylaxis in Dock8^−/− mice.

(A–C) Number of jejunal MCs (A, left), baseline serum MCPT-1 (A, right), POA with drop in body temperature (B, left) and serum MCPT-1 concentrations 60 min post-challenge (B right), and Il4 expression by LP cells determined by qPCR (C) in 8-week-old Mcpt5-Cre/Il4ra^flox/floxDock8^−/− mice and Mcpt5-Cre/Il4ra^flox/+Dock8^−/− controls, n = 4–6 mice/group.

(D) Serum concentration of HRP in 8-week-old Mcpt5-Cre/Il4ra^flox/floxDock8^−/− mice relative to controls 20 min after oral administration of HRP (D), n = 4–6 mice/group. A representative experiment out of 3 is shown in (A)–(C) and out of 2 for (D). For qPCR analysis, results are expressed as fold change of the Il4:B2m mRNA ratio relative to control. Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001 by t test.

IL-4 signaling in Treg cells inhibits the generation of intestinal RORγt⁺ Treg cells in Dock8^−/− mice and promotes intestinal Il4 expression, intestinal MC expansion, and oral anaphylaxis

Treg cells suppress MC proliferation and activation as well as food allergy.^59–61 This function is primarily exerted by RORγt⁺ Treg cells.¹⁵ The percentage of CD4⁺FOXP3⁺ Treg cells among CD4⁺ cells, as well as the FOXP3 expression by Treg cells, were decreased in the jejunum of Dock8^−/− mice compared with WT controls (Figure 7A). The fraction of RORγt⁺ Treg cells among jejunal FOXP3⁺ Treg cells was markedly decreased in Dock8^−/− mice compared with WT controls, whereas the fraction of GATA3⁺ Treg cells was increased (Figure 7B). The intestinal microbiota drive transforming growth factor (TGF)-β1 expression by intestinal Treg cells and Treg-derived TGF-β1 drives the differentiation of RORγt⁺ Treg cells in an autocrine manner.⁶¹ Tgfb1, but not Il10 expression, was lower in CD4⁺GFP⁺(FOXP3⁺) Treg cells sorted from the jejunal LP of Foxp3^eGFPDock8^−/− mice compared with Foxp3^eGFP controls (Figure 7C).

Figure 7. Selective deficiency of IL-4R in Tregs reverses jejunal MC expansion and susceptibility to oral anaphylaxis in Dock8^−/− mice.

(A–C) Percentage CD4⁺FOXP3⁺ Tregs out of CD4⁺ T cells (A, left), MFI of FOXP3 expression in Treg cells (A, right), and percentage of GATA-3⁺ and RORγt⁺ cells among Tregs (B) in jejunal LP cells from unmanipulated 8-week old WT and Dock8^−/− mice. Tgfb1 and Il10 expression by sorted CD4⁺eGFP⁺Tregs (C) in the jejunum of 8-week-old Foxp3-eGFP/Dock8^−/− mice and Foxp3-eGFP controls, n = 4–7 mice/group.

(D) Representative histograms (left) and percent suppression (right) of IgE-induced surface mobilization of LAMP-1 in BMMCs from Dock8^−/− mice and WT controls in the absence or presence of sorted splenic CD4⁺eGFP(FOXP3)⁺ Tregs from Foxp3-eGFP/Dock8^−/− mice or Foxp3-eGFP controls, n = 4–5 mice/group.

(E–H) Percentage CD4⁺Foxp3⁺ Tregs out of CD4⁺ T cells (E, left) and MFI of FOXP3 expression in Treg cells (E, right), percentage of GATA-3⁺ and RORγt⁺ cells among Tregs (F), Tgfb1 expression in sorted YFP⁺ jejunal CD4⁺YFP(FOXP3)⁺ Tregs (G), and suppression of IgE-induced surface mobilization of LAMP-1 in WT BMMCs by sorted splenic CD4⁺YFP⁺(FOXP3⁺) Tregs from Foxp3-YFP-Cre/Il4ra^flox/floxDock8^−/− mice and Foxp3-YFP-Cre/Il4ra^flox/+ Dock8^−/− controls (H), n = 4–5 mice/group.

