The objective of this application is to develop a unique strategy for the treatment of cancer using gene therapy. The applicant has selected carboxypeptidase A (CPA) as a therapeutic gene based on its ability to convert the prodrug methotrexate-alpha-phenylalanine (MTX-Phe) to methotrexate (MTX). The advantage of the proposed strategy compared to systemic administration is that therapeutic concentrations of MTX can be achieved within the tumor while avoiding systemic toxicity due to tumor specific expression of the activating enzyme. Since CPA is normally secreted from cells as a zymogen and requires cleavage by trypsin in the digestive system for functional activation, the applicants will generate CPA forms that can be expressed in mammalian cells as active enzyme independent of cleavage by trypsin. They will also create a plasma membrane associated form of CPA. The modified forms of CPA will be biochemically and enzymatically characterized in comparison to the authentic trypsin activated CPA in specific aim 1. Evaluation of the modified enzymes expressed in tissue culture cell lines will be performed in specific aim 2. The applicants will determine the cytotoxicity of MTX-Phe on CPA non-expressing and expressing cells. One of the biggest advantages of an enzyme/prodrug approach to gene therapy is that fact that the toxic metabolite (drug) produced by a few transduced tumor cells is able to kill neighboring cells (the bystander effect ). The extent of the bystander effect will be investigated using the CPA/MTX-Phe system. Since MTX is generated extracellularly in this case, it should be able to efficiently kill CPA non-expressing cells within a tumor. This will be studied in a mouse model in specific aim 3.