While we and others have prospectively identified potential origins of leukaemic stem cells (LSCs) resulting in phenotypically identical myeloid leukaemia, the functional differences among these origin-specific LSCs have not been vigorously investigated. Recently, our lab has shown that β-catenin, which is dispensable for normal HSCs, is essential for MLL LSCs, highlighting it as a potential therapeutic target for selective eradication of LSCs while sparing normal HSCs. Here, I investigated the role of β-catenin in LSCs of distinctive cellular origins using both RTTA (in vitro) and bone marrow transplantation (in vivo) assays. I demonstrated that conditional deletion of β-catenin abolished the leukaemogenic potential of LSK-Meis1-Hoxa9 pre-LSCs whereas CMP-Meis1-Hoxa9 pre-LSCs were able to develop into LSCs in vivo regardless of β-catenin status. Interestingly, conditional inactivation of β-catenin abolished the in vivo leukaemogenic property of GMP-MLL pre-LSCs whereas LSK-MLL pre-LSCs were still able to induce leukaemia in vivo in the absence of β-catenin, revealing functional differences in origin-specific LSCs in spite of their ability to induce phenotypically identical leukaemia. Since a major function of β-catenin is to mediate stem cell self-renewal, I hypothesized the presence of multiple alternative self-renewal pathways in normal HSCs that may allow the LSK/HSC-origin specific LSCs overcoming inactivation of β-catenin. Hox genes have been previously shown by others and us as key players in mediating self-renewal in normal and malignant haematopoiesis. Consistently, Hoxa9 exhibited functions as β-catenin in mediating developing of MLL pre-LSCs originated from GMP but not HSCs. To test if Hoxa9 may compensate the loss of β-catenin in LSK-MLL pre-LSCs, I have created a conditional β-catenin/Hoxa9 knockout model where a combination of these two molecules can be inactivated in normal HSCs and various myeloid progenitors. As a result, I demonstrated that specific inactivation of β-catenin or Hoxa9 abolished the oncogenic potentials of GMP-MLL pre-LSCs, but not LSK-MLL pre-LSCs. Strikingly, inactivation of both pathways suppressed oncogenic transformation by LSKMLL pre-LSCs and resulted in down-regulation of several Meis1-Hoxa9 target genes. Among them is Prmt1, which has been identified as a critical epigenetic modifying enzyme associated with an oncogenic MLL fusion complex. Prmt1 knockdown inhibited clonogenic activity of LSKMLL-ENL LSC with Hoxa9 or β-catenin knockout, suggesting Prmt1 as a key modulator for transformation of LSK-MLL-ENL LSC. Together, I have demonstrated, for the first time, specific molecular and functional differences among origin-specific LSCs, in which crosstalk between multiple self-renewal pathways inherited from the cell of origins may determine the biology of the disease and mediates resistance to potential targeted therapies.