Cyclosporin A Rescues Thymocytes from Apoptosis Induced by Very Low Concentrations of Thapsigargin: Effects on Mitochondrial Function

EXPERIMENTAL CELL RESEARCH ARTICLE NO.

227, 264–276 (1996)

0276

Cyclosporin A Rescues Thymocytes from Apoptosis Induced by Very Low Concentrations of Thapsigargin: Effects on Mitochondrial Function PAUL WARING1

JOANNE BEAVER2

AND

John Curtin School of Medical Research, P.O. Box 334, Canberra City, Canberra, ACT 2601, Australia

Raising intracellular calcium levels can induce apoptosis or programmed cell death in many cells. While early rises in intracellular calcium are not universally associated with apoptotic cell death, calcium clearly plays a key role in many of the biochemical events which occur during apoptosis. In this paper we have determined intracellular calcium rises induced by 2, 10, and 100 nM thapsigargin in mouse thymocytes. These concentrations cause increases in cytosolic calcium of 100–250, 400–600, and ú1000 nM, respectively. These rises are sustained for at least 85 min and the ratio between the maximum rise caused by 10 nM compared to 2 nM thapsigargin is 2.1 { 0.4 (n Å 6). Both 2 and 10 nM thapsigargin cause apoptosis at 24 h as shown by DNA fragmentation and morphology when examined by electron microscopy. Cyclosporin A (CsA) inhibits apoptosis caused by 2 nM thapsigargin but not that caused by 10 nM thapsigargin. Electron microscopy of thymocytes treated with 2 nM thapsigargin at 24 h shows intact mitochondria although with altered morphology. There is no loss of ATP or decrease in the ATP/ADP ratio in these cells over 12 h. Mitochondria in cells treated with 10 nM thapsigargin, however, are swollen by 6 h and many are lost by 24 h. These cells show greatly diminished ATP content by 12 h and a decrease in ATP/ADP ratio. Examination of the effects of PMA, an activator of the plasma membrane calcium ATPase pump, on cells treated with 10 nM thapsigargin suggests that two pools of calcium may be responsible for the differential effects of the two calcium levels in the cells. Probing of the mitochondrial membrane potential (MMP) by rhodamine 123 staining of live cells shows that the collapse of the MMP caused by 10 nM thapsigargin is unaffected by CsA. The MMP is also reduced in cells treated with 2 nM thapsigargin but this is restored by CsA. Cells are also rescued from apoptosis caused by 2 nM thapsigargin by incubation with FK506. This immunosuppressive agent has no ef1 To whom correspondence and reprint requests should be addressed at Division of Cell Biology, John Curtin School of Medical Research, P.O. Box 334, Canberra City, ACT 2601, Australia. E-mail: [email protected]. 2 Current address: Biomedical Science, Faculty of Applied Science, University of Canberra, P.O. Box 1, Belconnen, ACT 2616, Australia.

ECR 3248

/

6i12$$$461

INTRODUCTION

Thapsigargin is a sesquiterpene which is a potent inhibitor of the endoplasmic reticulum (ER) calcium ATPase pump [1]. Treatment of cells with thapsigargin results in a sustained rise in intracellular calcium, presumably because of the constant leakage from calcium stores associated with the ER, and all effects of thapsigargin are thought to be related to its ability to raise intracellular calcium levels [2, 3]. This has made thapsigargin a useful tool in the study of controlled intracellular calcium changes and the subsequent effects on cells. Apoptosis is a well-defined mode of cell death with definitive biochemical and morphological features [4, 5]. The role of calcium in the induction of apoptosis has been the focus of numerous studies. Evidence has been presented for intracellular calcium rises during activation-induced cell death [6], glucocorticosteroidinduced apoptotic cell death [7], and apoptosis induced by toxins [8]. In contrast, a number of studies have shown that calcium is by no means universally associated with apoptotic cell death [9, 10]. We have shown that glucocorticosteroid-induced apoptosis and gliotoxin-induced apoptosis in 10-day-old mouse thymocytes are not associated with any early intracellular calcium rises [11]. We have also shown that while thapsigargin does result in apoptosis in murine thymocytes, the kinetics are different from that observed with glucocorticosteroid-induced apoptosis

264

0014-4827/96 $18.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.

AID

fect on the membrane permeability transition induced in isolated mitochondria. These results suggest that very low rises in intracellular calcium in thymocytes cause activation-induced cell death inhibited by CsA and FK506 and are without effect on ATP levels and therefore do not involve irreversible mitochondrial damage. Exceeding these calcium levels by only twofold results in apoptosis accompanied by reduced ATP levels and mitochondrial damage, although apoptotic cell death in this instance is unaffected by the classic inhibitor of mitochondrial permeability transition, CsA. q 1996 Academic Press, Inc.

09-03-96 10:12:58

eca

AP: Exp Cell

EFFECTS OF CYCLOSPORIN A ON THAPSIGARGIN-INDUCED APOPTOSIS

[12]. The latter results in significant apoptosis at 6 h while thapsigargin-induced apoptosis occurs over 18 – 24 h, suggesting a different mechanism. During investigation of thapsigargin-induced apoptosis, we observed mitochondrial swelling in P815 cells and thymocytes treated with ú10 nM thapsigargin (11, Beaver and Waring, submitted for publication). In the presence of calcium and phosphate ions, isolated mitochondria undergo a so-called permeability transition in the membrane, resulting in the loss of small molecules such as glutathione and calcium [13–15]. This transition and the subsequent dysfunction of mitochondria have been shown to be inhibited by cyclosporin A (CsA)3 [16–18]. CsA has also been shown to inhibit activation-induced cell death (AICD) in T cells presumably by affecting the repertoire of cytokines expressed during AICD [19–21]. Because of this we undertook an examination of the effect of CsA on apoptosis induced by thapsigargin in murine thymocytes. Our results demonstrate that the two pathways to apoptosis induced by the same calcium mobilizing agent in the same cell type depend on the relative magnitude of the calcium rise. EXPERIMENTAL PROCEDURES Materials. Thymocytes were harvested from 10-day-old mice in F15 media with 5% FCS, passed through a stainless steel mesh, filtered through Nybolt nylon monofilament gauze (60 mm) and centrifuged at 300g for 5 min at 07C. Cells were suspended in fresh F15 media with 5% FCS at 1 1 106/ml unless otherwise stated. Cells were incubated in 24-well plates or in culture flasks at 377C and 5% CO2 . PMA, ATP, ADP, propidium iodide, RNAase dexamethasone, and INDO-1 were obtained from Sigma. Cyclosporin A was provided by Sandoz, Basel, Switzerland and kept as 1 mM solutions in ethanol at 0207C and diluted as required. Thapsigargin was supplied from Calbiochem/Novabiochem (San Diego, CA) and kept as 3 mM solutions in DMSO at 0207C and diluted as required. Dilute (õ3 mM) solutions of thapsigargin lost activity even when stored at 0207C and were discarded after a single use. Anti-CD3 (from the 145-2C11 clone) was a kind gift of Dr. P. Hodgkin, Centenary Institute, Sydney, Australia. The FK506 was a kind gift from the laboratory of Dr. A. Dulhunty, John Curtin School of Medical Research. All other reagents were the purest available. Measurement of apoptosis. Apoptosis was estimated by DNA content using propidium iodide staining of fixed cells and electron microscopy. Cells were treated at 1 1 106/ml in 2 ml of F15 with 5% FCS, pelleted, and fixed in 70% ethanol overnight. Cells were then washed three times in PBS, suspended in 1 ml of PBS, and treated with propidum iodide (4 mg/ml) and RNAase (200 mg/ml) for at least 30 min before analysis by flow cytometry. The subdiploid population

