The transcriptional repressor ICER and cAMP-induced programmed cell death

The transcriptional repressor ICER and cAMP-induced programmed celldeath

Sandrine Ruchaud1, Paule SeiteÂ 1, Nicholas S Foulkes2, Paolo Sassone-Corsi2 and Michel Lanotte1

1INSERM U301, Centre G. Hayem, HoÃpital St Louis, 1 avenue Claude Vellefaux, 75010-Paris, France; 2Institut de GeÂneÂtique etde Biologie MoleÂculaire et Cellulaire, 1, rue Laurent Fries, 67404-Illkirch, France

The cAMP pathway plays a central role in the responseto hormonal signals for cell proliferation, di�erentiationand apoptosis. In IPC-81 leukaemia cells, activation ofthe cAMP pathway by prostaglandin E1 treatment, orother cAMP-elevating agents, induces apoptosis within4 ± 6 h. Inhibition of mRNA or protein synthesis duringthe ®rst 2 h of cAMP induction protects cells fromapoptosis, suggesting a requirement for early geneexpression. cAMP-dependent protein kinase phosphor-ylates a class of nuclear factors and thereby regulatesthe transcription of a speci®c set of genes. Here we showthat CREM (cAMP Responsive Element Modulator)expression is induced rapidly upon prostaglandin E1treatment of IPC-81 cells. The induced transcriptscorrespond to the early product ICER (Inducible cAMPEarly Repressor). ICER expression remains elevateduntil the burst of cell death. Protein synthesis inhibitorswhich prevent cAMP-induced apoptosis also block denovo ICER synthesis. Transfected IPC-81 cell lines,constitutively expressing high level of ICER are resistantto cAMP-induced cell death. In these transfected cells,cAMP fails to upregulate the ICER transcriptsdemonstrating that ICER exerts strongly its repressorfunction on CRE-containing genes. That an earlyexpression of ICER blocks apoptosis, suggests that generepression by endogenous ICER in IPC-81 is unsu�cientor occurs too late to protect cells against death. ICERtransfected cells rescued from cAMP-induced apoptosisare growth arrested. It shows for the ®rst time thatCREM activation directly participates to the decision ofthe cell to die. ICER, by sequentially repressing distinctsets of CRE-containing genes could modulate cell fate.

Keywords: apoptosis; leukaemia; cAMP-signalling;transcription factors; CREM; ICER

Introduction

The cAMP pathway plays a central role in the responseto hormonal signals for cell proliferation, differentia-tion and apoptosis (reviewed in Taylor et al., 1990;McKnight, 1991; Spaulding, 1993; Francis and Corbin,1994; Gjerstsen and Doskeland, 1996). Apoptosis is aphysiological mechanism which can be triggered byexternal signals or programmed by the cell itself(Wyllie et al., 1980; Potten, 1988; Wyllie, 1993;Williams and Smith, 1993; Steller, 1995; White, 1996).Exogenous stimuli which activate speci®c receptors

(reviewed in Nagata and Golstein, 1995), or removal ofgrowth factors, as in the case of hematopoieticprogenitor cells (Williams et al., 1990), all induceapoptosis. Apoptosis can also be arti®cially induced bytreatment with hormones or chemotherapeutic drugs(Fisher, 1994; Thompson, 1995). Signal transductionpathways appear to play a key role in apoptosis. Thesecond messengers Ca2+, cAMP and phorbol esterswhich activate protein kinases such as calmodulinkinase, PKA, PKC and also tyrosine kinases (Pratt andMartin, 1975; Kizaki et al., 1989, 1990; McConkey etal., 1990; Susuki et al., 1990; Lanotte et al., 1991;Dowd and Miesfeld, 1992) have all been implicated inapoptotic processes. The mechanisms whereby signal-ling pathways direct apoptosis appear to be complexand cell type-speci®c. Indeed, cAMP induces apoptosisin thymocytes (Kizaki et al., 1990; McConkey et al.,1992; Susuki et al., 1990), while it stimulatesproliferation of epithelial cells and neuronal celldi�erentiation (Freidman, 1976; Dumont et al., 1989).Importantly, cAMP synergises with glucocorticoidsand PKC signalling in inducing apoptosis (McConkeyet al., 1990; Susuki et al., 1990; Groul and Altschmeid,1993; reviewed in McConkey et al., 1992), while itcooperates with retinoids in the di�erentiation ofteratocarcinoma cells (Hu and Gudas, 1990; Weiler-Guetter et al., 1992) and of various leukaemia cells(Olsson et al., 1982; Gianni et al., 1995; Ruchaud et al.,1994).The diversity of action of cAMP signalling in the

hemopoietic system, particularly the capacity of cAMPto modulate hemopoietic cell development has beenknown for more than a decade. cAMP-elevating agentsexert negative and positive control on proliferation anddi�erentiation of murine and human myeloid, lym-phoid and erythroid progenitor cells (Rossi et al., 1980;Taetle and Koessler, 1980; Lanotte et al., 1986). cAMPinduces monocytic di�erentiation of M1 mousemyeloid leukaemia (Homma et al., 1978) and of asubclone of HL60 cells (Tortora et al., 1988). Itpotentiates granulocytic di�erentiation of retinoid-induced maturation of human myeloid leukaemia cells(Olsson et al., 1982; Ruchaud et al., 1994). cAMPsignalling triggers apoptosis in normal and leukaemicthymocytes and in myeloid leukaemia cells. The IPC-81cell line (Lanotte et al., 1986), derived from a ratmyeloid leukaemia so far represents one of the best-de®ned cell systems to study cAMP-induced apoptosis.We have previously demonstrated that treatment withagents elevating intracellular cAMP levels (choleratoxin, prostaglandins (PGEs), IBMX and cAMP-analogues) rapidly induces IPC-81 cell death (Lanotteet al., 1991). More than 90% of the cells showmorphological changes and internucleosomal DNA

Correspondence: M LanotteReceived 27 December 1996; revised 2 May 1997; accepted 2 May1997

