(M1430-01-01) Discrepancies in Adeno-Associated Virus Serotype and Isotype Characterization Positivity Rates in Cynomolgus Serum

Purpose: Considering the high prevalence of adeno-associated virus (AAV) anti-drug antibodies (ADA) in the population, including potential cross‑reactivity between AAV serotypes, a significant number of individuals must be screened to determine a negative population to be used during immunogenicity cut point analysis.  In current literature, there is a wide range of pre-existing immunity rates between different AAV serotypes.  In the following experiments, individuals were evaluated and compared for positivity rates of AAV9 and AAV8 total antibody (TAb) and IgG isotype (IgG) assays.  Individuals’ results were also compared in both types of assay formats to better understand concordance between TAb and IgG characterization.  During isotype characterization, it was expected that a majority subset of confirmed TAb positive individuals would be IgG positive, while those TAb negative would remain negative when evaluated for IgG.

Methods: AAV9 and AAV8 TAb assays were evaluated with a sequential sandwich format where samples, positive controls (PCs), and negative control (NC) were added to an MSD Standard Bind plate coated with AAV capsid.  Any ADA present in the cynomolgus monkey serum bound to the AAV on the plate.  This complex was then detected by Sulfo-Tag-AAV. Due to high prevalence of observed coefficient of variation between duplicates above twenty percent and to eliminate potential non-specific matrix interference, AAV9 and AAV8 IgG assays were evaluated with a BEAD format.  Samples, PCs, and NC were incubated with biotin-AAV and any ADA present in the cynomolgus monkey serum would form a bridge with the biotin-AAV molecules.  Streptavidin beads were added to form a streptavidin-biotin-AAV ADA complex.  The bead complex was then acid treated to elute the ADA from the bead complexes.  The acid treated samples were incubated with biotin-AAV neutralization solution.  The neutralized samples were bound to a blocked MSD-Streptavidin (MSD-SA) plate and then detected by Sulfo-Tag-chicken anti-human IgG Fc antibody. Samples that contained anti-AAV antibodies bound to AAV generated a signal when voltage was applied.  The resulting RLU signal (relative light units) is directly proportional to the amount of ADA present in the sample.  Individuals are considered positive above the 5% false positive rate cut point in the TAb assays, or above the 1% false positive rate cut point in the IgG assays.

Results: Overall, 43 out of 80 unfortified individuals (approx. 54.0%) screened above the 1.20 S/N cut point for AAV9 Tab ADA, whereas 75 out of 101 individuals (approx. 75.0%) screened above the 1.19 S/N cut point for AAV8 Tab ADA.  These results indicate higher pre-existing immunity to AAV8 than AAV9 in the cynomolgus population. In AAV9 IgG analysis, the cut point was 2.14 S/N and 43 out of 80 individuals were positive, which matches the expected results based on AAV9 TAb positivity rate.  However, nine (11.3%) TAb-negative individuals were positive in the IgG assay and nine (11.3%) TAb-positive individuals were negative in the IgG assay. In AAV8 IgG analysis, the cut point was 9.55 S/N and 88 out of 101 individuals (approx. 88%) were positive, resulting in a higher positivity rate than the AAV8 TAb assay.  Additionally, 15 (14.9%) TAb-negative individuals were positive in the IgG assay whereas two (2.0%) TAb-positive individuals were negative in the IgG assay.

Conclusion: It was observed that there is a higher rate of pre-existing ADA for AAV8 serotype compared to AAV9 in cynomolgus serum individuals, indicating more individuals need to be screened in AAV8 assays to obtain a negative population for cut point analysis.  While it was expected that a majority of TAb‑positive individuals would be IgG positive, it was also expected that TAb-negative individuals would remain negative in the IgG assays.  However, it was observed that AAV TAb-negative individuals were positive for IgG, indicating isotyping results for AAV ADA may be unreliable with current methodology.  The discordance between AAV TAb and IgG results could potentially be due to non‑specific detection or undetected ADA positive individuals in the TAb assays.  Given the current data, it is suggested that all samples should be analyzed in both AAV TAb and IgG assays, regardless of TAb sample outcome, to fully understand sample positivity results.


Total number of positive and negative individuals evaluated in the AAV TAb and IgG assays. In the AAV9 assays, 80 individuals were evaluated and compared. In the AAV8 assays, 101 individuals were evaluated and compared. These results indicate higher pre-existing immunity to AAV8 compared to AAV9.


Individuals were compared between the TAb and IgG assays for AAV9 and AAV8. Individuals are listed as TAb- or IgG-negative (TAb - or IgG -) and TAb- and IgG-positive (TAb + or IgG +). Positivity rates in AAV9 are similar in the TAb and IgG assays, whereas unexpected higher positivity is seen in the AAV8 IgG assay compared to the AAV8 TAb assay.