Addiction and depression are chronic relapsing disorders with devastating consequences for individuals and their social environment (1). Chronic exposure to drugs of abuse, as well as prolonged abstinence from these drugs, is strongly associated with lowered mood and a negative affective state. Conversely, in some individuals, depressed mood potently drives the consumption of euphoric psychoactive substances, a process referred to as self-medication. Accordingly, epidemiological studies have clearly demonstrated a marked comorbidity between addiction and depression (2, 3). This comorbidity is accompanied by greater functional disability, longer illness duration, less social competence, and higher service utilization. Therefore, understanding pathophysiological mechanisms underlying comorbidity has important therapeutic implications.

The present review will discuss numerous lines of evidence that have accumulated to document the kappa opioid receptor (KOR) as an important substrate in comorbidity between addictive and depressive disorders. The KOR belongs to the opioid system, a neuromodulatory system that is widely expressed throughout the central and peripheral nervous systems. The opioid system is composed of three G protein-coupled opioid receptors: mu (MOR), delta (DOR), and kappa (KOR), which under physiological conditions are activated by a family of endogenous peptides to inhibit neuronal activity. Among opioid peptides, dynorphins (encoded by the Pdyn gene) primarily activate the KOR and have very low affinity for MOR or DOR. Conversely, the other opioid peptides (endorphin and enkephalins) poorly interact with the KOR. Therefore, the dynorphin/KOR signaling pathway forms a distinct process within the opioid system (4, 5).

Opioid receptors tightly regulate motivational processes, and are identified as important players in psychiatric disorders characterized by reward dysfunction, such as addiction and depression (6, 7). Several exhaustive reviews have recently summarized data on the role of MOR and DOR in these disorders, and will be briefly mentioned when appropriate (6, 8–11). Our goal is to provide the reader with a historical and neuro-anatomical perspective on where, when, and how KORs are recruited in rodent models of addiction and stress-related psychopathology (12–16). First, we will summarize how the KOR progressively emerged as an anti-reward system that encodes dysphoria and limits motivational properties of drugs of abuse. Secondly, we will show that the KOR is recruited and activated during stressful experiences, thereby contributing to the emergence of depressive states (10, 17, 18). Finally, we will discuss two main aspects of how these roles of KOR in addiction, stress-related behaviors and depression have important implications for the understanding of comorbidity. On one hand, we will show that stress-induced recruitment of KOR signaling is a potent trigger of drug-seeking and relapsing behaviors. On the other hand, we will summarize data on KOR function in the particular context of depressive-like behaviors that emerge during chronic exposure to drugs of abuse, as well as during drug abstinence (19–21).

As a pathophysiological substrate of comorbidity, the KOR represents a promising pharmacological target (10, 18). Clinical trials are currently on-going to assess KOR antagonists as a treatment for depression (22), in particular in the context of addicted patients suffering from comorbid depressive conditions (23, 24). Building on rodent studies, we will discuss the potential of therapeutic strategies targeting the KOR.

Interest in KOR pharmacology historically stemmed from the hope of developing analgesic compounds devoid of the classical abuse potential of MOR agonists, such as morphine. Unfortunately, early human studies exploring properties of KOR agonists reported potent dysphoric and psychomimetic effects (25, 26). While these results clearly decreased the therapeutic potential of KOR in the treatment of pain, they also urged preclinical researchers to explore these intriguing dysphoric effects.

Activation of the MOR is known to induce euphoria in human, and to produce reinforcement in animal models. Therefore, researchers hypothesized that MOR and KOR may have opposite effects in the regulation of motivational processes, potentially through the modulation of common neuronal pathways. This framework was initially explored using conditioned place preference (CPP) or conditioned place avoidance (CPA). In this Pavlovian conditioning paradigm, a drug is repeatedly paired with a set of environmental stimuli that progressively acquire positive (CPP) or negative (CPA) motivational properties. Following repeated conditioning sessions, the animal subsequently exhibits preference or avoidance on re-exposure to the environmental stimuli (in the absence of the drug), a behavior that depends on learning, motivational, and hedonic mechanisms. The seminal rat study by Shippenberg and Hertz (27) reported that, as hypothesized, systemic administration of the KOR agonist U69593 or morphine yielded opposite effects, respectively producing CPA and CPP. While morphine-induced CPP reflects its reinforcing properties, KOR-induced CPA suggested that this receptor might be an anti-reward mechanism that contributes to a bi-directional regulation of motivation and hedonic tone.

