| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Institut für Immunologie, Universitätsklinikum Schleswig-Holstein Campus Kiel, Kiel, Germany; and
Abteilung Immunologie und Zellbiologie, Forschungszentrum Borstel, Germany
1Correspondence: Institut für Immunologie, Universitätsklinikum Schleswig-Holstein Campus Kiel, Kiel, Germany. E-mail: dadam{at}email.uni-kiel.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: tumor necrosis factor • RIP1 • mitochondria
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. In addition, caspase-independent PCD has been implicated in the homeostasis of an organism by, for instance, contributing to the negative selection of lymphocytes, embryonic removal of interdigital webs, or the death of chondrocytes controlling the longitudinal growth of bones (1)
. In vivo, relevance for caspase-independent PCD has been demonstrated directly for TNF-induced hyperacute shock (2)
and is discussed for the deletion of T cells (1
, 3)
. At the clinical level, characterization of caspase-independent death pathways promises new options to prevent death processes in neurodegenerative disease or to kill tumor cells that have developed strategies to evade apoptotic death signals (1
, 4)
.
The sphingolipid ceramide has been described as one of the first mediators of TNF-induced cell death. However, with regard to apoptosis, the relevance of ceramide—either being an essential inducer or a mere by-product generated in consequence of cell death—has been the topic of intense discussions (e.g., refs 5
, 6
). With regard to caspase-independent PCD, it has been demonstrated that the killing of NB16 neuroblastoma cells by exogenous C2-ceramide is not inhibited by broad-spectrum caspase inhibitors (7)
. Likewise, addition of exogenous ceramide analogs or induction of intracellular ceramide accumulation has been observed in conjunction with caspase-independent PCD in several cell types (8
9
10
11
12
13)
. Except for these studies, however, the role of ceramide has remained largely uninvestigated. Here, we have focused on the function of ceramide in caspase-independent PCD and establish the relevance of ceramide as a general mediator of this form of PCD.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cell culture
L929, NIH3T3, and Jurkat wild-type cells were originally obtained from the American Type Culture Collection (Manassas, VA, USA). Fas-associated death domain protein (FADD)-, RIP1-deficient, and Bcl-2-overexpressing Jurkat cells were kindly provided by J. Blenis, B. Seed, and S. Korsmeyer (all Harvard Medical School, Boston, MA, USA). Cells were maintained in a mixture of Click’s/RPMI 1640 (50/50% v/v) supplemented with 10% v/v FCS, 10 mM glutamine, and 50 µg/mL each of streptomycin and penicillin in a humidified incubator containing 5% w/v CO2. L929 cells overexpressing the full-length acid ceramidase (AC) cDNA have been described (14)
. Lung fibroblasts from acid sphingomyelinase (A-SMase) -deficient and wild-type mice (15)
kindly provided by E. Gulbins (Essen, Germany) were generated essentially as described (16)
except cells were cultivated in DME supplemented with 20% v/v FCS. Lung fibroblasts were immortalized by stable transfection with the SV40 large T antigen expression vector pMSSVLT (17)
. For reconstitution of A-SMase deficiency, 293 cells were transfected with the A-SMase expression construct pRK5-ASM by the calcium phosphate method; 48 h later the culture supernatant was collected, filtered, diluted with fresh medium (1:2), and added to lung fibroblasts for 16 h before stimulation.
Ceramide quantitation
Ceramide was quantitated in duplicate by the diacylglycerol kinase assay essentially as described by Dressler and Kolesnick (18)
following recommendations of Perry and Hannun (19)
. Alternatively, ceramide was quantitated by the charring method following high-performance thin layer chromatography exactly as described (20)
. Autoradiographs and thin layer chromatography plates were scanned and analyzed using the software package PCBAS (Raytest).
Cytotoxicity assays
For flow cytometric measurement of cell death, cells were seeded in 6-well plates at 5 x 105 cells/well. After treatment detached and adherent cells were collected, followed by centrifugation. The cells were resuspended in PBS/5 mM EDTA containing 2 µg/mL propidium iodide (PI), and the red fluorescence was measured on a FACSCalibur flow cytometer (BD Biosciences, San Diego, CA, USA). Immortalized lung fibroblasts were alternatively analyzed by staining with crystal violet as follows: 104 cells were seeded in flat-bottomed 96-well plates. After stimulation, adherent cells were washed twice with PBS and incubated for 10 min at 37°C in 50 µL of staining solution (0.5% w/v crystal violet, 4% w/v formaldehyde, 30% v/v ethanol, and 0.17% w/v NaCl). The staining solution was washed away with tap water and cells were dried for 1 h at 50°C. Stained cells were dissolved in acetic acid (33% v/v) and the intensity of the staining was colorimetrically determined at 570 nm in a microplate reader (Tecan, Medford, MA, USA).
