Innate immune responses to systemically administered adeno-associated virus (AAV) gene therapy vectors have been implicated as potentiators of serious adverse events, and are a significant safety concern. We recently reported that a subpopulation of AAVs which is generated during vector production and associated within or on the surface of extracellular vesicles (EV-AAV) are less immune stimulatory than standard AAV counterparts. Specifically, compared to conventional AAV, EV-AAV incubated with anti-AAV antibodies reduced hallmark type I interferon-a (IFNa) release up to 93% after exposure on human peripheral blood mononuclear cells (PBMCs) in our novel assay. Furthermore, EV-association conferred as much as 48% reductions in CTL activation (CD69+) when compared to matched controls, strongly suggesting a potential for mitigating critical immune response challenges now faced in AAV gene therapy.
EV-AAVs have been shown to evade neutralizing antibodies. Moreover, IFNa release in our assay is dependent upon increased Fc receptor mediated vector uptake by dendritic cells due to AAV-antibody immunocomplex formation. Surprisingly, decreased IFNa release in EV-AAV stimulated conditions did not correlate with decreased vector uptake. Additionally, EVs display phosphatidyl serine on their surface, a lipid moiety enveloped viruses employ to activate Tyro3/Axl/Mer receptors on dendritic cells and dampen type I IFN signaling. We therefore hypothesized that the observed immunomodulatory advantages might be conferred by EVs independent of AAV association.
We purified EVs by size exclusion chromatography, and verified endotoxin concentrations were less than 5 EU/mL. Next, we isolated human PBMCs from random healthy donors and exposed them to antibody-complexed AAV2-CAG>GFP preincubated with increasing doses of purified EV or controls. Baseline IFNa release into supernatant on untreated PBMCs was nearly undetectable. Similarly, IFNa stimulation by AAV2-GFP without antibodies was nearly undetectable, likely due to insufficient internalization by plasmacytoid dendritic cells. However, antibody-complexed AAV vectors induced robust 275, 374, and 415 pg/mL IFNa release (N=3). Importantly, the addition of EVs dramatically decreased IFNa concentrations in a dose dependent manner in all three donors, reaching 100% reduction when 1.6e9 EVs/well were added. This study reveals a potentially substantial role for EVs as an immunomodulatory agent during AAV administration.