Spatial transcriptomics combined with single-cell RNA-sequencing unravels the complex inflammatory cell network in atopic dermatitis

Background: Atopic dermatitis (AD) is the most common chronic inflammatory skin disease with complex pathogenesis for which the cellular and molecular crosstalk in AD skin has not been fully understood. Methods: Skin tissues examined for spatial gene expression were derived from the upper arm of 6 healthy control (HC) donors and 7 AD patients (lesion and nonlesion). We performed spatial transcriptomics sequencing to characterize the cellular infiltrate in lesional skin. For single-cell analysis, we analyzed the single-cell data from suction blister material from AD lesions and HC skin at the antecubital fossa skin (4 ADs and 5 HCs) and full-thickness skin biopsies (4 ADs and 2 HCs). The multiple proximity extension assays were performed in the serum samples from 36 AD patients and 28 HCs. Results: The single-cell analysis identified unique clusters of fibroblasts, dendritic cells, and macrophages in the lesional AD skin. Spatial transcriptomics analysis showed the upregulation of COL6A5, COL4A1, TNC, and CCL19 in COL18A1-expressing fibroblasts in the leukocyte-infiltrated areas in AD skin. CCR7-expressing dendritic cells (DCs) showed a similar distribution in the lesions. Additionally, M2 macrophages expressed CCL13 and CCL18 in this area. Ligand–receptor interaction analysis of the spatial transcriptome identified neighboring infiltration and interaction between activated COL18A1-expressing fibroblasts, CCL13- and CCL18-expressing M2 macrophages, CCR7- and LAMP3-expressing DCs, and T cells. As observed in skin lesions, serum levels of TNC and CCL18 were significantly elevated in AD, and correlated with clinical disease severity. Conclusion: In this study, we show the unknown cellular crosstalk in leukocyte-infiltrated area in lesional skin. Our findings provide a comprehensive in-depth knowledge of the nature of AD skin lesions to guide the development of better treatments.