logy 548 (2006) 21–28www.elsevier.com/locate/ejphar

European Journal of Pharmaco

Modulation of agmatine on calcium signal in morphine-dependent CHO cellsby activation of IRAS, a candidate for imidazoline I1 receptor

Ning Wu a, Rui-Bin Su a, Yin Liu a, Xin-Qiang Lu a, Jian-Quan Zheng a, Bin Cong b, Jin Li a,⁎

a Beijing Institute of Pharmacology and Toxicology, 27th Taiping Road, Beijing 100850, Chinab Hebei Medical University, Shijiazhuang 050017, China

Received 18 March 2006; received in revised form 14 July 2006; accepted 17 July 2006Available online 22 July 2006

Abstract

The present study investigated the effects of agmatine action on imidazoline I1 receptor antisera-selected protein (IRAS), a candidate forimidazoline I1 receptor, on prolonged morphine-induced adaptations of calcium signal and long-lasting alterations in gene expression to furtherelucidate the role of IRAS in opioid dependence. Two cell lines, Chinese hamster ovary cells expressing μ opioid receptor alone (CHO-μ) andexpressing μ opioid receptor and IRAS together (CHO-μ/IRAS), were used. After chronic treatment with morphine for 48 h, naloxone induced asignificant elevation of intracellular calcium concentration ([Ca2+]i) in CHO-μ and CHO-μ/IRAS cells. Agmatine (0.01–3 μM) concentration-dependently inhibited the naloxone-precipitated [Ca2+]i elevation when co-pretreated with morphine in CHO-μ/IRAS, but not in CHO-μ.Efaroxan, an imidazoline I1 receptor-preferential antagonist, completely reversed the effect of agmatine in CHO-μ/IRAS. Agmatine (1–10 μM)administration after chronic morphine exposure for 48 h partially decreased the [Ca2+]i elevation in CHO-μ/IRAS which was entirely antagonizedby efaroxan, but not in CHO-μ. In addition, agmatine (1 μM) co-pretreated with morphine attenuated the naloxone-precipitated increases ofcAMP-responsive element binding protein and extracellular signal-regulated kinase 1/2 phosphorylations and c-Fos expression in CHO-μ/IRAS.These effects were blocked by efaroxan as well. Taken together, these results indicate that the agmatine-IRAS action system attenuates the up-regulations of Ca2+ signal and its downstream gene expression in morphine-dependent model in vitro, providing additional evidence to support thecontribution of IRAS to opioid dependence.© 2006 Elsevier B.V. All rights reserved.

Keywords: Opioid dependence; IRAS; Imidazoline I1 receptor; Agmatine; Intracellular Ca2+ concentration; CREB phosphorylation; ERK phosphorylation; c-Fosexpression

1. Introduction

Opioid dependence is a chronic, relapsing disorder in thebrain, which involves a long-lasting neuroadaptation in thecentral nervous system. Recently, it is suggested that intracel-lular signal pathways play an important role in opioid de-pendence. It has been demonstrated that Ca2+ signal pathwayadaptations are crucial to opioid dependence. The significantincrease in the intracellular Ca2+ level has been observed invitro and in vivo in case of chronic exposure to opioids oropioids withdrawal (Yamamoto et al., 1978; Stiene-Martin etal., 1993; Spencer et al., 1997; Xie et al., 2002; Kong et al.,

⁎ Corresponding author. Tel.: +86 10 66932681; fax: +86 10 68211656.E-mail address: jinli9802@yahoo.com (J. Li).

0014-2999/$ - see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.ejphar.2006.07.013

2004; Zang et al., 2000). The disruption in intracellular Ca2+

homeostasis leads to activations of mitogen-activated proteinkinase/extracellular signal-regulated kinase (MAPK/ERK) cas-cade and certain transcription factors, such as cAMP-responsiveelement binding protein (CREB) and activator protein 1 (AP-1)(Bilecki and Przewlocki, 2000; Przewlocki, 2004). Throughcertain transcription factors, the transient responses of Ca2+

second messenger transform into long-lasting changes in geneexpression that underlies the neural plasticity and behavioraladaptations in opioid dependence.

Agmatine, an endogenous ligand to imidazoline receptors (I1and I2 subtypes), was initially isolated from the mammaliancentral nervous system in 1994 (Li et al., 1994). Endogenously,agmatinemay play a role as a neurotransmitter or neuromodulator(Reis and Regunathan, 2000). It acts by antagonizing N-methyl-

22 N. Wu et al. / European Journal of Pharmacology 548 (2006) 21–28

D-aspartate (NMDA) receptor, inhibiting nitric oxide synthase(NOS), and binding to imidazoline receptors and α2-adrenocep-tors (Reis and Regunathan, 2000). Recently, agmatine has beenfound involved in attenuating opioid dependence in vivo(Aricioglu-Kartal and Uzbay, 1997; Li et al., 1999; Morgan et al.,2002; Wei et al., 2005). On the other hand, prolonged exposureto morphine decreased endogenous agmatine level and itsbiosynthetic enzyme (arginine decarboxylase) activity in ratbrain, both of which were further decreased by naloxoneprecipitation (Aricioglu-Kartal and Regunathan, 2002). Al-though a few studies showed that idazoxan, an imidazolinereceptors/α2-adrenoceptors-mixed antagonist, could reverseagmatine's inhibitory effects on opioid dependence (Li et al.,1999; Wei et al., 2005), it does not ascertain that imidazolinereceptors mediated these effects, because it is well known thatα2-adrenoceptor is closely associated with opioid dependence.Thus, whether imidazoline receptors, especially imidazoline I1receptor, participate in the modulation by agmatine of opioiddependence remains unclear.

