Effect of adenovirus-mediated nitric oxide synthase gene transfer on vasospasm after experimental subarachnoid hemorrhage.

Objective: Evidence indicates that vasospasm after subarachnoid hemorrhage (SAH) is caused in part by a decrease in the vasodilator nitric oxide (NO), which is produced mainly in endothelial cells. This study tested whether intracisternal injection of adenovirus-expressing endothelial NO synthase (eNOS) would decrease vasospasm in dogs.

Methods: In 12 dogs, baseline cerebral angiography was performed, and then SAH was produced by two injections of blood into the cisterna magna. The dogs were randomized (n = 6/group) to intracisternal injection of adenovirus-expressing lacZ (Ad327beta-Gal) or eNOS (AdCD8-NOS), administered immediately after the first blood injection. Angiography was repeated on Day 7, and then L-arginine (50 mg) was administered intracisternally, and angiography was repeated. Cerebrospinal fluid aspirated from the cisterna magna on Days 2 and 7 was analyzed for levels of NO metabolites. The dogs were killed, and their basilar arteries were removed and studied pharmacologically. Four control dogs underwent angiography on Day 0, followed by virus injection (n = 2/group). Angiography was repeated on Day 7, and the control dogs were killed. Transgene expression was detected in tissue removed on Day 7 by histochemical staining for lacZ, by polymerase chain reaction for messenger ribonucleic acid for eNOS, and by measurement of NO metabolites in cerebrospinal fluid.

Results: Angiography showed significant vasospasm in each group (Ad327beta-Gal, -54 +/- 7% reduction in basilar artery diameter; AdCD8-NOS, -53 +/- 7%), with no significant difference between groups. Injection of L-arginine caused an insignificant increase in arterial diameter in each group. In dogs without SAH, Ad327beta-Gal caused a reduction in basilar artery diameter (-13 +/- 10%, P = 0.42; paired t test), whereas injection of AdCD8-NOS caused an increase in diameter (14 +/- 16%, P = 0.77; paired t test). Histological examination and beta-galactosidase staining of dogs given injections of Ad327beta-Gal showed staining in inflammatory cells in the subarachnoid space, in the adventitia of the cerebral vessels, and in the liver and lungs. Messenger ribonucleic acid for eNOS was detected in the leptomeninges of dogs given injections of AdCD8-NOS. Under isometric tension, basilar arteries from each group demonstrated similar relaxation to L-arginine, but arteries exposed to eNOS demonstrated significantly greater relaxation to L-arginine plus tetrahydrobiopterin than arteries exposed to lacZ. Cerebrospinal fluid levels of NO and its metabolites were significantly higher in dogs treated with AdCD8-NOS than those treated with Ad327beta-Gal 2 days after SAH.

Conclusion: These results demonstrate that adenovirus vectors can be used to transfer genes to cells in the subarachnoid space of dogs. Enough NO can be produced in the absence of SAH to dilate the basilar artery. After SAH, however, NO plus a cofactor can dilate arteries in vitro, but not enough NO is generated in the subarachnoid space to prevent vasospasm, perhaps owing to the scavenging of NO by hemoglobin.