Development and Pathology of OUD

The horrors of the opioid epidemic extend beyond the estimated ~ 45,000 annual deaths from opioid overdose. Firstly, for every fatal opioid overdose there are many more non-fatal overdoses, some of which inevitably lead to hospitalization, intubation, coma, and permanent disability. Additionally, common sequelae afflicting the roughly 2.1 million people with OUD spans infections such as hepatitis C, HIV/AIDS, abscesses, and endocarditis (a bacterial infection of the heart tissue that can quickly lead to sepsis, heart failure, and death), trauma and injury, worsening of psychiatric conditions such as anxiety and depression, job loss and decreased worker productivity, child custody cases, relationship instability, and criminal justice involvement.^{Reference Williams, Compton and Manseau8} As a result, it is estimated that the current epidemic costs the nation hundreds of billions of dollars a year.^{Reference Leslie, Ba, Agbese, Xing and Liu9}

Given these widely publicized and dramatic risks and outcomes, it is striking how so many individuals have nonetheless developed OUD in the past two decades. Although much remains to be discovered about addiction, scientific researchers have determined several key pathways involved in the development of “substance use disorders” as they are called in the field of psychiatry's Diagnostic and Statistical Manual, 5th edition (DSM-5). Since the 1970s, the centrality of dopamine release in reward pathways has been emphasized as a critical step for intoxicating substances that hijack the brain,^{Reference Koob10} especially among persons with low basal levels of dopamine activity.^{Reference Wise11} Dopamine release in the Central Nervous System following self-administration of highly potent addictive substances greatly surpasses that from natural pleasurable activities such as eating rich foods, social activity, sex, and adventure sports. Thus, it has been postulated that some individuals are at risk for over-learning the habitual use of drugs. It is thought this occurs among approximately 15-25% of regular users of any given class of intoxicants (e.g. alcohol, cannabis, opioids, nicotine, stimulants).^{Reference Lopez-Quintero, de los Cobos and Hasin12}

Recent discovery stemming from the boom in neuroimaging in the 1990s has added much more nuance to the dopamine-hijacking-the-brain hypothesis. There is now greater appreciation for disruptions between neural circuits in the frontal lobes that control executive functioning, inhibition and impulse control, and higher-level reasoning and the aggressive emotional circuits of the lower brain (sometimes referred to as the reptilian brain as it is structurally shared across most non-human animal species) that regulate drives for survival such as hunger, defense, and procreation.^{Reference Volkow, Koob and McLellan13} With longer periods of frequent heavy drug use, top-down control by the frontal lobes (e.g. orbitofrontal cortex, anterior cingulate, dorsolateral prefrontal cortex) over circuitry associated with strong drives (e.g. limbic system including the ventral tegmental area (VTA) and the striatum) weakens.^{Reference Koob, Galanter, Kleber and Brady14} As a result, many individuals experience increasingly strong drives to continuously use a substance while becoming less inhibited and less aware of mounting consequences.^{Reference Williams, Olfson and Galanter15} Over time, use can become compulsive and uncontrollable. The role of dopamine release and impairments in top-down control are thought to be involved in the development of addiction, spanning all major classes of drugs.¹⁶

Who is at risk? In some ways this question has been answered. Individuals with family histories of drug addiction are clearly at risk, such as members of families with intergenerational alcoholism or someone with multiple biological siblings with nicotine use dis-order.¹⁷ These individuals would clearly be at increased risk of developing, respectively, alcohol use disorder or nicotine use disorder with repeated exposure. Largely, this shared risk among family members reflects gene activity, likely impacting the structure and distribution of relevant cell surface receptors and signaling molecules (i.e. glutamate transmission or nicotinic acetylcholine receptors) encoded by those regions of DNA. For instance, epidemiological studies in monozygotic (i.e. identical) and dizygotic (i.e. fraternal) twins have shown that genetic factors account for ~ 60% of the variance of drug abuse and addiction.^{Reference Kendler, Myers, Prescott, Agrawal and Lynskey18}

