Angiogenesis and remodelling in human thoracic aortic aneurysms

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Angiogenesis and remodelling in human thoracicaortic aneurysmsKetty Kessler1,2*, Luciano F. Borges3,4, Benoıt Ho-Tin-Noe1, Guillaume Jondeau5,Jean-Baptiste Michel1,2*, and Roger Vranckx1

1Univ Paris Diderot, Sorbonne Paris Cite, LVTS, UMR-S1148, F-75018 Paris, France; 2INSERM Unit 1148, Hopital Xavier Bichat, Secteur Claude Bernard, 46 rue Henri Huchard, FR-75877 Pariscedex 18, France; 3Heart Institute (InCor), Hospital das Clınicas da Faculdade de Medicina da Universidade de Sao Paulo, Sao Paulo, Brazil; 4Department of Morphology, Universidade Federal deMinas Gerais, Belo Horizonte, Brazil; and 5Centre National de Reference pour le syndrome de Marfan et apparentes, Hopital Xavier Bichat, Paris, France

Received 14 February 2014; revised 9 August 2014; accepted 12 August 2014

Time for primary review: 32 days

Aims Human thoracic aneurysm of the ascending aorta (TAA) is a chronic disease characterized by dilatation of the aortic wall,which can progress to vessel dissection and rupture. TAA has several aetiologies, but all forms present common features,including tissue remodelling. Here, we determined and characterized the angiogenic process associated with TAA and itsrelation with wall remodelling.

Methodsand results

Immunostaining for blood vessels showed an increased density of microvessels originating from the adventitia in the ex-ternal medial layer of TAA compared with healthy aortas. Proteomic array analysis of 55 angiogenic factors in medial andadventitial layers showed different expression profiles in both tissue compartments between aneurysmal and healthyaortas. Quantification by ELISA confirmed that all forms of TAA contained higher levels of several pro- and anti-angiogenic factors, including angiopoietin-1 and -2, fibroblast growth factor-acidic, and thrombospondin-1, than thatof healthy aortas. However, all groups showed comparable levels of vascular endothelial growth factor-A. QuantitativeRT-PCR demonstrated that angiopoietins were overexpressed in TAA media. Immunostaining and electron microscopyrevealed that neovessels had defective endothelial junctions and poor mural cell coverage. This incomplete structure wasassociated with the accumulation of plasminogen and albumin in the media of TAA.

Conclusion We describe, for the first time, leaky neovessel formation in TAA media in association with an imbalance of angiogenicfactor levels. Although the initiating mechanisms of neo-angiogenesis in TAA and the potential aetiology-related differ-ences remain to be determined, our results suggest that neo-angiogenesis could participate in TAA wall remodelling andweakening through deposition of blood-borne zymogens.

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Keywords TAA † Marfan † Angiopoietins † Angiogenic balance † Thrombospondins

1. IntroductionThoracic aneurysm of the ascending aorta (TAA) is a chronic and multi-factorial disease in which important remodelling of the arterial wall takesplace. Several aetiologies of TAA have been characterized: monogenicforms with mutations in different gene, FBN1, TGFBR1/2, SMAD3,MYH11, and a-SMA, forms associated with a bicuspid aortic valve(BAV) and the degenerative form associated with aging.1 TAA of thesedifferent aetiologies present common features, among which areelastic fibre and extracellular matrix (ECM) degradation occurring byelastolysis and collagenolysis in the medial layer. Moreover, areas ofcystic medial degeneration are present in TAA wall characterized by

smooth muscle cell (SMC) apoptosis, particularly around microvessels,and accumulation of mucoid substance. These areas, enriched in glyco-saminoglycans, show structural disorganization and characterize patho-logical tissue remodelling in TAA. Recent work has shown thataneurysmal SMCs are epigenetically reprogrammed to constitutively in-crease Smad2 expression and phosphorylation with the consequence ofstimulating a protective effect by the overexpression of anti-proteases inthe medial layer.2 –4

Surprisingly, while angiogenesis is a well-described feature leading tolesionprogression in atheromatous arteries andabdominal aortic aneur-ysm, there is no reported evidence that such a process takes place inTAA. However, if one considers some of the TAA characteristics

* Corresponding author. Tel: +33140257526; fax: +33 140258602, Email: [email protected] (K.K.); Tel:+33 140257526; fax: +33 140258602, Email: [email protected] (J.-B.M.)

Published on behalf of the European Society of Cardiology. All rights reserved. & The Author 2014. For permissions please email: [email protected].

Cardiovascular Researchdoi:10.1093/cvr/cvu196

Cardiovascular Research Advance Access published September 4, 2014 at IN

SER

M on Septem

ber 22, 2014http://cardiovascres.oxfordjournals.org/

Dow

nloaded from

http://cardiovascres.oxfordjournals.org/

cited above, the possibility that angiogenesis occurs in TAA becomeshighly plausible. In fact, both the mechanical properties and protein com-position of the ECM have been shown to regulate the angiogenic prop-erties of mesenchymal cells.5,6 Therefore, modifications of these twoparameters through degradation of the ECM could modify the produc-tion of angiogenic factors by SMCs in TAA. This hypothesis is furthersupported by a recent publication, showing that a fibrillin mutationand the consequent elastin fragmentation in hypercholesterolaemicapoE2/2 mice significantly enhanced intraplaque neovascularization.7

Furthermore, we have shown recently that angiogenesis in humanatheromatous disease is associated with substantial changes in thephenotype of medial SMCs. Indeed, SMCs underlying intimal lesions dis-played an increased production of vascular endothelial growth factor-A(VEGF-A) conferring pro-angiogenic properties to them that persistedlong after removal of the initial trigger of this phenotypic change,the atheroma-derived agonists of peroxisome proliferator-activatedreceptor-g.Onecould thushypothesize thatTAA-associatedreprogram-ming of SMCs could, in a similar manner, support an angiogenic response.

For these reasons, the aim of this study was to investigate whetherneo-angiogenesis occurs in TAA. Our results show that TAA displaysan increased density of structurally fragile vessels in the media in associ-ation with an imbalance in angiogenic factor levels. Our data furthersuggest that neovessels in the aneurysmal medial layer could participatein aortic wall remodelling by reinforcing the accumulation of blood-borne proteases in the vessel wall.

