Association Between Osteopontin and Human Abdominal Aortic Aneurysm

【摘要】  Objectives- In vitro and animal studies have implicated osteopontin (OPN) in the pathogenesis of aortic aneurysm. The relationship between serum concentration of OPN and variants of the OPN gene with human abdominal aortic aneurysm (AAA) was investigated.

Methods and Results- OPN genotypes were examined in 4227 subjects in which aortic diameter and clinical risk factors were measured. Serum OPN was measured by ELISA in two cohorts of 665 subjects. The concentration of serum OPN was independently associated with the presence of AAA. Odds ratios (and 95% confidence intervals) for upper compared with lower OPN tertiles in predicting presence of AAA were 2.23 (1.29 to 3.85, P =0.004) for the population cohort and 4.08 (1.67 to 10.00, P =0.002) for the referral cohort after adjusting for other risk factors. In 198 patients with complete follow-up of aortic diameter at 3 years, initial serum OPN predicted AAA growth after adjustment for other risk factors (standardized coefficient 0.24, P =0.001). The concentration of OPN in the aortic wall was greater in patients with small AAAs (30 to 50 mm) than those with aortic occlusive disease alone. There was no association between five single nucleotide polymorphisms or haplotypes of the OPN gene and aortic diameter or AAA expansion.

Conclusions- Serum and tissue concentrations of OPN are associated with human AAA. We found no relationship between variation of the OPN gene and AAA. OPN may be a useful biomarker for AAA presence and growth.

In this study the association of osteopontin (OPN) with human AAA was investigated. Serum OPN concentrations were related to AAA presence and growth; however, 5 polymorphisms in the OPN gene were not associated with AAA. Serum OPN may be a useful biomarker for AAA.

【关键词】  abdominal aortic aneurysm genetic polymorphisms osteopontin

Introduction

Abdominal aortic aneurysm (AAA; OMIM#100070) is associated with considerable morbidity and mortality in the Western world. 1 The identification, monitoring, and selection of patients for treatment presently rely on aortic imaging and clinical judgment. 2 The cause of AAAs remains unclear. Identified risk factors for AAA include increasing age, male gender, smoking, coronary heart disease (CHD), hypertension, dyslipidemia, and a positive family history. 3-5 A number of studies have examined the association between circulating and genetic factors with the presence or expansion of AAAs; however, there have been no replicated findings to date. 6-9 Because accumulation of macrophages and lymphocytes is a consistent observation in biopsies of human AAA, circulating markers of inflammation may play a role in the identification of patients with AAA or those more likely to expand to a size requiring treatment. 9,10 See page 439

Osteopontin (OPN, SPP1, OMIM*166490) is a phosphorylated acidic glycoprotein that has been implicated in a large number of physiological and pathological processes including bone remodelling, vascular calcification, and tumor metastasis. 11,12 Experimental studies support a role for OPN in promoting inflammation, proteolysis, and atherosclerosis, which are all integral processes in AAA. 12-18 In fact, OPN has been specifically linked with development of AAA in an animal model. 19 In humans, common variants of the OPN gene have been associated with immunologic diseases and also related to the serum concentrations of the protein. 20-23 The aim of the present study was to assess the relationship between serum OPN concentrations, polymorphisms of the OPN gene, and AAA presence and growth in humans.