Beyond what are likely combinations of high-risk genes that actively increase risk of uncontrollable use of a given class of drug, there are other risk factors that have been repeatedly associated with the development of substance use disorders including biological characteristics, psychological traits and symptoms, and social determinants of health.¹⁹ Foremost is trauma and early adverse childhood events. Especially among women with substance use disorders, rates of early life trauma are high.^{Reference Hien, Cohen, Miele, Litt and Capstick20} In addition to the profound and sometimes lifelong psychological disruptions from experiencing trauma, it is now increasingly appreciated that early life adversity can change neurochemical messaging and network activity in the brain.^{Reference Dube, Felitti and Dong21} As a result, affected individuals may have greater mood reactivity, lower distress tolerance, increased difficulty attaching to others in healthy ways, and be more prone to outbursts of anger and mood lability.²² Relatedly, individuals, especially adolescents, with untreated psychiatric disorders such as anxiety and depression have higher rates of substance use and addiction treatment admissions.²³ Such psychiatric symptoms can both contribute to, and be exacerbated by, personality traits such as thrill seeking, externalizing behaviors, delay discounting (over valuing short term rewards above long-term satisfaction), and impulsivity.

Individual risk factors for developing a substance use disorder (e.g. OUD), such as genes, family history, trauma, psychiatric conditions, and personality traits can be amplified by certain environmental factors. For instance, individuals predisposed to addiction have much worse outcomes when they live in environments with easy access to intoxicating substances (i.e. ubiquitous prescription of opioids, advertisement of pharmaceuticals, cheap black market prices) and minimal constraints on their use (societal acceptability, peer pressure, lack of employment opportunities).^{Reference Fischer, Russell, Sabioni, van den Brink and Le Foll24}

Cannabis as a Gateway Drug

While gateway theories about cannabis can be dated back to the Reefer Madness of the 1920s and beyond, contemporary work on the gateway hypothesis is largely attributable to Eric and Denise Kandel's pioneering investigations spanning lines of animal, human, and epidemiological research. In a seminal article in the New England Journal of Medicine (2014), the Kandels demonstrate how “one drug affects the circuitry of the brain in a manner that potentiates the effects of a subsequent drug.”²⁵ In this case, the Kandels showed in early work how administering nicotine to mice for seven days before exposing them to cocaine changed (i.e. amplified) their response to subsequent cocaine administration. How could this be possible?

Much of the explanation lies in learning and memory formation at the molecular level. In short-term memory, an external stimulus, such as painful shock to a rodent's tail, causes the sensory neuron detecting the stimulus to send information in the form of neurochemical signals across the synapse to a motor neuron which instructs the rat's tail to move out of the way in order to reduce pain. In short-term (i.e. temporary) memory, the structure of neither cell is changed. In long-term memory however, such as with repeated stimuli, a gene transcription factor, cyclic AMP (cAMP) response-element–binding protein (CREB), is activated to turn on downstream genes, resulting in the growth of new synapses in the sensory cell, thus strengthening its connection with motor neurons.²⁶ This long-term memory is also referred to as long-term potentiation and in the case of facilitating a rodent's ability to avoid pain would be considered adaptive.

Long-term potentiation is relevant to nicotine's gateway effect on cocaine use because nicotine directly impacts CREB's ability to turn on downstream genes. The Kandels explain (2014):

Genes associated with long-term memory have two regions: a regulatory (promoter) region and a coding (early genes and late genes) region. Early genes refer to genes that are turned on early in the process of switching on longterm memory and in turn lead to the activation of late genes, which encode for proteins essential for the growth of new synaptic connections. DNA is wrapped around nucleosomes (octamers of histones). The combination of DNA and nucleosomes is called chromatin. The nucleosomes cluster around the promoter region and need to be modified (acetylated) for the transcription factors to bind to the promoter regions so that memory is stored. This modification is referred to as the acetylation of chromatin structure.²⁷

In summary, they demonstrated in animal models how exogenous chemicals (in this case nicotine) can alter gene expression affecting long-term potentiation in response to other drugs of abuse (in this case cocaine) through increased global acetylation of chromatin in the striatum, the same region of the brain directly implicated in the development of addiction.²⁸ The Kandels found in their studies that nicotine changed the ability of cocaine to affect gene expression based on its impact on CREB activity. In follow up studies, they further showed that this robust effect was minimized among mice bred with genetic mutations leading to the under production of CREB.

