From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
Opposite Effects of Tumor Necrosis Factor Q! on the SphingomyelinCeramide Pathway in Two Myeloid Leukemia Cell Lines: Role of Transverse
Sphingomyelin Distribution in the Plasma Membrane
By Ali Bettaieb, Michel Record, Marie-George CBme, Anne-Claire Bras, Hugues Chap, Guy Laurent,
and Jean-Pierre Jaffrezou
Tumor necrosis factor a (TNFcu) mediates proliferation, functional activation, and apoptoticcell death dependingon the
target cell type. Although sphingomyelin (SPM) hydrolysis
and ceramide generation may functionas an important mediator in TNFa signaling, the molecular mechanisms of the
signaling pathway(s1 are still not well understood. The aim
of the present study is t o compare the effect of TNFa on
SPM metabolism and cell growth in two myeloid leukemic
cell lines (U937 and KGla) thatdiffer in their sensitivity t o
TNFa. Our results show that TNFa induced apoptosis in
U937 but not in KGla cells. TNFa triggered in KGla cells
neither SPM hydrolysis nor ceramide generation, but induced SPM synthesis and ceramide breakdown as well as
dose-dependent cell proliferation. In contrast, TNFa induced
in U937 SPMhydrolysis and ceramide generation as well as
dose-dependent cell death. Synthetic cell permeant ceramide, as well as natural ceramide, generated by treatment
with bacterial sphingomyelinase (SPMase), all induced
apoptosis in both U937 and KGla cells. These findings indicate that theSPM-ceramide pathway isaltered in KGla cells
upstream of the ceramide generation. Analysis of thetransverse distribution of SPM in the plasma membrane showed
that the SPM pool involved in cell signaling (inner leaflet)
was markedly reduced in KGla cells; it is 7-fold lower than
that found in theinner leaflet of U937 cells. Therefore, our
study stronglysuggests that the
different responses induced
by TNFa in myeloid cells are dependent on theSPM plasma
membrane transverse asymmetry.
0 1996 b y The American Society of Hematology.
T
are mainly located within the outer leaflet of the cell plasma
membrane, whereas phosphatidylserine, phosphatidylethanolamine, and probably phosphatidylinositol are essentially
within theinner monolayer (for review, see DevauxJ6).However, while lipid asymmetry has been firmly established in
erythrocytes and to a lesser degree in some other eukaryotic
cells, it is important to note that SPM distribution is not
universal. The percentage of SPM in the outer layer of
plasma membranes canvary from 100% inrat redblood
cells to only 46% in KrebsI1 ascites cells.17 Therefore, since
SPM distribution appears to be variable depending on the
origin of the cells, it is conceivable that cells that exhibit a
high degree of transverse SPM segregation, thereby displaying lower SPM levels in the inner leaflet, present a
smaller hydrolysable SPM pool compared with cells in
which the SPM is more equally distributed between the
plasma membrane leaflets. If confirmed, such a hypothesis
would imply thatSPM organization in the plasma membrane
maysignificantlyinfluencethe
ceramide generation and,
consequently, the sensitivity to TNFa.
In this study, we have compared the SPM distribution, as
wellas the effect of TNFa on SPM metabolism, intwo
UMOR NECROSIS FACTOR a (TNFa), a multifactorial cytokine, has been previously reported to play an
important role in the control of both proliferation and differentiation of normaland malignant myeloidcells.’.’Ithas
been shown that TNFa exerts variable effects on myeloid
cell proliferation. Indeed, the proliferation of some established cell lines and fresh cells can be strongly inhibited by
TNFa, whereas other leukemic myeloid cell populations are
inherently less sensitive to TNFa.’,4For example, it has been
previously reported that U937 and HL-60 cells were highly
sensitive to TNFa, whereas HEL, KG1, and KGla cells were
muchmore r e ~ i s t a n t . The
~ , ~ basis for these differences is
presently unknown.
Recent studies have shown that TNFa triggers apoptosis
in U937 and HL-60 cell^.^.^ In these cells, TNFa induces a
sphingomyelin (SPM) cycle in which a sphingomyelinase
(SPMase) is activated and hydrolyses SPM to produce ceramide leading to activation of transcriptional factors such as
NF-kB.7 In this signaling pathway, ceramide seems to act as
second messenger and addition of synthetic ceramide analogs or exposure to bacterial SPMase mimicked the biological effects of TNFa.*-” The molecular basis of ceramideinduced apoptosis is poorly understood. However, multiple
targets for the direct or indirect action of ceramide have been
identified, including a membrane kinase, mitogen activated
protein kinase (MAPK),I2a Ser/Thr phosphatase of the PP2A
subfamily, termed ceramide-activated protein phosphatase
(CAPP),” and protein kinase C<.’4Regardless of the mechanism by which ceramide induces apoptosis, it is conceivable
that either insufficient production of ceramide or modificationin ceramide activity on its specific target(s) may decrease TNFa-induced cytotoxicity in myeloid cells. Altematively, the SPM-ceramide pathway may be impaired due to
a lack of SPMase activation or a decrease in cellular SPM,
which is involved in the SPM cycle. Indeed, Linardic and
HannunI5 have identified in HL-60 and U937 cells a unique
signaling pool of SPM that undergoes significant hydrolysis
in response to TNFa. This pool of SPM appears to be localized in the inner leaflet of the plasma membrane.
It is generally accepted that SPM and phosphatidylcholine
Blood, Vol 88, No 4 (August 15), 1996: pp 1465-1472
From the CJF INSERM 9503, Centre Claudius Re‘gaud, INSERM
Unite‘ 326, Hdpital Purpan, and the Service d’He‘matologie, Centre
Hospitalier Universitaire Putpan, Toulouse, France.
