Biochimica et Biophysica Acta 1793 (2009) 546–560

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r.com/ locate /bbamcr

Role of BNIP3 in TNF-induced cell death — TNF upregulates BNIP3 expression

Saeid Ghavami a,b,c, Mehdi Eshraghi a, Kamran Kadkhoda a, Mark M. Mutawe b,c, Subbareddy Maddika a,1,Graham H. Bay a, Sebastian Wesselborg e, Andrew J. Halayko b,c, Thomas Klonisch d, Marek Los f,g,⁎a Manitoba Institute of Cell Biology, and Department of Biochemistry and Medical Genetics, University of Manitoba, Canadab Department of Physiology, University of Manitoba, Canadac Manitoba Institute of Child's Health, University of Manitoba, Canadad Department of Human Anatomy and Cell Sciences, and Manitoba Institute of Child Health, Winnipeg, Canadae Department of Internal Medicine I, University of Tübingen, Tübingen, Germanyf BioApplications Enterprises, Winnipeg, Manitoba, Canadag Interfaculty Institute for Biochemistry, University of Tübingen, D-72076 Tübingen, Germany

Abbreviations: TNF, tumor necrosis factor alpha; AEndo-G, Endonuclease G; tBid, truncated bid; BNIP3, Bcl2TM, trans-membrane; PTP, permeability transition pofactor-alpha-1; NO, nitric oxide; ROS, reactive oxygen spthiazolyl)-2,5-diphenyl-2H-tetrazolium bromide; PARPase-1; DAF-2DA, di-aminofluorescein-2 diacetate; GAphate dehydrogenase; HSP60, heat shock protein 60;HtrA2, high-temperature requirement A2; Smac, secocaspases; DIABLO, direct inhibitor of apoptosis binding(methylamidino)-L-ornithine acetate salt, JC-1, 5,5′,6,6′-benzimidazolylcarbocyanine iodide; DHR-123, dihydrochondrial membrane potential change⁎ Corresponding author. BioApplications Enterprise

Canada. Tel./fax: +1 204 334 5192.E-mail address: mjelos@gmail.com (M. Los).

1 Current address: Department of Therapeutic Radiolog

0167-4889/$ – see front matter © 2009 Elsevier B.V. Aldoi:10.1016/j.bbamcr.2009.01.002

a b s t r a c t

a r t i c l e i n f o

Article history:

Tumor necrosis factor alpha Received 17 September 2008Received in revised form 8 December 2008Accepted 5 January 2009Available online 15 January 2009

Keywords:CaspaseCathepsinFlow cytometryLysosomeROS

(TNF) is a cytokine that induces caspase-dependent (apoptotic) and caspase-independent (necrosis-like) cell death in different cells. We used the murine fibrosarcoma cell line model L929and a stable L929 transfectant over-expressing a mutated dominant-negative form of BNIP3 lacking the C-terminal transmembrane (TM) domain (L929-ΔTM-BNIP3) to test if TNF-induced cell death involved pro-apoptotic Bcl2 protein BNIP3. Treatment of cells with TNF in the absence of actinomycin D caused a rapid fall inthe mitochondrial membrane potential (ΔΨm) and a prompt increase in reactive oxygen species (ROS)production, which was significantly less pronounced in L929-ΔTM-BNIP3. TNF did not cause the mitochondrialrelease of apoptosis inducing factor (AIF) and Endonuclease G (Endo-G) but provoked the release of cytochromec, Smac/Diablo, and Omi/HtrA2 at similar levels in both L929 and in L929-ΔTM-BNIP3 cells. We observed TNF-associated increase in the expression of BNIP3 in L929 that was mediated by nitric oxide and significantlyinhibited by nitric oxide synthase inhibitor N5-(methylamidino)-L-ornithine acetate. In L929, lysosomal swellingand activation were markedly increased as compared to L929-ΔTM-BNIP3 and could be inhibited by treatmentwith inhibitors to vacuolar H+-ATPase and cathepsins −B/−L. Together, these data indicate that TNF-induced celldeath involves BNIP3, ROS production, and activation of the lysosomal death pathway.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Apoptosis and necrosis are two classical forms of cell death but amixed death mode that shares elements of both characteristics mayalso occur. In multicellular organisms, apoptosis can be triggered bymultiple events, including cancer therapeutics, and its purpose is to

IF, apoptosis inducing factor;/E1B 19kD interacting protein;re; HIF-1α, hypoxia inducingecies; MTT, 3-(4,5-dimethyl-2--1, poly(ADP-ribose)polymer-PDH, glyceraldehyde-3-phos-HDAC1, histone deacytalase 1;nd mitochondrial activator ofprotein of low PI; LNMMA, N5-tetrachloro-1,1′,3,3′-tetraethyl-rhodamine-123; ΔΨm, mito-

s, Winnipeg, MB, R2V 2N6,

y, Yale School Of Medicine, USA.

l rights reserved.

eliminate damaged or undesirable cell [1–3]. By contrast, necrosis is alargely uncontrolled cell death and can result from exposure toextremely noxious stimuli [4]. Apoptosis and necrosis are distin-guished by morphological and biochemical criteria. Morphologicalcharacteristics of apoptosis include condensed, fragmented cell nuclei,a reduction in cell volume and cell membrane blebbing. Typicalmorphological characteristics of necrosis include pronounced swel-ling of the mitochondria and cell membrane, which leads to cellrupture. Biochemical characteristics unique to apoptosis include anincrease in outer-mitochondrial membrane permeability, inter-nucleosomal fragmentation of chromosomal DNA, and exposure ofphosphatidylserine on the outer surface of the cell [3,5].

Originally identified as a factor causing hemorrhagic necrosis inestablished tumors [6,7], the cytokine tumor necrosis factor alpha(TNF) has since been shown to play a crucial role in the pathogenesisof acute and chronic inflammatory diseases [8,9]. TNF can induceapoptotic (caspase-dependent) or necrotic (caspase-independent) celldeath in vitro, depending on the cell type used [10–12]. Inflammatoryresponses include components of both TNF-induced apoptosis andnecrosis but in the presence of the pan-caspase/apoptosis inhibitorand enhancer of necrosis, zVAD shows increased TNF toxicity in vivo

547S. Ghavami et al. / Biochimica et Biophysica Acta 1793 (2009) 546–560

[13]. The different effects induced by TNF are mediated via two cellsurface TNF receptors known as high affinity TNF-R1 (55 kD) and lowaffinity TNF-R2 (75 kD) [14]. TNF binding to its receptors activate acascade of biochemical reactions including caspase activation, activa-tion of protein kinases [15], and the generation of ROS [16].

Upon binding to its high-affinity receptor, TNF-family memberstrigger a cascade of events, which lead to the recruitment of adaptorproteins and the activation of caspase-8. The binding of CD95 to itsreceptor activates caspase-8. Activated caspase-8 starts a mitochon-dria-independent caspase cascade directly relaying on the activationof caspase-3. Caspase-8 may also cleave and hyper-activate the pro-apoptotic Bid molecule. Truncated bid (tBid) translocates to mito-chondria and activates the mitochondrial death pathway [17,18].Thus, mitochondria may serve as a signal amplifier leading to theactivation of caspase-9, -6, -8 and -3 [19]. BNIP3 (Bcl2/E1B 19kDinteracting protein) over-expression induces an atypical, mixedcaspase-independent form of cell death. BNIP3 was discovered in ayeast two-hybrid screen as the protein that interacts with adenovirusE1B 19K, which is homologous to Bcl2 [20]. All Bcl2 family proteinspossess at least one of four Bcl2 homology domains (BH1 to BH4)which determine the ability of these proteins to induce or inhibitapoptosis [21]. BNIP3 belongs to the BH3-only subfamily and has a C-terminal trans-membrane (TM) domain [22]. Over-expression ofBNIP3 leads to the opening of the mitochondrial permeabilitytransition pore (PTP), thereby abolishing the proton electrochemicalgradient. This activates a chain of events culminating in chromatincondensation and DNA fragmentation [23,24]. Although chromatincondensation is an established marker of apoptosis, it has beenproposed that BNIP3 induces a novel necrosis-like form of cell death.BNIP3-induced cell death is independent of caspases and the nucleartranslocation of AIF, a mitochondrial flavoprotein. Furthermore, therelease of cytochrome c from mitochondria is not involved in BNIP3-mediated cell death [24]. Thus, BNIP3-mediated cell death resemblesTNF-induced cell demise that combines hallmarks of necrosis andapoptosis, in L929 cells [12].

Here we report that TNF-triggered cell death induced in theabsence of actinomycin D (commonly used accelerator of TNF-inducedcell death) is partially inhibited by the dominant negative mutant ofBNIP3, L929-ΔTM-BNIP3, which lacks the C-terminal BNIP3 domainessential for insertion of BNIP3 into the mitochondrial membrane andexecution of apoptotic function. TNF induced both a HIF-1αindependent but NO-dependent increase of BNIP3 expression, andtransfer of this Bcl2-familymember from the nucleus to mitochondria.TNF-triggered cell death involved ROS-production and the activationof the lysosomal pathway. The protective effect of ΔTM-BNIP3 wasassociated with decreased mitochondrial ROS production.

