Non-Invasive Vagus Nerve Stimulation Protects Neurons in the Perihematomal Region and Improves the Outcomes in a Rat Model of Intracerebral Hemorrhage

Abstract

Background

Intracranial hemorrhage (ICH) is a devastating stroke subtype with a high rate of mortality and disability. Therapeutic options available are primarily limited to supportive care and blood pressure control, while the surgical approach remains controversial.

Objective

In this study, we explored the effects of noninvasive vagus nerve estimation (nVNS) on hematoma volume and the outcome in a rat model of collagenase-induced ICH.

Methods

Adult male Wistar rats were randomized into two study groups: 1) ICH-treated (rats treated with five 2-minute nVNS) and 2) ICH-control (ICH with sham nVNS). Each group received either 0.1 U or 0.2 U collagenase dose. After assessing neurological function, rats were euthanized at 24 h for spectrophotometric hemoglobin assay, hematoma volume measurements, and histological studies.

Results

The ICH-treated group that received 0.1 collagenase dose demonstrated significantly smaller hematoma volume and improved motor function compared with the ICH-control with the same dose. Furthermore, the pooled data for the ICH-treated groups (both 0.1 U and 0.2 U collagenase) revealed a reduction in neuronal loss in the perihematomal region in the histopathological studies. This effect was not significant for the group that received a 0.2 collagenase dose.

Conclusion

nVNS therapy in acute settings may provide a neuroprotective effect and limit hematoma expansion in smaller volumes, improving neurological function post-ICH.

Introduction

Stroke is a major public health concern globally, ranking as the second-leading cause of death and the third-leading cause of disability worldwide. Intracerebral hemorrhage (ICH) constitutes 27.9% of all types of strokes and represents a significant portion (44%) of stroke-related mortality.¹ Furthermore, ICH patients suffer a high rate of long-term disability.^{2, 3} Unfortunately, current therapeutic options are mostly limited to supportive care, including blood pressure control, glucose control, fever treatment, and anticoagulation reversal.⁴ No specific medical treatment has been clearly beneficial^{5, 6}, and the surgical approach for hematoma evacuation remains controversial.^7–9 In addition, ICH phase III randomized clinical trials have failed to significantly improve outcomes among the treatment arms.

Vagus nerve stimulation (VNS) is a neuromodulation technique that provides electrical stimulation to the cervical vagal nerve segment and has been applied in various neurological disorders.^10–12 VNS was initially delivered through a device that requires surgical implantation of the electrodes around the cervical vagus nerve and connecting them to a stimulator implanted under the anterior chest wall.¹³ The surgical procedure increases costs, and its use is limited to chronic settings; additionally, invasive VNS (iVNS) has been related to adverse effects, including infection, vocal cord paresis, and cardiac events.¹⁴ Recently developed non-invasive VNS (nVNS) have been demonstrated to be effective and safe, does not require a surgical procedure, and facilitates a dose-adjusted stimulation on demand.^15–17 The nVNS was initially approved by the Food and Drug Administration (FDA) as an adjunctive therapy in drug-resistant epilepsy¹⁸ , and since then, due to its properties of neuromodulation and immune-modulation^{19, 20}, it has been studied in several brain disorders.²¹ However, the preclinical and clinical evidence on the use of VNS for ICH is scarce.²² Given the available preclinical and clinical evidence of the utility of VNS as an adjunctive therapy for various neurological diseases, such as migraine, headaches, seizures, and ischemic stroke, investigating the effect of VNS on improving the outcomes of ICH is warranted.^{11, 12, 23, 24}

This study aimed to evaluate the use of nVNS as a therapeutic approach to control hematoma volume and improve the outcomes of ICH in a rat model of collagenase-induced ICH.

Materials and Methods

We used male Wistar rats (Charles River Laboratories, Wilmington, MA) for the experiment. Animals were maintained under standard conditions (25 ± 2°C, the humidity of 50 ± 6%, 12 h light/dark cycle) with unlimited access to standard food and water for at least one week for acclimatization before neurobehavioral training. During the acclimation period, rats were handled by the trainer for habituation. Before inducing ICH, animals were randomly assigned to two different study groups: 1) ICH-treated (rats treated with five 2-minute nVNS) and 2) ICH-control (ICH with sham nVNS). We used the GraphPad randomization algorithm (GraphPad Software Inc., La Jolla, CA, https://www.graphpad.com/quickcalcs/randomize1.cfm) to randomize the two study groups. Each group received either 0.1 U or 0.2 U collagenase, which resulted in two different hematoma volumes. All measurements/assessments were performed in a blinded fashion to reduce study bias.

