the airway epithelium. In this scenario, trans -epithelial electrolyte trans- 
port likely modulates the quantity and composition of respiratory tract epithe- 
lial fluid. Thus, an inability to actively transport Cl' from the submucosa to 
the epithelial surface in a normal fashion would limit H 2 0 from being secreted 
from the epithelium in a normal fashion. Consequently, the airway secretions 
would be relatively dehydrated, providing a rational scenario for the subse- 
quent development of mucus obstruction, infection and inflammation. Second, 
the respiratory disease results from processes internal to the respiratory 
epithelial cells. In this scenario, there may be abnormalities in the pH of 
intracellular organelles causing dysfunction of enzymes modifying mucins or 
abnormal expression of cell components that interact with microorganisms such 
as Pseudomonas . In either scenario, demonstration of normal CFTR mRNA ex- 
pression in the airway epithelium would constitute a significant advance in 
developing gene therapy for CF. 
It is also not known whether all airway epithelial cells are involved in the 
pathogenesis of the airway disease, or only a subset. If the "exterrfal elec- 
trolyte milieu" hypothesis is correct, it is likely that most, if not all of 
the epithelium is involved, consistent with the widespread (albeit low) ex- 
pression of the CFTR gene in the surface airway epithelium. Alternatively, if 
the "internal milieu" hypothesis holds, a subset of cells, such as mucus pro- 
ducing cells, may play a dominant role. Independent of the mechanisms leading 
to the abnormal mucus, bacterial colonization and inflammation, there is gen- 
eral agreement that the intense neutrophil -dominated inflammation deranges the 
airway epithelium (11,12). While a variety of mediators are likely involved, 
injury to the epithelium is mediated to a large extent by neutrophil elastase 
and neutrophil -generated oxidants (11,13,58). 
In summary, the available evidence presents an overwhelming case for the fun- 
damental abnormality in the lung of CF patients to be an abnormality in ex- 
pression of the CFTR gene in the airway epithelium. This strongly supports the 
concept that it is rational to use gene transfer directly to airway epithelial 
cells to correct the CF defect that ultimately causes the major clinical mani- 
festation of the disease. One strategy to accomplish this is to utilize a 
replication deficient recombinant adenovirus that contains an active promoter 
and a normal CFTR cDNA (59). In addition to the ability to accommodate a large 
(up to 7.5 kb) exogenous cDNA, the adenovirus has the advantage of being trop- 
ic for respiratory epithelium and capable of transferring recombinant genes 
into non-proliferating cells. The recombinant adenoviral approach has been 
successful in transferring the human al- antitrypsin gene to the respiratory 
epithelium of experimental animals in vivo (60) and directly relevant to this 
protocol, recombinant adenovirus containing a normal CFTR cDNA is successful 
in mediating jin vivo transfer and expression of the human CFTR gene to the 
respiratory epithelium of the lungs of experimental animals (59). Further, a 
recombinant adenovirus containing the human CFTR cDNA will correct the Cl' 
secretory abnormality of epithelial cells of individuals with CF in vitro. In 
contrast to the adenovirus, there is currently no other vector system capable 
of transferring the CFTR cDNA to the airway epithelium in vivo with efficacy 
and efficiency necessary to treat the disease. The only other vector system 
that has been shown to mediate in vivo transfer of the human CFTR cDNA to the 
respiratory epithelium jji vivo in experimental animals is double stranded 
circular DNA plasmids combined with liposomes (61) , and more recently with an 
adeno-associated vector (62). However, the extent of expression with these 
[ 202 ] 
Recombinant DNA Research, Volume 20 
