Modified Intracranial Aneurysm Rupture Rat Model with Angiographic Imaging

Abstract

Intracranial aneurysms (IAs) are a major cause of spontaneous subarachnoid hemorrhage (SAH) and are associated with high morbidity and mortality. Current IA rodent models often exhibit low rupture rates and limited imaging capabilities, restricting their translational utility. This study introduces a modified elastase-based rat model that incorporates angiographic imaging to overcome these challenges. IAs were induced in 7-week-old female Sprague–Dawley rats using a combination of surgical and pharmacological interventions, including carotid artery and renal artery ligation, bilateral ovariectomy, high-salt diet, and two elastase injections into the basal cistern. Digital subtraction angiography (DSA) was employed to assess aneurysm formation and rupture rate. Histological and immunohistochemical analyses were conducted to characterize aneurysm morphology and the inflammatory response. The modified model achieved a high rate of IA formation (85%) and rupture (60%) within 28 days. DSA enabled visualization of vessel tortuosity and flow dynamics, features relevant to human IA development, which often occurs in areas subjected to hemodynamic stress, and the tortuosity of intracranial vessels affects their rupture ^[1]. Histological analysis indicated structural degradation of the aneurysm wall, while immunohistochemistry showed neutrophil infiltration, potentially implicating inflammation in IA rupture. This improved IA model offers a reliable method for inducing and visualizing IAs with a high rupture rate, making it a valuable tool for studying the pathophysiology and therapeutic interventions of IAs. Enhanced by DSA, this model has the potential to advance therapeutic research by enabling the real-time monitoring of aneurysm development and rupture.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12975-025-01366-w.

Introduction

Intracranial aneurysms (IAs) rupture accounts for 85% of spontaneous subarachnoid hemorrhages (SAHs), with a global incidence of 9.1 per 100,000 individuals annually. This rate is even higher in countries such as Finland (19.7 per 100,000) and Japan (22.7 per 100,000) [2]. Among those affected, 8.3% experience fatal outcomes at onset without receiving medical care, and a significant proportion of these are young individuals. Even with timely treatment, early mortality remains at 12.1%, and 29.7% are left with severe morbidity (modified Rankin Scale > 2) [3].

Despite the catastrophic consequences of IA rupture, the current preventative strategies are limited to surgical intervention. The advancement of pharmacological treatment modalities beyond surgical intervention is greatly sought after, and the need for a suitable animal model of the disease has resulted in the development of numerous IAs animal aneurysm rupture models [4, 5]. However, a significant number of these models suffer from critical limitations, including low rupture rate and unpredictable timing of aneurysm development, which impede their translational utility. Furthermore, current rodent IA models rely on MRI imaging followed by histological confirmation, which still has disadvantages owing to its relatively low resolution, inability to visualize small vessels and small IAs, prolonged examination time, and limited availability of MRI [8, 9]. Although conventional angiography has been applied in rodent IA models, immediate euthanasia precludes proper analysis using this procedure [10, 11].

This study aimed to establish and validate a novel rat model of IA rupture by integrating surgical manipulation, elastase injection, and angiographic evaluation to enhance our understanding of IA pathophysiology and therapeutic interventions.

Methods

Study Approval of Animal Experiments

All animal experiments were conducted in accordance with the National Institute of Health’s Guide for the Care and Use of Laboratory Animals and complied with the ARRIVE guidelines (https://arriveguidelines.org). Approval was obtained from the Institutional Animal Care and Use Committee of Jikei University School of Medicine (Approval Number #2023–004). We performed the experiments with the utmost recommended care to minimize distress and any incidental suffering, and particular care was taken for animals with neurological deficits.

Rodent IA Models and Histological Analyses

Seven-week-old female Sprague–Dawley rats (Slc: SD; Japan SLC, Shizuoka, Japan) were used in this study. Female rats were selected based on previous reports indicating that sexual maturation begins at approximately 6 weeks of age [12, 13], and based on our previous findings that rupture rates are higher in female rats [4, 6]. Animals were maintained under a 12-h light/dark cycle with ad libitum access to food and water.

