IntroductionDuring development of the nervous system, a wide range ofcellular processes such as proliferation, survival,differentiation, synaptogenesis and maturation are driven bymultiple extracellular growth factors known as neurotrophicfactors. Their effects are usually mediated by membranereceptors belonging to the receptor with tyrosine kinaseactivity (RTK) family (Blume-Jensen and Hunter, 2001;Manning et al., 2002). Generally, binding of the ligand to theRTK extracellular region leads to receptor dimerizationfollowed by activation of its intracellular PTK domain byconformational changes of the catalytic loop associated withphosphorylation of key tyrosine residues (Schlessinger, 2000).Phosphorylated tyrosine residues can bind different adaptorsubstrate proteins that subsequently trigger various signalingcascades such as those of MAP kinase, PI 3-kinase or PLC�.The nature and degree of activation of these pathways largely

depends on the involved RTK and on the cell type in which itis expressed. This variability could explain the broad range ofsubsequent neural cellular responses (Bibel and Barde, 2000;Huang and Reichardt, 2001; Huang and Reichardt, 2003;Kaplan and Miller, 2000; Schlessinger, 2000). Theestablishment of the rat pheochromocytoma cell line PC12 hasconsiderably facilitated the study of action of someneurotrophic factors (Greene and Tischler, 1976). These cellscan either differentiate from chromaffin-like cells tosympathetic neuron-like cells when treated with nerve growthfactor (NGF) or proliferate upon epidermal growth factor(EGF) stimulation (Huff et al., 1981), although thecorresponding receptors all belong to the RTK family.

Anaplastic lymphoma kinase (ALK) is a novel RTKbelonging to the insulin receptor subfamily. ALK has beencharacterized in humans, mouse and, more recently, inDrosophila (Iwahara et al., 1997; Loren et al., 2001; Morris et

5811

Anaplastic lymphoma kinase (ALK) is a receptor tyrosinekinase essentially and transiently expressed in specificareas of the developing central and peripheral nervoussystems. We previously demonstrated that a membrane-bound and constitutively active form of the ALK proteintyrosine kinase (PTK) domain induced the neuron-likedifferentiation of PC12 cells through specific activation ofthe mitogen-activated protein kinase (MAP kinase)pathway. Its PTK domain had been originally identified ina nucleo-cytosolic and constitutively active transformingprotein, NPM-ALK. Downstream targets involved inoncogenic proliferation and survival processes have beenproposed to include phospholipase C�� (PLC��),phosphoinositide 3-kinase (PI 3-kinase)/AKT, STAT 3/5and Src. We therefore postulated that activation of specificsignaling pathways leading to differentiation orproliferation can be differently controlled depending on thesubcellular localization of ALK PTK domain. To increaseknowledge of its physiological role in the nervous system,we focused in the present study on the influence of itssubcellular localization on neuronal differentiation. Toachieve this goal, we characterized biological responses andtransduction pathways in PC12 cells elicited by variousconstructs encoding membrane-bound (through

transmembrane or myristyl sequences) or cytosolic ALK-derived proteins. In order to control the activation of theirPTK domain, we used an inducible dimerization system.Here, we demonstrate that membrane attachment of theALK PTK domain, in PC12 cells, is crucial for initiation ofneurite outgrowth and proliferation arrest through adecrease of DNA synthesis. Furthermore, we show that thisdifferentiation process relies on specific and sustainedactivation of ERK 1/2 proteins. By contrast, activation ofthe cytosolic form of this domain fails to induce MAPkinase activation and cell differentiation but promotes a PI3-kinase/AKT-dependant PC12 cell proliferation. Thesedata indicate that subcellular localization of the ALK PTKdomain was a determinant for the control and specificity ofdownstream transduction cascades and was crucial fordeciding the fate to which the neuronal cell will becommitted.

Supplementary material available online athttp://jcs.biologists.org/cgi/content/full/119/24/5811/DC1

Key words: ALK, FK506-binding protein, PC12 cells, MAP kinase,Neurite outgrowth, Cell proliferation

Summary

Role of the subcellular localization of ALK tyrosinekinase domain in neuronal differentiation of PC12cellsJean Y. Gouzi, Christel Moog-Lutz, Marc Vigny* and Nicole Brunet-de CarvalhoINSERM, U706, Institut du Fer à Moulin, 17 rue du Fer à Moulin; and UPMC, 4 Place Jussein, Paris, F-75005, France*Author for correspondence (e-mail: vigny@fer-a-moulin.inserm.fr)

Accepted 15 September 2005Journal of Cell Science 118, 5811-5823 Published by The Company of Biologists 2005doi:10.1242/jcs.02695

Research Article

Jour

nal o

f Cel

l Sci

ence

5812

al., 1997). ALK possesses a classical structure shared withother RTKs: a large extracellular domain, a singletransmembrane domain and an intracellular domain containingthe tyrosine kinase catalytic site. Recently, pleiotrophin (PTN)and midkine (MK) have been proposed as potential ligands ofALK (Stoica et al., 2001; Stoica et al., 2002), although severalstudies do not confirm this hypothesis (Dirks et al., 2002;Miyake et al., 2002; Moog-Lutz et al., 2005; Motegi et al.,2004). However, a recent report suggests that a truncated formof PTN (and not the full-length PTN) could activate ALK (Luet al., 2005). In addition, the protein Jelly Belly (Jeb) has beenidentified as the ligand of Drosophila ALK (Englund et al.,2003; Lee et al., 2003) and Jeb is distinct from the Drosophilahomologs of PTN and MK, Miple 1 and 2 (Lee et al., 2003).Finally, no obvious vertebrate homolog has been identified forJeb in the sequence database.

In situ hybridization and northern blot studies in mammalsrevealed that alk transcripts were essentially and transientlyexpressed in specific regions of the developing central andperipheral nervous systems (Iwahara et al., 1997; Morris et al.,1997). In Drosophila, regulated DAlk mRNA and DAlk proteinexpressions were also described in the developing brain andventral nerve cord (Loren et al., 2001). This distributionstrongly suggested that ALK could play an important role inthe normal development and function of the nervous system.In agreement with this hypothesis, we previously demonstratedthat a membrane-bound and constitutively active form of theALK PTK domain (a Fc-ALK chimera in which a Fc fragmentof mouse IgG is substituted with the ALK extracellulardomain) induced the neuron-like differentiation of PC12 cells(Souttou et al., 2001). Analysis of the signaling pathwaysinvolved in this process pointed to an essential role of the MAPkinase cascade. Recently, we and others (Moog-Lutz et al.,2005; Motegi et al., 2004) reported that ALK activationtriggered by specific monoclonal antibodies promoted neuriteoutgrowth of neuronal cell lines through activation of thispathway.

The intracellular domain of ALK had been previouslyidentified in some anaplastic large cell lymphomas (ALCL)resulting from abnormal chromosomal translocations that leadto the expression of oncogenic fusion kinases (reviewed byPulford et al., 2004). The most characteristic is the nucleo-cytosolic NPM-ALK protein, which contains the completeALK intracellular domain fused to the N-terminal part ofnucleophosmin [NPM (Morris et al., 1994)]. NPMoligomerization leads to the constitutive activation of the ALKkinase domain (Bischof et al., 1997) and PLC�, PI 3-kinase/AKT, STAT 3/5 and Src have been proposed asdownstream targets of NPM-ALK involved in oncogenicproliferation and survival processes. It had also been reportedthat cytoplasmic localization, but not nuclear localization, ofNPM-ALK was required for its transforming activity (Bai etal., 1998; Bai et al., 2000; Bischof et al., 1997; Fujimoto et al.,1996; Nieborowska-Skorska et al., 2001; Slupianek et al.,2001; Zamo et al., 2002).

On the basis of the above-mentioned previous results, wepostulated that activation of specific signaling pathwaysleading to differentiation or proliferation can be differentlycontrolled depending on the subcellular localization of theALK kinase domain. It is noteworthy that subcellularlocalization of a PTK domain as a factor that determines the

signal transduction specificity has been poorly documented sofar and could have a profound impact on cellular processes(Schlessinger, 2000). Here, taking account of the cellularexpression of ALK during development, we particularlystudied the role of the subcellular localization of its PTKdomain in neuronal differentiation. To achieve this goal, weexpressed different constructs encoding membrane-bound orcytosolic ALK-derived proteins in PC12 cells. In order tocontrol the activation of their intracellular PTK domain, weused an inducible dimerization system. Here, we show thatmembrane attachment of the ALK PTK domain is crucial forinitiation of neurite outgrowth and proliferation arrest of P12cells through a decrease of DNA synthesis. Furthermore, weshow that this differentiation process relies on specific andsustained activation of the MAP kinase pathway. By contrast,activation of the cytosolic form of this domain fails to induceMAP kinase activation and cell differentiation but promotes aPI 3-kinase/AKT-dependent PC12 cell proliferation.

