Abstract
BACKGROUND:
Subarachnoid hemorrhage (SAH) represents a severe injury to the brain and is associated with a high mortality (40%). Several experimental SAH models are described in the literature requiring specialized equipment and a high degree of surgical expertise. Our goal was to validate a simplified, cost-effective model to permit future studies of SAH.
METHODS:
SAH was induced by injection of homologous blood into the cisterna magna. Perfusion-fixation then perfusion of gelatinous India ink was performed. Brains and brainstems were collected and imaged for analysis of cerebral vasospasm. Triphenyl tetrazolium chloride (TTC) staining was used to analyze brain tissue cell death 24 hours following stroke. A composite neuroscore was utilized to assess SAH-related neurologic deficits.
RESULTS:
Anterior cerebral artery and basilary artery diameters were significantly reduced at 24 hours post SAH induction. Middle cerebral artery diameter was also reduced; however, the results were not significant. TTC staining showed no infarcted tissue. Neuroscores were significantly lower in the SAH mice, indicating the presence of functional deficits.
CONCLUSIONS:
This simplified model of SAH elicits pathological changes consistent with those described for more complex models in the literature. Therefore, it can be used in future preclinical studies examining the pathophysiology of SAH and novel treatment options.
Keywords
Introduction
Rationale
Stroke represents the second leading cause of death worldwide [1]. While hemorrhagic stroke (13%) accounts for fewer strokes than ischemic (87%), it has a significantly higher mortality (40%) and morbidity [2]. In particular, subarachnoid hemorrhage (SAH), a subtype of hemorrhagic stroke, is associated with high mortality [1]. Cerebral vasospasm (CVS) and delayed cerebral ischemia following SAH contribute to post-stroke mortality and morbidity [1]. In humans, acute CVS occurs immediately after hemorrhage and can be observed for up to 21 days post-stroke [3]. In mouse models of stroke, CVS has been observed as soon as 1 hour after SAH and has been found to persist for 3–4 days post-stroke [4]. CVS is mostly described for the anterior cerebral artery (ACA), middle cerebral artery (MCA), and basilar artery (BA) [4].
Although sophisticated and highly technical models of murine subarachnoid hemorrhage are ubiquitous in the literature, these require expensive specialized equipment, and a high degree of surgical expertise. Here, the objective was to validate a feasible model which would be cost effective and reproducible to permit future studies of SAH.
The intracisternal injection of blood model involves the injection of autologous or homologous blood directly into the subarachnoid space via the cisterna magna or the prechiasmatic cistern. This model has been performed in rodents and larger animals. Although the injection model is easily reproducible, it has been found to produce effects that are not as severe as the endovascular perforation model [5].
The model was chosen as the basis for this study due to its low mortality risk, high reproducibility, and feasibility. The goal of the study was to establish a simplified model of subarachnoid hemorrhage in mice. Behaviour scores, 2,3,5-triphenyltetrazolium chloride (TTC) staining, and perfusion-fixation with India ink were used to determine whether stroke occurred. We hypothesized that animals in the SAH group will show lower behaviour scores and reduced cerebral vessel diameter.
Materials and methods
Experimental design
Fifteen wild type male C57BL/6 mice (Charles River Laboratories, Wilmington, Massachusetts, USA) were used for experiments. Mice were between 8 and 16 weeks old and weighed between 17 and 32 grams. Mice were housed in the Carleton Animal Care Facility at Dalhousie University. Upon arrival at the facility, animals had a one-week acclimatization period prior to experimentation. Mice were fed a standard diet and had constant access to food and water. Mice were group housed prior to treatment and placed in individual cages during recovery. Experimental procedures were approved by the University Committee on Laboratory Animals (Protocol #21-089).
Mice were assigned to 2 groups. Mice assigned to the sham group (n = 6) had sham surgery with saline injected into the intracisternal space. Mice assigned to the SAH group (n = 6) received homologous blood injection into the intracisternal space. Following surgery, survival time was 24 hours. After 24 hours, animals underwent India ink perfusion (n = 10) or TTC staining (n = 2).
Subarachnoid hemorrhage induction
Homologous blood collection and blood injection system
Homologous blood from a donor mouse was collected from inferior vena cava and loaded to an injection system for experimental blood injection. Briefly, the donor mouse was anesthetized using an intraperitoneal injection of pentobarbital (1 : 1 pentobarbital of 54 mg/ml; 90 mg/kg). The mouse was placed back into its cage until anesthesia was achieved. Next, the mouse was placed in a supine position to prepare for surgery. Before surgery began, pedal reflex was checked to ensure adequate depth of anesthesia. A 1 cm midline incision was made just below the ribcage of the mouse. Saline-soaked cotton tips were used to move intestines to the left and the liver upward to locate the section of the inferior vena cava (IVC) between the kidneys. About 0.5 ml blood was collected from IVC via the heparinized 30 G needle connected to the injection system as described below. The donor mouse was sacrificed by exsanguination. Blood from a donor mouse would be used for 3 animals undergoing SAH.
