Protective Effect of Clusterin on Blood–Retinal Barrier Breakdown in Diabetic Retinopathy

Abstract

Purpose.: To investigate whether clusterin attenuates blood–retinal barrier (BRB) breakdown in diabetic retinopathy.

Methods.: Mice with streptozotocin-induced diabetes and advanced glycation end product–treated human retinal microvascular endothelial cells (HRMECs) were used to determine the effect of clusterin on vascular permeability and tight junction protein expression, through perfusion of retinal vessels with FITC-bovine serum albumin, a [³H]sucrose permeability assay, a cell viability assay, Western blot analysis, immunocytochemistry, immunohistochemistry, and terminal deoxynucleotidyl transferase biotin-dUTP nick-end labeling.

Results.: Up to 20 μg/mL of clusterin, which is 20 times the effective therapeutic concentration, did not affect the viability of the HRMECs. Moreover, it caused no toxicity in the retina. It effectively inhibited vascular endothelial growth factor–induced hyperpermeability in the HRMECs and the retinas. The antipermeability activity of clusterin was related to the restoration of tight junction proteins. Finally, it was shown to reduce leakage from the vessels in the diabetic retinas and to restore the expression of the tight junction proteins.

Conclusions.: The data suggest that clusterin, a well-known antipermeability factor naturally secreted by cells, may have therapeutic potential in the treatment of diabetic BRB breakdown.

Diabetic retinopathy (DR) is one of the leading causes of blindness. It is clinically classified as having two stages: nonproliferative and proliferative. ¹ Although the exact mechanism by which capillary nonperfusion occurs has not been elucidated, nonprogressive disease becomes progressive with the increase in retinal ischemia, which leads to growth of new vessels and potential severe and irreversible vision loss. ² However, the main cause of vision loss in DR is macular edema, which occurs at any stage and results from blood–retinal barrier (BRB) breakdown, characterized by vascular leakage due to increased vascular permeability. ^1,2  

Diabetes alters the structure and function of most cell types in the retina, including the vasculature and neural network, ³ which is closely related to BRB breakdown in the early stage of DR. ⁴ As shown in our previous reports, the cellular interactions that regulate the blood–neural barrier by modulating both brain angiogenesis and tight junction formation ⁵ also play a critical role in retinal barrier genesis, ⁶ and barrier function in retinal vessels is modulated by the retinal endothelial junction structure. ^6,7 Recently, we have shown that zonula occludens (ZO)-1 and occludin are components of tight junctions in retinal endothelial cells. ^6,8,9 In particular, ZO-1 expression is inversely related to the permeability of the BRB, ^6,8–10 as well as occludin. ¹¹  

Clusterin, also known as testosterone-repressed prostate message-2, apolipoprotein J, or sulfated glycoprotein-2, is a major secretory glycoprotein that is composed of two disulfide-linked, 35- to 40-kDa subunits (α and β) encoded by a single gene. ¹² It has been implicated in diverse physiological functions including that of an extracellular chaperone that stabilizes stressed proteins in a folding-competent state. ^13,14 Recently, we have demonstrated that clusterin is upregulated in the developing retina, especially in retinal endothelial cells, protects cells from stress, and promotes cell viability, ⁸ which is mediated by Akt activation. ⁹ We have also provided definite evidence of a protective role of clusterin in ischemia-induced BRB breakdown. ⁹ Therefore, given that diabetic macular edema is closely related to ischemic injury to the retina and clusterin protects the BRB from ischemia-induced breakdown, it is reasonable to apply clusterin to the prevention of diabetes-induced BRB breakdown. 

In the present study, clusterin effectively prevented BRB breakdown in a mouse model of DR. In addition to protecting against ischemia-induced retinal endothelial cell death and tight junction protein loss, ⁹ we suggest that clusterin, a well-known antipermeability factor naturally secreted from cells, has therapeutic potential for the treatment of diabetic BRB breakdown. 

