Hepatocyte Growth Factor Protects RPE Cells from Apoptosis Induced by Glutathione Depletion

Abstract

purpose. To study the mechanism of the protective effect of hepatocyte growth factor (HGF) in oxidative injury to RPE cells induced by glutathione (GSH) depletion.

methods. RPE cells were treated with HGF for 24 hours (20 ng/mL) and then were treated with DL-buthionine-(S,R)-sulfoximine (BSO) for an additional 24 hours. Cell death, apoptosis, and GSH levels were measured. Levels of intracellular reactive oxygen species (ROS) and their cellular localization were assessed by confocal microscopy. Expression of Bcl-2 and release of cytochrome c from mitochondria were quantified. The effect of BSO on caspase-3 activation and expression was determined. Gene expression of key enzymes of GSH metabolism by real-time PCR and regulation and translocation of the transcription factor NF-E2–related factor (Nrf2) by BSO were examined.

results. Treatment with BSO-induced apoptosis in RPE caused a significant decrease in intracellular GSH and in GSH/GSSG ratios. Marked increases in lipid peroxidase (LPO), ROS, and mitochondrial cytochrome c release and a decrease in Bcl-2 expression were observed. Elevated GSH/GSSG ratio (especially in mitochondria), decreased LPO and ROS, attenuation of apoptosis, and partial restoration of Bcl-2 expression were found in the HGF-pretreated cells. BSO activated caspase-3, and this effect was significantly blocked by HGF. Both HGF and BSO induced anti-oxidant gene expression. Nrf2 translocated to the nuclear region after treatment with BSO, whereas HGF did not induce such translocation.

conclusions. The protective effect of HGF may be attributed in part to the elevation of mitochondrial GSH. BSO and HGF act in concert to enhance GSH-related gene expression in stressed RPE cells.

Retinal pigment epithelium (RPE) cell dysfunction and death play a vital role in the pathogenesis of age-related macular degeneration (AMD). ¹ Ample evidence in the literature suggests that oxidative stress may be a contributing factor to RPE dysfunction in AMD. ² In vitro studies show that oxidant-treated RPE cells undergo apoptosis, a possible mechanism by which RPE cells are lost in AMD. ²  

The generation of lipid peroxidation products and reactive oxygen species (ROS) has been documented in AMD and other retinal diseases. Significantly higher concentrations of lipid peroxidation products in patients with AMD have recently been reported as compared to controls. ³ Everiklioglu et al. ⁴ showed that nitric oxide and lipid peroxidation products increased in AMD and that this increase has been associated with decreased enzyme activity of the antioxidative enzymes superoxide dismutase, glutathione peroxidase (GPX), and catalase. Human RPE lipofuscin, the accumulation of which can lead to retinal disease, contains proteins that are modified by lipid peroxidation products and advanced glycation end products. ⁵ The inhibition of lysosomal and antioxidant activity in RPE by lipofuscin has also been shown. ⁶ Recently, the highly reactive lipid peroxidation product 4-HNE was found to significantly induce the angiogenic factor vascular endothelial growth factor in RPE. ⁷ Thus, preventing the formation of lipid peroxidation products or the scavenging of reactive oxidative metabolites in RPE is of considerable therapeutic importance. 

In this context, it is becoming increasingly evident that the redox status of RPE cells plays a critical role in combating oxidant stress. ² Cellular redox potential is largely determined by glutathione (GSH), which accounts for >90% of cellular nonprotein thiols. ⁸ Most GSH is found in the cytosol; however, GSH in mitochondria, which comprises 10% to 20% of the cellular pool, is considered essential for cellular integrity and function. ⁸ The protective role of GSH and its amino acid precursors from oxidant-induced apoptosis in vitro has been reported. ^{9 10 11} Retinal redox status was significantly altered by intense light and was partially normalized by exogenous thiol replenishment via N-acetylcysteine. ¹² Furthermore, GSH and several enzymes of GSH metabolism (e.g., GSH peroxidase, GSH-S-transferase) have been reported to offer protection to the retina. ^{2 13} This includes the counteraction of photic injury, ¹⁴ modulation of chloride conductance and volume regulation, ¹⁵ and partial reattachment of experimentally detached retinas. ¹⁶  

