Development and validation of prechiasmatic mouse model of subarachnoid hemorrhage to measure long-term neurobehavioral impairment

Controllable and reproducible animal models of aneurysmal subarachnoid hemorrhage (SAH) are crucial for the systematic study of the pathophysiology and treatment of this debilitating condition. Despite the variety of animal models of SAH currently available, attempts to translate promising therapeutic strategies from preclinical studies to humans have largely failed. This failure is likely due, at least in part, to poor replication of pathology and disabilities in these preclinical models, especially the long-term neurocognitive deficits that drive poor quality of life / return to work in SAH survivors. Therefore, there is an unmet need to develop experimental models that reliably replicate the long-term clinical ramifications of SAH – especially in mice where genetic manipulations are straightforward and readily available. To address this need, we developed a standardized mouse model of SAH that reproducibly produced significant and trackable long-term neurobehavioral deficits. SAH was induced by performing double blood injections into the prechiasmatic cistern – a simple modification to the well-characterized single prechiasmatic injection mouse model of SAH. Following SAH, mice recapitulated key characteristics of SAH patients including long-term cognitive impairment as observed by a battery of behavioral testing and delayed pathophysiologic processes assayed by neuroinflammatory markers. We believe that this new SAH mouse model will be an ideal paradigm for investigating the complex pathophysiology of SAH and identifying novel druggable therapeutic targets for treating SAH-associated long-term neurocognitive deficits in patients.


Introduction
Spontaneous subarachnoid hemorrhage (SAH) from rupture of an intracranial aneurysm leads to signi cant mortality and long-term neurological de cits, with ~ 50% of SAH patients suffering major cognitive de cits precluding their return to work or resumption of daily activities [1][2][3].Several methods for modeling SAH have been developed in an effort to provide an experimental foundation for mechanistic and therapeutic investigation in the preclinical setting.Most of these animal models recapitulate some, but not all, of the acute pathophysiological events that occur after SAH, including development of secondary brain injury from Early Brain Injury (EBI) and Delayed Cerebral Ischemia (DCI) [4].Less well characterized are the long-term consequences of SAH, especially its deleterious impact on cognitive function.
The most frequently used animal models of SAH are: (i) Endovascular perforation model, which involves advancing a suture into the internal carotid artery (ICA) until it perforates a vessel within the Circle of Willis leading to extravasation of blood into the basal cisterns [5][6][7][8][9]; and (ii) Direct injection of blood into the cisterna magna or prechiasmatic cistern [10][11][12], where the former results in a blood clot primarily localized around vessels of the posterior circulation and the latter around vessels of the anterior circulation [13,14].Each of these models has their own unique set of advantages and disadvantages, but none fully recreate the pathologic conditions of human SAH [15][16][17].One of the many reasons for this situation could be that rodent models of SAH are technically demanding and have been di cult to standardize for variability in hemorrhage severity [18].
Regarding long-term cognitive de cits after experimental SAH, a variety of induction methods, rodent species, and timepoints have been examined.To date, the most consistent and trackable neurocognitive de cits have been demonstrated in rat models of SAH including endovascular perforation [19], prechiasmatic injection [20], and cisterna magna injection [20,21] techniques.All three induction methods produce signi cant neurocognitive de cits in rats up to 1-month after ictus.However, rats are more expensive than mice and targeted genetic manipulation is signi cantly more challenging.In contradistinction, long-term neurocognitive de cits in mouse models of SAH have proven sign cantly more challenging.For prechiasmatic injection and cisterna magna injections mouse models, neurocognitive de cits have only been demonstrated up to 2-week after ictus (Table 1) [12,22,23].For the endovascular perforation mouse model, disparate results have been reported with Regnier-Golanov et al. reporting signi cant neurocognitive de cits 1-month after ictus using a battery of behavioral tests including Morris Water Maze [24], while Milner et al. showed no neurocognitive de cits 1-month after ictus as assessed by Morris Water Maze [25] and Fanizzi et al. showed minimal long-term neurobehavioral de cits 1month after ictus using a battery of behavioral tests including Morris Water Maze [26].Development of a mouse model of SAH that produces robust and consistent long-term neurocognitive de cits is highly desirable, as it would accelerate experiments, lower costs, and leverage the tremendous power of genetically manipulated mice.
To address this important investigational gap, we developed a "prechiasmatic double injection" mouse model of SAH in which heterologous arterial blood is injected twice into the prechiasmatic cistern 24h apart.Using this model, we performed a validation study using a battery of well-characterized behavioral tests to de ne long-term neurocognitive de cits (> 30 days post-SAH) and corroborated these de cits with chronic neuroin ammatory changes as assessed by immunohistochemistry.

