Dopamine D2 autoreceptor interactome: targeting the receptor complex as a strategy for treatment of substance use disorder

Abstract

Dopamine D2 autoreceptors (D₂ARs), located in somatodendritic and axon terminal compartments of dopamine (DA) neurons, function to provide a negative feedback regulatory control on DA neuron firing, DA synthesis, reuptake and release. Dysregulation of D₂AR-mediated DA signaling is implicated in vulnerability to substance use disorder (SUD). Due to the extreme low abundance of D₂ARs compared to postsynaptic D₂ receptors (D₂PRs) and the lack of experimental tools to differentiate the signaling of D₂ARs from D₂PRs, the regulation of D₂ARs by drugs of abuse is poorly understood. The recent availability of conditional D₂AR knockout mice and newly developed virus-mediated gene delivery approaches have provided means to specifically study the function of D₂ARs at the molecular, cellular and behavioral levels. There is a growing revelation of novel mechanisms and new proteins that mediate D₂AR activity, suggesting that D₂ARs act cooperatively with an array of membrane and intracellular proteins to tightly control DA transmission. This review highlights D₂AR-interacting partners including transporters, G-protein-coupled receptors, ion channels, intracellular signaling modulators, and protein kinases. The complexity of the D₂AR interaction network illustrates the functional divergence of D₂ARs. Pharmacological targeting of multiple D₂AR-interacting partners may be more effective to restore disrupted DA homeostasis by drugs of abuse.

1. Introduction

Substance use disorder (SUD) is a chronic relapse disease characterized by compulsive and impulsive drug-seeking and taking behavior, inability to limit drug intake and emergence of negative affectivity during withdrawal (Koob & Volkow, 2010). SUD is a serious public health problem with devastating consequences to society. Opioid abuse has emerged as the deadliest drug crisis in the American history. There is an urgent need to improve our understanding of the neurobiological complexity underpinning SUD so that new, better pharmacotherapies can be developed.

Drugs of abuse exert their effects by disrupting signal transduction of multiple neurotransmitter systems and thus producing profound changes in the expression, localization, and function of many G-protein coupled receptors (GPCRs), transporters and ion channels [see review in (Ross & Peselow, 2009; Vetulani, 2001)]. The growing revelation of new proteins and novel mechanisms implicated in SUD has expanded the repertoire of pharmacological targets. However, monodrug therapies targeting individual proteins have not been effective. In recent years it has become apparent that receptors, transporters and ion channels do not operate alone in native tissues. It remains debatable whether or not these membrane proteins are assembled into functional and stable heterodimers or oligomers on the plasma membrane (Frederick, et al., 2015; Fuxe, et al., 2012; Kasai, et al., 2018; Lambert, 2010; Milligan, 2009). However, they do form functional complexes with other membrane proteins as well as cytoplasmic proteins including scaffold proteins and protein kinases to regulate intracellular signal transduction (Natarajan & Berk, 2006; Sager & Torres, 2011; Seagar & Takahashi, 1998). Disruption of the protein-protein interaction network induces aberrant cellular activity and signaling of specific proteins within the network. For example, it has been demonstrated in cultured cells and in vivo that there are reciprocal interactions between adenosine A2 receptors (A₂ARs) and dopamine (DA) D₂ receptors (D₂Rs) (Kull, et al., 1999; Salim, et al., 2000; Shen, et al., 2013; Taura, et al., 2018). Chronic cocaine exposure abolishes the functional interaction between D₂Rs and A₂ARs in rat striatum resulting in abnormal signaling for both DA and adenosine (Borroto-Escuela, et al., 2017; Pintsuk, et al., 2016). Moreover, human methamphetamine (METH) abusers show persistent deficits in DA function including reduced striatal D₂R availability (Volkow, et al., 1999). Interestingly, genetic variations in human ADORA2A, the gene encoding A₂ARs, are associated with METH dependence (Kobayashi, et al., 2010) and anxiety following amphetamine (AMPH) consumption (Hohoff, et al., 2005). These data suggest that responses to METH are at least in part dependent on the balance between DA and adenosine signaling mediated by D₂Rs and A₂ARs, respectively. Understanding the function of protein complexes at the molecular, cellular and behavioral levels will potentially provide valuable information for applying drug combinations to target multiple components of the protein complexes and their signaling pathways for future treatment of SUD.

A major neurotransmitter system involved in SUD is the DA system, which consists the mesocorticolimbic, nigrostriatal and tuberoinfundibular DA pathways. The mesocorticolimbic DA pathway originates from the ventral tegmental area (VTA) and projects to the nucleus accumbens (NAc), amygdala and prefrontal cortex (PFC). Altered DA signaling in the mesocorticolimbic DA pathway has been classically associated with the actions of drugs of abuse especially psychostimulants (Di Chiara & Imperato, 1988; Wise, 1996). The ability of cocaine and AMPH to increase midbrain DA neuron firing and augment synaptic DA concentrations in the NAc and PFC is the primary underlying mechanism for their reinforcing and rewarding properties. In addition, the mesocorticolimbic DA pathway regulates cognitive function and emotional behavior. The nigrostriatal DA pathway, projecting from the substantia nigra (SN) to the caudate putamen (CPu), is not only critically involved in motor function but is also implicated in brain-stimulation reward, intravenous drug self-administration and reward learning [see review in (Wise, 2009)]. Although these two DA pathways share common roles in regulation of reward function, the VTA/NAc and SN/CPu regions are interconnected with distinct subpopulations of neurons, and thus contribute differentially to DA-mediated activity (Haber, 2014; Wise, 2009). Lastly, DA neurons in the hypothalamus project via the tuberoinfundibular tract to the infundibulum and anterior pituitary, and DA within this pathway inhibits the release of prolactin (Moore, 1992).

Extracellular DA binds to DA receptors to induce diverse intracellular signaling events that control DA-related physiological functions including DA synthesis, locomotor activity, reward, learning and memory (Dalley & Everitt, 2009). The DA receptors are categorized into two classes with distinct pharmacological properties and signal transduction pathways: the D₁-like receptors (D₁Rs and D₅Rs) and the D₂-like receptors (D₂Rs, D₃Rs and D₄Rs). The D₁-like receptors are coupled to the heterotrimeric G-proteins Gαs or Golf to stimulate the activity of adenylyl cyclase (AC) and increase the synthesis of cyclic adenosine monophosphate (cAMP). The D₁R-like receptors are located on non-DA neurons, and modulate neuron excitability and neurotransmitter release via various mechanisms including cAMP-mediated activation of the second messengers including protein kinase A (PKA) and phosphoinositide, calcium mobilization, activation of calcium and sodium channels and induced trafficking of ionotropic glutamate receptors (Sidhu, 1998; Undieh, 2010). The D₂-like receptors are coupled to the Gαi/o-protein to inhibit the activity of AC and thus reduce intracellular cAMP production (Beaulieu, et al., 2015). The D₂-like receptors are present on DA and non-DA neurons and control neuron excitability and the release of neurotransmitters including DA, glutamate, GABA and acetylcholine.

Exposure to drugs of abuse such as cocaine produces marked changes in the expression, localization and function of D₁- and D₂-like receptors and these changes are linked to vulnerability to relapse (Anderson & Pierce, 2005; Ashok, et al., 2017). Thus, both D₁Rs and D₂Rs are potential therapeutic targets for treatment of SUD. The D₂Rs are categorized into two subtypes: D₂ autoreceptors (D₂ARs) and postsynaptic D₂ receptors (D₂PRs; or heteroreceptors) based on their cellular localization and functionality. Tremendous efforts have been devoted to delineate the role of striatal D₂PRs in the regulation of SUD [see review in (Ashok, et al., 2017)]; however, emerging evidence sheds light on an equivalent, important contribution of D₂ARs to SUD. The important role of D₂ARs in regulation of DA signaling has been elegantly reviewed (Ford, 2014). The objective of the current review is two-fold: a) to summarize the association of dysfunctional D₂AR with addiction-like behavior; and b) to describe the diverse functional interactions of D₂ARs with other membrane and intracellular proteins that mediate DA signaling and behavioral responses to drugs.

2. D₂ARs: Distribution and Function

2.1. D2ARs: Cellular and Anatomic Distribution

The D₂ARs are abundantly expressed in the somatodendritic compartments of midbrain DA neurons and presynaptically on their projections to the ventral and dorsal striatum. However, the D₂ARs are absent at terminal sites of DA neurons projected from the midbrain to the cortex (Chiodo, et al., 1984; Lammel, et al., 2008; Talmaciu, et al., 1986). The D₂PRs are concentrated in neurons receiving DA afferents in various brain regions including GABA neurons and acetylcholinergic interneurons in the striatum and cortex (Sesack, et al., 1994).

The D₂Rs exist in the long (D₂LR) and short (D₂SR) form by an alternative splicing at exon 6 of the same gene (Tan, et al., 2003). The D₂LRs differ from D₂SRs by the addition of 29 amino acids in the third intracellular loop, which contributes to their differential coupling to Gα subunits and downstream signal transduction shown in COS-7 cells (Montmayeur, et al., 1991). Because they are splice variants, the mRNA for both D₂SR and D₂LR is postulated and further confirmed to be expressed in the same DA neurons (Dal Toso, et al., 1989; Jang, et al., 2011; Montmayeur, et al., 1991; Neve, et al., 2013; Neve, et al., 1991) and non-DA neurons (Neve, et al., 2013). However, it is technically challenging to determine whether the proteins for D₂SRs and D₂LRs are equally expressed in the same neurons due to a lack of distinguishable antibodies and radioactive ligands for these two isoforms.

2.2. D₂ARs: Physiological Function

The D₂Rs belong to a subfamily of 7-transmembrane domain GPCRs that couple to Gαi/o proteins. Activation of the D₂Rs induces diverse downstream signaling via both Gαi/o and Gβγ subunits. Inhibition of AC activity, potassium channel activity, and mitogen-activated protein kinase activity is mediated by Gαi/o subunits whereas activity of ion channels, protein kinases and phospholipases is regulated by dissociated Gβγ subunits (Neve, et al., 2004). Due to their differences in cellular localization in vivo, the D₂ARs and D₂PRs have distinct functions in regulation of DA neuron firing, DA release and synthesis. The D₂ARs operate through a negative feedback mechanism to tightly control DA transmission. Somatodendritic D₂ARs directly couple to G-protein-coupled inwardly rectifying potassium (GIRK) channels via Gβγ subunits (Pillai, et al., 1998). Activation of D₂ARs opens GIRK channels, which increases the membrane potassium conductance and thus inhibits DA neuron firing (Beckstead, et al., 2004; Lacey, et al., 1987; Pucak & Grace, 1994). Axon terminal D₂ARs regulate exocytotic DA release via the voltage-gated Kv1 potassium channels. Activation of D₂ARs increases Kv1 currents, and inhibits neuron terminal excitability and probability for DA release (Cubeddu & Hoffmann, 1982; Schmitz, et al., 2003; Wolf & Roth, 1990). The D₂ARs modulate the opening of mouse striatal Kv1 channels via Gβγ subunits as blockade of Gβγ subunits abolishes D₂AR-mediated DA overflow and potentiation of Kv1 channel conductance (Fulton, et al., 2011). For D₂PRs on non-DA neurons, receptor activation regulates the release of neurotransmitters including GABA (Centonze, et al., 2002), glutamate (Bamford, et al., 2004) and acetylcholine (Bamford, et al., 2004), which in turn influences presynaptic DA release. It is important to note that D₂PRs act synergistically with D₁Rs to regulate cellular and behavioral responses to direct and indirect DA receptor agonists (Hasbi, et al., 2011); however, this is not the case for D₂ARs because D₁Rs are not expressed in DA neurons.

