Can a broken barrier be repaired?

Abstract

Life threatening cerebral oedema can occur following stroke or as a consequence of a brain tumour, and although its resolution is critical there is as yet no agreed therapeutic strategy. Corticosteroids have been used with some success in treating tumour-generated oedema (Kaal & Vecht, 2004), but the results following stroke are equivocal at best (Poungvarin, 2004). This is disappointing as well as unexpected because several cell culture and whole animal models have shown that dexamethasone, or hydrocortisone can improve blood–brain barrier function, especially following an ischaemic insult. A paper by Forster et al. (2006) in this issue of The Journal of Physiology is the latest of these, and demonstrates elegantly that glucocorticoid treatment of mice results in increased occludin and claudin-5 expression in brain, but not heart, capillaries. This was mirrored by the presence of the glucocorticoid receptor in the brain vessels and its absence in those of the heart, and by their nuclear translocation in cerebral, but not myocardial, endothelial cell lines. Glucocorticoids, however, have numerous targets that are often referred to as side-effects, but are more likely to dominate the outcome of prolonged treatment with high doses. It should be instructive to consider the different mechanisms that can come into play when a glucocorticoid is administered, and the dose at which it is given. Glucocorticoid administration reduced vasoactive endothelial growth factor (VEGF) secretion from 9L glioma cells implanted in mouse brain (Heiss et al. 1996) and the cerebral oedema associated with the implant. These effects were also glucocorticoid receptor dependent. As VEGF is an important mediator of cerebral oedema following ischaemia–reperfusion via opening the blood–brain barrier (Paul et al. 2001) it is likely that dexamethasone improves the blood–brain barrier in the vicinity of the tumour by reducing VEGF production. Similarly, Forster et al. (2006) found that removing fetal calf serum, once the culture was confluent, resulted in barrier tightening. It is reasonable to suppose that this was likely to be due to removing the growth factors that are in the serum, especially VEGF. The water channel aquaporin-4 has been implicated in regulating brain water balance (Manley et al. 2004), and the absence of this water channel, which lies on the endothelial-facing membrane of astrocytes, affects the processes of resolving brain water overload. So oedema that results from water intoxication and early phase of focal ischaemia was much improved in mice that lacked this water channel, but they fared much worse from procedures that resulted in blood–brain barrier disruption. It seems that aquaporin-4 can form part of a protective mechanism, because it is rapidly up-regulated following trauma (Sun et al. 2003). Dexamethasone reduces aquaporin-4 expression when administered to adult sheep at human therapeutic doses (Ron et al. 2005) and may be part of an explanation for barrier enhancing effects being vitiated. Recent work from Martha O'Donnell's laboratory at UC Davis points to another factor that may contribute to the lack of effectiveness of glucocorticoid therapy. They have shown over the last few years that the Na-K-Cl cotransporter is activated in brain endothelium by low PO2 conditions (Foroutan et al. 2005). That this effect contributes to oedema that follows ischaemia was demonstrated by administering the cotransporter inhibitor bumetanide just before middle cerebral artery occlusion in rats: the resulting oedema and infarction were reduced (O'Donnell et al. 2004). Earlier work from that laboratory had shown the mechanism by which corticosteroid therapy can lead to glaucoma. The trabecular meshwork cells are richly supplied with the Na+–K+–Cl− cotransporter and low concentrations of dexamethasone resulted in greater transporter activity in cultured monolayers, and also transiently increased transporter protein expression (Putney et al. 1997). The therapeutic potential for any treatment that activates the glucocorticoid receptor depends on whether its activation is involved in either of the deleterious effects of dexamethasone treatment, and there is no information on this point, which brings us back to considering the key questions that have been raised by the work that Forster et al. have presented in this issue. What is it about the myocardial endothelium that renders its glucocorticoid receptor immune from the advances of dexamethasone, while its cousin in the brain responds so well? Do the cerebral and myocardial endothelial cells show differences in Na+–K+–Cl− contransporter activation and aquaporin-4 inhibition? If these can be answered there may be a new target for developing a therapeutic tool.