Molecular Medicine

, Volume 20, Issue 1, pp 697–706 | Cite as

Gene Expression Study in Positron Emission Tomography-Positive Abdominal Aortic Aneurysms Identifies CCL18 as a Potential Biomarker for Rupture Risk

  • Audrey CourtoisEmail author
  • Betty V. Nusgens
  • Roland Hustinx
  • Gauthier Narnur
  • Pierre Gomez
  • Helena Kuivaniemi
  • Jean-Olivier Defraigne
  • Alain C. Colige
  • Natzi Sakalihasan
Open Access
Research Article


Rupture of abdominal aortic aneurysm (AAA) is a cause of significant mortality and morbidity in aging populations. Uptake of 18-fluorodeoxyglucose (FDG) detected by positron emission tomography (PET) is observed in the wall of 12% of AAA (A+), with most of them being symptomatic. We previously showed that the metabolically active areas displayed adventitial inflammation, medial degeneration and molecular alterations prefacing wall rupture. The aim of this study was to identify new factors predictive of rupture. Transcriptomic analyses were performed in the media and adventitia layers from three types of samples: AAA without FDG uptake (A0) and with FDG uptake (A+), both at the positive spot (A+Pos) and at a paired distant negative site (A+Neg) of the same aneurysm. Follow-up studies included reverse-transcriptase-polymerase chain reaction (RT-PCR), immunohistochemical staining and enzyme-linked immunosorbent assay (ELISA). A large number of genes, including matrix metalloproteinases, collagens and cytokines as well as genes involved in osteochondral development, were differentially expressed in the A+Pos compared with A+Neg. Moreover, a series of genes (notably CCL18) was differentially expressed both in the A+Neg and A+Pos compared with the AO. A significant increase of CCL18 was also found at the protein level in the aortic wall and in peripheral blood of A+ patients compared with A0. In conclusion, new factors, including CCL18, involved in the progression of AAA and, potentially, in their rupture were identified by a genome-wide analysis of PET-positive and -negative human aortic tissue samples. Further work is needed to study their role in AAA destabilization and weakening.


Rupture of abdominal aortic aneurysm (AAA) is a leading cause of death in western societies. Because AAA is generally asymptomatic, the evaluation of rupture risk and its prediction are challenging. Until now, the most commonly used factor to predict AAA rupture is its size (1, 2, 3), and a general rule is to perform surgery when the diameter is >55 mm. Nevertheless, growth of AAA is not linear and highly unpredictable (2,4). In the current practice, a conservative approach is often considered for patients with small AAA. However, they may also rupture, whereas many undiscovered aneurysms may grow to a considerable size without rupture. Therefore, diameter is not a fully reliable or specific predictor of the risk of rupture, necessitating the search for new markers.

Uptake of 18-fluorodeoxyglucose (FDG) visualized by positron emission tomography (PET) is currently used to detect hypermetabolic activity of cells seen in tumors and inflammatory processes. Several studies have also shown that FDG-PET can reliably detect leukocyte activities in atherosclerosis (5,6). We previously reported that a focal uptake of FDG is usually observed in the aortic wall of patients with rapidly expanding and symptomatic aneurysms that are prone to rupture. Our data also suggest that the FDG-PET-positive spots may correspond to the future site of rupture (7). They are further characterized by the presence of a dense inflammatory infiltrate mainly localized in the adventitia, sometimes organized in tertiary lymphoid organs, and by a marked loss of smooth muscle cells in the media and by an increased expression of molecules involved in the wall remodeling, notably matrix metalloproteinases (MMPs) (7).

This study aims to identify genes involved in this metabolic activation potentially related to rupture. A genomewide transcriptomic analysis was performed on biopsies collected from patients with an FDG uptake both in the positive spot (A+Pos) and in a distant negative site of the same aneurysm (A+Neg). These paired biopsies were further compared with samples collected from AAA patients with no FDG uptake (A0) to discriminate biological alterations associated with FDG uptake, to detect new systemic biomarkers correlated with a higher risk of rupture and to identify new pathways involved in the progression and rupture of AAA.

Materials and Methods

Patients and PET Computed Tomography Image Acquisition

The investigation conformed to the principals outlined in the Declaration of Helsinki. All patients gave written informed consent for the use of the aortic tissue and serum samples for research. The collection of the human samples was approved by the Ethics Committee of the University Hospital of Liège.

PET computed tomography (PET-CT) data were acquired in two hospitals (CHU and CHC), using either a Gemini BB (16-slice CT; Philips, Bruxelles, Belgium) or a Discovery LS (16-slice CT; GE Healthcare, Little Chalfont, UK). The image acquisition protocol was performed as previously described (7). The standardized uptake value (SUV) of the positive spot was expressed as the ratio between the uptake in the AAA and the liver (rSUV), with each patient acting as his or her own control. Patients with known connective tissue disorders or thoraco-AAAs were excluded from the study.

Tissue and Blood Collection

The number of samples available for the various analytical procedures was variable depending on clinical or practical reasons (no operation, biopsy not available, sample of poor quality or of too small size to allow all types of analyses).