Materials and Methods

Mice

C57BL/6 mice were purchased from Samtako (Seoul, Korea). Care, use, and treatment of all animals in this study were in strict agreement with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. C57BL/6 mice were kept in standard 12-hour dark–light cycles and approximately 23°C room temperature. 

Cell Culture

Human retina microvascular endothelial cells (HRMECs) were purchased from the Applied Cell Biology Research Institute (Kirkland, WA) and grown on attachment-factor–coated plates in complete medium (Cell Systems, Kirkland, WA) or in M199 medium supplemented with 20% fetal bovine serum (FBS), 3 ng/mL basic fibroblast growth factor (bFGF; Millipore, Bedford, MA), and 10 U/mL heparin (Sigma-Aldrich, St. Louis, MO). The HRMECs were taken from passages 4 to 6. Advanced glycation end product (AGE; Calbiochem, Darmstadt, Germany) treatment (10 μg/mL) was performed on cells cultured in serum-free M199 supplemented with 1% (vol/vol) penicillin-streptomycin. 

Affinity Purification of Clusterin from Human Serum

Clusterin was purified from fresh normal human plasma as in our previous description, which complied with the Declaration of Helsinki. ^8,9,15 Human plasma supplemented with 0.5 mM phenylmethylsulfonyl fluoride (PMSF) was precipitated using 12% polyethylene glycol (PEG, MW 3350; Sigma-Aldrich) overnight at 4°C, and after centrifugation, the supernatant was reprecipitated with 23% PEG. This precipitate was dissolved in 10 mM Tris buffer (pH 7.4), 0.5 mM PMSF, subjected to DEAE Sepharose column chromatography (Fast Flow; GE Health Care Life Sciences, Buckinghamshire, UK), and equilibrated with 10 mM Tris buffer (pH 7.4). Fractions were obtained by eluting with a linear gradient from 0 to 0.5 M NaCl, and the pool of positive fractions containing clusterin (validated by immunoblot analysis with clusterin antibody; M18; Santa Cruz Biotechnology, Santa Cruz, CA) were subjected to heparin Sepharose column chromatography (Fast Flow; GE Health Care Life Sciences) on a system equipped with a 20-mM potassium phosphate buffer (pH 6.0) pre-equilibrated column. Proteins bound to heparin Sepharose were fractionated by using a linear gradient of NaCl (0–2 M), and fractions positive for clusterin were sorted by immunoblot analysis with anti-clusterin M18. The serum clusterin obtained was finally purified by affinity chromatography (Cyanogen Bromide-Activated Sepharose 4B; Sigma-Aldrich) covalently conjugated with anti-clusterin monoclonal antibody (1G8), which was generated in our laboratory with recombinant human full-length clusterin expressed in Escherichia coli as an antigen. The positive pool of clusterin obtained by heparin Sepharose column chromatography was then applied to a 1G8 affinity column at 4°C. The column was initially washed with 10 mM potassium phosphate buffer (pH 7.4) containing 0.5 M NaCl and 1% Triton X-100 and then rewashed with 10 mM potassium phosphate buffer (pH 7.4) containing 0.5 M NaCl. Pure clusterin remaining on the column was collected by eluting with 2 M guanidine-HCl in 0.5 M NaCl. The eluted protein was dialyzed against 5 mM potassium phosphate (pH 6.5) and lyophilized before being stored at −80°C. 

Induction of Diabetes in Mice

As described in our previous report, ¹⁶ 10-week-old male mice were intraperitoneally injected with 180 mg/kg streptozotocin (Sigma-Aldrich) to induce diabetes. If plasma glucose concentration was >300 mg/dL at 24 hours after streptozotocin injection, the mice were considered to be diabetic. To assess the antipermeability activity of clusterin, we administered an intravitreal injection of 1 μg/mL clusterin in 1 μL PBS or of PBS alone (control) into diabetic mice 7 days after streptozotocin injection. 

