Comparison of bioanalytical methods for the quantitation of PEGylated human insulin

Purpose: The quality of bioanalytical data is dependent upon selective, sensitive, and reproducible analytical methods. With evolving technologies available, bioanalytical scientists must assess which is most appropriate for their molecule through proper method validation. For an early stage PEGylated insulin program, the characteristics of four platforms, ELISA, ECL, Gyrolab, and LC-MS/MS, were evaluated using fit-for-purpose method development and validation, while also evaluating costs.

Method: Methods selected for validation required acceptable performance based on satisfaction of a priori criteria prior to proceeding to subsequent stages of validation. LBA pre-validation included reagent selection, evaluation of matrix interference, and range determination. LC-MS/MS pre-validation included selection of a signature peptide; optimization of sample preparation, HPLC, and LC-MS/MS conditions; and calibration range determination. Pre-study validation tested accuracy and precision (mean bias criteria±30%; precision≤30%). Pharmacokinetic (PK) parameters were estimated for an in vivo study with WinNonlin noncompartmental analysis. Statistics were performed with JMP using ANOVA and Tukey-Kramer post hoc analysis. A cost analysis was performed for a 200-sample PK study using the methods from this study.

Results: All platforms, except Gyrolab, were taken through validation. However, a typical Gyrolab method was included for the cost analysis. Ranges for the ELISA, ECLA, and LC-MS/MS were 8.52-75, 2.09-125, and 100-1000 ng/mL, respectively, and accuracy and precision fell within a priori criteria. PK samples were analyzed in the 3 validated methods. PK profiles and parameters are similar for all methods, except LC-MS/MS, which differed at t=24h and with AUC0-24. Further investigation into this difference is warranted. The cost analysis identified the Gyrolab platform as the most expensive and ELISA as the least expensive, with method specific consumables attributing significantly to costs.

Conclusions: ECLA had a larger dynamic range and sensitivity, allowing accurate assessment of PK parameters. Although this method was more expensive than the ELISA, it was the most appropriate for the early stage PEGylated insulin program. While this case study is specific to PEGylated human insulin, it highlights the importance of evaluating and selecting the most appropriate platform for bioanalysis during drug development.