Reduced Collagen Biosynthesis Is the Hallmark of Cerebral Aneurysm
Contribution of Interleukin-1β and Nuclear Factor-κB
Background— Reduced extracellular matrix is a prominent feature of cerebral aneurysms (CAs). We previously reported excessive ECM degradation in CA walls. In the present study, we examined collagen biosynthesis in CA walls and the molecular mechanisms underlying it in CA progression.
Methods and Results— RT-PCR and immunohistochemistry showed reduced expression of procollagen type I, III, and lysyl oxidase (LOX) in CA walls. Treatment with the LOX inhibitor β-aminopropionitrile resulted in enhanced progression of CA. Expression of procollagen type I, III, and LOX was inhibited by interleukin-1β (IL-1β) in cultured rat aortic smooth muscle cells (RASMCs) in vitro. Nuclear factor κ-B (NF-κB) was activated in IL-1β-stimulated RASMCs, and treatment with NF-κB decoy oligodeoxynucleotides (ODN) restored reduced expression of procollagen type I, III, and LOX in vitro. NF-κB decoy ODNs ameliorated the expression of procollagen type I, III, and LOX in CA walls in vivo.
Conclusions— Collagen biosynthesis was significantly inhibited at the transcriptional level and in the posttranscriptional enzymatic modification in CA walls through upregulated expression of IL-1β and the NF-κB pathway. Reduced collagen biosynthesis may contribute to CA progression, and inhibition of this process may lead to the prevention of the progression and rupture of CAs.
Cerebral aneurysms (CAs) are a major cause of subarachnoid hemorrhage, which has a 30-day mortality of 45% and a prevalence of moderate-to-severe disability of 30%.1 Pathologically, CA is characterized by decreased collagen content and chronic inflammatory response in aneurysmal walls.2,3 In aneurysmal walls, proteinases such as matrix metalloproteinases (MMPs) and cysteine cathepsins are upregulated and cause excessive degradation of the extracellular matrix (ECM).4–6 A simultaneous imbalance between proteinases and their endogenous inhibitors, such as tissue inhibitor of MMP (TIMPs) and cystatin C, promotes degenerative changes in aneurysmal walls.5,7 ECM biosynthesis in CA walls and its role in the formation and progression of CA is incompletely understood.
Among several types of collagens, collagen types I and III are the most abundant and have an important role in maintaining arterial stiffness. Collagens are formed via a complex biosynthetic pathway involving intracellular and extracellular posttranslational modifications by various enzymes.8 Lysyl oxidase (LOX) oxidizes lysine residues in elastin and collagen, stabilizing these fibrous proteins by crosslinking.9,10 Some inflammatory cytokines, including tumor necrosis factor-α (TNF-α) and interleukin-1β (IL-1β), downregulate procollagen biosynthesis at the transcriptional level in various types of cells.11–14 Activation of nuclear factor-kappa B (NF-κB) is associated with downregulation of the procollagen α1(I) gene in vitro.15,16 We reported that upregulation of IL-1β17 and NF-κB activation18 contributed to CA progression. These findings suggest that IL-1β and NF-κB may regulate collagen biosynthesis in CA walls.
We examined collagen biosynthesis in CA walls and the role of IL-1β and NF-κB in collagen biosynthesis using an experimentally induced CA model and cultured rat vascular smooth muscle cells.
Induction of Experimental CAs in Rats and β-Aminopropionitrile Treatment
Animal experiments complied with Japanese community standards on the care and use of laboratory animals.
One hundred forty-two Sprague-Dawley rats (age, 7 weeks; Oriental Bioservice, Osaka, Japan) were used. CAs were induced as described by Nagata et al.19 After the induction of anesthesia (pentobarbital, 50 mg/kg, i.p.), the left common carotid artery and posterior branches of the bilateral renal arteries were simultaneously ligated with 10-0 nylon. Rats were fed a high salt diet containing 8% sodium chloride. Blood pressure was measured by a tail-cuff method without anesthesia at least twice in each animal.
