British Journal of Haematology, 1997, 98, 673–685

Inhibition of autophagy abrogates tumour necrosis factor a induced apoptosis in human T-lymphoblastic leukaemic cells L I J I A , R O B ERT R. D O UR M A S H KI N ,* PAU L D. A L L E N , A L A N B. G RAY,† A DR I A N C. N E W L A N D Departments of Haematology, *Medical Microbiology and †Pathology, St Bartholomew’s and the Royal London School of Medicine and Dentistry, London

AND

S T E P H E N M. K E L S E Y

Received 11 December 1996; accepted for publication 29 May 1997

Summary. The pattern and the sequence of tumour necrosis factor-a (TNFa) induced cell death in the acute T-lymphoblastic leukaemic cell line CCRF-CEM and its vinblastineresistant subline CEM/VLB100 have been studied. Previously, we found that the CEM/VLB100 cell line was more sensitive to TNFa-induced killing than its parental CCRF-CEM cell line. TNFa-induced cell death showed an apoptotic pattern, as detected by agarose electrophoresis, flow cytometry and transmission electron microscopy (TEM). TEM images revealed that autophagy and condensed mitochondria occurred earlier than nuclear fragmentation. The specific inhibitor of autophagy, 3-methyladenine (3MA), inhibited

the formation of autophagosomes. TNFa-induced DNA fragmentation and cytolysis were completely inhibited by 10 mM 3MA. Inhibition of the fusion of lysosomes with autophagosomes by asparagine did not block TNFa-induced apoptosis. In addition, amino acid and protein deprivation enhanced TNFa-induced autophagy but not apoptosis. We propose that the early stages of autophagy are required for, but do not necessarily result in, TNFa-induced apoptosis.

Apoptosis is an active process of programmed cell death which is currently best identified by a series of morphological changes in dying cells accompanied by nuclear fragmentation. Cell shrinkage and loss of normal contents, dense chromatin condensation, cellular budding and fragmentation, and rapid phagocytosis by professional phagocytes or adjacent cells occur in a fixed sequence (Hockenbery, 1995). The shrinkage of the cytoplasm in the early stages of apoptosis does not involve lysosomal digestion of cytoplasmic constituents (Kerr et al, 1972; Majno & Joris, 1995). Classic autophagy is a general mechanism whereby eukaryotic cells degrade parts of their own cytoplasm, including organelles (except the nucleus). Structural and biochemical studies have indicated that portions of cytoplasm are first sequestered by a double membrane-bound vacuole (autophagosome), originating from ribosome-free regions of the rough endoplasmic reticulum (RER). The nascent autophagic vacuole is loaded with various acid hydrolases by fusion with pre-existing lysosomes or Golgi apparatus-derived vesicles to form a single membrane-bound degradative vacuole or autolysosome (Dunn, 1990a, b; Aubert et al,

1996). Early activation of lysosomes and formation of autophagosomes have been described as predominant features of programmed cell death during metamorphosis in embryonic tissues (Hinchliffe, 1981; Bowen et al, 1996) and during regression of tumours (Gullino, 1980). This autophagic cell death has been named type II cell death or ‘autophagic suicide’ as opposed to non-autophagic apoptotic or type I cell death (Zakeri et al, 1995; Bursch et al, 1996). TNFa binds to specific receptors on the surface of some tumour cells to induce apoptosis via a signalling pathway which involves interleukin 1b-converting enzyme (ICE) and/ or ICE-like proteases (Fraser & Evan, 1996). The precise mechanisms by which nuclear fragmentation are initiated are unknown. In addition, many tumour cells are resistant to induction of cell death by TNFa and the mechanism by which these cells acquire such resistance are unclear. Human acute T-lymphoblastic leukaemic cells (T-ALL) are known to contain abundant lysosomes and an active Golgi complex. The strongly positive acid phosphatase activity and the large numbers of autophagosomes contribute to the differentiation of T-ALL from other types of leukaemia by cytochemical or ultrastructural studies (Rozman et al, 1992). In previous studies we found that the vinblastine-resistant

Correspondence: Dr Stephen M. Kelsey, Department of Haematology, Royal London Hospital, Whitechapel, London E1 1BB. q 1997 Blackwell Science Ltd

Keywords: apoptosis, autophagy, 3-methyladenine (3MA), tumour necrosis factor a (TNFa), transmission electron microscopy (TEM).

673

674

Li Jia et al

Fig 1. Morphological definition of the autophagosomes. (A) A phase 1 autophagosome in a CCRF-CEM non-apoptotic cell after treatment with TNFa shows a mitochondrion sequestered by RER (arrow). (B) Phase 2 autophagosomes in a non-apoptotic cell after treatment with TNFa shows a myelin-like membranous structure in the autophagosomes (arrow head). (C) Phase 3 autophagosomes or residual bodies in apoptotic cell after treatment with TNFa and showing exocytosis (arrow head). Bars ¼ 0.5 mm.

