Site-specific elevation of interleukin-1β and matrix metalloproteinase-9 in the Willis circle by hemodynamic changes is associated with rupture in a novel rat cerebral aneurysm model

Abstract

The pathogenesis of subarachnoid hemorrhage remains unclear. No models of cerebral aneurysms elicited solely by surgical procedures and diet have been established. Elsewhere we reported that only few rats in our original rat aneurysm model manifested rupture at the anterior and posterior Willis circle and that many harbored unruptured aneurysms at the anterior cerebral artery-olfactory artery bifurcation. This suggests that rupture was site-specific. To test our hypothesis that a site-specific response to hemodynamic changes is associated with aneurysmal rupture, we modified our original aneurysm model by altering the hemodynamics. During 90-day observation, the incidence of ruptured aneurysms at the anterior and posterior Willis circle was significantly increased and the high incidence of unruptured aneurysms at the anterior cerebral artery-olfactory artery persisted. This phenomenon was associated with an increase in the blood flow volume. Notably, the level of matrix metalloproteinase-9 associated with interleukin-1β was augmented by the increase in the blood flow volume, suggesting that these molecules exacerbated the vulnerability of the aneurysmal wall. The current study first demonstrates that a site-specific increase in interleukin-1β and matrix metalloproteinase-9 elicited by hemodynamic changes is associated with rupture. Our novel rat model of rupture may help to develop pharmaceutical approaches to prevent rupture.

Introduction

Cerebral aneurysms are a major cause of subarachnoid hemorrhage (SAH). Despite the catastrophic consequences of aneurysmal rupture, not all mechanisms underlying the formation, progression, and rupture of cerebral aneurysms are fully understood.

In their rat aneurysm model, Hashimoto et al.¹ induced hemodynamic changes by right common carotid artery (CCA) ligation and renal hypertension by ligation of the bilateral posterior renal arteries. They fed their rats a saline and β-aminopropionitrile diet to elicit disruption of the vascular wall. As epidemiological studies have shown that the incidence of aneurysmal SAH is higher in post- than premenopausal women and in men,² we modified their model and created a cerebral aneurysm model in oophorectomized rats.³ In these animals, the incidence of aneurysms at the anterior cerebral artery-olfactory artery bifurcation (ACA-OA) is high.⁴ Elsewhere we reported aspects of the pathogenesis of aneurysm formation. We documented that estrogen deficiency and hypertension augmented the elevation of oxidative stress, exacerbated endothelial damage induced by hemodynamic changes alone, and resulted in an increase in vascular degradation molecules.^5,6 Some drugs effectively reduced the incidence of cerebral aneurysm formation.^7–9 Despite their high number, no aneurysms at the ACA-OA, and only a few located at the anterior and posterior Willis circle (AW and PW) ruptured. Our recent study¹⁰ showed that in our model rats exposed to hyperhomocysteinemia (HHcy), the increase in rupture of aneurysms at the AW and PW was accompanied by an imbalance in matrix metalloproteinase (MMP)-9 and the tissue inhibitor of metalloproteinase (TIMP)2 and by an increase in interleukin (IL)-6; both were abrogated by treating these rats with folic acid. These findings indicate that aneurysmal rupture is associated with vascular degradation due to an imbalance in MMP-9 and TIMP2. However, we have no direct evidence for the site-specificity of aneurysmal rupture and the mechanisms leading to rupture remain to be elucidated.

Although the most important risk factor for rupture in humans is thought to be the aneurysm size and site,^11–13 little is known about the site-specificity of rupture. Hemodynamic flow patterns in aneurysms and surrounding vessels play a role in aneurysmal rupture.¹⁴ Carotid artery occlusion leads to significant hemodynamic changes in the cerebral circulation and the collaterals in the circle of Willis and to an increased demand on the collateral circulation. Furthermore, carotid artery occlusion due to atherosclerosis, iatrogenic ligation, or agenesis of the internal carotid artery (ICA) resulted in cerebral aneurysm formation.^15–17 While hemodynamic changes may be implicated in the formation and rupture of aneurysms, there is no direct evidence for their role in such events.

