Skin damage signals mediate allergic sensitization to spatially unlinked antigen

Abstract

Our current understanding of immunity to pathogens suggests that anatomic coupling of antigens with danger signals is a required feature for the formation of immune memory. However, in the context of pathogen-independent inflammation, the stringency of this anatomical coupling is unclear. Here, we demonstrate that multiple modes of skin injury were sufficient to induce a humoral response to antigens introduced in the gut. Skin damage induced a narrow subset of endocrine cytokines that were necessary and sufficient for the priming of antigens introduced at various distal tissues. Thus, in addition to "local priming" of antigen entering through damaged skin, there also exists another paradigm of "remote priming" where anatomical coupling is not essential because of the dissemination of damage-associated intermediaries. Our findings have implications for understanding the fundamental mechanisms of the formation of humoral memory with wide implications for diseases such as food allergy and in vaccinology.

Fig. 1.. Skin inflammation can prime humoral responses to bystander antigens in the gut.

(A) Schematic representation of skin injury–induced oral sensitization. Mice were subjected to one of various acute injury modalities [tape stripping (TS), intradermal acetone injection (Ace), punch biopsy (PB), severe UV exposure, or chronic atopic dermatitis (MC903)]. OVA was administered intragastrically (ig) immediately after the acute injury events or topical application of chronic inflammatory agent on days 0, 3, and 6. Plasma was collected on day 10, and mice were challenged with OVA intravenously (iv) on day 12. (B) Heatmap of OVA-specific humoral responses as log₂ fold change over respective controls [depilated mice for TS, PB, and UV; intradermal (id) PBS for ACE; and epicutaneous (ec) ethanol (EtOH) for MC903]. (C to H) Temperature loss induced by intravenous challenge for each skin damage modality. (H) Mice were exposed to UV, given OVA or vehicle (Veh) intragastrically, and kept cohoused on wire-bottom cages during sensitization. (I) Temperature loss induced by oral challenge in BALB/c mice on day 12. **P < 0.01, ***P < 0.001, and ****P < 0.0001, using two-way analysis of variance (ANOVA) with the [(C) to (H)] Šidák or (I) Tukey multiple comparisons correction. Data are presented as means ± SEM (n = 4 to 10 mice per group) and are representative of at least two independent experiments, with the exception of (F), which are pooled from two experiments.

Fig. 2.. Circulating skin injury–derived cytokines mediate remote priming in a coded fashion.

(A) Multiplex cytokine analysis from plasma 1 hour after skin injury or topical treatment on days 0 and 3. IL-33, TSLP, IL-1β, and IL-18 were measured by ELISA. LIF, leukemia inhibitory factor; RANTES, regulated on activation, normal T cell expressed and secreted; MIP, macrophage inflammatory protein; GM-CSF, granulocyte-macrophage colony-stimulating factor; VEGF, vascular endothelial growth factor; MIG, monokine induced by interferon-gamma [also known as CXCL9 (C-X-C motif chemokine ligand 9)]; IP-10, interferon-gamma inducible protein 10 kDa (also known as CXCL10); KC, keratinocyte chemoattractant (also known as CXCL1); CSF, colony-stimulating factor; M-CSF, macrophage colony-stimulating factor; and LIX, lipopolysaccharide-induced CXC chemokine (also known as CXCL5). Values reflect the log₂ fold change over the average of depilated mice or respective vehicle controls [PBS intradermally or ACE and EtOH epicutaneously for MC903]. (B) Schematic representation of systemic cytokine-induced distal sensitization. Mice were sensitized with 0.5 to 1 μg of recombinant cytokine iv and 10 μg of OVA subcutaneously (sc) on days 0, 3, and 6; bled on day 9; and challenged with OVA intravenously on day 10. (C) Relative OVA-specific immunoglobulin concentrations reflected as log₁₀ fold change over the average of the PBS controls. (D to G) Temperature loss after 100-μg OVA intravenous challenge on day 10 after the indicated skin injury–induced sensitization in wild-type (WT) and ST2–TSLPR–IL-25 triple-knockout (YRS TKO) mice. ‡ denotes death of group during challenge. (H and I) OVA-specific immunoglobulins after sensitization with 1 μg of IL-1β iv and 10 μg of OVA sc in WT and TKO mice. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001, using [(D), (E), and (I)] two-way ANOVA with the Tukey multiple comparisons correction, [(F) and (G)] unpaired t test at first time point, or (H) one-way ANOVA with the Tukey multiple comparisons correction. Data are presented as means ± SEM (n = 2 to 13 mice per group). With the exception of the screens in (A) and (B), which were performed once, data are representative of at least two independent experiments.