DA synthesis is regulated by D₂ARs through its inhibition of the activity of tyrosine hydroxylase (TH), the rate-limiting enzyme in DA synthesis (Bunney, et al., 1973). Although axon terminal D₂ARs have been considered to play a major role in DA synthesis in the striatum, somatodendritic D₂ARs are also implicated (Argiolas, et al., 1982). TH activity is primarily determined by its phosphorylation (Haycock & Haycock, 1991; Lindgren, et al., 2000). Activation of D₂ARs by quinpirole reduces basal and forskolin-stimulated TH phosphorylation at Ser⁴⁰ but not Ser¹⁹ and Ser³¹, suggesting that D₂ARs can inhibit TH phosphorylation via decreasing AC activity (Lindgren, et al., 2001). However, a recent study shows that quinpirole stimulates TH phosphorylation at Ser⁴⁰ in mice that have either D₂AR deletion in DA neurons or D₂PR deletion in striatal GABA neurons (Anzalone, et al., 2012), indicating that both D₂ARs and D₂PRs are involved in the feedback mechanism in regulation of DA synthesis.

It appears that the D₂SRs and D₂LRs function equivalently as autoreceptors. Both isoforms have similar binding affinities for DA and share the same canonical signaling pathways in HEK293 cells (Leysen, et al., 1993). Electrophysiological recordings on Xenopus oocytes co-expressing GIRKs and D₂SRs or D₂LRs are similar in DA inhibition of GIRK activation (Sahlholm, et al., 2008). Moreover, in rat primary mesencephalic neurons transfected with D₂SRs or D₂LRs, both isoforms inhibit neuronal firing and neurotransmitter release with similar potencies (Jomphe, et al., 2006). Lastly, these two isoforms are equally functional as autoreceptors in inhibition of GIRK currents in the SN of D₂R knockout mice expressing either D₂SRs or D₂LRs (Neve, et al., 2013). However, several lines of evidence also suggest that D₂SRs and D₂LRs are molecularly and functionally distinct from each other. First, D₂SRs are more susceptible to agonist-induced internalization than D₂LRs in cultured cell lines and primary cultured mouse DA neurons (Ito, et al., 1999; Itokawa, et al., 1996; Liu, et al., 1992; Morris, et al., 2007; Thibault, et al., 2011). Second, these two isoforms differ in their posttranslational modifications and intracellular transport in CHO cells (Fishburn, et al., 1995). The D₂LRs are processed to the matured form (fully glycosylated, ~70 kDa) much slower than are D₂SRs. A notable portion of the intracellular D₂LR pool is partially glycosylated and remains in the endoplasmic reticulum and Golgi without reaching to the plasma membrane. Therefore, these two isoforms are coupled to different proteins that mediate post-translational processing and intracellular compartmental transport, which may contribute to their differential responses to agonist-induced signaling transduction and internalization. Third, D₂SRs are more susceptible to desensitization than D₂LRs. Acute cocaine exposure reduces D₂SR inhibition of GIRK currents in the SN of D₂LR knockout mice but has no effect on D₂LR activity in D₂SR knockout mice (Robinson, et al., 2017b). Moreover, D₂SR-, but not D₂LR-mediated desensitization of GIRK currents is calcium dependent (Gantz, et al., 2015). Finally, using D₂LR and D₂SR knockout mice, the Borrelli group elegantly delineates the differential roles of these two isoforms in regulation of pre- and post-synaptic DA signaling (Radl, et al., 2018). While D₂SRs function as autoreceptors on DA neurons to control DA synthesis via TH phosphorylation, D₂LRs are necessary for postsynaptic D₂R-mediated locomotor response to stimuli (e.g. cocaine) and appear to act through regulation of striatal acetylcholine and activity of medium spiny neurons. These observations are in agreement with a previous report that the D₂SRs are predominantly expressed in the cell bodies and proximal and distal dendrites of DA neurons whereas the D₂LRs are postsynaptic, mostly present in the striatal neurons innervated by DA fibers (Khan, et al., 1998)

3. D₂ARs and Vulnerability to SUD

It has been well established that there is an inverse relationship between the availability of striatal D₂PRs and the vulnerability for drug intake in humans (Martinez, et al., 2004; Shumay, et al., 2012; Volkow, et al., 1990; Volkow, et al., 1999), non-human primates (Nader & Czoty, 2005; Nader, et al., 2006), rats (Maggos, et al., 1998; Tsukada, et al., 1996), and mice (Bailey, et al., 2008). However, accumulating evidence also suggests an apparent association of D₂AR expression and activity with vulnerability to drug abuse. A recent publication clearly shows that a partial Drd2 deletion in rat VTA causes a global elevation in functional brain activity (Martin, et al., 2020), demonstrating the significant impact of altered activity of D₂Rs in the midbrain on brain networks that are involved in processing reward, motor, cognition and emotional information.

3.1. D₂ARs and Novelty-seeking

Novelty seeking is commonly used to assess individual vulnerability for drug-taking behavior (Piazza, et al., 1989). Animals can be characterized as high- or low-responders based on novelty-induced locomotor activity. High-responders exhibit a greater and longer duration of locomotor activity than low-responders do, and are more likely self-administer abused drugs including cocaine (Flagel, et al., 2016; Marinelli & White, 2000), AMPH (Piazza, et al., 1989; Piazza, et al., 1998), morphine (Xigeng, et al., 2004), and ethanol (Nadal, et al., 2002).

Novelty-seeking behavior is mediated in part by D₂AR-mediated DA transmission. In humans, individuals high on novelty-seeking possess low availability of midbrain D₂Rs/D₃Rs (mostly autoreceptors) measured by positron emission tomography (PET) of [¹⁸F]fallypride, a D₂R/D₃R antagonist (Zald, et al., 2008). Ex vivo autoradiography of [¹⁸F]Fallypride binding shows that high-responding rats display lower levels of D₂Rs in the midbrain when compared to low-responding rats (Tournier, et al., 2013). Moreover, D₂AR knockout mice exhibit hyperlocomotor activity in a novel environment (Bello, et al., 2011). Low D₂AR activity in high novelty-seeking animals contributes to high basal firing and bursting activity of DA neurons in these animals (Bello, et al., 2011; Marinelli & White, 2000). There is no difference in the binding of the D₃R-preferring antagonist [³H]PHNO in rat midbrain between high- and low-responders (Tournier, et al., 2013), suggesting novelty-seeking behavior is primarily mediated by D₂ARs instead of D₃ARs.

High novelty-seekers are associated with increased axon terminal DA release presumably mediated by subsensitive presynaptic D₂ARs. Low doses of quinpirole preferentially activate D₂AR on DA terminals resulting in inhibition of DA release and suppression of locomotor activity induced by novel environment (Eilam & Szechtman, 1989; Hu & Wang, 1988; White & Wang, 1986). Thus, novelty-induced locomotor suppression by quinpirole is commonly used as a measure for presynaptic D2AR activity. High-responders are less sensitive to quinpirole inhibition of novelty-induced locomotor activity than are low-responders (Tournier, et al., 2013). High-responders also display higher striatal basal DA levels and greater DA release by AMPH than low responders do (Bradberry, et al., 1991; Braff, et al., 1992; Tournier, et al., 2013). Furthermore, in drug naïve healthy humans, the amount of striatal DA release induced by AMPH treatments correlates with the intensity of novelty-seeking, drug craving and cognition deficits (Leyton, et al., 2002; Riccardi, et al., 2006). Collectively, these data suggest that high novelty-seeking is associated with low levels and/or reduced activity of D₂ARs. The D₂AR subsensitivity confers elevated DA signaling, which likely subserves novelty-seeking and propensity to drug-taking behavior.

3.2. D₂ARs and Impulsivity for Drugs

A prominent behavioral phenotype of drug addicts is impulsive behavior or lack of self-control. Impulsivity is characterized by inability to wait, altered time perception and insensitivity to long-term rewards despite the negative consequence (de Wit, 2009). Impulsivity is considered to be an important factor contributing to relapse (Volkow, et al., 2010). Experimentally, impulsivity is frequently measured by delay discounting task which assesses preference for smaller immediate rewards than larger delayed rewards (Bickel & Marsch, 2001). High impulsivity is associated with choice for an immediate reward. Human patients addicted to cocaine (Garcia-Rodriguez, et al., 2013), METH (Hoffman, et al., 2006), nicotine (Bickel, et al., 1999), opioids (Kirby, et al., 1999), and alcohol (Mitchell, et al., 2005) all favor smaller immediate rewards over larger delayed rewards as compared to healthy subjects. High impulsivity measured by delay discounting is a clinical relevant model for understanding why drug addicts choose the transient effect of drug reward without thinking about the negative long-term consequences.

DA signaling in the PFC is implicated in impulsivity. During the delay discounting performance, there is an increase in the DA metabolite 3,4-dihydroxyphenylacetic acid in rat orbitofrontal cortex (Winstanley, et al., 2006), suggesting DA utilization in decision making of delay discounting. In agreement with this notion, acute AMPH treatment to healthy humans deceases discounting of delayed reward when compared to placebo treatment (de Wit, et al., 2002). However, high impulsivity in humans is also associated with greater AMPH-induced DA release in the striatum (Buckholtz, et al., 2010). Thus, DA transmission in both the mesolimbic and mesocortical pathways appears to mediate impulsivity although through different mechanisms. Low levels of D₂ARs in the midbrain are linked to high impulsivity and drug craving in human (Buckholtz, et al., 2010), presumably due to diminished control of axon terminal DA release and midbrain DA neuron excitability. A recent study further reveals that there is a direct causal, inverse relationship between midbrain D₂AR levels and impulsivity (Bernosky-Smith, et al., 2018). When Drd2 was decreased in the VTA by an adeno-associated virus (AAV) encoding shRNAs against Drd2, rats exhibited greater choice impulsivity for the smaller, immediate reward as compared to their corresponding control animals.