Oncogene (1997) 15, 827 ± 836 1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00

fragmentation typical of apoptosis within 6 h ofinduction. Furthermore, speci®c 28S ribosomal RNAcleavage has recently been identi®ed as a typicalcharacteristic of apoptosis in these cells (Houge et al.,1993, 1995). It has been established that cAMPinduced apoptosis of IPC-81 cells is mediated byisoform I of the PKA holoenzyme. Indeed, a subclonederived from IPC-81 cells (IPC-RID336) carrying a pointmutation in the RIa regulatory subunit at the cAMPbinding site is resistant to cAMP induced apoptosis(Duprez et al., 1993). Intracytoplasmic microinjectionof puri®ed PKA catalytic subunit commits the cells todeath (Vintermyr et al., 1993). Furthermore, cell deathis delayed by treating IPC-81 cells with actinomycin Dor cycloheximide. This treatment suppresses morpho-logical changes typical of apoptosis, as well as DNAfragmentation and 28S rRNA cleavage (Lanotte et al.,1991; Houge et al., 1993, 1995). This suggest a directinvolvement of PKA-directed gene expression. Here wehave analysed whether a speci®c nuclear response tocAMP exists during apoptosis in hemopoietic cells.cAMP-activated PKA phosphorylates speci®c nucle-

ar transcription factors which include CREB, CREMand ATF-1 (Sassone-Corsi, 1994). These factors bindas dimers to the speci®c consensus sequence TGACGT-CA, CRE (cAMP responsive element) (Montminy etal., 1986). CREM appears to play a privileged roleamongst CRE regulatory factors in a variety ofdi�erent physiological systems. By a process ofextensive alternative splicing, the CREM gene gener-ates a family of both activator and repressor isoforms(Delmas et al., 1992; Foulkes et al., 1991, 1992a,b,Laiode et al., 1993). Using an alternative intronicpromoter the CREM gene also generates the powerfulrepressor, ICER (Molina et al., 1993; Stehle et al.,1993). The ICER speci®c promoter (P2) containstandemly repeated CRE-like elements and is rapidlyand strongly induced upon activation of the cAMPpathway. The kinetics of cAMP induction place ICERas a key player in the early cellular response tohormonal stimuli. It has been shown that ICER ispredominantly expressed in neuroendocrine tissues(Molina et al., 1993; Stehle et al., 1993; Monaco etal., 1995; Lalli and Sassone-Corsi, 1995).Here we show that ICER is rapidly induced by

prostaglandin E1 and cAMP treatment of the IPC-81leukaemic cell line and that CREM activation directlyparticipates in the decision of the cell to di�erentiate ordie. Transfection experiments show that ICER acts asa repressor of cAMP-induced apoptosis.

Results and Discussion

Prostaglandin treatment induces apoptosis of IPC-81cells

We have previously shown that agonistic stimulation ofadenylate cyclase coupled membrane receptors ortreatment with cholera toxin and cAMP analoguesinduce IPC-81 cells to undergo apoptosis (Lanotte etal., 1991).Prostaglandins have a crucial role in the physiolo-

gical regulation of lymphoid and myeloid celldevelopment (Taetele and Koessler, 1980; Hayari etal., 1985; Monaco et al., 1995). Prostaglandin E1

(PGE1) treatment of IPC-81 cells causes a dramaticmorphological change which is reminiscent of cAMP-induced cell death (Figure 1a,b); and in Lanotte et al.,1991). Cells treated with PGE1 show high sensitivitywith an IC50 of 0.2 mM (Figure 1c). PGE1 also causesDNA fragmentation (Figure 1d lower panel) typical ofapoptosis as described in other cells (Duke et al., 1983;McConkey et al., 1990).Prostaglandin E have been shown to activate

signalling pathways coupled to adenylate cyclase(Hittleman and Butcher, 1973; Sonnenburg andSmith, 1988; Mori et al., 1989) and phosphorylationby PKA. Therefore, we wanted to know whether PGE1also mediated a cAMP-dependent nuclear response.We tested whether CREM gene expression may bemodulated in IPC-81 cells by PGE1. In Figure 1d weshow that CREM mRNA levels are dramaticallyinduced, reaching a peak by 2 ± 3 h after PGE1

treatment and decreasing by 6 ± 8 h, accompanyingapoptosis, as shown by DNA fragmentation.

CREM induction in IPC-81 cells upon apoptosis

To mimic the transduction of an external signal wherecAMP is the second messenger, we used analogs of thismolecule which directly stimulate PKA. Here we haveused 8CPT-cAMP because of its lipophilicity, a�nityfor the R1 subunit of PKA and stability (Braumannand Jarstor�, 1995; Ruchaud et al., 1995). In addition,8CPT-cAMP was shown to e�ciently induce apoptosisin IPC-81 cells (Ruchaud et al., 1995).The level of Cyclic-AMP Responsive Element

Binding (CREB) mRNA is a�ected by cAMPtreatment since a marked decrease occurs at 90 minafter treatment (Figure 2). The decrease in CREBmRNA expression correlates with the onset ofapoptosis indicated by the 28S rRNA cleavage(Houge et al., 1993, 1995) (see lower panel, lanes 13and 14), and also follows a marked decrease in b-actinRNA which accompanies apoptosis (lane 14). Incontrast, CREM mRNA expression (Figure 2) isundetectable in untreated cells (lane 7), but is thenstrongly induced as early as 30 min after cAMP-treatment, to then peak after 2 ± 3 h (lanes 12 and13). The strong CREM mRNA upregulation persisteduntil 4 h after treatment. Importantly, the induction ofCREM occurred in a dose-dependent manner (lanes1 ± 6). Similar to cell viability in culture which shows a-all or non-response (Lanotte et al., 1991, Ruchaud etal., 1995), CREM mRNA levels shifted from low tomaximum levels when cAMP molarity increased from24 mM to 50 mM. Thus, CREM inducibility shows astrict dose-dependent response. This feature is strik-ingly similar to the dose-dependent, cAMP-inducedapoptosis (Lanotte et al., 1991).cAMP treatment results in the induction of various

