Use of response surface methods and path of steepest ascent to optimize ligand-binding assay sensitivity

Abstract

Response surface methods (RSM) combined with a steepest ascent approach is a powerful technique to optimize assay performance. In this case, a ligand-binding assay (LBA) to quantitate a peptide biotherapeutic was optimized for improved sensitivity using this technique. Conditions were elucidated to enable pg/mL quantitation of the peptide in human plasma using steepest ascent to efficiently optimize assay factors. Instead of relying solely on assay development experience and intuition to improve assay sensitivity, this systematic approach takes advantage of a predictive mathematical model generated through response surface methods that defines a specific path towards greater predicted assay sensitivity. The actual response observed along the steepest ascent path was in good agreement with the model for several steps, until reagent concentrations moved beyond the physical limits of the system, and model breakdown occurred. RSM combined with steepest ascent method proved a useful tool for sensitivity optimization in three ways: (1) The required LBA sensitivity performance (approximately 200 pg/mL), measured as a signal-to-noise ratio (SNR) at the targeted lower limit of quantitation (LLOQ), was efficiently achieved in only two optimization experiments; (2) Steepest ascent confirmed that the desired sensitivity was found within the initial RSM design space, and no further gain in sensitivity was found venturing beyond this design space along the steepest ascent path; (3) The desired assay sensitivity was maintained over a reasonable range of reagent concentrations along the steepest ascent path, indicating assay robustness for this parameter.