Introduction

Drug addiction is a heterogeneous disorder characterized by cyclic periods of drug use, withdrawal and abstinence, and drug-craving and recurrence of use (Koob and Volkow, 2016). Addiction is highly prevalent in our society, with an estimated 35 million people world-wide and 19.3 million people in the United States (US) currently meeting diagnostic criteria for a substance use disorder (SUD) (Figure 1A; Substance Abuse and Mental Health Services Administration, 2019; United Nations Office on Drugs and Crime, 2019). Additionally, epidemiological surveys suggest that, in a person’s lifetime, there is a ∼10% prevalence of a SUD (Grant et al., 2016; Substance Abuse and Mental Health Services Administration, 2019). Drug addiction is also one of the largest public health problems in the US, with an annual financial burden of $740 billion in costs related to treatment, lost work productivity, healthcare, and crime (National Institute on Drug Abuse, 2020). These numbers are likely to increase as illicit drug use is rising, with a quarter of a billion people worldwide reporting use in the past year (United Nations Office on Drugs and Crime, 2019). Within the US, over 17 million people aged 12 and above are estimated to initiate drug use annually. Rates of opioid use, in particular, are continuing to climb, with 53 million past-year opioid users worldwide and ∼11 million people in the US reporting opioid misuse within the past year (United Nations Office on Drugs and Crime, 2019). This is especially alarming, as the number of deaths in the US involving opioids has increased 6-fold from 1999 to 2017, with ∼130 Americans dying from use per day (Centers for Disease Control, 2018).

Although the majority of research on SUDs has focused on individual substances in isolation, with a multiple drug use history often considered an exclusion criterion for clinical studies, it is important to recognize that many drug users engage in polysubstance use. For instance, 30–80% of heroin users have been reported to also use cocaine (Leri et al., 2003a), and deaths involving both cocaine and opioids in the US more than doubled between 2010 and 2015 (Centers for Disease Control, 2018). A person is considered a polysubstance user if they use more than one substance, including use of multiple drugs on separate occasions (sequential use) or at the same time (concurrent/simultaneous). Limiting studies to individual drugs risks overlooking interactions between substances, decreases translatability of preclinical research, and can impede the efficacy of identified treatments for SUDs. Indeed, polysubstance use has consistently been associated with worse treatment outcomes, including poorer treatment retention, higher rates of relapse, and a three-fold higher mortality rate compared to mono-substance use (Williamson et al., 2006; Staiger et al., 2013; de la Fuente et al., 2014). This review seeks to combine the current knowledge of the mechanisms and consequences of individual drug use with the most up-to-date research on polysubstance use, making sure to note, when possible, if polysubstance use is concurrent/simultaneous, sequential, or a combination of these patterns. We will first provide an overview of public health trends regarding single and polysubstance use, as well as the impact of polysubstance history on metrics of substance use severity. This will be followed by a discussion of findings from preclinical studies, and their translatability to real-world substance use, outlining considerations to be made when designing polysubstance studies. We then detail the pharmacology of individual substances and some of their effects on the cortical-basal ganglia-thalamic (C-BG-T) circuitry, which sets the groundwork for understanding how polysubstance use may change the neuropathology of addiction. Care will be given to discussing the differences between brain alterations in single versus polysubstance use, highlighting the most common combinations of polysubstance use. For clarity and in order to avoid duplication in our discussions, sections are organized by primary used substance, with consideration for the consequences that result when additional drugs are combined with a primary drug. Finally, we offer suggestions and highlight potential methods to move forward with the important task of examining polysubstance disorders.

Figure 1. Public health trends in drug use. (A) Drug use in the United States from 1990 to 2019. Data from the National Household Survey on Drug Abuse (Substance Abuse and Mental Health Services Administration, 1993; Substance Abuse and Mental Health Services Administration, 1995; Substance Abuse and Mental Health Services Administration, 1997; Substance Abuse and Mental Health Services Administration, 2003) and the National Survey on Drug Use and Health (Substance Abuse and Mental Health Services Administration, 2005; Substance Abuse and Mental Health Services Administration, 2007; Substance Abuse and Mental Health Services Administration, 2009; Substance Abuse and Mental Health Services Administration, 2011; Substance Abuse and Mental Health Services Administration, 2013; Substance Abuse and Mental Health Services Administration, 2015; Substance Abuse and Mental Health Services Administration, 2017; Substance Abuse and Mental Health Services Administration, 2019). (B) Unspecified polysubstance use in treatment-seeking drug users in Finland from 1997 to 2008. Top: Primary (left) and secondary (right) drugs used by treatment-seeking drug users, shown as percent of total users. Bottom: Percent of users reporting exclusive misuse of one drug (white bars) or misuse of a given drug along with polysubstance use of another (colored bars). Data from Onyeka et al. (2012).

Public Health Trends in Substance Use

Drug addiction is both pervasive and deadly, with ∼585,000 drug use-related deaths occurring each year worldwide (United Nations Office on Drugs and Crime, 2019). Nonetheless, although drug addiction and its impacts are often centered around individual drugs, drug misuse is largely found to involve multiple substances (Gjersing et al., 2013; Roy et al., 2013; Substance Abuse and Mental Health Services Administration, 2016). Indeed, drug-dependent individuals report an average use of 3.5 substances (Onyeka et al., 2012), including both simultaneous and sequential polydrug use (Figure 1B). In addition, the likelihood of developing comorbid substance dependencies is high in clinical populations (Leri et al., 2004; Lorvick et al., 2018). Although combinations of co-used substances vary, primary drug dependencies are typically found for alcohol, opioids, amphetamine, and methamphetamine, while cannabis and cocaine are more often reported as secondary-or tertiary-used substances (Substance Abuse and Mental Health Services Administration, 2016). The high prevalence of polysubstance use is particularly concerning given the impact it can have on both SUD severity and treatment outcomes. For example, a polysubstance history is associated with greater unmet physical and mental health care needs, increased risk behavior, violence, and increased overdose and mortality risk compared to single substance use (Pennings et al., 2002; Gilmore et al., 2018; Lorvick et al., 2018). In this section, we will discuss the public health consequences surrounding single substance use, as well as polysubstance use in relation to secondary substance combinations. This overview will aim to address overarching patterns of polydrug use, including the substance combinations and patterns of use that most commonly occur. However, given that data is limited for specific drug use patterns, types of classification, and differences in definition of polydrug combinations across studies, it is unlikely to capture all combinations, histories, and patterns of use.

Psychostimulants

Psychostimulants are the second-most widely used class of drugs, with 18 million current cocaine users and 29 million current prescription stimulant users worldwide (United Nations Office on Drugs and Crime, 2019). Worldwide prevalence of psychostimulant use has remained relatively stable from 1990 to 2017, with 7.38 million reported to meet criteria for an amphetamine use disorder and 5.02 million reported to meet criteria for a cocaine use disorder (Degenhardt et al., 2018). However, the number of drug-related overdose deaths involving psychostimulants has continued to climb, especially in the US, with a 2.6-fold increase in the cocaine overdose death rate and 3.6-fold increase in the methamphetamine overdose death rate from 2000 to 2017 (Degenhardt et al., 2018).

Notably, cocaine and amphetamine users are predominantly polysubstance users, with one study reporting 74 and 80% incidence of polysubstance history, respectively (Kedia et al., 2007). Specifically, cocaine use and developing a cocaine use disorder is associated with concurrent heroin, cannabis, tobacco, and alcohol use (Kedia et al., 2007; Roy et al., 2013; John and Wu, 2017). Similarly, amphetamine users exhibit several types of polysubstance use, with high probabilities of alcohol, tobacco, and cannabis use. In addition, other classes of amphetamine polysubstance users exhibit higher probabilities of heroin and other opioid use. Across groups, lower probabilities of cocaine use with amphetamine compared to other drug classes used with amphetamine are reported as well (Darke and Hall, 1995; Kelly et al., 2017). Polysubstance use is common among stimulant users with both concurrent and sequential drug consumption patterns. For instance, simultaneous use of psychostimulants and opioids is seen with both cocaine (“speedball”) and methamphetamine (“bombita”). Sequential use of psychostimulants and opioids is also common, including the use of cocaine or amphetamine to avoid opioid-related somatic withdrawal symptoms (Hunt et al., 1984; Ellis et al., 2018) and the use of opioids to reduce overexcitation following cocaine use (Kreek, 1997). Additionally, there is an increased likelihood of same-day methamphetamine use with alcohol consumption (Bujarski et al., 2014).

Though these studies did not specify the patterns of polydrug use, a meta-analysis of reports on concurrent versus simultaneous cocaine use found a 24–98% range of simultaneous cocaine and alcohol use and 12–76% incidence of simultaneous cannabis use (Liu et al., 2018). Rates of concurrent use were 37–96% for cocaine and alcohol use, 43–94% for cocaine and cannabis use (Liu et al., 2018), 70–80% for cocaine and nicotine use (Budney et al., 1993; Weinberger and Sofuoglu, 2009), and 85–95% for amphetamine and nicotine use (Brecht et al., 2007; Grant et al., 2007). The high variability in reported frequencies highlights the complexity in identifying drug use patterns, which can vary across demographics, study periods, study structure, and definitions of concurrent and simultaneous use.

Polydrug use involving psychostimulants poses significant public health risks. For example, one study showed that amphetamine users were 21 times more likely to have a concurrent cannabis use disorder and 7 times more likely to have past-year concurrent cocaine use, compared to those with no prior history of amphetamine use (Massaro et al., 2017). In addition, nearly one-third of overdose deaths involved both psychostimulants and opioids, such as heroin and fentanyl (Kariisa et al., 2019). The hazards of psychostimulant co-use also extend to other substances, with combined cocaine and cannabis use resulting in higher standardized death rates in emergency department (ED) visits, suggesting elevated mortality risks with this combination (Gilmore et al., 2018). Additionally, combining cocaine and alcohol use increases the risk for cardiotoxicity compared to either drug alone (Pennings et al., 2002).

Nicotine

Although the use of tobacco (i.e. the dried leaves of the tobacco plant containing nicotine) has declined since the early 2000s, nicotine is still one of the most commonly used drugs, with 58.8 million people aged 12 or above reporting past-month nicotine use in the US (Substance Abuse and Mental Health Services Administration, 2019). It is also commonly used with many other substances, as 17% of nicotine users also used cannabis, 4.7% also used opioids, 2.6% also used cocaine, and 1.4% also used psychostimulants in the past month. In contrast, nonsmokers had much lower percentages of past-month substance use (3.7, 1.2, 0.2, and 0.3% for the aforementioned substances, respectively), (Moeller et al., 2018). This difference is notable, given that people with a nicotine use disorder are 3–4 times more likely to have a second SUD (Chou et al., 2016). In addition, it was found that past-year tobacco use was significantly associated with opioid use disorders, as well as comorbidities for cannabis and alcohol use and use disorders, and cocaine use in samples of primary care patients (John et al., 2019). Tobacco use severity (i.e. frequency of use and number of cigarettes smoked) has also been significantly correlated with onset of heroin and cocaine use (Frosch et al., 2000). Historically, nicotine has primaryily been used by smoking tobacco cigarettes. However, new advances in technology have led to the development of electronic (e-) cigarettes, designed to deliver nicotine in a toxin-free manner. The marketing of e-cigarettes as a safer alternative to traditional tobacco cigarettes is concerning, as it has led to an increase in the probability of nicotine use and a resurgence in the potential for nicotine addiction. For example, e-cigarette use among middle and high school students has increased from 2012 to 2016 (Gentzke et al., 2019), and a spike in use was observed among young adults (18–24 years old) around 2013 to 2014, when e-cigarette products were introduced (Gentzke et al., 2019). Despite delivery of lower doses of nicotine, the safety of commercial e-cigarettes has been debated, since compensatory “puffing” behaviors or high voltage settings leads to the production of carcinogenic agents (Jensen et al., 2015). The potential danger of use is further compounded by the variable amounts of nicotine provided across e-cigarette manufacturers (Goniewicz et al., 2013). The unique influence of vaping on the development of nicotine dependence and how this differentially contributes to polysubstance use disorders remains largely unknown and should be studied in the coming years.

Opioids

The prevalence of opioid misuse (i.e. use outside of prescribed use) has risen dramatically in recent years, with ∼53 million adults (1.1% of global population) reporting past-year non-medical use of an opioid (United Nations Office on Drugs and Crime, 2019). In the US alone, 11 million people reported past-year opioid misuse in 2016 (Substance Abuse and Mental Health Services Administration, 2017); however, this estimate is conservative as it does not include homeless or incarcerated individuals with disproportionately higher levels of opioid use. In addition, the rate of first-time heroin users rose in parallel with non-medical use of prescription opioids from 2002 to 2011 (United Nations Office on Drugs and Crime, 2019), a reflection that suggests individuals with past-year prescription opioid misuse are 19 times more likely to initiate heroin use than those without such a history (Muhuri et al., 2013; Cicero et al., 2018). Studies have investigated polydrug use among heroin and prescription opioid misusers and found higher frequencies of opioid use in people that also use cocaine (>33%) or methamphetamine (>20%) (Wang et al., 2017; Hedegaard et al., 2018), but reduced prevelance for primary opioid use in those that have secondary alcohol or cannabis use (Wang et al., 2017; but see Cicero et al., 2020). In addition, first-time methamphetamine use is more prevalent following past-month opioid use (Cicero et al., 2020). Of those entering treatment for heroin use, it has been found that 91% of people reported a lifetime history of cocaine use (Williamson et al., 2006). Additionally, a study in the United Kingdom found that 54% of opioid users in treatment between 2017 and 2018 also had a comorbid crack cocaine use disorder (Public Health England, 2018). With respect to patterns of multi-drug use, simultaneous use of heroin with alcohol and/or cannabis is more common than with psychostimulants (Kelly et al., 2017; Bobashev et al., 2018), and a sequential pattern of drug use is preferred for opioids and psychostimulants.

In the US, opioid use is a national public health emergency responsible for more than 1.6 million years of life lost from 2001 to 2016 (Gomes et al., 2018). Moreover, opioid overdose deaths are currently the leading cause of accidental death among US adults, with 68% of all drug overdose deaths involving an opioid (United Nations Office on Drugs and Crime, 2019). Given that nearly 80% of fatal opioid overdoses also involved another substance, it appears that there is a greater risk of death when opioids are used in combination with other opioids and/or other drugs (Jones et al., 2018). Specifically, of these deaths, 78% involved another opioid, 21.6% involved cocaine, 11.1% involved alcohol, and 5.4% involved a psychostimulant other than cocaine (Jones et al., 2018). Furthermore, opioid-related ED visits also involved tobacco (51.1%), cocaine (36.9%), other stimulants (22.6%), cannabis (25.1%), or alcohol (16.9%). Substantial polysubstance use of three or more of these substances has also been reported for opioid-related ED visits (Liu and Vivolo-Kantor, 2020), and the likelihood of these visits has been associated with the degree of severity of other SUDs (Zale et al., 2015; John et al., 2019). Taken together, these reports suggest that combining opioid use with use of other substances can exacerbate the deleterious consequences of opioid use. In addition to overdose risk, opioid users experience very high rates of relapse, with 59% of individuals relapsing in the first week and 80% relapsing in the first month of abstinence (Smyth et al., 2010). Past use of other substances, including the degree of cocaine use, increases relapse susceptibility (Williamson et al., 2006). Methamphetamine use among those seeking treatment for opioid use has also been on the rise (United Nations Office on Drugs and Crime, 2019), and recent reports indicate that methamphetamine use is associated with a discontinuation of buprenorphine treatment in people with an opioid use disorder (Tsui et al., 2020). Thus, a better understanding of the impact of polysubstance use in the context of opioids is crucial for more successful emergency responses and long-term treatment outcomes.

Cannabinoids

It is estimated that 188 million individuals 12 years or older use cannabis worldwide (United Nations Office on Drugs and Crime, 2019), including 43.5 million individuals in the US (Substance Abuse and Mental Health Services Administration, 2019). Beginning in 2012 with Washington and Colorado, 11 states and the District of Columbia have legalized recreational cannabis, making it legally accessible to ∼328 million people. The number of cannabis users in the US has risen with its gradual decriminalization and legalization, from 4.1% in 2002, to 9.9% in 2007, to 15.9% in 2018 (Hasin et al., 2015; Substance Abuse and Mental Health Services Administration, 2019). Frequency of cannabis use is also high, with reports of 40% of individuals being daily or near-daily users (Substance Abuse and Mental Health Services Administration, 2019).

Cannabis users are reported to have high rates of past month tobacco, alcohol, and/or amphetamine use (Connor et al., 2013). One of the most common combinations is simultaneous use of alcohol and cannabis (McCabe et al., 2006), along with simultaneous alcohol, cocaine, and cannabis use (Liu et al., 2018). The impact of concurrent cannabis is notable, as this pattern of use is associated with more alcoholic drinks per day, suggesting facilitation of alcohol use with coexisting cannabis consumption (Aharonovich et al., 2005; Subbaraman et al., 2017). Polydrug use is particularly prevalent in younger populations. Among adolescent cannabis users, 27.5% reported additional drug use within the same year of starting cannabis use, and nearly 67% use two or more other drugs (Subbaraman and Kerr, 2015). Cannabis is frequently used during treatment for other SUDs (Connor et al., 2013; Subbaraman et al., 2017), and this has been associated with reduced treatment efficacy. For example, cannabis use has been found to result in shorter periods of alcohol abstinence (Subbaraman et al., 2017), as well as greater incidence of relapse to cocaine (Aharonovich et al., 2005; Mojarrad et al., 2014). In addition, polydrug use among cannabis users has been correlated with reduced socioeconomic mobility, financial instability, and relationship difficulties (Aharonovich et al., 2005; Cerdá et al., 2016; Subbaraman et al., 2017), a heightened degree of mood disorder symptom severity, decision-making deficits, social difficulties, and self-harm (Subbaraman and Kerr, 2015; Lopez-Quintero et al., 2018). Although these data suggest that the consequences of drug use are enhanced by concurrent cannabis use, it should be noted that clinical outcomes can vary for studies examining polydrug use among cannabis users. For example, some studies suggest a nuanced impact of polysubstance use that is dose-dependent, with no synergistic effects of cannabis and alcohol at low doses of either drug (Ballard and De Wit, 2011), and a lack of association of cannabis use in heroin relapse (Aharonovich et al., 2005).

Alcohol

Alcohol is one of the most commonly used drugs, with up to 290 million people diagnosed with an alcohol use disorder worldwide (United Nations Office on Drugs and Crime, 2019), including 15 million people in the US (Substance Abuse and Mental Health Services Administration, 2019). Alcohol is frequently used with other substances, with reports indicating that 5.6% of US adults have used both alcohol and another illicit drug within the past year, and 1.1% have met diagnostic criteria for both an alcohol use disorder and another SUD (Falk et al., 2006). The most commonly reported substance co-used with alcohol is cannabis (10%), with less common comorbidities found with opioids (2.4%), cocaine (2.5%), and amphetamine (1.2%) (Falk et al., 2006). Although simultaneous use of alcohol and cannabis or alcohol and prescription opioids is most common (McCabe et al., 2006), simultaneous use is also seen with cocaine (Liu et al., 2018).

Polydrug use increases the risk of developing an alcohol use disorder (Grant et al., 2015, 2016), particularly in young adults, men, and American Indians/Alaskan Natives (Falk et al., 2006). Polydrug use that includes alcohol is associated with additional comorbidities, including higher prevalence of mood disorders, anxiety disorders, more intense drinking, and more intense drug consumption and drug-craving (Preston et al., 2016; Saha et al., 2018). The negative consequences of alcohol polydrug use are also highlighted by data indicating that 21% of ED visits for patients 12–24 years old involved both alcohol and drugs. These visits were also more likely to require treatment for injuries, and had higher rates of inpatient admittance (Naeger, 2017). In addition, 17% of substance treatment admissions were related to both alcohol and drug use, representing 45% of primary alcohol admissions and 33% of drug misuse admissions (White et al., 2011; Substance Abuse and Mental Health Services Administration, 2016; Naeger, 2017). The rate of hospitalizations involving alcohol polydrug use has been increasing, particularly in young adults, with reports suggesting a 76% rise in inpatient admittance between 1998 and 2008, compared to either drug or alcohol overdoses alone (White et al., 2011). While the polysubstance users in these surveys were primarily white and male, recent trends indicate a rise in ED visits relating to alcohol and drug combinations in females (Naeger, 2017), suggesting a change in the demographics of polysubstance combinations that include alcohol.

Behavioral Models of Addiction in Polydrug Studies

Behavioral models of drug addiction are used to examine the neurobiological underpinnings of the development, maintenance and relapse to drug use. The most commonly used models are locomotor sensitization (a progressive and persistent increase in locomotor responses to the same dose of a drug), conditioned place preference (CPP; a test of drug reward measured as an increase in time spent in a drug-paired chamber) and drug self-administration (response-contingent intake of drug) (see Panlilio and Goldberg, 2007; Spanagel, 2017; Kuhn et al., 2019 for review). Experimental designs using these models vary across a number of pharmacological and non-pharmacological parameters including contingency of drug use, amount of access to drug, context associated with drug use, and routes of administration. Here, we describe how these models have been used with polydrug combinations, and how this work has informed our understanding of polydrug use and addiction.

Initial polysubstance studies largely used noncontingent models of drug administration, particularly CPP and cross-sensitization models, whereby the impact of priming doses of one drug on side preference or motor activity, respectively, of another drug are determined (Shippenberg et al., 1998; Lu et al., 2002; Cole et al., 2003; Leri et al., 2003b; Liang et al., 2006). More recently, studies have been examining how drug self-administration history impacts subsequent drug choice preference and/or drug-craving via responding to drug-associated cues following extinction and/or withdrawal (Leri and Stewart, 2001; De Luca et al., 2019; Rubio et al., 2019; Crummy et al., 2020). Preclinical polysubstance models involving simultaneous administration of multiple drugs, such as alcohol and nicotine or cocaine and heroin combinations, have also been used frequently (Mello and Newman, 2011; Mello et al., 2013; Sentir et al., 2020). In clinical models, both concurrent and sequential polysubstance use is assessed in subjects via scoring of affective measures to drug-taking, drug-craving following visual cues, and autonomic response measures such as blood pressure and heart rate (Foltin et al., 1993; Greenwald et al., 2010; Giasson-Gariépy et al., 2017).

More recently, studies are comparing single versus polysubstance self-administration to determine the effect of drug history on drug-induced molecular and circuit alterations (Briggs et al., 2018; Stennett et al., 2020; Zhu et al., 2020). Additionally, paradigms based on behavioral economic principles can determine the preferred level of drug intake (i.e. no-cost intake; Q₀), as well as the amount of effort an animal is willing to exert to defend Q0 before consumption and responding begins to decline (i.e. price; P_max) (Oleson and Roberts, 2009). These paradigms are especially powerful in that they can use Q₀ and P_max to generate normalized measures of value (i.e. essential value; α) and price (_nP_max), which have been used to compare price sensitivity, effort, and value across different drug and non-drug rewards in polysubstance models in several species (e.g. in rats, rhesus monkeys, and human participants) (Petry and Bickel, 1998; Ward et al., 2006; Wade-Galuska et al., 2007, 2011; Crummy et al., 2020). In particular, these studies permit examination of the relative reinforcing properties of different doses and classes of drugs (Wade-Galuska et al., 2007; Cooper et al., 2010; Huskinson et al., 2015), as well as alterations in cost valuation for a drug following pre-exposure to another drug (Cooper et al., 2010; Hofford et al., 2016; Morris et al., 2018). Direct quantification of the assigned value of these drugs across different polysubstance histories and drug doses is very useful for assessing the impact of polysubstance history on relative reinforcer value. Furthermore, these measures can be used to compare how polysubstance users value drug rewards across different experimental parameters (e.g. differences in priming dose of one drug, environmental context, pattern of drug use). Finally, clinical studies are using questionnaires or controlled laboratory environments to investigate the behavioral effects of a polysubstance history. Specifically, these studies use monetary choice procedures that compare assigned value of drugs at different doses (Greenwald et al., 2010), how assigned value changes for one drug with a change in price of another (Petry and Bickel, 1998; Petry, 2001; Sumnall et al., 2004; Chalmers et al., 2010), or how relative value of one drug changes with perceived change in subjective quality of another available drug (Cole et al., 2008). Additionally, progressive ratio tests for a single drug or drug combinations to study motivation (Greenwald et al., 2010), and delay-discounting rates for money and drug rewards to study decision-making, (Strickland et al., 2019) have also been performed. These studies permit comparison of perceived value across multiple drugs in participants with histories of single or polysubstance use (for further review, see Heinz et al., 2012).

Effects of Polydrug Use on Addictive Behaviors

Given the unique neurobiological alterations that can occur with exposure to multiple drugs, along with the high prevalence of polysubstance use disorders, there is a strong need to develop polydrug paradigms that have high translational value. These paradigms are critical for fully understanding the behavioral changes and addiction-related phenotypes that develop following polydrug use. However, given the vast number of potential substance combinations and the variability in methodologies that exist across studies, there are currently mixed results and interpretations regarding the impact of polydrug history on addiction-related behaviors. Nonetheless, some general trends in drug consumption, drug preference and drug-seeking have been demonstrated in commonly investigated substance combinations (Figure 2).

Figure 2. Summary of the effects of specific polydrug combinations on assays of addiction-like behaviors. Studies are organized into X/Y polydrug combos (columns) and behavioral assays (rows), with subcolumns for the effect of drug Y on drug X (left subcolumn) and the effect of drug X on drug Y (right subcolumn). Symbols represent the net effect of the X/Y polydrug combo on a given behavior, with color depicting the specific drugs tested. Sensitization: locomotor sensitization; Conditioned place preference: acquisition, expression; Drug intake: self-administration; Motivation: progressive ratio, behavioral economics; Drug craving: reinstatement, cue reactivity. Data from Mello and Mendelson (1978), Mello et al. (1980), Mello et al. (2014), Huston-Lyons et al. (1993), Foltin et al. (1993), Aspen and Winger (1997), Ranaldi and Wise (2000), Valverde et al. (2001); De Vries et al. (2001), Parker et al. (2004); Solinas et al. (2005), Liang et al. (2006); Biala and Budzynska (2006), Ward et al. (2006); Panlilio et al. (2007), Panlilio et al. (2013), Winger et al. (2007), Lê et al. (2010), Lê et al. (2014), Levine et al. (2011), Cortright et al. (2011), Pomfrey et al. (2015), Maguire and France (2016); Mahmud et al. (2017), Fredriksson et al. (2017); Giasson-Gariépy et al. (2017), Griffin et al. (2017); Schwartz et al. (2018), Winkler et al. (2018); Manwell et al. (2019), Ponzoni et al. (2019), Crummy et al. (2020), and Stennett et al. (2020).

Psychostimulants

Some of the more commonly studied polydrug combinations include administration of cocaine with other drugs (Francesco et al., 2003; Leri et al., 2003b; Substance Abuse and Mental Health Services Administration, 2016). However, in contrast to human reports (Heil et al., 2001; Williamson et al., 2006; Staiger et al., 2013; Preston et al., 2016; Lorvick et al., 2018; Kariisa et al., 2019), an increase in addiction-like severity has not been observed in preclinical studies, suggesting a need for models with greater translational relevance that can capture the enhanced severity seen in human polysubstance users. Specifically, animal studies of sequential cocaine and alcohol or cocaine and heroin use have not found differences in drug intake or reinstatement of drug-seeking as a function of single versus polydrug use (Pattison et al., 2014; Fredriksson et al., 2017; Crummy et al., 2020; Stennett et al., 2020). These effects were observed despite variance in the use of contingent and non-contingent drug administration, drug doses, and species, including rats (Crummy et al., 2020; Stennett et al., 2020) and rhesus monkeys (Aspen and Winger, 1997). Consistent with this work, intermittent alcohol exposure has not been shown to affect cocaine self-administration (Aspen and Winger, 1997; Fredriksson et al., 2017) or the reinforcing properties of cocaine measured via demand curves in rhesus monkeys (Winger et al., 2007). In addition, intermittent alcohol exposure has not been shown to affect progressive ratio tests of motivation for cocaine (Mateos-García et al., 2015), or the long-term reconsolidation of preference for cocaine in drug-paired contexts (Zhu et al., 2020) in rats. In contrast, adolescent alcohol exposure has been shown to have long-lasting effects on cocaine self-administration and reward, suggesting that this population is particularly susceptible to the effects of polysubstance use. For example, adolescent alcohol exposure increases motivation for cocaine (Mateos-García et al., 2015), enhances the development of a cocaine CPP in both mice (Molet et al., 2013) and rats (Hutchison and Riley, 2012; Mateos-García et al., 2015), and weakens cocaine-induced taste aversion (Busse et al., 2005). In addition, simultaneous heroin and psychostimulant administration increases the motivation to self-administer cocaine (Ward et al., 2006) and methamphetamine (Ranaldi and Wise, 2000), and both simultaneous and sequential administration of morphine and methamphetamine have been shown to be rewarding, as measured by the development of a CPP (Briggs et al., 2018). Pretreatment with an opioid also enhances methamphetamine-induced psychomotor sensitization (Liang et al., 2006). These findings suggest that opioids can enhance the rewarding and motivational properties of psychostimulants, particularly when administered simultaneously.