CREM transcripts, 2.5 kb, 1.8 kb and 1.5 kb (Figure 2).The CREM gene generates a family of activators andrepressors by alternative splicing, translation initiationand use of an alternative intronic promoter (Delmas etal., 1992; Foulkes et al., 1992; Laoide et al., 1993;Molina et al., 1993). In order to determine the type ofCREM transcripts which are induced in IPC-81 cells weused RNAse protection analysis (Figure 3a, b). ThepCREM p75 probe visualises and distinguishes betweenICER and the other CREM transcripts (Stehle et al.,

Transcription repressor ICER and cAMP-induced apoptosisS Ruchaud et al

828

1993). RNA from cAMP-treated IPC-81 cells protectstwo fragments (122 nt and 95 nt) which correspond toICER isoforms generated from the P2 promoter(Molina et al., 1993) using the S1 and S2 transcriptionstart sites, respectively. The protected fragmentcorresponding to the CREMa,b,g or t transcripts(40 nt) derived from the P1 promoter reveals a verylow and non-induced expression (not seen in Figure 3b;detectable only after a prolonged exposure).Western blot analysis of IPC-81 cellular extracts

with an anti-CREM antibody shows (Figure 4) thatICER proteins are detected as trace amounts inuntreated cultures (see in Figure 4 and legend; panela overexposed to show ICER expression). cAMPtreatment resulted in the induction of ICER proteincorresponding to a doublet migrating around 16 ±18 kDa and a major 13 kDa band. When the cells aretreated with cAMP for 2 h the proteins are upregu-lated, then strongly expressed after 3 h (Figure 4a). Allthree ICER bands are upregulated by cAMP in a dose-dependent manner. These protein bands show a similarincrease, irrespective of their distinct initial levels.ICER proteins persisted at very high levels untilapoptosis was observed (4 ± 6 h). The kinetics of

induction of ICER proteins are consistent with thoseof the ICER transcript (Figure 2) and with the kineticsof induction of apoptosis by cAMP (not shown). SinceICER transcripts and ICER proteins are the onlyCREM gene products found upregulated by cAMP inIPC cells, it will then be referred to ICER only, in thefollowing sections of this paper.

cAMP triggers both induction and repression ofapoptosis

The block of protein synthesis by cycloheximide (CHX)during the early phase of cAMP-induction confersresistance to cAMP-induced cell death, as shown byinhibition of DNA fragmentation (Lanotte et al., 1991),28S rRNA cleavage (Houge et al., 1993, 1995; and inthis paper, Figure 5a (lower panel), compare lanes 4 ± 6(without CHX) and lanes 9 ± 11 (with CHX). In theseexperiments CHX treatment alone is not inducingrRNA cleavage (Figure 5a (lower panel) lanes 12 ± 15)and has little e�ect on IPC-81 cell viability (Figure 5b,14% of apoptotic cells after 8 h of CHX treatment).Importantly, cotreatment of CHX and cAMP does notalter the timing of ICER mRNA induction (Figure 5a;

a

b

c

C 1 2 3 4 6 8

INCUBATIONS (h)

28S

18S

β actin

CREM

622 bp — 404 bp — 242 bp —

1 2 3 4 5 6 7

d

Figure 1 Apoptosis and CREM induction in PGE1 treated IPC-81 cells. IPC-81 cells were treated with PGE1 (25 mM) for 8 h;May ±Grunwald staining, control culture (a) and treated cells (b). Note the typical morphology of apoptotic cells, chromatincondensation, nuclear fragmentation, plasma membrane blebbing and formation of apoptotic bodies (b). Cell viability was measuredby MTT assay after a 48 h incubation of cultures with PGE1. Dose-response to PGE1 (c). Northern blot analysis of CREMtranscripts upon PGE1 treatment: kinetics of CREM expression (d, upper panel). RNA loading was calibrated by staining the geland measuring the 18S ribosomal band. Notice the coordinated decrease of CREM and b-actin mRNA when apoptosis peaked (6 ±8 h). These decreases correspond to the burst of apoptosis in culture (b) and the peak of DNA fragmentation evaluated byelectrophoretic migration on agarose gels (d, lower panel)


829

lanes 2 and 7). A signi®cant increase in the levels of theICER transcript (Figure 5a; compare lane 4 to lane 9,then lane 6 to lane 11) is obtained by CHX treatment. Itcon®rms that CHX rescues cells from cAMP-inducedcell death, but it also shows that ICER mRNA remainshigh in the rescued cells. To precisely determine whenCHX has to be present in culture to rescue cells fromapoptosis, CHX addition was progressively delayedafter cAMP-triggering, then cell viability measured after8 h of cAMP-treatment. Figure 5c shows that apoptotic

200

100

50

25

12.5

6.

2

C

15

30

60

90

120

180

240

7Kb —

2.5Kb —

1.8Kb — 1.5Kb —

28S

18S

28S

18S

)))

CREB

CREM

β actin

CONCENTRATIONS (µM)

INCUBATIONS (mn)

1 2 3 4 5 6 7 8 9 10 11 12 13 14

Figure 2 Expresion of CREB and CREM mRNA during cAMP-induced cell death. RNA loading was con®rmed by staining andmeasuring the 18S ribosomal bands, since both b-actin mRNAand 28S rRNA varied during apoptosis. Lane 7, untreated controlculture; lanes 1 ± 6, dose-response to 8-CPT-cAMP evaluated after2 h. Lanes 8 ± 14, kinetics of CREM induction (8-CPT-cAMP;200 mM). CREB, CREM and b-actin mRNAs were analysedsuccessively by hybridisation with the respective labelled probeson the same membrane. On this membrane also the 28S rRNAethidium bromide ¯uorescence featured the rRNA cleavage as ainternal marker for apoptosis. Notice again the coordinated latedecrease in CREM, CREB and b-actin RNAs when apoptosispeaked (lane 14)

C .5 1 2 3 4 5 6 7

INCUBATIONS (h)