2. Materials and methods

2.1. Reagents

Cell culture media were purchased from Sigma (Oakville, ON,Canada) or Gibco (Canada). Phosphate buffered saline (PBS; pH=7.4),N5-(methylamidino)-L-ornithine acetate salt (L-NMMA), 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT),and murine anti-human BNIP3 were from Sigma. Cell culture plasticsware obtained from Nunc Co. (Canada). Anti-human/murine/rabbitBNIP3, rabbit anti-human/murine/rat glyceraldehyde-3-phosphatedehydrogenase (GAPDH), murine anti-human/murine/rat tubulin,rabbit anti-human/murine/rat Omi/HtrA2, murine anti-human/murine/rat cytochrome c, rabbit anti-human/murine/rat HIF-1α,and goat anti-human/murine/rat Endo-G were obtained from SantaCruz Biotechnologies (USA). 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethyl-benzimidazolylcarbocyanine iodide (JC-1), and DAF-2DA wereobtained from Invitrogen Molecular Probes (Canada). Murine anti-human/murine cleaved caspase-3 (Asp175), was obtained from R&D,

Canada. Rabbit anti-human/murine cleaved PARP-1 were purchasedfrom Cell Signaling, Canada. Caspase-Glo®-8, and -9 assay systemswere obtained from Promega, Canada.

2.2. Cell culture

Murine fibroblast cell lines L929 and stable transfectant of L929with a dominant negative mutant of BNIP3, L929-ΔTM-BNIP3, werecultured in DMEM supplemented with 10% fetal calf serum, 100 U/mlpenicillin and 100 μg/ml streptomycin. Cells were incubated at 37 °C ina humidified CO2 incubator with 5% CO2 and maintained underlogarithmic growth conditions.

2.3. Development of L929-ΔTM-BNIP3 cells

L929 cells were grown in 100 mm dishes. After reaching 80%confluence, cells were collected (106 cells) and transfected withpCDNA3-BNIP3-ΔTM (5 μg) using electroporation. After 48 h trans-fected cells were treated with selection media containing Geneticin(G418) (1000 μg/ml) and kept in selection media for 3 months toproduce cells stably over-express ΔTM-BNIP3 protein. The expressionof ΔTM-BNIP3 protein was controlled by Western blot.

2.4. MTT cytotoxicity assay

Colorimetric MTT assay was employed to evaluate the cytotoxicityeffect of TNF on L929 and L929-ΔTM-BNIP3 cell lines [25]. Briefly, cells(1.5×104 cells/ml) were grown for 24 h in 96-well culture platescontaining 200 μl of culture medium. Defined concentrations of TNFwere added and incubated for different time intervals, as shown in thefigures, followed by the incubationwith 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide. The percentage of cell viabilitywas calculated using the equation: [mean optical density (OD) oftreated cells /mean OD of control cells]×100.

2.5. Measurement of apoptosis by flow cytometry

Apoptosis was measured using the Nicoletti method [26,27].Briefly, cells grown in 12 well plates were treated with S100A8/A9(100 μg/ml) for the indicated time periods. After scrapping, cells werepelleted by centrifugation at 800 g for 5 min, washed once with PBS,and resuspended in hypotonic propidium iodide (PI) lysis buffer (1%sodium citrate, 0.1% Triton X-100, 0.5 mg/ml RNase A, 40 μg/mlpropidium iodide). Cell nuclei were incubated for 30 min at 30 °C andthe nuclei were subsequently analyzed by flow cytometry. Nuclei tothe left of the G1 peak containing hypo-diploid DNA were consideredto be apoptotic.

2.6. ROS measurement using flow cytometry

L929 and L929-ΔTM-BNIP3 cells were treated with TNF withindicated concentrations for different time points in 6 well plates.Following treatment, cells were incubated in 1 μM dihydro-rhodamine-123 (DHR-123; 1 μM) in DMSO (final concentration 0.1%v/v). Incubation with DHR-123 was for 30 min at 37 °C. To preventlight accelerated oxidation, samples were maintained in the darkprior to and during analysis. Fluorescence within dye-loaded cellswas analyzed by flow cytometry (FACS Calibur; Becton Dickinson,Mississauga, ON, Canada. For flow cytometry, the excitation wave-length for rhodamine-123 fluorescence was 488 nm and emission at530 nm (FL-1). Signals were processed using a logarithmic amplifierand fluorescence distributions were plotted on a 4-decade logarith-mic scale. 10,000 events were counted and, as a semi-quantitativeindicator for ROS levels, the geo-mean and mean-linear fluorescencevalues were calculated using Cell Quest Software (Becton Dickinson,Mississauga, ON, Canada).

548 S. Ghavami et al. / Biochimica et Biophysica Acta 1793 (2009) 546–560

2.7. Measurement of mitochondrial membrane potential

To measure the Ψm of cells, the fluorescent probe JC-1 (5,5,6,6 -tetrachloro-1,1,3,3 -tetraethylbenzimidazole carbocyanide iodide) wasused. JC-1 exists as a monomer at low values of Ψm (greenfluorescence), while it forms aggregates at highΨm (red fluorescence).Thus, mitochondria with normal Ψm concentrate JC-1 into aggregates(red fluorescence), but with de-energized or depolarized Ψm, JC-1forms monomers (green fluorescence). Briefly, L929 and L929-ΔTM-BNIP3 (5×105 cells/well were treated with TNF (0.2, 0.8 and 1 ng/ml)for 8 and 12 h, washed in PBS (pH=7.4), and incubated for 15 min at37 °C with 2.5 μg/ml JC-1. Cells were pelleted at 400 g for 5 min inroom temperature, washed in PBS, and analyzed by flow cytometry.The analyzer threshold was adjusted on the forward light scatterchannel to exclude most of the sub-cellular debris. When excitedsimultaneously by a 488 nm argon-ion laser source, JC-1 monomersand JC-1-aggregates were separated into the flow cytometer FL1 andFL2 channels, respectively. Mean fluorescence intensity values for FL1and FL2 expressed as relative linear fluorescence channels wereobtained for at least 15,000 events in all experiments. The decrease inΨm was determined employing a 3-D plot with FL2, FL1, and cellcounts depicted on the x/y/z axis, respectively [28].

2.8. Cell fractionation

Cytosolic, mitochondrialand and nuclear fractions were generatedusing a digitonin-based subcellular fractionation technique asdescribed previously [29,30]. Briefly, 107 cells were harvested bycentrifugation at 800 g, washed in PBS pH 7.2, and re-pelletted. Cellswere digitonin-permeabilized for 5 min on ice at a density of 3×107/ml in cytosolic extraction buffer (250mM sucrose, 70mMKCl,137mMNaCl, 4.3mMNa2HPO4,1.4mMKH2PO4 pH 7.2,100 μMPMSF,10 μg/mlleupeptin, 2 μg/ml aprotinin, containing 200 μg/ml digitonin). Plasmamembrane permeabilization of cells was confirmed by staining in a0.2% trypan blue solution. Cells were then centrifuged at 1000 g for5 min at 4 °C. The supernatants (cytosolic and mitochondria fractions)were saved and the pellets solubilized in the same volume of nuclearlysis buffer, followed by pelletting at 12,500 g for 10 min at 4 °C. Themitochondria was separated from cytosolic fraction pelletting13,000 g and the pellets solubilized in equal amount of mitochondriallysis buffer (50mMTris pH 7.4,150mMNaCl, 2mMEDTA, 2mMEGTA,0.2% Triton X-100, 0.3% NP-40, 100 μM PMSF, 10 μg/ml leupeptin, 2 μg/ml aprotinin). For the detection of proteins, equal amount of fractionsprotein were supplemented with 5× SDS-PAGE loading buffer,subjected to standard 12% SDS-PAGE and transferred to nitrocellulosemembranes.

2.9. Immunoblotting

Upon treatment of cells with TNF (1 ng/ml) for indicated timepoints, BNIP3, cytochrome c, and caspase-3 were detected byimmunoblotting in total cell lysates and fractionated cell compart-ments, respectively. Briefly, harvested cells were washed once withcold PBS and cells were re-suspended for 20 min on ice in lysis buffer:20 mM Tris–HCl (pH 7.5), 0.5% Nonidet P-40, 0.5 mM PMSF, and 0.5%protease inhibitor cocktail (Sigma). Then, experimental procedureswere followed exactly as described previously [28].

2.10. Quantitative PCR measurement of BNIP3 mRNA

Cells were treated with TNF (1 ng/ml), TNF solvent (PBS) andmedium (control). Total cellular RNA was isolated using the RNeasyPlus Mini Kit (Qiagen) according to the manufacturer's instructions.1 μg of total RNA was reverse transcribed by using the QuantiTectReverse Transcription Kit (Qiagen) according to the manufacturer'srecommendations. mRNA expression of murine BNIP3 in L929 cells

was quantified on a Applied Biosystems 7500 Real-Time PCR Systeminstrument using Power SYBR Green PCR Master Mix (AppliedBiosystems) [31]. A dissociation curve was generated at the end ofeach PCR cycle to verify that a single product was amplified. Therelative expression levels of the BNIP3 gene was calculated as folddifference=2(−Ct), where Ct is threshold cycle normalized to 18Sinternal control. Oligonucleotide primerswere as follows for BNIP3, 5′-GCTCCCAGACACCACAAGAT-3′ (forward) and 5′-TGAGAGTAGCTG-TGCGCTTC-3′ (reverse), and for 18S: 5′-CGCCGCTAGAGGTGAAATTC-3′ (forward) and 5′-TTGGCAAATGCTTTCGCTC-3′ (reverse).

2.11. Nitric oxide measurement using DAF-2DA

Recently, a newfluorometricmethod based ondi-aminofluorescein-2 diacetate (DAF-2 DA) has been developed for the direct detection ofNO production. The DAF-2DA penetrates the cell membrane and is thenhydrolysed by cytosolic esterases, producing DAF which traps NO toproduce the green-fluorescence signal fluorescein. The high sensitivityof DAF and its broad concentration response range make this system ahighly sensitive tool to detect even low amounts of NO produced by theconstitutive enzymes [32]. Briefly, DAF-2DA stock solution was diluted(2.5 μM) inpre-warmedHBSSbuffer (Hank's Balanced Salt solutionwith20mMHepes bufferwithCa2+ andMg2+ (pH=7.4) and added to the cellstreated with TNF (0–24 h, 1 ng/ml) and control cells (treated withmedium and TNF solvent, PBS). The cells were incubatedwith DAF-2DAfor 40min and thenwashedwithHBSSone time to remove excess probeand incubated in HBSS for more 30 min to allow complete de-esterification of the intracellular diacetates. Fluorescence excitationand emission were calculated in 495 and 515 nm, respectively.

