Reversal of Morphine Antinociceptive Tolerance and

Ca 2ϩ/Calmodulin-Dependent Protein Kinase II □

S

Lei Tang,Pradeep K.Shukla,Lili X.Wang,and Zaijie Jim Wang

Department of Biopharmaceutical Sciences (L.T.,P.K.S.,L.X.W.,Z.J.W.),and Cancer Center (Z.J.W.),University of Illinois,Chicago,Illinois;and Feinberg School of Medicine,Northwestern University,Chicago,Illinois (L.X.W.)

Received October 26,2005;accepted February 23,2006

ABSTRACT

Previous studies have suggested that Ca 2ϩ/calmodulin-depen-dent protein kinase II (CaMKII)can modulate opioid tolerance and dependence via its action on learning and memory.In this study,we examined whether CaMKII could directly regulate opioid tolerance and dependence.CaMKII activity was in-creased after the treatment with morphine (100mg/kg s.c.or 75mg s.c.of morphine/pellet/mouse);the effect exhibited a tem-poral correction with the development of opioid tolerance and dependence.In mice treated with morphine (100mg/kg s.c.),morphine tolerance and dependence developed in 2to 6h.An acute supraspinal administration of KN93[2-[N -(2-hydro-xyethyl)]-N -(4-methoxybenzenesulfonyl)]amino-N -(4-chlorocin-namyl)-N -methylbenzylamine)],a CaMKII inhibitor,was able to dose-dependently reverse the already-established antinoci-ceptive tolerance to morphine (p Ͻ0.001for 15–30nmol;not significant for 5nmol).KN92[2-[N -(4-methoxybenzenesul-fonyl)]amino-N -(4-chlorocinnamyl)-N -methylbenzylamine](30

nmol i.c.v.),a kinase-inactive analog of KN93,did not affect opioid tolerance.Neither KN92nor KN93affected basal noci-ception or acute morphine antinociception (1–10nmol i.c.v.).Likewise,dependence on morphine was abolished by the acute administration of KN93,but not KN92,in a dose-dependent manner.Pretreatment of mice with KN93also prevented the development of morphine tolerance and dependence.The ef-fect of acute CaMKII inhibition was not limited to the particular experimental model,because KN93also acutely reversed the established opioid tolerance and dependence in mice treated with morphine (75mg/pellet/mouse s.c.)for 6days.Taken together,these data strongly support the hypothesis that CaMKII can act as a key and direct factor in promoting opioid tolerance and dependence.Identifying such a direct mecha-nism may be useful for designing pharmacological treatments for these conditions.

Opioids,such as morphine,are clinically used primarily as analgesics.Drug tolerance and dependence are two of the major problems associated with these drugs,which greatly limit their effectiveness and usage.The molecular mecha-nisms underlying opioid tolerance and dependence are not entirely understood.The current study aims to test the hy-pothesis that Ca 2ϩ/calmodulin-dependent protein kinase II (CaMKII)can directly regulate opioid tolerance and depen-dence.

CaMKII is a multifunctional Ca 2ϩ/calmodulin-activated protein kinase,whose ␣and ␤isoforms are abundant in the central nervous system (Hudmon and Schulman,2002).A vast amount of information is available for the interaction of CaMKII ␣isoform and N -methyl-D -aspartate (NMDA)receptors in generating long-term potentiation in hippocampal neurons,which is critical for learning and memory (e.g.,Mayford et al.,1996).Inhibition or disrup-tion of CaMKII impairs spatial learning (Silva et al.,1992).Interestingly,antagonists of the NMDA receptor effectively inhibit the development of opioid tolerance and dependence (Marek et al.,1991;Trujillo and Akil,1991).It also has been reported that chronic microinjection of CaMKII inhibitors into hippocampus was able to prevent the development of opioid tolerance (Fan et al.,1999).The same CaMKII inhibitors were ineffective when chronically

This work was supported in part by National Institutes of Health Grant DA005050and funds from the University of Illinois and American Foundation for Pharmaceutical Education (AFPE).

Article,publication date,and citation information can be found at http://jpet.aspetjournals.org.doi:10.1124/jpet.105.097733.□S The online version of this article (available at http://jpet.aspetjournals.org)contains supplemental material.ABBREVIATIONS:CaMKII,Ca 2ϩ/calmodulin-dependent protein kinase II;CREB,cAMP-response element-binding protein;KN93,2-[N -(2-hydroxy-ethyl)]-N -(4-methoxybenzenesulfonyl)]amino-N -(4-chlorocinnamyl)-N -methylbenzylamine);KN92,2-[N -(4-methoxybenzenesulfonyl)]amino-N -(4-chloro-cinnamyl)-N -methylbenzylamine;MK801,(5R ,10S )-(ϩ)-5-methyl-10,11-dihydro-5H -dibenzo[a,d]cyclohepten-5,10-imine hydrogen maleate;NMDA,N -methyl-D -aspartate;␮OR,␮opioid receptor;MPE,maximum possible effect;pCaMKII,phosphorylated CaMKII;MS,morphine.

0022-3565/06/3172-901–909$20.00

T HE J OURNAL OF P HARMACOLOGY AND E XPERIMENTAL T HERAPEUTICS

Vol.317,No.2Copyright ©2006by The American Society for Pharmacology and Experimental Therapeutics 97733/3113104JPET 317:901–909,2006

Printed in U.S.A.

901

However,these studies did not address the possibility that CaMKII could directly(i.e.,not necessarily relying on learning and memory)modulate opioid tolerance and dependence.Identifying such a direct mechanism for CaMKII is not only important to our understanding but will also have a profound impact on designing therapeutic interventions for opioid tolerance and dependence.Studies have suggested that the NMDA system may be directly involved in promoting opioid tolerance and dependence (Gutstein and Trujillo,1993).Similar direct mechanisms have not been proposed for CaMKII.

