Molecular mechanisms of the formation and progression of intracranial aneurysms.

Until recently, only a little was understood about molecular mechanisms of the development of an intracranial aneurysm (IA). Recent advancements over the last decade in the field of genetics and molecular biology have provided us a wide variety of evidences supporting the notion that chronic inflammation is closely associated with the pathogenesis of IA development. In the field of genetics, large-scale Genome-wide association studies (GWAS) has identified some IA susceptible loci and genes related to cell cycle and endothelial function. Researches in molecular biology using human samples and animal models have revealed the common pathway of the initiation, progression, and rupture of IAs. IA formation begins with endothelial dysfunction followed by pathological remodeling with degenerative changes of vascular walls. Medical treatments inhibiting inflammatory cascades in IA development are likely to prevent IA progression and rupture. Statins and aspirin are expected to suppress IA progression by their anti-inflammatory effects. Decoy oligodeoxynucleotides (ODNs) inhibiting inflammatory transcription factors such as nuclear factor kappa-B (NF-κB) and Ets-1 are the other promising choice of the prevention of IA development. Further clarification of molecular mechanisms of the formation and progression of IAs will shed light to the pathogenesis of IA development and provide insight into novel diagnostic and therapeutic strategies for IAs.

Introduction

Intracranial aneurysm (IA) is a common lesion with a prevalence ranging from 1% to 5% in large autopsy studies[1)] and a major cause of subarachnoid hemorrhage (SAH) with high rates of morbidity and mortality. Despite the catastrophic sequelae of aneurysmal SAH, most IAs are asymptomatic and clinically silent during a course of a patient’s lifetime. Recent prospective large trials revealed a low rupture risk for patients with small unruptured IAs.[2,3)] In contrast, a considerable number of patients present with SAH because of rupture of a small IA in clinical practice. Although prophylactic surgical or endovascular treatments should be indicated only for selected patients with unruptured IAs, determinants for treatment selection are confined to aneurysm size, aneurysm location, and history of SAH, and other predisposing conditions leading to IA enlargement and rupture have not been fully elucidated.

IA is a pathological outward bulging of intracranial arteries that may have a locally thinned wall prone to rupture. Under physiological circumstances, there is a homeostatic balance between hemodynamic stress and arterial wall integrity. IA is characterized by loss of integrity in arterial walls including endothelium dysfunction, intimal hyperplasia, decreased cellular components, and disorganized extracellular matrix (ECM). Theoretically, IA is believed to develop as a disruption of a homeostatic balance in arterial walls by excessive hemodynamic stress. ECM provides a structural framework essential for the functional properties of vessel walls. Aneurysmal enlargement arises from an imbalance between ECM protein synthesis and degradation. The degenerated aneurysmal wall becomes too fragile to resist the hemodynamic stress, and finally ruptures. The process leading to the initiation, progression, and rupture of IA is involved in dynamic structural changes of the arterial wall.

Recent advancement in genetics and molecular biology for IA shed light to the molecular mechanisms of IA formation and rupture. More comprehensive understanding of the molecular events in the IA wall helps us to assess the risk of rupture of individual IA and provide an optimizing treatment to individual patients. In this review, the author presents the most updated findings from genetic and molecular biological studies of IAs, and discusses the implications of these findings for future diagnostic and therapeutic management.

Genetic Factors for the Development of an IA

Ample evidence suggests the contribution of genetic factors in the pathogenesis of IA. Some heritable connective tissue disorders including autosomal dominant polycystic kidney disease, Ehlers-Danlos syndrome type IV and neurofibromatosis type I are associated with IA.[4,5)] A familial clustering of SAH and IA also supports the importance of genetic factors in IA formation and rupture.[6,7)]

Several susceptibility loci and genes associated with IA have been reported in studies of genetic linkage and single nucleotide polymorphism (SNP) for familial IAs.[8–17)] Unfortunately, only few of them have been replicated among studies. For example, linkage to 7q11 was identified in a study analyzing 104 Japanese affected sib-pairs [nonparametric logarithm of odds score (NPL) 3.2][12)] and confirmed by a study on 13 families with IA in Utah [logarithm of odds score (LOD) 2.3].[9)] However, another two Japanese linkage studies did not replicate linkage to the 7q11 locus.[10,18)] The largest genome-wide linkage analysis of IA families throughout North America, New Zealand, and Australia using results in the full sample of 333 IA pedigrees demonstrated linkage to the chromosome 4q32.2 (LOD 2.6) and 12p12.3 (LOD 3.1).[8)] However, the study did not detect the evidence of linkage to the chromosomal loci identified in previous studies.

