Annals of Vascular Surgery
Volume 21, Issue 1 , Pages 61-66, January 2007

Cyclooxygenase-2 Expression and Its Association with Increased Angiogenesis in Human Abdominal Aortic Aneurysms

  • K.S. Chapple, FRCS

      Affiliations

    • Academic Unit of Vascular Surgery, University of Leeds, St. James's University Hospital, Leeds, UK
    • Corresponding Author InformationCorrespondence to: K.S. Chapple, FRCS, Department of Colorectal Surgery, The Alfred Hospital, Commercial Road, Melbourne, Vic 3004, Australia
    • ,
  • D.J. Parry, FRCS

      Affiliations

    • Academic Unit of Vascular Surgery, University of Leeds, St. James's University Hospital, Leeds, UK
    • ,
  • S. McKenzie, MRCS

      Affiliations

    • Academic Unit of Vascular Surgery, University of Leeds, St. James's University Hospital, Leeds, UK
    • ,
  • K.A. MacLennan, MRCPath

      Affiliations

    • Department of Histopathology, University of Leeds, St. James's University Hospital, Leeds, UK
    • ,
  • P. Jones, PhD

      Affiliations

    • Molecular Medicine Unit, University of Leeds, St. James's University Hospital, Leeds, UK
    • ,
  • D.J.A. Scott, FRCS

      Affiliations

    • Academic Unit of Vascular Surgery, University of Leeds, St. James's University Hospital, Leeds, UK

Leeds, United Kingdom

Article Outline

Although the mechanism whereby non-steroidal anti-inflammatory drugs may reduce abdominal aortic aneurysm (AAA) development is unknown, one potential route is via inhibition of the cyclooxygenase (COX) enzyme. Despite the fact that evidence from animal models suggests a role for the COX-2 isoform in promotion of AAA development, only very limited data exist on COX-2 expression in human AAAs. Semiquantitative immunohistochemistry for COX-2 was performed on a series of formalin-fixed, paraffin-embedded human AAAs (n = 49). Associated clinicopathological data, including the degree of inflammatory cell infiltration and neorevascularization, were obtained. COX-2 protein was detected in 46 of 49 (94%) human AAAs. Expression of COX-2 protein varied widely between AAAs. COX-2 protein localized to cells in the inflammatory infiltrate with a morphology characteristic of macrophages. COX-2 expression increased with the extent of inflammatory cell infiltration (P < 0.001) and with the degree of AAA neorevascularization (P < 0.001). Logistic regression analysis identified neorevascularization (P < 0.001) as the only significant independent predictor of COX-2 positivity in human AAAs. COX-2 protein is present at increased levels in the majority of human AAAs and is expressed by mononuclear cells in the inflammatory cell infiltrate. Promotion of angiogenesis by COX-2 may play a role in AAA development.

 

Back to Article Outline

Introduction 

Despite the considerable morbidity and mortality associated with abdominal aortic aneurysm (AAA),1 its pathogenesis is incompletely understood. The inflammatory response may be important since the presence of numerous inflammatory cells, in particular macrophages, within the aortic wall is a notable feature.2 Commensurate with this, evidence from in vivo studies suggests that non-steroidal anti-inflammatory drugs (NSAIDs) may reduce expansion of AAAs.3, 4 Although the mechanism of this effect remains unclear, one potential pathway is via inhibition of the cyclooxygenase (COX) enzyme. Two main isoforms of COX exist, COX-1 and COX-2. COX-1 is expressed constitutively in most normal human tissues, whereas COX-2 is expressed at low levels in inflammatory cells but is highly inducible by cytokines and growth factors.5 COX-2 expression is absent in normal human aorta6 but is upregulated in human atherosclerotic lesions,6 experimentally induced AAAs,4 and human AAAs.6, 7, 8, 9 However, despite the fact that COX-derived prostanoids such as prostaglandin E2 (PGE2) have a number of effects that are associated with AAA expansion,8, 10, 11, 12 COX-2 expression in human AAAs has only been studied in a few studies utilizing small numbers of AAAs.6, 7, 8, 9 Therefore, we performed an immunohistochemical study of COX-2 expression in a large cohort of human AAAs and analyzed its association with a number of clinicopathological variables, including neorevascularization.

