Flavonoids Protect Retinal Ganglion Cells from Oxidative Stress–Induced Death

Abstract

purpose. Mounting evidence suggests that oxidative stress contributes to the pathogenesis of many ocular diseases, including glaucoma and diabetic retinopathy. No treatments are available to prevent the neuronal degeneration that occurs in these disorders. The purpose of this study was to determine whether flavonoids, natural products that are abundant in fruits and vegetables, can protect retinal ganglion cells from oxidative stress–induced death.

methods. The ability of flavonoids to protect an immortalized retinal ganglion cell line, RGC-5 cells, was tested using three model systems of oxidative stress–induced cell death: glutathione (GSH) depletion, t-butyl peroxide (t-BOOH) treatment, and hydrogen peroxide (H₂O₂) treatment.

results. GSH depletion causes retinal ganglion cell death by a pathway involving the production of endogenous reactive oxygen species (ROS), whereas the other two treatments use exogenous sources of ROS. It was found that specific flavonoids could protect retinal ganglion cells from cell death initiated by all three of the inducers of oxidative stress with high levels of potency and low toxicity. In the case of GSH depletion, different flavonoids could be shown to act at distinct steps in the cell death pathway. Several of the protective flavonoids also induced the synthesis of the transcription factor NF-E2–related factor 2 and phase 2 antioxidant enzymes such as heme oxygenase 1, markers for the activation of the antioxidant response element.

conclusions. These results suggest that certain flavonoids can function as potent and effective neuroprotective agents for retinal ganglion cells.

Flavonoids (Fig. 1)comprise a large family of plant-derived polyphenolic compounds widely distributed in fruits and vegetables and, therefore, regularly consumed in the human diet. ^{1 2 3} They are particularly abundant in beverages derived from plants such as tea, cocoa, and red wine. A number of physiologic benefits have been attributed to flavonoids, including protection from cardiovascular disease and cancer, and many of these effects are thought to stem from their potent antioxidant and free-radical scavenging properties. Flavonoids have been shown to reduce the growth of malignant carcinomas and to inhibit tumor angiogenesis, and they are thought to be responsible for the low rates of cancer and atherosclerosis in people who consume soy-based diets. ^{1 3}  

Flavonoids have multiple properties that are potentially of benefit for the prevention and treatment of ocular diseases, particularly those that involve the loss of nerve cells. First, flavonoids can protect central nervous system (CNS)–derived nerve cells from oxidative stress–induced death, ⁴ which has been linked to the nerve cell loss seen in multiple ocular diseases. ^{5 6 7} Second, several studies have shown that certain flavonoids can induce the activity and expression of phase 2 detoxification proteins. ^{8 9 10} The phase 2 detoxification proteins include enzymes associated with glutathione (GSH) biosynthesis and metabolism and redox-sensitive proteins such as heme oxygenase 1 (HO-1). ¹¹ The transcriptional activation of these and other genes encoding phase 2 detoxification proteins is mediated by a cis-acting enhancer termed the antioxidant response element (ARE). By inducing the expression of antioxidant defense enzymes, these flavonoids have the potential to have long-lasting effects on cellular function. This, in turn, could be highly beneficial to cells exposed to chronic oxidative stress. In addition, flavonoids have been shown to induce neurite outgrowth, ¹² reduce inflammation, ¹³ and inhibit endothelial cell proliferation, ¹⁴ all properties that could have additional benefits for the treatment of ocular diseases. Finally, in animal studies, flavonoids have generally shown low levels of toxicity over a wide range of doses. 

Recently, we characterized the molecular and cellular events underlying oxidative stress–induced cell death in retinal ganglion cells using RGC-5 cells, an immortalized retinal ganglion cell line. ¹⁵ Although in the course of these studies we identified several compounds that were able to protect the cells from oxidative stress–induced death, for a variety of reasons, including significant levels of toxicity, none of these compounds is likely to be useful for the long-term treatment of ocular diseases. 

Because flavonoids have a number of qualities that could make them useful for the long-term treatment of ocular diseases, particularly those that involve the loss of retinal ganglion cells, we examined their ability to protect the immortalized retinal ganglion cell line, RGC-5 cells, from oxidative stress–induced death. The results presented here are the first demonstration that specific flavonoids can prevent oxidative stress–induced death in retinal ganglion cells with high potency and low toxicity. 