— 145

— 115

— 67

— 34

ICER (S1)

ICER (S2)

1 2 3 4 5 6 7 8 9 10

b

a

122

95

40

Figure 3 The induced CREM isoform corresponds to ICER.CREM transcripts induced by 8-CPT-cAMP (100 mM) duringapoptosis were analysed by RNAse protection. The bandscorrespond to various CREM isoforms: ICER-S1 (122 nt),ICER-S2 (95 nt), CREMa,b,g,t (40 nt). The main inducibletranscript corresponding to ICER-S1 is already maximallyinduced after 2 h. A weak 40 nt signal corresponding to CREMtranscripts generated from the P1 promoter (not detectable in this®gure) can also be seen when the ®lm is overexposed

C 2 3 4 5 6

18 kDa

13 kDa

C 50

75

100

200

CONCENTRATIONS (µM)

INCUBATIONS (Hours)

Figure 4 Induction of ICER proteins during cAMP-induced apoptosis. Total protein extracts from IPC-81 cells triggered toapoptosis by cAMP were analysed by 15% SDS±PAGE and ICER proteins were detected by a speci®c rabbit antibody, asdescribed in Methods. (a) Kinetics of ICER induction (8-CPT-cAMP, 100 mM); (b) Dose-response to cAMP measured after 3 h ofincubation with 8-CPT-cAMP


830

cell counts remain close to 20% when CHX was addedduring the ®rst 2 h of induction (C2 give the percent ofcell death with cells treated with cAMP alone). Whenadded later, CHX fails to protect cells from apoptosis.Interestingly, CHX delivered after the ®rst 3 h ofcAMP-treatment synergised with cAMP in inducingcell death (see in Figure 5c, after 3 h). The biphasicnature of the CHX-response curve (Figure 5c) isconsistent with the notion that the cells synthesise anearly e�ector to trigger cell death. The synergistic e�ectof CHX allows us to hypothetise that a distinctcomponent may be produced to counteract cAMP-

induced cell death, and that the synthesis of thiscomponent is also inhibited by CHX treatment (after3 h). Accordingly, in IPC-81 cells the e�ector ofapoptosis is generated before the gene product whichcounteracts cell death, and the balance between theexpression of the inducer and the repressor maydetermine the cell fate.

ICER expression and cell fate

The IPC-RID336 mutant cells carry a mutation in theRIa cAMP binding site of PKA (see in Duprez et al.,

C 1 2 4 6 8 1 2 4 6 8 2 4 6 8

8CPT-cAMP 8CPT-cAMP + CHX

CHX

18S

28S

18S

ICER

β actin

INCUBATION (h)

a

1 2 3 4 5 6 7 8 9 10 11 12 13 1415

b c

Figure 5 E�ects of CHX treatments on cell viability, rRNA cleavage and ICER induction. (a) E�ect of CHX and cAMP on ICERinduction and 28S rRNA cleavage. Lane 1, untreated culture; lanes 2 ± 6, 8-CPT-cAMP treatment (100 mM); lanes 7 ± 11, combinedtreatment of 8-CPT-cAMP (100 mM) and CHX (28 mg/ml); lanes 12 ± 15, CHX treatments (28 mg/ml). Notice that CHX alone is notinducing apoptosis at this concentration. (b) E�ect of CHX and cAMP on IPC-81 cell viability. Cells (105 cell/ml) were cultured in24 well culture plates for 0 to 8 h with either 8-CPT-cAMP alone, CHX alone, or with 8CPT-cAMP plus CHX (& 8CPT-cAMPalone (100 mM); * CHX alone (28 mg/ml); & 8CPT-cAMP (100 mM) plus CHX (28 mg/ml)). Apoptosis was evaluated bymorphological analysis of at least 500 cells on MGG stained slides. (c) E�ect of delayed treatments with CHX on cAMP-inducedapoptosis. CHX was added in cultures at di�erent times after the induction of apoptosis by 8CPT-cAMP (100 mM) and cell viabilitywas measured, 8 h after cAMP induction, in all cultures (*). Cell viability was also measured in cultures when CHX alone wasadded at the same times (~). C1 shows the cell viability in untreated cultures; C2 shows the cell viability in cultures treated withcAMP alone. Values below the C2 line indicate the protective e�ects on CHX when delivered during the early phase of apoptosisinduction; values above C2 show the synergistic e�ect of cAMP and CHX on apoptosis, when CHX is delivered only during the latephase of apoptosis


831

1993). In these cells, the CREM gene is only weaklyupregulated at the highest concentrations of 8CPT-cAMP (1000 mM), as shown by ICER mRNA levels(Figure 6a, lanes 10 ± 12. These cells also show aremarkable resistance to cAMP-induced cell death(Figure 6b, lanes 9 ± 15), as demonstrated by a lackof DNA fragmentation. This strengthens the correla-tion between a strong ICER upregulation and celldeath. However, the lack of CREM gene inducibility inthe mutant IPC-81 cells resistant to apoptosis does notindicate what role ICER could play in the control ofcell death.As shown above, ICER is consistently strongly

upregulated in IPC-81 cells at cAMP concentrationswhich trigger cell death. We wanted to analyse theresponse of cells to the lowest cAMP concentrationable to induce a signi®cant ICER upregulation. Whencells were treated with 20 ± 30 mM 8-CPT-cAMP theyresponded to cAMP by a slight ICER upregulation(for instance, such an experimental condition isfeatured in Figure 2, lane 4) and most cells showedonly little morphological alteration after 8 h, but asfew as 20% of the cells were found viable after 48 h.These surviving cells are all growth arrested (no mitosiswere found, not shown) and clearly show morpholo-gical signs of maturation (Figure 7).We have shown that a gradation in the strength of

cAMP signal which can be obtained with either theRIa mutation in IPC-RID336 (Duprez et al., 1993) orwith a gradual increase in the concentration of a stablecAMP-analogue in the culture can change the fate ofIPC-81 cells. When ICER is not induced, no detectablecell response is observed. When a low expression ofICER is observed, cells either undergo apoptosis (after8 h) or embark in cell di�erentiation (48 h). When apowerful signal is delivered (PGE1 25 mM) or 8-CPT-cAMP (200 mM)) it causes a strong and sustainedICER expression, accompanied in a few hours by

massive death of immature cells, with all the features ofprogrammed cell death.Programmed cell death is classically divided into

distinct phases, the transduction of a death signal byintracellular signalling pathways (signalling), thecascade of (positive and negative) regulation whichresults in cell death triggering, then ®nally anirreversible process of cell destruction (execution) (seein recent reviews by Ho�man and Lieberman, 1994;Steller, 1995; Martin and Green, 1995; White, 1996;Rowan and Fischer, 1997). If in IPC-81 cells ICERfunctions as a strong repressor of the cAMP signal,one would expect that it could repress cAMP-induced

