Noradrenergic Circuits and Signaling in Substance Use Disorders

Abstract

The central noradrenergic system innervates almost all regions of the brain and, as such, is well positioned to modulate many neural circuits implicated in behaviors and physiology underlying substance use disorders. Ample pharmacological evidence demonstrates that α₁, α₂, and β adrenergic receptors may serve as therapeutic targets to reduce drug –seeking behavior and drug withdrawal symptoms. Further, norepinephrine is a key modulator of the stress response, and stress has been heavily implicated in reinstatement of drug taking. In this review, we discuss recent advances in our understanding of noradrenergic circuitry and noradrenergic receptor signaling in the context of opioid, alcohol, and psychostimulant use disorders.

INTRODUCTION

Norepinephrine (NE) is one of the most abundant neurotransmitters in the brain and periphery, and plays an important role in a variety of behavioral processes including, arousal, attention, appetite, and stress (Aston-Jones and Cohen, 2005; Berridge and Waterhouse, 2003; Rinaman, 2011; Sara and Bouret, 2012). The central noradrenergic system consists of six nuclei: A1/Rostroventrolateral medulla (RVLM), A2/Nucleus of the tractus solitarius (NTS), A5, A6/Locus coeruleus (LC), A7, and subcoeruleus. The cell bodies of all noradrenergic nuclei that innervate the brain are located in the medulla or pons. The A6/LC contains the largest number of central noradrenergic neurons and they project to almost all areas of the forebrain, including but not limited to the cortex, hippocampus, and central nucleus of the amygdala (CeA), via the dorsal noradrenergic bundle (DNB). The A1/RVLM and A2/NTS nuclei provide noradrenergic innervation to the paraventricular nucleus of the insular cortex, hypothalamus, bed nucleus of the stria terminalis (BNST), CeA, and basolateral amygdala (BLA) via the ventral noradrenergic bundle (VNB) (Moore and Bloom, 1979; Robertson et al., 2013).

NE modulates neurons via three main families of G-protein coupled receptors: the α₁, α₂, and β adrenergic receptors (ARs). The α₁ family consists of three different subtypes, α_1a, α_1b, and α_1D, all of which signal canonically through Gα_q proteins and recruit phospholipase C and Ca²⁺ signaling. The α₂ family consists of 3 subtypes, α_2a, α_2b, and α_2c, which signal through the Gα_i pathway and have the ability to suppress cAMP accumulation and/or alter the conductance of K⁺ and Ca²⁺ channels. The β family also consists of three subtypes, named β₁, β₂, and β₃, which all signal canonically through Gα_s proteins to enhance cAMP signaling promoting activation of protein kinase A. All three AR subtypes are found on postsynaptic neurons, where they modulate neuronal activity. However, α₂ ARs are also found on presynaptic terminals where they suppress the release of NE in an autocrine manner. α₂ ARs are also found presynaptically on glutamatergic and GABAergic terminals, where they suppress glutamate or GABA release respectively as heteroreceptors (Shields et al., 2009).

In recent history, the involvement of NE in substance use disorders (SUDs) has been overlooked in favor of dopamine (DA). However, several decades of research have clearly demonstrated that NE plays a critical role in some components of addiction-related behavior (Weinshenker and Schroeder, 2007). NE plays a significant role in stress-induced reinstatement of drug-seeking behaviors as well as withdrawal processes (Mantsch et al., 2016). Pharmacological data also support the importance of NE in substance use disorder (SUD), as drugs targeting α and/or β ARs modulate drug-seeking behaviors (Weinshenker and Schroeder, 2007). Indeed, recent clinical trials support the efficacy of α₁ AR antagonists and α₂ AR agonists in SUD (Kahn et al., 1997; Pergolizzi et al., 2019; Simpson et al., 2015; Simpson et al., 2018). Improving our understanding of how NE modulates neural circuitry during SUD may lead to the identification of new circuit interactions and novel therapeutics. This review will focus on recent advances in our understanding of how noradrenergic circuits modulate use of opioids, alcohol, and psychostimulants.

NOREPINEPHRINE, STRESS, AND SUBSTANCE USE DISORDER

SUD has been linked to brain systems that mediate both reward and stress. Logically this makes sense as both external and internal stress drives natural reward seeking (i.e. thirst drives the seeking of water.) Activation of these brain stress systems, however, is thought to be key to the negative emotional states (dysphoria, anxiety, and irritability) associated with drug misuse, and especially drug dependence, thus driving drug seeking through negative reinforcement learning (Koob and Le Moal, 1997; Koob and Le Moal, 2008). Life stress is also a well-known risk factor for the development of SUD and greatly increases the risk of relapse, and stress and anxiety disorders are highly comorbid with SUDs (Sinha, 2008).

NE is a key component of brain stress systems in large part due to NE’s direct and indirect influence over the hypothalamic-pituitary-adrenal axis (HPA)(Herman et al., 2016). In brief, activation of the paraventricular nucleus of the hypothalamus causes synthesis and release of corticotropin releasing factor (CRF), which in turn stimulates adrenocorticotropic hormone (ACTH) production by the anterior pituitary gland. ACTH then enters the bloodstream to mediate production of glucocorticoids by the adrenal cortex. Glucocorticoids subsequently mediate the cellular stress response. Peripherally, NE is a key component in the sympathetic nervous system, which also mediates the stress response. In response to a stressful stimulus, preganglionic neurons originating in the interomediolateral gray matter of the spinal cord release acetylcholine, which drives peripheral NE release by either directly activating postganglionic NE neurons or stimulates epinephrine and norepinephrine production and release into the bloodstream by the adrenal medulla (Goldstein, 1987). Preganglionic sympathetic neurons also receive direct inputs from brain areas such as the paraventricular nucleus of the hypothalamus and insular cortex, as well as the noradrenergic A2/NTS and A1/RVLM areas (Fisher et al., 2009), and are indirectly influenced by other fear- and stress-related brain regions, such as the amygdala and BNST (Buijs and Van Eden, 2000; Crestani et al., 2013; Lebow and Chen, 2016). In an acute setting, peripheral NE signaling via the sympathetic nervous system, in concert with central NE signaling upon the HPA axis, instigate the “flight or fight” response. However, when chronically activated, central and peripheral signaling can lead to the development of neuropsychiatric disorders and can contribute to SUD (Sinha, 2008).

