Opioid addiction and overdose are at record levels in the United States. This is driven, in part, by their widespread prescription for the treatment of pain, which also increased opportunity for diversion by sensation-seeking users. Despite considerable research on the neurobiology of addiction, treatment options for opioid abuse remain limited. Mood disorders, particularly depression, are often comorbid with both pain disorders and opioid abuse. The endogenous opioid system, a complex neuromodulatory system, sits at the neurobiological convergence point of these three comorbid disease states. We review evidence for dysregulation of the endogenous opioid system as a mechanism for the development of opioid addiction and/or mood disorder. Specifically, individual differences in opioid system function may underlie differences in vulnerability to opioid addiction and mood disorders. We also review novel research, which promises to provide more detailed understanding of individual differences in endogenous opioid neurobiology and its contribution to opioid addiction susceptibility.
Keywords: affect, endogenous opioid, depression, opiate, painOpiate drugs have been a part of the human experience since the Neolithic period (Merlin 2003), with cultivation of the opium poppy and religious, medicinal, and recreational use of its resin by nearly every culture throughout history (Brownstein 1993, Schiff 2002). However, many of the root causes of the current opiate crisis are uniquely modern. Beginning with the isolation of morphine from opium in 1804 (Schmitz 1985), the synthesis of thousands of opioid compounds throughout the twentieth century has led to an enormous field of drugs with diverse pharmacological characteristics. In the late nineteenth and early twentieth centuries, there was no regulation of morphine and other opiates. For example, the Sears catalog sold heroin and morphine preparations by mail order (Inciardi 1986). This led to widespread dependence and addiction, triggering massive regulation of narcotics and a sentiment of opiophobia in the medical community, which persisted through most of the twentieth century (Morgan 1985).
In the last decades of the twentieth century, the pendulum of public and medical opinion on the medical use of opioids began to swing in the opposite direction. Multiple factors led to a massive increase in prescription and consumption of opioids in the United States. These included the wide array of available pharmaceuticals, complaints of undertreatment of pain (Max 1990), the widespread and inaccurate belief that use of opiates for pain rarely culminates in addiction (Porter & Jick 1980), the successful but fraudulent marketing of OxyContin as having ‘greatly reduced addiction potential’ (Van Zee 2009), and the American Pain Society’s 1995 recommendation to record pain as “the fifth vital sign” and eliminate it using opioid analgesics (Campbell 1996) This increase in the number of people exposed to opioids, often through prescriptions, coupled with the increase in very potent opioid drugs, has led to a new wave of opioid use disorder (OUD), which has recently been characterized as a national public health emergency (Jones et al. 2018).
We should note, however, that only a small percentage of those taking opioids for chronic pain convert to addiction (Vowles et al. 2015). Moreover, treatment for pain is by no means the only path to OUD. Temperamental and personality traits have been implicated in the propensity to seek drugs in general and opioids in particular (Amirabadi et al. 2015, Milivojevic et al. 2012). Although the majority of heroin addicts report beginning with a prescription drug (Am. Soc. Addict. Med. 2016), it is unclear whether these were legitimately prescribed to them, and prescription opioids are often diverted by others who are not suffering from physical pain (Ford et al. 2019). It is critical, therefore, to consider the heterogeneity of paths to opioid addiction as we develop strategies for fundamental research into the treatment and prevention of this devastating disorder.
Historically, our understanding of the mechanisms of opiate drug action lagged behind the proliferation of pharmaceutical compounds. However, thanks to advances in molecular biology and neuroscience, there has been an explosion of knowledge since we last reviewed the opioid field for this journal (Akil et al. 1984). This includes cloning and crystal structure of the opioid receptors, elucidation of structure and function of their peptide ligands, and elaboration of the functional neuroanatomy of the opioid system. While many gaps remain in our understanding of this complex system, the knowledge that has been gained provides insights into the underlying biology of addiction and associated disorders. However, this understanding has not yet translated into meaningful strategies for treatment or prevention of OUD. This lack of success suggests that some critical factors are not being considered. We suggest that these factors involve unique biological and psychological characteristics of individuals exposed to opioids that impact propensity to transition from use to misuse and addiction. We also suggest that these differences result in different biological responses to the drugs, and these distinct consequences require different treatment strategies. Thus, OUD demands a precision medicine approach, which strives to tailor treatment based on unique features of each patient’s disease process, particularly individual differences in opioid system biology.