To examine the role of LOX, some rats were fed food containing 0.12% β-aminopropionitrile (BAPN) (Tokyo Chemicals), a LOX inhibitor. Three months after aneurysm induction with (n=19) or without (n=10) BAPN treatment, the anterior cerebral artery/olfactory artery (ACA/OA) bifurcation was stripped and observed under light microscopy after Elastica van Gieson (EvG) staining. Aneurysm refers to an outward bulging of the arterial wall detected by light microscopy. We analyzed media thickness and aneurysm size to evaluate the pathological changes in aneurysmal walls. Media thickness was expressed as the ratio of the minimal thickness of the media in aneurysmal walls to that in the surrounding normal arterial walls. Aneurysm size was calculated as the mean of the maximal longitudinal diameter and the maximal transverse diameter of aneurysms.
A polyclonal antibody against LOX was generated by immunizing rabbits with keyhole limpet hemocyanin (KLH)-conjugated synthesized peptide (FPQRVKNQGTSDFL), which was preserved among rat, mouse, and human LOX. After 6 times of immunization, antisera were collected and purified using an affinity column. Other primary antibodies used were: goat polyclonal antiprocollagen type I antibody (Santa-Cruz) goat polyclonal antiprocollagen type III antibody (Santa-Cruz); rabbit polyclonal anticollagen type I antibody (Santa-Cruz); rabbit polyclonal anticollagen type III antibody (Chemicom); mouse monoclonal antismooth muscle α-actin antibody (Laboratory Vision); Cy3-conjugated mouse monoclonal antismooth muscle α-actin antibody (Sigma); mouse monoclonal anti-α-tubulin antibody (Sigma); and mouse monoclonal anti-NF-κB p65 subunit antibody that recognizes only DNA-binding form (Chemicon).20
One or 3 months after aneurysm induction, rats were subjected to immunohistochemistry as previously described.4,18 Age-matched male Sprague-Dawley rats were used as controls. The area of positive fields was measured by Scion Image (Scion).
One month (n=6) or 3 months (n=6) after aneurysm induction, Western blotting was performed as previously described.21 Densitometric analyses included 6 independent experiments.
Quantitative (Real-Time) Polymerase Chain Reaction
One month (n=6) or 3 months (n=6) after aneurysm induction, rats were subjected to quantitative RT-PCR using total RNA from the whole circle of Willis as previously described.7,18 A second derivate maximum method was used for crossing point determination, using LightCycler Software 3.3 (Roche, Basel, Switzerland).
The primer sets were: forward 5′-atgctcagctttgtggatacg-3′, reverse 5′-gaccatcaacaccatctctgc-3′ for procollagen α1(I), forward 5′-gcctcccagaacattacatacc-3′, reverse 5′-agactgtcttgctccattcacc-3′ for procollagen α1(III), forward 5′ -cacgcagcagaagaatgg-3′, reverse 5′-cgcagtaccagcctcagc-3′ for LOX, forward 5′-aagtccctcaccctcccaaaag-3′, reverse 5′-aagcaatgctctgagcttccc-3′ for β-actin. The conditions for RT-PCR were: 42 cycles of 95°C 20s for denaturation, 53°C 20 s for annealing, and 72°C 20 s for extension.
Measurement of LOX Activity in Aneurysmal Walls
Total protein from the whole circle of Willis was extracted using Complete Lysis-M kit (Roche) per manufacturer’s instructions (n=5). LOX activity was measured as previously described.22 Briefly, the oxidative product of N-acetyl-3,7-dihydroxyphenoxazine (Amplex Red; Invitrogen) was measured by a spectrophotometer at 570 nm. Background fluorescence was measured and subtracted.
