Mapping MLL-AF4 fusion protein-binding sites in human leukemia cells. (A) Schematic diagram of strategy for mapping MLL-AF4 fusion protein-binding sites. SEM precursor B acute leukemia cells express the MLL-AF4 fusion protein. REH precursor B acute leukemia cells express only endogenous AF4 and MLL1. The N terminus of MLL (blue) is recognized by ChIP antibody anti-MLL-N (blue) and immunoprecipitates both wild-type MLL and MLL-AF4 fusion protein in SEM cells. The C terminus of AF4 (red) is recognized by ChIP antibody anti-AF4-C and immunoprecipitates both wild-type AF4 and MLL-AF4 fusion protein in SEM cells. Wild-type AF4 and MLL-N are immunoprecipitated by anti-AF4-C and anti-MLL-N, respectively. (B) Binding of AF4 (red) and MLL-N (black) in SEM cells (top panels) and REH cells (bottom panels) as determined by ChIP-seq. Binding profiles are shown across an 800-kb portion of the genome surrounding the PROM1 gene (gene models shown in black below graph; a black arrow indicates transcription start sites). MLL-AF4 fusion protein binding is indicated by a red bar.

MLL-AF4 target genes are enriched for early developmental regulators. (A) Signals for AF4 (red) and MLL-N (black) reflecting binding of the presumptive MLL-AF4 fusion protein in SEM cells as determined by ChIP-seq. Binding profiles are shown across a 15- to 150-kb portion of the genome surrounding the HOXA9, TWIST1, HMGA2, and RUNX2 genes (gene models shown in black below graph; a black arrow indicates transcription start sites). Size of MLL-AF4-enriched region is indicated by top brackets. A detailed description of data analysis methods is provided in the Supplemental Material. (B) Composite AF4-C terminus (red) and MLL-N terminus (black) binding profiles for all MLL-AF4 target genes. The start site and direction of transcription of the average gene are indicated by an arrow. (C) ChIP-seq density heat map of AF4-C terminus (red) and MLL-N terminus (black) for all MLL-AF4 target genes. The genomic region from −5kb to +10kb relative to the transcription start site of each gene is shown. Gene order is determined by highest average MLL/AF4 read density from top to bottom. The start site and direction of transcription of the genes are indicated by an arrow. (D) Selected results of GSEA (http://www.broad.mit.edu/gsea) of MLL-AF4 target genes .

MLL-AF4 target genes define MLL-linked leukemia in vivo. Hierarchical clustering of relative expression levels of 42 genes occupied by MLL-AF4 fusion protein target regions >10 kb. Comparisons were made across the SEM and REH cell lines and 132 peripheral blood samples of patients diagnosed with leukemia. Each row corresponds to a gene that is bound by MLL-AF4 for which expression data were available. Each column corresponds to a single gene expression microarray. For each gene, expression is shown relative to the average expression level of that gene across all samples, with shades of red indicating higher than average expression and green lower than average expression. Columns and rows were ordered by unsupervised hierarchical clustering. A detailed description of data analysis methods is provided in the Supplemental Material.

Mistargeting of chromatin modifications occur as epigenetic lesions at MLL-AF4 target regions. (A) Binding of H3K79me2 (green) and H3K4me3 (blue) in SEM cells as determined by ChIP-seq. Binding profiles are shown across an 800-kb portion of the genome surrounding the PROM1 gene. Gene models shown in black below the graph; a black arrow indicates transcription start sites. MLL-AF4 fusion protein-binding regions are indicated by red bars. (B) Binding of H3K79me2 ChIPs (green) and H3K4me3 ChIPs (blue) in SEM cells as determined by ChIP-seq. Binding profiles are shown across 15- to 150-kb portions of the genome surrounding the HOXA9, TWIST1, HMGA2, and RUNX2 genes. Gene models are shown in black (below graph). A black arrow indicates transcription start sites. MLL-AF4 fusion protein binding is indicated by a red bar. (C) H3K79me2 (green) and H3K4me3 (blue) binding profiles for all MLL-AF4 target genes in SEM cells. Genes are ordered as in Figure 2C. (D) Composite H3K79me2 ChIP enrichments (green) and H3K4me3 ChIP enrichments (blue) for all genes (left) and all MLL-AF4 target genes (right). The start site and direction of transcription of the average gene are indicated by an arrow.

Model of MLL-AF4-mediated gene activation in leukemia. Schematic diagram of transcriptional misregulation in MLL-AF4-induced leukemia. MLL-AF4 associates with elongation proteins (pTEFb, ENL, and DOT1) at sites of epigenetic lesions. Phosphorylated and elongating RNA Pol II shown in yellow. Blue (H3K4me3) and green (H3K79me2) circles represent histone modifications at aberrant chromatin domains. Genes annotated as bound by MLL-AF4 and playing a role in hematopoietic stem cell function are shown to the right and subdivided into functional categories.