ry lavage fluid, posterior pharynx and rectal swab samples using the cul- 
ture techniques described in the previous paragraph. Neither control animal 
administered vehicle or the animals administered AvlCFl , had adenovirus 
detectable in any sample (data analyzed to 42 days post-administration) . 
In rhesus study 5C, AvlCFl (n=2) was delivered to the right mainstem bron- 
chus at day zero, followed by Ad 5 (10* pfu/kg) delivered to the mainstem 
bronchus at day 7 and an additional 2 animals received only Ad 5 (same 
dose) at day 7. Samples were obtained for evaluation as described for group 
4C. Adenovirus could not be detectable in any samples from these animals 
(data analyzed to 35 days post-administration). 
In summary, replication deficient adenovirus was infrequently detected in 
secretions, typically at less than one week after administration and no 
longer than 15 days after administration. 
To evaluate the possibility that respiratory administration of a recombi- 
nant vector would gain access to the gastrointestinal tract and be subse- 
quently shed [something observed with replication competent Ad infection in 
which fecal shedding is typically observed (Fox et al . , 1969)], rhesus 
(n=3; rhesus study El) were given a replication deficient recombinant 
adenovirus containing the human al- antitrypsin cDNA (see Rosenfeld et al., 
1991 for details) in enteric release capsules (10 8 - 10 9 pfu/kg , n=3 rhesus, 3 
total doses each) . 
The capsules were administered to 3 animals housed in open caging in the 
presence of 2 non- treated, sentinel animals. Sentinel animals were housed 
next to or across from treated animals. All animals were in the same room 
maintained under negative airflow (relative to the hall) , and a dry bedding 
system, changed three times weekly, was used. Recombinant adenovirus, in 
capsules designed for enteric release, was administered on days 1,4 and 7. 
Secretions (posterior pharynx and rectum fecal samples) for adenovirus 
detection (by culture on 293 cells) were sampled from treated and sentinel 
animals one day after each dose, and at days 14, 28, and 43. All treated 
animals had adenovirus detected by culture from secretions (oral and/or 
fecal) one day after at least two of the adenovirus doses. However, no 
sentinel animal had adenovirus detected in secretions at any time. This 
suggests that inadvertent transfer of a recombinant virus from a treated to 
an untreated individual via this route is unlikely. 
4.8 Are there Risks of Having Some of E3 Deleted From the Vector? 
The Ad5 E3 region includes 3456 bp (nucleotides 27576 to 31032, map units 
76.6 to 86.2) and codes for at least 6 proteins. In AvlCFl (which is based 
on Addl327) , and AdCFTR (which is based on Addl324) , 66% of E3 is deleted 
(nucleotides 28,592 to 30,470 in Ad5 ; map units 78.4 to 84.7). In wild type 
adenovirus, the E3 mRNAs arise by alternative splicing of a common precur- 
sor with two polyadenylation sites, E3A and E3B (Cladaras and Wold, 1985; 
Chow et al . , 1979; Wold and Gooding, 1991). The E3A products include: 
gpl9K, 11. 6K, 6.7K, and 3.6K. The E3B products include: 7.5K, 10. 4K, 12. 5K, 
14. 5K, and 14. 7K. The predicted 3.6K, 7 . 5K and 12. 5K products have not been 
identified. 
[716] 
Recombinant DNA Research, Volume 16 
