A small number of genes of Drosophila have been identified by mutations which cause overgrowth of ectodermally derived tissue. Mutations in the best studied gene of this type, lethal giant larvae (lgl), cause loss of normal tissue structure, cell death and loss of capacity to differentiate of organs derived from skin ectoderm (imaginal discs) and neural ectoderm (imaginal neuroblasts). Imaginal discs derived from lgl mutants when transplanted into normal hosts, can grow extensively but do not invade host tissues. Neuroblasts derived from lgl mutants when transplanted into normal hosts, however, do invade host tissues. The invasiveness of such cells suggests that these mutant cells might be analogous to metastatic mammalian tumor cells. It is quite likely that mammalian cells have a gen homologous to lgl because antibodies raised against the Drosophila lgl protein cross react with a cell surface protein of mouse cells. We propose to compare the altered properties of these mutant Drosophila cells to the altered properties of metastatic mammalian tumor cells in order to evaluate whether invasive tumors of Drosophila re relevant models of metastatic tumors of mammals. To do so, we will compare transplanted mutant Drosophila tissue to transplanted non-mutant tissue using antibodies to proteins that have been found by work in the laboratory of Lance Liotta to distinguish metastatic from non-metastatic mammalian cells. These marker proteins include laminin receptor, type IV collagenase, tissue inhibitor of metallo-proteinase-2 and Nm23. The host tissue will be homozygous for a mutation in the aldehyde oxidase gene. the donor tissue will express aldehyde oxidase and so will be easily distinguished from host tissue by histochemical staining. This will also allow us to investigate whether there is any specificity as to which host tissues are invaded. Drosophila invasive tumors could become a valuable model of metastasis because of the power of genetic analysis. We also propose to perform pilot mutagenesis screens to recover mutations which suppress the lgl phenotype. Among the mutations recovered, we hope to identify genes whose products are required for expression of the invasive phenotype caused by mutations in the lgl gene. Once such mutations are recovered their phenotypes will be examined to discover their effect in individuals with wildtype levels of lgl product. The ultimate goal of this approach is to clone genes identified in this way and use them as probes to isolate their mammalian homologues.