Polydrug studies with cocaine and nicotine have largely reported additive and/or synergistic effects of the two drugs. In particular, co-administration of cocaine and nicotine increases drug intake in rhesus monkeys (Mello et al., 2014) and rats (Bechtholt and Mark, 2002), enhances locomotor sensitization and the development of a CPP in mice (Levine et al., 2011), and induces a cross-sensitized drug-craving (Reid et al., 1998; Weinberger and Sofuoglu, 2009; Cortright et al., 2012). Additionally, chronic nicotine pretreatment facilitates the acquisition of cocaine self-administration (Horger et al., 1992; Bechtholt and Mark, 2002; Linker et al., 2020), increases motivation under a progressive ratio schedule, impairs extinction learning, and enhances drug-primed reinstatement for amphetamine following amphetamine self-administration (Cortright et al., 2012). Conversely, prior nicotine treatment history reduces demand elasticity for cocaine (Schwartz et al., 2018). Notably, the effects of nicotine and psychostimulant polydrug use are largely dose-dependent as pretreatment with a smaller dose of nicotine [0.3 mg/kg, subcutaneous (sc)] increases motivation to take cocaine under a progressive ratio schedule in rats, whereas a larger dose of nicotine (0.6 mg/kg, sc) has the opposite effect (Bechtholt and Mark, 2002). Additionally, simultaneous administration of methamphetamine [2.0 mg/kg, intraperitoneal (ip)] and nicotine (1.0 mg/kg ip) induces a conditioned place aversion in mice, while sequential administration of the same dose induces a CPP (Briggs et al., 2018). Together, these studies demonstrate that the effects of nicotine on psychostimulant motivation, intake, and reward are heavily impacted by the parameters surrounding nicotine delivery (e.g. dose, route, pattern, etc.), which should be carefully considered when comparing study results and developing preclinical polydrug paradigms. Notably, adolescent exposure to nicotine has no effect on the subsequent development of a cocaine CPP, cocaine-induced taste aversion, cocaine self-administration, extinction, or reinstatement of cocaine-seeking in adulthood (Pomfrey et al., 2015), although it has been shown to enhance cocaine self-administration in adolescent rats (Linker et al., 2020). These data suggest that, unlike alcohol, early exposure to nicotine does not lead to increases in addiction-like behavior to cocaine in animals.

In humans, simultaneous cocaine and cannabis use produces feelings of “stimulated” and “high” that last longer than either drug alone (Foltin et al., 1993), and cue-induced drug-craving in individuals who co-use cocaine and cannabis lasts longer than for those who only use cocaine (Giasson-Gariépy et al., 2017). In contrast, THC reduces the motivation to self-administer cocaine in rodents (Panlilio et al., 2007). Although this suggests a differential regulation of cocaine’s effects in humans and rodents, further work is necessary to ensure that animal models of increased addiction severity cannot, in fact, be developed. Notably, however, neither cocaine and cannabis nor cocaine and alcohol co-administration in humans produces subjective effects that are different from cocaine, cannabis, or alcohol alone (Foltin et al., 1993). Similarly, THC pretreatment in rodents has no effect on psychostimulant reward or self-administration, nor does it potentiate the development of a CPP to amphetamine (Panlilio et al., 2007; Cortright et al., 2011; Keeley et al., 2018), indicating a unique effect of cocaine and cannabis on drug-craving. Interestingly, CBD has no effect on cocaine self-administration, motivation, or cue-induced reinstatement of cocaine-seeking (Mahmud et al., 2017), suggesting the effects of cannabis on cocaine craving are likely due to THC, rather than CBD. However, CBD treatment has been found to reduce motivation to self-administer methamphetamine on a progressive ratio schedule and to reduce methamphetamine-primed reinstatement of drug-seeking (Hay et al., 2018). The additive effects of cannabinoid and psychostimulant polydrug use appear to be dependent on both the amount of drug consumed and the age range during use. For example, acute THC weakens psychomotor sensitization, but repeated THC administration promotes tolerance to the acute effects, increasing amphetamine-induced stereotypy and locomotor activity (Gorriti et al., 1999; Cortright et al., 2011). Additionally, adolescent THC exposure accelerates acquisition of cocaine self-administration and increases intake of low doses of cocaine (Friedman et al., 2019), indicating long-lasting changes in reward circuitry following adolescent THC use, similar to alcohol.

Nicotine

Limited work has focused on the effects of polydrug use on nicotine-induced addiction behaviors. However, it has been found that THC pretreatment can enhance nicotine consumption and price inelasticity (measured by α) in behavioral economic tests in rats (Panlilio et al., 2013), and heroin intake has been found to increase cigarette consumption in people (Mello et al., 1980). In addition, pre-exposure to alcohol or simultaneous access to both alcohol and nicotine decreases nicotine self-administration in rodent studies (Lê et al., 2010, 2014). Although access to alcohol has no effect on responding for nicotine under extinction conditions, a priming dose of alcohol does reinstate nicotine-seeking (Lê et al., 2010). Finally, systemic co-administration of methamphetamine and nicotine produces a conditioned place aversion in rats (Briggs et al., 2018), whereas pretreatment with either amphetamine or morphine increases the rewarding properties of nicotine as shown with lowered intracranial self-stimulation thresholds (Huston-Lyons et al., 1993). These studies further emphasize the need to consider use patterns and dose in interpretation of polydrug use effects.

Opioids

Similar to psychostimulant polydrug studies, sequential use of heroin and cocaine has not been found to alter heroin self-administration or reinstatement of heroin-seeking (Crummy et al., 2020). Although alcohol pretreatment can prevent the long-term reconsolidation of preference for morphine in drug-paired contexts (Zhu et al., 2020), adolescent alcohol exposure enhances the development of a morphine CPP (Molet et al., 2013). This finding indicates that the long-term effects of adolescent alcohol exposure are generalizable to multiple drug classes. Interestingly, co-administration of morphine and THC prevents the development of the analgesic tolerance that normally accompanies long-term exposure to either drug alone (Cichewicz and McCarthy, 2003; Cox et al., 2007; Smith et al., 2007). In addition, the analgesic effects of THC and oxycodone co-administration are additive to oxycodone alone (Nguyen et al., 2019). Moreover, administration of either THC or both THC and CBD attenuates naloxone-precipitated withdrawal without impacting the development of a morphine CPP (Lichtman et al., 2001; Valverde et al., 2001). These data suggest a potential role for cannabinoids in regulating a physical dependence to opioids without altering their reinforcing properties. In support of this, repeated THC administration has no effect on breakpoint during a PR test of heroin self-administration (Solinas et al., 2004, but see Nguyen et al., 2019) or relapse to heroin-seeking, although it produces a small reduction in both heroin (Maguire and France, 2016) and oxycodone (Nguyen et al., 2019) intake in fixed-ratio self-administration sessions. The effects of opioid and cannabis polydrug use, however, appear to be dose-dependent as systemic administration of THC prior to heroin self-administration reduces responding for large doses of heroin, but has no effect on responding for lower doses in both monkeys and rats (Solinas et al., 2004; Maguire and France, 2016).

Cannabinoids

As with psychostimulants and opioids, administration of nicotine with THC augments the effects of either drug alone when measured in tests of locomotion, analgesia, and hypothermia (Valjent et al., 2002). In addition, THC and nicotine co-administration exacerbates the somatic symptoms of THC withdrawal (Valjent et al., 2002). However, after repeated nicotine treatment and 2 weeks of drug abstinence, nicotine re-administration attenuates THC-induced decreases in locomotor activity, increases in anxiety measures (when assessed in the elevated-plus maze), and changes in social interaction (Manwell et al., 2019). These findings suggest that nicotine enhances the negative symptoms of THC when administered concurrently or in close temporal proximity. Although nicotine pretreatment enhances the rewarding effects of subthreshold doses of THC (Ponzoni et al., 2019), cocaine pretreatment heightens THC-induced anxiogenic behaviors (Panlilio et al., 2007). Cannabis and alcohol polydrug use is relatively common in humans, and individuals report reduced alcohol consumption when cannabis is available (Mello and Mendelson, 1978), suggesting a role for cannabinoids in alcohol intake. However, drug-induced cognitive and physical impairments in humans, as assessed in a driving simulation, were found to be more severe after use of THC and alcohol compared to either drug alone (Downey et al., 2013). Conducting polydrug studies of combinations of THC or CBD with other drug classes is therefore necessary to understand the differential effects resulting from these drugs.

Unfortunately, due to long-term restrictions on cannabis research in the US and past difficulties in modeling cannabis use with self-administration models in rodents (Panlilio et al., 2015), much less is known about the impact of cannabis relative to other drugs. The development of novel methods of cannabis self-administration in animals, including oral self-administration of Δ9-tetrahydrocannabinol (THC)-containing gelatin (Kruse et al., 2019), self-administration of vaporized THC and cannabidiol (CBD) (Freels et al., 2020), and intravenous self-administration of THC and CBD (Neuhofer et al., 2019) will help facilitate the preclinical study of cannabis use disorder, as well as enable us to better understand the consequences of polydrug use involving cannabis.

Alcohol

Polydrug use of alcohol and nicotine produce mixed phenotypes in relation to addiction behaviors. For example, pre-exposure to alcohol or simultaneous access to both alcohol and nicotine increases alcohol self-administration, but not when nicotine is administered prior to alcohol each day (Lê et al., 2010, 2014). Chronic nicotine treatment also enhances alcohol preference, an effect that persists through nicotine withdrawal (Blomqvist et al., 1996). However, although access to nicotine impairs extinction learning to alcohol responding, it has no effect on reinstatement of drug-seeking, as rats respond similarly on alcohol and nicotine-associated levers following a priming injection of nicotine (Lê et al., 2010). Nonetheless, another study found that priming doses of alcohol, but not nicotine, were capable of reinstating alcohol-seeking following self-administration of both nicotine and alcohol (Sentir et al., 2020). Studies have not systematically examined the effects of other drugs on alcohol use and addiction.

Neurobiology of Addiction

The development and maintenance of addiction behaviors arises in part from maladaptive neuroplasticity within the neural circuits responsible for decision-making, learning, motivation, and reward processing. In particular, alterations in the cortico-basal ganglia-thalamic (C-BG-T) network are known to contribute to drug-taking and drug-seeking behaviors, as well as the persistence of SUDs (Koob and Volkow, 2016). The C-BG-T is a heavily interconnected network that integrates sensory and interoceptive cues to drive motivated behavioral output. The striatum, which serves as an interface of the C-BG-T, receives extensive glutamatergic input from cortical (e.g. prefrontal) and subcortical (e.g. amygdala, hippocampus, thalamus) regions, along with dopaminergic input from the midbrain [substantia nigra (SN)/ventral tegmental area (VTA)] (Gerfen and Surmeier, 2011; Calabresi et al., 2014). Integration of glutamatergic and dopaminergic inputs with local inhibition in the striatum contributes to the initiation or suppression of behavioral output, and imbalanced signaling between the two striatal output pathways (i.e. the direct and indirect) can drive addictive behaviors (Kravitz et al., 2010; Ferguson et al., 2011; O’Neal et al., 2019). It is beyond the scope of the current review to fully explore all of the neurobiological changes that occur in the C-BG-T with drug use. We will instead focus on one microcircuit within the C-BG-T that is central to the acute effects of drugs with addictive potential, and in some of the persistent changes that develop following long-term drug use: The prefrontal cortex (PFC) – nucleus accumbens (NAc) – VTA network (Figure 3). Following a review of the microcircuitry and connectivity of these regions, we will discuss disruptions that occur within this network following both acute and long-term exposure to different classes of drugs, with emphasis on the similarities and/or differences of effects relative to polydrug combinations.

Figure 3. Neural circuitry targeted by potentially addictive drugs. Simplified schematic emphasizing local and distal connections between the PFC, NAc, and VTA that are targeted by potentially addictive drugs. Left: NAcMSNs receive excitatory glutamatergic inputs from PFCGLU neurons, dopaminergic inputs from VTADA neurons, and inhibitory GABAergic inputs from other NAcMSNs. Right: VTADA neurons are maintained under tonic inhibition by local VTAGABA interneurons and NAcMSNs and receive excitatory inputs from PFCGLU neurons. DA: dopamine; GLU: glutamate; MSN: medium spiny neuron; NAc: nucleus accumbens; PFC: prefrontal cortex; VTA: ventral tegmental area.

Prefrontal Cortex

The PFC is centrally involved in reward learning, decision-making, and outcome valuation (Garcia et al., 2018). It is a highly heterogenous structure, which adds to the complexity in understanding its role in cognition, as well as how its dysregulation contributes to drug use and addiction. In general, the medial prefrontal cortex (mPFC) – encompassing the anterior cingulate (ACC), prelimbic (PrL), and infralimbic (IL) cortices – regulates motivation and seeking of both natural and drug rewards via excitatory glutamatergic projections to the NAc and VTA (Koob and Volkow, 2016). Notably, projections from the mPFC to the NAc are topographically organized, with the PrL innervating the NAc core and the IL innervating the NAc shell (Brog et al., 1993). In contrast, the orbitofrontal cortex projects more heavily to the dorsal striatum and SN and is primarily involved in outcome and probability valuation (Padoa-Schioppa and Conen, 2017). Hypoactivity in the PFC contributes to drug craving and seeking despite negative consequences in preoccupation stages of addiction, with dysregulated connectivity to the striatum and VTA contributing to cue sensitivity and motivated drug-taking (Volkow and Boyle, 2018). The PFC is comprised of six layers, each with unique connectivity patterns and distinct cell types. Specifically, a majority of PFC neurons are large pyramidal output cells (75%), as well as several subtypes of interneurons (∼25%) (Santana and Artigas, 2017). Pyramidal cells in layers II/III send local projections within cortex, while those in layers V-VI send projections throughout the C-BG-T, including to the striatum, midbrain, amygdala, hippocampus, and thalamus (Gabbott et al., 2005; Santana and Artigas, 2017). Pyramidal cells can be further subdivided based on physiology and connectivity (Morishima and Kawaguchi, 2006; Brown and Hestrin, 2009; Reiner et al., 2010; Shepherd, 2013; Kim E. J. et al., 2015). Recent studies have begun to characterize the anatomical, electrophysiological, and molecular profiles of each of these cell types (Kalmbach et al., 2015; Kim E. J. et al., 2015; Saiki et al., 2018; Chen et al., 2019; Winnubst et al., 2019), though how they each regulate behavior remains poorly understood. Finally, the PFC contains multiple populations of interneurons that heavily regulate cortical output via projections to both pyramidal neurons and interneurons (van Versendaal and Levelt, 2016; Batista-Brito et al., 2017).

Nucleus Accumbens

The striatum is a heterogeneous structure comprised primarily of two interspersed populations of GABAergic medium spiny neurons (MSNs) that can bidirectionally regulate behavioral output. Direct pathway MSNs (dMSNs) express dopamine D1-like (D₁) receptors and the neuropeptides dynorphin and substance P, project directly to the midbrain, and can promote behavioral output by serving as a “go” signal. Conversely, indirect pathway MSNs (iMSNs) express dopamine D2-like (D₂) receptors and the neuropeptide enkephalin, project indirectly to the midbrain via the pallidum (GPe and VP), and can suppress behavioral actions by serving as a “stop” signal (Kravitz et al., 2010; Gerfen and Surmeier, 2011). Drug use promotes increased phasic dopamine from D₁ activation, prompting reward attribution to drug use during binge/intoxication phases of the addiction cycle, conditioning, and incentive salience attribution to drug-taking contexts (Volkow et al., 2011; Koob and Volkow, 2016). The striatum contains dorsal and ventral compartments, with further subdivisions based on connectivity and function. The ventral striatum – comprised of the olfactory tubercule, NAc core, and NAc shell – receives dopaminergic modulation from the VTA and glutamatergic input from the PFC, as well as thalamic, hippocampal, and amygdala nuclei (Li et al., 2018). In general, the ventral striatum regulates motivated behavior and reward learning. However, it has been hypothesized that an ascending loop between the ventral and the dorsal striatum facilitates information consolidation during learning, whereby habitual behaviors transition from the ventral striatum to the dorsal striatum, contributing to compulsive drug-seeking (Dobbs et al., 2016; Koob and Volkow, 2016; Burke et al., 2017). Importantly, while dMSNs and iMSNs have historically been differentiated by downstream targets and expression of dopamine receptors, ventral striatal dMSNs send collaterals to the VP (Cazorla et al., 2014; Kupchik et al., 2015), and D₁ and D₂ receptors are co-expressed to some degree in the NAc core (6–7%), and NAc shell (12–15%) (Bertran-Gonzalez et al., 2008; Gagnon et al., 2017). In addition to MSNs, the striatum contains large, tonically active cholinergic interneurons and multiple subtypes of GABAergic interneurons with distinct electrophysiological properties and peptide expression patterns (Burke et al., 2017). MSNs also receive cholinergic modulation from other projections (Dautan et al., 2014), though the relevance of these inputs to local or network dynamics and the role of cholinergic striatal neurons in the addiction cycle remains uncertain. Interestingly, each MSN receives ∼5000–15000 excitatory inputs in addition to ∼1200–1800 GABAergic inputs from other MSNs, therefore, modulation of MSN activity via cholinergic and dopaminergic inputs appears necessary for signal integration and effective synaptic plasticity (Moyer et al., 2007; Burke et al., 2017). Indeed, MSNs exhibit bi-stability, residing almost exclusively in either a down-state (−80 mV) or an up-state (−50 mV, near threshold) in the absence of external input. In addition, the maintenance of bi-stability and intrinsic excitability relies on the activity of cation channels that are under robust dopaminergic and cholinergic modulation (Plenz and Kitai, 1998; Moyer et al., 2007). Maintenance of intrinsic excitability within MSNs is critical for normal regulation of behavioral output, and dysregulation of striatal microcircuitry contributes to the development and expression of addiction behaviors (Bock et al., 2013; Stefanik et al., 2013; O’Neal et al., 2019).

Ventral Tegmental Area

The VTA sends dopaminergic projections to cortical, striatal, and subcortical (e.g. hippocampus, amygdala, thalamus) areas to modulate C-BG-T network activity (Koob and Volkow, 2010). Dopaminergic projections from the VTA to the NAc regulate goal-directed behaviors and have been heavily implicated in the binge/intoxication phase of SUDs (Koob and Volkow, 2016). Recently identified subtypes of VTA^DA neurons with unique molecular profiles preferentially project to the NAc core or NAc shell, and regulate reward learning or motivation, respectively. However, co-activation of both populations appears to be necessary for robust reinforcement (Heymann et al., 2019). VTA^DA neurons receive dense glutamatergic input from the PFC and several midbrain structures, as well as GABAergic input from a variety of sources (reviewed in Morales and Margolis, 2017), including the NAc and local VTA^GABA neurons. VTA^GABA neurons maintain tonic DA levels via inhibition of VTA^DA neurons, and brief disinhibition of VTA^DA neurons results in phasic DA release into the NAc. VTA^GABA neurons receive glutamatergic input from the PFC and a number of subcortical nuclei, as well as GABAergic input from throughout the brain, including the NAc (Morales and Margolis, 2017). The VTA also contains glutamatergic neurons that project to striatal interneurons (Brown et al., 2012; Qi et al., 2016), though the inputs to and behavioral relevance of these neurons is unknown. Notably, the VTA contains subpopulations of DA neurons that can release glutamate and/or GABA (Kim J. I. et al., 2015; Berrios et al., 2016; Morales and Margolis, 2017), allowing the VTA to modulate local C-BG-T activity at multiple levels and time scales. Within the NAc, the activity of VTA^DA neurons is modulated by dopamine D₂ autoreceptors on VTA^DA terminals and cholinergic interneurons (Morales and Margolis, 2017).

Mechanisms of Addiction: Acute Drug Effects

A unique feature of all potentially addictive drugs is the ability to reinforce the binge/intoxication phase of the addiction cycle via evoked phasic DA release into the NAc, yet the underlying mechanisms vary across drugs (Figure 4; Volkow and Boyle, 2018). Psychostimulants disrupt DA reuptake into VTA^DA terminals (Pontieri et al., 1995), nicotine and alcohol directly activate VTA^DA neurons (Pidoplichko et al., 1997; Brodie et al., 1999), and opioids and cannabinoids disinhibit VTA^DA neurons (Pidoplichko et al., 1997; Cheer et al., 2000). The synergistic and antagonistic interactions between different drugs within the C-BG-T lends complexity to the study of polydrug use, and little is known about the mechanisms underlying the acute effects of specific polydrug combinations. However, in vivo extracellular recordings in rats that alternatively self-administered cocaine and heroin in the same session found that only ∼20% of PFC and NAc neurons responded similarly to both drugs (Chang et al., 1998), indicating divergent engagement of the C-BG-T by these drugs. Thus, examining the synergistic and antagonistic mechanisms of different drugs can guide our understanding of how specific polydrug combinations may disrupt C-BG-T network dynamics and contribute to the manifestation of addiction behaviors.

Figure 4. Primary mechanisms of action of potentially addictive drugs. Potentially addictive drugs increase DA release into the NAc, but different drugs act via distinct mechanisms. Top: Opioids and cannabinoids disinhibit VTADA neurons via presynaptic inhibition of VTAGABA and NAcMSN inputs through four notable mechanisms: Inhibition of VG Ca2+ channels, activation of GIRKs, inactivation of AC, and inhibition of GABA release. Nicotine activates VTADA neurons via direct activation of somatodendritic nAChRs and activation of presynaptic PFCGLU inputs. Alcohol directly activates VTADA cell bodies, but the mechanism is not understood. Bottom: Psychostimulants impair DA reuptake by blocking DAT (cocaine) or reversing the activity of DAT and facilitating DA release (amphetamine), leading to increased dopaminergic tone in the NAc. Alcohol also targets PFCGLU inputs to the NAc, but the net effect on PFCGLU activity is unknown. AC: adenylyl cyclase; ATP: adenosine triphosphate; cAMP: cyclic adenosine monophosphate; CB1R: cannabinoid 1 receptor; DA: dopamine; DAT: dopamine transporter; D1R: dopamine D1 receptor; GABA: gamma- aminobutyric acid; GIRK: G protein-coupled inwardly rectifying K+ channel; GLU: glutamate; mGluR: metabotropic glutamate receptor; μOR: mu opioid receptor; NAc: nucleus accumbens; nACh: nicotinic acetylcholine receptor; NMDA: N-methyl-D-aspartate receptor; PFC: prefrontal cortex; SNARE: soluble N-ethylmaleimide-sensitive factor attachment protein receptor; VG Ca2+: voltage-gated Ca2+ channel; VTA: ventral tegmental area.

Psychostimulants

All psychostimulants directly enhance striatal DA release via disruption of dopamine transporter (DAT) activity (Pontieri et al., 1995), though they do so via distinct mechanisms. Cocaine blocks DAT-mediated reuptake of DA, while amphetamine reverses DAT activity and induces DA release from VTA^DA terminals (Hyman et al., 2006; Kelly et al., 2008). Psychostimulants also acutely increase glutamate transmission in the PFC, NAc, and VTA (Reid and Berger, 1996; You et al., 2007; Shin et al., 2016), indicating broad increases in activity throughout the C-BG-T network following psychostimulant use. The combination of enhanced glutamatergic and dopaminergic input to the NAc facilitates the transition of MSNs to the up- state, and activation of DA-dependent signaling cascades (McFarland and Kalivas, 2001; Feltenstein and See, 2008). For example, psychostimulants acutely increase activation of the immediate early gene Fos in striatal dMSNs and iMSNs (Badiani et al., 1999; Uslaner et al., 2001; Ferguson and Robinson, 2004). Fos encodes a number of proteins, including ΔFosB, that have been widely implicated in addiction pathology, and enhanced activation in the NAc is thought to contribute to long-term disruptions in normal C-BG-T activity (Nestler et al., 2001). Co-administration of nicotine, alcohol, or heroin enhances psychostimulant-induced DA release into the NAc (Bunney et al., 2001; Mello et al., 2014; Pattison et al., 2014), though specific combinations do so via divergent mechanisms. For example, administration of cocaine and nicotine simultaneously activates VTA^DA neurons and disrupts DA reuptake (Mello et al., 2014; De Moura et al., 2019), resulting in a greater magnitude of DA release into the NAc than that evoked from either drug alone. Notably, some polydrug combinations that include psychostimulants have divergent and opposing effects on the C-BG-T circuit. Co-administration of alcohol with cocaine induces hepatic production of cocaethylene, which can target DAT and presynaptic D₂ autoreceptors to amplify DA release (Bunney et al., 2001). However, alcohol also prevents cocaine-induced glutamate transmission in the NAc core (Stennett et al., 2020). Lastly, polydrug studies with psychostimulants have identified an exacerbation of drug-induced cellular toxicity compared to psychostimulant use alone. Specifically, co-administration of heroin with cocaine decreases metabolic activity, increases intracellular Ca²⁺ signaling, and decreases mitochondrial membrane potential (Cunha-Oliveira et al., 2008). Collectively, these effects contribute to enhanced caspase 3-dependent apoptotic activity and subsequent cell death compared to either drug alone (Cunha-Oliveira et al., 2008).

Nicotine

Unlike psychostimulants, nicotine enhances DA release via activation of nicotinic acetylcholine (nACh) receptors within the VTA. nACh receptors are non-selective cation channels, with permeability to Na⁺, K⁺, and Ca²⁺, and their activation leads to depolarization and enhanced neurotransmitter release (Benowitz, 2009). The reinforcing effects of nicotine are primarily due to nicotine’s activity on somatodendritic nACh receptors located on VTA^DA neurons and on presynaptic nACh receptors located on PFC^GLU inputs (Nisell et al., 1994). Acute nicotine exposure activates both presynaptic PFC inputs (Fu et al., 2000; Picciotto and Kenny, 2013) and VTA^DA cell bodies (Calabresi et al., 1989; Pidoplichko et al., 1997), triggering phasic DA release into the NAc (Dani and De Biasi, 2001; Picciotto and Mineur, 2014). Notably, nicotine-evoked DA release is occluded with blockade of glutamate receptors or activation of GABA_B receptors (Fadda et al., 2003; Kosowski et al., 2004), highlighting the extensive regulation of VTA^DA neuron activity by local VTA microcircuitry. Studies have not examined how the acute effects of nicotine are changed by polydrug use.

Opioids

Although opioids are differentiated by their origin, potency, and receptor bias factor (Schmid et al., 2017), all opioids exert their rewarding effects via activation of mu opioid (μO) receptors. μO receptors are expressed both somatodendritically and axonally (Arvidsson et al., 1995), but the primary mechanism of opioid-induced DA release relies on presynaptic inhibition of VTA^GABA neurons. μO receptors are inhibitory G protein-coupled receptors (GPCRs), and their activation reduces neuronal excitability via four mechanisms: (1) Gα-mediated inhibition of cAMP-dependent signaling cascades (e.g. PKA, CREB), (2) Gβγ-mediated activation of G protein-coupled inwardly rectifying K⁺ (GIRK) channels, (3) Gβγ-mediated inactivation of voltage-gated Ca²⁺ channels, and (4) Gβγ-mediated inhibition of SNARE-dependent vesicle release (Bourinet et al., 1996; Blanchet and Lüscher, 2002; Blackmer et al., 2005; Al-Hasani and Bruchas, 2011; Zamponi and Currie, 2013). Acute exposure to opioids inhibits VTA^GABA neurons (Johnson and North, 1992; Corre et al., 2018), resulting in subsequent disinhibition of VTA^DA neurons and phasic DA release into the NAc (Hemby et al., 1995; Pontieri et al., 1995). Nonetheless, although opioids facilitate DA release into the striatum, whether this DA transmission is necessary for opioid reward remains a point of debate (Badiani et al., 2011). Opioids activate VTA^DA neurons in vivo and increase DA in the NAc (Di Chiara and Imperato, 1988; Johnson and North, 1992), but neither lesions of the NAc nor systemic antagonism of DA receptor blockades have an effect on opioid reward (Ettenberg et al., 1982; Van Ree and Ramsey, 1987; Olmstead and Franklin, 1997; Sellings and Clarke, 2003). Given the divergent mechanism of action for opioids compared to psychostimulants and nicotine, it is not surprising that co-administration of these drugs augments the acute effects of opioids. Indeed, simultaneous administration of opioids and psychostimulants produces an additive increase in DA release in the NAc, and prolongs elevated levels of DA and its metabolites, DOPAC and HVA (Zernig et al., 1997). Similarly, simultaneous administration of opioids and nicotine enhances opioid-evoked DA release in the NAc and dorsal striatum (Vihavainen et al., 2008). Cross-tolerance to opioid-mediated analgesia has also been shown following pre-exposure to nicotine and cannabis (Schmidt et al., 2001), and chronic nicotine treatment dose-dependently reduces analgesic tolerance to opioids in a nACh receptor-dependent manner (Haghparast et al., 2008; De Moura et al., 2019). Interestingly, pretreatment with Ca²⁺ channel blockers or naloxone prevents this tolerance, suggesting a complex pharmacological interaction between opioids and nicotine (Biala and Weglinska, 2006). Notably, cross-tolerance to opioid analgesia is mediated via divergent mechanisms for different polydrug combinations. Nicotine and opioid cross-tolerance is mediated by μO and nACh receptors (Haghparast et al., 2008; De Moura et al., 2019) while cannabinoid and opioid cross-tolerance is mediated by μO receptors and cannabinoid 1 (CB₁) receptors (Pugh et al., 1994, 1996).

