Inhibition of Notch1-mediated inflammation by intermedin protects against abdominal aortic aneurysm via PI3K/Akt signaling pathway

Notch1 signaling was activated in murine and human AAA

Notch1 signaling participates in local inflammation in the vessel wall, and inhibition of Notch1 signaling reduced AAA in mice by attenuating macrophage-mediated inflammation [14]. First, we detected Notch1 signaling markers in AngII-induced AAA mouse models. The mRNA levels of Notch1 and its ligand DLL4 were higher by 12.3- and 16.7- fold (both P<0.05) with than without AngII treatment. Meanwhile, Notch1 signaling target genes hes1 and hey1 were increased 30.5- and 5.92-fold (both P<0.05) with AngII treatment to induce AAA (Figure 1A). The protein levels of hes1 and the Notch1 active fragment NICD were increased by 76.4% and 67.1% (both P<0.05) with AAA (Figure 1B, 1C). Immunohistochemistry results were consistent with western blot findings, and hes1 and NICD protein were mainly located in vascular adventitia (Figure 1D, 1E). Furthermore, levels of Notch1 signaling markers NICD, hes1 and inducible nitric oxide synthase (iNOS) were obviously higher in human AAA aortas than in normal aortas (Figure 2A–2C).

Figure 1. Notch1 signaling was activated in murine AAA. (A) Quantitative real-time PCR of Notch1, DLL4, hes1 and hey1 mRNA expression in aortas of mice with control (saline), AngII, AngII plus IMD and AngII plus DAPT. (B) Western blot analysis of hes1 and Notch1 signaling pathway marker NICD in aortas with Control, AngII, AngII+IMD and AngII+DAPT treatment. β-actin was a control for protein loading. Results are from 1 representative experiment of 3. (C) Quantification is shown as a ratio of β-actin expression. (D, E) Immunohistochemistry of the protein expression of NICD and hes1 in aortas of mice. Scale bar, 500 μm, 100 μm. Boxes and arrows show enlarged areas. Data are mean ± SD. *P<0.05, **P<0.01 vs. Control. ^#P<0.05, ^##P<0.01 vs. AngII.

Figure 2. Notch signaling and M1 polarization were activated in human aorta of AAA. Control, Patients with non-abdominal aortic aneurysm; AAA, Patients with abdominal aortic aneurysm; Negative-con, Negative control group of patients with non-abdominal aortic aneurysm; Negative-AAA, Negative control group of patients with abdominal aortic aneurysm. Immunohistochemistry of (A, B) the protein expression of NICD and hes1 in patient aortas and (C) the protein expression of M1-type macrophage marker iNOS in patient aortas.

IMD alleviated AngII-induced AAA by blocking Notch1 signaling

Then, we tested the effects of IMD using AngII-induced AAA model. Negative control animals treated with saline did not show AAA features. The morbidity with AngII treatment was 75%, which was consistent with previous studies [28]. IMD treatment significantly reduced the occurrence of AAA to 46.2% (P<0.05) and N-[N-(3,5-difluorophenacetyl)-Lalany]-S-phenylglycine t-butyl ester (DAPT) (Notch1 pathway inhibitor) had a similar effect as IMD, with 53.3% AAA morbidity (P<0.05) (Figure 3A, 3B). The maximum abdominal aorta diameter of AngII-treated mice was markedly dilated as compared with saline treatment and was reversed with IMD and DAPT treatment (Figure 3B). AAA aortas showed intramural thrombus formation, severe aortic dilatation (Figure 3C), elastin degradation (Figure 3D) and increased MMP activity (Figure 3E), which were all alleviated with IMD infusion. DAPT also had a similar inhibitory effect as IMD.

Figure 3. IMD alleviated AngII-induced AAA in mice by blocking Notch1 signaling. (A) Representative photographs of macroscopic features of AAA in male ApoE-/- mice. Control, saline infusion; AngII, AngII (1000 ng/kg/min) infusion; AngII+IMD (300 ng/kg/min) infusion; AngII+DAPT (10 mg/kg) gavage for 4 weeks. The arrows indicate typical AAA in ApoE-/- mice. Scale bars, 5 mm. (B) Incidence of AngII-induced AAA in animals and maximal abdominal aortic diameter of survivor mice at the end of 28 days. (C) Representative HE staining of aortas. Scale bar, 200 μm, 100 μm. (D) Representative VVG staining. Scale bar, 500 μm. (E) Representative in situ zymography of MMP activity in suprarenal aortas from ApoE-/- mice. Scale bar, 100 μm. Data are mean ± SD. *P<0.05, **P<0.01 vs. Control. #P<0.05, ##P<0.01 vs. AngII.