C C500

250

125

62

31

16

8 1000

50

0 25

0 12

5 62

31

4.4Kb — 2.3Kb — 2.0Kb —

.56Kb —

18S

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

IPC-81 IPC-RI D336

CONCENTRATION (µM)

a

b

ICER

β actin

Figure 6 Lack of ICER induction in the apoptosis-resistant IPC-R1D336 cells. IPC-81 cells (wild-type) and the apoptosis-resistantIPC-R1D336 cells were cultured in the same conditions withincreasing concentrations of 8-CPT-cAMP for 3 h (a). While IPC-81 cells show the previously observed induction (lanes 1 ± 8), theapoptosis-resistant cells IPC-RID336 show little or not ICERinducibility (lanes 9 ± 15). Apoptosis in the same cultures wasevaluated by DNA fragmentation analysis (b)

a

b

c

Figure 7 Granulocytic maturation of apoptosis-resistant IPC-81cells by treatment with low cAMP concentrations. (a) Controlculture; (b) morphology of IPC-81 cell treated by 8CPT-cAMP(100 mM) for 5 h. (c) morphology of IPC-81 cells having resistedto apoptosis by a 25 mM 8-CPT-cAMP treatment for 48 h. Noticethat this treatment induced a low ICER upregulation in IPC-81cells at 3 ± 5 h, while at the same time, little or no DNAfragmentation is observed


832

cell death. To be e�cient the transcriptional repressionof the death signal by ICER has to occur before self-destruction had started. An indication of the action ofICER as a repressor is shown by its e�ects on its owntranscription. As shown, ICER downregulation(repression) does not occur before 4 ± 6 h, whilecAMP-induced cell death becomes rapidly irreversiblebecause of self destruction has started earlier (3 h). Asa matter of fact, our results support the hypothesis thatICER neither represses its own expression, nor the celldeath programme, before death signalling has pro-gressed beyond transcriptional regulation and apopto-sis has been triggered.

A constitutive expression of ICER protects IPC-81 cellsagainst cAMP-induced programmed cell death

IPC-81 cells were transfected by electroporation of theMSCVpkg expression vector constructs containing theICERIIg coding sequence (ICER-S) under the control

of the MSCV promoter. As experimental controls, asimilar expression vector with a ICER anti-sense (AS)insert or an èmpty vector' were also used to transfectcells. Several cell lines were selected for puromycinresistance and tested for exogenous ICER-S and ASmRNA expression. Èmpty vector' transfectants(puromycin-resistant controls) had unchanged ICERexpression and responses to cAMP compared tonontransfected control cells. In ICER-AS transfec-tants constitutively expressing antisense mRNA, asmeasured by Northern blot analysis (not shown), therewas no modi®cation in the cAMP-induced expressionof ICER protein suggesting that in these cells ICER-AS mRNA did not interfere speci®cally with ICERtranslation. Furthermore, these transfectants were notaltered in their apoptotic response to cAMP (notshown). ICER-S clones show a strong constitutiveexpression of a 4.5 Kb ICER mRNA transcriptderived from the MSCV construct (Figure 8a). Inthese ICER-S transfected cells the endogenous 2.5 kb,1.8 kb and 1.5 kb CREM transcripts are no longerinducible by cAMP (Figure 8a, compare lanes 4 and 5and lanes 11 and 12) while CREM induction in theèmpty vector' transfectants remains similar to thatobserved in IPC-81 cells (Figure 2). In addition theclones show constitutive expression of a 13 kDa ICERprotein by Western blot analysis using anti-CREMantibody (Figure 8b). The 18 kDa ICER doubletremained at the basal expression level in ICER-Stransfected cells, while it is upregulated in thepuromycin-resistant control cells con®rming thatexogenous ICER repressed both the expression ofendogenous ICER mRNA transcripts and protein.Morphological analysis and viability assay indicatethat ICER-S transfected cells are resistant to cAMP-induced apoptosis. However, a tenfold increase in thecAMP concentration can still trigger cell death: theIC50 (35 mM for the puromycin resistant controls) isshifted to 400 mM for ICER-S transfectants (Figure8c). cAMP (200 mM) induces no signi®cant changes inthe cell morphology of the transfected cells. Althoughgrowth arrest is observed, no granulocytic maturationcomparable to that observed at low cAMP concentra-tions occurred. This demonstrates that ICER func-tions as a repressor of cAMP-induced programmedcell death.

Conclusion

Here we document for the ®rst time cAMP inducibilityof the CREM gene in a myeloid leukaemia cell line.The induced CREM isoform corresponds to thepowerful transcriptional repressor ICER, indicatingthat down-regulation of the CRE-containing genes islikely to follow this induction.A strong ICER induction accompanies apoptosis

triggered by treatment with the prostaglandin E1 orby cAMP analogs. However, at low cAMP concentra-tions ICER is only slightly expressed and a variablenumber of IPC-81 cells, not triggered to apoptosis,embark on granulocytic cell maturation. That ICERfailed to block cAMP-induced apoptosis could be dueto the fact that CRE-containing immediate early geneswhich trigger apoptosis are not repressed by ICERbecause its expression is too low or occurs too late.

