Oxidative Damage to Human Lens Epithelial Cells in Culture: Estrogen Protection of Mitochondrial Potential, ATP, and Cell Viability

Abstract

purpose. Epidemiologic studies demonstrate a higher incidence of cataracts in estrogen-deprived postmenopausal women, but the mechanism for the increased risk of cataracts is unclear. An elevated level of H₂O₂ in aqueous humor and whole lenses has been associated with cataractogenesis. In the present study, for the first time, the protective effect of estrogens against oxidative stress were tested in cultured human lens epithelial cells (HLECs).

methods. To investigate the involvement of 17β-estradiol (17β-E₂) in protection against oxidative stress, HLECs were exposed to insult with H₂O₂ at a physiological level (100 μM) over a time course of several hours, with and without pretreatment with 17β-E₂. Cell viability was measured by calcein AM assay, and 2′,7′-dichlorofluorescein diacetate (DCFH-DA) was used to determine intracellular reactive oxygen species (ROS). Intracellular adenosine triphosphate (ATP) level was quantified with a luciferin- and luciferase-based assay and mitochondrial potential (ΔΨ_m) was monitored by a fluorescence resonance energy-transfer technique.

results. H₂O₂ caused a dose-dependent decrease in mitochondrial membrane potential, intracellular ATP levels, and cell viability. Dose-dependent increases in cell viability and intracellular ATP level were observed with pretreatment of 17β-E₂ for 2 hours before oxidative insult. At 1 nM, 17β-E₂ increased cell viability from 39% ± 4% to 75% ± 3%, and at 100 nM or higher, it increased survival to greater than 95%. The level of intracellular ATP approached normal with 17β-E₂ at 100 nM or higher. Pretreatment with 17β-E₂ did not diminish intracellular ROS accumulation after exposure to H₂O₂. Moreover, two nonfeminizing estrogens, 17α-E₂ and ent-E₂, both of which do not bind to either estrogen receptor α or β, were as effective as 17β-E₂ in the recovery of cell viability. The estrogen receptor antagonist, ICI 182,780, did not block protection by 17β-E₂. Both 17β- and 17α-E₂ moderated the collapse of ΔΨ_m in response to either H₂O₂ or excessive Ca²⁺ loading.

conclusions. The present study indicates that both 17α- and 17β-E₂ can preserve mitochondrial function, cell viability, and ATP levels in human lens cells during oxidative stress. Although the precise mechanism responsible for protection by the estradiols against oxidative stress remains to be determined, the ability of nonfeminizing estrogens, which do not bind to estrogen receptors, to protect against H₂O₂ toxicity indicates that this conservation is not likely to be mediated through classic estrogen receptors.

Age-related cataracts are a leading cause of visual impairment and blindness, and an ever-increasing health problem with the aging of the world population. In the United States, approximately 1.35 million cataract surgeries are performed annually at a cost of more than $3 billion. ¹ Cataract represents a large financial burden on health-care systems, and there remains a need to develop effective pharmaceuticals for the prevention or treatment of cataract. 

There is a higher incidence of cataract in postmenopausal women than in age-matched men, which leads to the notion that the absence of estrogens may contribute to the increased risk. ^{2 3 4 5 6} Indeed, epidemiologic studies indicate beneficial effects of hormone replacement therapy (HRT) against cataract in postmenopausal women. ^{7 8 9 10 11 12} For example, the Beaver Dam Eye Study ¹⁰ and the Salisbury Eye Evaluation Project ⁷ have both found protective associations between hormone use and lens nuclear opacity. In addition, another large cross-sectional study, the Blue Mountains Eye Study, ⁸ found that HRT was associated with reduced cortical opacity in lens. Recent epidemiologic reevaluation of the Blue Mountains Eye Study determined a significant trend for increasing incidence of nuclear cataract in postmenopausal women. ¹¹ Weintraub et al., ¹² recently evaluated HRT and lens opacities in a population of 480 postmenopausal women and determined that “current use of estrogen-only preparations was associated with a 49% decreased risk of nuclear opacities compared with never use.” Studies using tissue culture and animal models also suggest beneficial effects of estrogen in lens. In a lens culture system, estrogen protected lenses against cataracts induced by transforming growth factor (TGF)-β. ¹³ Estrogen has also been reported to exert protective effects in a rat model of age-related cataracts induced by methylnitrosourea (MNU). ¹⁴  

Several studies have demonstrated the beneficial effects of the antioxidant activity of estrogen and, further, that the hormone’s action is independent of classic receptor-dependent mechanisms. Our laboratory has shown that estradiol at physiological concentrations can block membrane oxidation. ¹⁵ Estrogen treatment has been shown to reduce lipid peroxidation induced by glutamate and further to attenuate the acceleration of intracellular peroxide production resulting from exposure to H₂O₂ ¹⁶ and by mitochondrial electron transport inhibitors. ¹⁷ Consistent with these data are studies showing that estrogen inhibits formation of lipid peroxyls and oxidation of low-density lipoproteins in vitro. ^{18 19} In vivo studies have demonstrated that estrogen replacement therapy provided by transdermal patch reduces low-density lipoproteins. ²⁰ These effects of estrogen do not appear to require estrogen receptors (ERs), ^{21 22 23} suggesting that estrogen exerts antioxidant activities through ER-independent mechanisms. 

