Cocaine use disorder (CUD) remains a major public health concern and no FDA-approved pharmacotherapies are currently available. CUD is a highly heterogenous disorder, partially characterized by personal health risks, societal and legal implications, and high relapse rates (Degenhardt et al., 2013; Kampman, 2019; Volkow and Boyle, 2018). Craving is a robust predictor of relapse (Hunt et al., 1971; Paliwal et al., 2008; Sinha, 2011), and both stress and drug-associated cues significantly increase the likelihood of re-initiated cocaine use following abstinence (Preston et al., 2009; Preston et al., 2018; Sinha et al., 1999; Sinha et al., 2006). Neuroimaging studies demonstrate that cocaine cues activate brain regions involved in reward, memory, and stress regulation (Fox et al., 2006; Grant et al., 1996; Kilts et al., 2001; Kilts et al., 2004; Waldrop et al., 2010). Importantly, polysubstance use is the clinical norm: approximately 80% of individuals with CUD also use nicotine (Foltin et al., 2021), and nearly 90% co-use alcohol (van Amsterdam et al., 2024). Epidemiological data indicate that a substantial subset of individuals with CUD also concurrently use alcohol and nicotine, highlighting a clinically significant pattern of polysubstance use that remains largely unmodeled in preclinical research (Allen et al., 2024b; Cougle et al., 2016; Mantsch et al., 2016). This gap underscores the need for models that more accurately reflect the complexity of human substance use.

Cocaine and ethanol co-use result in the formation of cocaethylene, a metabolite with a longer half-life than cocaine and similar potency in inhibiting dopamine reuptake (Bradberry et al., 1993; Landry, 1992). Cocaethylene substituted for cocaine in drug discrimination studies (Katz et al., 1992; Woodward et al., 1991) and maintained responding under various schedules of reinforcement with equal potency to cocaine in rhesus monkeys (Allen et al., 2023; Jatlow et al., 1991). In humans, cocaethylene resulted in cocaine-like subjective effects despite being less potent than cocaine (Hart et al., 2000; Perez-Reyes et al., 1994). Although cocaethylene has been evaluated in self-administration models, its contribution to relapse-related behaviors has not been examined preclinically. Understanding how cocaethylene interacts with other substances such as nicotine may clarify mechanisms that heighten relapse vulnerability in polysubstance users. Nicotine is the most prevalent co-used substance with cocaine, and evidence suggests that it can potentiate cocaine’s reinforcing effects (Allen et al., 2024b; Weinberger et al., 2017). Given the high prevalence of nicotine co-use among individuals with CUD, defining its impact on relapse related behaviors is an important translational goal.

Reinstatement paradigms are widely used to model relapse because they allow systematic evaluation of stimuli that trigger drug seeking, including drug primes, cues, and stressors (Bossert et al., 2013; Nicolas et al., 2022; Shaham et al., 2003). Drug-induced reinstatement can model the extent to which cocaine-like subjective effects promote relapse. In preclinical models of cocaine-induced reinstatement, priming doses of cocaine reliably increase drug-seeking behavior, emphasizing the discriminative stimulus effects of cocaine (McFarland and Kalivas, 2001; Roberts et al., 2025; Schmidt et al., 2005). Both preclinical (Jing et al., 2022) and clinical studies (Childress et al., 1999) demonstrate that drug-associated cues reinstate cocaine seeking via mesocorticolimbic circuitry and activate limbic regions linked to craving and reward. Although reinstatement models provide valuable and translational tools for understanding CUD, important gaps remain in their ability to capture the complexity needed for improved clinical relevance.

