Search

CD19 Chimeric Antigen Receptor (CAR) therapy represents a breakthrough in the treatment of relapsed/refractory acute lymphoblastic leukaemia and early-phase clinical trials with CAR-modified-T cells have shown unprecedented responses. A CAR is a recombinant receptor that combines a single chain variable fragment (scFv) against a tumour-associated antigen and an intracellular activation domain of the T cell receptor (TCR), recognizing membrane-bound antigen, typically with a higher affinity compared to TCR-pMHC affinity. We compared the influences of differences in on- and off-rates underlying the binding kinetics of CD19 CARs on downstream CAR T cell responses and constructed a novel CAR derived from the CAT hybridoma demonstrating a lower affinity as a result of a greater off-rate, whilst maintaining an on-rate nearly identical to FMC63, the scFv most-used in clinical trials. Using in vitro pre-clinical models we demonstrated that CAT-CAR+ T cells showed increased proliferation, higher cytotoxic responses, and increased cytokine production, as well as a greater number of interactions between CAT-CAR+ T cells and target cells and a greater motility in comparison to FMC63. In vivo CAT-CAR+ T cells showed an enhanced ability to clear disease, proliferate and produce cytokines, displaying markers characteristic for improved T cell fitness. We believe that, analogous to the natural TCR, a lower overall binding affinity might be mitigated by a relatively faster off-rate in the setting of a constant on-rate and propose that this may enhance CAR T cell function through serial triggering, where increased target:effector interaction time, may lead to exhaustion and activation induced cell death. These data have important implications for the development of future CARs.

Myeloma remains largely incurable and there is an unmet need for new therapies. B-cell maturation antigen (BCMA) is an attractive therapeutic target due to its restricted expression on normal and malignant plasma cells(PC). A natural ligand for BCMA is a proliferating ligand (APRIL) which also naturally binds another antigen expressed on myeloma cells: Transmembrane activator and CAML interactor (TACI). APRIL is an attractive binder as it is a compact, self-protein which binds both BCMA and TACI with high affinity. BCMA and TACI expression was quantified on tumour cells from a large number of patients to explore the clinical utility of targeting these antigens. Tumour BCMA expression was found in all patients, persisted with relapse and extramedullary spread and low levels was associated with improved patient outcome. However BCMA expression was variable, often low and TACI was co-expressed on tumour from 78% of patients. Recognising that dual antigen targeting of BCMA and TACI may increase target antigens on tumour and reduce the risk of antigen negative disease escape, an APRIL based chimeric antigen receptor(ACAR) was optimised and in vitro activity against BCMA and TACI targets demonstrated. Target kill was seen at low levels of antigen and low effector to target ratios and further, potent ACAR efficacy was demonstrated in vivo in an intramedullary murine myeloma model. A targeted mutagenesis strategy was also used to generate a BCMA specific APRIL mutant. When APRIL mutant was used in a CAR(ACAR-mut), specificity was confirmed for BCMA targets, however there was no significant T cell expansion against primary tumour. In a second strategy, APRIL based bispecific molecules were engineered consisting of a CD3 specific single chain variable fragment linked to an APRILWT(APRILiTE). Preliminary studies using unpurified and unquantified protein in co-cultures of target and activated T cells, demonstrated cytolysis of BCMA expressing cells and primary myeloma cells using APRILiTE. Therefore, APRIL based therapeutic strategies provide a novel and promising approach to immunotherapy in myeloma.

