| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,1
* INSERM U466, Laboratoire de Biochimie, Maladies Métaboliques, Institut Louis Bugnard, C.H.U. Rangueil, 31403 Toulouse, France;
INSERM E9910, Institut Claudius Régaud, 31052 Toulouse, France; and
Section of Hematology/Oncology, University of Illinois at Chicago, Chicago, Illinois, USA
2Correspondence: INSERM U. 466, Laboratoire de Biochimie, ‘Maladies Métaboliques’, Institut Louis Bugnard, Bât. L3, C.H.U. Rangueil, 1 Avenue Jean Poulhès, F-31403 Toulouse Cedex 4, France. E-mail: levade{at}rangueil.inserm.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
nor anti-CD95 induced
the degradation to ceramide of a natural SM that had been first
introduced selectively into acidic compartments. 3)
Stimulation of SV40-transformed fibroblasts by TNF- or CD40 ligand
resulted in apoptosis equally well in cells derived from control
individuals and from patients affected with Farber disease, having a
genetic defect of acid ceramidase activity leading to lysosomal
accumulation of ceramide. Also, induction of apoptosis using anti-CD95
(Fas) or anti-CD40 antibodies, TNF-, daunorubicin, and ionizing
radiation was similar in control and Farber disease lymphoid cells. In
all cases, apoptosis was preceded by a comparable increase of
intracellular ceramide levels. 4) Retroviral-mediated
gene transfer and overexpression of acid ceramidase in Farber
fibroblasts, which led to complete metabolic correction of the ceramide
catabolic defect, did not affect the cell response to TNF- and CD40
ligand.— Ségui, B., Bezombes, C., Uro-Coste, E., Medin, J. A., Andrieu-Abadie, N., Augé, N., Brouchet, A., Laurent, G.,
Salvayre, R., Jaffrézou, J.-P., Levade, T. Stress-induced
apoptosis is not mediated by endolysosomal ceramide.
Key Words: sphingomyelin • sphingomyelinase • Farber disease • ceramidase • lysosome • signal transduction
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
. Ceramide also derives
from the stepwise degradation of all sphingolipids through the action
of specific hydrolases, the best documented of which are located in the
lysosomes. In particular, ceramide can originate from the breakdown of
the major sphingolipid, sphingomyelin (SM) by an acid sphingomyelinase.
Then, in the lysosomal compartment, acid ceramidase degrades ceramide
to liberate sphingosine and a free fatty acid (2
, 3)
.
A wide array of biological responses to extracellular stimuli have been
described to be mediated by the intracellular generation of ceramide
(for reviews, see refs 4
5
6
7
8
9
10
). Among these effects is the
induction, in a variety of cell types, of apoptotic cell death
triggered by various stress agents, including ligation of the 55 kDa
tumor necrosis factor (TNF) and CD95 (Fas/APO1) receptors, ionizing and
UV radiation, anticancer drugs, heat shock, and exposure to hydrogen
peroxide (5
, 11
12
13
14
15)
. Ceramide has been proposed to
mediate apoptosis induced by these agents because 1) the
stimuli prompt the production of ceramide from SM hydrolysis,
2) the stimuli activate a sphingomyelinase, and
3) exogenously added ceramide (usually, synthetic
short-chain, cell-permeant analogs of ceramide) mimics the cytotoxic
effect.
Whereas different sphingomyelinases having distinct subcellular
locations may implement different biological effects of ceramide
(7
, 9)
, an apoptotic function has been attributed by some
authors to the ceramide produced in acidic organelles through the
action of an acid sphingomyelinase. An increase in sphingomyelinase
activity measured at acidic pH, which preceded the onset of apoptosis,
has been reported in extracts of cells treated with TNF- or
anti-CD95 antibodies (16
, 17)
. In addition, lymphoid cells
derived from patients affected with Niemann-Pick disease, an inborn
disorder characterized by the deficient activity of lysosomal
sphingomyelinase (18
, 19)
, failed to generate ceramide and
to undergo apoptosis on irradiation (20)
or CD95 ligation
(21)
. Similar observations were described on endothelial
cells from acid sphingomyelinase knock-out mice that were irradiated
(20)
or exposed to lipopolysaccharide (22)
.