Methods

Study Design

We tested the following hypotheses in this study: (1) Serum (a) and tissue (b) OPN concentrations are associated with the presence of AAA; (2) Serum OPN concentration is associated with the growth rate of AAA; (3) Selected single nucleotide polymorphisms (SNPs) in the OPN gene are associated with the presence of AAA; (4) Selected single nucleotide polymorphisms (SNPs) in the OPN gene are associated with the growth rate of AAA. To examine these hypotheses we assessed patients from two different populations. Firstly, we used patients from the Western Australia Screening Study. 5 DNA was available from 4227 subjects of which 689 had an AAA (maximum aortic diameter 30 mm) and 3538 did not (maximum aortic diameter <30 mm). This sample was used to examine hypothesis three. Serum was available from 233 patients with AAA from this population. We selected a similar number of controls (n=233) from the screened population who did not have an AAA. Selection of these subjects was based on aortic diameter (19 to 22 mm) and a similar age range. A previous study suggested normal aortic diameter in this population and age group for men is 19 to 22 mm. 5 This group of screen detected patients with AAA, and matched controls were used to assess hypothesis 1a. To further examine the value of OPN as a biomarker for AAA we examined a second population. This cohort was a consecutive series of patients referred to a tertiary vascular referral center for aortic assessment. 132 of this referral population were found to have an AAA, and 67 did not have an aortic aneurysm. This population of referral detected AAAs (n=132), and controls (n=67) were also used to assess hypothesis 1a. From this referral population 24 patients underwent open aortic surgery and were used to assess hypothesis 1b. Hypothesis 4 was examined in a subset of patients with AAA (n=644) from the screened population in which 6 monthly ultrasound surveillance was carried out for a minimum of 12 months (median 54 months). Hypothesis 2 was examined in a subset of patients in which serum was available, who had been identified from the screened (n=146) or referral (n=52) populations in which ultrasound follow-up was available for a minimum of three years. The study was approved by the relevant ethics committees, and informed consent was obtained from participating patients.

Clinical Data and Definitions

To adjust for other risk factors for AAA development and progression we collected a variety of clinical characteristics including: age, gender, hypertension, diabetes, dyslipidemia, coronary heart disease (CHD), smoking history, peripheral arterial disease (PAD), family history of AAA, and blood pressure (BP) (see supplementary methods for definitions available online at http://atvb.ahajournals.org).

Assessment of Aortic Diameter

The maximum transverse and anteroposterior diameter of the infrarenal abdominal aorta was measured by an experienced vascular sonographer at each site utilizing a 3.75 MHz probe and their own validated ultrasound equipment (Toshiba Capasee, Philips HDI 5000; GE Logic 9). The reproducibility of aortic measurements is regularly assessed in these vascular laboratories. For example, the reproducibility of aortic readings was recently assessed in one laboratory by two repeat readings on 10 patients with AAA. The coefficient of repeatability was 1.2 mm, and the limit of agreement -0.8 to 1.6 mm.

Analysis of Serum and Aortic OPN

Blood was collected from patients after an overnight fast. Serum was stored at -80°C until later batch assessment of OPN concentrations using ELISA according to manufacturer?s instructions and expressed as ng/mL (R&D Systems). 15 This assay was selected because a previous study demonstrated excellent recovery and intra- and interassay reproducibility in our laboratory. 24 In a subgroup of patients undergoing open aortic surgery for occlusive disease (n=6) or AAA (n=18), biopsies taken from the anterior wall of the aorta opposite the inferior mesenteric artery were available for measurement of tissue OPN concentration. The cohort included 6 patients with small AAAs (<50 mm) in which the decision to carry out aortic intervention was also influenced by concurrent symptoms of occlusive iliac atheroma. Aortic biopsies were ground under liquid nitrogen (LN 2 ), and proteins were extracted and quantified as previously described. 25 The concentration of OPN within aortic biopsies was measured using ELISA (R&D Systems) and expressed as pmol/L per mg of protein. Western analysis was used to confirm the results of ELISA as previously described. 25 The localization of OPN in aortic biopsies was investigated with immunohistochemistry as previously reported. 15,25

C-Reactive Protein Assay

C-reactive protein (CRP) was measured by a high-sensitivity assay, with the use of the particle-enhanced immunonephelometry system on the BNII analyzer (Dade Behring). 9

Genotyping

Fifty-five single nucleotide polymorphisms (SNPs) have been identified in the human OPN gene (SPP1, http://www.ncbi.nlm.nih.gov/projects/SNP/). Many of these are within noncoding regions of the gene and have not been related to any functional effects. We selected 5 SNPS (rs4754, location exon 6, rs9138, location 3'-UTR, rs11226616, location exon 7, rs1126772, location 3'-UTR, and rs11730582, location promoter) based on previous association with autoimmune diseases and ability to determine haplotypes predicting serum OPN concentrations. 20-23 Genotyping was performed using real-time quantitative polymerase chain reaction (PCR) utilizing custom TaqMan probes (see supplemental methods, available online at http://atvb.ahajournals.org, and Table 1 ). Error rate was assessed by duplicate analysis of 10% of the samples and found to be <0.05%.