The results are reported based on the ARRIVE 2.0 Guidelines for Reporting Animal Research ²⁵. The study was approved and conducted according to the guidelines set by the Institutional Animal Care and Use Committee at the University of New Mexico.

ICH Induction

The rats were anesthetized in an induction chamber with 3% isoflurane. Next, the animal was mounted in the prone position in a stereotaxic frame (David Kopf Instruments, Tujunga, CA) housing a nose cone to deliver 2.0–2.5% isoflurane throughout the procedure. To alleviate pain, infection, and dehydration, buprenorphine (0.03 mg/kg), meloxicam (2 mg/kg), Baytril (10 mg/kg), and 5 ml of normal saline were administered subcutaneously at the start of the surgery. The collagenase-induced ICH model used for this study has been previously described in detail.²⁶ To induce ICH, the scalp was shaved and disinfected. Using a floor surgical microscope (Zeiss, OMPI^® CS on NC-2 stand), a midline incision was performed to expose the skull and retract the fascia. A 1-mm burr hole (3.0 mm lateral and 0.2 mm rostral to bregma) was drilled on the right side of the skull without damaging the dura. Next, a 26 G needle attached to a micro-syringe (10 μl gastight syringe, model 1701, Hamilton, Reno, NV) was inserted stereotactically into the right basal ganglia through the burr hole. At first, the syringe needle was lowered to a depth of 6.0 mm at a rate of 1mm/min. After 30 seconds, the needle was retracted 0.5 mm and positioned at a 5.5 mm depth. Each animal received 1 or 2 μl normal saline containing 0.1 U or 0.2 U (0.1 U/μl concentration) of collagenase (Type VII, 0.2 um filtered, high purity, C2399, Sigma-Aldrich, Inc., Saint Louis, MO) in the right striatum infused at a rate of 2 μl/minute using an infusion pump (Nonomite, Harvard Apparatus, Holliston, MA). The needle position was maintained for another 10 minutes to allow collagenase to be fully absorbed by the surrounding tissue before withdrawing it at a 1mm/min rate. After the needle withdrawal, the burr hole on the skull was sealed with bone wax, and the skin incision was sutured. Body temperature was monitored with a rectal probe and maintained at 37±0.5 °C using a heating pad. The heart rate and %SpO2 were monitored using a hand-held pulse oximeter (model 2500A VET, Nonin Medical, Inc., Plymouth, MN).

The nVNS Protocol

The nVNS therapy protocol has been previously described in detail ²⁷. We used a gammaCore device (electroCore, Inc., NJ, USA) to deliver stimulations (24 V, 60 mA, and 1 ms duration bursts of 5 kHz sinewaves, repeated at 25 Hz). The stimulation therapy was initiated 30 minutes post-ICH, with five 2-minute sessions (each 10 minutes apart), with the highest intensity/amplitude (24 V) for the ICH-treated group (see Figure 1). For the ICH-control group, the device was attached but not turned on. Animals in both groups were kept under anesthesia for the same duration based on the ICH-treated group’s (during the surgery + 30-minute interval post-injury + 50 minutes nVNS) to avoid study bias based on anesthesia duration. After the nVNS therapy, the rats were moved to an intensive care system (ThermoCare Intensive Care Unit, Daisy Products LLC, dba: ThermoCare, Paso Robles, CA) for recovery from anesthesia. The animals were sacrificed at 24 hours for spectrophotometric hemoglobin assay, hematoma volume measurements, and histological studies.

Figure 1:

We used a modified gammaCore device (A) with a miniature collar (B). The nVNS protocol for the experiment, with five 2-minute stimulations and 10-minute intervals between each stimulation starting 30 minutes post-ICH (C).

