Chromatin Modifications Induced by the AML1/ETO Fusion Protein Reversibly Silence Its Genomic Targets Through AML1 and Sp1 Binding Motifs

Abstract 2422

The AML1/ETO fusion protein, which is present in 10% to 15% of cases of acute myeloid leukemia (AML), is known to repress myeloid differentiation genes through DNA binding and recruitment of chromatin-modifying proteins and transcription factors in target genes. Previous studies have shown that both histone H4 deacetylation and K9 trimethylation of histone H3 (H3K9me3), respectively catalyzed by enzymes such as histone deacetylase 1 (HDAC1) and lysine methyltransferase enzyme suppressor of variegation 3–9 homolog 1(SUV39H1), are direct AML1/ETO effects on specific locus, but to date no global analysis of the AML1/ETO functional repressive chromatin modifications have been performed. ChIP-chip analysis of human hematopoietic stem/progenitor cells transduced with the AML1/ETO fusion gene (HSPC-AE) enabled us to identify 1168 AML1/ETO target genes, 103 of which were co-occupied by HDAC1 with a lost of hyperacetylation at histone H4, and 264 showed an increase H3K9me3 (p(X)<0.001; Normalized Log Ratio>0.7). We found a significant enrichment of genes involved in hematopoietic differentiation and in specific signaling pathways (i.e. TGF-β and Wnt/β-catenin) in AML1/ETO epigenetically modified target gene sets. Among them we identified novel genes (i.e. CTCF, MAPK1, SIRT1, YES1, AML1, MLLT3, and RPS19) previously described to be involved on the hematopoietic development, which epigenetic repression could implicate relevant steps in AML development induced by t(8;21). Gene Set Enrichment Analysis (GSEA) showed that expression of the 103 AML1/ETO-HDAC1 and 264 AML1/ETO-H3K9me3 target genes is highly diminished on HSCP-AE cells compared to the normal HSPC (p<0.0001 and p<0.001 respectively), suggesting that transduction with AML1/ETO directly causes this transcriptional silencing by the induction of chromatin modifications. Individual analysis confirmed a significant downregulation (p< 0.0001) of 6 out of 7 selected candidate genes. Furthermore, by using a Cre-LoxP AML1/ETO system we showed that this transcriptional repression is reversible, as the expression levels of SIRT1, AML1, CTCF, RPS19, YES1, GSK3A and MAPK1 genes were significantly restored upon AML1/ETO knockout. The clinical importance of these target genes was further analyzed in a set of 15 t(8;21) AML primary samples. All genes were found downregulated on the primary samples compared to the control HSCP, although different levels of significance were observed between them. Interestingly, AML1 binding sites were absent on a large number of the 1168 AML1/ETO target genes identified, and a Sp1 binding site was found in over 50% of them. AML1/ETO association with other motifs (e.g. E-Box motifs) has been already described and, in fact, Sp1 transcription factor is essential in hematopoietic differentiation and is known to interact with AML1/ETO. A significant enrichment of Sp1 presence in all the promoters studied was confirmed in the HSPC-AE samples compared to normal HSPC (p<0.1).Therefore, this study provides a global analysis of AML1/ETO functional chromatin modifications and identifies the important role of Sp1 in the DNA-binding pattern of AML1/ETO, suggesting a role for Sp1 targeted therapy in this leukemia subtype.