Detailed list of proteins differentially expressed between AAA and PAO samples.
Abdominal aortic aneurysms (AAAs) are a potentially fatal disorder. Atherosclerotic changes followed by accelerated degradation of collagen and elastin, the main components of the extracellular matrix (ECM) in the vascular wall, cause the development of AAAs in deteriorating aortic walls that dilate progressively and may eventually rupture. Other alterations associated with AAAs include marked changes in the cellular composition of the aortic wall, especially the infiltration of macrophages and T-lymphocytes into the adventitia and a major reduction in the population of vascular smooth muscle cells (SMCs) (Lopez-Candales et al., 1997; Henderson et al., 1999).
AAA appear to be a consequence of complex mechanisms involving several potential factors: immunological, inflammatory, chemotactic, apoptotic, protease-related, angiogenic and fibrinolytic (Ailawadi et al., 2003). Recently, interleukin-8 and monocyte chemoattractant protein 1 expression were shown to be raised in the AAA biopsies compared to the abdominal aorta of controls, suggesting that pathways involving these proteins may be involved in AAA pathologies (Middleton et al., 2009). AAA is a disease associated with chronic inflammation in the aortic wall and leukotrienes are powerful lipid mediators released by inflammatory cells (Samuelsson, 1983). A recent study showed the increased expression of leukotriene C4 synthase together with the predominant formation of cysteinyl-leukotrienes in human AAA, linked with matrix metalloproteinases (MMP) production (Di Gennaro et al., 2010). This suggests a mechanism by which leukotrienes may promote matrix degradation in the AAA wall.
Despite reports that major atherosclerotic risk factors are related to AAA (Iribarren et al., 2007), we still lack the information about the process of aneurysmal degeneration (Wassef et al., 2007, Golledge et al., 2008) that we need to develop non-invasive approaches for its early detection. In aneurysm formation, collagen degradation exceeds its synthesis and together with excessive degradation of other ECM macromolecules, notably elastin, ultimately promotes rupture of these aneurysms. Increased local production of proteases, that is, enzymes capable of degrading collagen and elastin ECM proteins, have been identified in aneurysm sites (Dobrin & Mrkvicka, 1994; Knox et al., 1997; Thompson, 2006). The source of these proteases in humans is unclear, however, although many studies have measured their expression in cells from biopsies or tissues obtained after surgery (Wilson et al, 2006).
2. Purpose of the study
The aim of the study was to conduct a global comparison of the mRNA and protein expression profiles in the vascular cells to determine the genes involved in the pathological progression of aneurysmal disease by comparing the transcriptomic and proteomic profiles of SMCs and macrophages of patients with AAAs and those who have peripheral arterial occlusion (PAO) but no AAAs (verified by imaging). Using samples from two groups of patients in LILAS (the Lille Anevrysmal Study, described below), those with AAAs and others with PAO, we isolated SMCs from aortic tissue and monocytes (that were differentiated into macrophages) from blood. Although these diseases have common mechanisms, inflammatory cells essentially infiltrate the media and adventitia of the arterial wall in AAAs, whereas they are mostly in the intima in PAO. Because the factors that influence these distinctions are unknown, we designed a study that would allow us to distinguish “pro-aneurysmal” risk factors from those that are associated with atherogenesis and stenosis. We choose PAO disease as control diseases as Thompson et al., (Thompson, 2002) described the separation of AAA disease from other atherosclerotic diseases like peripheral vascular disease and carotid disease.
Following recommendations for large-scale parallel quantitation of proteins (Kingsmore, 2006), we have constructed a protein microarray that contained on one slide 42 different antibodies representing proteins found by transcriptomic and proteomic analysis to be differentially modulated in macrophages and SMC from AAA and PAO patients. Finally, we performed the same analysis of plasma from the same patients to identify proteins involved in the aneurysmal process that are more easily detectable, as hypothesized that circulating biomarkers can reflect inflammation and degeneration in the AAA wall (Golledge et al., 2008).
