Molecular mechanism: ERK signaling, drug addiction and behavioral effects

Abstract

Addiction to psychostimulants has been considered as a chronic psychiatric disorder, characterized by craving and compulsive drug seeking and use. Over the past two decades, accumulating evidence has demonstrated that repeated drug exposure causes long-lasting neurochemical and cellular changes that results in enduring neuroadaptation in brain circuitry and underlie compulsive drug consumption and relapse. Through intercellular signaling cascades, drugs of abuse induce remodeling in the rewarding circuitry that contributes to the neuroplasticity of learning and memory associated with addiction. Here, we review the role of the extracellular signal-regulated kinase (ERK), a member of the mitogen-activated protein kinase, and its related intracellular signaling pathways in drug-induced neuroadaptive changes that are associated with drug-mediated psychomotor activity, rewarding properties and relapse of drug seeking behaviors. We also discuss the neurobiological and behavioral effects of pharmacological and genetic interferences with ERK-associated molecular cascades in response to abused substances. Understanding the dynamic modulation of ERK signaling in response to drugs may provide novel molecular targets for therapeutic strategies to drug addiction.

1. Introduction

Drug addiction is a chronic brain disease characterized by high relapse rates and compulsive drug use despite negative consequences. To date, there is no effective treatment for drug addiction. Understanding the neurobiological aspects underlying substance abuse provides a basis for developing potential therapeutic strategies targeting to drug addiction. Accumulating evidence demonstrates that drugs of abuse alter dopamine (DA) and glutamate (Glu) neurotransmission in the mesocorticolimbic system to exert their molecular and behavioral effects^1–3. DA neurons in the ventral tegmental area (VTA) and their descending projections to the nucleus accumbens, prefrontal cortex (PFC) and other limbic regions, including the hippocampus (HIPP) and amygdala (Amy), comprise the mesocorticolimbic system⁴, which is crucial for reward and reinforcement processing, motivation, and goal-directed behavior^5,6. The NAc and VTA also receive Glu output from the PFC. In addition, a reciprocal Glu connection is found between the PFC and Amy. The nigrostriatal pathway containing the DA projection from the substantia nigra to the caudate putamen (CPu/dorsal striatum) has also been implicated in molecular events, rewarding effects and habitual behavior of drug addiction^7,8.

The extracellular signal-regulated kinases (ERK) cascade, one of isoforms of mitogen-activated protein kinases (MAPK), is associated with the pathology of diseases due to its role in cell proliferation, differentiation, survival, and death^9,10. ERK contains two isoforms, ERK1 and ERK2, which are indicated as “ERK” throughout the current review. Otherwise, individual subtype of ERK is specified as ERK1 or ERK2. In 1996, the activation of ERK was first identified in the VTA after chronic morphine or cocaine administration¹¹. Thereafter, several lines of studies have focused ERK-mediated molecular signaling in response to various drugs of abuse during the last two decades. Herein, we review the alterations of ERK signaling induced by abused substances including cocaine, amphetamine, methamphetamine, marijuana, nicotine, and alcohol. In addition, most of these drugs have been shown to induce psychomotor changes, the ERK-associated molecular changes underlying drug-induced behaviors is also discussed. Further, due to the critical role of ERK on the neuroplasticity of learning and memory associated with addiction¹², its influence on the reinforcing, rewarding, and relapse/reinstatement of drug addiction is also described.

2. ERK signaling pathway

Initially, intracellular ERK signaling has been characterized to respond to extracellular stimuli and regulate cell proliferation and differentiation¹³. For example, once activated by growth factors or neurotrophins, the tyrosine kinase receptors recruit Ras family G-proteins and lead to sequential activation of Raf (MAPK kinase kinase), MEK (MAPK kinase), and ERK. Once ERK is activated, the phosphorylated (p)ERK protein can translocate to the nucleus¹⁴, where they phosphorylate the ternary complex factor Elk-1^15,16. The activated Elk-1 and other ternary complex factors associate with serum response factor, bind to the serum response element site, and promote immediate early gene (IEG) transcription related to neuroadaptation^17–19. In addition to Elk-1, through phosphorylating ribosomal S6 kinases and mitogen- and stress-activated protein kinases (pRSKs and pMSKs, respectively), ERK has been shown to indirectly result in cAMP response element-binding protein phosphorylation (pCREB), a transcription factor that has been shown to regulate gene expression^20–24. Increasing evidence shows a glutamate linkage to ERK signaling in neurons both in vivo and in vitro. For instance, through the elevation of intracellular calcium (Ca²⁺)/calmodulin (CaM)/CaM kinases (CaMK), the activation of the Glu N-methyl-D-aspartate receptor (NMDA-R) can increase the phosphorylation of MEK (pMEK)/ERK/Elk-1 in hippocampal slices, neuronal culture^25–27, cortical cultured neurons²⁸, and striatal cultured neurons^29–31. Inhibition of ERK activation attenuates Glu-mediated pElk-1 in the striatal slice³², striatum including the CPu^33–35 (wrong article for ref 35; should be Sgambato V. et al., J. Neurosci. 1998 18(21):8814–8825), and the HIPP¹⁷. Alternatively, in PC12 cells, Ca²⁺ may increase the intracellular cAMP through Ca²⁺/CaM-sensitive adenylyl cyclase (AC) leading to the activation of PKA. Increase of cAMP and PKA induces pMEK via the activation of Rap1/Raf^36,37. Consistent with these studies, pharmacological activation of DA D1 receptor (D1-R) or the AC markedly stimulates ERK activity and its phosphorylation in various neuronal cells^33,38–41. In addition, activation of group 1 metabotropic Glu receptors (mGluR1/5) has been shown to increase the intracellular Ca²⁺ and activate ERK signaling^42–45. Although the activation of DA D2 receptor (D2-R) inhibits PKA activity, D2-R stimulation also increases ERK signaling through PKC activation⁴⁶.

There are several families of ERK-related protein phosphatases (PPs). Among them, PP2A and striatal-enriched protein tyrosine phosphatase (STEP) are the best characterized. PP2A is a major serine/threonine phosphatase containing two regulatory subunits and one catalytic subunit. PP2A mediates a rapid inactivation of pERK in vitro. STEP is another phosphatase that regulates ERK activation. Although it is enriched in the CPu and NAc, STEP is expressed abundantly in the mesocorticolimbic system^47,48. Through direct interaction of a kinase-interacting motif, STEP and its non-neuronal homologues have been shown to dephosphorylate pERK and prevent its nuclear translocation^49,50. Phosphorylation of STEP (pSTEP) reduces its activity and its capacity to inhibit pERK⁴⁹. STEP is regulated through D1-R/PKA/DARPP-32 signaling⁵¹. In vitro, D1-R activation has been shown to activate pThr34 and inhibit pThr75 DARPP-32 via PKA-activated PP2A⁵², respectively. The activation of pThr34 DARPP-32 subsequently inhibits PP1 and thereby increasing pSTEP⁵³. In addition, stimulation of NMDA-R has been reported to induce Ca²⁺-activated PP2A and PP2B which inhibit DARPP-32 signaling^52,54,55 and indirectly modulate ERK activity. Therefore, the PPs of pERK are regulated by DA- and Glu-mediated transmission. Further, dual specificity MAPK phophatases 1 and 3 (MKP-1/3) are also implicated in pERK de-activation. Both in vitro and in vivo studies indicated that MKP-1/3 expression and activation is dependent on ERK signaling. Once induced and activated, MKP-1/3 reduces the ERK activation as an inhibitory feedback loop^34,56–61. Furthermore, there is evidence demonstrating that MKP-1 is phosphorylated (pMKP-1) by pERK leading to MKP-1 protein stabilization without altering its ability to dephosphorylate pERK⁶².

