Inbred mouse strain differences in alcohol and nicotine addiction-related phenotypes from adolescence to adulthood

Abstract

Understanding the genetic basis of a predisposition for nicotine and alcohol use across the lifespan is important for public health efforts because genetic contributions may change with age. However, parsing apart subtle genetic contributions to complex human behaviors is a challenge. Animal models provide the opportunity to study the effects of genetic background and age on drug-related phenotypes, while controlling important experimental variables such as amount and timing of drug exposure. Addiction research in inbred, or isogenic, mouse lines has demonstrated genetic contributions to nicotine and alcohol abuse- and addiction-related behaviors. This review summarizes inbred mouse strain differences in alcohol and nicotine addiction-related phenotypes including voluntary consumption/self-administration, initial sensitivity to the drug as measured by sedative, hypothermic, and ataxic effects, locomotor effects, conditioned place preference or place aversion, drug metabolism, and severity of withdrawal symptoms. This review also discusses how these alcohol and nicotine addiction-related phenotypes change from adolescence to adulthood.

1. Introduction

Nicotine and alcohol are considered to be the first and third most prevalent preventable causes of death in the United States, respectively (Mokdad et al., 2004). In the US alone, nicotine use accounts for over 480,000 deaths and over $380,000,000,000 in costs to America’s economy annually (CDC, 2014; Xu et al., 2021). Alcohol is responsible for an estimated 95,000 deaths and a cost of $249,000,000,000 to the United States’ economy annually (Esser, 2020; Sacks et al., 2015). Both drugs have been associated with a myriad of health concerns including cardiovascular disease and cancer (England et al., 2017; WHO, 2019; Mishra et al., 2015; Rehm et al., 2006). Thus, understanding factors that influence the use and abuse of nicotine and alcohol is of critical importance for national health efforts.

Numerous factors are responsible for individual variability in the likelihood to use and abuse drugs. Nicotine and alcohol dependence are heritable (Verhulst et al., 2015; Vink et al., 2005), but there are different ways through which genetic factors can impact addiction. Cognitive and personality traits, such as deficits in executive function and increased impulsivity, are associated with increased risk of drug dependence (Ersche et al., 2012). Additionally, sensitivity to the positive and negative effects of drugs can impact the likelihood of developing an addiction (Schuckit, 1994). Experiencing greater positive effects of drugs, such as euphoria, and fewer negative effects of drugs, such as hangover, can support patterns of regular drug use (King et al., 2021; Piasecki et al., 2010, 2012). The genetic factors that affect these phenotypes (cognitive function, impulsivity, and drug sensitivity) most likely interact to shape a propensity for addiction and Substance Use Disorders (SUD).

Age of initiation of substance use is another factor that contributes to developing a SUD. Alcohol and nicotine use during adolescence may be particularly problematic. Adolescents who use nicotine or alcohol are more likely to later use drugs and become dependent on these drugs in adulthood (Chassin et al., 1990; Grant et al., 2006). Among youth aged 12–20 in the United States, 7,046,000 report alcohol use and, of them, 4,222,000 report binge alcohol use within the last month (McCance-Katz, 2019). In other words, over half of adolescent alcohol users consume alcohol in dangerous, binge-like quantities. Additionally, in 33 % of adolescents using tobacco intermittently, as infrequent as once a month, signs of nicotine addiction emerged within two months (DiFranza et al., 2002). That percentage increased to 70 % before beginning to smoke daily. In one year, of youth aged 12–20, 6,062,000 reported using tobacco products and 12,765,000 reported consuming alcohol (McCance-Katz, 2019). Long-term consequences of nicotine and alcohol during adolescence (Chassin et al., 1990; Grant et al., 2006) combined with the high rate of experimentation and use clearly illustrate the need for intervention and prevention during adolescence in order to curb the devastating impact of drug abuse and addiction (Rahman and Bell, 2021). While rates of cigarette use has declined in adolescents, youth e-cigarette use now exceeds use of traditional cigarettes (Arrazola et al., 2015). This is concerning because nicotine, by itself, can have negative health impacts (England et al., 2017; Mishra et al., 2015) and vaping nicotine is highly addictive (Jankowski et al., 2019).

Mouse genetic models are powerful tools that enable the dissection of factors that influence phenotypes in order to identify genetic contributions. Whereas the genetic diversity within an inbred strain does not model human genetic diversity, researchers can use techniques such as inbred mouse strain comparisons and mapping quantitative trait loci to identify genes of interest associated with aspects of addiction-related behaviors (Buck and Finn, 2001; Crabbe et al., 1999). While the genetic bases of traits may vary between mice and humans, these genetic findings in mice can be translated at the level of gene, protein, or neurobiological network in humans. These models hold the potential to aid in the understanding of how drug exposure interacts with genetic background to influence adaptive and maladaptive outcomes, in both adolescents and adults. For this reason, many studies have investigated phenotypes related to substance use in rodent models. Outbred (genetically heterogeneous) and inbred (isogenic) rodent lines have been utilized to assess phenotypes across genetically diverse or similar populations (Chesler, 2014).

This review will highlight phenotypes related to substance use in inbred populations, due to their utility in examining genetic substrates and the extensive work done characterizing them. Mouse lines are considered to be inbred when they have undergone 20 generations of inbreeding (e.g., sister × brother, parent × offspring mating). A population is considered to be a substrain through inbreeding separate from the parent population for many generations, or from suspected genetic differences noted from the parent strain (The Jackson Laboratory, 2022). This can also result from genetic drift when breeding a parent strain in different locations (or vendors) for many generations. Parent strains are named with letters and/or numbers only, and substrains are named with that parent name, a forward slash, and other letters and/or numbers to indicate a unique identity (The Jackson Laboratory, 2022). To aid in the interpretation of strain differences, Table 1 lists the relationships between parent strains and substrains. Substrains are listed throughout the text where possible, but in some cases, substrains were not specified in the reviewed literature and strain names are listed as they appeared. Readers who are interested in specific phenotypic differences across strains and substrains that are not discussed in this review are encouraged to visit the Mouse Phenome Database (https://phenome.jax.org), which extensively catalogues reported phenotypic differences across strains (Bogue et al., 2020).

Table 1.

Inbred mouse parent strains and substrains mentioned throughout the review.

Parent strain	Substrains mentioned in this review
129	129S1/SvlmJ, 129P3/J, 129/J, 129/SvEv, 129S6/SvEvTac
A	A/J, A/HeN
AHe	AHe/J
AKR	AKR/J
BALB	BALB/cJ, BALB/cByJ, BALB/cAnN, BALB/Ibg
BTBR	BTBR + T(tf/tf), BTBR T + tf/J
BUB	BUB/BnJ
C3H	C3H/HeJ, C3H/HeNMTV, C3H/HeN, C3H/Ibg
C57	C57BL/6J, C57BL/6NJ, C57BL/6NCrl, C57BLKS/J, C57L/J, C57BL/ 10N, C57BR/cdJ, C57BL/6N, C57BL/6lbg, C57BL/6NTac
C58	C58/J
CAST	CAST/EiJ
CBA	CBA/J
CE	CE/J
DBA	DBA/2J, DBA/1J, DBA/2N, DBA/2Ibg
FVB	FVB/NJ
LP	LP/J
MA	MA/MyJ
NMRI
NOD	NOD/ShiLtJ
NZB	NZB/B1NJ
NZM	NZM/M1NJ
NZO	NZO/HtLtJ
PERA	PERA/Ei
PL	PL/J
PWK	PWK/PhJ
SEX	SEX/1ReJ
SJL	SJL/J
SM	SM/J
SPRET	SPRET/Ei
ST	ST/bJ
SWR	SWR/J
WSB	WSB/EiJ

Alcohol and nicotine have many distinct and some overlapping biological targets that produce the unique phenotypes discussed in this manuscript. Alcohol has a wide range of targets, including GABAergic, glutamatergic, dopaminergic, serotonergic, opioid, and cholinergic systems (Chastain, 2006; Miller and Kamens, 2020). Nicotine acts more specifically on cholinergic receptors, modulating other downstream targets, including glutamatergic and dopaminergic systems (Benowitz, 2009). The described strain differences in alcohol and nicotine behavioral phenotypes likely represent complex, interactive strain differences in these (and other) biological systems. Because of the known importance of mesolimbic dopamine signaling in reward and addiction, dopamine-associated phenotypes will be emphasized throughout this review.

The phenotypes to be reviewed include measures of voluntary consumption and self-administration, which are gold-standard assays for addiction, measures of initial sensitivity to drug action (e.g. hypothermia, ataxia, locomotor activity and anxiety) that can affect the likelihood of continued use of the drug, conditioned place preference and aversion to evaluate the rewarding and aversive properties of drug use, and the effect of withdrawal on multiple phenotypes. While not exhaustive, focusing on these phenotypes will aid in understanding genetic contribution to the different factors known to contribute to SUD in humans. Where studies are available, we will include reviews of data highlighting genetic differences in patterns of consumption and drug effects in adolescents and adults.

2. Alcohol phenotypes

2.1. Background

Studies of human populations have consistently demonstrated the importance of genetic background in determining patterns of alcohol use and misuse. Alcohol Use Disorders (AUD) are estimated to be 50 % heritable (Begleiter and Kissin, 1995; Schuckit, 2014; Verhulst et al., 2015), but the way that genetic background influences AUD risk is complex. Separable risk factors for and characteristics of AUD may be influenced by different genetic variants (Tawa et al., 2016). Identification of these variants can be difficult in humans, where additional factors (such as social setting and prior experience) can modulate genetic effects. However, modeling specific alcohol-related phenotypes in mice can allow for dissociation of genetic factors influencing each phenotype (Crabbe et al., 1999). In addition, these models allow for the investigation of how genetic background may interact with adolescent age to produce unique alcohol phenotypes. Existing reviews have examined how alcohol effects vary across adolescent, adult, and aged rodents (Matthews and Imhoff, 2021), but the following sections will emphasize differences in adolescent and adult phenotypes across genetic (strain) backgrounds.

2.2. Voluntary consumption

Understanding the factors that drive alcohol consumption and escalation of drinking behavior can inform understanding of alcohol abuse and AUD. Voluntary consumption of ethanol has been studied extensively in inbred mice. Various paradigms have been developed to assess alcohol consumption during continuous and intermittent access, in an effort to understand consumption of moderate amounts of alcohol and drinking to intoxication.

Two-bottle choice procedures allow for simple assessment of alcohol consumption and preference. Generally, subjects are acclimated over multiple days to two sipper bottles in their home cages. One bottle contains water, and the other contains a diluted ethanol solution. The volume or dose of ethanol consumed is monitored by observing liquid levels on graduated drinking bottles or by weighing the drinking bottle. Drinking in this paradigm generally does not lead to intoxication. In this model a notably high-drinking strain is C57BL/6J (Bagley et al., 2021; Yoneyama et al., 2008). However, recent data shows that the PWK/PhJ strain, a founder strain of the Collaborative Cross (CC) and Diversity Outbred (DO) populations, consumes even more ethanol than C57BL/6J in this paradigm (Bagley et al., 2021). Across two studies examining strain differences in ethanol consumption (Yoneyama et al., 2008; male and female; 3–10 % v/v EtOH; and Bagley et al., 2021; male and female; 20 % w/v EtOH), the relatively high-consuming strains included C58/J and C57BLKS/J, and relatively low-consuming strains included DBA/2J, BUB/BnJ, BTBR + T(tf/tf), and NZO/HtLtJ. Sex effects, while noted, were inconsistent across strain backgrounds, suggesting genetic background interacts with sex to produce unique ethanol consumption patterns (Bagley et al., 2021; Yoneyama et al., 2008).

