Ephrin type B receptor 4 (EphB4, Eph receptor) selectively binds ephrin B2 (Eph ligand). EphB4/ephrin B2 is involved in embryonic vessel development, vascular remodelling and pathological vessel formation in adults (including tumour angiogenesis). Binding of vascular endothelial growth factor (VEGF)-A to the endothelial-specific receptor VEGF receptor-2 is the main extracellular signal triggering angiogenic response. Little is known about the role of EphB4/ephrin B2 during angiogenesis and arteriovenous plasticity in cerebral arteriovenous malformation (cAVM). This study investigated EphB4 and ephrin B2 expression in cAVM.
Methods
Haemorrhagic (H-AVM) and nonhaemorrhagic (NH-AVM) specimens of AVM nidus, obtained after microsurgical cAVM resection, and normal superficial temporal artery (STA) specimens, were analysed retrospectively. VEGF-A, EphB4 and ephrin B2 expression were studied by immunohistochemistry and immunoblotting.
Results
In cAVM (10 H-AVM; 10 NH-AVM), VEGF-A was immunocytochemically localized to endothelial cells; strong endothelial cell staining was found for EphB4 in veins and ephrin B2 in arteries. Normal STA (n = 10) did not express EphB4 or ephrin B2. EphB4 and ephrin B2 expression was greater in H-AVM than in NH-AVM.
Conclusions
Endothelial cells are more active in H-AVM than NH-AVM. EphB4 and ephrin B2 play important roles in neovascularization and arteriovenous differentiation/plasticity. These data provide new insights into the aetiology of cAVM and lay a foundation for further study. The notch pathway induced by VEGF-A may be a key signalling pathway in this process.
Introduction
Ephrin type B receptor 4 (EphB4) belongs to the receptor tyrosine kinase (RTK) family. RTKs are a diverse group of transmembrane proteins that, on receiving an external stimulus, respond by transmitting a signal to the inside of the cell and thus control the cell’s shape, proliferation, differentiation and migration.1,2 RTKs have key roles in normal physiology and in oncogenesis.3 EphB4 selectively binds ephrin B2 and no other ephrin B ligands.4 Unlike other families of RTKs, which bind soluble ligands, Eph receptors interact with ephrin ligands bound to the cell surface; these ligands can transmit signals on receptor engagement in a phenomenon known as bidirectional signalling. Ephrin B ligands, which possess a transmembrane and a cytoplasmic domain, are tyrosine phosphorylated on receptor binding, which induces ephrin reverse signalling in addition to receptor forward signalling.5 This type of signalling establishes the mechanistic basis of junction formation between arterial and venous endothelial cells through the restriction of cellular intermingling.6,7 EphB4 and its cognate ligand ephrin B2 play important roles in embryonic vessel development and vascular remodelling.8
Cerebral arteriovenous malformations (cAVMs) comprise tangles of abnormally formed blood vessels, and are an important cause of intracranial haemorrhage and seizure in young adults.9 They are histopathologically characterized by the presence of a nidus, consisting of a conglomerate of tortuous vessels lacking capillaries, an arterial feeder and a draining vein.10 cAVMs have long been thought to arise from developmental derangements (at the embryonic, fetal or early postnatal stage of vessel formation), whereby abnormal vascular formation results in continued vascular remodelling due to a chronic increase in blood flow during postnatal growth.11–13 Their course cannot easily be predicted; they may remain static, grow or even regress. Angiogenesis and arteriovenous plasticity therefore play important roles in the formation and development of cAVMs. Several reports suggest that EphB4/ephrin B2 is also involved in pathological vessel formation in adults, including tumour angiogenesis,14–16 but little is known about its role during angiogenesis or in arteriovenous plasticity in cAVM. The purpose of this retrospective study was to explore the expression of EphB4 and ephrin B2 in cAVM.
Patients and methods
This study was approved by the Ethics Committee of Beijing Tiantan Hospital, Capital Medical University, Beijing. All individuals providing data for use in this study were registered as inpatients of Beijing Tiantan Hospital. Written consent to obtain clinical specimens and to publish data obtained in the study was obtained from the participants, or from the guardians of participants who were <18 years of age.
