Abstract
Lysyl oxidases (Lox), which are members of the amine oxidase family, are involved in the maturation of elastic lamellae and collagen fibers. Modifications of amine oxidases in idiopathic annulo-aortic ectasia disease (IAAED) have never been investigated. Our aim was to examine the expression of several proteins that might interfere with elastic fiber organization in control (n = 10) and IAAED (n = 18) aortic tissues obtained at surgery. Expression of amine oxidases and semicarbazide-sensitive amine oxidase (SSAO), and cellular phenotypic markers were examined by immunohistopathology and confocal microscopy. The expression of these proteins was assessed in relation to clinical and histomorphological features of the arterial wall. In control aorta, SSAO staining was expressed along elastic lamellae, whereas in aneurysmal areas of IAAED, SSAO was markedly decreased, in association with severe disorganization of elastic lamellae. Smooth muscle myosin heavy chain was also decreased in IAAED compared with controls, indicating smooth muscle cell dedifferentiation. Multiple regression analysis showed that elastic lamellar thickness (ELT) was correlated positively with the SSAO: elastin ratio and negatively with the Lox: elastin ratio, and that the clinical features of IAAED (aneurysm, thoracic aorta diameter, and aortic insufficiency) were positively correlated with ELT but not with SSAO. The relationship between SSAO expression and ELT suggests that this amine oxidase may be involved in elastic fiber organization. However, in advanced IAAED, the deficit in SSAO expression could be secondary to the decrease and fragmentation of elastic fibers and/or to vascular smooth muscle cell dedifferentiation.
A
Aneurysms may result from destruction of the extracellular matrix (ECM) via inflammatory proteases, such as those observed in complicated atherosclerosis (Fontaine et al. 2002), or from an alteration in elastic fiber organization, which occurs in cases of quantitative and/or qualitative defects of some molecular components of elastic or collagen fibers, the major ECM constituents of the arterial wall. However, a genetically determined deficit in elastin leads to VSMC proliferation during development and arterial stenosis (Li et al. 1997, 1998). By contrast, Marfan syndrome, a genetic disorder caused by mutations in the gene for fibrillin-1, a component of the microfibrils, is accompanied by aortic aneurysms due to a lack of elastic fiber integrity (Dietz et al. 1991; Kainulainen and Peltonen 1991). Some thoracic aortic aneurysms associated with aortic insufficiency (aortic regurgitation as defined by Doppler echocardiography), which fit into none of the above-mentioned categories, are classified as idiopathic annulo-aortic ectasia disease (IAAED) and remain pathologically undefined (Savunen and Aho 1985; Savunen 1986). Because the etiology of IAAED is not related to any known genetic disorder, it is an interesting model in which to study possible, as yet undiscovered, intrinsic structural disorders of the ECM.
During embryonic arterial development and growth, functional elastic fibers are formed by cross-linking of tropoelastin. The first step in this process is the formation of δ-aldehyde, allysine, through oxidation of ∊-amino groups by the secreted copper-dependent enzyme lysyl-oxidase (Lox) (Bedell-Hogan et al. 1993; Li et al. 2000; Kagan and Li 2003). Subsequent nonenzymatic condensation of modified and unmodified lysines leads to formation of characteristic bifunctional and tetrafunctional cross-links specific to insoluble elastin (Smith-Mungo and Kagan 1998; Maki et al. 2002; Kagan and Li 2003). Assembly of collagen molecules into fibrils also requires the formation of crosslinks by Lox. Therefore, the process of cross-linking is essential for correct elastic and collagen fiber function. In humans, the importance of the cross-linking is illustrated by the Menkes syndrome, a functional deficit in Lox related to a mutation in the gene coding for a copper-transporting ATPase, that results in a fragmentation of elastic fibers (Smith-Mungo and Kagan 1998).