Materials and MethodsPlasmid constructionsThe pC4M-Fv2E plasmid (ARIAD Pharmaceuticals) encodes a proteinthat we named myr-FH (Fig. 1B) composed of an N-terminalmyristoylation signal, two tandem Fv domains (derived from theFKBP domain) followed by a C-terminal hemagglutinin (HA) tag.The cDNA encoding the ALK intracellular domain (1058-1620 aa)was cloned by PCR amplification of the human ALK cDNA plasmid(Souttou et al., 2001) using the primers forward 5�-GGCCTC-TAGAGTGTACCGCCGGAAGCACCAG-3� and reverse 5�-CCG-GTCTAGAGGGCCCAGGCTGGTTCATGCT-3�, then sequencedand subcloned into the XbaI site of the pC4M-Fv2E vector to make aconstruct expressing the myrIA-FH fusion protein (Fig. 1B). Toconstruct the pC4-Fv2E vector, we inserted the XbaI-XhoI Fv fragmentof the pC4M-Fv2E plasmid into the XbaI site of the pC4-Fv1E plasmid(ARIAD Pharmaceuticals). The cDNA coding for the ALKintracellular domain was cloned into the XbaI site of this vector tomake a construct expressing the cytoIA-FH fusion protein (Fig. 1B).We performed directed mutagenesis on the ALK cDNA to create aSmaI site in order to remove the extracellular domain of the protein(42-1026 aa) and a XbaI site (replacing the stop codon). This cDNAwas sequenced and inserted into the EcoRI-XbaI sites of the pC4M-Fv2E vector, yielding expression of the tmbIA-FH protein (Fig. 1B).By an exchange of KpnI fragments between this mutated cDNA andthe wild-type ALK cDNA (821-5619 bp), we constructed the vectorexpressing the ALK-FH protein (Fig. 1B). A construct encoding akinase-defective form of the full-length human ALK, in which theinvariant lysine residue located in the ATP-binding portion of thecatalytic domain was changed to arginine, was previously inserted inthe mammalian expression vector pcDNA 3.1 (Souttou et al., 2001).From this construct, we generated the different constructs encodingthe kinase-defective forms of ALK-FH, myrIA-FH and cytoIA-FHproteins. In addition, inserts encoding cytoIA-FH and myrIA-FH weresubcloned into the pCBC vector (Moog-Lutz et al., 2005) usingdirected mutagenesis and constructs were verified by DNAsequencing.

Cell culture and transfectionRat pheochromocytoma (PC12) cells (kindly provided by T. Galli,Institut Jacques Monod, Paris, France) were grown in a completemedium (RPMI 1640; Gibco Life technologies) supplemented with10% horse serum (HS) and 5% fetal calf serum (FCS) at 37°C in anatmosphere containing 5% CO2. Cells were cultured in flasks orplastic dishes coated with collagen (1 �g/cm2). PC12 cells were

Journal of Cell Science 118 (24)

Jour

nal o

f Cel

l Sci

ence

5813ALK and PC12 cell differentiation

electroporated as previously described (Moog-Lutz et al., 2005) andimmediately transferred to fresh complete medium for 1 day. Then,unless otherwise specified, cells were serum starved in a low-serummedium (1% HS without FCS) for 16 hours before the addition of thedimerizer. For experiments with various kinase inhibitors, cells werepre-treated with the agents for 30-60 minutes before the addition ofthe dimerizer.

Human 293 embryonic kidney (HEK 293) cells were cultivated andtransfected using the calcium phosphate precipitation method. Cellsstably expressing cytoIA-FH and myrIA-FH were established usingpCBC vectors as previously described (Moog-Lutz et al., 2005).

Reagents and antibodiesThe dimerizer (AP20187) was obtained from ARIADPharmaceuticals. U0126 was purchased from Cell SignalingTechnology. LY294002 and wortmannin were from Calbiochem. 5-Bromodeoxyuridine (BrdU) and mouse anti-BrdU FITC-conjugatedantibody were from Roche Diagnostics. Antibodies and their sourceswere as follows: mouse monoclonal anti-HA-tag antibody (12CA5;generous gift of Y. Frobert, CEA, Saclay, France); rat anti-HA-tagantibody (3F10; Roche Diagnostics); mouse anti-PY antibody (4G10)and rabbit anti-ERK antibody (Upstate Biotechnology); mouse anti-P-ERK antibody (Sigma); rabbit anti-AKT phosphoserine-473

antibody and anti-AKT antibody (Cell Signaling Technology); mouseanti-transferrin receptor antibody (H 68.4; Zymed); and rabbit anti-stathmin antibody (STC3; kindly provided by A. Sobel, INSERMU706, Paris, France).

Immunocytochemistry and confocal microscopyPC12 cells transiently transfected with the different constructs weregrown on collagen-coated plastic dishes for 2 days in a completemedium, fixed for 15 minutes at room temperature with pre-warmed2% formaldhehyde/30 mM sucrose and washed three times with PBS.Cells were then permeabilized in 0.5% PBS-Triton X-100 for 5minutes and washed with 0.1 M PBS-glycine for 15 minutes. After 1hour of blocking in PBS containing 1.5% BSA, cells were incubatedin the same buffer with monoclonal mouse anti-HA-tag (12CA5)antibody (0.9 �g/ml) to visualize cells expressing ALK-FH-derivedproteins. The cells were then washed five times with PBS before andafter incubation with anti-mouse IgG Alexa Fluor 488-conjugatedsecondary antibody (Molecular Probes). In order to visualize nuclei,cells were incubated with 25 nM Sytox Orange nucleic acid stain(Molecular Probes) for 2 minutes and washed with water. The cellswere then mounted in Mowiol 4-88 (Calbiochem). Confocal lasermicroscopy was performed using a TCS SP2 confocal microscope(Leica). Images were assembled using Adobe Photoshop software.

Subcellular fractionationElectroporated PC12 cells grown for two days were rapidly washedwith cold PBS and resuspended in 200 �l hypotonic buffer [10 mMTris-HCl, pH 7.5, 5 mM MgCl2, 2 mM EGTA, 25 mM �-glycerolphosphate, 5 mM sodium fluoride, 2 mM sodium pyrophosphate, 1mM sodium orthovanadate and protease inhibitor mixture ‘complete’(Roche)]. Cells were disrupted by three cycles of freezing followedby thawing, after each of which a 25-gauge needle was used to disruptthe material further. Whole-cell lysates were spun at 800 g for 10minutes at 4°C to remove cell nuclei and crude debris, and thesupernatants were submitted to ultracentrifugation at 100,000 g for 45minutes at 4°C in a TLA 100.1 rotor (Optima Beckman) to generatesupernatant (cytosol) and pellet (membrane) fractions. Membranefractions were then resuspended in a RIPA buffer and clarified bycentrifugation at 21,000 g for 10 minutes at 4°C.

Neurite outgrowth assayElectroporated PC12 cells were treated at day 2 with the dimerizer.At day 4, cells were fixed and processed for immunofluorescence asdescribed above. Cells were first incubated with monoclonal mouseanti-HA-tag (12CA5) antibody (0.9 �g/ml) and then with anti-mouseIgG FITC-conjugated secondary antibody (Jackson ImmunoResearchLaboratories) to visualize cells expressing ALK-FH-derived proteins.Then cells were mounted in Mowiol 4-88 supplemented with Hoechst33258 nucleic acid stain (0.5 �g/ml, Molecular Probes) in order tovisualize nuclei. Conventional fluorescence microscopy wasperformed on a Leica microscope equipped with a MicroMax CCDcamera (Princeton Instruments). Images were assembled using AdobePhotoshop software. The Metamorph software was used (RoperScientific) for quantification, and 100 transfected cells were countedand cells bearing neurites longer than twice the diameter of the cellbody were scored as differentiated. The experiments were performedin triplicate.

BrdU-incorporation assayElectroporated PC12 cells were treated with the dimerizer for only 12hours; 6 hours after dimerizer application, cells were incubated withBrdU (10 �M) for 6 hours, then fixed and processed forimmunofluorescence as described above. After DNA denaturation

A

2x FKBPv

"dimerizer"

cytoplasm

ALK-FH

PP

HA-tag

TK TK TKTK

ALK-FH220 kDa

tmbIA-FH97 kDa

myrIA-FH89 kDa

cytoIA-FH87 kDa

myr-FH23 kDa

PTKECD TM HA

2x FKBPv

PTKTM

myrPTK

PTK

myr

B

Fig. 1. ALK-FH-derived proteins and the inducible dimerizationsystem. (A) Schematic representation of the induced dimerization ofALK-FH protein by the synthetic lipophylic and bifunctionaldimerizer AP20187. (B) Schematic representation of the full-lengthALK-FH protein and its intracellular domain as membrane-bound(transmembrane, tmbIA-FH; or myristoylated, myrIA-FH) orcytosolic (cytoIA-FH) proteins. ECD, extracellular domain; TM,transmembrane domain; PTK, protein tyrosine kinase domain; 2�FKBPv, tandem copies of modified form of the intracellular FK506-binding protein; HA, hemagglutinin tag; myr, myristoylationsequence from Src.

