Staphylococcus aureus δ-toxin present on skin promotes the development of food allergy in a murine model

Abstract

Background: Patients with food allergy often suffer from atopic dermatitis, in which Staphylococcus aureus colonization is frequently observed. Staphylococcus aureus δ-toxin activates mast cells and promotes T helper 2 type skin inflammation in the tape-stripped murine skin. However, the physiological effects of δ-toxin present on the steady-state skin remain unknown. We aimed to investigate whether δ-toxin present on the steady-state skin impacts the development of food allergy.

Material and methods: The non-tape-stripped skins of wild-type, Kit^W-sh/W-sh, or ST2-deficient mice were treated with ovalbumin (OVA) with or without δ-toxin before intragastric administration of OVA. The frequency of diarrhea, numbers of jejunum or skin mast cells, and serum levels of OVA-specific IgE were measured. Conventional dendritic cell 2 (cDC2) in skin and lymph nodes (LN) were analyzed. The cytokine levels in the skin tissues or culture supernatants of δ-toxin-stimulated murine keratinocytes were measured. Anti-IL-1α antibody-pretreated mice were analyzed.

Results: Stimulation with δ-toxin induced the release of IL-1α, but not IL-33, in murine keratinocytes. Epicutaneous treatment with OVA and δ-toxin induced the local production of IL-1α. This treatment induced the translocation of OVA-loaded cDC2 from skin to draining LN and OVA-specific IgE production, independently of mast cells and ST2. This resulted in OVA-administered food allergic responses. In these models, pretreatment with anti-IL-1α antibody inhibited the cDC2 activation and OVA-specific IgE production, thereby dampening food allergic responses.

Conclusion: Even without tape stripping, δ-toxin present on skin enhances epicutaneous sensitization to food allergen in an IL-1α-dependent manner, thereby promoting the development of food allergy.

Keywords: IL-1α; IgE; Staphylococcus aureus δ-toxin; epicutaneous sensitization; food allergy; murine model.

Figure 1

δ-toxin present on the non-tape-stripped skin strongly induced food allergic responses following epicutaneous sensitization to food allergens in a murine model. (A) Experimental design to investigate the occurrence of for food allergy after intragastric administration of OVA in WT mice that had been epicutaneously treated or not with OVA ± δ-toxin once a week for six weeks. Blood samples were taken, and small intestines were isolated on day 56. (B) Frequency of diarrhea in OVA-challenged mice after epicutaneous treatment with OVA ± δ-toxin on the non-tape-stripped skin or after non-treatment. (C–E) Serum levels of (C) OVA-specific IgE, (D) OVA-specific IgG1, and (E) MCPT-1 in the mice after the final administration of OVA. (F) The numbers of jejunum mast cells of the mice after the final administration of OVA. (G) Jejunum sections stained with chloroacetate esterase (scale bars, 100 μm). Mast cells stain red. (B, F) Data are pooled from two independent experiments. (C–E) Data are representative of two independent experiments. Means ± SD have been plotted. *, P < 0.05, **, P < 0.01.

Figure 2

IL-33-ST2 signaling is dispensable for OVA-challenged food allergic responses after epicutaneous treatment with OVA and δ-toxin on the non-tape-stripped skin. (A, D) Frequency of diarrhea in OVA-administered mice after epicutaneous treatment with OVA ± δ-toxin on the non-tape-stripped skin of (A) WT and Kit^W-sh/W-sh mice and (D) WT and ST2 knockout (KO) mice. (B, E) Serum levels of OVA-specific IgE in (B) WT and Kit^W-sh/W-sh mice and (E) WT and ST2 KO mice after the final administration of OVA. (C, F) The numbers of jejunum mast cells in (C) WT and Kit^W-sh/W-sh mice and (F) WT and ST2 KO mice after the final administration of OVA. (A–F) Data are representative of two independent experiments. (B, C, E, F) Means ± SD have been plotted. **, P < 0.01.