3. ERK signaling and drug addiction

ERK signaling is responsive to various abused drugs in the mesocorticolimbic system. Both acute and chronic exposure to drugs results in alteration of ERK-mediated signaling in specific brain regions underlying neuronal plasticity and drug-induced behavioral changes. Therefore, we focus on the effects of most prevalent abused substances on ERK signaling and its relationship of drug-mediated behavioral changes across different paradigms including locomotor activity/sensitization, conditioned place preference (CPP), and self-administration (SA), if applicable. Since pharmacological and genetic approaches have been used to interfere with the ERK signaling cascade, their effects on abused drug-mediated behaviors were summarized in Table 1 and Table 2, respectively.

Table 1.

Effects of MEK inhibitors on drugs-induced behaviors

Drugs	MEK inhibitors (dose, area)	Behavioral effects	References
None	SL327 (50 mg/kg, i.p.)	↑ basal locomotor activity	152
	SL327 (50–100 mg/kg, i.p.)	↓ basal locomotor activity	92,156,254,255
	PD98059 (50 μM, continuous infusion into the PFC)	↑ basal locomotor activity↓	154
Cocaine	SL327 (50 mg/kg, i.p.)	↓ acute cocaine-induced locomotion	78
	SL327 (30 mg/kg, i.p.); PD98059 (10 μM, VTA)	↓ development of locomotor sensitization (inhibitors were injected/infused before each cocaine injection)	92,94
	SL327 (40 mg/kg, i.p.); PD98059 (2 μg) or U0126 (1 μg, NAc)	↓ expression of locomotor sensitization (inhibitors were injected/infused before cocaine challenge)	80,93
	SL327 (30 mg/kg, i.p.)	↓ conditioned locomotor response (inhibitor was injected before each cocaine injection during conditioning)	92
	SL327 (50 mg/kg, i.p.); U0126 (0.1 μg, VTA)	↓ development of CPP (inhibitors injected/infused before each cocaine injection during conditioning)	78,100
	U0126 (1 μg, NAc core)	↓ expression of CPP (inhibitor was infused before CPP test)	107
	SL327 (30 mg/kg, i.p.); PD98059 (2 μg ) or U0126 (1 μg, NAc core); U0126 (1 μg, BLA)	↓ context- and cocaine priming-induced expression of CPP and ↓ context-induced reinstatement after SA by impairing memory reconsolidation (inhibitors were injected/infused either before or after reconsolidation phase)	82,107,112,113
	U0126 (1 μg, CeA)	↓ context+cues-induced relapse after abstinence from SA (inhibitor was infused before relapse testing)	116
	U0126 (1 μg, VTA)	↓ BDNF/GDNF-enhanced relapse by context+cues after abstinence from SA (infusions were conducted immediately after the end of the last SA session)	118,119
	U0126 (0.5 μg, dmPFC)	↓ BDNF’s inhibitory effect on context-, cues- and cocaine priming-induced drug seeking after abstinence/extinction of SA (infusions were conducted immediately after the end of the last SA session)	121
Amphetamine	SL327 (50–100 mg/kg, i.p.)	↓ acute amphetamine-induced locomotion	78,140,141
	PD98059 (50 μM, continuous infusion into the PFC)	↑ acute amphetamine-induced locomotor activity	154
	SL327 (40 mg/kg, i.p.)	↓ expression of locomotor sensitization (inhibitors were injected/infused before amphetamine challenge)	80
	SL327 (30 mg/kg, i.p.)	↓ conditioned locomotor response (inhibitor was injected before each amphetamine injection during conditioning)	92
	PD98059 (2.5 μg, NAc )	↓ development of intra-NAc amphetamine-induced CPP (inhibitor was infused before or after each intra-NAc amphetamine infusion during conditioning)	160
	PD98059 (2 μg, NAc )	↓ expression of amphetamine-CPP (inhibitor was infused before CPP testing)	180
Marijuana (THC)	SL327 (50 mg/kg, i.p.)	↓ development of THC-induced locomotion tolerance ( inhibitor was injected before each THC administration)	200
	SL327 (50 mg/kg, i.p.)	↓ development of THC-CPP (inhibitor was injected before each conditioning session)	196
Alcohol	PD98059 (30 or 90 μg, i.c.v )	↓ development of ACD-CPP (inhibitor was infused before each conditioning session)	253
	SL327 (30 mg/kg, i.p.)	↑ ethanol SA (inhibitor was injected before SA session)	255
	U0216 (0.5 μg, VTA)	↓ GDNF’s inhibitory effect on ethanol SA (infusions were conducted before SA session)	259

i.p. = intraperitoneal injection; i.c.v = intracerebroventricular infusion; ↑ = enhancing effect; ↓ = inhibiting effect

Table 2.

Effects of interfering ERK signaling-related genes/proteins on drug-induced behaviors

Target genes/proteins	Behavioral effects	References
Ca2+-stimulated AC1/AC8 (KO)	↑ acute cocaine-induced locomotion ↓ development of cocaine locomotor sensitization	124
Ras-GRF-1 (KO)	↓ development and expression of cocaine locomotor sensitization ↓ cocaine-CPP ↓ repeated THC-induced behavioral tolerance	127,199,200
Ras-GRF-1 (OE)	↑ development and expression of cocaine locomotor sensitization ↑ cocaine-CPP	127
Ras-GRF-2 (KO)	↓ ethanol intake and preference (two bottle free choice task)	242
ERK1 (KO)	↑ basal locomotor activity ↑ acute amphetamine-induced locomotion ↑ development of cocaine locomotor sensitization ↑ cocaine-CPP	34,128,152,153
ERK1 (KD in the PFC)	↑ basal locomotor activity ↑ acute amphetamine-induced locomotion	154
ERK2 (OE in the VTA)	↑ development and expression of cocaine locomotor sensitization ↑ cocaine-CPP	130
ERK2 (KD in the VTA)	↓ development and expression of cocaine locomotor sensitization ↓ cocaine-CPP	130
MSK-1 (KO)	↓ development and expression of cocaine locomotor sensitization ↑ cocaine-CPP	79
Inhibition of pElk-1	↓ development and expression of cocaine locomotor sensitization ↓ the establishment of cocaine-CPP	74

KO = knockdown; KD = knockdown; OE = overexpression; ↑ = enhancing effect; ↓ = inhibiting effect