In the current study, we tested, for the first time, the protective effects of estrogens against oxidative stress using in vitro cultured human lens epithelial cells (HLECs). Elevated levels of H₂O₂ are found in the lenses and aqueous humor of patients with cataract, ^{24 25 26} and it is held that H₂O₂ is a major oxidant that contributes to formation of cataract. ²⁷ In the present study, we assessed the ability of 17β-estradiol (17β-E₂) to protect against the adverse effects of H₂O₂ on mitochondrial membrane potential (ΔΨ_m), intracellular adenosine triphosphate (ATP) levels, and cell viability in HLECs. Further, we attempted to assess the role of estrogen receptors (ERs) in this action of estrogens on lens viability by using two nonfeminizing estrogens that do not bind to ERs, 17α-estradiol (17α-E₂) and Ent-estradiol (ent-E₂), and an ER antagonist, ICI 182,780. 

Material and Methods

Chemicals

17β-E₂ and 17α-E₂ were purchased from Steraloids, Inc. (Wilton, NH). ICI 182,780 was purchased from Tocris (Ellisville, MO). The complete enantiomer of 17β-E2_, Ent-E₂, was synthesized by methods that we have previously described. ²⁸  

All steroids and ICI 182,780 were dissolved in ethanol at a final concentration of 10 mM and diluted to appropriate concentration in culture medium as required. Unless otherwise stated, steroid treatment to cell cultures involved a 2-hour preincubation followed by continued administration of the steroid in the presence of H₂O₂. Those cells receiving vehicle (in place of estradiol) pretreatment were maintained in fresh culture medium at the same final ethanol concentration. Control cells were maintained in culture medium with appropriate changes of fresh medium. In experiments involving the ER antagonist ICI 182,780, it was added 30 minutes before addition of 17β-E₂. 

H₂O₂ was purchased from Mallinckrodt Baker Inc. (Paris, KY). H₂O₂ was diluted with culture medium to final concentration before using. Calcein AM, 2,7-dichlorofluorescin diacetate (DCFH-DA), and ATP determination kits were purchased from Molecular Probes (Eugene, OR). 

Cell Culture

HLE-B3 cells, a human epithelial cell line immortalized by simian virus (SV)-40 viral transformation, ²⁹ were obtained from Usha Andley (Washington University School of Medicine, Department of Ophthalmology, St. Louis, MO) and cultured in Eagle’s minimal essential medium (MEM) supplemented with 20% fetal bovine serum (Hyclone Laboratories, Logan, UT) and 20 μg/mL gentamicin (Sigma, St. Louis, MO) in 150-cm² culture flasks at 37°C and 5% CO₂ and 95% air. All experiments were performed with HLE-B3 cells between passages 18 and 25. 

Measurement of Reactive Oxygen Species

The extent of cellular oxidative stress was estimated by monitoring the generation of reactive oxygen species (ROS) using the fluorescent dye DCFH-DA. Cells were plated 24 hours before initiation of the experiment at a density of 5000 cells per well in 96-well plates. Cells were loaded with DCFH-DA at a final concentration of 50 μM for 45 minutes. After incubation, DCFH-DA was removed, and cells were washed twice with 1× PBS (pH 7.4) and incubated with MEM containing 20% FBS with a bolus dose of H₂O₂ (50 and 100 μM) for 10 to 60 minutes, DCFH-DA fluorescence was determined at an excitation of 485 nm and an emission of 538 nm, by microplate-reader (model FL600; Biotek, Highland Park, VT). Values were normalized to the percentage in untreated control groups. It should be noted that, “DCFH-DA is taken up by cells and tissues, usually undergoing deacetylation by esterase enzymes. Oxidation of DCFH within cells leads to fluorescent dichlorofluorescein, which can easily be visualized (strong emission at 525 nm with excitation at 488 nm). This technique is becoming popular as a means of visualizing ‘oxidative stress’ in living cells. In addition to peroxidase/H₂O₂, several species cause DCFH oxidation, probably including RO₂·, RO·,OH·, HOCl, and ONOO⁻, but not O₂·⁻or H₂O₂. Hence, this ‘fluorescent imaging’ is an assay of generalized ‘oxidative stress’ rather than of production of any particular oxidizing species, and it is not a direct measure of H₂O₂.” ³⁰  

Calcein AM Assay

Cells were plated 24 hours before the initiation of the experiment, at a density of 5000 cells per well in 96-well plates. Cells were exposed to two doses of H₂O₂ (50 and 100 μM) from 1 to 24 hours. After exposure to H₂O₂, cells were rinsed with 1× PBS (pH 7.4), and viability was assessed by the addition of 25 μM calcein AM, as previously described. ²⁸ Calcein AM fluorescence was determined at an excitation of 485 nm and an emission of 538 nm with the microplate reader (FL600; Biotek). Percentage viability was calculated by normalization of all values to the H₂O₂-free control group (100%). Calcein staining was visualized by fluorescence microscope (Diaphot-300; Nikon, Tokyo, Japan), and cells were photographed for qualitative documentation. Four random fields of cells were examined, and photographs were taken of cells in 96-well plates. 

