Effects of Noninvasive Facial Nerve Stimulation in the Dog Middle Cerebral Artery Occlusion Model of Ischemic Stroke
Background and Purpose—Facial nerve stimulation has been proposed as a new treatment of ischemic stroke because autonomic components of the nerve dilate cerebral arteries and increase cerebral blood flow when activated. A noninvasive facial nerve stimulator device based on pulsed magnetic stimulation was tested in a dog middle cerebral artery occlusion model.
Methods—We used an ischemic stroke dog model involving injection of autologous blood clot into the internal carotid artery that reliably embolizes to the middle cerebral artery. Thirty minutes after middle cerebral artery occlusion, the geniculate ganglion region of the facial nerve was stimulated for 5 minutes. Brain perfusion was measured using gadolinium-enhanced contrast MRI, and ATP and total phosphate levels were measured using 31P spectroscopy. Separately, a dog model of brain hemorrhage involving puncture of the intracranial internal carotid artery served as an initial examination of facial nerve stimulation safety.
Results—Facial nerve stimulation caused a significant improvement in perfusion in the hemisphere affected by ischemic stroke and a reduction in ischemic core volume in comparison to sham stimulation control. The ATP/total phosphate ratio showed a large decrease poststroke in the control group versus a normal level in the stimulation group. The same stimulation administered to dogs with brain hemorrhage did not cause hematoma enlargement.
Conclusions—These results support the development and evaluation of a noninvasive facial nerve stimulator device as a treatment of ischemic stroke.
The currently available emergency treatments for ischemic stroke focus on removing the occlusive blood clot. However, another means for improving cerebral blood flow (CBF) in ischemic stroke is available, namely, dilating the cerebral arteries. Numerous animal studies have demonstrated that facial nerve stimulation with electric current dilates cerebral arteries and increases CBF (reviewed in Borsody et al1). Additional studies have shown that electric stimulation of autonomic branches of the facial nerve improves ischemic stroke measures2,3 and that lesions of those branches worsen ischemic stroke severity.4,5 However, accessing the facial nerve in the clinical setting so as to apply electric current to it would be difficult.
Previously, we demonstrated that noninvasive stimulation of the facial nerve trunk with magnetic energy in normal dogs and sheep increases CBF and causes vasodilation of cerebral arteries.1 Herein, we report on the effect of such stimulation in a clinically relevant dog ischemic stroke model. Additionally, we assess the effect of the same stimulation administered to dogs with brain hemorrhage caused by puncture of the intracranial internal carotid artery (ICA) as a test of safety of this potential treatment.
Ethical Approval and Facial Nerve Stimulation
All experimental procedures were approved by the Ethics Committee of the Universidad Autónoma Metropolitana of Mexico City. Facial nerve stimulation was performed with a modified transcranial magnetic stimulator (MagPro R30; MagVenture, Atlanta, GA) equipped with a fluid-cooled 6.5 cm figure-8 stimulation coil. Stimulation power was set at 1.8 T at the coil surface. Stimulation was administered with 280 μs biphasic pulses delivered at 10 Hz for a 5-minute period in all experiments. Placement of the stimulation coil was performed as described in Figure I in the online-only Data Supplement.
Ischemic Stroke Experiments
A total of 12 adult mongrel dogs weighing 15 to 37 kg were kept with ad libitum access to food and water until 8 hours before the experimental procedure, at which time they were food-restricted. Dogs were randomly assigned to either stimulation (n=6) or control (n=6) groups. Anesthesia was induced with intramuscular injection of Zoletil (tiletamine/zolazepam, 1:1; 7 mg/kg), propofol (2.5 mg/kg), and fentanyl (2 μg/kg), and anesthesia was maintained with propofol (10 mg/kg per hour). Dogs were intubated after induction and mechanically ventilated throughout the experiment. Vital sign monitoring included heart rate, blood pressure, and arterial blood gas measurement. Ventillation was titrated according to oxygen saturation and end-tidal CO2 levels.
Middle cerebral artery (MCA) occlusion was induced in all 12 animals by the injection of an autologous blood clot into an ICA through an endovascular catheter under digital subtraction angiography. After clot injection, MCA occlusion was confirmed by digital subtraction angiography, MR angiography, and loss of CBF in the MCA distribution on baseline imaging. In 1 dog allocated to the stimulation group, no loss of CBF was identified despite MCA occlusion on digital subtraction angiography and MR angiography; this dog was excluded from further analysis. At the end of each experiment, dogs were euthanized by intravenous injection of potassium chloride under general anesthesia.
