FXII is an emerging target for thrombosis, yet several questions remain to be addressed. Firstly, from drug discovery perspective, level of enzyme occupancy needed for efficacy (which largely dictates potency and selectivity requirement for small molecule inhibitors) is unclear, as most reported active site inhibitors have some level of off-target activities. Secondly, from disease treatment perspective, it is unclear whether FXIIa inhibition will be a safe strategy for stroke treatment or prevention, as it was recently reported that FXIIa inhibition destabilized the subocclusive thrombi in a plaque rupture model. In this presentation, we set out to address these questions using a previously described molecule, Inf4mut15. We generated the human albumin (HA)-Inf4mut15 fusion protein (Mut-inf) for our studies. In vitro Mut-inf displayed comparable potency as the widely used wild-type HA-Infestin4 (WT-inf) (human FXIIa Ki = 73 and 120 pM, respectively). Both infs acted as competitive reversible active site inhibitors of FXIIa, with no binding to FXII zymogen, hence same mode of action as certain small molecule inhibitors. Mut-inf, however, was much more selective against plasmin compared to WT-inf (20,000- and 75-fold Ki separation, respectively), consistent with results from the functional tPA-induced TEG assay, where Ly60 was reduced dose-dependently by WT- but not Mut-inf. Mut-inf aPTT doubling concentration was 15 uM and FXIIa Ki in 30% plasma was 3.5 nM. Calculated enzyme occupancy for Mut-inf for doubling human aPTT is thus 99.9%. In the rabbit model of cerebral microembolic signals (MES) induced by FeCl 3 injury of the carotid artery, treatment with vehicle (n=7), WT-, and Mut-inf (1mg/kg and n=5 each) produced arterial thrombus of 6.0±0.4, 1.9±0.6, and 0.2±0.1 mg, respectively; incidence of MES detected in the middle cerebral artery was 4.1±1.3, 1.8±0.6, and 0.0±0.0, respectively. In summary, our studies demonstrated that very high enzyme occupancy will be required for achieving a putative aPTT doubling target in human for FXIIa active site inhibitors, highlighting the challenge with the small molecule modality. Our MES studies suggest that targeting FXII may offer a safe strategy for stroke prevention and/or other thromboembolic disorders.