Next we explored whether IMD protected against AAA formation by inhibiting Notch1 signaling activation. Notch1 signaling markers were upregulated during AAA induction (Figure 1). After IMD treatment, the mRNA expression of Notch1, DLL4, hes1 and hey1 were decreased in abdominal aortas 81.3%, 86.8%, 91.2% and 70.1%, respectively (all P<0.05) (Figure 1A), as compared with AngII treatment. Protein levels of hes1 and NICD were reduced 45.6% and 29.6% (both P<0.05) as compared with AngII treatment (Figure 1B, 1C). Immunohistochemistry showed similar results (Figure 1D, 1E). DAPT had similar effects as IMD.

IMD inhibited macrophage-mediated inflammation by Notch1 signaling

Notch1 signaling-mediated macrophage inflammation plays a critical role in AAA pathogenesis [14]. Since activated Notch1 signaling markers were mainly located in vascular adventitia (Figure 1D, 1E), so we supposed that IMD could inhibit AAA by alleviating macrophage-mediated inflammation. First, we detected the expression of representative inflammatory factors in mouse aortas. The mRNA levels of IL-6, monocyte chemotactic protein-1 (MCP-1) and interferon (IFN-γ) were significantly elevated after AngII infusion; IMD and DAPT reversed their expression (Figure 4A). However, the expression of tumor necrosis factor-α (TNF-α) was not changed. It was reported that IL-6 and MCP-1 were the strongest inflammatory factors involved in macrophage migration [3], so we examined the infiltration of macrophages in the vascular wall. Immunofluorescence assay showed that the expression of CD68, a typical macrophage marker, was higher in the vascular wall with AngII than control treatment; IMD and DAPT remarkably reduced the CD68 fluorescence intensity (Figure 4B, 4D). In vitro, using wound-healing assay, lipopolysaccharide (LPS) treatment significantly increased the migration distance of peritoneal macrophages as compared with control treatment; IMD pretreatment could dose-dependently inhibit the migration of macrophages (Figure 4C, 4E).

Figure 4. IMD inhibited macrophage-mediated inflammation via Notch1 signaling. (A) Quantitative real-time PCR of mRNA levels of IL-6, MCP-1, IFN-γ and TNF-α in aortas of Control, AngII, AngII+IMD and AngII+DAPT mice. (B) Representative immunofluorescence analysis of the protein expression of CD68 (green) and DAPI (blue) in aortas of mice. Scale bar, 100 μm. (C) Scratch-wound test to detect the transfer ability of mouse peritoneal macrophages in vitro. PBS, control group; LPS, stimulated group (100 ng/mL, 12 h). IMD was pretreated with a concentration gradient (1, 10, 100 nmol/L, 1 h). Cell migration distance was observed at 12 h after scratch wounding. Scale bar, 200 μm. n=3. (D) Quantification of (B). n=3, Data are mean ± SD. **P<0.01 vs. Control. ^##P<0.01 vs. AngII. (E) Quantification of scratch-wound test. n=3, Data are mean ± SD. *P<0.05 vs. Control. ^#P<0.05 vs. LPS.

IMD inhibited NOD-like receptor family pyrin domain containing 3 (NLRP3) inflammasome activation and M1 polarization via Notch1 signaling

NLRP3 inflammasome activation leads to production of mature IL-1β and IL-18, then recruitment of a large number of inflammatory cells to the vascular wall, which plays an important role in the initiation of inflammatory responses in the early stages of various diseases [29]. To investigate whether the protective effects of IMD against AAA formation are also mediated by NLRP3 inflammasome activation, we examined the apoptosis-associated speck-like protein containing CARD (ASC) expression. Immunohistochemical results showed that the expression of ASC in abdominal aortic tissues was increased in the AngII aneurysm group and reduced with IMD and DAPT treatment (Figure 5A). Next, we further explored the effect of IMD on NLRP3 inflammasome activation in a macrophage line Raw264.7. The protein level of NLRP3, ASC, caspase-1, IL-18, and IL-1β was upregulated during AAA formation, and IMD and DAPT inhibited NLRP3 inflammasome activation (Figure 5B–5D).

