Angiogenesis in the Outer Membrane of Chronic Subdural Hematomas through Thrombin-Cleaved Osteopontin and the Integrin α9 and Integrin β1 Signaling Pathways

Background: A chronic subdural hematoma (CSDH) is considered to be an inflammatory and angiogenic disease. The CSDH outer membrane, which contains inflammatory cells, plays an important role in CSDH development. Osteopontin (OPN) is an extracellular matrix protein that is cleaved by thrombin, generating the N-terminal half of OPN, which is prominently involved in integrin signal transduction. We explored the expression of the N-terminal half of OPN in CSDH fluid and the expression of integrins α9 and β1 and the downstream components of the angiogenic signaling pathways in the outer membrane of CSDHs. Methods: Twenty samples of CSDH fluid and eight samples of CSDH outer membrane were collected from patients suffering from CSDHs. The concentrations of the N-terminal half of OPN in CSDH fluid samples were measured using ELISA kits. The expression levels of integrins α9 and β1, vinculin, talin-1, focal adhesion kinase (FAK), paxillin, α-actin, Src and β-actin were examined by Western blot analysis. The expression levels of integrins α9 and β1, FAK and paxillin were also examined by immunohistochemistry. We investigated whether CSDH fluid could activate FAK in cultured endothelial cells in vitro. Results: The concentration of the N-terminal half of OPN in CSDH fluid was significantly higher than that in the serum. Western blot analysis confirmed the presence of these molecules. In addition, integrins α9 and β1, FAK and paxillin were localized in the endothelial cells of vessels within the CSDH outer membrane. FAK was significantly phosphorylated immediately after treatment with CSDH fluid. Conclusion: Our data suggest that the N-terminal half of OPN in CSDH fluid promotes neovascularization in endothelial cells through integrins α9 and β1. The N-terminal half of OPN, which is part of the extracellular matrix, plays a critical role in the promotion of CSDHs.


Introduction
Chronic subdural hematomas (CSDHs) occur in senior citizens who have suffered from mild head trauma. Episodes of mild head trauma are frequently unrecognized for elderly people. Computerized tomography (CT) scanning often reveals CSDHs. A CSDH is a neovascularized inflammatory disease. However, the pathogenesis of CSDHs has not been fully clarified. The dura is lined with a layer of connective tissue cells named "dural border cells", which can develop into fibro-cellular connective tissue [1]. After mild head trauma, laceration of dural border cells develops into the inner and outer membranes. diaminetetraacetate, 0.2 mM ethylene glycol bis(aminoethyl ether)tetraacetate, 0.2 mM phenylmethylsulfonyl fluoride, 1.25 µg/mL pepstatin A, 0.2 µg/mL aprotinin, 1 mM sodium orthovanadate, 50 nM sodium fluoride, 2 mM sodium pyrophosphate and 1% Nonidet P-40. The homogenates were centrifuged at 12,000× g for 10 min at 4 • C. The proteins in the supernatants were separated using 7.5% sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis.