3 Abbreviations used: CsA, cyclosporin A; ATP, adenosine 5*-triphosphate; ADP, adenosine 5*-diphosphate; AICD, activation-induced cell death; FCS, fetal calf serum; MMP, mitochondrial membrane potential; ER, endoplasmic reticulum; PBS, phophate-buffered saline; PMA, phorbol-12-myristate-13 acetate; DMSO, dimethylsulfoxide; FACS, fluorescence-activated cell sorting; HPLC, high-performance liquid chromatography; EGTA, (ethylene-dioxy) diethylenedinirilo-tetraacetic acid; PKC, protein kinase C.

AID

ECR 3248

/

6i12$$$462

09-03-96 10:12:58

265

of cells was taken as the fraction of apoptotic cells. All data are the mean of three to six separate incubations. For electron microscopy, cells were pelleted after treatment and washed once in PBS, pelleted again, and fixed with 2% glutaraldehyde in 0.1 M sodium cacodylate (pH 7.4) for 2 h. The preparations were postfixed in osmium tetroxide for 1.5 h and en bloc stained in 2% uranyl acetate for 1 h. They were then dehydrated in ethanol and embedded in Spurrs resin. Sections were cut and viewed on a Phillips 301 electron microscope. Cells were scored for apoptosis based on morphological evidence of condensed chromatin as seen in Fig. 2. Magnification was adjusted to give 50–60 cells per field and at least four separate fields were counted for each determination. Calcium measurement. Intracellular calcium levels were measured using the INDO-1 method. Thymocytes were loaded with INDO-1 (0.3 mg/ml) at 377C at 2 1 107/ml in 400 ml for 15–20 min in F15 with 1% FCS. After being loaded with dye, 100 ml of the cells were diluted into 1.9 ml of F15 containing 1% FCS followed by immediate analysis using a Becton Dickinson FACStar instrument. Cells were kept at 377C during the course of data collection. Data were collected at the rate of 150–200 cells/s and 300,000–800,000 events were collected. Data were collected for 30–80 min depending on the experiment. Indo-1 was excited with UV light at 351–363 nm. Fluorescence at 405 and 485 nm and the 405/485 ratio were monitored. The 405/485 ratio (R) was set to a value of 1 at channel 100 before each run. The ratio was monitored for exactly 5 min before rapid (õ10 s) addition of thapsigargin. PMA was included in the final solution or was added after a predetermined time. Vanadate was added before thapsigargin addition. Ratios were determined at various time points including 1–2 min from the start of a run and were referred back to time 0, R Å 1 for calculation of each ratio. Values of Rmax and Rmin and Sf2/Sb2 were calculated according to the method described in Ref. [22] for each experiment. Intracellular calcium changes were calculated using [Ca/2]i Å Kd 1 {[R 0 Rmin]/ [Rmax 0 R]} 1 Sf2/Sb2. [22]. Values of Rmax were typically 4.5–4.9 and were obtained by addition of ionomycin. Rmin was 0.12–0.14 and Sf2/Sb2 was 1.1 to 1.4; these values were calculated using bulk fluorometry as recommended [22]. A value of 250 nM was used for the Kd . Mitochondrial membrane potential. For rhodamine 123 staining [23], after treatment 2 1 106 cells were pelleted and suspended in 2 ml PBS containing rhodamine 123 at 10 mg/ml for 30 min at room temperature. Cells were then washed by being suspended in 2 ml of PBS and being left at room temperature for 30 min. This washing was repeated a total of three times. Complete washing is essential to reduce background fluorescence due to cytosolic dye. Fluorescence was collected in fluorescence detector 1 (Fl1), with a filter 430/30 nm bandpass on a Becton Dickinson FACScan. All flow cytometry data were analyzed using WinMDI software kindly provided by Joseph Trotter, Salk Institute, La Jolla, California. Measurement of ATP and ADP. ATP and ADP were determined using HPLC. Cells at 1 1 106/ml in 10–20 ml of F15 with 5% FCS were pelleted after treatment and washed quickly in cold PBS. Cells were suspended in 250 ml PBS and 25 ml taken for protein determination and/or used for measuring apoptosis. Cells were then lysed by the addition of 250 ml of 20% trichlororacetic acid (TCA) and left on ice for 10 min. Precipitated protein was removed by centrifugation and TCA removed by extracting three times with water-saturated ether. Nucleotides were separated on a Waters SAX strong anion exchange column on a Beckman System Gold HPLC instrument. Elution was by a 0–85% convex gradient of 500 mM KH2PO4 , pH 5.0, at 1 ml/min. The initial solvent was 10 mM KH2PO4 , pH 6.6. Detection was at 260 nm. Typical retention times for ADP and ATP were 36 and 56 min, respectively. Standards were run following the analysis of unknowns. Protein was determined using the Bio-Rad protein assay method. Between three and five separate incubations

eca

AP: Exp Cell

266

WARING AND BEAVER

FIG. 1. The effect of cyclosporin A (CsA) on apoptosis induced in thymocytes at 24 h by 2 and 10 nM thapsigargin. (A) Untreated cells; (B) 10 nM thapsigargin; (C) 10 nM thapsigargin / 50 nM CsA; (D) 2 nM thapsigargin; (E), 2 nM thapsigargin / 50 nM CsA. M1 indicates the subdiploid apoptotic cells. The X axis is a measure of DNA content. Most cells in the untreated population are in G1 or G0 as expected of unstimulated thymocytes.