In cellular models of opioid tolerance,␮opioid receptor (␮OR)desensitization was enhanced when a constitutively active form of CaMKII was also expressed in Xenopus oocytes (Mestek et al.,1995;Koch et al.,1997).The effect was absent if the native receptor was replaced with a mutated receptor lacking the consensus CaMKII phosphorylation sites(Koch et al.,1997).On the other hand,intracellular Ca2ϩ,calmod-ulin,and CaMKII can all be regulated by opioids.Cytosolic free Ca2ϩwas increased after the treatment with opioids (Fields and Sarne,1997;Smart et al.,1997;Spencer et al., 1997;Quillan et al.,2002).Likewise,chronic treatments with opioids have been found to increase calmodulin activity (Nehmad et al.,1982)and mRNA levels(Niu et al.,2000). Indeed,CaMKII activity was increased in opioid tolerance in vivo(Lou et al.,1999;Wang et al.,2003;Liang et al.,2004). Anatomically,CaMKII and␮OR were found to be colocalized in dorsal root ganglia sensory neurons and superficial layers of spinal cord dorsal horn(Bruggemann et al.,2000).More-over,upon activation,␮OR was found to internalize to intra-cellular locations where CaMKII proteins were located (Bruggemann et al.,2000).Therefore,cellular and biochem-ical evidence support the possibility that CaMKII and opioid system can directly interact with each other,leading to cel-lular opioid desensitization.In a preliminary study,we found that tolerance to morphine was abolished by an acute spinal treatment with a single dose of KN93[2-[N-(2-hydroxy-ethyl)]-N-(4-methoxybenzenesulfonyl)]amino-N-(4-chlorocin-namyl)-N-methylbenzylamine)],a CaMKII inhibitor(Niki et al.,1993),consistent with the hypothesis that CaMKII can directly modulate opioid tolerance(Wang et al.,2003).How-ever,we caution that multiple factors can confound this interpretation.The current study extends our previous find-ings by carefully applying a different degree of CaMKII in-hibition using multiple doses of KN93and a negative control, KN92[2-[N-(4-methoxybenzenesulfonyl)]amino-N-(4-chloro-cinnamyl)-N-methylbenzylamine](Tombes et al.,1995),in two rodent models of opioid tolerance.In addition,in this study,we tested whether opioid dependence can be directly regulated by CaMKII.

Materials and Methods

Materials.Morphine and placebo pellets were obtained from the National Institute on Drug Abuse(Rockville,MD).Morphine sulfate was from Abbott Laboratories(North Chicago,IL).

KN93and KN92were purchased from Calbiochem(San Diego,CA).Naloxone and all other chemicals were from Sigma(St.Louis, MO).ICR mice(20–25g;Harlan Laboratories,Indianapolis,IN) were maintained on a12-h light/12-h dark cycle and provided with food and water ad libitum before experimental procedures.All experiments were performed in accordance with the National Institutes of Health guidelines and after approval by the Animal Care and Use Committee of the University of Illinois at Chicago.

Drug Administration.Intracerebroventricular injections,under light ether anesthesia,were made into the left lateral ventricles as described previously(Bilsky et al.,1996;Wang et al.,2001).Stan-dard procedures were used for i.p.and s.c.injections.Placebo and morphine pellets were implanted in the s.c.space as described pre-viously(Way et al.,1969;Patrick et al.,1975).

Tests for Antinociception.Basal nociception and morphine-induced antinociception were evaluated using the tail-immersion test as described previously(Wang et al.,2001).In brief,the test was performed by dipping the distal1/3of the tail into a water bath maintained at52°C and recording the latency to a rapid tail-flick response.Morphine-induced(i.c.v.)antinociception was tested at the time of peak drug response after the injection of morphine and expressed as the percentage of maximal possible effect(MPE)ac-cording to the following formula:%MPEϭ100ϫ(postdrug latencyϪpredrug latency)/(cut-off-predrug latency).A12-s cut-off was applied to prevent tissue injury.The time of peak drug response was deter-mined to be20min in our studies and was not altered by opioid tolerance(see Supplemental Data).

Acute Opioid Tolerance and Dependence.Separate groups of eight ICR mice(20–25g)were made acutely tolerant to and dependent on opioids by the administration of a large dose of morphine(100mg/kg s.c.)(Wang et al.,1994;Bilsky et al.,1996). We have reported that maximal morphine tolerance and depen-dence developed over2to6h(Bilsky et al.,1996).Control mice received the same volume of saline.Tolerance to opioids was studied in these mice4.5h later by measuring the antinociceptive effect exhibited by a test dose of morphine(1–10nmol i.c.v.).A significant reduction of antinociceptive effect signified the pres-ence of tolerance to morphine.To assess dependence,mice were given naloxone(1–10mg/kg i.p.)5h after the administration of morphine(100mg/kg s.c.)and were immediately placed inside glass cylinders.Vertical jumps were recorded for15min.To determine the effect of CaMKII inhibition,separate groups of mice were given the CaMKII inhibitor KN93(5–30nmol i.c.v.)(Niki et al.,1993)or its kinase-inactive structural analog,KN92(30nmol i.c.v.)(Tombes et al.,1995),15min before naloxone or the test dose of morphine.To test whether KN93or KN92prevented the development of tolerance and dependence,KN93or KN92was coadministered with morphine(100mg/kg s.c.)