(I–K) Jejunal MC numbers (I, left), baseline serum MCPT-1 (I, right), POA with drop in body temperature (J, left), and serum MCPT-1 concentrations 60 min post-challenge (J, right), and Il4 expression in LP cells (K) in 8-week-old Foxp3-YFPCre/Il4ra^flox/floxDock8^−/− mice and controls, n = 3–5 mice/group. Data shown are representative of 3 independent experiments in (A)–(K). Gene expression results are expressed as fold change of the cytokine gene:B2m mRNA ratio relative to the mean of controls. Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001 by t test in (A)–(K).

Treg cell suppression of IgE-mediated MC degranulation is mediated by TGF-β1.⁶¹ Sorted CD4⁺GFP⁺(FOXP3⁺) Treg cells from Foxp3^eGFPDock8^−/− mice and Foxp3^eGFP controls were compared for their ability to suppress the degranulation of WT BMMCs preincubated with IgE anti-DNP and challenged with DNP-BSA. Suppression by DOCK8-deficient Treg cells was impaired compared with DOCK8-sufficient WT Treg cells (Figure 7D). BMMCs from Dock8^−/− mice were normally susceptible to inhibition of IgE-mediated degranulation by DOCK8-sufficient Treg cells but poorly suppressed by DOCK8-deficient Treg cells (Figure 7D).

IL-4 signaling in Treg cells dampens their Tgfb1 expression and inhibits their differentiation into RORγt⁺ Treg cells.^60,61 We investigated the effect of disrupting IL-4 signaling in the Treg cells of Dock8^−/− mice. We compared Foxp3-YFP-Cre/Il4ra^flox/floxDock8^−/− mice, which selectively lack IL-4Rα in Treg cells, to Foxp3-YFP-Cre/Il4ra^flox/+Dock8^−/− controls. IL-4Rα deletion in Treg cells had no effect on the percentage of Treg cells among CD4⁺ cells in the intestine of Dock8^−/− mice or on FOXP3 expression but resulted in a marked increase in the fraction of intestinal RORγt⁺ Treg cells, with no alteration in the fraction of intestinal GATA3⁺ Treg cells (Figures 7E and 7F). It also caused increased Tgfb1 expression by intestinal Treg cells with no significant alteration in Il10 expression (Figures 7G and S7A). Further, Treg cells from Foxp3-YFP-Cre/Il4ra^flox/floxDock8^−/− mice demonstrated improved ability to suppress IgE-mediated degranulation of WT BMMCs (Figure 7H). IL-4Rα deletion in Treg cells markedly reduced jejunal MC load and serum MCPT-1 levels (Figure 7I) and attenuated POA in Dock8^−/− mice (Figure 7J). It also decreased Il4 expression and the percentage of jejunal CD4⁺IL-4⁺ cells (Figures 7K and S7B). Thus, IL-4 signaling impairs the capacity of intestinal Treg cells in Dock8^−/− mice to produce TGF-β1, express RORγt, and suppress intestinal IL-4 production, intestinal MC expansion and oral anaphylaxis.

DISCUSSION

We demonstrate that intestinal dysbiosis and insufficient production of IL-17 cytokines synergize with Treg cell dysfunction to drive intestinal MC expansion in DOCK8-deficient mice, thereby increasing intestinal permeability and promoting susceptibility to oral anaphylaxis.

We observed that patients with DOCK8 deficiency who are highly susceptible to food anaphylaxis had elevated levels of serum tryptase. Findings in Dock8^−/− mice recapitulated those in the patients. Dock8^−/− mice were highly susceptible to oral anaphylaxis following cutaneous, as well as enteric sensitization, demonstrated elevated baseline serum levels of MCPT-1 and developed higher levels of IgE Abs following sensitization compared with WT controls, mirroring the high levels of IgE Abs to food antigens found in DOCK8-deficient patients.⁶² This may have contributed to the exaggerated oral anaphylaxis in actively sensitized Dock8^−/− mice. Following passive sensitization, Dock8^−/− mice demonstrated selective susceptibility to oral anaphylaxis independent of the levels and affinity of IgE Ab.

Dock8^−/− mice demonstrated increased MC load in the small intestine, but not lungs or skin. This increase involved primarily MMCs located in the small intestinal epithelium and LP and was associated with increased intestinal permeability, which likely contributed to the increased susceptibility to oral anaphylaxis. DOCK8 deficiency in T cells, but not in MCs, was sufficient to cause intestinal MC expansion and increased susceptibility to oral anaphylaxis.

