Abdominal Aortic Aneurysm Rupture Is Associated With Increased Medial Neovascularization and Overexpression of Proangiogenic Cytokines

【摘要】  Objective- Matrix metalloproteinase (MMP) activity has been linked to abdominal aortic aneurysm (AAA) rupture. Medial neovascularization (MNV), a histopathologic characteristic of AAAs, involves proteolytic degradation of extracellular matrix by MMPs to facilitate endothelial cell migration. The role of MNV in aneurysm rupture is unknown. This study investigated whether MNV is increased in aneurysm rupture.

Methods and Results- Biopsy samples from aneurysm rupture edge were compared with control biopsy samples from aneurysm wall at the level of rupture and from anterior sac in 12 ruptured AAAs. Further controls were obtained from anterior sac of 10 nonruptured AAAs. MNV, microvessel diameter, maturity index, and inflammatory infiltrate were quantified using morphometric analyses following immunohistochemistry. Expression of proangiogenic mediators was quantified using quantitative real-time-polymerase chain reaction. Compared with anterior sac and aneurysm wall at level of rupture, MNV was increased ( P <0.001) in rupture edge biopsy samples and consisted of smaller diameter ( P <0.001) and more immature microvessels ( P <0.001). mRNA expression of v -integrin, vascular endothelial growth factor, vascular endothelial-cadherin, monocyte chemoattractant protein-1, and vimentin was increased ( P <0.05) in rupture edge biopsy samples.

Conclusions- This study demonstrated increased medial neovascularization and overexpression of proangiogenic cytokines at aneurysm rupture edge. Further investigations into whether this angiogenic response was a causative factor of aneurysm rupture are needed.

This study investigated the extent of neovascularization in abdominal aortic aneurysm rupture and demonstrated increased medial neovascularization and overexpression of proangiogenic cytokines at aneurysm rupture edge. Further investigations into the role of angiogenesis in aneurysm rupture may open novel therapeutic avenues to prevent aneurysm rupture.

【关键词】  aneurysm angiogenesis metalloproteinases neovascularization rupture

Introduction

The natural history of abdominal aortic aneurysms (AAA) is expansion and rupture. Recent insights into the biological processes causing aneurysm expansion have led to translational research investigating the use of novel pharmacotherapeutic agents aimed at retarding aneurysm growth. 1-3 Any medical treatment of AAA must address aneurysm rupture as well as expansion. In contrast to the expansion of AAA, the biological processes causing aneurysm rupture have received little attention and current knowledge relating exclusively to the vascular biology of AAA rupture is lacking. Prospects for the development of pharmacological inhibition of aneurysm rupture therefore remain poor.

Aneurysm rupture was historically considered to be a simple physical process that occurred when wall stress from the circulation exceeded the tensile strength of the aortic wall. 4 Focusing on aortic wall stress as the cause of rupture has led to the view that aneurysm rupture was regulated solely by mechanical factors. Just as the complexity of the atherosclerotic plaque has become apparent, it is now recognized that AAA rupture is a multifaceted biological process involving biochemical, cellular, and proteolytic influences in addition to biomechanical factors. 5

Proteolytic activities of MMPs have been implicated in aneurysm wall weakening and rupture. 5-7 Previous work have demonstrated that high levels of MMPs-8 and MMP-9 were localized to aneurysm rupture edge in humans 5 and experimental studies in rats have shown that inhibition of MMP activity by tissue inhibitor of MMP (TIMP)-1 prevented aneurysm rupture. 6 The MMP family is closely involved in the process of neovascularization 8 and play key proangiogenic roles such as the proteolytic degradation of extracellular matrix (ECM) to facilitate endothelial cell migration during angiogenesis, 9 detachment of pericytes from microvessels undergoing angiogenesis, release of ECM-sequestered angiogenic growth factors, 10 exposure of cryptic proangiogenic integrin binding sites in the ECM, cleavage of vascular endothelial-cadherin endothelial cell-cell adhesions and generation of promigratory ECM fragments. 11 The medial layer of the infrarenal human abdominal aorta, unlike the thoracic aorta, is not normally supplied by microvessels from the adventitial vasa vasorum. 12 Neovascularization of the medial layer is a consistent feature of established AAAs; 13 however, the role of this pathological angiogenic response in aneurysm rupture has never been investigated. 14,15 As pathological neovascularization occurs in close spatial proximity to the collagen and elastin network of the aortic media, the expression of MMPs during such angiogenic response may lead to proteolytic disruption and degradation of aortic media. In view of these considerations, this study assessed the hypothesis that increased medial neovascularization is associated with aneurysm rupture. The extent of medial neovascularization and the gene expression of proangiogenic mediators were evaluated at aneurysm rupture edge and compared with the aneurysm wall at the level of rupture (3 cm away from the rupture edge) and with the anterior aneurysm sac of ruptured and non-ruptured aneurysms.