A strong candidate for imidazoline I1 receptor, known asimidazoline I1 receptor antisera-selected protein (IRAS), hasbeen cloned from human hippocampus (Piletz et al., 2000).Several lines of evidences support the identity of nativeimidazoline I1 receptor and cloned IRAS in tissue distributions,ligand binding properties and some cellular functions (Piletzet al., 2000, 1999; Dontenwill et al., 2003; Dupuy et al., 2004).Furthermore, our recent study showed the signal pathwayscoupled to IRAS were similar to those coupled to imidazoline I1receptor (Li et al., 2006). We have established a preliminaryrelationship between IRAS and opioid dependence obtainedfrom CHO cell line that stably co-expresses μ opioid receptorand IRAS (Wu et al., 2005). This cell line lacks endogenous α2-adrenoceptors, I2-imidazoline receptor, NMDA receptors andNOS (Fraser et al., 1989; Zhao et al., 2004; Uchino et al., 2001;Wu et al., 2005). Thus, the interferences from agmatine actionwith these targets can be ruled out. In this cell model, we havefound that IRAS is necessary for agmatine to attenuate thecompensatory up-regulation in cAMP signal pathway in mor-phine dependence (Wu et al., 2005).

However, the modulation by agmatine action on IRAS ofother signal adaptations and gene expressions during opioiddependence is unknown. To ascertain the role of IRAS, orimidazoline I1 receptor, in opioid dependence, more evidence isneeded. The purpose of the present study, therefore, is to in-vestigate the effects of agmatine-IRAS system on prolongedmorphine-induced adaptations of calcium signal, another im-portant pathway in opioid dependence, and its downstreamfactors (ERK, CREB, and c-Fos) to further elucidate the role ofIRAS in opioid dependence.

2. Materials and methods

2.1. Materials

The creations of CHO-μ and CHO-μ/IRAS cell lines havebeen described previously (Wu et al., 2005). RPMI 1640 me-dium and geneticin were purchased from Invitrogen Corpora-

tion (Gibco™, Grand Island, NY, USA). Hygromycin B waspurchased from Roche Diagnostics Gmbh (Roche, Mannheim,Germany). Fetal bovine serum was purchased from HyClone-Pierce (HyClone®, Sonth Logan, UT, USA). Fura-2/AM wasthe product of Biotium (Glowing Products for Science™,Hayward, CA, USA). Rabbit anti-phosphoSer131-CREB anti-body was obtained from Upstate (Upstate, New York, USA).Mouse anti-phosphoTyr204-ERK1/2, rabbit anti-ERK1/2 andanti-c-Fos, goat anti-actin antibodies and enhanced chemilu-minscence (ECL) detection kit were purchased from Santa CruzBiotechnology (Santa Cruz, CA, USA). Horseradish peroxi-dase-conjugated secondary antibodies were purchased fromZhongshan Biotechnology Co (Beijing, China). Nitrocellulosemembrane was purchased from Amersham Biosciences(Hybond, Sweden). Agmatine, naloxone, probenecid andprotease inhibitors were purchased from Sigma ChemicalCompany (Sigma®, St. Louis, MO, USA). Efaroxan waspurchased from Research Biochemicals International (RBI,Natick, MA, USA). Morphine was purchased from QinghaiPharmaceutic Factory (Xining, China).

2.2. Cell cultures

CHO-μ cells were cultured in RPMI 1640 supplementedwith 10% heat-inactivated fetal bovine serum, 100 U/mlpenicillin, 100 U/ml streptomycin, and 50 μg/ml hygromycinB at 37 °C with humidified atmosphere consisting of 95% airand 5% CO2. The medium for CHO-μ/IRAS cells was thesame as that for CHO-μ cells except for 200 μg/ml geneticincontained.

2.3. Measurement of intracellular calcium concentrations

Cells were cultured in 6-well plates for 48 h to reach 80%confluence. If necessary, vehicle (saline) or drugs were added inthe meantime. After 48 h, cells were harvested by exposure to0.02% EDTA in phosphate-buffered saline. After being washedtwice, cells were loaded with 4 μM fura-2/AM at 37 °C for30 min. Subsequently, cells were washed three times and re-suspended in Hepes-buffered balanced salt solution (145 mMNaCl, 5 mM KCl, 1.3 mM MgCl2, 1.2 mM CaCl2, 1.2 mMNaH2PO4, 10 mM glucose, 20 mM Hepes, pH 7.4) containing1 mM probenecid and 0.1% bovine serum albumin. Probenecidwas added to prevent loss of fura-2 via the anion transporter (DiVirgilio et al., 1990). And then, cells were incubated in naloxonefor 15 min to induce withdrawal and intracellular calciumconcentration ([Ca2+]i) was measured on a fluorescence spectro-photometer (F4500, Hitachi, Japan) with 340 and 380 nmexcitation wavelengths and 510 nm emission wavelength. [Ca2+]iwas calculated as [Ca2+]i=Kd* (R−Rmin) / (Rmax−R)*Fmin(380) /Fmax(380), with a value of 224 nM for Kd. R is the ratio of thefluorescence intensities excited by 340 and 380 nM, beingcorrected for background fluorescence. Fmax was obtained byadding triton X-100 with 0.1% final concentration to saturate thecytoplasmic fura-2 with Ca2+, and Fmin was obtained by addingEGTA with 10 mM final concentration (pHN8.5) to completelychelate Ca2+.