Patients
The study retrospectively included samples from patients with cAVM, with or without haemorrhage, admitted to the Department of Neurosurgery at Beijing Tiantan Hospital between January 2011 and July 2011; the patients included both emergency admissions and routine admissions. The samples and data included in this study were randomly selected for inclusion using a computer program. Digital subtraction angiography, computed tomography and magnetic resonance imaging were used for evaluation before and after surgery, using standard techniques. The Spetzler–Martin grading system (www.neurosurgic.com) for arteriovenous malformations was used to confirm that the surgical risk for participating patients with cAVM was acceptable. All patients underwent surgical cAVM resection, which was a routine aspect of their care and was undertaken using standard techniques. After surgery, the diagnosis of cAVM was confirmed pathologically. Also included in the study were head trauma patients with structurally normal superficial temporal arteries (STAs), who served as controls; the controls were not matched to the treatment group. Head trauma patients with cerebrovascular diseases, tumours or other systemic diseases were excluded.
Tissue specimens
Tissue specimens were obtained from each patient after microsurgical cAVM resection. The cAVM nidus was dissected away from adjacent brain tissue in the operating room, and a representative portion of malformed AVM nidal tissue was immediately placed in an aluminium container and submerged in liquid nitrogen. Frozen tissues were stored at −80℃ until analysis, which was undertaken ∼2 months after excision. Specimens of structurally normal STA tissue were obtained from the control patients and were collected, frozen and stored in the same manner as the cAVM specimens.
Immunohistochemistry
The frozen tissue samples were cut into 5–6 µm-thick sections for haematoxylin and eosin staining, and for immunohistochemical staining for EphB4, ephrin B2 and vascular endothelial growth factor (VEGF). After holding for 20 min at room temperature, frozen sections were fixed in standard acetone for 10 min. Procedures for immunohistochemistry were as described elsewhere.17,18 Briefly, anti-EphB4 antibody (rabbit monoclonal, dilution 1:150; Santa Cruz Biotechnology, CA, USA), anti-ephrin B2 antibody (mouse polyclonal, dilution 1:300; Santa Cruz Biotechnology) and anti-VEGF-A antibody (mouse polyclonal, dilution 1:500; Santa Cruz Biotechnology) were used for immunohistochemical staining; antirabbit and antimouse secondary antibodies were used to detect these antibodies. Tissue sections were counterstained with nuclear fast blue. Microscopy examinations were undertaken with a DMI4000 microscope (Leica, Wetzlar, Germany); magnifications ranged between ×50 and ×200, and 10 scopes were visualized in every tissue section; the average (mean ± SD) was then calculated for each sample. Immunohistochemistry results were recorded as scores for the percentage of cells staining positively, and scores for staining intensity. The percentage of cells staining positively was scored 0 (≤10% positive cells), 1 (11–25%), 2 (26–50%), 3 (51–75%) or 4 (>75%). Staining intensity was scored 0 (negative), 1 (weak), 2 (medium) or 3 (strong). The product of the two scores was regarded as the total score for protein expression; total scores were categorized as − (score 0), + (<6), or ++ (≥6).
Immunoblotting assays
Equal amounts (10 µg) of protein from tissue lysates were separated by sodium dodecyl sulphate–polyacrylamide gel electrophoresis, transferred onto nitrocellulose membranes and probed with primary and secondary antibodies. Additional antibodies used were anti-EphB4 antibody (rabbit monoclonal; Santa Cruz Biotechnology; dilution 1:500) and anti-ephrin B2 antibody (mouse polyclonal; Santa Cruz Biotechnology; 1:500). Band intensity was analysed using a Kodak® Gel Logic imaging system (Kodak Scientific Imaging Systems, Rochester, NY, USA) and Quantity One® (Bio-Rad, Hercules, CA, USA) software. β-Actin (dilution 1:1500; Cell Signaling Technology, Danvers, MA, USA) was used as a control.
Statistical analyses
All results were expressed as mean ± SD. SPSS® statistical software, version 13.0 (SPSS Inc. Chicago. IL., USA) was used for analyses. The Kolmogorov–Smirnov test was used to test for normal distribution of all quantitative data; χ2-test was used for comparisons between two sets of data. Single-factor analysis of variance and the Kruskal–Wallis test were used for comparisons among three sets of data. P < 0.05 was considered statistically significant.