Detailed descriptions of MRI protocols are available in the online-only Data Supplement. Tissue perfusion was assessed using standard blood flow thresholds for large animal stroke models.6,7 CBF and tissue perfusion were calculated by computer algorithm; blinding of images was not possible because of pronounced flush of blood to the extracranial tissues that identified stimulated animals (Figure 1).1 Average normalized perfusion index, ischemic core volumes, and ATP/total phosphate ratio of the stimulation group were compared with those of the control group with repeated-measures ANOVA (NCSS software package), using data sets beginning at the poststroke time point. Based on the expectation of increasing CBF with facial nerve stimulation, 1-sided tests were used.
Brain Hemorrhage Experiments
Three dogs as described above were used in these experiments. Intracranial hemorrhage was induced by puncture of the ICA with a Touhy needle advanced through a frontal craniotomy under neuronavigation guidance. On successful rupture of the ICA, intracranial hematoma was monitored with T1 and T2 MRI every 15 minutes until it was judged stable by the study staff on 3 consecutive imaging studies, after which facial nerve stimulation was delivered as described above. Follow-up T1 and T2 MRI and perfusion imaging were then performed immediately (t=0), 30 minutes, and 60 minutes poststimulation. Hematoma size as demonstrated on T1 and T2 imaging was judged qualitatively by the study staff that included a neurologist (M.K.B.) and a neurosurgeon (F.C.P.).
Effect of Facial Nerve Stimulation in Dogs With Ischemic Stroke
Five minutes of facial nerve stimulation was administered beginning 30 minutes after confirmation of MCA occlusion by angiography and poststroke MRI evaluation. Facial nerve stimulation with pulsed magnetic energy did not cause nystagmus in any animal during or after stimulation, nor was salivation or lacrimation manifestly increased in stimulated animals. Facial nerve stimulation did not affect vital signs (Table I in the online-only Data Supplement).
Figure 1 shows an example of perfusion images in representative dogs from both stimulation and control groups. CBF as measured by perfusion index was increased in the region of ischemia when stimulation was administered 30 minutes after MCA occlusion. The effect of facial nerve stimulation was durable, lasting ≥90 minutes poststimulation. No such increase in CBF in ischemic brain region was observed in control animals.
In all dogs receiving facial nerve stimulation, blood flow to extracranial tissues was also increased, although this effect was only observed immediately poststimulation (t=0). No movement of mastication muscles was observed during stimulation, indicating that the activation of trigeminal motor system did not account for increased blood flow to or through extracranial tissues.
Figure 2 shows the group average effect of facial nerve stimulation on CBF (perfusion index), comparing stimulation and control groups. Average CBF was decreased to ≈70% of baseline levels in the ischemic hemisphere region of interest, and perfusion stayed at those depressed levels in the control group, whereas it was returned to normal by facial nerve stimulation (P<0.01). Improvement in CBF after facial nerve stimulation was found to be durable for ≥90 minutes poststimulation. CBF was not reduced in surrounding brain regions after stimulation, including nonischemic frontal, parietal, and occipital regions both in ipsilateral and contralateral hemispheres (Figure II in the online-only Data Supplement).
Facial nerve stimulation also reduced ischemic core volume (Figure 3). The reduction in ischemic core volume after facial nerve stimulation did not come at the expense of an enlargement of penumbral volume, and the size of ischemic core was statistically smaller in the stimulation group in comparison to the control group (P<0.01) that showed an enlargement of ischemic core volume over time. By the 90-minute time point, however, it appeared that ischemic core volume began to increase in the stimulation group.
Figures 2 and 3 show data from 11 dogs that had MRI measures taken out to 90 minutes poststimulation. Data from a larger group of 18 dogs, which includes the aforementioned 11 dogs, that had MRI measures out to 60 minutes poststimulation are shown in Figures III and IV in the online-only Data Supplement. The larger data set demonstrates the same responsiveness to facial nerve stimulation as presented above.
As shown in Figure 4, the ATP/total phosphate ratio was similar in both groups before and after stroke onset, but the ratio showed a significant drop in the control group after 60 minutes post–sham stimulation, whereas in the stimulation group, the ratio at that time was comparable if not greater than baseline (P<0.05).
Effect of Facial Nerve Stimulation in Dogs With Brain Hemorrhage
Puncture of intracranial ICA with a Touhy needle did not kill any of the 3 dogs subjected to the procedure, and it produced a mixture of subarachnoid hemorrhage with intraventricular and intraparenchymal extension. Using the same parameters administered in ischemic stroke experiments, facial nerve stimulation ipsilateral to arterial puncture in dogs with stable intracranial hematomas did not kill any of the dogs, nor did it cause gross enlargement of the hematoma (Figure 5). CBF did not seem to change from baseline in this small group of dogs, and in fact, a large decrease in blood flow to extracranial tissues was observed (compare bottom panels of Figures 5 with immediately post-stimulation panel in Figure 1A).