The injection system was made of a 30G needle connected to a 10 ml syringe through microbore extension tubing. The injection system was heparinized and rinsed with saline (0.9% sodium chloride, Baxter Healthcare Corporation, Illinois) and filled with 7 ml saline before loading the blood from the needle point. The syringe was used to move blood along the microbore extension tubing, the saline pushes the blood through the needle thus the blood would only occupy about 6.5 cm of the tubing behind the needle. In essence, the saline contained in the 10 ml syringe would act as a hydraulic fluid reservoir so when the saline is pushed out of the syringe and into the tubing, the extremely small volume of blood is forced out of the needle and into the target area. The injection system loaded with donor blood was attached to an automatic pump (GenioTouch, Kent Scientific, CT, USA) and was programmed to inject 60μL of blood over 60 seconds.
SAH
The experimental mouse was anesthetized in an induction chamber with a concentration of 4–5% isoflurane in 0.8 liters per minute (LPM) of oxygen. Once anesthetized, isoflurane was reduced to 2–3% concentration and a nose cone was used to maintain anesthesia. The mouse was then placed in a prone position on a feedback-controlled heating pad (37degC). The mouse was placed on the heating pad with a small 3D printed wedge under the neck. The wedge was used to create a 120deg angle between the mouse’s snout and its body. The pedal reflex was checked periodically for the duration of the experiment to ensure adequate depth of anesthesia. A rectal temperature probe was used to monitor temperature during the experiment. Eye lube was placed on the eyes and the eyes were covered with a small piece of gauze. A piece of tape was placed over the snout and eyes to old the mouse in place.
The occipital crest (the part of the skull protruding immediately above the neck muscles) was used as a reference point –this point is roughly found between the ears. A 1 cm midline incision was made on the posterior scalp between the occipital bone and the arch of the first cervical vertebrae. Cotton tips were used to control any bleeding that arose. Forceps were used to bluntly dissect superficial connective tissue and occipital muscles until the atlantooccipital membrane was exposed and the cisterna magna was located. The tissues were held in place with the membrane exposed using a small retractor (Fig. 1). Next, magnification was used to precisely visualize the location of the cisterna magna –a blood vessel running directly down the brainstem called Arteria Dorsalis Spinalis [6] used as a landmark. Once the cisterna magna was located, the blood-loaded needle connected with pre-set tubing and 10 ml syringe system was inserted at a 45deg angle caudally and an assistant started the pump as described above. The needle was kept still for the entire 60 seconds during the blood infusion. Animals in the sham groups were injected with saline instead of blood. Immediately after the injection, the animal was placed at a 45deg angle with its head downward for 15 minutes to encourage clot formation. While the animal was in the head-down position, the neck muscles were approximated, and tight sutures (5-0 sutures) were used to ensure no cerebrospinal fluid (CSF) leakage. The mouse was then placed in a cage alone, on a heating pad to recover. Mice were given a food and water mixture on the bottom of their cage. Post-operative analgesia was not necessary; however, we used topical 2% lidocaine gel on the incision site.
Exposed cisterna magna. Image shows tissues held in place with retractor during surgery. The position of the cisterna magna is marked by an asterix.
India ink-stained cerebral vessels. Images showing the mouse cerebra vessels stained with India ink. (A) Full brain and stained cerebral vessels. (B) Bifurcation of the internal carotid artery and the anterior and middle cerebral arteries. (C) Basilar artery and bifurcation of the posterior arteries.
Vessel measurement technique. Image showing the technique used to measure vessel diameter. Red line indicates 100μm from the bifurcation of the posterior cerebral arteries. Yellow line shows the measurement of the vessel diameter.
Neuroscoring was performed at 22 hours following SAH induction or sham surgery. A composite neuroscore (Appendix A) for evaluating functional deficits in mice subjected to SAH was used [5]. The composite score consists of 8 subtests –observation of spontaneous activity, climbing, balance, side stroking, vibrissae touch, visual, forelimb use, and hindlimb use. Low scores are indicative of increased neurological deficits, while high scores are indicative of normal behaviour.
