While superoxide dismutase (SOD) may be useful in treating inflammation, the problems of getting it into the blood in the right concentration, for long enough periods, and to the intended organ, have limited its translation into human clinical medicine. None of the three naturally occurring forms of human SOD is well suited for use as a therapeutic agent. SOD1 and SOD2 are normally intracellular enzymes and are rapidly cleared by the kidney. SOD3 occurs outside cells, but binds so tightly to cell surfaces or to collagen fibrils in the intracellular matrix that it remains largely in the few organs that secrete it. The "stickiness" of SOD3 results from a positively charged region in the hydrophilic C-terminus of each subunit. We have genetically engineered a hybrid chimeric SOD called SOD2/3 with greatly improved pharmacological properties. It has the sequence encoding the mature human SOD2 fused to the C-terminus of human SOD3. This hybrid SOD2/3 is highly expressed and easily purified. The molecule binds to endothelial cells, but less tightly than SOD3, and circulates well enough to become widely attached to extracellular surfaces, presumably in many tissues. The loose binding appears to produce a buffering effect on enzyme concentration, effectively eliminating bell-shaped dose-response curves. Single IV injections of SOD2/3 have protected experimental animals against a variety of models involving inflammation or ischemia/reperfusion.