Genome-wide linkage analyses have also identified a positive association with positional candidate genes, the function of which is assumed to be involved in IA development, including elastin (ELN),[12,19,20)] collagen type 1 A2 (COL1A2),[21)] endoglin (ENG),[22)] perlecan (HSPG2),[23)] versican (CSPG2),[24)] and TNFRSF13B.[25)] However, there have been conflicting results on the association of ELN,[26–28)] COL1A2,[23)] and ENG[29–31)] with IA. The perlecan gene coding for a heparan sulfate proteoglycan 2 and the versican gene coding for chondroitin sulfate proteoglycan 2 play an important role in ECM assembly.[23,24)] A haplotype of TNFRSF13B SNPs coding for trans-membrane activator and calcium modulator ligand interactor was also found to be associated with IA.[25)] Although these genes are attractive candidates for IA susceptible genes, confirmatory studies are needed for the association of polymorphism haplotypes. By the method of linkage analysis, no single gene has been consistently identified as a candidate gene. These results suggest that multiple chromosomal loci or genes may be associated with IA development. The discrepancy among studies may be derived from the difference of genetic and environmental factors among populations, and suggests the complexity and multiplicity of genetic determinants of IA.

Genome-wide association studies (GWAS) provide an alternative approach to elucidate the influence of genetic variants in complex, multifactorial disorders. In hypertension[32)] and ischemic stroke,[33)] GWAS has already revealed previously unknown genes which provide a novel insight into future diagnostic and therapeutic advancements. In 2008, Bilguvar et al. first published a large-scale GWAS of IA using Finnish, Dutch, and Japanese cohorts including 2,196 IA cases and 8,085 controls, and identified SNPs in loci 2q33.1 [odds ratio (OR), 1.24; P = 4.4 × 10–8], 8q11.23 (OR, 1.36; P = 1.4 × 10–10), and 9p21.3 (OR, 1.29; P = 1.4 × 10–10) that were associated with sporadic and familial IAs independent of exogenous risk factors.[34)] Subsequently, 5 GWASs of IA were published to date (Table 1).[35–39)] The 9p21.3 locus was consistent in 4 of 6 GWAS. Within this genomic region, there are multiple genes which may be related to the pathogenesis of IA. CDKN2A codes for the p16INK4a tumor suppressor gene, spliced variants of which inhibit cyclin-dependent kinase 4 (CDK4) or p53 degradation, and may play vital roles in cell cycle progression at the G1 stage. CDKN2B codes for p15INK4b, which prevents cellular proliferation through induction by transforming growth factor beta (TGF-β). CDKN2BAS is a noncoding ribonucleic acid (RNA) region involved in the transcriptional repression of these genes. Targeted deletion of the 9p21 locus in rat vascular smooth muscle cells (SMCs) showed uncontrolled proliferation compared to controls.[40)] Thus, the 9p21 gene cluster is considered to play an essential role in cellular proliferation via various molecular pathways and regulators.

Table 1

Genome wide association studies of intracranial aneurysms

Study (year)	Cohort	Gene locus	SNP	Candidate genes	Potential mechanism associated with IA	OR (95%CI)	P value
Bilguvar et al. (2008)[34)]	2,196 cases and 8,085 controls (European and Japanese)	2q33.1	rs700651	BOLL, PLCL1	Angiogenesis	1.24 (1.15–1.34)	4.4 × 10–8
		8q11.23	rs10958409	SOX17	Endothelial function	1.36 (1.24–1.49)	1.4 × 10–10
		9p21.3	rs1333040	CDKN2A-CDKN2B	Cellular proliferation	1.29 (1.19–1.40)	1.4 × 10–10
Yasuno et al. (2010)[39)]	5,891 cases and 14,181 controls (European and Japanese)	8q11.23	rs10958409	SOX17	Endothelial function	1.17 (1.10–1.25)	9.0 × 10–7
		8q12.1	rs9298506	SOX17	Endothelial function	1.28 (1.20–1.38)	1.3 × 10–12
		9p21.3	rs1333040	CDKN2A-CDKN2B	Cellular proliferation	1.32 (1.25–1.39)	1.5 × 10–22
		10q24.32	rs12413409	CNNM2	Cellular proliferation	1.29 (1.19–1.40)	1.2 × 10–9
		13q13.1	rs9315204	STARD13	Cellular proliferation	1.20 (1.13–1.28)	2.5 × 10–9
		18q11.2	rs11661542	RBBP8	Cellular proliferation	1.22 (1.15–1.28)	1.1 × 10–12
Akiyama et al. (2010)[35)]	1,027 cases and 853 controls (Japanese)	1q21	rs7550260	ARHGEF11	Actin cytoskeleton remodeling	1.32 (1.15–1.50)	4.9 × 10–5
		3p25.2	rs9864101	IQSEC1	Actin cytoskeleton remodeling	1.49 (1.23–1.80)	3.6 × 10–5
		7p21.2	rs4628172	TMEM195	Unknown	1.3 (1.14–1.48)	1.3 × 10–5
		9q31.2-31.3	rs1930095	(intergenic region)		1.44 (1.22–1.71)	1.3 × 10–5
Yasuno et al. (2011)[38)]	3,111 cases and 1,666 controls (Japanese)	4q31.23	rs6841581	EDNRA	Endothelin signaling	1.22 (1.14–1.31)	2.2 × 10–8
		12q22	rs6538595	NDUFA12/NR2C1/FGD6/VEZT	Unknown	1.16 (1.10–1.23)	1.1 × 10–7
		20p12.1	rs1132274	RRBP1	Unknown	1.2 (1.11–1.28)	6.9 × 10–7
Low et al. (2012)[37)]	2,431 cases and 12,696 controls (Japanese)	4q31.22	rs6842241	EDNRA	Endothelin signaling	1.25 (1.16–1.34)	9.6 × 10–9
		9p21.3	rs10757272	CDKN2BAS	Cellular proliferation	1.21 (1.13–1.30)	1.6 × 10–7
Foroud et al. (2012)[36)]	1,483 cases and 1,683 controls (European)	8q11.23	rs1072737	SOX17	Endothelial function	1.25 (NR)	8.7 × 10–5
		9p21.3	rs6475606	CDKN2BAS	Cellular proliferation	1.35 (NR)	3.6 × 10–8

