Promotes Gene Expression by Disrupting the DNA Binding of the AML1/ETO and HDAC1 Repressor Complex.

Abstract

In t(8;21) AML, the AML1/ETO fusion protein promotes leukemogenesis by recruiting histone deacetylase (HDAC) and silencing target genes important for hematopoiesis. In addition to its anticonvulsant properties, valproic acid (VPA), an eight-carbon branched-chain fatty acid, has shown a significant antitumor activity mediated, at least in part, by inhibition of the HDAC enzymatic activity. The molecular mechanisms through which VPA inhibits HDACs and restore gene expression in cancer cells remain, however, to be fully elucidated. Herein, we showed that in AML1/ETO-positive cells, VPA modulates HDAC activity by releasing HDAC1 and AML1/ETO from the DNA binding sites at concentrations attainable in pts. Nuclear and whole cells extracts were obtained from AML1/ETO-positive Kasumi-1 cells untreated or treated with VPA (0.3 to 3 mM) for 24 hours and subjected to immunoblotting with AML1, ETO or HDAC1 antibody. Decreased levels of HDAC1 was noticed in the nuclear extract, but not in the unfractionated cell lysates following VPA treatment. These changes correlated with a dose-dependent inhibition of HDAC enzymatic activity and global hyperacetylation of histone H3 and H4. Decrease of HDAC1 and AML1/ETO in the nucleus was likely due to dissociation of the repressor complex proteins from DNA and relocation in the cytoplasm as indicated by a “band-depletion” assay. Untreated or VPA treated Kasumi-1 cells were cross-linked with 1% formaldehyde to stabilize the binding of proteins to their target DNA sequences, and subjected to immublotting with HADC1, AML1 or ETO antibody. While free protein can be resolved on a polyacrylamide gel by immunoblotting, high-molecular weight DNA-protein complexes cannot. Following formaldehyde cross-linking, in untreated cells, the AML1/ETO-HDAC1 complex was stabilized onto the DNA binding sites thereby forming a low-mobility DNA-protein complex which resulted in the absence of the corresponding gel band. In contrast, following VPA treatment, high-intensity AML1/ETO and HDAC1 bands were visualized, suggesting a decrease in DNA binding and an increase in the free-pools of the AML1/ETO and HDAC1 proteins. No changes were observed in HDAC1 and AML1/ETO RNA or protein levels in cells not cross-linked with formaldehyde, thereby excluding upregulation of the corresponding genes following VPA treatment. Similar results were confirmed at the promoter of specific AML1/ETO-target genes (i.e., IL-3). Following VPA exposure, decreased levels of HDAC1 and AML1/ETO, and histone H3 and H4 hyperacetylation were detected on the IL-3 promoter by chromatin immunoprecipitation. The functional relevance of these changes was supported by detection of mRNA expression of IL-3, which was otherwise silenced in untreated cells. Notably, the activity of VPA to HDAC1 and AML1/ETO appeared quite specific, as levels of other factors that initiate histone acetylation and/or gene transcription such as HAT1 and Pol II did not decrease, rather increased on the IL-3 promoter. Further, following VPA exposure, we observed posttranslational changes of specific H3 and H4 lysine residues associated with gene transcription (H3K9 and H4K16 acetylation and H3K4 methylation), partial cell differentiation, cell cycle arrest via upregulation of p21 and caspase-induced apoptosis. Taken together, these data suggest VPA as a promising compound for targeting molecular subgroups of AML, especially those in which leukemogenesis is promoted by aberrant HDAC activity.