The apoptosis/necrosis transition in cerebellar granule cells depends on the mutual relationship of the antioxidant and the proteolytic systems which regulate ROS production and cytochrome c release en route to death

Authors


Address correspondence and reprint requests to Ersilia Marra, Istituto di Biomembrane e Bioenergetica CNR, Via Amendola 165/A, 70126 Bari, Italy. E-mail: csmmaa08@area.ba.cnr.it

Abstract

We investigate the death route induced by potassium depletion in cerebellar granule cells in 0–15 h time range and study whether and how mutual relationship occurs between the cell antioxidant and proteolytic system. To achieve this, we incubated cells in the absence or presence of inhibitors of the antioxidant system, including superoxide dismutase and catalase, and of the proteolytic system, consisting of proteasomes and caspases, and investigated whether and how (i) cell survival, (ii) reactive oxygen species (ROS) production and (iii) antioxidant enzyme and caspase-3 activity change as a function of time after the apoptotic stimulus. The involvement of both antioxidant and proteolytic system on cytochrome c release was also investigated. Cell survival was found to increase in the presence of either proteasome or caspase inhibitors. On the contrary, as a result of the antioxidant system impairment, shift from apoptosis to necrosis occurs. We show that the antioxidant system, which exhibits a huge activity increase up to 3 h after apoptosis induction, is subjected to the proteasome-dependent proteolysis and that the increase in the antioxidant system found in the absence of proteasome activity is accompanied by ROS production decrease. Consistently, the early ROS-dependent release of cytochrome c was found to be prevented when the activity of the antioxidant system increased. Finally, caspase-3 activation was prevented by the inhibitors of both antioxidant system and proteasome.

Abbreviations used
Act D

actinomycin D

AOS

antioxidant system

BME

basal medium Eagle

CP

captopril

CGCs

cerebellar granule cells

cyt c

cytochrome c

C-GNT-cells

control glutamate-treated cells

7DIV

7 days in vitro

GDH

glutamate dehydrogenase

GNT

glutamate neurotoxicity

GNT-cells

glutamate-treated cells

GSH

reduced glutathione

GSSG

glutathione disulfide

LDH

lactate dehydrogenase

MK801

(+/–)-5-methyl-10,11-dihydro-5H-dibenzo(a,d)cyclohepten-5,10-imine hydrogen maleate

NH2-TZ

NH2-triazole

No-AOS

S-K5 cells in the presence of AOS inhibitors

No-AOS/CASPASE

S-K5 cells in the presence of AOS and caspase inhibitors

No-AOS/PROTEASOME

S-K5 cells in the presence of AOS and proteasome inhibitors

No-CASPASE

S-K5 cells in the presence of caspase inhibitor

No-PROTEASOME

S-K5 cells in the presence of proteasome inhibitor

No-PROTEASOME/CASPASE

S-K5 cells in the presence of proteasome and caspase inhibitors

O2·

superoxide anion

PBS

phosphate-buffered saline medium

ROS

reactive oxygen species

S-K25 cells

control cells

S-K5 cells

apoptotic cells

SOD

superoxide dismutase

z-VAD

z-VAD-fmk

Cultured cerebellar granule neurones deprived of depolarizing levels (25 mm) of extracellular potassium are a model for studying cell death as they undergo apoptosis (D'Mello et al. 1993; Galli et al. 1995; Vitolo et al. 1998) in a manner interpreted as the in vitro counterpart of deafferentation of these neurones (Borsello et al. 2000).

At present, the elucidation of the processes by which signalling pathways lead cells to death, via apoptosis/necrosis, and by which death occurs is the subject of many studies dealing with a variety of apoptosis-related events. In this regard, reactive oxygen species (ROS) generated during apoptosis have been recognized as mediators of intracellular apoptotic signalling cascades (Greenlund et al. 1995; Schulz et al. 1996; Cai and Jones 1998; Esteve et al. 1999; Valencia and Moran 2001; Curtin et al. 2002). Consistently, overexpression of superoxide dismutase (SOD) was found to prevent apoptosis both in neuronal cell lines (Rabizadech et al. 1995) and in primary neuronal cultures (Greenlund et al. 1995), while glutathione depletion proved to result in an increase of ROS levels (Tan et al. 1998). ROS generation proved to occur also in necrosis (Lafon-Cazal et al. 1993; Gunasekar et al. 1995; Atlante et al. 1997). On the other hand, cytochrome c (cyt c) release occurs in cerebellar granule neurones both during potassium depletion and in glutamate-necrosis (Bobba et al. 1999; Atlante et al. 1999). In this case, cyt c release is largely prevented if ROS production is reduced by antioxidants (Atlante et al. 2000, 2001). Furthermore, in the early phase of apoptosis, cyt c release from mitochondria is prevented by SOD (Fujimura et al. 1999, 2000), as well as by MG132 (Bobba et al. 2002), a proteasome inhibitor, but not by the caspase inhibitor, z-VAD (Bobba et al. 1999). Thus, the question arises as to whether and how the cell antioxidant system (AOS) and consequently ROS production are regulated during apoptosis by the cell proteolytic system and whether such a regulation may modify the death route. In this paper, we assay the antioxidant enzymes and the proteolytic system at different times after the apoptotic stimulus and show that there is mutual modulation of the components of the cell antioxidant and proteolytic system which play a major role in determining both ROS production and cyt c release. Interestingly, apoptosis/necrosis shift occurs in cerebellar granule cells when the antioxidant system is totally impaired.