Methods

Chemicals

2′,7′-Dichlorodihydrofluorescein diacetate (H₂DCF-dA), indo-acetoxymethylester (Indo-1), Pluronic 127, 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethyl-benzimidazolcarbocyanine iodide (JC-1), DAPI, and rhodamine-123 were purchased from Molecular Probes (Eugene, OR). Flavonoids were purchased from Alexis or Sigma/Aldrich (St Louis, MO). All other chemicals were from Sigma. 

Cell Culture and Viability Assays

RGC-5 cells were obtained from N. Agarwal ¹⁶ and were grown on tissue culture dishes in Dulbecco’s modified Eagle’s medium (DMEM)–low glucose supplemented with 10% fetal calf serum (FCS). Cell viability was determined by a modified version of the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay based on the standard procedure. ¹⁷ To obtain consistent results with RGC-5 cells, the cells were plated into 35 mm dishes at 7 × 10⁴ cells/dish in complete medium. The next day, the medium was replaced with DMEM-low glucose supplemented with 5% dialyzed FCS (DFCS), and the cells were treated with the flavonoids for 30 minutes before the addition of the oxidant (10 mM glutamate + 500 μM BSO; 0.5 mM t-butyl peroxide [t-BOOH] or 650 μM H₂O₂) except for the experiments shown in Figure 5 . In preliminary studies, a range of doses from 1 to 100 μM for each flavonoid was tested. Later studies used the dose of each flavonoid that was most effective at preventing cell death. Twenty-four hours after the addition of the oxidant, the cell culture medium in each dish was aspirated and replaced with 500 μL DMEM–low glucose with 5% DFCS containing 2.5 μg/mL MTT. After 4 hours of incubation at 37°C, the cells were solubilized by the addition of 500 μL of a solution containing 50% dimethylformamide and 20% SDS (pH 4.7). The absorbance at 560 nm was measured on the following day with a microplate reader (Flow Titertec Multiskan PLUS, Mk 11; ICN, Costa Mesa, CA). Results obtained from the MTT assay correlated directly with the extent of cell death as confirmed visually (see, for example, Fig. 2 ). Controls using dishes without cells and cells without the protective agents were used to determine the effects of the agents on the assay chemistry or cell viability, respectively. 

Microscopy

A light microscope (Inverted Microscope Diaphot-TMD; Nikon, Tokyo, Japan) equipped with a phase-contrast condenser (Phase contrast-2 ELWD 0.3; Nikon), a 10× objective lens, and a digital camera (Coolpix 990; Nikon) was used to capture the images with the manual setting. 

Total Intracellular GSH/GSSG

Total intracellular GSH/oxidized GSH (GSSG) was determined using whole cell lysates from untreated, glutamate plus BSO-treated, flavonoid-treated and glutamate plus BSO + flavonoid-treated cells, as described ¹⁵ and normalized to total cellular protein. For each flavonoid, the concentration most effective at preventing cell death was used. 

Reactive Oxygen Species Level

The intracellular accumulation of reactive oxygen species (ROS) in the RGC-5 cells was determined using H₂DCF-dA, as described. ¹⁵ The increase in ROS levels after 18 hours of treatment with 10 mM glutamate plus 500 μM BSO was compared with the increase after treatment with glutamate plus BSO in the presence of the different effective flavonoids. For each flavonoid, the concentration most effective at preventing cell death was used. Cells treated with flavonoids alone were used to determine the effects of the flavonoids on the assay chemistry. Results represent the average ± SD of 2 to 3 independent experiments, with each treatment performed in duplicate. 

Calcium Measurement

The cytosolic accumulation of calcium in the RGC-5 cells was measured using the dye indo-acetoxymethylester (Indo-1) and flow cytometry, as described. ¹⁵ The increase in calcium levels after 19 hours of treatment with 10 mM glutamate plus 500 μM BSO was compared with the increase after treatment with glutamate plus BSO in the presence of the different effective flavonoids. For each flavonoid, the concentration most effective at preventing cell death was used. Calcium data were plotted as dot plots, and fluorescence above a certain ratio was designated as high. The number of cells with high fluorescence was calculated for each sample, and the ratiometric increase of cells with high calcium was determined with respect to the control value. Cells treated with flavonoids alone were used to determine the effects of the flavonoids on the assay chemistry. 