2. Methods

2.1 Patients and aortic specimensThe clinical research protocol was approved by the local Ethics Committee(CPP 05 04 32, Ambroise Pare, Boulogne, France, April 2005; updated inMarch 2008) and was in accordance with the principle of the Declarationof Helsinki.8 All patients gave informed consent. Aneurysmal ascendingaortic specimens were collected during aortic surgery (Hopital Bichat).Ninety specimens were divided into three groups according to their clinicalfeatures and genetic background: Marfan (n ¼ 27, mean age: 39+13 years),TAA associated with BAVs (n ¼ 37, mean age: 55+ 15 years), and the de-generative form (n ¼ 26, mean age: 71+ 8 years). The clinical data asso-ciated with this series of patients are reported in Table 1. All specimenswere from aneurysms of .50 mm in diameter. Normal ascending aortaswereobtained fromorgan transplantdonors (n ¼ 50)with the authorizationof the French Biomedicine Agency (PFS09-007) and in accordance with theDeclaration of Helsinki. Aneurysmal tissues were sampled in the outercurvature, the most dilated part of the ascending aorta, and in the sinus ofValsalva for Marfan patients. A small piece of each aortic sample was fixedin paraformaldehyde (PFA) for histology. Aortic tissue preparationconsistedof an immediatedissection of the aorticwallwith the separation of the medialand adventitial layers at the junction of the external elastic lamina (EEL) andthe adventitia.

2.2 Immunohistology of human aortasThe fixed sections were treated for antigen retrieval by heating with citratebuffer, pH 6. For fibroblast growth factor-acidic (FGF-a), plasminogen, anderythrocyte staining, endogenous peroxidase was quenched with 3% (v/v)hydrogen peroxide. Non-specific binding was blocked by incubation with1% (v/v) normal foetal calf serum for 30 min at 378C. Slides were incubatedfor 1 h 30 min at 378C with the primary antibody: anti-vWF (5 mg/mL;DAKO), anti-TSP-1 (5 mg/mL; Abcam), anti-Ang-1 (10 mg/mL; R&DSystems), anti-Ang-2 (20 mg/mL; R&D Systems), anti-FGF-a (5 mg/mL;Santa Cruz), anti-albumin (1/1000 dilution; DAKO), anti-plasminogen(10 mg/mL; Abcam), anti-VE-cadherin (5 mg/mL; Bender Medsystems),

anti-collagen IV (2 mg/mL; Abcam), anti-a-SMA (2 mg/mL; DAKO), oranti-a2b3-integrin (2 mg/mL; home made9). Nuclei were stained withHoechst stain and visualized in blue. Detection of FGF-a and plasminogenwas performed with Histogreen (Histoprime Linaris), and slides were coun-terstained with nuclear red.

For the details of histopathological evaluation, image acquisition, andtransmission electron microscopy, see Supplementary material online.

2.3 Whole-mount stainingAortic specimens were fixed in 4% PFA overnight. Free aldehyde groupswere quenched with glycine at 100 mM pH 7, and the aortic tissue was per-meabilized in Triton 2% in DMSO 20% containing 2% foetal calf serum. Theantibody to MYH11 (10 mg/mL; Millipore) and lectin UEA rhodamine (5 mg/mL; VectorLab) were incubated simultaneously followed by incubation withthe secondary antibody (Alexa Fluor 488). Tissue samples were then dehy-drated in Tetrahydrofuran (Sigma-Aldrich) baths at 25, 50, 75, and 100%,cleared in 1 vol benzyl alcohol with 2 vol benzyl benzoate (Sigma-Aldrich),and examined by fluorescence microscopy in a glass petri dish, withoutmounting.

2.4 Aortic tissue samplesAortic tissue extractions (protein, mRNA) were performed directly fromfrozen aortic media stored at 2808C. The aortic samples were first cryogeni-cally pulverized in liquid nitrogen, using a freezer mill (6870 Spex SamplePrep).Between30and150 mgofcrushedaorticpowderwasused foreachextraction.

Small pieces of aortic media were incubated for 24 h in RPMI culturemedium containing 1% L-glutamine and 1% penicillin, streptomycin, andamphotericin at 378C (5% CO2) to studyextracellularly secreted or releasedproteins (tissue-conditioned medium). The volume of culture medium wasadjusted to the sample wet weight (6 mL/g). Each tissue-conditionedmedium was collected and centrifuged (3000 g, 15 min, at 208C).

For the details of protein extraction, proteomic array, ELISA, immuno-blotting, RT-qPCR, cell culture, and cell migration assay, see Supplementarymaterial online.

2.5 Cell permeability measured by thexCELLigence systemThe xCELLigence system (Roche Applied Science) has three components:an analyzer, a device station, and a 16-well E-plate.The dimensionless param-eter cell index (CI) was used to quantify cell status based on the measuredcell–electrode impedance.10 A CI is a quantitative measure of cell status ina well, including the cell number, cell viability, adhesion degree, and morph-ology. ‘Normalized cell index’ at a specific time point is acquired by dividingthe CI value by the value at a reference time point.

2.6 Statistical analysisValues areexpressed as medians (dot plot graphics)or as means+ SEM. Thesignificance of differences between groups was tested using the Kruskall–Wallis non-parametric test followed by Dunn’s post hoc test, the pairedt-test, the Mann–Whitney test, or a two-way analysis of variance followedby the Bonferroni post hoc test, when appropriate. Correlations were evalu-ated using the Spearman test. A value of P ≤ 0.05 was considered significantand *P , 0.05, **P , 0.01, ***P , 0.001.

3. Results

3.1 Blood vessels are located in the mediaof TAADuring the separation of human TAA media from adventitia, weobserved that blood-red areas were present on the external medialsurfaceof thecurvature, suggesting possible angiogenic vessels penetrat-ing into the media (Figure 1A). Such areas werenot seen in healthy aortas.

K. Kessler et al. of 13 at IN

SER

M on Septem


Dow

nloaded from

http://cardiovascres.oxfordjournals.org/lookup/suppl/doi:10.1093/cvr/cvu196/-/DC1




To address the possible angiogenic nature of these areas, immunostainingfor von Willebrand Factor (vWF), a marker of blood vessels, was per-formed on transverse sections of TAA and healthy aortas. The resultsshowed numerous blood vessels inwardly penetrating into the externalmedia of TAA (Figure 1B–D). The quantification of vWF-positive bloodvessels showed that they were increased in TAA walls compared withthose of undilated healthy aortas (P , 0.001; Figure 1C). The neovesseldensity was significantly different in Marfan and degenerative groups com-pared with healthy aortas (P , 0.001 and ,0.01, respectively), in contrasttoBAV.Microvesseldensitywasevaluatedaccording to their location in themedial layer, which was subdivided into four parts. Thefirst corresponds totheareavery close to theEEL, at theborderbetween themedial andadven-titial layers. The rest of the medial layer was then separated into three equalparts: the external 1/3 closest to the adventitia, the middle section, and theinner1/3,closest tothe intima.Theresults showedthatbloodvesseldensitywas significantly increased in TAAs compared with healthy aortas (P ,

