CB1 Cannabinoid Receptor-MediatedNeurite Remodeling in MouseNeuroblastoma N1E-115 Cells

Dan Zhou and Z.H. Song*Department of Pharmacology and Toxicology, School of Medicine, University of Louisville, Kentucky

The morphological remodeling of neuronal cells influ-ences neurogenesis and brain functions. We hypothesizethat psychoactive and neurotoxic effects of cannabi-noids may be mediated, at least in part, by their morpho-regulatory activities. In the present study, mouse neuro-blastoma N1E-115 cells were used as an in vitro model toinvestigate cannabinoid-induced neurite remodeling ef-fects and to identify the involvement of cannabinoid re-ceptors in this neurite remodeling process. Using reversetranscription-polymerase chain reaction and immunoflu-orescence microscopy, the endogenously expressedCB1, but not CB2, cannabinoid receptors were detectedin morphologically differentiated N1E-115 cells. Activa-tion of these natively expressed CB1 cannabinoid recep-tors by cannabinoid agonist HU-210 led to aconcentration-dependent inhibition of adenylate cyclaseactivity. Importantly, HU-210 treatment induced neuriteretraction in a concentration-dependent manner. Pre-treatment of N1E-115 cells with a CB1 antisense oligode-oxynucleotide (ODN) suppressed HU-210-induced inhi-bition of forskolin-stimulated cAMP accumulation,indicating that the knocking down of functional CB1 can-nabinoid receptor expression was achieved. AntisenseODN pretreatment also abolished HU-210-induced neu-rite retraction, demonstrating the involvement of CB1cannabinoid receptors in mediating the neurite remodel-ing effects of HU-210. In addition, reversing HU-210-induced intracellular cAMP declination by 8-Br-cAMPpartially prevented HU-210-induced neurite retraction,indicating the involvement of cAMP-dependent signalingpathways in mediating the neurite remodeling function ofCB1 cannabinoid receptors in N1E-115 cells. These datademonstrate that neurite remodeling is a newly discov-ered function of CB1 cannabinoid receptors. This morpho-regulatory function of CB1 cannabinoid receptors mightbe a new mechanism that mediates the psychoactiveand neurotoxic effects of cannabinoids in developing andadult brain. J. Neurosci. Res. 65:346–353, 2001.© 2001 Wiley-Liss, Inc.

Key words: CB1 cannabinoid receptor; antisense;cAMP; neurites; retraction

INTRODUCTIONThe discovery of cannabinoid receptors and endog-

enous cannabinoid ligands has drawn much attention tocannabinoid research (Howlett, 1995; Pertwee, 1997). Sofar, two cannabinoid receptors have been cloned andcharacterized. The CB1 cannabinoid receptor is foundpredominantly on neurons in the central and peripheralnervous systems (Matsuda et al., 1990, 1993), whereas theCB2 cannabinoid receptor is expressed mainly on the cellsof the immune system (Munro et al., 1993). It has alreadybeen shown that the CB1 cannabinoid receptor couples tomultiple signal transduction pathways, including adenylatecyclase, mitogen-activated protein kinase (MAPK), andion channels (Howlett, 1995; Pertwee, 1997). The diver-sity of signaling pathways that couple to the CB1 canna-binoid receptor indicates multiple physiological functionsthat may be mediated by this receptor. In the brain, theCB1 cannabinoid receptor is distributed predominantly inthe cortex, basal ganglia, cerebellum, and hippocampus(Herkenham et al., 1991; Matsuda et al., 1993; Tsou et al.,1998). The endogenous cannabinoid system has been pro-posed, based on the anatomical distribution of CB1 can-nabinoid receptors and the pharmacological effects of can-nabinoid compounds, to be involved in the control ofmovement, cognition, learning, memory, and brain de-velopments (Felder and Glass, 1998).

The effects of cannabinoids on rat brain develop-ment have been examined (Fernandez-Ruiz et al., 2000).It was found that feeding the mother rats with cannabinoidcompounds during pregnancy or lactation or both couldmodify the maturation of neurotransmitter systems andrelated behaviors of their fetuses. These effects are prob-ably mediated by the activation of cannabinoid receptorsthat are expressed in the early stages of the developingbrain. Recently, by using magnetic resonance imaging(MRI) and positron emission tomography (PET), it was

Contract grant sponsor: National Institutes of Health; Contract grant num-ber: DA-11511.