Based on these findings, we modified our rat aneurysm model and tested the hypothesis that a site-specific response to hemodynamic changes elicits an imbalance in MMP-9 and TIMP2 and an increase in pro-inflammatory cytokines, thereby leading to vascular degradation and aneurysmal rupture.

Here we show that rats in our modified model manifest a high incidence of ruptured aneurysms under experimentally induced pathophysiological conditions and that their rupture in estrogen-deficient hypertensive rats is associated with a site-specific increase in the level of IL-1β and vascular degradation molecules induced by hemodynamic changes in the Willis circle.

Materials and methods

All experiments and protocols were approved by the ethics committee of the Institute of Biomedical Sciences, Tokushima University Graduate School and conducted in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. Experiments were reported according to the ARRIVE guidelines. Before any procedures, the rats were anesthetized by 2–4% isoflurane inhalation.

Female Sprague-Dawley rats purchased from Charles River Laboratories Japan Inc. (Yokohama, Japan) were housed in a temperature- and humidity-controlled room (about 23℃ and 50%, respectively) under a 12-h light cycle and allowed free access to food and water.

Aneurysm induction

Ten-week-old female rats (230–260 g, n = 64) were divided into two equal groups. All rats were anesthetized by inhaling 2–4% isoflurane and by a subcutaneous injection of 0.25% bupivacaine. Group 1 underwent ligation of the left CCA; in group 2, we additionally ligated the right pterygopalatine- and the right external carotid artery (PPA, ECA) (Supplemental Fig. 1A). Immediately after left CCA ligation, we performed oophorectomy. Subsequently, the rats received a high salt diet (8% sodium chloride). Two weeks later, an investigator blinded to their group membership performed bilateral posterior renal artery ligation (RAL) to induce hypertension. The cause of death within 30 days was unclear in five group 1 and six group 2 rats; they were excluded from this study. Survivors exhibiting abnormal behavior or a drastic weight loss were euthanized by 4% isoflurane inhalation. Those, and rats that died, were used by two blinded observers to assess aneurysmal rupture 30–90 days after the last procedure. Ruptured aneurysms were inspected under a stereomicroscope after removing the blood coagulum. We identified saccular aneurysms at the AW and PW that were 1.5 times larger than the parent artery. They were categorized as AW aneurysms when they had arisen at the anterior communicating artery (AcomA), the ACA, the ICA, or the right middle cerebral artery (MCA) and as PW aneurysms when they were located at the proximal portion of the left posterior cerebral artery (P1). As shown in Figure 1(a), we prepared vascular corrosion casts 90 days after RAL and examined the unruptured aneurysms.¹⁸ The right ACA-OA and the Willis circle on the casts were inspected under a scanning electron microscope (SEM, VE8800, Keyence, Osaka, Japan). Based on morphological findings, the vascular wall surface was recorded as harboring a cerebral aneurysm when there was moderate outward evagination or an obvious saccular formation at the ACA-OA bifurcation.

Figure 1.

Relationship between hemodynamic changes and the incidence of ruptured aneurysms at the anterior and posterior Willis circle. Representative vascular corrosion cast of a group 2 rat (a). Subarachnoid hemorrhage (SAH) after aneurysmal rupture was assessed stereomicroscopically in rats that died, exhibited abnormal behavior, or manifested a drastic loss in body weight in the course of 90 days. We inspected ruptured aneurysms at the anterior and posterior Willis circle (AW, PW) under a stereomicroscope after removing the blood coagulum (b). SAH-free survival was calculated based on the incidence of SAH (c) (**p < 0.01, log-rank test). On day 90, we prepared corrosion casts and calculated the total incidence of unruptured- and ruptured aneurysms at the AW and PW (d). **p < 0.01, ^†p < 0.01, Fisher’s exact test. ACA: anterior cerebral artery; BA: basilar artery; ICA: internal carotid artery; MCA: middle cerebral artery; OA: olfactory artery; P1: proximal portion of posterior cerebral artery.

Transcranial duplex ultrasonography

The bilateral ICA (n = 7 per group) and the basilar artery (BA, n = 4 per group) were examined 4 weeks after RAL by simple random sampling without replacement.