0.0001) and in the external part compared with the internal part of theaortic media (P , 0.0001; see Supplementary material online, Table S2A).In healthy aortas, blood vessels were mainly located close to the adventitiaand consisted of typical vasa vasorum (see Supplementary material online,Table S2B). In TAA, neovessels were increased close to the adventitia butmany extended inwardly across the external 1/3, reaching the middlepart of the medial layer with a significantly different distribution betweenthe external 1/3 and the internal 1/3 of the TAA media, particularly inMarfan (P , 0.05) and degenerative groups (P , 0.01), whereas

BAV-associated TAA was not significantly different from healthy aortas(see Supplementary material online, Table S2C–E). Nevertheless, theseneovessels never reached the intima. Whole-mount immunostaining wasperformed with a lectin specific for endothelial cells, and the TAA tissuewas cleared to permit visualization across the whole aortic wall from theadventitia to the intima.Theresults confirmedthat thebloodvesselsorigin-ate from the adventitia and penetrate into the media (Figure 1E). Thus, thenumber of blood vessels was increased in TAA media with a negative gra-dient from the adventitia to the mid-media. Taken together, these resultsshowed that an angiogenic process occurs in TAA, and that neovesselspenetrating the media likely derive from adventitial vessels.

To evaluate the relationship between neovessel density and theprogression of dilatation, a correlation was established between themaximal diameter of the TAA and the number of vessels per mm2 oftissue. (r2 ¼ 0.23, P , 0.03; see Supplementary material online, Figure S1).

3.2 Differential protein expression ofangiogenic factors in TAA compared withhealthy aortas: identification andlocalizationTo study the equilibrium between angiogenic factors, the protein levelsof 55 factorswere measured from pooled aortic extracts of either mediaor adventitia and from pooled samples of culture medium conditionedby incubation with media or adventitia from TAA and healthy aortas

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 1 Clinical characteristics of patients

Healthy aorta TAA (n 5 90)

Marfan BAV Degenerative

N 50 27 37 26

Age, years 51.0+13.3 38.8+12.5 54.7+14.9 71.3+7.8

Sex, % male 64.0 77.8 83.8 88.5

Aortic diameter, mm – 54.1+9.5 51.2+4.4 56.9+8.0

Risk factors

Hypertension, n – 5 12 13

Diabetes, n – 2 7 2

Dyslipidaemia, n – 2 6 10

Smokers, n – 5 13 8

Not determined, n 9 6 4

Medical treatment

Statins, n – 1 7 8

b-Blockers, n – 16 13 10

Aspirin, n – 2 5 9

ARA2, n – 3 5 1

ACEI, n – 1 4 6

Unknown, n 8 9 7

Mutations

FBN1, n 20

TGFBR1, n 2

TGFBR2, n 1

Unknown, n 4

The cohort of patients studied presented two groups: patients with healthy aortas and those with aneurysmal thoracic ascending aortas of three potential aetiologies: Marfan syndrome, BAV,and the degenerative form. Every aneurysm had an aortic diameter above 50 mm. Results are expressed as means+ SD.BAV, bicuspid aortic valve; TAA, thoracic aneurysm of the ascending aorta; ARA2, angiotensin II receptor antagonist; ACEI, angiotensin-converting enzyme inhibitor; FBN1, fibronectin 1;TGFBR1/2, transforming growth factor receptor 1/2.

Angiogenesis in TAA of 13 at IN

SER

M on Septem


Dow

nloaded from









(the screening of angiogenic factors is shown in Supplementary materialonline, Figure S2). Four types of samples were thus defined: the medialextract (1), the adventitial extract (2), the medial conditioned medium(3), and the adventitial conditioned medium (4). Among these 55factors, we found that 34 were dysregulated in TAA compared withhealthy aortas. Expression of some of these proteins was found inonly one compartment and was not detectable in the other compart-ments, e.g. angiopoietins 1 and 2 (Ang-1 and -2), latency-associatedprotein (LAP), persephin, and thrombospondin-2 (TSP-2) were foundonly in the TAA medial extract, and dipeptidyl peptidase IV (DPPIV)was present only in the TAA adventitial extract. Also, urokinase-typeplasminogen activator (u-PA) was found only in the TAA adventitialconditioned medium.

A differential relative expression factor for each compartment wasdetermined, defined as the ratio between TAA values and values ofhealthy aorta for each factor in the extract and the conditionedmedium. Factors can be classified into four groups as illustrated inFigure 2A: Group I contains factors that were greatly increased in TAA

compared with healthy aortas (beyond 10 500) (in red), Group II com-prises factors that were more moderately increased (between 2 and10 500) (in orange and yellow), and Group III comprises factors thatwere decreased in TAA compared with healthy aortas (,0.9; ingreen). This group includes factors with a role in angiogenesis (e.g. endo-glin) but also factors with an anti-protease role, such as the tissue inhibi-tors of metalloproteinases, TIMP-1 and TIMP-4. A fourth group includedfactors with a ratio not significantly different between TAA and healthyaortas (ratio between 0.9 and 1.9; in grey; Figure 2A).

Results can be analysed by comparing the compartmental ratio(value in TAA/value in healthy aorta) two by two for Groups I (mostlyred) and II (mostly orange and yellow; Figure 2A). Comparing medialand adventitial extracts reveals that plasminogen, Ang-1, Ang-2, TSP-1,platelet-derived endothelial cell growth factor (PD-ECGF), insulin-likegrowth factor-binding protein-1 (IGFBP-1), endocrine gland-derivedvascular growth factor (EG-VEGF), LAP, TSP-2, persephin, matrixmetalloproteinase-9 (MMP-9), and pentraxin-3 were all more abundantin TAA media.

