Adoptive T cell therapy to treat cancer in combination with re-directing specificity through T cell receptor (TCR) gene transfer, represents an effective therapeutic option. However, reduced effector responses due to the immunosuppressive tumour microenvironment and insufficient long-term engraftment of transferred cells represent two potential limitations. Tumours often employ mechanisms to inhibit T cell responses including secretion of TGFβ and depleting the tumour microenvironment of amino acids. The main aim of this PhD project was to develop a strategy to enhance T cell function for tumour therapy. The mammalian target of rapamycin (mTOR) pathway regulates CD8 T cell differentiation such that high mTOR activation leads to enhanced effector whilst low mTOR activation leads to increased T cell memory formation. Two retrovirus constructs have been designed whereby one expresses the positive mTOR regulator Rheb and the other expresses the negative mTOR regulator Pras40. Rheb transduction into CD8 T cells resulted in enhanced activation of mTOR, increased effector functions and partial resistance to TGFβ and low arginine concentrations. Pras40 overexpression led to a decrease in the activation of mTOR and reduced effector functions. Rheb transduced CD8 T cells expanded efficiently upon antigen encounter in vivo, followed by pronounced T cell contraction. Pras40 transduced T cells were unable to expand in vivo, but persisted at low numbers and acquired a central memory phenotype. Tumour bearing mice treated with TCR re-directed CD8 T cells transduced with Rheb showed improved tumour protection. Pras40 overexpression resulted in the loss of the protective function of TCR re-directed T cells. Together, the data show that gene transfer can be used to regulate mTOR activity in T cells. Enhancing mTOR activity led to improved tumour control despite reducing memory formation. Permanent mTOR inhibition, on the other hand, preserved some memory characteristcs of T cells but deteriorated their tumour protective functions.

The events during papillomavirus lesion-formation are not well understood, but are likely to differ between high and low risk HPV types, which have different effects on the infected basal layer. These differences most likely reflect differences in protein function and gene expression patterns that have evolved to support the different biologies of the two virus groups. While high-risk types such as HPV16 or 18 can drive cell proliferation in the basal and suprabasal layers, low-risk types such as HPV 6 and 11 appear not to require this function. In order to compare the two virus groups we have introduced the different HPV genomes into a genetically identical keratinocyte background and examined their effect on functions required for lesion formation following epithelial trauma. The non-immortal keratinocyte cells have normal differentiation properties, are near-diploid and are sensitive to contact inhibition. They are currently used in the clinic to prepare skin substitutes for burns victims. In this model, high-risk HPV types increase cell growth (but not migration) rate when cells have space to grow, such as would occur during wound healing. They can also overcome normal cell-cell contact inhibition as the cells pack-up, and this is manifest in organotypic rafts as an increase in basal and parabasal cell division similar to that seen in neoplasia. This growth advantage would allow a single infected cell to outgrow its uninfected neighbours following infection or during lesion expansion after wounding. Interestingly, the low-risk HPV types appear to have negative effect on growth rate, allowing the more rapidly dividing uninfected cells to predominate. This suggests a fundamental difference in the biology of the two HPV groups, and supports the idea that low-risk types may reside in long-lived slow-cycling cell such as stem cell. For the high-risk types this model may not hold true, with lesion-formation being directed actively through functional changes in the infected basal layer.

Activation of the NF-κB pathway is linked to cancer development and progression. Kaposi’s sarcoma-associated herpesvirus (KSHV/HHV8) encodes vFLIP which binds to the NEMO/IKKγ subunit of IKK and constitutively activates NF-κB, leading to tumourigenesis. Cellular FLIPs, which share sequence homology with KSHV vFLIP and induce NF-κB activation, are upregulated in a variety of malignancies and are therefore promising targets for anti-cancer therapies. The cFLIP family consists of three splice variant isoforms (cFLIPL, cFLIPS and cFLIPR) and two proteolytic fragments (p43-FLIP and p22-FLIP). Much is known about how cFLIPs regulate apoptosis but the mechanisms by which they activate NF-κB are not well understood. Potential similarities to vFLIP-induced activation have been suggested but not investigated. Here we show that, unlike KSHV vFLIP, cFLIP variants are not found in stable complexes with NEMO and all require upstream events to mediate signalling to IKK. By mutational analysis on NEMO and protein expression knockdowns, we demonstrate that all cFLIP isoforms require the ubiquitin binding domain (UBD) of NEMO while it is redundant for vFLIP’s function. Similarly, our data reveals that TAK1 is essential for induction of IKK by cFLIP isoforms but not vFLIP. We further show that different cFLIP isoforms have different requirements for IKK activation. While cFLIPL needs LUBAC to activate NF-κB, cFLIPS and p22-FLIP require FADD and RIP1. Contrary to existing reports, our results suggest that processing of cFLIPL to p22-FLIP or p43-FLIP fragments by caspase-8 is not necessary for its IKK activation. Finally, we propose that vFLIP-mediated activation of IKK is most likely to occur through induction of multimerisation and re-orientation of the IKK complexes within higher order IKK assemblies that lead to autophosphorylation of the enzymatic subunits, IKKα and β. In conclusion, the work in this thesis provides evidence that vFLIP, cFLIPL, cFLIPS and p22-FLIP have specific and different mechanisms of inducing IKK activation. This has implications for the design of therapeutics to block pathological NF-κB activation in viral and non-viral tumours.

