Mechanisms of Apoptosis in Human Retinal Pigment Epithelium Induced by TNF-α in Conditions of Heavy Metal Ion Deficiency

Abstract

purpose. To investigate the mechanism underlying apoptosis in retinal pigment epithelium (RPE) induced by TNF-α in conditions of heavy metal ion deficiency.

methods. Apoptotic morphology was assessed with Hoechst 33342. FITC-VAD-fmk or antibody specific to cleaved caspase 3 was used to detect in situ activated caspases or cleaved caspase 3, respectively. JC-1 and carboxylated dichlorodihydrofluorescein diacetate were used as probes to measure mitochondrial membrane potential (Δψ_m) and intracellular reactive oxygen species (rOx).

results. The apoptotic response of RPE cells was markedly enhanced when TNF-α plus actinomycin D (act-D) was coapplied with N,N,N′,N′-tetrakis(2-pyridylmethyl)ethylenediamine (TPEN), a heavy metal ion chelator. The apoptosis was caspase dependent, and a blockade with cyclosporin A (CsA), an inhibitor of the mitochondrial permeability transition (MPT), but not FK506, a calcineurin inhibitor, abolished caspase activation and subsequent apoptosis, suggesting that apoptosis requires the MPT, and that caspase activation is downstream to the MPT. MPT, observed in situ as Δψ_m loss, was prevented when cells were pretreated with either CsA or the pan-caspase inhibitor z-VAD-fmk. This result suggests that apoptotic signaling is initiated by the MPT and further amplified by downstream caspases, probably through a feedback loop. An increase in rOx was observed in cells exposed to TNF-α+act-D+TPEN, and rOx production did not require MPT or caspase activation. In addition, application of antioxidants significantly inhibited rOx production, Δψ_m loss, and apoptosis. These data suggest that the rOx may play a role as a proapoptotic signal.

conclusions. The data suggest that intracellular heavy metal ions play a role in determining the apoptosis induction threshold of the inflammatory response to TNF-α in RPE.

Given the facts that many cytokines, such as TNF-α, mediate both proapoptotic and prosurvival signals ¹ and that many cells expressing the TNF family of receptors have the required components to execute apoptosis given appropriate apoptotic stimuli, these cells must be endowed with devices to regulate the apoptotic machinery to ensure that inappropriate activation of the apoptotic process does not occur. Findings in studies suggest that this can be achieved at multiple sites along apoptosis-signaling pathways such as through the inhibition of caspase activation by FLICE inhibitory protein (cFLIP), or by inhibitors of apoptosis proteins (IAPs), ² or by the suppression of mitochondrial membrane permeability by up- and/or downregulation of expression of pro- or antiapoptotic bcl-2 family members. ³  

In addition to the repression of apoptosis through protein–protein interactions, many studies indicate that intracellular heavy metals may play a role in determining the apoptosis induction threshold. Investigators in many studies using chelators to study the role of heavy metal ion deficiency in apoptosis reached an important conclusion that the cellular content of heavy metal ions determines the induction threshold of apoptosis (that is, the depletion of intracellular heavy metal ions may lead to apoptosis, or increase vulnerability of cells to other cell stresses. ^{4 5 6 7 8} Although these studies showed that during the apoptotic process, many biochemical events occur (e.g., cytochrome c [cytc] translocation, caspase activation, cellular binding of annexin V, DNA fragmentation, and cleavage of several cellular targets), the apoptotic signaling involved in the premitochondrial and mitochondrial phases of the apoptotic process was not determined. ³  

In this study, using cultured human RPE as a model, we investigated the molecular events underlying apoptosis induced by TNF-α in the presence of N,N,N′,N′-tetrakis(2-pyridylmethyl)ethylenediamine (TPEN; at a sublethal concentration) to deplete the intracellular content of heavy metal ions. Our data suggest that reactive oxygen species (rOx) evoked in response to inflammatory cytokine stimulation under conditions of heavy metal ion deficiency induces the mitochondrial permeability transition (MPT) and, as a consequence, the apoptotic cell death of RPE. Through a molecular approach, our study provides a unique perspective into the mechanisms of how signals evolve from inflammation, and how heavy metal ion deficiency may become integrated into a proapoptotic signal leading to apoptosis in the RPE. 

