Autonomous Rexinoid Death Signaling Is Suppressed by Converging Signaling Pathways in Immature...

Autonomous Rexinoid DeathSignaling Is Suppressed byConverging Signaling Pathways inImmature Leukemia Cells

G. R. Benoit, M. Flexor, F. Besancon, L. Altucci, A. Rossin,J. Hillion, Z. Balajthy*, L. Legres†, E. Segal-Bendirdjian,H. Gronemeyer, and M. Lanotte

INSERM U-496 (G.R.B., M.F., J.H, Z.B., L.L.,E.S.-B., M.L.)Centre G. HayemHopital Saint-Louis75010 Paris, France

INSERM U-365 (F.B.)Institut Curie75248 Paris Cedex 05, France

Institut de Genetique et de Biologie Moleculaire et Cellulaire(L.A., A.R., F.H.G.)Centre Nationale de la Recherche Scientifique/INSERM/ULPBP 163, 67404 Illkirch CedexC. U. de Strasbourg, France

Istituto di Patologia generale e Oncologia (L.A.)Seconda Universita degli studi di NapoliPiazzetta S. Andrea delle Dame 280138, Napoli, Italy

On their own, retinoid X receptor (RXR)-selectiveligands (rexinoids) are silent in retinoic acid recep-tor (RAR)-RXR heterodimers, and no selective rexi-noid program has been described as yet in cellularsystems. We report here on the rexinoid signalingcapacity that triggers apoptosis of immature pro-myelocytic NB4 cells as a default pathway in theabsence of survival factors. Rexinoid-inducedapoptosis displays all features of bona fide pro-grammed cell death and is inhibited by RXR, butnot RAR antagonists. Several types of survival sig-nals block rexinoid-induced apoptosis. RARa ago-nists switch the cellular response toward differen-tiation and induce the expression of antiapoptosisfactors. Activation of the protein kinase A pathwayin the presence of rexinoid agonists induces mat-uration and blocks immature cell apoptosis. Addi-tion of nonretinoid serum factors also blockscell death but does not induce cell differentiation.Rexinoid-induced apoptosis is linked to neitherthe presence nor stability of the promyelocyticleukemia-RARa fusion protein and operates also in

non-acute promyelocytic leukemia cells. Togetherour results support a model according to which rexi-noids activate in certain leukemia cells a defaultdeath pathway onto which several other signalingparadigms converge. This pathway is entirely distinctfrom that triggered by RAR agonists, which controlcell maturation and postmaturation apoptosis.(Molecular Endocrinology 15: 1154–1169, 2001)

INTRODUCTION

Retinoids regulate complex physiological events dur-ing development, control maintenance of homeosta-sis, and induce or inhibit cellular proliferation, differ-entiation, and death. Due to their strong differentiativeand antiproliferative activity, retinoids are used as can-cer therapeutic agents and may be able to exertcancer-preventive activities (1–3). The prototype of acancer that can be successfully treated with retinoidsis acute promyelocytic leukemia (APL), but also thetreatment of squamous cell carcinoma of the cervixand the skin (4) and Kaposi sarcoma (5) have beenreported. Moreover, retinoids can suppress oral pre-malignancy and prevent second primary head-and-neck tumors (6). Most, if not all, biological responses

0888-8809/01/$3.00/0Molecular Endocrinology 15(7): 1154–1169Copyright © 2001 by The Endocrine SocietyPrinted in U.S.A.

1154 on July 1, 2005 mend.endojournals.orgDownloaded from

http://mend.endojournals.org

to retinoids originate from the transcriptional control ofgene programs by the cognate nuclear receptors (7).Malfunction due to genetic defects associated withthese receptors or their downstream mediators, whichmay alter or interrupt retinoid signaling, can causemajor pathologies and may account for therapeuticfailures. Consequently, the comprehension of retinoidsignaling has been a major task in cell biology duringthis decade (reviewed in Ref. 7).

The highly pleiotropic effects of retinoids result fromthe combinatorial action of their six receptors [retinoicacid receptors (RARa, b, and g), and retinoid X recep-tors (RXRa, b, and g)], which can heterodimerize andact as ligand-inducible transcription-regulatory fac-tors. In addition, RXRs can also form heterodimerswith various other nuclear receptors [e.g. vitamin Dreceptor (VDR), thyroid hormone receptor (TR), perox-isome proliferator activated receptor (PPAR), and or-phan receptors], thereby modulating multiple signalingpathways (reviewed in Refs. 8 and 9). To assess thecontributions of the individual heterodimeric partnersand to investigate whether both RAR and RXR canautonomously induce cognate signaling pathways,retinoid panagonists/antagonists or receptor-selectiveagonists/antagonists have been developed (10–13).Using such reagents, RXR, which was previously con-sidered a nonautonomous signaling partner in theRXR-RAR heterodimer (14, 15), we have shown re-cently that rexinoids can signal autonomously in thecontext of an activated protein kinase A (PKA) path-way (16). This novel signaling paradigm operates in-dependently of RARs and promyelocytic leukemia(PML)-RARa, even in the presence of RAR antago-nists, and triggers the maturation not only of promy-elocytic NB4 (17) but also of retinoid-resistant NB4-R2cells (18, 19), thus bypassing the genetic defects of theresistant cells (16).

Significant insight has been gathered in recent yearsin the genetic basis of APL. Due to a t(15;17) chromo-somal translocation, a fusion protein between the reti-noic acid receptor a (RARa) and PML (20), the functionof which is still poorly understood (21–23), is formedand causes a differentiation block at the promyelocyticstage. It is believed that this fusion protein acts as adominant-negative mutant that impairs the action ofresidual RARa expressed from the second allele, butpharmacological doses of retinoic acid lead to a de-stabilization of the fusion protein and/or relieve itsdominant negative activity, with concomitant differen-tiation of the leukemic blasts (reviewed in Refs. 24–26).In APL cells, all-trans-retinoic acid (ATRA)-inducedmaturation is followed by late cell death process,which exhibits all the hallmarks of apoptosis (reviewedin Refs. 27–30), but whether maturation and apoptoticcell death are triggered by the same or distinct signal-ing pathways has remained elusive (21). The formationof PML-RARa may affect the kinetics/efficiency ofpostmaturation apoptosis in APL cells: in fact, theectopic expression of PML induces apoptosis (31–33),and it is conceivable that PML-RARa impairs this func-

tion of PML. Indeed, PML-RARa has been shown toexert antiapoptotic effects (33–35) and PML-RARadegradation induced by ATRA may facilitate the onsetof apoptosis observed after terminal maturation ofAPL cells (36). Accordingly, maturation-resistant APLcells might be restrained in their ability to embark onthe apoptosis program, even when the correspondingmachinery is functional. Apoptosis and maturation arelikely mechanistically coupled since numerous genespotentially involved in the apoptotic process are reg-ulated during maturation (c-myc, p53, Bcl-2, Bcl-xL,PML-RARa, PML). Although several reports indicatethat retinoids induce apoptosis in cells defective formaturation of APL cells (37, 38) or non-APL cells (39),convincing evidence that retinoids induce cell deathindependently of cell maturation in retinoid-responsivecells is lacking.

Here we provide the first evidence for an autono-mous rexinoid-induced default apoptosis programthat is operative in immature NB4 APL cells (17) and isentirely distinct from RAR agonist-controlled cell mat-uration and subsequent postmaturation apoptosis.Moreover, we demonstrate that rexinoid signaling isintegrated in, and controlled by, contextual signalingparadigms that affect NB4 cell growth and differenti-ation. Altogether our results strongly support the ideathat an autonomous rexinoid pathway for apoptosisexists in APL cells that operates independently of theRAR agonist-dependent pathway for cell maturationand postmaturation apoptosis. Rexinoid-induced ap-optosis is not an isolated feature of NB4 cells, whichwere derived from an APL patient classified FAB M3(17), as we observed it also in PLB985 cells that havebeen established from a patient with myelomonocyticleukemia (FAB M4) (40). Together, our results suggestthe existence of a novel RXR-selective cell biologicalactivity that could correspond to a basal death-by-default program of APL, and possibly other cell types.Apparently, cell life and proliferation in the presence ofrexinoid agonists require survival signals of very dif-ferent characteristics, three of which we have identi-fied. It thus appears that RXR may be a valuable phar-macological target for anticancer therapy.