3. Population study
The LILle Anevrysmal Study (LILAS) was a case-control study that enrolled 42 men, either AAA (case patients) or peripheral arterial occlusion (PAO) (control patients) who needed a vascular surgery or endovascular treatment at the Lille University Hospital Centre (Lille, France) (Lamblin et al., 2010). The ethics committee of the Lille University Hospital Center (France) approved the study (CCPPRB n° CP03/47 of 06 May 2003), and each patient provided written informed consent. The protocol required a blood sample (80 ml) to be taken at inclusion, before the surgery. All patients underwent the planned surgical treatment and, when possible, surgeons removed the thrombus and took a sample of the aneurysm wall in the case patients (n=24) and a sample of the aortic wall during the bypass grafting for the control PAO patients (n=18).
The mean age of subjects was 65.6 ± 6.9 years but AAA patients were older (68.0 ± 6.1 years versus 62.3 ± 6.6 years). No other cardiovascular risk factors differed between the groups. In both groups, one third of the patients had hypertension and half dyslipidemia. Diabetes mellitus was found in 17% of the AAA group and 22% of the PAO group, an insignificant difference. Both groups also had similar percentages of family members with a history of heart disease (myocardial infarction and/or acute coronary syndrome): 29% in AAA patients and 35% in PAO patients. The AAA case patients (n=24) had a mean maximal external aortic diameter of 56.1 ± 11.3 [range: 46-95] mm. All the PAO patients had undergone recent vascular imaging of the aorta and lower limbs before surgery: none had AAAs or iliac, femoral or popliteal aneurysms. The groups did not differ concerning their cardiovascular treatment (Table 2, Lamblin et al, 2010). Before surgery, fewer than two thirds had antiplatelet treatment. Fewer AAA than PAO patients received ACE inhibitor treatment (21% versus 44%). Three of the 18 PAO patients took non-steroidal anti-inflammatory drugs and no AAA patients..
The difference of age between the two groups (AAA and PAO) of patients recruited in this study is consistent with literature on the mean age of onset of the disease in patients with similar risk factors. Also consistent with the literature is the finding that 30% of AAA patients and 50% of PAO patients have family members who have had a myocardial infarction or acute coronary syndrome.
4.1. Tissue biopsies from AAA and PAO patients
Aortic samples (Fig. 1) were obtained for 20 of the 24 AAA patients of the LILAS study. Because treatment procedures for PAO changed during the study period, from surgical to endovascular interventions, tissues were available for only the 4 of the 18 patients who had surgical treatment (Fig. 1).
4.2. Primary smooth muscle cells cultures
Human SMCs were obtained from all of the AAA tissue samples (n=20) and only one of the four samples from PAO patients. Individual cultures of human aortic SMCs were established for each patient by dissection of a residual segment of abdominal aorta and then cutting out the media into 1- to 2-mm3 pieces and culturing them on dishes coated with collagen type 1. Cells began to migrate from the explant after 10 days and confluence was reached in one month.
4.3. Primary cultures of human monocytes-derived macrophages
Primary cultures of human macrophages are the cell models used most often in vitro for human macrophages, because only a simple venous blood sample is required to investigate qualitative or quantitative macrophage modifications in a case-control study. A highly-standardized primary culture had to be devised to take into account the individual genetic variability of each blood donor, as well as the behavioral and environmental influences, and hidden abnormalities (Korke et al., 2002). Primary cultures of human macrophages were prepared as previously described (Pinet et al., 2003), with a technique adapted from Boyum (Boyum, 1968).