In this review, we summarize the current literature regarding the biology of the endogenous opioid system, in particular the highly interdependent nature of its roles in pain, depression, and addiction. We then examine how this knowledge informs the etiology and treatment of OUD, as well as discussing promising areas of research that can move the field toward a precision medicine approach.
The opioid system is a highly diverse peptide neurotransmitter system, composed of three main neuropeptide families, the β-endorphins derived from pro-opiomelanocortin, the enkephalins derived from proenkephalin and prodynorphin (PDYN) precursors, and the dynorphins derived from PDYN. Each family contains multiple member peptides with diverse binding characteristics (Mansour et al. 1995) ( Figure 1a ).
(a) The three canonical opioid receptors demonstrate relatively similar degrees of affinity for various cleavage products of the endogenous opioid propeptides. The kappa opioid receptor (KOR) displays the highest degree of ligand specificity, with high degrees of affinity for prodynorphin (PDYN) peptides and very low affinity for β-endorphin or Leu/Met-Enk. Note that despite extremely low affinity for Leu/Met-Enk, the KOR shows high degrees of affinity for other proenkephalin (PENK) cleavage products. The biological relevance of this remains inconclusive, as the degree to which these peptides are produced from proenkephalin (PENK) and the conditions under which they are released in the brain remain understudied but raise the possibility that KOR-expressing regions are responsive to PENK-producing cells. Likewise, all three receptors show comparable binding affinity for PDYN peptides, indicating that cells expressing any opioid receptor type are likely to respond to PDYN release. Binding affinities data (indicated in nanomolar concentrations) in panel a are from Mansour et al. (1995). (b) There are high degrees of overlapping expression between the three canonical opioid receptors in a variety of brain regions established to be relevant to pain, mood/affect, and reward/motivation. When considered with the relative nonselectivity between receptors and signal peptides (as illustrated in panel a), it becomes apparent that any alteration in the volume and/or specific make-up of opioid peptide release is capable of complex, difficult-to-anticipate effects on a wide variety of neurobiological and behavioral outputs, especially regarding the modulation of pain, affect, and reward, often concurrently. Figure adapted from Mansour et al. (1988).
The system also includes three canonical receptors, the mu, delta, and kappa opioid receptors. A fourth receptor-peptide pair, nociceptin and its receptor, the nociceptin receptor or opioid-receptor-like 1, is also part of the opioid system. However, this pair was only recently discovered in comparison to the other three (Meunier et al. 1995, Mollereau et al. 1994). Their role in pain, affect, and addiction are still under investigation and will not be reviewed, but note that contemporary research indicates a role in pain and motivation (Di Cesare Mannelli et al. 2015, Kiguchi et al. 2016, Parker et al. 2019).
All three canonical receptors are members of the G protein–coupled receptor family, coupling to inhibitory Gi/oα subunits. The importance of the opioid system is underscored by its evolutionary age, appearing in essentially modern form in all jawed vertebrates and thus already established in a common ancestor over 450 million years ago (Dreborg et al. 2008). Furthermore, opioid receptors are among the most widely expressed receptors in the brain, as they are found throughout the neocortex, hippocampus, striatum, thalamus, hypothalamus, substantia nigra, and several nuclei of the brainstem and have widespread expression in the spinal cord and peripheral neurons.
The peptides have historically been portrayed to exhibit selectivity for particular receptors; however, it is important to recognize that the opioid peptides are fairly promiscuous, particularly the dynorphins. There is not a 1:1 correspondence between a signal peptide and receptor; members of all three peptide families are capable of activating the three receptors to varying degrees ( Figure 1a ). The opioid receptors also show a high degree of overlap in their regional expression patterns, particularly in regions relevant to pain, affect, and reward (Mansour et al. 1988) ( Figure 1b ). This results in a system whereby all the opioid receptor types within a brain region can be simultaneously activated by different peptides or even by the same peptide.
