Neurobehavioral testing in subarachnoid hemorrhage: A review of methods and current findings in rodents

Abstract

The most important aspect of a preclinical study seeking to develop a novel therapy for neurological diseases is whether the therapy produces any clinically relevant functional recovery. For this purpose, neurobehavioral tests are commonly used to evaluate the neuroprotective efficacy of treatments in a wide array of cerebrovascular diseases and neurotrauma. Their use, however, has been limited in experimental subarachnoid hemorrhage studies. After several randomized, double-blinded, controlled clinical trials repeatedly failed to produce a benefit in functional outcome despite some improvement in angiographic vasospasm, more rigorous methods of neurobehavioral testing became critical to provide a more comprehensive evaluation of the functional efficacy of proposed treatments. While several subarachnoid hemorrhage studies have incorporated an array of neurobehavioral assays, a standardized methodology has not been agreed upon. Here, we review neurobehavioral tests for rodents and their potential application to subarachnoid hemorrhage studies. Developing a standardized neurobehavioral testing regimen in rodent studies of subarachnoid hemorrhage would allow for better comparison of results between laboratories and a better prediction of what interventions would produce functional benefits in humans.

Introduction

Subarachnoid hemorrhage (SAH) resulting from rupture of saccular intracranial aneurysms is associated with severe morbidity and mortality.¹ The incidence of aneurysmal SAH (aSAH) in the United States is estimated to be 9.7–14.5 per 100,000 person-year.^1–3 Several clinical trials over the past decade have been conducted to improve neurological outcomes in patients afflicted with cerebral vasospasm after aneurysmal SAH.^4–7 These studies were formulated on the basis of preclinical and clinical studies that showed efficacy in the reduction of arterial vasospasm and possible neuroprotective properties.⁸ Despite signs of some improvement in angiographic vasospasm, none of these trials showed a significant clinical benefit in neurological outcomes.^9,10 These disappointing findings have led to calls for an evaluation of the translational experimental designs in the hopes of enhancing the potential for clinical success. Among these initiatives, the use of neurobehavioral testing as an objective outcome measure constitutes an exciting opportunity. Preclinical studies directed toward demonstrating enhanced behavioral function will improve the predictive value of animal models of SAH for clinical efficacy of novel therapeutic agents.

Neurobehavioral tests have long been used to evaluate treatments for multiple diseases including neonatal hypoxic brain injury, ischemic stroke and traumatic brain injury in preclinical animal studies.^11–22 Functional recovery is the key end-point in clinical studies because morphological changes do not always correspond well with functional deficits. However, until recently similar testing was not widespread in experimental SAH studies. Outcomes in animal studies of SAH have typically relied on histological measurements and neurological scoring.^23–26 The severity and resolution of arterial vasospasm was typically measured on histological specimens by digital morphometry or vessel caliber was calculated from subtraction angiography images.^27,28 After randomized, double-blinded, controlled clinical trials such as those for the endothelin receptor antagonist Clazosentan and Tirilazad showed improvement in angiographic vasospasm but failed to produce a benefit in neurological function,^6,10,29 consideration for more rigorous methods of neurobehavioral testing became of critical importance to evaluate the benefit of the proposed treatments. While many recent preclinical studies of SAH have incorporated neurobehavioral testing, standardized methodology applicable to SAH is yet to be developed. This is in contrast to well-established outcome measures that are used in animal models of neurotrauma.³⁰ The goal of this review is to discuss different types of neurobehavioral tests and their relevance to rodent models of SAH.

General considerations

The cumulative injury caused by SAH is characterized by two distinct stages: early brain injury (EBI) and delayed cerebral ischemia (DCI). EBI refers to the acute brain injury following aneurysmal rupture.³¹ When blood is released into the subarachnoid space, intracranial pressure raises sharply, which results in decreased cerebral blood flow (CBF) and transient global ischemia associated with temporary loss of consciousness. Transient ischemia is also thought to cause endothelial cell death and blood–brain barrier (BBB) disruption, cerebral edema development, and neuronal apoptosis in the hippocampus, brain stem and cerebellum.³¹ At the molecular level, decreased NO and ATP, increased potassium, elevated endothelin-1 levels, persistent inflammation as observed by increased endothelial expression of cytokines and cell adhesion molecules, and activation of platelets, thrombin, matrix metalloproteinases (MMPs), and pro-apoptotic pathways are seen during the EBI phase of SAH.^31,32

