Dynorphin and its role in alcohol use disorder

Abstract

The dynorphin / kappa opioid receptor (KOR) system has been implicated in many aspects that influence neuropsychiatric disorders. Namely, this system modulates neural circuits that primarily regulate reward seeking, motivation processing, stress responsivity, and pain sensitivity, thus affecting the development of substance and alcohol use disorder (AUD). The effects of this system are often bidirectional and depend on projection targets. To date, a majority of the studies focusing on this system have examined the KOR function using agonists and antagonists. Indeed, there are studies that have examined prodynorphin and dynorphin levels by measuring mRNA and tissue content levels; however, static levels of the neuropeptide and its precursor do not explain complete and online function of the peptide as would be explained by measuring dynorphin transmission in real time. New and exciting methods using optogenetics, chemogenetics, genetic sensors, fast scan cyclic voltammetry are now being developed to detect various neuropeptides with a focus on opioid peptides, including dynorphin. In this review we discuss studies that examine dynorphin projections in areas involved in AUD, its functional involvement in AUD and vulnerability to develop AUD at various ages. Moreover, we discuss dynorphin’s role in promoting AUD by dysregulation motivation circuits and how advancements in opioid peptide detection will further our understanding.

1. Introduction to dynorphin

Dynorphin, the endogenous peptide for the kappa opioid receptor, is a complex neuropeptide known to exist in many forms, yet we do not have a good grasp of the function of each of these forms or fragments. Dynorphin was first purified, sequenced and named Dynorphin A (1–17)(Cox et al., 1975; Goldstein et al., 1981, 1979). Interestingly, the amino-acid sequence was found to be the same as leu-enkephalin (YGGFL) but with a basic carboxy-terminal extension. This is one of reasons why detecting peptide changes in vivo has been so challenging. The discovery of Dynorphin B (13 amino acids) or rimorphin followed, in conjunction with the c-terminus extended form of dynorphin B, Big Dynorphin and leumorphin (Fischli et al., 1982; Kilpatrick et al., 1982; see Ferre et al., 2019 for sequences). This was preceded by discovery of the endogenous morphine-like substances and opioid peptides methionine and leucine enkephalin (Cox et al., 1976, 1975; Goldstein et al., 1971; Hughes et al., 1975; Li and Chung, 1976; Teschemacher et al., 1975; Tperenius and Wahlström, 1975). This era has been fondly described as the ‘golden era of neuropeptide discovery’ (Chavkin, 2013) a critical period during which our understanding of neuropeptide function was enhanced following new methods of neuropeptide detection and purification. In response to this and in the period of time that followed there was a drive to better understand the biochemistry and structural biology of these opioid peptides and as a result the cDNA was cloned and sequenced for prodynorphin, the precursor for dynorphin (Kakidani et al., 1982).

During the same period there was also a push to define and better understand the existence of specific opioid receptors mu, sigma and kappa (Martin et al., 1976), as well as mu-, delta-, kappa-type action using in vitro bioassays (Lord et al., 1977). There were a number of limitations with these assays at the time that prevented determination of receptor selectivity (Chavkin, 2013) until the development of selective-receptor protection strategy using the nitrogen mustard analog of naltrexone, β-chlornaltrexamine (β-CNA) (Portoghese et al., 1978), which ultimately resulted in confirmation that dynorphin A was an endogenous ligand for the ‘physically distinct, non-converting kappa opioid receptor’ (Chavkin, 2013; Chavkin et al., 1982). Following the discovery that different opioid receptors exist and have selective endogenous peptides, the focus of the field shifted primarily towards using this knowledge to ascertain the function of these opioid receptor systems in analgesia and addiction (discussed in detail further in the review).

Importantly the cloning of the opioid receptor genes (Chen et al., 1993; Evans et al., 1992; Kieffer et al., 1992; Mollereau et al., 1994; Yasuda et al., 1993), very much advanced our knowledge of the dynorphin/KOR signaling system as part of the G protein-coupled receptor (GPCR) superfamily, specifically Gi/Go-coupled (for a detailed review see (Al-Hasani and Bruchas, 2011). For example, we now know that a distinct characteristic of KOR is the high negative potential at extracellular loop 2, that has been shown experimentally to increase the surrounding concentration of dynorphin due to its highly positive charge (Björnerås et al., 2014; Wang et al., 1994).

To date we know that prodynorphin is primarily expressed in the central nervous system, where upon depolarization of prodynorphin-containing neurons, protein convertase 2 processing occurs to release dynorphin from within these presynaptic neurosecretory vesicles. Dynorphin in turn primarily activates KOR thereby modulating neurotransmitter release and postsynaptic activity (Chavkin et al., 1983; Civelli et al., 1985; Ferré et al., 2019; Molineaux and Cox, 1982; Pickel et al., 1993; Wagner et al., 1993, 1991). However, we do not have a clear grasp of the recycling and degradation kinetics of dynorphin following release. Furthermore, it is not clear which or when specific dynorphin fragments are released or necessary for signaling. Interestingly, it seems that the content/distribution of unprocessed prodynorphin and dynorphin peptides may differ in axon terminals depending on the brain region (Yakovleva et al., 2006), which suggests that regional differences in trafficking and processing of the protein exists, further complicating matters.

Over the years research efforts have flip flopped between focusing on better understanding the function and dynamics of dynorphin versus KOR receptor function and binding to allow us to better understand their mechanisms of action. In the last 10 or so years there has been increased effort to better visualize and solve the structure of not only the KOR receptor in their active and inactive state but GPCRs in general, using crystallography and electron microscopy (Che et al., 2018; Wu et al., 2012). Despite these impressive advances it is now more apparent than ever that in conjunction with a better understanding of KOR receptor function there are limitations in our ability to study the function of dynorphin in vivo, which the field is currently trying to address and is discussed below.

2. Dynorphin projections

Initial traces of dynorphin were observed as early as embryonic age 14 in mouse brain (Rius et al., 1991). In the adult brain, dynorphin is found in various regions. Relatively high levels of dynorphin were measured in the amygdala, entorhinal cortex, dentate gyrus, hypothalamus, midbrain, striatum, hippocampus, and medulla-pons (Fallon and Leslie, 1986; Goldstein and Ghazarossian, 1980; Schwarzer, 2009). In comparison, lower levels of dynorphin were observed in the cerebellum and cortex (Goldstein and Ghazarossian, 1980). These studies indicate that dynorphin is widely distributed in the brain, particularly in regions that are critical in the affective modulation of a number of neuropsychiatric diseases including addiction and mood disorders. In the following section we will explore dynorphin projections in the mesolimbic regions, including the extended amygdala.