Figure 5. NLRP3 inflammasome activation was inhibited by IMD via Notch1 signaling. (A) Immunohistochemistry analysis of the protein expression of ASC in aortas of Control, AngII, AngII+IMD and AngII+DAPT mice. Control, saline infusion; AngII, AngII (1000 ng/kg/min) infusion; AngII+IMD (300 ng/kg/min) infusion; AngII+DAPT (10 mg/kg) gavage for 4 weeks. Scale bar, 500 μm, 100 μm. Boxes and arrows show enlarged areas. (B) Western blot analysis of protein levels of NLRP3, ASC, caspase-1, IL-1β, IL-18 in Raw264.7 macrophage cell line. The cell line was exposed to PBS, AngII (100 nm, 1 h) and DAPT (20 nm, 1 h). (C, D) Quantification of (B). n=3, Data are mean ± SD. *P<0.05 vs. Control. ^#P<0.05 vs. AngII.

Under different microenvironments, macrophages can be polarized to assume different phenotypes and show diverse functions. Generally, M1-type macrophages are considered to have pro-inflammatory phenotypes and M2-type macrophages anti-inflammatory phenotypes [30]. M1 and M2 macrophages polarization contribute to AAA formation [2, 3]. Therefore, we detected the effect of IMD on macrophage polarization in bone-marrow-derived macrophages (BMDMs). IFN-γ and interleukin-4 (IL-4) were used to induce BMDM M1 and M2 polarization, respectively. The mRNA expression of the M1 polarization markers IL-12, IL-1β, TNF-α and iNOS was greatly upregulated after IFN-γ treatment; IMD inhibited M1 polarization marker expression (Figure 6A). Similarly, IL-4 promoted the expression of the M2 polarization markers arginase 1 (Arg1), IL-10 and CD206, and IMD further facilitated their expression (Figure 6B). We further explored the protein level of M1 and M2 markers. IFN-γ increased the M1 marker protein levels of iNOS, IL-1β and CD16, and IMD suppressed the protein levels (Figure 6C, 6D). In contrast, IFN-γ suppressed the M2 protein levels of Arg1, IL-10 and CD206, and IMD restored the levels (Figure 6E, 6F). In mice AAA tissue, the M1 macrophage polarization marker iNOS was increased in Ang II group compared with control, IMD and DAPT could decrease the iNOS expression; M2 macrophage polarization marker Arg1 was activated in Ang II group, which increased more in IMD and DAPT group (Supplementary Figure 2A, 2B).

Figure 6. IMD inhibited macrophage M1 polarization. Control, saline treatment group; LPS+ IFN-γ, M1-type macrophage induction group; IFN-γ+IMD, M1-type macrophages pretreated with IMD; IL-4, M2-type macrophage induction group; IL-4+IMD, M2-type macrophages pretreated with IMD. Quantitative real-time PCR analysis of mRNA levels of (A) M1-type macrophage marker IL-12, IL-1β, TNF-α and iNOS in BMDMs and (B) M2-type macrophage marker Arg1, IL-10 and CD206 in BMDMs. Western blot analysis of (C, D) M1-type macrophage markers IL-12, IL-1β and CD16 and (E, F) M2-type macrophage markers Arg1, IL-10 and CD206 in BMDMs. β-actin was a control for protein loading. n=3, Data are mean ± SD. *P<0.05 vs. Control. ^#P<0.05 vs. IFN-γ.

IMD restrained Notch1 signaling activation by suppressing the expression of ADAM10

Next we explored the molecular mechanism of IMD inhibiting Notch1 signaling activation. ADAM10 and ADAM17 are key proteases during Notch1 signaling pathway activation [11]. In AngII-induced mice abdominal aortas, the mRNA expression of ADAM10 and ADAM17 was increased by 11.9- and 15.1-fold (both P<0.05). IMD decreased ADAM10 and ADAM17 mRNA expression by 68.7% and 70.1% (both P<0.05) (Figure 7A). Results for the protein level of ADAM10 were consistent with mRNA findings (Figure 7B).