We used the dog model of embolic MCA occlusion with autologous blood clots in our experiments because it resembles the clinical condition of human embolic stroke. Autologous blood clot was injected into distal ICA from where it was carried by the blood flow into cerebral arteries; typically, the clot lodges in the MCA, but the precise site of occlusion cannot be controlled.8,9 As a result, the degree of loss of CBF, involvement of anterior communicating and cerebral arteries with MCA occlusion, and size and location of the infarct can exhibit considerable variability among the dogs, which is similar to the clinical reality of ischemic stroke. However, we did not experience much of this variability. Instead, we noted the failure of complete MCA occlusion to result in any region of reduced CBF on contrast-enhanced perfusion MRI in 1 dog—the likely effect of robust collateralization around the proximal MCA in this species.10
Despite the variability of the model, the effect of noninvasive magnetic stimulation of facial nerve was evident and statistically significant with only 11 experimental subjects. We demonstrate that facial nerve stimulation increased CBF when administered 30 minutes after MCA occlusion, and the CBF response outlasted the 5-minute stimulation period by ≥90 minutes after cessation of stimulation. The improvement in CBF in the region of interest in ischemic hemisphere was not attributable to a steal of blood from surrounding brain regions, which, to the contrary, exhibited a small degree of increased CBF themselves, consistent with a diffuse vasodilation of arteries of anterior circulation in response to facial nerve stimulation.11
The improvement in CBF after facial nerve stimulation translated into a reduction in ischemic core volume. Over a period of several hours, the volume of ischemic core is expected to increase as tissue in the penumbral zone succumbs to prolonged ischemia,12,13 but the acute time frame of our experiments may not have allowed the full area of hypoperfusion to manifest because both ischemic core and penumbra volumes increased in the control group. Indeed, within the 90-minute post–sham stimulation time frame in the control group, the total area of hypoperfusion seems to be reaching a plateau. In comparison, the group subjected to facial nerve stimulation demonstrated a significant reduction in ischemic core volume. This change may reflect the shifting of tissue satisfying an ischemic core definition to slightly improved perfusion, leading it to be classified as ischemic penumbra, and similarly the shifting of ischemic penumbra to normal perfusion; such a change would be consistent with a global improvement in perfusion or with a robust network of collateralization that can feed an ischemic brain region from multiple sides. The latter, in fact, seems to be available to the MCA in dogs.10
Tissue perfusion measures further suggest that the effect of facial nerve stimulation may be temporary, because the volume of ischemic core began to increase in the stimulation group 90 minutes after stimulation. This may reflect reocclusion of the MCA after gradual loss of a vasodilatory drive from the facial nerve, although the CBF response did not seem to weaken at that time point. However, tissue perfusion in ischemic core may not have been fully captured by the hemispheric region of interest used to measure CBF, allowing the 2 measures to be in discordance. If vasodilation in affected portions of the MCA were to weaken faster than in unaffected arteries of the brain, reinstitution of ischemic core could occur while CBF outside of ischemic core region remains elevated.
But we do not view this data and question why the effect of facial nerve stimulation only lasts between 60 and 90 minutes; rather, we are curious about how the effect of 5 minutes of facial nerve stimulation persists for as long as it did in our experiments. In normal (nonstroke) animals, Goadsby14 applied direct electric stimulation to the facial nerve in posterior fossa, and by applying 1 minute of continuous stimulation at 10 Hz, the increase in CBF began to return to baseline immediately after cessation of stimulation. Inconsistent evidence exists that facial nerve stimulation may have longer effects in animals with ischemic stroke than in normal animals. In 1 study of rats with permanent MCA occlusion published only in abstract form,3 direct electric stimulation of sphenopalatine ganglion (which receives some of the vasoactive petrosal branches from the facial nerve) for a total of 10 minutes was reported to increase CBF by >40% for ≥24 hours. However, in another study in the same rat model, a total of 8 minutes of otherwise comparable sphenopalatine ganglion stimulation produced a notably smaller increase in CBF that returned to baseline within a matter of few minutes. In our experiments with noninvasive magnetic facial nerve stimulation, we observed increases in CBF lasting 20 to 30 minutes after a 5-minute stimulation period in normal sheep and dogs that were dependent on targeting the geniculate ganglion region of the facial nerve.1 We think that the longer effect of stimulation seen in our ischemic stroke experiments may reflect either persistent vasodilation or dissolution of the blood clot, and indeed, vasodilation may facilitate breakdown of blood clot by the physical force of partially restored blood flow.15 Spontaneous lysis of occluding clot in some dogs, but not all, might explain the suddenly increased variability observed in ischemic core volumes in dogs 90 minutes after stimulation. These questions will be addressed by upcoming experiments.