Immunoblotting

For immunoblotting of HO-1, untreated and flavonoid-treated RGC-5 cells from cultures of the same density used for the cell death assays were washed twice in cold phosphate-buffered saline (PBS) then scraped into lysis buffer containing 50 mm HEPES, pH 7.4, 150 mm NaCl, 50 mm NaF, 1.5 mm MgCl₂, 1 mm EGTA, 10% glycerol, 1% Triton X-100, 10 mm sodium pyrophosphate, 1 mm Na₃VO₄, 1 mm phenylmethylsulfonyl fluoride (PMSF), 15 μg/mL aprotinin, 1 μg/mL pepstatin, and 5 μg/mL leupeptin. Lysates were incubated at 4°C for 30 minutes, then cleared by centrifugation at 14,000 rpm for 10 minutes For immunoblotting of NF-E2–related factor 2 (Nrf2), nuclear extracts were prepared as described ¹⁸ from untreated and flavonoid-treated cells. For each flavonoid, the concentration most effective at preventing cell death was used. Protein concentrations were determined (BCA protein assay; Pierce, Rockford, IL). Equal amounts of protein were solubilized in 2.5× SDS sample buffer, separated on 10% SDS-polyacrylamide gels, and transferred to nitrocellulose. Transfers were processed as described ¹⁵ and were developed with a reagent (Super Signal; Pierce). Primary antibodies used were anti-Nrf2 (SC13032; 1/1000) from Santa Cruz Biotechnology (Santa Cruz, CA) and anti-heme oxygenase-1 (HO-1) (SPA-896; 1/5000) from Stressgen. 

Statistical Analysis

Experiments presented were repeated at least two times with duplicate or triplicate samples. Data are presented as the mean ± SD. An unpaired Student’s t-test was used to compare the data obtained. 

Results

Because the sources of oxidative stress in RGCs in vivo are unknown, we used three different models of oxidative stress in combination with the RGC-5 cell line to examine the ability of different flavonoids to prevent oxidative stress–induced death. Two of these models, t-BOOH treatment and H₂O₂ treatment, use the addition of an exogenous source of ROS to induce oxidative stress, whereas the third model, induction of oxidative stress by GSH depletion, relies on ROS produced endogenously by mitochondria. To determine whether flavonoids can protect the RGC-5 cells from death in any of these three different models of oxidative stress, the cells were exposed to 10 mM glutamate plus 500 μM BSO to induce GSH depletion, 0.05 mM t-BOOH, or 650 μM H₂O₂ ¹⁵ in the presence of representatives from several of the different classes of flavonoids (Fig. 1)at concentrations shown to be effective in CNS-derived nerve cells, ⁴ Twenty-four hours later, later cell viability was examined by light microscopy. These concentrations of oxidants were chosen because they induce similar levels of cell death (∼85%) in the RGC-5 cells. As shown in Figure 2 , different flavonoids showed distinct patterns of protection in the three models of oxidative stress. The flavonol quercetin was able to protect the RGC-5 cells from death induced by all three oxidative insults. In contrast, the flavanone taxifolin prevented the cell death induced by GSH depletion and t-BOOH but not by H₂O₂ treatment, whereas kaempferol was effective only against oxidative stress induced by GSH depletion. 

To directly compare the protective efficacy of different flavonoids in the three different models of oxidative stress, RGC-5 cells were exposed to 10 mM glutamate plus 500 μM BSO, 0.05 mM t-BOOH, or 650 μM H₂O₂ in the presence of varying concentrations of each flavonoid and after 24-hours cell viability was determined with the MTT assay. For these studies, we tested multiple members of each of the six classes of flavonoids. Whereas some flavonoids—such as the flavones baicalein and luteolin, the flavonols 3,6-dihydroxyflavone, 3,7-dihydroxyflavone, galangin, fisetin, and quercetin, and the flavanone eriodictyol—were effective at protecting the RGC-5 cells in all three models of oxidative stress (Fig. 3and Table 1 ), other flavonoids were only effective against oxidative stress induced by GSH depletion and t-BOOH treatment (taxifolin, isorhamnetin) or just GSH depletion (3-hydroxyflavone, apigenin, kaempferol, morin, genistein, epicatechin). A number of flavonoids, including dietary flavonoids such as myricetin and epigallocatechin gallate (EGCG), were completely ineffective against all three oxidative insults. As shown in Figure 3 , flavonoids that were protective in all three models of oxidative stress showed several different dose–response patterns. In general, lower concentrations of the effective flavonoids were required to provide significant protection against oxidative stress induced by GSH depletion than against that induced by treatment with either t-BOOH or H₂O₂. However, some of the most effective flavonoids, such as 3,6-dihydroxyflavone, 3,7-dihydroxyflavone, and luteolin, protected in the low micromolar range against all three oxidative insults (Table 1) . We also examined the toxicity of those flavonoids protective against oxidative stress. Many of the protective flavonoids showed a 5- to 10-fold difference between EC₅₀ and LD₅₀ (Table 1) . Exceptions included apigenin, isorhamnetin and genistein, each of which showed significant toxicity at concentrations within the same range of the protective concentrations. 