Ample epidemiological data links stress disorders, including post-traumatic stress disorder (PTSD), generalized anxiety disorders, and early childhood and adolescent exposure to traumatic events or abuse, to an increased risk for developing SUD (Sinha, 2008). This is likely explained by positive feedback interactions between activation of central and peripheral stress systems and the midbrain dopaminergic system, which is a brain region that directly mediates the reinforcing properties of many drugs of abuse (Di Chiara and Imperato, 1988; Volkow et al., 2007). Increased CRF and cortisol levels due to activation of the HPA axis during stress can directly modulate dopaminergic signaling and release in the ventral tegmental (VTA) and nucleus accumbens (NAcc) (Kalivas and Duffy, 1995; Piazza and Le Moal, 1996; Saal et al., 2003; Thierry et al., 1976). Further, both drug withdrawal and exposure to drug cues activates the HPA-axis. For example, in recently abstinent individuals with cocaine use disorder, exposure to a stress or drug cue results in a long-lasting activation of the HPA-axis, with increases in plasma cortisol, ACTH, norepinephrine, and epinephrine levels (Sinha et al., 2003). These studies demonstrate a pathway by which exposure to stress or drug withdrawal can activate central and peripheral stress systems to instigate or reinstate drug consumption through modulation of midbrain dopaminergic systems. In future sections, we will discuss studies investigating the role of the NE system in stress- or cue-induced reinstatement of drug consumption along with other roles of the noradrenergic system in other phases of SUD.

NOREPINEPHRINE IN OPIOID USE DISORDER

Significant pharmacological evidence links noradrenergic signaling and circuitry with opioid use disorder (OUD). The α₁ AR antagonist prazosin reduces heroin self-administration (SA) in rats and blocks the somatic withdrawal symptoms of morphine in mice (Greenwell et al., 2009; Solecki et al., 2019b). Noradrenergic drugs are also used clinically to treat the symptoms of opioid withdrawal. The α₂ agonists clonidine and lofexidine are effective at reducing opioid withdrawal symptoms (Gish et al., 2010; Kahn et al., 1997; Pergolizzi et al., 2019). Taken together this suggests that reducing postsynaptic noradrenergic signaling by blocking α₁ ARs or inhibiting release of norepinephrine by activating presynaptic α₂ ARs may be important for alleviating opioid withdrawal. Additional studies have helped to better characterize the circuitry that mediates the effects of NE on opioid consumption and withdrawal.

Extended amygdala

The BNST, along with the CeA and the shell of the NAcc constitute the extended amygdala (Alheid, 2003). The BNST functions as an important integrator of stress and reward information (Jennings et al., 2013; Shackman and Fox, 2016). In the context of SUD, the BNST been implicated in withdrawal and is involved in stress-induced reinstatement of drug seeking behavior (Koob, 2009; Vranjkovic et al., 2017).

A combination of behavioral and physiological studies has implicated BNST noradrenergic signaling in many aspects of opioid use. Both mice lacking dopamine β-hydroxylase (DBH), the enzyme responsible for norepinephrine synthesis, and α1_b AR knockout mice do not develop conditioned place preference (CPP) in response to morphine (Drouin et al., 2002; Olson et al., 2006). Restoration of DBH expression in the A2/NTS, but not the A6/LC, restores CPP in response to morphine (Olson et al., 2006). Due to the fact that the A2/NTS sends an extremely strong projection to the extended amygdala, and especially the BNST, these data suggest that NE signaling in the BNST may be necessary for CPP. This is highly probable because the BNST is strongly implicated in the seeking of cocaine, and ethanol, and in stress induced reinstatement to morphine seeking (Pina et al., 2015; Sartor and Aston-Jones, 2012; Wang et al., 2006). However, other A2/NTS targets like the basolateral amygdala and hypothalamus could also play a role in mediating this behavior.

Substantial evidence has linked NE signaling to stress-induced opioid reinstatement. Lesioning of the VNB, which carries most of the noradrenergic afferents from the A2/NTS and A1/RVLM, or infusion of the α₂ AR agonist clonidine into the BNST blocks stress-induced reinstatement of morphine CPP (Wang et al., 2001). However, microinjection of clonidine into the LC or lesion of the DNB, which carries most of the noradrenergic afferents from the LC, did not impact stress-induced maintenance of CPP in response to foot shock (Wang et al., 2001). In congruence, microinjection of clonidine into the LC did not impact reinstatement of heroin seeking in rats (Shaham et al., 2000). Taken together this suggests that the BNST and NE projections originating from the A2/NTS nucleus are necessary for stress induced reinstatement of opioid CPP and SA, but NE projections originating in the A6/LC are not.

NE released in the BNST is also a critical mediator of the somatic and affective components of opioid withdrawal (Delfs et al., 2000). Abstinence from morphine enhances activation of the ventrolateral BNST and its afferents from the A2/NTS (Aston-Jones and Harris, 2004; Hamlin et al., 2004; Harris and Aston-Jones, 2007). In fact, NE release in the BNST co-occurs during specific withdrawal behaviors in rats including, teeth chattering and swallowing movements, and withdrawal behaviors were exacerbated after co-administration of naloxone and the α₂ AR antagonist idazoxan following a single dose of morphine (Fox et al., 2017). This suggests that the amount of NE released in the BNST may correlate with the severity of withdrawal and that noradrenergic signaling in the BNST may contribute to the negative reinforcement and aversive learning associated with opioid withdrawal. Further, opioid withdrawal can induce plasticity in adrenergic circuitry itself. Chronic morphine withdrawal induces adrenergic plasticity in the BNST whereby α₂ AR become less effective at inhibiting NE release and norepinephrine transporter (NET) reuptake activity is decreased in rats (McElligott et al., 2013). We have observed similar reduced efficacy in α₂ ARs, and enhanced BNST NE release in chronically stressed mice (restraint, (Schmidt et al., 2019), demonstrating that across species opioid withdrawal and stress may result in common pathological mechanisms. In combination these effects may enhance BNST NE levels and potentially exacerbate symptoms of opioid withdrawal. As OUD is a chronically relapsing disorder, and α₂ AR antagonism is one of the main in-patient therapeutic treatments for withdrawal, this type of receptor desensitization could limit the therapeutic efficacy of these treatments. Individuals are now experiencing naloxone precipitated withdrawal to treat overdose at increasing rates with the availability of over-the-counter Narcan administration, Moreover the rates of patients receiving multiple doses of naloxone from emergency medical service (EMS) providers is on the rise (up 26% from 2012–2015, (Faul et al., 2017). This trend has only been exacerbated during the COVID-19 global pandemic. For example, EMS naloxone administration increased 57.8% and repeat administration increased 84.8% as compared to the pre-pandemic period in Guilford County, NC (Khoury et al., 2021). While, we emphasize that naloxone is life-saving treatment and should be readily available, we also wish to emphasize that the field needs to have a better mechanistic understanding of how both spontaneous, and rapid, precipitated opioid withdrawal alter neural circuits and plasticity.

