Lithium blocks the PKB and GSK3 dephosphorylation induced by

ceramide through protein phosphatase-2A

Alfonso Moraa,*, Guadalupe Sabioa, Ana Marıa Riscoa, Ana Cuendab, Juan C. Alonsoa,German Solera, Francisco Centenoa,1

aDepartamento de Bioquımica y Biologıa Molecular, Facultad de Veterinaria, Universidad de Extremadura, Avenida Universidad s/n, 10071 Caceres, SpainbMRC Protein Phosphorylation Unit, Department of Biochemistry, University of Dundee, Dundee DD1 4HN, UK

Received 12 September 2001; accepted 13 November 2001

Abstract

The biochemical mechanism of apoptosis induced by ceramide remains still unclear, although it has been reported that dephosphorylation

of PKB at Ser-473 may be a key event. In this article, we show that C2-ceramide (N-acetyl-sphingosine) induces the dephosphorylation of

both protein kinase B (PKB) and glycogen synthase kinase-3 (GSK3) in cerebellar granule cells (CGC). We also show that lithium protects

against the apoptosis induced by C2-ceramide by blocking the dephosphorylation of both kinases. Since lithium inhibits in vivo the observed

protein phosphatase-2A (PP2A) activation induced by ceramide, we hypothesise that the neuroprotective action of lithium may be due to the

inhibition of the PP2A activation by apoptotic stimuli. D 2002 Elsevier Science Inc. All rights reserved.

Keywords: Apoptosis; C2-ceramide; Cerebellar granule cells; GSK3; Lithium; PKB; PP2A

1. Introduction

Apoptosis is a form of cell death critical in the nervous

system for normal development and also implicated in

neurodegenerative diseases [1]. Although the biochemical

mechanism of the apoptotic process still remains unclear,

the implication of ceramide has been shown in many

different systems [2]. With regard to the way in which

ceramide induces apoptosis, it has been described that it can

regulate both protein kinase [3] and protein phosphatase

activities [4,5]. Among the described phosphatase activities

induced by ceramide, protein phosphatase-2A (PP2A) has

been reported [6,7]. Recently, it has been demonstrated

that phosphatidylinositol 3-kinase (PI3K)/protein kinase B

(PKB) pathway, which can be inactivated by PP2A [8], is

one of the main targets implicated in the induction of

apoptosis by ceramide [9]. In fact, short-chain cell per-

meable analogues, C2- and C6-ceramides, induced apoptosis

in several cell lines through dephosphorylation of PKB at

Ser-473 but without the inhibition of PI3K activity [10,11].

Among the described mechanisms, PKB regulates apop-

tosis by phosphorylation of glycogen synthase kinase-3

(GSK3) at Ser-9 in GSK3b and Ser-21 in GSK3a [12]

and then inactivates GSK3. In this regard, there are many

emerging data that show the proapoptotic implication of

GSK3 [13,14]. In fact, selective small-molecule inhibitors of

GSK3 activity protect primary neurones from death [15].

Lithium is widely used as a mood-stabilising drug to treat

manic–depressive disorders [16], and recently it is emer-

ging as a neuroprotective agent [14,17–23]. Although this

property has been explained by its uncompetitive inhibition

of GSK3 [24,25], our group has demonstrated a new action

mechanism: this ion may inhibit a serine–threonine phos-

phatase activate by potassium deprivation because it was

able to prevent the dephosphorylation of PKB and GSK3

induced by the apoptotic insult [26].

In this article, we check this hypothesis using another

apoptotic paradigm, C2-ceramide (N-acetyl-sphingosine),

0898-6568/01/$ – see front matter D 2002 Elsevier Science Inc. All rights reserved.