Several immunologic abnormalities were identified in the small intestine of adult Dock8^−/− mice, including decreased RORγt⁺ Th17 cells and ILC3s, increased GATA3⁺ ILC2s and Th2 cells, decreased expression of Il17a, Il17f, and Il22, and increased expression of Il4 and Il13. The decreased expression of type 17 cytokines is likely due to the role DOCK8 plays in STAT3 activation following T cell receptor (TCR) activation and IL-23R ligation.^26,27 It is in line with the previously demonstrated impairment of Th17 differentiation of naive murine and human DOCK8-deficient T cells, decreased numbers of intestinal ILC3s in Dock8^pri/pri mice, and diminished numbers of circulating ILC3s in DOCK8-deficient patients.^26,27,62,63

Dock8^−/− mice demonstrated decreased expression of AMPs in the intestinal epithelial layer and significant shifts in the gut microbiome. These include an increase in SFB similar to IL-17A-deficient mice,²¹ a decrease in Lactobacillus, thought to protect from food allergy,⁶⁴ and in Turicibacter, Parabateroides, and Alloprovetella, all of which play an immunoregulatory role.^65,66 Moreover, Dock8^−/− mice demonstrated increased intestinal tuft cells and Il25 expression, likely driven by the products of a dysbiotic microbiome⁴⁸ and possibly by elevated intestinal expression of Il4 and Il13.⁵⁰

Intervening at 3–4 weeks of age, when both intestinal microbiome and intestinal MC load were unperturbed, demonstrated a critical role for type 17 cytokine deficiency as well as intestinal dysbiosis in the intestinal MC expansion and increased susceptibility to oral anaphylaxis of adult Dock8^−/− mice and placed IL-17A/IL-22 deficiency and dysbiosis upstream of the increased intestinal expression of Il25 and Il4. Following IL-22 + IL-17A treatment of Dock8^−/− mice, Turicibacter, Parabacteroides, and Romboutsia emerged as associated with protection from anaphylaxis. These taxa respectively play a role in bile acid metabolism and protect from intestinal infection,⁶⁷,⁶⁸ increase the expression of tight junction proteins, strengthen the gut barrier,⁶⁹ activate Toll-like receptor 2 (TLR2)/nuclear factor κB (NF-κB) signaling in IECs, and improve the gut microbiota.⁷⁰ Future experiments will seek to define the exact mechanisms by which these taxa may protect from oral anaphylaxis. Altogether, the results suggest that intestinal dysbiosis drives intestinal MC expansion and susceptibility to oral anaphylaxis in Dock8^−/− mice. However, FMT from Dock8^−/− mice was insufficient to further increase the pre-existing heightened susceptibility phenotype of GF WT mice.

Both IL-25 and IL-4 blockades reduced intestinal MC expansion, attenuated oral anaphylaxis, and reduced intestinal Il4 expression in Dock8^−/− mice. Reduction of intestinal Il4 expression by IL-25 blockade strongly suggests that IL-25 drives increased intestinal Il4 expression in Dock8^−/− mice. Reduction of Il4 expression by IL-4 blockade is consistent with the critical role IL-4 plays in upregulating its own expression by T cells.⁷¹

IL-4 directly targeted MCs to cause their expansion in the intestine of Dock8^−/− mice. IL-4Rα deletion in MCs significantly reduced intestinal MC expansion and oral anaphylaxis in Dock8^−/− mice. It significantly reduced intestinal permeability in Dock8^−/− mice, indicating an important role for IL-4-driven MC activation in the exaggerated oral anaphylaxis in these mice.

Defective Treg cell function played a critical role in the intestinal MC expansion and susceptibility to oral anaphylaxis of Dock8^−/− mice. Treg cell number and expression of FOXP3 were decreased in the intestine of these mice. The proportion of GATA3⁺ Treg cells was increased, whereas the proportion of RORγt⁺ Treg cells was decreased. Moreover, expression of Tgfb1, the major Tgfb isoform expressed in Treg cells, was decreased. Consistent with the role of Treg-cell-derived TGF-β1 in suppressing MC degranulation,^72,73 Treg cells from Dock8^−/− mice failed to suppress the degranulation of BMMCs.