Taken together, these findings suggested that diverse stress agents can
mediate their cytotoxic action through the generation of ceramide in
endosomes/lysosomes. However, conflicting data have been published,
including studies using the same cell lines and the same agonist.
Indeed, not only has activation of a nonacidic sphingomyelinase
frequently been reported subsequent to cell injuries (see Discussion),
but also significant ceramide production and a normal apoptotic
response have been observed in cells genetically lacking functional
acid sphingomyelinase (23
, 24)
.
The present study was designed to elucidate the role of the natural
ceramide present or formed in acidic compartments in the transduction
of apoptotic signals. We speculated that if the ceramide liberated by
acid lysosomal sphingomyelinase is involved in apoptosis signaling, its
accumulation in acidic compartments should result in cytotoxicity
and/or sensitivity to apoptotic stress agents. To test this hypothesis,
we first examined the response of cells loaded with ceramide under
conditions that allow lysosomal targeting of the exogenous lipid.
Secondly, we examined the response to apoptotic signals of cells
derived from patients affected with Farber disease. Farber disease
(lipogranulomatosis; McKusick 22800) is a rare, autosomal recessive,
lysosomal storage disorder, which is characterized by the intracellular
accumulation of ceramide as a result of a deficiency in the activity of
acid lysosomal ceramidase (or N-acylsphingosine deacylase, E.C.
3.5.1.23) (25
, 26)
. We postulated that in these cultured
Farber cells the levels of lysosomal ceramide would be augmented, and,
consequently, if this ceramide plays a role in apoptosis, these cells
might have an altered sensitivity to stress-induced apoptosis. This
study demonstrates that endosomal/lysosomal ceramide does not mediate
the apoptotic cell death induced by diverse stress agents and that acid
ceramidase is not essential in this signaling pathway.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Radiolabeled ceramide was
obtained by hydrolysis of [ceramide-3H]SM using
B. cereus sphingomyelinase; N-palmitoyl-D-sphingosine,
[palmitoyl-1-14C] (55 mCi/mmol) was purchased
from A.R.C. (St Louis, Missouri). [
-32P]ATP
(7000 Ci/mmol) was purchased from ICN (Orsay, France), and
[methyl-3H]thymidine from Amersham (Les Ulis,
France). Sphingomyelinase (B. cereus), C2-ceramide
(N-acetyl-sphingosine), and dioleoylphosphatidylglycerol were supplied
from Sigma (Lisle d’Abeau, France). Recombinant E. coli
diacylglycerol kinase, octyl-ß-glucoside (Ultrol grade) and
concanamycin A were from Calbiochem (Meudon, France); silica gel 60 TLC
plates (Art. 5721) were from Merck (Darmstadt, Germany). Recombinant
human TNF- was purchased from PeproTech (Tebu, France). Antibodies
to human CD95 (clone CH-11) and CD40 (clone M3) were from Upstate
Biotechnology-Euromedex (Strasbourg, France) and Genzyme (Cergy,
France), respectively. Soluble recombinant human CD40 ligand was kindly
provided by Immunex (Seattle, Wash.). Daunorubicin was obtained from
the National Cancer Institute Drug Repository. All solvents and other
reagents obtained from Merck or SDS (Peypin, France) were of analytical
grade. DMEM Glutamax, RPMI 1640 Glutamax, penicillin, streptomycin, and
trypsin-EDTA were from Gibco BRL (Cergy-Pontoise, France); fetal calf
serum was from Boehringer Ingelheim (Gagny, France).
Cell lines and cell culture
Human SV40-transformed skin fibroblasts were derived from normal
individuals, from a patient affected with Niemann-Pick disease type A
(line Ber), or from a patient affected with Farber disease, who died at
3 days of age (line Moh) (28)
. The cells were routinely
grown in a humidified 5% CO2 atmosphere at
37°C in DMEM medium containing Glutamax (2 mmol/l), penicillin (100
U/ml), streptomycin (100 µg/ml), and heat-inactivated fetal calf
serum (10%), as previously reported (28
, 29)
. Fibroblasts
were transduced using a recombinant retroviral vector carrying the
full-length human acid ceramidase cDNA as described (30)
.