Shortly after the EBI period concludes, DCI ensues. DCI encompasses angiographic vasospasm, cortical spreading ischemia, microthrombosis, and microcirculation constriction, among other mechanisms.³¹ Increasing evidence suggests that inflammation and, more specifically, leukocyte-endothelial cell interactions as well as peri-vascular and parenchymal inflammation, depleted NO, and increased endothelin-1 play a critical role in the pathogenesis of vasospasm after SAH.³³ Glutamate excitotoxicity and waves of cortical spreading depression are also implicated in seizure activity after SAH.³⁴ These injuries are thought to be responsible for the long-term sequelae of SAH.

SAH survivors often present with symptoms of depression, anxiety, and impaired cognitive function in addition to motor weakness, sensory changes, or speech difficulties.^35,36 Neurocognitive deficits studied in patients recovering from SAH also include attention and working memory deficits.^37,38 These types of neuropsychological deficits have been correlated with atrophy of temporomesial structures as well as atrophic enlargement of ventricles and sulci with voxel-based morphological studies using MRI.^39–41 Moreover, changes in hippocampal volume was very significantly associated with neuropsychiatric deficits one year after SAH.⁴¹ Although a multitude of scoring systems including Hunt & Hess (H&H), World Federation of Neurological Societies (WFNS), and Fisher grade have been used to predict patient outcomes, the impact of SAH on neurocognitive function is seldom rigorously assessed in routine clinical practice.

Although grading systems similar to the Fisher scale are being developed in animal models to evaluate the bleeding scale of induced SAH in vivo and ex vivo,^26,42,43 evaluating neurobehavioral deficits post-SAH induction with scales similar to H&H or WFNS is not easily accomplished. Extrapolating from human subjects, it can be proposed that animals with induced SAH dying within 24 h of injury likely represent the high grade SAH subjects and could be comparable to H&H 4–5 grade patients; animals surviving beyond 24 h after injury likely represent less severe hemorrhages and could be comparable to H&H Grade 1–3 patients. These animals with longer survival constitute an ideal target for neurobehavioral testing.

Better characterization of functional outcomes after SAH may lead to studies with a higher potential for translation into clinical settings. Neurobehavioral tests can provide more precise, accurate, and objective results that, in combination with traditionally used outcome measures, would enhance the quality of the preclinical data used to design further clinical trials.

Handling and behavioral testing methodology in animal models of SAH

Handling can drastically affect neurobehavioral testing results; therefore, proper handling is mandatory for accurate results.⁴⁴ Handling methods for mice and rats are similar, but there are subtle differences between the species, and methodology should be tailored to the species.⁴⁵ Several methods can be used to hold or restrain animals, all of which can have different effects on their anxiety levels.⁴⁴ The most commonly used method to capture or restrain a rodent is holding the animal by the tail. This method involves holding the base of the tail between the thumb and forefinger and lifting the animal while using the opposite gloved hand to support the body. Animals may also be restrained by holding the loose skin of the scruff.^35,44 Two other methods previously described involve the use of a tunnel apparatus or cupping of the hands. Tunnel handling involves guiding the animal into a plastic tube and then using the plastic tube to lift the animal out of the cage or testing apparatus.^35,44 With cupping, an animal is scooped up from the cage using one or both hands and allowed to sit or explore the palms of the hands for a period of time. Cupping is more appropriate for mice.⁴⁴ To prevent jumping behavior, an animal can be cupped loosely with both hands to habituate it to human touch for the first few handling sessions. Similarly, hands can be used to close both ends of the tunnel to prevent the animal from escaping during the initial handling period. Detailed description of housing, husbandry and handling can be found elsewhere.⁴⁵