IL-4Rα deletion in Treg cells of Dock8^−/− mice caused a marked increase in intestinal RORγt⁺ Treg cells and Tgfb1 expression by intestinal Treg cells in agreement with the observations that IL-4 suppresses Rorc expression and Tgfb gene auto-amplification.^74,75 Further, it restored the ability of DOCK8-deficient Treg cells to suppress MC degranulation and reduced intestinal Il4 expression, MC load, serum MCPT-1 levels, and attenuated oral anaphylaxis. Although Th2 skewing is characteristic of DOCK8 deficiency, DOCK8-deficient naive T cells do not exhibit increased Th2 polarization in vitro. In contrast, DOCK8-deficient induced Treg (iTreg) cells display increased instability with IL-4 treatment in vitro and in skin undergoing allergic inflammation.⁷⁶ Our data strongly suggest that Th2 skewing in DOCK8 deficiency results from defective Treg cell control of the Th2 response. The dysfunction of DOCK8-deficient Treg cells likely involves their inability to make effective contact with their targets as well as impaired IL-2 driven STAT5 phosphorylation and reduced FOXP3 expression as we previously demonstrated.^23,24 Subsequent unregulated IL-4 production further aggravates Treg cell dysfunction in a vicious cycle.

In summary, we provide evidence that intestinal dysbiosis and impaired IL-17 cytokine production synergize with Treg cell dysfunction to increase susceptibility to oral anaphylaxis in mice with DOCK8-deficiency. Our findings suggest that reconstitution with type 17 cytokines, non-absorbable antibiotics, and IL-4Rα blockade may help mitigate food anaphylaxis in DOCK8-deficient patients. Some of these strategies could apply to other patients with food allergy and an increased intestinal MC load such as patients with atopic dermatitis.

Limitations of the study

This study does not address the composition of the intestinal microbiome or the cellular and cytokine milieu of the small intestine in patients with DOCK8 deficiency. The role of individual taxa in the intestinal pathology of DOCK8-deficient mice is not defined. The contributions of systemic IL-4 derived fromTh2 cells versus Treg cells to the increased intestinal MC load and susceptibility to anaphylaxis of DOCK8-deficient mice remain unknown. The contribution of the previously demonstrated impaired IL-2-driven STAT5 phosphorylation to the Treg cell dysfunction, which underlies the increased susceptibility of DOCK8-deficient mice to anaphylaxis, as well as whether this susceptibility is reversed by administration of normal Treg cells, needs investigation.

RESOURCE AVAILABILITY

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Raif S. Geha (raif.geha@childrens.harvard.edu).

Materials availability

This study did not generate new unique reagents.

Data and code availability

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

STAR★METHODS

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Sample donors and collection

Plasma was collected after informed consent in a research protocol approved by the Committee on Clinical Investigation at Boston Children’s Hospital and the collaborators’ institutions. DOCK8-deficient patient information is provided in Table S1.

Mice

Dock8^−/− and Dock8^flox/flox mice were described in references Janssen et al.^23,24 Dock8^pri/pri mice were a gift from C. Goodnow.^24,43 Igh7^−/− mice were a gift from H. Oettgen.⁷⁷ Mcpt5-Cre mice were described in reference Scholten et al.⁷⁸ Cd4-Cre^Tg mice were from Taconic. Rag2^−/− mice were from Taconic. Foxp3-eGFP and Foxp3-YFP-Cre mice were from the Jackson Laboratory. Il4ra^flox mice were described in reference Herbert et al.⁷⁹ Il4-Il13^flox and Rora-Cre^Tg mice were described in reference Leyva-Castillo et al.⁸⁰ All transgenic mice, except Igh7^−/− mice, were on a C57BL/6 background. WT C57BL6 and BALB/c mice were from the Jackson Laboratory. Both female and male mice were studied with similar results. All mice listed above were kept in an SPF environment with ad libitum access to water and food. WT C57BL6 GF mice were obtained from the Boston Children’s Hospital Animal Research Core and moved to SPF conditions prior to FMT. Mice were studied at 8-weeks of age unless otherwise noted in the text. All mouse studies were approved and performed in accordance with Boston Children’s Hospital and Michigan Medicine Institutional Animal Research and Care Committee.

