Combined Th2 and Th17/22 cytokines models atopic dermatitis pathophysiology in vitro

In atopic dermatitis (AD), epidermal disease hallmarks are driven by a complex cutaneous inflammatory milieu that varies between patients. How these variable inflammatory signatures translate to patient-specific disease endotypes is poorly understood, which hampers the personalized use of targeted drugs. We here employed primary and immortalized keratinocyte-derived human epidermal equivalents stimulated with T helper (Th)2, -17 and -22 cytokines, to unravel which AD-associated cytokines drive specific epidermal disease hallmarks. Th2 cytokines IL-4 and IL-13 were main inducers of a pro-inflammatory and hyperproliferative response. The presence of IL-17A or IL-22 in the Th2 milieu caused spongiosis, impaired keratinocyte differentiation (loss of FLG, IVL, KRT2) and epidermal barrier defects (altered CLDN1, CLDN4, transepithelial impedance). Transcriptomic analyses revealed upregulation of proliferation genes by Th2 cytokines, and activation of different immune pathways upon Th2 cytokines alone and combination of Th2 and Th17/22 cytokines. Downregulation of epidermis development was most apparent upon combined exposure to Th2 cytokines and IL-22. Single-cell spatial transcriptomics showed expansion of keratinocytes expressing high level of proliferation genes in all epidermal layers, and downregulation of terminal differentiation genes in the upper epidermal layers. Our in vitro AD model using Th2 + IL-22 showed the highest transcriptome resemblance with in vivo AD lesional epidermis and we identified IL-22 specific alteration of epidermis development by lipid mediators like ACER1 and AKR1C3. Finally, we revealed that Aryl Hydrocarbon Receptor (AHR) ligands partially restored the barrier defects caused by Th2 and IL-22, whereas combination of AHR ligand Tapinarof with Janus Kinase (JAK) inhibitor I had an even greater restoring effect. Thus, specific combinations of cytokines upregulated in AD skin drive differential disease hallmarks highlighting the importance of personalized medicine and the need for prediction models in the targeted treatment of AD patients.