Deciphering hypoxia signaling pathways at the transcriptional and post-transcriptional level in leukaemogenesis

Haematopoietic stem cells (HSC) reside at the top of a hierarchy capable of reconstituting the blood system of an organism throughout its life. HSCs inhabit the bone marrow where the low partial pressure of oxygen and support cells form the HSC niche. Acute myeloid leukaemia (AML) is an aggressive cancer of the haematopoietic stem and progenitor cells with only 20% to 25% of AML patients surviving 5 years or longer following diagnosis, and relapse occurring in about 50% of patients who achieve remission following initial treatment. Leukaemic stem cells (LSC) in AML are capable of self-renewal and generation of leukaemic cells, competing with healthy HSCs in the HSC niche and remodelling the niche to support LSCs at the expense of normal haematopoiesis. Current therapies often fail to eradicate LSCs leading to relapse. N6-methyladenosine (m6A) is a prevalent internal RNA modification in mammalian cells which is implicated in both cancers and hypoxia responses. I developed a novel pipeline for the analysis of m6A-Seq utilising currently available tools and compared the pipeline to the popular MACS2 method for m6A-Seq analysis. The novel pipeline includes a means to test for di'erentially methylated genes, employing a linear model to filter transcripts whose m6A changes correlate closely with RNA template changes. This pipeline was validated in silico using data generated by the Kranc group and publicly available data. With this pipeline, I analysed m6A-Seq data from leukaemic and healthy cKit+ cells generated by the Kranc group and demonstrated di'erential m6A methylation in key genes implicated in leukaemogenesis. Combining published half-life data with results from m6A-Seq, I also found a correlation between width of peaks, their position, and half life of the transcript, providing a new angle of research. I analysed data generated by the Kranc lab in preleukaemic cells to produce a gene expression signature associated with hypoxia response. This gene signature was used to establish a diminished hypoxia response in leukaemic cells deficient in YTHDF2, an m6A reader which mediates the degradation of m6A-methylated mRNA. Combining my findings from leukaemic cells deficient in HIFs (hypoxia inducible factors), I showed despite a down-regulation of hypoxia response genes in YTHDF2-deficient LSCs, predicted HIF1 targets are upregulated, and YTHDF2-deficient LSCs specifically downregulate HIF1 targets which are upregulated under hypoxia responses. I also demonstrated that HIF1 and HIF2 targets overlap and the observed synergistic dysregulation of genes in HIF1 and HIF2 double deficient LSCs upon hypoxia response are mostly due to redundancy of HIF function. Finally, computational analyses of gene and protein expression combined with m6A methylation results from my pipeline revealed processes contributing to the suppression of leukaemia in YTHDF2-deficient LSCs, and I proposed a new model for the mechanisms underlying YTHDF2-deficiency in LSCs which is consistent with published literature, linking m6A regulation, hypoxia responses, and the unfolded protein response.