In addition to G protein–dependent signaling pathways, the opioid receptors signal via a wide assortment of second- and third-messenger systems (Al-Hasani & Bruchas 2011) and exhibit ligand-directed signaling (Pradhan et al. 2012). Importantly, although biased signaling is mainly studied with regard to divergent responses to exogenous drugs, it appears to be an intrinsic feature of the natural function of the system, enabling nuanced receptor responses to different signal peptide variants. For example, three PDYN products, dynorphin A, dynorphin B, and α-neoendorphin, display highly divergent capacities to desensitize and internalize the kappa receptor despite similar binding affinity, receptor occupancy, and intrinsic efficacy to engage G protein activity (Chen et al. 2007).
An additional layer of complexity is the existence of splice variants of the three opioid receptors (Pasternak & Pan 2013, Piltonen et al. 2019, Wei et al. 2000), which show altered affinity for the peptide ligands (and exogenous opioids) and potentially different biases for G protein–dependent versus –independent signaling (Margolis et al. 2017, Pasternak 2004). The activities of the opioid receptors are also influenced by positive and negative allosteric modulators (Livingston & Traynor 2018), which include other endogenous neurotransmitters/neurohormones (Kathmann et al. 2006, Meguro et al. 2018). Thus, opioid signaling can be tuned and refined based on concurrent activity of other systems.
The endogenous opioid system interfaces with a number of other key neurotransmitter systems, including the endocannabinoid system, the serotonergic system, the oxytocin and vasopressin systems, and the glucocorticoid axis (Bicknell 1985, Diniz et al. 2018, Pfeiffer & Herz 1984, Robledo et al. 2008, Tao & Auerbach 2005, Watson et al. 1982, Young et al. 1986). This interplay between the opioid system and other pathways warrants a separate review, and we have included here only a handful of examples. For a more comprehensive look at the neurobiology of the endogenous opioid system, we refer the reader to Gutstein & Akil (2005).
The opioid system is a highly integrated system, which depends on fine-tuned balance, with many mechanisms for compensation and regulation (e.g., Cawley et al. 2016). In most cases, this complexity is highly adaptive, permitting fine-tuned control over the circuitry modulated by the opioid system. But it also makes the system vulnerable to disturbances, and given its broad influence, this can lead to significant and lasting maladaptive outcomes. For example, the various enzymes that convert opioid precursors to their final products function at different rates, leading to a different mix of neuropeptides being released as a function of neural activity (Bronstein et al. 1990). Exposure to chronic morphine inhibits activity of neurons that synthesize the highly potent β-endorphin(1–31). Reduced release allows for its conversion to the less potent form, β-endorphin(1–27). Upon withdrawal from chronic morphine, the endogenous system suffers from not only downregulated receptor function but also a shift toward a less potent opioid neurotransmitter signal, likely contributing to negative affect and increased sensitivity to pain.
Many factors are known to dysregulate the opioid system, ranging from genetic differences to perturbations from environmental sources such as the experience of childhood adversity (Lutz et al. 2018), chronic pain (Llorca-Torralba et al. 2019), migraines (Jassar et al. 2019), or alterations resulting from exposure to opioid drugs (Trujillo et al. 1995), alcohol (Shibasaki et al. 2013), and other addictive drugs (Mongi-Bragato et al. 2018; Turchan et al. 1998, 2002). Dysregulation of the opioid system has in turn been linked to neuropsychiatric diseases, including depression (Lutz & Kieffer 2013, Pecina et al. 2019), personality disorders (Bandelow & Wedekind 2015, Prossin et al. 2010), post-traumatic stress disorder (PTSD), Alzheimer’s disease (Torres-Berrio & Nava-Mesa 2019), and anxiety (Bruchas et al. 2010, Sher 1998).
While the breadth and complexity of this system result in widespread involvement in a variety of areas, we limit the scope of the remaining overview to the role of the opioid system in the overlapping disorders of addiction and affective disorders, particularly depression. However, these disorders do not occur in a vacuum. The opioid system sits at the point of convergence between many systems, including addiction, stress, affect/mood, feeding, sexual behavior, immune function, and others, and influences them all simultaneously. Thus, dysregulations in the endogenous opioid system could manifest as disturbances in any of these.