Methods

Sample Collection

This study was approved by the Local Research Ethics Committee and informed consent was obtained from patients. Twelve consecutive patients who underwent emergency open repair for ruptured AAA (rAAA) (10 males; mean age, 73; range, 61 to 82) were recruited. During aneurysm repair, the rupture site was identified and a strip of aortic wall including the aneurysm rupture site was carefully excised. Nine of the 12 rupture sites were posterior and 3 anterior. The edge of rupture was clearly identified in these strips of aortic wall and subjected to immunohistochemical and molecular analyses. From these strips of aortic wall, tissues at the level of rupture which were 3 cm away from the rupture edge were used as controls. As additional controls, a further strip of aortic wall from the anterior aneurysm sac was excised 2 cm distal to the left renal vein.

Anterior aneurysm sac biopsies (also 2 cm distal to the left renal vein) were similarly obtained from 10 patients who underwent elective open repair for nonruptured AAAs (nrAAA) (10 males; mean age, 75; range, 68 to 84). These nrAAA control tissues were used for immunohistochemical analysis only.

Tissue Processing for Immunohistochemistry

Specimens were cut into &50 mg fragments, embedded in OCT (Raymond & Lamb Ltd, UK) and stored at -80°C.

Tissue Processing for Molecular Analysis

Immediately after excision, samples were placed in RNA Late (Ambion, UK). Samples were washed with RNA Later to remove blood. Any adherent mural thrombus and excess adventitia were carefully dissected from the aortic wall and excluded from analysis. These samples were stored at -20°C until extraction of RNA.

Immunohistochemistry

OCT mounted tissue was cryosectioned (10 µm) onto Superfrost charged slides (VWR International; Leuven, Belgium). Immunohistochemistry was performed with a labeled streptavidin-biotin method using antibodies to CD31 (Clone JC70A diluted 1:50; DAKO, UK), CD68 (Clone EBM11 diluted 1:50; DAKO, UK), -actin smooth muscle ( -SMA) (Clone 1A4 diluted 1:100; DAKO, UK) and v -integrin (Clone P2W7 [CD51] diluted to 8 µg/mL; R&D Systems, UK). Clone P2W7 recognizes the integrin V subunit, which forms heterodimers with ß1 (CD29), ß3 (CD61), ß5, ß6, and ß8 integrin subunits. For each specimen, nonrelevant isotype-matched monoclonal negative controls and no primary antibody negative controls were also performed. Please see http://atvb.ahajournals.org for online expanded methods.

Double immunostaining was performed for CD31 and -actin smooth muscle using DakoCytomation EnVision Doublestain system according to the manufacturer?s protocol.

Morphometric Analysis

Microvessel counts of CD31-positive microvessels were made by light microscopy with x 400 objective (high-power field ) as adapted from previous published studies. 16,17 Quantification of microvessel density (microvessels per HPF) was evaluated within the tunica media in cross-sectional profile. The medial layer was identified by abundance of -SMA-positive smooth muscle cells. Please see http://atvb.ahajournals.org.

Quantification of degree of inflammatory infiltrate and v -integrin immunohistochemical staining was performed by a "semiautomatic" quantification technique 18 using the image analysis program National Institutes of Health Image (US National Institutes of Health).

Total RNA Extraction and cDNA Synthesis

The optimal tissue weight for RNA extraction was between 30 to 60 mg. Samples were ground to a fine powder in liquid nitrogen (LN 2 ) and total RNA was extracted following the manufacturer?s protocol (RNeasy Fibrous Tissue RNA extraction kit; Qiagen, UK). The integrity of the extracted total RNA was confirmed using spectrophotometry and Agilent 2100 bioanalyzer. These samples were reverse transcribed using high-capacity cDNA archive kit (Applied Biosystems, UK) according to manufacturer?s recommended protocol. The resulting cDNA mixture was stored at -80°C until further use.

