Acute myeloid leukaemia (AML) is an aggressive subgroup of leukaemia characterised by the accumulation of immature blasts. Prognosis of AML is often poor due to dose-limiting toxicities and unresponsiveness to standard chemotherapy. Differentiation therapy describes therapeutics which act to release the characteristic differentiation block of AML blasts, facilitating differentiation and the natural clearance of cells, yet clinical success to this point has been limited by high rates of relapse. Recent research conducted in murine cells has identified the capacity for mature AML-derived cells of persistent lineages to de-differentiate, highlighting a potential avenue for relapse. Here we use leukaemias driven by a class I NRAS mutation and a Doxycycline repressible class II MLL-AF9 fusion protein to investigate AML differentiation therapy response. We have characterised the differentiation response of our leukaemias both in vitro and in vivo. Upon repression of MLL-AF9 expression in both in vitro and in vivo settings, immature leukaemia cells bifurcate into two mature lineages: short-lived neutrophil-like cells that are rapidly cleared, or long-lived macrophage-like cells that contribute exclusively to residual disease. We have recently obtained transcriptomic and secretomic data for these leukaemias, which has returned a large pool of candidates potentially driving neutrophilic and monocytic differentiation. Using CRISPR we hope to validate macrophage-depleting candidates. In a complementary approach, we are currently depleting AML-derived macrophages in leukaemic mice following differentiation therapy by using clodronate liposomes, a well-established method for eliminating macrophage populations in vivo. We anticipate that targeted clearance of AML-derived macrophages will reduce minimal residual disease and relapse. Ultimately we hope to translate these findings to existing therapeutics to minimise relapse following differentiation therapy.