Collaborative study for the validation of an improved HPLC assay for recombinant IFN-alfa-2

The current European Pharmacopoeia (Ph. Eur.) texts for Interferon (IFN)-alfa-2 include a nonspecific photometric protein assay using albumin as calibrator and a highly variable cell-based assay for the potency determination of the protective effects. A request was expressed by the Official Medicines Control Laboratories (OMCLs) for improved methods for the batch control of recombinant interferon alfa-2 bulk and market surveillance testing of finished products, including those formulated with Human Serum Albumin (HSA). A HPLC method was developed at the Medical Products Agency (MPA, Sweden) for the testing of IFN-alfa-2 products. An initial collaborative study run under the Biological Standardisation Programme (BSP; study code BSP039) revealed the need for minor changes to improve linearity of the calibration curves, assay reproducibility and robustness. The goal of the collaborative study, coded BSP071, was to transfer and further validate this improved HPLC method. Ten laboratories participated in the study. Four marketed IFN-alfa-2 preparations (one containing HSA) together with the Ph. Eur. Chemical Reference Substance (CRS) for IFN-alfa-2a and IFN-alfa-2b, and in-house reference standards from two manufacturers were used for the quantitative assay. The modified method was successfully transferred to all laboratories despite local variation in equipment. The resolution between the main and the oxidised forms of IFN-alfa-2 was improved compared to the results from the BSP039 study. The improved method even allowed partial resolution of an extra peak after the principal peak. Symmetry of the main IFN peak was acceptable for all samples in all laboratories. Calibration curves established with the Ph. Eur. IFN-alfa-2a and IFN-alfa-2b CRSs showed excellent linearity with intercepts close to the origin and coefficients of determination greater than 0.9995. Assay repeatability, intermediate precision and reproducibility varied with the tested sample within acceptable ranges. Test accuracy estimated by comparing the values obtained by the participants to the declared contents determined by the manufacturers was good despite the absence of a common reference preparation. In conclusion, the present study showed that the new method is suitable, reproducible and transferable. Proposals for the revision of Ph. Eur. texts are presented.