Expression of Procollagens and LOX in Cultured Rat Aortic Smooth Muscle Cells and Cultured Rat Aortic Fibroblasts In Vitro
Rat aortic smooth muscle cells (RASMCs) were purchased from DS Pharma Biomedical. Cells were cultured in Dulbecco modified Eagle medium (DMEM; Sigma) with 10% fetal bovine serum (FBS). Rat aortic fibroblasts were primarily cultured as previously established. Briefly, adventitia tissue of the aorta of 3-week-old SD rats was dissected and dispersed in medium. Cells were collected and incubated in a fibronectin-coated dish (Becton Dickinson) and SmGm medium (Takara) containing epidermal growth factor (EGF), fibroblast growth factor (FGF), insulin, and FBS. After 4 hours of serum starvation, 0, 1, 10, or 100 ng/mL of rat recombinant IL-1β (Abcam) was added to the medium and incubated for 24 hours. Cells were subjected to quantitative PCR and Western blotting as previously described (n=6).
RASMCs cultured on slides (Becton Dickinson) were stimulated by 100 ng/mL of IL-1β for 24 hours. Cells were then subjected to immunohistochemical analysis as described above. Nuclei were stained by 4′, 6-diamino-2-phenylindole (DAPI).
Electrophoretic Mobility Shift Assay
Nuclear protein from IL-1β-stimulated RASMCs was extracted using Qproteome Nuclear Protein Kit (Qiagen) per manufacturer’s instructions. Electrophoretic mobility shift assay (EMSA) for NF-κB binding motif was done by LightShift Chemiluminescent EMSA Kit (Pierce) as previously described.18 Five micrograms of nuclear extract was incubated with 20 fmol of biotin 3′ end-labeled oligonucleotides containing the κB sequence (5′-GGGATTTCCC-3′). After electrophoresis, transfer and crosslinking, the signal was detected by a peroxidase/luminol system (Chemiluminescent Nucleic Acid Detection Module, Pierce). A 200-fold excess amount of nonlabeled κB oligonucleotides (cold probe) was added to confirm specificity.
NF-κB Decoy Oligodeoxynucleotide
NF-κB decoy oligodeoxynucleotide (ODN; 5′-ccttgaagggatttccctcc-3′ and 5′-ggagggaaatcccttcaagg-3′) was synthesized as previously described.4,18,23 Scrambled decoy ODN (5′-ttgccgtacctgacttagcc-3′ and 5′-ggctaagtcaggtacggcaa-3′) was the control. For in vitro study, 100 nmol/L of NF-κB or scrambled decoy ODN was added to the culture medium of RASMCs at the same time as the stimulation by 100 ng/mL of IL-1β. For in vivo study, 40 μg of NF-κB or scrambled decoy ODN in 60 μL PBS was injected into the cisterna magna every 2 weeks under general anesthesia after CA induction. The dose of NF-κB decoy ODN in vitro and in vivo was determined according to a previous study.18
Immunohistochemistry for Human Samples
Human CA samples were obtained from 7 patients who underwent neck clipping for unruptured aneurysms with informed consent. As a control, we used the middle cerebral artery (MCA) obtained at the STA-MCA bypass surgery (n=5). Immunohistochemistry was performed as previously described.5,18
Data (mean±SD) were analyzed by the Mann-Whitney U test for a 2-group comparison, and Kruskal-Wallis 1-way ANOVA on ranks followed by the Turkey-Kramer test for a multiple comparison. P<0.05 was considered statistically significant.
Collagen Amounts in CA Walls
In immunohistochemistry, normal cerebral arterial walls contained a large amount of collagen type I and III (supplemental Figure I, available online at http://atvb.ahajournals.org). Collagen amounts were reduced with CA development (type I 0 months versus 1 month, P<0.01, n=6; type III 1 month versus 3 months, P<0.01, n=6). Western blotting also showed decreased collagen amounts in CA walls (type I 0 months versus 1 month, P<0.01, n=6; type III 1 month versus 3 months, P=0.043, n=6).