Figure 1 Microvessel density and distribution in TAA media. (A) The aortic wall from a patient with thoracic aortic aneurysm before and after the sep-aration between medial and adventitial layer (n . 25). (B) Immunolocalization of blood vessels in healthy (n ¼ 14) and aneurysmal aortas (n ¼ 25).vWF-stained vessels (in red) were located in the adventitia in healthy aortas and in the media in TAA. Media/adventitia separation is represented by adotted line. Elastic fibres appear in green by autofluorescence. Cell nuclei are stained in blue. Scale bar ¼ 200 mm (C). Median number of vWF-positivevessels per mm2 of media within whole longitudinal aortic sections in healthy (n ¼ 14), Marfan (n ¼ 9), BAV (n ¼ 8), and degenerative aortas (n ¼ 8).A non-parametric Kruskall–Wallis test was performed (P ¼ 0.0005) followed by Dunn’s post hoc test (*P , 0.05, **P , 0.01, ***P , 0.001). (D) Repre-sentative images of healthy (n ¼ 14) and TAA aortas (n ¼ 25) showing tissue morphology by Masson’s trichrome staining with the following: cytoplasm inpink, collagen in green, nuclei in black, and erythrocytes in red (a and b). Microvessels (arrows) were immunostained for vWF (in red) and nuclei withHoechst stain (in blue) (c and d ). Scale bars ¼ 500 mm. Each layer is separated by a dotted line. (E) An aneurysmal aortic sample before and after whole-mount immunostaining and clearing. Representative images of 50 mm thick maximal intensity projections of serial optical sections in 2 mm steps, showingthe invasion of adventitial microvessels (lectin in red) into the media (characterized by smooth muscle cell staining by MYH11, in green), layers separated bya dotted line. Details of the inset are shown below at higher magnification. This experiment was repeated on three different patients. Scale bars ¼ 50 mm.


SER

M on Septem


Dow

nloaded from




Figure 2 Disequilibrium in expression profiles of angiogenic factors between healthy and TAA aortas. (A) Differential expression profile of angiogenicproteins identified by proteomic array. A pool of aortic samples from six healthy patients was compared with a pool of samples from nine TAA patients.Results are expressedby the ratio of the pixel density measured in TAA to that measured in healthy aortas. The pixel density wasmeasured in the media andthe adventitia in tissue extracts and in conditioned medium (C.M.). The scale bar represents a colour code corresponding to the ratio values. (B–F) Proteincontent per gram tissue measured in medial and adventitial layers from healthy, Marfan, BAV, and degenerative aortas (n ¼ 18, 11, 10, and 13 for media andn ¼ 9, 9, 10, and 9 for adventitia, respectively). Non-parametric Kruskall–Wallis tests were performed (P , 0.01 for all angiogenic factors, except forVEGF-A: NS) followed by Dunn’s post hoc tests (*P , 0.05, **P , 0.01, ***P , 0.001).


SER

M on Septem


Dow

nloaded from


Comparing the medial extract with the medial conditioned mediumrevealswhich factors are over-released by the TAA media. The followingdiffusible factors were increased in TAA: plasminogen, interleukin-8(IL-8), FGF-a, and monocyte chemoattractant protein-1 (MCP-1).

In the adventitial layer, IL-8 and DPPIV were found in the extract butnot in the corresponding conditioned medium in contrast to plasmino-gen, u-PA, FGF-a, MCP-1, and placental growth factor (PlGF), whichwere found in both adventitial tissue extracts and conditionedmedium, indicating that they are diffusible factors (Figure 2A).

The analysis of medial and adventitial conditioned medium reveals anincreased diffusion of plasminogen, FGF-a, and MCP-1 from samples ofboth TAA layers. Interestingly, MMP-8 secretion was decreased in bothTAA layers compared with those of healthy aortas (Figure 2A).

To confirm the semi-quantitative differential expression from a poolof patients, quantitative studies using ELISA were performed on individ-ual proteins selected from the data in Figure 2A. Results showedthat protein levels of VEGF-A and PlGF were similar in TAA andhealthy aortas, both in the media and in the adventitia (except forPlGF which increased in degenerative tissue; Figure 2B; see Supple-mentary material online, Figure S3 shows PlGF content in aortas). More-over, the transcriptional expression study in medial extracts showed asimilar level of VEGF-A, PlGF, and platelet-derived growth factors(PDGF-A and PDGF-B) in TAA and healthy aortas, thus strengtheningthe proteomic array results (Figure 3A; quantification of PlGF andPDGF-A and -B mRNA in media is shown in Supplementary materialonline, Figure S4).

Among the angiogenic factors, which were differentially expressed inTAA, we chose to study further Ang-1, Ang-2, TSP-1, and FGF-a as can-didates for the angiogenic switch. Their protein levels werequantified byELISA in individual aortic extracts and in conditioned medium. Resultsshowed an increase for all four factors in TAA compared with healthyaortas in the both medial and adventitial tissue extracts (Figure 2C–F ).However, Ang-1, FGF-a, and TSP-1 levels were mostly similar in the con-ditioned medium of TAA and healthy aortas (see Supplementary mater-ial online, Table S3).

At the transcriptional level, further analysis demonstrated that TSP-1and FGF-a mRNA expression was similar in aneurysmal and healthymedial samples (Figure 3D and E). But interestingly, the expression ofangiopoietins was increased in TAA medial tissue compared withhealthy media, particularly angiopoietin-1 (Figure 3B and C ). However,their transcriptional expression profile was similar in the adventitia inall four groups (as well as FGF-a and TSP-1; data not shown). Addition-ally, the expression of angiopoietins was also studied in SMCs culturedin vitro, but their overexpression observed in TAA tissue was nolonger observed (data not shown).

Histological studies were performed on aortic serial sections to de-termine tissue structure, vessel localization, and angiogenic factor pres-ence and distribution in the media. Staining with Masson’s trichromeshowed the presence of tissue remodelling in TAA media with elastic la-mellar disorganization (Figure 4). On the contrary, healthy media hadparallel and continuous elastic lamellae. Immunostaining for vWF, specif-ic for endothelial cells, demonstrated that neovessels were mostlylocated in the mucoid areas in TAA. The angiogenic factors TSP-1,Ang-1, Ang-2, and FGF-a were increased in TAA, particularly in theECM. It is interesting to note that Ang-1 and Ang-2 accumulated in theneovessel endothelial cells.

In summary, Ang-1, Ang-2, FGF-a, and TSP-1 were increased in TAAmedia, particularly in the ECM, and more specifically in the areas ofmucoid degeneration and neovessel endothelial cells.

3.3 Hypoxia and contractile SMC phenotypein TAAHypoxia was studied in TAA as a potential stimulus of angiogenesis inTAA. The hypoxia markers, hypoxia-inducible factor-1 alpha (HIF-1a)and sirtuine-1 (Sirt-1), were analysed by western blotting from tissueextracts. No difference was revealed between healthy and aneurysmalaortas (data not shown).

Since SMC dedifferentiation into myofibroblasts is able to promoteangiogenesis,11,12 SMC phenotype was investigated in TAA. We thusperformed immunostaining and western blotting on aortic tissues forMYH11 and its isoforms, SM1 and SM2. MYH11 levels were identicalin TAA and healthy aortas whatever the isoform studied (data notshown). These results indicate that SMCs in TAA maintain their con-tractile phenotype and do not show signs of marked dedifferentiationinto myofibroblasts.