Naïve laboratory animals generally seek to avoid human contact and are stressed by restraint. Several recent studies have shown that picking up by the tail is associated with increased aversion and stress levels compared to using a tunnel or cupping by open hand.^35,44 These latter modes result in lower stress levels, better acceptance and willingness of the animal to interact with the experimenter.^35,44 Therefore, handling the animals by more gentle methods is crucial to minimize stress levels and avoid its confounding effects. These methods are particularly important for cognitive tests.⁴⁵ Tests that measure anxiety levels such as the elevated plus maze have been shown to be directly affected by handling.³⁵ Although there are no reported data for the optimum duration of handling before the start of the experiment to achieve maximal acclimatization to the experimenter, minimum of 15 min of handling and 45 min of apparatus habituation and pretraining over the span of 2–3 days is considered acceptable.⁴⁵ It is also important to keep in mind that different strains of rodents respond differently to restraint and exhibit variable anxiety levels towards humans. Different strains also exhibit variable motivation and learning ability to perform behavioral tests.⁴⁵ In addition to proper handling methodology, gloves (and laboratory gowns/coats) may be rubbed with the soiled bedding of the animals to further reduce stress and help to habituate the subjects before each handling session.⁴⁴ Manual sanitation of the equipment using 70% ethanol or other disinfectant agents followed by either air drying or drying with paper towels can be used to clean the apparatus before each animal is introduced to the testing space to avoid distraction by the pheromones or excretions left by previous subjects.

Neurological severity scoring

Several neurological severity scales involve grading animals on points scored during spontaneous activity or after induction of cranial nerve reflexes by an examiner blinded to group identity. One of the most commonly used grading scales, developed by Garcia and colleagues for ischemic stroke in 1995, evaluates spontaneous activity and motor responses to stimulation.⁴⁶ It consists of a six-part test with a maximum score of 18 and its modified version for SAH can be used to evaluate neurobehavioral outcomes in rats and mice.⁴⁷ Components of the tests are listed in Table 1. We have reviewed the literature using the keywords “subarachnoid hemorrhage,” “neurological scoring,” “neuroscore,” and “neurological severity,” for mice and rats separately (Supplementary Tables 1 and 2).

Table 1.

Modified Garcia score.

Test	0 Points	1 Point	2 Points	3 Points
Spontaneous activity (in cage for 5 min)	No movement	Barely moves	Moves but does not approach 3 walls of cage	Moves and approaches ≥3 walls of cage
Spontaneous movement of all limbs	No movement	Slight limb movements	Moves all limbs slowly	Moves all limbs same as pre-SAH
Movement of forelimbs (outstretching when held by tail)	No outreaching	Slight outreaching	Outreach is limited and less than pre-SAH	Outreach same as pre-SAH
Climbing wall of wire cage	Falls from slope	Fails to climb	Climbs weakly	Normal climbing
Reaction to touch on both sides of trunk		No response	Weak response	Normal response
Response to vibrissae touch		No response	Weak response	Normal response

Neurological scores are thought to be most useful in the early EBI phase of SAH.⁴⁷ In our literature review, we found that endovascular perforation model was the most commonly used in mouse studies which used neurological scoring; however, the wide variety of point systems used makes the comparison between the efficacy of each point score test challenging (Figure 1). At 1 and 3 days after SAH, the 18-point score was the most widely used neurological score scale in rats with positive results in majority of the studies (Figure 2). Although neurological severity grading scales such as the modified Garcia score have proven useful as complementary tools when testing differences among treated groups in translational studies, inter-observer reproducibility has not been rigorously assessed, the extent of examiner bias is unknown, and the lack of objective, precisely measurable variables, pose significant questions about the validity of the scale.

Figure 1.

Results of SAH studies using neurological scoring within 72 h in mice. (+ sign denotes list of studies with positive results, − sign denotes list of studies with negative results).

Figure 2.

Results of SAH studies using neurological scoring within 72 h in rats. (+ sign denotes list of studies with positive results, − sign denotes list of studies with negative results).