Materials and Methods

Cell Culture

The adult human RPE cell line ARPE-19 from American Type Tissue Collections (Manassas, VA) was cultured in Dulbecco’s modified essential medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and penicillin/streptomycin. Cultures were maintained in an air–5% CO₂ incubator at 37°C and constant humidity. Most of the experiments were conduced when culture density reached 70% to 80% of confluence, usually 2 days after passage. Right before the experiments, the culture medium was replaced with DMEM containing 1% FBS. Under this culture condition, cells did not show significant changes of morphology within 48 hours. All culture reagents were from Invitrogen (Carlsbad, CA). 

Immunocytochemical Detection of Activated Caspase 3

Immunolabeling was performed on cultured cells as follows: (1) fixation with 10% buffered formalin for 10 minutes at room temperature (RT); (2) three 3-minute rinses with phosphate-buffered saline (PBS); (3) permeabilization with 1% Triton X-100 in PBS for 10 minutes at RT; (4) three 3-minute rinses with PBS; (5) incubation in 10% normal goat serum in PBS (GSP) for 20 minutes at RT; (6) incubation in antibodies against human cleaved caspase 3 (BD-PharMingen, San Diego, CA) in 1.5% GSP overnight at 4°C; (7) three 3-minute rinses with PBS; and (8) incubation with Alexa Fluor 488 secondary antibodies (Molecular Probes, Eugene, OR) in 1.5% GSP for 60 minutes at 37°C. Immunofluorescence was visualized with an inverted microscope (Axiovert S100; Carl Zeiss Meditec, Dublin, CA) equipped with a cooled, digital charge-coupled device (CCD) camera (Roper Scientific, Trenton, NJ). 

In Situ Staining of Activated Caspases

After drug treatments, cells were stained for 1 hour with FITC-VAD-fmk (Promega, Madison, WI), according to the protocol provided by the manufacturer, and then were washed to remove unbound dyes. Images were obtained with a cooled, digital CCD camera (Roper Scientific) with a band-pass filter set: excitation (E _x) and emission (E _m) = 485 and 530 nm, respectively. 

Apoptosis Assay

After drug treatments, cells were treated for 10 minutes with Hoechst 33342 (Molecular Probes) to stain the nuclei. Imaging was performed with a cooled, digital CCD camera (Roper Scientific). Apoptotic cells were quantified by employing the characteristic that apoptotic nuclei have a higher integrated pixel intensity due to chromatin condensation. The counting of nuclei was performed on computer (MetaMorph; Universal Imaging, West Chester, PA). 

Δψ_m Measurement

To monitor mitochondrial membrane potential (Δψ_m), 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolylcarbocyanine iodide (JC-1; Molecular Probes) was used. The green fluorescent JC-1 probe exists as a monomer at low Δψ_m. However, at higher potentials, JC-1 forms red-fluorescent J aggregates. Thus, the emission of this cyanine dye can be used as a sensitive measure of Δψ_m. The ratio of red-to-green JC-1 fluorescence, indicated as R, is dependent only on the membrane potential and not on other factors that may influence single-component fluorescence signals, such as mitochondrial size, shape, and density. Cells were first loaded with JC-1 for 1 hour and then were exposed to different treatments. Images were obtained with a cooled, digital CCD camera (Roper Scientific) with a pair of narrow band-pass filter sets: E _x/E _m = 485/530 nm and 520/610 nm. Alternatively, fluorescence intensity measurements were made with the plate reader, and, during these measurements, the cells were maintained at 37°C in DMEM without phenol red but with 10 mM HEPES added. 

Fluorescence Detection of rOx

Cells were loaded with 50 μM carboxylated dichlorodihydrofluorescein diacetate (carboy-H₂DCFDA; Molecular Probes) for 1 hour at 37°C and then were washed to remove the dye remaining in the solution. After loading, cells were exposed to different treatments, followed by fluorescence intensity measurements made by a plate reader (Bio-Teck, Winooski, VT) using a narrow band-pass filter set: E _x/E _m = 480/530 nm. During the measurements, the cells were maintained at 37°C in DMEM without phenol red but with 10 mM HEPES added. 