Cultured skin fibroblasts were used after reaching confluency. Human
Epstein-Barr virus-transformed lymphoid cell lines were derived from
control subjects, from a patient affected with Niemann-Pick disease
type A (line Tre), or from patients with Farber disease (lines GM5748,
obtained from the Human Genetic Mutant Cell Repository, Camden, N.J.;
or line Moz). These cell lines have previously been characterized with
respect to SM and ceramide metabolism (27
, 31)
. They were
routinely grown in a humidified 5% CO2
atmosphere at 37°C in RPMI 1640 medium containing Glutamax (2
mmol/l), penicillin (100 U/ml), streptomycin (100 µg/ml), and
heat-inactivated fetal calf serum (10%). Jurkat T cells were grown
under the same conditions as the lymphoid cells. Irradiations were
performed using a 60Co source (1.25 MeV, Alcyon,
General Electric) at a dose rate of ~1 Gy/min.
Incubation of intact cells with sphingomyelin
Cells were incubated at 37°C for the indicated periods of
time with medium containing 10% fetal calf serum and
[ceramide-3H]SM (~106
dpm/ml, which was added as an ethanolic solution) (27)
.
Next, cells were briefly washed with fresh medium and further incubated
in the presence or absence of TNF- or anti-CD95. At the end of
incubation, cells were washed three times with phosphate-buffered
saline (PBS) containing bovine serum albumin (2 mg/ml) and twice with
PBS alone, and harvested using a rubber policeman. The cell pellets
were stored at -20°C.
Lipid extraction and analyses
Cell pellets were suspended in 0.6 ml distilled water and
sonicated for 2 x 15 s (Soniprep MSE sonicator). After an
aliquot was taken for protein determination (32)
, the
lipids were extracted (33)
.
For analysis of the intracellular distribution of [ceramide-3H]SM metabolites, lipids were resolved by analytical TLC developed in chloroform/methanol/water (100:42:6, by vol.) up to two-thirds of the plate and then in chloroform/methanol/acetic acid (94:1:5, by vol.). The distribution of the radioactivity on the plate was analyzed using a Berthold LB2832 radiochromatoscan. Unlabeled or radioactive lipid standards were used to identify the various metabolic products, which were scraped and quantified by liquid scintillation.
Total intracellular SM levels were determined on cells metabolically
labeled for 48 h with [3H]choline (1
µCi/ml) using the previously described procedure (34)
.
Ceramide levels were quantitated in the lipid extracts essentially as
described (35)
, using E. coli diacylglycerol
kinase and [-32P]ATP. Radioactive
ceramide-1-phosphate was isolated by TLC using
chloroform/acetone/methanol/acetic acid/water (50:20:15:10:5, by vol.)
as the developing solvent. Alternatively, ceramide was quantitated
after metabolic labeling of the cells for 48 h with 1 µCi/ml
[9, 10-3H]palmitic acid (53.0 Ci/mmol;
Amersham) (14
, 36)
.
Acid ceramidase assay
Fibroblasts were harvested and washed in cold 250 mM
sucrose/50 mM NaCl/10 mM Tris/HCl, pH 7.5. Cell pellets were suspended
and disrupted in 250 mM sucrose/1 mM EDTA by brief sonication.
Ceramidase activity was assayed as described previously
(37)
using
[palmitoyl-1-14C]sphingosine
(105 dpm/assay) as the substrate along with
detergents in the presence of 125 mM citrate-phosphate buffer, pH 4.
After a 2 h incubation at 37°C, the liberated radioactive fatty
acid was isolated by phase partition, followed by TLC separation in
chloroform/methanol/acetic acid (94:1:5, by vol.). After exposure to
iodine vapors and/or radiochromato-scanning to locate the fatty acid on
the plate, the radioactive spots were scraped and counted by liquid
scintillation.
Sphingomyelinase assay
Neutral and acid sphingomyelinase activities were
determined on freshly isolated cell pellets essentially as described
(14
, 38)
, using
[choline-methyl-14C]SM (NEN; 100,000 dpm/assay)
as substrate.
Morphological studies, cell viability, and DNA fragmentation
assays
Lymphoid cells were centrifuged at low speed on glass slides and
examined under the light microscope after May-Grünwald-Giemsa
(MGG) staining. Cell membrane permeability was assessed by the trypan
blue dye exclusion test. In some experiments,
[3H]thymidine incorporation was monitored as
previously reported (34)
. Fibroblast viability was
estimated directly on the culture flasks by coincubation with propidium
iodide (6 µM) and Syto-13 (1 µM; Molecular Probes, Leiden, The
Netherlands). The percentage of apoptotic cells (having a condensed and
fragmented nucleus) was evaluated by counting cells under a Leica
fluorescence-equipped inverted microscope.