Locus Coeruleus

The A6/LC provides noradrenergic input to most forebrain structures, and, as such, changes to A6/LC function are well positioned to modulate many different brain regions. Systemic or intracoeruleur administration of morphine directly inhibits LC neurons (Aghajanian, 1978; Bird and Kuhar, 1977; Korf et al., 1974; Valentino and Wehby, 1988). However, chronic opioid consumption induces tolerance which prevents morphine-induced LC neuron inhibition due to compensatory changes in cAMP/PKA signaling pathways (Ivanov and Aston-Jones, 2001; Kogan et al., 1992). Interestingly, chronic morphine exposure causes internalization of μ opioid receptor (OR) from the membrane of LC neurons in male rats, but not in females (Enman et al., 2019). This suggests there may be critical sex differences in LC plasticity in response to opioid exposure which may affect tolerance and/or withdrawal.

Like the BNST, most research into the role of the LC has focused on the withdrawal phase of opioid use disorder. The LC is activated above baseline levels by opioid withdrawal and opioid withdrawal alters excitability of LC neurons, which is likely a result of long-lasting changes in cAMP/PKA/CREB signaling pathways (Fig 1) (Han et al., 2006; Hayward et al., 1990; Mazei-Robison and Nestler, 2012; Rasmussen et al., 1990). Down regulation of CREB in the LC/A6 using viral approaches attenuated some withdrawal behaviors associated with morphine withdrawal including, ptosis, wet dog shakes, and lacrimation, while upregulation of CREB enhanced those same withdrawal behaviors (Han et al., 2006). Further, a study using a LC slice culture model demonstrated that LC activity normalizes back to baseline levels following chronic morphine exposure, but LC activity is significantly enhanced following bath application of naloxone or washout of morphine. These effects on LC neuron activity are directly mediated by μ OR activation, suggesting an adaptive change in the valence of μ OR signaling (Cao et al., 2010). Overall, these findings conflict with early studies that found that lesion of the DNB, which contains NE projections from the LC, or LC proper did not attenuate somatic withdrawal behavior nor conditioned place aversion (CPA) after opioid withdrawal, suggesting that LC projections do not mediate all withdrawal-associated symptoms (Caillé et al., 1999; Delfs et al., 2000). However, these earlier studies may need to be reconsidered as more recent anatomical studies using modern genetic approaches have demonstrated that some LC neurons innervate the BNST and that the LC and A2/NTS innervate each other (Robertson et al., 2013). It is important to note, however, that there also may be species differences between rat and mouse that could account for these discrepancies.

Fig 1.

Plasticity in LC neuron following opioid dependence and tolerance. In the non-opioid dependent LC, activation of μ ORs normally decreases cAMP/PKA signaling and depresses LC excitability. The LC also receives glutamatergic input from the paragigantocellularis nucleus, lateral habenula, and prefrontal cortex and GABA/CRF input from the CeA. Under normal conditions, glutamatergic and CRF inputs enhance LC neuron activity. In the opioid dependent LC, the valence of μ OR activity changes, whereby activation of μ ORs enhances cAMP/PKA and CREB signaling. Glutamatergic inputs to the LC are also enhanced and drive increased LC excitability. There is also an increase in CRFR1 receptors which may potentiate the activity on CRF on LC excitability.

The altered plasticity observed in the LC after chronic morphine exposure and withdrawal, may be a consequence of changes in inputs to the LC. There is an increase in excitatory input to the LC in chronically morphine-treated animals, as demonstrated by an increased frequency of spontaneous EPSCs (Torrecilla et al., 2008). However, the specific glutamatergic inputs to the LC that are enhanced is unknown, as the LC receives glutamatergic input from the paragigantocellularis nucleus, lateral habenula, and prefrontal cortex (Aston-Jones et al., 1986; Herkenham and Nauta, 1979; Jodo and Aston-Jones, 1997). Opioid use and withdrawal also modulate peptidergic inputs to the LC, including CRF and orexin. Stimulating the LC via orexin receptors promotes behaviors associated with opioid withdrawal in both morphine dependent and control rats, while blocking orexin receptor-1 (HCRTR1) prevents withdrawal behaviors (Davoudi et al., 2016; Ghaemi-Jandabi et al., 2017). Chronic morphine use sensitizes the LC to the effects of CRF and stress by recruiting CRF receptor 1 to the plasma membrane of LC neurons (Enman et al., 2019; Xu et al., 2004). Blocking CRF receptor 1 prevents both morphine CPP and withdrawal behaviors and also reduces NE turnover in the cerebral cortex (Funada et al., 2001; Lasheras et al., 2015). This demonstrates a mechanism by which stress can enhance LC activity and perhaps potentiate opioid-seeking activity and withdrawal.

Taken together, these preclinical rodent studies demonstrate a few anatomical locations where systemic administration of noradrenergic compounds may work to reduce the symptoms of opioid withdrawal in humans. Both blockade of α₁ ARs or activation of α₂ ARs in the BNST may be an important site of action to diminish the symptoms of opioid withdrawal. However, the observed desensitization of α₂ ARs in the BNST following repeated bouts of morphine withdrawal in rats may suggest that α₂ AR agonists potentially work in other brain regions as well (McElligott et al., 2013). Further, the association between opioid withdrawal and enhanced LC activation suggests that reducing postsynaptic noradrenergic signaling in brain regions that receive LC input, including CeA or mPFC, may play an important role in the therapeutic action of α₁ ARs antagonists or α₂ ARs agonists.

NOREPINEPHRINE AND ALCOHOL USE DISORDER

Noradrenergic signaling is also a critical mediator of the effects of ethanol. In clinical populations, reduced serum DBH activity is associated with alcohol use disorder (AUD) (Köhnke et al., 2002). Further, the DBH*444G/A allele, which causes a significant reduction in serum DBH activity, is associated with AUD but not with withdrawal symptom severity (Köhnke et al., 2006).

Alcohol consumption activates noradrenergic circuitry. Intraperitoneal injection and binge consumption (>80 mg/dL) of ethanol induces expression of the early immediate protein cFos in noradrenergic neurons in the LC and A2/NTS (Burnham and Thiele, 2017; Thiele et al., 2000). Acute ethanol exposure enhances NE release in the medial prefrontal cortex, basal forebrain, and BNST (Jadzic et al., 2021; Jaime et al., 2020). However, the increase in NE release in the medial prefrontal cortex (mPFC) was attenuated after one week of ethanol SA, which suggests adaptive changes in LC neuron activity, and/or terminal release mechanisms, following long-term ethanol use (Jaime et al., 2020). NE levels are elevated in both the CNS and periphery during alcohol withdrawal in mice (Kovács et al., 2002). In alcohol dependent patients undergoing alcohol withdrawal, NE levels are elevated in the cerebrospinal fluid and in plasma (Fitzgerald, 2013; Hawley et al., 1981; Patkar et al., 2003).