*Correspondence to: Z.H. Song, Department of Pharmacology and Tox-icology, School of Medicine, University of Louisville, Louisville, KY40292. E-mail: zhsong@louisville.edu

Received 16 February 2001; Revised 30 April 2001; Accepted 2 May 2001

Journal of Neuroscience Research 65:346–353 (2001)

© 2001 Wiley-Liss, Inc.

found that human subjects who started using marijuana atan early adolescent age showed a smaller whole brain andpercentage of cortical gray matter as well as a large per-centage of white matter volume (Wilson et al., 2000). Inaddition, it has been detected that long-term treatment ofrats with WIN55212-2 could induce morphologicalchanges in hippocampus (Lawston et al., 2000). Takentogether, these data suggest a regulatory function of can-nabinoids on neuronal development and morphogenesis.The effects of cannabinoids on cytoskeletal componentsand on cellular morphology have been observed in differ-ent cell types (Kiosses et al., 1990; Tahir and Zimmerman,1991; Tahir et al., 1992). However, these previous reportsused relatively high concentrations of cannabinoid ligands.It is unknown whether cannabinoid-induced cytoskeletaland morphological changes are mediated by specific can-nabinoid receptors or that these changes reflect a nonspe-cific cellular toxicity.

Mouse neuroblastoma N1E-115 cells have been usedwidely as a model system to study neurite genesis andremodeling (Hirose et al., 1998; Kranenburg et al., 1999).In the present study, experiments were performed to clar-ify whether functional cannabinoid receptors are expressedendogenously in differenciated N1E-115 cells andwhether these receptors mediate HU-210-induced neuriteremodeling in this cell line.

MATERIALS AND METHODS

Cell Culture and Treatments

Mouse neuroblastoma N1E-115 cells were obtained fromthe American Type Culture Collection (ATCC, Rockville,MD). Cells were maintained in Dulbecco’s modified Eagle’smedium (DMEM)-10% fetal calf serum (FBS) supplementedwith penicillin and streptomycin (Gibco BRL, Rockville, MD)at 37°C in an atmosphere of humidified air and 5% CO2.

For neurite induction, N1E-115 cells were cultured inserum-free DMEM for 24–48 hr. More than 85% of the cellsbecame flattened and about 40% extended neurites under theseconditions. Neurite retraction was evoked by incubating cellswith HU-210 for 15 min with or without preincubation with8-Br-cAMP (Sigma, St. Louis, MO) for 30 min.

Antisense treatment was performed according to a proce-dure published by Hunter and Burstein (1997), using phospho-rothioate oligodeoxynucleotide (ODN; Integrated DNA Tech-nologies, Inc., Coralville, IA) corresponding to the sequences ofmouse CB1 cannabinoid receptor cDNA. The antisense probe isan 18 mer 59-GTACTGAATGTCATTTGA-39 complemen-tary to positions 73–90 of the mouse CB1 cannabinoid receptormRNA codon. The corresponding 18-mer sense fragment,59-TCAAATGACATTCAGTAC-39, was synthesized and usedas control. The antisense sequence was compared to sequencesdeposited in the GenBank/EMBL data banks and revealed noidentity with other rodent sequences. N1E-115 cells were pre-treated with 3.0 mM antisense or sense control ODN for 48 hrin serum-free DMEM at 37°C in an atmosphere of humidifiedair and 5% CO2. Following antisense ODN pretreatment, thecells were used in cAMP accumulation and morphological as-says.

Reverse Transcription-Polymerase Chain ReactionDetection of Cannabinoid Receptor mRNA Expression

