Soluble Vascular Adhesion Protein-1 Accumulates in Proliferative Diabetic Retinopathy

Abstract

Purpose.: Vascular adhesion protein (VAP)-1, a multifunctional molecule with adhesive and enzymatic properties, is expressed at the surface of vascular endothelial cells of mammals. It also exists as a soluble form (sVAP-1), which is implicated in oxidative stress via its enzymatic activity. This study explores a link between increased level of sVAP-1 and oxidative stress in proliferative diabetic retinopathy (PDR) with a focus on mechanistic components to form sVAP-1 by shedding from retinal endothelial cells.

Methods.: Protein levels of sVAP-1 and N epsilon-(hexanoyl)lysine (HEL), an oxidative stress marker, in the vitreous samples from patients with PDR or non-PDR were measured by ELISA. The mechanism of VAP-1 shedding under diabetic condition, exposure to high glucose and/or inflammatory cytokines, was explored using cultured retinal capillary endothelial cells.

Results.: Protein level of sVAP-1 was increased and correlated with HEL in the vitreous fluid of patients with PDR. Retinal capillary endothelial cells released sVAP-1 when stimulated with high glucose or inflammatory cytokines, such as TNF-α and IL-1β in vitro. Furthermore, matrix metalloproteinase-2 and -9, type IV collagenases, were the key molecules to mediate the protein cleavage of VAP-1 from retinal capillary endothelial cells.

Conclusions.: Our data for the first time provide evidence on the link between sVAP-1 and type IV collagenases in the pathogenesis of PDR.

Introduction

Vascular adhesion protein (VAP)-1, a homodimeric sialylated glycoprotein discovered by Salmi and Jalkanen in 1992, ¹ is a multifunctional molecule that possesses enzymatic activity and is involved in the leukocyte recruitment cascade. VAP-1 has a large homology with semicarbazide-sensitive amine oxidase (SSAO), which oxidizes aliphatic and aromatic primary monoamines, such as methylamine and aminoacetone, and VAP-1/SSAO converts them to the corresponding aldehydes with the release of hydrogen peroxide and ammonia. ² In addition, as an adhesion molecule, VAP-1 regulates the extravasation step during leukocyte recruitment. ³  

VAP-1 is exclusively expressed in the cellular membrane of vascular endothelial cells, smooth muscle cells, and adipocytes of mammals. ⁴ In the ocular tissues of humans ⁵ and rodents, ^6,7 VAP-1 is localized on the endothelial cells of retinal and choroidal vessels. On inflammation, VAP-1 is known to facilitate the accumulation of inflammatory cells into the inflamed tissues in concert with other leukocyte adhesion molecules. ⁸ In fact, previous studies have elucidated that VAP-1 is crucial in the pathology of systemic inflammatory diseases, including rheumatoid arthritis, inflammatory bowel diseases, atherosclerosis, and diabetes. ^4,9,10 As for ocular diseases, we recently reported, using animal models, that VAP-1 is involved in the molecular mechanisms of acute ocular inflammation, ⁶ inflammation-associated ocular angiogenesis, ⁷ and leukostasis under diabetic conditions. ¹¹ These findings have indicated that, as a leukocyte adhesion molecule, VAP-1 plays a critical role not only in systemic disorders, but also in ocular diseases associated with inflammation. 

VAP-1 also exists as a soluble form in plasma. It is believed that the soluble form of VAP-1 (sVAP-1) is released by proteolytic cleavage at the transmembrane anchor of membrane-bound VAP-1 on the cell surface, ^12–14 and there is some experimental evidence that matrix metalloproteinases (MMPs) cut off the extracellular part of membrane-bound VAP-1 in cultured adipocytes. ¹² However, the exact mechanism by which MMPs, proteinases responsible for degradation of extracellular matrix, contribute to sVAP-1 release is unknown. In addition, much attention has been paid to the elevated serum concentration of sVAP-1 in patients with diabetes, and a growing body of evidence has indicated that hyperglycemia and/or hyperinsulinemia stimulate sVAP-1 release with its sustained enzymatic activity. ^15,16 Moreover, diabetic patients with retinopathy display significantly higher plasma VAP-1/SSAO activities. ¹⁷ Metabolites generated by VAP-1/SSAO (i.e., hydrogen peroxide and aldehydes) are known to be involved in cellular oxidative stress and advanced glycation end product formation, both of which are crucial for the pathogenesis of diabetic retinopathy. ^18–20 Therefore, sVAP-1 may promote mechanisms responsible for diabetic complications in the eye, proliferative diabetic retinopathy (PDR), and the mechanism(s) to generate sVAP-1 in PDR are of great interest. 