The next step was to investigate underlying neurochemical substrates, with early studies exploring how the KOR may regulate the mesolimbic pathway (Figure 1). This pathway is composed of dopaminergic (DA) neurons that are located in the midbrain ventral tegmental area (VTA) and project to forebrain limbic structures, including the ventral striatum [or nucleus accumbens (NAc)] and prefrontal cortex (PFC). Animal and human data have clearly demonstrated that drug addiction (and mood disorders, see Part 2) associate with major disruptions of the brain’s DA reward circuitry (28), which normally acts to predict and encode the salience of environmental stimuli and natural rewards. The now classical “unitary” theory of addiction postulates that essentially all drugs of abuse enhance DA transmission in the NAc, an effect that is central to their rewarding properties (29). Within this line, many studies have consistently shown that acute reinforcing effects of morphine rely on disinhibition, i.e., the activation of DA neurons. This disinhibition occurs through the activation of MOR expressed by GABAergic interneurons located mainly in the tail of the VTA [tVTA, or RMTg, see in Ref. (30, 31)], but also in the VTA and NAc (32). In contrast, decreased DA signaling was hypothesized to be responsible for the encoding of KOR-mediated aversion. Using microdialysis, Spanagel and colleagues (33) showed that DA release in the NAc was decreased by infusion of a KOR agonist into the NAc, but not into the VTA (pharmacological agents used in every study discussed in the present review are summarized in Tables 1–3). In addition, infusion of a KOR antagonist (nor-BNI) into the NAc increased DA release, suggesting that dynorphins tonically reduce DA neurotransmission in this region. The most compelling evidence implicating DA neurons in KOR-induced aversion came recently (12) from the use of genetically modified mice using the Cre-lox recombination system (34). Bals-Kubik et al. took advantage of a knockin mouse expressing the Cre-recombinase under transcriptional control of the endogenous promoter of the DA transporter [DAT, a specific marker of DA neurons (35)]. These mice were bred with another knockin mouse harboring a conditional “floxed” KOR allele, thereby achieving the specific deletion of KOR in DA neurons (DAT KOR-cKO). At the behavioral level, KOR-induced CPA was abolished in DAT KOR-cKO mice, and restored upon virally mediated KOR re-expression in the VTA (12).

In parallel, investigators undertook a brain-wide analysis of regions where recruitment of the KOR may potentially encode aversion. The effect of local KOR activation was assessed in several areas using the CPA paradigm (49). Infusion of the KOR agonist U50,488H in the NAc was sufficient to induce a robust CPA, consistent with the notion that KOR activation in this region decreases DA release. Surprisingly, infusions in the PFC, lateral hypothalamus, and VTA (but not in the substantia nigra and dorsal striatum) had similar effects, suggesting that multiple KOR pools may regulate motivation and hedonic tone. These results also indicate that activation of VTA KOR induces CPA in the absence of any change in NAc DA release [see aforementioned neurochemical data (33)], implying the involvement of another brain region receiving DA innervation (i.e., the PFC, see below). In addition to the regulation of DA transmission, KOR expression and function is now under investigation in many other brain regions using rodent assays relevant to reward and mood [e.g., bed nucleus of the stria terminalis, BNST, amygdala, locus coeruleus (LC), see below].