Cell culture

We used human brain vascular smooth muscle cells (HBVSMCs; ScienCell Research Laboratories, Carlsbad, CA, USA) from passages 5 to 8. They were grown in smooth muscle cell medium (ScienCell) in 10-mm polystyrene plates. After 24-h treatment with recombinant human IL-1β (Wako, Osaka, Japan) (1 ng/ml), the cells were harvested for quantitative real-time polymerase chain reaction (qRT-PCR) assay. Cell proliferation was determined with the Cell Counting Kit-8 (Dojindo Laboratories, Kumamoto, Japan) in 24-well plates.

RNA isolation and qRT-PCR assay

Samples from another set of sham-operated-, group 1 and group 2 rats (n = 8 each) were subjected to qRT-PCR assay. Five weeks after RAL, these rats were euthanized and total RNA from the right ACA-OA, the right P1 and the left P1 was isolated and extracted. At the time of sampling, no aneurysms had ruptured. We examined the level of MMP-9, TIMP2, IL-1β, IL-6, tumor necrosis factor (TNF)-α, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH). The results were normalized to GAPDH.

Statistical analysis

Power estimates were calculated based on α = 0.05, 1 -β = 0.8, and a surgery-related drop-out rate of 10–20% to obtain group sizes appropriate for detecting an effect size of 0.4 in vivo based on a preliminary experiment using G*Power 3.1. The highest failures were death after bilateral RAL. The survival rate was analyzed with the log-rank test, the incidence of aneurysms with the Fisher exact test, and sequentially obtained data with analysis of variance (ANOVA) followed by the Tukey–Kramer test for multiple comparisons. The mRNA expression levels were determined with Kruskal–Wallis test followed by the Wilcoxon signed-rank test with Bonferroni correction for comparing samples from the same individuals and the Mann–Whitney U-test with Bonferroni correction for three-group comparisons. We calculated the Spearman rank correlation (r) to characterize the correlation strength between IL-1β and MMP-9. Statistical analyses were performed using IBM SPSS Statistics 22. Data are shown as the mean ± SD. Differences were considered statistically significant at p < 0.05.

Results

Additive hemodynamic changes increased the incidence of ruptured AW and PW aneurysms without affecting the blood pressure and the incidence of ACA-OA aneurysms

In our model, rats with aneurysms induced under pathophysiological conditions elicited by surgical procedures and diet alone, we detected SAH from the rupture of AW and PW aneurysms (Figure 1(b)); no aneurysms at the ACA-OA ruptured. We euthanized rats manifesting a 10% loss in body weight, hemiparesis, or chronic convulsions during the observation period and looked for ruptured aneurysms. Vascular casts¹⁸ were prepared on day 90 to examine unruptured AW-, PW-, and ACA-OA aneurysms.

In the course of 90 days, the incidence of SAH was significantly higher in group 2 than group 1 rats with ruptured aneurysms (p < 0.01, Figure 1(c)). Significantly more group 2- than group 1 rats suffered aneurysmal rupture (50% vs. 11%, p < 0.01, Figure 1(d)). The total incidence of ruptured and unruptured aneurysms was also significantly higher in group 2 (58% vs. 22%, p < 0.01), suggesting that hemodynamic changes were involved. The rupture rate (the number of ruptured aneurysms/total number of aneurysms at the AW and PW) was high in group 2 (81%) and group 1 rats (50%), indicating that AW and PW aneurysms were unstable and prone to rupture in both groups. As shown in Figure 2(a), on corrosion casts, unruptured aneurysms were remarkably larger at the AW and PW than at the ACA-OA. The incidence of ACA-OA aneurysms was similar in both groups and high (around 80%). However, ACA-OA aneurysms did not rupture, indicating that they were stable (Figure 2(b)). These findings point to site-specificity of aneurysmal growth and rupture.

Figure 2.

Relationship between hemodynamic changes and the incidence of unruptured aneurysms at the ACA-OA bifurcation. On day 90, we prepared corrosion casts. The right ACA-OA and the Willis circle on the casts were inspected at 2.5 kV under a scanning electron microscope. Representative aneurysms at the AW and PW were larger than aneurysms at the ACA-OA (a). There was no difference in the incidence of unruptured ACA-OA aneurysms between the two groups (b).