Hematoma volume measurement

Each rat was transcardially perfused with 200 ml of cold phosphate-buffered saline (PBS) to measure hematoma volume. The brain was immediately harvested, 2 mm thick sliced, and photographed from both anterior and posterior views. The color images were converted to grayscale values using Image Processing and Analysis in Java (ImageJ, NIH Image) software. Then, the observed area of the hematoma on the ipsilateral hemisphere was outlined on both anterior and posterior views. The area on each view was calibrated by multiplying the area by the average grayscale. The hematoma volume was calculated as the average of the two views (anterior and posterior) areas multiplied by 2 mm slice thickness. The total volume was calculated as the sum of the volumes for each slice.

Spectrophotometric Hemoglobin Assay

We have previously reported the spectrophotometric hemoglobin assay method to quantify the amount of hematoma.²⁶ We used the ipsilateral photographed brain slices for hematoma volume measurements to calculate the hemoglobin concentration. One ml sterilized distilled water was added to the ipsilateral brain hemisphere and homogenized (Biogen Series, Model Pro200, Pro Scientific, Inc., Oxford, CT) for 30 seconds and sonicated (QSPNICA sonicator, Q Sonica, LLC, Newtown, CT) for another 60 seconds. Each sample was transferred to a 1.5 ml Eppendorf tube and centrifuged at 13,000 rpm at - 4 °C for 30 minutes. The hemoglobin-containing supernatant was transferred to a new tube and stored at −80°C for later analysis. After the completion of the study, the supernatant samples were mixed with Drabkin’s reagent (using a 1 to 4 ratio by volume) to obtain 1200 μl solution samples for spectrophotometry at 540 nm wavelength using the Ultrospec^™ 2100 UV/Visible spectrophotometer (GE Healthcare, formerly Amersham Biosciences). A standard curve was also created by preparing 11 samples containing a known concentration of bovine hemoglobin ranging from 0 to 3 mg/ml.

Neurological and Motor Function Assessments

The neurobehavioral and motor assessments were conducted in a blinded fashion. Motor function training and assessment procedures used in our laboratory have been described previously.^28–30

Beam Walking

Rats were trained on the beam walking apparatus for 5 days before the injury. Assessments were performed at baseline and 24 hours post-ICH. For the beam walking, the time to cross the 1-meter distance on a 3.5 centimeters thick beam, the number of footfalls, and falling from the beam (if applicable) were recorded and scored. For post-ICH assessments, the time to cross the 1-meter distance and the footfall counts were adjusted for the baseline performances by subtracting the baseline values from the ones at 24 hours. Per our protocol, an arbitrary 15 seconds was assigned to the crossing time if the rats fell from the beam before completing the task (see Figure 2 for the beam walking apparatus used for the study).

Figure 2:

The modified beam-walking apparatus used for the experiment.

Rotarod Test

Balance and motor function were also assessed using a rotarod test (Rotamex-5; Columbus Instruments, Columbus, OH). The rats were trained on the apparatus until they could hold their balance on the rotating rod for 120 seconds in multiple trials with a rotational speed from 1 to 25 rpm and an acceleration of 1 rpm every 2 s (cutoff time at 120 seconds). For post-ICH assessments, rats were given 2 trials, 30 minutes apart, where the latency to fall was recorded. We used the best time from the two trials as the result. Two rats did not reach the 120-second cutoff during the training period. Therefore, the 24-hour values were adjusted for the baseline performance.

Garcia Test

Sensorimotor deficits were evaluated using Garcia’s test, a composite neurological test.^31–34 Baseline assessments were conducted before the injury, followed by post-ICH assessments at 24 hours. Garcia’s test includes 6 components: spontaneous activity, symmetry in the movement of limbs, forepaw outstretching, climbing, body proprioception, and response to vibrissae touch. The possible range of combined neurological scores is from 3 (most severe) to 18 (no deficit).

