91. Universal Purification of AAV Serotypes 1-5 Modified To Contain a Heparin Binding Epitope

To directly evaluate the utility of distinct serotypes of adeno-associated virus (AAV) for gene therapy applications, it will be necessary to purify each in a similar manner. The methodology to obtain highly purified rAAV2 for use in clinical trials has been established using heparin column affinity binding in conjunction with iodixanol gradients. With the exception of AAV3, the other serotypes of AAV do not bind heparin. Recently the amino acids responsible for the ability of AAV2 to bind to heparin were identified (Kern et al 2003, J Virol 77:11072-81) and tested in the context of rAAV5 (Opie et al 2003, J Virol 77:6995-7006). To assess whether all serotypes could be modified to bind heparin, the AAV2 amino acids R585 R588 and A590 were substituted into the homologous positions in serotypes 1, 3, 4, and 5 to generate the pxr1RRA, pxr3RRA, pxr4RRA, and pxr5RRA plasmids. These helper vectors were then used to produce recombinant eGFP virus which were initially purified by either cesium chloride or iodixanol gradients followed by dialysis.1 × 1010 viral genome-containing particles from serotypes 1-5 and the analogous RRA-containing serotypes were applied to three types of commercially available heparin agarose (Sigma: Heparin type I (cat # H6508), II-S (cat # H3025), and III-S (cat # H1277)). As expected, rAAV2 and rAAV3 bound all three types of heparin with rAAV2 binding type III-S and rAAV3 binding type I heparin most efficiently. As expected AAV1, 4, and 5 did not bind to most types of heparin; however, there were some exceptions. Most notably, approximately 70% of rAAV4 bound type III-S heparin while approximately 40% of AAV1, 4, and 5 bound to and eluted from type II-S heparin. These results suggest that other types of heparin should be considered in the optimal purification of AAV serotypes.AAV serotypes modified to contain the RRA epitope bound and eluted from all types of heparin agarose tested in a profile similar to rAAV2. The binding affinity of rAAV5 RRA was the least efficient with approximately 60% of total virus eluting from the columns. rAAV RRA 1, 3, and 4 bound with greater efficiency (between 80-85%). The ability to purify these AAV serotypes in similar manners will allow more accurate comparisons to be made regarding tissue tropisms. In addition, since every purification method utilized for clinical trials must undergo its own certification, a universal purification scheme for all AAV serotypes would eliminate this need.