for each time point and treatment were made, and the data are the means of each determination. All errors are standard deviations. Immunoprecipitation. Thymocytes were washed three times in phosphate-free Krebs–Ringer bicarbonate and resuspended at 2 1 107/ml. Carrier-free [32P]phosphate was added (20 mCi/2 1 106 cells) and the cells were incubated at 377C for 2 h. Cells were then suspended in complete medium at 2 1 106/ml and either treated with 1 mM PMA for 1–3 min or left untreated. Cells were then pelleted and the supernatant was discarded. Cells were lysed in 300 ml RIPA buffer (PBS, 1% NP40, 0.5% sodium deoxycholate, 0.1% SDS) containing PMSF (100 mg/ml), aprotinin (60 mg/ml), and sodium orthovanadate (1 mM) on ice for 30 min. The lysate was precleared with agarose–protein G conjugate and the Ca2/-ATPase was precipitated with antibody to human plasma membrane Ca2/-ATPase (5F10 clone, Sigma) followed by agarose–protein G. The pellet was washed three time with ice-cold RIPA buffer with inhibitors, treated with standard electrophoresis sample buffer, and run in 7.5% polyacrylamide gels using Bio-Rad equipment. Autoradiography was carried out using Kodak X-OMAT.

a small increase in apoptosis (Table 1). Higher concentrations of CsA had the same effect. When estimated using flow cytometry, no more than 50% apoptosis was observed induced in thymocytes by 2 nM thapsigargin. TABLE 1 Apoptosis Induced in Thymocytes by Thapsigargin Treatment by FACSb 1

2

RESULTS

Thapsigargin at 2 nM and 10 nM induces apoptosis in thymocytes differentially inhibited by CsA. Figure 1 shows the effects of CsA on apoptosis induced by 2 and 10 nM thapsigargin over 24 h. We initially measured apoptosis using propidium iodide staining of fixed cells [24]. In this method the loss of small DNA fragments due to apoptosis is revealed as a subdiploid population of cells. There were two clearly separable effects of CsA. At 2 nM thapsigargin, CsA significantly inhibited apoptosis at 24 h. Cyclosporin A at 50 nM was found to be the optimal concentration for inhibition. Cells were preincubated for 15 min before thapsigargin treatment. Conversely, cyclosporin A at 50 nM did not inhibit apoptosis induced by 10 nM thapsigargin and in many experiments we consistently observed

AID

ECR 3248

/

6i12$$$462

09-03-96 10:12:58

by FACSb 3

by EMc 4

% Apoptosisa

Control 2 nM 2 nM / 50 nM CsA 10 nM 10 nM / 50 nM CsA

19.6 47.0 25.5 88 96

{ { { { {

0.8 1.5 2.2 0.8 1

Control 2 nM 2 nM / 50 nM CsA 2 nM / 50 nM CsA / 1 uM PMA 10 nM 10 nM / 1 mM PMA 10 nM / 50 nM CsA / 1 uM PMA

22.2 61.8 29.6 75.2 63.1 51.1 69.6

{ { { { { { {

1 4 5 2 0.5 1.5 1.4

Control 2 nM 2 nM / 50 nM FK506

21 { 1 60 { 2.5 34 { 4

Control 2 nM 2 nM / 50 nM CsA 10 nM 10 nM / 50 nM CsA

14 { 7 94 { 3 46 { 7 ú95 ú95

a

Estimated at 25 h. Estimated using propidium iodide staining of fixed cells, mean of four separate determinations. c Estimated by scoring for morphology of sections using electron microscopy. b

eca

AP: Exp Cell

EFFECTS OF CYCLOSPORIN A ON THAPSIGARGIN-INDUCED APOPTOSIS

267

FIG. 2. Morphology of thymocytes treated with thapsigargin for 24 h. (A) Untreated; (B) 2 nM; (C) 2 nM / 50 nM CsA; (D) 10 nM; (E) 10 nM / 50 nM CsA; (F) 10 nM at 6 h. Original magnification, approximately 1550.

Estimation by morphological criteria, however, showed in fact that almost all cells had the features of apoptosis when treated with either 2 or 10 nM thapsigargin (Fig. 2) and analysis by flow cytometry consistently underestimated the proportion of apoptotic cells, particularly with 2 nM thapsigargin. Again, cyclosporin A resulted in significant protection with 2 nM thapsigargin (Fig. 2C and Table 1) but provided no protection with 10 nM thapsigargin. Cells treated with 10 nM thapsigargin at 24 h showed extensive secondary necrosis while cells treated with 2 nM thapsigargin had most organelles intact (Fig. 2). Cells treated with 10 nM thapsigargin (Fig. 2F) or 2 nM thapsigargin (not shown) show no apoptosis at 6 h. This is consistent with our earlier observations on the kinetics of apoptosis induced in thymocytes by thapsigargin [12]. Because we examined the effects of PMA on calcium levels (see below), Table 1 also includes data for cells treated with 1 mM PMA. The presence of PMA abolishes the protective effect of CsA on cells treated with 2 nM thapsigargin. There also appeared to be a slight protective effect of PMA on cells treated with 10 nM thapsigargin alone but this

AID

ECR 3248

/

6i12$$$463

09-03-96 10:12:58

again is lost following simultaneous treatment with CsA. Incubation of cells with FK506 also provided protection against apoptosis caused by 2 nM thapsigargin (Table 1). Cells treated with 2 nM thapsigargin do not have swollen mitochondria. Cells treated with 10 nM thapsigargin showed swollen mitochondria at 6 h when examined under higher magnification (Fig. 3C, compare with normal mitochondria in Fig. 3A). We have already shown that thapsigargin can result in swollen mitochondria in thymocytes [11] and that treatment with other agents which can inhibit the endoplasmic reticulum calcium ATPase pump will also result in swollen mitochondria and apoptosis (Beaver and Waring, submitted for publication). Almost all cells treated with 10 nM thapsigargin had at least some mitochondria swollen typical of Fig. 3C. Not all mitochondria in treated cells were swollen, however. Swelling was typified by a three- to fourfold increase in apparent volume and dissolution of the fine structure of the organelles. These results are typical of mitochondrial swelling seen in isolated organelles when exposed to calcium and

eca

AP: Exp Cell

268

WARING AND BEAVER

FIG. 3. Swollen mitochondria in thymocytes caused by 10 but not 2 nM thapsigargin. (A) Mitochondria in normal cells at 24 h; (B) mitochondria following treatment of cells with 2 nM thapsigargin for 24 h. Note the expanded matrix space typical of activated cells; (C) mitochondria in cells treated with 10 nM thapsigargin for 6 h. Organelles are swollen two to three times their normal size and internal detail is lost. Very few mitochondria could be detected under these conditions by 24 h. Original magnification, approximately 17000.