TCR gene therapy allows redirection of the antigen specificity of T cells by the introduction of novel TCR α and β chains by retroviral transduction. These TCR gene modified T cells can be adoptively transferred to target defined tumour antigens. The majority of TCR gene therapy studies has focused on the adoptive transfer of CD8+ T cells but there is increasing recognition of a central role for CD4+ T cells in effective immunotherapy protocols. The use of CD4+ T cells has been limited by the lack of well defined class II restricted TCR and also because the majority of tumours don’t express class II MHC. As a result research has focused on introducing class I restricted TCR into CD4+ T cells. Initial work has demonstrated that class I restricted CD4+ T cells often have reduced functional avidity compared to the parental CD8+ T cell. In particular, CD4+ T cells transduced with CD8 dependent TCRs are often of much lower functional avidity when introduced in the absence of a CD8 co-receptor. In order to improve the functional avidity of class I restricted CD4+ T cells, murine CD4+ T cells were co-transduced with F5 TCR (specific for influenza peptide, NP, in the context of H2-Kb) and additional CD3 molecules. The amount of CD3 within in a cell is rate limiting for the expression of introduced TCR and thus when cells are transduced with additional CD3 it removes this rate limiting step and thus enhances the surface expression of the TCR. TCR surface expression is one of the key determinants of T cell functional avidity. CD4+ T cells co-transduced with F5- TCR and CD3 had increased surface expression of F5-TCR and increased pentamer binding. This translated in vitro into increased functional avidity compared to CD4+ T cells transduced with F5-TCR only. When adoptively transferred in vivo into irradiated tumour bearing syngeneic recipients, F5- TCR + CD3 CD4+ T cells had greater expansion and persistence and trafficked to the tumour site at higher and faster rates than F5-TCR only CD4+ T cells. In addition, F5-CD3 CD4+ T cells demonstrated superior control of tumour growth. Unexpectedly mice that received adoptive transfer of F5-TCR + CD3 CD4+ T cells developed marked lethal toxicity. Further experiments to try to determine the nature of this toxicity suggest a multifactorial cause including mispairing of the introduced TCR α and β chains with the endogenous TCR and development of autoreactive T cells in the presence of additional CD3 mediated either by upregulation of the introduced TCR or the endogenous TCR.

Background: Autosomal dominant Hyper IgE Syndrome (AD-HIES) is a rare primary immunodeficiency characterized by high serum IgE levels, eosinophilia, and skin and lung infections. Additional features of AD-HIES include characteristic facial appearance, scoliosis, retained primary teeth, and joint hyperextensibility. Recently, AD-HIES has been associated with heterozygous dominant negative mutations in the signal transducer and activator of transcription 3 (STAT3) which plays a key role in the signal transduction of a broad range of cytokines, and is crucial for IL-6-mediated regulation of Th17 cells. Objective: We aimed to characterize patients with the clinical diagnosis of ADHIES, to identify STAT3 mutations, and to assess the frequency and functional consequences of these mutations. Furthermore, we studied STAT3-dependent signalling pathways in patients with an AD-HIES phenotype but no STAT3 mutation. Methods: We sequenced STAT3 in 153 patients with a strong clinical suspicion of AD-HIES and further components of the IL-6 signalling pathway in patients found to be STAT3 wild type. The impact of the mutations on immune cell function was assessed by measurement of cytokine release by immune cells, T cell phenotyping and STAT1 phosphorylation assays. Results: About 60% of the AD-HIES patients revealed mutations in STAT3. All mutations found were heterozygous, clustered mainly in the DNA-binding or the SH2 domain and exerted dominant-negative effects. Functional analysis of mutations affecting different domains of STAT3 revealed that some mutations might have a less severe impact on functionality of STAT3, About 40% of our cohort of patients presenting with AD-HIES phenotype harboured wild type STAT3 and may carry mutations in other genes of either the same or closely related signalling pathways. Nevertheless, we ruled out mutations affecting the IL-6 pathway iIn five Sardinian patients with wild type STAT3. Impact of findings: The results lead to a better characterization of heterozygous STAT3 mutations and of the pathogenesis of AD-HIES.