RESULTS

Rexinoids Induce Apoptosis of Immature NB4Cells in Low Serum Cell Culture Conditions

That rexinoids can cross-talk in an RAR-independentmanner with other signaling pathways to induce celldifferentiation (16) could imply the existence of anautonomous RXR signaling pathway, the activity ofwhich is regulated positively or negatively by othersignals. Thus, the apparent absence of any biologicaleffects of rexinoids in cell culture systems could bedue to the masking by other signaling pathways. Tolimit the impact of such pathways and to exclude atthe same time possible interference from serum-borne

Rexinoids Autonomously Induce Maturation-Independent Apoptosis 1155

on July 1, 2005 mend.endojournals.orgDownloaded from


Fig. 1. RARa and RXR Agonists Induce Distinct Biological Responses in NB4 Promyelocytic Leukemia Cells Cultured underLimiting Serum Conditions

A, Dose-response (2 nM to 500 nM) to BMS753 (RARa agonist; green curve) and SR11237 (RXR agonist; red curve) of NB4 cellgrowth measured after 72 h. Cell viability was measured by the WST-1 colorimetric assay. Data (mean values of triplicates) wereexpressed in percent of the untreated control. At concentrations ranging from 2 to 20 nM, SR11237 has no significant effect oncell proliferation and viability, while at 100 nM, SR11237 induces complete cell death. BMS753 causes growth arrest and cellmaturation at concentrations greater than 100 nM. No cell death was observed for the highest concentration used (500 nM). Thecell morphologies (May-Grunwald Giemsa staining) corresponding to the indicated treatments are shown in insets. Apoptotic cellsexhibited massive nuclear fragmentation and chromatin spreading followed by cell disintegration. B, Electrophoretic analysis of

MOL ENDO · 2001 Vol. 15 No. 71156



retinoic acids, NB4 cells were adapted and perma-nently grown in low serum media (0.5% instead of10% FCS) supplemented with retinoid-free essentialgrowth regulators (see Materials and Methods).

Notably, while rexinoids have no apoptogenic effecton NB4 cells grown in high serum, they induce rapidand massive cell death with all typical features ofapoptosis when serum factor(s) are limiting. Cell deathinduced by the rexinoid agonist SR11237 (0.125 mM)occurred between 60 and 72 h of treatment with asequence of events typical for apoptosis: cell shrink-age, nuclear fragmentation, altered cytoskeleton ar-chitecture, and sudden cell disruption (Fig. 1 and datanot shown). DNA fragmentation was confirmed byclassical agarose gel electrophoresis (Fig. 1B) andflow cytometry analysis (TUNEL, Fig. 1C). Cell mor-phology changes and DNA cleavage were observedconcomitantly with caspase 3 activation and poly-ADP-ribose polymerase (PARP) cleavage (not shown;see also Fig. 4D). Rexinoid-induced apoptosis re-quired a transcriptionally active RXR, as the rexinoidantagonist BMS287 rescued NB4 cells from SR11237-induced apoptosis in a dose-dependent manner (Fig.1D), also demonstrating that the rexinoid is not toxicper se to these cells; we also did not notice any toxicitywith other cell types (data not shown). No sign ofsignificant cell maturation, as assessed by nitrobluetetrazolium (NBT) staining or CD11c cell surfacemarker positivity, accompanied rexinoid (SR11237)-induced cell death (Fig. 1E, lanes 5). Degradation ofPML-RARa, a hallmark of retinoid-induced maturationof APL cells, was not observed upon rexinoid treat-ment of NB4 cells in low serum (Fig. 1F; compare thePML-RARa and DPML-RARa bands in ATRA, rexi-noid-treated cells, and controls). In keeping with thesedata we did not find any alteration in the micropunc-

tate staining of PML nuclear bodies during rexinoidexposure (data not shown). We conclude that rexi-noid-induced NB4 cell apoptosis (in low serum condi-tions) occurs without prior differentiation and resultsfrom (bona fide) RXR-mediated gene programming inthe absence of transcriptionally active RARs.

In contrast to rexinoid signaling, retinoid action wasnot affected by low serum concentrations. Natural reti-noids (ATRA, preferentially binding to RARs, and 9-cisRA, binding to both RARs and RXRs), as well as RARa-specific agonists (BMS753), induced NB4 cell maturationsimilarly as in 10% serum (Fig. 1, A and E; see Ref. 16 forhigh serum data). Importantly, no apoptosis could beobserved after 60 h or 72 h treatment with the retinoidBMS753, whereas under identical conditions massiveapoptosis occurred in rexinoid-treated cells (Fig. 1, Band C). Note, however, that postmaturation apoptosisbecomes apparent after prolonged exposure to retinoidagonists (data not shown). The above data show that thechange in the concentration of serum factors affectsRXR/rexinoid but not RAR/retinoid-mediated signaling.Correspondingly, several events associated with post-maturation apoptosis of NB4 cells, such as Bcl-2 down-regulation, were not observed during rexinoid apoptosis,while they could be seen after retinoid treatment of thesecells also in low serum (data not shown).

RAR Antagonists Turn pan-RAR/RXR Agonistsinto Apoptotic Inducers

Given that distinct biological activities are induced byRAR (maturation) and RXR (apoptosis) agonists inconditions of limiting serum factors, we tested whetherblocking the RAR activity of the pan-RAR/RXR agonist9-cis RA would switch between the two responses.Indeed, while 9-cis-RA enhanced proliferation at low

DNA fragmentation during rexinoid-induced NB4 cell death on agarose gels. Apoptotic chromatin cleavage was monitored by theformation of DNA ladders to compare the apoptogenic activities of BMS753 (500 nM) and SR11237 (125 nM). No DNAfragmentation was detected after a treatment for 60 h, while SR11237-induced DNA fragmentation became apparent as early as48 h and was massive after 60 h ligand exposure. C, Flow cytometry analysis of DNA fragmentation by the TUNEL method. Cellswere treated as indicated in Fig. 1B and analyzed at 72 h. At this time 28% of the SR11237 (125 nM)-treated cells displayed DNAlabeling indicative of apoptosis (untreated control, 2%). However, due to massive apoptosis and cell disintegration the flowcytometry underestimates apoptosis, as disrupted cells and debris are lost during the cell washes. No DNA fragmentation wasdetected in BMS753 (500 nM)-treated cells. D, RXR-dependent induction of apoptosis by the RXR agonist SR11237. NB4 cellswere treated with increasing concentrations of the RXRa agonist SR11237 and the RXR antagonist BMS287, as indicated. Cellviability was analyzed at 72 h as described in Fig. 1A, using O.D. units in the WST-1 colorimetric assay for representation. Notethat at 500 nM the RXR antagonist significantly neutralizes the activity of 250 nM SR11237 (30%). About 60% of the SR11237activity is abolished by BMS287 (500 nM) when the agonist concentration is lowered to 125 nM. This 4:1 ratio is in keeping withthe differences in binding affinity for RXR of the two compounds used in competition. E, Rexinoid apoptosis occurs without cellmaturation. Cell differentiation was not observed by morphological criteria (not shown), surface membrane markers (CD11cexpression), or by functional assay (NBT reduction). Analyses were performed at 48 h (first signs of apoptosis in the culture) and72 h. Values correspond to the percentage of positive cells. Lane 1, Untreated control; lane 2, 9-cis RA (200 nM); lane 3, ATRA(200 nM); lane 4, BMS753 (500 nM); lane 5, SR11237 (125 nM). In lane 5 (72 h) the low counts further indicate massive apoptosis.F, Rexinoid apoptosis in low serum condition does not involve PML-RARa proteolysis. NB4 cells were incubated for 36 h in eitherlow (0. 5%; L) or high serum (10%; H) culture media with ATRA (1 mM), cAMP (200 mM), ATRA (1 mM) 1 cAMP (200 mM); SR11237(0.2 mM) and SR11237 (0.2 mM) 1 cAMP (200 mM). Cell responses were determined by morphological examination of stainedmicroscope slides. (M, maturation; A, apoptosis; “—”, neither maturation nor apoptosis). Cell extracts were analyzed bySDS-PAGE and membranes were probed with a specific antiserum raised against human RARa [RPa (F)]. The D-PML-RARaspecific degradation (97-kDa band) is only detected after ATRA or ATRA 1 cAMP treatment inducing cell maturation in both lowand high serum conditions; no PML-RARa degradation was observed during rexinoid-induced maturation in either condition.




Fig. 2. The RARa Antagonist BMS614 Converts the panRAR,RXR Agonist into a Death InducerA, Apoptosis dose-response of 9-cis RA in the presence of the RARa antagonist BMS614 (2 mM). Control cultures comprised

untreated cells (black square, 100%); BMS614 (2 mM) treated cultures (white square); dose-response to 9-cis-RA (from 2 nM

MOL ENDO · 2001 Vol. 15 No. 71158



concentrations (,100 nM) and induced growth arrestand cell maturation at high concentrations as early as48 h (Fig. 2A) without any sign of cell death (Fig. 2B),the cotreatment with 9-cis RA and the RARa antago-nist BMS614 induced rapid cell death (Fig. 2, A and B)in the absence of any cell maturation. Neither morpho-logical changes, nor up-regulation of CD11c, nor NBTreduction was observed (data not shown). Note thaton its own BMS614 was neither inhibiting cell growthnor inducing apoptosis (Fig. 2, A and B).