phosphate and are thought to be due to osmotic effects following permeability changes to the mitochondrial membrane [13–18]. No swollen mitochondria were observed in cells treated with 2 nM thapsigargin at 6 h (not shown) or at 24 h (Fig. 3B). By 24 h mitochondria in cells treated with 10 nM thapsigargin were difficult to find. Mitochondria in cells treated with 2 nM thapsigargin at 24 h did show changes typical of activated cells (Fig. 3B). These changes include an increase in matrix volume and changes in the appearance of mitochondrial cristae. Cells activated for hormone production, for example, have been described as having mitochondria similar to those seen in Fig. 3B [25]. Intracellular calcium increases caused by 2, 10, and 100 nM thapsigargin and comparison with anti-CD3. All concentrations of thapsigargin tested caused a sustained increase in intracellular calcium following a lag period of 2–3 min (Fig. 4). In addition, all concentrations caused greater than 90% of all thymocytes to respond with an increase in intracellular calcium levels. Thapsigargin at 2 nM resulted in a maximum increase of 150–250 nM calcium while 10 nM consistently produced an increase approximately twice this of 400–600 nM. The ratio of the increased calcium level achieved with 10 nM thapsigargin to that achieved by 2 nM thapsigargin at 30 min was 2.1 { 0.4 (n Å 6). Each experiment described in the figures was carried out on cells from a single animal on the same day. In normal media ([Ca2/]: 1.3 mM) the increased calcium levels were sustained for at least 85 min. All results are typical of six separate experiments. We also measured the effects of 100 nM thapsigargin on calcium levels. This resulted in changes in intracellular calcium of greater than 1000 nM. Thapsigargin at 100 nM gave apparently less apoptosis than it did at 10 nM but this is typical of a switch to necrotic cell death by 100 nM thapsigargin seen with many toxins when a threshold

AID

ECR 3248

/

6i12$$$463

09-03-96 10:12:58

concentration is exceeded [11; and data not shown]. We used calcium-depleted media (õ120 mM) in the presence and absence of increased EGTA (0–100 mM) to examine the effects of low extracellular calcium on calcium levels inside the cell. Using 50–100 mM EGTA and calcium-depleted media, this generally resulted in the expected effect of both a lower peak response and a diminution of the response with time with all concentrations of thapsigargin (data not shown). This is expected as maintenance of the sustained calcium levels requires entry of external calcium stimulated by depletion of the ER calcium stores [26]. More importantly, in calcium-depleted media alone (no EGTA) we consistently observed two populations of cells corresponding to high and low calcium rises when cells were treated with 10 nM thapsigargin (Fig. 5). These populations

FIG. 4. Increases in intracellular calcium in thymocytes treated with thapsigargin. Open squares, 2 nM; closed squares, 10 nM; open triangles, 100 nM. Results are typical of at least six separate experiments. The increase in calcium was sustained for at least 85 min, the longest time recorded in these experiments.

eca

AP: Exp Cell

EFFECTS OF CYCLOSPORIN A ON THAPSIGARGIN-INDUCED APOPTOSIS

269

FIG. 5. Two populations of cells with increased calcium are observed in calcium-depleted media. (A) Calcium flux caused by 10 nM thapsigargin in calcium-depleted media. (B) 10 nM in normal media; (C) 2 nM in normal media. The arrows indicate the low and high calcium populations. The time axis extends over 40 min while the calcium axis indicates a change in the 405/485 ratio as an indication of calcium increase.

corresponded closely to those seen with 10 and 2 nM thapsigargin in normal media. Cyclosporin A had no effect on calcium increases caused by 2 and 10 nM thapsigargin (data not shown). Using calcium-depleted media with or without EGTA, we were unable to inhibit apoptosis caused by thapsigargin. This is consistent with reports that low extracellular calcium and EGTA will themselves induce apoptosis [27] and is testament to the complicated relationship between apoptosis and calcium. We also measured calcium rises induced in thymocytes with anti-CD3 antibody. Engagement of CD3 has been shown to induce calcium rises and result in apoptosis in thymocytes and T-cell clones [6]. Although only just under 10% of the cells actually showed a calcium increase, those cells which responded showed an increase comparable to that seen with 2 nM thapsigargin. These experiments were performed in solution and low numbers of responding cells may be due to lack of sufficient cell surface cross linking and/or accessory cells [28]. Addition of further anti-CD3 20 min after the addition of 2 nM thapsigargin caused no further rise in calcium, indicating that the same pools of calcium were mobilized (Fig. 6). In contrast, incremental addition of thapsigargin at concentrations of 2 nM and higher caused a corresponding incremental calcium increase as would be expected from progressive inhibition of the ER calcium ATPase pump (data not shown). PMA treatment results in abrogation of the high calcium population of cells and phosphorylation of the plasma membrane Ca2/ ATPase pump. PMA is an activator of protein kinase C [29], which itself results in varied pleiotropic effects within the cell, including inducing apoptosis in thymocytes in its own right [30]. A further consequence of activation of PKC is activation of the plasma membrane calcium ATPase pump in

AID

ECR 3248

/

6i12$$$463

09-03-96 10:12:58

erythrocytes and the subsequent decrease in internal calcium levels in these cells [31]. Because this phorbol ester has been shown to have varying effects on apoptosis, we examined the effect of PMA at 1 mM on calcium levels following thapsigargin treatment. Initially we added PMA following attainment of the sustained calcium rise induced by 10 nM thapsigargin. Figure 7 shows that this rapidly led to a fall in intracellular calcium levels approaching those observed in the presence of 2 nM thapsigargin. Preincubation with PMA for 15 min prior to addition of thapsigargin led to a decrease in calcium rises to levels produced by 2 nM thapsigargin alone. PMA at 0.1 mM had a moderate effect on calcium levels, reducing them to about 75% of that caused by 10 nM thapsigargin. The maximum reduction in calcium levels occurred at 0.2 to 1 mM PMA. Increasing PMA to 3 mM has no further effect. PMA at all concentrations had no effect on 2 nM thapsigargin-induced calcium rises. In addition PMA had no effect on calcium rises induced by 100 nM thapsigargin. The addition of orthovanadate, a known inhibitor of the calcium plasma membrane pump [32], reverses the effect of PMA on calcium levels (Fig. 8). This is consistent with the effects on calcium being modulated by activation of the plasma membrane pump [31]. Activation of the pump was confirmed by the demonstration that PMA treatment results in increased phosphorylation of the plasma membrane pump (Fig. 9). Effect of thapsigargin on ATP and ADP levels in thymocytes. Apoptosis is thought to be an energy-requiring process and it was thought to proceed without overt mitochondrial damage; early work demonstrated that no gross mitochondrial changes occur during apoptosis [33]. However, it has been reported that agents which have a direct effect on mitochondrial function will induce apoptosis [34]. Table 2 shows the results of an

eca

AP: Exp Cell

270

WARING AND BEAVER

FIG. 6. Thapsigargin at 2 nM mobilizes the same pool of calcium as anti-CD3 treatment of thymocytes. (A) Cells treated with anti-CD3 show an increase followed by a decline over 10–15 min. (B) Cells treated with 2 nM thapsigargin show a sustained increase in calcium with a peak slightly greater than anti-CD3. Addition of anti-CD3 at 20 min produces no further increase. Addition of further thapsigargin at this point causes a corresponding incremental increase in calcium (not shown), indicating that calcium is still available in thapsigarginsensitive stores.