Measurement of ATP Levels

Cells were plated at a density of 5 × 10⁵ cells per well in 12-well plates. After 48 hours, cells were exposed to various doses of H₂O₂ from 15 minutes to 8 hours. Cellular ATP levels were quantified with a luciferin and luciferase-based assay. ³¹ Cells were washed with PBS once and lysed with ATP-releasing buffer (100 mM potassium phosphate buffer [pH 7.8]: 1% Triton X-100, 2 mM EDTA, and 1 mM dithiothreitol [DTT]). Ten microliters of the lysate was added to 96-well plates (InterMed, Naperville, IL) . ATP concentrations in lysate were quantified using an ATP-determination kit according to the manufacturer’s instruction. The 96-well plates were then read (SpectraMAX GeminiXS plate reader; Molecular Devices, Sunnyvale, CA). A standard curve was generated with solutions of known ATP concentrations. Protein concentration of samples were determined by Bradford assay. ³² ATP levels were calculated as nanomolar ATP per milligram protein and normalized to levels in untreated control cultures. 

Monitoring ΔΨ_m

ΔΨ_m was recorded in intact and digitonin-permeabilized cells with an assay based on fluorescence resonance energy transfer (FRET) between two dyes: nonyl acridine orange (NAO; Molecular Probes), which stains cardiolipin, lipid found exclusively in the mitochondrial inner membrane, and tetramethylrhodamine (TMRE; Molecular Probes), a potentiometric dye taken up by mitochondria in accord with Nernstian dictates potential and concentration. The presence of TMRE quenches NAO emission in proportion to ΔΨ_m, whereas loss of ΔΨ_m with consequent efflux of TMRE dequenches NAO. ^{33 34} The high specificity of NAO staining; selective monitoring of the fluorescence emitted by NAO, not TMRE; and the stringent requirement for colocalization of both dyes within the mitochondrion, all act in concert to allow the FRET assay to report ΔΨ_m, unconfounded by background signal arising from potentiometric dye responding to plasma membrane potential_.  

Twenty-four hours before assay, cells were trypsinized and plated in clear-bottomed, black-walled, 96-well plates (Costar 3606; Corning International, Corning, NY). Cells were plated at 60,000 per well for use in high-throughput screening protocols, as described previously. ^{16 28}  

Statistical Analyses

Effects on ΔΨ_m were quantified by calculating the area under the curve (AUC) after either Ca²⁺ challenge or addition of H₂O₂. To partially correct for variations in cell density, staining and optical aberrations in these plates, all wells were normalized to the initial relative fluorescence units (RFU) reading in each well using a fluorescence-imaging plate reader (FLIPR; with accompanying software; Molecular Devices). In a variation of the analysis, the AUC was divided by the amount of initial quenching of NAO, which is an alternate technique to compensate for variability in cell density and staining and for optical aberrations. Dose–response data were fitted on computer by nonlinear regression analysis (sigmoid equations; Prism, ver. 3.00 for Windows; GraphPad Software, San Diego, CA). One way ANOVA and Bonferroni post hoc testing were performed with the same software. 

The significance of differences among groups was determined by one-way ANOVA. Planned comparisons between groups were determined by the Tukey test. For all tests, P < 0.05 was considered significant. 

Results

Effects of H₂O₂

ROS Accumulation.

As shown in Figure 1 a and as measured by DCF fluorescence intensity, exposure to H₂O₂ caused an increase in intracellular ROS accumulation in the cultured HLECs. By 60 minutes after administration, 50 μM H₂O₂ initiated a modest but linear increase in ROS content (172% ± 11%) over control cells. A higher dose of H₂O₂ (100 μM) prompted a biphasic accumulation of ROS in the cultured cells. After 40 minutes, there was an ROS buildup resulting in a moderate intracellular augmentation by 40 minutes (200% of control), but accumulation was amplified to approximately 500% of control by 60 minutes. 

Intracellular ATP Content.

Administration of both 50 and 100 μM H₂O₂ triggered a rapid decrease in intracellular ATP levels (Fig. 1b) . Within 60 minutes, 50 and 100 μM H₂O₂ reduced ATP by 68% ± 8% and 50% ± 5% of control, respectively. Whereas the ATP levels appeared to slightly recover 2 hours after administration of H₂O₂, the overall trend was for ATP levels to decline at either H₂O₂ dose to approximately 40% of control over the remainder of the time course. 