Submitted December 4, 1995; accepted April 16, 1996.
Supported in part by I’INSERM (CRE n”920411, G.L.), I’ARC
(grants 6749, G.L.; and 2069, J.P.J.),and la Fe‘diration Nationale
des Centres de Lutte contre le Cancer (J.P.J.). A.B. isthe recipient
of a grant from La Fondation Contre la Leuce‘mie (Paris) and la
Ligue Contre le Cancer du Gers (Auch), France.
Address reprint requests to GuyLaurent, MD, Service d’Hhatologie Clinique, CHU Purpan, place du Dr Baylac, 31059 Toulouse,
France.
The publication costsof this article were defrayedin part by page
chargepayment. This article must therefore be hereby marked
“advertisement” in accordance with 18 U.S.C. section 1734 solely to
indicate this fact.
0 1996 by The American Society of Hematology.
0006-4971/96/8804-0013$3.00/0
1465
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
BETTAIEB ET AL
1466
human leukemic cell lines, KGla and U937, which differ in
their sensitivity to TNFa. We report that, in contrast to U937
cells, TNFa triggers neither apoptosis nor SPM hydrolysis
in KG1a cells but rather induces SPM synthesis and ceramide
breakdown. Furthermore, we found that the SPM pool localized in the inner leaflet of the plasma membrane was dramatically reduced in KGla cells, compared with U937 cells,
suggesting that SPM distribution may profoundly influence
the signaling pathway of TNFa. The potential implications
of these studies for pharmacologic manipulation of cell death
are discussed.
MATERIALS ANDMETHODS
Materials. RPMI 1640, fetal calf serum (FCS), and L-glutamine
were purchased from Eurobio (Les Ulis, France), and Iscove's Modified Dulbecco's medium (IMDM), penicillin, and streptomycin from
Gibco BRL (Cergy-Pontoise, France). Human recombinant TNFa
was supplied from PeproTech-Tebu (Le Perray-en-Yvelines,
France). (MethyL'H) thymidine (79 Cilmmol), and [9,10 (n)-'H]
palmitic acid (53 Ci/mmol) were purchased from Amersham (Les
Ulis, France) and choline chloride ( m e t h ~ l - ~ H
(81
) Ci/mmol) from
Du Pont-NEN (Les Ulis, France). Sphingomyelinase (Staphylococcus aureus), phospholipase A, (PLA2)(Naja naja venom), phospholipase C (PLC) type XI (Bacillus cereus), phospholipase D (PLD)
(Cabbage), N-hexanoyl-D-sphingosine(C,-ceramide), I4C-Concanavalin A, other reagents and all solvents were obtained from Sigma.
The N-acetylsphingosine (C,-ceramide) was purchased from
Matreya (Pleasant Gap, PA).
Cell culture. The human myeloblastic cell line KGla (purchased
from the ATCC) was cultured in IMDM supplemented with 20%
FCS. The human monocytic cell line U937 was purchased from the
ATCC and maintained in culture in RPMI 1640 medium supplemented with 2 mmol/L glutamine, 100 U/mL penicillin, 100 pg/mL
streptomycin, and 10% FCS. Cells were maintained at densities
between 1 X 1Os and 1 X 10' cells/mL in a fully humidified incubator
containing 5% CO, at 37°C. Immunophenotypic analysis showed
that KGla cells exhibited an early (immature) phenotype (CD34',
CD38-, CD33-), whereas U937 cells showed a mature phenotype
(CD34-, CD38'. CD33+).'*Both cell lines expressed both p55 and
p75 TNF receptors.'
Cell morphology. Cell morphology was evaluated using MayGriinwald-Giemsa staining. Treated or untreated cells were collected, washed, and mounted on glass slides by cytospin centrifugation. The cells were stained and observed under microscope. Extent
of apoptosis was determined by the number of cells undergoing
chromatin condensation and fragmentation, cytoplasmic volume
contraction, and zeiosis.
Determination of DNA fragmentation. DNA fragmentation analysis was performed as previously described4 with slight modifications. Cells, resuspended in RPMI or IMDM supplemented with
10% FCS at 2.5 X lO'lmL, were incubated with 0.5 pCi1lO' cells
of [methyl-'H]-thymidine and cultured at 37°C. After 24 hours incubation, the radioactive medium was removed and cells were resuspended in fresh medium and treated with the different agents and
harvested. Cell pellets were lysed in Tris-HCI 15 mmoVL, EDTA
20 mmoVL, Triton-X-l00 0.5%, pH 8.0. After incubation on ice for
30 minutes, lysates were centrifuged for 30 minutes at 20,OOOg. The
supernatant containing the low molecular weight DNA and the pellet
(intact chromatin DNA) were collected into scintillation fluid and
counted. The percentage of DNA release was calculated as follows:
% DNA fragmentation = [cpm (supernatant)/cpm (supernatant +
pellet)] X 100. Specific fragmentation was calculated as follows:
% specific fragmentation = [(% DNA fragmentation - % spont
fragmentation)/(lOO - % spont fragmentation)] X 100 (% spont
fragmentation: % spontaneous DNA fragmentation in untreated
cells).
DNA isolation and gel electrophoresis. Low molecular weight
DNA was purified as previously described" with minor modifications. Treated and untreated cells ( 5 X 10') were pelleted and lysed
with 600 pL of Tris-HCI 15 mmoUL, EDTA 20 mmol/L, Triton X100 O S % , pH 8.0. After incubation at 4°C for 30 minutes, the lysate
was centrifuged at 20,000g for 30 minutes. The supematant was
twice subjected to extraction with phenol/chlorophorm/isoamylalcohol (25:24: 1) followed by ethanol precipitation. The precipitate was
dissolved in 20 pL Tris-EDTA buffer and incubated for 30 minutes
with RNAse A (250 pg/mL) at 37°C. DNA fragments were electrophoresed in a 2% agarose gel, and then visualized by ethidium
bromide staining.
