Impaired VE-Cadherin/β-Catenin Expression Mediates Endothelial Cell Degeneration in Dilated Cardiomyopathy
Background— The cross-talk between vascular endothelial growth factor (VEGF)-A, angiopoietin (Ang), and VE-cadherin coregulates endothelial cell (EC) survival. Cardiac expression of VEGF-A but not its receptor KDR is blunted in dilated cardiomyopathy (DCM). Whether VE-cadherin/Ang function is affected in DCM is unknown.
Methods and Results— The myocardial expression of VE-cadherin/β-catenin, Ang-1, Ang-2, and their receptor Tie-2 was examined in DCM, ischemic cardiomyopathy (ICM), and in control subjects through the use of real-time RT-PCR, Western blotting, and immunocytochemistry. EC degeneration was quantified by TEM. RNA interference against VE-cadherin and VEGF deprivation and stimulation were applied to cultured DCM myocardium and human microvascular ECs to examine the interplay between VEGF, VE-cadherin/β-catenin, and Ang-2. Analysis of tissue sections with similar rates of EC degeneration in both patient groups showed that VE-cadherin/β-catenin expression was downregulated in DCM only (P<0.05). Although Ang-1 was not changed, Ang-2 expression was downregulated and Tie-2 protein expression was upregulated both in DCM and ICM (P<0.05). The ratio of degenerated to normal ECs was significantly higher in DCM versus ICM (P<0.05). Targeted VE-cadherin gene silencing in cultured human ECs resulted in similar degenerative effects observed in myocardial ECs of DCM patients. In vitro experiments indicated that VE-cadherin/β-catenin expression is independent of VEGF.
Conclusions— These results indicate for the first time that the EC survival is impaired in myocardium of patients with DCM involving VE-cadherin/β-catenin, probably independent of VEGF. Targeting VE-cadherin might be of benefit to counteract the selective EC pathology in DCM.
Received April 23, 2002; de novo received May 14, 2003; accepted July 9, 2003.
The genetic mechanisms underlying dilated cardiomyopathy, a disorder with left ventricular dilation often leading to congestive heart failure, are slowly being unraveled.1 Mutation of multiple genes encoding for proteins that interact with the cytoskeleton have been described as causing dilated cardiomyopathy (DCM) in some patients, yet more genes are expected to be identified (genetic heterogeneity).2 Although end-stage DCM and ischemic cardiomyopathy (ICM) share many clinical features, chronic hypoxia, which is considered a major trigger of ICM, is absent in DCM.3 Accordingly, in mice knocked out for VEGF-A164 and VEGF-A188, ICM develops,4 although expression of VEGF-A165 and VEGF-A189 is downregulated in myocardium of patients with DCM.5
VEGF has various functions on endothelial cells (ECs), primarily the induction of proliferation, differentiation, and survival. The survival function of VEGF depends on an interaction between VEGF-R (KDR), PI3-kinase, and Akt.6,7 Ang-1 also promotes, whereas angiopoietin (Ang)-2 suppresses, endothelial survival, at least in the absence of angiogenic stimuli.6,8 Gene truncation studies in the mouse indicate that VE-cadherin controls EC survival by involving its cytoplasmic domain, β-catenin, and KDR.9 VE-cadherin specifically contributes to EC differentiation and is expressed only at endothelial adherens junctions.10,11 The extracellular domain of VE-cadherin mediates initial cell adhesion, whereas the cytosolic tail is required for interaction with the cytoskeleton and junctional strength, mediated through β-catenin.10–13 Absence of VE-cadherin prevents a normal organization of new vessels, suggesting that VE-cadherin is a key mediator of angiogenesis.10
It appears that the cross-talk between VEGF/Ang/VE-cadherin coregulates EC survival. The actions of these systems overlap, yet as different vascular beds have specific survival mechanisms,6 and since the specific VEGF-survival mediator KDR is not affected in myocardium of patients with DCM,5 this study sought to elucidate the role of Ang/VE-cadherin in endothelial survival in DCM.