Effect of KN93and KN92on Basal Nociception and Mor-phine Antinociception.To determine whether CaMKII inhibitor KN93or its inactive structural analog,KN92,affected tail-flick latency and interfered with morphine antinociception,groups of eight mice were given KN93(30nmol/5␮l i.c.v.),KN92(30nmol/5␮l i.c.v.),or saline(5␮l i.c.v.)15min before the administration of morphine(1–10nmol i.c.v.)or saline(5␮l i.c.v.).Latencies to tail-flick responses were determined20min later.

Chronic Model of Opioid Tolerance and Dependence.On day0,separate groups of six male ICR mice(20–25g)were implanted subcutaneously with morphine pellets(1pellet/mouse; each pellet contains75mg of morphine base)to induce opioid tolerance and dependence.Control mice received placebo pellets(1 pellet/mouse;a placebo pellet is made of the same excipients but contains no morphine).Tolerance and dependence developed over the course of2to6days(Ho et al.,1975;Patrick et al.,1975).To determine morphine tolerance,mice were injected with a test dose of morphine(10nmol i.c.v.).A significant reduction of morphine-antinociceptive effect signified the presence of tolerance to mor-phine.The presence of opioid dependence was revealed by chal-lenging morphine-treated mice with naloxone(10mg/kg i.p.)on

902Tang et al.

Day 6.Immediately after the administration of naloxone,mice were placed inside glass cylinders,and the number of withdrawal jumps was recorded for 15min.To determine the effect of acute CaMKII inhibition on tolerance or dependence,separate groups of mice were given KN93(5–45nmol i.c.v.)15min before naloxone or the test dose of morphine on day 6.

Western Blot Analysis.Western blotting procedures were performed as described previously (Wang et al.,2001).In brief,brain cortices were dissected and quickly frozen on dry ice.Tis-sues from three mice of the same experimental group were pooled and homogenized with a glass homogenizer in a buffer [1%Non-idet P-40,0.5%sodium deoxycholate,0.1%sodium dodecyl sulfate,and 5mM EDTA in phosphate-buffered saline,pH 7.4](3ml/g wet tissue)in the presence of protease inhibitors (0.05mg/ml bestatin,0.05mg/ml leupeptin,0.05mg/ml pepstatin,and 0.1mg/ml phe-nylmethylsulfonyl fluoride).The homogenates were incubated on a rotator at 4°C for 2h,and the soluble fraction was collected after separation by centrifugation (45,000g ,60min).Protein content was determined by a modified Bradford method (Pierce Biotech-nology,Rockford,IL).Samples (15␮g of protein)were separated by 10%SDS-polyacrylamide gel electrophoresis and electrotrans-ferred onto nitrocellulose membrane.The membrane was pre-blocked in 5%nonfat milk in 20mM Tris-buffer saline (pH 7.6)with 0.1%Tween 20and probed with rabbit antibodies recogniz-ing CaMKII (1/1,000;Santa Cruz Biotechnology,Santa Cruz,CA),the activated form of CaMKII [antiphosphorylated CaMKII (pCaMKII)antibody,1/1000;Promega,Madison,WI],or the acti-vated form of cAMP-response element-binding protein

(CREB;

Fig.1.Effect of KN93on morphine-antinociceptive tolerance and CaMKII activity.A,the dose-response curve of KN93in reversing the established opioid tolerance.Groups of eight male ICR mice received morphine or an equal volume of saline (Saline).Four hours later,KN93(5,15,or 30nmol,dissolved in 5␮l of saline i.c.v.)was given to several groups of morphine-treated mice.The remaining groups of mice received an i.c.v.injection of saline (5␮l).Fifteen minutes later,morphine (1,3,or 10nmol i.c.v.)was administered to mice to determine its antinociception by the tail-flick assay.Data are expressed in %MPE (mean ϮS.E.M.).B,comparison of the effects of KN93and KN92on opioid antinociceptive tolerance.In morphine-treated mice (100mg/kg s.c.),tolerance to morphine (10nmol i.c.v.)was reversed by the acute treatments with KN93(15and 30nmol i.c.v.)but not by the lowest dose of KN93(5nmol i.c.v.)or KN92(30nmol i.c.v.).C,the time course of the activation of CaMKII by morphine.Groups of three mice were injected s.c.with saline or morphine (100mg/kg).Brain samples were taken at the indicated time points to determine supraspinal CaMKII activity.The activated CaMKII was determined by the Western blotting method using an antibody specific for the Thr286-pCaMKII.Histograph data,expressed as mean ϮS.E.M.,were constructed from the representative figure shown and three other experiments.D,inhibition of CaMKII activation by KN93.Groups of three morphine-treated mice (100mg/kg s.c.for 4h)were injected i.c.v.with either saline (5␮l)or KN93(30nmol in 5␮l of saline).Fifteen minutes later,brain samples were taken for the analyses of supraspinal CaMKII activity as described above.Histogram data,expressed as mean ϮS.E.M.,were constructed from the representative figure shown and three other experiments.ءء,p Ͻ0.01;ءء,p Ͻ0.01;ءءء,p Ͻ0.001compared with the Saline group;###,p Ͻ0.001compared with the MS group.

CaMKII in Opioid Tolerance and Dependence

903

Statistical Analysis.Data are expressed as meanϮS.E.M.Dif-ferences in responses between the treatment groups were deter-mined using analysis of variance followed by Student’s t(two groups) or Dunnett’s t(multiple groups)tests.Statistical significance was established at95%.