Metabolic labeling, extraction, and analysis of cellular phospholipids. For phospholipid labeling, cells were incubated in RPMI
medium containing 1% FCS and [9,1O(n)-'H]palmitic acid ( I pCi1
mL). After 48 hours incubation, the radioactive medium was removed and cells were incubated for 2 to 4 hours in 1 % FCS enriched
medium. Cells ( S X 10' resuspended in 3 mL of RPMI 1640 medium
supplemented with 25 mmol/L HEPES) were treated with 20 ng/
mL TNFa at various times and harvested. Lipids were extracted by
the method of Bligh and Dyer2" and were separated by thin layer
chromatography (TLC) using chloroform/methanol/water(100:42:6,
by vol) followed by a second step using petroleum ether1diethylether1
acetic acid (80:20:1, by vol) or hexane/diethylether/formic acid
(55:45:1, by vol) as developing solvent systems. Radioactive lipid
spots, detected with Berthold radiochromatoscan and upon exposure
to iodine vapor, were scraped into scintillation fluidand counted.
The positions of ceramide on TLC plates were determined by comparison with a concomitantly run ['HI-lipid extract from U937 and
KGla cells treated with exogenous bacterial SPMase ( 100 mU/I .S
X IOh cells).
[.'H]-thymidine incorporation and proliferution of U937 and
KGla cells. KGla and U937 cells (2-4 X IO' cells1200 pL per
well) were resuspended in IMDM supplemented with l % FCS and
incubated with various concentrations of TNFa in a fully humidified
incubator containing 5% CO, at 37°C for 3 days. ['Hjthymidine (0.5
pCi/mL) was then added 16 hours before harvesting. The incorporationof radioactivity was determined by scintillation counting. For
measuring cell proliferation, U937 and KGla cells treated with or
without TNFa were collected after 4 days incubation. The numbers
of viable cells were counted using hemocytometer in triplicate in
the presence of trypan blue. Statistical analysis was performed by
Student's t test.
Preparation of nuclear extracts. Cells ( 5 X 10' in 5 mL of
FCS-free IMDM) were treated separately with the different effectors
(TNFa, bacterial SPMase, and synthetic ceramide) for 1 hour after
which nuclear extracts were prepared as previously described.' Cells
were then washed twice with cold phosphate-buffered saline (PBS)
and suspended in 400 pL of lysis buffer (10 mmol/L HEPES, pH
7.8, I O mmol/L KC], 0.1 mmol/L EDTA, 0.1 mmol/L EGTA, I
mmol/L dithiothreitol (DTT), 1 mmoUL phenylmethylsulfonyl fluoride (PMSF), 2 pmol/L pepstatin, 0.6 pmol/L leupeptin, 1 pg/mL
aprotinin, and 0.6% nonidet P-40. After 15 minutes on ice, the
nuclear pellet was recovered after centrifugation at 1200g and resuspended in 50 pL of 20 mmol/L HEPES, pH 7.9, 0.4 moVL NaC1,
1 mmoUL EDTA, 1 mmoUL EGTA, and the tubes were incubated
on ice for 30 minutes with intermittent mixing. These nuclear extracts were then centrifuged 30 minutes at 20,0001: and supernatants
were either used immediately or stored at -20°C for later use.
Electrophoretic mobility shift assa.ys (EMSAs). Electrophoretic
mobility shift assays were performed by incubating 4 pg of nuclear
extract with 50,OOO cpm of'*P-end labeled NF-kB (5'AG'M'GAGGGGCTTTCCCAGGC-3') consensus oligonucleotides in a reaction buffer (2 mmol/L HEPES, pH 7.5, 50 mmollL NaCI, 0.5 mm011
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
1467
SPHINGOMYELINSEGREGATION AND TNFa SIGNALING
L EDTA, 1 m m o K MgCI2, 4% glycerol, 0.5 m m o m D m , 1 pg
poly [dI-dC], and 2 pg BSA). After 20 minutes incubation, the DNAprotein complex formed was separated from free oligonucleotide by
electrophoresis through a low ionic strength 4% native polyacrylamide gel. Electrophoresis was carried out in buffer containing 6.7
mmom Tris-HC1 (pH 7.9). 3.3 mmol/L sodium acetate, 2 mmol/L
EDTA at 10 V/cm for 3 hours at room temperature with buffer
recirculation. The gel was dried and autoradiographed with intensifying screens at -70°C.
Plasma membrane isolation. Isolation of cell plasma membrane
was performed according to Record et al." This method has been
applied to various cell models, such as Krebs-I1 ascitic cells, HL60 cells, A431 cells, and n e ~ t r o p h i l s . ~ ~Briefly,
~ * " ~ ~cells (SO to 100
X lo6) resuspended in 5 mL PBS were treated with 0.5 pCi I4CConcanavalin A (ConA) for 10 minutes at room temperature. 14CConA surface-labeled cells, resuspended in 100 mmol/L KCl, 5
mmol/L MgCL, 1 mmol/L ATP, 25 mmol/L Tris-HC1 (pH 9.6),
were homogenized in a nitrogen cavitation bomb (Kontes, Vineland,
NJ) after a 5 minute period with 40 atm of NZat 4°C. Four milliliters
of postnuclear supernatant was loaded onto a mixture of 1 1 mL
Percoll, 2.2 mL distilled water, buffered with 4.8 mL of 400 mmoll
L KCI, 20 mL MgCI,, 100 mmol/L Tris-HC1 (pH 9.6). Membranes
were then separated by centrifugation at 1.29 X 10'' rad' X S", ie,
160,000g for 10 minutes at plateau in a Beckman rotor 60 Ti. Fractions of 2 mL were harvested from the top of each tube, diluted to
adjust the pH to 7.4, and
recentrifuged using a 50 Ti rotor at 200,OOOg
for 45 minutes to eliminate Percoll. Plasma membranes, corresponding to the peak of l4C-ConAacross the gradient were then collected.