Thymocyte-expressed molecule involved in selection 2 (Themis2) is the second member of the Themis family. Recently, the first member of the Themis family, Themis, has been reported to be part of the TCR signalling cascade and its deletion severely affects thymocyte progression from the double positive to the single positive stage. All family members share similar domains and high sequence similarity and show tissue specific expression with Themis2 being expressed in B lymphocytes, macrophages and dendritic cells. THEMIS2 associates with BCR signalling molecules such as GRB2, VAV or LYN and is phosphorylated in response to BCR stimulation. For these reasons I hypothesised that Themis2 might have an important role in B cell development or activation. I show that Themis2 is expressed throughout the B cell lineage and exclude redundant expression of other Themis family members. After B cell activation Themis2 expression is downregulated. Analysis of a newly created Themis2-deficient mouse strain showed that B cell development proceeds normally in the absence of THEMIS2. Experiments on in vitro cultured Themis2-deficient primary B cells demonstrated that proliferation and survival, BCR internalisation and antigen presentation as well as expression of activation markers and cytokines were unaffected. RNA sequencing revealed only minor changes in transcription in follicular B cells, even after activation. Similarly, antibody levels to in vivo immunisation with T-dependent or T-independent antigens or challenge with influenza virus did not suggest that Themis2 is required for antibody responses either. Reactions to a model of acute allergic airway inflammation showed only marginally reduced cell numbers in the bronchoalveolar lavage fluid yet all other markers of inflammation were all normal. In conclusion, I found that Themis2 is not required for B cell development, activation or antibody responses. Further studies will be required to define the role of Themis2 in the immune system.

Polycomb group proteins maintain cell identity by repressing developmental regulator genes specific for other cell types. There are two main complexes: Polycomb repressive complex 1 (PRC1) and 2 (PRC2). PRC2 methylates histone H3 lysine 27 (H3K27me3), creating a binding site for PRC1 that ubiquitinates H2AK119. Polycomb target genes are associated with stalled RNA polymerase II (RNAPII), and the initiation marker H3K4me3, known as bivalent chromatin. Our laboratory has demonstrated that short RNAs are transcribed from the promoter region of these genes in human T-cells, while the work carried out as part of the present thesis demonstrates that short RNAs are also transcribed in murine embryonic stem cells (ESCs). This indicates that they are conserved across different species and cell types. Northern blotting for RNAs ≤200 nucleotides extracted from murine ES cell deficient for PRC2 and PRC1 revealed that short RNA production is independent of Polycomb activity. When cells differentiate and Polycomb-target genes become activated, short RNAs are depleted. Given that PRC2 interacts with RNA, this loss of short RNAs might allow gene activation. Additionally, polycomb response elements (PRE) have been detected in Drosophila. These elements are necessary and sufficient for polycomb recruitment. A recently identified PRE, HOXD11.12, recruits PRC2 in human mesenchymal stem cells (MSC). It is hypothesized that PRE activity is due to the transcription of short RNAs. Blotting for RNA extracted from MSC identified short RNAs transcribed from D11.12. Moreover, these short RNAs can form the same secondary structure as the previously-identified short RNAs and are also located at a CpG island. Furthermore, RASL12 and YBX2 behave as PREs while D11.12 from active HOXD11 enhances gene expression, potentially also acting as a Trithorax response element (TRE).