extensive study of ATP and ADP levels in thymocytes treated with thapsigargin at 2, 10, and 100 nM. This confirms that there is a time-dependent decrease in ATP levels when cells are exposed to 10 and 100 nM thapsigargin. Using the ATP/ADP ratio as a measure of mitochondrial activity, it is also seen that this parameter is reduced by both 10 and 100 nM thapsigargin. We also observed a large increase in ADP levels in cells treated with 100 nM thapsigargin as early as 15 min while only a small increase in ADP is observed for 10 nM thapsigargin. This clearly establishes that ATP levels can drop in cells prior to undergoing apoptosis and that this is consistent with the early gross swelling of mitochondria seen in cells treated with 10 nM thapsigargin at 6 h. When cells were treated with 2 nM thapsigargin there was no decrease in ATP levels at 12 h. No changes in ATP were seen at earlier time points for 2 nM thapsigargin (not shown). We observed either an increase in ATP levels or an increase in the ATP/ADP ratio in cells treated with 2 nM thapsigargin (Table 2). Because PMA reduces the calcium levels in cells treated with 10 but not 100 nM thapsigargin, we measured the effect of PMA on ATP levels. It can be seen that 1 mM PMA results in a significant partial restoration of ATP levels in cells treated with 10 but not 100 nM thapsigargin correlating with the effect of PMA on calcium levels. This suggests that mitochondrial dysfunction resulting in ATP loss only occurs when intracellular calcium levels exceed 100–250 nM. Because of the general interest in mitochondrial function and apoptosis, we also measured ATP levels in thymocytes treated with dexamethasone—the archetypal model for in vitro apoptosis. Figure 10A shows that ATP levels fall rapidly 3–4 h after dexamethasone treatment. However, we detect DNA fragmentation at 3 h but no change in ATP at this time. Allowing for the fall in ATP in untreated

AID

ECR 3248

/

6i12$$$463

09-03-96 10:12:58

cells in culture (Fig. 10B), these data are consistent with loss of ATP either coincident with or following DNA fragmentation. Given that internucleosomal DNA fragmentation in apoptosis is itself preceded by 50- to 200-kb fragmentation, these data show that decreased ATP levels are not an early event in dexamethasoneinduced apoptosis. Effects of PMA and Cyclosporin A on mitochondrial membrane potential changes induced by 2 and 10 nM thapsigargin. Rhodamine 123 was used to probe the mitochondrial membrane potential (MMP) in cells treated with thapsigargin. Mitochondria in normal cells will accumulate this dye as a consequence of the membrane electrochemical gradient which is sustained by oxidative phophorylation. This results in a high percentage of brightly staining cells when untreated (Fig. 11A). Treatment with 10 nM thapsigargin results in a dramatic loss of MMP consistent with the effects of this concentration of thapsigargin on mitochondrial size and ATP content (Fig. 11B). Incubation of cells with PMA partially restores the MMP (Fig. 11C). Although 2 nM thapsigargin has no effect of nucleotides and does not result in overtly swollen mitochondria, it does have some apparent effect on the mitochondrial membrane potential at 24 h (Fig. 11D), which is not significantly affected by PMA (not shown). However, R123 staining of cells treated with 2 nM thapsigargin resembles that in cells treated with 10 nM thapsigargin and PMA, again suggesting that damage to mitochondria resulting in irreversible swelling occurs only when the calcium levels exceed 100–250 nM (compare Figs. 11C and 11D). Table 3 details the significant protective effect of PMA on MMP at 20 h. Some effect on MMP is seen as early as 3 h. When thymocytes are treated with thapsigargin at 2 or 10 nM together with cyclosporin A and examined for R123 staining a different picture emerges (Fig. 12

eca

AP: Exp Cell

EFFECTS OF CYCLOSPORIN A ON THAPSIGARGIN-INDUCED APOPTOSIS

271

FIG. 8. The effect of sodium orthovanadate at 500 mM on lowering of intracellular calcium by 1 mM PMA in thymocytes treated with 10 nM thapsigargin. Open squares, 10 nM thapsigargin. Open triangles, 10 nM thapsigargin / 1 mM PMA. Closed triangles, 10 nM thapsigargin / 1 mM PMA / 500 mM sodium orthovanadate.

FIG. 7. The presence of PMA reduces calcium levels in thymocytes treated with 10 nM thapsigargin but not 2 or 100 nM. (A) Addition of PMA when calcium levels have peaked reduces the effects of 10 nM (open squares) but not 100 nM thapsigargin (closed squares). Arrow indicates the time of addition of PMA. (B) Preincubation of cells with PMA for 1–3 min results in the reduction of calcium levels in cells treated with 10 nM thapsigargin (triangles) to that caused by 2 nM thapsigargin (squares). Closed squares are cells with 10 nM thapsigargin alone. PMA had no effect of calcium levels induced by 2 nM (data not shown).

cell death. Raised calcium levels may result in activation of the calcium-dependent nucleases responsible for apoptosis [36] but since apoptosis in thymocytes can occur in the absence of detectable calcium rises [11] and raised calcium levels do not result in apoptosis over the same time frame [11], there must be alternative explanations. Mitochondrial function in isolated organelles has been shown to be compromised by raised intracellular calcium levels initially

and Table 3). CsA at 50 nM enhances the collapse of the MMP and correlates with its effects on apoptosis caused by 10 nM thapsigargin. In contrast and similarly correlating with its effect on apoptosis, CsA protects thymocytes from loss of MMP when treated with 2 nM thapsigargin. DISCUSSION

Raised calcium levels in cells can potentially result in a number of diverse effects on cellular function including activation of proteases, lipases, and nucleases [35]. All of these actions can contribute to

