Green Tea Extract Treatment Alleviates Ocular Inflammation in a Rat Model of Endotoxin-Induced Uveitis

Green tea extract (GTE) ingested by rats exerted anti-oxidative activities in various ocular tissues as shown in our previous studies. The present work investigated anti-inflammatory effects of GTE on endotoxin-induced uveitis (EIU). EIU was generated in adult rats by a footpad injection of 1 mg/kg lipopolysaccharide (LPS). Oral administration of GTE (550 mg/kg) was given one, two or four times after LPS injection. Twenty-four hours later, LPS produced severe hyperemia and edema in the iris. Immunocytochemical examinations showed an accumulation of infiltrating cells in the aqueous humor that were immunopositive for cluster of differentiation 43 (CD43) and CD68, markers for leucocytes and macrophages, respectively. Analyses of the aqueous humor showed an increase in pro-inflammatory mediators including tumor necrosis factor-alpha (TNF-α), interleukin-6 (IL-6) and monocyte chemoattractant protein-1 (MCP-1). GTE treatments improved the clinical manifestations and reduced infiltrating cells and protein exudation in the aqueous humor, which were not observed under half dose of GTE (275 mg/kg). The number of CD68 positive macrophages residing in the iris and ciliary was also reduced. GTE suppressed production of TNF-α, IL-6 and MCP-1 in the aqueous humor, which was associated with a down-regulation of LPS receptor complex subunits, Toll-like receptor 4 (TLR-4) and CD14, and suppression of nuclear factor-kappa Bp65 (NF-κBp65) in the iris and ciliary body. Our findings show that GTE is a potent anti-inflammatory agent against the inflammation of EIU, and suggest a potential use in treatment of acute uveitis.


Introduction
Uveitis, an ocular inflammatory condition, accounts for approximately 10-15% cases of total blindness and up to 20% of legal blindness in developed world [1,2]. The goal of uveitis treatment is to suppress inflammation and achieve regression when it occurs [3]. However inflammation can recur with various complications, such as cataract and permanent cumulative damages [4,5]. Administrations of corticosteroid are standard therapeutic strategy, but they have many potential side effects such as intraocular pressure increase, cataract formation and increase in infection susceptibility [6]. Therefore, alternative treatments which are safer and more long lasting are needed.
The rat model of endotoxin-induced uveitis (EIU) has been widely used for evaluating potential ocular anti-inflammatory compounds since it was reported in 1980 [7][8][9][10][11][12]. EIU can be induced by systemic injection of lipopolysaccharide (LPS), which generates inflammatory responses largely in the anterior uvea and mild responses in the posterior segments of the eye, mimicking the pathological conditions in human acute uveitis [12][13][14]. It has been reported that LPS was recognized by membrane-bound cluster of differentiation 14 (mCD14) and Toll-like receptors (principally TLR-4) on the surface of macrophages. Receptors activation in these immune surveillance cells resulted in phosphorylation of nuclear factor-kappa B (NF-kB) and caused release of pro-inflammatory factors, such as tumor necrosis factor-a (TNF-a), interlukin-6 (IL-6) and monocyte chemoattractant protein-1 (MCP-1) [15][16][17][18][19][20][21]. As a crucial proximal mediator, TNF-a stimulates acute phase reaction of inflammation by influencing leukocyte activation and infiltration, and inducing production of other mediators such as IL-6, a major cytokine regulator of acute phase response [22,23]. MCP-1, a powerful chemotactic and activating factor [24], stimulates the activation of mitogen-activated protein kinases (MAPKs) to promote monocytes migration [25]. Another study shows that the blood-humor barrier is broken down two hours after LPS injection [26], which results in migration of polymorphonuclear and mononuclear cells into the aqueous humor. The inflammatory process reaches peak level at 18-24 hours after LPS injection [27].
Catechins, the major component in green tea extract (GTE), have been shown to exert anti-oxidative, anti-inflammatory, antiangiogenic and anti-carcinogenic effects [28,29]. In a pharmacokinetic study, we have shown that catechins reach peak level in the ocular tissues of normal rats at 1 to 2 hours after ingestion, and produce significant reduction in oxidative stress within these tissues [30]. We have addressed in this study the hypothesis that orally administered GTE could serve as effective anti-inflammatory agents to alleviate inflammatory responses in anterior segments of the eye triggered by a systemic injection of LPS.