Aquaporin-4 Enzyme-linked immunosorbent assay Test

For the rats with 0.2 U collagenous, photographed brain ipsilateral slices from hematoma volume measurements were transferred into 1.5 ml Eppendorf tubes with 200 μL of RIPA buffer. Samples were immediately homogenized, sonicated, and centrifuged at - 4 ° C at 10.000 RPM for 20 minutes. The supernatants were transferred into new 1.5 ml Eppendorf tubes and aliquoted into 35 μL samples to be stored at −80° C. At the completion of the experiments, samples were transferred to the Cytokine Reference Laboratory (CRL, University of Minnesota, CLIA’88 license #24D0931212) for aquaporin-4 (APQ-4, Novus Biologicals, LLC, Centennial, CO) Enzyme-linked immunosorbent assay (ELISA) test.³³ Samples were assayed according to the manufacturer’s instructions. The assay employed the quantitative sandwich enzyme immunoassay technique. The absorbance was measured on a microtiter plate reader (BioTek Synergy XL Multimode Reader, Agilent, Santa Clara, CA). The intensity of the color formed is proportional to the concentration of the sample. Samples were tested in duplicate, and values were interpolated from a log-log or 4PL-fitted standard curves. The staff member at CRL who performed the tests was blinded to the study groups.

Histopathological Assessments

Rats were transcardially perfused with phosphate-buffered saline (PBS) followed by 4% paraformaldehyde (PFA) in PBS. The harvested brains were fixed for 48 hours in 4% PFA in PBS, cryoprotected in 30% sucrose in PBS, and frozen in optimal cutting temperature (OCT) compound (Sakura Finetek, Torrance, CA, USA) using dry ice for hematoxylin and eosin (H&E) and Nissl staining. Sample rat brains from each treatment group were sectioned coronally using cryostat Leica CM1850 (Leica Biosystems, Wetzlar, Germany). Tissue morphology was displayed by Cresyl Violet (Nissl) and H&E staining protocols (Anatech LTD Harris hematoxylin and eosin-Y) on 10 μm thick sections. Images were taken using NanoZoomer 2.0-HT slide scanner and analyzed using Hamamatsu NDP.view2 software (Hamamatsu Photonics K.K., Hamamatsu City, Japan).

Histopathological assessment was performed by an investigator blinded to the treatment groups using the Cell Counter and Volume Calculator functions of the Analyze Plugin within the NIH ImageJ software. Neuronal counts were conducted in the perihematomal regions within a 100 μm radius from the hematoma border across consecutive Nissl-stained brain sections.

Statistical Analysis

Inter-group differences in lesion volume, Hemoglobin, rotarod time, MBW time, MBW anxiety index, and Garcia score were assessed with independent samples t-test or two-way ANOVA. Intra-reader agreement for lesion volume was assessed using intraclass correlation (ICC). The F-test was used to assess variance differences. Correlation analysis between lesion volume and neurobehavioral variables or between neurobehavioral variables was performed using Spearman (rank) correlation. The statistical significance threshold was set for P<0.05. All the statistical analysis was performed with MedCalc (v.22) or custom R-scripts.

Results

The data that support the findings of this study are available from the corresponding author upon reasonable request. For the study, we used 53 male Wistar rats (age = 11.36 weeks, 95% CI: 11.09–11.6 weeks and weight: 371 g, 95% CI: 361–381 g). The sample size was n = 20 rats (10 ICH-control and 10 ICH-treated rats) in the 0.1 U collagenous sample and n = 33 (17 ICH-control and 16 ICH-treated rats) in the 0.2 U collagenous sample. Of those 53 rats, 41 rats (21 controls and 20 VNS rats) were used for the lesion volume measurements.

Hematoma volume measurement

For the 0.1 U dose collagenous groups, the mean hematoma volume was 8.97 mm³ (95% CI: 6.48 −11.46) for the ICH-control group (n=7) and 5.51 mm³ (95% CI: 3.12 to 7.89) for the ICH-treated group (n=7). The t-test indicates a significant difference between ICH-treated and ICH-control groups (P=0.030). For the 0.2 U dose collagenous groups, the mean hematoma volume was 12.38 mm³ (95% CI: 10.99–13.78) for the ICH-control group (n=14) and 11.19 mm³ (95% CI: 8.61–13.77) for the ICH-treated (n=13). The t-test for unequal variances doesn’t find a significant difference (P=0.38). Figure 3 shows the graphs for the hematoma volume measurements and brain slices samples.

Figure 3:

A) Comparison of hematoma volumes vs. treatment groups and doses. Error bars indicate 1 standard deviation (SD). B & C) Brain slices sample images for ICH-control and ICH-treated groups for 0.1 U collagenase dose. D & E) Brain slices sample images for ICH-control and ICH-treated groups, respectively, for 0.2 U collagenase dose. For simplicity, only photographs from one side of the slices are shown.