2.1. Ventral Tegmental Area

The ventral tegmental area (VTA) provides dopaminergic innervation to many brain areas and receives afferents containing various neurotransmitters and neuromodulators from other regions. Neurons in the VTA have been known to be affected by ethanol. For example, acute ethanol directly excites dopamine neurons in the VTA (Brodie et al., 1999) and results in increased AMPA/NMDA ratio (Saal et al., 2003). Moreover, intra-VTA infusion of ethanol via self-administration is reinforcing and increases subsequent self-administration (Rodd et al., 2004). During withdrawal however, spontaneous activity dopamine neurons is reduced (Diana et al., 1992; Shen and Chiodo, 1993). Furthermore, acute alcohol exposure inhibits firing of GABA neurons in the VTA (Steffensen et al., 2009) followed by a large increase in GABA neuron activity during acute and chronic ethanol induced withdrawal (Gallegos et al., 1999; Steffensen et al., 2009). A recent study has shown that withdrawal following both acute and chronic ethanol modulate glutamate transmission in the VTA (Williams et al., 2018). Specifically, acute-ethanol induced withdrawal resulted in decreased spontaneous excitatory presynaptic currents (EPSCs) and amplitudes of miniature EPSCc; in contrast, withdrawal from chronic intermittent ethanol exposure resulted in greater frequency of presynaptic events (spontaneous and miniature EPSCs). Moreover, systemic administration of ethanol acutely results in elevation of dynorphin levels (Jarjour et al., 2009). Together, these studies show that the VTA is a crucial structure involve is ethanol-related processing and the overall responsivity of this brain region may influence the development of alcohol use disorder (AUD). Though not studied in detail, VTA receives dynorphinergic projections from the lateral hypothalamus and the striatum. In the following sub-sections, we will discuss the impact of that these afferents may have on reward seeking behaviors based on their synaptic targets within the VTA.

Lateral hypothalamic afferents

The VTA receives inputs from several regions containing prodynorphin expressing neurons. These neurons impinge upon VTA projection neurons and modulate reward processing and seeking behaviors. Neurons in the lateral hypothalamus, a region with a substantial projection to the VTA (Fallon et al., 1985; Iyer et al., 2018), have been recently implicated in reward seeking (España et al., 2011, 2010; Richardson and Aston-Jones, 2012). Within the VTA, these projection neurons make connections with dopaminergic cell bodies which in turn project to the medial and lateral nucleus accumbens shell (NAc shell) and the basolateral amygdala (BLA), and inhibit them via activation of its cognate receptor, the KOR (Baimel et al., 2017). Behaviorally, dynorphin release in the VTA via activation of these lateral hypothalamic neurons results in a reduction of reward seeking, possibly due to their inhibitory effects on mesolimbic dopamine projection neurons (Muschamp et al., 2014).

Striatal afferents

In addition to the lateral hypothalamus, striatal neurons also project to the VTA. The output neurons of dorsal and ventral striatum consist of medium spiny neurons (MSNs) expressing either dopamine-1 (D1) receptor or dopamine-2 (D2) receptor. Both, D1- and D2-MSNs are inhibitory and release GABA. D1-MSN neurons express prodynorphin and release dynorphin along with GABA (Al-Hasani et al., 2015). The striatum is divided in dorsal and ventral aspects consisting of the caudate-putamen and the nucleus accumbens (NAc), respectively. D1-MSN projections from dorsal striatum terminate in substantia nigra pars compacta and the VTA while the ones from ventral striatum only project to the VTA (Fallon et al., 1985). Ventral striatum, as referred to as the NAc, provides one of the most prominent feedback projections to the VTA (Watabe-Uchida et al., 2012). These striatal D1-MSN neurons preferentially impinge upon non-dopaminergic GABA neurons in the VTA resulting in disinhibition of dopamine neurons (Bocklisch et al., 2013; Chuhma et al., 2011; Xia et al., 2011). However, Yang and colleagues (Yang et al., 2018) recently showed a regionally specific projection of D1-MSNs from the NAc to the VTA. Specifically, they showed that D1-MSNs from medial NAc shell impinge upon VTA dopamine neurons predominantly projecting to medial NAc shell. In contrast, the same study showed that D1-MSNs from lateral NAc shell target VTA GABA neurons. Both projections were observed to exert inhibitory control over their respective targets, therefore resulting in opposing actions. Dopaminergic and non-dopaminergic neurons in the VTA express KORs; activation of these receptors results in a reduction of firing rate of these neurons (Ford et al., 2006; Margolis et al., 2006, 2005, 2003). Thus, even though D1-MSNs corelease GABA and dynorphin, it is possible that at least some of the inhibition is driven by dynorphin transmission. It is important to note however, that while VTA dopamine neurons projecting to NAc have been shown to not express KORs, KORs are present on VTA dopamine neurons projecting to prefrontal cortex (PFC) and the BLA (Margolis et al., 2006). Actions of dynorphin via KORs have been reported at both, pre- and post-synaptic locations (Ford et al., 2007).

In summary, dynorphin containing projections form the lateral hypothalamus and the NAc impinge upon both dopaminergic as well as GABAergic neurons in the VTA. Data from studies in the recent years have certainly added substantial understanding of how affective and reward-seeking behaviors are shaped by dynorphin and have helped identify targeted cell-types of these projection neurons. Nevertheless, more studies are necessary to fully identify targets of these projections, including topographical organization. Complete understanding at that level will enable us to clarify the role of dynorphin in addictive behaviors.

2.2. Nucleus Accumbens

Similar to VTA, the NAc is implicated in AUD based on findings of preclinical studies demonstrating changes in accumbal neuron activity following acute and chronic ethanol exposure. Acute systemic administration of ethanol (physiologically relevant doses, 0.5 – 2 g/kg) increases extracellular levels of dopamine in the NAc in ethanol naïve rodents (Imperato and Di Chiara, 1986; Karkhanis et al., 2016b, 2014; Yim et al., 2000) while very high doses of ethanol decrease dopamine levels (Imperato and Di Chiara, 1986; Yim and Gonzales, 2000). A similar effect, inhibition of dopamine at high concentrations of ethanol, is observed in ex vivo preparations (Yorgason et al., 2015, 2014); this effect is modulated by nicotinic acetylcholine receptors (Yorgason et al., 2015). Chronic intermittent ethanol exposure however, results in hypodopamine state in rodents and macaques possibly driving anhedonia, subsequently leading to excessive ethanol consumption (Karkhanis et al., 2016a; Rose et al., 2016; Siciliano et al., 2016a, 2016b). Furthermore, following chronic ethanol exposure low to moderate doses of ethanol administered systemically during acute withdrawal result in reduced extracellular levels of dopamine (Karkhanis et al., 2016a). Surprisingly, the opposite effect is observed in rats exposed to social isolation during adolescence, that is, systemic administration of ethanol following prolong adolescent social isolation, which also results in excessive ethanol consumption, augments ethanol-induced elevation in dopamine (Karkhanis et al., 2016b).

In addition to dopamine, glutamatergic transmission within the NAc also plays an important role in alcohol seeking and consumption. Activation of cortical glutamatergic afferents into the NAc is essential for cue-induced alcohol seeking (Keistler et al., 2017). In contrast, activation of glutamatergic projection from the basolateral amygdala to the NAc disrupted cue-induced alcohol seeking (Keistler et al., 2017). Moreover, glutamate receptors on medium spiny neurons in the NAc alter their morphology after chronic ethanol exposure, further affecting alcohol consumption and seeking (Laguesse et al., 2017). A recent study has shown increased c-Fos expression following chronic intermittent ethanol exposure during acute withdrawal in the NAc shell and core; this augmented c-Fos expression is present during protracted withdrawal (7 days post final ethanol exposure) in NAc shell (Smith et al., 2019).