The reverse transcription-polymerase chain reaction (RT-PCR) conditions for the detection of cannabinoid receptormRNA expression in N1E-115 cells were similar to thosedescribed by Zhou et al. (1999), with modifications. Briefly,total RNA was isolated from the cells using the Rneasy Mini Kit(Qiagen, Valencia, CA). One microgram of RNA was firsttreated with 0.5 U RNase free DNase-I (Promega, Madison,WI) and then reverse transcribed using 100 U of M-MLVreverse transcriptase (Ambion, Inc., Austin, TX). The reversetranscription system contained 5 mM oligo-dT primer, 10 Uplacental RNase inhibitors, 10 mM Tris-HCl (pH 8.3), 50 mMKCl, and 1.5 mM MgCl2. The prepared reaction mixture witha 20 ml final volume was incubated at 37°C for 1 hr. The PCRwas performed in a solution of 10 ml, containing 5 ml of reversetranscription product, 1 ml of 103 PCR buffer (100 mMTris-HCl, 500 mM KCl, 1.5 mM MgCl2, pH 8.3), 0.5 mM ofeach primer, 1 ml of 0.5 mM dNTP mix, and 0.5 U of TaqDNA polymerase (Ambion, Inc.). For the CB1 cannabinoidreceptor, the forward primer is 59-GTCACCAGTGTGCT-GTTGCT-39, and the reverse primer is 59-TGTCTCAGG-TCCTTGCTCCT-39. For the CB2 cannabinoid receptor, theforward primer is 59-TGCTGCTCATATGCTGGTTC-39,and the reverse primer is 59-CTTCTGACTCGGGCTGT-TTC-39. In a PTC-100 programmable thermal controller (MJResearch, Inc., Watertown, MA), the samples were denatured at94°C for 2 min and then subjected to 30 cycles at 94°C for0.5 min, 55°C for 0.5 min, and 72°C for 1 min, with a final10 min extension at 72°C. The expression of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as internal con-trol. The PCR fragments were separated on 1.5% agarose gelsand visualized by ethidium bromide staining. Gels were photo-graphed using a Polaroid camera. The amplified DNA productwas extracted from the gels with the QIA quick gel extractionkit (Qiagen, Inc.) and sequenced by fluorescence dye termina-tors on a Beckman CEQ 2000 DNA analysis system.

Cannabinoid Receptor Immunofluorescence Staining

N1E-115 cells were seeded at a density of 2.5 3 105 cellsper slide onto glass coverslips. After 24 hr of serum starvation,cells were fixed in 4% paraformaldehyde for 10 min at roomtemperature. The paraformaldehyde was quenched with100 mM glycine for 10 min, and the cells were blocked with10% normal goat serum in PBS for 20 min without permeabi-lization. Primary antibody used was rabbit polyclonal anti-CB1antibody (Cayman Chemical Co., Ann Arbor, MI) at 5 mg/ml.Cells were incubated with primary antibodies for 1.5 hr at 37°Cin a humidified chamber, washed, and incubated further for 1 hrat 37°C with fluorescein isothiocyanate (FITC)-conjugated goatanti-rabbit IgG antibody (Zymed, South San Francisco, CA).Slides were mounted in Vectashield mounting medium withDAPI (Vector Laboratories, Inc., Burlingame, CA). Themounted slides were viewed under a Nikon fluorescence mi-croscope (Eclipse E-800) and photographed with a digital cam-era (model SP401-115; Diagnostic, Inc., San Diego, CA).

CB1 Receptor and Neurite Remodeling 347

cAMP Accumulation Assay

cAMP accumulation assays were performed as describedpreviously, with slight modifications (Felder et al., 1992; Songand Bonner, 1996). Briefly, HU-210 was diluted with DMEMmedium containing 50 mg/ml fatty acid-free bovine serumalbumin (BSA) in silanized test tubes, and 25 ml solutions withdifferent drug concentrations were added to individual tubes.Forskolin was also added to the test tubes in a volume of 25 mlto a final concentration of 1.0 mM. The confluent cells werelifted from culture plates with phosphate-buffered saline (PBS)containing 0.5 mM EDTA. After washing with medium, thecells were incubated with phosphodiesterase inhibitor RO20-1724 for 10 min. The stimulation of cAMP synthesis wasinitiated by adding cells to test tubes containing forskolin andHU-210. The final reaction volume was 250 ml, with 1 3 106

cells per tube, and the reaction mixture was incubated at 37°Cfor 5 min. The reaction was stopped with the addition of anequal volume of 0.1 N HCl. An aliquot of 50 ml reactionsupernatant was removed for cAMP radioimmunoassay using akit from DuPont-NEN (DuPont-NEN, Wilmington, DE). TheEC50 value of HU-210 on inhibiting forskolin-stimulate cAMPaccumulation was calculated by nonlinear regression analysisusing Prism 2.01 (GraphPad Software, San Diego, CA).

Phalloidin Staining and Cell Counting

After serum starvation and different treatments, the cellswere fixed in 4% paraformaldehyde for 15 min at room tem-perature. The paraformaldehyde was quenched with 100 mMglycine for 10 min, and the cells were then permeabilized with0.3% Triton X-100 for 10 min. Subsequently, slides werewashed briefly with and blocked with 3% bovine serum albuminin PBS for 10 min. After briefly washing with PBS, Alexa-488-conjugated phalloidin (1.0 mg/ml; Molecular Probes, Eugene,OR) was added to stain F-actin. Slides were mounted inVectashield mounting medium with DAPI (Vector Laborato-ries, Inc.) and viewed under an Olympus IX50 fluorescencemicroscope. More than 200 cells were identified in randomlychosen fields in five views, and the percentage of cells thatretracted neurites after HU-210 application was calculated. Sta-tistical analysis was performed using one-way ANOVA (Prism2.01; GraphPad Software, Inc.).

