Search

Chimeric antigen receptor (CAR) T cell therapy represents a step-change improvement in outcomes for patients with treatment-resistant haematological malignancies. Despite this success, these results have not been replicated when CAR T cells are directed against solid tumours. The reasons for this are multifactorial, however clinical trials consistently report that a small minority of patients treated with CAR T cells for solid tumours will achieve favourable clinical responses. This suggests that while current CAR T cell designs are insufficient for treating solid tumours, there remains scope for optimisation that will improve the efficacy of this therapy. In line with this concept, multiple different 'armoured' CAR T cell designs have been described that aim to overcome the factors that limit CAR T cell anti-tumour activity. The work in this thesis describes efforts to genetically manipulate the metabolism of CAR T cells in order to improve their anti-tumour activity. Initially, the aim of these efforts was to enhance T cell oxidative phosphorylation (OXPHOS) in order to promote the persistence of CAR T cells after adoptive transfer. Knockdown of the metabolic regulatory proteins Drp1 or Rictor failed to achieve this, however expression of a constitutively AMPK mutant was successful at enhancing OXPHOS in CAR T cells. Despite this, no improved anti-tumour activity was observed by CAR T cells expressing this AMPK mutant, possibly due to the absolute requirement for T cells to conduct anaerobic glycolysis to fuel effector functions. As an alternative approach, efforts were made to enable CAR T cells to overcome metabolic competition for low glucose concentrations in the solid tumour microenvironment (TME). In vitro modelling of CAR T cell effector functions at low glucose concentrations revealed that key effector functions are inhibited by low glucose concentrations mimicking those found in the TME. Attempts to armour CAR T cells to overcome this by overexpressing glucose transporters or glycolytic enzymes were unsuccessful, however two approaches to engineer CAR T cells to utilise an alternative carbon source (ACS) to glucose, trehalose and fructose, were successfully validated in vitro. These data provide the foundation for future work to demonstrate that engineering CAR T cells to metabolise an ACS to glucose will result in improved anti-tumour activity in in vivo human tumour models.

T-cells engineered with a second generation CD19 targeted chimeric antigen receptors (CARs) have recently shown great promise for the treatment of haematological malignancies, with three therapies now approved for the treatment of acute lymphoblastic leukaemia and diffuse large B-cell lymphoma. Despite the success of these and other CARs in haematological malignancies, CAR therapy for solid tumours has encountered numerous obstacles. Even with a number of trials ongoing in solid tumours using second-generation CAR T-cells, it is apparent that these CARs are generally not effective in achieving an anti-tumour response. This indicates that there is a need for further development of CARs to improve their efficacy in the treatment of solid tumours. One way to achieve this is to re-engineer these receptors to deliver increased costimulation. This may help to increase the persistence, survival and anti-tumour activity of therapeutically infused CAR T-cells. In this thesis, I set out to deliver enhanced costimulation using a novel CAR system known as a parallel CAR (pCAR). Parallel CARs comprise the co-expression of a second-generation CAR and a chimeric co-stimulatory receptor (CCR) which delivers a different co-stimulatory signal. To direct cells against multiple solid tumour types, pCAR T-cells were targeted against the pan-ErbB family. CAR and CCR specificity was directed against pan-ErbB using the EGF-TGFα chimeric protein, T1E, as the targeting moiety. Using mutated versions of T1E with reduced affinity for the epidermal growth factor receptor (EGFR; ErbB1), I investigated the relationship between the affinity of CAR and CCR for their respective target antigens. Decreasing the affinity of T1E in the CAR position decreased the anti-tumour activity of the CAR T-cells. Competition for the target antigen was ruled out by replacing the T1E CCR with a peptide targeted against a second tumour-associated antigen, αvβ6 integrin. When tested in vitro, pCARs with wildtype 3 T1E in both CAR and CCR or an αvβ6-targeted peptide in the CCR position showed better activity compared to a second-generation CAR. However, in vivo experiments were less conclusive, suggesting that further experimentation is warranted. To test the pCAR format in a different ErbB-expressing tumour model, an scFv derived from the EGFR-specific antibody ICR62 was cloned into the pCAR format with T1E as a CCR. This was evaluated in models of glioblastoma since EGFR was identified as a valid target in this disease. In vitro testing again showed superiority of the pCAR over an equivalent second-generation CAR, however, again, further experimentation is required to confirm if there is any superiority of the pCAR in an in vivo model of glioblastoma. In summary, pCARs show potential as a further evolution of CAR T-cells, providing greater persistence and cytokine support in vitro with potentially better persistence in vivo, however individual optimisation for each target is required.