C1 1 2 3 4 5 6 C2 1 2 3 4 5 6INCUBATIONS (hrs)

IPC-81 E.V. IPC-81 ICER

28S —

18S —

— Exogenous ICER(4.5kb) — — —

— β actin

1 2 3 4 5 6 7 8 9 10 11 12 13 14

C1 2 4 6 C2 2 4 6

INCUBATIONS (hrs)

IPC-81 E.V. IPC-81 ICER

30 kDa —

21 kDa — 14 kDa —

18 kDa

13 kDa

Endogenous ICER

a

b

c

Figure 8 Biological response to cAMP of ICER-transfected IPC-81 cells. (a) Transfected cells (IPC-81E.V.: IPC-81 Empty Vector;IPC-81ICER: IPC-81 ICER sense) were analysed comparatively onNorthern-blot for ICER mRNA expression upon cAMPtreatment. RNA loading was calibrated by staining the gel andmeasuring the 18S ribosomal RNA band. (b) Total proteinextracts from the two IPC-81 transfected cell lines triggered bycAMP were analysed by 15% SDS±PAGE and ICER proteinswere detected by a speci®c rabbit antibody (see in Methods). (c)Cell viability was measured by WST-1 assay after a 24 hincubation of IPC-81E.V. (*) and IPC-81ICER (&) cell cultureswith 8-CPT-cAMP (8-CPT-cAMP, 100 mM)


833

This hypothesis is supported by the action of CHXand the features of ICER transfectants expressingconstitutively ICER. A strong expression of ICER inthese transfected cells, before cAMP-triggering, canblock the upregulation of the endogenous CREMtranscripts and also block cAMP-induced apoptosis.The e�ects of ICER on cell fate remain di�cult toelucidate when expressed at low level, both using theinducible IPC-81 cells, or the transfected clonesconstitutively expressing ICER. We cannot excludethat ICER acts as an inducer of cell death (repressionof survival (anti-apoptotic) genes), and as a repressorof apoptosis (repressor of genes essential in theapoptotic machinery), depending on whether it isexpressed at low or high levels. For this reason, thedevelopment of an inducible system, where ICERexpression is precisely tuned, is now required.Altogether our data suggest that the fate of the cells± granulocytic maturation and apoptosis ± isdetermined by both the intensity of the cAMP-signaland the capacity of these cells to rapidly counteractthis signal by inducing ICER.These ®ndings represent an important step in the

understanding of the physiological processes linked tothe regulation of the CREM gene. Namely, they extendthe property of CREM inducibility to the non-neuroendocrine cells. Also, it is the ®rst time thatCREM activation is linked to the decision of a cell todi�erentiate or to die. Thus, CREM participates in thecomplex regulatory mechanisms by which signallingpathways modulate cell fate.Signal-induced cell death is modulated essentially at

two distinct levels, by acting on the signal transductionmachinery ± on the expression or the function of itse�ector components ± or by acting on the triggering ofapoptosis which depends upon the action of positiveand negative regulators (Ho�man and Lieberman,1994; Steller, 1995). The action of CHX on cAMP-induced cell death in IPC-81 cells served to preciselydemarcate a narrow window of time during which thetranscriptional machinery is required to triggerapoptosis. An interesting feature of CHX action inthese cells is an early blockade of cell death, then a latesynergy between CHX and cAMP in inducing celldeath, suggesting that cAMP induces sequentially theinducer of apoptosis and a repressor. The transfectionexperiments have permitted to identify ICER as arepressor of cAMP-induced cell death. It remains toelucidate whether ICER acts only by repressing thecAMP-signalling or also by repressing some CRE-containing genes playing important roles as inducers ofcell death. Identi®cation of the gene regulatory targetsof ICER in hemopoietic cells should thereby shed lighton the basic molecular mechanisms underlying cell fatedecision.

Materials and methods

Cell culture and cell viability assay

The rat promyelocytic leukaemia cell line IPC-81 wascultured as previously described (Lanotte et al., 1991). TheIPC-RID336 cAMP-resistant clone was isolated as describedby Duprez et al. (1993). Cell morphology was analysedusing the May ±GruÈ nwald-Giemsa staining. Cell viabilitywas assessed by morphological integrity of cells under

phase-contrast microscopy and by the dimethylthiazoldi-phenyl tetrazolium bromide colorimetric assay (MTT) formitochondrial dehydrogenase enzymatic activity (Mos-mann, 1983) or WST-1 assay (Boehringer Mannheim).

DNA fragmentation

Cells in exponential phase of growth were pelleted andimmediately disrupted in lysis bu�er (10 mM Tris-HCl pH8, 100 mM EDTA, 10 mM EGTA, 0.5% SDS), according toa protocol adapted from (Williams, 1987). DNAse-freeRNAse (Sigma) was added to 20 mg/ml of lysate, whichwas incubated for 2 h at 378C. The cell lysates were thenincubated at 568C with proteinase K (Sigma) at 100 mg/mlfor 1 h. The DNA was extracted with phenol, pelleted with2 vol. ethanol and 0.1 vol. 10 M ammonium acetate,dissolved in TE-bu�er (10 mM Tris-HCl pH 8, 1 mM

EDTA) and separated on a 1.5% agarose gel (FMCbioproducts, USA). DNA fragments were visualised ongels after ethidium bromide staining.

Isolation of RNA and Northern blot analysis

Total RNA was isolated according to the proceduredescribed by Chomczynsky and Sacchi (1987). 20 mg oftotal RNA were electrophoresed in a 1.1% agarose/10%formaldehyde gel and then blotted on a nylon membrane.28S rRNA cleavage was shown by ethidium bromidestaining of total RNA as previously described (Houge etal., 1995).

RNAse protection analysis

RNA probes, labelled with [a-32P]UTP, were preparedusing reagents from a riboprobe kit (Promega). Theseprobes were used for RNAse protection studies, followinga standard protocol previously described (Ausubel et al.,1987).