Cell Viability.

H₂O₂ induced a time- and dose-dependent decline in cell viability (Fig. 1c) . At 50 μM, it decreased cell viability to 65% ± 2% of control by 12 hours after exposure. Exposure to 100 μM H₂O₂ had a significantly greater effect on viability of HLECs, with the surviving of cells declining to 45% ± 4% by 2 hours and almost complete cell death by 8 hours after H₂O₂ treatment. 

Effects of 17β-E₂ against H₂O₂ Exposure on HLECs

Because 100 μM H₂O₂ had a significant impact on ROS accumulation, intracellular ATP content, and cell viability, this concentration was selected for further studies with estrogens. 

H₂O₂-Induced ROS Increase in HLECs.

As in Figure 1a , 100 μM H₂O₂ progressively and significantly increased intracellular ROS over the 60-minute observation period. Concentrations of 17β-E₂ ranging from 1 nM to 10 μM did not modify intracellular accumulation of ROS (Fig. 2) . 

H₂O₂-Induced ATP Decline in HLECs.

17β-E₂ was examined for its protective capacity against the decline of intracellular ATP caused by exposure of cultured HLECs to exogenous H₂O₂ (Fig. 3) . A 2-hour pretreatment with 17β-E₂ reversed the decline of intracellular ATP brought about by H₂O₂ in a dose-dependent manner. Whereas a low concentration of 17β-E₂ (10 nM) did not alleviate the ATP loss in response to H₂O₂ treatment, ATP levels were restored from 75% ± 4% (peroxide treatment) to 93% ± 9% of control with 100 nM 17β-E₂. At concentrations above 1 μM, 17β-E₂ completely normalized the intracellular ATP pool. 

H₂O₂-Induced Cell Death in HLECs.

17β-E₂ protected against HLE-B3 cell loss due to H₂O₂ toxicity in a dose-dependent manner. H₂O₂ (100 μM) resulted in a 61% ± 4% decline in cell survival. A 2-hour preincubation with 17β-E₂ (1 nM) increased cell survival from 39% ± 4% to 75% ± 3%. At 100 nM and higher, 17β-E₂ completely protected against H₂O₂ toxicity (Fig. 4) . 

Effects of 17α- and 17β-E₂ on Ca²⁺- and H₂O₂-Mediated Collapse of ΔΨ_m

As is the case with other cell types, ^{33 34} acute increases in cytosolic Ca²⁺ induced through the ionophore ionomycin caused a dose-dependent collapse of ΔΨ_m in intact HLECs (Fig. 5) . Even a brief 5-minute incubation with either 17α- or 17β-E₂ at 0.5 μM substantially reduced the magnitude of this ionomycin-induced ΔΨ_m collapse (Fig. 5) , reflected by an increase in EC₅₀ from 0.95 μM to 1.6 and 2.3 μM for 17α- and 17β-E₂, respectively. Thus, it required more Ca²⁺ to induce comparable ΔΨ_m collapse when the estradiols were present, or conversely, under comparable Ca²⁺ loading, a larger portion of the mitochondrial population retained its membrane potential in the presence of estradiols. Note also that the magnitude of the ΔΨ_m collapse in the presence of 17β-E₂, expressed as total AUC, was lower than with buffer and 17α-E₂, even at the highest ionomycin concentrations (Fig. 5) . 

Similar responses were observed when HLECs were exposed to 2.5 mM H₂O₂, as an acute cytotoxic stimulus (Fig. 6) . As might be expected, the amount of H₂O₂ needed to collapse ΔΨ_m acutely (response within seconds) was substantially more than the concentration necessary in the long-term cytotoxicity studies. In any event, the magnitude of ΔΨ_m collapse induced by H₂O₂ was moderated by both 17β- and 17α-E₂ at 0.5 μM (preincubation for 30 minutes), although this occurred predominantly at higher H₂O₂ concentrations (0.1 and 1 mM). For example, EC₅₀ was not significantly different between control and estradiol treatments, but one-way ANOVA of the total AUC during the response reveals significant moderation of ΔΨ_m collapse by both 17β- and 17α-E₂ (F = 3.9, P < 0.03). 