TABLE 1. Comparison of Patients With and Without AAA Undergoing Serum OPN Assessment in the Population Cohort (n=466)

Statistical Analyses

The characteristics of patients with AAA and controls were compared with Pearson Chi-Squared for nominal variables and Mann-Whitney U test for continuous variables ( Tables 1 and 2 ). We used binary logistic regression analysis to examine the relationship between serum OPN and CRP with AAA allowing for other known determinants of aneurysm ( Table 3 ). The ability of serum OPN and CRP to distinguish the patients with AAA was investigated using receiver operating characteristic (ROC) curves and area under the curve (AUC). The association of aortic OPN concentration with aortic diameter group was examined with Kruskal Wallis and Dunnett T3 posthoc tests ( Figure 1 ). Trend analysis revealed that change in aortic diameter over time approximated linear, therefore the association of serum OPN with aortic expansion was assessed with linear multiple regression analysis to allow for other known determinants of AAA progression (initial aortic diameter, diabetes, and smoking history). SNPs were tested for deviations from Hardy-Weinberg equilibrium using an exact Markov-Chain Monte Carlo (MCMC) test. 26 The association between OPN genotype and haplotype with AAA presence, aortic diameter, and growth were examined using a Generalized Linear Model framework accounting for covariates in the software package SimHap vBeta2.1. 27,28 Linkage disequilibrium (LD) between SNPs as measured by the D' and r 2 metric was assessed and haplotype frequencies were estimated using the Bayesian-MCMC framework implemented in Phase v2.1.1 and statistical significance was assessed by means of permutation testing. 29

TABLE 2. Comparison of Patients With and Without AAA Undergoing Serum OPN Assessment in the Referral Cohort (n=199)

TABLE 3. Independent Determinants of AAA in 233 Cases and 233 Controls From a Population Cohort (n=466)

Figure 1. Concentration of OPN in aortic biopsies in relation to aortic diameter. Shown are box plots representing median and interquartile range of OPN concentrations from biopsies of aortic occlusive disease (n=6) and small AAA (n=6) and large AAA (n=12). OPN concentrations were significantly greater in biopsies from patients with small AAAs.

Results

Serum OPN and AAA

We examined the association between serum OPN and AAA in 2 cohorts. Firstly, we compared serum OPN in 233 patients with AAA (220 patients with 30 to 49 mm diameter AAAs and 3 patients with 50 to 54 mm diameter AAAs) and 233 controls with normal aortic diameter (19 to 22 mm) identified from a population screening study ( Table 1 ). Secondly, we examined serum from a cohort of 199 patients referred for aortic assessment at a tertiary vascular center (31 patients with 50 mm diameter AAAs, 101 patients with 30 to 49 mm diameter AAAs, and 67 patients with aortic diameter <30 mm; Table 2 ). Serum OPN was significantly higher in patients with AAA ( Tables 1 and 2 ). OPN was independently associated with the presence of AAA after adjusting for other risk factors in both cohorts. Subjects with serum OPN in the highest tertile were more than twice as likely to have an AAA compared with those with serum OPN in the lowest tertile after allowing for other risk factors in the population cohort ( Table 3 ). Independent odds ratio for upper and middle OPN tertiles were 4.08 (95% CI 1.67 to 10.00, P =0.002) and 3.06 (95% CI 1.30 to 7.23, P =0.01) in the referral cohort. Serum CRP was also associated with the presence of AAA in the population cohort only ( Table 1 ). The association of serum OPN with AAA was independent of CRP level ( Table 3 ). Mean serum OPN was 77.4±43.5 ng/mL in patients with 30 to 49 mm AAAs compared with 77.3±44.5 ng/mL in those with 50 to 80 mm AAAs in the referral cohort, P =0.99. The AUC from ROC analyses of serum OPN in determining the presence of AAA was 0.65 (95% CI 0.60 to 0.70, P <0.0001) compared with 0.61 (95% CI 0.56 to 0.66, P <0.0001) for CRP in the population cohort. Combining serum OPN and CRP did not improve accuracy in detecting AAA (AUC 0.66, 95% CI 0.61 to 0.70). An OPN concentration of 57.2 ng/mL gave the most accurate diagnosis of AAA but only had a sensitivity of 73% and a specificity of 52%. For screening purposes we require a test with high sensitivity. An OPN concentration of 37 ng/mL gave a sensitivity of 95% but a specificity of only 16% for AAA.