RNA interference
The predesigned siRNAs Smpd1_1, Smpd1_2, and Smpd1_3 specific for murine A-SMase, the predesigned siRNAs Ripk1_1, Ripk1_2, Ripk1_3, and the custom-made siRNA Ripk1_4 (target sequence 5'-CCACUAGUCUGACUGAUG-3') specific for murine RIP1 as well as the negative control siRNA were obtained from Ambion (Austin, TX, USA). To target expression, 106 cells were transfected with 150 pmol of siRNA by Amaxa nucleofection, using solution V and program T-20 (A-SMase) or solution R and program A-24 (RIP1). The efficiency of the transfection procedure was monitored by nucleofection of pmaxGFP (Amaxa) and routinely yielded 70–90% of green fluorescent cells. Cells were subsequently cultured for 24 to 72 h before further analysis.
Measurement of A-SMase activity
The micellar SMase assay using exogenous radiolabeled sphingomyelin was performed as described (21)
. Briefly, cells were homogenized in 100 µL 0.2% v/v Triton X-100. Radioactive phosphocholine produced from [N-methyl-14C]-sphingomyelin (labeled in the choline moiety, Amersham CFA566) was identified by TLC and routinely determined in the aqueous phase by scintillation counting.
Immunoblots
Cells were collected and lysed in TNE buffer (50 mM Tris pH 8.0, 1% v/v NP-40, 2 mM EDTA) containing 10 µg/mL pepstatin/aprotinin/leupeptin, 1 mM sodium orthovanadate, and 5 mM NaF. 25 µg of cell protein per lane were resolved by electrophoresis on SDS polyacrylamide gels. After electrophoretic transfer to nitrocellulose, reactive proteins were detected using antisera specific for RIP1 (BD Biosciences) and the ECL detection kit (Amersham Biosciences, Arlington Heights, IL, USA). Expression of RIP1 protein was quantified after scanning of the autoradiographs using the software package PCBAS (Raytest, Wilmington, NC, USA).
Measurement of ROS
Cells were seeded at a density of 2 x 105 cells per well and stimulated as indicated, with addition of 1 µM dihydrorhodamine 123 for a total of 1 h. Subsequently, the cells were harvested, washed once with cold PBS, resuspended in PBS, and analyzed by flow cytometry.
Flow cytometric analysis of 
m
Cells were treated as indicated with the addition of 150 nM CMTMRos (Molecular Probes, Eugene, OR, USA) for the last 30 min of stimulation. After harvesting, the cells were washed once in cold PBS, fixed by incubation in paraformaldehyde (4% w/v) for 15 min at room temperature in the dark and analyzed on a FACSCalibur flow cytometer (BD Biosciences).
Analysis of cytochrome crelease
After treatment, cells were harvested at 4°C and resuspended in 1 mL of a buffer containing 20 mM HEPES pH 7.5, 10 mM KCl, 1.5 mM MgCl2, 1 mM EDTA, 0.1 mM phenylmethylsulfonyl fluoride and Complete protease inhibitor mixture (Roche, Nutley, NJ, USA) as recommended by the manufacturer. Cells were lysed by passaging through a 27-gauge needle and homogenates were centrifuged at 1000 x g for 5 min at 4°C. Supernatants were further centrifuged at 10000 x g for 15 min at 4°C. The resulting mitochondrial pellets were resuspended in 40 µL of cold cell lysis buffer (20 mM Tris, pH 7.5, 100 mM NaCl, 1% v/v Triton X-100, 1 mM phenylmethylsulfonyl fluoride, Complete protease inhibitor mixture). The supernatants from the final centrifugation were centrifuged at 6000 x g for 65 min at 4°C using Centricon YM-10 centrifugal filter devices (Millipore, Bedford, MA, USA) to prepare concentrated cytosolic extracts. For detection of mitochondrial release of cytochrome c in immunoblots, 5–20 µg of cytosolic or mitochondrial extract (depending on the cell line used) was analyzed using cytochrome c antibody (BD Biosciences).