To model less acute oxidative stress that is likely to be more physiologically relevant, HLECs were incubated for 6 hours with 10 μM H₂O₂, with or without 17β- or 17α-E2, and then challenged with a Ca²⁺ load through application of ionomycin described earlier. 17β- or 17α-E₂ moderated ΔΨ_m collapse, repressing the magnitude of the response more than its rate (Figs. 7 8) . In all three replicates, the estradiols consistently increase the EC₅₀ for ionomycin (Fig. 8 ; Table 1 ). To control for potential effects of the compounds on cell growth or loss during the prolonged (6-hour) preincubation, the AUC data from each well were normalized to the cell density in that well at the start of the observations by dividing the AUC by the magnitude of quenching of the initial fluorescence signal by the potentiometric dye (A, initial signal; B, signal after quenching). Thus, AUC/(A − B) provides an index of compound efficiency that is independent of effects on cell viability. ³⁴  

Evidence that 17β-E₂ Protection against Oxidative Stress Is Not Mediated by Binding to Estrogen Receptors

To exclude a possible connection for ER-dependent binding with the protection afforded by 17β-E₂ against H₂O₂ induced cytotoxicity, two nonfeminizing estrogens that exhibit marginal (17α-E₂) or no (Ent-E₂) binding capacity to ERs were examined. Compared with the potent natural estrogen 17β-E₂, 17α-E₂ binds weakly to ERs, and the 17α-E₂–ER complex only transiently binds to the estrogen-responsive element. ^{28 35 36 37} Ent-17β-E₂, the enantiomer of 17β-E₂, has identical physiochemical properties as 17β-E₂, with the crucial exception that it is incapable of interacting with other stereospecific molecules, such as the ERs. ²⁸ At 100 nM, 17β-E₂ improved cell survival from 32% ± 3% to 76% ± 3%. Both 17α-E₂ and Ent-E₂ displayed equivalent effectiveness against H₂O₂-induced cytotoxicity. 17α-E₂ increased cell survival to 73% ± 3% and Ent-E₂ enhanced cell viability to 79% ± 4% (Fig. 9) , suggesting that the protective action of 17β-E₂ is not mediated by classic ERs. 

To further exclude a role for ERs in estrogen protection against H₂O₂-induced cytotoxicity, an ER antagonist ICI 182,780 was examined because of its ability to prevent 17β-E₂ from binding to ERs. The structure of ICI 182,780 is similar to 17β-E₂, and it effectively competes with 17β-E₂ for binding to ERs. As shown in Figure 10 , 100 μM H₂O₂ killed 79% ± 8% of cells, whereas 100 nM 17β-E₂ protected 77% ± 5% of cells from the H₂O₂ insult. Coadministration of ICI 182,780 and 17β-E₂ did not block the protection provided by17β-E₂ against H₂O₂ insult. At the same time, administration of 100 nM ICI 182,780 alone increased cell survival to 65% ± 6%, a result not entirely unexpected, because ICI 182,780 is structurally similar to 17β-E₂. Taken together, these data support our contention that the protection against H₂O₂ toxicity afforded by17β-E₂ is not mediated by ERs. 

Discussion

We report, for the first time, that the cell death induced in HLECs by H₂O₂ is associated with production of intracellular ROS, collapse of ΔΨ_m, and profound depletion of ATP and that estrogens potently protected against collapse of ΔΨ_m, ATP depletion, and cell death without affecting production of ROS such as H₂O₂, RO₂·, RO·, OH·, HOCl, and ONOO⁻. Collectively, these data suggest that the reported protection from cataracts afforded by HRT in postmenopausal women ^{7 8 9 10 11 12} is due to these cytoprotective effects against H₂O₂ toxicities in lens epithelial cells. 

H₂O₂ is a potent diffusible pro-oxidant that initiates a series of oxidative events in cells. ^{38 39} We observed that HLECs adaptively responded to a low concentration (50 μM) of H₂O₂, as evidenced by a modest increase in ROS, maintenance of stable, albeit lower, concentrations of ATP, and relative resistance to cell death. In contrast, the higher (100 μM) concentration of H₂O₂ was associated with a delayed but profound increase in ROS, a rapid and marked decline in ATP concentrations, and nearly complete cell death within 8 to 12 hours. This higher concentration of H₂O₂ is clearly a pathologic insult, from which cells failed to recover under the conditions of this study. 

Our first assessment of the effects of 17β-E₂ was on production of ROS, using an indicator that detected soluble ROSs. In this assay, 17β-E₂ at concentrations ranging from low physiological (1 nM) to pharmacologic (10 μM) were ineffective in changing the ROS response to H₂O₂. Clearly, the cytoprotective effects of 17β-E₂ in this cell line are not dependent on its ability to affect the production or clearance of soluble ROS. A possibility that we have not explored in the present study is that estrogens may affect lipid peroxidation, without affecting soluble ROS. Estrogens are lipid soluble and preferentially penetrate in cellular membranes. ⁴⁰ In neuronal cultures, physiological concentrations of 17β-E₂ effectively reduce lipid peroxidation. ^{22 41} However, it is clear from our data that in HLECs, soluble ROS are not influenced by exposure to estrogen. 