In summary, the modified expression of angiogenic factors in TAAappears to occur without the involvement of hypoxia or a change ofSMC phenotype.

3.4 Neovessels have abnormal morphologyin TAAThe paradoxical increase in both pro- and anti-angiogenic factors in TAAmedia might induce abnormal angiogenesis. We thus investigated thepossibility of an incomplete structure of neovessels by performingco-immunostaining for vWF and different vessel structural components.Neovessels found in TAA media were compared with adventitial vessels(in both TAA and healthy aortas) as controls (Figure 5). The mature,structurally complete vessels in the adventitia were composed of endo-thelial cells lying on a basal lamina composed of collagen IV andconnected to each other via adherent junctions, which stained forVE-cadherin. In addition, a-actin-positive cells (SMCs/pericytes) sur-rounded the vessels forming one or several layer(s) (Figure 5A, B, andH ). In contrast, co-immunostaining in TAA media showed that neoves-sels can lack mural cells (Figure 5C, D, and H ) and adherent junctions(Figure 5D, E, and H ), and may also have a discontinuous or absentbasal lamina (Figure 5C, D, G, and H ). Some erythrocytes visualized byphase contrast microscopy were located outside the medial neovessels(Figure 5C, D, and I ).

An ultrastructural study of neovessels demonstrated pericytes withmany cytoplasmic vesicles surrounding endothelial cells containingvacuoles (Figure 5F). The basal lamina displayed discontinuities and thelumen was encroached upon by endothelial cell bulging (Figure 5F and G).

To assess vessel permeability to circulating cells, immunostaining forplatelets with a specific antibody against a2b3-integrin was performed.We observed platelet aggregates exclusively in the adventitia and insideblood vessels in all four groups (data not shown). However, immunos-taining for plasma proteins, i.e. albumin and also the protease precursor,plasminogen, was observed in the TAA media but not in that of healthyaorta (Figure 6). These data indicate that neovessels are permeable toplasma proteins, thus leading to their accumulation in the TAA media.We further investigated protease accumulation in media extracts andquantified protein levels of MMP-3, -7, -9, and -19, known to be involvedin tissue remodelling (Figure 6E–H ). MMP-9 and -19 proteins wereincreased in TAA media extracts compared with healthy aortas, al-though their expression by medial cells was similar (Figure 6G and H,and see Supplementary material online, Figure S5C and D). Medial cellsexpressed MMP-2, -3 (at very low level), -9, and -19 mRNA similarlyin healthy and aneurysmal medial cells as shown in Supplementary


SER

M on Septem


Dow

nloaded from










material online, Figure S5. However, MMP-7 and -12 (specific of macro-phage expression) were not expressed by medial cells (data not shown).

The effect of secretedand diffusible angiogenic factorson an endothe-lial cell monolayer was evaluated. Impedance measurements reflectingpermeability were performed in vitro in the presence of aortic layer-conditioned medium. As a positive control, endothelial cells wereincubated with thrombin, a serine protease, which is known to alterendothelial barrier function13 (Figure 7A and B). Results showed thatan endothelial cell monolayer incubated with conditioned medium(from either TAA or healthy aortas) had decreased impedance, demon-strating that endothelial cell permeability was increased. We exploredfurther the capacity of migration and proliferation of endothelial cellsin vitro in the presence of conditioned medium of healthy and aneurysmalaortas (Figure 7C and D). Results showed that endothelial cell incubationwith healthy aorta-conditioned medium inhibited their migration/proliferation compared with the unstimulated condition and also incu-bation with TAA-conditioned medium. Taken together, these resultssuggest that endothelial cell incubation with media-conditioned med-ium increases the cell permeability in all four groups, but only the TAA-conditioned medium favours endothelial cell migration/proliferation.

4. DiscussionIn this study, we have investigated and characterized angiogenesis inTAA. We demonstrated the presence of neovessels in the externaltwo-thirds of the TAA media, likely originating from pre-existing adven-titial vessels. They became scarcer approaching the intima and neoves-sels were totally absent in the inner third of the media. This aspect

differs from observations in atheroma where neovessels can reach theintimal plaques and induce intraplaque haemorrhages.14

One important point in this study is the fact that TAA angiogenesisoccurs in the absence of modifications in VEGF-A levels. This contrastswith descriptions in atheromatous aortas where the angiogenic re-sponse was shown to be associated with a lipid-dependent stimulationof VEGF-A expression by SMCs.15 Although VEGF-A is the mostcommon mediator of angiogenesis, new VEGF-independent mechan-isms have been identified associated with vascular diseases.16 Similarly,VEGF-independent mechanisms have also been reported in somecancers.17

Our study was non-exhaustive but clearly demonstrated a disequilib-rium of pro- and anti-angiogenic factors in TAAs compared with healthyaortas. In particular, both Ang-1 and Ang-2 were overexpressed in an-eurysmal medial tissue. It has been demonstrated that mice overexpres-sing Ang-1 in the skin produced larger, more numerous, and more highlybranched blood vessels.18 Ang-2 is involved in vascular remodelling,19

and our results showed that Ang-2 mainly co-localized with the endo-thelial cells. Moreover, Ang-2, associated with VEGF-A, can induce anangiogenic response and an increase in vessel permeability.20 Takentogether, these data suggest that, together, Ang-1 and Ang-2 couldpromote angiogenesis and participate in tissue remodelling in TAA.

In contrast toSmad2,which is epigenetically up-regulated inTAA,3 theoverexpression of angiopoietins was lost when SMCs were culturedin vitro, implicating a microenvironment-dependent transcriptionalregulation.

It is noteworthy that some angiogenic factors were more greatlyincreased in the degenerative form of TAA compared with BAV and

Figure 3 Ang-1 and -2 are overexpressed by cells in TAA media. Quantification of mRNA expression from medial layer cells normalized to b-actin ex-pression in healthy aortas, Marfan, BAV, and degenerative TAA (n ¼ 9, 7, 10, and 4, respectively). Non-parametric Kruskall–Wallis tests were performed(P , 0.01 for Ang-1 and -2, and NS for VEGF-A, TSP-1, and FGF-a) followed by Dunn’s post hoc tests (*P , 0.05, **P , 0.01, ***P , 0.001).


SER

M on Septem


Dow

nloaded from



Marfan forms, particularly for PlGF, Ang-1, Ang-2, and FGF-a. Thus,despite the fact that different forms of TAA present many commonfeatures, each aetiology can have its own particularity.