Quantitative Real-Time Polymerase Chain Reaction

The mRNA levels of candidate angiogenic mediators (target genes) (supplemental Table I, available online at http://atvb.ahajournals.org) were determined by quantitative real-time-polymerase chain reaction (PCR) using TaqMan Gene Expression Assays (Assays-on-Demand Gene Expression Products, Applied Biosystems, UK), which have a FAM reporter dye at the 5' end of the Taqman MGB probe and a nonfluorescent quencher at the 3'-end of the probe. QRT-PCR was performed using an Mx4000 Multiplex Quantitative PCR System (Stratagene, UK).

Statistics

All data were reported as mean±SEM. Differences in morphometric analyses were assessed using 1-way analysis of variance (ANOVA) 2 groups and paired 2-tailed Student t test for comparisons of rupture edge to paired anterior sac or paired level of rupture. Bland Altman analyses were used to compare intra- and inter-observer agreements on microvessel counts. The distributions of mRNA levels of all candidate angiogenic mediators were skewed and logarithmic transformation was used for normalization. Differences in mRNA expression of candidate angiogenic mediators were measured using repeated measures ANOVA. Correlation between mRNA levels of angiogenic mediators and microvessel density was determined by Pearson correlation coefficient test. A value of P <0.05 was considered significant.

Results

Medial Neovascularization

Overall, mean microvessel density in the medial layer was significantly increased ( P <0.001) at rAAA rupture edge compared with nrAAA anterior sac, rAAA anterior sac, and rAAA level of rupture ( Table 1 and Figure 1 ). No intramural hematoma was observed in the medial layer of tissue cross-sections. The mean diameter of medial microvessels was significantly smaller ( P <0.001) at rAAA rupture edge compared with nrAAA anterior sac, rAAA anterior sac, and rAAA level of rupture ( Table 1 and Figure 1 ).

TABLE 1. Summary of Results of Morphometric Analyses

Figure 1. Low-power ( x 50) (top row) and high-power ( x 400) (bottom row) photomicrographs show higher density and smaller diameter CD31-positive microvessels in tunica media of rAAA rupture edge (posterior rupture) compared with nrAAA anterior sac, rAAA anterior sac, or rAAA level of rupture (posterior rupture).

The majority of ruptures (n=9) occurred in the posterior aspect of aneurysm. To confirm that the differences in microvessel density and diameter were not merely the result of anatomic differences between posterior and anterior aneurysm wall, data were analyzed using only the subset of 9 posterior aneurysm ruptures ( Table 1 ). Similar results were obtained in that mean microvessel density was increased ( P <0.001) and mean microvessel diameter was smaller ( P <0.0001) at the posterior rupture edges compared with posterior aneurysm wall biopsy samples from the level of rupture.

The recruitment of smooth muscle cells around nascent vessels is essential to the maturation of the primitive vascular network. 19 The maturity of microvessels was assessed by double immunostaining for the presence of CD31 and -actin smooth muscle (a marker expressed by smooth muscle cells), as adapted from Cao et al. 20 Mature microvessels were defined as those demonstrating an outer smooth muscle cell coat. Both mature and immature microvessels were noted in the medial layers throughout the anterior aneurysm sac and rupture edge biopsies but overall, rupture edge biopsy samples demonstrated significantly less mature microvessels; microvessel maturity index (percentage of mature microvessels): rAAA anterior sac, 83.3±2.3% versus rAAA rupture edge, 47.6±5.4% (paired t test, P <0.001) ( Figure 2 ).

Figure 2. A, Medial microvessel maturity index as percentages of microvessels coated with smooth muscle cells; * P <0.001, paired Students t test. B, x 400 magnification of anterior sac and rupture edge demonstrating double immunostaining for CD31-positive endothelial cells (red: alkaline phosphatase) and -actin smooth muscle (brown: DAB). Small arrows point toward immature microvessels lacking smooth muscle cell coat and larger arrows point toward mature blood vessels with smooth muscle cell coat.

Inflammatory Infiltration

Immunohistochemical staining also revealed a spatial correlation between medial neovascularization and inflammatory infiltration, of which most cells were monocyte/macrophages (identified by antibodies to CD68, data not shown) and lymphocytes (morphological appearance). However, the extent of inflammatory infiltrate in areas of medial neovascularization was not significantly different ( P =0.49) in all representative samples ( Table 1 ).

