Lead and Calcium Produce Rod Photoreceptor Cell Apoptosis by Opening the Mitochondrial Permeability Transition Pore*

, Calcium overload is suggested to play a fundamental role in the process of rod apoptosis in chemical-induced and inherited retinal degenerations. However, this hy-pothesis has not been tested directly. We developed an in vitro model utilizing isolated rat retinas to determine the mechanisms underlying Ca 2 1 - and/or Pb 2 1 -induced retinal degeneration. Confocal microscopy, histological, and biochemical studies established that the elevated [Ca 2 1 ] and/or [Pb 2 1 ] were localized to photoreceptors and produced rod-selective apoptosis. Ca 2 1 and/or Pb 2 1 induced mitochondrial depolarization, swelling, and cytochrome c release. Subsequently caspase-9 and caspase-3 were sequentially activated. Caspase-7 and caspase-8 were not activated. The effects of Ca 2 1 and Pb 2 1 were additive and blocked completely by the mitochondrial permeability transition pore (PTP) inhibitor cyclosporin A, whereas the calcineurin inhibitor FK506 had no effect. The caspase inhibitors carbobenzoxy-Leu-Glu-His-Asp-CH 2 F and carbobenzoxy-Asp-Glu-Val-Asp- CH 2 F, but not carbobenzoxy-Ile-Glu-Thr-Asp-CH 2 F, dif- ferentially blocked post-mitochondrial events. The levels of reduced and oxidized glutathione and pyridine nucleotides in rods were unchanged. Our results demonstrate that rod mitochondria are the target site for Ca 2 lead-induced rod-selective apoptosis and to elucidate the cellular and biochemical mechanisms underlying the rod-selective apoptosis. Our results show that exposure of isolated adult rat to Ca 2 1 and/or Pb 2 1 resulted in rod-selective apoptosis and suggest that Ca the internal Me 2 1 binding of the mitochondrial thereby opening the PTP and initiating the cytochrome c cascade of in rods. show that approximately twice (1.7–2.2x) the amount of processed or cleaved fragments were detected in retinas incubated with Ca 2 1 and Pb 2 1 ( lane 4 ) compared with either Ca 2 1 ( lane 3 ) or Pb 2 1 alone. In E and F , retinas were incubated in buffers contain- ing Ca 2 1 and Pb 2 1 in the absence ( lane 1 ) or presence of the caspase-3 inhibitor ( lane 2, 1 n M DEVD-fmk) or the mitochondrial PTP inhibitor ( lane 3, 10 m M CsA) and procaspase-9 and procaspase-3 processing were assessed. Immunoblots show that approximately the same the amount of p35 fragment was present in the absence ( E, lane 1 ) or presence ( E, lane 2 ) of DEVD-fmk. These blots are representative of three to four experiments for each treatment condition. c ) is released from the mitochondria to cytosol where it interacts with apoptotic protease activating factor-1 ( Apaf-1 ), dATP/dADP, and procaspase-9, and where caspase-9 is pre- sumably activated. Caspase-9 activates procaspase-3. Caspase-3 subsequently activates endonucleases that cleave DNA into HMW fragments and this results in the occurrence of morphological rod apoptosis.

Apoptosis is an active mode of cell death that is induced by a variety of physiological and pathological stimuli. Convergent evidence suggests that mitochondria and caspases play a central and fundamental role in the effector or executioner phase of apoptosis (1). Early during the effector phase, the mitochondrial permeability transition pore (PTP), 1 a megachannel in the inner mitochondrial membrane, is opened by a variety of apoptotic inducers such as elevated matrix Ca 2ϩ , pro-oxidants, and thiol-reactive agents (1)(2)(3). This leads to mitochondrial depolarization (decrease in ⌬⌿ m ) and subsequently to the release of cytochrome c and/or apoptosis-inducing factor from mitochondria to cytosol or nuclei (1,4,5). Caspases are activated, which cleave downstream death substrates and activate endonucleases that cleave genomic DNA into fragments resulting in the apoptotic nuclear morphology (6,7). The opening of the mitochondrial PTP and the apoptotic process can be inhibited by a diverse group of agents such as Bcl-2, Bcl-x L , bongkrekic acid, and CsA (1).
Sustained increases in intracellular [Ca 2ϩ ] trigger apoptosis in a diverse array of in vivo and in vitro systems (1,8). Results from several studies suggest that elevated rod photoreceptor [Ca 2ϩ ] plays a key role in the process of apoptotic rod cell death in humans and animals during inherited retinal degenerations, retinal diseases and injuries, and chemical exposure. These include patients with retinitis pigmentosa and cancer-associated retinopathy (9,10), lead-exposed rats (11)(12)(13), retinal degeneration mice (13,14), rats injected with anti-recoverin monoclonal antibodies (15), rats with hypoxic-ischemic injury (16), and rats with light induced damage (17).
Several neurotoxic heavy metals, transition metals, and organometals produce neuronal apoptosis in whole animals and cultured cells (11, 12, 18 -25). For example, low to moderate level Pb 2ϩ exposure produces apoptotic rod and bipolar cell death in developing and adult rats (11,12) and apoptotic neuronal cell death in primary cultured cells (23,24). Similarly, Cu 2ϩ and methyl mercury produce apoptosis in the developing and mature olfactory epithelium and cerebellum, respectively (18,22). Although the molecular mechanisms underlying the apoptosis induced by these different metals are unknown, there are two possible, though not mutually exclusive, triggering mechanisms: one is Ca 2ϩ overload and the other is the generation of ROS. Both in vivo and in vitro lead exposure increase rod and retinal intracellular [Ca 2ϩ ] during cell death (26,27). Hg 2ϩ , tributyltin, and methyl mercury also cause Ca 2ϩ overload in cultured neuronal cells prior to cell death (19,20,28). Moreover, Hg 2ϩ , Mn 2ϩ , Cu 2ϩ , and methyl mercury can induce apoptosis in neurons and other tissues by increasing ROS (21, 29 -31).
As noted above, most studies of neuronal apoptosis are conducted using cell cultures or cell-free systems. The goals of this study were to establish an in vitro retinal model of in vivo lead-induced rod-selective apoptosis and to elucidate the cellular and biochemical mechanisms underlying the rod-selective apoptosis. Our results show that exposure of isolated adult rat retinas to Ca 2ϩ and/or Pb 2ϩ resulted in rod-selective apoptosis and suggest that Ca 2ϩ and Pb 2ϩ bind to the internal Me 2ϩ binding site of the mitochondrial PTP, thereby opening the PTP and initiating the cytochrome c-caspase cascade of apoptosis in rods.