Materials and methods

Materials

Tissue culture medium and fetal calf serum were purchased from Gibco (Grand Island, NY, USA) and tissue culture dishes were from Nunc (Taastrup, Denmark). All enzymes and biochemicals were from Sigma Chemicals Co. (St Louis, MO, USA). The inhibitor z-VAD-fmk was purchased from Calbiochem (La Jolla, CA, USA). Anti-cyt c antibodies (7H8–2C12) were purchased from Pharmigen (San Diego, CA, USA); antiglutamate dehydrogenase antibodies were kindly supplied by Dr F. Rothe (Institut fur Medizinische Neurobiologie, University of Magdeburg, Magdeburg, Germany).

Cell cultures

Primary cultures of cerebellar granule cells (CGCs) were obtained from dissociated cerebellar of 7-day-old Wistar rats as in Levi et al. (1984). Cells were plated in basal medium Eagle (BME) supplemented with 10% fetal calf serum, 25 mm KCl, 2 mm glutamine and 100 µg/mL gentamicin on dishes coated with poly l-lysine. Cells were plated at 2 × 106/35-mm dish or 6 × 106/60-mm dish. 1β-Arabinofuranosylcytosine (10 µm) was added to the culture medium 18–22 h after plating to prevent proliferation of non-neuronal cells.

Induction of apoptosis

Apoptosis was induced as in D'Mello et al. (1993): at 6–7 days in vitro (DIV), cells were washed twice and switched to a serum-free BME (S-), containing 5 mm KCl and supplemented with 2 mm glutamine and 100 µg/mL gentamicin for the time reported in the figure legends. Apoptotic cells are referred to as S-K5 cells. Control cells were treated identically but maintained in serum-free BME medium supplemented with 25 mm KCl for the indicated times and are referred to as S-K25 cells.

The addition of inhibitors or other compounds was made at the apoptosis induction time. The final concentration of dimethyl sulphoxide (DMSO), when required, was kept below 0.1%. Corresponding controls were treated either with the same inhibitor or with the same concentration of DMSO.

Glutamate neurotoxicity induction

Glutamate exposure was performed 7 days after plating. Primary cultures were exposed for 30 min to glutamate (100 µm) at 25°C in Locke's solution (154 mm NaCl, 5.6 mm KCl, 3.6 mm NaHCO3, 2.3 mm CaCl2, 5.6 mm glucose, 10 mm HEPES pH 7.4) in the presence of 1 µm glycine added in order to fully activate NMDA-sensitive glutamate recognition sites (Johnson and Ascher 1987). Cells were then replenished with BME containing 25 mm KCl, 2 mm glutamine and gentamicin (100 µg/mL), and put in the incubator. For the quantitative assessment of glutamate neurotoxicity (GNT), cell integrity and count were measured, as described below, after 12–24 h. Glutamate-treated cells and control cells are referred to as GNT- and C-GNT cells, respectively.

Assessment of neuronal viability

Viable CGCs were quantified by counting the number of intact nuclei after dissolving the cells in detergent-containing solution as described in Volontèet al. (1994). Apoptosis or neurotoxicity was expressed as the percentage of intact cells with respect to control cells kept under the same respective experimental conditions. In control experiments, 95–97% integrity was found after 24 h.

DNA fragmentation analysis

Fragmentation of DNA was performed as in Hockenbery et al. (1990). Briefly CGCs (6 × 106) were plated in poly l-lysine-coated 60-mm tissue culture dishes, collected with cold phosphate-buffered saline (PBS, pH 7.2) and, after removal of the medium and washing once with cold PBS, CGCs were centrifuged at 3500 g for 5 min. The pellet was lysed in 10 mm Tris–HCl, 10 mm EDTA, 0.2% Triton X-100 (pH 7.5). After 30 min on ice, the lysates were centrifuged at 17 000 g for 10 min at 4°C. The supernatant was digested with proteinase K and then extracted twice with phenol–chloroform/isoamyl alcohol (24 : 1). The aqueous phase, containing soluble DNA, was recovered and nucleic acids were precipitated with sodium acetate and ethanol overnight. The pellet was washed with 70% ethanol, air-dried and dissolved in TE buffer (10 mm Tris–HCl, 1 mm EDTA, pH 7.5). After digestion with RNase A (50 ng/mL at 37°C for 30 min), the sample was subjected to electrophoresis in a 1.8% agarose gel and visualized by ethidium bromide staining. Soluble DNA from equal numbers of cells was loaded in each lane.

Lactate dehydrogenase activity

The activity of lactate dehydrogenase (LDH) released into the culture medium was determined spectrophotometrically according to (Bernt and Bergmeyer 1963) and expressed as percentage of maximum LDH released at 24 h after induction of glutamate neurotoxicity. An aliquot (100 µL) of culture medium (2 × 106 cells/2 mL culture medium) was added to 2 mL of 50 mm Tris–HCl buffer pH 7.4 in the presence of 0.2 mm NADH. The assay reaction was started by adding 0.6 mm pyruvate.

Superoxide anion detection in CGCs

Superoxide anion (O2·) was detected, as in Atlante et al. (1997) and Atlante and Passarella (1999), according to the Fe3+-cyt c method (Forman and Fridovich 1973). The newly formed O2· gave an increase in absorbance at 550 measured using a Perkin-Elmer LAMBDA-5 spectrophotometer equipped with a thermostated holder (Perkin-Elmer, Foster City, CA, USA). A calibration curve was made by using an O2· producing system, i.e. xanthine plus xanthine oxidase, and an O2· detection system, i.e. Fe3+-cyt c, that, in the presence of O2·, gave Fe2+-cyt c with 1 : 1 stoichiometry.