The ATP depletion induced by H₂O₂ treatment no doubt reflects at least two actions of the pro-oxidant: interruption of oxidative production of ATP with the concomitant depletion of the energy-containing molecule as a result of attempts to repair damage caused by H₂O₂. Either or both mechanisms may be involved in the depletion of ATP, although we have observed that H₂O₂ causes a profound downregulation in several oxidative phosphorylation enzyme transcripts (Cammarata PR, Moor AN, unpublished observations, 2002), an effect that would certainly help undermine ATP production. Moreover, we do not exclude the well-known dramatic inactivation of glyceraldehyde-3-phosphate dehydrogenase by H₂O₂ as part of the explanation for the decline in intracellular ATP. ⁴² Whether estradiol prevents the inactivation of glyceraldehyde-3-phosphate dehydrogenase by peroxide, thereby contributing to the restoration of intracellular ATP, is currently under investigation. 

On shorter time scales, it is well known that oxidative insult readily represses electron transport efficiency and oxidative phosphorylation, primarily by inactivating both the Fe-S reaction centers of several of the electron transport respiratory centers, and heme moieties in the cytochromes. ^{39 43} In addition, even without the downregulation of mitochondrial gene expression noted earlier, the data show that H₂O₂ directly collapses ΔΨ_m in HLECs, an event that not only eliminates the driving force for mitochondrial ATP production, but that also exacerbates subsequent free radical production. ^{39 43}  

Damage to mitochondria can lead to deficiency in ATP production and to a concomitant increase in production of ROS that can overwhelm cellular antioxidant defense systems. Under conditions of oxidative stress, mitochondria undergo a catastrophic, irreversible loss of the impermeability of the inner mitochondrial membrane that causes a complete collapse of ΔΨ_m, a process called permeability transition (PT). ⁴⁴ Accelerated mitochondrial radical production compromises cellular and mitochondrial integrity by inducing peroxidation of membrane lipids and impeding oxidative phosphorylation. The resultant acute loss of ATP causes the transmembrane ion-dependent ATPases to fail, thereby precipitating cell death from osmotic failure. ^{39 45}  

17β-E₂ was effective in protecting cellular ATP levels and in protecting HLECs from death. A dose-dependent increase in cellular ATP levels was observed from 100 nM to 10 μM. These data suggest that at pharmacologic concentrations of estrogens, cellular ATP is preserved. Inasmuch as ATP is essential for normal cellular function, including its survival, this action of 17β-E₂ may be necessary, but insufficient for the observed cytoprotective effects. Indeed, it appears that the cytoprotective effects of 17β-E₂ occur at concentrations lower than those needed for ATP maintenance. As such, other actions of 17β-E₂ are involved in its cytoprotective effects. Although we do not know the precise mechanism of the cytoprotective effects of estrogens in HLECs, in neurons a plethora of cellular responses to the steroid have been reported, including the protection of mitochondrial function, stimulation of antiapoptotic proteins, and stimulation of protective signaling pathways. ^{46 47 48 49 50 51}  

The data in the current study indicate that both 17α- and β-E₂ equipotently increased the amount of Ca²⁺ or H₂O₂ necessary to collapse ΔΨ_m in HLECs, effectively stabilizing mitochondrial integrity and preserving function under pathogenic conditions. This effect does not require prolonged exposure to the estradiols—it became apparent in 5- and 30-minute incubations—yet it was also apparent after a 6-hour preincubation. The result is that, at a given Ca²⁺ or oxidative load, a larger portion of the mitochondrial population retains ΔΨ_m and hence continues to function. Such a response readily explains preservation of ATP levels by estradiols during exposure to H₂O₂, as well as repression of cell death through both necrosis and apoptosis under these conditions. 

The mitoprotective effects of the estradiols shown in the ΔΨ_m assay could be due to any combination of the mechanisms of action known for this class of compounds ^{46 47 48 49 50 51} such as membrane stabilization, ⁴⁰ which is particularly germane to the retention of ΔΨ_m. Indeed, the moderation by estrogens of ΔΨ_m collapse could be due to a repression of Ca²⁺ uptake into the mitochondria through the uniporter, to increased Ca²⁺ efflux from the mitochondria, or to a direct membrane-stabilization effect, all of which would yield similar-appearing responses in this assay. In addition, although both 17α- and β-E₂ equipotently repressed the magnitude of ΔΨ_m collapse induced by either Ca²⁺ or H₂O₂ in the FRET assay (Figs. 5 6 7 8) , the data do not permit distinguishing a modest effect in most of the mitochondrial population from a profound effect in a smaller mitochondrial subpopulation. It is apparent, however, that in the absence of additional stressors, the estrogens alone do not dissipate or hyperpolarize ΔΨ_m, as is reflected by comparable amounts of initial quenching, seen as coincidence of the curves between 100 and 300 seconds before the addition of ionomycin (Fig. 7) . At the very least, this suggests that the estrogens do not exert their protective effects by uncoupling electron transport, a mechanism known to protect neuronal cells from oxidative stress and Ca²⁺ loading associated with glutamate excitotoxicity. ⁵²  

