Methyl‐β‐cyclodextrin concurs with interleukin (IL)‐4, IL‐13 and IL‐25 to induce alterations reminiscent of atopic dermatitis in reconstructed human epidermis

Reconstructed human epidermis (RHE) mimic normal human in vivo epidermis in terms of histology, distribution of differentiation markers and barrier functionality.1 A typical transcriptional profile and the activation of signalling pathways reminiscent of atopic dermatitis (AD) lesional skin can be obtained in RHE upon incubation with methyl-β-cyclodextrin (MβCD),2, 3 a molecule that extracts cholesterol from plasma membranes, thereby disrupting lipid microdomains. However, barrier function and morphology remain unaltered in those conditions, requiring further refinement of the model. Because of the crucial role played by Th2 immune response in AD, a mixture of interleukins linked to this Th2 response (IL-4, IL-13 and IL-25) was used in addition to MβCD, in an attempt to induce most of the epidermal AD features in RHE. Ultimately, a valid RHE model of acute AD would allow studying the epidermal component of pathogenesis. Reconstructed human epidermis was incubated for 2 hours with MβCD in order to induce cholesterol depletion (Fig. S1), then for 48 hours with the interleukin mix and compared to control tissues, and to tissues incubated with either MβCD or interleukins. When allowed to recover in fresh culture medium after incubation with MβCD for 2 hours, RHE revealed no obvious histological alteration when compared with untreated RHE (Fig. 1a). Conversely, RHE treated with IL-4, IL-13 and IL-25 for 48 hours displayed intercellular space widening similar to spongiosis (already reported for IL-4 and IL-13 4-6) and hypogranulosis, two histological hallmarks of lesional AD skin (Fig. S2). These morphological alterations were enhanced when keratinocytes were challenged by MβCD before being incubated with the three interleukins (Fig. 1a). Barrier function, weakened in AD, was studied using two assays: measurement of trans-epithelial electrical resistance (TEER) and assessment of permeability to the fluorescent dye lucifer yellow (LY) through the RHE. A significant decrease in TEER was observed in RHE incubated with the interleukins, and worsened upon membrane cholesterol depletion (Fig. 1b). Accordingly, permeability of RHE towards LY was significantly increased after combined treatments, whereas cholesterol depletion alone or incubation with interleukins only was insufficient to elicit an effect (Fig. 1c). Then, the expression of atopic dermatitis markers was investigated. Gene expression of filaggrin (FLG) and loricrin (LOR), epidermal differentiation genes usually downregulated in lesional AD, was analysed through RT-qPCR (Fig. 2a). FLG and LOR exhibited reduced mRNA levels in RHE incubated with IL-4, IL-13 and IL-25, as already reported in the literature regardless of the different concentrations and timings used.5 This decrease was exacerbated and became significant when keratinocytes were challenged by cholesterol depletion before incubation with the interleukins. Simultaneously, relative expression levels of carbonic anhydrase II (CA2) and neural epidermal growth factor-like 2 (NELL2), two genes upregulated in AD,7 were significantly enhanced in RHE in response to incubation with interleukins (Fig. 2a). Immunohistological analysis confirmed decreased expression of LOR and increased expression of CA2 in AD lesions (Fig. S3a). Similar changes were observed in RHE treated with interleukins and became even more evident in tissues previously treated with MβCD (Fig. S3b). Expression levels for hyaluronan synthase 3 (HAS3) were found elevated in RHE exposed to interleukins and further increased after combined treatments (Fig. 2a), in agreement with the upregulation observed in AD.4, 8 Though, no significant changes were found with respect to HAS1 expression levels (data not shown), unlike data collected from AD lesions.8 Fluorescent detection of hyaluronic acid (HA) revealed increased staining of intercellular spaces between keratinocytes in AD lesions (Fig. S4a). HA was also more strongly detected in RHE incubated with interleukins, particularly after previous challenging by MβCD (Fig. S4b). Accordingly, increased HA concentrations were measured in culture medium of treated RHE (Fig. S4c). Thymic stromal lymphopoïetin (TSLP) is a cytokine, upregulated in AD, which contributes to Th2 immune response activation, promotes itch in skin and the “atopic march” in general.9 However, despite significant induction in response to 2 hours of MβCD (Fig. S1c), mRNA expression levels for TSLP were no longer above baseline 48 hours later (Fig. 2a). However, in the meantime, over the 6 hours period following cholesterol depletion (Fig. 2b), data illustrate that interleukins and MβCD concur to significantly enhance TSLP expression in challenged RHE. This study illustrates that challenging RHE through cholesterol depletion, or by incubation with IL-4, IL-13 and IL-25, results in multiple epidermal alterations. But interestingly, a combination of those two treatments has additive effects, allowing mimicking an AD-like epidermal phenotype in vitro in the absence of immune cells. Indeed, morphological alterations such as tissue spongiosis and hypogranulosis, alterations in mRNA expression levels and histological localizations of typical AD and differentiation markers, modulations of epidermal HA synthesis and epidermal barrier weakening represent hallmarks of AD epidermis. In vivo, such alterations become responsible for activation of the immune system because they promote penetration of pathogens or allergens, thereby creating some vicious circle likely responsible for AD lesions development (Fig. 2c). Most probably, challenging of keratinocytes by MβCD models alterations in cell signalling through disorganization of specific lipid microdomains in this cell type.3 Interestingly, gene expression levels for IL-13Rα2, IL-13Rα1, IL-4Rα and IL-17RA (subunit of IL-25 receptor) were upregulated in keratinocytes incubated with MβCD.3 This indicates that cholesterol-containing lipid microdomains could regulate signalling through these receptors and could therefore explain the additive effects observed in this study. In conclusion, this study confirms that membrane cholesterol depletion in keratinocytes concurs with Th2-related cytokines to elicit an AD-like phenotype. The present model could be used to study other features encountered in AD epidermis, but also evaluate compounds intending to relieve, prevent AD lesions or restore keratinocyte functions. We thank B. Balau, V. De Glas, K. De Swert and D. Van Vlaender for technical help. EDV, SG and MB performed the study. EDV, JM and AM analysed the data. YP, AC and MS designed the study. AN contributed the AD biopsies. EDV, CLdR and YP wrote the manuscript. AC and MS are employees of Straticell. Financial support was provided by the Région Wallonne (BAREPI-convention-1217660) and FRFC 2.4.522.10F grant to YP. The authors stated that they have no conflict of interest. Data S1 Supplementary methods and material Data S2 Supplementary references Figure S1 Effect of incubation with MbCD for 2 hours on RHE analysed at day 11 of tissue reconstruction Figure S2 Morphology of healthy and lesional AD skin biopsies Figure S3 Labeling of atopic dermatitis-associated markers Figure S4 Hyaluronic acid (HA) staining and concentration in the culture medium Table S1 Oligonucleotides used to perform the RT-qPCR Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.