TSP-1, also increased in TAA media, is an adhesive glycoproteinwhich mediates cell-to-cell and cell-to-matrix interactions.21 It actsas a powerful endogenous inhibitor of angiogenesis both indirectly,

Figure 4 Angiogenic factors are increased in TAA media. Serial sections were used to localize different angiogenic factors in the same area of the wall inhealthy Marfan, BAV, and degenerative aortas (n ≥ 5 in each group). Masson’s Trichrome showsgeneral tissue morphology with cytoplasm in pink, collagenin green, nuclei in black, and erythrocytes in red. Black stars indicate the mucoid areas. In the same area, microvessels are immunostained for vWF in red.TSP-1, Ang-1, and Ang-2 appear also in red in serial sections. Elastic lamellae are green due to their autofluorescence. FGF-a immunostaining appears ingreen with pink cytoplasm, red nuclei, and dark pink erythrocytes (nuclear red counterstaining). Scale bar ¼ 50 mm.


SER

M on Septem


Dow

nloaded from


Figure 5 Microvessels inTAAmediahavean incomplete structure.Thestructureofvessels in theadventitial layerof healthyaorta (n ¼ 5)andTAA(n ¼ 6) (Aand B) was compared with that in the aneurysmal medial layer (n ¼ 6) (C–E). Vessel structure was characterized in serial sections by co-immunostaining withvWF (in red) and specificproteins involved in vessel formation (in green): VE-cadherin (adherent junctions between endothelial cells), collagen IV (basal lamina),and a-SMA (a-smooth muscle actin, which stains pericytes and smooth muscle cells). Tissue morphology was determined by phase contrast microscopyshowingelastic fibres anderythrocytes. Arrows showerythrocytes outside vessels. Cell nuclei appear in blue.Details are shown in insets at higher magnification.Scale bar¼ 20 mm. (F and G). Electron micrographs illustrating the ultrastructure of microvessel in TAA media. Pericytes (Pc) surrounding the endothelial cells(EC) contain cytoplasmic vacuoles (white arrows). The junctions between Pc and ECs are indicated with black arrows. The abnormal vessels have a fragmentedbasal lamina (white circle) andECs bulge (white arrowhead) into the lumen (L). (H ) Quantification of medial vessels presenting co-immunostaining for vWF andneovessel ultrastructural components: a-SMA (51%), collagen IV (90%), or VE-cadherin (76%) in TAA. Medial microvessels were compared with adventitialvessels as control (100%, n ¼ 6 different patients). A non-parametric Kruskall–Wallis test was performed (P , 0.0001) followed by Dunn’s post hoc test(*P , 0.05, **P , 0.01, ***P , 0.001). (I) Erythrocytes, detected using diaminobenzidine (DAB), were present outside the microvessels (inset).


SER

M on Septem


Dow

nloaded from


by sequestering angiogenic growth factors and effectors in the ECM,and directly, by inducing an anti-angiogenic programme in endothelialcells.11 Its increase and sequestration in TAA media could promote

neovessel permeability and could favour the retention of angiogenicfactors in the ECM. In this case, its role within the TAA mediawould be protective by trapping angiogenic factors and temporarily

Figure6 Microvessel leakage in TAA media. (A) Representativepatternsof albumin immunofluorescence staining (in purple) in healthy, Marfan, BAV, anddegenerative aortas. Five sections, each from a different patient, were analysed for each group. The green signal corresponds to elastic fibre autofluores-cence. Intima, media, and adventitia layers are separated by a dotted line. Scale bar ¼ 500 mm. (B) Representative plasminogen staining on healthy (n ¼ 5),Marfan (n ¼ 3), BAV (n ¼ 3), and degenerative (n ¼ 3) aortas. Insets represent higher magnification. Scale bar ¼ 40 mm. (C) Examples of average fluores-cence intensity staining profiles for albumin measuredalong line scans (n ¼ 15 random line scans drawn perpendicular to the adventitial border towards theluminal border for each aetiology) in healthy, Marfan, BAV, and degenerative aortas. (D) Mean percentage of plasminogen-positive staining (expressed as %of pixels stained normalized by the total number of pixels per field) at five different fields taken in healthy (n ¼ 5), Marfan (n ¼ 3), BAV (n ¼ 3), and de-generative (n ¼ 3) aortas. A non-parametric Kruskall–Wallis test was performed (P , 0.0001) followed by Dunn’s post hoc test; results are expressedas mean+ SEM. (E and F) MMPs expression in media of aneurysmal and healthy aortas (n ≥ 4 in both groups). MMP-9 and -19 (*P , 0.05, **P , 0.01,***P , 0.001) are increased in TAA media compared with control according to a Mann–Whitney test; results are expressed as mean+ SEM.


SER

M on Septem


Dow

nloaded from


avoiding their release. Thus, despite the fact that the quantity ofangiogenic factors was increased in TAA, their mRNA expression bymedial cells was similar whatever the origin of the TAA, suggestingthat angiogenic factors were retained within the tissue (e.g. FGF-a).This retention in the ECM, which also includes Ang-1and Ang-2,could be controlled by TSP-1.

Angiogenesis requires destruction of the basement membrane andlocal degradation of the ECM to allow migration and proliferation ofendothelial cells and the release of ECM-bound growth factors.Indeed, it has been recently demonstrated that the presence of a Marfan-like fibrillin mutation in atheroma-prone apoE2/2 mice enhancedintraplaque neovascularization, elastic fibre fragmentation, and finallyatherothrombotic complications in this experimental model.7 Inaccord with these recent experimental observations, we observed anincreased angiogenic tendency in human Marfan TAA compared withthe other aetiologies. Nevertheless, comparing the different aetiologiesrequires further specific studies. Also, we show that the permeability ofneovessels is associated with a lack of mural cells, low VE-cadherin levelssuggesting decreased endothelial cell interactions, reduced collagen IV

levels indicating anomalies of the basal lamina, and with many endothelialcytoplasmic vacuoles. Notably, endothelial permeability could beincreased by the conditioned medium from healthy and also from aneur-ysmal aortas (Figure 7A and B). However, only TAA-conditioned mediumwas in favour of migration and proliferation of endothelial cells (Figure 7Cand D). These results suggest that while in healthy media the microenvir-onment inhibits the formation and growth of neovessels, this inhibition islifted in TAA media, thus leading to the formation of incomplete andleaky neovessels. The hyperpermeability could participate in the in-crease in plasma proteins within the TAA wall. It is known that severalproteases are increased in TAA: MMPs, but also proteases of theplasminergic system: t-PA and u-PA, which are produced by theSMCs.4,12,22 Indeed, our results showed an increase in proteases:plasminogen, MMP-9, and MMP-19 in TAA media as well as u-PA inTAA adventitia (Figure 6G and H and see Supplementary materialonline, Figure S2). These proteases might increase neovessel permeabil-ity by weakening the neovessel ultrastructural components such aspericytes and adherent junctions, especially since tissue metallopro-teinase inhibitors, TIMP-1 and TIMP-4, were decreased in TAA. Thus,

