To maintain a balanced production of all mature haematopoietic lineages and at the same time secure a stem cell reservoir, intricate regulatory programs have evolved to control multi-lineage differentiation and self-renewal in haematopoietic stem and progenitor cells. Leukaemogenic mutations commonly disrupt these regulatory programs causing a block in differentiation with simultaneous enhancement of proliferation. AML is a genetically heterogeneous disease characterized by chromosomal rearrangements that often produce fusion proteins with aberrant transcriptional regulatory activities. Among those, t(11;19) generates the MLL-ENL fusion protein, an aberrant transcriptional factor with an unusual mechanism of transcriptional activation targeting the elongation step of transcription. MLL rearrangements develop a unique category of AML and ALL, mainly affecting infants and paediatric patients, whose transcriptional profile differentiates them from other types of leukaemias. While multiple mechanisms driving this disease have been reported, early cellular and molecular consequences of MLL-ENL expression are still poorly understood.

The first aim of this thesis was the generation of a new in vitro mouse model of MLL-ENL driven AML. Since even highly purified haematopoietic stem/progenitor cell populations are recognised to be heterogeneous, it is impossible to define a precise wild-type parental control in conventional retroviral transduction leukaemia models. To circumvent this problem, the strategy devised here was based on the mouse haematopoietic progenitor cell line Hoxb8-FL. Hoxb8-FL cells are able to proliferate indefinitely in vitro in the presence of Flt3 ligand (Flt3L) and an estrogen (β-estradiol)-regulated form of Hoxb8, and show lympho-primed multipotent progenitor (LMPP)-like differentiation capacity both in vitro and in vivo. Transduction experiments generated MLL-ENL expressing Hoxb8-FL cells which, following cytokine enriched culture conditions, generated a preleukaemic cell line able to cause AML development in vivo.

The second aim of the thesis was to investigate the transcriptional consequences of MLL-ENL expression at early stages of preleukaemic evolution. The main approach applied was scRNA-seq since cell fate choices, including the stepwise acquisition of a malignant phenotype, are made by single cells. These studies were carried out culturing cells either in a self-renewal condition or in differentiating condition since the previously reported requirement of myeloid differentiation for AML development.

The third aim of the thesis was to explore whether any of the early leukaemogenic events identified, could represent genetic vulnerabilities specifically associated with MLL-ENL expression. To this end, genome-wide CRISPR-Cas9 screening was performed on the new AML model developed. The data obtained were integrated with scRNA-seq results in order to identify and validate new potential therapeutic targets associated with early MLL-ENL driven transcriptional changes.

In summary, the approaches developed and validated within this thesis have generated a new in vitro murine AML model which has been exploited to gain novel insights into the early steps leading to MLL-ENL driven leukaemogenesis and to validate new candidate therapeutic strategies.