In A0 patients, a biopsy was collected at the time of surgery in the anterior aneurysmal wall. For A+ patients, two fragments were collected: one at the site of FDG uptake (A+Pos), as localized using anatomical landmarks and CT images as detailed previously (7), and a second taken at a distant site negative for FDG uptake (A+Neg). A full thickness slice of the samples was fixed in 4% paraformaldehyde for histological and immunohistochemical analysis. When available, the remaining tissue was dissected to separate the media and adventitia, snap-frozen in liquid nitrogen and kept at −80°C until use.

Blood samples were collected from all patients undergoing a PET-CT exam (n = 33) and from 12 healthy individuals. Serum was separated by centrifugation for 10 min at 3,000g.

Microarray Analyses

Among the 12 AAA patients enrolled in the transcriptomic study (Table 1 and Supplementary Table S1), 6 patients had no visible FDG uptake (A0), whereas 6 others presented positive spots (A+Pos) localized within a larger negative area (A+Neg). Demographic information, AAA size and rSUV on these patients are presented in Supplementary Table S1. Tissue samples were ground at liquid nitrogen temperature (Dismembrator apparatus; B. Braun, Berlin, Germany). Total RNA was extracted with TriPure extraction reagen (Roche Diagnostics Corporation, Indianapolis, IN USA) or the RiboPure™ kit (Ambion [Thermo Fisher Scientific Inc., Waltham, MA, USA]) according to the manufacturer’s instructions. Total RNA was quantified (NanoDrop; Isogen Life Science, De Meern, the Netherlands), and its quality (expressed by the RNA quality indicator) was assessed by the Experion Automated Electrophoresis System (Bio-Rad). Three to five RNA samples from each group (adventitia and media of A0, A+Pos or A+Neg samples) with RNA quality indicator values ≥7 were pooled in equal concentration (Table 1). A total of 500 ng pooled RNA was labeled by using the GeneChip Expression 3′-Amplification One-Cycle Target Labeling Kit (Affymetrix, Santa Clara, CA, USA) following the manufacturer’s protocol. The cRNA was hybridized to the whole human genome U133Plus2.0 chip (Affymetrix) according to the manufacturer’s protocol. The data were generated with the MAS 5.0 algorithm included in the GeneChip Operating Software. Validations of microarray data were performed by real-time quantitative reverse-transcriptase-polymerase chain reaction (qRT-PCR) on individual RNA samples using gene-specific primer pairs (Supplementary Table S2).
Table 1

Clinical features of patients without (AO) and with FDG uptake (A+) used for transcriptomic study

Patient ID

AAA size (mm)


RNA pool (media layer)

RNA pool (adventitia layer)

































Median (IQR)

56.5 (55.3–59.3)

0.40 (0.35–0.54)


































Median (IQR)

54.5 (52.5–56.5)

1.00a (0.91–1.18)


Standardized uptake value ratio (rSUV): aortic SUVmax/hepatic SUVmax.

ap < 0.001 using Mann-Whitney U test.

The negative (A+Neg) and positive (A+Pos) sites of A+ patients were compared with each other and with samples of the A0 patients. Each comparison was performed separately for the media and the adventitia. A fold-change of ≥2 was used to select differentially expressed genes across the six comparisons. Lists of differentially expressed genes were imported into the DAVID software for analysis (8,9) and examined for biologically relevant associations by using the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genome (KEGG) pathway analysis (10). Enriched KEGG pathways were considered to be significant when the modified Fisher exact p value (Ease Score) was <0.05. National Center for Biotechnology Information (NCBI) official gene symbols available at were used for annotations.

Real-Time qRT-PCR

Microarray data were validated for selected genes of interest by real-time qRT-PCR on RNA from individual samples. Briefly, 500 ng total RNA was reverse-transcribed using SuperScript II Reverse Transcriptase (Invitrogen [Thermo Fisher Scientific]). Real-time PCR was performed in a final volume of 20 µL containing cDNA corresponding to 1 ng mRNA, 2 µmol/L of each primer and 10 µL of the qPCR MasterMix Plus for SYBR-Green (Eurogentec, Liège, Belgium) in the ABI Prism 7300 Sequence Detection System (Applied Biosystems [Thermo Fisher Scientific]). A dilution curve of a reference cDNA was performed for each gene product to calculate the reaction efficiency. Data were further standardized by using the geometrical mean of GAPDH and β2 microglobulin expression. These two genes were similarly expressed in all samples (data not shown). Primer sequences are listed in Supplementary Table S2.

Immunological Assays

For immunohistochemical analyses, sections from 22 paraffin-embedded AAA specimens (10 A0, 5 A+Neg and 7 A+Pos) were stained by using antibodies against CCL18 (Abcam ab104867; 1/500; Abcam, Cambridge, UK) and collagen type X (Abcam ab49945; 1/1,000). Briefly, antigen retrieval was achieved with 2.5% trypsin solution (Gibco, Invitrogen [Thermo Fisher Scientific]) (10 min at 37°C), and the samples were incubated with the primary antibody overnight at 4°C (in 10% bovine serum albumin). Peroxidase-conjugated secondary antibody (Dako, Glostrup, Denmark) was detected using AEC substrate (Dako). The CCL18 labeling was quantified by counting the number of positive cells in 10 fields at high magnification (×400). The collagen × staining was evaluated by computer-assisted image analysis (ImageJ software; National Institutes of Health, Bethesda, MD; [11]). The results were expressed in arbitrary units per surface unit.