DMQD-CHO was a gift from Dr. Atsushi Takahashi (Kyoto University, Kyoto, Japan). Anti-caspase-3 and anti-caspase-9 antibodies were gifts from Dr. Donald W. Nicholson (Merck Frosst Center for Therapeutic Research, Quebec, Canada). The anti-caspase-7 antibody was a gift from Dr. Gerald M. Cohen (MRC Toxicology Unit, Leicester, UK). FK506 was a gift from Fujisawa Pharmaceuticals (Tokyo, Japan).
Experimental Animals and Retinal Incubations-All experimental and animal care procedures were in compliance with the principles of the American Physiological Society and the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Adult (90 -120 days old) female Long-Evans hooded light-adapted rats (Harlan Sprague-Dawley, Indianapolis, IN) were used in all experiments. The in vitro neural retina incubation procedure was modified slightly from Medrano and Fox (27). Briefly, for control experiments two neural retinas were incubated for 15 min at 37°C in 10 ml of Tris buffer (30 mM Tris-HCl, 120 mM NaCl, 5 mM KCl, 3 mM MgCl 2 , 10 mM glucose, pH 7.4, 310 milliosmolar) with or without 0.5 mM IBMX. IBMX was used to increase the number of open cGMP-gated channels in rods so that increased amounts of Ca 2ϩ and Pb 2ϩ could enter these cells (27,32,33). There were no differences, on any measure, between retinas incubated with or without IBMX, so the data were combined and are presented as control buffer. For experiments assessing the effects of Ca 2ϩ and/or Pb 2ϩ , two neural retinas were incubated in a Tris-IBMX buffer for 15 min with or without 0.1-2.0 mM free Ca 2ϩ , 10 nM to 10 M free Pb 2ϩ , or 0.5 mM free Ca 2ϩ plus 1 M free Pb 2ϩ . For time course studies, retinas were incubated for 5, 10, or 15 min in control buffers, or in IBMX plus 0.5 mM Ca 2ϩ and 1 M Pb 2ϩ . Because IBMX was used in all experiments with Ca 2ϩ and/or Pb 2ϩ , it will not be noted further in the text. As described (27), the Calcon computer program (version 9.4) was used to calculate the amounts of CaCl 2 and PbCl 2 required to produce the desired concentrations of free [Ca 2ϩ ] and [Pb 2ϩ ] for each condition, and the concentrations were verified with ion-selective electrodes.
Detection of HMW DNA Fragmentation-To detect cleavage of genomic DNA into HMW fragments, single cell retinal suspensions were made from each pair of retinas (n ϭ 5-7 pairs of retinas/condition) by incubating retinas in 500 l of activated papain (26 units/ml) at 37°C for 5 min followed by trituration. Field inversion gel electrophoresis was carried out as described (38). After field inversion gel electro-phoresis and staining with EtBr, DNA fragments of 50, 300, and 600 kbp were visualized by UV fluorescence.
Detection and Quantification of Apoptosis-To quantify the number of apoptotic and necrotic cells and to determine the location and type(s) of retinal cells undergoing apoptosis, retinas were incubated in Ca 2ϩ and/or Pb 2ϩ buffers (n ϭ 4 -8 retinas/condition) and stained with AO/ EtBr essentially as described (39). Retinas were stained with AO/EtBr (20 g/ml), whole mounted on slides, and viewed on an Olympus IX-70 inverted microscope equipped for epifluorescence. All layers of the central and mid-peripheral retinas were scanned. and stained nuclei were counted in ten visual fields in four different retinal areas using stereological procedures (40). The techniques for tissue preparation, fixation, and electron microscopy studies were as described (11,12). Values represent mean Ϯ S.E. of the combined number of early and late apoptotic nuclei/mm 2  ), flame atomic absorption spectrometry (n ϭ 7-8 retinas/ condition) was used as described (27). Fluo-3 Ca 2ϩ imaging and confocal laser scanning microscopy were used as described (41) to localize the distribution and to determine the relative concentrations of free Ca 2ϩ and Pb 2ϩ in retinas (n ϭ 3 retinas/condition). Retinas (180 -200-m thick) were mounted photoreceptor side down and scanned every 2 m. Ninety to 100 images were aligned and stacked using the three-dimensional software program in NIH Image (version 1.62).
Determination of Mitochondrial Membrane Potential and Swelling-The techniques for utilizing JC-1 were essentially as described (42). Briefly, retinas (n ϭ 4 -8/condition) were stained with JC-1 (10 g/ml), whole mounted on slides, and examined with the fluorescence microscope. The red fluorescence (585 nm) from the J-aggregate (hyperpolarized mitochondria) and the green fluorescence (520 nm) from the JC-1 monomer (depolarized mitochondria) were viewed with a rhodamine optical filter set and a fluorescein optical filter set, respectively. Pairs of digitized and processed images of the same photoreceptor cells from intact retinas showing red and green fluorescence were captured, and the intensities of fluorescence were quantified using NIH Image Software.
Mitochondrial swelling (n ϭ 3-4 experiments/condition) was estimated from the decrease of light scattering at 540 nm. Retinal mitochondria were isolated as described (27)  Enzymatic Assays for Caspase Activity-Caspase activity was measured using three different tetrapeptide-pNA substrates (DEVD-pNA, DMQD-pNA, and LEHD-pNA) as described (43). The incubated retinas were swelled for 10 min on ice in cell lysis buffer, homogenized with a Polytron PT1200, and centrifuged (10,000 ϫ g, 20 min, 4°C). 50 l of supernatant (15-20 mg/ml protein) was incubated with 50 l of reaction solution (150 mM NaCl, 0.1% CHAPS, 15 mM dithiothreitol) and 2 l of tetrapeptide-pNA (100 M) at 37°C for 30 -120 min. The reaction was linear for its duration and was stopped at 60 min by adding 400 l of water. Absorbance was measured at 405 nm. For negative controls, each supernatant was incubated with reaction buffer without tetrapeptide-pNA. For positive controls, each supernatant was preincubated with 2 M of the respective caspase inhibitor before adding substrate.
The arbitrary values are presented as mean Ϯ S.E. for three to seven experiments/condition/substrate.
Measures of Redox Status and Oxidative Stress in Rods-Following the incubations, 4 -6 retinas were pooled and rod outer or rod outer and inner segments were isolated as described (45,46) and frozen in liquid nitrogen. The concentrations of reduced and oxidized glutathione (GSH and GSSG) and pyridine nucleotides (NADH 2 and NAD ϩ ) were measured in rod outer and inner segments as described (47,48). The rod outer segment fatty acid concentration, phospholipid yield, and conjugated diene content were measured as described (45). The latter are formed during the oxidation of polyunsaturated fatty acids and are an accepted measure of lipid peroxidation similar to malonaldehyde (45). Values represent the mean Ϯ S.E. from three to four experiments/measurement/condition.
