Neutralization of recombinant adeno-associated virus (AAV) gene transfer vectors by pre-existing antibodies is a significant barrier to clinical gene transfer using systemic administration. In a recent clinical trial for hemophilia B, while efficient gene transfer and therapeutic levels of hFIX were achieved in a human subject with low pre-existing antibodies to AAV2, no gene transfer was observed in a subject with a modest pre-existing AAV2 antibody titer of 1:17 who received the same dose (

Manno et al.
Nature Med.
). With the objective to achieve consistent and efficient transduction in the presence of AAV antibodies, ‘humanized’ AAV vectors were generated by covalent attachment of human albumin (HSA) to the surface of purified AAV2 by cross linking the single free thiol (Cys34) of HSA to lysine residues on the vector surface using suflosuccinimidyl 4 [N-maleimidomethyl] cyclohexane-1-caboxylate (Sulfo-SMCC). Variable levels of HSA substitution up to 65 HSA molecules per virion were achieved, corresponding to a maximum of approximately one HSA molecule per virion capsid protein. Analysis by SDS-PAGE of the modified vector revealed the presence of a 130kDA polypeptide species in the conjugated vector, not observed in the unmodified vector, that was consistent with predicted molecular weight obtained by covalent attachment of HSA with AAV capsid protein VP3. Dynamic light scattering demonstrated an average radius of 16.6 nm for the HSA-conjugated vector compared with 11.6 nm for unconjugated vector, consistent with coating of the vector surface with HSA. Analysis by flow cytometry demonstrated that the binding of AAV2-specific monoclonal antibody A20 to unmodified vectors was progressively reduced to background levels in HSA-conjugated vectors with increasing HSA substitution. Quantitative analysis by Biacore indicated that the association rate and available sites per virion for antibody binding were reduced > 10- fold in HSA-conjugated vectors. The conjugated vectors were significantly more resistant than unmodified vectors to neutralization by AAV antibody, demonstrating comparable transduction in the presence of 16-fold higher concentration of MAb A20 following transduction of cultured human hepatocytes (HepG2). In conclusion, covalent coupling of HSA to the surface of recombinant AAV2 is demonstrated as a feasible method to modify AAV vectors to increase their resistance to neutralizing antibodies. ‘Humanized’ AAV vectors such as HSA-coated AAV that effectively mask viral epitopes recognized by antibodies and display human epitopes may enable more consistent transduction following systemic administration in vivo, and enable vector readministration.

Disclosures: Wright:Genzyme: Consultancy; Tacere: Consultancy.