3.3. D₂AR, Drug Self-administration and Reinstatement

A deficit in D₂AR activity results in increased DA release and enhanced response to cocaine. Conditional Drd2 autoreceptor knockout mice are more sensitive to cocaine-induced locomotor activity and cocaine-conditioned place preference as compared to controls (Bello, et al., 2011). These mice acquire cue-triggered self-administration of cocaine faster and show enhanced salience of cocaine- but not sucrose-paired cues. However, conditional knockout of Drd2 does not change the total cocaine intake in responding to 1 mg/kg/infusion cocaine in these mice (Holroyd, et al., 2015). In contrast, Drd2 knockdown in rat VTA via lentivirus-mediated shRNA delivery dose-dependently increases cocaine self-administration at doses between 0.02–0.56 mg/kg/infusion (Chen, et al., 2018). The differences in the animal species, the potential compensatory changes resulting from the complete Drd2 knockout, and the doses of cocaine may contribute to the observed discrepancy between these two studies. Interestingly, Drd2 knockdown in only the NAc increases cocaine intake at lower doses (0.02–0.07 m/kg/infusion) and has no effect when cocaine doses are higher than 0.07 m/kg/infusion. These data suggest the dissociable roles of D₂Rs in the VTA and the NAc in limiting high doses of cocaine consumption.

Relapse to drug-seeking behavior is in part mediated by neuroadaptations in mesocorticolimbic DA signaling following chronic drug exposure (Koob & Volkow, 2010). The activity of VTA DA neurons controls the strength of DA signaling in the mesocorticolimbic circuit and plays a key role in drug reinstatement. Blockade of ionic glutamate receptor-mediated DA neuron activation in the VTA prevents cocaine reinstatement (Sun, et al., 2005; Xue, et al., 2011). Because DA neuron activity is tightly controlled by D₂AR, deficiency in D₂AR activity would precipitate relapse. Microinjection of quinpirole into rat VTA dose-dependently decreases cocaine reinstatement and this process is reversed by co-administration of the D₂R/D₃R antagonist eticlopride (Xue, et al., 2011). Furthermore, prior exposure to AMPH in rat VTA desensitizes D₂ARs and potentiates AMPH-seeking behavior under a progressive ratio self-administration schedule (Vezina, et al., 2002). Interestingly, the persistency of D₂AR adaptation to chronic drug exposure is brain region- and function-specific. The reduced presynaptic D₂AR activity in inhibition of striatal DA release following chronic AMPH treatment appears to be transient because it only lasts 3–4 days (Wolf, et al., 1993). In contrast, AMPH treatment (i.p. 5 mg/kg, once or twice daily for 7 days) produces a marked reduction in D₂AR-mediated inhibition of DA neuron firing in rat VTA after 8 days of abstinence although to a lesser degree when compared to 24–32 hrs withdrawal from the final administration of AMPH (White & Wang, 1984).

Drug-induced midbrain D₂AR subsensitivity is likely the first step in a cascade of signaling events which, when repeated, might produce enduring changes in DA transmission in the midbrain and DA neuron projection areas. The D₂AR subsensitivity is attributed by many factors. Because D₂ARs are coupled to Gαi/o to signal, chronic drug exposure may attenuate the coupling strength between D₂ARs and Gαi/o proteins. For example, AMPH self-administration diminishes D₂R (mostly D₂AR)-stimulated G-protein activation in rat midbrain assessed by an antibody-capture [³⁵S]GTPγS scintillation proximity assay (Calipari, et al., 2014). Chronic exposure to drugs of abuse disrupts the lipid composition of plasma membrane and alters the compartmentalization of membrane proteins [see review in (Luessen & Chen, 2016)], which likely contributes to decoupling of D₂ARs from their cognate G-protein subunits. Alternatively, chronic drug exposure reduces D₂AR activity by accelerating agonist-induced receptor internalization and delaying receptor recycling. Although it is not feasible to investigate D₂AR internalization in vivo, it has been shown that D₂Rs undergo robust β-arrestin dependent internalization upon agonist stimulation in cultured cells (Kim, et al., 2004; Lan, et al., 2009; Namkung, et al., 2009a). However, a recent study also demonstrates that GFP-D₂ARs in mouse DA neurons desensitize but do not internalize in response to agonist treatments (Robinson, et al., 2017a). It would be interesting to investigate if chronic exposure to drugs of abuse (e.g. cocaine & AMPH) preferentially targets Gα-mediated over β-arrestin-mediated signal transduction of D₂ARs, which many explain the commonly observed D₂AR desensitization as a consequence of drug exposure.

4. The D₂AR Interactome

In recent years, many proteins have been identified to interact with D₂ARs including transporters, GPCRs, ion channels, intracellular signaling modulators and protein kinases. These interactions suggest that the function, trafficking and compartmentalization of D₂ARs can be regulated via a wide variety of molecular mechanisms (see Fig.1). Understanding these regulations will have important implications for appreciation of reward, motor, cognitive and affective functions tied to D₂AR-mediated DA signaling.

Figure 1.

The dopamine D₂ autoreceptor interactome. (A) In the somatodendritic compartment of dopamine (DA) neurons, the dopamine D₂ autoreceptor (D₂AR) inhibits DA neuron firing by activating GIRK channel activity via Gβγ subunit. Activation of the neurotensin receptor type 1/2 (NTS_1/2R) inhibits D2_AR-mediated GIRK currents. Activation of the nicotinic receptor (nAChR) increases DA neuron firing which is antagonized by D₂ARs. When DA neurons are depolarized, DA released via exocytosis in the midbrain produces an inhibitory postsynaptic current that is dependent on voltage-sensitive calcium channel (VGCC). (B) In the axon terminal compartment of DA neurons, D₂AR inhibits Kv1 channel-mediated DA release via Gβγ subunits. Activation of nAChR, angiotensin receptor type 1 (AT₁R) and NTS1_/2R attenuate D₂AR inhibition of evoked DA release. In contrast, activation of trace amine-associated receptor1 (TAAR1) promotes D₂AR activity. The VGCC-mediated neuron excitability is inhibited by D₂AR. Activation of D₂AR increases dopamine transporter (DAT) surface expression and reuptake function. In both the somatodendritic and axon terminal compartment of DA neurons, D₂AR is subject to phosphorylation by G-protein coupled receptor kinase (GRK), protein kinase C (PKC) and PKA. The D₂AR couples to calcium-sensitive proteins including neuronal calcium sensor-1 (NCS-1) which recruits GRK2 to the receptor in the presence of calcium. Lastly, D₂AR-mediated G protein signaling is regulated by regulator of G-protein signaling (RGS) proteins including RGS2, RGS4 and RGS6.

4.1. Dopamine Transporters (DATs)

The D₂ARs and DATs are the two primary mechanisms to terminate synaptic DA transmission and maintain synaptic DA homeostasis. The D₂Rs (either D₂SRs or D₂LRs) and DATs are both expressed in DA neurons and can physically and functionally interact with each other to control DA signaling. The DAT N-terminus directly couples to the third intracellular loop of D₂Rs (Lee, et al., 2007). Co-expression of D₂Rs and DATs in cultured cells increases DAT surface expression and DA reuptake (Bolan, et al., 2007; Lee, et al., 2007). Moreover, activation of D₂Rs by agonists increases surface DAT levels and DAT reuptake function in cultured cells (Bolan, et al., 2007; Chen, et al., 2013; Lee, et al., 2007). In vitro activation of D₂ARs in mouse dorsal striatum (Gowrishankar, et al., 2018) or whole striatum (Chen, et al., 2013) also increases DAT surface expression and function. In agreement with these observations, mice carrying a mutant DAT construct A599V exhibit constitutive D₂AR hyperactivity as well as increased basal surface DAT expression in the dorsal striatum (Gowrishankar, et al., 2018). This A599V mutation in DAT is associated with increased phosphorylation of Thr53 at the N-terminus of DATs, suggesting that altered DAT phosphorylation may underlie the disrupted interaction between D₂ARs and DATs.

The D₂R-stimulated increase in DAT surface expression results from accelerated DAT recycling instead of decreased DAT internalization and this process involves protein kinase Cβ (PKCβ) and extracellular signal-regulated kinase (ERK) (Chen, et al., 2013). In neuroblastoma 2a (N2A) cells, inhibition of PKCβ activity abolishes D₂R-activated DAT trafficking to the surface. Moreover, quinpirole-stimulated DAT trafficking is completely abolished in PKCβ knockout mice. D₂R stimulation increases phosphorylation of PKCβII but not PKCβI in striatal synaptosomes that are Percoll-purified to eliminate postsynaptic membranes. These data suggest that D₂AR-mediated DAT trafficking requires activation of PKCβ activity. Significantly, the laboratory of Gnegy has shown that activation of PKC, especially PKCβ, promotes rapid DAT trafficking to the surface in N2A cells upon short-term exposure to DAT substrates including DA and AMPH (Chen, et al., 2009; Furman, et al., 2009; Johnson, et al., 2005a). Therefore, PKCβ appears to act as a functional linker to transduce and propagate the signaling of D₂ARs to DAT in control of D₂R-mediated DAT trafficking and activity. In addition, inhibition of ERK activity prevents D₂R-stimulated DAT trafficking to surface in cultured cells (Bolan, et al., 2007; Chen, et al., 2013; Lee, et al., 2007) and mouse striatum (Chen, et al., 2013). Interestingly, deletion of PKCβ diminishes quinpirole-stimulated ERK phosphorylation whereas ERK inhibition does not influence PKCβII phosphorylation by quinpirole (Chen, et al., 2013), indicating that PKCβ is upstream of ERK. To summarize, these data suggest the PKCβ-ERK signaling cascade is required for D₂R-medidated DAT insertion on the plasma membrane.

Disruption of the D₂AR-DAT interaction in vivo alters D₂AR activity and DA-related behavior. The Dat1 knockout mice have a complete loss of D₂AR-mediated inhibition of DA neuron firing, evoked striatal DA release, and DA synthesis, and display hyperlocomotor activity (Jones, et al., 1999). Moreover, treatment with a membrane penetrant peptide that disrupts the D₂AR-DAT interaction in mice decreases synaptosomal DA reuptake and increases locomotor activity (Lee, et al., 2007). Accordingly, acute depletion of Drd2 in rat SN decreases DAT reuptake function, increases basal locomotor activity, and reduces locomotor responsiveness to high doses of cocaine (Budygin, et al., 2016). However, these observations are in contrast to a report that conditional Drd2 knockout mice do not show a change in DA reuptake kinetics (Bello, et al., 2011), suggesting potential compensatory changes resulting from a complete, persistent deletion of Drd2.