Statistical Analysis-All values are presented as the mean Ϯ S.E. All data analyses were performed on untransformed data and analyzed using the appropriate analysis of variance (ANOVA) and Fisher's protected least significant difference posthoc comparisons (StatView, Abacus Concepts, Inc., Berkeley, CA). Differences from controls were regarded as significant if p Ͻ 0.05.

RESULTS
Retinal DNA Is Cleaved into HMW Fragments-Cleavage of DNA into HMW fragments is considered one of the hallmarks of apoptosis (38). Retinas incubated for 15 min in control buffer produced a minimal amount of 600-and 300-and no 50-kbp DNA fragments (Fig. 1, lane 1). Retinas incubated with 0.1-2 mM Ca 2ϩ or 10 nM to 10 M Pb 2ϩ exhibited a concentration-dependent increase in the amount of HMW fragments (data not shown). The concentrations of Ca 2ϩ and Pb 2ϩ that produced an apparent 50% increase in the amount of HMW fragments (300 Ͼ 600 Ͼ Ͼ 50 kbp) were 0.5 mM Ca 2ϩ and 1 M Pb 2ϩ (Fig.  1, lanes 2 and 3). These concentrations were employed for all other experiments. The amount of HMW fragments from retinas incubated with Ca 2ϩ and Pb 2ϩ was greater than with either divalent cation alone, and the pattern of fragmentation was different (300 Ͼ 50 Ͼ 600 kbp; Fig. 1, lane 4) suggesting an additive effect of these divalent cations. Moreover, the Ca 2ϩand Pb 2ϩ -induced changes were not the result of nonspecific divalent cation effects as the pattern of HMW DNA fragmentation for retinas incubated in buffers containing Zn 2ϩ (1-3 mM) or no Mg 2ϩ was not different than controls shown in Fig.  1, lane 1 (data not shown). The production of HMW DNA fragments by Ca 2ϩ and Pb 2ϩ was inhibited completely by 1 nM DEVD-fmk (Fig. 1, lane 5), 1 M CsA (Fig. 1, lane 6), or 10 M CsA (data not shown). In contrast, it was not blocked by 100 M YVAD-cmk or 100 nM FK506 (data not shown), indicating that group I caspases and calcineurin, respectively, were not involved under these conditions.
Retinal Apoptosis Induced by Ca 2ϩ and/or Pb 2ϩ Is Rod Photoreceptor-specific-AO/EtBr staining enabled us to distinguish between viable, early apoptotic, late apoptotic, and necrotic cells (39). All early and late apoptotic nuclei were identified as rod photoreceptor cells as determined by their localization within the outer nuclear layer and nuclear diameter. Electron microscopy confirmed that the Ca 2ϩ -and/or Pb 2ϩ -induced retinal cell death was rod-specific and apoptotic. Fig. 2A shows a typical apoptotic nucleus with fragmented and marginated chromatin, whereas the surrounding rod nuclei appear normal. Retinas incubated in control buffer had ϳ25 apoptotic early and late rod cells/mm 2 of retina (Fig. 2B). The number of apoptotic rods in retinas incubated with Ca 2ϩ or Pb 2ϩ in-creased 3-4-fold, whereas retinas incubated with Ca 2ϩ and Pb 2ϩ showed an 8-fold increase in apoptotic rods indicating an additive effect of both divalent cations similar to that observed with the HMW fragmentation. The Ca 2ϩ -and Pb 2ϩ -induced rod apoptosis was inhibited completely by 1 nM DEVD-fmk or 10 M CsA, but not by 100 nM FK506 (Fig. 2B). Moreover, AO/EtBr staining revealed that DEVD-fmk and CsA were cytoprotective, as the rod nuclei from these treated retinas were uniformly green indicating that these cells were viable and did not die instead by necrosis. The number of necrotic cells did not vary with any incubation condition: 1.9 Ϯ 0.8 necrotic cells/ mm 2 of retina. Taken together, these data show that Ca 2ϩ and Pb 2ϩ produce rod-selective apoptosis and that caspases are required for both nuclear apoptosis and cell death. Pb 2ϩ produces a concentration-dependent decrease in the Ca 2ϩ -induced fluo-3 fluorescence (50). Therefore, we reasoned that the intracellular distribution of Pb 2ϩ might be localized by its ability to quench the Ca 2ϩ -fluo-3 fluorescence. In controls, the fluo-3 signal was observed mainly in the photoreceptors (Fig. 3A). In retinas incubated with Ca 2ϩ , the fluo-3 fluorescence was localized almost exclusively in the photoreceptors and was markedly increased as evidenced by saturation of the high affinity dye in several rods (Fig. 3B). In similar preparations, the free [Ca 2ϩ ] in rods ranged from 5 to 10 M (32). In retinas incubated with Ca 2ϩ and Pb 2ϩ (Fig. 3C) or Ca 2ϩ followed by Pb 2ϩ (data not shown), the Ca 2ϩ -induced fluo-3 fluorescence was quenched to a level slightly below that observed in controls (Fig. 3A). These results indicate that Pb 2ϩ readily entered the retina, was localized to the photoreceptors, and did not block the entry of Ca 2ϩ . The addition of 1 nM DEVD-fmk to the retinal incubation buffers did not affect the Ca 2ϩ -induced fluo-3 fluorescence (data not shown) indicating that caspase-3 did not influence divalent cation entry into the photoreceptors.
Overall, the fluo-3 fluorescence was lower in the inner than outer retina and there were no treatment-related changes in the inner retinal [Ca 2ϩ ]. Retinas incubated with Pb 2ϩ alone had slightly lower levels of fluo-3 fluorescence in the photoreceptors than in controls (data not shown).