Back to Article Outline

Materials and Methods 

Human AAAs 

Patients undergoing AAA repair in a single center over a 3-year period were included in the study. During surgery a 1 × 1 cm full-thickness segment of aneurysmal anterior aortic wall was excised. Tissue was formalin-fixed and paraffin-embedded prior to routine histopathological analysis.

Clinicopathological Data 

A prospective database of clinicopathological variables was maintained for all patients undergoing AAA repair. The data are described in Table I.

Table I. Clinicopathological variables collected for all patients undergoing AAA repair
Clinicopathological featureNumber (%)
Gender42 male:7 female
Age range (years)52-90
Pathological type of AAA
Atherosclerotic46 (94)
Inflammatory3 (6)
Mode of presentation
Elective42 (86)
Emergency7 (14)
Cerebrovascular disease
Asymptomatic carotid stenosis1 (2)
Previous transient ischemic attack3 (6)
Stroke with recovery2 (4)
Coronary artery disease
Stable angina12 (25)
Myocardial infarction7 (14)
Smoking habit
Nonsmoker17 (34)
Ex-smoker19 (39)
Current smoker13 (27)
Hypertension20 (41)
Congestive cardiac failure7 (14)
Cardiac valvular disease7 (14)
Cardiac arrhythmia6 (12)
Hyperlipidemia13 (27)
Diabetes mellitus2 (4)
Respiratory disease13 (27)
Renal impairment8 (16)
Concurrent renal dialysis2 (4)

COX-2 Immunohistochemistry 

After routine histopathological analysis, COX-2 immunohistochemistry was performed on formalin-fixed, paraffin-embedded sections by a modification of a technique previously described.13 Briefly, after blocking of endogenous peroxidase activity with 0.6% (v/v) hydrogen peroxide in 100% methanol for 15 min at room temperature, sections underwent antigen retrieval by pressure cooking (100°C, pressure 15 psi) in 10 mmol/L citrate buffer, pH 6.0, for 60 sec prior to cooling in tap water and incubation with the primary antibody (affinity-purified rabbit polyclonal anti-human COX-2 immunoglobulin G [IgG, 100 μg/mL; IBL, Gunma, Japan], diluted 1:25 in phosphate-buffered saline). The secondary antibody was biotinylated swine anti-rabbit IgG (Dako, Carpinteria, CA; dilution 1:200). Following incubation with streptavidin/biotin-horseradish peroxidase complex (Dako), sections were visualized using 3,3′-diaminobenzidine tetrahydrochloride (0.7 mg/mL). Negative controls, including antibody preadsorption with its cognate peptide, were performed as described previously.13 Human sporadic colorectal carcinoma was used as a positive tissue control.

Semiquantitative Analysis of Pathological Variables 

COX-2 immuoreactivity was determined by a semiquantitative method based on both degree and intensity of staining. The percentage of immunoreactive cells was combined with an estimate of the staining intensity to produce a scale of 0-3 (0, no staining; 1, small numbers [<10%] of COX-2-positive cells; 2, 10-50% of cells positive for COX-2; 3, >50% of cells COX-2-positive with intense cellular staining).13 Scoring was performed separately for the tunica intima, tunica media, and tunica adventitia; and these scores were combined to produce an overall AAA COX-2 expression score (scale 0-9). The degree of inflammatory infiltrate was also assessed using a similar scoring system. Inflammatory cell scores for the tunicae intima, media, and adventitia (0-3 for each area) were combined to produce an overall AAA inflammatory cell score (scale 0-9).