For this purpose, we profiled gene and protein expression of target molecules in myocardium of patients with DCM and ICM as well as in control nonfailing hearts, through the use of gene array, real-time RT-PCR, Western blotting, and immunocytochemistry. Electron microscopy of sectioned tissues served to assess EC morphology. To address possible interaction of VEGF and VE-cadherin in ECs, gene silencing was performed in cultured human dermal microvascular ECs (HDMVECs) as well as in cultured myocardial tissue using RNA-interference (RNAi) and a fluorescent reporter gene. Our results indicate that VE-cadherin/β-catenin but not Ang-1 cardiac expression is downregulated in DCM, independent of VEGF-A, and is associated with increased EC degeneration.
Patients and Tissue Sampling
A total of 84 patients scheduled for transplantation, including 60 with DCM and 24 with ICM, who gave informed consent to be enrolled, were studied. The control group included 12 donors whose hearts could not be transplanted due to surgical reasons. Table summarizes the demographic and clinical data of the patients. The results of clinical tests of coded patients were documented. Patients with arterial hypertension, myocardial infarction (6 months before surgery), or valve diseases were excluded. Multiple myocardial biopsy specimens were taken from the anterior left ventricular wall. The selection criteria for biopsy specimens were (1) preoperative: The biopsy was taken 2 cm distal to the diagonal branch and 2 cm from the LAD. If the artery was occluded, resulting in myocardial infarction, confirmed by angiography and echocardiography, the patient was excluded. Patients with ICM had stenoses in the LAD ranging from 75% to 90%. These preoperative criteria helped us to exclude necrotic and to include ischemic myocardium into the study (2) histological criteria: If unexpected scar regions were detected in TEM despite the preoperative tests, the patient was excluded.
Poly(A+)-RNA was isolated with the use of Oligotex-dT kit (Quiagen).14 First-strand cDNA synthesis was performed with 2 μg poly(A+) RNA and an AMV-RT kit (Promega). The RNA strand within the DNA-RNA duplex was degraded,15 and the products were purified on a Sephadex G-50 spun-column (Pharmacia Corp). For reverse-strand priming, first-strand cDNA was used to generate [α-32P]dCTP-labeled second-strand cDNA,15 which was used for the cDNA arrays (Cardiovascular Array, Clontech). cDNA hybridization signals were quantified through the use of ImageQuant software (Molecular Dynamics).
TEM EC Morphology
From each heart, two biopsy specimens were embedded in Epon; four ultrathin sections were made per specimen and analyzed in TEM at ×1000. The upper right and the lower left quadrant in each grid (resulting in an equal morphometry area) were studied. In each quadrant, two fields were used, yielding 32 fields for each heart and a total of 1920 fields for DCM, 768 fields for ICM, and 384 fields for control. In each field, 20 capillaries were studied. The endothelium was categorized as degenerated if the endothelial membrane was damaged, for example, discontinuous or fringed; if mitochondria were degenerated16; if vesicles were formed inside the cytoplasm; and if the nuclear membrane was damaged. Images were obtained and analyzed with the use of a morphometry software (Lucia G, Optoteam). The average number of normal and degenerated ECs was calculated for each field. Mean values were calculated for DCM, ICM, and control.
Sampling for RNA and Protein Analysis
One frozen biopsy specimen per heart was divided into 10 slices, and the latter were transferred onto specimen holders on dry ice and covered by PBS. Semithin sections (500 nm) were cut from each slice (Ultracut T equipped with cryo-attachment; Leica). Sections were prepared for TEM and then coded. In each block, the upper right and the lower left quadrant and in each quadrant, two fields were analyzed. Ten capillaries were studied per field, and the average number of degenerated ECs was calculated for each block. Blocks containing 20% to 30% degenerated ECs were selected. Consecutive sections from these selected blocks were pooled for each patient and used for molecular analyses.