In addition to the neurotransmitter changes highlighted above, neuropeptides, such as dynorphin and met-enkephalin are also altered by ethanol in the NAc. For example, systemic administration of ethanol elevates dynorphin and met-enkephalin levels (Marinelli et al., 2006, 2005). Interestingly, the dopamine inhibiting and augmented elevating effects of dopamine following chronic intermittent ethanol exposure and social isolation, respectively, are reversed following KOR blockade with norBNI (Karkhanis et al., 2016b, 2016a). These studies indicate that the dynorphin/KOR system interacts with neurotransmitters in the NAc and are involved in shaping the progression towards maladaptive reward and alcohol seeking behaviors as well as AUD. NAc consists of local and projecting dynorphin neurons; we will discuss these projections and the influence dynorphin may have in shaping synaptic transmission of various neurotransmitters within the NAc.

Lateral hypothalamic afferents

Like the VTA, NAc receives dynorphin projections from lateral hypothalamus. In fact, many of the dynorphin expressing neurons projecting from the lateral hypothalamus to the VTA reside in the same topographical location and are similar in response properties to the ones that project to NAc (Iyer et al., 2018). In fact, this study showed that about 21% of the neurons from the lateral hypothalamus project to both NAc and the VTA. Topographical survey of the terminal location of these lateral hypothalamic neurons indicates that these neurons terminate in the NAc core as well as the medial and lateral aspects of NAc shell (Baldo et al., 2003; Brown et al., 2013; Iyer et al., 2018), however information identifying the target neurons of these hypothalamic projections within the NAc is lacking.

Local collaterals on dopamine terminals

D1-MSNs residing in the NAc co-release dynorphin and GABA. As mentioned in section 2.1, these neurons project to the VTA and impinge upon dopamine projection neurons and local GABAergic interneurons within the VTA. Axon collaterals from these neurons also synapse onto local dopamine terminals arising from the VTA and other cell types within the NAc. Thus, dynorphin containing cell bodies and axon terminals are present in the NAc (Van Bockstaele et al., 1995). The KORs are located presynaptically on dopamine terminals in the NAc (Svingos et al., 2001; Werling et al., 1988) and their activation reduces dopamine release (Britt and McGehee, 2008; Karkhanis et al., 2016a, 2016b; Thompson et al., 2000) while increasing dopamine uptake, leading to an overall decrease in extracellular dopamine levels (Thompson et al., 2000). Likewise, dopamine transmission in the NAc regulates dynorphin; ablation of mesostriatal dopamine neurons with 6-hydroxydopamine resulted in a reduction of dynorphin mRNA levels and increasing dopamine levels by apomorphine administration increased dynorphin mRNA levels (Gerfen et al., 1991). Together, studies showing bidirectional relationship between dopamine and dynorphin imply that while dynorphin release-mediated excitation of KORs can inhibit dopamine release, dopamine transmission can control the excitability of these same dynorphin containing D1-MSNs.

Local collaterals on other neurotransmitter terminals

In addition to presynaptic location on dopamine terminals, KORs are also expressed presynaptically on glutamatergic and serotoninergic neurons, and pre- and post-synaptically on GABAergic neurons (Land et al., 2009; Svingos et al., 2001; Tao and Auerbach, 2002). KORs on serotonin terminals in the NAc promote serotonin transporter insertion following repeated exposure to stress, thus resulting in a reduction in serotonin transmission, driven by endogenous dynorphin release (Schindler et al., 2012). NAc receives most of its glutamatergic afferents from the PFC, ventral hippocampus, and the basolateral amydgala (BLA) (Britt et al., 2012). KORs inhibit glutamate presynaptically in a pathway specific manner; this inhibitory control results in attenuation of D1-MSN excitation specifically by activation of KORs on BLA glutamate projection but not on hippocampal glutamate projections (Tejeda et al., 2017). Effects of KOR activation on cortical glutamate afferents into the NAc are still unknown. Serotonin afferents into the NAc arise from the dorsal raphe and these terminals express KORs (Land et al., 2009; Tao and Auerbach, 2002). KORs are present on both D1- and D2-MSNs and their activation inhibit D2-MSNs collaterals; however, this inhibitory effect is minimal in comparison to KOR mediated inhibition of D1-MSNs (Tejeda et al., 2017). Therefore, the effect of dynorphin in the overall excitation of output D1-MSNs of NAc is the sum of inhibition and disinhibition of glutamate, serotonin, and dopamine transmission.

Regional differences

Previous literature indicates that there are regional differences in dynorphinergic and KOR control over hedonic and dysphoric behavioral responses. Studies have shown that local activation of both D1-MSNs and KORs is essential to modulate aggression and rejection of conspecifics in prairie voles (Aragona et al., 2006; Resendez et al., 2012) and local activation of dynorphin-expressing D1-MSNs in ventral NAc shell drives aversive behavioral response (Al-Hasani et al., 2015). Surprisingly, photostimulation of dynorphin-expressing D1-MSNs in dorsal NAc shell resulted in positively reinforced motivated behaviors (Al-Hasani et al., 2015). Both the aversive and reinforcing responses were inhibited by KOR blockade, therefore implying that these behavioral differences were produced by dynorphin (Al-Hasani et al., 2015). Differences in “liking” and “wanting” have also been observed along the rostro-caudal axis within NAc shell. Specifically, stimulation of KORs in the rostro-dorsal quadrant of NAc shell resulted in hedonic reactions, while stimulation of KORs in the caudal half of NAc shell resulted in a suppression of these hedonic reactions (Castro and Berridge, 2014). These regional differences and the collective effect of dynorphin on various neurotransmitter systems shape affective and motivated behaviors, which are grossly impacted by stress (Castro and Bruchas, 2019). The complexity of the nature of interaction of all these systems highlight the importance of using methods that enable us to selectively modulate specific circuits.

2.3. Bed Nucleus of Stria Terminalis

The bed nucleus of stria terminalis (BNST) is a part of the extended amygdala and plays a major role in stress-induced intake of ethanol. Inhibition of excitatory neural activity within the BNST reduced the magnitude of conditioned place preference for ethanol (Pina et al., 2015). Furthermore, ethanol-induced conditioned place preference, which involves contextual learning and repeated exposures to ethanol, is associated with overall augmented neuronal excitability within the ventral BNST; in contrast, ethanol-mediated conditioned place aversion results in overall attenuation of presynaptic glutamatergic transmission with no change in GABAergic transmission (Pati et al., 2019). Studies have also shown modulation in catecholamine transmission following ethanol administration. Specifically, acute systemic administration of ethanol elevates extracellular levels of dopamine (Carboni et al., 2000) and norepinephrine (Jadzic et al., 2019). Chronic intermittent exposure to ethanol results in increased c-Fos expression in the dorsal and ventral BNST during acute withdrawal; however, this effect is observed only after 26 hours into withdrawal in dorsal BNST (Smith et al., 2019).