Histological Examinations
For the analysis of cellular expression of integrin α9, integrin β1, FAK and paxillin, immunohistochemical staining was performed on samples from three out of 20 patients at room temperature using the avidin-biotinylated peroxidase complex (ABC) technique. To preserve the outer membrane of the CSDH samples, the membranes were incubated in 10 mL of ice-cold 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) for 3 h and then embedded in paraffin.
In this study, 10-µm-thick sections were prepared with a microtome and mounted onto MAS-coated glass slides (Matsunami Glass, Kishiwada, Japan). The sections were deparaffinized in xylene, hydrated through an ethanol gradient, and then fully rehydrated in water. Endogenous peroxidase activity was blocked with 0.3% H 2 O 2 in 100% methanol for 20 min. All sections for immunostaining were processed for microwave-enhanced antigen retrieval. Slide-mounted sections immersed in 0.01 M sodium citrate buffer (pH 6.0) were placed for 15 min in a 700 W microwave oven at maximum power.
Nonspecific immunoreactivity was blocked by incubation with goat or donkey serum for 30 min, depending on the primary antibody. The samples were treated with primary antibodies against integrin α9 (#MAB4574, R&D Systems) at a dilution of 1:150, integrin β1 (#34971, Cell Signaling Technology) at a dilution of 1:100, paxillin (#125891, Gene Tex, Gene Tex, Irvine, CA, USA) at a dilution of 1:500 and FAK (#3285, Cell Signaling Technology) at a dilution of 1:500 for 36 h at 4 • C. After several rinses in PBS, the samples were incubated with secondary biotinylated antibodies (anti-goat IgG 1:200, anti-rabbit IgG 1:200; Santa Cruz Biotechnology) at room temperature for 2 h. After several more rinses in PBS, the samples were incubated with Vectastain ABC reagent (Vectastain ABC Kit; Vector Laboratories, Burlingame, CA, USA) for 1 h. After several more rinses in PBS, the bound peroxidase was visualized by incubating the sections with a solution containing 0.05% 3,3 -diaminobenzidine tetrahydrochloride (Sigma Aldrich) and 0.01% H 2 O 2 in 0.05 M Tris-HCl (pH 7.4) for 10 min. After several rinses in water, the immunostained sections were dehydrated and cover-slipped with Entellan new (Merck, Kenilworth, NJ, USA).

Cultured Vascular Endothelial Cells
Endothelial cells of the mouse brain (b. End3) were obtained from the HPA Cultured Collection (London, UK). The endothelial cells were cultured in Dulbecco's modified Eagle's medium (Nissui, Tokyo, Japan) with 10% fetal bovine serum at 37 • C and 5% CO 2 .

Effect of CSDH Fluid on FAK and Extracellular Signal-Regulated Kinase (ERK)
We obtained CSDH fluid during trepanation surgery and centrifuged it to remove any debris. b.End3 cells were incubated with a serum-containing medium that contained CSDH fluid. The volumes of the medium and CSDH fluid were 7.5 mL and 2.5 mL per culture dish, respectively. Protein lysates were prepared from the cells harvested at 5 min, 15 min and 60 min (n = 3 per group). We used b.End3 cells treated with media alone as the control (n = 3). Total cell lysates were subjected to Western blotting analysis using antibodies against p-FAK at Tyr397 (#129840, Gene Tex), FAK (#3285, Cell Signaling Technology), p-ERK at Thr202/Tyr204 (#4695, Cell Signaling Technology), ERK (#9102, Cell Signaling Technology) and β-actin as discussed above. All antibodies were used at a 1:750 dilution except p-FAK at Tyr397 at a 1:5000 dilution. The band intensities were quantitated using densitometry with ImageQuant software (GE Healthcare).

Statistical Analysis
Data are expressed as the mean ± standard error. The Mann-Whitney U test was used for the analysis of differences between the two groups. Statistical analyses were performed using a one-way analysis of variance (ANOVA) followed by Fisher's post hoc test, as appropriate. Significance was indicated when p < 0.05.

Concentration of N-Terminal Half OPN in CSDH Fluid and Serum
First, we examined whether the N-terminal half of OPN exists in CSDH fluid and serum to determine whether the ECM is involved in the development of CSDH. The concentration of the N-terminal half of OPN in CSDH fluid (29,451.5 ± 8146.5 pg/mL) was significantly higher than that in serum (365.8 ± 73.2 pmol/L) based on the Mann-Whitney U test (p < 0.01, Figure 1).

Cultured Vascular Endothelial Cells
Endothelial cells of the mouse brain (b. End3) were obtained from the HPA Cultu Collection (London, UK). The endothelial cells were cultured in Dulbecco's modi Eagle's medium (Nissui, Tokyo, Japan) with 10% fetal bovine serum at 37 °C and 5% C

Effect of CSDH Fluid on FAK and Extracellular Signal-Regulated Kinase (ERK)
We obtained CSDH fluid during trepanation surgery and centrifuged it to rem any debris. b.End3 cells were incubated with a serum-containing medium that contai CSDH fluid. The volumes of the medium and CSDH fluid were 7.5 mL and 2.5 mL culture dish, respectively. Protein lysates were prepared from the cells harvested min, 15 min and 60 min (n = 3 per group). We used b.End3 cells treated with media al as the control (n = 3). Total cell lysates were subjected to Western blotting analysis us antibodies against p-FAK at Tyr397 (#129840, Gene Tex), FAK (#3285, Cell Signa Technology), p-ERK at Thr202/Tyr204 (#4695, Cell Signaling Technology), ERK (#91 Cell Signaling Technology) and β-actin as discussed above. All antibodies were used 1:750 dilution except p-FAK at Tyr397 at a 1:5000 dilution. The band intensities w quantitated using densitometry with ImageQuant software (GE Healthcare).