Throughout the observation period, until rupture onset, the blood pressure (Supplemental Fig. IB) and body weight (data not shown) of the experimental rats did not change significantly.

Histopathologically, the ruptured rat PW aneurysms resembled human saccular cerebral aneurysms

The wall of ruptured saccular cerebral aneurysms in humans is characterized by a disruption of the internal elastic lamina in the intima and a decellularized, degenerated matrix in the media.¹⁹ We histopathologically assessed the ruptured PW aneurysm from a euthanized group 2 rat that exhibited drastic weight loss 40 days after RAL. We observed a poorly organized luminal thrombus in the lumen, disruption of the internal elastic lamina in the intima, loss of mural cells, a degenerated matrix, and hyalinization in the thickened media of the ruptured PW aneurysm (Figure 3). The features of the rat ruptured PW aneurysms resembled those seen in the wall of human ruptured saccular cerebral aneurysms.

Figure 3.

Representative elastica van Gieson (EvG) stain of ruptured PW aneurysms in group 2 rats. EvG stain of cross-sections of a ruptured PW aneurysm demonstrating a poorly organized luminal thrombus (T), disruption of the internal elastic lamina (DEL) in the intima, a decellularized (De), degenerated matrix (DM), and hyalinization (Hy) in the thickened media. The normal cerebral artery (CA) exhibits a continuous internal elastic lamina (EL) [(original magnification, × 40 (a), × 100 (b)].

The higher incidence of ruptured PW aneurysms in group 2 rats was associated with an increase in the blood flow volume (BFVo) in the BA

To confirm the presence of hemodynamic changes, we obtained ultrasonograms in group 1 and 2 rats that had undergone carotid artery ligation 6 weeks earlier. The antegrade peak systolic velocity (PSV) and the end-diastolic velocity (EDV) in the right ICA and the BA were significantly higher and the retrograde PSV in the left ICA was lower in both experimental groups than in sham-operated rats (Figure 4(a) and (b)). Therefore, ultrasonography confirmed hemodynamic changes (Figure 4(d)). The rise in the mean flow velocity (MFV) in the right ICA and the BA of both groups was significantly higher than in the shams (Figure 4(c)). It was somewhat higher in group 2 than group 1; the difference was not statistically significant.

Figure 4.

Cerebral blood flow changes mirrored on ultrasonograms. Peak systolic velocity (PSV) (a) and end-diastolic velocity (EDV) (b) in the right ICA, the left ICA, and the BA were assessed ultrasonographically. The mean flow velocity (MFV) was calculated based on the PSV and the EDV (c). Representative ultrasonographic images mirrored the antegrade blood flow in the right ICA and the BA and the retrograde flow in the left ICA of experimental rats (d). *p < 0.05, **p < 0.01 by Tukey–Kramer test.

As shown in Figure 5(a), the diameter of the right ICA and the left P1 was increased in both experimental groups. Notably, the diameter of the BA was larger in group 2 than in group 1. The BFVo²⁰ was calculated from the MFV and the vascular diameter using the equations:

$$\mathrm{MFV}(v)=(\mathrm{PSV}-\mathrm{EDV})/3+\mathrm{EDV}\mathrm{BFVo}=1/4\pi r^{2}v$$

where r = the vessel diameter.

Figure 5.

Relationship between the site and the incidence of ruptured aneurysms and the blood flow volume (BFVo). We determined the vascular diameter (sham, n = 13; group 1, n = 24; group 2, n = 13) on corrosion casts (a) after a 90-day observation period and calculated the BFVo in each artery. The BFVo of the BA was significantly greater in group 2 than group 1 rats (b). *p < 0.05, **p < 0.01, Tukey–Kramer test. On day 90, the incidence of rupture of PW aneurysms was significantly higher in group 2 (n = 26) than group 1 rats (n = 27) (c). *p < 0.05, Fisher’s exact test.

