GPA33 forms a distinct diagnostic target class to Claudin 18.2 in oesophageal adenocarcinoma enabling the development of a novel GPA33 antibody-based detection platform.

BACKGROUND: Oesophageal adenocarcinoma (OAC) is a cancer of high unmet clinical need. Because of tumour heterogeneity, it is likely that OAC will be stratified into several subtypes. Claudin 18.2 antibodies form emerging novel therapeutics in patients with a subtype of OAC. A large-scale proteogenomics screen in OAC identified glycoprotein A33 (GPA33) protein as a dominating cancer-specific target. We set out to determine whether GPA33 is distinct from or overlaps with Claudin 18.2 as a theranostic target in OAC. METHODS: A microarray from n = 106 patients, composed of cancer, normal squamous tissue, normal gastric tissue, and metastatic lymph nodes, was used to compare the expression of GPA33 and Claudin 18.2. A single-chain variable fragment (scFv)-phage display library was screened against recombinant GPA33 protein to isolate novel monoclonal antibodies. Next-generation complementarity-determining region 3 (CDR3) DNA sequencing (NGS) and enzyme-linked immunosorbent assay (ELISA) were both used to measure efficacy of antibody enrichment during biopanning. RESULTS: GPA33 exhibits superior tumour-specific expression compared with Claudin 18.2, the latter of which is expressed in normal gastric tissue. GPA33 and Claudin 18.2 exhibit statistically significant mutually exclusive expression in cancer tissue cores; 36% of cancers are GPA33(+)/Claudin 18.2(-), whilst 22% are GPA33(-)/Claudin 18.2(+). GPA33 therefore forms a novel target for theranostics in a significant number of patients. A monoclonal antibody (RSE-05) targeting GPA33 was isolated from a scFV-phage display library. The antibody required a di-sulphide bridge to maintain its epitope on the antigen. Epitope mapping was performed using di-sulphide bridge mutagenesis, peptide-phage display, and XL-MS. The dominant epitope resides in the V-type IgG domain of GPA33 at residues 27-29 and structural amino acids S17 and K65. This di-sulphide bridge-constrained epitope defines a novel monoclonal antibody binding interface. The RSE-05 monoclonal antibody can be adapted and used as a capture-sensor tool to measure GPA33 protein in liquid phase using a two-site sandwich ELISA format. CONCLUSIONS: GPA33 exhibits elevated cancer-specific expression relative to Claudin 18.2, indicating that GPA33 can also form the basis for a cancer diagnostic. Claudin 18.2 and GPA33 generally exhibit mutually exclusive expression suggestive of two different OAC development pathways. Thus, GPA33 forms a novel target that captures the Claudin 18.2-negative patient class, and the monoclonal antibody we describe forms the basis for novel diagnostic and therapeutic tools for development in OAC.