2.12. Immunocytochemistry and confocal imaging

Cells were grown overnight on coverslips and then treated withTNF for 12 h. Cells were washed in PBS, fixed in 4% paraformaldehydein PBS at room temperature, and permeabilized with 0.25% Triton X-100 in PBS. To detect BNIP3, AIF, Endo-G, Omi/HtrA2, and Smac/Diablotranslocation, cells were incubated with anti-BNIP3 murine IgG(Sigma; diluted 1:200) anti-AIF murine IgG (Santa Cruz Biotechnol-ogy; diluted 1:500), anti-Endo-G rabbit IgG (Biovision, 1:50 dilution),anti-Omi/HtrA2 rabbit IgG, and anti-Smac/Diablo rabbit IgG (SantaCruz Biotechnology; both at 1:150). Slides were washed three timeswith PBS and incubated with Cy5-conjugated secondary antibody(1:1000; BNIP3, AIF, and Endo-G) or FITC-conjugated secondaryantibody (1:50; BNIP3, Omi/HtrA2 and Smac/Diablo). To visualizenuclei, cells were stained with DAPI (10 μg/ml). The mitochondriawere stained with Mitotracker Red CMXRos (Molecular Probes;200 nM) in DMEM culture medium for 15 min prior to fixation. Thecell-permeant MitoTracker probes contained a mildly thiol-reactivechloromethyl moiety that appeared to be responsible for keeping thedye associated with the mitochondria after fixation. To labelmitochondria, cells were incubated with sub-micromolar concentra-tions of a MitoTracker probe, which passively diffused across theplasma membrane and accumulated in active mitochondria. Once themitochondriawere labeled, cells were treated with an aldehyde-basedfixative to allow further processing of the sample. Because most of theMitoTracker probes were retained after permeabilization withdetergents or organic solvents, the sample continued to exhibit thefluorescent staining pattern characteristic of live cells duringsubsequent processing steps (e.g., immunocytochemistry, in situhybridization, or electron microscopy) [26]. Fluorescent images wereanalyzed using an Olympus-IX81 multi-laser confocal microscope.

2.13. Caspase activity assays

Luminometric assays Caspase-Glo®-8, -9 (Promega, Canada, Nepean,ON) were used to measure the proteolytic activity of Caspase-8 (IETD-

Fig. 1. TNF induced cell death in L929 cells. (A) L929 cells were treated with differentconcentrations of TNF for 0–24 h and cell viability was assessed by MTT assay. For eachTNF concentration the treated cells were compared with control cells which had beentreated with the TNF solvent PBS. Results are expressed as percentage of correspondingcontrol and represent the means±SD of three independent experiments. TNFsignificantly induced cell death in L929 cells in different concentration and time points(Pb0.001). (B) DNA histogram of L929 cells treated with TNF. The Nicoletti method wasused for sample processing, typical DNA-histogram s are shown. M2 (statistical marker)has been placed to mark sub-diploid DNA. TNF- treatment (24 h, 1 ng/ml) inducedapoptosis only a fraction of cells. G1 and G2 peaks are clearly visualized and a sub-diploid peak corresponding to apoptotic cells was clearly visible to the left from bothpeaks that represent normal cells. (C) L929 cells were treated with TNF (1 ng/ml) forindicated times, then total cell extracts were harvested, resolved by SDS-PAGE, andcleaved PARP-1 was detected by Western blot. Protein loading was controlled withGAPDH, (parallel experiment performed with the same protein extracts).

549S. Ghavami et al. / Biochimica et Biophysica Acta 1793 (2009) 546–560

ase) and -9 (LEHD-ase). The assays were performed according tomanufactures instruction. Briefly, cells cultured in 96-well plate (15000cells/well), were treated with indicated concentrations of TNF fordifferent time points. The freshly prepared caspase reagents (100 μl perwell) contained either z-LETD-Luciferin alone or z-LEHD-Luciferin andwhole protein cell lysate extract buffer. For each experiment, negativecontrol cells (treated cells in medium without caspase reagent andreagent blank containing medium only) were included. The measure-ments were then performed exactly as described previously [28].

2.14. Statistical analysis

The results were expressed as means±SD and statistical differ-ences were evaluated by one-way and two-way ANOVA followed byTukey's post-hoc test, using the software package SPSS 11.05 andGraph Pad Prism 4.0, Pb0.05 was considered significant.

3. Results

3.1. TNF induced cell death in L929 cells

L929 cells were treated with different concentrations of TNF indifferent time points and TNF cytotoxic effect was measured usingMTTassay. In each concentration/time point cell viabilitywas accessedcomparing to corresponding control (cells treated with TNF solvent,PBS). To avoid experimental artifacts, all experiments were performedwithout the transcriptional inhibitor actinomycin D. Our resultsshowed that TNF induced significant cell death in all concentration/time point in L929 cells (Pb0.001) (Fig. 1A). To confirm that theobserved form of cell death was apoptosis, the experiments wereindependently repeated using an apoptosis-specific flow cytometricmethod (Nicoletti), which detects apoptosis-typical hypodiploidnuclei. TNF-treated cells showed apoptotic cell death (Fig. 1B). It waspreviously shown that TNF-induced apoptosis caused PARP-1 cleavage[12]. In our experimental model we also show that TNF induced PARP-1 cleavage in L929 cells, thus, providing additional confirmation ofapoptosis induction by TNF (Fig. 1C).

3.2. TNF induced time and concentration-dependent BNIP3 expression inL929 cells

We investigated the effect of TNF treatment on the regulation ofBNIP3 expression. Time-kinetics experiments indicate that BNIP3protein expression (Fig. 2A, B) and BNIP3-mRNA content as assayed byquantitative PCR (Fig. 2C), were increased in L929 cells treated withTNF. BNIP3 protein expression also showed dependence on TNFconcentration. Thus, higher concentrations of TNF induced higherexpression of BNIP3 (Fig. 2D). Therefore for the further experimentswe have been mostly using the higher (1 ng/ml) TNF concentration.

3.3. Dominant-negative mutant of BNIP3 partially inhibited TNF toxicityand TNF-induced BNIP3 mitochondrial translocation without affectingcytochrome c release and caspase-9 activation

To study the role of BNIP3 in TNF cytotoxicity, we compared TNF-triggered changes between L929 and a stable transfectants with thedominant-negative mutant of BNIP3 that lacks the trans-membranedomain, which is critical for its association with mitochondria (L929-ΔTM-BNIP3). Due to the strong toxicity of BNIP3 even in transienttransfection experiments, we were unable to provide data on cellseven transiently overexpressing BNIP3. MTT assay revealed that L929-ΔTM-BNIP3 cells were significantly more resistant towards TNF ascompared to the parental L929 cell line (Pb0.01) (Fig. 3A). To avoidexperimental artifacts, all experiments were performed without thetranscriptional inhibitor actinomycin D. During the initiation of celldeath BNIP3 can associate with mitochondria and interact with Bcl2

Fig. 2. TNF induced time and concentration dependent BNIP3 expression in L929 cells. (A) Western blot detection of BNIP3 expression in L929 cells treated with 1 ng/ml TNF for 0, 8,16 and 24 h. Tubulinwas included as loading control. 30 and 60 kD signals of BNIP3were significantly increased after 8 and 16 h. In control samples “C”, the cells were treatedwith theTNF solvent PBS. The Western blot signals are representative of three independent experiments. (B) Quantification of BNIP3 expression in TNF treated. Western blots of experimentsperformed as described in (A) were quantified as indicated in the method section (C) TNF induced an increase in BNIP3 mRNA expression. L929 cells were treated with TNF forindicated times and BNIP3 mRNA was measured using qPCR. TNF induced a significant increase in BNIP3 mRNA (Pb0.05). The data represent triplicates of three independentexperiments. (D) Western blot of BNIP3 expression in L929 cells treated with 0.75 and 1 ng/ml TNF for 12 h. BNIP3 expression was TNF-concentration dependent. Tubulin wasincluded as loading control. The Western blot data is representative for of three independent experiments.

550 S. Ghavami et al. / Biochimica et Biophysica Acta 1793 (2009) 546–560

and Bcl-XL [23,33,34]. The predicted molecular weight of BNIP3 is21.5 kD. However, in SDS-PAGE it migrates as a monomer of 30 kD anda homodimer of ∼60 kD. In some models, homodimerization appearsto be a feature of mitochondrial localization [24] and TM-domainmutants of BNIP3 including BNIP3Δ179–185 and point mutations atL179 and G180 fail to homodimerize, as does a C-terminal deletionmutant (BNIP3Δ184–194) [34]. Furthermore, exclusive mitochondriallocalization for BNIP3 was only shown for some tissues (i.e. muscle),but not for others (i.e. the brain) [24]. Interestingly, the BH3 domain ofBNIP3 does not appear to be essential for BNIP3-dependent cell deathinduction [34]. We have investigated the intracellular BNIP3 distribu-tion upon TNF treatment in L929 and L929-ΔTM-BNIP3 cells. Inuntreated L929 cells, BNIP3 was mainly localized in nuclei (Fig. 3B),whereas in L929-ΔTM-BNIP3 cells negligible amounts of nuclearBNIP3 were observed (Fig. 3B). In both L929 and L929-ΔTM-BNIP3cells, TNF-treatment resulted in an increased presence of BNIP3 in themitochondrial fraction, while BNIP3 content decreased in the nuclearfraction in L929 (Fig. 3B). The data obtained by cell fractionation wereconfirmed by immunohistochemistry followed by confocal micro-scopy (Fig. 3C, D).