CI: confidence interval, IA: intracranial aneurysm, NR: not reported, OR: odds ratio, SNP: single nucleotide polymorphism.

SOX17 located in the vicinity of associated SNPs at 8q12.1 has also been strongly implicated in most GWAS of IAs. SOX17, a transcription factor of the SOX family, is involved in hematopoietic development from endothelial cell (EC)-derived embryonic and pluripotent stem cells.[41)] Overexpression of SOX17 in mouse tumor ECs promoted tumor angio-genesis and vascular abnormalities,[42)] suggesting that SOX17 may be closely related to the maintenance of vascular endothelial function.

A recent GWAS found an SNP on chromosome 4q31.23, coding for the endothelin receptor type A (EDNRA) gene, was significantly associated with IA in Dutch, Finnish, and Japanese populations (OR 1.22; 95% CI 1.14–1.31; P = 1.95 × 10–8).[38)] Another GWAS in a Japanese cohort also showed that the other SNP on 4q31.22 near the EDNRA gene, was significantly associated with IA [OR 1.25; 95% confidence interval (CI) 1.16–1.34; P = 9.583 × 10–9].[37)] EDNRA is a G-protein-coupled receptor for endothelins and expressed in vascular SMCs. Endothelins are potent vasoconstrictors in tissue remodeling after vascular wall injury.[43)] Alternatively, down-regulation of EDNRA signaling could lead to defective vascular repair after injury, allowing IA development.

In a meta-analysis on 61 studies including 32,887 IA cases and 83,683 controls, the strongest associations were found for the SNPs on the above mentioned 3 loci: 9p21 within the CDKN2BAS gene (rs10757278: OR 1.29; 95% CI 1.21–1.38; and rs1333040: OR 1.24; 95% CI 1.20–1.29), 8q11 near the SOX17 gene (rs9298506: OR 1.21; 95% CI 1.15–1.27; and rs10958409: OR 1.19; 95%CI 1.13–1.26), and 4q31.23 near the EDNRA gene (rs6841581: OR 1.22; 95% CI 1.14–1.31).[44)]

GWAS yielded valuable insight into genetic factors associated with IA development, implying a substantial role of cell cycle dysfunction and endothelial dysfunction in the pathogenesis of IA. However, the collective results only represent associations with IA, and cannot demonstrate that the abnormalities of the genes identified are causative of IA development. For comprehensive understanding of molecular mechanisms of IA development, detailed molecular cascades leading to the formation, progression, and rupture of IA should be elucidated.

Accumulation of Inflammatory Cells in IAs

The infiltration of inflammatory cells is a hallmark of IA.[45–47)] The main component of inflammatory cells infiltrating into the IA wall is macrophage. Although macrophages infiltrate into the wall of both unruptured and ruptured IAs, the degree of macrophage infiltration was more prominent in ruptured IAs,[46)] suggesting that vascular inflammation is closely associated with IA rupture. The degree of leukocyte infiltration also correlated with degenerative changes in the IA wall such as loss of medial SMCs and ECM degradation.[47)] Macrophages infiltrated into aneurysmal walls of experimentally induced IAs in mice[48)] and rats[49)] at the early stage of IA formation, and the number of macrophages infiltrated in the IA wall increased with IA development.[49)] In macrophage-depleted mice, IA formation was dramatically decreased compared to control mice.[48)] These data strongly suggest that macrophages accumulating into the IA wall play a role in IA development. Recently, Hasan et al. reported their preliminary results of macrophage imaging in the IA wall using ferumoxytol-enhanced magnetic resonance imaging (MRI).[50)] This imaging technique may be able to estimate the degree of inflammation in the IA wall and the risk of rupture in the future.