Cell homogenate preparation

Before each experiment, the culture medium was removed and the plated CGCs, i.e. S-K5 or S-K25 cells in the absence or presence of different compounds, were washed with PBS and then collected by gentle scraping in a final volume of 4 mL PBS/90-mm diameter dish. Suspended granule cells showed full viability, even though they lacked the morphological organisation present in culture dishes such as cell–cell and cell–substrate contacts as well as neuritis. Cell integrity, which remains rather constant for 3–5 h, was quantitatively assessed by checking the inability of cells to oxidize externally added succinate, which cannot enter intact cells (Berry et al. 1991), by checking the ability of ouabain to block glucose transport in cells (Atlante et al. 1996) and by counting dead cells, identified as large phase-bright cell bodies, as in Volontèet al. (1994). The final cell suspension routinely contained 85–95% intact cells and was prepared after 6–7 DIV. Cell homogenate from the cell suspension was obtained by 10 strokes with a Dounce potter at room temperature. With this procedure, LDH is released and subsequent treatment with Triton X-100 does not cause further release.

The cell protein assay was determined according to Waddel and Hill (1956), with bovine serum albumin used as a standard.

Antioxidant enzyme activities

The activities of SOD and catalase were determined with cell homogenate (about 0.1 mg cell protein).

SOD (E.C.1.15.1.1) activity was measured by the inhibition of xanthine oxidase/cyt c system reaction as described by Kirby and Fridovich (1982). In this assay, xanthine oxidase, acting on xanthine in the presence of oxygen, generates superoxide anion, O2· which reduces cyt c, and this reduction is inhibited by SOD. One enzymatic unit of SOD is the amount of enzyme required to inhibit the rate of reduction of cyt c by 50%: then the activity of SOD is expressed as the percentage inhibition of the control reaction. CGC homogenate was suspended at 25°C in 1.5 mL PBS buffer in the presence of Fe3+-cyt c (10 µm) plus xanthine (50 µm). The reaction was started with xanthine oxidase addition and the absorbance increase at 418 nm was monitored.

Catalase (E.C.1.11.1.6) determination was performed by a spectrophotometric assay based on the catalysed decomposition of H2O2, essentially according to Aebi (1984). The peroxide decomposition rate is directly proportional to the enzyme activity. To determine the catalase activity, cell homogenate was suspended at 25°C in 1.5 mL PBS buffer in a quartz cuvette. After reading the initial absorbance at 240 nm, 100 µL H2O2 (final concentration 10 mm) was added and the decrease in absorbance monitored. The slope of the absorbance versus time plot is directly proportional to the activity of the sample and is expressed as ΔA240/min × 106 cells.

Glutathione disulfide/reduced glutathione ratio measurement

Glutathione (GSH) or glutathione disulfide (GSSG) were assayed in cell homogenate, according to Akerboom and Sies (1981). Briefly, GSH in the presence of methylglyoxal (2 mm) and glyoxalase I (6 e.u.) was specifically converted into S-lactoyl-GSH which could be monitored directly at 240 nm; GSSG amount was assayed in the same cuvette by measuring the stoichiometric conversion of NADPH (10 µm) spectrophotometrically at 340 nm in the presence of glutathione reductase (1 e.u.).

Caspase-3 activity

Caspase activity was measured by using the Clontech ApoAlert Caspase-3 Assay Kit following manufacturer's instructions. DEVD-pNA was used as a colourimetric substrate. The increase in protease activity was determined by the spectrophotometric detection at 405 nm of the chromophore p-nitroanilide (pNA) after its cleavage by caspase-3 from the labelled caspase-3-specific substrate (DEVD-pNA).

Polarographic measurements to detect cyt c release

O2 consumption was measured polarographically by means of a Gilson 5/6 oxygraph using a Clark electrode, as described in Atlante et al. (1996, 1998). Instrument sensitivity was set to a value in order to follow rates of O2 uptake as low as 0.5 natoms/min/mg protein. The cell homogenate in PBS (about 0.2 mg protein) was incubated in a thermostated (25°C) water-jacketed glass vessel (final volume equal to 1.5 mL). In order to detect the cyt c released in the extramitochondrial phase, the ability of CGC homogenate to oxidise ascorbate was checked. Briefly, as ascorbate cannot permeate per se the outer mitochondrial membrane (Alexandre and Lehninger 1984), thus its oxidation can occur as a result of the release from mitochondria of a component that can oxidize ascorbate and then reduce oxygen via cytochrome c oxidase, i.e. cyt c (Atlante et al. 1999, 2000).

Immunoblot analysis

Immunoblot analysis was performed on cytosolic and mitochondrial fractions from control and apoptotic cultures, as in Bobba et al. (1999).

Statistical analysis and computing

Data are reported as mean values ± standard errors for the indicated experiments. When necessary statistical analysis was carried out according to the Student's t-test. Experimental plots were obtained by means of GraFit (Erithacus Software, Horley, Surrey, UK).