Figure 7 Exposure to conditioned medium from healthy and TAA aortas increases endothelial cell permeability. (A) Endothelial cells from human aortaswere incubated until confluence. Conditioned medium from medial and adventitial layers of healthy and TAA aortas (n ¼ 4 for each group) was added toconfluent endothelial cells. As apositive control, thrombinwasadded to cells at 0.5 nM. Impedancewasmeasuredevery 5 min ineachcondition. Results areexpressed as normalized cell index+ SEM for one experiment in duplicate and are representative of three separate experiments. (B) Mean percentage ofnormalized cell index (normalized by that measured in the presence of thrombin) of endothelial cells incubated with thrombin (0.5 nM), endothelial cellmedium(control), healthyaorta, orTAA-conditioned medium (CM)at10 min.Results areexpressedasmean+ SEM (paired t-testwasperformed). (C andD) Representative images and percentage of endothelial cells stained with Hoechst after 24-h incubation. Cells were incubated with culture medium RPMI(n ¼ 3), with healthy aorta-CM (n ¼ 5), or with TAA-CM (n ¼ 6). Results are expressed as the mean difference between cell-covered areas at t ¼ 0 h(represented by a red dotted circle) and t ¼ 24 h (represented by a white circle)+ SEM. Scale bar ¼ 1 mm (*P , 0.05, **P , 0.01, ***P , 0.001).


SER

M on Septem


Dow

nloaded from




the hyperpermeability of neovessels in TAA could be the resultof several events: an imbalance of pro-/anti-angiogenic factors and ofproteases/anti-proteases, and the presence of incomplete neovessels.

The imbalance between deleterious proteases and protective anti-proteases is clearly in favour of a degradation of the elastic lamellaeand the formation of mucoid areas. Our previous work showed alsothat overexpression of the serine protease inhibitors, proteasenexin-1 and plasminogen activator inhibitor-1, in TAA SMCs was pro-tective, in favour of a progressive aneurysmal dilation, avoiding acutewall dissection.4

We have observed that neovessels were mostly located in the mucoiddegeneration areas, suggesting that ECM degradation facilitates neoves-sel invasion. We postulate that the imbalance of expression betweenproteases and TIMPs induces ECM degradation leading to the releaseand activation of factors promoting angiogenesis such as MMPs. More-over, angiogenic factors mainly located in mucoid areas can increasethe migration and proliferation of endothelial cells from the adventitiaand promote angiogenesis, forming areas of weakness and importanttissue remodelling.

The correlation study showed that the extent of angiogenesis appearsto increase with the advancement of TAA dilatation. Thus, we hypothe-size that angiogenesis in TAA may have a permissive and deleteriouseffect and could participate in the disease evolution.

Our data demonstrated that angiogenic activity was increased adja-cent to the TAA adventitial layer. This is in accord with the results ofGomez and co-workers showing that the inactive form of TGF-b1(LAP) was increased in TAA, in particular close to the adventitia, andco-localized with the degraded elastic lamellae.23 Furthermore, weshowed that the angiogenesis-related factors, DDPIV and IL-8, areexpressed exclusively in the TAA adventitial compartment. More inves-tigations are required to confirm our semi-quantitative data and theirbiological implications in TAA angiogenesis and remodelling, but to-gether our results suggest that the adventitial layer also plays an import-ant role in the progression of dilatation.

Finally, we have shown that the angiogenesis in TAA does not appearto be related to environmental hypoxia. On the contrary, it is probablethat the depletion of SMCs and elastic fibres makes the aneurysmalmediamorepermissive to the diffusionofO2 and free radicals.Aputativeregulation pathway such as NF-kB may be implicated: the NF-kB tran-scription factors could alter gene expression involved in the angiogenicprocess in TAA as observed in solid tumours.16 To support this postu-late, the quantification of angiogenesis-related proteins revealedincreased IL-8, MMP-9, and IGFBP, which can all be regulated by theNF-kBsignallingpathway.24 However, the precisemolecular mechanismleading to the overexpression of angiopoietins in TAA media remainscurrently unclear.

In conclusion, our study has demonstrated for the first time that anangiogenic process takes place in the TAA media. Until now, themedia were considered to be the principal wall layer involved in TAApathogenicity. But we have demonstrated here that the adventitiallayer may also play an important role in disease evolution by promotingthe sprouting of neovessels towards the TAA media, highlighting the im-portance of interactions between each of the vascular compartmentsand the need to consider the whole tissue in context. Angiogenesisappears tobe inducedbyan imbalancebetweenpro- andanti-angiogenicfactors, and the neovessels were shown to be leaky. Their permeabilityled to leakage of plasma proteins and promoted plasma zymogen accu-mulation in the aneurysmal wall. The tissue microenvironment is favour-able to ECM degradation, formation of mucoid areas, and endothelial

cell proliferation/migration. Neovessel invasion into the media, togetherwith the presence of mucoid areas, plays an important role in tissue re-modelling and probably participates in weakening the aortic wall byimpairing its structural integrity and potentially promoting dissection.Finally, more investigations are necessary to identify the precise molecu-lar mechanisms promoting this angiogenesis in TAA and the potentialdifferences between aetiologies.

Supplementary materialSupplementary material is available at Cardiovascular Research online.

AcknowledgementsThe authors thank surgeons of the Departments of Cardiac and VascularSurgery of Bichat Hospital for providing normal and pathological aorticsamples, and Mary Osborne-Pellegrin for help in editing. We thankAngelique Levoye, Morgane Lannoy, and Laurence Walch for helpful sci-entific discussion. We also thank Liliane Louedec, Catherine Deschildre,and Nadir Cheurfa for their technical support.

Conflict of interest: none declared.

FundingThis work was supported by E.U FP-7 project Fighting Aneurysmal Project(www.fighting-aneurysm.org/) and by the French National ResearchAgency (GDMP2). K.K. was recipient of fellowship from the French Ministryfor Higher Education and Research and from the Medical ResearchFoundation (grant no. FDT20130928022).