Semicarbazide-sensitive amine oxidase (SSAO), also called vascular adhesion protein-1 (VAP-1), is another copper-containing amine oxidase. A high activity of SSAO is associated with VSMCs localized in the media of the mammalian aorta (Lyles 1996; Jalkanen and Salmi 2001), and its expression is dependent on the level of maturation of VSMCs (El Hadri et al. 2002). Using a pharmacological approach, Langford et al. (1999) have shown that chronic blockade of vascular SSAO activity leads to a striking disorganization of the elastic fiber architecture within the aortic media, accompanied by a decrease in the mature elastin content and an increase in collagen, leading to aortic dilatation (Langford et al. 1999). Therefore, it is conceivable that, like Lox, SSAO may contribute to the cross-linking processes that participate in the organization of elastin and collagens. Very recently, it has been reported (Göktürk et al. 2003) that in a murine transgenic model, overexpression of human SSAO in smooth cells leads to abnormal structure of the aortic elastic laminae. These studies have suggested that like Lox, SSAO may contribute to the organization of the elastin network. SSAO has been shown to be expressed by human adult aortic VSMCs but is not implicated in atherosclerotic or inflammatory diseases (Jaakkola et al. 1999). To our knowledge, no study has evaluated SSAO expression in human aneurysmal disease.
The aim of our study was to examine the expression of several proteins that could be involved in elastic fiber organization at the aneurysmal site in IAAED. These molecules were analyzed in parallel with elastic lamellar thickness (ELT) considered as an index of elastic fiber structural integrity. Expression of the amine oxidases (Lox and SSAO) was evaluated by immunohistochemistry (IHC) and confocal microscopy. We also investigated alterations in SMC differentiation using monoclonal antibodies (MAbs) specific to sm-myosin heavy-chain (sm-MHC) and sm-α-actin. We observed that although SSAO and Lox were easily detectable in the vicinity of elastic fibers of normal aorta, only SSAO expression was dramatically reduced in the media of IAAED, even after adjustment for decreased elastic fiber content. SSAO content was positively correlated with ELT in both control and IAAED subjects. Whether this marked decrease in SSAO expression in IAAED is simply the consequence of elastic fiber alterations and SMC dedifferentiation, or whether it itself contributes to the disorganization of elastic lamellae, cannot be concluded from this study and requires further investigation.
Materials and Methods
Patients
Our histomorphological study concerns 18 patients with IAAED. Clinical diagnosis of thoracic aortic aneurysm was confirmed by transthoracic echocardiography and aortography. The presence of aortic insufficiency (aortic regurgitation) was examined by Doppler echocardiography. Age, sex, weight, height, blood pressure, personal and family medical history, and vascular risk factors were recorded. Patients underwent dermatological, rheumatological, and ophthalmological examinations to detect any ECM disease, and skin biopsy to identify and exclude Marfan syndrome (Beighton et al. 1988).
Normal aortic tissues were obtained from patients who had died in accidents and who were used as heart graft donors, in accordance with the requirements of the local Ethical Committee (CHU Bordeaux). All the controls were sex- and age-matched and free of coronary artery disease and of major vascular risk factors including diabetes, hypercholesterolemia, hypertension, and smoking. This study was performed in a clinical-research hospital project (PHRC-CHU, Bordeaux, France).
Histomorphometry, Immunohistochemistry, and Confocal Microscopy
Aortic samples were obtained during Bentall's surgery. Tissues were placed in Dulbecco's modified Eagle's medium (Life Technologies EUROBIO, Courtaboeuf, France). Samples were then separated into two fragments, one fixed in paraformaldehyde for histochemical studies and one freeze dried for immunohistology. Aortic sections already paraformaldehyde-fixed and paraffin-embedded were stained with Weigert's resorcin/fuschin and Masson's trichrome for elastin fiber, collagen content, and SMC identification (Michel et al. 1994; Sauvage et al. 1997). Thickness of elastic lamellae was measured across the entire media on 10 different fields of each Weigert's-stained section to take into account all elastic lamellae of an aneurysmal area.