NE also modulates ethanol consumption. DBH KO mice have reduced ethanol consumption in a SA paradigm (Weinshenker et al., 2000). However, it is unclear if this reduction in ethanol consumption is due to a decrease in the rewarding properties of ethanol per se, or an increase in the aversive and/or interoceptive effects of ethanol, as DBH KO mice are more sensitive to ethanol-induced sedation and hypothermia (Weinshenker et al., 2000). It is worth noting that one of the only FDA approved treatments for AUD, the alcohol dehydrogenase inhibitor disulfiram, also is a potent inhibitor of DBH (for detailed discussion see: (Weinshenker and Schroeder, 2007) . Further, just as has been demonstrated with other misused substances, treatment with an α₂ AR agonist decreases alcohol SA while α₂ AR antagonist treatment increases alcohol consumption (Lê et al., 2005). However, these results should be interpreted with some caution because the α₂ antagonist used in the study, yohimbine, has significant off-target effects (Davis et al., 2008). Treatment with α₂ agonists also decreases stress-induced alcohol seeking behaviors due to footstock stress (Lê et al., 2005). A more recent study found that the α2 AR agonist clonidine reduced alcohol consumption in a genetically alcohol-preferring rats (Rasmussen et al., 2014). Clinical data supporting the use of α2 AR agonists is mixed. In a small clinical trial, clonidine did not significantly improve withdrawal symptoms relative to placebo (Adinoff, 1994). A clinical trial for the α2 AR agonist guanfacine for AUD in women is currently ongoing (NCT03137082), which is an interesting avenue to explore given that there are known sex differences in efficacy for drugs targeting the noradrenergic system (Fox et al., 2014; Fox and Sinha, 2009). Taken together these data suggest that reducing NE transmission with an α2 AR agonist may be an effective target for reducing alcohol consumption.

Blocking α₁ ARs has also been suggested as a therapeutic target in multiple clinical and preclinical studies. In fact, the α₁ AR antagonist prazosin has demonstrated efficacy in reducing the number of drinks and number of drinking days in a double-blind randomized clinical trial in alcohol use disorder (Simpson et al., 2018). A recent clinical trial found that prazosin was more efficacious in patients suffering from more severe withdrawal symptoms. These patients experienced a significant reduction in heavy drinking days and alcohol cravings, while those with low or no withdrawal symptoms did not experience a benefit from prazosin (Sinha et al., 2021). Prazosin was also tested in AUD patients in a trial that used imagery exposure to stress and alcohol cues. Prazosin significantly decreased alcohol craving induced by both stress and alcohol cues, which suggests that prazosin may help to mitigate stress associated with early alcohol withdrawal (Fox et al., 2012; Milivojevic et al., 2020). A further clinical trial assessed the efficacy of prazosin in patients with comorbid AUD and PTSD. While prazosin reduced drinking outcomes in this study, it did not affect PTSD symptoms (where there have been promising but conflicting studies, see: (Raskind et al., 2018, 2013, 2003) and resulted in significant hypotensive side effects (Simpson et al., 2015). An additional small clinical study found that the α₁ AR antagonist, doxazosin, was more effective in reducing alcohol consumption in patients with high starting systolic blood pressure, which suggests that individual differences in sympathetic tone may account for the success or failure of α₁ AR antagonists in the clinic (Haass-Koffler et al., 2017). In preclinical studies, blocking α₁ ARs reduced alcohol self-administration in dependent animals and reduced alcohol consumption in a chronic intermittent access paradigm (Skelly and Weiner, 2014; Walker et al., 2008). Overall, these clinical and preclinical studies provide strong evidence that blocking α₁ ARs may be effective in AUD. However, the potential positive effects on alcohol consumption need to be balanced against potential negative hypotensive effects that may make these compounds undesirable for clinical use.

There is significantly less literature regarding the involvement of β ARs in AUD. Blocking central, but not peripheral, β ARs also reduced alcohol consumption in alcohol-dependent rats (Gilpin and Koob, 2010). Clinically, one early study found that the β AR antagonist propranolol was equally effective as diazepam at alleviating moderate withdrawal symptoms in AUD (Bailly et al., 1992). Two larger double-blind placebo control trial found that the β AR antagonist, atenolol, reduced withdrawal symptoms, alcohol cravings, and treatment failure in patients with mild withdrawal (Gottlieb et al., 1994; Horwitz et al., 1989). In the below sections, we will discuss circuit mechanisms that may mediate the effects of these noradrenergic drugs.

Extended and basolateral amygdala

Noradrenergic signaling in the extended amygdala has been specifically implicated in alcohol consumption. Acute ethanol administration increases NE output in the BNST (Jadzic et al., 2021). Microinjection of pindolol, a combined β AR antagonist and 5-HT_1A/B agonist, into the BLA reduced long term ethanol consumption (Patkar et al., 2019; Patkar et al., 2017). Further, injection of the β AR antagonist propranolol into the BLA prevented reinstatement of alcohol seeking behaviors in a self-administration paradigm in rats (Chesworth and Corbit, 2018). This suggests that NE signaling in the BLA is a critical modulator of ethanol consumption, although the precise β AR subunit responsible for these effects is unknown.

Orbitofrontal cortex

Cortical regions, including the orbitofrontal cortex (OFC), have also been implicated in alcohol use disorder. The OFC mediates some components of choice behavior, reversal learning, and anticipation of expected outcomes in response to appetitive and aversive stimuli (Bissonette et al., 2008; Roesch and Olson, 2007; Schoenbaum et al., 1998). OFC dysfunction has also been implicated in neuropsychiatric diseases, including alcohol use disorder (Fortier et al., 2008; Verdejo-García et al., 2006). A recent study demonstrated that in non-alcohol exposed animals, monoamine receptors, including D2 DA receptors, α₂ AR, and 5-HT_1A receptors reduce OFC neuron excitability, mainly through activation of G-protein inwardly rectifying potassium (GIRK) channels. However, in chronic intermittent ethanol-treated mice, OFC neurons become more excitable and do not respond to D2 DA receptor, α₂ AR, or 5-HT_1A receptor activation (Nimitvilai et al., 2017). This demonstrates a mechanism by which chronic alcohol consumption can cause widespread disruption of cortical function.

Locus Coeruleus

NE signaling in the LC has also been implicated in alcohol use disorder. A recent study found that noradrenergic inputs from the LC to the VTA are required for the reinforcing effects of ethanol. Administration of α₁ AR agonists into the posterior VTA enhanced alcohol self-administration, while α₂ AR agonists decreased alcohol self-administration (Shelkar et al., 2017). Further, silencing of LC inputs to the posterior VTA with lidocaine blocked alcohol self-administration (Shelkar et al., 2017).