Cannabinoids

Cannabis contains two principal cannabinoids with varying affinity for cannabinoid (CB) receptors: THC is a partial agonist with moderate affinity for both CB₁ and CB₂ receptors whereas CBD has extremely low affinity for CB₁ and CB₂ receptors and signals through an unknown mechanism (Pertwee, 2008). Interestingly, pretreatment with a range of CBD doses has no effect on THC self-administration (Wakeford et al., 2017), indicating non-overlapping signaling pathways for each cannabinoid. CB₁ receptors are expressed on presynaptic terminals throughout the CNS and are responsible for the psychoactive effects of cannabis, while CB₂ receptors are primarily expressed on immune cells of the CNS and PNS and are primarily responsible for the antinociceptive and anti-inflammatory effects of cannabis (Pertwee, 2008). Both subtypes signal through inhibitory GPCR signaling pathways (similar to opioids), and activation of the receptors results in presynaptic inhibition (via activation of GIRK channels and inhibition of VG Ca²⁺ channels) and downregulation of cAMP-dependent signaling cascades. CB₁ receptors are heavily expressed on presynaptic terminals of VTA^GABA neurons as well as NAc dMSNs that target VTA^DA neurons (Szabo et al., 2002; Lupica et al., 2004), and activation of CB₁ receptors reduces GABA-mediated inhibitory postsynaptic currents in VTA slices (Cheer et al., 2000). CB₁-mediated disinhibition of VTA^DA neurons results in an increase in burst firing (French et al., 1997; Diana et al., 1998) and subsequent release of DA into the NAc (Ton et al., 1988; Chen et al., 1990; Fadda et al., 2006), and these effects are blocked by systemic or intra-VTA naloxone (Chen et al., 1990; Tanda et al., 1997). Similar to other drugs, the acute effects of cannabinoids on C-BG-T network activity are broad and engage multiple neurotransmitter systems. For example, acute THC increases both DA and glutamate signaling in the NAc and PFC (Pistis et al., 2002a, b), but decreases GABA signaling in the VTA and PFC (Cheer et al., 2000; Pistis et al., 2002a). Collectively, these alterations in signaling reduce inhibitory feedback within the C-BG-T and facilitate behavioral output. Alcohol consumption prior to cannabis use enhances plasma THC levels and increases self-reported euphoria in humans (Lukas and Orozco, 2001), indicating synergistic effects between the two drugs. Moreover, simultaneous administration of cannabinoids with psychostimulants or opioids enhances activation of VTA^DA neurons (Pistis et al., 2004), and simultaneous administration of THC and nicotine increases cFos activation throughout the C-BG-T (Valjent et al., 2002).

Alcohol

Despite its widespread use, much less is known about the mechanisms underlying the acute effects of alcohol. Alcohol activates dissociated VTA^DA neurons (Brodie et al., 1999) and induces DA release into the NAc (Weiss et al., 1993; Pontieri et al., 1996; Lecca et al., 2006), similar to other potentially addictive drugs. However, GABAA receptors are known to play a central role in the effects of alcohol (Hyytiä and Koob, 1995; Lobo and Harris, 2008). For example, alcohol potentiates GABA_A signaling both in cortical slices and neuronal cultures (Aguayo, 1990; Reynolds and Prasad, 1991; Reynolds et al., 1992; Tatebayashi et al., 1998). Given that NAc dMSNs selectively inhibit VTA^GABA neurons via GABA_A-mediated signaling (Edwards et al., 2017), it is possible that alcohol facilitates phasic DA release from VTA^DA neurons via inactivation of local VTA^GABA neurons. In support of this hypothesis, alcohol inhibits VTA^GABA neurons (Steffensen et al., 2009), and intra-VTA infusion of GABA_A agonists dose-dependently increase DA release (Kalivas et al., 1990). In addition to GABA, the acute effects of alcohol are dependent on glutamatergic signaling within the C-BG-T (Grant and Colombo, 1993; Krystal et al., 1994). Alcohol increases glutamate release in the NAc and VTA via activation of presynaptic D₁ receptors (Nie et al., 1994; Xiao et al., 2009), suggesting that alcohol engages a feedforward loop for activation of VTA^DA neurons. Polydrug use with alcohol produces synergistic effects throughout the C-BG-T, likely as a result of alcohol’s unique pharmacological profile. For example, chronic pretreatment with nicotine enhances acute alcohol-induced DA release in the NAc (Johnson et al., 1995; Blomqvist et al., 1996), and elevated levels of DA, DOPAC, and HVA persist for over an hour (Tizabi et al., 2002, 2007; Ding et al., 2012). Additionally, alcohol and nicotine co-administration acutely increase production of BDNF and GDNF in the NAc (Truitt et al., 2015), along with increases in glutamatergic signaling in the VTA and PFC (Deehan et al., 2015; Engle et al., 2015). Notably, this wide activation of the C-BG-T network is absent following administration of either drug alone, demonstrating a unique mechanism of action for alcohol and nicotine polydrug use.

Mechanisms of Addiction: Long-Term Alterations

Long-term use of psychostimulants, nicotine, opioids, cannabinoids, and alcohol results in widespread and disparate changes throughout the C-BG-T network, yet there are notable alterations that are shared across drugs (Figure 5). These long-term adaptations contribute to transitions from binge/intoxication phases to withdrawal and negative affect, followed by preoccupation and compulsive drug-seeking (Koob and Volkow, 2016). For example, acute withdrawal produces a transient reduction in tonic DA levels in the NAc (Weiss et al., 1992; Diana et al., 1993, 1998; Hildebrand et al., 1998). This is followed by a persistent increase in excitability of VTA^DA neurons (Mansvelder and McGehee, 2000; Bloomfield et al., 2016; Creed et al., 2016; Langlois and Nugent, 2017; You et al., 2018), which contributes to enhanced cue-evoked phasic DA release during abstinence (Volkow et al., 2011). Conversely, withdrawal also produces a persistent reduction in long-term depression (LTD) and intrinsic excitability in NAc MSNs, as well as a reduction in striatal D₂ receptor binding (Trifilieff and Martinez, 2013). As described earlier, striatal D₂ receptors are primarily expressed on iMSNs, can serve as a “stop” signal on the C-BG-T circuit, and reduced D₂ availability has been ubiquitously linked to a range of addictive diseases, including drug addiction and obesity (Volkow and Morales, 2015; Kravitz et al., 2016; Friend D.M. et al., 2017). Finally, these drugs all drive a persistent increase in ΔFosB expression in cortical pyramidal cells and NAc dMSNs (Lobo et al., 2013), which has been linked to drug-seeking during abstinence (Nestler et al., 2001). Importantly, although chronic use of any of these drugs results in a multitude of other transient and/or persistent changes across the C-BG-T network, it is beyond the scope of the current review to provide an exhaustive summary. Rather, the focus will be on changes in plasticity, morphology, and connectivity within the VTA, NAc, and PFC following both single and polydrug use, with an emphasis on how the antagonistic and synergistic effects of these drugs can differentially disrupt C-BG-T network dynamics.

Figure 5. Persistent disruptions in synaptic and structural plasticity caused by long-term use of potentially addictive drugs. Studies are organized by cell type (columns) and type of disruption (rows), with symbols depicting the net change in plasticity and color depicting which drug was tested. Data from Bonci and Williams (1996), Kang et al. (1996, 1998), Robinson and Kolb (1997), Robinson and Kolb (2004), Badiani et al. (1999), Mansvelder and McGehee (2000), Dahchour and De Witte (2000), Uslaner et al. (2001), Brown and Kolb (2001), Robinson et al. (2002), Saal et al. (2003), Amantea et al. (2004), Hamilton and Kolb (2005), Nasif et al. (2005), Kolb et al. (2006), Kolb et al. (2018), Huang et al. (2007), Zhou et al. (2007), Van Den Oever et al. (2008), Kalivas et al. (2009), Niehaus et al. (2010), Russo et al. (2010), Bowers et al. (2010), Spiga et al. (2010), Levine et al. (2011), Dacher and Nugent (2011), Kroener et al. (2012), Lobo et al. (2013), Trifilieff and Martinez (2013), Mello et al. (2014), Peterson et al. (2015), Bloomfield et al. (2016), Creed et al. (2016), Ehlinger et al. (2016), Hearing et al. (2016), Morud et al. (2016), Friend L. et al., 2017), Langlois and Nugent (2017), Edwards et al. (2017), Spencer et al. (2018), You et al. (2018), Hwang and Lupica (2019), Kruse et al. (2019), McDevitt et al. (2019), Neuhofer et al. (2019), Pickel et al. (2019), and Ponzoni et al. (2019).

Psychostimulants

Psychostimulants produce long-term disruptions in glutamate homeostasis and alterations in neuronal morphology throughout the C-BG-T network (Kalivas, 2009; Badiani et al., 2011). Repeated cocaine administration weakens GABA_A-mediated inhibition of prelimbic (PrL) pyramidal neurons, increasing their excitability and augmenting excitatory drive to the NAc (Nasif et al., 2005; Huang et al., 2007). Similarly, chronic cocaine increases glutamatergic input to the VTA (Saal et al., 2003; Bowers et al., 2010) and weakens GABA_B-mediated inhibition of VTA^DA neurons (Bonci and Williams, 1996; Edwards et al., 2017), leading to a facilitation of VTA^DA neuron activity. Additionally, psychostimulants generate silent synapses on dMSNs via synaptogenesis (Boudreau et al., 2007; Graziane et al., 2016) and depress glutamate release from PrL inputs to the NAc core (Bamford and Wang, 2019), weakening striatal output. Importantly, cocaine-silenced synapses on dMSNs can be unsilenced during withdrawal via recruitment of AMPA receptors (Boudreau et al., 2007; Graziane et al., 2016), and a low dose psychostimulant challenge restores glutamate release into the NAc core (Boudreau et al., 2007; Bamford and Wang, 2019), suggesting the promotion of allostasis. Finally, repeated psychostimulant exposure increases expression of ΔFosB in PFC pyramidal neurons and NAc dMSNs (Perrotti et al., 2005), which contributes to an increase in dendritic branching that persists for at least 30 days of abstinence (Robinson and Kolb, 1997, 2004; Russo et al., 2010). Studies have not examined how these effects of psychostimulants are changed by polydrug use.

Nicotine

Similar to psychostimulants, long-term nicotine administration reduces GABA_B-mediated signaling in the mPFC and NAc (Amantea et al., 2004), dampening inhibitory drive onto the C- BG-T network. Additionally, nicotine enhances glutamatergic input to VTA^DA neurons (Mansvelder and McGehee, 2000; Saal et al., 2003), and sensitizes evoked DA release into the NAc (Benwell et al., 1995). Repeated nicotine use also leads to an upregulation of nicotinic acetylcholine receptors throughout the C-BG-T, including the PFC and midbrain (Marks et al., 1983; Benwell et al., 1988; Breese et al., 1997). Conversely, chronic nicotine exposure weakens both glutamatergic and GABAergic inputs to the NAc during abstinence, potentially via enhanced sensitivity of NAc D₂ receptors (Morud et al., 2016). Because NAc output is heavily regulated via local GABAergic microcircuitry, increased activity of iMSNs would likely dampen overall NAc output. This is supported by the upregulation of Ca²⁺-permeable AMPA receptors in the NAc that are capable of conducting Ca²⁺ in the absence of NMDA receptor activation (Ponzoni et al., 2019). Long-term nicotine use also produces a persistent increase in mPFC pyramidal cell and NAc MSN dendritic branching (Brown and Kolb, 2001; Hamilton and Kolb, 2005; Ehlinger et al., 2016), similar to psychostimulants. Moreover, adolescent nicotine use enhances synaptic pruning, microglial activation, and inflammatory cytokine expression throughout the C-BG-T via a D₂ receptor-mediated mechanism (Linker et al., 2020). Repeated co-administration of psychostimulants and nicotine augments long-term potentiation (LTP) induction and ΔFosB in the NAc core (Levine et al., 2011; Mello et al., 2014). Interestingly, these effects are absent following sequential administration of these drugs (Levine et al., 2011), indicating a unique pharmacological profile for concurrent nicotine and psychostimulant administration.

Opioids

Similar to psychostimulants and nicotine, long-term exposure to opioids strengthens glutamatergic input to VTA^DA neurons (Bonci and Williams, 1996; Saal et al., 2003). They also weaken GABAergic inhibition of VTA^DA neurons (Mansvelder and McGehee, 2000; Niehaus et al., 2010; Dacher and Nugent, 2011) via a reduction of dMSN-mediated GABA_B inhibition (Bonci and Williams, 1996). This effect is largely driven by MSNs in the NAc (Yang et al., 2018), indicating a persistent disruption in NAc output. Unlike psychostimulants and nicotine, however, long-term exposure to opioids decreases dendritic branching and spine density in NAc MSNs and mPFC pyramidal cells (Badiani et al., 1999; Robinson et al., 2002; Van Den Oever et al., 2008). Within the NAc, long-term opioid exposure weakens glutamatergic input to NAc shell iMSNs (Hearing et al., 2016; McDevitt et al., 2019) and induces silent synapses on iMSNs via AMPA internalization (Graziane et al., 2016). Moreover, during withdrawal, opioid-generated silent synapses on iMSNs are eliminated (Graziane et al., 2016), and the intrinsic excitability of iMSNs is weakened (McDevitt et al., 2019). Given that MSNs create a dense network of lateral inhibition within the NAc, with ∼30% of iMSNs synapsing onto other iMSNs or dMSNs (Taverna et al., 2008), it is possible that opioid-induced disruptions in iMSN signaling could disrupt local NAc microcircuitry and facilitate aberrant C-BG-T network dynamics. Studies have not examined how these effects of opioids are changed by polydrug use.

Cannabinoids

Repeated THC administration weakens glutamatergic signaling from the mPFC to the NAc (Spencer et al., 2018; Hwang and Lupica, 2019; Neuhofer et al., 2019) but strengthens input from the basolateral amygdala and ventral hippocampus to the NAc shell (Hwang and Lupica, 2019), suggesting a rewiring of excitatory input to the NAc following long-term cannabinoid exposure. Notably, repeated THC administration also occludes CB₁-mediated LTD on VTA^GABA neurons (Friend L. et al., 2017), indicating a loss of local inhibitory drive on VTA^DA neurons. Additionally, adolescent THC administration followed by abstinence reduces intrinsic excitability of PrL neurons (Pickel et al., 2019) and weakens glutamatergic input to the VTA (Kruse et al., 2019), indicating the presence of a hypoglutamatergic state induced by cannabinoids. THC exposure also produces long-lasting changes in morphology throughout the C-BG-T, increasing dendritic spine length and branching in the mPFC and NAc shell (Kolb et al., 2006) and dendritic spine density in the NAc shell (Kolb et al., 2018), but reducing the size of VTA^DA neurons (Spiga et al., 2010; Behan et al., 2012). Repeated co-administration of THC and nicotine enhances expression of ΔFosB in the NAc, and acute withdrawal dysregulates glutamatergic input to the NAc and PFC (Ponzoni et al., 2019).

Alcohol

Similar to opioids, extended alcohol exposure decreases dendritic branching and spine density in the NAc shell and mPFC (Zhou et al., 2007; Peterson et al., 2015), and withdrawal from alcohol reduces tonic DA levels in the NAc (Rossetti et al., 1992; Diana et al., 1993). However, alcohol withdrawal also causes distinct changes throughout the C-BG-T network, including reduced GABAergic signaling in the NAc and hippocampus (Kang et al., 1996, 1998; Dahchour and De Witte, 2000) and increased glutamatergic signaling in the NAc and PFC (Dahchour and De Witte, 2000; Kroener et al., 2012). Notably, this increased glutamatergic signaling is due to disrupted glutamate reuptake, rather than enhanced glutamate release (Pati et al., 2016). In vitro polydrug exposure to alcohol and nicotine induces a 2.5-fold increase in caspase-3 activation, elevating apoptotic cascades and driving cell death (Ramlochansingh et al., 2011). Interestingly, however, alcohol or nicotine withdrawal-induced neurodegeneration is less severe following co-administration of both drugs (Oliveira-da-Silva et al., 2010), indicating a unique molecular pathology following nicotine and alcohol polydrug use.

Conclusion

Improving the translatability and mapping of behavioral measures in preclinical models to accurately reflect polysubstance history and dependency in clinical populations is essential. This is particularly true, as human imaging studies are limited by an inability to control for intake history, making behavioral models in other species advantageous for assessing polydrug history under controlled intake conditions. However, these models are limited in their capacity to fully encompass the complex social and environmental contexts that contribute to the unique use patterns for multiple substances with addiction potential. Nevertheless, when designing experiments in clinical or preclinical populations, factors such as time of day of intake, temporal proximity of intake, and environmental preferences for administration must be consideredfor each substance class. For instance, cocaine use is predominantly favored outside of home environments, whereas heroin use is greater in “home” contexts in both humans and rodents (Caprioli et al., 2009; Badiani and Spagnolo, 2013; De Pirro et al., 2018; De Luca et al., 2019). In addition, route of drug administration (e.g. oral ingestion, injection, and inhalation) is especially important to factor into studies, as it leads to unique patterns of polysubstance history that may impact the developmentand severity of addiction behaviors (Roy et al., 2013).

Accounting for temporal patterns of drug use is also essential. Although many studies have focused on simultaneous drug combinations, sequential patterns of polydrug consumption are more frequently reported (Leri et al., 2004; Roy et al., 2013) and produce unique circuit adaptations following acute and repeated drug exposure (Cunha-Oliveira et al., 2008). Understanding sex differences in frequency and pattern of polydrug use (McClure et al., 2017), drug discrimination (Spence et al., 2016), and circuit alterations (Canterberry et al., 2016; Manwell et al., 2019) is also necessary to fully understand the interactions and impacts of polydrug use in clinical populations. Furthermore, although powerful behavioral economic models allow comparisons across drug classes, these experiments must be designed with consideration of different scales of intake and indifference points for drug valuation in order to accurately model parameters and interpret data (Newman and Ferrario, 2019). As these factors, coupled with the interplay of socioeconomic background, social environment, and access strategies, affect frequency and susceptibility to polydrug use (Gjersing et al., 2013; Hernández-Serrano et al., 2018), encompassing these influences in preclinical models is crucial for translational relevance of findings.

It is important to appreciate that novel preclinical paradigms are continuously being developed to model different routes of drug administration and to study relapse under clinically relevant conditions. For instance, recently established models allow for voluntary control over self-administration of vaporized ethanol (de Guglielmo et al., 2017; Kimbrough et al., 2017), cannabis (McLaughlin, 2018; Freels et al., 2020), and nicotine (Marusich et al., 2019). In addition, multiple models for inducing drug abstinence are being introduced. For instance, voluntary abstinence is achieved via pairing of drug-taking with adverse consequences such as shock-lever pairings or electrified barriers placed in front of levers (Krasnova et al., 2014; Venniro et al., 2016), andthrough choice procedures involving a drug versus an alternative reinforcer (e.g. palatable food or social interaction) (Venniro et al., 2016, 2017, 2019). These paradigms are notable for their high translational value and may be powerful for understanding the neural basis of therapies such as contingency management, in which abstinence from drug use results in a monetary reward. Combining these models with behavioral economic paradigms could clarify how the relative value of drugs and alternative reinforcers changes with polydrug use to improve treatment efficacy. Addiction is a complex disease with multiple, highly variable factors contributing to the initiation and maintenance of drug use, as well as relapse. Given the widespread prevalence of polydrug use among drug users, it is critical that we incorporate this variablility into human studies and animal models. This will help to determine if polysubstance use exacerbates SUD severity, if increased SUD severity drives polysubstance use, or if there is a bidirectional relationship between the two. Though challenging, understanding the behavioral, genetic, and environmental contributions to polysubstance use and addiction, as well as the mechanisms that underlie addiction-severity and relapse, will aid in developing efficacious treatment and policy strategies to combat this ongoing public health crisis.

Introduction

Drug addiction represents a complex disorder marked by cycles of drug consumption, withdrawal, abstinence, and a strong desire for the drug leading to relapse. This condition is widespread, affecting millions globally and within the United States, with a significant lifetime prevalence. The financial burden associated with drug addiction is substantial, encompassing costs for treatment, lost productivity, healthcare, and crime. These figures are projected to rise as illicit drug use increases, particularly involving opioids, which have seen a dramatic rise in misuse and associated deaths.

While much research on substance use disorders (SUDs) has focused on individual substances, it is crucial to recognize that many individuals engage in polysubstance use, involving more than one substance either at different times or simultaneously. Limiting studies to single drugs risks overlooking the complex interactions between substances, which can reduce the applicability of research findings and hinder the effectiveness of treatments. Polysubstance use is consistently linked to poorer treatment outcomes, including higher relapse rates and increased mortality. This document aims to integrate current understanding of individual drug use mechanisms with recent research on polysubstance use, highlighting different patterns of combined substance use.

Public Health Trends in Substance Use

Substance addiction poses a severe public health threat, contributing to hundreds of thousands of deaths worldwide each year. Although the impacts of addiction are often discussed in relation to single drugs, misuse commonly involves multiple substances. Individuals with drug dependence typically report using several substances, often in both simultaneous and sequential patterns. The likelihood of developing dependencies on multiple substances is high in clinical populations. While primary dependencies often involve alcohol, opioids, and stimulants, cannabis and cocaine are frequently reported as secondary or tertiary substances. The high prevalence of polysubstance use is particularly concerning due to its impact on the severity of SUDs and treatment success. For instance, a history of polysubstance use is linked to greater unmet healthcare needs, increased risky behaviors, violence, and a higher risk of overdose and death compared to using a single substance. This section provides an overview of public health patterns for both single and polysubstance use, focusing on common substance combinations and usage patterns, acknowledging that comprehensive data for all patterns and definitions may be limited.

Psychostimulants

Psychostimulants represent the second most commonly used class of drugs globally, with millions of users of cocaine and prescription stimulants. While the worldwide prevalence of psychostimulant use has remained stable over recent decades, the number of overdose deaths involving these substances has significantly climbed, particularly in the United States.

A notable characteristic of cocaine and amphetamine users is their predominant engagement in polysubstance use. Reports indicate a high incidence of a polysubstance history among these individuals. Cocaine use is frequently associated with concurrent use of heroin, cannabis, tobacco, and alcohol. Similarly, amphetamine users often engage in polysubstance use with alcohol, tobacco, and cannabis, and in some cases, with heroin and other opioids. Simultaneous use of psychostimulants and opioids, such as "speedball" (cocaine and heroin) or "bombita" (methamphetamine and heroin), is observed. Sequential use is also common, where psychostimulants might be used to alleviate opioid withdrawal symptoms, or opioids to reduce over-excitement after cocaine use. The reported frequencies of concurrent versus simultaneous use vary significantly across studies, reflecting the complexity of identifying precise drug use patterns influenced by demographics, study periods, and definitions of use.

Polydrug use involving psychostimulants presents considerable public health risks. For example, amphetamine users demonstrate a significantly higher likelihood of having concurrent cannabis or cocaine use disorders compared to those without a history of amphetamine use. A substantial proportion of overdose deaths involve both psychostimulants and opioids. Combining cocaine and cannabis has been associated with elevated mortality risks during emergency department visits, and the combination of cocaine and alcohol increases the risk of heart toxicity beyond that of either drug alone.

Nicotine

Despite a decline in tobacco use since the early 2000s, nicotine remains one of the most widely used substances. A significant percentage of nicotine users also consume cannabis, opioids, cocaine, or psychostimulants, a rate considerably higher than among non-smokers. Individuals diagnosed with a nicotine use disorder are several times more likely to have another substance use disorder. Tobacco use severity has also been linked to the initiation of heroin and cocaine use.

Technological advancements have led to the development of electronic cigarettes (e-cigarettes), marketed as a less harmful alternative to traditional tobacco products. This marketing has contributed to increased nicotine use, particularly among middle and high school students and young adults. While e-cigarettes may deliver lower nicotine doses, concerns exist regarding their safety, as compensatory vaping behaviors or high voltage settings can produce carcinogenic agents. The varying nicotine content across different e-cigarette products further complicates safety assessments. The unique impact of vaping on the development of nicotine dependence and its contribution to polysubstance use disorders requires further investigation.

Opioids

Opioid misuse has dramatically increased globally in recent years, with millions of adults reporting non-medical use annually. In the United States, a substantial number of individuals report past-year opioid misuse. The rise in first-time heroin users has paralleled the increase in prescription opioid misuse, with a strong link between prior prescription opioid misuse and heroin initiation. Studies on polydrug use among opioid misusers indicate higher frequencies of opioid use when cocaine or methamphetamine are also consumed, but lower prevalence when alcohol or cannabis are secondary substances.

Opioid use is a national public health crisis in the United States, accounting for a large proportion of accidental deaths. A significant majority of fatal opioid overdoses involve other substances, highlighting the increased risk of death when opioids are combined with other drugs or other opioids. Opioid-related emergency department visits also frequently involve tobacco, cocaine, other stimulants, cannabis, or alcohol, with substantial reports of polysubstance use involving three or more of these substances. The likelihood of these visits is linked to the severity of other substance use disorders. These findings suggest that combining opioid use with other substances exacerbates the harmful consequences of opioid use. Opioid users also experience very high rates of relapse. Previous use of other substances, including cocaine and methamphetamine, increases relapse susceptibility and can affect treatment adherence. Therefore, a comprehensive understanding of the impact of polysubstance use in the context of opioids is critical for more effective emergency responses and improved long-term treatment outcomes.

Cannabinoids

Millions of individuals worldwide and in the United States use cannabis. Its gradual decriminalization and legalization in various states have contributed to an increase in the number of users and the frequency of cannabis consumption.

Cannabis users often report high rates of co-use with tobacco, alcohol, and amphetamines. One of the most common combinations involves the simultaneous use of alcohol and cannabis. The impact of concurrent cannabis use is notable, as it is associated with increased daily alcohol consumption. Polydrug use is particularly prevalent in younger populations, with a significant percentage of adolescent cannabis users reporting the use of additional drugs. Cannabis is frequently used during treatment for other substance use disorders, which has been linked to reduced treatment efficacy, including shorter periods of alcohol abstinence and a higher incidence of cocaine relapse. Polydrug use among cannabis users has also been correlated with negative socioeconomic outcomes, financial instability, relationship difficulties, heightened mood disorder symptoms, impaired decision-making, social challenges, and self-harm. However, clinical outcomes can vary; some studies suggest a nuanced, dose-dependent impact, with no synergistic effects at low doses or a lack of association with heroin relapse.

Historically, research on cannabis has faced limitations due to restrictions and difficulties in modeling its use in animal studies. However, the development of novel methods for cannabis self-administration in animals, including oral, vaporized, and intravenous administration of Δ9-tetrahydrocannabinol (THC) and cannabidiol (CBD), is poised to significantly advance preclinical studies of cannabis use disorder and our understanding of polysubstance use involving cannabis.

Alcohol

Alcohol is one of the most commonly used drugs globally, with millions diagnosed with an alcohol use disorder. It is frequently consumed with other substances, with significant percentages of adults reporting past-year use of both alcohol and an illicit drug, or meeting diagnostic criteria for both an alcohol use disorder and another substance use disorder. Cannabis is the most commonly co-used substance with alcohol, followed by opioids, cocaine, and amphetamine. Simultaneous use of alcohol with cannabis or prescription opioids is most prevalent.

Polydrug use that includes alcohol increases the risk of developing an alcohol use disorder, particularly among young adults, men, and specific racial/ethnic groups. Such polydrug use is associated with additional comorbidities, including a higher prevalence of mood and anxiety disorders, more intense drinking, and increased drug consumption and craving. The negative consequences of alcohol polydrug use are also evident in emergency department visits, which are more likely to involve both alcohol and drugs, often requiring treatment for injuries and leading to higher rates of inpatient admissions. The rate of hospitalizations involving alcohol polysubstance use has been increasing, especially in young adults, with a reported rise compared to overdoses involving either drug or alcohol alone. Recent trends indicate a shift in the demographics of alcohol-related polysubstance combinations, with a rise in emergency department visits among females. Systematic examination of the effects of other drugs on alcohol use and addiction remains an area needing more research.

Behavioral Models of Addiction in Polydrug Studies

Behavioral models of drug addiction serve as crucial tools for investigating the neurobiological underpinnings of drug use initiation, maintenance, and relapse. Commonly employed models include locomotor sensitization, conditioned place preference (CPP), and drug self-administration, with experimental designs varying across pharmacological and non-pharmacological parameters such as drug contingency, access duration, environmental context, and administration routes. These models have been adapted to study polydrug combinations, providing insights into complex use patterns.

Initially, polysubstance studies predominantly utilized noncontingent drug administration models, like CPP and cross-sensitization. More recently, research has incorporated drug self-administration history to assess subsequent drug choice preference or craving, often in response to drug-associated cues following periods of extinction or withdrawal. Preclinical models frequently employ simultaneous administration of multiple drugs, such as alcohol with nicotine or cocaine with heroin. Clinical models, conversely, assess concurrent and sequential polysubstance use in human subjects by evaluating affective responses to drug-taking, drug-craving elicited by visual cues, and autonomic responses like blood pressure and heart rate. Further advancements include comparisons of single versus polysubstance self-administration to determine drug history's impact on molecular and circuit alterations. Paradigms based on behavioral economic principles are also used to quantify drug intake preferences and the effort individuals are willing to expend for drugs, providing normalized measures of value and price sensitivity across various drug and non-drug rewards in multiple species. These approaches facilitate the direct comparison of drug valuation across different polysubstance histories and doses, offering insights into the impact of drug pre-exposure. Clinical studies complement these efforts by employing questionnaires or controlled laboratory environments to investigate behavioral effects of polysubstance history, using monetary choice procedures, progressive ratio tests, and delay-discounting rates to compare perceived value and decision-making across multiple drugs in individuals with single or polysubstance use histories.

Effects of Polydrug Use on Addictive Behaviors

Understanding the behavioral changes and addiction-related phenotypes resulting from polydrug use is essential, particularly given the unique neurobiological alterations that can occur with exposure to multiple drugs and the high prevalence of polysubstance use disorders. However, due to the vast number of potential substance combinations and methodological variability across studies, findings regarding the impact of polydrug history on addiction-related behaviors are often mixed. Nevertheless, some general trends in drug consumption, preference, and seeking have been observed for commonly investigated substance combinations.

Psychostimulants

Studies on polydrug combinations involving cocaine are common. However, in contrast to human observations of increased addiction severity, preclinical studies have not consistently shown an escalation in addiction-like severity following sequential cocaine and alcohol or cocaine and heroin use, suggesting a need for more clinically relevant models. These findings persist despite variations in administration methods, drug doses, and species. For instance, intermittent alcohol exposure has not been found to affect cocaine self-administration or its reinforcing properties. In contrast, adolescent alcohol exposure has shown lasting effects, increasing motivation for cocaine and enhancing the development of cocaine-induced conditioned place preference in rodents. Simultaneous administration of heroin with psychostimulants can increase the motivation to self-administer these drugs, and both simultaneous and sequential administration of opioids and psychostimulants can be rewarding.