Mice Retinal Tissue Preparation

One day after the intravitreal injection of 1 μg/mL clusterin in 1 μL PBS or of PBS only into mice with streptozotocin-induced diabetes of 7 days' duration or mice with retinal vascular leakage that was induced by intravitreal injection of 20 ng/mL VEGF (Sigma-Aldrich) in 1 μL PBS, the mice were carefully killed, and the eyes were enucleated and hemisected at the ora serrata. The retinas were gently teased off the sclera with a fine brush. Contamination by retinal pigment epithelial cells was reduced to a minimum. Whole retinal proteins were extracted with lysis buffer (50 mM Tris [pH 7.6], 150 mM NaCl, 1% Triton X-100, 0.1% SDS, protease inhibitor cocktail [Sigma-Aldrich]), and 1 mM PMSF) on ice for 20 minutes and centrifuged at 14,000 rpm for 20 minutes. The supernatants were then harvested and stored at −80°C. 

Leakage Assessment by Perfusion of Retinal Vessels with FITC-BSA

Retinal vascular leakage was assessed by perfusion of the retinal vessels with FITC-bovine serum albumin (FITC-BSA; Sigma-Aldrich) in a method modified from our previous report. ¹⁷ Briefly, 1 day after intravitreal injection of 1 μg/mL clusterin in 1 μL PBS or of PBS alone into mice with streptozotocin-induced diabetes of 7 days' duration or mice with VEGF-induced retinal vascular leakage, deeply anesthetized mice were perfused through the tail vein with FITC-BSA dissolved in PBS. After a 1-hour perfusion, the eyes were enucleated and fixed in 4% paraformaldehyde for 2 hours. The retinas were dissected, flatmounted (Dako mounting medium; DakoCytomation, Glostrup, Denmark), and viewed by fluorescence microscopy (BX50; Olympus, Tokyo, Japan) at 100× magnification. The experiments were repeated at least three times. For quantification of retinal vascular leakage, after a 1-hour FITC-BSA perfusion, the eyes were enucleated, embedded in OCT medium, and immediately frozen in liquid nitrogen. The plasma was collected and assayed for fluorescence with a fluorescence spectrophotometer (SPEX; Molecular Devices, Sunnyvale, CA) based on the standard curves of FITC-BSA obtained in normal mouse plasma. Frozen retinal sections (5 μm thick) collected every 30 μm were viewed with a fluorescence microscope (BX50; Olympus), and six images from the nonvascular retina (200 μm²) were collected in each section. Conditions and exposure durations were identical among all photos included in the comparison. FITC-BSA fluorescence intensity was quantified (Q-win computer software; Leica, Wetzlar, Germany) and normalized to plasma fluorescence intensity for each animal. 

Cell Viability Assay

Cell viability was evaluated with the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. HRMECs (1 × 10⁵ cells) were plated in 96-well plates and cultured overnight. The cells were treated with clusterin (0.1–20 μg/mL) for 48 hours. The medium was then replaced with fresh medium containing 0.5 mg/mL MTT for 4 hours. After incubation, the medium was carefully removed from the plate, and DMSO was added to solubilize the formazan produced from MTT by the viable cells. Absorbance was measured at 540 nm with a microplate reader (Molecular Devices). 

[³H]sucrose Permeability Assay

With some modifications of our published method, ¹⁸ HRMECs (1 × 10⁵ cells) were plated onto a permeable filter (Transwell; Corning Costar, Cambridge, MA). After reaching confluence, the HRMECs were treated for 6 hours with 20 ng/mL VEGF or 1 μg/mL clusterin, and 50 μL (0.8 μCi/mL) [³H]sucrose (1 μCi/μL; Amersham Pharmacia, Buckinghamshire, UK) was added to the upper compartment of the permeable filter. After 30 minutes, the amount of radioactivity that had diffused into the lower compartment was determined with a liquid scintillation counter (Perkin Elmer/Wallac, Gaithersburg, MD). 