Our results argue against a primary involvement of ERs in the observed effects of estrogens. However, we have, in fact, confirmed the positive presence of ER-α and -β in the cultured HLEC strain HLE-B3 by RT-PCR analysis and subsequent verification of the PCR products by sequence analysis and Southern blot with specific internal oligonucleotides to the directed primer pairs, as well as by immunofluorescence techniques (manuscript in preparation). Three estrogens, 17β-, 17α-, and Ent-E₂, that differ by as much as 32-fold in their affinity for either ER-α or -β, ^{28 35 36 37} have equivalent effects on HLEC survival and the action of 17β-E₂ was not antagonized by the prototypic ER antagonist ICI 182,780. The ICI compound itself exerted cytoprotective activity, probably the result of its phenolic A ring, a requirement for cytoprotection by estrogens. ^{21 23 45} In this respect, a recent study by Han et al., ⁵³ demonstrated that the protective potency of various estrogens was dependent on the precise estrogenic structure. Whereas 17α-E₂, a phenolic ring estrogen, acted similar to the antioxidants taurine and vitamin C against the peroxide-induced damage to cultured rabbit renal proximal tubule cells, 17β-E₂, a catecholic estrogen, behaved in a manner similar to the iron chelators deferoxamine and phenanthroline. In this regard, it is important to further point out that superoxide dismutase mimics, in particular, TEMPOL (Sigma), ⁵⁴ prevents Fe⁺²–mediated generation of the damaging hydroxyl radical, by reacting with superoxide, thus preventing recycling of Fe⁺³ to Fe⁺², while deferoxamine chelates intracellular Fe⁺³ and prevents its reduction to Fe⁺². We cannot at this time rule out the possibility that 17β-E₂, like the superoxide dismutase (SOD) mimic TEMPOL or deferoxamine, acts by limiting the availability of Fe⁺² and thereby prevents certain damaging effects of H₂O₂. Irrespective of the precise mode of action of 17β-E₂, the study by Han et al., ⁵³ raises the interesting possibility that various estrogens have differential cytoprotective potential, and by inference, disparity in their mechanisms of action. Of immediate relevance to our studies, however, Han et al., like us, conclude that “these cytoprotective effects of estrogens are not dependent on classical estrogen receptors.” ⁵³  

In contrast, a recent study by Davis et al., ⁵⁵ argues that estrogen protection in the eye may result from direct interactions with its ocular ERs. Studies in their transgenic mouse model, which express ER-Δ, a dominant-negative form of ER-α that inhibits ER-α function, show spontaneous cortical cataracts that progress with age in transgene-positive women after puberty. 

Collectively, our studies demonstrate that estrogens are potent cytoprotectants that preserve mitochondrial function during oxidant insult in HLECs in culture. These results indicate that estrogens may be useful therapies for the prevention of cataracts in postmenopausal women and that nonfeminizing estrogens could provide similar protection in men. 

 Submitted for publication August 16, 2002; revised November 1 and November 25, 2002; accepted December 5, 2002.

 Disclosure: X. Wang, None; J.W. Simpkins, None; J.A. Dykens, MitoKor (E); P.R. Cammarata, 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.

 

(a) Effects of H₂O₂ on ROS accumulation in HLECs. The control group was not treated with H₂O₂ after loading DCFH-DA. Depicted mean ± SEM (n = 8) percentages of DCF fluorescence normalized to the control (no H₂O₂) at each sampling time. (b) Time course of the effects of H₂O₂ (50 and 100 μM) on intracellular ATP levels in HLE-B3 cells. Data are expressed as a percentage of control levels (non–H₂O₂-treated cells at each sampling time) and represent the mean ± SEM of determinations made in four to six cultures per group. (c) Time course of the effects of H₂O₂ on viability of HLE-B3 cells. Data are expressed as a percentage of control group (non–H₂O₂-treated cells at each sampling time) and represent the mean ± SEM (n = 8). In all panels, when SEM bars are not shown, they are obscured by the symbol.

Figure 2.

 

Effect of 2 hours of 17β-E₂ pretreatment on ROS production in HLE-B3 cells after 100 μM H₂O₂ treatment. ROS production was measured at indicated time points. Data are expressed as a percentage of vehicle control levels and represent mean ± SEM (n = 8).

Figure 3.

 

Effect of 2 hours of pretreatment with 17β-E₂ on ATP levels in HLE-B3 cells treated for 90 minutes with 100 μM H₂O₂. The vehicle group was pretreated with the equal amount of vehicle. Data are expressed as a percentage of normal levels and represent the mean ± SEM of determinations made in six to eight cultures per group. *P < 0.05 versus respective vehicle group; **P < 0.05 versus respective H₂O₂ group; ***P < 0.05 versus respective H₂O₂ group.

Figure 4.