Figure 3

δ-toxin present on the non-tape-stripped skin enhanced epicutaneous sensitization to food allergen in a murine model. (A) Experimental design for epicutaneous sensitization. WT mice were epicutaneously treated or not with OVA ± δ-toxin once a week for six weeks on the non-tape-stripped abdominal skin. On day 42, blood samples were obtained, and skins and small intestines were isolated. (B) Serum levels of OVA-specific IgE. (C) The thickness of epidermis. (D) Skin sections stained with hematoxylin and eosin (scale bars, 100 μm). (E, F) The numbers of mast cells in the (E) skin and (F) jejunum. (B, F) Data are pooled from two independent experiments. (C, E) Data are representative of two independent experiments. Means ± SD have been plotted. *, P < 0.05, **, P < 0.01. ns, not significant.

Figure 4

δ-toxin present on the non-tape-stripped skin strongly induced the translocation of OVA-loaded cDC2 from skin to draining LN in murine model. (A) Experimental design for analyzing dendritic cells in skin and axillary LNs. WT mice were epicutaneously treated or not with OVA-AF647 ± δ-toxin on the non-tape-stripped abdominal skin on days 0 and 7. Samples of skin were isolated on Day 8 and axillary LNs were isolated on Day 8 or 11. (B) The percentage of OVA-AF647-positive cells among skin cDC2 from the mice 24 h after the final treatment. (C, D), (C) Total cells and (D) AF-647-positive cDC2 in axillary LN of mice 24 h after the final treatment. (E, F) Axillary LN cells purified from the mice 96 h after the final treatment were re-stimulated with 25 μg/mL OVA for 4 days. Concentrations of (E) IL-4 and (F) IL-13 in the culture supernatants of axillary LN cells. (B–F) Data are representative of two independent experiments. Means ± SD have been plotted. **P < 0.01.

Figure 5

Murine keratinocytes released IL-1α in response to stimulation with δ-toxin. Murine keratinocytes were stimulated with different concentrations of δ-toxin for 2 or 24 h, as indicated. (A, B) Concentrations of (A) IL-1α and (B) TSLP in the culture supernatants. (C) Relative expression levels of mRNA of IL-1α and IL-36α in the δ-toxin-stimulated keratinocytes. (D) The percentage of dead cells. (A–D) Data are representative of three independent experiments. Means ± SD have been plotted. *P < 0.05 or **P < 0.01.

Figure 6

Pretreatment with anti-IL-1α Ab decreased the δ-toxin-mediated translocation of OVA-loaded cDC2 from skin to draining LN in murine model. (A) Experimental design for analyzing the cytokine levels in skin tissues. (B, C) Protein levels of IL-1α in skin tissue homogenates (B) and mRNA levels of IL-36α in skin tissues (C) obtained from the mice 6 h after the first or second epicutaneous treatment with OVA ± δ-toxin on the non-tape-stripped abdominal skin. (D) Experimental design for analyzing dendritic cells in skin and axillary LN. Non-tape-stripped abdominal skin of WT mice were epicutaneously treated with OVA-AF647 ± δ-toxin on days 0 and 7. The effects of intraperitoneal administration of anti-IL-1α Ab or control Ab were examined. Skins and axillary LNs were isolated on day 8. (E) The percentage of OVA-AF647-positive cells among skin cDC2 from the mice 24 h after the last treatment. (F) AF-647-positive cDC2 in axillary LN from the mice 24 h after the last treatment. (B, C, E, F) Data are representative of two independent experiments. Means ± SD have been plotted. *, P < 0.05, **, P < 0.01.

Figure 7

Pretreatment with anti-IL-1α Ab dampened δ-toxin-mediated, OVA-induced food allergic responses in murine model. (A) Experimental design for food allergy of intragastrically OVA-administered WT mice that had been epicutaneously treated with OVA + δ-toxin once a week for six weeks on the non-tape-stripped abdominal skin. The effects of intraperitoneal administration of anti-IL-1α Ab or control Ab were examined. Blood samples were taken, and small intestines were isolated on day 57 (B) Frequency of diarrhea in OVA-administered mice. (C–E) The serum levels of (C) OVA-specific IgE and (D) MCPT-1 in the mice after the last challenge with OVA. (E) The numbers of jejunum mast cells of the mice after the last challenge with OVA. (F) Jejunum sections stained with chloroacetate esterase (scale bars, 100 μm). Mast cells stain red. (B-F) Data are representative of two independent experiments. Means ± SD have been plotted. *P < 0.05 or **P < 0.01.