Primary mouse cell cultures

MCs were cultured from bone marrow collected from 8-week-old mice as detailed in reference Galand et al.⁸¹

METHOD DETAILS

Serum concentrations of mast cell tryptases

Concentrations of β-tryptase in human serum and MCPT-1 and MCPT-4 in mouse serum were determined using ELISA commercial kits (EZ0898 from Boster Bio for β-tryptase, 88–7503 from ThermoFisher for MCPT-1, and LS-F44860 from LSBio for MCPT4) and per the manufacturer’s directions.

Anaphylaxis

For active oral anaphylaxis, 8-week-old mice were sensitized epicutaneously (EC) or orally. For EC sensitization, back skin was subjected to six cycles of tape stripping 2 days apart followed by application of 200 μg OVA (A5503, Sigma-Aldrich) or saline secured with a film dressing (TegadermTM, 3M). The skin was tape stripped 6 times in the first cycle and twice in subsequent cycles.⁸² For oral sensitization, mice were sensitized by gavage of 5 mg of OVA and 10 μg CT (#100B, List Biological Technologies) in 150 μL of PBS weekly for 8 weeks.³⁷ Mice were rested for 7 days after sensitization, then challenged with 150 mg OVA in 200 μL PBS. 24 h before challenge, mice were injected with an implantable temperature transponder (IPTT-300, BioMedic Data Systems). Core body temperatures were measured serially for 60 min after challenge using a DAS-6001 Smart Probe (Bio Medic Data Systems). After 60 min, blood was collected for serum analysis.

For passive oral anaphylaxis, mice were passively sensitized by i.p., injection of 10 μg IgE anti-trinitrophenyl (TNP) monoclonal antibody, a gift from Dr. Fred Finkelman.⁸³ The following day, mice were challenged by oral gavage of 2.5 mg TNP-BSA (T-5050, Bio-search) or i.p. injection of 8 μg TNP-BSA. Temperature monitoring and blood collection were done as detailed above.

Serum concentrations of IgE anti-OVA antibodies

For detection of serum IgE anti-OVA antibodies, 96-well plates (ThermoFisher) were coated overnight at 4°C with rat anti-mouse IgE (clone R35–72, BD Biosciences) at 2 μg/mL overnight. The plate was blocked with 0.5% gelatin for 1 h. Serum samples were incubated for 2 h. Biotinylated-OVA at 12 μg/mL was used for detection. Bound OVA-biotin was detected with avidin-HRP (ThermoFisher) and color change after incubation with TMB substrate (ThermoFisher) was measured using Biotek ELx808 plate reader.

Histology

For chloroacetate esterase (CAE) staining, 1 cm pieces of the mouse mid-jejunum were fixed with 4% paraformaldehyde for 1 hour before transfer to PBS. Samples were embedded in paraffin and sectioned. Sections were stained with hematoxylin and eosin (H&E) and CAE. CAE⁺ cells were counted by two blinded investigators. Ten HPFs were counted per slide. Slides were imaged using a bright field microscope (Nikon).

Analysis of gene expression

RNA was isolated using a RNAeasy Plus micro kit (Qiagen). Complementary DNA was reverse transcribed using an iScript cDNA synthesis kit (Biorad). The Ion AmpliSeq Transcriptome Mouse Gene Expression Kit was used to prepare bar-coded libraries and sequenced by using an Ion S5 next-generation sequencer. The AmpliSeqRNA plug-in (ThermoFisher Scientific) was used to calculate differential gene expression analysis. qPCR was performed with TaqMan Universal Master Mix II, no UNG, and TaqMan Probes (all from Life Technologies) using a QuantStudio 3 Real-Time PCR machine (Applied Biosystems, Thermo Fisher). Gene expression was normalized to β2-microglobulin. Relative mRNA expression was quantified using the 2^−ΔΔCt method.

Jejunal, lung and skin cell suspensions

Cells were isolated from the jejunum as previously described.⁸¹ Briefly, the jejuna were harvested and flushed with PBS and 2% fetal calf serum (FCS) before being cut longitudinally and into 1 cm segments. The pieces were incubated in HBSS supplemented with 10 mM EDTA, 1.5 mM DTE, and 0.5% FCS shaking at 37°C for 40 min. Epithelial cells were collected from the supernatant and digested with Liberase DL (Roche) and DNase I (Sigma-Aldrich). Tissue pieces were then digested in HBSS with 20% FCS and 100 U/ml of collagenase VIII (Sigma-Aldrich) with shaking at 37°C for 30 min. Immune cells from the LP were further purified on a 40% Percoll gradient (GE Healthcare). To isolate cells from the lungs, single cell suspensions were generated by gently disrupting tissue between two frosted slides (VWR). Cells were then passed through a 70 μm cell strainer. For skin cell suspensions, 1 cm² skin pieces were obtained and cell isolation was performed as we previously described.⁸⁴