Quantitative DNA fragmentation was determined by the
spectrofluorometric DAPI procedure as described previously (39
, 40)
.
PARP cleavage assay
Analysis of poly(ADP-ribose)polymerase (PARP) proteolysis
was assessed by resuspending cells in sample buffer (62.5 mM Tris, pH
6.8, 4 M urea, 10% glycerol, 2% SDS, 5% ß-mercaptoethanol, and
0.04% bromophenol blue). Samples were boiled for 5 min, loaded onto a
10% SDS-polyacrylamide gel, electrophoresed, and transferred to a
nitrocellulose membrane. PARP and its cleaved fragment were detected by
using a rabbit polyclonal antiserum (Boerhinger-Mannheim, Meylan,
France) and a donkey anti-rabbit secondary antibody (Immunotech,
Marseille, France). The signal was visualized by enhanced
chemiluminescence (Amersham, Buckinghamshire, U.K.).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
by incubating the living cells with human
LDL that were pretreated with a bacterial sphingomyelinase to degrade
virtually all the lipoprotein SM to ceramide (Fig. 1C
). As shown in Fig. 1A
, although an agonistic anti-CD95
antibody induced cell death, loading Jurkat T cells with LDL-ceramide
(see Fig. 1D
) did not significantly alter their viability.
In contrast to the natural compound delivered to acidic compartments,
synthetic short-chain ceramides (i.e., N-acetyl- or
N-hexanoyl-sphingosines), which readily enter cells, were cytotoxic
(Fig. 1A
).
|
To test whether accumulation of ceramide in endosomes could affect cell
viability, cells were coincubated with LDL-ceramide and concanamycin A,
an antibiotic that blocks the transport from endosomes to lysosomes as
a result of inhibition of vacuolar proton-ATPase and neutralization of
the pH of acidic organelles. Fig. 1B
shows that inhibition
of ceramide traffic to lysosomes did not alter the viability of Jurkat
cells. Identical results were obtained on EBV-transformed lymphoid
cells and on the leukemic U937 cell line (data not shown). These
findings strongly suggest that the natural ceramide introduced into
endosomes and lysosomes does not signal cell death in cellular models
where ceramide has been described to serve as an apoptotic mediator.
TNF- or CD95 stimulation does not result in degradation of
lysosomal sphingomyelin
Because ceramide has been proposed to be generated in acidic
organelles through TNF- or CD95 receptor-mediated activation of an
acid sphingomyelinase (12
, 42)
, and because we and others
have found that acid sphingomyelinase-deficient (Niemann-Pick
disease) cells do not exhibit defects in the apoptotic response nor in
the activation of the ceramide pathway (see Discussion), we examined
whether TNF- or CD95 could stimulate the degradation of SM present
in the endosomal/lysosomal compartment. Lymphoid cells and
SV40-transformed fibroblasts derived from patients with Niemann-Pick
disease type A, which accumulate SM intracellularly, were loaded with
radioactive SM under conditions that target this SM to acidic
organelles (27
, 31)
, and then treated with TNF- or with
an agonistic anti-CD95 antibody (CH-11). As seen in Fig. 2
, the radiolabeled SM accumulated in endosomes/lysosomes because of the
defect in acid sphingomyelinase. Notably, neither stimulus induced a
detectable, time-dependent hydrolysis of the SM present in acidic
compartments (compare with Figs. 3
, 4
, and 7
), indicating that lysosomal SM is not accessible to the
sphingomyelinase activated (even in Niemann-Pick cells) by apoptotic
agents.
|
|
|
|
Accumulation of ceramide in Farber disease cell lines does not
affect cell growth
To further evaluate the role of lysosomal ceramide and acid
ceramidase in apoptosis, we used both SV40-transformed skin fibroblasts
and EBV-transformed lymphoid cell lines derived from patients affected
with Farber disease. As illustrated in Fig. 3
, these mutant cells
exhibited a 3- to 10-fold increase in intracellular ceramide
concentration (A, B) and failed to degrade the
ceramide produced intra-lysosomally from the hydrolysis of a
radiolabeled SM (C–F) that was targeted to lysosomes
through lipoprotein receptor-mediated endocytosis (27
, 31
, 41)
. These results are in accordance with the previously
reported defect in acid ceramidase activity measured in cell extracts
(28
, 43)
.
