Purpose: Polysorbates (PS) are excessively used as nonionic surfactants in biopharmaceutical formulations. As a result of such ubiquitous use, PS degradation and its impact on overall product quality is receiving increased attention. A common analytical setup that permits detailed study of PS degradation consists of chromatographic separation followed by mass spectrometry (LC-MS). However, data analysis is often limited to overall chromatogram changes or selected species. Considering that PS is a highly heterogeneous product and that, as a result, single MS spectra can contain hundreds of species, this is unsatisfactory. An automated approach to comprehensively identify PS species and eliminate the laborious manual data analysis is clearly needed.
Methods: Data Set Generation
Final concentrations of 0.02% PS 80 and 1.5mM AAPH were dissolved in oxygen saturated ultra-purified water and a volume of 1ml was pipetted into glass vials. The vials were stoppered, sealed and incubated at 40°C. Samples were collected in triplicates after 0, 6, 12, 24, and 48 hours, spiked with 150ul of 200mM L-Methionine and stored at -20C until LC-MS analysis. For LC-MS analysis, 1ug of PS80 was injected onto a Prevail C18 150mm x 2.1mm HPLC column; mass spectrometric analyses were performed on a Waters Q-TOF Premier mass spectrometer collecting MS1 spectra in positive electrospray ionization mode.
Using in-house Visual Basic routines to access MassLynx 4.1 functions, centroid MS1 data were 2-point calibrated for known m/z markers. TIC chromatograms were divided into “scan extraction windows” based on peak retention times and all scans within a window were summed to produce a spectrum. In-house Matlab routines were used for mass binning, data enrichment, and plotting. Data enrichment was performed by cross-matching experimental m/z data with theoretical m/z data of PS 80 species.
Results: Degradation generally progressed faster for higher-order esters. Individual PS 80 species revealed clear differences in their degradation patterns (continuous degradation vs. build-up of degradation products followed by degradation) and degradation rates.
Conclusion: The automated data analysis approach presents a significant time reduction over manual data analysis and successfully identified the degradation patterns of the full range of PS 80 species.