Polydrug studies involving cocaine and nicotine generally report additive or synergistic effects. Co-administration increases drug intake, enhances locomotor sensitization and conditioned place preference, and induces cross-sensitized drug-craving. Chronic nicotine pretreatment can facilitate the acquisition of cocaine self-administration, increase motivation, impair extinction learning, and enhance drug-primed reinstatement for amphetamine. Conversely, prior nicotine treatment history has been shown to reduce demand elasticity for cocaine. These effects are often dose-dependent, with smaller doses of nicotine increasing motivation for cocaine, while larger doses may have the opposite effect. The pattern of administration also matters; simultaneous administration of methamphetamine and nicotine can induce a conditioned place aversion, whereas sequential administration of the same doses may induce a conditioned place preference. These studies highlight that the effects of nicotine on psychostimulant motivation, intake, and reward are heavily influenced by delivery parameters such as dose, route, and pattern. Adolescent nicotine exposure has not been shown to increase addiction-like behavior to cocaine in adult animals, though it can enhance cocaine self-administration in adolescent rats.

In humans, simultaneous cocaine and cannabis use can lead to prolonged feelings of "stimulation" and "high," and cue-induced craving for cocaine can last longer in individuals who co-use cocaine and cannabis. However, some rodent studies indicate THC may reduce the motivation to self-administer cocaine. Furthermore, co-administration of cocaine with cannabis or alcohol in humans may not produce subjective effects different from either drug alone. In rodents, THC pretreatment has not affected psychostimulant reward or self-administration, nor has it potentiated conditioned place preference to amphetamine. CBD, another cannabinoid, has generally shown no effect on cocaine self-administration or craving, but has been found to reduce motivation and reinstatement for methamphetamine. The additive effects of cannabinoid and psychostimulant polydrug use appear to be dependent on the amount of drug consumed and the age of use. Acute THC can weaken psychomotor sensitization, but repeated administration leads to tolerance, increasing amphetamine-induced effects. Additionally, adolescent THC exposure accelerates the acquisition of cocaine self-administration and increases intake of low cocaine doses, suggesting long-lasting changes in reward circuitry similar to those observed with alcohol.

Nicotine

Limited research has focused specifically on how polydrug use impacts nicotine-induced addictive behaviors. However, it has been observed that THC pretreatment can enhance nicotine consumption and alter its valuation in behavioral economic tests in rats. In humans, heroin intake has been linked to increased cigarette consumption. In rodent studies, pre-exposure to alcohol or simultaneous access to both alcohol and nicotine can decrease nicotine self-administration. While alcohol access may not affect responding for nicotine under extinction conditions, a priming dose of alcohol has been shown to reinstate nicotine-seeking. Furthermore, systemic co-administration of methamphetamine and nicotine can produce a conditioned place aversion in rats, whereas pretreatment with either amphetamine or morphine can increase the rewarding properties of nicotine. These studies underscore the importance of considering use patterns and drug doses when interpreting the effects of polydrug use.

Opioids

Similar to psychostimulant polydrug studies, sequential use of heroin and cocaine has not been found to alter heroin self-administration or reinstatement of heroin-seeking. While alcohol pretreatment can prevent the long-term reconsolidation of preference for morphine in drug-paired contexts, adolescent alcohol exposure has been shown to enhance the development of a morphine-induced conditioned place preference. This suggests that the long-term effects of adolescent alcohol exposure may generalize across different drug classes.

Interestingly, co-administration of morphine and THC can prevent the development of analgesic tolerance that typically occurs with long-term exposure to either drug alone. Additionally, the analgesic effects of THC and oxycodone co-administration can be additive to those of oxycodone alone. Administration of THC or a combination of THC and CBD has been shown to attenuate naloxone-precipitated withdrawal symptoms without affecting the development of a morphine-induced conditioned place preference, suggesting a potential role for cannabinoids in modulating physical dependence on opioids without altering their reinforcing properties. Repeated THC administration has not significantly affected motivation for heroin or relapse to heroin-seeking, though it may cause a small reduction in heroin and oxycodone intake in fixed-ratio self-administration sessions. The effects of opioid and cannabis polydrug use appear to be dose-dependent, as systemic THC administration prior to heroin self-administration reduces responding for large doses of heroin but has no effect on lower doses.

Cannabinoids

Co-administration of nicotine with THC has been observed to amplify the effects of either drug alone in tests of locomotion, analgesia, and hypothermia. Furthermore, THC and nicotine co-administration can exacerbate the somatic symptoms of THC withdrawal. However, after repeated nicotine treatment and a period of abstinence, nicotine re-administration can attenuate THC-induced decreases in locomotor activity, increases in anxiety, and changes in social interaction, suggesting that nicotine can enhance the negative symptoms of THC when administered concurrently or in close temporal proximity. While nicotine pretreatment can enhance the rewarding effects of subthreshold doses of THC, cocaine pretreatment has been shown to heighten THC-induced anxiety behaviors.

Cannabis and alcohol polydrug use is relatively common in humans, with individuals sometimes reporting reduced alcohol consumption when cannabis is available, suggesting a potential role for cannabinoids in modulating alcohol intake. However, drug-induced cognitive and physical impairments in humans, such as those assessed in driving simulations, were found to be more severe after combined use of THC and alcohol compared to either drug alone. Therefore, conducting polydrug studies of combinations involving THC or CBD with other drug classes is crucial to understand the distinct effects resulting from these complex interactions.

Alcohol

Polydrug use involving alcohol and nicotine produces varied outcomes related to addiction behaviors. For example, pre-exposure to alcohol or simultaneous access to both alcohol and nicotine can increase alcohol self-administration, but this effect is not observed if nicotine is administered prior to alcohol each day. Chronic nicotine treatment has also been shown to enhance alcohol preference, an effect that can persist through nicotine withdrawal. While access to nicotine may impair extinction learning related to alcohol responding, it has not consistently affected the reinstatement of drug-seeking. However, other studies have found that priming doses of alcohol, but not nicotine, were capable of reinstating alcohol-seeking following self-administration of both substances. Research has not systematically examined the effects of other drugs on alcohol use and addiction to the same extent as other combinations.

Neurobiology of Addiction

The development and persistence of addiction behaviors stem, in part, from maladaptive neuroplasticity within the neural circuits governing decision-making, learning, motivation, and reward processing. Specifically, alterations within the cortico-basal ganglia-thalamic (C-BG-T) network are recognized contributors to drug-taking, drug-seeking behaviors, and the enduring nature of substance use disorders. The C-BG-T network is highly interconnected, integrating sensory and internal cues to drive motivated behavior. The striatum, a central component of this network, receives extensive excitatory inputs from cortical and subcortical regions, along with dopaminergic input from the midbrain. The integration of these inputs, along with local inhibition within the striatum, influences the initiation or suppression of behavior. Imbalances between the striatum's two main output pathways can drive addictive behaviors. This section will focus on the prefrontal cortex (PFC), nucleus accumbens (NAc), and ventral tegmental area (VTA) network, which is central to the acute effects of addictive drugs and some persistent changes following long-term use.

Prefrontal Cortex

The prefrontal cortex plays a central role in reward learning, decision-making, and assessing the value of outcomes. It is a highly diverse structure, which complicates a full understanding of its cognitive functions and how its dysregulation contributes to drug use and addiction. Generally, the medial prefrontal cortex (mPFC), encompassing areas like the anterior cingulate, prelimbic, and infralimbic cortices, regulates motivation and the pursuit of both natural and drug rewards through excitatory projections to the NAc and VTA. Hypoactivity in the PFC contributes to drug craving and seeking despite negative consequences during preoccupation stages of addiction. Its dysregulated connectivity with the striatum and VTA can lead to increased sensitivity to cues and heightened motivated drug-taking.

The PFC is organized into six layers, each with distinct connectivity patterns and cell types. Most PFC neurons are large pyramidal output cells, along with various subtypes of interneurons. Pyramidal cells in deeper layers project throughout the C-BG-T network, including to the striatum, midbrain, amygdala, hippocampus, and thalamus. These cells can be further categorized based on their physiological and connectivity profiles. The PFC also contains multiple populations of interneurons that significantly regulate cortical output by projecting to both pyramidal neurons and other interneurons.

Nucleus Accumbens

The striatum is a complex structure primarily composed of two intermingled populations of GABAergic medium spiny neurons (MSNs), which can bidirectionally regulate behavioral output. Direct pathway MSNs (dMSNs) are associated with promoting behavior, while indirect pathway MSNs (iMSNs) tend to suppress actions. Drug use increases a type of dopamine release that activates dMSNs, attributing reward to drug use during intense phases of the addiction cycle, and conditioning environments with drug-taking.

The ventral striatum, comprising the olfactory tubercule, NAc core, and NAc shell, receives modulatory dopaminergic input from the VTA and excitatory input from the PFC, as well as other brain regions. This area generally regulates motivated behavior and reward learning. It has been hypothesized that a progressive loop between the ventral and dorsal striatum facilitates the transition of habitual behaviors, moving from the ventral to the dorsal striatum and contributing to compulsive drug-seeking. In addition to MSNs, the striatum contains other important neurons, including large, continuously active cholinergic interneurons and various subtypes of GABAergic interneurons. The activity of MSNs is critical for normal regulation of behavioral output, and their dysregulation within the striatum contributes significantly to the development and expression of addictive behaviors.

Ventral Tegmental Area

The ventral tegmental area (VTA) sends dopaminergic projections to cortical, striatal, and subcortical areas, modulating the activity of the C-BG-T network. Dopaminergic projections from the VTA to the NAc specifically regulate goal-directed behaviors and are heavily implicated in the initial intense phase of substance use disorders.

VTA dopaminergic neurons receive significant excitatory input from the prefrontal cortex and other midbrain structures, as well as inhibitory input from various sources, including the NAc and local VTA GABAergic neurons. These local GABAergic neurons maintain baseline dopamine levels by inhibiting VTA dopaminergic neurons. Brief reductions in this inhibition can lead to rapid dopamine release into the NAc. The VTA also contains glutamatergic neurons that project to striatal interneurons, and some dopaminergic neurons within the VTA can co-release glutamate or GABA. This multifaceted capability allows the VTA to modulate local C-BG-T activity across multiple levels and time scales.

Mechanisms of Addiction: Acute Drug Effects

A defining characteristic of all potentially addictive drugs is their capacity to reinforce the initial, intense phase of the addiction cycle by inducing rapid dopamine release into the nucleus accumbens. However, the specific underlying mechanisms differ across various drug classes. For example, psychostimulants disrupt dopamine reuptake, while nicotine and alcohol directly activate VTA dopaminergic neurons. Opioids and cannabinoids, conversely, disinhibit VTA dopaminergic neurons. The synergistic and antagonistic interactions between different drugs within the C-BG-T network add complexity to the study of polydrug use. Limited information is available regarding the precise mechanisms behind the acute effects of specific polydrug combinations. However, research suggests that various drugs engage the C-BG-T network in distinct ways. Understanding these synergistic and antagonistic mechanisms can provide crucial insights into how particular polydrug combinations might disrupt C-BG-T network dynamics and contribute to the manifestation of addictive behaviors.

Psychostimulants

All psychostimulants acutely enhance striatal dopamine release by interfering with dopamine transporter (DAT) activity. Cocaine achieves this by blocking DAT-mediated dopamine reuptake, whereas amphetamine reverses DAT activity and promotes dopamine release from VTA dopaminergic terminals. Psychostimulants also acutely increase glutamate transmission in the prefrontal cortex, nucleus accumbens, and ventral tegmental area, indicating widespread increases in activity across the C-BG-T network. This combination of enhanced glutamatergic and dopaminergic input to the NAc facilitates the activation of medium spiny neurons and dopamine-dependent signaling pathways.

The co-administration of nicotine, alcohol, or heroin enhances psychostimulant-induced dopamine release into the NAc, though through distinct mechanisms. For instance, simultaneous administration of cocaine and nicotine activates VTA dopaminergic neurons and disrupts dopamine reuptake, leading to a greater magnitude of dopamine release than either drug alone. Some polydrug combinations involving psychostimulants can have divergent or opposing effects on the C-BG-T circuit. Co-administration of alcohol with cocaine produces cocaethylene, which can amplify dopamine release; however, alcohol also prevents cocaine-induced glutamate transmission in the NAc core. Furthermore, polydrug studies with psychostimulants have identified an exacerbation of drug-induced cellular toxicity compared to psychostimulant use alone. Specifically, co-administration of heroin with cocaine has been shown to reduce metabolic activity, increase intracellular calcium signaling, and decrease mitochondrial membrane potential, contributing to enhanced cell death.

Nicotine

Unlike psychostimulants, nicotine increases dopamine release primarily through the activation of nicotinic acetylcholine (nACh) receptors within the VTA. These receptors are non-selective cation channels that, upon activation, lead to depolarization and enhanced neurotransmitter release. The reinforcing effects of nicotine are mainly attributed to its action on nACh receptors located on VTA dopaminergic neurons and on presynaptic nACh receptors found on glutamatergic inputs from the prefrontal cortex. Acute nicotine exposure activates both presynaptic inputs and VTA dopaminergic cell bodies, triggering rapid dopamine release into the NAc. Notably, nicotine-induced dopamine release can be blocked by certain manipulations of glutamate or GABA receptors, highlighting the complex regulation of VTA dopaminergic neuron activity by local microcircuitry. The acute effects of nicotine when combined with other drugs in polydrug use have not been extensively studied.

Opioids

Although opioids vary in origin, potency, and receptor bias, they all exert their rewarding effects through the activation of mu opioid (μO) receptors. These receptors are expressed both on cell bodies and axons, but the primary mechanism of opioid-induced dopamine release involves the presynaptic inhibition of VTA GABAergic neurons. Mu opioid receptors are inhibitory G protein-coupled receptors; their activation reduces neuronal excitability through various mechanisms, including inhibiting signaling pathways, activating specific potassium channels, inactivating calcium channels, and inhibiting neurotransmitter release. Acute exposure to opioids inhibits VTA GABAergic neurons, leading to a subsequent disinhibition of VTA dopaminergic neurons and rapid dopamine release into the NAc.

Given the distinct mechanism of action for opioids compared to psychostimulants and nicotine, it is not surprising that their co-administration can amplify the acute effects of opioids. Simultaneous administration of opioids and psychostimulants, for example, produces an additive increase in dopamine release in the NAc and prolongs elevated levels of dopamine and its metabolites. Similarly, combining opioids with nicotine enhances opioid-evoked dopamine release in both the NAc and dorsal striatum. Cross-tolerance to opioid-mediated pain relief has also been observed following pre-exposure to nicotine and cannabis. This cross-tolerance involves complex pharmacological interactions and is mediated by different receptor systems depending on the specific polydrug combination.

Cannabinoids

Cannabis contains two main cannabinoids with differing affinities for cannabinoid (CB) receptors: THC, a partial agonist for both CB1 and CB2 receptors, and CBD, which has very low affinity for these receptors and signals through an unknown mechanism. CB1 receptors are located on presynaptic terminals throughout the central nervous system and are responsible for the psychoactive effects of cannabis. CB2 receptors are primarily found on immune cells and mediate antinociceptive and anti-inflammatory effects. Both receptor subtypes signal through inhibitory pathways similar to opioids, leading to presynaptic inhibition and downregulation of specific signaling cascades.

CB1 receptors are densely expressed on presynaptic terminals of VTA GABAergic neurons and NAc dMSNs that target VTA dopaminergic neurons. Activation of CB1 receptors reduces inhibitory currents in VTA slices, leading to disinhibition of VTA dopaminergic neurons, an increase in their firing, and subsequent dopamine release into the NAc. These effects can be blocked by certain opioid antagonists. Like other drugs, the acute effects of cannabinoids on C-BG-T network activity are broad and involve multiple neurotransmitter systems. Acute THC, for instance, increases both dopamine and glutamate signaling in the NAc and PFC, while decreasing GABA signaling in the VTA and PFC. These alterations collectively reduce inhibitory feedback within the C-BG-T and facilitate behavioral output. Alcohol consumption prior to cannabis use has been shown to enhance plasma THC levels and increase self-reported euphoria in humans, indicating synergistic effects between the two drugs. Moreover, simultaneous administration of cannabinoids with psychostimulants or opioids can enhance the activation of VTA dopaminergic neurons, and combining THC and nicotine has been observed to increase cellular activation throughout the C-BG-T.

Alcohol

Despite its widespread use, less is understood about the precise mechanisms underlying the acute effects of alcohol. Alcohol activates dissociated VTA dopaminergic neurons and induces dopamine release into the NAc, similar to other potentially addictive drugs. GABAA receptors are known to play a central role in alcohol's effects, as alcohol can enhance GABAA signaling. It is hypothesized that alcohol may facilitate rapid dopamine release from VTA dopaminergic neurons by inactivating local VTA GABAergic neurons. Supporting this, alcohol inhibits VTA GABAergic neurons, and infusing GABAA agonists into the VTA can dose-dependently increase dopamine release.

In addition to GABA, alcohol's acute effects are dependent on glutamatergic signaling within the C-BG-T. Alcohol can increase glutamate release in the NAc and VTA, suggesting it engages a feedforward loop for activating VTA dopaminergic neurons. Polydrug use involving alcohol often produces synergistic effects throughout the C-BG-T, likely due to alcohol's unique pharmacological profile. For example, chronic pretreatment with nicotine enhances acute alcohol-induced dopamine release in the NAc, with elevated dopamine levels persisting for an extended period. Furthermore, co-administration of alcohol and nicotine acutely increases the production of certain growth factors in the NAc, along with increases in glutamatergic signaling in the VTA and PFC. This widespread activation of the C-BG-T network is typically absent when either drug is administered alone, demonstrating a unique mechanism of action for alcohol and nicotine polydrug use.

Mechanisms of Addiction: Long-Term Alterations

Prolonged use of psychostimulants, nicotine, opioids, cannabinoids, and alcohol leads to extensive and distinct changes across the C-BG-T network, yet some notable alterations are shared among these drugs. These long-term adaptations contribute to the progression through the addiction cycle, from intense use to withdrawal, negative emotional states, and ultimately, preoccupation with drug-seeking. For instance, acute withdrawal often results in a temporary reduction in baseline dopamine levels in the NAc, followed by a persistent increase in the excitability of VTA dopaminergic neurons, which contributes to enhanced cue-evoked dopamine release during abstinence. Conversely, withdrawal can also lead to a lasting reduction in certain forms of synaptic plasticity and intrinsic excitability in NAc MSNs, as well as a decrease in the availability of dopamine D2 receptors in the striatum. Reduced D2 receptor availability has been broadly linked to various addictive disorders. Finally, chronic use of these substances drives a persistent increase in the expression of specific proteins in cortical pyramidal cells and NAc dMSNs, which has been associated with drug-seeking during abstinence. While chronic use of any of these drugs results in numerous other transient or persistent changes across the C-BG-T network, this discussion will focus on alterations in plasticity, morphology, and connectivity within the VTA, NAc, and PFC following both single and polydrug use, emphasizing how the combined effects of these drugs can uniquely disrupt C-BG-T network dynamics.

Psychostimulants

Long-term psychostimulant use leads to enduring disruptions in glutamate balance and alterations in neuronal structure throughout the C-BG-T network. Repeated cocaine administration weakens inhibitory signaling in certain prefrontal cortex neurons, increasing their excitability and enhancing excitatory input to the NAc. Similarly, chronic cocaine increases excitatory input to the VTA and weakens inhibitory signaling of VTA dopaminergic neurons, facilitating their activity. Additionally, psychostimulants can create inactive synaptic connections on dMSNs and reduce glutamate release from prefrontal inputs to the NAc core, potentially weakening striatal output.

It is important to note that cocaine-inactivated synapses on dMSNs can become active again during withdrawal through the recruitment of specific receptors. A low dose of psychostimulant can restore glutamate release into the NAc core, suggesting a process of compensatory adjustment. Furthermore, repeated psychostimulant exposure increases the expression of certain proteins in PFC pyramidal neurons and NAc dMSNs, which contributes to an increase in dendritic branching that persists for at least a month into abstinence. Research has not yet examined how these specific effects of psychostimulants are altered by polydrug use.

Nicotine

Similar to psychostimulants, long-term nicotine administration reduces certain inhibitory signaling pathways in the mPFC and NAc, thereby dampening inhibitory control over the C-BG-T network. Nicotine also enhances excitatory input to VTA dopaminergic neurons and sensitizes dopamine release into the NAc. Repeated nicotine use leads to an increase in nicotinic acetylcholine receptors throughout the C-BG-T, including in the PFC and midbrain. Conversely, chronic nicotine exposure can weaken both excitatory and inhibitory inputs to the NAc during abstinence, potentially through increased sensitivity of NAc D2 receptors. This is supported by the upregulation of certain calcium-permeable receptors in the NAc.

Long-term nicotine use also leads to a persistent increase in dendritic branching in mPFC pyramidal cells and NAc MSNs, similar to the effects of psychostimulants. Moreover, adolescent nicotine use has been linked to enhanced synaptic pruning, activation of immune cells in the brain, and increased inflammatory signaling throughout the C-BG-T, mediated by dopamine D2 receptors. Repeated co-administration of psychostimulants and nicotine can enhance the strengthening of synaptic connections and increase the expression of specific proteins in the NAc core. Interestingly, these effects are not observed after sequential administration of these drugs, indicating a unique pharmacological profile for concurrent nicotine and psychostimulant use.

Opioids

Similar to psychostimulants and nicotine, long-term exposure to opioids strengthens excitatory input to VTA dopaminergic neurons. They also weaken GABAergic inhibition of VTA dopaminergic neurons, primarily through a reduction of inhibition mediated by dMSNs. This effect suggests a persistent disruption in NAc output. However, unlike psychostimulants and nicotine, long-term opioid exposure decreases dendritic branching and spine density in NAc MSNs and mPFC pyramidal cells.

Within the NAc, long-term opioid exposure weakens excitatory input to NAc shell iMSNs and induces inactive synaptic connections on iMSNs. During withdrawal, opioid-generated inactive synapses on iMSNs are eliminated, and the intrinsic excitability of iMSNs is reduced. Given that MSNs form a dense network of lateral inhibition within the NAc, opioid-induced disruptions in iMSN signaling could disrupt local NAc microcircuitry and facilitate aberrant C-BG-T network dynamics. Research has not yet examined how these specific effects of opioids are altered by polydrug use.

Cannabinoids

Repeated administration of THC weakens glutamatergic signaling from the mPFC to the NAc but strengthens input from other brain regions to the NAc shell, suggesting a reorganization of excitatory input to the NAc following long-term cannabinoid exposure. Notably, repeated THC administration also blocks a specific form of synaptic plasticity on VTA GABAergic neurons, indicating a loss of local inhibitory control over VTA dopaminergic neurons. Additionally, adolescent THC administration followed by abstinence has been shown to reduce the intrinsic excitability of certain prefrontal neurons and weaken glutamatergic input to the VTA, suggesting a state of reduced glutamate activity induced by cannabinoids.

THC exposure also produces long-lasting changes in the morphology of neurons throughout the C-BG-T, increasing dendritic spine length and branching in the mPFC and NAc shell, and increasing dendritic spine density in the NAc shell, while reducing the size of VTA dopaminergic neurons. Repeated co-administration of THC and nicotine enhances the expression of specific proteins in the NAc, and acute withdrawal from this combination can dysregulate glutamatergic input to the NAc and PFC.

Alcohol

Similar to opioids, extended alcohol exposure decreases dendritic branching and spine density in the NAc shell and mPFC. Withdrawal from alcohol also reduces baseline dopamine levels in the NAc. However, alcohol withdrawal also causes distinct changes throughout the C-BG-T network, including reduced GABAergic signaling in the NAc and hippocampus and increased glutamatergic signaling in the NAc and PFC. This increased glutamatergic signaling is due to disrupted glutamate reuptake rather than enhanced release.

In laboratory studies, polydrug exposure to alcohol and nicotine can significantly increase cellular apoptosis, leading to cell death. Interestingly, however, alcohol or nicotine withdrawal-induced neurodegeneration may be less severe following co-administration of both drugs, indicating a unique molecular pathology when nicotine and alcohol are used in combination.

Conclusion

Improving the applicability of findings from preclinical models to accurately reflect polysubstance use history and dependence in human populations is critical. While human imaging studies are limited by the inability to control intake history, animal models offer advantages for assessing polydrug history under controlled conditions. However, these models have limitations in fully encompassing the complex social and environmental contexts that contribute to unique patterns of multiple substance use. Factors such as the time of day of intake, temporal proximity of consumption, environmental preferences for administration, and the route of drug administration are important considerations, as they can influence the development and severity of addiction behaviors.

Accounting for temporal patterns of drug use is also essential, as sequential patterns of polydrug consumption are often reported more frequently than simultaneous patterns and can lead to distinct neural adaptations. Understanding sex differences in the frequency and patterns of polydrug use, drug discrimination, and circuit alterations is also necessary to fully comprehend the interactions and impacts of polydrug use in clinical populations. Furthermore, while powerful behavioral economic models allow comparisons across drug classes, experiments using these models must be designed with consideration for different scales of intake and indifference points for drug valuation to accurately model and interpret data. The interplay of socioeconomic background, social environment, and access strategies also affects the frequency of and susceptibility to polydrug use. Incorporating these influences into preclinical models is crucial for ensuring the translational relevance of findings.

Novel preclinical paradigms are continuously being developed to model different routes of drug administration and study relapse under clinically relevant conditions. For instance, voluntary self-administration models for vaporized ethanol, cannabis, and nicotine have been established. Additionally, various models for inducing drug abstinence are being introduced, such as pairing drug-taking with adverse consequences or choice procedures involving a drug versus an alternative reinforcer. These paradigms hold high translational value and can be powerful tools for understanding the neural basis of therapies like contingency management. Combining these models with behavioral economic paradigms could clarify how the relative value of drugs and alternative reinforcers changes with polydrug use, potentially improving treatment efficacy. Addiction is a complex disease with multiple, highly variable factors contributing to its initiation, maintenance, and relapse. Given the widespread prevalence of polydrug use among individuals with substance use disorders, it is critical to incorporate this variability into both human studies and animal models. This will help determine whether polysubstance use exacerbates the severity of substance use disorders, whether increased disorder severity drives polysubstance use, or if there is a bidirectional relationship between the two. Although challenging, gaining a deeper understanding of the behavioral, genetic, and environmental contributions to polysubstance use and addiction, as well as the mechanisms underlying addiction severity and relapse, will be instrumental in developing effective treatment and policy strategies to combat this ongoing public health crisis.

Introduction

Drug addiction is a complex disorder characterized by repeating cycles of drug use, withdrawal, periods of not using drugs, intense cravings, and a return to using drugs. This condition is widespread. Globally, an estimated 35 million people meet the criteria for a substance use disorder (SUD). In the United States, this number is about 19.3 million people. Surveys suggest that approximately 10% of individuals will experience a SUD in their lifetime. Drug addiction also poses a major public health challenge in the United States, with an annual financial cost of $740 billion related to treatment, lost work productivity, healthcare, and crime. These costs are likely to grow as illegal drug use increases, with a quarter of a billion people worldwide reporting use in the past year. In the United States, over 17 million people aged 12 and older are estimated to begin using drugs each year. Opioid use, in particular, continues to rise, with 53 million past-year opioid users worldwide and about 11 million people in the U.S. reporting opioid misuse in the last year. This trend is alarming, as opioid-involved deaths in the U.S. increased six-fold from 1999 to 2017, with approximately 130 Americans dying daily from opioid use.

While most research on SUDs has focused on individual substances, often excluding those with a history of using multiple drugs from studies, it is important to recognize that many drug users engage in polysubstance use. For example, 30–80% of heroin users have also reported using cocaine, and deaths involving both cocaine and opioids in the U.S. more than doubled between 2010 and 2015. A person is considered a polysubstance user if they use more than one substance, whether at different times (sequential use) or simultaneously. Limiting studies to individual drugs risks missing important interactions between substances, reduces how well preclinical research translates to human experience, and can hinder the effectiveness of treatments for SUDs. Indeed, polysubstance use has consistently been linked to poorer treatment outcomes, including worse treatment retention, higher relapse rates, and a three-fold higher mortality rate compared to using only one substance. This review aims to combine current knowledge about the mechanisms and effects of individual drug use with the latest research on polysubstance use, noting whenever possible if polysubstance use is concurrent, sequential, or a combination. The discussion begins with an overview of public health trends for single and polysubstance use, as well as the impact of polysubstance history on the severity of substance use. This is followed by findings from preclinical studies and their relevance to real-world substance use, along with considerations for designing polysubstance studies. The pharmacology of individual substances and some of their effects on the cortical-basal ganglia-thalamic (C-BG-T) circuitry are then detailed, setting the foundation for understanding how polysubstance use may change the brain pathology of addiction. Care is taken to discuss differences in brain alterations between single and polysubstance use, highlighting the most common combinations. For clarity, sections are organized by the primary substance used, considering the consequences when additional drugs are combined. Finally, suggestions and potential methods for advancing the critical task of examining polysubstance disorders are offered.

Public Health Trends in Substance Use

Drug addiction is both widespread and lethal, contributing to approximately 585,000 drug use-related deaths worldwide each year. Although discussions about drug addiction and its effects often center on individual drugs, drug misuse commonly involves multiple substances. In fact, drug-dependent individuals report using an average of 3.5 substances, including both simultaneous and sequential polydrug use. Additionally, the likelihood of developing multiple substance dependencies is high among those seeking clinical help. While combinations of co-used substances vary, primary drug dependencies are typically seen for alcohol, opioids, amphetamine, and methamphetamine. Cannabis and cocaine are more often reported as secondary or tertiary substances used. The high occurrence of polysubstance use is particularly concerning due to its impact on SUD severity and treatment outcomes. For instance, a history of polysubstance use is linked to greater unmet physical and mental healthcare needs, increased risky behaviors, violence, and a higher risk of overdose and death compared to single substance use. This section discusses the public health consequences of single substance use, as well as polysubstance use in relation to secondary substance combinations. This overview aims to address common patterns of polydrug use, including the substance combinations and use patterns that occur most frequently. However, given that data is limited for specific drug use patterns, classification types, and differences in the definition of polydrug combinations across studies, it is unlikely to capture all combinations, histories, and patterns of use.