3.1 Cocaine

Numerous studies have demonstrated that acute cocaine administration increases pERK in the CPu, NAc, PFC, central and basolateral Amy (CeA and BLA, respectively), HIPP, and bed nucleus of the striatal terminals (BNST)^63–77. The increased pERK and its downstream targets including pMSK-1, pElk-1, pCREB, phosphorylation of GluN2B (pGluN2B) and IEGs by acute cocaine are dependent on the activation of MEK, D1-R/DARPP-32, and NMDA-R^{51,63,66,67,70,71,74,76,78–80}. In addition to pMSK-1 induction, the pRSKs in the CPu are also increased by acute cocaine leading to the indirect activation of CREB by pERK^77,79. In terms of PPs of pERK, acute cocaine has been shown to result in an increase of MKP-1 mRNA in the CPu and cortex⁸¹. In addition, depending on D1- and NMDA-Rs, the phosphorylation of MKP-1 was also enhanced in the CPu and NAc 45–60 min after acute cocaine, contributing to the transient pERK induction⁷⁶. Further, the pSTEP was also downregulated after acute cocaine in the CPu with corresponding pERK induction⁷⁷. Together, in a time dependent manner, the activation and inactivation of PPs is critical for controlling the acute cocaine-augmented pERK. Behaviorally, the acute cocaine-induced locomotor activity was not affected by the MEK inhibitor, SL327 (30 or 40mg/kg), but partially inhibited or not altered with a higher dose injection (50mg/kg) which has non-specific sedative effect on basal locomotion^{51,78,80,82,83}. Similar to acute cocaine, MEK/ERK activation is necessary for the chronic cocaine-induced IEG expression in the CPu, NAc and Amy in a time-dependent manner^66,67. In cocaine-sensitized animals, 7–21 days but not 1 day withdrawal resulted in increased Glu aminomethyl phosphonic acid receptor (AMPA-R) subunit surface insertion and NDMA-R subunit expression with paralleled pERK induction in the NAc^84–87 (also see⁸⁸). AMPA-R expression in the NAc after prolonged withdrawal from repeated cocaine injection is dependent on the activation of both GluN2B and pERK, which contributes to the development of behavioral sensitization⁸⁶. This conclusion is further supported by a study demonstrating that D1-R/Src kinase-mediated pGluN2B is necessary for the pERK induction in response to repeated cocaine administration⁷⁰. In addition, cocaine challenge after withdrawal from repeated cocaine administration also resulted in sensitized pERK in the CPu and NAc compared to the acute cocaine effect^72,89,90. The cocaine behavioral sensitization-induced pERK and pCERB in the NAc is dependent on ERK activation⁹¹. Further, the induction and expression of cocaine behavioral sensitization can be inhibited by systemic SL327 injection or intra-NAc MEK inhibitor infusion^80,92,93. Similarly through MEK activation, the pERK induction in the VTA is required for the development of behavioral sensitization to cocaine^11,94. Lastly, studies have indicated that, in response to D1- and NMDA-Rs activation, pERK induced by cocaine is responsible for the chronic cocaine-enhanced dendritic spine density and dendritic length in the CPu and NAc^95,96 providing the morphological evidence mediated by ERK signaling after repeated cocaine administration. Repeated pairing a specific environment with drug administration leads to a memory association between contextual cues and the drug rewarding effect. Subsequently, the context itself directly motivates drug-seeking behavior as a measurement of the reinforcing effect of the drug^97,98, which is associated with ERK signaling. For example, the acquisition of cocaine-CPP is accompanied by pERK induction in the NAc and PFC in a D1-R dependent manner⁹⁹. Systemic pre-administration of SL327 (50mg/kg) and a GluN2B antagonist prevented the development of cocaine-CPP^70,78, indicating the requirement of NMDA-R-mediated ERK activation in the formation of context-drug association memory. ERK activation in the VTA is necessary for the development of cocaine-CPP¹⁰⁰. Cocaine challenge in the drug-paired environment resulted in pERK and pCREB induction in the subset of neurons of the NAc¹⁰¹. In animals with repeated cocaine administration, the saline challenge enhanced pERK induction in the D1-positive neurons in NAc and CPu indicating context conditioning-induced ERK activity⁷². Similarly, after the establishment of CPP, CPP testing or re-exposure to the cocaine-associated context induced pERK, pCREB and/or ΔFosB in the CPu, HIPP, VTA and NAc as well as in D1-R containing neurons of the NAc^{100,102–106}. The CPP test-induced pERK expression in the VTA is dependent on mGluR1 activation and protein synthesis¹⁰⁶. Further, Miller and Marshall demonstrated that CPP test-elevated pERK and drug-seeking behavior were blocked by intra-NAc core infusion of U0126 (2μg/side) ¹⁰⁷. In the cocaine SA paradigm, context-induced relapse is also associated with enhanced pERK in the NAc core and CPu¹⁰⁸. Altogether, these results imply that, through ERK signaling, the NAc core and VTA are important for the memory formation of context-drug association. pERK in the NAc core and CPu also involve the retrieval of CPP memory and a general motor activation driven by drug-associated context, respectively.

Memory reconsolidation occurs when well-established drug-associated memories are recalled by re-exposure to drug associated context, cues, or the drug itself during which memories can be destabilized by adding new information or subjected to manipulation^109–111. The ability to disrupt drug-related memories provides an opportunity to promote treatment outcome and prevent relapse. The general procedure to test the memory reconsolidation on drug seeking behavior contains two phases: re-exposing animals to drug-associated context (phase 1) followed by testing drug seeking behavior after withdrawal (phase 2). A previous study demonstrated that, before or immediately after phase 1, intra-NAc core MEK inhibition through U0126 (1μg/side) or PD98059 (2μg/side) reduced cocaine-CPP during the phase 2. The expression of pERK, pCREB, pElk-1 and c-Fos induced by phase 2 is also attenuated with inhibiting ERK during phase 1¹⁰⁷. Systemic SL327 injection after phase 1 also decreased subsequent context-induced CPP in animals conditioned by escalating doses of cocaine¹¹². Similar to reactivation of CPP memory by context, the memory reconsolidation in response to cocaine is also accompanied by ERK activation in the PFC, NAc, and CPu. With or without cocaine priming, the systemic SL327 (20mg/kg) pre-treatment before phase 1 blocks the subsequent drug seeking behavior⁸². However, the effect of ERK on cocaine-induced memory reconsolidation is still dependent on the presence of context. Thus, the contribution of cocaine itself on memory reconsolidation is still ambiguous. After the establishment of cocaine SA, U0126 (1μg/side) infusion into the BLA immediately after phase 1 prevented context-induced reinstatement and the pERK induction after phase 2¹¹³. Taken together, these studies indicate that ERK signaling activated during memory reconsolidation is necessary for cocaine seeking behavior. However, a critical time window, 6 hr after the reactivation of memories, has been documented during which the memory is susceptible to alteration in the fear conditioning paradigm¹¹⁴. The pre-treatment before phase 1 may influence the memory retrieval instead of reconsolidation. If the ERK signaling actually involves in drug-related memory reconsolidation, the difference should be found when treatment is conducted within and beyond the critical time window in terms of both behavioral and molecular aspects.