The role of the endogenous opioid system in the mechanisms directly underlying addiction have been well studied. By contrast, the role of this system in the regulation of affect and this contribution to the development and maintenance of addictive states, including OUD, are understudied. In the human clinical population, mood disorders, particularly depression, are often comorbid with opioid use/misuse (Goesling et al. 2015). This relationship is bidirectional, with OUD and depression being risk factors for the emergence of the other (Manchikanti et al. 2007, Scherrer et al. 2014). Pain is often (though not always) a mediating factor in this relationship. Opioid use can also trigger opioid-induced hyperalgesia, which in turn can precipitate or worsen depression.
The opioid system is, among other functions, a regulator of the emotional circuitry of the brain, responsible for moment-to-moment fine tuning of affective state as well as emotional responses to both positive and negative experiences (Drolet et al. 2001, Kennedy et al. 2006, Koepp et al. 2009, Ribeiro et al. 2005, Tejeda & Bonci 2019). This regulation is constantly being refined by changes in peptide tone, composition of the peptide milieu (which specific combinations of peptides are released), and the level and composition of receptor expression, as described above.
Different subsystems within the opioid system mediate distinct components of the affective response. For example, the dynorphin/kappa subsystem has been shown to mediate the aversive affective component of pain (Cahill et al. 2014, Liu et al. 2019, Massaly et al. 2019). In contrast, the mu opioid receptor in affective circuits acts to induce or increase hedonic pleasure and to blunt aversive emotional responses (Inui & Shimura 2017, Pecina & Berridge 2005). Thus, it can be broadly hypothesized that the basal functional profile of kappa versus mu receptor systems may mediate trait-level affective valence and intensity of responses to emotionally relevant stimuli, and this functional profile might be tunable through various mechanisms. It is worth noting, however, that viewing the mu and kappa systems as always mutually antagonistic is simplistic, as it appears these systems can also work together to mediate hedonic responses in some brain regions such as the nucleus accumbens shell (Castro & Berridge 2014). The complex relationship between kappa and mu receptor biology, where kappa receptor agonism is sometimes functionally antimu and sometimes not, is likely mediated by nuances in receptor expression profiles in particular brain areas, functional selectivity of particular combinations of peptide transmitters and receptor variants, and the temporal dynamics of activation (acute versus chronic) (Emery & Eitan 2019).
The opioid system has long been described like a seesaw, dynorphin/kappa on one side and mu-delta/β-endorphin-enkephalin on the other, in balance. Imbalance toward the mu side leads to increased vulnerability to addiction, while imbalance in favor of kappa signaling drives negative affect and mood disorder. However, hypotheses rooted in such a model have not been borne out consistently in the literature. In place of the seesaw model, the opioid system can instead be imagined as a carnival plate-spinner; a complex act balancing synthesis, processing, and release of peptides and the expression profile of receptors (and splice variants), as well as a balance between functional receptors, expressed but desensitized receptors, internalized receptors in reserve, and internalized receptors tagged for degradation. Unlike the seesaw model, where balance can be restored by readjusting either side of a mutually opponent system, even a slight nudge toward imbalance anywhere can set the whole system off-balance, much like our plate-spinner when a single plate among many begins to wobble. The relatively slow-responding regulatory biology of the opioid system leaves the system prone to overcorrection, leading to swings from one state of imbalance to another. If the opioid system is chronically disturbed (e.g., via drug abuse), it may be difficult for the body to reestablish homeostasis, requiring a restabilizing influence from without.
Repeated use of opioids creates a vicious cycle whereby the resulting affective dysregulation can be relieved by opiate drugs, driving the individual to continue seeking and taking opiates to restore allostatic balance to the now-compromised system. This interpretation aligns with the addiction model offered by Koob (2015), whereby continued use of addictive drugs is driven largely by attempts to ameliorate the negative physiological and psychological consequences of drug withdrawal. This vicious cycle may manifest not only as relapse to drug taking but also through any of the behaviors impacted by endogenous opioid imbalance, such as a mood disorder, which in turn may reactivate drug seeking.
Depression represents an excellent example of the close interplay between addiction, mood, and the opioid system. Half of patients with depression report that their symptoms are not adequately managed by available interventions, and about 35% are completely resistant to interventions (Akil et al. 2018). The neurocircuitry of depression contains many regions with high degrees of opioid expression, including the hippocampus, amygdala, nucleus accumbens, prefrontal cortex, and hypothalamus (Akil et al. 2018, Mansour et al. 1988). The opioid system has also been functionally implicated in mood disorders (Lutz & Kieffer 2013). While conventional antidepressants target the monoamine systems, there is a growing appreciation of the contribution of the opioid system and interest in targeting it for novel antidepressants, and these efforts, while preliminary, have been promising.