Results

Cytotoxic Effect of TNF-α+Act-D in the Presence of TPEN

To reduce the intracellular content of heavy metal ions in RPE, TPEN, a membrane-permeable chelator of heavy metal ions (e.g., Zn²⁺, Cu²⁺) was used. Our preliminary experiments showed that TPEN at ≤2 μM, in our culture conditions, did not elicit any toxicity, whereas a dose-dependent cytotoxicity was observed when the concentration of TPEN was increased to ≥3 μM. 

It has been shown in various cell systems that the cytotoxic effect of TNF-α is expressed only when an inhibitor of transcription or translation or a blocker of NF-κB activation is coapplied. ⁹ To access how heavy metal ion depletion may influence the cytotoxicity of TNF-α in RPE, cultured RPE cells were exposed to TNF-α (10 ng/mL) plus act-D (0.5 μg/mL) in the presence or absence of 1 μM TPEN. As shown in Figure 1 , when exposed to TNF-α+act-D (Figs. 1B 1G) , TNF-α plus TPEN (Figs. 1D 1G) , act-D+TPEN (Figs. 1E 1G) , or TPEN alone (Figs. 1F 1G) , cells displayed ≤5% apoptotic nuclei (i.e., chromatin condensation, nuclear fragmentation) as indicated by Hoechst 33342 staining of the nuclei. In contrast, the number of apoptotic cells was markedly increased in a dose-dependent manner when TNF-α+act-D was coadministrated with TPEN (Figs. 1C 1G 1H) ; coapplication of TNF-α+act-D with 1 μM TPEN increased the level of apoptosis from 5% ± 5% to 74% ± 20% (mean ± SE, n = 3). Note that the culture medium (DMEM) contains millimolar levels of Ca²⁺ and Mg²⁺, and so the enhancing effect of TPEN on the apoptotic response of RPE cells to TNF-α+act-D cannot be ascribed to its low binding affinity for these divalent ions. 

Role of Caspase Activation in TNF-α+act-D+TPEN–Induced Nuclear Alteration

To determine whether the apoptosis induced by TNF-α+act-D+TPEN was caspase dependent, independent, or both ¹⁰ the experiment shown in Figure 1Cwas conducted in the presence of a pan caspase inhibitor, z-VAD-fmk. Cells were treated with 10 μM z-VAD-fmk 1 hour before and during exposure to TNF-α+act-D+TPEN, and the level of apoptosis was examined 18 hours after treatment. As seen in Figure 2 , z-VAD-fmk (A3) effectively suppressed the apoptosis induced by TNF-α+act-D+TPEN (A2) in a dose-dependent manner (Fig. 2B) , in medium containing 3 μM z-VAD-fmk, the level of apoptosis was reduced from 80% ± 20% to 10% ± 10% (mean ± SE; n = 3), indicating that caspase activation is necessary for apoptosis induced by TNF-α+act-D+TPEN. This conclusion was consistent with immunofluorescence analysis. Antibodies specific to the cleaved, but not the intact, form of caspase 3 detected a significant level of fluorescence staining of cleaved caspase 3 in cells exposed to TNF-α+act-D+TPEN (Fig. 3D1) , but not in the control specimens (Fig. 3C1) . 

Effect of CsA, an MPT Blocker, on Cleavage of Procaspase 3 and Subsequent Nuclear Alternation

In other apoptotic cell models in which TNF-α+act-D was used as an inducer, it was shown that the cascade leading to caspase activation is initiated by the release of apoptogenic factors from mitochondria into the cytosol through a mechanism called the MPT. ⁹ Therefore, we examined whether MPT was involved in the apoptosis induced by TNF-α+act-D+TPEN. Cells were pretreated with CsA, an MPT inhibitor, 1 hour before and during the treatment with TNF-α+act-D+TPEN, and apoptosis was examined 18 hours after treatment. CsA markedly reduced apoptosis in a dose-dependent manner (Figs. 3A3 3B) ; with 10 μM CsA, the number of apoptotic cells was reduced from 78% ± 18% to 4% ± 3% (mean ± SE; n = 3). Further, immunofluorescence analysis of cleaved caspase 3 revealed that the number of cells with positive staining of activated caspase 3 was noticeably reduced when 10 μM CsA was included in the medium (Fig. 3D1vs. 3E1 ). Taken together, these data indicate the involvement of MPT in apoptosis induced by TNF-α+act-D+TPEN. 