For enzyme-linked immunosorbent assay (ELISA), proteins were extracted from separated media and adventitia layers of the aortic wall samples (n = 10 for A0 and n = 6 for A+ patients at the A+Pos and A+Neg sites). Tissues were homogenized in a buffer containing 0.05 mol/L Tris-HCl (pH 7.2), 0.15 mol/L NaCl, 0.002 mol/L CaCl2 and 2 mol/L urea. Serum samples were collected from 12 healthy donors and 33 AAA patients (17 A0 and 16 A+). CCL18 protein levels were measured in tissue extracts and serum by ELISA according to the manufacturer’s instructions (R&D Systems, Minneapolis, MN, USA).

Statistical Analysis

Results were expressed as the median and the interquartile range (IQR) and considered to be significant at the 5% critical level (p < 0.05) by using the Mann-Whitney U test for unpaired samples or the Wilcoxon signed-rank test for paired samples, as implemented in the Graphpad Prism software (GraphPad Software; GraphPad, La Jolla, CA, USA). Correlation between qRT-PCR and microarray data was calculated by using the Spearman test for nonparametric values as implemented in the GraphPad Prism software.

All supplementary materials are available online at .


Transcriptomic Analyses to Compare Expression in FDG-Positive and -Negative Aortic Tissue Samples Taken from the Media and Adventitia Layers

For the transcriptomic study, we used pooled RNA samples from patients with similar clinical features (Table 1 and Supplementary Table S1). In this population, six patients did not show any significant FDG uptake in the aneurysmal wall (A0 samples), whereas six others presented focal uptake spots (A+Pos samples) within a much larger region with no uptake (A+Neg samples). The median aortic diameter was similar in both groups (Table 1). These three types of samples were compared with each other (A+Pos/A0; A+Pos/A+Neg; A+Neg/A0) in the media and adventitia layers. In this experimental setting, genes that are modulated in a similar way both in the A+Pos and A+Neg compared with A0 likely reflect general alterations of the wall biology in patients with FDG uptake, whereas comparisons between the A+Pos and the A+Neg in the same aneurysm give insights into more local activation processes at sites potentially more prone to rupture. Using a twofold difference in expression as a cutoff, the six different comparisons identified several hundreds of genes (Figure 1 and Supplementary Table S3; detailed lists are provided in Supplementary Tables S4–S9).
Figure 1

Distribution of differentially expressed genes in the media (A) and adventitia (B) layers of AAA patients. Venn diagram to show the number of genes with differential expression with a fold-change ≥2 when comparing the A+Neg to the A0 patients (−/0), the A+Pos to the A0 (+/0) and the A+Pos to the A+Neg of the same patient (+/−), in the media (A) and the adventitia (B).

On the basis of pathophysiological knowledge and hypotheses, 13 differentially expressed genes were selected for follow-up studies. Primer pairs (Supplementary Table S2) were designed for validation of microarray data by qRT-PCR by using individual RNA samples of the media and adventitia layers. Among the 58 comparisons reported in Table S10 and S11, the microarray and qRT-PCR results were in agreement in that most of the ratios determined from microarray data were within the range of the values obtained by qRT-PCR. A significant Spearman correlation was found between the microarray and the qRT-PCR results (Figure 2). Altogether, these data demonstrate the reliability of the genome-wide transcriptomic analyses.
Figure 2

Correlation between the microarray expression data and real-time qRT-PCR data. A linear regression was established between the change in gene expression in the microarray result and the qRT-PCR for 57 comparisons. r = 0.75 and p < 0.0001 using the Spearman test. A.U., arbitrary unit.

Pathway Analysis of Differentially Expressed Genes

To determine the functions of the differentially expressed genes, we used the DAVID software to classify the genes into biological pathways. The analysis revealed that only approximately 25% of the differentially expressed genes could be assigned to KEGG pathways (5–15 depending on the comparisons). Many of the enriched KEGG pathways participate to related biological processes, explaining why a significant number of the differentially expressed genes can be found in several pathways.

Comparison between the A+Pos and the A+Neg showed a significant enrichment of seven pathways in the media and five in the adventitia samples (Tables 2 and 3). The identification of pathways related to the cardiovascular system was expected and further validated our data. The two pathways enriched both in the media and the adventitia (extracellular matrix [ECM] receptor interaction and focal adhesion) are related pathways and probably reflect remodeling of the ECM occurring in the FDG-positive samples. Also of interest was the identification of the cytokine-cytokine receptor interaction pathway potentially indicative of an immune reaction.
Table 2

Enriched KEGG pathways in the media of the A+Pos versus the A+Neg samples.