Results
This study included data from 20 patients (15 male and five female) with cAVM; their mean age was 28.3 years (range, 8–59 years). The first presentation was with haemorrhage in 10 patients and without haemorrhage in the remaining 10 patients; two patients haemorrhaged twice and three patients underwent radiosurgery before cAVM resection (Table 1). Also included in the study were control data from 20 patients (17 male and three female; mean age, 32.5 years [range, 12–56 years]), each of whom had head trauma with a structurally normal STA.
Table 1. Clinical data for patients with cerebral arteriovenous malformation (cAVM) who underwent microsurgical cAVM resection, included in a study investigating EphB4 and ephrin B2 expression in cAVM.
Range of grades, I–VI, correlating with likely operative outcome; grade VI represents an inoperable lesion (www.neurosurgic.com).
Haematoxylin and eosin staining of a frozen cAVM section showed partial fragmentation of the endothelial layer, vascular stenosis and deformation and endothelial cell hypertrophy (Figure 1A). Immunocytochemical analysis of VEGF-A in cAVM confirmed the localization of VEGF-A to endothelial cells (Figure 1B). All cAVM specimens revealed strong staining for EphB4 and ephrin B2 in vascular endothelial cells; EphB4 was located in the venous endothelium (Figure 1C, D), whereas ephrin B2 was located in the arterial endothelium (Figure 1E, F). Immunocytochemical analysis of EphB4 in frozen cAVM sections confirmed that EphB4 was expressed in mature and immature vessels (Figure 1D). Neither EphB4 nor ephrin B2 was expressed in normal STA (Figure 1G, H); the brown staining in figure 1G is not true staining in the vascular endothelial cell.
Figure 1. Frozen sections of cerebral arteriovenous malformation (cAVM) tissue taken from patients with the condition, and normal superficial temporal artery (STA) taken from control (head trauma) patients who had no other systemic diseases. (A) Haematoxylin and eosin staining of a section of a cAVM. There was partial fragmentation of the endothelial layer, vascular stenosis and deformation and endothelial cell hypertrophy (arrowheads). Original magnification, ×50. (B–H) Immunocytochemical analysis. (B) Vascular endothelial growth factor-A was localized to endothelial cells in cAVM (arrowheads). Original magnification, ×400. (C) EphB4 was localized to the endothelial layer of a cAVM vein (arrowhead). Original magnification, ×100. (D) In a section of a cAVM, EphB4 was expressed in an immature vessel (arrowhead). Original magnification, ×100. (E) EphrinB2 was localized to endothelial cells in a cAVM artery (arrowheads). Original magnification, ×400. (F) Ephrin B2 was localized to the endothelial layer in a cAVM section (arrowhead). Original magnification, ×100. (G) EphB4 was not expressed in normal STA. Original magnification, ×100; the brown staining shown is not true staining of the vasculare endothelial cell (H) Ephrin B2 was not expressed in normal STA. Original magnification, ×100. The colour version of this figure is available at: http://imr.sagepub.com.
In immunoblotting assays (Figure 2, a–c), expression of EphB4 was greater in both haemorrhagic cAVM (H-AVM) and nonhaemorrhagic cAVM (NH-AVM) than in normal STA. EphB4 expression was greater in H-AVM than in NH-AVM; band intensity percentages with respect to the control (β-actin) differed significantly among groups (STA, 15.51 ± 2.11%; NH-AVM, 48.75 ± 7.16%; H-AVM, 59.04 ± 9.74%, P = 0.000) (Figure 2b). Expression of ephrin B2 was greater in both H-AVM and NH-AVM than in normal STA, and was greater in H-AVM than in NH-AVM; band intensity percentages differed significantly among groups (STA, 21.02 ± 3.24%; NH-AVM, 58.19 ± 6.91%; H-AVM, 70.75 ± 7.81%; P = 0.000) (Figure 2c). In immunoblotting assays (Figure 3, a and b), expression of VEGF-A was greater in both H-AVM and NH-AVM than in normal STA, and was greater in H-AVM than in NH-AVM. VEGF-A band intensity relative to β-actin differed significantly among groups (STA, 76.42 ± 12.89%; NH-AVM, 135.07 ± 18.41%; H-AVM, 166.72 ± 16.22%; P = 0.002) (Figure 3b).