Psychostimulants

Psychostimulants are the second most widely used class of drugs, with 18 million current cocaine users and 29 million current prescription stimulant users worldwide. Global prevalence of psychostimulant use remained relatively stable from 1990 to 2017, with 7.38 million reported to meet criteria for an amphetamine use disorder and 5.02 million for a cocaine use disorder. However, the number of drug-related overdose deaths involving psychostimulants has continued to rise, especially in the U.S., with a 2.6-fold increase in the cocaine overdose death rate and a 3.6-fold increase in the methamphetamine overdose death rate from 2000 to 2017.

Significantly, cocaine and amphetamine users are predominantly polysubstance users; one study reported a 74% incidence of polysubstance history for cocaine users and 80% for amphetamine users. Specifically, cocaine use and the development of a cocaine use disorder are linked to concurrent heroin, cannabis, tobacco, and alcohol use. Similarly, amphetamine users show several types of polysubstance use, with high probabilities of alcohol, tobacco, and cannabis use. Additionally, other groups of amphetamine polysubstance users have higher probabilities of heroin and other opioid use. Across groups, lower probabilities of cocaine use with amphetamine compared to other drug classes used with amphetamine have also been reported.

Polysubstance use is common among stimulant users, with both concurrent and sequential drug consumption patterns. For example, simultaneous use of psychostimulants and opioids is seen with both cocaine ("speedball") and methamphetamine ("bombita"). Sequential use of psychostimulants and opioids is also common, including using cocaine or amphetamine to avoid opioid-related physical withdrawal symptoms and using opioids to reduce over-stimulation following cocaine use. There is also an increased likelihood of same-day methamphetamine use with alcohol consumption.

Although these studies did not specify the patterns of polydrug use, a review of reports on concurrent versus simultaneous cocaine use found a 24–98% range of simultaneous cocaine and alcohol use and 12–76% incidence of simultaneous cannabis use. Rates of concurrent use were 37–96% for cocaine and alcohol, 43–94% for cocaine and cannabis, 70–80% for cocaine and nicotine, and 85–95% for amphetamine and nicotine. The wide variation in reported frequencies highlights the complexity in identifying drug use patterns, which can differ based on demographics, study periods, study design, and definitions of concurrent and simultaneous use.

Polydrug use involving psychostimulants poses significant public health risks. For instance, one study showed that amphetamine users were 21 times more likely to have a concurrent cannabis use disorder and 7 times more likely to have past-year concurrent cocaine use, compared to those with no prior history of amphetamine use. Furthermore, nearly one-third of overdose deaths involved both psychostimulants and opioids, such as heroin and fentanyl. The dangers of psychostimulant co-use also extend to other substances, with combined cocaine and cannabis use leading to higher standardized death rates in emergency department (ED) visits, suggesting elevated mortality risks with this combination. Additionally, combining cocaine and alcohol use increases the risk for heart damage compared to either drug alone.

Nicotine

Although tobacco use, which contains nicotine, has declined since the early 2000s, nicotine remains one of the most commonly used drugs, with 58.8 million people aged 12 or above in the U.S. reporting past-month nicotine use. It is also commonly used with many other substances: 17% of nicotine users also used cannabis, 4.7% used opioids, 2.6% used cocaine, and 1.4% used psychostimulants in the past month. In contrast, non-smokers had much lower percentages of past-month substance use (3.7%, 1.2%, 0.2%, and 0.3% for the aforementioned substances, respectively). This difference is significant, as individuals with a nicotine use disorder are 3–4 times more likely to have a second SUD. Additionally, past-year tobacco use was strongly associated with opioid use disorders, as well as co-occurring cannabis and alcohol use and related disorders, and cocaine use in primary care patients. The severity of tobacco use (i.e., frequency of use and number of cigarettes smoked) has also been significantly correlated with the start of heroin and cocaine use.

Historically, nicotine has mainly been consumed by smoking tobacco cigarettes. However, new technological advances have led to the development of electronic (e-) cigarettes, designed to deliver nicotine in a way that avoids toxins. The marketing of e-cigarettes as a safer alternative to traditional tobacco cigarettes is concerning, as it has increased the likelihood of nicotine use and a resurgence in the potential for nicotine addiction. For example, e-cigarette use among middle and high school students increased from 2012 to 2016, and a spike in use was observed among young adults (18–24 years old) around 2013 to 2014 when e-cigarette products were introduced. Despite delivering lower doses of nicotine, the safety of commercial e-cigarettes has been debated, since compensatory "puffing" behaviors or high voltage settings can lead to the production of cancer-causing agents. The potential danger of use is further compounded by the variable amounts of nicotine provided by different e-cigarette manufacturers. The unique influence of vaping on the development of nicotine dependence and how this specifically contributes to polysubstance use disorders remains largely unknown and requires further study in the coming years.

Opioids

The prevalence of opioid misuse, meaning use outside of prescribed medical reasons, has sharply increased in recent years, with approximately 53 million adults (1.1% of the global population) reporting past-year non-medical use of an opioid. In the U.S. alone, 11 million people reported past-year opioid misuse in 2016; however, this estimate is conservative as it does not include homeless or incarcerated individuals, who have disproportionately higher levels of opioid use. Additionally, the rate of first-time heroin users rose alongside the non-medical use of prescription opioids from 2002 to 2011, suggesting that individuals with past-year prescription opioid misuse are 19 times more likely to initiate heroin use than those without such a history. Studies have investigated polydrug use among heroin and prescription opioid misusers and found higher frequencies of opioid use in individuals who also use cocaine (over 33%) or methamphetamine (over 20%), but reduced prevalence for primary opioid use in those with secondary alcohol or cannabis use. Furthermore, first-time methamphetamine use is more common after past-month opioid use. Among those entering treatment for heroin use, 91% reported a lifetime history of cocaine use. Additionally, a study in the United Kingdom found that 54% of opioid users in treatment between 2017 and 2018 also had a co-occurring crack cocaine use disorder. Regarding patterns of multi-drug use, simultaneous use of heroin with alcohol and/or cannabis is more common than with psychostimulants, and a sequential pattern of drug use is preferred for opioids and psychostimulants.

In the U.S., opioid use is a national public health emergency, responsible for more than 1.6 million years of life lost from 2001 to 2016. Moreover, opioid overdose deaths are currently the leading cause of accidental death among U.S. adults, with 68% of all drug overdose deaths involving an opioid. Given that nearly 80% of fatal opioid overdoses also involved another substance, there appears to be a greater risk of death when opioids are used in combination with other opioids and/or other drugs. Specifically, of these deaths, 78% involved another opioid, 21.6% involved cocaine, 11.1% involved alcohol, and 5.4% involved a psychostimulant other than cocaine. Furthermore, opioid-related emergency department (ED) visits also involved tobacco (51.1%), cocaine (36.9%), other stimulants (22.6%), cannabis (25.1%), or alcohol (16.9%). Significant polysubstance use of three or more of these substances has also been reported for opioid-related ED visits, and the likelihood of these visits has been associated with the severity of other SUDs. Taken together, these reports suggest that combining opioid use with other substances can worsen the harmful consequences of opioid use. In addition to overdose risk, opioid users experience very high rates of relapse, with 59% of individuals relapsing in the first week and 80% relapsing in the first month of abstinence. Past use of other substances, including the extent of cocaine use, increases susceptibility to relapse. Methamphetamine use among those seeking treatment for opioid use has also been increasing, and recent reports indicate that methamphetamine use is associated with discontinuation of buprenorphine treatment in individuals with an opioid use disorder. Therefore, a better understanding of the impact of polysubstance use in the context of opioids is crucial for more successful emergency responses and long-term treatment outcomes.

Cannabinoids

It is estimated that 188 million individuals aged 12 years or older use cannabis worldwide, including 43.5 million individuals in the U.S. Starting in 2012 with Washington and Colorado, 11 states and the District of Columbia have legalized recreational cannabis, making it legally accessible to about 328 million people. The number of cannabis users in the U.S. has increased with its gradual decriminalization and legalization, rising from 4.1% in 2002, to 9.9% in 2007, and to 15.9% in 2018. The frequency of cannabis use is also high, with reports indicating that 40% of individuals are daily or near-daily users.

Cannabis users are reported to have high rates of past-month tobacco, alcohol, and/or amphetamine use. One of the most common combinations is simultaneous use of alcohol and cannabis, along with simultaneous alcohol, cocaine, and cannabis use. The impact of concurrent cannabis use is notable, as this pattern is associated with more alcoholic drinks per day, suggesting that co-existing cannabis consumption facilitates alcohol use. Polydrug use is particularly common in younger populations. Among adolescent cannabis users, 27.5% reported additional drug use within the same year of starting cannabis use, and nearly 67% use two or more other drugs.

Cannabis is frequently used during treatment for other SUDs, and this has been linked to reduced treatment effectiveness. For example, cannabis use has been found to result in shorter periods of alcohol abstinence, as well as a greater incidence of relapse to cocaine. Additionally, polydrug use among cannabis users has been correlated with reduced socioeconomic mobility, financial instability, and relationship difficulties, as well as a heightened degree of mood disorder symptom severity, decision-making deficits, social difficulties, and self-harm. Although these data suggest that the consequences of drug use are enhanced by concurrent cannabis use, it should be noted that clinical outcomes can vary for studies examining polydrug use among cannabis users. For example, some studies suggest a nuanced impact of polysubstance use that is dose-dependent, with no synergistic effects of cannabis and alcohol at low doses of either drug, and a lack of association of cannabis use in heroin relapse.

Unfortunately, due to long-term restrictions on cannabis research in the U.S. and past difficulties in modeling cannabis use with self-administration models in rodents, much less is known about the impact of cannabis compared to other drugs. The development of new methods for cannabis self-administration in animals, including oral self-administration of Δ9-tetrahydrocannabinol (THC)-containing gelatin, self-administration of vaporized THC and cannabidiol (CBD), and intravenous self-administration of THC and CBD, will help advance preclinical studies of cannabis use disorder. These methods will also enable a better understanding of the consequences of polydrug use involving cannabis.

Alcohol

Alcohol is one of the most commonly used drugs, with up to 290 million people worldwide diagnosed with an alcohol use disorder, including 15 million people in the U.S. Alcohol is frequently used with other substances, with reports indicating that 5.6% of U.S. adults have used both alcohol and another illicit drug within the past year, and 1.1% have met diagnostic criteria for both an alcohol use disorder and another SUD. The most commonly reported substance co-used with alcohol is cannabis (10%), with less common co-occurring conditions found with opioids (2.4%), cocaine (2.5%), and amphetamine (1.2%). Although simultaneous use of alcohol and cannabis or alcohol and prescription opioids is most common, simultaneous use is also seen with cocaine.

Polydrug use increases the risk of developing an alcohol use disorder, particularly in young adults, men, and American Indians/Alaskan Natives. Polydrug use that includes alcohol is associated with additional co-occurring conditions, including a higher prevalence of mood disorders, anxiety disorders, more intense drinking, and more intense drug consumption and drug cravings. The negative consequences of alcohol polydrug use are also highlighted by data indicating that 21% of emergency department (ED) visits for patients aged 12–24 years old involved both alcohol and drugs. These visits were also more likely to require treatment for injuries and had higher rates of inpatient admission.

Additionally, 17% of substance treatment admissions were related to both alcohol and drug use, representing 45% of primary alcohol admissions and 33% of drug misuse admissions. The rate of hospitalizations involving alcohol polydrug use has been increasing, particularly in young adults, with reports suggesting a 76% rise in inpatient admission between 1998 and 2008, compared to either drug or alcohol overdoses alone. While the polysubstance users in these surveys were primarily white and male, recent trends indicate a rise in ED visits related to alcohol and drug combinations in females, suggesting a change in the demographics of polysubstance combinations that include alcohol.

Behavioral Models of Addiction in Polydrug Studies

Behavioral models of drug addiction are employed to investigate the brain mechanisms underlying the development, maintenance, and relapse of drug use. The most frequently used models include locomotor sensitization (a progressive and lasting increase in movement responses to the same drug dose), conditioned place preference (CPP; a test of drug reward measured as an increase in time spent in a drug-associated chamber), and drug self-administration (drug intake dependent on a response). These experimental designs vary in several ways, including whether drug use is contingent, the amount of drug access, the context of drug use, and the methods of administration. This section describes how these models have been used with polydrug combinations and how this research has advanced the understanding of polydrug use and addiction.

Early polysubstance studies largely used models where drug administration was not dependent on a behavior, particularly CPP and cross-sensitization models. In these, the impact of initial doses of one drug on preference for a location or motor activity, respectively, of another drug is determined. More recently, studies have been examining how a history of drug self-administration impacts subsequent drug choice preference and/or drug craving, measured by responses to drug-associated cues after extinction and/or withdrawal. Preclinical polydrug models involving simultaneous administration of multiple drugs, such as alcohol and nicotine or cocaine and heroin combinations, have also been frequently used. In human clinical models, both concurrent and sequential polysubstance use is assessed through measurements of emotional responses to drug-taking, drug craving following visual cues, and autonomic responses like blood pressure and heart rate.

More recently, studies are comparing single versus polysubstance self-administration to determine how drug history affects drug-induced molecular and circuit alterations. Additionally, paradigms based on behavioral economic principles can determine the preferred level of drug intake (i.e., no-cost intake; Q0), as well as the amount of effort an animal is willing to exert to maintain Q0 before consumption and responding begins to decline (i.e., price; Pmax). These paradigms are particularly powerful because they can use Q0 and Pmax to generate normalized measures of value (i.e., essential value; α) and price (nPmax). These measures have been used to compare price sensitivity, effort, and value across different drug and non-drug rewards in polysubstance models in several species, including rats, rhesus monkeys, and human participants. In particular, these studies allow for the examination of the relative reinforcing properties of different doses and classes of drugs, as well as changes in the perceived cost of a drug after prior exposure to another drug. Direct quantification of the assigned value of these drugs across different polysubstance histories and drug doses is very useful for assessing the impact of polysubstance history on the relative value of a reinforcer. Furthermore, these measures can be used to compare how polysubstance users value drug rewards across different experimental parameters (e.g., differences in the initial dose of one drug, environmental context, pattern of drug use). Finally, clinical studies use questionnaires or controlled laboratory environments to investigate the behavioral effects of a polysubstance history. Specifically, these studies use monetary choice procedures that compare the assigned value of drugs at different doses, how the assigned value changes for one drug with a change in the price of another, or how the relative value of one drug changes with a perceived change in the subjective quality of another available drug. Additionally, progressive ratio tests for a single drug or drug combinations to study motivation, and delay-discounting rates for money and drug rewards to study decision-making, have also been performed. These studies allow for the comparison of perceived value across multiple drugs in participants with histories of single or polysubstance use.

Effects of Polydrug Use on Addictive Behaviors

Given the unique brain changes that can occur with exposure to multiple drugs, along with the high prevalence of polysubstance use disorders, there is a strong need to develop polydrug paradigms that are highly relevant to human conditions. These paradigms are crucial for fully understanding the behavioral changes and addiction-related characteristics that develop following polydrug use. However, due to the vast number of potential substance combinations and the variability in methods across studies, current results and interpretations regarding the impact of polydrug history on addiction-related behaviors are mixed. Nevertheless, some general trends in drug consumption, drug preference, and drug-seeking have been demonstrated in commonly studied substance combinations.

Psychostimulants

Some of the more commonly studied polydrug combinations involve administering cocaine with other drugs. However, in contrast to human reports, an increase in addiction-like severity has not been observed in preclinical studies. This suggests a need for models with greater relevance to human conditions that can capture the increased severity seen in human polysubstance users. Specifically, animal studies of sequential cocaine and alcohol or cocaine and heroin use have not found differences in drug intake or the return to drug-seeking based on whether single or polydrug use occurred. These findings were observed despite variations in the use of contingent and non-contingent drug administration, drug doses, and species, including rats and rhesus monkeys. Consistent with this work, intermittent alcohol exposure has not been shown to affect cocaine self-administration or the reinforcing properties of cocaine, as measured by demand curves in rhesus monkeys. Additionally, intermittent alcohol exposure has not been shown to affect progressive ratio tests of motivation for cocaine, or the long-term reconsolidation of preference for cocaine in drug-paired contexts in rats. In contrast, adolescent alcohol exposure has been shown to have long-lasting effects on cocaine self-administration and reward, suggesting that this population is particularly susceptible to the effects of polysubstance use. For example, adolescent alcohol exposure increases motivation for cocaine, enhances the development of a cocaine conditioned place preference (CPP) in both mice and rats, and weakens cocaine-induced taste aversion. Furthermore, simultaneous heroin and psychostimulant administration increases the motivation to self-administer cocaine and methamphetamine, and both simultaneous and sequential administration of morphine and methamphetamine have been shown to be rewarding, as measured by the development of a CPP. Pretreatment with an opioid also enhances methamphetamine-induced psychomotor sensitization. These findings suggest that opioids can enhance the rewarding and motivational properties of psychostimulants, particularly when administered simultaneously.

Polydrug studies involving cocaine and nicotine have largely reported additive and/or synergistic effects of the two drugs. Specifically, co-administration of cocaine and nicotine increases drug intake in rhesus monkeys and rats, enhances locomotor sensitization and the development of a conditioned place preference (CPP) in mice, and induces a cross-sensitized drug craving. Additionally, chronic nicotine pretreatment facilitates the acquisition of cocaine self-administration, increases motivation under a progressive ratio schedule, impairs extinction learning, and enhances drug-primed reinstatement for amphetamine following amphetamine self-administration. Conversely, prior nicotine treatment history reduces demand elasticity for cocaine. It is important to note that the effects of nicotine and psychostimulant polydrug use are largely dose-dependent. For instance, pretreatment with a smaller dose of nicotine increases motivation to take cocaine in rats, whereas a larger dose has the opposite effect. Additionally, simultaneous administration of methamphetamine and nicotine induces a conditioned place aversion in mice, while sequential administration of the same dose induces a CPP. Together, these studies demonstrate that the effects of nicotine on psychostimulant motivation, intake, and reward are heavily impacted by the parameters surrounding nicotine delivery (e.g., dose, route, pattern), which should be carefully considered when comparing study results and developing preclinical polydrug paradigms. Notably, adolescent exposure to nicotine has no effect on the subsequent development of a cocaine CPP, cocaine-induced taste aversion, cocaine self-administration, extinction, or reinstatement of cocaine-seeking in adulthood, although it has been shown to enhance cocaine self-administration in adolescent rats. These data suggest that, unlike alcohol, early exposure to nicotine does not lead to increases in addiction-like behavior to cocaine in animals.

In humans, simultaneous cocaine and cannabis use produces feelings of "stimulated" and "high" that last longer than either drug alone, and cue-induced drug craving in individuals who co-use cocaine and cannabis lasts longer than for those who only use cocaine. In contrast, THC reduces the motivation to self-administer cocaine in rodents. Although this suggests a different regulation of cocaine's effects in humans and rodents, further work is necessary to confirm that animal models of increased addiction severity cannot, in fact, be developed. Notably, however, neither cocaine and cannabis nor cocaine and alcohol co-administration in humans produces subjective effects that are different from cocaine, cannabis, or alcohol alone. Similarly, THC pretreatment in rodents has no effect on psychostimulant reward or self-administration, nor does it promote the development of a conditioned place preference to amphetamine, indicating a unique effect of cocaine and cannabis on drug craving. Interestingly, CBD has no effect on cocaine self-administration, motivation, or cue-induced reinstatement of cocaine-seeking, suggesting the effects of cannabis on cocaine craving are likely due to THC, rather than CBD. However, CBD treatment has been found to reduce motivation to self-administer methamphetamine on a progressive ratio schedule and to reduce methamphetamine-primed reinstatement of drug-seeking. The additive effects of cannabinoid and psychostimulant polydrug use appear to depend on both the amount of drug consumed and the age range during use. For example, acute THC weakens psychomotor sensitization, but repeated THC administration promotes tolerance to the acute effects, increasing amphetamine-induced stereotypy and locomotor activity. Additionally, adolescent THC exposure accelerates acquisition of cocaine self-administration and increases intake of low doses of cocaine, indicating long-lasting changes in reward circuitry following adolescent THC use, similar to alcohol.

Nicotine

Limited research has focused on the effects of polydrug use on nicotine-induced addiction behaviors. However, it has been found that THC pretreatment can enhance nicotine consumption and price inelasticity in behavioral economic tests in rats, and heroin intake has been found to increase cigarette consumption in people. Additionally, pre-exposure to alcohol or simultaneous access to both alcohol and nicotine decreases nicotine self-administration in rodent studies. Although access to alcohol has no effect on responding for nicotine under extinction conditions, an initial dose of alcohol does reinstate nicotine-seeking.

Finally, systemic co-administration of methamphetamine and nicotine produces a conditioned place aversion in rats, whereas pretreatment with either amphetamine or morphine increases the rewarding properties of nicotine as shown by lowered intracranial self-stimulation thresholds. These studies further emphasize the need to consider use patterns and dose when interpreting the effects of polydrug use.

Opioids

Similar to psychostimulant polydrug studies, sequential use of heroin and cocaine has not been found to alter heroin self-administration or the return to heroin-seeking. Although alcohol pretreatment can prevent the long-term reconsolidation of preference for morphine in drug-paired contexts, adolescent alcohol exposure enhances the development of a morphine conditioned place preference (CPP). This finding indicates that the long-term effects of adolescent alcohol exposure are generalizable to multiple drug classes. Interestingly, co-administration of morphine and THC prevents the development of analgesic tolerance that normally accompanies long-term exposure to either drug alone. In addition, the pain-relieving effects of THC and oxycodone co-administration are additive to oxycodone alone.

Moreover, administration of either THC or both THC and CBD attenuates naloxone-precipitated withdrawal without impacting the development of a morphine CPP. These data suggest a potential role for cannabinoids in regulating physical dependence to opioids without altering their reinforcing properties. In support of this, repeated THC administration has no effect on breakpoint during a progressive ratio test of heroin self-administration or relapse to heroin-seeking, although it produces a small reduction in both heroin and oxycodone intake in fixed-ratio self-administration sessions. The effects of opioid and cannabis polydrug use, however, appear to be dose-dependent as systemic administration of THC prior to heroin self-administration reduces responding for large doses of heroin, but has no effect on responding for lower doses in both monkeys and rats.

Cannabinoids

As with psychostimulants and opioids, administration of nicotine with THC amplifies the effects of either drug alone when measured in tests of movement, pain relief, and hypothermia. Additionally, THC and nicotine co-administration worsens the physical symptoms of THC withdrawal. However, after repeated nicotine treatment and 2 weeks of drug abstinence, nicotine re-administration reduces THC-induced decreases in movement, increases in anxiety measures, and changes in social interaction. These findings suggest that nicotine enhances the negative symptoms of THC when administered concurrently or in close temporal proximity. Although nicotine pretreatment enhances the rewarding effects of subthreshold doses of THC, cocaine pretreatment heightens THC-induced anxiety-provoking behaviors.

Cannabis and alcohol polydrug use is relatively common in humans, and individuals report reduced alcohol consumption when cannabis is available, suggesting a role for cannabinoids in alcohol intake. However, drug-induced cognitive and physical impairments in humans, as assessed in a driving simulation, were found to be more severe after use of THC and alcohol compared to either drug alone. Conducting polydrug studies of combinations of THC or CBD with other drug classes is therefore necessary to understand the different effects resulting from these drugs.

Alcohol

Polydrug use of alcohol and nicotine produces mixed outcomes in relation to addiction behaviors. For example, pre-exposure to alcohol or simultaneous access to both alcohol and nicotine increases alcohol self-administration, but not when nicotine is administered prior to alcohol each day. Chronic nicotine treatment also enhances alcohol preference, an effect that persists through nicotine withdrawal. However, although access to nicotine impairs extinction learning for alcohol-seeking, it has no effect on the reinstatement of drug-seeking, as rats respond similarly on alcohol and nicotine-associated levers following an initial injection of nicotine. Nonetheless, another study found that initial doses of alcohol, but not nicotine, were capable of reinstating alcohol-seeking following self-administration of both nicotine and alcohol. Studies have not systematically examined the effects of other drugs on alcohol use and addiction.

Neurobiology of Addiction

The development and persistence of addiction behaviors partly arise from harmful changes in the brain circuits responsible for decision-making, learning, motivation, and reward processing. In particular, disruptions in the cortico-basal ganglia-thalamic (C-BG-T) network are known to contribute to drug-taking and drug-seeking behaviors, as well as the persistence of SUDs. The C-BG-T is a highly interconnected network that integrates sensory and internal cues to drive motivated behavior. The striatum, which acts as a central hub in the C-BG-T, receives extensive excitatory input from cortical areas (like the prefrontal cortex) and subcortical regions (like the amygdala, hippocampus, and thalamus), along with dopamine input from the midbrain (substantia nigra/ventral tegmental area). The integration of these excitatory and dopamine inputs with local inhibition in the striatum helps initiate or suppress behavior, and an imbalance in signaling between the two striatal output pathways (the direct and indirect) can lead to addictive behaviors. It is beyond the scope of this review to fully explore all brain changes that occur in the C-BG-T with drug use. Instead, the focus will be on one microcircuit within the C-BG-T that is central to the immediate effects of addictive drugs and to some of the lasting changes that develop after long-term drug use: the prefrontal cortex (PFC) – nucleus accumbens (NAc) – ventral tegmental area (VTA) network. After reviewing the microcircuitry and connections of these regions, disruptions that occur within this network after both acute and long-term exposure to different drug classes will be discussed, emphasizing the similarities and/or differences of effects relative to polydrug combinations.

Prefrontal Cortex

The prefrontal cortex (PFC) is centrally involved in reward learning, decision-making, and evaluating outcomes. It is a highly diverse structure, which adds to the complexity of understanding its role in cognition and how its dysfunction contributes to drug use and addiction. In general, the medial prefrontal cortex (mPFC)—which includes the anterior cingulate (ACC), prelimbic (PrL), and infralimbic (IL) cortices—regulates motivation and seeking of both natural and drug rewards through excitatory glutamatergic projections to the nucleus accumbens (NAc) and ventral tegmental area (VTA). Specifically, projections from the mPFC to the NAc are organized such that the PrL innervates the NAc core and the IL innervates the NAc shell. In contrast, the orbitofrontal cortex projects more strongly to the dorsal striatum and substantia nigra and is primarily involved in evaluating outcomes and probabilities. Reduced activity in the PFC contributes to drug craving and seeking despite negative consequences during preoccupation stages of addiction, with disrupted connections to the striatum and VTA contributing to cue sensitivity and motivated drug-taking.

The PFC consists of six layers, each with unique connectivity patterns and distinct cell types. Specifically, most PFC neurons are large pyramidal output cells (75%), along with several subtypes of interneurons (about 25%). Pyramidal cells in layers II/III send local projections within the cortex, while those in layers V-VI send projections throughout the C-BG-T, including to the striatum, midbrain, amygdala, hippocampus, and thalamus. Pyramidal cells can be further categorized based on their physiology and connectivity. Recent studies have begun to characterize the anatomical, electrophysiological, and molecular profiles of each of these cell types, though how each regulates behavior remains poorly understood.

Finally, the PFC contains multiple populations of interneurons that heavily regulate cortical output by projecting to both pyramidal neurons and other interneurons.

Nucleus Accumbens

The striatum is a diverse structure composed primarily of two intermingled populations of GABAergic medium spiny neurons (MSNs) that can regulate behavioral output in two directions. Direct pathway MSNs (dMSNs) express dopamine D1-like (D1) receptors and the neuropeptides dynorphin and substance P, project directly to the midbrain, and can promote behavior by serving as a "go" signal. Conversely, indirect pathway MSNs (iMSNs) express dopamine D2-like (D2) receptors and the neuropeptide enkephalin, project indirectly to the midbrain via the pallidum, and can suppress behavior by serving as a "stop" signal. Drug use promotes increased burst-like dopamine release from D1 activation, leading to the assignment of reward to drug use during binge/intoxication phases of the addiction cycle, conditioning, and associating drug-taking contexts with incentive salience.

The striatum has dorsal and ventral compartments, with further subdivisions based on connections and function. The ventral striatum—comprising the olfactory tubercle, nucleus accumbens (NAc) core, and NAc shell—receives dopamine modulation from the ventral tegmental area (VTA) and excitatory input from the prefrontal cortex (PFC), as well as thalamic, hippocampal, and amygdala nuclei. In general, the ventral striatum regulates motivated behavior and reward learning. However, it has been hypothesized that an ascending loop between the ventral and dorsal striatum facilitates information consolidation during learning, where habitual behaviors shift from the ventral striatum to the dorsal striatum, contributing to compulsive drug-seeking. Importantly, while dMSNs and iMSNs have historically been distinguished by their downstream targets and expression of dopamine receptors, ventral striatal dMSNs also send branches to the VP, and D1 and D2 receptors are co-expressed to some extent in the NAc core (6–7%) and NAc shell (12–15%).