To date, most studies have focused on understanding D₂AR regulation of DAT trafficking and reuptake function; however, whether and how D₂AR influences the ability of DAT to efflux DA remain largely unknown. Because phosphorylation of DAT N-terminus is required for reverse transport of DA mediated by DAT (Guptaroy, et al., 2009; Khoshbouei, et al., 2004), stimulation of protein kinases by D₂AR activation can alter DAT N-terminus phosphorylation and thus may influence DA efflux. Mice carrying the DAT-A559V mutation display hyperactive D₂AR activity and increased DA efflux that can be blocked by the D₂R antagonist raclopride (Bowton, et al., 2010). Interestingly, D₂AR hyperactivity in this DAT variant parallels an increase in the basal activity of Ca2+/calmodulin-dependent protein kinase II (CaMKII) and DAT N-terminus phosphorylation. These data indicate that the anomalous DA efflux in the DAT-A559V mutant is mediated by CaMKII activity via phosphorylation of DAT and a change in DAT conformation from outward- to inward-facing. It would be interesting to further elucidate the role of D₂AR in regulation of DAT-mediated DA efflux and identify additional protein kinases involved. The tight interaction between D₂ARs and DAT is an important mechanism by which extracellular DA is regulated. An abnormal interaction between D₂ARs and DATs has been implicated not only in drug abuse (Buckholtz, et al., 2010) but also in several other brain disorders including attention deficit and hyperactivity disorder (Mazei-Robison, et al., 2008), Parkinson’s disease (Seeman & Niznik, 1990), and schizophrenia (Lee, et al., 2009). Thus, intracellular modulators that can strengthen this D₂AR-DAT interaction may be beneficial to restore abnormal DA signaling in these disease states.

4.2. GPCRs

4.2.1. Neurotensin type 1 and 2 Receptors (NTS1/2Rs)

Neurotensin is a neuropeptide abundantly expressed throughout the brain (Hokfelt, et al., 1984). Neurotensin binds to G-protein coupled neurotensin type 1 (NTS₁Rs, couple to Gαq, Gαi/o and Gαs) and 2 receptors (NTS₂Rs, couple to Gαq) and non-G-protein coupled NTS₃R (Vincent, et al., 1999). The action of brain neurotensin is mostly mediated by NTS₁Rs and NTS₂Rs. These receptors are widely distributed in the brain and notably localized in midbrain DA cell bodies and dendrites and in striatal DA axon terminals (Nicot, et al., 1994; Palacios & Kuhar, 1981; Quirion, et al., 1987). Midbrain DA neurons are also innervated by neurotensin-positive fibers projected from the forebrain (Geisler & Zahm, 2006). The neurotensin system has long been known to modulate DA transmission via diverse mechanisms [see review in (Tschumi & Beckstead, 2019)]. Neurotensin depolarizes midbrain DA neurons through activation of a non-selective cation conductance (Jiang, et al., 1994). The effect of neurotensin on DA neuron excitability is mediated by the glutamate, GABA and DA systems. Neurotensin exerts a NTS₁R-dependent, biphasic effect on glutamate signaling in VTA DA neurons in that that low concentrations augment and high concentrations reduce the currents (Kempadoo, et al., 2013). Genetic deletion of Ntsr1, encoding NTS1Rs, decreases DA neuron firing (Dominguez-Lopez, et al., 2019), suggests an excitatory role of NTS₁Rs on neurotensin-mediated DA neuron activity. Neurotensin also increases DA neuron excitability by decreasing GABA signaling at GABA_B receptor synapses in the midbrain, and this effect is mediated by NTS₂Rs (Stuhrman & Roseberry, 2015; Tschumi & Beckstead, 2018). Finally, neurotensin excites midbrain DA neurons through depression of D₂AR-mediated synaptic currents (Piccart, et al., 2015; Shi & Bunney, 1990; Shi & Bunney, 1991). Neurotensin enhances DA release at both somatodendritic and axon terminal sites of DA neurons (Fawaz, et al., 2009; Okuma, et al., 1983). A direct application of neurotensin to midbrain slices increases calcium-dependent DA neuron firing (Jiang, et al., 1994; Seutin, et al., 1989; Werkman, et al., 2000), which can be blocked by a NTS₁R antagonist (St-Gelais, et al., 2004). In addition, blockade of D₂Rs/D₃Rs by sulpride prevents neurotensin-induced DA release in rat NAc (Fawaz, et al., 2009), suggesting the functional antagonism of neurotensin on D₂AR activity.

Activation of NTS₁Rs by neurotensin reduces the potency of D₂AR agonists that may contribute to the antagonism of neurotensin on D₂ARs. In HEK293 cells co-expressing D₂LRs and NTS₁Rs, the D₂R agonist FAUC326 displays a 34-fold reduction in the binding affinity for D₂Rs when compared to cells expressing D₂LRs alone (Koschatzky, et al., 2011). Moreover, a neurotensin receptor agonist (e.g. JMV449) markedly reduces quinpirole inhibition of forskolin-induced stimulation of cAMP response element-binding protein in HEK293T cells expressing D₂LRs and NTS₁Rs (Borroto-Escuela, et al., 2013). Recent evidence indicates the existence of a constitutive NTS₁R-D₂R (D₂SRs or D₂LRs) complex on the membrane in cultured cells and animal brain tissues measured by Bioluminiscence Resonance Energy Transfer (BRET), confocal microscopy and coimmunoprecipitation assays (Borroto-Escuela, et al., 2013; Koschatzky, et al., 2011). Although it has yet to be determined whether there is a direct physical interaction between these two receptors, the existence of the NTS₁R-D₂R complex suggests that the antagonistic action of neurotensin on D₂AR activity is likely modulated by neurotensin-stimulated, NTS₁R-mediated intracellular signaling, resulting in reduced D₂AR activity such as decoupling of D₂ARs from G-proteins and rapid receptor internalization. Acute treatment of HEK293 cells expressing either D₂SRs or D₂LRs with neurotensin produces receptor phosphorylation, internalization and heterologous desensitization through activation of PKC (Thibault, et al., 2011). The involvement of PKC activation in neurotensin signaling may underlie the increased potency of quinpirole to activate the intracellular ERK signaling pathway in the presence of neurotensin (Borroto-Escuela, et al., 2013). These data collectively suggest an antagonistic interaction on the plasma membrane and a synergistic cytoplasmic, PKC-dependent interaction between D₂ARs and NTS_1/2Rs.

Aberrant neurotensin signaling has been associated with drugs of abuse through modulation of DA signaling, especially D₂AR-mediated signaling. Microinjection of neurotensin in rat midbrain reduces intracranial self-stimulation threshold, suggesting the neurotensin peptide promotes motivated behavior (Rompre, et al., 1992). Both NTS₁Rs and NTS₂Rs are involved in neurotensin-mediated drug-seeking and taking behavior. Microinjection of the NTS₁R/NTS₂R antagonist SR142948A into mouse VTA delays METH acquisition, attenuates METH intake during the training and maintenance of self-administration, and reduces drug-seeking behavior during extinction, cue-induced reinstatement, and METH intake under the progressive ratio schedule (Dominguez-Lopez, et al., 2018). Genetic deletion of NtsR1 reduces METH intake and motivation for METH in mice (Dominguez-Lopez, et al., 2019). Additionally, intra-VTA pretreatment with or systemic administration of the selective NTS₁R antagonist SR142948A prevents AMPH-induced behavioral sensitization (Panayi, et al., 2005; Rompre & Perron, 2000). Systemic treatment with PD149136, a NTS₁R agonist, decreases METH self-administration (Sharpe, et al., 2017). These data suggest that neurotensin-mediated activation of NTS1/2R signaling promotes drug-seeking and taking behavior whereas inhibition of neurotensin signaling diminishes it. Modulation of the neurotensin signaling is a mechanism by which drugs of abuse exert their effects. For example, METH self-administration increases neurotensin levels in the midbrain and striatum (Hanson, et al., 2012; Wagstaff, et al., 1996). The increased neurotensin content may dampen D₂AR activity resulting in an imbalance between DA and neurotensin signaling and increased vulnerability to drug-seeking and taking behavior. Future studies are necessary to establish the functional relationship between D2ARs and NTS₁Rs/NTS₂Rs in animal models of drug abuse. It would be also interesting to investigate whether there is a functional reciprocal relationship between D2ARs and NTS₁/₂Rs.Thus, targeting both the neurotensin-NTS₁R and DA-D₂AR signaling could be a promising approach to normalize DA signaling.

4.2.2. Trace Amine-Associated Receptors 1

Trace amines including p-tyramine, β-phenylethylamine, octopamine and tryptamine are endogenous amines present in the brain with low concentrations (Pei, et al., 2016). They bind primarily to trace amine-associated receptors 1 (TAAR1s), which couple to Gαs and Gαq (Berry, 2004; Borowsky, et al., 2001; Bunzow, et al., 2001; Lindemann, et al., 2008). The Taar1 mRNA is ubiquitous but with notable presence in the mesolimbic and nigrostriatal DA pathways (Borowsky, et al., 2001; Lindemann, et al., 2008). In contrast to majority of GPCRs that are localized on the membrane, TAAR1s are predominantly intracellular with an abundant expression within the presynaptic terminals of monoamine neurons (Barak, et al., 2008; Bunzow, et al., 2001; Miller, 2011; Xie, et al., 2007). The TAAR1s may traffic to the membrane to respond to agonist stimulation. Alternatively, TAAR1s may functionally interact with amines intracellularly.

The TAAR1s play an important role in regulation of DA signaling. Deletion of Taar1 or blockade of TAAR1s with antagonists increases the DA neuron firing rate in mouse midbrain and augments evoked DA release in the NAc (Bradaia, et al., 2009; Leo, et al., 2014; Lindemann, et al., 2008). In contrast, activation of TAAR1s reduces the firing frequency of mouse DA neurons (Revel, et al., 2011). Interestingly, TAAR1 activation does not alter basal DA release and reuptake in rat NAc slices (Asif-Malik, et al., 2017), suggesting that TAAR1-mediated DA signaling is absent when DA tone is low. Moreover, TAAR1s regulate DA-related behavioral responses to psychostimulants. Activation of TAAR1s in rodents decreases cocaine-stimulated locomotor activity (Revel, et al., 2012), blocks cocaine self-administration (Revel, et al., 2012), prevents cocaine reinstatement (Pei, et al., 2014; Pei, et al., 2015), and attenuates cocaine-induced increase in intracranial self-stimulation (Pei, et al., 2015). Moreover, Taar1 gene variants are directly associated with METH-induced effects including consumption, hyperthermia and toxicity (Miner, et al., 2017; Mootz, et al., 2020; Shi, et al., 2016; Stafford, et al., 2019).