Gene Expression Levels of Candidate Mediators of Angiogenesis

Normalized values for expression of proangiogenic mediators in the rupture edge, aneurysm wall at the level of rupture and anterior sac are summarized in Table 2. The genes for v -integrin, vascular endothelial growth factor (VEGF), VE-Cadherin, monocyte chemoattractant protein-1 (MCP-1) and Vimentin were significantly upregulated at the rupture edge. No significant difference was observed for the gene expressions of transforming growth factor-1 (transforming growth factor-ß1), basic fibroblast growth factor, hepatocyte growth factor, VEGF receptor 2 (VEGFR2) or CD68.

TABLE 2. Summary of Gene Expressions (Arbitrary Units Normalized to 18S rRNA) of Candidate Angiogenic Mediators

Analyses were performed to determine if there was any correlation between microvessel density and gene expression of candidate proangiogenic cytokines. Tissues from rAAA anterior sac, rAAA level of rupture and rAAA rupture edge were used for these correlation studies. Microvessel density correlated positively with gene expression of VEGF (Pearson correlation coefficient r 2 =0.349, P =0.002), VE-cadherin (r 2 =0.178, P =0.01) and MCP-1 (r 2 =0.198, P =0.007) but not for the rest of the candidate proangiogenic cytokines (supplemental Table II).

v -Integrin Immunohistochemistry

Of the various proangiogenic mediators upregulated at the mRNA level, v -integrin was selected for quantitative assessment of protein expression. All 12 pairs of rAAA rupture edge and anterior sac were subjected to immunohistochemical staining with v -integrin. Figure 3 A demonstrates that the percentage area fraction of v -integrin staining was significantly higher at the rupture edge: rAAA rupture edge, 5.57±1.27% versus rAAA anterior sac, 2.39±0.53% ( P =0.034). Within the anterior sac, v -integrin staining was mainly limited to weak staining of the vasa vasorum within the media and adventitia ( Figure 3 B). Using serially sectioned aneurysm rupture edge specimens, there was significant colocalization of v -integrin with CD31 and -SMA on mature microvessels ( Figure 4A, 4B and 4 C). v -integrin also colocalized to areas of intense inflammatory infiltrate ( Figure 4 A) and with CD31 on immature microvessels that lacked outer smooth muscle cell coats ( Figure 4D, 4E, and 4 F).

Figure 3. A, Quantitative evaluation of v -integrin immunostaining (percentage area of v -integrin staining per X50 magnification); * P =0.034, paired Students t test. B, Low-power magnification ( x 50) photomicrographs show more intense staining with v -integrin at the rupture edge compared with the anterior sac in a ruptured AAA. Within the anterior sac, v -integrin staining was mainly limited to the vasa vasorum within the media (horizontal arrow) and adventitia (vertical arrow).

Figure 4. v -integrin colocalization study from the rupture edge of a ruptured AAA (posterior rupture). A and D, v -integrin. B and E, CD31. C and F, -SMA. Magnifications x 100 (top row) and x 400 (bottom row). Arrowheads indicate colocalization of v -integrin with CD31 and -SMA on mature microvessels. Circles indicate v -integrin staining in areas of intense inflammatory infiltrate. Vertical arrows point toward colocalization of v -integrin with CD31 on immature microvessels that lacked an outer smooth muscle cell coat (absence of staining with -SMA).

Discussion

This study has demonstrated that there were focal areas of increased medial neovascularization localized to aneurysm rupture edge, with a concomitant increase in gene expression of key angiogenic factors ( v -integrin, VEGF, VE-cadherin, MCP-1, and vimentin) and protein expression of v -integrin. The presence of small lumen and immature microvessels supported the notion of an ongoing active angiogenic environment at the rupture edge.

MMPs are essential participants during the angiogenic response. 8 Using similar methods of aneurysm rupture site and anterior sac biopsy samples, elevated concentrations of MMPs-8 and -9 at the rupture site compared with the anterior sacs of ruptured AAAs have previously been reported. 5 Immunolocalization studies have indicated that some MMPs produced in aneurysms were associated with proliferative microvessels. 13 Collectively, these earlier observations together with findings from the present study implicated a potential pathway for aneurysm rupture involving angiogenesis and its elaboration of MMPs. The localized elevation of these proteases in areas of active neovascularization may have a role in focal extracellular matrix degradation and AAA rupture. Indeed, experimental studies in rats have shown that inhibition of MMP proteolytic activity by TIMP-1 6 and plasminogen activator inhibitor (PAI)-1 7 had the potential to prevent aneurysm rupture.