Endotoxin-induced uveitis and GTE treatment
All experiments were conducted according to the Association for Research in Vision and Ophthalmology (ARVO) statement on the use of animals. Ethics approval for this study was obtained from the Animal Ethics Committee of the Chinese University of Hong Kong. Sprague-Dawley rats (about 250 g, 6-8 weeks old) were obtained from the Laboratory Animal Service Center of the Chinese University of Hong Kong. Ethics approval for this study was obtained from the Animal Ethics Committee of the University. All animals were housed at 25uC with 12/12 hour light-dark cycles, and were allowed to access freely to food and water. Before the experiment, animals were fasted overnight and body weight was recorded.
EIU was induced by injection of 0.1 mL of pyrogen-free saline dissolved LPS (from Salmonella typhimurium; Sigma Chemical, St. Louis, MO, USA) at the dose of 1 mg/kg into one footpad. The dosage was selected according to results of a preliminary study, which showed that LPS at 1 mg/kg was the optimal dose in inducing moderate inflammation in both eyes without causing obvious lesion in the liver and kidney. The GTE Theaphenon E was kindly provided by Dr. Y. Hara, which contains EGCG (epigallocatechin gallate, .65%), EGC (epigallate catechins, ,10%), EC (epicatechin, ,10%) and ECG (epicatechin gallate, ,10%) and other trace catechin derivatives. It was prepared as a 550 mg/kg GTE suspension in 0.5 mL distilled water and was fed intragastrically into the rat.
The rats were randomly divided into three treatment groups: i) GTE1, fed with GTE two hours after LPS injection (LPS+GTE1, n = 6); ii) GTE2, fed with GTE twice at two and eight hours after LPS injection (LPS+GTE2, n = 6); iii) GTE4, fed with GTE four times at two, five, eight and eleven hours after LPS injection (LPS+ GTE4, n = 6). Control groups consisted of: i) normal control, footpad injected with saline and fed with water two hours after injection (Saline+water, n = 3); (ii) LPS controls, footpad injected with LPS and fed with water (LPS+water, n = 6); (iii) Dxm controls, footpad injected with LPS and fed with Dexamethasone (Dxm) (1 mg/kg, distilled water suspension; Sigma Chemical, USA) two hours after LPS injection (LPS+Dxm, n = 6); and (iv) GTE controls, footpad injected with saline and fed with GTE four times as in GTE4 group (Saline+GTE4, n = 3). Another eighteen rats were used for histological studies.
In another experiment, the dosage effect of GTE was tested in 23 rats: i) normal control (n = 5), footpad injected with saline followed by oral administration of water at the 2 nd and 11 th hours after footpad injection; ii) LPS group (n = 6), footpad injection of LPS followed by feeding of water at the 2 nd and 11 th hour after footpad injection; iii) 550 mg/kg GTE (n = 6), oral administration of the dose at the 2 nd and 11 th hour after footpad injection of LPS; iv) half dose GTE (n = 6), oral administration of 275 mg/kg GTE at the 2 nd and 11 th hour after the LPS injection.
Twenty four hours after LPS injection, the rats were anesthetized with intraperitoneal injection of 4.0 mL of ketamine-xylazine mixture (1.5:1, Alfasan International B.V., Holland) for collections of ocular tissues. They were terminated immediately by drawing the whole blood through heart puncture.

Clinical manifestations scoring
Clinical features of ocular inflammation in both eyes were evaluated using a slit lamp and graded from score 0 to score 4 by a masked observer 24 hours after LPS injection as described previously [31]. The grading is assigned as: 0 = no obvious inflammatory response; 1 = discrete dilation of iris and conjunctival vessels; 2 = moderate dilation and iris and conjunctival vessels with moderate flare in the anterior chamber; 3 = intense iridal hyperemia with intense flare in the anterior chamber; 4 = same clinical signs as 3 with presence of fibrinoid exudation and miosis.