The BFVo was significantly larger in the right ICA and BA of both experimental groups than in sham rats. As the diameter of and the MFV in the right ICA and BA were somewhat greater in group 2 than in group 1, the BFVo in the BA was significantly greater in group 2 (p < 0.05, Figure 5(b)) and the BFVo in the right ICA was a little greater in group 2 than group 1 although the difference was not statistically significant. The incidence of ruptured AW and PW aneurysms was 3–4 times higher in group 2 than group 1 (Figure 5(c)). These observations show that, in both experimental groups, the BFVo in the right ICA and BA was increased and that additive ligation of the right PPA and ECA increased the BFVo in the BA. Although we could not directly assess the BFVo in the AcomA, the left P1, and other arteries in the Willis circle, we think that the increased incidence of ruptured AW and PW aneurysms may be associated with the increased BFVo in the right ICA and BA, respectively. Despite the higher incidence of rupture of AW and PW aneurysms, the incidence of unruptured ACA-OA aneurysms was not affected by hemodynamic changes. These observations confirm the site-specific vulnerability of vessels in the Willis circle exposed to hemodynamic changes.

The mRNA level of MMP-9 was higher in the vascular wall of the left P1 in group 2 than group 1 rats

To identify molecules related to the difference in the vulnerability to rupture of aneurysms in group 2 rats, we compared the mRNA level of inflammation-related and vascular degradation molecules at the left P1 of sham, group 1, and group 2 rats. As shown in Figure 6(a), in both experimental groups, the mRNA level of IL-1β and MMP-9 was significantly higher than in the shams (Figure 6(a)). Furthermore, the mRNA level of MMP-9 was higher in group 2 than group 1 rats. These findings suggest that the expression of IL-1β and MMP-9 was induced in group 1 and 2 rats and that additive carotid ligation resulted in a greater increase in MMP-9 in group 2 rats.

Figure 6.

Gene expression of interleukin (IL)-1β and matrix degradation molecules. The mRNA level of IL-1β, matrix metalloproteinase (MMP)-9, and the tissue inhibitor of metalloproteinase (TIMP)2 in the vascular wall at the left P1 of sham-, group 1-, and group 2 rats (n = 8 each group) was assessed by quantitative real-time polymerase chain reaction (qRT-PCR) assay (a). The expression of these molecules in the vascular wall at the right P1, ACA-OA, and left P1 in group 2 rats was also assessed by qRT-PCR (b). For normalization, we used the mRNA level of glyceraldehyde 3-phosphate dehydrogenase (GAPDH). *p < 0.05, **p < 0.01 by the Kruskal–Wallis test followed by the Mann–Whitney test or the Wilcoxon signed-rank test with Bonferroni correction. ns: not significant.

MMP-9 expression was higher in the vascular wall of the left P1 than in other arteries of group 2 rats

To further address the site-specific vulnerability in the Willis circle, we compared the mRNA level of IL-1β, MMP-9, and TIMP2 at the right and left P1, and the ACA-OA of group 2 rats. As shown in Figure 6(b), the mRNA level of IL-1β was significantly higher at the ACA-OA and the left P1 than at the right P1 (p < 0.01). The mRNA level of MMP-9 was highest at the left P1 and positively correlated with the mRNA level of IL-1β (r = 0.46, p < 0.05). Furthermore, there was an evident imbalance between MMP-9 and TIMP2 at the left P1. These findings suggest that the site-specific overwhelming expression of degradation molecules induced by hemodynamic changes in the Willis circle increased the propensity of PW aneurysms to rupture.

IL-1β upregulated the expression of MMP-9 and increased cell proliferation in HBVSMCs in vitro

MMP-9 activity is increased by IL-1β in cardiac fibroblasts in vitro.²¹ To confirm the effect of IL-1β on MMP-9, we used HBVSMCs. We found that the mRNA level of MMP-9, but not of TIMP2 was increased and that the proliferation of HBVSMCs exposed to IL-1β was increased (Figure 7).

Figure 7.

The effects of IL-1β on human brain vascular smooth muscle cells (HBVSMCs) in vitro and schematic illustration of our novel cerebral aneurysm model. HBVSMCs were exposed for 24 h to IL-1β (1 ng/ml) and the mRNA level of MMP-9 and TIMP2 was determined by qRT-PCR assay (a). For normalization, we used the mRNA level of GAPDH. After 24-h treatment with IL-1β (1 ng/ml), cell proliferation was determined with the Cell Counting Kit-8 (b). **p < 0.01 by the Mann–Whitney U-test. N.D.: not detectable. Increased levels of pro-inflammatory cytokines and matrix degradation molecules induced by estrogen deficiency, hypertension, and hemodynamic stress resulted in the rupture of aneurysms at the AW and PW (c).