q 1997 Blackwell Science Ltd, British Journal of Haematology 98: 673–685

Autophagy and Apoptosis T-lymphoblastic leukaemic cell line CEM/VLB100 and daunorubicin-resistant erythroleukaemic cell line K/DAU600 were more sensitive to TNFa-induced killing than their parental CCRF-CEM and K562 cl.6 cell lines ( Jia et al, 1995). The greater sensitivity of the drug-resistant cell lines to TNFa was not entirely due to the lower protective effects of endogenous TNFa expression and decreased manganese superoxide dismutase (MnSOD) activity. However, it was associated with increased mitochondrial electron transport chain (ETC) enzyme activity ( Jia et al, 1996). The relationship between mitochondrial ETC activity and TNFa-induced apoptosis was unclear, although preliminary evaluation of cells undergoing apoptosis by TEM suggested early involvement of mitochondria in autophagosome formation. We therefore wished to study whether mitochondrial autophagocytosis and lysosomal degradation were involved in the process of TNFa-induced apoptosis. In addition, we wished to evaluate whether autophagic activity could be involved in the differential sensitivity of CCRF-CEM and CEM/ VLB100 cell lines to TNFa. We have demonstrated that the CEM/VLB100 cell line has greater autophagic capacity and is more susceptible to TNFa-induced apoptosis than its parental CCRF-CEM cell line. TNFa induces both autophagy and apoptosis. 3-Methyladenine (3MA), a specific inhibitor of the early stages of autophagy, blocked TNFa-induced autophagy and completely abrogated TNFa-mediated apoptosis. However, the induction of autophagy per se does not necessarily induce apoptosis. We suggest that the early stages of autophagy, which may involve possible activation of a neutral endonuclease or nascent protease synthesis, are required for TNFa-induced apoptosis and the process is independent of the later stage of autolysis.

675

cells were incubated at 378C with 250 mg/ml RNase A and 50 mg/ml PI for 1 h. PI fluorescence of nuclei was measured by flow cytometry (EPICS Elite, Coulter, U.S.A.) with excitation at 488 nm and emission at 620 nm. Data analysis was carried out on cells gated on an integral channel v peak channel display to exclude cell debris and clumped cells. Preparation and electrophoretic analysis of DNA. DNA from 2 × 106 cells was prepared and analysed by 1% agarose (Type I) gel electrophoresis (Allen & Newland, 1996). Briefly, cells were lysed by incubation at 568C for 60 min in lysis buffer [10 mM EDTA, 50 mM Tris-HCl, pH 8.0, containing 0.5% (w/v) N-lauroylsarcosine and 0.5 mg/ml proteinase K]. RNase A was added to the sample to give a final concentration of 250 mg/ml and the incubation was continued for a further 60 min. The lysates were then extracted three times with chloroform/isoamyl alcohol (24 : 1), centrifuged at 13 000 g for 10 min and the fragmented DNA was precipitated from the supernatant in 70% ethanol at 48C overnight. The precipitated DNA was collected by centrifugation at 13 000 g for 10 min at 48C and dissolved in TE buffer (10 mM Tris-HCl, pH 8.0, 1 mM EDTA). The DNA concentration was measured by Gene Quant (Pharmacia). Before electrophoresis, loading buffer [0.1 M EDTA, pH 8.0, 0.05% (w/v) bromophenol blue and 40% (w/v) sucrose] was added to each sample at a 1 : 4 ratio. Approximately 20 mg of DNA was loaded into each well and electrophoresis was carried out at 70 V for 80 min in TBE buffer (2 mM EDTA, 89 mM Tris-borate, pH 8.0). A FX174 Hae III digested DNA was applied to each gel to provide a size marker. After electrophoresis, DNA was stained by soaking the gel in TBE buffer containing 1 mg/ml ethidium bromide for 10 min

MATERIALS AND METHODS Materials. Recombinant human TNFa (2 × 104 U/mg protein), 3-methyladenine (3MA), DL-asparagine (Asp), leupeptin (Leup), ribonuclease A (RNase A), proteinase K, DNA FX174 Hae III digested molecular weight markers, agarose (Type I and Type VII), propidium iodide (PI) and lactate dehydrogenase (LDH) assay kit were purchased from Sigma (U.K.). Sodium cacodylate, osmium tetroxide (OsO4), uranyl acetate, propylene oxide, lead acetate were from Agar Scientific Ltd (U.K.). TAAB 812 resin was from TAAB Laboratories Equipment Ltd (U.K.). Cell culture conditions. The human acute T-lymphoblastic leukaemic (T-ALL) CCRF-CEM cell line and its vinblastineresistant subline CEM/VLB100 were used in this study. The cells were cultured in RPMI 1640 medium (Sigma) supplemented with 10% heat-inactivated fetal calf serum (FCS), 25 mM HEPES, 2.0 mM L-glutamine, pH 7.4, penicillin (100 U/ml), streptomycin (100 mg/ml) at 378C in a 5% CO2 humidified incubator. Cell concentration was maintained at 3–10 × 105 cells/ml by subculturing every 2 or 3 d. DNA content analysis by flow cytometry. Leukaemic cells (5 × 105/ml) were seeded in 24-well plates and incubated with or without 250 U/ml TNFa. After different incubation times, 200 ml cells were taken out and permeabilized in 70% ethanol for 40 min at 48C. After washing twice with PBS,

Fig 2. CEM/VLB100 cells have more active autophagic function than CCRF-CEM cells. The number of autophagosomes in resting cells were counted for each cell from TEM micrographs. 86 CCRF-CEM and 87 CEM/VLB100 cells were used for this analysis. The number of autophagosomes per cell was plotted against the percentage of the cell population containing this number of autophagosomes. A significant difference between the two cell lines (P < 0.005) was statistically confirmed by Mann-Whitney U test.

q 1997 Blackwell Science Ltd, British Journal of Haematology 98: 673–685

676

Li Jia et al

Fig 3. TNFa-induced cell death showed an apoptotic pattern. (A) Kinetic analysis of TNFa-induced DNA fragmentation in the CCRF-CEM and CEM/VLB100 cell lines assessed by flow cytometry after incubation with 250 U/ml TNFa for 3–9 h. The cell cycle phases were: G0/G1 (B), S (C), G2M (E), and apoptotic cells (F).