Expression of Procollagen Type I and Type III in CA Walls
Procollagen type I and III were highly expressed in the control cerebral arterial wall according to histochemical analysis (Figure 1D and 1G). Their expressions were significantly downregulated in CA walls 1 month and 3 months after aneurysm induction (Figure 1E, 1F, 1H, and 1I). Western blotting revealed a remarkable reduction of procollagen expression in CA walls (type I control versus 1 month, P=0.040, n=6; type III control versus 1 month, P=0.043, n=6; Figure 1J, 1K, and 1L). mRNA expression of procollagen α1(I) and α1(III) were also dramatically downregulated 1 month after aneurysm induction (α1[I] control versus 1 month, P=0.039, n=6; α1[III] control versus 1 month, P=0.044, n=6; Figure 1M and 1N).
Expression and Activity of LOX in CA Walls
LOX was strongly expressed in the entire cerebral arterial wall, including the adventitia (Figure 2D). Immunohistochemistry showed downregulated expression of LOX particularly in the media of CA walls 1 month after aneurysm induction (Figure 2E). After 3 months, LOX expression was scarcely detected in the CA wall (Figure 2F). Western blotting revealed that the expression of LOX protein was reduced with CA progression (control versus 1 month, n=6, P=0.037; Figure 2G and 2H). Quantitative PCR demonstrated decreased mRNA expression of LOX in CA walls 1 month after aneurysm induction (control versus 1 month, n=6, P=0.022; Figure 2I). LOX activity was significantly decreased 1 month after aneurysm induction (control 100±26 [relative activity], 1 month 70.6±10.7, 3 months 56.5±17.4, n=5) (P=0.032, control versus 1 month, n=5; Figure 2J).
Expression of Procollagens and LOX in the Contralateral ACA/OA Bifurcation
In the contralateral ACA/OA bifurcation 3 months after aneurysm induction, expressions of procollagen type I, III, and LOX were decreased compared with the control (before CA induction; type I P=0.92, n=6; type III P=0.068, n=6; LOX P=0.043, n=6; supplemental Figure II). These expression levels were significantly higher than those in the aneurysmal side (type I P=0.038, n=6; type III P=0.016, n=6; LOX P<0.01, n=6).
Effects of the LOX Inhibitor BAPN on CA Formation in Rats
Aneurysm size was significantly larger in the BAPN-treated group (BAPN 90±26.6 μm, n=19; control 71.2±24.5 μm, n=10; P=0.047; Figure 3A). Media thickness of CA walls was significantly thinner in the BAPN-treated group (BAPN 0.27±0.14 μm, n=19; control 0.47±0.20, n=10; P=0.018; Figure 3B). Three months after aneurysm induction, the developmental rate of CA was higher in the BAPN-treated group (89%, n=19) than in the control (80%, n=10; Figure 3C). Blood pressure was not different between the 2 groups (BAPN 160.5±21.7 mm Hg, n=19; control 170.6±14.0 mm Hg, n=10; Figure 3D).
Effects of IL-1β on Procollagen Type I, Type III, and LOX Expression in RASMCs In Vitro
mRNA expression of procollagen α1(I), α1(III), and LOX was significantly downregulated by 10 ng/mL of IL-1β (α1[I] 1 ng/mL versus 10 ng/mL, P=0.027, n=6; α1[III] 1 ng/mL versus 10 ng/mL, P=0.034, n=6; LOX 1 ng/mL versus 10 ng/mL; P=0.036, n=6; Figure 4). Western blotting also showed decreased expression of procollagen type I (P=0.021, n=6), type III (P<0.01, n=6) and LOX (P=0.012, n=6) in response to 100 ng/mL of IL-1β (supplemental Figure III). In cultured rat aortic fibroblasts, downregulation of procollagens and LOX was not elicited by IL-1β (supplemental Figure IV).