Western Blot Analysis

Western blot analysis was performed by standard methods. The protein concentration was measured with a BCA protein assay kit (Pierce, Rockford, IL). Equal amounts of protein were separated by electrophoresis on 5% to 10% SDS-PAGE and transferred electrophoretically onto nitrocellulose membrane (Amersham, Little Chalfont, UK). The membranes were blocked for 30 minutes in 5% nonfat milk. After the reaction was blocked, the membranes were incubated overnight with anti-ZO-1 (1:1000; Zymed), anti-ZO-2 (1:2000; Zymed), and anti-occludin (1:1000; Zymed) at 4°C. After they were washed with phosphate-balanced solution (PBS)-T, the membranes were incubated for 1 hour at room temperature with horseradish peroxidase–conjugated anti-rabbit IgG or anti-mouse IgG (1:10,000; Pierce) in PBS-T and 1% nonfat milk. To ensure the equal loading of protein in each lane, the blots were stripped and reprobed with an antibody against β-actin. Intensities were normalized relative to control values. The blots were scanned with a flatbed scanner and the band intensity was analyzed (TINA software; Raytest, Staubenhardt, Germany). 

Immunohistochemistry

The enucleated mouse eyes used for immunohistochemistry were immersion fixed in 4% paraformaldehyde and subsequently embedded in paraffin. Serial sections (4 mm thick) were prepared from paraffin blocks. The sections were deparaffinized and hydrated by sequential immersion in xylene and graded alcohol solutions, treated with proteinase K for 5 minutes at 37°C, and then treated with normal serum obtained from the same species in which the secondary antibody was developed for 10 minutes to block nonspecific staining. Slides were incubated overnight at 4°C with anti-ZO-1 (1:100; Zymed) and anti-platelet/endothelial cell adhesion molecule (PECAM)-1 (1:100; Chemicon, Temecula, CA). AlexaFluor 546 donkey anti-goat IgG (1:400; Molecular Probes, Eugene, OR) and AlexaFluor 488 donkey anti-rabbit IgG (1:400, Molecular Probes) were used as secondary antibodies. The slides were mounted in aqueous mounting medium (Faramount; Dako) and observed by light microscopy (Carl Zeiss Meditec, Chester, VA). 

Terminal Deoxynucleotidyl Transferase Biotin-dUTP Nick-End Labeling Assay

Clusterin (20 μg/mL) in 1 μL PBS or in PBS alone was intravitreally injected into 10-week-old mice. The mice were killed 3 days after clusterin or PBS injection and the eyes enucleated. The globes were fixed in 4% paraformaldehyde in 0.1 M phosphate buffer for 24 hours and embedded in paraffin. TUNEL staining was performed with a kit (ApopTag; Intergen, Purchase, NY), according to the manufacturer's instructions. TUNEL-positive cells were evaluated by two masked and independent observers (Jeo-HK, JinHK) in 10 fields randomly selected from each slide at ×400 magnification by fluorescence microscope (BX50; Olympus). 

Immunocytochemistry

ZO-1 expression in the intercellular junctions was examined by an immunocytochemistry according to our published method. ¹⁹ Briefly, treated cells were fixed with 2% paraformaldehyde and permeabilized with 0.2% Triton X-100. After they were washed in PBS, the slides were blocked with 3% BSA for 1 hour, and the cells were incubated with anti-ZO-1 (1:1000; Zymed) at 4°C, followed by incubation with anti-goat IgG-rhodamine (Santa Cruz Biotechnology). The slides were then viewed by fluorescence microscopy (BX50; Olympus). 

Statistical Analysis

Differences between the groups were evaluated for significance with Student's paired t-test (SPSS for Windows, ver. 12.0; SPSS, Chicago, IL). The results are shown as the mean ± SD. P ≤ 0.05 was considered statistically significant. 

Results

Effect of Clusterin on the Viability of Retinal Endothelial Cells

To investigate the cytotoxic effect of clusterin on retinal endothelial cells, we performed an MTT assay on various concentrations of clusterin (0.1–20 μg/mL). The viability of the HRMECs treated with up to 20 μg/mL clusterin was not affected (Fig. 1A). In addition, retinal toxicity after intravitreal injection of 20 μg/mL clusterin, 20 times the effective therapeutic dose, ^8,9 was evaluated through histologic examination and TUNEL assay. As shown in Figure 1B, the retina was of normal thickness, and all retinal layers were clear without any inflammatory cells in the vitreous, retina, or choroid. Compared with the control, TUNEL-positive cells were not increased by clusterin treatment. 