Historically, reinstatement models have lacked social and biological complexity, yet emerging evidence shows that social and environmental factors can significantly alter reinstatement responding and relapse vulnerability (Rough et al., 2025b). Sex differences are well documented in both preclinical and clinical studies across multiple stages of the substance use cycle, including relapse (Dos Anjos Rosário et al., 2022; Morales-Silva et al., 2024; Nicolas et al., 2022). Findings from these studies reveal a complex and sometimes inconsistent pattern. Clinically, women often report greater stress-induced craving (Elman et al., 2001), although sex differences in cue- and drug-induced craving remain equivocal (Fox et al., 2006; Moran-Santa Maria et al., 2014). Preclinical work provides mechanistic insight into these effects. Estradiol enhanced cocaine reinforcement and reinstatement, whereas progesterone and its metabolites appeared to exert protective effects (Becker et al., 2017; Larson et al., 2007; Nicolas et al., 2022; Swalve et al., 2016b). In short, reinstatement varies across the estrous cycle, and estradiol’s rapid modulation of striatal dopamine release underscores the role of ovarian hormones in relapse susceptibility (Kokane and Perrotti, 2020). Beyond sex differences, social factors also shape drug seeking. In socially housed nonhuman primates, placement within the social hierarchy reflects a continuum in which subordination corresponds to chronic social stress and dominance is associated with social and environmental enrichment (Czoty et al., 2009; Kromrey et al., 2016; Nader et al., 2012a). This continuum has strong translational value and reliably influences cocaine self-administration (Allen and Nader, 2025; Czoty et al., 2017). These observations highlight that sex and social hierarchy may influence cocaine reinforcement and point to the importance of incorporating these factors into translational models that more accurately reflect clinical relapse risk.

Thus, the goal of the present study was to investigate how cocaine and cocaethylene priming doses influenced drug-induced reinstatement in socially housed male and female cynomolgus macaques. To address additional gaps in the polysubstance use and reinstatement literature, this study also examined whether co-administration of nicotine altered reinstatement outcomes and whether sex or social rank moderated these effects. By incorporating polysubstance use, the social environment, and individual differences, this model provides a more translational framework for understanding relapse and identifying variables that moderate cocaine- and cocaethylene-induced drug seeking.

The present study examined how sex, social rank, and nicotine exposure interact to influence cocaine- and cocaethylene-induced reinstatement under a food vs. drug choice procedure in socially housed cynomolgus monkeys. Across both sexes and social ranks, pretreatments with cocaine and cocaethylene reliably increased drug-associated choice relative to saline baselines, with dominant monkeys showing greater reinstatement than subordinates for cocaethylene and marginally greater reinstatement for cocaine. Cocaine was more potent than cocaethylene, producing significant reinstatement at lower doses. When nicotine was added, its effects were influenced by sex: nicotine co-administration increased cocaine-induced reinstatement in females, regardless of social rank. However, it did not impact cocaine-induced reinstatement in males and often reduced the effects of cocaine-induced reinstatement at the individual-subject level. Nicotine had more variable and generally weaker effects on cocaethylene-induced reinstatement, with only a trend toward increasing cocaethylene-induced reinstatement in females. Finally, nicotine alone reinstated drug-associated responding only in females. Together, these findings suggest that common polysubstance combinations (cocaine, alcohol/cocaethylene, and nicotine), the social environment, and biological sex interact to determine relapse-related behavior. Furthermore, they highlight heightened sensitivity to the effects of nicotine in females.

This study extended the comparison of cocaine and cocaethylene to a reinstatement model, where cocaine was more potent than cocaethylene and cocaine tended to result in greater reinstatement. Earlier research comparing the behavioral potency of cocaine and cocaethylene noted differences depending on the experimental paradigm. For example, Katz et al. (1992) found that cocaine was 4-times more potent than cocaethylene in increasing operant responding by mice and 5-times more potent than cocaethylene in cocaine discrimination studies in rats and squirrel monkeys. In contrast, Allen et al. (2023) reported that cocaine and cocaethylene were equipotent in monkeys responding under a progressive-ratio and concurrent choice contingencies. It has been speculated that reinstatement models are similar to studies involving drug discrimination (Katz and Higgins, 2003). The present study supports this premise by demonstrating that cocaine is also more potent than cocaethylene in a model of drug-induced reinstatement.