Here, we measured the vitreous level of sVAP-1 in patients with PDR and studied the relationship between sVAP-1 accumulation and oxidative stress in PDR. In addition, we explored the molecular mechanism that cleaves VAP-1 protein under the diabetic condition. 

Methods

Specimens

Vitreous samples were collected from 37 eyes of 37 patients with PDR (22 males and 15 females; mean age, 58.8 ± 1.4 years), who underwent pars plana vitrectomy for prolonged vitreous hemorrhage and tractional retinal detachment involving macular lesions. For a control, vitreous samples were obtained from 14 eyes of 14 patients with nondiabetic ocular diseases: epiretinal membrane (ERM) and idiopathic macular hole (MH) (5 males and 9 females; mean age, 61.7 ± 0.6 years). Undiluted vitreous samples were collected and were frozen rapidly at −80°C. Fibrovascular tissues surgically removed from other subjects with PDR (n = 5) were also used for immunofluorescence study. This study was conducted in accordance with the tenets of the Declaration of Helsinki and after receiving approval from the institutional review committee of Hokkaido University Hospital. Written informed consent was obtained from all patients after an explanation of the purpose and procedures of this study. 

ELISA

The protein level of sVAP-1 in the vitreous samples was measured using ELISA kits for human sVAP-1 (Bender MedSystems, Vienna, Austria). According to the manufacturer's protocol, the samples were processed with the reagents, and the optical density was determined at 450 and 650 nm using a micro plate reader (Sunrise; TECAN, Männedorf, Switzerland). N epsilon-(hexanoyl) lysine (HEL) is a lipid hydroperoxide–modified lysine residue, a useful marker of early lipid peroxidation-derived protein modification, that is oxidative stress. ^21–24 HEL levels were measured in the vitreous samples by a competitive ELISA kit (Nikken Seil Co., Ltd., Tokyo, Japan) according to the manufacturer's protocol. The optical density was determined at 450 nm using a micro plate reader (Sunrise; TECAN). 

Immunofluorescence Microscopy

Paraffin sections of fibrovascular tissues were deparaffinized and hydrated through exposure with xylene and graded alcohols followed by water. Thereafter, slides were incubated with 10% normal goat serum (Invitrogen, Carlsbad, CA) for 30 minutes to block nonspecific binding. Sections were then incubated with primary rabbit polyclonal antibody against VAP-1 (1/50, H-43; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) or isotype control antibody at 4°C overnight and reacted with Alexa Fluor 546 goat anti-rabbit IgG (1/200; Invitrogen) and fluorescein isothiocyanate (FITC)-conjugated anti-CD31 antibody (1/200; Abcam, Cambridge, MA). Photomicrographs were taken using a fluorescence microscope (BIOREVO, BZ-9000; Keyence Corporation, Osaka, Japan). 

Cell Culture

The rat retinal capillary endothelial cell line TR-iBRB2 was provided from Fact, Inc., Sendai, Japan. ²⁵ TR-iBRB2 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) containing 5.5 mM glucose, which corresponds to blood glucose concentration during normal condition, and L-glutamine supplemented with 10% (vol/vol) fetal bovine serum and 15 mg/L endothelial cell growth factor (ECGF; Roche, Mannheim, Germany) in type I collagen-coated culture flasks or dishes. 

For high glucose stimulation, TR-iBRB2 cells were cultured in DMEM containing 25 mM D-glucose, which corresponds to blood glucose concentration during the hyperglycemic condition, for 7 days. Cultured medium was changed every 2 days. For cytokine stimulation, cells were starved for 16 hours and treated with either rat recombinant TNF-α (1, 10, 100 ng/mL; R&D Systems, Minneapolis, MN) or rat recombinant IL-1β (0.1, 1, 10 ng/mL; R&D Systems) for 3 hours. 