AID

ECR 3248

/

6i12$$$463

09-03-96 10:12:58

FIG. 9. Phosphorylation of murine plasma membrane Ca2/ ATPase by PMA. Lane A, untreated thymocytes. Lane B, thymocytes treated with 1 mM PMA for 1–3 min. Thymocytes were prelabeled with [32P]phosphate for 2 h. Standard molecular weight markers are shown at 220 and 97.4 kDa. The protein at 144 kDa was immunoprecipitated using anti-human plasma membrane calcium ATPase and is the correct molecular weight for this protein.

eca

AP: Exp Cell

272

WARING AND BEAVER

TABLE 2 ATP and ADP Levels in Thymocytes Treated with Thapsigargin Treatment

Time

ADP a

ATP/ADP b

1

Control 10 nM 100 nM

2

Control 10 nM 100 nM

3

Control 10 nM 100 nM

4

Control 2 nM 10 nM 100 nM

12 12 12 12

5

Control 2 nM

12 h 12 h

14.1 { 0.9 13.6 { 0.9

0.75 { 0.2 0.55 { 0.04

16.5 { 3 20.6 { 0.9

6

Control 10 nM 10 nM / 1 mM PMA 100 nM 100 nM / 1 mM PMA

12 12 12 12 12

25.7 11.3 16.8 8.47 5.47

2.26 1.58 2.56 6.18 3.6

11.8 7.15 6.5 1.37 1.52

a b

15 min 15 min 15 min

ATP a 18.2 { 0.2 16.0 { 0.4 12.2 { 0.9

0.77 { 0.18 1.13 { 0.08 2.24 { 0.13

19.8 { 5 11.9 { 1.2 4.57 { 0.6

1h 1h 1h

15.2 { 0.08 13.3 { 0.3 10.4 { 0.6

0.92 { 0.1 2.86 { 0.3 6.73 { 0.15

16.5 { 0.3 4.65 { 0.2 1.6 { 0.2

4h 4h 4h

23.5 { 3.5 18.8 { 2 17.8 { 3.3

0.59 { 0.1 1.3 { 0.5 12.7 { 2.4

39 { 3 14 { 2 1.4 { 0.2

15.1 19.9 9.9 7.65

0.90 1.5 1.28 9.47

h h h h

h h h h h

{ { { {

{ { { { {

0.2 0.7 1 0.7

0.9 0.5 0.9 0.7 0.2

{ { { {

{ { { { {

0.08 0.15 0.06 1

0.6 0.05 0.2 0.3 0.1

14.2 11.2 6.51 0.68

{ { { {

{ { { { {

1.2 0.9 0.4 0.06 (n Å 5) P õ 0.05 (n Å 5)

3 0.6 0.12 0.3 0.1

Values are nanograms of ATP or ADP per microgram of total protein; means of three to four separate incubations. Calculated on a molar basis.

thought to be due to sequestration and precipitation of calcium as calcium phosphate [15]. There is now increased evidence that calcium- and inorganic phosphate-dependent changes in the permeability of mitochondrial membranes may result in more subtle mitochondrial changes due to loss of small molecules followed by irreversible swelling [13 – 18]. Significantly, these permeability changes can occur in the absence of mitochondrial swelling [37], indicating that the swelling is an irreversible end process of the permeability change. Early mitochondrial changes have also been claimed to occur during or prior to apoptotic cell death [38, 39]. Because of the known effects of CsA on this permeability change in isolated mitochondria, we initially examined the effect of CsA on apoptosis induced by thapsigargin since this latter agent has previously been shown to result in swollen mitochondria prior to apoptotic cell death (Ref. 11 and Beaver and Waring, submitted for publication). Here we show that 2 nM thapsigargin, which results in an increase in intracellular calcium of only 100 – 250 nM above resting levels, results in apoptosis which is significantly inhibited by cyclosporin A and FK506. Mitochondria in these cells were not swollen and had taken on an appearance typical of mitochondria in activated cells with increased matrix volume [40]. Unlike gross swelling, these changes are revers-

AID

ECR 3248

/

6i12$$$463

09-03-96 10:12:58

ible [41]. There were changes in MMP in cells treated with 2 nM thapsigargin but this is reversed by CsA. Apoptosis induced under these conditions thus has the characteristics of AICD, a form of cell death induced in thymocytes and certain T-cell lines by engagment of anti-CD3 and important for negative selection. This process is inhibited by cyclosporin A and is thought to be a result of inhibition of cytokine production necessary for progression of the death program [42]. Loss of MMP under these conditions appears to be a consequence of the cells undergoing apoptosis and is reversed by CsA. Importantly there is no decrease in ATP levels at 12 h, a time at which there is already significant apoptosis occurring with 2 nM thapsigargin (not shown). Thapsigargin at 2 nM mobilizes the same pools of calcium as anti-CD3 engagement and examination of ATP and ADP levels following treatment with 2 nM thapsigargin shows an increase either in ATP or in the ATP/ADP ratio, again indicating that this process is an active one [40]. Inhibition of apoptosis caused by 2 nM thapsigargin by FK506 provides evidence that protection is not due to any effect on mitochondria since FK506 does not inhibit the membrane permeability transition in these organelles (Dr. G. E. N. Kass, personal communication). Increasing thapsigargin to 10 nM consistently re-

eca

AP: Exp Cell

EFFECTS OF CYCLOSPORIN A ON THAPSIGARGIN-INDUCED APOPTOSIS

FIG. 10. The effect of dexamethasone on ATP levels in thymocytes. (A) Cells treated with 1 mM dexamethasone, Open triangles, ATP; open squares, % apoptosis. (B) Untreated cells.

sults in only a twofold increase in intracellular calcium. Under these conditions cell death, with the typical features of apoptosis including DNA fragmentation and morphology, is still seen, but is no longer inhibited by cyclosporin A. There is also a gross change in mitochondrial morphology and an early diminution of ATP and the ATP/ADP ratio, indicative of an irreversible effect on mitochondria. There is thus a clear qualitative difference between the effects of the two concentrations of thapsigargin even though the calcium levels change by only a factor of two. Experiments with PMA lead to a rapid lowering of intracellular calcium levels in cells treated with 10 nM thapsigargin to the levels caused by 2 nM thapsigargin. Pretreatment of thymocytes with PMA results in calcium rises in the presence of 10 nM thapsigargin exactly like those caused by 2 nM. The effect of vanadate, a known inhibitor of the plasma membrane calcium pump, in overcoming the PMA low-