Total genes in the pathway (n)

Modulated genes in A+Pos (n)


hsa04512: ECM receptor interaction




hsa05412: Arrhythmogenic right ventricular cardiomyopathy




hsa04060: Cytokine-cytokine receptor interaction




hsa05410: Hypertrophic cardiomyopathy




hsa04510: Focal adhesion




hsa04514: Cell adhesion molecules




hsa05414: Dilated cardiomyopathy




Significantly enriched KEGG pathways among the differentially expressed genes in the media of A+Pos compared with A+Neg were identified by using the DAVID software (8). The total number of genes in the pathway, the number of modulated genes observed in our microarray data and the modified Fisher exact p value (Ease Score) are listed.

Table 3

Enriched KEGG pathways in the adventitia of the A+Pos versus the A+Neg.


Total genes in the pathway (n)

Modulated genes in A+Pos (n)


hsa04512: ECM receptor interaction




hsa04510: Focal adhesion




hsa00380: Tryptophan metabolism




hsa04270: Vascular smooth muscle contraction




hsa00350: Tyrosine metabolism




Significantly enriched KEGG pathways among the differentially expressed genes in the adventitia of A+Pos compared with A+Neg were identified by using the DAVID software (8). The total number of genes in the pathway, the number of modulated genes observed in our microarray data and the modified Fisher exact p value (Ease Score) are indicated.

Although the A+Neg and the A0 samples are both metabolically inactive, their comparison identified 15 enriched KEGG pathways in the media (Supplementary Table S12) and eight in the adventitia (Supplementary Table S13). Again, many pathways are directly related to ECM remodeling, signaling by soluble factors and immune reaction.

Identification of Groups of Genes Relevant to Systemic or Local Alterations of AAA

Among the differentially expressed genes (Table S4 to S9), those involved in extracellular homeostasis (diverse macromolecules, collagens, proteases) and in inflammatory processes (cytokines) were evaluated more extensively because of their potential role in the growth and rupture of AAA. Insulinlike growth factors and related molecules were also of interest, since many members of this family were differentially expressed. Selected groups of genes showing systemic or local changes in expression are illustrated for the media (Figure 3) and the adventitia (Figure 4) by using a color coding. Grey boxes represent genes with no change in expression (fold-change <2), whereas red and green represent increasing levels of upregulation or downregulation, respectively.
Figure 3

Expression of selected gene families in the aortic media on the basis of microarray analysis. Color codes correspond to the fold-change between the three comparisons: A+Neg compared with A0 (−/0), A+Pos compared with A0 (+/0) and A+Pos compared with A+Neg (+/−). Arrows designate genes validated by real-time PCR.

Figure 4

Expression of selected gene families in the aortic adventitia on the basis of microarray analysis. Color codes correspond to the fold-change between the three comparisons: A+Neg compared with A0 (−/0), A+Pos compared with A0 (+/0) and A+Pos compared with A+Neg (+/−). Arrows designate genes validated by real-time PCR.

Comparisons of the A+Neg with A0 (−/0 column in Figures 3 and 4) showed many similarities for transcripts related to ECM turnover and homeostasis (metalloproteases, collagens, osteochondral development) in both the media and adventitia. In contrast, expression of these genes was strongly altered in A+Pos compared with the two other conditions (+/− and +/0 columns). For example, two collagenases (MMP1 and MMP13) were strongly upregulated in the media and the adventitia of the A+Pos samples, although with a high interindividual variability (Figures 3 and 4). The expression of elastin and fibulin 5, two major components of the elastic fibers, was reduced, whereas the expression of metalloelastase (MMP12) was strongly upregulated in the adventitia. Some genes related to basement membrane (LAMA1, LAMA2, LAMB3, NID1) were also differentially regulated in the A+Pos sites. The most striking differences were found in the group of collagens, mainly fibril-forming collagens I, III, V and XII. Genes associated with osteochondral development were also aberrantly regulated in the adventitia, including cartilage-specific collagens X and XI, periostin (POST; a matricellular protein regulating collagen fibril formation and biomechanical properties of connective tissues), transcription factors (RUNX2, FOXC1) and factors participating in the Wnt signaling (WNT2, WNT5A, FRZB). A different situation was observed for cytokine-related genes, with more similarities between A+Pos and A+Neg (more gray boxes in the +/− columns and more colored boxes in the +/0 and −/0 columns). Of special interest was the high overexpression of IL6 and CCL18 in both the media and adventitia in the entire wall of A+ compared with A0 patients, since these factors are involved in inflammatory processes and could be used as circulating markers.

Increased Protein Production of CCL18 and Collagen Type X in the FDG-Positive Aortic Wall Compared with the FDG-Negative Samples

Protein expression and tissue localization of type X collagen were determined by immunostaining on tissue samples, to verify the activation of the osteochondral development pathway in the A+Pos adventitia samples (Figure 5A). Type X collagen was localized mainly in the adventitia, and there was significantly more staining in the A+Pos tissues compared with the A0 samples (Figure 5C). However, differences were less pronounced than expected from transcriptomic analyses and did not reach statistical significance between A+Pos and A+Neg.
Figure 5

Immunostaining of collagen type X (A and C) and CCL18 (B and D). Full-thickness aortic wall sections from representative patients with a PET-negative (A0) or a PET-positive (A+) AAA, at the negative (A+Neg) and the positive site (A+Pos) are shown. Semiquantitative analysis in A0 (n = 10), A+Neg (n = 5) and A+Pos (n = 7) for the collagen type X (C) staining and CCL18 (D) was carried out. **p < 0.01 and ***p < 0.001 using Mann-Whitney U test or Wilcoxon signed rank test for paired samples. Scale bar in collagen type X sections = 200 µm and in CCL18 sections = 50 µm.