CsA-Sensitive Reduction of Δψ_m In Vitro

To monitor Δψ_m in living RPE cells, JC-1, a ψ_m-sensitive fluorescent dye, was used. Cells stained with JC-1 (0.3 μg/mL) showed punctuate fluorescent staining (Fig. 4B1) . On expose to carbonyl cyanide m-chlorophenylhydrazone (cccp), a proton ionophore used to reduce the proton gradient across the mitochondrial inner membrane, the red fluorescence intensity was markedly reduced concomitant with a slight increase in green fluorescence intensity (data now shown). These data indicate that JC-1 accumulated in the mitochondria of RPE in a potential-dependent manner. On depolarization, a decrease in the ratio of red-to-green JC-1 fluorescence (R) would be expected at the dose of JC-1 used in this study. To monitor time-dependent changes of Δψ_m in the RPE cells, the cells were first loaded with JC-1 for 1 hour and then exposed to different treatments. Fluorescence intensity measurements were made by a plate reader, and the fluorescence intensity ratio R was normalized (R/R _max) and then plotted versus time (Fig. 4A) . During recordings, a slow change of R/R _max, reflecting a fluctuation of Δψ_m, was observed in all treatments including the control. The cause of this slow change is not clear. However, a time-dependent reduction of R compared with the control was seen in cells exposed to TNF-α+act-D+TPEN, but not in those exposed to TNF-α+act-D. Between 4 and 8 hours, the R/R _max curve derived from cells exposed to TNF-α+act-D+TPEN started to decline, deviating from that of control cells; the R/R _max measured at 14 hours after treatment, from both the control and cells exposed to TNF-α+act-D+TPEN, was 0.75 and 0.30, respectively. The TNF-α+act-D+TPEN–induced reduction of Δψ_m was significantly inhibited when CsA was included in the medium. At 14 hours, the R/R _max measured from cells exposed to TNF-α+ act-D+TPEN with CsA was 0.70, comparable to the control. In summary, cells exposed to TNF-α+act-D+TPEN had a lower Δψ_m and the TNF-α+act-D+TPEN-induced reduction of Δψ_m was inhibited by CsA. 

Image analysis of cells under the same conditions is consistent with this finding. After an 18-hour incubation, cells exposed to TNF-α+act-D+TPEN had a reduced red fluorescence intensity of JC-1 on average (Fig. 4C1)compared with the control (Fig. 4B1) . Note that the TNF-α+act-D+TPEN-induced decrease of red fluorescence of JC-1 was seen in cells with both apoptotic and regular nuclei (Figs. 4C1 4C2) . By contrast, CsA inhibited the TNF-α+act-D+TPEN-induced loss of red fluorescence of JC-1 (Fig. 4D1) . Taken together, these data suggest that the Δψ_m reduction induced by TNF-α+act-D+TPEN is triggered by the MPT. 

A potential complication in using CsA as an MPT inhibitor is that CsA is also a potent inhibitor of calcineurin. CsA binds to and inhibits calcineurin, resulting in a redistribution of Bax/Bcl-x_L, and, thereby, suppressing apoptosis. To examine the possible involvement of calcineurin in the TNF-α+act-D+TPEN–induced apoptosis, we used FK506, a calcineurin inhibitor lacking an effect on the transition pores. As shown in Figure 5 , FK506 neither prevented Δψ_m loss nor suppressed apoptotic nuclear changes induced by TNF-α+act-D+TPEN when compared with CsA. These data suggest that the inhibitory effect of CsA on the apoptosis induced by TNF-α+act-D+TPEN is through its action on the MPT. 