Discussion

We first document the high, reproducible incidence of ruptured aneurysms at the AW, PW, and of unruptured aneurysms at the ACA-OA in rats exposed to hemodynamic changes under pathophysiological conditions elicited by surgical procedures and diet alone. We addressed the site-specific characteristics of these aneurysms and found that AW and PW aneurysms did, while ACA-OA aneurysms did not rupture. We show that rupture was at least partly associated with a site-specific elevation of IL-1β and an imbalance in MMP-9 and TIMP2 induced by hemodynamic changes.

Our earlier study³ found that estrogen-deficient hypertensive female rats subjected to ligation of the left CCA had a high incidence of aneurysms at the ACA-OA bifurcation. While none of those aneurysms ruptured, some aneurysms at the AW and PW ruptured. To induce a higher rupture rate, in the current study, we ligated, in addition to the left CCA, the right ECA, and the right PPA (group 2). As expected, the additive hemodynamic changes elicited in those rats resulted in a higher incidence in the formation and rupture of aneurysms at the AW and PW. Histopathologically, the ruptured PW aneurysms resembled saccular cerebral aneurysms in humans. Cai et al.²² reported that while hemodynamic changes after the ligation of the unilateral CCA, contralateral PPA, and the ECA induced aneurysms in normotensive male rats, their incidence was low. In an earlier study,^3,23 we induced cerebral aneurysms in male and female rats. We found that while few males developed aneurysms, in oophorectomized rats, we encountered a high incidence of cerebral aneurysms. Those findings suggested the protective role of estrogen against vascular damage. Besides the vascular degeneration elicited by exposing the current rats to estrogen deficiency, renal hypertension, and a high salt diet, the additive hemodynamic changes we induced may have contributed to the high incidence of aneurysmal rupture.

To assess the hemodynamic changes elicited by additive carotid artery ligation, we first recorded blood flow velocity (BFVe) changes observed on ultrasonograms, which are thought to be useful for evaluating the intracranial cerebral blood flow in rats.²⁴ We detected an increase in the BFVe in the cerebral arteries of our experimental rats. Then we used SEM to study the aneurysms and to determine the vascular diameter on corrosion casts. We observed the outward remodeling of the ICA and the left P1 in these rats. We suspect that the higher incidence of ruptured PW aneurysms in group 2 than group 1 rats was associated with the increased BFVo in the BA. Earlier studies in rabbits^25,26 suggested that an increase in the blood flow results in adaptive vascular remodeling and the formation of aneurysms. Quantitative computational hemodynamic analysis showed that aneurysmal growth is associated with an increased risk for rupture.^27,28 These findings support our hypothesis that hemodynamic changes are associated with aneurysmal growth and rupture. To further address the effects of altered hemodynamics, we compared the mRNA level of IL-1β, MMP-9, and its inhibitor, TIMP2, in the left P1 of our experimental and sham rats. We found that the increase in MMP-9 was associated with a higher incidence of PW aneurysms in group 2.

Despite the high incidence (around 80%) of ACA-OA aneurysms in both experimental groups, there was no increase in their rupture. This suggests a site-specific response to hemodynamic changes and a difference in the pathogenesis underlying the formation and rupture of cerebral aneurysms. To better understand the vulnerability of the left P1 in group 2 rats that manifested a high incidence of aneurysmal rupture, we assessed inflammatory-related and vascular degradation molecules and compared their levels at the vulnerable left P1 with seemingly less vulnerable sites, the ACA-OA bifurcation and the right P1. TNF-α is thought to play a highly important role in vascular inflammation which, in turn, has a pivotal role in aneurysmal progression.^9,29–32 Hemodynamic stress and induced hypertension increased the expression of TNF-α in rats²⁹ while the TNF-α inhibitors etanercept and 3,6′-dithiothalidomide (DTH) decreased the rate of aneurysm formation and rupture.^31,32 As we observed no difference in the mRNA level of TNF-α and IL-6 at the left P1 and other arteries in group 2 rats (data not shown), we think that it does not contribute to aneurysmal rupture.