To induce IA, animals were subjected to a series of surgical and pharmacological interventions under general anesthesia with isoflurane (induction, 5.0%; maintenance, 1.5 ~ 2.0%; #IYESC-0001, Pfizer Inc., New York, NY) and local anesthesia using 0.5% Xylocaine® (lidocaine hydrochloride; Sandoz, Tokyo, Japan). The surgical procedures included bilateral ovariectomy [14], ligation of the left common carotid artery (CCA), right external carotid artery (ECA), and right pterygopalatine artery (PPA). Systemic hypertension was achieved by the combination of a high salt diet and the ligation of the left renal artery [6]. Postoperatively, animals were fed chow containing 8% NaCl and 0.12% 3-aminopropionitrile (BAPN; Tokyo Chemical Industry, Tokyo, Japan), a lysyl oxidase inhibitor that catalyzes the cross-linking of collagen and elastin. A total of 100 milliunits of elastase dissolved in 20 µL of phosphate-buffered solution was stereotactically injected into the basal cisterns on postoperative day (POD) 7 and 14, using coordinates from the Rat Brain Atlas (6.5 mm anterior to bregma, 30° anterior inclination, 9.0 mm depth from the skull surface). Rats that died before the second injection were excluded from the analysis. The observation period concluded on POD28. Animals were included for analysis if they (1) exhibited no spontaneous response, (2) showed obvious neurological symptoms such as hemiparesis and unilateral circling, and (3) were still alive at POD28. At the endpoint, rats were sacrificed using isoflurane overdose. All deceased animals underwent autopsy to examine the onset of SAH. If SAH was observed, the ruptured IA lesions were harvested for detailed analysis. The circle of Willis was microscopically dissected for both SAH-positive and SAH-negative cases from the brain surface with eventual IA lesions for detailed examination. Histopathological examination of the dissected structures was performed using Elastica van Gieson staining.

In this procedure, rats that died because of procedural complications were excluded, resulting in a mortality rate of approximately 13% of all animals used.

In the preliminary study, 14 rats were used to determine the optimal number of elastase injections. Subsequently, 23 rats were included in the main experiment; therefore, a total of 37 rats were included in this study. In the preliminary phase, animals were randomized into three groups, each receiving different elastase injection protocol. This randomization was performed to minimize allocation bias. The primary aim of this exploratory experiment was to assess the feasibility and effects of injection number on IA formation prior to the main experiment.

Sample Size Calculation

An a priori power analysis was conducted using EZR software to determine the minimum number of animals required to detect a statistically significant difference in aneurysmal rupture rates between the original and modified models. The rupture rate of the original model (5.7%, 2 out of 35 animals) was based on previously reported data [4, 6]. Assuming a rupture rate of 60% in the modified model, a two-sided significance level of 0.05, and a desired power of 0.8, the minimum required sample size was calculated to be nine animals per group using Fisher’s exact test.

To enhance the reliability of the results while minimizing unnecessary animal use in accordance with ethical principles, 20 animals were included in the modified model group, slightly exceeding the calculated requirement.

Animal Angiography

Animals from POD21 to POD28 were evaluated by angiography using X-ray fluoroscopy (Artis zee, Siemens, Germany), primarily via the tail artery [15]. If the tail access failed, the right femoral artery was used as an alternative.

For tail access, animals were pre-warmed using a hair dryer and heating pad to induce vasodilation of the caudal ventral artery, as previously described [16, 17]. A 24-gauge venous indwelling needle (SR-OT2225C, TERUMO, Tokyo, Japan) was inserted in the caudal ventral artery 6–10 mm caudal to the tail root. After confirming backflow, the needle was removed by placing a 0.4-mm guidewire (FGW16-AG18S30, Toray Medical, Tokyo, Japan) and directly advancing it into the abdominal aorta under fluoroscopic guidance. The 24-gauge sheath was then exchanged in-line for a 22-gauge sheath containing a microcatheter (inner diameter (ID), 0.42 mm; outer diameter (OD), 0.55 mm; Kaneko Cord, Tokyo, Japan), while maintaining the guidewire position to prevent arterial damage.

If the femoral artery approach was required, a 2-cm incision was made on the right thigh under local anesthesia, exposing and clipping the femoral artery. After applying 1% lidocaine to induce vasodilation, a small sidewall incision was created, and a 22-gauge sheath with a microcatheter and guidewire was inserted and secured using a 6–0 Prolene loop suture. In total, two rats were subjected to the femoral artery approach.