4.4. Transcriptomic analysis of vascular cells from AAA and PAO patients
Specific oligonucleotides for 137 genes corresponding to 24 matrix metalloproteinases (MMP), 4 tissue inhibitors of metalloproteinases (TIMP), 20 ADAMs (a disintegrin and metalloproteinase), 20 ADAMTS (ADAMS with thrombospondin motifs), 65 other proteases and 4 control genes (hydroxymethylbilane synthase, beta-actin, NADPH oxidase 1, and Glyceraldehyde-3-phosphate dehydrogenase (GADPH)) were designed with OLIGOMER software (Mediagen, France) as previously described (Lamblin et al., 2010). Data were analyzed in comparison to a reference RNA, mixture from brain, colon, heart, placenta and testis, chosen for the expression of each protease and anti-protease in at least one of these tissues. We performed a self (Cy3)-self (Cy5) hybridization of the reference RNA and observed equivalent staining throughout the microarray, except for the renin gene which we excluded from the analysis (not shown). RNA samples were extracted from human SMCs and macrophages and only RNA with a RIN value >9 was analyzed. Each sample (5 μg) was analyzed with two slides according to a dye-swap strategy to account for labeling and detection differences between Cy5 and Cy3. Data analysis was performed with the statistical language R (v 2.0.1) (Ihaka & Gentelman, 1996), more specifically with the LIMMA library (Linear Models for MicroArray data) (Smyth et al, 2003). In accordance with the MIAME (Minimum Information About a Microarray Experiment) guidelines, we note the steps involved in data processing: 1) each microarray was scanned twice (at high and low intensity) for each wavelength of Cy3 and Cy5; 2) two separate normalization steps were conducted : first, pin-by-pin and second, a lowess fitness normalization. A moderated t-statistic with empirical Bayes shrinkage of the standard error was used to classify the statistically significant modulations (Lönnstedt & Speed, 2003). Because of multiple testing, p-values were corrected (Benjamini & Hochberg, 1995) to control for the false discovery rate. Genes with an adjusted p-value < 0.01 in at least three of five replications were considered further.
4.5. Proteomic analysis of vascular cells from AAA and PAO patients
Intracellular proteins from macrophages and SMCs of AAA and PAO patients were loaded (5 μg) on 2D gels. 2D-DIGE was performed as recently described (Dupont et al, 2008). Image analysis was performed with Progenesis SameSpots v2.0 software. The differences in protein spots were then analyzed. Spots were considered to have significantly different normalized spot volumes if the fold change was greater than 1.5 and the corresponding p value (one-way ANOVA analysis) was significant. The last step applied multivariate statistics to the selected spots by calculating q values (for the false discovery rate) and power. Spots corresponding to proteins differentially expressed between the different groups of samples were identified by MALDI-TOF following the protocol previously described (Acosta-Martin et al, 2009).
4.6. Protein microarray profiling of vascular cells and plasma from AAA and PAO patients
The protein arrays allowed us to monitor 42 protein-antibody pairs and detected the profile of differentially expressed proteins in aneurysmal disease. Four differentially expressed proteins were selected from 2D analysis of macrophages from the LILAS or an earlier study: HSP70, aconitase-1, GRP75 and beta-actin (Dupont et al, 2008). Four others were selected from microarray analysis, TIMP-3, ADAMTS10, fibronectin, and tenascin. Finally, four more were identified by the 2D-analysis of plasma from patients from LILAS (personal communications), apolipoprotein AII, vitronectin, transthyretin and factor H.
Recently, laser microdissection was used to analyze mRNA expression from macrophages and SMCs from aneurysmal tissue of a rat model of AAA created using elastase infusion (Sho et al., 2005). They found that modulation of at least 5 mRNAs differed according to cell type and flow conditions and thus concluded that a global analysis of aneurysmal tissue could mask cellular responses specific to inflammation and flow. These recent data reinforce our strategy to quantify mRNA and protein expression in two types of vascular cells from patients with different pathologies: macrophages and SMCs from AAA and PAO patients, instead of analysing the whole AAA or PAO biopsies.