An important next goal was to identify which neuronal cell types are controlled by the KOR. Electron microscopy approaches (50, 51) found that in the NAc, half of the axons that were KOR-immunoreactive also expressed DAT. Interestingly, this study found that almost one-third (29%) of these KOR-immunoreactive axons were DAT-negative but contacted pre-synaptic terminals of DAT-positive neurons, suggesting that the mesolimbic pathway is regulated at the level of the NAc by afferent neurons expressing the KOR. Based on recent evidence (16), it is likely that non-DAergic KOR-positive neurons are, at least in part, serotonergic (5-HT). Possibly, KORs expressed by 5-HT neurons may mediate DA/5-HT crosstalk in the NAc, and represent a mechanism that contributes to interactions between mood and reward, as well as between addiction and depression (see below). At the level of the NAc (52), there is also evidence for KOR-dependent modulation of glutamate release, suggesting that this receptor may be expressed pre-synaptically by glutamatergic cortical neurons that densely innervate the NAc. To our knowledge, the behavioral relevance of the latter KOR pool has not been addressed. Finally, in the PFC (53), the KOR was mainly located on pre-synaptic terminals, likely corresponding to DAergic inputs, although the neurochemical identity of these neurons was not assessed.

Other investigators used electrophysiology and immunohistochemistry to identify KOR-expressing neurons. Application of a KOR-selective agonist in the VTA decreased spontaneous firing activity of a sub-group of neurons (54). This KOR-mediated inhibition only occurred in DA cells, as indicated by immunoreactivity for tyrosine hydroxylase (the rate limiting enzyme for DA synthesis, and another marker of DA neurons). Electrophysiology, retrograde tracing and microdialysis were then combined to assess whether DA neurons projecting either to the NAc or to the PFC are differentially regulated by the KOR (55). These elegant experiments revealed that local KOR activation in the VTA hyperpolarized PFC-targeting DA neurons, but had no effect on NAc-targeting DA neurons. Accordingly, DA release was reduced in the PFC, but not in the NAc, upon VTA KOR activation.

Overall these results are consistent with previous CPA and microdialysis studies, and suggest a model whereby VTA KORs do not control NAc DA tone but rather modulate DA release in the PFC to produce CPA. The previously described DAT KOR-cKO mice recently provided strong evidence for the latter hypothesis. Tejeda et al. found that infusion of a KOR agonist in the PFC decreased DA overflow in wildtype (WT) but not in DAT KOR-cKO mice (13), confirming KOR-mediated control of DA transmission in the PFC. Importantly, the authors then directly tested the behavioral relevance of PFC KORs for dysphoria in rats. Infusion of a KOR antagonist into the PFC was sufficient to prevent KOR agonist-induced CPA, clearly identifying the limbic cortex as a necessary substrate for this behavioral effect.

Collectively, results from these various methodological approaches also suggest that NAc-projecting DA neurons express KOR in pre-synaptic terminals, but not in soma and dendrites (33), while PFC-projecting DA neurons express KOR in both compartments (13, 54, 55) (Figure 1). At the molecular level, it is currently unknown how DA neurons may control KOR trafficking to distinct cellular compartments as a function of their projection targets. We speculate that the type of electrophysiological feedback (excitatory from the cortex, inhibitory from striatal medium spiny neurons) provided to the VTA by each region may be implicated. Alternatively, cell-autonomous processes might be involved, with distinct transcriptomic profiles in NAc- and PFC-projecting DA neurons leading to distinct KOR post-translational modifications and trafficking. To experimentally address the latter hypothesis, technological advances now allow researchers to distinguish transcriptomes from neuronal populations sharing a common cell-body location but with distinct projections (56). Alternatively, retrograde tracing may be coupled with knockin reporter mice expressing opioid receptors in fusion with fluorescent proteins [such mice are currently available for mu and delta, but not kappa, opioid receptors (57, 58)].