In addition to changes in neurotransmitters, the dynorphin/KOR system within the BNST also regulates behaviors associated with ethanol self-administration. Dynorphin in the BNST has been shown to regulate anxiolytic phenotype; specifically, deletion of KORs on glutamate inputs in the BNST augments anxiety-like behaviors (Crowley et al., 2016). This alteration in anxiety-like behavior may, at least in part, further regulate ethanol-seeking behaviors. Activation of KORs by systemic administration of the agonist increases stress-aversive and anxiety-like behaviors and promotes ethanol intake and seeking as well as reinstatement of ethanol seeking; subsequent local inhibition of KORs within the BNST reverses these behaviors (Lê et al., 2018). Together these studies indicate that BNST is involved in AUD and dynorphin activity within the BNST may shape its contribution to the development and/or maintenance of AUD. In the following sub-sections, we will discuss neural circuits that are modulated by dynorphin within the BNST.

Dynorphin influence on GABA transmission

KORs are broadly expressed in the BNST (Poulin et al., 2009). It has been shown that high levels of expression of prodynorphin mRNA, and therefore dynorphin, were observed in fusiform, oval, rhomboid, and anterior lateral nuclei of the BNST, with moderate levels of expression in anteromedial and interfascicular nuclei (Crowley et al., 2016; Fallon and Leslie, 1986; Kash et al., 2014; Poulin et al., 2009). These dynorphin containing neurons also contain GABA (Crowley et al., 2016). High density of KORs were observed in the dorsomedial, principal, and ventral nuclei of the BNST (Marchant et al., 2007; Poulin et al., 2009). Interestingly, the expression patterns of dynorphin and the cognate KOR are complementary as labeling for prodynorphin and KORs was not observed in the same nuclei, except for rhomboid nucleus, where expression of both prodynorphin and KORs were observed (Poulin et al., 2009). The BNST receives a strong GABAergic projection, co-expressing dynorphin, from lateral central amygdala (CeA; Marchant et al., 2007) and the paraventricular hypothalamus (Crowley et al., 2016). Neurons in lateral CeA have been observed to express prodynorphin, and about 33% of these prodynorphin-expressing neurons, co-express corticotrophin releasing hormone (Marchant et al., 2007). An overall reduction in GABA transmission via presynaptic mechanisms was observed following KOR activation in the BNST; selective circuit manipulation using optogenetics revealed that this KOR-mediated reduction in GABA transmission was specific to the CeA→BNST projection (Li et al., 2012; Normandeau et al., 2018).

Dynorphin influence on glutamate transmission

In addition to GABAergic afferents, BNST receives glutamatergic projections from the BLA and the PFC (Kash, 2012). Surprisingly, KOR activation using an agonist as well as local dynorphin neuron stimulation within the BNST selectively inhibit BLA glutamate transmission but not PFC glutamatergic projection in the BNST (Crowley et al., 2016). The glutamate afferents from the BLA target both dynorphin positive and negative neurons; however, their control over dynorphin positive neurons is significantly greater than dynorphin negative neurons (Crowley et al., 2016).

In summary, of all the GABAergic and glutamatergic afferents in the BNST, the dynorphin/KOR system selectively inhibits GABA transmission in the CeA→BNST projections and glutamate transmission of the BLA→BNST projections. While we don’t know the exact net result of this mixed modulation, what we do know is that deletion of KORs from glutamatergic BLA terminals in the BNST results in augmentation of anxiolytic behavior (Crowley et al., 2016). These data imply that akin to the NAc, the effect of the dynorphin/KOR system is specific to targeted activation or inhibition.

In conclusion, dynorphin release and therefore KOR activation results in mixed behavioral effects even though the receptors couple to inhibitory G-protein therefore reducing neurotransmission. Specifically, activation of dynorphin containing neurons in dorsal NAc results in positively reinforced motivated behaviors while activation of dynorphin neurons in the ventral NAc promote aversive behaviors. Furthermore, opposing effects of KOR stimulation have been observed along the rostro-caudal axis in the NAc, with activation in rostral regions driving hedonic response and caudal regions suppressing these hedonic responses. In the BNST activation of KORs results in anxiolytic effects, which is the opposite of what is found in other areas. Collection of these studies highlights the importance of receptors, cell, and circuit specificity. While these are results that we observe in rodents naïve to experimental perturbations such as stress and drug exposure, the complexity of the system can only hint to differential responses that may be observed following chronic stress and drug exposure. The following sections will focus on these altered states.

3. Role of dynorphin in alcohol use disorder

The dynorphin/KOR system is modulated by both chronic and acute exposure to ethanol across various brain regions. A number of studies using the chronic intermittent ethanol (CIE) exposure paradigm using inhalation chambers have thoroughly examined the effect of CIE exposure on KOR function (Anderson et al., 2016a; Karkhanis et al., 2016a; Rose et al., 2016). These studies show that CIE exposure during adulthood results in increased KOR function in the NAc measured by recording KOR-activation mediated inhibition of dopamine release; these effects were observed immediately after and 72 hours after cessation of the last ethanol exposure (Karkhanis et al., 2016a; Rose et al., 2016). CIE-induced escalation in ethanol intake was inhibited by KOR blockade (Anderson et al., 2016a; Rose et al., 2016). Likewise, after six months of chronic ethanol consumption in monkeys, KORs were observed to be hyperfunctioning in the NAc (Siciliano et al., 2016a, 2018). An increase in the KOR gene, Oprk1, mRNA levels has been observed in the BNST following long-term ethanol self - administration in rats (Erikson et al., 2018). Various studies have shown that activation of KORs or augmented KOR function/signaling elevates ethanol consumption (Anderson et al., 2016b, 2016a; Kissler et al., 2014; Rose et al., 2016; Karkhanis et al., 2016b). Though much of the alcohol literature has focused on understanding the function and expression of KORs, it only explains ethanol’s effects partially. To fully understand the effects of ethanol on the dynorphin/KOR system, it is crucial to study the ligand. In fact, one study has shown that ethanol application to striatal neurons in vitro potentiates prodynorphin expression (Logrip et al., 2008). The following section focuses on studies that show changes in dynorphin or prodynorphin expression following acute and chronic ethanol exposure at various ontogenetic timepoints.

3.1. Ethanol exposure in adulthood

Acute ethanol exposure

Despite the difficulty involved in measuring dynorphin levels, a few studies have shown that acute and chronic ethanol exposure positively modulates dynorphin levels across various brain areas. A single ethanol exposure resulted in an upregulation of prodynorphin gene in the PFC and the amygdala (D’Addario et al., 2013). For example, single systemic administration of ethanol results in elevation in dynorphin levels in the NAc (Marinelli et al., 2006), VTA (Jarjour et al., 2009), CeA (Lam et al., 2008), and the paraventricular nucleus (Chang et al., 2007). The peak increase in dynorphin in the NAc was observed at 30 mins following ethanol administration, with a greater elevation at a high dose (3.2 g/kg) of ethanol (Marinelli et al., 2006). Importantly, complementary studies how that dopamine levels do not increase above baseline at high doses beyond 2.5 g/kg (Yim et al., 2000). Furthermore, at this timepoint, 30 mins post injection, dopamine levels have been observed to return to baseline (Karkhanis et al., 2014; Yim et al., 2000). Parametric studies of ethanol, however, have shown a dissociation between brain ethanol concentration and dopamine levels, where ethanol concentrations in the brain last for twice the amount of time after dopamine levels peak (Yim et al., 2000). Together, these studies suggest that the initial elevation of dopamine levels following acute systemic ethanol administration possibly activate D1-MSNs, thus increasing dynorphin levels, which then inhibits dopamine.

