The Importance of Examining Sex Differences in Animal Models Validated to Induce an addiction-like Phenotype

In 2021, females are still routinely left out of biomedical research with no justification or included but ignored in the analyses of data. A recent report that analyzed papers published in 2019 across 34 journals and nine biological disciplines showed that while the number of sex-inclusive research studies has increased (as compared to 2016), females and sex differences remain understudied given that only 49% of these studies included females and males, and of the articles that include both sexes, only 42% conducted sex-based analyses (Woitowich et al., 2020). As women in the field, we want to use this article to highlight the need for more research designed to study sex differences in substance use disorder since biological differences, such as sex, may be critical to developing successful treatments for this disease. We highlight a specific need to address sex differences in animal models validated to induce an addiction-like phenotype like that observed in humans with substance use disorder (compulsive use, an enhanced motivation to use the drug, vulnerability to relapse). This is an important distinction since, thus far, preclinical studies have focused predominantly on initial vulnerability to drug use as assessed under short-access drug self-administration conditions (i.e., 1–2 hours/day access). Extended-access drug self-administration procedures (i.e., ≥ 6-hr/day access to the drug) are the gold-standard for inducing addiction-like features in animals, including compulsive drug use, as measured by drug use despite negative consequences (e.g., foot shock), an enhanced motivation to obtain the drug, as assessed using a progressive-ratio schedule or a within-session threshold procedure, and enhanced drug-seeking/vulnerability to relapse, as assessed using extinction and reinstatement procedure (Lynch, 2018). These characteristics emerge following withdrawal from extended, but not short-access self-administration, and are associated with distinction neurobiological changes (Lynch 2018; Roberts et al., 2007; Ahmed and Koob, 1998; Wolf and Tseng 2012; Zimmer et al., 2012; Lynch et al. 2021). Thus, the overarching goal of this opinion paper is to discuss the importance of using animal models validated to induce the diseased state when investigating sex differences in development, neurobiological basis, and treatment of substance use disorder. We also offer some suggestions on how to design and interpret sex differences (or lack thereof) in these animal models.

The Importance of Using a Validated Animal Models of Substance Use Disorder

Preclinical studies on sex differences in addiction have focused predominantly on initial vulnerability to drug use, as assessed by rates of acquisition of drug self-administration and maintenance levels of drug intake under short-access self-administration conditions (1–2 hour/day access to the drug; Ahmed and Koob, 1998; Lynch, 2018; Kuhn et al., 2019). While these short-access self-administration studies have been useful for establishing sex differences in vulnerability to drug use, with evidence indicating greater vulnerability in females (Becker 2016; Lynch and Carroll, 1999; Carroll et al., 2004; Campbell and Carroll, 2000), the findings may not apply to the disease state of addiction considering that such procedures generally do not capture critical features of substance use disorder in humans (see review: Lynch 2018; Kuhn et al., 2019). This disconnect from the clinical situation is critical considering evidence from both human and animal studies indicating that the neurobiological mechanisms underlying drug use change over time as the disease develops (Koob and Volkow, 2016). Indeed, we argue that the use of extended-access drug self-administration procedures validated to induce addiction-like features like those observed in women and men with substance use disorder would increase the translational relevance of preclinical studies on addiction, and with regard to medications development, would help address the “translation failure” of these studies to predict clinical outcomes (e.g., Negus and Banks 2020). Similarly, we argue that in order to produce a more representative and translationally relevant body of knowledge on sex differences in the neurobiological basis and potential treatments for addiction, studies need to use animal models validated to induce addiction-like features like those observed in women and men with substance use disorder.