Activity of the dynorphin/KOR pathway on DA neurotransmission and in CPA has obvious implications for addiction-related behaviors, as observed for a variety of drugs of abuse in self-administration paradigms [see in Ref. (59) for an exhaustive review]. KOR agonists dose-dependently decrease morphine self-administration in rats and mice (60, 61). Similar inhibitory effects of KOR activation were found for ethanol (62–64), cocaine (61, 65–68), nicotine (69), and cannabis (70), and these were associated with reduced drug-induced DA release (for example, cocaine, see in Ref. (71); ethanol, see in Ref. (72)). Globally, these results provide robust evidence for an inhibitory effect of KOR on the rewarding effects of drugs of abuse, and recent findings suggest that natural rewards, such as social interactions, may also be affected. In prairie voles, a monogamous rodent species, maintenance of mating pair bonds relies on the expression of aggressive behaviors toward novel conspecifics. Interestingly, this form of “social aversion” has been shown to be mediated by KOR signaling within the NAc (73). In rodents, social play represents a highly studied, naturally occurring behavior that recruits DA neurons and triggers potent reinforcement. Systemic activation of KOR decreased social play in both rats (74, 75) and mice (76). These findings are relevant to our understanding of depression in human as anhedonia, or the altered perception of rewarding properties of everyday-life stimuli (including social interactions), is a hallmark of this condition. Therefore, while KOR-dependent modulation of DA and reward was initially conceptualized and explored in addiction paradigms, it is now becoming clear that it also has strong implications for mood disorders (17, 18). CPA reflects the interaction of several neurobiological mechanisms, corresponding to three psychological constructs: learning, motivation, and hedonia. Intra-cranial self-stimulation (ICSS) is another paradigm assessing these three aspects: in this operant conditioning, animals learn to self-administer brief electrical pulses into specific brain regions (usually the medial forebrain bundle (77)). Systemic activation of KOR was found to induce an anhedonic-like state in ICSS, as indicated by increased stimulation threshold (78). In the latter work, the stimulation electrode was placed in the lateral hypothalamus, strengthening previous evidence for KOR-dependent regulation of hedonic state occurring outside the NAc (49). Further, Potter et al studied the kinetics of KOR agonist-induced ICSS modulation following acute and repeated injections (79). The KOR agonist Salvinorin-A increased the stimulation threshold, and this acute effect persisted with daily injections over an 8-day period. Interestingly, repeated injections also triggered delayed and opposite effects, as evidenced by decreased ICSS stimulation threshold 24 h post-injection, suggesting that opponent processes (80, 81) had developed.

The neuronal pathway potentially linking hypothalamic KOR activity with DA transmission and reward has been poorly studied. Recent elegant data using electron microscopy, electrophysiology, ICSS, and cocaine self-administration (47, 48), suggest an antagonistic interplay between orexin and dynorphin peptidergic systems. The hypocretin/orexin system is composed of neurons originating in the lateral hypothalamus and projecting to several mesolimbic structures (82). Importantly, orexin and dynorphin were found to act as co-transmitters in neurons of the hypothalamus (47): the two peptides co-localize in synaptic vesicles, and are co-released upon electrical hypothalamic stimulation. The authors further showed that orexin and dynorphin act within the VTA to stimulate and inhibit, respectively, the excitability of DA neurons, thereby bi-directionally modulating reward (in ICSS experiments), and self-administration of cocaine (and potentially other drugs of abuse). In the VTA, most cells (65%) were found to be common targets for both orexin and dynorphin. Based on previous evidence, future experiments may test the hypothesis that VTA DA neurons expressing both KOR and orexin receptors project preferentially to the PFC rather than the NAc.

Overall, data on KOR function in the regulation of reward highlights the importance of assessing the full spectrum of peptides and neurotransmitters expressed along the mesolimbic pathway and associated neuronal circuits. Determining how this network dynamically evolves under chronic pathologic conditions will be an exciting endeavor.