Animals
All experimental procedures were carried out following the guidelines for the care and use of animals in research, which were established by the Institutional Animal Care and Use Committee at Washington University and approved by the National Institutes of Health.Male C57BL/6J mice between 12-14 weeks old were obtained from Jackson Laboratories (Bar Harbor, ME).The mice were provided ad libitum access to water and a standard mouse chow diet, and were acclimated to a controlled environment with a 12-hour light/dark cycle.The animals were randomly assigned to experimental groups.

SAH induction in mice
Experimental SAH was induced in mice using the prechiasmatic injection technique.Brie y, the mice were rst anesthetized with iso urane (2% induction, 1.5% maintenance) in a mixture of 70:30 N 2 O: O 2 , and their body temperature was maintained at 37°C with a rectal temperature probe and a thermo-regulated heating pad.The mice were then secured in a stereotaxic frame, and the scalp was incised to expose the surface of the dorsal skull.An intracranial pressure (ICP) probe was placed to monitor ICP during and after the procedure.A hole was made in the atlanto-occipital membrane using a 23G needle to insert the ICP probe.The probe was pulled slightly to ensure that it showed a pulsating curve ranging between 0-5 mmHg.
For SAH induction, an incision was made in the midline of the anterior scalp, exposing the transparent skull.The superior sagittal sinus was observed and avoided during SAH induction.A 0.7 mm burr hole was drilled in the skull 5 mm anterior to the bregma and slightly off the midline to avoid the sagittal sinus.At this point, 0.1 ml of fresh heparinized blood was collected from a littermate donor.A 27G spinal needle connected to a 1cc syringe was positioned with the side port facing up and angled 32° from the vertical axis.The needle was advanced carefully until contact was made with bone at a depth of about 7 mm, and then the needle was retracted 0.5-1.0mm.100µl of blood for SAH induction or arti cial cerebrospinal uid (CSF) for sham controls was injected through the spinal needle over 40-45s using a digital infusion pump.The needle was kept in place for an additional 2-3 min to prevent back ow of injected blood or CSF leakage, then slowly withdrawn.
The second blood injection was performed at the same site 24 h after the rst injection using the same protocol.Following surgery, any residual blood products or debris were cleared from the skull surface.The drilling point was sealed with bone wax, and the incision was closed with a sterile suture.The mice were allowed to recover from anesthesia in an incubator and then returned to their cages.
Mice were subjected to transcardial perfusion with heparinized PBS, and the brain was carefully removed to examine the blood distribution along subarachnoid vessels.No visible blood was observed in the sham-operated animals.In contrast, in the case of SAH, extravasated blood was consistently found covering the MCA and distributed along larger arteries that branch from the MCA.Any animals that died during or immediately after the surgical procedure, those from the SAH group without subarachnoid blood, those from the sham group with subarachnoid blood, and those with severe hemiparesis within 6h of surgery (indicative of inadvertent MCA occlusion) were excluded from the analysis.All SAH and sham-operated animals that survived and were able to complete neurobehavioral assessments were included in the analysis.

Morris Water Maze
The Morris water maze was performed in accordance with our previous studies [27][28][29].Brie y, two principal axes of the maze are the standard designation; with each line bisecting the maze is perpendicular to the another, dividing it into four imaginary cardinal poles: North (N) -opposite point of the experimenter, South (S) -experimenter's position, east (E)experimenter's right and West (W) -experimenter's left; thereby creating four quadrants.
A pool lled with opaque, non-toxic white paint was used for all experiments.An escape platform was placed in one of the quadrants of the pool beneath the water surface (0.5-1 cm), located halfway between the wall and the center of the pool.The water temperature was maintained at 22 ± 1°C to prevent the mouse from oating.Three cues of different shapes were placed around the water pool to help the mouse in nding the platform.Mice were randomly placed in the water, and an automated ANY-Maze Behavioral Tracking Software (Stoelting) was used to record the time taken to reach the escape platform (escape latency), swim speed, and total swim distance.The mice underwent four trials per day at each of the four cardinal drop points (north, south, east, west) in random order for eight consecutive days, with the platform hidden at the same location each day.If the mice failed to locate the escape platform within the 60-second trial length, mice were manually transferred to the platform for around 15 seconds.A single probe trial without a hidden platform was completed on the 9th day.To ensure no visual de cit due to SAH surgery, a visible platform test of the MWM was also performed.

