Chemokines have been convincingly implicated in actuating inflammatory leukocyte emigration. To affect the circulating leukocytes, tissue-derived chemokines have to traverse the endothelial cells (ECs). This was thought to be accomplished by chemokine diffusion through the intercellular gaps. On the contrary, we show by electron microscopy that the prototype chemokine IL-8 is internalized by venular ECs abluminally and transcytosed to the luminal surface. Here, it is presented to the adherent leukocytes on the EC membrane, predominantly in association with the EC projections. The intact C terminus of IL-8, the molecule's "immobilization" domain, is required for the EC binding, transcytosis, and consequently, the in vivo proemigratory activity of IL-8, indicating that the described subcellular interactions of IL-8 with the ECs are functionally relevant.

Chemokines have been convincingly implicated in driving leukocyte emigration in different inflammatory reactions. However, the cellular and molecular mechanisms of chemokine involvement in leukocyte emigration are not clear. We and others suggested that leukocyte adhesion to the endothelium and transmigration are induced by chemokines immobilized on the endothelial cell surface. This would require the presence of specific chemokine binding sites in this microanatomical location. Using an in situ binding assay we demonstrated the presence of binding sites for interleukin-8 (IL-8) and RANTES, but not monocyte inflammatory protein-1 alpha on the endothelium of postcapillary venules and small veins in human skin. In contrast, venules and veins in various anatomical locations showed dramatically differing IL-8 binding patterns. The subcellular distribution of IL-8 in the venular endothelial cells following its in vivo and ex vivo injections was studied by use of electron microscopy. Our results suggest that IL-8 was internalized by the endothelial cells, transported transcellularly via plasmalemmal vesicles, and released onto the luminal surface where it appeared located preferentially on tips of membrane protrusions. We were unable to study the endothelial IL-8 binding or transport in vitro because all the in vitro propagated endothelial cell lines and primary endothelial cells tested lacked IL-8 binding sites. This includes human umbilical vein endothelial cells (HUVECs), which also did not bind IL-8 in situ. However, HUVECs provided a satisfactory in vitro system to study the secretion of IL-8 by the endothelial cells. Two possible alternative pathways were described: secretion directly from the Golgi apparatus or following storage in Weibel-Palade bodies.

No abstract