Effects of NF-κB Decoy ODN on Procollagen Type I, Type III, and LOX Expression in RASMCs In Vitro
The DNA binding form of the NF-κB p65 subunit demonstrated NF-κB activation by IL-1β in RASMCs (supplemental Figure V). EMSA for the NF-κB binding motif showed one specific complex band in IL-1β-stimulated RASMCs in a dose-dependent manner (Figure 5A). This band was completely abolished by competition with the κB oligonucleotide but not with a mutated binding motif, confirming that the band was specific for NF-κB (Figure 5A). Downregulation of mRNA expression of procollagen α1(I) α1(III) and LOX by IL-1β in RASMCs was restored by NF-κB decoy ODN, but not by scrambled decoy ODN (α1[I] P=0.024, n=6; α1[III] P=0.048, n=6; LOX P=0.013, n=6) (Figure 5B through 5D). Downregulated protein expressions of procollagen type I, III, and LOX by IL-1β in RASMCs were ameliorated by NF-κB decoy ODN, but not by scrambled decoy ODN (supplemental Figure VI).
Effects of NF-κB Decoy ODN on Procollagen Type I, Type III, and LOX Expression in Experimentally Induced Rat CAs In Vivo
IL-1β expression and NF-κB activation were well merged in rat CA walls 3 months after aneurysm induction (Figure 6A and 6B). In the scrambled decoy ODN-treated group, mRNA expressions of procollagen α1(I), α1(III), and LOX were significantly reduced 1 month after aneurysm induction (α1[I] P<0.01, n=6; α1[III] P=0.028, n=6; LOX P<0.01, n=6) (Figure 6C through 6E). In the NF-κB decoy ODN-treated group, expression levels of these mRNAs were significantly higher than those in the scrambled decoy ODN-treated group (α1[I] P<0.01, n=6; α1[III] P=0.012, n=6; LOX P=0.039, n=6; Figure 6C through 6E).
Expression of Procollagen Type I, III, and LOX in Human CAs
Procollagen type I, type III, and LOX were abundantly expressed in control middle cerebral artery especially in the media (supplemental Figure VII). These expressions were decreased in human CA walls.
A reduced collagen amount in arterial walls is one of the prominent pathological features of CAs,2,3 which is confirmed also in the present study (supplemental Figure I). Excessive collagenolysis in CA walls is induced by the imbalanced expression between MMPs and TIMPs,7 and between cathepsins and cystatin C.5 Decreased collagen biosynthesis in CA walls, the other important aspect of degenerative changes in CA walls, was observed in this study. Expression of procollagen type I and III was downregulated in CA walls at the transcriptional level (Figure 1). Some studies showed a reduced level of type-III collagen in patients with CAs.24,25 Although no mutations were found in the type-III procollagen gene in patients with CAs,26,27 reduced production of type-III collagen may be derived from impaired regulation of gene expression of the type-III procollagen gene.
Collagens must be crosslinked to exhibit the normal physical properties of tensile strength. LOX catalyzes the final enzymatic step for it.10 LOX was abundantly expressed in normal cerebral arterial walls, whereas its expression was markedly reduced in CA walls (Figure 2). LOX was principally expressed in medial SMCs, although LOX expression was detected in the adventitia. Cultured fibroblasts are known to express LOX.28 In CA walls, LOX expression was reduced in the media 1 month after CA induction. Three months after CA induction, LOX expression in the adventitia was also reduced. LOX expressions in SMCs and fibroblasts in CA walls are likely to be differentially regulated in CA walls because LOX expression was not affected by IL-1β in cultured rat aortic fibroblasts (supplemental Figure IV). The role of fibroblasts in CA development remains to be elucidated. Treatment with the LOX inhibitor BAPN promoted CA enlargement and thinning of the CA wall in experimentally induced rat CAs (Figure 3). The difference in CA size between the control and the BAPN-treated group was not large (P=0.047), but that of media thickness was significant (P=0.018). Thinning of CA walls increases the susceptibility of CA rupture. These results indicate a role for LOX in maintaining the stiffness of CA walls. This notion is supported by genetic analyses revealing that one single nucleotide polymorphism of LOXL2, a member of LOX family genes, showed a strong association with CA susceptibility.29