Chronic or repeated ethanol exposure

Repeated and chronic exposure to ethanol also increases dynorphin expression in various brain regions. For example, tissue content of dynorphin in the NAc and the periaqueductal gray increased by 30 mins and remained high for 21 days after the final ethanol dose following a 13-day regimen of systemic ethanol administration (Lindholm et al., 2000). This continuous elevation of tissue content opposes the transient elevation observed by Marinelli and colleagues (2006) following a single systemic administration. This difference could be due to multiple exposures verses a single exposure; however, the difference could also be driven by the fact that extracellular levels of dynorphin were measured in one study using microdialysis followed by radioimmunoassay (Marinelli et al., 2006), while total content – intra- and extra-cellular – was measured in the other (Lindholm et al., 2000). While we don’t fully understand the exact differences between the availability of usable intracellular versus extracellular levels of dynorphin, it is possible that total content levels will not necessarily match extracellular levels of dynorphin. This may also be dependent on the brain region in question.

In addition to systemic administration, voluntary consumption of ethanol has also been shown to increase expression of dynorphin or mRNA encoding dynorphin and prodynorphin. Specifically, expression of mRNA encoding dynorphin has been reported to be elevated in the CeA and the hypothalamus in alcohol preferring rats compared to non-preferring rats following voluntary consumption (Zhou et al., 2013). During acute withdrawal, expression of mRNA encoding dynorphin and pro-dynorphin was observed to be increased in the NAc after one-month of voluntary ethanol consumption; the expression levels returned to baseline after 96 hours (Przewłocka et al., 1997). In addition to changes in the NAc, manipulation of dynorphin transmission in the CeA also regulated ethanol intake. Chemogenetic inhibition of dynorphin-containing neurons in the CeA reduced binge-like ethanol intake and activation of KORs using the agonist in the presence of dynorphin inhibition increased binge-like ethanol intake (Anderson et al., 2019). In comparison, chronic ethanol exposure augmented dynorphin expression in the CeA of dependent rats during acute withdrawal (Kissler et al., 2014). Likewise, dependence-induced heightened consumption of ethanol was reversed by KOR blockade, potentially by inhibiting the actions of dynorphin (Berger et al., 2013). This effect was not observed in non-dependent rats. Together these studies show that inhibition of dynorphin transmission in the CeA rescues compulsive ethanol consumption and excessive ethanol exposure augments dynorphin production. Furthermore, blocking dynorphin effects by systemic administration of KOR antagonist also reduces compulsive ethanol intake. It is important to note that in another study, KOR blockade in adult rats with limited access to ethanol (30 min access to 10% ethanol every other day for a total of 3 exposures), resulted in elevation of ethanol intake in male rats and a reduction in female rats (Morales et al., 2014). This study confirms results from other studies showing a lack of KOR-blockade effect in non-dependent male rat. A majority of the studies have been primarily in male animals, which highlights the paucity in literature examining effects of ethanol on the dynorphin/KOR system in females. Clearly, there are differences that need to be studied.

Since rodents do not readily consume ethanol to the point of intoxication or dependence, ethanol vapor exposure is generally employed to induce dependence. Most of the studies examining changes in the dynorphin/KOR system following CIE exposure have focused on examining KOR function and manipulating it. Therefore, results from these studies only implicate the indirect impact of CIE on dynorphin function and transmission. CIE-induced escalation in ethanol intake is reversed by KOR blockade in mice (Anderson et al., 2016a; Rose et al., 2016). Moreover, dependence-induced elevation in operant responding for ethanol is also reduced by KOR blockade in rats (Walker and Koob, 2008). The effect of an acute ethanol challenge on extracellular levels of dopamine in the NAc in dependent mice (CIE exposed), as measured by microdialysis is inhibitory; which is reversed by inhibition of KORs (Karkhanis et al., 2016a). In non-dependent rats, an acute ethanol challenge increases dopamine levels and KOR blockade has no effects on this ethanol-mediated elevation in dopamine (Karkhanis et al., 2016a). The reversal of inhibitory effect of ethanol challenge on dopamine levels and reversal of escalated ethanol consumption in dependent mice and rats following KOR blockade suggests that dynorphin effects, which were blocked, were essential for producing those effects.

Relapse following protracted abstinence

The studies mentioned in the acute and chronic alcohol exposure subsections above have investigated dynorphin role following acute withdrawal – 0 to 72 hours post last ethanol exposure. However, it is important to note that the KOR/dynorphin system may in fact be involved in promoting relapse drinking and/or reinstatement of alcohol seeking after protracted abstinence in non-dependent and dependent animals. For example, KOR activation in the BNST promoted reinstatement of alcohol seeking following seven days of extinction in rats with history of alcohol self-administration; this effect was blocked by inhibition of KORs with norBNI (Lê et al., 2018). Furthermore, KOR activation in dependent rats promoted alcohol seeking at a greater propensity compared to non-dependent animals after 10 days of abstinence (Funk et al., 2019). In addition, reinstatement of ethanol seeking driven by ethanol conditioned stimulus was blocked by naltrexone, an opioid antagonist (Liu and Weiss, 2002), indicating a possible involvement of dynorphin in promoting context dependent reinstatement.

It is known that high anxiety levels, depression, comorbid affective disorders, and exposure to stress increase the propensity and risk of AUD in humans (McCaul et al., 2017; Strine et al., 2012). This anxiety and depression can be a result of preexisting neuropsychiatric disorders or it can be driven by excessive alcohol consumption. Collection of results from rodent studies suggest that stress influences the progression of AUD; stress enhances alcohol reward and alcohol alleviates stress; chronic alcohol compromises behavioral coping and alters stress responsivity; together these changes increase vulnerability to stress triggered relapse (see Becker, 2017 for review). Inhibition of KORs following protracted abstinence has been shown to reduce depressive-like behavior (Jarman et al., 2018), possibly also reducing occurrence of relapse and reinstatement (Lê et al., 2018). Together these studies indicate that in adults, the relationship between dynorphin/KOR system and alcohol is bidirectional. While acute and chronic alcohol exposure modulate dynorphin and KOR function, alcohol seeking, craving, and consumption are in fact also driven by changes in the dynorphin/KOR system.