In the clinical setting, treatment for oxidative stress–induced injury may frequently commence after the stress has begun. Thus, for a flavonoid to have the best chance of being effective in vivo, it should be possible to add it after the initial insult and still achieve protection. Therefore, the latest times at which the most effective flavonoids could be added to the cells were determined and compared with the protection afforded by simply removing the insult. In the case of glutamate plus BSO, the cells are only fully committed to death after 14 to 16 hours of treatment (Fig. 4A) , a significant window of time in which to evaluate the ability of the flavonoids to rescue the cells from death. In contrast, commitment to t-BOOH–induced death requires only 3 to 4 hours of treatment (Fig. 4B)and commitment to H₂O₂-induced death requires just 1 to 2 hours of treatment (Fig. 4C) . We found that one of the protective flavonoids could be added to the RGC-5 cells as late as 16 to 18 hours after the addition of glutamate plus BSO and still provide significantly greater protection from death than that afforded by simply changing the medium (Fig. 4A) . Consistent with its ability to block the late influx of calcium, 3,7-dihydroxyflavone protected even when added 18 hours after the initial insult. In contrast, with t-BOOH, there was little or no difference between the effect of removing t-BOOH from the cells or adding flavonoids (Fig. 4B) . Interestingly, with H₂O₂, several of the flavonoids appeared to provide protection greater than that seen with the removal of the insult at 1 hour after the addition of H₂O₂ but not at later times (Fig. 4C) . Taken together, these data indicate that flavonoids can protect some of the cells when added after the insult, but the degree of this protection is highly dependent on the precise nature of the initial insult and on the flavonoid. Thus, although with glutamate plus BSO at least one of the flavonoids could rescue cells, with t-BOOH and H₂O₂ the protection afforded by the postinsult addition of flavonoids generally did not differ significantly from that seen when the insult was simply removed. 

We also examined the mechanisms underlying protection by the flavonoids using the model of oxidative stress induced by GSH depletion. The initiating event in the death of the RGC-5 cells induced by the combination of glutamate plus BSO is the loss of GSH from the cell. ¹⁵ After the loss of GSH, there is a gradual increase in the level of ROS that can eventually reach 50- to 100-fold the level in untreated cells. The peak of ROS is followed shortly by a sharp increase in intracellular calcium and cell death. In CNS-derived nerve cells, different flavonoids can block each of these steps. ⁴ To determine at which step the protective flavonoids were acting to block oxidative stress–induced death in the RGC-5 cells, we looked at the effects of the most effective neuroprotective concentrations of each of the protective flavonoids on GSH loss, ROS accumulation, and calcium influx after treatment with glutamate plus BSO. As expected from the design of the assay, none of the flavonoids prevented the initial loss of GSH (Table 2) . 

A number of the flavonoids significantly reduced the accumulation of ROS (Fig. 5) . The ability of the flavonoids to prevent ROS accumulation after treatment with glutamate plus BSO was tested using the concentrations of the flavonoids most effective at preventing cell death. A few of the flavonoids, including 3,6-dihydroxyflavone, 3,7-dihydroxyflavone, and hesperetin, did not block ROS accumulation but did prevent the critical influx of calcium (Table 2) , which is the final step in the cell death cascade. 