One example of an effective, novel antidepressant is tianeptine, which has recently been demonstrated to be a highly selective mu receptor agonist (Gassaway et al. 2014). Activity at the mu receptor was demonstrated to be necessary for the antidepressant effect of tianeptine, though the exact mechanism behind this effect is elusive (Samuels et al. 2017). Another example is ketamine, an NMDA receptor antagonist, which demonstrates rapid and long-lasting antidepressant effects (Zarate et al. 2006) and has recently been approved for treatment-resistant depression. Data suggest that activity at the opioid receptors is partially responsible (Williams et al. 2018). Interactions between the opioid and NMDA systems also play a role in pain (Trujillo & Akil 1991, 1994), hinting at another mechanism where pain and depression interact.
The use of prescription opioids for pain is regarded as a gateway mechanism (Am. Soc. Addict. Med. 2016, Compton et al. 2016, Miech et al. 2015) where pain puts individuals into initial contact with opioid medications, and chronic use of opioid medications can develop into dependence and addiction. However, the relationship may be more complex. The relationship between pain and depression implies an underlying biological dysregulation of the endogenous opioid system. Thus, pain, particularly chronic pain, or depression may indicate vulnerability to opioid abuse resulting from endogenous opioid dysregulation, even without previously identified genetic propensity or environmental risk.
The complex and delicate balance of the endogenous opioid system provides for several paths to the functional end state of addiction ( Figure 2 ). One such pathway is genetic propensity, which largely delineates the starting state of an individual’s opioid system. Indeed, a genetic component to addiction vulnerability has been well known for a long time (Merikangas et al. 1998) and leveraged by researchers for decades by employing rodent models with differing addiction vulnerability (Berrettini et al. 1994, Flagel et al. 2016, Murphy et al. 2001, Shoaib et al. 1995).
Genes and environment interact to provide multiple distinct behavioral and biological pathways that converge on the same behavioral output of opioid abuse/addiction. For example, a genetic propensity toward novelty- or sensation-seeking behavior would provide one pathway to opioid use and abuse, while attempts at self-medicating depression resulting from endogenous opioid system dysregulation represent another pathway, which is likely very different in its etiology and trajectory. The opioids themselves are likely to be the dominant factor in altering brain structure and function while the drug is being used, largely masking individual differences when examined during this period. However, the long-term consequences and associated vulnerabilities are not likely to be identical between the groups following cessation of drug use. For instance, individuals who arrived at chronic opioid use via these different pathways may display sensitivity to different triggers of relapse. This then would necessitate different strategies to support abstinence and prevent relapse dependent upon individual differences.
In humans, twin studies have revealed that about 50% of the susceptibility to opiate addiction is heritable (Kendler et al. 2003, Tsuang et al. 1998), comparable to the heritability estimates for alcohol abuse and nicotine addiction (Sharp & Chen 2019, Verhulst et al. 2015). A large fraction of the heritability is due to factors shared across addiction broadly, though a fair proportion is unique to opiates specifically (Tsuang et al. 2001). Addiction is highly polygenic (Hall et al. 2013), though a short list of genes accounts for a relatively large amount of the variance (Reed et al. 2014). However, at present, genome-wide association studies (GWASs) of opioid addiction are of considerably smaller scale than those targeting psychiatric disorders such as depression or schizophrenia. Thus, there remain several unanswered questions regarding the genetics of addiction.
In humans, well over one hundred single nucleotide polymorphisms (SNPs) in the mu opioid receptor alone have been characterized (Ikeda et al. 2005). The most frequent and best-studied SNP in the mu opioid receptor, A118G, has been strongly linked with increased susceptibility to addiction, both to opioids and to other addictive substances (Deb et al. 2010). This mutation also alters responsiveness of the receptor to endogenous opioid peptides (Bond et al. 1998). A homologous mutation in a mouse model exhibits altered reward sensitivity and addiction-like behaviors (Mague et al. 2009, Zhang et al. 2015). These results highlight the importance of genetically mediated initial system state to the development of opioid addiction.