Statistical Analysis
Data are expressed as the mean ± standard error. The Mann-Whitney U test used for the analysis of differences between the two groups. Statistical analyses w performed using a one-way analysis of variance (ANOVA) followed by Fisher's post test, as appropriate. Significance was indicated when p < 0.05.

Concentration of N-Terminal Half OPN in CSDH Fluid and Serum
First, we examined whether the N-terminal half of OPN exists in CSDH fluid serum to determine whether the ECM is involved in the development of CSDH. concentration of the N-terminal half of OPN in CSDH fluid (29,451.5 ± 8146.5 pg/mL) significantly higher than that in serum (365.8 ± 73.2 pmol/L) based on the Mann-Whit U test (p < 0.01, Figure 1). . The concentration of the N-terminal half of OPN in CSDH fluid was significantly higher than that in serum based on the Mann-Whitney U test. Data represent the median values and 25th and 75th percentiles with maximum/minimum whiskers with scatter plots. * p < 0.01 according to the Mann-Whitney U test.

Western Blot Analysis of Integrins and the Angiogenic Signaling Pathway
The N-terminal half of OPN binds to the integrin receptor on the cell surface. Signal transduction from the ECM to cytoplasmic actin occurs through integrin receptors. We explored whether these molecules are expressed in the CSDH outer membrane. Figure 2 shows the results of the Western blot analyses of integrins α9 and β1 and the downstream signaling pathway components. Nearly constant β-actin levels were detected in the CSDH outer membrane samples. Integrins α9 and β1, vinculin, talin-1, FAK, paxillin, α-actinin and Src were detected in almost all samples; however, in some cases, the signals were weak. Moreover, the activated form of FAK was detected in the CSDH outer membrane. Positive controls revealed that these molecules had been correctly detected. Figure 1. Concentrations of thrombin-cleaved osteopontin (N-terminal half of OPN) in serum (n = 5) and chronic subdural hematoma (CSDH, n = 20). The concentration of the N-terminal half of OPN in CSDH fluid was significantly higher than that in serum based on the Mann-Whitney U test. Data represent the median values and 25th and 75th percentiles with maximum/minimum whiskers with scatter plots. * p < 0.01 according to the Mann-Whitney U test.

Western Blot Analysis of Integrins and the Angiogenic Signaling Pathway
The N-terminal half of OPN binds to the integrin receptor on the cell surface. Signal transduction from the ECM to cytoplasmic actin occurs through integrin receptors. We explored whether these molecules are expressed in the CSDH outer membrane. Figure 2 shows the results of the Western blot analyses of integrins α9 and β1 and the downstream signaling pathway components. Nearly constant β-actin levels were detected in the CSDH outer membrane samples. Integrins α9 and β1, vinculin, talin-1, FAK, paxillin, α-actinin and Src were detected in almost all samples; however, in some cases, the signals were weak. Moreover, the activated form of FAK was detected in the CSDH outer membrane. Positive controls revealed that these molecules had been correctly detected. Integrins β1 and α9, vinculin, talin-1, focal adhesion kinase (FAK), FAK phosphorylated at Tyr397 (p-FAK at Tyr397), paxillin, α-actinin and Src were detected in almost all cases. Positive controls are shown in the right lanes and suggest that these molecules were correctly detected. RAW 264.7, murine leukemia macrophage cell line lysate; rat liver, rat liver whole cell lysate; A431 cell lysate, epidermoid carcinoma cell lysate; rat brain lysate, rat brain whole cell lysate.

