levels of gene transfer over a log range with respect to the dose of recombinant virus. In 
addition the experiments indicate that the adenoviral gene persists and the recombinant 
minigene is stably expressed for at least 3 weeks without apparent diminution. This model 
would not detect loss of adenoviral infected cells due to a destructive immune response because 
the xenografts are maintained in an immune deficient animal. 
This model also provides an excellent opportunity to address two important questions: Do the 
transducted cells express viral structural genes? Does the genetically reconstituted graft 
produce and shed virus? In each case the answer appears to be no. This suggests that 
replication of El and E3 deleted Adenovirus type 5 recombinants will not occur in a human 
airway. The xenograft is a very sensitive model of replication because the production and 
propagation of virus would not be blunted by an immune response, which would be present in 
patients but absent in the nu/nu mouse. 
II.B.6. Complementation in Human CF Xenografts 
In order to assess the feasibility of gene therapy it is important to determine the level of genetic 
reconstitution that is necessary for detectable improvement in the electrophysiological 
properties of the human airway. Furthermore, it is important to determine if this level of 
correction is possible with the virus under consideration for human trials (Ad.CB-CFTR). This 
issue was initially addressed by Johnson et al. in which they studied the bioelectric properties 
of epithelial sheets formed in vitro with mixtures of CFTR negative and CFTR positive cells 
[Johnson et al., 1992]. Their data indicate that as few as 6-10% genetically corrected cells 
within an epithelial sheet generated CL transcript properties similar to sheets comprised of 
100% corrected cells. 
We studied this question in CF xenografts exposed to Ad.CB-CFTR virus. CF grafts were exposed 
to either Ad.CMV-lacZ or Ad.CB-CFTR and subsequently evaluated in situ for correction of cAMP 
transport across the epithelium using a modification of the procedure described by Johnson et 
al. Nu/nu mice that carried xenografts were anesthetized with ketamine/xylazine in 
preparation for in situ functional assays. The basic concept is to measure transepithelial 
potential difference (PD) in response to various stimuli. The ground electrode is placed 
subcutaneously and the measuring electrode is placed in the graft lumen bathed in the indicated 
solution. The initial measurement is made in the presence of 120 mM CL. The luminal bath was 
subsequently changed in the following way: a) amiloride is added, b) CL is decreased from 120 
to 3 mM, and c) forskolin is added. 
Results obtained with xenografts derived from a non-CF and CF individuals are summarized in 
Figure 15 (panels A and B respectively). Introduction of the Na channel blocker amiloride to 
the lumen bath leads to a drop in PD (i.e., less negative) in both non-CF and CF grafts. The PD 
increases (i.e., become more negative) in a stepwise manner in the non-CF graft when luminal 
CL is decreased (representing basal conductance through CFTR) and when CFTR is stimulated 
with forskolin. 
CF xenografts exposed to Ad.CMV-lacZ or Ad.CB-CFTR were subjected to similar studies. The CF 
graft exposed to Ad.CMV-lacZ (Figure 15, panel D) was indistinguishable from the noninfected 
CF graft (Figure 15, panel B) while the CF graft exposed to CFTR adenovirus demonstrated 
significant increases in PD in response to changes in luminal CL and forskolin (Figure 15, 
panel C). 
These xenografts have been explanted and evaluated for efficiency of genetically reconstituted 
cells. Previous experience indicates that CFTR expression should be detected in 5-20% of the 
cells. Analyses of these grafts presented in Figure 15 are underway. 
Recombinant DNA Research, Volume 16 
[817] 