Y-maze
The Y-maze test is a method used to evaluate spatial working and reference memory in rodents, as described in previous studies [27][28][29].Brie y, The test involves a Y-shaped maze consisting of three white, opaque plastic arms at a 120° angle.The mice were placed in the center of the maze and allowed to explore two arms for 10 min during the training trial.The third arm (novel arm) was kept closed (Fig. 4a,b).After a 30-minute inter-trial interval, the blocked arm was opened, and the mice were placed again in the center of the maze to explore all three arms for 10 min during the test trial.Over the course of multiple arm entries, the animal should display a tendency to enter the third arm that was less visited.The number of arm entries and triads was recorded to calculate the percentage of alternation.An entry is de ned as when all four limbs are within the arm, and the total number of novel arm entries by mice is then calculated.

Novel Object Recognition (NOR)
The NOR test is a widely used method for assessing the memory of rodents.The test is based on the innate inquisitiveness of rodents to explore novel objects and retain their spatial memory [28][29][30][31].The test was conducted in an open-eld arena, to which the mice were rst habituated.On the next day, four objects made of similar materials but differing shapes were placed within the arena, spaced equidistantly with an empty space for the mice to explore.The mice were then introduced into the arena with two identical objects and allowed to explore for 10 minutes.The mice were then removed, and one of the two identical objects was replaced with a novel object.Following a 10-minute inter-trial period, the mice were returned to the arena, and their exploration behavior for novel objects was recorded using a video camera for 5 min.The mice were considered to be engaged in object exploration when their head was oriented within 45° of an object and within 4 cm of it.
Rearing with the head oriented upward was also included if at least one forepaw was on the object.However, climbing over or sitting on the objects was not included.Finally, the percentage of exploratory preference was calculated.
Fear Conditioning (FC) FC test was conducted to evaluate amygdala-and hippocampus-dependent memory in rodents, as described in our previous publications [27,29,31].The test was carried out in an acrylic square chamber (33 × 33 × 25 cm) with a stainless steel rod oor that was connected to a shock generator for delivering a footshock.A tone stimulus was delivered through a speaker.Before conditioning, the chamber was cleaned using a 1% Virkon solution to mask any former odor cues.On the day of conditioning, mice were transported from their home cage to a testing room and allowed to sit undisturbed in their cages for 10 minutes.Mice were then placed in the conditioning chamber and allowed to explore for 2 minutes before the onset of the tone (20 seconds, 2000 Hz).In the delay conditioning procedure, a 2-second and 0.5 mA shock was given in the last 2 seconds of the tone duration.Mice recieved ve delayed conditioning trials, with 120 seconds intertrial intervals (ITI).
One minute after the last shock, the mice were removed from the conditioning chamber and returned to their home cages.After 24 hours, the tone test was conducted in a triangular chamber situated in a different room.The chamber was geometrically different from the conditioning chamber, thereby enabling the assessment of the tone conditioning in the absence of the training context.The chamber was cleansed with 70% isopropyl solution after every mouse.During the tone test, three 20-second tones were given after a 2-minute baseline period, with each tone separated by a 120-second ITI.The mice were removed from the triangular chamber 1 minute after the last tone presentation and returned to their home cages.
The freezing response was measured using a time sampling procedure, in which an observer scored the presence or absence of the freezing response for each mouse at every 2-second interval.Twenty-four hours after the tone test, a context test was conducted by placing each mouse back in the original conditioning chamber for 5 minutes.During this test, freezing was scored for each mouse at every 5 second interval.The data were transformed into a percentage freezing score by dividing the number of freezing observations by the total number of observations and multiplying by 100.