We clarified the involvement of IL-1β and NF-κB in the downregulation of procollagen type I and III, and LOX in CA walls. IL-1β downregulates their expressions in a dose-dependent manner in RASMCs (Figure 4). We previously reported upregulated expression of IL-1β in CA walls and impaired progression of experimentally induced CAs in IL-1β-deficient mice.17 IL-1β induces apoptotic cell death in SMCs of arterial walls, thereby exacerbating thinning of the media. IL-1β inhibits ECM biosynthesis in each medial SMC, further promoting degenerative changes in CA walls. IL-1β activated NF-κB (Figure 5A), and NF-κB decoy ODN ameliorated the IL-1β-induced downregulation of procollagen type I, III, and LOX gene expression in vitro (Figure 5B and 5D). NF-κB decoy ODN restored their expressions in experimentally induced CAs in rats (Figure 6C through 6E). These results strongly suggest that IL-1β downregulates the gene expression of procollagens and LOX through NF-κB activation. Coexpression of IL-1β and the DNA binding form of the NF-κB p65 subunit in CA walls further supports this notion (Figure 6B). NF-κB acts as a repressor by binding to the promoter region in some situations. Roebuck et al identified 3 NF-κB binding sites within the human procollagen α1(I) gene promoter in osteoblasts.16 They also showed that titanium particles stimulated the binding of NF-κB subunits p65 and p50 to these sites. It is likely that the direct binding of NF-κB to the promoter region has a functional role in the downregulation of procollagens and LOX gene in the medial SMCs of CA walls.
It is controversial whether collagen content is decreased or increased in abdominal aortic aneurysms (AAAs). Although some studies showed decreased levels of collagen in AAAs,30 others showed increased levels of collagen,31,32 implicating the role of increased collagen in AAA formation. On the other hand, all previous studies reported decreased collagen contents in CAs.2,3 Decreased expression of procollagens and LOX was also demonstrated in human CAs (supplemental Figure VII), strongly suggesting the clinical relevance of results in the present study. This discrepancy may be derived from the difference of the pathophysiology between AAA and CA. CA is formed at an arterial bifurcation suffering from high shear stress, whereas AAA arises at a nonbranching site. Excessive hemodynamic stress attributable to anatomic architecture and hypertension is considered to be the first step of CA formation. Studies using an endothelium-SMC coculture system suggested that shear stress acts on vascular SMCs through endothelial interactions to modulate SMC gene expression.33 It is highly possible that excessive hemodynamic stress transmits inflammatory responses in SMCs, causing reduced ECM biosynthesis through endothelium-SMC interaction. Even in the contralateral side, which is influenced by only hemodynamic stress attributable to hypertension, expressions of procollagens and LOX were decreased compared with the ACA/OA bifurcation before CA induction (supplemental Figure II). Excessive hemodynamic stress affecting the ipsilateral ACA/OA bifurcation further inhibits collagen biosynthesis in arterial walls, thus promoting CA progression.
Fluid shear stress induces the activation of NF-κB, which mediates the proinflammatory cascade leading to ECM degradation18 and the suppression of ECM biosynthesis. In the present study, NF-κB decoy ODN was intrathecally administered and cannot be applied to patients with CAs. When NF-κB blockade will be feasible through an oral or transvenous administration in the future, it may be a promising candidate for prevention of the progression and rupture of CAs.
Sources of Funding
This work was supported by a Grant-in-Aid for Scientific Research (number 17390399) from the Ministry of Education, Science, and Culture of Japan.
Received August 14, 2008; revision accepted March 8, 2009.
Johnston SC, Selvin S, Gress DR. The burden, trends, and demographics of mortality from subarachnoid hemorrhage. Neurology. 1998; 50: 1413–1418.
Kataoka K, Taneda M, Asai T, Kinoshita A, Ito M, Kuroda R. Structural fragility and inflammatory response of ruptured cerebral aneurysms. A comparative study between ruptured and unruptured cerebral aneurysms. Stroke. 1999; 30: 1396–1401.
Aoki T, Kataoka H, Morimoto M, Nozaki K, Hashimoto N. Macrophage-derived matrix metalloproteinase-2 and -9 promote the progression of cerebral aneurysms in rats. Stroke. 2007; 38: 162–169.