T-lymphocytes also actively participate in the inflammatory reaction in the IA wall.[45,46)] Frosen et al. histologically compared 42 ruptured with 24 unruptured IAs and found the association of the number of T-lymphocytes with aneurysm rupture.[46)] Ishibashi et al. reported an increased number of mast cells in the IA wall of an experimentally induced rat IA model.[51)] Inhibitors of mast cell degranulation effectively inhibited the degenerative changes in the IA wall. The degranulation of mast cells induced the expression and activation of matrix metalloproteinase (MMP)-2, -9, and inducible nitric oxide synthase (iNOS) in primary cultured SMCs from rat IAs. Mast cells are involved in allergic inflammation through the release of mediators and cytokines. Based on these results, mast cells seem to constitute an integral part of the inflammatory response in the IA development.

Humoral immune reaction elicited by inflammatory cells seems to be associated with IA development. Antibodies (IgM and IgG) were found primarily in the luminal side of most human IAs.[45,52)] Tulamo et al. reported complement activation in human IA walls more abundantly in ruptured cases, and they advocated that complement activation was associated with IA rupture and wall degeneration.[53)] The activation of the complement system also causes the release of pro-inflammatory cytokines, suggesting that complement activation may be one of the triggers of inflammatory cascades in the IA wall.

Endothelial Dysfunction

The endothelial monolayer is the center of command of vascular homeostasis. ECs have various important function including anti-atherogenic properties, vasodilation, and anti-thrombotic effects. Morphological changes of ECs in the IA wall were well documented in the literature.[47,54–56)] Gap formation at the junctions of the ECs was one of the most obvious changes on the endothelial surface of the aneurysms.[55)] The expression of tight junction proteins such as occluding and zona occludens-1 (ZO-1) was down-regulated in the early stage of the development of experimentally induced rat IAs.[56)] Other morphological changes observed in the IA wall are the disruption of the arrangement of ECs, the adhesion of leukocytes to ECs, and partial loss of ECs, which are considered as endothelial damage.

In addition to morphological changes, endothelial dysfunction has drawn attention in the pathogenesis of IA development (Fig. 1). In a narrow sense, endothelial dysfunction means the induction of proinflammatory genes. Representative proinflammatory genes expressed in ECs in the IA wall are chemokines and cell adhesion molecules that cause macrophage infiltration. Monocyte chemotactic protein-1 (MCP-1) plays critical roles in monocyte/macrophage recruitment to affected sites in various vascular diseases.[57)] The gene expression of MCP-1 was up-regulated in the rat IA wall at the early stage of IA formation.[58)] MCP-1 deficient mice exhibited a significant decrease in IA formation,[48,58)] and the treatment with a plasmid DNA encoding a dominant negative mutant of MCP-1 (7ND) inhibited IA development in rats.[58)] For the recruitment of monocytes/macrophages, vascular cell adhesion molecule-1 (VCAM-1) also plays a role by mediating firm adhesion of monocytes to ECs. VCAM-1 expression in human IA samples was reported,[45)] and studies using gene expression profiling for IA showed increased expression of VCAM-1.[59–61)] Although functional importance of VCAM-1 in the pathogenesis of IA development remains to be elucidated, VCAM-1 seems to be one of the key molecules linking endothelial dysfunction to macrophage accumulation.

Fig. 1.

Inflammatory cascades in the development of intracranial aneurysms. COX-2: cyclooxygenase-2, ECM: extracellular matrix, eNOS: endothelial nitric oxide synthase, EP2: prostaglandin receptor 2, IL-1β: interleukin-1 beta, iNOS: inducible nitric oxide synthase, LOX: lysyl oxidase, MCP-1: monocyte chemotactic protein-1, MMP: metalloproteinase, NF-κB: nuclear factor kappa B, nNOS: neuronal nitric oxide synthase, PGE2: prostaglandin E2, ROS: reactive oxygen species, TLR4: toll-like receptor 4, TNF-α: tumor necrosis factor-alpha, TNFR1: tumor necrosis factor receptor 1, VCAM-1: vascular cell adhesion molecule-1.

Nuclear factor-kappa B (NF-κB) is a family of transcriptional factors regulating the expression of a variety of genes including VCAM-1[62)] and MCP-1[63,64)] in response to inflammatory mediators.[65)] In experimentally induced rat IAs, activation of NF-κB occurred in the intima of the IA wall in the early stage of IA formation.[66)] NF-κB p50 subunit deficient mice exhibited decreased incidence of IA formation with reduced number of macrophage infiltration into the arterial wall. NF-κB decoy oligodeoxynucleotides (ODNs) inhibited IA formation when it was administered at the early stage of IA formation in rats. Macrophage infiltration and expression of downstream genes were dramatically inhibited by NF-κB decoy ODN. These data indicate that NF-κB is involved in IA formation as a key regulator of proinflammatory genes. IAs preferentially locate at the arterial bifurcation or sharp curves,[67)] suggesting the role of hemodynamics in IA biology. ECs act as a sensor of complex forces and flows in the blood stream. In response to shear stress, ECs changes cell shape as a result of reorganization of cytoskeletal components.[68)] Fluid shear stress promotes the trans-location into the nucleus of NF-κB in cultured ECs by activating IκB kinase.[69)] Given the focal nature of IA location, NF-κB activation elicited by excessive hemodynamic stress is likely to be the first step of IA formation.