Mitochondrial ⌬⌿ Is Decreased in Retinas Incubated in Ca 2ϩ and/or Pb 2ϩ -Accumulating evidence suggests that mitochondria play an important role in regulating apoptosis (1). The rod ⌬⌿ m was monitored using JC-1 staining and fluorescence microscopy. As illustrated by a representative figure, control rod mitochondria throughout the retina displayed an intense fluorescence from J-aggregates (Fig. 4A) and very weak fluorescence from JC-1 monomers (Fig. 4F). Overall, the mitochondria from retinas incubated with Ca 2ϩ and/or Pb 2ϩ were depolarized as indicated by the decreased J-aggregate fluorescence and increased JC-1 monomer fluorescence. Representative images of rods from retinas incubated with Ca 2ϩ (Fig. 4, B and G) or Pb 2ϩ (Fig. 4, C and H) reveal markedly decreased J-aggregate formation (Fig. 4, B and C) and increased JC-1 monomers (Fig.  4, G and H). Closer examination of these figures reveals that a subpopulation of rod mitochondria was depolarized. Moreover, in retinas incubated in Ca 2ϩ and Pb 2ϩ the rod mitochondria were depolarized even further as shown by the large decrease in J-aggregate (Fig. 4D) and increase in JC-1 monomer (Fig.  4I). An electron micrograph of a transverse section through normal rod inner segments reveals that the circular pattern of mitochondria in a rod (Fig. 4K) is similar to the ring-shaped J-aggregate fluorescence observed in control retinas (Fig. 4A).
To determine the contribution of the mitochondrial PTP to the mitochondrial depolarization, retinas were incubated with 10 M CsA plus Ca 2ϩ and Pb 2ϩ . A representative pair of figures reveals that CsA markedly increased the rod ⌬⌿ m as evidenced by an increase in J-aggregate (Fig. 4, E versus D) and a decrease in JC-1 monomer (Fig. 4, J versus I) indicating that the Ca 2ϩ -and Pb 2ϩ -induced opening of PTP significantly contributed to the rod mitochondrial depolarization. Similar data were obtained with CsA in the presence of Ca 2ϩ or Pb 2ϩ alone (data not shown). Interestingly, many rod mitochondria were still depolarized in the presence of CsA (compare Fig. 4, J to F). This latter result is consistent with our findings that Ca 2ϩ and Pb 2ϩ FIG. 2. AO/EtBr staining and electron microscopy reveal that Ca 2؉ and/or Pb 2؉ produce a DEVD-fmk and CsA inhibitable rod-selective apoptosis. Retinas were incubated as described in the legend to Fig. 1. In A, a representative electron micrograph illustrates a typical apoptotic rod nucleus (arrowhead) found in the outer nuclear layer of retinas incubated in Ca 2ϩ and/or Pb 2ϩ buffers. Note the condensed and fragmented chromatin that is typical of apoptotic cells and that the surrounding rod nuclei appear normal. Original magnification, 7000ϫ. Bar indicates 2 m. In B, a quantitative analysis of the number of combined early and late apoptotic rod cells for each treatment condition is presented as the mean Ϯ S.E. of the number of apoptotic rod cells/mm 2 of retina for four to eight retinas for each treatment condition. Retinas were incubated in control buffers or in buffers containing Ca 2ϩ and/or Pb 2ϩ or Ca 2ϩ and Pb 2ϩ plus either 1 nM DEVD-fmk, 10 M CsA, or 100 nM FK506. All values with asterisks are significantly different from controls at: *, p Ͻ 0.05 and **, p Ͻ 0.02. FIG. 3. Fluo-3 Ca 2؉ imaging and confocal laser scanning microscopy show that Ca 2؉ and Pb 2؉ concentrations are increased in photoreceptor cells. Retinas were incubated as described in the legend to Fig. 1. As described under "Experimental Procedures," the retinas were whole mounted photoreceptor side down and examined with the confocal laser scanning microscopy following fluo-3 loading. The images were processed using NIH Image Software. The arrowheads mark the outer retinal (photoreceptor) region. In A, the fluo-3 signal in control retinas was localized mainly to the photoreceptors. In B, retinas incubated with Ca 2ϩ had a marked increase in fluo-3 signal localized almost exclusively to the photoreceptors. In C, the fluo-3 signal in retinas incubated with Ca 2ϩ and Pb 2ϩ was quenched at or below control levels indicating that both Ca 2ϩ and Pb 2ϩ were localized in the outer retina. These images are representative of three retinas for each treatment condition.
produce a concentration-dependent transient stimulation of state 4 (ADP-independent) respiration in isolated retinal mitochondria that is inhibited by ruthenium red but not by CsA (Ref. 27 and data not shown). Alternatively, it is possible that Ca 2ϩ and Pb 2ϩ produce an initial mitochondrial depolarization that contributes to opening the mitochondrial PTP. This is consistent with the findings of Bernardi (3) on Ca 2ϩ -induced opening of the PTP in liver mitochondrial populations. The temporal and spatial resolution of our techniques does not allow us to differentiate between these two alternatives. FK506 (100 nM) had no effect on the Ca 2ϩ -and Pb 2ϩ -induced depolarization of rod mitochondria (data not shown), indicating that calcineurin is not involved in the opening of the PTP. To determine the contribution of the caspase activation to the mitochondrial depolarization, retinas were incubated with 1-10 nM DEVD-fmk plus Ca 2ϩ and Pb 2ϩ . DEVD-fmk had no effect on the Ca 2ϩ -and/or Pb 2ϩ -induced mitochondrial depolarization (data not shown). Finally, verapamil (100 M) had no effect on the pattern of JC-1 fluorescence in mitochondria (data not shown), indicating that the effects of CsA were on the PTP and not on the multidrug resistance pump (51).
Rod Redox Status Is Not Altered in Retinas Incubated in Ca 2ϩ and/or Pb 2ϩ -Decreases in GSH and NADH levels and the production of ROS occur in many forms of apoptosis, are triggered by Ca 2ϩ overload, and can initiate rod photoreceptor degeneration (1,3,45,52). In addition, the probability of mitochondrial PTP opening is increased by the oxidation of glutathione and pyridine nucleotides (3). Therefore the concentrations of rod GSH, GSSG, NADH 2 , and NAD ϩ and the peroxidation status of rod lipids were determined in retinas incubated with Ca 2ϩ and/or Pb 2ϩ . In rods isolated from retinas incubated in control buffer, the mean Ϯ S.E. concentrations of GSH and GSSG were 31.8 Ϯ 1.8 and 4.1 Ϯ 0.4 nmol/mg protein, respectively, and for NADH 2 and NAD ϩ were 0.61 Ϯ 0.06 and 3.44 Ϯ 0.29 nmol/mg protein, respectively. The high NAD ϩ / NADH 2 ratio is similar to that found in rat, rabbit, and monkey retinas (53,54) and most likely results from the very high rate of retinal aerobic glycolysis (55). The mean Ϯ S.E. mol% of the major fatty acids in rods was 9. The values were not significantly different in nonincubated control rods indicating that incubation per se did not change rod redox or lipid peroxidation status. Moreover, the values for each of the above rod measures were not significantly different in retinas incubated in Ca 2ϩ and/or Pb 2ϩ compared with controls (96 Ϯ 5% of control). These results demonstrate that Ca 2ϩ and/or Pb 2ϩ did not produce oxidative stress.