Flow cytometry

Cell suspensions were washed, incubated with TruStain FcX (anti-CD19/CD32, Biolegend) and eF506 viability dye (ThermoFisher) to exclude dead cells. Monoclonal antibodies to the following cell surface antigens were used for flow cytometry as detailed in the key resource table. For intracellular staining cells were fixed and permeabilized using the eBioscience Transcription Factor Staining Buffer Set (ThermoFisher). All data was acquired on a BD LSRFortessa cell analyzer using FACSDiva software (BD Biosciences). Analyses were performed using FlowJo software (Tree Star, Inc.).

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
IgE anti-trinitrophenyl	Gift from Dr. Fred Finkelman (Cincinnati Children's Hospital)	N/A
mouse IgE	BD Biosciences	Cat#564207; RRID: AB_2738668
mouse IL-25	Biolegend	Cat# 514401; RRID: AB_2043954
IgG1 isotype control	BioXCell	Cat# BE0088; RRID: AB_1107775
mouse IL-4	BioXCell	Cat# BE0045; RRID: AB_1107707
mouse IgE DNP	Sigma-Aldrich	SPE-7/Cat# D8406
TruStain FcX (mouse CD19/CD32)	Biolegend	Cat# 101320; AB_1574975
mouse CD11b	BD Biosciences	Cat# 553309; RRID: AB_394773
mouse CD11c	Biolegend	Cat# 117303; RRID: AB_313772
mouse F4/80	Biolegend	Cat# 123105; RRID: AB_893499
mouse/human B220	Biolegend	Cat# 103204; RRID: AB_312989
mouse Gr-1	Biolegend	Cat# 108403; RRID: AB_313368
mouse NKp46	Biolegend	Cat# 137615; RRID: AB_11219387
mouse CD3	Biolegend	Cat# 100243; RRID: AB_2563946
mouse IgE	Biolegend	Cat# 406907; RRID: AB_493291
CD45	Biolegend	Cat# 103116; RRID: AB_312981
CD117 (c-kit)	Biolegend	Cat# 105811; RRID: AB_313220
Siglec-F	BD Pharmingen	Cat# 552126; RRID: AB_394341
CD31	Biolegend	Cat# 102419; RRID: AB_10612742
EpCAM	Biolegend	Cat# 118215; RRID: AB_1236477
CD45	Biolegend	Cat# 103112; RRID: AB_312977
GATA-3	Thermo Fisher	Cat# 53-9966-42; RRID: AB_2574493
RORgt	Thermo Fisher	Cat# 12-6981-82; RRID: AB_10807092
CD3	Thermo Fisher	Cat# 46-0032-80; RRID: AB_1834428
CD90.2	Biolegend	Cat# 105324; RRID: AB_2201291
CD4	Biolegend	Cat# 100451; RRID: AB_2564591
Foxp3	Thermo Fisher	FJK-16s/17-5773-82
T-bet	Biolegend	Cat# 644815; RRID: AB_10896427
CD107a (LAMP-1)	Biolegend	Cat# 121618; RRID: AB_2749905
Biological samples
Human plasma samples	Samples were collected after informed consent in a research protocol approved by the Committee on Clinical Investigation at Boston Children’s Hospital and the collaborators’ institutions.	N/A
Chemicals, peptides, and recombinant proteins
ovalbumin	Sigma-Aldrich	A5503
cholera toxin	ListBiological Technologies	#100B
trinitrophenyl-bovine serum albumin (TNP-BSA)	Biosearch	T-5050
avidin-HRP	Thermo Fisher	18-4200-89
TMB substrate	ThermoFisher	N301
Liberase DL	Roche	5989132001
DNase I	Sigma-Aldrich	DN25
Percoll	GE Healthcare	17089101
viability dye	Thermo Fisher	65-0866
eBioscience Transcription Factor Staining Buffer Set	Thermo Fisher	00-5523-00