Flow cytometric analysis of membrane integrity and phosphatidylserine externalization
106 cells were collected, washed with cold PBS, and resuspended in 100 µL of a buffer (10 mM HEPES pH 7.4, 140 mM NaCl, 5 mM CaCl2) containing 2% v/v of annexin-V-FLUOS labeling reagent (Roche) and 20% v/v 7-AAD reagent (BD Biosciences). The cells were incubated for 20 min in the dark at room temperature and analyzed by flow cytometry using a FACSCalibur Analyzer (BD Biosciences).
Clonogenic survival assays
To analyze the ability of cells to form colonies in soft agarose, after treatment 1000 viable cells (as determined by Trypan blue staining; 10,000 cells for Jurkat wild-type) were mixed with 2 mL of 0.4% Sea Plaque agarose (Cambrex) in complete medium, transferred into 6-well plates containing 3 mL of a 1% SeaKem LE agarose underlayer (Cambrex) in complete medium, and incubated at 37°C. Colony formation (>20 cells) was scored after 14 days of incubation on an inverted Zeiss Axiovert 100 microscope. For this purpose, viable cells were stained by adding 1 mL of 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT; Sigma, St. Louis, MO, USA; 2.5 mg/mL in PBS) and incubation for 2 h at 37°C to allow metabolization of MTT to blue MTT-formazan. For visual representation, NIH3T3 cells were plated in complete medium without agarose as above and stained with crystal violet after 7 days as described under "Cytotoxicity assays," except that all steps subsequent to washing with tap water were omitted.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. In contrast to L929 cells, where caspases are not activated by TNF (Supplementary Fig. 1A), analysis of TNF-induced caspase-independent PCD in most other cell systems requires active inhibition of interfering caspase-dependent responses. Therefore, we initially used the established L929 system to investigate whether ceramide is still crucial for TNF-induced death of these cells in the presence of caspase inhibitors. TNF in combination with the broad-spectrum caspase inhibitor zVAD-fmk induced a much more pronounced increase of intracellular ceramide than TNF alone (Fig. 1
A), corresponding to the much more rapid killing described for TNF/zVAD-treated L929 cells (ref 22
; Fig. 1B
). No caspase activity was detected under these conditions (Supplementary Fig. 1A). Analysis of nuclear morphology (Supplementary Fig. 2A) suggested that PCD induced by TNF/zVAD was "necrosis-like" (1)
.
|
In line with a causative role of ceramide in TNF/zVAD-induced PCD, two independent AC-overexpressing L929 clones (AC23 and AC52) displayed a higher resistance to TNF/zVAD than parental L929 cells. AC degrades ceramide generated by A-SMase in response to TNF, thereby protecting L929 cells against the cytotoxic action of TNF (14)
. Enhanced survival was consistently seen for different concentrations of TNF/zVAD as well as for different incubation times (Fig. 1C
).
Consistent with the above results, the pharmacological agents D609 and desipramine also enhanced resistance against TNF/zVAD in parental L929 cells (Fig. 1D
). D609 indirectly inhibits the TNF-dependent activation of A-SMase and thereby ceramide generation by this enzyme in the lysosomal compartment by impairing the activity of phosphatidylcholine-specific phospholipase C, an enzyme upstream of A-SMase (21
, 23)
. Desipramine causes a rapid and irreversible reduction of A-SMase activity by inducing proteolytic degradation of the enzyme (24)
. In contrast, neither scyphostatin nor fumonisin B1 could protect the cells (Fig. 1D
). Scyphostatin inhibits neutral sphingomyelinase (25)
, an enzyme distinct from A-SMase generating ceramide at the plasma membrane, whereas fumonisin B1 is an inhibitor of ceramide synthase that generates ceramide by de novo synthesis (26)
. With regard to interpreting results obtained with D609, however, a certain caution should be taken since murine and human forms of phosphatidylcholine-specific phospholipase C have not yet been cloned; therefore, evidence that D609 inhibits its target in vitro (i.e., directly) is still missing.