In addition to potency differences between cocaine and cocaethylene, these experiments demonstrated that social rank significantly influenced drug-induced reinstatement when examining the dose of drug that had the greatest effect on % drug-associated choice. In this study, dominant monkeys exhibited significantly higher % drug-associated choice following cocaethylene pretreatments and marginally higher % drug-associated choice following cocaine exposure when compared to their subordinate counterparts. This social rank-dependent difference underscores the importance of social context in modulating drug-seeking behavior. Previous research has demonstrated that chronic social subordination in monkeys is a well-validated model of social stress (Kaplan et al., 1986; Kaplan et al., 1999) and can increase susceptibility to a host of diseases including atherosclerosis (Kaplan et al., 1982), upper respiratory infection (Cohen et al., 1997), and reproductive dysfunction (Kaplan et al., 2010). Subordinate monkeys also exhibit heavier adrenal glands when compared to dominants (Shively and Kaplan, 1984), consistent with the view that social rank alters hypothalamic–pituitary–adrenal (HPA) axis physiology, a key regulator of stress responses (Henry and Stephens, 1977).

Prior work in our laboratory has shown that after long-term stable social housing when compared to dominant monkeys, subordinate males (Morgan et al., 2002) and females were more vulnerable to cocaine reinforcement on an acquisition procedure (Johnson et al., 2024; Nader et al., 2012b). However, a similar relationship between social rank and sensitivity to cocaine was not seen in this study. Instead, this study found that dominant monkeys had greater drug-associated choice with both cocaine and cocaethylene pretreatments. Previous work has shown that while subordinate monkeys are more vulnerable to cocaine initially, after long-term cocaine self-administration, response rates and cocaine intakes are no longer different between dominant and subordinate male monkeys who self-administer cocaine alone (Czoty et al., 2004). Hence, it is possible that the factors influencing initial vulnerability to cocaine reinforcement are not the same as those that influence other facets of drug use such as the reinstatement of drug-seeking. Subordination and chronic social stress have been linked to reduced basal dopaminergic tone and lower striatal dopamine D2/D3 receptor availability (Morgan et al., 2002; Nader et al., 2012b), which are neuroadaptations that promote initial vulnerability to cocaine acquisition (Morgan et al., 2002). On the other hand, social dominance has been associated with increased D2/D3 availability and lower vulnerability to cocaine reinforcement (Johnson et al., 2024; Morgan et al., 2002). However, under a reinstatement paradigm, higher D2/D3 receptor availability in dominant monkeys could plausibly enhance the salience of the interoceptive cues of cocaine and of drug-paired stimuli, thereby increasing the efficacy of pharmacological primes to reactivate drug-seeking in dominant animals (Nader and Czoty, 2005; Wolff and Saunders, 2023). In short, greater drug-induced reinstatement in dominant monkeys may suggest context-dependent differences in stress-mediated initiation (acquisition) and cue- or interoception-mediated relapse (reinstatement). The behavioral, neurochemical and neuropharmacological mechanisms underlying these social rank-related differences in drug-induced reinstatement deserve further investigation.

While social rank influenced cocaine- and cocaethylene-induced reinstatement, the addition of nicotine shifted the primary moderating factor to sex. Only female monkeys showed increased % choice of the drug-associated lever when nicotine was added to cocaine, and a similar, although non-significant, pattern emerged for cocaethylene. In follow-up analyses that included all nicotine doses tested, statistical analyses confirmed that the nicotine-related effects observed in the primary analyses were not an artifact of selecting the dose with the largest behavioral impact. Specifically, when comparing monkeys with the same nicotine dose, nicotine’s modulation of cocaine- and cocaethylene-induced reinstatement remained sex-dependent. Importantly, complete nicotine dose-response curves were determined with cocaine and cocaethylene and for some monkeys, lower nicotine doses resulted in greater reinstatement than higher nicotine doses. While the reason for these differences in sensitivity are not apparent, these data highlight important individual differences in the behavioral effects of nicotine. Furthermore, nicotine alone produced drug-induced reinstatement only in females, indicating that females may be more sensitive to the effects of nicotine in general and to the effects of nicotine on cocaine- and cocaethylene-induced reinstatement. These observations align with prior clinical and preclinical work showing sex differences in nicotine sensitivity. For instance, one clinical study found that women were more sensitive to the subjective effects of intravenous nicotine and reported enhanced ratings of “drug strength” and “head rush” when compared to men (Sofuoglu and Mooney, 2009). In rodents, females acquired nicotine reinforcement more rapidly under an FR schedule (Donny et al., 2000) and had higher peak breakpoints under a PR schedule when compared to males (Li et al., 2014).