3.2. Ethanol exposure during adolescence

While the above-mentioned studies show changes in dynorphin following acute or chronic ethanol exposure during adulthood, some studies have examined changes in dynorphin following adolescent ethanol exposure. Adolescence is an vulnerable period during which dynorphin mediated changes are heightened, for example, one study has shown that dynorphin-induced hyperpolarization in paraventricular nucleus neurons increases through puberty (approximately four weeks of age), followed by a return to levels observed at two weeks of age (Chen et al., 2015). Furthermore, ethanol exposure during adolescence promotes heightened ethanol consumption later in life (Spear, 2015). Adolescent ethanol exposure, in alcohol preferring rats, alters the mesolimbic dopamine system to be more sensitive and responsive to ethanol (Toalston et al., 2014), however mechanisms driving this high intake of ethanol are still unknown. A few studies have shown fluctuations in dynorphin levels following repeated ethanol exposure during adolescence. Orogastric exposure to ethanol (2 g/kg) for three consecutive days per week between four and nine weeks of age resulted in augmented dynorphin expression in the substantia nigra when examined three weeks after the last ethanol exposure, but not 2 hours post final exposure (Granholm et al., 2018). Surprisingly, this study did not show an effect of treatment (water vs. ethanol) on dynorphin expression in the NAc at either time points. Another study using voluntary ethanol intake in adolescent rats showed that KOR blockade did not reduce ethanol intake after a three-day repeated ethanol exposure regimen (10% ethanol in supersac solution, every other day); an effect that was observed in adult female rats (Morales et al., 2014). This may be due to housing effects as Morales and colleagues (2014) housed their rats singly (1 rat/cage) starting post-natal day 25. Interestingly, alcohol-induced effects on dynorphin in adolescent rats were dependent on housing condition: dynorphin expression levels were lower in single housed (each single housing session = 24 hours) adolescent rats in the amygdala when measured two hours after the final isolation session, but this effect was rescued by ethanol (20% ethanol, voluntary intake for 3 consecutive days of the week for six weeks; (Palm and Nylander, 2014). Surprisingly, no differences in dynorphin expression levels between group and single housed rats or between water and ethanol drinking single housed rats were found in the NAc (Palm and Nylander, 2014). Because many neural processes are changing during adolescence development, it is difficult to identify specific mechanisms that may drive the observed behavioral changes. The lack of changes in dynorphin observed in some studies may be due to, at least in part, continuous development of processes and pruning during this age.

3.3. Prenatal ethanol exposure

Similar to adolescent ethanol exposure, exposure to prenatal ethanol exposure promotes high ethanol consumption later in life. It is surprising however, that very few studies have actually examined these differences. One study has shown that repeated prenatal ethanol exposure (gestational day 17 – 20; 1 or 2 g/kg) results in significantly greater dynorphin mRNA levels following the 2 g/kg dose, and lower levels of protein content following 1 and 2 g/kg when examined on post-natal day 8 in the NAc (Bordner and Deak, 2015). In the same study, an increase in dynorphin mRNA was observed in the VTA following 2 g/kg when examined on post-natal day 4 (Brodner and Deak, 2015). This study indicates that prenatal ethanol exposure possibly results in an upregulation in dynorphin signaling. Similar results were observed in a separate study, which showed that prenatal (gestation day 17 – 20) ethanol exposure (2 g/kg) resulted in increased prodynorphin mRNA expression in the VTA and a reduction of DNA methylation at the prodynorphin gene of infant and adolescent rats (Wille-Bille et al., 2018). However, the opposite effect, a reduction in prodynorphin mRNA levels was observed in the NAc (Wille-Bille et al., 2018). This prodynorphin mRNA reduction observed in the NAc is likely due to a compensatory adaptation. Overall, rats in this study were observed to have an anxiety-like phenotype. Some adult studies have shown that repeated stress exposure during adolescence and adulthood result in reduced dynorphin mRNA (Donahue et al., 2015) and content (Karkhanis et al., 2016b) and increased anxiety-like behaviors along with facilitation of ethanol consumption (Karkhanis et al., 2016b). It is known that prenatal ethanol exposure promotes elevated ethanol consumption later in life. Thus, this augmented ethanol consumption may be driven by fluctuations and instability in the dynorphin/KOR system or by compensatory mechanisms yet to be identified. To date, no studies have examined the effect of ethanol on dynorphin transmission in animals exposed to ethanol prenatally.

In summary, ethanol and the dynorphin/KOR system have a circular relationship. Facilitation of dynorphin transmission promotes ethanol drinking behaviors and excessive ethanol consumption potentiates dynorphin expression and therefore dynorphin transmission. Chronic ethanol exposure, however, at least in adults downregulates dynorphin possibly promoting other compensatory mechanisms. Dynorphin is indeed regulated by other neuromodulatory systems, such as the corticotropin releasing factor (M. R. Bruchas et al., 2010; Koob, 2015; Land et al., 2008) and brain derived neurotropic factor (Logrip et al., 2008). These systems play a crucial role in shaping dynorphin responses to ethanol and stress and perhaps the modulatory effects are ontogeny specific. In the next section, we address the role of motivation and dynorphin in the modulation of reward seeking behaviors.

4. Role of dynorphin in motivation regulation

Drug addiction literature revolves around dopamine neurotransmission in the mesolimbic areas because numerous studies over decades have shown supraphysiological elevations of dopamine levels in the striatum following drug, including ethanol, exposure (Volkow et al., 2009). Increase in dopamine levels in the ventral striatum/NAc results in activation of D1-MSNs. These recurring supraphysiological surges of dopamine and activation and the subsequent activation of D1-MSNs promote various neuroadaptations. One such adaptation is change in upregulation of dynorphin expression via activation of the transcriptional factor CREB (cAMP response element-binding protein) within the neurons of NAc (Muschamp and Carlezon, 2013). This increase in dynorphin results in inhibition of dopamine release via activation of KORs, which are localized on dopamine terminals in most terminal regions and on cell bodies in the VTA. Due to this overall reduction in dopamine transmission, there is a shift in homeostatic point of dopamine, promoting hyperkatifeia, which is defined as hypersensitivity to emotional distress and pain as well as anhedonia, defined as the inability to experience pleasures (Koob, 2019; Shurman et al., 2010).

In order to fully understand the dysregulated motivation processes, it is important to examine behavior and neurobiology in animals that demonstrate vulnerability to excessive drug use. Mice lacking dynorphin have been shown to consume significantly more ethanol compared to their wildtype counterparts (Rácz et al., 2013). It is interesting that dynorphin is not necessary for ethanol preference, in fact, dynorphin deficiency may promote ethanol self-administration; however, it is crucial to reinstate stress-induced ethanol drinking (Rácz et al., 2013). In this study, mice lacking dynorphin did not demonstrate stress-induced facilitation of ethanol consumption. Moreover, rats genetically predisposed to consume alcohol appear to have reduced dynorphin tone. For example, dynorphin levels were lower in alcohol preferring rats compared to the alcohol non-preferring rats in the NAc at baseline (Nylander et al., 1994). Similarly, Lewis rats, an inbred strain, that readily consume ethanol express less dynorphin compared to the Fischer 344 inbred rat strain (Nylander et al., 1995a). Conversely, ethanol-avoiding DBA/2 mice were reported to have greater dynorphin expression in the striatum compared to the ethanol-preferring C57BL/6j mice (Jamensky and Gianoulakis, 1997; Ploj et al., 2000). Perhaps animals with genetic modifications that either reduce or eliminate dynorphin, are highly motivated to take drugs because the feedback signal necessary to reduce drug-induced dopamine release, is dysregulated or lacking.