T-cell lymphomas and leukemias are aggressive, treatment-resistant cancers with poor prognosis. Immunotherapeutic approaches have been limited by a lack of target antigens discriminating malignant from healthy cells. While treatment of B-cell cancers has been enhanced by targeting pan B-cell antigens, an equivalent approach is not possible for T-cell malignancies since profound T-cell depletion, unlike B-cell depletion, would be prohibitively toxic. We propose an immunotherapeutic strategy for targeting a pan T-cell antigen without causing severe depletion of normal T-cells. The α/β T-cell receptor (TCR) is a pan T-cell antigen, expressed on >90% of T-cell lymphomas and all normal T-cells. An overlooked feature of the TCR is that the β-constant region comprises 2 functionally identical genes: TRBC1 and TRBC2. Each T-cell expresses only one of these. Hence, normal T-cells will be a mixture of individual cells expressing either TRBC1 or 2, while a clonal T-cell cancer will express TRBC1 or 2 in its entirety. Despite almost identical amino acid sequences, we identified an antibody with unique TRBC1 specificity. Flow cytometry (FACS) of T-cells in normal donors (n = 27) and patients with T-cell cancers (n = 18) revealed all subjects had TRBC1 and 2 cells in both CD4 and CD8 compartments, with median TRBC1 expression of 35% (range 25-47%). In addition, we examined viral-specific T-cells in healthy volunteers, by generation of Epstein Barr virus-specific primary cytotoxic T-cell lines (3 donors) or by identification of cytomegalovirus- (3 donors) or adenovirus- (5 donors) specific T-cells by peptide stimulation. We demonstrated similar TRBC1: 2 ratios in viral-specific cells, suggesting that depletion of either subset would not remove viral immunity. Next, using FACS and immunohistochemistry, we showed that TCR+ cell lines (n = 8) and primary T-cell lymphomas and leukemias (n = 55) across a wide range of histological subtypes were entirely restricted to one compartment (34% TRBC1). As proof of concept for TRBC-selective therapy, we developed anti-TRBC1 chimeric antigen receptor (CAR) T-cells. After retroviral transduction of healthy donor T-cells, comprising mixed TRBC1/2 populations, 90% of T-cells expressed CAR on the cell surface. No detectable TRBC1 T-cells remained in the culture, suggesting selective depletion of this population. Anti-TRBC1 CAR T-cells secreted interferon-γ in response to TRBC1-expressing target cell lines (p < 0.001) or autologous normal TRBC1+ cells (p < 0.001), and not TRBC2 cell lines or autologous normal TRBC2 cells. Anti-TRBC1 CAR killed multiple TRBC1 cell lines (p < 0.001) and autologous normal TRBC1cells (p < 0.001), and not TRBC2 cell lines or autologous normal TRBC2 cells. These cell-line based findings were confirmed using primary cells from two patients with TRBC1+ adult T-cell leukaemia/lymphoma. We demonstrated specific tumour kill by allogeneic or autologous T-cells in vitro, despite partial downregulation of surface TCR by tumour cells. We developed a xenograft murine model of disseminated T-cell leukemia by engrafting engineered firefly luciferase+ TRBC1+ Jurkat cells in NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG) mice. Bioluminescent imaging and FACS of marrow at 5 days following IV T-cell injection showed that mice treated with anti-TRBC1 CAR T-cells and not non-transduced (NT) T-cells had disease clearance (p < 0.0001). In a further model, mice were engrafted with equal proportions of TRBC1-Jurkat and TRBC2-Jurkat cells. FACS analysis of bone marrow at 5 days post T-cells demonstrated specific eradication of TRBC1 and not TRBC2 tumour by anti-TRBC1 CAR (p < 0.001). In summary, we have demonstrated a novel approach to investigation and targeting of T-cell malignancies by distinguishing between two possible TCR β-chain constant regions. Using CART-cells targeting TRBC1 we have demonstrated proof of concept. Unlike non-selective approaches targeting the entire T-cell population, TRBC targeting could eradicate a T-cell tumour while preserving sufficient normal T-cells to maintain cellular immunity.

Adoptive Cell Therapy (ACT) is a novel approach to cancer treatment, which implements the use of the patient T-cells redirected to eradicate tumour cells. The T-cells are engineered to express a Chimeric Antigen Receptor (CAR) that bestows specificity for an arbitrary antigen. CAR-T cell clinical trials have shown tumour eradication of leukaemia, but modest results against solid tumours. There is an ongoing debate regarding the correlation of CAR affinity and antigen-density recognition threshold, which we sought to investigate. Although the TCRs sensitivity against low antigen targets is superior at low-affinity, the effect of affinity in CAR efficacy and sensitivity is controversial. In order to examine the correlation between CAR efficacy and affinity, we designed a CAR against an intracellular protein in melanoma. Tyrosinase-Related Protein 1 (Tyrp1) resides in the surface of the organelles called melanosomes. However, Tyrp1 is also trafficked to the surface at a low density. We mutated an anti-Tyrp1 single-chain Variable Fragment (scFv) to acquire eight mutants of various affinities. The affinity range was 0.74nM-54.3nM. Including the wild-type scFv, these nine affinity gradient CARs were challenged with a cell line expressing low, medium and high densities of cell-surface Tyrp1. Three mutant scFv CARs were superior to the wild-type scFV. However, the affinities of those superior CARs was variant. There was no pattern observed to suggest a direct correlation between the affinity and the CAR efficacy. In contrast to previous publications, this study shows that affinity does not play a key role in determining the CAR efficacy. We hypothesise that the difference in CAR efficacy depends on a combination of factors, such as scFv stability and affinity, CAR density, and antigen density. However, the complex interaction of these parameters, as well as further confounding factors, renders the deduction of a pattern challenging. In order overcome this complexity, a multivariate analysis of all the parameters together is necessary.

