that the presence of mutant p53 may be necessary for full expression of the 
transforming potential of the gene. Mutations of p53 are common in a wide 
spectrum of tumors 10 ' 13 ; they occur in both NSCLC and SCLC cell lines and 
fresh tumors 13,14 . 
Options for specific targeting of oncogenes include inhibition of expression of a 
dominant gene or replacement of a deleted or mutated tumor suppressor gene. 
Progress in the understanding of the critical genes involved in tumor 
development and in technology for altering gene expression logically led to our 
studies of techniques for achieving these options. Initially, a model for specific 
inhibition of K-ras was developed. We chose to work with K-ras because of the 
applicability of the findings to a large number of tumors, because of our 
previous work with K-ras, and because information on the genetic organization 
and sequence of the ras gene family was readily available. Advances in 
antisense and retroviral gene transfer technology suggested that application of 
these techniques may mediate specific inhibition of oncogene expression. 
Antisense mRNA, which is precisely complementary to the corresponding sense 
mRNA, inhibits translation. The mechanisms for this inhibition have not been 
completely defined but include inhibition of translation by ribosomes, 
degradation of sense-antisense duplexes by enzymes, and failure of export 
from the nucleus. Thus, specific targeting of a gene in a multigene family could 
occur if it possessed unique sequences in a region amenable to antisense 
inhibition, such as an initiation codon or splice site. 
The working hypothesis that we developed is that reversal of a single 
altered genetic event in the cancer cell can potentially reverse critical 
features of the malignant phenotype of that cell. This finding has important 
therapeutic implications. Cancer cells have multiple genetic alterations. Therapy 
directed toward oncogenes will be practical only if therapeutic effects occur with 
targeting of one or two genes. It is unlikely that any therapy targeting 
oncogenes or their products will be absolutely specific for cancer cells. If other 
genes can compensate for loss of normal function by a specific oncogene 
mediated by an antisense construct, the harmful effects of the therapy will be 
reduced. Studies from our laboratory indicate that reversal of a single genetic 
alteration has profound effects on the growth and tumorigenicity of lung cancer 
cells 15,16 . Additional support for this concept comes from a recent study by 
Soriano and coworkers 17 in which transgenic mice were created that lacked a 
functional c-src proto-oncogene. The resulting developmental defect in the 
mice was osteopetrosis. The ubiquity of c-src, its high degree of conservation 
among species, and its role in mitosis suggest that inactivation would be lethal, 
but this was not the case; viable mice were recovered. A possible explanation is 
that other closely related nonreceptor tyrosine kinases such as yes and fyn can 
compensate for loss of c-src. Introduction of a single copy of a wildtype tumor 
suppressor gene into normal cells would be unlikely to have adverse effects if it 
occurred during therapy directed at replacing inactivated tumor suppressor 
Recombinant DNA Research, Volume 16 
[419] 