Adoptive immunotherapy with autologous T lymphocytes transduced with anti-tumour antigen-specific T cell receptors (TCRs) has emerged as a promising anti-cancer therapy. In vitro studies have suggested that T cell function may be limited by tumour associated macrophages (TAMs) and myeloid-derived suppressor cells (MDSCs). This study aims to determine the influence of TAMs and MDSCs on the antigen-specific function of TCR-transduced T cells in vitro and in vivo. Two murine tumour models, the EL4 lymphoma and ID8 ovarian carcinoma were modified to ensure that CD8+ T cells transduced with the F5-TCR recognised tumour cells expressing a model target antigen, Influenza A Nucleoprotein (NP) in the presence of tumour-infiltrating TAMs and MDSCs. It was demonstrated that Gr-1+ MDSCs isolated from tumour-bearing mice failed to suppress F5-TCR CD8+ T cells in vitro. Moreover, in vivo depletion of Gr-1+ MDSCs following the administration of an anti-Gr-1 antibody did not affect the anti-tumour function of adoptively transferred F5-TCR CD8+ T cells. Interestingly, bone marrow-derived CD115+ monocytes potently suppressed T cell proliferation in vitro, although these cells failed to impair the anti-tumour function of F5-TCR CD8+ T cells in vivo when adoptively transferred to EL4-NP tumour-bearing mice. This suggested that Gr-1+ MDSCs and/or CD115+ monocytes may not significantly impair the anti-tumour efficacy of CD8+ T cells expressing high avidity TCR in the tumour models examined. Further experiments are warranted to explore their impact on tumour antigen-specific T cells at lower frequencies and/or with lower avidity.

Melanoma progression occurs despite evidence of melanoma-specific T cell activation. Chronic or repeated antigen stimulation can cause dysregulated T cell differentiation through upregulation of inhibitory receptors (immune exhaustion) or end-stage differentiation (immune senescence). This thesis therefore investigated the hypothesis that blood and skin derived T cells of melanoma patients are driven towards immune exhaustion and senescence. An increase in senescent CD8+ TEMRA cells was detected in the blood of old melanoma patients. These cells had high cytotoxic but low proliferative potential. Whilst it could not be determined whether they were melanoma specific, the TEMRA expansions occurred independently from persistent viral infections such as CMV, and their function could be boosted through p38 signalling blockade. Skin resident T cells of melanoma patients showed no increase in T cell differentiation but instead upregulation of exhaustion markers PD-1 and CTLA-4. Granzyme B and perforin, essential for granule mediated cell killing, remained low in these cells, suggesting insufficient cytotoxic function. Skin derived T cells from healthy individuals also expressed high levels of PD-1 and low levels of cytotoxic granule components. Exposure to IL-2, IL-15 and CD3/CD28 boosted perforin and granzyme expression in healthy skin cells. Conversely, PD-1 signalling blockade during CD3 stimulation increased granzyme B expression. In summary, melanoma associated immune dysfunctions were of a different nature in blood and skin T cells. Immunotherapies designed to boost immune function in patients might therefore have different efficacies in both organs.

The actomyosin motor that drives both motility and host cell invasion in the human malaria parasite Plasmodium falciparum is thought to be coupled through aldolase to the cytoplasmic domain (CTD) of type 1 membrane proteins of the thrombospondin-related anonymous protein (TRAP) family. Other type 1 membrane proteins including apical membrane antigen (AMA1) and members of the erythrocyte binding antigen (EBL) and reticulocyte binding homologue (Rh) protein families have been implicated in host cell binding and moving junction formation. We show that P. falciparum aldolase binds to P. falciparum actin. Using a kinetic assay, we show that TRAP family members also bind to aldolase. A direct binding method confirmed that merozoite TRAP (MTRAP) and TRAP bind aldolase and indicated that the interaction is mediated by more than just the C-terminal six amino acid residues identified previously. Single amino acid substitutions in the CTD abolished binding to aldolase. MTRAP CTD was phosphorylated by calcium dependent protein kinase 1 (CDPK1) and protein kinase A, kinases that are known to phosphorylate other motor proteins, and this modification increased the affinity of binding ten-fold. Similarly AMA1, EBA and Rh protein families also bound to aldolase, with the affinity also increased by CTD phosphorylation. Therefore other proteins involved in host cell recognition and invasion, in addition to members of the TRAP family, may be connected to the motor through aldolase. If this is the case their affinity for aldolase may also be important in addition to their host ligand specificity in determining the use of, and efficiency of alternate invasion pathways. These interactions also contributed to stimulate actin polymerization and enhance aldolase enzymatic activity, increasing the enzyme affinity for it’s substrate, and potentially to the compartmentalization of the glycolytic pathway and consequent ATP availability for motor complex function.