Most notably, even though these cells accumulate large amounts of
ceramide, their growth rate in either 1 or 10% fetal calf
serum-containing medium was similar to that of their normal
counterparts (data not shown), as was the basal number of apoptotic
cells (see Fig. 5
).
|
The sphingomyelin-ceramide pathway can be activated in Farber
lymphoid cells
Because ceramide has been proposed to be produced on treatment
with, and to transduce the apoptotic effect of, diverse stress agents,
we sought to determine whether the apoptosis of Farber cells induced by
these stimuli was accompanied by the activation of the ceramide
pathway. This was studied either by measuring the sphingomyelinase
activity or by following the SM breakdown and ceramide generation.
Fig. 4
demonstrates that all stress stimuli used here activated the ceramide
pathway equally well in Farber and control lymphoid cells. Indeed, CD95
stimulation (Fig. 4A
), as well as daunorubicin, TNF-, and
ionizing radiation treatment (Fig. 4C
), resulted in an
increase of cellular ceramide concentration and a concomitant increase
in neutral sphingomyelinase activity (Fig. 4D
). SM
hydrolysis was also stimulated similarly in normal and Farber cells by
anti-CD40 (Fig. 4B
). Acid sphingomyelinase activity was not
significantly affected after treatment with daunorubicin or anti-CD95
(data not shown). Of note, despite their deficient acid ceramidase
activity the Farber cells did not exhibit a higher amount of
agonist-induced ceramide than their normal counterparts, suggesting
little contribution of the lysosomal pathway in the production of this
ceramide.
Consistent with these findings, when radiolabeled SM loading
experiments (as in Fig. 2
) were performed on Farber fibroblasts or
lymphoid cells, neither TNF- nor anti-CD95 promoted an enhanced
catabolism of lysosomal ceramide (data not shown).
Control and Farber lymphoid cells are equally sensitive to
CD95-mediated apoptosis
Because CD95 ligation has been reported to activate the
SM-ceramide pathway and to promote apoptosis in EBV-transformed
lymphoid cell lines (21
, 24)
, we investigated the effect
of an agonistic anti-CD95 (CH-11) antibody on lymphoid cells derived
from Farber disease patients. As shown in Fig. 5
, triggering of the CD95 receptor led to apoptotic cell death of both
control and Farber cells. Indeed, the characteristic features of
apoptosis were observed morphologically (i.e., cell size reduction,
chromatin condensation, and membrane blebbing; Fig. 5A
), by
monitoring the DNA fragmentation (Fig. 5C
), and by examining
the cleavage of PARP (Fig. 5D
). These alterations were
followed by modifications of plasma membrane permeability as measured
by trypan blue dye uptake (data not shown), indicative of
postapoptotic necrotic features. Quantification of these phenomena
indicated that Farber lymphoid cells underwent apoptosis in a similar
fashion and with a similar time course (data not shown) as their normal
counterparts (Fig. 5B
).
Control and Farber lymphoid cells are equally sensitive to TNF-,
CD40, daunorubicin, and ionizing radiation-induced apoptosis
The sensitivity of Farber lymphoid cells to apoptotic agents other
than anti-CD95 was evaluated. Control and Farber lymphoid cells were
exposed to TNF-, daunorubicin, ionizing radiation, and anti-CD40,
all stimuli that have been shown to induce apoptosis. Fig. 6
demonstrates that the cytotoxic effect of these agents was similar in
control and Farber cells. Indeed, not only was the growth of control
and mutant cells similarly inhibited (Fig. 6A, B
), but also
PARP was cleaved (Fig. 6C
) on treatment by the stress
agents. Furthermore, morphological studies using MGG, DAPI, or TUNEL
staining indicated characteristic features of apoptosis (data not
shown). Finally, control and Farber cells were equally sensitive to
cell-permeant ceramides (data not shown), indicating that these
synthetic short-chain ceramides are cytotoxic irrespective of the
degree of accumulation of natural ceramide in acidic compartments.