The beneficial effects of facial nerve stimulation on CBF and tissue perfusion are further corroborated by the ATP/total phosphate ratio, a commonly used measure of energy state of brain tissue.16 In contrast to the decrease in ATP/total phosphate ratio in the control group, which implies loss of critical energy reserves, no such decrease was observed after stimulation of the facial nerve. In fact, a small increase in the ratio was observed. Because we see no parallels for this overshoot in the scientific literature, we think it represents nothing more than the variability in the measure. However, we must recognize that a significant volume of brain was still ischemic core at the time of the ATP/total phosphate ratio overshoot, and the only way to explain the coincidental observations would be to presume the overshoot to be real if not underestimating the energy state of noncore brain. The preservation of ATP stores is consistent with improvement of CBF and tissue perfusion that reduced the volume of ischemic core.17
We suggest that a facial nerve stimulator using pulsed magnetic energy directed at the facial nerve in the temporal bone would be safe and tolerable when used in humans. Allowing for its application in anesthetized animals in this report, facial nerve stimulation did not seem to have many of the potential adverse effects that would be expected based on the site of stimulation. No nystagmus nor excessive salivation or lacrimation were noted during or after stimulation, an observation confirmed in our normal animal experiments.1 We consider the absence of nystagmus an important finding, because the placement of stimulation coil over the ear interposes the vestibular apparatus of the inner ear in the pathway of magnetic field. Other aspects of neurological examination were not available in anesthetized, nonsurvival dog subjects, notably hearing, because the auditory structures of inner ear are similarly in the path of magnetic field. However, previous experiments1 did recover several dogs from the experimental procedure that had involved facial nerve stimulation with comparable parameters, and in those cases, no evidence of neurological impairment was observed by the study veterinarian. Furthermore, we have conducted first-in-man experiments in an unsedated healthy volunteer and found no vertigo, nystagmus, tinnitus, or pain during or after stimulation. Additional testing in normal subjects is warranted and planned for 2014. In these normal subject studies, stimulation coil designs that more specifically stimulate the facial nerve (versus surrounding structures) will be used so as to reduce the likelihood of side effects and magnetic field exposure to the brain.
Remarkably, pulsed magnetic stimulation of the facial nerve also appeared safe in dogs with brain hemorrhage, causing no fatalities in a small group of dogs and no enlargement of hematoma size after stimulation. Interestingly, perfusion imaging qualitatively demonstrated a pronounced decrease in extracranial blood flow in such animals and no difference in CBF versus baseline. These surprising findings may reflect an inherent property of facial nerve function because the stimulation parameters used were identical to those in ischemic stroke experiments. We currently have no mechanistic explanation for this observation, but we hypothesize that the autonomic components of the facial nerve are subject to inhibitory regulation triggered by increased intracranial pressure or blood products. Such regulation may be provided by the trigeminal nerve, which has sensory innervation of the meninges that is known to be responsive to various irritants18 and distortion of the dura19 and which connects to the brain stem nuclei from which vasoactive components of the facial nerve derive.20 Further studies into this curious finding are necessary to define its mechanism.
Based on the results of experiments in dogs, we think that a medical device based on pulsed magnetic stimulation of the facial nerve could be beneficial in the emergency treatment of patients with ischemic stroke based on the noninvasive nature of stimulation, general safety of stimulation, and prolonged and sizable effect on CBF and tissue perfusion seen after a relatively short stimulation period. In conjunction with further animal testing, we are now proceding with the development of a clinical prototype for assessing safety and efficacy in humans.
We would like to thank Ramon Torres, DVM, in charge of anesthesia and animal care, and to the members of the CI3M laboratory team—Rafael Lara, Fernanda Maldonado, and Andres Morón—for their general help with these experiments.
Sources of Funding
The experiments were supported by a generous grant from the Bugher Foundation (www.bugher.org) and by Northern Neurosciences Inc, a not-for-profit organization that supports neuroscience research (www.northern-neurosciences.com).
Drs Borsody and Sacristan are employed by, and shareholders in, Nervive Inc, which owns patents on the facial nerve stimulator. The other authors have no conflicts to report.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.113.003243/-/DC1.
- Received August 19, 2013.
- Revision received December 16, 2013.
- Accepted January 6, 2014.
- © 2014 American Heart Association, Inc.
- Yarnitski D,
- Lorian A,
- Dayan A,
- Avnon Y,
- Krakovsky M,
- Lamensdorf I
- Henninger N,
- Fisher M
- Kano M,
- Moskowitz MA,
- Yokota M
- Purdy PD,
- Devous MD Sr.,
- White CL III.,
- Batjer HH,
- Samson DS,
- Brewer K,
- et al
- Loftus CM,
- Greene GM,
- Detwiler KN,
- Baumbach GL,
- Heistad DD
- Goadsby PJ
- Kaube H,
- Hoskin KL,
- Goadsby PJ
- Noseda R,
- Monconduit L,
- Constandil L,
- Chalus M,
- Villanueva L