Histological Observations
Next, to determine where integrins α9 and β1, FAK and paxillin were expressed in the CSDH outer membrane, we performed an immunohistochemical analysis. The analysis revealed that integrin α9 ( Figure 3A,B), integrin β1 ( Figure 3C,D), FAK ( Figure 3E,F) and paxillin ( Figure 3G,H) were localized in the endothelial cells of blood vessels within the outer membrane. Higher magnification images distinctly showed that the endothelial cells expressed these molecules ( Figure 3B,D,F,H). The endothelial cells were consistently negative for the markers listed above in the controls without the primary antibodies ( Figure 3I). In the adjacent dura mater, there was no apparent staining for these anti- Integrins β1 and α9, vinculin, talin-1, focal adhesion kinase (FAK), FAK phosphorylated at Tyr397 (p-FAK at Tyr397), paxillin, α-actinin and Src were detected in almost all cases. Positive controls are shown in the right lanes and suggest that these molecules were correctly detected. RAW 264.7, murine leukemia macrophage cell line lysate; rat liver, rat liver whole cell lysate; A431 cell lysate, epidermoid carcinoma cell lysate; rat brain lysate, rat brain whole cell lysate.

Histological Observations
Next, to determine where integrins α9 and β1, FAK and paxillin were expressed in the CSDH outer membrane, we performed an immunohistochemical analysis. The analysis revealed that integrin α9 ( Figure 3A,B), integrin β1 ( Figure 3C,D), FAK ( Figure 3E,F) and paxillin ( Figure 3G,H) were localized in the endothelial cells of blood vessels within the outer membrane. Higher magnification images distinctly showed that the endothelial cells expressed these molecules ( Figure 3B,D,F,H). The endothelial cells were consistently negative for the markers listed above in the controls without the primary antibodies ( Figure 3I). In the adjacent dura mater, there was no apparent staining for these antibodies except for the endothelial cells in tortuous arteries that penetrated the dura mater (Supplemental Figure S1). bodies except for the endothelial cells in tortuous arteries that penetrated the dura mater (Supplemental Figure S1).

Activation of FAK and ERK in Endothelial Cells by CSDH Fluid
Furthermore, to determine the effect of the N-terminal half of OPN within CSDH fluid, we examined whether CSDH fluid induced phosphorylation of FAK and ERK in endothelial cells ( Figure 4A and 4B, respectively). Significantly higher levels of phosphorylated FAK (at Tyr397) and ERK (at Thr202/Tyr204) were achieved 5 min after the addition of CSDH fluid to cultured vascular endothelial cells than in the control (p < 0.05), whereas the levels of FAK, ERK and β-actin were not changed.
Panels (B,D,F,H), respectively. Note that these molecules are expressed in endothelial cells (B,D,F,H). Slices immunostained without primary antibodies are shown in (I). Scale bars = 100 µm.

Activation of FAK and ERK in Endothelial Cells by CSDH Fluid
Furthermore, to determine the effect of the N-terminal half of OPN within CSDH fluid, we examined whether CSDH fluid induced phosphorylation of FAK and ERK in endothelial cells ( Figure 4A and 4B, respectively). Significantly higher levels of phosphorylated FAK (at Tyr397) and ERK (at Thr202/Tyr204) were achieved 5 min after the addition of CSDH fluid to cultured vascular endothelial cells than in the control (p < 0.05), whereas the levels of FAK, ERK and β-actin were not changed.

Discussion
The expression of the N-terminal half of OPN in CSDH fluid was significantly higher than that in serum. Integrins and their downstream angiogenic pathway intermediates were detected in the outer membrane of CSDHs. Integrins α9 and β1, FAK and paxillin were expressed in the endothelium of blood vessels in the CSDH outer membrane, and CSDH fluid activated FAK in endothelial cells immediately after exposure.