Cells have a number of different endogenous antioxidant defense mechanisms. Induction of phase 2 detoxification proteins through activation of the ARE can provide long-term protection of cells against oxidative stress. Transcriptional activation of the ARE is dependent on the transcription factor Nrf2, a member of the Cap‘n’Collar family of bZIP proteins. ¹⁹ To determine whether any of the neuroprotective flavonoids can activate the ARE and induce the synthesis of phase 2 detoxification proteins, we first treated the RGC-5 cells for 30 minutes to 4 hours with the optimal effective concentrations of the different protective flavonoids or vitamin E and looked for an increase in the level of the transcription factor Nrf2 in the nuclei of cells. Only a subset of the protective flavonoids increased the levels of Nrf2 (Fig. 6and Table 2 ). Figure 6Ashows the time course of Nrf2 induction by quercetin, a flavonoid that activates the ARE and the lack of induction by luteolin, a flavonoid that does not activate the ARE in the RGC-5 cells. To determine whether the increases in Nrf2 levels translated into an increase in phase 2 detoxification proteins, we treated the RGC-5 cells for 24 hours with the same flavonoids and looked for an increase in HO-1. We chose HO-1 because its synthesis is generally dependent on the ARE ^{20 21} and robust antibodies are available. As shown in Figure 6Band Table 2 , some of the neuroprotective flavonoids are able to significantly increase the expression of HO-1. The increases in HO-1 levels correlated well with the increases in the levels of Nrf2 such that no flavonoids that did not induce Nrf2 expression induced HO-1 expression and vice versa. 

Discussion

The data presented in this report demonstrate that specific flavonoids can protect RGCs from oxidative stress–induced cell death. In these studies, we used three different models of oxidative stress because the precise causes of oxidative stress in glaucoma, diabetic retinopathy, and other ocular diseases involving the loss of retinal ganglion cells are unknown. Therefore, identifying compounds that can protect against multiple inducers of oxidative stress is likely to have the greatest therapeutic potential. We found that some of the flavonoids protect regardless of which model of oxidative stress was tested. Other flavonoids only protect against oxidative stress induced by GSH depletion, which is probably the mildest form of stress because of the slow accumulation of ROS but which may most closely approximate the conditions found in vivo. 

In these studies, we tested a wide range of flavonoids covering the six different structural classes (Fig. 1) . In general, our results are consistent with those of Ishige et al., ⁴ who looked at protection from oxidative stress induced by GSH depletion in the mouse hippocampal cell line HT22. They demonstrated requirements for the specific location of hydroxyl groups, unsaturation of the C ring, and high hydrophobicity for a flavonoid to be effective at preventing oxidative stress–induced cell death. These structure–activity relationships are in complete agreement with our data. Furthermore, consistent with their data, we found little or no protection by catechin, epicatechin, or EGCG, suggesting that though these compounds may work well in in vitro assays of oxidative stress, they are less effective in whole cells. However, we did find some differences in efficacy between the two types of cells. For example, taxifolin has an EC₅₀ of 30 μM for the protection of the RGC-5 cells from oxidative stress induced by GSH depletion but an EC₅₀ > 50 μM for protection of the CNS-derived mouse HT22 cells from a similar insult. ⁴  

We began our studies with the flavonoids by pretreating the cells for 30 minutes before the addition of the stress. However, in all likelihood, in vivo the oxidative stress will be present before treatment. Thus, it is important to know whether specific flavonoids can be added after the induction of stress and still provide neuroprotection. We identified several flavonoids that could be added long after treatment with glutamate plus BSO was begun, and some of these provided significantly greater protection than what was seen if the insult was simply removed from the cells. The ability of some flavonoids to rescue cells exposed to this insult is consistent with their abilities to quench endogenously generated ROS (quercetin, luteolin; Fig. 5 ) or to block the late influx of calcium (3,7-dihydroxyflavone; Table 2 ). These flavonoids were also protective when added after t-BOOH or H₂O₂ had already been added. However, given the more rapid time course of the cell death initiated by both these treatments, the timeframe for postoxidative stress protection by the addition of these flavonoids was more constrained. Furthermore, when compared with the protection afforded by simply changing the medium, the data suggest that in most cases, flavonoids prevent the death of cells that have not yet committed to die after treatment with t-BOOH or H₂O₂ rather than rescuing cells that have already begun the cell death process. Nevertheless, the ability of certain flavonoids to provide posttreatment protection, regardless of the insult, could still be very useful in the clinical setting since it is not usually possible to remove the insult in vivo in the same way it is in vitro. 