Results

In order to get some insight into the mechanism by which CGCs die as well as into the role played by certain cell components in the processes leading to death, 7DIV cultures were kept either in high potassium medium (S-K25 cells) or subjected to low potassium shift (S-K5 cells), which is responsible of triggering apoptosis (D'Mello et al. 1993), in the absence or presence of a variety of compounds designed to inhibit certain enzymes which participate in the processes leading to programmed cell death. Captopril (CP, 5 mm) was used to inhibit SOD (Jay 1998). As this compound, when used in different experimental systems, shows a variety of effects, including prevention of ROS dependent toxicity (Abd El-Aziz et al. 2001; Candan and Alagozlu 2001) and, importantly, of caspase-3 activation (Uhal et al. 1998), control was made that under our conditions it has no significant effect on caspase activity. Use was also made of NH2-triazole (NH2-TZ, 10 mm), which proves to have no effect on caspase activity, but inhibits catalase (Antonenkov et al. 1989; Dringen and Hamprecht 1997), MG132 (5 µm), a proteasome inhibitor (Figueiredo-Pereira et al. 1994), and z-VAD (100 µm), a broad-range inhibitor of caspases (Fenteany et al. 1995). Cells subjected to potassium shift to induce apoptosis and treated with each of the above-mentioned inhibitors will be referred to as ‘No-AOS’ cells, when antioxidant enzymes are inhibited by CP plus NH2TZ, ‘No-PROTEASOME’ cells, when proteasome activity is blocked by MG132, ‘No-CASPASE’ cells, when caspase activity is prevented by z-VAD and ‘No-PROTEASOME/CASPASE’ cells, when the inhibitor pair MG132 and z-VAD was used. Consistently, No-AOS/PROTEASOME and No-AOS/CASPASE conditions refer to apoptosis conditions in the presence of the respective inhibitors.

Effect of antioxidant enzyme system, proteasome and caspase inhibitors on the survival of CGCs subjected to potassium shift

The survival of S-K5, No-PROTEASOME, No-CASPASE and No-AOS cells was compared as measured up to 15 h after potassium shift (Fig. 1); as a control, it was checked that the inhibitors had no effect on S-K25 cell survival (not shown). In agreement with D'Mello et al. (1993), cell survival was reduced by about 20% (p < 0.0001, six experiments) and 50% (p < 0.0001, six experiments) at 8 h and 15 h after potassium shift (S-K5), respectively. As a control of the ROS involvement in the processes leading to cell death, a mixture of antioxidant compounds, including vitamin E (0.1 mm), vitamin C (0.1 mm) and GSH (1 mm), as well as SOD (50 e.u./mL) was found to prevent apoptosis largely (not shown, see Atlante et al. 1998). Under both No-PROTEASOME and No-CASPASE conditions, no cell death was found up to 8 h with about 80% (p < 0.001, six experiments) survival measured at 15 h in both cases (see also Canu et al. 2000; Valencia and Moran 2001).

Figure 1.

Survival of CGCs subjected to potassium shift. The effect of antioxidant system, proteasome and caspase inhibitors. Rat CGCs (106/well) at 7DIV were incubated in low potassium (S-K5, •) serum-free culture medium in the absence or presence of different inhibitors: MG132 (5 µm), i.e. No-PROTEASOME (□); z-VAD (100 µm), i.e. No-CASPASE (▵); or CP (5 mm) + NH2-TZ (10 mm), i.e. No-AOS (bsl00036). At different times cell viability was determined by counting the number of intact nuclei. Cell viability is expressed as the percentage of S-K25 cells to which a 100% value was given. Control values were 100 ± 10. Results are means ± standard errors of triplicate measurements and representative of six different experiments carried out with different cell preparations from different groups of animals.

As 29% (p < 0.0001, six experiments) and negligible cell survival was measured in No-AOS cells at 8 and 15 h, respectively, we tried to ascertain how CGCs die under these conditions. To achieve this early LDH release (Fig. 2a), DNA laddering (Fig. 2b) and the death sensitivity to MK801 (Fig. 2c), which allows for distinguishing between apoptosis and necrosis death, were checked up to 15 h after the potassium shift. Comparison was made among S-K5, S-K25, No-AOS and cells in which glutamate neurotoxicity (GNT), with its control (C-GNT), had been previously induced as in Atlante et al. (2000), with respect to the LDH release measured in 1- to 15-h time range. As expected, no significant LDH activity was found both in S-K5 and S-K25 culture medium up to 15 h (see D'Mello et al. 1993), while LDH activity was found to increase in GNT cells with respect to its control (C-GNT). Surprisingly, in No-AOS cell culture medium, LDH activity was found to increase further with respect to that measured in GNT, reaching the same activity measured in these cells at 24 h (not shown).

Figure 2.