Immunohistochemistry
Immunohistochemistry was performed as previously described [6, 8, 27-29, 32].Free-oating brain sections with a thickness of 40-µm were subjected to uorescent immunohistochemical staining using the following primary antibodies: anti-mouse brinogen (ab34269, 1:1000 dilution), anti-Iba1 monoclonal antibody (MA5-27726, 1:1000 dilution), anti-P2RY12 monoclonal antibody (MABN2593, 1:1000 dilution), and anti-glial acidic brillary protein (GFAP) (IF03L, 1:1000 dilution).The secondary antibodies used included Alexa Fluor 488, 647, and 594 Goat Anti-mouse IgG.In brief, free-oating brain sections were washed with TBS and then blocked in TBS (prepared in 0.3% Triton-X 100 and 3% goat serum) for 2h.Subsequently, the brain sections were incubated in primary antibodies while being shaken at room temperature for 24h.Three 10-minute washes were performed after 24h, and the brain sections were incubated in secondary antibodies for another 24h.After incubation, the brain sections were washed three times for 10 minutes each.Next, brain sections were then mounted on slides using Vectashield H-1000 (Vector Laboratory), and the slides were sealed with nail polish.Finally, the samples were imaged using a Nikon Eclipse ME600 digital video microscopy system (Nikon Instruments Inc., Melville, NY), and the images were analyzed with Image-J software.

Statistical analysis
Statistical analysis was performed using GraphPad Prism 9.0 (GraphPad Software Inc., CA, USA).We estimated sample sizes using effect sizes based on our previous publications [27][28][29]31] and pilot experiments.The sample size for the twosample t-test was four to ve per group.In addition, we recommend having a backup animal in case of possible accidents during the procedure, for example, skull damage during ear bar mounting, which can cause intracranial bleeding that may affect the nal outcome of the experiment.An independent two-tailed t-test was used to obtain statistically signi cant differences between the groups.All data were tested for normality using the Shapiro-Wilk test.Results represented as mean ± s.d.Statistical signi cance was set at P < 0.05.

Results
A standardized rodent model that closely replicates human pathophysiology of SAH including reliable and trackable longterm neurocognitive de cits is important to develop for use in preclinical studies.In this context, a new and improved version of the previously described prechiasmatic blood injection mouse model of SAH is presented -referred to as the "double prechiasmatic injection" mouse model of SAH.The model entails two discrete injections (24 hours apart) of determined volumes of heterologous whole blood into the prechiasmatic cistern.This SAH model has a distinct advantage over currently used experimental mouse models of SAH in that it causes highly reproducible injury severity leading to robust neurocognitive de cits 1-month after ictus, along with trackable chronic neuroin ammatory changes.The prechiasmatic double injection model eliminates possible confounding variables by allowing control over the initiation, volume, and rate of hemorrhage with low mortality.It also allows for variation of blood volumes and injection rates to optimize resulting neurocognitive de cits across mouse strains, conditions, and laboratories.

Prechiasmatic SAH-induced long-term neurobehavioral impairment
In this mouse model of SAH, a battery of different neurobehavioural paradigms, such as Morris water maze (MWM), Ymaze, novel object recognition (NOR), and fear conditioning (FC) were used for the assessment of long-term neurological functions.In the MWM, independent t-test revealed signi cant difference (p < 0.01) in the latency to reach the platform on day 8 between sham and SAH mice, indicating the spatial learning and memory impairment in SAH mice (Fig. 2b).Furthermore, the outcome from the probe trial showed a loss in spatial memory retention in SAH mice, compared to sham animals, as concluded by signi cantly less time spent by SAH mice in target quadrant compared to sham group (Fig. 2c).
The Y-maze test was performed to test the cognitive function that assesses the natural exploration behavior of rodents.Our results indicated that SAH led to memory impairment in mice compared to sham controls as evidenced by signi cantly (p < 0.01) less time spent by SAH mice in novel arm compared to sham-operated mice (Fig. 3a,b).
The NOR test was performed to evaluate the perirhinal cortex circuitry, which showed that the brain region is compromised in this mouse model of SAH (Fig. 3c,d).Compared to sham-operated mice, SAH mice showed a signi cant decline (p < 0.05) in the discriminative index, indicating cognitive impairment after SAH.
Lastly, we performed the FC test, the aversive associative learning and memory assessment of sham and SAH-operated mice.The test concluded impaired tone and contextual conditioning in SAH animals.SAH mice showed signi cantly reduced percent freezing compared to sham controls in tone test, indicating an injury-induced impairment of amygdala-and hippocampus-associated memory function (Fig. 3e,f).