In addition to MSNs, the striatum contains large, tonically active cholinergic interneurons and multiple subtypes of GABAergic interneurons with distinct electrical properties and peptide expression patterns. MSNs also receive cholinergic modulation from other projections, though the relevance of these inputs to local or network dynamics and the role of cholinergic striatal neurons in the addiction cycle remains uncertain. Interestingly, each MSN receives approximately 5000–15000 excitatory inputs in addition to about 1200–1800 GABAergic inputs from other MSNs; therefore, modulation of MSN activity via cholinergic and dopamine inputs appears necessary for signal integration and effective synaptic plasticity. Indeed, MSNs exhibit bi-stability, existing almost exclusively in either a down-state (−80 mV) or an up-state (−50 mV, near threshold) in the absence of external input. Furthermore, the maintenance of bi-stability and intrinsic excitability relies on the activity of cation channels that are under robust dopamine and cholinergic modulation. Maintaining intrinsic excitability within MSNs is critical for the normal regulation of behavioral output, and dysregulation of striatal microcircuitry contributes to the development and expression of addiction behaviors.

Ventral Tegmental Area

The ventral tegmental area (VTA) sends dopamine projections to cortical, striatal, and subcortical areas (e.g., hippocampus, amygdala, thalamus) to modulate cortical-basal ganglia-thalamic (C-BG-T) network activity. Dopamine projections from the VTA to the nucleus accumbens (NAc) regulate goal-directed behaviors and have been strongly implicated in the binge/intoxication phase of substance use disorders (SUDs). Recently identified subtypes of VTA dopamine (VTADA) neurons with unique molecular profiles preferentially project to the NAc core or NAc shell, and regulate reward learning or motivation, respectively. However, co-activation of both populations appears necessary for robust reinforcement. VTADA neurons receive dense excitatory input from the prefrontal cortex (PFC) and several midbrain structures, as well as inhibitory input from a variety of sources, including the NAc and local VTA GABA neurons.

VTAGABA neurons maintain tonic dopamine levels by inhibiting VTADA neurons, and brief disinhibition of VTADA neurons results in burst-like dopamine release into the NAc. VTAGABA neurons receive excitatory input from the PFC and a number of subcortical nuclei, as well as inhibitory input from throughout the brain, including the NAc. The VTA also contains glutamatergic neurons that project to striatal interneurons, though the inputs to and behavioral relevance of these neurons are unknown. Notably, the VTA contains subpopulations of dopamine neurons that can release glutamate and/or GABA, allowing the VTA to modulate local C-BG-T activity at multiple levels and time scales.

Within the NAc, the activity of VTADA neurons is modulated by dopamine D2 autoreceptors on VTADA terminals and cholinergic interneurons.

Mechanisms of Addiction: Acute Drug Effects

A unique feature of all potentially addictive drugs is their ability to strengthen the binge/intoxication phase of the addiction cycle through triggered burst-like dopamine release into the nucleus accumbens (NAc); however, the underlying mechanisms differ across drugs. Psychostimulants disrupt dopamine reuptake into VTA dopamine (VTADA) terminals, nicotine and alcohol directly activate VTADA neurons, and opioids and cannabinoids disinhibit VTADA neurons. The synergistic and antagonistic interactions between different drugs within the C-BG-T add complexity to the study of polydrug use, and little is known about the mechanisms underlying the immediate effects of specific polydrug combinations. However, in vivo extracellular recordings in rats that alternately self-administered cocaine and heroin in the same session found that only about 20% of prefrontal cortex (PFC) and NAc neurons responded similarly to both drugs, indicating divergent engagement of the C-BG-T by these drugs. Thus, examining the synergistic and antagonistic mechanisms of different drugs can guide the understanding of how specific polydrug combinations may disrupt C-BG-T network dynamics and contribute to the manifestation of addiction behaviors.

Psychostimulants

All psychostimulants directly enhance striatal dopamine (DA) release by disrupting dopamine transporter (DAT) activity, though they do so through distinct mechanisms. Cocaine blocks DAT-mediated reuptake of DA, while amphetamine reverses DAT activity and induces DA release from VTA dopamine (VTADA) terminals. Psychostimulants also immediately increase glutamate transmission in the prefrontal cortex (PFC), nucleus accumbens (NAc), and VTA, indicating broad increases in activity throughout the C-BG-T network following psychostimulant use. The combination of enhanced glutamatergic and dopaminergic input to the NAc facilitates the transition of medium spiny neurons (MSNs) to the up-state and activates DA-dependent signaling pathways. For example, psychostimulants immediately increase the activation of the immediate early gene Fos in striatal direct pathway MSNs (dMSNs) and indirect pathway MSNs (iMSNs). Fos encodes several proteins, including ΔFosB, which has been widely implicated in addiction pathology, and enhanced activation in the NAc is thought to contribute to long-term disruptions in normal C-BG-T activity.

Co-administration of nicotine, alcohol, or heroin enhances psychostimulant-induced DA release into the NAc, though specific combinations do so through different mechanisms. For example, simultaneous administration of cocaine and nicotine activates VTADA neurons and disrupts DA reuptake, resulting in a greater magnitude of DA release into the NAc than that evoked from either drug alone. Notably, some polydrug combinations that include psychostimulants have different and opposing effects on the C-BG-T circuit. Co-administration of alcohol with cocaine induces the liver to produce cocaethylene, which can target DAT and presynaptic D2 autoreceptors to amplify DA release. However, alcohol also prevents cocaine-induced glutamate transmission in the NAc core.

Lastly, polydrug studies with psychostimulants have identified an exacerbation of drug-induced cellular toxicity compared to psychostimulant use alone. Specifically, co-administration of heroin with cocaine decreases metabolic activity, increases intracellular calcium signaling, and decreases mitochondrial membrane potential. Collectively, these effects contribute to enhanced caspase 3-dependent apoptotic activity and subsequent cell death compared to either drug alone.

Nicotine

Unlike psychostimulants, nicotine enhances dopamine (DA) release by activating nicotinic acetylcholine (nACh) receptors within the ventral tegmental area (VTA). nACh receptors are non-selective ion channels, permeable to sodium, potassium, and calcium, and their activation leads to depolarization and enhanced neurotransmitter release. The reinforcing effects of nicotine are primarily due to nicotine's activity on nACh receptors located on VTA dopamine (VTADA) neurons and on presynaptic nACh receptors located on prefrontal cortex (PFC) glutamatergic inputs.

Acute nicotine exposure activates both presynaptic PFC inputs and VTADA cell bodies, triggering burst-like DA release into the nucleus accumbens (NAc). Notably, nicotine-evoked DA release is blocked by glutamate receptor blockade or activation of GABAB receptors, highlighting the extensive regulation of VTADA neuron activity by local VTA microcircuitry. Studies have not examined how the immediate effects of nicotine are changed by polydrug use.

Opioids

Although opioids are distinguished by their origin, potency, and receptor bias factor, all opioids exert their rewarding effects by activating mu opioid (μO) receptors. μO receptors are expressed both on the cell body and dendrites and on axons, but the primary mechanism of opioid-induced dopamine (DA) release relies on presynaptic inhibition of VTA GABA (VTAGABA) neurons. μO receptors are inhibitory G protein-coupled receptors (GPCRs), and their activation reduces neuronal excitability through four mechanisms: (1) Gα-mediated inhibition of cAMP-dependent signaling pathways (e.g., PKA, CREB), (2) Gβγ-mediated activation of G protein-coupled inwardly rectifying K+ (GIRK) channels, (3) Gβγ-mediated inactivation of voltage-gated calcium channels, and (4) Gβγ-mediated inhibition of SNARE-dependent vesicle release.

Acute exposure to opioids inhibits VTAGABA neurons, resulting in subsequent disinhibition of VTA dopamine (VTADA) neurons and burst-like DA release into the nucleus accumbens (NAc). Nonetheless, although opioids facilitate DA release into the striatum, whether this DA transmission is necessary for opioid reward remains debated. Opioids activate VTADA neurons in living organisms and increase DA in the NAc, but neither NAc lesions nor systemic blockade of DA receptors affect opioid reward.

Given the different mechanism of action for opioids compared to psychostimulants and nicotine, it is not surprising that co-administration of these drugs enhances the immediate effects of opioids. Indeed, simultaneous administration of opioids and psychostimulants produces an additive increase in DA release in the NAc and prolongs elevated levels of DA and its metabolites, DOPAC and HVA. Similarly, simultaneous administration of opioids and nicotine enhances opioid-evoked DA release in the NAc and dorsal striatum.

Cross-tolerance to opioid-mediated pain relief has also been shown following pre-exposure to nicotine and cannabis, and chronic nicotine treatment dose-dependently reduces pain-relieving tolerance to opioids in a nACh receptor-dependent manner. Interestingly, pretreatment with calcium channel blockers or naloxone prevents this tolerance, suggesting a complex pharmacological interaction between opioids and nicotine. Notably, cross-tolerance to opioid analgesia is mediated by different mechanisms for various polydrug combinations. Nicotine and opioid cross-tolerance is mediated by μO and nACh receptors, while cannabinoid and opioid cross-tolerance is mediated by μO receptors and cannabinoid 1 (CB1) receptors.

Cannabinoids

Cannabis contains two main cannabinoids with varying affinity for cannabinoid (CB) receptors: THC is a partial agonist with moderate affinity for both CB1 and CB2 receptors, whereas CBD has extremely low affinity for CB1 and CB2 receptors and signals through an unknown mechanism. Interestingly, pretreatment with a range of CBD doses has no effect on THC self-administration, indicating non-overlapping signaling pathways for each cannabinoid. CB1 receptors are expressed on presynaptic terminals throughout the central nervous system (CNS) and are responsible for the psychoactive effects of cannabis, while CB2 receptors are primarily expressed on immune cells of the CNS and peripheral nervous system (PNS) and are mainly responsible for the pain-relieving and anti-inflammatory effects of cannabis. Both subtypes signal through inhibitory G protein-coupled receptor (GPCR) signaling pathways (similar to opioids), and activation of the receptors results in presynaptic inhibition (via activation of GIRK channels and inhibition of voltage-gated calcium channels) and downregulation of cAMP-dependent signaling cascades.

CB1 receptors are heavily expressed on presynaptic terminals of VTA GABA (VTAGABA) neurons as well as nucleus accumbens (NAc) direct pathway medium spiny neurons (dMSNs) that target VTA dopamine (VTADA) neurons, and activation of CB1 receptors reduces GABA-mediated inhibitory postsynaptic currents in VTA slices. CB1-mediated disinhibition of VTADA neurons results in an increase in burst firing and subsequent release of dopamine into the NAc, and these effects are blocked by systemic or intra-VTA naloxone.

Similar to other drugs, the immediate effects of cannabinoids on C-BG-T network activity are broad and involve multiple neurotransmitter systems. For example, acute THC increases both dopamine and glutamate signaling in the NAc and prefrontal cortex (PFC), but decreases GABA signaling in the VTA and PFC. Collectively, these alterations in signaling reduce inhibitory feedback within the C-BG-T and facilitate behavioral output.

Alcohol consumption prior to cannabis use enhances plasma THC levels and increases self-reported euphoria in humans, indicating synergistic effects between the two drugs. Moreover, simultaneous administration of cannabinoids with psychostimulants or opioids enhances activation of VTADA neurons, and simultaneous administration of THC and nicotine increases cFos activation throughout the C-BG-T.

Alcohol

Despite its widespread use, much less is understood about the mechanisms underlying the immediate effects of alcohol. Alcohol activates dissociated VTA dopamine (VTADA) neurons and induces dopamine (DA) release into the nucleus accumbens (NAc), similar to other potentially addictive drugs. However, GABAA receptors are known to play a central role in the effects of alcohol. For example, alcohol enhances GABAA signaling both in cortical slices and neuronal cultures. Given that NAc direct pathway medium spiny neurons (dMSNs) selectively inhibit VTA GABA (VTAGABA) neurons via GABAA-mediated signaling, it is possible that alcohol facilitates burst-like DA release from VTADA neurons by inactivating local VTAGABA neurons. In support of this hypothesis, alcohol inhibits VTAGABA neurons, and intra-VTA infusion of GABAA agonists dose-dependently increases DA release.

In addition to GABA, the acute effects of alcohol depend on glutamatergic signaling within the C-BG-T. Alcohol increases glutamate release in the NAc and VTA by activating presynaptic D1 receptors, suggesting that alcohol engages a feedforward loop for the activation of VTADA neurons.

Polydrug use with alcohol produces synergistic effects throughout the C-BG-T, likely as a result of alcohol’s unique pharmacological profile. For example, chronic pretreatment with nicotine enhances acute alcohol-induced DA release in the NAc, and elevated levels of DA and its metabolites, DOPAC and HVA, persist for over an hour.

Additionally, alcohol and nicotine co-administration immediately increase the production of BDNF and GDNF in the NAc, along with increases in glutamatergic signaling in the VTA and PFC. Notably, this widespread activation of the C-BG-T network is absent following administration of either drug alone, demonstrating a unique mechanism of action for alcohol and nicotine polydrug use.

Mechanisms of Addiction: Long-Term Alterations

Long-term use of psychostimulants, nicotine, opioids, cannabinoids, and alcohol results in widespread and distinct changes throughout the C-BG-T network, yet there are notable alterations that are shared across drugs. These long-term adaptations contribute to transitions from binge/intoxication phases to withdrawal and negative affect, followed by preoccupation and compulsive drug-seeking. For example, acute withdrawal produces a temporary reduction in tonic dopamine (DA) levels in the nucleus accumbens (NAc). This is followed by a persistent increase in the excitability of VTA dopamine (VTADA) neurons, which contributes to enhanced cue-evoked burst-like DA release during abstinence.

Conversely, withdrawal also produces a persistent reduction in long-term depression (LTD) and intrinsic excitability in NAc medium spiny neurons (MSNs), as well as a reduction in striatal D2 receptor binding. As described earlier, striatal D2 receptors are primarily expressed on indirect pathway MSNs (iMSNs), can serve as a "stop" signal on the C-BG-T circuit, and reduced D2 availability has been consistently linked to various addictive diseases, including drug addiction and obesity. Finally, these drugs all drive a persistent increase in ΔFosB expression in cortical pyramidal cells and NAc direct pathway MSNs (dMSNs), which has been linked to drug-seeking during abstinence. Importantly, although chronic use of any of these drugs results in numerous other temporary and/or persistent changes across the C-BG-T network, providing an exhaustive summary is beyond the scope of this review. Instead, the focus will be on changes in plasticity, morphology, and connectivity within the VTA, NAc, and prefrontal cortex (PFC) following both single and polydrug use, with an emphasis on how the antagonistic and synergistic effects of these drugs can differently disrupt C-BG-T network dynamics.

Psychostimulants

Psychostimulants cause long-term disruptions in glutamate balance and changes in neuronal structure throughout the C-BG-T network. Repeated cocaine administration weakens GABAA-mediated inhibition of prelimbic (PrL) pyramidal neurons, increasing their excitability and boosting excitatory drive to the nucleus accumbens (NAc). Similarly, chronic cocaine increases glutamatergic input to the ventral tegmental area (VTA) and weakens GABAB-mediated inhibition of VTA dopamine (VTADA) neurons, leading to facilitated VTADA neuron activity.

Additionally, psychostimulants create silent synapses on direct pathway medium spiny neurons (dMSNs) through the formation of new synapses and reduce glutamate release from PrL inputs to the NAc core, weakening striatal output. Importantly, cocaine-silenced synapses on dMSNs can become active again during withdrawal by recruiting AMPA receptors, and a low dose psychostimulant challenge restores glutamate release into the NAc core, suggesting the promotion of allostasis (the process by which the body maintains stability through physiological or behavioral changes).

Finally, repeated psychostimulant exposure increases the expression of ΔFosB in prefrontal cortex (PFC) pyramidal neurons and NAc dMSNs, which contributes to an increase in dendritic branching that persists for at least 30 days of abstinence. Studies have not examined how these effects of psychostimulants are changed by polydrug use.

Nicotine

Similar to psychostimulants, long-term nicotine administration reduces GABAB-mediated signaling in the medial prefrontal cortex (mPFC) and nucleus accumbens (NAc), thereby dampening inhibitory drive onto the C-BG-T network. Additionally, nicotine enhances glutamatergic input to VTA dopamine (VTADA) neurons and sensitizes evoked dopamine release into the NAc. Repeated nicotine use also leads to an increase in nicotinic acetylcholine receptors throughout the C-BG-T, including the PFC and midbrain.

Conversely, chronic nicotine exposure weakens both glutamatergic and GABAergic inputs to the NAc during abstinence, potentially through enhanced sensitivity of NAc D2 receptors. Because NAc output is heavily regulated by local GABAergic microcircuitry, increased activity of indirect pathway medium spiny neurons (iMSNs) would likely dampen overall NAc output. This is supported by the upregulation of calcium-permeable AMPA receptors in the NAc that are capable of conducting calcium in the absence of NMDA receptor activation.

Long-term nicotine use also produces a persistent increase in mPFC pyramidal cell and NAc MSN dendritic branching, similar to psychostimulants. Moreover, adolescent nicotine use enhances synaptic pruning, microglial activation, and inflammatory cytokine expression throughout the C-BG-T via a D2 receptor-mediated mechanism. Repeated co-administration of psychostimulants and nicotine increases long-term potentiation (LTP) induction and ΔFosB in the NAc core. Interestingly, these effects are absent following sequential administration of these drugs, indicating a unique pharmacological profile for concurrent nicotine and psychostimulant administration.

Opioids

Similar to psychostimulants and nicotine, long-term exposure to opioids strengthens glutamatergic input to VTA dopamine (VTADA) neurons. They also weaken GABAergic inhibition of VTADA neurons through a reduction of direct pathway medium spiny neuron (dMSN)-mediated GABAB inhibition. This effect is largely driven by MSNs in the nucleus accumbens (NAc), indicating a persistent disruption in NAc output. Unlike psychostimulants and nicotine, however, long-term exposure to opioids decreases dendritic branching and spine density in NAc MSNs and medial prefrontal cortex (mPFC) pyramidal cells.

Within the NAc, long-term opioid exposure weakens glutamatergic input to NAc shell indirect pathway MSNs (iMSNs) and induces silent synapses on iMSNs via AMPA receptor internalization. Moreover, during withdrawal, opioid-generated silent synapses on iMSNs are eliminated, and the intrinsic excitability of iMSNs is weakened. Given that MSNs create a dense network of lateral inhibition within the NAc, with about 30% of iMSNs synapsing onto other iMSNs or dMSNs, it is possible that opioid-induced disruptions in iMSN signaling could disrupt local NAc microcircuitry and facilitate abnormal C-BG-T network dynamics.

Studies have not examined how these effects of opioids are changed by polydrug use.

Cannabinoids

Repeated THC administration weakens glutamatergic signaling from the medial prefrontal cortex (mPFC) to the nucleus accumbens (NAc) but strengthens input from the basolateral amygdala and ventral hippocampus to the NAc shell, suggesting a rewiring of excitatory input to the NAc following long-term cannabinoid exposure. Notably, repeated THC administration also blocks cannabinoid 1 (CB1)-mediated long-term depression (LTD) on VTA GABA (VTAGABA) neurons, indicating a loss of local inhibitory drive on VTA dopamine (VTADA) neurons.

Additionally, adolescent THC administration followed by abstinence reduces intrinsic excitability of prelimbic (PrL) neurons and weakens glutamatergic input to the VTA, indicating a state of low glutamate induced by cannabinoids. THC exposure also produces long-lasting changes in structure throughout the C-BG-T, increasing dendritic spine length and branching in the mPFC and NAc shell and dendritic spine density in the NAc shell, but reducing the size of VTADA neurons.

Repeated co-administration of THC and nicotine enhances expression of ΔFosB in the NAc, and acute withdrawal dysregulates glutamatergic input to the NAc and PFC.

Alcohol

Similar to opioids, extended alcohol exposure decreases dendritic branching and spine density in the nucleus accumbens (NAc) shell and medial prefrontal cortex (mPFC), and withdrawal from alcohol reduces tonic dopamine (DA) levels in the NAc. However, alcohol withdrawal also causes distinct changes throughout the C-BG-T network, including reduced GABAergic signaling in the NAc and hippocampus and increased glutamatergic signaling in the NAc and PFC.

Notably, this increased glutamatergic signaling is due to disrupted glutamate reuptake, rather than enhanced glutamate release. In vitro polydrug exposure to alcohol and nicotine induces a 2.5-fold increase in caspase-3 activation, elevating apoptotic cascades and driving cell death. Interestingly, however, alcohol or nicotine withdrawal-induced neurodegeneration is less severe following co-administration of both drugs, indicating a unique molecular pathology following nicotine and alcohol polydrug use.

Conclusion

Improving how well behavioral measures in preclinical models reflect polysubstance history and dependency in clinical populations is essential. This is particularly true because human imaging studies are limited by the inability to control for intake history, making behavioral models in other species advantageous for assessing polydrug history under controlled intake conditions. However, these models have limitations in fully capturing the complex social and environmental contexts that contribute to the unique use patterns for multiple addictive substances. Nevertheless, when designing experiments in clinical or preclinical populations, factors such as time of day of intake, temporal proximity of intake, and environmental preferences for administration must be considered for each substance class. For instance, cocaine use is predominantly favored outside of home environments, whereas heroin use is greater in "home" contexts in both humans and rodents. Additionally, the route of drug administration (e.g., oral ingestion, injection, and inhalation) is especially important to factor into studies, as it leads to unique patterns of polysubstance history that may impact the development and severity of addiction behaviors.

Accounting for temporal patterns of drug use is also essential. Although many studies have focused on simultaneous drug combinations, sequential patterns of polydrug consumption are more frequently reported and produce unique circuit adaptations following immediate and repeated drug exposure. Understanding sex differences in the frequency and pattern of polydrug use, drug discrimination, and circuit alterations is also necessary to fully understand the interactions and impacts of polydrug use in clinical populations. Furthermore, although powerful behavioral economic models allow comparisons across drug classes, these experiments must be designed with consideration of different scales of intake and indifference points for drug valuation to accurately model parameters and interpret data. As these factors, coupled with the interplay of socioeconomic background, social environment, and access strategies, affect the frequency of and susceptibility to polydrug use, incorporating these influences into preclinical models is crucial for the relevance of findings to human conditions.

It is important to appreciate that new preclinical paradigms are continuously being developed to model different routes of drug administration and to study relapse under clinically relevant conditions. For instance, recently established models allow for voluntary control over self-administration of vaporized ethanol, cannabis, and nicotine. In addition, multiple models for inducing drug abstinence are being introduced. For example, voluntary abstinence is achieved by pairing drug-taking with adverse consequences such as shock-lever pairings or electrified barriers placed in front of levers, and through choice procedures involving a drug versus an alternative reinforcer (e.g., palatable food or social interaction). These paradigms are notable for their high relevance to clinical practice and may be powerful for understanding the neural basis of therapies such as contingency management, in which abstinence from drug use results in a monetary reward. Combining these models with behavioral economic paradigms could clarify how the relative value of drugs and alternative reinforcers changes with polydrug use to improve treatment effectiveness. Addiction is a complex disease with multiple, highly variable factors contributing to the initiation and maintenance of drug use, as well as relapse. Given the widespread prevalence of polydrug use among drug users, it is critical that this variability be incorporated into human studies and animal models. This will help determine if polysubstance use exacerbates SUD severity, if increased SUD severity drives polysubstance use, or if there is a bidirectional relationship between the two. Though challenging, understanding the behavioral, genetic, and environmental contributions to polysubstance use and addiction, as well as the mechanisms that underlie addiction severity and relapse, will aid in developing effective treatment and policy strategies to combat this ongoing public health crisis.

Introduction

Drug addiction is a complex condition marked by cycles of drug use, withdrawal, and intense cravings, often leading to a return to use. This issue is widespread, affecting millions globally and in the United States. Lifetime surveys suggest about 10% of people will experience a substance use disorder. Drug addiction also poses a significant public health challenge in the U.S., costing hundreds of billions annually due to treatment, lost work, healthcare, and crime. These costs are likely to rise as illegal drug use increases worldwide, with millions starting drug use each year. Opioid use, in particular, continues to grow, leading to a dramatic increase in deaths.

While much research has focused on individual substances, it is important to recognize that many people who use drugs engage in polysubstance use, meaning they use more than one substance. This can involve using multiple drugs at different times or at the same time. Focusing only on individual drugs risks missing how substances interact, which can limit the effectiveness of treatments. Polysubstance use has consistently been linked to worse treatment outcomes, including shorter treatment periods, higher rates of relapse, and a greater risk of death compared to using only one substance. This document aims to combine current knowledge about how individual drugs affect the body with recent research on polysubstance use, noting different patterns of use when possible. It will review public health trends for single and polysubstance use, discuss findings from preclinical studies, and detail how individual substances affect brain circuits. Special attention will be paid to how polysubstance use changes the brain, highlighting common drug combinations. The goal is to provide suggestions for future research into polysubstance use disorders.

Public Health Trends in Substance Use

Drug addiction is both common and deadly, with hundreds of thousands of drug-related deaths occurring worldwide each year. While attention often focuses on individual drugs, drug misuse frequently involves multiple substances. Individuals dependent on drugs often report using several substances, either at the same time or one after another. It is also common for people to develop dependencies on more than one substance. Although the specific combinations vary, primary drug dependencies typically involve alcohol, opioids, and certain stimulants, while cannabis and cocaine are often used as secondary or tertiary substances.

The high rate of polysubstance use is concerning because it can worsen substance use disorders and make treatment less successful. For instance, a history of polysubstance use is linked to greater unmet needs for physical and mental healthcare, increased risky behaviors, violence, and a higher risk of overdose and death compared to using only one substance. This section will discuss public health trends related to both single substance use and polysubstance use, focusing on common drug combinations and usage patterns. However, it is important to note that available data is often limited regarding specific drug use patterns and how polysubstance combinations are defined across studies.

Psychostimulants

Psychostimulants are among the most commonly used classes of drugs globally. While the overall use of psychostimulants remained fairly stable for many years, the number of drug-related overdose deaths involving these substances has continued to rise significantly, especially in the United States.

Many people who use cocaine and amphetamines are also polysubstance users. Cocaine use and dependence are frequently linked to concurrent use of heroin, cannabis, tobacco, and alcohol. Similarly, amphetamine users often combine the drug with alcohol, tobacco, and cannabis. Some amphetamine users also report using heroin and other opioids. Simultaneous use of psychostimulants and opioids is common, such as combining cocaine with heroin (known as a "speedball") or methamphetamine with opioids (a "bombita"). People may also use psychostimulants and opioids sequentially, such as using cocaine or amphetamine to avoid opioid withdrawal symptoms, or using opioids to calm down after cocaine use. There is also an increased likelihood of using methamphetamine and alcohol on the same day.

Identifying exact patterns of polydrug use can be complex due to varying definitions and study methods. However, it is clear that polysubstance use involving psychostimulants carries significant public health risks. For example, amphetamine users are much more likely to have a concurrent cannabis use disorder or to have used cocaine in the past year. A large portion of overdose deaths involve both psychostimulants and opioids. Combining cocaine and cannabis has also been linked to higher death rates in emergency department visits. Additionally, using cocaine and alcohol together increases the risk of heart problems compared to using either drug alone.

Nicotine

Although tobacco use has decreased since the early 2000s, nicotine remains one of the most widely used drugs. Millions of people in the U.S. report using nicotine each month. It is also commonly used with many other substances; a significant percentage of nicotine users also use cannabis, opioids, cocaine, or psychostimulants in the same month. This difference is striking, as people with a nicotine use disorder are much more likely to have a second substance use disorder. Additionally, tobacco use has been strongly linked to opioid, cannabis, and alcohol use disorders, as well as cocaine use. The severity of tobacco use, such as how often someone uses it or how many cigarettes they smoke, has also been connected to the start of heroin and cocaine use.

Historically, nicotine was primarily consumed by smoking tobacco cigarettes. However, new technologies have led to the development of e-cigarettes, which are marketed as a safer alternative. This marketing has raised concerns because it has increased the likelihood of nicotine use and revived the potential for nicotine addiction, especially among young people. Despite claims of delivering lower doses of nicotine, the safety of commercial e-cigarettes is still debated, as some uses can lead to the production of harmful chemicals. The amounts of nicotine provided also vary widely among different e-cigarette products. The specific impact of vaping on nicotine dependence and how it contributes to polysubstance use disorders is largely unknown and requires further study.

Opioids

The non-medical use of opioids has increased dramatically in recent years, with millions worldwide reporting misuse in the past year. In the U.S., the number of first-time heroin users rose alongside the non-medical use of prescription opioids, suggesting that individuals who misuse prescription opioids are much more likely to begin using heroin. Studies on polysubstance use among those who misuse heroin and prescription opioids show higher rates of opioid use in people who also use cocaine or methamphetamine. However, there is less prevalence for primary opioid use among those who also use alcohol or cannabis. First-time methamphetamine use is also more common after recent opioid use. A large percentage of people entering treatment for heroin use report a history of cocaine use.

When it comes to patterns of multi-drug use, using heroin with alcohol or cannabis at the same time is more common than with psychostimulants. However, a sequential pattern of use is often preferred for opioids and psychostimulants.

Opioid use is a major public health crisis in the U.S., causing a significant loss of life. Opioid overdose deaths are now the leading cause of accidental death among U.S. adults. Given that a large majority of fatal opioid overdoses also involve another substance, the risk of death appears much higher when opioids are combined with other opioids or other drugs. Many opioid-related emergency department visits also involve tobacco, cocaine, other stimulants, cannabis, or alcohol. Combining opioid use with other substances can worsen the negative effects of opioid use. Opioid users also face very high rates of relapse. Previous use of other substances, including cocaine, increases the chance of relapse. Methamphetamine use among those seeking opioid treatment is also rising, and recent reports suggest it is linked to stopping buprenorphine treatment for opioid use disorder. Therefore, a better understanding of how polysubstance use affects opioids is essential for more successful emergency responses and long-term treatment.

Cannabinoids

Millions of people aged 12 or older use cannabis worldwide and in the U.S. The number of cannabis users in the U.S. has risen with its gradual decriminalization and legalization. Many individuals also use cannabis frequently, with a significant portion being daily or near-daily users.