As mentioned above, extended-access procedures are the gold-standard for inducing addiction-like features in animals. That said, extended-access procedures vary widely and not all have been validated to induce an addiction-like phenotype. For example, with some extended-access procedures animals are not first trained to self-administer the drug (e.g, Caprioli et al., 2015); thus, some of the extended-access period includes acquisition which varies between subjects and may confound sex differences in drug intake and the development of an addiction-like phenotype. In our hands, some animals take 7 or more days to acquire, even after lever-press pre-training, so if the extended-access period is only 10 or 14 days long it may not be sufficient to induce an addiction-like phenotype. Similarly, some extended-access procedures restrict the number of infusions available each hour (e.g., Venniro et al., 2017a) which may obscure/confound sex differences (and other individual differences) in levels of drug self-administration and may prevent the later development of the addiction-like phenotype. The need for validation is exemplified by a series of publications showing that rats given “extended-access” to opioids or stimulants continue to show a near exclusive preference for a non-drug reinforcer, such as palatable food or social interaction, over infusions of the drug (e.g, Caprioli et al., 2015; Venniro et al., 2018). Given that this phenotype is in stark contrast to the human condition where non-drug reinforcers are neglected in preference to the drug, we recommend that researchers verify that the extended-access conditions used induce an addiction-like phenotype either by comparing to a baseline obtained prior to extended self-administration and withdrawal or comparing to behavior observed in short-access controls. While numerous studies have used a time-point during early withdrawal to verify the later development of an addiction-like phenotype (e.g., day 1 to verify the incubation of craving; Caprioli et al., 2015), this comparison is not ideal for assessing sex differences given that behavioral phenotypes of addiction (enhanced drug-seeking and motivation for the drug) decrease to very low levels during early withdrawal (Corbett et al., 2012) with differential effects in males and females (Lynch and Taylor, 2005).

Unfortunately, conclusions of sex differences (or lack thereof) in substance use disorders are often made based on findings from animal models not validated to induce the diseased state. For example, in a recent review on sex differences in drug craving and relapse, Nicolas et al., (2021) concluded that “the available data do not support the notion of sex differences in craving and relapse/reinstatement for methamphetamine or heroin in rodent models”. However, we argue that this conclusion is premature considering that most of the evidence has been obtained from studies using either short-access or restrictive “extended-access” conditions not validated to induce an addiction-like phenotype (e.g., no acquisition period prior to extended-access self-administration, limits on daily intake).

In summary, there is a significant difference between recreationally drug use and substance use disorders, and while animal models are needed for both phases, we highlight a particular need for additional research on sex differences in substance use disorders using validated animal models. We further recommend that conclusions of sex differences in the neurobiological mechanisms and potential treatments for substance use disorder be based on models that have been verified to induce an addiction-like phenotype as discussed above.

Determining Sex Differences in Validate Animal Models of Substance Use Disorder

Sex differences, and other individual differences, are best studied under threshold conditions that induce a range of behavioral output (e.g., both high and low levels of drug-seeking/taking or that induce an addiction-like phenotype in some, but not all animals). A great example of this is the telescoping effect which reflects a faster time course to addiction in females. It was originally reported in women for alcohol use disorder, but has also been noted for other drug classes including, psychostimulants, opioids, and cannabis (Anglin et al.,1987; Brady and Randall 1999; Griffin et al.,1989; Hernandez-Avila et al.,2004; McCance-Katz et al.,1999; Westermeyer and Boedicker 2000; Mayo et al.,2019; Back et al.,2011a, b; Haas and Peters 2000; Hser et al.,1987a, b; Sartor et al., 2014). We and others have also demonstrated that a telescoping effect also occurs in females in rat models of cocaine use disorder (Kawa et al., 2019; Lynch and Taylor, 2004; Ramôa et al., 2013; Towers et al., 2021). We showed that this effect is revealed following extended-access self-administration and an intermediate period of withdrawal (i.e., 7 days), which was optimal for inducing an addiction-like phenotype in females (defined as an enhanced motivation for cocaine relative to baseline prior to extended-access self-administration), but threshold for males such that only a minority of males displayed this phenotype (Towers et al., 2021). We have noted similar effects for compulsive use (Towers et al., 2021), as well as cued-induced relapse, with results indicating that in females, these features are maximally expressed following 7 days of withdrawal, whereas, males require 14 days.