Vestibulomotor function tests

Rotarod

The Rotarod test was first described by Dunham and Miya in 1957.⁴⁸ It provides quantitative assessment of motor function and has been used in experimental SAH mice by Mesis et al.⁴⁹ The apparatus consists of a rotating cylinder with larger disks spaced so as to create compartments for individual animals. The device is typically set to accelerate at a constant speed. Often the rotation speed is set to increase from 4 rotations per minute (rpm) to 40 rpm in 300 s. This setting can be adjusted, depending on the severity of the injury, to make the task harder or easier. Rotarod latency is defined as the time until the animal falls off the cylinder or clings to the rotating rod for two consecutive rotations. The latency of response is recorded before SAH injury as baseline performance and at discrete time points after injury. Up to three trials are usually performed. Absolute time spent on the rotarod or percent change from baseline performance, using the average of trial scores or best score, can be used for analysis. Factors such as the diameter of the rotating rod, compartment spacing, pre-training variables (such as frequency and duration of pre-training), testing environments (under light or dark conditions) and testing times in relation to the daylight cycle of the animal, testing dates and associated “learning effects” due to increased frequency of testing procedures, pose a challenge for comparing data across different studies. Because it is a less time-consuming procedure, the Rotarod has gained popularity and its use has been more frequently reported in the SAH literature.^49–57 Results of all the studies using the Rotarod in SAH-induced rodents are reviewed in Supplementary Table 3.

A lack of comparison between sham and SAH animals was found in a multitude of studies that utilized rotarod testing.^{49–52,55,58} In other studies where sham-operated and SAH animals were compared, significant differences were found in both short- and long-term survival. For example, in a short-term study using an endovascular perforation model, Wu et al. showed a significant difference in rotarod perforamance between sham and SAH mice at 24 h.⁵⁹ Using the same model, Sherchan et al. found significant differences in rotarod latency between sham and SAH mice 21 days after SAH in a long-term study.⁵³ Similar results were found at similar testing points for rats. Rats subjected to the rotating rod 14 days after SAH did significantly worse compared to shams using the endovascular perforation model.⁵⁴ In another study in which both prechiasmatic blood injection and double cisternal blood injection models were compared, the authors showed significant differences between sham and SAH animals at 24 h after SAH, but failed to show any differences 7 and 35 days post-SAH.⁵⁷ Moreover, they found no significant difference in rotarod latencies in SAH animals between the two models.⁵⁷ Siler et al. also failed to find significant differences between sham and SAH animals 4–6 days post-injury.⁶⁰ These findings suggest that in both rats and mice, significant differences in Rotarod performance was documented in the acute (24 h) and long terms (e.g., 14 and 21 days) but not in the mid-term (4–7 days).

Open field test

The open field test is used to assess locomotor exploratory and anxiety-like behavior in animals that are exposed to a novel environment. Using a video camera or infrared beams, an animal’s spontaneous locomotor activity is recorded typically for 5 min. Distance travelled, resting time, stereotypic time, ambulatory time, and time spent in the center of the box can be analyzed with accompanying software. Thigmotaxis, the natural tendency of animals to stay close to the periphery and avoid the center of novel environments, can be used as a measure of anxiety, in addition to spontaneous locomotion.⁶¹

Open field testing has been used in relatively few studies and provides an exciting opportunity for behavioral testing in SAH. Our literature search using the keywords “subarachnoid hemorrhage” and “open field test” revealed only two studies using rats and one study using mice. Although Hollig et al. failed to show significant differences in distance travelled and ambulatory time in SAH-induced rats 24 h after SAH induction using the endovascular perforation model,⁶² Boyko and colleagues reported significantly decreased locomotor activity, mean velocity, time and distance travelled in the central part of the field after a double-blood injection model in rats 3 weeks after SAH induction.⁶³ However, in the same study, no statistical difference in open field test between control and SAH animals was found using in the single-blood injection model of SAH 3 weeks after blood injection.⁶³ In an endovascular perforation model of SAH in mice, significant reduction in travelled distance was found in SAH-induced compared to sham-operated mice 96 h after SAH.⁵⁶ Future studies can help further establish the behavioral profile of SAH animals in open field testing in various testing points using different models.