LDH release, DNA fragmentation and sensitivity of cell survival to MK801 in No-AOS and in cells undergoing necrosis. (a) LDH activity in the extracellular culture medium. Cells were switched to serum-free medium containing low K+ in the absence (S-K5) or presence of AOS inhibitors (No-AOS) for 1–15 h or treated with glutamate (see Materials and methods) for the same times (GNT). LDH activity was assayed, as reported in the Materials and methods, in No-AOS and GNT-cells as well as in the respective controls (S-K25 or C-GNT). Results represent mean ± standard errors of triplicate cultures and are expressed as percentage maximum LDH activity released in GNT at 24 h (100%). In control cultures LDH activity in the extracellular medium was negligible up to 96 h. (b) DNA fragmentation. Soluble DNA was extracted from either neurones switched to serum-free culture medium containing low K+ (5 mm) in the absence (lanes f–i) or presence of actinomycin D (Act D, 1 µg/mL; lane j) or AOS inhibitors (No-AOS; lanes l–o), or neurones treated with glutamate for the same times (lanes u–x). Lanes (b–e) and (q–t) contain DNA from either control cells maintained in high K+ (25 mm) or C-GNT-cells for 1–15 h time range, respectively. DNA from equal numbers of plated cells (6 × 106) was loaded in each lane. Size marker was HaeIII-digested Φx174 phage DNA (lanes a, k, p and z). (c) MK-801 effect on cell survival. Cells were switched to serum-free medium containing low K+ in the absence (S-K5) or presence of AOS inhibitors (No-AOS) for 1–15 h. Where indicated, MK-801 (1 µm) was present in the medium. At different times, cell viability was determined by counting the number of intact nuclei. Cell viability is expressed as the percentage of S-K25 cells to which a 100% value was given. Control values were 100 ± 5. Results are means + standard errors of triplicate measurements and representative of six different experiments carried out with different cell preparations from different groups of animals.

In Fig. 2(b), S-K25, S-K5, No-AOS, C-GNT and GNT cells were compared in 1- to 15-h time range with respect to DNA laddering, which is a specific hallmark of apoptosis. DNA laddering was found only in the 8- to 15-h time range in S-K5 (Fig. 2b, lanes h and i), but it was completely prevented in the presence of the transcriptional inhibitor actinomycin D (Act D, 1 µg/mL; Fig. 2b, lane j).

In another experiment, the effect of either Act D (1 µg/mL) or MK801 (1 µm), a selective NMDA receptor antagonist which inhibits glutamate-necrosis (Calissano et al. 1993; Atlante et al. 1997; Minervini et al. 1997) in CGCs, on cell viability was investigated (Fig. 2c). In agreement with Fig. 2(b), Act D was found to totally prevent S-K5 death, without any effect on No-AOS cell survival. On the other hand, No-AOS cell viability was largely restored by MK801 (29 vs. 70% at 8 h), which has no significant effect on S-K5 survival at the same time.

ROS production in S-K5, No-AOS, No-CASPASE and No-PROTEASOME cells en route to death

In the light of the crucial role played by ROS both in necrosis (Lafon-Cazal et al. 1993; Gunasekar et al. 1995; Atlante et al. 1997, 2001) and in apoptosis (Rabizadech et al. 1995; Greenlund et al. 1995; Schulz et al. 1996; Cai and Jones 1998; Esteve et al. 1999; Valencia and Moran 2001), we investigated ROS production as a function of the time after potassium shift (Fig. 3) under the conditions described above. In order to assess ROS production, use was made of the experimental procedure already developed for CGCs undergoing necrosis (Atlante et al. 1997; Atlante and Passarella 1999), which allows for early in situ O2· detection. As shown in Fig. 3, low O2· levels (0.05 nmol/106 cells) were measured in S-K25 cells with no significant variation up to 15 h. Contrarily, high O2· production occurs in CGCs undergoing apoptosis (S-K5): superoxide production increases up to 3 h, when about 4 nmol/106cells (p < 0.0001, four experiments) are formed, and remains rather constant for the following 15 h. As expected (Atlante et al. 1997, 2000), externally added SOD was found to partially prevent O2· production (not shown); when captopril and NH2TZ were used separately, a slight, but significant increase or no effect on ROS production was measured, respectively (not shown); consistently under No-AOS conditions O2· production increased further (up to 5 nmol/106 cells, p < 0.01, four experiments), while under No-PROTEASOME conditions, the amount of O2· was reduced by about 50% as compared to that measured in S-K5 cells. In No-CASPASE cells the superoxide amount was unchanged with respect to that found in S-K5 cells for the first 3 h incubation, but decreased thereafter in the subsequent hours with about 1.5 nmol/106cells (p < 0.05, four experiments) at 15 h. Interestingly, in No-PROTEASOME/CASPASE cells, O2· production was prevented mostly up to 3 h (about 1.8 nmol/106cells, p < 0.01, four experiments) and was lower than that found for either No-PROTEASOME or No-CASPASE cells in the 8- to 15-h time range. ROS production measured either in No-AOS/PROTEASOME or No-AOS/CASPASE was found not to differ to that measured in No-AOS.

Figure 3.

ROS production in CGCs undergoing apoptosis. The effect of antioxidant system, proteasome and caspase inhibitors. Rat CGCs (0.5 × 106/well) at 7DIV were incubated either in high potassium (S-K25, ○) or in low potassium (S-K5, ●) serum-free culture medium added with Fe3+-cyt c (10 µm) in the absence or presence of different inhibitors: MG132 (5 µm), i.e. No-PROTEASOME (□); z-VAD (100 µm), i.e. No-CASPASE (▵); CP (5 mm) + NH2-TZ (10 mm), i.e. No-AOS (bsl00036); MG132 (5 µm) + z-VAD (100 µm), i.e. No-PROTEASOME/CASPASE (bsl00165); MG132 (5 µm) + CP (5 mm) + NH2-TZ (10 mm), i.e. No-AOS/PROTEASOME (*); z-VAD (100 µm) + CP (5 mm) + NH2-TZ (10 mm), i.e. No-AOS/CASPASE (◊). At different times after apoptosis induction, the culture solution was taken and the increase in absorbance at 550 nm, due to Fe3+-cyt c reduction, was determined. The experimental data are reported as nanomoles ± standard errors of ROS formed per 106 cells, calculated on the basis of the stoichiometry of the reaction using the extinction coefficient determined under our experimental conditions (see Materials and methods). The experiment was repeated four times with different cell preparations.