India ink perfusion
The animals were anesthetized with isoflurane (as described above) at 23 hours post SAH induction or sham surgery in a fume hood. Perfusion-fixation with formalin and India ink was performed through cardiac puncture as described by Xue et al. 2014 [7]. Briefly, the chest was opened to expose the heart and an 18 G blunt fill needle connected to a microbore extension set was inserted to the left ventricle, and the right atrium was cut to allow blood/perfusion to go through. Three 10 mL syringes of the following were used for manual perfusion in sequence: phosphate-buffered saline (PBS), 10% formalin, India ink (10%) mixed with 2% gelatin (reference here).
After perfusion was complete, the mouse was left in the supine position and placed in a –20degC freezer for 20 minutes before brain collection.
Brain collection
The brain was collected from the perfused or euthanized mice for imaging or TTC staining. Briefly, using small dissecting scissors, the scull was cut along the sagittal suture to avoid damaging the cortex. The scull was carefully peeled away using forceps. The brain was lifted out of the scull using a small metal scoop. If the animal was perfused with India ink, the brain was stored in 10% formalin. If the brain was to be used for TTC staining, it was placed on a brain matrix and put into a –80degC freezer for 5 minutes to flash-freeze the tissue.
TTC staining of brain slices
2% TTC solution was made by dissolving 200 mg TTC powder (Alfa Aesar (A10870), Massachusetts, USA) in 9.8 mL of saline in an incubator 37degC for 10 minutes with a shaker until the TTC powder was completely dissolved.
The flash-frozen brain was placed on the brain matrix and five or six 1 mm coronal sections were made with a surgical blade. Each section was placed into a well on a 12 well plate containing TTC solution on a platform shaker in an incubator at 37degC for 10 minutes. After 10 minutes, slices were removed and placed onto a transparent grid plate for imaging. Images of the anterior and posterior sides of the slices were taken.
Infarct volume quantification
Brain slices were analyzed using Fiji (NIH, USA). The volume of issued infarcted was expressed as a perfect of total tissue to standardize the value for varying brain sizes. The total area of each brain was determined by tracing the outer edge of each brain slice and summing the surface areas. The infarct size was measured by tracing around the infarcted (white) tissue on each brain slice and measuring the area. Percent impacted was calculated by dividing the total infarcted area by the total brain area.
Cerebral vessel imaging
Brains were removed from formalin immediately prior to imaging and placed onto a microscope slide ventral side up for ease of positioning. A Kimwipe was used to blot dry the brain to reduce glare from the bright light of the microscope. Images were taken with MVX10 microscope, using various objectives (0.63, 1, 2, 2.5, 3.2, 4, 5, 6.3) as needed to image the vessels of interest. The selection of the highest possible magnification was based on maintaining optimal image quality. Images of the major cerebral vessels were taken using cellSens (Olympus Life Science, Japan).
Cerebral vessel measurement
Diameters of the ACA, MCA, and BA were measured using Fiji (NIH, USA). Prior to measurement, Fiji was calibrated using a stage micrometer. Measurements of ACA and MCA diameter were made 100μm distal to the bifurcation of the internal carotid artery. The BA diameter was measured in two places, 100μm distal to the bifurcation of the posterior cerebral arteries, and 100μm proximal to the bifurcation of the vertebral arteries. If measurements at both locations were obtained, the values were averaged. If measurements could be made in only one location, only that value was used. Measurements were made as follows: first, a 100μm line was drawn in the vessel lumen (parallel to the vessel walls), then a straight line was drawn perpendicularly across the diameter of the vessel.
Statistical analysis
Results are expressed as a mean±standard deviation. Calculations were done using GraphPad Prism (GraphPad Software, San Diego, California USA). All data was analyzed with Prism using a two-tailed unpaired t-test. Statistical significance was accepted at P < 0.05.
Results
General observations
Animals took approximately 10 minutes to resume movement in their cage after completion of surgery and removal of anesthetic gas. All animals behaved normally and ate and drank typically. No neurological deficits were explicitly visible prior to neuroscoring. No animals died following sham or SAH surgery. Two animals (one sham and one SAH) showed brief respiratory depression following injection of blood or saline into the intracisternal space. Both animals recovered in less than 1 minute and no significant concerns were evident. All animals showed minimal leakage of the fluid injected (blood or saline) in the last 10–15 seconds of the injection. Three animals (SAH group) showed pooling of blood in what is presumed to be the subarachnoid space. This could have equally occurred in sham animals; however, it was not visible due to the transparency of saline. There were no apparent signs of damage to the brain tissue caused by insertion of the needle into the intracisternal space.