Unlike pERK sensitization in cocaine-induced behavioral sensitization, immediately after the cessation of cocaine SA, there is an a dissociation between pERK induction and cocaine intake indicating the failure of developing pERK sensitization or tolerance, although with enhanced pERK expression in several brain regions¹¹⁵. However, ERK activation has been implicated in relapse after withdrawal. For example, the extinction test (conditioned cues + context) significantly increased pERK in the CeA and cocaine seeking behavior after 30 days withdrawal. Both enhanced pERK and relapse are dependent on MEK and NMDA-R activation¹¹⁶. Similarly, the pERK induction in the ventromedial PFC has been shown to mediate extinction test-induced cocaine seeking behavior after 1 or 30 days withdrawal from cocaine SA¹¹⁷. Through ERK activation, direct intra-VTA glial cell line-derived neurotrophic factor (GNDF) or brain derived-neurotrophic factor (BDNF) infusion immediately after the last session of cocaine SA induced robust drug seeking behavior after 3 or 10 days withdrawal^118,119. These results demonstrated that the potentiated ERK signaling underlies relapse behavior after cocaine SA. In contrast to augmented pERK induction in the PFC after 1 day abstinence of cocaine SA¹¹⁷, 2 hr after the last cocaine SA session, we have demonstrated a transient reduction of pERK in the PFC^120–122. The reduction of pERK is associated with an increase of STEP but not PP2A activity accompanied by decreased total GluN2B protein expression and phosphorylation, suggesting the inhibitory effect of STEP on pERK and NMDA-R¹²³. Through MEK activation and normalization of pERK in the PFC, direct BDNF infusion into the dorsomedial PFC immediately after the end of the last cocaine SA session resulted in a long-term inhibition on context-, cue- or cocaine-induced relapse¹²¹. Thus, it indicated that rescuing the ERK signaling or hypofunction in the PFC during early withdrawal might provide a potential therapeutic strategy for preventing cocaine relapse.

Several animal models have been used to dissect the ERK signaling cascade in cocaine-induced behavioral changes. For example, double knockout (KO) type 1 and type 8 Ca²⁺-stimulated AC resulted in a reduction of basal pERK in medium spiny neurons in the CPu with blunted acute cocaine-induced pERK, pMSK-1 and pCREB. Behaviorally, these double KO AC mice are supersensitive to low dose acute cocaine-induced locomotion and fail to develop behavioral sensitization in response to repeated cocaine administration¹²⁴. Ras-guanine nucleotide-releasing factors 1 (Ras-GRF1), the upstream activator of Ras, can increase ERK signaling. In the CPu and NAc, the protein expression of Ras-GRF-1 is increased by acute psychostimulants including cocaine^125,126. D1-R agonist and Glu-induced pERK is attenuated in the striatal slice of Ras-GRF-1 KO mice. The acute cocaine-induced pERK is downregulated and upregulated in Ras-GRF-1 KO and overexpressing (OE) mice, respectively. In addition, the development of cocaine behavioral sensitization and cocaine-CPP are attenuated in Ras-GRF-1 KO mice accompanied by a reduction of FosB/ΔFosB in the CPu and NAc. An opposite facilitation on behavior and FosB/ΔFosB was observed in Ras-GRF-1 OE mice in response to repeated cocaine¹²⁷. ERK1 KO mice exhibit higher responsibility to morphine¹²⁸. Similarly, in response to chronic cocaine exposure, ERK1 KO mice display enhanced behavioral sensitization and cocaine-CPP as well as c-fos mRNA induction in the CPu³⁴. This suggests that ERK1 acts as an inhibitor on ERK2 activation and a heightened stimulus- or cocaine-induced ERK2 signaling after ERK1 KO¹²⁹. In addition, selective ERK2 OE in the VTA resulted in an increase of sensitivity of cocaine-CPP and the repeated cocaine-mediated behavioral sensitization¹³⁰. In contrast, inhibition of ERK2 activity in the VTA attenuated the cocaine-CPP and the development and expression of cocaine-induced locomotor sensitization. Through activating MSKs, ERK leads to the increase of CREB activity. The acute cocaine-induced pCREB and IEGs as well as histone H3 phosphorylation were attenuated in the CPu and/or NAc of MSK-1 KO mice, indicating the role of MSK-1 in chromatin remodeling in response to cocaine. Although showing higher sensitivity to low dose cocaine-CPP, MSK-1 KO mice have reduced behavioral sensitization in response to repeated cocaine administration⁷⁹. Finally, systemic injection of the peptide inhibiting pElK-1 significantly inhibited acute cocaine-activated pElk-1, pElk-1 nuclear translocation, and histone H3 phosphorylation, as well as IEGs protein and mRNA expression in the CPu and NAc^74,131. Further, the inhibition of pElk-1 also resulted in an attenuation of repeated cocaine-induced dendritic plasticity in the NAc shell and prevented repeated cocaine-induced behavioral sensitization and CPP⁷⁴. Together, these studies demonstrated that ERK-associated signaling is important for the long-term cocaine-mediated behavioral alterations, rewarding effects and neuronal plasticity. Interestingly, the acute cocaine-mediated locomotor activity was not altered in animal models with manipulation of ERK1 or downstream molecular targets of ERK (e.g. MSK-1, ElK-1), further supporting that ERK signaling is not required for the acute cocaine-induced psychomotor effect.

Since both NMDA- and D1-Rs are implicated in cocaine-induced pERK, the direct protein-protein interaction between both receptors may underlie their effects on ERK activation^132–135. Previously, we have demonstrated the protein-protein interaction between D1-R and GluN1 of NMDA-R in the CPu. The D1-R/GluN1 complex is disrupted after acute cocaine administration which may underlie transient pERK induction by cocaine¹³⁶. The assumption is supported by a recent finding indicating that interference of D1-R/GluN1 association in vitro decreases D1 agonist- and NMDA-induced pERK induction. In addition, disrupting the protein-protein interaction in the NAc also attenuates acute cocaine-induced pERK induction and repeated cocaine-induced behavioral sensitization in the two injection protocol¹³⁷. Further, the receptor complex of sigma-1, histamine H₃, and D1-Rs has been found in the striatum. Through binding to sigma-1-R, cocaine results in a disinhibitory effect of histamine H₃ receptor on D1-Rs leading to pERK activation after either acute cocaine injection or cocaine SA¹³⁸. However, the impact of these receptor-receptor interactions on cocaine-induced behavioral alteration is still unknown.