The Ca 2ϩ -and/or Pb 2ϩ -induced Cytochrome c Release and Mitochondrial Swelling Are Inhibited by CsA but Not by DEVDfmk-Following a diverse array of apoptotic stimuli, cytochrome c is released from mitochondria to the cytosol where it participates in activating the caspase cascade leading to cell death (1,4). Because cytochrome c release is not obligatory for apoptosis (56,57), its role in Ca 2ϩ -and/or Pb 2ϩ -induced apoptosis was examined. Retinas incubated in control buffers had no detectable cytochrome c in the cytosolic fraction (Fig. 5, lane  1), whereas approximately equal amounts of cytochrome c were detected in the cytosolic fractions of retinas incubated in Ca 2ϩ (Fig. 5, lane 2) or Pb 2ϩ (Fig. 5, lane 3). Retinas incubated with both divalent cations released twice the amount of cytochrome c (Fig. 5, lane 4) compared with retinas incubated with either divalent cation. The maximum amount of cytochrome c released was 15-20% suggesting that Ca 2ϩ and/or Pb 2ϩ released cytochrome c from the large, not small, pool that is localized to the intercristal compartment (51). CsA completely blocked the Ca 2ϩ -and Pb 2ϩ -induced release of cytochrome c (Fig. 5, lane 6), whereas 100 nM FK506 (data not shown) or DEVD-fmk (Fig. 5, lane 5) did not. Interestingly, CsA did not completely block mitochondrial depolarization (Fig. 4) or the Ca 2ϩ -or Pb 2ϩinduced transient increase in state 4 respiration (Ref. 27 and data not shown) indicating that mitochondrial depolarization per se does not result in cytochrome c release; a similar result was obtained with rat brain mitochondria (57). These results also demonstrate that calcineurin or activation of group II caspases are not required for cytochrome c release. A mitochondrial fraction from control retinas was loaded as a positive control (Fig. 5, lane 7). The absence of cytochrome oxidase subunit IV in the cytosolic fractions (Fig. 5, lanes 1-6) indicates no contamination of mitochondria.
The findings that CsA, but not FK506, blocked cytochrome c release and reduced rod mitochondrial depolarization strongly

FIG. 4. The Ca 2؉ -and/or Pb 2؉ -induced opening of the mitochondrial PTP results in a decreased rod mitochondrial ⌬⌿.
Retinas were incubated as described in the legend to Fig. 1. Rod inner segment mitochondrial ⌬⌿ m was measured in intact rod photoreceptors using JC-1 staining and fluorescence microscopy as described under "Experimental Procedures." The red fluorescence (585 nm) from hyperpolarized mitochondria was viewed as J-aggregates (A-E), whereas the green fluorescence (520 nm) from depolarized mitochondria was viewed as JC-1 monomer (F-J). Retinas incubated in control buffer had intense red fluorescence from the J-aggregates (A) and weak green fluorescence from the JC-1 monomers (F) in individual rods, indicating that the rod mitochondria were hyperpolarized. Retinas incubated in buffer containing Ca 2ϩ (B and G) or Pb 2ϩ (C and H) showed decreased intensity of red fluorescence and increased green fluorescence indicating that some of the rod mitochondria were depolarized. Retinas incubated with Ca 2ϩ and Pb 2ϩ had very low intensity of red fluorescence (D) and strong green fluorescence (I) indicating that most of these rod mitochondria were depolarized. 10 M CsA partially inhibited the Ca 2ϩ -and Pb 2ϩinduced mitochondrial depolarization as shown by the increased red fluorescence (E) and decreased green fluorescence (J) compared with D and I, respectively. 100 nM FK506 had no effect on the Ca 2ϩ -and Pb 2ϩ -induced mitochondrial depolarization (data not shown). These images are representatives of four to eight retinas for each treatment condition. In K, a representative electron micrograph of a transverse section through rod inner segments illustrates that the ring pattern of mitochondria is similar to the fluorescent pattern observed in the hyperpolarized control retina (A). Original magnification, 18,000ϫ. Bar indicates 1 m.
suggest that PTP opening was involved in these processes. In addition, previous electron microscopy studies show that rod mitochondria from lead-exposed rats are swollen (11) suggesting that cytochrome c may be released as a result of mitochondrial swelling and rupture of the outer membrane. Therefore, light scattering studies were conducted using isolated retinal mitochondria exposed to 475 nM Ca 2ϩ and/or 585 pM Pb 2ϩ , concentrations that produce a transient stimulation of mitochondrial State 4 respiration and a 50 -75% inhibition of mitochondrial State 3 respiration (27). 2 Large amplitude mitochondrial swelling was induced by Ca 2ϩ (Fig. 6A) or Pb 2ϩ (Fig. 6B), and the effects were additive with Ca 2ϩ and Pb 2ϩ (Fig. 6B). Moreover, the effects of Ca 2ϩ and/or Pb 2ϩ were blocked completely by 1 mM EGTA or 0.5 M CsA (94 Ϯ 5% of control) and partially blocked by 0.1 M ruthenium red (65 Ϯ 5% of control) or 0.5 M ruthenium red (84 Ϯ 7% of control). Similar results with CsA were observed in a sucrose buffer, whereas N-ethylmaleimide did not block the effects of Ca 2ϩ and/or Pb 2ϩ (data not shown).