The TAAR1s modulate DA transmission via D₂ARs. The TAAR1-D₂R complex is confirmed in rat midbrain as TAAR1s are immunoprecipitated by an antibody against D₂Rs (D₂ARs) (Harmeier, et al., 2015). In HEK293 cells overexpressing TAAR1s and D₂LRs, the presence of D₂LRs promotes the trafficking of TAAR1s from the cytosol to the membrane (Harmeier, et al., 2015). This observation is consistent with a report showing co-localization of these two receptors on the plasma member in HEK293 cells (Espinoza, et al., 2011). Moreover, the presence of D₂Rs is reported to have no influence on the overall expression of TAAR1s (Harmeier, et al., 2015), which contradicts to the reported reduction in the total expression of TAAR1s in the presence of D2Rs (Espinoza, et al., 2011). The discrepancy may result from the variation in the expression levels of TAAR1s relative to D₂Rs between these two experiments. Moreover, the TAAR1-D₂R complex modulates the signaling of D₂Rs. When TAAR1s are stimulated, β-arrestin is recruited from D₂Rs to TAAR1s to initiate TAAR1-activated, β-arrestin-mediated intracellular signaling including Akt phosphorylation and subsequent activation of glycogen synthase kinase-3β (GSK3β) activity (Harmeier, et al., 2015). Because of a lesser association with β-arrestin, application of the D₂R agonist results in increased Gαi/o signaling, which explains the increased affinity of D₂Rs for agonists when both receptors are expressed.

A functional TAAR1-D₂R (D₂AR) interaction has been confirmed in vitro and in vivo. A TAAR1 agonist potentiates quinpirole inhibition of DA release in wildtype mice but not in Taar1 knockout mice (Leo, et al., 2014). When D₂AR activation is measured by paired-pulse stimulation in the same study, Taar1 deletion results in a reduction in D₂AR activity, confirming that TAAR1s mediate DA signaling through D₂ARs. Furthermore, application of RO5256390, a TAAR1 agonist, blocks cocaine-induced inhibition of DA clearance in rat NAc, and this blockade is reversed by co-treatment with L-741,626, a D₂R antagonist, and GSK3 inhibitors (Asif-Malik, et al., 2017). Collectively, these data suggest that D₂AR activation is necessary for TAAR1 modulation of cocaine’s effects.

4.2.3. Angiotensin II type 1 Receptors

Although angiotensin II (AngII) is present in the brain and the peripheral tissue, the AngII levels in the brain are much higher than in the peripheral circulation (Hermann, et al., 1984). AngII is notably present in the somatodendritic and axon terminal compartments of DA neurons (Garrido-Gil, et al., 2013). AngII binds to Gαq-coupled angiotensin II type 1 receptors (AT₁Rs), which form a complex with D₂Rs in HEK293 cells (Martinez-Pinilla, et al., 2015). This AT₁R-D₂R complex appears to be functional. In HEK293 cells co-expressing AT₁Rs and D₂LRs, the AT₁R antagonist candesartan blocks D₂LR-mediated cAMP production, ERK activation and β-arrestin recruitment whereas the D₂LR antagonist prevents AT₁R-mediated Gαq signaling (Martinez-Pinilla, et al., 2015). The observation of the cross-antagonism for AT₁Rs and D₂LRs suggest that these two receptors may share some common signaling pathways. However, Martinez-Pinilla et al. (2015) also reported that co-expression of D₂LRs with AT₁Rs does not alter D₂LR signaling but influences AT₁R signaling. In the presence of D₂LRs, the coupling efficiency between AT₁Rs and Gαq is attenuated and the potency of AngII to stimulate intracellular calcium release via AT₁Rs is reduced. This reduction is reversed by the D₂LR antagonist raclopride, suggesting that D₂LRs antagonize the activity of AT₁Rs likely through the intracellular signaling. This mechanism may be applicable in vivo given that both AT₁Rs and D₂ARs are expressed in the axon terminals of DA neurons. The AT₁Rs plays a role in regulation of presynaptic DA signaling. Perfusion of AngII to rat striatum in vivo or treatment with AngII to striatal slices triggers DA release that is reversed by AT₁R antagonists (Brown, et al., 1996; Mendelsohn, et al., 1993). AngII-stimulated DA release may be mediated through inhibition of striatal presynaptic D₂AR activity given the functional antagonism shown in HEK293 cells. It would be interesting to investigate the functional interaction between AT₁Rs and D₂ARs on DA release in vivo. The AT₁Rs are also involved in METH-seeking and taking behavior. Blockade of AT₁Rs attenuates METH self-administration and reinstatement in rats (Xu, et al., 2019), implicating increased AT₁R activity resulting from METH treatment. Whether increased AT₁R activity following METH self-administration is due to attenuated presynaptic D₂AR activity warrants further investigation.

4.3. Channels

4.3.1. G-protein-coupled Inwardly-rectifying Potassium Channels

Activation of the D₂ARs inhibits midbrain neuron firing and decreases somatodendritic DA release primarily through activation of GIRK channels (Inanobe, et al., 1999; Lacey, et al., 1987; Uchida, et al., 2000). The D₂ARs directly couple to GIRK channels via G-protein βγ subunits (Pillai, et al., 1998). Although midbrain DA neurons express both D₂ARs and D₃ autoreceptors (D₃ARs), D₃ARs play a minimum role in GIRK channel activation because D₂R knockout mice exhibit the absence of quinpirole-mediated GIRK currents in the SN DA neurons (Davila, et al., 2003). This notion is further demonstrated in Xenopus oocytes that DA activation of GIRK currents is mediated by either D₂SRs or D₂LRs, but not D₃Rs, and this regulation is voltage-dependent (Sahlholm, et al., 2008). Midbrain DA neurons express transcripts for GIRKs1–4; however, GIRK2 and GIRK3 channel subunits are the most abundant (Davila, et al., 2003). It appears that GIRK2 subunits are primarily responsible for most of the DA receptor-mediated synaptic currents (Beckstead, et al., 2004; McCall, et al., 2017). In contrast, GIRK3 subunits are critical for GABA_B-mediated GIRK currents in DA neurons (Munoz, et al., 2016).

The D₂ARs modulate GIRK channel activity and behavioral response to drugs. Decreasing GIRK current in VTA DA neurons increases cocaine-induced locomotor activity (McCall, et al., 2019) and cocaine self-administration (McCall, et al., 2017). Accumulating evidence indicates that chronic exposure to drugs of abuse alters GIRK currents in midbrain DA neurons, which is partially attributed to disrupted somatodendritic D₂AR activity. For example, withdrawal from chronic ethanol exposure (2 g/kg, i.p., three times daily for 7 days) enhances D₂AR-mediated GIRK current in a calcium-dependent manner in C57BL/6J mice (Perra, et al., 2011). This may contribute to the hypodopaminergic state during withdrawal. In contrast, 2 weeks of short-access (1hr/day) and long-access (4 hrs/day) METH self-administration (0.0015–0.15 mg/kg/infusion) markedly reduce D₂AR inhibition of GIRK currents in both VTA and SN DA neurons from DBA/2J mice (Sharpe, et al., 2014). Similarly, chronic phencyclidine administration (5 mg/kg, i.p., twice/day for 7 days) to DBA/2J mice attenuates D₂AR inhibition of GIRK currents in SN DA neurons (Piccart, et al., 2019). One likely underlying mechanism for the desensitization of D₂AR-mediated GIRK currents following METH and phencyclidine treatment is the reduced coupling efficiency between D₂ARs and GIRK channels mediated by Gβγ subunits. Chronic exposure to drugs alters the expression of G-proteins (Luessen, et al., 2017; Sun, et al., 2015), which in turn can influence the strength of the coupling between D₂ARs and GIRK channels. Alternatively, chronic exposure to drugs of abuse may change the affinity of DA for D₂ARs. Binge pattern cocaine administration (15 mg/kg/day, i.p., 7 or 14 days) reduces the apparent affinity of the substrates for striatal D₂Rs but not the number of binding sites (Tsukada, et al., 1996), suggesting that chronic cocaine treatment alters the conformation or stability of membrane D₂Rs and thus attenuates substrate binding affinity. The D₂AR-dependent activation of GIRK channel conductance is desensitized following exposure to other DA-related drugs including cocaine (Dragicevic, et al., 2014; Gantz, et al., 2015) and L-DOPA (Dragicevic, et al., 2014) in rodents. Furthermore, acute cocaine treatment induces desensitization of D₂SRs, but not the D₂LRs, in mouse SN DA neurons (Robinson, et al., 2017b). These data suggest that desensitization of D₂AR-GIRK currents in midbrain DA neurons may be a common mechanism underlying drugs of abuse.

4.3.2. Voltage-gated Potassium Channels

The presynaptic DA release is unlikely to be regulated by GIRK channels because they are not commonly detected on axon terminals (Luscher, et al., 1997). Instead, voltage-gated potassium channels such as Kv channels are localized at these sites and are involved in DA release. The wide spectrum Kv channel blockers such as 4-aminopyridine (4-AP) and tetraethylammonium (TEA) enhance electrically evoked [³H]DA release from rat striatal slices (Cass & Zahniser, 1991), confirming the functional presence of voltage-gated Kv-type potassium channel at DA neuron terminals. The action of these voltage-gated potassium channels is mediated by D₂ARs because 4-AP-and TEA-evoked DA release is attenuated in the presence of quinpirole in rat striatum (Cass & Zahniser, 1991). Moreover, in primary cultured DA neurons, action-potential evoked glutamate-mediated excitatory postsynaptic currents are inhibited by D₂ARs and this effect is largely abolished by 4-AP but not the GIRK channel blocker barium (Congar, et al., 2002), implying that terminal DA release is primarily mediated by Kv channels.

Members of Kv1 subfamily are abundantly expressed in various brain regions (Veh, et al., 1995) and function to regulate neuron excitability and neurotransmitter release (Robertson, 1997). The laboratory of Dr. Trudeau has elegantly delineated the distribution and the function of Kv1 subunits in DA neurons (Fulton, et al., 2011). The Kv1.1, Kv1.2, Kv1.6 are all expressed on DAT-positive mouse DA neuron axons. Moreover, these channels regulate D₂AR activity because α-dendrotoxin, a Kv1.1, Kv1.2 or Kv1.6 subunit-containing potassium channel blocker, diminishes D₂AR inhibition of evoked DA release. The Kv1.2 subunits and D₂Rs can be coimmunoprecipitated in cultured COS cells and primary cultured DA neurons. Maurotoxin (10 nM), which inhibits Kv1.2-containing channels at this dose, attenuates D₂AR inhibition of evoked DA release. The Kv1.2 knockout mice exhibit a pronounced reduction in D₂AR-mediated inhibition of evoked DA release in striatal slices, further implying a specific role of Kv1.2 in regulation of DA release. Furthermore, the Gβγ subunit contributes to the coupling between axon D₂ARs and Kv1 channels. In DA neurons, Kv3 family subunits are also expressed (Dufour, et al., 2014); however, blockade of Kv3.1, Kv3.2 and Kv3.4 channels fail to modulate D₂AR-mediated inhibition of DA release (Martel, et al., 2011).