TNF-induced apoptosis and necrosis both involve mitochondrialparticipation. Interaction of TNF with its TNF receptor-1 ultimatelyleads to the activation of caspase-8 as well as activation of themitochondrial, caspase-9 dependent death pathway. We investigatedcaspase activation upon TNF treatment in L929 and L929-ΔTM-BNIP3cells. TNF-triggered mitochondrial cytochrome c release, and theactivation of caspase-9. The release of cytochrome c from mitochon-dria and the extent of caspase-9 were similar in L929 and L929-ΔTM-BNIP3 (Fig. 3E, F). This indicated that the TNF-mediated activation ofthe mitochondrial death pathway did not require the presence offully-functional BNIP3. Furthermore, similar levels of caspase-3activation were detected in L929 and L929-ΔTM-BNIP3 cells (Fig.3G), again indicating that the caspase cascade was not affected byBNIP3 mutation.

3.4. TNF-mediated increase in BNIP3 expression involves NO-synthase

We investigated the effect of TNF treatment on the regulation ofBNIP3 expression. BNIP3 protein expression (Fig. 2A, B) andmRNA (Fig.2C), was increased in L929 cells treated with TNF. As HIF-1α is one ofthe transcription factors regulating BNIP3 expression [24,34,35], weinvestigated if HIF-1α played role in TNF-induced increase of BNIP3expression. TNF failed to induce HIF-1α expression, thus excludingHIF-1α transcriptional activity as a cause of the TNF-mediated BNIP3upregulation (data not shown). It was shown recently that BNIP3expression could be affected by nitric oxide [36,37]. Endogenouslyproduced or exogenously added NO activates the Bnip3 promoter andinduces expression of BNIP3 protein in RAW264.7 macrophages undernormoxic conditions [36]. TNF significantly induces NO and NO2

−

production in L929 cells (Pb0.01) (Fig. 4A). These cells express bothconstitutive and a TNF-inducible nitric oxide synthase (NO-synthase)[38,39]. Treatment with the competitive nitric oxide synthase (NOS)inhibitor (L-NMMA) prior to TNF stimulation significantly inhibitedTNF-induced cell death and upregulation of 30 kD BNIP3-expressionand cell death in L929 cells (Fig. 4B, C). In the presence of L-NMMA,there was a significant increase in 60 kD BNIP3 but 30 kD BNIP3significantly decreased (Pb0.05) (Fig. 4C, D). The NOS inhibitorattenuated the overall increase of BNIP3 content in TNF-treated cellsand shifted the equilibrium towards the 60 kD BNIP3 form. We havealso investigated if L-NMMA could prevent BNIP3 translocation tomitochondria. Cells were pre and then co-treated with L-NMMA andTNF (1 ng/ml). L-NMMA could not prevent BNIP3 translocation tomitochondria (data was not show).

3.5. Dominant-negative BNIP3 mutant protects from TNF-triggeredmitochondrial depolarization and ROS generation

The change in Ψm is a common cellular event during TNF-induced cell death [40–42]. TNF treatment induced changes in Ψm

Fig. 3. The expression of BNIP3 dominant negative mutant ΔTM-BNIP3 partially protects L929 against TNF-induced cell death via prevention of BNIP3 translocation to mitochondria,without significantly affecting in cytochrome c release and caspase-9, and -3 activation. (A) Effect of TNF on L929 and L929-ΔTM-BNIP3 cells. Cells were treated with differentconcentrations of TNF for 0–24 h and cell viability was assessed byMTT assay. For each TNF concentration the treated cells were compared with control cells which had been treatedwith the TNF solvent PBS. Results are expressed as percentage of corresponding control and represent the means±SD of three independent experiments. There is significantdifference between parental L929 and L929-ΔTM-BNIP3 cell in TNF cytotoxicity (Pb0.01). (B) TNF induced mitochondrial translocation of BNIP3. BNIP3 presence in mitochondrial,and nuclear fractions of L929 and L929-ΔTM-BNIP3 cells treated with TNF (1 ng/ml). BNIP3 signals were detected by Western blot. BNIP3 was mainly located in the nucleus in L929cells and translocated to mitochondria when cells were treated with TNF. In L929-ΔTM-BNIP3 cells, TNF induced less pronounced BNIP3 translocation to mitochondria. Proteinloading was controlled with HDAC1 (nuclear fraction), and HSP60 (mitochondrial fraction). The data is representative for three independent experiments. (C, D) Intracellularlocalization of BNIP3 in (C) L929 and (D) L929-ΔTM-BNIP3 cells was determined by confocal microscopy. Cells were either left untreated or treated with TNF (1 ng/ml) for 12 h. Cellswere immunostained with an anti-BNIP3 antibody followed by a FITC conjugated secondary antibody (green), and DAPI (nuclear DNA counterstain, blue). TNF-treated cells showedmitochondrial BNIP3 translocation which was less pronounced in L929-ΔTM-BNIP3 cells. The anti-BNIP3 antibodies used were unable to differentiate between wild type andmutated BNIP3. The magnification was 1000×. Representative micrographs are shown for all samples. (E) Cytochrome c detection by Western blot in mitochondrial and cytosolicfractions of L929 and L929-ΔTM-BNIP3 cells treated with TNF (1 ng/ml) for indicated time. Protein loading was controlled with GAPDH (cytosolic fraction) and HSP60 (mitochondrialfraction). The data represent representative blots of three independent experiments. (F) Caspase-9 activity in L929 and L929-ΔTM-BNIP3 cells treated with TNF (1 ng/ml) wasmeasured by a Caspase-Glo® luminometric assay. The caspase activity is presented as “fold increase” in comparison to the control of each time point. The data represent triplicates ofthree independent experiments. There was no significant difference in caspase-9 activity between the two cell lines (PN0.05). (G) L929 and L929-ΔTM-BNIP3 cells were treated withTNF (1 ng/ml) for indicated time periods, then total cell extracts were harvested, resolved by SDS-PAGE, and active subunits of caspase-3 were detected by Western blot. Proteinloading was controlled with GAPDH, (parallel experiment performed with the same protein extracts). GAPDH represented for L929 cells. The data represent triplicates of threeindependent experiments.

551S. Ghavami et al. / Biochimica et Biophysica Acta 1793 (2009) 546–560

Fig. 3 (continued).

552 S. Ghavami et al. / Biochimica et Biophysica Acta 1793 (2009) 546–560

in both wild type L929 and ΔTM-BNIP3-over-expressing cells (Fig.5A). These changes were dose and time dependent (data not shown)and revealed L929 that exhibited higher changes in Ψm than L929-ΔTM-BNIP3 transfectants (Fig. 5A). Employing the ROS-sensitivefluorogenic marker DHR-123, we investigated ROS generation byflow cytometry in untreated and treated (TNF at 1 ng/ml) L929 andL929-ΔTM-BNIP3 cells. When compared with wild type L929 cells,L929-ΔTM-BNIP3 transfectants produced significantly less ROS (Fig.5B, C) (Pb0.05). N-acetycysteine (NAC) (a clinically-applied anti-oxidant) interferes with increased ROS generation and TNF-induceddeath in L929 cells [25]. In our experimental system, NAC pre- andco-treatment partially prevented TNF-induced cell death in L929cells (Fig. 5D).

3.6. TNF treatment does not induce EndoG and AIF release frommitochondria in L929 cells

AIF is a phylogenetically conserved flavoprotein essential forapoptosis during embryonic development [43]. It is a redox-activeNADH-oxidase and, upon apoptosis induction, translocates frommitochondria to the nucleus to induce chromatin condensation [44].

Endo-G is a mitochondrial nuclease encoded by a nuclear gene.Once released from mitochondria into the cytosol, Endo-G trans-locates to the nucleus where it causes oligo-nucleosomal DNAfragmentation even in the presence of caspase inhibitors [45]. Inmammalian cells, Endo-G cooperates with exonuclease and DNase-Ito facilitate DNA processing [46]. We did not observe the release ofthese molecules from mitochondria upon TNF treatment (Fig. 6A, B).Thus, AIF and Endo-G do not appear to be components of TNF-triggered cell death pathway in L929 cells in our experimentalsystem. Similar data has also been observed for L929 ΔTM-BNIP3(not shown).

3.7. TNF-triggered mitochondrial release of Smac/Diablo and Omi/HtrA2is unaffected by ΔTM-BNIP3

To further examine the effect of TNF on mitochondria, wemonitored the TNF-triggered release of Smac/Diablo and Omi/HtrA2from mitochondria. Murine Smac and its human ortholog Diablo areproteolytically processed in mitochondria and released from theintermembrane mitochondrial space upon triggering apoptosis [47].The mammalian serine protease Omi, also known as high-

Fig. 4. TNF induces generation of reactive nitrogen species that are involved in TNF cytotoxicity and BNIP3 expression in L929 cells. (A) L929 cells were treated with TNF (1 ng/ml) forindicated time and NO production was measured using DAF-2A. TNF induced a significant increase in NO production (Pb0.05). The data represent triplicates of three independentexperiments. (B) NO-synthase inhibitor interfered with TNF-induced cell death. Pre-treatment of L929 cells for 1 h with L-NMMA (500, 2500, 5000 nM) reduced TNF cytotoxicity(TNF: 1 ng/ml for 24 h) as measured byMTTassay. Data represent values from quadruplicates of four independent experiments (Pb0.05). Cells treated with TNF solvent PBS served ascontrol. (C) Western blot analysis of BNIP3 expression in L929 cells upon pre-treatment with L-NMMA (500 nM for 1 h), with treatment with 1 ng/ml TNF for indicated time (rightpanel). Left panel, that represents the treatment with TNF alone served as a control. GAPDH was included as loading control. In each sets of data, cells treated with TNF solvent PBSand L-NMMA (500 nM and pre-treated for 1 h) served as an internal control “C”. (D) Quantification of BNIP3 expression in TNF treated cells pretreated with L-NMMA. NO synthaseinhibitor shifted the 30 kD/60 kD BNIP3 ratio toward the 60 kD BNIP3 form, whereas total BNIP3 expression remained unchanged. Data were adjusted to the “C” control signal andrepresent triplicates of three independent experiments.