Drinking to intoxication, another aspect of alcoholism, has been more difficult to model in rodents. A common way to encourage binge-like drinking behavior in mice is to use a Drinking in the Dark (DID) paradigm (Rhodes et al., 2005), where subjects are given access to ethanol bottles early in their dark phase when they are active. This can be sufficient to produce pharmacologically significant blood ethanol concentration (BEC) levels in C57BL/6J mice, representing intoxication. Similar to patterns seen in two-bottle choice studies, a study surveying strain differences by Rhodes et al. (2007) (male and female; 20 % v/v EtOH) found that C57BL/6J mice consumed large amounts of ethanol in this paradigm (~6.8 g/kg EtOH). Both BALB/cJ and BALB/cByJ strains consumed relatively large amounts (~5.9 and 4.4 g/kg EtOH, respectively), whereas LP/J and 129S1 strains consumed smaller amounts (~2.1 and 1.6 g/kg EtOH, respectively), and the DBA/2J strain consumed very little ethanol (~1.5 g/kg EtOH).

In humans, preference for alcoholic beverages can be motivated or altered by flavors or sweetness of the beverage. Thus, genetic variability in taste can change alcohol consumption patterns, and this can be observed in rodent models. Strain differences in alcohol consumption can be modulated by altering sweetness in voluntary consumption experiments. For example, adding saccharin to ethanol solutions increases voluntary consumption in most strains and changes relative drinking profiles (Yoneyama et al., 2008; male and female; 3–10 % v/v EtOH). Adding saccharin to the ethanol solutions did not increase drinking in 129S1/SvImJ, LP/J, A/J, BUB/BnJ, and DBA/2J strains but nearly quadrupled drinking in PERA/Ei, FVB/NJ, and BALB/cByJ strains. However, consumption of sweet saccharin solutions seems to be driven by a mechanism that is at least partially distinct from that which drives ethanol consumption. Comparisons of strain averages of consumption volumes demonstrated that consumption of sweetened ethanol solution correlated much more strongly with the saccharin-only solution than the unsweetened ethanol solution (Yoneyama et al., 2008). Additionally, another study (Fidler et al., 2011; male; 20 % v/v EtOH) found that intragastric ethanol self-administration patterns are sometimes unique from voluntary oral ethanol consumption patterns. In this study, subjects were allowed to self-administer intragastric ethanol for four days, and they found that DBA/2J mice consumed relatively large amounts of ethanol. These findings are in contrast to oral consumption studies, where DBA/2J mice consistently consume very little ethanol relative to other strains (Bagley et al., 2021; Rhodes et al., 2007; Yoneyama et al., 2008). Further, prior ethanol exposure created a greater increase in ethanol self-administration in DBA/2J mice relative to C57BL/6J mice, suggesting that genetic background determined the effect of prior experience on ethanol self-administration. This suggests that taste-related factors influence oral ethanol consumption in mice, and that the extent to which prior ethanol exposure encourages subsequent ethanol self-administration is dependent upon strain.

Continued drug use despite negative consequences is a key component of addiction. To test the strength of the drive to consume ethanol, voluntary consumption after self-administration with negative consequences was tested in a small strain panel (Halladay et al., 2017; male; 10 % EtOH). Subjects were trained to lever-press for ethanol, and eventually footshock punishments were administered in response to some lever presses. Testing measured self-administration after punishment sessions, with 129S1/SvImJ and BALB/cJ exhibiting the largest reductions in ethanol consumption after punishment, while this effect was context-dependent in C57BL/6J mice and not observed in DBA/2J mice. DBA/2J mice were also less sensitive to reduction of consumption by addition of a bitter tastant, denatonium benzoate. Cumulatively, these findings suggest that ethanol consumption is driven and deterred by different factors across strains.

While many previous studies note the unique high-drinking profile of the C57BL/6J strain, this phenotype differs among C57BL/6 substrains (Mulligan et al., 2008). This is an exciting area of research because the relatively small genetic differences between substrains with different ethanol responses provides a unique opportunity to discover causal genetic variants. One noteworthy study (Mulligan et al., 2008; male and female; 3–21 % w/v EtOH) found that two substrains, C57BL/6J and C57BL/6NCrl, differed significantly in ethanol consumption in a two-bottle choice paradigm, with C57BL/6J mice consuming more than C57BL/6NCrl mice. These strains did not differ in initial sensitivity to ethanol, measured as the latency to loss of righting reflex (LORR), duration of LORR, and BEC after recovery from LORR after ethanol treatment. Comparisons of alcohol-related brain transcriptome changes between the strains allowed authors to associate phenotypes with genetic loci that differed between strains. This led to the identification of candidate genes, including D14Ertd449e, H2afz, Psen1, Wdfy1, and Clu, likely to be involved in regulating ethanol consumption.

Age also seems to importantly influence ethanol consumption. Adolescent subjects may consume more or less ethanol in social settings, suggesting that social setting importantly affects ethanol consumption (Logue et al., 2014; Panksepp et al., 2017). Age effects on ethanol consumption can also vary across strain backgrounds. Comparisons across adolescence and adulthood in C57BL/6J and DBA/2J demonstrate the importance of both age and genetics in determining ethanol consumption (Moore et al., 2010; male and female; 20 % EtOH). In a DID paradigm modeling binge-like alcohol consumption by Moore and colleagues, adolescent C57BL/6J mice were found to voluntarily drink more than adult C57BL/6J mice. However, in the DBA/2J strain, adolescents and adults drank comparable volumes of ethanol. Thus, in the C57BL/6J strain, age is an important factor in determining voluntary binge-like ethanol consumption, but this is not the case in the DBA/2J strain, at least under the conditions tested. This study demonstrated that genetic background can interact with age to produce unique ethanol consumption phenotypes. Age and ethanol consumption may also interact with sex. Notably, Moore et al. (2010) also reported that females often consumed more ethanol than males, and human research suggests that adolescent binge drinking rates are increasing in females (Dir et al., 2017). This suggests that there are sex- or gender-dependent differences in adolescent alcohol consumption that warrant further study in rodent subjects in order to understand the interaction of age, sex, and genetics.

In summary, many studies have demonstrated variability in alcohol consumption behaviors across inbred strains. A well-studied strain that consistently drinks large quantities across drinking paradigms is C57BL/6J, although the PWK/PhJ strain has been shown to consume even more ethanol (Bagley et al., 2021; Rhodes et al., 2007; Yoneyama et al., 2008). The extreme drinking phenotype is not as apparent in a closely related C57BL/6 substrain, C57BL/6NCrL (Mulligan et al., 2008). The DBA/2J strain, however, consistently consumes little ethanol (Bagley et al., 2021; Rhodes et al., 2007; Yoneyama et al., 2008), and may be driven to drink by unique mechanisms (Fidler et al., 2011; Halladay et al., 2017). Other factors that can influence strain patterns of ethanol consumption include sweetness (Yoneyama et al., 2008), negative reinforcement (Halladay et al., 2017), and adolescent age (Moore et al., 2010).

2.3. Conditioned preference and aversion

Another component of SUD modeled in rodents is the rewarding and aversive associations formed between a drug stimulus and its associated context or cues. Conditioned place preference (CPP), conditioned place aversion (CPA), and conditioned taste aversion (CTA) paradigms all measure subjects’ propensity to seek out or avoid a context or cue that was associated with drug exposure. In ethanol CPP and CPA training, subjects are treated with a rewarding (CPP) or aversive (CPA) dose of ethanol and allowed to explore one chamber of a multi-chamber behavioral apparatus (Cunningham, 2014, 2019; Gremel and Cunningham, 2007). On testing day, subjects do not receive ethanol and are allowed to explore both chambers. The amount of time that subjects spend in the chamber they associated with the drug is compared to the time they spent in the other chamber, creating a preference or aversion score. In CTA, subjects are trained to associate consumption of a neutral or positive cue (conditioned stimulus; CS) with an aversive dose of ethanol (Chester et al., 1998; Moore et al., 2013). A negative change in consumption between pre- and post-conditioning sessions demonstrates aversion to the ethanol-associated cue. CPP, CPA, and CTA have been used to study the reinforcing potential of ethanol across genetic backgrounds (Crabbe et al., 1992).

CPP is useful for studying positive reinforcing effects of ethanol. One study examined ethanol-induced CPP in a panel of 15 common inbred mouse strains (Cunningham, 2014; male; 2 or 4 g/kg i.p., 20 % v/v EtOH). Preference was measured using two doses of ethanol in a 10-day protocol, consisting of one acclimation day, eight training days, and one testing day. On testing day, place preference was measured as percent time spent in the ethanol-paired chamber from training. Generally, strains that showed weak ethanol preference (close to 50 % time spent in ethanol-paired chamber) were C57L/J, BALB/cJ, C58/J, and C57BL/6J. Strains with notably strong ethanol preference (close to 70 % time in ethanol-paired chamber) were AHe/J, NZB/B1NJ, DBA/2J, and AKR/J. Some of these strain differences are counterintuitive when compared to ethanol self-administration. The C57BL/6J strain, known for consuming large amounts of ethanol, showed weak conditioned place preference (Bagley et al., 2021; Yoneyama et al., 2008). However, the DBA/2J strain, known for consuming small quantities of ethanol, showed strong conditioned place preference. This could suggest that either mice are self-administering ethanol for reasons additional to or other than reward, or that mice with weak CPP need to administer higher levels of ethanol to achieve reward.

CPA can demonstrate aversive effects of ethanol. Another study by Cunningham (2019; male; 2 or 4 g/kg i.p.,20 % v/v EtOH) also examined two doses of ethanol with a 10-day CPA protocol in the same 15 inbred strains. Strains that showed weak CPA (close to 50 % time in ethanol-paired chamber) were PL/J, C58/J, C57L/J, and NZB/B1NJ. It is notable that in the previous study (Cunningham, 2014), two of these strains (C57L/J and C58/J) also showed weak CPP, but one strain (NZB/B1NJ) had shown strong CPP. Strains that showed strong CPA (close to 30 % time in ethanol-paired chamber) were 129P3/J, SWR/J, CBA/J, and DBA/2J.

When interpreting CPP and CPA findings, it is important to consider that strain differences in other measures, such as activity levels, could interfere with interpretation of measured preference or aversion. A study by Gremel and Cunningham (2007; male; 2 g/kg i.p., 20 % v/v EtOH) examined ethanol CPP in DBA/2J and NZB/B1NJ strains, which both showed strong ethanol conditioned place preference. Additional experimental groups examined CPP of both strains when an ethanol injection was administered prior to testing. They found that pre-testing ethanol markedly increased activity levels in DBA/2J, and not NZB/B1NJ, mice, and this was sufficient to block expression of ethanol conditioned place preference in DBA/2J, but not NZM/M1NJ mice. This suggests that strain differences in activity levels may have the potential to interfere with the expression of ethanol CPP or CPA. Further, it is important to note that conditioned ethanol-related behaviors can also be altered by body temperature, ambient temperature, and even environmental enrichment (Cunningham et al., 1988; Dickinson and Cunningham, 1998; Pautassi et al., 2017), which highlights the complexity and sensitivity of these ethanol-induced behaviors.

CTA, like CPA, can be used to understand the aversive effects of ethanol. Unlike CPA, CTA studies measure aversion as reduced consumption of a drug-associated cue instead of reduced time spent in a drug-associated context. Thus, the expression of aversion in CTA can be affected by taste and consumption patterns. One notable study examined ethanol CTA across eight inbred strains, including both adolescent and adult subjects (Moore et al., 2013; 1.5, 2.25, & 3 g/kg i.p.). Aversion was measured as percent change in CS (saline) consumption between pre- and post-conditioning sessions. Among adult subjects, C3H/HeJ, 129S1/SvlmJ, DBA/2J, and FVB/NJ strains generally showed strong CTA (close to 100 % decrease in consumption). BTBR T + tf/J and C57BL/6J strains showed relatively weak CTA (close to 0 % change in consumption). Patterns among adolescent subjects were slightly different. Strains with strongest CTA were 129S1/SvlmJ, BTBR T + tf/J, FVB/NJ, DBA/2J, and C3H/HeJ. Strains with weaker adolescent CTA were C57BL/6J and BALB/cByJ. Across both age groups, C57BL/6J mice demonstrated weak CTA. While there was some variability in this pattern, generally, most strains had weaker CTA in adolescence than in adulthood. However, this study demonstrated that this may vary by genotype — notably, in the BTBR T + tf/J strain, adolescents show stronger CTA than adults. Additionally, genetic analyses of CTA have shown that these strain patterns may also be changed by ethanol pre-exposure. Specifically, some evidence suggests that C57BL/6J mice may be more susceptible than DBA/2J mice to attenuation of CTA by ethanol-preexposure (Risinger and Cunningham, 1995). Thus, it is important to consider how strain background may interact with other factors, such as age and ethanol preexposure, to determine CTA phenotypes.