CRF and noradrenergic signaling in the LC are critical components for stress and are required for stress induced drug reinstatement (Smith and Aston-Jones, 2008). LC noradrenergic neurons are among the first to be recruited during a stressful event and increased NE release in the CeA enhances CRF-expressing GABAergic projection neuron activity in the CeA (Borodovitsyna et al., 2020; Campese et al., 2017; Finnell et al., 2019). In turn, CRF-expressing GABAergic neurons in the CeA project back to the LC and enhanced CRF input to LC neurons enhances LC neuron activity and anxiety like phenotypes (Koob, 1999; McCall et al., 2015; Reyes et al., 2011; Snyder et al., 2012). This interaction between LC neurons and CRF-expressing neurons in the CeA is disrupted by chronic ethanol consumption and creates a feed- forward loop that is activated following alcohol withdrawal and stress (Funk et al., 2006; Gilpin et al., 2015; Koob, 2015) (Fig 2). A recent study found that CRF receptor agonist administration in the CeA decreases NE release in the CeA in slice, which provides a local circuit mechanism of NE release inhibition. However, ethanol itself had no effect on NE release in the CeA (Hedges et al., 2020). In contrast, a previous study using in vivo microdialysis found that local infusion of CRF receptor agonist enhanced NE transmission in the CeA, an effect that would likely exacerbate stress response in vivo (Su et al., 2015). The differences between these two studies likely reflect the recruitment of additional circuits, perhaps connections from the PFC to the CeA, which are also regulated by LC noradrenergic inputs (Borodovitsyna et al., 2020). These long-loop connections would not be intact in the ex vivo slice preparation.

Figure 2.

LC and CeA circuitry in stress and alcohol use disorder. The LC sends noradrenergic projections to the CeA, while the CeA sends GABAergic and CRF-positive projections to the LC. During stress or alcohol withdrawal, LC activity increases and sends enhanced noradrenergic input to the CeA. This in turn increases GABA/CRF neuron activity in the CeA resulting in a feedforward-loop causing enhanced LC and CeA activity in stress and alcohol withdrawal. This enhanced activity contributes to an increase in both anxiety-like behaviors and alcohol consumption in rodents.

Non-neuronal cells

Non-neuronal cell types, including astrocytes and microglia, have also been implicated in alcohol use disorder (Adermark and Bowers, 2016). A recent study used 2-photon microscopy to investigate how ethanol and NE modulate vigilance-included changes in astrocytic Ca²⁺ signaling. NE triggers Ca²⁺ transients in astrocytes through activation of α₁ ARs. However, acute ethanol exposure causes a decrease in locomotor-induced increases in Ca²⁺ transients in astrocytes in both cerebellum and V1 cortex. The decrease in astrocyte Ca²⁺ is not due to direct effects of ethanol and astrocytes, but rather a decrease in LC activity by ethanol and a corresponding decrease in global NE release (Ye et al., 2020). This demonstrates how alcohol-induced changes in NE release from LC neurons can have physiological ramifications throughout the brain on both neurons and glia.

Taken together, these preclinical rodent studies demonstrate a few anatomical locations where systemic administration of noradrenergic compounds may work to treat AUD. Blockade of β ARs in the BLA may help to reduce alcohol consumption, although curiously β AR antagonists are the least well studied target from a clinical perspective. Further, the clear association between enhanced LC activation and stress-induced reinstatement of alcohol consumption suggests that α₁ AR antagonists or α₂ AR agonists may reduce alcohol consumption by blocking noradrenergic signaling in brain regions that receive LC projections such as the CeA or mPFC. Targeting these receptors in the CeA may help break the feedforward loop involving CRF-signaling between the CeA and LC to help relieve some of the aversive effects of alcohol withdrawal as well as stress-induced reinstatement. Further, blocking LC to VTA noradrenergic inputs may be effective in reducing the rewarding effects of alcohol by indirectly modulating NAcc DA signaling.

NOREPINEPHRINE AND PSYCHOSTIMULANT USE DISORDER

Psychostimulants, including amphetamine, methamphetamine, and cocaine, inhibit the norepinephrine transporter (NET), dopamine transporter (DAT), and serotonin transporter to modulate monoamine uptake in the brain. Early self-administration studies found that NE was not a mediator of the maintenance phase of psychostimulant self-administration. Rather, self-administration of psychostimulants depended on DA signaling. DA receptor antagonists reduced psychostimulant self-administration, but NE antagonists generally had no effect (Yokel and Wise, 1975; Yokel and Wise, 1976). Lesion of the VNB also had no effect on the maintenance phase of cocaine self-administration (Roberts et al., 1977). Further, selective NET inhibitors are generally not self-administered, while DAT inhibitors are self-administered (Skjoldager et al., 1993; Wee et al., 2006; Woolverton, 1987). Together, these data suggest that NE is not involved in the maintenance of psychostimulant self-administration.

However, NE signaling has been implicated in other models of psychostimulant use disorder. NET knockout mice show reduced CPP to amphetamine, likely a result of enhanced extracellular NE levels (Ventura et al., 2003). NE may also play a role in regulating locomotor and behavioral sensitization to psychostimulants. DBH knockout mice appear to be more sensitive to the locomotor effects of cocaine and amphetamine and are more sensitive to stereotypy following psychostimulant administration (Schank et al., 2006; Weinshenker et al., 2002). NET knockout mice also demonstrate enhanced locomotor sensitization to cocaine and amphetamine (Xu et al., 2000). Some of these changes may be confounded by alterations in DA signaling in the NAc caused by DBH knockout. However, this strongly implicates that the NE system is regulating the mesolimbic DA system in profound ways.

Knockout studies have implicated the α₁ AR as a key postsynaptic mediator of behavioral sensitization to psychostimulants. Systemic or direct PFC administration of α₁ AR antagonists reduces cocaine or amphetamine locomotor sensitization (Drouin et al., 2002; Weinshenker et al., 2002; Wellman et al., 2002). Global knockout of α_1b ARs completely blocks psychostimulant induced locomotor activity (Auclair et al., 2004; Drouin et al., 2002; Salomon et al., 2006). In congruence, activating postsynaptic α₁ or α₂ ARs enhanced the locomotor effects of psychostimulants (Villégier et al., 2003). Taken together, these data suggest that enhanced NE signaling is critical for the locomotor effects of psychostimulants, although the effects on DA release caused by the manipulations in these experiments should be considered.