Jour

nal o

f Cel

l Sci

ence

5814

with 2N HCl for 30 minutes and neutralization with a borate buffer(0.1 M pH 8.5; three washes), cells were first incubated with mouseanti-BrdU FITC-conjugated (2 �g/ml) and rat anti-HA-tag (3F10; 0.5�g/ml) antibodies and then with anti-rat IgG TRITC-conjugatedsecondary antibody (Jackson ImmunoResearch Laboratories) tovisualize cells expressing ALK-FH-derived proteins and BrdU-positive cells. The cells were then mounted and conventionalfluorescence microscopy and quantification were performed asdescribed above. 100 transfected cells were counted and cells withpositive BrdU staining were scored as having undergone DNAreplication during the time of labeling. The experiments wereperformed in triplicate.

Cell lysates and immunoblotting analysisCell extracts were prepared by lysing the cells in a RIPA buffer [10mM NaPi buffer, pH 7.8, 60 mM NaCl, 1% Triton X-100, 0.5%deoxycholic acid, 0.1% SDS, 10% glycerol, 25 mM �-glycerolphosphate, 50 mM sodium fluoride, 2 mM sodium pyrophosphate,1 mM sodium orthovanadate and protease inhibitor mixture‘complete’ (Roche)] and analysed by immunoblotting as previouslydescribed (Moog-Lutz et al., 2005). Bound proteins were visualizedusing the ECL system (Amersham Bioscience) or the OdysseyImaging System (LI-COR bioscience). This latter was also used forquantification.

ResultsConstructs encoding ALK-derived proteinsWe used an inducible dimerization system recently applied tothe study of RTKs such as the platelet-derived growth factor(PDGF) and insulin receptors (Yang et al., 1998) or thehepatocyte growth factor (HGF) receptor (Boccaccio et al.,2002). This method is based on the use of a small, synthetic,lipophylic and bifunctional ligand known as the dimerizer(AP20187). The dimerizer binds in trans with high affinity toa modified form of the intracellular FK506-binding protein(FKBPv) and has no ability to bind endogenous FKBP. Itallows the dimerization of intracellular functional domains ofvarious proteins that have been fused to the FKBP modules(Clackson et al., 1998) as represented in Fig. 1A. Thus, weengineered several plasmid constructs encoding chimericproteins in which the C-terminus of the human ALKintracellular domain was fused to two tandem copies ofFKBPv (Fig. 1B) and a HA tag. These constructs allowed theexpression of the full-length receptor (ALK-FH), of the entireALK intracellular domain linked to the membrane eitherthrough the ALK transmembrane sequence (tmbIA-FH) orthe Src myristoyl sequence (myrIA-FH) and finally of thecomplete ALK intracellular domain without any membraneattachment sequence (cytoIA-FH). Several control constructswere also generated. One (myr-FH) encodes the twoFKBPv modules bound to the membrane after deletion of theentire ALK intracellular domain from the myrIA-FHconstruct. The other corresponds to kinase-defectivemutants of the various ALK constructs (schematicrepresentations not shown) in which the invariant lysineresidue located in the ATP-binding site of the catalyticdomain was changed to arginine (see Materials and Methods).This point mutation has been previously shown to inhibitcompletely the transforming capability of the NPM-ALKoncogenic protein (Bischof et al., 1997) and the PC12 neuriteoutgrowth induced by the expression of the constitutive active

form of the ALK tyrosine kinase domain (Souttou et al.,2001).

Induced kinase activation of full-length ALK resulted inPC12 cell neurite extension and ERK 1/2 activationTo test whether the inducible dimerization system we used wasfunctional with the full-length ALK, we first transfected PC12cells with the construct encoding ALK-FH and assayed forneurite outgrowth in the presence of increasing doses ofdimerizer (0-1 �M). As shown in Fig. 2A, the dimerizerincreased the proportion of cells expressing ALK-FHextending neurites in a dose-dependent manner from a basallevel of about 20% in the absence of dimerizer to a maximumlevel of about 50% obtained with the dimerizer at 20 nM. Thisoptimal concentration was used for the subsequentexperiments. Cells transfected with the ALK-FH kinase-defective mutant (dALK-FH; Fig. 2B) exhibited a basal levelof only 5% of transfected cells bearing neurites, significantlylower compared with the 15-20% obtained with ALK-FH-expressing cells and similar to that observed with control cellstransfected with the empty vector (not shown). In this case,addition of the dimerizer did not induce any differentiation.These data indicated that both the basal and induced neuriteoutgrowth described specifically relied on the kinase activityof the ALK-FH protein and furthermore indicated that thepresence of the FKBPv modules and/or the dimerizer did notinterfere with neurite extension.

To demonstrate that ALK-FH was actually activated wedirectly assayed its phosphorylation level by western blotanalysis of whole-cell lysates after various periods of dimerizertreatment (Fig. 2C). The anti-HA-tag antibody revealed twomajor species of ALK-FH as previously reported in rat andmouse brain (Iwahara et al., 1997; Morris et al., 1997) and inHEK 293 or NIH3T3 cells stably transfected with the ALKreceptor (Moog-Lutz et al., 2005; Motegi et al., 2004). In theabsence of dimerizer, both species of ALK-FH (220 kDa and140 kDa) exhibited a basal level of phosphorylation that wasstrongly increased upon dimerizer treatment to a level thatreached a maximum within 5 minutes and then remainedconstant for at least 30 minutes. No ALK phosphorylation wasobserved in cells expressing the kinase-defective mutant(dALK-FH), thus confirming that both basal and inducedphosphorylation were a result of its intrinsic kinase activity. Wethen assayed for ERK 1/2 phosphorylation (Fig. 2D) sinceMAP kinase activation had been previously shown to beessential in PC12 cell differentiation triggered by theexpression of the constitutively active form of ALK (Souttouet al., 2001). Immunoblotting with anti-phospho-ERK 1/2antibodies revealed that the ERK 1/2 proteins were activatedin the presence of dimerizer, reaching a maximumphosphorylation within 10 minutes. We further showed thattheir phosphorylation was sustained in a long-lasting mannersince it remained elevated for at least 1 hour and even after 24hours (data not shown). In addition, no ERK 1/2 activation wasdetected in cells expressing the kinase-defective mutant dALK-FH, which demonstrates that basal and induced ERKactivations were dependent on ALK kinase activity. Altogether,these results were in agreement with those obtained previouslywith the constitutively active Fc-ALK chimera, clearlyestablishing that the ALK-FH protein dimerized by its

Journal of Cell Science 118 (24)

Jour

nal o

f Cel

l Sci

ence

5815ALK and PC12 cell differentiation

intracellular part was functional, and therefore validating thedimerization system we used.

Normalized expression and subcellular localization ofALK-FH-derived proteinsSince the major aim of this study was to analyze the role of thesubcellular localization of ALK PTK domain in differentiationand/or proliferation of PC12 cells, we normalized theexpression of the different constructs and we analyzed thelocalization of the different ALK-FH-derived proteins. Thus,PC12 cells were initially transfected by electroporation withequal amounts of the different ALK-derived constructs (25 �gDNA for 5�106 cells). Immunoblotting analysis using the anti-HA-tag antibody revealed that ALK-derived proteins displaysimilar apparent molecular weight and that their expressionlevels were highly heterogeneous (data not shown). myrIA-FH

protein displayed the highest expression level whereas, incomparison, tmbIA-FH was weakly expressed. On the basis ofthe expression of the tmbIA-FH protein, we performed celltransfections with a range of decreasing concentrations ofDNA constructs encoding myrIA-FH and cytoIA-FH proteins,and established those required to obtain approximatelyidentical protein expression levels as shown in Fig. 3A.

The second step was to verify that all these proteins werecorrectly located in the cell by performing confocalmicroscopy analysis (Fig. 3B). tmbIA-FH and myrIA-FHproteins exhibited a potent enrichment at the periphery of thetransfected cells, suggesting a plasma membrane localization.A much less intense intracellular staining could also beobserved, probably corresponding to intracellular membranecompartments. By contrast, the cytoIA-FH protein showedessentially a uniform distribution throughout the cytoplasm andwithin the cell nucleus. Subcellular fractionation mainly

Fig. 2. Induced kinase activation of the full-length receptor ALK results in neurite extension and ERK 1/2 activation. (A) Dimerizer dose-dependent effect on neurite outgrowth induced by the activation of ALK-FH protein. PC12 cells were electroporated with the ALK-FHconstruct (25 �g), cultured overnight in a 1% horse serum medium and treated without or with increasing concentrations of dimerizer asindicated for 2 days. Then immunofluorescence assay was performed using the anti-HA-tag antibody (12CA5); 100 transfected cells werecounted and cells bearing neurites longer than twice the diameter of the cell body were scored as differentiated. The experiment was performedin triplicate and values are expressed as the mean ± s.e.m. (%). (B) Neurite outgrowth induced by ALK-FH dimerization is a result of its ownkinase activity. PC12 cells were electroporated with the indicated constructs (25 �g) and then submitted to the same protocol as described in(A). Cells were treated with a dimerizer concentration of 20 nM. The experiment was performed in triplicate and values are expressed as themean ± s.e.m. (%). Time course of (C) ALK-FH and (D) ERK 1/2 phosphorylation following dimerizer treatment. ALK-FH-transfected PC12cells were incubated with the dimerizer (20 nM) for the indicated periods. As controls, non-transfected cells (Mock) or cells transfected withthe dALK-FH mutant were treated with the dimerizer during 30 minutes. Cells were lysed in a RIPA buffer as described in the Materials andMethods and the lysates (10 �g) were submitted to western blot analysis using the anti-PY (4G10) or the anti-P-ERK 1/2 antibody (upperpanels) and then reprobed with the indicated antibodies (lower panels). The P-ERK 2 densitometry quantification was performed using theOdyssey Imaging System on three independent experiments. Statistical analyses were carried out by Student’s t-test (***P<0.005). Asterisk inC indicates a 160 kDa protein that was not revealed by the anti-HA-tag antibody and therefore not related to ALK.