Fig 3(B). TNFa-induced DNA fragmentation was confirmed by agarose gel electrophoresis after incubation with or without TNFa for 6 h. Lanes 1–4 refer to the CCRF-CEM cell line and lanes 5–7 to CEM/VLB100 cells. Lanes 1 and 5 are control cells. Lane 2 is cells treated with 1 mM ZnSO4. Lanes 3 and 6 are cells treated with TNFa only. Lanes 4 and 7 are cells pre-treated with 1 mM ZnSO4 for 30 min prior to the addition of TNFa. The standard DNA fragments (lane 8) are indicated on the right as numbers of base pairs. q 1997 Blackwell Science Ltd, British Journal of Haematology 98: 673–685

Autophagy and Apoptosis

677

Fig 4. TNFa-enhanced autophagy in the non-apoptotic cells. CCRF-CEM (A) and CEM/VLB100 (B) cells were treated with TNFa for 3 and 6 h and compared with the control cells. Autophagosomes in nonapoptotic cells were counted in TEM micrographs. The cell numbers used for analysis were: CCRF-CEM control cells ¼ 86, TNF 3 h ¼ 47, TNF 6 h ¼ 35; CEM/VLB100 control cells ¼ 87, TNF 3 h ¼ 33, TNF 6 h ¼ 38. Autophagosome numbers per cell were plotted against frequency. Significantly enhanced autophagy was observed in response to TNFa (3 h P < 0.0001, 6 h P < 0.05 in CCRF-CEM cells; 3 h P < 0.05 in CEM/ VLB100 cells) compared with control cells and confirmed by Mann-Whitney U test.

followed by destaining in TBE buffer and visualized on a UV transilluminator. Assessment of cell death. Leukaemic cells (106/ml) were incubated with or without 250 U/ml TNFa. 100 ml of cell suspensions were taken out for analysis. Cell death was kinetically quantified by measurement of lactate dehydrogenase (LDH) release into the medium. The released LDH was expressed as a percentage of total cellular LDH, which was determined after complete lysis of the cells with 0.1% Triton-X 100. The LDH activity was determined by LDH assay kit. Briefly, LDH catalyses the oxidation of lactate to pyruvate with simultaneous reduction of NADþ. The formation of NADH results in an increase in absorbance at 340 nm. The rate of increase in absorbance at 340 nm is directly proportional to LDH activity in the cells (Amador et al, 1963). The LDH activity was recorded kinetically at 308C for 30 s. The assay was performed using a spectrophotometer (LKB Biochem Ultrospec II) connected to Kinetics Software run on an Apple IIe Computer ( Jia et al, 1996). Transmission electron microscopy. CCRF-CEM and CEM/ VLB100 cells (106/ml) were incubated with or without

250 U/ml TNFa. Cells were taken out at indicated times. The cells which were suspended in medium were fixed for 1 h in 1.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.2–7.4) at room temperature, washed twice in cacodylate buffer, postfixed for 1 h on ice in 1% OsO4 in 0.1 M cacodylate buffer, then washed twice with distilled water. The fixed cells were stained with 0.5% uranyl acetate for 1 h. After washing once, cells were embedded in 2% (w/v) agarose (Type VII), dehydrated in a graded ethanol series and infiltrated with 100% propylene oxide. The specimens were further infiltrated in TAAB 812 resin embedded in suitable capsules and placed in an oven (608C) overnight. Silver-gold sections were cut on a Reichert ultramicrotome (Ultracut E) and mounted on copper grids. Lead citrate staining was performed under conditions free from carbon dioxide. The sections were viewed and photographed in a Philips EM 301 electron microscope (Dourmashkin et al, 1993). Morphological definitions. TNFa-induced morphological changes were viewed by TEM image. Micrographs of single cells were taken by random screening at 7–15k magnification. 40–100 films of each sample were used for morphological analysis. The autophagic process was divided into

q 1997 Blackwell Science Ltd, British Journal of Haematology 98: 673–685

678

Li Jia et al

three phases: phase 1 is the formation of autophagosomes containing portions of the cells own cytoplasm or organelles which are sequestered by a membranous border (Fig 1A); phase 2 occurs when the autophagosomes begin digestion of their contents including lysosomal granules (Fig 1B); phase 3, is seen as residual bodies with digested contents surrounded by a sealed membrane (Fig 1C) (Bainton, 1981; Pfeifer, 1987; Seglen, 1987). Statistical analysis. Statistical analysis was performed by using Statistica (Statsoft) and Excel programs for Windows. Non-parametric statistics were used for comparison of groups with different numbers of autophagosomes per cell because the data are positively skewed. Significant differences between two non-parametric groups were analysed by Mann-Whitney U test. Series data from multiple samples were compared by ANOVA. Comparison of cell death by percentage of LDH leakage was performed using a two-tailed paired t-test. A significant difference was set at a ¼ 0.05 for all comparisons. RESULTS CEM/VLB100 cells have more autophagic activity than their parental CCRF-CEM cells The morphological ultrastructure of CCRF-CEM and CEM/ VLB100 cells were analysed by TEM. Autophagosomes and lysosomal granules were often seen in both cell lines; however, CEM/VLB100 cells underwent more autophagy in the resting state than the parental cells (P < 0.005, MannWhitney U test) (Fig 2). TNFa-induced apoptosis was Zn2þ-inhibitable Flow cytometric analysis of cellular DNA content was used to quantify apoptosis (Barry & Eastman, 1993) (Fig 3A).