Efficiency of the plasma membrane isolation and purity ofthe preparation has already been reported by us.25.26
Analysis of cellular SPM contentandtransversedistribution.
Analysis of transverse SPM distribution in the plasma membrane
was performed in U937 and KGla cells according to the two step
procedure as previously de~cribed."~'~
In the first step, total cells (1.5
X 106/mL)were incubated with 100 mU/mL of bacterial SPMase at
37°C. The reaction was stopped by the addition of 10 mmom EDTA
on ice. After harvesting the cells at the indicated times, lipids were
extracted by the method of Bligh and Dye?' and separated on TLC
using chlorofonn/methanol/water (70:35:5, by vol). The various
spots detected after exposure to iodine vapor were determined for
phosphorus content according to BBttcher et al.27 The percentage of
SPM, compared with total cell phospholipids in untreated (a) and
SPMase-treated cells (b) were measured; the difference giving the
percentage of hydrolyzed SPM (c), (a - b = c). In the second
step, the percentages of SPM, compared with total phospholipids in
purified cell plasma membrane (d), as well asthe ratio of total
membrane phospholipid versus total cell phospholipid (e) were also
determined. The percentage of hydrolyzed SPM, compared with total
membrane phospholipids (outer leaflet SPM) was (f) = de. Thus,
the inner leaflet SPM was (g) = d - f. In summary, the inner leaflet
SPM content, compared with total membrane phospholipids, was
calculated according to the following formula: g = d - [(a - b)/e]
as in ref 17 (see Table 1).
RESULTS
Effect of TNFa on KG1 a and U937 DNA fragmentation.
KGla andU937 cells were treated with 10 to 30 ng/mL
TNFa for 6 hours and analyzed for DNA fragmentation
using the L3Hlthymidine release assay. In U937 cells, TNFa
caused 15% to 20% of DNA fragmentation (Fig 1A).In
contrast, TNFa (in the same concentration range) did not
induce DNA fragmentation in KGla cells (Fig 1A); the inability of TNFa to induce apoptosis in KGla cells was observed within a large range of concentrations (10 to 100 ng/
mL) (data not shown). Similar results were observed when
Table 1. SPM Content and Distribution in KGla and U937 Cells
Values (%)
K G l a Cells
U937 Cells
a
b
d
e
18.70
7.90
27.90
0.40
0.90
12.50
6.40
18.80
0.49
6.40
(1
Values: a, percent of SPM, compared with total cell phospholipids
in untreated cells; b, percent of SPM, compared with total cell phospholipids in SPMase-treated cells; d. percent of SPM, compared with
phospholipids of the cell plasma membrane; e, ratio of total plasma
membrane phospholipids versus total cell phospholipids; g, percent
of the inner-associated SPM content, compared with plasma membrane phospholipids, was calculated according to the following formula: g = d - [(a - b)/el.
extent of apoptosis was evaluated by morphology (Fig 1B).
Internucleosomal degradation of DNA was confirmedby the
appearance of DNA ladders (Fig 1C).
Effect of TNFa on SPM metabolism in K G l a and U937
cells. The lack of apoptotic response of KGla cells to
TNFa prompted us to evaluate the influence of TNFa on
SPM metabolism in these cells compared with U937. Cells
were prelabeled with [9,10-3H]palmiticacid to equilibrium
for 48 hours, then incubated with TNFa at 20 ng/mL, and
harvested at various times. As illustrated in Fig 2A, TNFa
induced in U937 cells SPM hydrolysis, which reached a
maximum (approximately 20%) at 15 to 20 minutes cell
treatment. A significant boost in intracellular ceramide concentrations was observed at approximately the same time as
SPM hydrolysis (Fig 2B). In contrast, in KGla cells, TNFa
(within the same concentration range) induced cellular SPM
synthesis and concomitant decrease of intracellular ceramide
(Fig 2, C and D). The increase of SPM reached a plateau at
about 20 minutes (Fig 2C). Ceramide content decreased rapidly within the first 10 minutes and then remained constant
up to 30 minutes (Fig 2D).
Exogenous SPMase and ceramide induce apoptosis in
KGla. In order toconfirm that the failure of ceramide
generation was the major responsible intermediate for the
lack of apoptotic response to TNFa in KGla, we treated
these cells for 6 hours with bacterial SPMase at 100 mu/
mL (generating natural ceramide) or with 20 pmol/L synthetic cell permeant C6-ceramide. As shown in Fig 3A and
B, SPMase as well as synthetic ceramide caused a significant
DNA fragmentation in both KGla and U937 cells, and cells
morphologically presented typical apoptotic features (data
not shown). The effect of SPMase was specific since phospholipase C (PLC) (Fig 3), PLAl, PLD, 1,2-dioctanoyl-snglycerol (DAG), or the 4,5 C*-dihydroceramide (data not
shown) did not induce DNA fragmentation. It is interesting
to note that bacterial SPMase and C6-ceramide were able to
induce higher DNA fragmentation in KGla than in U937
cells. Internucleosomal degradation of DNA was confirmed
by the appearance of DNA ladders (Fig 3C).