Figure 7. IMD inhibited Notch1 signaling activation by ADAM10. (A) Quantitative real-time PCR analysis of mRNA levels of ADAM10 and ADAM17 in aortas of Control, AngII, AngII+IMD and AngII+DAPT in mice. (B) Western blot analysis of protein level of ADAM10 in aortas of Control, AngII, AngII+IMD and AngII+DAPT mice. Western blot analysis of protein levels of (C, D) NICD, the intracellular fragment of Notch1 and the full length Notch1 receptor and (E) ADAM10 and ADAM17 in AngII-induced macrophage cell line Raw264.7. n=3, Data are mean ± SD. *P<0.05, **P<0.01 vs. Control. ^#P<0.05, ^##P<0.01 vs. AngII. (F) Western blot analysis of protein level of ADAM10 in AngII-induced macrophage cell line Raw264.7. Except for the first 3 groups, IMD receptor blocker IMD_17-47 group was added. n=3, Data are mean ± SD. *P<0.05 vs. Control. ^#P<0.05 vs. AngII. ^&P<0.05 vs. AngII+IMD. (G) Western blot analysis of protein level of ADAM10 in AngII-induced macrophage cell line Raw264.7. Except for the first 3 groups, cAMP/PKA inhibitor H89, cGMP/PKG inhibitor Ly83583 and PI3K/Akt inhibitor LY294002 were added. n=3, Data are mean ± SD. *P<0.05 vs. Control. ^#P<0.05 vs. AngII. ^$P<0.05 vs. AngII+IMD.

We further explored the mechanism in a macrophage lineage Raw264.7. First, we detected the protein levels of NICD and full-length Notch1 receptors in Raw264.7 macrophages by western blot analysis. AngII treatment increased NICD level by 2.19-fold (P<0.05), which was reversed by IMD pre-infusion; Notch1 receptor level was similar in 3 groups (Figure 7C, 7D). Also, ADAM10 and ADAM17 protein levels were elevated by AngII and reversed by IMD in macrophages (Figure 7E). The level of Akt phosphorylation showed the opposite trend as NICD, ADAM10 and ADAM17 (Supplementary Figure 1A). Furthermore, IMD17-47, an IMD receptor blocker, intervened in the effect of IMD on ADAM10 expression (Figure 7F). The PI3K/Akt pathway inhibitor LY294002 also blocked the effect of IMD on ADAM10 expression, but cyclic adenosine monophosphate/proteinkinase A (cAMP/PKA) and cGMP/PKG inhibitors had no effect on ADAM10 expression (Figure 7G).

Endogenous IMD alleviated CaCl₂-induced AAA formation

Our previous studies reported that exogenous IMD protected against AAA formation by inhibiting oxidative stress and endoplasmic reticulum stress [27, 28]. To further explore the role of endogenous IMD in AAA, we created IMD transgenic and knockout mice and induced aneurysm by incubation of CaCl₂ in the abdominal aorta. With CaCl₂ incubation, AAA was induced in 73% of mice in a C57 background (Figure 8A, 8B), the maximum abdominal aorta diameter was increased and the protein levels of NICD, hes1 and M1 macrophage marker iNOS were upregulated (Figure 8A, 8B, 8D, 8E, 8F). Moreover, 90% of IMD-knockout mice (P<0.05) developed AAA and showed more increased maximum abdominal aorta diameter (Figure 8A, 8B). However, AAA was induced in only 50% of IMD transgenic mice (P<0.05) and the maximum abdominal aorta diameter was decreased (Figure 8A, 8B). HE staining showed that CaCl₂ induction caused dilated abdominal aorta, vascular wall thickening and damaged vascular elastic plates, which was alleviated by IMD overexpression and aggravated by IMD deficiency (Figure 8C). Moreover, Notch1 pathway markers were upregulated in aortas of IMD-knockout mice and were downregulated in IMD transgenic mice (Figure 8D–8F).

Figure 8. Endogenous IMD alleviated CaCl₂-induced AAA and Notch1 signaling and M1 polarization. Control, wild type C57 mice with saline incubation; CaCl₂, wild-type C57 mice with CaCl₂ incubation; CaCl₂+IMD-Tg, IMD transgenic mice with CaCl₂ incubation to induce AAA; CaCl₂+IMD ko, IMD-knockout mice with CaCl₂ incubation to induce AAA. (A) Representative photographs of macroscopic features of aneurysms in male C57 mice at the end of 28 days. Scale bar, 5 mm. (B) Incidence of CaCl₂-induced AAA in animals and maximal abdominal aortic diameter of surviving mice at the end of 28 days. The vascular diameter > 50% of the average diameter of the control group was considered AAA. *P<0.05 vs. Control. ^#P<0.05 vs. CaCl₂. (C) Representative HE staining. Scale bar, 200 μm. (D–F) Immunohistochemistry of the protein expression of NICD, hes1 and iNOS in aortas of mice.