Ishige et al. ⁴ found that flavonoids could protect at three steps of the cell death pathway initiated by GSH depletion: maintaining GSH levels, inhibiting the accumulation of ROS, and blocking calcium influx. However, in our study with RGC-5 cells, none of the flavonoids we tested prevented the loss of GSH. GSH loss can occur through several different mechanisms, including direct export, oxidation to GSSG followed by export, and conjugation to a variety of molecules such as lipids and proteins. ^{22 23} Furthermore, in addition to being exported, GSSG itself can bind to proteins. In most cases, the replacement of lost GSH requires new GSH synthesis. The combination of glutamate plus BSO that we used to induce oxidative stress in the RGC-5 cells inhibits GSH biosynthesis both indirectly, through the inhibition of cystine uptake, and directly, through the inhibition of glutamate cysteine ligase (GCL), the rate-limiting enzyme in GSH biosynthesis. In contrast, Ishige et al. ⁴ used glutamate treatment alone to deplete GSH. Glutamate blocks the uptake of cystine, the rate-limiting amino acid in GSH biosynthesis, but does not inhibit GSH synthesis itself. Taken together, these data suggest that some flavonoids increase GSH levels by stimulating GSH biosynthesis through direct effects on GCL. This stimulation cannot occur in the presence of BSO, which inhibits GCL. 

Many of the flavonoids did prevent the accumulation of ROS. Several pieces of evidence suggest that at least some of them do so by acting directly as antioxidants. First, many have high trolox equivalent antioxidant capacity (TEAC) values consistent with those of an antioxidant. ^{4 24} Second, not only do these flavonoids protect against oxidative stress induced by GSH depletion and t-BOOH, they protect against H₂O₂ toxicity. However, there are a few flavonoids, such as kaempferol and taxifolin, that prevent the accumulation of ROS in response to treatment with glutamate plus BSO but are unlikely to act directly as antioxidants because they have low TEAC values and are not protective against H₂O₂. These flavonoids may block ROS production by mitochondria, as was shown for some tyrphostins. ²⁵ Finally, as noted earlier, a small number of flavonoids, such as 3,6- and 3,7-dihydroxyflavone, prevent calcium influx, the last step in the cell death process. These flavonoids do not prevent the accumulation of ROS but instead allow cells to survive in the presence of high levels of ROS. Given that calcium influx is the final, common step in oxidative stress–induced cell death initiated by a variety of different stimuli, ¹⁵ compounds that block this step could be particularly useful for protecting cells after the onset of the initial insult (e.g., Fig. 4 ). Exactly how specific flavonoids prevent calcium influx is unclear. 

It is likely that the most effective neuroprotective agents will not only have multiple activities, they will also induce the synthesis of antioxidant enzymes that can continue to function even when the compound is no longer present in the cell. Compounds that can activate the ARE have the potential to be particularly effective neuroprotective agents because they can induce the synthesis of phase 2 detoxification enzymes that are part of the endogenous antioxidant defense mechanism present in cells. We found that several of the protective flavonoids could induce the up-regulation of the ARE-specific transcription factor Nrf2 and the subsequent stimulation of HO-1 synthesis, which we used as a marker of Nrf2-dependent ARE activation. However, only a small subset of the protective flavonoids was effective in these assays, suggesting that this group might have the greatest potential clinical benefits. These flavonoids include fisetin, quercetin, and galangin. 

Several flavonoids particularly effective at protecting the RGCs from oxidative stress–induced death are found in relatively high quantities in specific fruits and vegetables. ²⁶ For example, quercetin is abundant in yellow onions, kale, apples, and blueberries and is also found in red wine and tea. Studies with human volunteers have shown that a single meal of fried onions can significantly increase the plasma levels of quercetin, which is then slowly eliminated over the next 17 hours. ²⁷ Thus, repeated intake of onions and other foods containing quercetin would lead to a build-up in the plasma concentration. In cattle, quercetin was also found to accumulate in the retina. ²⁸ Taken together, our results suggest that the consumption of a diet rich in these fruits and vegetables could have beneficial effects on the eye in pathologic conditions and in normal aging. 

In summary, working toward the long-term goal of identifying a neuroprotective compound that could be used clinically for the treatment of multiple retinal disorders, we have found that specific flavonoids can protect RGCs from oxidative stress–induced death with high levels of potency and low toxicity. These results have significant clinical potential because oxidative stress has been implicated in many types of ocular diseases, including glaucoma, diabetic retinopathy, and macular degeneration. Furthermore, compared with traditional antioxidants, flavonoids have unique advantages that make them especially attractive for clinical use. In particular, as we show, some flavonoids induce the expression of antioxidant proteins that can help protect cells from oxidative stress, thus potentially providing long-term protection of the eye. 

