Both EC types showed a constitutive release of sJAM-A into culture supernatants which was enhanced by pro-inflammatory stimulation. These in vitro data supported the blood level studies suggesting sJAM-A as a biomarker of vascular inflammation. In HUVEC, sJAM-A was found to be shedded from the cell surface by the disintegrin and metalloproteinases 10 and 17 upon pro-inflammatory stimulation. On a functional level, recombinant sJAM-A reduced adhesion of mononuclear cells to cultured HUVEC and HDMEC. Furthermore it reduced chemokine-triggered endothelial transmigration of CD4 + CD45RO + memory T cells across HUVEC under static and flow conditions. Finally, recombinant sJAM-A inhibited neutrophil extravasation in vivo in an air pouch model of vascular inflammation. Based on these results, it was suggested that sJAM-A might limit leukocyte extravasation at sites of vascular inflammation. However, it was also noted that all published experiments demonstrating a reduction of leukocyte extravasation by recombinant sJAM-A used concentrations too high to allow final conclusions about the pathophysiological function of sJAM-A in vivo. To date, the in vivo function of sJAM-A has not finally been addressed. To protect CNS homeostasis, EC forming the blood-brain barrier are highly specialized and differ in many molecular aspects from other EC types such as HUVEC. A potential role of JAM-A at the BBB under inflammatory conditions was indicated by a study reporting reduced brain endothelial JAM-A immune staining in the rat cortical cold injury model. Reduced endothelial JAM-A immune staining was also observed in active brain lesions of patients with multiple sclerosis. No major difference of brain EC JAM-A expression levels was noted between mice with cytokine-induced meningitis and sham-treated animals. The expression of sJAM-A at the BBB has not been investigated so far. Here we studied the regulation of JAM-A expression in cultured human brain microvascular EC and evaluated sJAM-A as a serum marker of BBB breakdown. While inflammatory stimulation of cultured HBMEC induced a subcellular redistribution of cell-bound JAM-A as previously described in other EC types, a detected constitutive release of sJAM-A was unexpectedly neither enhanced by pro-inflammatory nor by hypoxia/reoxygenation AZ 960 JAK inhibitor challenge. Accordingly, sJAM-A serum levels did not indicate a BBB breakdown in vivo. These results define a novel feature of human brain EC, distinguishing them from other EC types and possibly contributing to their particular anergy in the protection of CNS homeostasis. The major finding of this study is that in contrast to previously reported EC types, cultured human brain EC do not show an inducible release of sJAM-A under pro-inflammatory conditions.