Following the placement of the microcatheter and guidewire in the artery using either approach, the microcatheter was advanced into the right CCA or one of the vertebral arteries (VA), under fluoroscopy guidance. Once in position, the guidewire was gently withdrawn to avoid microcatheter displacement. A 0.5 mL bolus of contrast agent (Iohexol, 300 mg iodine/mL; Daiichi Sankyo, Tokyo, Japan) mixed with heparin (1 mg Iohexol: 10 units heparin) was injected rapidly to visualize the cerebral arterial system. The C-arm was adjusted to left anterior oblique (LAO) angles of 0°, 45°, and 80° to allow comprehensive imaging of vessel morphology, including arterial expansion or tortuosity.

To minimize radiation exposure, fluoroscopy time was kept to a minimum, and low-dose imaging settings were used during both catheter placement and contrast injection.

Elastica van Gieson Staining

For Elastica van Gieson staining, tissue sections were first deparaffinized using graded ethanol series. They were then sequentially incubated with hydrophobic Resorcin-Fuchsin solution to visualize the elastic lamina, Weigert’s iron hematoxylin solution to stain nuclei, and Fuchsin Acid Van Gieson solution to visualize collagen and extracellular matrix. Following staining, sections were dehydrated and mounted using a xylene-based mounting medium.

Immunohistochemical Analysis

After sacrifice, 5-µm-thick frozen paraffin-embedded sections were prepared from dissected brain specimens. After blocking with 3% donkey serum (#AB_2337258; Jackson ImmunoResearch, Baltimore, MD), slices were incubated with primary antibodies, followed by fluorescent dye-conjugated secondary antibodies (Jackson ImmunoResearch). Fluorescence images were acquired using a confocal fluorescence microscope (LSM880; Carl Zeiss, Gottingen, Germany).

The primary antibodies used were Cy3-conjugated mouse monoclonal anti-α-smooth muscle actin (SMA) antibody (#C6198, Sigma-Aldrich, St. Louis, MO) and rabbit monoclonal anti-myeloperoxidase (MPO) antibody (#ab9535, Abcam, Cambridge, UK).

The secondary antibody used was Alexa Fluor 647-conjugated donkey anti-rabbit IgG H&L (#A31573; Thermo Fisher Scientific).

Results

Effective Elastase Dose Establishment

To confirm the precision of the basal cistern coordinates applied in the procedure, Evans blue dye was injected and its distribution in the subarachnoid space and vasculature was visualized (Supplementary Fig.  1 A). A pilot study was then conducted to compare the effect of different elastase injection frequencies (one, two, or three times) on aneurysm induction and rupture rates in rats. An elastase injection volume of 100 milliunits in 10 µL was determined based on a previous mouse elastase model [18] modified according to the interspecies differences in brain weight and CSF volume [19]. We administered one injection (n = 5), two injections (n = 6), and three injections (n = 3) at a 7-day interval between each injection to minimize the effect of acute intervention-related mortality. Within 30 days post-surgery, the SAH rates were 0%, 50%, and 33% in the one, two-, and three-times injection groups, respectively. Therefore, the two-injection protocol was selected for the main study (Supplementary Fig. 1B). In the pilot study, a dose-dependent increase in rupture rate was not observed. This may be due to a plateau in elastase enzymatic activity with repeated administration, or to arachnoiditis around the optic chiasm cistern caused by repeated punctures, which may prevent the full amount of elastase from being delivered.

Digital Subtracted Angiography and Flow Study of the Animal Model

After the surgical procedure, immediate digital subtraction angiography (DSA) was performed to visualize the IA system, evaluate blood flow, and confirm the ligation of the left CCA. Since the right PPA and ECA were also subjected to ligation, a clear blood flow from the CCA to the internal carotid artery (ICA) on the right side without additional branch supply was confirmed. Territories originally supplied by the left CCA were instead perfused via collateral circulation by the right CCA, which redirected them toward the ACA territory. When the microcatheter was relocated to the left VA, the basilar artery (BA) and the posterior portion of the circle of Willis became visible. Additionally, distal perfusion of the left PPA was observed as contrast flow detoured into the left CCA towards the ligation level (Fig. 1B). The names and abbreviations of each vessel are shown in Fig. 1.