Histological examination of infiltrating cells
Under deep anesthesia, the rats were perfused intracardially with 0.01 M sterile phosphate buffer saline (PBS) followed by 4% paraformaldehyde. Both eyes were removed and immersed in 10% formalin for 24 hours at room temperature. Some connective tissues were maintained on each eye to facilitate orientation. The eyes were embedded in paraffin and sectioned in 5 mm thickness along the vertical meridian and the optic nerve head. After deparaffinization and rehydration, the sections were stained with Hematoxylin and Eosin (H&E). The anterior and posterior segments of the eyeball were examined under a light microscope (DMRB, Leica Microsystems, Wetzlar, Germany). Total number of infiltrating cells located in anterior segments (anterior/posterior chamber) and posterior segments (vitreous body around optic nerve head) were counted in a masked fashion by an ocular pathologist as described previously [32].
Some sections were selected for immunostaining in order to determine the identity of infiltrating cells. The slides were heat to induce epitope retrieval using a pressure cooker (Biocare Medical, Walnut Creek, CA). After blocking with 0.1% bovine serum at room temperature, mouse anti-rat monoclonal antibody CD43 (1:80 dilution, AbD Serotec, Kidlington, UK) or CD68 (1:100 dilution, AbD Serotec, Kidlington, UK) was applied separately to the sections and incubated at 4uC overnight, which binds . Control sections were processed as above without primary antibody. In another experiment, sections of the eye were processed for immunostaining of CD68 in order to determine whether GTE treatment reduced LPS-induced accumulation of macrophages in the stroma of iris and ciliary body. Four non-consecutive sections were collected from the eye of normal controls, and animals treated with LPS alone, LPS plus GTE or dexamethasone. The number of CD68 positive cells inside stroma of iris and ciliary body was counted in a masked fashion.

Cell count and protein assay in aqueous humor
Aqueous humor was collected by piercing the anterior chamber with a 30 gauge needle. One microliter aqueous humor was diluted with 9 ml 0.01 M PBS and suspended in an equal volume of Trypan-blue solution. The number of cells was counted by using a hemacytometer under a light microscope. Another portion of the aqueous humor was centrifuged at 2500 rpm for 15 minutes at 4uC. Cells-free supernatant was used for total protein assay in duplicate (Bio-Rad, Hercules, CA, USA). Blood samples were collected by heart puncture and clotted at room temperature for 2 hours in serum vial. They were centrifuged at 2500 rpm for 15 minutes at 4uC. The serum and cells-free aqueous humor were taken for determination of TNF-a and IL-6 (rat-ELISA kit, R&D Systems, Minneapolis, USA) and MCP-1 (rat-ELISA kit, Invitrogen, Camarillo, CA, USA), each in duplicate.

Quantification of CD14 and TLR-4 mRNA expression
Iris, ciliary body/process and retina were collected. After washing with 0.01 M cold sterile PBS, tissues were immersed in 300-500 ml Trizol reagent (Invitrogen, USA) and stored at 280uC until use. Total RNA were isolated and treated with RNase-free DNaseI according to the manufacturer's protocol (Qiagen, Hilden, Germany). RNA (0.5-1 mg) was reversely transcribed into cDNA using SuperScript III reverse transcriptase (Invitrogen, USA). PCR was performed using iCycler PCR instrument (Bio-Rad) and lightCycle 480 II real-time PCR system (Roche Applied Science, Penzberg, Germany). Sequences of the gene-specific primers were designed using the online Primer 3 Input Program (version 0.4.0) ( Table 1).
Quantitative real-time PCR (qPCR) was performed with SYRB green PCR mixture containing 10 ml of 26480 SYRB green I Master (Roche, USA), 0.55 ml of 10 nM primers (Invitrogen), 1 ml cDNA and 8 ml double distilled water. The parameters were as follows: pre-incubation 95uC for 10 minutes, followed by 40 amplification cycles each with denaturation at 95uC for 15 seconds, annealing/extension at 60uC for 1 minute. The melt-curve analysis was performed using default settings of the instrument. Final qPCR products were electrophoresed in 2.0% agarose gel. Relative CD14 and TLR-4 mRNA expression of each sample was calculated as described [33]. In brief, after the threshold cycle value of the target gene (C T target) was emendated with the value of internal control gene GAPDH (C T GAPDH ), the normalized expression value (2 2(CT target2CT GAPDH) ) was calculated, then the expression fold change due to treatment were obtained according to the normalized expression value.