The defenses ¹⁷ to minimize ROS levels include enzymes that are involved in ROS metabolism and biotransformation of xenobiotics. Examples of these enzymes include superoxide dismutases, glutathione peroxidases, gamma glutamylcysteine synthetase (γ-GCS; also called glutamine cysteine ligase catalytic), thioredoxins, and heme-oxygenases, as well as phase 2 enzymes, such as glutathione S-transferases. Basal and inducible expression of a number of these antioxidant genes are mediated in part by a cis-acting DNA element known as the antioxidant response element (ARE). ^{18 19} The nuclear transcription factors NF-E2-related factor (Nrf2) and c-Jun bind to the ARE and activate the gene expression of antioxidant genes. Nrf2 is retained in the cytoplasm by the repressor protein Keap1 in normal cells; treatment with xenobiotics and antioxidants causes dissociation of Nrf2 from Keap1. ²⁰ The free Nrf2 translocates into the nucleus, heterodimerizes with c-Jun, and binds to the ARE, resulting in the induction of antioxidant genes. ²⁰ The physiological requirements of Nrf2 in RPE, the nature and significance of compartmentalization in defense against oxidative stress, and mediation by growth factors are largely unknown. 

Hepatocyte growth factor/scatter factor (HGF/SF) has a variety of cellular functions, ²¹ and its biologic actions are mediated by the HGF/SF receptor c-met, a transmembrane tyrosine kinase. ²¹ Previously, we showed that cultured RPE cells express HGF receptor c-met and that they respond to HGF by proliferation and chemotactic migration. ²² HGF was also found to protect various cell types against apoptosis induced by a variety of stimuli ^{23 24 25 26} . Our recent in vivo rabbit study revealed that adenovirus-mediated overexpression of HGF in RPE has a strong neuroprotective effect on detached photoreceptor cells. ²⁷ Tsuboi ²⁸ found that HGF increased intracellular GSH by transcriptional activation of rate limiting γ-GCS in rat hepatocytes. 

In this study, we imposed oxidative stress on RPE cells by treatment with BSO, a GSH depletor. The induction of RPE apoptosis through oxy-radical production, cytochrome c release, Bcl-2 downregulation, and caspase-3 activation was demonstrated. Pretreatment of RPE cells with HGF markedly elevated the mitochondrial GSH and protected the RPE from BSO-induced apoptosis and caspase-3 activation. BSO induced Nrf2 nuclear translocation with increased antioxidant gene expression. HGF did not induce Nrf2 translocation; however, it did induce the expression of a complementary set of GSH-related antioxidant genes. 

Materials and Methods

Cell Culture and Treatment

Studies using cultured human RPE were approved by the Institutional Review Board of the University of Southern California. RPE cells were isolated from human donor eyes obtained from Advanced Bioscience Resources, Inc. (Alameda, CA) and were cultured in Dulbecco’s minimal Eagle’s medium (DMEM; Fisher Scientific, Pittsburgh, PA) with 2 mM l-glutamine, 100 U/mL penicillin, 100 μg/mL streptomycin (Sigma, St. Louis, MO), and 10% heat-inactivated fetal bovine serum (FBS; Irvine Scientific, Santa Ana, CA), as previously described. ²⁹ Purity of cultures was established by immunohistochemical staining of cytokeratin, a marker for RPE cells. Greater than 95% of cells were cytokeratin positive, indicating epithelial origin, whereas no cells were found positive for the macrophage marker CD11 or for the endothelial cell marker von Willebrand factor. ³⁰ Third- to fourth-passage cells at >90% confluence were used in all experiments. Cells were changed to 1% FBS DMEM for 24 hours. We used cells maintained in 1% FBS in all our experiments to restrict the contribution from serum antioxidants and growth factors. Our previous studies have found low serum conditions optimal for these studies; however, serum-free conditions cannot be used because they induce prominent cellular stress. ³⁰ The cells were pretreated for 24 hours with HGF (10–50 ng/mL; R&D Systems, Minneapolis, MN) in 1% FBS and then were treated continuously with BSO for 0, 6, 12, 24, and 48 hours. Shorter incubation periods for BSO (30 minutes) and HGF (6 hours) were used in confocal microscopic analyses. 