|
Untransduced and genetically corrected Farber fibroblasts are
equally sensitive to TNF- and CD40-induced apoptosis
To investigate whether the sensitivity of Farber cells to
apoptotic agents was restricted to lymphoid cells, we examined the
effect of stress inducers on another cell type (i.e., skin
fibroblasts). SV40-transformed fibroblasts have been reported to
undergo apoptotic cell death on stimulation of members of the TNF
receptor superfamily (e.g., TNF, CD95, and CD40 receptors)
(44)
. In addition, this response in SV40-transformed
fibroblasts is accompanied by the activation of the ceramide pathway
(Ségui, B., and Levade, T., unpublished results). As illustrated
in Fig. 7
(C–E), exposure of SV40-transformed fibroblasts from a
patient with Farber disease to TNF- or CD40 ligand led to a
reduction in thymidine incorporation, which was linked to apoptosis as
evidenced by morphological alterations. No difference was found in
cytotoxicity between control and Farber fibroblast cell lines. In
addition, TNF and CD40 receptor stimulation resulted in SM hydrolysis
in both mutant and control fibroblasts (Fig. 7B
).
To further examine the importance of acid ceramidase in this apoptotic
signaling pathway, we tested the ability of SV40-transformed
fibroblasts that overexpress acid ceramidase to respond to TNF- or
CD40 ligation. Toward this aim, we used Farber fibroblasts that had
been infected with a recombinant retroviral vector encoding acid
ceramidase (30)
. These transduced cells not only exhibit a
fully corrected catabolism of ceramide but also overexpress acid
ceramidase (~5x above the activity found in normal control cells;
Fig. 7A
). Despite their high acid ceramidase activity, these
transduced cells responded to TNF- and CD40 ligation exactly as
mock-transduced cells, by activating the ceramide pathway (Fig. 7B
) and by undergoing apoptosis (Fig. 7C-E
).
Similar observations were made on Farber lymphoid cells that were
challenged with anti-CD95, TNF-, daunorubicin, or ionizing radiation
after cross-correction of their metabolic deficiency (data not shown)
(i.e., after having been incubated with the culture medium of
genetically corrected Farber fibroblasts to normalize the acid
ceramidase activity) (30)
.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 5
, 7
8
9
10
, 45)
. In fact, the intracellular concentration of ceramide
has been shown to increase after cell stimulation with cytokines
(TNF-), antibodies (anti-CD95, anti-IgM), UV and ionizing radiation,
anticancer drugs (e.g., daunorubicin, vincristine), serum withdrawal,
heat, and hypoxia. The generation of ceramide was not merely the
consequence of general cell dysfunction and death, because inhibition
of apoptosis by certain caspase synthetic or viral protein inhibitors
did not abrogate the ceramide increase (5
, 10
, 46)
.
Rather, ceramide is believed to belong to the pathways of the
signaling/commitment phase of apoptosis (5
, 9
, 10)
.
Indeed, exogenously supplied ceramide or the ceramide formed at the
cell surface after treatment with bacterial sphingomyelinase can
activate multiple pathways that play a critical role in the apoptotic
process (e.g., the JNK/SAPK pathway, mitochondrial alterations, caspase
cascade, and NF-
B nuclear translocation).
While the targets of ceramide begin to be deciphered, controversy
exists as to the mechanisms of its generation, in particular the nature
and subcellular localization of the enzyme sphingomyelinase. Various
sphingomyelinases are present in mammalian cells that can be
distinguished by their subcellular localization and pH optimum
(1
, 2
, 18)
. Of these, the best characterized enzyme is the
acid sphingomyelinase. This enzyme is present in acidic organelles, and
its deficiency gives rise to the lipid storage disorder Niemann-Pick
disease (18
, 19)
. The implication of an acid
sphingomyelinase in signal transduction was first reported by
Krönke et al. as an intermediate for TNF-induced nuclear
translocation of NF-B in leukemic cells (42)
.