Discussion
The expression of the N-terminal half of OPN in CSDH fluid was significantly higher than that in serum. Integrins and their downstream angiogenic pathway intermediates were detected in the outer membrane of CSDHs. Integrins α9 and β1, FAK and paxillin were expressed in the endothelium of blood vessels in the CSDH outer membrane, and CSDH fluid activated FAK in endothelial cells immediately after exposure.
The extracellular matrix protein osteopontin is a glycoprotein involved in physiological and pathological events during inflammatory processes. The concentrations of thrombin-cleaved osteopontin in synovial fluid were well correlated with the severity of knee osteoarthritis [15]. Compared with full-length osteopontin, the N-terminal half of OPN induces markedly more cell attachment through integrin receptors [16]. Disease activity in lupus nephritis is correlated with the urine concentrations of the N-terminal half of OPN rather than full-length osteopontin, suggesting that the N-terminal half of OPN is an indicator of inflammation of the kidney [17]. A previous study revealed that excessive coagulation, generation of thrombin and increased fibrinolysis occur within CSDH fluid [18]. Given the results of previous studies and our data, osteopontin is cleaved by thrombin within CSDH fluid, and the N-terminal half of OPN plays a role in inflammatory reactions in the CSDH outer membrane. To the best of our knowledge, this study is the first to demonstrate the existence of the N-terminal half of OPN in CSDH fluid.
Integrins are α/β heterodimeric cell surface receptors that mediate cell-cell and cell-ECM interactions and orchestrate cell attachment, movement, growth, differentiation and survival. Integrin α9 is widely expressed in a variety of cell types, including the epithelium [19]. Integrin β1 is the main β subunit for α9 in these cells [19]. The β1 class of integrins participates in many aspects of vascular biology, particularly angiogenesis [20]. β1 integrins play an important role in endothelial cell adhesion, migration and survival during angiogenesis and vascular remodeling [21,22]. A deficit of endothelial β1 integrins prohibited endothelial cell maturation, migration and sprouting and induced endothelial cell apoptosis [23]. Thrombin-cleaved osteopontin can attach to integrin α9β1 via the sequence SVVYGLR, which is located between the arginine-glycine-aspartic acid (RGD) sequence and the thrombin cleavage site [24]. Based on our data, it is possible that thrombincleaved osteopontin, i.e., the N-terminal half of OPN, induces angiogenesis through integrin α9β1 in the endothelium of the outer membrane. However, it should be noted that CSDH fluid contains matrix metalloproteinases (MMPs), which contribute to inflammatory processes [25,26]. MMP-3 and MMP-7 can also cleave OPN [27], and this cleaved OPN can bind to integrin α9β1 [28], suggesting that other molecules might be involved in the activation of the integrin α9β1 receptor.
After these integrin receptors combine with the extracellular matrix, the formation of complex multiprotein structures occurs. FAK is a regulator of signals from the ECM to the cytoplasmic actin cytoskeleton ( Figure 5) [29]. Angiogenesis is mandatory for tumor development. FAK participates in endothelial cell proliferation, which has been revealed to control tumor angiogenesis in many cancers [30]. FAK promotes angiogenesis in a dose-dependent manner [31] when overexpressed in transgenic mice [32]. α-Actin is a highly conserved protein and a member of the actin cross-linking protein family. α-Actin is phosphorylated on tyrosine residues by FAK and binds to actin [33]. These molecules regulate the flow of signals from the extracellular matrix to the actin cytoskeleton and induce angiogenesis ( Figure 5).
Paxillin is an adaptor protein located at the interface between the actin cytoskeleton and the plasma membrane [34] and is one of the key components of integrin signaling ( Figure 5). The FAK/Src complex phosphorylates ERK and tyrosine and serine residues of paxillin, promotes cell migration and regulates adhesion turnover at the cell front through paxillin [35]. In alkali-burned corneas, neovascularization occurs from the corneal limbus to the cornea, where paxillin induces the migration of endothelial cells and promotes angiogenesis [36]. Netrin-1 is a laminin-like secreted protein that is thought to be an axon guidance molecule during neural development. Netrin-1 activates the FAK/Src/paxillin pathway and modulates angiogenesis, and these changes are accompanied by the upregulation of VEGF [37]. In human retinal angiogenesis, VEGF-induced FAK/Src/paxillin signaling plays an important role [38]. Both talin and vinculin also play important roles in cell growth, morphogenesis, and cell migration during development. Marked