Figure 2. (a) Immunoblotting assays for ephrin B2 and EphB4, performed in material collected from malformed vessels; (b, c) band intensity ratio as a percentage of β-actin intensity, mean ± SD; (b) EphB4. Analysis of variance, F = 16.83, P = 0.000. (c) Ephrin B2. Kruskal–Wallis test, χ2 = 15.834, P = 0.000. β-actin served as a control. STA, normal superficial temporal artery; NH-AVM, nonhaemorrhagic cerebral arteriovenous malformation; H-AVM, haemorrhagic cerebral arteriovenous malformation.
Figure 3. (a) Immunoblotting assays for vascular endothelial growth factor (VEGF)-A, performed in material collected from malformed vessels. (b) band intensity ratio as a percentage of β-actin intensity, mean ± SD; Statistical analysis was by single-factor analysis of variance (P = 0.002). β-actin served as a control. STA, normal superficial temporal artery; NH-AVM, nonhaemorrhagic cerebral arteriovenous malformation; H-AVM, haemorrhagic cerebral arteriovenous malformation.
Discussion
Angiogenesis involves a process in which trophoblastic blood vessels send out capillary sprouts to form new tube-like structures.9 This complex, multistep process begins with an increase in local or systemic angiogenic factors and is followed by breakdown of endothelial basement membrane, to facilitate endothelial cell migration and proliferation.9 Binding of VEGF-A to the endothelial-specific VEGF receptor-2 appears to be the main extracellular signal triggering the angiogenic response.9 Research characterizing cerebrovascular lesions has demonstrated increased expression of angiogenic factors, such as VEGF-A, at the protein and mRNA levels.19,20 In such lesions, normal adult cerebral vascular growth is stopped, endothelial cell proliferation and angiogenesis are in a relatively inhibited state and VEGF-A expression is low.21 Such findings may indicate that in cAVM angiogenesis is in an active state, especially in H-AVM, and that endothelial cell proliferation and angiogenesis are more active, which may also explain the causes of haemorrhage. A cAVM is a complex tangle of abnormal arteries and veins linked by one or more fistulae: cAVMs lack a capillary bed and the small arteries have a deficient muscularis layer. Such characteristics indicate that abnormal arteriovenous plasticity may be the main reason for the formation of cAVMs.
Ephrin B2 and EphB4 have been regarded as the primary marker molecules for arteriovenous differentiation, with ephrin B2 expressed exclusively by arterial endothelial cells and EphB4 by venous endothelial cells.8,22,23 This interpretation is reinforced by growing but not yet conclusive evidence that vessel marker proteins have roles other than establishing vessel identity in vessel sprouting (during vessel remodelling and angiogenesis), such as the modulation of cell signalling and the direction of cell-tip filopodial extension.24–26 Indeed, ephrin B2 can act as a proangiogenic factor in postnatal neovascularization, since it is capable of inducing neovascularization when delivered in a corneal micropocket assay.27
The present findings suggest extensive and ongoing roles for EphB4 and ephrin B2 in cAVM. We found that high ephrin B2 expression continued to mark arteries selectively in cAVM, but that there was no expression in normal STA; in line with other research discussed above,21 we postulate that EphB2 did not stain in STA because these were mature, normal vessels, whereas in cAVM EphB2 staining was observed because the vessels were malformed and/or immature. Furthermore, we found ephrin B2 expression only in the arterial endothelium, not in the surrounding arterial smooth muscle cells and pericytes. This conflicts with a report that ephrin B2 remained highly expressed in the endothelium and smooth muscle of adult arterial vessels.28 We also found that high EphB4 expression continued to mark venous endothelium selectively in cAVM, but was not expressed in normal STA. Together, these findings suggest that EphB4 and ephrin B2 might be involved, in an ongoing fashion, in regulating vascular endothelial formation, and arteriovenous differentiation and plasticity. Furthermore, in our view, selective expression of EphB4 and ephrin B2 in veins and arteries suggests that ligand–receptor pairs can act together, leading to abnormal angiogenesis in cAVM in the embryo and adult.