Ceramide accumulation elicits caspase-independent PCD in NIH3T3 fibroblasts and in Jurkat leukemic T cells
After confirming that ceramide (generated by A-SMase) is responsible for the TNF-induced dying of L929 cells not only in the absence but also in the presence of caspase inhibitors, we extended our investigations to other systems of caspase-independent PCD. In both murine NIH3T3 fibroblasts and human leukemic Jurkat T cells, TNF can trigger an alternative route to PCD provided caspase activity is actively inhibited or absent (27
, 28)
. While zVAD-fmk or TNF alone did not alter ceramide levels, corresponding to the inability of either substance to kill NIH3T3 cells when applied individually (27)
, treatment of NIH3T3 cells with TNF/zVAD led to a pronounced increase of intracellular ceramide comparable to results obtained in L929 cells (Fig. 2
A). Concomitantly, the viability of the cells was substantially reduced but could be restored by addition of D609. As in L929 cells, scyphostatin and fumonisin B1 did not protect the cells (Fig. 2B
). Although less pronounced than in L929 cells, changes in nuclear and cellular morphology in response to TNF/zVAD argued for a mode of caspase-independent death resembling necrosis-like PCD (Supplementary Fig. 2B).
|
In Jurkat wild-type T cells, Holler and co-workers have demonstrated that caspase-independent necrosis-like PCD in response to both TNF (see Supplementary Fig. 2C) and Fas ligand is substantially enhanced by addition of the protein biosynthesis inhibitor CHX, and have suggested this is due to the degradation of a short-lived inhibitor that selectively controls caspase-independent PCD in these cells (28)
. As we observed the same increase in TNF/zVAD-induced caspase-independent PCD after addition of CHX (Supplementary Fig. 3A), Jurkat wild-type cells were sensitized for TNF by addition of CHX in subsequent experiments. We verified that any caspase activity was completely suppressed in TNF/CHX/zVAD treated cells, confirming that death indeed occurred by caspase-independent PCD (Supplementary Fig. 1B). As shown in Fig. 2C
, TNF/CHX/zVAD-induced caspase-independent PCD in Jurkat wild-type cells was again accompanied by a pronounced accumulation of intracellular ceramide with a linear increase over time. Ceramide accumulation started well before the onset of cell death (Fig. 2D
), suggesting that ceramide represents a cause rather than a consequence of caspase-independent PCD. Cell death was once more reduced by the addition of D609, but not by fumonisin B1 or scyphostatin (Fig. 2E
).
We analyzed Jurkat cells genetically deficient for FADD (29)
. The absence of FADD prevents caspase activation by TNF (Supplementary Fig. 1C) but simultaneously renders these cells more sensitive to TNF (Supplementary Fig. 3B; ref 28
). Therefore, TNF kills FADD-deficient Jurkat cells exclusively by caspase-independent necrosis-like PCD, similar to L929 cells (Supplementary Fig. 2C). Confirming our results from Jurkat wild-type cells, FADD-deficient Jurkat cells displayed a pronounced linear accumulation of intracellular ceramide in response to TNF alone, again well before the onset of cell death (Fig. 2F, G
). In summary, these data corroborate the concept that TNF-induced intracellular accumulation of ceramide elicits caspase-independent PCD in cells of fibroblastoid as well as T cell origin.
Lung fibroblasts genetically deficient for A-SMase are protected from caspase-independent PCD
Since results obtained with pharmacological inhibitors such as D609 must be interpreted with a certain caution due to potential nonspecific effects, we sought to further substantiate the function of ceramide in caspase-independent PCD by selectively targeting A-SMase expression using RNA interference (RNAi). As shown in Fig. 3
A, 48 h after transfection A-SMase activity in L929 cells was reduced to a minimum of 30%, which was not further diminished by variation in the transfection conditions (e.g., time, Fig. 3A
or amount of siRNA, not shown). However, we did not detect a corresponding inhibition of TNF- or TNF/zVAD-induced caspase-independent PCD (Fig. 3B
). Parallel transfection of NIH3T3 cells likewise did not confer protection (not shown). As one possible explanation for this result, the achieved reduction of A-SMase activity might not yet be sufficient to inhibit the death response. Alternatively, this result might indicate lack of a role for A-SMase and leave the possibility that ceramide is generated by sources other than A-SMase.
|
We therefore immortalized and analyzed lung fibroblasts from A-SMase-deficient mice in a direct genetic approach. As shown in Fig. 3C
, these cells are completely devoid of any residual A-SMase activity. In consequence, treatment with TNF/zVAD induced ceramide accumulation in wild-type, but not in A-SMase-deficient fibroblasts (Fig. 3D
). In viability assays, A-SMase-deficient cells displayed a pronouncedly higher resistance to TNF/zVAD than wild-type cells (Fig. 3E
). Supernatant from cells transfected with an A-SMase expression construct contains secreted A-SMase and can restore A-SMase deficiency when added to cells (30)
. Reconstitution of the A-SMase defect in lung fibroblasts by this method did restore their sensitivity to TNF/zVAD to the level of wild-type cells (Fig. 3F
). In time course experiments, ceramide levels in wild-type cells increased well before the onset of cell death (not shown), again arguing for ceramide being an inducer rather than a consequence of caspase-independent PCD. Taken together, the results obtained in lung fibroblasts provide strong support for a causative role of ceramide in TNF-induced caspase-independent PCD in yet another cell type.