AID

ECR 3248

/

6i12$$$463

09-03-96 10:12:58

273

ering of calcium is additional evidence that PMA is activating this ion pump. Phosphorylation of the plasma membrane pump by PMA confirms activation of this enzyme [31]. We believe that this suggests two separate pools of calcium are mobilized by thapsigargin, one of which is sensitive to the action of the plasma membrane pump and one which is not. Alternatively it may be that concentrations of intracellular calcium resulting from 2 nM thapsigargin are not sufficiently high to be affected by the plasma membrane pump even when activated by PMA. Nevertheless calcium levels exceeding those caused by 2 nM thapsigargin, while still resulting in apoptosis, result in early irreversible damage to mitochondria and lack of inhibition by CsA. As shown by the effects on ATP and the mitochondrial membrane potential, this level of calcium may represent a threshold over which fewer physiological effects of calcium increase become manifest. In support of this, PMA results in significant restoration of ATP levels in cells treated with 10 nM thapsigargin as well as partially restoring the MMP. PMA does not rescue thymocytes from apoptosis but its long-term effects are complicated by the fact that it has been shown to induce apoptosis in its own right. The difference in the effects of 2 and 10 nM thapsigargin may be related to the physical location of the two putative calcium pools in the cell. Rizzuto et al. [43] for example have shown that calcium changes within mitochondria depend on the agent inducing the cytosolic rise, not just on the magnitude of the rise. This was attributed to the generation of local domains of high intracellular calcium close to inositol triphosphate sensitive channels. Such local high calcium rises may be triggered by very low thapsigargin concentrations, allowing entry of calcium into mitochondria similar to that which occurs during metabolic activity [40]. Alternatively, 2 nM thapsigargin may be mobilizing nuclear calcium. Further subcellular examination of the location of the calcium rises could address these questions. Treatment of cells with 10 nM thapsigargin and PMA might be expected to restore cyclosporin A sensitivity. We found no effect of CsA on such cells. However, inhibition of apoptosis by CsA caused by 2 nM thapsigargin was also abolished by PMA treatment and may be related to other effects of activation of PKC in the cell. The presence of two pools of calcium is also supported by the clear presence of two populations of cells with low and high calcium when treated with 10 nM thapsigargin in calcium-depleted media. Rises in intracellular calcium must be sustained by recruitment of extracellular calcium via capacitative filling of stores [26] or the increases are transient. Presumably in calciumdepleted media this does not occur efficiently enough for all thymocytes, allowing the lower population to be

eca

AP: Exp Cell

274

WARING AND BEAVER

FIG. 11. PMA partially restores mitochondrial membrane potential in thymocytes treated with 10 nM thapsigargin for 24 h. (A) Control; (B) 10 nM; (C) 10 nM / 1 mM PMA; (D) 2 nM alone. The region M1 defines the highly staining population of cells containing mitochondria with normal membrane potential.

seen during FACS analysis in the presence of 10 nM thapsigargin. PMA has no effect on calcium levels nor on reduced ATP levels caused by 100 nM thapsigargin which results in necrotic cell death as determined by morphological criteria (data not shown). However, the dramatically increased ADP levels in cells treated with 100 nM

TABLE 3 The Effect of Thapsigargin on Rhodamine 123 Staining of Thymocytes Treatment

% Bright cellsa

3h

Control 10 nM 10 nM / 0.2 mM PMA 10 nM / 1 mM PMA

82.2 73.1 76.6 75.0

{ { { {

1.6 2.3 2 1.2

20 h

Control 10 nM 10 nM / 0.2 mM PMA 10 nM / 1 mM PMA

75 18.6 40.3 35.0

{ { { {

1.5 1.7 1.6 2.4

24 h

Control 10 nM 10 nM / 0.05 mM CsA 2 nM 2 nM / 0.05 mM CsA 100 nM

80.0 21.0 3.5 36.3 60.5 4.7

{ { { { { {

0.8 1.1 0.5 1.5 7.5 0.4

a

Mean of four separate incubations.

AID

ECR 3248

/

6i12$$$464

09-03-96 10:12:58

thapsigargin probably reflect increased activation of the plasma membrane calcium ATPase by a calciumdependent pathway when levels exceed 1000 nM in the cytosol. Thus activation of the pump by PMA, if this were occurring, would be expected to have no further effect. This is supported by the lack of effect on ATP and calcium levels. Examination of ATP levels in dexamethasone-treated cells showed no early loss of ATP and is in agreement with earlier work which also shows that an alteration in membrane potential occurs after DNA damage [44]. In this paper we have defined two apparently separate pathways to apoptosis induced in thymocytes by thapsigargin, resulting from a relatively small difference in intracellular calcium increases. A small quantitative change in calcium levels has resulted in a qualitative change due to a switch in cell death from one with the apparent features of activation-induced cell death inhibited by CsA and FK506 to one which is not inhibited by CsA and showing early diminished ATP levels. In both cases, however, the cells ultimately display the morphology of apoptosis. Effects on MMP by 2 nM thapsigargin are reversed by CsA but this is likely due to inhibition of apoptosis per se. Concentrations of thapsigargin which double the rise in intracellular calcium are sufficient to result in irreversible mitochondrial damage and an entirely different route to apoptosis. The third scenario of excessive calcium increase caused by 100 nM thapsigargin results in necrosis.

eca

AP: Exp Cell

EFFECTS OF CYCLOSPORIN A ON THAPSIGARGIN-INDUCED APOPTOSIS

275

FIG. 12. Cyclosporin A restores mitochondrial membrane potential in thymocytes treated with 2 nM thapsigargin but not 10 nM thapsigargin for 24 h. (A) Control; (B) 10 nM; (C) 10 nM / 50 nM CsA; (D) 2 nM; (E) 2 nM / 50 nM CsA; (F) 100 nM. The region M1 defines the highly staining population of cells containing mitochondria with normal membrane potential.

This paper also demonstrates that gross mitochondrial changes and ATP decreases can occur either prior to DNA fragmentation as seen with 10 nM thapsigargin or following DNA fragmentation caused by dexamethasone. ATP changes similar to those caused by 10 nM thapsigargin occur with 100 nM thapsigargin which results in necrosis. Thapsigargin at 2 nM

AID

ECR 3248

/

6i12$$$464

09-03-96 10:12:58

causes no early ATP loss. In conclusion our data suggest that there is no simple correlation between mitochondrial function and the onset of apoptosis. The authors thank Professor Basil Roufogalis, Dr. Sek Chow, and Ms. Kathleen Doherty for most helpful comments and advice, Mr. Allan Sjaarda for excellent technical assistance, Mr. Geoff Osborne

eca

AP: Exp Cell

276

WARING AND BEAVER

and Ms. Sabine Gruninger of the JCSMR FACS laboratory for invaluable technical assistance with FACS analysis, and Ms. Kathy Gillespie for invaluable assistance with electron microscopy.

