Transcriptional landscape of epithelial and immune cell populations revealed through FACS-seq of healthy human skin

Human skin consists of multiple cell types, including epithelial, immune, and stromal cells. Transcriptomic analyses have previously been performed from bulk skin samples or from epithelial and immune cells expanded in cell culture. However, transcriptomic analysis of bulk skin tends to drown out expression signals from relatively rare cells while cell culture methods may significantly alter cellular phenotypes and gene expression profiles. To identify distinct transcriptomic profiles of multiple cell populations without substantially altering cell phenotypes, we employed a fluorescence activated cell sorting method to isolate keratinocytes, dendritic cells, CD4+ T effector cells, CD4+ Treg cells, and CD8+ T effector cells from healthy skin samples, followed by RNA-seq of each cell population. Principal components analysis revealed distinct clustering of cell types across samples, while differential expression and coexpression network analyses revealed transcriptional profiles of individual cell populations distinct from bulk skin, most strikingly in the least abundant CD8+ T effector population. Our work provides a high resolution view of cutaneous cellular gene expression and suggests that transcriptomic profiling of bulk skin may inadequately capture the contribution of less abundant cell types. Overall design: Transcriptomic profiles from keratinocyte, dendritic cell, CD4+ T cell, CD4+ Treg cells, and CD8+ T cell populations were obtained from surgical skin discards from 11 healthy adults. Cell populations from whole skin were sorted via FACS and transcripts generated using an Illumina HiSeq 2500 platform. RNA-seq data for the bulk control samples were originally deposited in GEO study GSE74697.

51

Ahn RS, Taravati K, Lai K, Lee KM, Nititham J, Gupta R, Chang DS, Arron ST, Rosenblum M, Liao W