Materials

Synthetic human IMD_1–53, rat IMD_17–47 and AngII were from Phoenix Pharmaceuticals (Belmont, CA, USA). Alzet Mini-osmotic pumps (Alzet model 1004) were from DURECT Corp. (Cupertino, CA, USA). DAPT, LPS, IFN-γ, IL-4, PI3K/Akt inhibitor LY294002, cAMP/PKA inhibitor H89 and cGMP/PKG inhibitor Ly83583 were from Sigma (MO, USA). Primary antibodies for Notch1 (ab52627), hes1 (ab71559), NICD (ab8925), iNOS (ab3523), NLRP3 (ab214185), caspase-1 (ab1872), CD68 (ab955), CD16 (ab203883), Arg1 (ab91279), CD206 (ab64693, ADAM10 (ab1997), ADAM17 (ab2051), IL-1β (ab2105), IL-10 (ab9969) and IL-18 (ab191860) were from Abcam (Cambridge, UK). Primary antibodies for p-Akt (Ser473) (4060) and Akt (9272) were from Cell Signaling Technology(CST, USA). Primary antibodies for β-actin (sc-1616), GAPDH (sc-25778), ASC (sc-514414), and all secondary antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Other chemicals and reagents were of analytical grade. Mouse macrophage line Raw264.7 were bought from American Type Culture Collection (ATCC, USA).

IMD transgenic and knockout mice

IMD transgenic and knockout mice were generated by Model Animal Research Center of Nanjing University. All animals were maintained at C57BL/6 genetic background, kept under specific pathogen-free conditions, approved by the Animal Care and Use Committee of Peking University Health Science Center. Male mice 8-12 weeks old were used for experiments.

AngII-induced AAA model

ApoE^-/- mice in a C57BL/6 background were from the Animal Center, Peking University Health Science Center (Beijing). All animal care and experimental protocols complied with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication, 8^th Edition, 2011). ApoE^-/- mice (5 months old, male) were randomly assigned to 4 groups for treatment: infusion of saline; AngII (1000 ng/kg/min) [7]; AngII plus IMD_1-53 (300 ng/kg/h) [28]; and AngII plus DAPT. AngII, saline and IMD_1-53 were infused for 28 days. DAPT was administered 3 days before AngII infusion. AngII and IMD_1-53 were administered subcutaneously via an Alzet Mini-osmotic pump. Four weeks later, surviving mice were intraperitoneally injected with pentobarbital (50 mg/kg) for anesthetization.

CaCl₂-induced AAA model

Briefly, as described previously [48], C57BL/6 and IMD transgenic and knockout mice were anesthetized with an intraperitoneal injection of pentobarbital (40 mg/kg) before undergoing laparotomy. The abdominal aorta between the renal arteries and bifurcation of the iliac arteries was isolated from the surrounding retroperitoneal structures. The AAA model was induced with extraluminal CaCl₂. After baseline measurements, 0.5 mol/L CaCl₂ was applied to the external surface of the aorta for 15 min. The aorta was rinsed with 0.9% sterile saline and the incision was closed. NaCl (0.9%) was substituted for CaCl₂ in sham control mice.

Cell culture

BM macrophages were extracted and cultured as previously described [7] with minor modification. Mouse tibias were collected and the BM was flushed with Dulbecco’s modified Eagle’s medium (DMEM; Gibco, USA) and differentiated into primary BMDMs. Cells were cultured in DMEM containing 10% fetal bovine serum (FBS; Hyclone, USA) and incubated at 37° C in a humidified atmosphere containing 5% CO₂. Macrophages were stimulated with IFN-γ and IL-4 to induce M1 and M2 macrophage polarization, respectively [30].

The mouse peritoneal macrophages were collected and cultured as previously described [49] with minor modification. The mouse were injected intraperitoneally with 2 ml of brewer modified thioglycollate medium (BD, USA). Three days later, cells were harvested by peritoneal lavage with 5 ml RPMI-1640 medium (Gibco, USA). Then cells were washed with cold RPMI-1640 medium and seeded on plastic plates. After two hours incubation, the non-adherent cells were washed with RPMI-1640 medium and the adherent cells were monolayer macrophages.

Wound healing assay

The mouse peritoneal macrophages was used in wound healing assay. Macrophages were cultured in 6-well plates and starved for 12 hours in medium supplemented with 0.5%FBS. Then macrophages were scraped, washed with PBS and incubated under 37° C in the presence of LPS or IMD_1-53 for 24 hours in an atmosphere containing 5% CO₂. Images were captured through a Olympus microscope.