The extent of neorevascularization in each AAA was determined on hematoxylin and eosin–stained sections using morphometric assessment. Each AAA was examined at ×200 magnification. The number of vessel-like structures with characteristic morphological features was noted. Any vessel containing red blood cells was included. Vessel lumens were not necessary to determine the presence of a microvessel. Neorevascularization was scored on a scale of 0-3 (0, no microvasculature present; 1, minimal numbers of microvessel present; 2, moderate numbers of microvessel present; 3, large numbers of microvessel present).

Scoring was performed by two observers, blinded to the origin of sections. Scoring was performed independently, and a consensus decision was reached for the small numbers of cases in which scoring differed. Unless otherwise stated, all scores are expressed as median (interquartile range [IQR]).

AAA Computed Tomography 

Where available, preoperative computed tomographic (CT) scans were assessed as to the maximum AAA abdominoposterior size and the presence or absence of any intraluminal thrombus. If present, the site of any thrombus (anterior, posterior, or circumferential) and the percentage of AAA lumen enclosed by thrombus were also assessed. All radiological scoring was performed by a single observer, blinded to clinicopathological data (including COX-2 expression scores).

Statistical Analysis 

Clinicopathological variables were entered into a logistic regression model with forward conditional selection. The significance of differences in COX-2 protein expression related to those clinicopathological factors identified as independent predictors by logistic regression was tested using the Mann-Whitney U-test. The significance of differences in inflammatory cell infiltrate between individual areas of each AAA was tested using the Kruskal Wallis one-way analysis of variance (ANOVA). The relationship between COX-2 expression score and both degree of inflammatory cell infiltrate and neorevascularization was tested with the Spearman rank correlation coefficient. Statistical significance was assumed if the P value was <0.05.

Back to Article Outline

Results 

Clinicopathological Features of Human AAAs 

Forty-nine formalin-fixed, paraffin-embedded human AAAs were studied from 49 patients. Clinicopathological features are described in Table I.

Radiological Characteristics of Human AAAs 

Preoperative CT scans were available for 41 (84%) patients. Median anteroposterior AAA size was 6.8 (IQR 6.0-7.7) cm. Intramural thrombus was present in 38 of 41 (93%) AAAs (anteriorly in 14 [34%], posteriorly in 3 [7%], and circumferentially in 21 [51%]). Thrombus occupied 39% (IQR 27-56%) of the lumen.

Inflammatory Infiltrate and Neorevascularization in AAAs 

Inflammatory cells were visualized in 47 of 49 (96%) AAAs. In 10 AAAs, the degree of cellular inflammatory infiltrate was uniform throughout the tunica intima, tunica media, and tunica adventitia. However, in the majority of AAAs (n = 37), there was a wide range in the extent of the inflammatory infiltrate, with some areas demonstrating a high inflammatory cell count while other areas within the same aneurysm had only minimal numbers of inflammatory cells. Overall, inflammatory cells were observed less frequently in the tunica intima than in the tunica media or adventitia (median tunica intima inflammatory cell score 1 [IQR 0-1], tunica media inflammatory cell score 2 [1-2], tunica adventitia inflammatory cell score 2 [2-3]; P < 0.001, Kruskal-Wallis ANOVA). Neorevascularization scores are described in Table II.

Table II. Neorevascularization scores for human AAAs
Neorevascularization scoreNumber of AAAs (%)
05 (10)
111 (22)
222 (45)
311 (22)