Molecules with expression restricted to either arterial or venous endothelial cells are referred to as arteriovenous markers, and play critical roles in determining the corresponding endothelial cell lineages during embryonic development.29 The expression level of the arterial endothelial cell marker ephrin B2 seems to be dependent on extracellular determinants, as it is up-regulated by VEGF-A and exposure to hypoxia.30,31 We speculate that hypoxia of the cerebral regional area due to artery–vein fistulae in cAVM stimulates VEGF-A production, which upregulates expression of ephrin B2. In the embryo, induction of arteriovenous-specific marker expression occurs when VEGF-A signals via the VEGF receptor-2/neuropilin-1 complex in arterial-fated angioblasts, leading to the downstream activation of notch and ERK signalling pathways, and the subsequent expression of the arterial marker ephrin B2. There is some evidence that transcription factors can inhibit the notch signalling pathway in veins, inducing the expression of the venous marker EphB4.32–36 The appropriate expression of EphB4 and ephrin B2 has been determined by Delta-like ligand (DLL)-1, a notch ligand and an upstream factor regulating embryonic artery differentiation and postnatal angiogenesis.37 Hainaud et al.38 found that upregulation of notch ligand DLL4 by VEGF-A could increase the activity of notch 4, leading to upregulation of ephrin B2 and downregulation of EphB4.37,38 The activation of notch 4 is necessary for the upregulation of ephrin B2 by VEGF-A and vascular endothelial cell differentiation.
The other notable finding of the present study was that expression of VEGF-A, EphB4 and ephrin B2 was higher in H-AVM than in NH-AVM, which suggests that endothelial cells are more active in H-AVM than in NH-AVM.
Vascular endothelial growth factor-induced cell signalling has been implicated in arteriovenous specification during embryonic development,39 suggesting that VEGF-A plays a role in arteriovenous specification in adult neovascularization. However, other studies showed that VEGF-A is not responsible for angiogenic sprouting,40,41 suggesting a different mechanism for the regulation of ephrin B2 and EphB4 expression, and arteriovenous identity specification. Arteriovenous specification in adult neovascularization may be driven by haemodynamic forces and perivascular cell activity that is associated with specific needs of the tissue, rather than by predetermined programming. To our knowledge, haemodynamic forces are the main reasons for haemorrhage in cerebral AVM, which could explain the higher expression of VEGF-A, EphB4 and ephrin B2 in H-AVM. Hypoxia due to regional haemorrhage also stimulates VEGF-A production and promotes neovascularization.42,43
In conclusion, cAVMs have long been thought to arise from abnormal vascular formation during embryonic development. The course of cAVMs cannot easily be predicted; they may remain static, grow or regress. Neovascularization and arteriovenous specification are the most important molecular events in cAVM. High expression of EphB4 and ephrin B2 continues to mark arteries and veins selectively in cAVM, and plays an important role in neovascularization and arteriovenous specification. The notch pathway induced by VEGF-A may be a key signalling pathway. Understanding the molecular mechanism underlying neovascularization, differentiation and arteriovenous specification is important for the development of future therapeutic approaches and for the prediction of bleeding in cAVM. Expression of VEGF-A in cAVM was the first step that we studied. We now plan to conduct studies with proliferation markers, notch pathway markers and with primary cultures defined from human specimens, to elucidate more about this key area of clinical research.
Declaration of conflicting interest
The authors had no conflicts of interest to declare in relation to this article.
Funding
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
References
1. Lamorte L, Park M. The receptor tyrosine kinases: role in cancer progression. Surg Oncol Clin N Am 2001; 10: 271–288.
7. Hamada K, Oike Y, Ito Y, et al. Distinct roles of ephrin-B2 forward and EphB4 reverse signaling in endothelial cells. Arterioscler Thromb Vasc Biol 2003; 23: 190–197.