Caspase-9 and Caspase-3 Are Sequentially Activated-Based on the K i values of DEVD-CHO for all the caspases (34), our results suggest that caspase-3, -7, and/or -8 may participate in the Ca 2ϩ -and Pb 2ϩ -induced rod apoptosis and that these caspases act downstream from the mitochondria. In addition, the apical caspase-9 interacts with cytochrome c to activate the group II executioner caspases (1). To determine if caspase-3, -7, -8, and/or -9 were activated during incubation with Ca 2ϩ and Pb 2ϩ and to delineate their sequence of activation, proteolytic activities associated with these caspases were measured by enzymatic assays and Western blot analysis. First, DEVD-pNA, which is selectively cleaved by caspase-3, -7, and -8 (35,59), was used as a substrate. Control retinas exhibited a minimal amount of DEVDase activity (Fig. 7A), whereas retinas incubated with Ca 2ϩ or Pb 2ϩ exhibited a 3-4-fold increase in DEVDase activity. DEVDase activity was increased ϳ7-fold in retinas incubated in both divalent cations again showing the additive effect of Ca 2ϩ and Pb 2ϩ (Fig. 7A). The Ca 2ϩ -and Pb 2ϩ -induced increase in DEVDase activity was inhibited completely by 1 nM DEVD-fmk or 10 M CsA, but not by 100 nM FK506 (Fig. 7A). These results indicate that PTP opening is upstream of caspase activation and that calcineurin is not involved. Second, to identify the specific caspases and their sequence of activation more selective caspase substrates and inhibitors were utilized. Retinas incubated in control buffers exhibited minimal caspase activity, as shown by the minimal cleavage of DEVD-pNA, DMQD-pNA, and LEHD-pNA (Fig.  7B). The cleavage of these three substrates was increased 5-7fold in retinas incubated with both Ca 2ϩ and Pb 2ϩ . DMQD-CHO, a selective caspase-3 inhibitor (36), completely inhibited the Ca 2ϩ -and Pb 2ϩ -induced increase in DEVDase and DMQ-Dase activity but not the LEHDase (caspase-9) activity. Similarly, 1 nM DEVD-fmk did not inhibit LEHDase activity (data not shown). In contrast, CsA and the caspase-9-selective inhibitor LEHD-fmk (35) inhibited completely the Ca 2ϩ -and Pb 2ϩinduced increase in DEVDase, DMQDase, and LEHDase activity, whereas the caspase-8 selective inhibitor IETD-fmk (34) did not (Fig. 7A). These results provide evidence that caspase-9 was activated following the opening of the mitochondrial PTP, caspase-3 was activated by caspase-9, and caspase-8 did not participate in this apoptotic cascade. The fact that caspase activity was inhibited by CsA and that high concentrations of EGTA (2.5 mM) and dithiothreitol (10 mM) were in the assay buffers eliminated the possibility that the Ca 2ϩ -and/or Pb 2ϩinduced increase in caspase activity was a direct effect of these divalent cations on the caspases. This is consistent with the findings that caspase activity is not affected by concentrations of Ca 2ϩ below 100 mM (43).
The processing of procaspase-9, -3, and -7 was examined using Western blot analysis. Following incubation in Ca 2ϩ or Pb 2ϩ , the 46-kDa proform of caspase-9 was processed into a 35-kDa fragment (p35) (Fig. 8A, lanes 2 and 3), and the amount of p35 doubled in the presence of both cations (Fig. 8A, lane 4). A quantitatively similar effect was observed for procaspase-3 such that its 32-kDa proform was processed into p17 fragments in retinas incubated in Ca 2ϩ and/or Pb 2ϩ (Fig. 8B). The 35-kDa proform of caspase-7 was detected in all retinas. However, it was not processed into either its p19 or p32 fragments (7) by Ca 2ϩ and/or Pb 2ϩ (data not shown), indicating that this group II caspase was not involved in the Ca 2ϩ -or Pb 2ϩ -triggered rod apoptosis. The effect of Ca 2ϩ and/or Pb 2ϩ on the processing of the death substrates PKC ␦ and PARP also was examined (Fig.  8, C and D). PARP is cleaved by group II and III caspases, whereas PKC ␦ is cleaved only by caspase-3 (60). Following incubation in buffers containing Ca 2ϩ and/or Pb 2ϩ , intact PKC ␦ (78 kDa) and PARP (116 kDa) were cleaved into their signature 42- (Fig. 8C) and 85-kDa (Fig. 8D) fragments, respectively. The amounts of p42 PKC ␦ and p85 PARP were doubled in retinas incubated with Ca 2ϩ and Pb 2ϩ compared with either divalent cation alone. FIG. 7. Caspase-9 and caspase-3 are sequentially activated by Ca 2؉ and/or Pb 2؉ . Retinas were incubated as described in the legend to Fig. 1. In A, DEVDase activity was measured as described under "Experimental Procedures." The effects of Ca 2ϩ and/or Pb 2ϩ as well as Ca 2ϩ and Pb 2ϩ in the presence of 1 nM DEVD-fmk, 10 M CsA, or 100 nM FK506 were determined. The arbitrary values represent the mean Ϯ S.E. of three to seven experiments for each treatment condition. All values with asterisks are significantly different from controls at: *, p Ͻ 0.05; and **, p Ͻ 0.02. In B, retinas were incubated in control buffer or with Ca 2ϩ and Pb 2ϩ in the absence or presence of the caspase-3 inhibitor (100 M DEVD-CHO), caspase-8 inhibitor (20 M IETD-fmk), caspase-9 inhibitor (20 M LEHD-fmk), or mitochondrial PTP inhibitor (10 M CsA). DEVDase activity, caspase-3-specific protease activation, and caspase-9-specific protease activation were measured using DEVD-pNA, DMQD-pNA, and LEHD-pNA as substrates, respectively, as described under "Experimental Procedures." In control retinas, the cleavage of all three substrates was minimal. The arbitrary values represent the mean Ϯ S.E. of three to five experiments for each treatment condition. All values with asterisks are significantly different from controls at p Ͻ 0.02. To corroborate the finding that Ca 2ϩ and Pb 2ϩ induced sequential activation of caspase-9 and caspase-3 activity, the processing of these caspases in the presence of 1 nM DEVD-fmk or 10 M CsA was examined using Western blot analysis. As noted above, incubation in Ca 2ϩ and Pb 2ϩ resulted in the formation of a caspase-9 p35 fragment (Fig. 8E, lane 1) and a caspase-3 p17 fragment (Fig. 8F, lane 1). DEVD-fmk blocked completely the formation of the p17 fragment (Fig. 8F, lane 2) but did not affect the formation of the p35 fragment (Fig. 8E,  lane 2). CsA, however, blocked the formation of both the p17 and p35 fragments (Fig. 8, E and F, lane 3). These results clearly demonstrate the Ca 2ϩ -and Pb 2ϩ -induced sequential activation of caspase-9 and caspase-3.