Across all CPP, CPA, and CTA studies, the DBA/2J strain in particular shows contrasting phenotypes, with strong conditioned preference and aversion to ethanol. Unique strain profiles suggest that ethanol CPP and CPA phenotypes are driven by distinct genetic mechanisms, which is further supported by the lack of correlation between CPP and CPA strain means (Cunningham, 2019). Various factors, such as ethanol pre-exposure and adolescent age, may interact with strain differences to influence ethanol CTA (Gremel and Cunningham, 2007; Risinger and Cunningham, 1995), and likely CPP and CPA.

2.4. Initial sensitivity: sedative (LORR), hypothermic, and ataxic responses to ethanol exposure

Initial sensitivity to ethanol, or the severity of acute ethanol effects, is thought to contribute importantly to the development of AUD. In humans, a low level of response to alcohol can predict alcoholism later in life (Heilig et al., 2010; Schuckit, 1994), and alcohol sensitivity is influenced by genetic background (Schuckit et al., 2001). For decades, rodent studies have been used to further examine the relationship between genetic background and ethanol sensitivity. Noteworthy early studies in a genetically heterogeneous mouse population examined genetic contributions by selectively mating subjects with similar high or low ethanol sensitivity (McClearn, 1981). Subjects were mated together after demonstrating comparable LORR (referred to as “Sleep”) duration after ethanol exposure, meaning that they would not turn over from their backs in similar amounts of time post-administration (e.g. latency to reacquire balance). Over time, this selective breeding exaggerated phenotypic differences and generated Long-Sleep and Short-Sleep groups with drastically different ethanol sensitivity phenotypes (Markel et al., 1996; Phillips and Dudek, 1991). These studies were important early demonstrations of genetic effects on initial sensitivity to ethanol.

Studies of inbred mice have been used to further understand genetic contributions to LORR after ethanol exposure. A 1983 study (Crabbe, 1983; male; 4 g/kg i.p., 20 % v/v EtOH) examined ethanol-induced LORR duration in a panel of 20 inbred mouse strains. Strain means for LORR duration varied from an average of 32.8 min (A/HeN) at the shortest and 140.4 min (129/J) at the longest. Strains with the shortest duration were A/HeN, C3H/HeNMTV, and C58/J, and strains with the longest duration were 129/J, PL/J, and DBA/1J. Later studies have found that LORR after ethanol varies not only across genetic (i.e., strain) background, but also by interactions between strain background and age. Specifically, comparisons of LORR after ethanol treatment in adolescent and adult C57BL/6J and DBA/2J mice show that there are different ontogenetic patterns of sensitivity across strains (Linsenbardt et al., 2009; male and female; 4 g/kg i.p.). Adult DBA/2J mice had longer LORR duration than adult C57BL/6J mice. Adolescents of both strains had a shorter LORR duration than their adult counterparts, suggesting that decreased sensitivity to this effect during adolescence may be conserved across genetic backgrounds. However, BECs were compared at LORR recovery to assess the role of pharmacokinetic differences in ethanol phenotypes, and this revealed a more complex relationship between age and genetic background. Adolescent DBA/2J mice had higher BECs than DBA/2J adults at the time of LORR recovery. This suggests that adult DBA/2J subjects are more sensitive to high BECs than adolescent DBA/2J subjects because adolescent subjects were able to recover their righting reflex with significantly higher BECs. This effect was only seen in DBA/2J, and not C57BL/6J, mice, demonstrating that age-dependent ethanol-induced LORR phenotypes can be dependent upon genetic background.

Hypothermic responses to ethanol vary by genetic background and appear to be related to ethanol’s depressant effects on the brain (Alkana et al., 1985). Hypothermia is associated with other ethanol sensitivity measurements, including LORR, ethanol-induced activity levels, and conditioned place preference and aversion (Alkana et al., 1985; Cunningham et al., 1991; Cunningham and Niehus, 1993). A study by Crabbe et al. (1982; male; 3 g/kg i.p., 10 % v/v EtOH) surveyed hypothermic responses to ethanol across 20 inbred mouse strains. They found substantial variation across strains in both initial measures of hypothermia and in the development of tolerance to ethanol-induced hypothermia over time. Strains that were particularly sensitive to ethanol (exhibiting the largest reductions in body temperature after ethanol) were CE/J, CBA/J, and A/HeN. The least sensitive strains included MA/MyJ, DBA/1J, and DBA/2N. In some strains, this predicted susceptibility to the development of tolerance to ethanol-induced hypothermia. CE/J, 129/J, and BALB/cAnN strains showed the strongest progressive tolerance, while PL/J, MA/MyJ, DBA/1J, and DBA/2J strains did not exhibit significant tolerance to repeated ethanol exposures. While adolescent hypothermia studies in mice are limited or nonexistent, studies in Sprague-Dawley rats suggest that adolescents may experience less severe or shorter duration of hypothermia after ethanol exposure (Ristuccia and Spear, 2004; Silveri and Spear, 2000). This would be consistent with a general tendency of adolescents to be less sensitive to many ethanol-related impairments (Spear, 2010), but it is not yet clear if adolescent insensitivity to hypothermia after ethanol is conserved across different genetic backgrounds in mice.

Alcohol intoxication in humans can lead to ataxia (i.e., the loss of full control of bodily movement), and this has been modeled in rodents. Studies have examined strain effects on ataxia after ethanol exposure using a variety of protocols. One study examined a panel of 19 strains for ataxia over the course of 1 h after an acute ethanol injection (Crabbe et al., 1982; male; 3 g/kg i.p.; 10 % v/v EtOH). Subjects were scored on a scale of 0 to 3, with 0 representing normal behavior, 1–2 representing increasing locomotor incoordination, and 3 representing LORR. Strains with the highest ataxia scores (most impaired) were 129/J, C57BL/10N, AKR/J, and A/HeN, and strains with the lowest scores (least impaired) were SWR/J, SJL/J, and C57BR/cdJ. Another study surveyed 20 inbred strains for performance in dowel (balance) and grid (ambulatory ataxia) tests (Crabbe, 1983; male; 2–3 g/kg i.p., 10–20 % v/v EtOH). In the dowel test, immediately after an ethanol injection (3 g/kg, i.p.; 10 % v/v EtOH), subjects were monitored for how long they could balance on a dowel before falling. Strains with the shortest latency to fall (most impaired) were DBA/1J, C57BL/6J, A/HeN, and BALB/cAnN, and strains with the longest latency to fall (least impaired) were C57BR/cdJ, SWR/J, and C58/J. In the grid test, subjects walk across a grid starting 10 min after an ethanol injection (2 g/kg, i.p.; 20 % v/v). The number of footslips through the grid is used as an index of ataxia. In this assay, strains with the most footslips (most impaired) were C57BL/10 N, MA/MyJ, and C57BL/6J, and strains with the fewest footslips (least impaired) were CE/J, CBA/J, and C57BR/cdJ. Later, Crabbe et al. (2003a) (male and female; 1–2.5 g/kg i.p.) also examined how small changes in experimental variables (such as drug dose, time between drug treatment and testing, balance beam width) influenced their observed strain differences in measures of motor incoordination after ethanol (balance beam and grid test). They found that small changes in these variables created strain-specific changes in observed ethanol sensitivity. The DBA/2J strain, for example, performed poorly on the narrowest balance beam after the highest ethanol dose, but it appeared to be the least impaired strain on the widest balance beam. This study confirmed that strain differences in these assays may be sensitive to small experimental manipulations and recommended that future studies use specific doses and a balance beam width to optimize generalizability and replicability of strain differences in these tasks. A separate study examined balance beam performance after ethanol across adult and adolescent C57BL/6J and DBA/2J subjects (Linsenbardt et al., 2009; male and female; 1.5 and 1.75 g/kg i.p.). While adolescents had higher footslip counts than adults at the higher dose, suggesting that adolescents were more sensitive to ethanol’s effects, there was little evidence of strain-dependent age effects in this test.

The studies described in this section provide examples of how LORR, hypothermia, and ataxia may vary significantly by strain after ethanol exposure. These differences may be due to differences in mechanisms of action at the level of neural circuitry and/or in phenotypic outcomes. Notably, the A/HeN is one of the least sensitive strains for LORR and one of the most sensitive strains for hypothermia (Crabbe, 1983; Crabbe et al., 1982). DBA/1J is very sensitive to LORR effects, while relatively insensitive to hypothermia effects. The C57BL/6J strain is consistently one of the most impaired strains on both the balance beam and grid tests of motor incoordination. For both LORR and hypothermia, adolescents are often less sensitive than adults (Linsenbardt et al., 2009; Ristuccia and Spear, 2004; Silveri and Spear, 2000). The opposite may be true for ataxic effects of ethanol, and adolescents may be more sensitive to these effects than adults. However, strain differences in these measures may be very sensitive to small changes in experimental variables, so it is important to be careful in interpretation of these effects (Crabbe et al., 2003b).

2.5. Locomotor effects

Locomotor effects of ethanol have been associated with mesolimbic dopamine responses (Meyer et al., 2009), making ethanol-induced locomotor activity assessments a simple and valuable model of dopamine-mediated alcohol effects in mice. Sensitization to this effect, where the same dose of ethanol can produce an enhanced locomotor response after repeated exposures, can also be used to study lasting neural adaptations after ethanol exposure.

An early study by Randall et al. (1975; male; 0.75–2.25 g/kg i.p) established strain differences in ethanol-induced locomotor activity between C57BL/6J and BALB/cJ mice. After three days of 30-min arena acclimation periods, subjects were treated with saline or ethanol and locomotor activity was observed for 1 h. The same dose range produced opposite effects in the two strains. As dose increased, activity increased in the BALB/cJ strain and decreased in the C57BL/6J strain. The opposite locomotor effects of the same alcohol doses across these two strains established the importance of genetic background in determining phenotypic outcome. This is in contrast to phenotypes like ethanol CPP, where C57BL/6J and BALB/cJ strains show similar ethanol responses (weak conditioned preference; Cunningham, 2014).

Studies examining more diverse strain panels have expanded upon these initial findings. Crabbe (1983; male; 1–2 g/kg i.p., 10–20 % v/v EtOH) surveyed locomotor increases and decreases after ethanol across 20 inbred mouse strains. Subjects were treated with a higher dose of ethanol (2 g/kg, i.p.) to evoke a decrease in locomotor activity among most strains. The C57BL/10N strain had a notably large decrease in locomotor activity at this dose (about twice as much as other high-responding strains). C57BL/6N, MA/MyJ, SEX/1ReJ, and C57BR/cdJ strains also had particularly large decreases in locomotor activity, while BUB/BnJ, C58/J, PL/J, BALB/cAnN, CE/J, AKR/J, and C3H/HeNMTV strains showed increases in locomotor activity at this dose. The lower dose (1 g/kg, i.p.) produced the largest increase in locomotor activity in C57BR/cdJ and C58/J strains. SJL/J and DBA/1J strains experienced locomotor depression at this dose. Notably, Crabbe et al. (1983) found that strain means for basal activity did not correlate well with ethanol-induced activity changes, suggesting that different genetic and neural substrates may influence basal activity and ethanol-induced activity changes. However, high basal activity levels were associated with fewer errors in a grid ambulatory ataxia test, suggesting that genes influencing basal activity may be involved in other ethanol-induced motor effects.

Like other ethanol phenotypes, initial ethanol-induced locomotor activity and sensitization to locomotor effects after repeated ethanol exposures may differ between adolescents and adults. In a study by Stevenson et al. (2008; male; 0–3 g/kg i.p.; 20 % v/v EtOH) adolescent DBA/2J mice were more sensitive to acute ethanol-induced locomotor activity increases but less sensitive to locomotor sensitization when compared to adult DBA/2J mice (Stevenson et al., 2008). This could indicate that adolescents experience greater dopamine-related immediate effects of ethanol, such as reward, but they may be less susceptible than adults to some longer-lasting neural changes as indicated by their decreased sensitization response. This is consistent with genetic studies that indicate different mechanisms mediating acute locomotor responses and sensitization to ethanol. For example, a QTL mapping study associated acute locomotor sensitivity and ethanol sensitization to separate loci, with only one commonly associated region (Phillips et al., 1995). However, it remains to be determined if this phenotype generalizes across diverse populations, and so future research could assess this developmental pattern of sensitivity across other strains.