Aortic Wall Concentrations of OPN

The concentration of OPN in the aortic wall was measured in 24 patients undergoing open aortic surgery for occlusive disease or AAA. The patients were divided into 3 groups based on aortic diameter as aortic occlusive disease only (diameter <30 mm, n=6, mean diameter 20.8±4.2 mm), small AAA and occlusive disease (diameter 30 to 50 mm, n=6, mean diameter 44.4±6.8 50 mm, n=12, mean diameter 65.5±16.4 mm). Aortic OPN concentration was associated with diameter group ( P =0.01), being higher in patients with small AAAs compared with those with aortic occlusive disease ( P =0.04) or large AAA ( P =0.008; Figure 1 ). The higher concentration in biopsies from small AAAs was confirmed by Western blotting (data not shown). Immunohistochemistry demonstrated that OPN was most markedly expressed within the intimal atheroma of aortic biopsies ( Figure 2 ). In biopsies taken from small AAAs OPN was also demonstrated within the media and adventitia ( Figure 2 b).

Figure 2. Immunohistochemistry for OPN in aortic biopsies. Shown are representative biopsies from an occluded aorta (a), a small AAA (b, 3.5 cm), and a large AAA (c, 8 cm). Brown staining for OPN is present in all layers of the small AAA biopsy but mainly within the intimal atheroma of the biopsies taken from the occluded aorta and large AAA.

Serum OPN and AAA Growth

We assessed the association of serum OPN concentration with aneurysm growth (mean increase in maximum aortic diameter over three years 3.5±3.3 mm, range -5 to 14.2 mm) in 198 patients (mean age 76.7±5.0 years) with small AAAs (mean aortic diameter 34.0±3.8 mm) followed by ultrasound surveillance for 3 years (complete follow-up). Serum OPN was correlated with aortic diameter change ( r =0.24, P =0.001). After adjusting for other known risk factors for aortic expansion (initial aortic diameter, diabetes mellitus, and smoking history) serum OPN predicted AAA growth (F=5.42, standardized coefficient=0.24, P =0.001).

OPN Genotype and AAA Presence and Growth

The characteristics of the 4227 patients genotyped for OPN genetic polymorphisms are shown in Table 4. All loci were in Hardy-Weinberg equilibrium in control samples and were therefore tested for association with AAA ( Table 5 ). No significant associations were found between any of the genotyped loci and AAA or aortic diameter (data not shown). Aortic expansion was not significantly associated with any of the SNPs. The mean LD between loci was D'=0.86 and r 2 =0.58 (supplemental Table II, available online at http://atvb.ahajournals.org). Estimated haplotype frequencies did not differ significantly between individuals with or without AAA and are shown in supplemental Table III (available online at http://atvb.ahajournals.org). No relationship was shown between polymorphisms or haplotypes in OPN and serum concentrations of OPN in the population cohort. For example, mean serum OPN concentration for subjects with polymorphisms at rs11226616 (previously associated with OPN concentration) were 68.08±28.62, 72.40±37.88 and 74.05±37.88 for those who were common homozygotes (n=245), heterozygotes (n=189) or rare homozygotes (n=32), P =0.47.