Disruption of Mitochondrial Membrane Integrity Precedes the Activation of Caspases-Retinas were incubated with Ca 2ϩ and Pb 2ϩ for 5, 10, or 15 min to determine the time-dependent sequence that links the following events: release of cytochrome c, decrease in mitochondrial ⌬⌿, increase in caspase activity, and rod-selective apoptosis. The results are summarized in Fig.  9. Significant mitochondrial depolarization occurred during the initial 5 min of incubation and continued until the end of incubation. Significant cytochrome c release also occurred during the initial 5 min of incubation and it increased almost linearly during the entire incubation period. In contrast, DEV-Dase activity increased slightly, but significantly, during the first 5-10 min of incubation and then increased 7-fold by 15 min. Finally, the number of apoptotic rods increased slightly after 10 min of incubation and then increased ϳ8-fold by 15 min. These results, coupled with the mitochondrial swelling and CsA data, demonstrate that alterations in mitochondrial function are early and critical events in the Ca 2ϩ -and Pb 2ϩinduced rod apoptosis. DISCUSSION The results from this study identified rod photoreceptors as selective targets of Ca 2ϩ and Pb 2ϩ and delineated the biochemical mechanisms and temporal sequence of events in the apoptotic signaling cascade. The rod-selective apoptosis produced by pathophysiologically relevant concentrations of Ca 2ϩ and/or Pb 2ϩ is similar to that observed in a wide variety of human and animal retinal degenerations where Ca 2ϩ overload appears to have a fundamental role (9 -15).
The Ca 2ϩ -and/or Pb 2ϩ -induced Apoptosis Is Rod-selective-Several independent biochemical and morphological measures were used to establish that the cell death produced by Ca 2ϩ and/or Pb 2ϩ was apoptotic, rod-selective, and that the effects of these divalent cations were additive. First, genomic DNA was cleaved into HMW fragments of 600, 300, and 50 kbp following incubation in buffers containing Ca 2ϩ or Pb 2ϩ . These discrete large DNA fragments result from changes in the integrity of the higher order chromatin structure and are considered one of the hallmarks of apoptosis (38). Second, AO/EtBr-staining studies revealed that the apoptotic cell death was rod-selective, the effects of Ca 2ϩ and Pb 2ϩ were additive, and DEVD-fmk and CsA were cytoprotective. Electron microscopic analysis confirmed that apoptosis was rod selective. Third, DEVDase activity and the processing of caspase-9 and caspase-3, selective biochemical markers of apoptosis (43), were increased to the same degree by Ca 2ϩ or Pb 2ϩ , and the effects of both divalent cations were additive. A scheme for the Ca 2ϩ -and Pb 2ϩ -induced apoptotic process in rods is presented in Fig. 10.
Following incubation in Ca 2ϩ alone or Ca 2ϩ and Pb 2ϩ , retinal [Ca] was increased to approximately the same level (23-37%) found in retinas of lead-exposed rats and retinal degeneration mice undergoing rod apoptosis (13,26,27). The finding that free Ca 2ϩ and free Pb 2ϩ were selectively accumulated in the photoreceptors is consistent with our observations that the apoptotic cell death was rod-selective. The rod selectivity is thought to result from a sustained increase in rod [Pb] and/or FIG. 9. The Ca 2؉ -and Pb 2؉ -induced mitochondrial depolarization and cytochrome c release precede caspase activation and chromatin condensation. Retinas were incubated in control buffers or with Ca 2ϩ and Pb 2ϩ for 5, 10, or 15 min. Detection and quantification of mitochondrial ⌬⌿, cytochrome c release, DEVDase activity, and a number of rod apoptotic cells were conducted as described under "Experimental Procedures." Quantitative analysis showed that significant mitochondrial depolarization and cytochrome c release occurred during the initial 5 min of incubation and increased further with 10 and 15 min of incubation. DEVDase activity was slightly, but significantly, increased during the first 5-10 min of incubation and then increased markedly by 15 min. The first significant increase in the number of apoptotic rods occurred at 10 min and then it increased dramatically with 15 min of incubation. The arbitrary values represent the mean Ϯ S.E. of three to seven experiments for each treatment condition.
FIG. 10. Schematic diagram of the cell signaling pathway mediating the Ca 2؉ -and/or Pb 2؉ -induced rod apoptosis. Ca 2ϩ and/or Pb 2ϩ enter the rod photoreceptor outer segment via the cGMP-gated channel and subsequently enter the rod mitochondria through the ruthenium red (RR)-sensitive Ca 2ϩ uniporter and bind to the internal metal (Me 2ϩ ) site of the mitochondrial permeability transition pore (PTP). This opens the PTP, which leads to mitochondrial depolarization, swelling, and osmosis-induced rupture of outer mitochondrial membrane. Cytochrome c (cyt c) is released from the mitochondria to cytosol where it interacts with apoptotic protease activating factor-1 (Apaf-1), dATP/dADP, and procaspase-9, and where caspase-9 is presumably activated. Caspase-9 activates procaspase-3. Caspase-3 subsequently activates endonucleases that cleave DNA into HMW fragments and this results in the occurrence of morphological rod apoptosis. DEVD, Ac-DEVD-fmk; DMQD, Ac-DMQD-fmk; LEHD, Ac-LEHD-fmk.
[Ca] compared with cones because the Na ϩ /Ca 2ϩ (K ϩ ) exchanger in the rods is 8 -10 times slower than in cones (33). The absence of free Ca 2ϩ and/or Pb 2ϩ in bipolar cells and the corresponding absence of bipolar cell apoptosis in the in vitro retinal preparation, compared with rats exposed to lead during development or adulthood (11,12), most likely results from the lack of inhibition of bipolar cell cGMP phosphodiesterase by IBMX (34) during the short in vitro incubation period. This is supported by findings that only the rod, but not bipolar, cell cGMP levels were elevated in isolated rat retinas incubated in IBMX (61). These results suggest that the intracellular levels of Ca 2ϩ and Pb 2ϩ produced during in vitro incubation are pathophysiologically relevant and that the preparation is useful for determining the biochemical mechanisms underlying rod-selective death induced by Ca 2ϩ overload and/or Pb 2ϩ neurotoxicity.