That RAR antagonists liberate the apoptotic activityassociated with pan-RAR/RXR agonists suggests thatcertain bifunctional ligands, i.e. RXR agonists that dis-play intrinsic RAR antagonistic activity, may be super-agonists for rexinoid-induced apoptosis. Such a ligandhas been described previously (BMS749; Ref. 16). In-deed, compared with SR11237, BMS749 displays aleft shift of two logs (EC50 400 nM and 8 nM, respec-tively, at 48 h) for its apoptotic activity (Fig. 2C). After72 h no surviving cells were found in cultures treatedwith BMS749 at 8 nM. Moreover, apoptosis occurredearlier after treatment with BMS749 than withSR11237 (48 h vs. 72 h). The different apoptogenicpotencies of BMS749 and SR11237 were also clearfrom the different kinetics of fragmentation of chromo-somal DNA (Fig. 2D) and TUNEL analysis (compareBMS749 in Fig. 2E, third panel from top, with SR11237in Fig. 1C, middle panel). Note that a second bifunc-tional compound displaying the same activity asBMS749, BMS772, acted also as a super death ago-nist in this system (Fig. 2C).

Rexinoids Induce Apoptosis in MyelomonocyticPLB985 Cells

To assess whether rexinoid-induced apoptosis is anisolated feature of APL cells or of the NB4 cell model,and whether it can operate independently of the pres-ence of the PML-RARa fusion protein, we adapted

myelomonocytic PLB985 cells (40) to low serum con-dition and exposed the cells to the rexinoid BMS749 orthe RARa agonist BMS753. BMS753 retarded moder-ately the proliferation of PLB985 cells and induceddifferentiation but exerted no apoptogenic effect, as isobvious from the absence of sub-G1 apoptotic bodies(Fig. 3A) and annexin positivity (Fig. 3B; see Fig. 3C forthe effect on proliferation). In contrast, whereas theBMS749 rexinoid had no effect on PLB985 cells in10% serum, an exposure of cells adapted to 1% se-rum resulted in a G1 block and more than 50% apo-ptosis after 3 days (Fig. 3, A–C; compare the cellsgrown in 1% serum in the absence and presence ofBMS749). Thus, rexinoids have apoptogenic potentialalso for non-APL cells, and apoptosis occurs indepen-dently of the PML-RARa fusion protein.

The RARa-Induced Terminal Maturation ProgramActs Dominantly over the Rexinoid-InducedProgram That Triggers Apoptosis of ImmatureBlasts

The observation that exposure of NB4 cells to thepan-RAR/RXR agonist 9-cis RA results in cell matura-tion suggests that the RAR activity (maturation) asso-ciated with this ligand can override the RXR activity(apoptosis). If true, this is an important aspect relevantto the design and use of rexinoids because most of therexinoids available to date possess (traces of) retinoidactivity that could potentially limit rexinoid action. Toaddress this issue directly, we carried out experimentsin which the pure RXR-specific agonist, SR11237, andthe RARa-selective agonist, BMS753, were mixed to-gether (Fig. 4). SR11237 was used at 250 nM, a con-centration not allowing any survival at 72 h (see Figs.1A and 4A, lane 2). BMS753, used at 500 nM, inducedNB4 cell maturation but no cell death could be de-tected at 72 h (Fig. 1C, bottom panel; Fig. 2E, secondpanel; Fig. 4A inset 9). Note that the decrease in viable

to 500 nM) (blue curve). NB4 cells were cultured for 48 h in presence of BMS614 (2 mM) plus increasing concentrations of 9-cis-RA(from 2 nM to 500 nM) (red curve). The estimated number of viable cells (O.D., arbitrary unit, means of triplicates) is given as percentof the untreated control (gray square; 100%). Under these conditions the values for 9-cis-RA were significantly above the controlfrom 6 nM to 50 nM indicating growth stimulation; growth inhibition associated with cell maturation was observed for concen-trations above 100 nM (blue curve); no apoptosis was detected. The cotreatment with 9-cis-RA (increasing concentrations) anda constant concentration of RARa antagonist (BMS614; 2 mM) shows a steep decrease in cell viability above 20 nM, associatedwith massive apoptosis (also apparent from cell morphology or DNA fragmentation, not shown). B, RARa antagonist converts amaturation-inducing retinoid into an apoptotic inducer. BMS614 and 9-cis-RA were used alone or combined as indicated in thefigure. The colors of histograms correspond to the color labels in Fig. 2A. Cultures were analyzed after 48 h of treatment. Cellviability was measured by the WST-1 assay (lanes 1–4). Cell morphology was analyzed after Giemsa staining (insets 1 to 4).BM614 (2 mM) affects neither cell proliferation nor cell viability. After 48 h (lane 4, inset 4) the combination of the two drugs inducesmassive cell death, making this combination more efficient that SR11237 (see Fig. 1). Note that 9-cis RA (0.1 mM) induces nogrowth arrest at 48 h when the first sign of morphological maturation is already visible (inset 2). C, Comparative analysis of theapoptogenic potential of RXR-specific ligands and bifunctional rexinoids in NB4 cells. Dose response to bifunctional rexinoids(BMS749, blue squares; BMS772, red squares) compared with the RXR-specific agonist SR11237 (green triangles). Cell viabilitywas evaluated as described above. D, Comparative analysis of the apoptotic potential of RXR specific ligands and bifunctionalrexinoids in NB4 cells. Electrophoretic analysis of DNA fragmentation during rexinoid-induced NB4 cell death on agarose gels.The experimental conditions are reported in the legend to Fig. 1B. (SR11237, 125 nM; BMS749, 50 nM). E, Flow cytometry analysisof DNA fragmentation in NB4 cells by the TUNEL method. Cells were treated (BMS753, 200 nM; BMS749, 200 nM) and analyzedat 48 h as indicated in Fig. 1B. The corresponding morphological features of cells are shown in panels at the right.




cell counts at this time (Fig. 1A; Fig. 4A, lane 9) reflectsthe growth inhibition associated with granulocyticmaturation (compare the cell morphologies depictedin insets 1 and 9 of Fig. 4A). Notably, increasing theconcentration of BMS753 efficiently inhibitedSR11237 rexinoid-induced apoptosis in a dose-de-pendent manner (Fig. 4A, lanes 2–8) with equimolarconcentrations of SR11237 and BMS753 resulting incell maturation (lane 7). We conclude that residualRARa activity masks or even blunts rexinoid signaling.

An initial screening of the activity of key factorsinvolved in the regulation of apoptosis revealed anincreased expression of several antiapoptosisgenes, such as bfl1, c-IAP1, c-IAP2, NAIP, as well asthe tumor suppressor p19ARF, in the presence ofRARa agonists (Fig. 4B), suggesting that these fac-tors may contribute to the antagonistic effect ofBMS753 on rexinoid-induced apoptosis. Indeed,bfl-1 (Fig. 4C) and the other above mentioned anti-apoptotic genes were induced when the cells wereexposed to both the BMS749 rexinoid and an ex-cess of the RARa agonist BMS753. No up-regula-

tion of these genes was seen with pure rexinoids(Fig. 3, B and C). Thus, in the presence of retinoids,the induction of an antiapoptotic gene program ap-parently counteracts the rexinoid-induced apopto-sis. In view of these results, it is possible that serum-borne retinoic acids have “disguised” rexinoidsignaling in cell culture systems, thus explainingwhy this pathway had not been detected earlier.

Given that RXR is a promiscuous heterodimerizationpartner for a great number of nuclear receptors, wewondered whether a ligand for any partner of RXR in aheterodimeric receptor complex could antagonizerexinoid-induced apoptosis similarly as retinoids, eventhough (with the exception of VDR) these receptors arenot involved in mediating NB4 cell maturation. How-ever, neither ligands specific for other RAR isotypes[RARb (BMS641), RARg (BMS961), nor for the VDR,TR, or PPARa, -b, and -g receptors had any antiapo-ptotic effect (Fig. 4D). These results suggest that reti-noids may simultaneously induce expression of themature phenotype and inhibit a default apoptosispathway triggered by RXR agonists.