There was a significant increase in MMP-9, a decrease in TIMP2, and an imbalance in these molecules at the left P1 of group 2 rats. In agreement with the current study, elsewhere¹⁰ we reported that in our oophorectomized aneurysm model rats treated with HHcy, AW and PW aneurysms were prone to rupture and we documented the disequilibrium in MMP-9 and TIMP2 and the increased infiltration of macrophages at the AW. Associated with the increase in MMP-9, the expression of IL-1β was increased at the left P1; this phenomenon was observed in neither our sham rats nor in the right P1 of experimental rats where no aneurysms were detected. IL-1β, one of the key inflammatory mediators in cerebrovascular inflammation,³³ cerebral aneurysms,³⁴ and aortic aneurysms, promotes extracellular matrix degradation by increasing the production of MMPs.³⁵ Zhang et al.³⁶ demonstrated that in humans, activation of the nod-like receptor (NLR) family, the pyrin-domain containing 3 (NLRP3) and the NLRP3 inflammasome-mediated production of mature IL-1β, is associated with the rupture of cerebral aneurysms. We found that in vitro treatment of HBVSMCs with IL-1β increased their expression of MMP-9 and their proliferation. This was consistent with histological findings of a thickened media and a disrupted matrix. These findings partly support our hypothesis that an increase in MMP-9 and IL-1β elicited a disruption of the aneurysmal wall and the rupture of aneurysms. Further studies are needed to confirm the causal relationship between these molecules and aneurysmal rupture.

Inflammatory changes due to estrogen deficiency, hypertension, and altered hemodynamics may render the vascular wall at the AW and PW vulnerable. Based on current and earlier findings, we posit that in our model rats, increased levels of IL-1β in the vascular wall affected the degradation molecule MMP-9 and rendered the vulnerable aneurysmal wall prone to rupture (Figure 7(c)). At present, we cannot explain why there is a site-specific increase in the level of IL-β and MMP-9 at the left P1. Studies are underway to determine whether the rate of aneurysmal rupture can be decreased by inhibiting the level of these molecules in group 2 rats.

In humans, carotid artery occlusion leads to significant hemodynamic changes in the cerebral circulation with collaterals in the circle of Willis and to the formation of intracranial aneurysms.^15–17 The most important risk factor for rupture is thought to be the aneurysm size and site.^11–13 The rupture of aneurysms we induced in our rats by eliciting hemodynamic changes was site-specific, therefore, their aneurysms may partly reflect the characteristics of human de novo aneurysms. Consequently, our modified rat aneurysm model may be useful for studying the formation, growth, and rupture of cerebral aneurysms in humans.

In summary, we provide new evidence that aneurysmal rupture in rats was partly attributable to a site-specific increase in MMP-9 associated with an increase in IL-1β induced by altered hemodynamics. Importantly, under the pathophysiological conditions we imposed, we observed aneurysmal rupture at sites in the circle of Willis of rats that are similar to those in humans. Furthermore, histopathologically, the ruptured rat aneurysms resembled human saccular cerebral aneurysms. Our group 2 rats may represent a good model for assessing the efficacy of pharmacological treatments to prevent aneurysmal rupture. Currently, we have no direct evidence for an association between aneurysms and pro-inflammatory and degradation molecules and studies are underway to identify other pathophysiological factors that contribute to the rupture of aneurysms.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Grant-in-Aid for Scientific Research (JSPS KAKENHI Grant Number JP15H04950), a Grant-in-Aid for Young Scientists (B) (JSPS KAKENHI Grant Number JP15K19972) a Grant-in-Aid for the Strategic Young Researcher Overseas Visits Program for Accelerating Brain Circulation from the Japan Society for the Promotion of Science (JSPS Grant Number JPS2407), and R01NS055876 (TH) and R01NS082280 (TH) from the National Institutes of Health (United States).

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

TM, DKK, and KTK took part in the conception, design, acquisition of data, and the drafting of the article. MK and Y. Kurashiki participated in the acquisition and the analysis or interpretation of data. KY, KS, YT, TK, Y. Kanematsu, JS, TH, and SN critically reviewed the paper for intellectual content and approved its submission.

Supplementary material

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