MAbs specific to α-sm-actin, sm-MHC, collagen type IV, and elastin were purchased from Sigma (St Louis, MO). Antibodies to Lox were a gift of Dr P. Sommer (IPCB CNRS; Lyon, France), while antibodies to metalloproteinase (MMP)-2 and to MMP-9 were from Chemicon (Temecula, CA) and from Santa Cruz Technology (Santa Cruz, CA), respectively. SSAO antibody was a kind gift of Dr S. Jalkanen (Medicity Research Laboratory, Department of Biochemistry and Pharmacy; Turku University, Finland).
Because antigenic preservation is sensitive to tissue fixation conditions, tissues were quickly deep frozen in liquid nitrogen, then freeze dried and embedded in paraffin as described previously (Louis et al. 2000). Bound primary antibodies were detected using biotinylated anti-mouse, anti-rat, anti-rabbit, or anti-goat secondary antibodies and the streptavidin/biotinylated horseradish peroxidase complex (Amersham Pharmacia Biotech; Orsay, France). The final complex was visualized using a DAB/peroxidase kit from Vector (Burlingame, CA). Sections were counterstained in Mayer's hemalun, dehydrated, and mounted in a EUKITT medium (Kindler; Freiburg, Germany). Controls were performed by omitting the primary antibody.
Histomorphometric data and immunostaining were quantified at high-power magnification by color video image analysis using an IBM PC (Daniel Lamazière et al. 1993). The video camera was connected to a microscope. The software Quancoul (Quant'Image; INSERM U441, Pessac, France) defined true colors on the basis of three independent parameters (hue, intensity, and saturation). The parameters were calibrated against a background of control antigen as previously described (Daniel Lamazière et al. 1993; Bézie et al. 1998). To account for the color intensity as an index of the amount of antigenic detection, we expressed the collagen density and antigenic content in arbitrary units (AU) as the product of the positively labeled surface by the hue intensity and optical density of the same positive pixels (Daniel Lamazière et al. 1993; Bézie et al. 1998). At least 10 sections for each aorta were measured to calculate the mean value for each sample. Because the surgical arterial biopsies did not allow study of the full thickness layer, all measurements were normalized to normal or injured arterial area. We controlled the validity of our quantitative immunological method in accordance with ELISA quantification (Figure 1).
For confocal examination, antigen detection was revealed by the use of a fluorescent (Cy3 red) species-specific secondary antibody. Elastin autofluorescence was examined with a confocal microscope (Nikon PCM 2000). Visualization was obtained by EZ 2000 software, and tridimensional reconstruction was performed using 20 images at 0.5-μm intervals with Imaris software (Bitplane; Zurich, Switzerland). This results in an appearance of perspective in a one-plane image.
Statistical Analysis
Results are expressed as means ± SD. Statistical comparisons between groups were made using the non-parametric Mann-Whitney test. Clinical parameters were analyzed after adjustment for age and sex. Differences in SSAO expression between IAAED and controls were analyzed after adjustment for elastin content using a general linear model. Univariate correlations and multiple regression analysis with stepwise selection were used to assess the determinants of ELT and clinical IAAED parameters. Statistical analysis was carried out using the NCSS 2000 software. p<0.05 was considered statistically significant.
Results
Clinical Data
All patients were selected on the criterion of absence of any known genetic disorders and any clinically detected arteriosclerosis. Major clinical parameters found in our 18 IAAED patients are summarized in Table 1. The control population and IAAED patients were similar with respect to age, body weight, height, and arterial blood pressure. The proportion of males and females studied was comparable in both groups. All clinical morphological parameters, as well as immuno-histology of skin biopsies, confirmed that these patients did not have a Marfan phenotype. The thoracic aorta diameter evaluated after aortography and the number of patients with aortic insufficiency were significantly higher in the IAAED group than in the control group. These differences remained significant even after adjustment for age and sex.
Antigen quantification, comparison of ELISA, and digital Image analysis. ELISA (
Clinical characteristics in IAAED and control subjects a
aIAAED, idiopathic annulo-aortic ectasia disease; BP, brachial blood pressure. Values are mean ± SD.