horseradish peroxidase	Sigma-Aldrich	P6782
QuantaBlu Fluorogenic peroxidase substrate	Thermo Fisher	15169
recombinant mouse IL-22	R&D systems	582
recombinant mouse IL-17A	R&D systems	7956
mouse IgG2a Fc protein	BioXCell	BE0097
ampicillin	Sigma-Aldrich	A9518
vancomycin	Sigma-Aldrich	V2002
neomycin	Sigma-Aldrich	PHR1491
metronidazole	Sigma-Aldrich	M1547
DNP albumin	Sigma-Aldrich	A6661
anti-mouse CD3/CD28 coated beads	ThermoFisher	11452
Critical commercial assays
Human tryptase/TPSAB1/B2 ELISA kit	Boster Bio	EZ0898
mouse MCPT1 ELISA	Thermo Fisher	88-7503
mouse MCPT4 ELISA	LSBio	LS-F44860
RNAeasy micro kit	Qiagen	74004
iScript cDNA synthesis kit	Biorad	1708890
The Ion AmpliSeq Transcriptome Mouse Gene Expression Kit	Thermo Fisher	A36553
TaqMan Universal Master Mix II, no UNG	Thermo Fisher	4440043
ZymoBIOMICS DNA Miniprep Kit	Zymo Research	D4300
Platinum SuperFi DNA Polymerase	Invitrogen	12351010
Experimental models: Organisms/strains
Mouse: Dock8-/-		N/A
Mouse: Dock8flox/flox		N/A
Mouse: Dock8pri/pri	Gift from Dr. Christopher Goodnow (Garvan Institute of Medical Research)	N/A
Mouse: Igh7-/-	Gift from Dr. Hans Oettgen (Boston Children’s Hospital)	N/A
Mouse: Mcpt5-Cre	Gift from Dr. Axel Roers Cologne, Germany	N/A
Mouse: CD4-CreTg	Taconic	RRID:IMSR_TAC:4196
Mouse: Rag2-/-	Taconic	RRID:IMSR_TAC:RAG2
Mouse: Foxp3-eGFP	Jackson Laboratory	RRID:IMSR_JAX:023800
Mouse: Foxp3-YFP-Cre	Jackson Laboratory	RRID:IMSR_JAX:016959
Mouse: Il4-Il13flox/flox	Jackson Laboratory	RRID:IMSR_JAX:015859
Mouse: Rora-CreTg	Gift from Dr. Dennis O’Leary, California, USA	N/A
Mouse: C57BL/6N	Charles River	RRID:MGI:2159965
Mouse: Balb/c	Charles River	RRID:MGI:2161072
Oligonucleotides
B2m Taqman Assays	Thermo Fisher	Mm00437762_m1
Il4 Taqman Assays	Thermo Fisher	Mm00445259_m1
Il13 Taqman Assays	Thermo Fisher	Mm00434204_m1
Mcpt1 Taqman Assays	Thermo Fisher	Mm00656886_g1
Il17a Taqman Assays	Thermo Fisher	Mm00439618_m1
Il17f Taqman Assays	Thermo Fisher	Mm00521423_m1
Il22 Taqman Assays	Thermo Fisher	Mm00444241_m1
Ifng Taqman Assays	Thermo Fisher	Mm01168134_m1
Dclk1 Taqman Assays	Thermo Fisher	Mm01284845_m1
Il25 Taqman Assays	Thermo Fisher	Mm00499822_m1
Reg3b Taqman Assays	Thermo Fisher	Mm00440616_g1
Reg3g1 Taqman Assays	Thermo Fisher	Mm00441127_m1
Tgfb1 Taqman Assays	Thermo Fisher	Mm01178820_m1
Il10 Taqman Assays	Thermo Fisher	Mm01288386_m1
Software and algorithms
FlowJo	Tree Star, Inc	Version 10.8
Prism	GraphPad	Version 10
MicrobiomeAnalyst	N/A	https://www.microbiomeanalyst.ca/

Intestinal permeability

Mice were gavaged with 400 μl of 5 mg/ml HRP (Sigma-Aldrich). Blood was obtained 20 min after gavage. Serum HRP content was measured by incubation with QuantaBlu Fluorogenic peroxidase substrate (ThermoFisher). Fluorescence intensity was measured by a BMG LABTECH FLUOstar Omega reader.