References1. Jondeau G, Michel JB, Boileau C. The translational science of Marfan syndrome. Heart

2011;97:1206–1214.2. Gomez D, Coyet A, Ollivier V, Jeunemaitre X, Jondeau G, Michel JB, Vranckx R. Epigen-

etic control of vascular smooth muscle cells in Marfan and non-Marfan thoracic aorticaneurysms. Cardiovasc Res 2011;89:446–456.

3. Gomez D, Kessler K, Michel JB, Vranckx R. Modifications of chromatin dynamics controlsmad2 pathway activation in aneurysmal smooth muscle cells. Circ Res 2013;113:881–890.

4. Gomez D, Kessler K, Borges LF, Richard B, Touat Z, Ollivier V, Mansilla S, Bouton MC,Alkoder S, Nataf P, Jandrot-Perrus M, Jondeau G, Vranckx R, Michel JB. Smad2-dependent protease nexin-1 overexpression differentiates chronic aneurysms fromacute dissections of human ascending aorta. Arterioscler Thromb Vasc Biol 2013;33:2222–2232.

5. Kniazeva E, Kachgal S, Putnam AJ. Effects of extracellular matrix density and mesenchy-mal stem cells on neovascularization in vivo. Tissue Eng Part A 2011;17:905–914.

6. Sullivan KE, Quinn KP, Tang KM, Georgakoudi I, Black LD III. Extracellular matrix remod-eling following myocardial infarction influences the therapeutic potential of mesenchy-mal stem cells. Stem Cell Res Ther 2014;5:14.

7. Van der Donckt C, Van Herck JL, Schrijvers DM, Vanhoutte G, Verhoye M, Blockx I, VanDer Linden A, Bauters D, Lijnen HR, Sluimer JC, Roth L, Van Hove CE, Fransen P,Knaapen MW, Hervent AS, De Keulenaer GW, Bult H, Martinet W, Herman AG, DeMeyer GR. Elastin fragmentation in atherosclerotic mice leads to intraplaque neovascu-larization, plaque rupture, myocardial infarction, stroke, and sudden death. Eur Heart J2014; doi:10.1093/eurheartj/ehu041.

8. World Medical Association Declaration of Helsinki. Recommendations guiding physi-cians in biomedical research involving human subjects. Cardiovasc Res 1997;35:2–3.

9. Pidard D, Montgomery RR, Bennett JS, Kunicki TJ. Interaction of AP-2, a monoclonal anti-body specific for the human platelet glycoprotein IIb-IIIa complex, with intact platelets.J Biol Chem 1983;258:12582–12586.

10. Sun M, Fu H, Cheng H, Cao Q, Zhao Y, Mou X, Zhang X, Liu X, Ke Y. A dynamic real-timemethod for monitoring epithelial barrier function in vitro. Anal Biochem 2012;425:96–103.

11. Sweetwyne MT, Murphy-Ullrich JE. Thrombospondin1 in tissue repair and fibrosis:TGF-beta-dependent and independent mechanisms. Matrix Biol 2012;31:178–186.

12. Narayanan N, Tyagi N, Shah A, Pagni S, Tyagi SC. Hyperhomocysteinemia during aorticaneurysm, a plausible role of epigenetics. Int J Physiol Pathophysiol Pharmacol 2013;5:32–42.

13. Ehringer WD, Edwards MJ, Miller FN. Mechanisms of alpha-thrombin, histamine, andbradykinin induced endothelial permeability. J Cell Physiol 1996;167:562–569.


SER

M on Septem


Dow

nloaded from


http://www.fighting-aneurysm.org/

http://www.fighting-aneurysm.org/


14. Michel JB, Virmani R, Arbustini E, Pasterkamp G. Intraplaque haemorrhages asthe trigger of plaque vulnerability. Eur Heart J 2011;32:1977–1985. 1985a, 1985b,1985c.

15. Ho-Tin-Noe B, Michel JB. Initiation of angiogenesis in atherosclerosis: smooth musclecells as mediators of the angiogenic response to atheroma formation. Trends CardiovascMed 2011;21:183–187.

16. Kim YW, Byzova TV. Oxidative stress in angiogenesis and vascular disease. Blood 2014;123:625–631.

17. Zeng W, Gouw AS, van den Heuvel MC, Zwiers PJ, Zondervan PE, Poppema S, Zhang N,Platteel I, de Jong KP, Molema G. The angiogenic makeup of human hepatocellularcarcinoma does not favor vascular endothelial growth factor/angiopoietin-drivensprouting neovascularization. Hepatology 2008;48:1517–1527.

18. Suri C, McClain J, Thurston G, McDonald DM, Zhou H, Oldmixon EH, Sato TN,Yancopoulos GD. Increased vascularization in mice overexpressing angiopoietin-1.Science 1998;282:468–471.

19. Gale NW, Thurston G, Hackett SF, Renard R, Wang Q, McClain J, Martin C, Witte C,Witte MH, Jackson D, Suri C, Campochiaro PA, Wiegand SJ, Yancopoulos GD.

Angiopoietin-2 is required for postnatal angiogenesis and lymphatic patterning, andonly the latter role is rescued by Angiopoietin-1. Dev Cell 2002;3:411–423.

20. Benest AV, Kruse K, Savant S, Thomas M, Laib AM, Loos EK, Fiedler U, Augustin HG.Angiopoietin-2 is critical for cytokine-induced vascular leakage. PLoS ONE 2013;8:e70459.

21. Tolsma SS, Volpert OV, Good DJ, Frazier WA, Polverini PJ, Bouck N. Peptidesderived from two separate domains of the matrix protein thrombospondin-1 haveanti-angiogenic activity. J Cell Biol 1993;122:497–511.

22. Borges LF, Gomez D, Quintana M, Touat Z, Jondeau G, Leclercq A, Meilhac O,Jandrot-Perrus M, Gutierrez PS, Freymuller E, Vranckx R, Michel JB. Fibrinolytic activityis associated with presence of cystic medial degeneration in aneurysms of the ascendingaorta. Histopathology 2010;57:917–932.

23. Gomez D, Al Haj Zen A, Borges LF, Philippe M, Gutierrez PS, Jondeau G, Michel JB,Vranckx R. Syndromic and non-syndromic aneurysms of the human ascending aortashare activation of the Smad2 pathway. J Pathol 2009;218:131–142.

24. Perkins ND, Gilmore TD. Good cop, bad cop: the different faces of NF-kappaB.Cell Death Differ 2006;13:759–772.


SER

M on Septem


Dow

nloaded from


Angiogenesis and remodelling in human thoracic aortic aneurysms

Documents

Transcript of Angiogenesis and remodelling in human thoracic aortic aneurysms