While their work is convincing that nicotine can serve as a gateway drug to cocaine in mice models, it cannot speak directly to the possible role of cannabis as a gateway drug for opioids in adolescent humans. However several lines of research are suggestive that similar processes may occur for a subset of individuals with genetic risk. For instance, the adolescent brain is on an ongoing state of pruning and remodeling. And it is known that the endocannabinoid system plays a pivotal role in adolescent brain developmental phases through action on neurotransmitter systems regulating neurogenesis, differentiation, and synaptogenesis.^{Reference Orr, Spechler and Cao29} These critical stages of pruning and remodeling strategically reduce the number of neurons and shape their connections during adolescence, a process that does not fully mature until around age 24. It is these processes and the extended period of time it takes for a human brain to fully myelinate that are thought to confer additional risk for addiction with younger ages of trauma exposure, adverse events, and drug use-because the brain is still actively developing and more vulnerable to environmental exposures.

Cannabis and cannabinoids impact adolescent neurodevelopment through many different mechanisms including the regulation of synapse plasticity, modification of adolescent emotion and cognition circuits, the reduction of whole dendrites, reduction in post-synaptic glutamate A2 AMPA receptors, disruption of signaling and impaired neurogenesis, spinogenesis and NMDA-receptor mediated memory formation, and regulation of CREB which directly regulates histone deacetylases controlling gene expression as detailed in the prior section on nicotine and cocaine.^{Reference Miller and Szutorisz30}

For instance, cannabinoid-opioid interactions have been detailed by studies showing that developmental exposure to cannabinoids among rodents increases subsequent heroin-induced conditioned place preference,^{Reference Biscaia, Fernandez, Higuera-Matas, Singh, McGregor and Mallet31} a classic hallmark of addiction severity. Pistis and colleagues have also shown that adolescent rodent exposure to novel cannabinoid agonists has been reported to induce tolerance to morphine.^{Reference Pistis, Perra and Pillolla32} In part this could be explained by the central role of pruning of excitatory synapses in adolescence for fortifying executive functioning, inhibition, and impulse control in adulthood-processes that underlie drug self-administration and the development of addiction.

Cannabinoid-opioid interactions are likely mediated by activation of cannabinoid type-1 (CB1) and μ-opioid receptors within synaptically-linked (or possibly even the same) neurons and afferent axons in the nucleus accumbens (NAc) shell. Co-expression of receptors suggests convergent actions of cannabinoids and opioids on downstream neural activity. For example, when naloxone and anandamide (a naturally occurring molecule in humans that has THC-like effects) are co-administered (i.e. in the same NAc microinjections), endocannabinoid “liking” was muted in the presence of naloxone, hedonic response to cannabinoids may require concurrent endogenous opioid signaling on some level.^{Reference Mitchell, Berridge and Mahler33}

Additionally, rodent investigations exploring the potential gateway effects of cannabis (studies examining adolescent rodent exposure to cannabinoid agonists or THC) provide evidence of enhanced opioid intake and adult sensitivity later in life to opioids.^{Reference Ellgren, Spano, Hurd, Tomasiewicz, Jacobs and Wilkinson34} Hurd and colleagues developed an experimental rodent model that could mimic periodic adolescent use of cannabis use, finding that adult male rats with low to-moderate THC exposure during adolescence exhibited enhanced heroin self-administration.³⁵

As the endocannabinoid system is dynamically altered during adolescence in brain areas central to reward, decision-making, and motivation (e.g. the limbic system VTA and striatum, mentioned earlier) suggests that cannabis use during these critical developmental phases may impact long-term potentiation in the mesocorticolimbic system (spanning the meso-cortex in the frontal lobes and the limbic system), impacting later-life drug use and addiction proneness.

Additional studies on genetic variation suggests these effects may be more likely among an undetermined subset of the population with genes controlling relevant pathways involved in the tightly linked opioid and endocannbinoid systems (such as Penk genes, the expression of FosB, and other messengers in the striatum) vulnerable to cannabis exposure. For instance, Cadoni and colleagues found that adult Lewis and Fischer rats differentially responded to heroin when pre-exposed in adolescence to THC (typically the most abundant and reinforcing cannabinoid in the cannabis plant). Lewis rats developed stronger conditioned place preference to heroin and resistance to decreasing use compared with Fischer rats when first exposed to THC.^{Reference Cadoni36} In other words, THC exposure induced strain-specific changes in the impact of heroin on conditioned place preference and in resultant behaviors provoked by subsequent heroin use. In conclusion, Hurd writes that, “vulnerability involves a delicate balance between factors that promote and protect against disease, and adolescent cannabis use, an environmental factor, may tip this balance in teens with high-risk genotypes and behavioral traits.”^{Reference Hurd37}