Fig. 3. Rexinoid-Induced Apoptosis Is Operative in Myelomonocytic PLB985 Cells That Are Devoid of PML-RARaA, Flow cytometry analysis (propidium-iodide staining) of PLB985 cells grown in 10%, or adapted to 1%, serum. Cells were

treated for 72 h with 1 mM BMS749 or BMS753, as indicated. The percentage of sub-G1 particles representing apoptotic bodiesare given. B, Percentage of annexin V-positive cells in high and low serum after 96 h exposure to the indicated ligands. C,Proliferation curve of PLB985 cells treated as indicated in 1% serum-containing medium.

MOL ENDO · 2001 Vol. 15 No. 71160



Rexinoid-Induced Apoptosis Is Operative inRetinoid-Resistant NB4-R2 Cells

Rexinoid signaling can still function in retinoid-resistantAPL cells, such as the NB4-R2 cell line in which resis-tance is due to a point mutation that truncates the ligand-binding domain of the PML-RARa fusion protein (16, 19).We therefore investigated whether rexinoid signalingwould operate in low serum conditions to induce apo-ptosis of NB4-R2 cells. Both SR11237 (data not shown)and the bifunctional RXR agonist RAR-antagonistBMS749 induced cell death under these conditions inboth NB4 and NB4-R2 cells involving caspase 3 activa-tion (Fig. 5A) and PARP cleavage (Fig. 5B). Importantly, inthe absence of a functional PML-RARa, the pan-RAR/RXR agonist 9-cis RA was also able to induce apoptosis(Fig. 5C, middle panel). This is in keeping with our resultsdemonstrating that RARa signaling overrides the rexi-noid apoptosis signaling.

PKA Agonists Switch the Rexinoid Responsefrom Apoptosis to Differentiation

The above data suggest that rexinoid-induced apoptosisof immature APL cells and retinoid-induced maturationof these cells are mutually exclusive phenomena. In viewof our recent demonstration of the existence of a novelNB4 cell maturation pathway that involves a cross-talkbetween rexinoids and PKA agonists (16), we investi-gated whether also this alternative differentiation path-way would be incompatible with rexinoid apoptosis un-der low serum conditions. Indeed, addition of PKAagonists, such as 8-chloro-phenyl-thio-cAMP (8CPT-cAMP), blunted BMS749-rexinoid-induced apoptosisand triggered maturation not only in NB4 but, notably,also in the retinoid-resistant NB4-R2 cells (Fig. 6). Theseresults indicate the existence of several independenttypes of “check points” or “controlling systems” thatallow the cell to switch on or off the rexinoid-dependentcell death or maturation pathways.

Rexinoid Apoptosis Is Rescued by Serum Factors

Our initial rationale for using low serum conditions in theexperiments described above was to, 1) exclude thepossible “contamination” with serum-borne retinoids tostudy “pure” rexinoid action, and 2) to exclude or limit apossible cross-talk between rexinoids and signalingpathways induced by serum factors. To reveal the pos-sible role of serum factors in rexinoid-induced apoptosisin the absence of any cell differentiation, we studied theeffect of increasing serum concentrations (0.5% to 10%FCS) on BMS749 (500 nM)-induced NB4 cell death. Se-rum efficiently rescued the cells from the apoptopic ac-tion of the BMS749 rexinoid (Fig. 7). This rescue oc-curred also in the presence of RAR antagonists, thusconfirming that it corresponded to a signaling phenom-enon different from that triggering RARa-dependentmaturation. Also serum depleted of hormones by char-coal treatment showed similar capacity to inhibit apo-

ptosis (not shown), indicating that the serum compo-nent(s) that gives rise to the “rescue” effect is not a smallmolecule that can be readily absorbed to active surfaces.Importantly, rexinoid apoptosis was inhibited by serum inthe absence of any sign of NB4 cell maturation (data notshown). We conclude that a nonretinoid activity in serumis able to suppress rexinoid apoptosis.

Nuclear Factor-kB (NF-kB) Is Activated duringRetinoid-Induced Maturation but Serum FactorsDo Not Use This Survival Pathway to SuppressRexinoid-Dependent Death Signaling

That rexinoid-induced death of immature NB4 cells isentirely different from that subsequent to retinoid-induced differentiation is strongly supported by theanalysis of the expression patterns of several apopto-sis-regulatory key genes. In particular, the expressionof a number of antiapoptosis genes that are induced byretinoids is not affected during rexinoid death signaling(Fig. 4, B and C) and only tumor necrosis factor-a (TNFa)expression was augmented in the panel of genes testedwhen NB4 cells were exposed to rexinoids under lowserum conditions (data not shown). To investigatewhether serum factors would determine the cell fate viaTNF-elicited nuclear factor-kB (NF-kB)-mediated signal-ing, the activation of NF-kB by retinoids, rexinoids, and P75 A agonists was tested in low and high serum condi-tions by electrophoretic mobility shift assay (EMSA) (Fig.8). Clearly, NF-kB activation correlated with cell matura-tion induced by either RARa agonists or rexinoid/PKAcross-talk, but not with rexinoid-induced cell death. Im-portantly, serum factors did not change this pattern ofNF-kB nuclear activation. Moreover, the observationsthat serum rescue from rexinoid apoptosis does not in-volve the induction of antiapoptotic genes and NF-kB, asis the case for RARa agonists, indicates that serum fac-tors use a distinct survival pathway to suppress rexinoid-dependent death signaling.

Together these results demonstrate that 1) retinoidand rexinoid signaling activate distinct biological pro-grams in NB4 cells and 2) rescue from rexinoid apo-ptosis by RARa agonists and serum factors involvesdistinct survival programs.

DISCUSSION

Rexinoids Induce an Autonomous Death Pathwayin Promyelocytic Leukemia Cells

Several lines of evidence indicate that we have iden-tified a novel rexinoid-dependent apoptogenic signal-ing pathway that is operative in immature NB4 cellsand is distinct from previously investigated postmatu-ration apoptosis. These conclusions are supported bythe following observations: 1) rexinoid apoptosis re-quires a transcriptionally active RXR independently ofprior cell differentiation, 2) differentiation blocks rexi-noid apoptosis of immature cells, 3) rexinoid apoptosis




Fig. 4. Rexinoid-Dependent Signaling for Cell Death Is Suppressed by RAR AgonistsA, Rescue of rexinoid-induced NB4 cell death by RARa ligands occurs concomitantly with the induction of cell maturation. Cell

viability was estimated by WST-1 assay and expressed in percent of the untreated control (as described above). Cell morphologywas assayed by MGG staining. Cells were cultured for 72 h and analyzed. Increasing concentrations of the RARa agonist BMS753

MOL ENDO · 2001 Vol. 15 No. 71162



is operative in retinoid-resistant cells and is even en-hanced in the presence of RARa antagonists that in-hibit cell differentiation, 4) serum factors block rexinoid

apoptosis but not retinoid-induced cell differentiationand postmaturation apoptosis, 5) rexinoid apoptosis isfully functional in retinoid-resistant cells that do notdifferentiate or undergo postmaturation apoptosis,and 6) the differential expression of known key genesregulating cell life and death indicates that rexinoidapoptosis and postmaturation death are two com-pletely distinct gene programs. This latter conclusionis further supported by the observation that retinoidsrescue cells from rexinoid apoptosis whereas theysynergize with RXR ligands for maturation and post-maturation death.

Note that rexinoid apoptosis is a signaling pathwaythat is entirely distinct from the apoptosis observed byso-called “pseudo-retinoids” such as CD437 or 4-HPR(41–43). This conclusion is based on the observationsthat 1) CD437 and 4-HPR have been reported to in-duce apoptosis in high serum (41–43); 2) CD437-in-duced NB4 cell apoptosis in low serum occurs with thesame potency as in high serum media; 3) neither PKAnor RARa agonists could diminish the CD437-depen-dent apoptosis of NB4 cells (G. Benoit and M. Lanotte,unpublished); and 4) both CD437 and 4-HPR are de-void of any measurable rexinoid activity in reporter cellassays (C. Gaudon and H. Gronemeyer, unpublished).

RXR within the RAR-RXR heterodimer is believed tobe silenced by apo-RAR (a phenomenon also termed“RXR subordination”) but may synergize with holo-RAR although the mechanistic basis of this phenom-enon is still a matter of controversy (12, 14, 44–47). Inaddition to its signaling through RAR-RXR het-erodimers, RXR homodimers and a great number ofalternative RXR heterodimers can signal in target cells(9). What could be the RXR signaling entity that trig-gers immature APL cell apoptosis? Our study does notsupport an implication of RAR-RXR heterodimers thatare believed to mediate differentiation and postdiffer-entiation apoptosis of NB4 and F9 cells (12, 48, 49),mainly because bifunctional ligands, such as BMS749,do apparently not generate a transcriptionally active

Fig. 4D.