As noted above, a major factor that modulates the initial vulnerability to OUD is temperament. Human studies show that some individuals are prone to externalizing behavior such as aggression, impulsivity, sensation seeking, and psychopathic behavior, and these tendencies are strong predisposing factors to substance abuse in general and OUD in particular (Amirabadi et al. 2015, Cloninger 1987, Milivojevic et al. 2012, Zuckerman & Kuhlman 2000). Others are prone to internalizing behavior and are more likely to exhibit anxiety, depression, and other mood disorders, and following stress they become vulnerable to substance use and likely account for overlap between mood disorders and OUD (Khan et al. 2005).
We have developed an animal model to capture these temperamental tendencies (Stead et al. 2006). At one extreme, rats that are externalizers and respond strongly to novelty [bred high responders (bHRs)] are highly prone to seeking drugs of abuse and becoming addicted to them (Flagel et al. 2016). On the other extreme are internalizers [bred low responders (bLRs)] that are highly prone to anxiety-like (Stead et al. 2006), depressive-like (Stedenfeld et al. 2011), and PTSD-like behaviors (Prater et al. 2017). These lines of animals capture two distinct paths to substance abuse: sensation seeking (bHRs) and negative affect (bLRs). Importantly, we have shown that these tendencies are highly genetically rooted. Indeed, only seven genetic loci account for two-thirds of the genetic variance and one-third of total variance of novelty-induced locomotor behavior in these animals (Zhou et al. 2019). It should be noted that novelty-induced locomotion has also been genetically linked to opiate seeking in various mouse strains (Ambrosio et al. 1995). We know a great deal about neural differences that emerged in these animals through selective breeding. Of relevance here is that the endogenous opioid system is clearly different, including differences in the relative balance between mu and kappa opioid receptors (Turner et al. 2019). This then provides a valuable genetic model for analyzing gene by environment interactions that lead to differential vulnerability to opioid use, addiction, and relapse.
Another variable that influences susceptibility—or resilience—to addiction is that of the environment. While genetic makeup provides the boundary conditions determining the range in which the opioid system can be tuned, environmental influences are the mechanisms that are responsible for the tuning within those boundary conditions. Environmental conditions can be broadly positive or negative, enhancing resilience or vulnerability, respectively. Enrichment of the physical environment provides a protective effect, reducing opioid self-administration (Hofford et al. 2017) and blunting conditioned place preference (CPP) to heroin (El Rawas et al. 2009). Environmental enrichment also aids in the establishment and maintenance of drug abstinence in dependent animals (Peck et al. 2015) and protects against cue-induced reinstatement of extinguished heroin seeking (Galaj et al. 2016).
The role of social environment per se on opioid addiction–like behaviors is distinct from enrichment of the physical environment. Social enrichment has been shown to blunt morphine CPP (Kennedy et al. 2012) and opioid self-administration (Bozarth et al. 1989). Additionally, when given a choice between heroin self-administration and social interaction with a conspecific within the drug-paired context, rats demonstrate a strong preference for social interaction. Furthermore, social interaction blocks incubation of heroin craving following forced abstinence in an opiate-dependent animal (Venniro et al. 2016). Conversely, social isolation increases the locomotor-activating effects of morphine (Coudereau et al. 1996). Paradoxically, social isolation blunts morphine CPP (Coudereau et al. 1997) and physical dependence on morphine (Adler et al. 1975). This is likely mediated through alteration of endogenous opioid function, reducing expression of mu opioid receptors throughout the brain, including areas responsible for the rewarding effects of morphine (Van den Berg et al. 1999).
The quality of social interaction also matters. Mice housed with drug-naïve conspecifics show reduced CPP, sensitization, analgesic tolerance, and opioid-induced hyperalgesia following morphine compared to mice housed with morphine-exposed conspecifics, despite an otherwise-similar social environment (Bates et al. 2014, 2016; Cole et al. 2013; Hodgson et al. 2010), implying alterations in endogenous opioid function. This effect is mediated in part by reduced positive physical interaction, particularly social grooming (Bates et al. 2017).