The AOS activity and the GSSG/GSH ratio in S-K5, No-PROTEASOME and No-CASPASE cells en route to death

Figure 4 shows the activity of SOD (Fig. 4a) and catalase (Fig. 4b) together with the GSSG/GSH ratio (Fig. 4c), which is a measure of thiol oxidation state, in cell homogenates obtained from S-K25, S-K5, No-PROTEASOME and No-CASPASE cells at different times after apoptosis induction. Interestingly, the above enzyme activities were found to be higher in S-K5 than S-K25 cell homogenate already 30 min after apoptosis induction. The maximum enzyme activity increase in S-K5 cell homogenate (about 300%) was measured 3 h after potassium shift, with progressive reduction up to 15 h when it was about 30% with respect to the controls. Under No-PROTEASOME conditions, the AOS enzyme activity proved to be higher than that measured in S-K5 cell homogenate at all the investigated times. In No-CASPASE cells, no AOS enzyme activity change occurred up to 3 h, whereas it remained constant and much higher than that of the S-K5 cells for incubations lasting up to 15 h. SOD and catalase activities were found to change in a quite similar manner when measured up to 3 h in No-PROTEASOME/CASPASE and in No-PROTEASOME cells: an increase in both the activities was found to occur to a higher extent than that found in S-K5 cells. A small decrease was measured at 5 h, with the activities remaining constant and at higher values than those measured in S-K5 cells. Interestingly, when the relative thiol oxidation state was monitored in No-PROTEASOME/CASPASE conditions, a further increase was observed up to 15 h.

Figure 4.

Antioxidant system activity and GSSG/GSH ratio measurements in CGCs undergoing apoptosis. The effect of proteasome and caspase inhibitors. Rat CGCs (2 × 106/well) at 7DIV were incubated either in high potassium (S-K25, ○) or in low potassium (S-K5, •) serum-free culture medium in the absence or presence of the inhibitors MG132 (5 µm), i.e. No-PROTEASOME (□); z-VAD (100 µm), i.e. No-CASPASE (▵); MG132 (5 µm) + z-VAD (100 µm), i.e. No-PROTEASOME/CASPASE (bsl00165). At different times cells were scraped, collected, homogenized and assayed for SOD (a), catalase (b) and GSSG/GSH ratio (c) (see Materials and methods). Each value represents the mean ± standard errors value of antioxidant system activity, expressed as percentage inhibition of Fe3+-cyt c reduction rate for SOD and as ΔΑ240/min for catalase, and of [GSSG]/[GSH] ratio value (see Materials and methods), obtained from five separate experiments carried out using different cell preparations.

The caspase-3 activity in S-K5, No-AOS and No-PROTEASOME cells en route to death

The caspase-3 activity was barely detectable in either S-K25 or S-K5 cells up to 3 h, but thereafter it progressively increased in S-K5 with the maximum (230% over control, p < 0.001, five experiments) reached at 8 h (Fig. 5). Under No-PROTEASOME conditions, only a slight increase over controls was detectable, with the activity remaining constant over the whole 15-h period, thus confirming that caspase-3 activation requires an early active role of proteasomes (Canu et al. 2000). Externally added SOD proved to inhibit partially caspase-3 activation (not shown). Interestingly, a strong reduction of the caspase-3 activity as compared to S-K5 cells was found in No-AOS cells in which apoptosis/necrosis shift occurs (see Fig. 2). As a control, we checked that captopril and NH2TZ, added either separately or together had no effect on the caspase activity in vitro, which is strongly inhibited by zVAD (see Fig. 5b).

Figure 5.

Caspase-3 activity in CGCs undergoing apoptosis. The effect of antioxidant system and proteasome inhibitors. (a) Cells (6 × 106/well) were incubated either in high potassium (S-K25, ○) or in low potassium (S-K5, •) serum-free culture medium in the absence or presence of the inhibitors MG132 (5 µm), i.e. No-PROTEASOME (□); CP (5 mm) + NH2-TZ (10 mm), i.e. No-AOS (bsl00036). Control cultures were switched to S-K25 medium. At different times, cultures were lysed and assayed for DEVD-pNA cleavage. The increase in protease activity was determined by the spectrophotometric detection, at 405 nm, of the cleaved chromophore p-nitroanilide (pNA). Each value represents the mean ± standard error values of DEVD-pNA cleavage, expressed as A405 value. The experiments were repeated five times. (b) Cells incubated either in high potassium (S-K25) or in low potassium (S-K5) serum-free culture medium for 8 h were lysed and assayed for DEVD-pNA cleavage in the absence or presence of either NH2-TZ (10 mm) or CP (5 mm), added alone or together, or z-VAD (100 µm).