Contribution of Caspase Activation to Δψ_m Loss

To determine the causal relationship between the MPT and caspase activation, the following experiments were conducted using a fluorescein isothiocyanate (FITC) conjugate of the cell-permeable caspase inhibitor VAD-fmk to detect intracellular activated caspases. Our preliminary test (with a different apoptosis model) showed that after exposure for 4 hours to HA14–1, an antagonist of bcl-2 known to induce cyt c release and, as a consequence, apoptosis, ¹¹ cells containing activated caspases could be easily identified with FITC-VAD-fmk used as a probe, and that CsA did not interfere with the binding of FITC-VAD-fmk to its targets, the activated caspases (data not shown). This demonstrated that FITC-VAD-fmk was a useful in situ marker for the detection of activated caspases in RPE. 

We then used this probe to examine intracellular caspase activation in RPE cells under TNF-α+act-D+TPEN exposure, with or without the addition of CsA. Cells exposed to TNF-α+act-D+TPEN showed intensive fluorescence staining with FITC-VAD-fmk (Fig. 6B1)compared with the control cells (Fig. 6A1) . In contrast, staining was abolished when CsA was included in the medium (Fig. 6C1) . These data indicate that the basal level of caspase activation before the MPT in RPE cells is low or undetectable, and that the MPT is necessary for significant caspase activation. 

Because downstream caspases, once activated, may amplify the apoptotic signaling by acting on the mitochondrial membranes, directly or indirectly, through the proapoptotic molecules of the bcl-2 family (e.g., Bid), ¹² we therefore examined whether caspase activation is involved in the TNF-α+act-D+TPEN–induced disruption of Δψ_m. Cells loaded with JC-1 were first treated with 10 μM z-VAD-fmk and then exposed to TNF-α+act-D+TPEN. During the experiments, JC-1 and z-VAD-fmk were continually present in the medium. Fluorescence intensity measurements were made by a plate reader and a typical result is shown in Figure 6D(n = 5). Treatment with z-VAD-fmk effectively prevented the Δψ_m reduction induced by TNF-α+act-D+TPEN. At 16 hours, the R/R _max measured from control cells, from cells exposed to TNF-α+ act-D+TPEN plus CsA, or from those to TNF-α+act-D+TPEN plus z-VAD-fmk were nearly the same (∼0.75), whereas the TNF-α+act-D+TPEN–exposed cells showed a reduction in R/R _max to 0.33. These results suggest that caspases, once activated, may play a role in regulating mitochondrial membrane permeability. ¹² Taken together, our data suggest that apoptotic signaling is initiated by the MPT and then amplified by downstream caspases, probably through a feedback loop. 

Effect of Exposure to TNF-α+act-D+TPEN on Generation of rOx

It has been shown that rOx play a role in the cellular events induced by TNF-α. ^{13 14} We therefore investigated the possible role of rOx in apoptosis induced by TNF-α+act-D+TPEN. We first examined whether rOx was generated after exposure to TNF-α+act-D+TPEN, by using carboxy-H₂DCFDA, a membrane-permeable, fluorescent dye that is sensitive to rOx. Carboxy-H₂DCFDA will not fluoresce until first hydrolyzed by intracellular esterases once inside the cell and then is oxidized by intracellular rOx (carboxy-H₂DCFDA→carboxy-DCF). Cells were first loaded with 50 μM carboxy-H₂DCFDA for 1 hour, washed with normal medium without dyes, and then exposed to different treatments. The fluorescence intensity measurements were made using a plate reader. As shown in Figure 7A , a continuous, slow increase of carboxy-DCF fluorescence appeared in cells exposed to TNF-α+act-D+TPEN, but not in those exposed to TNF-α+act-D. The increasing carboxy-DCF fluorescence detected in cells exposed to TNF-α+act-D+TPEN reflects an increase in rOx, because it was abolished when an antioxidant array (1 mM ascorbic acid, 1 mM Trolox [Hoffman-LaRoche, Nutley, NJ], and 500 μM 4-hydroxy-2,2,6,6-tetramethyl-piperidine-1-oxyl [4-OH-TEMPO]) was included in the medium (Fig. 7A) . Cotreatments with CsA or z-VAD-fmk did not decrease, but rather slightly increased, the carboxy-DCF fluorescence in cells exposed to TNF-α+act-D+TPEN (Fig. 7B) . These data indicate that the generation of rOx after the cells’ exposure to TNF-α+act-D+TPEN occurred before, and did not require, the MPT or caspase activation, suggesting a role for rOx as a proapoptotic signal in triggering apoptosis. In consistent with this suggestion, cotreatment with the antioxidant array significantly inhibited Δψ_m disruption (Fig. 7C)and apoptosis (Fig. 7D)induced by TNF-α+act-D+TPEN. 