Jour

nal o

f Cel

l Sci

ence

5816

confirmed these results and also revealed that, compared withthe pure cytosolic marker stathmin, a minor but significant partof the cytoIA-FH protein was associated with the membranefraction (Fig. S1, supplementary material).

Membrane attachment of ALK intracellular domain isrequired for the induction of neurite outgrowth of PC12cells and MAP kinase activationWe investigated the phenotypic effect induced after two daysof dimerizer treatment in cells expressing ALK-FH-derivedproteins (Fig. 4A). In the absence of dimerizer, control cellstransfected with the myr-FH construct showed essentially around shape and undifferentiated phenotype. By contrast, only

cells transfected with the constructs encoding the membrane-bound proteins (tmbIA-FH and myrIA-FH) exhibited asignificant level of basal differentiation (25% and 32%respectively, Fig. 4B), displaying short and thin neurites (Fig.4A). However, when treated with the dimerizer, the percentageof these cells extending neurites was greatly increased (55%and 75%, respectively) and they displayed a much larger andspread cell body. Moreover, they harbored thick neuriteextensions (reaching several folds the cell body size), whichcould be visible as early as the 12th hour of dimerizertreatment. By contrast, even in the presence of dimerizer, fewerthan 10% of the cells expressing the cytosolic form of the ALKintracellular domain (cytoIA-FH) extend neurites (Fig. 4B).Thus, they essentially retained their undifferentiatedphenotype. This very weak differentiation level probablyresulted from the activation of the tiny cytoplasmic proteinfraction associated with the membrane (Fig. S4, supplementarymaterial). Control cells expressing membrane-bound FKBPvmodules alone did not trigger any neurite outgrowth.

As we previously showed that induced activation of thetransmembrane full-length receptor ALK-FH led to PC12 celldifferentiation and sustained activation of ERK 1/2 proteins(Fig. 2D), we investigated how activation of the membrane-bound or cytosolic ALK-FH-derived proteins could induce theactivation of ERK proteins. We focused our comparative studyon the two most closely related proteins: myrIA-FH andcytoIA-FH. Immunoblot analysis of whole-cell extracts faintlyrevealed a dimerizer-induced phosphorylation of myrIA-FHand cytoIA-FH (Fig. 4C). Taking into account the strongactivation of the downstream pathways induced by thedimerizer (see below), one could have expected a strongerincrease in the level of tyrosine phosphorylation of ALK-FH-derived proteins as observed for the full-length receptor (Fig.2C). The use of stable HEK 293 clones expressing the sameALK-derived proteins at a much lower level allowed us todetect a stronger increase in their dimerizer-induced tyrosinephosphorylation level, which was revealed inimmunoprecipitation experiments (Fig. S2A, supplementarymaterial). Similar HA immunoprecipitates were alsoperformed using PC12 cell lysates. We did not detect a strongerchange in tyrosine phosphorylation in the immunoprecipitatesthan in the whole-cell lysates. Thus, the basal phosphorylationof these proteins probably resulted from their spontaneousdimerization in the absence of any treatment, as suggested bythe absence of phosphorylation in kinase-defectivecorresponding proteins (Fig. S2B, supplementary material).Note that the levels of basal and induced phosphorylation ofthe cytoIA-FH protein were slightly lower than that of themyrIA-FH protein. As shown in Fig. 4C, the myrIA-FH proteininduced a significant increase of ERK 1/2 phosphorylationupon only 2 minutes of dimerizer treatment, which wassustained in a long-lasting manner as it remained constant afterat least 1 hour and even after 24 hours (not shown). By contrast,the expression of the cytosolic protein (cytoIA-FH) resulted ina weak ERK phosphorylation. This much weaker ERKactivation could be explained by the lower level ofphosphorylation of this protein. To test this hypothesis, weoverexpressed the cytoIA-FH protein to obtain a degree oftyrosine phosphorylation similar to that of the myrIA-FHprotein. In these conditions, this weak level of ERK activationwas identically observed (data not shown). Note that either

Journal of Cell Science 118 (24)

Fig. 3. Cellular expression of ALK-FH-derived proteins:normalization of the expression levels and subcellular localization.(A) Western blot analysis of the expression of ALK-FH-derivedproteins after normalization of their expression levels. PC12 cellswere electroporated with the indicated quantities of the DNAconstructs, in order to obtain comparable levels of proteinexpression. Cells were lysed after two days in a RIPA buffer asdescribed in the Materials and Methods, and the lysates (10 �g) weresubmitted to a western blot analysis using the anti-HA-tag (12CA5)monoclonal antibody. (B) Immunofluorescence confocal microscopyanalysis of the subcellular localization of ALK-FH-derived proteinsrevealed with the anti-HA-tag antibody (green) in permeabilizedPC12 cells (representative fields are shown). Cell nuclei were labeledwith the Sytox Orange nucleic acid stain (red). Bar, 10 �m.

Jour

nal o

f Cel

l Sci

ence

5817ALK and PC12 cell differentiation

basal or induced ERK phosphorylations triggered by themyrIA-FH and cytoIA-FH proteins were a result of theirintrinsic kinase activities (Fig. S2B, supplementary material).Thus, all these data provided evidence for the first time thatmembrane attachment of the ALK PTK domain was requiredfor the induction of neurite outgrowth of PC12 cells and forthe activation of MAP kinase.

Induced neurite extension of PC12 cells is dependent onERK 1/2 but not PI 3-kinaseIt had previously been shown that MAP kinase but not PI 3-kinase activation was essential to the neuron-likedifferentiation of PC12 cells expressing the constitutivelyactive Fc-ALK chimera (Souttou et al., 2001). We thereforeasked whether the sustained ERK 1/2 activation followinginduced activation of the membrane-bound ALK proteins (Fig.2B, Fig. 4B) was actually required for the induction of theneurite outgrowth process. Thus, the effect of U0126, one ofthe selective inhibitors of MEK 1/2 (MAP kinase kinase)proteins (Favata et al., 1998) was first tested on neuriteoutgrowth induced by activation of the full-length ALK-FHprotein. As shown in Fig. 5A, and as expected, U0126treatment led to a progressive inhibition of the induced neuriteextension of transfected cells in a dose-dependant manner,reaching a weak level of about 10% of these cells bearingneurites at a concentration of 25 �M of U0126 that we chosefor subsequent analysis. This concentration was actually shownto inhibit either the basal or induced ERK phosphorylation inthese cells completely (Fig. S3, supplementary material). Cellsexpressing the membrane-bound proteins tmbIA-FH or myrIA-FH were treated or not with 20 nM of dimerizer in the presenceor absence of U0126. As shown in Fig. 5B, the U0126treatment almost completely abolished both the basal anddimerizer-induced neurite outgrowth, which also reached aweak differentiation level of about 10%. By contrast,wortmannin (Fig. 5C), one of the widely used pharmacologicalinhibitors of PI 3-kinase activity (Nakanishi et al., 1992; Powiset al., 1994), did not trigger any neurite outgrowth inhibitionwhen used at the active dose of 50 nM (see below). Thus, the

neurite extension process induced by activation of themembrane-bound proteins ALK-FH, tmbIA-FH and myrIA-FH required activation of the MAP kinase pathway but not thePI 3-kinase pathway.

The differentiation process induced by activation of themembrane-bound ALK intracellular domain isconcomitant with DNA synthesis arrestIt has previously been described (Greene and Tischler, 1976;

Fig. 4. Membrane attachment of the ALK intracellular domain isrequired to induce neurite extension of PC12 cells and MAP kinaseactivation. (A) PC12 cells were electroporated with normalizedquantities of the indicated constructs, cultured overnight in a 1%horse serum medium and treated with the dimerizer (20 nM) for 2days. Then cells were fixed, permeabilized and immunofluorescenceassay was performed. Cells expressing the ALK-FH-derived proteinswere labeled with the anti-HA-tag antibody (12CA5) and cell nucleiwere colored with the nucleic acid stain Hoechst 33258. Bar,100 �m. (B) Quantification of the effect of the induced activation ofthe different proteins on neurite outgrowth, as indicated in theMaterials and Methods. The experiment was performed in triplicateand values are expressed as the mean ± s.e.m. (%). (C) Time courseof ERK 1/2 phosphorylation following dimerizer treatment. PC12cells expressing the indicated proteins were cultured overnight in a1% horse serum medium and then incubated with the dimerizer(20 nM) for the indicated periods. Cells were lysed in RIPA buffer asdescribed in the Materials and Methods and the lysates (10 �g) weresubmitted to western blot analysis using the anti-HA-tag or the anti-P-ERK antibody and then reprobed with the anti-PY (4G10) or theanti-ERK antibody.