RESULTS

Identification of CB1 Cannabinoid ReceptormRNA and Protein in N1E-115 Cells

RT-PCR experiments were performed to examinethe expression of mRNA for CB1 and CB2 cannabinoidreceptors in N1E-115 cells. By using specific primers thatrecognize CB1 cannabinoid receptor mRNA, the PCRamplification of the first-strand cDNA yielded a singleband that matched the expected molecular size of 394 basepair. Sequencing of this PCR product confirmed that thisband indeed represents the CB1 cannabinoid receptor. Incontrast, primers designed specifically for the CB2 canna-binoid receptor failed to detect the expression of CB2mRNA in these cells. The amplicon of the internal con-trol, GAPDH, presented a single band around the ex-pected molecular size of 395 bp (Fig. 1A). In addition, in

undifferentiated N1E-115 cells, RT-PCR experimentsalso detected mRNA for CB1, but not CB2, receptor(data not shown).

To detect CB1 cannabinoid receptor protein expres-sion, serum-starved N1E-115 cells were labeled with arabbit anti-CB1 cannabinoid receptor polyclonal anti-body, and the cells were subsequently reacted with FITC-conjugated anti-rabbit IgG. The expression of the CB1cannabinoid receptor molecules was found on both cellbodies and neurite formations (Fig. 1B).

CB1 Cannabinoid Receptor Signalingin N1E-115 Cells

The synthetic cannabinoid agonist HU-210 wasfound to inhibit forskolin-stimulated cAMP accumulationin N1E-115 cells. The HU-210-induced inhibition ofcAMP accumulation was concentration dependent (Fig.2). The EC50 value that was calculated by nonlinear re-gression is 0.14 6 0.05 nM, which is consistent withpreviously reported values (Felder et al., 1992; Song andBonner, 1996).

Fig. 1. Detection of CB1 cannabinoid receptors in mouse neuroblas-toma N1E-115 cells. A: RT-PCR detection of CB1 cannabinoidreceptor mRNA in N1E-115 cells. Total RNA was isolated fromN1E-115 cells and reverse transcribed into cDNA. The cDNA wasfurther amplified by PCR using specific primers that recognize CB1 orCB2 cannabinoid receptor. CB1, but not CB2, cannabinoid receptormRNA was amplified at the expected size. GAPDH was used asinternal control. B: Immunofluorescent staining of cell surface CB1cannabinoid receptors. CB1 cannabinoid receptors were found both onthe cell body and on neurite formations of N1E-115 cells. Cells werephotographed at 3100.

348 Zhou and Song

Neurite Retraction Induced by HU-210 Treatmentin N1E-115 Cells

N1E-115 cells exhibit neurite outgrowth in responseto serum starvation. In the absence of serum, roundedN1E-115 cells initially became more flattened and thenproduced spiky structures around their circumference,which gradually polarized to one or two discrete regionsfrom which the neurites grew out in the following 24–48 hr.Under our culture conditions, more than 35% of the cellsexhibited neurite bearing after 24 hr of serum starvation,which is similar to the published observations (Hirose etal., 1998; Kranenburg et al., 1999). Treating these serum-starved N1E-115 cells with HU-210 at different concen-trations for 15 min induced a concentration-dependentretraction of preformed neurites (Fig. 3A,B). No signifi-cant differences in cell viability between control and HU-210-treated groups were found through DAPI stainingand MTT assay (data not shown), indicating that theobserved neurite retraction is not a result of cell death inthese experiments.

Inhibition Of CB1 Cannabinoid ReceptorSignaling and HU-210-Induced Neurite Retractionby a Specific CB1 Antisense ODN Treatment

An antisense “knock-down” strategy was used toconfirm the involvement of CB1 cannabinoid receptors inmediating the neurite remodeling effect of HU-210. Incells that were pretreated with CB1 antisense ODN, HU-210 (10 nM) failed to inhibit forskolin-stimulated cAMPaccumulation. However, in sense control, ODN-pretreated cells, HU-210 retained its effect of inhibitingforskolin-stimulated cAMP accumulation (Fig. 4). In ad-

dition, ODNs alone did not cause any changes in intra-cellular cAMP levels (data not shown).