Discussion

In summary, our data show that (1) the apoptotic response of RPE cells to TNF-α+act-D is manifestly enhanced when intracellular heavy metal ion content is reduced; (2) the induced apoptotic cell death is caspase-dependent; (3) the caspase activation and subsequent apoptotic events are initiated by the MPT; (4) downstream caspases, once activated, are involved in mitochondrial dysfunction (e.g., Δψ_m loss), suggesting the presence of a mitochondria–caspase feedback loop; and (5) on exposure to TNF-α+act-D+TPEN, rOx is generated, and its production does not require the MPT or caspase activation, suggesting that rOx may be a proapoptotic signal in triggering apoptosis. Taken together, our data suggest that a synergistic interaction between TNF-α and a metal ion deficiency may play a role in triggering apoptosis. 

A consequence of the MPT is the release of cyt c from the mitochondria into the cytosol, where it combines with apaf-1 and procaspase to form an apoptosome, which then triggers caspase activation. ³ Caspases, once activated, may cause further cyt c release through a mitochondrial amplification as a result of extensive cyt c release. ^{15 16 17} In addition, the activated caspases may attack complex I and II of the electron transport chain, causing a further deterioration of mitochondrial electron transport. ¹⁴ Collectively, these apoptotic events, especially cyt c release and a blockade of electron transport by activated caspase, significantly contribute to the Δψ_m loss during the apoptotic process. Therefore, observing a blockade in the reduction of Δψ_m using CsA would strongly suggest that the MPT is the initial step in triggering these apoptotic events. Although the MPT in RPE cells has not yet been demonstrated, our data obtained using CsA suggest that the MPT may be present in RPE cells, and, once induced, may lead to apoptosis. A direct study of the MPT using isolated mitochondria prepared from RPE cells will help to clarify this question. 

The involvement of a mitochondria-caspase amplification feedback loop in the apoptotic process has been demonstrated in many cellular systems. ^{15 16 17} It is accepted that a positive feedback loop ensures synchronization of the signals derived from stochastic, local events, eventually leading to a global effect such as apoptosis. Our data suggest that the initial apoptotic event, the MPT, may occur in a small population of mitochondria, indicating that the MPT induction threshold varies among mitochondria in RPE cells. This heterogeneity among mitochondria has been shown in other cell systems. ^{18 19} It is not yet clear what mechanisms lead to this heterogeneity; however, Δψ_m may contribute. It is known that Δψ_m is the driving force for mitochondrial Ca²⁺ uptake, ²⁰ and mitochondrial [Ca²⁺] and Δψ_m play important roles in determining the probability of opening of the transition pores. ³ Using JC-1 as a probe, we observed a wide variability in Δψ_m among mitochondria in every RPE cell inspected (Yang JH, unpublished results, 2003). The question of how this is related to the overall apoptotic process warrants further investigation. 