Cerebral vessel diameter
ACA diameter
Fig. 4 shows the ACA diameters in SAH and sham brains. ean diameter of ACA in the SAH group was 89.82±21.46μm. The mean diameter of ACA vessels in the sham group was 180.80±59.93μm. The ACA diameter in mice subjected to SAH was reduced significantly (about 50%) when compared to sham animals at 24 hours post-surgery (P = 0.0183).
ACA vessel diameter. Effect of SAH on ACA vessel diameter. Data represent the diameter of the ACA in micrometers. Data is presented as the mean±SD (SAH, n = 4; sham, n = 7); *p < 0.05.
The effects of SAH on MCA diameter are shown in Fig. 5. MCA vessels showed vasoconstriction 24 hours following SAH. The MCA mean diameter for sham animals was 157.0±38.87μm. That of the SAH mice was 97.72±20.46μm. MCA vessels in mice subjected to SAH showed a 38% decrease in diameter. The change in MCA vessel size because of SAH was not statistically significant (P = 0.0846). A one-tailed t-test yielded a statistically significant value (P = 0.0423).
MCA vessel diameter. Effect of SAH on MCA vessel diameter. Data represent the diameter of the MCA in micrometers. Data is presented as the mean±SD (SAH, n = 2; sham, n = 7); data is not significant (ns).
Fig. 6 shows the effects of SAH on BA diameter. At 24 hours post-stroke, SAH animals showed marked vessel constriction. The mean BA vessel diameter of SAH mice was 136.6±9.143μm. The mean BA diameter in sham animals was 172.9±18.33μm. SAH mice showed a 21% reduction in vessel diameter. The vessel constriction was statistically significant (P = 0.0122).
BA vessel diameter. Effect of SAH on BA vessel diameter. Data represent the diameter of the BA in micrometers. Data is presented as the mean±SD (SAH, n = 4; sham, n = 4); *p < 0.05.
Fig. 7 shows mouse brains stained with TTC. No dead tissue was observed in both sham and SAH mouse brains as evidenced by the absence of white tissue (Fig. 7). As such, infarct size was determined to be zero in both sham and SAH mice.
TTC stained brains. Brains stained using TTC. Red tissue is healthy. White tissue (not pictured) is infarcted tissue. (A) TTC stained brain from a sham mouse. (B) TTC stained brain from an SAH mouse.
Neuroscores. Data represents scores obtained using a composite neuroscoring system. Data is presented individual data points and mean±SD (n = 12); *p < 0.05.
Animals were scored using the composite neuroscore (Appendix A) 22 hours after surgery. Lower behaviour scores are indicative of increased functional deficits following stroke. The mean score for mice in the sham group was 23±1.265 points. The mean behaviour score for mice in the SAH group were 20.83±1.472 points. Mice subjected to SAH showed significantly lower behaviour scores (P = 0.0210). Variability was seen in the following subtests: climbing, balance, visual, vibrissae touch, and hindlimb use. The greatest variability was seen in the climbing subtest –6 mice were scored a ‘2’, indicating presence of functional deficits.
Discussion
We evaluated the effects of SAH on ACA, MCA and BA vessel diameter, infarct size, and neuroscore in attempt to validate a new simplified model of SAH in mice. We found that the simplified model causes constriction of MCA vessels and significant constriction of ACA and BA vessels 24 hours post-SAH. Additionally, mice subjected to SAH showed functional deficits when evaluated with the composite neuroscore (Appendix A). No infarcted tissue was observed at 24 hours post SAH.
The effect of SAH on vessel diameter in mice has been evaluated in several studies using different models of SAH [8–13]. Additionally, the time course of cerebral vasospasm establishment and resolution has been investigated. Using the endovascular puncture model of SAH, Kamii et al. [14] found a 57% reduction in MCA diameter no change in BA diameter when comparing sham and SAH groups. Interestingly, these results do not align with those produced by our simplified model of SAH (we saw an insignificant change in MCA diameter and a significant change in BA diameter). Vessel recovery (re-establishment of normal lumen size) was observed at 7 days post-SAH. The same model has yielded significant neurological deficits at 24 hours; however, here the Modified Garcia Score was used [11].
The intracisternal blood injection model (which our simplified model was based upon) has been found to elicit vasospasm in ACA, MCA, and BA vessels. Lin et al. [4] perfused vessels with India ink and used a computer-assisted image analysis system (MetaMorph) to automatically measure vessel diameters. Vasospasm was observed as early as 1-hour post-SAH and was found to persist until 7 days post-SAH. In our comparable simplified model of SAH, similar results were observed; however, we only examined vessels at one time point (24 hours). Additionally, Lin and colleagues did not observe any functional deficits following SAH; however, a scoring system was not employed.