Overall, the mean lesion volume for both treatment groups was significantly larger (t-test P=0.0002) for the 0.2 U dose group: 11.81 mm³ (95% CI: 10.46–13.16) than the 0.1 U dose group: 7.24 mm³ (95% CI: 5.45–9.03). The intra-reader agreement for the lesion volumes was excellent, with an intra-class correlation (ICC) of 0.99 (95% CI: 0.988–0.996).

Spectrophotometric Hemoglobin Assay

For the 0.1 U dose groups, the mean ipsilateral hemoglobin concentration was 1.08 mg/ml (95% CI: 0.94 to 1.21) for the ICH-control group (n=6) and 0.930 mg/ml (95% CI: 0.58 to 1.28) for the ICH-treated group (n=6). For the 0.2 U dose groups, the ipsilateral hemoglobin concentration was 1.36 mg/ml (95% CI: 1.01 to 1.72) for the ICH-control group (n=8) and 1.44 mg/ml (95% CI:1.04 to 1.84) for the ICH-treated group (n=7). The t-test did not reveal any significant pairwise differences in ipsilateral hemoglobin concentration between the ICH-control and ICH-treated groups for either collagenous dose despite a visible lower hemoglobin concentration for the ICH-treated group at 0.1 U dose (see Figure 4). Moreover, two-way ANOVA showed a significant effect for the dose factor only (P= 0.009) with higher ipsilateral hemoglobin concentration for the 0.2 U dose.

Figure 4:

Ipsilateral Hemoglobin concentration (mg/ml) vs. treatment groups and doses. Error bars indicate 1 SD.

Beam Walking

For the 0.1 U dose groups, the mean beam walking time change at 24 hours was 5.51 seconds (−0.4 to 11.5 seconds) for the ICH-control group (n=10) and 2.1 seconds (−0.4 to 4.5 seconds) for the ICH-treated group (n=10). The t-test for unequal variances did not reveal any significant difference in time between the two groups. However, the F-test for inequality of variances was significant (variance ratio: 5.84, P=0.015). For the 0.2 U dose groups, the mean beam walking time change at 24 hours post-ICH was 10.7 sec (5.2 to 16.2 seconds) for the ICH-control group (n=17) and 4.8 seconds (0.9 to 8.6 seconds) for the ICH-treated group (n=16), see Figure 5. No significant difference in the beam walking time was observed (P=0.07). Similarly, no significant difference in variance was observed using the F-Test (variance ratio: 2.21). No significant difference was observed for the adjusted number of footfalls between the two treatment groups at dose 0.1 (median =1, IQR: 0 to 1 for the control group versus median = 0.5, IQR: 0 to 2 for the nVNS group) and at dose 0.2 (median = 1, IQR: 0 to 2 for the control group versus median = 0, IQR: −1 to 1 for the nVNS group), or between the dose groups (Mann-Whitney tests).

Figure 5:

Beam Walking time (seconds) for the two collagenase doses for the ICH-treated and ICH-control groups. There were no significant pairwise differences, except for beam walking time variance differences on day 1 for 0.1 U dose (P=0.015). Error bars indicate 1 SD.

Rotarod Test

For the 0.1 dose groups, the mean rotarod time change at 24 hours was −67.1 seconds (−81.1 to −53.1 seconds) for the ICH-control group (n=10) and −29.2 seconds (−57.4 to −0.96 seconds) for the ICH-treated group (n=10). The t-test for unequal variances showed a significant difference in rotarod time change between the two groups (P=0.0175). The variances between treatment groups were also significantly different (F-Test: p=0.048). For the 0.2 dose groups, the rotarod time change at 24 hours was −60.4 seconds (−79.8 to −40.9 seconds) for the control group (n=17) and −74.6 seconds (−89.4 to −59.8 seconds) for the ICH-treated group (n=15). No significant difference in rotarod times change or variances were observed between the two treatment groups for the 0.2 U dose (see Figure 6).

Figure 6:

Comparison of adjusted rotarod times (seconds) at 24 hours post-ICH vs. treatment groups and doses. Error bars indicate 1 SD.