Fig. 1. Chronic agmatine administration prevented morphine-pretreated naloxone-precipitated [Ca2+]i elevation. (A) The effects of agmatine co-pretreated withmorphine on naloxone-induced [Ca2+]i elevation in CHO-μ and CHO-μ/IRAScells. Cells were pretreated with vehicle plus morphine or agmatine plus morphinefor 48 h. After the load of fura-2, cells were incubated in naloxone for 15 min andmeasured fluorescence. n=9 for CHO-μ and n=10 for CHO-μ/IRAS. ##Pb0.01,significantly different from vehicle-treated control group by t-test; **Pb0.01,significantly different from morphine-treated group (agmatine 0 μM) by one-wayANOVA followed Dunnett's t test. (B) Antagonism by efaroxan of the inhibitoryeffect of agmatine on [Ca2+]i elevation in CHO-μ/IRAS. Efaroxan was added withmorphine and agmatine together. n=10. **Pb0.01, significantly different fromefaroxan absence group by t-test.

23N. Wu et al. / European Journal of Pharmacology 548 (2006) 21–28

2.4. Western blots

Protein phosphorylations of CREB and ERK1/2 and ex-pression of c-Fos were analyzed by Western blots. Cells wereplated into 10 cm dishes and treatedwith vehicle (saline) or drugsfor 48 h. After removal of drugs, naloxone was added for 20 min.The stimulation was terminated by two quick washes in ice-coldphosphate-buffered saline and the addition of lysis buffer(50 mM Tris–HCl, 150 mM NaCl, 1% NP40, 1 mM EDTA,1 mM EGTA, 1 μg/ml aprotinin, 1 μg/ml pepstetin, 1 μg/mlleupeptin, 1 mM phenylmethyl sulfonylfluoride (PMSF), 1 mMdithiothreitol, 1 mM NaF, 1 mM Na3VO4, pH 7.4). Cell lysateswere cleared by centrifugation (4 °C, 12,000 g) for 20 min. Aftermeasurement of protein concentration by Brad-ford method, celllysates were denatured with 6× Laemmli buffer, and frozen at−70 °C. Proteins were separated with 12% polyacrylamide gelelectrophoresis and transferred to nitrocellulose membranes.Membranes were blocked in 3% dry milk powder and 0.05%tween-20 in Tris-buffered saline, and incubated with the primaryantibodies (anti-phospho-CREB, anti-phospho-ERK1/2 or anti-c-Fos antibody) at 4 °C overnight. Parallel immunoblots wereprepared to detect the actin or total ERK1/2 to confirm the equalamount of protein per lane. After incubation with correspondingsecondary antibodies, bands were visualized using the ECL kitaccording to the manufacturer's instructions. Western blots werescanned and analyzed by beta4.0.2 of Scion Image software.

2.5. Data analysis

Data were expressed as mean±S.D. and n referred to sep-arate cultures and experiments. Significant differences weredetermined by t-test or one-way analysis of variance (ANOVA)with Dunnett's t test for two groups or multiple comparisons,respectively.

3. Results

3.1. Effects of agmatine on the resting [Ca2+]i

We first investigated the effects of agmatine on [Ca2+]i in theabsence of morphine. In CHO-μ/IRAS cells and CHO-μ cells,agmatine at 10 nM–10 μMdid not alter the resting [Ca2+]i within5min (data not shown), suggesting no acute stimulating action ofagmatine on [Ca2+]i. When agmatine (10 nM–3 μM) pretreatedcells for 48 h, the resting [Ca2+]i did not change in CHO-μ, whileit decreased slightly in CHO-μ/IRAS at agmatine 3 μM.

3.2. Agmatine inhibiting chronic morphine-pretreatednaloxone-precipitated [Ca2+]i elevation via IRAS

After administration of morphine (10 μM) for 48 h, naloxone(100 μM, 15 min) elicited a significant elevation of [Ca2+]i by33.1% in CHO-μ and 33.8% in CHO-μ/IRAS, respectively.When extracellular Ca2+ was omitted, the [Ca2+]i elevation wasabolished completely, suggesting that extracellular Ca2+ influxplays a key role in naloxone-precipitated [Ca2+]i elevation inmorphine-dependent cells.

Chronic concurrent pretreatment of CHO-μ/IRAS cells withagmatine (10 nM–3 μM) during morphine exposure for 48 hdecreased naloxone-induced [Ca2+]i elevation in a concentra-tion-dependent manner (Fig. 1A). A complete return to the basallevel was obtained in the presence of 1 and 3 μM agmatine. InCHO-μ cells, however, agmatine from 10 nM to 3 μM failed to

Fig. 2. Acute agmatine administration inhibited morphine-pretreated naloxone-precipitated [Ca2+]i elevation. (A) The effects of agmatine treated aftermorphine onnaloxone-induced [Ca2+]i elevation in CHO-μ and CHO-μ/IRAS cells. Cells werepretreated with vehicle or morphine for 48 h. After the load of fura-2, cells weretreated by agmatine for 1 min before incubation with naloxone. n=8 for CHO-μand n=9 for CHO-ì/IRAS. ##Pb0.01, significantly different from vehicle-treatedcontrol group by t-test; *Pb0.05 and **Pb0.01, significantly different frommorphine-treated group (agmatine 0 ìM) by one-way ANOVA followed Dunnett'st test. (B) Antagonism by efaroxan of the inhibitory effect of agmatine on [Ca2+]ielevation in CHO-ì/IRAS. Efaroxan was added for 1 min before agmatine. n=8.**Pb0.01, significantly different from Mor+Agm group by t-test.