In parallel to these studies on reward, recent data have demonstrated that the KOR also controls emotional responses, in particular during stressful experiences. Pharmacological studies in rodents indicate that the dynorphin/KOR system regulates mood-related behaviors. In rats, systemic administration of KOR agonists and antagonists showed pro- and antidepressant-like effects, respectively, in the forced swim (FS) and learned helplessness (LH) tests [see in Ref. (10, 41–45) for a review]. Consistent with CPA studies, systemic KOR activation decreased DA release in ventral (44, 83), dorsal (84, 85), and striatal regions, while local NAc injection of a KOR agonist mimicked the prodepressant-like effect of systemic treatment (86). These data further confirm that KOR-dependent modulation of DA is implicated in both mood- and addiction-related behaviors (28). Interestingly, KOR-dependent prodepressant-like effects may be modulated by gender (87), an important aspect considering that the prevalence of depression is higher in women. Using ICSS, the authors found that the KOR agonist-induced increase in ICSS stimulation threshold was higher in male than female rats. This effect was independent from circulating levels of gonadal hormones, and was not accounted for by sex differences in pharmacokinetics of the agonist. Rather, sex differences in KOR agonist-induced neuronal activation, as revealed by c-fos staining, were found in the BNST and PVN, but not in the NAc or amygdala. Therefore, in addition to the mesolimbic pathway, sex-specific KOR-dependent regulation of hedonic tone may also occur at the level of the BNST and PVN, two structures controlling stress-responses and emotions.

Adding to pharmacological studies targeting KOR, there is also evidence that dynorphins provide an endogenous tonic regulation of mood-related traits (43, 88). In the NAc, medium spiny neurons expressing the DA D1-receptor are known to synthesize and release dynorphins under the control of the cAMP response element binding protein (CREB). Accordingly, prodynorphin levels were decreased in the NAc of transgenic mice overexpressing a dominant negative form of CREB. This effect was associated with decreased behavioral despair in the LH paradigm. Consistently, a recent study reported that Pdyn knock-down (by viral expression in the NAc of an anti-Pdyn short hairpin RNA) decreased depressive-like behavior in the FS test (46).

Beyond baseline emotional responses, data indicate that activity of the dynorphin/KOR system is potentiated by stress. Acute, but not chronic, restraint stress was shown to sensitize KOR-dependent CPA (89). Also, repeated exposure to FS stress produced a prodepressant-like effect that was blocked by the KOR antagonist nor-BNI, and was absent in Pdyn KO mice (90). Dynorphins were further demonstrated to modulate repeated stress-dependent aversive conditioning (37). Mice trained to associate a given odor with FS stress robustly avoided that odor. This avoidance behavior was not observed in Pdyn KO mice, and was blocked in WT mice by pre-treatment with a KOR antagonist. Similarly, a context repeatedly paired with footshocks was aversive in WT mice; but again, this effect was absent in Pdyn KO mice, and prevented by KOR antagonist pre-treatment. Importantly, the authors showed that corticotropin releasing factor (CRF) release in the central nervous system is likely the primary event responsible for stress-induced recruitment of the dynorphin/KOR system. Results indicated that systemic injection of CRF triggered KOR phosphorylation, as revealed using a phospho-KOR antibody. Further, stress-induced CPA (mimicked by systemic or intracerebroventricular injection of CRF or the CRF₂-receptor agonist Urocortin III) was absent in Pdyn KO mice, and blocked by nor-BNI pre-treatment. Following stress exposure, KOR activation, and phosphorylation was identified in several brain structures, including the basolateral amygdala, hippocampus, dorsal raphe, VTA, and NAc. Altogether, these data show that dysphoric aspects of stress behaviorally manifest when CRF stimulates dynorphin release, yielding KOR activation (37).