However, studies using reinstatement procedures in rodents have yielded mixed results, with some reporting no sex differences (Feltenstein et al., 2012), others finding greater nicotine-induced reinstatement in males (Swalve et al., 2016a) and still others reporting enhanced nicotine-induced reinstatement in females (Kilty et al., 2024). Our data add to this literature by showing that nicotine resulted in drug-induced reinstatement only in females. Furthermore, the premise that females are more sensitive to the effects of nicotine when combined with cocaine is also consistent with prior work from our laboratory, which showed that the addition of nicotine to cocaine potentiated cocaine choice to a greater extent in females compared to males, regardless of social rank (Allen et al., 2024b). Future studies will be needed to determine how nicotine and cocaethylene interact under self-administration procedures and how those interactions relate to the current reinstatement findings.

In the present study, only females showed significant reinstatement with nicotine alone, and one possibility to explain this finding is that the increase in cocaine- and cocaethylene-induced reinstatement with nicotine co-administration reflects an additive effect of the two drugs. However, this interpretation has an important caveat: in several monkeys, nicotine alone did not increase drug-associated responding, yet it still significantly increased % drug-associated choice when combined with cocaine or cocaethylene. This pattern suggests that nicotine’s potentiation of drug-induced reinstatement is more complex than simple additivity. Because many females exhibited irregular menstrual cycles during this study, we could not assess its impact on the effect of nicotine on drug-induced reinstatement. Therefore, it remains possible that nicotine’s effects varied across cycle phase. The menstrual cycle irregularities noted within our cohort of female macaques could be due to a variety of factors. It is well established in preclinical and clinical studies that stressors, including social stress, can alter menstrual cycles in women (Barsom et al., 2004) and female animals (Bethea et al., 2008). Along with social stressors, chronic cocaine use has also been shown to alter menstrual cycles in preclinical and clinical studies (Mello and Mendelson, 1997; Mello et al., 1997). In fact, one study found that cocaine self-administration in rhesus monkeys was related to low progesterone levels and luteal phase dysfunction. Moreover, it resulted in amenorrhea which persisted even when cocaine self-administration was discontinued (Mello et al., 1997). In one study from our laboratory, we examined whether acquisition of cocaine self-administration under a fixed-ratio schedule was influenced by menstrual cycle and did not find differences (Nader et al., 2012b). It is certainly the case that the role of menstrual cycle on measures of the behavioral and reinforcing effects of cocaine will be influenced by experimental parameters, social factors and drug history.

Although the mechanisms behind the irregular menstrual cycles are not fully understood, one study (Mello et al., 1997) hypothesized that it could be due to changes in baseline levels of anterior pituitary hormones that regulate menstrual cycles such as luteinizing hormone (LH), follicle-stimulating hormone (FSH), adrenocorticotropic hormone (ACTH), and prolactin. One important point is that several studies have shown that progesterone may decrease the reinforcing effects of numerous drugs including nicotine (Lynch and Sofuoglu, 2010; Quinones-Jenab and Jenab, 2010). Thus, one possibility is that chronic cocaine self-administration in this cohort of female monkeys resulted in prolonged progesterone depletion which potentiated nicotine + drug-induced reinstatement. Future work should explicitly test whether variability in nicotine sensitivity in monkeys is related to menstrual cycle phase, hormonal differences, and/or other neurobiological differences between males and females.