Fig. 3.. Circulating IL-33 is sufficient to induce high-affinity IgE to novel antigens at any barrier site.

(A, C, and E) OVA-specific IgG1 and (B, D, and F) OVA-specific IgE after distal sensitization with [(A) and (B)] increasing doses of recombinant IL-33 intravenously and 10 μg of OVA sc in the footpad, [(C) and (D)] 100 ng of IL-33 iv with increasing doses of OVA via subcutaneous footpad injection, or [(E) and (F)] 100 ng of IL-33 iv and rested for the indicated times before 10 μg of OVA via sc footpad injection. (G) BSA-NP-2–specific (high-affinity) and BSA-NP-27–specific (low-affinity) IgE from T_FH cell–sufficient (Bcl6^fl/fl) or T_FH cell–deficient mice (Cd4cre⁺Bcl6^fl/fl) that were immunized with 100 ng of IL-33 iv and 10 μg of NP-20–coated OVA sc. Plasma was assayed at 1:10 dilutions for all samples. (H) OVA-specific IgE of mice distally sensitized with 500 ng of IL-33 iv and either 50 mg of OVA ig or 100 μg of OVA intranasally (in), as indicated. (I) OVA-specific IgG1 and (J) OVA-specific IgE of mice immunized with 100 ng of IL-33 iv and 50 mg of OVA ig on day 0 (1×); days 0 and 3 (2×); or days 0, 3, and 6 (3×), as indicated. (K) Temperature loss induced by the fourth oral challenge of BALB/c mice remotely primed with 100 ng of IL-33 iv and 50 mg of OVA ig. (L) OVA-specific IgE of naïve or OVA-“tolerized” mice immunized with 100 ng of IL-33 iv and 50 mg of OVA ig. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001, using one-way ANOVA with the [(A), (B), (H), (I), (J), and (L)] Tukey or (E) Dunnet multiple comparisons correction, [(C), (D), and (F)] Kruskal-Wallis test with the Dunn multiple comparisons correction, or two-way ANOVA with the (G) Tukey or (K) Šidák multiple comparisons correction. Data are presented as means ± SEM (n = 3 to 9 mice per group) and are representative of at least two independent experiments.

Fig. 4.. IL-33 remodels the immune landscape to promote allergic sensitization, in part, via ILC2s.

(A) Quantification of CD49b⁺FcεRIα⁺ basophils, CD11b⁺Siglec-F⁺ eosinophils, and Lin⁻KLRG1⁺Thy1⁺ST2⁺GATA3⁺ ILC2s as percent of CD45⁺ cells in indicated organs and days after sensitization with 100 ng of IL-33 iv and 10 μg of OVA sc. n = 2 to 4 mice per group. D, days. (B, D, and I) OVA-specific IgE from the indicated mice on day 9 after sensitization with 100 ng of IL-33 iv and 50 mg of OVA ig. (C and E) T_FH cell percentages of CD62L⁻CD44⁺ effector CD4 T cells in indicated organs at day 6 after sensitization with 100 ng of IL-33 iv and 10 μg of OVA sc. (F and G) IL-4, IL-5, and/or IL-13 measured in the supernatant of cells after ex vivo stimulation on day 10 with (F) IL-33 or (G) anti-CD3. (H) Quantification of IL-4– or IL-13–expressing Lin⁻KLRG1⁺ST2⁺ ILC2s in indicated organs on day 10 after sensitization with 100 ng of IL-33 iv and 10 μg of OVA sc. (J) Local priming involves the introduction of antigen at the same place, such as via damaged or inflamed skin. Remote priming occurs when damage or inflamed skin induces the release of damage-associated cytokines with adjuvant potential, namely IL-1β, IL-18, TSLP, and IL-33, which can act systemically to initiate a humoral response to antigens that enter at any site, such as the gastrointestinal tract. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001, using [(B), (D), and (I)] unpaired t test, two-way ANOVA with the (C) Šidák correction or [(E) to (G)] Tukey multiple comparisons correction, (H) the Kruskal-Wallis H test with the Dunn multiple comparisons correction (for IL-4⁺), or (H) the Brown-Forsythe and Welch test (for IL-13⁺). Data are presented as means ± SEM (n = 2 to 11 mice per group) and are representative of at least two independent experiments with the exception of (I), which are pooled from two experiments. LN, lymph node; mLN, mesenteric lymph node; dLN, draining lymph node; ndLN, nondraining lymph node; Spln, spleen.