PII: S0898 -6568 (01 )00282 -0

Abbreviations: C2-ceramide, N-acetyl-sphingosine; CGC, cerebellar

granule cells; GSK3, glycogen synthase kinase-3; MTT, 3-[4,5-dime-

thylthiazol-2-yl]-2,5-diphenyltetrazolium bromide; PI3K, phosphatidylino-

sitol 3-kinase; PKB, protein kinase B; PP2A, serine– threonine protein

phosphatase-2A

* Corresponding author. Tel./fax: +34-92-7257-160.

E-mail addresses: almoraco@unex.es (A. Mora), pacenten@unex.es

(F. Centeno).1 Also corresponding author.

www.elsevier.com/locate/cellsig

Cellular Signalling 14 (2002) 557–562

where lithium also protects cerebellar granule cells (CGC)

against apoptosis [20]. We observed that C2-ceramide

induced dephosphorylation of PKB and GSK3, and lithium

prevented the dephosphorylation of both kinases induced by

the apoptotic insult. As C2-ceramide induced the activation

of PP2A, which is blocked by lithium, these results corrob-

orate a new model for the neuroprotective effects of lithium:

in addition to being an uncompetitive inhibitor of GSK3,

this new point of view reveals that lithium may be an

inhibitor of a serine–threonine phosphatase activated by

the apoptotic stimuli.

2. Materials and methods

2.1. Materials

Complete protease inhibitor cocktail tablets were from

Roche. Tissue culture reagents and cytosine arabinoside

were from Sigma. Secondary antirabbit and antisheep

IgG antibodies coupled with horseradish peroxidase were

from Pierce. ECL reagents and protein G–Sepharose

were from Amersham Pharmacia Biotech. PKI (TTYAD-

FIASGRTGRRNAIHD), the specific peptide inhibitor of

cAMP-dependent protein kinase, and other peptides were

synthesised by F.B. Caudwell in the MRC Protein Phos-

phorylation Unit.

2.2. CGC culture, drug treatment and cell viability

measurement

Primary cultures of CGC were obtained from 7- to 8-day-

old Wistar rats as previously described [23]. Briefly, CGC

were seeded at a density of 2.5� 105 cells/cm2 and grown

on polylysine-coated plates, in DMEM containing 10%

foetal calf serum, 25 mM KCl, 2 mM glutamine,

0.37� 10 � 3 mg/ml of insulin, 5 mM glucose, 7 mM

p-aminobenzoic acid, 0.1 mg/ml sodium pyruvate, 50 U/ml

penicillin and 50 mg/ml streptomycin in a humidified

atmosphere of 5% CO2 at 37 �C. Cytosine arabinoside

(10 mM) was added after 24 h to arrest the growth of

nonneuronal cells. After 7 days in culture, C2-ceramide and

LiCl were added to the cultures and cell viability was

measured after 24 h by the MTT (3-[4,5-dimethylthiazol-

2-yl]-2,5-diphenyltetrazolium bromide) assay [27]. The per-

centage of survival (Eq. (1)) was calculated from the

absorbance values at 500 nm as follows:

% Survival ¼ O:D: treated� O:D: ceramide

O:D: control� O:D: ceramide� 100 ð1Þ

where O.D. treated is the O.D. at 500 nm of the different

experimental conditions with ceramide plus lithium and

O.D. ceramide corresponds to the cultures treated with C2-

ceramide 50 mM.

2.3. Western blots

The culture medium was removed after 30 min of

cotreatment with drugs and the cells were lysed in ice-cold

lysis buffer (50 mM Tris, pH 7.5, 0.1% Triton X-100, 2 mM

EDTA, 2 mM EGTA, 1 mM microcystin, 1 mM Na3VO4,

50 mM NaF, 10 mM sodium b-glycerophosphate, 5 mM

sodium pyrophosphate, 0.1% (v/v) b-mercaptoethanol and

1 tablet/50 ml of complete protease inhibitor cocktail). After

centrifugation at 15,000� g for 5 min, proteins were

resolved by 10% SDS-polyacrylamide gel electrophoresis

(SDS-PAGE) and transferred to nitrocellulose membranes.