Semiquantitative Analysis of COX-2 Expression in Human AAAs 

Overall, COX-2 protein was detected in 46 of 49 (94%) formalin-fixed, paraffin-embedded human AAAs. Immunoreactive COX-2 was localized to cells in the tunica adventitia in 44 (90%) cases, to the tunica media in 46 (94%) cases, and to the tunica intima in 32 (65%) cases. Positive staining for COX-2 was also evident on the endothelial cell lining of capillaries in the aortic wall and on the endothelial lining of the tunica intima. The extent of cellular staining varied markedly between AAAs. In some cases (n = 12), only small numbers of COX-2-positive cells were present throughout the AAA (combined COX-2 expression score for all areas 1-3), while in other AAAs (n = 10) large numbers of COX-2-positive cells occupying most of the cellular infiltrate were evident (combined COX-2 expression score 7-9, Fig. 1A). The close proximity of COX-2-positive cells to capillaries was a prominent feature (Fig. 1B). COX-2-positive cells often exhibited a phenotype characteristic of macrophages (mononuclear cells with large ovoid nuclei, Fig. 1C).

  • View full-size image.
  • Fig. 1 

    COX-2 immunohistochemistry on formalin-fixed, paraffin-embedded human AAAs. A Large numbers of COX-2-positive cells (arrow) within the tunica intima. COX-2 expression score 3. Magnification ×200. B COX-2-positive cells were often aggregated around capillaries (arrow). Note that the capillary endothelial cells are also COX-2-positive. Magnification ×200. C COX-2-positive cells (arrow) were observed to have a morphology characteristic of macrophages. Magnification ×400.

Clinicopathological Features Associated with COX-2 Expression in Human AAAs 

Logistic regression identified neorevascularization (P < 0.001) as the only significant independent predictor of COX-2 positivity in human AAAs. Univariate analysis demonstrated increased COX-2 expression in AAAs exhibiting neorevascularization (median COX-2 expression score 6 [IQR 3–6], n = 44) than in AAAs with no neorevascularization (median COX-2 expression score 1 [IQR 0–4], n = 5; P = 0.006, Mann-Whitney U test). The degree of AAA neorevascularization correlated significantly with the overall COX-2 expression score (r = 0.76, P < 0.001, Spearman rank correlation coefficient).

The overall AAA COX-2 expression score correlated with the degree of inflammatory infiltrate (r = 0.77, P < 0.001, Spearman rank correlation coefficient). No significant differences in AAA COX-2 expression scores were identified related to gender, mode of presentation, pathological type of AAA, presence of cerebrovascular or coronary artery disease, use of renal dialysis, preoperative renal function, smoking habit, AAA size, presence or site of intramural thrombus, or presence of hypertension, congestive cardiac failure, cardiac valvular disease, cardiac arrhythmias, hyperlipidemia, diabetes, or respiratory disease.

Back to Article Outline

Discussion 

Expression of COX-2 in human AAAs has only been previously investigated in a few studies utilizing limited numbers of AAAs.6, 7, 8, 9 Therefore, we performed an immunohistochemical study of COX-2 protein expression in a large series of formalin-fixed human AAAs using a previously validated anti-human COX-2 antibody.13 COX-2 protein was expressed, at variable levels, by the majority (94%) of AAAs. Previous reports have noticed COX-2 expression in all of the small numbers of AAAs studied.6, 7, 8, 9 It is interesting to speculate why a small number (6%) of AAAs did not contain immunoreactive COX-2 protein. Reduction of COX-2 gene transcription has been described by antioxidants,14 corticosteroids,15 and aspirin.16 We did not obtain data on concurrent use of corticosteroids or NSAIDs and, therefore, cannot discount that this may have biased COX-2 expression data.

COX-2 protein was expressed predominantly by cells in the inflammatory cell infiltrate within the aortic wall, which is in agreement with previous studies.6, 7, 8, 9 Endothelial cells also stained strongly for COX-2, as has been previously noted.6, 8, 9 Currently, only limited data exist on the nature of the COX-2-expressing cellular population in AAAs. In vitro, COX-2 can be induced in a variety of cell types, including vascular endothelial cells,17 vascular smooth muscle cells,18 and transformed monocytes.19 However, in elastase-induced AAAs, COX-2 protein expression appears to be restricted to ED2-positive mononuclear cells in the inflammatory cell infiltrate.4 Using in situ hybridization, Holmes et al.7 noted restriction of COX-2 mRNA expression to macrophage-like cells in the inflammatory infiltrate in human AAAs. Macrophages isolated from human AAAs using enzymatic dispersion and anti-CD14 immunomagnetic separation have also been noted to express COX-2 protein.8 Immunohistochemical studies of COX-2 protein expression in small numbers of human AAAs have also identified the presence of COX-2 in macrophage-like cells, predominantly at the medial/adventitial junction.6, 8, 9 Coimmunofluorescence studies, utilizing a macrophage marker such as CD68, would confirm that macrophages are the predominant COX-2-expressing cell type in human AAAs.