Histology and immunofluorescence of mice

Mice were killed and the abdominal aortas were harvested, fixed for 12 hr and embedded in paraffin, and cross-sections (6 μm) were prepared. Paraffin sections were stained with hematoxylin and eosin (HE) and Verhoeff-van Gieson (VVG).

For immunofluorescence assay, freshly cut OCT-embedded aorta sections were incubated with the antibodies anti-CD68 (1:100 dilution) for 1 hr at 37° C, then at 4° C overnight, washed with PBS, then incubated with secondary TRITC-conjugated goat anti-rabbit IgG (1:300 dilution). IgG was a negative control. DAPI was used to stain the nuclei.

Immunohistochemistry of human and mice

Human and mouse AAA aortas were used in immunohistochemistry assay. Surgical specimens were obtained from AAA patients undergoing open surgical repair at Beijing An Zhen Hospital. The control aortic samples were obtained from heart transplantation donors (without AAA) at Beijing An Zhen Hospital. Informed consent was obtained from all patients according to protocols approved by the Medical Ethical Committee of Beijing An Zhen Hospital that comply with the principles outlined in the Declaration of Helsinki.

First, abdominal aortas were fixed in paraformaldehyde for 12 hr and embedded in paraffin, and cross-sections (6 μm) were prepared. For immunohistochemistry, paraffin sections from human or mouse aortas were incubated at 37° C for 1 hr, then at 4° C overnight with primary antibodies for hes1 (1:300 dilution), NICD (1:100 dilution), iNOS (1:200 dilution), Arg1 (1:300 dilution) or ASC (1:100 dilution). On day 2, arterial sections were rinsed with 0.01 mol/L phosphate buffered saline (PBS), then incubated with horseradish peroxidase-conjugated secondary antibodies diluted in 0.1 mol/L PBS for 1 hr at room temperature and redyeing the nucleus with hematoxylin. Normal serum was a negative control.

In situ zymography

Freshly OCT-embedded mouse aorta sections (7.00 um) were used to detect MMP activity. Sections were incubated in a dark and humid chamber at 37° C for 24 hr with MMP fluorogenic substrate DQ–gelatin–FITC (Invitrogen), then washed with PBS. Negative controls were incubated with PBS.

Western blot analysis

Aortic tissue or cells were homogenized in lysis buffer. Then equal amounts of total protein were resolved by 10% or 15% SDS-PAGE (15% for small proteins). Proteins were transferred to a nitrocellulose membrane and blocked with 5% nonfat dried milk for 1 hr, then incubated with the primary antibodies β-actin (1:4000 dilution), GAPDH (1:2000), NLRP3 (1:1000), caspase-1 (1:500), ASC (1:1000), Arg1 (1:1000), IL-1β (1:1000), IL-18 (1:1000), IL-10 (1:200), CD16 (1:300), CD206 (1:1000), Notch1 (1:500), hes1 (1:1000), NICD (1:500), iNOS (1:500), ADAM10 (1:500), ADAM17 (1:500), p-Akt (1:1000) or Akt (1:1000) overnight at 4° C, then with horseradish peroxidase-conjugated secondary antibody for 1 hr. Protein expression was analyzed by using NIH ImageJ and normalized to β-actin expression. All experiments were repeated at least 3 times.

Quantitative real-time PCR

Total RNA from aortic tissue and BMDMs was extracted with Trizol (Applygen, Beijing) and reverse transcribed by using a reverse transcription system (Applygen, Beijing). Real-time PCR amplification included Applied Biosystems 7500 fast PCR System (Life Technologies, USA) and SYBR Green I reagent (Tiangen Biotech, Beijing). The cycle threshold (Ct) was determined as the number of PCR cycles required for a given reaction to reach an arbitrary fluorescence value within the linear amplification range. Relative quantification was performed according to the 2^-ΔΔCt method, with GAPDH as a reference. The primers for real-time PCR were in Table 1.

Statistical analysis

GraphPad Prism v5.00 for Windows (GraphPad Software Inc, San Diego, CA, USA) was used for analysis; data are expressed as mean ± SD. First, Kolmogorov-Smirnov test was used to evaluate the normality of data distribution (P>0.1) and F-test was used to compare variances (P>0.1 take for equal variance). Then Student t test was used to compare 2 groups. One-way ANOVA was used to compare more than 2 groups, and when significant (P<0.05), the Tukey HSD test was used to test for differences between groups. P<0.05 was considered statistically significant. Fisher’s exact test was used for the incidence of aneurysm. P<0.05 was considered statistically significant.