Immunostaining also showed CCL18-positive cells mainly in the adventitia (Figure 5B) and in the inner part of the media. A semiquantitative analysis confirmed that there was a significantly increased number of CCL18-positive cells in the A+Neg and A+Pos groups (Figure 5D). These data were validated by ELISA in tissue extracts (Figures 6A, B). The level of CCL18 was not altered in the media of the different groups, whereas it was significantly increased in the A+Neg adventitia compared with the A0 patients and was even higher in A+Pos than in A+Neg. The circulating level of CCL18 (Figure 6C) was also measured by ELISA in the serum of 12 healthy donors (clinical characteristics in Supplementary Table S14) and 33 AAA patients (17 A0 and 16 A+). CCL18 concentrations were significantly higher in the AAA patients compared with the healthy population. A small but significant difference was also observed between A0 and A+. The circulating level of CCL18 was not correlated with the diameter of AAA (r = 0.9, data not shown).
Figure 6

CCL18 protein levels in the aortic wall (A and B) and serum (C) of AAA patients measured by ELISA. Box and whiskers plots of protein levels of CCL18 are shown. CCL18 was measured in the media (A) and adventitia (B) of AO (n = 11), A+Neg (n = 6) and A+Pos (n = 7) aortic wall samples. Circulating CCL18 (C) was measured in the serum of healthy people (n = 12) and AAA patients (n = 33; 17 AO and 16 A+). *p < 0.05 and ***p < 0.001 using Mann-Whitney U test or Wilcoxon signed rank test for paired samples.


Pathobiology of AAA is complex and not yet completely understood, although many pathogenic mechanisms have been suggested to participate in AAA progression and rupture (12). Because the majority of AAA patients are asymptomatic, aneurysms are often detected incidentally. However, aneurysms that produce symptoms, such as rapid enlargement and/or severe back or flank pain, are at increased risk of rupture, an often fatal complication. Finding new biological pathways and markers associated with risk for aneurysm rupture by using genome-wide approaches has been the focus of intense research for several years (10,13, 14, 15). These findings will allow better understanding of rupture risk and predicting this ultimate event. FDG, a metabolic marker used in PET imaging, accumulates preferentially in cells with high metabolic activity, such as tumor cells, activated macrophages and lymphocytes (16,17). Our cohort of patients can be divided into patients with or without significant FDG uptake in the aneurysmal wall. It is worth noting that most of the A0 patients were asymptomatic, whereas a significant number of A+ patients presented warning signs of impending rupture. The positive FDG uptake sites have been shown by us and others to correspond to the maximal peak wall stress (18) and to the site of rupture (19). The novelty of this study was to analyze media and adventitia layers separately from three types of surgical specimen collected either from aneurysms without any significant FDG uptake or from positive aneurysms, at both positive and negative uptake sites.

Enriched KEGG pathways were identified in the different types of samples. In the media, the cytokine-cytokine receptor interaction pathway was significantly enriched, suggesting the involvement of inflammatory mediators in AAA progression. It is worth noting that some cytokines, such as IL6 and CCL18, were upregulated not only in the A+Pos but also in the A+Neg compared with the A0. This observation suggests that environmental events and/or systemic factor(s) might affect the whole aneurysmal wall and activate pathways leading to wall degradation and weakening. In agreement with this hypothesis, the majority of A+ patients presented a significantly increased preoperative circulating C-reactive protein (CRP) levels in comparison to A0 patients (7). CCL18 was highly overexpressed throughout the surgical samples of A+ patients (A+Pos and A+Neg sites in the media and the adventitia). Its overexpression was validated by real-time qRT-PCR in all the patients of our series and confirmed at the protein level by ELISA. Immunohistochemical analyses of the aortic wall by an anatomopathologist showed that the main CCL18-producing cells are macrophages, with a few endothelial cells and fibroblasts. CCL18 is chemotactic for T and B cells (20), is detected in atherosclerosis plaques in regions rich in macrophages (21) and was also reported as a fibro-inflammatory mediator able to stimulate collagen production by lung and dermal fibroblasts (22). In this study, the plasma levels of CCL18 were significantly higher in the AAA population compared with healthy individuals, with the highest concentrations being measured in A+ patients. These data suggest that CCL18 could be used as a factor for evaluating the presence and the metabolic activity of AAA in combination with potential markers such as MMP9, which was previously shown to be correlated with the growth of AAA in some studies (23) or with C-reactive protein, which is increased in A+ patients (7).