Results

Effect of the Acute CaMKII Inhibition on Acute Mor-phine Tolerance.We first tested our hypothesis in an acute model of opioid tolerance.Tolerance to morphine was estab-lished in2to6h after the administration of morphine(100 mg/kg s.c.)(Bilsky et al.,1996),as evidenced by the signifi-cant reduction of i.c.v.morphine-induced antinociception (Fig.1,A and B).Brain CaMKII activity increased over a time course of1to4h after s.c.morphine administration (Fig.1C)correlated with the time course of the development of tolerance(Bilsky et al.,1996).KN93(15–30nmol i.c.v.),a selective CaMKII inhibitor,administered35min before the antinociceptive test(15min before the i.c.v.test dose of morphine),reversed the already-established tolerance to morphine.The effect was KN93-dose-dependent,as KN93at higher doses(15and30nmol)was able to significantly re-verse the established tolerance to morphine(pϽ0.001)while ineffective at a lower dose(5nmol)(Fig.1,A and B).In contrast,KN92,a kinase-inactive structural analog of KN93, had no effect on morphine-antinociceptive tolerance(Fig. 1B).The Western blotting experiments verified that acutely administered KN93(i.c.v.)significantly reduced the supra-spinal CaMKII activity in morphine-treated mice(Fig.1D). Effect of KN93and KN92on Basal Nociception and Morphine Antinociception.One potential problem in in-terpreting the above data was that KN93might directly affect basal nociception or interfere with the antinociceptive effect of morphine.To account for this possibility,we tested the effects of KN93and KN92on basal nociception and morphine antinociception.KN93or KN92did not alter the basal tail-flick withdrawal latencies(data not shown).In all three of the morphine doses examined,morphine-induced antinociception was not affected by the administration of KN93or KN92(Fig.2),indicating that KN93or KN92did not affect acute morphine antinociception.

Effect of the Acute CaMKII Inhibition on Acute Mor-phine Dependence.In morphine-treated(100mg/kg s.c.) mice,dependence on opioids developed in2to6h(Bilsky et al.,1996).Challenging these mice with naloxone i.p.(5h after morphine administration)precipitated withdrawal jumps in a naloxone dose-dependent manner,which was largely absent in saline-treated control mice(Fig.3,A and B). KN93(30nmol i.c.v.),given15min before naloxone,was able to completely suppress the naloxone-induced withdrawal jumping(Fig.3,A and B).Withdrawal jumping was signifi-cantly attenuated(pϽ0.01)at a lower dose(15nmol), whereas KN93at the lowest dose used(5nmol)did not significantly affect naloxone-induced withdrawal jumping. Nor did the negative control compound,KN92(30nmol), show a significant effect(Fig.3B).

Prevention of Opioid Tolerance and Dependence by the CaMKII Inhibition.We next tested whether a broad supraspinal inhibition of CaMKII could prevent the develop-ment of tolerance to and dependence on opioids.In these studies,KN93or KN92(30nmol i.c.v.)was administered immediately before the injection of morphine(100mg/kg s.c.).Five hours later,mice received morphine or“morphine plus KN92”developed antinociceptive tolerance to morphine, whereas tolerance was absent in mice receiving“morphine plus KN93”[pϽ0.01compared with morphine(MS)group; not significantly different from the control group](Fig.4A). Similar cotreatment with KN93,but not KN92(not signifi-cantly different from the MS group),prevented the develop-ment of opioid dependence by significantly reducing the num-bers of naloxone-induced withdrawal jumps(pϽ0.01 compared with MS group;not significantly different from the control group)(Fig.4B).These data indicated that a broad supraspinal inhibition of CaMKII was effective in preventing the development of both opioid tolerance and dependence in mice.

KN93Reversed Opioid Tolerance in a Chronic Model of Opioid Tolerance.To determine that the pharmacolog-ical effect of CaMKII inhibition was not limited to a partic-ular animal model of opioid tolerance,we further tested our hypothesis in a chronic model of opioid tolerance.Mice de-veloped tolerance to opioids over2to6days after receiving the s.c.implantation of morphine pellets(Ho et al.,1975; Patrick et al.,1975).The antinociception produced by mor-phine(10nmol i.c.v.)before the implantation of pellets was 91.0Ϯ5.0%MPE.On day6,the same test dose of morphine produced a significantly reduced antinociceptive response in morphine-pelleted mice(11.5Ϯ2.0%MPE,pϽ0.001)while remaining fully active in mice implanted with placebo pellets (93.1Ϯ 4.0%MPE)(Fig.5A),indicating the presence of antinociceptive tolerance in morphine-pelleted mice.

Su-Fig.2.Effects of KN93and KN92on MS antinociception.Groups of eight male ICR mice received i.c.v.injections of KN93(30nmol in5␮l of saline),KN92(30nmol in5␮l of saline),or saline(5␮l).Fifteen minutes later,morphine(1,3,or10nmol i.c.v.)was administered to mice to determine its antinociception by the tail-flick assay.Neither KN93nor KN92altered acute morphine antinociception(pϾ0.05).Data are ex-pressed as%MPE(meanϮS.E.M.).

904Tang et al.

praspinal CaMKII expression and activity increased signifi-cantly over the time course of 6days as the tolerance devel-oped (Fig.5,B–D).When KN93(15–45nmol i.c.v.)was given 15min before morphine-antinociceptive tests,KN93was able to significantly reverse the established tolerance in these mice (p Ͻ0.001for all three doses)(Fig.5A).KN93at the lowest dose used (5nmol)was not effective.Compared with the acute model,a higher dose of KN93(45nmol)was needed to achieve a complete reversal of tolerance (not significantly different from the “placebo-pelleted”group).These results were in agreement with those from the acute model,suggest-ing that the acute inhibition of supraspinal CaMKII effec-tively disrupted the established opioid-antinociceptive toler-ance.