3.2 Amphetamine

Acute amphetamine (AMPH) has been shown to increase pERK in the CPu, NAc, PFC, and VTA^{51,80,92,139–143}. Multiple upstream receptors and molecular activators have been implicated in acute AMPH-induced ERK signaling in a brain region specific manner. For instance, acute AMPH-induced pMEK and pERK in the striatum is regulated by D1-R/DARPP-32 and NMDA-R activation⁵¹. In contrast, pERK induction in the PFC by acute AMPH is dependent on NMDA-R, adrenoceptors and serotonin receptors but not D1- or D2-Rs¹⁴⁴. Blockade of mGluR1/5 or mGluR5 specifically in the CPu attenuated acute AMPH-induced pERK, pElk-1, pCREB, and Fos immunoreactivity^145–147. The activation of Ca²⁺/calmodulin-dependent protein kinases II (CaMK II) in the CPu is also necessary for acute AMPH-augmented pERK, pElk-1, and pCREB¹⁴⁵. Direct MEK inhibition via systemic SL327 (20–100 mg/kg) administration or intra-CPu U0216 (2μg/side) infusion attenuated acute AMPH-elevated pERK and pCREB protein expression in the CPu and NAc, and IEGs including preproenkephalin, preprodynorphin, and c-fos mRNA in the CPu^{80,140,141,148}. However, the differential pERK induction profile in the CPu in response to acute psychostimulants is determined by the environment: acute AMPH and cocaine induced pERK expression mainly in D1-R-expressing neurons^51,72,149, whereas, in a novel environment, AMPH dominantly increases pERK in D2-R-containing neurons of the striatum¹⁴⁸. In line with cocaine, PPs have been shown to be induced by acute AMPH administration which may control ERK activity after AMPH stimulation. For example, in the CPu, acute AMPH significantly increases pSTEP in a DARPP-32 dependent manner⁵¹. In addition, acute AMPH increases the gene encoding PP2B in the striatum including the CPu and NAc¹⁵⁰ relevant to MKP-1 mRNA expression and DARPP-32/STEP activity^53,151.

Behaviorally, similar to their enhanced response to rewarding properties of morphine and cocaine, ERK1 KO mice exhibit higher hyperlocomotion after acute AMPH injection^34,128,152. ERK1 KO mice display increased basal locomotor activity accompanied by a reduction of pRSK expression in the PFC and striatum^128,152,153, indicating a blunted ERK-mediated signaling after ERK1 ablation. The increased basal and acute AMPH-induced locomotion as well as the reduction of pRSK can be replicated by chronic and continuous infusion of MEK inhibitor, PD98059 (50μM), and selective knockdown of ERK1 in the PFC¹⁵⁴. Although the predominant hypothesis indicates that enhanced stimuli-activated ERK2 signaling in the CPu and NAc in ERK1 KO mice is responsible for increased behavioral responses to abused drugs^34,128, the reduction of ERK-mediated molecular cascade, at least in the PFC, may also contribute to both basal and drug-induced behavioral phenotype due to a general inhibition of ERK1 and ERK2 activity by MEK inhibitor. The latter assumption is supported by our recent finding demonstrating that rats raised in enriched environment have an augmented basal pERK induction in the PFC associated with lower basal and repeated nicotine-induced locomotion compared to control animals¹⁵⁵. The acute AMPH-induced hyperactivity was not altered by SL327 (30–40 mg/kg) but attenuated by high doses of SL327 (50–100 mg/kg) with a potentially inhibitory effect on basal locomotion^{80,92,140,141,156}. Although inhibiting acute AMPH-induced locomotor activity, acute systemic MEK inhibition by SL327 (50 mg/kg) resulted an enhancement to the basal locomotion¹⁵⁷. The discrepancy may be accounted for experimental procedure, since a potentiated acute AMPH-activated locomotor activity was documented after pERK suppression in the CPu of rats without habituating to the behavioral apparatus¹⁴⁷.

In a D1- and D2-Rs dependent manner, AMPH challenge after withdrawal from repeated AMPH exposure resulted in behavioral sensitization which is associated with pERK and pCREB sensitization in the CPu^158,159. The chronic AMPH-augmented pERK and pCREB induction is attenuated by D1- but not D2-Rs antagonist. Thus, although antagonism of both D1- and D2-Rs can inhibit the expression of behavioral sensitization, only D1-R-mediated ERK and CREB activation is critical for the expression of behavioral sensitization of the AMPH challenge. In contrast to the CPu, the expression of AMPH-induced behavioral sensitization is required for ERK’s inhibitory effect on CREB activity modulated by Ca²⁺ voltage-gated channels in the NAc⁸⁰. However, in the VTA, withdrawal from repeated AMPH exposure results in elevated MKP-1 and PP2B protein expression to downregulate the AMPH-mediated pERK induction¹⁴³. Systemic administration of SL327 (30 or 40 mg/kg) dose-dependently prevents the development and expression of behavioral sensitization as well as the acquisition of conditioned locomotor response to AMPH administration^80,92. A previous study demonstrated that intra-NAc AMPH infusion led to pERK and the establishment of CPP¹⁶⁰. The AMPH-CPP was prevented by direct intra-NAc PD98059 (2.5 μg/side) infusion either before or after each conditioning session, suggesting the role of ERK on memory acquisition and consolidation of association of contextual rewarding effect of AMPH. However, the enhanced locomotor response by intra-NAc AMPH infusion is not affected by MEK inhibition. Altogether, it seems that ERK plays an important role in chronic AMPH-induced behavioral alterations ranging from behavioral sensitization, conditioned locomotor response to CPP. However, dynamic molecular mechanisms underlying behaviors including ERK-mediated downstream targets and the modulatory effect of ERK-related protein phosphatases should be further elucidated in specific brain regions associated to AMPH.

3.3 Methamphetamine

Methamphetamine (METH) is a highly addictive psychostimulant causing a serious and growing worldwide problem associated with medical, socioeconomic, and legal domains^161,162. Although accumulating evidence has implicated the Glu and DA neurotransmission in METH-induced behavioral changes^163–167, a direct exploration of their downstream target, ERK signaling, is limited. Acute METH (3 mg/kg) injection significantly increases pERK in the striatum which is attenuated in serine racemase KO mice¹⁶⁸. Serine racemase is an enzyme synthesizing D-serine, an endogenous co-agonist of NMDA-R, thereby, partially supporting the requirement of NMDA-R for acute METH-induced pERK. In contrast, a recent study demonstrated that acute METH (2 mg/kg) did not affect pERK in either CPu or NAc¹⁶⁹. The dose of METH, routes of administration, or the timing of collecting tissue may contribute to the discrepancy.

METH challenge after withdrawal from repeated METH exposure has been shown to induce behavioral sensitization related to pERK induction in both CPu and NAc as well as ΔFosB expression in the CPu^122,169,170. The development and expression of METH behavioral sensitization and challenge-augmented pERK induction were inhibited by levo-tetrahydropalmatine, an antagonist of D1- and D2-Rs^169,171,172, suggesting the involvement of DA receptors in chronic METH-induced pERK and behavioral sensitization. However, the METH challenge-elevated pERK is associated to the consequences of acute stimulation, since the pERK protein expression in the NAc is transiently increased during early withdrawal or not altered after long-term abstinence^122,173. In agreement with the increase of pERK induction in the NAc shell after 1 day withdrawal from METH sensitization¹²², 2 hr withdrawal from METH SA resulted in elevated D1-R, pCREB, and ΔFosB protein expression as well as transcriptional regulating genes including CREB, Elk-1, and Fos family in the CPu^174,175. Genes associated with dual-specificity phosphatases 12 and protein tyrosine phosphatase were also upregulated, implying an inhibitory mechanism to dampen ERK signaling during the early phase of withdrawal from METH SA^175–177. In both D1- and NMDA-Rs dependent manners, acute or chronic METH administration results in increases of MKP-1 and MKP-3 mRNA in several brain regions including the PFC, orbital cortex, CPu, NAc, and HIPP^178,179. Therefore, the ERK-driven MKPs expression and other phosphatases represent a positive feedback to gate the transient ERK activation in response to acute or chronic METH exposure.