Fig. 3. Effect of agmatine chronic co-pretreated with morphine on naloxone-precipitated phosphorylated-CREB increase in CHO-μ/IRAS. Cells were treatedwith drugs for 48 h. After removal of drugs, naloxone was added for 15 min.n=6. **Pb0.01, significantly different from morphine group by t-test;+Pb0.05, significantly different from morphine+agmatine group by t-test.

24 N. Wu et al. / European Journal of Pharmacology 548 (2006) 21–28

inhibit morphine-pretreated naloxone-precipitated [Ca2+]i ele-vation (Fig. 1A). This result suggests that IRAS might benecessary for agmatine to prevent naloxone-precipitated [Ca2+]ielevation in morphine-dependent cells. For lack of endogenousα2-adrenoceptors in CHO cells (Fraser et al., 1989), an I1/α2

receptors-mixed antagonist efaroxan was used to furtherexamine the role of IRAS in mediating the effects of agmatineon the [Ca2+]i elevation. Pretreatment of CHO-μ/IRAS cellswith efaroxan (50 μM, 48 h) had no effect on either resting [Ca2+]i or morphine-pretreated naloxone-precipitated [Ca2+]i eleva-tion, but it entirely blocked the inhibitory effect of agmatine asmentioned above (Fig. 1B).

After chronic morphine (10 μM) exposure for 48 h, agmatine(10 nM–10 μM) administration for 1 min before naloxone didnot alter the naloxone-induced [Ca2+]i elevation in CHO-μ cells(Fig. 2A). But in CHO-μ/IRAS cells, agmatine at 1 and 10 μMattenuated the [Ca2+]i elevation (Fig. 2A). In contrast to 35.2%increase in the absence of agmatine, the increase of [Ca2+]i inthe presence of 1 and 10 μM agmatine was reduced to 23.4%and 15.8%, respectively. And this inhibitory effects were com-pletely reversed by 50 μM efaroxan (pretreatment for 1 minbefore agmatine) (Fig. 2B). Taken together, these findingsindicate that agmatine by activation of IRAS not only preventsthe development of [Ca2+]i elevation, but partially inhibits theexpression of [Ca2+]i elevation in prolonged morphine-treatedcells as well.

3.3. Effects of chronic agmatine administration on thephosphorylations of CREB and ERK1/2 in morphine-depen-dent CHO-μ/IRAS cells

Opioids regulate CREB phosphorylation level and binding tosome promoters of several genes implicated in drug addition.Both cAMP pathway and calcium pathway can phosphorylateCREB to up-regulate its transcriptional activity (Bilecki andPrzewlocki, 2000; Przewlocki, 2004). Baseline experimentswere conducted in the absence of morphine. Neither agmatine(1 μM) nor efaroxan (50 μM) treatment alone for 48 h alteredthe level of phosphorylated CREB in CHO-μ/IRAS cells (datanot shown). Administration of morphine (10 μM) for 48 h hadno effect on CREB phosphorylation, but naloxone precipitation(100 μM, 20 min) significantly increased the phosphorylatedCREB to three times that of control (Fig. 3). Agmatine (1 μM)co-pretreated with morphine for 48 h decreased the naloxone-induced CREB phosphorylation (Fig. 3). Efaroxan (50 μM,48 h) did not alter morphine-pretreated naloxone-precipitatedCREB phosphorylation, but it entirely reversed the effect of

Fig. 5. Effect of agmatine chronic co-pretreated with morphine on naloxone-precipitated c-Fos expression increase in CHO-μ/IRAS. Cells were treated withdrugs for 48 h. After removal of drugs, naloxone was added for 15 min. n=5.*Pb0.05, significantly different from morphine group by t-test; +Pb0.05,significantly different from morphine+agmatine group by t-test.

25N. Wu et al. / European Journal of Pharmacology 548 (2006) 21–28

agmatine (Fig. 3). This indicates that transfected IRAS mediatesthis inhibitory effect of agmatine in this cell line.

MAPK cascade is another important pathway in opioid de-pendence. It can integrate signals from secondmessenger systems.ERK1/2, themost abundantmembers ofMAPK family in neurons,is activated by Ca2+–CaMKs–Ras–Raf–MEK pathway (Przew-locki, 2004). In CHO-μ/IRAS cells, neither the total ERK1/2 levelnor the phosphorylated ERK1/2 level was changed when exposedto agmatine (1 μM), efaroxan (50 μM) or morphine (10 μM) for48 h (data not shown). Like CREB phosphorylation, ERK1/2phosphorylation in morphine-pretreated CHO-μ/IRAS cells in-creased to 367% after naloxone precipitation, while the totalERK1/2 remained unchanged (Fig. 4). Concurrent administrationof agmatine (1 μM) and morphine for 48 h reduced the phos-phorylation of ERK1/2, and this response was entirely abolishedby imidazoline I1 receptor antagonist efaroxan (Fig. 4).