 

(a) Dose-dependent effects of 2 hours of pretreatment with 17β-E₂ on viability of HLE-B3 cells at 8 hours after 100 μM H₂O₂ treatment. The vehicle group (−H₂O₂ and 17β-E₂) was pretreated with the equal amount of vehicle. Data are expressed as a percentage of cells surviving and represent the mean ± SEM (n = 8). ***P < 0.001 versus the H₂O₂ group. (b) Photomicrographs of the dose-dependent effects of 2 hours of pretreatment with 17β-E₂ in HLE-B3 cells at 8 hours of 100 μM H₂O₂ treatment in calcein AM–loaded cells.

Figure 5.

 

Effect of 17β-E₂ (▵) and 17α-E₂ (□) on the collapse of ΔΨ_m, resulting from ionomycin-induced calcium loading. Both estradiols (0.5 μM, 5 minutes of preincubation) significantly moderated the magnitude of ΔΨm collapse compared with the control (▪), as reflected by the EC₅₀ of 2.31 and 1.55 μM for 17β-E₂ and 17α-E₂, respectively, compared with control EC₅₀ of 0.95 μM. Data are expressed as the AUC after addition of ionomycin and represent the mean ± SEM of four wells per group. When SEM bars are not shown, they are obscured by the symbol.

Figure 6.

 

Effect of 17β-E₂ (▵) and 17α-E₂ (□) on the acute collapse of ΔΨ_m, resulting from H₂O₂ exposure. Both estradiols (0.5 μM with 30 minutes of preincubation) significantly moderated the magnitude of ΔΨ_m collapse compared with the control (▪), as reflected by significant differences in the AUCs. In this case, there were no significant differences between the EC₅₀ values (0.8, 1.1, and 1.5 μM for 17β-E₂, 17α-E₂, and control, respectively), although one-way ANOVA of the AUCs reveals that both estradiols significantly suppressed the ΔΨ_m collapse caused by H₂O₂ (F = 3.9, P < 0.03). This suggests that the differences between the treatments are dominated by divergence at the highest H₂O₂ concentrations. Regression coefficients (r ²) are 0.96, 0.97, and 0.96 for 17β-E₂, 17α-E₂, and control, respectively. Data are expressed as the AUC after addition of H₂O₂ and represent the mean ± SEM of four wells per group. When SEM bars are not shown, they are obscured by the symbol.

Figure 7.

 

Time course of the effects of 17β-E₂ (▵), 17α-E₂ (□) or control buffer (no estrogen, ▪) on ionomycin-mediated ΔΨ_m collapse in H₂O₂-treated cells, using the FRET assay. To more closely mimic the in vivo situation, before Ca²⁺ challenge, the lens cells were incubated for 6 hours with 0.5 μM estradiols plus 10 μM H₂O₂. Addition of a potentiometric dye, TMRE (arrow at 50 seconds) quenched the initial NAO signal (A), whereas addition of ionomycin (5 μM, arrow at 320 seconds) collapsed ΔΨ_m, thereby releasing the dye from the mitochondria. Efflux of the dye dequenched the NAO, as is seen by the sharp recovery of the NAO signal. ^{32 33} Data are expressed as mean RFU. Treatments were compared using the AUCs after Ca²⁺ (or H₂O₂) challenge.

Figure 8.

 

Effects of estradiol on ΔΨ_m collapse after long-term exposure to H₂O₂. Before Ca²⁺ challenge, the lens cells were treated as in Figure 7 . In this protocol, both 17β-E₂ (▵) and 17α-E₂ (□) moderated the acute collapse of ΔΨ_m resulting from exposure to ionomycin, compared with the control (▪). This was reflected by a consistent increase in EC₅₀ over the control (Table 1) , indicating mitochondrial stabilization by estradiols. Regression coefficient values are in Table 1 . Data are expressed as AUC/A − B after addition of ionomycin and represent the mean ± SEM of four wells per group. When SEM bars are not shown, they are obscured by the symbol.

Table 1.

 

EC₅₀ and Regression Correlation Coefficient for Effect of Estrogens in Long-Term Exposure to H₂O₂

Exp.	Buffer	17α-E2	17β-E2
1	0.8 (0.89)	1.1 (0.96)	1.4 (0.97)
2	2.3 (0.93)	2.8 (0.96)	3.0 (0.94)
3	2.1 (0.92)	3.1 (0.85)	3.7 (0.84)

Figure 9.

 

Effects of 2 hours of pretreatment with 17β-E₂, 17α-E₂, and Ent-E₂ on cell viability in HLE-B3 cells for 8 hours at 100 μM H₂O₂ treatment. Data are expressed as a percentage of the vehicle group and represent the mean ± SEM (n = 8). ***P < 0.001 versus the H₂O₂ group.

Figure 10.

 

Effects of ICI 182,780 on 17β-E₂ protection in HLE-B3 cells after 8 hours of 100 μM H₂O₂ treatment. Data are expressed as a percentage of cells surviving compared with the vehicle group and represent the mean ± SEM (n = 7–8). *P < 0.05 and ***P < 0.001 versus the H₂O₂ group.