5.1. Primary smooth muscle cells cultures
All the cultured human SMCs had the same elongated, spindle-shaped morphology, and at confluence all cultures assumed a hill-and-valley pattern that was maintained throughout all subcultures (Fig. 2).
Protein expression profiles determined by western blot analysis showed that all the cultures from both patient groups, expressed α-smooth muscle actin, vinculin and metavinculin. This finding indicates the presence of differentiated SMCs, as well as fibronectin, a component of basement membrane, that is indicative of a proliferative and secretory SMC phenotype (Fig. 2).
5.2. Primary cultures of human monocytes-derived macrophages
After, 12 days of culture, monocytes are differentiated into macrophages as shown by phase-contrast microscopy (Fig. 3).
The quality of macrophage cultures was evaluated by characterization of two markers, ECE-1 (Endothelin converting enzyme-1) mRNA and MMP-9 (Matrix metalloproteinase-9). ECE-1 mRNA was detected by RT-PCR on total RNA isolated from macrophages and MMP-9 activity was determined in culture medium by gelatin zymography (Pinet et al, 2003). Sixteen primary MDMs cultures, met these criteria and we observed no differences between macrophages obtained from AAA or PAO patients (Fig. 3).
5.3. Transcriptomic analysis of smooth muscle cells and macrophages from AAA and PAO patients
The importance of proteases as mediators of ECM degradation in vascular disease has been proved (Wight, 2005), as the role of other proteases and their inhibitors in AAA (Choke et al., 2005). Our global analysis on a dedicated array of proteases and anti-proteases was not limited by advance theories or hypotheses, precisely because it was intended to generate hypotheses. This technology enables more efficient selection of the possible genes involved in AAAs, as shown in two studies that used AAA specimens and a cDNA microarray (Tung et al., 2001) or quantitative RT-PCR (Higashikata et al., 2004).
We tested mRNA from SMCs of 12 different AAA patients and 2 mRNA samples from one PAO patient, and we compared them to the reference RNA. For human macrophages, mRNA expression ratio of AAA/PAO genes was tested from 4 different AAA patients and 5 different PAO patients.
The microarray analysis indicated that three mRNA sample species appeared to be significantly modulated with a q value (moderated p-value) <0.01 in SMCs from AAA compared with PAO patients: angiotensin converting enzyme (ACE), ADAMTS5 and ADAMTS8. In the AAA patients, ACE (AAA/PAO ratio: - 2.8 fold) was down-regulated, and ADAMTS5 (AAA/PAO ratio: 1.2 fold) and ADAMTS8 (AAA/PAO ratio: 1.3 fold) were very slightly but significantly up-regulated (Fig. 4A). The graphs of individual mRNA expression (AAA/PAO ratio) for these three genes show the variability between patients (Fig. 4B). Of the 137 genes analyzed first by microarrays and then by Q-PCR for selected genes, all showed significant and concordant up- or down-regulation of expression levels for CML samples. The down-regulation of ACE mRNA in AAA patients is not surprising as a recent meta-analysis showed that ACE I/D polymorphisms are associated with a significant risk of AAAs (Thompson et al., 2008), especially in patients with hypertension (Korcz et al., 2009). Similarly, we found that ADAMTS5 and ADAMTS8 were up-regulated in SMC of AAA compared with PAO patients. These two members of the ADAMTS family are involved in vascular lesion development and can degrade different proteoglycans present in blood vessels (Yao et al., 1994).