REFERENCES 1. Thastrup, O., Cullen, P. J., Drobak, B., Hanley, M. R., and Dawson, A. P. (1990) Proc. Natl. Acad. Sci. USA 87, 2466–2470. 2. Gosh, T. K., Bian, J. H., Short, A. D., Rybak, S. L., and Gill, D. L. (1991) J. Biol. Chem. 266, 24690–24697. 3. Furuya, Y., Lundmo, P., Short, A. D., Gill, D. L., and Isaacs, J. T. (1994) Cancer Res. 54, 6167–6175. 4. Kerr, J. F. R., Searle, J., Harmon, B. V., and Bishop, C. J. (1987) in Perspectives on Mammalian Cell Death (Potten, C. S., Ed.), pp. 93–128, Oxford Univ. Press, Oxford. 5. Wyllie, A. H., Morris, R. G., Smith, A. L., and Dunlop, D. (1984) J. Pathol. 142, 555–556. 6. McConkey, D. J., Hartzell, P., Amador-Perez, J. F., Orrenius, S., and Jondal, M. (1989) J. Immunol. 143, 1801–1806. 7. McConkey, D. J., Nicotera, P., Hartzell, P., Bellomo, G., Wyllie, A. H., and Orrenius, S. (1989) Arch. Biochem. Biophys. 269, 365–370. 8. McConkey, D. J., Hartzell, P., Duddy, S. K., Hakansson, H., and Orrenius, S. (1988) Science 242, 256–259. 9. Alnemri, E. S., and Litwack, G. (1990) J. Biol. Chem. 265, 17323–17333. 10. Duke, R. C., Witter, R. Z., Nash, P. B., Young, J. D-E., and Ojcius, D. M. (1994) FASEB J. 8, 237–246. 11. Beaver, J. P., and Waring, P. (1994) Immunol. Cell Biol. 72, 489–499. 12. Beaver, J. P., and Waring, P. (1995) Eur. J. Cell Biol. 68, 47– 54. 13. Chappell, J. B., and Crofts, A. R. (1961) Biochem. J. 95, 378– 386. 14. Gunter, T. E., and Pfeiffer, D. R. (1990) Am. J. Physiol. 258, C755–C786. 15. Carafoli, E. (1987) Annu. Rev. Biochem. 56, 395–433. 16. Savage, M. K., Jones, D. P., and Reed, D. J. (1991) Arch. Biochem. Biophys. 290, 51–56. 17. Griffins, E. J., and Halestrap, A. P. (1991) Biochem. J. 274, 611–614. 18. Broekemeier, K. M., Carpenter-Deyo, L., Reed, D. J., and Pfeiffer, D. R. (1992) FEBS Lett. 304, 192–194. 19. Shi, Y., Sahai, B. M., and Green, D. R. (1989) Nature 339, 625– 626. 20. Mercep, M., Noguchi, P. D., and Ashwell, J. D. (1989) J. Immunol. 142, 4085–4092.

21. Sigal, N. H., and Dumont, F. J. (1992) Annu. Rev. Immunol. 10, 519–560. 22. June, C. H., and Rabinovitch, P. S. (1994) in Methods in Cell Biology, Chap. 10, pp. 149–174, Academic Press, New York. 23. Darzynkiewicz, Z., Staiano-Coico, L., and Melamed, M. R. (1981) Proc. Natl. Acad. Sci. USA 78, 2383–2387. 24. Darzynkiewicz, Z., Bruno, S., Del Bino, G., Gorczyca, W., Holtz, M. A., Lassota, P., and Traganos, F. (1992) Cytometry 13, 795– 808. 25. Munn, E. A. (1974) in The Structure of Mitochondria, Chap. 2, pp. 70–98, Academic Press, London. 26. Takemura, H., Hughes, A. R., Thastrup, O., and Putney, J. W. (1989) J. Biol. Chem. 264, 12266–12271. 27. Lindenboim, L., Haviv, R., and Stein, R. (1995) J. Neurochem. 64, 1054–1063. 28. Clevers, H., Alarcon, B., Wileman, T., and Terhorst, C. (1988) Annu. Rev. Immunol. 6, 629–662. 29. Castagna, M., Takai, Y., Kaibuchi, K., Sano, K., Kikkawa, U., and Nishizuka, Y. (1982) J. Biol. Chem. 257, 7847–7851. 30. Kizaki, H., Tadakuma, T., Odaka, C., Muramatsu, J., and Ishimura, T. (1989) J. Immunol. 143, 1790–1794. 31. Wright, L. C., Chen, S., and Roufogalis, B. D. (1993) Arch. Biochem. Biophys. 306, 277–284. 32. Lytton, J., Westlin, M., Burke, S. E., Shull, G. E., MacLennan, D. H., and Hanley, M. R. (1991) Biophys. J. 59, 249a. 33. Wyllie, A. H. (1981) in Cell Death in Biology and Pathology (Bowen, I. D., and Lockshin, R. A., Eds.), Chap. 1, pp. 9–24, Chapman–Hall, London. 34. Wolvetang, E. J., Johnson, K. L., Krauer, K., Ralph, S. J., and Linnane, W. W. (1994) FEBS Lett. 339, 40–44. 35. Orrenius, S., Burkitt, M. J., Kass, G. E. N., Dypbukt, J. M., and Nicotera, P. (1992) Ann. Neurol. Suppl. 32, S33–S42. 36. Cohen, J. J., and Duke, R. C. (1984) J. Immunol. 132, 38–42. 37. Savage, M. K., and Reed, D. J. (1994) Arch. Biochem. Biophys. 315, 142–152. 38. Zamzami, N., Marchetti, P., Castedo, M., Zanin, C., Vayssiere, J-L., and Petit, P. X. (1995) J. Exp. Med. 181, 1661–1672. 39. Petit, P. X., Lecoeur, H., Zorn, E., Dauguet, C., Mignotte, B., and Gougeon, M-L. (1995) J. Cell. Biol. 130, 157–167. 40. McCormack, J. G., and Denton, R. M. (1993) Dev. Neurosci. 15, 165–173. 41. Hackenbrock, C. R. (1966) J. Cell Biol. 30, 269–297. 42. Schreiber, S. L., and Crabtree, G. R. (1992) Immunol. Today 13, 136–141. 43. Rizzuto, R., Bastianutto, C., Brini, M., Murgia, M., and Pozzan, T. (1994) J. Cell Biol. 126, 1183–1194. 44. Cossarizza, A. C., Kalashnikova, G., Grassilli, E., Chiappelli, F., Salviola, S., Capri, M., Barbieri, D., Troiano, L., Monti, D., and Franceschi, C. (1994) Exp. Cell Res. 214, 323–330.

Received February 16, 1996 Revised version received May 20, 1996

AID

ECR 3248

/

6i12$$$464

09-03-96 10:12:58

eca

AP: Exp Cell