Garcia Test

For the 0.1 U dose groups, the mean Garcia score at 24 hours was 14.3 (12.61 to 15.99) for the ICH-control group (n=10) and 15.8 (14.74 to 16.86) for the ICH-treated group (n=10). The t-test did not reveal any significant difference in Garcia’s score between the groups (P=0.11), but a possible trend should be considered, considering the small sample size. For the 0.2 U dose groups, the mean Garcia score at 24 hours was 14.35 (13.26 to 15.44) for the ICH-control group (n=17) and 14.63 (13.63to 15.61) for the ICH-treated group (n=16). No significant difference in Garcia score was observed (P=0.70), see Figure 7.

Figure 7:

Garcia score at 24 hours versus groups. Error bars indicate 1 SD.

Aquaporin-4 Enzyme-linked immunosorbent assay Test

We used samples from the ICH-treated (n=7) and ICH-control (n=7) from the 0.2 U dose group for the ELISA test. The AQP-4 mean value was 1.62 ± 0.70 ng/mL in the ICH-control group and 2.90 ± 1.44 ng/mL in the ICH-treated group. The difference in AQP-4 levels in the ipsilateral brain tissue between the two groups was near significant (P = 0.0530). The current study pooled effect size for AQP-4 mean inter-group difference was large (1.14), but our power analysis showed that the sample size would require at least 46 rats (23 rats per group) to detect a significant difference.

Histopathological Assessments

We used a total of 12 samples (n=3 for the ICH-treated and ICH-control groups for the 0.1 U and 0.2 U doses). Due to small sample sizes, we did not observe a statistically significant difference between the two groups for either collagenous dose. However, once we pooled the results for the 0.1 U and 0.2 U doses (n=6), the perihematomal neuronal counts were significantly different between the two groups (P=0.0265, as shown in Figure 8). The effect size using pooled SD was 1.5292 (95% CI: 0.26–2.79).

Figure 8:

A) Perihematomal neuronal counts for the ICH-control and ICH-treated groups. The results for the two collagenase doses are pooled. B & C) Nissl staining samples for ICH-control and ICH-treated groups, respectively, for 0.1 U collagenase dose. D & E) Nissl staining samples for ICH-control and ICH-treated groups, respectively, for 0.2 U collagenase dose. Note: Images are selected based on where the maximum area hematoma was observed.

Correlation analysis (Rho Spearman – rank correlation)

• Correlations between lesion volume (mm³) and ipsilateral hemoglobin concentration (mg/ml): Rho: 0.87, P<0.0001

• Hematoma volume and rotarod time at 24 hours: rho: −0.43. P=0.0061. Moderate correlation.

• Hematoma volume and Garcia test at 24 hours: rho: −0.42. P=0.0063. Moderate correlation.

• Hematoma volume and MBW time at 24 hours: rho: 0.36. P=0.0217. Weak correlation.

Significant correlations between adjusted neurobehavioral variables:

• Rotarod time and Garcia: rho = 0.67. Strong correlation.

• Beam walking time and Garcia score: rho = −0.67. Strong correlation.

• Beam walking and Garcia score: rho = −0.67. Strong correlation.

• Beam Walking time and rotarod: rho = −0.60. Strong correlation.

• Beam walking time and footfall count: rho = 0.36. Weak correlation.

Discussion

In this study, we evaluated the effect of nVNS using a collagenase-induced ICH model with two different doses of type VII collagenase (0.1 U and 0.2 U). Animals were assigned to ICH-treated (nVNS) or ICH-control (sham nVNS) group. In the ICH volume analysis, a 0.1 U dose resulted in a smaller hematoma for the group that received nVNS. Likewise, the hemoglobin concentration tended to be lower in this group, although not statistically significant. ICH volume and hemoglobin concentration were not different between the intervention and control group for those rats that received 0.2 U collagenase. Regarding neurological assessment of balance and motor coordination, the ICH-treated group performed better than the controls on the rotarod apparatus. However, the difference was only statistically significant for 0.1 U collagenase. Both groups did worse in the rotarod test at 24 hours post-ICH compared to the baseline, confirming ICH’s negative impact on neurological function. Evaluation of limb coordination and gait through the beam walking test could not demonstrate a significant effect of nNVS. However, we observed a more homogenous performance for the ICH-treated group compared to controls for the 0.1 U collagenase dose, with some animals in the control group performing far worse. The Garcia test showed a trend toward better scores in the ICH-treated group. However, the small size of the sample could have led to invalid results. Similarly, AQP-4 levels tended to be greater in the intervention group, however a larger sample might be needed to confirm these results. The histopathological assessment showed a greater neuronal loss in the perihematomal regions in the ICH-control group compared with the ICH-treated (nVNS) group. Finally, there were moderate but significant correlations between the hematoma volume and the neurobehavioral outcomes, which confirms the consistency between the severity of the pathology and its functional sequelae. Similarly, the strong correlation observed between the neurobehavioral variables was expected, knowing the overlap of these tests regarding the sensorimotor performances.

