Abstract
Background: The chromosomal translocation t(7;11)(p15,p15) encodes the oncogenic transcription factor NUP98-HOXA9 which results in a fusion of the nucleoporin 98kDa (NUP98) and homeobox A9 (HOXA9) genes. The oncogenic mechanisms underlying this translocation remain poorly understood and patients are currently inadequately served by traditional cytotoxic chemotherapy regimens.
Aims:To decipher the underlying biology of the NUP98-HOXA9 fusion protein and develop rational therapeutic strategies targeting its oncogenic mechanism.
Methods: Human cellular models expressing NUP98-HOXA9, HOXA9 wt or NUP98 wt were established by retroviral transduction of HEK293FT human cell line and human hematopoietic progenitors (CD34+, hHP) isolated from donor cord blood. Chromatin immunoprecepitation experiments followed by sequencing (ChIP-seq) and quantitative ChIP (qChIP) were used to define fusion specific binding locations. Cloning regulatory regions of selected target genes in a luciferase vectorconfirmed the direct involvement of NUP98-HOXA9 in their regulation. RTQ-PCR and gene expression microarrays were used to evaluate expression levels. Co-Immunoprecipitation experiments validated protein-protein interactions and drug treatments were performed at IC50. Cell viability was analysed by apoptosis, proliferation and Colony Forming Unit assays.
Results:Comparison of ChIP-seq data from HEK293FTmodels of NUP98-HOXA9, HOXA9 wt or NUP98 wt respectively, identified 4,471 target genomic regions of the fusion protein (FDR < 0.05), located within +4/-4 kb from the annotated Transcription Start Site (TSS) of 1,363 genes, with 399 genes common to HOXA9 wt and 5 to NUP98 wt. The NUP98-HOXA9 binding sites included enhancers of MEIS1, HOXA9 and PBX3 (PBX3 and HOXA9 were common to NUP98 wt and MEIS1 to HOXA9 wt). Together these transcription factors form a key activator complex that regulates the expression of genes involved in leukemogenesis and its overexpression is significant related to adverse prognosis in AML. Luciferase assays showed that the upregulation of this leukemic axis was directly induced by the interaction of NUP98-HOXA9 with the corresponding enhancer regions of MEIS1, HOXA9 and PBX3. Treatment of cells with HXR9, a specific peptide inhibitor of HOXA9 and PBX3 interaction, led to a selective decrease in the proliferation of hHP expressing NUP98-HOXA9, confirming the relevance of these target genes to its oncogenic mechanism. Combining ChIP-seq and gene expression data of three independent clones of hHP expressing NUP98-HOXA9 and patient samples (n = 5) harbouring t(7;11)(p15,p15) revealed a dual regulatory role of the fusion protein, in both repressing and activating target gene transcription where, for example, MEIS1, HOXA9, PBX3 and AFF3 were found overexpressed and BIRC3, SMAD1, FILIP1L and PTEN downregulated. Interactions of NUP98-HOXA9 with p300 and HDAC1 were shown to drive this transcriptional activation and repression, respectively. We found using qChIP experiments that p300 bound to the regulatory regions of the overexpressed genes only when NUP98-HOXA9 was present, whereas we observed significant enrichment of HDAC1 binding to the promoter regions of the downregulated genes when the fusion protein was expressed. Taking advantage of this latter observation, we demonstrated a dramatic inhibitory effect on the viability of hHP expressing NUP98-HOXA9after the treatment with subtherapeutic doses (IC50 = 4nM) of the HDAC inhibitor LBH-589 (Panobinostat) with no effect in control hHP transduced with an empty vector.
Conclusion: An improved understanding of the pathobiology underlying recurrent translocation events in AML is a critical first step for the development of rational, targeted therapies. Here, we identify upregulation of the targetable MEIS1-HOXA9-PBX3 complex underpinning the leukemogenic activity of NUP98-HOXA9. Its activity in repressing transcription mediated through interaction with HDAC1, has been shown to be also a key pathogenic mechanism that can be exploited through use of HDAC inhibitors and potentially lead to a promising new therapy for this high-risk group of patients.
Disclosures
No relevant conflicts of interest to declare.
From enhanceropathies to the epigenetic manifold underlying human cognition