In the series of experiments using antisense strategyto prevent the HU-210-induced neurite retraction, thecells that were pretreated with CB1 antisense ODN dem-onstrated less neurite retracting response to HU-210 treat-ment compared to the control group. In contrast, nosignificant suppression of HU-210-induced neurite re-modeling was found in the sense ODN pretreatmentgroup (Fig. 5). Furthermore, neither antisense nor senseODN pretreatment alone caused any neurite remodeling(Fig. 5).

Suppression of HU-210-Induced NeuriteRetraction by a cAMP Analogue

One of the major signaling pathways of cannabinoidreceptors is the inhibition of adenylate cyclase. To test thehypothesis that HU-210-induced neurite remodeling in-volves the decreased intracellular cAMP levels, 8-Br-cAMP, a membrane-permeable cAMP analogue, wasused. Pretreatment of N1E-115 cells with 8-Br-cAMP for30 min partially inhibited HU-210-induced neurite re-traction in a concentration-dependent manner, whereas8-Br-cAMP by itself did not stimulate significant morpho-logical changes (Fig. 6). At the highest concentrationsused, 8-Br-cAMP was able to reduce by 75% the HU210-induced neurite remodeling.

DISCUSSIONNumerous studies have found that morphological

changes play an important role in mediating the neuro-toxic effects of major drugs of abuse (Levitt et al., 1997;Huether et al., 1997). The morphological disruption ofnormal hippocampal formation has also been suggested tocontribute to the learning and short-term memory deficitsrelated to marijuana and cannabinoid usage (Scallet, 1991).Additionally, a morphological degeneration has beenobserved in the hippocampal CA1 structure of rats thatwere treated with a synthetic cannabinoid analogue,WIN55212-2 (Lawston et al., 2000). Based on these pre-vious reports, we postulate that psychoactive and neuro-toxic effects of cannabinoids may be mediated partially bytheir morphoregulatory activities. In the current study, themouse neuroblastoma N1E-115 cells were used to inves-tigate the involvement of cannabinoid receptor and itsnegatively coupled adenylate cyclase pathway in mediatingneurite remodeling, which might play a role in the psy-choactive and neurotoxic effects of cannabinoids.

A well-known response of mouse neuroblastomaN1E-115 cells to serum withdrawal is morphological dif-ferentiation by means of flattening and neurite extension.Addition of serum or agents such as lysophosphatidic acid(LPA) will induce retraction of preformed neurites, so thatthe cells become rounded again. Because of these well-established characteristics, N1E-115 cells have been usedas an in vitro model to study the receptor-mediated neu-rite remodeling processes (Hirose et al., 1998; Kranenburget al., 1999). Using RT-PCR and immunofluorescencemicroscopy techniques, the current study demonstrated

Fig. 2. Inhibition of forskolin-stimulated cAMP accumulation inmouse neuroblastoma N1E-115 cells by HU-210. N1E-115 cells werepreincubated with RO-20-1724 for 10 min at 37°C, and then the cellswere incubated with HU-210 plus 2.0 mM forskolin for 10 min. ThecAMP levels were measured by radioimmunoassays. Seven concentra-tions of HU-210, ranging from 3 3 10–10 to 3 3 10–8 M, were used.Data points represent mean 6 SEM of a triplicate experiment. Theexperiment was repeated twice with similar results.

CB1 Receptor and Neurite Remodeling 349

Fig. 3. HU-210-induced neurite remodeling in mouse neuroblastomaN1E-115 cells. A: N1E-115 neuroblastoma cells undergoing morpho-logical changes in response to serum starvation and HU-210 treatment.Cells were photographed at 340. Ia,Ib: Cultured N1E-115 cells dem-onstrate a rounded morphology in regular culture medium containing10% fetal bovine serum. IIa,IIb: Serum withdrawl for 48 hr stimulatedmorphological differentiation by means of flattening and neurite ex-tension. IIIa,IIIb: HU-210 treatment at 10 nM for 15 min inducedneurite retraction. B: HU-210-induced neurite retraction in serum-

starved N1E-115 cells was concentration dependent. Morphologicallydifferentiated N1E-115 cells were incubated with HU-210 at concen-trations ranging from 0.01 to 100 nM. After 15 min of incubation, thecells were fixed and stained with Alexa-488 fluorescent-labeled phal-loidin. In A, Ia, IIa, and IIIa are the results from Alexa-488 fluorescent-labeled phalloidin staining, and Ib, IIb, and IIIb are phase-contrastimages. Cells were counted, and the data shown in B are mean 6 SEMof at least three experiments (*P , 0.05, **P , 0.01).

that CB1, but not CB2, cannabinoid receptors are ex-pressed endogenously in N1E-115 cells. Furthermore, theEC50 value of HU-210 in inhibiting forskolin-stimulatedcAMP accumulation is consistent with previous reports,indicating that these natively expressed CB1 cannabinoidreceptors are indeed functionally active.