Jour

nal o

f Cel

l Sci

ence

5818

Rudkin et al., 1989; van Grunsven et al., 1996a; van Grunsvenet al., 1996b; Yan and Ziff, 1995; Yan and Ziff, 1997) that,upon NGF treatment, PC12 cells undergo a differentiationprocess and a proliferation arrest in the G0-G1 phase of thecell cycle by decreasing their proliferation rate and DNAsynthesis. In the present study, we show that induced activationof the membrane-bound proteins tmbIA-FH and myrIA-FHleads to neurite outgrowth of PC12 cells (Fig. 4A,B).Therefore, we asked whether activation of these proteins wasaccompanied by DNA synthesis arrest. This process wasmonitored through a BrdU-incorporation assay using PC12cells grown asynchronously in the presence of serum. Asshown in Fig. 6, induced activation of tmbIA-FH and myrIA-FH proteins significantly decreased the proportion oftransfected cells incorporating BrdU during the time oflabeling, revealing an arrest of DNA synthesis. By contrast,cells expressing the myrIA-FH-related kinase-defective mutant(dmyrIA-FH) did not exhibit such an effect. This indicates thatthis DNA synthesis arrest process was specifically a result ofthe kinase activity of the membrane-bound ALK-FH-derivedproteins. Furthermore, we showed that this process indeedrequired ERK 1/2 activation since it was abolished by U0126treatment. In conclusion, our results show for the first time thatthe induced activation of these membrane-bound proteins leadsto differentiation (see above) and concomitantly to an arrest ofthe continued proliferation of transfected PC12 cells, and thatboth of these processes are controlled by the ERK pathway.

Activation of the cytosolic form of the ALK intracellulardomain induces DNA synthesis through involvement ofthe PI 3-kinase/AKT pathway As the cytosolic form of the ALK intracellular domain wasshown to fail to induce any neurite outgrowth of PC12 cells,

Journal of Cell Science 118 (24)

Fig. 5. Activation of the MAP kinase pathway, but not the PI 3-kinasepathway, is required to induce neurite extension of PC12 cells.(A) MEK inhibitor dose-dependent inhibition of neurite outgrowthinduced by the activation of the full-length ALK-FH protein. PC12cells were electroporated with ALK-FH construct (25 �g), culturedovernight in a 1% horse serum medium and treated with the dimerizer(20 nM) for two days in the presence of increasing concentrations ofthe MEK inhibitor U0126 dissolved in DMSO. As a control, cellswere treated without DMSO (No vehicle). DMSO control (0 �M)showed no adverse effects. (B,C) Inhibition of neurite outgrowth bythe MEK inhibitor U0126 but not by the PI 3-kinase inhibitorwortmannin. PC12 cells were electroporated with normalizedquantities of the indicated constructs, cultured as described above andtreated with the dimerizer (20 nM) in the presence or not of U0126(25 �M) (B) or wortmannin (50 nM) (C) for two days. All the data(A-C) were obtained following immunofluorescence detection withthe anti-HA-tag antibody (12CA5), neurite outgrowth of transfectedcells being scored as indicated in the Materials and Methods. Theexperiments were performed in triplicate and values are expressed asthe mean ± s.e.m. (%).

B

U0126 :dimerizer :

exte

ndin

g ne

urite

s (%

)T

rans

fect

ed c

ells

+ +---

+ +- -- - ++++-

tmbIA-FH myrIA-FH

90

010

30

50

70

C

Wortmannin :+ +-- + +- -

- - - - ++++tmbIA-FH myrIA-FH

90

010

30

50

70

dimerizer :

exte

ndin

g ne

urite

s (%

)T

rans

fect

ed c

ells

A

ALK-FH

0

10

20

30

40

50

60

70

0 10 20 30 40 50No vehicle

U0126 (µM)

exte

ndin

g ne

urite

s (%

)T

rans

fect

ed c

ells

Fig. 6. Neurite outgrowth of PC12 cells mediated by ALK-FH-derived protein is accompanied by arrest of DNA synthesis. Cellswere electroporated with normalized quantities of the indicatedconstructs, continually cultured in a normal serum medium (10%horse serum, 5% fetal calf serum) in order to study their normalasynchronous growth and treated with the dimerizer (20 nM) for 12hours in the presence or not of U0126 (25 �M). At 6 hours afterdimerizer application, cells were incubated with BrdU (10 �M) for 6hours of incorporation. Then immunofluorescence assay wasperformed using the anti-HA-tag (3F10) and the FITC-conjugatedanti-BrdU antibodies; 100 transfected cells were counted and cellswith positive BrdU staining were scored as having undergone DNAreplication during the time of labeling. The experiment wasperformed in triplicate and values are expressed as the mean ± s.e.m.(%). Statistical analyses were carried out by Student’s t-test(*P<0.05; **P<0.01).

tmbIA-FH myrIA-FH dmyrIA-FH

dimerizer : +- +- +-U0126 :

30

0

20

10

40

Tra

nsfe

cted

cel

ls

inco

rpor

atin

g B

rdU

(%

)

- +++- +- +-

tmbIA-FH myrIA-FH dmyrIA-FH

∗∗ ∗∗

∗∗

- -- -- + + + +

Jour

nal o

f Cel

l Sci

ence

5819ALK and PC12 cell differentiation

we looked for its potential functional role. Since the oncogenicand cytoplasmic NPM-ALK protein has been shown to triggercell transformation and cell growth (Bai et al., 2000; Bischofet al., 1997; Slupianek et al., 2001), we investigated whetheractivation of the cytoplasmic cytoIA-FH protein could controlPC12 cell growth. We therefore monitored DNA synthesisfollowing activation of cytoIA-FH in PC12 cells synchronizedin low-serum medium, as usually required to study DNAsynthesis induction (Pardee, 1989). As shown in Fig. 7A,induced activation of the cytoIA-FH protein significantlyincreased the proportion of transfected cells incorporatingBrdU (from 27% to 37%), revealing an increase of DNA

synthesis. By contrast, cells expressing its related kinase-defective mutant (dcytoIA-FH) or the membrane-boundmyrIA-FH protein did not exhibit such an effect, indicating thatthis induced DNA synthesis was specifically a result of thekinase activity of the cytosolic protein and that cytosoliclocalization was required for this effect. We then analyzed thesignal transduction pathways involved in this process. Weinvestigated whether this induced DNA synthesis was PI 3-kinase/AKT dependent, as it has been previously shown forNPM-ALK-induced cell proliferation (Bai et al., 2000;Slupianek et al., 2001). Thus, we tested the effect of twoselective and widely used inhibitors of PI 3-kinase activity,

Fig. 7. Activation of the cytosolic form of the ALK intracellular domain induces DNA synthesis through the PI 3-kinase/AKT pathway.(A) Induced activation of the cytoIA-FH but not of the myrIA-FH protein leads to DNA synthesis in PC12 cells. Cells were electroporated withnormalized quantities of the indicated constructs and were submitted to the same protocol as in Fig. 6 except that they were cultured overnightin a 1% horse serum medium before dimerizer treatment. ***P<0.005 (Student’s t-test). (B) PI 3-kinase inhibitor inhibition of DNA synthesisinduced by the activation of the cytoIA-FH protein. PC12 cells transiently expressing the cytoIA-FH protein were submitted to the sameprotocol as in A. They were treated with the dimerizer (20 nM) for 12 hours in the presence of the indicated increasing concentrations of PI 3-kinase inhibitors wortmannin or LY294002 dissolved in DMSO. As a control, cells were treated without DMSO (No vehicle). DMSO controls(0 nM or �M) showed no adverse effects. The experiments were performed in triplicate and values are expressed as the mean ± s.e.m. (%).(C) Induced DNA synthesis of PC12 cells expressing the cytoIA-FH protein is dependent on PI 3-kinase but not on ERK 1/2. Cells transientlyexpressing the cytoIA-FH protein were submitted to the same protocol as in A. U0126 (25 �M) and/or wortmannin (50 nM) were added to themedium 60 and 30 minutes, respectively, before dimerizer treatment. **P<0.01; ***P<0.005 (Student’s t-test). (D) Time course of AKTphosphorylation following dimerizer treatment. PC12 cells transfected with the indicated constructs were cultured overnight in 0% horse serummedium and then incubated with the dimerizer (20 nM) for the indicated periods. Cells were lysed in RIPA buffer as described in the Materialsand Methods and the lysates (10 �g) were submitted to western blot analysis using the anti-AKT phosphoserine-473 antibody and thenreprobed with the anti-AKT antibody. (E) MAP kinase pathway inhibition revealed a potential activation effect of myrIA-FH protein on DNAsynthesis through the PI 3-kinase pathway. Cells transiently expressing the myrIA-FH protein were submitted to the same protocol as in C.**P<0.01 (Student’s t-test).