Apoptosis Assays

The cleavage of DNA, which characteristically occurs in apoptosis, was measured with a TUNEL kit (In-Situ Cell Death Detection Kit, Fluorescein; Roche Molecular Biochemicals, Indianapolis, IN). Briefly, after treatment, floating and adherent cells (released by trypsin) were collected and fixed with freshly prepared 4% paraformaldehyde in phosphate-buffered saline (pH 7.4) for 30 minutes at room temperature, then permeabilized with 0.1% Triton X-100 in 0.1% sodium citrate on ice for 2 minutes. To label DNA strand breaks, cells were incubated with a 50-μL TUNEL reaction mixture containing TdT and fluorescein-dUTP in the binding buffer for 1 hour at 37°C in a humidified atmosphere. Cells were then washed and analyzed by flow cytometry. 

Cell viability was determined with a commercial kit that uses a two-color fluorescence cell viability assay (LIVE/DEAD viability/toxicity kit; Molecular Probes, Eugene, OR) based on the simultaneous determination of live and dead cells with two probes that measure intracellular esterase activity and plasma membrane integrity. ³¹  

Measurement of Cellular and Mitochondrial GSH

Cellular GSH was measured using a colorimetric GSH assay kit (Oxford Biomedical Research, Oxford, MI). This kit makes use of a kinetic enzymatic recycling assay based on the oxidation of GSH by 5,5′-dithiobis-2-nitrobenzoic acid (DTNB) and glutathione reductase to measure the total GSH content in cells. Addition of NADPH2 initiates the progressive reduction of DTNB by GSH, causing a color increase that is monitored at 405 nm. The rate of color change, typically monitored over a 4-minute period, is proportional to the total GSH concentration. This assay is specific for GSH; other thiols do not cause interference in the assay. Mitochondrial-enriched fractions were isolated using a Mitochondria/Cytosol Fractionation Kit (BioVision Inc., Mountain View, CA) ³² ; enriched fractions were then used to measure the mitochondrial GSH, using the procedure described for cellular GSH, and cytochrome c released to the cytosol. 

ROS Determination

ROS accumulation in the RPE cells was detected with a carboxy-H₂-DCFDA (Molecular Probes) staining method. This assay is based on the principle that the nonpolar, nonionic H₂-DCFDA crosses cell membranes and is enzymatically hydrolyzed into nonfluorescent H₂-DCF by intracellular esterases. In the presence of ROS, H₂-DCF is rapidly oxidized to become highly fluorescent DCF. ³³ At the end of the treatments (which were performed simultaneously), cells were incubated at 37°C for 1 hour with 5 μM carboxy-H₂-DCFDA dissolved in the culture medium. To determine the compartmentalized accumulation of ROS, the mitochondria were visualized by a cell-permeable, mitochondria-specific fluorescent dye, CMXRos (Molecular Probes); this provides dual staining if ROS accumulate in the mitochondria. The cells were then incubated with CMXRos at a final concentration of 500 nM for 30 minutes after incubation with the carboxy-H₂-DCFDA for 30 minutes. 

Determination of Lipid Peroxides

Lipid peroxide levels in RPE cells were determined with a lipid peroxidase kit (Biotech LPO-586; OxisResearch, Portland, OR) according to the manufacturer’s protocol. This assay is based on the reaction of a chromogenic reagent, N-methyl-2-phenylindole (R1), with MDA and 4-hydroxyalkenals. One molecule of either MDA or 4-hydroxyalkenal reacts with two molecules of reagent R1 to yield a stable chromophore with maximal absorbance at 586 nm. 

Western Blot Analyses

Cells were lysed, supernatants were collected, and proteins were resolved on Tris-HCl polyacrylamide gels (Ready Gel; Bio-Rad, Hercules, CA) at 120 V. The proteins were transferred to a polyvinylidene difluoride (PVDF) blotting membrane (Millipore, Bedford, MA), and the membranes were probed with mouse anti–cytochrome c monoclonal antibody, rabbit polyclonal anti-caspase-3 antibody, and rabbit polyclonal anti–Bcl-2 antibody (all purchased from PharMingen, San Diego, CA). Details of the secondary antibodies used and their detection by chemiluminescence (Amersham Pharmacia Biotech, Cleveland, OH) have been previously described. ³⁴ Equal loading was verified by determining protein concentration by Bio-Rad protein assay (Bio-Rad) and using the same amount of protein from protein lysates for electrophoretic analysis. GAPDH was used as standard, and the ratio between the analyzed protein and GAPDH from quantitative densitometric analysis was used for comparison of control and experimental samples. 

