Cold-injury of the cerebral cortex: immunolocalization of cellular proteins and blood-brain barrier permeability studies.

The occurrence of blood-brain barrier (BBB) permeability alterations and neovascularization are well documented in the cerebral cortical cold-injury model. This model was used to determine whether the glucose transporter (glutI) protein was present in endothelium of cerebral vessels with breakdown of BBB to protein and when regenerating endothelial cells become immunoreactive for glutI protein. Secondly, the protein products of c-fos and c-jun were localized to determine whether these early immediate genes are activated in this model. Observations were made over a period of 12 hours to 14 days after the cold-injury. Blood-brain barrier permeability was assessed using horseradish peroxidase (HRP) as a tracer. Since HRP may not be able to enter thrombosed vessels within the cold lesion, immunohistochemistry was used to detect extravasation of endogenous serum proteins using antisera to rat serum proteins. The proteins-glut1, GFAP, c-fos and c-jun-were localized by immunohistochemistry. Endothelium of vessels which were permeable to protein, whether in the cold-injury site or in the perilesional area, all contained glut1 protein; hence, the presence of glut1 did not appear to correlate with an intact BBB to protein. An interesting point is that in the process of neovascularization, regenerating endothelial cells become immunoreactive for glut1 at 5 days and this coincides with the presence of tight junctions in these cells. Immunoreactivity for c-fos was observed in regenerating endothelium within the lesion site, in astrocytes, and to a lesser extent in endothelial cells and neurons in the perilesional area. Few astrocytes showed immunoreactivity for c-jun at 4 and 5 days. Possibly, the growth factors generated to promote angiogenesis and repair led to activation of the c-fos gene with deposition of c-fos protein. The results suggest that during nervous system development or endothelial regeneration, the presence of glut1 in cerebral endothelium coincides with the presence of an intact BBB to protein and protein tracers. However, in pathological states presence of glut1 in cerebral endothelium does not appear to correlate with an intact BBB to protein. This model lends itself to the study of angiogenesis and repair processes in the cerebral cortex in an environment unaffected by ischemia and thus the findings may be relevant to traumatic injuries of the human cerebral cortex.