Page 37 - Haematologica August 2018

CARD10 regulates granulocytic differentiation
at intergenic (37%) and intronic regions (46%) and only 10% of the sites located in core promoter regions (Figure 1A). We integrated the gene expression profile of granulo- cytes from Cebpe WT and KO mice (RNA-seq) and CEBPE ChIP-seq analysis. A total of 312 genes were downregulat- ed in the absence of CEBPE. Of these, 93 genes had a CEBPE promoter peak, 103 genes had an intronic peak and 116 genes had an intergenic peak. Of the 116 peaks, 18 genes had a peak within 10kb from the transcriptional start site (TSS), while 98 genes had a peak beyond 10kb from the TSS (Online Supplementary Table S1).
As genome-wide distribution of enhancer and promoter associated histone modifications provides a reliable signa- ture of cell identity, we further analysed the 312 genes for histone marks in the hematopoietic lineage using a previ- ously published data set.22 We used ChIP-seq data for three histone marks; H3K4me3 (mark of active promot- ers), H3K4me1 (mark of poised enhancers) and H3K27ac (mark of poised and active enhancers) and analysed the distribution of these marks for the 312 genes in long term hematopoetic stem cells (LT-HSC), common myeloid pro- genitors (CMP), granulocyte-monocyte progenitor (GMP) and granulocytes (Gr). Expectedly, comparison of
H3K27ac, H3K4me1 and H3K4me3 profiles showed that the majority of genes with promoter binding of CEBPE had a transcriptional activation signature in the granulo- cytes but not in (LT-HSC), myeloid precursors (CMP and GMP populations) (Figure 1B). Cumulative assessment of Z scores revealed that in granulocytes, an overall increased signal intensity for all three active histone marks was detected in genes with either intronic or intergenic occu- pancy within 10kb of the TSS, as compared with genes with peaks beyond 10kb of the TSS (Online Supplementary Figure S1).
This analysis further emphasizes the role of CEBPE as a transcription factor essential for granulopoiesis and identi- fies a set of genes that might be regulated by CEBPE, with ChIP-seq peaks at non-promoter regions.
Card10/Carma3 is regulated by CEBPE in the granulocytic population
Among the 18 genes that had an intergenic binding of CEBPE within 10kb of the TSS, Card10 was an interesting target which harbors active histone marks and is exclu- sively expressed in granulocytes and monocytes (Online Supplementary Figure S2A).
AB
CDE
F
Figure 3. Card10 is a direct target of CEBPE. (A) Illustration of CEBPE binding peaks at the mouse Card10 locus. Binding sites located around -1.5kb and -7kb of the Card10 transcription start site are depicted in red. (B) FPKM values (RNA-seq) of Cebpe and Card10 in LT-HSC, CMP, GMP and Granulocytes (Gr) (https://www.ebi.ac.uk/gxa/experiments/E-MTAB-3079). (C) RT-PCR validation of transcript levels of Card10 in sorted immature granulocytes from wildtype and Cebpe KO mice. (D) ChIP-PCR validates binding of CEBPE to the -7kb region of Card10 gene using three primer pairs. Data are presented as percentage of input. (E) pGL4 basic luciferase vector containing a 500bp fragment harboring the -7kb CEBPE binding region was co-transfected with different amounts (0.2, 0.5, 0.8, and 1 μg) of pcDNA-Cebpe into NIH/3T3 cells. Luciferase activities were assayed 24 hours after transfection. Results represent fold induction of relative luciferase activity after normalization to Renilla control of two independent experiments, each done in triplicate. (F) Left panel; EMSA was performed with wild-type CEBPE binding sequence (CARD10 oligo), mutant CEBPE binding sequences (mutant oligo) and consensus CEBP binding sequences (CEBP oligo). Biotin-labelled probe was mixed with protein extracts from 293T cells transfected with either an empty vector or an expression vector for CEBPE, and the reaction mixtures were resolved on native 10% polyacrylamide-TBE gel. Cold competition was carried out with 100- fold molar excess of unlabelled oligo. Right panel; Western blot of 293T cell lysate +/- trans- fection with CEBPE expression vector used for EMSA assay. Alpha tubulin (TUBA1A) was used as loading control. ****P<0.0001, *P<0.05
haematologica | 2018; 103(8)
1273