Figure 4C shows that the mRNA samples identified as differentially expressed in macrophages from AAA and PAO patients are different from those differentially modulated in SMCs. Differential modulation was observed in eight macrophage mRNA samples. Two were down-regulated in AAA compared with PAO patients: SCARB1 (scavenger receptor class B, member 1) AAA/PAO(ratio: 1.8) and BACE2 (beta-site APP-cleaving enzyme 2) (AAA/PAO ratio: 1.2). Six were up-regulated in AAA patients: caveolin 1 (AAA/PAO ratio: 1.8), ADAMTS10 (AAA/PAO ratio: 1.7), APP (amyloid precursor protein) (AAA/PAO ratio: 2.3), MMP-27 (AAA/PAO ratio: 2.3), ADAM12 (AAA/PAO ratio: 1.9) and TIMP3 (AAA/PAO ratio: 2.3). Microarray analysis of macrophages showed the modulation of other proteases and anti-proteases, but we were unable to validate these results by RT-PCR because of the limited amount of RNA obtained from cultures. Interestingly, TIMP-3, which we found to be up-regulated in macrophages from AAA patients, has been shown to be a potent inhibitor of ADAMTS5 (Kashiwagi et al., 2001). Nonetheless, our results are consistent with the data reported (Sho et al., 2005) and thus demonstrate the interest of analyzing cell-specific compared with tissue-specific responses.
5.4. Proteomic analysis of smooth muscle cells and macrophages from AAA and PAO patients
Proteins from macrophages and SMCs of AAA and PAO patients were analysed on 2D gels. An average image was established from the scanned gels for macrophages (Fig. 5A) and for SMC (Fig. 5B) from AAA and PAO patients.
This differential analysis found twenty-one polypeptidic spots with differential abundance levels and a statistically reproducible difference over the series of gels. Four spots were up-regulated (spots 16, 17, 18 and 19) and 17 down-regulated (spots 1, 2, 3, 4, 5, 6, 7, 8, 9; 10, 11, 12, 13, 14, 15, 18 and 19) in macrophages from AAA compared with PAO patients (Fig. 5A). The intensity of each spot was calculated as the mean ± SD and expressed as a percentage of normalized volume. The same approach was used for the SMC proteins, although only two 2D gels were used for the PAO SMCs (Fig. 5B). This differential analysis found seven polypeptidic spots with differential abundance levels and a statistically reproducible difference over the series of gels. Two spots were up-regulated (spots 2 and 5) and 5 down-regulated (spots 1, 3, 4, 6, and 7) in SMC from AAA compared with PAO patients.
Table 1 summarizes the identity and factor of variation of proteins differentially expressed by macrophages and SMCs from AAA and PAO patients. Of the 21 polypeptidic spots differentially expressed for macrophages, 17 were identified by mass spectrometry and corresponded to 13 non-redundant proteins. Unfortunately, we were unable to identify proteins for SMC due to either the low intensity signal on the mass spectrometry or a low probability score.
Of the macrophage proteins identified, three are components of cytoskeleton: microtubule-associated protein tau (spot 2), beta-actin (spots 17 and 18) and macrophage-capping protein (spot 15) and cathepsin S (spot 18), has protease activity.
This dual transcriptomic and proteomic analysis of vascular cells from AAA and PAO patients revealed both the differences and the complementarity of microarray and 2D-electrophoresis technologies. Proteomic analysis identified only one protease, cathepsin S as differentially expressed in macrophages from AAA and PAO patients. This is consistent with previous reports showing an increase of cathepsin S activation in the wall of AAAs (Abisi et al., 2007; Abdul-Hussein et al., 2007). Still more recently, leukocyte cathepsin S was shown to be involved in macrophage apoptosis and increased SMC content and collagen deposition (De Nooijer et al., 2009). Interestingly, of the 13 proteins identified overall, three (microtubule-associated protein tau, beta-actin and macrophage-capping protein) are components of cytoskeleton, while the others are involved in protein metabolism and stress response. Unfortunately, we were unable to identify the SMC proteins corresponding to the polypeptidic spots differentially expressed.
Because of the different size of the proteome and transcriptome, we did not conduct a global correlation analysis of proteins and mRNA levels of SMCs and macrophages from these patients. Our strategy was to use specific and dedicated tools to identify the modulation of proteins and mRNAs in SMC and macrophages from AAA and PAO patients and to apply the same technology to their plasma samples.