Fig. 1.

The vessel flows after establishing the model. A The anatomy of the intracranial vessels. B Digital subtraction images of the intracranial vessel immediately after establishing the model. The images in the upper line represent the anterior circulation and the images in the lower lines represent the posterior circulation. C A schematic of the blood flow. Abbreviation: ACA, anterior cerebral artery; BA, basilar artery; ICA, internal cerebral artery; PCA, posterior cerebral artery; Pcom, posterior communicating artery; PPA, pterygopalatine artery; VA, vertebral artery

Incidence of IA and Ruptured IA

Twenty-three rats were assigned to the study, with three rats excluded due to mortality before the second elastase injection on POD14 (one expired during the procedure and two expired soon after the first injection). Of the remaining 20 rats that met inclusion criteria, 8 expired before the observation cutoff point, two suffered from significant neurological deficits and were sacrificed for analysis, and 10 survived until POD28. SAH was grossly identified at autopsy in 12 rats, representing 60% of the included animals (Supplementary Fig. 2).

Histopathological analysis revealed a total of 20 aneurysms (unruptured and ruptured IAs) on 17 animals (85%) were identified. Among these, 10 aneurysms (50%) were identified around the anterior portion of the circle of Willis (including the right ACA, left ACA, and common ACA), 4 (20%) were identified around the posterior portion of the circle of Willis, 3 (15%) were discovered at the right MCA branch, and 3 (15%) were at the right ICA/posterior communicating artery (PcomA) junction (Fig. 2A).

Fig. 2.

Number and anatomical distribution of unruptured and ruptured IAs. The animals were maintained for 4 weeks after establishing the model or autopsied if they died during the observation period. The upper panel shows the number and location of the unruptured IAs. The lower panel shows the number and location of the ruptured IAs. Abbreviation: ACA, anterior cerebral artery; BA, basilar artery; ICA, internal cerebral artery; PCA, posterior cerebral artery; Pcom, posterior communicating artery; PPA, pterygopalatine artery; VA, vertebral artery

Of the 12 animals with SAH, histological rupture point was identified in the anterior portion of the circle of Willis in six (50%) and two animals (16.7%) each for the posterior circle of Willis, right MCA, and right ICA/PcomA junction (Fig. 2B). In comparison of rupture rates by IA location, 60% (6/10) of aneurysms were identified around the anterior portion of the circle of Willis, 66% (2/3) of aneurysms were identified around the posterior portion of the circle of Willis, and 66% (2/3) of aneurysms were identified at the right MCA branch. No statistically significant association was observed between the rupture site and anatomical location. From the perspective of surgical invasiveness, multiple DSA procedures were not feasible, and imaging data were collected at a single time point. Consequently, time series analysis to determine the timing of IA formation and its progression to rupture was not possible. Although no significant association was found between aneurysm rupture and its anatomical location, this may be attributable to uneven elastase administrations across all cerebral vessels following repeated injections.

Digital Subtraction Angiography Study of the IA

Eight animals underwent DSA imaging, with six of them via ventral caudal artery and two via femoral approach. When compared to images of a healthy rat IA system, immediate identification of aneurysms after DSA was possible in five animals, with berry-like bulging out from the parent artery that resembled human saccular IAs on DSA images (Fig. 3). By POD28, autopsy analysis confirmed aneurysms in seven of the eight animals that underwent DSA, five of which were associated with SAH caused by IA rupture. In all seven animals with IA, a severe tortuosity pattern was observed in the circle of Willis (Supplementary Fig. 2).

Fig. 3.