Previous investigations have demonstrated that there was no significant difference in the concentrations of MMP-1, -2, -3, -8, -9, and -13, and TIMP-1 and -2 between the anterior aneurysm sac of ruptured and non-ruptured AAA. 5 Similarly, this study demonstrated that there was no difference in angiogenic activity between anterior sac of ruptured and nrAAAs. Taken together, these data suggested that any factors predisposing to rupture were not ubiquitous within the aortic wall. Because there was no global change in proteolytic activity in ruptured and nrAAAs, mechanisms of aortic rupture were not easily defined from studies comparing wall tissue from ruptured and non-ruptured AAAs. Instead, data from the present study demonstrated the presence of microenvironments localized to aneurysm rupture edges where angiogenesis was increased and gene expression of angiogenic cytokines upregulated. This concept of microenvironments within aneurysms was also supported by Vallabhaneni et al, 21 who demonstrated heterogeneity of tensile strength and MMP activity in aneurysmal walls within patients, suggesting that that there were localized "hot spots" of MMP hyperactivity and discrete areas of weakened aneurysmal wall.

In contrast to previous observations that chronic inflammation in established AAAs was closely correlated with angiogenesis, 14 this study found that the increased medial neovascularization at aneurysm rupture edge was independent of inflammatory infiltrate. Therefore, the development of neovascularization at the rupture edge was likely to be more than a response to tissue injury.

This study reported that both mature and immature microvessels were present at aneurysm rupture edge and anterior sac, although rupture edge biopsy samples demonstrated increased density of small and immature microvessels. If one were to accept that the larger mature microvessels at the anterior sac, presumably at some stage, developed from small and immature microvessels, the absence of progression to rupture at this site presents a dilemma for the hypothesis that neovascularization plays a role in aneurysm rupture. However, as this was an observational study based on end-stage tissue biopsy samples, the temporal sequence of events surrounding the development of neovessels remains conjectural. Clearly, further studies are needed to provide insight into sequential pathogenic factors leading to aneurysm rupture.

Before molecular analysis, care was taken to macroscopically remove adherent mural thrombus and excise adventitia from specimens. Nevertheless, the possibility of including microscopic quantities of these layers during evaluation of gene expression cannot be excluded. This consideration therefore limited the interpretation of the regional origin of the various candidate genes studied. Another limitation of this study was the small sample size.

In conclusion, this study demonstrated localized increased medial neovascularization and overexpression of proangiogenic cytokines at aneurysm rupture edge. The issue of whether the observed increased angiogenesis was a causative factor in aneurysm rupture or simply a consequence of the disease process merits further experimental work. Ongoing work involves investigations into whether VEGF-induced angiogenesis has any influence on aneurysm rupture in angiotensin II-induced aneurysms in apoE -/- hyperlipidemic mice. The pathological features of this AAA model closely mimic human AAAs, although the localized medial dissection and accumulation of intramural hematoma that precede aneurysm rupture in this AAA model are not observed in human aneurysm rupture. Insights obtained from these experiments may open novel therapeutic avenues to prevent AAA rupture.

Acknowledgments

We are grateful to Robert McFarland, Thomas Loosemore, and Josh Derodra for assistance with collecting the tissue specimens, and to Jan Poloniecki for expert help with statistical analysis.

Source of Funding

This work was supported by a research fellowship from the Royal College of Surgeons of England. The authors would like to acknowledge the Biomic Centre of St. Georges, and thank the staff for their help.

Disclosures

None.

【参考文献】
  Walton LJ, Franklin IJ, Bayston T, Brown LC, Greenhalgh RM, Taylor GW, Powell JT. Inhibition of prostaglandin E2 synthesis in abdominal aortic aneurysms: implications for smooth muscle cell viability, inflammatory processes, and the expansion of abdominal aortic aneurysms. Circulation. 1999; 100 (1): 48-54.

Boyle JR, McDermott E, Crowther M, Wills AD, Bell PR, Thompson MM Doxycycline inhibits elastin degradation and reduces metalloproteinase activity in a model of aneurysmal disease. J Vasc Surg. 1998; 272: 354-361.

Thompson RW, Baxter BT. MMP inhibition in abdominal aortic aneurysms. Rationale for a prospective randomized clinical trial. Ann N Y Acad Sci. 1999; 878: 159-178.