TABLE 4. Clinical Characteristics of 4227 Subjects Genotyped

TABLE 5. Genotype Frequencies and Hardy-Weinberg Equilibrium (n=4227)

Discussion

In this study we provide evidence to support an association between OPN and human AAA. Serum concentrations of OPN were significantly higher in subjects with AAA, and the association between OPN level and AAA was independent of other risk factors for AAA. We compared OPN with the inflammatory marker CRP in terms of predicting AAA presence ( Tables 1 through 3 ). We found that unlike CRP, OPN predicted AAA presence in both cohorts examined. Serum OPN was also associated with faster growth of small AAAs, which we have previously demonstrated is not the case for CRP. 9 Using ROC curves we found serum OPN concentration to be better at identifying individuals with AAA than CRP but the sensitivity and specificity achieved were too poor to support its value as a screening test. The concentration of OPN within biopsies of small AAAs was greater than demonstrated in occlusive aortic disease or large AAAs. However, no association between variants of the OPN gene, including haplotypes, and AAA was found.

OPN has been shown to promote inflammation by a number of mechanisms, including supporting macrophage and T cell chemotaxis and adhesion, prolonging lymphocyte survival, and enhancing cell mediated immunity. 13,14 OPN is also involved in the activation of proteolytic pathways such as paracrine augmentation of pro-MMP9 activity. 18 Deficiency of and therapy to reduce OPN protects against aneurysm formation in an animal model. 19,30 These findings coupled with the serum studies outlined in this project suggest a role for this cytokine in AAA development and progression. It is possible that the high serum concentrations of OPN in patients with AAA simply reflect a response to AAA formation or are related to the environmental risk factors for the disease. The lack of association between genetic polymorphism for OPN and AAA could be interpreted as supporting this view. However, given the complex pathophysiology of AAA, it is likely that any single genetic variant will only have a weak effect on the development or progression of the condition. The demonstration of high concentrations of OPN in biopsies of small but not large AAA by comparison to patients with aortic occlusive disease alone also brings in to question the importance of OPN in aneurysm progression. It should be noted that the majority of the patients in this study had small rather than large AAAs. Thus the association of serum OPN with AAA refers to this population. We postulate that OPN plays a role in the early development and progression of AAA. High concentrations of OPN in the aorta of some patients with aortic occlusive disease could encourage the excessive positive remodelling and medial destruction typical of AAA. A larger sample of patients with aortic biopsies and serum will be required to investigate this in more detail.

In conclusion our findings support animal data suggesting a role for OPN in the initial development of AAA. Further studies are required to assess whether serum OPN combined with other biomarkers may be useful in identifying patients with AAA or predicting disease progression. We found no evidence to support the importance of genetic polymorphism in OPN in the pathogenesis of AAA.

Acknowledgments

Thanks to the participants and staff involved in the Western Australian AAA Screening Study and Health in Men Study.

Sources of Funding

This project is supported by The Townsville Hospital Private Practice Fund and grant number RO1 HL080010-01 from the National Institutes of Health and NHMRC Practitioner Fellowships (431503 and 458505 to J.G. and P.E.N.).

Disclosures

None.

【参考文献】
  Gillum RF. Epidemiology of aortic aneurysm in the United States. J Clin Epidemiol. 1995; 48: 1289-1298.

Thompson MM, Bell PR. ABC of arterial and venous disease. Arterial aneurysms. BMJ. 2000; 320: 1193-1196.

Lederle FA, Johnson GR, Wilson SE, Chute EP, Littooy FN, Bandyk D, Krupski WC, Barone GW, Acher CW, Ballard DJ. Prevalence and associations of abdominal aortic aneurysm detected through screening. Aneurysm Detection and Management (ADAM) Veterans Affairs Cooperative Study Group. Ann Intern Med. 1997; 126: 441-449.