3.4. Effect of chronic agmatine administration on c-Fosexpression in morphine-dependent CHO-μ/IRAS cells

As is known, c-Fos, a component of transcription factor AP-1,is one of the targets modulated by phosphorylated CREB. It has

Fig. 4. Effect of agmatine chronic co-pretreated with morphine on naloxone-precipitated phosphorylated-ERKs increase in CHO-ì/IRAS. Cells were treatedwith drugs for 48 h. After removal of drugs, naloxone was added for 15 min.n=5. *Pb0.05, significantly different from morphine group by t-test; +Pb0.05,significantly different from morphine+agmatine group by t-test.

been reported that prolonged morphine administration induces aweak expression in c-Fos, while naloxone-precipitated withdraw-al significantly increases its expression in several brain areas(Bilecki and Przewlocki, 2000; Przewlocki, 2004). Chronicexposure to agmatine (1 μM), efaroxan (50 μM) or morphine(10 μM) for 48 h also had no effect on c-Fos expression in CHO-μ/IRAS cells (data not shown). After pretreatment with morphinefor 48 h, the c-Fos expression in naloxone precipitation was twicethat of control (Fig. 5), which is consistent with the previousstudies. Agmatine (1 μM) pretreatedwithmorphine recovered theincreased c-Fos to the basal level (Fig. 5). This effect was blockedby imidazoline I1 receptor antagonist efaroxan as well, suggestingthe participation of IRAS.

4. Discussion

More and more studies demonstrate that changes in calciumsignal and some related gene expressions are crucial to opioiddependence. Our present study showed firstly that agmatine byactivation on IRAS, a candidate for imidazoline I1 receptor,attenuated the adaptive changes in calcium signal and itsdownstream gene expressions in morphine-dependent CHO-μ/IRAS cells. This further elucidates the role of IRAS in opioiddependence.

It has been reported that prolonged morphine treatment ormorphine withdrawal by naloxone precipitation elicited [Ca2+]iin vivo or in vitro (Yamamoto et al., 1978; Stiene-Martin et al.,1993; Spencer et al., 1997; Xie et al., 2002; Kong et al., 2004;Zang et al., 2000). In chronic morphine-treated CHO-μ andCHO-μ/IRAS cells, we found naloxone significantly increased[Ca2+]i. Some studies showed that the mobilization ofintracellular Ca2+ stores results in [Ca2+]i elevation (Spenceret al., 1997; Xie et al., 2002), but others demonstrated thatextracellular Ca2+ influx might play a key role in the elevation(Stiene-Martin et al., 1993; Kong et al., 2004). In our present

26 N. Wu et al. / European Journal of Pharmacology 548 (2006) 21–28

study, naloxone-induced [Ca2+]i elevation was abolished by theomission of extracellular Ca2+, suggesting an influx of extra-cellular Ca2+. Because of the absence of endogenous NMDAreceptor in CHO cells (Uchino et al., 2001), the naloxone-precipitated extracellular Ca2+ influx seems to be throughvoltage-dependent Ca2+ channels. In fact, dihydropyridine-sensitive L-type rather than ω-conotoxin-sensitive N-type Ca2+

channel current was observed in native CHO cells (Skryma et al.,1994). So we presume that the response of [Ca2+]i to morphinewithdrawal was mediated by L-type Ca2+ channels in our cellmodels. Compared to previous reports, the degree of naloxone-precipitated [Ca2+]i elevation was lower than that in neurons orneuroglial cells. This may be explained by the low density of L-type Ca2+ channels in CHO cells (Skryma et al., 1994).

In IRAS-negative CHO-μ cells exposure to morphine for 48h, co-pretreatment with agmatine failed to prevent the [Ca2+]ielevation induced by naloxone precipitation. In contrast, inIRAS-positive CHO-μ/IRAS cells, agmatine concentration-dependently inhibited the [Ca2+]i elevation and this effect wasentirely blocked by imidazoline I1 receptor antagonist efaroxan.Additionally, after chronic morphine exposure, agmatinepartially inhibited the [Ca2+]i elevation in CHO-μ/IRAS cells,which was antagonized by efaroxan as well. Thus, our presentstudy provides the evidence for the participation of agmatineactivation on IRAS in suppressing development and expressionof calcium signal adaptation in morphine dependence. Amongthe targets of agmatine action, NMDA receptors, α2-adreno-ceptors, and NOS have been proved to be closely related toopioid dependence. As a putative imidazoline I1 receptor ligand,agmatine displays relatively high affinity to imidazoline I1receptor with 0.7 μM of IC50 value in rat cerebral cortex and33 nM of Ki value in human platelet (Li et al., 1994; Piletz et al.,1996). While the IC50 value of agmatine displacing [3H]PACfrom α2-adrenoceptors was 4 μM in rat cerebral cortex (Li et al.,1994), the Ki values of agmatine for three subtypes of α2-adrenoceptor ranged from 7 μM to 164.4 μM(Piletz et al., 1996).The IC50 value of agmatine blocking NMDA receptor currentwas more than 100 μM in rat hippocampal neurons (Yang andReis, 1999), and the Ki values of agmatine for three isoforms ofNOS were 660 μM, 220 μM and 7.5 mM (Galea et al., 1996).Because of the higher affinity to imidazoline I1 receptor than toα2-adrenoceptors, NMDA receptor and NOS, we think IRAS, orimidazoline I1 receptor, might play a relatively important role inagmatine's inhibitory effects on morphine dependence. Ourprevious and present studies showed both cAMP overshoot and[Ca2+]i elevation in chronic morphine-treated naloxone-precip-itated cells could be attenuated by agmatine action on IRASwithrelatively low concentrations (Wu et al., 2005).