RIP1 function is indispensable for ceramide-induced caspase-independent PCD
RIP1 appears to be essential for TNF-induced caspase-independent PCD of Jurkat cells (28)
. To gain further insight into the signaling pathways by which ceramide mediates caspase-independent PCD, we used Jurkat cells deficient for RIP1 expression (31)
. We confirmed that, in the presence of zVAD-fmk, these cells are completely resistant against TNF/CHX-induced caspase-independent PCD (although they could still be killed by caspase-dependent mechanisms in the absence of zVAD-fmk; Supplementary Fig. 1D, 3C; ref 28
). Correspondingly, these cells showed an exclusively apoptotic morphology (in contrast to the mixed apoptotic/necrosis-like morphology of dying Jurkat wild-type cells or to the solely necrosis-like morphology of dying FADD-deficient cells, Supplementary Fig. 2C). As shown in Fig. 4
A, ceramide accumulation was massively impaired in TNF/CHX/zVAD-treated, RIP1-deficient cells when compared with wild-type Jurkat cells, suggesting that, in Jurkat cells, RIP1 is essential for the induction of ceramide production by TNF and that RIP1 and ceramide mediate their death signals through the same pathway.
|
To extend our analyses of RIP1 beyond the Jurkat T cell line, we neutralized RIP1 function by a pharmacological approach. The inhibitor geldanamycin (GA) has been used in Jurkat cells to limit the function of RIP1 (28)
. GA disrupts the function of the heat-shock protein HSP90 by preventing the release of client proteins such as RIP1 undergoing refolding from HSP90. The stabilized complex is then ubiquitinated and degraded (32)
. Both GA and the structurally unrelated HSP90 inhibitor radicicol (RC) pronouncedly decreased RIP1 protein levels in L929 cells (Supplementary Fig. 4A). Moreover, electrophoretic mobility shift assays revealed a complete block of RIP1-dependent nuclear factor-
B translocation in GA-pretreated cells in response to TNF, confirming the functional inhibition of RIP1 (Supplementary Fig. 4B). In consequence, GA- or RC-treated L929 cells proved completely resistant against TNF/zVAD-induced caspase-independent PCD (Fig. 4B
). Identical results were obtained in NIH3T3 cells or in Jurkat wild-type cells treated with TNF/CHX/zVAD (Fig. 4B
).
To account for potential nonspecific effects of the above inhibitors, we selectively targeted RIP1 expression in NIH3T3 cells by RNAi. As shown in Fig. 4C
, down-regulation of RIP1 protein levels by independent siRNAs specific for RIP1 clearly diminished caspase-independent PCD in response to TNF/zVAD, further corroborating the idea that RIP1 is essential for induction of ceramide-mediated caspase-independent PCD. In summary, the above data suggest that ceramide and RIP1 mediate caspase-independent PCD through signaling pathways that are conserved across different cell types.
Participation of mitochondria in ceramide-mediated caspase-independent PCD
Mitochondria have been implicated as an additional factor in caspase-independent PCD and have been described as targets of ceramide in the induction of cell death (1)
. As shown in Fig. 5
A, Jurkat cells overexpressing the antiapoptotic mitochondrial protein Bcl-2 (33)
clearly displayed a higher resistance to TNF/CHX/zVAD-induced caspase-independent PCD than their parental counterparts. Likewise, overexpression of Bcl-2 can delay the onset of TNF-induced caspase-independent PCD in L929 cells (34)
, suggesting that mitochondria do indeed represent another component of the signaling cascade by which ceramide elicits caspase-independent PCD.
|
With regard to the molecular mechanisms by which mitochondria might elicit caspase-independent PCD, reactive oxygen species (ROS) represent a likely candidate. We therefore investigated the importance of ROS for ceramide-mediated caspase-independent PCD. In line with previous results (22)
, pretreatment with the radical scavenger BHA protected L929 cells from TNF/zVAD-induced caspase-independent PCD, potentially arguing in favor of a causative function of ROS in the ceramide death pathway. However, the structurally closely related radical scavenger BHT did not protect L929 cells (Fig. 5B
). The same results have been observed in NIH3T3 cells (27)
. Moreover, neither BHA nor BHT could protect Jurkat wild-type cells from TNF/CHX/zVAD-induced caspase-independent PCD (Fig. 5B
), although both radical scavengers were equally effective in scavenging ROS in both L929 and Jurkat wild-type cells (Fig. 5C
). Clearly, additional studies are required to fully unravel the mechanism by which ROS participate in the ceramide death pathway.