The filters were blocked for 1 h with 5% skimmed milk

in 1� Tris-buffered saline, 0.5% Tween 20, followed by

a 1-h incubation with the primary antibody diluted in

the same blocking solution: 0.5 mg/ml of sheep polyclonal

anti-PKB antisera specific for the C terminus, 0.5 mg/ml of

sheep polyclonal anti-GSK3a antisera, 1 mg/ml of sheep

polyclonal anti-phospho-GSK3a antisera, or 1/1000 of

phospho-PKB (Ser-473) antibody (Cell Signaling Tech-

nology, Beverly, MA). In order to increase the specificity,

the antibodies against phospho-PKB and phospho-GSK3

were incubated with their respective dephosphopeptides

(KHFPQFSYSAS for phospho-PKB and RARTSSFAEPG

for phospho-GSK3). The secondary antibodies were 5000-

fold diluted and incubated for 1 h. Detection was performed

by enhanced chemiluminescence. The signal detected

by chemifluorescence was quantified with Band Leader

2.01 Software.

2.4. PKB activity

After the cells were lysed in ice-cold lysis buffer,

endogenous PKB activity was measured in the supernatants

(15,000� g for 5 min). Cell extracts corresponding to

300 mg of protein were incubated for 1 h at 4 �C with

agitation with 4 mg of PKB antibody (against PH domain)

coupled to 10 ml of protein G–Sepharose. The immune

complexes were washed twice with lysis buffer containing

0.5 M NaCl and 0.1% b-mercaptoethanol, followed twice

with buffer A (50 mM Tris, pH 7.5, 0.1 mM EGTA, 0.1%

b-mercaptoethanol). In vitro kinase assays were performed

for 30 min at 30 �C in 50 ml of reaction volume containing

20 ml of immunoprecipitate in kinase buffer (50 mM

Tris–HCl, pH 7.5, 0.1% b-mercaptoethanol, 0.1 mM

EGTA, 10 mM PKI, 1 mM microcystin, 100 mM 32Pg-ATP

(106 cpm/nmol), 10 mM magnesium acetate) with 50 mM of

the peptide GRPRTSSFAEG as substrate [12].

2.5. Protein phosphatase assay

The protein phosphatase activity of cellular lysates was

determined by measuring the generation of free PO4 from

the phosphopeptide RRA(pT)VA using the molybdate–

malachite green–phosphate complex assay as described by

the manufacturer (Promega, Madison, WI). Cell lysates

A. Mora et al. / Cellular Signalling 14 (2002) 557–562558

were prepared in a low-detergent lysis buffer (0.25% Non-

idet P40, 50 mM Tris, pH 7.4, 150 mM NaCl, 1 mM PMSF,

10 mg/ml aprotinin, 10 mg/ml leupeptin). The phosphatase

assay was performed in a PP2A-specific reaction buffer [7]

(final concentration 50 mM imidazole, pH 7.2, 0.2 mM

EGTA, 0.02% b-mercaptoethanol, 0.1 mg/ml BSA) using

100 mM phosphopeptide substrate and 5 mg of protein from

supernatants (15,000� g for 5 min). After a 15-min incuba-

tion at 30 �C, molybdate dye was added, and free phosphate

was measured by optical density at 630 nm. A standard

curve with free phosphate was used to determine the amount

of free phosphate generated. Phosphatase activity was

defined as number of picomoles of free PO4 generated per

microgram of protein multiplied by time in minutes.

2.6. Other methods

Protein concentration was determined by the method of

Bradford [28] using BSA as standard.

Values shown are the mean ± S.E.M. of at least three

independent experiments made in triplicate. Statistical sig-

nificance of the data was evaluated after the calculation of a

two-tailed P value (unpaired t test) using the GraphPad

Prism 2.01 program (GraphPad Software). P < .05 was set as

the threshold for statistical significance.