Table I. The CEM/VLB100 cell line is more sensitive to TNFa-induced apoptosis than the CCRF-CEM cell line.

Percentage of apoptotic cells (%) Time (h)

CCRF-CEM

CEM/VLB100

0 3 6 9

2.3 6 0.7 6.8 6 2.7 16.6 6 3.0 17.5 6 3.4

2.9 6 0.5 17.8 6 1.7 28.7 6 1.1 31.9 6 2.3

The data shown were obtained from DNA histograms as displayed in Fig 3 by four separate experiments. Significant differences between the two cell lines after 250 U/ml TNFa treatment were analysed by ANOVA (P < 0.01).

CEM/VLB100 cells were more sensitive to TNFa-induced apoptosis than CCRF-CEM cells. Apoptotic cells appeared in the sub G0/G1 fraction after 3 h in the CEM/VLB100 cell line and after 6 h in the CCRF-CEM cell line (Table I and Fig 3). This was confirmed by agarose gel electrophoresis of DNA from these cells which demonstrated double-stranded fragments of DNA migrating in a ‘ladder’ pattern (multiples of 185 bp) (Kerr & Harmon, 1991) (Fig 3B). When both CCRFCEM and CEM/VLB100 cells were pre-treated by 1 mM ZnSO4 for 30 min, TNFa-induced DNA fragmentation was completely inhibited. This was confirmed by cytospin staining. TNFa failed to induce nuclear fragmentation after ZnSO4 treatment. This suggests that the pathway of TNFa-induced

Table II. TNFa-enhanced the autophagic process in non-apoptotic cells.

Conditions

Total cells counted

Total no. per cell

Phase 1 (%)

Phase 2 (%)

Phase 3 (%)

CEM control* CEM þ TNF 3 h* CEM þ TNF 6 h* CEM þ TNF 3 h† CEM þ 3MA þ TNF† CEM þ Asn þ TNF* CEM þ Leup þ TNF* CEM/VLB control* CEM/VLB þ TNF 3 h* CEM/VLB þ TNF 6 h* CEM/VLB þ TNF 3 h× CEM/VLB þ 3MA þ TNF† CEM/VLB þ Asn þ TNF* CEM/VLB þ Leup þ TNF*

82 47 35 60 53 34 38 87 33 36 41 61 31 35

2 (0–5) 7 (4–10) 5 (2-6) 9 (4.5–12) 4.5 (0.5–7) 4 (1–6) 4 (2–7) 4 (2–8) 6 (4–8) 3 (1–5) 11 (7–14.5) 5 (1–7) 3 (1–6) 3 (2–5)

10 9 6 4 2 17 7 13 5 2 6 0.3 26 8

23 43 43 40 23 49 34 25 36 17 23 22 35 11

67 48 51 56 75 34 59 62 59 81 71 78 39 81

Cells were incubated with 250 U/ml TNFa in culture medium (*) or in HBSS (†). The test groups pre-treated with autophagic inhibitors (3MA, Asn or leupeptin) and exposed to TNFa were fixed at 3 h. The numbers of autophagosomes in each phase were counted by using TEM micrographs according to the definition described in the Materials and Methods. Data shown are the percentage number of autophagosomes (%). Total no. per cell represents the median and interquartile range (in parentheses) for the numbers of autophagosomes in each cell counted and analysed by non-parametric statistics. q 1997 Blackwell Science Ltd, British Journal of Haematology 98: 673–685

Autophagy and Apoptosis

679

Fig 5. TNFa-induced autophagy in apoptotic cells. (A) An apoptotic CCRF-CEM cell which had been treated with TNFa for 3 h. The chromatin becomes pyknotic and packed into a half-moon-like mass against the nuclear membrane. Arrows show the phase 1 autophagosomes sequestered by RER. Bar ¼ 1 mm.

apoptosis in these cells, possibly the endonuclease, was Zn2þinhibitable. TNFa enhanced autophagy in apoptotic and non-apoptotic cells The autophagic process was notably enhanced by TNFa, in

both apoptotic and non-apoptotic cells. The number of autophagosomes in non-apoptotic cells was significantly increased in both cell lines by TNFa treatment for 3 h (P < 0.0001, Mann-Whitney U test) (Fig 4A and B). The membranous border of the autophagosomes in these cell

q 1997 Blackwell Science Ltd, British Journal of Haematology 98: 673–685

680

Li Jia et al

Fig 5(B). Apoptotic CCRF-CEM cell treated with TNFa for 6 h shows a dense rounded mass containing two pyknotic nuclear fragments. The apoptotic cell demonstrates exocytosis of phase 3 autophagosomes (arrow heads). Bar ¼ 1 mm.

lines was derived from ribosome-bound RER (Fig 1A). The percentage of phase 2 autophagosomes (Fig 1B) was increased in the non-apoptotic cells by TNFa (Table II and Fig 1B). After 6 h of exposure to TNFa the number of autophagosomes in each cell was less than at 3 h, although still higher than in the resting state in the CCRF-CEM (P < 0.05, MannWhitney U test) but not in the CEM/VLB100 cell lines (Fig 4). At this time point the numbers of autophagosomes in the non-apoptotic CEM/VLB100 cells (median ¼ 3, interquartile range 1–5) were less than in CCRF-CEM cells (median ¼ 5, interquartile range 2–6). The numbers of phase 1 and phase 2 autophagosomes in CEM/VLB100 cells were markedly

