One barrier to clinical use of gene therapy is the lack of methods to target gene delivery to specific cell types. Retroviral vectors have desirable attributes including transgene integration for long term expression. However, current formulations use broad host range envelope proteins (Env), e.g., amphotropic MLV or VSV G. Previous attempts to restrict transduction by modifying retroviral Env failed for unknown reasons. In previous Aim 1, we identified the defect that prevents gene delivery by Env with ligand sequences inserted at the beginning or end of the receptor-binding domain (RBD), the two most used sites. The primary flaw was failure of target receptor binding to fully activate chimeric Env. In light of these results, we initiated a novel approach to chimeric Env design based on the hypothesis that activation upon receptor binding will be achieved if the ligand sequences replace the natural virus receptor binding site (RBS). Based on this concept, a peptide ligand, somatostatin (Sst), was used to construct a prototype chimeric MLV Env referred to as Sst-RBS. In new preliminary studies, control chimeras in which Sst sequences were inserted in the amino-terminus, the PRR and after residue 230 were also examined. The ability of Sst-RBS to target infection was evaluated by two criteria - acquisition of entry specifically through the Sst receptor and loss of entry via the natural virus receptor. Sst-RBS gave 40,000 TU/ml of unconcentrated virus stock and transduction of 50% of cells expressing Sst receptor, whereas, control chimeras gave almost no infection. Sst-RBS vector showed only rare infection of mouse NIH 3T3 cells. New preliminary studies also show that a two-fold increase in targeted infection was obtained by addition of an arginine 95-to-aspartate change in the Env sequences of Sst-RBS. This variant of Sst-RBS shows complete loss of NIH 3T3 infection, consistent with loss of the ability to interact with the natural receptor. To build on these results we propose to: (i) examine methods for increasing targeted infection of Sst-RBS vectors; (ii) determine if RBS replacement is generally applicable to other peptide ligands, including ones that lack thiol pairs; and (iii) determine if tumor-specific delivery is obtained in vivo in Sst receptor positive neuroblastomas in an immune-deficient setting (scid mouse tumor xenoplant model) as compared to an immune competent setting (the TH-MYCN transgenic mouse model for spontaneous neuroblastoma).