The increase of pERK, pElk-1, pCREB, and/or ΔFosB protein expression in the CPu, NAc or PFC is related to METH-induced CPP^180,181. Specifically, the acquisition of CPP and pERK induction in the NAc by METH-CPP require D1-R but not NMDA-R activation. Intra-NAc infusion of MEK inhibitor, PD98059 (2 μg/side), also prevents the expression of METH-CPP and pERK induction¹⁸⁰. Therefore, this demonstrates the importance of the activation of D1-R/MEK/ERK/pElk-1 in the NAc on the development and expression of METH-CPP. In contrast, the METH-CPP testing reduced pERK and pCREB in the NAc after a single pairing session with 2 days withdrawal¹⁸², suggesting either a compensatory reduction in response to overactivation of ERK signaling during conditioning and withdrawal or other molecular cascades are required for the initial acquisition of METH-CPP. Both assumptions should be further deciphered to identify molecular mechanisms underlying the difference between single and multiple condition session-mediated METH-CPP.

Chronic METH use causes cognitive deficits associated with altered neurotransmission^183–186. In animal studies, repeated METH administration leads to spatial learning and memory impairment, which is associated with reduced total ERK in the PFC¹⁸⁷. In addition, deficits in spatial working memory and novel object recognition (NOR) are accompanied by an inability of pERK induction in the HIPP and PFC by the learning process or stimuli^188–191. Interestingly, intra-PFC infusion of PD98509 (2 μg/side) mimics the METH-induced cognitive impairment in NOR¹⁸⁸, indicating that a reduced pERK signaling is responsible for the cognitive dysfunction after long term METH exposure. Several drugs have been demonstrated to ameliorate the cognitive deficit by METH through ERK signaling. For example, depending on nicotinic acetylcholine receptors (nAChRs), D1-R and MEK activation, galantamine, a drug used to treat Alzheimer’s disease by inhibiting acetylcholinesterase and allosterically modulating nAChRs, alleviates NOR impairment through pERK induction in the PFC¹⁹⁰. Similarly, modafinil, a cognitive enhancer with a weak DA transporter inhibiting effect, also activates pERK in the PFC to rescue the NOR deficit^191,192 probably through increasing extracellular DA levels. Finally, clozapine, an atypical antipsychotic medication, reverses dysfunctional pERK signaling in the HIPP with an attenuating effect on spatial working memory impairment induced by chronic METH¹⁸⁹. Taken together, these results demonstrate that the cognitive impairment induced by chronic METH is attributed to the downregulation of ERK signaling during learning and memory-a potential therapeutic molecular biomarker for future drug development.

3.4 Marijuana

Δ9-tetrahydrocannabinol (THC) is the main psychoactive component of marijuana which is one of the most used illicit drugs¹⁹³. Cannabinoid receptors 1 and 2 (CB1-R and CB2-R) have been identified and located mainly in neuronal and peripheral tissues, respectively. Activation of the CB1-R leads to the closing of Ca²⁺ and the opening of potassium channel, subsequently inhibiting AC and activating protein kinase including ERK¹⁹⁴. Acute low dose THC injection (1 mg/kg) has been demonstrated to increase pERK expression in the mesocorticolimbic system^63,195. Specifically, in the CPu and NAc, the THC-activated pERK is mediated by CB1-, D1-, D2- and NMDA-Rs indicating a synergistic action among cannabinoid, DA and Glu neurotransmission. Acute THC-induced ERK downstream targets, pElk-1 and zif268 mRNA, were inhibited by D1-R antagonist and the MEK inhibitor, SL327 (100 mg/kg). Further, in response to repeated low dose of THC injection, the development of THC-CPP was attenuated by SL327 (50 mg/kg), suggesting that ERK-regulated signaling is involved in THC rewarding effect¹⁹⁶. Similarly, acute THC- (1 mg/kg) induced transient pERK induction in the HIPP was dependent on the activation of CB1- and NMDA-Rs. SL327 (100 mg/kg) pre-treatment also inhibited the acute THC-induced IEG expression (c-Fos protein, Zif268 and BDNF mRNAs) in the HIPP¹⁹⁷. However, the relevance between acute THC-induced behavioral changes and ERK signaling should be further elucidated, since there is no significant locomotor activity alteration by low dose THC¹⁹⁸.

In contrast to the low dose of THC, high dose of THC (≥10 mg/kg) acutely resulted in hypolocomotor activity in the CPu and cerebellum with pERK, pCREB, and c-fos induction depending on CB1-R and Ras-GRF1^195,199,200. Although the ERK signaling in both brain regions is distinct from the acute THC-induced hypolocomotor, it is necessary for the development of behavioral tolerance, a gradually behavioral recovery from the initial hypolocomotor by acute THC, after repeated THC injection^199–201. In an ERK dependent manner, the behavioral tolerance is mediated by recruiting G protein-coupled receptor kinases and β-arrestins to desensitize and internalize CB1-R in the CPu and cerebellum. The chronic THC-mediated cerebellar synaptic transmission and plasticity as well as reduced sensitivity of CB1-R activation were also prevented in Ras-GRF1 KO mice²⁰². In addition, chronic THC exposure-induced pCREB and FosB protein expression in the PFC and HIPP is inhibited by either SL327 (50 mg/kg) or in Ras-GRF1 null mice²⁰¹. Taken together, the results demonstrated that, in response to a high dose of THC, the activation of pERK-mediated signaling in the CPu and cerebellum is critical for the development of behavioral tolerance. In the PFC and HIPP, the ERK-associated molecular cascade may underlie the addicted state for THC. However, the latter assumption needs to be examined due to a similar analgesic tolerance effect after a chronic high dose of THC. Since THC-induced behavioral sensitization and self-administration have been documented^203–205, it will be worthwhile to determine the role of ERK on reinforcing/rewarding effects of THC in specific brain region(s).

In addition to the THC action on CB1-R, dopamine agonist and psychostimulants have been shown to increase endocannabinoid release^206–208. A previous study has indicated that acute cocaine-induced pERK, and c-Fos in the CPu and NAc was inhibited by CB1-R antagonist pretreatment, and mice with CB1-R KO or conditional deletion in the forebrain neurons⁶⁸. In addition, the elevated pERK protein expression-induced by chronic cocaine in the VTA is dependent on CB1-R activation. The development of cocaine-CPP and underlying pERK induction were also inhibited by intra-VTA CB1-R antagonist infusion¹⁰⁰, implicating the role of CB1-R and endocannabinoids in regulating the rewarding effect of cocaine mediated by ERK signaling activation.