(2–500 nM) were combined with a constant concentration of the RXR agonist (SR11237; 250 nM). B, Retinoids and rexinoids inducethe expression of distinct sets of proliferation and (anti)apoptosis mediators. Modulation of the mRNA levels of a number of keyfactors (denoted at the right) known to be involved in the regulation of proliferation (p19) and apoptosis (TRAFs, IAPs, Bfl1) asassessed by multiplex RNAse protection assays. NB4 cells grown in low serum conditions were exposed to the agents displayedat the top for 0, 12, 24, 36, and 48 h (subsequent lanes for each treatment). Only sections of the corresponding gels are shown;the bottom panel gives a representative example of the expressions of L32 and GAPDH used as the invariant internal controlsfor calibration. Note that the invariant controls were equivalent to the one shown in all cases displayed here. C, Retinoid-dependent rescue from rexinoid-induced apoptosis correlates with the induction of antiapoptogenic gene programs. MultiplexRNAse protection assays with bcl2 family members. Expression of the antiapoptotic bfl-1 gene is induced when an excess of theRARa agonist BMS753 is added to NB4 cells exposed to the rexinoid BMS749 (lanes 6–9). Exposure times were 0 (lane 1) and12 h, 24 h, 36 h, and 48 h (lanes 2–5, 6–9, and 10–13). “Probe” corresponds to the nondigested multiplex probe; lines point tothe smaller gene expression-indicative fragments after hybridization and RNAse treatment. D, Action of the ligands of various RXRpartners on rexinoid-induced apoptosis. Open circles, Untreated control; black filled squares, dose-response to ligands for theRXR partner; red squares and red curves, dose-response to the RXR agonist (SR11237); the arrowed red square indicates theresponse to 200 nM SR11237; this concentration (200 nM SR11237) is used together with increasing concentration of the variousligands for the RXR partners (open blue squares and blue curve). The concentrations of the various ligands for the RXR partners[RARa, BMS753 (panel 1); RARb, BMS641 (panel 2); RARg, BMS961 (panel 3); VDR, vitamin D3, (panel 4); TR, T3 (panel 5); PPARpan agonists, BMS 990 (panel 6), BMS530 (panel 7), BMS972 (panel 8)] ranged from 2 nM to 500 nM. Cell viability was evaluatedafter 72 h of treatment using the WST-1 assay (O.D. arbitrary unit, means of triplicates, values in % of the untreated control).




RAR-RXR heterodimer. The observation that a tran-scriptionally active RXR is required for rexinoid apo-ptosis suggests an implication in this phenomenon ofeither RXR homodimers, for which so far neither aseparate signaling pathway nor cognate target genes(or so-called “permissive” heterodimers) have beenidentified. Clearly further genetic studies, involving, forexample, mutants that are deficient in RXR homo- orheterodimerization, are required to provide evidence

for the existence of a potentially existing RXR ho-modimer death signaling pathway.

Rexinoid Death Signaling: A Default Pathway inDisguise?

The observation that the knockout of RXRa generatesa lethal phenotype indicates that RXR is more thansimply a silent heterodimerization partner (50). Our

Fig. 5. Rexinoid-Induced Apoptosis in the Retinoid-Resistant NB4-R2 CellsA, Caspase 3 activity as measured by cleavage of the colorimetric substrate DEVD-pNA (see Materials and Methods). B,

Western blot analysis of caspase-3 and PARP cleavages in response to rexinoid treatment in NB4 and NB4-R2 cells. C, Flowcytometry analysis of DNA fragmentation in NB4-R2 cells during retinoid and rexinoid treatments. The experimental conditionsare those decribed in Fig. 1C. Note that 9-cis-RA induces apoptosis in NB4-R2 cells. As mentioned in legend to Fig. 1C, apoptosisin BMS749- treated cultures is underestimated due to cell disruption. In contrast to the situation in wt cells, BMS753 did notabrogate BMS749-induced apoptosis in NB4-R2 cells.

MOL ENDO · 2001 Vol. 15 No. 71164



results strongly support the implication of RXR (li-gands) in the regulatory mechanisms controlling thecell life and death balance (Fig. 9), which had probablygone unnoticed because of the suppressive action ofserum factors that are present in virtually all experi-ments done with cultured cells or because its actionwas masked by retinoids and/or other signaling path-ways. In this respect, it will be of interest to assess therexinoid responsivity of hematopoietic and nonhema-topoietic cells other than NB4 under conditions oflimiting serum factors: notably, that RXR is required forapoptosis in cells of nonhematopoietic origin, which isbased on the observation that retinoid-induced apo-ptosis is blunted in mouse F9 embryo carcinoma cellslacking RXRa (49, 51). In this case, however, it hasremained unclear whether RXR apoptogenic signalingis autonomous. Also in non-APL HL60 cells, RXR li-gands were required for postmaturation apoptosis butonly after prior exposure of the cells to retinoids andsubsequent cell maturation (52). Again, it will be inter-esting to assess whether rexinoids have the capacityto signal autonomously in these cells. It is worth notingthat retinoid-rexinoid signaling in HL60 is apparentlydistinct from that in APL cells where RARa agonistssuffice to induce maturation and postmaturation apo-ptosis (Fig. 9). The underlying mechanism(s) account-ing for distinct action of retinoids and rexinoids inthese two cell lines are not yet elucidated but may be

linked to the differential expression of the PML-RARafusion protein.

It is tempting to speculate that (endogenous) rexi-noids may even correspond to death inducers, de-pending on the signaling context (e.g. hormones,cytokines, extracellular matrix; see Ref. 16 for PKAaction) of a cell at a certain time in development orposition within the cell lineage. Indeed, evidence forthe possible existence of endogenous rexinoids hasbeen obtained with transgenic “reporter” mice (53, 54).Thus, RXRs may correspond to attractive targets fordrug design, possibly in combination with compoundsthat alter the inhibitory activity of retinoid agonists(preferably in the form of a bifunctional retinoid, suchas BMS749) or endogenous signals that correspond tothe unknown serum factors observed in this study.Further investigation to determine the cascade ofevents downstream from the RXR-dependent tran-scriptional regulation and of the nature of the interfer-ing signals in vivo should provide new insights on hownatural retinoids control cell fate in cells and develop-ing organisms.

MATERIALS AND METHODS

Reagents and Drugs

ATRA, 9-cis retinoic acid (9-cis RA), and phorbol 12-myristate13-acetate (PMA) were purchased from Sigma (St. Louis,MO). The BMS753, BMS649 (SR11237), BMS614, BMS493,

Fig. 6. Antiapoptotic Action of cAMP Analog in CulturesNB4 and NB4-R2 cells were cultured in similar conditions,

with a constant concentration of the RXR agonist BMS749(500 nM) and with increasing concentration of 8-CPT-cAMP inculture medium [RPMI1640 supplemented with essential fac-tors in HY supplement (1% vol/vol), FCS (0.5% vol/vol)]. Cellviability was evaluated using WST-1 assay (as describedabove). Data (O.D., arbitrary unit, means of triplicates) areexpressed in percent of the untreated control in basal media.In the absence of 8-CPT-cAMP, NB4 and NB4-R2 cell cul-tures showed no viable cells (gray square). The curves (NB4,black squares; NB4-R2, white squares) show the death res-cue by increasing concentration of cAMP. Note that viablecell counts also reflect growth arrest associated with matu-ration, as also observed for the action of BMS753 (see inFig. 4A).

Fig. 7. Antiapoptotic Action of Serum Factors from Serum inCultures

NB4 cells were cultured with increasing concentrations (%)of serum supplement in culture medium [RPMI 1640 supple-ment with essential factors in HY supplement (1% vol/vol),defined as basal media]; opened squares (curve A). In asecond series of cultures, in conditions similar to panel A, afixed concentration of the RXR agonist BMS749 (500 nM) wasadded (curve B). Cell viability was evaluated using WST-1assay (as described above). Data (O.D., arbitrary unit, meansof triplicates) were expressed as the percentage of the un-treated control in basal media. The inset shows the value ofthe ration B/A at identical serum concentration in cultures.




BMS009, BMS287, BMS772, and BMS749, provided byBristol-Myers Squibb (Princeton, NJ), are synthetic retinoidswith receptor selectivity, the features of which have beenreported previously (16).

