Isolation and characterisation of antibodies against human and canine EphA2_ targets for comparative studies in brain cancer

Abstract

Glioblastoma multiforme (GBM) is one of the high-grade gliomas, with a patient median survival of around 15 months from the time of diagnosis. Current treatments for this brain cancer initially comprise of surgery and radiotherapy, followed by administration of Temozolomide. Despite many advances in preclinical studies of GBM in mice, the outcomes have not been translated to the clinical level for humans. This is because the mouse model of GBM does not recapitulate the human disease and is not a suitable model for comparative studies of spontaneous tumours in humans. The canine model represents a powerful, large animal model of gliomas, since canine brain tumours occur spontaneously, with an incidence rate and patient age profile quite similar to human populations. EphA2 is one of the multi-domain receptors of the most significant tyrosine kinase family of receptors, and is highly expressed in both human and canine glioblastoma. Overexpression of EphA2 has been shown to correlate with tumour stage, progression and patient survival. Since EphA2 is not expressed in normal brain tissue, but is over-expressed on GBM cells, it is potentially a highly useful receptor for antibody-based therapy of brain cancer in both humans and dogs. Thus, monoclonal antibodies (mAbs) which cross-react with both human and canine EphA2 would be valuable molecular entities with diagnostic and therapeutic potential. To generate mAbs specific for EphA2 with cross-reactivity to both canine and human EphA2 receptors, phage display technology was employed using various strategies for the presentation of the EphA2 antigen to the phage antibody library. This included biopanning against firstly recombinant whole extracellular domain (ECD) of both human and canine EphA2, secondly recombinant ligandbinding domain (LBD) of human and canine EphA2, and lastly against cells expressing the native EphA2. Biopanning against recombinant EphA2-ECD generated one promising antibody (mAb AGH001) which showed cross-reactivity to both recombinant human and canine EphA2-ECD. However, mAb AGH001 was shown to also bind to EphA2 Knock Down (KD) cells, suggesting mAb AGH001 may bind to a domain within EphA2-ECD, which is conserved among EphA family receptors, other than the unique LBD of EphA2. To isolate specific anti-EphA2 antibodies for the LBD on EphA2-ECD, the human and canine LBD was utilised as an antigen for a biopanning campaign. Utilising this strategy of biopanning against the LBD of EpahA2, mAb HuB1 was isolated and showed binding to both human and canine EphA2-LBD and EphA2-ECD as well; however, mAb HuB1 also showed binding to the negative EphA2 cell line. This could be due to partially overlapping epitopes on other Eph receptors. mAb AGH001 and a positive control antibody, murine mAb 4B3 did not bind to EphA2-ECD subdomains. It can be concluded that the recombinant ECD sub-domains did not recapitulate the native conformation of EphA2. Furthermore, we demonstrated that mAb AGH001 has a conformational III epitope, therefore relies on a structural conform...