Optimising CRISPR for targeting of primary haematopoietic stem cells

CRISPR (clustered regularly interspaced short palindromic repeats) offers an unprecedented array of genetic tools for treating and modelling haematological disorders. However, there are concerns about its safety. CRISPR-based editing can induce upregulation of TP53 associated pathways, a finding particularly important for clinical applications. The impact of CRISPR-based editing on haematopoietic stem cells (HSCs) has not, however, been interrogated in detail. In the work described in this thesis, I initially optimised the CRISPR approach in mouse haematopoietic stem and progenitor cells (HSPCs) by targeting the fluorescent reporter tdTomato, as this enabled lineage tracking of edited HSCs. I then sorted edited HSCs for RNA-Sequencing (RNA-Seq) 24 hours post-editing. This allowed me to assess the immediate transcriptional consequences of genome editing. Using in vivo transplantation, I serially tracked the contribution of edited cells to all blood lineages. I carried out RNA-Seq on edited HSCs 18 weeks post-transplant which allowed me to assess whether the short-term changes in transcriptional profiles were maintained long-term. To demonstrate the applicability of the genome editing approach I had optimised to a disease model setting, I targeted the NF-1 gene which harbours loss of function mutations in 15% of JMML patients. While mouse models exist for this disorder, most lack a human fetal context. I first optimised targeting of the NF-1 locus in a GM-CSF dependent cell line (TF-1). I then carried out in vitro NF-1 targeting studies using human cord blood (CB) to establish a GM-CSF hypersensitivity phenotype. Finally, I set up in vivo transplants of NF-1 edited human CB to establish whether this could propagate the myeloid biased expansion characteristic of JMML. In this way, I applied the CRISPR genome editing approach to study both normal and malignant haematopoiesis: first to establish the molecular and functional consequences of genome editing on normal haematopoietic output and second to model the consequence of NF-1 loss of function in a human fetal-like context.