A number of factors may impact the detection of sex differences in animal models of substance use disorder; therefore, there needs to be serious consideration of the experimental design and conditions prior to concluding a lack of sex differences. For example, in the context of the telescoping effect drug access during the extended-access period is likely critical. Specifically, in the most commonly used extended-access procedure, the long-access procedure developed by Ahmed and Koob (1998), animals first undergo an acquisition phase and then have unlimited fixed-ratio 1 (FR1) access to a drug for 6 or more hours/day and show high levels of drug intake which escalates over time. This procedure has been validated relative to short access controls to induce the development of an addiction-like phenotype, including compulsive use, an enhanced motivation for the drug, and high levels of drug-seeking/vulnerability to relapse when assessed following a period of protracted withdrawal (Ahmed et al., 2000; Barbier et al., 2013; Lenoir and Ahmed, 2006; Vendruscolo et al., 2011). However, previous studies indicate that the addiction-like phenotype emerges sooner during withdrawal and following less drug exposure when animals are given extended-access to drug using intermittent-access conditions (either in discrete trials or in discrete 5-min bins; Morgan et al., 2005; Zimmer et al., 2012; Allain et al., 2015). These differences are important to consider from a sex difference perspective since, as mentioned above, the time-course for the development of an addiction-like phenotype is different in females and males. Therefore, with the long-access procedure a later withdrawal time-point may be necessary to detect ex difference reported using the intermittent access procedure

Sex differences are also more likely to be revealed under low versus high dose conditions since these conditions engender greater variability in levels of intake (Carroll et al., 2004). High drug doses are also more likely than low doses to induce non-specific or aversive effects which can obscure sex differences in the reinforcing effects of the drug (Lynch and Carroll 2001; Torres et al., 2014). Environmental stressors may also interact with sex to obscure sex differences in the development of an addiction-like phenotype. For example, environmental stressors, such as neonatal isolation stress or food restriction, can obscure and even reverse the female enhanced vulnerability during both initial and later stages of the addiction process (Lynch et al., 2005; Campbell and Carroll, 2000; Swalve et al. 2016).

Lastly, statistical power and the sensitivity of the statistical analyses performed are additional key considerations that impact the ability to detect and draw conclusions of sex differences in rat models of substance use disorder. The sample size needs to be large enough to test the overall and interactive effects of sex and the main dependent variable (Murray et al., 2017). This requires a sample size larger than a study designed to investigate the main effect of a dependent variable. The sample size for females may also need to be larger than that of males in order to determine the effect of estrous cycle as phase of the estrous cycle is known to impact different aspects of the disease process (Kerstetter et al.,2008; Peterson et al.,2014; Lynch et al.,2019). For example, cue-induced relapse responding is similar between males and females following extended-access self-administration and withdrawal yet responding is highest in females tested during estrus as compared to males and females tested during non-estrus phases (Corbett et al., 2012; Bakhti-Suroosh et al., 2020; Nicolas et al., 2019). Thus, sex differences may be dependent on the percentage of animals tested during estrus versus non-estrus phases and larger samples may be required to assess effects at different estrous cycle phases, particularly for assessments conducted in one session on one day, as opposed to over several sessions on multiple days. It is also important to consider that the estrous cycle of female rodents singly housed has the ability to become synchronized (McClintock, 1978) and drug self-administration may result in disruptions of estrous cyclicity (Broderick and Malave, 2014), which can result in a greater percentage of females tested during estrus versus non-estrus phases; such effects may obscure or magnify sex differences.

In conclusion, we encourage researchers in the field to specifically examine sex differences in the development, neurobiological basis, and treatment of substance use disorder since we believe that such knowledge is necessary to move the science forward and produce a more representative and translationally relevant body of knowledge. However, such studies require careful consideration in terms of design and analysis beginning with a model that is validated to induce an addiction-like phenotype and that engenders individual differences. We also strongly recommend that conclusions of a lack of sex difference on particular aspects of substance use disorder be made only after thorough parametric evaluation since such conclusions may have important clinical implications. Instead, in cases where both sexes are included and similar effects observed, unless the study was designed and powered to detect sex differences, it is more appropriate to conclude that similar effects occurred in males and females rather than there was no sex difference. This will hopefully help further understand the disease process in females and males with the goal to provide equitable treatment strategies to women and men.