Cerebral blood flow was measured and compared in 10 symmetrical brain regions following unilateral trigeminal ganglionectomy (n = 13), sham operation (n = 6), or trigeminal root section (rhizotomy) (n = 8) in cats. Multiple determinations were obtained in anesthetized and paralyzed animals using radiolabeled mict-ospheres during (i) normocapnia-normotension, (ii) hypercappia (5% C02/95% room air), A third major pathway was recently defined innervating cerebral blood vessels (7). Perivascular sensory axons projecting from the trigeminal ganglia were demonstrated by retrograde (7-9) and orthograde (10) axonal transport techniques as well as by Wallerian degeneration techniques (11) or by surgical lesioning combined with immunohistochemistry (12, 13) or radioimmunoassay (14, 15). Fibers primarily from the ophthalmic division (7, 8) join sympathetic and parasympathetic fibers within the cavernous plexus (11) from which they distribute to the ipsilateral internal carotid artery and the circle of Willis. Trigeminovascular fibers are mainly unmyelinated (16), reside within the adventitial layer, transmit nociceptive information (17, 18), and contain substance P (12), neurokinin A (19), calcitonin gene-related peptide (13), and cholecystokinin 8 (20).In this report, we examined the possibility that sensory axons participate in the regulation of cerebral blood flow based on knowledge that (i) substance P (21, 22), neurokinin A (23), and calcitonin gene-related peptide (13) are cerebrovasodilating molecules; (ii) peptides are released from pial vessels in vitro by depolarizing stimuli (24); and (iii) release mediates vasodilation and/or increases in blood flow following electrical stimulation of sensory nerves in other tissues (25-27). Blood flow was examined during perturbations previously shown to involve nonneurogenic (hypercapnia) and neurogenic mechanisms (at least in part). We conclude that increases in brain blood flow during severe hypertension are not totally passive to the perfusion pressure. Furthermore, increases in flow associated with generalized seizures or hypertension are mediated in part by the trigeminal nerve via local "axon reflex-like" mechanisms involving neuropeptide release from perivascular nerve endings.
MATERIALS AND METHODSSham and Lesioning Procedures. Male and female adult mongrel cats (3.0-4.5 kg) were anesthetized with ketamine hydrochloride (30 mg/kg, i.m.) and sodium pentobarbital (10-15 mg/kg, i.p.) and mechanically respirated with halothane anesthesia (0.5-2.0%) as needed. Animals were subjected to unilateral trigeminalectomy as described (15). In this procedure, trigeminal ganglia cell bodies are destroyed and perivascular fibers degenerate accordingly. Shamoperated animals were subjected to craniectomy, exposure of the trigeminal ganglia, as well as an equivalent amount of brain retraction as required for the ganglionectomy.The trigeminal root (that section between the ganglion and the brainstem) was approached by a left suboccipital craniectomy, which extended laterally and ...