RNA was isolated and PCR was performed with the use of the FastStart DNA Master SYBR Green kit (Roche).5 The primer sequences (sense/antisense) were VE-cadherin: 5′-AACTTCCCCTTCTTCACCC-3′/5′-AAAGGCTGCTGGAAAATG-3′; β-catenin: 5′-CAAGTGGGTGG-TATAGAGG-3′/5′-CAATGGGAGAATAAAGCAG-3′; Ang-1: 5′-TTCAGAACCACACGGCTAC-3′/5′-TTCCTCTCTTTTTCCTCCC-3′; Ang-2: 5′-TGCAGCTACACTTTCCTC-3′/5′-TCTATCATCACAGCC-GTC-3′; Tie-2: 5′-TTCCAACATTACACACTCC-3′/5′-GCTATAAGC-AGCATCTTCC-3′; CD-31: 5′-GCTCTCTTGATCATTGCG-3′/5′-GAGGACACTTGAACTTCC-3′; GAPDH: 5′-AAATCCCATCA-CCATCTT-3′/5′-TTCCACGATACCAAAGTT-3′. LCDA Version 3.1.102 was used for PCR data analysis (Roche). The specificity of the amplification product was determined by performing a melting curve analysis. Standard curves for expression of each gene were generated by serial dilution of known quantities of the respective cDNA gene template. Relative quantification of the signals was done by normalizing the signals of the different genes with the GAPDH signal. Then, VE-cadherin and Tie-2 expression levels were normalized to the EC density with the use of CD-31.
Tissue lysates (50 μg/lane) were separated by SDS-PAGE before electrophoretic transfer onto Hybond C Super (Amersham).5 The blots were probed with antibodies against VE-cadherin (polyclonal; Bender MedSystems), β-catenin (monoclonal; Calbiochem), Ang-1 (polyclonal, Santa Cruz Biotechnology), Ang-2 (monoclonal, Calbiochem), Tie-2 (polyclonal; Santa Cruz Biotechnology), and CD-31 (monoclonal, TCS) before incubation with horseradish peroxidase–conjugated secondary antibodies (Amersham). Proteins were immunodetected with the use of chemiluminescence (Supersignal-West-Femto, Pierce), and protein bands were quantified by Easy Plus Win 32 software (Herolab). VE-cadherin and Tie-2 expression levels were normalized to the EC density with CD-31.
EC Culture and Gene Silencing
HDMVECs were cultured in EGM2 medium (Clonetics) plus 10% FBS and EC growth supplement (50 μg/mL), penicillin (100 U/mL), and streptomycin (100 μg/mL). Dermal tissue was treated with trypsin and passed through a nylon mesh, and HDMVECs were purified with CD-31-coated Dynabeads (Dynal).17 Cells were grown (37°C, 5% CO2) to 80% to 90% confluency, and the third passage was grown to 60% confluency for experiments. Cells were grown in EGM-2 medium without VEGF (2 hours), then stimulated with 100 ng/mL recombinant VEGF (Promokine) at time intervals. Cells were transfected with green fluorescent protein (GFP; Clontech) and 20 pmol/mL VE-cadherin double-stranded small interfering (si)RNAs with LipofectAMINE and Plus Reagent (Invitrogen), incubated for 48 hours, followed by RNA isolation. The siRNAs used were VE-cadherin: sense/antisense: 5′-AGAUGCAGAGGCUCAUGAUTT-3′/5′-AUCAUGAGCCUCUGCAUCUTT-3′. For annealing of siRNAs, 20 μmol/L single strands were incubated in annealing buffer (100 mmol/L potassium acetate, 30 mmol/L HEPES-KOH at pH 7.4, 2 mmol/L magnesium acetate) for 1 minute at 90°C and 1 hour at 37°C. Experiments were performed in triplicate.
Freshly isolated myocardium (2 to 3 mm3) was incubated in DMEM containing 10% FCS (100 pieces per 6-well plate). After 6 hours, wells were treated with 50 pmol/mL siRNA against VE-cadherin or stimulated with 100 ng/mL VEGF (Promokine) followed by incubation for 24 hours and RNA isolation. Experiments were performed in triplicate.
