Development of highly sensitive methods for defining off target mutations to enable safe gene editing of haematopoietic cells for transplantation

RNA guided endonucleases (RGENs), such as the CRISPR-Cas9 system have revolutionised our ability to edit the genome because they allow us to cut and edit the genome at the sequence defined by a RNA guide, which can be rapidly designed and manufactured. This has had a huge impact on molecular biology; allowing the genome of a wide variety of different cell types and organisms to be edited. The next challenge is to use this technology to edit the genome of human cells to treat disease. Haematopoiesis is one of the first areas in which this technology will be successfully implemented in the clinic because haematopoietic stem cell (HSC) transplantation has been performed for over 40 years and facilities for collecting, purifying and transplanting human HSCs are well established.

One of the major barriers to using genome editing technology in patients is the potential risk of malignant transformation from unintended mutations during the editing process. In order to improve genome editing techniques to the point at which they can be safely used in human trials it is critical that assays are developed to define mutations in primary cells that have undergone genome editing. These off target effects need to be defined accurately in an unbiased genome wide fashion and the assays need to be highly sensitive so that rare off target mutations can be detected because malignancy arises from clonal transformation of single cells. To date, several methods have been described for determining off target effects but none of the techniques is able to define mutations sensitively and accurately in primary cells. This is a key stumbling block to the clinical application of these techniques because without rapid and cost effective methods to compare the accuracy of the huge variety of different techniques available for genome editing it will be difficult to develop safe methods of editing for use in human trials.

I will develop a novel method that initially identifies all potential sites of in vitro RGEN in naked DNA (this has previously been shown to be a highly predictive of potential sites of off target activity). Subsequently these sites will be sequenced at great depth using biotinylated oligonucleotides designed to capture DNA at the target sites identified in vitro. Using this it should be relatively straightforward to make the assay is at least 100 times more sensitive than the best available methods. I will also use this approach to look for rare sites of unintended insertion of viral vectors and template sequences.

The overall burden of mutations that occur during the editing process will also be investigated. This will be done by performing whole genome sequencing on haematopoietic stem and progenitor cells purified by flow cytometry both before and after the editing process. Methods will be used to minimise sequencing errors and software will be written to define the overall burden of mutations from the variability of the sequences derived from sequencing.

These methods will be used to optimise two models of genome editing for the treatment of thalassaemia and sickle cell disease for potential clinical use. First a model has been developed in the host laboratory which aims to cure patients with transfusion dependent HbE beta thalassaemia, which accounts for around 50% of transfusion dependent thalassaemia, by deleting key transcription factor binding sites at the alpha globin gene. In addition an established method will be set up, which uses an RGEN to insert a DNA template to correct the beta globin gene in situ and simultaneously express a cell surface marker that allows purification of corrected cells. This will allow me to test the methods for quantification of off target effects on the two main strategies for using RGENs for editing cells.

Sites of off target CRISPR-Cas9 nuclease activity will initially be defined in purified genomic DNA. Digenome-seq shows that RGEN cleavage in vitro is a good predictor of off target sites but it uses whole genome sequencing to identify cut sites, making it prohibitively expensive and insensitive. The novel method involves ligating a customized sequencing adaptor to RGEN cut sites to allow them to be specifically sequenced.

I have previously developed oligonucleotide capture technology for extremely deep sequencing (>100,000 fold coverage). Biotinylated oligonucleotides will be designed to pull down all of the potential sites of Cas9 cleavage identified in vitro and these will then be used to define with high sensitivity whether off target mutations occur during in vivo editing of primary cells. In addition oligonucleotides will be designed to the viral vector and the sequences to be inserted by HDR so that off target sits of integration can be identified. I plan to investigate whether a chromosome conformation capture based method (Capture-C) I pioneered, can improve the sensitivity of the assay detect to sites of off target integration.

Off target effects will be studied in two models of genome editing for the treatment of thalassaemia and sickle cell disease. The first model has been developed in the host laboratory to treat HbE beta thalassemia and uses a single CRISPR-Cas9 cut and the non homologous end joining pathway to delete key transcription factor binding sites at the main enhancer of the alpha globin gene. A previously published method will also be set up, which uses homologous recombination to correct the beta globin gene in situ and simultaneously insert a cell surface marker for purification of corrected cells. This uses an adeno-associated virus (AAV) 6 vector to deliver the template for HDR and I plan to use a second AAV-6 vector to deliver Cas9 controlled by an inducible promoter to study how Cas9 expression alters off target effects.

The outcomes of this research would be of particular benefit to researchers in the field of genome editing and clinicians in the field of bone marrow transplantation and cellular therapy. However, it has the potential to have a much broader impact. The development of methodology to define off target effects will further our understanding of the potential risks of genome editing and will allow us to develop safe methods for editing the genome of potentially any cell type or organism. Once developed, the methodology will be patented as it is likely that it will be of commercial value. Commercial exploitation of intellectual property will be achieved through the Oxford University Innovation Ltd. licensing and ventures team.

Importantly, I envisage that the work will also lead to a method of quantifying the general mutational burden in haematopoietic stem cells. This could also have a broad impact on current medical practice because it will allow us to define the damage caused by different cytotoxics and chemotherapy regimens to haematopoietic stem cells and the ancillary risk of secondary myelodysplasia.

Development of methods for safe genome editing of haematopoietic stem cells and lymphocytes has potential the to have a therapeutic impact on a very wide range of diseases. It is likely that long term cure for inherited genetic diseases of haematopoiesis, such as sickle cell disease, thalassaemia, Wiskott-Aldrich syndrome and Fanconi anaemia will be achieved within the next decade. At present autologous bone marrow transplantation is becoming an important treatment modality for some autoimmune diseases, in particular multiple sclerosis. Genome editing of autologous haematopoietic stem cells therefore has the potential to provide long term cure for a wide variety of autoimmune conditions. Engineered Chimeric Antigen Receptor (CAR) T cells are currently an important treatment modality for several forms of haematological malignancy including chronic lymphocytic leukaemia and acute lymphoblastic leukaemia. Interestingly, it has recently been demonstrated that targeting of chimeric antigen receptors to the endogenous T cell receptor locus using CRISPR-Cas9 improves the efficiency of CAR T cells. The power of this technology is likely to improve and it is likely to be applicable to a broader range of malignancies in the future. Finally it is also likely that engineered haematopoietic cells will be used to treat infectious diseases. It has been shown that allogeneic transplantation with a CXCR5 negative donor can cure HIV. Although trials of engineered lymphocytes with CXCR5 knockdown were unsuccessful for treating HIV, autologous transplantation with HSCs with knockdown of CXCR5 has the potential to be curative analogous to allogeneic transplantation.

The work to be developed in this project is likely to provide a rapid method to identify mutations resulting from genome editing and this has the potential to allow the field to improve the genome editing technology more rapidly. Since developing genome editing technology is an international effort, the research is likely to have a significant impact on the UK profile in this area. If successful the impact of the research could be very rapid indeed (within 12 months).