Cannabis users often report recent use of tobacco, alcohol, and/or amphetamines. One of the most common combinations is using alcohol and cannabis at the same time, along with concurrent alcohol, cocaine, and cannabis use. The impact of concurrent cannabis use is notable, as this pattern is associated with more alcoholic drinks per day, suggesting it facilitates alcohol consumption. Polydrug use is particularly common among younger cannabis users. Cannabis is frequently used during treatment for other substance use disorders, which has been linked to reduced treatment success. For example, cannabis use has been found to shorten periods of alcohol abstinence and increase the likelihood of relapsing to cocaine use. Polydrug use among cannabis users has also been associated with reduced financial stability, relationship difficulties, worse mood disorder symptoms, poor decision-making, and self-harm. While these data suggest that the negative consequences of drug use are enhanced by concurrent cannabis use, it is worth noting that clinical outcomes can vary. Some studies suggest a nuanced impact, with no combined effects of cannabis and alcohol at low doses, and a lack of association between cannabis use and heroin relapse in some cases.

Unfortunately, less is known about the impact of cannabis compared to other drugs, partly due to past restrictions on cannabis research and difficulties in modeling its use in animal studies. However, the development of new methods for animals to self-administer cannabis, such as vaporized or intravenous forms, will help advance the study of cannabis use disorder and the consequences of polysubstance use involving cannabis.

Alcohol

Alcohol is one of the most commonly used drugs, with hundreds of millions of people worldwide diagnosed with an alcohol use disorder, including millions in the U.S. Alcohol is often used with other substances; a significant percentage of U.S. adults have used both alcohol and another illicit drug in the past year, and many meet the diagnostic criteria for both an alcohol use disorder and another substance use disorder. The most commonly reported substance used with alcohol is cannabis, with less common co-use found with opioids, cocaine, and amphetamines. While simultaneous use of alcohol and cannabis or alcohol and prescription opioids is most common, concurrent use with cocaine is also seen.

Polydrug use that includes alcohol increases the risk of developing an alcohol use disorder, particularly in young adults, men, and certain Native American populations. Polydrug use involving alcohol is also associated with additional health issues, including a higher prevalence of mood and anxiety disorders, more intense drinking, and greater drug consumption and craving. The negative effects of alcohol polysubstance use are further highlighted by data showing that a significant portion of emergency department visits for young patients involve both alcohol and drugs. These visits are also more likely to require treatment for injuries and have higher rates of inpatient admission.

The rate of hospitalizations involving alcohol polysubstance use has been rising, especially in young adults, indicating a significant increase in severe cases compared to overdoses from either drug or alcohol alone. While polysubstance users in past surveys were mainly white males, recent trends suggest a rise in emergency department visits related to alcohol and drug combinations in females, indicating a shift in who is affected by these combinations.

Behavioral Models of Addiction in Polydrug Studies

Behavioral models of drug addiction are used to study the brain processes that lead to the development, maintenance, and relapse of drug use. Common models include locomotor sensitization, which measures increased movement in response to a drug; conditioned place preference, which assesses drug reward by measuring time spent in a drug-associated environment; and drug self-administration, where animals can control their drug intake. These experimental designs vary in how drugs are administered, the amount of drug available, the environment where drugs are used, and the ways drugs are given. Researchers use these models to understand polysubstance use and addiction.

Early polysubstance studies often used models where drugs were given without the subject's control, such as to see how one drug affected the preference or activity related to another. More recent studies examine how a history of self-administering one drug affects the choice and craving for another drug. Models that involve simultaneous administration of multiple drugs, like alcohol and nicotine or cocaine and heroin combinations, have also been frequently used. In human studies, both concurrent and sequential polysubstance use are evaluated through measures of mood, craving in response to visual cues, and physical responses like blood pressure.

More recently, studies compare single versus polysubstance self-administration to see how a drug history affects brain changes. Paradigms based on economic principles can also determine how much drug an animal prefers to take and how much effort it is willing to exert for the drug. These methods are powerful because they allow for direct comparisons of the value and effort associated with different drugs and non-drug rewards in polysubstance models across various species, including humans. Such studies can assess how a history of using one substance impacts the perceived value of another, offering valuable insights into the impact of polysubstance use on reward valuation.

Effects of Polydrug Use on Addictive Behaviors

Given the unique brain changes that can occur with exposure to multiple drugs and the high prevalence of polysubstance use disorders, there is a strong need for studies that can accurately reflect real-world polysubstance use. These studies are crucial for fully understanding the behavioral changes and addiction-related patterns that develop after polysubstance use. However, due to the vast number of possible substance combinations and different study methods, the findings on how polysubstance use affects addiction behaviors are currently mixed.

Despite these challenges, some general trends in drug consumption, preference, and seeking have been observed for commonly investigated substance combinations. For psychostimulants, human reports often show an increase in addiction severity with polydrug use, though animal studies have not always replicated this, suggesting a need for more relevant models. Simultaneous use of psychostimulants and opioids or nicotine generally increases motivation and drug intake, and enhances drug-related behaviors. However, the effects of nicotine and psychostimulant co-use depend heavily on factors like dose and pattern of delivery. For opioids, simultaneous use with psychostimulants and nicotine often increases dopamine release in the brain and can lead to complex interactions regarding drug tolerance. Cannabinoids in combination with other drugs show varied effects, sometimes enhancing the impact of other substances and other times not, and more research is needed in this area. Alcohol in combination with nicotine can also produce mixed results on addiction behaviors, sometimes increasing alcohol intake or craving, but sometimes leading to less severe neurodegeneration than either drug alone. Overall, these findings highlight that the impact of polydrug use on addictive behaviors is highly complex and depends on the specific drugs, their doses, and how they are used together.

Neurobiology of Addiction

The development and persistence of addiction behaviors result partly from harmful brain changes within neural circuits that control decision-making, learning, motivation, and reward. Specifically, changes in the cortico-basal ganglia-thalamic (C-BG-T) network are known to contribute to drug use and seeking, as well as the chronic nature of substance use disorders. This network is highly interconnected and integrates sensory and internal cues to drive motivated behaviors. Key regions include the prefrontal cortex (PFC), which is involved in reward learning and decision-making; the nucleus accumbens (NAc), which regulates motivated behavior and reward learning; and the ventral tegmental area (VTA), which sends dopamine signals to other brain areas to modulate network activity. Imbalances in signaling within these regions can drive addictive behaviors.

All potentially addictive drugs share the ability to reinforce drug use by increasing dopamine release in the NAc, though they do so through different mechanisms. Psychostimulants interfere with dopamine reuptake, while nicotine and alcohol directly activate dopamine neurons. Opioids and cannabinoids increase dopamine by reducing the activity of neurons that normally inhibit dopamine release. These combined and opposing effects of different drugs within the C-BG-T network add to the complexity of studying polysubstance use. For example, simultaneously using cocaine and nicotine can lead to a greater increase in dopamine release than either drug alone. Some drug combinations can even lead to increased cellular toxicity and cell death compared to using a single drug.

Long-term use of psychostimulants, nicotine, opioids, cannabinoids, and alcohol leads to widespread changes throughout the C-BG-T network. These adaptations contribute to the progression of addiction, from initial drug use to withdrawal, negative emotional states, and then compulsive drug-seeking. For instance, withdrawal from drugs can cause a temporary drop in dopamine levels, followed by a lasting increase in the excitability of VTA dopamine neurons. This increased excitability contributes to greater dopamine release when exposed to drug-related cues during abstinence. Additionally, chronic drug use can reduce the sensitivity of certain dopamine receptors in the striatum, which is linked to various addictive disorders. All these drugs also cause a persistent increase in a specific protein (ΔFosB) in key brain regions, which is linked to drug-seeking during abstinence.

While single drug use causes many specific changes, polydrug use can lead to different or amplified disruptions in brain circuits. For example, repeated co-administration of psychostimulants and nicotine can enhance certain brain changes related to learning and reward, effects not seen with sequential use. Long-term opioid use tends to decrease the branching of neurons in some areas, unlike psychostimulants and nicotine which can increase it. Cannabis use also causes long-lasting changes in brain structure and signaling. Alcohol, when combined with nicotine, can lead to increased cell death, but surprisingly, withdrawal from both drugs together might result in less severe neurodegeneration than withdrawing from either drug alone. These unique patterns of change underscore how polysubstance use can alter brain dynamics differently than single drug use.

Conclusion

Improving how research models in laboratories accurately reflect the complex history and dependence of polysubstance use in people is vital. While human imaging studies are limited because researchers cannot control past drug use, animal models offer a controlled way to study polydrug use. However, these models cannot fully capture the complex social and environmental factors that influence how people use multiple substances. When designing studies, whether with people or animals, it is important to consider factors such as the time of day drugs are taken, how close together drug uses occur, and the environments where drug administration takes place. The way a drug is taken, such as oral ingestion, injection, or inhalation, is also crucial to consider, as it can affect how polysubstance use develops and its severity.

It is also essential to account for the timing and patterns of drug use. Although many studies have focused on simultaneous drug combinations, people more often report using drugs sequentially, one after another, which can lead to unique brain adaptations. Understanding differences between sexes in polysubstance use frequency, patterns, drug discrimination, and brain changes is also necessary to fully grasp how polydrug use impacts people. Furthermore, while powerful models based on economic principles allow for comparisons across drug classes, these experiments must be designed carefully to accurately value drugs and interpret data. Since these factors, along with socioeconomic background, social environment, and access to drugs, all affect how often and how susceptible someone is to polysubstance use, including these influences in laboratory models is crucial for research to be relevant to real-world situations.

New laboratory models are constantly being developed to study different ways drugs are taken and to examine relapse under conditions similar to those in a clinical setting. For instance, new models allow animals to voluntarily self-administer vaporized alcohol, cannabis, and nicotine. Additionally, various models for inducing drug abstinence are being introduced, such as linking drug use with negative consequences or offering a choice between a drug and an alternative reward like food or social interaction. These new approaches are highly valuable for translating research findings to real-world applications and could help understand the brain basis of therapies where abstinence is rewarded.

Addiction is a complex disease with many variable factors contributing to its beginning, ongoing use, and relapse. Given how common polysubstance use is among people who use drugs, it is critical to include this variability in both human studies and animal models. This will help determine if polysubstance use makes substance use disorders worse, if increased disorder severity leads to polysubstance use, or if there is a back-and-forth relationship between the two. Though challenging, understanding the behavioral, genetic, and environmental contributions to polysubstance use and addiction, as well as the mechanisms behind addiction severity and relapse, will assist in developing effective treatment and policy strategies to combat this ongoing public health crisis.

Introduction

Drug addiction is a complex problem. It involves repeated cycles of using drugs, stopping for a time, strongly wanting drugs, and then starting to use them again. This problem affects many people. It is thought that about 35 million people worldwide and 19.3 million people in the United States currently have a drug use problem. Also, studies show that about 10 out of 100 people will have a drug use problem at some point in their lives. Drug addiction is a big public health issue in the United States, costing about $740 billion each year. This money goes to treatment, lost work, healthcare, and dealing with crime. These numbers are likely to grow as more people use illegal drugs. About 250 million people around the world reported using illegal drugs in the past year. In the United States, over 17 million people aged 12 and older start using drugs each year. Opioid use is especially increasing, with 53 million people worldwide and about 11 million people in the United States reporting opioid misuse in the past year. This is very worrying because the number of deaths in the United States due to opioids went up 6 times from 1999 to 2017. About 130 Americans die from opioid use every day.

Most studies on drug use problems have looked at one drug at a time. Often, people who use many drugs are not included in these studies. But it is important to know that many drug users take more than one substance. For example, 30 to 80 out of 100 heroin users also use cocaine. Also, deaths involving both cocaine and opioids in the United States more than doubled between 2010 and 2015. A person is considered a polysubstance user if they use more than one drug. This includes using different drugs at different times or using several drugs at the same time. Focusing only on one drug in studies might mean missing how drugs affect each other. It can also make research less useful in the real world and make treatments for drug use problems less effective. In fact, using many drugs has often led to worse treatment results. People are less likely to stay in treatment, more likely to start using drugs again, and have a much higher chance of dying compared to those who use only one drug. This paper will combine what is known about how single drugs affect the body and mind with the latest research on using many drugs. It will note when possible if people use drugs at the same time, at different times, or both. First, there will be an overview of common drug use trends and how using many drugs affects the seriousness of drug use problems. Then, findings from animal studies and how they relate to real-world drug use will be discussed. This will include things to think about when planning studies on using many drugs. Next, the paper will explain how different drugs affect the brain, especially parts linked to addiction. This will help understand how using many drugs might change the brain and how addiction works. Care will be taken to discuss differences in brain changes when one drug is used versus many drugs, highlighting the most common drug combinations. For clarity, and to avoid repeating information, sections are arranged by the main drug used. The effects of adding other drugs to a main drug will also be considered. Lastly, suggestions will be made on how to study drug problems involving many substances.

Public Health Trends in Substance Use

Drug addiction is common and can cause death. About 585,000 deaths worldwide each year are related to drug use. While drug addiction and its effects are often thought of in terms of one drug, drug misuse often involves many substances. In fact, people who depend on drugs often report using about 3.5 different substances, sometimes at the same time and sometimes one after another. Also, people who are in treatment for drug problems often have problems with many drugs. Even though the mixes of drugs people use can change, the main drugs people depend on are usually alcohol, opioids, amphetamines, and methamphetamines. Cannabis and cocaine are more often used as second or third drugs. The high number of people using many drugs is a concern because it can make drug use problems more severe and make treatment less successful. For example, people who use many drugs often have more unmet health needs, take more risks, are involved in more violence, and have a higher chance of overdose and death compared to those who use only one drug. This section will talk about the public health effects of using single drugs and using many drugs, especially when other drugs are combined. This overview will try to cover common ways people use many drugs, including the most common drug mixes and patterns of use. However, because there is limited data on specific drug use patterns, types of classification, and different ways that drug mixes are defined across studies, it may not cover all combinations, past uses, and patterns of use.

Psychostimulants

Psychostimulants are the second most used type of drugs. Around the world, 18 million people currently use cocaine and 29 million use prescription stimulants. The number of people using psychostimulants worldwide has stayed about the same from 1990 to 2017. About 7.38 million people meet the criteria for an amphetamine use problem, and 5.02 million meet the criteria for a cocaine use problem. However, the number of overdose deaths involving psychostimulants has continued to go up, especially in the United States. Cocaine overdose deaths increased 2.6 times, and methamphetamine overdose deaths increased 3.6 times from 2000 to 2017.

It is important to note that most people who use cocaine and amphetamines also use many other drugs. One study found that 74 out of 100 cocaine users and 80 out of 100 amphetamine users had used many drugs in the past. Specifically, cocaine use and developing a cocaine use problem are linked to using heroin, cannabis, tobacco, and alcohol at the same time. Similarly, amphetamine users often use alcohol, tobacco, and cannabis. Other amphetamine users are more likely to also use heroin and other opioids. Across different groups, there is less chance of cocaine being used with amphetamine compared to other drugs used with amphetamine. Using many drugs is common among stimulant users, with people taking drugs both at the same time and one after another. For example, psychostimulants and opioids are used at the same time in "speedball" (cocaine and opioids) and "bombita" (methamphetamine and opioids). Using psychostimulants and opioids one after another is also common. This includes using cocaine or amphetamine to avoid the body's bad feelings when stopping opioids, and using opioids to calm down after using cocaine. Also, there is a higher chance of using methamphetamine and alcohol on the same day.

While these studies did not specify the exact patterns of using many drugs, one review of reports on using cocaine at different times versus at the same time found that 24 to 98 out of 100 people used cocaine and alcohol at the same time, and 12 to 76 out of 100 people used cocaine and cannabis at the same time. The numbers for using drugs at different times were 37 to 96 out of 100 for cocaine and alcohol, 43 to 94 out of 100 for cocaine and cannabis, 70 to 80 out of 100 for cocaine and nicotine, and 85 to 95 out of 100 for amphetamine and nicotine. The wide range in these numbers shows how hard it is to figure out drug use patterns. These patterns can change based on different groups of people, study times, study setups, and how "at the same time" or "different times" are defined.

Using many drugs that include psychostimulants creates big public health risks. For example, one study showed that amphetamine users were 21 times more likely to also have a cannabis use problem and 7 times more likely to have used cocaine in the past year, compared to those who had never used amphetamines. Also, almost one-third of overdose deaths involved both psychostimulants and opioids, like heroin and fentanyl. The dangers of using psychostimulants with other substances also include other drugs. For instance, combining cocaine and cannabis led to higher death rates in emergency room visits, meaning this combination is more dangerous. Also, combining cocaine and alcohol increases the risk of heart problems compared to using either drug alone.

Nicotine

Even though tobacco use has gone down since the early 2000s, nicotine is still one of the most commonly used drugs. In the United States, 58.8 million people aged 12 or older reported using nicotine in the past month. Nicotine is also often used with many other substances. In the past month, 17 out of 100 nicotine users also used cannabis, 4.7 out of 100 used opioids, 2.6 out of 100 used cocaine, and 1.4 out of 100 used psychostimulants. In contrast, people who do not smoke had much lower rates of past-month drug use. This difference is important because people with a nicotine use problem are 3 to 4 times more likely to have a second drug use problem. Also, past-year tobacco use was strongly linked to opioid use problems, as well as problems with cannabis and alcohol, and cocaine use in groups of patients getting primary care. How much tobacco a person uses (how often and how many cigarettes) has also been strongly linked to when they start using heroin and cocaine. Historically, nicotine was mainly used by smoking tobacco cigarettes. However, new technology has led to electronic cigarettes (e-cigarettes), which are made to deliver nicotine without toxins. The way e-cigarettes are advertised as a safer choice than regular tobacco cigarettes is a concern, as it has led to more nicotine use and a return of the potential for nicotine addiction. For example, e-cigarette use among middle and high school students went up from 2012 to 2016. A sharp rise in use was seen among young adults (18–24 years old) around 2013 to 2014, when e-cigarette products became available. Even though they deliver lower doses of nicotine, the safety of commercial e-cigarettes is debated. This is because people may take more puffs or use high voltage settings, which can create cancer-causing substances. The possible danger of use is made worse by the different amounts of nicotine in e-cigarettes from different makers. How vaping uniquely affects developing nicotine dependence and how this specifically adds to problems with using many drugs is still largely unknown and needs more study in the coming years.

Opioids

The number of people misusing opioids (using them outside of what is prescribed) has gone up sharply in recent years. About 53 million adults (1.1 out of 100 people worldwide) reported using an opioid for non-medical reasons in the past year. In the United States alone, 11 million people reported misusing opioids in the past year in 2016. However, this number might be lower than the actual amount because it does not include people who are homeless or in jail, who have much higher rates of opioid use. Also, the rate of people using heroin for the first time increased at the same time as non-medical use of prescription opioids from 2002 to 2011. This suggests that people who misused prescription opioids in the past year are 19 times more likely to start using heroin than those without such a history. Studies have looked at how people misuse many drugs among heroin and prescription opioid users. They found higher rates of opioid use in people who also use cocaine (more than 33 out of 100) or methamphetamine (more than 20 out of 100). However, there was less main opioid use in those who also used alcohol or cannabis. Also, using methamphetamine for the first time is more common after using opioids in the past month. Of those starting treatment for heroin use, it was found that 91 out of 100 people reported having used cocaine at some point in their lives. Additionally, a study in the United Kingdom found that 54 out of 100 opioid users in treatment between 2017 and 2018 also had a crack cocaine use problem. Regarding how people use multiple drugs, using heroin at the same time as alcohol or cannabis is more common than using it with psychostimulants. A pattern of using opioids and psychostimulants one after another is preferred.

In the United States, opioid use is a major public health crisis. It was responsible for more than 1.6 million years of life lost from 2001 to 2016. Also, opioid overdose deaths are currently the main cause of accidental death among adults in the United States, with 68 out of 100 of all drug overdose deaths involving an opioid. Since almost 80 out of 100 fatal opioid overdoses also involved another substance, it seems there is a greater risk of death when opioids are used with other opioids or other drugs. Specifically, of these deaths, 78 out of 100 involved another opioid, 21.6 out of 100 involved cocaine, 11.1 out of 100 involved alcohol, and 5.4 out of 100 involved a psychostimulant other than cocaine. Furthermore, opioid-related emergency room visits also involved tobacco (51.1 out of 100), cocaine (36.9 out of 100), other stimulants (22.6 out of 100), cannabis (25.1 out of 100), or alcohol (16.9 out of 100). Many people using three or more of these substances have also been reported for opioid-related emergency room visits. The chance of these visits has been linked to how severe other drug use problems are. All these reports suggest that using opioids with other substances can make the harmful effects of opioid use worse. Besides the risk of overdose, opioid users have very high rates of starting to use again. About 59 out of 100 people start using again in the first week, and 80 out of 100 start again in the first month of not using. Past use of other substances, including how much cocaine was used, increases the chance of starting again. Methamphetamine use has also been increasing among those seeking treatment for opioid use. Recent reports show that methamphetamine use is linked to stopping buprenorphine treatment in people with an opioid use problem. So, a better understanding of how using many drugs affects opioid use is key for better emergency responses and long-term treatment outcomes.

Cannabinoids

It is believed that 188 million people aged 12 or older use cannabis worldwide, including 43.5 million people in the United States. Starting in 2012 with Washington and Colorado, 11 states and the District of Columbia have made recreational cannabis legal. This means it is legally available to about 328 million people. The number of cannabis users in the United States has grown as it has slowly become less illegal and more legal. The rate went from 4.1 out of 100 people in 2002, to 9.9 out of 100 in 2007, to 15.9 out of 100 in 2018. People also use cannabis often, with reports that 40 out of 100 people use it daily or almost daily.

Cannabis users often report using tobacco, alcohol, and/or amphetamines in the past month. One of the most common combinations is using alcohol and cannabis at the same time, along with using alcohol, cocaine, and cannabis at the same time. The effect of using cannabis at the same time is notable. This pattern of use is linked to drinking more alcoholic drinks per day, suggesting that using cannabis at the same time helps alcohol use. Using many drugs is especially common in younger people. Among young cannabis users, 27.5 out of 100 reported using other drugs within the same year they started using cannabis, and almost 67 out of 100 use two or more other drugs. Cannabis is often used during treatment for other drug use problems, and this has been linked to less effective treatment. For example, cannabis use has been found to lead to shorter times of not drinking alcohol, as well as a greater chance of starting cocaine use again. Also, using many drugs among cannabis users has been linked to less upward social and economic movement, money problems, and relationship difficulties. It is also linked to more severe mood problems, trouble making decisions, social problems, and self-harm. While this information suggests that the negative effects of drug use are made worse by using cannabis at the same time, it is worth noting that treatment results can vary in studies looking at using many drugs among cannabis users. For example, some studies suggest that using many drugs has a specific impact that depends on the dose. There are no combined effects of cannabis and alcohol at low doses of either drug, and no link between cannabis use and heroin use again.

Alcohol

Alcohol is one of the most commonly used drugs. Up to 290 million people worldwide are diagnosed with an alcohol use problem, including 15 million people in the United States. Alcohol is often used with other substances. Reports show that 5.6 out of 100 adults in the United States have used both alcohol and another illegal drug in the past year, and 1.1 out of 100 have met the requirements for both an alcohol use problem and another drug use problem. The most commonly reported substance used with alcohol is cannabis (10 out of 100). Less common combinations are with opioids (2.4 out of 100), cocaine (2.5 out of 100), and amphetamine (1.2 out of 100). While using alcohol and cannabis or alcohol and prescription opioids at the same time is most common, using alcohol and cocaine at the same time is also seen.

Using many drugs increases the risk of developing an alcohol use problem, especially in young adults, men, and Native Americans/Alaska Natives. Using many drugs that include alcohol is linked to other problems, such as more mood problems, anxiety problems, more intense drinking, and more intense drug use and drug cravings. The bad effects of using alcohol with many drugs are also shown by data that says 21 out of 100 emergency room visits for patients aged 12-24 involved both alcohol and drugs. These visits were also more likely to need treatment for injuries and had higher rates of being admitted to the hospital. In addition, 17 out of 100 admissions to substance treatment were related to both alcohol and drug use. This makes up 45 out of 100 of primary alcohol admissions and 33 out of 100 of drug misuse admissions. The rate of hospital stays involving alcohol and many drugs has been going up, especially in young adults. Reports suggest a 76 out of 100 increase in hospital admissions between 1998 and 2008, compared to drug or alcohol overdoses alone. While the people in these surveys who used many drugs were mainly white and male, recent trends show an increase in emergency room visits related to alcohol and drug combinations in females. This suggests a change in which groups of people are using alcohol with other drugs.

Behavioral Models of Addiction in Polydrug Studies

Behavioral models of drug addiction are used to study how the brain works in the development, staying power, and return to drug use. The most common models are those that look at increased movement when using a drug, how much an animal prefers a place where they got a drug, and how much an animal works to give itself a drug. Experiments using these models change in many ways, such as how drugs are given, how much drug is available, the place linked to drug use, and how drugs are taken. Here, we describe how these models have been used with mixes of drugs and how this work has helped us understand using many drugs and addiction.

Early studies on using many drugs mainly used models where the drug was given without the animal doing anything to get it. These models looked at how one drug affected the preference for a place or how active an animal was with another drug. More recently, studies have been looking at how a past history of self-administering drugs affects future drug choices or wanting drugs again after stopping. Animal models that involve giving multiple drugs at the same time, such as alcohol and nicotine or cocaine and heroin, have also been used often. In human studies, both using many drugs at the same time and one after another are looked at by checking feelings after taking drugs, how much a person wants drugs when they see clues, and body responses like blood pressure and heart rate.

More recently, studies are comparing using one drug versus many drugs to see how a history of drug use changes the brain at a tiny level. Also, methods based on how people make choices about goods can figure out the most preferred amount of drug taken when it costs nothing. They can also see how much effort an animal will put in to get that preferred amount before it stops trying. These methods are very useful because they can use these measures to create ways to compare the value and cost of different drug and non-drug rewards when many drugs are used. This has been done in rats, monkeys, and humans. In particular, these studies help examine how rewarding different doses and types of drugs are, and how the value of a drug changes after being exposed to another drug. Directly measuring the value given to these drugs across different histories of using many drugs and different drug doses is very helpful for seeing how using many drugs affects the value of the reward. Also, these measures can be used to compare how people who use many drugs value drug rewards under different experiment conditions (for example, different starting doses of one drug, environment, or pattern of drug use). Finally, human studies are using surveys or controlled lab settings to look at the behavior effects of having used many drugs. Specifically, these studies use money choice tasks that compare the value given to drugs at different doses, how the value given to one drug changes when the price of another drug changes, or how the value of one drug changes when the quality of another available drug is thought to change. Also, tests that make animals work harder for a single drug or drug mixes have been done to study motivation, and tests that look at how quickly people discount future money and drug rewards have been done to study decision-making. These studies allow for comparing the perceived value across multiple drugs in people who have used single drugs or many drugs.

Effects of Polydrug Use on Addictive Behaviors

Because unique brain changes can happen when a person uses multiple drugs, and because using many drugs is so common, there is a strong need to create ways to study many drugs that are useful for understanding real-world problems. These ways of studying are key for fully understanding the behavior changes and addiction-like problems that happen after using many drugs. However, because there are so many possible drug combinations and different ways studies are done, the results and ideas about how a history of using many drugs affects addiction behaviors are currently mixed. Still, some general trends in drug use, drug preference, and drug-seeking have been shown in common drug combinations.

Psychostimulants

Some of the most studied combinations of many drugs include using cocaine with other drugs. However, unlike human reports, animal studies have not shown an increase in how severe addiction is. This suggests a need for models that are more like real-world human experience and can show the increased severity seen in people who use many drugs. Specifically, animal studies of using cocaine and alcohol or cocaine and heroin one after another have not found differences in how much drug is taken or how likely a person is to start seeking drugs again, whether they used one drug or many drugs. These effects were seen even though the studies used different ways of giving drugs, different drug amounts, and different animals, including rats and monkeys. In line with this, using alcohol from time to time has not been shown to affect cocaine self-administration or how rewarding cocaine is, as measured in monkeys. Also, using alcohol from time to time has not been shown to affect tests of how motivated an animal is for cocaine, or the long-term memory of preferring cocaine in places linked to the drug in rats. In contrast, alcohol use during the teenage years has been shown to have lasting effects on cocaine self-administration and reward. This suggests that young people are especially likely to be affected by using many drugs. For example, alcohol use during teenage years increases motivation for cocaine, makes it easier to develop a preference for places linked to cocaine in both mice and rats, and makes animals less likely to dislike the taste of cocaine. Also, using heroin and psychostimulants at the same time increases the motivation to give oneself cocaine and methamphetamine. Both using morphine and methamphetamine at the same time and one after another have been shown to be rewarding. Giving an opioid beforehand also makes the body more sensitive to methamphetamine. These findings suggest that opioids can make psychostimulants more rewarding and increase motivation for them, especially when used at the same time.

Studies on using cocaine and nicotine together have largely reported that the effects of the two drugs add up or work together to make the effects stronger. Specifically, using cocaine and nicotine together increases how much drug is taken in monkeys and rats. It also makes movement more sensitive to the drug and helps animals develop a preference for places linked to the drug in mice. It also causes a stronger drug craving. Also, long-term nicotine use beforehand helps animals start giving themselves cocaine, increases motivation when they have to work harder for the drug, makes it harder to stop a drug-seeking habit, and increases the chance of starting amphetamine use again after stopping. On the other hand, a history of nicotine treatment beforehand reduces how much a person wants cocaine when the price goes up. It is important to note that the effects of nicotine and psychostimulant use with many drugs depend a lot on the amount of drug used. For example, giving a smaller dose of nicotine beforehand increases the motivation to take cocaine in rats, while a larger dose of nicotine has the opposite effect. Also, giving methamphetamine and nicotine at the same time causes animals to avoid a place that was linked to the drugs, while giving them one after another causes them to prefer that place. Together, these studies show that how nicotine affects psychostimulant motivation, intake, and reward is greatly influenced by how nicotine is given (for example, dose, way of taking it, pattern of use). This should be carefully thought about when comparing study results and creating animal models of using many drugs. It is worth noting that nicotine exposure during the teenage years has no effect on developing a cocaine preference later, disliking the taste of cocaine, self-administering cocaine, stopping cocaine-seeking, or starting cocaine-seeking again in adulthood. However, it has been shown to increase cocaine self-administration in teenage rats. This information suggests that, unlike alcohol, early exposure to nicotine does not lead to more addiction-like behavior with cocaine in animals.