Cytochrome c release in S-K5 in No-AOS, No-PROTEASOME and No-CASPASE cells en route to death

As ROS production causes cyt c release from mitochondria (Atlante et al. 2000), we investigated whether and how cyt c release occurs in S-K25, S-K5, No-AOS, No-PROTEASOME, No-CASPASE and No-PROTEASOME/CASPASE as well as in No-AOS/PROTEASOME and No-AOS/CASPASE cells (Fig. 6). The presence of cyt c was checked in the extramitochondrial phase both by means of polarographic measurement of the activation of cyt c-dependent ascorbate oxidation as in (Atlante et al. 1999, 2000; Bobba et al. 1999, 2002; Fig. 6a), and with immunological analysis performed with western blots (Fig. 6b). The mitochondrial fractions were also investigated immunologically with respect to their cyt c content with the glutamate dehydrogenase used as a control to normalize the amount of protein loaded onto the gel (Fig. 6b′). In each case, measurements were made 3 h after apoptosis induction, when mitochondria are still coupled (Bobba et al. 2002) and maximum of cyt c release (Bobba et al. 1999), proteasome activation (Canu et al. 2000), ROS production and antioxidant enzyme activity (see Figs 3 and 4) occur.

Figure 6.

Cytochrome c release in CGCs undergoing apoptosis. The effect of antioxidant system, proteasome and caspase inhibitors. (a) Polarografic measurement of cyt c. Rat CGCs (2 × 106/well) at 7DIV were incubated either in high potassium (S-K25) or in low potassium (S-K5) serum-free culture medium in the absence or presence of the inhibitors MG132 (5 µm), i.e. No-PROTEASOME; z-VAD (100 µm), i.e. No-CASPASE; MG132 (5 µm) + z-VAD (100 µm), i.e. No-PROTEASOME/CASPASE; CP (5 mm) + NH2-TZ (10 mm), i.e. No-AOS; CP (5 mm) + NH2-TZ (10 mm) + z-VAD (100 µm), i.e. No-AOS/CASPASE; CP (5 mm) + NH2-TZ (10 mm) + MG132 (5 µm), i.e. No-AOS/PROTEASOME; SOD (5 e.u./well), i.e. S-K5 + SOD. At different times cells were scraped, collected and homogenized. Cell homogenate (about 0.2 mg protein) was incubated at 25°C in the presence of rotenone (3 µm); antimycin (0.8 µm) and myxothiazole (6 µm) in a water-jacketed glass vessel and cyt c release was detected by means of oxygen consumption caused by externally added ascorbate (ASC, 5 mm; see Materials and methods; inset). Results, expressed as nanoatoms of oxygen/min/mg cell protein, are the means ± standard errors of triplicate measurements and are representative of at least six different experiments carried out with different cell preparations obtained from different groups of animals. (b) Western-blot analysis of cyt c. Either cytosolic (b) or respective mitochondrial (b′) fractions from either control (S-K25) or apoptotic CGCs in the absence (S-K5) or presence of the inhibitors above reported were analysed by western blotting analysis, as described in the Materials and methods. Antibodies against mitochondrial GDH were used to normalize the protein amount loaded onto the gel.

Apoptosis induction caused the release of fully functioning cyt c as revealed by the greater rate of ascorbate oxidation by cell homogenate than by S-K25 cells (see inset Fig. 6a for a typical experiment) as well as by western blots (Fig. 6b). No significant cyt c release was found in No-PROTEASOME cells as in Bobba et al. (2002) and in No-PROTEASOME/CASPASE cells, but it occurred in No-CASPASE, no-AOS and No-AOS/CASPASE cells as in S-K5 cells. Interestingly, in No-AOS/PROTEASOME cells, cyt c release was found to be about 70% that released in both No-AOS/CASPASE and S-K5 cells. As a control, the occurrence of partial prevention of cyt c release in S-K5 cells in the presence of externally added SOD was also found.

As expected, the amount of cyt c in the corresponding mitochondrial fractions (Fig. 6b′), evaluated in the same cellular preparations, was found in good agreement with respect to the relative cytosolic amount, being in an inverse interrelationship (see Bobba et al. 1999).

Discussion

Because in cell death, ROS production can trigger many processes, including the cyt c release crucial in leading cells to programmed death (Liu et al. 1996; Kluck et al. 1997; Higuchi et al. 1998; Skulachev 1998; Atlante et al. 1999), in this paper we investigate whether and how cell survival and ROS production is affected by certain CGC components, namely the cell antioxidant system (which includes SOD, catalase and the thiol oxidation state) and the proteolytic system (proteasomes and caspases), with special attention to their time-dependent activity changes as well as to their mutual inter-relationship. In order to determine the relative contribution of the antioxidant and proteolytic systems to the mechanism of cell death, we used specific inhibitors of the different component activity and monitored both cell survival and ROS production as a function of time after potassium deprivation. In this investigation we have examined how cyt c release depends on the above systems.

In agreement with Valencia and Moran (2001), this paper not only shows that AOS plays a crucial role in the death route, but it has the added dimension of showing that ROS level depends on AOS activity, which, in turn, is regulated by the proteolytic systems and regulates the cytochrome c release. Moreover, we show that ROS threshold occurs, related to apoptosis/necrosis transition. In this regard, we show that in cells in which the antioxidant system is impaired, death occurs via necrosis, as demonstrated by the occurrence of the LDH release, the absence of DNA laddering and the death sensitivity to MK801 which inhibits necrosis without affecting apoptosis (see Fig. 2).