The volume of the hematoma in ICH has been associated with prognosis in terms of survival, hospital stay, and disability.³⁵ ICH volume, along with other variables such as age, Glasgow Coma Scale and location of the hemorrhage, have been used to build and validate prognostic scores that could be applied in clinical practice.^36–38 The detrimental effect of the ICH in the acute phase can be related to the mass effect within the parenchyma, causing displacement and disruption of the perihematomal tissue.³⁹ Furthermore, greater ICH volume can lead to higher intracranial pressure (ICP), which compromises tissue perfusion.⁴⁰ Acute hypertensive response is common within the first few hours after ICH and might be associated with hematoma expansion and deterioration.⁴¹ Early control of blood pressure, in addition to blood pressure stability, has been proposed as a strategy to limit hematoma growth and improve functionality.^{42, 43} INTERACT3 trial demonstrated that a care bundle that includes lowering blood pressure, glucose control, fever control, and rapid reversal of anticoagulation (when indicated) improved functional outcomes for patients with ICH.⁴ Our study demonstrated lower hematoma volume in the nVNS group. The lower hemoglobin concentration in the treatment group did not reach significance. We attribute this to the fact that part of the blood of the hematoma was smeared on the plate during photography, confounding the results. However, the very high correlation between hemoglobin concentration and lesion volume (0.87, P<0.0001) indicates that both lesion volume and hemoglobin are consistent overall.

Disruption of the blood-brain barrier (BBB) after ICH in the perihematomal region is a critical event frequently followed by vasogenic edema, inflammation, and secondary injury.^{44, 45} Preserving BBB integrity could alleviate hematoma expansion, perihematoma edema, and inflammation, ultimately improving neurological function.^{45, 46} Our results indicate that nVNS contributes to the attenuation of hematoma growth in addition to the modulation of BBB permeability and edema, which is suggested by the regulation of AQP-4 levels and reduced neuronal loss in the intervention group. Altogether, nNVS could be a potential non-invasive therapy in the hyperacute phase of ICH that could be applied even in the prehospital setting for both hemorrhagic and ischemic stroke.^{12, 47, 48} Potential concerns exist regarding the safety of nVNS in ICH due to its effect on regional cerebral blood flow (rCBF).^{49, 50} It is important to note that the rCBF changes signify brain activations. In one post-traumatic stress disorder (PTSD) study⁵¹, rCBF/perfusion was measured using HR-PET imaging. The nVNS treatment increased anterior cingulate and hippocampus blood flow, indicating a reversal of neurobiological changes with PTSD consistent with improved autonomic control. The brain region activation leads to increased rCBF to a more normal state and does not augment normal physiological rCBF. Furthermore, the risks of increasing hematoma volume have been refuted by the present study and previous clinical and animal studies that have confirmed the safety and feasibility of using nVNS in hemorrhagic stroke.^52–54 Positive effects of rapid and intensive blood pressure lowering post-ICH as part of treatment⁵⁵ could be counterbalanced by the compromise of cerebral perfusion.^{56, 57} The use of nVNS along with blood pressure control could have a synergistic effect given the modulatory cardiovascular properties of nVNS and its ability to selectively induce rCBF changes that might be protective to certain areas of the brain.^58–60