Strains with notable locomotor responses to ethanol include C57BL/10N, which has shown extreme locomotor depression compared to other strains, and C57BR/cdJ and C58/J strains, which show relatively high locomotor activity after a low dose of ethanol (Crabbe, 1983). It seems that adolescents and adults differ in locomotor responses to ethanol, but age effects on acute locomotor stimulation and sensitization may be unique and warrant further study (Stevenson et al., 2008). In summary, these findings suggest that strain and age alter locomotor responses to ethanol, implicating an important role for genetic background and age in mediating dopamine-related responses to ethanol.

2.6. Withdrawal severity

Cessation of alcohol use after a long duration of chronic or intermittent alcohol exposure can lead to the development of withdrawal symptoms. The intensity of withdrawal symptoms is variable between individuals and is related to the likelihood of relapse to alcohol use during a quit attempt (for review on withdrawal and relapse, see Becker, 2008). The more unpleasant withdrawal symptoms and cravings after alcohol cessation, the more difficult it is to continue abstaining from alcohol. For this reason, understanding the genetic factors influencing alcohol withdrawal severity can improve understanding of the genetic factors influencing AUD more broadly.

In humans and mice, severe alcohol withdrawal symptoms can include seizures and convulsions (Hoffman and Tabakoff, 1994). In mice, the intensity of handling-induced convulsions (HIC) after cessation of ethanol exposure is often used to study the severity of ethanol withdrawal (Crabbe et al., 1983; Metten et al., 2010; Metten and Crabbe, 2005). A scale can be used to represent HIC severity, with larger numbers representing the most severe convulsions (tonic-clonic, quick onset, long duration, occurring spontaneously or from mild stimulus) and 0 representing no convulsions. A 1983 study (Crabbe et al., 1983; male; EtOH vapor exposure) and follow-up 2005 study (Metten and Crabbe, 2005; male; EtOH vapor exposure) both examined HIC during withdrawal across inbred mouse strains. To produce dependence, mice were given ethanol vapor over 3 days, and withdrawal testing occurred 0–25 h after cessation of treatment. Strains varied in latency to peak HIC score and in the value of peak HIC score. Across both studies, AKR/J, C58/J, C57BL/6J, PL/J, and BALB/cByJ strains generally exhibited the least severe HIC response, and DBA/2N, SWR/J, C3H/HeNMTV-, DBA/2J, and C3H/HeJ strains generally had the most severe HIC during ethanol withdrawal. Later studies have expanded on these findings by diversifying patterns of ethanol exposure used to produce dependence. Metten et al. (2010) examined HIC during ethanol withdrawal after an intermittent ethanol exposure. Subjects were exposed to ethanol vapor over 3 days, consistent with their previous study (Metten and Crabbe, 2005), but subjects were removed for 8-h breaks from ethanol each day. This intermittent pattern was chosen to better represent non-continuous patterns of alcohol use and sensitization to withdrawal reported in humans (Booth and Blow, 1993). Strain patterns of withdrawal severity were somewhat different after the intermittent paradigm when compared to previous chronic exposure studies. Strains with less severe HIC indices were C3H/HeJ, C57BL/6J, CBA/J, and SJL/J, and strains with highest HIC indices were DBA/1J, BTBR T + tf/J, CE/J, and C57BR/cdJ. These studies importantly demonstrate that ethanol withdrawal severity is dependent upon genetic background and on the pattern of previous ethanol exposures.

Other studies have conducted smaller strain comparisons examining other ethanol withdrawal phenotypes. One study examined affect-related ethanol withdrawal symptoms across C57BL/6 substrains, C57BL/6J and C57BL/6NJ (Hartmann et al., 2020; male and female; EtOH vapor exposure). Mice of both substrains were exposed to seven days of intermittent ethanol vapor, and behavior was assessed after acute withdrawal and long-term abstinence. Experimenters examined sucrose preference and forced swim test performance for anhedonia or depressive-like behavior, light-dark box performance for anxiety-like behavior, and HIC severity. Generally, there was not much evidence of C57BL/6J and C57BL/6NJ substrain differences in HIC severity or affective (depressive- and anxiety-like) behaviors due to abstinence from ethanol vapor. Another study compared acute ethanol withdrawal-associated fear conditioning deficits across C57BL/6J and DBA/2J strains (Tipps et al., 2015; male; acute withdrawal from 4 g/kg i.p., 20 % v/v EtOH). Subjects were trained in delay or trace fear conditioning procedures. Both procedures train subjects to associate an auditory cue and a context with an unconditioned stimulus (US), or footshock, but delay conditioning involves a co-terminating cue and US, while trace conditioning involves a short temporal separation (trace interval) between cue and US presentation. Acute ethanol withdrawal generally impaired performance in contextual learning tests and enhanced performance in cued learning tests after delay and trace training paradigms. The only exception to this effect was that the C57BL/6J strain did not show an enhancement in cued test performance after delay training procedures. Specifically, in the C57BL/6J strain, cued learning outcomes were comparable between saline-and ethanol-treated groups after delay training. These results demonstrate that cognitive deficits during ethanol withdrawal vary by genetic background. However, collectively, studies by Hartmann et al. (2020) and Tipps et al. (2015) suggest that effects of strain may vary between withdrawal phenotypes.

While age differences in ethanol withdrawal phenotypes have not been studied thoroughly in inbred mice, they are reported to be different between adolescent and adult rats (Doremus et al., 2003; Varlinskaya and Spear, 2004). Generally, adolescents are less sensitive to negative effects of ethanol withdrawal than adults. For example, a study in Sprague-Dawley rats found that during acute ethanol withdrawal (18 h after treatment), adolescent subjects do not show anxiety-like behaviors in an elevated plus maze like adult subjects (Doremus et al., 2003; male; 4 g/kg i.p., 20 % v/v EtOH). Additionally, adolescent Sprague-Dawley rats experience less severe social suppression during acute ethanol withdrawal (18 h after 4 g/kg, i.p.) than adults (Varlinskaya and Spear, 2004; male and female; 4 g/kg i.p., 20 % v/v EtOH). Instead, during recovery from acute withdrawal, they actually experience social facilitation not seen in adults. Adolescent insensitivity to negative effects of alcohol may increase the likelihood of adolescent alcohol use (Spear, 2010). Thus, it will be important to examine whether these age-dependent sensitivities are conserved in other genetically heterogeneous populations.

Across studies, DBA strains (DBA/1J, DBA/2J, DBA/2N) usually show severe ethanol withdrawal-related phenotypes (Crabbe et al., 1983; Metten et al., 2010; Metten and Crabbe, 2005; Tipps et al., 2015). The C57BL/6J strain generally shows less severe ethanol withdrawal phenotypes (Crabbe et al., 1983; Metten et al., 2010; Metten and Crabbe, 2005; Tipps et al., 2015). Studies considering adolescent age effects generally find that adolescents are less susceptible to negative symptoms of ethanol withdrawal (Doremus et al., 2003; Varlinskaya and Spear, 2004), but more research could be done to understand whether this varies by genotype and if there are exceptions to this trend.

2.7. Ethanol metabolism

It is important to note that small differences in ethanol metabolism across strains can be responsible for large differences in behavioral responses to ethanol. Studies examining ethanol-related behaviors, such as voluntary ethanol consumption, often include some kind of pharmacokinetic analysis to address the potential confound of metabolic differences.

A study by Crabbe et al. (2003a) (male and female; 1–2.5 g/kg i.p., 20 % v/v EtOH) measured BECs of 21 strains at various timepoints 30–150 min after ethanol treatment. Some of these data were used to calculate Widmark’s β, which represents the slope of the blood ethanol elimination curve (listed as mg/ml/h). Strains with the steeper slope, or faster ethanol elimination, were PERA/Ei, SPRET/Ei, SM/J, DBA/2J, and FVB/NJ. Strains with the least steep slopes, or slower ethanol elimination, were BALB/cByJ, C57L/J, PL/J, SJL/J, and 129S1/SvlmJ. Additionally, males generally achieved higher BECs than females, suggesting a robust sex differences across genetic backgrounds.

A notable study measured ethanol consumption and pharmacokinetics across Collaborative Cross (CC) recombinant inbred strains, the Diversity Outbred (DO) population, and the 8 founder strains from which they are derived (Bagley et al., 2021; male and female; 1.5 g/kg, i. p.; 20 % w/v EtOH). Different subjects were used to study ethanol consumption and metabolism. The pharmacokinetics group was treated with ethanol (1.5 g/kg, i.p.) and BECs were measured from samples collected 30 min after treatment. The study found that BEC measures were heritable. Of the eight inbred founder strains included in the study, NZO/HILtJ, A/J, and 129S1/SvlmJ had the highest BECs, and WSB/EiJ, CAST/EiJ, and NOD/ShiLtJ had the lowest BECs. While both ethanol metabolism and consumption were heritable, they did not find an association between ethanol metabolism and consumption measures.

While few studies have assessed interactions of age and genetic background in determining ethanol metabolism, adolescent mice may have faster ethanol absorption and/or metabolism than adult mice. A study examining ethanol metabolism across adult and adolescent C57BL/6J and DBA/2J mice found that adolescents of both strains exhibited lower BECs (suggesting faster metabolism) across multiple timepoints after ethanol treatment (Linsenbardt et al., 2009; male and female; 4 g/kg i.p.). Thus, it is likely that ethanol is metabolized faster in adolescents of diverse genetic backgrounds.

Overall, strain variation in ethanol metabolism is an important variable to consider in all studies of strain differences in ethanol responses. Strain and age differences in ethanol metabolism have been reported consistently (Bagley et al., 2021; Crabbe et al., 2003a; Linsenbardt et al., 2009), although they are not always related to strain differences in ethanol responses (Bagley et al., 2021). Determining whether ethanol metabolism is responsible for group differences is important in pinpointing the genetic factors involved in alcohol abuse and AUD.

2.8. Summary

In studies examining alcohol phenotypes across inbred strains, patterns emerge for differences in phenotypes, but those patterns are not entirely consistent. Phenotypic differences between C57BL/6J, C58/J, DBA/2J, and 129S1 mice will be used to highlight the complexity of behavioral responses. C57BL/6J and C58/J mice are high consumers of alcohol whereas DBA/2J and 129S1 mice are low consumers of alcohol. This could suggest that C57BL/6J and C58/J mice will show similar patterns of alcohol-related endophenotypes, as would DBA/2J and 129S1 mice, if endophenotypes assess related factors that contribute to alcohol-related disorders. However, this is not always the case. C57BL/6J mice show low sensitivity to the sedative effects of alcohol as measured by LORR but show high levels of sensitivity to alcohol-induced ataxia and decreased locomotor effects. Similarly, C58/J mice show low sensitivity to the sedative effects of alcohol as measured by LORR. However, in contrast to C57BL/6J mice, they also show low sensitivity to alcohol-induced ataxia while showing alcohol-induced locomotor stimulatory effects. This suggests that the ataxia and locomotor effects associated with alcohol involve different genetic substrates than the self-administration effects and the sedative effects. Finally, C57BL/6J and C58/J mice both show weak CPP, CPA, and CTA, which is in contrast to both strains showing higher levels of alcohol self-administration. This suggests that the genetics of aversion and preference and the genetics of self-administration for alcohol are either different or share similar substrates that regulate the phenotypes in different directions. That is, high-self administration does not mean high preference or aversion.

Both DBA/2J and 129S1 mice show low alcohol consumption but show high CPA and CTA while the DBA/2J mice also showed high CPP. This is the mirror image of what was seen in C57BL/6J and C58/J mice. However, the 129S1 mice are similar to C57BL/6J mice in the effects of alcohol on ataxia and 129S1 and DBA/2J mice differ in alcohol metabolism. These differences restrict the ability to make simple associations, such as alcohol consumption is driven by the rate of alcohol metabolism, and instead demonstrate that the contributing factors are complex.