Prostaglandin E2 (PGE2) - prostaglandin E receptor 2 (EP2) signaling also functions as a link between hemodynamic stress and IA formation through the activation of NF-κB. PGE2 is synthesized from arachidonic acid by sequential actions of two enzymes, cyclooxygenase (COX) and PGE synthase (PGES).[70)] Among several isoforms of two enzymes, the expression of COX-2 and microsomal PGES1 (mPGES1), that are induced in various inflammatory diseases, was up-regulated in the rat[71)] and human[72)] IA wall together with that of EP2.[71)] Both EP2 deficiency and COX-2 inhibition significantly suppressed NF-κB activation and IA formation.[71)] Conversely, NF-κB inhibition by decoy ODN suppressed COX-2 expression in the IA wall,[71)] suggesting the presence of a positive feedback loop between COX-2 and NF-κB via EP2. As shear stress induced COX-2 and EP2 in cultured ECs,[71)] it seems that this positive feedback loop stimulated by excessive shear stress contributes to the maintenance of chronic inflammation in the IA wall.

Tumor necrosis factor-alpha (TNF-α) is a cytokine that induces signaling pathway associated with inflammation and apoptosis. TNF-α triggers endothelial dysfunction with increased monocyte recruitment in atherogenesis.[73)] Jayaraman et al. revealed increased TNF-α expression in human IA samples by reverse transcription polymerase chain reaction (RT-PCR) and Western blotting.[74)] Increased TNF-α content was confirmed in experimentally induced rat IAs with increased activity of TNF-α converting enzyme (TACE), an enzyme responsible for TNF-α release.[75)] In TNF receptor superfamily member 1a (TNFR1)-deficient mice, the incidence of IA formation was significantly reduced compared to control mice.[75)] TNF-α expression correlated with increased expression of toll-like receptor (TLR).[61)] TRL-4, which stimulates NF-κB activation in arterial walls in atherosclerosis,[76)] was expressed in the endothelia cell layers in human and mouse IAs.[77)] The TLR4 expression coincided well with NF-κB activation,[77)] supporting the involvement of the TNF-α/LR4/NF-κB pathway in IA development. The above-mentioned findings strongly suggest the crucial role of TNF-α in the pathogenesis of IA formation.

For the maintenance of vascular homeostasis against the continuous hemodynamic force, nitric oxide (NO) is produced by endothelial nitric oxide synthase (eNOS) in the endothelium. Through the action of NO as the endothelium-derived relaxing factor (EDRF), eNOS has a protective effect on various arterial diseases including atherosclerosis and abdominal aortic aneurysm (AAA).[78)] Therefore, reduced eNOS expression is the hallmark of endothelial dysfunction. The expression of eNOS was decreased in rat IAs compared to the contralateral cerebral arterial wall without IA.[79)] Abruzzo et al. reported that eNOS deficient mice exhibited increased IA formation in female rats.[80)] On the other hand, Aoki et al. demonstrated the incidence of IA formation in male eNOS deficient mice was similar to that in wild-type mice.[79)] In the IA wall in eNOS deficient mice, neuronal nitric oxide synthase (nNOS) was up-regulated in a compensatory manner, and eNOS and nNOS double knockout mice exhibited an increased incidence of IA formation accompanied by pronounced macrophage infiltration.[79)] These contradictory results may be derived from gender difference. Estrogen is well known to have a vaso-protective and anti-inflammatory effect through the NF-κB signaling pathway.[81)] In the vascular wall of rat IAs induced by hypertension and oophorectomy, the expression level of both eNOS and estrogen receptor α (ERα) was decreased in immunohisto-chemistry and RT-PCR.[82)] The incidence of rat IA formation increased threefold by oophorectomy,[83)] and morphological changes in ECs in the IA wall was accelerated by estrogen deficiency.[31)] These findings may partly explain the predisposition to a higher incidence of IA development in postmenopausal women.

Pathological Changes in Vascular SMCs

In healthy vascular walls, vascular SMCs principally engage in contraction. In response to several environmental stimuli, SMCs undergo phenotypic modulation to a synthetic phenotype, in which SMCs proliferate and synthesize ECM. The phenotypic modulation of SMCs was found in the IA wall with decreased expression of contractile proteins.[84)] As phenotypic modulation was more pronounced in ruptured IAs,[84)] it appears to be related to the pathological remodeling of the IA wall and to the rupture mechanism.