Subsequent studies by these authors indicated that only the ceramide
liberated by the acid sphingomyelinase in endosomes/lysosomes could
mediate the TNF-induced activation of NF-B and the induction of cell
death (38)
. In this model, TNF-dependent activation of an
acid sphingomyelinase is believed to be signaled by the death domain
region of the TNF p55 receptor and involves the recruitment of the
TRADD and FADD proteins (9
, 47
, 48)
. Other observations,
still based on in vitro measurements of enzyme activity,
have suggested a signaling function for acid sphingomyelinase in
apoptosis triggered by anti-CD95 (12
, 16)
and anti-IgM
(49)
. Moreover, cultured cells derived from patients with
Niemann-Pick disease or certain cell types from acid
sphingomyelinase-deficient mice were reported to be resistant to
ionizing radiation (20)
, doxorubicin (23)
,
lipopolysaccharide (22)
, or CD95 stimulation
(21)
. Overall, these findings gave support to the idea
that the ceramide formed in acidic compartments serves as a second
messenger for diverse stress stimuli.
However, this concept is challenged by several observations that
strongly argue against this view (50)
. First, activation
of a neutral, but not acid, sphingomyelinase has been associated with
induction of apoptosis by a number of stress agents, including TNF-
(17
, 38
, 51
, 52)
, anti-CD95 (53)
, anti-IgM
(54)
, anti-CD40 (Ségui, B., Andrieu-Abadie, N., and
Levade, T., unpublished results), daunorubicin (14)
,
Ara-C, cis-platinum, taxol, mitoxantrone (55
, 56)
, serum withdrawal (57)
, ionizing radiation
(58)
, and hypoxia (59)
. Second, cells having
a genetic deficiency of acid sphingomyelinase (i.e., Niemann-Pick
cells) were found to generate ceramide and to undergo apoptosis after
CD95 stimulation (24)
or exposure to anti-class I
antibodies (60)
. We recently extended these findings to
mutant cells exposed to a variety of stress stimuli (Bezombes, C.,
Ségui, B., Bruno, A.P., Uro-Coste, E., Andrieu-Abadie, N.,
Laurent, G., Jaffrézou, J.P., Levade, T., unpublished results).
Moreover, hydrolysis of SM was observed in TNF-stimulated Niemann-Pick
cells (34)
, and diverse biological responses supposed to
be mediated by ceramide were noted in these deficient cells, ranging
from NF-B (61
62
63)
, ERK or JNK activation (23
, 64)
to cytokine production (65
, 66)
. Even when
apoptosis was monitored in acid sphingomyelinase-deficient mice, only
some cell types resisted the stress agent (20)
. Consistent
with these findings is the marked absence of similitude between the
Niemann-Pick condition and the phenotypes of CD95 or CD95
ligand-defective animals (50
, 67)
. Third, a series of
experiments have indicated that the signal-induced hydrolysis of SM
occurs in (or very close to) the inner leaflet of the plasma membrane
but not in endosomes (36
, 52
, 68
, 69)
. Finally, natural
ceramide formed or introduced into the acidic compartments appears to
be unable to escape these organelles (41)
, which makes it
difficult to understand how a ceramide present in the lumen of
endosomes/lysosomes could activate protein targets present in other
subcellular compartments.
In addition to the above observations, the present study strongly argues against an essential role for endosomal/lysosomal ceramide in stress-induced apoptosis. Two experimental approaches have been used to investigate directly the possible signaling function of endosomal/lysosomal ceramide. One approach was to selectively introduce natural ceramide in acidic organelles of cells (like Jurkat) that are very sensitive to ceramide-induced cell death. We demonstrate that ceramide loading of acidic organelles failed to result in apoptosis, even when the transport between endosomes and lysosomes was blocked. In addition, we show that apoptotic agents known to trigger the SM-ceramide pathway did not stimulate the breakdown of a SM initially targeted to the endosomal/lysosomal compartment of Niemann-Pick cells where it accumulated owing to the deficient activity of acid sphingomyelinase. In the same cells, those agents induced an apoptotic cell death that was preceded by neutral sphingomyelinase activation and hydrolysis of SM (as followed by labeling of all cellular SM pools) (Bezombes, C., Ségui, B., Bruno, A.P., Uro-Coste, E., Andrieu-Abadie, N., Laurent, G., Jaffrézou, J.P., Levade, T., unpublished results).
