Generation and application of recombinant antibody fragments for prostate cancer detection.

Prostate cancer (PCa) is the most prevalent adenocarcinoma and the second highest cause of cancer-related deaths in men. Early diagnosis is required to identify the development of PCa to reduce the risk of the disease metastasising to different regions of the body. Multiple biomarkers in serum have been identified for the diagnosis of PCa. However, prostate-specific antigen (PSA) and its isoforms remain the gold standard test for PCa detection. Many antibodies have been generated for the detection of PSA with good sensitivities and limits of detection. To ensure reliable assay performance a number of biomarkers for PCa are required to increase specificity and sensitivity thus reducing the number of false positive results upon initial examination. The overall aims of the work presented in this thesis were to generate, select and characterise antibodies to PSA, an important biomarker in PCa. The generation and selection of murine and avian recombinant antibody fragments to free PSA (fPSA) was performed. Animals were immunised, RNA extracted and cDNA synthesised. The antibody variable genes were amplified by PCR, gel-purified and cloned into a phagemid vector. Electrocompetent cells were transformed and library sizes of 2.02 x 106 cfu/mL and 3.10 x 108 cfu/mL were obtained for the murine and avian clones, respectively. Specific clones were isolated by ‗biopanning‘ which involved a selection process on fPSA-coated immunotubes, enrichment and affinity maturation. Five and three rounds of selection were performed on the murine and avian anti-fPSA single chain variable fragments (scFv) recombinant antibody libraries, respectively. Preliminary analysis by ELISA indicated a large panel of positive binding clones for both screened libraries with approximately 60% of the analysed clones binding to fPSA. The murine recombinant antibody fragments were reformatted and expressed as single chain antibodies (scAb). This construct greatly enhanced soluble protein expression and stability while also facilitating the rapid kinetic analysis of candidate antibodies by Biacore. A differential expression profile of an anti-fPSA scFv and reformatted scAb was performed by ELISA. Individual antibodies isolated from the ‗biopanning‘ process were then characterised. The murine anti-fPSA antibodies were further analysed using the Biacore 3000 in a capture assay format. This facilitates correct affinity ranking of the selected antibodies where the antibody fragment was captured using the appropriate secondary tag antibody (anti-human constant kappa Ab) and a fixed concentration of fPSA (1 g/mL) passed over the captured antibody fragment. Ninety four antibody fragments were evaluated and ranked on their percentage complex stability values after a 5 minute dissociation period. Eleven of the ninety four were subjected to further analysis and their affinities ranged from 3.99 x 10-9 M to 3.62 x 10-10 M. They were then expressed in host Escherichia coli (E. coli) cells on a large scale and purified by immobilised metal affinity chromatography (IMAC). The cell lysates were characterised by ELISA, SDS-PAGE and Western blotting. The antibodies were sequenced and analysed for variability and similarities between the coding regions of the variable domains. Inhibition ELISA-based assays with inter and intra-day analysis were performed on the purified antibodies and a reproducible and accurate assay format was demonstrated with a limit of detection of 2.6 ng/mL. The avian clones were analysed, ranked and affinities evaluated by Dr. Paul Leonard using the Biacore A100 employing the same assay format used for analysis of the murine antibody fragments. Imperative to assay design is the selection of optimal antibody candidates for improved performance. For assay development antibody pairs which bind different regions of the fPSA molecule were determined using the capture assay format on the Biacore 3000. One murine scAb, two avian Fabs and one avian scFv were shown to bind different epitopes of the fPSA molecule with no non-specific binding evident. Following the successful purification of the murine and avian antibody fragments immunohistochemistry studies were performed on PCa tissue samples. These antibodies were shown to bind to different zones of the prostate gland (epithelial and stromal regions) confirming that the antibodies were binding to different epitopes of the fPSA molecule. The purified murine anti-fPSA scAb was biotinylated and this was employed in a sandwich ELISA assay whereby the murine anti-fPSA biotinylated scAb was captured on a neutravidin-coated ELISA plate and fPSA sandwiched using the avian antibody fragments as the detection antibodies. The assay sensitivity was improved using this format with a limit of detection of 1.8 ng/mL in PBST solution. Female serum was spiked with fPSA and inter-day studies performed by sandwich ELISA. Patient serum samples with diagnosed PCa were evaluated by sandwich ELISA to determine levels of fPSA in the sample and data were compared to documented tPSA levels which were comparable to the levels obtained from a commercially available sandwich assay