Beam balance and beam walking test

The beam balance test quantifies motor and vestibular functioning in animals by scoring their balancing and/or traversing performance.⁶⁴ The animal is placed in the center of a wooden beam with its long axis parallel to the beam. The beam is elevated above a table, and a foam pad is positioned to protect the animals when they fall off the beam. Dimensions of the beam may vary. Scoring of performance is typically based on whether the animal can maintain balance and traverse the beam.⁶⁵ A scoring system of 0 to 2 can be used if both tasks are evaluated: 0 when the animal cannot maintain balance on the beam at all; 1 when the animal can maintain balance but not traverse the beam; 2 when the animal can maintain balance and traverse the beam.⁶⁶ To score the animals purely on balancing performance, scores are based on how long the animal is able to maintain balance, with points assigned to ranges of time.⁶⁵

Rats undergoing SAH via a cisternal blood injection model showed poor performance on the beam balance test on the first day after surgery before their motor performance improved.⁶⁴ In other studies, deficits in beam balance score and times were seen after cisternal blood injection on days 1 and 2.^67–69 Beam traversing times were shown to be higher on days 1 to 4 after SAH induction in rats.^67–69 There were no significant differences in the beam walking test in an endovascular perforation model of SAH 3, 7, 14, and 21 days after SAH in rats; however, the parameter used (% impaired steps) was different from that used in the previous studies.⁷⁰ The beam balance test offers flexibility in measurements that can make experiments conducted in different laboratories difficult to compare.

Other potential vestibulomotor tests

Grip strength and CatWalk gait analysis are other two well-established tests that can be utilized for assessment of vestibulomotor function in SAH-induced animals. Forelimb and hindlimb grip strength measurement is a technique first described in 1979⁷¹ and measures the maximal force an animal applies on a specially designed pull bar attached to a sensor. The dual sensor model allows the animal to grasp the pull bar with its paws and measures the maximum force attained when the animal is gently tugged back along a straight line leading away from the sensor until it releases the bar.¹⁹ This type of device has been used to assess forelimb neuromuscular function and integrity in experimental models of neuromuscular disorders including amyotrophic lateral sclerosis,⁷² transient ischemic attack,⁷³ permanent ischemic stroke¹⁹ and TBI⁷⁴ with variable results. The utility of this test in SAH models is not well established.

The CatWalk system is used to collect detailed gait analysis data rodents.⁷⁵ As the animal traverses a walkway, a video camera records light deflection by the paws. Only the points of contact on the walkway are illuminated and the intensity of illumination reflects pressure of the applied paws.⁷⁶ CatWalk software can further analyze up to 50+ parameters of gait including stride length, paw print intensity, paw print size, and swing phase (time between consecutive paw prints).⁷⁷ SAH animal models have not been thoroughly studied using the CatWalk gait analysis system. However, it has been used in various animal models of neurological disease such as neuropathic pain, spinal cord contusion injury, sciatic nerve injury, Parkinson’s disease, and ischemic stroke.^76,78–82 The CatWalk can provide valuable insights on altered gait in SAH animals and should be pursued further in SAH rodent studies. For blood injection models in which the femoral artery is used as a source of arterial blood, manipulation of the femoral nerve and possible hind limb ischemia induced with the blood withdrawal procedure should be kept in mind when evaluating hind limb deficits.

Cognitive tests

Morris water maze

Spatial learning, reference memory and working memory can be evaluated with a variety of tasks in rodents with SAH using the Morris Water Maze (MWM). This technique uses a circular pool of opaque water with a hidden platform available for escape. The platform is hidden by placing it just below the water surface. Opacity of the water is controlled with powdered milk or non-toxic white paint.⁸³ The hidden platform and opaque water provide no local visual cues for escape to the platform; thus the MWM specifically assesses navigation and working memory skills.⁸⁴ Hippocampal lesions impair initial post-operative navigation, further confirming the role of working memory in the task.⁸⁵

Latency to escape, velocity, and distance swum are the most common measures of performance in the MWM.⁸⁶ Animals are tested with repeated trials to track spatial learning over time. Travelling patterns can also be recorded by labeling the pool with North, South, East, and West markers and/or using a video camera placed above the pool. The platform can be moved between quadrants and may also be removed entirely to track free-swim patterns in what is known as a “probe trial.” Animals who perform well in the probe trial spend most of their time in the platform’s quadrant or swimming through or around the platform’s original location.⁸³

Spatial learning is best assessed with four trials daily with random start locations, while the hidden platform remains in the same location within the tank.⁸⁶ This task is known as spatial acquisition. To assess reference memory after spatial learning, a probe trial is given 24 h after the final spatial acquisition trial. As mentioned previously, no platform is present in the tank and the amount of time spent in each quadrant of the tank is the most important measure during this phase of testing. Changing the platform location each day and allowing two trials starting from the same location in the tank evaluates spatial working memory. The escape latency difference between the two trials is the main outcome in this type of testing.⁸⁶

