Novel treatment strategies such as gene therapy are warranted in view of the failure of current treatment approaches to cure a high percentage of patients with advanced bladder cancers. Testing of the hypothesis that blocking the angiogenic switch may keep tumour growth in check has been facilitated by the discovery of endogenous inhibitors of angiogenesis and has also added another research dimension to the field of cancer gene therapy. Consequently, the concept of targeting the tumour vasculature with anti-angiogenic agents has emerged as an attractive new strategy in the treatment of cancer. Targeted biological therapies that selectively interfere with tumour angiogenesis could improve survival among patients with bladder cancer. Endostatin is a tumour-derived angiogenesis inhibitor and is the first endogenous inhibitor of angiogenesis to be indentified in a matrix protein. Gene therapy represents an attractive approach to treat cancers and other chronic diseases. The development of an effective delivery system is absolutely critical to the usefulness and safety of gene therapy. At present, the adeno-associated virus (AAV) vector has the most promising potential in view of its non-pathogenicity, wide tropisms and long-term transgene expression in vivo. Gene therapy studies using different serotypes of recombinant AAV (rAAV) as delivery vehicles have proved rAAVs to be an effective modality of cancer gene therapy. In the present study, an IgG fragment was inserted at the start of the sequence coding for endostatin with the aim of enabling continuous secretion of endostatin the serum. We also investigated the suppression effect of AAV-mediated endostatin expression on endothelial cells and in mice xenograft models of bladder cancer. Our data demonstrates that rAAV-endostatin controlled tumour cell growth and achieves strong anti-tumour efficacy in vivo.