Immunoprecipitation and Western Blotting

For immunoprecipitation (IP), culture medium was collected and centrifuged at 1000g for 10 minutes at 4°C. After centrifugation, the supernatants were transferred to a new tube with protease inhibitor cocktail (Complete mini; Roche). The medium was precleared with Protein A-Agarose (Roche) at 4°C on a rocking platform overnight and then incubated with 1 μg/mL of anti-VAP-1 antibody (H-43) for 1 hour. The immunocomplexes were collected with protein A-agarose. To confirm whether the loaded samples contain equal amounts of protein, we loaded supernatant solution and stained them with Coomassie Brilliant Blue. 

For Western blotting of cell lysates, cells were washed twice with ice-cold PBS. Cells were then lysed in 1× Laemmli sample buffer, sonicated 4 times for 5 seconds each on ice, and centrifuged at 20,630g at 4°C for 10 minutes. The supernatants were boiled at 95°C for 3 minutes. The samples were separated by SDS-PAGE and electroblotted to polyvinylidene fluoride (PVDF) membranes (GE Healthcare, Buckinghamshire, UK). To block the nonspecific binding, the membranes were incubated with 5% skim milk and subsequently incubated with either a rabbit polyclonal antibody against VAP-1 (1/1000, H-43) or a rabbit monoclonal antibody against β-actin (1/2000; Cell Signaling Technology, Beverly, MA) at 4°C overnight, followed by incubation with a horseradish peroxidase–conjugated goat anti-rabbit IgG (1:2000; Jackson ImmunoResearch, West Grove, PA). Signals were visualized with chemiluminescence (ECL plus; GE Healthcare) according to the manufacturer's protocol. 

RT-PCR

Total RNA was extracted by RNeasy mini kit (Qiagen, Hilden, Germany) and reverse transcribed into cDNA by using PrimeScript II first strand cDNA Synthesis Kit (Takara, Tokyo, Japan) according to the manufacturer's protocol. The primer sequences used for PCR and the expected size of the amplification products are as follows: 5′-TGC CGA GCA CAC ACT GGG CAC T-3′ (sense) and 5′-TGC GGT AGC CTC GCC TGT GAC C-3′ (antisense) for VAP-1; 281 bp, 5′-TAC TGA ACT TCG GGG TGA TCG GTC C-3′ (sense) and 5′-CAG CCT TGT CCC TTG AAG AGA ACC-3′ (antisense) for TNF-α; 295 bp, 5′-GCT GAT ACT GAC ACT GGT ACT G-3′ (sense) and 5′-CAA TCT TTT CTG GGA GCT C-3′ (antisense) for MMP-2; 216 bp, 5′-CAT CCG TAA AGA CCT CTA TGC CAA C -3′ (sense) and 5′-ATG GAG CCA CCG ATC CAC A-3′ (antisense) for ACTb; 171 bp. Rat IL-1β/ IL-1F2 primer pair (RDP-174; R&D Systems) was used for Il-1β (584 bp). Each PCR reaction was cycled at 98°C for 10 seconds, 61°C for 30 seconds (VAP-1), 60°C for 30 seconds (MMP-2), 58°C for 30 seconds (TNF-α), or 55°C for 45 seconds (IL-1β), and 72°C for 30 seconds for 30 to 40 cycles, followed by 72°C for 7 minutes. The products were electrophoresed in 2% agarose gels, stained with SYBR safe (Invitrogen, Carlsbad, CA), and visualized with UV irradiation. 

Inhibitor Treatment and RNA Interference for Proteases

Broad-spectrum MMP inhibitor BB94 (British Biotech Pharmaceuticals Ltd., Oxford, UK), a disintegrin and metalloproteinase (ADAM) inhibitor KB-R7789, ²⁶ inhibitor for both MMP-2 and MMP-9 (Merck, Frankfurter, Germany), and inhibitor for MMP-9 (Merck) were used in this study. TR-iBRB2 cells were seeded at a concentration of 4 × 10⁴ cells/6-cm culture dish in DMEM containing 5.5 or 25 mM D-glucose. After 5 days of culture, media were changed into fresh DMEM containing 5.5 or 25 mM D-glucose and protease inhibitors (1 μM BB94, 1 μM KB-R7785, 10 μM MMP-2/9 inhibitor, and 1 μM MMP-9 inhibitor). After 48-hour incubation, culture media were collected and subjected to IP 

Statistical Analysis

All results are expressed as the mean ± SEM as indicated. Student's t-test was used for statistical comparison between groups. Differences between the means were considered statistically significant when the probability values were less than 0.05. Pearson correlation coefficient was used to examine correlations. 