reduced at 6 h, with increased numbers in phase 3. The percentage of each phase in the CCRF-CEM cell line remained similar to that seen at 3 h (Table II). This indicated that autophagy in the CEM/VLB100 cell line was passing from phase 1 to phase 2, and from phase 2 to phase 3 more rapidly than in the CCRF-CEM cell line. Nascent autophagic bodies which were sequestered by RER and the nuclear envelope were prominent in apoptotic cells (Fig 5A). The folded double membrane which sequestered cytoplasm or mitochondria was connected to the nuclear envelope. This was only observed in the apoptotic cells and appeared distinct from the phase 1 autophagosomes in the non-apoptotic cells. The number and the phase

q 1997 Blackwell Science Ltd, British Journal of Haematology 98: 673–685

Autophagy and Apoptosis of autophagic bodies in the apoptotic cells could not be accurately assessed due to release of autophagosomes by exocytosis (Figs 1C and 5B). The membranes of phase 2 and phase 3 autophagosomes were intact (Figs 1C and 5B). 3-Methyladenine (3MA), a specific inhibitor of autophagy, completely prevented TNFa-induced DNA fragmentation and cytolysis Three inhibitors of autophagy were used to elucidate the relationship between autophagy and apoptosis. 3MA is a specific inhibitor of autophagy which has been demonstrated to block autophagic sequestration, i.e. the first step of the autophagic process (Seglen & Gordon, 1982; Kopitz et al, 1990). To assess the effect of 3MA on TNFa-induced autophagy and apoptosis, cells were cultured under amino acid and protein-deficient conditions (Seglen, 1983; Sandvig & van Deurs, 1992). Protein and amino acid deprivation enhanced TNFa-induced autophagy in the CEM/VLB100 cell

681

line (P < 0.005, Mann-Whitney U test) but did not increase apoptosis as analysed by both TEM and flow cytometry (data not shown). TNFa-enhanced autophagy was significantly inhibited in the presence of 3MA (P < 0.0001, MannWhitney U test) (Fig 6), as demonstrated by the lack of RER sequestration or the formation of phase 1 autophagosomes. The numbers of phase 2 autophagosomes were similar to control cells, but more residual bodies (phase 3) were seen in the presence of 3MA (Table II). The effect of 3MA on TNFa-induced apoptosis was assessed by flow cytometry after 9 h exposure to TNFa. TNFa-induced apoptosis in both the CCRF-CEM and CEM/ VLB100 cell lines was completely abrogated by 10 mM 3MA (Fig 7A). The inhibitory effect of 3MA was confirmed by absence of DNA ladders as detected by agarose gel electrophoresis and absence of apoptotic morphological change by TEM (data not shown). 3MA also inhibited TNFa-induced cytolysis after 9 h exposure to TNFa as

Fig 6. The inhibitory effect of 3MA on TNFa-enhanced autophagy. CCRF-CEM (A) and CEM/VLB100 (B) cells were suspended in HBSS, pre-treated with 10 mM 3MA for 30 min and incubated with 250 U/ml TNFa for 3 h. Autophagosomes in non-apoptotic cells were counted in TEM micrographs. The cell numbers used for this analysis were: CCRF-CEM þ TNF 3 h in HBSS ¼ 60, 3MA þ TNF 3 h ¼ 53; CEM/VLB100 þ TNF 3 h in HBSS ¼ 41, 3MA þ TNF 3 h ¼ 61. Autophagosome numbers per cell were plotted against frequency of occurrence. A significant inhibitory effect of 3MA on autophagy (P < 0.0001) was compared with the cells treated with TNFa alone in HBSS and statistically confirmed by Mann-Whitney U test. q 1997 Blackwell Science Ltd, British Journal of Haematology 98: 673–685

682

Li Jia et al

Fig 7. 3MA abrogated TNFa-induced apoptosis and cytolysis. (A) 3MA completely inhibited TNFa-induced DNA fragmentation as assessed by flow cytometry after incubation with 10 mM 3MA and TNFa for 9 h. TNFa-induced apoptosis (sub-G0/G1 population) was absolutely abrogated by 3MA. See legend to Fig 3A for cell cycle phases.

assessed by LDH leakage (Fig 7B). Experiments could not be extended beyond 9 h as cell viability in HBSS in which the 3MA was dissolved was not sufficient. Asparagine (Asn) and leupeptin did not inhibit TNFa-induced apoptotic cell death Asn has been described to inhibit phase 2 of autophagy, the fusion of autophagosome with lysosome, or ‘amphisome’ formation (Kopitz et al, 1990). Leupeptin, an inhibitor of serine and cysteine proteinase (Seglen, 1983), inhibits lysosomal protein degradation (Kopitz et al, 1990). The inhibitory effects of Asn and leupeptin on autophagy were confirmed in this study. Cells were pre-treated with 10 mM Asn or 300 mM leupeptin for 30 min, respectively, in the culture medium and further incubated with 250 U/ml TNFa for 3 h. The total number of autophagosomes and the number of cells with autophagosomes were significantly less than those with TNFa alone (P < 0.01, Mann-Whitney U test). The effect of Asn on autophagy was demonstrated by blocking the transfer of autophagosomes from phase 1 to phase 2, and from phase 2 to phase 3 (Table I). Leupeptin reduced the number of autophagosomes (P < 0.005, Mann-