553S. Ghavami et al. / Biochimica et Biophysica Acta 1793 (2009) 546–560

temperature requirement A2 (HtrA2), is released from the mito-chondria to the cytosol during apoptosis where it contributes bothto caspase-dependent and caspase-independent apoptosis [48].Both Smac/Diablo and Omi/HtrA2 bind and antagonize the actionsof inhibitors of apoptosis (IAPs) [49]. L929 cells were treated withTNF (1 ng/ml, 12 h) and the subcellular location of Smac/Diablo andOmi/HtrA2 was examined by confocal imaging (Fig. 6C, D).Immunoreactive Smac/Diablo and Omi/HtrA2 were present inmitochondria of cells treated with medium alone (Fig. 6C, D,control panel). Upon TNF-treatment, these proteins were releasedfrom the mitochondria to the cytosol (Fig. 6C, D) indicating thatthese proteins are involved in TNF-induced cell death. Over-expression of ΔTM-BNIP3 did not significantly inhibit the TNF-mediated release of Smac/Diablo and Omi/HtrA2 (Fig. 6C, D). Thismight be explained by the TNF-induced increase in wild-type BNIP3expression, which may compensate for the presence of dominantnegative mutant ΔTM-BNIP3.

3.8. Dominant-negative BNIP3 mutant counteracts the activation oflysosomal death pathway by TNF

Lysosomes are important organelles for the control of caspase-independent cell death. In the event of cell death, lysosomal acidhydrolases of the cathepsin (cath) family frequently translocate fromlysosomes into the cytosol. In some experimental systems cath-B andcath-D are rate limiting for death induced by interferon-γ, TNF, p53, orpro-oxidants [50]. The vacuolar H+-ATPase inhibitor bafilomycin A1(Baf A1; 0.05 μM), the irreversible inhibitor of Cath-B and cath-L zFF-fmk (100 μM), and the cath-B inhibitor CA-074-Me (10 μM), partiallyinhibited TNF-induced cell death in L929 and L929-ΔTM-BNIP3 cells(Fig. 7A–C) (Pb0.05). Staining of cells with the acidophilic lysosomalprobe LTR revealed that TNF caused an increase in lysosomal volumein L929 cells (Fig. 7D), whichwas absent in L929-ΔTM-BNIP3 cells (Fig.7E). Pre-treatment with bafilomycin A1, zFF-fmk, CA-074-Me, orantioxidant N-acetyl-L-cysteine (5 mM) abolished the differences in

Fig. 5. The over-expression of dominant-negative form of BNIP3 decreased TNF-induced ROS production and Δψm. (A) ΔTM-BNIP3 over-expression partly protected from TNF-triggered Δψm as shown by cytofluorimetric analysis of ψm in L929 (upper panel) and a derivative that over-expresses ΔTM-BNIP3 (lower panel). Cells were treated for 12 h withmedium alone (left diagrams), or with TNF (1 ng/ml). TNF treatment triggered marked changes in mitochondrial membrane potential that were partially abolished in ΔTM-BNIP3over-expressing cells. A representative experiment of four is shown. (B, C) In L929 (B) TNF (1 ng/ml) triggered marked ROS production, whereas in (C) L929-ΔTM-BNIP3 cells TNFtreatment elicited only minimal (∼4–15%) induction of ROS (Pb0.05). Cells were treated with TNF for the indicated time periods and ROS was measured using dihydrorhodamine-123. Experiments were repeated four times and mean ROS values are shown. (D) Broad-spectrum anti-oxidant, NAC, partially inhibited TNF induced cell death. L929 cells were pre-treated with indicated concentrations of NAC for 3 h and than co-treated with TNF (1 ng/ml) and indicated concentration of NAC for 24 h. NAC partially inhibited TNF induced celldeath (Pb0.05). The data represents triplicates of three independent experiments.

554 S. Ghavami et al. / Biochimica et Biophysica Acta 1793 (2009) 546–560

555S. Ghavami et al. / Biochimica et Biophysica Acta 1793 (2009) 546–560

lysosomal volume between L929 and L929-ΔTM-BNIP3 cells inresponse to TNF treatment (data not shown). Thus, the higher levelsof ROS observed in L929 as compared to L929-ΔTM-BNIP3 cellsappeared to contribute to lysosomal damage in TNF-treated L929 cellsand this may involve the release of cathepsins into the cytoplasm[51,52].

4. Discussion

In the present study, we have investigated the role of the pro-apoptotic Bcl2-family member BNIP3 in TNF-induced cell death. Weemployed the rodent fibrosarcoma cell line L929, which has beenwidely used to study the mechanisms of TNF cytotoxicity. TNF inducescell death in L929, which may resemble necrosis or apoptosis [50,53].BNIP3 is a pro-death member of the Bcl2-family of apoptotic proteins[54]. Its mechanism of action is not fully understood. It is acceptedthat, at least in part, it exerts its activity by inserting into themitochondrial membrane. This leads to the opening of the mitochon-drial permeability pore with resultant membrane depolarization and

Fig. 6. TNF did not affect themitochondrial release of caspase-independent apoptosis inducingrelease of caspase-dependent apoptosis inducing proteins (Smac/DIABLO, Omi/HtrA2). (A, Bapoptoticmitochondrial proteins AIF and Endo-G from L929. Cellular localization of AIF (A) an(1 ng/ml) for 12 h prior to immunostaining with anti-AIF and anti-Endo-G antibodies folcounterstained with DAPI (blue) and mitotracker (red) to visualize nuclei and mitochondriaapoptotic mitochondrial proteins Omi/HtrA2 and Smac/Diablo in L929 and L929-ΔTM-BNmitochondrial proteins Omi/HtrA2 and Smac/Diablo. Cellular localization of Omi/HtrA2 (microscopy. Cells were treatedwith 1 ng/ml TNF for 12 h prior to immunostaining with anti-Osecondary antibody (green). Cells were counterstainedwith DAPI (blue) andmitotracker (red1000×. Representative micrographs are shown for all samples.

subsequent ROS generation [35]. This rapid and profound mitochon-drial dysfunction does not involve the release of mitochondrialcytochrome c, caspase activation, or the nuclear translocation of AIF[24]. The resulting cell death is accompanied by an increase in plasmamembrane permeability [34]. To address the role of BNIP3 in TNF-induced cell death, we employed wild-type BNIP3 and a deletionmutant of BNIP3 which lacks the transmembrane domain (ΔTM-BNIP3) known to be important in BNIP3-induced cell death [55]. TNF-stimulated cell death in L929 was BNIP3-dependent and the presenceof ΔTM-BNIP3 resulted in significantly higher TNF-resistance. BNIP3-mediated cell death is independent of caspase activation and therelease of cytochrome c from mitochondria [24]. Likewise, caspase-3and -9 activity and cytochrome c release were similar in the presenceof wild type BNIP3 and ΔTM-BNIP3 in L929. Thus, TNF triggeredcaspase-independent cell death, which was partially inhibited byΔTM-BNIP3. Some Bcl2-like proteins contain a single C-terminal TMhelix to facilitate their membrane targeting [56]. Upon induction ofcell death, BNIP3 integrates into the mitochondrial outer membranewith the N-terminus (residues 1–162) pointing into the cytoplasm and

proteins (AIF, Endo-G) andΔTM-BNIP3 over-expression did not affect themitochondrial) TNF-treatment (in the absence of actinomycin D) did not trigger the release of pro-d Endo-G (B) were determined by confocal microscopy. L929 cells were treatedwith TNFlowed by detection with Cy-5-conjugated secondary antibody (magenta). Cells were. (C, D) TNF-treatment (in the absence of actinomycin D) triggered the release of pro-IP3 cells. ΔTM-BNIP3 over-expression did not prevent the release of pro-apoptotic

C) and Smac/Diablo (D) in L929 and L929-ΔTM-BNIP3 cells was probed by confocalmi/HtrA2 and anti-Smac/Diablo antibodies, followed by detectionwith FITC-conjugated). In control experiments, cells were treatedwith TNF solvent PBS. Themagnificationwas

Fig. 6 (continued).

556 S. Ghavami et al. / Biochimica et Biophysica Acta 1793 (2009) 546–560

Fig. 7. ΔTM-BNIP3 over-expression counteracts TNF-triggered activation of lysosomal cell death pathway. Pretreatment for 1 h with the inhibitors of lysosomal death pathway (A) BafA1 (0.05 μM), (B) z-FF-fmk (100 μM), and (C) CA-074-Me (20 μM) reduced TNF-cytotoxicity (TNF at 1 ng/ml for 24 h) in L929 but not in L929-ΔTM-BNIP3. L929-ΔTM-BNIP3 cells werepartly protected by the expression of the dominant-negative mutant of BNIP3. TNF cytotoxicity was assessed byMTTassay. Data represent average values from quadruplicates of fourindependent experiments. In control experiment, cells were either treated with TNF solvent (PBS) or with the inhibitors alone at the same concentration as used in the experiments.(D, E) TNF triggered swelling of lysosomes in L929, but not in L929-ΔTM-BNIP3. L929 (D) and L929-ΔTM-BNIP3 cells (E) were treated with TNF (1 ng/ml for 24 h) in the absence ofactinomycin D and stained with the acidophilic lysosomal probe LysoTracker Red (LTR). TNF caused an increase in volume and frequency of cytoplasmic granule staining with LTRwhich was partially inhibited by ΔTM-BNIP3 over-expression (E). Nuclei were visualized by DAPI (blue). In control experiment, cells were treated with PBS. The magnificationwas 1000×. Representative micrographs are shown for all samples.