defects in focal adhesions and embryonic death occur with the loss of either talin or vinculin in mice [39,40]. Talin is also a key regulator of the interaction between the cytoskeleton and integrins, having multiple interaction sites for other adhesome components ( Figure 5) [41]. Talin-1 is essential for endothelial proliferation and postnatal angiogenesis [42]. Furthermore, vinculin is a key regulator of cell adhesion by engaging in direct interactions with talin and actin ( Figure 5) [43,44]. After human corneal limbal epithelial cells were exposed to hypoxic conditions, the expression of talin, paxillin and vinculin differed from that under normoxic conditions, affecting cell migration [45]. Our data revealed that all of these molecules were present, as indicated by Western blot analysis, and were located in the endothelium of the CSDH outer membrane, as indicated by immunohistochemistry. Moreover, FAK and ERK in endothelial cells were activated by CSDH fluid. The N-terminal half of OPN is activated by thrombin signals to the actin cytoskeleton via integrin α9β1 located on the cell surface and induces angiogenesis within the CSDH outer membrane. Paxillin is an adaptor protein located at the interface between the actin cytoskeleton and the plasma membrane [34] and is one of the key components of integrin signaling ( Figure 5). The FAK/Src complex phosphorylates ERK and tyrosine and serine residues of paxillin, promotes cell migration and regulates adhesion turnover at the cell front through paxillin [35]. In alkali-burned corneas, neovascularization occurs from the corneal limbus to the cornea, where paxillin induces the migration of endothelial cells and promotes angiogenesis [36]. Netrin-1 is a laminin-like secreted protein that is thought to be an axon guidance molecule during neural development. Netrin-1 activates the FAK/Src/paxillin pathway and modulates angiogenesis, and these changes are accompanied by the upregulation of VEGF [37]. In human retinal angiogenesis, VEGF-induced FAK/Src/paxillin signaling plays an important role [38]. Both talin and vinculin also play important roles in cell growth, morphogenesis, and cell migration during development. Marked defects in focal adhesions and embryonic death occur with the loss of either talin or vinculin in mice [39,40]. Talin is also a key regulator of the interaction between the cytoskeleton and integrins, having multiple interaction sites for other adhesome components ( Figure 5) [41]. Talin-1 is essential for endothelial proliferation and postnatal Angiogenesis is a complex process regulated by numerous receptors, growth factors, ECM-cell interactions, etc. The concentration of VEGF in hematoma fluid has been reported to be an essential mechanistic factor in the pathophysiological progression and development of CSDH and angiogenesis [3,26]. The angiogenic effect of VEGF depends on the presence of integrin β1 [46]. FAK and Src participate in angiogenesis induced by VEGF [47,48]. We previously revealed that activation of mitogen-activated protein kinases (MAPKs) by VEGF occurs in CSDH outer membranes and plays a critical role in the angiogenesis of CSDHs [9,49]. Phosphorylated c-Jun N-terminal kinase (JNK) is expressed in the vascular endothelium of the CSDH outer membrane. FAK activates JNK through an extraordinary mechanism involving the recruitment of paxillin to the plasma membrane [50]. Overall, this work could serve as a basis for further consideration of the N-terminal half of the OPN/integrin pathway as a potential therapeutic target for CSDH.
In the present study, we note several limitations. First, from our limited number of patients, we could not detect correlations among the concentrations of the N-terminal half of OPN in CSDH fluid, the data from Western blot analyses and the development stage of CSDH. Further studies including more patients will be necessary to clarify this relationship. Second, integrins α9 and β1 and subsequent angiogenic signaling molecules were found only in the outer membrane of CSDHs. We need to determine whether these signaling molecules are activated during the development of CSDH. In our in vitro study, we should treat endothelial cells with sera from healthy people and not medium alone as a control treatment.

Conclusions
For the first time, we detected the expression of the N-terminal half of OPN in CSDH fluid and integrins α9 and β1, FAK, paxillin, vinculin and the subsequent angiogenic signaling pathway in the CSDH outer membrane. Significantly high concentrations of the N-terminal half of OPN in CSDH fluid might play an important role in angiogenesis and inflammation in CSDH, resulting in the growth of the hematoma. This angiogenic signaling pathway through integrins α9 and β1 might be an alternative therapeutic target for the treatment of refractory CSDH.