Figure 1.

 

Clusterin did not affect the viability of retinal endothelial cells and induced no retinal toxicity. (A) Various concentrations of clusterin (0.1–20 μM) were incubated with HRMECs for 2 days. Cell viability was measured by MTT assay. The data represent the mean ± SE of results in three independent experiments (*P < 0.05). (B) Clusterin (20 μg/mL) was intravitreally injected, and the globes were enucleated 3 days after treatment. TUNEL-positive cells were randomly measured in 10 randomly selected fields in each slide at ×400 magnification. Data are representative of results in three independent experiments. GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer. Scale bar, 50 μm.

Effect of Clusterin on VEGF-Induced Hyperpermeability in Retinal Endothelial Cells and Retinal Vessels

To evaluate the antipermeability effect of clusterin on retinal endothelial cells, we conducted a [³H]sucrose permeability assay in HRMECs. [³H]sucrose permeability was significantly increased by VEGF, an effect that was completely prevented by clusterin treatment of the HRMECs (Fig. 2A). 

Figure 2.

 

Clusterin inhibited VEGF-induced hyperpermeability in retinal endothelial cells and retinal vessels. (A) HRMECs were treated for 6 hours with 20 ng/mL VEGF or 1 μg/mL clusterin. The results of a [³H]sucrose permeability assay are expressed as counts per minutes (cpm). Each value represents the mean ± SE of three independent experiments (*P < 0.05). (B) Vascular leakage in the retina was evaluated by measuring FITC-BSA fluorescence. At 1 day after intravitreal injection of 1 μg/mL clusterin in 1 μL PBS or PBS alone in mice with VEGF-induced retinal vascular leakage, a wholemount retinal preparation was examined after a 1-hour perfusion of FITC-BSA. FITC-BSA fluorescence intensity was measured by image analysis in serial retinal sections. The average retinal FITC-BSA fluorescence intensity was calculated and normalized to plasma fluorescence intensity. These experiments were repeated three times with similar results. Images were selected as representative of three independent experiments. *P < 0.005. Scale bar, 100 μm.

Leakage assessment with FITC-BSA was used to determine the antipermeability activity of clusterin in retinal vessels with VEGF-induced hyperpermeability. As shown in Figure 2B, with intravitreal injection of VEGF, fluorescein-conjugated dextran, recognized as diffuse fluorescence around retinal vessels, easily infiltrated the vessel walls and diffused into the retina. The clusterin significantly inhibited VEGF-induced leakage from the retinal vessels was confirmed by the clearance of diffuse fluorescence in the retina. FITC-BSA fluorescence intensity was measured by image analysis in serial retinal sections. The average retinal FITC-BSA fluorescence intensity was calculated and normalized to plasma fluorescence intensity. The retinal FITC-BSA fluorescence that was significantly increased (3.22 ± 0.23, P < 0.05) in VEGF-treated retinas compared with that in the control retinas (0.63 ± 0.11) decreased significantly with clusterin treatment (0.51 ± 0.35, P < 0.05). 

Attenuation of the Loss of Tight Junction Proteins in Retinal Endothelial Cells under Diabetic Conditions

To investigate the effect of clusterin on the expression of tight junction proteins in retinal endothelial cells under diabetic conditions, we assessed the expression of ZO-1 and -2 in AGE-treated HRMECs after treatment with clusterin. The levels of ZO-1 and -2 were measured, as in our previous reports. ^6,9,10 As demonstrated in Figure 3A, clusterin effectively prevented loss of ZO-1 and -2 in AGE-treated HRMECs (P < 0.05). In addition, restoration of ZO-1 expression was reconfirmed by immunocytochemistry. In confluent HRMECs, ZO-1 is a cytoplasmic protein arranged along intercellular junctions. In AGE-treated HRMECs, the level of ZO-1 at intercellular junctions markedly decreased, but treatment with clusterin completely blocked the effect (Fig. 3B). 