The effects of drugs of abuse on the expression, trafficking and function of voltage-gated potassium channels are largely unknown. A recent report shows that withdrawal from METH self-administration increases the gene expression and membrane protein localization of Kv1.1 and Kv1.3 in rat NAc (Jayanthi, et al., 2020). These changes are also accompanied by altered DNA methylation at the CpG-rich sites near the promoter region of Kcna1 and Kcna3, genes encoding Kv1.1 and Kv1.3. Additionally, some reports indicate that members of Kv1 channels are involved in behavioral and neurochemical responses to cocaine (Kourrich, et al., 2013), AMPH (Ghelardini, et al., 2003), METH (Qu, et al., 2014), and sucrose (O’Donovan, et al., 2019). Future research is necessary to further delineate the direct role of each Kv1 isoforms in regulation of D₂AR-mediated DA signaling and behavioral responses to drugs of abuse.

4.3.3. Calcium and Calcium Channels

The internalization, phosphorylation and function of D₂Rs are regulated either directly or indirectly by calcium or calcium channels. D₂Rs can physically and functionally interact with the neuronal calcium sensor-1 (NCS-1), a member of the neuronal calcium sensor family of EF hand calcium-binding proteins. The N-terminus of NCS-1 directly binds to the C-terminus of D₂Rs demonstrated in the yeast two-hybrid, GST-pulldown and coimmunoprecipitation assays (Kabbani, et al., 2002). Functionally, overexpression of NCS-1 in HEK293 cells markedly reduces basal D₂R phosphorylation, and attenuates DA-induced D₂R internalization without altering the basal surface receptor levels. The D₂Rs form a complex with NCS-1 in a calcium-independent manner in HEK293 cells and primary neurons; however, GRK2 is recruited to this complex in response to receptor activation in a calcium-dependent manner. Structural biology analysis suggests that NCS-1 has structural flexibility at its C-terminus and acts as a small scaffold protein by allowing both D2Rs and NCS-1 to bind simultaneously in the presence of calcium (Pandalaneni, et al., 2015). This notion is further validated in vitro. Activation of Cav1.3, a L-type calcium channel expressed in SN DA neurons (Olson, et al., 2005; Sinnegger-Brauns, et al., 2009), increases intracellular calcium levels, resulting in recruitment of GRK2 to the NCS-1-D₂R complex to modulate D₂R phosphorylation and internalization (Dragicevic, et al., 2014). Since NCS-1 modulates calcium-dependent neurotransmitter release (Chen, et al., 2001; Guild, et al., 2001; McFerran, et al., 1998) and calcium channel activity (Tsujimoto, et al., 2002; Weiss, et al., 2000; Weiss & Burgoyne, 2001), this physical and functional interaction between NCS-1 and D₂Rs mediates D₂AR-mediated, calcium-dependent DA release and DA neuron excitability. It would be interesting to investigate if chronic exposure to drugs of abuse alters this D₂AR-NCS-1 interaction and related DA signaling. In addition to NCS-1, D₂Rs form a complex with other calcium-sensitive proteins including S100B (Dempsey & Shaw, 2011; Liu, et al., 2008) and calmodulin (Bofill-Cardona, et al., 2000; Liu, et al., 2007), and these complexes regulate D₂R signaling although through different mechanisms.

When DA neurons are depolarized, calcium influx through voltage-gated channels promotes somatodendritic DA release, which in turn activates D₂ARs and inhibits GIRK currents. Prolonged activation desensitizes D₂AR and thus increases D₂AR-dependent GIRK currents. When intracellular calcium is chelated in mouse midbrain DA neurons, the amplitude of the D₂AR-GIRK current is greater and the rate of D₂AR desensitization is slower (Beckstead & Williams, 2007; Perra, et al., 2011; Sharpe, et al., 2014). Moreover, the calcium-dependent D₂AR desensitization involves activation of phospholipase C and CaMKII (Perra, et al., 2011). It is further revealed that D₂SR-, but not D₂LR-, mediated inhibition of GIRK currents in DA neurons requires calcium (Gantz, et al., 2015; Robinson, et al., 2017b). Drug exposure dysregulates calcium-dependent D₂AR activity. For example, desensitization of the D₂AR-GIRK current in mouse DA neurons is abolished by METH self-administration (Sharpe, et al., 2014) and repeated ethanol exposure (Perra, et al., 2011). Moreover, a single cocaine injection reduces calcium-dependent desensitization of D₂SRs, not D₂LRs, when each D₂R isoform is overexpressed in D₂R knockout mice (Robinson, et al., 2017b).

In the midbrain, DA neurons form dendrodendritic synapses with other DA neurons (Groves & Linder, 1983; Nirenberg, et al., 1996; Wilson, et al., 1977). When DA neurons are excited, exocytotically released DA in the midbrain produces the inhibitory postsynaptic current (IPSC) that is tetrodotoxin sensitive, calcium dependent and inhibited by D₂Rs but not D₃Rs (Beckstead, et al., 2004). Moreover, the DA-mediated IPSC is dependent on voltage-sensitive calcium channels as cadmium (a blocker for all types, 300 μM) completely blocks the IPSC. Although application of ω-agatoxin IVA (blocker for P/Q-type) has no effect on IPSC, a combination of blockers of P/Q-type and N-type (ω-conotoxin GVIA) in the presence or absence of nimodipine (a blocker for L-type) reduces the amplitude of IPSC. These data suggest that more than one voltage-sensitive calcium channel is involved in IPSC. Further, prolonged DA exposure produces a calcium-dependent, long-term depression of IPSC and this depression is mediated through desensitization of postsynaptic D₂ARs in midbrain DA neurons (Beckstead & Williams, 2007).

Ample evidence also suggests that intracellular calcium levels and voltage-gated calcium channels control DA volume transmission [see review in (Rice & Patel, 2015)]. The extent of calcium and calcium channel regulation is brain region-dependent, with nearly a full order of magnitude difference between the values of EC₅₀ for somatodendritic regulation of DA release (μM Ca2⁺ range) in the midbrain vs axon terminal DA release (mM Ca2⁺ range) in the striatum (Chen, et al., 2011). However, voltage-gated calcium channel blockers alone or in combination are less effective at reducing somatodendritic DA release in the SN compared to the striatal axonal release (Bergquist, et al., 1998; Chen, et al., 2006). Moreover, calcium and calcium channels can indirectly determine the activation of D₂ARs in both somatodendritic and axonal release region via control of DA volume transmission. As already discussed, the amount of DA in these regions determines activation of D₂ARs, which themselves are differentially distributed in expression (Hurd, et al., 1994) and function (Cragg & Greenfield, 1997) across regions for corresponding control of DA neuron activity and DA release.

4.3.4. Ionotropic Nicotinic Acetylcholine Receptors

Nicotine binds to nicotinic acetylcholine receptors (nAChRs), which are ligand-gated cationic channels. Neuronal nAChRs are heteromeric pentamers made of a heterogeneous family of eight different α (α_3–7) and β (β_2–4) subunits. These receptors are categorized into homomeric α7-containing nAChRs (α7nAChRs) and heteromeric non-α7 nAChRs. The mRNA of nAChR subunits is expressed throughout the VTA and other brain regions (Azam, et al., 2002; Collins, et al., 2009). The β2-containing nAChRs are present in the cell bodies and axon terminals of DA neurons in the SN (Jones, et al., 2001; Wonnacott, et al., 2000; Zoli, et al., 2002). The nAChRs comprising exclusively of the α7 subunit (α7nAChRs) are also present in DA neurons of rat midbrain (Pidoplichko, et al., 1997; Wu, et al., 2004; Yang, et al., 2009). Electron microscopic imaging of dual immunolabeling of α7-containing nAChRs and D₂ARs in the VTA reveals that these two receptors co-localize in the soma, axon and terminal of DA neuron as well as in other types of neurons (Garzon, et al., 2013). The α6-containing nAChRs are the other major nAChR subtype that are strongly expressed in midbrain DA neurons (Azam, et al., 2002; Mackey, et al., 2012). Activation of these nAChRs on DA neurons increase neuron firing which is antagonized by D₂ARs.

Selective activation of striatal cholinergic interneurons elicits DA release via nAChRs located on DA cell bodies and does not require VTA mediated cell firing (Cachope, et al., 2012; Threlfell, et al., 2012). This nAChR-mediated facilitation of DA release is inherent to ex vivo voltammetry work that uses electrical stimulation to elicit DA because of simultaneous activation of cholinergic interneurons and a supply of ACh at DA axon terminals. This explains how nAChR antagonists, or desensitizing concentrations of nicotine, reduce DA release to single pulse or low frequency electrical stimulations in this preparation. The magnitude of DA release is restored (and sometimes exceeds pre-drug, baseline levels) when higher frequency stimulations are applied to the slice. Notably, this nAChR-mediated, frequency-dependent release of DA is similar to what is observed following application of the D₂R agonist quinpirole (Zhang & Sulzer, 2004). Thus, D₂R agonism and nAChR antagonism facilitate paired-pulse ratios by selectively reducing the first stimulation in slice preparations, and may model amplification of phasic bursts relative to tonic DA release.

Perfusion of nicotine to rat NAc persistently increases the extracellular level of DA [see review in (Subramaniyan & Dani, 2015)]. Mice lacking the β2 subunit of nAChRs display striking hyperactivity in the open field, which suggests disrupted DA neurotransmission (Avale, et al., 2008). Given the prominent role of D₂AR in regulation of axon terminal DA release, there might be a functional crosstalk between D₂ARs and nAChRs in striatal DA neuron terminals. Nicotine-stimulated striatal DA release is mediated by both presynaptic nAChRs and D₂ARs as dihydro-β-erythroidine (DHβE), a non-α7 nAChR antagonist, or the D₂R/D₃R agonist quinpirole blocks DA release (Quarta, et al., 2007), suggesting that D₂ARs modulate nicotine-induced, non-α7 nAChR-mediated DA release. Moreover, DhβE abolishes the D₂R/D₃R antagonist raclopride-induced DA release, indicating that D₂ARs inhibit nicotine-stimulated DA release by blockade of non-α7 nAChR activity. This functional interaction is further supported by the existence of a protein complex between D₂ARs and β2 subunit of the nAChRs in HEK293 cells and rat brain (Quarta, et al., 2007). Immunocytochemistry analysis show that D₂ARs and non-α7 nAChRs are co-localized in the terminals of DA cells. Additionally, in rat striatal tissues, D₂Rs can be immunoprecipitated by α4 subunit in the presence of the β2 subunit of nAChRs and vise verse. It is worth noting that D₂ARs and the β2 subunit of nAChRs are also co-localized in the soma and dendrites of DA neurons, suggesting that these heteromeric receptor complexes in the VTA might play a functional role in regulation of DA neuron excitability.