Several studies report varying results in MWM performance following the endovascular perforation model of SAH. For example, Hu et al. showed significant deficits in SAH rats 17–21 days after injury, although a treatment effect with hyperbaric oxygen was not seen in MWM performance.⁸⁷ Using the same model, Sherchan et al. documented deficits between SAH and sham rats 21–25 days after injury.⁵³ In this same study, a treatment effect was seen with minocycline, which improved the outcome.⁵³ Similarly, using the same model, escape latency and distance swum was found to be increased for SAH-induced rats when the platform was moved to another location.⁸⁸ Although deficits in MWM performance are reported with rats in the endovascular perforation model, Milner et al. found no significant difference between sham and SAH mice 17–21 days after SAH induction.⁸⁹ It is not known whether the same model may produce deficits at different time points in mice. Moreover, no hemispherical or hippocampal volume changes were detected in SAH mice in a study by Atangana et al. using MRI, suggesting that current models of SAH may not induce significant radiological and functional deficits in hippocampus volume and neurocognitive tests respectively in SAH mice.⁹⁰ Further research is needed to improve surgical SAH induction methods and assessment of neurocognitive deficits at different time points.

Autologous cisternal blood injection induced significant deficits on escape latencies 4–5 days after SAH in rats, and this effect was significantly alleviated by dimethylfumarate. However, the same blood injection did not produce changes in cued learning tasks.⁹¹ Using the cisternal double injection model in rats, Feng and colleagues demonstrated significant cognitive deficits in rats 5 days after SAH induction, and ceftriaxone improved these deficits.⁹² Using the same model, Takata et al. showed increased latencies and swimming distances in SAH rats at later time points, specifically days 29–35.⁹³ A treatment effect was described with long-term simvastatin administration.⁵¹ In a double hemorrhage SAH model, simvastatin given for five weeks after surgery significantly improved escape latency and swimming distance despite a decrease in swim speed.⁹⁴ MWM performance has not been studied in cisternal blood injection models of SAH in mice.

A prechiasmatic injection model of SAH in rats significantly reduced MWM performance in two different studies on postoperative days 4–5 and only day 5, respectively.^95,96 Similarly, prechiasmatic injection to induce SAH did not induce deficits in cued learning, although deficits in spatial learning were evident 4–5 days after SAH.^97–99 A treatment effect with tamoxifen, tert-butylhydroquinone (tBHQ), an Nrf2 activator, and YC-1, a hypoxia-inducible factor 1 (HIF-1) inhibitor, were also observed on those days.^97–99 Sasaki et al. found a similar trend between the prechiasmatic injection model of SAH and the double blood injection model of SAH, reporting that in both experimental models, swimming speed was significantly increased and escape latency was decreased over time in rats tested 29–32 days after SAH induction.⁵⁷

Elevated T maze and elevated plus maze tests

The elevated T-maze is a three-armed apparatus used to analyze memory and anxiety in laboratory animals. In the apparatus, two arms are open, and 40-cm high walls enclose the third; the closed arm is connected perpendicularly to the open arms. The apparatus is elevated above the ground by 50 cm.¹⁰⁰ The elevated plus maze, on the other hand, has two open arms and two closed arms all perpendicular to each other.¹⁰¹ In both tests, the animals are placed in the center of the maze and allowed to explore it freely for a predetermined period of time. The animals can then be timed for their withdrawals from the ends of both open and closed arms and number of arm entries.¹⁰¹ The elevated T-maze presents an aversive experience to animals exploring the open arms because rats inherently fear openness and height.¹⁰² Inhibitory avoidance, representing conditioned fear, is measured when animals are placed at the end of the enclosed arm. Over trials, the animals learn inhibitory avoidance after repeatedly reaching the intersection of the maze and seeing the two open arms.¹⁰³ Animals placed at the ends of open arms escape to the closed arm. Unconditioned fear is assessed in the second task. The elevated T-maze differs from the elevated plus-maze because it distinctly separates two types of fear.¹⁰⁴ In healthy animals, withdrawal times from the closed arm is expected to increase over trials, while escape times from the open arms are expected to decrease over trials.