At present, direct evidence to support a role for COX-2 in the promotion of AAA development is lacking. However, the COX inhibitor indomethacin is recognized to reduce the rate of expansion of elastase-induced AAAs;3, 4 and in a small case-control study of 15 patients, Walton et al.8 noted a reduction in AAA growth rate in association with NSAID usage. Since inflammation is recognized to be a crucial factor in AAA development,20 our finding that COX-2 was consistently expressed at higher levels as the inflammatory cell infiltrate increased is consistent with the promotion of human AAA development via a COX-2-dependent pathway, although we cannot exclude the possibility that COX-2 may be merely a surrogate marker of inflammatory cell recruitment.

The nature of potential activating stimuli for COX-2 in human AAAs remains uncertain. COX-2 protein localized to lining endothelial cells of the tunica intima, which is consistent with fluid shear stress acting as a stimulus for COX-2 expression.21, 22 We did not note any association between smoking habit and COX-2, although benzo[a]pyrene, a polycyclic aromatic hydrocarbon found in tobacco smoke, is recognized to induce COX-2 in vascular smooth muscle cells.23 In addition, COX-2 expression was not related to the presence of intramural thrombus. This is surprising considering that within AAAs macrophages are more concentrated in areas covered by thrombus24 and that intramural thrombus causes localized hypoxia,25 which is recognized to promote COX-2 expression.26

We noted an association between the degree of neorevascularization and increased COX-2 expression in human AAAs. In addition, COX-2-expressing cells tended to aggregate in areas of high microvessel density. These observations are consistent with recent evidence that angiogenesis plays a key role in the pathogenesis of AAAs. Angiogenic factors such as vascular endothelial cell growth factor and keratinocyte growth factor27 are expressed at increased levels in AAAs. Compared to the normal aortic wall, neorevascularization is increased in all layers of an AAA.28, 29, 30, 31, 32 The degree of neorevascularization in AAAs is recognized to correlate with the overall extent of the inflammatory cell infiltrate,30, 31 with inflammatory cells tending to aggregate in areas of focal neorevascularization.28, 29 Data from in vitro and in vivo studies suggest that inflammatory cytokines themselves may promote angiogenesis.32 However, there is increasing evidence that COX-2 may have a role in this angiogenic process. Selective COX-2 inhibition has been demonstrated to attenuate angiogenesis in several in vitro33, 34, 35 and in vivo35, 36, 37 models. Moreover, interleukin-1β-dependent angiogenesis is markedly reduced in Cox-2 knockout mice.35 In conjunction with our data, these observations suggest that the angiogenic response in AAAs may be a COX-2-dependent process.

In summary, we demonstrate that COX-2 protein is expressed in the majority of human AAAs by macrophage-like cells within the inflammatory cell infiltrate. COX-2 expression is associated with neorevascularization, suggesting a novel mechanism whereby COX-2 may promote AAA development.