Jour

nal o

f Cel

l Sci

ence

5820

wortmannin and LY294002 (Vlahos et al., 1994), on BrdUincorporation induced by dimerizer activation of the cytoIA-FHprotein. As shown in Fig. 7B, both PI 3-kinase inhibitorsinduced progressive inhibition of DNA synthesis of transfectedcells in a dose-dependent manner from 35% to about 17% ofthese cells incorporating BrdU. Note that, in the absence ofdimerizer, wortmannin reduced BrdU incorporation to a similarlevel (Fig. 7C). By contrast and as expected, cells did not showany significant inhibition of BrdU incorporation when treatedwith 25 �M of the inhibitor of ERK 1/2 activity U0126 (Fig.7C). Thus, DNA synthesis induced by activation of thecytosolic protein (cytoIA-FH) required activation of PI 3-kinasebut not the MAP kinase pathway. As AKT protein is one of themajor downstream effectors of PI 3-kinase, we studied the time-course activation of AKT by dimerizer treatment in cellstransfected with either the cytoIA-FH or myrIA-FH constructs.As shown in Fig. 7D in cells transfected with the cytoIA-FHconstruct, AKT phosphorylation was increased after 10 minutesof dimerizer treatment and was maximal after 30 minutes. Thissignificantly induced phosphorylation was completely inhibitedin the presence of 50 nM wortmannin (Fig. S4, supplementarymaterial). In good agreement with the preceding results, thesedata confirm that, following dimerizer stimulation, cellsexpressing the cytosolic protein actually displayed activation ofthe PI 3-kinase/AKT pathway, which was required to triggercell growth through an increase of DNA synthesis.

It is also noteworthy that in the same treatment conditionscells expressing the membrane-bound protein myrIA-FH (Fig.7D) also displayed an increase of AKT phosphorylation.Interestingly, we observed that activation of myrIA-FH proteinled to a significant increase in BrdU incorporation in thepresence of U0126 (Fig. 7E). This increase was abolished bythe dual addition of U0126 and wortmannin. These resultsstrongly suggested that the activation of the MAP kinasepathway masked the potential proliferation process that couldbe triggered by the PI 3-kinase pathway. Thus, activation of themembrane-bound proteins led to the activation of both theMAP kinase and PI 3-kinase/AKT pathways, whereas neuriteoutgrowth and proliferation arrest processes observed in thiscase essentially relied on the activation of the MAP kinasecascade.

DiscussionThe aim of the present report was to provide further elementsto the comprehension of the role of ALK in neuronaldifferentiation and/or proliferation and to analyze the role andmechanisms of action of the subcellular localization of its PTKdomain in relation to these processes and their related signalingpathways. Indeed, subcellular localization of a PTK domain asa factor that determines the signal transduction specificity hasbeen poorly documented so far (Schlessinger, 2000). In thisreport, we established for the first time that membraneattachment of the ALK PTK domain is crucial for induction ofneurite outgrowth of PC12 cells and their proliferation arrest.We further showed that these two processes resulted fromspecific and sustained activation of the MAP kinase pathway.By contrast, we demonstrated that activation of the cytosolicform of this domain failed to induce MAP kinase activationand cell differentiation but rather promoted a PI 3-kinase/AKT-dependent cell growth.

We first showed that the dimerizer induced thephosphorylation of the full-length ALK-FH protein andtriggered a MAP-kinase-dependent neurite outgrowth of PC12cells. This indicates that the induced activation of ALK byintracellular dimerization was functional. These data wereconsistent with previous studies reporting that this pathwaywas required for the ALK-dependent differentiation of neuron-like cells through dimerization of the Fc-ALK chimera or ALKextracellular domains (Moog-Lutz et al., 2005; Motegi et al.,2004; Souttou et al., 2001). Our data also indicated that MAP-kinase-dependent neurite outgrowth of PC12 cells was notdependent on the presence of the ALK extracellular domain ascomparable neurite outgrowth efficiencies were foundfollowing activation of full-length ALK-FH or truncatedtmbIA-FH proteins. MAP-kinase-dependent neurite outgrowthof PC12 cells was also not dependent on the type of membraneanchoring (transmembrane versus myristoyl). However, somenoticeable differences were detected. The myristoylatedprotein appeared more efficient at promoting both the basal anddimerizer-induced differentiation of PC12 cells. An attractivehypothesis would be that the myrIA-FH protein, like othermyristoylated proteins, is concentrated into the lipid rafts ofthe plasma membrane. This concentration by itself couldfacilitate the basal or induced dimerization of the protein. Inaddition, the compartmentalized assembly of MAP kinasesignaling proteins within lipid rafts might provide a higherefficiency for the myrIA-FH protein to activate these signalingproteins than for the transmembrane form tmbIA-FH. Forinstance, several reports pointed out the crucial role of theFRS2 adaptor in the NGF- or FGF-induced differentiation ofPC12 cells and established that FRS2 was associated withinlipid rafts through myristoyl anchors (Ong et al., 2000; Ridyardand Robbins, 2003).

The sustained activation of the ERK 1/2 proteins triggeredby the membrane-bound form of ALK also deservescomments. This typical kinetics pattern is in agreement withthose reported in several previous studies stating that asustained activation of the MAP kinase pathway induced byvarious neurotrophic factors such as NGF or FGF wasessential to promote PC12 cell differentiation (Marshall,1995). In fact, it was also previously shown in PC12 cells thatoverexpression of the EGF or insulin receptors in the presenceof their respective ligands led to a sustained activation of theMAP kinase cascade and subsequently to neuronaldifferentiation, whereas only a transient activation leading tocell proliferation has been described when these receptorswere expressed at a physiological level (Dikic et al., 1994;Traverse et al., 1994). This indicates that the level of receptorexpression could be crucial for the induction of the MAPkinase cascade and cell differentiation. As in transientlytransfected cells, proteins are usually overexpressed, whichcould explain the observed sustained activation of the MAPkinase pathway induced by the dimerizer. However, ourunpublished results obtained with stable HEK 293 clonesexpressing the same ALK-derived proteins at a much lowerlevel confirmed that ERK proteins also displayed a sustainedphosphorylation kinetics following dimerizer treatment (datanot shown). In addition, it had also been reported thatendogenously expressed ALK receptor in neuroblastoma cellscould promote neurite outgrowth through sustained MAPkinase activation (Motegi et al., 2004). Altogether, these data

Journal of Cell Science 118 (24)

Jour

nal o

f Cel

l Sci

ence

5821ALK and PC12 cell differentiation

suggest that ligand-induced activation of the endogenous ALKreceptor in physiological conditions could trigger a sustainedactivation of the MAP kinase cascade leading to subsequentneuronal differentiation in vivo.

The difference between the membrane-bound and cytosolicproteins in the ERK cascade activation could result from thecrucial role played by specific signaling complexes located atthe cell membrane. Indeed, activated RTKs could recruit to themembrane different cytosolic adaptors (Sos, Grb2, Shc, etc.)or directly link membrane-anchored docking proteins (FRS2).These protein complexes facilitate subsequent activation ofGTPase proteins of the Ras family (Ras, Rap1, etc.) that arealso associated to the membrane through lipid anchors(Schlessinger, 2000). Thus, these confined proteinassemblages bring together signal proteins, and organize andcoordinate the function of a large part of the MAP kinasesignaling cascade at the membrane. Thus, in contrast to themembrane-bound forms of ALK, the cytosolic form probablyfailed to activate these membrane-bound protein complexes,resulting in the lack of MAP kinase activation. In this context,it is noteworthy that the activation of the MAP kinase pathwayby the NPM-ALK protein or by other cytosolic forms of ALKresulting from various chromosomal translocations had neverbeen reported (reviewed by Pulford et al., 2004). Interestingly,the gene encoding moesin (MSN) at Xq11-12 has beenrecently reported as a new partner of ALK in rare cases ofanaplastic large cell lymphoma (ALCL). The fusion proteinMSN-ALK (Tort et al., 2001) exhibited a distinctivemembrane-restricted labeling pattern. This particularmembrane localization of an ALK oncogenic form ispresumed to reflect association of moesin with cell membraneproteins. Therefore, it would be interesting to know whetherthis membrane-bound form of ALK is able to activate theMAP kinase pathway.

Here, we cultured PC12 cells using the culture conditionsof low-serum medium (1% horse serum, overnight) beforedimerizer stimulation. We chose this particular cell culturecondition because it had been previously reported that NGF-induced neurite extension occurred more rapidly in PC12cells arrested in G0 phase than in asynchronously growingcells (Rudkin et al., 1989). In addition, synchronized cellsare also usually required to study DNA synthesis induction.Indeed, when deprived of serum, cells continue to cycle untilthey terminate mitosis, at which point they exit into the G0state (Pardee, 1989). Then, they can be reintroduced into thecell cycle by re-addition of growth factor. In agreement withthese data, we showed here that, when cultured in low-serummedium, cells expressing the cytosolic ALK-derived formunderwent DNA synthesis in the presence of dimerizer,probably reflecting their reintroduction into the cell cycle.Such an effect was also observed with cells cultured in aserum-containing medium, but DNA synthesis induction wasless significant. We also demonstrated that this phenomenonmainly required activation of the PI 3-kinase/AKT pathway.This is in full agreement with several previous reportsshowing that PI 3-kinase/AKT activation was required andsufficient for cell-cycle entry and DNA synthesis (Roche etal., 1994; Valius and Kazlauskas, 1993) and that thispathway was essential to enhance proliferation of cellsexpressing the cytosolic NPM-ALK protein (Bai et al., 2000;Slupianek et al., 2001). By using the engineered cytosolic

ALK-derived protein, we essentially reproduced the resultsobtained in NPM-ALK-positive cells. Thus, our model is toa certain extent a relevant transposition of the oncogenicsystem.