Although the reinstatement paradigm has been used for decades as a preclinical model of relapse, it has important limitations. Perhaps most notably, many preclinical reinstatement procedures show limited functional equivalence, and in some cases even negative associations, with relapse outcomes in the clinical literature (Katz and Higgins, 2003). One key issue is that relapse outside the laboratory typically occurs following drug exposure during ongoing access, rather than after experimenter-delivered “pretreatments” [or what Katz and Higgins (62) referred to as an “initiator” of relapse]. Despite these limitations, extending reinstatement models to a choice procedure, where behavior is not extinguished but instead reallocated to an alternative is a meaningful step toward improving translational relevance.

A second limitation, particularly relevant to comparisons of cocaine and cocaethylene potency, is the monkeys’ drug histories. All animals had substantially longer cocaine self-administration histories than cocaethylene histories prior to this experiment. This matters because saline was delivered on the same operant switch and paired with the same discriminative stimuli as the drug condition, and those cues were likely more strongly associated with cocaine than cocaethylene. Given this, the finding that cocaine was more potent than cocaethylene in inducing drug-induced reinstatement should be interpreted in the context of this important caveat. Future studies should evaluate the role that drug history plays in modifying the variables examined in this study. Moreover, because cocaine was more potent that cocaethylene, it is possible that cocaine had a greater maximal effect on % drug-associated choice than cocaethylene because high enough doses of cocaethylene were not tested. This is an important limitation to consider when interpreting the findings of this study.

Another consideration is that in some monkeys, the addition of nicotine to cocaine or cocaethylene reduced the total number of trials completed, suggesting potential rate-decreasing effects at higher dose combinations. However, in the primary group-level analyses (Figures 2A,B), the cocaine + nicotine and cocaethylene + nicotine doses included were rarely those that produced decreases in total trials. In fact, for the cocaethylene + nicotine analysis, only data from F-8555 were included beyond the point at which trial completion declined (Supplementary Figure 6A), and for the cocaine + nicotine analysis, only data from F-8551 were included at that point in the dose–response curve (Supplementary Figure 5A). Thus, rate-decreasing effects of these combinations likely had minimal impact on the group-level statistics reported. Finally, our model differs from typical human routes of self-administration which impacts drug pharmacokinetics. Cocaethylene is a metabolite formed in the liver when alcohol is consumed prior to cocaine use (Harris et al., 2003; Landry, 1992). In this study, cocaethylene was available intravenously, which may produce neuropharmacological effects that differ from the in vivo formation of cocaethylene. The same concern applies to nicotine: in humans it is most commonly inhaled and can be taken sequentially with cocaine or cocaine plus alcohol rather than simultaneously. By not modeling the route and temporal dynamics of nicotine or cocaine + alcohol self-administration during co-use, our procedure may not capture all clinically relevant features of polysubstance use.

Despite these limitations, the present design incorporated key variables, including sex, social rank and polysubstance use, and generated several testable hypotheses. Most notably, the effects of nicotine on cocaine- and cocaethylene-induced reinstatement differed by sex. Clinically, nicotine use is not routinely considered in trials for CUD, and smokers and non-smokers are often not analyzed separately (Anderson et al., 2009; Dackis et al., 2012; Kampman, 2019). Our findings suggest that continued nicotine exposure may be especially consequential for women with CUD and should be considered when designing and interpreting treatment studies. Importantly, while several studies suggest that alcohol use may reduce the efficacy of treatments for CUD (Anderson et al., 2009; Rough et al., 2025a), the impact of continued alcohol use on relapse probability for CUD has not been systematically investigated. This remains an important direction for future work. Overall, these data highlight the translational value of incorporating common polysubstance-use patterns, sex, and environmental variables into preclinical models to better capture factors that influence treatment outcomes for CUD.