Immunofluorescence Staining of Nrf2

RPE cells were cultured in multichamber slides and treated for 6 hours with 20 ng/mL HGF or 1 hour with 500 μM BSO, or they were pretreated with HGF for 6 hours followed by 1-hour BSO treatment. Cells were then fixed with methanol and probed with 1:200 dilution of polyclonal Nrf2 antibody (Vector Laboratories, Burlingame, CA) at 4°C overnight, followed by 1:200 dilution of FITC-conjugated anti-rabbit IgG for 30 minutes. DAPI staining was used to visualize the nucleus. 

Cytochrome c Release

Cytochrome c release from the mitochondria into the cytosol was measured by Western blot. ³⁵ Proteins that were extracted from the cytosol and mitochondria were isolated using a mitochondria/cytosol fractionation kit, separated by 15% SDS-PAGE, and transferred onto nitrocellulose membranes. 

Caspase-3 Activation

Caspase-3 activation was determined using an FITC-conjugated cell-permeable caspase inhibitor VAD-FMK (CaspACE FITC-VAD-FMK In Situ Marker; Promega, Madison, WI). Cells were collected and incubated with FITC-VAD-FMK for 60 minutes and measured by flow cytometry using the FL-1 setting. Ten thousand events were recorded in each analysis. Western blot analysis of active caspase-3 was performed using a specific polyclonal antibody, as described. 

Expression of GSH-Related Enzymes

Relative quantitative expression of GSH-related enzymes was evaluated using real-time PCR (LightCycler; Roche Molecular Biochemicals). An RNA extraction reagent (TRIzol; Invitrogen Life Technologies, Carlsbad, CA) was used to isolate RNA, and contaminating genomic DNA was removed with a kit (DNA-free; Ambion, Austin, TX). Reverse transcription was performed using 1 μg total RNA, oligo(dT)₁₅ primer (Promega), and AMV reverse transcriptase (Promega). Polymerase chain reactions were performed in a 20-μL mixture containing 2 μL green fluorescent dye with Taq DNA polymerase, and reaction buffer (LightCycler FastStart DNA Master SYBR Green I; Roche Molecular Biochemicals). β-actin, the housekeeping gene, served as an internal control; GPX1, GST 5.8, and γGCS were the GSH-related enzymes. The following primer sets were designed (Primer Express software; Applied Biosystems, Foster City, CA): β-actin 5′, GPX 5′-GACCGACCC CAAGCTCATC-3′ and 5′-TTCTCAAAGTTCCAGGCAACATC-3′; GST 5.8, 5′-CTTC ATTGTGGGAGACCAGATCT-3′ and 5′-GGGCTAGGACCTCA TGGATCA-3′; γ-GCS, 5′-CAGTTGGCTACTATCTGTC-3′ and 5′-GTCTATTGAGT CATATCGGG. The detection of product formation was set in the center of the linear portion of PCR amplification, the cycle at which each reaction reached the set threshold (C _T) was determined, and relative multiples of change in mRNA expression were determined by calculation of ΔΔ C _T. ^{36 37} Results are reported as the mean difference in relative multiples of change in mRNA expression ± SEM. 

Statistical Analysis

Results are expressed as mean ± SEM. An unpaired, two-tailed Student’s t-test was used to determine the statistical difference between two group means. Differences were considered statistically significant at P < 0.05. 

Results

BSO-Induced RPE Cell Apoptosis

BSO (500 μM) treatment decreased the viability of RPE to 74.3% at 24 hours (P < 0.05) and to 53.2% at 48 hours (P < 0.01) when compared with the controls (Fig. 1A) . Viability was significantly higher in the HGF-pretreated cells than in the BSO-treated cells at 24 hours and 48 hours (P < 0.05). The mechanism of cell death in BSO-treated RPE was assessed with the TdT-mediated dUTP nick-end labeling (TUNEL) assay. BSO (500 μM) caused 25.1% (P < 0.05) and 43.6% (P < 0.05) of cells to undergo apoptosis at 24 and 48 hours of treatment, respectively. Pretreatment with HGF significantly decreased the apoptosis caused by BSO at 24 and 48 hours (P < 0.01). 