The inflammatory environment, especially exacerbated in the A+Pos, certainly contributes to the intense ECM remodeling occurring during AAA progression (24) and ultimately leading to its rupture. According to this hypothesis, Dihlmann et al. (25) demonstrated recently that the inflammasome process was activated in the AAA wall at the site identified as high rupture risk areas by finite element analysis (25). ECM receptor interaction and focal adhesion KEGG pathways were modified both in the media and in the adventitia of AAA, mostly due to the differential expression of collagens and adhesion molecules such as integrins. On the basis of these data and on previous publications demonstrating the involvement of MMP2 and MMP9 in the growth and rupture of AAA (26, 27, 28, 29), the expression of genes involved in ECM accumulation and remodeling were studied in more detail. We found that the expressions of several MMPs, notably collagenases (MMP1, MMP8 and MMP13), elastases (MMP9 and MMP12), matrilysin (MMP7) and MT1-MMP, the main activator of MMP2, are all upregulated in the A+Pos samples. Of special interest, we measured in the adventitia a >10-fold increase of MMP1 in the A+Pos compared with the A+Neg, and the expression of MMP13, which was absent in the AO and weakly expressed in the A+Neg, was also strongly upregulated in the A+Pos. MMP13 has been shown to play a crucial role in chronic inflammatory diseases (30) and is increased in AAA (31). The presence of these collagenases at the FDG uptake site could contribute to the degradation of fibrillar collagens I and III, which provide the ultimate tensile strength to the aorta. This scenario could result in a progressive weakening of the wall and finally its rupture (32).

A tentative repair process by the adventitial cells has been previously suggested by the increased expression of COL1A1, a main fibrillar structural component of the adventitia (33,34). We have extended this observation in our microarray analysis. Besides increased expression of type I, III and V fibril-forming collagens, many others collagen types were also overexpressed in the A+Pos, notably COL10A1 and COL11A1, two collagen types predominantly localized in cartilage and vitreous body. COL10A1 is expressed by hypertrophic chondrocytes during endochondral ossification (35), whereas COL11A1, besides its cartilaginous expression, has been found in peritumoral myofibroblasts (36). Interestingly a number of other genes involved in cartilage and bone development were also upregulated in the A+Pos in the adventitia. Among them, periostin (POSTN), a matricellular protein involved in cell adhesion, was first identified in osteoblastic cells (37) and was recently described to be involved in AAA development (38). POSTN localizes mainly in bone but has been also found in fibrous connective tissues subjected to mechanical stress such as heart valves and tendons and in several types of tumors (rev. in 37). It is also overexpressed during vascular injury by proliferative smooth muscle cells to facilitate their migration in the adventitia (39). In a recent proteomic study POSTN was reported to be increased in AAA (40). Altogether, these original data suggest that cells in the adventitial tissue with a positive FDG uptake possess a specific phenotype leading to the production of an inappropriate ECM. The presence and increased levels of collagen type X protein observed by immunostaining in AAA walls further support this hypothesis. The nature of these cells and their relationship to progenitor lineage need further investigation.


The global gene expression analysis performed here on FDG-positive and -negative aortic tissue samples identified several genes, such as CCL18, potentially involved in aneurysm progression and rupture. Further studies on larger populations of patients are needed to replicate the findings and assess their clinical value.


The authors declare that they have no competing interests as defined by Molecular Medicine, or other interests that might be perceived to influence the results and discussion reported in this paper.



We thank the surgeons and nurses of the Department of Cardiovascular and Thoracic Surgery of the CHU of Liège, the CHC St Joseph and CHR Citadelle for their contribution to the collection of AAA specimens. We thank Benoit Hennuy and the GenoTranscriptomics platform of the GIGA for carrying out the microarray experiments. We thank Joan Somja for analysis of the histological sections. This work was supported by the FP7 European Program Fighting Aneurysmal Diseases number 200647, the Belgian Fonds de la Recherche Scientifique (grant 3.4518.07) and a grant from the University of Liège.

Supplementary material

10020_2014_2001697_MOESM1_ESM.pdf (562 kb)
Supplementary material, approximately 562 KB.


