M.J. Welsh and A.E. Smith, RAC Application 
mutant protein was misfolded and therefore recognized by the cellular quality control 
mechanism located in the endoplasmic reticulum and was degraded. Early data on this 
hypothesis was conflicting (29-31,39). More recently, we evaluated the hypothesis in 
airway epithelia using a quantitative assay to detect CFTR in the apical membrane (32). 
CFTR was present in the apical membrane of normal airway epithelia, but none was 
detected in the apical membrane of CF epithelia bearing the AF508 mutation. Studies in 
the sweat gland duct epithelium have also shown that in CF patients with the AF508 
mutation, the mutant protein is mislocalized and is probably not present in the plasma 
membrane (59). The conclusion that the AF508 mutation causes defective protein 
processing was also supported by our studies that tested the effect of temperature on 
processing of CFTRAF508; a reduction in incubation temperature can " corre ct" the 
processing of some mutant proteins. We found that the processing of CFTRAF508 reverted 
toward that of wild-type as the incubation temperature was reduced (60). When the 
processing defect was corrected, cAMP-regulated chloride channels appeared in the plasma 
membrane. This result also explained some earlier data, obtained from Xenopus oocytes 
(oocytes are grown at 19°C); that data suggested that CFTR containing the AF508 mutation 
was present in the plasma membrane (39). 
The results of these studies indicate that the CFTR chloride channel is either missing from 
the apical membrane of CF cells bearing the AF508 mutation and some other mutations 
(32,57,58), or it is present at much reduced levels. Thus the results explain the lack of 
apical chloride permeability in most CF epithelia. Unfortunately, because the protein is 
mislocalized, therapeutic attempts designed to alter the function of mutant CFTR are not 
expected to be successful for the majority of patients. On the other hand, pharmacological 
attempts to relocate mutant CFTR could be attempted. 
b. The mutated protein may have abnormally reduced chloride channel function. Some 
CF-associated mutations generate proteins that appear to be processed correctly and hence 
are presumably correctly located in the plasma membrane (57). Nevertheless, they do not 
function normally (58). Examples include: the G551D mutation, which has no detectable 
function in halide efflux assays (58); the G1349D mutation, which produces a chloride 
channel that has a markedly reduced probability of being open (48,58); and the S1255P 
mutation, which has a significantly reduced stimulation by intracellular ATP (48). It is also 
possible that some mutations could lead to a defect in both intracellular trafficking and 
function. 
c. Some mutations fail to produce a complete protein. Some mutations produce frame 
shifts, splicing mutations, or premature termination of translation (3,55,56). Such 
mutations would be expected to fail to generate a complete protein. Recent data suggest 
that some mutations that encode a premature stop codon lead to markedly reduced levels of 
mRNA and thus presumably little or no protein (61). 
3.0 CONSIDERATIONS FOR GENE THERAPY OF CF 
Gene therapy of CF presents several considerations that are different from those addressed 
in most previous proposals submitted to the RAC. 
3.1 Target Tissue 
Because 95% of CF patients die of lung disease, the lung will be the main target for gene 
therapy. The hallmark abnormality of the disease is defective electrolyte transport by the 
epithelial cells that line the airways. Numerous investigators (reviewed in 8) have 
[860] 
Recombinant DNA Research, Volume 16 