Alterations of RPE Redox Status by BSO

Figure 2shows the total intracellular GSH and mitochondrial GSH in BSO-treated RPE and the effect of HGF pretreatment on these levels. Intracellular GSH levels decreased significantly (>90%; P < 0.01 versus controls) with 500 μM BSO exposure for 24 hours, and pretreatment with HGF increased cellular GSH levels twofold from the depleted state (Fig. 2A) . Mitochondrial GSH levels, which constitute a small portion of total cellular GSH in RPE cells, ¹⁰ increased significantly (P < 0.01) with HGF pretreatment, reaching levels similar to those of the mitochondria of untreated cells (Fig. 2B) . Preincubation of untreated RPE cells with HGF alone did not significantly affect either cellular GSH or mitochondrial GSH levels. 

ROS Production by BSO Treatment

The compartmentalized accumulation of ROS in RPE cells was verified by a dual staining method using an ROS-specific dye, carboxy-H₂-DCFDA, and a mitochondria-specific fluorescent dye, CMXRos. A remarkable increase in the levels of ROS could be seen with BSO treatment when compared with the control, and the ROS accumulation was primarily found in the mitochondria (Fig. 3) . Pretreatment with HGF significantly decreased the generation of ROS in the mitochondria. In separate studies, the production of ROS was measured by quantifying lipid peroxide (LPO) and MDA either individually or in combination. BSO treatment induced a twofold increase in LPO level compared with control levels (Fig. 4) . When RPE cells were preincubated with HGF (20 ng/mL), LPO generation was significantly reduced to levels almost the same as those of the control RPE (Fig. 4) . 

Release of Cytochrome c and Activation of Caspase-3 by BSO Treatment

To determine whether exposure to BSO results in the release of cytochrome c from mitochondria, we measured cytochrome c levels by Western blot in mitochondria and cytosolic fractions; this revealed an increased release of cytochrome c to the cytoplasm with 24 hours of BSO treatment. Pretreatment with HGF resulted in a marked reduction of cytochrome c released to the cytosol (Fig. 5) . No significant release of cytochrome c was detected in the untreated control cells. Evidence for the activation of caspase-3 with BSO by flow cytometry and Western blot analysis is shown in Figure 6 . In flow cytometric analysis, BSO treatment increased the number of active caspase-3–positive cells from 7.8% to 42.5% (P < 0.01), whereas a HGF-pretreatment of the BSO-treated cells decreased the number of active caspase-3–positive cells to 20.4% (P < 0.01 versus BSO-only group; Fig. 6 , top panel). Western blot analysis confirmed a partial inhibition of caspase-3 activation by BSO with HGF pretreatment (Fig. 6 , bottom panel). 

Bcl-2 Regulation with BSO and HGF

Figure 7shows that Bcl-2 expression was significantly reduced (P < 0.01 versus control) with 24 hours of BSO treatment. In HGF (20 ng/mL)–pretreated cells, a partial restoration of Bcl-2 expression was observed (Fig. 7) . 

Gene Regulation of Key GSH-Related Enzymes with BSO and HGF Treatments

The effect of BSO treatment on the regulation of gene expression of glutathione-S-transferase (hGST5.8), heavy subunit of γ-GCS, and glutathione peroxidase-1 (GPX1) and changes in expression with HGF pretreatment were studied by quantitative real-time PCR. Figure 8shows that BSO treatment increased hGST 5.8 and γ-GCS gene expression approximately 2.5-fold above control expression, whereas GPX1 expression remained unchanged. Preincubation with HGF caused a further significant increase in hGST 5.8 over BSO treatment and upregulated GPX1; γ-GCS was not affected. When RPE cells were incubated with HGF alone, no significant change in the expression of GST 5.8 and γ-GCS was observed when compared with untreated control expression, whereas GPX1 showed a significant (P < 0.05) increase (Fig. 8) . 

Localization of Nrf2 and Effects of BSO and HGF Treatments

Figure 9shows confocal images of the localization of Nrf2 in RPE and the effects of BSO and HGF pretreatment on this localization. Nrf2 was expressed diffusely in control RPE, whereas 500 μM BSO treatment for 1 hour caused increased expression and translocation of Nrf2 to the nuclear and perinuclear regions. Conversely, HGF treatment alone did not significantly alter the intensity or staining pattern of RPE when compared with untreated controls. HGF pretreatment followed by BSO showed moderate Nrf2 staining, mostly in the nuclei. 

Discussion

The present study has shown that oxidative stress induced by GSH deficiency leads to apoptosis in RPE and that HGF partially blocks the induction of apoptosis by upregulating cellular redox status and by inhibiting caspase-3–dependent cell death. Nuclear translocation of the transcriptional regulator Nrf2 by oxidative stress and the complementary induction of several antioxidant genes by BSO and by HGF were also found. 