Cryosections were incubated with 1.5% goat serum to block nonspecific binding. Sections were incubated with a cocktail containing primary antibodies against VE-cadherin (rabbit polyclonal, 1:50 dilution, Bender MedSystems) and CD-31 (rat monoclonal, 1:100 dilution, RDI) for 2 hours. A cocktail containing Texas Red–anti-rabbit IgG and fluorescein-anti-rat-IgG (Vector, 1:50 dilution) detected positive staining. Additional cryosections were fixed with acetone and incubated with the Ang-2 primary antibody (monoclonal, Calbiochem). A secondary anti-mouse antibody (DAKO EnVision) and a biotinylated anti-goat antibody (Vector Labs, dilution 1:200) detected Ang-2 positive staining. The ABC peroxidase kit (Vector Labs) was used to detect biotin; diaminobenzidine was applied to develop color. Images were generated on a microscope (Leica), with a Nikon Coolpix 5000 digital camera.
At the end of the study, the key was broken and the expression levels of molecules and EC morphology were compared between the groups by means of ANOVA. Data are expressed as mean values, and statistical significance was set at P<0.05.
Cardiac VE-Cadherin/β-Catenin Expression Is Selectively Downregulated in DCM
Gene profiling in arrays indicated lower expression levels for VE-cadherin/β-catenin in cardiac biopsy specimens of patients with DCM but not ICM versus control subjects (Figure 1, A), whereas Ang-1 gene expression was not markedly changed (not shown). After this gene profiling, the expression of target molecules was studied in pooled tissue sections with similar rates of EC degeneration in both patient groups. VE-cadherin/β-catenin mRNA and protein expression was significantly downregulated in cardiac biopsy specimens of patients with DCM versus control subjects (P<0.05), whereas their expression levels were not significantly changed in ICM (Figure 1, B and C). Correlating to the expression profiles in tissue homogenates, immunocytochemistry of double-stained tissue sections showed that VE-cadherin but not CD-31 expression was discontinuous in myocardial vascular ECs of patients with DCM, whereas both VE-cadherin and CD-31 were expressed equally strong in the entire endothelial lining of myocardial vessels in control subjects (Figure 1, D and E). These studies indicated that VE-cadherin/β-catenin expression is blunted in DCM.
Ang-2 Expression Is Declined in DCM
In both DCM and ICM, Ang-1 expression was unchanged. Whereas Ang-2 mRNA and protein expression was downregulated in DCM, in ICM only, Ang-2 protein expression was decreased versus nonfailing hearts (Figure 2). At the receptor level, Tie-2 mRNA expression was not significantly changed in all patients, whereas its protein expression was upregulated in both DCM and ICM (P<0.05; Figure 2, A and B). These data indicated that only Ang-2 expression is affected in DCM.
Since VEGF-A is downregulated in DCM,5 the following studies should indicate whether VE-cadherin/β-catenin and Ang-2 downregulation is VEGF-dependent.
VE-Cadherin/β-Catenin Expression Is Independent of VEGF
Silencing of the VE-cadherin gene neither affected VEGF-A, β-catenin, Ang-2, nor EC-marker CD-31 expression; VEGF-A stimulation did not affect VE-cadherin, β-catenin, and CD-31 expression, whereas Ang-2 expression was upregulated in DCM cardiac cultures (Figure 3A). Similarly, VEGF-A deprivation or stimulation in cultured ECs did not affect VE-cadherin or CD-31 expression (Figure 3B). These studies indicated that blunted VE-cadherin/β-catenin but not Ang-2 expression in DCM is independent of VEGF. The following experiments were performed to show the effect of blunted VE-cadherin in ECs. Gene silencing was selected to reduce but not completely block gene expression, similar to DCM in vivo conditions.
VE-Cadherin Gene Silencing Results in Degeneration of Human ECs
Cotransfection of GFP with VE-cadherin siRNA visualized transfection rate of cells. After transfection of HDMVECs with VE-cadherin siRNA, gene expression of the target molecule decreased significantly (Figure 3C). This resulted in formation of cytoplasmic vesicles and projections and disruption of the EC monolayer (Figure 3D), indicating the contribution of blunted VE-cadherin to EC degeneration.