557S. Ghavami et al. / Biochimica et Biophysica Acta 1793 (2009) 546–560

the C terminus (163–185) projecting inwards [24]. Delineating themolecular mechanisms of BNIP3 actions is very important. It has beenfound that both the pro-apoptotic activity of this protein and itsinteraction with other members of the Bcl2 family is sensitive to thepresence of its TM segment [57,58]. The functional importance of TMsegment was also confirmed by the fact that the enforced expressionof a BNIP3 mutant lacking this domain partially blocks hypoxia-

induced cell death [58]. The results of our study also confirm theimportance of TM domain in BNIP3 induced cell death.

Reactive nitrogen intermediates, including NO, NO2− and NO3−, areinvolved in TNF-mediated cytolysis [38] but the mechanism remainedunknown. Here we show that NO intermediates can promote thetranscriptional upregulation of BNIP3. This BNIP3 gene activation isabolished after treatment with the nitrogen oxide synthase inhibitor

558 S. Ghavami et al. / Biochimica et Biophysica Acta 1793 (2009) 546–560

(L-NMMA) in L929. Thus, we concluded that nitric oxide participatesin TNF-induced cytotoxicity by the upregulation of the expression ofBNIP3. The increased level and cellular translocation of BNIP3 from thenucleus to mitochondria may then trigger mitochondrial dysfunctionfrequently observed upon TNF-treatment. L929 cells over-expressingΔTM-BNIP3 were partially protected from TNF-induced changes toΨm and this coincided with higher residual Ψm in L929-ΔTM-BNIP3than observed in parental L929 cells. Thus, our finding providedfurther evidence for a role of TNF and BNIP3 in mitochondrialmalfunction as part of the TNF-induced cell death.

Mitchell's chemi-osmotic theory [59] identifiedΔΨm as the drivingforce for mitochondrial synthesis of ATP, an essential component forcell viability and the execution of apoptosis. Ψm vary depending onthe physiological status of cells and is increased in activated cells andtumor cells [60,61]. This suggests that higher Ψm, as observed withL929-ΔTM-BNIP3 cells, may contribute to increased resistance againstapoptosis. TNF treatment causes a rapid and significant ATP depletion[12,55,62]. The TNF-inducedmitochondrial translocation of BNIP3 andthe subsequent loss of mitochondrial membrane potential may, inpart, explain this effect on cellular ATP levels and stresses theimportance of the TM of BNIP3 for its pro-apoptotic function. Bycontrast,ΔTM-BNIP3 over-expression partially attenuated the effect ofTNF on ΔΨm suggesting that this deletion variant may act as a partialantagonist of BNIP3.

Apoptosis involves the synchronized release [63] with similarkinetics [64] of cytochrome c, Smac/Diablo, and Omi/HtrA2. This wasalso observed in TNF-treated L929 and L929-ΔTM-BNIP3. Theseproteins reside in the mitochondrial inter-membrane space whereasAIF is anchored in the inner membrane [65], and Endo-G likely mainlylocalized in the matrix [42]. This would explain the lack of AIF andEndo-G in TNF-induced cell death in our experimental model.Insertion of BNIP3 into the mitochondrial outer membrane andchanges inΨm may participate in the release of Smac/Diablo and Omi/HtrA2 during TNF-induced cell death. Indirect evidence in support ofthis interpretation came from L929 over-expressing ΔTM-BNIP3.When treated with TNF, these cells showed similar levels of releasefor cytochrome c, Smac/Diablo, and Omi/HtrA2 as in L929, emphasiz-ing the functional importance of the BNIP3 TM domain, and theinability of ΔTM-BNIP3 to interfere with this release.

Over-expression of ΔTM-BNIP3 counteracted lysosomal damageduring TNF-mediated cell death. Loss of lysosomal integrity withsubsequent activation of cell death cascades has been implicated inapoptosis during oxidative stress [59,66], growth factor starvation, Fasactivation, α-tocopheryl-succinate-mediated apoptosis in Jurkat T-cells [60,61], 6-hydroxydopamine-associated death of cultured micro-glia [67], apoptosis induced by the synthetic retinoid CD437 in humanleukemia HL-60 cells [68], TNF- and bile acid-mediated hepatocyteapoptosis [69,70], and p53-mediated apoptosis of M1-t-p53 myeloidleukemic cells [71]. Recently, TNF-cytotoxicity was shown to associatewith the permeabilization of lysosomes, release of cath-B into thecytosol [72], and subsequent cath-B-induced activation of themitochondrial death pathway [69,73]. TNF-mediated apoptosis ismarkedly attenuated by alkalinizing acidic vesicles, employing cath-Binhibitors, or in cath-B knockout mice [73,74]. Furthermore, cath-Bdeficient mice are resistant to TNF-mediated liver injury [75]. Weshow here that pre-treatment of L929 cells with the vacuolar H+-ATPase inhibitor bafilomycin A1 and the inhibitors of cath-B/cath-LzFF-fmk and CA-074-Me attenuates the enhanced TNF-mediated celldeath in L929. Moreover, TNF-induced enlargement of lysosomalvolume was inhibited by ΔTM-BNIP3 over-expression and coincidedwith decreased levels of ROS in L929-ΔTM-BNIP3 cells. Alterations inmitochondrial structure and function, as observed with TNF-treatedL929, and subsequent ROS formation seem to be an important step inTNF cytotoxicity [75]. Inhibition of ROS by pre-treatment with NACabolished the increase in lysosomal volume in TNF-exposed L929.Thus, BNIP3 appears to be involved in the activation of the lysosomal

pathway in TNF-mediated cell death. In part, the actions of BNIP3 aremediated indirectly via increased production of ROS as a result ofΔΨm. ROS act as locally active messenger and signal mitochondrialdamage to lysosomes which respond by swelling and cathepsinrelease, thereby, executing cell death [76].

In conclusion, we show that BNIP3 plays an important role in TNF-induced cell death. TNF induced an NO-mediated increase in BNIP3expression and the translocation of BNIP3 to mitochondria, whichresulted in perturbed mitochondrial function and increased ROSproduction. Both effects were antagonized by over-expression ofΔTM-BNIP3 and consequently ΔTM-BNIP3 protected against lysoso-mal activation in TNF-treated L929 cells. Our data provide a functionalbasis and link between BNIP3, a mediator of atypical cell death, andTNF, known for decades to induce amixed necro-apoptotic form of celldemise. We used TNF in the absence of the transcriptional inhibitoractinomycin D, which is frequently employed among TNF-researchers.This allowed for the detection of TNF-induced changes in geneexpression, which let to our discovery of TNF-induced and NO-mediated upregulation of BNIP3 as another mechanism contributingto TNF-induced cell death.

Acknowledgements

S.G. thankfully acknowledges fellowships from MHRC and CCMF,MICH, NTPAA, and CLA/GSK/CIHR. M.E. is thankful for fellowshipsfrom CCMF and MICH. G.H.B. is thankful for the support through theBSc-Med program of the Univ. Manitoba. M.L. thankfully acknowl-edges support through CFI-Canada Research Chair program, MHRC-,CIHR, and MICH-founded programs. S.W. acknowledges support fromDFG (We 1801/2-4, GRK1302, SFB 685); the German Federal Ministryof Education, Science, Research and Technology (Hep-Net); the IZKFTübingen (IZKF; Fö. 01KS9602); the Wilhelm Sander-Stiftung(2004.099.1); and the Landesforschungsschwerpunktprogramm ofthe Ministry of Science, Research and Arts of the Land Baden-Wuerttemberg.

References

[1] G. Brouckaert, M. Kalai, X. Saelens, P. Vandenabeele, Apoptotic pathways and theirregulation, In: M. Los, S.B. Gibson (Eds.), Apoptotic Pathways as Target for NovelTherapies in Cancer and Other Diseases, Springer Academic Press, New York, 2005.

[2] L. Cherlonneix, From life without death to permanent delaying of death in life (Dela vie sans mort à la Vie en suspens), Rev. Hist. Épistémol. Sci. de. la. Vie. 11 (2004)139–161.

[3] I. Rashedi, S. Panigrahi, P. Ezzati, S. Ghavami,M. Los, Autoimmunity and apoptosis—therapeutic implications, Curr. Med. Chem. 14 (2007) 3139–3151.

[4] M. Jaattela, J. Tschopp, Caspase-independent cell death in T lymphocytes, Nat.Immunol. 4 (2003) 416–423.

[5] N. Festjens, T. Vanden Berghe, P. Vandenabeele, Necrosis, a well-orchestrated formof cell demise: signalling cascades, important mediators and concomitantimmune response, Biochim. Biophys. Acta 1757 (2006) 1371–1387.

[6] E.A. Carswell, L.J. Old, R.L. Kassel, S. Green, N. Fiore, B. Williamson, An endotoxin-induced serum factor that causes necrosis of tumors, Proc. Natl. Acad. Sci. U. S. A.72 (1975) 3666–3670.

[7] L.J. Old, Tumor necrosis factor (TNF), Science 230 (1985) 630–632.[8] B. Beutler, A. Cerami, Cachectin and tumour necrosis factor as two sides of the

same biological coin, Nature 320 (1986) 584–588.[9] B. Beutler, A. Cerami, Tumor necrosis, cachexia, shock, and inflammation: a

common mediator, Annu. Rev. Biochem. 57 (1988) 505–518.[10] R. Beyaert, W. Fiers, Molecular mechanisms of tumor necrosis factor-induced

cytotoxicity. What we do understand and what we do not, FEBS Lett. 340 (1994)9–16.

