Staphylococcus aureus Exploits Epidermal Barrier Defects in Atopic Dermatitis to Trigger Cytokine Expression

Abstract

Patients with atopic dermatitis (AD) have an abnormal skin barrier and are frequently colonized by S. aureus. In this study we investigated if S. aureus penetrates the epidermal barrier of subjects with AD and sought to understand the mechanism and functional significance of this entry. S. aureus was observed to be more abundant in the dermis of lesional skin from AD patients. Bacterial entry past the epidermis was observed in cultured human skin equivalents and in mice but was found to be increased in the skin of cathelicidin knockout and ovalbumin-sensitized filaggrin mutant mice. S. aureus penetration through the epidermis was dependent on bacterial viability and protease activity, because killed bacteria and a protease-null mutant strain of S. aureus were unable to penetrate. Entry of S. aureus directly correlated with increased expression of IL-4, IL-13, IL-22, thymic stromal lymphopoietin, and other cytokines associated with AD and with decreased expression of cathelicidin. These data illustrate how abnormalities of the epidermal barrier in AD can alter the balance of S. aureus entry into the dermis and provide an explanation for how such dermal dysbiosis results in increased inflammatory cytokines and exacerbation of disease.

Conflict of interest statement

The authors state no conflict of interest.

Copyright © 2016 The Authors. Published by Elsevier Inc. All rights reserved.

Figures

Figure 1. Dysbiosis of the subepidermal compartments from skin of AD patients

(a) qPCR results for relative abundance of DNA for 16S rRNA, (b)S. aureus (c) and S. epidermidis detected in epidermal or dermal compartments, isolated by laser-capture microdissection (LCM), of skin from normal skin of non-AD subjects, nonlesional and lesional skin of patients with AD. Non-tissue controls (NTC) were simultaneously processed. Data represent mean ± SEM of 11 subjects. *P<0.05, **P<0.01. Epi=epidermis, Derm=dermis, UD=undetectable. (d) 16S rRNA pyrosequencing results from samples isolated by LCM of the epidermis and dermis of nonlesional (NL) and lesional (L) skin of AD subjects. Each bacterial phylum is shown in different color. (e–g) Immunofluorescence for S. aureus and keratin-14 in lesional (e) or nonlesional (f) skin of AD subject. Staining with isotype control (g). (h,i) Immunofluorescense for S. aureus and CD11c in lesional skin of AD subject. Staining with isotype control (i). Arrows indicate S. aureus staining detected outside of CD11c+ immune cells. Immunostaining shown is a representative of 3 biopsies from different donors. Scale bar= 20μm (white) or 200μm (yellow).

Figure 2. S. aureus actively penetrates human skin organotypic equivalents

Time-dependent entry of S. aureus across the epidermis of a human skin organotypic construct. Viable S. aureus (1×106 CFU) (a–c) or UV-killed S. aureus (1×107 CFU) (d) were applied on the stratum corneum surface and individual constructs fixed at the indicated time after bacterial application. Skin constructs were then sectioned and stained with anti-S. aureus (green) to visualize bacteria. Keratinocyte nuclei were counter stained with DAPI (4′,6-diamidino-2-phenylindole) (blue). Scale bar=20 μm. Arrows indicate immunoreactivity for S. aureus under the epidermal surface when live bacteria were applied to the surface.

Figure 3. Cathelicidin inhibits S. aureus entry into the mouse dermis

(a–c)S. aureus (ATCC35556) or vehicle (mock) were loaded in agar discs and applied on dorsal skin of Camp−/− or WT mice for 20 hrs. Skin was then excised and DNA extracted from epidermis (a), dermis (b) or adipose tissue (c) isolated by LCM. Relative colony forming units (rCFU) of S. aureus DNA was determined by real-time qPCR by comparison to a standard of known CFUs of S. aureus (ATCC35556). (d) As negative control a non-tissue control (NTC) was simultaneously processed with the same reagents from embedding material adjacent to each tissue section. The data were normalized against tissue volume excised by LCM. Data represent mean±SEM of results from 4 independent experiments.*P<0.05. UD=undetectable

Figure 4. A loss-of-function mutation in filaggrin increases S. aureus entry into the mouse dermis after ovalbumin sensitization

(a) Transepidermal water loss (TEWL) determined before (−) or after tape-stripping (+) on the dorsal skin of FLGft/ft Balb/c or WT mice which were treated by repeated applications of OVA or PBS. (b–e) The backs of FLGft/ft Balb/c mice, or wild type (WT) Balb/c mice were treated with tape-stripping and OVA as described in panel (a). Abundance of S. aureus (ATCC3555) in epidermis (b), dermis (c) and adipose tissue (d) was measured by qPCR and LCM as described in Figure 3. (e) NTC was simultaneously processed as negative control. Data represent mean ± SEM of results from 6 independent experiments. *P<0.05, **P<0.01, ***P<0.001. UD=undetectable

Figure 5. S. aureus protease activity is required for penetration of the epidermis and induction of inflammatory cytokines

(a) Entry of WT or an extracellular protease-null mutant strain of MRSA into organotypic human skin constructs . (b–e) Entry of WT, extracellular protease-null mutant strain and UV-killed WT strain of MRSA into epidermis (b), dermis (c) and adipose tissue (d) of FLGft/ft Balb/c mice sensitized by OVA was tracked as described in Figure 3. NTC was processed as negative control (e). (f–o) To correlate entry of MRSA strains with cutaneous immune response, gene expression of indicated cytokines (f–n) and indicated AMPs (o) was measured in the same whole skin biopsies from panel (b–d). To compare relative expression level of each AMP, data was shown as relative to GAPDH expression. Data represent mean ± SEM of results from 5–6 independent experiments. *P<0.05, **P<0.01, ***P<0.001.

Figure 6. Skin barrier repair decreases S. aureus entry into the dermis and suppresses subsequent immune response in skin with FLG mutation

(a) Effect of the application of barrier repair formula of ceramide-triple lipid mixture on TEWL from FLGft/ft Balb/c mice sensitized with OVA. TEWL was measured 4 hrs after application of barrier repair lipids or vehicle. (b–e) Effect of the application of barrier repair lipid mixture on entry of S. aureus into the skin of FLGft/ft Balb/c mice with AD-like inflammation. Entry of S. aureus (ATCC3555) into epidermis (b), dermis (c) and adipose tissue (d) was tracked 4 hrs after application of barrier repair lipids or vehicle. NTC was simultaneously processed as negative control (e). Data represent mean ± SEM of results from 5–6 independent experiments. (f–i) Effect of the application of barrier repair lipid mixture on cytokine and Camp inductions after epicutaneous application of S. aureus on the OVA-sensitized skin of FLGft/ft Balb/c mice. Data represent mean ± SEM of results from 6–7 independent experiments. *P<0.05, **<0.01, ***P<0.001. UD=undetectable