The important role of heavy metal ions in numerous physiological functions has been known for many years. ^{21 22} Moreover, it has been shown in many cell systems that exogenous heavy metal ions can protect cells from apoptosis. By contrast, chronic exposure to the divalent heavy metals has been linked to the development of cellular dysfunctions and diseases. ²³ However, our understanding of the role these ions play in apoptotic process is very limited. In a cell-free model, Zn²⁺ at micromolar levels was shown to suppress caspase 3 activation. ²⁴ In intact cells, Zn²⁺ is shown to have multiple roles in apoptosis: (1) It can act like an antioxidant ^{25 26} ; (2) once taken up by the mitochondria, it induces the Δψ_m loss and the production of rOx ²⁷ ; and (3) it inhibits ethanol-induced liver apoptosis through its interference with Fas ligand activation. ²⁸ In addition, elevation of intracellular Fe²⁺ protects leukemia cells or tumor cells from NO-mediated Fe²⁺ depletion and subsequent apoptosis. ^{29 30} In pheochromocytoma cells, intracellular Zn²⁺ and Cu²⁺ are involved in NF-κB activation, thereby determining cell viability. ³¹ A recent study showed that Cd²⁺ directly induces the MPT in isolated mitochondria, and the mechanism of the MPT induced by Cd²⁺ is distant from that of the Ca²⁺-induced MPT. ³¹ Given that the gating properties of the MPT, and its sensitivity to pharmacological agents, are varied among cell types, ^{33 34 35} it would not be surprising if various heavy metals were shown to have differential effects on induction of the MPT or apoptotic signaling. 

The involvement of rOx in many diseases has been known for many years. Studies using a cellular model for AMD (i.e., the apoptosis induced by A2E) concluded that that the generation of singlet oxygen may be involved in the mechanism leading to the death of A2E-containing RPE cells after blue light illumination. ^{36 37} The identity of the rOx generated by exposure to TNF-α+act-D+TPEN was not examined in this study. However, the incomplete block of apoptosis by antioxidants, as shown in Figure 7 , may suggest that rOx generated in the RPE during exposure to TNF-α+act-D+TPEN is beyond the capacity of added antioxidants to act as rOx scavengers or that proapoptotic signaling may involved factor(s) other than rOx. 

Both heavy metal ion deficiency and apoptosis have been implicated in many age-related diseases. ^{21 22 38 39 40 41 42} However, the exact role played by these factors and their interrelationships in the initiation of these age-related disorders is not fully understood. Our study indicates that a reduction in the intracellular level of heavy metal ions may increase the risk of apoptosis in RPE cells due to cytokine(s) attack. In this study, TNF-α alone was used to induce apoptosis, whereas, in vivo, a more complicated pathway involving a network of cytokines and other components would be expected. ³⁸ A full understanding of the mechanism underlying apoptosis may help to design an effective antiapoptosis strategy, and, perhaps, effective treatments for ocular disorder caused by apoptosis in RPE due to inflammation. 

 Supported in part by grants from Research to Prevent Blindness, the International Retina Research Foundation, Lions Eye Bank Foundation of Houston, Retina Research Foundation of Houston, Core Grant P30EY02520 and Grant EY04446 from the National Eye Institute.

 Submitted for publication March 22, 2004; revised August 12, 2004; accepted October 4, 2004.

 Disclosure: J.-H. Yang, None; W.-D. Le, None; S.F. Basinger, None; S.M. Wu, None; C.-y. Yang, 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.

 

Cytotoxicity of TNF-α under conditions of heavy metal ion deficiency. (A–F) Fluorescence images of RPE cells stained with Hoechst 33342, showing nuclear morphology of control cells (A), cells treated with TNFα+act-D (B), TNF-α+act-D in the presence of TPEN (C), TNF-α+TPEN (D), act-D+TPEN (E), and TPEN alone (F). (G, H) A quantitative analysis of apoptosis. Cells were exposed for 18 hours to the treatments indicated, and the percentage of apoptosis was determined after nuclear staining with Hoechst 33342 (G). (H) RPE cells were treated for 18 hours with TNF-α+act-D in the presence of increasing concentrations of TPEN, and the percentage of apoptosis was determined as in (G).

Figure 2.

 

Caspase inhibitor blocked TNF-α+act-D+TPEN-induced apoptosis in cultured RPE cells. (A1–A3) Fluorescence micrographs of cells stained with Hoechst 33342 showing nuclear morphology of control cells (A1) and of cells treated with TNF-α+act-D+TPEN (A2) or with TNF-α+act-D+TPEN in the presence of 10 μM z-VAD-fmk (A3). (B) Cells first treated for 1 hour with the indicated concentrations of z-VAD-fmk, then exposed to TNF-α+act-D+TPEN for 18 hours, with z-VAD-fmk continually present. The percentage of apoptosis was determined as in Fig. 1G .