Alcorn HG, Wolfson SK, Jr., Sutton-Tyrrell K, Kuller LH, O?Leary D. Risk factors for abdominal aortic aneurysms in older adults enrolled in The Cardiovascular Health Study. Arterioscler Thromb Vasc Biol. 1996; 16: 963-970.

Jamrozik K, Norman PE, Spencer CA, Parsons RW, Tuohy R, Lawrence-Brown MM, Dickinson JA. Screening for abdominal aortic aneurysm: lessons from a population-based study. Med J Aust. 2000; 173: 345-350.

Sofi F, Marcucci R, Giusti B, Pratesi G, Lari B, Sestini I, Lo Sapio P, Pulli R, Pratesi C, Abbate R, Gensini GF. High levels of homocysteine, lipoprotein (a) and plasminogen activator inhibitor-1 are present in patients with abdominal aortic aneurysm. Thromb Haemost. 2005; 94: 1094-1098.

Lindholt JS, Vammen S, Fasting H, Henneberg EW, Heickendorff L. The plasma level of matrix metalloproteinase 9 may predict the natural history of small abdominal aortic aneurysms. A preliminary study. Eur J Vasc Endovasc Surg. 2000; 20: 281-285.

Strauss E, Waliszewski K, Gabriel M, Zapalski S, Pawlak AL. Increased risk of the abdominal aortic aneurysm in carriers of the MTHFR 677T allele. J Appl Genet. 2003; 44: 85-93.

Norman P, Spencer CA, Lawrence-Brown MM, Jamrozik K. C-reactive protein levels and the expansion of screen-detected abdominal aortic aneurysms in men. Circulation. 2004; 110: 862-866.

Shimizu K, Mitchell RN, Libby P. Inflammation and Cellular Immune Responses in Abdominal Aortic Aneurysms. Arterioscler Thromb Vasc Biol. 2006; 26: 987-994.

Giachelli CM, Speer MY, Li X, Rajachar RM, Yang H. Regulation of vascular calcification: roles of phosphate and osteopontin. Circ Res. 2005; 96: 717-722.

Rangaswami H, Bulbule A, Kundu GC. Osteopontin: role in cell signaling and cancer progression. Trends Cell Biol. 2006; 16: 79-87.

Denhardt DT, Noda M, O?Regan AW, Pavlin D, Berman JS. Osteopontin as a means to cope with environmental insults: regulation of inflammation, tissue remodeling, and cell survival. J Clin Invest. 2001; 107: 1055-1061.

O?Regan A, Berman JS. Osteopontin: a key cytokine in cell-mediated and granulomatous inflammation. Int J Exp Pathol. 2000; 81: 373-390.

Golledge J, McCann M, Mangan S, Lam A, Karan M. Osteoprotegerin and osteopontin are expressed at high concentrations within symptomatic carotid atherosclerosis. Stroke. 2004; 35: 1636-1641.

Matsui Y, Rittling SR, Okamoto H, Inobe M, Jia N, Shimizu T, Akino M, Sugawara T, Morimoto J, Kimura C, Kon S, Denhardt D, Kitabatake A, Uede T. Osteopontin deficiency attenuates atherosclerosis in female apolipoprotein E-deficient mice. Arterioscler Thromb Vasc Biol. 2003; 23: 1029-1034.

Isoda K, Kamezawa Y, Ayaori M, Kusuhara M, Tada N, Ohsuzu F. Osteopontin transgenic mice fed a high-cholesterol diet develop early fatty-streak lesions. Circulation. 2003; 107: 679-681.

Lai CF, Seshadri V, Huang K, Shao JS, Cai J, Vattikuti R, Schumacher A, Loewy AP, Denhardt DT, Rittling SR, Towler DA. An osteopontin-NADPH oxidase signaling cascade promotes pro-matrix metalloproteinase 9 activation in aortic mesenchymal cells. Circ Res. 2006; 98: 1479-1489.