We examined whether there was evidence for an early decrease in mitochondrial membrane potential (m) or for an early leakage of mitochondrial cytochrome c in ceramide-mediated caspase-independent PCD. Concurring with reports on TNF- or FasL-induced caspase-independent PCD in L929sAhFas and Jurkat wild-type cells (28
, 35)
, a decrease of m in TNF/zVAD- or TNF/CHX/zVAD-treated L929, NIH3T3, and Jurkat wild-type cells became apparent only in the late to end, but not the early, stages of cell death (Fig. 6
A), coinciding with loss of plasma membrane integrity (Figs. 1C
, 2B, E
).
|
Similarly, release of cytochrome c did not occur early in either L929, NIH3T3, or Jurkat wild-type cells (Fig. 6B
). Identical results have been obtained by us and others for caspase-independent PCD of L929/L929sAhFas cells induced by TNF alone, where cytochrome c release has been observed in the late phases of cell death but never in the early stages (ref 35
; not shown). This observation is also in line with the absence of detectable caspase-8 and -3 activity in TNF/zVAD-treated L929 cells (Supplementary Fig. 1). In summary, mitochondrial changes characteristic for caspase-dependent apoptosis such as early release of cytochrome c and early dissipation of m were not detected, suggesting that mitochondria contribute to ceramide-mediated caspase-independent PCD by mechanisms other than these.
Rapid externalization of phosphatidylserine is not essential for caspase-independent PCD induced by ceramide
To investigate more closely and precisely the role of ceramide in caspase-independent PCD, we monitored the kinetics of phosphatidylserine externalization vs. loss of plasma membrane integrity in L929, NIH3T3, and Jurkat wild-type cells. As observed above for loss of m, when undergoing ceramide-mediated caspase-independent PCD all three cell lines concurrently stained positive with both annexin V and 7-amino-actinomycin D (7-AAD) in the late phase of cell death (Fig. 7
). Therefore, the rapid externalization of phosphatidylserine that has been described in caspase-dependent apoptosis of these cells (28
, 35)
apparently is not crucial for caspase-independent PCD induced by ceramide.
|
Inhibition of caspases, but not cathepsins, enhances the ceramide death pathway
We investigated how zVAD-fmk enhances ceramide accumulation in TNF-induced caspase-independent PCD—specifically, whether this was a caspase-mediated effect or whether lysosomal cathepsins were implicated (zVAD-fmk inhibits both types of enzymes at the concentration of 20 µM employed in the previous experiments; ref 36
). L929 cells were treated with TNF in combination with zVAD-fmk, the cathepsin B/L inhibitor zFA-fmk, the cathepsin B inhibitor CA-074 Me, or the broad spectrum cysteine protease inhibitor E-64d (20 µM each). Of all inhibitors, only zVAD-fmk strongly amplified the cytotoxic response of TNF (Fig. 8
A). When added at 1 µM, a concentration that only affects caspases but not cathepsins (36)
, zVAD-fmk still substantially enhanced the cytotoxic effects of TNF (not shown), suggesting that the inhibition of caspases rather than cathepsins by zVAD-fmk is responsible for the observed enhancement of ceramide accumulation and for caspase-independent PCD.
|
To further specify the caspase responsible for this effect, we analyzed the impact of the inhibitors zIETD-fmk (specific for the initiator caspase-8) and zDEVD-fmk (inhibiting the effector caspase-3). Whereas zVAD-fmk still was the most potent enhancer of TNF cytotoxicity, zIETD-fmk and, to a lesser extent zDEVD-fmk, likewise accelerated TNF-dependent cell death (Fig. 8B
). Parallel experiments demonstrated accumulation of ceramide in L929, NIH3T3, and Jurkat cells after treatment with TNF in combination with low amounts of zVAD-fmk as well as with zDEVD-fmk or with zIETD-fmk (Fig. 8C
). Garcia-Calvo and co-workers have demonstrated that the DEVD tetrapeptide is a potent inhibitor not only for caspase-3, but also for caspase-8 (37)
. Therefore, the effect of zDEVD-fmk on caspase-independent PCD and ceramide accumulation may result from inhibition of the initiator caspase-8 rather than the effector caspase-3, implicating caspase-8 as a potential candidate responsible for suppression of ceramide-mediated caspase-independent PCD under physiological conditions.