The elevation of cellular GSH and GSH-related enzymes in ocular cells has been found to be protective against photic injury, volume regulation, and retinal detachment. ^{13 14 15 16} Furthermore, many chemotherapeutic agents have profound effects on the cellular redox status, which can play an important role in the induction of apoptosis. Apoptosis induced by oxidative stress has been implicated in retinal diseases, including AMD. ^{1 2} The induction of apoptosis by tert- butyl hydroperoxide and its suppression by GSH in vitro was reported recently. ³⁸ Our present study not only characterizes apoptosis, it provides insight into the sequence of biochemical signaling events that occur; it has also focused on the important role played by mitochondria. Furthermore, our results show that HGF pretreatment partially attenuates the development of apoptosis by inhibiting mitochondrial biochemical events that lead to cell death. 

Consistent with the postulate that mitochondrial ROS are important effector molecules in the induction of apoptosis, ³⁹ we observed a significant increase in ROS generation in RPE with BSO. Initiation of the death receptor pathway with an activation of caspase-3 and the release of cytochrome c into the cytosolic compartment was found. ⁴⁰ Our finding of increased cytochrome c release with GSH depletion in RPE is in agreement with a recent study of a human B-lymphoma cell line in which BSO caused a time-dependent increase in cytochrome c release. ⁴¹ However, the relationship among GSH depletion, cytochrome c release, and apoptosis is a matter of debate. It was reported that GSH extrusion is the first event to occur before cytochrome c translocation and apoptosis. ⁴² However, additional work has shown that cytochrome c release is not a terminal event leading to apoptosis but is a consequence of redox disequilibrium, which may not be associated with apoptosis under some conditions. ^{42 43} Thus, the regulatory role of the cellular redox state during apoptosis is still controversial. In recent studies, Armstrong and Jones ⁴⁴ found that ANT (one of the components of the poly-protein complex MPT) is a key player in cell death induced by mitochondrial ROS in HL60 cells exposed to BSO. These authors also found that endogenous ROS generated at the respiratory chain complex III induced MPT under GSH-depleted conditions in HL60 cells. ⁴⁵  

Bcl-2 is an antiapoptotic protein localized to membranes of the nucleus, endoplasmic reticulum, and mitochondria. ^{45 46} Its localization to mitochondria has been associated with the inhibition of the MPT and the release of mitochondrial cytochrome c, both of which are central to apoptosis. ^{47 48} In this study, we found that Bcl-2 is downregulated with GSH depletion and that HGF restored Bcl-2 expression to control levels. It was suggested that the suppression of Bcl-2 expression may increase the overall Bim/Bcl-2 ratio for cell survival under conditions of ROS downregulation of Bcl-2 in T cells. ⁴⁹ It is probable that HGF, by detoxifying the ROS in RPE, may reverse the ROS-induced decline in Bcl-2, thereby preventing apoptosis. 

Mitochondrial apoptotic signaling could have been caused by two different pathways, the damage-induced and the physiological pathways, both converging at the activation of caspase-3. ⁵⁰ The relative contribution of the two pathways in BSO-induced apoptosis in RPE and the effect of HGF on the two pathways are not known at present. It is of interest that apoptosis induced by tert-butyl hydroperoxide in RPE was preceded by the upregulation of Fas-ligand and that antioxidants inhibited Fas-ligand and apoptosis. ⁵¹ Our finding in this study that HGF causes the attenuation of ROS levels and the downregulation of caspase-3 with a partial elevation of cell GSH suggests that its action, at least in part, may be mediated by improvement in the redox status of RPE. Caspase-3–independent mechanisms of cell death in chemically induced oxidative injury have also been reported. Zhang et al. ⁵² showed that menadione-induced cell death differed between U937 cells and ARPE-19 cells in that the former exhibited caspase-3 activation, whereas in the latter, it was the apoptosis-inducing factor, not caspase-3, that was involved in cell death. In the present study, we primarily focused on the mitochondrial pathways of cell death, though evaluating other pathways would also prove to be of interest. 

