Immune sensing of food allergens promotes avoidance behaviour

The study addresses how the immune system modulates behavior in the context of allergy, specifically whether and how allergic sensitization can drive antigen-specific avoidance of foods. Allergic diseases have risen with industrialization, and type 2 immunity (Th2 cells, IgE, mast cells, eosinophils, ILC2s) mediates responses that, while protective against noxious substances (venoms, irritants, xenobiotics), can become pathological. Defensive reflexes (sneezing, itching, vomiting) are known, and avoidance behavior has been observed in allergic contexts, suggesting type 2 immunity could limit exposure to detrimental stimuli. The mechanistic links between type 2 immune responses and behavioral outputs remain unclear. This work tests the hypothesis that IgE-dependent allergic sensitization induces specific food allergen avoidance via immune–neural circuits and mediators, defining molecular and cellular requirements and neural correlates, and exploring genetic background effects.

Prior studies indicate immune system impacts on nervous system function and behavior (neuroimmunity). Allergic responses can induce defensive reflexes and avoidance behaviors. Historical work demonstrated immune-induced flavor aversion and conditioned taste aversion following anaphylaxis. TRPM5-dependent taste pathways mediate bitter/sweet/umami but may be dispensable for certain avoidance. Mast cells and IgE are central in allergic disease and GI symptoms; mast cell–neuronal interactions drive itch, pain, diarrhea, and visceral malaise. Leukotrienes and prostanoids modulate nociceptors and vagal neurons; CysLT receptors implicated in itch and emesis. GDF15 from stressed tissues acts on the area postrema/NTS to cause conditioned flavor avoidance and anorexia. The literature suggests candidate mediators (histamine, serotonin, substance P, CGRP, leukotrienes) and pathways (vagal, nociceptive DRG, area postrema) that could link immune activation to aversion but lacked direct mechanistic demonstration in food allergy.

• Animals and sensitization: BALB/c and C57BL/6 mice were sensitized subcutaneously with ovalbumin (OVA) plus aluminum hydroxide (alum) on days 0 and 7; controls received alum alone. Some cohorts used OVA plus cholera toxin (oral adjuvant) or OVA plus LPS (non-allergic inflammatory sensitization). Acclimation occurred in home cages with two lickometer-equipped water bottles.
• Two-bottle preference tests: After acclimation, one bottle was replaced with OVA solution (typically 1%). Lick counts and side preference were recorded over minutes to hours; bottle sides were switched on day 2. Time course, dose dependence, persistence (up to 48 weeks), and specificity (BSA control) were assessed. TRPM5 dependency was tested using Trpm5-null mice.
• Allergen oral challenges: In sensitized mice, intragastric OVA was administered according to schedules (e.g., five challenges) to model food allergy, with sham water gavages as controls. Outcomes included gut motility, serum antibodies, mast cell accumulation, and neuronal activation.
• Neural activation mapping: Ninety minutes post-oral OVA, brains were collected for FOS immunostaining to assess activation in the nucleus of tractus solitarius (NTS), external lateral parabrachial nucleus (elPBN), central amygdala (CeA), and control regions (area postrema, lateral hypothalamus, PVN).
• Genetic models: IgE-deficient (IgE-KO), IL-4 receptor alpha-deficient (IL-4RA-KO), high-affinity Fc epsilon receptor (FCER1)-deficient mice, CysLTR2-KO, LTC4S-KO, substance P-KO, TRPM5-KO, and RMB mice (for mast cell depletion) were used. Bone marrow chimeras were generated by transplanting WT or FCER1-KO marrow into irradiated WT recipients to restrict/reconstitute hematopoietic FCER1 expression.
• Mast cell manipulation: RMB mice received diphtheria toxin (DT) to ablate mast cells (protocol included three DT injections prior to preference testing). Efficacy verified in peritoneal and intestinal compartments; basophils unaffected.
• Pharmacology: Histamine H1 (loratadine) and H2 (famotidine) antagonists; serotonin synthesis inhibitor para-chlorophenylalanine and 5-HT3 antagonist (ondansetron); neuropeptide antagonists (substance P receptor antagonist aprepitant; CGRP receptor antagonist BIBN4096); cyclooxygenase inhibitor indomethacin; 5-lipoxygenase inhibitor zileuton; CysLTR1 antagonist montelukast; CysLTR2 antagonist HAMI3379. Drugs were administered acutely or over several days prior to testing or challenges.
• Leukotriene pathway assessment: Alox5 expression quantified by qPCR across GI regions (duodenum>jejunum>distal) and epithelial compartments after sensitization/challenge; dependence on IgE and mast cells assessed. Targeted mass spectrometry measured duodenal LTE4 levels in WT vs LTC4S-KO mice.
• Vagotomy: Subdiaphragmatic vagotomy performed; success confirmed by Fluoro-Gold tracer absence in dorsal motor nucleus of the vagus. Preference testing conducted 3 weeks post-surgery.
• GDF15 analyses: Serum GDF15 measured after serial oral or systemic OVA challenges across strains/genotypes and with leukotriene pathway inhibitors. Tissue Gdf15 mRNA profiled by qPCR (duodenum, colon) and RNAscope in situ hybridization with probes for Gdf15, Fcer1a, and Epcam to localize expression to epithelial cells and proximity to mast cells. Recombinant GDF15 (rGDF15) administered to mast cell–depleted mice immediately before preference tests in a dose–response manner. GDF15 function tested via neutralizing anti-GDF15 antibody vs isotype administered 5 h before preference testing.
• Immunological readouts: Serum total and OVA-specific IgE and OVA-specific IgG1 measured; intestinal mast cell and eosinophil accumulation quantified; corticosterone and mast cell protease 1 measured; GI transit and diarrhea assessed.
• Statistics: Mann–Whitney tests for two-group comparisons; ANOVA with multiple-comparisons for multi-group datasets; data reported as mean ± s.e.m.; P-value thresholds noted.

