Skip to main content
. 2019 Sep 12;134(24):2195–2208. doi: 10.1182/blood.2019001553

Figure 4.

Figure 4.

Cohesin deficiency severely impairs erythroid differentiation but only modestly alters HSPC homeostasis. (A) Experimental design to determine cohesin dynamics following perturbation of Stag2. (B) Average profiles of Rad21 binding at active promoters and active enhancers in Luc_HPC and shS2_HPC cells. To minimize global quantitative biases, the Rad21 signal was CPM normalized and internally calibrated to its signal at the insulator/other sites (see “Methods and materials”). Shown are means of 2 replicates, whereas each short hairpin RNA was considered an independent replicate (Luc1_HPC, Luc2_HPC, shS2_1 HPC, shS2_2_HPC). (C) Differential gene expression between Luc_HPC and shS2 HPC cells (3 biological replicates representing different knock-down clones per group: Luc1_HPC, Luc2_HPC, Luc3_HPC, shS2_1_HPC, shS2_2_HPC, shS2_3_HPC) demonstrates minimal gene expression changes. The blue shaded area consists of genes enriched in HPC, whereas the brown shaded area shows genes significantly expressed in shS2_HPC cells. The numbers at the upper corners show the counts of log2FC > 0.5, padj < 0.1, and (log2FC > 1, padj < 0.05) significant differential genes. For comparison, the significance and difference spreads were set intentionally to match those in panel F. (D) Representative flow cytometric analysis of erythroid differentiation following EPO for Luc_Ery and shS2_Ery at the indicated times. (E) Variance by principal component analysis of the shCohesin_HPC, Luc_HPC, shS2_ery, and Luc_ery RNAseq datasets. Each dot represents a different short hairpin RNA (Luc1-3_HPC, shS2_1-3_HPC, shR21_1-3_HPC, shSmc1a_1-2_HPC, Luc1-2_ery, shS2_1-2_ery). (F) Differential gene expression between Luc_ery and shS2_ery cells (2 biological replicates for each group: Luc1_ery, Luc2_ery, shS2_1_ery, shS2_2_ery). The numbers at the upper corners show the counts of log2FC > 0.5, padj < 0.1, and (log2FC > 1, padj < 0.05) significantly differentially expressed genes. (G) NES for significant (FDR > 0.05 datasets) hematopoiesis-related datasets as determined by GSEA. Input was a ranked list of all pajd < 0.1 genes from panel F. (H) Venn diagrams of gene expression changes displaying overlaps between erythroid or immature genes from Figure 2F and Luc_ery or shS2_ery genes from Figure 4F. CPM, counts per million; FDR, false discovery rate; NES, normalized enrichment score.