KN93Reversed Opioid Dependence in a Chronic Model of Opioid Dependence.We next tested the effect of an acute CaMKII inhibition on opioid dependence in a chronic opioid dependence model.On day 6after morphine-pelleting,naloxone precipitated a significant number of with-drawal jumps in morphine-implanted mice compared with placebo-pelleted mice,indicative of the presence of opioid dependence (Fig.6).Acutely administered KN93(i.c.v.15min before naloxone administration)dose-dependently de-creased the number of naloxone-withdrawal jumping in mor-phine-pelleted mice.At the highest dose (45nmol),KN93completely suppressed the withdrawal jumping (p Ͻ0.001compared with the morphine-pelleted mice;not significantly different from the placebo-pelleted mice).KN93at lower doses (5–30nmol)also significantly reduced the number of withdrawal jumping (Fig.6),suggesting that acute inhibition of CaMKII was effective in reversing established opioid de-pendence in the chronic mouse model of opioid dependence.Downstream Effectors of CaMKII.CaMKII affects a number of downstream effectors,including receptors and transcription factors.The transcriptional factor CREB,

a

Fig.3.Effect of acute CaMKII inhibition by KN93on morphine depen-dence.A,dose-response curve of KN93in reversing the established opioid dependence.Groups of eight male ICR mice received s.c.injections of MS or an equal volume of saline (Saline).All mice received i.p.injections of naloxone (1,3,or 10mg/kg)5h later.Fifteen minutes before the injection of naloxone,KN93(5,15,or 30nmol dissolved in 5␮l of saline i.c.v.)was given to several groups of morphine-treated mice.The remaining groups of mice received i.c.v.injections of saline (5␮l).The numbers of vertical jumps were recorded for each group of mice and expressed as mean ϮS.E.M.B,comparison of the effects of KN93and KN92on opioid depen-dence.In morphine-treated mice (100mg/kg s.c.),dependence on mor-phine was revealed by 10mg/kg i.p.naloxone-precipitated withdrawal jumping 5h later.The withdrawal jumping was completely blocked by an acute treatment with KN93at the highest dose (30nmol,i.c.v.)and attenuated by KN93at 15nmol (p Ͻ0.01).KN93at the lowest dose (5nmol,i.c.v.)or KN92(30nmol i.c.v.)did not significantly alter morphine dependence (p Ͼ0.05).ءءء,p Ͻ0.001compared with the Saline group;##,p Ͻ0.01;###,p Ͻ0.001compared with the MS

group.

Fig.4.Prevention of MS-antinociceptive tolerance and physical depen-dence by KN93.Groups of eight male ICR mice received KN93(30nmol in 5␮l of saline i.c.v.,designated as KN93ϩMS group),KN92(30nmol in 5␮l of saline i.c.v.,KN92ϩMS group),or saline (5␮l i.c.v.,MS group)immediately prior to the administration of morphine (100mg/kg s.c.).Control mice received only s.c.injection of saline.A,development of tolerance to morphine was prevented by KN93but not by KN92.B,development of morphine dependence,as revealed by 10mg/kg i.p.nal-oxone-precipitated withdrawal jumping,was prevented by KN93but not by KN92.Data are expressed as mean ϮS.E.M.ءءء,p Ͻ0.001compared with the Control group;##,p Ͻ0.01compared with the MS group.

CaMKII in Opioid Tolerance and Dependence

905

key CaMKII downstream effector,has been previously pro-posed to be important in opioid tolerance and dependence (Nestler,2001).We examined the activation of CREB (pCREB)in mice that have been treated with morphine.As expected,chronic treatment with morphine increased the levels of pCREB (Fig.7).

Discussion

The current study tested the hypothesis that CaMKII can directly regulate opioid tolerance and dependence.Previous studies using a rat model of opioid tolerance and dependence (10mg/kg s.c.morphine every 12h)demon-strated that opioid tolerance and dependence were pre-vented by chronically microinjecting into hippocampus

chemical CaMKII inhibitors or antisense oligodeoxynucle-otides (Fan et al.,1999;Lu et al.,2000).The same chronic treatments in striatum (tolerance)and amygdale (depen-dence)were not effective.Neither was the acute CaMKII inhibition in hippocampus able to affect opioid tolerance (Fan et al.,1999).Based on the hippocampus-specific ac-tion of CaMKII inhibition and the requirement of chronic administration,it has been suggested that CaMKII mod-ulated opioid tolerance and dependence through the learn-ing and memory pathways.This was not entirely surpris-ing since opioid tolerance and dependence have been hypothesized to involve learning and memory (Siegel,1976).Several antagonists of NMDA receptor,another member of a group of genes essential for long-term

poten-

Fig.5.A,reversal of morphine-antinociceptive tolerance by KN93in a chronic model of opioid tolerance.Groups of six male ICR mice were implanted s.c.with morphine pellets (1ϫ75mg of morphine/pellet/mouse)or placebo pellets (PB group)on day 0.On day 6,KN93(5–45nmol,dissolved in 5␮l of saline i.c.v.)was given to several groups of morphine-pelleted mice.The remaining groups of mice received an i.c.v.injection of saline (5␮l).Fifteen minutes later,morphine antinociception (10nmol i.c.v.)was determined.Data are expressed as %MPE (mean ϮS.E.M.).Antinociceptive tolerance to morphine was detected in mice that received only morphine pellets (MS).In contrast,in morphine-pelleted mice that also received KN93,tolerance was completely reversed (45nmol,p Ͼ0.01compared with PB group)or significantly attenuated (5–30nmol).ءءء,p Ͻ0.001compared with PB group;###,p Ͻ0.01compared with the MS group.B,activation of supraspinal CaMKII by morphine.Groups of three mice were implanted s.c.with morphine pellets (1morphine pellet/mouse)or placebo pellets (day 0group).Brain samples were taken at the indicated time points for the analyses of supraspinal CaMKII activity.The activated CaMKII was determined by the Western blotting method using an antibody specific for Thr286-pCaMKII.Data,expressed as mean ϮS.E.M.,were from the representative figure shown and three other experiments.ء,p Ͻ0.05compared with day 0.C,chronic treatment with morphine increased supraspinal CaMKII expression.Groups of three mice were implanted s.c.with morphine pellets (1morphine pellet/mouse)or placebo pellets (day 0group).Brain samples were taken at the indicated time points for the analyses of supraspinal CaMKII expression.Data,expressed as mean ϮS.E.M.,were from the representative figure shown and three other experiments.ء,p Ͻ0.05compared with day 0.D,activation of supraspinal CaMKII by morphine normalized by the expression of total CaMKII.Experiments were performed as described in B and C.Data from these panels were analyzed together to obtain the ratio of pCaMKII/total CaMKII.ء,p Ͻ0.05compared with day 0.