While NE does not appear to be involved in the maintenance phase of psychostimulant self-administration, NE is important for reinstatement of psychostimulant self-administration. Researchers have selectively bred a high responding rat line that exhibits enhanced levels of prolonged drug seeking behavior following long-term self-administration of cocaine, increased propensity for stress-induced reinstatement, and higher risk of relapse following abstinence (Flagel et al., 2016). These high responding rats have increased basal and cocaine-induced efflux of both DA and NE in the NAcc relative to low responding rats (Mabrouk et al., 2018). Intracerebroventricular injection of NE reinstates cocaine-seeking behavior (Brown et al., 2011), while inhibition of DBH with nepicastat reduced cocaine reinstatement following both stress and drug-associated cues (Schroeder et al., 2013). Further, systemic administration of α₂ AR agonists reduced stress induced reinstatement of cocaine seeking following foot shock (Erb et al., 2000). A recent clinical trial found that the α2 agonist guanfacine significantly reduced cocaine cravings and stress in female patients that were cocaine-dependent, while no effects were seen in cocaine-dependent male patients (Fox et al., 2014). The α₁ AR antagonist, doxazosin, also decreased cocaine use in a small clinical trial (Shorter et al., 2013). Doxazosin appears to be more effective in patients with lower noradrenergic tone caused by the (C1021T) variant in the DBH gene (Zhang et al., 2019). However, doxazosin is less effective in individuals with the (T1848A) variant in the ADRA1D gene, which encodes the α_1D AR (Shorter et al., 2020). These data demonstrate that enhanced NE release is an important mediator of stress-induced reinstatement. Further, the clinical trials for noradrenergic drugs demonstrate the importance of considering sex and genetic variants that may affect noradrenergic signaling/tone when evaluating the efficacy of these compounds.

Extended amygdala

Ample evidence has demonstrated that the extended amygdala is critical for the rewarding properties of psychostimulants (Li et al., 2015; Venniro et al., 2020). Both α and β ARs in the extended amygdala have been implicated in stress-induced reinstatement of cocaine seeking behaviors. Microinjection of β AR antagonists into either the BNST or the CeA blocks cocaine reinstatement following foot shock or swim stress but does not block drug-induced reinstatement (Leri et al., 2002). The β2 AR selective antagonist ICI-118,551 also reduced CRF mRNA expression in the BNST following swim stress (McReynolds et al., 2014). Blockade of β2 ARs in the ventral BNST suppressed the activity of a ventral BNST to VTA circuit, which in turn suppressed both CRF receptor activation in the VTA and stress induced reinstatement of cocaine consumption (Vranjkovic et al., 2014) (Fig 3). However, β AR antagonists did not influence cocaine-induced reinstatement of cocaine seeking, suggesting a specific role of NE in response to stress-induced – but not drug-induced reinstatement (Leri et al., 2002).

Figure 3.

vBNST to VTA circuitry in cocaine use disorder. In CUD, noradrenergic innervation of the vBNST is increased, which leads to increased activity of CRF-expressing GABAergic BNST neurons via activation of β₂ ARs. These BNST neurons project to the VTA and release elevated amounts of CRF. CRFR1 is activated which stimulates dopaminergic VTA neurons and results in increased DA release in the NAcc and mPFC which contributes to stress-induced reinstatement. This pathway is inhibited following β₂ AR antagonist treatment. CRF innervation from the vBNST to VTA is inhibited which decreases dopaminergic VTA neuron activity. This ultimately results in reduced DA release in the NAcc and mPFC and opposes stress-induced reinstatement of cocaine seeking.

α2 ARs have also been investigated in the context of cocaine use disorder in animal models, which is especially important given the interest in α2 AR agonists in the clinic. Postsynaptic excitatory α2_a ARs are located in the BNST. A recent study found that global deletion of α2_a ARs from the BNST blocked stress-induced cocaine reinstatement, but rescue of α2_a AR signaling by mimicking its effect with G_αi-DREADD expression specifically in the BNST restored stress-induced cocaine reinstatement (Perez et al., 2020). These data complicate the use of α2 AR agonists like guanfacine, because while guanfacine may inhibit stress-induced reinstatement by activating presynaptic α2 autoreceptors, it may also promote stress induced reinstatement by activation α2_a ARs heteroreceptors in the BNST. However, the authors found that low dose guanfacine (0.15 mg/kg) did not enhance BNST activity as assessed by cfos staining but did suppress stress-induced cocaine reinstatement (Perez et al., 2020).

Chronic cocaine administration also induced plasticity in the BNST. Chronic (100-day) self-administration of cocaine in Rhesus monkeys resulted in a roughly 50% increase in NET-binding as assessed by in vitro autoradiography (Macey et al., 2003). These changes in noradrenergic plasticity may represent a pathological mechanism that contributes to stress-induced reinstatement and may be corrected by α1 and/or β AR antagonists.

Medial prefrontal cortex

Noradrenergic signaling in the mPFC has been implicated in psychostimulant use disorder. Selective depletion of NE in the prefrontal cortex prevents the formation of CCP to amphetamine (Ventura et al., 2003). Microinjection of the α₁ AR antagonist terazosin directly into the mPFC blocked cocaine-induced reinstatement of cocaine-seeking behaviors, but direct injection into either the NAc or VTA had no effect (Schmidt et al., 2017). This suggests that many of the effects of α₁ AR antagonists observed following systemic administration may be due to action of the drug in the mPFC. A recent study further characterized interactions between stress and cocaine CPP in the mPFC. Exposure to acute restraint stress enhanced cocaine CPP in rats (Shinohara et al., 2019). This was the result of synchronized transmission of both NE and DA in the mPFC, where DA was required for CPP, but NE enhanced that preference. This is likely driven by enhanced by cortical pyramidal cell excitability, through an α₁ AR-mediated mechanism (Shinohara et al., 2020; Wada et al., 2020).

Ventral tegmental area

While DA release in the NAc core mediates the rewarding properties of psychostimulants, noradrenergic innervation of the VTA plays an important modulatory role in drug preference. Conditioned stimulus-induced drug seeking, where drug use is paired with a conditioned stimulus, depends on phasic DA signaling in the NAc core (Solecki et al., 2013; Stuber et al., 2005). A recent study demonstrated that blocking α₁ ARs in the VTA attenuated conditioned stimulus-induced reinstatement of cocaine seeking (Solecki et al., 2019a). NE also interacts with CRF to modulate VTA excitability. Co-injection of a CRF receptor agonist and an α₁ AR agonist enhanced CPP for cocaine. This effect is mediated, at least in part, by potentiation of glutamatergic inputs onto VTA DA neurons via enhanced IP3-Ca²⁺ and NMDA receptor activity (Tovar-Díaz et al., 2018).

