Recombinant adeno-associated virus (rAAV) is produced by transfecting cells with two constructs: the rAAV vector plasmid and the rep-cap plasmid. After subsequent adenoviral infection, needed for rAAV replication and assembly, the virus is purified from total cell lysates through CsCl gradients. Because this is a long and complex procedure, the precise titration of rAAV stocks, as well as the measure of the level of contamination with adenovirus and rep-positive AAV, are essential to evaluate the transduction efficiency of these vectors in vitro and in vivo. Our vector core is in charge of producing rAAV for outside investigators as part of a national network promoted by the Association Française contre les Myopathies/Généthon. We report here the characterization of 18 large-scale rAAV stocks produced during the past year. Three major improvements were introduced and combined in the rAAV production procedure: (i) the titration and characterization of rAAV stocks using a stable rep-cap HeLa cell line in a modified Replication Center Assay (RCA); (ii) the use of different rep-cap constructs to provide AAV regulatory and structural proteins; (iii) the use of an adenoviral plasmid to provide helper functions needed for rAAV replication and assembly. Our results indicate that: (i) rAAV yields ranged between 10(11) to 5 x 10(12) total particles; (ii) the physical particle to infectious particle (measured by RCA) ratios were consistently below 50 when using a rep-cap plasmid harboring an ITR-deleted AAV genome; the physical particle to transducing particle ratios ranged between 400 and 600; (iii) the use of an adenoviral plasmid instead of an infectious virion did not affect the particles or the infectious particles yields nor the above ratio. Most of large-scale rAAV stocks (7/9) produced using this plasmid were free of detectable infectious adenovirus as determined by RCA; (iv) all the rAAV stocks were contaminated with rep-positive AAV as detected by RCA. In summary, this study describes a general method to titrate rAAV, independently of the transgene and its expression, and to measure the level of contamination with adenovirus and rep-positive AAV. Furthermore, we report a new production procedure using adenoviral plasmids instead of virions and resulting in rAAV stocks with undetectable adenovirus contamination.

Recombinant adeno-associated virus vectors (rAAV) have been successfully used for long-term gene expression in animal models and in patients. However, while the therapeutic potential of rAAV appears promising, safety issues, including contaminants found in vector stocks, must be further evaluated. We previously reported that a cis-acting replication element present within the AAV-2 p5 promoter was responsible for the encapsidation of rep-cap sequences observed during rAAV production. In that study, we also noticed that plasmid-derived prokaryotic sequences (such as the ampicillin resistance gene) could be found packaged into AAV capsids. In this report, first we confirmed and extended the latter observation by analyzing rAAV stocks produced using different procedures. Second, we demonstrated that these plasmid-derived sequences were transferred and persisted in vivo after rAAV injection into different tissues. Third, our data showed that at least some of these packaged plasmid molecules were linked to the AAV ITRs and were present in vivo in a form that could be rescued through bacterial transformation. This study highlights the need for more stringent characterization of rAAV stocks and provides useful information on the development of rAAV production methods that are able to circumvent or limit the generation of such undesirable particles.

No abstract