Microbiome analysis

Stool samples from 8–10 week-old mice were acquired in the middle of the light period. Samples were immediately placed on dry ice and frozen at −80°C for storage. Genomic DNA was extracted from stools using ZymoBIOMICS DNA Miniprep Kit (Zymo Research, Cat D4300) and 16S rRNA gene V3-V4 region was amplified using Platinum SuperFi DNA Polymerase (Invitrogen, Cat 12351010) and primers with annealing sequences 5’-CCTACGGGNGGCWGCAG-3’ and 5’-GGACTACNVGGGTWTCTAAT-3’. Amplicon sequencing was performed in 300 bp paired-end using Miseq v3 platform, and amplicon sequence variant (ASV) table was generated using a previous bioinformatics pipeline.⁸⁵ Alpha-diversity, beta-diversity and bacterial genera with differential abundances were analyzed using MicrobiomeAnalyst.⁸⁶

Treatment with IL-22 +IL-17A, anti-IL-25, IL and anti-IL-4

Three-week-old Dock8^−/− mice were treated for 4 weeks prior to analysis. One set was i.p. injected weekly with 15 μg of IL-22-Fc (582, R&D systems) and IL-17A (7956, R&D systems) or IgG2a Fc protein (BE0097, BioXCell). Another set was i.p. injected weekly with 50 μg anti-IL25 (clone 35B, BioLegend) or rat IgG1 isotype control (BE0088, BioXCell). A third set was i.p. injected 3 times per week i.p. with 100 μg anti-IL4 (11B11, BioXCell), or IgG1 isotype control. 8–10 week old mice were also injected with anti-IL4 as detailed above.

Oral antibiotics treatment

At 3.5-weeks of age, mice were weaned and given water supplemented with 1 g/L ampicillin, 500 mg/L vancomycin, 1 g/L neomycin, and 1 g/L metronidazole (all from Sigma) for 4 weeks.

Fecal material transplantation

Mouse stool samples were acquired in the middle of the light period. For FMT, each fecal pellet was solubilized in 150 μl of PBS and kept on ice. The equivalent of 2 pellets were combined and gavaged into each GF recipient mouse within 60 min of fecal collection. After FMT, recipient mice were kept under SPF conditions with autoclaved food and bedding. Repeat FMT was done at 1 week later and the mice were analyzed at 2 weeks

LAMP-1 mobilization by BMMCs

BMMCs were incubated with 5–50 ng/ml of anti-DNP IgE (clone SPE-7, Sigma-Aldrich). After 24 h, BMMCs were stimulated with 0–100 ng/ml DNP albumin (Sigma Aldrich) and stained for surface LAMP-1 using a PE-conjugated anti-LAMP Ab (clone 1D4B, Biolegend). After 10 min, samples were washed with ice cold PBS with 0.5% BSA and 2 mM EDTA and analyzed by flow cytometry.

Treg cell suppression of MC degranulation

For Treg cell suppression of BMMC degranulation, CD4⁺CD25⁺eGFP⁺ Treg cells were sorted from Foxp3-eGFP and Foxp3-eGFP/Dock8^−/− mice using an MA900 Cell Sorter (Sony). Treg cells were added at equal numbers to BMMCs along with anti-CD3+anti-CD28 coated beads (11452, ThermoFisher) and IgE anti-DNP mAb. 24 hours later, BMMCs were stimulated with 50 ng/ml DNP albumin (Sigma Aldrich) and analyzed by flow cytometry as detailed above.

QUANTIFICATION AND STATISTICAL ANALYSIS

Comparisons were analyzed for statistical significance using unpaired two tailed Student’s t test, two-way ANOVA, and Mann-Whitney U test to determine the p value using Prism software (GraphPad Software, Inc.). A p value of <0.05 was considered significant.

Supplementary Material

SUPPLEMENTAL INFORMATION

Supplemental information can be found online at https://doi.org/10.1016/j.immuni.2025.06.004.

Highlights.

• DOCK8 deficiency is linked with food allergy and elevated serum tryptase levels

• DOCK8-deficient mice exhibit oral anaphylaxis and expanded mucosal mast cells (MMCs)

• IL-4-driven MC expansion results from reduced intestinal Th17 cells and dysbiosis

• DOCK8-deficient Treg cells fail to suppress intestinal IL-4 production and MC expansion