It has been reported that agmatine inhibited voltage-dependent Ca2+ channel currents in isolated neurohypophysialnerve terminals (Wang et al., 2002). Further studies showedagmatine attenuated L-type Ca2+ current in cultured rat hippo-campal neurons and ventricular myocytes (Weng et al., 2003; Liet al., 2002). But whether the effect is due to agmatine's directblockade in L-type Ca2+ channels or indirect modulation to L-type Ca2+ channels by acting on imidazoline I1 receptor isunclear. Contradictory results were observed in imidazoline I1

receptor stimulation on cellular Ca2+ (Ernsberger, 1999). How-ever, a conclusion that is consistent with the previous studies isthat imidazoline I1 receptor activation definitely does not elicitthe calcium spikes associated with receptors coupled to eitherphosphatidylionsitol hydrolysis or the opening of cation chan-nels. This is consistent with our present finding that agmatinedid not cause a rise of internal Ca2+ in CHO-μ/IRAS cells.Since agmatine can attenuate L-type Ca2+ channel currents, wethought the partial inhibition of acute agmatine to naloxone-induced [Ca2+]i elevation might result from the blockade of L-type Ca2+ channels. According to our study, the blockade of L-type Ca2+ channels by agmatine seems to depend on thetransfected IRAS. Our previous study showed that high con-centrations of agmatine might modulate L-type Ca2+ channelsin an IRAS-independent manner (Wu et al., 2005), but theeffects of agmatine higher than 10 μM could not be observed inthe present study because we found high concentrations ofagmatine interfered fura-2 fluorescence, which agrees with theprevious report (Shepherd et al., 1996). As for the inhibition ofprolonged agmatine concurrent treatment with morphine, itsmechanism might be due to modulating the up-regulation of L-type Ca2+ channel density in morphine dependence. Prolongedadministration of L-type Ca2+ channel antagonist decreases,rather than increases, the density of brain dihydropyridine-sensitive binding sites (Panza et al., 1985). Long-term treatmentwith morphine leads to an increase in the site density andconcurrent L-type Ca2+ channel antagonist treatment preventsthis increase (Ramkumar and El-Fakahany, 1988; Michaluket al., 1998). Since agmatine functionally inhibits Ca2+ chan-nels, it is reasonable to presume that agmatine might prevent theup-regulation of Ca2+ channels in morphine dependence. How-ever, further studies are needed.

The stability of the behavioral abnormalities that characterizeopioid dependence indicates that drug-induced changes in geneexpression may be involved. Transcription factors are certain toplay a crucial role in the development of opioid dependence.CREB is one of the most important factors linking the opioid-regulated secondary messenger systems to alterations in geneexpression. PKA, CaMKs or MAPK can phosphorylate CREBat Ser133 to increase its transcriptional activity (Bilecki andPrzewlocki, 2000; Przewlocki, 2004). c-Fos, one of the targetgenes of CREB, constitutes another transcription factorimplicated in drug addiction AP-1. CREB mutant mice showedattenuated withdrawal signs of morphine physical dependenceand did not respond to the reinforcing properties of morphine ina CPP paradigm (Maldonado et al., 1996; Walters et al., 2005),and suppression of c-Fos expression with antisense oligonu-cleotides prevented the acquisition of morphine induced CPP(Tolliver et al., 2000). Recent studies suggest the involvementof ERK signal in action of morphine dependence (Bilecki andPrzewlocki, 2000; Przewlocki, 2004). Ca2+/CaMKs pathwaycauses ERKs activation. In addition, cAMP/PKA signalactivates or inhibits ERKs phosphorylation in different celllines. In neurons, PKA may be a positive regulator (Grewalet al., 1999). For the up-regulation of cAMP pathway in chronicexposure to morphine, PKA-dependent ERKs phosphorylationmay play an important role in MAPK cascade activation.

27N. Wu et al. / European Journal of Pharmacology 548 (2006) 21–28

Therefore, ERK, CREB and its target gene c-Fos integrate Ca2+

and cAMP second messenger systems, and transform short-lasting acute opioid signals into long-lasting alternations at thelevel of gene transcription. So we further investigated the effectsof agmatine-IRAS system on transcription factors. In our pre-vious and present studies, besides modulating cAMP and Ca2+

second messengers separately, agmatine-IRAS also modulatedintegration points for the two signals in morphine dependence.In morphine-dependent CHO-μ/IRAS cells, we found IRASmediated the inhibitory responses of agmatine to the increasesof ERK and CREB phosphorylations and c-Fos expression. Byregulating transcription factor CREB and immediate early geneproduct c-Fos, agmatine-IRAS system might result in the al-ternations of gene expression in response to opioid.

Taken together, we found agmatine action on IRAS, a strongcandidate for imidazoline I1 receptor, attenuated the elevation ofintracellular Ca2+ levels in morphine-dependent cells. Agma-tine-IRAS also reduced the increases of ERK1/2 and CREBphosphorylations and c-Fos expression. According to our pre-sent and previous studies, we speculate that IRAS, or imidazo-line I1 receptor, contributes to morphine dependence.

Acknowledgement

This research was supported by a grant from the NationalBasic Research Program of China (2003CB515400).