The effects of cannabinoids on cell morphology andcytoskeleton have been reported for several cell types(Kiosses et al., 1990; Tahir and Zimmerman, 1991; Tahiret al., 1992). For example, in PC-12 cells, D9-THC hasbeen shown to disrupt microfilaments and microtubules(Tahir et al., 1992). However, since relatively high con-centrations of cannabinoids were used, and it was uncer-tain whether cannabinoid receptors were expressed inthese cell lines, the contribution of cannabinoid receptorsin mediating the morphological remodeling remains un-known from these previous studies. To our knowledge,the current study is the first to report that HU-210, aprototypical cannabinoid, can cause neurite retraction in aconcentration-dependent manner in N1E-115 cells, andthe concentrations of HU-210 applied to induce neuriteretraction were in the nanomolar range. To confirm theinvolvement of CB1 cannabinoid receptors in this process,a specific CB1 antisense ODN was employed. Pretreatingthe cells with the antisense ODN prevented the HU-210-induced neurite retraction. Because the inhibition offorskolin-stimulated cAMP accumulation by HU-210 wasalso blocked by the antisense pretreatment, these resultsdemonstrate strongly that functional CB1 cannabinoid

receptors were knocked down significantly by the anti-sense pretreatment and the neurite remodeling processinduced by HU-210 was mediated by specific CB1 can-nabinoid receptors.

Fig. 4. Abolishment of the inhibitory effect of HU-210 on forskolin-stimulated cAMP accumulation by CB1 antisense oligodeoxynucle-otide (ODN) in mouse neuroblastoma N1E-115 cells. N1E-115 cellswere pretreated for 48 hr with 3.0 mM antisense or sense ODNs andused for forskolin-stimulated cAMP accumulation assay. The cells werepreincubated with RO-20-1724 for 10 min at 37°C and then wereincubated with 10 nM HU-210 plus 2.0 mM forskolin for 10 min. ThecAMP levels were measured by radioimmunoassays. HU-210-inducedinhibition of forskolin-stimulated cAMP accumulation was detected incontrol cells (no ODN pretreatment) and sense ODN-pretreatedgroups (**P , 0.01), whereas the effect of HU-210 was not detectedin antisense ODN-pretreated cells (P . 0.05). Data represent mean 6SEM of three experiments.

Fig. 5. Inhibition of HU-210-induced neurite retraction by CB1 an-tisense ODN in mouse neuroblastoma N1E-115 cells. N1E-115 cellswere pretreated with 3.0 mM antisense or sense ODN for 48 hr.HU-210 at 10 nM was used to induce the neurite retraction followingantisense or sense ODN pretreatment. Pretreating the cells with anti-sense or sense ODN alone did not alter the serum starvation-inducedmorphological differentiation in N1E-115 cells (P . 0.05). HU-210-induced neurite remodeling was detected in control cells (no ODNpretreatment) and sense ODN-pretreated groups (**P , 0.01),whereas the effect of HU-210 was not detected in antisense ODN-pretreated cells (P . 0.05). Data shown are mean 6 SEM of threeexperiments.

Fig. 6. Inhibition of HU-210-induced neurite retraction by 8-Br-cAMP in mouse neuroblastoma N1E-115 cells. Morphologically dif-ferentiated N1E-115 cells were pretreated with the membrane-permeable cAMP analogue 8-Br-cAMP at 0.01, 0.10, and 1.00 mM for30 min at 37°C. HU-210 was then added into the culture medium ata final concentration of 10 nM. After 15 min of incubation, the cellswere fixed and stained with Alexa-488 fluorescent-labeled phalloidin.Cells were counted, and the data shown are mean 6 SEM of at leastthree experiments (*P , 0.05; **P , 0.01).