As discussed, one contributing factor is age at exposure. For alcohol self-administration, adolescent C57BL/6J mice consumed more than adult C57BL/6 mice. This fits a narrative in which adolescents consume more. However, this effect was not seen in DBA/2J mice, which suggests that an interaction of age and genetic background influences alcohol self-administration. Understanding the genetic substrates of this interaction could aid in identifying risk factors for developing an AUD. In contrast, studies suggested that adolescents may be less sensitive to some effects of alcohol but more sensitive to the impact of alcohol on motor function as measured by ataxia and motor activation. In general, adolescent mice were less sensitive to the sedative effects of alcohol and showed less CTA than adult mice, at least for the strains examined thus far. This profile fits with the idea that alcohol use in adolescents may be impacted by the fewer negative effects of alcohol consumption and this can include withdrawal symptoms since adolescent mice did not show the same anxiety-like behaviors exhibited by adult mice during withdrawal from alcohol. Some of these differences may be due to differences in metabolism, as adolescent mice were faster metabolizers of alcohol than adult mice.

Genetic background significantly influenced all of the discussed categories of ethanol-related phenotypes. Many ethanol studies have included DBA/2J and DBA/1J strains, and these strains often show unique responses to ethanol, such as severe withdrawal (Metten et al., 2010; Metten and Crabbe, 2005) and strong conditioned preference and aversion (Cunningham, 2014, 2019) when compared to other commonly used strains, such as C57BL/6J. Comparisons of closely related inbred mouse populations, like the C57BL/6 substrains, have revealed that even minor genetic variations can be responsible for differences in alcohol phenotypes, like ethanol consumption, across substrains. Thus, comparisons across C57BL/6 substrains show promise for discovering and modeling genetic variants underlying ethanol sensitivity (Mulligan et al., 2008). Many of the reviewed studies reported that small alterations to experimental procedures such as altering the taste of ethanol solutions for voluntary consumption (Halladay et al., 2017; Yoneyama et al., 2008) or changing the width of balance beams used for motor coordination tests (Crabbe et al., 2003b) can alter patterns of strain differences in assays of ethanol-related behaviors. These studies reinforce the importance of considering and accounting for how minor experimental variables may influence measures of alcohol phenotypes and their genetic underpinnings. Additionally, rates of ethanol metabolism differ across strains and ages (Bagley et al., 2021; Crabbe et al., 2003a; Linsenbardt et al., 2009), and this must be considered when studying differences in ethanol phenotypes across populations. While females were not included in all studies, sex differences were reported for some phenotypes, such as ethanol metabolism. Finally, for some phenotypes such as voluntary consumption and LORR, strain and age can interact to produce unique developmental patterns of behavior.

3. Nicotine phenotypes

3.1. Background

A large body of research has evaluated the impact of nicotine exposure on adolescent development and adult function. This is an important area of research because not every individual that is exposed to nicotine becomes dependent or addicted, suggesting that a variety of individual factors shape susceptibility to nicotine’s effects. Environment is likely to be one such factor as well as genetics that contributes to individual differences. Further, the role of genetics in individual differences and susceptibility to the maladaptive effects of adolescent nicotine exposure specifically is understudied. Identifying genetic factors that contribute to individual risk or protective factors for the maladaptive effects of nicotine exposure, particularly in adolescence, will aid in developing interventions for at risk individuals and groups as well as treatments. The following sections will examine strain differences (and when available, adolescent data) in nicotine consumption, sensitivity, and associated changes in locomotion, reward and aversion, withdrawal phenotypes, and metabolism. Unless indicated, nicotine doses are reported as a freebase dose.

3.2. Voluntary consumption and self-administration

Drug self-administration is often considered a gold standard in assessing addiction-like phenotypes in animal models. As discussed previously, voluntary consumption and self-administration studies can be used to study the reinforcing properties of a drug and motivation for the drug (Panlilio and Goldberg, 2007). As with any behavior, self-administration is a complex behavior and rates of self-administration can be influenced by factors other than reward. One such factor is the aversive properties of the drug. Nicotine is a paradox for self-administration. Considered to be one of the most addictive drugs, self-administration in rodent models has been difficult to establish compared to other drugs of abuse (Matta et al., 2007). This has resulted in the employment of different self-administration paradigms such as oral self-administration and intravenous self-administration. Both approaches have been used in genetic analysis.

In a dose-response comparison of oral nicotine consumption, male and female C57BL/6 mice consumed more nicotine than DBA/2 mice, with female C57BL/6 mice showing a stronger preference than males (McGlacken et al., 1995; male and female; 1–100 μg/ml nicotine (dose not stated as freebase); substrain designations not listed). This study also examined ethanol, amphetamine, and aspartame and found that for ethanol and amphetamine, C57BL/6 mice showed greater preference but DBA/2 mice had greater preference than C57BL/6 mice for aspartame. This suggests that C57BL/6 mice might have increased sensitivity to the reinforcing effects of drugs while DBA/2 mice find sweetness more reinforcing. This also suggests that lower drug preference in DBA/2 mice may not be due to a dysfunctional reward system as they consume greater amounts of aspartame.

Similar results suggesting a propensity for oral nicotine self-administration in C57BL/6 mice were found in an extended inbred strain survey (Robinson et al., 1996; male; 10–200 μg/ml nicotine). Mice started with a 10 μg/ml nicotine concentration that was increased in a stepwise manner to 200 μg/ml. The vehicle was tap water for one experiment and 0.2 % saccharin for a second experiment. Regardless of vehicle, C57BL/6 mice consumed the most followed by DBA/2, BUB, A and then similar levels for C3H and ST/b mice. Of note, differences between the C57BL/6 mice and the other mice were greatest at the higher end of the concentration curve. This suggests that C57BL/6 mice may be less sensitive to the aversive effects of higher doses of nicotine.

Using a two-bottle choice paradigm, tap water versus 5 μg/ml nicotine in tap water, one study assessed preference for nicotine in six inbred mouse strains (Aschhoff et al., 2000; male and female; 5 μg/ml nicotine (dose not stated as freebase); strains without the “J” designation were from Charles River). Comparing total nicotine consumed to total tap water consumed, C57BL/6 mice consumed the most nicotine followed by C3H, A/JXNMRI, DBA/2, NMRI, ST/bJ strains in that order. To assess if variation in oral nicotine consumption related to aversion to the taste, A/J, C57BL/6J, and DBA/2J mice were compared in a two-bottle choice paradigm of 75 μg/ml nicotine versus water and also tested for brief consumption of a range of nicotine doses, 5–400 μg/ml nicotine (Glatt et al., 2009; male and female; 5–400 μg/ml nicotine). Similar to the prior study, C57BL/6 mice consumed the most nicotine (75 μg/ml) but the three strains did not differ in aversion to nicotine across the doses tested (5–400 μg/ml). These results further suggest that C57BL/6 mice may be a good model for studying the reinforcing effects of nicotine. This study also found that for the C57BL/6 strain only, females consumed more nicotine than males.

Genetic differences associated with nicotine metabolism may contribute to vulnerability for greater nicotine intake. To examine if differences in CYP2A5, the main enzyme responsible to nicotine metabolism in mice and the homologue to CYP2A6 in humans, contribute to variability in oral nicotine self-administration, F2 C57BL/6XST/bJ mice were tested in two bottle choice (Siu et al., 2006; male and female; 5–100 μg/ml nicotine dose was not stated as freebase). Both the Robinson et al. (1996) and the Aschhoff et al. (2000) studies found that C57BL/6 mice had high levels of oral nicotine consumption while ST/bJ mice had low levels. Using an escalating dose response from 5 to 100 μg/ml of nicotine, Siu et al. (2006) found that high nicotine consumers had higher levels of CYP2A5 in male F2 mice. In female F2 mice, nicotine metabolism did not vary across higher and lower consumers. This suggests that genetic factors contributing to nicotine intake vary by sex and for males, greater levels of a nicotine metabolism-associated enzyme may increase nicotine consumption.

In a study examining the strain differences and nicotinic acetylcholine receptor (nAChR) involvement in oral nicotine choice self-administration, Bagdas et al. (2019; male and female; 10–960 μg/ml nicotine) examined common inbred strains and strains with modified nAChR subunit-encoding genes and found sex and strain differences. Overall, females consumed more nicotine and showed greater preference for nicotine than male mice. C57BL/6J mice had greater intake and preference for oral nicotine than DBA/2J mice. Both nAChR subunit β2 knockout (KO) mice and α6 KO mice had decreased oral nicotine consumptions and preference. Knocking out the α7 nAChR subunit had no effect on oral nicotine preference or consumption but α5 KO mice showed increased nicotine preference and consumption. This study further suggests that female mice may have a higher propensity to self-administer nicotine and C57BL/6 mice have higher preference for and consumption of nicotine than DBA/2J mice. Variants in nAChR subunit function and numbers could contribute to nicotine preference and self-administration as β2 and α6 subunits may support nicotine preference and self-administration whereas the α5 subunit may suppress preference and self-administration.

Single nucleotide polymorphisms (SNP) in genes encoding nAChRs could be another contributing factor to differences in nicotine self-administration. A SNP in Chrna4, the gene encoding the α4 nAChR subunit, results in an alanine/threonine substitution (Butt et al., 2005). Given the involvement of α4β2 nAChRs in nicotine addiction, Butt et al. (2005; male; 25–100 μg/ml nicotine) investigated if inbred mice differing in expression of the A529 and T529 variants of α4 would segregate by SNP for oral nicotine consumption. Fourteen lines of mice were tested, and strains with the A529 genotype consumed significantly less oral nicotine. C57BL/6J mice, T529 genotype, consumed the most nicotine across the three doses tested whereas C3H/Ibg mice, A529 genotype, consumed the least amount of nicotine at the two higher doses. This contrasts with a previous study (Aschhoff et al., 2000), where C3H mice were found to consume some of the largest volumes of nicotine. This study replicated past findings in C57BL/6 mice and provided evidence that differences in Chrna4 may be a contributing factor to differences in nicotine self-administration.

Using the differences in free-choice oral nicotine consumption between C57BL/6 and C3H mice, Li et al. (2007; male and female; 25–100 μg/ml nicotine) used quantitative trait loci (QTL) mapping in mice from an F2 cross of C57BL/6J and C3H/HeJ mice to identify chromosomal regions associated with oral nicotine consumption. The study identified four significant QTL, located on chromosomes 1, 4, 7, and 15. The largest logarithm of odds (LOD) score for nicotine consumption was located on chromosome 1, around 96 cM. Mouse chromosome 1 is syntenic with human chromosome 1, with the 96 cM region in mice related to the 169 cM area in humans. A study in humans found that the QTL with the greatest association with nicotine dependence was on chromosome 1 around 168 cM (Wang et al., 2005). The study by Li et al. (2007) further supports that genetic differences between C57BL/6 and C3H mice contribute to observed differences in nicotine consumption and that the genetic substrates identified in mice may translate to genetic substrates associated with nicotine addiction in humans.

Strain survey of age-dependent effects of adolescent nicotine self-administration is an area ripe for investigating the genetics underlying adolescent sensitivity to nicotine use. While the full genetics of adolescent nicotine self-administration remain to be identified, a study in outbred CD-1 mice examined oral nicotine self-administration differences at early (PND 24–25), middle (PND 37–48), and late (PND 50–61) adolescence (Adriani et al., 2002; male and female; 0–30 mg/l nicotine). Early adolescent mice showed the strongest preference for the nicotine solution while later adolescent mice trended toward avoidance. While not in inbred mice, this study suggests that genetic heterogeneity of this phenotype exists in mice. Further, this finding supports studies in humans that suggest the earlier the onset of nicotine use, the greater the likelihood of developing nicotine dependence (Kendler et al., 2013; Lanza and Vasilenko, 2015).