5.5. Protein microarray profiling of SMCs, macrophages and plasma from AAA and PAO patients
Because of the limited amount of proteins available from macrophages and SMCs of our LILAS study, we used protein antibody array technology that has been shown to provide protein information in a systematic, reproducible and high-throughput fashion in several diseases (Sanchez-Carbayo et al., 2006; Weber et al., 2007).
The AAA/PAO ratio for each protein modulated is presented in detail in Table 2.
First, we analysed macrophages and SMCs proteins from LILAS patients. The microarray, revealed up-regulation in AAA patients of three macrophage proteins and down-regulation of six, as well as up-regulation of six SMC proteins and down-regulation of eight. All the proteins up-regulated in macrophages from AAA patients were also up-regulated in their SMCs: fibronectin, β-actin, tenascin. Five of six proteins down-regulated in AAA macrophages were also down-regulated in their SMCs: TIMP-3, ADAMTS5, HSP70, aconitase, GRP75.
Plasma analysis, of course, restricted the number of proteins available that could be detected as differentially modulated between the two groups. Interestingly, analysis of plasma from the LILAS patients showed only five proteins differentially expressed in plasma from AAA and PAO patients, three already selected from the macrophage and SMC samples, ADAMTS8 (AAA/PAO ratio: 0.9), ADAMTS5 (AAA/PAO ratio: 1.11) and TIMP-3 (AAA/PAO ratio: 1.15) and two only in plasma: ApoCIII (AAA/PAO ratio: 0.78) and alpha1-antitrypsin (AAA/PAO ratio: 0.88).
Of particular interest, however, is the finding that TIMP-3, ADAMTS5 and ADAMTS8 were differentially expressed in proteins from macrophages, SMCs and plasma. ADAMTS8 is known to degrade ECM proteoglycans and has the highest specific activity for cleaving aggrecan (Porter et al., 2005). Our data are in agreement with a report (Theocharis et al., 2001), which showed a decrease in the concentration of versican in human AAA. Unfortunately, tools to measure ADAMTS activity need to be developed (Wagstäter et al., 2008). Interestingly, TIMP-3 was shown to be modulated in the two cell types involved in AAA, SMCs and macrophages and also in the plasma of the same patients, as shown with the overexpression of TIMP-3 in AAA tissue compared to aortic occlusive disease but they did not measure plasma TIMP-3 levels of the same patients as we did (Carell et al., 2002). The availability of TIMP assays allowed us to confirm the increased plasma levels of TIMP-3 in AAA patients (Lamblin et al., 2010).
The limitations of this study are the relatively few tissue specimens allowing study of SMCs from PAO patients, because of changes in their treatment strategy (endovascular protheses rather than surgery). To our knowledge, this is the first study to combine several global analyses to assess the changes in the expression of genes and proteins of vascular cells involved in aneurysmal disease. Its strengths include the largest and most detailed view of changes in proteases and anti-proteases expression in patients presenting AAA. The protein array techniques confirmed that the differentially expressed proteins could also be detected in plasma. We found a restricted number of proteins differentially expressed between AAA and PAO patients: TIMP-3, ADAMTS5 and ADAMTS8that differ significantly in plasma of AAA patients compared to PAO approach. Combining transcriptomic and proteomic is a valid approach to a better understanding of the pathophysiology of AAA but strengthen the need to investigate multiple circulating biomarkers.
This work was supported by grants from “Fondation pour la Recherche Médicale”, Région Nord Pas de Calais and a European FAD grant (Health-F2-2008-200647). All the authors declare that they have no conflicting financial interests.
- Nicolas Lamblin1,2,3,4, Philippe Ratajczak1,2,3, David Hot2,5,6, Emilie Dubois1,2,3, Maggy Chwastyniak1,2,3, Olivia Beseme1,2,3, Hervé Drobecq2,3,6, Yves Lemoine2,5,6, Mohammad Koussa4 and Philippe Amouyel1,2,3,4