Cell Lines, Cultures, and Analysis of Cell Maturationand Cell Viability

NB4, NB4-R2, and PLB985 cells (17, 18, 40) were adapted toculture conditions with minimal serum addition in the syn-thetic media and allowing optimal cell proliferation and/ordifferentiation and long-term survival. To this purpose, cellswere maintained in RPMI 1640 medium (Life Technologies,Inc., Gaithersburg, MD) supplemented with 1% (vol/vol) HY(Life Technologies, Inc.), 0.5% (NB4 and NB4R2) and 1%(PLB985) (vol/vol) FCS (Bayer Corp., Elkhart, IN), glutamine (2mM), and antibiotics in a humidified incubator at 37 C with 5%CO2. Morphological studies were performed on smearsstained with May-Grunwald-Giemsa (MGG; Sigma). Cell mat-uration was measured by NBT reaction. Results are ex-pressed as percentage of NBT-positive cells after a count on

300 cells. Cell viability was measured by the WST-1 colori-metric assay (Roche Molecular Biochemicals, Indianapolis,IN). Data (mean values of triplicates) were expressed in per-cent of the untreated control.

Analyses of Apoptotic Features

Apoptosis was assessed by the TUNEL method, propidiumiodide staining, or annexin V immunostaining using the an-nexin V detection kit (Roche Molecular Biochemicals); sam-ples were analyzed by as recommended by the supplier.Briefly, cells were incubated in buffer (10 mM HEPES/NaOH,pH 7.4, 140 mM NaCl, 2.5 mM CaCl2) containing annexin-V-fluorescein isothiocyanate (1 mg/ml) for 10 min in the dark.After resuspension in 1 ml labeled buffer, samples were an-alyzed using the FACScan flow cytometer. For the TUNELassays the Fluorescent In Situ Cell Death Detection kit(Roche Molecular Biochemicals) was used according to themanufacturer protocol except that cells were fixed inPBS-4% formaldehyde. Labeled cells were analyzed usingthe FACSCALIBUR. Internucleosomal DNA cleavage was vi-sualized after agarose gel electrophoresis. DNA was isolatedfrom 2 3 106 cells according to the procedure described byMiller et al. (55), modified as we previously described (56).Caspase activities of total cell extracts were measured usinga colorimetric procedure. Briefly, 2 3 106 cells were har-vested and lysed in buffer A [50 mM Tris-HCl, pH 7.5, 0.03%NP40, 1 mM dithiothreitol (DTT)] after washing in PBS, pH 7.2.Unsoluble material was removed by centrifugation at 14,000rpm for 15 min at 4 C. Protein concentration of the superna-tant was measured using the BCA assay reagent (PierceChemical Co., Rockford, IL). The reaction was set up in96-well plates by adding 0.2 mM of specific colorimetric sub-strate DEVD-pNa for Caspase-3 to 0.01 ml of lysate incaspase reaction buffer (100 mM HEPES, pH 7.5, 10% su-crose, 0.1% 3-([3-cholamidopropyl]dimethylammonio)-2-hydroxy-1-propanesulfonate, 10 mM DTT). The reaction wasincubated at 37 C, and release of pNa was measured byabsorbance reading at 405 nm once per hour during 5 h.Enzyme activity was measured as initial velocity of the enzy-matic kinetic.

Immunofluorescence Analysis and Flow CytometryAnalysis of Cell Surface Antigen

Immunofluorescence analysis of Bcl-2 expression was per-formed as described previously. Briefly, after treatment bythe indicated compound, cells were smeared on histologicalglass slides using a cytocentrifuge (Cytospin, Shandon). Afterovernight drying, cell smears were fixed in acetone at 4 C for10 min and allowed to air dry for 20 min. The slides were thensequentially incubated with PBS for 15 min, and monoclonalmouse antibody was raised against human Bcl-2 protein(DAKO Corp., Carpenteria, CA) at a dilution of 1:400 in PBSfor 1 h. After three washes in PBS, the slides were incubatedwith fluorescein-coupled antimouse antibody (Sigma) at adilution of 1:200 in PBS for 30 min. After three washes in PBS,the slides were mounted with 5 ml of fluorescent mountingmedium (DAKO Corp.) 0.2% DAPI. All incubations were atroom temperature. Preparations were examined by Fluores-cent Microscopy. Images were collected and digitalized us-ing a CCD color camera and QWIN software (Leica Corp.,Deerfield, IL). The expression of the membranous adhesionmolecule CD11c integrin was analyzed by direct immunoflu-orescence. After incubation with the indicated compounds,cells were washed in PBS and labeled with antihuman CD11cPE mouse monoclonal antibodies (Becton Dickinson and Co.,Franklin Lakes, NJ). Cells were then washed twice in PBS andfixed in 1% paraformaldehyde/PBS solution. Cells were an-alyzed using a FACSCALIBUR (Becton Dickinson and Co.)flow-cytometer.

Fig. 8. EMSA Measurement of NF-kB Activation by SerumRetinoids, Rexinoids, and cAMP in NB4 Cells

NB4 cells were treated with the indicated agents (serum,10% vol/vol; SR11237, 500 nM; 8-CPT-cAMP, 200 mM;BMS753, 500 nM) for 48 h with the exception of TNFa (5 h).EMSA was carried out as described in Materials and Meth-ods. Cell maturation and/or apoptosis was evaluated in par-allel on the same cultures. Biological responses are indicatedas an inset on the figure (no biological response detected (2);apoptosis (A); maturation (M). The migration shift of the probebound to NF-kB was shown by autoradiography. Autoradio-grams were scanned and analyzed with Image Quant com-puter program. Values (%) were expressed as increase ofbinding compared with the untreated cell control (lane 1).Lane 10 shows background control (no nuclear extract).

MOL ENDO · 2001 Vol. 15 No. 71166



Ribonuclease (RNAse) Protection Assays

Total RNA was extracted with the Trizol reagent (Life Tech-nologies, Inc., cat. 15596–018). The RNAse protection assaywas performed according to the supplier’s instructions(PharMingen, San Diego, CA). Briefly, the corresponding tem-plate sets (PharMingen) were labeled with [a-32P] uridinetriphosphate. RNA (4 mg) and 6 to 8 3 105 cpm of labeledprobes were used for hybridization. After RNAse treatments,the protected probes were resolved on a 5% urea-polyacryl-amide-bis-acrylamide gel.

Preparation of Nuclear Extracts and ElectrophoreticMobility Shift Assay

Cells were lysed in buffer A (10 mM HEPES, pH 7.9, 1 mM

EDTA, 60 mM KCl, 1 mM DTT, 0.05% NP40, 1 mM phenyl-methylsulfonyl fluoride, 2 mg/ml of aprotinin, antipain, andleupeptin) for 5 min on ice, and the cell lysate was centrifugedat 2,000 3 g for 5 min. The nuclear pellet was then washedin buffer A without NP 40, resuspended in buffer B (20 mM

Tris HCl, pH 8, 1.5 mM MgCl2, 600 mM KCl, 0.2 mM EDTA, 0.5mM DTT, 25% glycerol, 1 mM phenylmethylsulfonyl fluoride, 2mg/ml of aprotinin, antipain, and leupeptin), frozen at 280 C,and centrifuged at 10,000 3 g for 30 min to remove debris.Protein concentration of nuclear extracts was determined bythe Bradford assay. For EMSA, the double-stranded consen-

sus NF-kB probe, 59-AGT TGA GGG GAC TTT CCC AGGC-39; 39-TCA ACT CCC CTG AAA GGG TCC G-59, was end-labeled using [g-32P] ATP and T4 polynucleotide kinase.Binding reactions were carried out in a 20 ml binding reactionmixture [10 mM Tris-HCl, pH 7.5, 50 mM NaCl, 0.5 mM DTT,10% glycerol, 0, 2% NP40, and 4 mg of poly(dI-dC)(dI-dC)]containing 5 mg of nuclear proteins and 0.5 ng of the radio-labeled probe. Samples were incubated for 45 min on ice andfractionated by electrophoresis on a 6% nondenaturing poly-acrylamide gel in TAE buffer (7 mM Tris, pH 7.5, 3 mM sodiumacetate, 1 mM EDTA). Gels were run at 180 V for 2.5 h at 4 C,dried, and autoradiographed.