Inhibitory avoidance performance is reduced with the administration of the anxiolytics benzodiazepam, buspirone, ipsapirone, and ritanserin. The anxiogenic agents yohimbine, TFPP and mCPP facilitate inhibitory avoidance.¹⁰³ In a study using the endovascular perforation model of SAH in rats, animals were placed in the closed arm and timed on their choice of one arm. On day 21 post-surgery, intracranial pressure correlated with delayed decision in the elevated T-maze.¹⁰⁵ In another study by Sasaki et al., rats with SAH induced by prechiasmatic blood injection, showed decreased anxiety behavior on the elevated plus maze, whereas anxiety was increased in SAH rats undergoing cisternal double blood injection, suggesting that different models of SAH may induce different anxiety behavior in animals.⁵⁷ The elevated T-maze and elevated plus maze are often used to research psychiatric disorder treatments; however, they can also reliably evaluate anxiety and memory in animal models of SAH.

Other potential neurocognitive tests

The radial arm maze (RAM) is another neurocognitive test that can be used to assess neurocognitive function in experimental SAH. Olton and Samuelson developed the RAM test in 1976 as a method for assessing memory and spatial learning in rats.¹⁰⁶ With this device, both working and reference memory can be tested simultaneously with careful observation of rodent behavior during testing. To date, no studies have used the RAM to study the cognitive changes in rodent models of SAH. Briefly, the RAM consists of eight horizontal arms oriented around a central platform above the floor. Automated doors are located at the entrance of each arm and the rats or mice are placed on the central platform, from which they have to collect hidden food placed at the end of the arms. In the standard version of the RAM, animals are habituated to the environment by being placed on the central platform and allowed to explore the maze for 15 min each day. Food is scattered on the end of selected radial arms. On the last day of habituation (typically day 3), the number of food reinforcements is reduced to half, and the session ends when all eight arms have been visited. After this habituation period, the rats or mice are trained once daily for eight consecutive days. One piece of food is hidden at the end of each arm in a well, and the animal is allowed to freely explore the maze. Each session lasts until (a) all eight arms have been entered, (b) 10 min have passed since the start of the test, or (c) 2 min have passed since the animal’s last arm entrance. Arm entries are recorded for later analysis.

Sensory tests

Adhesive removal test

The Adhesive Removal Test (ART) was developed to assess sensorimotor asymmetry in rats with unilateral nigrostriatal damage.¹⁰⁷ Schallert et al. found that the animals quickly removed any mildly adhesive paper from their skin and that that latency to removal of the adhesive materials could be measured. They placed sticky tape bilaterally on the snout, forelimbs, and hind limbs of animals and measured latencies to contact the tape and to remove it. Bilateral comparison showed significant differences between contralateral and ipsilateral responses after unilateral 6-OHDA microinfusions.¹⁰⁷ The test has been used to investigate sensorimotor function in TBI and cerebral ischemia.^108,109 For ART testing, a testing box can be used to contain the animal. The animal is acclimated to the testing box for 60 s before applying the sticky tabs. The same experimenter should apply the adhesive each time to ensure that pressure during application does not vary. The animal is given 120 s to remove the tape and should be timed for latency to contact the adhesive and latency to remove it.¹¹⁰ The ART allows for observations to be recorded, with or without the use of video capture during testing, and also removes subjectivity from data collection with the latency measurement.

The literature search using the keywords “subarachnoid hemorrhage” and “adhesive removal test” has revealed a single study. In the study by Kooijman et al., sensorimotor behavior was significantly affected 21 days after induction of SAH via endovascular puncture in rats.⁸⁷