The structural integrity of vessel walls depends on a balance between synthesis and degradation of ECM in the media. Thinning of the media proceeds with IA progression. The arterial wall of IAs exhibited both enhanced degradation of ECM and attenuated biosynthesis of ECM. A family of matrix metalloproteinases (MMPs) degrades most of the arterial ECM components including elastin and collagen, and causes pathological vascular remodeling in atherosclerosis[85,86)] and AAA.[87,88)] In human IA samples, the expression of MMPs was demonstrated by immunohistochemistry and Western blotting.[89,90)] In experimentally induced rat IAs, MMP-2 and -9 was expressed principally in macrophages infiltrated into the IA wall, but SMCs also secreted MMP-2 and -9.[49)] Treatment with MMP inhibitors attenuated IA development in animal models,[49,91)] suggesting that MMPs contribute to IA development. At the initiation of IA formation, MMP-2 and -9 were preferentially expressed in SMCs adjacent to the damaged internal elastic lamina (IEL)[92)] in a rabbit model.[42)] This finding is interesting because disruption or loss of IEL is a characteristic feature seen in the early aneurysmal change[93)] and is considered to be the first step of pathological remodeling leading to IA formation. Although it is not clear what induces MMPs in SMCs, increased MMP expression in medial SMCs may be the trigger of IEL damage.

Tissue inhibitors of MMPs (TIMPs) regulate the proteinase activity of MMPs via forming complexes and are the most potent endogenous inhibitors of MMPs in the vascular wall. The expression level of TIMP-1 and TIMP-2 mRNA did not increase in the late stage of IA formation, whereas that of MMP-2 and MMP-9 mRNA continued to increase with IA progression.[94)] Both TIMP-1 deficient mice and TIMP-2 deficient mice showed enhanced IA progression with increased enzyme activity of MMPs,[94)] suggesting that TIMP-1 and TIMP-2 have a protective role for IA progression.

Cathepsins were originally identified as lysosomal or endosomal proteases, but is known to have elastolytic and collagenolytic properties outside lysosomes and endosomes.[95)] The expression of cathepsin B, K, and S was up-regulated in the late stage of the rat IA wall, while that of cystatin C, an endogenous inhibitor of cysteine cathepsins, was reduced with IA progression.[96)] In ruptured and unruptured human IAs, abundant expression of cathepsins was also observed.[47,96)] Treatment with NC-2300, a specific inhibitor for cysteine cathepsins, resulted in the decreased incidence of advanced IAs in rats.[96)] The imbalance between cathepsins and cystatin C, as well as that between MMPs and TIMPs, is likely to contribute to ECM degradation in the IA wall.

In addition to excessive ECM degradation, ECM biosynthesis is attenuated in the IA wall. Expression of procollagen type I, III, and lysyl oxidase (LOX) were reduced in SMCs in rat IA wall.[97)] LOX oxidizes lysine residues in elastin and collagen, stabilizing these fibrous proteins by crosslinking.[98)] Its inhibitor, β-aminopropionitrile (BAPN), has been used to induce IAs in rodent models.[99)] The down-regulation of procollagen type I, III, and LOX in IA walls is, at least in part, mediated by NF-κB activation in SMCs. NF-κB decoy ODNs ameliorated the expression of procollagen type I, III, and LOX in cultured SMCs in vitro and in the IA wall in vivo.[99)]

Ets-1 is the other transcription factor mediating vascular inflammation in SMCs in the IA wall. Ets-1 was expressed and activated principally in vascular SMCs in both experimentally induced rat IAs and human IA walls.[100)] One downstream gene of Ets-1 identified by chromatin immunoprecipitation (CHIP) analysis was MCP-1. Treatment with ets decoy ODNs resulted in the prevention of IA enlargement and diminished expression of MCP-1 in the IA wall.[100)] Other pathways may be involved in the process of IA development. For example, c-Jun N-terminal kinase (JNK), a selective inhibitor of which caused regression of established AAA,[101)] was highly phosphorylated in human IAs.[102)]

In the late stage of IA development, medial SMCs decrease in number, leading to further decrease of ECM biosynthesis. Apoptosis in medial SMCs was found in both human[103)] and rat IAs.[104)] Inducible NOS (iNOS) was highly expressed in the human and rat IA wall.[105)] iNOS has multiple biological effects such as the provocation of inflammatory reaction and the production of reactive oxygen species (ROS) that may be relevant to IA development. Based on experimental results demonstrating that both an iNOS inhibitor[105)] and iNOS deficiency[106)] attenuated IA development and the number of apoptotic cell death in the IA wall, iNOS contributes to IA development, at least in part, by inducing apoptosis in medial SMCs. Interleukin-1 beta (IL-1β) is an isoform of proinflammatory cytokine which functions in a broad array of normal and pathological inflammatory, hematopoietic, and immunologic situations. IL-1β was expressed principally in medial SMCs in the rat IA wall.[107)] In IL-1β deficient mice, apoptosis in SMCs was significantly reduced in number, and induced IAs were smaller in size than those in wild type mice.[107)] Because there is so far no direct evidence that IL-1β induces apoptosis in vascular SMCs, the role of IL-1β in apoptotic cell death of SMCs in IA development remains unclear. Similarly, TNF-α triggers a protease cascade leading to apoptosis through TNFR1 and caspase-8, but caspase-8 has not been shown in the IA wall. Recently, Laaksamo et al. reported that cleaved caspase-9 (intrinsic activator of apoptosis) was significantly increased in ruptured IA walls.[108)] Increased expression of hemeoxygenase-1 (HO-1) in ruptured IAs implies high oxidative stress as a probable cause of the intrinsic activation of cell death.

