Early Systems for Generating HDAds

Although, as mentioned above, the first Ad vectors were defective, helper-dependent viruses, with the development of E1-complementing 293 cells vector development focused on FGAd (E1 deleted) vectors. These were much easier to isolate and propagate than HDAds. The possibility of using HDAds for gene transfer was reexamined in studies reported by Mitani et al. (45). These investigators used a p-galactosidase-neomy-cin fusion gene to replace 7.3 kb of essential Ad sequences in an Ad genomic plasmid. Although this modification did not remove all the viral coding sequences, it did render the recombinant defective and helper dependent. This vector was rescued by cotransfection of 293 cells with purified Ad2 DNA, which provided helper functions. One percent to 5% of the resulting plaques turned blue following X-gal staining, indicating rescue of the recombinant virus. Blue plaque isolates were expanded by serial propagation on 293 cells and finally purified by CsCl ultracentrifugation. Fractionation through the gradient resulted in partial purification of the vector from the helper virus due to the difference in their buoyant densities. Significantly, the ability of the vector to transduce cells in vitro and express the reporter gene was demonstrated. However, the yield of vector was quite low. Furthermore, helper virus contamination remained rather high, at about 200- to 500-fold greater than the vector. In addition, the genome of the vector had undergone rearrangement.

Using a different strategy, Fisher et al. (46) constructed a plasmid bearing a 5.5-kb HDAd genome containing the Ad 5' ITR and packaging signal, LacZ reporter gene, and 3' ITR. 293 Cells were infected with a FGAd to serve as a helper virus and subsequently transfected with the HDAd plasmid DNA. The HDAd was amplified by serial coinfections and finally purified by CsCl ultracentrifugation. Using this method, partial purification of the vector could be achieved, but with the helper virus still in 10- to 100-fold excess. In addition, considerable genomic rearrangements in the form of concatemerization of the vector were observed, which were likely due to the small size of the vector's genome (47). Nevertheless, the vector was demonstrated to be capable of transducing cells in vitro. Using this strategy, a vector bearing cystic fibrosis transmembrane conductance regulator (CFTR) (46) and dystophin (48) expression cassettes was generated and shown to transduce cells in vitro.

Another strategy for generating HDAds was reported by Kumar-Singh and Chamberlain (49). In this system, 293 cells were cotransfected with a plasmid bearing the HDAd genome and purified Ad DNA to provide helper functions. The HDAd contained a LacZ reporter gene and dystrophin cDNA, whereas the helper virus contained the alkaline phosphatase (AP) reporter gene. The vector was amplified by serial propagation and purified by CsCl ultracentrifugation. This resulted in a final vector preparation with 4% helper virus contamination, as determined by AP:LacZ ratio. Importantly, this vector was capable of transducing myogenic cell cultures and express dystrophin in myotubes.

These early systems showed that HDAds could indeed be generated and could transduce a variety of target cells in vitro to direct transgene expression. However, they also emphasized the need to further improve production strategies because the high levels of helper virus contamination, low vector yield and, in many cases, vector genome rearrangement were clearly obstacles that needed to be addressed before the full potential of HDAd could be realized. In particular, relying solely on physical separation between the vector and the helper by CsCl ultracentrifugation was clearly inadequate to achieve the desired vector purities. Therefore, strategies to preferentially inhibit helper virus propagation, while not affecting its ability to trans-complement the HDAd, were required.

One system designed to specifically address preferential inhibition of helper virus propagation was reported in 1996 by Kochanek et al. (50). This strategy was based on early studies in Ad packaging by Grable and Hearing (51), which showed that deletion of 91 bp from the packaging signal severely reduced, but did not abolish, encapsidation of the Ad genome into virions while not affecting viral DNA replication. More important, Grable and Hearing observed a competition for packaging in cells coinfected with 2 Ads: one containing a wild-type packaging signal and the other having a mutant packaging signal. The former was packaged preferentially, almost to the exclusion of the latter. Taking advantage of these observations, Kochanek et al. deleted this 91-bp deletion from the Ad packaging signal of the helper viral genome, thus impairing its ability to be packaged but not affecting its ability to replicate and thus trans-complement the HDAd genome (50). In addition, because the HDAd genome contained the wild-type Ad packaging signal, it would be preferentially packaged over the helper viral genome. The combination of this strategy in conjunction with CsCl ultracentrifugation resulted in a final vector preparation with 1% helper virus contamination. A similar strategy, but using a different packaging signal mutation, was employed by Alemany et al. (52) to generate HDAds. These strategies represented a significant improvement over the previous methods in terms of lower levels of helper virus contamination. However, 1% helper virus contamination would still likely be too high for human use, especially if high doses of HDAds were required. Furthermore, the 91-bp packaging signal deletion resulted in 90-fold reduction in yield of helper virus. From a practical standpoint, this would render production of helper virus stocks problematic because large amounts would be needed to produce the considerable quantities of HDAd needed for clinical applications.

Another system designed to specifically address propagation of the helper virus was reported by Liber et al. (53). This system took advantage of the bacteriophage P1 Cre recombi-nase, which catalyzes site-specific recombination between 34 bp loxP. In this method, an FGAd was engineered to contain 2 parallel loxP sites, one immediately downstream of transgene expression cassette used to replace E1 and the other in E3. Following infection of 293 cells expressing Cre, 25 kb of Ad sequences was excised from the FGAd, leaving behind a 9kb HDAd. Thus, the unrecombined FGAd acted as the helper virus for the propagation of a 9-kb HDAd. Unfortunately, the small HDAd was unstable in vivo, providing limited duration transgene expression for reasons that are still not fully understood. Furthermore, the vector still retained Ad coding sequences and, because the HDAd originated from an FGAd, had limited cloning capacity. Clearly, further improvements were needed if HDAd were to progress into clinical applications.

0 0

Post a comment