References

Aricioglu-Kartal, F., Regunathan, S., 2002. Effect of chronic morphinetreatment on the biosynthesis of agmatine in rat brain and other tissues.Life Sci. 71, 1695–1701.

Aricioglu-Kartal, F., Uzbay, I.T., 1997. Inhibitory effect of agmatine onnaloxone-precipitated abstinence syndrome in morphine dependent rats. LifeSci. 61, 1775–1781.

Bilecki, W., Przewlocki, R., 2000. Effect of opioids on Ca2+/cAMP responsiveelement binding protein. Acta Neurobiol. Exp. 60, 557–567.

Di Virgilio, F., Steinberg, T.H., Silverstein, S.C., 1990. Inhibition of Fura-2sequestration and secretion with organic anion transport blockers. CellCalcium 11, 57–62.

Dontenwill, M., Pascal, G., Piletz, J.E., Chen, M., Baldwin, J., Ronde, P.,Dupuy, L., Urosevic, D., Greney, H., Tadeda, K., Bousquet, P., 2003. IRAS,the human homologue of nischarin, prolongs survival of transfected PC12cell. Cell Death Differ. 10, 933–935.

Dupuy, L., Urosevic, D., Greney, H., Quaglia, W., Pigini, M., Brasili, L.,Dontenwill, M., Bousquet, P., 2004. I1 imidazoline receptor-mediated effectson apoptotic processes in PC12 cells. Cell Death Differ. 11, 1049–1052.

Ernsberger, P., 1999. The I1-imidazoline receptor and its cellular signalingpathways. Ann. N.Y. Acad. Sci. 881, 35–53.

Fraser, C.M., Arakawa, S., McCombie, R.W., Venter, C.J., 1989. Cloning,sequence analysis, and permanent expression of a human α2-adrenergicreceptor in Chinese hamster ovary cells. Evidence for independent pathwaysof receptor coupling to adenylate cyclase attenuation and activation. J. Biol.Chem. 264, 11754–11761.

Galea, E., Regunathan, S., Eliopoulos, V., Feinstein, D.L., Reis, D.J., 1996.Inhibition of mammalian nitric oxide synthases by agmatine, an endogenouspolyamine formed by decarboxylation of arginine. Biochem. J. 316, 247–249.

Grewal, S.S., York, R.D., Stork, P.J.S., 1999. Extracellular-signal-regulatedkinase signalling in neurons. Curr. Opin. Neurobiol. 9, 544–553.

Kong, Q., Chen, J.H., Gong, Z.H., 2004. Effects of Nù-nitro-L-arginine methylester, MK-801, nifedipine on the intracellular calcium concentration alternatedby morphine in NG108-15 cells. Chin. J. Pharmacol. Toxicol. 18, 264–268.

Li, G., Regunathan, S., Barrow, C.J., Eshraghi, J., Cooper, R., Reis, D.J., 1994.Agmatine: an endogenous clonidine-displacing substance in the brain.Science 263, 966–999.

Li, J., Li, X., Pei, G., Qin, B.Y., 1999. Effects of agmatine on tolerance to andsubstance dependence on morphine in mice. Acta Pharmacol. Sin. 20,232–238.

Li, Q., Yin, J.X., He, R.R., 2002. Effect of agmatine on L-type calcium current inrat ventricular myocytes. Acta Pharmacol. Sin. 23, 219–224.

Li, F., Wu, N., Su, R.B., Zheng, J.Q., Xu, B., Lu, X.Q., Cong, B., Li, J., 2006.Involvement of phosphatidylcholine-selective phospholipase C in activationof mitogen-activated protein kinase pathways in imidazoline receptorantisera-selected protein. J. Cell. Biochem. 98, 1615–1628.

Maldonado, R., Blendy, J.A., Tzavara, E., Gass, P., Roques, B.P., Hanoune, J.,Schutz, G., 1996. Reduction of morphine abstinence in mice with a mutationin the gene encoding CREB. Science 273, 657–659.

Michaluk, J., Karolewicz, B., Antkiewicz-Michaluk, L., Vetulani, J., 1998. Effectsof various Ca2+ channel antagonists on morphine analgesia, tolerance anddependence, and on blood pressure in the rat. Eur. J. Pharmacol. 352, 189–197.

Morgan, A.D., Campbell, U.C., Fons, R.D., Carroll, M.E., 2002. Effects ofagmatine on the escalation of intravenous cocaine and fentanyl self-administration in rats. Pharmacol. Biochem. Behav. 72, 873–880.

Panza, G., Grebb, J.A., Sanna, E., Wright Jr., A.G., Hanbauer, I., 1985. Evidencefor down-regulation of 3H-nitrendipine recognition sites in mouse brain afterlong-term treatment with nifedipine or verapamil. Neuropharmacology 24,1113–1117.

Piletz, J.E., Zhu, H., Chikkala, D.N., 1996. Comparison of ligand bindingaffinities at human I1-imidazoline binding sites and the high affinity state ofalpha-2 adrenoceptor subtypes. J. Pharmacol. Exp. Ther. 279, 694–702.

Piletz, J.E., Jones, J.C., Zhu, H., Bishara, O., Ernsberger, P., 1999. Imidazolinereceptor antisera-selected cDNA and mRNA distribution. Ann. N.Y. Acad.Sci. 881, 1–7.