As mentioned previously, intravenous nicotine self-administration is difficult to achieve, though possible (Matta et al., 2007). A study examining intravenous nicotine self-administration in an F2 cross between C57BL/6J and C3H/J mice found no difference in intravenous nicotine self-administration for both male and female mice (Bilkei-Gorzo et al., 2008; male and female; 0.75 μg nicotine). However, greater reinstatement of intravenous nicotine self-administration after extinction of intravenous nicotine self-administration was reported in mice that showed high stress responsivity similar to C57BL/6J while mice with stress responsivity similar to C3H/J had lower levels of reinstatement. This study indirectly supports prior work suggesting differences in nicotine administration between C57BL/6J and C3H/J mice by demonstrating differences in reinstatement of nicotine self-administration while also highlighting the critical role of stress in nicotine addiction phenotypes.

Overall, nicotine self-administration studies in panels of inbred mice suggest that C57BL/6 mice show strong preference for nicotine self-administration while C3H mice have low levels of nicotine self-administration. Genetics contribute to these differences as associated QTL have been identified. Genetic factors associated with metabolic differences, nAChR subunit differences, and differences in stress reactivity may also be causal agents contributing to differences in nicotine intake.

3.3. Conditioned preference and aversion

While self-administration studies are important for understanding the reinforcing effects of nicotine, CPP can be used to study nicotine reward (Nader, 2016). To review, nicotine (or any drug) administration is associated with the context of one chamber and vehicle is associated with a second distinct chamber. Subjects are later tested for which chamber they prefer. Preference for the drug paired chamber is interpreted that the subjects found the drug rewarding and may be engaging in drug seeking behavior. Multiple studies have examined nicotine CPP in panels of inbred mice.

In one study, nicotine CPP was examined in four strains of inbred mice (Ise et al., 2014; male; 0.4, 0.8, 1.26 mg/kg dose not stated as freebase). Dose-dependent CPP was seen for DBA/2N and BALB/c mice with significant CPP seen for the two highest doses. C3H/HeN mice showed an inverted U-dose response curve with CPP for the 0.8 mg/kg dose and conditioned place aversion (CPA) for the 1.26 mg/kg dose. In contrast, C57BL/6N mice only showed CPA and this was at the highest dose. This result is counter to the self-administration results where C57BL/6 mice self-administered the most nicotine. This might be due to differences in the substrain of C57BL/6 mice used as C57BL/6J and C57BL/6N have phenotypic and genetic differences associated with differences in SNP variants, and insertions and deletions (Simon et al., 2013).

A study of CPP in eight strains of inbred mice reported a different pattern of CPP and CPA across inbred strains than the Ise et al. (2014) study (Kutlu et al., 2015; male; 0.35 mg/kg nicotine). Significant CPP was seen for C57BL/6J, CBA/J, and 129/SvEv mice. In contrast, A/J, BALB/cByJ, C3H/HeJ, and DBA/2J mice did not show CPP and DBA/1J mice showed CPA. The finding from this study that C57BL/6 found nicotine rewarding matches well with nicotine self-administration studies that report high levels of nicotine self-administration by male mice of this strain (Yan et al., 2012). However, the results of this study are in contrast to the results of Ise et al. (2014). Ise et al. (2014) found CPP in DBA/2N and BALB/c, which was not seen in the Kutlu et al. (2015) study, and CPA for C57BL/6 mice, the opposite of the Kutlu et al. (2015) study. The studies differed in mouse vendors and thus it is possible that genetic drift between the two vendors contributed to differences. Another possible difference resulted from the doses tested, though if Ise et al. (2014) reported doses in free base weight, the lowest dose 0.4 mg/kg is similar to the 0.35 mg/kg dose used by Kutlu et al. (2015) study and if Ise and colleagues reported doses in nicotine tartrate salt weight, the highest dose would convert to a dose of 0.41 mg/kg, which again would be close to 0.35 mg/kg.

In a study comparing CPP at multiple nicotine doses between C57BL/6J and DBA2/J mice, Jackson et al. (2009; male; 0.1, 0.3, 0.5, 0.7, 1, or 1.5 mg/kg nicotine) found that C57BL/6J, but not DBA2/J, mice developed CPP for nicotine. Of the six doses tested, C57BL/6J mice showed CPP for nicotine at 0.3, 0.5, and 0.7 mg/kg doses. These results are matched by the results from Kutlu et al. (2015) but counter to results from Ise et al. (2014). While there is not a clear reason for the discrepancy of results, the findings from Jackson et al. (2009) and Kutlu et al. (2015) align with prior studies reviewed that suggest that C57BL/6 mice are more sensitive to the effects of nicotine in nicotine-addiction associated phenotypes.

Overall, the literature demonstrates a clear influence of genetics on nicotine-induced CPP. This effect, however, might be influenced by substrain or vendor differences as studies that used C57BL/6J versus C57BL/6N mice had divergent results for nicotine CPP. Results largely suggest the C57BL/6J mice have a stronger preference for nicotine than DBA2/J mice. This preference data aligns with self-administration data indicating that C57BL/6 mice consume more nicotine, which indicates that this line of mouse may model individuals that are more likely to consume nicotine.

3.4. Initial sensitivity

The acute, or initial effects of drugs of abuse may contribute to susceptibility to developing addiction (Gould, 2010). Greater initial effects of drugs of abuse could contribute to repeated use, increasing the likelihood of developing addiction. In addition, an ability of early drug use to facilitate maladaptive learning could lead to the formation of drug-context and drug-stimuli associations that sustain drug seeking and addiction (Hyman, 2005). In a study examining the dose-dependent effects of acute nicotine administration in four strains of inbred mice, all strains displayed dose-dependent reduction of body temperature. Sensitivity to the effects of acute nicotine on body temperature varied by strain, with BALB/cBy mice most sensitive and C3H/lb mice the least sensitive (Marks et al., 1983; male and female; 0–2.0 mg/kg nicotine). No sex effects were detected. The effects of acute nicotine on acoustic startle were also examined and only C3H/lb mice showed enhanced startle after acute nicotine administration. This study demonstrated that genetic influence on the acute effects of nicotine varies by phenotype as C3H/lb. mice were most sensitive to the acute effects of nicotine on acoustic startle but least sensitive to the effects of acute nicotine on hypothermia.

C57BL/6J and DBA/2J mice were examined for dose-dependent effects of acute nicotine on locomotor activity, elevated plus maze (used to model anxiety), body temperature, and sensitivity to pain (tail-flick and hot plate) (Jackson et al., 2009; male and female; nicotine doses varied by task). C57BL/6J mice showed greater responses to acute nicotine for locomotor activity, antinociceptive effects of nicotine in the tail flick assay, and greater nicotine-induced hypothermia than DBA/2J. In the elevated plus maze, C57BL/6J displayed reduced anxiety-like behavior while DBA/2J mice showed increased anxiety-like behavior. These results show that for the measured responses, C57BL/6J mice are more sensitive to the acute effects of nicotine than DBA/2J mice. For elevated plus maze, the results suggest both strains have a large effect but in the opposite direction. C57BL/6J mice show anxiolytic effects of acute nicotine while DBA/2J mice show anxiogenic effects of acute nicotine. While the C57BL/6J strain was more sensitive to acute nicotine effects in many cases, this study showed that increased sensitivity of the C57BL/6J strain is phenotype-specific, with DBA/2J mice being more susceptible to anxiogenic effects.

A study examining the influence of genetics on contextual fear conditioning in eight strains of inbred mice found that the dose dependent effects of acute nicotine varied by genetic background (Portugal et al., 2012; male; 0.023–0.18 mg/kg nicotine). In all strains that acute nicotine had an effect, contextual fear conditioning was enhanced. This is in contrast to cued fear conditioning for which no dose of acute nicotine altered cued fear conditioning. Contextual and cued fear conditioning involve multiple brain regions but differ in dorsal hippocampus involvement; dorsal hippocampus is critically involved in contextual, but not cued, fear conditioning (Logue et al., 1997). Mice from the 129/SvEv, A/J, BALB/cByJ, C57BL/6J, DBA/1J, and DBA/2J strains all showed enhancement of contextual fear conditioning after treatment with acute nicotine. The effective dose varied by strain with only the C57BL/6J and DBA/2J mice sensitive to multiple doses. Fear conditioning in the C3H/HeJ and CBA/J strains was not altered by acute nicotine. These results demonstrate that genetic differences contribute to the effects of acute nicotine on hippocampus-dependent learning as assessed by contextual fear conditioning.

Common to studies by Jackson et al. (2009) and Portugal et al. (2012), BALB/cByJ and C57BL/6J mice were sensitive to the acute effects of nicotine on hypothermia and contextual fear conditioning. Future research may elaborate on age-specific differences in initial sensitivity to nicotine.

3.5. Locomotor activity

As described previously, the locomotor stimulatory effects of drugs of abuse are thought to be endophenotypes for addiction (Robinson and Berridge, 1993). This is because the locomotor stimulatory effects are mediated by increased dopamine release (Engelhard et al., 2019; Johnson et al., 1996) just as are the reinforcing effects of drugs of abuse (Volkow et al., 2019). Therefore, mouse inbred strain panel research that seeks to identify genetic substrates associated with locomotor stimulatory effects of nicotine may also inform the genetics of the reinforcing properties of nicotine and on nicotine addiction.

Assessing the effects of nicotine on locomotor activity in an open field arena, Marks et al. (1983) (male and female; 0–1.5 mg/kg) found that nicotine dose-dependently decreased locomotor activity for BALBcByJ, C57BL/6Ibg, and DBA/2Ibg mice but in contrast, low doses of nicotine increased activity in C3H/Ibg mice. No sex effects were found. Locomotor activity was assessed in a chamber that mice were habituated to that was also under dim lighting provided by red light. Thus, while open field measurements can reflect changes in locomotor activity and anxiety, the contribution of anxiety to the phenotype assessed under these conditions was likely reduced. If the locomotor stimulator effects of a drug predict abuse liability, these results could suggest C3H mice would have a higher propensity for nicotine addictive-like phenotypes such as nicotine self-administration. However, altered propensity to self-administer nicotine, as examined in the previous section, could complicate this interpretation for C3H mice.

Another study investigating genetic contributions to the stimulatory effects of nicotine examined mice from a A/J × C57BL/6J F2 cross and 21 B6.A chromosome substitution strains, where each strain contained a different A/J chromosome on a C57BL/6J background (Boyle and Gill, 2009; male and female; 0.28 mg/kg nicotine). Open field testing was performed in dim red light and mice were treated with one dose of nicotine. Two QTLs emerged. An allele on chromosome 11 was associated with lower nicotine-induced locomotor activity compared to homozygous C57BL/6J mice while an allele at chromosome 16 was associated with increased nicotine-induced locomotor activity. In the B6.A chromosome substitution strains, mice with A/J chromosomes 2, 14, 16, or 17 on a C57BL/6J background had greater nicotine-induced activity than homozygous C57BL/6J mice. These findings demonstrate genetic influence on the psychostimulatory effects of nicotine and that either chromosomal region from A/J mice is associated with increased psychostimulatory effects of nicotine or that in the transfer of chromosome 2, 14, 16, or 17 to a C57BL/6J background, epistatic interactions occur leading to greater nicotine-induced locomotor activity.

Unlike the extensive strain surveys reporting the effects of acute ethanol on locomotor activity and repeated administration of ethanol on sensitization of the locomotor effect, few strains, as discussed above, have been evaluated for the acute effects of nicotine on locomotor activity and only two substrains of C57BL/6, C57BL/6J and C57BL/6N, have been evaluated with repeated administration of nicotine. Akinola et al. (2019) (male; 0.5 or 1.0 mg/kg nicotine, s.c.) tested adult C57BL/6J and C57BL/6N mice in an open field for 30 min after saline injections for 3 days followed by nicotine treatment for 5 days. Then after a 2-day washout period the mice were challenged again with nicotine in a single open field session. On the first day of nicotine administration, C57BL/6J mice showed decreased locomotor activity at 0.5 and 1.0 mg/kg dose while C57BL/6N mice only showed decreased locomotor activity at 1.0 mg/kg dose. The subsequent 4 days of nicotine administration produced similar levels of locomotor activity so there was no indication of sensitization to the repeated administration of nicotine in either substrain. Finally following 2-days without nicotine, 1.0 mg/kg nicotine still decreased locomotor activity in both substrains and 0.5 mg/kg only decreased locomotor activity in the C57BL/6J substrain but not the C57BL/6N substrain.

