Activation of tissue factor (TF)-dependent coagulation is an early event in the pathogenesis of sepsis, responsible for microvascular thrombosis and consequent organ injury. We hypothesized that sepsis-induced TF procoagulant activity may be paralleled by a decreased expression or function of its natural inhibitor, tissue factor pathway inhibitor (TFPI). To test this hypothesis we used a non-human primate sepsis model in which adult baboons were infused with live E. coli at 109 CFU/kg (sublethal dose), 1010 CFU/kg (lethal dose) or saline (control group) and the animals were sacrified after 2, 8, or 24 hours. Lung tissue was snap frozen for protein and mRNA extraction or fixed and processed for light and electron microscopy. Tissue cryosections were immunostained using double/multiple fluorescence labeling approaches for TF, TFPI, factor VII/FVIIa and fibrin, in conjunction with cell markers for endothelial cells (CD31 or von Willebrand factor), leukocytes (PSGL-1), macrophages (CD68), PMN (myeloperoxidase) and platelets (gpIIb/IIIa). Large amounts of TF were detected in leukocytes, endothelial cells and platelet-rich microthrombi, starting from 2 hours and throughout the examined period. Concomitantly, confocal and electron microscopy analysis revealed increased leukocyte infiltration, platelet aggregates and fibrin deposition in the intravascular and interstitial compartments. In addition, TF induction was documented by semiquantitative RT-PCR, ELISA, western blot and factor Xa activation assays. Whereas TFPI mRNA showed only a modest increase, tissue-associated TFPI protein was found considerably decreased, especially during the first eight hours post E. coli infusion. Moreover, TFPI inhibitory activity of lung extracts from septic animals was 6–8 fold lower comparing to controls. TF activity measurements in the presence of inhibitory anti TFPI antibodies showed that only a very small fraction of endogenous TF was inhibited by tissue-associated TFPI, suggesting that most of the active TFPI available in the vascular compartment was depleted. The decrease of TFPI inhibitory potency cannot be exclusively explained by its proteolytic degradation, as we did not find significant amounts of truncated TFPI on western blots. In conclusion, our studies demonstrate that the exposure to septic and inflammatory stimuli lead to a decrease of TFPI-dependent endothelial anticoagulant potential, simultaneous with a strong TF-dependent procoagulant response. Activation of TF-dependent coagulation pathway not adequately countered by TFPI may have important roles in the pathogenesis of sepsis-associated disseminated intravascular coagulation. Strategies aimed to restore the physiological anticoagulant function of TFPI may help preventing sepsis-induced multiple organ dysfunction syndrome and death.