[11] W. Fiers, R. Beyaert, W. Declercq, P. Vandenabeele, More than one way to die:apoptosis, necrosis and reactive oxygen damage, Oncogene 18 (1999) 7719–7730.

[12] M. Los, M. Mozoluk, D. Ferrari, A. Stepczynska, C. Stroh, A. Renz, Z. Herceg, Z.-Q.Wang, K. Schulze-Osthoff, Activation and caspase-mediated inhibition of PARP: amolecular switch between fibroblast necrosis and apoptosis in death receptorsignaling, Mol. Biol. Cell 13 (2002) 978–988.

[13] A. Cauwels, B. Janssen, A. Waeytens, C. Cuvelier, P. Brouckaert, Caspase inhibitioncauses hyperacute tumor necrosis factor-induced shock via oxidative stress andphospholipase A2, Nat. Immunol. 4 (2003) 387–393.

[14] D.T. Humphreys, M.R. Wilson, Modes of L929 cell death induced by TNF-alpha andother cytotoxic agents, Cytokine 11 (1999) 773–782.

[15] J. Van Lint, P. Agostinis, V. Vandevoorde, G. Haegeman, W. Fiers, W. Merlevede, J.R.Vandenheede, Tumor necrosis factor stimulates multiple serine/threonine protein

559S. Ghavami et al. / Biochimica et Biophysica Acta 1793 (2009) 546–560

kinases in Swiss 3T3 and L929 cells. Implication of casein kinase-2 andextracellular signal-regulated kinases in the tumor necrosis factor signaltransduction pathway, J. Biol. Chem. 267 (1992) 25916–25921.

[16] K. Schulze-Osthoff, A.C. Bakker, B. Vanhaesebroeck, R. Beyaert, W.A. Jacob, W.Fiers, Cytotoxic activity of tumor necrosis factor is mediated by early damage ofmitochondrial functions. Evidence for the involvement of mitochondrial radicalgeneration, J. Biol. Chem. 267 (1992) 5317–5323.

[17] D.R. Green, Apoptotic pathways: the roads to ruin, Cell 94 (1998) 695–698.[18] H. Li, H. Zhu, C.J. Xu, J. Yuan, Cleavage of BID by caspase 8 mediates the

mitochondrial damage in the Fas pathway of apoptosis, Cell 94 (1998) 491–501.[19] C. Scaffidi, S. Fulda, A. Srinivasan, C. Friesen, F. Li, K.J. Tomaselli, K.M. Debatin, P.H.

Krammer, M.E. Peter, Two CD95 (APO-1/Fas) signaling pathways, EMBO J. 17(1998) 1675–1687.

[20] J.M. Boyd, S. Malstrom, T. Subramanian, L.K. Venkatesh, U. Schaeper, B. Elangovan,C. D'Sa-Eipper, G. Chinnadurai, Adenovirus E1B 19 kD and Bcl-2 proteins interactwith a common set of cellular proteins, Cell 79 (1994) 341–351.

[21] V.S. Marsden, A. Strasser, Control of apoptosis in the immune system: Bcl-2, BH3-only proteins and more, Annu. Rev. Immunol. 21 (2003) 71–105.

[22] G. Chen, R. Ray, D. Dubik, L. Shi, J. Cizeau, R.C. Bleackley, S. Saxena, R.D. Gietz, A.H.Greenberg, The E1B 19K/Bcl-2-binding protein Nip3 is a dimeric mitochondrialprotein that activates apoptosis, J. Exp. Med. 186 (1997) 1975–1983.

[23] J.Y. Kim, J.J. Cho, J. Ha, J.H. Park, The carboxy terminal C-tail of BNip3 is crucial ininduction of mitochondrial permeability transition in isolatedmitochondria, Arch.Biochem. Biophys. 398 (2002) 147–152.

[24] C. Vande Velde, J. Cizeau, D. Dubik, J. Alimonti, T. Brown, S. Israels, R. Hakem, A.H.Greenberg, BNIP3 and genetic control of necrosis-like cell death through themitochondrial permeability transition pore, Mol. Cell. Biol. 20 (2000) 5454–5468.

[25] S. Ghavami, C. Kerkhoff, M. Los, M. Hashemi, C. Sorg, F. Karami-Tehrani,Mechanism of apoptosis induced by S100A8/A9 in colon cancer cell lines: therole of ROS and the effect of metal ions, J. Leukoc. Biol. 76 (2004) 169–175.

[26] S. Ghavami, C. Kerkhoff,W.J. Chazin, K. Kadkhoda,W. Xiao, A. Zuse, M. Hashemi, M.Eshraghi, K. Schulze-Osthoff, T. Klonisch, M. Los, S100A8/9 induces cell death via anovel, RAGE-independent pathway that involves selective release of Smac/DIABLOand Omi/HtrA2, Biochim. Biophys. Acta 1783 (2008) 297–311.

[27] M. Hashemi, S. Ghavami, M. Eshraghi, E.P. Booy, M. Los, Cytotoxic effects of intra-and extracellular zinc chelation on human breast cancer cells, Eur. J. Pharm. 557(2007) 9–19.

[28] S. Ghavami, A. Asoodeh, T. Klonisch, A.J. Halayko, K. Kadkhoda, T.J. Kroczak, S.B.Gibson, E.P. Booy, H. Naderi-Manesh, M. Los, Brevinin-2R(1) semi-selectively killscancer cells by a distinct mechanism, which involves the lysosomal-mitochondrialdeath pathway, J. Cell. Mol. Med. 12 (2008) 1005–1022.

[29] C. Adrain, E.M. Creagh, S.J. Martin, Apoptosis-associated release of Smac/DIABLOfrom mitochondria requires active caspases and is blocked by Bcl-2, EMBO J. 20(2001) 6627–6636.

[30] S. Maddika, S.R. Ande, E. Wiechec, L.L. Hansen, S. Wesselborg, M. Los, Akt-mediated phosphorylation of CDK2 regulates its dual role in cell cycle progressionand apoptosis, J. Cell. Sci. 121 (2008) 979–988.

[31] P. Sharma, T. Tran, G.L. Stelmack, K. McNeill, R. Gosens, M.M. Mutawe, H. Unruh,W.T. Gerthoffer, A.J. Halayko, Expression of the dystrophin-glycoprotein complex is amarker for human airway smooth muscle phenotype maturation, Am. J. Physiol.,Lung Cell. Mol. Physiol. 294 (2008) L57–L68.

[32] M.O. Lopez-Figueroa, C. Caamano, M.I. Morano, L.C. Ronn, H. Akil, S.J. Watson,Direct evidence of nitric oxide presence within mitochondria, Biochem. Biophys.Res. Commun. 272 (2000) 129–133.

[33] T. Imazu, S. Shimizu, S. Tagami, M. Matsushima, Y. Nakamura, T. Miki, A. Okuyama,Y. Tsujimoto, Bcl-2/E1B 19 kD-interacting protein 3-like protein (Bnip3L) interactswith bcl-2/Bcl-xL and induces apoptosis by altering mitochondrial membranepermeability, Oncogene 18 (1999) 4523–4529.

[34] R. Ray, G. Chen, C. Vande Velde, J. Cizeau, J.H. Park, J.C. Reed, R.D. Gietz, A.H.Greenberg, BNIP3 heterodimerizes with Bcl-2/Bcl-X(L) and induces cell deathindependent of a Bcl-2 homology 3 (BH3) domain at both mitochondrial andnonmitochondrial sites, J. Biol. Chem. 275 (2000) 1439–1448.

[35] L. Lamy, M. Ticchioni, A.K. Rouquette-Jazdanian, M. Samson, M. Deckert, A.H.Greenberg, A. Bernard, CD47 and the 19 kD interacting protein-3 (BNIP3) in T cellapoptosis, J. Biol. Chem. 278 (2003) 23915–23921.

[36] Y.H. Yook, K.H. Kang, O. Maeng, T.R. Kim, J.O. Lee, K.I. Kang, Y.S. Kim, S.G. Paik, H.Lee, Nitric oxide induces BNIP3 expression that causes cell death in macrophages,Biochem. Biophys. Res. Commun. 321 (2004) 298–305.

[37] R. Zamora, Y. Vodovotz, B. Betten, C. Wong, B. Zuckerbraun, K.F. Gibson, H.R. Ford,Intestinal and hepatic expression of BNIP3 in necrotizing enterocolitis: regulationby nitric oxide and peroxynitrite, Am. J. Physiol. Gastrointest. Liver Physiol. 289(2005) G822–G830.

[38] S. Hauschildt, P. Scheipers, W.G. Bessler, A. Mulsch, Induction of nitric oxidesynthase in L929 cells by tumour-necrosis factor alpha is prevented by inhibitorsof poly(ADP-ribose) polymerase, Biochem. J. 288 (Pt 1) (1992) 255–260.

[39] M. Salter, R.G. Knowles, S. Moncada, Widespread tissue distribution, speciesdistribution and changes in activity of Ca(2+)-dependent and Ca(2+)-independentnitric oxide synthases, FEBS Lett. 291 (1991) 145–149.

[40] K. De Vos, V. Goossens, E. Boone, D. Vercammen, K. Vancompernolle, P.Vandenabeele, G. Haegeman, W. Fiers, J. Grooten, The 55-kD tumor necrosisfactor receptor induces clustering of mitochondria through its membrane-proximal region, J. Biol. Chem. 273 (1998) 9673–9680.

[41] S. Ko, T.T. Kwok, K.P. Fung, Y.M. Choy, C.Y. Lee, S.K. Kong, Tumour necrosis factorinduced an early release of superoxide and a late mitochondrial membranedepolarization in L929 cells. Increase in the production of superoxide is notsufficient to mimic the action of TNF, Biol. Signals Recept. 10 (2001) 326–335.