In summary, several converging lines of inquiry have shown that adolescent and young adult (i.e. through age 24) brain development is key to executive functioning and behavioral control, that cannabis can change adolescent gene expression and alter these key periods of neurodevelopment, that genes can predict the priming impact of cannabis on opioids and that there is likely individual variation in the risk of cannabis use in adolescence having a deleterious effect on adolescent brain maturation and downstream vulnerability to opioid exposure and addiction.

Population-Level Studies

Although the prior section largely drew from animal lab studies, consistent findings exist in the human literature at the population level, highly suggestive of interactions between the endogenous opioid and endocannabinoid systems. For instance, it has long been known that there are high rates of cannabis use among methadone patients with OUD. Human clinical trials have shown that intermittent cannabis use has been associated with attenuated opioid withdrawal, that medical cannabis participants report reduced use of prescription opioids, and innovative human lab studies have shown that low dose cannabis and low dose opioids when co-administered can provide superior analgesia than the use of either alone at higher doses.^{Reference Cooper38} These findings have prompted some advocates and policymakers to push for the availability of medical cannabis for patients with OUD.^{Reference Voelker39} It is premature, however, to suggest there is conclusive evidence for cannabis benefiting patients with OUD or the risk for developing OUD.⁴⁰ Further, with increasing prevalence of high potency cannabis, up to 90+% for concentrates, it is unclear what population-level effects will be observed in the coming years.^{Reference Williams and Hill41}

For instance, what if the use of cannabis actually worsens outcomes among these populations of interest? Large national studies suggest this may be the case, or at least that the many touted protective benefits of cannabis for people using opioids may be overblown in the media. Olfson and colleagues found that among NESARC respondents in waves I and II (the same group of roughly 35,000 individuals were re-interviewed three years apart in 2001-2002 and 2004-2005) that cannabis users were less likely to decrease opioid use than non-users.^{Reference Olfson42} In part due to these findings and other studies, Humphreys and Saitz recently cautioned in JAMA that, “aggregate population associations (eg, between medical cannabis and opioid overdose) may be opposite of those seen within individuals.⁴³ In the only individual-level analysis, which included 57,146 people aged 12 and older, of a nationally representative sample, medical cannabis use was positively associated with greater use and misuse of prescription opioids.^{Reference Caputi and Humphreys44} Their suggestion is that while states with medical cannabis programs may have been associated with slower increases in rates of opioid overdose following their implementation,^{Reference Bachhuber, Saloner, Cunningham and Barry45} that this may be an ecological fallacy unrelated to individuals' experiences with cannabis and the progression of opioid use and OUD. For instance, a recent analysis that replicated Bachhuber's design⁴⁶ found negative associations casting doubt on the earlier publication.^{Reference Shover, Davis, Gordon and Humphreys47}

Additional concerns regard the impact of adolescent cannabis use on psychiatric symptomatology. Among addiction specialists, long-term and high potency use of cannabis is well known to cause or worsen symptoms of anxiety, psychosis, depression, and PTSD.^{Reference Cooper, Williams, Montoya and Weiss48} Cannabis withdrawal can also cause protracted symptoms of anxiety, irritability, insomnia, and pain,^{Reference Di Forti, Marconi and Carra49} all of which can be associated with opioid self-administration. A recent systematic review and meta-analysis in JAMA Psychiatry (2019) of 11 studies including more than 23,000 individuals, found that adolescent cannabis consumption was associated with increased risk of developing depression and suicidal behavior later in life, even in the absence of a premorbid condition.^{Reference Fine, Moreau and Karcher50} Although this association did not replicate for subsequent anxiety disorders, their findings are certainly not suggestive of a protective effect of cannabis on psychiatric comorbidity. Given high rates of psychiatric conditions, especially depressive disorders, among patients with OUD, if cannabis has a gateway effect on subsequent opioid use or the development of OUD it may be indirectly mediated by worsening mental health burdens (e.g. depressive symptoms) among already at risk adolescents.