Back to Article Outline

References 

  1. Ashton HA, Buxton MJ, Day NE, et al. The Multicentre Aneurysm Screening Study (MASS) into the effect of abdominal aortic aneurysm screening on mortality in men: a randomised controlled trial. Lancet. 2002;360:1531–1539
  2. Freestone T, Turner RJ, Coady A, et al. Inflammation and matrix metalloproteinases in the enlarging abdominal aortic aneurysm. Arterioscler Thromb Vasc Biol. 1995;15:1145–1151
  3. Holmes DR, Petrinec D, Wester W, et al. Indomethacin prevents elastase-induced abdominal aortic aneurysms in the rat. J Surg Res. 1996;63:305–309
  4. Miralles M, Wester W, Sicard GA, et al. Indomethacin inhibits expansion of experimental aortic aneurysms via inhibition of the COX2 isoform of cyclooxygenase. J Vasc Surg. 1999;29:884–893
  5. Herschman HR. Prostaglandin synthase 2. Biochim Biophys Acta. 1996;1299:125–140
  6. Hong BK, Kwon HM, Lee BK, et al. Coexpression of cyclooxygenase-2 and matrix metalloproteinases in human aortic atherosclerotic lesions. Yonsei Med J. 2000;41:82–88
  7. Holmes DR, Wester W, Thompson RW, et al. Prostaglandin E2 synthesis and cyclooxygenase expression in abdominal aortic aneurysms. J Vasc Surg. 1997;25:810–815
  8. Walton LJ, Franklin IJ, Bayston T, et al. Inhibition of prostaglandin E2 synthesis in abdominal aortic aneurysms. Circulation. 1999;100:48–54
  9. Franklin IJ, Walton LJ, Greenhalgh RM, et al. The influence of indomethacin on the metabolism and cytokine secretion of human aneurysmal aorta. Eur J Vasc Endovasc Surg. 1999;18:35–42
  10. Pentland AP, Shapiro SD, Welgus HG. Agonist-induced expression of tissue inhibitor of metalloproteinases and metalloproteinases by human macrophages is regulated by endogenous prostaglandin E2 synthesis. J Invest Dermatol. 1995;104:52–57
  11. Hinson RM, Williams JA, Shacter E. Elevated interleukin 6 is induced by prostaglandin E2 in a murine model of inflammation: possible role of cyclooxygenase-2. Proc Natl Acad Sci USA. 1996;93:4885–4890
  12. Bayston T, Ramessur S, Reise J, et al. Prostaglandin E2 receptors in abdominal aortic aneurysm and human aortic smooth muscle cells. J Vasc Surg. 2003;38:354–359
  13. Chapple KS, Cartwright EJ, Hawcroft G, et al. Localization of cyclooxygenase-2 in human sporadic colorectal adenomas. Am J Pathol. 2000;156:545–553
  14. Chinery R, Beauchamp RD, Shyr Y, et al. Antioxidants reduce cyclooxygenase-2 expression, prostaglandin production, and proliferation in colorectal cancer cells. Cancer Res. 1998;58:2323–2327
  15. DeWitt DL, Meade EA. Serum and glucocorticoid regulation of gene transcription and expression of the prostaglandin H synthase-1 and prostaglandin H synthase-2 isozymes. Arch Biochem Biophys. 1993;306:94–102
  16. Xu X-M, Sansores-Garcia L, Chen X-M, et al. Suppression of inducible cyclooxygenase 2 gene transcription by aspirin and sodium salicylate. Proc Natl Acad Sci USA. 1999;96:5292–5297
  17. Jones DA, Carlton DP, McIntyre TM, et al. Molecular cloning of human endoperoxide synthase type II and demonstration of expression in response to cytokines. J Biol Chem. 1993;268:9049–9054
  18. Young W, Mahboubi K, Haider A, et al. Cyclooxygenase-2 is required for tumour necrosis factor-alpha- and angiotensin II-mediated proliferation of vascular smooth muscle cells. Circ Res. 2000;86:906–914