3. Results

As shown in Fig. 1, lithium partially prevents the cell

viability decrease induced by C2-ceramide. These results

confirm our previous work [20] where we demonstrated that

C2-ceramide decreased cell viability in a dose-dependent

manner by inducing apoptosis with an EC50 of about

50–60 mM. The neuroprotective action of lithium at these

concentrations has been also previously shown in other

cellular systems and apoptotic stimuli [17–23,26].

In order to study the mechanisms by which ceramide

induces apoptosis in our experimental conditions, we next

studied the effect of ceramide on PKB phosphorylation

level. As shown in Fig. 2A, the incubation of the cells with

50 mM C2-ceramide for 30 min decreased the level of PKB

phosphorylation at Ser-473. In the same set of experiments,

Fig. 2. Lithium blocks the C2-ceramide-induced dephosphorylation and

inactivation of PKB. Cells were cotreated with 5 mM LiCl and 50 mM C2-

ceramide. Cell lysates were obtained after 30 min of each treatment. (A)

Twenty micrograms of protein were resolved in 10% SDS-PAGE,

transferred to nitrocellulose membranes and revealed against total or

phospho-PKB (Ser-473). (B) Three hundred micrograms of protein were

used in the PKB activity measurement. Each value represents the

mean ± S.E.M of three experiments made in triplicate. *P< .05 compared

to untreated (control) cells.

Fig. 1. Partial protection elicited by lithium against C2-ceramide-induced

apoptosis. Cells were cotreated with 50 mM C2-ceramide and different

concentrations of LiCl. Cell viability was measured after 24 h of C2-ceramide

treatment. Each value represents the mean ± S.E.M. of three experiments

made in triplicate. *P < .05 compared to 50 mM C2-ceramide-treated cells.

A. Mora et al. / Cellular Signalling 14 (2002) 557–562 559

lithium prevented PKB dephosphorylation induced by

C2-ceramide. Dephosphorylation of PKB at Ser-473

induced by C2-ceramide might be reflected in PKB activity

because phosphorylation at this residue is required for the

activation of this kinase [12]. As expected, C2-ceramide

inhibits PKB activity (Fig. 2B), keeping a good correlation

between the inhibition rate and the observed decrease in

the phosphorylation state of PKB. In Fig. 2B, it is also

observed that lithium precludes the inhibition of PKB

activity induced by C2-ceramide. This result suggests that

the partial protection of lithium against C2-ceramide-

induced apoptosis could be due to the inhibition of the

serine–threonine phosphatase activity responsible of PKB

dephosphorylation and inactivation.

Since one of the main substrates of PKB is GSK3 [12],

and this kinase has been related to apoptotic induction when

active [13–15], we decided to check the effect of C2-

ceramide on the phosphorylation state of GSK3 at Ser-21,

the residue that induces the inactivation of the kinase when

phosphorylated. Fig. 3 shows that C2-ceramide induces

dephosphorylation of GSK3 and the coincubation with

lithium prevents this dephosphorylation, blocking its proa-

poptotic effect.

These results suggest the activation of a serine–threonine

phosphatase induced by C2-ceramide, which could be

implicated in the dephosphorylation of PKB and GSK3.

As the activation of PP2A by ceramide was demonstrated

[6,7], we decided to check the possible implication of this

protein in this apoptotic model and its relation with the

neuroprotective action of lithium.

As shown in Fig. 4, C2-ceramide induces the activation

of PP2A in a concentration-dependent manner. Moreover,

Fig. 3. Lithium blocks the C2-ceramide-induced dephosphorylation of

GSK3. Cells were cotreated with 5 mM LiCl and 50 mM C2-ceramide. (A)

Cell lysates were obtained after 30 min of each treatment. Twenty

micrograms of proteins were resolved in 10% SDS-PAGE, transferred to

nitrocellulose membranes and revealed against total or phosho-GSK3

(Ser-21). The bands were quantified by densitometry (B). Each value

represents the mean ± S.E.M of three experiments made in tripli-

cate. *P< .05 compared to untreated (control) cells.