Protein Extraction and Western-Blot Analysis

Total protein extracts were prepared. Briefly, cultured cellswere washed in PBS and pelleted by centrifugation at 400 3g for 5 min. Pellets of 2 3 106 cells were immediately lysed byadding 100 ml of a boiling Laemmli solution containing b-mer-captoethanol and disrupted with a pestle. Samples were thenboiled for 5 min and insoluble material was removed bycentrifugation at 13,000 rpm for 5 min. Protein amount wasquantified by a Coomassie Blue staining. Protein extracts (10mg) were loaded on SDS-polyacrylamide gels, electropho-resed, and blotted onto polyvinylidene fluoride membranes(Millipore Corp., Bedford, MA). After transfer, proteins werevisualized with Ponceau S (Sigma) to confirm equal loading of

Fig. 9. Two Default Signaling Pathways Triggered by Retinoids and Rexinoids Determine Life and Death of NB4 PromyelocyticLeukemia Cells

Simplified schematic illustration of retinoid and rexinoid signaling pathways that affect NB4 cell maturation, survival, andapoptosis. The rexinoid pathway (top), activated by rexinoid (i.e. RXR selective) agonists and mediated by either so-called“permissive” (see text for details and references) RXR heterodimers (RXR-“X”) with an unknown nuclear receptor heterodimer-ization partner or RXR homodimers [(RXR)2], leads by default to immediate apoptosis of immature NB4 cells. This pathway ischaracterized by its insensitivity toward RARa antagonists. Several alternative signaling options can rescue NB4 cells fromrexinoid-induced apoptosis, including RARa and PKA agonists, both of which activate pathways that lead to cell maturation, andpresently uncharacterized serum factors that induce survival. The second default signaling pathway (bottom) is dependent onRARa agonists, abrogated by RARa antagonists, and leads to cell maturation followed by postmaturation apoptosis. The receptorspecies involved in this signaling have not been unequivocally determined and may involve RXR-RARa or RXR-PML-RARaheterodimers or oligomers (57) of PML-RARa [(PML-RARa)x]. Note that coordinate activation of RARa and RXR leads tosynergistic activation of cell maturation.




protein. Membranes were blocked with 5% non-fat dry milk inPBS, pH 7.6, 0.1% Tween 20 (PBS-T), and then incubatedwith a specific antiserum raised against the indicated proteinin PBS-T 0.5% milk for 18 h at 4 C. Membranes were incu-bated with horseradish peroxidase-coupled antibody (TheJackson Laboratory, Bar Harbor, ME) for 30 min at 25 C. Eachof these steps was followed by three washes for 10 min inPBS-T 0.5% milk. Labeling was performed as described inthe ECL detection kit (Amersham Pharmacia Biotech, Arling-ton Heights, IL).

Acknowledgments

We thank the Bristol-Myers-Squibb chemists for providingthe synthetic retinoids. PLB985 cells were generously pro-vided by Dr. Y. E. Cayre (Paris). L.A., A.R., and H.G. thankMichele Lieb for technical help and Emmanuelle Wilhelm fortechnical support and expert advice.

Received November 10, 2000. Revision received February14, 2001. Accepted March 12, 2001.

Address requests for reprints to: M. Lanotte, INSERMU-496, Centre G. Hayem, Hopital Saint-Louis, 1, AvenueClaude Vellefaux, 75010 Paris, France. E-mail: [email protected] or H. Gronemeyer, IGBMC, BP 163,67404 Illkirch CEDEX, CU de Strasbourg, France. E-mail:[email protected].

This work was supported by funds from the Institut Na-tional de la Sante et de la Recherche Medicale, the CentreNational de la Recherche Scientifique, the Hopital Universi-taire de Strasbourg, Bristol-Myers-Squibb, and grants to M.L.from the Ligue Nationale contre le Cancer and Associationpour la Recherche contre le Cancer (ARC).

* Present address: Department of Biochemistry, UniversityMedical School, Debrecen, Nagyerdei krt. 98. H-4012,Hungary.

† Present address: Service d’Anatomie-Pathologie, CentreG. Hayem, Hopital Saint-Louis, 1, Avenue Claude Vellefaux,75010 Paris, France.

REFERENCES

1. Lotan R 1996 Retinoids in cancer chemoprevention.FASEB J 10:1031–9

2. Sporn MB, Roberts AB, Goodman DS 1994 TheRetinoids: Biology, Chemistry and Medicine. RavenPress, New York

3. Lippman SM, Lotan R 2000 Advances in the develop-ment of retinoids as chemopreventive agents. J Nutr130:479S–482S

4. Lippman SM, Parkinson DR, Itri LM, Weber RS, SchantzSP, Ota DM, Schusterman MA, Krakoff IH, GuttermanJU, Hong WK 1992 13-cis-retinoic acid and interferona-2a: effective combination therapy for advanced squa-mous cell carcinoma of the skin. J Natl Cancer Inst84:235–241

5. Bonhomme L, Fredj G, Averous S, Szekely AM, EcsteinE, Trumbic B, Meyer P, Lang JM, Misset JL, Jasmin C1991 Topical treatment of epidemic Kaposi’s sarcomawith all-trans-retinoic acid. Ann Oncol 2:234–235

6. Hong WK, Lippman SM, Itri LM, Karp DD, Lee JS, ByersRM, Schantz SP, Kramer AM, Lotan R, Peters LJ 1990Prevention of second primary tumors with isotretinoin insquamous-cell carcinoma of the head and neck. N EnglJ Med 323:795–801

7. Chambon P 1996 A decade of molecular biology of reti-noic acid receptors. FASEB J 10:940–954

8. Gronemeyer H, Laudet V 1995 Transcription factors 3:nuclear receptors. Protein Profile 2:1173–1308

9. Mangelsdorf DJ, Thummel C, Beato M, Herrlich P,Schutz G, Umesono K, Blumberg B, Kastner P, Mark M,Chambon P, Evans M 1995 The nuclear receptorsuperfamily: the second decade. Cell 83:835–839

10. Lehmann JM, Jong L, Fanjul A, Cameron JF, Lu XP,Haefner P, Dawson MI, Pfahl M 1992 Retinoids selectivefor retinoid X receptor response pathways. Science 258:1944–1946

11. Apfel CM, Kamber M, Klaus M, Mohr P, Keidel S,LeMotte PK 1995 Enhancement of HL-60 differentiationby a new class of retinoids with selective activity onretinoid X receptor. J Biol Chem 270:30765–30772

12. Chen JY, Clifford J, Zusi C, Starrett J, Tortolani D,Ostrowski J, Reczek PR, Chambon P, Gronemeyer H1996 Two distinct actions of retinoid-receptor ligands.Nature 382:819–822

13. Sun SY, Yue P, Dawson MI, Shroot B, Michel S, LamphWW, Heyman RA, Teng M, Chandraratna RA, Shudo K,Hong WK, Lotan R 1997 Differential effects of syntheticnuclear retinoid receptor-selective retinoids on thegrowth of human non-small cell lung carcinoma cells.Cancer Res 57:4931–9

14. Vivat V, Zechel C, Wurtz JM, Bourguet W, Kagechika H,Umemiya H, Shudo K, Moras D, Gronemeyer H, Cham-bon P 1997 A mutation mimicking ligand-induced con-formational change yields a constitutive RXR that sensesallosteric effects in heterodimers. EMBO J 16:5697–709

15. Schulman IG, Li C, Schwabe JWR, Evans RM 1997 Thephantom ligand effect: allosteric control of transcriptionby the retinoid X receptor. Gene Dev 11:299–308

16. Benoit G, Altucci L, Flexor M, Ruchaud S, Lillehaug J,Raffelsberger W, Gronemeyer H, Lanotte M 1999 RAR-independent RXR signaling induces t(15;17) leukemiacell maturation. EMBO J 18:7011–7018

17. Lanotte M, Martin-Thouvenin V, Najman S, Balerini P,Valensi F, Berger R 1991 NB4, a maturation inducible cellline with t(15;17) marker isolated from a human acutepromyelocytic leukemia (M3). Blood 77:1080–1086

18. Ruchaud S, Duprez E, Gendron MC, Houge G, GenieserHG, Jastorff B, Doskeland SO, Lanotte M 1994 Twodistinctly regulated events, priming and triggering, duringretinoid-induced maturation and resistance of NB4 pro-myelocytic leukemia cell line. Proc Natl Acad Sci USA91:8428–8432

19. Duprez E, Benoit G, Flexor M, Lillehaug JR, Lanotte M2000 A mutated PML/RARA found in the retinoid matu-ration resistant NB4 subclone, NB4–R2, blocks RARAand wild-type PML/RARA transcriptional activities. Leu-kemia 14:255–261

20. de The H, Lavau C, Marchio A, Chomienne C, Degos L,Dejean A 1991 The PML-RAR a fusion mRNA generatedby the t(15;17) translocation in acute promyelocytic leu-kemia encodes a functionally altered RAR. Cell 66:675–684

21. Wang ZG, Delva L, Gaboli M, Rivi R, Giorgio M, Cordon-Cardo C, Grosveld F, Pandolfi PP 1998 Role of PML incell growth and the retinoic acid pathway. Science 279:1547–1551

22. Matera AG 1999 Nuclear bodies: multifaceted subdo-mains of the interchromatin space. Trends Cell Biol9:302–309

23. Seeler JS, Dejean A 1999 The PML nuclear bodies: ac-tors or extras? Curr Opin Genet Dev 9:362–367

24. Grignani F, Fagioli M, Alcalay M, Longo L, Pandolfi PP,Donti E, Biondi A, Lo Coco F, Grignani F, Pelicci PG 1994Acute promyelocytic leukemia: from genetics to treat-ment. Blood 83:10–25