To address the question of whether in consequence, activators of caspases also suppress ceramide formation, we examined ceramide levels in apoptotic L929, NIH3T3, and Jurkat wild-type cells undergoing caspase-dependent cell death. As shown in Fig. 8D
, induction of apoptosis in all three cell lines did not suppress but rather induced ceramide formation, in line with a large body of evidence implicating the accumulation of intracellular ceramide in apoptotic cell death (38
, 39)
. Thus, while activation of caspases such as caspase-8 may suppress ceramide-mediated caspase-independent PCD (and therefore ceramide production through the caspase-independent death pathway), it simultaneously induces ceramide production as part of the apoptotic death pathway. Although an apparent paradox at first glance, the induction of ceramide by either death pathway may assist a cell in committing suicide regardless of whether this is accomplished by caspase-dependent or -independent mechanisms.
Ceramide-mediated caspase-independent PCD reduces the clonogenic survival of tumor cells
We wanted to determine whether induction of ceramide-mediated caspase-independent PCD represents a viable strategy for the elimination of tumor cells that have become refractory to drugs eliciting apoptotic cell death. We therefore analyzed the clonogenic survival of untreated L929, NIH3T3, and Jurkat wild-type cells in comparison to cells that were treated with TNF/zVAD (Jurkat wild-type cells: TNF/CHX/zVAD). This treatment simulated apoptosis resistance of cells by inhibiting caspases through zVAD-fmk while simultaneously inducing ceramide-dependent caspase-independent PCD. As shown in Fig. 9
, clonogenicity was clearly decreased in all three cell lines, confirming that induction of caspase-independent PCD by manipulation of intracellular ceramide levels may represent a promising additional option for the development of novel tumor therapies.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 39)
, the issue whether ceramide represents an essential inducer or is generated merely in consequence of apoptosis has been, and still is, subject to debate (5
,6
). There is no doubt that ceramide is an essential inducer of apoptosis in certain tissues, e.g., in the lung (39)
or in oocytes (40)
, but apparently this is not the case in others, e.g., in the thymus (39)
. Although the role of ceramide in apoptosis has been studied extensively by many groups for many years, it remains a complex issue that probably will not be definitely clarified in the near future.
Here, we have focused on investigating the relevance of ceramide in caspase-independent PCD. In line with caspase-independent PCD being the more ancient form of cell death clearly predating the appearance of caspases in evolution (1
, 41)
, ceramide signaling likewise predates apoptosis (40)
. Therefore, it is tempting to speculate that in the ancient pathways to cell death, ceramide might have represented one of the central mediators. Subsequently, with the evolution of caspase-dependent apoptosis, the original system, including ceramide, may have partially or completely transferred its functions to caspase-dependent mechanisms, resulting in the differential relevance of ceramide for apoptotic cell death seen in current organisms. However, under conditions where the classical apoptosis machinery fails (e.g., in tumor cells that have become resistant to death signals inducing apoptosis), intracellular generation of ceramide may regain importance by acting as part of an ancient backup system that enables a cell to still commit suicide by caspase-independent PCD.
A pivotal role of ceramide in caspase-independent PCD is supported by our observation that ceramide consistently accumulates well before the onset of caspase-independent death in multiple cell systems. Interference with ceramide generation by multiple approaches such as overexpression of AC, pharmacological intervention, RNAi, or genetic ablation of A-SMase or RIP1 uniformly conferred protection regardless of the cell type studied. Moreover, A-SMase-deficient cells could be resensitized for caspase-independent PCD by reconstitution of A-SMase, providing compelling evidence for ceramide as a central mediator of caspase-independent PCD.
Unlike a previous study reporting that fumonisin B1 can protect L929 cells against TNF cytotoxicity (42)
, we did not observe protection against TNF/zVAD in this or against TNF in a previous (14)
study. This may be explained by differences of the L929 cells used. However, our results using fumonisin B1 are consistent within L929, NIH3T3, and Jurkat wild-type cells and in line with the data using AC-overexpressing or A-SMase-deficient cells, implicating A-SMase as the main enzyme responsible for the observed accumulation of ceramide.
