et al. normal adult volunteers were immunized to Ad5 capsid antigens and subsequently exposed 
to an inoculum of live Ad5 into the respiratory tract. Ten of 19 individuals who received a 
placebo immunization and subsequently live Ad5 had a self limited febrile illness characterized 
by fever and moderate pharyngitis. No significant untoward effects were noted. 
An additional consideration relates to the malignant potential of these gene transfer reagents. In 
contrast to retroviruses which integrate into the genome, adenoviruses are maintained in the 
genome as extrachromosomal elements which theoretically cannot cause a secondary malignancy 
due to insertional mutagenesis. In support of this is the fact that no human malignancy has been 
directly linked to adenoviruses. 
Psychological and social implications. It is likely that the initiation of these experiments will 
be associated with tremendous attention by the media. We will follow a strict code of 
confidentiality as mandated by the RAC and inform the patients of this potential problem. 
V.B. Risks to others 
The only potential risk to others is the possibility that the recombinant virus could be 
transmitted to another individual, take up residence, and cause disease. We believe this is 
highly unlikely for a variety of reasons. 
One very important aspect of this scenario relates to the likelihood that transmission of Ad.CB- 
lacZ will cause disease. We obviously do not think this is likely based on the fact that we are 
proposing to use it therapeutically. Toxicity could be due to one of two mechanisms: expression 
of CFTR or pathology due to the inherent proporties of Ad5. As we argued above, ectopic or 
unregulated overexpression of CFTR does not appear to be toxic. Furthermore, pathology of the 
recombinant due to its properties as an infectious virus is unlikely because all patients will 
have had previous immunity to the virus and the recombinant should have a tremendous growth 
disadvantage due to the deletion of El . 
Another possibility is that the Ad.CB-CFTR will undergoe a recombination with homologous wild 
type Ad to form a more toxic structure. As demonstrated below the most likely recombination 
will be between the large stretch of homology 3' to the CFTR minigene. This will form two 
recombinants: one containing the CFTR minigene with E3 sequences (this exceeds the wild type 
genome size; our attempts to actually make this virus have failed because it cannot be packaged) 
and a wild-type hybrid genome missing E3 sequences (there is no reason to expect this to be 
more pathogenic than either parent). 
E3 deletion 
Map Units 
Recombinant DNA Research, Volume 16 
[ 835 ] 