The time course of both antioxidant and proteolytic systems merits a detailed discussion with respect to their involvement in ROS production, cyt c release and cell survival. We report the fast increase in the activity of SOD and catalase and in the thiol oxidation state up to 3 h when cells survive even though ROS production increases. This is consistent with the fact that the steady-state ROS level in CGCs derive from the cell balance between the ROS-producing and the ROS-scavenging systems, neither of them being fully established to date. Later, in the route leading to death, the antioxidant activity decreases: as we have shown that in cells with inhibited proteasome an increase in the activity of the antioxidant system occurs, then we first propose that increase in the AOS occurs in the absence of proteasome activity with a consequent ROS production decrease. Such a proposal requires that the antioxidant enzymes are subjected to both the proteasome- and caspase-dependent proteolysis. On the other hand, the capability of the proteolytic system to cause ROS production per se can be ruled out because cells in which either antioxidant system or both antioxidant system and proteasome are inhibited do not differ from each other with respect to ROS production.

Moreover, the slight but statistically significant difference found in the prevention of ROS decrease found in cells in which proteasome and caspase, separately or together, were blocked with respect to S-K25 cells, suggests that when both proteasome and caspase are active, they participate to the AOS proteolysis in a different manner, being their effects additive. In particular, because AOS activity remains high in cells without proteasome activity in the 3- to 15-h time range, when caspases are active, we conclude that the caspase-dependent AOS degradation requires previous proteasome activity.

We investigated the caspase-3 activity as a function of time. It was found to increase in apoptosis, with the maximum found at 8 h. Interestingly, caspase-3 activation does not occur in cells with no proteasome activity. Because under these conditions no cyt c release takes place (see Fig. 6), we suggest that this is the cause of the lack of caspase-3 activation. It should be considered, in fact, that cyt c release was shown to trigger caspase cascade (Kluck et al. 1997; Li et al. 1997). Surprisingly, in No-AOS we have no caspase-3 activation, nonetheless cells die. This occurs owing the apoptosis/necrosis shift which takes place when ROS production is high, not regulated by cell AOS.

The picture emerging from the above findings with respect to the cyt c release is the following: at 3 h maximum cyt c release occurs (Bobba et al. 1999), no caspase activity is required in this release (Bobba et al. 1999; Fig. 5), and ROS production is regulated by the antioxidant systems modulated by the proteasome. Figures 4–6 show that the extra-AOS activity in cells without proteasome is found in cells in which cyt c release does not occur, probably preventing the caspase cascade. Given that cyt c release, in cells in which the antioxidant system is impaired, is reduced by preventing proteasome activity, we suggest that cyt c release depends also on other protein(s) that are sensitive to the increased proteasome activity. The caspase cascade (Budihardjo et al. 1999; Lee and Wei 2000) contributes to AOS degradation, thus increasing ROS production and leading to programmed cell death.

In this paper, we propose a possible apoptosis time-course in CGCs. In our scheme (Fig. 7) we consider that, following the new mRNA and protein synthesis which occurs upstream ROS production (Schulz et al. 1996), apoptosis can occur in two separate phases: an EARLY-phase ranging from 0 to 3 h after the apoptotic stimulus and a LATER-phase ranging between 3 and 15 h. In the EARLY-phase, when ROS production increases (Fig. 3), the cell components, including AOS and proteasomes, are evoked to maintain sufficient ROS production for the correct application of the programmed death (Fig. 4 and Canu et al. 2000). The caspases have no role in this phase (Fig. 5). In this situation, cyt c release occurs from still coupled (Bobba et al. 2002) mitochondria (Fig. 6), possibly in order to begin the activation of the caspase cascade (Li et al. 1997; Zou et al. 1997, 1999; Cai et al. 1998; Du et al. 2000) as well as to drive ATP synthesis (Waterhouse et al. 2001; Wigdal et al. 2002). In the LATER-phase proteasome activity decreases slowly (Canu et al. 2000; Bobba et al. 2002), but caspase-3 activity increases (Fig. 5). This causes a further reduction in AOS enzyme activity (Fig. 4) with cyt c degradation (Bobba et al. 1999). As a result of this, ROS levels are still high (Fig. 3) and the cells are committed to death (Figs 1 and 2). However, when the antioxidant system is totally blocked, necrosis occurs (Figs 1 and 2).

Figure 7.

The apoptosis time course. The time course of apoptosis is described as occurring in two phases: (1) EARLY-phase (0–3 h), (1.1) ROS production (Fig. 3; Greenlund et al. 1995; Schulz et al. 1996; Valencia and Moran 2001); (1.2) cytochrome c release from mitochondria (Fig. 6; Liu et al. 1996; Kluck et al. 1997; Higuchi et al. 1998; Skulachev 1998; Atlante et al. 1999, 2000, 2001; Bobba et al. 1999); (1.3) antioxidant system activity activation (Fig. 4); (1.4) proteasome activity activation (see Canu et al. 2000; Bobba et al. 2002). (2) LATER-phase (3–8 h), (2.1) steady-state ROS level (Fig. 3); (2.2) cytochrome c-dependent caspase activation (see Li et al. 1997; Cai et al. 1998; Zou et al. 1999; Du et al. 2000); (2.3) proteasome/caspase-dependent antioxidant system degradation (Fig. 4); (2.4) caspase-dependent cytochrome c degradation (Bobba et al. 1999); (2.5) caspase-dependent proteasome degradation (see Canu et al. 2000). For further details see the text.

Acknowledgements

This work was partially financed by PRIN ‘Metabolismo Energetico e Trasporto Mitocondriale nella Morte Cellulare ed in altre Condizioni Fisio-Patologiche’ (MURST), Fondi di Ricerca di Ateneo del Molise to SP, and Ministero della Salute grant ‘Alzheimer Project’ to PC. The authors thank Mr Richard Lusardi for linguistic consultation.

Ancillary