Adoptive transfer of antigen specific regulatory T (Treg) cells was shown to be beneficial for the suppression of autoimmunity. In this study, we engineered antigen-specific Treg-like cells that recognise mouse MHC-class II molecules. The restricted expression of MHCII to professional antigen presenting cells and to inflammatory sites, allow the employment of MHCII specific Treg for suppression of unwanted immune inflammation. A mouse MHCII-specific Chimeric Antigen Receptor (CAR) was used to redirect the specificity of T cells via retroviral transduction, while Foxp3 gene transfer into purified CD4+ cells leads to the acquisition of a Treg-like phenotype. Incorporation of a suicide mechanism within the engineered Treg-like cells for their in vivo selective deletion will prevent long-term immune suppression. The functionality of the generated CAR and the specificity of responses elicited by CAR bearing T cells were validated in vitro. In preliminary in vivo experiments, intravenous injection of C57BL/6 mice with syngeneic mouse MHCII-specific CD4+ T cells, led to GVHD-like toxicity, while no signs of toxicity were observed when Treg-like cells of the same specificity were transferred alone or in a 1:1 mix with mouse MHCII specific CD4+ T cells. These data suggest the suppressive potential of the engineered Treg-like cells.

Blockade of Cytotoxic T lymphocyte antigen-4 (CTLA-4) can enhance anti-tumour responses in preclinical models. A monoclonal antibody targeting human CTLA-4 (αCTLA4 mAb), is FDA approved for clinical use against metastatic melanoma. Paradigms suggest that αCTLA4 mAb blocks CTLA-4 dually on CTLA-4hi Tregs (restricting regulatory activity) and CTLA-4int-hi CD4 Teff (enhanced expansion) resulting in intratumoural Teff dominance and a net gain in anti-tumour activity. More recently, antibody-dependent cellular cytotoxicity (ADCC)-mediated depletion of CTLA-4hi Tregs in the tumour has been flagged as a critical factor driving anti-tumour immune responses. We explored the ADCC hypothesis in an adoptive cell transfer model of murine melanoma. We compared intratumoural Treg numbers in groups receiving CTLA-4+/+ or CTLA-4-/- tumour-specific T cells with/without exogenous αCTLA4 mAb. Exogenous αCTLA4 mAb resulted in diminished intratumoural tumour-specific Treg numbers in groups receiving CTLA-4+/+ but not CTLA-4-/- T-cells. Endogenous CTLA-4+/+ Tregs were depleted intratumourally with αCTLA4 mAb, irrespective of the CTLA-4 status of the adoptive cells. This data supports the proposed ADCC mechanism of αCTLA4 mAb and its dependence upon target cell CTLA-4 expression. Furthermore, we engineered tumour-specific T-cells to secrete αCTLA-4 mAb on IgG2a isotype (high ADCC activity) or IgG1 isotype (low ADCC activity) for adoptive transfer into melanoma-bearing mice. We found that transduced primary murine T-cells could engraft, infiltrate tumours and secrete biologically active αCTLA4 mAb. Secreted IgG2a αCTLA-4 mAb was strongly associated with intratumoural depletion of CTLA-4hi Tregs, but unexpectedly CTLA-4-replete transduced Teff were also vulnerable to depletion, and this significantly compromised the anti-tumour potential of this therapy. Intratumoural depletion of CTLA-4hi T-cells was not associated with (secreted) IgG1 αCTLA-4 mAb or (secreted) isotype control mAb. We believe that depletion of intratumoural Tregs is an important area for cancer immunotherapy, but our experience of αCTLA-4 mAb underlines the importance of careful immunomodulatory target selection to avoid toxicitiy.

Live, attenuated measles virus (MV) has demonstrated tumour-specific replication and anti-tumour activity in murine models of hematological and non-hematological malignancy, leading to a number of published or on-going clinical trials. However, the potentially overriding issue of how to achieve successful administration of oncolytic measles virotherapy in patients with intact humoral immunity remains. I have examined mechanisms for optimizing delivery of oncolytic MV to acute lymphoblastic leukaemia (ALL) – a disseminated haematological malignancy. Anti MV antibody titres in adult patients with ALL were evaluated before and after leukaemia treatment regimens including high dose corticosteroids, cyclophosphamide and anti-B cell monoclonal antibodies. Antibody titres were unaffected by induction and intensification treatment protocols, suggesting the need for consideration of optimal virotherapy delivery strategies to achieve therapeutic success. Here, I demonstrate that human bone marrow-derived mesenchymal stromal cells (BM-MSCs) can be used effectively as virus delivery vehicles, permitting ex-vivo cellular virus loading and intracellular virus amplification, delivery of virus to distant sites of disease following systemic infusion, and virus hand-off to precursor B lineage ALL cell targets in the presence of pre-existing anti-MV antibodies. In vivo modelling using SCID mice bearing disseminated pre-B ALL xenografts demonstrated enhanced survival of passively immunized animals following IV treatment with BM-MSC-delivered MV versus naked MV or BM-MSCs alone. In order for vaccine MV to be safely and rationally utilized as a novel therapeutic for ALL, a detailed mechanistic understanding of how the virus exerts its oncolytic effect is paramount. In this thesis, I have utilized a previously characterized model of cellular transformation generated using human BM-MSCs to characterize the phenomenon of relative tumour cell specificity by oncolytic MV, in terms of infectivity, productivity and cell killing. Furthermore, this thesis begins to explore some of the potential mechanisms that confer vaccine MV its tumour-tropic and anti-cancer properties.