The first article in a series of highlights from the 2021 AAPS PharmSci 360
Adeno-associated (AAV) gene replacement therapy has shown great promise for the treatment of many diseases, particularly rare disorders with unmet medical needs. However, the exposure-response relationship in AAV gene therapy (GTx) has not been fully established.
Understanding the link between dose and eventual transgene expression and dose and clinical efficacy is crucial to achieving efficacy, said Nessy Tania, Ph.D., in her presentation, Developing a Robust QSP Model of AAV Based Gene Therapy for Clinical Applications.
Achieving the correct first dose of AAV GTx is essential because too low a dose could lessen efficacy, and a too high dose could generate a strong immune response to the transduced cells, said Tania, a Principal Quantitative Systems Pharmacologist at Pfizer.
To improve understanding about exposure-response, Tania and her colleagues developed a mechanistic Quantitative Systems Pharmacology (QSP) model for AAV GTx that integrates available pre-clinical and clinical data on AAV and GTx with this vector.
Their model, which incorporates mechanistic steps linking dose and viral vector distribution to transgene expression and clinical efficacy, was developed, and calibrated based on published scientific data about AAV8 vectors, pre-clinical studies of liver-targeted AAV8 GTxs in laboratory mice, and the Pfizer-sponsored Phase 1/2A clinical trial of Fidanacogene elaparvovec, AAVSpark100 GTx targeting hemophilia B, caused by mutations in the coagulation factor IX (FIX) gene.
Researchers used allometric scaling in their species translational approach. Pre-clinical data provided an initial dose response relationship for model calibration. Additionally, adjustments for human physiological parameters such as volume of compartment and flow rates were made, Tania said.
A comparison of the model predictions with emerging data from the Fidanacogene elaparvovec clinical trial identified a 2-to-3-week lag in FIX activity dynamics, which had not been predicted by the initial QSP model calibrated to pre-clinical data.
The researchers evaluated several possible mechanisms that could cause the slower dynamics of FIX activity. Tania said one possible explanation is a combination of slow AAV processing rates. Allometric scaling imposed on several AAV intracellular processing rates also can explain the slower dynamics, she added.
“The previous translation of only adjusting synthesis rate in the model yielded faster than observed dynamics of FIX activity,” said Tania. “We analyzed several underlying model assumptions and incorporated different mechanistic hypotheses to obtain consistent agreements with observed clinical data.”
Tania added, “To capture the observed slower dynamics of FIX activity, two parameter adjustments on the QSP model for hemophilia B GTx can be made for model translation that still allows for model identifiability.” The two parameters are the protein synthesis rate and protein secretion rate.
After developing a full translation model for hemophilia B, the researchers explored the applicability of the model to GTx clinical trials for hemophilia A, caused by mutations in the the coagulation factor VIII (FVIII) gene.
Emerging clinical data revealed a non-linear dose-response in the hemophilia A GTx. In contrast, a linear or dose proportional response is observed in the hemophilia B clinical trial, Tania said.
Additionally, over several months, circulating FVIII levels decreased in several clinical trial patients, while in hemophilia B patients, FIX expression has been stable for up to four years since transfusion. Tania and colleagues proposed that studying differences between hemophilia A and B using mechanistic modeling may be key in explaining these observed differences.