As was the case with OUD and AUD, clinical and preclinical studies suggest that α₁ AR antagonists, α₂ AR agonists, β AR antagonists may be useful for the treatment of psychostimulant use disorder. Data from rodent studies suggests that α₁ AR antagonists may work by blocking postsynaptic NE signaling in the mPFC or VTA. In particular, reducing NE signaling in the LC to mPFC pathway may be important for reducing cocaine-induced reinstatement f cocaine use. Further, blocking LC to VTA noradrenergic inputs may be effective in reducing the rewarding effects of cocaine and other psychostimulants by indirectly modulating NAcc DA signaling. α₂ AR agonists may work by acting presynatpically to reduce NE release in the BNST, which may impact downstream targets, including the VTA. While there is no clinical data investigating the use of β AR antagonists for psychostimulant use disorder, preclinical data in rodents suggests that blocking β₂ ARs in the BNST or CeA may be a useful target.

CONCLUSIONS AND FUTURE DIRECTIONS

The studies discussed in this review clearly demonstrate that NE signaling and circuitry is relevant for multiple misused drugs. NE is important for the rewarding properties of opiates, as demonstrated by alterations in opiate CPP when NE is altered experimentally, and NE is critical for stress-induced reinstatement of opioid seeking and opioid withdrawal. NE is necessary for voluntary consumption of ethanol and stress-induced reinstatement of ethanol seeking. NE is also necessary for psychostimulant CPP and stress-induced reinstatement.

Numerous pharmacological studies have demonstrated that ARs are critical mediators of stress-induced reinstatement of drug consumption and withdrawal from opioids. Further, the recent success of α₁ AR antagonists and α2 AR agonists in clinical trials suggests that preclinical animal studies are translatable to human substance use disorder (clinical trials summarized in Table 1) (Hendrickson and Raskind, 2016; Simpson et al., 2018). While ample clinical and preclinical evidence supports the use of noradrenergic compounds, there are still significant barriers to their clinical use. All of the noradrenergic drugs tested in clinical trials for SUD are also used to reduce blood pressure clinically, and, as such, can produce undesirable side effects associated with a reduction in blood pressure, including fainting, in SUD patients (Kahn et al., 1997; Keaney et al., 2001; Lin et al., 1997). As such, noradrenergic compounds may be more effective in SUD patients with co-occurring hypertension. Further, additional drug discovery efforts are needed to identify noradrenergic signaling modulators that do not have hypotensive effects. Additionally, sex and pharmacogenetic effects, including polymorphisms in DBH and ADRA1D, also influence the efficacy of noradrenergic drugs in SUD, and should be considered when evaluating the success or failure of clinical trials (Fox et al., 2014; Shorter et al., 2020; Zhang et al., 2019).

Table 1.

Clinical trials of noradrenergic drugs in SUD

Substance use disorder studied	Type of study	Number of subjects	Compound tested	Outcome	Reference
Opioid Use Disorder	Double-blind trial	28	Lofexidine Clonidine	Decreased self-reported withdrawal symptoms with both compounds. Fewer adverse hypotensive events with lofexidine.	(Kahn et al., 1997)
	Double-blind randomized trial	80	Lofexidine Clonidine	Equal decrease in withdrawal symptoms for clonidine & lofexidine. Fewer hypotensive effects with lofexidine with better treatment retention	(Lin et al., 1997)
	Randomized multi-center double-blind placebo-controlled trial	264	Lofexidine	Significant decrease in withdrawal scoring in lofexidine vs placebo. Improved treatment retention in lofexidine vs placebo	(Gorodetzky et al., 2017)
	Randomized multi-center double-blind placebo-controlled trial	68	Lofexidine	Significant decrease in withdrawal scoring in lofexidine vs placebo. Improved treatment retention in lofexidine vs placebo	(Yu et al., 2008)
Alcohol Use Disorder	Double-blind randomized Placebo controlled trial	180	Atenolol	Reduced percentage of treatment failure and alcohol cravings in patients that received atenolol versus placebo.	(Horwitz et al., 1989)
	Double-blind randomized placebo-controlled trial	120	Atenolol	Reduced withdrawal symptoms, alcohol cravings, and treatment failure in patients that received atenolol versus placebo.	(Gottlieb et al., 1994)
	Double-blind comparative trial	28	Propranolol	Propranolol and diazepam were equally effective in diminishing alcohol withdrawal symptoms And anxiety	(Bailly et al., 1992)
	Double-blind randomized placebo-controlled trial	72	Chlordiazepoxide + lofexidine	Chlordiazepoxide + lofexidine group had significantly worse withdrawal symptoms, more adverse hypotensive effects, and similar treatment retention compared to chlordiazepoxide alone	(Keaney et al., 2001)
	Double-blind randomized Placebo-controlled trial	30 (AUD + PTSD)	Prazosin	Prazosin had fewer drinking days and fewer heavy drinking days per week than placebo. No effect on PTSD symptoms.	(Simpson et al., 2015)
	Double-blind randomized placebo-controlled trial	92	Prazosin	Prazosin had a decrease in number of drinks and heavy drinking days over the 12-week course of treatment than placebo. Moderate drowsiness in prazosin group.	(Simpson et al., 2018)
	Double-blind randomized placebo-controlled trial	100	Prazosin	Individuals with severe withdrawal symptoms had a significant reduction in drinking days, heavy drinking days, and drinks per week after prazosin treatment. Individuals with low or no withdrawal symptoms had no benefit from prazosin.	(Sinha et al., 2021)
	Double-blind randomized placebo-controlled trial	25	Clonidine vs diazepam or alprazolam	Prazosin did not significantly improve withdrawal symptoms versus placebo. Both benzodiazepines significantly improved withdrawal symptoms.	(Adinoff, 1994)
	Double-blind randomized placebo-controlled trial	40	Prazosin	Prazosin reduced stress cue-induced alcohol craving, alcohol cue-induced anxiety, basal cortisol, and cue-induced increases in cortisol versus placebo.	(Milivojevic et al., 2020)
	Double-blind randomized placebo-controlled trial	41	Doxazosin	Doxazosin reduced alcohol consumption and heavy drinking days versus placebo. Doxazosin was most effective in individuals with high blood pressure.	(Haass-Koffler et al., 2017)
Cocaine Use Disorder	Double-blind randomized placebo-controlled trial	40	Guanfacine	Guanfacine significantly decreased cocaine cravings and negative emotions in response to a drug cue in females but had no effect in males.	(Fox et al., 2014)
	Double-blind randomized placebo-controlled trial	35	Doxazosin	Doxazosin significantly increased the number of drug free days versus placebo.	(Shorter et al., 2013)
	Double-blind randomized placebo-controlled trial	76	Doxazosin	Doxazosin was less effective in individuals with the (T1848A) variant in the ADRA1D gene.	(Shorter et al., 2020)
	Double-blind randomized placebo-controlled trial	76	Doxazosin	Doxazosin was more effective in patients with lower noradrenergic tone caused by the (C1021T) variant in the DBH gene	(Zhang et al., 2019)