Results

sVAP-1 in Vitreous Samples of Patients with PDR

To determine whether sVAP-1 is increased in PDR eyes, the protein levels of sVAP-1 in the vitreous samples were measured by ELISA system. The sVAP-1 was detectable in all the PDR and non-PDR vitreous samples and was significantly elevated in the vitreous fluids of PDR patients (7.9 ± 0.9 ng/mL, n = 37, P < 0.001) compared with those of non-PDR patients (1.0 ± 0.2 ng/mL, n = 14, Fig. 1A). In all five fibrovascular tissue samples, immunofluorescence microscopy showed that VAP-1 protein was exclusively localized in the vessels stained with CD31 (Fig. 1B). 

Figure 1. 

 

sVAP-1 and oxidative stress marker HEL in PDR eyes. (A) Levels of sVAP-1 in the PDR and control vitreous samples. *P < 0.05. (B) Representative fluorescent micrographs of VAP-1 staining in fibrovascular tissues. (a) Red, VAP-1 (Alexa Fluor 546); (b) green, CD31 (FITC); (c) blue, counterstaining for the nuclei with 4′,6-diamidino-2-phenylindole (DAPI); (d) merged image (arrows indicate the localization of VAP-1 in endothelial cells; scale bar, 50 μm); (e) isotype control. (C) Correlation between sVAP-1 and HEL in the vitreous samples obtained from PDR eyes. The equation of the regression line is y = 0.3077x + 4.6828. n = 37.

Next, to study the correlation between sVAP-1 and oxidative stress in PDR eyes, we measured the level of HEL, an oxidative stress marker, in the vitreous samples obtained from patients with PDR. The level of sVAP-1 showed a moderate correlation with HEL concentration in the vitreous samples of PDR patients (ρ = 0.422, P < 0.05, Fig. 1C). 

Increase of sVAP-1 Release from Retinal Capillary Endothelial Cells with High Glucose Stimulation

To investigate whether high glucose facilitates release of sVAP-1 from retinal endothelial cells, we stimulated TR-iBRB2 cells with different concentrations of glucose and measured the levels of sVAP-1 in culture media 7 days after stimulation. IP western blotting revealed that the exposure of endothelial cells to higher glucose increased the protein level of sVAP-1 in culture media (Fig. 2A) and, compared with 5.5 mM dose of glucose, the level of sVAP-1 showed approximately 1.5-fold increase in the supernatants of TR-iBRB2 cells when stimulated with 25 mM glucose (144.6% ± 6.8%, n = 3, P < 0.01, Fig. 2B). By contrast, VAP-1 mRNA expression was not altered with high glucose stimulation (Fig. 2C), indicating that the increase of sVAP-1 in the supernatants is via shedding of preexisting VAP-1 protein in TR-iBRB2 cells, but not de novo synthesis. In accord with the previous reports, ^27,28 expression levels of inflammatory cytokines such as TNF-α and IL-1β were upregulated when stimulated with high glucose (Fig. 2C). 

Figure 2. 

 

Impact of high glucose stimulation on VAP-1 shedding in retinal capillary endothelial cells. (A) IP Western blotting of sVAP-1 in the culture media from TR-iBRB2 cells stimulated with different concentrations of glucose (5.5 and 25 mM). (B) Densitometric analysis of the bands. Values are expressed as the mean ± SEM (n = 3 in each group). **P < 0.01. (C) RT-PCR amplification of VAP-1, TNF-α, and IL-1β mRNA after high glucose stimulation.