Whitney U test) but did not have a demonstrable effect on any specific phase. Neither Asn nor leupeptin showed an inhibitory effect on TNFa-induced apoptosis and cytolysis (data not shown). DISCUSSION Autophagy is the mechanism for the sequestration of cytoplasm or unwanted organelles into lysosomes. It is a catabolic pathway for the degradation of intracellular membranes, cytosolic components and secretory products by lysosomes. Autophagy is an ongoing process in the steady state but can be induced under pathological conditions such as starvation and decline of the supply of respiratory substrates to the mitochondria (Aubert et al, 1996). It was recently demonstrated that a novel antigen, 7A6, appears on the mitochondrial outer membrane when cells undergo apoptosis, suggesting that the 7A6 molecule may be involved in the molecular cascade of apoptosis (Zhang et al, 1996). In addition, Reipert et al (1995, 1996) reported that autophagy during apoptosis is triggered by mitochondrial alterations. Autophagic capacity varies from cell to cell. It has been

q 1997 Blackwell Science Ltd, British Journal of Haematology 98: 673–685

Autophagy and Apoptosis

Fig 7(B). The inhibitory effect of 3MA on TNFa-induced cytolysis. Cells were suspended in HBSS with or without TNFa for 9 h. Cell death was assessed by LDH leakage into the medium. Data shown are mean 6SD of five independent experiments. A significant inhibitory effect of 3MA was observed by comparison with cells treated with TNFa alone (statistically analysed by paired two-tailed t-test). *P < 0.05, **P < 0.01.

suggested that release of lysosomal contents into the cytosol during the tertiary stages of autophagy results directly in apoptosis due to the presence, and subsequent release, of endonuclease II (DNase II) (Gullino, 1980; Hinchliffe, 1981). As a result, the presence of active autophagy has been associated with a greater tendency to undergo autophagyassociated apoptosis, or autophagic suicide (Schwarze & Seglen, 1985; Seglen, 1987; Sandvig & van Deurs, 1992). Acute T lymphoblastic leukaemic cells, including the CCRF-CEM cell line, have readily detectable lysosomal activity. Preliminary studies in our laboratory suggested that lysosomal enzyme activity and autophagy were temporally associated with the initiation of TNFa-induced apoptosis. In this study we have confirmed that TNFa induces death by apoptosis in cells of the CCRF-CEM and CEM/VLB100 leukaemic sublines, and that the CEM/VLB100 cell line is more sensitive to TNFa-induced apoptosis than its parental CCRF-CEM cells ( Jia et al, 1996). We have also shown that TNFa-induced DNA fragmentation is inhibitable by Zn2þ, a feature of a pH-neutral endonuclease (Shiokawa et al, 1994). This neutral endonuclease locates to either the RER or nuclear envelope (Peitsch et al, 1993) and is distinct from the lysosomal DNase II which requires intracellular acidification for DNA cleavage (Barry & Eastman, 1993). The inhibitory effect of Zn2þ is downstream of mitochondrial permeability transition (Zamzami et al, 1996). Quantitative analysis of autophagy by TEM revealed that CEM/VLB100 cells had more numerous autophagic bodies in the resting state than the parental CCRF-CEM cell line. In addition, exposure to TNFa increased autophagic activity in both cell lines within 3 h, prior to the detection of DNA fragmentation. In supporting this phenomenon, autophagic

683

elimination of altered mitochondria in an etoposide-treated haemopoietic stem cell line, FDCP-Mix, preceded nuclear apoptosis (Reipert et al, 1996). After 6 h, non-apoptotic cells contained fewer autophagic bodies than resting cells. These results suggest that increased autophagic activity is in some way associated with susceptibility to, and the process of, TNFa-induced apoptosis. Cells which survived the apoptotic stimulus of TNFa were either those that had less autophagic activity initially, or had down-regulated the autophagic process in some way. This observation suggests that an earlier stage of the autophagic process, possibly involving RER, was primarily associated with TNFa-induced apoptosis, rather than lysosomal release of DNase II or other acid hydrolases. We do not deny the lysosome release of DNase II can induce apoptosis, as defined type II apoptosis. In our study we found that phase III autophagosomes were released by exocytosis (Figs 1C and 5B) rather than broken in the cytosol, in agreement with Reipert et al (1996). This is also consistent with the involvement of a Zn2þ-inhibitable, RER-associated endonuclease. Initiation of autophagy by depriving cells of amino acids did not, however, result in an increase in apoptosis. Therefore the association between autophagic activity and apoptosis is dependent on a specific stimulus such as TNFa and is not primarily dependent on the process of autophagy per se. In order to establish a causal association between autophagy and TNFa-induced apoptosis, three inhibitors of autophagy were used. 3MA inhibited the initial autophagic sequestration of organelles by RER although the biochemical mechanism is unclear. 3MA inhibited autophagosome formation and completely inhibited TNFa-induced apoptosis in the CCRF-CEM and CEM/VLB100 cell lines. It was also found that 3MA inhibited autophagy and apoptosis in human mammary carcinoma cells, MCF-7, induced by the anti-oestrogen tamoxifen (Bursch et al, 1996). The later stages of autophagy were inhibited by Asn and leupeptin although no effect on TNFa-induced apoptosis was observed. We therefore suggest that autophagic sequestration of organelles by RER is a critical step in TNFa-induced apoptosis. DNA fragmentation is not dependent on lysosomal release of endonuclease or other acid hydrolases. 3MA, even at a relatively high concentration of 10 mM, has been reported to be a specific inhibitor of the autophagic process (Seglen, 1983, 1987; Kopitz et al, 1990). However, we cannot rule out the possibility that it has other effects on cellular biochemistry at this concentration which contribute to its ability to inhibit TNFa-induced apoptosis. We observed that the RER, which is responsible for autophagic sequestration, consisted of ribosome bound membrane rather than ribosome-free regions of RER (Dunn, 1990a, b). The RER is responsible for protein synthesis, but acts as an intracellular pool of Zn2þinhibitable neutral DNase (Peitsch et al, 1993) and of Ca2þ (Lam et al, 1994). Recent work has demonstrated that one of the major steps in the morphologic changes which occur during apoptosis is the activation of one or more proteases (Martin & Green, 1995; Fraser & Evan, 1996). The activation occurs in several stages: first, cytosolic caspase-3