• Allergic sensitization induces rapid, specific, and persistent avoidance of the sensitized food antigen:
  • Sensitized BALB/c mice decreased preference for OVA solution in a dose-dependent manner within 10 min, persisting on day 2 despite side switch, and lasting at least 48 weeks. Controls preferred OVA over water. Avoidance was antigen-specific (no aversion to BSA) and TRPM5-independent.
  • Brain activation consistent with aversive processing (FOS+) observed in NTS, elPBN, and CeA 90 min after oral OVA in sensitized mice (NTS P=0.043; elPBN P=0.0159; CeA P=0.0317); no changes in area postrema, lateral hypothalamus, or PVN.
• Genetic background influences behavior: BALB/c exhibited robust avoidance; C57BL/6 showed weak avoidance, correlating with lower IgE induction, minimal GI transit change, and limited mast cell activation.
• IgE is required for allergen avoidance:
  • Sensitized IgE-KO mice did not avoid OVA and instead increased preference compared to unsensitized IgE-KO controls (preference difference P=0.0003); WT sensitized mice avoided OVA (controls vs sensitized P not significant in IgE-KO; WT showed significant avoidance).
  • FCER1 deficiency or absence of FCER1 on hematopoietic cells abolished avoidance: chimeras with FCER1-KO marrow into WT recipients failed to avoid OVA despite normal OVA-specific IgE titres; hematopoietic FCER1 expression was necessary.
  • IL-4Rα signaling required (IL-4/IL-13 axis), as IL-4RA-KO sensitized mice did not avoid OVA.
  • Central FOS activation by oral allergen was IgE-dependent.
  • Even non-allergic OVA+LPS sensitization led to IgE-dependent avoidance with low but elevated IgE.
  • Avoidance to bitter tastant denatonium remained intact in IgE-KO mice (specificity of the deficit).
• Mast cells are necessary for avoidance:
  • RMB mice depleted of mast cells showed higher OVA consumption and preference than controls despite similar IgE levels (e.g., preference P=0.0124 increased in mast cell-depleted sensitized mice).
• Cysteinyl leukotrienes mediate avoidance:
  • 5-LOX inhibitor zileuton increased OVA preference in sensitized mice (reducing avoidance; P=0.0094) without affecting controls.
  • Alox5 expression induced in GI tissue (duodenum highest) post-challenge; induction lost in IgE-KO and mast cell–depleted mice; epithelium contributed to expression.
  • Duodenal LTE4 increased 7–25-fold in allergic WT vs controls; absent in LTC4S-KO.
  • LTC4S-KO sensitized mice showed increased cumulative OVA licks and preference vs WT (reduced avoidance), with unchanged IgE.
  • CysLTR2 antagonism (HAMI3379) or CysLTR2 deficiency increased OVA preference on day 1 in sensitized mice (less avoidance; pharmacology recapitulated LTC4S effects). CysLTR1 effects were confounded in controls (montelukast altered intake).
  • Subdiaphragmatic vagotomy did not significantly affect avoidance, suggesting non-vagal or redundant afferent pathways (e.g., nociceptive DRG) may mediate leukotriene sensing.
• GDF15 is required and downstream of IgE–mast cell–leukotriene signaling:
  • Serum GDF15 increased with oral and systemic OVA challenges in BALB/c, scaling with challenge number; induction required IL-4Rα, IgE, and FCER1-expressing cells and was reduced by zileuton, montelukast, or HAMI3379 pretreatment.
  • Gdf15 transcripts induced mainly in colon (also duodenum) epithelium (EPCAM+), with minimal overlap with Fcer1a; epithelial cells in colonic crypt/transit zones localized near mast cells.
  • Recombinant GDF15 rescued avoidance in mast cell–depleted mice in a dose-dependent manner; higher doses (0.1 mg/kg) matched WT sensitized avoidance.
  • Neutralizing anti-GDF15 increased OVA preference (reduced avoidance) on day 2 in sensitized mice (e.g., OVA preference increased; P=0.014–0.0021), without altering IgE/IgG1 titres.
• Collectively, allergen-specific IgE and mast cells drive avoidance via cysteinyl leukotrienes (notably LTC4S/CysLTR2 axis) and epithelial GDF15, with activation of central aversion circuits (NTS, elPBN, CeA).