Statistical analysis
All data were analyzed by nonparametric Kruskal-Wallis Tests. Two group comparisons were done by Mann-Whitney Tests. The p-value less than 0.05 was considered statistically significant. All statistical analyses were performed using SPSS statistical software package (version 20.0).

Clinical manifestations of inflammation in the eye
Slit lamp examination showed development of hyperemia and edema associated with miosis and fibril formation in the iris 24 hours after LPS injection (Fig. 1B-C). These clinical features were not observed in the normal controls (Fig. 1A), and appeared less severe in the GTE treated rats (Fig. 1D-E) and rats treated with dexamethasone (Fig. 1F). Quantitative evaluation of these clinical scores showed a significant reduction in animals treated with 1 to 4 times oral administration of GTE (p,0.05) when compared with animals fed with water after LPS insult (Fig. 1G).

Infiltrating cells in anterior segments of the eye
Injection of LPS induced substantial accumulation of infiltrating cells in the anterior chamber and posterior chamber of the eye ( Fig. 2A). Giemsa staining showed that these infiltrating cells in aqueous humor were variable in size and polymorphic in nuclear shape (Fig. 2B). Immunohistochemical studies on paraffin sections of LPS treated eye showed that most infiltrating cells in anterior and posterior chamber were CD43 positive leucocytes (Fig. 2D-E), whereas a few were CD68 positive macrophages (Fig. 2G-H).
A few CD43 and CD68 positive cells were found in the stroma of iris and ciliary body ( Fig. 2D and G). No staining was detected in sections processed with the absence of primary antibody (Fig. 2C and F).
The effects of GTE treatment on ocular inflammation were investigated in paraffin sections of the eye. LPS induced an accumulation of infiltrating cells and proteineous substances in the anterior and posterior chamber (Fig. 3B), which was not observed in normal control animals (Fig. 3A). Treatment of GTE reduced substantially the infiltration of cells and accumulation of protein in the anterior and posterior chamber (Fig. 3C), which was further reduced in animals treated with Dxm (Fig. 3D). Similar observations were presented in their corresponding vitreous body around the optic nerve head (Fig. 3E-H), where the accumulation of infiltrating cells was highly reduced in both GTE (Fig. 3G) and Dxm (Fig. 3H) treated animals. Counting of infiltrating cells on these sections (in anterior and posterior chamber, and in vitreous  Table 2. TNF-a, IL-6 and MCP-1 concentration in the serum after different treatments.  Figure 5. Effects of GTE on cytokine and chemokine production in aqueous humor. The pro-inflammatory factors TNF-a (A), IL-6 (B) and MCP-1 (C) were all reduced significantly after GTE treatments when compared with LPS+water (***p,0.01). Within group comparisons, however, did not show obvious differences among GTE treated groups (p.0.05). Sensitivity of the assays: 5 pg/mL for TNF-a, 21 pg/mL for IL-6, less than 8.0 pg/mL for MCP-1. Saline+water, n = 3; LPS+water, n = 6; LPS+GTE1, n = 6; LPS+GTE2, n = 6; LPS+GTE4, n = 6; LPS+Dxm, n = 6; Saline+water, n = 3. Data were shown as mean 6 SE. doi:10.1371/journal.pone.0103995.g005 space) showed a significant reduction in GTE treated animals (p, 0.05) when compared with the vehicle treated LPS group (LPS+ water) (Fig. 3I).

Cell infiltration and protein exudation in aqueous humor
The changes in infiltrating cells and protein content in aqueous humor were also analyzed. Treatments of GTE produced significant decrease in infiltration cells (p,0.05) and accumulation of protein (p,0.05) when compared with vehicle treated LPS group (Fig. 4A-B), agreeing with the results in histological analyses, though these effects were not as potent as Dxm (p, 0.01). No differences were observed among animals treated with one, two or four times of oral administrations of GTE, agreeing with the findings in the histological study. The reason of lack of effect for multiple dosage was unclear, but may be caused by the rapid decline of anti-oxidative property of GTE in the aqueous humor [30]. Treatment with GTE alone without injection of LPS produced neither an accumulation of cells nor an increase in protein exudation in the aqueous humor.
The dosage effect was tested in another experiment, in which the rats were fed with either 550 or 275 mg/kg GTE, 2 and 11 hours after LPS injection. The results showed that both infiltrating cell number and protein level in aqueous humor were significantly reduced at the dose of 550 mg/kg (p,0.05), but the effect became insignificant when the dose was reduced to half ( Fig. 4C-D), indicating that 550 mg/kg is an effective dose in suppressing these inflammatory responses.

