xenografts and corresponding areas of adjacent airway are shown in Figures 6 and 7. There was 
no obvious differences in the overall structure or ultrastructural features of the various 
samples except that the native structure occasionally appeared taller. The distribution of cell 
types did not significantly differ between the non-CF bronchus and non-CF xenograft or the 
non-CF and CF xenograft. The CF bronchus had a significantly increased distribution of goblet 
cells when compared to the CF xenograft suggesting environmental stimuli may be responsible, 
in part, for goblet cell hyperplasia in CF (32% in bronchus vs. 24% in graft). 
A detailed description of the electrophysiologic properties of the non-CF and CF xenografts are 
presented in section II.B.6. Manuscripts describing the use of the xenograft model are contained 
in Appendices C, D, and E. 
II.B.5. Safety and Efficacy Studies in Human Non-CF Xenografts 
Introduction. Xenografts generated from non-CF cells were used to address several issues 
critical to the evaluation of the feasibility and safety of adenoviral-mediated gene transfer to the 
human airway. 
Xenografts repopulated with a fully differentiated epithelium were infected with purified stocks 
of a recombinant virus expressing the easily detectable marker gene lacZ. This virus was 
selected because the product of lacZ is easily detected in situ and the overall structure of the 
virus is similar to the Ad.CB-CFTR virus. This recombinant called Ad.RSVIacZ has a lacZ 
minigene expressed from a CMV promoter in place of El sequences as well as deletion of the E3 
region. Into the lumen of the grafts were injected the following viruses: 1) Ad.RSVIacZ, 2) El 
deleted adenovirus (AdEldelta), 3) wild type Ad5,and 4) PBS ("mock"). The virus was 
expelled and the xenografts were subjected to the analyses summarized in Figure 8. Specific 
details of the experiments are provided below. Appendix E contains a manuscript describing 
these experiments. 
Efficiency and persistence of adenoviral-mediated oene transfer in human bronchial epithelium. 
Xenografts were infected with concentrated purified stocks of Ad.RSVIacZ adenovirus at 10 12 
pfu/ml, 10 9 pfu/ml, and 10 8 pfu/ml for a total of 1 hour. The El deleted adenovirus 
(AdEldelta) at 10 12 pfu/ml was used as a negative control. The grafts were harvested, fixed, 
stained in X-gal and visualized enface through a dissecting microscope in order to assess the 
overall efficiency of infection (Figure 9). Xenografts infected with the AdEldelta demonstrated 
no X-gal positive cells when harvested 3 days after exposure to virus (Figure 9D, inset), while 
large areas of lacZ expression were demonstrated in grafts exposed to Ad.RSVIacZ at 10 12 
pfu/ml and harvested at day 3 (Figure 9D). Morphometric analysis of GMA sections of this 
xenograft indicated gene transfer in 11% of the epithelial cells (Figure 10A and 10B). 
Xenografts infected with Ad.RSVIacZ at 10 9 and 10 8 pfu/ml demonstrated gene expression in 
1.9% and <0.1% of the total cells of the epithelium, respectively, when harvested at three days. 
Enface micrographs of these xenografts are provided in Figure 9A (10 8 pfu/ml) and Figure 9B 
( 1 0 9 pfu/ml). To analyze the stability of adenoviral lacZ expression within the bronchial 
epithelium, xenografts were also harvested 21 days following infection with 10 9 pfu/ml 
Ad.RSVIacZ adenovirus (Figure 9C). No changes in the percentage of X-gal positive cells were 
seen between 3 days and 21 days postinfection (compare Figure 9B with 9C, respectively). 
Similar experiments were performed with Ad.CB-CFTR virus to assess the efficiency of 
recombinant CFTR expression in the xenografts. Non-CF xenografts were exposed to 10 10 
pfu/ml of virus and subsequently explanted, frozen, cryosectioned, fixed and immunostained 
using the CFTR COOH terminal antibody. Figure IT shows immunocytochemical analyses of Ad. 
Recombinant DNA Research, Volume 16 
[815] 