8. Wang HU, Chen ZF, Anderson DJ. Molecular distinction and angiogenic interaction between embryonic arteries and veins revealed by ephrin-B2 and its receptor Eph-B4. Cell 1998; 93: 741–753.
9. Alkadhi H, Kol1ias SS, Crelier GR, et al. Plasticity of the human motor cortex in patients with arteriovenous malformations: a functional MR imaging study. AJNR Am J Neuroradiol 2000; 21: 1423–1433.
12. Hatva E, Jääskeläinen J, Hirvonen H, et al. Tie endothelial cell-specific receptor tyrosine kinase is upregulated in the vasculature of arteriovenous malformations. J Neuropathol Exp Neurol 1996; 55: 1124–1133.
13. Lawton MT, Jacobowitz R, Spetzler RF. Redefined role of angiogenesis in the pathogenesis of dural arteriovenous malformations. J Neurosurg 1997; 87: 267–274.
16. Miyazaki T, Kato H, Fukuchi M, et al. EphA2 overexpression correlates with poor prognosis in esophageal squamous cell carcinoma. Int J Cancer 2003; 103: 657–663.
18. Hamada K, Oike Y, Takakura N, et al. VEGF-C signaling pathways through VEGFR-2 and VEGFR-3 in vasculoangiogenesis and hematopoiesis. Blood 2000; 96: 3793–3800.
22. Oike Y, Ito Y, Hamada K, et al. Regulation of vasculogenesis and angiogenesis by EphB/ephrin-B2 signaling between endothelial cells and surrounding mesenchymal cells. Blood 2002; 100: 1326–1333.
23. Gale NW, Baluk P, Pan L, et al. Ephrin-B2 selectively marks arterial vessels and neovascularization sites in the adult, with expression in both endothelial and smooth-muscle cells. Dev Biol 2001; 230: 151–160.
24. Taylor AC, Murfee WL, Peirce SM. EphB4 expression along adult rat microvascular networks: EphB4 is more than a venous specific marker. Microcirculation 2007; 14: 253–267.
28. Gale NW, Baluk P, Pan L, et al. Ephrin-B2 selectively marks arterial vessels and neovascularization sites in the adult, with expression in both endothelial and smooth-muscle cells. Dev Biol 2001; 230: 151–160.
32. Lawson ND, Vogel AM, Weinstein BM. Sonic hedgehog and vascular endothelial growth factor act upstream of the Notch pathway during arterial endothelial differentiation. Dev Cell 2002; 3: 127–136.
37. Kim YH, Hu H, Guevara-Gallardo S, et al. Artery and vein size is balanced by Notch and ephrin B2/EphB4 during angiogenesis. Development 2008; 135: 3755–3764.
38. Hainaud P, Contrerès JO, Villemain A, et al. The role of the vascular endothelial growth factor-Delta-like 4 ligand/Notch4-ephrin B2 cascade in tumor vessel remodeling and endothelial cell functions. Cancer Res 2006; 66: 8501–8510.
40. Nunes SS, Greer KA, Stiening CM, et al. Implanted microvessels progress through distinct neovascularization phenotypes. Microvasc Res 2010; 79: 10–20.
41. Chang CC, Nunes SS, Sibole SC, et al. Angiogenesis in a microvascular construct for transplantation depends on the method of chamber circulation. Tissue Eng Part A 2010; 16: 795–805.
42. Ng I, Tan WL, Ng PY, et al. Hypoxia inducible factor-1alpha and expression of vascular endothelial growth factor and its receptors in cerebral arteriovenous malformations. J Clin Neurosci 2005; 12: 794–799.
43. Sho E, Komatsu M, Sho M, et al. High flow drives vascular endothelial cell proliferation during flow-induced arterial remodeling associated with the expression of vascular endothelial growth factor. Exp Mol Pathol 2003; 75: 1–11.
This article is distributed under the terms of the Creative Commons Attribution-NonCommercial 3.0 License (http://www.creativecommons.org/licenses/by-nc/3.0/) which permits non-commercial use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access page(http://www.uk.sagepub.com/aboutus/openaccess.htm).
I can access personal subscriptions, purchases, paired institutional access and free tools such as favourite journals, email alerts and saved searches.