The glial water channel AQP-4 has been shown to facilitate astrocyte swelling (cytotoxic edema) and is responsible for the reabsorption of extracellular edema (vasogenic edema).⁶¹ Perivascular AQP-4 increase post-traumatic brain injury (TBI) has been shown to be attenuated by VNS, leading to decreased brain edema.⁶² VNS (both iVNS and nVNS) has been shown to reduce cortical edema in TBI models.^{27, 63} Regarding our model, the down-regulation of AQP-4 in the affected brain region may be one of the underlying mechanisms for VNS-induced attenuation of brain edema, which in turn alleviates neuronal loss, compression of intracranial structures and ischemia, ultimately leading to better neurological recovery and function. Despite a decrease in AQP-4 levels in the ICH-treated group, we didn’t observe a significant reduction in hematoma size in the 0.2 group. However, the role of AQP-4 seems to be more complex than previously thought. Studies suggest that AQP-4 functionally decreases brain edema and BBB permeability by suppressing endothelial tight junction (TJ) opening and endothelial cell swelling, besides increasing the expression of TJ proteins occludin and zonula occluden-1.⁶⁴ On the other hand, brain edema and BBB disruption have been reported to decrease with reduced AQP-4 levels.⁶⁵ AQP-4 may contribute to brain edema and astrocyte swelling in the post-TBI acute phase and facilitate fluid reabsorption during recovery.⁶⁶

The inflammatory response in the early stage of ICH can affect hematoma volume, edema, and BBB disruption.⁶⁷ In general, blood products in the brain parenchyma and other cellular components can act like Damage Associated Molecular Patterns (DAMP), activating resident immune cells (astrocytes, microglia) part of the innate immune response, and leading to the release of inflammatory cytokines, reactive oxygen species and migration of peripheral immune cells which contributes to BBB disruption, edema, and autophagy.^{68, 69} nVNS has immunomodulatory properties through vagal innervation.⁷⁰ The vagal afferent fibers have IL-1β receptors that carry information to the brainstem’s nucleus of tractus solitarius (NTS). NTS projects this information to the hypothalamus, which stimulates the adreno-corticotrophin hormone production by the hypophysis and leads to the release of glucocorticoids in the adrenal gland and the modulation of systemic inflammation.⁷¹ Additionally, a cholinergic anti-inflammatory pathway has been described where vagal efferent fibers release acetylcholine (Ach) that binds α‐7‐nicotinic ACh receptors (α7nAChR) of macrophages inhibiting the release of proinflammatory cytokines such as TNF-α.⁷²

The study shortcomings include small sample size, the lack of more rigorous neurobehavioral and immunohistochemical assessments, using only male rats, and the use of autologous blood vs. collagenase that is warranted to arrive at a concrete assessment of the role of nVNS in improving the outcome of ICH. Previously, we have shown five 2×2-minute nVNS therapy is more effective than five 2-minute nVNS in reducing the severity of TBI in a rat model of cortical control impact.²⁷ Therefore, future studies should include different nVNS intensities and dosses initiated at different intervals post-ICH to fully understand the therapy’s utility for ICH.

Overall, nVNS might limit hematoma expansion and improve neurological function post-ICH, particularly for smaller hematoma volume. The results did not remain significant for the larger hematomas, suggesting that the clinical severity and size of the ICH might influence the efficacy of the therapy or that a different intensity of treatment may be indicated.

Conclusion

To conclude, our study demonstrated the potential ability of nVNS therapy to limit the size of the hematoma, modulate edema and neuronal loss, and improve neurological function 24 hours post-ICH. The nNVS therapy in the acute phase of ICH can attenuate hematoma expansion and limit the extent of neuronal loss in the perihematomal region. This is also reflected in better neurological function, especially for smaller hematomas. nNVS is a potential therapeutic strategy in the acute stage of ICH that could be applied non-invasively in the prehospital and/or acute setting due to its neuroprotective effects. Translation of these findings into the clinical scenario is warranted in order to confirm its safety and efficacy.

1. We confirm that the complies with all instructions to the authors.

2. We confirm authorship requirements have been met, and all authors approved the final manuscript.

3. The manuscript has not been published elsewhere and is not under consideration by another journal.

4. Reporting checklist: ARRIVE

5. Compliance with ethical standards: The study was approved by the Institutional Animal Care and Use Committee (IACUC) at the University of New Mexico Health Sciences Center (Protocol number: 201170)

Figure 9:

Significant rank correlations (Spearman Rho – P<0.05) between outcome variables