Although fewer strains of mice have been evaluated for locomotor response to nicotine, there is a clear genetic contribution to locomotor responses with nicotine dose dependently decreasing locomotor activity in BABLcBYJ, C57BL/6Ibg, DBA/2Ibg, and C57BL/6J mice while only high dose nicotine decreased locomotor activity in C57BL/6N. Only one strain, C3H/Ibg, showed increased locomotor activity and that was at low doses of nicotine.

3.6. Withdrawal

The hedonic effects of drugs of abuse may be the initial pathway into repeated use and eventual addiction, but the propensity to develop withdrawal effects may maintain addiction, drug seeking, and self-administration long after the hedonic effects have diminished (Wise and Koob, 2014). Nicotine withdrawal is characterized by changes in affect, cognition, and somatic signs. Strain surveys of inbred mice have examined all three groups of withdrawal symptoms.

Studies in mice have examined nicotine withdrawal in two different ways. One is cessation of chronic nicotine treatment while the other common method is treating a mouse that is receiving chronic nicotine with a nAChR antagonist. Cessation of chronic nicotine treatment to induce withdrawal models smoking and vaping cessation in humans. Withdrawal precipitated with a nAChR antagonist allows for examination of the contribution of specific nAChRs to nicotine withdrawal symptoms. A study in outbred ICR mice found that spontaneous nicotine withdrawal increased somatic signs, increased anxiety-like behavior in the elevated plus maze, and increased pain sensitivity (Damaj et al., 2003; male; 24 mg/kg/day nicotine for 14 days). In the same study, withdrawal was precipitated in ICR mice chronically treated with 24 mg/kg/day nicotine on day 15 using the general nAChR antagonist mecamylamine; dihydro-beta-erythroidine (DHβE), which has higher specificity for α4β2 nAChRs; and methyllycaconitine, which has higher specificity for α7 nAChRs. The results demonstrated that different nAChRs mediate different withdrawal symptoms with anxiety symptoms associated with α4β2 nAChRs and hyperalgesia associated with α7 nAChRs. The study also compared 2 mg/kg mecamylamine precipitated withdrawal in inbred C57BL/6J and 129/SvEv mice chronically treated with 24 mg/kg/day nicotine on day 15. Mecamylamine precipitated withdrawal was associated with increased somatic signs, increased anxiety-like behavior, and hyperalgesia in C57BL/6J mice whereas no significant changes were seen in 129/SvEv mice. Overall, C57BL/6J mice were more susceptible to nicotine withdrawal than 129/SvEv mice and results from the outbred ICR mice suggest that different nicotine withdrawal symptoms are mediated by different nAChR subtypes.

In another study examining both spontaneous and precipitated nicotine withdrawal, C57BL/6J and BALB/cByJ mice were treated chronically with nicotine (Stoker et al., 2008; sex not specified; 10–40 mg/kg/day nicotine for 14 or 28 days). No spontaneous nicotine withdrawal changes in light-dark box performance, prepulse inhibition of the acoustic startle reflex (PPI), acoustic startle, or somatic signs were seen after 14 days of treatment, which is counter to other studies. However, extending nicotine treatment to 28 days at 40 mg/kg/day in just the C57BL/6J strain resulted in spontaneous withdrawal symptoms in the light-dark box (anxiety), somatic signs, and inter cranial self-stimulations (ICSS: assess hedonia vs anhedonia). Mecamylamine and DHβE precipitated nicotine withdrawal was assessed in C57BL/6J mice treated chronically with 40 mg/kg/day nicotine for 14 days. Both mecamylamine and DHβE precipitated nicotine withdrawal increased the threshold for ICSS. This study demonstrated the importance of dose and treatment duration and that the method chosen to model nicotine withdrawal impacts outcomes.

A study compared the effects of withdrawal from nicotine administered for 14 days on somatic signs, nociceptive effects, anxiety-like behavior in the plus maze (affect), and locomotor activity in C57BL/6J and DBA/2J mice (Jackson et al., 2009; male; 36 mg/kg/day for 14 days). DBA/2J mice showed more somatic withdrawal signs than C57BL/6J mice. C57BL/6J mice showed withdrawal-associated increases in anxiety phenotype while DBA/2J mice did not show an associated withdrawal-related change in anxiety. Both strains demonstrated increased sensitivity to pain during nicotine withdrawal while only C57BL/6J mice showed increased locomotion during withdrawal. This study provides support for the contention that nicotine withdrawal is not a single phenotype but instead a constellation of phenotypes involving different neural and genetic substrates. The results also suggest that C57BL/6J mice may be more sensitive to nicotine withdrawal, at least for changes in affect and locomotion.

Another study examining strain differences in the effects of nicotine withdrawal on anxiety-like phenotypes compared the effects of nicotine withdrawal on light-dark box performance, acoustic startle response, and PPI in C57BL/6J and DBA/2J mice (Jonkman et al., 2005; male; 12 or 24 mg/kg/day nicotine). C57BL/6J mice had increased anxiety-like behavior in the light-dark box during nicotine withdrawal whereas DBA/2J mice did not. Acoustic startle reflex and PPI were unchanged for either strain. These results demonstrate that nicotine withdrawal does not affect all behaviors and further support data that suggest the C57BL/6J have increased sensitivity to nicotine withdrawal.

Changes in cognition are a common nicotine withdrawal-associated symptom in smokers (Ashare et al., 2014). The effect of genetic background on differences in nicotine withdrawal-related changes in cognition were examined in a panel of eight inbred strains (Portugal et al., 2012; male; 3–12 mg/kg/day nicotine for 14 days). Changes in contextual and cued fear conditioning were assessed 24 h after cessation of nicotine treatment. C57BL/6J and C3H/HeJ mice showed the largest deficits and deficits at two doses for contextual fear conditioning. A/J mice also showed contextual fear conditioning deficits during withdrawal from the two higher doses. BALB/cByJ and DBA/1J mice only showed contextual fear conditioning deficits after withdrawal from the highest dose of nicotine. The other strains did not show nicotine withdrawal deficits in contextual fear conditioning. No strains showed nicotine withdrawal deficits in cued fear conditioning. As the dorsal hippocampus is critically involved in contextual fear conditioning but cued fear conditioning is not critically dependent on the dorsal hippocampus (Logue et al., 1997), these results suggest that the dorsal hippocampus may be sensitive to the effects of nicotine withdrawal and that these effects are mediated by genetics.

A follow up study examined the effects of nicotine withdrawal on contextual and cued fear conditioning and nicotine binding affinity for high-affinity (e.g., α4β2 nAChRs) nAChRs in the dorsal and ventral hippocampus (Wilkinson et al., 2013; male; 18 mg/kg/day nicotine for 12 days). Only C57BL/6NTac mice showed nicotine withdrawal associated deficits in contextual fear conditioning, and no strain showed deficits in cued fear conditioning. Chronic nicotine increased high-affinity nAChR binding in the dorsal but not ventral hippocampus of C57BL/6NTac mice. No significant changes were seen in 129S6/SvEvTac mice, while B6129SF1/Tac mice showed an increase in ventral hippocampus nAChR binding but no change in dorsal binding. Other work has suggested that nAChR upregulation in the dorsal hippocampus is related to the onset of cognitive nicotine withdrawal symptoms (Gould et al., 2014). Together, these results suggest that genetic differences in chronic nicotine induced high affinity nAChR upregulation specific to the dorsal hippocampus may contribute to susceptibility for nicotine withdrawal effects on hippocampus-dependent learning seen in C57BL/6 mice.

Examining the effects of nicotine withdrawal across inbred strains illustrates the complexity of nicotine withdrawal and also the important contribution of genetics. Withdrawal phenotype varied by genetic background. Expression of one nicotine withdrawal phenotype in a mouse strain did not always indicate expression of other nicotine withdrawal phenotypes. This demonstrates that it is unlikely that nicotine withdrawal is a solitary symptom but instead a host of symptoms for which expression is determined by different genetics. Consistent across studies, C57BL/6 mice were more susceptible to nicotine withdrawal symptoms. The withdrawal symptoms may be mediated by changes in nAChR receptor levels and potentially their function, as suggested by precipitated withdrawal studies.

3.7. Nicotine metabolism

Another factor that could contribute to nicotine sensitivity and vulnerability to nicotine addiction is genetic differences in nicotine metabolism. In humans, genetic differences in nicotine metabolism are associated with altered risk for nicotine dependence (Pianezza et al., 1998). A study comparing nicotine metabolism in C57BL/6 and DBA/2 mice found that elimination of nicotine was similar but higher concentrations of the primary metabolite cotinine were seen in DBA/2 mice than in C57BL/6 mice (Siu and Tyndale, 2007; male; 1 mg/kg nicotine s. c.; from Charles River). In addition, the metabolism of cotinine into 3-hydroxycotinine was decreased in DBA/2 mice. The faster metabolism of cotinine in C57BL/6 compared to DBA/2 mice could explain some of the increased nicotine addiction phenotypes found in C57BL/6 mice. Rubinstein et al. (2008) report that increased risk of nicotine addiction may be seen in faster nicotine metabolizing adolescent smokers.

Strain differences in nicotine metabolism were also examined in a larger panel of inbred mice (Portugal et al., 2012; male). For an acute dose of nicotine, 0.09 mg/kg, plasma nicotine levels were similar across strains but DBA/2J mice showed the highest levels of plasma cotinine followed by 129/SvEv and BALB/cByJ mice. Following treatment with 6.3 mg/kg/day of chronic nicotine for 12 days, strain differences were seen in both plasma nicotine and plasma cotinine. DBA/1J mice had the highest levels of plasma nicotine followed by A/J mice and for plasma cotinine levels associated with chronic nicotine treatment, 129/SvEv mice, followed by DBA/1J and then DBA/2 mice had the next highest levels. The rank order of plasma cotinine levels associated with chronic nicotine was 129SvEv > DBA/1J > DBA/2J > A/J > CBA/J > BALB/cByJ > C3H/HeJ > C57BL/6J.

These results further indicate that the strain that shows the greatest number of phenotypes associated with nicotine addiction (i.e., the C57BL/6J strain) is also the strain that tends to have faster metabolism of nicotine metabolites. This does not mean that nicotine metabolism is the causal factor for nicotine addiction but indicates that genetics associated with nicotine addiction phenotypes may covary with the genetics of nicotine metabolism and that the rate of nicotine metabolism may be one factor contributing to nicotine addiction.

3.8. Summary

In studies examining endophenotypes of nicotine addiction, strain differences along with sex differences are seen and these effects can be moderated by non-genetic factors. For oral nicotine consumption, C57BL/6 and C3H mice consumed more than DBA/2 and ST/b mice at 10 μg/ml, but C3H mice were low consumers at a 5 μg/ml dose of nicotine. This demonstrates that genetic effects are sensitive to environmental and non-genetic factors such as dose. Similarly, sex can influence nicotine consumption as female mice consumed more nicotine than male mice. That this sex effect was perhaps greater in the C57BL/6 strain also demonstrated that sex and genotype can interact to shape phenotype.

As C57BL/6 and DBA/2 mice differed in nicotine consumption, it might be expected that they differ in other nicotine phenotypes. For CPP and CPA, C57BL/6 had greater preference and aversion than DBA/2 mice and for acute nicotine effects on anxiety, DBA/2 mice had increased anxiety while C57BL/6 mice had reduced anxiety. The differences in nicotine associated anxiety effects could explain why DBA/2 mice consume less nicotine than C57BL/6 mice if consuming nicotine is anxiogenic for DBA/2 mice. Interestingly, DBA/2 mice showed greater somatic withdrawal symptoms than C57BL/6 mice but C57BL/6 had greater withdrawal associated anxiety than DBA/2 mice. This suggests there is not a universal withdrawal phenotype but instead different withdrawal phenotypes that have different genetic substrates. The results also suggest that the genetics of nicotine effects on anxiety are different for acute and chronic/withdrawal, such that acute was anxiogenic in DBA/2 mice whereas withdrawal from chronic nicotine was anxiogenic in C57BL/6 mice.