Myocardial Vessels Have a High Rate of EC Degeneration in DCM
Degenerated ECs had discontinuous or fringed membranes, degenerated mitochondria and vesicles, and discontinuous nuclear membranes both in DCM and ICM (Figure 4A). By quantitative analysis, however, the rate of EC degeneration was significantly higher in cardiac biopsy specimens of patients with DCM versus patients with ICM. Although 30% of ECs and 31% of their nuclei were degenerated in DCM (not significant versus normal ECs within the DCM group), 20% of ECs and 10% of their nuclei (P<0.05 versus normal ECs within the ICM group) were degenerated in ICM (Figure 4B). These studies indicated that myocardial ECs have a high rate of degeneration in DCM.
We describe a VE-cadherin/β-catenin and Ang-2 downregulation in DCM. Accumulating evidence suggests that KDR transmits an EC survival signal of VEGF-A through PI3-kinase, Akt kinase, and Bcl-2.7 This signaling also involves VE-cadherin as a deficiency of VE-cadherin induced endothelial apoptosis and abolished transmission of the EC survival signal by VEGF-A to Akt kinase and Bcl-2 through reduced complex formation with KDR, β-catenin, and PI3-kinase.8 Our results showed degenerated ECs after VE-cadherin downregulation by RNAi in vitro; antibody studies in mice suggest that VE-cadherin exerts a relevant and specific activity in the maintenance of vascular integrity.18 The stabilization of the adherens-type junctions by VE-cadherin depends on the interaction of its tail with the cytoskeleton.10,11,13 Thus, the molecular pathology observed for VE-cadherin/β-catenin in DCM may affect EC survival and the stability of endothelial cell-cell contacts,19 as indicated by increased number of degenerated ECs, damage of EC membranes, and the resulting decreased number of patent adherens junctions.10
Downregulated VE-cadherin/β-catenin observed in our study appears to be selective for DCM and VEGF-independent for the following reasons. First, when tissue sections with a similar rate of EC degeneration were compared between ICM and DCM, VE-cadherin/β-catenin was blunted only in DCM. Second, VEGF-A deprivation and stimulation in cultured myocardium and HDMVECs did not change VE-cadherin/β-catenin expression. Third, VEGF-A expression was not affected by RNAi-induced VE-cadherin downregulation. Fourth, in ICM, VEGF-A expression is upregulated, although VE-cadherin/β-catenin is unchanged.
During angiogenesis, β-catenin is translocated from the plasma membrane to the cytoplasm, where it causes EC proliferation.19 Additionally, β-catenin and plakoglobin bind to the intracellular domain of the cadherins, are connected to the cytoskeleton through α-catenin; and the actin-based cytoskeleton is connected to the sarcomeric structure and the external basement membrane. Dystrophin, a cytoskeletal protein residing at the inner face of the sarcolemma, plays a critical role in the establishment of connections to the sarcolemma.20 Therefore, the possibility exists that catenin interacts with cytoskeletal proteins, for example, dystrophin. This interplay might constitute a link between the EC- and muscular20-related pathology and be in concert with the hypothesis that mutations of cytoskeletal protein genes could be a common pathway in DCM, at least in its familial form.20