Aoki T, Kataoka H, Ishibashi R, Nozaki K, Hashimoto N. Cathepsin B, K, and S are expressed in cerebral aneurysms and promote the progression of cerebral aneurysms. Stroke. 2008; 39: 2603–2610.
Aoki T, Kataoka H, Moriwaki T, Nozaki K, Hashimoto N. Role of TIMP-1 and TIMP-2 in the progression of cerebral aneurysms. Stroke. 2007; 38: 2337–2345.
Roebuck KA, Vermes C, Carpenter LR, Fritz EA, Narayanan R, Glant TT. Down-regulation of procollagen alpha1[I]] messenger RNA by titanium particles correlates with nuclear factor kappaB (NF-kappaB) activation and increased rel A and NF-kappaB1 binding to the collagen promoter. J Bone Miner Res. 2001; 16: 501–510.
Moriwaki T, Takagi Y, Sadamasa N, Aoki T, Nozaki K, Hashimoto N. Impaired progression of cerebral aneurysms in interleukin-1beta-deficient mice. Stroke. 2006; 37: 900–905.
Aoki T, Kataoka H, Shimamura M, Nakagami H, Wakayama K, Moriwaki T, Ishibashi R, Nozaki K, Morishita R, Hashimoto N. NF-kappaB is a key mediator of cerebral aneurysm formation. Circulation. 2007; 116: 2830–2840.
Nagata I, Handa H, Hashimoto N, Hazama F. Experimentally induced cerebral aneurysms in rats: Part VI. Hypertension Surg Neurol. 1980; 14: 477–479.
Aoki T, Kataoka H, Ishibashi R, Nozaki K, Egashira K, Hashimoto N. Impact of monocyte chemoattractant protein-1 deficiency on cerebral aneurysm formation. Stroke. 2009; 40: 942–951.
Nakashima H, Aoki M, Miyake T, Kawasaki T, Iwai M, Jo N, Oishi M, Kataoka K, Ohgi S, Ogihara T, Kaneda Y, Morishita R. Inhibition of experimental abdominal aortic aneurysm in the rat by use of decoy oligodeoxynucleotides suppressing activity of nuclear factor kappaB and ets transcription factors. Circulation. 2004; 109: 132–138.
Kuivaniemi H, Prockop DJ, Wu Y, Madhatheri SL, Kleinert C, Earley JJ, Jokinen A, Stolle C, Majamaa K, Myllyla VV. Exclusion of mutations in the gene for type III collagen (COL3A1) as a common cause of intracranial aneurysms or cervical artery dissections: results from sequence analysis of the coding sequences of type III collagen from 55 unrelated patients. Neurology. 1993; 43: 2652–2658.
van den Berg JS, Pals G, Arwert F, Hennekam RC, Albrecht KW, Westerveld A, Limburg M. Type III collagen deficiency in saccular intracranial aneurysms. Defect in gene regulation? Stroke. 1999; 30: 1628–1631.
Akagawa H, Narita A, Yamada H, Tajima A, Krischek B, Kasuya H, Hori T, Kubota M, Saeki N, Hata A, Mizutani T, Inoue I. Systematic screening of lysyl oxidase-like (LOXL) family genes demonstrates that LOXL2 is a susceptibility gene to intracranial aneurysms. Hum Genet. 2007; 121: 377–387.
Rahkonen O, Su M, Hakovirta H, Koskivirta I, Hormuzdi SG, Vuorio E, Bornstein P, Penttinen R. Mice with a deletion in the first intron of the Col1a1 gene develop age-dependent aortic dissection and rupture. Circ Res. 2004; 94: 83–90.
Chiu JJ, Chen LJ, Chang SF, Lee PL, Lee CI, Tsai MC, Lee DY, Hsieh HP, Usami S, Chien S. Shear stress inhibits smooth muscle cell-induced inflammatory gene expression in endothelial cells: role of NF-kappaB. Arterioscler Thromb Vasc Biol. 2005; 25: 963–969.