Figure 3.

 

Clusterin attenuated the loss of tight junction proteins in retinal endothelial cells under diabetic conditions. (A) HRMECs were incubated for 12 hours with or without clusterin (1 μg/mL) in AGE and assayed for the expression of ZO-1 and -2. Each value represents the mean ± SE of results in three independent experiments (*P < 0.05). β-Actin served as the loading control. (B) ZO-1 expression in the intercellular junction was examined by immunocytochemistry. The data and images are representative of results in three independent experiments. Scale bar, 10 μm.

Attenuation of Vascular Leakage in the Diabetic Retina and Restoration of Tight Junction Proteins

To investigate the effect of clusterin on vascular permeability in the diabetic mouse retina, we assessed wholemount retinal preparations from mice with streptozotocin-induced diabetes. Eight days after induction of diabetes, the mice were injected intravitreally with 1 μg/mL of clusterin and then perfused with fluorescein-conjugated dextran for 1 hour. As shown in Figure 4A, clusterin markedly inhibited the diffuse leakage from vessels in the diabetic retinas. FITC-BSA fluorescence was significantly increased in diabetic retinas (3.34 ± 0.37 pixels; P < 0.05) compared with that in control retinas (0.64 ± 0.28), but the fluorescence decreased significantly with the administration of clusterin (1.09 ± 0.33; P < 0.05). To examine the effect of clusterin on levels of the tight junction proteins ZO-1 and -2, we analyzed the proteins in retinas treated with clusterin 8 days after diabetes induction, and expression was found to be restored (Fig. 4B). 

Figure 4.

 

Clusterin attenuated vascular leakage of diabetic retina, accompanied by restoration of tight junction proteins. (A) Vascular leakage in the retina was evaluated by leakage assessment using an FITC-BSA fluorescence assay. At 1 day after intravitreal injection of 1 μg/mL clusterin in 1 μL PBS or of PBS alone in diabetic mice with streptozotocin-induced diabetes of 7 days' duration, wholemount retinas were prepared after a 1-hour perfusion of FITC-BSA. The average retinal FITC-BSA fluorescence intensity was calculated and normalized to the plasma fluorescence intensity. The experiments were repeated three times with similar results. Images are representative of those in three independent experiments. *P < 0.005. (B) At 8 days after streptozotocin injection, retinal proteins of diabetic mice were analyzed by Western blot analysis using antibodies to ZO-1 and -2. Each value represents the mean ± SE of results in three independent experiments (*P < 0.05). β-Actin served as the loading control. (C) Immunohistochemistry for ZO-1 and PECAM-1 was performed in diabetic retinas, with or without intravitreal injection of clusterin (1 μg/mL). Images were selected as representative of those in three independent experiments. GCL, ganglion cell layer; INL, inner nuclear layer. Scale bar: (A) 100 μm; (C) 50 μm.

Furthermore, we addressed whether clusterin could restore expression of tight junction proteins in retinal vessels in diabetic retinas. Immunohistochemistry of ZO-1 for the tight junction proteins and PECAM-1 for the endothelial cell proteins was performed in the diabetic retina after clusterin treatment. ZO-1 expression in the diabetic retinal vessels was markedly decreased compared with that in the control retinas, but it was restored with clusterin treatment (Fig. 4C). 

