Acute myeloid leukaemia (AML) remains a deadly disease, largely due to the persistence of drug-resistant leukaemia-initiating cells (LICs) which promote relapse. Therefore, effective therapies must target LICs. Patient-derived xenografts (PDXs) are valuable for testing new therapies, though establishing AML PDX models is challenging. Two humanized mouse strains, MISTRG and NRGS, have been developed for this purpose. In this study, we show both are suitable strains for the development of AML PDXs
however, MISTRG-derived PDXs contain 10 times higher LIC frequencies than NRGS-derived PDXs. These differences have crucial implications for preclinical AML therapy testing and modelling relapse models of the disease.