7 based Ad in which the hepatitis surface B antigen was inserted 
into the E3 region demonstrated that the E3~ recipients shed less 
virus than the wild type virus (Tacket et al . , 1992). 
(5) Studies of co- infection of cells in vitro with an E3 + and El- 
adenovirus demonstrate that the E3~ mutant suppresses the growth 
of the wild type E3 + Ad (see Berkner and Sharp, 1983 and section 
4.9) . 
In conclusion, it is preferable not to allow the virus to subvert these 
inflammatory/immune related processes associated with clearing replication 
competent adenovirus. Further, the limited human data available suggests 
the E3“ deletion in a replication competent adenovirus is associated with 
less shedding, i.e., an E3“ vector is likely safer for the individual and 
the environment than is an E3 + vector. 
4.9 Can Recombination or Complementation of the Vector Occur In Vivo Fol- 
lowing Respiratory Administration of the Vector, and If So, Does This 
Pose a Risk to the Patient and/or Environment? 
Although AvlCFl and AdCFTR will not replicate in freshly isolated human 
airway epithelial cells (see Sections 2.3, 4.1), the fact that the vectors 
are fabricated in a fashion that uses complementation to produce the final 
replication deficient recombinant vector implies that the same could occur 
in vivo in respiratory epithelial cells following administration of the 
vector. To demonstrate that complementation of a vector of this design can 
occur in human epithelial cells, HeLa cells were co- infected in vitro with 
AvlCFl and replication competent Ad5 . Whereas AvlCFl DNA did not replicate 
at an MOI of 30 pfu/cell, addition of Ad5 resulted in de novo synthesis of 
AvlCFl DNA (Figure 4.9-A). The same phenomenon of complementation has been 
observed in freshly isolated normal human airway epithelial cells as in- 
fected with AvlCFl followed by Ad5 (not shown). Thus, under the right 
circumstances, complementation can occur. 
In regard to recombination, the vector design is such that a 
recomb inational event including the intact CFTR cDNA in a replication 
competent virus is highly unlikely because of the difficulty in packaging 
such a viral genome ( Ghosh -Choudhury et al . , 1987). It is theoretically 
possible that the 5' end of the large CFTR cDNA could be packaged in such a 
fashion, but this would likely be akin to a stop codon or splice junction 
mutation in CFTR, a circumstance that produces no protein [and usually 
milder respiratory disease than the common CFTR mutations (Cutting et al . , 
1990; Hamosh et al . , 1991)]. 
If complementation were to occur, the result would be more AdCFTR. This 
should not be a problem in terms of over-expression (see Section 4.12), 
although the consequences of an increased dose of vector per se is not 
known (see Section 4.4). Alternatively this could be viewed as positive 
from the therapeutic point of view, in that chronicity of expression, and 
hence chronicity of efficacy, might be extended. 
If recombination were to occur, the overwhelming likely result is a repli- 
18] 
Recombinant DNA Research, Volume 16 
