GRHL3 chromatin binding and the super-enhancer landscape are reorganized in different functional states of epidermal keratinocytes

While the genomic mechanisms underlying progressive, irreversible cell lineage commitments are well-studied, we know little about the chromatin changes during transient cell states such as cell migration. Interestingly, a large number of SEs in NHEK-D and NHEK-M overlap genes encoding transcription factors with important roles in promotion of epidermal differentiation, including GRHL3, TP63, RUNX1, NOTCH3 and FOS. To test the role of these SE-associated transcription factors in a systematic manner, and to place GRHL3 in the context of other keratinocyte differentiation regulators, we used siRNAs to individually knock down GRHL3 and 50 other transcriptional regulators in NHEK-D. These transcriptional regulators were selected based on expression changes during human keratinocyte differentiation. Many have been previously implicated in epidermal differentiation in mice and humans; the genes encoding 22 of them are associated with SEs. To assess the effect of the knockdowns on keratinocyte differentiation, we monitored the expression of approximately 14,000 genes with custom-made Agilent microarrays. Among the 14,000 genes whose expression we monitored were all genes expressed in human keratinocytes and all transcriptional regulators. This 51 x 14,000 gene expression matrix provided a rich dataset to explore gene regulatory networks in epidermal differentiation. 50 transcription factors with known or suspected roles in epidermal differentiation were selected for siRNA knockdown. Each factor showed a significant change in expression during epidermal differentiation and had a well-defined DNA binding motif. Pooled siRNA for each factor (Dharmacon) were used for knockdown in NHEK. 24 hours after knockdown, cells were induced to differentiate by the addition of high calcium media, and collected 24 hours later. RNA was extracted and run on custom designed arrays. Each factor was knocked down as a single replicate, and four samples of scramble control siRNA were also included for comparison. Differential gene expression was determined by CyberT. R was used for PCA analysis, and hierarchical clustering of the data. Gene Enrichment Analysis was performed using GSEA. Weighted Gene Correlation Analysis (WGCNA) was used to develop network level descriptions of the siRNA data, and to link factors to the genes they are predicted to regulate. Analysis was limited to genes that showed differential expression across epidermal differentiation.