Animals exposed to chronic stress exhibit elevated alcohol consumption; for example, Rats and mice exposed to adolescent social isolation stress and repeated forced swim stress consume more ethanol compared to their stress naïve counterparts; an effect reversed by KOR blockade (Anderson et al., 2016a; Karkhanis et al., 2016b). Acute stress exposure, such as single exposure to social defeat, three-hours of immobilization stress, and one force swim stress exposure, elevates dynorphin expression in NAc shell and core, as well as the hippocampus for up to two hours following the cessation of stress exposure (Shirayama et al., 2004). This increase in dynorphin is transient, as the levels were different from those in stress naïve rats when examined five hours after stress exposure terminated (Shirayama et al., 2004). However, repeated exposure to stress or adversity attenuates prodynorphin mRNA levels and tissue content in the NAc (Donahue et al., 2015; Karkhanis et al., 2016b). Together, these studies indicate that in fact, low levels of dynorphin may increase vulnerability to develop maladaptive behaviors, such as, compulsive drug taking behaviors. In fact, it has been shown that exposure to chronic stress results in increased cocaine seeking (Brodnik et al., 2019; Groblewski et al., 2015) and ethanol consumption (Anderson et al., 2016a; Karkhanis et al., 2016b). This may in fact be due to altered reinforcing efficacy of ethanol and reward valency experienced by the animal as dynorphin is no longer as effective in regulating ethanol-induced reinforcement. There needs to be more work in order to elucidate the detailed mechanism involved here.

It is important to note that chronic ethanol exposure, voluntary or involuntary, as well as chronic stress exposure, both result in increased consumption of ethanol and the mechanisms driving these differences may in fact be bidirectional. Most of the studies examining dynorphin content or expression have reported changes during acute withdrawal, but no changes during protracted withdrawal. Although an indirect measure, ethanol mediated inhibition of dopamine in the NAc, which then reverses with KOR blockade (Karkhanis et al., 2016a), indicates that it is possible that the observed inhibition of dopamine is promoted at least in part due to a surge of ethanol-induced dynorphin release. This dynorphin release may occur from lateral hypothalamic neurons innervating the NAc. The stress response system is critically involved in driving addictive behaviors. It is known that chronic or repeated stress exposure attenuates dopamine tone, potentially due to hyperactive KORs, causing a hypodopaminergic state which promotes anhedonic and dysphoric behaviors. Thus, it increases vulnerability to substance and alcohol use disorder and facilitates relapse and reinstatement of drug use mostly via dysregulation of motivation-related neural processes. Stress and addiction are bidirectionally related, while stress promotes addictive and maladaptive behaviors, excessive drug consumption and cycles of withdrawal are stressful. In both situations, motivation and reward-seeking processes are dysregulated. While animals exposed to ethanol chronically may exhibit compulsive drinking to abolish negative reinforcement, it is possible that animals exposed to chronic stress may drink compulsively for the positive reinforcing values of the drug.

5. Sex differences in dynorphin function

A majority of the studies in the literature have been conducted in male animal models; some studies, however, have reported sex differences in the expression and function of dynorphin and KORs (see Chartoff and Mavrikaki, 2015 for review). Expression levels of dynorphin are sex-dependent (Becker et al., 2012). These disparities are based on differential chromosomal makeup and other variables. For example, one study has shown that the expression of prodynorphin mRNA is greater in the NAc and dorsal striatum of female mice with intact XX sex chromosomes compared to male XY mice and XO mice (Chen et al., 2009). Furthermore, expression levels of dynorphin fluctuate during the course of the estrous; while the expression of dynorphin A and B is reported to be stable throughout the estrous cycle in the NAc, expression of dynorphin A and B is observed to be lower in substantia nigra and dynorphin A is observed to be lower in caudate putamen during the estrous phase (Roman et al., 2006). Together, these studies indicate that sex differences may be involved in differential effects of KOR-activation on motivation and drug seeking observed in studies discussed in the following paragraph.

The greater levels of prodynorphin and dynorphin observed in females in areas involved in reinforcement and motivation processing may explain the heightened vulnerability of females to drug dependence. The idea here is that greater availability of dynorphin would potentially augmenting KOR activation, subsequently leading to stress-aversive effects and drug seeking. Indeed, human studies have shown evidence that stress-related disorders, including anxiety and mood disorders are twice as frequently diagnosed in females as males, and are more prone to developing drug dependence (World Health Organization, 2014). Though it is intuitive that greater dynorphin levels would enhance KOR-activation therefore resulting in greater stress-aversive states, preclinical literature demonstrates a different picture. Specifically, female rats are less sensitive than males to the KOR-activation-mediated increases in intra-cranial self-stimulation threshold of the medial forebrain bundle suggesting that females are less sensitive to depression-like effects mediated by KOR-activation (Russell et al., 2014). Activation of c-Fos following KOR activation with its agonist is observed to be greater in BNST of female compared to male rats (Russell et al., 2014). The BNST consists of a heterogeneous population of neurons; therefore, this non-specific increase in c-Fos activation is inconclusive. This same study has shown that KOR-activation augments c-Fos activation in the corticotropin releasing factor containing neurons in the paraventricular nucleus, suggesting that the enhanced depressive effects observed in females is likely driven by dynorphin/KOR system interaction with the corticotropin releasing factor system.

With respect to drug addiction and self-administration, prior activation of KORs in male rodents enhances rewarding and hyperactivity effects of cocaine and reinstates cocaine seeking (Chartoff et al., 2016; McLaughlin et al., 2006; Redila and Chavkin, 2008; Sershen et al., 1998); these effects were not observed in female mice (Sershen et al., 1998). Continuous activation of KORs with U50,488 in male rhesus monkeys augments cocaine choice (Negus, 2004). Together, these studies suggest that a KOR-mediated enhancement in relative reinforcing effects of cocaine across species. KOR modulation affects consumption of other drug classes. For example, KOR-inhibition increases ethanol intake in adult females while reducing it in adult males (Morales et al., 2014). Conversely, another study has shown a reduction in ethanol intake and preference following a combination treatment of KOR agonist Mesyl Salvanorin B or Nalfurafine in conjunction with Naltrexone, a mu opioid receptor antagonist with a lower affinity to KORs in male and female mice (Zhou et al., 2017; Zhou and Kreek, 2019a). Combination of Nalfurafine and Naltrexone also reduces relapse-like drinking in male and female mice (Zhou and Kreek, 2019b). In rodents exposed to chronic stress or ethanol, KOR-inhibition results in a reduction in ethanol intake (Anderson et al., 2016a; Karkhanis et al., 2016b; Rose et al., 2016). In the human literature, a PET imaging study has shown greater KOR availability in males compared to females in the anterior cingulate cortex, insula and the ventral pallidum among other areas (Vijay et al., 2016). A recent human study examining mechanisms via which naltrexone reduced alcohol intake shows that naltrexone-induced reduction in alcohol consumption is negatively correlated with availability of KORs in the cingulate cortex, pallidum, and the striatum (de Laat et al., 2019). Although this study enlisted male and female participants, they do not address sex differences.