Fig. 4. Lithium blocks the C2-ceramide-induced activation of PP2A. Cells

were cotreated with different concentrations of C2-ceramide with or without

5 mM LiCl. Cell lysates were obtained after 30 min of each treatment. Five

micrograms of protein were used in the PP2A activity measurement. Each

value represents the mean ± S.E.M of three experiments made in triplicate.

*P < .05 compared to untreated (control) cells.

A. Mora et al. / Cellular Signalling 14 (2002) 557–562560

the incubation with lithium blocks this activation, suggest-

ing that lithium protects against C2-ceramide by precluding

the activation of a PP2A activity in CGC. This PP2A-like

activity was not inhibited in vitro by lithium (data not

shown) and fully inhibited in vitro in the presence of

0.1 mM microcystin LR (data not shown), a potent inhibitor

of PP2A [29].

4. Discussion

The PKB dephosphorylation by C2-ceramide is in good

agreement with previous studies [10,11] and also suggests

that in our experimental conditions ceramide may induce

apoptosis by activating an unknown serine–threonine phos-

phatase activity [4,5,11]. This inactivation of PKB is one of

the described proapoptotic effects of ceramide, and could

explain the partial protection due to lithium. This effect of

lithium on cell viability is related to the blockade of PKB

dephosphorylation induced by C2-ceramide. This result

could be explained by the activation of PI3K by lithium.

However, this possibility can be excluded because lithium

also protects CGC against apoptosis in the presence of PI3K

inhibitors [23]. Moreover, we have previously shown that

lithium did not directly increase PKB activity [26]; there-

fore, our results suggest that the effect of lithium on PKB

activity may be due to a blockade in the PKB dephospho-

rylation induced by C2-ceramide.

These results are consistent with our previous work

where we demonstrated that lithium also prevented the

dephosphorylation of PKB at Ser-473 induced by potassium

deprivation [26]. Taken together, these results suggest that

both apoptotic stimuli, C2-ceramide and potassium depriva-

tion, share a common mechanism to trigger apoptosis, the

activation of a serine–threonine phosphatase that dephos-

phorylates PKB, and lithium may protect against apoptosis

by inhibiting this phosphatase activity.

It is important to note that the neuroprotective effect of

lithium is due to its blockade of the inactivation of PKB and

also to its inhibition of GSK3 by keeping it phosphorylated.

This model contrasts with its property as in vitro inhibitor of

GSK3 activity. In this regard, it is to be noted that in the

presence of C2-ceramide and lithium, GSK3 is phospho-

rylated at Ser-21, and therefore, it is inhibited. The same

results have been obtained when apoptosis is induced by

potassium deprivation of CGC, where GSK3 is dephos-

phorylated even in the presence of insulin, and lithium

prevents this effect [26].

Although there are experiments that suggest the activa-

tion of PP2A by ceramide [6,7], another discards the

activation of this phosphatase [9]. However, its is clear that

the protein phosphatase participates in the dephosphoryla-

tion of PKB by ceramide. Our results clearly show the

activation of a PP2A-like activity in the cultures stimulated

with C2-ceramide. Since this activation is blocked by the

coincubation with lithium, we can suggest that the neuro-

protective effect of lithium is due to the blockade of the

activation of this phosphatase activity by a mechanism still

unknown but not related to the direct inhibition of PP2A

activity by the ion.

In conclusion, our results demonstrate that C2-ceramide

induces the dephosphorylation of both PKB and GSK3,

probably by the activation of a PP2A-like activity. These

three events are blocked by lithium, suggesting that the

neuroprotective mechanism of the ion is related with the

inhibition of a PP2A-like protein phosphatase activated by

different apoptotic stimuli. This conclusion gives a new

property to lithium that explains its neuroprotective actions

that are not related with the inhibition of GSK3 activity in

vitro [24,25].