Stress is a complex physiological process that has a primarily adaptive value, but that can trigger pathological events during prolonged and excessive stressful experiences. Recently, interactions between stress and the KOR have been investigated using more sophisticated and ethologically relevant models of depression. In nature, confrontation among conspecific animals potentially generates significant consequences in terms of control over resources, access to mates, and social positions. For example, the resident–intruder social defeat paradigm (91) is a naturalistic model characterized by potent aggressive interactions that are unpredictable and inescapable, thereby inducing several anhedonia-like symptoms such as diminished sexual pursuit and decreased sucrose preference (52). McLaughlin and colleagues were the first to reveal the role of dynorphins and KOR in transducing the effects of social stress (92). Mice exposed to repeated social defeats over 3 days showed a characteristic defeated postural response, as well as an increased nociceptive threshold, or stress-induced analgesia (SIA, observed in a tail withdrawal latency assay). Both aspects were prevented in mice pre-treated with a KOR antagonist, or lacking the Pdyn gene. Another important feature of the social defeat model is that its effects show high inter-individual variability, both in rats and among inbred mice, such that animals can be typically separated into susceptible and resilient groups (93). Along this line, according to Bérubé et al. (94), expression levels of dynorphins in the NAc differ among susceptible and resilient Sprague–Dawley rats. Increased dynorphin mRNA levels (measured by qPCR) were found in the ventral striatum of susceptible rats (NAc shell, coherent with previous mice data), while surprisingly increased levels were observed in the dorsal striatum of resilient individuals, suggesting that the regulation of DA and mood by dynorphin and KOR may be more complex than anticipated. Additional studies will be necessary to further substantiate this hypothesis. In contrast, another study reported no change in dynorphin levels in VTA or NAc of socially defeated Long–Evans rats (95). Discordance between these two studies might be explained by the different strains used, or the absence of a distinction between resilient and susceptible Long–Evans rats in the latter study. Of note, Nocjar et al. (95) found decreased dynorphin-A, as well as decreased orexins A and B, in the hypothalamus of defeated rats. Therefore, combined regulation of VTA DA neurons activity by these two antagonistic peptides might mediate defeat-induced KOR-dependent social aversion, and be impaired following social defeat.

We previously discussed (Part 1) how the KOR may display differential cellular localization across the two populations of VTA DA neurons projecting to the NAc or to the PFC. A recent report suggests that this anatomical dissociation may have relevance for the understanding of the effects of chronic social defeat. Chaudhury and colleagues (96) showed that the selective inhibition of VTA DA neurons projecting either to the NAc or to the PFC, respectively, promoted resilience or susceptibility to repeated social defeat. Due to its selective cellular localization, it is tempting to speculate that the KOR may mediate prodepressant-like symptoms induced by the inhibition of the VTA-PFC DA pathway.

In addition to DAergic signaling, new findings suggest that 5-HT transmission may also be modulated by KOR in stress- and social defeat-based models of depression. Electrophysiology experiments (97, 98) initially demonstrated that the KOR regulates 5-HT neurons at the level of the dorsal raphe nucleus (DRN), a main 5-HT brain nucleus. Importantly, rescue experiments showed that the selective re-expression of KOR in the DRN of KOR KO mice is sufficient to restore the CPA induced by infusion of a KOR agonist in the NAc (38). Together with previous findings, these results indicate that KOR in the PFC, and KOR expressed by neurons present in the DRN, which target the NAc (that are likely to be 5-HT neurons), are necessary and sufficient, respectively, for the expression of KOR agonist-induced aversion. At the molecular level, acute social defeat was shown to trigger phosphorylation of KOR and the p38α kinase in the DRN (14). Recruitment of p38α in 5-HT neurons is essential, as defeat-induced social avoidance was abolished in cKO mice in which p38α is specifically deleted from serotonin transporter (SERT)-expressing neurons (p38α-cKO^SERT). Phosphorylated p38α in turn promotes SERT translocation to the plasma membrane, thereby increasing 5-HT reuptake and likely mediating social avoidance. Electrophysiological recordings in brain slices (40) also showed that KOR activation dampens excitability of DRN 5-HT neurons through two mechanisms: the pre-synaptic inhibition of glutamatergic inputs, and the post-synaptic stimulation of G-protein-gated inwardly rectifying potassium channels (GIRKs). Repeated exposure to FS stress impairs post-synaptic, but not pre-synaptic, effects of KOR activation. Importantly, stress-induced inhibition of KOR-mediated GIRK currents was abolished in p38α-cKO^SERT mice. Finally, recent evidence suggests that KOR regulation of DA and 5-HT neurons may converge at the level of the NAc to produce dysphoric and depressive-like effects. Repeated FS stress selectively increased cell-surface expression of SERT in the ventral striatum, but not in other regions examined (dorsal striatum, hippocampus, PFC, amygdala, or DRN). This effect of stress on SERT was prevented by pharmacological blockade of KOR signaling in the NAc, but not in the DRN (16). Altogether, stressful experiences appear to recruit a CRF-dynorhin-KOR-p38α-GIRK signaling cascade within DRN 5-HT neurons, as well as KOR activation in the NAc. These molecular adaptations in turn lead to up-regulation of SERT function in the NAc, and ultimately affect DA function to produce behavioral symptoms. Whether similar DRN signaling is also involved in more prolonged mood-related deficits, in particular in the context of chronic exposure to drugs of abuse (Part 3), has yet to be determined.

