Inflammation is characterized by the local accumulation of blood leukocytes, plasma proteins, and fluid, usually at an extravascular site of injury, infection, or antigenic stimulation. Under normal circumstances, inflammatory reactions are an essential part of host defense, playing a critical role in the eradication of infectious agents. However, inflammatory processes are intrinsically destructive to the surrounding tissues, and they can, in certain circumstances, do far more harm than good. For the transplant biologist, allograft rejection is a clear example of inflammation gone awry, and new strategies for the management of rejection may develop from an understanding of the mechanisms of inflammation. The purpose of this overview is to recount certain new insights as to how inflammatory reactions develop. In particular, we will highlight several ways in which vascular endothelial cells may actively participate in these processes and suggest that endothelium is a potential target for therapeutic intervention. Hemodynamic effects in inflammation. In Cohnheim's classical descriptions of acute inflammation, the first observable events of the inflammatory process were hemodynamic changes (1). Vessels at a site of incipient inflammation become exten-sively dilated (although in some systems, vasodilatation may be preceded by transient vasoconstriction). As a consequence of local vasodilatation, there is increased blood flow and deliv-ery of increased numbers of leukocytes to the tissue site. Local injection of vasodilators, which are not in themselves intrinsically inflammatory, markedly enhance both leukocyte accumulation and vascular leakage when combined with neutrophil chemoattractants such as the peptides f-met-leu-phe or C5a (2), Indeed, chemoattractants in the absence of vasodilators are quite limited in their ability to cause inflammation, sug-gesting that inflammation is in part dependent upon vasodila-tation.

Vasodilatation results from relaxation of vascular smooth muscle cell tone. Some substances, such as nitrates, directly relax vascular smooth muscle cells. However, most vasodilators are now known to act indirectly, stimulating vascular endothelial cells to release mediators that, in turn, relax underlying vascular smooth muscle cells (3). Two well-described endothe-lial factors are thought to be particularly important in vasodilatation. The first of these is prostacyclin, also known as PGI. This arachidonic acid metabolite, produced through the cyclớoxygenase pathway, is synthesized and secreted by endothelial cells in response to agonists such as thrombin, histamine, or leukotriene C4 (LTC).* These agonists act on endothelial