In humans, using cocaine and cannabis at the same time creates feelings of being "stimulated" and "high" that last longer than either drug alone. Also, drug craving caused by seeing drug cues in people who use both cocaine and cannabis lasts longer than for those who only use cocaine. In contrast, a substance in cannabis called THC reduces the motivation to give oneself cocaine in rodents. While this suggests that cocaine's effects are controlled differently in humans and rodents, more work is needed to make sure animal models can show increased addiction severity. However, using cocaine and cannabis or cocaine and alcohol at the same time in humans does not create feelings that are different from using cocaine, cannabis, or alcohol alone. Similarly, giving THC beforehand in rodents has no effect on psychostimulant reward or self-administration, nor does it make animals more likely to develop a preference for places linked to amphetamine. This points to a unique effect of cocaine and cannabis on drug craving. Interestingly, a substance in cannabis called CBD has no effect on cocaine self-administration, motivation, or drug craving caused by cues. This suggests that the effects of cannabis on cocaine craving are likely due to THC, not CBD. However, CBD treatment has been found to reduce the motivation to give oneself methamphetamine and to reduce methamphetamine-caused cravings. The effects of using cannabinoids and psychostimulants together seem to depend on both the amount of drug used and the age range during use. For example, one-time THC use weakens how sensitive the body is to movement, but repeated THC use leads to the body getting used to the effects, increasing amphetamine-caused repetitive movements and activity. Also, THC exposure during the teenage years speeds up how quickly animals start giving themselves cocaine and increases intake of low doses of cocaine. This shows lasting changes in the brain's reward system after teenage THC use, similar to alcohol.

Nicotine

Not much work has looked at how using many drugs affects addiction behaviors caused by nicotine. However, it has been found that giving THC beforehand can increase nicotine use and make people less sensitive to its price in economic tests in rats. Heroin use has been found to increase cigarette smoking in people. Also, being exposed to alcohol beforehand or having access to both alcohol and nicotine at the same time reduces how much nicotine rodents give themselves. While having access to alcohol has no effect on working for nicotine when the drug is not available, a small dose of alcohol can cause animals to start seeking nicotine again. Finally, giving methamphetamine and nicotine to the whole body at the same time causes rats to avoid a place where they got the drugs. But giving amphetamine or morphine beforehand increases how rewarding nicotine is, as shown by lower electrical brain stimulation needed. These studies further show the need to consider use patterns and doses when understanding the effects of using many drugs.

Opioids

Similar to studies on psychostimulants and many drugs, using heroin and cocaine one after another has not been found to change how much heroin is self-administered or how likely a person is to start seeking heroin again. Although giving alcohol beforehand can prevent the long-term memory of preferring morphine in places linked to the drug, alcohol exposure during the teenage years makes it easier to develop a preference for places linked to morphine. This finding shows that the long-term effects of alcohol exposure in teenage years apply to many different types of drugs. Interestingly, giving morphine and THC together prevents the body from getting used to the pain-relieving effects that normally happen with long-term use of either drug alone. Also, the pain-relieving effects of THC and oxycodone together are stronger than oxycodone alone. Furthermore, giving either THC or both THC and CBD reduces the bad feelings of withdrawal without affecting the development of a morphine preference. This information suggests that cannabinoids might help control physical dependence on opioids without changing how rewarding they are. Supporting this, repeated THC use has no effect on how much effort animals put in during a test for heroin self-administration or on starting to seek heroin again. However, it does cause a small reduction in both heroin and oxycodone intake in regular self-administration sessions. The effects of using opioids and cannabis together, however, seem to depend on the dose. Giving THC to the whole body before heroin self-administration reduces how much animals work for large doses of heroin but has no effect on how much they work for lower doses in both monkeys and rats.

Cannabinoids

As with psychostimulants and opioids, giving nicotine with THC makes the effects of either drug alone stronger when measured in tests of movement, pain relief, and body temperature. Also, using THC and nicotine together makes the body's bad feelings of THC withdrawal worse. However, after repeated nicotine treatment and 2 weeks of not using drugs, taking nicotine again reduces THC-caused decreases in movement, increases in anxiety (when tested in a specific maze), and changes in social interaction. These findings suggest that nicotine makes the negative symptoms of THC worse when used at the same time or close in time. Although nicotine treatment beforehand makes the rewarding effects of very low doses of THC stronger, cocaine treatment beforehand makes THC-caused anxiety behaviors worse. Using cannabis and alcohol together is fairly common in humans, and people report drinking less alcohol when cannabis is available. This suggests that cannabinoids might play a role in how much alcohol is consumed. However, drug-caused thinking and body problems in humans, as seen in a driving simulation, were found to be more severe after using THC and alcohol compared to either drug alone. Therefore, conducting studies on mixes of THC or CBD with other drug types is needed to understand the different effects that result from these drugs.

Unfortunately, due to long-term limits on cannabis research in the United States and past problems with creating animal models of cannabis use, much less is known about the effects of cannabis compared to other drugs. The development of new ways to get animals to self-administer cannabis, including giving them THC in gelatin by mouth, giving them vaporized THC and CBD, and giving them THC and CBD directly into the veins, will help move forward animal studies of cannabis use problems. This will also help us better understand the effects of using many drugs that include cannabis.

Alcohol

Using alcohol and nicotine together causes mixed results in addiction behaviors. For example, being exposed to alcohol beforehand or having access to both alcohol and nicotine at the same time increases how much alcohol is self-administered, but not when nicotine is given before alcohol each day. Long-term nicotine treatment also increases alcohol preference, and this effect stays even after stopping nicotine. However, although having access to nicotine makes it harder to stop responding for alcohol, it has no effect on starting to seek drugs again. This is because rats respond similarly on levers linked to alcohol and nicotine after a small dose of nicotine. Still, another study found that small doses of alcohol, but not nicotine, could cause animals to start seeking alcohol again after they had given themselves both nicotine and alcohol. Studies have not systematically looked at how other drugs affect alcohol use and addiction.

Neurobiology of Addiction

The development and continuation of addiction behaviors come partly from unhealthy changes in brain circuits. These circuits are responsible for making choices, learning, motivation, and how rewards are processed. In particular, changes in a brain network called the cortico-basal ganglia-thalamic network are known to play a part in drug-taking and drug-seeking behaviors, as well as how long drug use problems last. This network has many connections and brings together sensory information and body signals to drive motivated behavior. A key part of this network is the striatum. It gets signals from other parts of the brain (like the front of the brain, the amygdala, the hippocampus, and the thalamus), and also signals from a midbrain area. These signals work together in the striatum to either start or stop behavior. If the signals are out of balance in the striatum, it can lead to addictive behaviors. It is too much for this paper to cover all the brain changes that happen with drug use. Instead, the focus will be on one small circuit within this larger network that is central to the immediate effects of addictive drugs and some of the lasting changes that happen after long-term drug use: the network involving the prefrontal cortex, nucleus accumbens, and ventral tegmental area. After describing how these brain parts are connected, we will talk about what goes wrong in this network after both short-term and long-term use of different kinds of drugs, highlighting how these effects are similar or different when many drugs are used together.

Prefrontal Cortex

The prefrontal cortex (PFC) is very involved in learning about rewards, making choices, and deciding how valuable something is. It is a very complex brain area, which makes it harder to understand its role in thinking and how problems with it lead to drug use and addiction. In general, the inner part of the PFC helps control motivation and the seeking of both natural rewards and drug rewards. It does this by sending signals to other brain areas, including the nucleus accumbens and the ventral tegmental area. It is important that the signals from the inner PFC to the nucleus accumbens are organized in a specific way. In contrast, another part of the PFC, the orbitofrontal cortex, sends more signals to other areas and is mainly involved in deciding how valuable things are and how likely they are to happen. When the PFC is not active enough, it can lead to strong drug cravings and drug-seeking, even when there are bad consequences, especially in later stages of addiction. Problems with its connections to other brain areas can make people more sensitive to drug cues and more motivated to take drugs. The PFC has six layers, each with its own unique connections and cell types. Most PFC neurons are large output cells, as well as several types of smaller helper cells. These large cells send signals within the cortex and to other brain areas like the striatum, midbrain, amygdala, hippocampus, and thalamus. These large cells can be further divided based on how they work and how they are connected. Recent studies have started to describe the physical features, electrical activity, and chemical makeup of each of these cell types, but how each of them controls behavior is still not well understood. Finally, the PFC contains many groups of helper cells that strongly control how the cortex sends signals by connecting to both the large output cells and other helper cells.

Nucleus Accumbens

The striatum is a complex brain structure mainly made of two types of mixed nerve cells that can control behavior. These cells can either promote behavior by acting as a "go" signal or stop behavior by acting as a "stop" signal. Drug use causes more dopamine to be released, which makes drugs seem more rewarding during the binge/intoxication phases of addiction. It also helps people learn to link drug use to certain places and makes them feel a strong desire for drugs when they see cues. The striatum has top and bottom parts, with further smaller parts based on their connections and what they do. The bottom part of the striatum gets dopamine signals from the VTA and other signals from the PFC, as well as from the thalamus, hippocampus, and amygdala. In general, the bottom striatum controls motivated behavior and learning about rewards. However, it has been thought that a rising loop between the bottom and top striatum helps in learning, where habits move from the bottom striatum to the top striatum, leading to compulsive drug-seeking. Importantly, while these nerve cells have historically been separated by their targets and how they respond to dopamine, some cells in the bottom striatum also send signals to other areas, and dopamine receptors are found together in some parts of the striatum. Besides these main nerve cells, the striatum also contains large, always-active helper cells and many types of other helper cells with unique electrical properties and chemical patterns. The main nerve cells also get signals from other areas. However, how important these signals are for local brain activity and the role of these helper cells in the addiction cycle is still unclear. Interestingly, each main nerve cell gets many signals from other cells, so controlling their activity through other signals seems necessary for combining information and for healthy brain changes. Keeping the normal activity within these main nerve cells is key for healthy control of behavior, and problems with the tiny circuits in the striatum contribute to the development and expression of addiction behaviors.

Ventral Tegmental Area

The ventral tegmental area (VTA) sends dopamine signals to areas in the brain like the cortex, striatum, and other deeper brain parts. This helps control the activity of the main brain network involved in decision-making and reward. Dopamine signals from the VTA to the nucleus accumbens control goal-directed behaviors and are strongly linked to the binge/intoxication phase of drug use problems. Recently, new types of VTA dopamine cells have been found. They have unique chemical profiles and mainly send signals to specific parts of the nucleus accumbens. These cells control learning about rewards or motivation. However, both groups of cells need to be active for a strong rewarding effect. VTA dopamine cells get many signals from the prefrontal cortex and several midbrain structures, as well as signals that stop their activity from various sources, including the nucleus accumbens and local VTA helper cells. VTA helper cells keep dopamine levels steady by stopping the activity of VTA dopamine cells. A brief stopping of these helper cells leads to a quick release of dopamine into the nucleus accumbens. VTA helper cells get signals from the prefrontal cortex and a number of deeper brain areas, as well as signals that stop their activity from throughout the brain, including the nucleus accumbens. The VTA also contains other types of cells that send signals to helper cells in the striatum, though what signals they receive and their importance for behavior are not known. It is important that the VTA has groups of dopamine cells that can also release other chemicals. This allows the VTA to control brain activity at many levels and over different time spans. Within the nucleus accumbens, the activity of VTA dopamine cells is controlled by dopamine receptors on VTA dopamine cell endings and by helper cells that use another chemical.

Mechanisms of Addiction: Acute Drug Effects

A unique feature of all drugs that can cause addiction is their ability to make the binge/intoxication phase of the addiction cycle stronger. They do this by causing a quick release of dopamine into the nucleus accumbens. However, the ways they do this are different for each drug. Psychostimulants stop dopamine from being reabsorbed into VTA dopamine nerve endings. Nicotine and alcohol directly activate VTA dopamine cells. Opioids and cannabinoids remove the stopping effect on VTA dopamine cells. The way different drugs interact, sometimes working together and sometimes against each other, makes studying using many drugs very complex. Not much is known about how specific combinations of many drugs affect the brain right away. However, recordings in rats that used cocaine and heroin one after another in the same session found that only about 20 out of 100 neurons in the prefrontal cortex and nucleus accumbens reacted similarly to both drugs. This shows that these drugs affect the brain network in different ways. So, looking at how different drugs work together or against each other can help us understand how specific combinations of many drugs might mess up brain network activity and lead to addiction behaviors.

Psychostimulants

All psychostimulants directly increase dopamine release in the striatum by changing how a specific protein (dopamine transporter) works, though they do this in different ways. Cocaine blocks the reabsorption of dopamine, while amphetamine reverses the action of the dopamine transporter and causes dopamine to be released from VTA dopamine nerve endings. Psychostimulants also quickly increase activity in the prefrontal cortex, nucleus accumbens, and VTA, showing wide increases in activity throughout the brain network after psychostimulant use. The combination of increased signals to the nucleus accumbens helps nerve cells there become more active and activates processes that depend on dopamine. For example, psychostimulants quickly increase the activation of a certain gene in striatal nerve cells. This gene creates proteins that are strongly linked to addiction problems, and increased activation in the nucleus accumbens is thought to lead to long-term problems with normal brain network activity. Using nicotine, alcohol, or heroin at the same time makes psychostimulant-caused dopamine release in the nucleus accumbens even stronger, though specific combinations do so in different ways. For example, using cocaine and nicotine at the same time activates VTA dopamine cells and stops dopamine reabsorption, leading to a greater amount of dopamine release into the nucleus accumbens than either drug alone. It is important to note that some combinations of many drugs that include psychostimulants have different and opposing effects on the brain circuit. Using alcohol with cocaine at the same time causes the liver to produce a substance that can affect dopamine reabsorption and receptors, making dopamine release stronger. However, alcohol also stops cocaine-caused activity in a part of the nucleus accumbens. Lastly, studies on many drugs with psychostimulants have found that drug-caused cell damage is worse compared to using psychostimulants alone. Specifically, using heroin with cocaine at the same time lowers cell activity, increases signals inside cells, and lowers the power of cell energy parts. Together, these effects lead to more cell death compared to either drug alone.

Nicotine

Unlike psychostimulants, nicotine increases dopamine release by activating specific receptors within the VTA. These receptors are channels that let charged particles pass through, and their activation causes cells to become more active and release more chemicals. The rewarding effects of nicotine are mainly due to its action on these receptors located on VTA dopamine cells and on signals coming from the prefrontal cortex. Acute nicotine exposure activates both the prefrontal cortex signals and VTA cells, causing a quick release of dopamine into the nucleus accumbens. It is important to note that nicotine-caused dopamine release is stopped if certain other receptors are blocked or activated, showing how much the activity of VTA dopamine cells is controlled by the tiny circuits within the VTA. Studies have not looked at how the immediate effects of nicotine are changed by using many drugs.

Opioids

While opioids differ in their origin, strength, and how they affect receptors, all opioids create their rewarding effects by activating specific opioid receptors. These receptors are found both on the cell body and on the nerve endings. But the main way opioids cause dopamine release is by stopping the activity of VTA helper cells. Opioid receptors are inhibitory receptors, and their activation reduces how active nerve cells are through four ways: (1) stopping certain cell signals, (2) activating certain channels, (3) stopping certain calcium channels, and (4) stopping the release of chemicals from nerve endings. Short-term exposure to opioids stops VTA helper cells, which then removes the stopping effect on VTA dopamine cells and causes a quick release of dopamine into the nucleus accumbens. Still, even though opioids help dopamine release into the striatum, whether this dopamine release is needed for opioid reward is still debated. Opioids activate VTA dopamine cells in living animals and increase dopamine in the nucleus accumbens. However, damage to the nucleus accumbens or blocking dopamine receptors throughout the body has no effect on opioid reward. Given that opioids work differently than psychostimulants and nicotine, it is not surprising that using these drugs together makes the immediate effects of opioids stronger. In fact, using opioids and psychostimulants at the same time causes an added increase in dopamine release in the nucleus accumbens and keeps dopamine levels high for over an hour. Similarly, using opioids and nicotine at the same time increases opioid-caused dopamine release in the nucleus accumbens and a part of the striatum. The body getting used to opioid pain relief has also been shown after being exposed to nicotine and cannabis beforehand. Also, long-term nicotine treatment reduces how much the body gets used to opioid pain relief, depending on the dose and how a specific receptor is affected. Interestingly, giving certain blockers or another drug beforehand prevents this tolerance, suggesting a complex chemical interaction between opioids and nicotine. It is important that getting used to opioid pain relief works differently for different combinations of many drugs. Nicotine and opioid tolerance is caused by opioid and nicotine receptors, while cannabinoid and opioid tolerance is caused by opioid and cannabinoid receptors.

Cannabinoids

Cannabis has two main chemicals that affect specific receptors: THC partly activates both CB1 and CB2 receptors, while CBD has a very low effect on these receptors and works in an unknown way. Interestingly, giving a range of CBD doses beforehand has no effect on THC self-administration, showing that each chemical works through different paths. CB1 receptors are found on nerve endings throughout the brain and spinal cord and are responsible for the mind-altering effects of cannabis, while CB2 receptors are mainly found on immune cells and are mainly responsible for the pain-relieving and anti-inflammatory effects of cannabis. Both types of receptors work by stopping cell signals (similar to opioids). Activating these receptors leads to stopping nerve activity (by activating certain channels and stopping certain calcium channels) and lowering certain cell signals. CB1 receptors are heavily found on nerve endings of VTA helper cells as well as nucleus accumbens nerve cells that target VTA dopamine cells. Activating CB1 receptors reduces stopping signals in VTA brain slices. This stopping of VTA helper cells by CB1 then allows VTA dopamine cells to become more active and causes dopamine to be released into the nucleus accumbens. These effects are blocked by giving a drug that affects opioids in the body or directly into the VTA. Similar to other drugs, the immediate effects of cannabinoids on the main brain network are wide and involve many chemical systems. For example, short-term THC use increases both dopamine and other signals in the nucleus accumbens and prefrontal cortex, but decreases stopping signals in the VTA and prefrontal cortex. Together, these changes in signals reduce stopping feedback within the main brain network and make behaviors easier to do. Drinking alcohol before cannabis use increases THC levels in the blood and increases feelings of "euphoria" in humans, showing that the two drugs work together to make effects stronger. Also, using cannabinoids with psychostimulants or opioids at the same time increases the activation of VTA dopamine cells. Using THC and nicotine at the same time increases the activation of a certain gene throughout the brain network.

Alcohol

Despite its widespread use, much less is known about how alcohol affects the body right away. Alcohol activates certain VTA dopamine cells and causes dopamine to be released into the nucleus accumbens, similar to other drugs that can cause addiction. However, specific receptors are known to play a central role in the effects of alcohol. For example, alcohol makes certain signals stronger in brain slices and cell cultures. Given that certain nerve cells in the nucleus accumbens only stop VTA helper cells through these signals, it is possible that alcohol helps dopamine release from VTA dopamine cells by stopping local VTA helper cells. Supporting this idea, alcohol stops VTA helper cells, and putting certain activating drugs into the VTA increases dopamine release depending on the dose. In addition to these signals, the immediate effects of alcohol depend on other signals within the main brain network. Alcohol increases the release of a certain chemical in the nucleus accumbens and VTA by activating specific receptors, suggesting that alcohol starts a loop that activates VTA dopamine cells. Using many drugs with alcohol creates stronger effects throughout the main brain network, likely because alcohol has unique chemical properties. For example, long-term treatment with nicotine beforehand increases immediate alcohol-caused dopamine release in the nucleus accumbens, and high levels of dopamine and its related chemicals stay for over an hour. Additionally, using alcohol and nicotine at the same time quickly increases the production of certain brain growth factors in the nucleus accumbens, along with increases in other signals in the VTA and prefrontal cortex. It is important to note that this wide activation of the main brain network is not seen after using either drug alone, showing a unique way alcohol and nicotine work together when used as many drugs.

Mechanisms of Addiction: Long-Term Alterations

Long-term use of psychostimulants, nicotine, opioids, cannabinoids, and alcohol leads to wide and different changes throughout the brain network, yet there are important changes that are shared across drugs. These long-term changes contribute to the shift from binge/intoxication phases to withdrawal and negative feelings, followed by strong cravings and compulsive drug-seeking. For example, stopping drug use suddenly causes a temporary drop in normal dopamine levels in the nucleus accumbens. This is followed by a lasting increase in how easily VTA dopamine cells can be activated, which adds to stronger dopamine release when triggered by cues during the time of not using drugs. On the other hand, stopping drug use also causes a lasting reduction in how much certain nerve cells in the nucleus accumbens can change and become active. It also reduces the number of certain dopamine receptors in the striatum. As described earlier, these striatal dopamine receptors are mainly found on cells that can act as a "stop" signal in the brain circuit, and fewer of these receptors have been widely linked to many addiction problems, including drug addiction and obesity. Finally, these drugs all cause a lasting increase in a certain protein in the brain cells of the cortex and nucleus accumbens. This has been linked to drug-seeking during the time of not using drugs. Importantly, while long-term use of any of these drugs causes many other temporary or lasting changes across the brain network, it is too much for this paper to list every change. Instead, the focus will be on changes in how the brain adapts, its shape, and its connections within the VTA, nucleus accumbens, and prefrontal cortex after using one drug or many drugs. The paper will highlight how the opposing and working-together effects of these drugs can differently disrupt brain network activity.

Psychostimulants

Psychostimulants cause lasting problems with the balance of a certain brain chemical and changes in the shape of nerve cells throughout the brain network. Repeated cocaine use makes it harder for certain signals to stop nerve cells in the prelimbic area of the brain, making them more active and increasing signals to the nucleus accumbens. Similarly, long-term cocaine use increases signals to the VTA and weakens the stopping effect on VTA dopamine cells, leading to easier activity of VTA dopamine cells. Additionally, psychostimulants create new connections on certain nerve cells and reduce the release of a brain chemical from signals from the prelimbic area to the nucleus accumbens, weakening the output from the striatum. Importantly, cocaine-silenced connections on nerve cells can become active again during withdrawal, and a low dose of psychostimulant can restore the release of a brain chemical into the nucleus accumbens, suggesting the body tries to return to balance. Finally, repeated exposure to psychostimulants increases the presence of a certain protein in brain cells of the PFC and nucleus accumbens. This leads to more branches on nerve cells that last for at least 30 days of not using drugs. Studies have not looked at how these effects of psychostimulants are changed by using many drugs.

Nicotine

Similar to psychostimulants, long-term nicotine use reduces a certain stopping signal in the mPFC and nucleus accumbens, lessening the stopping effect on the main brain network. Additionally, nicotine increases a certain brain signal to VTA dopamine cells and makes dopamine release in the nucleus accumbens more sensitive. Repeated nicotine use also leads to more nicotine receptors throughout the brain network, including the PFC and midbrain. On the other hand, long-term nicotine exposure weakens both certain brain signals and stopping signals to the nucleus accumbens during the time of not using drugs, possibly by making certain dopamine receptors in the nucleus accumbens more sensitive. Because the output of the nucleus accumbens is strongly controlled by local stopping circuits, increased activity of certain nerve cells would likely reduce the overall output of the nucleus accumbens. This is supported by the increase in certain receptors in the nucleus accumbens that can carry certain ions without the need for other receptor activation. Long-term nicotine use also causes a lasting increase in the branching of mPFC brain cells and nucleus accumbens nerve cells, similar to psychostimulants. Moreover, adolescent nicotine use increases pruning of connections, activation of certain brain cells, and signs of inflammation throughout the brain network through a certain receptor-controlled process. Repeated use of psychostimulants and nicotine together increases the strength of certain brain connections and a certain protein in the nucleus accumbens. Interestingly, these effects are not seen after using these drugs one after another, showing a unique chemical effect when nicotine and psychostimulants are used at the same time.

Opioids

Similar to psychostimulants and nicotine, long-term exposure to opioids strengthens a certain brain signal to VTA dopamine cells. They also weaken signals that stop VTA dopamine cells by reducing a certain type of stopping signal from certain nerve cells. This effect is largely driven by nerve cells in the nucleus accumbens, showing a lasting problem with nucleus accumbens output. However, unlike psychostimulants and nicotine, long-term exposure to opioids decreases the branching and number of connections on nucleus accumbens nerve cells and mPFC brain cells. Within the nucleus accumbens, long-term opioid exposure weakens certain brain signals to certain nerve cells in the nucleus accumbens shell and creates new silent connections on these nerve cells. Moreover, during withdrawal, opioid-created silent connections on nerve cells are removed, and the basic activity of these nerve cells is weakened. Given that nerve cells create a dense network of stopping each other within the nucleus accumbens, it is possible that opioid-caused problems in nerve cell signaling could disrupt the local tiny circuits of the nucleus accumbens and lead to abnormal brain network activity. Studies have not looked at how these effects of opioids are changed by using many drugs.

Cannabinoids

Repeated THC use weakens a certain brain signal from the mPFC to the nucleus accumbens but strengthens signals from other brain areas to the nucleus accumbens shell. This suggests a rewiring of how signals enter the nucleus accumbens after long-term cannabinoid exposure. It is important to note that repeated THC use also blocks certain long-term changes on VTA helper cells, showing a loss of local stopping effect on VTA dopamine cells. Additionally, THC use during teenage years followed by not using drugs reduces the basic activity of certain brain cells and weakens certain brain signals to the VTA, showing that cannabinoids can cause a state of low activity in a certain brain chemical. THC exposure also causes lasting changes in the shape of nerve cells throughout the brain network, increasing the length and branching of connections in the mPFC and nucleus accumbens shell and the number of connections in the nucleus accumbens shell. However, it reduces the size of VTA dopamine cells. Repeated use of THC and nicotine together increases the presence of a certain protein in the nucleus accumbens, and sudden withdrawal messes up certain brain signals to the nucleus accumbens and PFC.

Alcohol

Similar to opioids, long-term alcohol exposure decreases the branching and number of connections on nerve cells in the nucleus accumbens shell and mPFC. Withdrawal from alcohol reduces normal dopamine levels in the nucleus accumbens. However, alcohol withdrawal also causes distinct changes throughout the brain network, including reduced stopping signals in the nucleus accumbens and hippocampus, and increased activating signals in the nucleus accumbens and PFC. Notably, this increased activating signal is due to problems with reabsorbing the brain chemical, rather than increased release. In lab studies, using alcohol and nicotine together increases cell death. Interestingly, however, nerve cell damage caused by alcohol or nicotine withdrawal is less severe after using both drugs together. This shows a unique way alcohol and nicotine work together when used as many drugs.

Conclusion

It is very important to make animal study results more useful for understanding real-world situations and to link behavior measures in animal models to how drug use problems with many substances affect people. This is especially true because human brain imaging studies are limited in their ability to control what a person has used in the past. This makes animal models helpful for studying a history of using many drugs under controlled conditions. However, these models cannot fully capture the complex social and environmental factors that lead to unique ways people use multiple addictive substances. Still, when planning studies in people or animals, factors like time of day drugs are taken, how close in time they are taken, and preferred environments for taking drugs must be thought about for each drug type. For example, cocaine use is mostly preferred outside of home environments, while heroin use is more common in "home" settings in both humans and animals. Also, how a drug is taken (for example, by mouth, injection, or inhaling) is especially important to consider in studies, as it leads to unique patterns of using many drugs that can affect how drug use problems develop and how severe they become.

Considering the timing of drug use patterns is also essential. While many studies have focused on using drugs at the same time, using many drugs one after another is more often reported and causes unique brain changes after short-term and repeated drug exposure. Understanding sex differences in how often and how people use many drugs, how they can tell drugs apart, and brain changes is also necessary to fully understand how using many drugs affects people. Furthermore, while powerful models that look at choices can compare different drug types, these experiments must be designed by considering different amounts of drug taken and points where people don't care about the drug's value. This is needed to correctly model factors and understand the information. Since these factors, along with how social background, social surroundings, and ways to get drugs interact, affect how often and how likely a person is to use many drugs, including these influences in animal models is crucial for the findings to be useful in the real world.

It is important to appreciate that new animal models are always being developed to show different ways of taking drugs and to study relapse under real-world conditions. For example, recently created models allow voluntary control over taking vaporized alcohol, cannabis, and nicotine. Also, many models for causing animals to stop drug use are being introduced. For instance, animals choose to stop using drugs by linking drug-taking with bad outcomes like electric shocks, or by having physical barriers. They also stop by choosing between a drug and something else rewarding (like tasty food or social interaction). These ways of studying are notable because they are very useful in the real world and can help understand the brain basis of treatments where stopping drug use leads to a reward. Combining these models with models that look at choices could clarify how the value of drugs and other rewards changes when many drugs are used, which could improve treatment effectiveness. Addiction is a complex disease with many highly varied factors that lead to starting and continuing drug use, as well as starting again. Given how common it is for drug users to use many drugs, it is critical that we include this variety in human studies and animal models. This will help determine if using many drugs makes drug use problems worse, if more severe drug use problems lead to using many drugs, or if they affect each other both ways. While challenging, understanding how behavior, genes, and environment contribute to using many drugs and addiction, as well as the ways that addiction severity and relapse happen, will help in developing effective treatment and policy plans to fight this ongoing public health crisis.