3.5 Nicotine

Cigarette smoking is the largest preventable cause of death and diseases worldwide with an estimated 6 million deaths each year²⁰⁹. It has been shown that nicotine, through activation of the DA- and Glu-related signaling in the mesocorticolimbic system, exerts its reinforcing effects^210–212. Several in vitro studies have demonstrated that activation of ERK and CREB by acute and chronic administration of nicotine depends on nAChRs, CaMKs, PKA, and MEK activity^213–217. A genome-wide expression analysis revealed acute nicotine exposure, through activation of ERK signaling, induced alterations of gene expression²¹⁸. Similarly, acute nicotine induced transient ERK activation through nAChRs, Ca²⁺ voltage-gated channels, CaMKs, and MEK in primary cortical and hippocampal neurons^219,220; however, only PKA is required for pERK induction by nicotine in the hippocampal neurons, suggesting differential upstream activators for ERK activity in distinct neuronal types. Chronic nicotine exposure in mesencephalic dopaminergic neuronal culture resulted in increases of dendritic length and soma size through nAChRs- and D3-R-recrutied ERK signaling²²¹, demonstrating that ERK involves nicotine-mediated structural neuronal plasticity. In vivo, acute nicotine administration increases pERK levels in the NAc, CPu, PFC, Amy, and BNST^51,63 (also see^222,223). In both the CPu and NAc, acute nicotine-induced pERK is mediated by D1-R/PKA/DARPP-32 signaling pathways^63,64, indicating the relevance of dopaminergic neurotransmission in response to nicotine. After chronic oral consumption of nicotine, the levels of pERK and pCREB were increased in the PFC, but pCREB was decreased in the NAc²²², suggesting an increase of PFC excitatory output into the NAc. Indeed, pERK was increased in the NAc of nicotine-induced CPP animals²²³, supporting the role of PFC-NAc projection in the conditioned rewarding effect of nicotine. Interestingly, a direct protein-protein interaction between α7nAChR and GluN2A has been identified in the HIPP, which can be upregulated by chronic nicotine exposure²²⁴. After nicotine self-administration, disruption of the α7nAChR-NMDA-R complex decreased ERK activity and blocked cue-induced reinstatement of nicotine seeking behavior²²⁴. Taken together, these results demonstrate that the ERK signaling pathway is a key integrator of the DA/D1-R and Glu/NMDA-R signaling that induces long-term cellular alterations and behavioral adaptation in response to nicotine exposure. However, a direct manipulation on ERK is warranted to examine its effect on nicotine-induced behavioral changes.

Environment is an important factor affecting the vulnerability for drug abuse^225–227. Exposure to an environmental enrichment paradigm results in neurobiological adaptations, particularly in the PFC of the mesocorticolimbic dopaminergic system^{155,228–230}. Our recent study has demonstrated that the basal level of pERK was higher in animals housing in an enriched environment condition compared with animals housing in an impoverished condition, which was negatively correlated with their respective baseline locomotor activities¹⁵⁵. After nicotine sensitization or nicotine SA, the pERK induction was significantly increased in the PFC of rats raised in either impoverished or standard condition. In contrast, due to their higher basal ERK activity in the PFC, nicotine did not alter pERK protein levels in animals raised in an enriched environment condition with a decreased sensitivity in response to chronic nicotine¹⁵⁵. Regardless of raising conditions, the pERK induction is positively correlated to the amount of nicotine intake during nicotine SA. Thus, these results suggest that pERK induction in the PFC may underlie the rewarding effect of nicotine which is consistent with a previous study demonstrating a preference for ERK-mediated signaling pathway activation in the PFC after nicotine SA²³¹.

3.6 Alcohol (ethanol)

In a time dependent manner, acute injection with higher doses of ethanol (EtOH, 2.5–4.7g/kg) reduced pERK and pCREB in the PFC, NAc, CPu, Amy, HIPP, cerebellum, and BNST in various ages of rodents^232–235. In contrast, acute administration of a lower dose of EtOH (1g/kg) significantly increased pERK in various regions including the NAc, and CeA in the D1-R and neuropeptide S receptor dependent manner^236,237. The acute EtOH-induced c-fos induction in the medial Amy was inhibited by the MEK inhibitor, U0126²³⁸. Similarly, acute acetaldehyde (ACD), the first and main metabolite of EtOH, enhanced pERK in the NAc, CeA and BNST through activation of D1-R and opioid receptors^239,240. Behaviorally, the low dose of acute EtOH (1mg/kg) is associated an anxiolytic response accompanied by the rapid increase of spine density in the CeA and medial Amy (MeA) through the BDNF-mediated TrkB phosphorylation and pERK/pElk-1/pCREB and Arc induction²⁴¹. In addition to the acute EtOH-mediated pERK signaling, its intrinsic activation state may contribute to the alcohol intake or preference. For example, despite the mixing results of acute EtOH-mediated pERK level in alcohol preferring animals, they have higher basal pERK level in the PFC and NAc as well as Ras-GRF2 expression, the upstream activator of MEK, in the brain compared to their alcohol non-preferring counterparts^242–245. The Ras-GRF2 KO mice exhibited lower EtOH intake associated with an aberrant DA transmission in the VTA-NAc projection mediated by ERK activation²⁴², revealing a functional role of ERK on acute EtOH-mediated DA signaling underlying the preference of alcohol.

The effect of chronic EtOH exposure on pERK is heavily dependent on administering paradigms, time of withdrawal and brain regions. Immediate cessation of repeated EtOH oral consumption and vaporized EtOH exposure has been demonstrated to decrease pERK in the PFC, NAc, CPu, Amy, and HIPP, although with an enhanced pERK induction after 7–11 hr withdrawal^235,246,247. In contrast, the chronic EtOH-attenuated neuronal plasticity during early withdrawal (e.g. within 1 day withdrawal) is associated with the downregulation of pERK and the inability of pERK induction in response to stimulus in the HIPP²⁴⁸. Similarly, a desensitization/tolerance of pERK or c-fos expression in response to EtOH re-exposure or challenge after withdrawal from repeated EtOH has been found in the PFC and HIPP^238,247. Further, a paralleled attenuated phosphorylation of GluN1 and CaMKII is also documented immediately after chronic EtOH exposure^235,247. Taken together, these results demonstrate that a reduction of Glu receptor-mediated ERK activity during early withdrawal leads to the desensitization of subsequent EtOH-induced pERK signaling. In contrast, Pandey and colleagues demonstrated that 24hr withdrawal from repeated EtOH consumption produced anxiety-like behavior followed by blunted BDNF/TrkB/pERK/pElk-1 and Arc protein expression with reduced spine density in the CeA and MeA²⁴¹. Intra-CeA BDNF infusion restored the early withdrawal-induced ERK signaling dysfunction and inhibited the anxiety-like behavior. Similarly, knockdown of the BDNF-mediated ERK signaling in the CeA and MeA induces anxiety and promotes EtOH intake²⁴⁹. This suggests that withdrawal syndrome after chronic EtOH consumption accompanied by physical signs and negative emotional state (e.g. anxiety, depression and irritability^250,251) may precipitate the relapse of EtOH intake. In addition, withdrawal from repeated EtOH has been demonstrated to result in an enhancement of fear conditioning depending on pERK activation in the BLA by NMDA-R and MEK activation²⁵². Thus, the ERK signaling in the Amy complex is important for the development and further acquisition of the negative affective state underlying the vulnerability for subsequent alcohol seeking behavior after withdrawal.