Preclinical animal studies have greatly expanded our understanding of how noradrenergic signaling in specific brain regions including, BNST, VTA, mPFC, and LC, modulates different stages of SUD (summarized in Table 2). However, it is significantly less clear which specific cell types in these brain regions mediate these effects and how connections between these brain regions impact substance use disorders. For example, while β AR antagonist administration into the BLA reduces stress-induced reinstatement of alcohol-seeking behavior it is unclear which specific outputs of the BLA are affected by this treatment (Chesworth and Corbit, 2018). Likewise, while noradrenergic inputs from the LC/A6 to the BLA have been studied extensively, noradrenergic inputs from the A2/NTS have garnered significantly less scrutiny, even though A2/NTS inputs to the BLA are more extensive (Robertson et al., 2013). Furthermore, it is unclear if LC neurons projecting to the prefrontal cortex differ in their response to opioids/opioid withdrawal from the projections to the BLA (Uematsu et al., 2017). Likewise, α₂ AR agonists show great therapeutic promise for opioid use disorder. However, it is not clear which precise input or outputs in the BNST mediate their effects on stress-induced reinstatement of opioid seeking. Accordingly, future studies can use chemogenetic, optogenetic, fluorescent imaging, and other viral approaches to make further advances in our understanding of the precise circuitry that mediates the effects of noradrenergic drugs on substance use disorders.

Table 2.

Summary of animal studies examining noradrenergic system in SUD

Substance Use Disorder Type	Model	Brain region(s) targeted	Manipulation/drug used	Consumption	CPP	Stress-induced reinstatement	Drug cue-induced reinstatement	Withdrawal	Reference
Opioid Use Disorder	Mouse	Global	DBH KO	n/a	↓	n/a	n/a	n/a	(Olson et al., 2006)
	Mouse	Global	ADRA1A KO	n/a	↓	n/a	n/a	n/a	(Drouin et al., 2002)
	Mouse	A2/NTS	Restoration DBH expression in DBH KO mouse	n/a	↑	n/a	n/a	n/a	(Olson et al., 2006)
	Rat	VNB	Lesion	n/a	n/a	↓	n/a	n/a	(Wang et al., 2001)
	Rat	BNST	Clonidine	n/a	n/a	↓	n/a	n/a	(Wang et al., 2001)
	Rat	DNB	Lesion	n/a	n/a	−	n/a	n/a	(Wang et al., 2001)
	Rat	LC	Clonidine	n/a	n/a	−	n/a	n/a	(Shaham et al., 2000)
	Rat	Global	Idazoxan following morphine naloxone	n/a	n/a	n/a	n/a	↑	(Fox et al., 2017)
	Rat	LC	Downregulation of CREB	n/a	n/a	n/a	n/a	↓	(Han et al., 2006)
	Rat	LC	Orexin 1 receptor antagonist (SB-334867)	n/a	n/a	n/a	n/a	↓	(Davoudi et al., 2016)
Alcohol Use Disorder	Mouse	Global	DBH KO	↓	n/a	n/a	n/a	n/a	(Weinshenker et al., 2000)
	Rat	Global	Lofexidine	↓	n/a	↓	n/a	n/a	(Lê et al., 2005)
	Rat	Global	Clonidine	↓	n/a	n/a	n/a	n/a	(Rasmussen et al., 2014)
	Rat	Global	Prazosin	↓	n/a	n/a	n/a	n/a	(Walker et al., 2008)
	Rat (CIE)	Global	Prazosin	↓	n/a	n/a	n/a	n/a	(Skelly and Weiner, 2014)
	Rat	Global	Propranolol	↓	n/a	n/a	n/a	n/a	Gilpin and Koob, 2010)
	Mouse	BLA	Pindolol	↓	n/a	n/a	n/a	n/a	Patkar et al., 2019)
	Rat	BLA	Propranolol	↓	n/a	n/a	↓	n/a	Chesworth and Corbit, 2018)
	Rat	VTA	Prazosin	↓	n/a	n/a	n/a	n/a	Shelkar et al., 2017)
Psychostimulant Use Disorder	Mouse	Global	NET KO	n/a	↓	n/a	n/a	n/a	Ventura et al., 2003)
	Rat	Global	Nepicastat (DBH inhibitor)	↓	n/a	↓	↓	n/a	Schroeder et al., 2013)
	Rat	Global	Clonidine Lofexidine Guanabenz	n/a	n/a	↓	n/a	n/a	Erb et al., 2000)
	Rat	BNST or CeA	Betaxolol & ICI-118,551 (β2 AR antagonist)	n/a	n/a	↓	−	n/a	Leri et al., 2002)
	Rat	BNST	ICI-118,551 (β2 AR antagonist)	n/a	n/a	↓	n/a	n/a	Vranjkovic et al., 2014)
	Mouse	BNST	Rescue of ADRA2A KO with DREADD	n/a	n/a	↑	n/a	n/a	Perez et al., 2020)
	Rat	mPFC	Terazosin	n/a	n/a	n/a	↓	n/a	Schmidt et al., 2017)
	Rat	NAcc	Terazosin	n/a	n/a	n/a	−	n/a	Schmidt et al., 2017)
	Rat		Terazosin	n/a	n/a	n/a	−	n/a	Schmidt et al., 2017)
	Rat	VTA	Terazosin	n/a	n/a	−	↓	n/a	Solecki et al., 2019a)

Highlights.

• Anatomical and pharmacological discussion of the noradrenergic system and circuits.

• Discussion of the role of various components of the noradrenergic system in models of opioid, alcohol, and psychostimulant use disorders.

• Discussion of the clinical studies that have targeted noradrenergic receptors and systems to alleviate substance use disorders.

Abbreviations:

NE

norepinephrine

RVLM

rostroventrolateral medulla

NTS

nucleus of the tractus solitaries

LC

locus coeruleus

CeA

central nucleus of the amygdala

DNB

dorsal noradrenergic bundle

BNST

bed nucleus of the stria terminalis

BLA

basolateral amygdala

VNB

ventral noradrenergic bundle

AR

adrenergic receptor

DA

dopamine

SUD

substance use disorder

HPA

hypothalamic-pituitary-adrenal axis

CRF

corticotrophin releasing factor

ACTH

adrenocorticotropic factor

PTSD

post-traumatic stress disorder

VTA

ventral tegmental area

NAcc

nucleus accumbens

SA

self-administration

OUD

opioid use disorder

DBH

dopamine β-hydroxylase

CPP

conditioned place preference

NET

norepinephrine transporter

OR

opioid receptor

CPA

conditioned place aversion

AUD

alcohol use disorder

mPFC

medial prefrontal cortex

OFC

orbitofrontal cortex