Discussion

Clusterin is upregulated in response to diverse pathophysiological stresses. ²⁰ Across species, it maintains a high level of sequence homology (70%–80% between mammals). ²¹ Its wide distribution and sequence conservation indicates a fundamental biological importance in its function. However, its role in pathologic conditions is controversial. For instance, clusterin, produced and released by Müller cells, may play an important role in the pathogenesis of ischemic injury in the rat retina, ²² whereas it protects against ischemia-induced retinal endothelial cell death and tight junction protein loss. ⁹  

In addition, the mode of action of clusterin in these situations remains unknown. The exact mechanism of the possible direct effect of clusterin on the BRB is unknown. One hypothesis involves a direct receptor-mediated intracellular penetration of clusterin and intracellular processing for biological activity. For example, LRP-2, the clusterin receptor, mediates the clearance of clusterin through brain endothelial cells, which could be involved in the neuronal accumulation of amyloid-β peptide and clusterin. ²³ However, this effect occurs only via the transcellular transport of clusterin through the barrier, without intracellular processing. Further investigation of the intracellular processing of clusterin is needed. In addition to this direct effect, it has been postulated that clusterin exerts its effects indirectly as an extracellular chaperone. ^13,14 Based on the activity of clusterin as an extracellular chaperone, we have demonstrated that it is upregulated in retinal endothelial cells undergoing hypoxia; and, in developing retinal vessels, it protects cells from stress and promotes cell viability. ⁸ Moreover, we have found that clusterin protects retinal endothelial cells from ischemia-induced apoptotic cell death and loss of tight junction proteins. ⁹ In the course of our research to determine the cytoprotective role of clusterin in retinal endothelial cells, we found that it effectively prevents BRB breakdown in the mouse model of DR. 

Increased vascular permeability in DR is accompanied by a decrease in tight junction proteins in the retinal endothelial cells. ⁴ The ZO family and occludin are well-characterized components of tight junctions in retinal endothelial cells. ^6,8,9 Particularly, the ZO family could cause changes in permeability, because it is closely linked to occludin, whose phosphorylation contributes to the regulation of permeability. ²⁴ ZO-1 and -2 are junctional proteins associated with the cytoplasmic surface of the tight junction and are localized to the points of membrane contact with the fibrils. They are involved in the formation of tight junctions, and their cellular localization is closely related to the permeability of vascular endothelial cells. ²⁵ We have suggested that the ZO family is a component of tight junctions in retinal endothelial cells ^6,8,9 with expression that is inversely related to the permeability of the BRB. ^6,8–10,16 Herein, we demonstrated that clusterin effectively inhibited VEGF-induced hyperpermeability and leakage from vessels of the diabetic retina and restored the expression of ZO-1 and -2. As in our previous reports, these results indicate that ZO-1 and -2 ensure the tightness of the BRB as well as occludin. ²⁶  

During the treatment with clusterin, we observed no systemic toxicity in the mice. In addition, clusterin showed no association with the cell viability of HRMECs and no retinal toxicity up to 20 μg/mL, which is equivalent to 20 times the effective dose (1 μg/mL). This means that clusterin can be safely applied to attenuate diabetic BRB breakdown with restoration of the expression of tight junction proteins without toxic effect on retinal endothelial cells. 

In summary, up to 20 μg/mL clusterin, 20 times the effective therapeutic dose, ^8,9 did not affect the viability of HRMECs or cause toxic effects in the retina. It effectively inhibited VEGF-induced hyperpermeability in HRMECs and the retina. The antipermeability activity of clusterin was related to the restoration of tight junction proteins, as shown in our previous studies. ^6,9 Finally, clusterin reduced leakage from vessels in the diabetic retina, which was accompanied by restoration of tight junction proteins. Based on the available evidence, we suggest that clusterin could attenuate BRB breakdown in DR. Furthermore, it may be applicable to other retinopathies that involve VEGF-mediated BRB breakdown, such as ischemia and inflammation. 

Footnotes

 Supported by Grant 03-2006-017-0 from the Seoul National University Hospital Research Fund and by Bio-Signal Analysis Technology Innovation Program 2009-0090895 of MEST/NRF (Ministry of Education, Science and Technology/National Research Foundation).

Footnotes

 Disclosure: J.-H. Kim, None; J.H. Kim, None; Y.S. Yu, None; B.H. Min, None; K.-W. Kim, None

The authors thank Hyoung-Oh Jun for technical help.