Severe sepsis is normally caused by bacterial infection, and can result in organ failure. Increased concentrations of inflammatory cytokines contribute to vascular and organ damage. Leucocytes are activated and adhere to the vessel wall. Leucocyte clumping contributes to impaired blood flow and oxidant injury to blood vessels. Blood pressure drops. Vascular permeability is increased, leading to oedema. In severe cases, there is disseminated intravascular coagulation (figure). Most patients who die from sepsis have low concentrations of anticoagulants protein C and antithrombin, indicating an active coagulation process even in the abscence of disseminated intravascular coagulation. Therefore, replacement therapy with these proteins might be effective. Many challenge this strategy, however, since patients who die, generally do not have extensive thrombosis at necropsy. However, an understanding of the protein C pathway, through animal experiments, provides a basis for this strategy. Protein C is activated by the thrombin-thrombomodulin complex on the endothelium. The thrombomodulin concentration, and hence the rate of protein C activation, is higher in the microvasculature than in the macrovasculature because of a high endothelial cell-to-blood volume ratio. Activated protein C (APC) inactivates factors Va and VIIIa inhibiting thrombin generation. Hinshaw and Taylor showed that dogs placed on an extracorporeal pump developed a circulating anticoagulant, APC, probably caused by thrombin generated by the pump. When these animals were challenged with a lethal dose of endotoxin, they survived. Seemingly paradoxically, we found that thrombin infusion before infusion of a lethal dose of endotoxin could also protect animals from severe sepsis. APC was then shown to protect baboons from sepsis. APC not only blocked the coagulant response, but also reduced the concentrations of various inflammatory markers, minimised the drop in blood pressure, and was protective when given hours after infusion of bacteria. Thrombin activation of protein C therefore probably plays a part in defense against severe sepsis. Studies from Hancock’s laboratory provided evidence for a cell surface receptor for APC on monocytes. Engagement of this receptor blocks endotoxin mediated calcium fluxes in the cell and prevents CD14 down-regulation. Thus, APC attenuates over-expression of some of the cytokines that contribute to organ dysfunction. Joyce and colleagues have shown that APC decreases expression of adhesion molecules responsible for leucocyte interaction with the endothelium. APC also altered gene expression in a pattern that would prevent apoptosis. The prevention of leucocyte adhesion to the endothelium might decrease endothelial cell injury in severe sepsis. Prevention of apoptosis would also decrease the endothelial cell dysfunction seen in the disease. The endothelial cell protein C receptor might be involved in cellular signalling. The receptor translocates from the plasma membrane into the nucleus and can carry APC in the process. Nuclear translocation seems to modulate gene expression, but the relative importance of the contributions of nuclear translocation and cell surface signalling to the protective functions of APC on the endothelium are unknown. Investigation of the protein C pathway has identified possible new treatments for severe sepsis, and combination therapies for sepsis might be possible. These include the addition of anticytokines, which by themselves have little or no protective effects, or inhibitors of platelet activating factor, which have shown promise in clinical studies. Since these agents work at different stages of the disease, combination therapies might improve survival without increased risk. Severe sepsis induces release of cytokines, oxidants, and proteases from leucocytes, injuring the vessel, triggering clotting, and inhibiting the protein C pathway (A). APC limits many of these events, protecting the vessels and organs from injury (B).