Arterial Wall Gross Morphology
The gross morphology of control elastic lamellae is shown in Figure 2A with typical staining of the elastin network. Figure 2B shows severe disorganization of elastic lamellae with areas of complete disappearance of the elastic fiber network in IAAED specimens. In contrast, the collagen density was similar in normal and IAAED biopsies (34.36 ± 9.24 for controls vs 30.44 ± 12.12 AU/cm2 for IAAED; p>0.05). The elastin density was greatly decreased (−42.7%) in the IAAED population in comparison to the control group (Table 2). Overall, the mean ELT was 52.3% lower in IAAED than in controls (Figure 2B; Table 2).
Elastic Fiber Structure and Amine Oxidase Expression
Figures 2C-2F and 3 show the immunohistological distribution of SSAO. There was a marked decrease in SSAO immunolabeling in IAAED specimens (Figures 2D, 2F, and 3) in comparison to controls (Figures 2C and 2E). Confocal microscopy shows (after 3D reconstruction) that in control aortic tissue (Figure 2E), SSAO (red) was located in the close vicinity of the elastic lamellae (green-yellow). This association was greatly decreased at the aneurysmal site (Figures 2F and 3). We selected a transition area (Figure 3) between the aneurysmal site and the area in which the organization of the arterial wall was comparable to that of controls. In this area, there was a negative gradient of SSAO (red) and ELT (green) toward the aneurysmal site. We describe three zones (Figure 3): in zone 1, we found a pattern of well-organized elastic fibers with large amounts of SSAO in close vicinity, similar to that observed in control arteries (Figure 2E); in zone 2, there was a marked decrease in SSAO staining and ELT was slightly decreased; in zone 3, there was an important decrease in both SSAO and ELT.
Table 2 shows that SSAO immunolabeling was significantly decreased at the aneurysmal site. This reduction was significant even after adjustment for elastin content. The SSAO:elastin ratio was significantly lower in IAAED than in control subjects (0.16 ± 0.13 vs 0.38 ± 0.09; p<0.001). Table 2 also compares Lox amine-oxidase immunolabeling between IAAED patients and controls. Both populations had a very low level of Lox, but this level was significantly higher in IAAED patients than in controls. Figure 4 shows the distribution of individual values of both amine oxidases in the two populations. In IAAED, the variability of Lox was greater than that of SSAO and a large overlap was observed between control and IAAED groups.
VSMC Differentiation
We next examined SMC differentiation using two smooth muscle-specific antibodies, sm-α-actin and sm-MHC. There was no modification of sm-α-actin at the aneurysmal site in IAAED patients. However, sm-MHC was decreased 2.7-fold (p<0.001) compared with control values (Figure 5A). This SMC dedifferentiation was not associated with an elevation in the number of VSMCs.
To investigate whether elastic fiber degradation is involved in the decrease in fiber thickness, we studied the level of metalloproteinases in IAAED. Figure 5B shows that MMP-2 and MMP-9 are not increased in IAAED.
Regression Analysis
Figure 6 shows that ELT was positively correlated with SSAO content in both the IAAED group (R 2 = 0.81; p<0.0001) and in controls (R 2 = 0.48; p<0.03). There was no significant correlation between ELT and Lox in either the IAAED group or controls. There was no significant correlation between sm-α-actin and sm-MHC content and the amount of SSAO in either group, or between ELT and sm-MHC content.
Multiple regression analysis showed that ELT was positively correlated with the SSAO:elastin ratio and negatively with the Lox:elastin ratio, representing 69% of the variability of ELT (Table 3). Other parameters, including collagen and sm-MHC, were not significantly correlated with ELT. We next analyzed the more significant clinical IAAED parameters, including aneurysm, aortic dilatation, and aortic insufficiency. IAAED aneurysm (R 2 = 0.88; p<0.00001), thoracic aortic dilatation (R 2 = 0.30; p<0.00001), and aortic insufficiency (R 2 = 0.54; p<0.00001) were positively correlated with ELT (not shown).