The results demonstrate that immune detection of a dietary antigen through IgE and mast cell activation can rapidly and specifically alter feeding behavior to avoid the sensitized allergen. This behavior occurs prior to overt intestinal inflammation, indicating a protective, anticipatory defense strategy. Central nervous system activation in prototypical aversion circuits (NTS, elPBN, CeA) links peripheral immune sensing to central processing of aversive stimuli. Mechanistically, mast cell-derived cysteinyl leukotrienes (LTC4/LTD4/LTE4) engage CysLTR2-dependent pathways to drive avoidance, and concurrently stimulate epithelial GDF15 production. GDF15, acting via the area postrema/NTS axis, is necessary for maintaining avoidance, and exogenous GDF15 can restore avoidance in the absence of mast cells. The partial temporal dependence on CysLTR2 and requirement for GDF15 suggest multi-pathway integration: leukotrienes may act both directly on sensory afferents (vagal or DRG nociceptors) and indirectly via induction of GDF15. Strain differences highlight the importance of genetic background: BALB/c mice mount stronger IgE responses and mast cell activation and exhibit robust avoidance, whereas C57BL/6 show weaker IgE responses and attenuated avoidance. This aligns with known differences in type 2 immunity between strains. These findings extend the concept of immune–nervous system crosstalk, positioning IgE-mediated sensing as a means to expand the organism’s environmental appraisal, guiding behavioral avoidance of harmful foods. The model reconciles beneficial early-phase roles of IgE (driving avoidance) with pathological consequences upon chronic exposure (diarrhea, motility, inflammation).

This study establishes that allergic sensitization generates antigen-specific food avoidance behavior in mice via an IgE- and mast cell–dependent pathway. Allergen ingestion activates central aversion circuits and requires cysteinyl leukotrienes (notably LTC4S/CysLTR2) and epithelial GDF15 to promote and sustain avoidance. Genetic background modulates the magnitude of behavior, correlating with IgE and mast cell responses. Main contributions include: (1) demonstration of immune-driven, antigen-specific avoidance distinct from taste pathways; (2) identification of IgE, mast cells, leukotrienes, and GDF15 as key mediators; (3) mapping of relevant CNS activation; and (4) delineation of strain-dependent variability. Future work should dissect afferent neural pathways (vagal vs DRG nociceptors), define cellular sources and targets of leukotrienes and GDF15 at single-cell resolution, test genetic models of leukotriene receptors beyond CysLTR2, investigate synergy between leukotrienes and GDF15, and assess translational relevance in human food allergy and therapeutic modulation (e.g., leukotriene/GDF15 pathways) to manage exposure and symptoms.

• Pharmacological inhibitor studies (histamine, serotonin, CGRP, substance P, leukotriene receptor antagonists) are subject to dose/timing limitations; negative results for some mediators may reflect suboptimal regimens and warrant genetic validation.
• Montelukast (CysLTR1 antagonist) confounded interpretation by affecting OVA intake in non-sensitized mice; thus, the role of CysLTR1 remains unresolved.
• Subdiaphragmatic vagotomy severs both afferent and efferent fibers, potentially confounding assessment of vagal contributions; redundancy with nociceptive DRG pathways may mask effects.
• GDF15 neutralization affected avoidance only on day 2, and recombinant GDF15 required higher-than-physiological doses to restore avoidance, indicating additional synergistic signals are likely required.
• Strain-dependent findings (BALB/c vs C57BL/6) suggest genetic variability; generalizability across strains and to humans remains to be established.
• While behavior persisted long-term, mechanistic studies focused on acute/short-term windows; chronic adaptations and circuit plasticity were not fully explored.