906

Tang et al.

tiation and learning and memory,have also been previ-ously found to prevent the development of opioid tolerance and dependence (Marek et al.,1991;Trujillo and Akil,1991).Further studies found that the effect of MK801,an NMDA receptor antagonist,did not depend entirely on its ability to interfere with associative learning (Gutstein and Trujillo,1993).Instead,MK801prevented morphine-an-tinociceptive tolerance in spinalized rats (Gutstein and Trujillo,1993),indicating that the NMDA receptor directly (i.e.,independent of learning and memory)affected opioid tolerance and dependence.

In this study,we tested whether such a direct role ex-isted for CaMKII.Because the acute inhibition of hip-pocampal CaMKII did not affect opioid tolerance (Fan et al.,1999),it would suggest that the acute inhibition of CaMKII had no or minimal impact on learning and mem-ory.We took the study paradigm a step further to study the acute supraspinal inhibition of CaMKII in two models of opioid tolerance and dependence in mice.An effect by the acute supraspinal inhibition served as an indication for a direct effect by CaMKII.

Unlike the acute hippocampal CaMKII inhibition,the acute supraspinal CaMKII inhibition by KN93i.c.v.,a selec-tive CaMKII inhibitor (Niki et al.,1993),effectively reversed the established opioid tolerance and dependence in an acute model of opioid tolerance/dependence.Whereas morphine-treated groups of mice showed antinociceptive tolerance,acutely administered (15min before)KN93was able to dose-dependently reverse the antinociceptive tolerance.The effect was not observed in mice acutely treated with KN92,a ki-nase-inactive structural analog of KN93(Tombes et al.,1995).Moreover,the effect of KN93was not due to any direct effect on nociception or antinociception,because KN93did not produce antinociception by itself or interfere with acute

morphine antinociception.A previous study also did not find an interference by KN93i.c.v.on subcutaneous morphine-produced antinociception or hyperlocomotion (Narita et al.,2004).The Western blotting analysis,on the other hand,confirmed the inhibition of CaMKII activity by the acute i.c.v.administration of KN93.

A direct action of CaMKII in opioid tolerance and depen-dence was further supported by studies in a chronic model (1ϫ75mg of morphine pellet/mouse for 6days)of opioid tolerance and dependence in mice.The acute CaMKII inhi-bition by KN93reversed the already-established morphine tolerance and dependence,indicating that the effect of acute CaMKII inhibition was not limited to a particular model.We have previously found that acute spinal inhibition of CaMKII reversed the established morphine tolerance in rats (Wang et al.,2003),suggesting that both spinal and supraspinal CaMKII are essential for the maintenance of opioid tolerance and dependence.

The time course of CaMKII activation correlated with the development of opioid tolerance and dependence in both of the models that we tested (Ho et al.,1975;Patrick et al.,1975;Bilsky et al.,1996).These data were in agreement with previous findings that CaMKII can be activated in different CNS regions after the treatment with morphine (Fan et al.,1999;Wang et al.,2003;Liang et al.,2004).

The direct action of CaMKII in opioid tolerance and dependence was supported by data from cellular studies.In cellular models,desensitization of ␮opioid receptor was found to be affected by CaMKII.DAMGO ([D -Ala 2,N -Me-Phe 4,Gly 5-ol]-enkephalin)-induced activation of G-protein-gated inwardly rectifying potassium channels was

signifi-

Fig.6.Reversal of morphine dependence by KN93in a chronic model of opioid dependence.Groups of six male ICR mice were implanted s.c.with morphine pellets (1morphine pellet/mouse)or placebo pellets (PB group)on day 0.On day 6,KN93(5–45nmol,dissolved in 5␮l saline i.c.v.)was given to several groups of morphine-pelleted mice.The remaining groups of mice received an i.c.v.injection of saline (5␮l).Fifteen minutes later,opioid dependence was revealed by 10mg/kg i.p.naloxone-induced with-drawal jumping in mice treated only with morphine pellets (MS group).In morphine-pelleted mice that also received KN93,naloxone-induced withdrawal jumping was completely blocked (45nmol,p Ͼ0.05compared with PB group)or significantly attenuated (5–30nmol).ءء,p Ͻ0.01;ءءء,p Ͻ0.001compared with the PB group;#,p Ͻ0.05;###,p Ͻ0.001compared with the MS

group).

Fig.7.Activation of supraspinal CREB by morphine.Groups of three mice were implanted s.c.with morphine pellets (1morphine pellet/mouse)or placebo pellets (day 0group).Brain samples were taken at the indicated time points for the analyses of supraspinal CREB activity.The activated CREB was determined by the Western blotting method using an antibody specific for phosphorylated CREB (pCREB).Data,expressed in mean ϮS.E.M.,were from the representative figure shown and three other experiments.ء,p Ͻ0.05,compared with day 0.