Discussion
The structural integrity of the arterial wall is dependent on both SMC differentiation and the ECM network, two factors that ensure tonicity and maintenance of mechanical properties. In IAAED, the expression of cellular or matrix proteins that may be involved in elastic fiber disorganization remains largely unknown. We particularly focused on the expression of two amine oxidases, Lox and SSAO, in which interest in regard to cardiovascular diseases has recently emerged (Jaakkola et al. 1999; Jalkanen and Salmi 2001; Kagan and Li 2003). Our main finding was a strong expression of SSAO along the elastic fiber in control subjects and a marked decrease in SSAO expression at the aneurysmal site in IAAED patients, correlated with a reduction in ELT. These data represent the first demonstration of changes in SSAO expression in diseased human arteries. However, our experimental approach does not allow us to establish a causal relationship between decreased SSAO content and elastic lamellar disorganization.
Morphological aspect of elastic fibers stained by the Weigert technique and nuclear counterstaining with hemalum in control (
The decrease in SSAO expression in IAAED patients and its positive correlation with ELT should be interpreted with caution. One must bear in mind that our study was performed on IAAED specimens that were probably at a late stage of the evolution of the disease. It is therefore impossible to determine whether during the natural history of IAAED the SSAO decrease represents an early event that precedes the elastic fiber disorganization. Because elastin turnover is very slow (Brown-Augsburger et al. 1996), this hypothesis can be considered only if SSAO contributes in some way to the postnatal stabilization of the elastin network.
Elastic fiber characteristics in IAAED and control subjects a
aAU, arbitrary units; IAAED, idiopathic annulo-aortic ectasia disease; SSAO, semi-carbazide-sensitive amine oxidase; Lox, lysyl oxidase. Values are mean ± SD.
Several lines of evidence in our experimental data and those of others suggest that SSAO may be involved in elastic fiber organization: (
Confocal microscope images, 3D reconstructed, of the thoracic aorta in one IAEED. This area (233 × 233 μm) is close to an aneurysm and shows a gradient of SSAO (red) immunostaining with elastic fiber (green) disorganization. Zone 1 is the junction with the normal part of the arterial wall. Zone 2 is a transitory area where elastic fibers are slightly thinner but with a marked decrease in SSAO content. Zone 3, in the aneurysmal site, displays important elastic fiber disorganization and a marked SSAO reduction.
Comparison of amine oxidase distribution in IAAED patients (closed circles) and controls (open circles). Lox content was increased in some IAAED patients, but there was a large overlap between the two populations. In contrast, the SSAO contents of each population were clearly distinct.
At a biochemical level, mechanisms by which SSAO might contribute to the organization of the ECM network remain speculative. Lox, which is structurally related to SSAO, is involved in collagen and elastin cross-linking and in chemotaxis of VSMCs (Kagan et al. 1984; Smith-Mungo and Kagan 1998; Li et al. 2000; Kagan and Li 2003). This amine oxidase can use NH2 in lysine side chains in proteins as a substrate. SSAO is generally considered to metabolize primary amines, such as benzylamine, methylamine, or aminoacetone, to generate the corresponding aldehyde, hydrogen peroxide, and ammonia. Interestingly, a recent study (Salmi et al. 2001) has shown by molecular modeling that a polypeptide can fit into a groove on the surface of SSAO at a position that overlies the catalytic center of SSAO. Overall, in endothelial cells, a lysine-containing peptide (GGGGKGGGG) fitting into the groove decreases SSAO-mediated hydrogen peroxide production in response to benzylamine, indicating that it interferes with enzyme catalytic activity (Salmi et al. 2001). By contrast, irrelevant lysine-containing peptides have no effect on SSAO activity. It is therefore conceivable that in addition to soluble primary amines, Lox and SSAO may act on amino acids included in matrix proteins. In addition, considering our methodological approach and the size of arterial samples, it is not possible to establish the relationship between SSAO tissue expression and activity, but this has previously investigated in the work of Jaakkola et al. (1999) and Andres et al. (2001). In this context, further morphological, biochemical, and molecular investigations will be required to determine this relationship and the exact nature of the SSAO substrate(s) in the blood vessels.