The effect of social interaction upon addiction liability is not independent of its role on opioid neurobiology vis-à-vis mood and/or pain. The influence of the endogenous opioid system on social bonding in mammals, as well as influencing human feelings of social connection, is well established (Inagaki 2018, Panksepp et al. 1994, Resendez et al. 2016) and is perturbed by the use of opiate drugs (Ragen et al. 2015, Rubin & Bridges 1984, Wang et al. 2018). Thus, the social network has a large impact on the biology mediating desire for, and response to, opioid drugs both directly and through intermediary states of pain (physical and emotional) and depression. Indeed, anecdotal clinical reports indicate that social isolation, abandonment, and despair increase opiate seeking, while social support provides a strong ameliorating effect (Rosenthal 2009).
Another facet of environment is the specific opioid used. Different opioids demonstrate different capabilities to foster sensitization, opioid-induced hyperalgesia, and rate/degree of tolerance at initially equianalgesic doses (Barwatt et al. 2013; Emery et al. 2015, 2016, 2017a, 2017b). Thus, the specific opioid drugs used/misused should be considered as another component that contributes to vulnerability to opiate addiction and/or outcomes influenced by opiate use, such as depression. This is particularly important when reporting adverse clinical effects and epidemiological statistics, where often categories such as opioids are not further subdivided.
Recent findings suggest that opioid drugs may not only drive the endogenous system to superphysiological levels but evoke signaling responses that are qualitatively different from those evoked by endogenous opioids (Stoeber et al. 2018). This is thought to be due partially to the fact that exogenous opioid drugs are membrane permeable, enabling them to activate intracellular opioid receptors. The degree to which endogenous peptides cross the cell membrane remains controversial (Ganapathy & Miyauchi 2005, Marinova et al. 2005); however, such membrane penetration is likely to have very different kinetics compared to small-molecule opioids. Thus, the use of opioid drugs may cause a unique physiological state that the opioid system was not designed to encounter often or at all (i.e., significant and sustained activation of cell-interior opioid receptors). These variables (genetic predisposition, environmental factors, and their interplay with the nature of the drug itself) should not be considered independently of each other, as they interact to influence the probability of addiction and severity.
In spite of the scale of the current epidemic, the treatment options for OUD are very limited. In a recent report by the US National Academies (US Natl. Acad. Sci. Eng. Med. 2019), only three drugs were listed as currently approved by the US Food and Drug Administration and effective for OUD treatment—methadone, buprenorphine, and naltrexone, all of which target the endogenous opioid system (Hedegaard et al. 2018). Significant challenges are associated with each, including lack of compliance with naltrexone because of negative affect and limited access to methadone and buprenorphine, as they require special clinics and/or licensing to dispense them. The social stigma associated with replacement therapy is also a major barrier to treatment, reducing compliance and contributing to relapse. Finding new strategies for treatment that lead to greater compliance and better long-term success is urgent. This, in our view, cannot be successful without taking into account the multiple paths that have led to the OUD and tailoring treatments accordingly. A precision medicine approach should be grounded in a better understanding of the stable and genetically mediated temperamental features of the individual, coupled with the history and current status of that individual, beyond medical history and including their affective state, the chronicity and psychological features of their pain, and the availability of social support.
At one extreme may be the patient with a propensity for internalizing behavior who is presenting with a pain condition. Given that the opioid system mediates affective state, it can be assumed that in a depressed patient, this system is in a state of imbalance. Since pain itself triggers negative affect, a pain patient with a history or current incidence of depression may be at particular risk for opiate abuse, where the opiate temporarily relieves their affective misery while simultaneously amplifying the underlying dysregulation of their endogenous opioid biology.
The other extreme includes people who have temperamental tendencies for sensation-seeking and antisocial behavior, who have a history of abusing drugs, and who might be seeking the rewarding aspect of opioids. Here again, the use of opioids will trigger adaptations that could lead to an OUD but with distinctly different neurobiological sequelae, likely requiring different approaches to treatment.
In order to move toward more strategies for precision treatments of OUD, as well as better approaches to prevention by considering the patient prior to pain treatment, there are several gaps that need to be addressed in the opioid field.
