Spotlight session: The need for an unbiased assay to detect and quantify replication competent AAV in clinical vector products
Replication-competent AAVs (rcAAV) are gene therapy product-related impurities that could have associated safety risks for patients. rcAAVs are conventionally detected in cell-based assays consisting of successive rounds of amplification in the presence of an adenovirus with a qPCR readout on viral gene sequences. The absence of rcAAV detection in 1E+08 vg has generally been used as an acceptance criterion.
Our data show that this cell-based method can yield significantly different results between testing facilities using different assay protocols. Moreover, we showed that positive or negative results are highly dependent on the capacity of AAV serotypes to infect the permissive cell line in vitro, which could be quite different from what is observed in vivo.
An analytical method that allows the unbiased detection and quantification of rcAAVs is thus needed. We have developed a method based on a multiplex digital PCR (dPCR) that targets 3 junctions specific to rcAAV genomes: ITR-Rep, ITR-Cap and Rep-Cap. It accounts for the presence of potential exogenous sequences and it takes advantage of the partitioning of the dPCR to isolate particles containing triple positive genomes. Several recombinant AAV vector batches of various serotypes, positive or negative for rcAAV in the cell-based assay showed similarly extremely low levels of triple positive genomes (less than 0.00008% of genomes quantified by qPCR) in comparison to wild type (wt)AAV. Interestingly the latter showed that only 3% to 7% of their genomes were triple positive. Results were independent of the HEK293/transfection manufacturing protocol. However, no triple positive entities were found in baculovirus-generated preparations likely due to the modifications in constructs of the vector system.
Our data suggest that there are likely no HEK293/transfection-derived vector AAV preparations devoid of rcAAV genomes. In this context our efforts constitute the first step toward the goal of changing the way rcAAVs are detected and quantified. More results from other sponsors using AAV vectors are needed to solidify our preliminary data set and to engage in a field-wide discussion, including regulators, on how to generate specifications and acceptance criteria for this novel analytical method. This will help to better reflect on and provide mitigation strategies for potential safety risks associated with rcAAVs until means to eliminate them from vector batches have been identified.
