Preclinical modelling of childhood acute lymphoblastic leukaemia

Abstract

BackgroundGreatly improved outcomes for children with leukaemia have come at the cost of intensive and toxic treatment regimens. New precision drugs are needed and must be validated in the most representative preclinical models before they can be assessed in clinical trials for children. We have developed an orthotopic patient-derived xenograft model of acute lymphoblastic leukaemia (ALL) that allows real time in-vivo tracking of disease engraftment, candidate gene analysis, and therapeutic validation. MethodsPatient-derived leukaemic blasts were transduced with a lentivector expressing enhanced green fluorescent protein (in-vitro analysis) and firefly luciferase (in-vivo tracking). Transduced cells were injected intrafemorally into highly immunocompromised NOD/LtSz-scid IL-2Rγ-/- (NSG) mice. Stable expression of the lentivector allowed bioluminescent tracking of disease progression over serial transplantations. Modification of the lentivector to include short hairpin RNA sequences has allowed simultaneous in-vivo tracking and gene knockdown. Furthermore, this model has been applied to the validation of a BCL2 inhibitor, ABT-737, as a novel targeted therapy in BCR-ABL positive ALL. FindingsKnockdown of the fusion oncogene MLL/AF4 impeded engraftment of the infant ALL cell line SEM. Treatment of mice with ABT-737 over 6 weeks substantially impaired leukaemia growth and prolonged the median survival from 151 to 190 days (p=0·0004). InterpretationIn this study, we have developed a lentiviral approach to the in-vivo preclinical validation of gene function, coupled with subsequent analysis of novel therapies. It combines use of patient-derived material with a suitable host microenvironment to minimise the risk of false-positive identification of new therapies. This approach minimises the risk to children enrolled into early-phase clinical studies and will accelerate development of effective treatments. FundingUK Medical Research Council, Cancer Research UK grant (C27943/A12788). The IVIS spectrum was funded by Welcome Trust grant 087961.