Role of Macrophages Accumulating in IAs

As macrophage-depleted mice had a substantial lower risk of IA progression,[48)] macrophages are definitely one of the key players in IA development. However, only a part of accelerating effects of macrophages on IA development has been revealed so far. Macrophages are the main source of MMP-2 and -9 production.[49)] Although not completely characterized, macrophages seem to express proinflammatory cytokines such as TNF-α and IL-1β. The other important role of macrophages are the source of ROS, a major inflammatory mediator in vascular diseases. ROS is produced through enzymatic reactions mainly by nicotinamide adenine dinucleotide phosphate oxidase (NADPH), HO-1 and iNOS, and is eliminated by antioxidants such as super oxide dismutase-1 (SOD-1). Expression of NADPH oxidase, HO-1 and iNOS was up-regulated in the rat IA wall, whereas expression of SOD-1 was down-regulated.[109)] Deletion of p47phox, a subunit of NADPH oxidase in neutrophils and macrophages, resulted in attenuated IA development, indicating that macrophage-derived ROS promotes IA development.

Therapeutic Modulation of Vascular Inflammation in the Wall of IAs

Accumulating evidences mentioned above provide support for the role of inflammation in the formation, progression, and rupture of IAs. Based on a hypothesis that the inhibition of inflammatory cascades leads to the prevention of enlargement and rupture of IAs, several therapeutic strategies have been investigated in animal models (Table 2).

Table 2

Therapeutic modulation of vascular inflammation in the wall of intracranial aneurysms

Agent	Pharmacological action	Model	Efficacy	Study
Doxycyclin	Nonspecific MMP inhibitor	Rat model	No effect	Kaufmann et al. (2006)[110)]
Tolylsam	Inhibitor for MMP-2, -9 and -12	Rat model	Decrease in the incidence of advanced IAs	Aoki et al. (2007)[49)]
NC-2300	Specific inhibitor for cysteine cathepsins	Rat model	Decrease in the incidence of advanced IAs	Aoki et al. (2008)[96)]
7ND	Dominant negative mutant DNA of MCP-1	Rat model	Decrease in IA size	Aoki et al. (2009)[58)]
Edaravone	Free radical scavenger	Rat model	Decrease in IA size	Aoki et al. (2009)[109)]
Celecoxib	COX-2 inhibitor	Rat model	Decrease in the incidence of advanced IAs	Aoki et al. (2011)[71)]
Tranilast	Mast cell degranulation inhibitor	Rat model	Decrease in IA sizeIncrease in media thichness	Ishibashi et al. (2010)[51)]
Candesartan	Angiotensin receptor blocker	Rat model	Decrease in IA formation	Tamura et al. (2009)[82)]
Valsartan	Angiotensin receptor blocker	Rat model	Decrease in IA formation	Aoki et al. (2009)[118)]
Olmesartan	Angiotensin receptor blocker	Rat model	No effect	Kimura et al. (2010)[115)]
Ibudilast	Phosphodiesterase-4 inhibitor	Rat model	Decrease in IA stage	Yagi et al. (2010)[117)]
Simvastatin	Pleiotrophic effects of statins	Rat model	Decrease in IA sizeIncrease in media thichness	Aoki et al. (2008)[113)]
Pitavastatin	Pleiotrophic effects of statins	Rat model	Decrease in IA sizeIncrease in media thichness	Aoki et al. (2009)[114)]
Pravastatin	Pleiotrophic effects of statins	Rat model	Decrease in IA formation	Kimura et al. (2010)[115)]
Pravastatin	Pleiotrophic effects of statins	Rat model	Decrease in IA formation at low doseIncrease in IA formation and rupture at high dose	Tada et al. (2011)[116)]
Asprin	Inhibitor of COX-2	Human	Decrease in the incidence of SAHSuppression of macrophage accumulation	Hasan et al. (2011)[119)]Hasan et al. (2013)[120)]
NF-κB decoy ODN	Inhibitor of NF-κB	Rat model	Decrease in IA formation	Aoki et al. (2007)[66)]
Ets decoy ODN	Inhibitor of Ets-1	Rat model	Decrease in IA sizeIncrease in media thichness	Aoki et al. (2010)[100)]
Chimeric decoy ODN	Inhibitor of NF-κB and Ets-1	Rat model	Regression of preexisting IAs	Aoki et al. (2012)[121)]

COX-2: cyclooxygenase-2, DNA: deoxyribonucleic acid, IA: intracranial aneurysm, MCP-1: monocyte chemotactic protein-1, MMP: metalloproteinase, NF-κB: nuclear factor kappa B, ODN: oligodeoxyribonucleotide, SAH: subarachnoid hemorrhage.