Taken together, the above results indicate that, in U937
cells, apoptotic cell death induced by TNFa was mediated
by ceramide released from SPM hydrolysis. In contrast, the
failure of TNFa to induce apoptosis was to be related to the
lack of ceramide generation.
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
BETTAIEBETAL
-TNF
TA
Fig 1. Detection of DNA fragmentation in U937 and KGla cells
after TNFa treatment. Detection of DNA fragmentation was analyzed
by three different methods: the PHlthymidine release assay (A), cell
morphology analysis using May-Grunwald-Giemsastaining (B), and
agarose gel electrophoresis IC). Cells were left untreated, or stimulated for 6 hours with 20 ng/mL TNFa. Extent of apoptosis was determined as described in Materials and Methods. Results are mean f
SD of three independent experiments.
<
KGla
U937
U
.r(
5 15Y
3
alO-
a
yc
O
S
50 .
'KGla
KG la
U937
11937 c
"
E e c t of TNFa, SPMase, and synthetic ceramide on KGIa
and U937 cell growth. Since we identified substantial differences between KGla and U937 cells in TNFa-activated
signaling pathways, we expected different biological effects
of TNFa on proliferation and viability of KGla and U937
cells. In fact, using the trypan blue exclusionassay, we found
that TNFa enhanced KGla cell growth with an optimal effectat25 ng/mL. The TNFa-proliferative effect persisted
for TNFa concentrations up to 200 ng/mL (Fig 4). TNFa
also induced an increase of ['Hlthymidineincorporation
reaching a maximum of 150% at 100 ng/mL of TNFa. In
contrast, TNFa induced a cytotoxic effect on U937 cells, as
well as a reduction of ["Hlthymidine incorporation (Fig 4).
Effect of TNFa, SPMase, and synthetic cerarnide on NFkB activation. Since TNFa and ceramide can activate transcriptional factors such as NF-kB,'weexaminedwhether
TNFa and the second messenger molecule ceramide are able
to induce NF-kB activation in the two cell lines. KGla and
U937 cells were incubated with TNFa (20 ng/mL), SPMase
(100 mU/mL), and synthetic C6-ceramide (20 pmol/L) for
60 minutes at 37°C and then examined for NF-kB activation
by electrophoretic mobility shift assay. The results in Fig 5
indicate that in both cell lines NF-kB was activated equally
well either with TNFa or SPMase and permeant ceramide.
Transverse distribution of SPM in KGla and U937 cell
lines. The lack of apoptotic response of KGla cells to
TNFa, and the ability of exogenous ceramide to restore
apoptosis in these cells indicate that the SPM-ceramide pathway is altered in KGla cells upstream of the ceramide generation. Therefore, we hypothesized that the cellular content
and/or plasma membrane distribution of SPM could be different between KGla andU937 cells. The totalmass of
cellular SPM was 389.60 2 35.6 and 264.37 ? 9.24 nmol
Pi/l Ox cells for KG 1a and U937 cells, respectively. The percentage of SPM, compared withtotal cell phospholipids,
was about 1.5 times higher in KGla cells (18.70%) than in
U937 cells (12.50%). The percentage of SPM, compared
withplasma membrane phospholipids, was also higher in
KGla cells (27.9%) than in U937 cells (18.8%). Moreover,
the analysis of transverse SPM distribution across plasma
membranes demonstrated that cells with the highest SPM
content (KGla) exhibited the lowest amount of SPM within
the inner leaflet. Indeed, as shown in Table 1, the values of
inner leaflet SPM content were 0.9% and 6.4%of membrane
phospholipids for KGla and U937cells, respectively. Therefore, in KGla cells, the percentage of SPM in the inner
leaflet, compared with plasma membrane, was [0.9/27.9] X
100 = 3.2%. whereas the inner leaflet SPM was as high as
34% ([6.4/18.8] X 100) in U937 cells (Fig 6).
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
SPHINGOMYELIN SEGREGATION AND TNFw SIGNALING
1469
140 1
Fig 2. Induction of SPM metabolism in U937 (A
and B) and KGla (C and D) cells treated with TNFa.
[9,10-JHlpalmitic acid labeled cell lines were treated
with TNFa (20 nglmL) for the time intervals indicated. Aliquots were then collected and lipid extracted. Labeled SPM (closed circles) and ceramide
(open circles) were resolved by analytical thin-layer
chromatographyas described in Materials and Methods.Results are expressed as percentage ofuntreated controls.For U937 cells (A and B), control
SPM and cerrmide counts were 54,000 2 11,250 and
4,166 f 520 cpm, respectively;the data are from one
representative expariment of two independent experiments. For KGla cells (C and Dl, control SPM and
ceramide counts were 51,000 ? 7,500 and 5,390 ?
950 cpm, respectively; results are mean 2 SEM of
four independent experiments.
0
10
20
30
40
50
0
-I
110
ILU
m
I
40
0
10
DISCUSSION
Several studies have indicated that TNFa triggers
apoptosis through a SPM pathway signaling. Ceramide is
emerging as a major mediator candidate for the action of
cytotoxic effectors including TNFa, ionizing radiation, and
cytotoxic drugs. The aim of the present investigation was to
compare the effects of TNFa on SPM metabolism in
apoptosis-resistant KGla cells as compared with apoptosissensitive U937 cells.
TNFa induced neither DNA fragmentation nor morphologic changes in KGla cells, whereas TNFa triggered
apoptosis in U937 cells as expected from previous studies6
The lack of apoptotic response of KGla cells was not due
to the lack of type I ( p 5 3 or type I1 ( p 7 3 TNF receptor^.^.^'
Synthetic C2-or C,-cell permeant ceramide as well as natural
ceramide generated by bacterial SPMase, resulted in internucleosomal DNA fragmentation in both KGla and U937cells.