  1. 1.
    Sakalihasan N, Limet R, Defawe OD. (2005) Abdominal aortic aneurysm. Lancet. 365:1577–89.CrossRefPubMedGoogle Scholar
  2. 2.
    Limet R, Sakalihassan N, Albert A. (1991) Determination of the expansion rate and incidence of rupture of abdominal aortic aneurysms. J. Vasc. Surg. 14:540–8.CrossRefPubMedGoogle Scholar
  3. 3.
    Brown PM, Zelt DT, Sobolev B. (2003) The risk of rupture in untreated aneurysms: the impact of size, gender, and expansion rate. J. Vasc. Surg. 37:280–4.CrossRefPubMedGoogle Scholar
  4. 4.
    Kurvers H, et al. (2004) Discontinuous, staccato growth of abdominal aortic aneurysms. J. Am. Coll. Surg. 199:709–15.CrossRefPubMedGoogle Scholar
  5. 5.
    Sakalihasan N, Michel JB. (2009) Functional imaging of atherosclerosis to advance vascular biology. Eur. J. Vasc. Endovasc. Surg. 37:728–34.CrossRefPubMedGoogle Scholar
  6. 6.
    Rudd JH, Hyafil F, Fayad ZA. (2009) Inflammation imaging in atherosclerosis. Arterioscler. Thromb. Vasc. Biol. 29:1009–16.CrossRefPubMedPubMedCentralGoogle Scholar
  7. 7.
    Courtois A, et al. (2013) 18F-FDG uptake assessed by PET/CT in abdominal aortic aneurysms is associated with cellular and molecular alterations prefacing wall deterioration and rupture. J. Nucl. Med. 54:1740–7.CrossRefPubMedGoogle Scholar
  8. 8.
    Huang da W, Sherman BT, Lempicki RA. (2009) Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat. Protoc. 4:44–57.CrossRefGoogle Scholar
  9. 9.
    Huang da W, et al. (2007) DAVID Bioinformatics Resources: expanded annotation database and novel algorithms to better extract biology from large gene lists. Nucleic Acids Res. 35:W169–75.CrossRefPubMedPubMedCentralGoogle Scholar
  10. 10.
    Lenk GM, et al. (2007) Whole genome expression profiling reveals a significant role for immune function in human abdominal aortic aneurysms. BMC Genomics. 8:237.CrossRefPubMedPubMedCentralGoogle Scholar
  11. 11.
    Schneider CA, Rasband WS, Eliceiri KW. (2012) NIH Image to ImageJ: 25 years of image analysis. Nat. Methods. 9:671–5.CrossRefPubMedPubMedCentralGoogle Scholar
  12. 12.
    Michel JB, et al. (2011) Novel aspects of the pathogenesis of aneurysms of the abdominal aorta in humans. Cardiovasc. Res. 90:18–27.CrossRefPubMedGoogle Scholar
  13. 13.
    Choke E, et al. (2009) Whole genome-expression profiling reveals a role for immune and inflammatory response in abdominal aortic aneurysm rupture. Eur. J. Vasc. Endovasc. Surg. 37:305–10.CrossRefPubMedGoogle Scholar
  14. 14.
    Hinterseher I, et al. (2013) Novel pathways in the pathobiology of human abdominal aortic aneurysms. Pathobiology. 80:1–10.CrossRefPubMedGoogle Scholar
  15. 15.
    Tromp G, Kuivaniemi H. (2009) Developments in genomics to improve understanding, diagnosis and management of aneurysms and peripheral artery disease. Eur. J. Vasc. Endovasc. Surg. 38:676–82.CrossRefPubMedGoogle Scholar
  16. 16.
    Newsholme P, Curi R, Gordon S, Newsholme EA. (1986) Metabolism of glucose, glutamine, long-chain fatty acids and ketone bodies by murine macrophages. Biochem. J. 239:121–5.CrossRefPubMedPubMedCentralGoogle Scholar
  17. 17.
    Bos R, et al. (2002) Biologic correlates of (18)fluorodeoxyglucose uptake in human breast cancer measured by positron emission tomography. J. Clin. Oncol. 20:379–87.CrossRefPubMedGoogle Scholar
  18. 18.
    Xu XY, et al. (2010) High levels of 18F-FDG uptake in aortic aneurysm wall are associated with high wall stress. Eur. J. Vasc. Endovasc. Surg. 39:295–301.CrossRefPubMedGoogle Scholar
  19. 19.
    Sakalihasan N, Hustinx R, Limet R. (2004) Contribution of PET scanning to the evaluation of abdominal aortic aneurysm. Semin. Vasc. Surg. 17:144–53.CrossRefPubMedGoogle Scholar
  20. 20.
    Schutyser E, Richmond A, Van Damme J. (2005) Involvement of CC chemokine ligand 18 (CCL18) in normal and pathological processes. J. Leukoc. Biol. 78:14–26.CrossRefPubMedPubMedCentralGoogle Scholar
  21. 21.
    Wortmann M, Peters AS, Hakimi M, Bockler D, Dihlmann S. (2014) Glyoxalase I (Glo1) and its metabolites in vascular disease. Biochem. Soc. Trans. 42:528–33.CrossRefPubMedGoogle Scholar
  22. 22.
    Atamas SP, et al. (2003) Pulmonary and activation-regulated chemokine stimulates collagen production in lung fibroblasts. Am. J. Respir. Cell Mol. Biol. 29:743–9.CrossRefPubMedGoogle Scholar
  23. 23.
    Golledge J, Tsao PS, Dalman RL, Norman PE. (2008) Circulating markers of abdominal aortic aneurysm presence and progression. Circulation. 118:2382–92.CrossRefPubMedPubMedCentralGoogle Scholar
  24. 24.
    Liapis CD, Paraskevas KI. (2003) The pivotal role of matrix metalloproteinases in the development of human abdominal aortic aneurysms. Vasc. Med. 8:267–71.CrossRefPubMedGoogle Scholar
  25. 25.
    Dihlmann S, et al. (2014) Increased expression and activation of absent in melanoma 2 inflammasome components in lymphocytic infiltrates of abdominal aortic aneurysms. Mol. Med. 20:230–7.CrossRefPubMedPubMedCentralGoogle Scholar
  26. 26.
    Sakalihasan N, Delvenne P, Nusgens BV, Limet R, Lapiere CM. (1996) Activated forms of MMP2 and MMP9 in abdominal aortic aneurysms. J. Vasc. Surg. 24:127–33.CrossRefPubMedGoogle Scholar
  27. 27.
    Petersen E, Wagberg F, Angquist KA. (2002) Proteolysis of the abdominal aortic aneurysm wall and the association with rupture. Eur. J. Vasc. Endovasc. Surg. 23:153–7.CrossRefPubMedGoogle Scholar
  28. 28.
    Wilson WR, et al. (2006) Matrix metalloproteinase-8 and -9 are increased at the site of abdominal aortic aneurysm rupture. Circulation. 113:438–45.CrossRefPubMedGoogle Scholar
  29. 29.
    Aziz F, Kuivaniemi H. (2007) Role of matrix metalloproteinase inhibitors in preventing abdominal aortic aneurysm. Ann. Vasc. Surg. 21:392–401.CrossRefPubMedPubMedCentralGoogle Scholar
  30. 30.
    Wernicke D, Seyfert C, Gromnica-Ihle E, Stiehl P. (2006) The expression of collagenase 3 (MMP-13) mRNA in the synovial tissue is associated with histopathologic type II synovitis in rheumatoid arthritis. Autoimmunity. 39:307–13.CrossRefPubMedGoogle Scholar
  31. 31.
    Tromp G, et al. (2004) Elevated expression of matrix metalloproteinase-13 in abdominal aortic aneurysms. Ann. Vasc. Surg. 18:414–20.CrossRefPubMedGoogle Scholar
  32. 32.
    Loftus IM, Thompson MM. (2002) The role of matrix metalloproteinases in vascular disease. Vasc. Med. 7:117–33.CrossRefPubMedGoogle Scholar
  33. 33.
    Defawe OD, et al. (2003) TIMP-2 and PAI-1 mRNA levels are lower in aneurysmal as compared to athero-occlusive abdominal aortas. Cardiovasc. Res. 60:205–13.CrossRefPubMedGoogle Scholar
  34. 34.
    Sakalihasan N, Heyeres A, Nusgens BV, Limet R, Lapiere CM. (1993) Modifications of the extracellular matrix of aneurysmal abdominal aortas as a function of their size. Eur. J. Vasc. Surg. 7:633–7.CrossRefPubMedGoogle Scholar
  35. 35.
    Gouttenoire J, et al. (2004) Modulation of collagen synthesis in normal and osteoarthritic cartilage. Biorheology. 41:535–42.PubMedGoogle Scholar
  36. 36.
    Garcia-Ocana M, et al. (2012) Characterization of a novel mouse monoclonal antibody, clone 1E8.33, highly specific for human procollagen 11A1, a tumor-associated stromal component. Int. J. Oncol. 40:1447–54.PubMedGoogle Scholar
  37. 37.
    Merle B, Garnero P. (2012) The multiple facets of periostin in bone metabolism. Osteoporos. Int. 23:1199–212.CrossRefPubMedGoogle Scholar
  38. 38.
    Yamashita O, et al. (2013) Periostin links mechanical strain to inflammation in abdominal aortic aneurysm. PLoS One. 8:e79753.CrossRefPubMedPubMedCentralGoogle Scholar
  39. 39.
    Lindner V, Wang Q, Conley BA, Friesel RE, Vary CP. (2005) Vascular injury induces expression of periostin: implications for vascular cell differentiation and migration. Arterioscler. Thromb. Vasc. Biol. 25:77–83.CrossRefPubMedGoogle Scholar
  40. 40.
    Didangelos A, et al. (2011) Extracellular matrix composition and remodeling in human abdominal aortic aneurysms: a proteomics approach. Mol. Cell. Proteomics. 10:M111.008128.CrossRefPubMedPubMedCentralGoogle Scholar