Choke E, Cockerill G, Wilson WR, Sayed S, Dawson J, Loftus I, Thompson MM. A review of biological factors implicated in abdominal aortic aneurysm rupture. Eur J Vasc Endovasc Surg. 2005; 30: 227-244.

Wilson WR, Anderton M, Schwalbe EC, Jones JL, Furness PN, Bell PR, Thompson MM. Matrix metalloproteinase-8 and -9 are increased at the site of abdominal aortic aneurysm rupture. Circulation. 2006; 113: 438-445.

Allaire E, Forough R, Clowes M, Starcher B, Clowes AW. Local overexpression of TIMP-1 prevents aortic aneurysm degeneration and rupture in a rat model. J Clin Invest. 1998; 102: 1413-1420.

Allaire E, Hasenstab D, Kenagy RD, Starcher B, Clowes MM, Clowes AW. Prevention of aneurysm development and rupture by local overexpression of plasminogen activator inhibitor-1. Circulation. 1998; 98: 249-255.

Cockerill GW, Gamble JR, Vadas MA. Angiogenesis: models and modulators. Int Rev Cytol. 1995; 159: 113-160.

Herron GS, Werb Z, Dwyer K, Banda MJ. Secretion of metalloproteinases by stimulated capillary endothelial cells. I. Production of procollagenase and prostromelysin exceeds expression of proteolytic activity. J Biol Chem. 1986; 261: 2810-2813.

Kalluri R. Basement membranes: structure, assembly and role in tumour angiogenesis. Nat Rev Cancer. 2003; 3: 422-433.

Rundhaug JE. Matrix metalloproteinases and angiogenesis. J Cell Mol Med. 2005; 9: 267-285.

Wolinsky H, Glagov S. Comparison of abdominal and thoracic aortic medial structure in mammals. Deviation of man from the usual pattern. Circ Res. 1969; 25: 677-686.

Herron GS, Unemori E, Wong M, Rapp JH, Hibbs MH, Stoney RJ. Connective tissue proteinases and inhibitors in abdominal aortic aneurysms. Involvement of the vasa vasorum in the pathogenesis of aortic aneurysms. Arterioscler Thromb. 1991; 11: 1667-1677.

Thompson MM, Jones L, Nasim A, Sayers RD, Bell PR. Angiogenesis in abdominal aortic aneurysms. Eur J Vasc Endovasc Surg. 1996; 11: 464-469.

Holmes DR, Liao S, Parks WC, Thompson RW. Medial neovascularization in abdominal aortic aneurysms: a histopathologic marker of aneurysmal degeneration with pathophysiologic implications. J Vasc Surg. 1995; 21: 761-771.

Vacca A, Ribatti D, Roncali L, Lospalluti M, Serio G, Carrel S, Dammacco F. Melanocyte tumor progression is associated with changes in angiogenesis and expression of the 67-kilodalton laminin receptor. Cancer. 1993; 72: 455-461. <a href="/cgi/external_ref?access_num=10.1002/1097-0142(19930715)72:2

Weidner N, Folkman J, Pozza F, Bevilacqua P, Allred EN, Moore DH, Meli S, Gasparini G. Tumor angiogenesis: a new significant and independent prognostic indicator in early-stage breast carcinoma. J Natl Cancer Inst. 1992; 84: 1875-1887.

Furness PN, Rogers-Wheatley L, Harris KP. Semiautomatic quantitation of macrophages in human renal biopsy specimens in proteinuric states. J Clin Pathol. 1997; 50: 118-122.

Carmeliet P, Conway EM. Growing better blood vessels. Nat Biotechnol. 2001; 19: 1019-1020.

Cao R, Brakenhielm E, Pawliuk R, Wariaro D, Post MJ, Wahlberg E, Leboulch P, Cao Y. Angiogenic synergism, vascular stability and improvement of hind-limb ischemia by a combination of PDGF-BB and FGF-2. Nat Med. 2003; 9: 604-613.

Vallabhaneni SR, Gilling-Smith GL, How TV, Carter SD, Brennan JA, Harris PL. Heterogeneity of tensile strength and matrix metalloproteinase activity in the wall of abdominal aortic aneurysms. J Endovasc Ther. 2004; 11: 494-502.

作者单位：Department of Cardiovascular Science, Academic Vascular Surgery Unit, St George?s, University of London, UK.