Increase of sVAP-1 Release from Retinal Capillary Endothelial Cells with Inflammatory Cytokine Stimulation

To further investigate the mechanism of sVAP-1 release induced with high glucose stimulation, we examined whether inflammatory cytokines upregulated with high glucose (i.e., TNF-α and IL-1β) increase the release of sVAP-1 from cultured retinal capillary endothelial cells. The protein level of sVAP-1 in culture media was found to increase in a dose-dependent manner when TNF-α was added to TR-iBRB2 cells (Fig. 3A). The level of sVAP-1 showed 1.9- and 2.3-fold increase when stimulated with 10 ng/mL (P < 0.01) and 100 ng/mL (P < 0.05) TNF-α, respectively (n = 3, Fig. 3B). Inversely, the level of membrane-bound form of VAP-1 in cell lysates was decreased by TNF-α stimulation with dose dependency, possibly via acceleration of VAP-1 shedding (Fig. 3C). VAP-1 mRNA expression remained unchanged 2 and 24 hours after stimulation with different concentrations of TNF-α (Fig. 3D). 

Figure 3. 

 

Impact of TNF-α stimulation on VAP-1 shedding in retinal capillary endothelial cells. (A) IP Western blotting of sVAP-1 in the culture media from TR-iBRB2 cells stimulated with different concentrations of TNF-α (10, 100 ng/mL). (B) Densitometric analysis of the bands. Values are expressed as the mean ± SEM (n = 3 in each group). *P < 0.05, **P < 0.01. (C) Western blotting of VAP-1 and β-actin expression in cell lysates of TR-iBRB2 cells stimulated with different concentrations of TNF-α. (D) RT-PCR amplification of VAP-1 at 2 and 24 hours after stimulation with different concentrations of TNF-α.

Similarly, the level of sVAP-1 showed 2.1- and 2.0-fold increase when stimulated with 1 ng/mL (P < 0.05) and 10 ng/mL (P < 0.01) IL-1β, respectively (n = 3, Figs. 4A, 4B), decreased the level of the membrane-bound form of VAP-1 in cell lysates (Fig. 4C), and showed no effect on VAP-1 mRNA expression (Fig. 4D). 

Figure 4. 

 

Impact of IL-1β stimulation on VAP-1 shedding in retinal capillary endothelial cells. (A) IP Western blotting of sVAP-1 in the culture media from TR-iBRB2 cells stimulated with different concentrations of IL-1β (1, 10 ng/mL). (B) Densitometric analysis of the bands. Values are expressed as the mean ± SEM (n = 3 in each group). *P < 0.05, **P < 0.01. (C) Western blotting of VAP-1 and β-actin expression in cell lysates of TR-iBRB2 cells stimulated with different concentrations of IL-1β. (D) RT-PCR amplification of VAP-1 at 2 and 24 hours after stimulation with different concentrations of IL-1β.

In accord with the blotting data, the immunofluorescence study showed the decrease of VAP-1 expression in TR-iBRB2 cells after stimulating with TNF-α or IL-1β (Fig. 5). 

Figure 5. 

 

VAP-1 protein in retinal capillary endothelial cells after stimulation with inflammatory cytokines. Representative micrographs of TR-iBRB2 cells stimulated with PBS control (left), TNF-α (middle), and IL-1β (right) immunostained for VAP-1 (Alexa Fluor 546, red). Nuclei were counterstained with DAPI (blue). Scale bar, 50 μm.

Shedding of sVAP-1 from Retinal Capillary Endothelial Cells

To better understand the mechanism of the accelerated sVAP-1 release with high glucose and inflammatory cytokine stimulation, we sought to identify proteinase(s) that cleave sVAP-1 from retinal capillary endothelial cells. Because it was previously reported that the broad-spectrum inhibitor for MMPs, Batimastat (BB94), blocks VAP-1 shedding in adipocytes, ¹² we hypothesized that sVAP-1 release from retinal capillary endothelial cells is attributed to proteinase cleavage of the membrane-bound form of VAP-1 and thus focused on MMPs and ADAMs, both of which belong to the metalloproteinase family. In accord with the previous report, ¹² BB94 blocked sVAP-1 release from TR-iBRB2 cells stimulated with high glucose, whereas KB-R7785, an ADAM inhibitor, showed negligible effect on the sVAP-1 release in culture media (Fig. 6A). Furthermore, a specific inhibitor for both MMP-2 and MMP-9 decreased the release of sVAP-1 from TR-iBRB2 cells by 55.2% in comparison with the control (n = 3, P < 0.01, Figs. 6B, 6C), and a specific inhibitor for MMP-9 alone was less effective (35%) in the inhibition of VAP-1 shedding than the inhibitor for both MMP-2 and MMP-9 (n = 3, P < 0.05, Figs. 6B, 6C). 