CB1 Receptor and Neurite Remodeling 351

A consequence of CB1 cannabinoid receptor activa-tion is the inhibition of adenylate cyclase activity andlowering of intracellular cAMP levels. Previously it hasbeen demonstrated that intracellular cAMP is an importantfactor for regulating neurite formation and stability(Richter-Landsberg and Jastorff, 1986; Wang and Zheng,1998). In this study, a membrane-permeable cAMP ana-logue, 8-Br-cAMP, was used to test the hypothesis thatthe decline of intracellular cAMP levels contributes toHU-210-induced neurite remodeling. The HU-210-induced neurite retraction was inhibited concentrationdependently by 8-Br-cAMP. This result suggests that theneurite retraction can be mediated, at least in part, throughthe HU-210-induced reduction of cAMP levels. 8-Br-cAMP was not able to block completely the neurite re-traction induced by HU-210, indicating that signal trans-duction pathways other than that of adenylate cyclasemight also be involved. Obviously, the signal transductionpathways downstream to and in parallel with the loweredintracellular cAMP levels remain to be elucidated in futurestudies.

One of the cellular mechanisms of neurogenesis andneuronal information processing, such as that in neuronaldevelopment, learning, and memory, is the formation ofnew synapses or the remodeling of existing ones (Chenand Tonegawa, 1997; Toni et al., 1999; Martin et al.,2000). In turn, it would be expected that the disturbanceof synaptogenesis and/or synaptic remodeling in a givenbrain area, such as the hippocampus, would cause learningand memory deficiency. By using receptor autoradiogra-phy, in situ hybridization, and immnunohistochemistrytechniques, high levels of CB1 cannabinoid receptor ex-pression have been found in the hippocampus (Herken-ham et al., 1991; Matsuda et al., 1993; Tsou et al., 1998).These previous localization data, plus the neurite remod-eling results from the current study, suggest that the CB1cannabinoid receptor plays a role in disturbing the synaptictargeting and remodeling in hippocampus. The develop-mental expression pattern of the CB1 cannabinoid recep-tor has also been investigated (Buckley et al., 1998;Fernandez-Ruiz et al., 2000). The fact that CB1 canna-binoid receptors are expressed in the embryonic centralnervous system in different developmental stages, togetherwith our new finding that these receptors can mediateneurite remodeling, implicates the biological function ofthis receptor in regulating neuronal morphogenesis duringdevelopment.

In summary, this study demonstrates that in N1E-115 cells the cannabinoid agonist HU-210 induces neuriteretraction. This effect of HU-210 is concentration depen-dent, can be blocked by antisense treatment, and is inhib-ited by 8-Br-cAMP treatment. These data indicate thatHU-210-induced neurite retraction is mediated by CB1cannabinoid receptors partially through inhibiting adenyl-ate cyclase activity. This newly discovered morphoregu-latory function of CB1 cannabinoid receptors might be anovel mechanism that mediates the neurotoxic effects ofcannabinoids in developing and adult brain. The down-

stream signaling pathways and the pharmacological/neurotoxicological significance of this CB1 cannabinoidreceptor-mediated neurite remodeling remain to be stud-ied further.

REFERENCESBuckley NE, Hansson S, Harta G, Mezey E. 1998. Expression of the CB1

and CB2 receptor messenger RNAs during embryonic development inthe rat. Neuroscience 82:1131–1149.

Chen C, Tonegawa S. 1997. Molecular genetic analysis of synaptic plas-ticity, activity-dependent neural development, learning, and memory inthe mammalian brain. Annu Rev Neurosci 20:157–184.

Felder CC, Glass M. 1998. Cannabinoid receptors and their endogenousagonists. Annu Rev Pharmacol Toxicol 38:179–200.

Felder CC, Veluz JS, Williams HL, Briley EM, Matsuda LA. 1992. Can-nabinoid agonists stimulate both receptor- and nonreceptor-mediatedsignal transduction pathways in cells transfected with and expressingcannabinoid receptor clones. Mol Pharmacol 42:838–845.

Fernandez-Ruiz J, Berrendero F, Hernandez ML, Ramos JA. 2000. Theendogenous cannabinoid system and brain development. Trends Neurosci23:14–20.

Herkenham M, Lynn AB, Johnson MR, Melvin LS, de Costa BR, RiceKC. 1991. Characterization and localization of cannabinoid receptors inrat brain: a quantitative in vitro autoradiographic study. J Neurosci11:563–583.

Hirose M, Ishizaki T, Watanabe N, Uehata M, Kranenburg O, MoolenaarWH, Matsumura F, Maekawa M, Bito H, Narumiya S. 1998. Moleculardissection of the Rho-associated protein kinase (p160ROCK)-regulatedneurite remodeling in neuroblastoma N1E-115 cells. J Cell Biol 141:1625–1636.