The role of ERK signaling pathway in EtOH-mediated rewarding effect has been documented. In the CPP model, the alcohol metabolite ACD-CPP is dependent on D1-R activation and the development of ACD-CPP can be attenuated by the MEK inhibitor, PD98059²⁵³. In contrast, a previous study has indicated that the systemic SL327 administration in the ERK-independent learning mechanism in EtOH-CPP did not affect the acquisition, expression and extinction of EtOH-CPP (2g/kg) as well as pERK after acute EtOH administration (2.5g/kg)²⁵⁴. However, EtOH-CPP has been shown to be established by the lower dose of EtOH (1g/kg) in D1-R dependent manner²⁵³, which is sufficient to induce pERK induction as described above. Based on the D1-R-activated pERK and significant pERK induction after acute EtOH, it is required to further test the effect of D1-R/ERK signaling on EtOH-CPP by the lower dose of EtOH. The ERK signaling is also implicated in the operant rewarding effect of EtOH. For instance, systemic MEK inhibition resulted in an increase of EtOH SA²⁵⁵, indicating the antagonism of acute pharmacological effect of alcohol promoting the drug taking behavior. After abstinence form EtOH SA, re-exposure to conditioned cues induced alcohol seeking behavior accompanied by pERK and c-Fos expression in the BLA²⁵⁶. The ERK signaling is also critical for the Glu transmission-mediated alcohol seeking. After extinction from EtOH SA, systemic mGluR5 inhibition attenuated cue-induced reinstatement and the cue-induced pERK expression in the BLA and NAc shell in alcohol-preferring rats²⁵⁷. Probably through restoring the Glu transmission, L-cysteine prevents EtOH SA and EtOH-primed-induced drug seeking²⁵⁸. In addition, the reinstatement-induced pERK in the NAc shell is also inhibited by the L-cystine pre-treatment. Finally, ERK activation in the VTA has been demonstrated to mediate the inhibitory effect of GDNF in preventing EtOH intake and reacquisition of EtOH SA after extinction²⁵⁹.

4. Conclusion and future direction

Drug addiction is a significant public health problem and has been considered as a chronic psychiatric disorder, characterized by craving and compulsive drug seeking and use. The main obstacle in drug addiction treatment is the cycle of relapse/reinstatement from drugs of abuse. This review summarizes the current understanding on the role of ERK signaling and its associated intracellular signaling pathways in drug-induced neuroadaptive changes underlying the rewarding and reinforcing mechanisms in response to abused drugs. Despite the differential regulatory pathways in which all drugs of abuse can affect ERK signaling, one evolving theme in all cases is the regulation of the ERK molecules at the phosphorylation level. It is therefore important to understand the precise mechanisms that underlie the regulation of ERK phosphorylation by different drugs of abuse. The ERK signaling pathway may play a critical role in the early intervention during withdrawal from chronic drug administration. For example, our recent studies demonstrated that normalizing the prefrontal ERK signaling pathway during the early withdrawal from repeated cocaine exposure leads to a long-term inhibitory effect on cocaine relapse^120,121 and restores the extracellular glutamate dysregulation in the NAc²⁶⁰. In contrast, after prolonged withdrawal from cocaine, an increase of PKA-mediated signaling is dominant in the PFC and NAc responsible for cocaine seeking^{85,121,261–263}, implicating that ERK activity in the PFC-NAc projection is dynamically regulated by multiple intracellular pathways. Future studies will evaluate the ability of novel therapeutic interventions to restore normal ERK signaling activity in the brain for inhibiting addictive drug-seeking behavior. On the other hand, environmental factors can also influence vulnerability to drug addiction. We demonstrated that environmental enrichment induces compensatory alterations of D1-R/DARPP-32 and ERK signaling pathways in the PFC, which may contribute to environmental enrichment-dependent reduction of susceptibility to nicotine^155,264. Although current knowledge of multiple factors regulating ERK activity has greatly expanded, many aspects of this regulation remain to be elucidated. For example, overexpression of microRNA-221 attenuates nicotine-induced pERK (unpublished data), whereas activation of ERK can regulate microRNA-221 expression²⁶⁵. Furthermore, ERK signaling has been associated with epigenetic mechanisms including chromatin remodeling through histone methylation and DNA methylation associated with drug addiction^266–268, which are critical for the regulation of gene expression, neuronal plasticity, and drug-induced behavioral alteration. In summary, several molecular signaling pathways are involved in the complexity of drug addiction; with ERK being the most highly characterized during the past two decades. Herein, we provide the general role of ERK-mediated molecular cascade in response to various abused drugs, but it is by no means exhaustive. Future studies are warranted to dissect the ERK signaling pathway providing a better understanding for the development of feasible and potential therapeutic strategies for drug addiction and related disorders.

Abbreviations

AMPH

amphetamine

DA

dopamine

Glu

glutamate

VTA

ventral tegmental area

PFC

prefrontal cortex

HIPP

hippocampus

Amy

amygdala

CPu

caudate putamen

NAc

nucleus accumbens

ERK

extracellular signal-regulated kinase

MAPK

mitogen-activated protein kinase

MEK

MAPK kinase

CB1-R

Cannabinoid receptor 1

CB2-R

Cannabinoid receptor 2

pERK

phosphorylated ERK

IEG

immediate early gene

RSK

ribosomal S6 kinase

MSK

mitogen and stress-activated protein kinase

CREB

cAMP response element-binding protein

pCREB

phosphorylated CREB

Ca²⁺

calcium

CaM

calcium/calmodulin

CaMK

CaM kinase

pMEK

phosphorylation of MEK

AC

adenylyl cyclase

mGluR1/5

metabotropic glutamate receptor-1/5

D1-R

dopamine D1 receptor

METH

methamphetamine

D2-R

dopamine D2-R

PKA

Protein Kinase A

PKC

Protein Kinase C

PP2A

Protein phosphatase 2A

PP2B

proteinphosphatase 2B

STEP

striatal-enriched protein tyrosine phosphatase

pSTEP

phosphorylation of STEP

DARPP-32

dopamine and cAMP regulated phosphoprotein-32

pThr75 DARPP-32

phosphorylation of DARPP-32 at Threonine 75

MKP-1/3

MAPK phosphatases 1 and 3

CPP

conditioned place preference

SA

self-administration

BNST

bed nucleus of the striatal terminals

pGluN2B

phosphorylation of glutamate receptor, ionotropic, N-methyl D-aspartate 2B

BDNF

brain derived-neurotrophic factor

Ras-GRF-1

Ras-guanine nucleotide-releasing factors 1

nAChRs

nicotinic acetylcholine receptors

THC

Δ9-tetrahydrocannabinol