  19. Arias-Negrete S, Keller K, Chadee K. Proinflammatory cytokines regulate cyclooxygenase-2 mRNA expression in human macrophages. Biochem Biophys Res Commun. 1995;208:582–589
  20. Ailawadi G, Eliason JL, Upchurch GR. Current concepts in the pathogenesis of abdominal aortic aneurysm. J Vasc Surg. 2003;38:584–588
  21. Doroudi R, Gan LM, Selin Sjogren L, et al. Effects of shear stress on eicosanoid gene expression and metabolite production in vascular endothelium as studied in a novel biomechanical perfusion model. Biochem Biophys Res Commun. 2000;269:257–264
  22. Inoue H, Taba Y, Miwa Y, et al. Transcriptional and posttranscriptional regulation of cyclooxygenase-2 expression by fluid shear stress in vascular endothelial cells. Arterioscler Thromb Vasc Biol. 2002;22:1415–1420
  23. Yan Z, Subbaramaiah K, Camilli T, et al. Benzo[a]pyrene induces the transcription of cyclooxygenase-2 in vascular smooth muscle cells. J Biol Chem. 2000;275:4949–4955
  24. Kazi M, Thyberg J, Religa P, et al. Influence of intraluminal thrombus on structural and cellular composition of abdominal aortic aneurysm wall. J Vasc Surg. 2003;38:1283–1292
  25. Vorp DA, Lee PC, Wang DH, et al. Association between intraluminal thrombus in abdominal aortic aneurysm with local hypoxia and wall weakening. J Vasc Surg. 2001;34:291–299
  26. Ji YS, Xu Q, Schmedtje JF. Hypoxia induces high-mobility-group protein I(Y) and transcription of the cyclooxygenase-2 gene in human vascular endothelium. Circ Res. 1998;83:295–304
  27. Kobayashi M, Matsubara J, Matsushita M, et al. Expression of angiogenesis and angiogenic factors in human aortic vascular disease. J Surg Res. 2002;106:239–245
  28. Holmes DR, Liao S, Parks WC, et al. Medial neovascularization in abdominal aortic aneurysms; a histopathologic marker of aneurysmal degeneration with pathophysiologic implications. J Vasc Surg. 1995;21:761–771
  29. Schneiderman J, Bordin GM, Engelberg I, et al. Expression of fibrinolytic genes in atherosclerotic abdominal aortic aneurysm wall. J Clin Invest. 1995;96:639–645
  30. Thompson MM, Jones L, Nasim A, et al. Angiogenesis in abdominal aortic aneurysms. Eur J Vasc Endovasc Surg. 1996;11:464–469
  31. Satta J, Soini Y, Mosorin M, et al. Angiogenesis is associated with mononuclear inflammatory cells in abdominal aortic aneurysms. Ann Chir Gynaecol. 1998;87:40–42
  32. Bernardini G, Ribatti D, Spinetti G, et al. Analysis of the role of chemokines in angiogenesis. J Immunol Methods. 2003;273:83–101
  33. Tsujii M, Kawano S, Tsuji S, et al. Cyclooxygenase regulates angiogenesis induced by colon cancer cells. Cell. 1998;93:705–716
  34. Hull MA, Thomson JL, Hawkey CJ. Expression of cyclooxygenase 1 and 2 by human gastric endothelial cells. Gut. 1999;45:529–536
  35. Kuwano T, Nakao S, Yamamoto H, et al. Cyclooxygenase 2 is a key enzyme for inflammatory cytokine-induced angiogenesis. FASEB J. 2004;18:300–310
  36. Sawaoka H, Tsuji S, Tsujii M, et al. Cyclooxygenase inhibitors suppress angiogenesis and reduce tumor growth in vivo. Lab Invest. 1999;79:1469–1477
  37. Masferrer JL, Leahy KM, Koki AT, et al. Antiangiogenic and antitumor activities of cyclooxygenase-2 inhibitors. Cancer Res. 2000;60:1306–1311

PII: S0890-5096(06)00016-1

doi:10.1016/j.avsg.2006.10.008

Annals of Vascular Surgery
Volume 21, Issue 1 , Pages 61-66, January 2007

Article Tools

* Image Usage & Resolution