Comparison of smooth muscle differentiation and metal-loprotease expression. (
Linear regressions of SSAO vs elastin fiber thickness in controls and IAAED patients. Open circles for controls (y = 1.011 × + 1.5529; R 2 = 0.4831; p<0.03); closed circles for IAAED (y = 1.9754× − 5.4653; R 2 = 0.8112; p<0.001). See Materials and Methods for statistical analysis.
Stepwise regression analysis of the determinants of elastic lamellar thickness a
aSSAO, semicarbazide-sensitive amine oxidase; Lox, lysyl oxidase; MMP, metalloproteinase; sm-MHC, sm-myosin heavy-chain.
However, the advanced state of disorganization of the arterial wall observed in IAAED does not allow us to establish this potential role of SSAO. Indeed, SMC dedifferentiation and elastin fragmentation and degradation could themselves account for the marked reduction in SSAO expression. Our results show that in contrast to atheroslerotic lesions, there is no increase in MMP2 and MMP9 expression in IAAED, suggesting that they do not play a role in elastolysis in this disease. However, we cannot exclude a modification in MMP activities as previously studied in metastatic rat liver using zymography and IHC (Mook et al. 2003). Although the relationship between MMP expression and activity has been well documented in atherosclerotic aneurysms, this relationship has never been explored in IAAED. In addition, other elastolytic enzymes exist, some of which can be produced by the VSMCs. In this context, the decrease in sm-MHC immunolabeling demonstrated that the differentiation process of VSMCs is altered in IAAED. This result is in agreement with our study in vitro (El Hadri et al. 2002) showing that rat VSMCs in the undifferentiated state express low levels of SSAO, whereas VSMC differentiation is accompanied by a large increase in SSAO gene and protein expression. Alternatively, recent studies have shown that elastin plays a direct role in controlling the VSMC phenotype. Therefore, it is conceivable that the reduction in elastin promotes VSMC dedifferentiation, thus contributing to SSAO downregulation in the vascular wall.
In contrast to SSAO, we observed an increase in Lox expression in IAAED, but Lox expression did not correlate with ELT in the controls or the IAAED population. However, using multiple regression analysis on the whole population, a decreased Lox:elastin ratio appeared as a significant determinant of ELT, but with a lower contribution than the increased SSAO:elastin ratio. In light of the importance of Lox in the biogenesis of the ECM by cross-linking both collagen and elastin (Smith-Mungo and Kagan 1998), we cannot exclude the possibility that this increase in Lox expression in IAAED represents an adaptive mechanism in response to ECM disorganization. Whether a crosstalk mechanism exists between Lox and SSAO expression and/or function in the arterial wall remains an open question.
In conclusion, the present study demonstrates that SSAO is expressed along elastic lamellae in intact arterial wall. In IAAED, SSAO, in contrast to Lox, was markedly decreased, associated with severe disorganization of the elastic lamellar network. The relationship between SSAO expression and ELT suggests that this amine oxidase may be involved in elastic fiber organization. However, in advanced IAAED, the deficit in SSAO expression could be secondary to elastic fiber fragmentation and reduction and/or to VSMC dedifferentiation. Experiments using SSAO gain or loss of function will be required to establish the potential interaction between SSAO and elastic fibers.
Footnotes
Acknowledgements
Supported by grants from INSERM and University Victor Segalen Bordeaux II. Specific grants were given by the European Community (5 PCRD) contract no. QLK6-CT-2001-00332 and by the Etablissement Public Régional of Aquitaine.
We thank Michel Safar and Mary Osborne-Pellegrin for helpful discussion.