Potency testing for a recombinant protein vaccine early in clinical development: Lessons from the Schistosoma mansoni Tetraspanin 2 vaccine

• Main Authors: Guangzhao Li, Lara Hoeweler, Brian Keegan, Jin Peng, Larissa Scholte, Peter Hotez, Maria Elena Bottazzi, David Diemert, Jeffrey Bethony
• Format: Article
• Language: English
• Published: Elsevier 2021-08-01
• Series: Vaccine: X
• Subjects: Vaccine; Potency; Compliance approach; Least-squares regression; Control charting; Bootstrap modeling
• Online Access: http://www.sciencedirect.com/science/article/pii/S2590136221000176

Introduction: As a primary stability-indicating parameter, potency should be strategically evaluated during each phase of vaccine development. Herein, we present potency testing during the early clinical development of the Schistosoma mansoni (Sm) Tetraspanin-2 vaccine formulated on Alhydrogel (Sm-TSP-2/Al). As Sm-TSP-2/Al does not induce sterilizing immunity against its target pathogen (Sm) in animal models, potency is measured by “serological substitution”, a method that can add significant variation to the potency metric, especially when used in a compliance (or ‘single data point’) approach. Methods: Potency data were analyzed using the compliance approach to determine if two clinical lots of Sm-TSP-2/Al retained potency over 84 and 36 months post-release, respectively. These same data were also analyzed by: i) least-squares regression with a joinpoint regression analysis; ii) control charting of stability slopes; and iii) bootstrap modeling. Nested-regression and bootstrapping were used to compare the potency of the first (#11-69F-003) and second (#1975) clinical lots of Sm-TSP-2/Al. Results: Despite significant variability in the immune assay, both clinical lots of Sm-TSP-2/Al remained potent for 84 and 36 months, respectively, in all four statistical approaches. The first lot of Sm-TSP-2/Al showed a gain in potency starting at 36 months post-release as captured by joinpoint regression. The two clinical lots of Sm-TSP-2/Al had comparable long-term potency. Conclusion: While a compliance approach can monitor the long-term stability of Sm-TSP-2/Al, it risks putting this critical stability-indicating parameter out of specification with each time point tested due to statistical multiplicity. Alternative statistical methods, such as joinpoint regression or bootstrapping, do not have this limitation and offer even more precise estimations of potency, with the added benefit of also providing predictive analytics. Nested regression and bootstrapping were shown to be a viable alternatives for lot-to-lot comparisons of the stability of Sm-TSP-2/Al. Instructions for implementing both these potency testing approaches are provided.