Figure 6. 

 

Effect of metalloproteinase blockade on VAP-1 shedding in retinal capillary endothelial cells. (A) IP Western blotting of sVAP-1 in the culture media from TR-iBRB2 cells treated with broad MMP inhibitor, BB94, or ADAM inhibitor, KB-R7785, under high glucose condition. (B) IP Western blotting of sVAP-1 in the culture media from TR-iBRB2 cells treated with a specific inhibitor for both MMP-2 and MMP-9 or inhibitor for MMP-9 under high glucose condition. (C) Densitometric analysis of the bands. Values are expressed as the mean ± SEM (n = 3 in each group). *P < 0.05, **P < 0.01.

Discussion

In the present study, we demonstrated that (1) sVAP-1, which possesses enzymatic activity to generate hydrogen peroxide, is increased and correlated with oxidative stress in the vitreous fluid of patients with PDR; (2) retinal capillary endothelial cells produce the membrane-bound form of VAP-1 and release sVAP-1 when stimulated with high glucose or inflammatory cytokines such as TNF-α and IL-1β; and (3) MMP-2 and MMP-9 are the proteinases predominantly responsible for VAP-1 shedding from retinal capillary endothelial cells. To our knowledge, this is the first report regarding the molecular mechanism of sVAP-1 accumulation in PDR eyes. 

It has been demonstrated that soluble forms of leukocyte adhesion molecules, such as selectins, intercellular adhesion molecule (ICAM)-1, and vascular cell adhesion molecule (VCAM)-1, are increased in the vitreous of PDR. ²⁹ Our data have revealed for the first time that the vitreous level of sVAP-1, a leukocyte adhesion molecule with enzymatic function as an SSAO, is also elevated in PDR. Previously, it was shown that methylamine, an endogenous substrate for VAP-1/SSAO, was present in intraocular fluid, for instance, in the aqueous humor. ³⁰ Methylamine is a metabolic end product of diverse compounds, including epinephrine, creatine, nicotine, and sarcosine. ³¹ In addition, aminoacetone, an intermediate in the metabolism of threonine and glycine, is also an endogenous substrate for VAP-1/SSAO. ³² Therefore, it is most likely that the substrates for VAP-1/SSAO are abundant in intraocular fluid. It has been proposed that these aliphatic amines become potentially harmful and cause cellular damage via oxidative stress in the presence of VAP-1/SSAO. ³³ Our data indicate that elevated sVAP-1 may increase deamination of aliphatic amines and, in turn, participates in increase of oxidative stress in the vitreous of PDR eyes. Indeed, in the current study the oxidative stress marker HEL, which was reportedly increased in the vitreous of PDR, ³⁴ correlated with the vitreous level of sVAP-1 in PDR. Oxidative stress promotes the production of VEGF and MMPs, both of which are key molecules in neovascularization ^35,36 and are therefore known to be crucial in the molecular mechanism of diabetic vascular complications. ³⁷ Our findings suggest that sVAP-1 may play a potential role in the increase of oxidative stress via its enzymatic activities in PDR eyes. 

