Comparison of cell-based and non-cell-based assay platforms for the detection of clinically relevant anti-drug neutralizing antibodies for immunogenicity assessment of therapeutic proteins.

Anti-drug neutralizing antibodies (NAbs) formed due to unwanted immunogenicity of a therapeutic protein point towards a mature immune response. NAb detection is important in interpreting the therapeutic's efficacy and safety in vivo. In vitro cell-based NAb assays provide a physiological system for NAb detection, however are complex assays. Non-cell-based competitive ligand binding (CLB) approaches are also employed for NAb detection. Instead of cells, CLB assays use soluble receptor and conjugated reagents and are easier to perform, however have reduced physiological relevance. The aim of this study was to compare the performance of CLB assays to established cell-based assays to determine the former's ability to detect clinically relevant NAbs towards therapeutics that (i) acted as an agonist or (ii) acted as antagonists by binding to a target receptor. We performed a head-to-head comparison of the performance of cell-based and CLB NAb assays for erythropoietin (EPO) and two anti-receptor monoclonal antibodies (AMG-X and AMG 317). Clinically relevant NAb-positive samples identified previously by a cell-based assay were assessed in the corresponding CLB format(s). A panel of 12 engineered fully human anti-EPO monoclonal antibodies (MAbs) was tested in both EPO NAb assay formats. Our results showed that the CLB format was (i) capable of detecting human anti-EPO MAbs of differing neutralizing capabilities and affinities and (ii) provided similar results as the cell-based assay for detecting NAbs in patient samples. The cell-based and CLB assays also behaved comparably in detecting NAbs in clinical samples for AMG-X. In the case of anti-AMG 317 NAbs, the CLB format failed to detect NAbs in more than 50% of the tested samples. We conclude that assay sensitivity, drug tolerance and the selected assay matrix played an important role in the inability of AMG 317 CLB assays to detect clinically relevant NAbs.

A method to quantitate the neutralizing capacity of anti-therapeutic protein antibodies in serum and their correlation to clinical impact.

A robust, quantitative method for assessing the neutralizing capacity of anti-therapeutic protein antibodies was developed and tested using 4 analytical assay platforms typically used for detection of anti-drug neutralizing antibodies. The method described here utilized titration of increasing concentrations of therapeutic protein into serum containing anti-therapeutic protein antibodies, either positive control antibodies or clinical samples. Neutralizing capacities were calculated by determining the EC50 from the titration curves. The neutralizing capacity of purified anti therapeutic protein antibodies was expressed in terms of "µg of drug neutralized per µg of anti-TP antibody" present. In the case of serum originating from clinical study subjects, the neutralizing capacity of the samples was expressed as "µg of drug neutralized per mL of serum". A relative shift in EC50 values was observed as the amount of serum or antibody was changed resulting in a proportional shift of the calculated neutralizing capacity. Application of this approach using different assay platforms was consistent providing evidence for its potential to be a useful approach to characterize, qualify and compare neutralizing positive control antibody preparations or clinical samples. Using this methodology, we were able to draw a clear correlation between the neutralizing capacity and the effect of these antibodies on a clinical pharmacodynamic (PD) marker. Determination of the neutralizing capacity of antibody positive samples from subjects in a clinical study indicated direct correlation of pharmacodynamic results with non-response, partial response and response to high, mid and low neutralizing capacities respectively.