Isolating novel antibodies against difficult targets using next generation antibody phage display

Phage display bypasses the immune response by bringing the entire immune repertoire to a target antigen, in vitro. The coupling of antibody phage display and next generation sequencing (NGS) has the potential to revolutionise the discovery of novel antibodies. In this study, next generation antibody phage display (NGAPD) strategies were devised to identify novel antibodies against conformational epitopes on the enzyme phytase, and against the hapten phytic acid. These were notably difficult targets and previous immunisation attempts for phytic acid have not demonstrated any immune response for hybridoma or the production of immune phage display libraries. This study used naïve human scFv and camelid VHH phage display libraries and it was therefore more difficult to identify potential binders due to a lack of bias towards the targets within a high level of antibody diversity. Following phage panning, the read depth provided by NGS facilitates the identification of novel ligands from within a diverse sub-library, and thus the study also aimed to optimise the use of NGS with naïve phage display libraries. The NGS approach was used in comparison with conventional colony picking. The colony picking approach successfully identified phage-fusion binders against the three targets, however when expressed as antibody-MBP fusions the binding activities were lost. The NGS data analysis involved ranking candidates using a Z score followed by further, highly stringent sorting analyses designed to remove nonspecific ligands. Major challenges involved rescuing scFv candidates from their sub-libraries; this was later resolved with the use of the camelid VHH library that uses a single antibody framework. The NGAPD concluded with a range of NGS candidates synthesised from the final VHH experiment. The candidates were cloned in a single step into a pMAL expression system using inverse PCR. Finally, the NGS candidates were expressed as MBP fusions; however no binding activity was detected. The final section of the study considered alternative bioinformatics analysis strategies for NGS data that have been used in the literature for both antibodies and peptides and also suggested novel enrichment profiles as a means of identifying selection dependant ligands. The VHH NGS data was reassessed based solely on candidate frequency and enrichment throughout panning. A selection of antibody candidates were identified by both bioinformatics methods for phytase, suggesting that the original NGS analysis strategy was successfully identifying promising candidates that were more likely to be conformation specific