Copyright information

© The Author(s) 2014

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, and provide a link to the Creative Commons license. You do not have permission under this license to share adapted material derived from this article or parts of it.

The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

To view a copy of this license, visit (

Authors and Affiliations

  • Audrey Courtois
    • 1
    • 2
    • 3
    Email author
  • Betty V. Nusgens
    • 2
  • Roland Hustinx
    • 4
  • Gauthier Narnur
    • 4
    • 5
  • Pierre Gomez
    • 5
  • Helena Kuivaniemi
    • 6
    • 7
  • Jean-Olivier Defraigne
    • 3
  • Alain C. Colige
    • 2
  • Natzi Sakalihasan
    • 1
    • 4
  1. 1.Surgical Research Center, GIGA-Cardiovascular Science UnitUniversity of LiègeLiègeBelgium
  2. 2.Laboratory of Connective Tissues Biology, GIGA-Research (GIGA-R)University of LiègeLiègeBelgium
  3. 3.Department of Cardiovascular and Thoracic SurgeryUniversity Hospital of Liège (CHU), University of LiègeLiègeBelgium
  4. 4.Department of Nuclear MedicineUniversity Hospital of Liège (CHU), University of LiègeLiègeBelgium
  5. 5.Department of Nuclear MedicineSt Joseph Hospital (CHC)LiègeBelgium
  6. 6.The Sigfried and Janet Weis Center for ResearchGeisinger Health SystemDanvilleUSA
  7. 7.Department of SurgeryTemple University School of MedicinePhiladelphiaUSA

Personalised recommendations