25. Melnick A, Licht JD 1999 Deconstructing a disease:RARa, its fusion partners, and their roles in the patho-genesis of acute promyelocytic leukemia. Blood 93:3167–3215

MOL ENDO · 2001 Vol. 15 No. 71168



26. Minucci S, Pelicci PG 1999 Retinoid receptors in healthand disease: co-regulators and the chromatin connec-tion. Semin Cell Dev Biol 10:215–225

27. Rowan S, Fisher DE 1997 Mechanisms of apoptotic celldeath. Leukemia 11:457–465

28. Vaux DL, Korsmeyer SJ 1999 Cell death in development.Cell 96:245–254

29. McCubrey JA, May WS, Duronio V, Mufson A 2000Serine/threonine phosphorylation in cytokine signaltransduction. Leukemia 14:9–21

30. Rich T, Watson CJ, Wyllie A 1999 Apoptosis: the germsof death. Nat Cell Biol 1:E69–71

31. Quignon F, De Bels F, Koken M, Feunteun J, AmeisenJC, de The H 1998 PML induces a novel caspase-independent death process. Nat Genet 20:259–265

32. Wang ZG, Ruggero D, Ronchetti S, Zhong S, Gaboli M,Rivi R, Pandolfi PP 1998 PML is essential for multipleapoptotic pathways. Nat Genet 20:266–272

33. Zhong S, Salomoni P, Ronchetti S, Guo A, Ruggero D,Pandolfi PP 2000 Promyelocytic leukemia protein (PML)and Daxx participate in a novel nuclear pathway forapoptosis. J Exp Med 191:631–640

34. Grignani F, Ferrucci PF, Testa U, Talamo G, Fagioli M,Alcalay M, Mencarelli A, Peschle C, Nicoletti I, Pelicci PG1993 The acute promyelocytic leukemia-specific PML-RAR a fusion protein inhibits differentiation and pro-motes survival of myeloid precursor cells. Cell 74:423–431

35. Fu S, Consoli U, Hanania EG, Zu Z, Claxton DF, AndreeffM, Deisseroth AB 1995 PML/RARalpha, a fusion proteinin acute promyelocytic leukemia, prevents growth factorwithdrawal-induced apoptosis in TF-1 cells. Clin CancerRes 1:583–590

36. Nason-Burchenal K, Takle G, Pace U, Flynn S, AllopennaJ, Martin P, George ST, Goldberg AR, Dmitrovsky E 1998Targeting the PML/RAR a translocation product triggersapoptosis in promyelocytic leukemia cells. Oncogene17:1759–1768

37. Delia D, Aiello A, Lombardi L, Pelicci PG, Grignani F,Formelli F, Menard S, Costa A, Veronesi U, et al 1993N-(4-hydroxyphenyl)retinamide induces apoptosis ofmalignant hemopoietic cell lines including those unre-sponsive to retinoic acid. Cancer Res 53:6036–6041

38. Bruel A, Benoit G, De Nay D, Brown S, Lanotte M 1995Distinct apoptotic responses in maturation sensitive andresistant t(15;17) acute promyelocytic leukemia NB4cells. 9-cis retinoic acid induces apoptosis independentof maturation and Bcl-2 expression. Leukemia9:1173–1184

39. Monczak Y, Trudel M, Lamph WW, Miller WH 1997 In-duction of apoptosis without differentiation by retinoicacid in PLB-985 cells requires the activation of both RARand RXR. Blood 90:3345–3355

40. Tucker KA, Lilly MB, Heck L, Rado TA 1987 Character-ization of a new human diploid myeloid leukemia cell line(PLB-985) with granulocytic and monocytic differentiat-ing capacity. Blood 70:372–378

41. Sun SY, Yue P, Chandraratna RA, Tesfaigzi Y, Hong WK,Lotan R 2000 Dual mechanisms of action of the retinoidCD437: nuclear retinoic acid receptor-mediated sup-pression of squamous differentiation and receptor-independent induction of apoptosis in UMSCC22Bhuman head and neck squamous cell carcinoma cells.Mol Pharmacol 58:508–514

42. Li H, Kolluri SK, Gu J, Dawson MI, Cao X, Hobbs PD, LinB, Chen G, Lu J, Lin F, Xie Z, Fontana JA, Reed JC,Zhang X 2000 Cytochrome c release and apoptosis in-

duced by mitochondrial targeting of nuclear orphan re-ceptor TR3. Science 289:1159–1164

43. Formelli F, Barua AB, Olson JA 1996 Bioactivities ofN-(4-hydroxyphenyl) retinamide and retinoyl b-glucuro-nide. FASEB J 10:1014–1024

44. Westin S, Kurokawa R, Nolte RT, Wisely GB, McInerneyEM, Rose DW, Milburn MV, Rosenfeld MG, Glass CK1998 Interactions controlling the assembly of nuclear-receptor heterodimers and co-activators. Nature 395:199–202

45. Kurokawa R, Soderstrom M, Horlein A, Halachmi S,Brown M, Rosenfeld MG, Glass CK 1995 Polarity-spe-cific activities of retinoic acid receptors determined by aco-repressor. Nature 377:451–454

46. Li C, Schwabe JWR, Banayo E, Evans RM 1997 Coex-pression of nuclear receptor partners increases their sol-ubility and biological activities. Proc Natl Acad Sci USA94:2278–2283

47. Chen ZP, Shemshedini L, Durand B, Noy N, Chambon P,Gronemeyer H 1994 Pure and functionally homogeneousrecombinant retinoid X receptor. J Biol Chem 269:25770–25776

48. Roy B, Taneja R, Chambon P 1995 Synergistic activationof retinoic acid (RA)-responsive genes and induction ofembryonal carcinoma cell differentiation by an RA recep-tor alpha (RAR a)-, RAR b-, or RAR g-selective ligand incombination with a retinoid X receptor-specific ligand.Mol Cell Biol 15:6481–6487

49. Clifford J, Chiba H, Sobieszczuk D, Metzger D, ChambonP 1996 RXRa-null F9 embryonal carcinoma cells areresistant to the differentiation, anti-proliferative andapoptotic effects of retinoids. EMBO J 15:4142–4155

50. Mascrez B, Mark M, Dierich A, Ghyselinck NB, Kastner P,Chambon P 1998 The RXRalpha ligand-dependent acti-vation function 2 (AF-2) is important for mouse develop-ment. Development 125:4691–4707

51. Chiba H, Clifford J, Metzger D, Chambon P 1997 Specificand redundant functions of retinoid X Receptor/Retinoicacid receptor heterodimers in differentiation, prolifera-tion, and apoptosis of F9 embryonal carcinoma cells.J Cell Biol 139:735–747

52. Nagy L, Thomazy VA, Shipley GL, Fesus L, Lamph W,Heyman RA, Chandraratna RA, Davies PJ 1995 Activa-tion of retinoid X receptors induces apoptosis in HL-60cell lines. Mol Cell Biol 15:3540–3551

53. Solomin L, Johansson CB, Zetterstrom RH, BissonnetteRP, Heyman RA, Olson L, Lendahl U, Frisen J, PerlmannT 1998 Retinoid-X receptor signalling in the developingspinal cord. Nature 395:398–402

54. de Urquiza AM, Liu S, Sjoberg M, Zetterstrom RH, Grif-fiths W, Sjovall J, Perlmann T 2000 Docosahexaenoicacid, a ligand for the retinoid X receptor in mouse brain.Science 290:2140–2144

55. Miller SA, Dykes DD, Polesky HF 1988 A simple saltingout procedure for extracting DNA from human nucleatedcells. Nucleic Acids Res 16:1215

56. Segal-Bendirdjian E, Jacquemin-Sablon A 1995 Cisplatinresistance in a murine leukemia cell line is associatedwith a defective apoptotic process. Exp Cell Res 218:201–212

57. Minucci S, Maccarana M, Cioce M, De Luca P, GelmettiV, Segalla S, Di Croce L, Giavara S, Matteucci C, GobbiA, Bianchini A, Colombo E, Schiavoni I, Badaracco G, HuX, Lazar MA, Landsberger N, Nervi C, Pelicci PG 2000Oligomerization of RAR and AML1 transcription factorsas a novel mechanism of oncogenic activation. Mol Cell5:811–820




Autonomous Rexinoid Death Signaling Is Suppressed by Converging Signaling Pathways in Immature...

Documents

Transcript of Autonomous Rexinoid Death Signaling Is Suppressed by Converging Signaling Pathways in Immature...