Figure 4.
Loss of miR-146a alters epigenetic state consistent with altered HSC proliferation, differentiation, primitiveness, and function. (A) Distance matrix of pairwise cell similarity within HSC regulatory genes between single-cell DNA methylation profiles of miR-146a−/− ESLAMs, WT ESLAMs, and WT LSKs (n = 43, 84, 64 cells, respectively). Unsupervised clustering relationships and Fisher’s exact test odds ratio are shown. (B) Multidimensional scaling of distance matrix in panel A, showing relationships between cluster (top) and cell type (bottom). (C) Heat map of percentage methylation of DM-TFBSs in single cells from panel A. TFBS names with the suffix .1 indicate that 2 TFBS ChIP-seq profiles for the same TF, collected in different cell types, were identified. Unsupervised clustering relationships and Fisher’s exact test odds ratio are shown. (D) Principal component analysis (PCA) of single-cell DM-TFBS methylation profiles in panel C. (E) TFBS contributions to PCA in panel D. Based on cell type contributions to dimension 1 shown in panel D, TFBSs hypermethylated in miR-146a−/− ESLAMs appear on the right, and the hypomethylated ones appear on the left. (F) GSEA plot showing enrichment of NF-κB target genes in miR-146a−/− vs WT LSK HSPCs. ChIP-seq, chromatin immunoprecipitation sequencing; FDR, false discovery rate; NES, normalized enrichment score.