The Mechanisms of HOXA9-mediated Acute Myeloid Leukaemia Maintenance

Abstract

Acute myeloid leukaemia (AML) is an aggressive haematological malignancy, which, despite recent advances in treatment, results in only 50% long-term survival for the younger and lower than 10% for the older patients. Understanding the molecular mechanisms underlying leukaemogenesis and leukaemia maintenance is essential for the development of novel targeted therapies and improving AML outcomes. Work described in this thesis examines the molecular mechanisms through which developmental transcription factor HOXA9 supports the AML maintenance. Beyond its normal roles in embryonic development and maintenance of the haematopoietic progenitor cells, HOXA9 is overexpressed in 50-70% of AMLs, is a harbinger of poor treatment outcomes and is an established therapeutic target: KMT2A-fusion, NUP98-fusion and mutated NPM1 driven AMLs not only highly express HOXA9 but also depend on HOXA9’s expression for survival, whereby downregulation of HOXA9 is associated with differentiation and apoptosis. Despite its established causative roles in leukaemia development and maintenance, we know surprisingly little about how HOXA9 works either in normal or malignant haematopoiesis. Therefore, elucidation of mechanisms through which HOXA9 exerts its leukaemogenic function could lead to novel therapeutic approaches in treating HOXA9-driven AML. This thesis describes the development of the endogenously tagged 3xFLAG-HOXA9 human AML cell lines, which allowed me to overcome the bottleneck in the study of HOXA9 – the absence of specific anti-HOXA9 antibodies. Utilising the newly established and validated 3xFLAG-HOXA9 system, I describe a comprehensive HOXA9 interactome, identify high confidence HOXA9 binding sites in AML and demonstrate that HOXA9 does not show features of an RNA-binding protein. I find that HOXA9 binds to many expected, but novel, transcriptional regulators and epigenetic modifiers. Unexpectedly, I discover that HOXA9 also binds to a large cluster of DNA-damage response (DDR) proteins and demonstrates a high degree of functional co-dependency with interacting partners involved in replication stress relief. As a transcription factor in the human AML setting, HOXA9 binds not only to active enhancers, as reported previously, but also demonstrates extensive binding to active promoters, primed enhancers and chromatin regions lacking activating marks.To study the non-transcriptional roles of HOXA9 in DDR, I engineered a bifunctional endogenous HALO-HOXA9 system in human AML cell lines that enabled both the acute degradation of HOXA9 as well as fluorescent tracking of the HOXA9 protein. I first utilised the acute degradation properties of this system to describe the role of HOXA9 in regulating chromatin accessibility and to identify the direct transcriptional targets of HOXA9, while avoiding secondary effects bias. HOXA9 differentially regulates the expression of over 3000 genes. Transcriptome profiling following acute HOXA9 depletion demonstrates that HOXA9 drives proliferation transcriptional programmes, while suppressing the expression of pivotal DNA damage signal transducers. Work laid out in this thesis demonstrates that HOXA9 upregulates the expression of multiple TFs, required for leukaemia maintenance, and sits at the apex of the AML core regulatory circuitry. Through integration of CHIP-seq, HOXA9-degradation RNA-seq analyses with developmental cis regulatory elements, I demonstrate that in AML HOXA9 binds and supports the activity of leukaemia-specific and developmental enhancers and plays a core role in the maintenance of HSC enhancer activity. By integrating the experimentally determined HOXA9 transcription targets with external CRISPR functional screens, I identify the HOXA9 downstream targets necessary for the maintenance of HOXA9-expressing and HOXA9-dependent AML. In addition to multiple already established oncogenic AML dependencies, such as BCL2, DOT1L, MLLT1, JMJD1C, KAT7, I identify several other novel downstream HOXA9 targets and susceptibilities in HOXA9-driven AML that can be explored therapeutically. Utilising the acute degradation of HOXA9, I was able to disentangle the transcriptional and protein functions of HOXA9 and uncover a non-transcriptional function of HOXA9 in preventing replication stress and DNA damage. Through HOXA9 degradation and RNA:DNA hybrid mapping using DRIP-seq, I demonstrated that HOXA9 reduces the R-loop accumulation over bound promoters and active enhancers, irrespective of its transcriptional function. Using break labelling in situ and sequencing (sBLISS), I mapped sites of dsDNA breaks to demonstrate that HOXA9 reduces dsDNA breaks around its binding regions, particularly enhancers. These findings were further supported by the increased DDR protein binding to HOXA9 chromatin binding sites upon HOXA9 degradation. Finally, validation experiments of HOXA9 interactors demonstrate that HOXA9-dependent AML cell lines are specifically dependent on the BTR complex and identifies RECQL family helicase BLM as both functional and interacting partner of HOXA9 in AML maintenance. The overarching theme from work covered in this thesis emerges that HOXA9 supports leukaemia maintenance by promoting cell growth and proliferation programs transcriptionally, while simultaneously protecting cells from oncogene-induced stress through transcriptional and non-transcriptional means. Characterisation n of HOXA9 downstream targets and binding partners identifies novel vulnerabilities in HOXA9-driven AML that can be explored therapeutically.