In metozoans, most of the key events that lead to cell activation and inhibition are controlled by tyrosine phosphorylation. Extracellular signals are transmitted by membrane bound receptors, which have intrinsic kinase activity or themselves recruit intracellular kinases to specialised inhibitory or activating phosphorylation motifs. In this way, the pattern of kinase activation creates its own turnover and can rapidly generate amplified signals by positive feedback, or recruit inhibitory proteins to counteract the signals. This process of inhibition is also constitutive since it requires continuous counter-inhibition by phosphatases at the cell surface and intracellularly even in the absence of ligands. The absence of phosphatase activity results in unbridled protein phosphorylation and form this and other data it has been proposed that the triggering of the T cell receptor and other co-receptors may result simply by physical exclusion of the large phosphatases such as CD45 from the vicinity of the receptors. Superagonist monoclonal antibodies may work in a similar way, by binding receptors close to the plasma membrane and excluding extracellular phosphatases. The work described in this thesis seeks to discover if antibody superagonists can be generated against the T cell inhibitory cell surface receptor PD-1 and test if this approach can attenuate the immune response. Using in vitro assays of lymphocyte activation and a mouse model expressing human PD-1, this study characterises a series of anti-PD-1 antibodies and shows how patterns of inhibitory activity varying according to binding sites. The inhibitory effects of the anti-PD1 antibodies are seen in the humoral, cellular and transplant immune responses. Agonistic anti-PD1 antibodies induce regulatory T cells and may have role in suppression of autoimmune disease. The thesis suggests that superagonism may be harnessed clinically to dampen the immune response, through activation of inhibitory receptors.

The normal physiological activation of naive T cells requires the engagement of both the T cell receptor (TCR) and the co-stimulatory molecule, CD28. However, a group of monoclonal antibodies (mAbs) have been identified that are able to activate T cells in vitro and in vivo via CD28 engagement alone. Two defining characteristics found in all CD28 superagonist mAbs are their membrane proximal CD28 epitopes and the requirement for mAb immobilisation. To investigate whether agonistic mAbs to similar cell molecules could be identified based on epitope position alone, mAbs to the inhibitory receptor PD-1 were generated and characterised. Using a drastic mutation-based epitope mapping technique, one mAb was identified with a membrane proximal epitope along with two other mAbs with membrane distal epitopes. These mAbs were tested for triggering activity in a hybridoma stimulation assay. mAb stimulation was observed with all three mAbs but only in cells expressing a PD-1 chimera that associated with the TCR and the strength of activation was dependent on epitope location. Cross-linking of a monomeric PD-1/CD28 chimera with a pair of anti-PD-1 mAbs resulted in signalling in this system, however, suggesting a role for ligand aggregation in addition to epitope position in mAb signalling. To further investigate the role of epitope position in CD28 superagonism, a cell line expressing a chimeric form of CD28 was created wherein the superagonistic mAb epitope was moved to a membrane distal position. When stimulated with a CD28 superagonist mAb signalling was no longer observed. However stimulation with another mAb that had an epitope to a membrane proximal location on the chimera resulted in superagonistic effect. These results show that epitope location is the dominant cause of T cell stimulation observed by CD28 superagonist mAbs and that epitope dependent mAb signalling is possible in other T cell surface molecules. The work described in this thesis has implications for both the development of immune modulating mAb therapeutics and for the general mechanism of triggering of cell surface receptors dependent on extrinsic tyrosine kinases.

Numerous previous studies of GABAAR ligands have suggested that GABAAR agonists must be zwitterionic and feature an intercharge separation similar to that of GABA (approx. 4.7-6.0 Ã ). We have demonstrated that monomeric, homodimeric and heterodimeric non-zwitterionic GABA amides are partial, full, or superagonists at the murine GABAA receptor (GABAAR). The agonism of these GABA amides is comparable to that of THIP, as shown by in vitro assay results. The assay data indicate that the agonism of GABA amides is tether length-dependent. Optimum agonism is achieved with a tether length of four methylenes in GABA amide dimers and in GABA amides bearing pendant amide or amino groups. We have further investigated the structure-activity relationship for GABA amides on the GABAAR by performing structural modifications to both the superagonist 2c and the agonist 6c. Synergism and [3H]muscimol binding experiments show that 2c binds to the same sites as GABA. Structural modification of 2c demonstrated that partial rigidification of the tether eliminated agonism and caused ligands to behave as weak competitive antagonists. We have also investigated the agonism of four ZAPA derivatives in 36Cl- uptake functional assay. Two of them are found to be as potent as GABA. In our studies of 1,4-benzodiazepines, our goal was to synthesize three different subtypes of quaternary 1,4-benzodiazepines by use of the memory of chirality (MOC) strategy. Disappointingly, most of the deprotonation/alkylations failed, due to various reasons. The failure of the reactions of (S)-alanine-derived tetrahydro-1,4-benzodiazepin-3-ones was probably due to either the unexpected side reactions or the steric hindrance of enolate alkylation. In the case of tetrahydro-1,4-benzodiazepin-2-ones, computational studies suggested that steric hindrance by both the benzo ring and N4-allyl group might retard deprotonation at C3 by bulky bases like KHMDS or LDA. Finally, (S)-serine-derived 1,4-benzodiazepin-2-ones and their elimination products (ï ¡-methylene benzodiazepines) were prepared. These proved unreactive towards deprotonation/alkylations and conjugate additions, respectively. The low reactivity of the ï ¡-methylene benzodiazepines towards nucleophiles was attributed to highly delocalized LUMOs that failed to direct nucleophiles to the ï ¢-carbons. / Ph. D.