With regard to the associated signal transduction pathways, RIP1 appears to be a central initiator since blockade of RIP1 function by genetic ablation, RNAi, or pharmacological intervention protected from TNF-induced caspase-independent PCD. Since inhibition of ceramide accumulation clearly diminished caspase-independent PCD but not as completely as inhibition of RIP1, ceramide obviously represents a central, but most likely not the only, factor transmitting the death signals generated by RIP1 in response to TNF.
Our own results and data from others using different Bcl-2 overexpressing cell lines implicate mitochondria in the ceramide death pathway (34
, 35)
. In line, mitochondria have been described as targets of ceramide in vitro (43)
and in vivo (1)
. The generation of ROS represents one of the mechanisms that have been suggested for the induction of caspase-independent PCD by mitochondria (3)
. Cauwels and co-workers recently demonstrated that zVAD-fmk does not alleviate, but rather exacerbates, TNF toxicity by a caspase-independent, ROS-mediated pathway in mice, providing strong evidence for a role of ROS in caspase-independent PCD in vivo (2)
. In accordance, our own data demonstrate protection of L929 cells by the radical scavenger BHA. However, the structurally very similar scavenger BHT did not protect L929 cells, and Jurkat wild-type cells were not protected by either scavenger. The different response of these two cell lines indicates that the involvement of mitochondria in ceramide-mediated caspase-independent PCD may not be universal. Therefore, the precise contribution of ROS to caspase-independent PCD in different cell types will have to be addressed in further studies. With regard to cytochrome c, our data clearly do not support a role for this molecule in caspase-independent PCD. The slow dissipation of m (and likewise, the late positive annexin V-staining) observed in TNF(/CHX)/zVAD-treated L929, NIH3T3, and Jurkat wild-type cells coinciding with loss of plasma membrane integrity has also been described in caspase-independent PCD induced by other stimuli (TNF or FasL; refs 28
, 35
). This suggests that loss of m (as well as positive staining for phosphatidylserine) is a consequence rather than the cause of ceramide-mediated caspase-independent PCD. Therefore, aside from ROS as potential mediators, mitochondria may participate in the ceramide death pathway by mechanisms and factors other than early release of cytochrome c or early dissipation of m. The function of apoptosis-inducing factor as such a mitochondrial candidate molecule is subject of current investigation. As for the mechanism by which Bcl-2 protects from ceramide-mediated caspase-independent PCD, Denecker and co-workers have implicated complexation of the proapoptotic Bcl-2 family member BNIP3 by Bcl-2, and alternatively, the ability of Bcl-2 to prolong the integrity of mitochondrial oxidative phosphorylation (35)
.
Our data indicate that inhibition of caspases (rather than cathepsins) appears to be responsible for the amplification of TNF-mediated caspase-independent PCD by zVAD-fmk in L929 cells, in line with previous data. In addition to zVAD-fmk, both zIETD-fmk and zDEVD-fmk induced accumulation of ceramide and accelerated TNF-dependent cell death in L929 cells. Since zDEVD inhibits caspase-8 in addition to caspase-3 (37)
, our data argue for the initiator caspase-8 as a likely candidate molecule whose inhibition (e.g., by infecting viruses attempting to block apoptotic suicide) might amplify the execution of the caspase-independent death program, thereby accelerating rather than blocking cell suicide. A similar amplification of caspase-independent PCD has been described after transfection of a dominant negative caspase-8 construct (27)
. Alternatively, inhibition of caspase-8 may inhibit a signaling pathway that protects from caspase-independent PCD. For instance, caspase-8 is able to proteolyze RIP1. Therefore, inhibition of caspase-8 might prevent cleavage of RIP1 and thus propagate caspase-independent PCD (44
, 45)
.
Finally, our results demonstrate that induction of ceramide-mediated caspase-independent PCD leads to a pronounced reduction of clonogenic survival, providing additional evidence for the concept that modulation of ceramide metabolism may represent an attractive strategy to augment conventional tumor therapies. This may be especially valuable for the elimination of tumor cells that have become resistant to drugs inducing apoptotic death signals (46
, 47)
.
|
ACKNOWLEDGMENTS |
|---|
Received for publication January 26, 2005. Accepted for publication August 10, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