These results strongly suggested that the lack of apoptotic
response to TNFa by KGla cells was most likely linked to
a lack of TNFa-induced ceramide generation. In order to
test this hypothesis, we evaluated the effect of TNFa on
SPM metabolism in KGla cells, compared with U937 cells.
In U937 cells, we found that TNFa induced SPM hydrolysis and ceramide generation as expected.6 In contrast, in
KGla cells we found that TNFa induced neither SPM hydrolysis nor ceramide generation, but rather SPM synthesis with
a concomitant decrease of the cellular ceramide content. The
mechanism which underlies this unexpected effect on SPM
metabolism is unclear. One could speculate that, in KGla
cells, TNFa induced SPM synthesis through the transfer of
phosphorylcholine head group from PC to ceramide. The
lack of SPM hydrolysis and subsequent ceramide generation
may explain the lack of apoptotic response of KGla cells to
TNFa. Several hypotheses can be put forward to explain
this finding. For example, the inability of TNFa to activate
20 30 40
Time (min)
50
cl
10
20
30
40
50
D
T
m
0
10
20 30 40
Time (min)
50
SPMase and/or a decrease of SPMase activity may account
for this observation. Alternatively, since it has been shown
that the SPM pool used for TNFa signaling is predominantly
located in the inner leaflet of the plasma membrane, the
SPM availability and its levelin the inner leaflet maybe
accountable for the lack of its hydrolysis by TNFa. Indeed,
we found that the size of this SPM pool was 7-fold lower
in KGla cells, compared with U937 cells.
The fact that TNFa activated different signaling pathways
and induced opposite cellular effects in KGla andU937
cells, prompted us to evaluate the action of TNFa on NFkI3, a transcriptional factor that was identified as reflecting
cellular response to TNFa. In both cell lines, we found that
TNFa aswell as ceramide activated N F - B , as expected
from previous studies.’ This observation suggests that at least
part of the TNF-receptor signal transduction mechanism is
shared by both cell lines.
It is becoming apparent that TNF-signal transduction is
accomplished via association with an emerging class of
noveland
diverse signaling molecules such as FADD/
MORT 1, TRADD, and IAP, which are capable of regulating
apoptosis when overexpressed in cell^.*^-^' In addition, a
recent study showed that a dominant-negative derivative of
FADD blocked TNFa-induced apoptosis as well as ceramide
generation in a lymphoma cell line, indicating that ceramide
occurred after FADD a~tivation.~’
Our results do not overide
the implication of these molecules in the cell death machinery. However, we can hypothesize that the SPM content at
the inner leaflet of plasma membranes maybe a limiting
factor in the interaction between selected members of the
TNFa-receptor superfamily and their associated proteins.
Even though KGla cells and plasma membranes contain
more SPM than those of U937 cells, the distribution of SPM
across the plasma membrane appears to be a regulatory factor
for ceramide generation. In that respect, KGla cells treated
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
BETTAIEB ET AL
1470
50
0
0
W
SPMase
PLC
Ceramide
10
*
in-"'
8n
KGla
U937
0
KGla
L
I
Q)
U
.-
-0 8 E
8
45
L
n
225 250
KGla
I
c
100 125 150 175
200
U937
L7937
L
c,
25
50
75
I
1
Q)
U
.F
t
8 e
(L,
0
cn
Fig 3. Detection of apoptosis in U937 and KGla cells after bacterial
SPMase and synthetic ceramide treatment. Cells were left untreated,
or stimulated for 6 hours with bacterial SPMase (100 mU/mL), PLC (100
mUlmL), or thecell permeant &-synthetic ceramides (20 pmol/L), and
harvested. Detection of DNA fragmentation was analyzed by three different methods the ['Hlthymidine release assay (A), cell morphology
analysis using May-GrOnwald-Giemsastaining (B), and agarose gel electrophoresis of DNA IC)as described in Materials and Methods. Resutts
are mean t SD of at least three independent experiments.
TNFa concentration
(ng/mL)
Fig 4. Effect of TNFa on KGla (closed circles) and U937 (open
circles) cell growth. (A) Cells (2-4 x 104/200p L per well) were incubated for4 days with various concentrations of TNFa. After incubation, cells were collected and counted under the microscope in the
presence of trypan blue. (B) Cells (2-4 x 104/200 p L per well) were
incubated for3 days with various concentrations of TNFrr, and then
0.5 pCi of ['Hlthymidine were added for an additional 16 hours incubation and harvested as described in Materials and Methods. The
amount of ['Hlthymidine incorporated in untreated KGla and U937
cells were 179,000 ? 10,240 and 136,000 2 8,650 cpm, respectively.
Results are the average of triplicate samples and are expressed as
percent of controls. Data represent the mean ? SD of three independent experiments. Asterisks indicate P c .01 compared with pretreatment values.
with exogenous SPMase generated more ceramide than
SPMase-treated U937 cells (data not shown). Subsequently,
more DNA fragmentation was obtained in SPMase-treated
KGla cells as compared with SPMase-treated U937 cells.