Induced AKT phosphorylation was also noticeable withcells expressing the membrane-bound myrIA-FH protein.Therefore, PI 3-kinase/AKT activation also occurs when theALK PTK domain is located at the membrane. Thus, activationof the membrane-bound protein led to activation of both theMAP kinase and PI 3-kinase pathways. Neurite outgrowth andproliferation arrest processes visualized in this case essentiallyrelied on the activation of the ERK pathway. Nevertheless,inhibition of the ERK pathway revealed a potentialproliferation effect of the myrIA-FH protein that is controlledby the PI 3-kinase/AKT pathway. These results fit differentreports analyzing the NGF effects through the MAP kinase andPI 3-kinase/AKT pathways in PC12 cells (Klesse et al., 1999).We showed here for the first time that induced activation of themembrane-bound forms of ALK triggered PC12 celldifferentiation and proliferation arrest. In contrast to our resultsand the NGF effects on PC12 cells, a recent study showed thatactivation of endogenous ALK expressed in the SK-N-SHneuroblastoma cell line induced both neurite outgrowth andcell proliferation (Motegi et al., 2004). A possible explanationis that the level of expression of the ALK receptor in these cellsis crucial as previously discussed. Furthermore, as pointed outby these authors, this difference of biological effects might bea result of cell-type specificity. Thus, depending on the levelsof activation of the ERK and PI 3-kinase pathways, the cellcould either differentiate and/or proliferate. A completeanalysis of both effects and inhibition of these two pathwayswould certainly be informative. The availability of monoclonalantibodies (Moog-Lutz et al., 2005) will now allow us to studydirectly the biological effects and related signaling pathwaystriggered by the activation of ALK during development, inparticular in primary neuronal cell cultures endogenouslyexpressing this receptor.

In conclusion, our results strongly support our initialhypothesis postulating that membrane attachment of the ALKPTK domain could be a determinant for the control andspecificity of the downstream transduction cascades. Thismembrane attachment was crucial for promotion of neuron-likedifferentiation and cell proliferation arrest of PC12 cellsthrough specific activation of the MAP kinase pathway. Inaddition, our model allowed for the first time a directcomparison between the full-length membrane-bound receptorand a cytosolic form of its PTK domain that strongly parallelsthe oncogenic forms of ALK resulting from variouschromosomal translocations. Indeed, our data showed thatsubcellular localization of the ALK PTK domain was crucialfor deciding the fate to which the neuronal cell will becommitted.

This work was supported in part by institutional funding fromINSERM and Université Paris 6, as well as by grants from theAssociation pour la Recherche sur le Cancer (ARC) and AssociationFrançaise contre les Myopathies (AFM). We thank ARIADPharmaceuticals (http://www.ariad.com/regulationkits) forproviding vectors and AP20187. We thank J. Gavard, H. Enslen andJ. Degoutin for fruitful discussions. We are particularly grateful toA. Sobel for his continual support and helpful comments on themanuscript.

Jour

nal o

f Cel

l Sci

ence

5822

ReferencesBai, R. Y., Dieter, P., Peschel, C., Morris, S. W. and Duyster, J. (1998).

Nucleophosmin-anaplastic lymphoma kinase of large-cell anaplasticlymphoma is a constitutively active tyrosine kinase that utilizesphospholipase C-gamma to mediate its mitogenicity. Mol. Cell. Biol. 18,6951-6961.

Bai, R. Y., Ouyang, T., Miething, C., Morris, S. W., Peschel, C. andDuyster, J. (2000). Nucleophosmin-anaplastic lymphoma kinase associatedwith anaplastic large-cell lymphoma activates the phosphatidylinositol 3-kinase/Akt antiapoptotic signaling pathway. Blood 96, 4319-4327.

Bibel, M. and Barde, Y. A. (2000). Neurotrophins: key regulators of cell fateand cell shape in the vertebrate nervous system. Genes Dev. 14, 2919-2937.

Bischof, D., Pulford, K., Mason, D. Y. and Morris, S. W. (1997). Role ofthe nucleophosmin (NPM) portion of the non-Hodgkin’s lymphoma-associated NPM-anaplastic lymphoma kinase fusion protein in oncogenesis.Mol. Cell. Biol. 17, 2312-2325.

Blume-Jensen, P. and Hunter, T. (2001). Oncogenic kinase signalling. Nature411, 355-365.

Boccaccio, C., Ando, M. and Comoglio, P. M. (2002). A differentiationswitch for genetically modified hepatocytes. FASEB J. 16, 120-122.

Clackson, T., Yang, W., Rozamus, L. W., Hatada, M., Amara, J. F., Rollins,C. T., Stevenson, L. F., Magari, S. R., Wood, S. A., Courage, N. L. et al.(1998). Redesigning an FKBP-ligand interface to generate chemicaldimerizers with novel specificity. Proc. Natl. Acad. Sci. USA 95, 10437-10442.

Dikic, I., Schlessinger, J. and Lax, I. (1994). PC12 cells overexpressing theinsulin receptor undergo insulin-dependent neuronal differentiation. Curr.Biol. 4, 702-708.

Dirks, W. G., Fahnrich, S., Lis, Y., Becker, E., MacLeod, R. A. and Drexler,H. G. (2002). Expression and functional analysis of the anaplastic lymphomakinase (ALK) gene in tumor cell lines. Int. J. Cancer 100, 49-56.

Englund, C., Loren, C. E., Grabbe, C., Varshney, G. K., Deleuil, F.,Hallberg, B. and Palmer, R. H. (2003). Jeb signals through the Alkreceptor tyrosine kinase to drive visceral muscle fusion. Nature 425, 512-516.

Favata, M. F., Horiuchi, K. Y., Manos, E. J., Daulerio, A. J., Stradley, D.A., Feeser, W. S., Van Dyk, D. E., Pitts, W. J., Earl, R. A., Hobbs, F. etal. (1998). Identification of a novel inhibitor of mitogen-activated proteinkinase kinase. J. Biol. Chem. 273, 18623-18632.

Fujimoto, J., Shiota, M., Iwahara, T., Seki, N., Satoh, H., Mori, S. andYamamoto, T. (1996). Characterization of the transforming activity of p80,a hyperphosphorylated protein in a Ki-1 lymphoma cell line withchromosomal translocation t(2;5). Proc. Natl. Acad. Sci. USA 93, 4181-4186.

Greene, L. A. and Tischler, A. S. (1976). Establishment of a noradrenergicclonal line of rat adrenal pheochromocytoma cells which respond to nervegrowth factor. Proc. Natl. Acad. Sci. USA 73, 2424-2428.

Huang, E. J. and Reichardt, L. F. (2001). Neurotrophins: roles in neuronaldevelopment and function. Annu. Rev. Neurosci. 24, 677-736.

Huang, E. J. and Reichardt, L. F. (2003). Trk receptors: roles in neuronalsignal transduction. Annu. Rev. Biochem. 72, 609-642.

Huff, K., End, D. and Guroff, G. (1981). Nerve growth factor-inducedalteration in the response of PC12 pheochromocytoma cells to epidermalgrowth factor. J. Cell Biol. 88, 189-198.

Iwahara, T., Fujimoto, J., Wen, D., Cupples, R., Bucay, N., Arakawa, T.,Mori, S., Ratzkin, B. and Yamamoto, T. (1997). Molecularcharacterization of ALK, a receptor tyrosine kinase expressed specificallyin the nervous system. Oncogene 14, 439-449.

Kaplan, D. R. and Miller, F. D. (2000). Neurotrophin signal transduction inthe nervous system. Curr. Opin. Neurobiol. 10, 381-391.

Klesse, L. J., Meyers, K. A., Marshall, C. J. and Parada, L. F. (1999). Nervegrowth factor induces survival and differentiation through two distinctsignaling cascades in PC12 cells. Oncogene 18, 2055-2068.

Lee, H. H., Norris, A., Weiss, J. B. and Frasch, M. (2003). Jelly belly proteinactivates the receptor tyrosine kinase Alk to specify visceral musclepioneers. Nature 425, 507-512.

Loren, C. E., Scully, A., Grabbe, C., Edeen, P. T., Thomas, J., McKeown,M., Hunter, T. and Palmer, R. H. (2001). Identification andcharacterization of DAlk: a novel Drosophila melanogaster RTK whichdrives ERK activation in vivo. Genes Cells 6, 531-544.