Bruemmer D, Collins AR, Noh G, Wang W, Territo M, Arias-Magallona S, Fishbein MC, Blaschke F, Kintscher U, Graf K, Law RE, Hsueh WA. Angiotensin II-accelerated atherosclerosis and aneurysm formation is attenuated in osteopontin-deficient mice. J Clin Invest. 2003; 112: 1318-1331.

D?Alfonso S, Barizzone N, Giordano M, Chiocchetti A, Magnani C, Castelli L, Indelicato M, Giacopelli F, Marchini M, Scorza R, Danieli MG, Cappelli M, Migliaresi S, Bigliardo B, Sabbadini MG, Baldissera E, Galeazzi M, Sebastiani GD, Minisola G, Ravazzolo R, Dianzani U, Momigliano-Richiardi P. Two single-nucleotide polymorphisms in the 5' and 3' ends of the osteopontin gene contribute to susceptibility to systemic lupus erythematosus. Arthritis Rheum. 2005; 52: 539-547.

Mochida S, Hashimoto M, Matsui A, Naito M, Inao M, Nagoshi S, Nagano M, Egashira T, Mishiro S, Fujiwara K. Genetic polymorphims in promoter region of osteopontin gene may be a marker reflecting hepatitis activity in chronic hepatitis C patients. Biochem Biophys Res Commun. 2004; 313: 1079-1085.

Chiocchetti A, Indelicato M, Bensi T, Mesturini R, Giordano M, Sametti S, Castelli L, Bottarel F, Mazzarino MC, Garbarini L, Giacopelli F, Valesini G, Santoro C, Dianzani I, Ramenghi U, Dianzani U. High levels of osteopontin associated with polymorphisms in its gene are a risk factor for development of autoimmunity/lymphoproliferation. Blood. 2004; 103: 1376-1382.

Chiocchetti A, Comi C, Indelicato M, Castelli L, Mesturini R, Bensi T, Mazzarino MC, Giordano M, D?Alfonso S, Momigliano-Richiardi P, Liguori M, Zorzon M, Amoroso A, Trojano M, Monaco F, Leone M, Magnani C, Dianzani U. Osteopontin gene haplotypes correlate with multiple sclerosis development and progression. J Neuroimmunol. 2005; 163: 172-178.

Clancy P, Oliver L, Jayalath R, Buttner P, Golledge J. Assessment of a serum assay for quantification of abdominal aortic calcification. Arterioscler Thromb Vasc Biol. 2006; 26: 2574-2576.

Moran CS, McCann M, Karan M, Norman P, Ketheesan N, Golledge J. Association of osteoprotegerin with human abdominal aortic aneurysm progression. Circulation. 2005; 111: 3119-3125.

Wigginton JE, Cutler DJ, Abecasis GR. A Note on exact Tests of Hardy-Weinberg Equilibrium. Am J Hum Genet. 2005; 76: 887-893.

Carter KW, McCaskie PA, Palmer LJ. JLIN: a java based linkage disequilibrium plotter. BMC Bioinformatics. 2006; 7: 60.

Lindstrom MJ, Bates DM. Nonlinear Mixed Effects Models for repeated measures data. Biometrics. 1990; 46: 673-687.

Stephens M, Smith N, Donnelly P. A new statistical method for haplotype reconstruction from population data. Am J Hum Genet. 2001; 68: 978-989.

Gavrila D, Li WG, McCormick ML, Thomas M, Daugherty A, Cassis LA, Miller FJ Jr., Oberley LW, Dellsperger KC, Weintraub NL. Vitamin E inhibits abdominal aortic aneurysm formation in angiotensin II-infused apolipoprotein E-deficient mice. Arterioscler Thromb Vasc Biol. 2005; 25: 1671-1677.

作者单位：Vascular Biology Unit (J.G., J.M., P.C., C.M.), School of Medicine, James Cook University, Townsville; the Laboratory for Genetic Epidemiology (N.S., L.J.P.), Western Australian Institute for Medical Research and UWA Centre for Medical Research, University of Western Australia, Perth; the Department