Therefore, even if other factors interfere with ceramide release in TNFa-treated KGla cells, our study indicates that
the transverse SPM segregation across the plasma membrane
influence ceramide generation by modifying the ratio of hydrolysable versus nonhydrolysable SPM pools. To the best
of our knowledge, the asymmetrical organization of phospholipids in the membrane of human leukemic cells had not
been investigated before (for review see Zachowski"). It is
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
1471
SPHINGOMYELINSEGREGATION AND TNFu SIGNALING
important to note that KG 1 a and U937 cells are derived from
distinct granulomonocytic lineages. KG 1 a cells display an
immature phenotype (CD34' CD38- CD33-), whereas U937
cells are more mature (CD34- CD38' CD339.l' Interestingly, other TNFa-resistant myeloid cells such as KG1 and
HEL cells' also display an immature phenotype, and respond
similarly to TNFa as K G l a cells'x (ie, these cell lines do
not undergo apoptosis when treated with TNFa). Furthermore, these cell lines seem to present distinct SPM distribution compared with the mature cell lines (CD34-) such as
U937 and HL-60, whichreadily undergo apoptosis when
treated with TNFa. Indeed, we found that bacterial SPMasetreated KG1 and HEL cells generated more ceramide than
the more mature cell lines U937 and HL-60. Although the
characterization of SPM asymmetry has not yet been fully
completed, these observations reflect the greater SPM content in the outer leaflet of KGla and HEL cells.
TNFa displayed different biological effects on KG la and
U937 cells. Indeed, whereas TNFa-induced apoptosis in
U937 cells, it triggered KGla cell proliferation in a dosedependent manner. It has been recently reported that TNFa
stimulates the less mature multipotential progenitors in
mouse long-term bone marrow cultures, and inhibits the further development of committed myeloid progenitors, suggesting that TNFa may exert opposite effects on cellular
proliferation of myeloid cells depending on their differentiation status.34Since KG l a and U937 cells are putative representatives of immature and mature myeloid differentiation
stages, respectively, our study suggests that the effects of
TNFa on myeloid cell proliferation depend on SPM metabolism and distribution.
Finally, our study provides evidence that TNFa may induce different responses in myeloid cells depending on their
relative cellular levels of ceramide. Among all the possible
1
2
3
4
5
A
B
m
d.
Fig 5. Effect of TNFu and ceramide on NF-kBactivation. U937 and
KGla cells (5 x 106)were incubated with TNFa (20 nglmL). SPMase
(100 mUlmLI, or synthetic C6-ceramide (20 pmollL1 for 60 minutes
at 37°C. Nuclearextracts were isolated, and EMSAs were performed,
as described in Materials and Methods. (A) Nuclear extracts from
U937 treated with TNFa (lane 3), SPMase (lane 4). or synthetic C6ceramide (lane 5). (B) Nuclear extracts from KGla treated with TNFu
(lane 3). SPMase (lane 41. or synthetic C6-ceramide(lane 5). The control in lane 1 indicates untreated cells.Specificity of DNA-protein
complex formation was demonstrated by competition with excess
unlabeled NF-kB probe (lane 2).
1
Outer leaflet
Inner leaflet
-1uu
KGla
U937
Fig 6. Transverse distribution of SPM in plasma membrane of
KGla and U937 cells. Results are expressed as a percentage of the
total SPM present in plasma membranes.
metabolic regulations of ceramide generation, we report here
that a structural membrane organization of SPM may constitute a limiting step for generation of this second messenger.
ACKNOWLEDGMENT
The authors would like to thank DrsB. Hill (Pierre Fabre, Castres).
T. Levade (CJF INSERM 92-06, Toulouse), M. Allouche, A.QuilletMary, and D. Lautier (CJF INSERM 9503, Toulouse) for critical
comments and C. Bordier for technical assistance.
REFERENCES
I . Peetre C, Gullberg U, Nilsson E. Olsson I: Effects of recombion proliferation and differentiation of
leukemic andnormal
hemopoietic cells in vitro. J Clin Invest
78: 1694. 1988
2. Johnson CS. Chang MJ, Furmanski P:In vivo hematopoietic
effects of tumor necrosis factor-a in normaland erythroleukemic
mice: Characterization and therapeutic applications. Blood 72: 1875,
1988
3. Munker R, DiPersio J, KoefflerHP: Tumor necrosisfactor:
Receptors on hematopoietic cells. Blood 70:1730, 1987
4. Higuchi M, Aggarwal BB: Differential rolesoftwo types of
the TNF receptor in TNF-induced cytotoxicity, DNA fragmentation,
and differentiation. J Immunol 152:4017, 1994
5. Kim MY, Linardic C. Obeid L, Hannun YA: Identification of
sphingomyelin turnover as an effector mechanism for the action of
tumor necrosis factor and y-interferon. J Biol Chem 266:484, 1991
6 . Obeid LM, Linardic CM.KarolakLA,
Hannun YA:Programmed cell death induced by ceramide. Science 259:1769. 1993
7. Schiitze S, Potthoff K, Machleidt T, BerkovicD,Wiegmann
K, Kronke M: TNF activates NF-kB by phosphatidylcholine-specific
phospholipase C-induced "acidic" sphingomyelin breakdown. Cell
7 1:765, I992
8. Johns LD, Sarr T, Ranges G: Inhibition of ceramide pathway
does not affect ability of TNF a to activate nuclear factor-& J
lmmunol 1525877, 1994
9.KolesnickRN,
Golde DW: The sphingomyelin pathway in
tumor necrosis factor and interleukin-l signaling. Cell 77:325, 1994
IO. Hannun YA: The sphingomyelin cycle and the second messenger function of ceramide. J Biol Chem 269:3215. 1994
11. Heller RA, Kronke M: Tumor necrosis factor receptor-mediated signaling pathway. J Cell Biol 1265, 1994
nant tumor necrosis factor
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
1472
12. Raines MA, Kolesnick RN, Golde DW: Sphingomyelinase
and ceramide activate mitogen-activated protein kinase i n myeloid
HL-60 cells. J Biol Chem 268:14572, 1993
13. Dobrowsky RT, Hannun YA: Ceramide stimulates a cytosolic
protein phosphatase. J Biol Chem 2675048, 1992
14. Lozano J, Bema E, Municio MM, Diaz-Meca MT, Dominguez
I, Sanz L, Moscat J: Protein kinase isoform is critical for kBdependent promoter activation by sphingomyelinase. J Biol Chem
269:19200, 1994
15. Linardic CM, HannunYA: Identification of a distinct pool
of sphingomyelin involved in the sphingomyelin cycle. J Biol Chem
269:23530, 1994
16. Devaux PF: Static and dynamic lipid asymmetry in cell membranes. Biochemistry 30:1163, 1991
17. Record M, El Tamer A, Chap H, Douste-Blazy L: Evidence
for a highly asymmetric arrangement of ether- and diacyl-phospholipid subclasses in the plasma membrane of Krebs I1 ascites cells.