The comparison of blood vessels in this model and a healthy rat. The animals were maintained for 4 weeks after establishing the model or autopsied if they died during the observation period. The upper panels show the right anterior circulation in this model (right panel) and in healthy rats (left panel). The lower figures show the posterior circulation of this model (right panel) and healthy rats (left panel). Abbreviation: ACA, anterior cerebral artery; BA, basilar artery; ICA, internal cerebral artery; PCA, posterior cerebral artery; Pcom, posterior communicating artery; PPA, pterygopalatine artery; VA, vertebral artery

Histological Assessment of Intracranial Aneurysms and Their Rupture

Light microscopic examination using Elastica van Gieson staining revealed histological characteristics similar to those observed in human IA, including a lack of elastic lamina and thinning of the smooth muscle layer. Microvessels (vasa vasorum-type) and inflammatory cells were observed near the rupture point (Fig. 4A).

Fig. 4.

Histological and immunohistochemical analysis of ruptured intracranial aneurysms. The upper lines show images of Elastica van Gieson staining of the ruptured IA (right panel). This image shows the lack of elastic lamina in the aneurysmal wall and inflammatory cells infiltrating the wall. Scale bar; 50 µm. In the lower panel, immunohistochemical images show that these inflammatory cells were myeloperoxidase-positive and appeared to infiltrate the vasa vasorum. Scale bar; 50 µm. Abbreviation: SMA, smooth muscle actin; MPO, myeloperoxidase

On immunohistochemical examination, accumulation of MPO-positive cells was observed in ruptured IA lesions. There were also some vasa vasorum with SMA-positive cells, specifically around the rupture site (Fig. 4B).

Discussion

We used rupture models that did not include elastase injection to investigate the mechanism of rupture, but low rupture rates and prolonged rupture timelines significantly reduced experimental efficiency [4, 20]. In contrast, this model effectively addresses the challenges of low rupture rates and prolonged rupture timelines without significantly increasing mortality (three animals were excluded owing to technical complications). As a result, the rupture rate improved compared to existing models, and the time to rupture was shortened. Therefore, this model holds promise for evaluating therapeutic agents aimed at preventing IAs rupture. In this model, we incorporated a specialized diet containing 3-aminopropionitrile, an irreversible inhibitor of lysyl oxidase that catalyzes the cross-linking of collagen and elastin. In human IAs, the elastic lamina is deficient or absent in the walls. Aoki et al. described the disruption of the elastic plate as an initial step in the development of IAs [6]. Therefore, this diet contributes to an increased incidence of IAs. Moreover, elastase likely promotes disruption of the elastic lamina, leading to an increased incidence of IAs, which in turn increases the rupture rate. Our model differs from previous techniques [21–23] that combined surgical intervention with elastase injections. Instead, it directly modifies blood flow by ligating the left ECA and PPA. This ensured that the unilateral CCA solely supplied the circle of Willis, inducing focused hemodynamic stress within the IA system. Another advantage of our model is its use of rats which have brains approximately four times larger than those of the mice [19]. This allowed a more detailed image of the circle of Willis without compromising the merits of murine animal models. Notably, our model exhibited a similar distribution of IA locations to the observed in human studies particularly, a relatively high portion of IA formation(50%) in the ACA region, consistent with clinical reports [26, 27]. Additionally, we also added ovariectomy procedures to our model. The Japanese database of SAH in humans suggests that postmenopausal women are more prone to developing SAH [28, 29]. As ovariectomy induces a menopause-like state in rats and this may also contribute to an increase in rupture rates.

Additionally, angiographic visualization showed enlarger arteries in experimental rats compared to healthy controls, suggesting arterial expansion and increased blood flow caused by elastase administration. This flow modification likely contributed to the elevated wall shear stress (WSS), which could have played a significant role in both the incidence and anatomical distribution of IA formation within the circle of Willis in this model [30]. Thus, increased blood flow resulting from arterial expansion offers a promising advantage for future applications of this model in studies focusing on flow-related mechanisms and interventions. Notably, all seven animals confirmed IA exhibited severe vessel tortuosity in the circle of Willis. This finding is consistent with previous studies demonstrating a strong correlation between vessel tortuosity and the presence of IA [8]. Tortuosity has been proven to promote matrix metalloproteinase (MMP) activation and induction of a series of inflammatory response mechanisms due to the accumulation of blood cells in relatively low WSS areas [31, 32], potentially accelerating IA pathogenesis in our model. Furthermore, Li et al. reported that the vascular tortuosity of ruptured IAs was significantly higher than that of unruptured IAs in humans [1]. This result suggests that our model replicates certain aspects of the pathophysiology of ruptured brain aneurysms in humans.