Repeated activation of nAChRs via nicotine influences D₂AR activity. For example, somatodendritic D₂AR availability is increased in female, but not male, human smokers (Okita, et al., 2016). In contrast, chronic nicotine exposure to mice has shown no effect on D₂AR sensitivity determined by D₂AR-mediated locomotor suppression (Tammimaki, et al., 2006), or decreased D₂AR sensitivity in ex vivo studies measuring stimulated [³H]DA release (Harsing, et al., 1992). Most research that has been published does not differentiate between nicotine’s effect on presynaptic D₂ARs and postsynaptic D₂PRs on GABA or cholinergic interneurons. For example, either nicotine self-administration (0.5 mg/kg/day) or administration via osmatic minipumps (1.5 mg/kg/day) for 14 days increases high-affinity D₂R expression in the striatum (Novak, et al., 2010). These receptor levels are normalized and even drop below the baseline following a period of prolonged withdrawal. Moreover, Drd2 knockout mice showed no conditioned place preference or locomotor stimulation to nicotine (0.5 mg/kg) (Wilar, et al., 2019). These effects likely describe the interaction of nAChRs with D₂PRs rather than D₂ARs.

4.4. Regulator of G-protein Signaling (RGS) Proteins

All RGS proteins contain a RGS domain which binds directly to activated Gα subunits to accelerate the rate of GTP hydrolysis, rapidly terminating G-protein signaling and receptor activation (Hepler, 1999; Watson, et al., 1996). More than 20 subtypes of RGS proteins are distributed in a brain region- and neuron-dependent manner (Gold, et al., 1997; Hooks, et al., 2008), suggesting that modulation of GPCR signaling by RGS proteins is receptor-type and brain region-specific. Accumulating evidence shows that drugs of abuse alter the expression of RGS proteins, which likely contributes to disrupted GPCR signaling and aberrant behavioral responses to drugs [see review in (Traynor, 2010)]. Due to a lack of specific antibodies and inhibitors for majority of RGS subtypes, the in vivo functions of most RGS proteins are largely unknown. Among RGS subtypes, the long splice isoform of RGS9 (RGS9–2) is selectively expressed in striatal medium spiny neurons and co-localizes with D₂Rs (Kovoor, et al., 2005; Rahman, et al., 2003; Zachariou, et al., 2003). RGS9–2 controls the Gαi/o signaling activated by striatal D₂PRs (Cabrera-Vera, et al., 2004; Rahman, et al., 2003), and deficiency in RGS9–2 is associated with aberrant DA transmission, enhanced cocaine reward and motor dysfunction in mice (Kovoor, et al., 2005; Rahman, et al., 2003). Although there is yet a report on the direct regulation of D₂AR activity by RGS proteins, a few RGS subtypes (e.g. RGS2, RGS4 and RGS6) are present in DA neurons and may potentially regulate D₂AR function.

RGS2 was hypothesized to have a selective affinity for Gαq when single-turnover GTPase assays are performed in solutions with purified proteins (Heximer, et al., 1997; Nance, et al., 2013). Although putative residues within the RGS domain of RGS2 proteins are necessary for RGS2 specificity towards Gαq or Gαi/o, the structural property of the Gα helical domain critically governs the Gα-RGS2 specificity and is influenced by the experimental environment (Kasom, et al., 2018). Accumulating evidence strongly suggests that RGS2 also has an affinity for activated Gαi/o in biological complex systems. For example, RGS2 stimulates GTPase activity of activated Gαi1 only when Gαi/o-coupled M2 muscarinic acetylcholine receptors (mAChR) are reconstituted in phospholipid vesicles (Ingi, et al., 1998). Additionally, RGS2 couples to activated Gαi/o proteins (e.g. Gαi2) by aluminum fluoride in N2A cells and rat striatum in an immunoprecipitation assay (Luessen, et al., 2016). Functionally, RGS2 inhibits ERK phosphorylation mediated by M1 and M2 mAChRs in CHO cells (Ingi, et al., 1998). Moreover, RGS2 suppresses β2 adrenergic receptor-mediated Gαi/o signaling in HEK293 cells and cardiomyocytes (Chakir, et al., 2011). Lastly, RGS2 negatively modulates D₂R-mediated G-protein signaling (Luessen, et al., 2016). Knockdown of RGS2 via siRNAs in N2A cells stably expressing D2SRs increases quinpirole-stimulated [³⁵S]GTPγS binding and enhances inhibition of forskolin-stimulated cAMP production confirming the ability of RGS2 to antagonize D₂R-mediated Gαi/o signaling in a biological relevant system. Single-cell gene analysis shows that Rgs2 mRNA is predominantly expressed in DA neurons over GABA neurons in rat VTA (Labouebe, et al., 2007). Immunocytochemistry further confirms the presence of RGS2 proteins in TH-positive neurons in rat VTA (Calipari, et al., 2014). Although it has yet to be investigated if RGS2 directly modulates D₂AR activity in DA neurons in vivo, AMPH self-administration causes increased RGS2 membrane translocation and reduced D₂AR inhibition of evoked DA release (Calipari, et al., 2014). The increased RGS2 membrane translocation likely dampens D₂AR-mediated G-protein signaling and thus reduces D₂AR activity. Given that AMPH self-administration produces the most pronounced change in RGS2 mRNA levels, compared to other RGS subtypes, in rat VTA (Sun, et al., 2015), it would be interesting to investigate if RGS2 proteins are a potential target to restore D₂AR activity disrupted by drugs of abuse and thus prevent drug-seeking and taking behavior.

RGS4 has been identified as a key mediator of opioid reward and analgesia primarily through interactions with Gαi/o-coupled μ-opioid receptors (Traynor, 2012). However, RGS4 also controls Gαi/o signaling stimulated by D₂Rs in HEK293 cells (Min, et al., 2012). The N-terminus of RGS4 proteins binds to the third intracellular loop of D₂Rs and this direct coupling is required for RGS4 to inhibit quinpirole-stimulated Gαi/o activation and downstream signaling. The Rgs4 mRNA is present in midbrain DA neurons as well as in GABA neurons (Labouebe, et al., 2007). Whole body deletion of Rgs4 attenuates cocaine-induced conditioned place preference in male mice (Rorabaugh, et al., 2018) although this effect might be mediated by both D₂ARs and D₂PRs.

RGS6 is present in both SN and VTA DA neurons, albeit with a greater abundance in the SN (Bifsha, et al., 2014; Luo, et al., 2019). The Rgs6 knockout mice exhibit enhanced sensitivity to quinpirole-induced suppression of locomotor activity in the novel environment (Luo, et al., 2019), suggesting hyperactive D₂AR activity. The increased D₂AR activity is also consistent with enhanced TH phosphorylation and DA synthesis in these mice. Moreover, Rgs6 knockout mice consume less ethanol when given free access, show reduced ethanol-induced reward measured by conditioned place preference, and are less susceptible to alcohol-induced withdrawal (Stewart, et al., 2015). Given that ethanol increases or decreases D₂AR activity in DA neurons depending on the species, doses and treatment regimen (Karkhanis, et al., 2015; Mrejeru, et al., 2015), RGS6 may regulate the action of ethanol in part through modulation of D₂AR activity.

In addition to the canonical role as a GTPase accelerator for many RGS proteins, newly emerging evidence suggests that RGS proteins can regulate GPCR trafficking via their non-canonical functions. For example, RGS9–2 inhibits DA-mediated internalization of D₂Rs in HEK293 cells via its RGS and DEP (Dishevelled, Egl-10 and Pleckstrin) domains (Celver, et al., 2010). RGS2 siRNA knockdown in N2A cells abolishes quinpirole-induced D₂R internalization in parallel with persistent membrane localization of β-arrestin (Luessen, et al., 2016). RGS2 GTPase activity may influence receptor phosphorylation and binding to β-arrestin. Alternatively, RGS2 may alter phosphorylation of β-arrestin which is necessary for β-arrestin to be released from the membrane (Alloway & Dolph, 1999). This is an important question to address in the future because D₂AR endocytosis is critical for receptor desensitization.

4.5. Protein Kinases

The canonical mechanism underlying D₂R internalization and desensitization involves receptor phosphorylation by protein kinases. In heterologous expression systems, stimulation of D₂Rs by agonists results in receptor phosphorylation followed by receptor uncoupling from activated G-proteins, arresting binding, receptor internalization and recycling [see review in (Lachowicz & Sibley, 1997)]. However, D₂R desensitization can be independent of receptor internalization. In primary cultured DA neurons, quinpirole stimulation of D₂ARs produces desensitization without altering receptor internalization (Robinson, et al., 2017a). Regardless, D₂R phosphorylation is an early event that triggers receptor internalization, desensitization and resensitization. The D₂Rs are phosphorylated primarily by G-protein-coupled receptor kinases (GRKs) and the second messenger activated protein kinases including PKC, PKA and CaMKII. The functional consequence of D₂R phosphorylation on receptor internalization, desensitization, recycling and resensitization relies on protein kinase-specific phosphorylation residues on D₂Rs.

4.5.1. GRKs

In heterologous expression systems, GRK2 and GRK3 can phosphorylate D₂Rs when activated by agonists (Ito, et al., 1999; Kim, et al., 2001; Li, et al., 2015b; Namkung, et al., 2009a); however, GRK4, GRK5 and GRK6 have no effect (Li, et al., 2015a; Namkung, et al., 2009a). Mutation of the serine/threonine residues within the third intracellular loop of D₂Rs that are substrates for GRK2/3, abolishes agonist-stimulated D₂R phosphorylation (Namkung, et al., 2009a), further suggesting that GRK2 and GRK3 are the primary GRKs to regulate D₂R phosphorylation.