To further analyze the role of endosomal/lysosomal ceramide, we used a
genetic approach. Cultured cells having a genetic defect in acid
ceramidase activity (Farber disease) were tested for their ability to
respond to various stress stimuli. These cells, which were of the
fibroblast or lymphoid type, accumulated ceramide as a result of the
block in its lysosomal degradation. However, despite the ceramide
storage in acidic compartments, they did not manifest any cell
viability abnormality, either in the absence or presence of a stress
agent. When stimulated with TNF-, anti-CD95 or CD40 antibodies,
anthracycline, or ionizing radiation, as well as cell-permeant
ceramides, Farber cells underwent apoptosis just as control cells, both
quantitatively and with the same dose-dependency and kinetics. In
addition, Farber fibroblasts were equally sensitive to oxidized LDL
(data not shown), which has been previously shown to activate the
ceramide pathway (70)
and to induce apoptosis
(71)
. These results agree with the notion that Farber
disease is characterized clinically by symptoms that do not evoke any
anomaly of apoptotic processes (26)
and that cultured
cells derived from patients affected with this inborn disorder do not
exhibit particular growth difficulties.
While the above observations provide evidence for the absence of
signaling functions of endosomal/lysosomal ceramide, other findings
support the idea that functional acid ceramidase is not required for
and does not influence transduction of apoptotic signals. Indeed, cells
having a genetic deficiency of acid ceramidase are shown here to
activate the SM pathway and to generate ceramide in response to various
stress agents in a similar way as control cells. Furthermore, stable
overexpression of acid ceramidase (up to 500% of the activity in
normal cells) was not accompanied by any alteration in cell viability
or in stress-induced apoptotic response and hydrolysis of SM. These
results are in agreement with those of Boesen de Cock et al.
(24)
, showing that acid sphingomyelinase is fully
dispensable for CD95-induced cell death. They also demonstrate that
metabolic transformation of lysosomal ceramide (e.g., production of
sphingosine) is not necessary for apoptosis signaling, suggesting that
if sphingosine formation is an important step for mediating apoptosis
(45)
, this sphingosine is not generated in acidic
organelles.
Taken together, these findings strongly argue against
endosomal/lysosomal ceramide playing a role in stress-induced apoptosis
and further strengthen the notion that the ceramide produced in cells
stimulated by apoptotic agents does not result from an
endosomal/lysosomal sphingomyelinase activity. Whereas some studies
have described the implication of an acid sphingomyelinase based on
in vitro measurements of enzyme activity or on observations
on genetically acid sphingomyelinase-deficient cells, this
investigation presents topological arguments that exclude the
endosomal/lysosomal compartment as a site for generation of apoptotic
ceramide. Whether the signaling ceramide can be formed in another
acidic compartment by a yet undefined sphingomyelinase remains to be
clarified, as does the role of the secretory form of acid
sphingomyelinase, which is also encoded by the gene mutated in
Niemann-Pick disease (72)
.
|
ACKNOWLEDGMENTS |
|---|
The authors thank Dr. K. Harzer for providing the primary culture of Farber fibroblasts (line Moh), and S. Carpentier, J. P. Basile, and J. C. Thiers for acid ceramidase assays and iconography. Financial support by INSERM, Faculté de Médecine-Rangueil Université Paul Sabatier Toulouse, Association pour la Recherche sur le Cancer, Ligue Nationale contre le Cancer, Vaincre les Maladies Lysosomales, and Centre National d’Etudes Spatiales is gratefully acknowledged.
|
FOOTNOTES |
|---|
2 Abbreviations: SM, sphingomyelin; TNF, tumor
necrosis factor; EBV, Epstein-Barr virus; LDL, low-density lipoprotein;
MGG, May-Grünwald-Giemsa; PARP, poly(ADP-ribose)polymerase; TLC,
thin-layer chromatography.
Received for publication April 22, 1999. Revised for publication September 10, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
B activation. J. Biol. Chem. 270,24518-24524B by phosphatidylcholine-specific phospholipase C-induced ‘acidic’ sphingomyelin breakdown. Cell 71,765-776[Medline]
: CrmA and Bcl-2 target distinct components in the apoptotic pathway. J. Exp. Med. 185,481-490 in the synthesis of cholesteryl esters in human fibroblasts. J. Biol. Chem. 269,879-882-induced ceramide generation and apoptosis in resistant human leukemia KG1a cells by the P-glycoprotein blocker PSC833. FASEB J 12,101-109B activation. Int. Immunol. 6,1269-1272 activates NF-B in acid sphingomyelinase-deficient mouse embryonic fibroblasts. J. Biol. Chem. 272,10904-10909