The Mitochondrial PTP Is the Target Site of Ca 2ϩ and/or Pb 2ϩ -Our results reveal that mitochondria are the initial target site in rods responsible for mediating the Ca 2ϩ -and Pb 2ϩ -induced apoptotic rod cell death. Both Ca 2ϩ and Pb 2ϩ rapidly entered the rod mitochondria through the Ca 2ϩ uniporter as evidenced by the ruthenium red blockade of divalent cation uptake and the transient stimulation of state 4 respiration (27). 2 Ca 2ϩ and Pb 2ϩ had an additive effect on decreasing ⌬⌿ m , producing mitochondrial swelling and releasing cytochrome c into the cytosol. These were early and critical events in the rod cell apoptotic process. CsA, but not DEVDfmk or FK506, completely blocked the Ca 2ϩ -and/or Pb 2ϩ -induced cytochrome c release and markedly reduced the mitochondrial depolarization produced by these divalent cations. The molecular mechanisms of action of Ca 2ϩ and Pb 2ϩ on the rod mitochondrial PTP are not known. There are two classes of PTP pore agonists/antagonists: one modulates the membrane and surface potential of the voltage sensor, and the other modulates cyclophilin-D binding (3). Although Ca 2ϩ and Pb 2ϩ have diverse effects on mitochondria (2, 3, 27, 62), our current results suggest a few unifying, though not mutually exclusive, mechanisms. First, Ca 2ϩ binds directly to the divalent metal (Me 2ϩ ) binding site on the matrix side of the mitochondrial PTP and induces its opening. These effects are inhibited by CsA (3, 51) but not by FK506 (37). In addition to CsA, which mediates its effect through cyclophilin-D, several group IIA and VIIA divalent cations competitively inhibit the Ca 2ϩ -induced PTP opening at this Me 2ϩ site. However, the divalent cation-induced inhibition of PTP occurs only when the Me 2ϩ /Ca 2ϩ ratio exceeds 10 (63). The findings that the Pb 2ϩ /Ca 2ϩ ratios in our experiments were 0.001-0.002, that the effects of Ca 2ϩ and Pb 2ϩ on all measures were additive and inhibited by CsA, and that Pb 2ϩ can function as a potent Ca 2ϩ agonist (62, 64) suggest one strong candidate mechanism for PTP opening is the occupation of the internal Me 2ϩ binding site by Ca 2ϩ and Pb 2ϩ . Second, it is possible that Ca 2ϩ and Pb 2ϩ opened the PTP by oxidizing GSH and/or NADH resulting in the oxidation of the sensitive vicinal dithiol present in the "S-site" and "P-site" of the PTP, respectively (3). Based on our findings of no Ca 2ϩand/or Pb 2ϩ -induced alterations in rod redox status or ROS production, we conclude that oxidative stress is not the mechanism underlying the proapoptotic action of these divalent cations. This is in contrast to the proapoptotic effects of Hg 2ϩ , Mn 2ϩ , Cu 2ϩ , and methyl mercury that produce ROS (21, 29 -31). Finally, it is possible that Pb 2ϩ functions independently of Ca 2ϩ and opens the PTP by cross-linking the dithiol in the S-site like the trivalent arsenite (1-3). However, preliminary experiments in isolated mitochondria using N-ethylmaleimide, a well-known blocker at both the S-sites and P-sites (3), showed no protective effect against Pb 2ϩ -or Ca 2ϩ -induced mitochon-drial swelling. 3 Altogether, these results indicate that both Ca 2ϩ and Pb 2ϩ bind to the matrix Me 2ϩ binding site of the mitochondrial PTP and open the PTP.
Results from several different experimental systems suggest that cytochrome c is released from mitochondria either by a PTP-dependent mechanism (1) or by one of two PTP-independent mechanisms (4,65). Our JC-1 staining and cytochrome c release studies are consistent with the PTP-dependent mechanism. Therefore, light scattering (mitochondrial swelling) experiments were conducted in isolated retinal mitochondria incubated in the presence of respiratory substrates, Mg 2ϩ and P i . Ca 2ϩ and/or Pb 2ϩ induced rapid and large amplitude mitochondrial swelling that was blocked by EGTA, ruthenium red, and CsA, indicating that the swelling was because of the opening of the mitochondrial PTP. In contrast, recent data showed that isolated rat brain mitochondria only exhibited Ca 2ϩ -induced swelling and permeability transition in the absence, but not presence, of Mg 2ϩ and ATP (58). The reason for these tissue differences is unknown especially because 1-3 log units lower free [Ca 2ϩ ] or CaCl 2 /mg of protein was used in the retinal mitochondrial preparation (27) compared with the brain mitochondrial preparation (58).
Caspase-9 Is the Initiator Caspase and Caspase-3 Is the Executioner Caspase-Results from our pharmacological experiments establish that caspase-9 is the initiator caspase, whereas caspase-3 is the executioner caspase in Ca 2ϩ -and/or Pb 2ϩ -induced rod apoptosis. Specifically, the caspase-3 selective inhibitor DMQD-CHO inhibited DEVDase activity but did not inhibit LEHDase activity, whereas the caspase-9 selective inhibitor LEHD-fmk inhibited the activity of all three tetrapeptide-pNA substrates indicating that caspase-9 was activated upstream of caspase-3. Interestingly, incubation with 1 nM DEVD-fmk did not inhibit the Ca 2ϩ -and Pb 2ϩ -induced LEH-Dase activity but it totally inhibited DEVDase activity, HMW DNA fragmentation, and nuclear chromatin condensation. This shows that activation of caspase-9 alone was not enough to cause apoptosis, whereas activation of caspase-3 was required for the cleavage of death substrates, activation of endonucleases to cleave DNA into HMW fragments, and morphological apoptosis. In vitro experiments have shown that caspase-9 activates procaspase-3 (6,7,66) and that this produces a feedback amplification loop that further activates procaspase-9 (6). In retinas incubated with Ca 2ϩ and Pb 2ϩ , CsA blocked completely the activation of procaspase-9 and procaspase-3, whereas neither DMQD-CHO nor DEVD-fmk decreased caspase-9 activity or cleavage, respectively. These results conclusively demonstrate the Ca 2ϩ and Pb 2ϩ induced sequential activation of caspase-9 and caspase-3. In addition, the presence of similar amounts of the caspase-9 p35 fragment in retinas incubated with Ca 2ϩ and Pb 2ϩ in the absence or presence of DEVD-fmk suggests that there was no or limited activation of caspase-9 by caspase-3. Finally, there was no evidence of Ca 2ϩand Pb 2ϩ -induced caspase-8 activity or processing of caspase-7. These findings are consistent with recent results showing that caspase-9 and caspase-3 are activated sequentially in chemicaland cytochrome c-induced apoptosis (7, 66), whereas caspase-8 is activated in receptor-mediated apoptosis (7).
In summary, we have demonstrated that elevation of intracellular Ca 2ϩ and/or Pb 2ϩ in rat rod photoreceptors produces rod-selective apoptosis and that the effects of Ca 2ϩ and Pb 2ϩ are additive. Based on the results, we have proposed a scheme for the Ca 2ϩ -and Pb 2ϩ -induced apoptotic process in rods (Fig.  10). A similar apoptotic cell-signaling cascade may underlie rod-selective or neuronal apoptosis observed in humans and animals with different retinal and neural degenerations resulting from Ca 2ϩ overload and/or Pb 2ϩ neurotoxicity. The mechanistic knowledge gained from the delineation of the apoptotic cascade in rod photoreceptors will be useful for the development of specific neuroprotective strategies.