GRKs play a role in D₂R trafficking including internalization and recycling. GRK2 promotes DA-induced D₂R internalization in HEK293 and COS-7 cells (Cho, et al., 2010; Iwata, et al., 1999; Namkung, et al., 2009b). GRK5 also facilitates D₂R internalization and may act in a phosphorylation-independent manner given that there is no evidence demonstrating direct phosphorylation of D₂Rs by GRK5 (Ito, et al., 1999). Furthermore, elimination of GRK-mediated phosphorylation residues within the third intracellular loop of D₂Rs has no influence on agonist-induced receptor internalization and functional coupling to G-proteins; instead, this mutation accelerates degradation of internalized D₂Rs and thus prevents receptor recycling to the membrane (Cho, et al., 2010; Namkung, et al., 2009a). Although these observations seem contradictory to the canonical role of GRK2 in regulation of D₂R phosphorylation and subsequent internalization, elimination of multiple GRK phosphorylation sites within D2Rs may change structural properties of D₂Rs and thus alter the sensitivity of the receptors to the effects of receptor agonists. The slow D₂R recycling in GRK phosphorylation-deficient mutants is believed to be mediated by dissociation of receptors with β-arrestin because receptor phosphorylation and β-arrestin are necessary for receptor resensitization once desensitized and internalized (Cho, et al., 2010). These data collectively suggest that GRK-dependent phosphorylation via GRK2/3 promotes the recycling of internalized D₂Rs to the membrane and thus responsible for resensitization.

Among GRK isoforms, GRK2, GRK3 and GRK6 are ubiquitously expressed in the brain whereas GRK5 is limited to the limbic and mesolimbic regions (Arriza, et al., 1992; Erdtmann-Vourliotis, et al., 2001). Given that GRK5 and GRK6 are not active in D₂R phosphorylation shown in cultured cells and the fact that GRK3 expression is notably weak in the brain, GRK2 has been considered as a major player in regulation of D₂R activity in vivo. Mice with selective GRK2 deletion in DA neurons exhibit increased basal locomotor activity and reduced sensitivity to acute and chronic cocaine-induced locomotor stimulation (Daigle, et al., 2014). Moreover, mice with GRK2 deficiency in D₂R-containing medium spiny neurons show increased basal locomotor activity and a marked reduction in sensitivity to cocaine stimulation. Interestingly, these mice show decreased basal DA release and hyperactive presynaptic D₂AR activity that can be blocked by sulpride, suggesting that the reduced cocaine effects are likely mediated through presynaptic D₂ARs instead of D₂PRs. The increased basal locomotor activity in both conditional GRK2 knockout mice is most likely mediated by dysregulated DAT function resulting from altered D₂AR and D₂PR activity in the absence of GRK2. In contrast to the regulation of D₂ARs primarily by GRK2, striatal D₂PR activity is critically modulated by GRK6 as GRK6 knockout mice show a greater coupling of striatal D₂PRs to G-proteins and enhanced sensitivity to the locomotor-stimulating effect by direct D₂R agonists, cocaine and AMPH (Gainetdinov, et al., 2003).

4.5.2. PKC

The phosphorylation of D₂Rs can occur in the absence of GRKs in HEK293 cells (Kim, et al., 2001), suggesting that other mechanisms are involved. Activation of PKC by treatment with phorbol 12-myristate 13-acetate (PMA) or 12-O-tetradecanoyl 4beta-phorbol 13 alpha-acetate produces a marked increase in D₂R phosphorylation and this effect is reversed by PKC inhibition (Morris, et al., 2007; Namkung & Sibley, 2004). Moreover, PKCβ plays a significant role in regulation of D₂R phosphorylation as overexpression of PKCβ increases basal and PMA-stimulated receptor phosphorylation (Namkung & Sibley, 2004). In contrast, other PKC isoforms including PKCδ, PKCμ and PKCζ have no roles in D₂R phosphorylation when overexpressed. Multiple sites within the second and third intracellular loops of D₂Rs have been identified as the substrate for PKC and deletion of these sites abolishes PMA-mediated D₂R phosphorylation (Morris, et al., 2007; Namkung & Sibley, 2004). In addition, PKC-mediated, site-specific phosphorylation of D₂Rs is involved in receptor internalization (Namkung & Sibley, 2004).

The D₂R function is modulated by PKC in heterologous expression systems. PKC phosphorylation of D₂Rs desensitizes the receptors by decreasing the potency of agonists to inhibit cAMP accumulation in HEK293T cells (Namkung & Sibley, 2004). PKC activation switches the preferential coupling of D₂Rs from inhibition of AC to facilitation of arachidonic acid release in CHO cells (Di Marzo, et al., 1993), suggesting that D₂R phosphorylation by PKC may alter the efficacy of D₂R coupling to different Gαi/o subunits and recruitment of different downstream effectors. Moreover, prolonged activation of D₂Rs leads to heterologous sensitization of AC responsiveness by decoupling of Gαi/o from AC and this sensitization is further enhanced when PKC is activated in HEK293 cells (Beazely & Watts, 2005). Lastly, PKC differentially modulates desensitization of D₂SRs and D₂LRs in the fibroblast and GH4 cells (Morris, et al., 2007). Activation of D₂Rs promotes Gβγ-mediated, phospholipase C-dependent calcium mobilization and this effect is inhibited by PKC when D₂SRs, but not D₂LRs, are activated. Because D₂SRs function primarily as autoreceptors and D₂LRs are postsynaptic (Radl, et al., 2018), the greater sensitivity of D₂SRs to PKC-mediated calcium immobilization likely renders the autoreceptors more susceptible to desensitization.

The D₂AR function is modulated by PKC, especially by PKCβ, both in vitro and in vivo. In rat striatal membranes treated with purified PKC, D₂R-mediated [³⁵S]GTPγS binding is abolished (Rogue, et al., 1990), although both D₂ARs and D₂PRs are activated. Regardless, these data suggest that PKC-dependent mechanisms can regulate the coupling between D₂Rs and G-proteins. Accordingly, downregulation of PKC activity by a prolonged treatment with a PKC activator attenuates D₂AR inhibition of the [³H]DA release from rat striatal slices (Iannazzo, et al., 1997). D₂AR inhibition and PKC activation can synergistically increase [³H]DA release from the striatum of rats and rabbits (Dong, et al., 1995). Moreover, PKCβ knockout increases D₂AR inhibition of endogenous DA release evoked by electrical stimulation and treatment with the potassium channel blocker 4-AP (Luderman, et al., 2015), indicating that the PKCβ isoform is involved in D₂AR-stimulated Kv1 channel activity to regulate evoked DA release. The enhanced D₂AR activity in PKCβ knockout mice is accompanied by an increase in surface receptor expression. In addition, ruboxistaurin, a selective inhibitor of PKCβ, attenuates cocaine-mediated DA overflow and locomotor activity, and this effect is abolished in the presence of the D₂R/D₃R antagonist raclopride (Zestos, et al., 2019). Lastly, because PKCβ is responsible for D₂R-stimulated DAT recycling shown in both cultured cells and mouse striatal synaptosomes (Chen, et al., 2013), these data suggest that PKC, especially PKCβ, modulates DA signaling in vivo by targeting both D₂AR and DAT activity.

The D₂ARs are also subject to regulation by a few other protein kinases. There are multiple consensus recognition sites for PKA within D₂Rs. PKA can directly phosphorylate purified D₂R proteins (Elazar & Fuchs, 1991), but has no effect on D₂R phosphorylation in HEK293 cells expressing D₂Rs (Namkung & Sibley, 2004). PKA phosphorylation of D₂Rs from striatal membranes and purified proteins reduces the efficacy of agonist binding by decoupling of D₂Rs from G-proteins (Elazar & Fuchs, 1991). Moreover, prolonged stimulation-induced D₂AR desensitization in rat SN is attenuated when AMP-activated protein kinase (AMPK) is activated (Yang, et al., 2019). Lastly, withdrawal from repeated ethanol exposure causes desensitization of D₂AR-dependent activation of GIRK channel conductance in mouse VTA DA neurons, which is mediated by IP3-induced calcium release from intracellular store and activation of CaMKII (Perra, et al., 2011). The structural biology modeling suggests that CaMKII interacts with D₂Rs and modulates DA binding affinity for D₂Rs. (Laoye, et al., 2016).

5. Conclusion & Perspectives

This review highlights many D₂AR-interacting partners including transporters, GPCRs, ion channels, signaling molecules and protein kinases (Fig.1), suggesting the functional divergence of the receptors. It is important to note that the D₂AR-interacting proteins involved in regulating midbrain DA neuron firing and axon terminal DA release are distinct. Although chronic drug exposure often alters D₂AR activity in both somatodendritic and axon terminal compartments of DA neurons, it appears that changes in somatodendritic D₂AR-mediated DA neuron firing are more enduring than the changes in axon terminal D₂AR-mediated DA release. The more persistent reduction of somatodendritic D₂AR activity induced by chronic exposure to drugs perhaps results from an irreversible disruption of the cellular D₂AR interaction network in the midbrain. Additionally, chronic exposure to drugs of abuse often downregulates surface receptor expression, which will likely diminish the efficacy and potency of receptor agonists/antagonists because of the reduced accessibility of the receptors by these drugs. This is particularly interesting because monodrug therapy has been the main approach for treatment of SUD and ineffective. Thus, concurrent targeting of D₂AR-interacting proteins at the membrane and intracellular levels along with direct receptor agonists or antagonists are a more promising approach to effectively normalize disrupted DA transmission. Future research on spatiotemporal analysis of the D₂AR interaction network will open up avenues to discover novel, dynamic D₂AR-interacting proteins, define new signal transduction pathways, and improve our understanding of the actions of drugs of abuse.

Abbreviations:

AAV

adeno-associated virus

A₂AR

adenosine A2 receptor

AC

adenylyl cyclase

4-AP

4-aminopyridine

AMPK

AMP-activated protein kinase

AMPH

amphetamine

AT₁Rs

angiotensin II type 1 receptors

BRET

bioluminescence resonance energy transfer

CaMKII

Ca2+/calmodulin-dependent protein kinase II

CPu

caudate putamen

cAMP

cyclic adenosine monophosphate

DA

dopamine

D₂R

dopamine D₂ receptors

D₂AR

dopamine D₂ autoreceptor

D₂PR

postsynaptic D₂ receptor

DAT

dopamine transporter

EPSCs

glutamate-mediated excitatory postsynaptic currents

ERK

extracellular signal-regulated kinase

GPCRs

G-protein coupled receptors

GIRK

G-protein-coupled inwardly rectifying potassium

GSK-3

glycogen synthase kinase-3

METH

methamphetamine

mAChR

muscarinic acetylcholinergic receptors

NAc

nucleus accumbens

nAChR

nicotinic acetylcholinergic receptors

NCS-1

neuronal calcium sensor 1

NTS₁R

neurotensin type 1 receptors

PFC

prefrontal cortex

PKA

protein kinase A

PKC

protein kinase C

PMA

phorbol 12-myristate 13-acetate

RGS

regulator of G-protein signaling

RGS9–2

the long splice form of RGS9

SN

substantia nigra

SUD

substance use disorder

TAAR1

trace amine-associated receptor 1

TEA

tetraethylammonium

TH

tyrosine hydroxylase

VTA

ventral tegmental area