Western blot analysis of protein expression

Cultured cells were washed in PBS and pelleted bycentrifugation at 400 g for 5 min. Pellets of 56105 cellswere immediately lysed by adding 100 ml of a boilingLaemmli solution containing b-mercaptoethanol anddisrupted with a pestle. Samples were then boiled for5 min and insoluble material removed by centrifugation at13 000 g for 5 min. Proteins were quanti®ed by a Coomasiecolorimetric titration. Protein extracts (30 mg) were loadedon 15% SDS-polyacrylamide gel, electrophoresed, andblotted onto nitrocellulose membranes (Schleicher &Schuell, Germany). After transfer to the nitrocellulosemembrane, proteins were visualised with Ponceau S(Sigma) to con®rm equal loading of protein. Membraneswere blocked with 5% unfatted milk in PBS pH 7.6,137 mM NaCl, then incubated with a speci®c rabbitpolyclonal antiserum (dil: 1/100) raised against recombi-nant CREMt protein (Delmas et al., 1993) in PBS/3% milkfor 18 h at 48C. Membranes were incubated with horse-radish peroxidase-linked protein A (Amersham) for 30 minat 258C. Each of these steps were followed by three washesfor 10 min in PBS/3% milk. Labelling was performed asdescribed in the ECL protocol (Amersham).

Plasmids, cell transfection, selection of ICER-transfected IPC-81 cells

The MSCVpkgICERIIgS and AS vectors were obtained byinserting the complete ICER IIg rat cDNA sequence in senseand antisense position respectively into the EcoRI restric-tion site of a MSCVpkg vector containing a puromycinresistance gene (Hawley et al., 1994; generously given by DrO Bruland, Bergen, NO). The constructs, MSCVpkgICER-IIgS and AS and MSCVpkg as control (each 2 mg) were


834

linearized with Sca I and used to electroporate 26106 IPC-81 cells (25 mF; 600 V on a BIORAD gene pulser system).Selection of stable transfected clones was performed bymaintaining the culture at a 0.8 mg/ml puromycin concen-tration during six weeks. After ampli®cation of survivingcells, proteins and RNA were extracted and analysed onNorthern and Western blots.

AcknowledgementsWe are grateful to Drs J Lillehaug, O Bruland (UnivBergen, NO) for providing us expression vectors and G

Houge for his advice in cell electroporation. Authors alsowish to thank Drs SO Doskeland, O Vintermyr, OTBrustugun, E Lalli and M Lames for discussions. Workin PS-C laboratory is supported by grant from INSERM,CNRS, Fondation pour la Recherche Medicale, Associa-tion pour la Recherche Medicale and RhoÃ ne-PoulencRorer. Work in ML laboratory was supported byINSERM, Fondation contre la LeuceÂ mie (Fondation deFrance), Association pour la Recherche contre le Cancer(ARC, France) and Ligue Nationale contre le cancer.Supported also by EEC grants (Biomed I; Human Capital& Mobility Program). SR is recipient of a Ph.D. fellowshipfrom Ligue Nationale contre le cancer.

References

Ausubel FM, Brent R, Kingston RE, Moore DD, SeidmanJG, Smith JA and Struhl K. (1987). Current protocols inMolecular Biology (N.Y: John Wiley & Sons).

Braumann TJ and Jastor� B. (1995). J. Chromatog., 350,105 ± 118.

Chomczynski P and Sacchi N. (1987). Anal. Biochem., 162,156 ± 159.

Delmas V, Laoide BM, Masquilier D, de Groot RP, FoulkesNS and Sassone-Corsi P. (1992). Proc. Natl. Acad. Sci.USA, 89, 4226 ± 4230.

Delmas V, van der Hoorn F, MellstroÈ m B, JeÂ gou B andSassone-Corsi P. (1993).Mol. Endocrinol., 10, 1502 ± 1514.

Dowd DR and Miesfeld RL. (1992). Mol. Cell. Biol., 12,3600 ± 3608.

Duke RC, Chervenak R and Cohen JJ. (1983). Proc. Natl.Acad. Sci. USA, 80, 6361 ± 6365.

Dumont JE, Jauniaux JC and Roger PP. (1989). TrendsBiochem., 14, 67 ± 71.

Duprez E, Gjertsen BT, Bernard O, Lanotte M andDoeskeland SO. (1993). J. Biol. Chem., 268, 8332 ± 8340.

Fisher DE. (1994). Cell, 78, 539 ± 542.Foulkes NS, Borrelli E and Sassone-Corsi P. (1991). Cell, 64,739 ± 749.

Foulkes NS, MellstroÈ m B, Benusiglio E and Sassone-Corsi P.(1992a). Nature, 335, 80 ± 84.

Foulkes NS and Sassone-Corsi P. (1992b). Cell, 68, 411 ±414.

Francis SH and Corbin JD. (1994). Annu. Rev. Physiol., 56,237 ± 272.

Friedman DL. (1976). Physiol. Rev., 56, 652 ± 708.Gianni M, Terao M, Norio P, Barbui T, Rambaldi A andGarattini E. (1995). Blood, 85, 3619 ± 3625.

Gjertsen BT and Doskeland SO. (1996). Biochim. Biophys.Acta, 1269, 187 ± 199.

Green DR and Martin SJ. (1996). Cell Death Di�., 3, 339 ±340.

Groul DJ and Altschmied J. (1993). Mol. Endocrinol., 7,104 ± 113.

Hawley RG, Lieu PH, Fong AZ and Hawley TS. (1994).Gene Ther., 1, 136 ± 138.

Hayari Y, Kukulansky T and Globerson A. (1985). Eur. J.Immunol., 15, 43 ± 48.

Hittelman KJ and Butcher RW. (1973). cAMP and themechanism of action of the prostaglandins. In TheProstaglandins, MF Cuthberg. (ed.). Lippincott, Philadel-phia. p. 151.

Ho�man B and Liebermann DA. (1994). Oncogene, 9, 1807 ±1812.

Homna V, Kasukabe T and Hozumi M. (1978). Biochem.Biophys. Res. Commun., 82, 1246 ± 1250.