CaMKII in Opioid Tolerance and Dependence

907

Besides opioid receptors,CaMKII may also interact with the NMDA receptors,leading to opioid tolerance and depen-dence.Ca2ϩinflux via the activation of NMDA receptors results in the activation and autophosphorylation of CaMKII at position Thr286(Fukunaga et al.,1992;Strack et al., 2000).Activated CaMKII,in turn,can phosphorylate and activate NMDA receptors,leading to more Ca2ϩinflux through the channel(Kitamura et al.,1993).Therefore, CaMKII and the NMDA receptor can interact with each other in a feed-forward fashion.As noted above,the NMDA recep-tor has been shown to directly affect opioid tolerance and dependence(Gutstein and Trujillo,1993).

In addition to opioid receptors and NMDA receptors,numer-ous other downstream effectors can be affected by CaMKII, including transcription factors,such as CREB(Sheng et al., 1991),activating transcript factor1(Shimomura et al.,1996), serum response factor(Misra et al.,1994),and CAAT-enhancer-binding protein␤(Wegner et al.,1992).We found that activa-tion of CREB(pCREB)was closely related to the activation of CaMKII in mice that were tolerant to and dependent on mor-phine.These data are in agreement with previous findings that CREB is important for opioid tolerance and dependence(Nes-tler,2001;Valverde et al.,2004).

In summary,our data provided the initial evidence for a critical role of CaMKII in directly promoting opioid toler-ance and dependence.This mechanism is in addition to its impact on tolerance and dependence via learning and memory.Identifying such a direct mechanism will not only have a profound impact on our understanding but may also lead to pharmacological interventions targeting the CaMKII pathway for the attenuation of opioid tolerance and dependence.Recently,we found that trifluoperazine, an orally available antipsychotic drug,was capable of dis-rupting opioid tolerance by inhibiting CaMKII(Tang et al., 2006).It is conceivable that CaMKII inhibitors,such as phenothiazine antipsychotics,should be tested in humans for the prevention and/or treatment of opioid addiction and tolerance.

References

Bilsky EJ,Bernstein RN,Wang Z,Sadee W,and Porreca F(1996)Effects of naloxone and D-Phe-Cys-Tyr-D-Trp-Arg-Thr-Pen-Thr-NH2and the protein kinase inhibitors H7and H8on acute morphine dependence and antinociceptive tolerance in mice. J Pharmacol Exp Ther277:484–490.

Bruggemann I,Schulz S,Wiborny D,and Hollt V(2000)Colocalization of the mu-opioid receptor and calcium/calmodulin-dependent kinase II in distinct pain-processing brain regions.Brain Res Mol Brain Res85:239–250.

Fan GH,Wang LZ,Qiu HC,Ma L,and Pei G(1999)Inhibition of calcium/calmodulin-dependent protein kinase II in rat hippocampus attenuates morphine tolerance and dependence.Mol Pharmacol56:39–45.

Fields A and Sarne Y(1997)The stimulatory effect of opioids on cyclic AMP produc-tion in SK-N-SH cells is mediated by calcium ions.Life Sci61:595–602. Fukunaga K,Soderling TR,and Miyamoto E(1992)Activation of Ca2ϩ/calmodulin-dependent protein kinase II and protein kinase C by glutamate in cultured rat hippocampal neurons.J Biol Chem267:22527–22533.

Gutstein HB and Trujillo KA(1993)MK-801inhibits the development of morphine tolerance at spinal sites.Brain Res626:332–334.Ho IK,Brase DA,Loh HH,and Way EL(1975)Influence of L-tryptophan on mor-phine analgesia,tolerance,and physical dependence.J Pharmacol Exp Ther193: 35–43.

Hudmon A and Schulman H(2002)Structure-function of the multifunctional Ca2ϩ/ calmodulin-dependent protein kinase II.Biochem J3:593–611.

Kitamura Y,Miyazaki A,Yamanaka Y,and Nomura Y(1993)Stimulatory effects of protein kinase C and calmodulin kinase II on N-methyl-D-aspartate receptor/ channels in the postsynaptic density of rat brain.J Neurochem61:100–109. Koch T,Kroslak T,Mayer P,Raulf E,and Hollt V(1997)Site mutation in the rat mu-opioid receptor demonstrates the involvement of calcium/calmodulin-dependent protein kinase II in agonist-mediated desensitization.J Neurochem 69:1767–1770.

Liang D,Li X,and Clark JD(2004)Increased expression of Ca2ϩ/calmodulin-dependent protein kinase II alpha during chronic morphine exposure.Neuro-science123:769–775.

Lou L,Zhou T,Wang P,and Pei G(1999)Modulation of Ca2ϩ/calmodulin-dependent protein kinase II activity by acute and chronic morphine administration in rat hippocampus:differential regulation of alpha and beta isoforms.Mol Pharmacol 55:557–563.

Lu L,Zeng S,Liu D,and Ceng X(2000)Inhibition of the amygdala and hippocampal calcium/calmodulin-dependent protein kinase II attenuates the dependence and relapse to morphine differently in rats.Neurosci Lett291:191–195.

Marek P,Ben-Eliyahu S,Gold M,and Liebeskind JC(1991)Excitatory amino acid antagonists(kynurenic acid and MK-801)attenuate the development of morphine tolerance in the rat.Brain Res547:77–81.

Mayford M,Bach ME,Huang YY,Wang L,Hawkins RD,and Kandel ER(1996) Control of memory formation through regulated expression of a CaMKII trans-gene.Science(Wash DC)274:1678–1683.

Mestek A,Hurley JH,Bye LS,Campbell AD,Chen Y,Tian M,Liu J,Schulman H, and Yu L(1995)The human mu opioid receptor:modulation of functional desen-sitization by calcium/calmodulin-dependent protein kinase and protein kinase C. J Neurosci15:2396–2406.

