Modelling the in-utero initation of ETV6-RUNX1 in childhood acute lymphoblastic leukaemia using human pluripotent stem cells

Abstract

BackgroundWe hypothesise that the clinical differences between adult and childhood acute lymphoblastic leukaemia (cALL) arise partly through their origin in developmentally distinct target cells. Since cALL frequently initiates in utero we aimed to characterise the earliest stages of lymphoid development in human fetal liver. In parallel we used human pluripotent stem cells (hPSC) to recapitulate fetal liver lymphopoiesis and model the impact of the fusion oncogene ETV6-RUNX1, the commonest genetic aberration in cALL. MethodsH1 human embyronic stem cells and MIFF3 human induced hPSCs (provided by University of Sheffield Centre for Stem Cell Biology) were maintained in vitro with mTeSR1/Matrigel. Vectors were produced by recombineering an ETV6 bacterial artificial chromosome with a custom DNA cassette (GeneArt, ThermoFisher, Waltham MA, USA). Vectors were transfected (Nucleofector II, Lonza, Basle, Switzerland) and G418-selected clones were screened by Southern blot. hPSCs were differentiated by sequential OP9/MS5 coculture. Populations sorted by fluorescence-activated cell sorting were analysed by single cell real-time PCR (Biomark 48.48, Fluidigm, San Fransisco, CA, USA) and 200 cell RNA sequencing. Human fetal livers were donated under informed consent with approval of Lund University Ethical Review Board and the Swedish National Board of Health and Welfare. FindingsB lymphopoiesis in the fetal liver was distinct from that in adult bone marrow or neonatal cord blood: the earliest fetal liver B cells did not fully express mature lymphoid effectors and the earliest identified lymphoid-capable progenitors coexpressed both lymphoid and myeloid gene expression programmes. Global and single cell gene expression profiling of B lymphopoietic progenitors generated by in vitro differentiation of hPSCs showed that they share the same transcriptional programme. hPSCs that were CRISPR-engineered to express ETV6-RUNX1 were partly arrested in B cell differentiation at the level of the lymphomyeloid progenitor. B cells that passed this block shared the global gene expression signature of the fetal lymphomyeloid progenitor, and single cell real-time PCR showed that they aberrantly coexpressed B and myeloid lineage genes. InterpretationOur results identify a B lymphoid progenitor in human fetal liver characterised by coexpression of lymphoid and myeloid gene expression programmes and suggest that this progenitor's B myeloid signature provides a permissive transcriptional context for ETV6-RUNX1 to effect a partial differentiation arrest. The developmental specificity of the gene expression signature of this cell could offer unique therapeutic targets, and the hSPC model might provide a novel drug-screening platform. FundingWellcome Trust Research Training Fellowship and NIHR Academic Clinical Fellowship (SR); Swedish Childhood Cancer Foundation (CB); Bloodwise, Cancer Research UK, Children with Cancer, and Great Ormond Street Hospital Children's Charity (TE).