Piletz, J.E., Ivanov, T.R., Sharp, J.D., Ernsberger, P., Chang, C.H., Pickard, R.T.,Gold, G., Roth, B., Zhu, H., Jones, J.C., Baldvin, J., Reis, D.J., 2000.Imidazoline receptor antisera-selected (IRAS) cDNA: cloning and charac-terization. DNA Cell Biol. 19, 319–329.

Przewlocki, R., 2004. Opioid abuse and brain gene expression. Eur. J.Pharmacol. 500, 331–349.

Ramkumar, V., El-Fakahany, E.E., 1988. Prolonged morphine treatmentincreases rat brain dihydropyridine binding sites: possible involvement indevelopment of morphine dependence. Eur. J. Pharmacol. 146, 73–83.

Reis, D.J., Regunathan, S., 2000. Is agmatine a novel neurotransmitter in brain?Trends Pharmacol Sci. 21, 187–193.

Shepherd, R.M., Hashmi, M.N., Kane, C., Squires, P.E., Dunne, M.J., 1996.Elevation of cytosolic calcium by imidazolines in mouse islets ofLangerhans: implications for stimulus–response coupling of insulin release.Br. J. Pharmacol. 119, 911–916.

Skryma, R., Prevarskaya, N., Vacher, P., Dufy, B., 1994. Voltage-dependent ionicconductances inChinese hamster ovary cells. Am. J. Physiol. 267, C544–C553.

Spencer, R.J., Jin, W., Thayer, S.A., Chakrabarti, S., Law, P.Y., Loh, H.H., 1997.Mobilization of Ca2+ from intracellular stores in transfected neuro2a cellsby activation of multiple opioid receptor subtypes. Biochem. Pharmacol. 54,809–818.

Stiene-Martin, A., Mattson, M.P., Hauser, K.F., 1993. Opiates selectivelyincrease intracellular calcium in developing type-1 astrocytes: role ofcalcium in morphine-induced morphologic differentiation. Dev. Brain Res.76, 189–196.

Tolliver, B.K., Sganga, M.W., Sharp, F.R., 2000. Suppression of c-fos inductionin the nucleus accumbens prevents acquisition but not expression ofmorphine-conditioned place preference. Eur. J. Neurosci. 12, 3399–3406.

Uchino, S., Watanabe, W., Nakamura, T., Shuto, S., Kazuta, Y., Matsuda, A.,Nakajima-Iijima, S., Kudo, Y., Kohsaka, S., Mishina, M., 2001. Establishmentof CHO cell lines expressing four N-methyl-D-aspartate receptor subtypes andcharacterization of a novel antagonist PPDC. FEBS Lett. 506, 117–122.

Walters, C.L., Godfrey, M., Li, X., Blendy, J.A., 2005. Alterations in morphine-induced reward, locomotor activity, and thermoregulation in CREB-deficient mice. Brain Res. 1032, 193–199.

Wang, G., Gorbatyuk, O.S., Dayanithi, G., Ouyang, W., Wang, J., Milner, T.A.,Regunathan, S., Reis, D.J., 2002. Evidence for endogenous agmatine in

28 N. Wu et al. / European Journal of Pharmacology 548 (2006) 21–28

hypothalamo-neurohypophysial tract and its modulation on vasopressinrelease and Ca2+ channels. Brain Res. 932, 25–36.

Wei, X.L., Su, R.B., Lu, X.Q., Liu, Y., Yu, S.Z., Yuan, B.L., Li, J., 2005.Inhibition by agmatine on morphine-induced conditioned place preferencein rats. Eur. J. Pharmacol. 515, 99–106.

Weng, X.C., Gai, X.D., Zheng, J.Q., Li, J., 2003. Agmatine blocked voltage-gated calcium channel in cultured rat hippocampal neurons. Acta Pharmacol.Sin. 24, 746–750.

Wu, N., Su, R.B., Xu, B., Lu, X.Q., Liu, Y., Zheng, J.Q., Piletz, J.E., Li, J., Qin,B.Y., 2005. IRAS, a candidate for I1-imidazoline receptor, mediatesinhibitory effect of agmatine on cellular morphine dependence. Biochem.Pharmacol. 70, 1079–1087.

Xie, Y., Yu, Z.P., Zhu, G.X., Fang, Q., Jiang, H.H., 2002. Mechanisms ofmorphine-evoked changes of intracellular calcium in primarily culturedhippocampal neurons. Chin. J. Appl. Physiol. 18, 124–127.

Yamamoto, H., Harris, R.A., Loh, H.H., Way, E.L., 1978. Effects of acute andchronic morphine treatments on calcium localization and binding in brain.J. Pharmacol. Exp. Ther. 205, 255–264.

Yang, X.C., Reis, D.J., 1999. Agmatine selectively blocks the N-methyl-D-aspartate subclass of glutamate receptor channels in rat hippocampalneurons. J. Pharmacol. Exp. Ther. 288, 544–549.

Zang, M., Shen, Q., Sun, Y., Wang, Q., Liu, J., 2000. Mechanisms of opioidreceptor-induced elevation in intracellular calcium by confocal laserscanning microscopy. Acta Acad. Med. Sin. 22, 254–258.

Zhao, Y., Su, R.B., Wu, N., Xu, B., Liu, Y., Lu, X.Q., Li, J., 2004. Stableexpression of imidazoline-1 receptor in mammalian cells. Bull. Acad. Mil.Med. Sci. 28, 333–336.