Taken together, this collection of studies examining drug taking and seeking behaviors, along with other affective behaviors is rather inconclusive. Unpublished evidence from our laboratories suggests that there are individual differences within dynorphin/KOR system. These individual differences of dynorphin and KORs (both numbers and affinity) likely drive disparities observed in studies mentioned in this section. More research directly comparing males and females is crucial to advance our knowledge regarding sex differences in order to develop refined pharmacotherapies to treat neuropsychiatric disorders, including addiction.

Most of the studies presented in this review thus far have examined the KORs, therefore indirectly studying the effects of dynorphin on behavior and neurotransmission. Although extremely informative, studies that have measured tissue content or expression levels of dynorphin, prodynorphin, and their respective mRNA are not able to explain the effects of and on dynorphin to the fullest as they are not measuring online transmission of dynorphin. Methods enabling direct measure of dynorphin release and transmission are currently being developed.

6. Where do we stand with dynorphin detection?

Neuropeptides have proven very difficult to detect as they can be rapidly cleaved by peptidases and undergo posttranslational modifications. They are also found at much lower concentrations compared to classical neurotransmitters. Most of our knowledge on neuropeptide changes following perturbation in rodent models has been through by proxy looking at mRNA or using immunoaffinity-based techniques due to their high sensitivity (Kendrick, 1990; Maidment et al., 1991, 1989; Marinelli et al., 2005; Nieto et al., 2002; Nylander et al., 1995b), however selectivity remains an issue. In addition to all dynorphin isoforms deriving from prodynorphin, α- and β-neoendorphin, which are leu-enkephalin-based opioid peptides are also derived from prodynorphin (Chavkin et al, 1983, Kakidani et al, 1982, Weber et al, 1982, Zamir et al, 1984). This in turn makes selectivity a continuous issue.

At a number of different time points through history, groups have tried to measure opioid peptide changes in vivo with somewhat limited success. A number of groups were able to measure enkephalins using a combination of microdialysis and mass spectrometry but this was primarily in rat models (Baseski et al., 2005; DiFeliceantonio et al., 2012; Gobaille et al., 1994; Haskins et al., 2001; Lisi and Sluka, 2006; Mabrouk et al., 2011; Shen et al., 1997). At the time one of the major concerns was the damage created by the implantation of the microdialysis probe. Since then microdialysis probe size has decreased and laboratories have been working to optimize detection methods to allow for detection of opioid peptides at lower concentrations from smaller volumes (Zhou et al., 2015). In general, the field has been more successful in detecting enkephalins as opposed to dynorphin in rats versus mice for the following reasons.

Dynorphin has been described to come online following a perturbation/insult to the system, such as in the event of a stressful stimuli and is thought to act to regulate mood states (Bruchas et al., 2010). Furthermore, we know dynorphin exists in many different forms and fragments, yet we do not have a clear idea of the dynamics/role of these fragments in the context of different behavioral states. Interestingly, early work in the 1980’s studying the structural analysis of the dynorphin sequence gave some insight into how specific amino-acid residues may be responsible for KOR selectivity. For example, the strongly basic residues in the C-terminal domain of dynorphin, arginine-7, lysine-11 and lysine-13, seem to be critical for KOR selectivity in Dynorphin-A (Chavkin, 2013; Chavkin and Goldstein, 1981). Furthermore, it was shown that Dynorphin-B also has strong basic residues in similar positions that influence KOR selectivity (James and Goldstein, 1984). Dynorphin fragments, described as natural fragments, such as Dynorphin a (1–8), do not have the C-terminal lysines and as a result have lower KOR potency and selectivity (Chavkin and Goldstein, 1981; James and Goldstein, 1984). These findings are important and likely overlooked but should certainly be taken into account as the field pushes towards developing methods to detect dynorphin in vivo.

With the rise of optogenetic and chemogenetic tools, there has been an increased effort to use these tools, in conjunction with microdialysis/LC-MS detection, in mouse models to improve existing detection methods (Al-Hasani et al., 2018; Patterson et al., 2015). Most recently, a study showed increases in dynorphin release in two regions of the NAc shell following optogenetic stimulation of dynorphin cells using pdyn-cre transgenic mouse line (Al-Hasani et al., 2018). In addition to the observed changes in dynorphin release in vivo, both leu- and met-enkephalin were also measured, as well as dopamine, GABA and glutamate (Al-Hasani et al., 2018). The ability to optogenetically stimulate release of peptides allows for method development and optimization that would be difficult when trying to measure more subtle changes in peptides, particularly dynorphin, during chronic behavioral manipulations for example. Interestingly, a group developed a complementary approach that does not use a microdilaysis/LC/MS approach but uses multiple-scan-rate waveform to enable real-time voltammetric detection of tyrosine containing neuropeptides, specifically met-enkephalin allowing for more spaciotemporal resolution (Calhoun et al., 2018), which holds promise in its applicability detecting dynorphin.

The field is beginning to see an increase in alternative and complementary methods to not only detect in vivo release of dynorphin and opioid peptides but to learn more about their release dynamics and function. For example, photoactivable opioids were developed to achieve rapid and spatially delimited delivery of neuropeptides in mammalian brain tissue (Banghart and Sabatini, 2012). Analogs of [Leu⁵]-enkephalin and the 8 amino acid form of Dynorphin A (Dyn-8) were developed, which when exposed to ultraviolet light causes release of leu-enkephalin and dynorphin (Banghart and Sabatini, 2012).

With the improvement of genetically encoded sensors, such as dLight (Patriarchi et al., 2018), there has been a push to developing opioid receptor-based sensors to allow for the same robust, spatiotemporal resolution and detection of physiologically and behaviorally relevant changes in peptides as has been shown with dopamine (Patriarchi et al., 2018). Along a similar vein, earlier this year Neuropeptide Release Reporters (NPRRs) were developed using the Drosophila larval neuromuscular junction (NMJ) as a model (Ding et al., 2019). These are genetically-encoded sensors with high temporal resolution and genetic specificity to better inform us of trafficking and packaging of native neuropeptides as real-time changes in fluorescence intensity (Ding et al., 2019). In the future, this technology could perhaps be applied to detecting opioid peptides. In summary, though we still have a long way to go before we can reliably and consistently measure and distinguish all fragments of in vivo dynorphin release during acute and chronic behavioral manipulations, the field has made impressive strides in developing new technological approaches and the future of dynorphin and opioid peptide in vivo detection looks very promising.

7. Summary

In this review we have highlighted the integral role of dynorphin, the endogenous peptide for KOR, in AUD. This system is evidently extremely complex in modulation of AUD and its role seems very much dependent on the model of AUD, exposure paradigm and age of exposure. It is also becoming very clear that part of the limitation to our ability to interrogate dynorphin’s role in AUD is due to a lack of accessible techniques to measure real-time in vivo changes following behavioral manipulations and alcohol exposure. We are amidst an exciting time where multiple groups are working on a number of complementary methods that will allow the field to more readily measure changes in both opioid peptides and neuropeptides in general.