Although our studies show that HGF attenuates LPO production, ROS accumulation, and cytochrome c release in RPE cells with BSO treatment, the magnitude of reversal with HGF varied among these indices of oxidant injury. Restoration of LPO and ROS with HGF in BSO-treated cells was higher than the recovery seen for caspase 3 activation, Bcl-2, and cytochrome c. This finding may be explained by the analytical variability and sensitivity of different tests and the possibility of additional HGF-mediated antioxidant effects. ³⁴  

The delineation of the exact mechanism(s) of the protective action of HGF in oxidant stress in RPE remains unexplored. Data from our previous work and other studies have shown that HGF itself triggers antioxidant genes, such as catalase. ³⁴ In this study, the expression of GPX-1 increased with HGF treatment alone, whrereas γGCS and GST expression were not significantly altered. In cells with oxidative injury induced by BSO, we found that antioxidant genes GST-5.8 and γGCS were upregulated, whereas GPX-1 expression compared with that of controls did not change. When GSH-depleted cells were preconditioned with HGF, GPX-1 and GST further increased in their expression than they did with HGF treatment alone. This may suggest that additional non–ARE-related mechanisms may play a role. It is conceivable that the complementary effects of BSO and HGF on GSH-related antioxidant gene expression are key to the improved thiol status found in these cells. HGF action may also be mediated by other redox-related proteins, such as thioredoxin. ⁵³ Our results showed that oxidative stress stimulated by GSH depletion is consistent with the stress-related upregulation reported recently by others. ¹³ Frank et al., ⁵⁴ however, did not find glutathione peroxidase immunoreactivity in the RPE of healthy controls or of patients with AMD. 

Alterations in other GPX family members (GPX2, GPX3, and GPX4) in response to HGF or BSO and the differential response to transcription factors such as NF-κB and AP-1 cannot be excluded at this time. ^{55 56 57} In this context, the expression of the novel transcription factor Nrf2, an important factor in the expression and coordinated induction of detoxifying enzyme genes in response to antioxidants and xenobiotics, is of considerable significance. ⁵⁸ Oxidative stress has been shown to cause a significant translocation of Nrf2 to the nucleus and to trigger increased expression of antioxidant genes. ¹⁸ Consistent with this, we found nuclear translocation of Nrf2 in RPE after treatment with BSO. HGF pretreatment on BSO-induced Nrf2 activity suggested the suppression of Nrf2 by HGF, which may involve modulation of the cooperative assembly of the activated transcription complex at the promoter of target genes. ^{58 59} Studies with Nrf2 reporter vectors or Nrf2 (−/−) mice ⁵⁹ on redox signaling are likely to shed more light on the mechanism. It may also be postulated that Nrf2 accumulation and subsequent activation is caused by decreased proteasomal degradation. ⁶⁰  

³  Contributed equally to the work and therefore should be considered equivalent authors.

 Supported in part by Grants EY02061 and EY03040 from the National Institutes of Health, a grant from the Arnold and Mabel Beckman Foundation, and a grant to the Department of Ophthalmology from Research to Prevent Blindness, Inc. (New York, NY).

 Submitted for publication March 18, 2005; revised May 26, 2005; accepted September 8, 2005.

 Disclosure: M.L. Jin, None; J. Yaung, None; R. Kannan, None; S. He, None; S.J. Ryan, None; D.R. Hinton, 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.

 

Inhibition of buthionine-(S,R)-sulfoximine (BSO)-induced cell death by HGF in RPE. (A) Cell viability was decreased by BSO treatment (500 μM) for 24 h and 48 h (P < 0.05). (B) Apoptosis was prominently increased after BSO treatment (500 μM) for 24 h and 48 h (P < 0.05). Pretreatment with hepatocyte growth factor (HGF) blocked this BSO effect. *P < 0.05 compared with BSO-treated group. Data are mean ± SEM (n = 4).

Figure 2.

 

Levels of cellular glutathione (GSH) and mitochondrial GSH with buthionine-(S,R)-sulfoximine (BSO) and hepatocyte growth factor (HGF) treatments. Intracellular GSH was measured by a colorimetric assay. Significant decreases were observed in GSH with BSO exposure compared with untreated controls. Pretreatment with 20 ng/mL HGF partially blocked this effect (A). Mitochondria GSH levels decreased significantly with 24 h BSO treatment, and HGF pretreatment restored mitochondria GSH to that of untreated controls (B).

Figure 3.