Howlett AC. 1995. Pharmacology of cannabinoid receptors. Annu RevPharmacol Toxicol 35:607–634.

Huether G, Zhou D, Ruther E. 1997. Causes and consequences of the lossof serotonergic presynapses elicited by the consumption of 3,4-methylenedioxymethamphetamine (MDMA, “ecstasy”) and its conge-ners. J Neural Transm 104:771–794.

Hunter SA, Burstein SH. 1997. Receptor mediation in cannabinoid stim-ulated arachidonic acid mobilization and anandamide synthesis. Life Sci60:1563–1573.

Kiosses BW, Tahir SK, Kalnins VI, Zimmerman AM. 1990. The effects ofdelta-9-tetrahydrocannabinol on actin microfilaments. Cytobios 63:23–29.

Kranenburg O, Poland M, van Horck FP, Drechsel D, Hall A, MoolenaarWH. 1999. Activation of RhoA by lysophosphatidic acid andGalpha12/13 subunits in neuronal cells: induction of neurite retraction.Mol Biol Cell 10:1851–1857.

Lawston J, Borella A, Robinson JK, Whitaker-Azmitia PM. 2000. Changesin hippocampal morphology following chronic treatment with the syn-thetic cannabinoid WIN 55,212-2. Brain Res 877:407–410.

Levitt P, Harvey JA, Friedman E, Simansky K, Murphy EH. 1997. Newevidence for neurotransmitter influences on brain development. TrendsNeurosci 20:269–274.

Martin SJ, Grimwood PD, Morris RG. 2000. Synaptic plasticity andmemory: an evaluation of the hypothesis. Annu Rev Neurosci 23:649–711.

Matsuda LA, Lolait SJ, Brownstein MJ, Young AC, Bonner TI. 1990.Structure of a cannabinoid receptor and functional expression of thecloned cDNA. Nature 346:561–564.

Matsuda LA, Bonner TI, Lolait SJ. 1993. Localization of cannabinoidreceptor mRNA in rat brain. J Comp Neurol 327:535–550.

Munro S, Thomas KL, Abu-Shaar M. 1993. Molecular characterization ofa peripheral receptor for cannabinoids. Nature 365:61–65.

Pertwee RG. 1997. Pharmacology of cannabinoid CB1 and CB2 receptors.Pharmacol Ther 74:129–180.

352 Zhou and Song

Richter-Landsberg C, Jastorff B. 1986. The role of cAMP in nerve growthfactor promoted neurite outgrowth in PC12 cells. J Cell Biol 102:821–829.

Scallet AC. 1991. Neurotoxicology of cannabis and THC—a review ofchronic exposure studies in animals. Pharmacol Biochem Behav 40:671–676.

Song ZH, Bonner TI. 1996. A lysine residue of the cannabinoid receptoris critical for receptor recognition by several agonists but notWIN55212-2. Mol Pharmacol 49:891–896.

Tahir SK, Zimmerman AM. 1991. Influence of marijuana on cellularstructures and biochemical activities. Pharm Pharmacol Biochem Behav40:617–623.

Tahir SK, Trogadis JE, Stevens JK, Zimmerman AM. 1992. Cytoskeletalorganization following cannabinoid treatment in undifferentiated anddifferentiated PC12 cells. Biochem Cell Biol 70:1159–1173.

Toni N, Buchs PA, Nikonenko I, Bron CR, Muller D. 1999. LTPpromotes formation of multiple spine synapses between a single axonterminal and a dendrite. Nature 402:421–425.

Tsou K, Brown S, Sanudo-Pena MC, Mackie K, Walker JM. 1998.Immunohistochemical distribution of cannabinoid CB1 receptors in therat central nervous system. Neuroscience 83:393–411.

Wang Q, Zheng JQ. 1998. cAMP-mediated regulation of neurotrophin-induced collapse of nerve growth cones. J Neurosci 18:4973–4984.

Wilson W, Mathew R, Turkington T, Hawk T, Coleman RE, ProvenzaleJ. 2000. Brain morphological changes and early marijuana use: a magneticresonance and positron emission tomography study. J Addict Dis 19:1–22.

Zhou D, Apostolakis EM, O’Malley BW. 1999. Distribution of D(5)dopamine receptor mRNA in rat ventromedial hypothalamic nucleus.Biochem Biophys Res Commun 266:556–559.

CB1 Receptor and Neurite Remodeling 353