Prostate cancer (PCa) is the most common malignancy in men in the UK. However, the 5-year survival rate of stage IV PCa is 30% and the median survival for the castrate resistant phenotype is less than 2 years. This illustrates a clinical need for improved therapies for metastatic castrate resistant prostate cancer. Chimeric antigen receptor (CAR) T cell therapy has shown unprecedented success in blood cancers. The application of this platform in solid cancers, including PCa, has been less efficacious. A 2nd generation CAR, P28z, has previously been developed to retarget T cells to PSMA. When co-expressed with the IL-4 responsive 4αβ chimeric cytokine receptor, the resultant P4 transduced T cells proliferate and enrich in the presence of IL-4. The aim of my project was to develop an IL-4 immunocytokine targeted to the tumour microenvironment to provide P4 CAR T cells with a proliferative signal and thereby increase anti-tumour efficacy. Fibroblast activation protein (FAP) was chosen as the target for the immunocytokine because of its overexpression in the stroma of epithelial cancers. An immunocytokine designed from a published anti-FAP single chain variable fragment (scFv) sequence, eFAP-4, showed FAP specificity and signalling capabilities through the 4αβ receptor. For drug development, PCa cell lines were characterised and developed to express PSMA and reporter genes. Additionally, FAP+ stromal cell lines were characterised and engineered with a reporter gene. When P4 CAR T cells were cultured on PSMA+FAP+ PCa/stroma monolayers in the presence of eFAP-4, increased T cell proliferation, cytokine secretion and persistence were observed. However, when PCa xenografts were established in vivo, stromal engraftment was insufficient. To model this interaction, PCa cells were transduced with FAP. In a pilot study using this PSMA+FAP+ PCa xenograft model, increased survival and tumour control was seen in the group treated with P4 + eFAP-4. This study has indicated the potential for the improvement of CAR T cell immunotherapy of PCa using tumour microenvironment targeted immunocytokines.

Cancer immunotherapy using chimeric antigen receptor (CAR) T-cells has shown exceptional promise in the treatment of patients with refractory B-cell malignancy. In this approach, patient-derived peripheral blood T-cells are engineered to express a cell surface receptor, which confers specificity for a tumour-associated (TA) antigen. Mucin-1 (MUC1) is a large transmembrane glycoprotein that is overexpressed in 90% of breast cancers. A further important characteristic of this mucin is the fact that it is under-glycosylated in cancer cells. This holds the potential for CAR T-cell mediated targeting of MUC1 epitopes in tumour-cells which are not exposed in normal cells. Antibodies such as TAB004 and HMFG2 are considered to bind preferentially to TA-MUC1. The aim of this PhD project was the development of a CAR Tcell approach for MUC1-positive breast cancer. Herein, the anti-tumour potential of a novel 2nd generation MUC1-specific CAR, named TAB28z, has been investigated. The binding domain of this CAR is derived from the TAB004 anti-MUC1 antibody. TAB28z is being compared with two other previously developed MUC1-specific CARs, H28z and HDF28z, both of which are derived from the HMFG2 antibody. TAB28z CAR T-cells demonstrated significant anti-tumour activity against MUC1-positive breast cancer cell lines in the in vitro setting. Nevertheless, it became apparent throughout this project that MUC1 expressed on activated T-cells is detected by both HMFG2-based and TAB004-based CAR T-cells during the T-cell expansion period. This background recognition resulted in tonic signalling by CARs, which was accompanied by constitutive production of IFN-γ, CAR T-cell enrichment, reduced T-cell expansion and a trend towards upregulation of T-cell activation and exhaustion markers. Despite these observations, the activity of the three MUC1-specific signalling CARs was investigated in two different breast cancer xenograft models. No significant anti-tumour responses were observed in either of the two models, which could possibly be attributed to the effects of tonic signalling.