Although nonspecific MMP inhibition with doxycycline was not effective in preventing IA formation in a rat model,[110)] a selective inhibitor for MMP-2, -9, and -12, Tolylsam, significantly decreased the incidence of advanced IAs in rats.[49)] Treatment with NC-2300, a specific inhibitor for cysteine cathepsins, also resulted in the decreased incidence of advanced IAs in rats.[96)] DNA encoding a dominant negative mutant of MCP-1 (7ND),[58)] a free radical scavenger (Edaravone),[109)] a COX-2 inhibitor (celecoxib),[71)] and a mast cell degranulation inhibitor (Tranilast)[51)] inhibited IA enlargement in a rat model as well. Inhibiting a specific factor in inflammatory cascades resulted in significant but modest suppression of IA development, implying the need for simultaneous inhibition of multiple factors.

Statins have vascular protective effects known as “pleiotropic” effects in addition to cholesterol lowering effect.[111,112)] Statins protect the endothelial function by increasing NO bioavailability and by inhibiting the expression of proinflammatory cytokines through the inactivation of NF-κB. Statins also inhibit the expression of MMPs and iNOS in both SMCs and macrophages. Various kinds of statins have been demonstrated to have suppressive effects on degenerative changes in the IA wall and IA enlargement in rodent models.[113–115)] Pitavastatin inhibited NF-κB activation and subsequent up-regulation of downstream gene such as MCP-1, VCAM-1, iNOS, IL-1β, and MMP-9 in the rat IA wall.[114)] In contrast to these results, Tada et al. reported that a high dose of pravastatin promoted IA growth and rupture through an increase of apoptosis in the IA wall.[116)] The dose-dependent disparate effects of statins warns us about the risk of overdose of statins. One of the other promising drugs is an inhibitor of phosphodiesterase-4 (Ibudilast), a cyclic adenosine monophosphate-specific enzyme involved in various inflammatory diseases.[117)] Angiotensin receptor blockers (ARBs) were also tested, but showed conflicting results.[82,115,118)]

In the International Study of Unruptured Intracranial Aneurysms (ISUIA), patients with unruptured IAs who used aspirin (acetylic acid) had a lower risk of hemorrhage compared to those who never used aspirin.[119)] Aspirin exerts its antiplatelet and anti-inflammatory actions by irreversible acetylation of COX-1 and -2. It appears that the protective effect of aspirin against IA rupture is mediated in part by inhibition of COX-2 and mPGES1. In fact, expression of COX-2 and mPGES-1 was lower in the aspirin-treated group than in the control group.[120)] Moreover, after 3 months of treatment with aspirin, the signal intensity corresponding to the uptake of ferumoxytol by macrophages in the IA wall was attenuated, showing the anti-inflammatory effect of aspirin on the IA wall.

Treatment with NF-κB decoy ODNs caused a dramatic decrease in the inflammatory response and IA incidence in rats. However, the inhibitory effect was not remarkable when the administration started at the late stage.[66)] Ets decoy ODNs also inhibited IA enlargement and wall thinning of IAs.[100)] Ets-1 is predominantly activated in medial SMCs, whereas activation of NF-κB mainly occurred in both ECs and SMCs. In the next step, we investigated the effect of simultaneous inhibition of NF-κB and Ets-1 by chimeric decoy ODNs and demonstrated a regressive effect of chimeric decoy ODNs in rat IAs.[121)] Chimeric decoy ODNs is the first and only agent which can regress IA development. Unfortunately, decoy ODN is degraded in a few hours by nucleases if it is transvenously or trasnorally administered. For clinical use, the drug form of decoy ODN must be modified in order to have a potent resistance to nucleases.

Future Perspective

Despite remarkable advancements in the field of IA research, the exact mechanisms of flow-driven inflammation remain unknown. There is no doubt that alterations in local flow have an influence on endothelial dysfunction and pathological remodeling in the IA wall. Computational flow dynamics (CFD) emerged in the 2000s in the field of IA research as a powerful tool for analyzing flow pattern in IAs and mechanical properties in the wall.[122–124)] In the future, combining experimental techniques of molecular biology and engineering will bring a breakthrough in the solution of the effect of flow patterns on biological changes in the IA wall.

Mechanisms of IA rupture also remain to be elucidated. Some histopathological features of ruptured IAs have been reported by the comparison between ruptured IAs and unruptured IAs. Pronounced inflammatory cell infiltration, degeneration of the ECM in the wall, and loss of mural cells are characteristic findings observed in ruptured IAs.[46,47)] In contrast, inflammation and degeneration in the vascular wall settle down in the healing process in most unruptured IAs. Positive feedback loops sustaining inflammatory cascades in the IA wall may be clue to the molecular basis of IA rupture. Macrophage imaging by MRI or other molecular imaging indicating active inflammation in the IA wall will become a diagnostic measure selecting dangerous IAs prone to rupture in the future.