 

Evidence by confocal microscopy for reactive oxygen species (ROS) accumulation in retinal pigment epithelium (RPE) cells. RPE cells were untreated (A, B, C), treated with buthionine-(S,R)-sulfoximine (BSO) for 24 h (D, E, F), or pretreated with 20 ng/mL hepatocyte growth factor (HGF) for 24 h prior to BSO (G, H, I) treatment. Carboxy-H₂DCFDA (for labeling ROS, panels B, E, H) and CMXRos (for mitochondria staining, panels A, D, G) were added to the cultures for 60 and 30 minutes before the end of the treatment, respectively. Double labeling for ROS and mitochondria was shown in merged images (C, F, I). BSO treatment induced abundant ROS production (E), and ROS accumulation in mitochondria (F). HGF effectively blocked ROS accumulation in mitochondria (H, I).

Figure 4.

 

Measurement of lipid peroxide (LPO) production in retinal pigment epithelium (RPE) cells. Malondialdehyde (MDA) and 4-hydroxy nonenal (4-HNE) levels in RPE cells were determined with a colorimetric kit. Buthionine-(S,R)-sulfoximine (BSO) treatment significantly increased lipid peroxide (LPO) production, and LPO returned to control level by hepatocyte growth factor (HGF) pretreatment. Data are mean ± SEM (n = 3). One and two asterisks indicate significant difference at P < 0.01 versus control and BSO groups, respectively.

Figure 5.

 

Release of cytochrome c (cyt c) from mitochondria. Protein isolated from cytosol and mitochondria was subjected to SDS-polyacrylamide gel electrophoresis and probed by anti-cytochrome c antibody as described in Materials and Methods. Buthionine-(S,R)-sulfoximine (BSO) treatment induced cyt c release to cytosolic compartment, and hepatocyte growth factor (HGF) treatment inhibited this effect.

Figure 6.

 

Determination of caspase-3 activation in retinal pigment epithelium (RPE) cells with buthionine-(S,R)-sulfoximine (BSO) treatment. Caspase-3 activation was measured by staining with FITC-VAD-FMK. Cells with increased caspase-3 like activity showed higher fluorescence intensity. Treatment with BSO induced a significant activation of caspase-3 that was partially blocked by hepatocyte growth factor (HGF) pretreatment. The figure also shows the induction of active caspase with BSO and partial block of activation with HGF pretreatment.

Figure 7.

 

Effect of buthionine-(S,R)-sulfoximine (BSO) treatment on Bcl-2 expression. Retinal pigment epithelium (RPE) cells were treated with 200 uM BSO for 24 h with or without pretreatment overnight with 20 ng/mL hepatocyte growth factor (HGF). Bcl-2 expression was examined using a rabbit polyclonal anti-Bcl-2 antibody as described in Materials and Methods. BSO caused a significant downregulation of Bcl-2 expression, which was partially restored by HGF treatment.

Figure 8.

 

Buthionine-(S,R)-sulfoximine (BSO)-induced changes in glutathione (GSH)-related genes and effect of hepatocyte growth factor (HGF). Changes in the gene expression of glutathione S-transferase (hGST5.8), gamma glutamylcysteine synthetase (γ-GCS) and glutathione peroxidase-1 (GPX1) were quantitated by real time-PCR. The details on generation of primers are described in Materials and Methods. In the case of γ-GCS, primers for the heavy sub unit (hGCS) and light subunit (lGCS) were generated, but the figure shows the results only from hGCS. The pattern of changes with lGCS was similar. Not shown are the data from hepatocyte growth factor (HGF) treatment of retinal pigment epithelium (RPE), which did not differ from control value for GST5.8 and γ-GCS, and which showed a slight increase for GPX. The P values for different comparisons with significant differences are shown.

Figure 9.

 

Effect of buthionine-(S,R)-sulfoximine (BSO) treatment on the expression and translocation of NF-E2-related factor (Nrf2). Details on the polyclonal Nrf2 antibody, the FITC conjugated secondary antibody and confocal microscopic analysis are described in Materials and Methods. DAPI staining was used to visualize the nucleus. FITC images for the four conditions viz. untreated control (A), hepatocyte growth factor (HGF) (20 ng/mL for 6 hours) (B), BSO (500 μM) alone for 1 hour (C), and HGF pretreatment (6 hours) with BSO (1 hour) treatment (D) are shown. This figure represents results from one of three donors studied. The Nrf2 expression varied from donor to donor and in one donor, it was more prominent with HGF treatment alone than what is presented.

The authors thank Christine Spee and Ernesto Barron for technical support.