There is controversy regarding the source of the soluble form of leukocyte adhesion molecules in PDR eyes. In this study, we excluded the vitreous samples with gross hemorrhage; nevertheless, it is possible that sVAP-1 was originated from blood contamination due to vitreous hemorrhage or vascular leakage in PDR. However, it is of note that sVAP-1 was also detected in the vitreous fluid of non-PDR patients (i.e., MH and ERM), indicating that vitreous sVAP-1 is present in ocular tissues other than circulating sVAP-1 contamination. Thus, we sought to find the cellular source of VAP-1 in ocular tissues. Previously, membrane-bound VAP-1 was detected in retinal vessels ⁵ and SSAO activity was found in retinal homogenates. ³⁸ In the current study, immunofluorescence microscopy revealed that VAP-1 protein was present in vascular endothelial cells of fibrovascular tissues. Furthermore, consistent with the immunofluorescence data, the cultured retinal capillary endothelial cells, TR-iBRB2, produced both the membrane-bound and the soluble form of VAP-1. Therefore, the previous and current data indicate that sVAP-1 in PDR is derived, at least in part, from vascular endothelial cells in fibrovascular tissues and retinal vessels. 

In PDR, glucose concentration is increased in aqueous humor and vitreous, ³⁹ and TNF-α and IL-1β are elevated in the vitreous fluid and fibrovascular tissues. ^40,41 High glucose is also known to increase the level of IL-1β in retinal capillary cells. ²⁷ In the present study, we have demonstrated that shedding of membrane-bound VAP-1 on retinal capillary endothelial cells is accelerated with stimulation of high glucose or inflammatory cytokines such as TNF-α and IL-1β. In addition, we have shown that high glucose stimulation upregulates the expression of TNF-α and IL-1β in retinal capillary endothelial cells. Altogether, these findings indicate that high glucose concentration aggravates the inflammatory condition in PDR via upregulation of proinflammatory cytokines, such as TNF-α and IL-1β, followed by VAP-1 shedding from retinal capillary endothelial cells. 

The inflammatory cytokines TNF-α and IL-1β are known to upregulate MMP-2 and MMP-9. ^42,43 Furthermore, our group previously reported that MMP-2 and MMP-9 are increased and activated in PDR eyes. ⁴⁴ MMP-2 and MMP-9 can degrade type IV collagen, laminin, and fibronectin, the main constituents of the basement membrane; thereby, MMPs play a crucial role in degradation of basement membrane during angiogenesis. ^45,46 Accumulated lines of evidence have further indicated that MMP-2 and MMP-9 have a potential to cleave a variety of nonextracellular matrix proteins, including cell surface adhesion molecules. ^47,48 Thus, in this study we examined the possibility that MMP-2 and MMP-9 may shed the membrane-bound form of VAP-1 from retinal capillary endothelial cells. Indeed, we have demonstrated that BB-94, a broad MMP inhibitor, suppresses shedding of the membrane form of VAP-1 in retinal capillary endothelial cells. Whereas an MMP-9 specific inhibitor reduced the VAP-1 shedding, the inhibitor for both MMP-2 and MMP-9 was more effective in the inhibition than the MMP-9–specific inhibitor alone. Therefore, it is most likely that MMP-2 and MMP-9 are the proteinases primarily responsible for VAP-1 shedding from retinal capillary endothelial cells, and they may act independently on the shedding. However, which MMP is predominantly responsible for the cleavage of the sVAP1 is still unclear. Further evaluation is required to elucidate the detailed mechanism. 

In summary, high glucose and inflammatory cytokines such as TNF-α and IL-1β facilitate shedding of membrane-bound VAP-1 from retinal capillary endothelial cells through the action of MMP-2 and MMP-9, both of which are also crucially involved in the formation of fibrovascular tissues. Furthermore, sVAP-1 accumulated in the vitreous fluid presumably induces oxidative stress in PDR eyes. These results provide information on the link between type IV collagenases and sVAP-1, ectoenzyme for both leukocyte recruitment and oxidative stress. 

Acknowledgments

The authors thank Ikuyo Hirose and Shiho Nanba (Hokkaido University Graduate School of Medicine) for their skillful technical assistance. We are indebted to Fact, Inc. for the generous provision of TR-iBRB2 cells used in this project. 

Footnotes

 Supported by Mishima Saiichi Memorial Ophthalmic Research Japan Foundation Award (KN) and the Eye Research Foundation for the Aged (KN).

Footnotes

 Disclosure: M. Murata, None; K. Noda, None; J. Fukuhara, None; A. Kanda, None; S. Kase, None; W. Saito, None; Y. Ozawa, None; S. Mochizuki, None; S. Kimura, None; Y. Mashima, None; Y. Okada, None; S. Ishida, None