 Supported by National Institutes of Health Grant NS28121 (PM) and by funding from the Mericos–TSRI Neurobiology and Vision Science Program.

 Submitted for publication March 29, 2005; revised May 31, July 11, and July 28, 2005; accepted September 26, 2005.

 Disclosure: P. Maher, None; A. Hanneken, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Chemical structures of the different classes of flavonoids used in this study. The positions of the hydroxyl groups and common names are listed in Table 1 .

Figure 5.

 

Flavonoids prevent the accumulation of ROS. RGC-5 cells were untreated or treated with 10 mM glutamate plus 500 μM BSO in the absence or presence of the flavonoids 3,7-dihydroxyflavone (3,7-DHF; 15 μM), hesperetin (50 μM), apigenin (10 μM), baicalein (25 μM), isorhamnetin (15 μM), kaempferol (15 μM), luteolin (15 μM), quercetin (25 μM), taxifolin (100 μM), or the antioxidant propyl gallate (PG; 10 μM). After 18 hours, the cells were loaded with H2DCF-dA and processed for fluorescence-activated cell sorter analysis. Cellular levels of ROS were measured using DCF fluorescence, and the fold increase in the median DCF fluorescence intensity with respect to that of the control was plotted. Results are the means of duplicate determinations ± SD. Similar results were obtained in two to three independent experiments. *Significantly different from glutamate plus BSO alone.

Figure 2.

 

Flavonoids protect RGC-5 cells from oxidative stress induced by treatment with glutamate plus BSO, t-BOOH, and H₂O₂. RGC-5 cells were treated with glutamate plus BSO (10 mM + 500 μM), t-BOOH (0.05 mM), or H₂O₂ (650 μM) in the absence or presence of the flavonoids quercetin (25 μM), kaempferol (15 μM), or taxifolin (100 μM). Twenty-four hours later the cells were photographed. Similar results were obtained in two to three independent experiments.

Figure 3.

 

Dose–response curves for protection of RGC-5 cells from oxidative stress. RGC-5 cells were treated with glutamate plus BSO (10 mM + 500 μM), t-BOOH (0.05 mM), or H₂O₂ (650 μM) in the absence or presence of the indicated doses of the flavonoids (A) luteolin, (B) 3,6-dihydroxyflavone, or (C) quercetin. Twenty-four hours later cell survival was measured by the MTT assay. Results presented are the average ± SD of three to five independent experiments.

Table 1.

 

Protection of RGC-5 Cells by Flavonoids

Flavonoid	Free Hydroxyl Positions	Common Name	Glutamate Plus BSO EC50 (μM)*	t-BOOH EC50 (μM)	H2O2 EC50 (μM)	LD50 (μM), †
Flavone	—	—	No, ‡	—, §	—, §	—, §
	3	Flavonol	6	No	No, ‡	>50
	3-methoxy	—	No, ‡	—, §	—, §	—, §
	6	—	No, ‡	—, §	—, §	—, §
	6-methoxy	—	No, ‡	—, §	—, §	—, §
	7	—	No, ‡	—, §	—, §	—, §
	3,6	—	2.5	8	13	>25
	3,7	—	7	10	10	>25
	5,7	Chrysin	No, ‡	—, §	—, §	—, §
	3,5,7	Galangin	10	50	100	80
	5,6,7	Baicalein	3	10	25	>25
	4′,5,7	Apigenin	15	No	No, ‡	25
	3,4′,5,7	Kaempferol	1	No	No, ‡	110
	3′,4′,5,7	Luteolin	2	7	8	>25
	3,3′,4′,7	Fisetin	15	10	15	>25
	3,3′,4′,5,7	Quercetin	17	18	20	>25
	3,3′,4′,5,7	Rutin	No, ‡	—, §	—, §	—, §
	3,3′methoxy,4′,5,7	Isorhamnetin	5	15	No, ‡	25
	2′,3,4′,5,7	Morin	50	No	No, ‡	>100
	3,3′,4′,5,5′,7	Myricetin	No, ‡	—, §	—, §	—, §
Isoflavone	4′,5,7	Genistein	100	No	No, ‡	100
Flavanone	4′,5,7	Naringenin	No, ‡	—, §	—, §	—, §
	3′,4′,5,7	Eriodictyol	5	25	25	>25
	3′,5,7,4′-methoxy	Hesperetin	50	No	No, ‡	>100
	3,3′,4′,5,7	Taxifolin	30	60	No, ‡	>100
Flavanol	3,3′,4′,5,7	Catechin	No, ‡	—, §	—, §	—, §
	3,3′,4′,5,7	Epicatechin	50	No	No, ‡	>100
		EGCG	No, ‡	—, §	—, §	—, §
Anthocyanidin	3,3′,4,4′,5,7	Cyanidin	No, ‡	—, §	—, §	—, §