Acknowledgments

This work was supported by grants PB98-0992 (from

DGICYT) and IPR00C009 and 00/05 (from Junta de

Extremadura) and by the Medical Research Council

(A.C.). A.M. and G.S. were supported by the Spanish

MEC predoctoral fellowships. We thank J. Leitch for sheep

antibodies and D.R. Alessi, M.J. Lorenzo and R.M. Biondi

for suggestions.

References

[1] Oppenheim RW. Annu Rev Neurosci 1991;14:453–501.

[2] Testi R. Trends Biochem Sci 1996;21:468–71.

[3] Zhang Y, Yao B, Delikat S, Bayoumy S, Lin XH, Basu S, McGinley

M, Chan-Hui PY, Lichenstein H, Kolesnick R. Cell 1997;89:63–72.

[4] Dobrowsky RT, Hannun YA. J Biol Chem 1992;267:5048–51.

[5] Chalfant CE, Kishikawa K, Mumby MC, Kamibayashi C, Bielawska

A, Hannun YA. J Biol Chem 1999;274:20313–7.

[6] Kowluru A, Metz SA. FEBS Lett 1997;418:179–82.

[7] Ruvolo PP, Deng X, Ito T, Carr BK, May WS. J Biol Chem 1999;

274:20296–300.

[8] Millward TA, Zolnierowicz S, Hemmings BA. Trends Biochem Sci

1999;24:186–91.

[9] Zhou H, Summers SA, Birnbaum MJ, Pittman RN. J Biol Chem

1998;273:16568–75.

[10] Schubert KM, ScheidMP, DuronioV. J Biol Chem 2000;275:13330–5.

[11] Salinas M, Lopez-Valdaliso R, Martın D, Alvarez A, Cuadrado A.

Mol Cell Neurosci 2000;15:156–69.

[12] Cross DAE, Alessi DR, Cohen P, Andjelkovich M, Hemmings BA.

Nature 1995;378:785–9.

[13] Pap M, Cooper GM. J Biol Chem 1998;273:19929–32.

[14] Bijur GN, DeSarno P, Jope RS. J Biol Chem 2000;275:7583–90.

[15] Cross DAE, Culbert AA, Chalmers KA, Facci L, Skaper SD, Reith

AD. J Neurochem 2001;77:94–102.

[16] Manji HK, Potter WZ, Lenox RH. Arch Gen Psychiatry 1995;

52:531–43.

[17] D’Mello SR, Anelli R, Calissano P. Exp Cell Res 1994;211:332–8.

[18] Chen RW, Chuang DM. J Biol Chem 1999;274:6039–44.

[19] Chalecka-Franaszek E, Chuang DM. Proc Natl Acad Sci USA

1999;96:8745–50.

[20] Centeno F, Mora A, Fuentes JM, Soler G, Claro E. NeuroReport

1998;9:4199–203.

[21] Munoz-Montano JR, Moreno FJ, Avila J. FEBS Lett 1997;411:183–8.

A. Mora et al. / Cellular Signalling 14 (2002) 557–562 561

[22] Nonaka S, Chuang DM. NeuroReport 1998;9:2081–4.

[23] Mora A, Gonzalez-Polo RA, Fuentes JM, Soler G, Centeno F. Eur J

Biochem 1999;266:886–91.

[24] Klein PS, Melton DA. Proc Natl Acad Sci USA 1996;93:8455–9.

[25] Stambolic V, Ruel L, Woodgett JR. Curr Biol 1996;6:1664–8.

[26] Mora A, Sabio G, Gonzalez-Polo RA, Cuenda A, Alessi DR,

Alonso JC, Fuentes JM, Soler G, Centeno F. J Neurochem 2001;

78:199–206.

[27] Mossman TR. J Immunol Methods 1983;65:55–63.

[28] Bradford MM. Anal Biochem 1976;72:248–54.

[29] Honkanen RE, Zwiller J, Moore RE, Daily SL, Khatra BS, Dukelow

M, Boynton AL. J Biol Chem 1990;265:19401–4.

A. Mora et al. / Cellular Signalling 14 (2002) 557–562562