This thesis examines the use of protein engineering to create antibody-based molecules for cancer treatment. The targeting unit used for these molecules was the single chain Fv antibody fragment MFE-23, which is directed against the tumour-associated marker carcinoembryonic antigen (CEA). It was hypothesised that implementation of molecular design features such as humanisation, high affinity, multivalency and mannose glycosylation to accelerate systemic clearance would result in the favourable in vivo performance of the molecules. Bioinformatic and recombinant approaches were applied to test this hypothesis. First the molecular interaction between the MFE-23 and its cognate antigen, CEA, was investigated. A recombinant form of the N-terminal N-A1 domain pair of CEA was expressed in bacteria and its specific interaction with MFE-23 was confirmed. Mass spectrometry in combination with proteolysis of the MFE-23/ CEA (N-Al) complex was used to obtain detailed sequence information on the putative region where MFE-23 binds to CEA, supporting structural predictions and the future design of MFE-23-based molecules. A series of divalent MFE-based molecules were then created. These were genetically linked by polypeptide chains or human serum albumin (HSA). A humanised, high affinity version of MFE-23 (M10b) was also investigated as part of the molecular design. In vivo studies demonstrated that the M10b-HSA assemblies, termed HSAbodies, specifically localised to tumour cells whilst clearing rapidly from normal tissues, resulting in improved tumour: tissue ratios compared to an anti-CEA IgG and higher overall uptake in tumour compared to monovalent MFE-23. Glycosylation sites were engineered into the HSA linker, which resulted in accelerated systemic clearance, further improving tumour: tissue ratios. The combination of functional affinity and controllable clearance resulted in a superior CEA-targeting molecule. These results support the proposed hypothesis and suggest a clear therapeutic and diagnostic potential for HSAbodies in both native and glycosylated form.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.