Biochem Biophys Acta 778:449, 1984
18. Bailly J-D, Muller C,Jaffrkzou J-P, Demur C, Cassar G,
Bordier C, Laurent G: Lack of correlation between expression and
function of P-glycoprotein in acute myeloid leukemia cell lines.
Leukemia 9:799, 1995
19. Ishida Y, Agata Y, Shibahara K, Honjo T: Induced expression
of PD-1, a novel member of the immunoglobulin gene superfamily,
upon programmed cell death. EMBO J 11:3887, 1992
20. Bligh EG, Dyer WJ: A rapid method of total lipid extraction
and purification. Can J Biochem Physiol 37:911, 1959
21. Record M, Laharrague P, Fillola G , Thomas J, Ribbes G,
Fontan P, Chap H, Corberand J, Douste-Blazy L: A rapid isolation
procedure of plasma membranes from human neutrophils using selfgenerating Percoll gradients. Importance of pH in avoiding contamination by intracellular membranes. Biochem Biophys Acta 819:1,
1985
22. Ribbes G, Ninio E, Fontan P, Record M, Chap H, Benveniste
J, Douste-Blazy L: Evidence that biosynthesis of platelet activating
factor (PAF-acether) by human neutrophils occurs in an intracellular
membrane. FEBS Lett 191:195, 1985
23.Vallari S, Record M, Snyder F: Conversion of alkylacetylglycerol to Platelet-Activating Factor in HL-60 cells and subcellular
localization of the mediator. Arch Biochem Biophys 171:548, 1990
BETTAIEB ET AL
24. Payrastre B, Plandavid M, Etievan C, Ribbes G, Carratero C,
Chap H, Douste-Blazy L: Characterization of plasma membranes
from A43 1 cells, isolated by self-generating percoll gradient: A rapid
isolation procedure toobtain plasma membranes with functional
epidermal growth factor receptors. Biochem Biophys Acta 939:355,
1988
25. TercC F, Record M, Ribbes G, Chap H,Douste-Blazy L:
Intracellular processing of cytidylyltransferase in Krebs I1 cells during stimulation of phosphatidylcholine synthesis. Evidence that a
plasma membrane modification promotes enzyme translocation specifically to the endoplasmic reticulum. J Biol Chem 263:3142, 1988
26. Record M, Bes SC, Chap H, Douste-Blazy L: Isolation and
characterization of plasma membranes from Krebs-I1 ascite cells
using Percoll gradients. Biochem Biophys Acta 699:57, 1982
27. Bottcher CJF, van Gent CM, Pries CA: Rapid and sensitive
sub-micro phosphorus determination. Anal Chim Acta 24:203, 1961
28. Tartaglia LA, Ayres TM, Wong GHW, Goeddel DV: A novel
domainwithinthe
55 kd TNF receptor signals cell death. Cell
74:845, 1993
29. Chinnaiyan AM, O’Rourke K, Tewari M, Dixit VM: FADD,
a novel death domain-containing protein, interacts with the death
domain of Fas and initiates apoptosis. Cell 81:505, 1995
30. Hsu H, Xiong J, Goeddel DV: The TNF receptor l-associated
protein signals cell death and NF-kB activation. Cell 81:495, 1995
31. ListonP,Roy N, Tamai K, Lefebvre C, Baird S , ChertonHorvat G, Farahani R, McLean M, Ikeda J-E, MacKenzie A, Korneluk RG: Suppression of apoptosis in mammalian cells by NAIP
and a related family of IAP genes. Nature 379:349, 1996
32. Chinnaiyan AM, Tepper CG, Seldin MF, O’Rourke K, Kischkel FC, Hellbardt S, Krammer PH, Peter ME, DixitVM:FADD/
MORT I is a common mediator of CD95 (FaslAPO- 1) and tumor
necrosis factor receptor-induced apoptosis. J Biol Chem 27 1 :496 I ,
I996
33. Zachowski A: Phospholipids in animal eukaryotic membranes: Transverse asymmetry and movement. Biochem J 294:l.
1993
34. Rogers JA, Berman JW: TNF-a inhibits the further development of committed progenitors while stimulating multipotential progenitors in mouse long-term bonemarrow cultures. J lmmunol
153:4694, l994
From www.bloodjournal.org by guest on June 17, 2017. For personal use only.
1996 88: 1465-1472
Opposite effects of tumor necrosis factor alpha on the
sphingomyelin- ceramide pathway in two myeloid leukemia cell lines:
role of transverse sphingomyelin distribution in the plasma
membrane
A Bettaieb, M Record, MG Come, AC Bras, H Chap, G Laurent and JP Jaffrezou
Updated information and services can be found at:
http://www.bloodjournal.org/content/88/4/1465.full.html
Articles on similar topics can be found in the following Blood collections
Information about reproducing this article in parts or in its entirety may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests
Information about ordering reprints may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#reprints
Information about subscriptions and ASH membership may be found online at:
http://www.bloodjournal.org/site/subscriptions/index.xhtml
Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American
Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.
Copyright 2011 by The American Society of Hematology; all rights reserved.