The Ang family that mediates EC survival binds exclusively to the Tie-2 receptor tyrosine kinase, and the Tie receptors are required for the angiogenic remodeling and vessel stabilization that occur subsequent to the initial vasculogenic actions of Flt-1 and KDR.21,22 Ang-1 phosphorylates tyrosine in Tie-2, is chemotactic for ECs, induces sprouting, and potentiates VEGF.23 Ang-1 can stabilize ECs,24and low levels of phosphorylated Tie-2 have been detected in the adult quiescent vasculature, indicating involvement of Tie-2 in vascular maintenance.6 Ang-1 can also facilitate PI3-kinase–dependent EC survival through stimulation of Akt6,25 in addition to the survival factor VEGF; the addition of VEGF can augment the antiapoptotic effect of Ang-1 in ECs.24 Thus, ECs appear to have evolved to respond to distinct survival cues, which might reflect the expression levels of Ang-1, VEGF, and VE-cadherin in particular vascular beds. Although Ang-1 functions as the activating ligand for Tie-2, Ang-2 is a context-dependent, competitive antagonist.26 In adults, Ang-2 is expressed primarily at sites of vascular remodeling, where it is thought to block the constitutive stabilizing or maturing action of Ang-1.26 We found unchanged Ang-1, reduced Ang-2, and increased Tie-2 expression in both DCM and ICM, indicating that Ang-1 signaling is not affected. Therefore, upregulation of Tie-2 in DCM might be a reaction of ECs to compensate for downregulated VE-cadherin–mediated survival and stabilization signals. In addition, increased Tie-2 levels in ICM might have been reactive to chronic hypoxia.27 Ang-2 might allow the vessels to revert to and remain in a more plastic state, in which they may be more responsive to a sprouting signal provided by VEGF.26 Whether downregulated Ang-2 expression is adaptive to blunted VEGF expression5 remains uncertain. The fact that destabilization by Ang-2 in the absence of VEGF leads to vessel regression, whereas such destabilization in the presence of VEGF facilitates the angiogenic response,26 might favor the adaptation hypothesis.
Together, it appears that the EC survival is impaired in myocardium of patients with DCM involving VE-cadherin/β-catenin, independent of VEGF. Targeting VE-cadherin might be of benefit to counteract the EC pathology in DCM.
This study was supported by the Austrian National Bank, grant 8823 (Dr Aharinejad).
Swynghedauw B. Molecular mechanisms of myocardial remodeling. Physiol Rev. 1999; 79: 215–262.
Abraham D, Hofbauer R, Schaefer R, et al. Selective downregulation of VEGF-A(165), VEGF-R(1), and decreased capillary density in patients with dilative but not ischemic cardiomyopathy. Circ Res. 2000; 87: 644–647.
Gerber HP, McMurtrey A, Kowalski J, et al. Vascular endothelial growth factor regulates EC survival through the phosphatidylinositol 3′-kinase/Akt signal transduction pathway: requirement for Flk-1/KDR activation. J Biol Chem. 1998; 273: 30336–30343.
Kim I, Kim HG, So JN, et al. Angiopoietin-1 regulates endothelial cell survival through the phosphatidylinositol 3′-Kinase/Akt signal transduction pathway. Circ Res. 2000; 86: 24–29.
Esser S, Lampugnani MG, Corada M, et al. Vascular endothelial growth factor induces VE-cadherin tyrosine phosphorylation in endothelial cells. J Cell Sci. 1998; 111: 1853–1865.
Müllner EW, Garcia-Sanz JA. Preparation of RNA: Manual of Immunological Methods. Vol 1. London: Academic Press; 1997: 339–406.
Zoellner H, Hofler M, Beckmann R, et al. Serum albumin is a specific inhibitor of apoptosis in human endothelial cells. J Cell Sci. 1996; 109: 2571–2580.
Corada M, Mariotti M, Thurston G, et al. Vascular endothelial-cadherin is an important determinant of microvascular integrity in vivo. Proc Natl Acad Sci U S A. 1999; 96: 9815–9820.
van Gijn ME, Daemen MJAP, Smits JFM, et al. The wnt-frizzled cascade in cardiovascular disease. Cardiovasc Res. 2002; 55: 16–24.
Gale NW, Yancopoulos GD. Growth factors acting via endothelial cell-specific receptor tyrosine kinases: VEGFs, angiopoietins, and ephrins in vascular development. Genes Dev. 1999; 13: 1055–1066.
Maisonpierre PC, Suri C, Jones PF, et al. Angiopoietin-2, a natural antagonist for Tie2 that disrupts in vivo angiogenesis. Science. 1997; 277: 55–60.
Willam C, Koehne P, Jurgensen JS, et al. Tie2 receptor is stimulated by hypoxia and proinflammatory cytokines in human endothelial cells. Circ Res. 2000; 87: 370–377.