Detection of Protein Aggregates by Sedimentation Velocity Analytical Ultracentrifugation (SV-AUC): Sources of Variability and Their Relative Importance KELLY K. ARTHUR, 1 JOHN P. GABRIELSON, 1 BRENT S. KENDRICK, 1 MICHAEL R. STONER 2 1 Analytical Sciences, Amgen Inc., 4000 Nelson Rd., Longmont, Colorado 80503 2 Formulations, Genencor, 925 Page Mill Rd., Palo Alto, California 94304 Received 7 July 2008; revised 15 October 2008; accepted 7 November 2008 Published online 7 January 2009 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.21654 ABSTRACT: Sedimentation velocity analytical ultracentrifugation (SV-AUC) has found application in the biopharmaceutical industry as a method of detecting and quantifying protein aggregates. While the technique offers several advantages (i.e., matrix-free separation and minimal sample handling), its results exhibit a high degree of variability relative to orthogonal size-sensitive separation techniques such as size exclusion chromatography (SEC). The goal of this work is to characterize and quantify the sources of variability that affect SV-AUC results, particularly size distributions for a monoclonal antibody monomer/dimer system. Contributions of individual factors to the overall variability are examined. Results demonstrate that alignment of sample cells to the center of rotation is the most significant contributing factor to overall variability. The relative importance of other factors (e.g., temperature equilibration, time-invariant noise, meniscus misplacement, etc.) are quantified and discussed. ß 2009 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 98:3522–3539, 2009 Keywords: analytical ultracentrifugation; sedimentation velocity; protein aggrega- tion; SEDFIT INTRODUCTION Sedimentation velocity analytical ultracentrifu- gation (SV-AUC) has emerged within the bio- pharmaceutical industry as an attractive way to determine protein size distributions. The techni- que’s advantages include size separation without matrix interaction, minimal sample handling, and, in some cases, the ability to directly analyze therapeutic proteins in product formulation with- out dilution. 1 However, compared to orthogonal sizing techniques such as size exclusion chroma- tography (SEC) and field flow fractionation (FFF), SV-AUC is plagued by relatively large and poorly understood variability. 2,3 In biopharmaceutical applications, it is increas- ingly necessary to build analytical and statistical formalisms (e.g., limit of quantitation (LOQ), limit of detection (LOD) and equivalence limits) around the results obtained from SV-AUC experiments. This is an exceedingly difficult problem; the interplay of numerous and diverse hardware-, software-, and procedure-related factors that contribute to the technique’s variability remains unclear. Previous work in our laboratory has focused on limitations of software and data fitting approaches, 4,5 primarily using simulated data to quantify the capabilities of the SEDFIT/c(s) data analysis software. 6,7 Pekar and Sukumar 8 exam- ined overall variability and justified an apparent experimental LOQ derived from replicate mea- surements. Correspondence to: Kelly K. Arthur (Telephone: 303-401- 7519; Fax: 303-401-4403; E-mail: kmacdona@amgen.com) Journal of Pharmaceutical Sciences, Vol. 98, 3522–3539 (2009) ß 2009 Wiley-Liss, Inc. and the American Pharmacists Association 3522 JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 10, OCTOBER 2009