Figure 3.

 

CsA, an MPT blocker, inhibited TNF-α+act-D+TPEN–induced apoptosis and caspase 3 activation in RPE cells. (A1–A3) Fluorescence micrographs of cells stained with Hoechst 33342 showing nuclear morphology of control cells (A1) and of cells treated with TNF-α+act-D+TPEN (A2) or with TNF-α+act-D+TPEN in the presence of 10 μM CsA (A3). (B) Cells treated with the indicated concentrations of CsA for 1 hour exposed to TNF-α+act-D+TPEN for 18 hours with CsA continually present. The percentage of apoptosis was determined as in Figure 1F . (C1, D1, E1) Fluorescence micrographs showing immunofluorescence staining of cleaved caspase 3 in control cells (C1) and in cells treated with TNF-α+act-D+TPEN (D1) or TNF-α+act-D+TPEN in the presence of 10 μM CsA (E1). (C2, D2, E2) Corresponding fluorescence images of (C1), (D1), and (E1), respectively, showing nuclear morphology after staining with Hoechst 33342.

Figure 4.

 

TNF-α+act-D+TPEN induced loss of Δψ_m. (A) Cells loaded with JC-1 were exposed to TNF-α+act-D (t+a) in the absence or presence of TPEN (t+a+tp) with or without CsA. The fluorescence intensity ratios (R: fluorescence intensity of J-aggregate over that diffuse monomer of JC-1) were measured and normalized to their maximum value (R/R_max). (B1, C1, D1) Fluorescence micrographs showing J-aggregate staining in control cells (B1), cells treated with TNF-α+act-D+TPEN (C1), and TNF-α+act-D+TPEN in the presence of 10 μM CsA (D1). (B2, C2, D2) Corresponding fluorescence images of (B1), (C1), and (D1), respectively, showing nuclear morphology after staining with Hoechst 33342.

Figure 5.

 

FK506 had no effect on Δψ_m loss or apoptosis induced by TNF-α+act-D+TPEN. (A) Cells loaded with JC-1 exposed to TNF-α+act-D+TPEN in the absence or presence of CsA (t+a+tp+CsA) or FK506 (t+a+tp+FK506). Fluorescence intensity ratios were determined as in Figure 4A . (B) Cells were exposed for 18 hours to the indicated treatments and the percentage of apoptosis determined as in Figure 1G .

Figure 6.

 

Caspase activation was downstream to the MPT and contributed to loss of Δψ_m. (A1, B1, C1) Fluorescence micrographs of cells stained with FITC-VAD-fmk to localize the activated caspase in untreated cells (A1), and cells exposed to TNF-α+act-D+TPEN in the absence (B1) or presence (C1) of CsA. (A2, B2, C2) Corresponding fluorescence images of (A1), (B1), and (C1), respectively, showing nuclear morphology after staining with Hoechst 33342. (D) Cells loaded with JC-1 exposed to TNF-α+act-D+TPEN in the absence or presence of CsA (t+a+tp+CsA) or z-VAD-fmk (t+a+tp+z-VAD-fmk). The fluorescence intensity ratios were determined as in Figure 4A .

Figure 7.

 

rOx involvement in apoptosis induced by TNF-α+act-D+TPEN. (A, B) Cells loaded with carboxy-H₂DCFDA exposed to TNF-α+act-D or TNF-α+act-D+TPEN (t+a+tp) in the absence or presence of antioxidants (A), or to TNF-α+act-D+TPEN in the absence or presence of CsA or z-VAD, respectively (B), followed by fluorescence intensity measurements. F/F₀ represents the normalization of fluorescence intensity measured at the times indicated (F) with that measured at zero time (F₀). (C) Cells loaded with JC-1 exposed to TNF-α+act-D+TPEN in the absence or presence of antioxidants or CsA. The fluorescence intensity ratios were determined as in Figure 4A . (D) Cells exposed for 18 hours to the treatments indicated, and the percentage of apoptosis determined as in Figure 1G .