MICA: Multiplexed genome editing for stealth and persistence of hypoimmunogenic 'universal' CAR T cells

Immunotherapy has greatly improved outcomes in patients who have relapsed with refractory leukaemias. This has relied mostly on manipulating patient's own immune cells, known as T cells, and arming them with a chimeric antigen receptor (CAR) to recognise and kill tumour. However, obtaining enough cells from patients or of sufficient quality, especially in from children, remains a key obstacle driving the search for alternative cell sources from healthy donors. One major issue is the need to match donors/recipients to prevent immune rejection.

Recent developments in genome editing have allowed for 'universal' CAR T cells from healthy volunteers to be made by removing flags from the surface of T cells that would otherwise be recognised as foreign. A single 'universal' CAR T cell product has the potential to treat >20 patients. thereby eliminating variability of bespoke manufacture, reducing costs to the healthcare system and importantly increasing timely accessibility to multiple patients who have otherwise failed all other treatments. Early phase clinical trials at UCL GOSH have demonstrated their potential to eliminate B cell leukaemia in several patients. Initial versions used TALEN genome editing tools to create targeted breaks in the DNA for the removal of the T cell receptor, a key driver of rejection, and of a surface marker recognised by a therapeutic antibody used for pre-conditioning. The first modification acts to make the cells 'invisible' whereas the second protects the CAR T cells from being destroyed by the antibody. We have generated newer versions using CRISPR/Cas9 technology that are currently being tested in clinic. While mostly successful, in some cases the 'universal' CAR T cells were cleared by the patient before they could eliminate the leukaemia.

Increasing stealth by removing additional foreign flags from their surface could help tackle these hurdles and allow CAR T cells to persist long enough to clear disease. We now have cutting-edge genome editing tools, known as base editors, that offer a safer way of making multiple modifications to the DNA without compromising on safety. This project will test this strategy and compare different routes to efficiently and safely manufacture stealthy 'universal' CAR T cells for application in leukaemias and solid tumours.

Off-the-shelf genome edited allogeneic CAR T cells have been hindered by instances of rejection or short term persistence suggesting a remaining level of visibility from possible interplay between HLA Class I and II. Newer base converting genome editing tools allow for the seamless disruption of multiple genes simultaneously within the same cell. Base editors can be employed for the targeted disruption of TCR, HLA class I & II to create next generation hypoimmunogenic stealthy CAR T cells against both liquid and solid tumours. The technology will be developed as a core strategy to prevent GVHD and promote stealth. The strategy will be modelled using new CAR targets against CD70, an emerging candidate in immunotherapy for haematological malignancies and renal cell carcinoma.

WP1 Enhancing stealth through base editing of surface HLA molecules
sgRNA against TCR, HLA Class I & II will be designed and synthesised and efficiency and molecular fidelity in primary T cells will be determined by validated in vitro assays.

WP2 Modelling anti-CD70 CAR
Anti-CD70 scFv sequences will be designed and cloned into a CRISP-CAR vector. In vitro and in vivo cytotoxic potency will be evaluated in models of B-ALL/AML/RCC.

WP3 Allo-resistance of hypoimmunogenic 'universal' anti-CD70 CAR T cells
Scaled manufacture of CRISPR-CAR70 cells with multiplexed base edits for stealth and persistence. In vitro and in vivo evaluation of enhanced protective advantage conveyed to TCRab- CAR T cells.

WP4 High throughput interrogation of editing fidelity - characterisation of safety
Interrogation of on- and off-target edits to evaluate editing fidelity vs promiscuity by high throughput techniques.

WP5 Mitigation strategies
Non-canonical HLA genes will be incorporated to address 'missing self' NK mediated killing. Base editor versions with increased fidelity and lower indel frequency will be evaluated. Suicide gene will be incorporated to mitigate against adverse toxicity events.