First-in-patient dose prediction for adeno-associated virus-mediated hemophilia gene therapy using allometric scaling

Abstract

Rational first-in-patient (FIP) dose selection is critical for successful clinical development of gene therapy. In this study, the author compared the performance of two allometric scaling approaches and body weight-based dose conversion approach for FIP dose prediction for adeno-associated virus (AAV)-mediated hemophilia gene therapy. The performance of the three approaches was examined using preclinical and clinical efficacy data of nine AAV vectors. In general, body weight-based direct conversion of effective doses in monkeys or dogs was more likely to underestimate FIP dose but worked better for one bioengineered vector with a high transduction efficiency specifically in humans. In contrast, allometric scaling between gene efficiency factor (logGEF) and body weight (logW) was likely to overestimate FIP dose but worked better than the other two approaches when vector capsid-specific T-cell responses were detected in patients. The third approach, allometric scaling between logGEF and W-0.25 was appropriate for FIP dose prediction in the absence of T-cell responses to AAV vectors or a dramatic difference in vector transduction efficiency between animals and humans. A decision tree was tentatively proposed to facilitate approach selection for FIP dose prediction in AAV-mediated hemophilia gene therapy.