q 1997 Blackwell Science Ltd, British Journal of Haematology 98: 673–685

684

Li Jia et al

(CPP32-like protease) is cleaved and activated in a reaction that appears to be triggered by cytochrome c release from mitochondria (Liu et al, 1996; Yang et al, 1997); second, activated caspase-3 interacts with an other cytosolic protein, called DNA fragmentation factor (DFF), to activate DNase I and generate DNA fragmentation (Liu et al, 1997). Even though apoptosis can occur in the absence of the nucleus ( Jacobson et al, 1994), and the apoptotic process can be blocked at any level of the apoptotic cascade, such as the mitochondria (Zamzami et al, 1996; Kroemer et al, 1997; Yang et al, 1997), ICE-like proteases or caspase-like proteases (Zhivotovsky et al, 1997), the cleavage of chromatin into nucleosomal fragments is still the biochemical hallmark of apoptosis (Liu et al, 1997). We therefore propose that the initial phase of autophagy is required, but is not sufficient for, TNFa-induced apoptosis. The effect may be mediated via activation of RER, with subsequent release of Ca2þ or a neutral endonuclease. ACKNOWLEDGMENTS The authors thank Professor D. N. Landon (Department of Neurocytology, Institute of Neurology, University of London) for his help in identification of TEM morphology. We also thank Dr D. Syndercombe Court for her statistical advice. REFERENCES Allen, P.D. & Newland, A.C. (1996) Apoptosis detection by DNA analysis. Molecular Diagnosis of Cancer (ed. by F. E. Cotter), pp. 207– 213. Humana Press, New Jersey. Amador, E., Dorfman, L.E. & Wacker, W.E.C. (1963) Serum lactate dehydrogenase: an analytical assessment of current assays. Clinical Chemistry, 9, 391–395. Aubert, S., Gout, E., Bligny, R., Marty-Mazars, D., Barrieu, F., Alabouvette, J., Marty, F. & Douce, R. (1996) Ultrastructural and biochemical characterization of autophagy in higher plant cells subjected to carbon deprivation: control by the mitochondria with respiratory substrates. Journal of Cell Biology, 133, 1251–1263. Bainton, D.F. (1981) The discovery of lysosomes. Journal of Cell Biology, 91, 66s–76s. Barry, M.A. & Eastman, A. (1993) Identification of deoxyribonuclease II as an endonuclease involved in apoptosis. Archives of Biochemistry and Biophysics, 300, 440–450. Bowen, I.D., Mullarkey, K. & Margen, S.M. (1996) Programmed cell death during metamorphosis in the blow-fly Calliphora vormitoria. Microscope Research Techniques, 34, 202–217. Bursch, W., Ellinger, A., Kienzl, H., Torok, L., Pandey, S., Sikorska, M., Walker, R. & Hermann, R.S. (1996) Active cell death induced by the anti-estrogens tamoxifen and ICI 164 384 in human mammary carcinoma cells (MCF-7) in culture: the role of autophagy. Carcinogenesis, 17, 1595–1607. Dourmashkin, R.R., Bucher, D. & Oxford, J.S. (1993) Small virus-like particles bud from the cell membranes of normal as well as HIVinfected human lymphoid cells. Journal of Medical Virology, 39, 229–232. Dunn, W.A., Jr (1990a) Studies on the mechanisms of autophagy: formation of the autophagic vacuole. Journal of Cell Biology, 110, 1923–1933. Dunn, W.A., Jr (1990b) Studies on the mechanisms of autophagy: maturation of the autophagic vacuole. Journal of Cell Biology, 110, 1935–1945.