In addition to WSS, recent advancements in IA research have highlighted the significant role of various microenvironmental factors in the process of IA formation and rupture, including inflammatory factors [18, 34, 35] and vasa vasorum development under hypoxic conditions [6, 20]. In our model, SMA-positive vasa vasorum were observed near the rupture site, and MPO-positive neutrophils were detected adjacent to the aneurysm wall. These findings are consistent with prior reports from the original model, which indicated that vasa vasorum development and neutrophil infiltration increased as rupture approached. Using this model, we obtained results similar to those of a previous study, confirming that the phenotype of the previous model was maintained [4, 6]. These results emphasize the potential of this innovative model to contribute to future investigations at the microenvironmental level, providing insights into the pathophysiology of IA and identifying potential therapeutic targets.

To the best of our knowledge, this is the first study to investigate blood flow via angiography after surgical ligation in IAs rodent models. This evaluation method has great potential for future applications in IA-related studies. Screening of each animal for IAs was achievable (by DSA) in our method, unlike previous rodent model that relied solely on autopsy and histopathological analysis for IA confirmation [18, 21]. Furthermore, the dynamics of blood flow through the unilateral CCA and its balance with the intact bilateral VAs have not been well discussed until now; therefore, the visualization of blood flow contributes to a more comprehensive understanding of the arterial changes in this IA model.

The percutaneous puncture with the 24-gauge sheath and indwelling needle via the ventral caudal artery, followed by switching to a 22-gauge sheath with intra-arterial guidewire placement, represents a minimally invasive refinement of the “Ohta method” [16]. This in-line exchange technique proved to be effective and minimally traumatic. Its future application may include more frequent follow-up angiography to understand the role of blood flow modification in IA genesis, precise and targeted drug administration, and even micro-scale endovascular interventions if we do further development in our equipment.

Despite the valuable findings demonstrated by our model, several limitations warrant further investigation. First, we lacked sufficient data to determine whether there was a dose-dependent effect of elastase, highlighting the need for future studies to explore this aspect. Additionally, establishing an optimal observational timepoint for assessing IA genesis and rupture could enhance the utility of this model. Although redirected blood flow through surgical ligation significantly increased aneurysm incidence, no clear association was observed between the location of aneurysm formation and rupture site. This suggests that the relationship between the diverted blood flow and rupture in our model remains unclear. Furthermore, the link between elastase injection and the natural process of IA genesis has not been fully demonstrated, raising concerns about the potential to overemphasize certain mechanistic pathways. Another concern involves the stress caused by surgery may increase systemic inflammatory levels. However, based on our initial reports using sham-operated controls [4, 6], no ruptures or similar complications were observed, suggesting that the impact of surgical invasiveness was likely minimal. These limitations should be carefully considered in future studies, particularly for the clinical translation and development of therapeutic applications.

Conclusion

Our newly developed model, which integrates surgical blood flow modification, dual elastase injections, hypertension induction via renal artery ligation, a high-salt and BAPN-enriched diet, and ovariectomy, presents a promising model design for IA research with a relatively high incidence and rupture rate of IA and improved blood flow understanding without attenuating the pathological features of IA. The introduction of DSA via a modified caudal ventral artery approach enables precise, real-time visualization of IA without sacrificing the animal, allowing for multiple imaging attempts, better detection of aneurysm progress, and increasing the reproducibility and reliability of the results. Potential future applications of this model and the study protocol could improve the efficiency and translational relevance of preclinical aneurysm research.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contribution

W.L.P. and M.I. are cotributed equally. Author’s contributions W.L.P., M.I., K.K., H.K., T.A., and Y.M. planned the experiments. M.I., T.K., H.O., T.A, and Y.M., prepared the compounds. W.L.P., M.I., A.O. and Y. M. acquired the data. W.L.P, M.I., K.K. and T.A. interpreted the data. W.L.P., M.I., and T. A. wrote the manuscript. K.K., T.K., H.O., H.K., H.J.O, Y.A., T.A., and Y.M., critically reviewed and modified the manuscript. All authors read and approved the final manuscript.

Data Availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.

Consent for Publication

Not applicable.