In addition to 5-HT and DA circuits, other possible sites of KOR-dependent mood regulation notably include hippocampal neurogenesis and noradrenergic (NA) transmission. One report found that in rats, the antidepressant-like effect of the KOR antagonist nor-BNI (99) associated in the hippocampus, as well as in other structures (e.g., frontal cortex, amygdala, hippocampus, and endopiriform cortex), with increased mRNA levels of BDNF, a neurotrophic factor controlling synaptic plasticity and neurogenesis. Further studies are required to better understand the relevance of this KOR/BDNF interaction.

We have summarized the role of KOR in the regulation of reward processes (Part 1), and in the modulation of stress-responses and affective states (Part 2). Based on these data, several groups have recently explored how the KOR may mediate the interplay between addiction and depression. The relationship between these two disorders is likely bi-directional: addicts show a strong lifetime risk for anxiety or depressive disorders, while, conversely, depressed patients frequently abuse drugs to self-medicate their depressive symptoms. Both aspects are currently being addressed in animal models.

A major challenge in the future will be to unravel dynamic adaptations of the endogenous dynorphin/KOR system as mood and reward disruption emerge and evolve. This issue is of significant clinical relevance considering the chronicity of these two conditions. In particular, available evidence indicates that the KOR exerts multiple controls over the main monoamines in rodents. Interestingly, addiction research suggests that repeated exposure to drugs of abuse disrupts mutual inhibitory feedback mechanisms between monoaminergic nuclei, which may mediate long-term behavioral dysfunction (163, 164). Whether such mechanisms also impair KOR-dependent mood regulation is an intriguing hypothesis in the context of comorbidity.

Accumulating evidence in the KOR field has recently prompted clinicians to undertake brain imaging studies and clinical trials (22). Very recently, the first PET-Scan study using a radioactive KOR antagonist was able to demonstrate significant and widespread disruption of KOR in vivo availability in subjects suffering from fear and dysphoric symptoms following severe trauma exposure (165). While results are nicely consistent with animal data on KOR and the mesolimbic pathway, they also suggest that other brain regions, currently poorly explored in preclinical settings, may be equally important (e.g., thalamus and insular cortex). Additional studies will be required to further assess KOR availability in well-characterized cohorts of depressed, addicted, and comorbid subjects. Finally, from a pharmacological point of view, the rapidly evolving field of biased agonism (or ligand-directed signaling) raises great hopes for KOR-targeting therapeutics (149). A major goal in the field of G-protein-coupled receptors is the identification of distinct signaling pathways that may operate to control specific behavioral responses. In the near future, such approaches will likely aid in the development of antidepressants acting as KOR antagonists and devoid of potentially associated adverse effects (e.g., hyperalgesia).

In conclusion, we have summarized in the present review the large body of evidence supporting the role of KOR in regulating reward and mood. We have also described how this receptor is ideally placed to mediate strong interactions between two frequent and severe psychiatric disorders, addiction, and depression. Altogether, preclinical research on the KOR exemplifies how transversal studies across multiple animal models have the potential to identify brain mechanisms that contribute to transdiagnostic pathophysiological processes, and therefore represent key therapeutic targets for the management of comorbidity, one of the most prominent global issues in mental health.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.