Genetics and genomics. While OUD appears highly heritable, the size of GWASs aimed at understanding the contribution of genetic variants is remarkably small. Other psychiatric disorders, including schizophrenia, bipolar disorder, and major depressive disorders, have profited from very large samples of subjects allowing identification of contributing genetic loci (Bergen & Petryshen 2012, Liu et al. 2011). While they explain a limited percentage of the variance, they have been valuable in pointing to unexpected genes and loci of interest that can be studied in animal models. In addition, the relationship to personality traits and temperament merits further exploration at the genetic level. There are recent studies that have focused on features such as impulsivity and neuroticism (de Wit 2009, Terracciano et al. 2008), both of which are clearly relevant to the two paths to OUD that we have described above. Finally, the pharmacogenomics associated with various opioid drugs is worthy of further exploration, as it might inform treatment of patients to modulate pain while minimizing OUD.
The impact of opioids on the human brain. While animal studies on the neurobiology of opioids have been extensive, we know rather little about the impact of OUD on the human brain. A recent literature search of “human postmortem brain” combined with various drugs of abuse was severely underrepresented compared to other disorders ( Figure 3 ). While the reasons for this are complex, one possibility is the assumption that these consequences are more easily, reliably, and cheaply studied in animal models. However, animal studies likely fail to capture the heterogeneity of human addiction and comorbidity with other disorders that might interact with and amplify the effects of opioids. Thus, genetics, gene expression, and anatomical studies of human postmortem tissue from opioid addicts, healthy controls, and individuals with a history of depression or pain are likely to be highly informative. As importantly, this will likely uncover novel players that might serve as biomarkers or novel targets for the treatment of these diseases.
Despite a growing appreciation of addiction as a neurobiological disease, human postmortem studies of the neuropathology of addiction currently lag behind postmortem studies of other psychopathologies. This limits our understanding and appreciation of individual differences in addiction pathology. Furthermore, this limits the ability to determine the degree to which genetic and neurobiological factors are universal as opposed to substance specific.
Reverse and forward translation. Results from human studies, whether genetic or postmortem, should be reverse translated to animal models that more completely represent the complex gene–environment interaction leading to human OUD. Such models will result in much more reliable preclinical findings and, hopefully, more successful translation.
Refinement of our understanding of endogenous opioid and other interacting systems. In particular, understanding the dynamic regulation and interplay of opioid system elements in the context of known circuits and in behaving animals is a critical backdrop to understanding the impact of various conditions such as pain and treatments such as exogenous opioids. Given the overlapping interactions between multiple opioid peptides and receptors anatomically and functionally, this may prove critical in understanding the changes elicited by various conditions, including pain and addiction. There has been recent progress in our ability to detect the real-time release kinetics of opioid peptides in response to stimuli and their effect on downstream circuit function. This includes the refinement of fast-scan cyclic voltammetry to detect enkephalins and dynorphins with high spatiotemporal precision in vivo (Calhoun et al. 2019). Another exciting approach to tackle this problem is the refinement of microdialysis liquid chromatography/mass spectrometry to identify release of specific peptides. Coupled with optogenetic stimulation, this allows examination of specific cell populations on motivated behavior while simultaneously monitoring which specific opioid peptides are responsible for such effects (Al-Hasani et al. 2018).
The combination of the above approaches across levels of analyses and between animals and humans is essential in developing the prevention and treatment strategies that recognize the great heterogeneity of this drug epidemic.
In summary, opioid addiction, depression, and pain are highly comorbid diseases. They are likely all manifestations of biological dysfunctions sharing a common factor: the endogenous opioid system. The endogenous opioid system is itself regulated by a complex dance between genetics and environmental factors. The latter includes the individual’s physical and social context, past and current emotional state, pain state, and exposure to drugs. While there are multiple genetic and environmental paths leading to a common state we term opioid addiction, the heterogeneity of the causes implies a heterogeneity of consequences both biological and psychological. They therefore require different treatments. Deeper and more systematic exploration of these variables will prove vital to developing better treatment and prevention strategies not only for addiction but for all the related neuropsychiatric ailments involving the endogenous opioid system.
The authors would like to thank A. Parsegian and C.A. Turner for their helpful discussions regarding this manuscript. This work was supported by the National Institute on Drug Abuse U01DA043098, the National Institutes of Health R01MH104261, the Office of Naval Research N00014–12-1–0366 and 00014–19-1–2149, the Hope for Depression Research Foundation, and the Pritzker Neuropsychiatric Research Consortium.
The authors are not aware of any affiliations, memberships, funding, or financial holdings that might be perceived as affecting the objectivity of this review.