[42] A.D. Uboldi, N. Savage, The adenylate cyclase activator forskolin partially protectsL929 cells against tumour necrosis factor-alpha-mediated cytotoxicity via a cAMP-independent mechanism, Cytokine 19 (2002) 250–258.

[43] N. Joza, S.A. Susin, E. Daugas, W.L. Stanford, S.K. Cho, C.Y. Li, T. Sasaki, A.J. Elia, H.Y.Cheng, L. Ravagnan, K.F. Ferri, N. Zamzami, A. Wakeham, R. Hakem, H. Yoshida, Y.Y.Kong, T.W. Mak, J.C. Zuniga-Pflucker, G. Kroemer, J.M. Penninger, Essential role ofthe mitochondrial apoptosis-inducing factor in programmed cell death, Nature410 (2001) 549–554.

[44] H. Ye, C. Cande, N.C. Stephanou, S. Jiang, S. Gurbuxani, N. Larochette, E. Daugas, C.Garrido, G. Kroemer, H. Wu, DNA binding is required for the apoptogenic action ofapoptosis inducing factor, Nat. Struct. Biol. 9 (2002) 680–684.

[45] L.Y. Li, X. Luo, X. Wang, Endonuclease G is an apoptotic DNase when released frommitochondria, Nature 412 (2001) 95–99.

[46] P. Widlak, W.T. Garrard, Ionic and cofactor requirements for the activity of theapoptotic endonuclease DFF40/CAD, Mol. Cell. Biochem. 218 (2001) 125–130.

[47] A.M. Verhagen, P.G. Ekert, M. Pakusch, J. Silke, L.M. Connolly, G.E. Reid, R.L. Moritz,R.J. Simpson, D.L. Vaux, Identification of DIABLO, a mammalian protein thatpromotes apoptosis by binding to and antagonizing IAP proteins, Cell 102 (2000)43–53.

[48] G. van Loo, M. van Gurp, B. Depuydt, S.M. Srinivasula, I. Rodriguez, E.S. Alnemri, K.Gevaert, J. Vandekerckhove, W. Declercq, P. Vandenabeele, The serine proteaseOmi/HtrA2 is released from mitochondria during apoptosis. Omi interacts withcaspase-inhibitor XIAP and induces enhanced caspase activity, Cell Death Differ. 9(2002) 20–26.

[49] M. van Gurp, N. Festjens, G. van Loo, X. Saelens, P. Vandenabeele, Mitochondrialintermembrane proteins in cell death, Biochem. Biophys. Res. Commun. 304(2003) 487–497.

[50] V. Goossens, K. De Vos, D. Vercammen, M. Steemans, K. Vancompernolle, W. Fiers,P. Vandenabeele, J. Grooten, Redox regulation of TNF signaling, Biofactors 10(1999) 145–156.

[51] K. Ollinger, U.T. Brunk, Cellular injury induced by oxidative stress is mediatedthrough lysosomal damage, Free Radic. Biol. Med. 19 (1995) 565–574.

[52] K. Roberg, U. Johansson, K. Ollinger, Lysosomal release of cathepsin D precedesrelocation of cytochrome c and loss of mitochondrial transmembrane potentialduring apoptosis induced by oxidative stress, Free Radic. Biol. Med. 27 (1999)1228–1237.

[53] V. Goossens, G. Stange, K. Moens, D. Pipeleers, J. Grooten, Regulation of tumornecrosis factor-induced, mitochondria- and reactive oxygen species-dependentcell death by the electron flux through the electron transport chain complex I,Antioxid. Redox Signal. 1 (1999) 285–295.

[54] S. Kothari, J. Cizeau, E. McMillan-Ward, S.J. Israels, M. Bailes, K. Ens, L.A.Kirshenbaum, S.B. Gibson, BNIP3 plays a role in hypoxic cell death in humanepithelial cells that is inhibited by growth factors EGF and IGF, Oncogene 22(2003) 4734–4744.

[55] M. Yasuda, P. Theodorakis, T. Subramanian, G. Chinnadurai, Adenovirus E1B-19K/BCL-2 interacting protein BNIP3 contains a BH3 domain and a mitochondrialtargeting sequence, J. Biol. Chem. 273 (1998) 12415–12421.

[56] M.J. Pinkoski, N.J. Waterhouse, D.R. Green, Mitochondria, apoptosis andautoimmunity, Curr. Dir. Autoimmun. 9 (2006) 55–73.

[57] M.G. Hinds, M. Lackmann, G.L. Skea, P.J. Harrison, D.C. Huang, C.L. Day, Thestructure of Bcl-w reveals a role for the C-terminal residues in modulatingbiological activity, EMBO J. 22 (2003) 1497–1507.

[58] Y.A. Vereshaga, P.E. Volynsky, J.E. Pustovalova, D.E. Nolde, A.S. Arseniev, R.G.Efremov, Specificity of helix packing in transmembrane dimer of the cell deathfactor BNIP3: a molecular modeling study, Proteins 69 (2007) 309–325.

[59] P. Mitchell, J. Moyle, Chemiosmotic hypothesis of oxidative phosphorylation,Nature 213 (1967) 137–139.

[60] S.D. Bernal, T.J. Lampidis, R.M. McIsaac, L.B. Chen, Anticarcinoma activity in vivo ofrhodamine 123, a mitochondrial-specific dye, Science 222 (1983) 169–172.

[61] S. Davis, M.J. Weiss, J.R. Wong, T.J. Lampidis, L.B. Chen, Mitochondrial and plasmamembrane potentials cause unusual accumulation and retention of rhodamine123 by human breast adenocarcinoma-derived MCF-7 cells, J. Biol. Chem. 260(1985) 13844–13850.

[62] I. Ubarretxena-Belandia, D.M. Engelman, Helical membrane proteins: diversity offunctions in the context of simple architecture, Curr. Opin. Struck. Biol. 11 (2001)370–376.

[63] C. Adrain, S.J. Martin, The mitochondrial apoptosome: a killer unleashed by thecytochrome seas, Trends Biochem. Sci. 26 (2001) 390–397.

[64] C. Munoz-Pinedo, A. Guio-Carrion, J.C. Goldstein, P. Fitzgerald, D.D. Newmeyer, D.R. Green, Different mitochondrial intermembrane space proteins are releasedduring apoptosis in a manner that is coordinately initiated but can vary induration, Proc. Natl. Acad. Sci. U. S. A. 103 (2006) 11573–11578.

[65] H. Otera, S. Ohsakaya, Z. Nagaura, N. Ishihara, K. Mihara, Export of mitochondrialAIF in response to proapoptotic stimuli depends on processing at theintermembrane space, EMBO J. 24 (2005) 1375–1386.

[66] D. Arnoult, B. Gaume, M. Karbowski, J.C. Sharpe, F. Cecconi, R.J. Youle,Mitochondrial release of AIF and EndoG requires caspase activation downstreamof Bax/Bak-mediated permeabilization, EMBO J. 22 (2003) 4385–4399.

[67] K. Roberg, K. Ollinger, Oxidative stress causes relocation of the lysosomal enzymecathepsin D with ensuing apoptosis in neonatal rat cardiomyocytes, Am. J. Pathol.152 (1998) 1151–1156.

[68] Y. Zang, R.L. Beard, R.A. Chandraratna, J.X. Kang, Evidence of a lysosomal pathwayfor apoptosis induced by the synthetic retinoid CD437 in human leukemia HL-60cells, Cell Death Differ. 8 (2001) 477–485.

[69] M.E. Guicciardi, J. Deussing, H. Miyoshi, S.F. Bronk, P.A. Svingen, C. Peters, S.H.Kaufmann, G.J. Gores, Cathepsin B contributes to TNF-alpha-mediated hepatocyte

560 S. Ghavami et al. / Biochimica et Biophysica Acta 1793 (2009) 546–560

apoptosis by promoting mitochondrial release of cytochrome c, J. Clin. Invest. 106(2000) 1127–1137.

[70] L.R. Roberts, H. Kurosawa, S.F. Bronk, P.J. Fesmier, L.B. Agellon, W.Y. Leung, F. Mao,G.J. Gores, Cathepsin B contributes to bile salt-induced apoptosis of rathepatocytes, Gastroenterology 113 (1997) 1714–1726.

[71] X.M. Yuan, W. Li, H. Dalen, J. Lotem, R. Kama, L. Sachs, U.T. Brunk, Lysosomaldestabilization in p53-induced apoptosis, Proc. Natl. Acad. Sci. U. S. A. 99 (2002)6286–6291.

[72] L. Foghsgaard, D. Wissing, D. Mauch, U. Lademann, L. Bastholm, M. Boes, F.Elling, M. Leist, M. Jaattela, Cathepsin B acts as a dominant execution proteasein tumor cell apoptosis induced by tumor necrosis factor, J. Cell Biol. 153 (2001)999–1010.

[73] M.E. Guicciardi, H. Miyoshi, S.F. Bronk, G.J. Gores, Cathepsin B knockout mice areresistant to tumor necrosis factor-alpha-mediated hepatocyte apoptosis and liverinjury: implications for therapeutic applications, Am. J. Pathol. 159 (2001)2045–2054.

[74] L. Monney, R. Olivier, I. Otter, B. Jansen, G.G. Poirier, C. Borner, Role of an acidiccompartment in tumor-necrosis-factor-alpha-induced production of ceramide,activation of caspase-3 and apoptosis, Eur. J. Biochem. 251 (1998) 295–303.

[75] K. Schulze-Osthoff, R. Beyaert, V. Vandevoorde, G. Haegeman, W. Fiers, Depletionof the mitochondrial electron transport abrogates the cytotoxic and gene-inductive effects of TNF, EMBO J. 12 (1993) 3095–3104.

[76] K.F. Ferri, G. Kroemer, Organelle-specific initiation of cell death pathways, Nat. CellBiol. 3 (2001) E255–E263.