Endostatin can inhibit angiogenesis and tumor growth in mice. A potential limitation of endostatin as an antitumor agent in humans is the short serum half-life of the protein that may decrease effective concentration at the site of tumor and necessitate frequent dosing. In an effort to improve antitumor activity, endostatin was fused to an antibody specific for the tumor-selective HER2 antigen to create an antibody-endostatin fusion protein (anti-HER2 IgG3-endostatin). Normal endostatin rapidly cleared from serum in mice (T(1/2)(2), = 0.6-3.8 hours), whereas anti-HER2 IgG3-endostatin had a prolonged half-life (90% intact; T(1/2)(2), 40.2-44.0 hours). Antigen-specific targeting of anti-HER2 IgG3-endostatin was evaluated in BALB/c mice implanted with CT26 tumors or CT26 tumors engineered to express the HER2 antigen (CT26-HER2). Radio-iodinated anti-HER2 IgG3-endostatin preferentially localized to CT26-HER2 tumors relative to CT26 tumors. Administration of anti-HER2 IgG3-endostatin to mice showed preferential inhibition of CT26-HER2 tumor growth compared with CT26. Anti-HER2 IgG3-endostatin also markedly inhibited the growth of human breast cancer SK-BR-3 xenografts in severe combined immunodeficient mice. Anti-HER2 IgG3-endostatin inhibited tumor growth significantly more effectively than endostatin, anti-HER2 IgG3 antibody, or the combination of antibody and endostatin. CT26-HER2 tumors treated with the endostatin fusion protein had decreased blood vessel density and branching compared with untreated CT26-HER2 or CT26 treated with the fusion protein. The enhanced effectiveness of anti-HER2 IgG3-endostatin may be due to a longer half-life, improved serum stability, and selective targeting of endostatin to tumors, resulting in decreased angiogenesis. Linking of an antiangiogenic protein, such as endostatin, to a targeting antibody represents a promising and versatile approach to antitumor therapy.