Fraser, A. & Evan, G. (1996) A license to kill. Cell, 85, 781–784. Gullino, P.M. (1980) The regression process in hormone-dependent mammary carcinomas. Hormone and Cancer (eds Iacobelli et al), pp. 271–279. Raven Press, New York. Hinchliffe, J.R. (1981) Cell death in embryogenesis. Cell Death in Biology and Pathology (ed. by I. D. Bowen and R. A. Lockshin), pp. 35–77. Chapman and Hall, London. Hockenbery, D. (1995) Defining apoptosis. American Journal of Pathology, 146, 16–19. Jacobson, M.D., Burne, J.F. & Raff, M.C. (1994) Programmed cell death and Bcl-2 protection in the absence of a nucleus. EMBO Journal, 13, 1899–1910. Jia, L., Jiang, X.R., Razak, K., Wu, Y.L., Newland, A.C. & Kelsey, S.M. (1995) TNF-mediated killing of human leukaemic cells: effects of endogenous antioxidant levels and TNFa expression in leukaemic cell lines. Leukaemia Research, 19, 187–194. Jia, L., Kelsey, S.M., Grahn, M.F., Jiang, X.R. & Newland, A.C. (1996) Increased activity and sensitivity of mitochondrial respiratory enzymes to TNFa-induced inhibition is associated with cytotoxicity in drug-resistant leukemic cell lines. Blood, 87, 2401– 2410. Kerr, J.F.R. & Harmon, B.V. (1991) Definition and incidence of apoptosis: an historical perspective. Apoptosis: the Molecular Basis of Cell Death (ed. by L. D. Tomei and F. O. Cope), pp. 5–19. Cold Spring Harbor Laboratory Press, New York. Kerr, J.F.R., Wyllie, A.H. & Currie, A.R. (1972) Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics. British Journal of Cancer, 26, 239–257. Kopitz, J., Kisen, G.O., Gordon, P.B., Bohley, P. & Seglen, P.O. (1990) Nonselective autophagy of cytosolic enzymes by isolated rat hepatocytes. Journal of Cell Biology, 111, 941–953. Kroemer, G., Zamzami, N. & Susin, S.A. (1997) Mitochondrial control of apoptosis. Immunology Today, 18, 44–51. Lam, M., Dubyak, G., Chen, L., Nunez, G., Miesfeld, R.L. & Distelhorst, C.W. (1994) Evidence that BCL-2 represses apoptosis by regulating endoplasmic reticulum-associated Ca2þ flux. Proceedings of the National Academy of Sciences of the United States of America, 91, 6569–6573. Liu, X., Kim, C.N., Yang, J., Jemmerson, R. & Wang, X. (1996) Induction of apoptosis programme in cell-free extracts: requirement of dATP and cytochrome c. Cell, 86, 147–157. Liu, X., Zou, H., Slaughter, C. & Wang, X. (1997) DFF, a heterodimeric protein that functions downstream of caspase-3 to trigger DNA fragmentation during apoptosis. Cell, 89, 175–184. Majno, G. & Joris, I. (1995) Apoptosis, oncosis and necrosis: an overview of cell death. American Journal of Pathology, 146, 3–15. Martin, S.J. & Green, D.R. (1995) Protease activation during apoptosis: death by a thousand cuts? Cell, 82, 349–352. Peitsch, M.C., Polzar, B., Stephan, H., Crompton, T., MacDonald, H.R., Mannherz, H.G. & Tschopp, J. (1993) Characterization of the endogenous deoxyribonuclease involved in nuclear DNA degradation during apoptosis (programmed cell death). EMBO Journal, 12, 371–377. Pfeifer, U. (1987) Functional morphology of the lysosomal apparatus. Lysosomes: their Role in Protein Breakdown (ed. by H. Glaumann and F. J. Ballard), pp. 3–60. Academic Press, London. Reipert, S., Berry, J., Hughes, M.T., Hickman, J.A. & Allen, T.D. (1995) Changes of mitochondrial mass in the hemopoietic stem cell line FDCP-Mix after treatment with etoposide: a correlation study by multiparameter flow cytometry and confocal and electron microscopy. Experimental Cell Research, 221, 281–288. Reipert, S., Hickman, J.A. & Allen, T.D. (1996) DNA inclusions within autolytic cytoplasmic vacuoles of hemopoietic stem cell line FDCP-Mix. Journal of Histochemistry and Cytochemistry, 44, 549–558.

q 1997 Blackwell Science Ltd, British Journal of Haematology 98: 673–685

Autophagy and Apoptosis Rozman, C., Woessner, S., Feliu, E., Lafuente, R. & Berga, L. (1992) Cell Ultrastructure for Haematologists. Ediciones Doyms. Sandvig, K. & van Deurs B. (1992) Toxin-induced cell lysis: protection by 3-methyladenine and cyclohexamide. Experimental Cell Research, 200, 253–262. Schwarze, P.E. & Seglen, P.O. (1985) Reduced autophagic activity, improved protein balance and enhanced in vitro survival of hepatocytes isolated from carcinogen-treated rats. Experimental Cell Research, 157, 15–28. Seglen, P.O. (1983) Inhibitors of lysosomal function. Methods in Enzymology, 96, 737–764. Seglen, P.O. (1987) Regulation of autophagic protein degeneration in isolated liver cells. Lysosomes: their Role in Protein Breakdown (ed. by H. Glaumann and F. J. Ballard), pp. 371–414. Academic Press, London. Seglen, P.O. & Gordon, P.B. (1982) 3-Methyladenine: a specific inhibitor of autophagic/lysosomal protein degradation in isolated rat hepatocytes. Proceedings of the National Academy of Sciences of the United States of America, 79, 1889–1892. Shiokawa, D., Ohyama, H., Yamada, T., Takahashi, K. & Tanuma, S.

685

(1994) Identification of an endonuclease responsible for apoptosis in rat thymocytes. European Journal of Biochemistry, 226, 23–30. Yang, J., Liu, X., Bhalla, K., Kim, C.N. & Wang, X. (1997) Prevention of apoptosis by Bcl-2: release of cytochrome c from mitochondria blocked. Science, 275, 1129–1132. Zakeri, Z., Bursch, W., Tenniswwod, M. & Lockshin, R. (1995) Cell death: programmed, apoptosis, necrosis or other? Cell Death and Differentiation, 2, 87–96. Zamzami, N., Susin, S.A., Marchetti, P., Hirsch, T., Go`mezMonterrey, I., Castedo, M. & Kroemer, G. (1996) Mitochondrial control of nuclear apoptosis. Journal of Experimental Medicine, 183, 1533–1544. Zhang, C., Ao, Z., Seth, A. & Schlossman, S.F. (1996) A mitochondrial membrane protein defined by a novel monoclonal antibody is preferentially detected in apoptotic cells. Journal of Immunology, 157, 3980–3987. Zhivotovsky, B., Burgess, D.H., Vanage, D.M. & Orrenius, S. (1997) Involvement of cellular proteolytic machinery in apoptosis. Biochemical and Biophysical Research Communications, 230, 481– 488.

q 1997 Blackwell Science Ltd, British Journal of Haematology 98: 673–685