Houge G, Doskeland SO, Boe R and Lanotte M. (1993).FEBS Lett., 315, 16 ± 20.

Houge G, Robaye B, Eikom TS, Golstein J, Mellgren G,Gjersten BT, Lanotte M and Doskeland SO. (1995). Mol.Cell. Biol., 15, 2051 ± 2062.

Hu L and Gudas LJ. (1990). Mol. Cell. Biol., 10, 391 ± 396.Kizaki H, Tadakuma T, Odaka C, Muramatsu J andIshimura Y. (1989). J. Immunol., 143, 1790 ± 1794.

Kizaki H, Suzuki K, Tadakuma T and Ishimura Y. (1990). J.Biol. Chem., 265, 5280 ± 5284.

Lalli E and Sassone-Corsi P. (1995). Proc. Natl. Acad. Sci.USA, 92, 9633 ± 9637.

Lanotte M, Hermouet S, Gombaud-Saintonge G and DoboI. (1986). Leuk. Res., 10, 1319 ± 1326.

Lanotte M, Riviere JB, Hermouet S, Houge G, WintermyrOK, Gjertsen BT and Doskeland SO. (1991). J. Cell.Physiol., 146, 73 ± 80.

Lanotte M, Gombaud-Saintonge G and Tertian G. (1986).Exp. Hematol., 14, 724 ± 731.

Laoide BM, Foulkes NS, Schlotter F and Sassone-Corsi P.(1993). EMBO J., 12, 1179 ± 1191.

Martin SJ and Green DR. (1995). Cell, 82, 349 ± 352.McConkey DJ, Orrenius S, Okret S and Jondal M. (1993).

FASEB J., 7, 580 ± 585.McConkey DL, Orrenius S and Jondal M. (1990). J.

Immunol., 145, 1227 ± 1230.McConkey DL, Jondal M and Orrenius S. (1992). Cellularsignaling in thymocyte apoptosis. Seminars in Immunol-ogy, Julius M. (ed.). Vol. 4, Acad. Press, London. pp.371 ± 377.

McKnight GS. (1991). Curr. Opin. Cell. Biol., 3, 213 ± 217.Molina CA, Foulkes NS, Lalli E and Sassone-Corsi P.(1993). Cell, 75, 875 ± 886.

Monaco LF, Foulkes NS and Sassone-Corsi P. (1995). Proc.Natl. Acad. Sci. USA, 92, 10673 ± 10677.

Montminy MR, Sevarino KA, Wagner JA, Mandel G andGoodman RH. (1986). Proc. Natl. Acad. Sci. USA, 83,6682 ± 6686.

Mori K, Oka S, Tani A, Ito S and Watanabe Y. (1989).Biochem. Biophys. Res. Comm., 162, 1535 ± 1540.

Mosmann T. (1983). J. Immunol. Meth., 65, 55 ± 61.Nagata S and Golstein P. (1995). Science, 267, 1449 ± 1456.Olsson IL, Breitman T and Gallo RC. (1982). Cancer Res.,42, 3922 ± 3928.

Potten CS. (1988). Cell, 53, 676 ± 678.Pratt RM and Martin GR. (1975). Proc. Natl. Acad. Sci.

USA, 72, 874 ± 877.Rossi GB, Migliaccio AR, Migliaccio G, Lettieri F, Di RosaM, Mastrobernardino G and Peshle C. (1980). Blood, 56,74 ± 79.

Rowan S and Fisher DE. (1997). Leukemia, 11, 457 ± 465.Ruchaud S, Zorn M, Davilar-villar E, Geneiser HG,Ho�mann C, Gjersten BT, Doskeland SO and LanotteM. (1995). Cell. Pharmacol., 2, 127 ± 140.

Ruchaud S, Duprez E, Gendron MC, Houge G, Geneiser H-G, Jastor� B, Doskeland SO and Lanotte M. (1994). Proc.Natl. Acad. Sci. USA, 91, 8428 ± 8432.

Sassone-Corsi P. (1994). EMBO J., 13, 4717 ± 4728.Sonnenburg WK and Smith WL. (1988). J. Biol. Chem., 263,6155 ± 6160.


835

Spaulding SW. (1993). Endocrine Rev., 14, 632 ± 650.Stehle JH, Foulkes NS, Molina CA, Simmoneaux V, PeÂ vet Pand Sassone-Corsi P. (1993). Nature, 365, 314 ± 320.

Steller H. (1995). Science, 267, 1445 ± 1449.Susuki K, Kizaki H, Tadakuma T and Ishimura Y. (1990).

Biochem. Biophys. Res. Comm., 171, 827 ± 831.Taetle R and Koessler A. (1980). Cancer Res., 40, 1223 ±1229.

Taylor SS, Buechler JA and Yonemoto W. (1990). Ann. Rev.Biochem., 59, 971 ± 1005.

Thompson CB. (1995). Science, 267, 1456 ± 1462.Tortora G, Tagliaferri P, Clair T, Colamonici O, NeckersLM, Robins RK and Cho-Chung YS. (1988). Blood, 71,230 ± 233.

Vintermyr OK, Gjersten BT, Lanotte M and Doskeland SO.(1993). Exp. Cell. Res., 206, 157 ± 161.

Weiler-Guettler H, Yu K, So� G, Gudas LJ and RosenbergRD. (1992). Proc. Natl. Acad. Sci. USA, 89, 2155 ± 2159.

Williams GT, Smith CA, Spooncer E, Dexter TM and TaylorDR. (1990). Nature, 343, 76 ± 79.

Williams GT. (1987). Gene, 53, 121 ± 126.Williams GT and Smith CA. (1993). Cell, 74, 777 ± 779.White E. (1996). Genes Dev., 10, 1 ± 15.Wyllie AH. (1993). Br. J. Cancer, 67, 205 ± 208.Wyllie AH, Kerr JF and Currie AR. (1980). Int. Rev. Cytol.68, 251 ± 305.


836

Documents

The transcriptional repressor ICER and cAMP-induced programmed cell death