TNF-a, IL-6, and MCP-1 in aqueous humor
The pro-inflammatory factors TNF-a, IL-6 and MCP-1 were determined in the aqueous humor using ELISA. LPS caused a surge in TNF-a, IL-6 and MCP-1, which were barely detected in normal controls (Fig. 5A-C). All GTE treatments caused a significant reduction of these factors when compared with the LPS group (LPS+water) (p,0.05). However, no significant difference was detected within the GTE groups (p.0.05), suggesting that multiple treatments did not produce additional effect in reducing these pro-inflammatory factors. The effect of Dxm was more potent than both GTE1 and GTE2 only for MCP-1 (GTE1 vs Dxm, p = 0.007; GTE2 vs Dxm, p = 0.002) but not for TNF-a and IL-6. GTE alone did not produce any detectable change in these factors. Analyses of the serum showed no significant reduction of pro-inflammatory factors in animals treated with either GTE when compared with those treated with LPS and vehicle ( Table 2). Dexamethasone treatment caused a significant reduction in MCP-1 (p,0.01) but not in IL-6. While TNF-a was beyond the detectable level in the serum, both GTE and Dxm treatments could not reduce the level of IL-6 and MCP-1 to that of normal control.
Expression of CD14 and TLR-4 mRNA, and NF-kBp65 activity in ocular tissues Quantitative real-time PCR was used to determine the expression of TLR-4 and CD14, the respective receptor and coreceptor for LPS, after LPS injection and GTE treatment ( Table 3). Expression of both TLR-4 and CD14 were up-regulated after LPS insult in the iris and ciliary body (Fig. 6A) and in the retina (Fig. 6B). These elevated expressions were suppressed significantly by GTE or Dxm treatment (p,0.01, Fig. 6A-B). GTE treatment alone without LPS injection did not caused obvious changes in TLR-4 and CD14 expression when compared with normal control. Table 3. CD14 and TLR-4 mRNA expression in the iris and ciliary body and the retina by quantitative real-time PCR analyses. The level of nuclear NF-kBp65, a downstream molecule of TLR-4 signaling, was examined in the iris and ciliary body after LPS and GTE treatments. Western blot showed that this protein was upregulated in the tissues after LPS injection, and its expression was decreased substantially after treatment of GTE4 and Dxm (Fig. 6C). Measurements of band intensity with reference to that of Lamin B confirmed a significant reduction of nuclear NF-kBp65 after GTE4 or Dxm treatment (p,0.05) (Fig. 6D).

Effect of GTE on macrophages within iris and ciliary body
To investigate the effect of GTE treatment on the macrophages accumulating in the iris and ciliary body, sections of the eye were immunostained with antibody against CD68. In normal control only a few CD68 positive macrophages were observed in the stroma of iris and ciliary body (Fig. 7A). The number of macrophages was increased substantially after LPS treatment ( Fig. 7B and E), and was reduced significantly after treatment of either GTE4 or Dxm (p,0.05) (Fig. 7C-E).