Lu, K. V., Jong, K. A., Kim, G. Y., Singh, J., Dia, E. Q., Yoshimoto, K.,Wang, M. Y., Cloughesy, T. F., Nelson, S. F. and Mischel, P. S. (2005).Differential Induction of Glioblastoma Migration and Growth by Two Formsof Pleiotrophin. J. Biol. Chem. 280, 26953-26964.

Manning, G., Whyte, D. B., Martinez, R., Hunter, T. and Sudarsanam, S.(2002). The protein kinase complement of the human genome. Science 298,1912-1934.

Marshall, C. J. (1995). Specificity of receptor tyrosine kinase signaling:transient versus sustained extracellular signal-regulated kinase activation.Cell 80, 179-185.

Miyake, I., Hakomori, Y., Shinohara, A., Gamou, T., Saito, M., Iwamatsu,A. and Sakai, R. (2002). Activation of anaplastic lymphoma kinase isresponsible for hyperphosphorylation of ShcC in neuroblastoma cell lines.Oncogene 21, 5823-5834.

Moog-Lutz, C., Degoutin, J., Gouzi, J. Y., Frobert, Y., Carvalho, N. B.-d.,Bureau, J., Creminon, C. and Vigny, M. (2005). Activation and inhibitionof anaplastic lymphoma kinase receptor tyrosine kinase by monoclonalantibodies and absence of agonist activity of pleiotrophin. J. Biol. Chem.280, 26039-26048.

Morris, S. W., Kirstein, M. N., Valentine, M. B., Dittmer, K. G., Shapiro,D. N., Saltman, D. L. and Look, A. T. (1994). Fusion of a kinase gene,ALK, to a nucleolar protein gene, NPM, in non-Hodgkin’s lymphoma.Science 263, 1281-1284.

Morris, S. W., Naeve, C., Mathew, P., James, P. L., Kirstein, M. N., Cui,X. and Witte, D. P. (1997). ALK, the chromosome 2 gene locus altered bythe t(2;5) in non-Hodgkin’s lymphoma, encodes a novel neural receptortyrosine kinase that is highly related to leukocyte tyrosine kinase (LTK).Oncogene 14, 2175-2188.

Motegi, A., Fujimoto, J., Kotani, M., Sakuraba, H. and Yamamoto, T.(2004). ALK receptor tyrosine kinase promotes cell growth and neuriteoutgrowth. J. Cell Sci. 117, 3319-3329.

Nakanishi, S., Kakita, S., Takahashi, I., Kawahara, K., Tsukuda, E., Sano,T., Yamada, K., Yoshida, M., Kase, H., Matsuda, Y. et al. (1992).Wortmannin, a microbial product inhibitor of myosin light chain kinase. J.Biol. Chem. 267, 2157-2163.

Nieborowska-Skorska, M., Slupianek, A., Xue, L., Zhang, Q., Raghunath,P. N., Hoser, G., Wasik, M. A., Morris, S. W. and Skorski, T. (2001).Role of signal transducer and activator of transcription 5 in nucleophosmin/anaplastic lymphoma kinase-mediated malignant transformation oflymphoid cells. Cancer Res. 61, 6517-6523.

Ong, S. H., Guy, G. R., Hadari, Y. R., Laks, S., Gotoh, N., Schlessinger, J.and Lax, I. (2000). FRS2 proteins recruit intracellular signaling pathwaysby binding to diverse targets on fibroblast growth factor and nerve growthfactor receptors. Mol. Cell. Biol. 20, 979-989.

Pardee, A. B. (1989). G1 events and regulation of cell proliferation. Science246, 603-608.

Powis, G., Bonjouklian, R., Berggren, M. M., Gallegos, A., Abraham, R.,Ashendel, C., Zalkow, L., Matter, W. F., Dodge, J., Grindey, G. et al.(1994). Wortmannin, a potent and selective inhibitor ofphosphatidylinositol-3-kinase. Cancer Res. 54, 2419-2423.

Pulford, K., Morris, S. W. and Turturro, F. (2004). Anaplastic lymphomakinase proteins in growth control and cancer. J. Cell Physiol. 199, 330-358.

Ridyard, M. S. and Robbins, S. M. (2003). Fibroblast growth factor-2-induced signaling through lipid raft-associated fibroblast growth factorreceptor substrate 2 (FRS2). J. Biol. Chem. 278, 13803-13809.

Roche, S., Koegl, M. and Courtneidge, S. A. (1994). Thephosphatidylinositol 3-kinase alpha is required for DNA synthesis inducedby some, but not all, growth factors. Proc. Natl. Acad. Sci. USA 91, 9185-9189.

Rudkin, B. B., Lazarovici, P., Levi, B. Z., Abe, Y., Fujita, K. and Guroff,G. (1989). Cell cycle-specific action of nerve growth factor in PC12 cells:differentiation without proliferation. EMBO J. 8, 3319-3325.

Schlessinger, J. (2000). Cell signaling by receptor tyrosine kinases. Cell 103,211-225.

Slupianek, A., Nieborowska-Skorska, M., Hoser, G., Morrione, A.,Majewski, M., Xue, L., Morris, S. W., Wasik, M. A. and Skorski, T.(2001). Role of phosphatidylinositol 3-kinase-Akt pathway innucleophosmin/anaplastic lymphoma kinase-mediated lymphomagenesis.Cancer Res. 61, 2194-2199.

Souttou, B., Carvalho, N. B., Raulais, D. and Vigny, M. (2001). Activationof anaplastic lymphoma kinase receptor tyrosine kinase induces neuronaldifferentiation through the mitogen-activated protein kinase pathway. J.Biol. Chem. 276, 9526-9531.

Stoica, G. E., Kuo, A., Aigner, A., Sunitha, I., Souttou, B., Malerczyk, C.,Caughey, D. J., Wen, D., Karavanov, A., Riegel, A. T. et al. (2001).Identification of anaplastic lymphoma kinase as a receptor for the growthfactor pleiotrophin. J. Biol. Chem. 276, 16772-16779.

Stoica, G. E., Kuo, A., Powers, C., Bowden, E. T., Sale, E. B., Riegel, A.

Journal of Cell Science 118 (24)

Jour

nal o

f Cel

l Sci

ence

5823ALK and PC12 cell differentiation

T. and Wellstein, A. (2002). Midkine binds to anaplastic lymphoma kinase(ALK) and acts as a growth factor for different cell types. J. Biol. Chem.277, 35990-35998.

Tort, F., Pinyol, M., Pulford, K., Roncador, G., Hernandez, L., Nayach, I.,Kluin-Nelemans, H. C., Kluin, P., Touriol, C., Delsol, G. et al. (2001).Molecular characterization of a new ALK translocation involving moesin(MSN-ALK) in anaplastic large cell lymphoma. Lab. Invest. 81, 419-426.

Traverse, S., Seedorf, K., Paterson, H., Marshall, C. J., Cohen, P. andUllrich, A. (1994). EGF triggers neuronal differentiation of PC12 cells thatoverexpress the EGF receptor. Curr. Biol. 4, 694-701.

Valius, M. and Kazlauskas, A. (1993). Phospholipase C-gamma 1 andphosphatidylinositol 3 kinase are the downstream mediators of the PDGFreceptor’s mitogenic signal. Cell 73, 321-334.

van Grunsven, L. A., Billon, N., Savatier, P., Thomas, A., Urdiales, J. L.and Rudkin, B. B. (1996a). Effect of nerve growth factor on the expressionof cell cycle regulatory proteins in PC12 cells: dissection of the neurotrophicresponse from the anti-mitogenic response. Oncogene 12, 1347-1356.

van Grunsven, L. A., Thomas, A., Urdiales, J. L., Machenaud, S., Choler,

P., Durand, I. and Rudkin, B. B. (1996b). Nerve growth factor-inducedaccumulation of PC12 cells expressing cyclin D1: evidence for a G1 phaseblock. Oncogene 12, 855-862.

Vlahos, C. J., Matter, W. F., Hui, K. Y. and Brown, R. F. (1994). A specificinhibitor of phosphatidylinositol 3-kinase, 2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one (LY294002). J. Biol. Chem. 269, 5241-5248.

Yan, G. Z. and Ziff, E. B. (1995). NGF regulates the PC12 cell cyclemachinery through specific inhibition of the Cdk kinases and induction ofcyclin D1. J. Neurosci. 15, 6200-6212.

Yan, G. Z. and Ziff, E. B. (1997). Nerve growth factor induces transcriptionof the p21 WAF1/CIP1 and cyclin D1 genes in PC12 cells by activating theSp1 transcription factor. J. Neurosci. 17, 6122-6132.

Yang, J., Symes, K., Mercola, M. and Schreiber, S. L. (1998). Small-molecule control of insulin and PDGF receptor signaling and the role ofmembrane attachment. Curr. Biol. 8, 11-18.

Zamo, A., Chiarle, R., Piva, R., Howes, J., Fan, Y., Chilosi, M., Levy, D. E.and Inghirami, G. (2002). Anaplastic lymphoma kinase (ALK) activates Stat3and protects hematopoietic cells from cell death. Oncogene 21, 1038-1047.

Jour

nal o

f Cel

l Sci

ence