Anti-drug Antibody Assay Conditions Significantly Impact Assay Screen and Confirmatory Cut-Points

Assays for the detection and confirmation of anti-drug antibodies (ADA) are commonly used tools for assessing the immunogenicity of drug candidates in both clinical and nonclinical studies. During the development of such assays, it is typical to optimize the assay conditions based on factors such as sensitivity or signal/noise ratio (S/N) and is commonly done using an assay positive control (PC). However, even carefully optimized methods often suffer with problems due to low cut-point factors and failure to distinguish assay “noise” from a true biological response. In this paper, we describe an approach to assay development in which the impacts of assay conditions on the response and variability, both analytical and biological, of drug-naive samples are tested by way of PC-independent assay condition optimization. Using two ADA methods as model systems, we examine the impact of minimum required dilution, assay reagent (labeled drug) concentrations, incubation time, assay, and wash buffer composition. We find that the choice of assay conditions, particularly the labeled drug concentration, can greatly affect the distribution of naive sample responses and thus impact screening and confirmatory assay cut-points. In two case studies presented, screening assay cut-point (SCP) varied from 1.38 to 2.20 and 1.04 to 1.20 while the confirmatory assay cut-point (CCP) varied from 58.5 to 95.6% and 26.2 to 16.2% depending on the conditions tested. Some of the conditions produced unacceptably high CCP values. It is proposed that the degree of the observed impact of the assay conditions on SCP and CCP values depends on the compound nature and assay matrix composition and is likely connected with the diversity of interactions between drug protein and matrix components. Because it was also observed that higher assay SCP can associate with a loss of the PC-based assay sensitivity, additional assessment of the assay conditions would be required to determine an overall assay performance acceptability, including assay PC-based sensitivity, drug, and target tolerance characteristics. In conclusion, it is suggested that by assessing performance of treatment-naive samples at various assay conditions, one can identify potential assay protocols that allow to avoid undesirably low screening (e.g., < 1.2) and confirmatory (e.g., < 25%) cut-points.