of cytologically malignant cells within the fluid. Progression of carcinomatous meningitis is manifested by 
altered mental status, cranial nerve abnormalities, signs of increased intracranial pressure, and increasing 
numbers of cytologically malignant cells within the cerebrospinal fluid. The patients with carcinomatous 
meningitis are the most clinically unstable and must be carefully selected to avoid studying patients who 
are likely to die quickly. Median survival of patients with carcinomatous meningitis is dependent on 
clinical status while median survival of patients with malignant ascites and malignant pleural effusion is 6- 
18 months. Because large numbers of such patients have been followed, the usual progression of the 
disease is predictable enough to allow for meaningful assessment of the proposed therapy .The effect of 
therapy can be followed by analyzing key endpoints of disease progression which are described in the 
protocol: including time to reaccumulation of fluid, abdominal girth, weight gain or loss. X-ray changes, 
and clinical signs. 
Preclinical data suggests the feasibility of in vivo retroviral vector-mediated gene therapy for the 
treatment of metastatic breast cancer involving meninges, peritoneum, or pleura. The model system of 
metastatic breast cancer growing within body fluids has several advantages of safety and efficacy for 
retroviral-mediated gene transfer into solid tumors. First, the pathology of metastasis into mesothelial-lined 
spaces consists of relatively thin tumor plaques with extravasation of cancer cells into the fluid: potentially 
allowing a reservoir for delivering retroviral vectors to malignant cells. This spreading pattern is in contrast 
to primary breast cancer and more solid metastatic sites where the breast cancer grows has a large 3- 
dimensional mass which could prevent delivery of retroviral vectors or other agents into the majority of 
tumor cells. In addition, infusion of retroviral vectors into these fluids should produce a greater effect on 
the growing cells within the fluid (predominantly cancer cells) then on proliferating cells in regions of the 
body which do not directly connect to the mesothelial-lined spaces. Within this malignant fluids, the breast 
cancer cells are the most mitotically active cells so few non-tumorous cells are at risk for infection. 
Second, breast cancer provides a model system in which a tissue-specific promoter (MMTV) may be 
employed to direct gene transfer effects toward the malignant cells without producing expression in non- 
breast cells ( 1 1-17). The use of breast-targeted retroviral vectors should provide specificity since the only 
cells within these malignant fluids which are expected to express the MMTV -regulated antisense genes are 
the breast cancer cells. Therefore the likelihood of selective gene transfer to the tumor cells is enhanced. 
The uptake and expression of the viral vectors can be readily assessed in these model systems because 
these fluids are readily accessible for cytologic, biochemical, and molecular analysis. 
II. Gene mutations in Cancer: Oncogenes and Tumor Suppressor Genes: 
One approach to gene therapy for cancer involves altering genes within cancer cells in order to “repair” 
or suppress their genetic defects. This approach is theoretically possible because cancer cells have well- 
described genetic defects. Although cancer is only rarely a hereditary disease, both animal and human 
studies strongly implicate acquired genetic mutations as the cause of most types of cancer. The mutations 
which cause or are correlated with cancer are of two types: 
1) oncogenes (genetically dominant), in which overexpression or altered function produces tumorigenesis 
2) tumor suppressor genes (genetically recessive), in which loss of the gene results in tumorigenesis 
The evidence that these gene mutations are implicated in cancer is quite convincing with numerous 
animal studies demonstrating that overexpression of dominant oncogenes in transgenic mice or deletion of 
tumor suppressor genes in “knock-out mice” (homologous recombination) produces cancer. Mutations in 
specific oncogenes and tumor suppressor genes are very common in human cancers: for example, 25-40% 
of colon cancers contain mutations in the ras dominant oncogene while 50-70% of colon cancers contain 
mutations in the p53 tumor suppressor gene [18]. The possibility of gene therapy to repair these defects is 
complicated by the notion that cancer involves a series of gene mutations. However, there are already 
several published studies in which inhibition of a dominant oncogene [19,20] or replacement of a missing 
tumor suppression gene [21 ] reversed the malignant phenotype. This suggests that altering even a single 
genetic defect in cancer cells may have important clinical consequences. Perhaps a more important 
[164] 
Recombinant DNA Research, Volume 20 