Possible applications of neurobehavioral tests in human clinical trials

The neurobehavioral tests described here have shown utility in different animal models and play a role in the preclinical research designs for functional outcome assessment for therapeutic drugs, especially for ischemic stroke and TBI. However, the methods of functional assessment for efficacy of a therapeutic agent in humans are far more simplistic and less precise than these strongly quantitative testing modalities in animals. The growing concern surrounding the failure of most of the clinical trials in fields of neurological diseases including SAH, calls for more precise outcome measures in humans.¹¹¹ A study from the SAHIT investigators discusses potential reasons for failure of randomized clinical trials in SAH, some of which could be related to problems in outcome measures and the multifactorial nature of delayed cerebral ischemia.¹¹¹ Other possible causes for failures in SAH trials are thought to stem from either the functional ineffectiveness of the tested therapies or the timing and dose of the treatment, inadequate sample size, insensitive or inappropriate outcome measures, the confounding effect of rescue therapies in placebo groups, treatment-associated side effects, and variations in practice across different centers.¹¹¹ Similarly, a recent review of the phase III progesterone trials suggests possible weaknesses in the testing parameters used to evaluate the efficacy of progesterone therapy following TBI, suggesting that outcome measures used in humans for neurological diseases need to improved.¹¹²

The Glasgow Outcome Scale (GOS), extended GOS (GOS-E) and Modified Rankin Scale (mRS) as well as incidence of delayed ischemic deficits and duration/severity of vasospasm are the most commonly used outcome measures in SAH.¹¹¹ However, many reviews question the clinical sensitivity and utility of the GOS, Disability Rating Scale (DRS), and the Functional Independence Measure (FIM).^113–115 Similarly, a ceiling effect is thought to occur with mRS, where most patients are in the favorable outcome group, thus making it challenging to show any improvement with treatment.¹¹¹

In the future, more sensitive tests for the evaluation of cognitive or neuropsychiatric outcomes as well as subtle testing tools for indices such as gait analysis, grip strength, and balance measures in humans should be used to better evaluate the efficacy of new therapies for debilitating neurological conditions such as SAH. Currently, gait analysis technology similar to that used in rodent models has been applied to human studies evaluating stroke patients.¹¹⁶ Grip strength and pinch strength have both been extensively evaluated in the follow-up care of stroke patients, and hand coordination testing has been applied to long-term follow-up of TBI patients.^117,118 Although not directly applicable at this time, correlates to many of the tests performed with animal models may help in the delineation of subtle benefits to new clinical therapies in human patients.

Although clinical trials are conducted for both male and female patient populations, preclinical studies are based overwhelmingly on male animals. Sex differences are known to affect the incidence, mortality and possibly also functional outcome after SAH in humans¹¹⁹ as well as the pathophysiology of SAH in rats.^120,121 However, there are very few studies the current literature investigating sex differences in functional outcome in experimental studies in SAH rodents using the neurobehavioral tests reviewed in this study.⁴³

Conclusions

Neurobehavioral tests can offer valuable information about sensorimotor and cognitive changes in experimental SAH models in rodents. These tests can also be used to evaluate the efficacy of therapeutic agents. The failure of the randomized clinical trials testing an endothelin receptor antagonist as well as tirilazad^5,29,122,123 supports the hypothesis that improvement in vasospasm alone, as an outcome measure, cannot predict functional improvement in patients.¹²⁴ Until recently, complex neurobehavioral tests to evaluate the effect of therapies on functional outcome have not been widely used in SAH. Studies to evaluate neurobehavioral profiles in different rodent models and implementation of these neurobehavioral tests in SAH preclinical research can improve the probability of translating functionally effective drugs to the bedside and help to ensure the success of future clinical trials.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Authors’ contributions

All authors certify that they have participated sufficiently in the work to take responsibility for the content, including participation in the concept, design, analysis, writing, revision or final approval of the manuscript. Study conception and design: Nefize Turan, Gustavo Pradilla; acquisition of data: Nefize Turan, Brandon A. Miller, Robert A. Heider, Maheen Nadeem; analysis and interpretation of data: Nefize Turan, Brandon A. Miller, Robert A. Heider, Maheen Nadeem; drafting of manuscript: Nefize Turan, Brandon A. Miller, Robert A Heider, Maheen Nadeem; critical revision: Brandon A. Miller, Iqbal Sayeed, Donald G. Stein, Gustavo Pradilla; final approval of the version to be published: Nefize Turan, Brandon A. Miller, Robert A. Heider, Maheen Nadeem, Iqbal Sayeed, Donald G. Stein, Gustavo Pradilla.

Supplementary material

Supplementary material for this paper can be found at http://jcbfm.sagepub.com/content/by/supplemental-data