A single-site mutant (M5) of native urokinase plasminogen activator (prouPA) induces effective thrombolysis in dogs with venous or arterial thrombosis with a reduction in bleeding complications compared to tPA. This effect, related to inhibition of two-chain M5 (tcM5) by plasma C1-inhibitor (C1I), thereby preventing non-specific plasmin generation, was augmented by the addition of exogenous C1I to plasma in vitro. In the present study, tPA, M5 or placebo +/- C1I were administered in two rat stroke models. In Part-I, permanent MCA occlusion was used to evaluate intracranial hemorrhage (ICH) by the thrombolytic regimens. In Part II, thromboembolic occlusion was used with thrombolysis administered 2 h later. Infarct and edema volumes, and ICH were determined at 24 h, and neuroscore pre (2 h) and post (24 h) treatment. In Part I, fatal ICH occurred in 57% of tPA and 75% of M5 rats. Adjunctive C1I reduced this to 25% and 17% respectively. Similarly, semiquantitation of ICH by neuropathological examination showed significantly less ICH in rats given adjunctive C1I compared with tPA or M5 alone. In Part-II, tPA, M5, and M5+C1I induced comparable ischemic volume reductions (>55%) compared with the saline or C1I controls, indicating the three treatments had a similar fibrinolytic effect. ICH was seen in 40% of tPA and 50% of M5 rats, with 1 death in the latter. Only 17% of the M5+C1I rats showed ICH, and the bleeding score in this group was significantly less than that in either the tPA or M5 group. The M5+C1I group had the best Benefit Index, calculated by dividing percent brain salvaged by the ICH visual score in each group. In conclusion, adjunctive C1I inhibited bleeding by M5, induced significant neuroscore improvement and had the best Benefit Index. The C1I did not compromise fibrinolysis by M5 in contrast with tPA, consistent with previous in vitro findings.