Differences in nicotine metabolism could explain some strain differences, though these effects may also be complex. C57BL/6 and DBA/2 mice showed similar levels of acute nicotine but differed in the metabolism of cotinine. DBA/2 mice were slower metabolizers of nicotine. For chronic nicotine, differences were seen for both nicotine and cotinine. DBA/2 mice were slower metabolizers of chronic nicotine and subsequent cotinine. These results demonstrate the metabolism has to be considered in context of drug administration procedures. In addition, it is possible that differences in metabolism between C57BL/6 and DBA/2 mice also contributes to differences in oral nicotine self-administration. That is, DBA/2 mice may consume less because they metabolize nicotine more slowly.

Few studies assessing the impact of age at time of nicotine administration on nicotine phenotypes have been reported. Age does influence nicotine consumption as early adolescent CD-1 mice consumed more nicotine than late adolescent mice; the genetics of this effect remains to be determined. Nicotine’s effects in adolescents in phenotypes of initial sensitivity, locomotor activity, anxiety, CPP, and CPA still need to be evaluated before comparisons can be made with adolescent phenotypes reported for alcohol consumption. The feasibility of assessing the effect of nicotine on various phenotypes is strong given the reported strain differences seen in adolescent mice in similar neuropsychiatric phenotypes (Eltokhi et al., 2020) and those reported for the effects of alcohol discussed earlier.

Overall, genetic background significantly contributes to phenotypic differences in nicotine effects. Across studies examining the acute effects of nicotine, C57BL/6 mice are consistently either the most sensitive or highly sensitive to the acute effects of nicotine. C3H and DBA/2 mice often, but not always, are less sensitive. For anxiety-like behaviors, acute nicotine has opposite effects in C57BL/6 mice and DBA/2 mice. For self-administration, C57BL/6 mice and DBA/2 again differ greatly. C57BL/6 mice consistently self-administered more nicotine than most strains while DBA/2 mice self-administer low levels of nicotine. When examined, the studies also found that female mice tended to self-administer more nicotine. Genetic differences in metabolism and in levels of CYP2A5 enzyme may account for some of the strain differences in nicotine self-administration but this may be specific for male mice, suggesting different genetic substrates are contributing to higher consumption in female mice. In addition to genetic differences in metabolism contributing to strain differences in nicotine self-administration, the studies in inbred mice demonstrated a contribution of genetic differences in nAChR subunits to phenotypic differences in nicotine self-administration. C57BL/6J mice have also been shown to be more sensitive to the rewarding effects of nicotine assessed in CPP paradigms. For nicotine withdrawal, severity of withdrawal varied by strain and withdrawal symptom measured. For somatic signs, DBA/2 had greater withdrawal symptoms but C57BL/6 mice had greater withdrawal symptoms for other withdrawal phenotypes. Different nAChRs may be selectively involved in specific nicotine withdrawal phenotypes with studies suggesting α4β2 nAChRs are associated with affective and cognitive withdrawal phenotypes. Differences in rate of metabolism and the underlying genetics could contribute to susceptibility to nicotine withdrawal phenotypes as C57BL/6 mice were fast metabolizers compared to DBA/2 mice, which were slow metabolizers, and C57BL/6 mice showed more withdrawal symptoms compared to DBA/2 mice. In summary, the described studies demonstrate that the use of inbred strains is instrumental in the characterization of nicotine phenotypes.

4. Combined alcohol and nicotine phenotypes

4.1. Background

Alcohol and nicotine are often co-administered (Falk et al., 2006). Anecdotally, it is common to hear stories of increased nicotine consumption when alcohol is consumed. This observation is supported by research as alcohol increases cigarette craving (Sayette et al., 2005) and tobacco use and nicotine dependence were positively associated with alcohol consumption rates (Falk et al., 2006). Potential reasons for the co-use and abuse of nicotine and ethanol are complex and include synergistic effects, amelioration of negative effects, and shared genetic and neural substrates, to name a few (Funk et al., 2006). Mouse models employing inbred strains of mice have identified some potential mechanisms contributing to co-use and abuse of alcohol and nicotine.

4.2. Phenotypes

One possible motivation for the co-use of nicotine and alcohol is that initially nicotine may reduce some negative symptoms associated with alcohol intake (Adams, 2017). Alcohol disrupts cognitive processes and this has been modeled in mice by examining the effects of alcohol on fear conditioning and plus-maze discriminative avoidance learning. Acute ethanol disrupts contextual and to a lesser extent cued fear conditioning, and ethanol disrupts plus-maze discriminative avoidance learning in C57BL/6J male and female mice (Davis et al., 2006; Gould, 2003; Gould and Lommock, 2003; Gulick and Gould, 2009b, 2011; Seemiller and Gould, 2021). Conversely, acute nicotine enhances contextual fear conditioning in C57BL/6J male mice (Davis et al., 2007; Gould and Higgins, 2003; Gould and Wehner, 1999). When administered together, nicotine prevents or ameliorates alcohol associated deficits in fear conditioning and plus-maze discriminative avoidance learning in C57BL/6J male mice (Gould and Lommock, 2003; Gulick and Gould, 2011). These results might suggest that the amelioration is simply an additive effect, that is nicotine enhances and ethanol disrupts so when administered together, the effects cancel out. However, acute nicotine enhances contextual fear conditioning through actions in the dorsal hippocampus (Davis et al., 2007) but the ameliorative effects of nicotine on alcohol-induced deficits in fear conditioning and in plus-maze discriminative avoidance learning in C57BL/6J mice were not ameliorated by direct infusion of nicotine into the hippocampus but by infusion of acute nicotine into the anterior cingulate cortex (Gulick and Gould, 2009a, 2011). Chronic nicotine, however, does not ameliorate ethanol-induced learning deficits in contextual fear conditioning, suggesting tolerance to the nicotine effect develops with chronic administration (Gulick and Gould, 2008). Together, these results suggest that at least for acute administration, nicotine may be co-administered with alcohol to ameliorate negative effects of alcohol and that the underlying neural processes are different than the processes involved in the effects of nicotine alone (Adams, 2017).

A study examining if nicotine and ethanol have additive effects examined the effects of nicotine on the hypnotic effects of ethanol in male C57BL/6J and DBA/2J mice (Slater et al., 2016). In both C57BL/6J and DBA/2J mice, acute nicotine enhanced the hypnotic effects of ethanol, as assessed by LORR, but the enhancement was greater in C57BL/6J mice. Nicotine did not alter blood alcohol concentrations in either strain. Tolerance for this effect of nicotine was seen, as well. These results suggest that genetics can modulate the ability of acute nicotine to enhance the hypnotic effects of ethanol as effects were greater in C57BL/6J mice. This ability of nicotine to enhance alcohol effects may be another reason alcohol and nicotine are co-administered.

In another study, sex and genotype differences in ethanol and nicotine intake were examined to test if ethanol and nicotine intake were genetically correlated (Li et al., 2005). In C57BL/6J mice, females consumed more ethanol and more nicotine, this sex effect was not seen for C3H/HeJ mice. In an F2 intercross, nicotine and ethanol consumption were correlated for both sexes and Chrna4, which encodes the α4 nAChR subunit, genotype influenced ethanol intake. Specifically, females homozygous for the C57BL/6J Chrna4 allele consumed less ethanol than females heterozygous or homozygous for the C3H/HeJ allele. This study demonstrated that genetics associated with the cholinergic system influence alcohol intake in a sex-specific manner and that alcohol and nicotine intake are positively correlated.

A strain survey of C57BL/6J, C3H/Ibg, BALB/Ibg, A/J, C58/J, and AKR/J male and female mice examined if ethanol increased nAChR ion influx induced by nicotine (Butt et al., 2003). A polymorphism in the gene encoding the α4 nAChR subunit was found that produces a variant with an alanine (A) for threonine (T) substitution at amino acid position 529. For mice with the A529 variant (A/J, AKR/J, C3H/Ibg), ethanol increased nicotine-evoked nAChR ion channel ion influx. These results demonstrated at the cellular level that alcohol and nicotine can have synergistic effects and that these effects vary by genotype.

Numerous studies have demonstrated additional links between nAChRs and ethanol consumption, and ethanol’s effects. In a study examining F2 mice created from a cross of C57BL/6J and DBA/2J mice, which represent high ethanol-preferring and ethanol-avoiding behavior respectively, polymorphisms in the Chrna5-Chrna3-Chrnb4 gene cluster were associated with ethanol preference (Symons et al., 2010). Mice with the C57BL/6J variant showed high ethanol preference; no sex differences were seen in ethanol preference. In addition, male transgenic mice with over expression of the Chrna5-Chrna3-Chrnb4 genes consumed less ethanol in a two-choice paradigm (Gallego et al., 2012), further suggesting that these nAChR-associated genes can contribute to the regulation of ethanol intake. In further support of this relationship, α4 nAChR subunit KO male mice had reduced ethanol-associated activation of ventral tegmental area dopaminergic neurons whereas ventral tegmental area dopaminergic neurons were activated with a subthreshold dose of ethanol in mice expressing a variant of the α4 nAChR hypersensitive to nicotine (Liu et al., 2013). The mice expressing a variant of the α4 nAChR hypersensitive to nicotine also developed ethanol CPP for a dose of ethanol subthreshold for CPP in wild-type mice. A study using male and female α6 nAChR subunit KO mice found that the KO mice had greater ethanol-induced LORR, which is used to assess the sedative effects of ethanol (Kamens et al., 2012). The study found that female mice consumed more ethanol than male mice but nAChR genotype did not modulate this effect. This suggests that the α6 nAChR subunit contributes to the suppression of some effects of alcohol.

Overall, these studies demonstrate that the systems and genetics mediating the effects of alcohol and nicotine do not work in isolation; instead, there are both crosstalk effects and there are common genetic substrates. These commonalities may influence increased co-consumption of nicotine and alcohol. The majority of studies have been conducted in adult subjects, but one study examined the relationship between nicotine and ethanol in adolescent mice. Adolescent female C57BL/6J mice were given 200 g/ml of nicotine for ten days via drinking water and ethanol intake was assessed on days 7–10 (Locker et al., 2016). Nicotine exposed mice consumed more ethanol and also had higher α4β2 nAChR levels in the frontal cortex. These results demonstrate similar relationships between nicotine and alcohol in adolescent mice as seen in adult mice and also suggest that nicotine and alcohol co-use in adolescence may contribute to increased risk for substance use and abuse during adolescence.

5. Conclusions

Given the complexity of substance abuse disorders, it is perhaps not surprising that the patterns of endophenotypes of alcohol and nicotine substance abuse disorders vary greatly across strains and across assays. This variability highlights the power of genetic models but also the intricacy of the interaction between genetic and environmental factors. Results across strains are influenced by non-genetic factors such as the method of delivery of the drug, the doses tested, and the pattern of exposure to the drug, to name a few, and by biological and behavioral effects such as sex, age, and changes in locomotion.

Overall, there are some similarities in strain effects for alcohol and nicotine. C57BL/6 mice were high consumers of both alcohol and nicotine whereas DBA/2 mice were low consumers of both. Similar patterns of consumption across nicotine and alcohol suggest similar genetic and neural substrates. In a study examining alcohol consumption in a cross of C57BL/6 and DBA/2 mice, polymorphisms in the Chrna5-Chrna3-Chrb4 gene cluster were associated with alcohol preference. Mice with the C57BL/6 variant consumed more alcohol. This directly demonstrated interaction between alcohol phenotypes and the cholinergic system and that this interaction is influenced by genetics. Cross talk between neural systems involved in the effects of alcohol and in the effects of nicotine may contribute to high co-use of both drugs, but even this effect is complex. Female C57BL/6 mice consumed more alcohol combined with nicotine than male mice, but this sex difference was not seen in C3H mice. The studies reviewed herein bear testimony to the complexity of alcohol and nicotine use disorders and the work that remains to be done. Continued parsing out of the interactive effects of genetic, age, sex, and environment on alcohol and nicotine use and co-use is critical for understanding these disorders and developing effective treatments and interventions that can be tailored for different genotypes.