Prechiasmatic SAH-induced thromboin ammation
The SAH-associated microvessel thrombosis and in ammation are described well in the acute and subacute phases but haven't been reported in the delayed phase following injury to date, nor has it been evaluated for concurrence with microthrombi, neuroin ammation, and delayed neurological deterioration.We are the rst to establish this correlation of SAH insult in our prechiasmatic double blood injection mouse model of SAH (Figs. 2,3, and 4).While microvessel thrombi have been observed in endovascular perforation models of SAH due to downstream mechanisms of vessel rupture,[6, [33][34][35] the blood infusion SAH models offer better insight into the SAH-induced thrombo-in ammation post-ictus.However, no such vessel rupture is present in these experimental models of SAH.The potential cause of microvessel thrombi within the cerebral vasculature following SAH is sustained in ammation.Our prechiasmatic double injection model of SAH showed microthrombi formation (Fig. 4a,b), microgliosis (Fig. 4c,d), and astrocytosis (Fig. 4g,h) in the different regions of the SAH brain at day 30 after injury, indicating the sustained post-ictus in uence on histopathological and neurological outcomes.

Discussion
Long-term neurocognitive de cits are the primary driver of poor quality of life and low rate of return to work in patients who suffer SAH.New treatment strategies to combat the pathophysiological contributors to these long-term cogntivie de cits (including EBI and DCI) are desperately needed, and animal models that more fully recapitulate the pathology and disability seen in SAH patients would greatly aid investigators' efforts.To date, the lack of a mouse model of SAH that consistently and reproducibly causes trackable long-term neurocogntivie de cits across independent laboratories has been a barrier to identi cation of novel, effective, and translatable therapies.The rich toolset of genetically modi ed mice makes this species particularly attractive for such studies.
In the present work, we sought to improve SAH disease modeling by modifying the previously reported pre-chiasmatic injection mouse model.The major modi cation was to inject heterologous blood into the prechiasmatic cistern in two stages, where the second blood injection was performed 24 h after the rst injection.We then conducted a series of neurobehavioral assessments and found that mice with SAH displayed several signi cant impairments.These impairments included (1) spatial learning and memory impairment in the MWM test, (2) a decline in cognitive ability in the Y-maze test, (3) a compromised recognition memory function in the NOR test, and (4) a signi cant decrease in memory function in the FC test, as indicated by a decrease in the percentage of freezing to tone and contextual conditioning compared to the sham group.
We also discovered that the neurobehavioral de cits present one month after SAH were strongly associated with chronic in ammation.This association was demonstrated by the percentage area immunostained by Iba-1 (microgliosis) and GFAP (astrocytosis).Additionally, we observed persistent microvascular thrombi after double injection SAH, suggestive of impaired brinolytic activity.In total, our ndings suggest our prechiasmatic double injection mouse model of SAH could be an important new experimental tool towards identifying and addressing the underlying mechanisms of SAH-induced neurocognitive dysfunction.
Our model has distinct advantages to other available mouse models of SAH.For single blood injection models into the prechiasmatic cistern or cisterna magna, neurocognitive de cits have only been observed up to 2-week after ictus (Table 1) [12,22,23], which limits translatability to the human condition.For the endovascular perforation mouse model, disparate results have been reported with one laboratory noting signi cant neurocognitive de cits one month post-ictus [24] and two other laboratories showing no or minimal neurocognitive de cits one month post-ictus [25] [26].Important advantages of our mouse model of SAH compared with existing methods are summarized in Table 1.

Conclusion
The proposed prechiasmatic double injection mouse model offers a promising means for testing the impact of genetically modi ed strains on the most important endpoint in SAH -long-term neurocognitive de cits.This model offers ease of application, reliability, and high spatiotemporal control of the injected blood volume, and it permits fair comparisons between experimental groups by reducing variability.If replicated by independent laboraties, this mouse model could become the standard for studies examining the pathophysiology of SAH and evaluating potential therapeutic avenues to improve long-term neurobehavioral outcomes.The implementation of such a model would, therefore, provide a valuable translational tool for advancing our understanding of SAH and its treatment.

Declarations
Tables Table 1 is available in the Supplementary Files section.
Overview of the protocol for developing and assessing prechiasmatic double blood injection mouse model of subarachnoid hemorrhage (SAH).The protocol includes preparation, pre-and post-procedure management of animals, behavioral assessments at certain time points post-SAH, and nally, euthanasia and injury-induced thromboin ammation analysis.