Discussion
Results of this study show that oral administration of GTE, at the dose of 550 mg/kg, produces anti-inflammatory effects against LPS-induced ocular inflammation. The major findings include: i) GTE alleviates clinical manifestations of ocular inflammation; ii) it reduces infiltration of leukocytes and macrophages, and leakage of protein into aqueous humor and vitreous body; iii) it also suppresses production of pro-inflammatory biomarkers TNF-a, IL-6 and MCP-1 in aqueous humor; iv) these anti-inflammatory effects are associated with down-regulation of LPS receptors, TLR-4 and CD14, and reduction of nuclear NF-kBp65, v) GTE also reduces accumulation of macrophages in the stroma of iris and ciliary body. These results provide evidence for the first time that GTE is a potent anti-inflammatory agent for acute ocular inflammation.
We have shown in this study that the levels of pro-inflammatory mediators, TNF-a, IL-6 and MCP-1, were elevated significantly in the aqueous humor twenty-four hours after LPS injection. TNF-a has been shown in rats to be the major player in LPS-induced leukocyte adhesion, vascular leakage and cell death in the rats' eyes [34]. In mice deficient of MCP-1, inflammatory responses are reduced in a model of EIU [32]. The surge of infiltrating leukocytes and macrophages in the aqueous humor is thus likely caused by an increased production of pro-inflammatory factors as a response to the LPS insult. LPS also induces an accumulation of macrophages in the stroma of iris and ciliary body, either by recruitment of circulating macrophages or activation of resident cells in these tissues. GTE exerts its anti-inflammatory actions through a suppression of production of these pro-inflammatory factors, and thereby reducing the infiltration of leucocytes and macrophages and exudation of protein into the aqueous humor, and recruitment or activation of macrophages residing in the iris and ciliary body. These anti-inflammatory effects seem to be specific to ocular tissues as shown by the results that significant reduction of IL-6 and MCP-1 after GTE treatment is observed only in the aqueous humor but not in the serum. The specific changes agree with previous findings that TNF-a, IL-6 and MCP-1 are elevated in aqueous humor of patients with infectious or noninfectious uveitis but not in the serum [35,36]. Reduction of TNF-a was detected only in the aqueous humor but not in the serum, possibly because its properties of brief production [37] and high hepatic clearance [38] in the circulatory system.
Juxtaposition of CD14 and TLR-4 in the anterior uvea contributes to the sensitivity of iris and ciliary body to LPS [39]. CD14 does not have a trans-membrane segment and so needs TLR-4, a trans-membrane protein, to transduce the LPS stimulation. LPS binds primarily to resident macrophages or epithelial cells lining the iris and ciliary processes [39,40]. When LPS-CD14-TLR4 cluster activation is inhibited by lipid raftdisrupting drug or lipid A mimetic antagonists, LPS-induced cellular cascading reaction is interrupted [41,42]. We have shown in this study that GTE suppresses LPS induced elevation of CD14 and TLR-4 mRNA in the anterior uvea and the retina, suggesting that GTE treatment may alleviate ocular inflammation by suppressing formation of the LPS receptor complex. These findings are consistent with previous reports showing that EGCG is able to block the interaction between LPS and TLR-4 and downregulate TLR-4 mRNA expression in vitro [43,44]. However, other receptor components, such as HSP 70/90 (heat-shock proteins 70 and 90), CXCR4 (chemokine receptor 4), and DAF (decay-accelerating factor) also participate in the LPS-CD14-TLR4 cluster activation, probably acting as additional LPStransfer molecules [45]. Moreover, EGCG has also been shown to suppress inflammatory responses by binding to the surface molecule 67-kDa laminin receptor (76LR), which results in reduction of expression of pro-inflammatory mediators, such as TNF-a, IL-6 and cyclooxygenase-2 [43]. Whether GTE may modulate function of these molecules remains to be investigated.
Anti-inflammatory properties of GTE have been reported in other experimental disease models. Epigallocatechin-3-gallate (EGCG), the major constituent of GTE, has shown to reduce expression of inflammatory biomarkers in human cell lines such as chondrocytes and corneal epithelium, and experimental models of dry eye [46][47][48]. These effects are related to a suppression of IkB kinase-b (IKKb) and TANK-binding kinase-1 (TBK1) [49], or inhibition of the proteasome-mediated degradation pathway that induces accumulation of NF-kB inhibitors, p27 Kip1 and IkB-a [50]. We have shown in the current study that LPS induced nuclear NF-kBp65 is significantly reduced after GTE treatment, probably caused by a reduced level of TLR-4-CD14 receptor complex, and leads eventually to a reduced expression of genes coding for cytokines and chemokines.
To our knowledge, this is the first report to demonstrate antiinflammatory effects of GTE on acute ocular inflammation. The findings in this study support strongly that GTE is a potent therapeutic agent for treatment of acute anterior uveitis.