*  Half-maximal effective concentration (EC₅₀) for protection was determined by exposing RGC-5 cells to different doses of each flavonoid in the presence of 10 mM glutamate plus 500 μM BSO, 0.05 mM t-BOOH, or 650 μM H₂O₂. Cell viability was determined after 24 hours by the MTT assay.

†  Half-maximal lethal doses were determined by exposing RGC-5 cells to increasing concentrations of each flavonoid for 48 hours. Cell viability was determined after 48 hours by the MTT assay.

‡  These compounds did not protect significantly from glutamate plus BSO, t-BOOH, or H₂O₂.

§  Not tested.

Figure 4.

 

Flavonoids protect even when added after the initial oxidative insult. RGC-5 cells were treated with (A) glutamate plus BSO (10 mM + 500 μM), (B) t-BOOH (0.05 mM), or (C) H₂O₂ (650 μM) alone or in the presence of the flavonoids fisetin (25 μM), 3,7-dihydroxyflavone (15 μM; 3,7-DHF), baicalein (25 μM), luteolin (15 μM), or quercetin (25 μM). Flavonoids were added either at the same time as glutamate plus BSO, t-BOOH, or H₂O₂ or at different times thereafter. In some experiments, rather than adding flavonoids, the medium was simply changed at the same time points (medium change). The dotted line at the bottom of each graph indicates the level of survival if the oxidant is left in the medium throughout the course of the experiment (i.e., no media change) without the addition of any flavonoids. Twenty-four hours later cell survival was measured by the MTT assay. Results presented are the average ± SD of three to five independent experiments. (A) *Significantly different from medium change (P = 0.02). **Significantly different from medium change (P < 0.005). (C) **Significantly different from medium change (P < 0.005).

Table 2.

 

Protective Mechanisms for Different Flavonoids

	GSH Metabolism	ROS Scavenger	Anti-Ca+2 Influx	ARE Activation
3,6 Dihydroxyflavone	−	−	+	−
3,7 Dihydroxyflavone	−	−	+	−
Galangin	−	+	−	+
Baicalein	−	+	−	−
Luteolin	−	+	−	−
Fisetin	−	+	−	+
Quercetin	−	+	−	+
Isorhamnetin	−	+	−	−
Eriodictyol	−	+	−	−
Taxifolin	−	+	−	−

 Sites of action against oxidative stress. Shown are data for flavonoids found to protect from multiple forms of oxidative stress (glutathione depletion, t-BOOH toxicity, and H₂O₂ toxicity). Total GSH was measured chemically, using glutathione depletion as the oxidative insult. Similar results were seen in two to three independent experiments (data not shown). ROS and intracellular calcium levels were measured using FACS analysis as shown in Figure 5for ROS using glutathione depletion as the oxidative insult. Similar results were seen in two to three independent experiments. ARE activation was determined by examining both the induction of Nrf2 and the expression of HO-1, as shown in Figure 6 . Similar results were seen in two to three independent experiments.

Figure 6.

 

Flavonoids induce the expression of Nrf2, the ARE-specific transcription factor, and HO-1, a phase 2 detoxification protein. (A) RGC-5 cells were untreated (ct) or treated with 25 μM quercetin or 15 μM luteolin for 1, 2, or 4 hours. Nuclei were prepared, and equal amounts of protein were analyzed by SDS-PAGE and immunoblotting with anti-Nrf-2 antibodies. (B) RGC-5 cells were untreated (ct) or treated for 24 hours with 25 μM quercetin (Q), 15 μM luteolin (lut), 15 μM kaempferol (K), 100 μM taxifolin (tax), or 10 μM vitamin E. Cell lysates were prepared, and equal amounts of cellular protein were analyzed by SDS-PAGE and immunoblotting with anti–HO-1 antibodies. Immunoblotting with anti-actin is shown as a loading control. Similar results were obtained in two to three independent experiments.