Misra RP,Bonni A,Miranti CK,Rivera VM,Sheng M,and Greenberg ME(1994) L-type voltage-sensitive calcium channel activation stimulates gene expression by a serum response factor-dependent pathway.J Biol Chem269:25483–25493. Narita M,Matsumura Y,Ozaki S,Ise Y,Yajima Y,and Suzuki T(2004)Role of the calcium/calmodulin-dependent protein kinase II(CaMKII)in the morphine-induced pharmacological effects in the mouse.Neuroscience126:415–421. Nehmad R,Nadler H,and Simantov R(1982)Effects of acute and chronic morphine treatment of calmodulin activity of rat brain.Mol Pharmacol22:3–394. Nestler EJ(2001)Molecular neurobiology of addiction.Am J Addict10:201–217. Niki I,Okazaki K,Saitoh M,Niki A,Niki H,Tamagawa T,Iguchi A,and Hidaka H (1993)Presence and possible involvement of Ca/calmodulin-dependent protein kinases in insulin release from the rat pancreatic beta cell.Biochem Biophys Res Commun191:255–261.

Niu S,Kuo CH,Gan Y,Nishikawa E,Sadakata T,Ichikawa H,and Miki N(2000) Increase of calmodulin III gene expression by mu-opioid receptor stimulation in PC12cells.Jpn J Pharmacol84:412–417.

Patrick GA,Dewey WL,Spaulding TC,and Harris LS(1975)Relationship of brain morphine levels to analgesic activity in acutely treated mice and rats and in pellet implanted mice.J Pharmacol Exp Ther193:876–883.

Quillan JM,Carlson KW,Song C,Wang D,and Sadee W(2002)Differential effects of mu-opioid receptor ligands on Ca2ϩsignaling.J Pharmacol Exp Ther302:1002–1012.

Sheng M,Thompson MA,and Greenberg ME(1991)CREB:a Ca(2ϩ)-regulated transcription factor phosphorylated by calmodulin-dependent kinases.Science (Wash DC)252:1427–1430.

Shimomura A,Ogawa Y,Kitani T,Fujisawa H,and Hagiwara M(1996)Calmodulin-dependent protein kinase II potentiates transcriptional activation through acti-vating transcription factor1but not cAMP response element-binding protein. J Biol Chem271:17957–17960.

Siegel S(1976)Morphine analgesic tolerance:its situation specificity supports a Pavlovian conditioning model.Science(Wash DC)193:323–325.

Silva AJ,Paylor R,Wehner JM,and Tonegawa S(1992)Impaired spatial learning in alpha-calcium-calmodulin kinase II mutant mice.Science(Wash DC)257:206–211.

Smart D,Hirst RA,Hirota K,Grandy DK,and Lambert DG(1997)The effects of recombinant rat mu-opioid receptor activation in CHO cells on phospholipase C, [Ca2ϩ]i and adenylyl cyclase.Br J Pharmacol120:1165–1171.

Spencer RJ,Jin W,Thayer SA,Chakrabarti S,Law PY,and Loh HH(1997)Mobi-lization of Ca2ϩfrom intracellular stores in transfected neuro2a cells by activation of multiple opioid receptor subtypes.Biochem Pharmacol54:809–818.

Strack S,McNeill RB,and Colbran RJ(2000)Mechanism and regulation of calcium/ calmodulin-dependent protein kinase II targeting to the NR2B subunit of the N-methyl-D-aspartate receptor.J Biol Chem275:23798–23806.

Tang L,Shukla PK,and Wang ZJ(2006)Trifluoperazine,an orally available clinically used drug,disrupts opioid antinociceptive tolerance.Neurosci Lett 397:1–40.

Tombes RM,Grant S,Westin EH,and Krystal G(1995)G1cell cycle arrest and apoptosis are induced in NIH3T3cells by KN-93,an inhibitor of CaMK-II(the multifunctional Ca2ϩ/CaM kinase).Cell Growth Differ6:1063–1070.

Trujillo KA and Akil H(1991)Inhibition of morphine tolerance and dependence by the NMDA receptor antagonist MK-801.Science(Wash DC)251:85–87. Valverde O,Mantamadiotis T,Torrecilla M,Ugedo L,Pineda J,Bleckmann S,Gass P,Kretz O,Mitchell JM,Schutz G,et al.(2004)Modulation of anxiety-like behavior and morphine dependence in CREB-deficient mice.Neuropsychopharma-cology29:1122–1133.

Wang Z,Bilsky EJ,Porreca F,and Sadee W(1994)Constitutive mu opioid receptor activation as a regulatory mechanism underlying narcotic tolerance and depen-dence.Life Sci54:L339–350.

908Tang et al.Wang Z,Gardell LR,Ossipov MH,Vanderah TW,Brennan MB,Hochgeschwender U, Hruby VJ,Malan TP Jr,Lai J,and Porreca F(2001)Pronociceptive actions of dynorphin maintain chronic neuropathic pain.J Neurosci21:1779–1786.

Wang ZJ,Tang L,and Xin L(2003)Reversal of morphine antinociceptive tolerance by acute spinal inhibition of Ca(2ϩ)/calmodulin-dependent protein kinase II.Eur J Pharmacol465:199–200.

Way EL,Loh HH,and Shen FH(1969)Simultaneous quantitative assessment of morphine tolerance and physical dependence.J Pharmacol Exp Ther167: 1–8.Wegner M,Cao Z,and Rosenfeld MG(1992)Calcium-regulated phosphory-lation within the leucine zipper of C/EBP beta.Science(Wash DC)256: 370–373.

Address correspondence to:Dr.Zaijie Jim Wang,Department of Biophar-maceutical Sciences,University of Illinois,833South Woods Street,Chicago, IL60612.E-mail:zjwang@uic.edu

CaMKII in Opioid Tolerance and Dependence909