Although both models of murine SAH have been found to produce cerebral vasospasm, the endovascular perforation model is associated with significantly higher mortality. These results indicate that the intracisternal injection model of subarachnoid hemorrhage yield more widespread cerebral vasospasm (reduction in mean vessel diameter was observed in ACA, MCA, and BA). Our finding that the largest reduction in vessel diameter was observed in the ACA aligns with the findings of Lin et al [4]. This suggests that the intracisternal injection model elicits the greatest effects on the anterior cerebral circulation.
Staining of mouse brains with TTC has been used in several models of stroke. In ischemic stroke, TTC staining can be used to measure infarct volume [15]. Here, TTC staining was used to investigate cerebral ischemia following SAH. We found no evidence of tissue death. In the literature, TTC staining post-SAH has yielded varying effects with some groups successful visualizing cerebral infarction and others unsuccessful [16]. The lack of effect observed here could be due to the timepoint at which staining was done. Delayed cerebral ischemia entails vasospasm, spreading ischemia, and delayed complications including infarctions [17]. Since the intracisternal injection model of SAH has shown sustained cerebral vasospasm for as long as 7 days, it is possible that our 24 hours timepoint was not long enough to show the effects of delayed cerebral ischemia.
Our study was the first to assess the neurological effects of the intracisternal injection model of SAH using the newly developed composite neuroscore [5]. This new scoring system was developed in response to the need to improve the usefulness of preclinical studies of stroke. The composite scoring system specifically assesses the deficits caused by SAH in mice. Although other groups have found no functional deficits in mice following SAH, the composite neuroscore criteria allowed us to find significant differences between the sham and SAH groups. According to the composite score, mice subjected to SAH in our study showed increased functional deficits when compared to the sham group. Our finding of functional deficits in mice following SAH is consistent with what would be expected considering the clinical effects of SAH observed in human patients (specifically, depressed level of consciousness and palsy) [18].
Upon surveying the literature for a suitable model of murine SAH to institute at our lab, we found published models to be highly complex with varying levels or reproducibility. Our simplified model shows several benefits, while still eliciting the effects expected following SAH (vasospasm and functional deficits). First, using 3D printed wedges, our model eliminates the need for a stereotaxic frame. In previous models, the stereotaxic frame is used to angle the mouse’s head downward to ease access to the atlantooccipital membrane. 3D printed wedges achieved the same effect and allowed for more flexibility in positioning the mouse. Additionally, other models make use of an operating microscope to identify the location of the cisterna magna. Our model eliminates the need for an operating microscope by instead relying on consistent positioning of the mouse (achieved with 3D printed wedges) and the use of anatomical landmarks under a magnification glasses. A previously published protocol for the perfusion of cerebral vessels with India ink was highly technical and required sophisticated equipment including pressure sensors [7]. We simplified this protocol immensely and removed the need for controlled perfusion pressure or for a syringe pump. Instead, we used a two-experimenter system and microbore tubing to allow the second experimenter to monitor perfusion speed while the first was preoccupied with the maintaining the position of the needle in the left ventricle. Although we had some challenges with perfusing all major vessels, this also occurred in the previously described, and more complex model. Finally, image analysis was done manually using technology downloaded free of charge (Fiji, NIH, USA). This allowed image analysis to be done at any time with minimal training required.
Future studies aiming to further validate this model should aim to investigate both earlier and later time points. Evaluating the effects of SAH at earlier timepoints (<24 hours post-surgery) will allow for further elucidation of the timeline of development of cerebral vasospasm and functional deficits. This information could be crucial in later studies assessing treatment options for SAH. Furthermore, investigating the effects of SAH at later time points (>24 hours post-surgery) will allow the effects of delayed cerebral ischemia to be assessed. Specifically, at which timepoint cerebral vasospasms resolve, functional deficits are no longer observable, and whether infarction can occur as a result of this model of SAH.
Furthermore, future studies could employ histology in order toto examine the structural impacts of vasospasm on cerebral vessels. Changes in the vessel structure such as corrugation of the elastic lamina, have previously been reported [19, 20]. Immunohistochemistry and cytokine analysis could be used to help elucidate the mechanisms of cerebral vasospasm –information which could be valuable in the development of treatments for SAH.
Footnotes
Appendix A
Score range is from 0 (maximum deficits) to 24 (no deficits). Notes:
From: Matsumura, K. et al. Neurobehavioral Deficits After Subarachnoid Hemorrhage in Mice: Sensitivity Analysis and Development of a New Composite Score. J Am Heart Assoc 8, (2019).